Downloaded When Monitors Were Returned to the Study Centre at the Baseline and Week 6 Visits

i Health eefits of o-utritive food opoets

This thesis is presented for the degree of Doctor of Philosophy to The University of Western Australia

2015

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Rihard L. Prie, Professor Joatha M Hodgso, Assoiate Professor Deorah Kerr

Foreword i Health eefits of o-utritive food opoets

Foreword and contents

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Rihard L. Prie, Professor Joatha M Hodgso, Assoiate Professor Deorah Kerr

Foreord and contents ii SUMMARY

This thesis is presented as a series of epidemiological, intervention and review papers which aimed to explore the health benefits of non-nutritive food components.

Flavonoids and probiotics represent two diverse groups of non-nutritive food components, that are thought to provide benefits to human health.

Flavonoid compounds

Flavonoids are a diverse group of compounds which share a common flavan nuclear structure. Major classes of flavonoids include flavonols, proanthocyanidins, flavanols, flavones, flavanones and isoflavones. There is strong in vitro and intervention data suggesting health benefits of flavonoid intake. However, data from epidemiological studies are less clear. As such, relationships between flavonoid intake and health outcomes were explored in a population of elderly postmenopausal women.

We found that consumption of flavonols from both tea and non-tea sources was associated with reduced risk of atherosclerotic vascular disease mortality. Similarly, consumption of proanthocyanidins was associated with better renal function and lower risk of adverse renal outcomes. The beneficial associations of flavonoids appeared to extend beyond intakes of isolated flavonoid classes. We observed a beneficial association between total-flavonoid intake and reduced risk of all-cause mortality.

When exploring our ability to assess flavonoid intake in populations, the two major flavonoid food composition databases, the United States Department of Agriculture and

Phenol-Explorer databases, yielded highly correlated intake estimates for total- flavonoids, flavanols , flavanones and anthocyanidins. However, the poorer correlation between flavonol and flavone intake estimates derived from each database was likely

Foreord iii due to differences in United States Department of Agriculture (USDA) and Phenol-

Explorer (PE) methodologies.

Probiotic bacteria

Probiotics are defined as live microorganisms, which when administered in adequate amounts, confer a health benefit to the host. Despite a long history of safe use, incorporation of probiotics into current therapeutic guidelines, and strong mechanistic data, there is an identified lack of statistically significant data supporting beneficial effects of probiotics.

In order to investigate the health benefits of daily consumption of yoghurt and its probiotics, a 6-week randomised, controlled, double blind, factorial trial in 156 overweight men and women was implemented. We found that supplementation with probiotic yoghurt is more efficacious than probiotic capsules at improving fecal probiotic content. However, this trial, which supplemented with the strains

Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp. lactis BB12, did not improve type 2 diabetes and cardiovascular disease risk factors.

With the aim of summarising current evidence for the hypocholesterolaemic benefits of probiotic supplementation, a Cochrane approved systematic review and meta-analysis was performed. It was concluded that daily probiotic supplementation improves total cholesterol and low density lipoprotein cholesterol concentrations. Furthermore, in a population based study of elderly postmenopausal women, habitual high probiotic yoghurt consumption was associated with improved common carotid artery intima media thickness; a risk factor for atherosclerotic vascular disease.

Foreord i Concluding remark

Taken as a whole, the results of the studies presented in this thesis support the concept that the non-nutritive food components, probiotics and flavonoids, may play a role in improving cardio-metabolic health.

Foreord ACKNOWLEGDEMENTS

There are times in one’s life, where things go right, great minds come together, and friendships are formed. My PhD candidature has been one such time. The problem, however, arises articulating my gratitude to those who have contributed to this time. On major lesson I have learnt during my candidature is to stand on the shoulders of the giants before you. With this in mind I will turn to the great minds of the past in order to articulate the contribution my supervisors, collaborators and friends have made to the last three years of my life.

The first person I must acknowledge is my primary supervisor, Professor Richard Prince, who has not only inspired me to be a better scientist, but more importantly, a better person. Intelligence plus character-that is the goal of true education Martin Luther King Jr. Educating the mind without educating the heart is no education at all Aristotle If you would be a real seeker after truth, it is necessary that at least once in your life you doubt, as far as possible, all things René Descartes True education does not consist merely in the acquiring of a few facts of science, history, literature, or art, but in the development of character David O. McKay

My secondary supervisor was Professor Jonathan Hodsgon, who with his wisdom and sensibility has inspired, encouraged and enabled me to develop my own research path. The mediocre teacher tells. The good teacher explains. The superior teacher demonstrates. The great teacher inspires William Arthur Ward Do not train a child to learn by force or harshness; but direct them to it by what amuses their minds, so that you may be better able to discover with accuracy the peculiar bent of the genius of each Plato

My PhD candidature has provided me with the opportunity to collaborate with many talented people from a wide variety of backgrounds. Thank you to the following people for sharing their impressive knowledge, opening my eyes to new fields, and making my learning opportunities immensely pleasurable; Deborah Kerr (supervisor), Wai Lim (renal physiology), Kevin Croft (chemistry), Catherine Rawlinson (metabolomics), and Rongchang Yang (microbiology). Education: that which reveals to the wise, and conceals from the stupid, the vast limits of their knowledge Mark Twain Play is the highest form of research Albert Einstein Educated men are so impressive William Shakespeare

The traditional path of many Dietitians is to pursue purely clinical research. The following people have shared their talent, wisdom and time to teach me important lab based skills that will forever make me a better scientist: Catherine Rawlinson (GCMS), Rongchang Yang (DNA extraction and PCR), as well as Jenny Wang, Anthony Buzzai and Lawrence Liew (general lab skills). Give a bowl of rice to a man and you will feed him for a day. Teach him how to grow his own rice and you will save his life Confucius

Foreord i Thank you to Helena Moneta, Nancy Lin, Blagica Stojceski, Fiona Edmonds, Jenny Wang, Lawrence Liew and Anthony Buzzai, who have made my PhD candidature an immensely happy and successful experience. The making of friends, who are real friends, is the best token we have of a man's success in life Edward Everett Hale Happiness is the highest good Aristotle

Thank you to Benjemin Williams, for understanding me, accepting me and helping me to grow. A friend is one that knows you as you are, understands where you have been, accepts what you have become, and still, gently allows you to grow William Shakespeare

Thank you to my parents and brother, for supporting me. There is no school equal to a decent home and no teacher equal to a virtuous parent Mahatma Gandhi

And finally, I return once again to my fantastic supervisor Richard, who has become and will continue to be one of my greatest friends. Perhaps the most delightful friendships are those in which there is much agreement, much disputation, and yet more personal liking George Eliot

Foreord ii STATEMENT OF CANDIDATE CONTRIBUTION

I declare that this thesis contains published work and work prepared for publication, which have been co-authored. None of the material herein has been presented for the purpose of obtaining any other degree. The candidate contributions are detailed below.

For all publications arising from this thesis, in conjunction with co-authors, Miss Ivey devised the pre-specified data analysis protocol and undertook the manuscript review. Miss Ivey was solely responsible for implementation of all data analysis, literature review, manuscript preparation and manuscript submission.

In regards to the randomised controlled trial presented in Appendix A, Miss Ivey, in collaboration with Professors’ Prince and Hodgson, devised the pre-specified study protocol and reviewed grant submissions. In conjunction with Dr Rongchang Yang (Murdoch University, Perth, Australia), Miss Ivey performed the fecal DNA extractions, Nanodrop, quantitative PCR, and digital PCR. Miss Ivey was solely responsible for the literature review, implementation and management of the trial, direct supervision of staff and students, data analysis, and preparation of grant applications.

Student signature

Coordinating supervisor signature

Date 28 January 2014

Foreord iii GRANTS AWARDED DURING PhD CANDIDATURE

Sir Charles Gairdner Group Research Advisory Committee (2012) Prince RL, Thompson PL, Hodgson JM, Kerry L Ivey. Biochemical analysis and implementation of a randomised controlled trial of yoghurt and probiotics on the metabolic syndrome.

$46,182

Foreord ix AWARDS RECEIVED DURING PhD CANDIDATURE

2013 Travel Award Nutrition Society of Australia; Kent Town, Australia.

2013 Travel Award University of Western Australia; Perth, Australia.

2012 Best Poster Award Combined Biological Sciences Meeting; Perth, Australia.

2012 Young Investigator Award Sir Charles Gairdner Group Research Advisory Committee; Perth, Australia.

2011 University Postgraduate Award University of Western Australia; Perth, Australia.

2011 University of Western Australia Top-Up Scholarship University of Western Australia; Perth, Australia.

Foreord x PRESENTATIONS DURING PhD CANDIDATURE

Nutrition Society of Australia annual scientific meeting (2013, oral) Kerry L Ivey, Lewis JR, Prince RL, Hodgson JM. Probiotic bacteria and glycaemic control: a randomised controlled trial.

Nutrition Society of Australia annual scientific meeting (2013, poster) Kerry L Ivey, Lewis JR, Prince RL, Hodgson JM. Association of proanthocyanidin intake with renal function and clinical outcomes.

Combined Biological Sciences Meeting (2013, oral) Kerry L Ivey, Hodgson JM, Lewis JR, Kerr DA, Thompson PL, Prince RL. Effect of probiotic bacteria on glycaemic control.

Combined Biological Sciences Meeting (2013, poster) Kerry L Ivey, Lewis JR, Prince RL, Hodgson JM. Atherosclerotic Vascular Disease and flavonols: what’s the relationship?

Nutrition Society of Australia symposium (2013, oral) Kerry L Ivey, Hodgson JM, Lewis JR, Kerr DA, Thompson PL, Prince RL. Probiotics and glycaemia.

Nutrition Society of America annual scientific meeting (2013, poster) Kerry L Ivey, Hodgson JM, Lewis JR, Kerr DA, Thompson PL, Prince RL. The independent and additive effects of yoghurt and its probiotics on serum lipid profile: a meta-analysis and randomised controlled trial.

Nutrition Society of America annual scientific meeting (2013, poster) Kerry L Ivey, Lewis JR, Prince RL, Hodgson JM. Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women.

Combined Biological Sciences Meeting (2013, poster) Kerry L Ivey, Lewis JR, Lim WH, Lim EM, Hodgson JM, Prince RL. Association between proanthocyanidin intake, renal function and clinical outcomes in elderly.

Combined Biological Sciences Meeting (2011, oral) Kerry L Ivey, Lewis JR, Hodgson JM, Zhu K, Dhaliwal SS, Thompson PL, Prince RL. Association between yogurt, milk, and cheese consumption and common carotid artery intima-media thickness and cardiovascular disease risk factors in elderly women.

Foreord xi PUBLICATIONS AND MANUSCRIPTS SUBMITTED DURING PhD CANDIDATURE

Cochrane Database of Systematic Reviews (in preparation) Kerry L Ivey, Hodgson JM, Kim S, Woodman RJ, Prince RL. Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Full review).

British Journal of Nutrition (submitted) Kerry L Ivey, Hodgson JM, Kerr DA, Thompson PL, Prince RL. The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial

American Journal of Clinical Nutrition (submitted) Kerry L Ivey, Prince RL, Ryan U, Hodgson JM, Lin SY, Yang R. Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial.

American Journal of Clinical Nutrition (in preparation) Kerry L Ivey, Croft KD, Prince RL, Hodgson JM. Comparison of flavonoid intake assessment methods

American Journal of Clinical Nutrition (under review) Kerry L Ivey, Hodgson JM, Croft KD, Lewis JR, Prince RL. Flavonoid intake and all-cause mortality.

European Journal of Clinical Nutrition (in press) Kerry L Ivey, Hodgson JM, Kerr DA, Lewis JR, Thompson PL, Prince RL. The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial.

PlosOne (2013. 8:8, e71166) Kerry L Ivey, Lewis JR, Lim WH, Lim EM, Hodgson JM, Prince RL. Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women.

Cochrane Database of Systematic Reviews (2013, 3) Kerry L Ivey, Hodgson JM, Dhaliwal SS, Woodman RJ, Prince RL. Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Protocol).

British Journal of Nutrition (2013, 110, 1648-55) Kerry L Ivey, Lewis JR, Prince RL, Hodgson JM. Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women.

American Journal of Clinical Nutrition (2011, 94:1, 234-9)

Foreord xii Kerry L Ivey, Lewis JR, Hodgson JM, Zhu K, Dhaliwal SS, Thompson PL, Prince RL. Association between yogurt, milk, and cheese consumption and common carotid artery intima-media thickness and cardiovascular disease risk factors in elderly women.

Foreord xiii CHAPTER OVERVIEW

Chapter 1 Overview.

Chapter 2 Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women.

Chapter 3 Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women.

Chapter 4 Flavonoid intake and all-cause mortality.

Chapter 5 Comparison of flavonoid intake assessment methods.

Chapter 6 Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial.

Chapter 7 The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial.

Chapter 8 The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial.

Chapter 9 Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review).

Chapter10 Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women.

Chapter 11 Concluding statements.

Foreord xi TABLE OF CONTENTS

Page

Foreword

Summary……………………………………………………..…………………………. p ii

Acknowledgements………………………………………….………………………….. p iv

Statement of candidate contribution……………….…………………………………… p vi

Grants awarded during PhD candidature….……………………….…………………... p vii

Awards received during PhD candidature...………………………….………………… p viii

Presentations during PhD candidature.………………………………….……………... p ix

Publications and manuscripts submitted during PhD candidature……………………... p x

Chapter overview……...……………………………………………………….………. p xii

Table of contents…………………………..……………………………………..…….. p xiii

List of tables………………………..……………….……………………………..…… p xxviii

List of figures…………………………….….……………………………………..…... p xxxi

List of abbreviations…………………….....…………………………………………… p xxxv

Foreord x

Page

Chapter 1: overview…………………………….……………………………………….. Ch1, p 1

1.1 Foreword………………………………………………………………………………. Ch1, p 2

1.2 Non-nutritive food components Ch1, p 3

1.2.1 Definition of nutrients

1.2.2 Definition of non-nutritive food components

1.2.3 Types of non-nutritive food components

1.2.3 Health promoting properties of nutrients and whole foods

1.2.4 Health promoting properties of non-nutritive food components

1.3 Flavonoids……………………………………………………………………………… Ch1, p 6

1.3.1 Flavonoid structure

1.3.2 Dietary sources of flavonoids

1.3.3 Flavonoid absorption and metabolism in humans

1.3.4 Health benefits of flavonoid consumption

1.3.5 Investigating flavonoid-disease relationships

1.3.6 Current gaps in flavonoid knowledge

1.4 Probiotic bacteria……………………………………………………………………... Ch1, p 12

1.4.1 Definition of probiotics

1.4.2 Health benefits of probiotics

1.4.3 Assessing probiotic-disease relationships

1.4.4 Current gaps in probiotic knowledge

1.5 Aims and objectives…………………………………………………………………… Ch1, p 19

1.5.1 Overarching aims of thesis and relevance to field

1.5.2 Research questions and hypotheses explored in order to achieve aim

1.6 Structure of thesis……………………………………………………………………... Ch1, p 20

1.7 Chapter 1 references…………………………………………………………………... Ch1, p 22

1.8 Tables…………………………………………………………………………………... Ch1, p 30

Foreord xi

Page Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women………………………………………….. Ch2, p 1

2.1 Foreword……………………………………………………………………………… Ch2, p 2

2.2 Abstract………………………………………………………………………………. Ch2, p 3

2.3 Introduction………………………………………………………………………….. Ch2, p 4

2.4 Methods……………………………………………………………………………….. Ch2, p 5 2.4.1 Participants 2.4.2 Atherosclerotic vascular disease mortality 2.4.3 Baseline vascular disease risk assessment 2.4.4 Dietary assessment 2.4.5 Flavonoid intake 2.4.6 Statistics

2.5 Results………………………………………………………………………………… Ch2, p 11 2.5.1 All-source flavonoid consumption 2.5.2 Consumption of tea and non-tea flavonoids 2.5.3 Individual dietary flavonols 2.5.4 Dietary confounders 2.5.5 Sensitivity analysis

2.6 Discussion…………………………………………………………………………….. Ch2, p 14

2.7 Chapter 2 references………………………………………………………………… Ch2, p 17

2.8 Tables…………………………………………………………………………………. Ch2, p 20

Foreord xii

Page Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women..…………………………………………….…... Ch3, p 1

3.1 Foreword……………………………………………………………………………… Ch3, p 2 3.1.1 Chronic kidney disease 3.1.2 Proanthocyanidins, vascular health and chronic kidney disease

3.2 Abstract………………………………………………………………………………. Ch3, p 4

3.3 Introduction………………………………………………………………………….. Ch3, p 5

3.4 Subjects and methods.……………………………………………………………….. Ch3, p 7 3.4.1 Participants 3.4.2 Renal function 3.4.3 Estimated glomerular filtration rate 3.4.4 Renal disease events 3.4.5 Baseline chronic kidney disease risk assessment 3.4.6 Dietary assessment 3.4.7 Proanthocyanidin intake 3.4.8 Statistics

3.5 Results………………………………………………………………………………… Ch3, p 12 3.5.1 Renal function by serum cystatin C 3.5.2 Renal function by egfr using the CKD-EPI equation (creatinine and cystatin C) 3.5.3 Chronic kidney disease and clinical outcomes 3.5.4 Potential dietary confounders 3.5.5 Non-proanthocyanidin flavonoids

3.6 Discussion…………………………………………………………………………….. Ch3, p 15

3.7 Chapter 3 references………………………………………………………………… Ch3, p 18

3.8 Tables…………………………………………………………………………………. Ch3, p 23

Foreord xiii

Page Chapter 4: Flavonoid intake and all-cause mortality…………….………………. Ch4, p 1

4.1 Foreword……………………………………………………………………………… Ch4, p 2

4.2 Abstract………………………………………………………………………………. Ch4, p 3

4.3 Introduction………………………………………………………………………….. Ch4, p 4

4.4 Subjects and methods.……………………………………………………………….. Ch4, p 6

4.4.1 Participants 4.4.2 Mortality 4.4.3 Baseline risk assessment 4.4.4 Dietary assessment 4.4.5 Flavonoid intake 4.4.6 Statistics

4.5 Results………………………………………………………………………………… Ch4, p 11

4.5.1 Total-flavonoid intake and all-cause mortality 4.5.2 Flavonoid class intake and all-cause mortality 4.5.3 Total flavonoid intake and mortality from cancer and cardiovascular disease

4.6 Discussion…………………………………………………………………………….. Ch4, p 14

4.7 Chapter 4 references………………………………………………………………… Ch4, p 18

4.8 Tables and figures……………………………………………………………………. Ch4, p 23

Foreord xix

Page Chapter 5: Comparison of flavonoid intake assessment methods……………... Ch5, p 1

5.1 Foreword……………………………………………………………………………… Ch5, p 2

5.2 Abstract………………………………………………………………………………. Ch5, p 3

5.3 Introduction………………………………………………………………………….. Ch5, p 4

5.4 Methods: methodological comparison…...………………………………………….. Ch5, p 6 5.4.1 Methodological comparison 5.4.2 Food composition comparison 5.5 Methods: flavonoid intake comparison………………………….…………………. Ch5, p 8 5.5.1 Participants 5.5.2 Dietary assessment 5.5.3 Flavonoid intake 5.5.4 Statistics

5.6 Results: methodological comparison ……………………………………………… Ch5, p 10 5.6.1 Data sources 5.6.2 Chemistry of included compounds 5.6.3 Food content data 5.6.4 Imputed food composition

5.7 Results: flavonoid intake comparison………………………………………………. Ch5, p 13 5.7.1 Flavonoid intake estimates 5.7.2 Extent to which the USDA and PE intake estimates are linearly related 5.7.3 Relationship between level of disagreement and mean estimated intake 5.7.4 Extent to which the USDA and PE intake estimates classify participants as low, moderate or high flavonoid consumers

5.8 Discussion…………………………………………………………………………….. Ch5, p 15

5.9 Chapter 5 references………………………………………………………………… Ch5, p 20

5.10 Tables and figures……………………………………………………………………. Ch5, p 24

5.11 Supplementary tables…..……………………………………………………………. Ch5, p 34

Foreord xx

Page Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial……………………....……………………………………... Ch6, p 1

6.1 Foreword……………………………………………………………………………… Ch6, p 2

6.2 Abstract………………………………………………………………………………. Ch6, p 3

6.3 Introduction………………………………………………………………………….. Ch6, p 4

6.4 Methods……………………………………………………………………………….. Ch6, p 6 6.4.1 Subjects 6.4.2 Intervention 6.4.3 Compliance 6.4.4 Baseline measurements 6.4.5 Measurements of bacterial count 6.4.6 Fecal DNA extraction 6.4.7 Quantitation of bacteria numbers in probiotic capsules using droplet digital PCR 6.4.8 Quantitation of bacteria numbers in faecal samples using conventional qPCR 6.4.9 Blinding and statistical analysis

6.5 Results………………………………………………………………………………… Ch6, p 11 6.5.1 Baseline cohort characteristics 6.5.2 Week 6 cohort characteristics 6.5.3 Effect of intervention on Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis BB12

6.6 Discussion…………………………………………………………………………….. Ch6, p 13

6.7 Chapter 6 references………………………………………………………………… Ch6, p 17

6.8 Tables…………………………………………………………………………………. Ch6, p 21

Foreord xxi

Page Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial.……………………. Ch7, p 1

7.1 Foreword……………………………………………………………………………… Ch7, p 2

7.2 Abstract………………………………………………………………………………. Ch7, p 3

7.3 Introduction………………………………………………………………………….. Ch7, p 4

7.4 Subjects and methods..……………………………………………………………….. Ch7, p 6 7.4.1 Subjects 7.4.2 Intervention 7.4.3 Compliance 7.4.4 Baseline measurements 7.4.5 Measurements of glycaemic control 7.4.6 Blinding and statistical analysis

7.5 Results………………………………………………………………………………… Ch7, p 10 7.5.1 Participant characteristics and compliance 7.5.2 Effect of intervention on biomarkers of glycaemic control 7.5.3 Exploratory analyses

7.6 Discussion…………………………………………………………………………….. Ch7, p 12

7.7 Chapter 7 references………………………………………………………………… Ch7, p 15

7.8 Tables…………………………………………………………………………………. Ch7, p 19

Foreord xxii

Page Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial.………………………………... Ch8, p 1

8.1 Foreword……………………………………………………………………………… Ch8, p 2

8.2 Abstract………………………………………………………………………………. Ch8, p 3

8.3 Introduction………………………………………………………………………….. Ch8, p 4

8.4 Subjects and methods…..…………………………………………………………….. Ch8, p 6 8.4.1 Subjects 8.4.2 Study design 8.4.3 Baseline and week 6 measurements 8.4.4 Home blood pressure monitoring 8.4.5 Measurements of serum lipid profile 8.4.6 Statistical analysis

8.5 Results………………………………………………………………………………… Ch8, p 10 8.5.1 Baseline cohort characteristics 8.5.2 Physical activity, energy and nutrient intakes 8.5.3 Effect of probiotics on blood pressure 8.5.4 Effect of probiotics on serum lipid profile 8.5.5 Secondary analysis

8.6 Discussion…………………………………………………………………………….. Ch8, p 12

8.7 Chapter 8 references………………………………………………………………… Ch8, p 14

8.8 Tables…………………………………………………………………………………. Ch8, p 17

Foreord xxiii

Page Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review)…….……….. Ch9, p 1

9.1 Foreword……………………………………………………………………………… Ch9, p 2

9.2 Abstract………………………………………………………………………………. Ch9, p 3

9.3 Plain language summary…………….……………………………………………….. Ch9, p 5 9.3.1 Research question 9.3.2 Background 9.3.3 Study characteristics 9.3.4 Key results 9.3.5 Quality of the evidence

9.4 Background……………...…………………………………………………………… Ch9, p 7 9.4.1 Description of the condition 9.4.2 Description of the intervention 9.4.3 Adverse effects of the intervention 9.4.4 How the intervention might work 9.4.5 Why it is important to do this review

9.5 Objectives…………………………….……………………………………………….. Ch9, p 11

9.6 Methods: criteria for considering studies in this review…...……………………… Ch9, p 12 9.6.1 Types of studies 9.6.2 Types of participants 9.6.3 Types of interventions 9.6.4 Types of outcome measures

9.7 Methods: search methods for identification of studies…………….……………….. Ch9, p 16 9.7.1 Electronic searches 9.7.2 Searching other resources

9.8 Methods: data collection and analysis….…………………………………………… Ch9, p 17 9.8.1 Selection of studies 9.8.2 Data extraction and management 9.8.3 Assessment of risk of bias in included studies 9.8.4 Measures of treatment effect 9.8.5 Unit of analysis issues 9.8.6 Dealing with missing data 9.8.7 Assessment of heterogeneity 9.8.8 Assessment of reporting bias 9.8.9 Data synthesis

Foreord xxi

9.8.10 Subgroup analysis and investigation of heterogeneity 9.8.11 Sensitivity analysis 9.9 Main results: results of the search….……………………………………………….. Ch9, p 23 9.9.1 Search results 9.9.2 Excluded studies

9.10 Main results: included studies.……………………………………………………… Ch9, p 24 9.10.1 Source of data 9.10.2 Overview of study populations 9.10.3 Settings 9.10.4 Participants 9.10.5 Study design 9.10.6 Comparisons 9.10.7 Interventions 9.10.8 Outcomes 9.10.9 Risk of bias in included studies

9.11 Main results: effects of interventions on primary outcomes………………………. Ch9, p 31 9.11.1 Adverse events

9.12 Main results: effects of interventions on secondary outcomes……………………. Ch9, p 32 9.12.1 Total cholesterol 9.12.2 Low density lipoprotein cholesterol 9.12.3 High density lipoprotein cholesterol 9.12.4 Triglyceride

9.13 Discussion…………………………….……………………………………………….. Ch9, p 41 9.13.1 Summary of main results 9.13.2 Overall completeness and applicability of evidence 9.13.3 Quality of the evidence 9.13.4 Potential biases in the review process 9.13.5 Authors’ conclusions

9.14 Tables……..……………...…………………………………………………………… Ch9, p 44

9.15 Figures………………….…………….……………………………………………….. Ch9, p 73

Foreord xx

9.16 Appendices…..…………...…………………………………………………………… Ch9, p 113 9.16.1 Appendix A: search strategies 3.16.2 Appendix B: description of interventions 9.16.3 Appendix C: baseline characteristics (1) 9.16.4 Appendix D: baseline characteristics (2) 9.16.5 Appendix E: matrix of study endpoints (publications) 9.16.6 Appendix F: examination of outcome reporting bias 9.16.7 Appendix G: outcome assessment 9.16.8 Appendix H: adverse events

9.17 Chapter 9 studies and references...... ……………………………………………….. Ch9, p 150

Foreord xxi

Page Chapter 10: Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women…………………………………………………………. Ch10, p 1

10.1 Foreword……………………………………………………………………………… Ch10, p 2

10.2 Abstract………………………………………………………………………………. Ch10, p 3

10.3 Introduction………………………………………………………………………….. Ch10, p 4

10.4 Subjects and methods….…………………………………………………………….. Ch10, p 5 10.4.1 Participants 10.4.2 Baseline vascular disease risk assessment 10.4.3 Assessment of dairy consumption 10.4.4 Clinical measurements 10.4.5 Statistics

10.5 Results……………………………………………………………………………… Ch10, p 9 10.5.1 Dairy product intake and CCA-IMT 10.5.2 Yogurt and CCA-IMT 10.5.3 Carotid atherosclerotic plaques

10.6 Discussion…………………………………………………………………………….. Ch10, p 11

10.7 Chapter 10 references……………………………………………………………… Ch10, p 14

10.8 Tables and figures…………………………………………………………………… Ch10, p 16

Foreord xxii

Page

Chapter 11: Concluding statements………………………………………………… Ch11, p 1

11.1 Foreword……………………………………………………………………………… Ch11, p 2

11.2 Findings in relation to research questions and thesis aim…………………………. Ch11, p 3

11.3 Implications of thesis findings for research………………………………………… Ch11, p 5

11.4 Implications of thesis findings for clinical practice…..……………………………. Ch11, p 7

11.5 Chapter 11 references……..……………………….………………………………… Ch11, p 8

11.6 Tables…………………………………………………………………………………. Ch11, p 10

Foreord xxiii

Page Appendix A: Randomised controlled trial of yoghurt and its probiotics: study protocol……………………………………………………………………………... A, p 1

A.1 Foreword……………………………………………………………………………… A, p 2

A.2 Study design rationale………………………………………………………………. A, p 3 A.2.1 Recruitment procedure A.2.2 Randomisation procedure A.2.3 Study design A.2.4 Test articles

A.3 Participants….……………………………………………………………………… A, p 6 A.3.1 Participants A.3.2 Sample size A.3.3 Recruitment

A.4 Study design …………………………………………………………………………. A, p 8 A.4.1 Screening assessment A.4.2 Lead in period A.4.3 Randomisation A.4.4 Intervention period

A.5 Primary outcome variables………………………………………………………… A, p 13 A.5.1 Biochemical measurements A.5.2 Blood pressure measurements

A.6 Secondary outcome variables presented in thesis…………………………………. A, p 15 A.6.1 Fecal bacterial content

A.7 Secondary outcomes not presented in this thesis ….……………………………… A, p 18 A.7.1 Metabolomic profile A.7.2 Gastrointestinal symptoms A.7.3 Functional health and wellbeing, physical and mental health A.7.4 Quality of life A.7.5 Biochemical measurements related to bone metabolism and inflammation A.7.6 Inflammatory markers A.7.7 Fasting urine sample

A.8 Measurements of exposure and confounding effects………………………………. A, p 20 A.8.1 Anthropometry A.8.2 Dietary intake A.8.3 Physical activity A.8.4 Compliance A.8.5 Medication use A.8.6 Co-morbidities, past medical history and adverse events

A.9 Appendix A references……...………………………………………………………… A, p 23

Foreord xxix LIST OF TABLES

Page

Chapter 1…………………………………………………………………………………

Table 1: Thesis research questions and hypotheses used to explore them Ch1, p 30

Chapter 2…………………………………………………………………………………

Table 1: Baseline characteristics of the cohort stratified by atherosclerotic vascular disease mortality Ch2, p 20

Table 2: Baseline flavonoid class intake according to dietary source Ch2, p 21

Table 3: Relationship between total flavonoid class intake groups and 5-year atherosclerotic vascular disease mortality Ch2, p 22

Table 4: Relationship between tea and non-tea flavonol intake groups and 5-year atherosclerotic vascular disease mortality Ch2, p 23

Chapter 3…………………………………………………………………………………

Table 1: Baseline, lifestyle and cardiovascular risk factors by tertiles of proanthocyanidin intake Ch3, p 23

Table 2: Baseline cystatin C concentration according to groups of proanthocyanidin intake Ch3, p 24

Table 3: Relationship between proanthocyanidin intake and 5-year hospitalisation or death renal failure events Ch3, p 25

Chapter 4…………………………………………………………………………………

Table 1: Baseline characteristics of the cohort stratified by total-flavonoid intake group. Ch4, p 23

Table 2: Association of total-flavonoid intake group and risk of all-cause mortality Ch4, p 24

Table 3: Association of total-flavonoid group and risk of mortality from cardiovascular disease and cancer Ch4, p 25

Foreord xxx

Chapter 5…………………………………………………………………………………

Table 1: Structure and chemical name of flavonoid classes included in this review Ch5, p 24

Table 2: Relative flavonoid class concentrations in foods, imputed from United States Department of Agriculture (USDA) and Phenol-Explorer (PE) data Ch5, p 25

Table 3: Daily consumption of total-flavonoid and flavonoid class, as estimated with the United States Department of Agriculture and Phenol-Explorer databases Ch5, p 26

Table 4: Tertiles of Phenol-Explorer (PE) flavonoid intake expressed against United States Department of Agriculture (USDA) flavonoid intake tertiles Ch5, p 27

Supplementary Table 1: Flavonol compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases Ch5, p 34

Supplementary Table 2: Flavanol compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases Ch5, p 36

Supplementary Table 3: Flavone compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases Ch5, p 38

Supplementary Table 4: Flavanone compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases Ch5, p 40

Supplementary Table 5: Anthocyanidin compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases Ch5, p 41

Chapter 6…………………………………………………………………………………

Table 1: Baseline characteristics of cohort, stratified by treatment group Ch6, p 21

Table 2: Treatment group summary statistics of bacterial count present in 1 g fecal DNA Ch6, p 22

Table 3: Main effect model of probiotic yoghurt supplementation on absolute bacterial count at 6-weeks Ch6, p 23

Chapter 7…………………………………………………………………………………

Table 1: Baseline characteristics by treatment group Ch7, p 19

Table 2: Treatment group summary statistics of glycaemic parameters at baseline and 6-weeks Ch7, p 20

Table 3: Main effect model of probiotic yoghurt supplementation on biomarkers of glycaemic control at 6-weeks Ch7, p 21

Table 4: Main effect model of probiotic capsule supplementation on biomarkers of glycaemic control at week 6 Ch7, p 22

Foreord xxxi

Chapter 8…………………………………………………………………………………

Table 1: Baseline characteristics of participants by treatment group Ch8, p 17

Table 2: Seven day home blood pressure parameters stratified by treatment group Ch8, p 18

Table 3: Serum lipid concentrations stratified by treatment group Ch8, p 19

Table 4: Main effect model of yoghurt and probiotic supplementation on home blood pressure parameters Ch8, p 20

Table 5: Main effect model of yoghurt and probiotic supplementation on serum lipid parameters Ch8, p 21

Chapter 9…………………………………………………………………………………

Table 1: Characteristics of included studies Ch9, p 44

Table 2: Characteristics of excluded studies Ch9, p 67

Table 3: Summary of findings table Ch9, p 68

Table 4: Study characteristics Ch9, p 69

Table 5: Genus and species of intervention probiotic bacteria Ch9, p 72

Chapter 10………………………………………………………………………………

Table 1: Baseline demographics and cardiovascular risk factors of participants Ch10, p 16

Table 2: Relationship of baseline dairy intake (g/day) to cardiovascular disease risk factors Ch10, p 17

Table 3: Baseline, lifestyle and cardiovascular risk factors by yogurt consumption group Ch10, p 18

Table 4: Baseline dietary factors by yogurt consumption group Ch10, p 19

Table 5: Common carotid artery intima-media thickness (mm) according to yogurt consumption group Ch10, p 20

Chapter 11………………………………………………………………………………

Table 1: Thesis research questions and hypotheses used to explore them Ch11, p 10

Appendix A………………………………………………………………………………

Table 1: Outline of intervention groups in the full factorial study design A, p 9

Foreord xxxii LIST OF FIGURES

Page

Chapter 1………………………………………………………………………………

Figure 1: Chemical structure of the flavan (2-phenylchroman) nucleus of flavonoid molecules Ch1, p 6

Chapter 4…………………………………………………………………………………

Figure 1: Proportional reduction in 5-year all-cause mortality incidence between low, moderate, and high total-flavonoid consumption as estimated from the United States Department of Agriculture database Ch4, p 26

Figure 2: Proportional reduction in 5-year all-cause mortality incidence between low, moderate, and high total-flavonoid consumption estimated from the Phenol-Explorer database Ch4, p 27

Chapter 5…………………………………………………………………………………

Figure 1: chemical structure of the flavan (2-phenylchroman) nucleus of flavonoid molecules. Ch5, p 29

Figure 2: Level of agreement between United States Department of Agriculture (USDA) and Phenol-Explorer (PE) total-flavonoid and flavonoid class intake estimates Ch5, p 30

Figure 3: Relationship between level of disagreement and mean estimated total-flavonoid and flavonoid class intake Ch5, p 32

Foreord xxxiii

Chapter 9…………………………………………………………………………………

Figure 1: Flow diagram of included studies Ch9, p 73

Figure 2: Risk of bias graph Ch9, p 74

Figure 3: Risk of bias summary Ch9, p 75

Figure 4 (analysis 1.8): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.8 Mild adverse events Ch9, p 76

Figure 5 (analysis 1.1): Forest plot of comparison: 1 Probiotic versus placebo - whole data set, outcome: 1.1 Total cholesterol Ch9, p 77

Figure 6 (analysis 1.1): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.1 Total cholesterol Ch9, p 78

Figure 7 (analysis 1.2): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.3 High density lipoprotein cholesterol Ch9, p 79

Figure 8 (analysis 1.2): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.2 Total cholesterol - excluding outliers Ch9, p 80

Figure 9 (analysis 2.1): Forest plot of comparison: 2 Probiotic versus placebo - sensitivity analysis excluding unpublished studies, outcome: 2.1 Total cholesterol Ch9, p 81

Figure 11 (analysis 4.1): Forest plot of comparison: 4 Probiotic versus placebo - sensitivity analysis based on number of participants included in analysis, outcome: 4.1 Total cholesterol Ch9, p 82

Figure 12 (analysis 5.1): Forest plot of comparison: 5 Probiotic versus placebo - sensitivity analysis using Random Effects model, outcome: 5.1 Total cholesterol. Ch9, p 83

Figure 13 (analysis 6.1): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.1 Total cholesterol. Ch9, p 84

Figure 14 (analysis 7.1): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.1 Total cholesterol. Ch9, p 85

Figure 15 (analysis 7.2): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.2 Total cholesterol - Borderline high studies split into two groups (above and below the median baseline total-cholesterol). Ch9, p 86

Figure 16 (analysis 1.3): Forest plot of comparison: 1 Probiotic versus placebo - whole data set, outcome: 1.2 Low density lipoprotein cholesterol. Ch9, p 87

Figure 17 (analysis 1.3): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.3 Low density lipoprotein cholesterol. Ch9, p 88

Figure 18 (analysis 1.4): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.4 Low density lipoprotein cholesterol - excluding outliers. Ch9, p 89

Figure 19 (analysis 1.4): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.4 Low density lipoprotein cholesterol - excluding outliers. Ch9, p 90

Foreord xxxi

Figure 20 (analysis 2.2): Forest plot of comparison: 2 Probiotic versus placebo - sensitivity analysis excluding unpublished studies, outcome: 2.2 Low density lipoprotein cholesterol. Ch9, p 91

Figure 22 (analysis 4.2): Forest plot of comparison: 4 Probiotic versus placebo - sensitivity analysis based on number of participants included in analysis, outcome: 4.2 Low density lipoprotein cholesterol. Ch9, p 92

Figure 23 (analysis 5.2): Forest plot of comparison: 5 Probiotic versus placebo - sensitivity analysis using Random Effects model, outcome: 5.2 Low density lipoprotein cholesterol. Ch9, p 93

Figure 24 (analysis 6.2): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.2 Low density lipoprotein cholesterol. Ch9, p 94

Figure 25 (analysis 7.3): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.3 Low density lipoprotein cholesterol. Ch9, p 95

Figure 26 (analysis 7.4): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.4 Low density lipoprotein cholesterol - Borderline high studies split into two groups (above and below the median baseline total-cholesterol). Ch9, p 96

Figure 27 (analysis 1.5): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.5 High density lipoprotein cholesterol. Ch9, p 97

Figure 28 (analysis 1.5): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.5 High density lipoprotein cholesterol. Ch9, p 98

Figure 29 (analysis 1.6): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.6 High density lipoprotein cholesterol - excluding outliers. Ch9, p 99

Figure 30 (analysis 1.6): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.6 High density lipoprotein cholesterol - excluding outliers. Ch9, p 100

Figure 31 (analysis 2.3): Forest plot of comparison: 2 Probiotic versus placebo - sensitivity analysis excluding unpublished studies, outcome: 2.3 High density lipoprotein cholesterol. Ch9, p 101

Figure 33 (analysis 4.3): Forest plot of comparison: 4 Probiotic versus placebo - sensitivity analysis based on number of participants included in analysis, outcome: 4.3 High density lipoprotein cholesterol. Ch9, p 102

Figure 34 (analysis 5.3): Forest plot of comparison: 5 Probiotic versus placebo - sensitivity analysis using Random Effects model, outcome: 5.3 High density lipoprotein cholesterol. Ch9, p 103

Figure 35 (analysis 6.3): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.3 High density lipoprotein cholesterol. Ch9, p 104

Figure 36 (analysis 7.5): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.5 High density lipoprotein cholesterol. Ch9, p 105

Figure 37 (analysis 1.7): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.7 Triglyceride. Ch9, p 106

Foreord xxx

Figure 38 (analysis 1.7): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.7 Triglyceride. Ch9, p 107

Figure 39 (analysis 2.4): Forest plot of comparison: 2 Probiotic versus placebo - sensitivity analysis excluding unpublished studies, outcome: 2.4 Triglyceride. Ch9, p 108

Figure 41 (analysis 4.4): Forest plot of comparison: 4 Probiotic versus placebo - sensitivity analysis based on number of participants included in analysis, outcome: 4.4 Triglyceride. Ch9, p 109

Figure 42 (analysis 5.4): Forest plot of comparison: 5 Probiotic versus placebo - sensitivity analysis using Random Effects model, outcome: 5.4 Triglyceride. Ch9, p 110

Figure 43 (analysis 6.4): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.4 Triglyceride. Ch9, p 111

Figure 44 (analysis 7.6): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.6 Triglyceride. Ch9, p 112

Chapter 10………………………………………………………………………………

Figure 1: Proportional reduction in common carotid artery intima-media thickness (CCA-IMT) between low, moderate and high yogurt consumption Ch10, p 21

Appendix A………………………………………………………………………………

Table 1: Outline of intervention groups in the full factorial study design A, p 8

Foreord xxxi LIST OF ABBREVIATIONS

ANCOVA Analysis of covariance ANOVA Analysis of variance ASVD Atherosclerotic vascular disease BB12 Bifidobacterium animalis subsp. lactis BB12 BMI Body mass index

CAIFOS Calcium Intake Fracture Outcome Study CARES Calcium Intake Fracture Outcome Age Related Extension Study CCA-IMT Common carotid artery intima-media thickness CI Confidence interval

CKD Chronic kidney disease CVD Cardiovascular disease DBP Diastolic blood pressure ddPCR Digital polymerase chain reaction DG Droplet generator DNA Deoxyribonucleic acid eGFR Estimated glomerular filtration rate FFQ Food frequency questionnaire

HbA1c Glycated haemoglobin HDLC High density lipoprotein cholesterol HMG-CoA 3-hydroxy-3-methylglutaryl-coenzyme A HOMA-IR Homeostasis Model Assessment of Insulin Resistance HR Hazard ratio

ICD-9-CM International Classification of Diseases, Injuries and Causes of Death: Clinical Modification ICD-10-AM International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Australian Modification IQR Interquartile range LJLa1 Lactobacillus johnsonii La1 LA5 Lactobacillus acidophilus La5

LDLC Low density lipoprotein cholesterol LSD Least significant differences

Foreord xxxii

MET Metabolic equivalent of task OR Odds ratio PE Phenol-Explorer qPCR Quantitative polymerase chain reaction SBP Systolic blood pressure Scys Serum cystatin C Scr Serum creatinine SD Standard deviation

SEM Standard error of the mean SPSS Statistical Package for the Social Sciences tgl Triglyceride UHT Ultra high temperature WADLS Western Australian Data Linkage System WHO World Health Organisation

Foreord Chapter 1: Page 1 Health eefits of o-utritive food opoets

Chapter 1: Overview

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Rihard L. Prie, Professor Joatha M Hodgso, Assoiate Professor Deorah Kerr

Chapter 1: Overview Chapter 1: Page 2

1.1 FOREWORD

The purpose of this overview chapter is to briefly outline the scientific and intellectual basis of my approach for this PhD thesis. The aim of this thesis was to explore the health benefits of consuming non-nutritive food components, specifically flavonoids and probiotics. As such, this chapter provides a brief description of non-nutritive food components, and then highlights current knowledge, and gaps in knowledge, regarding health benefits of flavonoid compounds and probiotic bacteria. This chapter then concludes with the presentation of the thesis aim and research questions, and an overview of thesis chapters.

Chapter 1: Overview Chapter 1: Page 3 1.2 NON-NUTRITIVE FOOD COMPONENTS

1.2.1 Definition of nutrients

Nutrients are food substances that are essential for growth, reproduction and the maintenance of life. A current definition of nutrients is as follows: “a fully characterized (physical, chemical, physiological) constituent of a diet, natural or designed, that serves as a significant energy yielding substrate, or a precursor for the synthesis of macromolecules or of other components needed for normal cell differentiation, growth, renewal, repair, defense and/or maintenance or a required signaling molecule, cofactor or determinant of normal molecular structure/function and/or a promoter of cell and organ integrity” (1). As such, the term nutrient encompasses both macronutrients, including carbohydrate, protein and fat, and micronutrients comprising of vitamins and minerals.

1.2.2 Definition of non-nutritive food components

In addition to nutrients, foods and beverages also contain factors, which are not essential for growth, reproduction or the maintenance of life. These factors are commonly referred to as non-nutritive food components (2). Following the discovery of non-nutritive food components with biological activities, recent research has focused on exploring the function of food components beyond and in addition to the supply of nutrients to maintain life.

1.2.3 Types of non-nutritive food components

Some non-nutritive food components are thought to have detrimental effects on human health. For example, the heavy metal lead does not feature in any known essential metabolic pathways in humans, and therefore is not considered a nutrient. However, as a result of food system contamination, lead appears in many food products throughout the

Chapter 1: Overview Chapter 1: Page 4 world. Overconsumption of this non-nutritive food component can result in lead toxicity with detrimental effects on the hemopoietic central nervous systems (3).

On the other hand, some non-nutritive food components do not appear to have any effect on human health, such as many of the artificial flavouring agents currently used in food manufacturing. Artificial compounds are often added to in food processing to enhance the flavour and palatability of the processed food products. The Joint

FAO/WHO Expert Committee on Food Additives concluded that many flavouring agents are innocuous to human metabolism (4).

There are also many non-nutritive food components that are believed to provide a beneficial effect to human health. The topic of this thesis is to explore health benefits of two classes of non-nutritive food components; flavonoid compounds and probiotic bacteria. Flavonoids and probiotics represent two diverse groups of food constituents that are not essential for human growth or survival, and thus defined as non-nutritive food components, that are thought to provide health benefits to humans (5, 6).

1.2.3 Health promoting properties of nutrients and whole foods

It is established that diet plays a major role in determining risk, survival and outcomes of preventable diseases such as cardiovascular disease (7), cancer (8) and chronic kidney disease (9). As chronic diseases are the leading cause of death in the world (10), there is considerable interest in identifying optimal diets, and components of diets, that can prevent or reduce chronic disease severity.

To date, dietary guidelines have focused on identifying particular nutrients and nutrient intake profiles which reduce risk of chronic disease (11, 12). Current nutrient reference values are based on ensuring nutrient requirements for growth and survival are met, and nutrient intake patterns for chronic disease prevention are followed (13, 14). However,

Chapter 1: Overview Chapter 1: Page 5 little direct attention is paid to non-nutritive food components in dietary guidelines for health or chronic disease prevention.

1.2.4 Health promoting properties of non-nutritive food components

Foods rich in flavonoids, such as tea, chocolate, wine, fruit, and vegetables elicit health benefits which do not appear to be explained simply by their nutrient profile (15-20).

Similarly, there is strong evidence to suggest that fermented milks containing probiotic bacteria also have beneficial effects on human health (21). Therefore, this thesis will explore in further detail the health benefits of flavonoids and probiotics.

Chapter 1: Overview Chapter 1: Page 6 1.3 FLAVONOIDS

1.3.1 Flavonoid structure

Flavonoids represent a diverse group of water soluble polyphenolic compounds derived from a flavan (2-phenylchroman ) nucleus (Figure 1). Derivations of this basic structure arise due to alterations in the 2(3) carbon-carbon bond, the formation of a ketone at carbon 4, and hydroxylation of carbons at various locations on the flavan backbone

(22). It is these derivations that allow the over 4000 flavonoid molecules to be grouped into one of five main flavonoid classes; Flavonols, Flavanols, Flavones, Flavanones, and Anthocyanidins (23-25). A sixth group, the isoflavones are often considered a class of flavonoid compounds despite having an 3-phenylchroman, rather than 2- phenylchroman, nucleus (26).

’ 4’ ’

8 5’ 7 6’

3 6

4 5

Chapter 1, Figure 1: Chemical structure of the flavan (2-phenylchroman) nucleus of flavonoid molecules

Figure sourced and adapted from Phenol-Explorer electronic database (27).

Each flavonoid class is comprised of numerous individual compounds with varying degrees of polymerisation, glycosylation, hydroxylation and esterification. The chemical structure of the flavan nucleus together with the numerous functional groups to which it is attached explains the numerous biological roles of flavonoids. In plants, flavonoids are involved in vital roles such as 1) pigmentation for vector mediated pollination (28-30); 2) pollen development, germination and pollen tube development

Chapter 1: Overview Chapter 1: Page 7 (31); 3) absorption of light waves in the ultraviolet spectrum, protecting from radiation induced DNA damage (32-34); 4) the promotion of a symbiotic relationship between nitrogen fixing bacteria and legumes(35); and 5)protection from environmental stress

(36) and disease resistance (37).

The biological roles of flavonoids are not limited to the plants from which they originate. Following consumption, flavonoid compounds perform many metabolic functions in humans.

1.3.2 Dietary sources of flavonoids

The flavonoid content of foods is complex. Different food products contain different amounts of different flavonoid compounds with varying degrees of hydroxylation and glycosylation. Furthermore, different varieties of the same food item can also have vastly different flavonoid content and composition (38). In the human diet, important sources of flavonoids include tea (both green and black), red wine, fruits, vegetables and chocolate (24, 39-41).

The complexity of the flavonoid content of foods is further exacerbated by the effect of variety, growing conditions, maturation, processing, and storage on flavonoid content

(42-44). As such, it is particularly difficult to easily specify the flavonoid content of food, making flavonoid intake particularly difficult to assess in population settings. This thesis will explore the complexities of flavonoid intake assessment, and how it is reflected in population based investigations.

1.3.3 Flavonoid absorption and metabolism in humans

Following consumption, the digestion, absorption and metabolic processes that flavonoids undergo in the gastrointestinal tract depends on the structure of the flavonoid compound present in the food at time of consumption, and the resident microflora

Chapter 1: Overview Chapter 1: Page 8 present in the gastrointestinal tract. This section provides a general overview of the absorption and metabolism of flavonoids.

Some flavonoid glycosides can be absorbed intact via the sodium dependent glucose transporter 1 (45), monocarboxylate transporter, multidrug-associated proteins 1 and 2, and the P-glycoprotein (46). However, in other circumstances, glycones are converted to aglycones in the gastrointestinal tract lumen by lactase phlorizin hydrolase. The aglycone compound can then passively diffuse across the brush border into the circulation (45).

Resident gastrointestinal tract microflora also play a role in modifying flavonoid structure, including ring scission, demethylation, dehydroxylation, and removal of glycoside moieties, which alters flavonoid absorption efficiency. However, the exact absorptive processes for all flavonoid compounds are not fully understood, and proportions of flavonoids absorbed are highly variable (47).

Following absorption, flavonoids play a role in numerous metabolic pathways. Many flavonoids modulate the expression and activity of cytochrome P450 enzymes (48-50), which structurally modifies the participating flavonoid compound (50). The structure of flavonoids also enables them to alter the function of many other physiologically important enzymes such as kinases, phospholipase A2, ATPase, and reverse transcriptase (51).

Flavonoids have also been shown to possess antioxidant activity by 1) inhibiting lipid peroxidation at the aqueous lipid interface of cells (52); 2) scavenging peroxyl (53), alkylperoxyl (54), superoxide (55) and peroxynitrite (56) radicals; 3) preventing DNA damage (57-60); and 4) inducing glutathione S-transferase (61). However, health benefits of flavonoids are unlikely to be primarily due to systemic antioxidant activity in vivo.

Chapter 1: Overview Chapter 1: Page 9 Not only is there variation in mechanisms promoting flavonoid absorption and metabolism, but there is also variation in the pathways of excretion. Some flavonoid compounds are excreted in the urine (62, 63). Following hepatic modification, some flavonoids are excreted in bile (64, 65). Some flavonoids and flavonoid derivatives circulate in the plasma, and there may be some tissue accumulation (51).

Due to the high degree of variability in flavonoid absorption, metabolism and excretion, there is no reference standard biomarker for flavonoid intake. As such, this thesis implements food composition data to ascertain flavonoid intake in populations.

1.3.4 Health benefits of flavonoid consumption

Consumption of flavonoids and flavonoid rich foods has been associated with lower risk of some chronic diseases (66). The beneficial associations of flavonoid intake with risk of cancer and cardiovascular disease associated outcomes (15, 16, 67-77) are likely mediated, at least in part, through the role of flavonoids in improving endothelial function, nitric oxide status (78, 79), blood pressure (80-82), and platelet function, and by reducing local oxidative stress and inflammation (83).

1.3.5 Investigating flavonoid-disease relationships

As previously discussed, flavonoid absorption, metabolism and excretion varies amongst individuals and flavonoid compounds. Because no single reference biomarker for flavonoid intake has been defined (51), alternative methods for determining flavonoid exposure for individuals needs to be implemented.

Therefore, the most appropriate method for assessing flavonoid consumption in populations is to combine traditional dietary assessment methods with flavonoid food composition databases, to obtain estimates of flavonoid intake in grams per day (68,

84). It is these computed estimates of flavonoid intake which are used as the

Chapter 1: Overview Chapter 1: Page 10 independent variable to investigate diet-disease relationships in population based settings. This method has several limitations, but remains the best method for estimation of flavonoid intake in populations.

The earliest flavonoid food composition database was developed by the United States

Department of Agriculture (USDA) (85, 86). Early versions of the USDA databases were restricted to the flavonol and flavone classes. As such, early epidemiological investigations of flavonoid-disease relationships were limited to these classes, and the role of other major flavonoid classes remained uncertain. With advances in analytical technology, quantification of flavanols, isoflavones, flavanones and proanthocyanidins has been made possible. Therefore, studies presented in this thesis are based on USDA databases that present food composition data for all major flavonoid classes. The flavonoid databases implemented in each chapter are reflective of the available data at time of writing.

More recently, the Phenol-Explorer (PE) (27) database has emerged, providing an additional high quality summary of the flavonoid content of commonly consumed foods. Unlike the USDA database which was compiled by the US Agricultural Research

Service, the PE database was compiled by the French National Institute for Agricultural

Research in collaboration with French Agency for Food, Environmental and

Occupational Health & Safety, the International Agency for Research on Cancer,

Universities of Alberta, Barcelona, and In Siliflo. Flavonoid investigations conducted in the latter part of the thesis implemented data from both the USDA and PE databases.

Chapter 1: Overview Chapter 1: Page 11 1.3.6 Current gaps in flavonoid knowledge

There is strong mechanistic and intervention data to demonstrate that flavonoids and foods rich in flavonoids display health promoting properties (78-83). Due to limitations and variations in food composition databases, the importance of diets high in flavonoid compounds habitually consumed over long periods of time remains unclear. Some studies that show particular flavonoid classes are associated with reduced risk of morbidity and mortality (15, 16, 67-77). However, these findings are not replicated in other studies in similar populations.

A major contributor to the lack of coherence in previous epidemiological studies of flavonoids is variations in flavonoids classes assessed when determining flavonoid intake. A strength of this thesis is that intakes of all major flavonoid classes are assessed, providing a comprehensive investigation of the role that total flavonoid and individual flavonoid classes play in diet-disease relationships.

With the emergence of the PE database, and the release of the 2013 USDA flavonoid food composition data, there is the opportunity to derive the most comprehensive flavonoid intake estimates. However, the appropriateness of each database for use in epidemiological studies has yet to be investigated. Therefore, this thesis will implement, compare and contrast data from the most up-to-date and comprehensive flavonoid food content databases.

Chapter 1: Overview Chapter 1: Page 12 1.4 PROBIOTIC BACTERIA

1.4.1 Definition of probiotics

Yoghurt is a dairy product produced by the bacterial fermentation of milk. The majority of yoghurt products consumed in Australia contain probiotic cultures of Lactobacillus acidophilus, Bifidobacteria and Lactobacillus casei (87). The definition of probiotics has evolved to reflect developments in the understanding of probiotic actions. Early definitions characterised probiotics as bacteria that improved intestinal microflora composition. More recent definitions have broadened to encompass additional health benefits of probiotic consumption beyond the colonisation of the gastrointestinal tract

(88). The World Health Organisation (WHO) (89) defines probiotics as ‘live microorganisms which, when administered in adequate amounts, confer a health benefit on the host’. Interestingly, this definition does not stipulate that probiotics improve colonic micro flora composition, suggesting that metabolites of probiotics may exert health benefits, independent of gastrointestinal colonisation.

The WHO definition encompasses both the bacteria and food products containing bacteria, but it remains unclear whether probiotic food products possess greater probiotic action than the isolated probiotic bacteria. Biologically active peptides released from proteins during the bacterial fermentation of milk are thought to contribute the metabolic benefits associated with probiotic consumption and aids in gastrointestinal colonisation of the probiotic species (90). However there is limited intervention data investigating this hypothesis. This thesis will report results of a randomised factorial study which directly identifies the independent and additive metabolic and gastrointestinal effects of supplementation of isolated probiotic bacteria and supplementation of probiotic bacteria in the whole-food matrix of yoghurt.

Chapter 1: Overview Chapter 1: Page 13 1.4.2 Health benefits of probiotics

Stemming from work in the early 20th century, there is now a well-established role of probiotic yoghurt in improving gastrointestinal conditions such as antibiotic-associated and C. difficile-associated diarrhoea (91, 92). Focus has now turned toward investigating the metabolic benefits of probiotic consumption. The cardiovascular benefits of yoghurt have been investigated since the early 1970s (93), with evidence of beneficial effects (94, 95), and promising results indicate the potential role of probiotics in improving glycaemic control (96, 97). However, due to variations in study design and lack of power to detect observed relationships, the data is not conclusive, with studies showing little or no metabolic benefit of probiotic bacteria (98-102). Therefore, this thesis presents data from a carefully designed and appropriately controlled randomised study exploring the health benefits of daily probiotic consumption.

1.4.2.1 Probiotics and serum lipid profile

Numerous mechanisms to explain the hypocholesterolaemic effect of probiotics have been hypothesised, including the role of probiotic bacteria in increasing: 1) bile acid deconjugation through the action of bile salt hydrolase (103-106); 2) cholesterol and fatty acid assimilation into probiotic bacteria membranes (107); and 3) conversion of cholesterol to coprostanol in the gastrointestinal tract (107).

The current understanding of probiotic action is insufficient to determine whether any hypocholesterolaemic benefits of probiotic fermented milk products are due to the mode of delivery (fermented milk), or the probiotic strain used. A study by Saxelin et al. (108) investigated the effect mode of probiotic delivery (cheese, yoghurt and capsule) had on serum lipid profile and microflora composition. They found that despite results indicating yoghurt consumption results in greater improvements in microflora composition, none of the test articles significantly altered serum lipids. This is likely

Chapter 1: Overview Chapter 1: Page 14 due to underpowering of the study and the combination of probiotic strains used in the study. Therefore, in an appropriately powered study, this thesis will present data which investigates the efficacy of L. acidophilus La5 and B. lactis Bb12 (probiotic strains with demonstrated previously hypocholesterolaemic effects) provided as either whole food or supplements (capsules).

The hypocholesterolaemic effects of probiotics have been investigated in several small randomised controlled trials, the results of which have been largely conflicting (109-

114). In order to summarise and draw conclusions about the ability of probiotics to improve serum lipid concentrations, a Cochrane approved systematic review and meta- analysis is presented in this thesis.

1.4.2.2 Probiotics and blood pressure

Results of population studies indicate that high intake of dairy foods is associated with lower incidence of hypertension (115). However, an intervention study failed to demonstrate blood pressure lowering associated with milk supplementation (116). Lack of power to detect small but clinically important effects on blood pressure is likely to have been a major issue in most previous trials. Systolic and diastolic blood pressure lowering has been observed with fermented milk containing bioactive peptides in hypertensive subjects (95, 117, 118), but results are not conclusive (119), and the effect of strain and administration form is yet to be investigated.

Effects of probiotic fermentation of milk to release bioactive peptides that inhibit ACE activity may explain the antihypertensive role of fermented milk (90). This has been demonstrated in hypertensive rats, where administration of fermented milk containing

ACE-inhibitory bioactive tripeptides, had a dose dependant hypotensive effect. It is also hypothesised that the presence of opioid like bioactive peptides in yoghurt may play an additional antihypertensive role (120). Therefore, this thesis will present results of a

Chapter 1: Overview Chapter 1: Page 15 study which investigated the hypothesis that that yoghurt and its probiotics have independent and additive effects to reduce blood pressure.

1.4.2.3 Probiotics and diabetes

The role of probiotics in improving glycaemic control has been explored in a RCT of probiotic supplementation and dietary education in normoglycaemic pregnant women

(121). This study found that in addition to dietary counselling, probiotic supplementation resulted in significantly lower glucose and insulin concentrations, reduced risk of elevated blood glucose level, and higher insulin sensitivity. Similarly, in fructose induced type 2 diabetic rats, probiotic fermented milk supplementation delayed the onset of glucose intolerance, hyperglycaemia, hyperinsulinaemia, dyslipidaemia and oxidative stress (97). By lowering glycated haemoglobin levels in diabetic rats, it appears probiotics may also play a role in long term glycaemic control (122).

The hypoglycaemic mechanisms of probiotics have been explored in animal models.

Improved glycaemic control may be a result of increased CD3+ and CD4+ T cells, and interferon-y and interlekin-2 delaying the onset of systemic inflammatory diabetes (96).

Another mechanism for the glycaemic benefits of probiotic supplementation was hypothesised by Yamano et al. (123) in a study that found intraduodenal injection of

Lactobacillus johnsonii La1 (LJLa1) in urethrane anaesthetised rats inhibited increases in plasma glucagon and glucose. It is likely this effect was due to suppression of adrenal sympathetic nerve activity and stimulation of gastric vagal nerve activity; possibly by interaction of LJLa1, or its metabolites, with afferent vagal nerve endings, decreasing adrenalin secretion thus reducing plasma glucagon and glucose concentrations (123). It is also hypothesised that the high ferulic acid esterase activity of LJ (124) has antihyperglycaemic and antioxidant effects (125).

Chapter 1: Overview Chapter 1: Page 16 There is substantial evidence in animal models to suggest the role of probiotics in improving glycaemic control, and even delaying the onset of type 2 diabetes. However there is inadequate data to confirm this effect in human subjects. Therefore, this thesis presents results of an investigation into the effects of probiotics provided as either the whole food or in capsules on biomarkers of glycaemic control.

1.4.3 Assessing probiotic-disease relationships

The World Health Organisation and Food and Agriculture Organisation of the United

Nations food and nutrition report entitled ‘Probiotics in food: health and nutritional properties and guidelines for evaluation’ (126) outlines recommendations for investigating and defining the probiotic effects of these bacteria in humans.

The first recommendation relates to strain specificity of health claims. There are numerous strains of probiotic bacteria, many of which are used in yoghurt manufacturing. It has been clearly established that the health benefits of probiotic bacteria are highly strain specific, due to the differing ability of the bacteria to survive and colonise the gastrointestinal tract and the differing metabolic functions exerted by each strain (89). As such, it is essential that health claims of probiotic products be directly linked to the particular probiotic strain present in the test article, and that health benefits of specific strains cannot be extrapolated to other strains without experimentation (126). In order to develop nutraceuticals and promote the health benefits of yoghurt consumption, it is important to establish a causal health benefit associated with consumption of probiotic strains commonly used in yoghurt manufacturing. Therefore, the randomised controlled trial presented in this thesis administered probiotic strains commonly used in the dairy and supplement industries.

However, it is acknowledged that several other bacterial strains are also commonly used.

Chapter 1: Overview Chapter 1: Page 17 The report also states that in order for a microorganism to be considered probiotic, it must be able to confer a health benefit in the actual product vehicle that will be made available to humans (126). It is desirable if test articles administered in randomised controlled trials are commercially available food or supplement products. Therefore, the randomised controlled trial presented in this thesis implemented commercially available probiotic yoghurt and probiotic capsules.

Finally, it is also recommended that probiotic products are made more widely available for populations at high risk of morbidity and mortality (126). As such, it is important to explore benefits of probiotics in high risk populations. Therefore, the randomised controlled trial presented in this thesis explores the health benefits of yoghurt and its probiotics in the elderly and in men and women with features of the metabolic syndrome.

1.4.4 Current gaps in probiotic knowledge

There is a lack of statistically significant probiotic efficacy data in humans (126). There are numerous metabolic studies of probiotics with a small sample size (95, 127, 128).

However, there is a lack of concurrency in their findings. Implementation of studies with appropriate power to detect observed relationships will aide in clarifying the precise role of probiotics in improving chronic disease risk factors. Therefore, the sample size of the randomised controlled trial presented in this thesis was determined by prespecified sample size calculations, which make it the largest randomised controlled trial on the topic to date.

The results of randomised controlled trials suggest a potential role of probiotics to elicit metabolic benefits. However due to under-powering of studies, as well as variability in findings and study design, it is necessary to apply a systematic approach to confirm health benefits in humans (126). Therefore, this thesis will also present data of a

Chapter 1: Overview Chapter 1: Page 18 systematic review and meta-analysis to summarise the benefits of probiotic bacteria on serum lipid profile.

Chapter 1: Overview Chapter 1: Page 19 1.5 AIMS AND OBJECTIVES

1.5.1 Overarching aims of thesis and relevance to field

There is a body of evidence to suggest that components of food which are not nutrients

(non-nutritive food components) may add to the health promoting effects of foods.

Despite this, current dietary guidelines focus on nutrient profile of foods in determining the components of a health promoting diet, and do not directly focus on non-nutritive food components.

As such, this thesis aims to explore the health benefits of consuming non-nutritive food components, specifically flavonoids and probiotics. It is anticipated that the findings of this thesis will add to the scientific dossier of non-nutritive food components, to further clarify whether foods high in flavonoids and probiotics should be endorsed as important contributors to a health promoting diet.

1.5.2 Research questions and hypotheses explored in order to achieve aim

In order to achieve the thesis aim, epidemiological, clinical and systematic review studies were implemented. Table 1 provides a summary of the research questions considered in the thesis, and the specific hypotheses used to answer them.

Chapter 1: Overview Chapter 1: Page 20 1.6 STRUCTURE OF THESIS

This thesis is, in accordance with the University of Western Australia Higher Degree by

Research Rules Act, presented as a series of scientific papers that resulted from the PhD candidature. Therefore, each thesis chapter contains an independent introduction, literature review, methods, results and discussion section. As such, some degree of overlap of papers presenting data from the same study is unavoidable. A general conclusion chapter closes out the thesis.

In the current introductory chapter, the aims, background, and aims of the thesis are introduced. This chapter has provided the basic impetus for the research, and outlined the fundamental hypotheses that will be addressed.

The vascular benefits of flavonoid consumption will be explored in epidemiological investigations looking at associations of flavonoid intake with atherosclerotic vascular disease (Chapter 2) and renal function (Chapter 3) outcomes. Chapter 2 will also explore whether the benefits ascribable to flavonol intake can be separated from the benefits attributable to tea; the major dietary source of flavonoids. Chapter 4 will then explore the association of flavonoid intake with all-cause mortality.

Chapter 5 uses the most up-to-date and comprehensive flavonoid food composition databases to provide a methodological summary and interpretation of these databases in an epidemiological context.

In addition to epidemiological studies, randomised controlled trials are another method used to assess diet-disease relationships in humans. Appendix A outlines the methods of a randomised controlled factorial study, exploring the health benefits of daily consumption of yoghurt and its probiotics. Chapters 6, 7 and 8 present the faecal

Chapter 1: Overview Chapter 1: Page 21 microflora, glycaemic, and blood pressure and cholesterol results of the trial, respectively.

With the aim of summarising current evidence for the hypocholesterolaemic benefits of probiotic supplementation, Chapter 9 presents results of a systematic review and meta- analysis on the topic. With a return to epidemiological investigations, Chapter 10 concludes the thesis with an exploration of the cardioprotective role probiotic yoghurt plays in a population setting.

The final chapter of this thesis (Chapter 11) will relate study findings to research questions and thesis aims, and outline the theoretical and practical implications of this research.

Chapter 1: Overview Chapter 1: Page 22 1.7 CHAPTER 1 REFERENCES

1. Young VR. Human nutrient requirements: the challenge of the post-genome era. Journal of Nutrition 2002;132(4):621-9. 2. Elmadfa I, Meyer AL. Importance of food composition data to nutrition and public health. European Journal of Clinical Nutrition 2010;64:S4-S7. 3. Papanikolaou NC, Hatzidaki EG, Belivanis S, Tzanakakis GN, Tsatsakis AM. Lead toxicity update. A brief review. Medical science monitor 2005;11(10). 4. Joint FAO/WHO Expert Committee on Food Additives. Evaluation of certain food additices: seventy-sixth report of the Joint FAO/WHO Expert Committee on Food Additives. WHO technical report series; no 974. Geneva, Switzerland, 2012. 5. Goldin BR. Health benefits of probiotics. British Journal of Nutrition 1998;80:S203-S7. 6. Yao LH, Jiang YM, Shi J, et al. Flavonoids in food and their health benefits. Plant Foods for Human Nutrition 2004;59(3):113-22. 7. Verschuren WMM. Diet and cardiovascular disease. Current Cardiology Reports 2012;14(6):701-8. 8. Donaldson MS. Nutrition and cancer: a review of the evidence for an anti-cancer diet. Nutrition Journal 2004;3(1):19. 9. Ash S, Campbell K, MacLaughlin H, et al. Evidence based practice guidelines for the nutritional management of chronic kidney disease. Nutrition and Dietetics 2006;63:S33-S45. 10. Yach D, Hawkes C, Gould CL, Hofman KJ. The global burden of chronic diseases. JAMA: The Journal of the American Medical Association 2004;291(21):2616-22. 11. Joint FAO/WHO Expert Consultation. Diet, nutrition and the prevention of chronic diseases. WHO technical report series 2003;916. 12. Lichtenstein AH, Appel LJ, Brands M, et al. Diet and lifestyle recommendations revision 2006: a scientific statement from the American Heart Association nutrition committee. Circulation 2006;114(1):82-96. 13. Australian Government, Department of Health and Ageing, National Health and Medical Research Foundation. Nutrient reference values for Australia and New Zealand. Canberra: Commonwealth of Australia, 2005. 14. U.S. Department of Agriculture, U.S. Department of Health and Human Services. Dietary guidelines for Americans, 2010. 7 ed. Washington: U.S. Governement Printing Office, 2010. 15. Arab L, Liu W, Elashoff D. Green and black tea consumption and risk of stroke. Stroke 2009;40(5):1786-92. 16. Wang Z-M, Zhou B, Wang Y-S, et al. Black and green tea consumption and the risk of coronary artery disease: a meta-analysis. American Journal of Clinical Nutrition 2011;93(3):506-15. 17. Lewis JR, Prince RL, Zhu K, Devine A, Thompson PL, Hodgson JM. Habitual chocolate intake and vascular disease: a prospective study of clinical outcomes in older women. Archives of Internal Medicine 2010;170(20):1857-8. 18. Buitrago-Lopez A, Sanderson J, Johnson L, et al. Chocolate consumption and cardiometabolic disorders: systematic review and meta-analysis. British Medical Journal 2011;343.

Chapter 1: Overview Chapter 1: Page 23 19. Wollin SD, Jones PJH. Alcohol, red wine and cardiovascular disease. Journal of Nutrition 2001;131(5):1401-4. 20. Liu RH. Health benefits of fruit and vegetables are from additive and synergistic combinations of phytochemicals. American Journal of Clinical Nutrition 2003;78(3):517S-20S. 21. Parvez S, Malik KA, Ah Kang S, Kim HY. Probiotics and their fermented food products are beneficial for health. Journal of Applied Microbiology 2006;100(6):1171-85. 22. Moss G, Smith P, Tavernier D. Glossary of class names of organic compounds and reactive intermediates based on structure. International Union of Pure and Applied Chemistry 1995;67(8):1307-75. 23. Aherne SA, O’Brien NM. Dietary flavonols: chemistry, food content, and metabolism. Nutrition 2002;18(1):75-81. 24. Cook N, Samman S. Flavonoids: chemistry, metabolism, cardioprotective effects, and dietary sources. Journal of Nutritional Biochemistry 1996;7(2):66- 76. 25. Das DK. Naturally occurring flavonoids: Structure, chemistry, and high- performance liquid chromatography methods for separation and characterization. Edtion ed. In: Lester P, ed. Methods in Enzymology: Academic Press, 1994:410- 20. 26. US Department of Agriculture. USDA database for the isoflavone content of selected foods; release 2.0. Maryland, 2008. 27. Neveu V, Perez-Jiménez J, Vos F, et al. Phenol-Explorer: an online comprehensive database on polyphenol contents in foods. Database 2010. 28. Koes RE, Quattrocchio F, Mol JNM. The flavonoid biosynthetic pathway in plants: function and evolution. BioEssays 1994;16(2):123-32. 29. Goto T, Kondo T. Structure and molecular stacking of anthocyanins: flower color variation. Angewandte Chemie International Edition in English 1991;30(1):17-33. 30. Hondo T, Yoshida K, Nakagawa A, Kawai T, Tamura H, Goto T. Structural basis of blue-colour development in flower petals from Commelina communis. Nature 1992;358:515-8. 31. Ylstra B, Touraev A, Moreno RMB, et al. Flavonols stimulate development, germination, and tube growth of tobacco pollen. Plant Physiology 1992;100(2):902-7. 32. Matern U, Heller W, Himmelspach K. Conformational changes of apigenin 7-O- (6-O-malonylglucoside), a vacuolar pigment from parsley, with solvent composition and proton concentration. European Journal of Biochemistry 1983;133(2):439-48. 33. Schmelzer E, Jahnen W, Hahlbrock K. In situ localization of light-induced chalcone synthase mRNA, chalcone synthase, and flavonoid end products in epidermal cells of parsley leaves. Proceedings of the National Academy of Sciences 1988;85(9):2989-93. 34. Stapleton AE. Ultraviolet radiation and plants: burning questions. Plant Cell 1992;4(11):1353. 35. Subramanian S, Stacey G, Yu O. Distinct, crucial roles of flavonoids during legume nodulation. Trends in Plant Science 2007;12(7):282-5. 36. Treutter D. Significance of flavonoids in plant resistance and enhancement of their biosynthesis. Plant Biology 2005;7(6):581-91.

Chapter 1: Overview Chapter 1: Page 24 37. Schijlen EGWM, Ric de Vos CH, van Tunen AJ, Bovy AG. Modification of flavonoid biosynthesis in crop plants. Phytochemistry 2004;65(19):2631-48. 38. US Department of Agriculture. USDA database for the flavonoid content of selected foods release 1. Maryland, 2003. 39. Hooper L, Kroon PA, Rimm EB, et al. Flavonoids, flavonoid-rich foods, and cardiovascular risk: a meta-analysis of randomized controlled trials. American Journal of Clinical Nutrition 2008;88(1):38-50. 40. Bondonno CP, Yang X, Croft KD, et al. Flavonoid-rich apples and nitrate-rich spinach augment nitric oxide status and improve endothelial function in healthy men and women: a randomized controlled trial. Free Radical Biology and Medicine 2002;52(1):95-102. 41. Ras RT, Zock PL, Draijer R. Tea consumption enhances endothelial-dependent vasodilation; a meta-analysis. PLoS One 2011;6(3.). 42. Howard L, Pandjaitan N, Morelock T, Gil M. Antioxidant capacity and phenolic content of spinach as affected by genetics and growing season. Journal of Agricultural and Food Chemistry 2002;50(21):5891-6. 43. van der Sluis AA, Dekker M, de Jager A, Jongen WM. Activity and concentration of polyphenolic antioxidants in apple: effect of cultivar, harvest year, and storage conditions. Journal of Agricultural and Food Chemistry 2001;49(8):3606-13. 44. Goldman I, Kader A, Heintz C. Influence of production, handling, and storage on phytonutrient content of foods. Nutrition Reviews 1999;57(9):46-52. 45. Day AJ, Gee JM, DuPont MS, Johnson IT, Williamson G. Absorption of quercetin-3-glucoside and quercetin-4′-glucoside in the rat small intestine: the role of lactase phlorizin hydrolase and the sodium-dependent glucose transporter. Biochemical Pharmacology 2003;65(7):1199-206. 46. Vaidyanathan JB, Walle T. Cellular uptake and efflux of the tea flavonoid (- )epicatechin-3-gallate in the human intestinal cell line Caco-2. Journal of Pharmacology and Experimental Therapeutics 2003;307(2):745-52. 47. Walle T. Absorption and metabolism of flavonoids. Free Radical Biology and Medicine 2004;36(7):829-37. 48. Hodek P, Trefil P, Stiborová M. Flavonoids-potent and versatile biologically active compounds interacting with cytochromes P450. Chemico-Biological Interactions 2002;139(1):1-21. 49. Kellis JT, Vickery LE. Inhibition of human estrogen synthetase (aromatase) by flavones. Science 1984;225(4666):1032-4. 50. Buening MK, Chang RL, Huang M-T, Fortner JG, Wood AW, Conney AH. Activation and inhibition of benzo(a)pyrene and aflatoxin B1 metabolism in human liver microsomes by naturally occurring flavonoids. Cancer Research 1981;41(1):67-72. 51. Middleton E, Kandaswami C, Theoharides TC. The effects of plant flavonoids on mammalian cells: implications for inflammation, heart disease, and cancer. Pharmacological Reviews 2000;52(4):673-751. 52. Erlejman AG, Verstraeten SV, Fraga CG, Oteiza PI. The interaction of flavonoids with membranes: potential determinant of flavonoid antioxidant effects. Free Radical Research 2004;38(12):1311-20. 53. Dugas AJ, Castañeda-Acosta J, Bonin GC, Price KL, Fischer NH, Winston GW. Evaluation of the total peroxyl radical-scavenging capacity of flavonoids: structure−activity relationships. Journal of Natural Products 2000;63(3):327-31.

Chapter 1: Overview Chapter 1: Page 25 54. Sawa T, Nakao M, Akaike T, Ono K, Maeda H. Alkylperoxyl radical- scavenging activity of various flavonoids and other phenolic compounds: implications for the anti-tumor-promoter effect of vegetables. Journal of Agricultural and Food Chemistry 1999;47(2):397-402. 55. Robak J, Gryglewski RJ. Flavonoids are scavengers of superoxide anions. Biochemical Pharmacology 1988;37(5):837-41. 56. Haenen GRMM, Paquay JBG, Korthouwer REM, Bast A. Peroxynitrite scavenging by flavonoids. Biochemical and Biophysical Research Communications 1997;236(3):591-3. 57. Kootstra A. Protection from UV-B-induced DNA damage by flavonoids. Plant Molecular Biology 1994;26(2):771-4. 58. Sestili P, Guidarelli A, Dachà M, Cantoni O. Quercetin prevents DNA single strand breakage and cytotoxicity caused by tert-butylhydroperoxide: free radical scavenging versus iron chelating mechanism. Free Radical Biology and Medicine 1998;25(2):196-200. 59. Duthie SJ, Collins AR, Duthie GG, Dobson VL. Quercetin and myricetin protect against hydrogen peroxide-induced DNA damage (strand breaks and oxidised pyrimidines) in human lymphocytes. Mutation Research/Genetic Toxicology and Environmental Mutagenesis 1997;393(3):223-31. 60. Wilms LC, Hollman PCH, Boots AW, Kleinjans J. Protection by quercetin and quercetin-rich fruit juice against induction of oxidative DNA damage and formation of BPDE-DNA adducts in human lymphocytes. Mutation Research/Genetic Toxicology and Environmental Mutagenesis 2005;582(1):155- 62. 61. Sahu SC, Gray GC. Pro-oxidant activity of flavonoids: effects on glutathione and glutathione S-transferase in isolated rat liver nuclei. Cancer letters 1996;104(2):193-6. 62. Krogholm KS, Haraldsdóttir J, Knuthsen P, Rasmussen SE. Urinary total flavonoid excretion but not 4-pyridoxic acid or potassium can be used as a biomarker for the intake of fruits and vegetables. Journal of Nutrition 2004;134(2):445-51. 63. Choudhury R, Srai SK, Debnam E, Rice-Evans CA. Urinary excretion of hydroxycinnamates and flavonoids after oral and intravenous administration. Free Radical Biology and Medicine 1999;27(3):278-86. 64. Das NP, Sothy SP. Studies on flavonoid metabolism. Biliary and urinary excretion of metabolites of (+)-[U-4 C] catechin. Biochemistry Journal 1971;125:417-23. 65. Chen J, Lin H, Hu M. Metabolism of flavonoids via enteric recycling: role of intestinal disposition. Journal of Pharmacology and Experimental Therapeutics 2003;304(3):1228-35. 66. Knekt P, Kumpulainen J, Järvinen R, et al. Flavonoid intake and risk of chronic diseases. American Journal of Clinical Nutrition 2002;76(3):560-8. 67. Huxley RR, Neil HA. The relation between dietary flavonol intake and coronary heart disease mortality: a meta-analysis of prospective cohort studies. European Journal of Clinical Nutrition 2003;57(8):904-8. 68. Mink PJ, Scrafford CG, Barraj LM, et al. Flavonoid intake and cardiovascular disease mortality: a prospective study in postmenopausal women. American Journal of Clinical Nutrition 2007;85(3):895-909.

Chapter 1: Overview Chapter 1: Page 26 69. Mursu J, Voutilainen S, Nurmi T, Tuomainen T-P, Kurl S, Salonen JT. Flavonoid intake and the risk of ischaemic stroke and CVD mortality in middle- aged Finnish men: the Kuopio Ischaemic Heart Disease Risk Factor Study. British Journal of Nutrition 2008;100(4):890-5. 70. Mursu J, Nurmi T, Tuomainen TP, Salonen JT, Pukkala E, Voutilainen S. Intake of flavonoids and risk of cancer in Finnish men: the Kuopio Ischaemic Heart Disease Risk Factor Study. International Journal of Cancer 2008;123(3):660-3. 71. Bosetti C, Spertini L, Parpinel M, et al. Flavonoids and breast cancer risk in Italy. Cancer Epidemiology Biomarkers and Prevention 2005;14(4):805-8. 72. Liu L-Z, Fang J, Zhou Q, Hu X, Shi X, Jiang B-H. Apigenin inhibits expression of vascular endothelial growth factor and angiogenesis in human lung cancer cells: implication of chemoprevention of lung cancer. Molecular Pharmacology 2005;68(3):635-43. 73. Garcia-Closas R, Agudo A, Gonzalez CA, Riboli E. Intake of specific carotenoids and flavonoids and the risk of lung cancer in women in Barcelona, Spain. Nutrition and Cancer 1998;32(3):154-8. 74. Zava DT, Duwe G. Estrogenic and antiproliferative properties of genistein and other flavonoids in human breast cancer cells in vitro. Nutrition and Cancer 1997;27(1):31-40. 75. So FV, Guthrie N, Chambers AF, Moussa M, Carroll KK. Inhibition of human breast cancer cell proliferation and delay of mammary tumorigenesis by flavonoids and citrus juices. Nutrition and Cancer 1996;26(2). 76. Peterson J, Lagiou P, Samoli E, et al. Flavonoid intake and breast cancer risk: a case–control study in Greece. British Journal of Cancer 2003;89(7):1255-9. 77. Bushman JL. Green tea and cancer in humans: a review of the literature. Nutrition and Cancer 1998;31(3):151-9. 78. Loke WM, Hodgson JM, Proudfoot JM, McKinley AJ, Puddey IB, Croft KD. Pure dietary flavonoids quercetin and (-)-epicatechin augment nitric oxide products and reduce endothelin-1 acutely in healthy men. American Journal of Clinical Nutrition 2008;88(4):1018-25. 79. Schroeter H, Heiss C, Balzer J, et al. (–)-Epicatechin mediates beneficial effects of flavanol-rich cocoa on vascular function in humans. Proceedings of the National Academy of Sciences 2006;103(4):1024-9. 80. Edwards RL, Lyon T, Litwin SE, Rabovsky A, Symons JD, Jalili T. Quercetin reduces blood pressure in hypertensive subjects. Journal of Nutrition 2007;137(11):2405-11. 81. Egert S, Bosy-Westphal A, Seiberl J, et al. Quercetin reduces systolic blood pressure and plasma oxidised low-density lipoprotein concentrations in overweight subjects with a high-cardiovascular disease risk phenotype: a double-blinded, placebo-controlled cross-over study. British Journal of Nutrition 2009;102(07):1065-74. 82. Brown AL, Lane J, Coverly J, et al. Effects of dietary supplementation with the green tea polyphenol epigallocatechin-3-gallate on insulin resistance and associated metabolic risk factors: randomized controlled trial. British Journal of Nutrition 2009;101(06):886-94. 83. Loke W, Hodgson J, Croft K. The biochemistry behind the potential cardiovascular protection by dietary flavonoids. Edtion ed. In: Fraga CG, ed. Plant phenolics and human health: biochemistry, nutrition and pharmacology. New Jersey: John Wiley & Sons, 2009.

Chapter 1: Overview Chapter 1: Page 27 84. Chun OK, Chung SJ, Song WO. Estimated dietary flavonoid intake and major food sources of U.S. adults. Journal of Nutrition 2007;137(5):1244-52. 85. US Department of Agriculture. USDA database for the flavonoid content of selected foods; release 2.1. Maryland, 2007. 86. Nutrient Data Laboratory, Agricultural Research Service, US Department of Agriculture. USDA database for proanthocyanidin content of selected foods. In: USDA, ed. Beltsville, MD, 2004. 87. Shortt C. The probiotic century: historical and current perspectives. Trends in Food Science and Technology 1999;10(12):411-7. 88. Schrezenmeir J, de Vrese M. Probiotics, prebiotics, and synbiotics: approaching a definition. The American Journal of Clinical Nutrition 2001;73(2):361S-4S. 89. Joint FAO/WHO Expert Consultation on Evaluation of Health and Nutritional Properties of Probiotics in Food. Health and nutritional properties of probiotics in food including powder milk with live lactic acid bacteria. Argentina, 2001. 90. Meisel H. Overview on milk protein-derived peptides. International Dairy Journal 1998;8(5-6):363-73. 91. Anukam KC, Reid G. Probiotics: 100 years (1907-2007) after Elie Metchnikoff's observation. Communicating Current Research and Educational Topics and Trends in Applied Microbiology 2008:466-74. 92. Marteau PR, Vrese Md, Cellier CJ, Schrezenmeir J. Protection from gastrointestinal diseases with the use of probiotics. American Journal of Clinical Nutrition 2001;73(2):430S-6S. 93. Mann GV, Spoerry A. Studies of a surfactant and cholesteremia in the Maasai. American Journal of Clinical Nutrition 1974;27(5):464-9. 94. Aihara K, Kajimoto O, Hirata H, Takahashi R, Nakamura Y. Effect of powdered fermented milk with Lactobacillus helveticus on subjects with high-normal blood pressure or mild hypertension. Journal of the American College of Nutrition 2005;24(4):257-65. 95. Hata Y, Yamamoto M, Ohni M, Nakajima K, Nakamura Y, Takano T. A placebo-controlled study of the effect of sour milk on blood pressure in hypertensive subjects. American Journal of Clinical Nutrition 1996;64(5):767- 71. 96. Matsuzaki T, Yamazaki R, Hashimoto S, Yokokura T. Antidiabetic effects of an oral administartion of Lactobacillus casei in a non-insulin-dependent diabetes mellitus (NIDDM) model using KK-Ay mice. Endocrine Journal 1997;44(3):357-65. 97. Yadav H, Jain S, Sinha PR. Antidiabetic effect of probiotic dahi containing Lactobacillus acidophilus and Lactobacillus casei in high fructose fed rats. Nutrition 2007;23(1):62-8. 98. Asemi Z, Samimi M, Tabassi Z, et al. Effect of daily consumption of probiotic yoghurt on insulin resistance in pregnant women: a randomized controlled trial. European Journal of Clinical Nutrition 2012. 99. Zarrati M, Shidfar F, Nourijelyani K, et al. Lactobacillus acidophilus La5, Bifidobacterium BB12, and Lactobacillus casei DN001 modulate gene expression of subset specific transcription factors and cytokines in peripheral blood mononuclear cells of obese and overweight people. BioFactors 2013;39(6):633-43. 100. Bukowska H, Pieczul-Mróz J, Jastrzebska M, Chełstowski K, Naruszewicz M. Decrease in fibrinogen and LDL-cholesterol levels upon supplementation of diet

Chapter 1: Overview Chapter 1: Page 28 with Lactobacillus plantarum in subjects with moderately elevated cholesterol. Atherosclerosis 1998;1998(137):2. 101. Naruszewicz M, Johansson M-L, Zapolska-Downar D, Bukowska H. Effect of Lactobacillus plantarum 299v on cardiovascular disease risk factors in smokers. American Journal of Clinical Nutrition 2002;76(6):1249-55. 102. Sanggaard K, Holst J, Rehfeld J, Sandstrom B, Raben A, Tholstrup T. Different effects of whole milk and a fermented milk with the same fat and lactose content on gastric emptying and postprandial lipaemia, but not on glycaemic response and appetite. British Journal of Nutrition 2004;92:447-59. 103. Cardona ME, Vanay VdV, Midtvedt T, Norin E. Probiotics in gnotobiotic mice Conversion of cholesterol to coprostanol in vitro and in vivo and bile acid deconjugation in vitro. Microbial Ecology in Health and Disease 2000;12(4):219-24. 104. Pereira DIA, McCartney AL, Gibson GR. An in vitro study of the probiotic potential of a bile-salt-hydrolyzing Lactobacillus fermentum strain, and determination of its cholesterol-lowering properties. Applied and Environmental Microbiology 2003;69(8):4743-52. 105. McAuliffe O, Cano RJ, Klaenhammer TR. Genetic analysis of two bile salt hydrolase activities in Lactobacillus acidophilus NCFM. Applied and Environmental Microbiology 2005;71(8):4925-9. 106. Moser SA, Savage DC. Bile salt hydrolase activity and resistance to toxicity of conjugated bile salts are unrelated properties in Lactobacilli. Applied and Environmental Microbiology 2001;67(8):3476-80. 107. Lye HS, Rusul G, Liong MT. Removal of cholesterol by lactobacilli via incorporation and conversion to coprostanol. Journal of Dairy Science 2010;93(4):1383-92. 108. Saxelin M, Lassig A, Karjalainen H, et al. Persistence of probiotic strains in the gastrointestinal tract when administered as capsules, yoghurt, or cheese. International Journal of Food Microbiology 2010;144(2):293-300. 109. Bertolami MC, Faludi AA, Batlouni M. Evaluation of the effects of a new fermented milk product (Gaio) on primary hypercholesterolemia. European Journal of Clinical Nutrition 1999;53(2):97. 110. Schaafsma G, Meuling WJ, Van Dokkum W, Bouley C. Effects of a milk product, fermented by Lactobacillus acidophilus and with fructo- oligosaccharides added, on blood lipids in male volunteers. European Journal of Clinical Nutrition 1998;52(6):436. 111. Ataie-Jafari A, Larijani B, Alavi Majd H, Tahbaz F. Cholesterol-lowering effect of probiotic yogurt in comparison with ordinary yogurt in mildly to moderately hypercholesterolemic subjects. Annals of Nutrition and Metabolism 2009;54(1):22-7. 112. Agerbaek M, Gerdes LU, Richelsen B. Hypocholesterolaemic effect of a new fermented milk product in healthy middle-aged men. European Journal of Clinical Nutrition 1995;49(5):346-52. 113. Hatakka K, Mutanen M, Holma R, Saxelin M, Korpela R. Lactobacillus rhamnosus LC705 together with Propionibacterium freudenreichii ssp shermanii JS administered in capsules is ineffective in lowering serum lipids. Journal of the American College of Nutrition 2008;27(4):441-7.

Chapter 1: Overview Chapter 1: Page 29 114. Simons LA, Amansec SG, Conway P. Effect of Lactobacillus fermentum on serum lipids in subjects with elevated serum cholesterol. Nutrition, Metabolism and Cardiovascular Diseases 2006;16(8):531-5. 115. Garcia-Palmieri M, Costas R, Jr, Cruz-Vidal M, Sorlie P, Tillotson J, Havlik R. Milk consumption, calcium intake, and decreased hypertension in Puerto Rico. Puerto Rico Heart Health Program study. Hypertension 1984;6(3):322-8. 116. Wennersberg MH, Smedman A, Turpeinen AM, et al. Dairy products and metabolic effects in overweight men and women: results from a 6-mo intervention study. American Journal of Clinical Nutrition 2009;90(4):960-8. 117. Jauhiainen T, Vapaatalo H, Poussa T, Kyronpalo S, Rasmussen M, Korpela R. Lactobacillus helveticus Fermented milk lowers blood pressure in hypertensive subjects in 24-h ambulatory blood pressure measurement. American Journal of Hypertension 2005;18(12):1600-5. 118. Seppo L, Jauhiainen T, Poussa T, Korpela R. A fermented milk high in bioactive peptides has a blood pressure lowering effect in hypertensive subjects. American Journal of Clinical Nutrition 2003;77(2):326-30. 119. Engberink MF, Schouten EG, Kok FJ, van Mierlo LAJ, Brouwer IA, Geleijnse JM. Lactotripeptides show no effect on human blood pressure: results from a double-blind randomized controlled trial. Hypertension 2008;51(2):399-405. 120. Jauhiainen T, Korpela R. Milk Peptides and Blood Pressure. The Journal of Nutrition 2007;137(3):825S-9S. 121. Laitinen K, Poussa T, Isolauri E. Probiotics and dietary counselling contribute to glucose regulation during and after pregnancy: a randomised controlled trial. British Journal of Nutrition 2009;101(11):1679-87. 122. Tabuchi M, Ozaki M, Tamura A, et al. Antidiabetic effect of Lactobacillus GG in streptozotocin-induced diabetic rats. Bioscience Biotechnology and Biochemistry 2003;67(6):1421-4. 123. Yamano T, Tanida M, Niijima A, et al. Effects of the probiotic strain Lactobacillus johnsonii strain La1 on autonomic nerves and blood glucose in rats. Life Sciences 2006;79:1963-7. 124. Lai KK, Lorca GL, Gonzalez CF. Biochemical properties of two cinnamoyl esterases purified from a Lactobacillus johnsonii strain isolated from stool samples of diabetes-resistant rats. Applied and Environmental Microbiology 2009;75(15):5018-24. 125. Balasubashini MS, Rukkumani R, Viswanathan P, Menon VP. Ferulic acid alleviates lipid peroxidation in diabetic rats. Phytotherapy Research 2004;18(4):310-4. 126. Joint FAO/WHO Expert Consultation. Probiotics in food: health and nutritional properties and guidelines for evaluation. 2006. 127. de Roos N, Schouten G, Katan M. Yoghurt enriched with Lactobacillus acidophilus does not lower blood lipids in healthy men and women with normal to borderline high serum cholesterol levels. European Journal of Clinical Nutrition 1999;53(4):277-80. 128. Mizushima S, Ohshige K, Watanabe J, et al. Randomized controlled trial of sour milk on blood pressure in borderline hypertensive men. American Journal of Hypertension 2004;17(8):701-6.

Chapter 1: Overview Chapter 1: Page 30 1.8 TABLES Chapter 1, Table 1: Thesis research questions and hypotheses used to explore them

Research question: Can the benefits of flavonoid consumption be separated from the benefits attributable to their major dietary whole food sources?

Hypothesis: High habitual intake of flavonols from tea and non-tea sources will both be associated with a reduced risk of atherosclerotic vascular disease mortality in a population of elderly women

Research question: Given strong evidence for the vascular benefits of flavonoid consumption, is flavonoid intake associated with better functioning of highly vascularised organs?

Hypothesis: High habitual proanthocyanidin intake is associated with better renal function and reduced risk of clinical renal outcomes in a population of elderly women.

Research question: Given the well documented benefits of flavonoid consumption, is the consumption of flavonoids from any class associated with death from any cause?

Hypothesis: Using two comprehensive food composition databases to assess flavonoid intake, total-flavonoid consumption will be inversely associated with risk of 5-year all-cause mortality.

Research question: Given a lack of gold standard biomarker to reflect flavonoid intake, is flavonoid intake able to be adequately assessed in population settings?

Hypothesis: There will be strong agreement between flavonoid intake estimates derived from United States Department of Agriculture and Phenol-Explorer flavonoid food content databases.

Research question: Is supplementation of the diet with probiotic bacteria able to alter faecal microflora profile?

Hypothesis: Daily probiotic supplementation, from capsules and the whole food (yoghurt) form, will improve faecal counts of Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis BB12.

Research question: Are probiotic bacteria able to improve glycaemic control?

Hypothesis: The probiotic bacteria L. acidophilus La5 and B. animalis subsp lactis Bb12, supplemented in a whole food (yoghurt) or isolated (capsules) form, will improve biomarkers of glycaemic control.

Research question: Using an appropriately powered study design, does daily supplementation with probiotic bacteria improve cardiovascular disease risk factors? Hypothesis: Six week supplementation with Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12, provided in either yoghurt or capsule form, will improve home blood pressure and serum lipid profile men and women with features of the metabolic syndrome Research question: Is it possible to draw conclusions about the cardioprotective role of probiotic bacteria?

Hypothesis: Using a Cochrane Library approved review of literature, supplementation with probiotic fermented milk and isolated probiotic bacteria will improve serum lipid profile in adults.

Research question: Is probiotic yoghurt beneficially associated with cardiovascular health in the population?

Hypothesis: Habitual intake of yoghurt, but not other dairy products, will be beneficially associated with common carotid artery intima-media thickness.

Chapter 1: Overview Chapter 2: Page 1

Health benefits of non-nutritive food components

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 2

2.1 FOREWORD

There is a strong body of population based data suggesting the vascular benefits of consuming flavonols; a major class of flavonoids (16). However, like all investigations involving whole foods, there are many potential confounding factors which make the interpretation of current data difficult.

The majority of flavonols in the diet often come from tea consumption (17). Like

Flavonols, there is a strong body of epidemiological data suggesting the vascular benefits of tea consumption (18, 19). Tea contains high concentrations of Flavonols and many other polyphenolic compounds (17). Therefore, it is possible that population based studies showing beneficial associations of Flavonol intake may be a reflection of the beneficial effect of tea intake and the components other than flavonols that tea contains. This chapter aims to explore the association between habitual Flavonol intake and atherosclerotic vascular disease, and to investigate the roles that Flavonols from both tea and non-tea sources play.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 3

2.2 ABSTRACT

British Journal of Nutrition (2013, 110, 1648-55) Kerry L Ivey, Lewis JR, Prince RL, Hodgson JM

Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women

Epidemiological studies indicate that dietary flavonoids generally, and flavonols specifically, may contribute to cardiovascular health. Tea consumption, which is often the main dietary source of flavonoids and flavonols, is associated with reduced risk of cardiovascular outcomes.

Our primary objective was to explore the association of habitual intake of flavonols from tea and non-tea sources with the risk of atherosclerotic vascular disease mortality in a population of elderly women.

1 063 women aged over 75 y, were randomly selected from ambulant Caucasian women living in Perth, Western Australia. Flavonoid consumption was assessed using the United States Department of Agriculture Flavonoid, Flavone and Proanthocyanidin databases. Atherosclerotic vascular disease mortality was assessed over 5-years of follow-up through the Western Australia Data Linkage System.

During follow-up, 64 women died from atherosclerotic vascular disease. Women in the highest compared to the lowest tertile of flavonol intake had lower risk [OR (95% CI)] of atherosclerotic vascular disease death [0.27 (0.13-0.59): P  0.01 for trend in multivariate-adjusted models]. Similar relationships were observed for flavonol intake derived from both tea [0.38 (0.18-0.79): P<0.01] and non-tea [0.41 (0.20-0.85): P=0.05] sources. Tea was the main contributor to flavonol intake (65%), and the intakes of flavonols from tea and non-tea sources were not significantly correlated.

In conclusion, increased consumption of flavonols was independently associated with lower risk of atherosclerotic vascular disease mortality. Both tea and non-tea sources of flavonols were independently associated with this benefit.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 4

2.3 INTRODUCTION

Flavonoids include a diverse group of over 6,000 compounds (5). There are seven major classes of flavonoids in the human diet, including flavonols, flavan-3-ols, proanthocyanidins, flavones, flavanones, anthocyanidins and isoflavones (6-8). Major dietary sources of flavonoids include tea, fruit and vegetables, chocolate, and red wine.

There is increasing evidence that dietary flavonoids generally and flavonols specifically contribute to cardiovascular health (9). A number of population studies have investigated the relationships of specific flavonoid classes with cardiovascular disease risk. Early population studies that assessed flavonol and flavone intakes indicated a significant reduction in coronary heart disease mortality with higher flavonol intake. A high flavonol intake has been previously associated with a 20% lower risk of fatal coronary heart disease (1). More recent studies have investigated relationships of all seven major classes of flavonoid with cardiovascular disease outcomes (10, 11). This has been made possible by recent improvements to food composition databases which have allowed assessment of all seven major classes of flavonoids (6-8).

Tea is frequently the main source of flavonoids in the diet. It often contributes more than half of total flavonoids. Tea is particularly rich in flavan-3-ols, but also provides a significant contribution to flavonols, and makes a less important contribution to other flavonoid classes. More than 40% of flavonols and more than 90% of flavan-3-ols in the diet (2). Meta-analyses of population-based studies indicate that a higher tea consumption is associated with lower risk of cardiovascular disease (3, 4). The relationships of flavonols specifically derived from tea and non-tea sources with cardiovascular outcomes has yet to be directly explored.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 5

Therefore, our primary objective was to explore the association of habitual intake of flavonols from tea and non-tea sources with the risk of atherosclerotic vascular disease mortality in a population of elderly women. Relationships for the intake of other major classes of flavonoids were also investigated.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 6

2.4 SUBJECTS AND METHODS

2.4.1 Participants

All participants were women originally recruited to a 5-year prospective, randomized, controlled trial of oral calcium supplements to prevent osteoporotic fractures (12). The women were then invited to take part in an observational follow-up study beginning in

2003: the Calcium Intake Fracture Outcome Age Related Extension Study (CARES).

All women were older than 75 y at baseline (2003) for this study. A total of 1 063 participants had complete food frequency and beverage intake data at baseline. This study was conducted according to the guidelines laid down in the Declaration of

Helsinki and all procedures involving human participants were approved by the Human

Ethics Committee of the University of Western Australia. Written informed consent was obtained from all participants.

2.4.2 Atherosclerotic vascular disease mortality

The primary outcome of interest was atherosclerotic vascular disease mortality.

Atherosclerotic vascular disease deaths data were retrieved from the Western Australian

Data Linkage System for each of the study participants from 2003 until 2008.

Atherosclerotic events were defined using diagnosis codes from the International

Classification of Diseases, Injuries and Causes of Death: Clinical Modification

(ICD-9-CM) (13) and the International Statistical Classification of Diseases and Related

Health Problems, 10th Revision, Australian Modification (ICD-10-AM) (14). These codes included ischemic heart disease including myocardial infarction (ICD-9-CM codes 410–414 and ICD-10-AM codes I20–I25); heart failure (ICD-9-CM code 428 and

ICD-10-AM code I50); cerebrovascular disease, excluding haemorrhage (ICD-9-CM codes 433–438 and ICD-10-AM codes I63–69, G45.9); and peripheral arterial disease,

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 7

(ICD-9-CM codes 440–444 and ICD-10-AM codes I70–I74). The search for atherosclerotic vascular ICD codes included all available diagnostic information that comprised parts 1 and 2 of the death certificate and the principal diagnosis in the inpatient data. All diagnosis text fields from the death certificate were used to ascertain the cause(s) of recent deaths where these data were not yet available from the Western

Australian Data Linkage System.

2.4.3 Baseline vascular disease risk assessment

The participants provided their previous medical history which was coded using the

International Classification of Primary Care – Plus method (15), as previously described in Ivey et al. (16). Previous atherosclerotic vascular disease was determined using verified hospitalisations from 1980-2003 from the Western Australian Data Linkage

System. The use of anti-hypertensive medication was also recorded. Because many participants were already on these medications, it was considered that these medications would give a better estimate of prevalent hypertension; a risk factor classically used in calculating cardiovascular risk (17). Smoking status was coded as non-smoker or ex-smoker/current smoker if they had smoked more than 1 cigarette per day for more than 3 months at any time in their life.

Physical activity was assessed using a previously validated questionnaire in which participants reported the time in hours of involvement in up to four sports, recreational activities and other forms of regular physical activity including walking, that were undertaken in the past 3 months. Energy expenditure (kJ/d) for these activities was calculated with the use of published energy costs (18, 19). Baseline weight was assessed using digital scales with participants wearing light clothes and no shoes. Baseline height

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 8 was assessed using a stadiometer, and the body mass index (BMI) was calculated in kg/m2.

2.4.4 Dietary assessment

A validated semi-quantitative food frequency questionnaire (FFQ) developed by the

Anti-Cancer Council of Victoria was used to assess baseline (2003) dietary intake (20).

The process of collection was identical, whereby a research assistant supervised the completion of the questionnaire in small groups. Food models, cups, spoons and charts for frequency were provided. Energy and nutrient intakes were estimated based on frequency of consumption and an overall estimate of usual portion size (21).

Participants also completed a beverage intake questionnaire which quantified habitual beverage consumption during the preceding year. Specifically they reported average consumption over the past 12 months of the number of cups (250 mL) per day or week of tea and coffee.

2.4.5 Flavonoid intake

Estimates of the flavonoid content of foods in the FFQ and beverage questionnaire were derived from the Flavonoid 2.1 (7), Isoflavone 2.0 (8) and Proanthocyanidin (6) food content databases developed by the United States Department of Agriculture. The method of computing flavonoid content of foods was similar to that outlined in Mink et al. (10). Specifically, for each food, we computed the sum of assessed flavonoids for each flavonoid class by summing the individual compounds of each flavonoid class, with the exception of isoflavones, where the total isoflavone value from the database was used.

The flavan-3-ol content of foods was considered to represent the average of total flavan-3-ol and proanthocyanidin monomer content (10). For foods where only Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 9 flavan-3-ol or proanthocyanidin monomer content was available, the single value provided was used to represent flavan-3-ol content. The proanthocyanidin content of foods was calculated by summing the proanthocyanidin dimers, trimers, 4-6mers,

7-10mers and polymers.

Where multiple varieties of a food listed in the FFQ were reported in the databases, the average flavonoid content of all similar varieties was computed, consistent with the descriptors used in the FFQ output. Foods in the FFQ that were not in the flavonoid databases were assumed to contain no flavonoids.

Intakes of flavonoid classes in mg/d were calculated by multiplying the estimated intake

(g edible portion/d) from FFQ and beverage questionnaire, with the flavonoid class content (mg/g edible portion) of each food item on the questionnaires.

2.4.6 Statistics

Before commencing statistical analysis, a pre-specified analytical protocol was produced. SAS (Version 9, SAS Institute Inc., Chicago, IL) was used to identify and categorise the mortality data from the Western Australian Data Linkage System. SPSS

(version 20; IBM, New York, NY) was then used for all further analyses.

Atherosclerotic vascular disease death odds ratio (OR) and 95% confidence intervals

(CI) were obtained using binary logistic regression of flavonoid intake by standard deviation (SD) scores and tertiles of consumption of each flavonoid class.

Three models were used: unadjusted; age and energy-adjusted; and multivariate-adjusted. This included pre-specified baseline risk factors significantly different in ANOVA and chi-squared test stratified by 5-year atherosclerotic vascular disease mortality, including age, energy expended in physical activity, previous atherosclerotic vascular disease, previous diabetes, and history of smoking. The Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 10 multivariate analysis included 1,008 participants due to missing data for one or more of the atherosclerotic vascular disease risk factors for 55 participants.

Stepwise logistic regression of flavonoid class intake SD score and atherosclerotic vascular disease mortality was performed. The multivariable candidate variables included age, energy intake, BMI, previous atherosclerotic vascular disease, energy expended in physical activity, previous diabetes, anti-hypertensive medication use, history of smoking, and intakes of saturated fat, fibre, protein, starch, vitamin C and alcohol at baseline. Sensitivity analysis was performed by repeating logistic regression analysis in participants without previous atherosclerotic vascular disease and diabetes at baseline.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 11

2.5 RESULTS

Over a 5-year period, 64 women died from atherosclerotic vascular disease. Participants who died from atherosclerotic vascular disease were more likely to have smoked, and have a history of atherosclerotic vascular disease and diabetes at baseline (Table 1).

Mean total intakes of individual flavonoids, classified according to chemical structure, from tea and non-tea sources, are shown in Table 2. The mean tea consumption was 2.9

(±2.0) cups/d.

2.5.1 All-source flavonoid consumption

In unadjusted and fully adjusted models, the risk [multivariate adjusted OR (95% CI) per SD] of atherosclerotic vascular disease mortality was significantly negatively associated with intake of all-source flavonol [0.54 (0.39-0.74)], flavan-3-ol [0.62

(0.46-0.85)], flavone [0.68 (0.48-0.96)], and flavanone [0.70 (0.50-0.98)]; but not with proanthocyanidin, anthocyanidin and isoflavone intakes (P>0.05).

These relationships were explored by dividing the population according to tertiles of the major flavonoid classes (Table 3). The mortality rate in the highest tertile of all-source flavonol intake was 3%, compared to 9% in the lowest tertile of consumption. A similar relationship was observed across tertiles of all-source flavan-3-ol intake.

2.5.2 Consumption of tea and non-tea flavonoids

Tea contributed 59% of total flavonoid intake, but was a major contributor to only two classes of flavonoids: flavonols (65%) and flavan-3-ols (93%) (Table 2). The intake of flavonols derived from tea and non-tea sources were associated with significantly reduced risk [multivariate adjusted OR (95% CI) per SD] of atherosclerotic vascular disease mortality: 0.63 (0.46-0.86) and 0.58 (0.41-0.81), respectively. These

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 12 relationships were explored by dividing the population according to tertiles of consumption (Table 4). Participants in the highest tertile of flavonol intake from either tea or non-tea sources had significantly lower risk of atherosclerotic vascular disease mortality compared with those in the lowest tertile. The intake of flavonols from tea and non-tea sources was not significantly correlated [r=-0.02, P=0.48].

Non-tea flavan-3-ol intake was not significantly associated with risk [multivariate adjusted OR (95% CI) per SD] of atherosclerotic vascular disease mortality: 0.89

(0.64-1.26).

2.5.3 Individual dietary flavonols

To investigate components of the relationship between flavonol intake and atherosclerotic vascular disease mortality, we repeated logistic regression analysis using the major flavonol compounds: quercetin (18±8 mg/d), myricetin (6±4 mg/d), and kaempferol (7±4 mg/d). Tea contributed 56% of total quercetin intake, and contributed over 70% of daily myricetin and kaempferol intake. Higher total quercetin [multivariate adjusted OR (95% CI) per SD: 0.52 (0.37-0.71)], myricetin [OR=0.59 (0.43-0.81)], and kaempferol [OR=0.58 (0.42-0.79)] intakes were significantly associated with reduced risk of atherosclerotic vascular disease mortality. A similar relationship was observed with consumption of quercetin [OR=0.58 (0.42-0.82)] from non-tea sources.

2.5.4 Dietary confounders

To account for diet-related potential confounders, a stepwise logistic regression model of atherosclerotic vascular disease mortality including all the baseline atherosclerotic vascular disease risk factors and dietary variables outlined in Table 1 was performed.

For all-source flavonol intake, the most parsimonious model consisted of flavonol intake SD score (P<0.01), previous atherosclerotic vascular disease (P<0.01), and age Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 13

(P<0.01). The most parsimonious model for non-tea flavonol consumption consisted of non-tea flavonol intake SD score (P<0.01), previous diabetes (P=0.02), previous atherosclerotic vascular disease (P<0.01), and age (P=0.01). The most parsimonious model for flavonol intake from tea sources consisted of flavonol intake from tea

(P<0.01), previous atherosclerotic vascular disease (P<0.01), age (P<0.01), vitamin C intake (P<0.01), and previous diabetes (P=0.05).

For all-source flavan-3-ol intake, the most parsimonious model consisted of flavan-3-ol intake SD score (P<0.01), previous atherosclerotic vascular disease (P<0.01), age

(P<0.01), vitamin C intake (P=0.01), and previous diabetes (P=0.05).

2.5.5 Sensitivity analysis

Multivariate sensitivity analysis was performed after excluding participants with previous atherosclerotic vascular disease and/or diabetes at baseline [n=839; 33 (4%) deaths from atherosclerotic vascular disease]. In this analysis, consumption of both all-source flavonols and non-tea flavonols remained significantly associated with mortality risk (data not shown). Although the estimates of risk reduction were similar for most other flavonoid classes, the relationships were attenuated such that they were no longer significant. This is likely to be due to the reduced number of deaths in this population.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 14

2.6 DISCUSSION

Higher consumption of flavonols from all, tea and non-tea sources and flavan-3-ols from all-sources were associated with reduced risk of atherosclerotic vascular disease mortality. These relationships remained after adjustment for baseline atherosclerotic vascular disease risk factors, and were independent of dietary factors. These results are consistent with benefit of flavonols and flavan-3-ols on cardiovascular health.

Participants with habitual flavonol consumption of ≥ 36 mg/d had 72% lower risk of atherosclerotic vascular disease mortality than low flavonol consumers. All three major dietary flavonols contributed to this association. A number of population studies over the past decade have now investigated flavonol intake in relation to cardiovascular outcomes (11, 22). These studies indicate that high flavonol intake is related to lower risk of cardiovascular disease (1).

The recent availability of improved food composition databases has allowed investigation of relationships of all seven classes of flavonoids with cardiovascular outcomes. We found that participants with high flavan-3-ol consumption had a 59% lower risk of atherosclerotic vascular disease mortality. However, in many populations it is difficult to explore this relationship independent of the benefits of tea consumption, which often supplies almost all of the flavan-3-ols in the diet (7).

To our knowledge, this is the first study to investigate the relationship between tea and non-tea flavonoid sources and atherosclerotic vascular disease mortality. The association of flavonol intake with atherosclerotic vascular disease mortality remained when excluding tea-derived flavonols from the analysis. We cannot rule out co-linearity between non-tea food sources of flavonols, such as apples, and dietary and lifestyle factors linked with cardiovascular disease. However, because consumption of non-tea

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 15 flavonols were not correlated with tea intake in our cohort, it appears that the cardiovascular benefits of flavonol consumption are independent of any benefits ascribable to tea, and that flavonols may contribute to the cardiovascular benefits of tea consumption. Benefits of non-tea flavonols are also supported by results of randomised controlled trials that have assessed effects of flavonols or flavonol-rich foods on mechanisms and risk factors associated with cardiovascular disease (23-25).

We did not observe a relationship between non-tea flavan-3-ols and mortality risk.

However, it is still possible that flavan-3-ols contribute to cardiovascular health. In the current population of older women, 93% of flavan-3-ol intake was derived from tea sources. It is therefore difficult to dissociate flavan-3-ol consumption from consumption of tea in this population, and the ability to observe a relationship with non-tea flavan-3-ol intake is limited. There is indirect evidence that flavan-3-ols from non-tea sources may contribute to cardiovascular health. We (26) and others (27) have demonstrated an inverse association of chocolate intake with cardiovascular outcomes.

The main flavonoids present in chocolate are flavan-3-ols and proanthocyanidins (28).

Our results suggesting that flavonols and flavan-3-ols contribute to cardiovascular health are further supported by results of randomised controlled trials investigating potential mechanisms and pathways. There is now strong evidence that flavonols, flavan-3-ols, and foods and beverages rich in these compounds, including tea (29), chocolate/cocoa (30) and apples (31), can improve endothelial function and augment nitric oxide status (32, 33). There is also evidence that these effects result in lowered blood pressure (24, 25, 34). Although less robust, there is also evidence that these compounds can influence inflammation, oxidative stress and platelet function (35). We did not observe consistent associations for other flavonoid classes. This may be due to limited intake of particular flavonoid classes and structural differences between the

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 16 different flavonoid classes. Apparently minor structural difference in flavonoids can have a large impact on their bioactivity (36).

It should be noted that the causality of the relationship between flavonol consumption and atherosclerotic vascular disease mortality cannot be established due to the observational nature of the study. Also, despite the inclusion of baseline, dietary and lifestyle risk factors into statistical models, residual or unmeasured confounders cannot be ruled out. Specifically, the potential impact of co-linearity of flavonoid intake with other potential dietary confounders such as sodium and potassium cannot be quantified, and as such further investigations of this relationship are warranted. Identification of causality is further limited by the complexity of flavonol compounds and the variability of flavonol content of foods. In particular the database used for estimation of the flavonoid content of foods is based on United States data. As such, the regional variation of flavonoid content of foods has not been accounted for in this investigation.

However, the strength of the association is such that despite these factors, the association remains significant even after adjustment for baseline, dietary and lifestyle risk factors.

In this cohort of elderly women, total flavonol and flavan-3-ol intakes were associated with reduced risk of atherosclerotic vascular disease mortality. This provides further support for the suggestion that flavonols and flavan-3-ols can contribute to cardiovascular health. The cardio-protective benefits of flavonols appear to be independent of the benefits ascribable to tea consumption, suggesting a habitual diet high in flavonols may play a role in stroke and coronary artery disease prevention.

Ultimately, in order to make public health recommendations regarding flavonol intake, further observational and intervention trials are necessary to establish the clinical benefits of flavonol consumption, independent of tea.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 17

2.7 CHAPTER 2 REFERENCES

1. Huxley RR, Neil HA. The relation between dietary flavonol intake and coronary heart disease mortality: a meta-analysis of prospective cohort studies. European Journal of Clinical Nutrition 2003;57(8):904-8. 2. Chun OK, Chung SJ, Song WO. Estimated dietary flavonoid intake and major food sources of U.S. adults. J Nutr 2007;137(5):1244-52. 3. Arab L, Liu W, Elashoff D. Green and black tea consumption and risk of stroke. Stroke 2009;40(5):1786-92. 4. Wang Z-M, Zhou B, Wang Y-S, et al. Black and green tea consumption and the risk of coronary artery disease: a meta-analysis. American Journal of Clinical Nutrition 2011;93(3):506-15. 5. Corradini E, Foglia P, Giansanti P, Gubbiotti R, Samperi R, Laganà A. Flavonoids: chemical properties and analytical methodologies of identification and quantitation in foods and plants. Nat Prod Res 2011;25(5):469-95. 6. Nutrient Data Laboratory, Agricultural Research Service, US Department of Agriculture. USDA database for proanthocyanidin content of selected foods. In: USDA, ed. Beltsville, MD, 2004. 7. US Department of Agriculture. USDA database for the flavonoid content of selected foods; release 2.1. Maryland, 2007. 8. US Department of Agriculture. USDA database for the isoflavone content of selected foods; release 2.0. Maryland, 2008. 9. Kawaguchi K, Matsumoto T, Kumazawa Y. Effects of antioxidant polyphenols on TNF-alpha-related diseases. Curr Top Med Chem 2011;11(14):1767-79. 10. Mink PJ, Scrafford CG, Barraj LM, et al. Flavonoid intake and cardiovascular disease mortality: a prospective study in postmenopausal women. American Journal of Clinical Nutrition 2007;85(3):895-909. 11. Mursu J, Voutilainen S, Nurmi T, Tuomainen T-P, Kurl S, Salonen JT. Flavonoid intake and the risk of ischaemic stroke and CVD mortality in middle-aged Finnish men: the Kuopio Ischaemic Heart Disease Risk Factor Study. British Journal of Nutrition 2008;100(4):890-5. 12. Prince RL, Devine A, Dhaliwal SS, Dick IM. Effects of calcium supplementation on clinical fracture and bone structure: results of a 5-year, double-blind, placebo-controlled trial in elderly women. Archives of Internal Medicine 2006;166(8):869-75. 13. World Health Organization. Manual of the international statistical classification of diseases, injuries, and causes of death : based on the recommendations of the ninth revision conference, 1975, and adopted by the twenty-ninth World Health Assembly. 1975 revision. ed. Geneva: World Health Organization, 1977. 14. National Centre for Classification in Health (Australia). The International statistical classification of diseases and related health problems, 10th revision, Australian modification (ICD-10-AM). 1st ed. Sydney: National Centre for Classification in Health, 1998. 15. Britt H. A new coding tool for computerised clinical systems in primary care-ICPC plus. Aust Fam Physician 1997;26 Suppl 2:S79-82. 16. Ivey KL, Lewis JR, Hodgson JM, et al. Association between yogurt, milk, and cheese consumption and common carotid artery intima-media thickness and Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 18

cardiovascular disease risk factors in elderly women. Am J Clin Nutr 2011;94(1):234-9. 17. D’Agostino RB, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care. Circulation 2008;117(6):743-53. 18. McArdle WD, Katch FI, Katch VL. Energy, nutrition and human performance. Philadelphia: Lea & Febiger, 1991. 19. Pollock ML, Wilmore JH, Fox SM. Health and fitness through physical activity. New York: Wiley, 1978. 20. Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Australia New Zealand Journal of Public Health 2000;24(6):576-83. 21. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pacific Journal of Clinical Nutrition 1994;3:19-31. 22. Keli SO, Hertog MGL, Feskens EJM, Kromhout D. Dietary flavonoids, antioxidant vitamins, and incidence of stroke: the Zutphen Study. Archives of Internal Medicine 1996;156(6):637-42. 23. Perez-Vizcaino F, Duarte J, Andriantsitohaina R. Endothelial function and cardiovascular disease: Effects of quercetin and wine polyphenols. Free Radical Research 2006;40(10):1054-65. 24. Egert S, Bosy-Westphal A, Seiberl J, et al. Quercetin reduces systolic blood pressure and plasma oxidised low-density lipoprotein concentrations in overweight subjects with a high-cardiovascular disease risk phenotype: a double-blinded, placebo-controlled cross-over study. British Journal of Nutrition 2009;102(07):1065-74. 25. Edwards RL, Lyon T, Litwin SE, Rabovsky A, Symons JD, Jalili T. Quercetin reduces blood pressure in hypertensive subjects. Journal of Nutrition 2007;137(11):2405-11. 26. Lewis JR, Prince RL, Zhu K, Devine A, Thompson PL, Hodgson JM. Habitual chocolate intake and vascular disease: a prospective study of clinical outcomes in older women. Arch Intern Med 2010;170(20):1857-8. 27. Buitrago-Lopez A, Sanderson J, Johnson L, et al. Chocolate consumption and cardiometabolic disorders: systematic review and meta-analysis. BMJ 2011;343. 28. Steinberg FM, Bearden MM, Keen CL. Cocoa and chocolate flavonoids: implications for cardiovascular health. J Am Diet Assoc 2003;103(2):215-23. 29. Ras RT, Zock PL, Draijer R. Tea consumption enhances endothelial-dependent vasodilation; a meta-analysis. PLoS One 2011;6(3.). 30. Hooper L, Kroon PA, Rimm EB, et al. Flavonoids, flavonoid-rich foods, and cardiovascular risk: a meta-analysis of randomized controlled trials. American Journal of Clinical Nutrition 2008;88(1):38-50. 31. Bondonno CP, Yang X, Croft KD, et al. Flavonoid-rich apples and nitrate-rich spinach augment nitric oxide status and improve endothelial function in healthy men and women: a randomized controlled trial. Free Radical Biology and Medicine 2002;52(1):95-102.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 19

32. Loke WM, Hodgson JM, Proudfoot JM, McKinley AJ, Puddey IB, Croft KD. Pure dietary flavonoids quercetin and (-)-epicatechin augment nitric oxide products and reduce endothelin-1 acutely in healthy men. American Journal of Clinical Nutrition 2008;88(4):1018-25. 33. Schroeter H, Heiss C, Balzer J, et al. (–)-Epicatechin mediates beneficial effects of flavanol-rich cocoa on vascular function in humans. Proceedings of the National Academy of Sciences 2006;103(4):1024-9. 34. Brown AL, Lane J, Coverly J, et al. Effects of dietary supplementation with the green tea polyphenol epigallocatechin-3-gallate on insulin resistance and associated metabolic risk factors: randomized controlled trial. British Journal of Nutrition 2009;101(06):886-94. 35. Loke W, Hodgson J, Croft K. The biochemistry behind the potential cardiovascular protection by dietary flavonoids. Edtion ed. In: Fraga CG, ed. Plant phenolics and human health: biochemistry, nutrition and pharmacology. New Jersey: John Wiley & Sons, 2009. 36. Loke WM, Proudfoot JM, Stewart S, et al. Metabolic transformation has a profound effect on anti-inflammatory activity of flavonoids such as quercetin: Lack of association between antioxidant and lipoxygenase inhibitory activity. Biochemical Pharmacology 2008;75(5):1045-53.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 20

2.8 TABLES

Chapter 2, Table 1: Baseline characteristics of the cohort stratified by atherosclerotic vascular disease mortality 1

Characteristics Participants with Participants without P value 5-year 5-year atherosclerotic atherosclerotic vascular disease vascular disease mortality mortality

Number of subjects 64 (6) 999 (94)

Atherosclerotic vascular disease risk factors History of smoking 24 (38) 259 (26) 0.04 Previous atherosclerotic vascular disease 26 (41) 155 (16) < 0.01 Previous diabetes 8 (12.5) 53 (5) 0.02 Anti-hypertensive medication use 32 (50) 405 (40) 0.14 Body mass index (kg/m2) 2 27.7 ± 6.5 27.1 ± 4.5 0.36 Age (years) 81.0 ± 2.8 80.0 ± 2.6 < 0.01 Physical activity (kJ/d) 3 327 ± 507 573 ± 653 < 0.01 Energy intake (kJ/d) 6584 ± 2 975 6898 ± 2512 0.29

Dietary intake Saturated fat (g/d) 25 ± 14 25 ± 13 0.86 Protein (g/d) 73 ± 33 78 ± 33 0.25 Starch (g/d) 88 ± 50 90 ± 38 0.57 Fibre (g/d) 20 ± 9 22 ± 8 0.04 Alcohol (mg/d) 6 ± 10 6 ± 9 0.77 Vitamin C (mg/d) 102 ± 52 129 ± 62 < 0.01

1Results are presented as mean ± SD or n (%) by ANOVA or chi-squared test where appropriate (n=1063). 2 n = 1 023. 3 n = 1 026.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 21

Chapter 2, Table 2: Baseline flavonoid class intake according to dietary source 1

Flavonoid class Compounds Major dietary Flavonoid class intake (mg/d) contributors (mg/d) All Tea Non-tea sources sources sources

Flavonol Isorhamnetin, kaempferol, Tea, pear, apple, 31 ± 14 20 ± 14 11 ± 5 myricetin, and quercetin onion

Flavan-3-ol Catechins and gallic acid Tea, green beans, 431 ± 279 401 ± 279 30 ± 32 esters of catechins, chocolate epicatechins and gallic acid esters of epicatechins, theaflavins and gallic acid esters of theaflavins, thearubigins, and proanthocyanidin monomers.

Proanthocyanidin Dimers, trimers, 4-6mers, Apple, chocolate, 215 ± 147 30 ± 21 184 ± 146 7-10mers, polymers tea, fruit juice

Flavone Apigenin and luteolin Tea, rice, orange, 3 ± 2 <1 2 ± 2 celery

Flavanone Eriodictyol, hesperetin, and Oranges, fruit 53 ± 38 9 ± 6 45 ± 38 naringenin juice, tea

Anthocyanidin Cyanidin, delphinidin, Fruit juice, 37 ± 26 <1 37 ± 26 malvidin, pelargonidin, banana, cabbage peonidin, and petunidin

Isoflavone Diadzein, genestein and Soy milk, 5 ± 6 <1 5 ± 6 glycitein breakfast cereal, tofu

1 Results are presented as mean ± SD (n=1 063).

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 22 Chapter 2, Table 3: Relationship between total flavonoid class intake groups and 5-year atherosclerotic vascular disease mortality 1 Low consumption Moderate High P value consumption consumption

Subjects [n (%)] 354 (33) 355 (33) 354 (33)

Flavonol (mg/d) < 24 24 -35 > 35 ASVD deaths [n (%)] 33 (9) 21 (6) 10 (3) Unadjusted 1.00 (reference) 0.61 (0.35-1.08) 0.28 (0.14-0.58) 0.01 Age and energy-adjusted 1.00 (reference) 0.62 (0.34-1.10) 0.29 (0.14-0.61) <0.01 Multivariate-adjusted 2 1.00 (reference) 0.58 (0.31-1.08) 0.27 (0.13-0.59) <0.01

Flavan-3-ol (mg/d) < 296 296 - 563 > 563 ASVD deaths [n (%)] 33 (9) 18 (5) 13 (4) Unadjusted 1.00 (reference) 0.52 (0.29-0.94) 0.37 (0.19-0.72) 0.01 Age and energy-adjusted 1.00 (reference) 0.53 (0.29-0.96) 0.39 (0.20-0.77) 0.01 Multivariate-adjusted 2 1.00 (reference) 0.51 (0.27-0.98) 0.34 (0.17-0.70) 0.01

Proanthocyanidin (mg/d) < 140 140 - 229 > 229 ASVD deaths [n (%)] 29 (8) 17 (5) 18 (5) Unadjusted 1.00 (reference) 0.56 (0.30-1.04) 0.60 (0.33-1.10) 0.07 Age and energy-adjusted 1.00 (reference) 0.57 (0.30-1.08) 0.63 (0.32-1.24) 0.18 Multivariate-adjusted 2 1.00 (reference) 0.61 (0.31-1.18) 0.62 (0.32-1.20) 0.22

Flavone (mg/d) < 2 2 - 3 > 3 ASVD deaths [n (%)] 32 (9) 16 (4) 16 (4) Unadjusted 1.00 (reference) 0.48 (0.26-0.88) 0.48 (0.26-0.88) 0.02 Age and energy-adjusted 1.00 (reference) 0.51 (0.27-0.96) 0.54 (0.28-1.05) 0.06 Multivariate-adjusted 2 1.00 (reference) 0.66 (0.34-1.26) 0.56 (0.29-1.12) 0.20

Flavanone (mg/d) < 32 32 - 61 > 61 ASVD deaths [n (%)] 31 (9) 18 (5) 15 (4) Unadjusted 1.00 (reference) 0.56 (0.30-1.02) 0.46 (0.24-0.87) 0.03 Age and energy-adjusted 1.00 (reference) 0.56 (0.31-1.03) 0.49 (0.26-0.96) 0.06 Multivariate-adjusted 2 1.00 (reference) 0.62 (0.33-1.19) 0.55 (0.28-1.09) 0.16

Anthocyanidin (mg/d) < 23 23 - 41 > 41 ASVD deaths [n (%)] 24 (7) 23 (6) 17 (5) Unadjusted 1.00 (reference) 0.95 (0.53-1.72) 0.69 (0.37-1.32) 0.49 Age and energy-adjusted 1.00 (reference) 1.00 (0.55-1.84) 0.71 (0.35-1.43) 0.55 Multivariate-adjusted 2 1.00 (reference) 1.09 (0.58-2.07) 0.67 (0.33-1.34) 0.35

Isoflavone (mg/d) < 3 3 - 4 > 4 ASVD deaths [n (%)] 26 (7) 16 (4) 22 (6) Unadjusted 1.00 (reference) 0.60 (0.31-1.13) 0.84 (0.46-1.50) 0.28 Age and energy-adjusted 1.00 (reference) 0.69 (0.35-1.36) 1.05 (0.52-2.09) 0.42 Multivariate-adjusted 2 1.00 (reference) 0.66 (0.33-1.33) 0.87 (0.46-1.64) 0.51 1 Results are presented as odds ratio (95% CI) from logistic regression models. (n=1 063). 2 Multivariate-adjusted model adjusted for age, previous cardiovascular disease, previous diabetes, energy expended in physical activity, and history of smoking.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 2: Page 23

Chapter 2, Table 4: Relationship between tea and non-tea flavonol intake groups and 5-year atherosclerotic vascular disease mortality 1

Low consumption Moderate consumption High consumption

Flavonol intake from tea (mg/d) < 14 14 - < 27 ≥ 27 Subjects [n (%)] 267 (25) 408 (38) 388 (36) ASVD deaths [n (%)] 23 (9) 26 (6) 15 (4) Unadjusted 1.00 (reference) 0.72 (0.40-1.29) 0.43 (0.22-0.83) Age and energy-adjusted 1.00 (reference) 0.72 (0.40-1.30) 0.45 (0.23-0.88) Multivariate-adjusted 2 1.00 (reference) 0.76 (0.40-1.42) 0.38 (0.18-0.79)

Non-tea flavonol intake (mg/d) < 8 8 - < 12 ≥ 12 Subjects [n (%)] 354 (33) 355 (33) 354 (33) ASVD deaths [n (%)] 32 (9) 19 (5) 13 (4) Unadjusted 1.00 (reference) 0.57 (0.32-1.02) 0.38 (0.20-0.74) Age and energy-adjusted 1.00 (reference) 0.56 (0.31-1.02) 0.38 (0.18-0.78) Multivariate-adjusted 2 1.00 (reference) 0.67 (0.36-1.24) 0.41 (0.20-0.85)

1 Results are presented as odds ratio (95% CI) from logistic regression models. (n=1 063). 2 Multivariate-adjusted model adjusted for age, previous cardiovascular disease, previous diabetes, energy expended in physical activity, and history of smoking.

Chapter 2: Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women Chapter 3: Page 1

Health benefits of non-nutritive food components

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 2

3.1 FOREWORD

The preceding chapter explored the beneficial association between flavonol intake from tea and non-tea sources and atherosclerotic vascular disease mortality. The vascular benefits of flavonoids extend beyond the flavonol class, and this thesis aims to explore if regular consumption of flavonoids from the proanthocyanidins class is associated with better functioning of the kidneys; a highly vascularised organ.

3.1.1 Chronic kidney disease

It is estimated that the prevalence of chronic kidney disease (CKD) in women rises dramatically with age; reaching rates of 8.8% for CKD stages 1-2 and 34.6% for CKD stages 3-5 in people over the age of 65. With a 26% increase in prevalence between

2000 and 2007, end stage kidney disease is becoming a serious public health problem

(1). This highlights the importance of strengthening the primary prevention focus for early CKD.

Overt kidney failure is associated with high cost, complex medical intervention, and a severe decline in the quality of life for affected individuals. However, there is surprisingly little information on the role nutrition plays in the pathogenesis of declining renal function in the elderly. This is especially relevant in light of the lengthening life expectancy occurring in Australia and other developed countries. Information is therefore urgently required regarding the relationship that nutrient intake has with declining kidney function. These insights can only be obtained from a detailed comprehensive prospective cohort study, for which there is currently limited data in the literature.

It is established that nutrition plays an important role in affecting survival and outcomes of patients with advanced CKD in a clinical setting. The current best practice guidelines

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 3 for dietary management of CKD patients (2) focus on maintaining biochemical parameters within normal ranges and preventing malnutrition of patients with end stage

CKD in a clinical setting. However, the extent to which these clinical approaches to diet therapy extend to individuals with earlier stages of renal dysfunction remains unclear and largely uninvestigated.

Unlike cardiovascular disease and cancer, where there is comprehensive data on dietary factors to reduce the risks of developing the disease, little is known about the dietary intake profile which can delay or prevent the onset of CKD.

3.1.2 Proanthocyanidins, vascular health and chronic kidney disease

Kidneys are highly vascularised organs, and vascular health and function is a major determinant of renal function and progression to CKD (3). Therefore, factors that improve vascular function are also likely candidates for improving renal health.

Proanthocyanidins are an oligomeric class of flavonoid compounds with demonstrated benefits to vascular health and function (4-10). Results of in vitro and animal model studies suggest that the vascular benefits of proanthocyanidin intake extend to improving renal function and outcomes (11-15). As such, we hypothesised that high habitual proanthocyanidin intake would be associated with better renal function and outcomes.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 4

3.2 ABSTRACT

PlosOne (2013. 8:8, e71166) Kerry L Ivey, Lewis JR, Lim WH, Lim EM, Hodgson JM, Prince RL

Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women

Background: Progression to chronic renal failure involves accelerated atherosclerosis and vascular calcification. Oxidative stress and endothelial dysfunction play a role in renal failure pathophysiology. In addition to improving vascular health and function, proanthocyanidins have been shown to exert renoprotective effects in animal models. Thus we hypothesize that proanthocyanidins may contribute to the maintenance of healthy renal function.

Objective: Determine the association of habitual proanthocyanidin intake with renal function and the risk of clinical renal outcomes in a population of elderly women.

Design: 948 women aged over 75 y, free of prevalent renal disease at baseline, were randomly selected from ambulant Caucasian women. Proanthocyanidin consumption was determined using a validated food frequency questionnaire and the United States Department of Agriculture proanthocyanidin food content database. Fasting serum cystatin C and creatinine were assessed at baseline. Renal failure hospitalisations and deaths were assessed over 5 years of follow-up through the Western Australia Data Linkage System.

Results: Compared to participants with low consumption, participants in the highest tertile of proanthocyanidin intake had a 9% lower cystatin C concentration (P<0.001). High proanthocyanidin consumers were at 50% lower risk of moderate chronic kidney insufficiency, and 65% lower risk of experiencing a 5-year renal disease event (P<0.05). These relationships remained significant following adjustment for renal disease risk factors and diet-related potential confounders.

Conclusion: Increased consumption of proanthocyanidins was associated with better renal function and substantially reduced renal associated events, which has been supported by mechanistic and animal model data. Proanthocyanidin intake should be further examined as a dietary contributor to better renal health.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 5

3.3 INTRODUCTION

Chronic kidney disease (CKD) represents a growing public health issue (16). The pathophysiology of CKD involves several mechanisms that are analogous to cardiovascular disease (17). Oxidative stress, atherogenesis, nitric oxide homeostasis, and endothelial function play important roles in the pathogenesis of these diseases

(18-22). Ageing is associated with structural and functional changes in the kidneys (23), resulting in impaired renal function (24). Reduced glomerular filtration rate (GFR) is a risk factor for atherosclerotic vascular disease (25, 26).

A recent meta-analysis has shown that when compared to creatinine, serum cystatin C may be a more accurate measurement of GFR (27), especially in the elderly (28).

Cystatin C provides early indications of renal dysfunction (29) and is less affected by muscle mass, weight, height, age and sex (30). Recognised risk factors for elevated cystatin C levels include traditional atherosclerotic vascular disease risk factors (31, 32).

However, the effect of dietary constituents on cystatin C levels remains uncertain.

Proanthocyanidins are a diverse group of plant-derived oligomeric compounds, and are members of the flavonoid group of molecules. There is mounting evidence that proanthocyanidins and proanthocyanidin-rich foods and beverages contribute to vascular health and reduce risk of vascular outcomes (4, 5) through acting as free radical scavengers, reducing platelet aggregation and blood pressure, and improving nitric oxide homeostasis and endothelial function (6-10). Although these vascular benefits of flavonoids are not limited to the proanthocyanidin class of flavonoids, based on mounting mechanistic and animal model data showing improved renal function and outcomes with proanthocyanidin supplementation (11-15), it appears the renoprotective benefit is limited to the specific flavonoid class of proanthocyanidins. As such, we hypothesise that via similar mechanisms, proanthocyanidins may contribute to

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 6 maintenance of healthy renal function and slows GFR decline over time. Therefore this study aimed to explore the association of habitual intake of proanthocyanidins with renal function and the risk of CKD and renal failure events in a population of elderly women.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 7

3.4 SUBJECTS AND METHODS

3.4.1 Participants

Following completion of a 5-year prospective, randomized, controlled trial of oral calcium supplements to prevent osteoporotic fractures (33), the women were then invited to take part in a follow-up study beginning in 2003: the Calcium Intake Fracture

Outcome Age Related Extension Study. At baseline (2003), the women were older than

75 y, and a total of 948 had complete proanthocyanidin, cystatin C and renal outcome data, and did not have prevalent renal disease at baseline. The Human Research Ethics

Committee of the University of Western Australia approved the study, and written informed consents were obtained from all participants.

3.4.2 Renal function

Serum was collected after an overnight fast, and serum cystatin C was quantified using a fully automated particle-enhanced immunoturbidimetric assay with Sentinel

Diagnostics reagents (Sentinel CH, Milan, Italy) on the Architect ci 16200 System

(Abbott Laboratories, Illinois, USA) according to the manufacturer instructions;

(intra-assay CV <1.5%, inter-assay CV <1%). Baseline serum creatinine was assessed in

918 participants and was analysed in 2005 using an isotope dilution mass spectrometry traceable Jaffe kinetic assay for creatinine on a Hitachi 917 analyzer (Roche Diagnostics

GmbH, Mannheim Germany).

3.4.3 Estimated glomerular filtration rate

In order to evaluate renal function, the estimated GFR (eGFR) was calculated in mL/min/1.73m2 from serum creatinine and cystatin C using the methods recently published in the New England Journal of Medicine (34). The CKD-EPI

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 8 creatinine-cystatin C equations for participants with a serum cystatin C (Scys) ≤

0.8mg/L were: serum creatinine (Scr) ≤ 0.7 mg/dl, eGFR = 130 x (scr/0.7)-0.248 x

(Scys/0.8)-0.375 x 0.995Age or Scr > 0.7 mg/dl, eGFR = 130 x (scr/0.7)-0.601 x

(Scys/0.8)-0.375 x 0.995Age. Estimated GFR for participants with a Scys > 0.8mg/L were calculated using the following equations: Scr ≤ 0.7 mg/dl, eGFR = 130 x (scr/0.7)-0.248 x

(Scys/0.8)-0.711 x 0.995Age or Scr > 0.7 mg/dl, eGFR = 130 x (scr/0.7)-0.601 x

(Scys/0.8)-0.711 x 0.995Age. Using this equation, moderate chronic kidney insufficiency was defined as eGFR < 60 mL/min/1.73m2.

3.4.4 Renal disease events

5-year incidence of acute or chronic renal failure events causing hospitalization or death was retrieved from the Western Australian Data Linkage System (WADLS) for each of the study participants from baseline. WADLS provides a complete validated record of every participant’s primary diagnosis hospitalizations and cause of death, if applicable, from the coded records of the death certificate. Renal failure events were defined using primary and additional diagnosis codes from the International Classification of Diseases,

Injuries and Causes of Death Clinical Modification (ICD-9-CM) (35) and the

International Statistical Classification of Diseases and Related Health Problems, 10th

Revision, Australian Modification (ICD-10- AM) (36). These codes included; glomerular diseases (ICD-9-CM codes 580 – 583, ICD-10- AM codes N00-08); renal tubulo-interstitial diseases (ICD-9-CM codes 593.3 – 593.5, 593.7 and 590-591,

ICD-10- AM codes N09-16); renal failure (ICD-9-CM codes 584 – 586, ICD-10- AM codes N17-19); and hypertensive renal disease (ICD-9-CM code 403, ICD-10- AM codes I12). The search for renal failure death ICD codes included all available diagnostic information that comprised Parts 1 and 2 of the death certificate and the principal diagnosis in the inpatient data. All diagnosis text fields from the death

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 9 certificate were used to ascertain the cause(s) of deaths where these data were not yet available from the WADLS.

3.4.5 Baseline chronic kidney disease risk assessment

Baseline medical histories were obtained from all participants and were coded using the

International Classification of Primary Care – Plus method (37), as previously described in Ivey et al. (38). Previous atherosclerotic vascular disease was determined using verified hospitalisations from 1980-2003 from the Western Australian Data Linkage

System. Participants maintained on anti-hypertensive medications at baseline were considered to have prevalent hypertension.

Smoking status was coded as non-smoker or ex-smoker/current smoker if they had smoked more than 1 cigarette per day for more than 3 months at any time in their life.

Baseline weight was assessed using digital scales with participants wearing light clothes and no shoes. Baseline height was assessed using a stadiometer, and the body mass index (BMI) was calculated in kg/m2.

3.4.6 Dietary assessment

Baseline dietary intake was assessed by a validated semi-quantitative food frequency questionnaire (FFQ) developed by the Anti-Cancer Council of Victoria (39). Energy and nutrient intakes were estimated based on frequency of consumption and an overall estimate of usual portion size (40). A beverage intake questionnaire (41) which quantified habitual beverage consumption during the preceding year, was also completed at baseline. Specifically participants reported average daily tea and coffee consumption over the past 12 months.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 10

3.4.7 Proanthocyanidin intake

Estimates of the proanthocyanidin content of foods in the FFQ and beverage questionnaire were derived from the Proanthocyanidin food content database (42). The method of computing proanthocyanidin content of foods was similar to that outlined in

Sesso et al. (43). Specifically, for each food, we computed the sum of assessed proanthocyanidins by summing the proanthocyanidin dimers, trimers, 4-6mers,

7-10mers and polymers. When multiple varieties of a food listed in the FFQ database were reported, the average proanthocyanidin content of all similar varieties was computed, consistent with the descriptors used in the FFQ output. Foods in the FFQ that were not in the USDA proanthocyanidin database were assumed to contain no proanthocyanidins. Intake of proanthocyanidins in mg/d was calculated by multiplying the estimated intake (g edible portion/d) from FFQ and beverage questionnaire, with the proanthocyanidin class content (mg/g edible portion) of each food item on the questionnaires.

A similar method was adopted to calculate an estimate of the intake of non-proanthocyanidin flavonoids. The Flavonoid 2.1 and Isoflavone 2.0 (44) food content databases were used to determine flavonoid content of food items, by summing mg/g edible portion for each of the individual compounds in the flavonol, flavan-3-ol, flavone, flavanone, anthocyanidin and isoflavone flavonoid classes present in each of the databases.

3.4.8 Statistics

Before commencing statistical analysis, a pre-specified analytical protocol was produced. The relationship between proanthocyanidin intake and baseline cystatin C concentration was examined in regression analysis using unadjusted and

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 11 multivariate-adjusted models. This included continuous variables energy and protein intake, BMI and age, and dichotomous variables antihypertensive use, prevalent cardiovascular disease (CVD), diabetes and history of smoking. The multivariate analysis included 932 participants due to missing data for one or more of the atherosclerotic vascular disease risk factors.

Participants were then divided into 3 groups based on tertiles of proanthocyanidin intake for further analysis by analysis of variance (ANOVA). Renal disease event odds ratio (OR) and 95% confidence intervals (CI) were obtained using binary logistic regression of flavonoid intake by standard deviation (SD) scores, and multivariable ANCOVA of tertiles of proanthocyanidin consumption.

Post hoc comparisons were only made after the main effect of the factor was found to be significant in the multivariable analyses. Stepwise linear regression of proanthocyanidin intake and cystatin C and Stepwise logistic regression of flavonoid class intake and renal disease events were used to account for potential covariance of independent variables. The multivariable candidate variables included antihypertensive use, BMI, prevalent CVD and diabetes, history of smoking, age, and intakes of energy, non-proanthocyanidin flavonoids, protein, fluid, phosphate, calcium, carbohydrate, and saturated fat at baseline. P ≤ 0.05 was the level of significance used to determine which multivariate candidate variables were included in the final model. The data were analysed using SPSS (version 15; SPSS Inc,

Chicago, IL) and SAS (Version 9, SAS Institute Inc., Chicago, IL).

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 12

3.5 RESULTS

Mean total proanthocyanidin intake was 215 ± 147 mg/d, range 18-1728 mg/d. Over 50% of total proanthocyanidin intake came from fruit (89 ± 63 mg/d), chocolate (43 ± 75 mg/d), and alcoholic beverages (32 ± 86 mg/d).

The baseline characteristics of the cohort are shown in Table 1. Mean cystatin C concentration at baseline was 1.18 (±0.29) mg/L, and over the 5-year follow-up period,

60 (6%) of participants experienced a renal disease event.

3.5.1 Renal function by serum cystatin C

The concentration of cystatin C was inversely associated with intake of proanthocyanidins; unadjusted standardised ß = -0.086, P = 0.008. This association remained significant in the fully adjusted model which included age, antihypertensive use, energy and protein intake, body mass index, prevalent atherosclerotic vascular disease and diabetes and smoking history; standardized ß = -0.134, P<0.001.

This relationship was explored further by dividing the population into tertiles of proanthocyanidin consumption (Table 2). In unadjusted and multivariate-adjusted models, participants in the highest tertile of proanthocyanidin intake had a lower cystatin C concentration than those in the lowest tertile; 7% and 9% reduction respectively. A similar trend was obtained when using serum creatinine concentration as an indicator of renal function; multivariate adjusted standardized ß = -0.073 and P =

0.052.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 13

3.5.2 Renal function by eGFR using the CKD-EPI equation (creatinine and cystatin C)

The mean eGFR of participants was 61.7 ± 13.7 ml/min/1.73m2, and 367 (39%) participants had moderate chronic kidney insufficiency at baseline, as defined by an eGFR < 60 ml/min/1.73m2. In unadjusted and fully adjusted models, the risk

[multivariate adjusted OR (95% CI) per SD] of CKD was significantly associated with intake of proanthocyanidins; 0.76 (0.63-0.91), P = 0.003. Participants in the highest tertile of proanthocyanidin intake had a 50% lower risk of having moderate chronic kidney insufficiency than those in the lowest tertile (Table 3).

3.5.3 Chronic kidney disease and clinical outcomes

Compared to the lowest tertile, participants in the highest tertile of proanthocyanidin intake were at 65% lower risk of experiencing a 5-year renal disease event (Table 3).

Five-year renal disease hospitalisation incidence was significantly different across proanthocyanidin consumption tertiles by chi-squared test (P=0.015). In the lowest tertile of proanthocyanidin consumption, there were 28 (9%) renal hospitalisations, compared to 15 (5%) and 12 (4%) in the moderate and high proanthocyanidin consumption groups, respectively. Although not significant (P=0.087), a similar trend was observed with renal disease associated mortality. There were 9 renal failure deaths

(3%) over the 5 year follow up in the lowest proanthocyanidin consumers, whereas in the moderate and high consumption groups, there were 3 (1%) and 3 (1%) deaths, respectively. The lack of significant association with renal disease mortality is likely due to lack of power to detect the association.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 14

3.5.4 Potential dietary confounders

To account for additional diet related potential confounders, a stepwise linear regression model of cystatin C concentration that included proanthocyanidin intake, and baseline renal disease risk factors and dietary intake variables outlined in Table 1 was performed.

The most parsimonious model consisted of: proanthocyanidin intake, body mass index, age, anti-hypertensive medication use, previous atherosclerotic vascular disease and intakes of non-proanthocyanidin flavonoids, saturated fat and total fluid intake.

Similarly, the addition of proanthocyanidin intake significantly improved the logistic regression model predictions for renal outcomes. In stepwise logistic regression for risk of chronic kidney disease, the most parsimonious model included proanthocyanidin intake SD score, age, antihypertensive medication use, body mass index, and previous atherosclerotic vascular disease. In stepwise logistic regression for renal event, the most parsimonious model was proanthocyanidin intake SD score, age, and histories of both atherosclerotic vascular disease and diabetes.

3.5.5 Non-proanthocyanidin flavonoids

A similar approach was used to identify the relationship between intake of other flavonoids and renal outcomes. Despite being significantly associated with cystatin C concentration (multivariate adjusted standardised ß = -0.105, P = 0.001), intake of non-proanthocyanidin flavonoids was not significantly associated with risk of moderate chronic kidney insufficiency or renal event; multivariate adjusted OR per SD = 0.899 (P

= 0.163) and multivariate adjusted OR per SD 0.783 (P=0.109), respectively.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 15

3.6 DISCUSSION

This study of elderly women is the first prospective study to suggest a potential role of proanthocyanidins in maintaining renal function and preventing renal disease events.

Total proanthocyanidin intake was beneficially associated with cystatin C concentration.

Proanthocyanidin consumption was also inversely associated with risk of moderate chronic kidney insufficiency and renal failure event in this cohort.

Total proanthocyanidin consumption of ≥ 141 mg/d was associated with a significantly better cystatin C concentration. This relationship was sufficiently robust to remain after adjustment for identified baseline and dietary risk factors. Compared to subjects with low proanthocyanidin intake, those with high consumption had a 7% lower cystatin C concentration. This difference in cystatin C concentration is likely to be of clinical significance as a 0.18 mg/L lower cystatin C concentration has been associated with a

33% lower risk of mortality (45).

The clinical significance of this relationship is further supported by our findings that participants with habitual high proanthocyanidin consumption had lower risks of moderate chronic kidney insufficiency and renal failure events. These results are reinforced by recent meta-analyses showing that moderate intake of wine (46) and chocolate (47), both rich sources of proanthocyanidins (42), are associated with reduced risk of cardiovascular disease; an independent risk factor for impaired kidney function and renal disease (48, 49). The ability of this relationship to extend to public health outcomes has been demonstrated by results of a controlled trial showing that when compared to a protein restricted diet, a polyphenol rich diet, low in carbohydrates and iron, was 40-50% more effective at reducing risk of renal events (50).

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 16

Our results showing that proanthocyanidins may contribute to renal health are supported by results of in vitro studies, studies using animal models, and randomised controlled trials in humans investigating potential mechanisms and pathways (7, 10, 11, 15, 51-57).

There is now direct evidence that proanthocyanidins specifically can improve renal health in animal models by reducing oxidative stress, improving antioxidant defence potential, and reducing oxidative renal injury (11, 15, 51). Animal models have also shown that proanthocyanidins and proanthocyanidin rich foods improve renal function and reduce apoptosis of tubular and interstitial cells (52, 53).

Another possible mechanism for renal protection by proanthocyanidins is by augmenting nitric oxide status and improving endothelial function (7, 10, 54, 55). More direct evidence that proanthocyanidins may be responsible for improved endothelial function derives from trials using flavonoid-rich cocoa (56) and grape seed extract (57).

It is important to note that proanthocyanidins with more than 2 flavonoid units are not absorbed intact. Prior to absorption, proanthocyanidins are metabolised to phenolic acids by gut bacteria (58). As such, it is likely that phenolic acid compounds are responsible for any physiological effects of proanthocyanidin consumption (59).

The level of flavonoid intake varies greatly across geographical areas (60), and to our knowledge, this is the first summary of proanthocyanidin intake in elderly Australian women. The mean proanthocyanidin intake reported in this study is greater than that reported for women from other Western countries (61, 62). However, this may be due to relatively greater flavonoid intake in Australia (63).

The paper has the following limitations. Firstly as an epidemiological paper the outcomes are based on post hoc classifications of patients and thus does not allow true randomisation. Secondly the metrics used for assessment of proanthocyanidin consumption may not truly reflect consumption of this cohort as the technique uses

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 17 analytical assays of US foods which may differ from Australian food items, and as such, the regional variation of proanthocyanidin content of foods has not been accounted for in this investigation. However the analytical data on intake were calculated before examination of the relation to clinical outcome data and was subject to rigorous covariate analysis in an attempt to identify important co-correlates that may have accounted for the observed relations. Identification of causality is further limited by the complexity of proanthocyanidin compounds and the variability of the proanthocyanidin content of foods. However, the strength of the association is such that despite these factors, the association remains significant even after adjustment for baseline, dietary and lifestyle risk factors.

To our knowledge, this is the first study to investigate the relationship between proanthocyanidin intake and renal outcomes in humans. In this cohort of elderly women, proanthocyanidin intake was associated with improved renal function and reduced risk of CKD and renal disease associated events. The renoprotective benefits of proanthocyanidins appear to be independent of traditional risk factors and dietary variables known to affect renal health, suggesting a habitual diet high in proanthocyanidins may play a role in preventing renal function decline and renal diseases. In addition to being of clinical significance, it is likely these findings will be of public health relevance, as 141 mg/d proanthocyanidin consumption is equivalent to an approximate daily intake of 50 g beans, 60 g nuts or 20 g chocolate. Ultimately, in order to make public health recommendations regarding proanthocyanidin intake, further observational and intervention trials are necessary to establish the clinical benefits on renal health.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 18

3.7 CHAPTER 3 REFERENCES

1. Australian institute of health and welfare. An overview of chronic kidney disease in Australia. In: AIHW, ed., 2009. 2. Ash S, Campbell K, MacLaughlin H, et al. Evidence based practice guidelines for the nutritional management of chronic kidney disease. Nutrition and Dietetics 2006;63:S33-S45. 3. Moe SM, Chen NX. Pathophysiology of vascular calcification in chronic kidney disease. Circulation Research 2004;95(6):560-7. 4. Mink PJ, Scrafford CG, Barraj LM, et al. Flavonoid intake and cardiovascular disease mortality: a prospective study in postmenopausal women. American Journal of Clinical Nutrition 2007;85(3):895-909. 5. McCullough ML, Peterson JJ, Patel R, Jacques PF, Shah R, Dwyer JT. Flavonoid intake and cardiovascular disease mortality in a prospective cohort of US adults. American Journal of Clinical Nutrition 2012;95(2):454-64. 6. Fine A. Oligomeric proanthocyanidin complexes: history, structure, and phytopharmaceutical applications. Alternative Medicine Review 2000;5(2):144-51. 7. Hooper L, Kroon PA, Rimm EB, et al. Flavonoids, flavonoid-rich foods, and cardiovascular risk: a meta-analysis of randomized controlled trials. American Journal of Clinical Nutrition 2008;88(1):38-50. 8. Bondonno CP, Yang X, Croft KD, et al. Flavonoid-rich apples and nitrate-rich spinach augment nitric oxide status and improve endothelial function in healthy men and women: a randomized controlled trial. Free Radical Biology and Medicine 2002;52(1):95-102. 9. Sano T, Oda E, Yamashita T, et al. Anti-thrombotic effect of proanthocyanidin, a purified ingredient of grape seed. Thrombosis Research 2005;115(1– 2):115-21. 10. Côrtes SF, Valadares YM, de Oliveira AB, Lemos VS, Barbosa MPT, Braga FC. Mechanism of endothelium-dependent vasodilation induced by a proanthocyanidin-rich fraction from Ouratea semiserrata. Planta Med 2002;68(05):412-5. 11. Yanarates O, Guven A, Sizlan A, et al. Ameliorative effects of proanthocyanidin on renal ischemia/reperfusion injury. Renal Failure 2008;30(9):931-8. 12. Nakagawa T, Yokozawa T, Satoh A, Kim HY. Attenuation of renal ischemia-reperfusion injury by proanthocyanidin-rich extract from grape seeds. Journal of nutritional science and vitaminology 2005;51(4):283-6. 13. Avramovic V, Vlahovic P, Mihailovic D, Stefanovic V. Protective effect of a bioflavonoid proanthocyanidin-BP1 in glycerol-induced acute renal failure in the rat: Renal Stereological Study. Renal Failure 1999;21(6):627-34. 14. Stefanovic V, Savic V, Vlahovic P, Cvetkovic T, Najman S, Mitic-Zlatkovic M. Reversal of experimental myoglobinuric acute renal failure with bioflavonoids from seeds of grape. Renal Failure 2000;22(3):255-66. 15. Lee YA, Kim YJ, Cho EJ, Yokozawa T. Ameliorative effects of proanthocyanidin on oxidative stress and inflammation in

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 19

streptozotocin-induced diabetic rats. Journal of Agricultural and Food Chemistry 2007;55(23):9395-400. 16. Coresh J, Selvin E, Stevens LA, et al. Prevalence of chronic kidney disease in the United States. Journal of the American Medical Association 2007;298(17):2038-47. 17. Cozzolino M, Brancaccio D, Gallieni M, Slatopolsky E. Pathogenesis of vascular calcification in chronic kidney disease. Kidney International 2005;68(2):429-36. 18. Annuk M, Zilmer M, Lind L, Linde T, Fellström B. Oxidative stress and endothelial function in chronic renal failure. Journal of the American Society of Nephrology 2001;12(12):2747-52. 19. Kielstein JT, Frölich JC, Haller H, Fliser D. ADMA (asymmetric dimethylarginine): an atherosclerotic disease mediating agent in patients with renal disease? Nephrology Dialysis Transplantation 2001;16(9):1742-5. 20. Stenvinkel P, Heimburger O, Paultre F, et al. Strong association between malnutrition, inflammation, and atherosclerosis in chronic renal failure. Kidney International 1999;55(5):1899-911. 21. Bagi Z, Hamar P, Antus B, Rosivall L, Koller A. Chronic renal failure leads to reduced flow-dependent dilation in isolated rat skeletal muscle arterioles due to lack of NO mediation. Kidney Blood Press Res 2003;26(1):19-26. 22. Liao J, Bettmann M, Sandor T, Tucker J, Coleman S, Creager M. Differential impairment of vasodilator responsiveness of peripheral resistance and conduit vessels in humans with atherosclerosis. Circulation Research 1991;68(4):1027-34. 23. Anderson S, Eldadah B, Halter JB, et al. Acute kidney injury in older adults. Journal of the American Society of Nephrology 2011;22(1):28-38. 24. Silva FG. The aging kidney: a review – part I. International Urology and Nephrology 2005;37(1):185-205. 25. Foley RN, Murray AM, Li S, et al. Chronic kidney disease and the risk for cardiovascular disease, renal replacement, and death in the United States Medicare population, 1998 to 1999. Journal of the American Society of Nephrology 2005;16(2):489-95. 26. Lewis J, Lim W, Dhaliwal S, et al. Estimated glomerular filtration rate as an independent predictor of atherosclerotic vascular disease in older women. BMC Nephrology 2012;13(1):58. 27. Dharnidharka VR, Kwon C, Stevens G. Serum cystatin C is superior to serum creatinine as a marker of kidney function: a meta-analysis. American Journal of Kidney Diseases 2002;40(2):221-6. 28. Fliser D, Ritz E. Serum cystatin C concentration as a marker of renal dysfunction in the elderly. American Journal of Kidney Diseases 2001;37(1):79-83. 29. Bökenkamp A, Herget-Rosenthal S, Bökenkamp R. Cystatin C, kidney function and cardiovascular disease. Pediatric Nephrology 2006;21(9):1223-30. 30. Filler G, Bökenkamp A, Hofmann W, Le Bricon T, Martínez-Brú C, Grubb A. Cystatin C as a marker of GFR—history, indications, and future research. Clinical Biochemistry 2005;38(1):1-8.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 20

31. Stevens LA, Schmid CH, Greene T, et al. Factors other than glomerular filtration rate affect serum cystatin C levels. Kidney International 2008;75(6):652-60. 32. Köttgen A, Selvin E, Stevens LA, Levey AS, Van Lente F, Coresh J. Serum Cystatin C in the United States: the third National Health and Nutrition Examination Survey (NHANES III). American Journal of Kidney Diseases 2008;51(3):385-94. 33. Prince RL, Devine A, Dhaliwal SS, Dick IM. Effects of calcium supplementation on clinical fracture and bone structure: results of a 5-year, double-blind, placebo-controlled trial in elderly women. Archives of Internal Medicine 2006;166(8):869-75. 34. Inker LA, Schmid CH, Tighiouart H, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. New England Journal of Medicine 2012;367(1):20-9. 35. World Health Organization. International classification of diseases, manual of the international statistical classification of diseases, injuries and causes of death: Based on the recommendations of the ninth revision conference, 1975, and adopted by the twenty-ninth World Health Assembly. Basic tabulation list with alphabetical list: World Health Organization, 1978. 36. Brämer GR. International statistical classification of diseases and related health problems. Tenth revision. World Health Statistics quarterly Rapport Trimestriel de Atatistiques Sanitaires Mondiales 1988;41(1):32. 37. Britt H. A new coding tool for computerised clinical systems in primary care-ICPC plus. Australian Family Physician 1997;26:S79. 38. Ivey KL, Lewis JR, Hodgson JM, et al. Association between yogurt, milk, and cheese consumption and common carotid artery intima-media thickness and cardiovascular disease risk factors in elderly women. American Journal of Clinical Nutrition 2011;94(1):234-9. 39. Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Australia New Zealand Journal of Public Health 2000;24(6):576-83. 40. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pacific Journal of Clinical Nutrition 1994;3:19-31. 41. Devine A, Hodgson JM, Dick IM, Prince RL. Tea drinking is associated with benefits on bone density in older women. American Journal of Clinical Nutrition 2007;86(4):1243-7. 42. Nutrient Data Laboratory, Agricultural Research Service, US Department of Agriculture. USDA database for proanthocyanidin content of selected foods. In: USDA, ed. Beltsville, MD, 2004. 43. Sesso HD, Gaziano JM, Liu S, Buring JE. Flavonoid intake and the risk of cardiovascular disease in women. American Journal of Clinical Nutrition 2003;77(6):1400-8. 44. USDA Database for the Flavonoid Content of Selected Foods, Release 2.1. U.S. Department of Agriculture.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 21

45. Shlipak MG, Katz R, Sarnak MJ, et al. Cystatin C and prognosis for cardiovascular and kidney outcomes in elderly persons without chronic kidney disease. Annals of Internal Medicine 2006;145(4):237-46. 46. Di Castelnuovo A, Rotondo S, Iacoviello L, Donati MB, de Gaetano G. Meta-analysis of wine and beer consumption in relation to vascular risk. Circulation 2002;105(24):2836-44. 47. Buitrago-Lopez A, Sanderson J, Johnson L, et al. Chocolate consumption and cardiometabolic disorders: systematic review and meta-analysis. British Medical Journal 2011;343. 48. McClellan WM, Langston RD, Presley R. Medicare patients with cardiovascular disease have a high prevalence of chronic kidney disease and a high rate of progression to end-stage renal disease. Journal of the American Society of Nephrology 2004;15(7):1912-9. 49. Elsayed EF, Tighiouart H, Griffith J, et al. Cardiovascular disease and subsequent kidney disease. Archives of Internal Medicine 2007;167(11):1130-6. 50. Facchini FS, Saylor KL. A low-iron-available, polyphenol-enriched, carbohydrate-restricted diet to slow progression of diabetic nephropathy. Diabetes 2003;52(5):1204-9. 51. Bagchi D, Bagchi M, Stohs SJ, et al. Free radicals and grape seed proanthocyanidin extract: importance in human health and disease prevention. Toxicology 2000;148(2-3):187-97. 52. Shi S, Zheng S, Zhu Y, Jia C, Xie H. Inhibitory effect of tea polyphenols on renal cell apoptosis in rat test subjects suffering from cyclosporine-induced chronic nephrotoxicity. Chinese Medical Journal 2003;116(9):1345-50. 53. Ulusoy S, Ozkan G, Yucesan FB, et al. Anti-apoptotic and anti-oxidant effects of grape seed proanthocyanidin extract in preventing cyclosporine A-induced nephropathy. Nephrology 2012;17(4):372-9. 54. Fitzpatrick DF, Bing B, Maggi DA, Fleming RC, O'Malley RM. Vasodilating procyanidins derived from grape seeds. Annals of the New York Academy of Sciences 2002;957(1):78-89. 55. Ras RT, Zock PL, Draijer R. Tea consumption enhances endothelial-dependent vasodilation; a meta-analysis. PLoS One 2011;6(3.). 56. Karim M, McCormick K, Kappagoda CT. Effects of cocoa extracts on endothelium-dependent relaxation. Journal of Nutrition 2000;130(8):2105S-8S. 57. Clifton PM. Effect of grape seed extract and quercetin on cardiovascular and endothelial parameters in high-risk subjects. Journal of Biomedicine and Biotechnology 2004;2004(5):272-8. 58. Scalbert A, Morand C, Manach C, Rémésy C. Absorption and metabolism of polyphenols in the gut and impact on health. Biomedicine and Pharmacotherapy 2002;56(6):276-82. 59. Déprez S, Brezillon C, Rabot S, et al. Polymeric proanthocyanidins are catabolized by human colonic microflora into low-molecular-weight phenolic acids. Journal of Nutrition 2000;130(11):2733-8. 60. Chun OK, Lee SG, Wang Y, Vance T, Song WO. Estimated flavonoid intake of the elderly in the United States and around the world. Journal of Nutrition in Gerontology and Geriatrics 2012;31(3):190-205.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 22

61. Zamora-Ros R, Andres-Lacueva C, Lamuela-Raventós RM, et al. Estimation of dietary sources and flavonoid intake in a Spanish adult population (EPIC-Spain). Journal of the American Dietetic Association 2010;110(3):390-8. 62. Wang Y, Chung S-J, Song WO, Chun OK. Estimation of daily proanthocyanidin intake and major food sources in the U.S. diet. Journal of Nutrition 2011;141(3):447-52. 63. Johannot L, Somerset SM. Age-related variations in flavonoid intake and sources in the Australian population. Public Health Nutrition 2006;9(8):1045-54.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 23

3.8 TABLES

Chapter 3, Table 1: Baseline, lifestyle and cardiovascular risk factors by tertiles of proanthocyanidin intake

Low intake Moderate intake High intake < 141 mg/d 141 - < 229 mg/d ≥ 229 mg/d

Number of subjects 316 (33) 316 (33) 316 (33)

Renal disease risk factors Age (years) 80 ± 3 80 ± 3 80 ± 3 History of smoking [n (%)] 78 (25) 82 (26) 95 (30) Previous ASVD [n (%)] a,b 65 (21) 40 (13) 49 (16) Previous diabetes [n (%)] 22 (7) 15 (5) 16 (5) Antihypertensive medication use [n (%)] 183 (58) 160 (51) 173 (55) Body mass index (kg/m2) 27 ± 5 27 ± 5 27 ± 4 Energy intake (kJ/d)a 5612 ± 1604 6730 ± 1867 8262 ± 3145 Protein (g/d)a 64 ± 22 77 ± 28 93 ± 41

Dietary intake Non-proanthocyanidin flavonoids (mg/d)a 371 ± 222 508 ± 246 553 ± 261 Fluid (mL/d)a 2416 ± 793 2680 ± 750 2734 ± 887 Phosphate (mg/d)a 1178 ± 359 1413 ± 439 1655 ± 612 Calcium (mg/d)a 812 ± 290 917 ± 307 997 ± 344 Sodium (mg/d)a 1703 ± 559 2015 ± 704 2348 ± 1045 Saturated fat (g/d)a 21 ± 10 24 ± 11 30 ± 16 Carbohydrate (g/d)a 149 ± 43 180 ± 47 217 ± 80

Results are mean ± SD or n (%) where appropriate. (n = 948) a represents significantly different (P<0.05) by ANOVA or chi-squared test where appropriate. b ASVD: atherosclerotic vascular disease

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 24

Chapter 3, Table 2: Baseline cystatin C concentration according to groups of proanthocyanidin intake

Low intake Moderate intake High intake P value < 141 mg/d 141 - < 229 mg/d ≥ 229 mg/d

Number of subjects 316 (33) 316 (33) 316 (33)

Cystatin C (mmol/L) Unadjusted1 1.23 ± 0.02a 1.17 ± 0.02b 1.14 ± 0.02b 0.001 Age-adjusted2 1.23 ± 0.02a 1.17 ± 0.02b 1.14 ± 0.02b <0.001 Multivariate-adjusted2 1.24 ± 0.02a 1.18 ± 0.02b 1.13 ± 0.02c <0.001

1 Results are mean ± SEM. 2 Results are least-squared mean ± SEM by ANCOVA. a,b,c represents significantly different by LSD (P<0.05). Multivariate-adjusted model: antihypertensive use, energy and protein intake, BMI, prevalent CVD and diabetes, history of smoking and age.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 3: Page 25

Chapter 3, Table 3: Relationship between proanthocyanidin intake and 5-year hospitalisation or death renal failure events

Low intake Moderate intake High intake P value < 141 mg/d 141 - < 229 mg/d ≥ 229 mg/d

Number of subjects 316 (33) 316 (33) 316 (33) eGFR < 60 ml/min/1.73m2 [n(%)] 151 (49) 117 (38) 99 (33) Unadjusted 1.00 (referent) 0.64 (0.47-0.89)* 0.50 (0.36-0.70)* <0.001 Age-adjusted 1.00 (referent) 0.63 (0.45-0.87)* 0.48 (0.34-0.67)* <0.001 Multivariate-adjusted 1.00 (referent) 0.61 (0.43-0.87)* 0.44 (0.30-0.65)* <0.001

Renal failure events [n(%)] 32 (10) 16 (5) 12 (4) Unadjusted 1.00 (referent) 0.47 (0.25-0.88)* 0.35 (0.18-0.69)* 0.004 Age-adjusted 1.00 (referent) 0.47 (0.25-0.88)* 0.35 (0.17-0.68)* 0.003 Multivariate-adjusted 1.00 (referent) 0.53 (0.27-1.05) 0.40 (0.18-0.89)* 0.048

Results are OR (95% CI) by logistic regression. *represents significantly different from referent (P<0.05). Multivariate-adjusted model: antihypertensive use, energy and protein intake, BMI, prevalent CVD and diabetes, history of smoking and age.

Chapter 3: Associations of proanthocyanidin intake with renal function and clinical outcomes in elderly women Chapter 4: Page 1

Health benefits of non-nutritive food components

Chapter 4: Flavonoid intake and all-cause mortality

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 2 4.1 FOREWORD

Chapters 2 and 3 explored the association of individual flavonoid classes (flavonols and proanthocyanidins) with individual disease outcomes (atherosclerotic vascular disease and chronic kidney disease). However, dietary patterns and disease states are complex. Flavonoid classes are not consumed in isolation, and causes of mortality are often multifactorial in nature. With this in mind, we explored the association of total flavonoid intake from all classes with mortality from any cause.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 3 4.2 ABSTRACT

American Journal of Clinical Nutrition (under review) Kerry L Ivey, Hodgson JM, Croft KD, Lewis JR, Prince RL

Flavonoid intake and all-cause mortality

Background: Flavonoids are bioactive compounds found in foods such as tea, chocolate, red wine, fruit and vegetables. Higher intakes of specific flavonoids and flavonoid rich foods have been linked to reduced mortality from specific vascular diseases and cancers. However, the importance of flavonoids in preventing all-cause mortality remains uncertain.

Objective: To explore the association between flavonoid intake and risk of 5-year mortality from all causes, using two comprehensive food composition databases to assess flavonoid intake.

Design: The study population included 1 063 randomly selected women aged over 75 y. All-cause, cancer and cardiovascular mortalities were assessed over 5-years of follow-up through the Western Australia Data Linkage System.

Two estimates of flavonoid intake (total-flavonoidUSDA and total-flavonoidPE) were determined using food composition data from the United States Department of Agriculture (USDA), and the Phenol-Explorer (PE) databases.

Results: During the 5-year follow-up period, 129 (12%) deaths were documented. Participants with high total-flavonoid intake were at lower risk [multivariate-adjusted HR (95% CI)] of 5-year all-cause mortality when compared to those with low levels of total-flavonoid consumption; total-flavonoidUSDA HR: 0.37 (0.22,0.58) and total-flavonoidPE HR: 0.36 (0.22,0.60). Similar beneficial relationships were observed for both cardiovascular disease mortality [total-flavonoidUSDA HR: 0.34 (0.17,0.69), flavonoidPE HR: 0.32 (0.16,0.61)] and cancer mortality [total-flavonoidUSDA HR: 0.25 (0.10,0.62), flavonoidPE HR: 0.26 (0.11,0.62)].

Conclusions: Using the most comprehensive flavonoid databases, we provide evidence that flavonoids may reduce mortality in older women. The benefits of flavonoids may extend to the etiology of cancer and cardiovascular disease.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 4 4.3 INTRODUCTION

Flavonoids are the largest dietary polyphenolic class. They represent a diverse group of over 4 000 different biologically active compounds synthesized during plant metabolism

(1, 2). Flavonoids can be divided up into 6 main flavonoid classes: Flavonols,

Flavan-3-ols, Flavones, Flavanones, Anthocyanins, and Isoflavones (3). A variety of biological effects have been identified for individual flavonoids, but the value of considering the class as a whole is that this can provide a biologically plausible way of considering the health effects of flavonoids with similar structural characteristics.

Following consumption, flavonoids may contribute to a variety of beneficial biological activities in humans. There is now strong and consistent data showing that flavonoids can maintain and augment nitric oxide status, and improve endothelial function (4-7).

There is also evidence that these compounds can influence blood pressure, oxidative damage, inflammation, platelet function and thrombosis, blood lipids and glucose metabolism (8, 9). These effects may help to explain the findings that flavonoids and flavonoid rich foods exhibit cardioprotective and antineoplastic properties (10-16).

However, despite being associated with all-cause cardiovascular disease mortality (12), flavonoid class intake has not been found to be associated with all-cause cancer mortality (17).

Particularly high dietary sources of flavonoids include tea (18), chocolate (19), fruit (20), and red wine (1). Meta-analyses of population based studies have found that consumption of these flavonoid- rich foods are associated with reduced risk of cancer

(21, 22) and cardiovascular disease (23, 24), which are the two leading causes of mortality (25). These whole foods have also been shown in prospective cohort studies to be associated with reduced risk of all-cause mortality (26-29). However, the role that intake of flavonoids specifically play in prevention of all-cause mortality is less clear.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 5 Previous studies exploring this hypothesis (30, 31), have been limited by the lack of comprehensive food composition data. In recent years, the quality food composition databases have improved exponentially, both in terms of the flavonoid compounds and food items they provide data on. The emergence and continued development of these databases the opportunity to explore the relationship of all-cause mortality and intake of total flavonoids and 6 common flavonoid classes in foods (12, 32). The two most comprehensive databases describing the flavonoid content of foods are the United States

Department of Agriculture (USDA) (33-35) and the Phenol-Explorer (PE) (36) food content databases. This study aims to use this data to explore the pre-specified hypothesis that total-flavonoid intake is beneficially associated with risk of all-cause mortality in a population of elderly women.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 6 4.4 SUBJECTS AND METHODS

4.4.1 Participants

1 136 postmenopausal women above the age of 75 years were recruited into the

Calcium Intake Fracture Outcome Age Related Extension Study in 2003. These participants had previously completed a 5-year prospective, randomized, controlled trial of oral calcium supplements to prevent osteoporotic fractures (37). Although from higher socioeconomic groups, participants in this cohort had similar disease burden and pharmaceutical consumption to whole populations of this age (38). A total of 1 063 participants had complete food frequency and beverage intake data at baseline. This study was approved by the Human Ethics Committee of the University of Western

Australia, and written informed consent was obtained from all participants.

4.4.2 Mortality

The primary outcome of interest was all-cause mortality. The mortality data from baseline (2003) until year-5 (2008) was retrieved from the Western Australian Data

Linkage System in 100% of participants.

Cardiovascular and cancer events were defined using diagnosis codes from the

International Classification of Diseases, Injuries and Causes of Death: Clinical

Modification (ICD-9-CM)(39) and the International Statistical Classification of

Diseases and Related Health Problems, 10th Revision, Australian Modification

(ICD-10-AM) (40). Cardiovascular codes included ICD9 390-459 and ICD10 I00-I99.

Cancer codes included ICD 9 140-239 and ICD 10 C00-C97. The search for cardiovascular and cancer ICD codes from the Western Australian Data Linkage System mortality registry included all available diagnostic information that comprised parts 1 and 2 of the mortality records. All diagnosis text fields from the death certificate were

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 7 used to ascertain the cause(s) of recent deaths where these data were not yet available from the Western Australian Data Linkage System.

4.4.3 Baseline risk assessment

Using the aforementioned ICD codes, previous cardiovascular disease and cancer at baseline was determined using verified hospitalizations from 1980-2003 from the

Western Australian Data Linkage System. Smoking status was coded as currently smoking or not-currently smoking if participants reported smoking at least 1 cigarette per day for a minimum of 3 months preceding their baseline visit.

Physical activity was assessed using a questionnaire in which participants reported the time of involvement for up to four physical activities of moderate intensity undertaken in the preceding 3 months (41, 42). Baseline weight was assessed using digital scales with participants wearing light clothes and no shoes. Baseline height was assessed using a stadiometer, and the body mass index (BMI) was calculated in kg/m2.

4.4.4 Dietary assessment

Baseline (2003) dietary intake was assessed using a validated semi-quantitative food frequency questionnaire (FFQ) developed by the Anti-Cancer Council of Victoria

(43-45). Energy and nutrient intakes were estimated based on frequency of consumption and an overall estimate of usual portion size (46). A beverage questionnaire was used to assess average tea and coffee consumption over the past 12 months.

4.4.5 Flavonoid intake

There are two major sources summarizing the flavonoid content of foods: the

Phenol-Explorer (PE) (36), and the United States Department of Agriculture (USDA)

Flavonoid 2.1 (33), Isoflavone 2.0 (34) and Proanthocyanidin (35) food content databases. Both of these databases were used to derive two separate estimates of

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 8 total-flavonoid intake: flavonoid intake based on food composition data from the USDA

(FlavonoidUSDA) and flavonoid intake based on food composition data from the PE

(FlavonoidPE).

The method of computing flavonoid content of foods has been are similar to those described in Ivey et al. (47), and extraction procedures for the 2 databases were identical.

Specifically, for each food, we computed the sum of assessed flavonoids for each flavonoid class by summing the individual compounds of each flavonoid class in the form expressed in each individual database. The exception was the IsoflavoneUSDA data, where the total Isoflavone value from the USDA database was used rather than the individual compounds.

The terminology and classification systems used by each database varied; therefore, a standardized classification system was adopted throughout this study. The terms

Anthocyanins and Isoflavones in this study refer to the Phenol-Explorer Anthocyanin and Isoflavonoid classes respectively. The term Flavanol in this study encapsulates both the Flavan-3-ol and Proanthocyanin classes in the USDA database. As such,

FlavanolUSDA represented the sum of Proanthocyanin polymer values in conjunction with total flavan-3-ol content.

The chalcone, dihydrochalcone and dihydroflavonol content of foods are described in the PE but not the USDA database. As these compounds are typically considered a precursor to many flavonoid compounds, and not a flavonoid specifically, they were not used to estimate total-flavonoid intake. They were instead summed and investigated for a relationship with all-cause mortality in a separate exploratory analysis.

Intakes of flavonoid classes in mg/d were calculated by multiplying the estimated intake

(g edible portion/d) from the FFQ and beverage questionnaires, with the flavonoid class content (mg/g edible portion) of each food item on the questionnaires. Where multiple varieties of a food listed in the FFQ were reported in the databases, the average

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 9 flavonoid content of all similar varieties was computed, consistent with the descriptors used in the FFQ output. For each database, foods in the FFQ that were not represented were entered as zero values for analysis for that particular database. Of the 93 individual food items specified in the FFQ, 19 foods in the USDA computation and 47 in the PE computation were recorded as not containing flavonoids in the respective database, or did not appear in the database. This includes foods such as Vegemite, butter, dairy products and sugar.

4.4.6 Statistics

Before commencing statistical analysis, a pre-specified analytical protocol was produced. SAS (Version 9, SAS Institute Inc., Chicago, IL) was used to identify and categorize the hospital and mortality data from the Western Australian Data Linkage

System. SPSS (Version 20; IBM, New York, NY) was then used for all further analyses.

Five-year all-cause mortality Hazard ratios (HR) and 95% confidence intervals (CI) were obtained using Cox regression of total-flavonoid intake by standard deviation (SD) scores and tertiles of total-flavonoid intake, using unadjusted and multivariate-adjusted models.

The multivariate-adjusted model included: age, prevalent cardiovascular disease, prevalent cancer, as well as the presence of assessed all-cause mortality risk factors identified by the World Health Organization (48). These included: overweight or obesity

(BMI > 25 kg/m2), low fruit and vegetable intake (less than 600g of total fruit and vegetable intake per day), physical inactivity (less than 2.5 hours of moderate-intensity physical activity per week), current cigarette smoker, and current alcohol consumer. The multivariate-adjusted model included 1 022 participants, due to missing variables in 41 women.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 10 Post-hoc analyses using tertiles of flavonoid-class intake was performed to identify which flavonoid-classes may contribute to the observed relationships. In order to address a potential source of outcome bias, we conducted a Cox regression analysis of total-flavonoid intake with the two major mortality causes, CVD and cancer (25). In order to take into account multiple comparisons, the level of significance for post-hoc comparisons was set at P<0.0056.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 11 4.5 RESULTS

The mean age of participants was 80 (±3) years, with a mean body mass index of 27 (±5) kg/m2 (Table 1). Daily consumption of alcohol 8 (±10) mg/d, and mean fruit and vegetable intake was 520 (±206) g/d. Mean total energy consumption was 1,643 (±607) kJ/d. Mean daily total flavonoid consumption was 696 ± 322 mg (median, IQR =

668,468-889) and 674 ± 326 mg (median, IQR = 648, 449-872) as computed by the

USDA and PE databases, respectively. During the 5-year follow-up period, 129 (12%) deaths were documented.

4.5.1 Total-flavonoid intake and all-cause mortality

Total flavonoid intake was associated with a lower risk of all-cause mortality in unadjusted analyses; FlavonoidUSDA unadjusted HR per SD = 0.64 (0.52,0.78),

FlavonoidPE unadjusted HR per SD = 0.66 (0.54,0.80). The association remained significant following multivariate adjustment; FlavonoidUSDA multivariate adjusted HR per SD = 0.67 (0.54,0.82), FlavonoidPE multivariate-adjusted HR per SD = 0.68

(0.55,0.84).

To explore this association further we trichotomised the cohort into groups based on tertiles of total-flavonoid intake. Irrespective of food composition database used to assess total-flavonoid intake, high consumers were at lower risk of all-cause mortality

(Table 2), and there were 62% fewer deaths in high compared to low total flavonoid consumers (Figure 1 and 2).

Analysis was repeated in 1 055 participants with estimated energy intake between 500 and 4000 kcal/d. Excluding participants with extreme dietary intake estimates did not alter the relationship between total-flavonoid intake and all-cause mortality;

FlavonoidUSDA multivariate adjusted HR per SD = 0.67 (0.54,0.83), FlavonoidPE multivariate-adjusted HR per SD = 0.68 (0.55,0.84). Similarly, exclusion the 13

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 12 participants who died in the first year of the study did not ameliorate the total-flavonoid and all-cause mortality relationship; FlavonoidUSDA multivariate adjusted HR per SD =

0.67 (0.54,0.84), FlavonoidPE multivariate-adjusted HR per SD = 0.69 (0.55,0.86).

4.5.2 Flavonoid class intake and all-cause mortality

Flavanols, derived mainly from consumption of black tea, were the major contributor to total flavonoid intake, comprising 82% (FlavanolUSDA) and 59% (FlavanolPE) of total daily flavonoid consumption. Flavonols, derived mainly from the consumption of tea, apples, pears and onions provided 4% and 26% of total USDA and PE intakes, respectively. Oranges and fruit juice were the predominant dietary sources of

Flavanones; FlavanoneUSDA: 8%, FlavanonePE: 8%. A total of less than 8 mg/d was supplied from the remaining classes Anthocyanins, Isoflavones and Flavones.

In order to identify which flavonoid classes contribute to the total-flavonoid and all-cause mortality relationship, the Cox regression analysis was repeated using tertiles of flavonoid class intake.

When compared to low consumers as the referent group, high consumers of Flavanols reduced risk (multivariate-adjusted HR for high tertile) of all-cause mortality;

FlavanolUSDA HR: 0.43 (0.26,0.70) P=0.001, FlavanolPE HR: 0.37 (0.23,0.61) P<0.001.

Similarly, compared to low consumers, high Flavonol consumers were at lower risk of all-cause mortality; FlavonolUSDA HR: 0.39 (0.23,0.65) P<0.001, FlavonolPE HR: 0.43

(0.27,0.70) P=0.001. A similar association was observed with tertiles of FlavonePE consumption; 0.51 (0.32,0.82) P=0.005. However the relationship with FlavoneUSDA intake was ameliorated following multivariate-adjustment; 0.60 (0.35,1.02) P=0.057.

In multivariate-adjusted models, consumption of Flavanones, Anthocyanins and

Isoflavones were not associated with all-cause mortality (P>0.05). In post-hoc

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 13 exploratory analysis, intake of the chalcone, dihydrochalcone and dihydroflavonol combined classes was also not associated with all-cause mortality.

4.5.3 Total flavonoid intake and mortality from cancer and cardiovascular disease

There were 78 (60%) deaths attributed to CVD and 49 (38%) cancer mortalities over the

5-year follow-up period. In order to explore the contribution of specific disease processes to the all-cause mortality relationship, we performed Cox regression analysis with total-flavonoid intake groups and mortality sub-types. There was at least 64% fewer CVD mortalities and 73% fewer cancer mortalities in the high compared to the low total flavonoid consumers (Table 3).

Of the 129 deaths, both CVD and cancer appeared on the death certificate in 17 (13%) of cases. In order to account for potential confounding, we repeated the analysis excluding these participants, which did not alter observed relationships (data not shown).

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 14 4.6 DISCUSSION

This prospective cohort study found that participants with higher total flavonoid consumption were at reduced risk of all-cause mortality. Despite differences in intake estimates between the PE and USDA databases, the relationships were similar and remained after adjustment for pre-specified mortality risk factors, and were irrespective of the database used to estimate flavonoid intake. The highest compared to the lowest tertile of flavonoid intake was associated with ~60% lower risk of all-cause mortality.

Furthermore, a higher flavonoid intake of ~350 mg, equivalent to approximately 2 cups of tea, was associated with ~40% lower risk of all-cause mortality. The protective relationship extended to both major causes of mortality. High total-flavonoid consumers displayed a 40–50% reduced risk of CVD and cancer mortality, compared to those with the lowest intake.

When exploring contributors to this relationship, the association with all-cause mortality appeared to be class-dependent. Those with high intakes of Flavanols and Flavonols experienced a lower mortality rate, whereas associations for the remaining flavonoid-classes appeared to be attenuated by adjustment for pre-specified risk factors.

The varying structures and bioactivities of the different flavonoid classes may help explain the lack of concurrency across different flavonoid classes (49, 50). However, the lack of beneficial association for the Flavone, Isoflavone, and Anthocyanin classes may also be explained by their low intake levels, suggesting the dose may be inadequate to confer health benefits in this population. This dosing effect may also explain the lack of association of total-flavonoid intake and mortality in the study by Mink and colleagues

(12), as the median flavonoid intake in the Iowa Women’s Health Study (51) of 239.2 mg/d was nearly 3 times lower than the 696 mg/d median flavonoid intake assessed in this cohort using USDA data. These intake estimates were based on USDA flavonoid and isoflavone databases that preceded the databases used in the current study.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 15 Therefore, in addition to differing dietary patterns amongst different geographical locations, variations in food content databases may also explain variations in flavonoid intake observed between studies. The differences in database utilization may also explain the differences in flavonoid intake estimates between our study and other

Australian data (52). Phenol-Explorer intake estimates in this cohort are larger than another study utilizing PE data in a younger French population (53). This difference between reported intakes is likely explained by the age and geographic differences between the cohorts (54).

This study was conducted in elderly postmenopausal women, where relationships between flavonoid intake and all-cause mortality were quite strong. Using a similar postmenopausal population, the Iowa Women’s Health Study also found similar associations of flavonoid intake and mortality (12). However, these findings have not been reproduced in elderly men (55), or in women and men of a younger age demographic (30), suggesting the benefits of habitual flavonoid consumption may be limited to postmenopausal women.

Our findings of a cardioprotective association with flavonoids are generally consistent with the findings of previous studies of flavonoids and CVD (56-58). Intervention studies have shown that flavonoids and flavonoid rich foods can improve CVD risk factors such as blood pressure (15, 16, 59), endothelial function (4, 5), and platelet formation (60). It is likely that these protective functions explain the benefits of flavonoids in protecting against subcategories of CVD including coronary heart disease

(61), stroke (62), and myocardial infarction (63). Further support for a vascular benefit of flavonoids come from our previous data showing an inverse association between the intake of Flavonols and atherosclerotic vascular disease, a subcategory of CVD (47).

To our knowledge, this is the first study to identify a beneficial relationship of total-flavonoid intake with cancer mortality. In fact, there are few previous studies of

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 16 total-flavonoid intake and all-cause cancer mortality, which to date have not observed a beneficial association (17, 64, 65). The lack of agreement of our findings with those from Knekt (65) and Hertog (17, 64) may be due to lower flavonoid intakes in some populations or the use of earlier versions of food composition databases. Similarly, the lack of association with all-cause cancer mortality observed by Cutler and colleagues

(51) is most likely explained by the aforementioned low total-flavonoid consumption in their cohort. Despite the lack of agreement in current all-cause cancer data, although not conclusive (51), there is epidemiological and in vitro data to suggest the protective role of flavonoids on particular cancer types such as lung cancer (66-69) and breast cancer

(70-72). There are also data linking intake of high flavonoid foods such as green tea with particular cancer types (73).

The flavonoid content of foods are affected by variety, growing conditions, and maturation, processing, storage (74-76), making flavonoid intake particularly difficult to assess. This issue is further exacerbated by the inability of the FFQ utilized in this study to identify all food items expressed in the databases. Our data were further limited by the strong likelihood of some degree of measurement error in flavonoid exposure, as evidenced by the discrepancies in intake estimates derived from the USDA and PE databases. This type of measurement error would typically tend to weaken our findings, and increase the likelihood of type 2 statistical errors. However, as the flavonoid intake estimates were based on the same frequency of food consumption data, the bias is introduced only because of the databases, and not because of measurement error in the food intake assessment. This highlights the major strength of this study, where we utilized the two most comprehensive food composition databases to derive two separate estimates of flavonoid intake in order to minimize bias in flavonoid intake estimations and to confirm our findings.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 17 Thearubigins are present in the USDA database, but not in the PE. The contribution of thearubigins to total-flavonoidUSDA intake is likely largely describes the discrepancies between FlavanolUSDA and FlavanolPE estimates. Similarly, the larger Flavonol intake estimates by the PE database are likely to arise primarily due to the expression of food composition data in glycoside and aglycoside Flavonol forms, whereas the USDA database provides corrected aglycoside data only. There is the possibility that due to the differences in USDA and PE food composition data, the subsequent different flavonoid intake estimates based on this data may have different relationships with outcome variables. This highlights the importance of utilizing two distinctly different databases when estimating flavonoid intake. The strength of this approach is that despite the many differences between the two databases, we observed similar relationships with all-cause mortality.

In summary, we found a beneficial relation between the dietary intake of total-flavonoids and the risk of mortality from all-causes, which was attributable primarily to the consumption of Flavanols and Flavonols, the main sources of flavonoids in the present population. It appears the benefits of flavonoids extend to the etiology of both cancer and cardiovascular disease.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 18

4.7 CHAPTER 4 REFERENCES

1. Cook N, Samman S. Flavonoids: chemistry, metabolism, cardioprotective effects, and dietary sources. Journal of Nutritional Biochemistry 1996;7(2):66-76. 2. Das DK. Naturally occurring flavonoids: Structure, chemistry, and high-performance liquid chromatography methods for separation and characterization. Edtion ed. In: Lester P, ed. Methods in Enzymology: Academic Press, 1994:410-20. 3. Aherne SA, O’Brien NM. Dietary flavonols: chemistry, food content, and metabolism. Nutrition 2002;18(1):75-81. 4. Loke WM, Hodgson JM, Proudfoot JM, McKinley AJ, Puddey IB, Croft KD. Pure dietary flavonoids quercetin and (-)-epicatechin augment nitric oxide products and reduce endothelin-1 acutely in healthy men. American Journal of Clinical Nutrition 2008;88(4):1018-25. 5. Schroeter H, Heiss C, Balzer J, et al. (–)-Epicatechin mediates beneficial effects of flavanol-rich cocoa on vascular function in humans. Proceedings of the National Academy of Sciences 2006;103(4):1024-9. 6. Karim M, McCormick K, Kappagoda CT. Effects of cocoa extracts on endothelium-dependent relaxation. Journal of Nutrition 2000;130(8):2105S-8S. 7. Clifton PM. Effect of grape seed extract and quercetin on cardiovascular and endothelial parameters in high-risk subjects. Journal of Biomedicine and Biotechnology 2004;2004(5):272-8. 8. Hodgson JM, Croft KD. Tea flavonoids and cardiovascular health. Molecular Aspects of Medicine 2010;31(6):495-502. 9. Rice-Evans C. Flavonoid antioxidants. Current medicinal chemistry 2001;8(7):797-807. 10. Androutsopoulos VP, Mahale S, Arroo RRJ, Potter G. Anticancer effects of the flavonoid diosmetin on cell cycle progression and proliferation of MDA-MB 468 breast cancer cells due to CYP1 activation. Oncology Reports 2009;21(6):1525-8. 11. Pettit GR, Hoard MS, Doubek DL, et al. Antineoplastic agents 338. The cancer cell growth inhibitory. Constituents of Terminalia arjuna (Combretaceae). Journal of Ethnopharmacology 1996;53(2):57-63. 12. Mink PJ, Scrafford CG, Barraj LM, et al. Flavonoid intake and cardiovascular disease mortality: a prospective study in postmenopausal women. American Journal of Clinical Nutrition 2007;85(3):895-909. 13. Mursu J, Voutilainen S, Nurmi T, Tuomainen T-P, Kurl S, Salonen JT. Flavonoid intake and the risk of ischaemic stroke and CVD mortality in middle-aged Finnish men: the Kuopio Ischaemic Heart Disease Risk Factor Study. British Journal of Nutrition 2008;100(4):890-5. 14. Perez-Vizcaino F, Duarte J, Andriantsitohaina R. Endothelial function and cardiovascular disease: Effects of quercetin and wine polyphenols. Free Radical Research 2006;40(10):1054-65. 15. Egert S, Bosy-Westphal A, Seiberl J, et al. Quercetin reduces systolic blood pressure and plasma oxidised low-density lipoprotein concentrations in overweight subjects with a high-cardiovascular disease risk phenotype: a

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 19 double-blinded, placebo-controlled cross-over study. British Journal of Nutrition 2009;102(07):1065-74. 16. Edwards RL, Lyon T, Litwin SE, Rabovsky A, Symons JD, Jalili T. Quercetin reduces blood pressure in hypertensive subjects. Journal of Nutrition 2007;137(11):2405-11. 17. Hertog MG, Feskens EJ, Hollman PC, Katan MB, Kromhout D. Dietary flavonoids and cancer risk in the Zutphen Elderly Study. Nutrition and Cancer 1994;22(2):175-84. 18. Ras RT, Zock PL, Draijer R. Tea consumption enhances endothelial-dependent vasodilation; a meta-analysis. PLoS One 2011;6(3.). 19. Hooper L, Kroon PA, Rimm EB, et al. Flavonoids, flavonoid-rich foods, and cardiovascular risk: a meta-analysis of randomized controlled trials. American Journal of Clinical Nutrition 2008;88(1):38-50. 20. Bondonno CP, Yang X, Croft KD, et al. Flavonoid-rich apples and nitrate-rich spinach augment nitric oxide status and improve endothelial function in healthy men and women: a randomized controlled trial. Free Radical Biology and Medicine 2002;52(1):95-102. 21. Sun C-L, Yuan J-M, Koh W-P, Mimi CY. Green tea, black tea and breast cancer risk: a meta-analysis of epidemiological studies. Carcinogenesis 2006;27(7):1310-5. 22. Tang N, Wu Y, Zhou B, Wang B, Yu R. Green tea, black tea consumption and risk of lung cancer: a meta-analysis. Lung Cancer 2009;65(3):274-83. 23. Arab L, Liu W, Elashoff D. Green and black tea consumption and risk of stroke. Stroke 2009;40(5):1786-92. 24. Wang Z-M, Zhou B, Wang Y-S, et al. Black and green tea consumption and the risk of coronary artery disease: a meta-analysis. American Journal of Clinical Nutrition 2011;93(3):506-15. 25. Yach D, Hawkes C, Gould CL, Hofman KJ. The global burden of chronic diseases. JAMA: The Journal of the American Medical Association 2004;291(21):2616-22. 26. Kuriyama S, Shimazu T, Ohmori K, et al. Green tea consumption and mortality due to cardiovascular disease, cancer, and all causes in Japan. JAMA: The Journal of the American Medical Association 2006;296(10):1255-65. 27. Buijsse B, Feskens EJ, Kok FJ, Kromhout D. Cocoa intake, blood pressure, and cardiovascular mortality: the Zutphen Elderly Study. Archives of Internal Medicine 2006;166(4):411. 28. Grønbæk M, Becker U, Johansen D, et al. Type of alcohol consumed and mortality from all causes, coronary heart disease, and cancer. Annals of Internal Medicine 2000;133(6):411-9. 29. Steffen LM, Jacobs DR, Stevens J, Shahar E, Carithers T, Folsom AR. Associations of whole-grain, refined-grain, and fruit and vegetable consumption with risks of all-cause mortality and incident coronary artery disease and ischemic stroke: the Atherosclerosis Risk in Communities (ARIC) Study. The American Journal of Clinical Nutrition 2003;78(3):383-90. 30. Knekt P, Jarvinen R, Reunanen A, Maatela J. Flavonoid intake and coronary mortality in Finland: a cohort study. British Medical Journal 1996;312:478-81. 31. Hertog MG, Kromhout D, Aravanis C, et al. Flavonoid intake and long-term risk of coronary heart disease and cancer in the seven countries study. Archives of Internal Medicine 1995;155(4):381.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 20 32. Zamora-Ros R, Jiménez C, Cleries R, et al. Dietary flavonoid and lignan intake and mortality in a Spanish cohort. Epidemiology 2013;24(5):726-33. 33. US Department of Agriculture. USDA database for the flavonoid content of selected foods; release 2.1. Maryland, 2007. 34. US Department of Agriculture. USDA database for the isoflavone content of selected foods; release 2.0. Maryland, 2008. 35. Nutrient Data Laboratory, Agricultural Research Service, US Department of Agriculture. USDA database for proanthocyanidin content of selected foods. In: USDA, ed. Beltsville, MD, 2004. 36. Neveu V, Perez-Jiménez J, Vos F, et al. Phenol-Explorer: an online comprehensive database on polyphenol contents in foods. Database 2010. 37. Prince RL, Devine A, Dhaliwal SS, Dick IM. Effects of calcium supplementation on clinical fracture and bone structure: results of a 5-year, double-blind, placebo-controlled trial in elderly women. Archives of Internal Medicine 2006;166(8):869-75. 38. Bruce DG, Devine A, Prince RL. Recreational Physical Activity Levels in Healthy Older Women: The Importance of Fear of Falling. Journal of the American Geriatrics Society 2002;50(1):84-9. 39. World Health Organization. Manual of the international statistical classification of diseases, injuries, and causes of death : based on the recommendations of the ninth revision conference, 1975, and adopted by the twenty-ninth World Health Assembly. 1975 revision. ed. Geneva: World Health Organization, 1977. 40. National Centre for Classification in Health (Australia). The International statistical classification of diseases and related health problems, 10th revision, Australian modification (ICD-10-AM). 1st ed. Sydney: National Centre for Classification in Health, 1998. 41. Lewis JR, Calver J, Zhu K, Flicker L, Prince RL. Calcium supplementation and the risks of atherosclerotic vascular disease in older women: results of a 5-year RCT and a 4.5-year follow-up. Journal of Bone and Mineral Research 2011;26(1):35-41. 42. Zhu K, Devine A, Lewis JR, Dhaliwal SS, Prince RL. 'Timed up and go' test and bone mineral density measurement for fracture prediction. Archives of Internal Medicine 2011;171(18):1655. 43. Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Australia New Zealand Journal of Public Health 2000;24(6):576-83. 44. Xinying PX, Noakes M, Keogh J. Can a food frequency questionnaire be used to capture dietary intake data in a 4 week clinical intervention trial? Asia Pacific Journal of Clinical Nutrition 2004;13(4):318-23. 45. Ambrosini GL, van Roosbroeck SAH, Mackerras D, Fritschi L, de Klerk NH, Musk AW. The reliability of ten-year dietary recall: implications for cancer research. Journal of Nutrition 2003;133(8):2663-8. 46. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pacific Journal of Clinical Nutrition 1994;3:19-31.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 21 47. Ivey KL, Lewis JR, Prince RL, Hodgson JM. Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women. The British journal of Nutrition 2013:1-8. 48. Murray CJ, Jamison DT, Lopez AD, Ezzati M, Mathers CD. Global Burden of Disease and Risk Factors: Washington, DC: World Bank and Oxford University Press, 2006. 49. Rice-Evans CA, Miller NJ, Paganga G. Structure-antioxidant activity relationships of flavonoids and phenolic acids. Free Radical Biology and Medicine 1996;20(7):933-56. 50. Loke WM, Proudfoot JM, Stewart S, et al. Metabolic transformation has a profound effect on anti-inflammatory activity of flavonoids such as quercetin: Lack of association between antioxidant and lipoxygenase inhibitory activity. Biochemical Pharmacology 2008;75(5):1045-53. 51. Cutler GJ, Nettleton JA, Ross JA, et al. Dietary flavonoid intake and risk of cancer in postmenopausal women: The Iowa Women's Health Study. International Journal of Cancer 2008;123(3):664-71. 52. Johannot L, Somerset SM. Age-related variations in flavonoid intake and sources in the Australian population. Public Health Nutrition 2006;9(8):1045-54. 53. Pérez-Jiménez J, Fezeu L, Touvier M, et al. Dietary intake of 337 polyphenols in French adults. The American Journal of Clinical Nutrition 2011;93(6):1220-8. 54. Zamora-Ros R, Knaze V, Luján-Barroso L, et al. Differences in dietary intakes, food sources and determinants of total flavonoids between Mediterranean and non-Mediterranean countries participating in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. British Journal of Nutrition 2013;109(08):1498-507. 55. Hertog MGL, Feskens EJM, Kromhout D, Hollman PCH, Katan MB. Dietary antioxidant flavonoids and risk of coronary heart disease: the Zutphen Elderly Study. Lancet 1993;342(8878):1007-11. 56. Yochum L, Kushi LH, Meyer K, Folsom AR. Dietary flavonoid intake and risk of cardiovascular disease in postmenopausal women. American Journal of Epidemiology 1999;149:943-9. 57. McCullough ML, Peterson JJ, Patel R, Jacques PF, Shah R, Dwyer JT. Flavonoid intake and cardiovascular disease mortality in a prospective cohort of US adults. American Journal of Clinical Nutrition 2012;95(2):454-64. 58. Cassidy A, Mukamal KJ, Liu L, Franz M, Eliassen AH, Rimm EB. High anthocyanin intake is associated with a reduced risk of myocardial infarction in young and middle-aged women. Circulation 2013;127(2):188-96. 59. Brown AL, Lane J, Coverly J, et al. Effects of dietary supplementation with the green tea polyphenol epigallocatechin-3-gallate on insulin resistance and associated metabolic risk factors: randomized controlled trial. British Journal of Nutrition 2009;101(06):886-94. 60. Loke W, Hodgson J, Croft K. The biochemistry behind the potential cardiovascular protection by dietary flavonoids. Edtion ed. In: Fraga CG, ed. Plant phenolics and human health: biochemistry, nutrition and pharmacology. New Jersey: John Wiley & Sons, 2009. 61. Huxley RR, Neil HA. The relation between dietary flavonol intake and coronary heart disease mortality: a meta-analysis of prospective cohort studies. European Journal of Clinical Nutrition 2003;57(8):904-8.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 22 62. Keli SO, Hertog MGL, Feskens EJM, Kromhout D. Dietary flavonoids, antioxidant vitamins, and incidence of stroke: the Zutphen Study. Archives of Internal Medicine 1996;156(6):637-42. 63. Geleijnse JM, Launer LJ, van der Kuip DA, Hofman A, Witteman JC. Inverse association of tea and flavonoid intakes with incident myocardial infarction: the Rotterdam Study. The American Journal of Clinical Nutrition 2002;75(5):880-6. 64. Hertog MG, Kromhout D, Aravanis C, Blackburn H, Buzina R, Fidanza F. Flavonoid intake and long-term risk of coronary heart disease and cancer in the seven countries study. Archives of Internal Medicine 1995;155:381-6. 65. Knekt P, Järvinen R, Seppänen R, et al. Dietary flavonoids and the risk of lung cancer and other malignant neoplasms. American Journal of Epidemiology 1997;146(3):223-30. 66. Mursu J, Nurmi T, Tuomainen TP, Salonen JT, Pukkala E, Voutilainen S. Intake of flavonoids and risk of cancer in Finnish men: the Kuopio Ischaemic Heart Disease Risk Factor Study. International Journal of Cancer 2008;123(3):660-3. 67. Bosetti C, Spertini L, Parpinel M, et al. Flavonoids and breast cancer risk in Italy. Cancer Epidemiology Biomarkers and Prevention 2005;14(4):805-8. 68. Liu L-Z, Fang J, Zhou Q, Hu X, Shi X, Jiang B-H. Apigenin inhibits expression of vascular endothelial growth factor and angiogenesis in human lung cancer cells: implication of chemoprevention of lung cancer. Molecular Pharmacology 2005;68(3):635-43. 69. Garcia-Closas R, Agudo A, Gonzalez CA, Riboli E. Intake of specific carotenoids and flavonoids and the risk of lung cancer in women in Barcelona, Spain. Nutrition and Cancer 1998;32(3):154-8. 70. Zava DT, Duwe G. Estrogenic and antiproliferative properties of genistein and other flavonoids in human breast cancer cells in vitro. Nutrition and Cancer 1997;27(1):31-40. 71. So FV, Guthrie N, Chambers AF, Moussa M, Carroll KK. Inhibition of human breast cancer cell proliferation and delay of mammary tumorigenesis by flavonoids and citrus juices. Nutrition and Cancer 1996;26(2). 72. Peterson J, Lagiou P, Samoli E, et al. Flavonoid intake and breast cancer risk: a case–control study in Greece. British Journal of Cancer 2003;89(7):1255-9. 73. Bushman JL. Green tea and cancer in humans: a review of the literature. Nutrition and Cancer 1998;31(3):151-9. 74. Howard L, Pandjaitan N, Morelock T, Gil M. Antioxidant capacity and phenolic content of spinach as affected by genetics and growing season. Journal of Agricultural and Food Chemistry 2002;50(21):5891-6. 75. van der Sluis AA, Dekker M, de Jager A, Jongen WM. Activity and concentration of polyphenolic antioxidants in apple: effect of cultivar, harvest year, and storage conditions. Journal of Agricultural and Food Chemistry 2001;49(8):3606-13. 76. Goldman I, Kader A, Heintz C. Influence of production, handling, and storage on phytonutrient content of foods. Nutrition Reviews 1999;57(9):46-52.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 23 4.8 TABLES AND FIGURES

Chapter 4, Table 1: Baseline characteristics of the cohort stratified by total-flavonoid intake group 1

Low intake Moderate intake High intake

Number (%) 354 (33) 355 (33) 354 (33)

2 FlavonoidUSDA Age (years) 80 ± 3 80 ± 3 80 ± 3 Prevalent cardiovascular disease 125 (35) 97 (27) 107 (30) Prevalent cancer 103 (29) 100 (28) 115 (32) Overweight or obese 4 215 (64) 222 (65) 228 (66) Low fruit and vegetable intake 5 297 (84) a 262 (74) b 197 (56) c Physical inactivity6 163 (46) 142 (40) 151 (43) Current smoker 13 (4) 8 (2) 11 (3) Current alcohol consumer 255 (72)a 287 (81)b 264 (75)b

3 FlavonoidPE Age (years) 80 ± 3 80 ± 3 80 ± 3 Prevalent cardiovascular disease 123 (35) 98 (28) 108 (30) Prevalent cancer 101 (28) 105 (30) 112 (32) Overweight or obese 4 222 (65) 214 (64) 229 (66) Low fruit and vegetable intake 5 295 (83) a 261 (74) b 200 (56) c Physical inactivity6 160 (45.2) 142 (40) 154 (44) Current smoker 12 (3) 10 (3) 10 (3) Current alcohol consumer 260 (73) 285 (80) 261 (73)

1Results are mean ± SD or n (%) where appropriate. n = 1 063 a,b,c Row values with different subscripts differ significantly by Bonferroni (P<0.05). 2 FlavonoidUSDA: low (<547 mg/d), moderate (547-<813 mg/d), high (≥813mg/d). 3 FlavonoidPE: low (<525 mg/d), moderate (525-<788 mg/d), high (≥788 mg/d). 4 Body mass index ≥ 25 kg/m2. n = 1 023. 5 Low fruit and vegetable intake: <600g total fruit and vegetable intake per day. 6 Physical inactivity: < 2.5 hours moderate-intensity physical activity per week.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 24

Chapter 4, Table 2: Association of total-flavonoid intake group and risk of all-cause mortality

Low intake Moderate intake High intake

Number (%) 354 (33) 355 (33) 354 (33)

1 FlavonoidUSDA Unadjusted 1.00 (referent) 0.78 (0.53,1.14) 0.37 (0.23,0.59)* Multivariate 1.00 (referent) 0.82 (0.54,1.25) 0.38 (0.22,0.64)*

2 FlavonoidPE Unadjusted 1.00 (referent) 0.57 (0.39,0.85)* 0.36 (0.23,0.57)* Multivariate 1.00 (referent) 0.52 (0.34,0.81)* 0.36 (0.22,0.60)*

Results are HR (95% CI) and n(%) where appropriate. * Represents significantly different from referent group by Cox regression (P<0.05). Multivariate adjusted model includes: age, prevalent cardiovascular disease and cancer, overweight or obesity, low fruit and vegetable intake, physical inactivity, current cigarette smoking and alcohol consumption. 1 FlavonoidUSDA: low (<547 mg/d), moderate (547-<813 mg/d), high (≥813mg/d). 2 FlavonoidPE: low (<525 mg/d), moderate (525-<788 mg/d), high (≥788 mg/d).

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 25

Chapter 4, Table 3: Association of total-flavonoid group and risk of mortality from cardiovascular disease and cancer

Low intake Moderate intake High intake

Number (%) 354 (33) 355 (33) 354 (33)

Cardiovascular disease mortality

1 FlavonoidUSDA CVD mortality [n(%)]3 36 (10) 29 (8) 13 (4) Unadjusted 1.00 (referent) 0.79 (0.48,1.28) 0.34 (0.18,0.64)* Multivariate 1.00 (referent) 0.81 (0.47,1.40) 0.34 (0.17,0.69)*

2 FlavonoidPE CVD mortality [n(%)]3 40 (11) 25 (7) 13 (4) Unadjusted 1.00 (referent) 0.60 (0.36,0.99)* 0.30 (0.16,0.57)* Multivariate 1.00 (referent) 0.47 (0.27,0.86)* 0.32 (0.16,0.61)*

Cancer mortality

1 FlavonoidUSDA Cancer mortality [n(%)] 26 (7) 16 (4) 7 (2) Unadjusted 1.00 (referent) 0.60 (0.32,1.23) 0.26 (0.11,0.59)* Multivariate 1.00 (referent) 0.63 (0.32,1.22) 0.25 (0.10,0.62)*

2 FlavonoidPE Cancer mortality [n(%)] 27 (8) 14 (4) 8 (2) Unadjusted 1.00 (referent) 0.50 (0.26,0.95)* 0.28 (0.13,0.62)* Multivariate 1.00 (referent) 0.44 (0.21,0.89)* 0.26 (0.11,0.62)*

Results are HR (95% CI) and n(%) where appropriate. * Represents significantly different from referent group by Cox regression (P<0.05). Multivariate adjusted model includes: age, prevalent cardiovascular disease and cancer, overweight or obesity, low fruit and vegetable intake, physical inactivity, current cigarette smoking and alcohol consumption. 1 FlavonoidUSDA: low (<547 mg/d), moderate (547-<813 mg/d), high (≥813mg/d). 2 FlavonoidPE: low (<525 mg/d), moderate (525-<788 mg/d), high (≥788 mg/d). 3 CVD: cardiovascular disease.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 26

Chapter 4, Figure 1: Proportional reduction in 5-year all-cause mortality incidence between low, moderate, and high total-flavonoid consumption as estimated from the United States Department of Agriculture database.

(a): Proportional reduction in 5-year all-cause mortality incidence between low, moderate, and high total-flavonoidUSDA consumption.

Low flavonoidUSDA intake: (<547 mg/d, n=354; moderate flavonoidUSDA intake: 547-<813mg/d, n=355; high flavonoidUSDA intake: ≥813mg/d, n=354.

(b): Cox proportional hazard model survival curve from baseline to 5-years follow-up stratified by total-flavonoidUSDA consumption.

Low flavonoidUSDA intake: <547mg/d, n=354.

Moderate flavonoidUSDA intake: 547-<813mg/d, n=355.

High flavonoidUSDA intake: ≥813mg/d, n=354.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 4: Page 27

Chapter 4, Figure 2: Proportional reduction in 5-year all-cause mortality incidence between low, moderate, and high total-flavonoid consumption estimated from the Phenol-Explorer database.

(a): Proportional reduction in 5-year all-cause mortality incidence between low, moderate, and high total-flavonoidPE consumption.

Low FlavonoidPE intake: <525 mg/d, n=354; moderate FlavonoidPE intake: 525-<788 mg/d, n=355; high FlavonoidPE intake: ≥788 mg/d, n=354.

(b): Cox proportional hazard model survival curve from baseline to 5-years follow-up stratified by total-flavonoidPE consumption.

Low FlavonoidPE intake: <525 mg/d, n=354.

Moderate FlavonoidPE intake: 525-<788 mg/d.

High FlavonoidPE intake: ≥788 mg/d, n=354.

Chapter 4: Flavonoid intake and all-cause mortality Chapter 5: Page 1

Health benefits of non-nutritive food components

Chapter 5: Comparison of flavonoid intake assessment methods

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 2 5.1 FOREWORD

The ability to assess flavonoid-disease relationships in epidemiological studies has been driven by the development of flavonoid food content databases. The United States

Department of Agriculture (USDA) created the first comprehensive flavonoid food content database (1). However, this was limited in terms of the number of flavonoid compounds assessed and the number of foods represented. As technology and available data advanced, so too did the subsequent releases of the USDA food content database

(2).

In 2009, the Phenol-Explorer food content database Version 1.0 was created with the support of the Institut National de la Recherche Agronomique, the NUTRIALIS program of the French Ministry of Research, Unilever, Danone and Nestlé. The current version of the Phenol-Explorer database now contains 37,636 food content values (3).

In 2013, the most upto date and revised USDA database for the flavonoid content of selected foods was released (4). As such, there are now two comprehensive flavonoid content databases available for implementation into epidemiological investigations.

However, it is unclear which database should be applied when estimating flavonoid intake in a population. Therefore, we aimed to compare and contrast methods used by each database to derive food content values, and to explore the level of agreement between the flavonoid intake estimates derived from USDA and Phenol-Explorer data.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 3 5.2 ABSTRACT

In preparation Kerry L Ivey, Croft KD, Prince RL, Hodgson JM

Comparison of flavonoid intake assessment methods

Background: Flavonoids are a diverse group of polyphenolic compounds, found in high concentrations in many plant foods and beverages. High flavonoid intake has been associated with reduced risk of chronic disease. To date, population based studies have used the United States Department of Agriculture (USDA) food content database to determine habitual flavonoid intake. More recently, a new flavonoid food content database, Phenol-Explorer (PE), has been developed. However, the level of agreement between the two databases has yet to be explored.

Aim: To compare the methods used by each database to derive food content values, and to explore the level of agreement between the flavonoid intake estimates derived from USDA and PE data.

Design: The study population included 1 063 randomly selected women aged over 75 y. Two separate estimates for total-flavonoid and each flavonoid class were determined using food composition data from the USDA and the PE databases.

Results: The two databases used similar criteria for inclusion of source data. Despite differences in net estimates, there was a strong level of agreement between total-flavonoid, flavanol, flavanone and anthocyanidin intake estimates derived from each database. Flavonol and flavone intake estimates showed the weakest level of agreement between databases. Several factors could contribute to differences in intake estimates. For almost all flavonoid compounds, USDA computed flavonoid content from aglycone molecular weights, whereas PE included the mass of functional groups in food content estimates. There were differences in the expression of flavonoid class content of individual food items, and thearubinigins were included in flavonol estimates in the USDA database, but not the PE.

Conclusion: In this population, application of USDA and PE source data yielded highly correlated intake estimates for total-flavonoids, flavanols, flavanones and anthocyanidins. For these classes, the USDA and PE databases can be used interchangeably in epidemiological investigations. There was poorer correlation between flavonol and flavone intake estimates due to differences in USDA and PE methodologies. When applying food content values to estimate flavonol and flavone, it is important to consider structure of interest (glycone versus aglycone) as well as the foods consumed and assessed in the cohort.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 4

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 5 5.3 INTRODUCTION

Flavonoids represent a diverse group of polyphenolic compounds derived from a flavan

(2-phenylchroman ) nucleus (Figure 1). Derivations of this basic structure arise due to alterations in the 2(3) carbon-carbon bond, the formation of a ketone at carbon 4, and hydroxylation of carbons at various locations on the flavan backbone (5). It is these derivations that give the over 4000 flavonoid molecules to be grouped into one of five main flavonoid classes; flavonols, flavanols, flavones, flavanones, and anthocyainidins

(Table 1) (6-8). Each flavonoid class is comprised of numerous individual compounds with varying degrees of polymerisation, glycosylation, hydroxylation and esterification.

Results of observational epidemiological studies of flavonoids and flavonoid rich foods suggest beneficial effects on cardiovascular disease and cancer (9-13). It is extremely difficult to assess the validity of flavonoid intake assessment tools as there is currently no gold standard biomarker of total-flavonoid intake. As such, the use of food composition databases in conjunction with traditional dietary assessment methods provides the most practical means for analysis of habitual dietary flavonoid consumption in the community. In previous epidemiological studies, the most widely adopted flavonoid food content database has been that developed by the United States Department of

Agriculture (USDA) (2, 14). Traditionally, food content data from the USDA database has been combined with food intake data from 24-hour recall and food frequency questionnaires (FFQ) to estimate flavonoid intake (9, 15). It is these computed estimates of flavonoid intake which are used as the independent variable to investigate diet-disease relationships in population based settings.

More recently, the Phenol-Explorer (3) database has emerged, providing an additional high quality summary of the flavonoid content of commonly consumed foods. However, the manner in which the two databases deal with the complexity of flavonoid structure

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 6 and content, and the degree to which these two databases agree when applied to a validated dietary assessment method has not yet been investigated. As such, this study aims to compare the methods used by each database to derive food content values, and to explore the level of agreement between the flavonoid intake estimates derived from

USDA and PE data.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 7 5.4 METHODS: METHODOLOGICAL COMPARISON

5.4.1 Methodological comparison

For this comparison, data from Microsoft Access databases were obtained from the

‘USDA database for the flavonoid content of selected foods release 3.1’ (4), ‘USDA database for the proanthocyanin content of selected foods’ (14) and Phenol-Explorer release 3.0 (3). Methodological data was obtained from associated documentation, and where information was missing or required clarification, the authors were contacted directly.

5.4.2 Food composition comparison

In order to compare and contrast the impact each database may have on food composition estimates, we sought to create a summary statistic to represent the estimated flavonoid content of foods reported in a commonly used and validated food frequency questionnaire (16-18). The method of computing flavonoid content of foods has been previously described in Ivey et al. (19), and is similar to the method adopted by many other investigators when assessing flavonoid intake (9).

Extraction procedures for the 2 databases (USDA and PE) were identical and were carried out by the same investigator. Both of these databases were used to derive two separate estimates of flavonoid intake: flavonoid intake based on food composition data from the USDA (FlavonoidUSDA) and flavonoid intake based on food composition data from the PE (FlavonoidPE) database.

The sum of assessed flavonoids for each flavonoid class was calculated by summing the individual compounds of each flavonoid class in the form expressed in each individual database. The terminology and classification systems used by each database varied; therefore, a standardized classification system was adopted throughout this study. The

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 8 term Anthocyanins in this study refer to the USDA anthocyanidin class. The term flavanol in this study encapsulates both the Flavan-3-ol and Proanthocyanin classes in the USDA database. As such, flavanolUSDA represented the sum of Proanthocyanin polymer values in conjunction with total flavan-3-ol content.

The chalcone, dihydrochalcone and dihydroflavonol content of foods are described in the PE but not the USDA database. As such, these compounds were omitted from this comparison study. Isoflavones are commonly included in epidemiological analyses of total-flavonoid intake (20), and are included in the flavonoid section of the PE database.

However, rather than sharing the nuclear structure of flavonoids (2-phenylchroman ), isoflavones have a 3-Phenylchroman base structure. As such, isoflavones did not meet the criteria for inclusion in this review.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 9 5.5 METHODS: FLAVONOID INTAKE COMPARISON

5.5.1 Participants

1,136 postmenopausal women above the age of 75 years were recruited into the

Calcium Intake Fracture Outcome Age Related Extension Study in 2003. These participants had previously completed a 5-year prospective, randomized, controlled trial of oral calcium supplements to prevent osteoporotic fractures (21). This study was approved by the Human Ethics Committee of the University of Western Australia, and written informed consent was obtained from all participants.

A total of 1,063 participants had complete food frequency and beverage intake data at baseline (2003). Participants had a mean age of 80 (±3) years, and a mean body mass index of 27 (±5) kg/m2.

5.5.2 Dietary assessment

Baseline dietary intake was assessed using a validated semi-quantitative (FFQ) developed by the Anti-Cancer Council of Victoria (16). Dietary intakes in g/day were estimated based on frequency of consumption and an overall estimate of usual portion size (22). A beverage questionnaire was used to assess average tea and coffee consumption over the preceding 12 months.

5.5.3 Flavonoid intake

Intakes of flavonoid classes in mg/d were calculated by multiplying the estimated intake

(g edible portion/d) from the FFQ and beverage questionnaires, with the flavonoid class content (mg/g edible portion) of each food item on the questionnaires. Where multiple varieties of a food listed in the FFQ were reported in the databases, the average flavonoid content of all similar varieties was computed, consistent with the descriptors

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 10 used in the FFQ output. Foods in the FFQ that were not in the flavonoid databases were assumed to contain no flavonoids.

5.5.4 Statistics

Paired sample t-test was performed in order to compare the mean total-flavonoid intake estimates derived from the two databases. The extent to which estimates from

USDA and PE databases are linearly related was explored with product-moment correlation coefficients. The relationship between measurement error and true value was gauged from the Bland and Altman plot. Spearman rank correlation coefficient was used to investigate the linear relationship between total-flavonoid intake groups based on tertiles of estimated intake from the two databases.

As a post hoc analysis, all investigations were repeated on the five flavonoid classes, in order to explore the potential contribution of individual classes to observed results. All data was analysed on SPSS (version 20; IBM, New York, NY) according to a pre-specified protocol.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 11 5.6 RESULTS: METHODOLOGICAL COMPARISON

5.6.1 Data sources

The PE database accessed only published data sources, whereas USDA also accessed unpublished data. In terms of the analytical methods used by data sources, both databases were based on chromatography values. However, PE also included spectrophometric data.

Phenol-Explorer excluded studies from the database if it was deemed that an inappropriate method of polyphenol extraction was used or lack on information was provided on the method. They also excluded studies if there was a lack of information on phenolic standards used for quantification, mean content values without a description of the number of samples analysed, or content values reported in a graph. It is unclear if these factors were considered in USDA database development. Both databases only included values for specific flavonoid compounds, and omitted summary values for total flavonoid or flavonoid class content of particular food items.

The PE database omitted data of non-edible parts of plants and non-commercial or experimental products. With the exception of low moisture content cereal products where standard moisture content data was used to convert results to a fresh weight, data sources that did not describe the moisture content of dried samples were excluded. A lack of descriptive data on the nature of the samples analysed was also a criteria for exclusion. USDA does list food sample criteria; however concurrence is inferred by methods of data display and aggregation.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 12 5.6.2 Chemistry of included compounds (Supplementary Tables 1-5)

With the exception of (-)-epigallocatechin and (+)-gallocatechin, the weight of flavonoid compounds in USDA food items was computed from the aglycone molecular weight. Conversely, the molecular weight of glycoside, glucoside and ester moieties of flavonoid molecules was included in the PE mass data. Cis and trans isomers are not separately identified in the PE, and are instead represented as a total value. USDA does not list isomer criteria.

The authors identify that the ‘USDA database for the proanthocyanin content of selected foods’ (14) is a provisional database, and has recognised that accuracy of food composition data in this area is limited by limitations in technical assessment methods.

5.6.3 Food content data

The flavonoid compound quantity of each food item is displayed as the mean weighted for the number of samples used to generate each original data. In calculating the mean content values, the PE also considered analytical methods used by the data sources by grouping the individual data points based on analytical technique prior to data aggregation. Food items are described using the USDA National Nutrient Database for

Standard Reference in the USDA database, and the PE database used LanguaL descriptors. In both databases, the flavonoid content of solid foods is expressed as mg/100g of fresh weight of edible portion of food. Beverage data is expressed as mg/100mL in the PE, and as mg/100g adjusted by specific gravity in the USDA. The

USDA standardised tea infusion data to 1% infusion strength, however did not adjust for brewing time. PE does not describe infusion strength considerations.

When data sources reported food items containing trace amounts of flavonoid compounds, the USDA estimated flavonoid content by multiplying the limit of

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 13 quantitation (if available) by 0.71. PE does not describe method for dealing with trace data. Zero values in both databases represent true zeroes or levels below the limit of detection, whereas missing values indicate an absence of available data.

5.6.4 Imputed food composition (Table 2)

When applied to standard food items appearing in the ACCV-FFQ to derive an estimate of the flavonoid-class content of food, it is apparent that the PE database provides a greater number of data points which can be used to create food content summaries.

However, data points that are absent in each database are generally classified by the opposing database as having a low concentration of the particular flavonoid-class in the food item. Chocolate is an exception where the USDA database does not report the flavonol, however the aggregated PE data identifies chocolate as a food item with high flavonol concentration.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 14 5.7 RESULTS: FLAVONOID INTAKE COMPARISON

5.7.1 Flavonoid intake estimates (Table 3)

Mean total-flavonoid intake of the cohort estimated from USDA was 25% greater than that estimated from the PE. Similarly, the USDA estimates were larger than the PE estimates for the flavanol and anthocyanidin classes. The greatest difference between the two databases was observed with the flavonol class, with PE estimates being nearly

6 times greater than the USDA flavonol intake estimates.

On an individual basis, there were no participants with identical intake estimates for total-flavonoid, or any of the flavonoid classes. From a population perspective, results of the paired sample t-test indicate that the mean consumption estimates for total-flavonoid, flavonol, flavanol, flavone and anthocyanidin, derived from the USDA database were different to those derived from the PE database.

5.7.2 Extent to which the USDA and PE intake estimates are linearly related

(Figure 2 and 3)

The estimates of total-flavonoid intake from USDA and PE databases were strongly, linearly and positively correlated. There was also a strong linear relationship between the 2 databases for flavonol, flavanol, flavanone, and anthocyanidin intake estimates.

The strength of linear relationship was weakest for the flavone intake estimates; Pearson correlation coefficient = 0.274, P < 0.001.

5.7.3 Relationship between level of disagreement and mean estimated intake

(Figure 3)

The Bland and Altman plot demonstrates that the variability between the total-flavonoid estimates two databases is related to the size of the mean flavonoid intake estimates.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 15 The size of the difference between the two databases was greatest at higher levels of mean total-flavonoid intake. Similarly, the difference between estimates from the two databases was also proportional to the mean estimate for all flavonoid classes.

5.7.4 Extent to which the USDA and PE intake estimates classify participants as low, moderate or high flavonoid consumers (Table 4)

In order to explore the ability of the two databases to rank intake of participants appropriately, we trichotomised the cohort into three levels of intake (low, moderate and high) based on tertiles of USDA and PE derived intake estimates. Despite significant differences in participant classification, the classification of participants based on intake estimates derived from both databases were significantly correlated.

Ranking over 80% of participants identically, the reliability of the intake rankings from the two databases was high for total-flavonoid, flavanol and flavone intake estimates.

With less than 50% placed into identical intake groups, the USDA and PE databases showed poorest ranking agreement for the flavone class.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 16 5.8 DISCUSSION

This study aimed to compare the methods used by the USDA and PE databases to derive food content values. By applying these values to a validated FFQ, we then aimed to explore the level of agreement between the flavonoid intake estimates derived from

USDA and PE databases. When applied to ACCV-FFQ food intake data, the mean total-flavonoidUSDA intake estimate being 25% greater than that derived from PE data.

Despite a high degree of heteroscedasticity, the intake values derived from the two databases were related in a linear fashion, in which 96% of the variation in intake estimates being explained by the relationship between USDA and PE data. These differences in total-flavonoid intake estimates can be explained by differences in flavonoid class intake estimates, that arise as a result of differences in methodologies adopted by each database to derive food content estimates, flavonoid compounds assessed by each database, as well as the food items expressed in each database.

The mean flavonolPE intake of the cohort was nearly 6 times greater than the flavonolUSDA estimate. The majority of this difference can be explained by a methodological difference where the USDA database expresses data as the mass of aglycone and aglycone converted molecules, whereas the PE expresses data which includes the mass of both the aglycone and glycoside moieties. This methodological difference has a substantial impact on food content estimates, as the molecular weight of flavonoids with functional groups attached can be as much as nearly 3 times greater than their aglycone counterparts. Although the PE database provides data for five additional flavonol sub-classes which were not expressed in USDA, they typically occur in very low quantities in foods, and as such likely have a very small impact on net-flavonol intake estimates. It is these methodological differences, along with the absence of flavonol content data for chocolate in the USDA database that are likely to

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 17 have contributed to the relatively poorer database agreement when compared to other flavonoid classes.

The major contributor to total-flavonoid intake was the flavanol class; comprising 80% of total-flavonoidUSDA and 59% total-flavonoidPE. With the estimated daily flavanolUSDA intake being 268mg greater than the flavanolPE estimate, there was substantial absolute difference in estimated flavanol intake. This is likely due to the absence of thearubinigin data in the PE data. Thearubigins make up 24% (158 mg/d) of total flavanol intake, and when thearubigins are excluded from analysis, mean flavanolUSDA is 508 mg/d, compared to the 398 mg/d estimated with PE data. The remainder of this difference is likely explained by the absence of functional group masses in the USDA database.

Despite differences in absolute estimates, there was good correlation in estimates from both databases, which likely arises due to the overrepresentation of tea in flavanol intake estimates. Flavanol from tea makes up 68% (452 mg/d) of flavanolUSDA and 58%

(230 mg/d) of flavanolPE intake. The large contribution of tea-flavanols to total flavanol intake means results in lower degree of variance attributable to differences between the two databases in flavanol content estimates of flavanol containing foods. As flavanols contribute substantially to total flavonoid intake, the high agreement between flavanol intake estimates is reflected in the high agreement in total-flavonoid intake estimates.

Flavones are the best described class in terms of food content estimates, likely because this class was included in the earliest flavonoid food content databases (1). Despite this, the flavonePE intake estimate is 5 times greater than the flavoneUSDA estimate, and there was poor agreement between the two estimates, with a substantial degree of heteroscedasticity. Glycosides are the predominant form of flavones in foods(23), which likely explains the different in absolute intake estimates, as well as the level of agreement between the two databases. By converting all flavone compounds to aglycone versions prior to computing the mass of flavone in a food, the USDA database

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 18 places equal weighting to all flavones irrespective of their chemical structure.

Conversely, by including different flavone compounds with different molecular weights, when computing flavone intake with PE data, foods high in high molecular weight flavone compounds will receive more weighting compared to those rich in low molecular weight flavones. This methodological difference results in different rankings of low and high flavone foods, and manifests in different ranking of participants as low, moderate and high flavone consumers across the two databases. It is unlikely that the absence of functional group masses in the USDA data would explain the all of the difference in intake estimates. The higher PE food content estimates for food such as tea and fruit juice likely contributed to the absolute differences and poor agreement between the two databases.

Flavanone intake between the two databases is strongly correlated and similar in terms of net estimates. The main reason for this is because the flavanone content of food is not well defined by each database. Therefore, variation attributable to differences in food content estimates is minimal. Furthermore, the effect of including functional groups in

PE estimates does not substantially increase PE values as the flavone glycones are typically of relatively low molecular weight(23).

Despite including functional group masses in anthocyanidinPE estimates, the daily

USDA anthocyanidin intake estimate is nearly 4 times greater than PE estimates. The difference in net intake estimates is likely explained by the higher anthocyanidin food content values in the USDA compared to the PE database. The absolute difference of 70 mg/d contributes to the higher total-flavonoidUSDA intake estimate. The two databases show moderately strong agreement in their ability to rank participants as low, moderate and high anthocyanidin consumers.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 19 The majority of methodologies adopted by both databases were comparable. Both databases did not specifically identify isomeric state of included compounds, primarily due to lack of source data for separate enantiomers(3). This omission is unlikely to affect the net estimates will not be affected as cis and trans isomeric versions are likely included in imputations of larger compound groups. However, absorption and bioactivity of flavonoids may be affected by isomeric state(24), as such the variability in bioactivity may not be reflected in current databases.

A notable difference between methodologies adopted by the two databases is in the data aggregation methods. In calculating the mean content values, the PE also considered analytical methods used by the data sources by grouping the individual data points based on analytical technique prior to data aggregation, whereas the USDA did not. The

USDA imputed content data containing trace amounts of flavonoid compounds. It is unclear how the PE deals with trace data. The databases used different methods for expressing flavonoid density of beverages, and the USDA adopted a standardised infusion strength for tea, whereas PE did not describe a standardisation method. Although potentially affecting content estimates of individual flavonoid compounds in individual food items, this lack of methodological congruency is unlikely to result in systematic differences between food content estimates, and is unlikely to result in net alterations in flavonoid intake estimates.

The comparison of flavonoid intake estimates derived from each database is affected by the foods included in the FFQ as well as the dietary intake pattern of the population investigated. We have previously used the ACCV-FFQ to identify cross-sectional and prospective flavonoid-disease relationships (18, 19), which has been validated in populations of similar age and geographical location to our cohort (25). However, dietary patterns vary across different geographical locations and age groups. As such, foods identified in FFQs typically reflect the dietary pattern of the population to which

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 20 it is administered. As such, the level of agreement between the USDA and PE databases observed in this study may be different to the level of agreement in cohorts of different nationality and age.

Despite differences in net estimates, there was a strong level of agreement between total-flavonoid, flavanol, flavanone and anthocyanidin intake estimates derived from each database. This agreement in ranking based on intake estimates makes the data appropriate for epidemiological studies which rely on ranking, rather than absolute values, for exploration of associations with outcomes. The weaker agreement between

USDA and PE estimates for both flavonol and flavone intake estimates arises due to differences in methodologies adopted by each database, and not because of a systematic error in either database. As such, when determining the appropriate database to calculate flavonol and flavone intake, it is important to consider whether aglycone or glycone structures are the compound of interest and which foods are important contributors to dietary intake in the population of interest.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 21

5.9 CHAPTER 5 REFERENCES

1. US Department of Agriculture. USDA database for the flavonoid content of selected foods release 1. Maryland, 2003. 2. US Department of Agriculture. USDA database for the flavonoid content of selected foods; release 2.1. Maryland, 2007. 3. Neveu V, Perez-Jiménez J, Vos F, et al. Phenol-Explorer: an online comprehensive database on polyphenol contents in foods. Database 2010. 4. Service. USDoAAR. Internet: Nutrient Data Laboratory home page: http://www.ars.usda.gov/nutrientdata/flav. 5. Moss G, Smith P, Tavernier D. Glossary of class names of organic compounds and reactive intermediates based on structure. International Union of Pure and Applied Chemistry 1995;67(8):1307-75. 6. Aherne SA, O’Brien NM. Dietary flavonols: chemistry, food content, and metabolism. Nutrition 2002;18(1):75-81. 7. Cook N, Samman S. Flavonoids: chemistry, metabolism, cardioprotective effects, and dietary sources. Journal of Nutritional Biochemistry 1996;7(2):66-76. 8. Das DK. Naturally occurring flavonoids: Structure, chemistry, and high-performance liquid chromatography methods for separation and characterization. Edtion ed. In: Lester P, ed. Methods in Enzymology: Academic Press, 1994:410-20. 9. Mink PJ, Scrafford CG, Barraj LM, et al. Flavonoid intake and cardiovascular disease mortality: a prospective study in postmenopausal women. American Journal of Clinical Nutrition 2007;85(3):895-909. 10. Sun C-L, Yuan J-M, Koh W-P, Mimi CY. Green tea, black tea and breast cancer risk: a meta-analysis of epidemiological studies. Carcinogenesis 2006;27(7):1310-5. 11. Tang N, Wu Y, Zhou B, Wang B, Yu R. Green tea, black tea consumption and risk of lung cancer: a meta-analysis. Lung Cancer 2009;65(3):274-83. 12. Arab L, Liu W, Elashoff D. Green and black tea consumption and risk of stroke. Stroke 2009;40(5):1786-92. 13. Wang Z-M, Zhou B, Wang Y-S, et al. Black and green tea consumption and the risk of coronary artery disease: a meta-analysis. American Journal of Clinical Nutrition 2011;93(3):506-15. 14. Nutrient Data Laboratory, Agricultural Research Service, US Department of Agriculture. USDA database for proanthocyanidin content of selected foods. In: USDA, ed. Beltsville, MD, 2004. 15. Chun OK, Chung SJ, Song WO. Estimated Dietary Flavonoid Intake and Major Food Sources of U.S. Adults. The Journal of Nutrition 2007;137(5):1244-52. 16. Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Australia New Zealand Journal of Public Health 2000;24(6):576-83. 17. Belobrajdic DP, Frystyk J, Jeyaratnaganthan N, et al. Moderate energy restriction-induced weight loss affects circulating IGF levels independent of dietary composition. European Journal of Endocrinology 2010;162(6):1075-82.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 22 18. Ivey KL, Lewis JR, Lim WH, Lim EM, Hodgson JM, Prince RL. Associations of Proanthocyanidin Intake with Renal Function and Clinical Outcomes in Elderly Women. PLoS ONE 2013;8(8):e71166. 19. Ivey KL, Lewis JR, Prince RL, Hodgson JM. Tea and non-tea flavonol intakes in relation to atherosclerotic vascular disease mortality in older women. The British journal of Nutrition 2013:1-8. 20. Chun OK, Chung S-J, Claycombe KJ, Song WO. Serum C-Reactive Protein Concentrations Are Inversely Associated with Dietary Flavonoid Intake in U.S. Adults. The Journal of Nutrition 2008;138(4):753-60. 21. Prince RL, Devine A, Dhaliwal SS, Dick IM. Effects of calcium supplementation on clinical fracture and bone structure: results of a 5-year, double-blind, placebo-controlled trial in elderly women. Archives of Internal Medicine 2006;166(8):869-75. 22. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pacific Journal of Clinical Nutrition 1994;3:19-31. 23. Manach C, Scalbert A, Morand C, Rémésy C, Jiménez L. Polyphenols: food sources and bioavailability. The American Journal of Clinical Nutrition 2004;79(5):727-47. 24. Brand W, Shao J, Hoek-van den Hil EF, et al. Stereoselective conjugation, transport and bioactivity of S-and R-hesperetin enantiomers in vitro. Journal of Agricultural and Food Chemistry 2010;58(10):6119-25. 25. Ambrosini GL, van Roosbroeck SAH, Mackerras D, Fritschi L, de Klerk NH, Musk AW. The reliability of ten-year dietary recall: implications for cancer research. Journal of Nutrition 2003;133(8):2663-8.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 23 5.10 TABLES AND FIGURES

Chapter 5, Table 1: Structure and chemical name of flavonoid classes included in this review.

Class Chemical name Characterising structure

Flavonol 3-hydroxy-2-phenylchromen-4-one

Flavanol 3-hydroxy-2-phenylchroman

Flavone 2-phenylchromen-4-one

Flavanone 2-phenylchroman-4-one

Anthocyanidin 4’,3,5,7-hydroxy-2-phenylchromenylium

Figures sourced and adapted from Phenol-Explorer electronic database.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 24

Chapter 5, Table 2: Relative flavonoid class concentrations in foods, imputed from United States Department of Agriculture (USDA) and Phenol-Explorer (PE) data.

Flavonol Flavanol Flavone Flavanone Anthocyanidin

USDA PE USDA PE USDA PE USDA PE USDA PE Fresh fruit Apple Apricot Avocado Banana Citrus Melon Peach/nectarine Pear Pineapple Strawberry Beverages Beer Chocolate drink Fruit juice Fortified wine Red wine White wine Tea Vegetables Beetroot Broccoli Cabbage/b.sprout Carrot Cauliflower Celery Cucumber Green bean Salad greens Onion/leek Other bean Pea Capsicum Tomato Spinach Zucchini Pumpkin Potato Miscellaneous Chocolate Nuts

Cut-off concentrations used for classifying foods as low, moderate or high flavonoid class sources from the USDA database are as follows: Flavonol: <5mg/100g, 5-20 mg/100g, >20mg/100g. Flavanol: <20mg/100g, 20-100 mg/100g, >100mg/100g. Flavone: <1mg/100g, >1-5 mg/100g, >5mg/100g. Flavanone: <5mg/100g, >5-20 mg/100g, >20mg/100g. Anthocyanidin: <5mg/100g, 5-20 mg/100g, >20mg/100g.

White represents foods and flavonoid classes which were not represented in the respective database.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 25

Chapter 5, Table 3: Daily consumption of total-flavonoid and flavonoid class, as estimated with the United States Department of Agriculture and Phenol-Explorer databases

Intake computed from USDA Intake computed from database PE database (mg/d) (mg/d)

Total-flavonoid a 834 ± 394 669 ± 326

Flavonoid classes Flavonol a 30 ± 17 175 ± 105 Flavanol a 666 ± 345 398 ± 206 Flavone a 4 ± 3 25 ± 14 Flavanone 40 ± 36 53 ± 53 Anthocyanidin a 88 ± 77 18 ± 19

Results are mean ± SD. a Results are different by paired sample t-test (P<0.05).

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 26

Chapter 5, Table 4: Tertiles of Phenol-Explorer (PE) flavonoid intake expressed against United States Department of Agriculture (USDA) flavonoid intake tertiles.

TOTAL-FLAVONOID b c Total-flavonoidPE Low Moderate High d Total-flavonoidUSDA Low intake 325 29 0 Classified identically: 948 (89%) Moderate intake 26 299 30 Classified differently: 115 (11%) High intake 3 27 324 Correlation coefficient: 0.912 a

FLAVONOL b e FlavonolPE Low Moderate High f FlavonolUSDA Low intake 289 65 0 Classified identically: 784 (74%) Moderate intake 50 223 82 Classified differently: 279 (26%) High intake 15 67 272 Correlation coefficient: 0.771 a

FLAVANOL b g FlavanolPE Low Moderate High h FlavanolUSDA Low intake 304 49 1 Classified identically: 891 (84%) Moderate intake 45 272 38 Classified differently: 172 (16%) High intake 5 34 315 Correlation coefficient: 0.866 a

FLAVONE b i FlavonePE Low Moderate High j FlavoneUSDA Low intake 174 115 65 Classified identically: 461 (43%) Moderate intake 91 131 133 Classified differently: 602 (57%) High intake 89 109 156 Correlation coefficient: 0.249 a

FLAVANONE b k FlavanonePE Low Moderate High l FlavanoneUSDA Low intake 340 14 0 Classified identically: 963 (91%) Moderate intake 14 305 36 Classified differently: 100 (9%) High intake 0 36 318 Correlation coefficient: 0.929 a

ANTHOCYANIDIN b m AnthocyanidinPE Low Moderate High n AnthocyanidinUSDA Low intake 260 77 17 Classified identically: 733 (69%) Moderate intake 81 205 69 Classified differently: 330 (31%) High intake 13 73 268 Correlation coefficient: 0.703 a

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 27

Results are n (%) or Spearman rank correlation coefficient. a Results are significantly correlated (P<0.001). b Results are significantly different by Pearson Chi-Square (P<0.001). c FlavonoidPE: low(<525mg/d),moderate(525-<788 mg/d),high (≥788mg/d). d FlavonoidUSDA: low(<646mg/d), moderate(646-<976mg/d), high(≥976mg/d). e FlavonolUSDA: low(<27mg/d), moderate(27-<41mg/d), high(≥41mg/d). f FlavonoidPE: low(<124mg/d), moderate(124-<221mg/d), high(≥221mg/d). g FlavanolPE: low(<302mg/d), moderate(302-<459mg/d), high(≥459 mg/d). h FlavanolUSDA: low(<502mg/d), moderate(502-<791mg/d), high(≥791mg/d). i FlavonePE: low(<18mg/d), moderate(18-<30mg/d), high(≥30 mg/d). j FlavoneUSDA: low(<3mg/d), moderate(3-<5mg/d), high(≥5mg/d). k FlavanonePE: low(<18mg/d), moderate(18-<48mg/d), high(≥48 mg/d). l FlavanoneUSDA: low(<18mg/d), moderate(18-<48mg/d), high(≥48mg/d). m AnthocyaninPE: low(<7mg/d), moderate(7-<19mg/d), high(≥19 mg/d). n AnthocyaninUSDA: low(<40mg/d), moderate(40-<98mg/d), high(≥98mg/d).

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 28 Chapter 5, Figure 1: chemical structure of the flavan (2-phenylchroman) nucleus of flavonoid molecules

Figure sourced and adapted from Phenol-Explorer electronic database.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 29 a) b) 3000 100

(mg/d) 2000 (mg/d)

50

1000 PE PE flavonol intake PE PE total-flavonoid intake

0 0 0 1000 2000 3000 0 50 100 USDA total-flavonoid intake (mg/d) USDA flavonol intake (mg/d) c) d) 3000 100

2000 (mg/d) (mg/d)

50

1000 PE PE intakeflavone PE PE flavanol intake

0 0 0 1000 2000 3000 0 50 100 USDA flavanol intake (mg/d) USDA flavone intake (mg/d) e) f) 400 300

300

(mg/d) 200 (mg/d)

200

100

PE intake PE flavanone 100 PE PE anthocyanidinintake

0 0 0 100 200 300 400 0 100 200 300 400 500 600 USDA flavanone intake (mg/d) USDA anthocyanidin intake (mg/d)

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 30

Chapter 5, Figure 2: Level of agreement between United States Department of Agriculture (USDA) and Phenol-Explorer (PE) total-flavonoid and flavonoid class intake estimates.

n = 1,063. : line of equality. a) Total-flavonoid intake estimates Pearson correlation coefficient= 0.963, P<0.001. Unstandardized B= 795±0.007, P<0.001. b) Flavonol intake estimates Pearson correlation coefficient= 0.828, P<0.001. Unstandardized B= 5.257±0.109, P<0.001. c) Flavanol intake estimates Pearson correlation coefficient= 0.822, P<0.001. Unstandardized B= 0.550 ± 0.007, P < 0.001. d) Flavone intake estimates Pearson correlation coefficient= 0.274, P<0.001. Unstandardized B= 1.524±0.164, P<0.001. e) Flavanone intake estimates Pearson correlation coefficient= 0.970, P<0.001. Unstandardized B= 1.422±0.011, P<0.001. f) Anthocyanidin intake estimates Pearson correlation coefficient= 0.817, P<0.001. Unstandardized B= 0.196±0.004, P<0.001.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 31 a) b) 1500 200

1000 0 0 200 400

500 -200 -flavonoid/d) total (mg (mg flavonol/d) (mg (mg 0 -400 Difference betweenUSDA and PE Difference betweenUSDA and PE 0 1000 2000 3000

-500 -600 Mean USDA and PE total-flavonoid intake Mean USDA and PE flavonol intake (mg/d) (mg/d) c) d) 1500 20 10

0 0 10 20 30 40 50 1000 -10

-20

-30 500 -40 (mg (mg flavone/d) (mg (mg flavanol/d) -50

-60 0 Difference betweenUSDA and PE Difference betweenUSDA and PE 0 500 1000 1500 2000 2500 -70

-80

-500 -90 Mean USDA and PE flavanol intake Mean USDA and PE flavone intake (mg/d) (mg/d) e) f) 100 500

400

0 300 0 50 100 150 200 250 300

200

(mg (mg flavanone/d) -100 100 (mg (mg anthocyanidin/d) Difference betweenUSDA and PE Difference betweenUSDA and PE 0 0 100 200 300 400

-200 -100 Mean USDA and PE flavanone intake Mean USDA and PE anthocyanidin intake (mg/d) (mg/d)

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 32 Chapter 5, Figure 3: Relationship between level of disagreement and mean estimated total-flavonoid and flavonoid class intake

n = 1,063. : line of equality. a) Bland and Altman plot incorporating total-flavonoid intake estimates b) Bland and Altman plot incorporating Flavonol intake estimates c) Bland and Altman plot incorporating Flavanol intake estimates d) Bland and Altman plot incorporating Flavone intake estimates e) Bland and Altman plot incorporating Flavanone intake estimates f) Bland and Altman plot incorporating Anthocyanidin intake estimates

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 33 5.11 SUPPLEMENTARY TABLES

Chapter 5, Supplementary Table 1: Flavonol compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases

Compounds Compounds Aglycone structure reported in reported in PE USDA Kaempferol; Kaempferol 3,7,4'-O-triglucoside; Kaempferol 3,7-O-diglucoside; Kaempferol 3-O-(2''-rhamnosyl-6''-acetyl-galactoside) 7-O-rhamnoside; Kaempferol 3-O-(2''-rhamnosyl-galactoside) 7-O-rhamnoside; Kaempferol 3-O-(6"-malonyl-glucoside) ; Kaempferol 7-O-glucoside; Kaempferide; Kaempferol 3-O-(6''-acetyl-galactoside) 7-O-rhamnoside; Kaempferol3-O-rutinoside; Kaempferol 3-O-acetyl-glucoside; Kaempferol 3-O-galactoside; Kaempferol Kaempferol 3-O-galactoside 7-O-rhamnoside; Kaempferol 3-O-glucoside; Kaempferol 3-O-glucosyl-rhamnosyl-galactoside; 6,8-Dihydroxykaempferol; Kaempferol 3-O-glucosyl-rhamnosyl-glucoside; Kaempferol 3-O-glucuronide; Kaempferol 3-O-rhamnoside; Kaempferol 3-O-rhamnosyl-rhamnosyl-glucoside; Kaempferol 3-O-sophoroside; Kaempferol 3-O-xylosyl-rutinoside; Kaempferol 3-O-sophoroside 7-O-glucoside; Kaempferol 3-O-xylosyl-glucoside.

Myricetin; Myricetin 3-O-arabinoside; Myricetin 3-O-galactoside; Myricetin Myricetin 3-O-glucoside; Myricetin 3-O-rhamnoside; Myricetin 3-O-rutinoside.

Isorhamnetin; Isorhamnetin 3-O-glucoside; Isorhamnetin 3-O-galactoside; Isorhamnetin 3-O-glucoside 7-O-rhamnoside; Isorhamnetin Isorhamnetin 3-O-glucuronide; Isorhamnetin 3-O-rutinoside; Isorhamnetin 4'-O-glucoside; Isorhamnetin 7-O-rhamnoside; Rhamnetin.

Quercetin; Quercetin 3,4'-O-diglucoside; Quercetin 3-O-(6"-malonyl-glucoside); Quercetin 3-O-(6"-malonyl-glucoside) 7-O-glucoside; Quercetin 7,4'-O-diglucoside; Quercetin 3-O-(6''-acetyl-galactoside) 7-O-rhamnoside; 3,7-Dimethylquercetin; Quercetin 3-O-acetyl-rhamnoside; Quercetin 3-O-xylosyl-rutinoside; Quercetin 3-O-galactoside; Quercetin 3-O-galactoside 7-O-rhamnoside; Quercetin 3-O-glucoside; Quercetin 3-O-glucosyl-rhamnosyl-galactoside; Quercetin Quercetin 3-O-glucosyl-rhamnosyl-glucoside; Quercetin 3-O-glucosyl-xyloside; Quercetin 3-O-glucuronide; Quercetin 3-O-rhamnoside; Quercetin 4'-O-glucoside; Quercetin 3-O-rhamnosyl-galactoside; Quercetin 3-O-xylosyl-glucuronide; Quercetin 3-O-rhamnosyl-rhamnosyl-glucoside; Quercetin 3-O-rutinoside; Quercetin 3-O-sophoroside; Quercetin 3-O-xyloside; Quercetin 3-O-arabinoside.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 34

Methylgalangin

Morin

Galangin

3-Methoxynobiletin; 3-Methoxysinensetin.

5,3',4'-Trihydroxy-3-methoxy-6:7-methylenedioxyflavone 4'-O-glucuronide; 5,4'-Dihydroxy-3,3'-dimethoxy-6:7-methylenedioxyflavone

4'-O-glucuronide; Patuletin 3-O-(2''-feruloylglucosyl)(1->6)-[apiosyl(1->2)]-glucoside; Patuletin 3-O-glucosyl-(1->6)-[apiosyl(1->2)]-glucoside; Jaceidin

4'-O-glucuronide; Spinacetin 3-O-(2"-feruloylglucosyl)(1->6)-[apiosyl(1->2)]-glucoside; Spinacetin 3-O-(2"-p-coumaroylglucosyl)(1->6)-[apiosyl(1->2)]-glucoside; Spinacetin 3-O-glucosyl-(1->6)-glucoside; Spinacetin 3-O-glucosyl-(1->6)-[apiosyl(1->2)]-glucoside. Figures sourced and adapted from Phenol-Explorer electronic database (3).

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 35

Chapter 5, Supplementary Table 2: Flavanol compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases

Compounds Compounds Aglycone structure reported in USDA reported in PE

(-)-epicatechin; (-)-epicatechin-(2a-7)(4a-8)-epicatechin (-)-epicatechin 3-O-galactoside; (-)-epicatechin 3-O-gallate.

(-)-epigallocatechin; (-)-epigallocatechin (-)-epigallocatechin; (-)-epigallocatechin 3-O-gallate. 3-gallate

(+)-catechin; (+)-catechin 3-O-gallate; (+)-catechin (+)-catechin 3-O-glucose.

(+)-gallocatechin (+)-gallocatechin; (+)-gallocatechin 3-O-gallate.

Theaflavin; Theaflavin; Theaflavin 3,3'-O-digallate; Theaflavin Theaflavin-3,3'-digallate; 3'-O-gallate; Theaflavin-3-gallate; Theaflavin 3-O-gallate. Theaflavin-3'-gallate.

Thearubigins

x n

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 36

02 mers; Procyanidin dimer B1; Procyanidin dimer B2; Procyanidin dimer B3; Procyanidin dimer B4; Procyanidin dimer B5; Procyanidin dimer B7; Dimers; Trimmers; Prodelphinidin dimer B3; 03 mers; Procyanidin trimer C1; 4-6mers; 7-10mers; Procyanidin trimer C2; Procyanidin trimer EEC; Procyanidin Polymers. trimer T2; Prodelphinidin trimer C-GC-C; x n Prodelphinidin trimer GC-C-C; Prodelphinidin trimer GC-GC-C; 04-06 mers; 07-10 mers; Polymers (>10 mers).

Cinnamtannin A2

x 4

Figures sourced and adapted from Phenol-Explorer electronic database (3).

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 37

Chapter 5, Supplementary Table 3: Flavone compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases

Compounds Compounds Aglycone structure reported in reported in PE USDA Apigenin; Apigenin 6,8-di-C-glucoside; Apigenin 6-C-glucoside; Apigenin 7-O-(6''-malonyl-apiosyl-glucoside); Apigenin 7-O-apiosyl-glucoside; Apigenin 7-O-diglucuronide; Apigenin Apigenin 7-O-glucoside; Apigenin 7-O-glucuronide; Apigenin arabinoside-glucoside; Apigenin galactoside-arabinoside; Rhoifolin; Rhoifolin 4'-O-glucoside; Isorhoifolin. Luteolin; 6-Hydroxyluteolin; 6-Hydroxyluteolin 7-O-rhamnoside; Luteolin 6-C-glucoside; Luteolin 7-O-(2-apiosyl-6-malonyl)-glucoside; Luteolin 7-O-(2-apiosyl-glucoside); Luteolin 7-O-diglucuronide; Luteolin 7-O-glucoside; Luteolin 7-O-glucuronide; Luteolin Luteolin 7-O-malonyl-glucoside; Luteolin 7-O-rutinoside; Pebrellin; Diosmin; Chrysoeriol 7-O-glucoside; Chrysoeriol 7-O-apiosyl-glucoside; Chrysoeriol 7-O-(6''-malonyl-glucoside); Chrysoeriol 7-O-(6''-malonyl-apiosyl-glucoside); Neodiosmin.

Cirsilineol; Nepetin; Sinensetin; Eupatorin; Jaceosidin.

Cirsimaritin; Scutellarein; Tetramethylscutellarein; Hispidulin.

Geraldone; 7,3',4'-Trihydroxyflavone.

Gardenin B; Tangeretin.

5,6-Dihydroxy-7,8,3',4'-tetramethoxyflavone.

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 38

7,4'-Dihydroxyflavone

Nobiletin

Chrysin

Baicalein

Figures sourced and adapted from Phenol-Explorer electronic database (3).

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 39

Chapter 5, Supplementary Table 4: Flavanone compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases

Compounds Compounds Aglycone structure reported in USDA reported in PE

Hesperetin Hesperetin; Hesperidin; Neohesperidin; Didymin; Poncirin.

Naringenin; 6-Geranylnaringenin; 6-Prenylnaringenin; 8-Prenylnaringenin; Naringenin 7-O-glucoside; Naringin; Naringin Naringenin 4'-O-glucoside; Naringin 6'-malonate; Narirutin; Narirutin 4'-O-glucoside; Sakuranetin; Isoxanthohumol.

Eriodictyol Eriodictyol; Eriodictyol 7-O-glucoside; Eriocitrin; Neoeriocitrin.

Pinocembrin

Figures sourced and adapted from Phenol-Explorer electronic database (3).

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 5: Page 40

Chapter 5, Supplementary Table 5: Anthocyanidin compounds reported in the United States Department of Agriculture (USDA) and Phenol-Explorer (PE) databases

Compounds Compounds Aglycone structure reported in reported in PE USDA Cyanidin; Cyanidin 3,5-O-diglucoside; Cyanidin 3-O-(6''-acetyl-galactoside); Cyanidin 3-O-(6''-acetyl-glucoside); Cyanidin 3-O-(6''-caffeoyl-glucoside); Cyanidin 3-O-(6''-dioxalyl-glucoside); Cyanidin 3-O-galactoside; Cyanidin 3-O-(6''-malonyl-3''-glucosyl-glucoside) ; Cyanidin 3-O-xylosyl-rutinoside; Cyanidin Cyanidin 3-O-(6''-malonyl-glucoside); Cyanidin 3-O-(6''-p-coumaroyl-glucoside); Cyanidin 3-O-(6''-succinyl-glucoside); Cyanidin 3-O-arabinoside; Cyanidin 3-O-glucosyl-rutinoside; Cyanidin 3-O-rutinoside; Cyanidin 3-O-glucoside; Cyanidin 3-O-sambubioside; Cyanidin 3-O-sambubiosyl 5-O-glucoside; Cyanidin 3-O-sophoroside; Cyanidin 3-O-xyloside. Delphinidin 3,5-O-diglucoside; Delphinidin 3-O-(6''-acetyl-galactoside);

Delphinidin 3-O-(6''-acetyl-glucoside); Delphinidin 3-O-(6''-malonyl-glucoside); Delphinidin 3-O-(6''-p-coumaroyl-glucoside); Delphinidin 3-O-sambubioside; Delphinidin Delphinidin 3-O-feruloyl-glucoside; Delphinidin 3-O-galactoside; Delphinidin 3-O-glucoside; Delphinidin 3-O-glucosyl-glucoside; Delphinidin 3-O-rutinoside; Delphinidin 3-O-xyloside; Delphinidin 3-O-arabinoside.

Malvidin 3,5-O-diglucoside; Malvidin 3-O-(6''-acetyl-galactoside); Malvidin 3-O-(6''-acetyl-galactoside); Malvidin 3-O-(6''-acetyl-glucoside); Malvidin Malvidin 3-O-(6''-caffeoyl-glucoside); Malvidin 3-O-(6''-p-coumaroyl-glucoside); Malvidin 3-O-arabinoside; Malvidin 3-O-galactoside; Malvidin 3-O-glucoside.

Peonidin; Peonidin 3-O-(6''-acetyl-galactoside); Peonidin 3-O-(6''-acetyl-glucoside); Peonidin 3-O-(6''-malonyl-glucoside); Peonidin Peonidin 3-O-(6''-p-coumaroyl-glucoside); Peonidin 3-O-arabinoside; Peonidin 3-O-galactoside; Peonidin 3-O-glucoside; Peonidin 3-O-rutinoside.

Petunidin 3,5-O-diglucoside; Petunidin 3-O-(6''-acetyl-galactoside); Petunidin 3-O-(6''-acetyl-glucoside); Petunidin 3-O-(6''-p-coumaroyl-glucoside); Petunidin 3-O-arabinoside; Petunidin Petunidin 3-O-galactoside; Petunidin 3-O-glucoside; Petunidin 3-O-rhamnoside; Petunidin 3-O-rutinoside; Pigment A; Pinotin A; Vitisin A.

Pelargonidin; Pelargonidin 3,5-O-diglucoside; Pelargonidin 3-O-sambubioside; Pelargonidin 3-O-(6''-acetyl-glucoside); Pelargonidin 3-O-(6''-malonyl-glucoside); Pelargonidin 3-O-(6''-succinyl-glucoside); Pelargonidin Pelargonidin 3-O-arabinoside; Pelargonidin 3-O-galactoside; Pelargonidin 3-O-glucoside; Pelargonidin 3-O-glucosyl-rutinoside; Pelargonidin 3-O-rutinoside; Pelargonidin 3-O-sophoroside. Figures sourced and adapted from Phenol-Explorer electronic database (3).

Chapter 5: Comparison of flavonoid intake assessment methods Chapter 6: Page 1

Health benefits of non-nutritive food components

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 2

6.1 FOREWORD

Preceding chapters have summarised the results of epidemiological studies to explore the associations of non-nutritive food compounds with health outcomes. In isolation, epidemiological studies allow statements regarding associations between exposure or risk factor and outcome to be made, with exploration of the interrelationships between different risk factors (1). When appropriately designed and implemented, these concluding statements are often extrapolated to the population from which the cohort was obtained. However, as known and unknown confounding factors cannot be adequately controlled for by statistical approaches, causality cannot be inferred from epidemiological studies (2).

Although less generalizable to larger populations, rigorously conducted and controlled randomised controlled trials provide unbiased estimates of treatment effect (2). To gain a deeper understanding of causal effects of non-nutritive food components, a randomised controlled trial of 6 week daily supplementation with probiotics from capsules and yoghurt was implemented. The details and rationale for study design are presented in Appendix A of this thesis. This chapter reports the microbiological results this randomised controlled trial, and explores the ability of yoghurt and probiotic supplementation to induce beneficial changes in fecal bacteria.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 3

6.2 ABSTRACT

American Journal of Clinical Nutrition (Submitted) Kerry L Ivey, Prince RL, Ryan U, Hodgson JM, Lin SY, Yang R

Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial

Background: Beneficial bacteria consumed by mouth (probiotics) have been shown to improve human health in many ways. Recent data also identify effects of the foods themselves on bowel bacterial numbers. Despite a long history of safe use and the incorporation of probiotics into current therapeutic guidelines, the ability of yoghurt and its probiotics to colonize the gastrointestinal tract, and the appropriate mode of administration in order to maximize probiotic efficacy remain uncertain.

Aim: To determine the effect of daily probiotic supplementation of Lactobacillus acidophilus La5 (La5) and Bifidobacterium animalis subsp lactis BB12 (BB12) from capsules and the whole food (yoghurt) form, on faecal counts of Lactobacillus acidophilus La5 (La5), Bifidobacterium animalis subsp lactis BB12 (BB12) and total Bifidobacteria in adults.

Methods: Following a 3-week washout period, 156 men and women over 55 y were randomized to a 6-week double-blinded parallel study. The four intervention groups were: A) probiotic yoghurt plus probiotic capsules; B) probiotic yoghurt plus placebo capsules; C) control milk plus probiotic capsules; and D) control milk plus placebo capsules. Outcome measurements including absolute level and concentration of fecal La5, BB12 and total Bifidobacteria were assessed at baseline and week 6.

Results: When compared to control milk, quantitative PCR (qPCR) analysis showed that probiotic yoghurt supplementation resulted in an increased BB12 count by more than 10,000%, and total Bifidobacteria count by more than 400%. When compared to placebo capsules, probiotic capsules did not affect the number of BB12 or total Bifidobacteria in the feces. Neither probiotic intervention improved La5 count.

Conclusion: Probiotic yoghurt substantially increased beneficial bacteria residing in the gastrointestinal tract. Probiotic capsules did not.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 4

6.3 INTRODUCTION

The human gastrointestinal tract is home to over 100 trillion bacteria, called the microbiome (3). These bacteria play vital roles in production and activation of vitamins, and the synthesis of health promoting molecules from undigested food. Also, by interacting with the immune and nervous systems, the microflora plays important roles in promoting immunity, preventing infection, and altering function of the immune system (4-8).

It is commonly understood that consumption of probiotic bacteria can alter microbiota composition. However, there is limited and varied data to support this hypothesis (9-11).

The health promoting benefits of probiotic consumption have been explored for over 40 years (12). Recent data suggests roles of particular genus and strains of probiotics in improving cardiometabolic and gastrointestinal health (13-17). Despite a long history of safe use and the incorporation of probiotics into current therapeutic guidelines (18, 19), the ability of probiotics to colonize the gastrointestinal tract and the appropriate mode of administration in order to maximize probiotic efficacy remains uncertain.

In industrialized countries, the majority of probiotic bacteria come from supplements, such as probiotic capsules, or bacterially fermented milk, such as yoghurt. It is thought that these different modes of administration may alter the efficacy of probiotic bacteria in colonizing the gastrointestinal tract (20). Biologically active peptides released during the bacterial fermentation of casein proteins in milk improve adhesion to hydrocarbons in the large intestine (21, 22). Conversely, the role of capsules in protecting probiotic bacteria from exposure to acid and digestive enzymes in the upper gastrointestinal tract is thought to improve survival and subsequent colonization of the colon (23). However, the role that mode of administration plays in determining the efficacy of probiotics remains uncertain. As such, this study aimed to determine the

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 5 effect of daily probiotic supplementation, from capsules and the whole food (yoghurt) form, on faecal counts of Lactobacillus acidophilus La5 (La5) and Bifidobacterium animalis subsp lactis BB12 (BB12) in elderly men and women.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 6

6.4 METHODS

6.4.1 Subjects

The cohort consisted of 156 men and women above 55 years, recruited using a population-based approach. Participants were considered eligible for the study if they had minimal usual probiotic intake (consuming less than 400 g yoghurt per week, and not taking probiotic supplements), body mass index ≥ 25 kg/m2, elevated waist circumference (≥ 94 cm in men and ≥ 80cm in women) and an office blood pressure ≥

120/80 mmHg. Participants were excluded if they were unable to complete the study, intolerant to dairy foods, or were taking antibiotics, immunosuppressive treatments or hypoglycemic agents.

6.4.2 Intervention

Prior to commencement of the trial, participants were asked to cease consumption of all foods and products containing probiotic bacteria during the 3-weeks preceding randomization. Subjects were then allocated to 1 of 4, 6-week interventions via computer-generated block randomization.

The four intervention groups in the triple blind, factorial study were: A) probiotic yoghurt plus probiotic capsules; B) probiotic yoghurt plus placebo capsules; C) control milk plus probiotic capsules; or D) control milk plus placebo capsules. Dairy products and capsules were consumed once daily, 30 minutes prior to the first meal of the day.

Both the probiotic yoghurt and probiotic capsules provided a minimum Lactobacillus acidophilus (La5) and Bifidobacterium animalis subsp lactis (Bb12) dose of 3.0 x 109

CFU/d. All capsules were identical in appearance, size, and color and were prepared by

Chr. Hansen (Australia). The probiotic yoghurt (prepared by Casa Dairy Products,

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 7

Australia) and control milk (prepared by Harvey Fresh, Australia) were similar in their nutritional composition.

Written informed consent was obtained in 100% of participants, and the Human

Research Ethics Committee of the University of Western Australia, Perth, Australia, approved the study. The study was carried out in accordance with the World Medical

Association Declaration of Helsinki, and was registered with the Australian New

Zealand Clinical Trials Registry prior to recruitment (ACTRN12612000033842).

6.4.3 Compliance

Compliance was assessed through a compliance diiry and by counting remaining capsules and weighing remaining dairy product at the completion of the study.

6.4.4 Baseline measurements

At baseline, standing height was measured by a wall-mounted stadiometer to the nearest

0.1cm, and body weight was measured by an electronic scale to the nearest 0.1 kg. Body mass index was calculated in kg/m2.

Dietary intake was assessed by a validated semi-quantitative food frequency questionnaire developed by the Anti-Cancer Council of Victoria and a beverage questionnaire (24). Energy and nutrient intakes were estimated based on frequency of consumption and an overall estimate of usual portion size (25).

6.4.5 Measurements of bacterial count

Fecal samples were collected at the end of the washout (baseline) and at the end of the

6-week intervention period (week 6). Participants were instructed to collect stool samples from the bowel motion preceding their baseline and week-6 visits, and stored at

-40C in fecal specimen jars, prior to delivery to the study center at their next visit.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 8

Participants were instructed to avoid contamination of the sample with urine during the collection procedure.

6.4.6 Fecal DNA extraction

Genomic DNA was extracted from 200 mg of each faecal sample using a PowerSoil

DNA purification Kit (MolBiol Carlsbad, California) according to the manufacturer’s instructions with a slight modification (prior to the DNA extraction, the samples were subjected to three cycles of freezing-thawing to ensure complete lysis of all bacteria). DNA was also extracted from individual probiotic capsules by diluting the contents of each capsule in 200 µl of water and then extracting the DNA using the

PowerSoil kit as described above. Extractions on probiotic capsules were performed in triplicate. The DNA concentration of all samples were then analysed (NanoDrop,

ND-1000) according to the user’s manual.

6.4.7 Quantitation of bacteria numbers in probiotic capsules using droplet digital

PCR

The absolute numbers of BB12, LA5, Bifidobacteria and total bacteria in each probiotic capsule was determined using droplet digital PCR (ddPCR). ddPCR was repeated on serial dilutions of capsule DNA from 1:10 to 1:10,000. This data was used as the standard curve for analog qPCR.

The primers and probes used for ddPCR quantitation of the concentrations of BB12, LA5,

Bifidobacteria and total bacteria in each capsule were as previously described (26), with the exception that the probe for LA5 was labelled with Joe at the 5’ end and BHQ1 at 3’ end. The probes for BB12, Bifidobacteria and total bacteria were all labelled with FAM at the 5’ end and BHQ2 at the 3’ end.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 9

Four separate ddPCR runs were conducted in triplicate using a QX100TM droplet digital

PCR system (BioRad, Gladesville, NSW, Australia) according to the manufacturer’s instructions. Briefly, the ddPCR reaction mixture consisted of 12.5 μl of a 2 × ddPCR master mix (Bio-Rad), 2 μl of primer/probe mix (12.5mM each of the primer and probe), 1 μl of

DNA (adjusted to 50 ng/µl) and 9.5 μl of H2O to make a final volume of 25 μl. Droplets were generated using the Droplet Generator (DG) with 70 µL DG Oil per well with a DG8 cartridge and cartridge holder, 25 µL PCR reaction mix, and DG8 gasket. Droplets were dispensed into the 96-well PCR plate by aspirating 40 µL from the DG8 cartridge into each well. The PCR plate was then heat-sealed with a foil seal and the sealed plate was placed in the PCR thermocycler. Cycling consisted of 95°C for 10 min, followed by 45 cycles of 94°C for 30 s and 58°C for 45 s, 1 cycle of 98°C for 10 min with a 12°C hold. After the reaction, the droplets were read using the Droplet Reader, and QuantaSoft software converted the data into the number of template copies per μl of PCR mixture. The number of copies in 1 μl of DNA solution was then calculated.

6.4.8 Quantitation of bacteria numbers in faecal samples using conventional qPCR

Analog (conventional) qPCR testing was used to quantitate the numbers of BB12, LA5,

Bifidobacteria and total bacteria in faecal samples using the standard curve generated from ddPCR (above).

The same primers and probes used in the ddPCR were also used for qPCR, with the exception that the probe for Bifidobacteria was labelled with Cy5 at the 5’ end and the probe for total bacteria was labelled with Rox at the 5’ end. The probes for BB12 and LA5 were labelled with FAM and Joe at their 5’ ends, respectively.

This enabled multiplex detection of all four reactions on a Rotor-Gene 6000 (Qiagen,

Victoria, Australia). The multiplex qPCR reaction mixture consisted of 10 μl of a 2× PCR

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 10 master mix (Roche, Castle, NSW Australia), 4 x 1 μl of each primer/probe mix (containing a

6.25 mM concentration of each primer and probe), 1 μl of faecal DNA extract (adjusted to 50 ng/µl) and H2O to a final volume of 20 μl. Cycling conditions were as follows: 10 min at

95°C, followed by 45 cycles of 95°C for 20 seconds and 60°C for 45 seconds. A standard curve and a negative control were included on all runs.

6.4.9 Blinding and statistical analysis

All study personnel and participants were blinded to treatment assignment for the duration of the study. The investigator involved in PCR analysis was also blinded to treatment allocation until all samples had been analyzed. Prior to analysis, the analytical protocol was prespecified, and all statistical analysis was undertaken using the SPSS statistical package (version 20).

The primary outcomes were absolute fecal La5 and BB12 count in 1 gram of fecal

DNA.Data was first assessed for normality. As a secondary outcome, the effect of the intervention on total Bifidobacteria and total fecal bacteria was explored. All outcome variables were non-normally distributed. As such, outcome data in tables and text was presented as median (inter quartile range). The week-6 outcomes were compared across intervention groups using a multivariable regression model of log transformed outcome variables, with adjustment for the log transformed baseline levels of each outcome, and for the effect of the other intervention (27). As a posthoc analysis, we assessed for interactions between both probiotic interventions. A within subject Wilcoxon Signed

Rank Test was used to explore individual changes in fecal bacteria counts from baseline to week 6.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 11

6.5 RESULTS

6.5.1 Baseline cohort characteristics (Table 1)

At baseline, participants had a mean age of 67 (± 8) years. Treatment groups were well matched at baseline, and there were no significant differences (P > 0.05) between groups for age, sex, body mass index, dietary intake, and bowel medication use.

The baseline faecal samples contained 21,632 (9,117 – 69,901) x 109 bacteria per gram of faecal DNA. La5 and BB12 were detected in the fecal samples of 133 (88%) and 124

(82%) participants, respectively. La5 and BB12 accounted for 2 (± 14) % and 3 (± 14) % of total fecal bacteria, respectively.

6.5.2 Week 6 cohort characteristics (Table 2)

The participant characteristics in Table 1 were similar between the intervention groups at week 6 (P > 0.05). Following the 6 week intervention period, La5 and BB12 were detected in 100% of faecal samples, and the number of total-bacteria in 1 gram of fecal

DNA was 43,267 (10,316 – 102,5985) x 109.

Fecal counts of La5, BB12 and Bifidobacterium increased (P < 0.05) from baseline to week 6 in all 4 intervention groups. The greatest magnitude of change was observed with BB12 in the participants receiving probiotic yoghurt.

6.5.3 Effect of intervention on Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis BB12 (Table 3)

When compared to control milk, probiotic yoghurt supplementation resulted in an increased BB12 and Bifidobacterium count by 83% and 6%, respectively. This effect was not observed after probiotic capsule supplementation (P > 0.05).

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 12

There were no significant differences in La5 count between probiotic and control for both probiotic yoghurt and probiotic capsules.

The interaction between the interventions was investigated as a secondary analysis, and was found to be non-significant for all outcomes (P>0.05). As such, the observed effects of probiotic yoghurt and probiotic capsules did not appear to be influenced by the presence or absence of the other probiotic test article.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 13

6.6 DISCUSSION

This study assessed the effects of probiotic yoghurt, and probiotics administered in capsule form, on faecal content of La5 and BB12, utilising digital PCR techniques. The results of this study showed that dairy supplementation altered fecal microflora composition. This study also found that probiotic yoghurt increased fecal BB12 and

Bifidobacteria count, but probiotic capsules did not.

During the six week intervention period, participants were randomized to consume daily doses of either yoghurt or UHT milk. The fecal content of La5, BB12 and

Bifidobacteria increased in all participants, irrespective of intervention group, from baseline to week 6. This study effect is likely a result of the properties of the dairy intervention articles consumed by all participants. However, the components responsible could be different for milk and yoghurt. The control milk product was pasteurised at ultra-high temperatures using an indirect method (UHT); a process which results in the formation of lactulose at high concentrations (28). Humans lack the digestive enzymes to absorb lactulose, and as such, it passes directly into the large intestine where it preferentially stimulates Bifidobacteria and Lactobacilli genera (29), and increases faecal bacterial counts (30-32). The probiotic yoghurt supplied a daily dose of La5 and

BB12. As such, the UHT form of the milk and the provision of probiotics by the yoghurt test articles may help to explain the presence of a treatment effect in the whole cohort, irrespective of dairy allocation.

Despite the finding that yoghurt and milk both have the ability to improve fecal bacteria content, yoghurt was more efficacious than milk in improving BB12 content; a bacterial strain which has beneficial effects on gastrointestinal and systemic immune function (33,

34). Our findings are in keeping with data from a study by Savard et al. (35) who found that supplementation with a yoghurt containing both La5 and BB12 improved BB12

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 14 content of the feces. However, unlike this study, we did not observe any beneficial effects of probiotic yoghurt on La5 concentrations. Unlike yoghurt, probiotic capsules did not increase the fecal BB12 count. This is supported by results of a randomized controlled trial which found that when compared to placebo, supplementation with capsules containing La5, BB12 as well as other probiotics and prebiotics, did not improve gastric colonization (36). Our data suggests that supplementation with yoghurt containing the probiotic strains La5 and BB12 is a more effective mode of administration for inducing beneficial changes to fecal BB12.

This beneficial change in BB12 following yoghurt supplementation was accompanied by an increase in total fecal Bifidobacteria; a bacterial genus with demonstrated ability to improve colonic transit time (37, 38) and bowel function (39, 40), and reduced gastrointestinal carcinogen exposure (41). The magnitude of the observed increase in total Bifidobacteria was in excess of the increase attributable to BB12; when compared to control milk, probiotic yoghurt increased BB12 count increased by 2,705, whereas the Bifidobacterium count increased by 10,771 bacteria per gram of fecal DNA. This bifidogenic effect of BB12 containing yoghurt may be explained by the BB12 bacteria promoting the abundance of other non-BB12 Bifidobacteria strains (42). However, this bifidogenic effect may also be explained by the yoghurt itself, as supplementation with yoghurt made from Streptococcus thermophiles and Lactobacillus bulgaricus has been shown to increase faecal Bifidobacteria counts, despite the yoghurt test article not containing Bifidobacteria (43). These data support our findings that yoghurt and/or

BB12 supplementation have the potential to increase faecal Bifidobacteria count, independently of the strains administered.

Previous studies of probiotic supplementation on fecal bacteria have utilised conventional qPCR methodologies (26). A limitation of methods applying only conventional qPCR is that external calibrators or normalization to endogenous controls Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 15 are required to estimate the concentration of target DNA in a sample (44). The data generated is only as accurate as the standard curve used to quantitate the samples.

Droplet digital PCR facilitates the accurate and precise quantitation of nucleic acid targets, and offers more accurate quantitation than conventional qPCR, without the need for calibration curves (44-46). Following serial dilutions, the numbers of BB12, LA5,

Bifidobacteria and total bacteria in each probiotic capsule were determined using ddPCR, allowing the generation of more accurate standard curve data.

Our study found that LA5 and BB12 probiotic yoghurt, but not probiotic capsules, increased faecal BB12 concentration, and neither intervention improved La5 concentration. These results are in contrast to a randomised controlled trial of 18-55 year old men and women where La5 and BB12 supplementation increased faecal counts of both strains (47). However, our results are in keeping with data from another human randomised controlled trial demonstrating that yoghurt is more effective than capsules at improving BB12 levels, but neither intervention is effective at improving

Lactobacillus counts (20).

Previous studies of probiotic supplementation on fecal bacteria have utilised conventional qPCR methodologies (26). A limitation of methods applying only conventional qPCR is that external calibrators or normalization to endogenous controls are required to estimate the concentration of target DNA in a sample (44). The data generated is only as accurate as the standard curve used to quantitate the samples.

Droplet digital PCR facilitates the accurate and precise quantitation of nucleic acid targets, and offers more accurate quantitation than conventional qPCR, without the need for calibration curves (44-46). Following serial dilutions, the numbers of BB12, LA5,

Bifidobacteria and total bacteria in each probiotic capsule were determined using ddPCR, allowing the generation of more accurate standard curve data.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 16

In conclusion, we found that supplementation with yoghurt and UHT milk alters microflora composition and that probiotic yoghurt containing BB12 and La5 improves fecal BB12 and Bifidobacteria content, but probiotic capsules do not. The microbiome is an important determinant of health (48), and both BB12 and Bifidobacteria have beneficial effects on gastrointestinal function and composition (33, 34, 37-41). As such, the results of this study suggest that yoghurt is more effective than capsules at promoting beneficial changes to fecal microflora composition and health. Future replication studies, particularly using different strains of probiotic bacteria, are indicated in order to clarify the role that mode of administration plays on efficacy of probiotic bacteria colonization.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 17

6.7 CHAPTER 6 REFERENCES

1. Victora CG, Huttly SR, Fuchs SC, Olinto MT. The role of conceptual frameworks in epidemiological analysis: a hierarchical approach. International journal of epidemiology 1997;26(1):224-7. 2. Kao LS, Tyson JE, Blakely ML, Lally KP. Clinical research methodology I: Introduction to randomized trials. Journal of the American College of Surgeons 2008;206(2):361. 3. Turnbaugh PJ, Ley RE, Hamady M, Fraser-Liggett CM, Knight R, Gordon JI. The human microbiome project. Nature 2007;449(7164):804-10. 4. Hooper LV, Midtvedt T, Gordon JI. How host-microbial interactions shape the nutrient environment of the mammalian intestine. Annual Review of Nutrition 2002;22(1):283-307. 5. Resta SC. Effects of probiotics and commensals on intestinal epithelial physiology: implications for nutrient handling. The Journal of Physiology 2009;587(17):4169-74. 6. Wong JM, de Souza R, Kendall CW, Emam A, Jenkins DJ. Colonic health: fermentation and short chain fatty acids. Journal of Clinical Gastroenterology 2006;40(3):235-43. 7. Pylkas AM, Juneja LR, Slavin JL. Comparison of different fibers for in vitro production of short chain fatty acids by intestinal microflora. Journal of Medicinal Food 2005;8(1):113-6. 8. Forchielli ML, Walker WA. The role of gut-associated lymphoid tissues and mucosal defence. British Journal of Nutrition 2005;93(1):41-8. 9. Kekkonen RA, Lummela N, Karjalainen H, et al. Probiotic intervention has strain-specific anti-inflammatory effects in healthy adults. World Journal of Gastroenterology 2008;14(13):2029-36. 10. Bartosch S, Woodmansey EJ, Paterson JCM, McMurdo MET, Macfarlane GT. Microbiological effects of consuming a synbiotic containing Bifidobacterium bifidum, Bifidobacterium lactis, and oligofructose in elderly persons, determined by Real-Time Polymerase Chain Reaction and counting of viable bacteria. Clinical Infectious Diseases 2005;40(1):28-37. 11. Klein A, Friedrich U, Vogelsang H, Jahreis G. Lactobacillus acidophilus 74-2 and Bifidobacterium animalis subsp lactis DGCC 420 modulate unspecific cellular immune response in healthy adults. European Journal of Clinical Nutrition 2008;62(5):584-93. 12. Mann GV, Spoerry A. Studies of a surfactant and cholesteremia in the Maasai. American Journal of Clinical Nutrition 1974;27(5):464-9. 13. Marteau PR, Vrese Md, Cellier CJ, Schrezenmeir J. Protection from gastrointestinal diseases with the use of probiotics. American Journal of Clinical Nutrition 2001;73(2):430S-6S. 14. Bertolami MC, Faludi AA, Batlouni M. Evaluation of the effects of a new fermented milk product (Gaio) on primary hypercholesterolemia. European Journal of Clinical Nutrition 1999;53(2):97. 15. Schaafsma G, Meuling WJ, Van Dokkum W, Bouley C. Effects of a milk product, fermented by Lactobacillus acidophilus and with

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 18

fructo-oligosaccharides added, on blood lipids in male volunteers. European Journal of Clinical Nutrition 1998;52(6):436. 16. Ataie-Jafari A, Larijani B, Alavi Majd H, Tahbaz F. Cholesterol-lowering effect of probiotic yogurt in comparison with ordinary yogurt in mildly to moderately hypercholesterolemic subjects. Annals of Nutrition and Metabolism 2009;54(1):22-7. 17. Agerbaek M, Gerdes LU, Richelsen B. Hypocholesterolaemic effect of a new fermented milk product in healthy middle-aged men. European Journal of Clinical Nutrition 1995;49(5):346-52. 18. Mowat C, Cole A, Windsor A, et al. Guidelines for the management of inflammatory bowel disease in adults. Gut 2011;60(5):571-607. 19. Mechanick JI, Kushner RF, Sugerman HJ, et al. American Association of Clinical Endocrinologists, the Obesity Society, and American Society for Metabolic and Bariatric Surgery medical guidelines for clinical practice for the perioperative nutritional, metabolic and nonsurgical support of the bariatric surgery patient. Obesity 2009;17(S1):S3-S72. 20. Saxelin M, Lassig A, Karjalainen H, et al. Persistence of probiotic strains in the gastrointestinal tract when administered as capsules, yoghurt, or cheese. International Journal of Food Microbiology 2010;144(2):293-300. 21. Conway PL, Gorbach SL, Goldin BR. Survival of Lactic Acid Bacteria in the Human Stomach and Adhesion to Intestinal Cells. Journal of Dairy Science 1987;70(1):1-12. 22. Meisel H. Overview on milk protein-derived peptides. International Dairy Journal 1998;8(5-6):363-73. 23. Krasaekoopt W, Bhandari B, Deeth H. Evaluation of encapsulation techniques of probiotics for yoghurt. International Dairy Journal 2003;13(1):3-13. 24. Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Australia New Zealand Journal of Public Health 2000;24(6):576-83. 25. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pacific Journal of Clinical Nutrition 1994;3:19-31. 26. Savard P, Lamarche B, Paradis M-E, Thiboutot H, Laurin É, Roy D. Impact of Bifidobacterium animalis subsp.lactis BB-12 and Lactobacillus acidophilus LA-5-containing yoghurt, on fecal bacterial counts of healthy adults. International Journal of Food Microbiology 2011;149(1):50-7. 27. Montgomery A, Peters T, Little P. Design, analysis and presentation of factorial randomised controlled trials. BMC Medical Research Methodology 2003;3(1):26. 28. Marconi E, Messia MC, Amine A, et al. Heat-treated milk differentiation by a sensitive lactulose assay. Food Chemistry 2004;84(3):447-50. 29. Salminen S, Salminen E. Lactulose, lactic acid bacteria, intestinal microecology and mucosal protection. Scandanavian Journal of Gastroenterology 1997;222:45-8. 30. Bouhnik Y, Attar A, Joly FA, Riottot M, Dyard F, Flourie B. Lactulose ingestion increases faecal bifidobacterial counts: a randomised double-blind

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 19

study in healthy humans. European Journal of Clinical Nutrition 2004;58(3):462-6. 31. Tuohy KM, Ziemer CJ, Klinder A, Knöbel Y, Pool-Zobel BL, Gibson GR. A human volunteer study to determine the prebiotic effects of lactulose powder on human colonic microbiota. Microbial Ecology in Health and Disease 2002;14(3):165-73. 32. Terada A, Hara H, Kataoka M, Mitsuoka T. Effect of lactulose on the composition and metabolic activity of the human faecal flora. Microbial Ecology in Health and Disease 1992;5(1):43-50. 33. Kabeerdoss J, Devi RS, Mary RR, et al. Effect of yoghurt containing Bifidobacterium lactis Bb12 on faecal excretion of secretory immunoglobulin A and human beta-defensin 2 in healthy adult volunteers. Nutrition Journal 2011;10(1):1-4. 34. Rizzardini G, Eskesen D, Calder PC, Capetti A, Jespersen L, Clerici M. Evaluation of the immune benefits of two probiotic strains Bifidobacterium animalis ssp. lactis BB-12 and Lactobacillus paracasei ssp. paracasei, L. casei 431 in an influenza vaccination model: a randomised, double-blind, placebo-controlled study. British Journal of Nutrition 2012;107(06):876-84. 35. Savard P, Lamarche B, Paradis M-E, Thiboutot H, Laurin É, Roy D. Impact of Bifidobacterium animalis subsp. lactis BB-12 and, Lactobacillus acidophilus LA-5-containing yoghurt, on fecal bacterial counts of healthy adults. International Journal of Food Microbiology 2011;149(1):50-7. 36. Anderson ADG, McNaught CE, Jain PK, MacFie J. Randomised clinical trial of synbiotic therapy in elective surgical patients. Gut 2004;53(2):241-5. 37. Marteau P, Cuillerier E, Meance S, et al. Bifidobacterium animalis strain DN-173 010 shortens the colonic transit time in healthy women: a double-blind, randomized, controlled study. Alimentary Pharmacology & Therapeutics 2002;16(3):587-93. 38. Bouvier M, Meance S, Bouley C, Berta J-L, Grimaud J-C. Effects of consumption of a milk fermented by the probiotic strain Bifidobacterium animalis DN-173 010 on colonic transit times in healthy humans. Bioscience and Microflora 2001;20(2):43-8. 39. Guyonnet D, Chassany O, Ducrotte P, et al. Effect of a fermented milk containing Bifidobacterium animalis DN-173 010 on the health-related quality of life and symptoms in irritable bowel syndrome in adults in primary care: a multicentre, randomized, double-blind, controlled trial. Alimentary Pharmacology and Therapeutics 2007;26(3):475-86. 40. Agrawal A, Houghton LA, Morris J, et al. Clinical trial: the effects of a fermented milk product containing Bifidobacterium lactis DN-173 010 on abdominal distension and gastrointestinal transit in irritable bowel syndrome with constipation. Alimentary Pharmacology and Therapeutics 2009;29(1):104-14. 41. Picard C, Fioramonti J, Francois A, Robinson T, Neant F, Matuchansky C. Bifidobacteria as probiotic agents: physiological effects and clinical benefits. Alimentary Pharmacology and Therapeutics 2005;22(6):495-512. 42. Mohan R, Koebnick C, Schildt J, et al. Effects of Bifidobacterium lactis Bb12 supplementation on intestinal microbiota of preterm infants: a double-blind, placebo-controlled, randomized study. Journal of Clinical Microbiology 2006;44(11):4025-31. Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 20

43. Bartram HP, Scheppach W, Gerlach S, Ruckdeschel G, Kelber E, Kasper H. Does yogurt enriched with Bifidobacterium longum affect colonic microbiology and fecal metabolites in health subjects? American Journal of Clinical Nutrition 1994;59(2):428-32. 44. Hindson BJ NK, Masquelier DA, Belgrader P, Heredia NJ, Makarewicz AJ, Bright IJ, Lucero MY, Hiddessen, AL, Legler TC, Kitano TK, Hodel MR, Petersen JF, Wyatt PW, Steenblock ER, Shah PH, Bousse LJ, Troup CB, Mellen JC, Wittmann DK, Erndt NG, Cauley TH, Koehler RT, So AP, Dube S, Rose KA, Montesclaros L, Wang S, Stumbo DP, Hodges SP, Romine S, Milanovich FP, White HE, Regan JF, Karlin-Neumann GA, Hindson CM, Saxonov S, Colston BW,. High-throughput droplet digital PCR system for absolute quantitation of DNA copy number. Analytical Chemistry 2011;83:8604-10. 45. Pinheiro LB CV, Hindson CM, Herrmann J, Hindson BJ, Bhat S, Emslie KR,. Evaluation of a droplet digital polymerase chain reaction format for DNA copy number quantification. Analytical Chemistry 2012;84:1003-11. 46. Vogelstein B, Kinzler KW. Digital PCR. Proceedings of the National Academy of Sciences 1999;96:9236-41. 47. Savard P, Lamarche B, Paradis M-E, Thiboutot H, Laurin É, Roy D. Impact of Bifidobacterium animalis subsp. lactis BB-12 and Lactobacillus acidophilus LA-5-containing yoghurt on fecal bacterial counts of healthy adults. International Journal of Food Microbiology 2011;149(1):50-7. 48. Round JL, Mazmanian SK. The gut microbiota shapes intestinal immune responses during health and disease. Nature Reviews Immunology 2009;9(5):313-23.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 21

6.8 TABLES

Chapter 6, Table 1: Baseline characteristics of cohort, stratified by treatment group

Dairy test article Probiotic yoghurt Control milk Capsule test article Probiotic Placebo Probiotic Placebo

Group A Group B Group C Group D

Number (%) 38 (26) 33 (22) 38 (26) 39 (26)

Dietary consumption Energy intake (MJ/d) 7.6 ± 2.6 8.7 ± 7.4 8.2 ± 2.5 7.4 ± 2.3 Fat intake (g/d) 69 ± 26 88 ± 95 81 ± 29 72 ± 27 Carbohydrate intake (g/d) 170 ± 66 178 ± 74 187 ± 59 166 ± 64 Protein intake (g/d) 87 ± 31 85 ± 29 94 ± 35 85 ± 29 Alcohol intake (g/d) 25 ± 25 17 ± 18 17 ± 17 17 ± 18 Fiber intake (g/d) 20 ± 8 21 ± 9 22 ± 8 20 ± 7 Fluid intake (L/d) 2.5 ± 0.9 2.5 ± 0.8 2.4 ± 0.8 2.6 ± 1.3 Caffeinated beverage intake (L/d) 0.8 ± 0.4 0.9 ± 0.5 1.0 ± 0.6 0.8 ± 0.4

Baseline characteristics Sex (male : female) 24 : 15 24 : 10 23 : 15 23 : 17 Age (years) 68 ± 8 68 ± 8 65 ± 7 65 ± 8 Body mass index (kg/m2) 31 ± 4 30 ± 4 31 ± 3 31 ± 3 Bowel medication use [n (%)] 3 (8) 4 (12) 3 (8) 1 (2)

Results are mean ± SD or n (%) where appropriate. No significant between group differences were identified by ANOVA or chi squared test, where appropriate (P>0.05).

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 22

Chapter 6, Table 2: Treatment group summary statistics of bacterial count present in 1 g fecal DNA

Dairy test article Probiotic yoghurt Control milk Capsule test article Probiotic Placebo Probiotic Placebo

Group A Group B Group C Group D

L. acidophilus La5 Baseline (n x 1012) 1 (0-5) 1 (0-14) 1 (0-7) 1 (0-15) Week 6 (n x 1012) 29 (11-89) 15 (2-89) 7 (1-3) 20 (1-69) P value 1 <0.001 <0.001 0.001 <0.001

B. lactis BB12 Baseline (n x 1012) 1 (0-23) 1 (0-5) 1 (0-3) 1 (0-16) Week 6 (n x 1012) 633 (75-4,290) 136 (3-1,922) 32 (4-182) 1 (0-3) P value 1 <0.001 <0.001 <0.001 0.228

Bifidobacterium Baseline (n x 1012) 257 (73-3,668) 132 (79-540) 497 (92-1114) 195 (65-960) Week 6 (n x 1012) 4,443 (901-13,069) 1042 (98-7871) 519 (180-1946) 376 (65-1373) P value 1 0.005 0.021 0.421 0.026

Results are median (IQR). 1 Within subject Wilcoxon signed rank test n: bacterial count per gram of total fecal DNA.

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 6: Page 23

Chapter 6, Table 3: Main effect model of probiotic yoghurt supplementation on absolute bacterial count at 6-weeks

Probiotic yoghurt 1 P value 2 No Yes

La5 (n x 1012) 13 (1-48) 24 (4-89) 0.076

Bb12 (n x 1012) 4 (1-74) 459 (18-2208) <0.001 Bifidobacteria (n x 1012) 512 (106-1,729) 2,220 (423-11,244) <0.001

Probiotic capsule 1 P value 2 No Yes

La5 (n x 1012) 18 (1-71) 16 (2-63) 0.902

Bb12 (n x 1012) 3 (1-228) 99 (17-1,481) <0.001 Bifidobacteria (n x 1012) 527 (92-3,744) 1,122 (414-8,837) 0.084

1 Results are week 6 median (IQR). 2 Univariate ANCOVA model of log transformed bacterial counts between yes and no, adjusted for log transformed baseline values and treatment n: bacterial count per gram of total fecal DNA. The interaction between the interventions was found to be non-significant (P>0.05).

Chapter 6: Effect yoghurt and its probiotics on fecal probiotic content; a randomized controlled trial Chapter 7: Page 1

Health benefits of non-nutritive food components

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial.

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 2

7.1 FOREWORD

The preceding chapter reported the effects of the probiotic intervention (Appendix

A) on faecal bacteria. Despite a well-established role of probiotic yoghurt in improving gastrointestinal conditions such as antibiotic-associated and C. difficile-

-associated diarrhoea (1, 2), the benefits of probiotics are not limited to the gastrointestinal tract. Focus has now turned toward investigating the possible metabolic benefits of probiotic consumption (3-7). This chapter explores the effect of yoghurt and its probiotics on glycaemic control in overweight men and women.

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 3

7.2 ABSTRACT

European Journal of Clinical Nutrition (In Press) Kerry L Ivey, Hodgson JM, Kerr DA, Lewis JR, Thompson PL, Prince RL

The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial.

Background: Evidence from animal and in vitro models suggest a role of probiotic bacteria in improving glycaemic control and delaying the onset of type 2 diabetes. However, the evidence from controlled trials in humans is limited.

Objective: To determine if the probiotic bacteria L. acidophilus La5 and B. animalis subsp lactis Bb12, supplemented in a whole food (yoghurt) or isolated (capsules) form, can improve biomarkers of glycaemic control.

Design: Following a 3-week washout period, 156 overweight men and women over 55 y (mean age: 67 ± 8 years; mean BMI: 31 ± 4 kg/m2) were randomized to a 6-week double-blinded parallel study. The four intervention groups were: A) probiotic yoghurt plus probiotic capsules; B) probiotic yoghurt plus placebo capsules; C) control milk plus probiotic capsules; and D) control milk plus placebo capsules. Outcome measurements including fasting glucose, insulin, glycated haemoglobin and Homeostasis Model Assessment of Insulin Resistance (HOMA-IR), were performed at baseline and week 6.

Results: Relative to the milk control group, probiotic yoghurt resulted in a significantly higher HOMA-IR (0.32 ± 0.15, P=0.038), but did not have a significant effect on the other three measures of glycaemic control (P>0.05). Relative to placebo capsules, probiotic capsules resulted in a significantly higher fasting glucose (0.15 ± 0.07 mmol/L, P=0.037), with no significant effect on the other three measures of glycaemic control (P>0.05). Further analyses did not identify other variables as contributing to these adverse findings.

Conclusion: Data from this study does not support the hypothesis that L. acidophilus La5 and B. animalis subsp lactis Bb12, either in isolated form or as part of a whole food, benefit short-term glycaemic control. Instead, there is weak data for an adverse effect of these strains on glucose homeostasis.

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 4

7.3 INTRODUCTION

At a population level, increased glycaemia is associated with increased risk of micro- and macro-vascular disease (8-11), even in the non-diabetic range (12). Thus population based approaches to improve glycaemia may reduce adverse vascular outcomes. The pathogenesis of impaired glucose tolerance and insulin resistance is complex and multifaceted. In addition to non-modifiable risk factors such as age, genetics and ethnicity, the worldwide epidemic of excessive body fat due to over-nutrition and physical underactivity, substantially contributes to type two diabetes prevalence (13-16).

Interactions between nutrition and the relative abundances of genera comprising the over 100 trillion microorganisms residing in the gastrointestinal tract (17) have also been associated with type two diabetes and related risk factors (18-24).

Recent experimental data provides impetus for further investigation into the role probiotic bacteria can play in improving insulin sensitivity and glucose tolerance (25).

Probiotic bacteria are microorganisms which, when administered in adequate amounts, as either isolated bacteria or in food products, confer a health benefit to the host (26).

The most commonly investigated and verified health benefits of probiotics is their beneficial effect on gastrointestinal outcomes (2). However, recently the effect of probiotic bacteria on metabolic outcomes has been studied (27-30).

The role of probiotics in improving glycaemic control has been explored in a RCT of probiotic supplementation and dietary education in normoglycaemic pregnant women

(31). This study found that in addition to dietary counselling, probiotic supplementation resulted in significantly lower glucose concentrations and reduced risk of elevated blood glucose level. Similarly, probiotic supplementation delayed the onset of glucose intolerance, hyperglycaemia, and hyperinsulinaemia in fructose induced type 2 diabetic

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 5 rats (7), and improved long-term glycaemic control in streptozotocin-induced diabetic rats (32). The glycaemic benefits of probiotics have been attributed to metabolites of these bacteria are which have been shown to affecting biological signalling pathways, modulate genes involved in ubiquitination and proteasomal processes, and alter autonomic nerve activity. (6, 33-37).

Overall the evidence from animal models suggests that probiotics may be useful in improving glycaemic control and delaying onset of type 2 diabetes. However, there is little data to confirm whether these effects are seen in humans. The proposed study aimed to investigate the effects of L. acidophilus La5 and B. animalis subsp lactis Bb12, provided in either yoghurt or capsules, on biomarkers of glycaemic control in overweight men and women.

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 6

7.4 SUBJECTS AND METHODS

7.4.1 Subjects

Between February 2012 and February 2013, 156 men and women were recruited using a population-based approach. A random selection of 8,000 men and women aged above

55 years, who were registered on the Western Australian electoral roll, received a letter inviting them to join the study.

Inclusion criteria included minimal usual probiotic intake (consuming less than 400 g yoghurt per week, and not taking probiotic supplements), body mass index (BMI) ≥ 25 kg/m2, elevated waist circumference (≥ 94 cm in men and ≥ 80cm in women) and an office blood pressure ≥ 120/80 mmHg. Exclusion criteria included: inability to complete the study, intolerance to dairy foods, and the use of antibiotics, immunosuppressive treatments or hypoglycaemic treatments. Of the 887 respondants screened, 156 were considered eligible and were randomised into the study (Figure 1). Prespecified sample size calculations concluded this sample was sufficient to detect a 5% change in fasting glucose concentrations, with 80% power at P=0.05.

7.4.2 Intervention

Participants were asked to cease consumption of all foods and products containing probiotic bacteria during both the 3-week washout and 6-week intervention periods.

Following washout, subjects were allocated to 1 of 4 study treatments via block randomization using computer-generated random numbers, devised by a statistician.

Participants were assigned to receive either: A) probiotic yoghurt plus probiotic capsules; B) probiotic yoghurt plus placebo capsules; C) control milk plus probiotic capsules; or D) control milk plus placebo capsules. Dairy products and capsules were consumed once daily, 30 minutes prior to the first meal of the day. Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 7

Both the probiotic yoghurt and probiotic capsules provided a minimum Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12 dose of 3.0 x 109

CFU/d. All capsules were identical in appearance, size, and colour and were prepared by

Chr Hansen (Australia). The probiotic yoghurt (prepared by Casa Dairy Products,

Australia) and control milk (prepared by Harvey Fresh, Australia) were similar in their nutritional composition. Participants in the control milk group received 8 g protein, 720 kJ, 4 g saturated fat, 12 g carbohydrate. Similarly, participants in the probiotic yoghurt group received 9 g protein, 650 kJ, 4 g saturated fat, 9 g carbohydrate from yoghurt per day.

Written informed consent was obtained in 100% of participants, and the Human

Research Ethics Committee of the University of Western Australia, Perth, Australia, approved the study. The study was carried out in accordance with the World Medical

Association Declaration of Helsinki, and was registered with the Australian New

Zealand Clinical Trials Registry prior to recruitment (ACTRN12612000033842). All data was collected at Sir Charles Gairdner Hospital, Perth, Australia.

7.4.3 Compliance

Compliance was assessed by counting remaining capsules and weighing remaining dairy product at the completion of the study. Adherence was further assessed by a compliance diary whereby participants kept a daily log of test article consumption throughout the intervention period.

7.4.4 Baseline measurements

At the end of the washout (baseline) standing height was measured by a wall-mounted stadiometer to the nearest 0.1cm, and body weight was measured by an electronic scale to the nearest 0.1 kg. Body mass index was calculated in kg/m2. Waist circumference Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 8 was measured by a tape measure to the nearest 0.1 cm at the narrowest part of the torso from the ventral view.

Dietary intake was assessed by a validated semi-quantitative food frequency questionnaire developed by the Anti-Cancer Council of Victoria (38). Energy and nutrient intakes were estimated based on frequency of consumption and an overall estimate of usual portion size (39), and the glycaemic load of the diet was estimated based on published values (40). The international Physical Activity Questionnaire was used to estimate the weekly energy expended in physical tasks, as represented by the metabolic equivalent of task (MET) score (41).

7.4.5 Measurements of glycaemic control

Fasting blood glucose, insulin and glycated haemoglobin (HbA1c) concentrations were assessed at the end of the washout (baseline) and at end of the 6-week intervention period (week 6).

In order to determine effects of the intervention on longer term glycaemic control (42),

HbA1c was measured by the Tina-quant Haemoglobin A1c Gen2 whole blood application (Roche Diagnostics for Integra 800 - [A1C-W, 2007-01, V 3]).

Serum glucose was measured by the Architect c16000 Analyser and serum insulin was measured on the Architect i2000SR Analyser. Glucose and insulin reagents were obtained from Abbott Diagnostics (Abbott Laboratories, Abbott Park, IL 60064, USA).

In order to determine the effect of the intervention on the responsiveness of peripheral tissues to insulin action, the Homeostasis Model Assessment of Insulin Resistance

(HOMA-IR) was calculated with the following formula: fasting serum insulin (mU/ml) x fasting plasma glucose (mmol/l) / 22.5 (43).

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 9

7.4.6 Blinding and statistical analysis

Participants were allocated to a study treatment via block randomization, using computer-generated random numbers (generated by a biostatistician who was not involved in the conduct of the study) sealed in opaque envelopes. All study personnel and participants were blinded to treatment assignment for the duration of the study. A senior investigator not involved in trial implementation held the randomisation code in a password protected folder, which was not broken until the trial had been completed and the analytical protocol had been finalised. All data was analysed according to a pre-specified protocol using SPSS (PASW version 18; IBM Corp., New York, NY,

USA).

The week-6 fasting glucose, insulin, glycated haemoglobin and HOMA-IR were compared across intervention groups using a multivariable regression model, with adjustment for the baseline levels of each outcome, and for the effect of the other intervention (44).

As a secondary analysis, the interaction between probiotic yoghurt and probiotic capsules was explored. Further multivariable regression analyses, adjusting for changes

(week-6 – Baseline) in BMI, waist circumference, physical activity level, glycaemic load, and intakes of energy, fat, carbohydrate and protein, were undertaken in order to explore factors which may contribute to the findings.

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 10

7.5 RESULTS

7.5.1 Participant characteristics and compliance

A total of 60 women and 96 men were randomised (Figure 1), with a mean age of 67 ±

8 years and a mean BMI of 31 ± 4 kg/m2. During the 6-week intervention, 5 participants withdrew from the study: 2 due to a death in the family, and 3 due to illnesses which did not appear to be as a result of the dairy products or capsules. Throughout the study period, all participants remained free of hypoglycaemic agent use, and median compliance was 100%.

Treatment groups were well matched at baseline (Table 1), and there were no significant differences between groups for age, sex, BMI, waist circumference, physical activity level, and dietary intake variables (P>0.05). Similarly, the biomarkers of glycaemic control (Table 2) were not different between intervention groups at baseline (P>0.05). A total of 5 (3%) of participants had a HbA1c value greater than 6.5 % at baseline.

7.5.2 Effect of intervention on biomarkers of glycaemic control

Probiotics from yoghurt or capsules did not significantly alter concentrations of either glycated haemoglobin or insulin relative to control treatments (Table 3). Probiotic yoghurt resulted in higher HOMA-IR (Table 3), whilst probiotic capsules did not significantly alter HOMA-IR (Table 4). Probiotic capsules resulted in higher fasting glucose concentration (Table 4), whereas probiotic yoghurt did not significantly alter fasting glucose (Table 3).

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 11

7.5.3 Exploratory analyses

The interaction between the interventions was investigated as a secondary analysis, and was found to be non-significant: interaction coefficient (HbA1c) = 0.001 (P=0.978); interaction coefficient (glucose) = 0.027 (P=0.871); interaction coefficient (insulin) =

0.684 (P=0.410); interaction coefficient (HOMA-IR) = 0.443 (P=0.507). As such, the observed effects of probiotic yoghurt and probiotic capsules did not appear to be influenced by the presence or absence of the other probiotic test article.

In order to assess how overall glycaemic control at baseline, as assessed by HbA1c, affects responsiveness to the probiotic interventions, we explored the interaction between the interventions and HbA1c. Inclusion of baseline HbA1c in the multivariable regression models did not alter interpretation of results (data not shown).

In order to identify factors which may explain observed results, the degree in which the hyperglycaemia risk factors changed during the intervention period were adjusted for in multivariable regression analyses. Inclusion of these variables in the models did not ameliorate or exacerbate the effect of the interventions on glycaemic outcomes.

Furthermore, we did not observe any significant difference across treatment groups in change (Week 6 – Baseline) of the modifiable risk factors outlined in Table 1 (data not shown).

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 12

7.6 DISCUSSION

Although data from animal studies suggest mechanisms whereby probiotics may benefit glycaemic control and insulin sensitivity (6, 33, 34), the present study suggests that supplementation with L. acidophilus La5 and B. animalis subsp lactis Bb12 does not improve glycaemic control, and may indeed have a slight detrimental effect.

When compared to the appropriate control, capsules containing L. acidophilus La5 and

B. animalis subsp lactis Bb12 marginally increased glycaemia, probiotic yoghurt increased insulin resistance, and no statistically significant effect on other biomarkers of glycaemic control was observed. These results are in contrast to the animal studies and other human trial data that suggest an acute and long-term hypoglycaemic effect of probiotic bacteria (7, 31, 32). However, our results are in keeping with data from other human randomised controlled trials demonstrating no effect of probiotic bacteria on glycaemia (45-47). The discrepancy between findings from human and animal trials may be due to innate biological differences between species, and the subsequent differences in maintenance of glucose homeostasis.

In this regard, the complexity of beneficial and detrimental probiotic-microbiome-host interactions should be recognised (48). In addition to increasing probiotic levels in the gastrointestinal tract, supplementation of probiotic bacteria can also result in proportional reductions in other genera (49). Furthermore, the activities of probiotic bacteria are highly variable and influenced by numerous factors. Gene expression of probiotic bacteria is not only affected by interactions with other bacteria residing in the gastrointestinal tract, but by the genotype of the host (50). This metabolic variance is further complicated by the effect host diet has on probiotic metabolism (51). The numerous factors affecting probiotic metabolism and activity, and the numerous factors

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 13 probiotic bacteria impact on are not yet fully understood, and may explain why in this study, L. acidophilus La5 and B. animalis subsp lactis Bb12 did not exert the hypothesised effects.

Another explanation for the discordance between findings from this and other studies may lie with the variations in study design. Studies observing glycaemic benefits of probiotic supplementation used models of induced diabetes or naturally occurring insulin resistance during pregnancy (7, 31, 32). Gestational state, although associated with insulin resistance, is almost certainly fundamentally different in physiology compared to non-pregnant individuals, primarily due to the variety of differences in hormonal status (52-54). In concert with the other negative probiotic studies (45-47), this cohort largely exhibited good glycaemic control. Therefore, despite not observing a mediating association of baseline long term glycaemic control in this primarily healthy population, we hypothesise that the beneficial effects of probiotics may be limited to pathological states of insulin resistance or type 2 diabetes.

An important but often overlooked factor affecting both metabolic outcomes and ability of the bacteria colonise the gastrointestinal tract, is the bacterial strain and combinations of strains used in probiotic products (55). To date, all the animal and human studies of probiotics on glycaemia have used different strains, combinations of strains or doses of probiotic bacteria, which may help explain the variation in reported glycaemic effects. A strength of this study design is that in addition to being commonly used in the yoghurt and supplement industries, the strains L. acidophilus La5 and B. animalis subsp lactis

Bb12 used in this study were chosen due to their demonstrated capacity to survive the harsh environment of the human gastrointestinal tract (49, 56, 57), adhere to hydrocarbons (58, 59), and exert metabolic benefits (29). However, despite this, glycaemic benefits of these strains were not observed.

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 14

We found that probiotic capsules and probiotic yoghurt had different effects on glycaemic biomarkers. The fasting glucose concentration was significantly higher in the participants taking probiotic capsules, but not the probiotic yoghurt group. The

HOMA-IR was significantly higher in the participants consuming probiotic yoghurt group, but not the probiotic capsules group. This apparent discrepancy in effects of probiotic yoghurt and capsules may be due to either a Type I or type II statistical error.

A post-hoc power calculation showed that the study had only 9% power to detect the observed difference in HOMA-IR for probiotic capsules, and 22% power to detect the observed difference in fasting glucose for probiotic yoghurt. Thus, it may be that there is a negative effect of probiotics on glycaemia, but the effect size is so small that we were underpowered to detect it.

In conclusion, data from this study does not support the hypothesis that L. acidophilus

La5 and B. animalis subsp lactis Bb12, either in isolated form or incorporated into a whole food, benefit short-term glycaemic control in men and women. The effect of probiotic bacteria on metabolism is complex due to both the complexity of host-microbiome interactions and the complexity of strains of probiotic bacteria. Future replication studies, particularly in diabetic patients, are indicated in order to clarify the role of probiotic strains on glycaemic control.

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 15

7.7 CHAPTER 7 REFERENCES

1. Anukam KC, Reid G. Probiotics: 100 years (1907-2007) after Elie Metchnikoff's observation. Communicating Current Research and Educational Topics and Trends in Applied Microbiology 2008:466-74. 2. Marteau PR, Vrese Md, Cellier CJ, Schrezenmeir J. Protection from gastrointestinal diseases with the use of probiotics. American Journal of Clinical Nutrition 2001;73(2):430S-6S. 3. Mann GV, Spoerry A. Studies of a surfactant and cholesteremia in the Maasai. American Journal of Clinical Nutrition 1974;27(5):464-9. 4. Aihara K, Kajimoto O, Hirata H, Takahashi R, Nakamura Y. Effect of powdered fermented milk with Lactobacillus helveticus on subjects with high-normal blood pressure or mild hypertension. Journal of the American College of Nutrition 2005;24(4):257-65. 5. Hata Y, Yamamoto M, Ohni M, Nakajima K, Nakamura Y, Takano T. A placebo-controlled study of the effect of sour milk on blood pressure in hypertensive subjects. American Journal of Clinical Nutrition 1996;64(5):767-71. 6. Matsuzaki T, Yamazaki R, Hashimoto S, Yokokura T. Antidiabetic effects of an oral administartion of Lactobacillus casei in a non-insulin-dependent diabetes mellitus (NIDDM) model using KK-Ay mice. Endocrine Journal 1997;44(3):357-65. 7. Yadav H, Jain S, Sinha PR. Antidiabetic effect of probiotic dahi containing Lactobacillus acidophilus and Lactobacillus casei in high fructose fed rats. Nutrition 2007;23(1):62-8. 8. Pettitt D, Lisse J, Knowler W, Bennett P. Development of retinopathy and proteinuria in relation to plasma-glucose concentrations in Pima Indians. The Lancet 1980;316(8203):1050-2. 9. Laakso M, Kuusisto J. Epidemiological evidence for the association of hyperglycaemia and atherosclerotic vascular disease in non-insulin-dependent diabetes mellitus. Annals of Medicine 1996;28(5):415-8. 10. Kuusisto J, Mykkänen L, Pyörälä K, Laakso M. NIDDM and its metabolic control predict coronary heart disease in elderly subjects. Diabetes 1994;43(8):960-7. 11. Stratton IM, Adler AI, Neil HAW, et al. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. British Medical Journal 2000;321(7258):405-12. 12. Balkau B, Shipley M, Jarrett RJ, et al. High blood glucose concentration is a risk factor for mortality in middle-aged nondiabetic men: 20-year follow-up in the Whitehall Study, the Paris Prospective Study, and the Helsinki Policemen Study. Diabetes Care 1998;21(3):360-7. 13. Alberti KGMM, Zimmet P, Shaw J. International Diabetes Federation: a consensus on Type 2 diabetes prevention. Diabetic Medicine 2007;24(5):451-63.

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 16

14. Drucker DJ, Nauck MA. The incretin system: glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase-4 inhibitors in type 2 diabetes. The Lancet;368(9548):1696-705. 15. Rains JL, Jain SK. Oxidative stress, insulin signaling, and diabetes. Free Radical Biology and Medicine 2011;50(5):567-75. 16. Kolb H, Mandrup-Poulsen T. The global diabetes epidemic as a consequence of lifestyle-induced low-grade inflammation. Diabetologia 2010;53(1):10-20. 17. Bäckhed F, Ley RE, Sonnenburg JL, Peterson DA, Gordon JI. Host-bacterial mutualism in the human intestine. Science 2005;307(5717):1915-20. 18. Larsen N, Vogensen FK, van den Berg FWJ, et al. Gut microbiota in human adults with type 2 diabetes differs from non-diabetic adults. PLoS ONE 2010;5(2):e9085. 19. Furet J-P, Kong L-C, Tap J, et al. Differential adaptation of human gut microbiota to bariatric surgery–induced weight loss: links with metabolic and low-grade inflammation markers. Diabetes 2010;59(12):3049-57. 20. Ding S, Chi MM, Scull BP, et al. High-fat diet: bacteria interactions promote intestinal inflammation which precedes and correlates with obesity and insulin resistance in mouse. PLoS ONE 2010;5(8):e12191. 21. Vijay-Kumar M, Aitken JD, Carvalho FA, et al. Metabolic syndrome and altered gut microbiota in mice lacking toll-like receptor 5. Science 2010;328(5975):228-31. 22. Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER, Gordon JI. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature 2006;444(7122):1027-131. 23. Bäckhed F, Ding H, Wang T, et al. The gut microbiota as an environmental factor that regulates fat storage. Proceedings of the National Academy of Sciences of the United States of America 2004;101(44):15718-23. 24. Yamanaka M, Nomura T, Kametaka M. Influence of intestinal microbes on heat production in germ-free, gnotobiotic and conventional mice. Journal of Nutritional Science and Vitaminology 1977;23(3):221. 25. Lye H, Kuan C, Ewe J, Fung W, Liong M. The improvement of hypertension by probiotics: effects on cholesterol, diabetes, renin, and phytoestrogens. International Journal of Molecular Sciences 2009;10(9):3755-75. 26. Joint FAO/WHO Expert Consultation on Evaluation of Health and Nutritional Properties of Probiotics in Food. Health and nutritional properties of probiotics in food including powder milk with live lactic acid bacteria. Argentina, 2001. 27. Bertolami MC, Faludi AA, Batlouni M. Evaluation of the effects of a new fermented milk product (Gaio) on primary hypercholesterolemia. European Journal of Clinical Nutrition 1999;53(2):97. 28. Schaafsma G, Meuling WJ, Van Dokkum W, Bouley C. Effects of a milk product, fermented by Lactobacillus acidophilus and with fructo-oligosaccharides added, on blood lipids in male volunteers. European Journal of Clinical Nutrition 1998;52(6):436. 29. Ataie-Jafari A, Larijani B, Alavi Majd H, Tahbaz F. Cholesterol-lowering effect of probiotic yogurt in comparison with ordinary yogurt in mildly to moderately hypercholesterolemic subjects. Annals of Nutrition and Metabolism 2009;54(1):22-7.

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 17

30. Agerbaek M, Gerdes LU, Richelsen B. Hypocholesterolaemic effect of a new fermented milk product in healthy middle-aged men. European Journal of Clinical Nutrition 1995;49(5):346-52. 31. Laitinen K, Poussa T, Isolauri E. Probiotics and dietary counselling contribute to glucose regulation during and after pregnancy: a randomised controlled trial. British Journal of Nutrition 2009;101(11):1679-87. 32. Tabuchi M, Ozaki M, Tamura A, et al. Antidiabetic effect of Lactobacillus GG in streptozotocin-induced diabetic rats. Bioscience Biotechnology and Biochemistry 2003;67(6):1421-4. 33. Yamano T, Tanida M, Niijima A, et al. Effects of the probiotic strain Lactobacillus johnsonii strain La1 on autonomic nerves and blood glucose in rats. Life Sciences 2006;79:1963-7. 34. Ljungberg M, Korpela R, Ilonen J, Ludvigsson J, Vaarala O. Probiotics for the prevention of Beta cell autoimmunity in children at genetic risk of type 1 diabetes—the PRODIA study. Annals of the New York Academy of Sciences 2006;1079(1):360-4. 35. Iyer C, Kosters A, Sethi G, Kunnumakkara AB, Aggarwal BB, Versalovic J. Probiotic Lactobacillus reuteri promotes TNF-induced apoptosis in human myeloid leukemia-derived cells by modulation of NF-κB and MAPK signalling. Cellular Microbiology 2008;10(7):1442-52. 36. Calcinaro F, Dionisi S, Marinaro M, et al. Oral probiotic administration induces interleukin-10 production and prevents spontaneous autoimmune diabetes in the non-obese diabetic mouse. Diabetologia 2005;48(8):1565-75. 37. Thomas CM, Versalovic J. Probiotics-host communication: Modulation of signaling pathways in the intestine. Gut Microbes 2010;1(3):148-63. 38. Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Australia New Zealand Journal of Public Health 2000;24(6):576-83. 39. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pacific Journal of Clinical Nutrition 1994;3:19-31. 40. Foster-Powell K, Holt SH, Brand-Miller JC. International table of glycemic index and glycemic load values: 2002. The American Journal of Clinical Nutrition 2002;76(1):5-56. 41. Craig CL, Marshall AL, Sjöström M, et al. International physical activity questionnaire: 12-country reliability and validity. Medicine and Science in Sports and Exercise 2003;35(8):1381-95. 42. World Health Organization. Use of glycated haemoglobin (HbA1c) in the diagnosis of diabetes mellitus. Geneva (Switzerland) 2011. 43. Matthews D, Hosker J, Rudenski A, Naylor B, Treacher D, Turner R. Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28(7):412-9. 44. Montgomery A, Peters T, Little P. Design, analysis and presentation of factorial randomised controlled trials. BMC Medical Research Methodology 2003;3(1):26. Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 18

45. Bukowska H, Pieczul-Mróz J, Jastrzebska M, Chełstowski K, Naruszewicz M. Decrease in fibrinogen and LDL-cholesterol levels upon supplementation of diet with Lactobacillus plantarum in subjects with moderately elevated cholesterol. Atherosclerosis 1998;1998(137):2. 46. Naruszewicz M, Johansson M-L, Zapolska-Downar D, Bukowska H. Effect of Lactobacillus plantarum 299v on cardiovascular disease risk factors in smokers. American Journal of Clinical Nutrition 2002;76(6):1249-55. 47. Sanggaard K, Holst J, Rehfeld J, Sandstrom B, Raben A, Tholstrup T. Different effects of whole milk and a fermented milk with the same fat and lactose content on gastric emptying and postprandial lipaemia, but not on glycaemic response and appetite. British Journal of Nutrition 2004;92:447-59. 48. Shenderov BA. Metabiotics: novel idea or natural development of probiotic conception. Microbial Ecology in Health and Disease 2013;24. 49. Savard P, Lamarche B, Paradis M-E, Thiboutot H, Laurin É, Roy D. Impact of Bifidobacterium animalis subsp. lactis BB-12 and, Lactobacillus acidophilus LA-5-containing yoghurt, on fecal bacterial counts of healthy adults. International Journal of Food Microbiology 2011;149(1):50-7. 50. Sonnenburg JL, Chen CTL, Gordon JI. Genomic and Metabolic Studies of the Impact of Probiotics on a Model Gut Symbiont and Host. PLoS Biol 2006;4(12):e413. 51. Louis P, Scott KP, Duncan SH, Flint HJ. Understanding the effects of diet on bacterial metabolism in the large intestine. Journal of applied microbiology 2007;102(5):1197-208. 52. Bell A, Bauman D. Adaptations of glucose metabolism during pregnancy and lactation. Journal of Mammary Gland Biology and Neoplasia 1997;2(3):265-78. 53. Bauman DE, Bruce Currie W. Partitioning of nutrients during pregnancy and lactation: a review of mechanisms involving homeostasis and homeorhesis. Journal of Dairy Science 1980;63(9):1514-29. 54. Kalkhoff R, Kissebah A, Kim H. Carbohydrate and lipid metabolism during normal pregnancy: relationship to gestational hormone action. Seminars in Perinatology, 1978:291. 55. Lye HS, Rusul G, Liong MT. Removal of cholesterol by lactobacilli via incorporation and conversion to coprostanol. Journal of Dairy Science 2010;93(4):1383-92. 56. Alander M, Mättö J, Kneifel W, et al. Effect of galacto-oligosaccharide supplementation on human faecal microflora and on survival and persistence of Bifidobacterium lactis Bb-12 in the gastrointestinal tract. International Dairy Journal 2001;11(10):817-25. 57. Fukushima Y, Kawata Y, Hara H, Terada A, Mitsuoka T. Effect of a probiotic formula on intestinal immunoglobulin A production in healthy children. International Journal of Food Microbiology 1998;42(1):39-44. 58. Schillinger U, Guigas C, Heinrich Holzapfel W. In vitro adherence and other properties of lactobacilli used in probiotic yoghurt-like products. International Dairy Journal 2005;15(12):1289-97. 59. Collado MC, Jalonen L, Meriluoto J, Salminen S. Protection mechanism of probiotic combination against human pathogens: in vitro adhesion to human intestinal mucus. Asia Pacific Journal of Clinical Nutrition 2006;15(4):570-5.

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 19

7.8 TABLES

Chapter 7, Table 1: Baseline characteristics by treatment group

Dairy test article Probiotic yoghurt Control milk Capsule test article Probiotic Placebo Probiotic Placebo Group A Group B Group C Group D

Number 40 37 39 40

Age (years) 68.4 ± 7.8 68.4 ± 8.7 64.7 ± 7.1 65.4 ± 8.4 Gender (M:F) 25:15 25:12 23:16 23:17 BMI (kg/m2) 30.6 ± 3.8 30.2 ± 4.3 30.8 ± 3.5 30.8 ± 3.5 Waist circumference (cm) 103 ± 10 101 ± 12 100 ± 9 100 ± 9 Physical activity (MET) 111 ± 7 109 ± 8 109 ± 8 111 ± 6 Energy intake (kJ/d) 7590 ± 2649 7473 ± 2433 8199 ± 2505 7367 ± 2299 Glycaemic load 83 ± 35 87 ± 33 94 ± 33 81 ± 34 Fat intake (g/d) 69 ± 26 73 ± 29 81 ± 29 72 ± 27 Carbohydrate intake (g/d) 169 ± 66 171 ± 60 187 ± 59 166 ± 64 Protein intake (g/d) 87 ± 31 94 ± 29 94 ± 35 85 ± 29

1 Results are mean ± SD or n where appropriate. No significant between group differences were identified by ANOVA (P>0.05).

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 20

Chapter 7, Table 2: Treatment group summary statistics of glycaemic parameters at baseline and 6-weeks

Dairy test article Probiotic yoghurt Control milk Capsule test article Probiotic Placebo Probiotic Placebo Group A Group B Group C Group D

Number 40 37 39 40

Fasting glucose Baseline (mmol/L) 5.53 ± 0.57 5.64 ± 1.01 5.59 ± 1.15 5.36 ± 0.55 Week 6 (mmol/L) 5.62 ± 0.65 5.47 ± 0.73 5.58 ± 1.29 5.18 ± 0.65 Change (mmol/L) 0.08 ± 0.39 -0.07 ± 0.43 -0.04 ± 0.36 -0.17 ± 0.50

Fasting insulin Baseline (mU/ml) 9.93 ± 4.75 9.63 ± 4.82 9.77 ± 4.59 9.99 ± 4.49 Week 6 (mU/ml) 11.20 ± 5.27 10.59 ± 6.74 9.74 ± 4.56 10.18 ± 5.36 Change (mmol/L) 1.32 ± 2.76 0.63 ± 3.83 -0.18 ± 3.33 0.18 ± 4.22

HOMA-IR Baseline 2.47 ± 1.27 2.48 ± 1.45 2.45 ± 1.18 2.44 ± 1.29 Week 6 2.85 ± 1.52 2.63 ± 1.91 2.41 ± 1.17 2.38 ± 1.39 Change (mmol/L) 0.39 ± 0.75 0.15 ± 1.04 -0.05 ± 0.87 -0.05 ± 1.011

HbA1c Baseline (%) 5.74 ± 0.41 5.86 ± 0.65 5.83 ± 0.67 5.56 ± 0.36 Week 6 (%) 5.69 ± 0.33 5.74 ± 3.19 5.78 ± 0.64 5.60 ± 0.34 Change (mmol/L) -0.05 ± 0.28 -0.04 ± 0.23 -0.05 ± 0.28 0.28 ± 0.28

Results are mean ± SD. No significant between group differences were identified by ANOVA (P>0.05).

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 21

Chapter 7, Table 3: Main effect model of probiotic yoghurt supplementation on biomarkers of glycaemic control at 6-weeks

Probiotic yoghurt1 Baseline adjusted difference2 P value No Yes

Number 79 77

HbA1c (%) 5.71 ± 0.03 5.69 ± 0.03 -0.02 ± 0.04 0.710 Glucose (mmol/L) 5.40 ± 0.05 5.52 ± 0.05 0.12 ± 0.07 0.094 Insulin (mU/ml) 9.97 ± 0.40 10.92 ± 0.42 0.95 ± 0.58 0.106 HOMA-IR 2.43 ± 0.10 2.75 ± 0.11 0.32 ± 0.15 0.038

1 Results are week 6 mean (± SE), adjusted for baseline values and treatment. 2 Mean difference (± SE) between yes and no. The interaction between the interventions was found to be non-significant (P>0.05).

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 7: Page 22

Chapter 7, Table 4: Main effect model of probiotic capsule supplementation on biomarkers of glycaemic control at week 6

Probiotic capsule1 Baseline adjusted difference2 P value No Yes

Number 77 79

HbA1c (%) 5.71 ± 0.03 5.69 ± 0.03 -0.02 ± 0.04 0.705 Glucose (mmol/L) 5.39 ± 0.05 5.54 ± 0.05 0.15 ± 0.07 0.037 Insulin (mU/ml) 10.37 ± 0.42 10.52 ± 0.40 0.15 ± 0.58 0.796 HOMA-IR 2.53 ± 0.11 2.65 ± 0.10 0.12 ± 0.15 0.419

1 Results are week 6 mean (± SE), adjusted for baseline values and treatment. 2 Mean difference (± SE) between yes and no. The interaction between the interventions was found to be non-significant (P>0.05).

Chapter 7: The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. Chapter 8: Page 1

Health benefits of non-nutritive food components

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial.

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 2

8.1 FOREWORD

Causing more than 17 million mortalities in 2008, cardiovascular diseases are the leading leading cause of deaths worldwide (1). Hypertension and hypercholesterolaemia are major risk factors for cardiovascular disease pathogenesis (2). Therefore, it is important to explore population based approaches to improving cardiovascular disease risk factors. factors.

The preceding chapter explored if supplementation with L. acidophilus La5 and B. lactis Bb12 can improve type 2 diabetes risk factors. This chapter continues the exploration of the metabolic effects of probiotics by exploring if daily supplementation with L. acidophilus La5 and B. lactis Bb12 can improve blood pressure and serum lipid concentrations in overweight men and women.

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 3

8.2 ABSTRACT

British Journal of Nutrition (submitted) Kerry L Ivey, Hodgson JM, Kerr DA, Thompson PL, Prince RL

The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial.

Background: Despite strong mechanistic data, and promising results from in vitro and animal studies, the ability of probiotic bacteria to improve blood pressure and serum lipid concentrations in humans remains uncertain.

Objective: To determine the effect of Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12, provided in either yoghurt or capsule form, on home blood pressure and serum lipid profile.

Subjects and methods: Following a 3-week washout period, 156 overweight men and women over 55 y were randomized to a 6-week double-blinded, factorial, parallel study. The four intervention groups were: A) probiotic yoghurt plus probiotic capsules; B) probiotic yoghurt plus placebo capsules; C) control milk plus probiotic capsules; and D) control milk plus placebo capsules. Each probiotic test article provided a minimum Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp. lactis Bb12 dose of 3.0 CFU/d. Home blood pressure monitoring, consisting of 7-day bi-daily repeat measurements, were collected at baseline and week 6. Fasting total cholesterol, low density lipoprotein cholesterol (LDLC), high density lipoprotein cholesterol (HDLC) and triglyceride were performed at baseline and week 6.

Results: When compared to control milk, probiotic yoghurt did not significantly alter blood pressure, heart rate or serum lipid concentrations (P > 0.05). Similarly, when compared to placebo capsules, supplementation with probiotic capsules did not alter blood pressure or concentrations of total cholesterol LDLC, HDLC, or triglycerides (P > 0.05).

Conclusions: The probiotic strains L. acidophilus La5 and B. animalis subsp. lactis Bb12 did not improve cardiovascular risk factors.

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 4

8.3 INTRODUCTION

Features of the metabolic syndrome (MetS) include elevated triglycerides, lowered high density lipoprotein cholesterol (HDLC), and elevated blood pressure (3). In 2010, 23% of US adults had MetS (4). Compared to their non-MetS counterparts, people with MetS are over 4.6 times more likely to have insulin resistance and 5.5 times more likely to have cardiovascular disease (5). Therefore it is important to identify effective, non-pharmacological, population-based approaches for prevention of MetS and its co-morbidities.

Yoghurt is a complex functional food produced by the probiotic bacterial fermentation of milk. The Wold Health Organisation (6) defines probiotics as live microorganisms which, when administered in adequate amounts, confer a health benefit on the host.

Interestingly, this definition does not stipulate that probiotics improve colonic microflora composition, suggesting that metabolites of probiotics may exert health benefits, independent of gastrointestinal colonisation. The cardiovascular benefits of yoghurt have been investigated since the early 1970s (7), with evidence of beneficial effects (8, 9). Despite this, there remains little understanding of the role of the whole food (yoghurt) in these relationships, and the role probiotic bacteria can play in improving blood pressure and serum lipid concentrations.

Yoghurt contains biologically active peptides produced in the bacterial fermentation of milk. Some evidence suggests these bioactive peptides have ACE-inhibitory and antithrombotic activity that may be responsible for the beneficial effects on features of

MetS, including reduction in blood pressure (9-12). An alternative mode of action of probiotic bacteria is through direct colonisation of the gastrointestinal tract. In this regard several mechanisms have been advanced to explain the hypocholesterolemic

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 5 effect of probiotics including the role of probiotic bacteria in increasing: 1) bile acid deconjugation through the action of bile salt hydrolase (13-16); 2) cholesterol and fatty acid assimilation into probiotic bacteria membranes (17); and 3) conversion of cholesterol to coprostanol in the gastrointestinal tract (17). Despite promising and consistent results from in vitro and animal models (13-17), the hypocholesterolemic effect of probiotics in humans remains uncertain (9, 18, 19). The major limitation of previous trials is the lack of power to detect small but clinically important effects on serum lipids.

This appropriately powered study aims to determine the effect of Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12, provided in either yoghurt or capsule form, on home blood pressure and serum lipid profile men and women with features of the metabolic syndrome.

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 6

8.4 SUBJECTS AND METHODS

8.4.1 Subjects

Participants were randomly recruited from the Perth general population aged over 55 y, using the Western Australian Electoral Roll (20). Inclusion criteria included: minimal usual probiotic intake (consuming < 400 g yoghurt / week, and not taking probiotic supplements), body mass index ≥ 25 kg/m2, elevated waist circumference (≥ 94 cm in men and ≥ 80cm in women), and an office blood pressure ≥ 120/80 mmHg. Exclusion criteria included: inability to complete the study, intolerance to dairy foods, and the use of antibiotics, immunosuppressive treatments or hypoglycaemic agents.

Written informed consent was obtained in 100% of participants, and the Human

Research Ethics Committee of the University of Western Australia, Perth, Australia, approved the study. The study was carried out in accordance with the World Medical

Association Declaration of Helsinki, and was registered with the Australian New

Zealand Clinical Trials Registry prior to recruitment (ACTRN12612000033842).

8.4.2 Study design

A six week randomized, controlled, parallel, double blind, factorial study was performed. Participants were asked to refrain from consumption of all foods and products containing probiotic bacteria from three weeks prior to their baseline visit, and for the duration of the study.

At baseline, participants were randomly assigned by computer-generated random numbers to 1 of 4 intervention groups. Participants were assigned to receive either: A) probiotic yoghurt plus probiotic capsules; B) probiotic yoghurt plus placebo capsules; C) control milk plus probiotic capsules; or D) control milk plus placebo capsules. During the 6 week intervention period, 5 participants withdrew from the study due to personal

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 7 reasons (n = 2) and illnesses unrelated to the test articles (n = 3); A = 1, B = 3, and C = 1

(20).

Dairy products and capsules were consumed once daily for 6 weeks, 30 minutes prior to the first meal of the day. Both the probiotic yoghurt and probiotic capsules provided a minimum Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp. lactis

Bb12 dose of 3.0 CFU/d. All capsules were identical in appearance, size, and colour and were prepared by Chr. Hansen (Australia).

The unflavoured, unsweetened probiotic yoghurt (prepared by Casa Dairy Products,

Australia) provided 650 kJ, 9 g protein, 4 g saturated fat, 9 g carbohydrate per day. The unflavoured, unsweetened control milk (prepared by Harvey Fresh, Australia) provided

720 kJ, 8 g protein, 4 g saturated fat, and 12 g carbohydrate per day.

8.4.3 Baseline and week 6 measurements

Data pertaining to past medical history, medication use and history of cigarette smoking were collected at recruitment, and updated at the 6 clinic visits throughout the 9 week study period. Estimates of energy and nutrient consumption were based on frequency of consumption and an overall estimate of usual portion size obtained from a validated semi-quantitative food frequency questionnaire (21, 22).

Standing height was measured by a wall-mounted stadiometer to the nearest 0.1cm, and body weight was measured by an electronic scale to the nearest 0.1 kg. Body mass index was calculated in kg/m2. The international Physical Activity Questionnaire was used to estimate the weekly energy expended in physical tasks, as represented by the metabolic equivalent of task (MET) score (23).

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 8

8.4.4 Home blood pressure monitoring

Participants performed bi-daily home blood pressure measurements during the 7 days preceding both the baseline and week 6 visits. A fully automated home blood pressure monitor (UA-767PC, A&D, Japan) was provided to each participant at visits held one week prior to baseline, and one week prior to week 6. During these visits, participants were guided through the correct measurement procedure and were asked to perform a blood pressure reading in front of the study coordinator. This was to ensure the reliability of the blood pressure data from home monitoring.

Participants recorded blood pressure twice daily; approximately 1 hour after breakfast and approximately 1 hour after the evening meal. In a quiet room, participants relaxed in a seated position for 5 minutes, and then performed 2 measurements over the following 5 minutes. The blood pressure cuff was positioned on bare skin, and located

2-3 cm above the antecubital position, and the arm was rested in a lateral position with the cuff in line with the heart.

Systolic blood pressure, diastolic blood pressure, pulse pressure and heart rate measurements were automatically recorded on the home blood pressure monitor, and were downloaded when monitors were returned to the study centre at the baseline and week 6 visits. In addition, participants also kept a diary where they manually recorded results of each home blood pressure measurement. In the case of missing or excess data recorded on the device, electronically recorded values were cross referenced with the manual records.

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 9

8.4.5 Measurements of serum lipid profile

At baseline and at 6 weeks, blood samples were collected from the antecubital fossa vein after an overnight fast. All biochemical analyses were performed in the PathWest laboratory at the Royal Perth Hospital, Western Australia.

Serum was analysed for total cholesterol, HDLC and triglycerides using routine analysis based on an enzymatic colorimetric test using the fully automated Architect c16000

Analyser (Abbott Diagnostics, Abbott Laboratories, Abbott Park, IL 60064, USA). Low density lipoprotein cholesterol (LDLC) concentrations were calculated using the

Friedewald formula (24).

8.4.6 Statistical analysis

A pre-specified sample size calculation determined that 68 participants in each main effect treatment arm (probiotic yoghurt or probiotic capsules) would be sufficient to detect a 3.6% change in LDLC (~0.13 mmol/L), with 80% power at P=0.05. With

P=0.05, 68 participants per main effect treatment arm was also calculated to provide >80% power to detect a 3.5 mmHg difference in mean systolic blood pressure

(measured twice daily over 7 days). This number was increased to allow for a predicted drop-out from treatment of around 10%.

All data was analysed according to a pre-specified protocol using SPSS (PASW version

18; IBM Corp., New York, NY, USA). Outcome measurements were compared across intervention groups using a multivariable regression model, with adjustment for the baseline levels of each outcome, and for the effect of the other intervention (25). As a secondary analysis, the interaction between probiotic yoghurt and probiotic capsules was explored.

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 10

8.5 RESULTS

8.5.1 Baseline cohort characteristics

A total of 60 women and 96 men were randomised, with a mean age of 67 ± 8 years and a mean BMI of 31 ± 4 kg/m2. The mean baseline total cholesterol of the cohort was 5.33

(± 1.15) mmol/L. Mean 7-day home systolic and diastolic blood pressures at baseline were 131 (± 12) and 74 (± 9) mmHg, respectively.

Treatment groups were well matched, and there were no significant differences between groups for baseline characteristics (P>0.05); Table 1. Similarly, the blood pressure, heart rate (Table 2), and serum lipid profile (Table 3) of participants were similar between intervention groups at baseline (P>0.05).

8.5.2 Physical activity, energy and nutrient intakes

The level of physical activity as well as intakes of energy and nutrients were similar between the intervention groups at baseline (Table 1) and week 6 (P > 0.05).

Furthermore, we did not observe any significant (P > 0.05) differences across treatment groups in change (Week 6 – Baseline) of these measures (data not shown).

8.5.3 Effect of probiotics on blood pressure

When compared to control milk, probiotic yoghurt did not significantly alter blood pressure or heart rate (P > 0.05). Similarly, when compared to placebo capsules, supplementation with probiotic capsules did not alter blood pressure or heart rate (P >

0.05); Table 4.

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 11

8.5.4 Effect of probiotics on serum lipid profile

When compared to control milk, probiotic yoghurt did not significantly alter lipid concentrations (P > 0.05). Similarly, when compared to placebo capsules, supplementation with probiotic capsules did not alter concentrations of total cholesterol

LDLC, HDLC, or triglycerides (P > 0.05); Table 5.

8.5.5 Secondary analysis

In a secondary analysis, we found no evidence for an interaction between probiotic yoghurt and probiotic capsules (P > 0.05). As such, the lack of significant effects of probiotic yoghurt and probiotic capsules did not appear to be influenced by the presence or absence of the other probiotic test article.

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 12

8.6 DISCUSSION

We assessed the effects of 6 week probiotic supplementation on blood pressure and serum lipid profile in 156 overweight or obese men and women. There was no evidence that blood pressure, heart rate, or lipid concentrations were altered as a result of consuming whole-fat yoghurt or capsules containing the probiotic strains L. acidophilus

La5 and B. animalis subsp. lactis Bb12.

We found that combination therapy of L. acidophilus La5 and B. animalis subsp. lactis

Bb12 had no effect on blood pressure. Our findings of lack of effect are supported by results of other randomised controlled trials which supplemented with the same combination of probiotic strains (26, 27). As the effects of probiotic bacteria are highly strain specific (28), it is not surprising that studies which supplemented with different probiotic strains and species found different results. Specifically, numerous randomised controlled trials of L. helveticus have demonstrated that supplementation with this species has an antihypertensive effect in humans (9, 11, 12). As such, the results of our study support the strain specificity of probiotic actions by confirming that unlike other probiotic preparations, L. acidophilus La5 and B. animalis subsp. lactis Bb12 co-therapy does not improve blood pressure.

Similar to results from other clinical trials (9, 18, 19), our study found that daily probiotic supplementation had no effect on serum lipid concentrations. However, these findings are in contrast to other human trial data demonstrating that probiotic supplementation improves serum lipid concentrations (29, 30). Our results are further opposed by strong in vitro and animal model data which clearly define mechanisms supporting the hypocholesterolaemic effect of probiotics (13-17).

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 13

The lack of cholesterol lowering effect in our, and other studies, may be explained by subject parameters. The mean baseline cholesterol concentration in our cohort cholesterol was 5 mmol/L. This is considered desirable / borderline high according to the third Report of the National Cholesterol Education Program Adult Treatment Panel

III (3).The mean baseline total cholesterol concentration of participants in the other randomised controlled trials that did not observe a beneficial effect was less than 5.4 mmol/L (9, 18, 19). Conversely, the mean baseline total cholesterol concentration of participants in trials which observed a hypocholesterolaemic effect of probiotics was ≥

5.7 mmol/L (29, 30). These findings suggest that the hypocholesterolaemic benefits of probiotics may be limited to populations with borderline high / high baseline total cholesterol levels.

In conclusion, the probiotic strains L. acidophilus La5 and B. animalis subsp. lactis

Bb12 did not improve 7-day home blood pressure. The lack of antihypertensive effect is likely due to the strain specificity of probiotic actions. Furthermore, probiotic supplementation did not improve serum lipid parameters. This lack of effect may be a result of the relatively good baseline cholesterol levels of the cohort. As such, future replication studies supplementing with L. acidophilus La5 and B. animalis subsp. lactis

Bb12 in hypercholesterolemic subjects are indicated.

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 14

8.7 CHAPTER 8 REFERENCES

1. Shanthi M, Pekka P, Bo N, eds. Global atlas on cadiovascular disease prevention and control. Geneva: World Health Organization, World Heart Federation, World Stroke Organization, 2011. 2. Fletcher B, Berra K, Ades P, et al. Managing abnormal blood lipids: a collaborative approach: cosponsored by the Councils on Cardiovascular Nursing; Arteriosclerosis, Thrombosis, and Vascular Biology; Basic Cardiovascular Sciences; Cardiovascular Disease in the Young; Clinical Cardiology; Epidemiology and Prevention; Nutrition, Physical Activity, and Metabolism; and Stroke; and the Preventive Cardiovascular Nurses Association. Circulation 2005;112(20):3184-209. 3. Third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III) final report. Circulation 2002;106(25):3143. 4. Beltrán-Sánchez H, Harhay MO, Harhay MM, McElligott S. Prevalence and trends of metabolic syndrome in the adult U.S. population, 1999–2010. Journal of the American College of Cardiology 2013;62(8):697-703. 5. Cameron AJ, Magliano DJ, Zimmet PZ, Welborn T, Shaw JE. The metabolic syndrome in Australia: prevalence using four definitions. Diabetes Research and Clinical Practice 2007;77(3):471-8. 6. Joint FAO/WHO Expert Consultation on Evaluation of Health and Nutritional Properties of Probiotics in Food. Health and nutritional properties of probiotics in food including powder milk with live lactic acid bacteria. Argentina, 2001. 7. Mann GV, Spoerry A. Studies of a surfactant and cholesteremia in the Maasai. American Journal of Clinical Nutrition 1974;27(5):464-9. 8. Aihara K, Kajimoto O, Hirata H, Takahashi R, Nakamura Y. Effect of powdered fermented milk with Lactobacillus helveticus on subjects with high-normal blood pressure or mild hypertension. Journal of the American College of Nutrition 2005;24(4):257-65. 9. Hata Y, Yamamoto M, Ohni M, Nakajima K, Nakamura Y, Takano T. A placebo-controlled study of the effect of sour milk on blood pressure in hypertensive subjects. American Journal of Clinical Nutrition 1996;64(5):767-71. 10. Meisel H. Overview on milk protein-derived peptides. International Dairy Journal 1998;8(5-6):363-73. 11. Jauhiainen T, Vapaatalo H, Poussa T, Kyronpalo S, Rasmussen M, Korpela R. Lactobacillus helveticus Fermented milk lowers blood pressure in hypertensive subjects in 24-h ambulatory blood pressure measurement. American Journal of Hypertension 2005;18(12):1600-5. 12. Seppo L, Jauhiainen T, Poussa T, Korpela R. A fermented milk high in bioactive peptides has a blood pressure lowering effect in hypertensive subjects. American Journal of Clinical Nutrition 2003;77(2):326-30. 13. Cardona ME, Vanay VdV, Midtvedt T, Norin E. Probiotics in gnotobiotic mice Conversion of cholesterol to coprostanol in vitro and in vivo and bile acid

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 15

deconjugation in vitro. Microbial Ecology in Health and Disease 2000;12(4):219-24. 14. Pereira DIA, McCartney AL, Gibson GR. An in vitro study of the probiotic potential of a bile-salt-hydrolyzing Lactobacillus fermentum strain, and determination of its cholesterol-lowering properties. Applied and Environmental Microbiology 2003;69(8):4743-52. 15. McAuliffe O, Cano RJ, Klaenhammer TR. Genetic analysis of two bile salt hydrolase activities in Lactobacillus acidophilus NCFM. Applied and Environmental Microbiology 2005;71(8):4925-9. 16. Moser SA, Savage DC. Bile salt hydrolase activity and resistance to toxicity of conjugated bile salts are unrelated properties in Lactobacilli. Applied and Environmental Microbiology 2001;67(8):3476-80. 17. Lye HS, Rusul G, Liong MT. Removal of cholesterol by lactobacilli via incorporation and conversion to coprostanol. Journal of Dairy Science 2010;93(4):1383-92. 18. de Roos N, Schouten G, Katan M. Yoghurt enriched with Lactobacillus acidophilus does not lower blood lipids in healthy men and women with normal to borderline high serum cholesterol levels. European Journal of Clinical Nutrition 1999;53(4):277-80. 19. Mizushima S, Ohshige K, Watanabe J, et al. Randomized controlled trial of sour milk on blood pressure in borderline hypertensive men. American Journal of Hypertension 2004;17(8):701-6. 20. Ivey KL, Hodgson JM, Kerr DA, Lewis JR, Thompson PL, Prince RL. The effects of probiotic bacteria on glycaemic control in overweight men and women: a randomised controlled trial. European Journal of Clinical Nutrition 2013;(In Press). 21. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pacific Journal of Clinical Nutrition 1994;3:19-31. 22. Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Australia New Zealand Journal of Public Health 2000;24(6):576-83. 23. Craig CL, Marshall AL, Sjöström M, et al. International physical activity questionnaire: 12-country reliability and validity. Medicine and Science in Sports and Exercise 2003;35(8):1381-95. 24. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clinical Chemistry 1972;18(6):499-502. 25. Montgomery A, Peters T, Little P. Design, analysis and presentation of factorial randomised controlled trials. BMC Medical Research Methodology 2003;3(1):26. 26. Asemi Z, Samimi M, Tabassi Z, et al. Effect of daily consumption of probiotic yoghurt on insulin resistance in pregnant women: a randomized controlled trial. European Journal of Clinical Nutrition 2012. 27. Zarrati M, Shidfar F, Nourijelyani K, et al. Lactobacillus acidophilus La5, Bifidobacterium BB12, and Lactobacillus casei DN001 modulate gene Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 16

expression of subset specific transcription factors and cytokines in peripheral blood mononuclear cells of obese and overweight people. BioFactors 2013;39(6):633-43. 28. Joint FAO/WHO Expert Consultation. Guidelines for the evaluation of probiotics in food. Canada, 2002. 29. Anderson JW, Gilliland SE. Effect of fermented milk (yogurt) containing Lactobacillus acidophilus L1 on serum cholesterol in hypercholesterolemic humans. Journal of the American College of Nutrition 1999;18(1):43-50. 30. Ataie-Jafari A, Larijani B, Alavi Majd H, Tahbaz F. Cholesterol-lowering effect of probiotic yogurt in comparison with ordinary yogurt in mildly to moderately hypercholesterolemic subjects. Annals of Nutrition and Metabolism 2009;54(1):22-7.

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 17

8.8 TABLES

Chapter 8, Table 1: Baseline characteristics of participants by treatment group

Dairy test article Probiotic yoghurt Control milk

Capsule test article Probiotic Placebo Probiotic Placebo

Group A Group B Group C Group D

Number 40 37 39 40

Baseline characteristics Age (years) 68 ± 8 68 ± 8.7 65 ± 7 65 ± 8 Gender (M:F) 25:15 25:12 23:16 23:17 Body mass index (kg/m2) 31 ± 4 30 ± 4 31 ± 4 31 ± 4 Hypolipidaemic agent use [n(%)] 25 (62) 21 (57) 20 (51) 27 (68) History of Cigarette smoking [n(%)] 21 (52) 21 (57) 21 (46) 21 (52) Prevalent cardiovascular disease [n(%)] 12 (30) 13 (35) 16 (41) 18 (45) Physical activity (MET) 111 ± 7 109 ± 8 109 ± 8 111 ± 6

Dietary intake Energy (kJ/d) 7.6 ± 2.6 7.4 ± 2.4 8.2 ± 2.5 7.4 ± 2.3 Saturated fat (g/d) 27 ± 12 29 ± 12 33 ± 14 30 ± 13 Cholesterol (g/d) 303 ± 133 297 ± 122 335 ± 159 308 ± 134 Fibre (g/d) 20 ± 8 21 ± 8 22 ± 8 20 ± 7 Sugar (g/d) 80 ± 34 74 ± 26 82 ± 23 73 ± 31

Results are mean ± SD or n (%) where appropriate. No significant between group differences were identified by ANOVA or chi squared test, where appropriate (P>0.05).

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 18

Chapter 8, Table 2: Seven day home blood pressure parameters stratified by treatment group

Dairy test article Probiotic yoghurt Control milk

Capsule test article Probiotic Placebo Probiotic Placebo

Group A Group B Group C Group D

Number 40 37 39 40

Systolic blood pressure Baseline (mmHg) 131 ± 13 130 ± 12 132 ± 12 130 ± 12 Week 6 (mmHg) 131 ± 14 129 ± 11 130 ± 12 129 ± 13

Diastolic blood pressure Baseline (mmHg) 74 ± 11 74 ± 7 76 ± 10 74 ± 7 Week 6 (mmHg) 74 ± 10 75 ± 7 75 ± 9 73 ± 8

Pulse pressure Baseline (mmHg) 57 ± 12 57 ± 11 56 ± 14 56 ± 12 Week 6 (mmHg) 57 ± 12 55 ± 10 56 ± 13 56 ± 14

Heart rate Baseline (bpm) 71 ± 9 71 ± 9 70 ± 14 72 ± 12 Week 6 (bpm) 70 ± 9 70 ± 9 70 ± 14 72 ± 11

Results are mean ± SD. No significant between group differences were identified by ANOVA (P>0.05).

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 19

Chapter 8, Table 3: Serum lipid concentrations stratified by treatment group

Dairy test article Probiotic yoghurt Control milk

Capsule test article Probiotic Placebo Probiotic Placebo

Group A Group B Group C Group D

Number 40 37 39 40

Total cholesterol Baseline (mmol/L) 5.52 ± 0.95 5.35 ± 1.40 5.20 ± 1.12 5.22 ± 1.12 Week 6 (mmol/L) 5.37 ± 0.88 5.32 ± 1.33 5.15 ± 1.10 5.21 ± 1.09

LDLC Baseline (mmol/L) 3.31 ± 0.77 3.22 ± 1.16 3.07 ± 0.92 3.10 ± 0.95 Week 6 (mmol/L) 3.18 ± 0.75 3.20 ± 1.18 3.03 ± 0.87 3.10 ± 0.95

HDLC Baseline (mmol/L) 1.42 ± 0.37 1.36 ± 0.32 1.41 ± 0.33 1.42 ± 0.35 Week 6 (mmol/L) 1.40 ± 0.34 1.36 ± 0.32 1.39 ± 0.33 1.44 ± 0.36

Triglycerides Baseline (mmol/L) 1.70 ± 0.77 1.70 ± 0.94 1.58 ± 0.74 1.53 ± 0.79 Week 6 (mmol/L) 1.71 ± 0.75 1.65 ± 0.77 1.59 ± 0.82 1.47 ± 0.62

Results are mean ± SD. No significant between group differences were identified by ANOVA (P>0.05).

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 20

Chapter 8, Table 4: Main effect model of yoghurt and probiotic supplementation on home blood pressure parameters

Treatment arm1 Baseline adjusted P value difference2 Yoghurt supplementation No Yes Systolic blood pressure (mmHg) 129 ± 1 130 ± 1 1.1 ± 1.8 0.548 Diastolic blood pressure (mmHg) 74 ± 1 75 ± 1 1.1 ± 1.3 0.422 Pulse pressure (mmHg) 56 ± 1 56 ± 1 -0.1 ± 1.8 0.933 Heart rate (bpm) 71 ± 1 70 ± 1 -0.6 ± 1.7 0.703

Probiotic supplementation No Yes Systolic blood pressure (mmHg) 129 ± 1 130 ± 1 0.9 ± 1.8 0.601 Diastolic blood pressure (mmHg) 74 ± 1 75 ± 1 0.8 ± 1.3 0.543 Pulse pressure (mmHg) 56 ± 1 56 ± 1 0.9 ± 1.8 0.615 Heart rate (bpm) 70 ± 1 70 ± 1 -0.1 ± 1.7 0.950

1 Baseline and treatment adjusted week 6 values (±SE). 2 Mean difference between yes and no (± SE).

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 8: Page 21

Chapter 8, Table 5: Main effect model of yoghurt and probiotic supplementation on serum lipid parameters

Treatment arm1 Baseline adjusted P value difference2 Yoghurt supplementation No Yes Total cholesterol (mmol/L) 5.29 ± 0.06 5.23 ± 0.06 -0.06 ± 0.09 0.477 LDLC (mmol/L) 3.16 ± 0.05 3.10 ± 0.05 -0.06 ± 0.07 0.410 HDLC (mmol/L) 1.42 ± 0.02 1.38 ± 0.02 -0.0.4 ± 0.02 0.126 Triglyceride (mmol/L) 1.58 ± 0.05 1.63 ± 0.06 0.05 ± 0.08 0.490

Probiotic supplementation No Yes Total cholesterol (mmol/L) 5.29 ± 0.06 5.24 ± 0.06 -0.06 ± 0.08 0.513 LDLC (mmol/L) 3.16 ± 0.05 3.10 ± 0.05 -0.06 ± 0.07 0.377 HDLC (mmol/L) 1.41 ± 0.02 1.39 ± 0.02 -0.02 ± 0.02 0.356 Triglyceride (mmol/L) 1.57 ± 0.06 1.64 ± 0.05 0.07 ± 0.08 0.376

1 Baseline and treatment adjusted week 6 values (±SE). 2 Mean difference between yes and no (± SE).

Chapter 8: The effect of yoghurt and its probiotics on blood pressure and serum lipid profile; a randomised controlled trial. Chapter 9: Page 1

Health benefits of non-nutritive food components

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review)

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 2

9.1 FOREWORD

The preceding three chapters have reported the results of a randomised controlled trial (Appendix A). As previously described, there are many benefits of conducting randomised controlled trials, namely the ability to draw causative conclusions regarding the results. However, randomised controlled trials are not free of bias.

In randomised controlled trials, bias is introduced when a systematic error in study design or implementation skews the outcomes of the study (1). One way of addressing issues of bias in individual randomised controlled trials, and to draw conclusions regarding treatment efficacy is to conduct a systematic review which draws together results from all available clinical trials. The Cochrane Collaboration is an international network that routinely publishes systematic reviews in human health care and policy aimed at providing the highest standard of evidence based health care (2). This Chapter presents a Cochrane approved systematic review and meta-analysis aimed at summarising all available randomised controlled trial evidence describing the efficacy of probiotics, in the form of isolated or fermented milk, in improving serum lipid concentrations.

The following manuscript has been submitted to the Cochrane Library of Systematic

Reviews. As such, this chapter is presented in the format of a full Cochrane systematic review and meta-analysis.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 3

9.2 ABSTRACT

Cochrane Library of Systematic Reviews (protocol: 2013,3) (full review: in preparation) Kerry L Ivey, Hodgson JM, Kim S, Woodman RJ, Prince RL.

Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults

Background: Dyslipidaemia is a significant risk factor for cardiovascular diseases. Probiotic bacteria are live microorganisms that improve human health when consumed. Supplementation with probiotic bacteria may be effective in improving serum lipid profile hence reducing risk and incidence of cardiovascular disease.

Objective: To assess the effect of probiotic bacteria, administered in the form of isolated bacteria or fermented dairy products for the primary prevention of cardiovascular disease in adults.

Search methods: We searched the Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library, MEDLINE Ovid and EMBASE Oviedo (all to June 2013) using an approved search protocol. We also searched databases of ongoing trials and reference lists of relevant studies and reviews.

Selection criteria: We considered randomised, controlled, parallel and crossover, clinical trials in the general adult population. Included studies compared the effect of probiotic bacteria, administered in the form of isolated bacteria or fermented dairy products, versus appropriate placebo on cardiovascular disease mortality or serum lipid concentrations; specifically total cholesterol, low density lipoprotein cholesterol (LDLC), high density lipoprotein cholesterol (HDLC) or triglyceride concentrations.

Data collection and analysis: Two review authors independently abstracted relevant population and intervention characteristics. We resolved any disagreements through discussion, or if required by a third party. The risk of bias of trials was assessed against key criteria: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other sources of bias.

For continuous variables, the data of interest was mean and SD of the change in lipid concentrations (mmol/L) from baseline to end of intervention for both probiotic and placebo intervention groups. If studies used more than one control or intervention group, we only analysed pairs of comparable test articles. Pooled mean differences with 95% confidence intervals were calculated, and meta-analyses were performed using Review Manager (version 5.2).

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 4

Main results: This review included 23 randomised controlled clinical trials consisting of 16 parallel group studies and 7 crossover studies with a median intervention period of 6 weeks (range 2 - 56 weeks). A total of 1,310 subjects participated in these studies.

The median intervention period was 6 weeks (range 2 - 56 weeks). The mean daily dose of fermented milk in included studies was 256 (range: 50 - 450) grams, and the mean daily dose of probiotic bacteria was 9.60 (range: 7.00 - 11.85) log colony forming units. The mean age of participants was 49 years. There was no outcome data on cardiovascular events, death from any cause, health-related quality of life and costs. When compared to placebo, probiotic supplementation did not increase risk of adverse events.

Based on low risk of bias in included studies, and high degree of homogeneity of the effect of probiotics, the quality of evidence provided by this review is considered to be high.

Compared with placebo, probiotic supplementation improved total cholesterol [mean difference: -0.16 (-0.23,-0.10) mmol/L, P < 0.00001] and LDLC [mean difference: -0.13 (-0.19, -0.07) mmol/L, P < 0.0001] concentrations. When compared to studies of participants with desirable baseline total-cholesterol concentrations, the beneficial effect on total cholesterol and LDLC was to be limited to participants with higher total cholesterol at baseline; P< 0.0001. Furthermore, when compared to probiotic bacteria from fermented milk, probiotic bacteria from capsules were more effective at lowering total cholesterol and LDLC concentrations; P < 0.0001.

Compared with placebo, probiotic supplementation had no effect on HDLC or triglyceride concentrations, and the lack of effect was not ameliorated by considering baseline total-cholesterol levels or mode of administration.

Authors' conclusions: Based on high quality evidence, we found that probiotic supplementation improved both total cholesterol LDLC concentrations, without eliciting adverse effects. Probiotics are more likely to have beneficial effects when supplemented in capsule form, or when administered to patients with higher total cholesterol concentrations at baseline. No conclusions regarding the effect of probiotics on cardiovascular disease mortality could be reached.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 5

9.3 PLAIN LANGUAGE SUMMARY

9.3.1 Research question

To assess the effect of probiotic bacteria, administered in the form of isolated bacteria or fermented dairy products primary prevention of cardiovascular disease in adults.

9.3.2 Background

High blood cholesterol levels increase the risk of heart disease in the community.

Therefore, it is important to find non-pharmacological population based approaches to improving cholesterol levels. Probiotics were more likely to have beneficial effects when supplemented in capsule form, or when administered to patients with higher total cholesterol concentrations at baseline.

Probiotics are the beneficial bacteria found in many supplement and fermented milk products. There is strong data from in vitro and animal studies that suggest probiotics may play a role in improving cholesterol levels. However, based on results of numerous small clinical trials, the role that probiotics play in humans is less clear.

9.3.3 Study characteristics

Twenty three randomised controlled clinical trials that provided data for this review. A total of 1,310 patients were assigned to daily probiotic supplementation, in isolated

(capsule) or whole food (fermented milk) form, versus placebo. The median intervention period was 6 weeks (range 2 - 56 weeks), and the mean age of participants was 49 years. Of the 23 included studies, 16 implemented a parallel design, and the remaining 7 were crossover studies.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 6

9.3.4 Key results

The data show that supplementation with probiotics reduces total cholesterol by 0.16

(-0.23, -0.10) mmol/L and low density lipoprotein cholesterol (LDLC) by 0.13 (-0.19,

-0.07) mmol/L. Probiotics are more likely to have beneficial effects when supplemented in capsule form, or when administered to patients with higher total cholesterol concentrations at baseline.

9.3.5 Quality of the evidence

Based on low risk of bias in included studies, and high degree of homogeneity of the effect of probiotics, the quality of evidence provided by this review is considered to be high.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 7

9.4 BACKGROUND

9.4.1 Description of the condition

Non-communicable diseases are a leading cause of death in the adult population the world over. Accounting for nearly 50% of all non-communicable disease mortalities, cardiovascular disease (CVD) contributes significantly to global burden of disease (3).

Despite current projections predicting CVD prevalence to increase by 10% by the year

2030 (4), CVD is preventable (5). Prevention of CVD is achieved through modulation of modifiable risk factors. Elevated serum cholesterol concentrations have been attributed to causing more than 2.6 million deaths worldwide (6). Therefore, the improvement of cholesterol concentrations, specifically circulating concentration of low density lipoprotein cholesterol (LDLC), is a primary target of CVD prevention strategies (7).

Elevated serum cholesterol concentrations contributes significant to global morbidity, mortality and disease burden (8). A reduction in serum cholesterol levels by 10%, which can be achieved by dietary interventions, may halve the risk of Ischaemic heart disease

(9) and would allow 75% of Americans to reach their LDL goals without the need for pharmacotherapy (10). With more than 50% of US adults reporting high blood cholesterol concentrations (11), it is important to identify effective, non-pharmacological, population-based approaches to improving cholesterol levels, and reducing CVD outcomes.

9.4.2 Description of the intervention

Probiotic bacteria are live microorganisms that improve human health when consumed.

Examples of the beneficial actions of probiotics include reducing the incidence of acute

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 8 upper respiratory tract infections (12), reducing duration and stool frequency in infectious diarrhoea (13) and persistent diarrhoea in children (14). The definition of probiotics has evolved to reflect developments in the understanding of probiotic actions.

Early definitions characterised probiotics as bacteria that improved intestinal microflora composition, however, more recent definitions have broadened to encompass additional health benefits of probiotic consumption (15). The World Health Organization (WHO)

(16) defines probiotics as live microorganisms which, when administered in adequate amounts, confer a health benefit on the host. Interestingly, this definition does not stipulate that probiotics improve colonic microflora composition, suggesting that metabolites of probiotics may exert health benefits, independent of gastrointestinal colonisation.

Bacterially fermented milk products, such as yogurt and sour milk, typically contain probiotic cultures of Lactobacillus acidophilus, Bifidobacteria and Lactobacillus casei

(17). Fermented foods account for approximately one-third of the world’s food intake

(18), and the production of yogurt throughout the world has been steadily increasing

(19). Therefore, any health benefits of yogurt consumption could have a significant impact on public health globally.

Stemming from work in the early 20th century, there is now a well-established role for probiotic yogurt in improving gastrointestinal conditions such as antibiotic-associated and Clostridium difficile-associated diarrhoea (20, 21). The focus has now turned toward investigating the metabolic benefits of probiotic consumption. The cardiovascular benefits of yogurt have been investigated since the early 1970s (22), with evidence of beneficial effects; despite this, there remains little understanding of the role the whole food (fermented milk), as compared to isolated probiotic bacteria, plays in these relationships. The area of probiotic research is complex, and the current scientific Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 9 literature is conflicting, likely due to differences in study design. This systematic review aims to provide a concise assessment of this difficult topic.

9.4.3 Adverse effects of the intervention

Commonly used probiotic genera such as Lactobacilli, Lactococci and Bifidobacteria are generally recognised as safe (23). Probiotics have a long history of use in commercial food products. No serious adverse events have been reported in randomised controlled trials in healthy adults. However, meta-analyses have found a very small proportion of probiotic supplementation can result in mild gastrointestinal changes such as increased thirst, constipation, vomiting, flatulence, nausea and loose bowels (12, 24).

9.4.4 How the intervention might work

Numerous mechanisms to explain the hypocholesterolaemic effect of probiotics have been hypothesised, including the role of probiotic bacteria in increasing: 1) bile acid deconjugation through the action of bile salt hydrolase (25-28); 2) cholesterol and fatty acid assimilation into probiotic bacteria membranes (29); and 3) conversion of cholesterol to coprostanol in the gastrointestinal tract (29).

The extent and ability of probiotics to perform these functions appears to be highly strain-specific (29), and in vivo is likely affected by host microflora and the ability of the particular strain to survive bile and acid conditions and colonise the gastrointestinal tract.

In vitro and animal model evidence suggests that bioactive peptides in fermented milk may exert hypocholesterolaemic effects. However, it is unclear whether when looking at the evidence as a whole, probiotic bacteria have a significant role to play in the prevention or management of dyslipidaemia. Furthermore, it is unclear whether the dose

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 10 or mode of administration (probiotic bacteria in fermented milk versus isolated probiotic bacteria) influences the hypocholesterolaemic effect.

9.4.5 Why it is important to do this review

The results of this systematic review will clarify the role regular probiotic intake plays in the prevention of cardiovascular disease. Through the use of conservative exclusion criteria, the results of this systematic review may translate into an applicable intervention strategy that is relevant at both a programme and an individual level. If it is established that probiotic supplementation improves serum lipid profile and provides direct and measurable reduction in CVD risk, the subsequent alteration in dietary advice would promote consumption to those at risk of hyperlipidaemia. Ultimately, what would be small and easily achieved changes to dietary patterns could affect serum lipid profile and CVD risk in populations.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 11

9.5 OBJECTIVES

To assess the effects of probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 12

9.6 METHODS: CRITERIA FOR CONSIDERING STUDIES FOR THIS

REVIEW

9.6.1 Types of studies

Randomised controlled clinical trials that reported incidence outcomes for cardiovascular disease events or serum lipid levels were considered for inclusion.

9.6.2 Types of participants

Only studies in the general adult population, over 18 years, were included. Studies were excluded if participants were children, pregnant, lactating, stoma patients or critically ill patients (i.e. undergoing recent surgery or experiencing acute illness) as they may react to the intervention in a different way; however some degree of CVD was permissible.

9.6.3 Types of interventions

We investigated the following comparisons of intervention versus control/comparator where the same letters indicate direct comparisons.

9.6.3.1 Interventions a) Probiotic bacteria supplementation, administered in the form of isolated bacteria (in

capsules, powders etc.) b) Fermented dairy products containing probiotic bacteria.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 13

9.6.3.2 Comparators a1) Placebo, defined as an identical capsule or powder not containing probiotic bacteria. b1) Placebo, defined as fermented milk that is either chemically fermented or fermented using traditional starter cultures, identical to those used in the test article.

Concomitant therapies had to be the same in the intervention and comparator groups.

However, studies were excluded if probiotic mode of administration is other than fermented milk or isolated bacteria, for example fruit juices. Studies were excluded if studies contain products known to affect serum lipid profile or intestinal microflora composition, including prebiotics (such as fructo-oligosaccharides) or soy protein.

Only trials with a control arm that allowed observed effects to be reasonably ascribed to probiotics were included.

9.6.4 Types of outcome measures

9.6.4.1 Primary outcomes

1) Cardiovascular events

2) Health-related quality of life

3) Adverse events

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 14

9.6.4.2 Secondary outcomes

1) All-cause mortality

2) Serum concentrations of total cholesterol, low-density lipoprotein cholesterol

(LDLC), high-density lipoprotein cholesterol (HDLC) and triglycerides

3) Socioeconomic costs

9.6.4.3 Method of outcome measurement

# Cardiovascular events: defined as hospitalisation or death from a coronary death

event, myocardial infarction, coronary insufficiency, angina, ischaemic stroke,

haemorrhagic stroke, transient ischaemic attack, peripheral artery disease and/or

heart failure.

# Adverse effects such as gastrointestinal discomfort.

# Serum total cholesterol, LDLC, HDLC and triglycerides

# Health-related quality of life: measured with a validated instrument

# All-cause mortality: verified using death certificates

# Socioeconomic costs

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 15

9.6.4.4 Timing of outcome measurement

If studies used more than one outcome analysis time point, we analysed the outcomes assessed at the full study (intervention) period. If studies used more than one control or intervention group, we only analysed pairs of comparable test articles.

We established a 'Summary of findings' table using the following outcomes listed according to priority:

1) Cardiovascular events

2) All-cause mortality

3) Health-related quality of life

4) Adverse effects

5) Serum total cholesterol concentration

6) Socioeconomic costs

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 16

9.7 METHODS: SEARCH METHODS FOR IDENTIFICATION OF STUDIES

9.7.1 Electronic searches

We used the following sources from inception to the present.

# Cochrane Library (27 June 2013).

# MEDLINE (until 27 June 2013).

# EMBASE (until 27 June 2013).

We also searched databases of ongoing trials (www.clinicaltrials.gov/), Current

Controlled Trials metaRegister (www.controlled-trials.com/), the EU Clinical Trials register (www.clinicaltrialsregister.eu/) and the WHO International Clinical Trials

Registry Platform (http://apps.who.int/trialsearch/). For detailed search strategies see

Appendix 1.

If additional key words of relevance had been detected during any of the electronic or other searches, we would have modified the electronic search strategies to incorporate these terms. We placed no restrictions on the language of publication when searching the electronic databases or reviewing reference lists in identified studies.

9.7.2 Searching other resources

We tried to identify additional studies by searching the reference lists of included trials and (systematic) reviews, meta-analyses and health technology assessment reports on the topic of probiotics and cardiovascular disease.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 17

9.8 METHODS: DATA COLLECTION AND ANALYSIS

9.8.1 Selection of studies

Two review authors (KI, JH) independently read the abstract, title or both sections of every record retrieved to determine the studies to be assessed further. We present a

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) flow-chart of study selection (Figure 1) (30).

9.8.2 Data extraction and management

For studies that fulfil inclusion criteria, two review authors (KI, JH) independently abstracted relevant population and intervention characteristics using standard tables in

Review Manager (version 5.2). All disagreements were be resolved by discussion, or, when required, by a third party.

We tried to find the protocol of each included study, either in databases of ongoing trials, in publications of study designs, or both.

In the event of duplicate publications, companion documents or multiple reports of a primary study, we maximised yield of information by collating all available data.

9.8.3 Assessment of risk of bias in included studies

Two review authors (KI, JH) assessed each trial independently. Disagreements were resolved by consensus, or with consultation of a third party.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 18

We assessed risk of bias using the Cochrane Collaboration’s tool (31, 32). We used the following bias criteria:

# Random sequence generation (selection bias).

# Allocation concealment (selection bias).

# Blinding (performance bias and detection bias), separated for blinding of

participants and personnel and blinding of outcome assessment.

# Incomplete outcome data (attrition bias).

# Selective reporting (reporting bias).

We assessed outcome reporting bias (33) by integrating the results of 'Examination of outcome reporting bias' (Appendix 7), 'Matrix of study endpoints (protocol/trial documents)' (Appendix 6) and section 'Outcomes (outcomes reported in abstract of publication)' of the 'Characteristics of included studies' table. This analysis formed the basis for the judgement of selective reporting (reporting bias).

We judged risk of bias criteria as 'low risk', 'high risk' or 'unclear risk' and evaluate individual bias items as described in the Cochrane Handbook for Systematic Reviews of

Interventions (32). We present a 'Risk of bias' graph (Figure 2) and a 'Risk of bias summary' (Figure 3).

For blinding of participants and personnel (performance bias), detection bias (blinding of outcome assessors) and attrition bias (incomplete outcome data) we intended to evaluate risk of bias separately for subjective and objective outcomes (34). We considered the implications of missing outcome data from individual participants.

We defined the following endpoints as subjective outcomes: Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 19

1) Adverse events.

2) Health-related quality of life.

We defined the following outcomes as objective outcomes:

1) All-cause mortality.

2) Cardiovascular events.

3) Serum total cholesterol concentration.

4) Serum lipid concentrations.

5) Socioeconomic costs.

9.8.4 Measures of treatment effect

For continuous variables, the data of interest was mean and SD of the change in lipid concentrations (mmol/L) from baseline to end of intervention for both probiotic and placebo intervention groups. Pooled mean differences with 95% confidence intervals were calculated, and meta-analyses were performed using Review Manager (version

5.2).

When only baseline and end of intervention data was provided, the mean and SD of the change was computed by two authors independently (KLI and SWK), in accordance with the methods outlined in the Cochrane Handbook for Systematic Reviews of

Interventions (32).

If studies used more than one control or intervention group, we only analysed pairs of comparable test articles. In some cases, where appropriate, data from different probiotic

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 20 or placebo intervention groups within the same studies were combined, to provide collated mean and SD of change for both probiotic and placebo intervention groups.

If outcome data was not able to be computed from the data provided in original publications, authors of the study were contacted for (for details see Appendix 12).

9.8.5 Unit of analysis issues

The data of interest was mean and SD of the change in lipid concentrations in mmol/L.

Where data was presented in publications as mg/dL, the conversion factors of 38.7 for cholesterol and for 88.5 triglycerides were applied.

9.8.6 Dealing with missing data

We investigated attrition rates, for example drop-outs, losses to follow-up and withdrawals, and critically appraised issues of missing data and imputation methods.

9.8.7 Assessment of heterogeneity

In the event of substantial clinical, methodological or statistical heterogeneity, we did not report study results as meta-analytically pooled effect estimates.

We identified heterogeneity by visual inspection of the forest plots and by using a standard Chi2 test with a significance level of α = 0.1, in view of the low power of this test. We examined heterogeneity using the I2 statistic, which quantifies inconsistency across studies to assess the impact of heterogeneity on the meta-analysis (35, 36), where an I2 statistic of 75% or more indicates a considerable level of inconsistency (32).

Had we found heterogeneity, we would have attempted to determine potential reasons for it by examining individual study and subgroup characteristics.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 21

We expected the following characteristics to introduce clinical heterogeneity:

# Differences in baseline cardiovascular disease risk.

# Age of participants and average age of experiencing a cardiovascular event during

follow-up.

# Different mode of administration

9.8.8 Assessment of reporting biases

As we included more than 10 studies for each lipid outcome, we used funnel plots to assess small study effects. Due to several explanations for funnel plot asymmetry we interpreted results carefully (37).

9.8.9 Data synthesis

We performed statistical analyses according to the statistical guidelines referenced in the latest version of the Cochrane Handbook for Systematic Reviews of Interventions

(32). Pooled mean differences with 95% confidence intervals were calculated using

Review Manager (version 5.2).

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 22

9.8.10 Subgroup analysis and investigation of heterogeneity

We carried out the following subgroup analyses and investigated interaction:

# Stratification of studies into 2 groups based on mode of test article administration

(fermented milk versus isolated probiotic bacteria).

# Stratification of studies into 3 groups based on mean baseline total cholesterol

concentration; desirable (<5.168 mmol/L), borderline high (5.168-6.176 mmol/L),

and high (> 6.176 mmol/L) (7).

9.8.11 Sensitivity analysis

We performed sensitivity analyses in order to explore the influence of the following factors on effect size:

# Restricting the analysis to published studies

# Restricting the analysis to very large studies (studies with greater than 100

participants included in analysis) to establish how much they dominate the results.

In order to test the robustness of the results, analyses were repeated using random-effects statistical models.

As all studies were considered to have a low degree of bias (see Assessment of risk of bias in included studies), sensitivity analyses taking into account risk of bias were not performed.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 23

9.9 MAIN RESULTS: RESULTS OF THE SEARCH

9.9.1 Search results

The initial search identified 3,651 records. Searching of clinical trial registries identified a further 6 studies for which we obtained the full text. After manually removing all duplicates, we pre-selected examined 2,386 abstracts on the basis of their titles or abstracts. We excluded 2,318 records because they did not meet the inclusion criteria or were not relevant to the question under study, and reviewed 68 full-text articles. After screening the full-text of the selected publications and hand searching reference lists, 23 studies (21 publications) met the inclusion criteria. See Figure 1 for the amended

PRISMA flow chart.

Each stage of selection of studies was followed by discussions with the group. The process resulted in agreement in the pre-selection stage, the quality of the assessment, and in the data extraction phase.

9.9.2 Excluded studies

Forty four studies had to be excluded after careful evaluation of the full publication (see

Figure 1). The main reasons for exclusion were that studies did not meet inclusion criteria based on interventions, control, participants or design. For further details, see

Characteristics of excluded studies.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 24

9.10 MAIN RESULTS: INCLUDED STUDIES

A detailed description of the characteristics of included studies is presented elsewhere

(see Characteristics of included studies and appendices). The following is a succinct overview:

9.10.1 Source of data

For this review, 22 randomised, controlled trials were identified through screening of the Medline, Embase and Cochrane Library search results and hand searching. One unpublished study was also included in this review (38).

9.10.2 Overview of study populations

A total of 1,310 participants were included in the 23 trials (Appendix 3). The individual sample size ranged from 13 (39) to 156 (38)

9.10.3 Settings

The study by (40) was conducted in a nursing home setting. All other studies were conducted in a community setting. Seventeen trials were conducted in economically developed countries. The remaining 6 trials were conducted in Slovakia (40), Iran

(41-43) and Estonia (39, 44).

9.10.4 Participants

For details of participating populations, see Appendix 3. The mean age of patients was

49 years. There were 569 males 681 females, and 60 patients without gender specification (45). Two studies (43, 46) did not include male participants, and the

Songisepp study 2 (44) only included 1 (6%) male. In three studies (47-49) no female individuals participated. Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 25

9.10.4.1 Baseline characteristics

For details of baseline characteristics, see Appendix 4. The mean total cholesterol of participants at baseline was 5.49 mmol/L, and one study (50) did not report baseline total cholesterol. The participants of 8 studies had desirable mean baseline total cholesterol (39, 42, 43, 46, 48, 51-53). Nine studies had borderline-high mean baseline total cholesterol (38, 40, 41, 44, 47, 54-57), The participants of the remaining 5 studies had a high mean baseline total cholesterol concentration (45, 49, 56, 58, 59).

The mean baseline LDLC of included participants was 3.52 mmol/L, and studies by (4,

47, 48, 55) did not report baseline LDLC concentrations.

The mean HDLC at baseline was 1.38 mmol/L, with 2 studies (50, 55) not reporting baseline HDLC concentration. The baseline triglyceride concentration was 1.47 mmol/L, and was reported in 100% of studies.

9.10.4.2 Inclusion criteria

For complete inclusion criteria of included studies, see Characteristics of included studies. Two studies included overweight or obese participants (38, 51) whereas 6 studies were limited to populations exhibiting some degree of hypercholesterolaemia

(41, 45, 56, 58-60). The study by Ejtahed et al. (42) only included participants with type 2 diabetes mellitus, and 5 of the studies (47, 50, 53-55) only included subjects exhibiting some degree of hypertension. One study (40) did not report any inclusion criteria.

9.10.4.3 Exclusion criteria

For complete exclusion criteria of included studies, see Characteristics of included studies. Generally, exclusion criteria were the presence of diabetes (39, 41, 43, 44, Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 26

50-53, 58, 59) and some degree of cardiovascular disease or previous cardiovascular event (39, 41, 44, 45, 50, 52, 53, 56, 59, 60). Eleven studies did not include participants using lipid lowering medications (39, 41-45, 52, 56, 58-60).

Two studies excluded participants with familial or secondary dyslipidaemia (45, 58).

Three studies did not report any exclusion criteria (40, 48, 49).

9.10.5 Study design

Studies were single centred randomised controlled trials and all studies used a placebo control. Seven trials adopted a cross-over design (39-41, 44, 46, 48, 58, 61), with cross over periods ranging from 2-4 weeks. Sixteen implemented a parallel designed trial (38,

42, 43, 45, 47, 49-57, 59, 60).

In terms of blinding, 15 studies were double-blinded for participants and personnel (38,

39, 42-45, 50-53, 56-58, 60), 3 studies were single-blinded for participants (41, 49, 55), and in 5 studies, blinding was not defined (40, 46-48, 54). No studies described blinding of outcome.

Date of intervention period was generally poorly described. However, publication dates ranged from 1996 - 2013.

The duration of interventions ranged from 2 to 56 weeks, with a median study period of

6 weeks. The durations of both the run-in and follow-up periods ranged from 0 to 4 weeks.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 27

9.10.6 Comparisons

Five studies investigated isolated probiotic supplements versus placebo (40, 45, 51, 56,

59). Seventeen studies investigated probiotic fermented milk versus placebo (39, 41-44,

46-50, 52-55, 57, 58, 60).

The study by Ivey et al. (38) used a factorial study design and investigated probiotics from both capsules and fermented milk, with appropriate placebo products for both capsules and milk.

9.10.7 Interventions

9.10.7.1 Probiotic interventions

In all studies of isolated probiotic bacteria, the bacteria were administered in capsule form (38, 40, 45, 56, 59). The two studies by Songisepp et al. (39, 44)supplemented with cheese as the fermented milk product, and the test articles implemented by

Kekkonen et al. (57) was a milk based fruit drink.

The study design of Agerholm-Larsen et al. (51) was such that there was three different probiotic intervention groups exploring 3 different probiotic species and combinations.

Ivey et al. (38) implemented 3 different probiotic intervention groups that involved different doses and modes of administration of identical probiotic strains. Usinger et al.

(53) implemented 2 probiotic intervention groups using 2 different daily doses of an identical probiotic fermented milk product.

Sixteen different species or combinations of species were used by the 23 studies (Table

2). The most commonly supplemented probiotic species or combinations of species used in the included studies were Lactobacillus fermentum (39, 44, 59), and the combination of Lactobacillus acidophilus and Bifidobacterium lactis (38, 41-43). Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 28

The mean daily probiotic dose was 9.60 (range: 7.00 - 11.85) log colony forming units

(CFU). Four reports did not specify the dose of probiotic test articles (47, 50, 53, 55).

Of the studies of fermented milk supplementation, the mean daily dose of probiotic fermented milk was 256 (range: 50 - 450) grams.

9.10.7.2 Control interventions

In all studies using probiotic capsules, the corresponding control product was a placebo capsule (38, 40, 45, 56, 59).

In studies of fermented milk, the corresponding control product for 9 studies was milk fermented with identical genus and species of starter cultures to that used in the probiotic fermented milk product (39, 41-44, 46, 49, 52, 58). Three reports did not provide sufficient information to determine whether starter cultures were identical in probiotic and control products (50, 57, 60). Acidified milk fermented was the control product for 5 studies (47, 51, 53-55). The control product implemented by Rizkalla et al.

(48) was a pasteurised fermented milk, and the control product corresponding to yogurt in the Ivey et al. study (38) was milk.

9.10.8 Outcomes

Only two studies explicitly stated primary or secondary endpoints in the trial documents.

For a summary of all outcomes assessed in each study, see Appendix 5; Appendix 6.

Cardiovascular events: No trial included information on cardiovascular events.

Health-related quality of life: No trial included information on health-related quality of life.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 29

Adverse events: Seven studies reported adverse events. All adverse events were considered mild.

All-cause mortality: No trial included information on all-cause mortality.

Serum lipid concentrations: All studies used standard/published methods for determining serum lipid concentrations.

Total cholesterol: Total serum cholesterol concentrations were reported in 22

studies (n=1,282).

Low-density lipoprotein cholesterol: Serum LDLC concentrations were reported in

19 studies (n=1,224).

High-density lipoprotein cholesterol: Serum HDLC concentrations were reported in

21 studies (n=1,247).

Triglycerides: Serum tgl concentrations were reported in 23 studies (n=1,381).

Socioeconomic costs: No trial included information on socioeconomic costs.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 30

9.10.9 Risk of bias in included studies

For an overview of review authors' judgments about each risk of bias item for individual studies and across all studies see Figure 2 and Figure 3.

9.10.9.1 Allocation (selection bias)

All included studies were randomised controlled clinical trials. However, an adequate description of sequence generation methods was reported in only 3 studies (38, 56, 58).

An adequate description of how allocation concealment was performed was observed in

2 reports (50, 56). The remainder of the studies either did not perform allocation concealment, or did not describe methods in the publications

9.10.9.2 Blinding (performance bias and detection bias)

Sixteen studies explicitly stated that blinding of the participants and personnel was undertaken (n = 1,111). Three studies reported that single blinding was undertaken, though it was unclear as to who and how this was achieved (n = 81). Four studies did not provide sufficient information about blinding procedures (n = 189).

9.10.9.3 Incomplete outcome data (attrition bias)

We did not observe attrition bias in any studies.

9.10.9.4 Selective reporting (reporting bias)

We did not observe attrition bias in any studies.

9.10.9.5 Other potential sources of bias

No other sources of bias were detected.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 31

9.11 MAIN RESULTS: EFFECTS OF INTERVENTIONS ON PRIMARY

OUTCOMES

None of the included studies included information on cardiovascular events or health-related quality of life.

9.11.1 Adverse events

For details of adverse event data, see Appendix 9.

No studies reported serious adverse events associated with the interventions.

Despite Hata et al. (54) reporting one participant from the probiotic group complaining of stomach ache, and two participants from placebo group complaining of diarrhoea, the authors state that "these symptoms were not attributed to the treatment". As such, this data was not incorporated into the meta-analysis of adverse events. Two studies (38, 44) recorded adverse event data, and reported that no adverse events occurred in the either the probiotic or control groups.

Of the 1,310 participants in this review, adverse events were only reported in13 (6%) participants receiving probiotic interventions, and in 4 (2%) participants receiving the placebo intervention groups. Mild gastrointestinal complaints accounted for all adverse events. Results of the meta-analysis of the 6 studies reporting adverse events (39, 49,

51-53, 59) showed that although there were more minor adverse events in those receiving probiotic supplementation, there increase in risk was not statistically significant [Odds Ratio 2.49 (0.89, 7.02)] with a high degree of homogeneity; I2 = 0%, P

= 0.56 (Analysis 1.8).

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 32

9.12 MAIN RESULTS: EFFECTS OF INTERVENTIONS ON

SECONDARY OUTCOMES

None of the included studies included information on all-cause mortality or socioeconomic costs.

9.12.1 Total cholesterol

9.12.1.1 Primary analysis

Meta-analysis of all 22 studies reporting total cholesterol outcomes (Analysis 1.1) revealed a significant cholesterol lowering effect of probiotics compared with placebo; mean difference: -0.16 (-0.23,-0.10) mmol/L, P < 0.00001 (Figure 5).The test for heterogeneity (P = 0.06) provides evidence that the different probiotic interventions share a common effect size.

9.12.1.2 Publication bias in primary analysis

The funnel plot of the primary total cholesterol analysis (Figure 6) was roughly symmetrical, with the exception of two outliers (41, 56) showing beneficial effects.

However, when these studies were excluded to improve symmetry of the funnel plot

(Figure 8), probiotic bacteria continued to have a beneficial effect on total cholesterol concentrations; mean difference -0.11 (-0.18, -0.04) mmol/L (Analysis 1.2).

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 33

9.12.1.3 Sensitivity analyses

Restricting the analysis to published studies

The study by Ivey et al. (38) was the only unpublished study. When this study was excluded from the meta-analysis, probiotic bacteria continued to have a beneficial effect on total cholesterol concentrations; mean difference -0.17 (-0.24, -0.10) mmol/L.

(Analysis 2.1)

Repeating analysis taking into account sample size

The effect size of probiotic supplementation on total cholesterol concentration was greater (P = 0.010) in studies with large sample size [mean difference -0.33 (-0.47, -0.19) mmol/L] when compared to studies with small sample size [-0.12 (-0.19, -0.04) mmol/L]. (Analysis 4.1)

Repeating analysis using Random Effect model

When primary analysis was repeated using a Random Effects model, probiotic bacteria continued to have a significant effect on cholesterol concentrations; mean difference

-0.16 (-0.25, -0.08) mmol/L. (Analysis 5.1)

9.12.1.4 Subgroup analyses

Comparison of effect size with probiotic bacteria from capsules and fermented milk

In this meta-analysis, probiotic bacteria were more effective (P = 0.003) at lowering total cholesterol when administered in capsules [mean difference -0.33 (-0.46, -0.20) mmol/L] compared with fermented milk; mean difference -0.10 (-0.18, -0.02) mmol/L

(Figure 13). The effect of isolated probiotics and probiotics from fermented milk were homogenous within each administration mode (P > 0.05). (Analysis 6.1) Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 34

Comparison of effect size across different baseline total concentrations

Probiotic interventions did not improve total cholesterol concentrations in participants with desirable cholesterol levels at baseline [mean difference -0.07 (-0.17, 0.03) mmol/L]. Conversely, probiotic supplementation did improve total cholesterol concentrations borderline-high and high total cholesterol at baseline; mean difference

-0.24 (-0.36, -0.13) mmol/L, and mean difference -0.22 (-0.37, -0.08) mmol/L, respectively (Figure 14). (Analysis 7.1)

Effect sizes were homogenous in the studies of participants with desirable and high baseline total cholesterol (P>0.05). Substantial heterogeneity was observed between the studies with borderline-high baseline total cholesterol; I2 = 59%, P = 0.01.

The median baseline total cholesterol concentration of studies with borderline-high total cholesterol was 5.39 mmol/L. To explore the source of heterogeneity, we dichotomised the studies of borderline-high baseline total cholesterol into 2 groups (Analysis 7.2); ≤ median (38, 47, 54, 55, 57) and > median (40, 41, 44, 56) baseline total cholesterol.

There was a significant difference between effect estimates of the two groups (P =

0.0001). Of the studies with borderline-high total cholesterol at baseline, probiotic supplementation had no effect on total cholesterol in studies with baseline total cholesterol ≤ median; mean difference -0.10 (-0.25, 0.04) mmol/L. Conversely, probiotic bacteria improved total cholesterol concentration in studies in the > median group; mean difference -0.45 (-0.62, -0.27) mmol/L (Figure 15).

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 35

9.12.2 Low density lipoprotein cholesterol

9.12.2.1 Primary analysis

A meta-analysis of all 19 studies reporting effects of interventions on LDLC concentration (Analysis 1.3) revealed a significant LDLC lowering effect of probiotics compared with placebo; mean difference: -0.13 (-0.19, -0.07) mmol/L, P < 0.0001

(Figure 16). Despite results being significantly heterogeneous (P = 0.04), the degree of inconsistency was considered low (I2 = 39%).

9.12.2.2 Publication bias in primary analysis

The funnel plot of the primary LDLC analysis (Figure 17) was roughly symmetrical, with the exception of two outlying studies showing beneficial effects (56, 60).

Exclusion of these studies improved symmetry of the funnel plot (Figure 19) and removed heterogeneity (P = 0.80). Following removal of these two studies, probiotic bacteria continued to lower LDLC concentrations; mean difference -0.08 (-0.14, -0.01) mmol/L (Analysis 1.4).

9.12.2.3 Sensitivity analyses

Restricting the analysis to published studies

Exclusion of the study by Ivey et al. (38) 2013 did not remove the beneficial effect of probiotic bacteria on LDLC concentrations; mean difference -0.14 (-0.21, -0.08) mmol/L. (Analysis 2.2)

Repeating analysis taking into account sample size)

There were no differences in effect size between studies of different sample size (P =

0.07). Probiotic supplementation improved LDLC in studies with small [mean Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 36 difference -0.09 (-0.17, -0.01) mmol/L] and large [-0.21 (-0.32, -0.10) mmol/L] sample size. (Analysis 4.2)

Repeating analysis using Random Effect model

When primary analysis was repeated using a Random Effects model, probiotic bacteria continued to have a significant effect on cholesterol concentrations; mean difference

-0.14 (-0.23, -0.05) mmol/L. (Analysis 5.2)

9.12.2.4 Subgroup analyses

Comparison of effect size with probiotic bacteria from capsules and fermented milk

When compared to probiotic bacteria from fermented milk [mean difference -0.07

(-0.15, 0.00) mmol/L], probiotic bacteria from capsules [mean difference -0.26 (-0.38,

-0.15) mmol/L] were more effective at improving LDLC concentrations; P = 0.006

(Figure 24). (Analysis 6.2)

Comparison of effect size across different baseline total concentrations

Probiotic interventions did not improve LDLC concentrations in participants with desirable cholesterol levels at baseline [mean difference -0.06 (-0.15, 0.04) mmol/L].

Conversely, probiotic supplementation did improve total cholesterol concentrations in studies of participants with borderline-high and high total cholesterol at baseline; mean difference -0.21 (-0.33, -0.09) mmol/L, and mean difference -0.25 (-0.39, -0.11) mmol/L, respectively (Figure 25). (Analysis 7.3)

Effect sizes were homogenous in the studies of participants with desirable and high baseline total cholesterol (P>0.05).

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 37

Substantial heterogeneity was observed between the studies with borderline-high baseline total cholesterol; I2 = 69%, P = 0.006. To explore the source of heterogeneity, we dichotomised the studies of borderline-high baseline total cholesterol into 2 groups

(Analysis 7.4); ≤ median and > median baseline total cholesterol. There was a significant difference between effect estimates of the two groups (P = 0.003). Probiotic supplementation in studies in the ≤ median total cholesterol at baseline group did not improve LDLC concentration [mean difference -0.04 (-0.20, 0.11) mmol/L]. However, with only two studies included in this analysis, there is insufficient data to draw conclusions from these results. Conversely, based on data of 4 studies, probiotics were effective at improving LDLC in the studies in the > median group; mean difference

-0.40 (-0.58, -0.23) mmol/L (Figure 26).

9.12.3 High density lipoprotein cholesterol

9.12.3.1 Primary analysis

Probiotic bacteria had no effect on HDLC in a meta-analysis of 21 studies (Analysis

1.5); mean difference: -0.00 (-0.03, 0.02) mmol/L (Figure 27). A low level of heterogeneity was observed between results; I2 = 39%, P = 0.03.

9.12.3.2 Publication bias in primary analysis

The funnel plot of the primary HDLC analysis (Figure 28) was roughly symmetrical, with the exception of the study by Fuentes et al. (45). Excluding this study from analysis improved funnel plot symmetry (Figure 30) and homogeneity (P = 0.29).

However, this analysis did not alter the lack of effect of probiotics on HDLC; mean difference -0.02 (-0.04, 0.01) mmol/L (Analysis 1.6).

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 38

9.12.3.3 Sensitivity analyses

Restricting the analysis to published studies

Exclusion of the study by Ivey et al. (38) did not alter the lack of association; mean difference -0.00 (-0.02, 0.03) mmol/L. (Analysis 2.3)

Repeating analysis taking into account sample size

Probiotics had no effect on HDLC in studies with small [mean difference 0.01 (-0.02,

0.03) mmol/L] and large [-0.03 (-0.07, 0.01) mmol/L] sample size. (Analysis 4.3)

Repeating analysis using Random Effect model

When analysis was repeated using a Random effects model probiotic bacteria continued to have no effect on HDLC in a meta-analysis of 21 studies on increasing HDLC levels; mean difference: -0.01 (-0.04, 0.02) mmol/L. (Analysis 5.3)

9.12.3.4 Subgroup analyses

Comparison of effect size with probiotic bacteria from capsules and fermented milk

There was no difference in effect size (P > 0.05) of studies supplementing with probiotic bacteria from capsules [mean difference 0.02 (-0.02, 0.06) mmol/L] and fermented milk

[mean difference -0.01 (-0.04, 0.02) mmol/L]; P = 0.84. (Analysis 6.3)

Comparison of effect size across different baseline total concentrations

There was a significant difference in effect estimate between studies with different baseline total cholesterol levels; P = 0.001. Probiotic interventions did not affect HDLC concentrations in participants with desirable or high total cholesterol levels at baseline; mean difference 0.02 (-0.01, 0.06) mmol/L, and mean difference 0.04 (-0.01, 0.09)

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 39 mmol/L, respectively. Conversely, probiotic supplementation did lower HDLC concentration in participants with borderline-high total cholesterol at baseline; mean difference -0.06 (-0.10, -0.02) mmol/L. (Analysis 7.5)

9.12.4 Triglycerides

9.12.4.1 Primary analysis

Probiotic supplementation did not affect triglyceride concentrations in a meta-analysis of 23 studies (Analysis 1.7); mean difference: -0.00 (-0.05, 0.05) mmol/L (Figure 37).

The results of the all of the studies included in the meta-analysis were highly homogenous; I2 = 0%, P = 0.96.

9.12.4.2 Publication bias in primary analysis

The funnel plot of the primary HDLC analysis (Figure 38) was symmetrical, suggesting a low degree of publication bias.

9.12.4.3 Sensitivity analyses

Restricting the analysis to published studies

Exclusion of the study by Ivey et al. (38) had no effect on observed results; mean difference: -0.00 (-0.05, 0.05) mmol/L. (Analysis 2.4)

Repeating analysis taking into account sample size

Probiotics had no effect on triglyceride concentrations in studies with small [mean difference -0.01 (-0.06, 0.05) mmol/L] and large [mean difference 0.04 (-0.09, 0.17) mmol/L] sample size. (Analysis 4.4)

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 40

Repeating analysis using Random Effect model

Repeated analysis using a Random Effects model had no effect on observed results; mean difference: -0.00 (-0.05, 0.05) mmol/L. (Analysis 5.4)

9.12.4.4 Subgroup analyses

Comparison of effect size with probiotic bacteria from capsules and fermented milk

Supplementation with probiotic bacteria from both capsules [mean difference -0.04

(-0.18, 0.10) mmol/L] and fermented milk [mean difference 0.02 (-0.03, 0.07) mmol/L] did not improve triglyceride concentrations. (Analysis 6.4)

Comparison of effect size across different baseline total concentrations

Probiotic supplementation did not improve triglyceride concentrations in participants with desirable [mean difference -0.01 (-0.07, 0.04) mmol/L], borderline high [mean difference 0.03 (-0.08, 0.14) mmol/L] or high [mean difference 0.03 (-0.13, 0.19) mmol/L] total cholesterol at baseline. (Analysis 7.6)

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 41

9.13 DISCUSSION

9.13.1 Summary of main results

We found 23 randomised controlled clinical trials consisting of 16 parallel group studies and 7 crossover studies with a median intervention period of 6 weeks. A total of 1,310 patients participated in these studies. There was no outcome data on cardiovascular events, death from any cause, health-related quality of life and costs. When compared to placebo, probiotic supplementation did not increase risk of adverse events.

Compared with placebo, probiotic supplementation improved total cholesterol and

LDLC concentrations. This beneficial effect on total cholesterol and LDLC appeared to be limited to participants with higher total cholesterol at baseline. When compared to probiotic bacteria from fermented milk, probiotic bacteria from capsules were more effective at lowering total cholesterol and LDLC concentrations.

Compared with placebo, probiotics showed no statistically significant effects on HDLC or triglycerides, and the lack of effect was not ameliorated by considering baseline total-cholesterol levels or mode of administration.

9.13.2 Overall completeness and applicability of evidence

The results of this review are considered highly generalisable. This review considered randomised controlled trials including participants from different countries and backgrounds, with varying serum lipid profiles at baseline. The majority of studies reported the use of a double blind study design. All but one of the studies were conducted in community settings in both economically developed and underdeveloped countries, enhancing generalisability of results. For all lipid outcomes, there was

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 42 sufficient data to allow implementation of meta-analyses, and application of all sensitivity and subgroup analyses.

With the exception of the study by Hlivak et al. (40) which had an intervention period of

56 weeks, the remaining studies in this review had intervention periods of less than 12 weeks. This prevents determination of the long-term efficacy of probiotic supplementation.

9.13.3 Quality of the evidence

Methodological quality varied considerably, and information from published reports was often inadequate to allow determination of the risk of bias for many of the domains.

Despite this, there was considerable homogeneity of results, and the quality of evidence for each of the comparisons was rated as high.

There was inadequate reporting on adverse events in the included trials. However, in general, probiotics were safe for supplementation in the healthy adult population, despite some minor gastrointestinal complaints.

9.13.4 Potential biases in the review process

The review followed guidelines set out by The Cochrane Collaboration (32). Two authors independently read all the candidate studies, assessed them for inclusion, and rated them for risk of bias. We discussed any disagreements with a third review author with the aim of reaching a consensus, thus minimising potential bias. Although a thorough search strategy was implemented, we cannot rule out that some randomised controlled trials were missed.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 43

All but one of the studies were published, thus we cannot rule out the possibility of publication bias. However, exclusion of the unpublished study (38) and outlying studies identified through funnel plots did not alter the significance of the results. The low possibility of publication bias is further supported by the observation that when all studies were included in the review, there was a high degree of homogeneity of observed effects. The low level of heterogeneity in the LDLC primary analysis was removed once mode of administration or baseline total cholesterol concentrations were taken into account.

9.13.5 Authors' conclusions

9.13.5.1 Implications for practice

Based on high quality evidence, we found that probiotic supplementation lowered total cholesterol by 0.16 mmol/L and LDLC by 0.13 mmol/L, without eliciting adverse effects. Probiotics are more likely to have beneficial effects when supplemented in capsule form, or administered to patients with higher total cholesterol concentrations.

The effects of probiotics are likely to be of clinical significance, as small changes in total cholesterol and LDLC are associated with large reductions in cardiovascular disease risk and outcomes (62-64).

9.13.5.2 Implications for research

We do not know the importance of daily probiotic supplementation on CVD outcomes.

As such, appropriately powered long term randomised controlled trials are indicated to clarify the clinical significance of probiotic supplementation.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 44

9.14 TABLES

Chapter 9, Table 1: Characteristics of included studies

Study: Agerholm-Larsen 2000 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): 4/8/NS Abstinence from additional probiotic sources during intervention period: Abstain additional fermented milk products Intervention groups excluded from review: Placebo capsule intervention group (n=10). Participants Condition: overweight/obese Number of participants completing study (male:female): 17:43 Inclusion criteria: Healthy; normal blood pressure; body mass index 25 - < 37.5 kg/m2; males and females; 18 - 55 y old; no medications (except birth control pills); normal alcohol habits Exclusion criteria: No special diets; pregnant or breast feeding; elite athletes; diabetes; kidney or liver disease; postmenopausal. Interventions 450 mL FM/d Intervention(s): Intervention group G: E. faecum (1 strain, human species), S. thermophillus (2 strains). Minimum probiotic dose (log CFU/d) = 11.66 Intervention group StLa: S. thermophillus (2 strains), L. acidophilus (2 strains). Minimum probiotic dose (log CFU/d) = 10.73 Intervention group StLr: S. thermophillus (2 strains), L. rhamnosus (1 strain). Minimum probiotic dose (log CFU/d) = 11.65 Comparator(s): Fermented with organic acid Outcomes Outcomes presented in abstract: LDLC, fibrinogen, blood pressure Outcomes presented in text: Body weight, fat mass, waist: hip ratio, sagittal height, TC, HDLC, tgl, tissue plasminogen activator activity, coagulation factor VIIc, C reactive protein Study details Study start date: NS Study completion date: NS Publication Guarantor: Arne Astrup details Stated aim of 1) To investigate the effect of GAIO in obese subjects with a strict dietary compliance. study 2) To examine whether the beneficial effects on risk factors previously observed in the GAIO group can be attributed to the macro- or micronutrients shared by GAIO® and the controlled product. 3) To examine any beneficial effect on the CVD risk factor is specific for GAIO or may be achieved by other bacterial cultured strains with favourable in vitro properties. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 45

Study: Anderson 1999 Methods Type of study: clinical trial Allocation: randomised Intervention model: cross-over assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/washout/follow-up): NS/4/2/NS Abstinence from additional probiotic sources during intervention period: abstinence from sweet Acidophilus milk. Participants Condition: primary hypercholesterolaemia Number of participants completing study (male:female):18:22 Inclusion criteria: Lactose tolerant; not consumed FM in the past 2 weeks; complete blood count, blood chemistry panel, thyroid stimulating hormone, and urinalysis within the normal range. Exclusion criteria: Familial or secondary hypercholesterolaemia; diabetes mellitus; hypothyroidism; hypocholesterolaemic medication within the last 3 weeks; myocardial infarction, angioplasty or stroke within the last 3 months; treated with anticoagulants, immunosuppressants, corticosteroids or estrogens; thyroid replacement; pregnant or had plans to become pregnant during the time course of the studies; lactating; recent history of alcohol abuse. Interventions 200 g FM/d Intervention(s): Starter: S. thermophillus MUH34. Probiotic: L. acidophilus L1. Minimum probiotic dose (log CFU/d) = 7.00 Comparator(s): Starter: S. thermophillus MUH34 Outcomes Outcomes presented in abstract: TC Outcomes presented in text: LDLC, HDLC, tgl, body weight Study details Study start date: NS Study completion date: NS Publication Final author: Stanley E Gilliand details Stated aim of To examine effects of consumption of 1 daily serving of fermented milk (yogurt) on serum lipids. study Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 46

Study: Ataie-Jafari 2009 Methods Type of study: clinical trial Allocation: randomised Intervention model: cross-over assignment Masking: single blind Primary purpose: prevention Duration in weeks (lead-in/intervention/washout/follow-up): 2/6/4/NS Abstinence from additional probiotic sources during intervention period: Abstain from yogurt and doogh. Participants Condition: mild to moderate hypercholesterolaemia Number of participants completing study (male:female): 4:10 Inclusion criteria: Total cholesterol between 5.17 and 7.76 mmol/L. Exclusion criteria: Coronary heart disease; diabetes; hypothyroidism; nephritic syndrome; obesity; using lipid-lowering drugs or pharmaceuticals known to affect the blood lipid metabolism. Interventions 300 g FM/d Intervention(s): Starter: S. thermophillus and L. delbrueckii subsp bulgaricus Probiotic: L. acidophilus and B. lactis. Minimum probiotic dose (log CFU/d) = 8.78 Comparator(s): Starter: S. thermophillus and L. delbrueckii subsp bulgaricus Outcomes Outcomes presented in abstract: TC Outcomes presented in text: LDLC, HDLC, tgl, LDLC:HDLC ratio Study details Study start date: NS Study completion date: NS Publication Final Author: Farideh Tahbaz details Stated aim of To Compare the effect of consuming of probiotic yogurt with that of ordinary yogurt on serum study cholesterol level in mildly to moderately hypercholesterolaemic subjects. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 47

Study: de-Roos 1999 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/washout/follow-up): 2/6/NS/NS Abstinence from additional probiotic sources during intervention period: Participants Condition: nil Number of participants completing study (male:female): 22:56 Inclusion criteria: healthy; between 18 and 65 y old. Exclusion criteria: Heart disease; diabetes; liver or kidney disease; use of medications known to affect blood lipid metabolism; TC> 8 mmol/L; tgl > 4 mmol/L Interventions 500 mL FM/d Intervention(s): Starter: S. thermophillus Probiotic: L. acidophilus L-1. Minimum probiotic dose (log CFU/d) = 9.68 Comparator(s): Starter: S. thermophillus Outcomes Outcomes presented in abstract: TC, LDLC, HDLC, tgl Outcomes presented in text: body weight Study details Study start date: NS Study completion date: NS Publication Final Author: MB Katan details Stated aim of To investigate whether intake of Lactobacillus acidophilus strain L-1 lowers serum cholesterol in study healthy men and women. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 48

Study: Ejtahed 2011 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/washout/follow-up): 1/6/NS/NS Abstinence from additional probiotic sources during intervention period: abstain from yogurt Participants Condition: Type 2 diabetes mellitus Number of participants included in analysis (male:female): 23:37 Inclusion criteria: Type 2 diabetes mellitus for at least 1 year; 30-60 y old; LDLC >= 2.6 mmol/L Exclusion criteria: Smoking; liver, inflammatory intestinal or kidney disease; use of cholesterol lowering medications; thyroid disorders; immunodeficiency disease; lactose intolerance; requiring insulin injections; taking nutritional supplements during the preceding 3 weeks; use of oestrogen; progesterone or diuretics; pregnant; breastfeeding; consuming probiotic products in the preceding 2 months; body mass index > 35 kg/m2. Interventions 300 g FM/d Intervention(s): Starter: S. thermophillus and L. bulgaricus Probiotic: L. acidophilus La5 and B. lactis Bb12. Minimum probiotic dose (log CFU/d) = 9.60 Comparator(s): Starter: S. thermophillus and L. bulgaricus Outcomes Outcomes presented in abstract: TC; LDLC, HDLC, tgl, TC:HDLD ratio, LDLC:HDLC ratio Study details Study start date: NS Study completion date: NS Publication Final author: MB Katan details Stated aim of To investigate the effects of probiotic and conventional yogurt on the lipid profile in type 2 study diabetic people. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 49

Study: Fabian 2006 Methods Type of study: clinical trial Allocation: randomised Intervention model: cross over assignment Masking: NS Primary purpose: prevention Duration in weeks (lead-in/intervention/washout/follow-up): 1/2/2/2 Abstinence from additional probiotic sources during intervention period: abstain from fermented milk Participants Condition: nil Number of participants completing study (male:female): 0:32 Inclusion criteria: healthy; normocholesterolaemic; female; aged between 22 and 29 Exclusion criteria: smoking Interventions 100 g FM/d Intervention(s): Starter: S. thermophillus and L. bulgaricus Probiotic: L. paracasei subsp paracasei. Minimum probiotic dose (log CFU/d) = 8.56 Comparator(s): Starter: S. thermophillus and L. bulgaricus Outcomes Outcomes presented in abstract: TC, TC:HDLC ratio, LDLC, HDLC, HDLC:LDLC ratio Outcomes presented in text: tgl Study details Study start date: NS Study completion date: NS Publication Final Author: Ibrahim Elmadfa details Stated aim of To verify and compare the effects of probiotic and conventional yogurt on the plasma lipid study profile of normocholesterolaemic women. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 50

Study: Fuentes 2013 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: double blinded Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): NS/12/4 Abstinence from additional probiotic sources during intervention period: Participants Condition: hypercholesterolaemia Number of participants completing study: 60 (male:female ratio not specified) Inclusion criteria: TC between 5.16-7.64 mmol/L; body mass index between 19-30 kg/m2; LDLC between 3.35 - 4.91 mmol/L male or female; aged between 18-65 years Exclusion criteria: tgl >= 3.85 mmol/L; cardiovascular event in the last 6 months; secondary dyslipidaemia related to thyroid dysfunction; use of drugs affecting lipid metabolism; pregnant; receiving a specific diet and institutionalised. Interventions capsule Intervention(s): L. plantarium (three strains). Minimum probiotic dose (log CFU/d) = 9.08 Comparator(s): Placebo capsule Outcomes Outcomes presented in abstract: TC, LDLC, oxidised LDLC Outcomes presented in text: tgl, HDLC, HDLC:LDLC ratio Study details Study start date: NS Study completion date: NS Publication Final Author: Jordi Cune details Stated aim of To investigate the effects of a mixture of three strains of Lactobacillus plantarium (CECT 7527, study CECT 7528 and CECT 7529) on cholesterol-lowering efficacy in hypercholesterolaemic patients. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 51

Study: Hata 1996 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: NS Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): 4/8/NS Abstinence from additional probiotic sources during intervention period: NS Participants Condition: hypertension Number of participants completing study (male:female): 8:22 Inclusion criteria: hypertension; 40-86 y old. Exclusion criteria: Secondary causes of hypertension; lactose intolerance; cow milk allergy; alcohol abuse; serious disease. Interventions 95 mL FM/d Intervention(s): L. helveticus and S. cerivasea. Minimum probiotic dose (log CFU/d) = 11.85 Comparator(s): Artificially acidified Outcomes Outcomes presented in abstract: blood pressure, pulse rate, body weight, TC, LDLC, HDLC, tgl Outcomes presented in text: tgl, glucose, potassium, sodium, chloride, blood urea nitrogen, uric acid, creatinine, total protein, glutamic oxaloacetic transaminase, glutamic pyruvic transaminase, y-glutamyl transpeptidase, lactate dehydrogenase, haemoglobin, red blood cell, platelet count, TC:HDLC ratio, HDLC:LDLC ratio Study details Study start date: NS Study completion date: NS Publication Final Author: Toshiaki Takano details Stated aim of NS study Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 52

Study: Hlivak 2005 Methods Type of study: clinical trial Allocation: randomised Intervention model: cross over assignment Masking: un-blinded Primary purpose: prevention Duration in weeks (lead-in/intervention/washout/follow-up): NS/56/NS/4 Abstinence from additional probiotic sources during intervention period: NS Participants Condition: elderly nursing home patients Number of participants completing study (male:female): 7:31 Inclusion criteria: NS Exclusion criteria: NS Interventions capsules Intervention(s): Enterococcus faecum M-74. Minimum probiotic dose (log CFU/d) = 9.30. Plus 50 ug organically bound selenium Comparator(s): placebo capsule Outcomes Outcomes presented in abstract: TC, LDLC, HDLC, tgl Study details Study start date: April 2001 Study completion date: May 2002 Publication Final Author: Z Mikes details Stated aim of To investigate the impact of long-term orally administered probiotic strain Enterococcus faecum study M-74 enriched with selenium on lipid profile (total cholesterol, LDH, HDL, and triglycerides) in humans. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 53

Study: Inoue 2003 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: single blind Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): NS/12/2 Abstinence from additional probiotic sources during intervention period: NS Participants Condition: mild hypertension Number of participants included in an analysis (male:female): 20:15 Inclusion criteria: WHO/ISH hypertension grade 1-2/level -II); > 20 y old Exclusion criteria: History of antihypertensive medication use; illness; pregnancy. Interventions 100 mL FM/d Intervention(s): L. casei shirota and Lactococcus lactis YIT 2027. Minimum probiotic dose (log CFU/d) = NS Comparator(s): acidified with L-lactic acid Outcomes Outcomes presented in abstract: BP, heart rate, body weight, glucosuria and proteinuria. Outcomes presented in text: red blood cell count, white blood cell count, platelets, haemoglobin, haematocrit, aspartate aminotransferase, alanine aminotransferase, blood urea, creatinine, total protein, blood glucose, uric acid, sodium, potassium, chloride, TC and tgl Study details Study start date: NS Study completion date: NS Publication Final Author: K Inoue details Stated aim of To study the effect of a new fermented milk product containing GABA (FMG) on the blood study pressure (BP) of patients with mild hypertension. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 54

Study: Ivey 2013 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/washout): 3/6/0 Abstinence from additional probiotic sources during intervention period: Abstain from probiotic containing products Participants Condition: overweight / obese and elevated waist circumference Number of participants completing study (male:female): 96:60 Inclusion criteria: consuming less than 400 g yogurt per week; not taking probiotic supplements; body mass index ≥ 25 kg/m2; waist circumference ≥ 94 cm in men and ≥ 80cm in women; office blood pressure ≥ 120/80 mmHg. Exclusion criteria: Inability to complete the study; intolerance to dairy foods; use of antibiotics, immunosuppressive treatments or hypoglycaemic treatments. Interventions Intervention(s): Intervention group 1: 200 mL FM/d PLUS capsules. Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12. Minimum probiotic dose (CFU/d) = 6x109 Intervention group 2: 200 mL FM/d. Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12. Minimum probiotic dose (CFU/d) = 3x109 Intervention group 3: Capsules. Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12. Minimum probiotic dose (CFU/d) = 3x109 Comparator(s): milk PLUS placebo capsules Outcomes Primary outcomes: LDLC, TC, HDLC, tgl, markers of cholesterol metabolism, blood pressure, glucose, insulin, glycated haemoglobin, homeostatic model of insulin resistance Secondary outcomes: Biochemical markers of inflammation, gastrointestinal health, quality of life and biomarkers of bone metabolism. Study details Study start date: February 2012 Study completion date: February 2013 Publication Responsible party/principal investigator: Richard Prince details Stated aim of 1) To examine the effects of yogurt and its probiotics on features of the metabolic syndrome, study intestinal health, bone metabolism and quality of life. 2) To determine dietary, lifestyle and health status determinants of baseline faecal microflora profile. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 55

Study: Jauhiainen 2005 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): 4/10/4 Abstinence from additional probiotic sources during intervention period: NS Participants Condition: hypertension Number of participants completing study (male:female): 34:60 Inclusion criteria: systolic blood pressure between 140-180 mm Hg; diastolic blood pressure between 90-110 mm Hg Exclusion criteria: Blood pressure lowering medication; unstable coronary artery disease; diabetes mellitus; malignant diseases; alcohol abuse; milk allergy; pregnancy Interventions 300 mL FM/d Intervention(s): L. helveticus. Minimum probiotic dose (log CFU/d) = NS Comparator(s): Normal mesophilic Lactococcus sp. mixed culture. NS Outcomes Outcomes presented in abstract: Blood pressure Outcomes presented in text: C-reactive protein, LDLC, tgl, TC : HDLC ratio, body weight, angiotensin converting enzyme activity. Study details Study start date: NS Study completion date: NS Publication Final Author: Riitta Korpela details Stated aim of To evaluate the BP-lowering effect and possible adverse events of Lactobacillus helveticus study LBK-16H fermented milk with a high tripeptide concentration on hypertensive subjects by using 24-h ABPM. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 56

Study: Jones 2012 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): 2/6/2 Abstinence from additional probiotic sources during intervention period: Participants Condition: hypercholesterolaemia Number of participants completing study (male:female): 40:74 Inclusion criteria: otherwise healthy male and female; aged between 18-74 years; LDLC > 3.4mmol/L with < 15% variation between successive screening visits; tgl < 4.0 mmol/L, BMI between 22-32 kg/m2; able to understand dietary procedures; motivated. Exclusion criteria: Use of statin or other cholesterol lowering prescription drugs within the last 6 months; use of plant sterols, n-3 fatty acids, fish oil, soya protein, soluble oat fibre, psyllium seed husk or other cholesterol lowering supplements within the last 3 months; history of chronic alcohol use > 2 drinks/d; use of systemic antibodies, corticosteroids, androgens or phytoin; myocardial infarction, coronary artery bypass or other surgical procedures within the last 6 months; lactose intolerance or allergies to dairy products; history of angina, congestive heart failure, inflammatory bowel disease, hepatic or biliary disease or cancer; chronic use of probiotics, fibre laxatives, or stimulant laxatives; history of eating disorders; energy expenditure > 16736 kJ/week; pregnancy, breastfeeding or intent to get pregnant. Interventions 400 mL FM/d Intervention(s): Starter culture: yogurt bacteria Probiotic bacteria: L. reuteri NCIMB 30242. Minimum probiotic dose (log CFU/d) = 10.45 Comparator(s): Starter culture: yogurt bacteria Outcomes Primary outcome: percent change from baseline in LDLC Outcomes presented in abstract: LDLC, TC, non-HDLC, tgl, HDLC, apoB-100 Outcomes presented in text: LDLC:HDLC ratio, faecal deconjugated bile acids, complete blood cell count, platelet, haematocrit, haemoglobin, urea, creatinine, alanine transaminase, aspartate aminotransferase, y-glutamyl transpeptidase, alkaline phosphatase, bilirubin, lipase, phosphate, potassium, sodium, chloride and hydrogen carbonate Study details Study start date: September 2008 Study completion date: April 2009 Publication Responsible Party: Micropharma Limited details Stated aim of Clinicaltrials.gov identifier: NCT01185795 study To evaluate the lipid lowering efficacy of a yogurt formulation containing microencapsulated Bile Salt Hydrolase (BSH)-active Lactobacillus reuteri cardioviva, taken twice per day over 6 weeks, in subjects with hypercholesterolaemia. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 57

Study: Jones 2012a Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): 4/9/NS Abstinence from additional probiotic sources during intervention period: NS Participants Condition: hypercholesterolaemia Number of participants completing study (male:female): 55:72 Inclusion criteria: Males and females, aged 20 to 75 years; LDLC > 3.4 mmol/L; tgl < 4.0 mmol/L; BMI between 22 and 32 kg/m²; Subject understands and accepts to follow the dietary recommendations advisable for hypercholesterolaemic patients (according to NCEP-ATP III guidelines); For subjects on statin monotherapy: dosage of statin must be stable for at least 3 months prior to the study beginning; compliant and motivated; Signed informed consent form prior to inclusion in the study; stable doses of thyroid hormone and anti-hypertensive agents; for female subjects: effective contraceptive methods used Exclusion criteria: Use of cholesterol lowering prescription drugs other than statin monotherapy within the last 6 months; use of plant sterols, omega 3, fish oil, soy protein, soluble oat fibre, psyllium seed husk, or other cholesterol lowering non-prescription food supplements within last 3 months; history of chronic use of alcohol (>2 drinks/d); use of systemic antibodies, corticosteroids, androgens, or phenytoin; subject having experienced any cardiovascular event (Myocardial infarction, coronary artery bypass, or other major surgical procedures) in the last 6 months; diabetes; receiving systemic treatment or topical treatment likely to interfere with evaluation of the study parameters; currently involved in a clinical trial or in an exclusion period following participation in another clinical trial; history of angina, congestive heart failure, inflammatory bowel disease, pancreatitis, gastrointestinal, renal, pulmonary, hepatic or biliary disease, or cancer (evidence of active lesions, chemotherapy or surgery in the past year); chronic user of probiotics or fibre laxative (greater than 2 doses/wk), or stimulant laxatives; history of eating disorders; exercise greater than 15 miles/wk or 4,000 kcal/wk; pregnancy, breast feeding, or intent to get pregnant Interventions Capsules Intervention(s): Probiotic bacteria: L. reuteri NCIMB 30242 (Cardioviva). Minimum probiotic dose (log CFU/d) = 9.60 Comparator(s): No probiotic bacteria Outcomes Outcomes presented in abstract: LDLC, TC, non-HDLC, apoB-100, LDLC:HDLC ratio, apoB-100:apoA-1 ratio, tgl, HDLC, High sensitivity C reactive protein, fibrinogen, plasma deconjugated bile acids, campesterol, sitosterol and stigmasterol Outcomes presented in text:: plant sterol, campesterol : TC ratio, sitosterol : TC ratio, stigmasterol : TC ratio, and plant sterol : TC ratio. Study details Study start date: NS Study completion date: NS Publication Final Author: Staya Prakash details Stated aim of FDA/OHRP IORG registration number: IORG0000612 study To evaluate the cholesterol-lowering efficacy and mechanism of action of bile salt hydrolase-active Lactobacillus reuteri NCIMB 30242 capsules in hypercholesterolaemic adults. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 58

Study: Kekkonen 2008 Methods Type of study: clinical trial Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/washout): 3/3/NS Abstinence from additional probiotic sources during intervention period: Abstain from probiotic-containing product Participants Condition: nil Number of participants completing study (male:female): 12:14 Inclusion criteria: regular exercise ≥ 3 times per week. Exclusion criteria: Participating in another clinical trial; milk allergy; chronic illness; antibiotic use during 2 months preceding the trial; gastrointestinal diseases and related medication; pregnancy or lactation. Interventions 250 mL milk based fruit drink Intervention(s): Probiotic bacteria: L. rhamnosus GG ATCC 53103. Minimum probiotic dose (log CFU/d) = 10.19 Comparator(s): NS (Similar placebo drink without probiotic bacteria) Outcomes Outcomes presented in abstract: lysophosphatidylcholines, sphingomyelins, glycerophosphatidylcholines, interleukin-6, tgl, tumour necrosis factor a, C-reactive protein Outcomes presented in text: TC; LDLC; HDLC, tgl Study details Study start date: NS Study completion date: NS Publication Final Author: Heikki Vapaatalo details Stated aim of To investigate effect of three weeks' intervention with the probiotic Lactobacillus rhamnosus GG study (LGG) bacteria on global serum lipidomic profiles and evaluate whether the changes in inflammatory variables (CRP, TNF- a and IL-6) a reflective in the global lipidomic profiles of healthy adults. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable 4 intervention groups (1 placebo, 3 probiotic intervention). Only data from 2 (placebo and 1 probiotic) groups was published.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 59

Study: Mizushima 2004 Methods Type of study: clinical study Allocation: randomised Intervention model: parallel assignment Masking: NS Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): NS/4/NS Abstinence from additional probiotic sources during intervention period: NS Participants Condition: hypertension Number of participants completing study (male:female): 42:0 Inclusion criteria: systolic blood pressure ≥ 130 mm Hg; diastolic blood pressure ≥ 85 mm Hg Exclusion criteria: history of antihypertensive medication use; history of lactose intolerance; cow milk allergy. Interventions 160 g FM/d Intervention(s): L. helveticus and S. cerevisiae. Minimum probiotic dose (log CFU/d) = NS Comparator(s): Artificially acidified Outcomes Outcomes presented in abstract: blood pressure Outcomes presented in text: angiotensin I, angiotensin II, angiotensin I:angiotensin II ratio, TC, HDLC, tgl Study details Study start date: NS Study completion date: NS Publication Final Author: Hirotsugu Ueshima details Stated aim of To assess the effect of sour milk, containing two tripeptides (valine-proline-proline and study isoleucine-proline-proline), on blood pressure (BP). Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 60

Study: Rizkalla 2000 Methods Type of study: clinical study Allocation: randomised Intervention model: cross-over assignment Masking:NS Primary purpose: prevention Duration in weeks (lead-in/intervention/washout/follow-up): NS/2.1/2.1/NS Abstinence from additional probiotic sources during intervention period: NS Participants Condition: nil Number of participants completing study (male:female): 24:0 Inclusion criteria: Healthy; male; aged 20-60 y old. Exclusion criteria: nil Interventions 500 g FM/d Intervention(s): L. bulgaricus and S. thermophillus. Minimum probiotic dose (log CFU/d) = 10 Comparator(s): Pasteurised probiotic FM Outcomes Outcomes presented in abstract: Plasma glucose area under the curve, insulin area under the curve, fatty acid area under the curve, fasting glucose, insulin, fatty acid, tgl, TC, butyrate, propionate, acetate, breath hydrogen Outcomes presented in text: HDLC, short chain fatty acids Study details Study start date: NS Study completion date: NS Publication Final Author: Gerard Slama details Stated aim of To compare the effects of chronic consumption of yogurt with (fresh) or without (heated) live study bacterial cultures (Lactobacillus bulgaricus and Streptococcus thermophillus) on plasma glucose, insulin, triacylglycerols, cholesterol, fatty acids, and short-chain fatty acids. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 61

Study: Sadrzadeh-Yeganeh 2010 Methods Type of study: clinical study Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): 1/6/NS Abstinence from additional probiotic sources during intervention period: avoid yogurt, probiotic and fermented products Intervention groups excluded from review: Control group whom did not consume any fermented and probiotic products for the duration of the study (n=29). Participants Condition: nil Number of participants completing study (male:female): 0:59 Inclusion criteria: Female; aged 19-49 y old; TC < 6.2 mmol/L; tgl < 2.3 mmol/L; body mass index < 30 kg/m2 Exclusion criteria: Smoking; kidney, liver or inflammatory intestinal diseases; thyroid disorders; diabetes; immunodeficiency diseases; lactose intolerance; supplement use; medication use; consuming probiotic yogurt or any probiotic-containing products in the preceding 2 months; elite athlete; pregnant; breastfeeding. Interventions 300 g FM/d Intervention(s): Starter: L. bulgaricus and S. thermophiles Probiotic: L. acidophilus La5 and B. lactis Bb12. Minimum probiotic dose (log CFU/d) = 7.89 Comparator(s): Starter: L. bulgaricus and S. thermophiles Outcomes Outcomes presented in abstract: tgl, TC, HDLC, LDLC, TC:HDLC ratio Study details Study start date: NS Study completion date: NS Publication Final Author: Maryam Chamary details Stated aim of To investigate the effect of probiotic yogurt containing Lactobacillus acidophilus La5 and study Bifidobacterium lactis Bb12 and conventional yogurt on the lipid profile in women. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 62

Study: Simons 2006 Methods Type of study: clinical study Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): 1/10/NS Abstinence from additional probiotic sources during intervention period: NS Participants Condition: nil Number of participants completing study (male:female): 16:28 Inclusion criteria: Aged 30-75 y old; TC ≥ 4.0 mmol/L; tgl ≤ 4.0 mmol/L Exclusion criteria: Using lipid modifying drugs; cardiac, renal or hepatic disease; diabetes; adverse reaction to lactose or milk products; using probiotic supplements Interventions Capsules Intervention(s): L. fermentum PCC. Minimum probiotic dose (log CFU/d) = 9.60 Comparator(s): Placebo capsule Outcomes Outcomes presented in abstract: tgl, TC, HDLC, LDLC, liver enzymes Outcomes presented in text: alanine aminotransferase, aspartate aminotransferase, gamma-glutyl transpeptidase, C reactive protein, creatinine, glucose Study details Study start date: September 2004 Study completion date: January 2005 Publication Final Author: Patricia Conway details Stated aim of To evaluate the effects of PCC® Lactobacillus fermentum in capsule form on LDL cholesterol study levels in hypercholesterolaemic subjects consuming their personally chosen diets. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 63

Study: Songisepp 2012: study 1 Methods Type of study: clinical study Allocation: randomised Intervention model: cross-over assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/washout): 4/3/2/NS Abstinence from additional probiotic sources during intervention period: avoid probiotic products Participants Condition: healthy Number of participants completing study (male:female): 5:8 Inclusion criteria: Desire to participate; no known health problems; 18-65 years Exclusion criteria: Antimicrobial treatment within the preceding 2 months; special dietary routine; unstable cardiopulmonary system; history of diabetes or malignancy; food allergy; acute infection; chronic renal or hepatic failure; gut surgery; acute illness within the preceding 4 weeks; medication use within the preceding 2 months; history of alcohol abuse; pregnant; breastfeeding; undiagnosed diabetes. Interventions 50g FM/d Intervention(s): Starter: C92 cheese starter Probiotic: L. fermentum Tensia. Minimum probiotic dose (log CFU/d) = 10.4 Comparator(s): Starter: C92 cheese starter Outcomes Primary Outcomes: To assess the safety of the novel probiotic Lactobacillus plantarum with antimicrobial properties and the strain containing cheese on healthy subjects. The survival of the probiotic strain in Gastrointestinal Tract (GIT) and its effect on faecal lactoflora, measured on 25/04/07, 16/05/07, 30/05/07 and 20/06/07. Secondary Outcomes: 1) To assess the health indices of healthy adults (body mass index, blood pressure), measured on 25/04/07, 16/05/07, 30/05/07 and 20/06/07; 2) The self-reported questionnaire was applied containing questions on welfare, nutritional habits, and habitual gastrointestinal symptoms (abdominal pain, flatulence, bloating, and stool frequency), measured once a week during the trial; 3) To determine haematological indices (haemoglobin, white blood cell count, red blood cells, platelets), plasma glucose, albumin, ferritin, Total Cholesterol (TC), Low-Density Lipoprotein cholesterol (LDL), High-Density Lipoprotein cholesterol (HDL), triglyceride and high-sensitive C-Reactive Protein (hsCRP), Interleukin 6 (IL-6), Immunoglobulins (IgA, IgM, IgG) levels, measured on 25/04/07, 16/05/07, 30/05/07 and 20/06/07; 4) To determine in urine the content of biogenic amines; 5) Faecal samples were analysed for the changes in the counts of clostridia (including C. difficile), total anaerobes, enterococci, E. coli and lactic acid bacteria, collected also at 25/04/07, 16/05/07, 30/05/07 and 20/06/07 and stored at -20°C. Faecal samples are analysed step-by-step during one year; 6) Denaturated Gradient Gel Electrophoresis (DGGE) was used to monitor changes in total faecal microflora after cheese consumption, analysis from faecal samples is performed in September 2007 Study details Study start – completion date: April – June 2007 Publication Final Author: Prof Marika Mikelsaar details Responsible party: Healthy Dairy Products Ltd (Estonia) - Bio-Competence Centre Stated aim of International Standard Randomized Controlled Trial number: ISRCTN38739209 study Hypothesis: The consumption of probiotic Lactobacillus plantarum-containing cheese has positive impact on intestinal microflora of healthy volunteers. Primary outcome: To assess the safety of cheese comprising the probiotic Lactobacillus plantarum strain in healthy elderly subjects. The survival of the probiotic strain in gastrointestinal tract (GIT) and its effect on intestinal microflora and clinical blood indices of healthy volunteers will be measured at the recruitment, after a 3-week of probiotic treatment, after a 2-week wash-out and after a 3-week placebo treatment. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 64

Study: Songisepp 2012: study 2 Methods Type of study: clinical study Allocation: randomised Intervention model: cross-over assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/washout/follow-up): 4/3/2/NS Abstinence from additional probiotic sources during intervention period: avoid probiotics Participants Condition: healthy Number of participants completing study (male:female): 1:17 Inclusion criteria: Desire to participate; no known health problems; > 65 years Exclusion criteria: Antimicrobial treatment within the preceding 2 months; special dietary routine; unstable cardiopulmonary system; history of diabetes or malignancy; food allergy; acute infection; chronic renal or hepatic failure; gut surgery; acute illness within the preceding 4 weeks; medication use within the preceding 2 months; history of alcohol abuse; pregnant; breastfeeding; undiagnosed diabetes. Interventions 50g FM/d Intervention(s): Starter: C92 cheese starter Probiotic: L. fermentum Tensia. Minimum probiotic dose (log CFU/d) = 8.17 Comparator(s): Starter: C92 cheese starter Outcomes Primary Outcomes: To assess the safety of cheese comprising the probiotic Lactobacillus plantarum strain in healthy elderly subjects. The survival of the probiotic strain in gastrointestinal tract (GIT) and its effect on intestinal microflora and clinical blood indices of healthy volunteers will be measured at the recruitment, after a 3-week of probiotic treatment, after a 2-week wash-out and after a 3-week placebo treatment. Secondary Outcomes: 1) Blood, urine and faecal samples are collected at the recruitment, after 3 weeks of probiotic treatment, after a 2-week wash-out period and after a 3-week placebo treatment; 2) The assessment of the health indices of healthy elderly (body mass index, blood pressure), measured at the recruitment, after 3 weeks of probiotic treatment, after a 2-week wash-out period and after a 3-week placebo treatment; 3) The self-reported questionnaire is applied containing questions on welfare, nutritional habits, and habitual gastrointestinal symptoms (abdominal pain, flatulence, bloating, and stool frequency), measured once a week during the trial; 4) To determine haematological indices (haemoglobin, white blood cell count, red blood cells, platelets), plasma glucose, albumin, total cholesterol (TC), low-density lipoprotein cholesterol (LDL), high-density lipoprotein cholesterol (HDL), triglyceride and high-sensitive C-reactive protein (hs-CRP), interleukin 6 (IL-6), immunoglobulins (IgA, IgM, IgG) levels will be measured at the recruitment, after 3 weeks of probiotic treatment, after a 2-week wash-out and after a 3-week placebo treatment; 5) Biogenic amines from urine samples will be measured at the recruitment, after 3 weeks of probiotic treatment, after a 2-week wash-out period and after a 3-week placebo treatment; 6) Faecal samples will be analysed for the changes in the counts of clostridia (including C. difficile), total anaerobes, enterococci, E. coli and lactic acid bacteria Study details Study start – completion date: January – April 2009 Publication Final Author: Prof Marika Mikelsaar details Stated aim of International Standard Randomized Controlled Trial number: ISRCTN45791894 study Hypothesis: The consumption of probiotic Lactobacillus plantarum-containing cheese has positive impact on intestinal microbiota and blood indices of healthy volunteers. Primary Outcomes: To assess the safety of cheese comprising the probiotic Lactobacillus plantarum strain in healthy elderly subjects. The survival of the probiotic strain in gastrointestinal tract (GIT) and its effect on intestinal microflora and clinical blood indices of healthy volunteers will be measured at the recruitment, after a 3-week of probiotic treatment, after a 2-week wash-out and after a 3-week placebo treatment. Notes FM/d: fermented milk per day; CFU/d: colony forming units / day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 65

Study: Usinger 2010 Methods Type of study: clinical study Allocation: randomised Intervention model: parallel assignment Masking: double blind Primary purpose: prevention Duration in weeks (lead-in/intervention/follow-up): NS/8/NS Abstinence from additional probiotic sources during intervention period: NS Participants Condition: Prehypertension or borderline hypertension Number of participants completing study (male:female): 50:40 Inclusion criteria: Systolic blood pressure between 130-149 mmHg or diastolic blood pressure between 85-95 mmHg Exclusion criteria: Cardiac or renal disease; diabetes; antihypertensive treatment; pregnancy; milk allergy Interventions FM Intervention(s): Intervention group 300 mL/d FM: L. helveticus Cardio-04. Minimum probiotic dose (log CFU/d) = NS Intervention group 150 mL/d FM: L. helveticus Cardio-04. Minimum probiotic dose (log CFU/d) = NS Comparator(s): Comparator group 300 mL/d FM: Artificially acidified Comparator group 150 mL/d FM: Artificially acidified Outcomes Primary Outcome: repeated 24 hour ambulatory measurements Outcomes presented in abstract: 24 hour ambulatory and office blood pressure, and heart rate Outcomes presented in text: TC, HDL, VLDL, LDL, tgl, creatine, potassium, sodium Study details Study start date: 2007 Study completion date: 2008 Publication Final Author: H Ibsen details Stated aim of 1) To investigate whether milk fermented with Lactobacillus helveticus; Cardi04 containing VPP study and IPP, reduces 24hBP in individuals with prehypertension and borderline hypertension. 2) To investigate the effect of FM on the office BP, heart rate and plasma concentrations of lipids. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 66

Study: Xiao 2003 Methods Type of study: clinical study Allocation: randomised Intervention model: parallel assignment Masking: Single blinded Primary purpose: prevention Duration in weeks (lead-in/intervention/follow up): 2/4/NS Abstinence from additional probiotic sources during intervention period: restricted consumption of fermented milk and pickles Participants Condition: nil Number of participants completing study (male:female): 32:0 Inclusion criteria: Healthy; adult; male; deemed suitable for participation on the basis of prior physical examination, serum cholesterol level and medical history Interventions 300 mL FM/d Intervention(s): Starter: S. thermophillus and L. delbrueckii subsp bulgaricus. Probiotic: B. longum BL1. Minimum probiotic dose (log CFU/d) = 8.61 Comparator(s): Starter: S. thermophillus and L. delbrueckii subsp bulgaricus. Outcomes Outcomes presented in abstract: TC Outcomes presented in text: HDL, LDL, tgl, glucose Study details Study start date: NS Study completion date: NS Publication Final Author: A Hiramatsu details Stated aim of To demonstrate the effect of milk fermented by B. longum strain BL1 on blood lipids in rats and study humans. Notes FM/d: fermented milk per day; CFU/d: colony forming units per day; TC: total cholesterol; LDLC: low density lipoprotein cholesterol; tgl: triglyceride; NS: not specified; NA: not applicable

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 67

Chapter 9, Table 2: Characteristics of excluded studies

Study identifier Reason for exclusion Al-Sheraji 2012 (65) Participants did not meet inclusion criteria Allen 2012 (66) Participants did not meet inclusion criteria Alonso 2009 (67) Control did not meet inclusion criteria Asemi 2012 (68) Participants did not meet inclusion criteria Ashar 2000 (69) Control did not meet inclusion criteria Chang 2011 (70) Intervention did not meet inclusion criteria Charlton 2008 (71) Intervention did not meet inclusion criteria De Leeuw 2009 (72) Control did not meet inclusion criteria Diop 2008 (73) Review outcomes not reported Enck 2008 (74) Review outcomes not reported Engberink 2008 (75) Intervention did not meet inclusion criteria Gemeir 2013 (76) Control did not meet inclusion criteria Hoo-Kil 2000 (77) Review outcomes not reported in humans Intorre 2011 (78) Control did not meet inclusion criteria Ishikawa 2002 (79) Control did not meet inclusion criteria Jayakumar 2012 (80) Participants did not meet inclusion criteria Jones 2012b (81) Review outcomes not reported Kajander 2008 (82) Review outcomes not reported Karlsson 2010 (83) Intervention did not meet inclusion criteria Kim 2011 (84) Intervention did not meet inclusion criteria Leber 2012 (85) Control did not meet inclusion criteria Lee 2007 (86) Intervention did not meet inclusion criteria Lewis 2005 (87) Outcomes did not meet inclusion criteria Linderborg 2012 (88) Intervention did not meet inclusion criteria Malaguarnera 2012 (89) Intervention did not meet inclusion criteria Marotta 2012 (90) Intervention did not meet inclusion criteria Massey 1984 (91) Control did not meet inclusion criteria Nestel 2012 (92) Intervention did not meet inclusion criteria Neyestani 2013 (93) Control did not meet inclusion criteria Pereg 2011 (94) Review outcomes not reported Pitnus 2013 (95) Intervention did not meet inclusion criteria Ranganathan 2009 (96) Control did not meet inclusion criteria Rossouw 1981 (97) Participants did not meet inclusion criteria Selinger 2013 (98) Review outcomes not reported Seppo 2003 (99) Intervention did not meet inclusion criteria Shab-Bidar 2011 (100) Intervention did not meet inclusion criteria Sialvera 2013 (101) Intervention did not meet inclusion criteria Szymanska 2012 (61) Intervention did not meet inclusion criteria Trautvetter 2012 (102) Control did not meet inclusion criteria Tripolt 2013 (103) Control did not meet inclusion criteria Troost 2008 (104) Review outcomes not reported Tuomilehto 2004 Review outcomes not reported van der Zander 2008a (105) Intervention did not meet inclusion criteria van Mierlo 2009 (106) Intervention did not meet inclusion criteria

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 68

Widhalm 2011 (107) Intervention did not meet inclusion criteria

Chapter 9, Table 3: Summary of findings table

Probiotic versus placebo - primary analysis for primary prevention of cardiovascular disease in adults Patient or population: patients with primary prevention of cardiovascular disease in adults Intervention: Probiotic versus placebo - primary analysis Outcomes Illustrative comparative risks* (95% CI) No of Quality of Participants the evidence Assumed risk Corresponding risk (studies) (GRADE) Control Probiotic versus placebo - primary analysis Total cholesterol The mean total cholesterol The mean total cholesterol in 1296 Scale from: -0.68 ranged across control the intervention groups was (22 studies) high to 0.10. groups from 0.16 lower ⊕⊕⊕⊕ Follow-up: 2 - 56 -0.51 to 0.10 mmol/L (0.23 to 0.1 lower) weeks Low density The mean low density The mean low density 1238 lipoprotein lipoprotein cholesterol lipoprotein cholesterol in the (19 studies) high cholesterol ranged across control intervention groups was ⊕⊕⊕⊕ Scale from: -0.51 groups from 0.13 lower to 0.20. -0.39 to 0.26 mmol/L (0.19 to 0.07 lower) Follow-up: 2 - 56 weeks High density The mean high density The mean high density 1261 lipoprotein lipoprotein cholesterol lipoprotein cholesterol in the (21 studies) high cholesterol ranged across control intervention groups was ⊕⊕⊕⊕ Scale from: -0.23 groups from 0 higher to 0.11. -0.10 to 0.15 mmol/L (0.03 lower to 0.02 higher) Follow-up: 2 - 56 weeks Triglyceride The mean triglyceride The mean triglyceride in the 1395 Scale from: -0.30 ranged across control intervention groups was (23 studies) high to 0.30. groups from 0 higher ⊕⊕⊕⊕ Follow-up: 2 - 56 -0.25 to 0.30 mmol/L (0.05 lower to 0.05 higher) weeks *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 69

Chapter 9, Table 4: Study characteristics

Study Intervention Sample Eligible Randomised ITT Finished Duration sizea study n n n n n(%) (weeks)b Agerholm- I1 (G) - - 16 - 16 (100) 8 Larsen 2000 I2 (StLa) - - 16 - 16 (100) 8 I3 (StLr) - - 15 - 14 (93) 8 C (PY) - - 15 - 14 (93) 8 total: - 80 63 60 (97) Anderson I1 - - - - - 4 1999 C1 - - - - - 4 total: - 48 40 - 40 (100) Ataie-Jafari I1 - - - - - 6 2009 C1 - - - - - 6 total: 18 18 - 14 (78) de-Roos 1999 I1 - - - 39 6 C1 - - - 39 6 total: 90 85 78 (92) Ejtahed 2011 I1 30 - 32 30 6 C1 30 - 32 30 6 total: 60 64 64 30 60 (94) Fabian 2006 I1 - - - - - 2 C1 - - - - - 2 total: 33 33 - 32 (97) Fuentes 2013 I1 - - 30 - 30 (100) 12 C1 - - 30 - 30 (100) 12 total: 60 60 (100) Hata 1996 I1 - - 20 - 17 8 C1 - - 16 - 13 8 total: 36 30 (83) Hlivak 2005 I1 - - - - - 56 C1 - - - - - 56 total: 43 38 (88) Inoue 2003 I1 - - 20 18 (89) 12 C1 - - 19 17 (90) 12 total: 39 39 35 (90)

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 70

Study Intervention Sample Eligible Randomised ITT Finished Duration sizea study n n n n n(%) (weeks)b Ivey 2013 I1 40 39 (98) 6 I2 37 32 (86) 6 I3 40 38 (95) 6 C1 39 39 (100) 6 total: 156 887 156 148 (95) Jauhiainen I1 50 - 53 - 47 (89) 10 2005 C1 50 - 55 - 47 (89) 10 total: 100 139 108 94 (87) Jones 2012 I1 60 - 60 - 56 6 C1 60 - 60 - 58 6 total: 120 - 120 114 114 (95) Jones 2012a I1 57 - - 66 - 9 C1 57 - - 61 - 9 total: 114 396 131 127 127 (97) Kekkonen I1 - - 11 - 11 (100) 3 2008 C1 - - 15 - 15 (100) 3 total: 68 26 26 (100) Mizushima I1 - - 23 - 22 (96) 4 2004 C1 - - 23 - 20 (87) 4 total: 62 46 46 42 (91) Rizkalla 2000 I1 - - - - - 2.1 C1 - - - - - 2.1 total: 24 24 - Sadrzadeh- I1 - - 30 - 29 (97) 6 Yeganeh 2010 C1 - - 30 - 30 (100) 6 total: 90 60 59 (98) Simons 2006 I1 - - 22 21 (95) 10 C1 - - 24 23 (96) 10 total: 46 44 (96) Songisepp I1 - - - - - 3 2012 1 C1 - - - - - 3 total: 8 13 13 (100)

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 71

Study Intervention Sample Eligible Randomised ITT Finished Duration sizea study n n n n n(%) (weeks)b Songisepp I1 - - - - - 3 2012 2 C1 - - - - - 3 total: 8 21 18 (86) Usinger 2010a I1 - - 32 - 30 (94) 8 I2 - - 32 - 29 (91) 8 C1 - - 15 - 15 (100) 8 C2 - - 15 - 15 (100) 8 total: 94 94 90 (96) Xiao 2003 I1 - - 16 - 16 (100) 4 C1 - - 16 - 16 (100) 4 total: 32 32 (100)

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 72

Chapter 9, Table 5: Genus and species of intervention probiotic bacteria

Genus and species Studies B. longum Xiao 2003 E. faecum Hlivak 2005 L. acidophilus Anderson 1999; de-Roos 1999 L. fermentum Simons 2006; Songisepp 2012: study 1; Songisepp 2012: study 2 L. helveticus Jauhiainen 2005; Usinger 2010 L. paracasei Fabian 2006 L. plantarium Fuentes 2013 L. reuteri Jones 2012; Jones 2012a L. rhamnosus Kekkonen 2008 L. acidophilus and B. lactis Ataie-Jafari 2009; Ejtahed 2011; Ivey 2013; Sadrzadeh-Yeganeh 2010 L. acidophilus and S. thermophillus Agerholm-Larsen 2000 L. bulgaricus and S. thermophillus Rizkalla 2000 L. casei and L. lactis Inoue 2003 L. helveticus and S. cerivasea Hata 1996; Mizushima 2004 L. rhamnosus and S. thermophillus Agerholm-Larsen 2000 E. faecum and S. thermophillus Agerholm-Larsen 2000

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 73

9.15 FIGURES

Chapter 9, Figure 1: Flow diagram of included studies

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 74

Chapter 9, Figure 2: Risk of bias graph

Random sequence generation (selection bias)

Allocation concealment (selection bias)

Blinding of participants and personnel (performance bias): Objective outcomes

Blinding of participants and personnel (performance bias): Subjective outcomes

Blinding of outcome assessment (detection bias): Objective outcomes

Blinding of outcome assessment (detection bias): Subjective outcomes

Incomplete outcome data (attrition bias): Objective outcomes

Incomplete outcome data (attrition bias): All outcomes

Selective reporting (reporting bias)

Other bias

0% 25% 50% 75% 100%

Low risk of bias Unclear risk of bias High risk of bias

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 75

Chapter 9, Figure 3: Risk of bias summary Random sequence generation (selection bias) (selection generation sequence Random bias) (selection concealment Allocation Objective outcomes bias): Blinding (performance and personnel of participants Subjective outcomes bias): Blinding (performance and personnel of participants outcomes Objective bias): (detection assessment outcome of Blinding outcomes Subjective bias): (detection assessment outcome of Blinding Objective outcomes bias): Incomplete outcome data (attrition All outcomes bias): Incomplete outcome data (attrition bias) (reporting Selective reporting bias Other

Agerholm-Larsen 2000 ? – + ? ? ? + + +

Anderson 1999 ? ? + ? ? ? + + +

Ataie-Jafari 2009 ? ? – ? ? ? + + +

de-Roos 1999 ? ? ? ? ? ? + + +

Ejtahed 2011 + ? + ? ? ? + + +

Fabian 2006 ? ? ? ? ? ? + + +

Fuentes 2013 ? ? + ? ? ? + + +

Hata 1996 ? ? ? ? ? ? + + +

Hlivak 2005 ? ? + ? ? ? + + +

Inoue 2003 ? ? – ? ? ? + + +

Ivey 2013 + ? + ? + ? + + +

Jauhiainen 2005 ? + + ? ? ? + + +

Jones 2012 ? ? + ? ? ? + + +

Jones 2012a + + + ? ? ? + + +

Kekkonen 2008 ? ? + ? ? ? + + +

Mizushima 2004 ? ? + ? ? ? + + +

Rizkalla 2000 ? ? ? ? ? ? + + +

Sadrzadeh-Yeganeh 2010 ? ? + ? + ? + + +

Simons 2006 ? ? + ? ? ? + + +

Songisepp 2012: study 1 ? ? + ? ? ? + + +

Songisepp 2012: study 2 ? ? + ? ? ? + + +

Usinger 2010 ? ? + ? ? ? + + +

Xiao 2003 ? ? – ? ? ? + + +

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 76

Chapter 9, Figure 4 (analysis 1.8): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.8 Mild adverse events.

Probiotic Placebo Odds Ratio Odds Ratio Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI

Agerholm-Larsen 2000 1 47 1 16 29.1% 0.33 [0.02, 5.54] de-Roos 1999 1 78 1 78 19.7% 1.00 [0.06, 16.28] Simons 2006 2 21 1 23 17.2% 2.32 [0.19, 27.59] Songisepp 2012: study 1 3 13 0 13 7.5% 9.00 [0.42, 194.07] Usinger 2010 1 60 0 30 12.9% 1.54 [0.06, 38.88] Xiao 2003 5 16 1 16 13.7% 6.82 [0.69, 66.90]

Total (95% CI) 235 176 100.0% 2.49 [0.89, 7.02]

Total events 13 4 Heterogeneity: Chi² = 3.90, df = 5 (P = 0.56); I² = 0% 0.001 0.1 1 10 1000 Test for overall effect: Z = 1.73 (P = 0.08) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 77

Chapter 9, Figure 5 (analysis 1.1): Forest plot of comparison: 1 Probiotic versus placebo - whole data set, outcome: 1.1 Total cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.08 0.3 46 0.06 0.64 14 3.7% -0.14 [-0.49, 0.21] Anderson 1999 -0.04 0.61 40 0.04 0.52 40 7.1% -0.08 [-0.33, 0.17] Ataie-Jafari 2009 -0.32 0.03 14 0.36 1.02 14 1.5% -0.68 [-1.21, -0.15] de-Roos 1999 -0.02 0.55 39 -0.07 0.54 39 7.5% 0.05 [-0.19, 0.29] Ejtahed 2011 -0.24 0.52 30 -0.04 0.44 30 7.4% -0.20 [-0.44, 0.04] Fabian 2006 0.1 0.57 17 0 0.53 16 3.1% 0.10 [-0.28, 0.48] Fuentes 2013 -0.87 0.57 30 -0.51 0.49 30 6.1% -0.36 [-0.63, -0.09] Hata 1996 -0.16 0.41 17 -0.14 0.61 13 3.0% -0.02 [-0.40, 0.36] Hlivak 2005 -0.72 0.92 20 -0.3 0.8 18 1.5% -0.42 [-0.97, 0.13] Inoue 2003 0.08 0.62 18 -0.01 0.82 17 1.9% 0.09 [-0.39, 0.57] Ivey 2013 -0.08 0.5 109 -0.01 0.67 39 8.3% -0.07 [-0.30, 0.16] Jones 2012 -0.71 0.9 56 -0.38 0.79 58 4.5% -0.33 [-0.64, -0.02] Jones 2012a -0.42 0.63 66 0.16 0.67 61 8.6% -0.58 [-0.81, -0.35] Kekkonen 2008 0 0.4 11 0.2 0.5 15 3.7% -0.20 [-0.55, 0.15] Mizushima 2004 -0.04 0.58 23 0.2 0.56 23 4.1% -0.24 [-0.57, 0.09] Rizkalla 2000 -0.15 0.99 24 -0.1 1.03 24 1.3% -0.05 [-0.62, 0.52] Sadrzadeh-Yeganeh 2010 -0.2 0.33 29 -0.13 0.36 30 14.2% -0.07 [-0.25, 0.11] Simons 2006 -0.4 0.72 23 -0.3 0.78 21 2.2% -0.10 [-0.54, 0.34] Songisepp 2012: study 1 0 0.78 12 0.3 0.63 12 1.4% -0.30 [-0.87, 0.27] Songisepp 2012: study 2 -0.1 0.6 18 -0.2 0.65 18 2.6% 0.10 [-0.31, 0.51] Usinger 2010 -0.05 0.74 60 0 0.57 30 5.7% -0.05 [-0.33, 0.23] Xiao 2003 -0.17 1.21 16 -0.02 1.45 16 0.5% -0.15 [-1.08, 0.78]

Total (95% CI) 718 578 100.0% -0.16 [-0.23, -0.10]

Heterogeneity: Chi² = 32.23, df = 21 (P = 0.06); I² = 35% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.78 (P < 0.00001) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 78

Chapter 9, Figure 6 (analysis 1.1): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.1 Total cholesterol.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 79

Chapter 9, Figure 7 (analysis 1.2): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.3 High density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.08 0.3 46 0.06 0.64 14 4.1% -0.14 [-0.49, 0.21] Anderson 1999 -0.04 0.61 40 0.04 0.52 40 7.9% -0.08 [-0.33, 0.17] de-Roos 1999 -0.02 0.55 39 -0.07 0.54 39 8.4% 0.05 [-0.19, 0.29] Ejtahed 2011 -0.24 0.52 30 -0.04 0.44 30 8.2% -0.20 [-0.44, 0.04] Fabian 2006 0.1 0.57 17 0 0.53 16 3.5% 0.10 [-0.28, 0.48] Fuentes 2013 -0.87 0.57 30 -0.51 0.49 30 6.8% -0.36 [-0.63, -0.09] Hata 1996 -0.16 0.41 17 -0.14 0.61 13 3.3% -0.02 [-0.40, 0.36] Hlivak 2005 -0.72 0.92 20 -0.3 0.8 18 1.6% -0.42 [-0.97, 0.13] Inoue 2003 0.08 0.62 18 -0.01 0.82 17 2.1% 0.09 [-0.39, 0.57] Ivey 2013 -0.08 0.5 109 -0.01 0.67 39 9.2% -0.07 [-0.30, 0.16] Jones 2012 -0.71 0.9 56 -0.38 0.79 58 5.1% -0.33 [-0.64, -0.02] Kekkonen 2008 0 0.4 11 0.2 0.5 15 4.1% -0.20 [-0.55, 0.15] Mizushima 2004 -0.04 0.58 23 0.2 0.56 23 4.5% -0.24 [-0.57, 0.09] Rizkalla 2000 -0.15 0.99 24 -0.1 1.03 24 1.5% -0.05 [-0.62, 0.52] Sadrzadeh-Yeganeh 2010 -0.2 0.33 29 -0.13 0.36 30 15.8% -0.07 [-0.25, 0.11] Simons 2006 -0.4 0.72 23 -0.3 0.78 21 2.5% -0.10 [-0.54, 0.34] Songisepp 2012: study 1 0 0.78 12 0.3 0.63 12 1.5% -0.30 [-0.87, 0.27] Songisepp 2012: study 2 -0.1 0.6 18 -0.2 0.65 18 2.9% 0.10 [-0.31, 0.51] Usinger 2010 -0.05 0.74 60 0 0.57 30 6.4% -0.05 [-0.33, 0.23] Xiao 2003 -0.17 1.21 16 -0.02 1.45 16 0.6% -0.15 [-1.08, 0.78]

Total (95% CI) 638 503 100.0% -0.11 [-0.18, -0.04]

Heterogeneity: Chi² = 13.67, df = 19 (P = 0.80); I² = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 3.17 (P = 0.002) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 80

Chapter 9, Figure 8 (analysis 1.2): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.2 Total cholesterol - excluding outliers.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 81

Chapter 9, Figure 9 (analysis 2.1): Forest plot of comparison: 2 Probiotic versus placebo - sensitivity analysis excluding unpublished studies, outcome: 2.1 Total cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.08 0.3 46 0.06 0.64 14 4.0% -0.14 [-0.49, 0.21] Anderson 1999 -0.04 0.61 40 0.04 0.52 40 7.8% -0.08 [-0.33, 0.17] Ataie-Jafari 2009 -0.32 0.03 14 0.36 1.02 14 1.7% -0.68 [-1.21, -0.15] de-Roos 1999 -0.02 0.55 39 -0.07 0.54 39 8.2% 0.05 [-0.19, 0.29] Ejtahed 2011 -0.24 0.52 30 -0.04 0.44 30 8.1% -0.20 [-0.44, 0.04] Fabian 2006 0.1 0.57 17 0 0.53 16 3.4% 0.10 [-0.28, 0.48] Fuentes 2013 -0.87 0.57 30 -0.51 0.49 30 6.6% -0.36 [-0.63, -0.09] Hata 1996 -0.16 0.41 17 -0.14 0.61 13 3.2% -0.02 [-0.40, 0.36] Hlivak 2005 -0.72 0.92 20 -0.3 0.8 18 1.6% -0.42 [-0.97, 0.13] Inoue 2003 0.08 0.62 18 -0.01 0.82 17 2.1% 0.09 [-0.39, 0.57] Jones 2012 -0.71 0.9 56 -0.38 0.79 58 5.0% -0.33 [-0.64, -0.02] Jones 2012a -0.42 0.63 66 0.16 0.67 61 9.3% -0.58 [-0.81, -0.35] Kekkonen 2008 0 0.4 11 0.2 0.5 15 4.0% -0.20 [-0.55, 0.15] Mizushima 2004 -0.04 0.58 23 0.2 0.56 23 4.4% -0.24 [-0.57, 0.09] Rizkalla 2000 -0.15 0.99 24 -0.1 1.03 24 1.5% -0.05 [-0.62, 0.52] Sadrzadeh-Yeganeh 2010 -0.2 0.33 29 -0.13 0.36 30 15.5% -0.07 [-0.25, 0.11] Simons 2006 -0.4 0.72 23 -0.3 0.78 21 2.4% -0.10 [-0.54, 0.34] Songisepp 2012: study 1 0 0.78 12 0.3 0.63 12 1.5% -0.30 [-0.87, 0.27] Songisepp 2012: study 2 -0.1 0.6 18 -0.2 0.65 18 2.9% 0.10 [-0.31, 0.51] Usinger 2010 -0.05 0.74 60 0 0.57 30 6.3% -0.05 [-0.33, 0.23] Xiao 2003 -0.17 1.21 16 -0.02 1.45 16 0.6% -0.15 [-1.08, 0.78]

Total (95% CI) 609 539 100.0% -0.17 [-0.24, -0.10]

Heterogeneity: Chi² = 31.56, df = 20 (P = 0.05); I² = 37% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.81 (P < 0.00001) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 82

Chapter 9, Figure 11 (analysis 4.1): Forest plot of comparison: 4 Probiotic versus placebo - sensitivity analysis based on number of participants included in analysis, outcome: 4.1 Total cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 4.1.2 Small sample size (< 100 participants)

Agerholm-Larsen 2000 -0.08 0.3 46 0.06 0.64 14 3.7% -0.14 [-0.49, 0.21] Anderson 1999 -0.04 0.61 40 0.04 0.52 40 7.1% -0.08 [-0.33, 0.17] Ataie-Jafari 2009 -0.32 0.03 14 0.36 1.02 14 1.5% -0.68 [-1.21, -0.15] de-Roos 1999 -0.02 0.55 39 -0.07 0.54 39 7.5% 0.05 [-0.19, 0.29] Ejtahed 2011 -0.24 0.52 30 -0.04 0.44 30 7.4% -0.20 [-0.44, 0.04] Fabian 2006 0.1 0.57 17 0 0.53 16 3.1% 0.10 [-0.28, 0.48] Fuentes 2013 -0.87 0.57 30 -0.51 0.49 30 6.1% -0.36 [-0.63, -0.09] Hata 1996 -0.16 0.41 17 -0.14 0.61 13 3.0% -0.02 [-0.40, 0.36] Hlivak 2005 -0.72 0.92 20 -0.3 0.8 18 1.5% -0.42 [-0.97, 0.13] Inoue 2003 0.08 0.62 18 -0.01 0.82 17 1.9% 0.09 [-0.39, 0.57] Kekkonen 2008 0 0.4 11 0.2 0.5 15 3.7% -0.20 [-0.55, 0.15] Mizushima 2004 -0.04 0.58 23 0.2 0.56 23 4.1% -0.24 [-0.57, 0.09] Rizkalla 2000 -0.15 0.99 24 -0.1 1.03 24 1.3% -0.05 [-0.62, 0.52] Sadrzadeh-Yeganeh 2010 -0.2 0.33 29 -0.13 0.36 30 14.2% -0.07 [-0.25, 0.11] Simons 2006 -0.4 0.72 23 -0.3 0.78 21 2.2% -0.10 [-0.54, 0.34] Songisepp 2012: study 1 0 0.78 12 0.3 0.63 12 1.4% -0.30 [-0.87, 0.27] Songisepp 2012: study 2 -0.1 0.6 18 -0.2 0.65 18 2.6% 0.10 [-0.31, 0.51] Usinger 2010 -0.05 0.74 60 0 0.57 30 5.7% -0.05 [-0.33, 0.23] Xiao 2003 -0.17 1.21 16 -0.02 1.45 16 0.5% -0.15 [-1.08, 0.78] Subtotal (95% CI) 487 420 78.6% -0.12 [-0.19, -0.04]

Heterogeneity: Chi² = 15.99, df = 18 (P = 0.59); I² = 0% Test for overall effect: Z = 3.04 (P = 0.002)

4.1.3 Large sample size (> 100 participants)

Ivey 2013 -0.08 0.5 109 -0.01 0.67 39 8.3% -0.07 [-0.30, 0.16] Jones 2012 -0.71 0.9 56 -0.38 0.79 58 4.5% -0.33 [-0.64, -0.02] Jones 2012a -0.42 0.63 66 0.16 0.67 61 8.6% -0.58 [-0.81, -0.35] Subtotal (95% CI) 231 158 21.4% -0.33 [-0.47, -0.19]

Heterogeneity: Chi² = 9.57, df = 2 (P = 0.008); I² = 79% Test for overall effect: Z = 4.50 (P < 0.00001)

Total (95% CI) 718 578 100.0% -0.16 [-0.23, -0.10]

Heterogeneity: Chi² = 32.23, df = 21 (P = 0.06); I² = 35% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.78 (P < 0.00001) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 6.67, df = 1 (P = 0.010), I² = 85.0%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 83

Chapter 9, Figure 12 (analysis 5.1): Forest plot of comparison: 5 Probiotic versus placebo - sensitivity analysis using Random Effects model, outcome: 5.1 Total cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI

Agerholm-Larsen 2000 -0.08 0.3 46 0.06 0.64 14 4.4% -0.14 [-0.49, 0.21] Anderson 1999 -0.04 0.61 40 0.04 0.52 40 6.6% -0.08 [-0.33, 0.17] Ataie-Jafari 2009 -0.32 0.03 14 0.36 1.02 14 2.2% -0.68 [-1.21, -0.15] de-Roos 1999 -0.02 0.55 39 -0.07 0.54 39 6.8% 0.05 [-0.19, 0.29] Ejtahed 2011 -0.24 0.52 30 -0.04 0.44 30 6.7% -0.20 [-0.44, 0.04] Fabian 2006 0.1 0.57 17 0 0.53 16 3.9% 0.10 [-0.28, 0.48] Fuentes 2013 -0.87 0.57 30 -0.51 0.49 30 6.0% -0.36 [-0.63, -0.09] Hata 1996 -0.16 0.41 17 -0.14 0.61 13 3.8% -0.02 [-0.40, 0.36] Hlivak 2005 -0.72 0.92 20 -0.3 0.8 18 2.2% -0.42 [-0.97, 0.13] Inoue 2003 0.08 0.62 18 -0.01 0.82 17 2.6% 0.09 [-0.39, 0.57] Ivey 2013 -0.08 0.5 109 -0.01 0.67 39 7.1% -0.07 [-0.30, 0.16] Jones 2012 -0.71 0.9 56 -0.38 0.79 58 5.1% -0.33 [-0.64, -0.02] Jones 2012a -0.42 0.63 66 0.16 0.67 61 7.3% -0.58 [-0.81, -0.35] Kekkonen 2008 0 0.4 11 0.2 0.5 15 4.4% -0.20 [-0.55, 0.15] Mizushima 2004 -0.04 0.58 23 0.2 0.56 23 4.7% -0.24 [-0.57, 0.09] Rizkalla 2000 -0.15 0.99 24 -0.1 1.03 24 2.0% -0.05 [-0.62, 0.52] Sadrzadeh-Yeganeh 2010 -0.2 0.33 29 -0.13 0.36 30 9.0% -0.07 [-0.25, 0.11] Simons 2006 -0.4 0.72 23 -0.3 0.78 21 3.0% -0.10 [-0.54, 0.34] Songisepp 2012: study 1 0 0.78 12 0.3 0.63 12 2.0% -0.30 [-0.87, 0.27] Songisepp 2012: study 2 -0.1 0.6 18 -0.2 0.65 18 3.4% 0.10 [-0.31, 0.51] Usinger 2010 -0.05 0.74 60 0 0.57 30 5.8% -0.05 [-0.33, 0.23] Xiao 2003 -0.17 1.21 16 -0.02 1.45 16 0.8% -0.15 [-1.08, 0.78]

Total (95% CI) 718 578 100.0% -0.16 [-0.25, -0.08]

Heterogeneity: Tau² = 0.01; Chi² = 32.23, df = 21 (P = 0.06); I² = 35% -1 -0.5 0 0.5 1 Test for overall effect: Z = 3.67 (P = 0.0002) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 84

Chapter 9, Figure 13 (analysis 6.1): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.1 Total cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 6.1.2 Capsules

Agerholm-Larsen 2000 -0.08 0.3 46 0.06 0.64 14 3.7% -0.14 [-0.49, 0.21] Fuentes 2013 -0.87 0.57 30 -0.51 0.49 30 6.1% -0.36 [-0.63, -0.09] Hlivak 2005 -0.72 0.92 20 -0.3 0.8 18 1.5% -0.42 [-0.97, 0.13] Ivey 2013 -0.01 0.47 39 -0.01 0.67 19 3.9% 0.00 [-0.34, 0.34] Jones 2012a -0.42 0.63 66 0.16 0.67 61 8.6% -0.58 [-0.81, -0.35] Simons 2006 -0.4 0.72 23 -0.3 0.78 21 2.2% -0.10 [-0.54, 0.34] Subtotal (95% CI) 224 163 26.0% -0.33 [-0.46, -0.20]

Heterogeneity: Chi² = 10.73, df = 5 (P = 0.06); I² = 53% Test for overall effect: Z = 4.94 (P < 0.00001)

6.1.3 Fermented milk

Anderson 1999 -0.04 0.61 40 0.04 0.52 40 7.2% -0.08 [-0.33, 0.17] Ataie-Jafari 2009 -0.32 0.03 14 0.36 1.02 14 1.5% -0.68 [-1.21, -0.15] de-Roos 1999 -0.02 0.55 39 -0.07 0.54 39 7.6% 0.05 [-0.19, 0.29] Ejtahed 2011 -0.24 0.52 30 -0.04 0.44 30 7.4% -0.20 [-0.44, 0.04] Fabian 2006 0.1 0.57 17 0 0.53 16 3.1% 0.10 [-0.28, 0.48] Hata 1996 -0.16 0.41 17 -0.14 0.61 13 3.0% -0.02 [-0.40, 0.36] Inoue 2003 0.08 0.62 18 -0.01 0.82 17 1.9% 0.09 [-0.39, 0.57] Ivey 2013 -0.04 0.45 32 -0.01 0.67 20 4.0% -0.03 [-0.36, 0.30] Jones 2012 -0.71 0.9 56 -0.38 0.79 58 4.6% -0.33 [-0.64, -0.02] Kekkonen 2008 0 0.4 11 0.2 0.5 15 3.7% -0.20 [-0.55, 0.15] Mizushima 2004 -0.04 0.58 23 0.2 0.56 23 4.1% -0.24 [-0.57, 0.09] Rizkalla 2000 -0.15 0.99 24 -0.1 1.03 24 1.4% -0.05 [-0.62, 0.52] Sadrzadeh-Yeganeh 2010 -0.2 0.33 29 -0.13 0.36 30 14.3% -0.07 [-0.25, 0.11] Songisepp 2012: study 1 0 0.78 12 0.3 0.63 12 1.4% -0.30 [-0.87, 0.27] Songisepp 2012: study 2 -0.1 0.6 18 -0.2 0.65 18 2.6% 0.10 [-0.31, 0.51] Usinger 2010 -0.05 0.74 60 0 0.57 30 5.8% -0.05 [-0.33, 0.23] Xiao 2003 -0.17 1.21 16 -0.02 1.45 16 0.5% -0.15 [-1.08, 0.78] Subtotal (95% CI) 456 415 74.0% -0.10 [-0.18, -0.02]

Heterogeneity: Chi² = 13.47, df = 16 (P = 0.64); I² = 0% Test for overall effect: Z = 2.48 (P = 0.01)

Total (95% CI) 680 578 100.0% -0.16 [-0.22, -0.09]

Heterogeneity: Chi² = 33.10, df = 22 (P = 0.06); I² = 34% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.65 (P < 0.00001) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 8.90, df = 1 (P = 0.003), I² = 88.8%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 85

Chapter 9, Figure 14 (analysis 7.1): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.1 Total cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 7.1.2 Desirable baseline total-cholesterol

Agerholm-Larsen 2000 -0.08 0.3 46 0.06 0.64 14 3.7% -0.14 [-0.49, 0.21] de-Roos 1999 -0.02 0.55 39 -0.07 0.54 39 7.5% 0.05 [-0.19, 0.29] Ejtahed 2011 -0.24 0.52 30 -0.04 0.44 30 7.4% -0.20 [-0.44, 0.04] Fabian 2006 0.1 0.57 17 0 0.53 16 3.1% 0.10 [-0.28, 0.48] Rizkalla 2000 -0.15 0.99 24 -0.1 1.03 24 1.3% -0.05 [-0.62, 0.52] Sadrzadeh-Yeganeh 2010 -0.2 0.33 29 -0.13 0.36 30 14.2% -0.07 [-0.25, 0.11] Songisepp 2012: study 1 0 0.78 12 0.3 0.63 12 1.4% -0.30 [-0.87, 0.27] Usinger 2010 -0.05 0.74 60 0 0.57 30 5.7% -0.05 [-0.33, 0.23] Subtotal (95% CI) 257 195 44.4% -0.07 [-0.17, 0.03]

Heterogeneity: Chi² = 3.64, df = 7 (P = 0.82); I² = 0% Test for overall effect: Z = 1.36 (P = 0.17)

7.1.3 Borderline high baseline total-cholesterol

Ataie-Jafari 2009 -0.32 0.03 14 0.36 1.02 14 1.5% -0.68 [-1.21, -0.15] Hata 1996 -0.16 0.41 17 -0.14 0.61 13 3.0% -0.02 [-0.40, 0.36] Hlivak 2005 -0.72 0.92 20 -0.3 0.8 18 1.5% -0.42 [-0.97, 0.13] Inoue 2003 0.08 0.62 18 -0.01 0.82 17 1.9% 0.09 [-0.39, 0.57] Ivey 2013 -0.08 0.5 109 -0.01 0.67 39 8.3% -0.07 [-0.30, 0.16] Jones 2012a -0.42 0.63 66 0.16 0.67 61 8.6% -0.58 [-0.81, -0.35] Kekkonen 2008 0 0.4 11 0.2 0.5 15 3.7% -0.20 [-0.55, 0.15] Mizushima 2004 -0.04 0.58 23 0.2 0.56 23 4.1% -0.24 [-0.57, 0.09] Songisepp 2012: study 2 -0.1 0.6 18 -0.2 0.65 18 2.6% 0.10 [-0.31, 0.51] Subtotal (95% CI) 296 218 35.1% -0.24 [-0.36, -0.13]

Heterogeneity: Chi² = 19.51, df = 8 (P = 0.01); I² = 59% Test for overall effect: Z = 4.26 (P < 0.0001)

7.1.4 High baseline total-cholesterol

Anderson 1999 -0.04 0.61 40 0.04 0.52 40 7.1% -0.08 [-0.33, 0.17] Fuentes 2013 -0.87 0.57 30 -0.51 0.49 30 6.1% -0.36 [-0.63, -0.09] Jones 2012 -0.71 0.9 56 -0.38 0.79 58 4.5% -0.33 [-0.64, -0.02] Simons 2006 -0.4 0.72 23 -0.3 0.78 21 2.2% -0.10 [-0.54, 0.34] Xiao 2003 -0.17 1.21 16 -0.02 1.45 16 0.5% -0.15 [-1.08, 0.78] Subtotal (95% CI) 165 165 20.5% -0.22 [-0.37, -0.08]

Heterogeneity: Chi² = 3.04, df = 4 (P = 0.55); I² = 0% Test for overall effect: Z = 2.98 (P = 0.003)

Total (95% CI) 718 578 100.0% -0.16 [-0.23, -0.10]

Heterogeneity: Chi² = 32.23, df = 21 (P = 0.06); I² = 35% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.78 (P < 0.00001) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 6.04, df = 2 (P = 0.05), I² = 66.9%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 86

Chapter 9, Figure 15 (analysis 7.2): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.2 Total cholesterol - Borderline high studies split into two groups (above and below the median baseline total-cholesterol).

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 7.2.1 Borderline-high baseline total-cholesterol - below the median

Hata 1996 -0.16 0.41 17 -0.14 0.61 13 8.5% -0.02 [-0.40, 0.36] Inoue 2003 0.08 0.62 18 -0.01 0.82 17 5.4% 0.09 [-0.39, 0.57] Ivey 2013 -0.08 0.5 109 -0.01 0.67 39 23.6% -0.07 [-0.30, 0.16] Kekkonen 2008 0 0.4 11 0.2 0.5 15 10.5% -0.20 [-0.55, 0.15] Mizushima 2004 -0.04 0.58 23 0.2 0.56 23 11.6% -0.24 [-0.57, 0.09] Subtotal (95% CI) 178 107 59.5% -0.10 [-0.25, 0.04]

Heterogeneity: Chi² = 1.83, df = 4 (P = 0.77); I² = 0% Test for overall effect: Z = 1.41 (P = 0.16)

7.2.3 Borderline-high baseline total-cholesterol - above the median

Ataie-Jafari 2009 -0.32 0.03 14 0.36 1.02 14 4.4% -0.68 [-1.21, -0.15] Hlivak 2005 -0.72 0.92 20 -0.3 0.8 18 4.2% -0.42 [-0.97, 0.13] Jones 2012a -0.42 0.63 66 0.16 0.67 61 24.4% -0.58 [-0.81, -0.35] Songisepp 2012: study 2 -0.1 0.6 18 -0.2 0.65 18 7.5% 0.10 [-0.31, 0.51] Subtotal (95% CI) 118 111 40.5% -0.45 [-0.62, -0.27]

Heterogeneity: Chi² = 8.94, df = 3 (P = 0.03); I² = 66% Test for overall effect: Z = 4.99 (P < 0.00001)

Total (95% CI) 296 218 100.0% -0.24 [-0.36, -0.13]

Heterogeneity: Chi² = 19.51, df = 8 (P = 0.01); I² = 59% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.26 (P < 0.0001) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 8.73, df = 1 (P = 0.003), I² = 88.5%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 87

Chapter 9, Figure 16 (analysis 1.3): Forest plot of comparison: 1 Probiotic versus placebo - whole data set, outcome: 1.2 Low density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.031 0.47 46 0.11 0.41 14 6.0% -0.14 [-0.40, 0.11] Anderson 1999 -0.1 0.64 40 0.038 0.57 40 5.5% -0.14 [-0.40, 0.13] Ataie-Jafari 2009 -0.24 0.52 14 0.26 0.98 14 1.1% -0.50 [-1.08, 0.08] de-Roos 1999 -0.02 0.64 39 -0.08 0.75 39 4.1% 0.06 [-0.25, 0.37] Ejtahed 2011 -0.25 0.56 30 -0.03 0.59 30 4.6% -0.22 [-0.51, 0.07] Fabian 2006 0 0.56 17 0 0.52 16 2.9% 0.00 [-0.37, 0.37] Fuentes 2013 -0.63 0.46 30 -0.35 0.46 30 7.2% -0.28 [-0.51, -0.05] Hlivak 2005 -0.76 1.03 20 -0.39 0.83 18 1.1% -0.37 [-0.96, 0.22] Ivey 2013 -0.06 0.4 109 0 0.54 39 11.3% -0.06 [-0.25, 0.13] Jauhiainen 2005 -0.06 0.51 50 -0.06 0.55 51 9.1% 0.00 [-0.21, 0.21] Jones 2012 -0.33 0.78 56 0.06 0.69 58 5.3% -0.39 [-0.66, -0.12] Jones 2012a -0.3 0.57 66 0.21 0.63 61 8.8% -0.51 [-0.72, -0.30] Kekkonen 2008 0.1 0.3 11 0.1 0.5 15 4.1% 0.00 [-0.31, 0.31] Sadrzadeh-Yeganeh 2010 -0.06 0.29 29 -0.03 0.24 30 21.0% -0.03 [-0.17, 0.11] Simons 2006 -0.3 0.86 23 -0.3 0.95 21 1.3% 0.00 [-0.54, 0.54] Songisepp 2012: study 1 0.1 0.93 12 0 0.65 12 0.9% 0.10 [-0.54, 0.74] Songisepp 2012: study 2 -0.1 0.71 18 -0.3 0.77 18 1.7% 0.20 [-0.28, 0.68] Usinger 2010 -0.1 0.86 60 0 0.65 30 3.8% -0.10 [-0.42, 0.22] Xiao 2003 -0.13 1.85 16 -0.07 1.91 16 0.2% -0.06 [-1.36, 1.24]

Total (95% CI) 686 552 100.0% -0.13 [-0.19, -0.07]

Heterogeneity: Chi² = 29.63, df = 18 (P = 0.04); I² = 39% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.16 (P < 0.0001) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 88

Chapter 9, Figure 17 (analysis 1.3): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.3 Low density lipoprotein cholesterol.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 89

Chapter 9, Figure 18 (analysis 1.4): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.4 Low density lipoprotein cholesterol - excluding outliers.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.031 0.47 46 0.11 0.41 14 7.0% -0.14 [-0.40, 0.11] Anderson 1999 -0.1 0.64 40 0.038 0.57 40 6.4% -0.14 [-0.40, 0.13] Ataie-Jafari 2009 -0.24 0.52 14 0.26 0.98 14 1.3% -0.50 [-1.08, 0.08] de-Roos 1999 -0.02 0.64 39 -0.08 0.75 39 4.7% 0.06 [-0.25, 0.37] Ejtahed 2011 -0.25 0.56 30 -0.03 0.59 30 5.3% -0.22 [-0.51, 0.07] Fabian 2006 0 0.56 17 0 0.52 16 3.3% 0.00 [-0.37, 0.37] Fuentes 2013 -0.63 0.46 30 -0.35 0.46 30 8.3% -0.28 [-0.51, -0.05] Hlivak 2005 -0.76 1.03 20 -0.39 0.83 18 1.3% -0.37 [-0.96, 0.22] Ivey 2013 -0.06 0.4 109 0 0.54 39 13.2% -0.06 [-0.25, 0.13] Jauhiainen 2005 -0.06 0.51 50 -0.06 0.55 51 10.6% 0.00 [-0.21, 0.21] Kekkonen 2008 0.1 0.3 11 0.1 0.5 15 4.7% 0.00 [-0.31, 0.31] Sadrzadeh-Yeganeh 2010 -0.06 0.29 29 -0.03 0.24 30 24.4% -0.03 [-0.17, 0.11] Simons 2006 -0.3 0.86 23 -0.3 0.95 21 1.6% 0.00 [-0.54, 0.54] Songisepp 2012: study 1 0.1 0.93 12 0 0.65 12 1.1% 0.10 [-0.54, 0.74] Songisepp 2012: study 2 -0.1 0.71 18 -0.3 0.77 18 1.9% 0.20 [-0.28, 0.68] Usinger 2010 -0.1 0.86 60 0 0.65 30 4.5% -0.10 [-0.42, 0.22] Xiao 2003 -0.13 1.85 16 -0.07 1.91 16 0.3% -0.06 [-1.36, 1.24]

Total (95% CI) 564 433 100.0% -0.08 [-0.14, -0.01]

Heterogeneity: Chi² = 11.10, df = 16 (P = 0.80); I² = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 2.25 (P = 0.02) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 90

Chapter 9, Figure 19 (analysis 1.4): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.4 Low density lipoprotein cholesterol - excluding outliers.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.031 0.47 46 0.11 0.41 14 7.0% -0.14 [-0.40, 0.11] Anderson 1999 -0.1 0.64 40 0.038 0.57 40 6.4% -0.14 [-0.40, 0.13] Ataie-Jafari 2009 -0.24 0.52 14 0.26 0.98 14 1.3% -0.50 [-1.08, 0.08] de-Roos 1999 -0.02 0.64 39 -0.08 0.75 39 4.7% 0.06 [-0.25, 0.37] Ejtahed 2011 -0.25 0.56 30 -0.03 0.59 30 5.3% -0.22 [-0.51, 0.07] Fabian 2006 0 0.56 17 0 0.52 16 3.3% 0.00 [-0.37, 0.37] Fuentes 2013 -0.63 0.46 30 -0.35 0.46 30 8.3% -0.28 [-0.51, -0.05] Hlivak 2005 -0.76 1.03 20 -0.39 0.83 18 1.3% -0.37 [-0.96, 0.22] Ivey 2013 -0.06 0.4 109 0 0.54 39 13.2% -0.06 [-0.25, 0.13] Jauhiainen 2005 -0.06 0.51 50 -0.06 0.55 51 10.6% 0.00 [-0.21, 0.21] Kekkonen 2008 0.1 0.3 11 0.1 0.5 15 4.7% 0.00 [-0.31, 0.31] Sadrzadeh-Yeganeh 2010 -0.06 0.29 29 -0.03 0.24 30 24.4% -0.03 [-0.17, 0.11] Simons 2006 -0.3 0.86 23 -0.3 0.95 21 1.6% 0.00 [-0.54, 0.54] Songisepp 2012: study 1 0.1 0.93 12 0 0.65 12 1.1% 0.10 [-0.54, 0.74] Songisepp 2012: study 2 -0.1 0.71 18 -0.3 0.77 18 1.9% 0.20 [-0.28, 0.68] Usinger 2010 -0.1 0.86 60 0 0.65 30 4.5% -0.10 [-0.42, 0.22] Xiao 2003 -0.13 1.85 16 -0.07 1.91 16 0.3% -0.06 [-1.36, 1.24]

Total (95% CI) 564 433 100.0% -0.08 [-0.14, -0.01]

Heterogeneity: Chi² = 11.10, df = 16 (P = 0.80); I² = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 2.25 (P = 0.02) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 91

Chapter 9, Figure 20 (analysis 2.2): Forest plot of comparison: 2 Probiotic versus placebo - sensitivity analysis excluding unpublished studies, outcome: 2.2 Low density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.031 0.47 46 0.11 0.41 14 6.8% -0.14 [-0.40, 0.11] Anderson 1999 -0.1 0.64 40 0.038 0.57 40 6.2% -0.14 [-0.40, 0.13] Ataie-Jafari 2009 -0.24 0.52 14 0.26 0.98 14 1.3% -0.50 [-1.08, 0.08] de-Roos 1999 -0.02 0.64 39 -0.08 0.75 39 4.6% 0.06 [-0.25, 0.37] Ejtahed 2011 -0.25 0.56 30 -0.03 0.59 30 5.2% -0.22 [-0.51, 0.07] Fabian 2006 0 0.56 17 0 0.52 16 3.2% 0.00 [-0.37, 0.37] Fuentes 2013 -0.63 0.46 30 -0.35 0.46 30 8.1% -0.28 [-0.51, -0.05] Hlivak 2005 -0.76 1.03 20 -0.39 0.83 18 1.2% -0.37 [-0.96, 0.22] Jauhiainen 2005 -0.06 0.51 50 -0.06 0.55 51 10.2% 0.00 [-0.21, 0.21] Jones 2012 -0.33 0.78 56 0.06 0.69 58 6.0% -0.39 [-0.66, -0.12] Jones 2012a -0.3 0.57 66 0.21 0.63 61 10.0% -0.51 [-0.72, -0.30] Kekkonen 2008 0.1 0.3 11 0.1 0.5 15 4.6% 0.00 [-0.31, 0.31] Sadrzadeh-Yeganeh 2010 -0.06 0.29 29 -0.03 0.24 30 23.6% -0.03 [-0.17, 0.11] Simons 2006 -0.3 0.86 23 -0.3 0.95 21 1.5% 0.00 [-0.54, 0.54] Songisepp 2012: study 1 0.1 0.93 12 0 0.65 12 1.1% 0.10 [-0.54, 0.74] Songisepp 2012: study 2 -0.1 0.71 18 -0.3 0.77 18 1.9% 0.20 [-0.28, 0.68] Usinger 2010 -0.1 0.86 60 0 0.65 30 4.3% -0.10 [-0.42, 0.22] Xiao 2003 -0.13 1.85 16 -0.07 1.91 16 0.3% -0.06 [-1.36, 1.24]

Total (95% CI) 577 513 100.0% -0.14 [-0.21, -0.08]

Heterogeneity: Chi² = 28.98, df = 17 (P = 0.03); I² = 41% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.19 (P < 0.0001) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 92

Chapter 9, Figure 22 (analysis 4.2): Forest plot of comparison: 4 Probiotic versus placebo - sensitivity analysis based on number of participants included in analysis, outcome: 4.2 Low density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 4.2.2 Small sample size (< 100 participants)

Agerholm-Larsen 2000 -0.031 0.47 46 0.11 0.41 14 6.0% -0.14 [-0.40, 0.11] Anderson 1999 -0.1 0.64 40 0.038 0.57 40 5.5% -0.14 [-0.40, 0.13] Ataie-Jafari 2009 -0.24 0.52 14 0.26 0.98 14 1.1% -0.50 [-1.08, 0.08] de-Roos 1999 -0.02 0.64 39 -0.08 0.75 39 4.1% 0.06 [-0.25, 0.37] Ejtahed 2011 -0.25 0.56 30 -0.03 0.59 30 4.6% -0.22 [-0.51, 0.07] Fabian 2006 0 0.56 17 0 0.52 16 2.9% 0.00 [-0.37, 0.37] Fuentes 2013 -0.63 0.46 30 -0.35 0.46 30 7.2% -0.28 [-0.51, -0.05] Hlivak 2005 -0.76 1.03 20 -0.39 0.83 18 1.1% -0.37 [-0.96, 0.22] Kekkonen 2008 0.1 0.3 11 0.1 0.5 15 4.1% 0.00 [-0.31, 0.31] Sadrzadeh-Yeganeh 2010 -0.06 0.29 29 -0.03 0.24 30 21.0% -0.03 [-0.17, 0.11] Simons 2006 -0.3 0.86 23 -0.3 0.95 21 1.3% 0.00 [-0.54, 0.54] Songisepp 2012: study 1 0.1 0.93 12 0 0.65 12 0.9% 0.10 [-0.54, 0.74] Songisepp 2012: study 2 -0.1 0.71 18 -0.3 0.77 18 1.7% 0.20 [-0.28, 0.68] Usinger 2010 -0.1 0.86 60 0 0.65 30 3.8% -0.10 [-0.42, 0.22] Xiao 2003 -0.13 1.85 16 -0.07 1.91 16 0.2% -0.06 [-1.36, 1.24] Subtotal (95% CI) 405 343 65.5% -0.09 [-0.17, -0.01]

Heterogeneity: Chi² = 10.41, df = 14 (P = 0.73); I² = 0% Test for overall effect: Z = 2.32 (P = 0.02)

4.2.3 Large sample size (> 100 participants)

Ivey 2013 -0.06 0.4 109 0 0.54 39 11.3% -0.06 [-0.25, 0.13] Jauhiainen 2005 -0.06 0.51 50 -0.06 0.55 51 9.1% 0.00 [-0.21, 0.21] Jones 2012 -0.33 0.78 56 0.06 0.69 58 5.3% -0.39 [-0.66, -0.12] Jones 2012a -0.3 0.57 66 0.21 0.63 61 8.8% -0.51 [-0.72, -0.30] Subtotal (95% CI) 281 209 34.5% -0.21 [-0.32, -0.10]

Heterogeneity: Chi² = 16.05, df = 3 (P = 0.001); I² = 81% Test for overall effect: Z = 3.88 (P = 0.0001)

Total (95% CI) 686 552 100.0% -0.13 [-0.19, -0.07]

Heterogeneity: Chi² = 29.63, df = 18 (P = 0.04); I² = 39% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.16 (P < 0.0001) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 3.17, df = 1 (P = 0.07), I² = 68.5%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 93

Chapter 9, Figure 23 (analysis 5.2): Forest plot of comparison: 5 Probiotic versus placebo - sensitivity analysis using Random Effects model, outcome: 5.2 Low density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI

Agerholm-Larsen 2000 -0.031 0.47 46 0.11 0.41 14 6.7% -0.14 [-0.40, 0.11] Anderson 1999 -0.1 0.64 40 0.038 0.57 40 6.4% -0.14 [-0.40, 0.13] Ataie-Jafari 2009 -0.24 0.52 14 0.26 0.98 14 2.0% -0.50 [-1.08, 0.08] de-Roos 1999 -0.02 0.64 39 -0.08 0.75 39 5.3% 0.06 [-0.25, 0.37] Ejtahed 2011 -0.25 0.56 30 -0.03 0.59 30 5.7% -0.22 [-0.51, 0.07] Fabian 2006 0 0.56 17 0 0.52 16 4.1% 0.00 [-0.37, 0.37] Fuentes 2013 -0.63 0.46 30 -0.35 0.46 30 7.4% -0.28 [-0.51, -0.05] Hlivak 2005 -0.76 1.03 20 -0.39 0.83 18 1.9% -0.37 [-0.96, 0.22] Ivey 2013 -0.06 0.4 109 0 0.54 39 9.1% -0.06 [-0.25, 0.13] Jauhiainen 2005 -0.06 0.51 50 -0.06 0.55 51 8.3% 0.00 [-0.21, 0.21] Jones 2012 -0.33 0.78 56 0.06 0.69 58 6.2% -0.39 [-0.66, -0.12] Jones 2012a -0.3 0.57 66 0.21 0.63 61 8.2% -0.51 [-0.72, -0.30] Kekkonen 2008 0.1 0.3 11 0.1 0.5 15 5.3% 0.00 [-0.31, 0.31] Sadrzadeh-Yeganeh 2010 -0.06 0.29 29 -0.03 0.24 30 11.2% -0.03 [-0.17, 0.11] Simons 2006 -0.3 0.86 23 -0.3 0.95 21 2.3% 0.00 [-0.54, 0.54] Songisepp 2012: study 1 0.1 0.93 12 0 0.65 12 1.7% 0.10 [-0.54, 0.74] Songisepp 2012: study 2 -0.1 0.71 18 -0.3 0.77 18 2.7% 0.20 [-0.28, 0.68] Usinger 2010 -0.1 0.86 60 0 0.65 30 5.1% -0.10 [-0.42, 0.22] Xiao 2003 -0.13 1.85 16 -0.07 1.91 16 0.4% -0.06 [-1.36, 1.24]

Total (95% CI) 686 552 100.0% -0.14 [-0.23, -0.05]

Heterogeneity: Tau² = 0.01; Chi² = 29.63, df = 18 (P = 0.04); I² = 39% -1 -0.5 0 0.5 1 Test for overall effect: Z = 3.10 (P = 0.002) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 94

Chapter 9, Figure 24 (analysis 6.2): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.2 Low density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 6.2.2 Capsules

Agerholm-Larsen 2000 -0.031 0.47 46 0.11 0.41 14 6.1% -0.14 [-0.40, 0.11] Fuentes 2013 -0.63 0.46 30 -0.35 0.46 30 7.2% -0.28 [-0.51, -0.05] Hlivak 2005 -0.76 1.03 20 -0.39 0.83 18 1.1% -0.37 [-0.96, 0.22] Ivey 2013 -0.01 0.35 39 0 0.54 19 5.5% -0.01 [-0.28, 0.26] Jones 2012a -0.3 0.57 66 0.21 0.63 61 8.9% -0.51 [-0.72, -0.30] Simons 2006 -0.3 0.86 23 -0.3 0.95 21 1.4% 0.00 [-0.54, 0.54] Subtotal (95% CI) 224 163 30.2% -0.26 [-0.38, -0.15]

Heterogeneity: Chi² = 10.75, df = 5 (P = 0.06); I² = 53% Test for overall effect: Z = 4.50 (P < 0.00001)

6.2.3 Fermented milk

Anderson 1999 -0.1 0.64 40 0.038 0.57 40 5.5% -0.14 [-0.40, 0.13] Ataie-Jafari 2009 -0.24 0.52 14 0.26 0.98 14 1.2% -0.50 [-1.08, 0.08] de-Roos 1999 -0.02 0.64 39 -0.08 0.75 39 4.1% 0.06 [-0.25, 0.37] Ejtahed 2011 -0.25 0.56 30 -0.03 0.59 30 4.6% -0.22 [-0.51, 0.07] Fabian 2006 0 0.56 17 0 0.52 16 2.9% 0.00 [-0.37, 0.37] Ivey 2013 -0.03 0.41 32 0 0.54 20 5.1% -0.03 [-0.31, 0.25] Jauhiainen 2005 -0.06 0.51 50 -0.06 0.55 51 9.1% 0.00 [-0.21, 0.21] Jones 2012 -0.33 0.78 56 0.06 0.69 58 5.3% -0.39 [-0.66, -0.12] Kekkonen 2008 0.1 0.3 11 0.1 0.5 15 4.1% 0.00 [-0.31, 0.31] Sadrzadeh-Yeganeh 2010 -0.06 0.29 29 -0.03 0.24 30 21.1% -0.03 [-0.17, 0.11] Songisepp 2012: study 1 0.1 0.93 12 0 0.65 12 0.9% 0.10 [-0.54, 0.74] Songisepp 2012: study 2 -0.1 0.71 18 -0.3 0.77 18 1.7% 0.20 [-0.28, 0.68] Usinger 2010 -0.1 0.86 60 0 0.65 30 3.9% -0.10 [-0.42, 0.22] Xiao 2003 -0.13 1.85 16 -0.07 1.91 16 0.2% -0.06 [-1.36, 1.24] Subtotal (95% CI) 424 389 69.8% -0.07 [-0.15, 0.00]

Heterogeneity: Chi² = 12.11, df = 13 (P = 0.52); I² = 0% Test for overall effect: Z = 1.86 (P = 0.06)

Total (95% CI) 648 552 100.0% -0.13 [-0.19, -0.07]

Heterogeneity: Chi² = 30.37, df = 19 (P = 0.05); I² = 37% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.02 (P < 0.0001) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 7.51, df = 1 (P = 0.006), I² = 86.7%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 95

Chapter 9, Figure 25 (analysis 7.3): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.3 Low density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 7.3.2 Desirable baseline total-cholesterol

Agerholm-Larsen 2000 -0.031 0.47 46 0.11 0.41 14 6.6% -0.14 [-0.40, 0.11] de-Roos 1999 -0.02 0.64 39 -0.08 0.75 39 4.5% 0.06 [-0.25, 0.37] Ejtahed 2011 -0.25 0.56 30 -0.03 0.59 30 5.0% -0.22 [-0.51, 0.07] Fabian 2006 0 0.56 17 0 0.52 16 3.1% 0.00 [-0.37, 0.37] Sadrzadeh-Yeganeh 2010 -0.06 0.29 29 -0.03 0.24 30 23.1% -0.03 [-0.17, 0.11] Songisepp 2012: study 1 0.1 0.93 12 0 0.65 12 1.0% 0.10 [-0.54, 0.74] Usinger 2010 -0.1 0.86 60 0 0.65 30 4.2% -0.10 [-0.42, 0.22] Subtotal (95% CI) 233 171 47.6% -0.06 [-0.15, 0.04]

Heterogeneity: Chi² = 2.72, df = 6 (P = 0.84); I² = 0% Test for overall effect: Z = 1.21 (P = 0.23)

7.3.3 Borderline high baseline total-cholesterol

Ataie-Jafari 2009 -0.24 0.52 14 0.26 0.98 14 1.3% -0.50 [-1.08, 0.08] Hlivak 2005 -0.76 1.03 20 -0.39 0.83 18 1.2% -0.37 [-0.96, 0.22] Ivey 2013 -0.06 0.4 109 0 0.54 39 12.4% -0.06 [-0.25, 0.13] Jones 2012a -0.3 0.57 66 0.21 0.63 61 9.7% -0.51 [-0.72, -0.30] Kekkonen 2008 0.1 0.3 11 0.1 0.5 15 4.5% 0.00 [-0.31, 0.31] Songisepp 2012: study 2 -0.1 0.71 18 -0.3 0.77 18 1.8% 0.20 [-0.28, 0.68] Subtotal (95% CI) 238 165 30.9% -0.21 [-0.33, -0.09]

Heterogeneity: Chi² = 16.16, df = 5 (P = 0.006); I² = 69% Test for overall effect: Z = 3.46 (P = 0.0005)

7.3.4 High baseline total-cholesterol

Anderson 1999 -0.1 0.64 40 0.038 0.57 40 6.1% -0.14 [-0.40, 0.13] Fuentes 2013 -0.63 0.46 30 -0.35 0.46 30 7.9% -0.28 [-0.51, -0.05] Jones 2012 -0.33 0.78 56 0.06 0.69 58 5.8% -0.39 [-0.66, -0.12] Simons 2006 -0.3 0.86 23 -0.3 0.95 21 1.5% 0.00 [-0.54, 0.54] Xiao 2003 -0.13 1.85 16 -0.07 1.91 16 0.3% -0.06 [-1.36, 1.24] Subtotal (95% CI) 165 165 21.5% -0.25 [-0.39, -0.11]

Heterogeneity: Chi² = 2.69, df = 4 (P = 0.61); I² = 0% Test for overall effect: Z = 3.45 (P = 0.0006)

Total (95% CI) 636 501 100.0% -0.15 [-0.21, -0.08]

Heterogeneity: Chi² = 27.91, df = 17 (P = 0.05); I² = 39% -1 -0.5 0 0.5 1 Test for overall effect: Z = 4.36 (P < 0.0001) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 6.34, df = 2 (P = 0.04), I² = 68.5%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 96

Chapter 9, Figure 26 (analysis 7.4): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.4 Low density lipoprotein cholesterol - Borderline high studies split into two groups (above and below the median baseline total-cholesterol).

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 7.4.1 Borderline-high total cholesterol - below the median

Ivey 2013 -0.06 0.4 109 0 0.54 39 40.2% -0.06 [-0.25, 0.13] Kekkonen 2008 0.1 0.3 11 0.1 0.5 15 14.5% 0.00 [-0.31, 0.31] Subtotal (95% CI) 120 54 54.6% -0.04 [-0.20, 0.11]

Heterogeneity: Chi² = 0.11, df = 1 (P = 0.74); I² = 0% Test for overall effect: Z = 0.54 (P = 0.59)

7.4.2 Borderline-high total cholesterol - above the median

Ataie-Jafari 2009 -0.24 0.52 14 0.26 0.98 14 4.1% -0.50 [-1.08, 0.08] Hlivak 2005 -0.76 1.03 20 -0.39 0.83 18 3.9% -0.37 [-0.96, 0.22] Jones 2012a -0.3 0.57 66 0.21 0.63 61 31.4% -0.51 [-0.72, -0.30] Songisepp 2012: study 2 -0.1 0.71 18 -0.3 0.77 18 5.9% 0.20 [-0.28, 0.68] Subtotal (95% CI) 118 111 45.4% -0.40 [-0.58, -0.23]

Heterogeneity: Chi² = 7.09, df = 3 (P = 0.07); I² = 58% Test for overall effect: Z = 4.55 (P < 0.00001)

Total (95% CI) 238 165 100.0% -0.21 [-0.33, -0.09]

Heterogeneity: Chi² = 16.16, df = 5 (P = 0.006); I² = 69% -1 -0.5 0 0.5 1 Test for overall effect: Z = 3.46 (P = 0.0005) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 8.97, df = 1 (P = 0.003), I² = 88.8%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 97

Chapter 9, Figure 27 (analysis 1.5): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.5 High density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.033 0.15 46 -0.06 0.34 14 1.7% 0.03 [-0.16, 0.21] Anderson 1999 0 0.24 40 0 0.24 40 5.0% 0.00 [-0.11, 0.11] Ataie-Jafari 2009 0.01 0.14 14 0.15 0.2 14 3.4% -0.14 [-0.27, -0.01] de-Roos 1999 0 0.21 39 0.03 0.25 39 5.3% -0.03 [-0.13, 0.07] Ejtahed 2011 0 0.17 30 -0.07 0.17 30 7.5% 0.07 [-0.02, 0.16] Fabian 2006 0.1 0.24 17 0 0.19 16 2.6% 0.10 [-0.05, 0.25] Fuentes 2013 0.07 0.12 30 -0.04 0.16 30 10.9% 0.11 [0.04, 0.18] Hata 1996 -0.14 0.29 17 0.09 0.4 13 0.8% -0.23 [-0.49, 0.03] Hlivak 2005 0.05 0.3 20 0.06 0.25 18 1.8% -0.01 [-0.18, 0.16] Ivey 2013 -0.02 0.15 109 0.02 0.14 39 20.4% -0.04 [-0.09, 0.01] Jones 2012 -0.05 0.34 56 -0.05 0.31 58 3.9% 0.00 [-0.12, 0.12] Jones 2012a 0 0.24 66 0.01 0.29 61 6.4% -0.01 [-0.10, 0.08] Kekkonen 2008 0 0.2 11 0.1 0.2 15 2.3% -0.10 [-0.26, 0.06] Mizushima 2004 -0.06 0.28 23 0.08 0.23 23 2.5% -0.14 [-0.29, 0.01] Rizkalla 2000 -0.1 0.21 24 -0.1 0.21 24 3.9% 0.00 [-0.12, 0.12] Sadrzadeh-Yeganeh 2010 0.11 0.15 29 0.07 0.17 30 8.3% 0.04 [-0.04, 0.12] Simons 2006 -0.1 0.24 23 0 0.3 21 2.1% -0.10 [-0.26, 0.06] Songisepp 2012: study 1 0 0.31 12 0.1 0.25 12 1.1% -0.10 [-0.33, 0.13] Songisepp 2012: study 2 -0.1 0.25 18 0 0.3 18 1.7% -0.10 [-0.28, 0.08] Usinger 2010 0 0.19 60 0 0.19 30 8.0% 0.00 [-0.08, 0.08] Xiao 2003 -0.07 0.92 16 -0.07 1 16 0.1% 0.00 [-0.67, 0.67]

Total (95% CI) 700 561 100.0% -0.00 [-0.03, 0.02]

Heterogeneity: Chi² = 32.99, df = 20 (P = 0.03); I² = 39% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.38 (P = 0.71) Favours placebo Favours probiotic

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 98

Chapter 9, Figure 28 (analysis 1.5): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.5 High density lipoprotein cholesterol.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 99

Chapter 9, Figure 29 (analysis 1.6): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.6 High density lipoprotein cholesterol - excluding outliers.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.033 0.15 46 -0.06 0.34 14 1.9% 0.03 [-0.16, 0.21] Anderson 1999 0 0.24 40 0 0.24 40 5.6% 0.00 [-0.11, 0.11] Ataie-Jafari 2009 0.01 0.14 14 0.15 0.2 14 3.8% -0.14 [-0.27, -0.01] de-Roos 1999 0 0.21 39 0.03 0.25 39 5.9% -0.03 [-0.13, 0.07] Ejtahed 2011 0 0.17 30 -0.07 0.17 30 8.4% 0.07 [-0.02, 0.16] Fabian 2006 0.1 0.24 17 0 0.19 16 2.9% 0.10 [-0.05, 0.25] Hata 1996 -0.14 0.29 17 0.09 0.4 13 0.9% -0.23 [-0.49, 0.03] Hlivak 2005 0.05 0.3 20 0.06 0.25 18 2.0% -0.01 [-0.18, 0.16] Ivey 2013 -0.02 0.15 109 0.02 0.14 39 22.9% -0.04 [-0.09, 0.01] Jones 2012 -0.05 0.34 56 -0.05 0.31 58 4.4% 0.00 [-0.12, 0.12] Jones 2012a 0 0.24 66 0.01 0.29 61 7.2% -0.01 [-0.10, 0.08] Kekkonen 2008 0 0.2 11 0.1 0.2 15 2.6% -0.10 [-0.26, 0.06] Mizushima 2004 -0.06 0.28 23 0.08 0.23 23 2.8% -0.14 [-0.29, 0.01] Rizkalla 2000 -0.1 0.21 24 -0.1 0.21 24 4.4% 0.00 [-0.12, 0.12] Sadrzadeh-Yeganeh 2010 0.11 0.15 29 0.07 0.17 30 9.3% 0.04 [-0.04, 0.12] Simons 2006 -0.1 0.24 23 0 0.3 21 2.4% -0.10 [-0.26, 0.06] Songisepp 2012: study 1 0 0.31 12 0.1 0.25 12 1.2% -0.10 [-0.33, 0.13] Songisepp 2012: study 2 -0.1 0.25 18 0 0.3 18 1.9% -0.10 [-0.28, 0.08] Usinger 2010 0 0.19 60 0 0.19 30 9.0% 0.00 [-0.08, 0.08] Xiao 2003 -0.07 0.92 16 -0.07 1 16 0.1% 0.00 [-0.67, 0.67]

Total (95% CI) 670 531 100.0% -0.02 [-0.04, 0.01]

Heterogeneity: Chi² = 21.96, df = 19 (P = 0.29); I² = 13% -1 -0.5 0 0.5 1 Test for overall effect: Z = 1.45 (P = 0.15) Favours placebo Favours probiotic

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 100

Chapter 9, Figure 30 (analysis 1.6): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.6 High density lipoprotein cholesterol - excluding outliers.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 101

Chapter 9, Figure 31 (analysis 2.3): Forest plot of comparison: 2 Probiotic versus placebo - sensitivity analysis excluding unpublished studies, outcome: 2.3 High density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 -0.033 0.15 46 -0.06 0.34 14 2.1% 0.03 [-0.16, 0.21] Anderson 1999 0 0.24 40 0 0.24 40 6.3% 0.00 [-0.11, 0.11] Ataie-Jafari 2009 0.01 0.14 14 0.15 0.2 14 4.3% -0.14 [-0.27, -0.01] de-Roos 1999 0 0.21 39 0.03 0.25 39 6.7% -0.03 [-0.13, 0.07] Ejtahed 2011 0 0.17 30 -0.07 0.17 30 9.5% 0.07 [-0.02, 0.16] Fabian 2006 0.1 0.24 17 0 0.19 16 3.2% 0.10 [-0.05, 0.25] Fuentes 2013 0.07 0.12 30 -0.04 0.16 30 13.7% 0.11 [0.04, 0.18] Hata 1996 -0.14 0.29 17 0.09 0.4 13 1.1% -0.23 [-0.49, 0.03] Hlivak 2005 0.05 0.3 20 0.06 0.25 18 2.3% -0.01 [-0.18, 0.16] Jones 2012 -0.05 0.34 56 -0.05 0.31 58 4.9% 0.00 [-0.12, 0.12] Jones 2012a 0 0.24 66 0.01 0.29 61 8.1% -0.01 [-0.10, 0.08] Kekkonen 2008 0 0.2 11 0.1 0.2 15 2.9% -0.10 [-0.26, 0.06] Mizushima 2004 -0.06 0.28 23 0.08 0.23 23 3.2% -0.14 [-0.29, 0.01] Rizkalla 2000 -0.1 0.21 24 -0.1 0.21 24 5.0% 0.00 [-0.12, 0.12] Sadrzadeh-Yeganeh 2010 0.11 0.15 29 0.07 0.17 30 10.5% 0.04 [-0.04, 0.12] Simons 2006 -0.1 0.24 23 0 0.3 21 2.7% -0.10 [-0.26, 0.06] Songisepp 2012: study 1 0 0.31 12 0.1 0.25 12 1.4% -0.10 [-0.33, 0.13] Songisepp 2012: study 2 -0.1 0.25 18 0 0.3 18 2.2% -0.10 [-0.28, 0.08] Usinger 2010 0 0.19 60 0 0.19 30 10.1% 0.00 [-0.08, 0.08] Xiao 2003 -0.07 0.92 16 -0.07 1 16 0.2% 0.00 [-0.67, 0.67]

Total (95% CI) 591 522 100.0% 0.00 [-0.02, 0.03]

Heterogeneity: Chi² = 30.76, df = 19 (P = 0.04); I² = 38% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.34 (P = 0.73) Favours placebo Favours probiotic

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 102

Chapter 9, Figure 33 (analysis 4.3): Forest plot of comparison: 4 Probiotic versus placebo - sensitivity analysis based on number of participants included in analysis, outcome: 4.3 High density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 4.3.2 Small sample size (< 100 participants)

Agerholm-Larsen 2000 -0.033 0.15 46 -0.06 0.34 14 1.7% 0.03 [-0.16, 0.21] Anderson 1999 0 0.24 40 0 0.24 40 5.0% 0.00 [-0.11, 0.11] Ataie-Jafari 2009 0.01 0.14 14 0.15 0.2 14 3.4% -0.14 [-0.27, -0.01] de-Roos 1999 0 0.21 39 0.03 0.25 39 5.3% -0.03 [-0.13, 0.07] Ejtahed 2011 0 0.17 30 -0.07 0.17 30 7.5% 0.07 [-0.02, 0.16] Fabian 2006 0.1 0.24 17 0 0.19 16 2.6% 0.10 [-0.05, 0.25] Fuentes 2013 0.07 0.12 30 -0.04 0.16 30 10.9% 0.11 [0.04, 0.18] Hata 1996 -0.14 0.29 17 0.09 0.4 13 0.8% -0.23 [-0.49, 0.03] Hlivak 2005 0.05 0.3 20 0.06 0.25 18 1.8% -0.01 [-0.18, 0.16] Kekkonen 2008 0 0.2 11 0.1 0.2 15 2.3% -0.10 [-0.26, 0.06] Mizushima 2004 -0.06 0.28 23 0.08 0.23 23 2.5% -0.14 [-0.29, 0.01] Rizkalla 2000 -0.1 0.21 24 -0.1 0.21 24 3.9% 0.00 [-0.12, 0.12] Sadrzadeh-Yeganeh 2010 0.11 0.15 29 0.07 0.17 30 8.3% 0.04 [-0.04, 0.12] Simons 2006 -0.1 0.24 23 0 0.3 21 2.1% -0.10 [-0.26, 0.06] Songisepp 2012: study 1 0 0.31 12 0.1 0.25 12 1.1% -0.10 [-0.33, 0.13] Songisepp 2012: study 2 -0.1 0.25 18 0 0.3 18 1.7% -0.10 [-0.28, 0.08] Usinger 2010 0 0.19 60 0 0.19 30 8.0% 0.00 [-0.08, 0.08] Xiao 2003 -0.07 0.92 16 -0.07 1 16 0.1% 0.00 [-0.67, 0.67] Subtotal (95% CI) 469 403 69.2% 0.01 [-0.02, 0.03]

Heterogeneity: Chi² = 30.65, df = 17 (P = 0.02); I² = 45% Test for overall effect: Z = 0.43 (P = 0.67)

4.3.3 LArge sample size (> 100 participants)

Ivey 2013 -0.02 0.15 109 0.02 0.14 39 20.4% -0.04 [-0.09, 0.01] Jones 2012 -0.05 0.34 56 -0.05 0.31 58 3.9% 0.00 [-0.12, 0.12] Jones 2012a 0 0.24 66 0.01 0.29 61 6.4% -0.01 [-0.10, 0.08] Subtotal (95% CI) 231 158 30.8% -0.03 [-0.07, 0.01]

Heterogeneity: Chi² = 0.56, df = 2 (P = 0.76); I² = 0% Test for overall effect: Z = 1.32 (P = 0.19)

Total (95% CI) 700 561 100.0% -0.00 [-0.03, 0.02]

Heterogeneity: Chi² = 32.99, df = 20 (P = 0.03); I² = 39% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.38 (P = 0.71) Favours placebo Favours probiotic Test for subgroup differences: Chi² = 1.79, df = 1 (P = 0.18), I² = 44.1%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 103

Chapter 9, Figure 34 (analysis 5.3): Forest plot of comparison: 5 Probiotic versus placebo - sensitivity analysis using Random Effects model, outcome: 5.3 High density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI

Agerholm-Larsen 2000 -0.033 0.15 46 -0.06 0.34 14 2.7% 0.03 [-0.16, 0.21] Anderson 1999 0 0.24 40 0 0.24 40 5.8% 0.00 [-0.11, 0.11] Ataie-Jafari 2009 0.01 0.14 14 0.15 0.2 14 4.5% -0.14 [-0.27, -0.01] de-Roos 1999 0 0.21 39 0.03 0.25 39 6.0% -0.03 [-0.13, 0.07] Ejtahed 2011 0 0.17 30 -0.07 0.17 30 7.2% 0.07 [-0.02, 0.16] Fabian 2006 0.1 0.24 17 0 0.19 16 3.7% 0.10 [-0.05, 0.25] Fuentes 2013 0.07 0.12 30 -0.04 0.16 30 8.5% 0.11 [0.04, 0.18] Hata 1996 -0.14 0.29 17 0.09 0.4 13 1.5% -0.23 [-0.49, 0.03] Hlivak 2005 0.05 0.3 20 0.06 0.25 18 2.9% -0.01 [-0.18, 0.16] Ivey 2013 -0.02 0.15 109 0.02 0.14 39 10.4% -0.04 [-0.09, 0.01] Jones 2012 -0.05 0.34 56 -0.05 0.31 58 5.0% 0.00 [-0.12, 0.12] Jones 2012a 0 0.24 66 0.01 0.29 61 6.7% -0.01 [-0.10, 0.08] Kekkonen 2008 0 0.2 11 0.1 0.2 15 3.4% -0.10 [-0.26, 0.06] Mizushima 2004 -0.06 0.28 23 0.08 0.23 23 3.7% -0.14 [-0.29, 0.01] Rizkalla 2000 -0.1 0.21 24 -0.1 0.21 24 5.0% 0.00 [-0.12, 0.12] Sadrzadeh-Yeganeh 2010 0.11 0.15 29 0.07 0.17 30 7.6% 0.04 [-0.04, 0.12] Simons 2006 -0.1 0.24 23 0 0.3 21 3.2% -0.10 [-0.26, 0.06] Songisepp 2012: study 1 0 0.31 12 0.1 0.25 12 1.9% -0.10 [-0.33, 0.13] Songisepp 2012: study 2 -0.1 0.25 18 0 0.3 18 2.7% -0.10 [-0.28, 0.08] Usinger 2010 0 0.19 60 0 0.19 30 7.4% 0.00 [-0.08, 0.08] Xiao 2003 -0.07 0.92 16 -0.07 1 16 0.2% 0.00 [-0.67, 0.67]

Total (95% CI) 700 561 100.0% -0.01 [-0.04, 0.02]

Heterogeneity: Tau² = 0.00; Chi² = 32.99, df = 20 (P = 0.03); I² = 39% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.67 (P = 0.50) Favours placebo Favours probiotic

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 104

Chapter 9, Figure 35 (analysis 6.3): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.3 High density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 6.3.2 Capsules

Agerholm-Larsen 2000 -0.033 0.15 46 -0.06 0.34 14 1.7% 0.03 [-0.16, 0.21] Fuentes 2013 0.07 0.12 30 -0.04 0.16 30 11.1% 0.11 [0.04, 0.18] Hlivak 2005 0.05 0.3 20 0.06 0.25 18 1.9% -0.01 [-0.18, 0.16] Ivey 2013 0 0.14 39 0.02 0.14 19 9.6% -0.02 [-0.10, 0.06] Jones 2012a 0 0.24 66 0.01 0.29 61 6.6% -0.01 [-0.10, 0.08] Simons 2006 -0.1 0.24 23 0 0.3 21 2.2% -0.10 [-0.26, 0.06] Subtotal (95% CI) 224 163 33.0% 0.02 [-0.02, 0.06]

Heterogeneity: Chi² = 9.73, df = 5 (P = 0.08); I² = 49% Test for overall effect: Z = 1.10 (P = 0.27)

6.3.3 Fermented milk

Anderson 1999 0 0.24 40 0 0.24 40 5.1% 0.00 [-0.11, 0.11] Ataie-Jafari 2009 0.01 0.14 14 0.15 0.2 14 3.5% -0.14 [-0.27, -0.01] de-Roos 1999 0 0.21 39 0.03 0.25 39 5.4% -0.03 [-0.13, 0.07] Ejtahed 2011 0 0.17 30 -0.07 0.17 30 7.7% 0.07 [-0.02, 0.16] Fabian 2006 0.1 0.24 17 0 0.19 16 2.6% 0.10 [-0.05, 0.25] Hata 1996 -0.14 0.29 17 0.09 0.4 13 0.9% -0.23 [-0.49, 0.03] Ivey 2013 0 0.14 32 0.02 0.14 20 9.3% -0.02 [-0.10, 0.06] Jones 2012 -0.05 0.34 56 -0.05 0.31 58 4.0% 0.00 [-0.12, 0.12] Kekkonen 2008 0 0.2 11 0.1 0.2 15 2.3% -0.10 [-0.26, 0.06] Mizushima 2004 -0.06 0.28 23 0.08 0.23 23 2.6% -0.14 [-0.29, 0.01] Rizkalla 2000 -0.1 0.21 24 -0.1 0.21 24 4.0% 0.00 [-0.12, 0.12] Sadrzadeh-Yeganeh 2010 0.11 0.15 29 0.07 0.17 30 8.5% 0.04 [-0.04, 0.12] Songisepp 2012: study 1 0 0.31 12 0.1 0.25 12 1.1% -0.10 [-0.33, 0.13] Songisepp 2012: study 2 -0.1 0.25 18 0 0.3 18 1.7% -0.10 [-0.28, 0.08] Usinger 2010 0 0.19 60 0 0.19 30 8.2% 0.00 [-0.08, 0.08] Xiao 2003 -0.07 0.92 16 -0.07 1 16 0.1% 0.00 [-0.67, 0.67] Subtotal (95% CI) 438 398 67.0% -0.01 [-0.04, 0.02]

Heterogeneity: Chi² = 19.84, df = 15 (P = 0.18); I² = 24% Test for overall effect: Z = 0.78 (P = 0.44)

Total (95% CI) 662 561 100.0% -0.00 [-0.02, 0.02]

Heterogeneity: Chi² = 31.39, df = 21 (P = 0.07); I² = 33% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.00 (P = 1.00) Favours placebo Favours probiotic Test for subgroup differences: Chi² = 1.82, df = 1 (P = 0.18), I² = 45.2%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 105

Chapter 9, Figure 36 (analysis 7.5): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.5 High density lipoprotein cholesterol.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 7.5.2 Desirable baseline total-cholesterol

Agerholm-Larsen 2000 -0.033 0.15 46 -0.06 0.34 14 1.7% 0.03 [-0.16, 0.21] de-Roos 1999 0 0.21 39 0.03 0.25 39 5.3% -0.03 [-0.13, 0.07] Ejtahed 2011 0 0.17 30 -0.07 0.17 30 7.5% 0.07 [-0.02, 0.16] Fabian 2006 0.1 0.24 17 0 0.19 16 2.6% 0.10 [-0.05, 0.25] Rizkalla 2000 -0.1 0.21 24 -0.1 0.21 24 3.9% 0.00 [-0.12, 0.12] Sadrzadeh-Yeganeh 2010 0.11 0.15 29 0.07 0.17 30 8.3% 0.04 [-0.04, 0.12] Songisepp 2012: study 1 0 0.31 12 0.1 0.25 12 1.1% -0.10 [-0.33, 0.13] Usinger 2010 0 0.19 60 0 0.19 30 8.0% 0.00 [-0.08, 0.08] Subtotal (95% CI) 257 195 38.5% 0.02 [-0.01, 0.06]

Heterogeneity: Chi² = 4.97, df = 7 (P = 0.66); I² = 0% Test for overall effect: Z = 1.20 (P = 0.23)

7.5.3 Borderline high baseline total-cholesterol

Ataie-Jafari 2009 0.01 0.14 14 0.15 0.2 14 3.4% -0.14 [-0.27, -0.01] Hata 1996 -0.14 0.29 17 0.09 0.4 13 0.8% -0.23 [-0.49, 0.03] Hlivak 2005 0.05 0.3 20 0.06 0.25 18 1.8% -0.01 [-0.18, 0.16] Ivey 2013 -0.02 0.15 109 0.02 0.14 39 20.4% -0.04 [-0.09, 0.01] Jones 2012a 0 0.24 66 0.01 0.29 61 6.4% -0.01 [-0.10, 0.08] Kekkonen 2008 0 0.2 11 0.1 0.2 15 2.3% -0.10 [-0.26, 0.06] Mizushima 2004 -0.06 0.28 23 0.08 0.23 23 2.5% -0.14 [-0.29, 0.01] Songisepp 2012: study 2 -0.1 0.25 18 0 0.3 18 1.7% -0.10 [-0.28, 0.08] Subtotal (95% CI) 278 201 39.5% -0.06 [-0.10, -0.02]

Heterogeneity: Chi² = 6.73, df = 7 (P = 0.46); I² = 0% Test for overall effect: Z = 3.08 (P = 0.002)

7.5.4 High baseline total-cholesterol

Anderson 1999 0 0.24 40 0 0.24 40 5.0% 0.00 [-0.11, 0.11] Fuentes 2013 0.07 0.12 30 -0.04 0.16 30 10.9% 0.11 [0.04, 0.18] Jones 2012 -0.05 0.34 56 -0.05 0.31 58 3.9% 0.00 [-0.12, 0.12] Simons 2006 -0.1 0.24 23 0 0.3 21 2.1% -0.10 [-0.26, 0.06] Xiao 2003 -0.07 0.92 16 -0.07 1 16 0.1% 0.00 [-0.67, 0.67] Subtotal (95% CI) 165 165 22.1% 0.04 [-0.01, 0.09]

Heterogeneity: Chi² = 7.53, df = 4 (P = 0.11); I² = 47% Test for overall effect: Z = 1.74 (P = 0.08)

Total (95% CI) 700 561 100.0% -0.00 [-0.03, 0.02]

Heterogeneity: Chi² = 32.99, df = 20 (P = 0.03); I² = 39% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.38 (P = 0.71) Favours placebo Favours probiotic Test for subgroup differences: Chi² = 13.76, df = 2 (P = 0.001), I² = 85.5%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 106

Chapter 9, Figure 37 (analysis 1.7): Forest plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.7 Triglyceride.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 0 0.63 46 -0.01 1.05 14 0.7% 0.01 [-0.57, 0.59] Anderson 1999 0.03 0.77 40 -0.048 0.69 40 2.2% 0.08 [-0.24, 0.40] Ataie-Jafari 2009 -0.22 0.68 14 -0.12 0.34 14 1.4% -0.10 [-0.50, 0.30] de-Roos 1999 -0.02 0.45 39 -0.05 0.38 39 6.7% 0.03 [-0.15, 0.21] Ejtahed 2011 0 0.63 30 0.09 0.39 30 3.3% -0.09 [-0.36, 0.18] Fabian 2006 -0.04 0.14 17 -0.01 0.14 16 25.0% -0.03 [-0.13, 0.07] Fuentes 2013 -0.33 0.56 30 -0.25 0.59 30 2.7% -0.08 [-0.37, 0.21] Hata 1996 -0.12 0.29 17 -0.15 0.72 13 1.3% 0.03 [-0.38, 0.44] Hlivak 2005 0.19 1 20 0.04 0.4 18 1.0% 0.15 [-0.33, 0.63] Inoue 2003 0.29 0.88 18 0.07 0.88 17 0.7% 0.22 [-0.36, 0.80] Ivey 2013 0 0.52 109 -0.04 0.5 39 6.7% 0.04 [-0.14, 0.22] Jauhiainen 2005 0.03 0.69 50 0 0.7 51 3.1% 0.03 [-0.24, 0.30] Jones 2012 0.3 1.22 56 0 0.75 58 1.6% 0.30 [-0.07, 0.67] Jones 2012a -0.02 0.71 66 0.13 1.04 61 2.3% -0.15 [-0.46, 0.16] Kekkonen 2008 0 0.6 11 0 0.5 15 1.2% 0.00 [-0.44, 0.44] Mizushima 2004 0.06 1.09 22 -0.07 1.04 22 0.6% 0.13 [-0.50, 0.76] Rizkalla 2000 0 0.24 24 0 0.38 24 7.1% 0.00 [-0.18, 0.18] Sadrzadeh-Yeganeh 2010 0 0.3 29 0.03 0.13 30 16.2% -0.03 [-0.15, 0.09] Simons 2006 -0.1 0.55 23 0 0.62 21 1.9% -0.10 [-0.45, 0.25] Songisepp 2012: study 1 0 0.41 12 0.3 0.48 12 1.8% -0.30 [-0.66, 0.06] Songisepp 2012: study 2 0 0.43 18 -0.1 0.36 18 3.4% 0.10 [-0.16, 0.36] Usinger 2010 0.075 0.31 60 -0.03 0.39 30 8.9% 0.10 [-0.06, 0.27] Xiao 2003 -0.08 1.89 16 0.03 1.91 16 0.1% -0.11 [-1.43, 1.21]

Total (95% CI) 767 628 100.0% -0.00 [-0.05, 0.05]

Heterogeneity: Chi² = 11.92, df = 22 (P = 0.96); I² = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.02 (P = 0.98) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 107

Chapter 9, Figure 38 (analysis 1.7): Funnel plot of comparison: 1 Probiotic versus placebo - primary analysis, outcome: 1.7 Triglyceride.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 108

Chapter 9, Figure 39 (analysis 2.4): Forest plot of comparison: 2 Probiotic versus placebo - sensitivity analysis excluding unpublished studies, outcome: 2.4 Triglyceride.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI

Agerholm-Larsen 2000 0 0.63 46 -0.01 1.05 14 0.7% 0.01 [-0.57, 0.59] Anderson 1999 0.03 0.77 40 -0.048 0.69 40 2.4% 0.08 [-0.24, 0.40] Ataie-Jafari 2009 -0.22 0.68 14 -0.12 0.34 14 1.5% -0.10 [-0.50, 0.30] de-Roos 1999 -0.02 0.45 39 -0.05 0.38 39 7.2% 0.03 [-0.15, 0.21] Ejtahed 2011 0 0.63 30 0.09 0.39 30 3.5% -0.09 [-0.36, 0.18] Fabian 2006 -0.04 0.14 17 -0.01 0.14 16 26.8% -0.03 [-0.13, 0.07] Fuentes 2013 -0.33 0.56 30 -0.25 0.59 30 2.9% -0.08 [-0.37, 0.21] Hata 1996 -0.12 0.29 17 -0.15 0.72 13 1.4% 0.03 [-0.38, 0.44] Hlivak 2005 0.19 1 20 0.04 0.4 18 1.1% 0.15 [-0.33, 0.63] Inoue 2003 0.29 0.88 18 0.07 0.88 17 0.7% 0.22 [-0.36, 0.80] Jauhiainen 2005 0.03 0.69 50 0 0.7 51 3.3% 0.03 [-0.24, 0.30] Jones 2012 0.3 1.22 56 0 0.75 58 1.8% 0.30 [-0.07, 0.67] Jones 2012a -0.02 0.71 66 0.13 1.04 61 2.5% -0.15 [-0.46, 0.16] Kekkonen 2008 0 0.6 11 0 0.5 15 1.3% 0.00 [-0.44, 0.44] Mizushima 2004 0.06 1.09 22 -0.07 1.04 22 0.6% 0.13 [-0.50, 0.76] Rizkalla 2000 0 0.24 24 0 0.38 24 7.6% 0.00 [-0.18, 0.18] Sadrzadeh-Yeganeh 2010 0 0.3 29 0.03 0.13 30 17.4% -0.03 [-0.15, 0.09] Simons 2006 -0.1 0.55 23 0 0.62 21 2.0% -0.10 [-0.45, 0.25] Songisepp 2012: study 1 0 0.41 12 0.3 0.48 12 1.9% -0.30 [-0.66, 0.06] Songisepp 2012: study 2 0 0.43 18 -0.1 0.36 18 3.6% 0.10 [-0.16, 0.36] Usinger 2010 0.075 0.31 60 -0.03 0.39 30 9.6% 0.10 [-0.06, 0.27] Xiao 2003 -0.08 1.89 16 0.03 1.91 16 0.1% -0.11 [-1.43, 1.21]

Total (95% CI) 658 589 100.0% -0.00 [-0.05, 0.05]

Heterogeneity: Chi² = 11.72, df = 21 (P = 0.95); I² = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.14 (P = 0.89) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 109

Chapter 9, Figure 41 (analysis 4.4): Forest plot of comparison: 4 Probiotic versus placebo - sensitivity analysis based on number of participants included in analysis, outcome: 4.4 Triglyceride.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 4.4.2 Small sample size (< 100 participants)

Agerholm-Larsen 2000 0 0.63 46 -0.01 1.05 14 0.7% 0.01 [-0.57, 0.59] Anderson 1999 0.03 0.77 40 -0.048 0.69 40 2.2% 0.08 [-0.24, 0.40] Ataie-Jafari 2009 -0.22 0.68 14 -0.12 0.34 14 1.4% -0.10 [-0.50, 0.30] de-Roos 1999 -0.02 0.45 39 -0.05 0.38 39 6.7% 0.03 [-0.15, 0.21] Ejtahed 2011 0 0.63 30 0.09 0.39 30 3.3% -0.09 [-0.36, 0.18] Fabian 2006 -0.04 0.14 17 -0.01 0.14 16 25.0% -0.03 [-0.13, 0.07] Fuentes 2013 -0.33 0.56 30 -0.25 0.59 30 2.7% -0.08 [-0.37, 0.21] Hata 1996 -0.12 0.29 17 -0.15 0.72 13 1.3% 0.03 [-0.38, 0.44] Hlivak 2005 0.19 1 20 0.04 0.4 18 1.0% 0.15 [-0.33, 0.63] Inoue 2003 0.29 0.88 18 0.07 0.88 17 0.7% 0.22 [-0.36, 0.80] Kekkonen 2008 0 0.6 11 0 0.5 15 1.2% 0.00 [-0.44, 0.44] Mizushima 2004 0.06 1.09 22 -0.07 1.04 22 0.6% 0.13 [-0.50, 0.76] Rizkalla 2000 0 0.24 24 0 0.38 24 7.1% 0.00 [-0.18, 0.18] Sadrzadeh-Yeganeh 2010 0 0.3 29 0.03 0.13 30 16.2% -0.03 [-0.15, 0.09] Simons 2006 -0.1 0.55 23 0 0.62 21 1.9% -0.10 [-0.45, 0.25] Songisepp 2012: study 1 0 0.41 12 0.3 0.48 12 1.8% -0.30 [-0.66, 0.06] Songisepp 2012: study 2 0 0.43 18 -0.1 0.36 18 3.4% 0.10 [-0.16, 0.36] Usinger 2010 0.075 0.31 60 -0.03 0.39 30 8.9% 0.10 [-0.06, 0.27] Xiao 2003 -0.08 1.89 16 0.03 1.91 16 0.1% -0.11 [-1.43, 1.21] Subtotal (95% CI) 486 419 86.2% -0.01 [-0.06, 0.05]

Heterogeneity: Chi² = 8.26, df = 18 (P = 0.97); I² = 0% Test for overall effect: Z = 0.25 (P = 0.81)

4.4.3 Large sample size (> 100 participants)

Ivey 2013 0 0.52 109 -0.04 0.5 39 6.7% 0.04 [-0.14, 0.22] Jauhiainen 2005 0.03 0.69 50 0 0.7 51 3.1% 0.03 [-0.24, 0.30] Jones 2012 0.3 1.22 56 0 0.75 58 1.6% 0.30 [-0.07, 0.67] Jones 2012a -0.02 0.71 66 0.13 1.04 61 2.3% -0.15 [-0.46, 0.16] Subtotal (95% CI) 281 209 13.8% 0.04 [-0.09, 0.17]

Heterogeneity: Chi² = 3.29, df = 3 (P = 0.35); I² = 9% Test for overall effect: Z = 0.55 (P = 0.58)

Total (95% CI) 767 628 100.0% -0.00 [-0.05, 0.05]

Heterogeneity: Chi² = 11.92, df = 22 (P = 0.96); I² = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.02 (P = 0.98) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 0.37, df = 1 (P = 0.54), I² = 0%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 110

Chapter 9, Figure 42 (analysis 5.4): Forest plot of comparison: 5 Probiotic versus placebo - sensitivity analysis using Random Effects model, outcome: 5.4 Triglyceride.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI

Agerholm-Larsen 2000 0 0.63 46 -0.01 1.05 14 0.7% 0.01 [-0.57, 0.59] Anderson 1999 0.03 0.77 40 -0.048 0.69 40 2.2% 0.08 [-0.24, 0.40] Ataie-Jafari 2009 -0.22 0.68 14 -0.12 0.34 14 1.4% -0.10 [-0.50, 0.30] de-Roos 1999 -0.02 0.45 39 -0.05 0.38 39 6.7% 0.03 [-0.15, 0.21] Ejtahed 2011 0 0.63 30 0.09 0.39 30 3.3% -0.09 [-0.36, 0.18] Fabian 2006 -0.04 0.14 17 -0.01 0.14 16 25.0% -0.03 [-0.13, 0.07] Fuentes 2013 -0.33 0.56 30 -0.25 0.59 30 2.7% -0.08 [-0.37, 0.21] Hata 1996 -0.12 0.29 17 -0.15 0.72 13 1.3% 0.03 [-0.38, 0.44] Hlivak 2005 0.19 1 20 0.04 0.4 18 1.0% 0.15 [-0.33, 0.63] Inoue 2003 0.29 0.88 18 0.07 0.88 17 0.7% 0.22 [-0.36, 0.80] Ivey 2013 0 0.52 109 -0.04 0.5 39 6.7% 0.04 [-0.14, 0.22] Jauhiainen 2005 0.03 0.69 50 0 0.7 51 3.1% 0.03 [-0.24, 0.30] Jones 2012 0.3 1.22 56 0 0.75 58 1.6% 0.30 [-0.07, 0.67] Jones 2012a -0.02 0.71 66 0.13 1.04 61 2.3% -0.15 [-0.46, 0.16] Kekkonen 2008 0 0.6 11 0 0.5 15 1.2% 0.00 [-0.44, 0.44] Mizushima 2004 0.06 1.09 22 -0.07 1.04 22 0.6% 0.13 [-0.50, 0.76] Rizkalla 2000 0 0.24 24 0 0.38 24 7.1% 0.00 [-0.18, 0.18] Sadrzadeh-Yeganeh 2010 0 0.3 29 0.03 0.13 30 16.2% -0.03 [-0.15, 0.09] Simons 2006 -0.1 0.55 23 0 0.62 21 1.9% -0.10 [-0.45, 0.25] Songisepp 2012: study 1 0 0.41 12 0.3 0.48 12 1.8% -0.30 [-0.66, 0.06] Songisepp 2012: study 2 0 0.43 18 -0.1 0.36 18 3.4% 0.10 [-0.16, 0.36] Usinger 2010 0.075 0.31 60 -0.03 0.39 30 8.9% 0.10 [-0.06, 0.27] Xiao 2003 -0.08 1.89 16 0.03 1.91 16 0.1% -0.11 [-1.43, 1.21]

Total (95% CI) 767 628 100.0% -0.00 [-0.05, 0.05]

Heterogeneity: Tau² = 0.00; Chi² = 11.92, df = 22 (P = 0.96); I² = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.02 (P = 0.98) Favours probiotic Favours placebo

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 111

Chapter 9, Figure 43 (analysis 6.4): Forest plot of comparison: 6 Probiotic versus placebo - capsules vs fermented milk, outcome: 6.4 Triglyceride.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 6.4.2 Capsules

Agerholm-Larsen 2000 0 0.63 46 -0.01 1.05 14 0.7% 0.01 [-0.57, 0.59] Fuentes 2013 -0.33 0.56 30 -0.25 0.59 30 2.7% -0.08 [-0.37, 0.21] Hlivak 2005 0.19 1 20 0.04 0.4 18 1.0% 0.15 [-0.33, 0.63] Ivey 2013 0.01 0.5 39 -0.04 0.5 19 3.0% 0.05 [-0.22, 0.32] Jones 2012a -0.02 0.71 66 0.13 1.04 61 2.3% -0.15 [-0.46, 0.16] Simons 2006 -0.1 0.55 23 0 0.62 21 1.9% -0.10 [-0.45, 0.25] Subtotal (95% CI) 224 163 11.6% -0.04 [-0.18, 0.10]

Heterogeneity: Chi² = 1.72, df = 5 (P = 0.89); I² = 0% Test for overall effect: Z = 0.54 (P = 0.59)

6.4.3 Fermented milk

Anderson 1999 0.03 0.77 40 -0.048 0.69 40 2.2% 0.08 [-0.24, 0.40] Ataie-Jafari 2009 -0.22 0.68 14 -0.12 0.34 14 1.4% -0.10 [-0.50, 0.30] de-Roos 1999 -0.02 0.45 39 -0.05 0.38 39 6.6% 0.03 [-0.15, 0.21] Ejtahed 2011 0 0.63 30 0.09 0.39 30 3.2% -0.09 [-0.36, 0.18] Fabian 2006 -0.04 0.14 17 -0.01 0.14 16 24.8% -0.03 [-0.13, 0.07] Hata 1996 -0.12 0.29 17 -0.15 0.72 13 1.3% 0.03 [-0.38, 0.44] Inoue 2003 0.29 0.88 18 0.07 0.88 17 0.7% 0.22 [-0.36, 0.80] Ivey 2013 -0.03 0.12 32 -0.4 0.5 20 4.6% 0.37 [0.15, 0.59] Jauhiainen 2005 0.03 0.69 50 0 0.7 51 3.1% 0.03 [-0.24, 0.30] Jones 2012 0.3 1.22 56 0 0.75 58 1.6% 0.30 [-0.07, 0.67] Kekkonen 2008 0 0.6 11 0 0.5 15 1.2% 0.00 [-0.44, 0.44] Mizushima 2004 0.06 1.09 22 -0.07 1.04 22 0.6% 0.13 [-0.50, 0.76] Rizkalla 2000 0 0.24 24 0 0.38 24 7.0% 0.00 [-0.18, 0.18] Sadrzadeh-Yeganeh 2010 0 0.3 29 0.03 0.13 30 16.1% -0.03 [-0.15, 0.09] Songisepp 2012: study 1 0 0.41 12 0.3 0.48 12 1.8% -0.30 [-0.66, 0.06] Songisepp 2012: study 2 0 0.43 18 -0.1 0.36 18 3.4% 0.10 [-0.16, 0.36] Usinger 2010 0.075 0.31 60 -0.03 0.39 30 8.8% 0.10 [-0.06, 0.27] Xiao 2003 -0.08 1.89 16 0.03 1.91 16 0.1% -0.11 [-1.43, 1.21] Subtotal (95% CI) 505 465 88.4% 0.02 [-0.03, 0.07]

Heterogeneity: Chi² = 19.70, df = 17 (P = 0.29); I² = 14% Test for overall effect: Z = 0.86 (P = 0.39)

Total (95% CI) 729 628 100.0% 0.02 [-0.03, 0.06]

Heterogeneity: Chi² = 22.05, df = 23 (P = 0.52); I² = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.62 (P = 0.53) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 0.63, df = 1 (P = 0.43), I² = 0%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 112

Chapter 9, Figure 44 (analysis 7.6): Forest plot of comparison: 7 Probiotic versus placebo - baseline cholesterol level, outcome: 7.6 Triglyceride.

Probiotic Placebo Mean Difference Mean Difference Study or Subgroup Mean SD Total Mean SD Total Weight IV, Fixed, 95% CI IV, Fixed, 95% CI 7.6.2 Desirable baseline total-cholesterol

Agerholm-Larsen 2000 0 0.63 46 -0.01 1.05 14 0.7% 0.01 [-0.57, 0.59] de-Roos 1999 -0.02 0.45 39 -0.05 0.38 39 6.9% 0.03 [-0.15, 0.21] Ejtahed 2011 0 0.63 30 0.09 0.39 30 3.4% -0.09 [-0.36, 0.18] Fabian 2006 -0.04 0.14 17 -0.01 0.14 16 25.8% -0.03 [-0.13, 0.07] Rizkalla 2000 0 0.24 24 0 0.38 24 7.3% 0.00 [-0.18, 0.18] Sadrzadeh-Yeganeh 2010 0 0.3 29 0.03 0.13 30 16.7% -0.03 [-0.15, 0.09] Songisepp 2012: study 1 0 0.41 12 0.3 0.48 12 1.8% -0.30 [-0.66, 0.06] Usinger 2010 0.075 0.31 60 -0.03 0.39 30 9.2% 0.10 [-0.06, 0.27] Subtotal (95% CI) 257 195 71.9% -0.01 [-0.07, 0.04]

Heterogeneity: Chi² = 5.33, df = 7 (P = 0.62); I² = 0% Test for overall effect: Z = 0.45 (P = 0.65)

7.6.3 Borderline high baseline total-cholesterol

Ataie-Jafari 2009 -0.22 0.68 14 -0.12 0.34 14 1.5% -0.10 [-0.50, 0.30] Hata 1996 -0.12 0.29 17 -0.15 0.72 13 1.4% 0.03 [-0.38, 0.44] Hlivak 2005 0.19 1 20 0.04 0.4 18 1.0% 0.15 [-0.33, 0.63] Inoue 2003 0.29 0.88 18 0.07 0.88 17 0.7% 0.22 [-0.36, 0.80] Ivey 2013 0 0.52 109 -0.04 0.5 39 6.9% 0.04 [-0.14, 0.22] Jones 2012a -0.02 0.71 66 0.13 1.04 61 2.4% -0.15 [-0.46, 0.16] Kekkonen 2008 0 0.6 11 0 0.5 15 1.2% 0.00 [-0.44, 0.44] Mizushima 2004 0.06 1.09 22 -0.07 1.04 22 0.6% 0.13 [-0.50, 0.76] Songisepp 2012: study 2 0 0.43 18 -0.1 0.36 18 3.5% 0.10 [-0.16, 0.36] Subtotal (95% CI) 295 217 19.3% 0.03 [-0.08, 0.14]

Heterogeneity: Chi² = 2.74, df = 8 (P = 0.95); I² = 0% Test for overall effect: Z = 0.50 (P = 0.62)

7.6.4 High baseline total cholesterol

Anderson 1999 0.03 0.77 40 -0.048 0.69 40 2.3% 0.08 [-0.24, 0.40] Fuentes 2013 -0.33 0.56 30 -0.25 0.59 30 2.8% -0.08 [-0.37, 0.21] Jones 2012 0.3 1.22 56 0 0.75 58 1.7% 0.30 [-0.07, 0.67] Simons 2006 -0.1 0.55 23 0 0.62 21 2.0% -0.10 [-0.45, 0.25] Xiao 2003 -0.08 1.89 16 0.03 1.91 16 0.1% -0.11 [-1.43, 1.21] Subtotal (95% CI) 165 165 8.9% 0.03 [-0.13, 0.19]

Heterogeneity: Chi² = 3.22, df = 4 (P = 0.52); I² = 0% Test for overall effect: Z = 0.34 (P = 0.73)

Total (95% CI) 717 577 100.0% -0.00 [-0.05, 0.05]

Heterogeneity: Chi² = 11.86, df = 21 (P = 0.94); I² = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.06 (P = 0.95) Favours probiotic Favours placebo Test for subgroup differences: Chi² = 0.57, df = 2 (P = 0.75), I² = 0%

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 113

9.16 APPENDICES

9.16.1 Appendix A: Search strategies

Search terms and databases: Unless otherwise stated, search terms are free-text terms; MeSH = Medical subject heading (MEDLINE medical index term); exp = exploded MeSH; the dollar sign ($) stands for any character(s); the question mark (?) substitutes one or no characters; tw = text word; pt = publication type; sh = MeSH; adj = adjacent (i.e. number of words within range of search term).

The Cochrane Library search strategy # 1 MeSH descriptor Cardiovascular diseases explode all trees # 2 MeSH descriptor Cholesterol explode all trees # 3 MeSH descriptor Triglycerides explode all trees # 4 MeSH descriptor Lipoproteins explode all trees # 5(cholesterol* in All Text or triglycerid* in All Text or lipoprotein* in All Text or triacylglycerol* in All Text) # 6 (HDL in All Text or LDL in All Text) # 7 ((cardiovascular in All Text near/6 diseas* in All Text) or (cardiovascular in All Text near/6 abnormalit* in All Text) or (cardiovascular in All Text near/6 infection* in All Text)) # 8 (#1 or #2 or #3 or #4 or #5 or #6 or #7) # 9 MeSH descriptor Lactobacillus explode all trees # 10 MeSH descriptor Probiotics explode all trees # 11 MeSH descriptor Fermentation explode all trees # 12 MeSH descriptor Dairy products explode all trees # 13 MeSH descriptor Bifidobacterium explode all trees # 14 MeSH descriptor Cultured milk products explode all trees # 15 (fermented in All Text and milk in All Text) # 16 (lactic in All Text and (acid* in All Text near/6 bacteri* in All Text) ) # 17(lactobacil* in All Text or probiotic* in All Text) # 18 ( (dairy in All Text and product* in All Text) or bifidobacteri* in All Text) # 19 ( ( (product in All Text near/6 cultured in All Text) and milk in All Text) or (product* in All Text near/6 fermented in All Text) ) # 20 (probiotic in All Text near/6 yogurt* in All Text) # 21 (#9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 or #17 or #18 or #19 or #20) # 22 (#8 and #21)

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 114

MEDLINE search strategy 1 exp Cardiovascular Diseases/ 2 exp Cholesterol/ 3 exp Triglycerides/ 4 exp Triacylglycerol/ 5 exp Lipoproteins/ 6 (cholesterol or triglycerid* or lipoprotein* or triacylglycerol).tw,ot. 7 (HDL or LDL).tw,ot. 8 (cardiovascular adj6 (diseas* or abnormalit* or infection*)).tw,ot. 9 or/1-8 10 exp Lactobacillus/ or exp Probiotics/ 11 Fermented milk.mp. 12 exp Fermentation/ 13 exp Dairy Products/ 14 exp Bifidobacterium/ 15 exp Cultured milk products/ 16 exp Yoghurt/ 17 (lactic acid* adj6 bacteri*).tw,ot. 18 (lactobacill* or probiotic*).tw,ot. 19 dairy product*.tw,ot. 20 bifidobacteri*.tw,ot. 21 (product* adj6 (cultured milk* or fermented)).tw,ot. 22 (probiotic adj6 yogurt*).tw,ot. 23 or/10-22 24 9 and 23 25 randomised controlled trial.pt. 26 controlled clinical trial.pt. 27 randomi?ed.ab. 28 placebo.ab. 29 drug therapy.fs. 30 randomly.ab. 31 trial.ab. 32 groups.ab. 33 or/25-32 34 Meta-analysis.pt. 35 exp Technology Assessment, Biomedical/ 36 exp Meta-analysis/ 37 exp Meta-analysis as topic/ 38 hta.tw,ot. 39 (health technology adj6 assessment$).tw,ot. 40 (meta analy$ or metaanaly$ or meta?analy$).tw,ot. 41 ((review$ or search$) adj10 (literature$ or medical database$ or medline or pubmed or embase or cochrane or cinahl or psycinfo or psyclit or healthstar or biosis or current content$ or systemat$)).tw,ot. 42 or/34-41 43 33 or 42 44 (comment or editorial or historical-article).pt. 45 43 not 44 46 24 and 45 47 (animals not (animals and humans)).sh. 48 46 not 47 49 limit 48 to adults

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 115

EMBASE search strategy 1 exp cardiovascular disease/ 2 exp cholesterol/ 3 exp triacylglycerol/ 4 exp lipoprotein/ 5 (cholesterol* or triglycerid* or lipoprotein* or triacylglycerol*).tw,ot. 6 (HDL or LDL).tw,ot. 7 (cardiovascular adj6 (diseas* or abnormalit* or infection*)).tw,ot. 8 or/1-7 9 exp Lactobacillus/ 10 exp probiotic agent/ 11 exp fermentation/ 12 exp dairy product/ 13 exp Bifidobacterium/ 14 exp yogurt/ 15 fermented milk.tw,ot. 16 (lactic acid* adj6 bacteri*).tw,ot. 17 (lactobacill* or probiotic*).tw,ot. 18 dairy product*.tw,ot. 19 bifidobacteri*.tw,ot. 20 (product* adj6 (cultured milk or fermented)).tw,ot. 21 (probiotic adj6 yogurt*).tw,ot. 22 or/9-21 23 8 and 22 24 exp Randomized Controlled Trial/ 25 exp Controlled Clinical Trial/ 26 exp Clinical Trial/ 27 exp Comparative Study/ 28 exp Drug comparison/ 29 exp Randomization/ 30 exp Crossover procedure/ 31 exp Double blind procedure/ 32 exp Single blind procedure/ 33 exp Placebo/ 34 exp Prospective Study/ 35 ((clinical or control$ or comparativ$ or placebo$ or prospectiv$ or randomi?ed) adj3 (trial$ or stud$)).ab,ti. 36 (random$ adj6 (allocat$ or assign$ or basis or order$)).ab,ti. 37 ((singl$ or doubl$ or trebl$ or tripl$) adj6 (blind$ or mask$)).ab,ti. 38 (cross over or crossover).ab,ti. 39 or/24-38 40 exp meta analysis/ 41 (metaanaly$ or meta analy$ or meta?analy$).ab,ti,ot. 42 ((review$ or search$) adj10 (literature$ or medical database$ or medline or pubmed or embase or cochrane or cinahl or psycinfo or psyclit or healthstar or biosis or current content$ or systematic$)).ab,ti,ot. 43 exp Literature/ 44 exp Biomedical Technology Assessment/ 45 hta.tw,ot. 46 (health technology adj6 assessment$).tw,ot. 47 or/40-46 48 39 or 47 49 (comment or editorial or historical-article).pt. 50 48 not 49 51 23 and 50 52 limit 51 to human 53 limit 52 to adult 'My NCBI' alert service (("probiotics"[MeSH Terms] OR "probiotics"[All Fields]) OR (("probiotics"[MeSH Terms] OR "probiotics"[All Fields] OR "probiotic"[All Fields]) AND ("microbiology"[Subheading] OR "microbiology"[All Fields] OR "bacteria"[All Fields] OR "bacteria"[MeSH Terms]))) AND ("cardiovascular diseases"[MeSH Terms] OR ("cardiovascular"[All Fields] AND "diseases"[All Fields]) OR "cardiovascular diseases"[All Fields] OR ("cardiovascular"[All Fields] AND "disease"[All Fields]) OR "cardiovascular disease"[All Fields]) AND Randomized Controlled Trial[ptyp]

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 116

9.16.2 Appendix B: Description of interventions

Study Intervention(s) Comparator(s) Design Blinding [strain, total dose/day] [strain, total dose/day]

Agerholm-Larsen I1 (G):450 mL FM/d. E. faecum (1 strain), C1: (PY):450 mL FM/d. Parallel Double 2000 S. thermophillus (2 strains). Minimum Fermented with organic blinded probiotic dose (log CFU/d) = 11.66 acid I2 (StLa):450 mL FM/d. S. thermophillus (2 strains), L. acidophilus (2 strains). Minimum probiotic dose (log CFU/d) = 10.73 I3 (StLr): 450 mL FM/d. S. thermophillus (2 strains), L. rhamnosus (1 strain). Minimum probiotic dose (log CFU/d) = 11.65 Anderson 1999 I1: 200 g FM/d. Starter: S. thermophillus C1: 200 g FM/d. Starter: Cross Double MUH34. Probiotic: L. acidophilus L1. S. thermophillus over blinded Minimum probiotic dose (log CFU/d) = MUH34 7.00 Ataie-Jafari 2009 I1:300 g FM /d. Starter: S. thermophillus C1:300 g FM /d. Starter: Cross Single and L. delbrueckii subsp bulgaricus S. thermophillus and L. over blinded Probiotic: L. acidophilus and B. lactis. delbrueckii subsp Minimum probiotic dose (log CFU/d) = bulgaricus 8.78 de-Roos 1999 I1:500mL FM/d. Starter: S. thermophillus. C1:500mL FM/d. Parallel Double Probiotic: L. acidophilus L-1. Minimum Starter: S. thermophillus blinded probiotic dose (log CFU/d) = 9.68 Ejtahed 2011 I1:300 g FM/d. Starter: S. thermophillus C1:300 g FM/d. Starter: Parallel Double and L. bulgaricus S. thermophillus and L. blinded Probiotic: L. acidophilus La5 and B. lactis bulgaricus Bb12. Minimum probiotic dose (log CFU/d) = 9.60 Fabian 2006 I1:100 g FM/d. Starter: S. thermophillus C1:100 g FM/d. Starter: Cross N/S and L. bulgaricus S. thermophillus and L. over Probiotic: L. paracasei subsp paracasei. bulgaricus Minimum probiotic dose (log CFU/d) = 8.56 Fuentes 2013 I1:capsule. L. plantarium (three strains). C1:capsule. Placebo Parallel Double Minimum probiotic dose (log CFU/d) = capsule blinded 9.08 Hata 1996 I1:95 mL FM/d. L. helveticus and S. C1:95 mL FM/d. Parallel N/S cerivasea. Minimum probiotic dose (log Artificially acidified CFU/d) = 11.85 Hlivak 2005 I1:capsules. Enterococcus faecum M-74. C1:capsules. placebo Cross nil Minimum probiotic dose (log CFU/d) = capsule over 9.30. Inoue 2003 I1:100 mL FM/d. L. casei shirota and C1:100 mL FM/d. Parallel Single Lactococcus lactis YIT 2027. Minimum acidified with L-lactic blinded probiotic dose (log CFU/d) = NS acid Footnotes: I: intervention. C: control. FM: fermented milk. CFU: colony forming units.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 117

Study Intervention(s) Comparator(s) Design Blinding [strain, total dose/day] [strain, total dose/day]

Ivey 2013 I1: 200 mL FM/d PLUS capsules. C1: milk PLUS placebo Parallel Double blinded Lactobacillus acidophilus La5 and capsules Bifidobacterium animalis subsp lactis Bb12. Minimum probiotic dose (log CFU/d) = 9.78 I2: 200 mL FM/d . Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12. Minimum probiotic dose (log CFU/d) = 9.48 I3: Capsules. Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12. Minimum probiotic dose (log CFU/d) = 9.48 Jauhiainen 2005 I1:300 mL FM/d. L. helveticus. C1:300 mL FM/d. Parallel Double blinded Minimum probiotic dose (log CFU/d) Normal mesophilic = NS Lactococcus sp. mixed culture. NS Jones 2012 I1:400 mL FM/d. Starter culture: C1:400 mL FM/d. Parallel Double blinded yogurt bacteria. Probiotic bacteria: L. Starter culture: yogurt reuteri NCIMB 30242. Minimum bacteria probiotic dose (log CFU/d) = 10.45 Jones 2012a I1: Capsules. Probiotic bacteria: L. C1:no probiotic Parallel Double blinded reuteri NCIMB 30242 (Cardioviva). bacteria Minimum probiotic dose (log CFU/d) = 9.60 Kekkonen 2008 I1:250 mL milk based fruit drink. C1:250 mL milk based Parallel Double blinded Probiotic bacteria: L. rhamnosus GG fruit drink. NS (Similar ATCC 53103. Minimum probiotic placebo drink without dose (log CFU/d) = 10.19 probiotic bacteria) Mizushima 2004 I1:160 g FM/d. L. helveticus and S. C1:160 g FM/d. Parallel N/S cerevisiae. Minimum probiotic dose Artificially acidified (log CFU/d) = N/S Rizkalla 2000 I1:500 g FM/d. L. bulgaricus and S. C1:500 g FM/d. Cross N/S thermophillus. Minimum probiotic Pasteurised probiotic over dose (log CFU/d) = 10 FM Sadrzadeh- I1:300 g FM/d. Starter: L. bulgaricus C1:300 g FM/d. Starter: Parallel Double blinded Yeganeh 2010 and S. thermophiles L. bulgaricus and S. Probiotic: L. acidophilus La5 and B. thermophiles lactis Bb12. Minimum probiotic dose (log CFU/d) = 7.89 Simons 2006 I1:Capsules. L. fermentum PCC. C1:Capsules.Placebo Parallel Double blinded Minimum probiotic dose (log CFU/d) capsule = 9.60 Songisepp 2012: I1: 50g FM/d. Starter: C92 cheese C1: 50g FM/d. Starter: Cross Double blinded study 1 starter C92 cheese starter over Probiotic: L. fermentum Tensia. Minimum probiotic dose (log CFU/d) = 10.4 Footnotes: I: intervention. C: control. FM: fermented milk. CFU: colony forming units.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 118

Study Intervention(s) Comparator(s) Design Blinding [strain, total dose/day] [strain, total dose/day]

Songisepp I1: 50g FM/d. Starter: C92 cheese C1: 50g FM/d. Starter: C92 cheese Cross Double 2012: study 2 starter starter over blinded Probiotic: L. fermentum Tensia. Minimum probiotic dose (log CFU/d) = 8.17 Usinger 2010 I1:300 mL/d FM: L. helveticus C1: 300 mL/d FM: Artificially Parallel Double Cardio-04. Minimum probiotic dose acidified blinded (log CFU/d) = NA C2: 150 mL/d FM: Artificially I2: 150 mL/d FM: L. helveticus acidified Cardio-04. Minimum probiotic dose (log CFU/d) = NA Xiao 2003 I1:300 mL FM/d. Starter: S. C1:300 mL FM/d. Starter: S. Parallel Single thermophillus and L. delbrueckii thermophillus and L. delbrueckii blinded subsp bulgaricus. subsp bulgaricus. Probiotic: B. longum BL1. Minimum probiotic dose (log CFU/d) = 8.61 Footnotes: I: intervention. C: control. FM: fermented milk. CFU: colony forming units.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 119

9.16.3 Appendix C: Baseline characteristics (1)

Study Interventions Co-interventions Duration in weeks Participating (lead-in/intervention/washout/ population follow-up) Agerholm- I1 (G) nil 4/8/NA/NS overweight/obese Larsen 2000 I2 (StLa) nil 4/8/NA/NS overweight/obese I3 (StLr) nil 4/8/NA/NS overweight/obese C1 (PY) nil 4/8/NA/NS overweight/obese all: nil 4/8/NA/NS overweight/obese Anderson I1 nil NS/4/2/NS primary 1999 hypercholesterolaemia C1 nil NS/4/2/NS primary hypercholesterolaemia all: nil NS/4/2/NS primary hypercholesterolaemia Ataie-Jafari I1 nil 2/6/4/NS mild to moderate 2009 hypercholesterolaemia C1 nil 2/6/4/NS mild to moderate hypercholesterolaemia all: nil 2/6/4/NS mild to moderate hypercholesterolaemia de-Roos 1999 I1 nil 2/6/NS/NS Healthy C1 nil 2/6/NS/NS Healthy all: nil 2/6/NS/NS Healthy Ejtahed 2011 I1 nil 1/6/NS/NS Type 2 diabetes mellitus C1 nil 1/6/NS/NS Type 2 diabetes mellitus all: nil 1/6/NS/NS Type 2 diabetes mellitus for at least one year Fabian 2006 I1 nil 1/2/2/2 Healthy C1 nil 1/2/2/2 Healthy all: nil 1/2/2/2 Healthy Fuentes 2013 I1 nil NS/12/4 hypercholesterolaemia C1 nil NS/12/4 hypercholesterolaemia all: nil NS/12/NA/4 hypercholesterolaemia Hata 1996 I1 nil 4/8/NA/NS hypertension C1 nil 4/8/NA/NS hypertension all: nil 4/8/NA/NS hypertension Hlivak 2005 I1 50 ug organically NS/56/NS/4 elderly nursing home bound selenium patients C1 nil NS/56/NS/4 elderly nursing home patients all: NS/56/NS/4 elderly nursing home patients

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 120

Study Interventions Co-interventions Duration in weeks Participating (lead-in/intervention/washout/ population follow-up) Inoue 2003 I1 nil NS/12/NA/2 mild hypertension C1 nil NS/12/NA/2 mild hypertension all: nil NS/12/NA/2 mild hypertension Ivey 2013 I1 nil 3/6/NA/0 overweight/obese and elevated waist circumference I2 Placebo capsules 3/6/NA/0 overweight/obese and elevated waist circumference I3 Milk 3/6/NA/0 overweight/obese and elevated waist circumference C1 nil 3/6/NA/0 overweight/obese and elevated waist circumference all: 3/6/NA/0 overweight/obese and elevated waist circumference Jauhiainen I1 nil 4/10/NA/4 hypertension 2005 C1 nil 4/10/NA/4 hypertension all: nil 4/10/NA/4 hypertension Jones 2012 I1 nil 2/6/NA/2 hypercholesterolaemia C1 nil 2/6/NA/2 hypercholesterolaemia all: nil 2/6/NA/2 hypercholesterolaemia Jones 2012a I1 nil 4/9/NA/NS hypercholesterolaemia C1 nil 4/9/NA/NS hypercholesterolaemia all: nil 4/9/NA/NS hypercholesterolaemia Kekkonen I1 nil 3/3/NA/NS Healthy 2008 C1 nil 3/3/NA/NS Healthy all: nil 3/3/NA/NS Healthy Mizushima I1 nil NS/4/NA/NS hypertension 2004 C1 nil NS/4/NA/NS hypertension all: nil NS/4/NA/NS hypertension Rizkalla I1 nil NS/2.1/2.1/NS Healthy 2000 C1 nil NS/2.1/2.1/NS Healthy all: nil NS/2.1/2.1/NS Healthy

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 121

Study Interventions Co-interventions Duration in weeks Participating (lead-in/intervention/washout/ population follow-up) Sadrzadeh-Y I1 nil 1/6/NA/NS Healthy eganeh 2010 C1 nil 1/6/NA/NS Healthy all: nil 1/6/NA/NS Healthy Simons 2006 I1 nil 1/10/NS Healthy C1 nil 1/10/NS Healthy all: nil 1/10/NS Healthy Songisepp I1 nil 4/3/2/NS healthy 2012: study 1 C1 nil 4/3/2/NS healthy all: nil 4/3/2/NS healthy Songisepp I1 nil 4/3/2/NS healthy 2012: study 2 C1 nil 4/3/2/NS healthy all: nil 4/3/2/NS healthy Usinger 2010 I1 nil NS/8/NS Prehypertension or borderline hypertension I2 nil NS/8/NS Prehypertension or borderline hypertension C1 nil NS/8/NS Prehypertension or borderline hypertension C2 nil NS/8/NS Prehypertension or borderline hypertension all: nil NS/8/NS Prehypertension or borderline hypertension Xiao 2003 I1 nil 2/4/NA/NS Healthy C1 nil 2/4/NA/NS Healthy all: nil 2/4/NA/NS Healthy

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 122

Study Intervention Co-morbidities Study Country Sex Age period M:F (years) Mean ± SD or range Agerholm- I1 (G) nil NS Denmark 4 : 12 38 ± 8 Larsen 2000 I2 (StLa) nil NS Denmark 4 : 12 39 ± 8 I3 (StLr) nil NS Denmark 4 : 10 38 ± 7 C1 (PY) nil NS Denmark 5 : 9 39 ± 7 all: nil NS Denmark 19:43 NS Anderson I1 nil NS USA 9:12 55 ± 14 1999 C1 nil NS USA 9:10 58 ± 13 all: nil NS USA 18 : 22 NS Ataie-Jafari I1 - NS Iran NS NS 2009 C1 - NS Iran NS NS all: - NS Iran 4:10 50 ± 7 de-Roos 1999 I1 - NS Netherlands 11 : 28 40 ± 8 C1 - NS Netherlands 11 : 28 40 ± 9 all: - NS Netherland 22 : 56 40 Ejtahed 2011 I1 LDLC ≥ 2.6 mmol/L NS Iran 11 : 19 51 ± 8 C1 LDLC ≥ 2.6 mmol/L NS Iran 12 : 18 51 ± 7 all: LDLC ≥ 2.6 mmol/L NS Iran 23 : 37 NS Fabian 2006 I1 nil NS Austria 0 : 17 24 ± 3 C1 nil NS Austria 0 : 16 24 ± 2 all: nil NS Austria 0 : 32 NS Fuentes 2013 I1 nil NS Spain 30 NS (sex NS) C1 nil NS Spain 30 NS (sex NS) all: nil NS Spain 60 NS (sex NS) Hata 1996 I1 nil NS Japan 4 : 13 76 ± 7 C1 nil NS Japan 4 : 9 73 ± 11 all: nil NS Japan 8 : 22 NS Hlivak 2005 I1 nil 2001-02 Slovakia 3 : 17 75 ± 7 C1 nil 2001-02 Slovakia 4 : 14 78 ± 7 all: nil 2001-02 Slovakia 7 : 31 75 ± 10 Inoue 2003 I1 nil NS Japan 10 : 8 56 ± 9 C1 nil NS Japan 10 : 7 54 ± 14 all: nil NS Japan 23 : 16 NS

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 123

Study Intervention Co-morbidities Study Country Sex Age period M:F (years) Mean ± SD or range Ivey 2013 I1 nil 2012-13 Australia 25 : 15 68 ± 8 I2 nil 2012-13 Australia 25 : 12 68 ± 9 I3 nil 2012-13 Australia 23 : 16 65 ± 7 C1 nil 2012-13 Australia 23 : 17 65 ± 8.4 all: nil 2012-13 Australia 96 : 60 67 ± 8 Jauhiainen I1 nil NS Finland 19 : 34 51 ± 12 2005 C1 nil NS Finland 17 : 38 55 ± 11 all: nil NS Finland 36 : 72 NS Jones 2012 I1 nil NS Canada 38 : 18 52 ± 14 C1 nil NS Canada 34 : 24 49 ± 13 all: nil NS Canada 72 : 42 NS Jones 2012a I1 nil NS Canada 28:38 50 ± 14 C1 nil NS Canada 27:34 48 ± 13 all: nil NS Canada 55:72 NS Kekkonen I1 nil NS Finland NS NS 2008 C1 nil NS Finland NS NS all: nil NS Finland 12 : 14 42 (range 23 - 55) Mizushima I1 nil NS Japan 23 : 0 44 ± 10 2004 C1 nil NS Japan 23 : 0 48 ± 9 all: nil NS Japan 46 : 0 NS Rizkalla I1 nil NS France 12 : 0 NS 2000 C1 nil NS France 12 : 0 NS all: nil NS France 24 : 0 NS Sadrzadeh-Y I1 nil NS Iran 0 : 29 36 eganeh 2010 C1 nil NS Iran 0 : 30 32 all: nil NS Iran 0 : 59 35 Simons 2006 I1 TC ≥ 4.0 mmol/L; 2004-05 Australia 8 : 13 53 ± 11 tgl ≤ 4.0 mmol/L C1 TC ≥ 4.0 mmol/L; 2004-05 Australia 8 : 15 50 ± 12 tgl ≤ 4.0 mmol/L all: TC ≥ 4.0 mmol/L; 2004-05 Australia 16 : 28 NS tgl ≤ 4.0 mmol/L Songisepp I1 nil 2009-09 Estonia NS NS 2012: study 1 C1 nil 2009-09 Estonia NS NS all: nil 2009-09 Estonia 5:8 29 ± 8 Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 124

Study Intervention Co-morbidities Study Country Sex Age period M:F (years) Mean ± SD or range Songisepp I1 nil 2009-09 Estonia NS NS 2012: study 2 C1 nil 2009-09 Estonia NS NS all: nil 2009-09 Estonia 1:17 70 ± 6 Usinger 2010 I1 nil 2007-08 Denmark 14 : 16 54 ± 12 I2 nil 2007-20 Denmark 22 : 8 52 ± 10 08 C1 nil 2007-20 Denmark NS NS 08 C2 nil 2007-20 Denmark NS NS 08 all: nil 2007-20 Denmark 50 : 40 NS 08 Xiao 2003 I1 nil NS Japan NS 44 ± 8 C1 nil NS Japan NS 44 ± 8 all: nil NS Japan 32 : 0 NS

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 125

9.16.4 Appendix D: Baseline characteristics (2)

Study Intervention Baseline total Baseline low Baseline high Baseline cholesterol density density triglyceride (mean ± SD lipoprotein lipoprotein (mean ± SD mmol/L) cholesterol (mean ± SD mmol/L) (mean ± SD mmol/L) mmol/L) Agerholm- I1 (G) 5.1 ± 0.9 3.0 ± 0.8 1.3 ± 0.4 1.6 ± 0.8 Larsen 2000 I2 (StLa) 5.1 ± 0.9 3.2 ± 0.8 1.3 ± 0.4 1.4 ± 0.6 I3 (StLr) 5.2 ± 1.0 3.1 ± 0.8 1.4 ± 0.3 1.6 ± 0.9 C1 (PY) 5.3 ± 1.2 3.1 ± 0.8 1.3 ± 0.3 2.0 ± 1.2 all: Anderson 1999 I1 6.5 ± 0.8 4.5 1.2 1.9 C1 6.3 ± 0.6 4.3 1.2 1.9 all: Ataie-Jafari I1 5.8 ± 0.5 3.7 ± 0.7 1.1 ± 0.3 2.2 ± 0.9 2009 C1 5.6 ± 0.7 3.7 ± 0.7 1.0 ± 0.2 2.0 ± 0.7 all: 5.7 ± 0.7 3.6 ± 0 .8 1.0 ± 0.3 2.3 ± 0.8 de-Roos 1999 I1 5.1 ± 0.7 3.2 ± 0.7 1.4 ± 0.3 1.1 ± 0.7 C1 5.2 ± 0.7 3.1 ± 0.8 1.5 ± 0.4 1.2 ± 0.5 all: Ejtahed 2011 I1 5.2 ± 0.7 3.2 ± 0.6 1.2 ± 0.3 1.7 ± 0.8 C1 4.9 ± 0.6 2.9 ± 0.6 1.3 ± 0.3 1.5 ± 0.6 all: Fabian 2006 I1 4.3 ± 0.7 2.5 ± 0.6 1.6 ± 0.4 0.7 ± 0.2 C1 4.5 ± 0.7 2.7 ± 0.6 1.7 ± 0.3 0.6 ± 0.2 all: Fuentes 2013 I1 6.4 ± 0.8 4.3 ± 0.6 1.1 ± 0.2 2.0 ± .8 C1 6.5 ± 0.6 4.4 ± 0.5 1.2 ± 0.2 2.1 ± 0.9 all: Hata 1996 I1 4.9 ± 0.9 NA 1.6 ± 0.5 1.3 ± 0.0 C1 5.6 ± 0.7 NA 1.7 ± 0.5 1.5 ± 0.9 all: NA Hlivak 2005 I1 5.9 ± 1.3 3.8 ± 1.2 1.2 ± 0.3 2.0 ± 1.4 C1 5.4 ± 1.1 3.6 ± 1.0 1.2 ± 0.3 1.5 ± 0.6 all: Inoue 2003 I1 5.3 ± 0.4 NA NA 1.6 ± 1.1 C1 5.5 ± 0.8 NA NA 1.5 ± 0.8 all: NA NA

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 126

Study Intervention Baseline total Baseline low Baseline high Baseline cholesterol density density triglyceride (mean ± SD lipoprotein lipoprotein (mean ± SD mmol/L) cholesterol (mean ± SD mmol/L) (mean ± SD mmol/L) mmol/L) Ivey 2013 I1 5.5 ± 1.0 3.3 ± 0.8 1.4 ± 0.4 1.7 ± 0.8 I2 5.2 ± 1.1 3.2 ± 1.2 1.4 ± 0.3 1.7 ± 0.9 I3 5.2 ± 1.1 3.1 ± 0.9 1.4 ± 0.3 1.6 ± 0.7 C1 5.3 ± 1.1 3.1 ± 1.0 1.4 ± 0.4 1.5 ± 0.8 total 5.3 ± 1.2 3.2 ± 1.0 1.4 ± 0.3 1.6 ± 0.8 Jauhiainen 2005 I1 NA 3.2 ± 0.9 NA 1.6 ± 1.0 C1 NA 3.1 ± 0.8 NA 1.6 ± 0.8 all: NA NA Jones 2012 I1 6.7 ± 0.9 4.4 ± 0.6 1.4 ± 0.4 1.6 ± 0.8 C1 6.6 ± 0.7 4.2 ± 0.5 1.5 ± 0.4 1.6 ± 0.8 all: Jones 2012a I1 6.4 ± 0.8 4.5 ± 0.6 1.3 ± 0.5 1.5 ± 0.8 C1 5.9 ± 0.6 4.1 ± 0.5 1.2 ± 0.3 1.5 ± 0.7 all: Kekkonen 2008 I1 5.4 ± 1.2 3.3 ± 1.0 1.5 ± 0.4 1.4 ± 1.1 C1 5.1 ± 1.1 3.1 ± 1.0 1.5 ± 0.4 1.0 ± 0.4 all: Mizushima 2004 I1 5.1 ± 0.6 NS 1.4 ± 0.4 1.8 ± 1.0 C1 5.3 ± 0.7 NS 1.4 ± 0.3 2.5 ± 2.0 all: NS Rizkalla 2000 I1 4.80 NS 1.2 0.85 C1 4.60 NS 1.10 1.0 all: NS Sadrzadeh- I1 4.8 ± 0.6 2.8 ± 0.5 1.3 ± 0.3 1.1 ± 0.4 Yeganeh 2010 C1 4.5 ± 0.7 2.6 ± 0.5 1.3 ± 0.2 0.9 ± 0.2 all: Simons 2006 I1 6.2 ± 1.0 4.0 ± 1.0 1.5 ± 0.3 1.6 ± 0.7 C1 6.3 ± 1.1 3.9 ± 1.1 1.8 ± 0.5 1.4 ± 0.9 all: Songisepp 2012 1 I1 4.6 ± 0.9 2.7 ± 0.8 1.7 ± 0.5 1.0 ± 0.6 C1 4.2 ± 0.6 2.6 ± 0.7 1.6 ± 0.4 0.9 ± 0.4 all: 4.6 ± 0.9 2.7 ± 0.8 1.7 ± 0.5 1.0 ± 0.6 Songisepp 2012 2 I1 5.7 ± 0.8 3.9 ± 0.8 1.7 ± 0.4 1.1 ± 0.6 C1 5.9 ± 0.9 4.1 ± 0.9 1.7 ± 0.4 1.2 ± 0.5 all: 5.8 ± 0.8 3.9 ± 0.8 1.7 ± 0.3 1.1 ± 0.6

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 127

Study Intervention Baseline total Baseline low Baseline high Baseline cholesterol density density triglyceride (mean ± SD lipoprotein lipoprotein (mean ± SD mmol/L) cholesterol (mean ± SD mmol/L) (mean ± SD mmol/L) mmol/L) Usinger 2010a I1 5.3 ± 0.7 3.5 ± 1.0 1.3 ± 0.3 1.0 ± 0.4 I2 5.2 ± 0.8 3.5 ± 0.7 1.3 ± 0.3 0.9 ± 0.4 C1 NS NS NS NS C2 NS NS NS NS all: Xiao 2003 I1 6.3 ± 0.4 4.0 ± 0.6 1.6 ± 0.4 1.6 ± 0.7 C1 6.3 ± 0.4 4.2 ± 0.4 1.4 ± 0.4 1.6 ± 0.7 all: Footnotes "NS" denotes not specified All studies were in a community setting, with the exception of Hlivlak 2005 which was conducted in a nursing home setting. No studies reported condition duration or ethnicity. C: comparator; I: intervention; SD: standard deviation

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 128

9.16.5 Appendix E: Matrix of study endpoints (publications)

Study Endpoint Time of reported in measurementa publication (weeks) Agerholm- Review's primary outcomes Larsen 2001 Cardiovascular events Health-related quality of life Adverse events Nausea, NS constipation Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, LDLC, TC, 0, 4, 8 HDL cholesterol and triglycerides HCLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b Blood pressure Body weight C reactive protein fat mass Fibrinogen FVIIc waist:hip ratio sagittal height tPA activity Anderson Review's primary outcomes 1999 Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 2, 3, 4 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b body weight

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 129

Study Endpoint Time of reported in measurementa publication (weeks) Ataie-Jafari Review's primary outcomes 2009 Cardiovascular events Health-related quality of life Adverse events adverse events - Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 6 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b LDLC:HDLC ratio de-Roos 1999 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events gastrointestinal - complaint Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 3, 6 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b body weight Ejtahed 2011 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events adverse events - Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 6 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b LDLC:HDLC ratio TC:HDLC ratio

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 130

Study Endpoint Time of reported in measurementa publication (weeks) Fabian 2006 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 4 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b HDLC:LDLC ratio TC:HDLC ratio Fuentes 2013 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 12 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b HDLC:LDLC ratio oxidised LDLC

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 131

Study Endpoint Time of reported in measurementa publication (weeks) Hata 1996 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events stomach-ache, - diarrhoea Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 8 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b blood urea nitrogen body weight chloride creatinine y-glutamyl transpeptidase glucose glutamic oxaloacetic transaminase glutamic pyruvic transaminase haemoglobin HDLC:LDLC ratio lactate dehydrogenase platelet count potassium red blood cell sodium TC:HDLC ratio total protein pulse rate blood pressure uric acid

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 132

Study Endpoint Time of reported in measurementa publication (weeks) Hlivak 2005 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 6, 12, 23, 44, HDL cholesterol and triglycerides HDLC, tgl 56 Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b Blood pressure HOMA hsCRP insulin TC:HDLC ratio

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 133

Study Endpoint Time of reported in measurementa publication (weeks) Inoue 2003 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, tgl 0, 12 HDL cholesterol and triglycerides Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b Blood pressure body weight glucosuria red blood cell count heart rate proteinuria white blood cell count platelets haemoglobin haematocrit aspartate aminotransferase alanine aminotransferase blood urea creatinine total protein blood glucose uric acid sodium potassium chloride

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 134

Study Endpoint Time of reported in measurementa publication (weeks) Ivey 2013 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events adverse events - Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 6 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b Blood pressure glycaemic control gastrointestinal symptoms Jauhiainen Review's primary outcomes 2005 Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, LDLC, tgl 0, 10 HDL cholesterol and triglycerides Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b Angiotensin converting enzyme activity Blood pressure Body weight C-reactive protein TC:HDLC ratio

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 135

Study Endpoint Time of reported in measurementa publication (weeks) Jones 2012 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 3, 6 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b apoB-100 faecal deconjugated bile acids LDLC:HDLC ratio non-HDLC complete blood cell count platelet haematocrit haemoglobin urea creatinine alanine transaminase aspartate aminotransferase y-glutamyl transpeptidase alkaline phosphatase bilirubin lipase phosphate potassium sodium chloride hydrogen carbonate

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 136

Study Endpoint Time of reported in measurementa publication (weeks) Jones 2012a Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0,3,6,9 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b apoB-100 apoB-100:apoA-1 ratio LDLC:HDLC ratio non-HDLC High sensitivity C reactive protein fibrinogen plasma deconjugated bile acids campesterol sitosterol stigmasterol Kekkonen Review's primary outcomes 2008 Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, LDLC, 0, 3 HDL cholesterol and triglycerides HDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b global lipidomic analysis C-reactive protein Interleukin-6 lysophospatidylcholine tumour necrosis factor-a sphingomyelin glycerophosphatidylcholine

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 137

Study Endpoint Time of reported in measurementa publication (weeks) Mizushima Review's primary outcomes 2004 Cardiovascular events Health-related quality of life Adverse events stomach-ache, - diarrhoea Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, HDLC, tgl 0, 2, 4 HDL cholesterol and triglycerides Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b Blood pressure Angiotensin I Angiotensin II Angiotensin I : Angiotensin II ratio Rizkalla 2000 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, HDLC, tgl 0, 2.1 HDL cholesterol and triglycerides Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b acetate breath hydrogen butyrate fatty acid glucose insulin propionate short chain fatty acids

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 138

Study Endpoint Time of reported in measurementa publication (weeks) Sadrzadeh- Review's primary outcomes Yeganeh 2010 Cardiovascular events Health-related quality of life Adverse events Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, tgl, TC, HDLC, 0, 3, 6 HDL cholesterol and triglycerides LDLC Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b TC:HDLC ratio Simons 2006 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Constipation, - flatus Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, tgl, TC, HDLC, 0, 4, 8, 10 HDL cholesterol and triglycerides LDLC Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b alanine aminotransferase aspartate aminotransferase C reactive protein creatinine gammaglutyl transpeptidase liver enzymes glucose

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 139

Study Endpoint Time of reported in measurementa publication (weeks) Songisepp Review's primary outcomes 2012 1 Cardiovascular events Health-related quality of life Adverse events welfare, adverse 0, 1, 2, 3 gastrointestinal symptoms (abdominal pain, flatulence, stool frequency, stool consistency) Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, tgl, TC, HDLC, HDL cholesterol and triglycerides LDLC Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b survival of the probiotic strain in Gastrointestinal Tract effect on faecal lactoflora body mass index blood pressure nutritional habits haematological indices (haemoglobin, white blood cell count, red blood cells, platelets) plasma glucose albumin ferritin high-sensitive C-Reactive Protein Immunoglobulins (IgA, IgM, IgG) urine the content of biogenic amines changes in the faecal counts of clostridia (including C. difficile), total anaerobes, enterococci, E. coli and lactic acid bacteria Interleukin 6

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 140

Study Endpoint Time of reported in measurementa publication (weeks) Songisepp Review's primary outcomes 2012 2 Cardiovascular events Health-related quality of life Adverse events welfare, adverse 0, 1, 2, 3 gastrointestinal symptoms (abdominal pain, flatulence, stool frequency, stool consistency) Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, tgl, TC, HDLC, HDL cholesterol and triglycerides LDLC Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b survival of the probiotic strain in Gastrointestinal Tract effect on faecal lactoflora body mass index blood pressure nutritional habits haematological indices (haemoglobin, white blood cell count, red blood cells, platelets) plasma glucose albumin ferritin high-sensitive C-Reactive Protein Immunoglobulins (IgA, IgM, IgG) urine the content of biogenic amines changes in the faecal counts of clostridia (including C. difficile), total anaerobes, enterococci, E. coli and lactic acid bacteria Interleukin 6

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 141

Study Endpoint Time of reported in measurementa publication (weeks) Usinger 2010a Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Gastrointestinal - pain, bloating Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, HDL, LDL, 0, 8 HDL cholesterol and triglycerides tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b Blood pressure Heart rate VLDL Xiao 2003 Review's primary outcomes Cardiovascular events Health-related quality of life Adverse events Faecal frequency - Review's secondary outcomes All-cause mortality Lipid serum levels of total cholesterol, LDL cholesterol, TC, HDLC, 0,4 HDL cholesterol and triglycerides LDLC, tgl Socioeconomic costs Other than review's primary/secondary outcomes reported in publication (classification: P/S/O)b glucose

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 142

9.16.6 Appendix F: Examination of outcome reporting bias

Study Clear that outcome Clear that outcome Clear that outcome Unclear whether the was measured and was measured and was measuredc [ outcome was analyseda analysedb measuredd Agerholm- low risk low risk low risk low risk Larsen 2000 Anderson 1999 low risk low risk low risk low risk Ataie-Jafari low risk low risk low risk low risk 2009 Bertolami low risk low risk low risk low risk 1999 de-Roos 1999 low risk low risk low risk low risk Ejtahed 2011 low risk low risk low risk low risk Fabian 2006 low risk low risk low risk low risk Fuentes 2013 low risk low risk low risk low risk Hata 1996 low risk low risk low risk low risk Hlivak 2005 low risk low risk low risk low risk Inoue 2003 low risk low risk low risk low risk Ivey 2013 low risk low risk low risk low risk Jauhiainen low risk low risk low risk low risk 2005 Jones 2012 low risk low risk low risk low risk Jones 2012a low risk low risk low risk low risk Kekkonen low risk low risk low risk low risk 2008 Klein 2008 low risk low risk low risk low risk Larsen 2006 low risk low risk low risk low risk Lewis 2005 low risk low risk low risk low risk Lin 1989 low risk low risk low risk low risk Mizushima low risk low risk low risk low risk 2004 Rizkalla 2000 low risk low risk low risk low risk Sadrzadeh- low risk low risk low risk low risk Yeganeh 2010 Simons 2006 low risk low risk low risk low risk Songisepp low risk low risk low risk low risk 2012 1 Songisepp low risk low risk low risk low risk 2012 2 Usinger 2010a low risk low risk low risk low risk Xiao 2003 low risk low risk low risk low risk

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 143

9.16.7 Appendix G: Outcome assessment

Study Total cholesterol Low density High density triglyceride lipoprotein lipoprotein cholesterol cholesterol Agerholm- Monotest Cholesterol Calculation using Precipitatint method: Test-Combination Larsen 2000 High Performance mmol/L units: LDLC magnesium-phosphor Triacylglycerol (CHOD-PAP) method = TC - (tgl / 2.2) - us-wolfram acid from (GPO-PAP) method from Boehringer HDLC Boehringer Mannheim from Boehringer Mannheim GmbH-Diagnostica Mannheim GmbH-Diagnostica (Germany). Measured GmbH-Diagnostica (Germany) on Cobas MiraS (Germany) Boehringer Mannheim GmbH-Diagnostica (Germany) Anderson Enzymatic method: Calculation using Precipitation method: Enzymatic method: 1999 Cholesterol ester mg/100mL units: Magnesium chloride microbial lipase and hydrolase and LDLC = TC - HDLC - and dextran sulphate. protease. Measured by cholesterol oxidase. (tgl / 5) Measured by an absorptiometry. Measured by enzymatic method absorptiometry with a discrete analyser Ataie-Jafari Enzymatic method: Calculation using Enzymatic method: Enzymatic method: 2009 ND (Pars Azmoon, mg/100mL units: ND (Pars Azmoon, ND (Pars Azmoon, Iran). Measured on a LDLC = TC - HDLC - Iran). Measured on a Iran). Measured on a Selectra II (tgl / 5) Selectra II Selectra II autoanalyser (Vita autoanalyser (Vita autoanalyser (Vita Lab, Finland) Lab, Finland) Lab, Finland) de-Roos 1999 Monotest Cholesterol ND Precipitation method: Enzymatic method: High Performance Magnesium chloride glycerol phosphate (CHOD-PAP) method and dextran sulphate. oxidase, peroxidase, from Boehringer Measured on: ND lipase, Mannheim 4-aminoantipyrine. GmbH-Diagnostica Measured by (Germany). Measured absorptiometry by high-performance thin-layer chromatography. Ejtahed 2011 Enzymatic method: Calculation using Precipitation method: Enzymatic method: cholesterol esterase mg/100mL units: ND. Enzymatic glycerol phosphate and cholesterol LDLC = TC - HDLC - method: cholesterol oxidase (Parasazmun oxidase (Parasazmun (tgl / 5) esterase and kit, Iran) kit, Iran) cholesterol oxidase (Parasazmun kit, Iran) Fabian 2006 Siedel 1983 Calculation using Burnstein 1970 Siedel 1993 mmol/L units: LDLC = TC - (tgl / 2.2) - HDLC Fuentes 2013 Enzymatic method: Enzymatic method: Enzymatic method: Enzymatic method: ND (Dade Behring, ND (Dade Behring, ND (Dade Behring, ND (Dade Behring, Siemens) Siemens) Siemens) Siemens) Hata 1996 ND ND ND ND Hlivak 2005 Reflotron (r) Plus Calculation using Reflotron (r) Plus Reflotron (r) Plus Meter (Roche mmol/L units: LDLC Meter (Roche Meter (Roche Diagnostics) analyser = TC - (HDL + tgl / Diagnostics) analyser Diagnostics) analyser 2.2) Inoue 2003 ND N/I N/I ND

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 144

Study Total cholesterol Low density High density triglyceride lipoprotein lipoprotein cholesterol cholesterol Ivey 2013 Enzymatic method: Calculated using Ultra HDL assay: Enzymatic method: Architect c16000 Friedewald equation: Architect c16000 Architect c16000 Analyser. Cholesterol LDLC * : T Chol – Analyser. Ultra HDL Analyser. Reagent (Abbott (HDLC + TG [÷2.19]) Reagent (Abbott Triglyceride Reagent Diagnostics, Abbott mmol/L Diagnostics, Abbott (Abbott Diagnostics, Laboratories, Abbott Laboratories, Abbott Abbott Laboratories, Park, IL 60064, USA) Park, IL 60064, USA) Abbott Park, IL 60064, USA) Jauhiainen Enzymatic method: Friedewald 1972 Enzymatic method: Enzymatic method: 2005 ND ND ND Jones 2012 Enzymatic method: Enzymatic method: Enzymatic method: Enzymatic method: ND ND ND ND Jones 2012a Measured using Measured using Measured using Measured using dimension RxL dimension RxL dimension RxL dimension RxL biochemistry analyser biochemistry analyser biochemistry analyser biochemistry analyser using reagent kits using reagent kits using reagent kits using reagent kits (Dade Behring, (Dade Behring, (Dade Behring, (Dade Behring, Siemens, Munich, Siemens, Munich, Siemens, Munich, Siemens, Munich, Germany) Germany) Germany) Germany) Kekkonen Enzymatic method: Friedewald 1972 Enzymatic method: Enzymatic method: 2008 Roche Diagnostics. Roche Diagnostics. Roche Diagnostics. Measured using an Measured using an Measured using an autoanalyser autoanalyser autoanalyser (Roche/Hitachi 912 (Roche/Hitachi 912 (Roche/Hitachi 912 Automatic Analyzer) Automatic Analyzer) Automatic Analyzer) Klein 2008 Enzymatic method: Precipitation method: Preciptation method: Enzymatic method: cholesterol electrophoretic magnesium and glycero-phosphate oxidase-peroxidase technique (Immuno phosphorwolframate. oxidase-peroxidase (Beckman, Krefeld, AG, Heildelberg, Enzymatic method: (Beckman, Krefeld, Germany). Germany). cholesterol Germany). Measured oxidase-peroxidase using CX 7 Synchron (Beckman, Krefeld, Beckman (Beckman, Germany). Krefeld, Germany). Larsen 2006 CHOD-PAP (no. Friedewald 1972 HDLC plus second GPO-PAP (no. 2016630) method. generation (no. 2016648) method. 3030024) method. Lewis 2005 ND ND ND ND Lin 1989 Boehringer Mannheim Boehringer Mannheim Boehringer Mannheim Boehringer Mannheim Diagnostics / Hitachi Diagnostics / Hitachi Diagnostics / Hitachi Diagnostics / Hitachi system H-704 system H-704 system H-704 system H-704 (Boehringer (Boehringer (Boehringer (Boehringer Mannheim Mannheim Mannheim Mannheim Diagnostics Division, Diagnostics Division, Diagnostics Division, Diagnostics Division, Indianapolis, IN) Indianapolis, IN) Indianapolis, IN) Indianapolis, IN) Mizushima ND N/I ND ND 2004 Rizkalla 2000 Labintest kit (Aix-en N/I Boehringer Mannheim Biomerieux kit Provence, France). kit (Meylan, France). (Marcy-l'Etoile, France)

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 145

Study Total cholesterol Low density High density triglyceride lipoprotein lipoprotein cholesterol cholesterol Sadrzadeh- Enzymatic method: Enzymatic method: Enzymatic method: Enzymatic method: Yeganeh 2010 cholesterol ND (Parsazmin kit, ND (Parsazmin kit, glycerol phosphate oxidase-p-aminophen DiaSys, Germany). DiaSys, Germany). oxidase-p-aminophen azon (CHOD-PAP) azon (GPO-PAP) (Parsazmin kit, (Parsazmin kit, DiaSys, Germany). DiaSys, Germany). Simons 2006 ND Friedewald 1972 ND ND Songisepp ND ND ND ND 2012 1 Songisepp ND ND ND ND 2012 2 Usinger 2010a ND ND ND ND Xiao 2003 7450 Automatic Friedewald 1972 7450 Automatic 7450 Automatic Analyzer (Hitachi Co. Analyzer (Hitachi Co. Analyzer (Hitachi Co. Ltd. Tokyo, Japan) Ltd. Tokyo, Japan) Ltd. Tokyo, Japan) Footnotes: ND: not defined; N/I: not investigated a Non-fasting blood samples used for determining serum lipid profile NB: Cardiovascular events, all-cause mortality, severe/serious adverse events, socioeconomic cost and health related quality of life were not assessed by any of the studies

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 146

9.16.8 Appendix H: Adverse events

Study Intervention Randomised Mild adverse events Left study due to adverse events [N] [N (%) ] [N (%)] Agerholm- I1 (G) 16 - - Larsen 2000 I2 (StLa) 16 - - I3 (StLr) 15 1 (7%) 1 (7%) (nausea and constipation) C (PY) 16 1 (6%) 1 (6%) (nausea and constipation) all: 63 2 (3%) 2 (3%) Anderson I1 - - 1999 C1 - - all: 40 - - Ataie-Jafari I1 - - 2009 C1 - - all: 14 - - de-Roos 1999 I1 39 - - C1 39 - - all: 78 1 (1%) 1 (1%) (gastrointestinal complaint) Ejtahed 2011 I1 30 - - C1 30 - - all: 60 - - Fabian 2006 I1 17 - - C1 16 - - all: 33 - - Fuentes 2013 I1 30 - - C1 30 - - all: 30 - - Hata 1996 I1 17 - - C1 13 - - all: 30 - - Hlivak 2005 I1 - - C1 - - all: 38 - - Inoue 2003 I1 18 - - C1 17 - - all: 39 - -

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 147

Study Intervention Randomised Mild adverse events Left study due to adverse events [N] [N (%) ] [N (%)] Ivey 2013 I1 38 - - I2 32 - - I3 39 - - C1 39 - - all: 148 - - Jauhiainen I1 53 - - 2005 C1 55 - - all: 108 Jones 2012 I1 56 - - C1 58 - - all: 114 - - Jones 2012a I1 66 - - C1 61 - - all: 127 - - Kekkonen I1 - - - 2008 C1 - - - all: 26 - - Mizushima I1 23 - - 2004 C1 23 - - all: 46 - - Rizkalla 2000 I1 12 - - C1 12 - - all: 24 - - Sadrzadeh- I1 29 - - Yeganeh 2010 C1 30 - - all: 59 - - Simons 2006 I1 21 2 (10%) - (constipation and increased flatus) C1 23 1 (4%) - (constipation and increased flatus) all: 44 3 (7%) -

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 148

Study Intervention Randomised Mild adverse events Left study due to adverse events [N] [N (%) ] [N (%)] Songisepp I1 - - - 2012 1 C1 - - - all: 13 3 (25%) - (flatulence) Songisepp I1 - - - 2012 2 C1 - - - all: 18 - - Usinger 2010a I1 30 1 (3%) - (mild gastrointestinal pain and bloating) I2 30 - - C1 - - - C2 - - - all: 90 1 (1%) - Xiao 2003 I1 16 5 (31%) - (increased faecal frequency) C1 16 1 (6%) - (increased faecal frequency) all: 32 6 (19%) -

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 149

9.16.9 Appendix 12: Survey of authors providing information on trials

Characteristic Study author Study Study author asked for Study author provided data contacted author additional information replied Lewis 2005 Yes Yes Yes: Request for mean No: geometric values are the only and SD of change data available. Author unable to variables access original dataset

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 150

9.17 CHAPTER 9 STUDIES AND REFERENCES

1. Pannucci CJ, Wilkins EG. Identifying and avoiding bias in research. Plastic and reconstructive surgery 2010;126(2):619.

2. Cochrane Collaboration. Internet: http://www.cochrane.org/.

3. World Health Organisation. Global status report on noncommunicable diseases 2010: description of the global burden of NCDs, their risk factors and determinants. World Health Organisation, Geneva 2011.

4. Heidenreich PA, Trogdon JG, Khavjou OA, Butler J, Dracup K, Ezekowitz MD, Finkelstein EA, Hong Y, Johnston SC, Khera A, et al. AHA policy statement: forecasting the future of cardiovascular disease in the United States: a policy statement from the American Heart Association. Circulation 2011;123:933-44.

5. Lloyd-Jones D, Adams RJ, Brown TM, Carnethon M, Dai S, Simone GD, Ferguson T, Ford E, Furie K, Gillespie C, et al. AHA statistical update: heart disease and stroke statistics - 2010 update: a report from the American Heart Association. Circulation 2010;121:e46-e215.

6. World Health Organisation. Global health risks: mortality and burden of disease attributable to selected major risks. World Health Organisation, Geneva 2009.

7. Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001;285(19):2486-97.

8. Roger V, Go A, Lloyd-Jones D, Adams R, Berry J, Brown T, et al. Heart disease and stroke statistics 2011 update: a report from the American Heart Association. Circulation 2011;123(4):e18-e209.

9. Law M, Wald N, Thompson S. By how much and how quickly does reduction in serum cholesterol concentration lower risk of ischaemic heart disease? BMJ 1994;308(6925):367-72.

10. Cohen J. A population-based approach to cholesterol control. American Journal of Medicine 1997;102(2):23-5.

11. Ghandehari H K-BSWND. Prevalence and extent of dyslipidemia and recommended lipid levels in US adults with and without cardiovascular comorbidities: The National Health and Nutrition Examination Survey 2003-2004. American Heart Journal 2008;156:112-9.

12. Hao Q, Lu Z, Dong BR, Huang CQ, Wu T. Probiotics for preventing acute upper respiratory tract infections. Cochrane Database of Systematic Reviews 2011(9).

13. Allen SA, Martinez EG, Gregorio GV, Dans LF. Probiotics for treating acute infectious diarrhoea. Cochrane Database of Systematic Reviews 2010(11).

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 151

14. Bernaola Aponte G, Bada Mancilla CA, Carreazo Pariasca NY, Rojas Galarza RA. Probiotics for treating persistent diarrhoea in children. Cochrane Database of Systematic Reviews 2010(11).

15. Schrezenmeir J, de Vrese M. Probiotics, prebiotics, and synbiotics: approaching a definition. The American Journal of Clinical Nutrition 2001;73(2):361S-4S.

16. Joint FAO/WHO Expert Consultation on evaluation of health and nutritional properties of probiotics in food. Health and nutritional properties of probiotics in food including powder milk with live lactic acid bacteria. World Health Organisations, Geneva, 2001.

17. Shortt C. The probiotic century: historical and current perspectives. Trends in Food Science and Technology 1999;10(12):411-7.

18. Campbell-Platt G. Fermented foods - a world perspective. Food Research International 1994;27(3):253-7.

19. Commission of the European Communities. Report from the commission to the council on the evolution of the market in milk products and competing products. Commission of the European Communities, 2007.

20. Anukam K, Reid G. Probiotics: 100 years (1907-2007) after Elie Metchnikoff's observation. Communicating Current Research and Educational Topics and Trends in Applied Microbiology 2008;21:166-74.

21. Marteau PR, Vrese Md, Cellier CJ, Schrezenmeir J. Protection from gastrointestinal diseases with the use of probiotics. American Journal of Clinical Nutrition 2001;73(2):430S-6S.

22. Mann G, Spoerry A. Studies of a surfactant and cholesterolemia in the Maasai. american Journal of Clinical Nutrition 1974;27(5):242-9.

23. Joint Fao/Who Working Group on Drafting Guidelines for the Evaluation of Probiotics in Food. Guidelines for the evaluation of probiotics in food: report of a Joint FAO/WHO Working Group on Drafting Guidelines for the Evaluation of Probiotics in Food. World Health Organization, 2002.

24. McFarland. Systematic review and meta-analysis of Saccharomyces boulardii in adult patients. World Journal of Gastroenterology 2010;16(14):2202-22.

25. Cardona M, Vanay V, Midvedt T, Norin E. Probiotics in gnotobiotic mice: conversion of cholesterol to coprostanol in vitro and in vivo and bile acid deconjugation in vitro. Microbial Ecology in Health and Disease 2000;12(4):219-24.

26. McAuliffe O, Cano R, Klaenhammer T. Genetic analysis of two bile salt hydrolase activities in Lactobacillus acidophilus NCFM. Applied Environmental Microbiology 2005;71(8):4925-9.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 152

27. Moser S, Savage D. Bile salt hydrolase activity and resistance to toxicity of conjugated bile salts are unrelated properties in Lactobacilli. Applied Environmental Microbiology 2001;67(8):3476-80.

28. Pereira D, McCartney A, Gibson G. An in vitro study of the probiotic potential of a bile-salt-hydrolyzing Lactobacillus fermentum strain, and determination of its cholesterol-lowering properties. Applied Environmental Microbiology 2003;69(8):4743-52.

29. Lye H, Rusul G, Liong M. Removal of cholesterol by lactobacilli via incorporation and conversion to coprostanol. Journal of Dairy Science 2010;93(4):1383-92.

30. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis Jpa et al. The PRISMA statement for reporting systematic and meta-analyses of studies that evaluate interventions: explanation and elaboration. PLoS Med 1999;6(7):1-28.

31. Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011;343:d5928-d.

32. Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. Cochrane Library, 2011.

33. Kirkham JJ, Dwan KM, Altman DG, Gamble C, Dodd S, Smyth R, et al. The impact of outcome reporting bias in randomised controlled trials on a cohort of systematic reviews. BMJ 2010;340:c365-c.

34. Hrobjartsson A, Thomsen AS, Emanuelsson F, Tendal B, Hilden J, Boutron I, et al. Observer bias in randomized clinical trials with measurement scale outcomes: a systematic review of trials with both blinded and nonblinded assessors. Canadian Medical Association Journal 2013;Epub ahead of print.

35. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in Medicine 2002;21:1539-58.

36. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analysis. BMJ 2003;327:557-60.

37. Sterne JA, Sutton AJ, Ioannidis JP, Terrin N, Jones DR, Lau J, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ 2011;343:d4002-d.

38. Ivey Kl, Hodgson JM, Kerr DA, Prince RL. The effects of yoghurt and its probiotics on features of the metabolic syndrome, intestinal function, bone metabolism and quality of life; a randomised controlled trial. Unpublished, 2013.

39. Songisepp E. Hutt P. Ratsep M. Shkut E. Koljalg S. Truusalu K. Stseptetova J. Smidt I. Kolk H. Zagura M. Mikelsaar M. Study 1: Saftey of a probiotic cheese

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 153

containing Lactobacillus plantarum Tensia according to a variety of health indices in different age groups. Journal of Dairy Science 2012;95(10):5495-509.

40. Hlivak PJ, Odraska J, Ferencik M, Ebringer L, Johnova E, Mikes Z. One-year application of probiotic strain Enterococcus faecium M-74 decreases serum cholesterol levels. Bratisl Lek Listy 2005;106(2):67-72.

41. Ataie-Jafari AL. Cholesterol-lowering effect of probiotic yogurt in comparison with ordinary yogurt in mildly to moderately hypercholesterolemic subjects. Annals of Nutrition and Metabolism 2009;54(1):22-7.

42. Ejtahed HS, Mohtadi-Nia J, Homayouni-Rad A, Niafar M, Asghari-Jafarabadi M, Mofid V, et al. Effect of probiotic yogurt containing Lactobacillus acidophilus and Bifidobacterium lactis on lipid profile in individuals with type 2 diabetes mellitus. Journal of Dairy Science 2011;94(7):3288-94.

43. Sadrzadeh-Yeganeh HEI. The effects of probiotic and conventional yoghurt on lipid profile in women. British Journal of Nutrition 2010;103(12):1778-83.

44. Songisepp E, Hutt P, Ratsep M, Shkut E, Koljalg S, Truusalu K, Stseptetova J, Smidt I, Kolk H, Zagura M, et al. Study 2: Saftey of a probiotic cheese containing Lactobacillus plantarum Tensia according to a variety of health indices in different age groups. Journal of Dairy Science 2012;95(10):5495-509.

45. Fuentes MC, Lajo T, Carrion JM, Cune J. Cholesterol-lowering efficacy of Lactobacillus plantarum CECT 7527, 7528 and 7529 in hypercholesterolaemic adults. British Journal of Nutrition 2013;109(10):1866-72.

46. Fabian EE. Influence of daily consumption of probiotic and conventional yoghurt on the plasma lipid profile in young healthy women. Annals of Nutrition and Metabolism 2006;50(4):387-93.

47. Mizushima SO. Randomized controlled trial of sour milk on blood pressure in borderline hypertensive men. American Journal of Hypertension 2004;17(8):701-6.

48. Rizkalla SWL. Chronic consumption of fresh but not heated yogurt improves breath-hydrogen status and short-chain fatty acid profiles: A controlled study in healthy men with or without lactose maldigestion. american Journal of Clinical Nutrition 2000;72(6):1474-9.

49. Xiao JZK. Effects of milk products fermented by Bifidobacterium longum on blood lipids in rats and healthy adult male volunteers. Journal of Dairy Science 2003;86(7):2452-61.

50. Jauhiainen TV. Lactobacillus helveticus fermented milk lowers blood pressure in hypertensive subjects in 24-h ambulatory blood pressure measurement. American Journal of Hypertension 2005;18(12):1600-5.

51. Agerholm-Larsen LR. Effect of 8 week intake of probiotic milk products on risk factors for cardiovascular diseases. European Journal of Clinical Nutrition 2000;54(4):288-97.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 154

52. de-Roos NM, Schouten G, Katan MB. Yoghurt enriched with Lactobacillus acidophilus does not lower blood lipids in healthy men and women with normal to borderline high serum cholesterol levels. European Journal of Clinical Nutrition 1999;53:277-80.

53. Usinger LJ. The antihypertensive effect of fermented milk in individuals with prehypertension or borderline hypertension. Journal of Human Hypertension 2010;24(10):678-83.

54. Hata YY. A placebo-controlled study of the effect of sour milk on blood pressure in hypertensive subjects. american Journal of Clinical Nutrition 1996;64(5):767-71.

55. Inoue KS. Blood-pressure-lowering effect of a novel fermented milk containing -aminobutyric acid (GABA) in mild hypertensives. European Journal of Clinical Nutrition 2003;57(3):490-5.

56. Jones ML, Martoni CJ, Parent M, S P. Cholesterol-lowering efficacy of a microencapsulated bile salt hydrolase-active Lactobacillus reuteri NCIMB. British Journal of Nutrition 2012;107(10):1505-13.

57. Kekkonen RAS. Effect of probiotic Lactobacillus rhamnosus GG intervention on global serum lipidomics profiles in healthy adults. World Journal of Gastroenterology 2008;14(20):3188-94.

58. Anderson JW, Gilliland SE. Effect of fermented milk (yogurt) containing Lactobacillus acidophilus L1 on serum cholesterol in hypercholesterolemic humans. Journal of the American College of Nutrition 1999;18(1):43-50.

59. Simons LAA. Effect of Lactobacillus fermentum on serum lipids in subjects with elevated serum cholesterol. Nutrition, Metabolism and Cardiovascular Diseases 2006;16(8):531-5.

60. Jones ML, Martoni CJ, Pietro ED, Simon RR, Prakash S. Evaluation of clinical safety and tolerance of a Lactobacillus reuteri NCIMB 30242 supplement capsule: a randomized control trial. Regulatory Toxicology and Pharmacology 2012;63:313-20.

61. Szymanska E, van Dorsten FA, Troost J, Paliukhovich I, van Velzen EJJ, Hendriks MMWB, Trautwein EA, van Duynhoven JPM, Vreeken RJ, Smilde AK. A lipidomic analysis approach to evaluate the response to cholesterol-lowering food intake. Metabolomics 2012;8:894-906.

62. Briel M, Ferreira-Gonzalez I, You JJ, Karanicolas PJ, Akl EA, Wu P, Blechacz B, Bass, er D, Wei X, et al. Association between change in high density lipoprotein cholesterol and cardiovascular disease morbidity and mortality: systematic review and meta-regression analysis. British Medical Journal 2009;338.

63. Chambless LE, Folsom AR, Davis V, Sharrett R, Heiss G, Sorlie P, Szklo M, Howard G, Evans GW. Risk factors for progression of common carotic atherosclerosis: the Atherosclerosis Risk in Communities Study, 1987 - 1998. American Journal of Epidemiology 2002;155(1):38-47.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 155

64. Pedersen TR, Olsson AG, Faergeman O, Kjekshus J, Wedel H, Berg K, Wilhelmsen L, Haghfelt T, Thorgeirsson G, Pyorala K, et al. Lipiprotein changes and reduction in the incidence of major coronary heart disease events in the Scandinavian Simvastatin Survival Study (4S). Circulation 1998;97:1453-60.

65. Al-Sheraji SH, Ismail A, Manap MY, Mustafa S, Yusof RM, Hassan FA. Hypocholesterolaemic effect of yoghurt containing Bifidobacterium pseudocatenulatum G4 of Bifidobacterium longum BB536. Food Chemistry 2012;135:356-61.

66. Allen AJ, Wareham K, Bradley C, Harris W, Dhar A, Brown H, Foden A, Cheung WY, Gravenor MB, Plummer S, et al. A multicentre randomised controlled trial evaluating lactobacilli and bifidobacteria in the prevention of antibiotic-associated diarrhoea in older people admitted to hospital: the PLACIDE study protocol. BMC Infectious Diseases 2012;12(108).

67. Alonso A, Zozaya C, Vazquez Z, Alfredo Martinez J, Martines-Gonzalez MA. The effect of low-fat versus whole-fat dairy product intake on blood pressure and weight in young normotensive adults. Journal of Human Nutrition and Dietetics 2009;22:336-42.

68. Asemi Z, Samimi M, Tabasi Z, Talebian P, Azarbad Z, Hydrazadeh Z, Esmaillzadeh. Effect of daily consumption of probiotic yoghurt on lipid profiles in pregnant women: a randomised controlled clinical trial. Journal of Maternal-Fetal and Neonatal Medicine 2012;25(9).

69. Ashar MN, Prajapati JB. Verification of hypocholesterolemic effect of fermented milk on human subjects with different cholesterol levels. Folia Microbiologica 2000;45(3):263-8.

70. Chang BJ, Park SU, Jang YS, Ko SH, Joo NM, Kim SI, Kim CH, Chang DK. Effect of functional yogurt NY-YP901 in improving the trait of metabolic syndrome. European Journal of Clinical Nutrition 2011;65:1250-5.

71. Chralton KE, Steyn K, Levitt NS, Peer N, Jonathan D, Gogela T, Rossouw K, Gwebushe N, Lombard CJ. A food-based dietary strategy lowers blood pressure in a low socio-economic setting: a randomised study in South Africa. Public health nutrition 2008;11(12):1397-406.

72. De Leeuw PW, Van Der Zander K, Kroon AA, Rennenberg RMW, Koning MMG. Dose-dependent lowering of blood pressure by dairy peptides in mildly hypertensive subjects. Blood Pressure 2009;18:44-50.

73. Diop L, Guillou S, Durand H. Probiotic food supplement reduces stress-induced gastrointestinal symptoms in volunteers: a double blind, placebo controlled, randomized trial. Nutrition Research 2008;28:1-5.

74. Enck P, Zimmermann K, Menke G, Muller-Lissner S, Martens U, Klosterhalfen S. A mixture of Escherichia coli (DSM 17252) and Enterococcus faecalis (DSM 16440) for treatment of the irritable bowel syndrome: a randomized controlled trial with primary care physicians. Neurogastroenterology and Motility 2008;20:1103-9. Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 156

75. Engberink MF, Schouten EG, Kok FJ, van Mierlo LAJ, Brouwer IA, Geleijnse JM. Lactotripeptides show no effect on human blood pressure: results from a double-blind randomized controlled trial. Hypertension 2008;51(2):399-405.

76. Gemeier C, Hackl E, Kedenko I, Leitner M, Kautzky-Willer A, Kedenko L, Paulweber B. Effects of the probiotic drink yakult light on parameters of lipid and glucose metabolism in patients with impaired metabolic function. Journal of Diabetes 2013;5:103-4.

77. Hoo-Kil J, Kim ER, Ji GE, Park JH, Cha SK, SL J. Comparative evaluation of probiotic activities of Bifidobacterium longum MK-G7 with commercial Bifidobacteria strains. Journal of Microbiology and Biotechnology 2000;10(2):147-53.

78. Intorre F, Foddai MS, Azzini E, Martin B, Montel MC, Catasta G, Toti E, Finotti E, Palomba L, Venneria E, et al. Differential effect of cheese fatty acid composition on blood lipid profile and redox status in normolipidemic volunteers: a pilot study. International Journal of Food Sciences and Nutrition 2011;62(6):660-9.

79. Ishikawa H, Akedo I, Umesaki Y, Tanaka R, Imaoka A, Otani T. Randomized controlled trial of the effect of Bifidobacteria-fermented milk on ulcerative colitis. Journal of the American College of Nutrition 2002;22(1):56-63.

80. Jayakumar S, Carbonneau M, Hotte N, Befus D, St Laurent C, Owen R, McCarthy M, Madsen K, Baily R, Ma M, et al. Probiotics do not reduce protal pressure in patients with decompensated cirrhosis. Hepatology 2012;56:744-.

81. Jones ML, Martoni CJ, Pietro ED, Simon RR, Prakash S. Evaluation of clinical safety and tolerance of a Lactobacillus reuteri NCIMB 30242 supplement capsule: a randomized control trial. Regulatory Toxicology and Pharmacology 2012;63:323-0.

82. Kajander K, Myllyluoma E, Rajilic-Stojanovic M, Kyronpalo S, Rasmussen M, Jarvenpaa S, Zoetendal EG, De Vos WM, Vapaatalo H, Korpela R. Clinical trial: multispecies probiotic supplementation alleviates the symptoms of irritable bowel syndrome and stabilizes intestinal microbiota. Alimentary Pharmacology and Therapeutics 2008;27:48-57.

83. Karlsson C, Ahrne S, Molin G, Berggren A, Palmquist I, Fredrikson GN, B J. Probiotic therapy with men with incipient arteriosclerosis initiates increased bacterial diversity in colon: a randomized controlled trial. Atherosclerosis 2010;208:228-33.

84. Kim EK, An SY, Lee MS, Kim TH, Lee HK, Hwang WS, Choe SJ, Kim TY, Han SJ, Kim HJ, et al. Fermented kimchi reduces body weight and improves metabolic parameters in overweight and obese patients. Nutrition Research 2011;31:436-43.

85. Leber B, Tripolt NJ, Eder M, Wascher TC, Pieber TR, Stauber R, Sourij H, Oettl K, Stadbauer V. The influence of probiotic supplementation on gut permeability

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 157

in patients with metabolic syndrome: an open label, randomized pilot study. European Journal of Clinical Nutrition 2012;66:1110-5.

86. Lee YM, Skurk T, Hennig M, H. H. Effect of a milk drink supplemented with whey peptides on blood pressure in patients with milk hypertension. European Journal of Clinical Nutrition 2007;46:21-7.

87. Lewis SJ, Burmeister S. A double-blind placebo-controlled study of the effects of Lactobacillus acidophilus on plasma lipids. European Journal of Clinical Nutrition 2005;59(6):776-80.

88. Linderborg KM, Jarvinen R, Lehtonen HM, Viitanen M, Kallio HPT. The fiber and/or polyphenols present in lingonberries null the glycaemic effect of the sugars present in the berries when consumed together with added glucose in healthy human volunteers. Nutrition Research 2012;32:471-8.

89. Malaguarnera M, Vacante M, Antic T, Giordano M, Chisari G, Acquaviva R, Mastrojeni S, Malaguarnera G, Mistretta A, Voli GL, et al. Bifidobacterium longum with fructo-oligosaccharides in patients with non alcoholic steatohepatitis. Digestive Diseases and Sciences 2012;57:545-53.

90. Marotta F, Yadav H, Kumari A, Catanzaro R, Jain S, Polimeni A, Lorenzetti A, Soresi V. Cardioprotective effect of a biofermented nutraceutical on endothelial function in healthy middle-aged subjects. Rejuvenation Research 2012;2(178-181).

91. Massey LK. Effect of Changing Milk and Yogurt Consumption on Human Nutrient Intake and Serum Lipoproteins1,2,3. Journal of Dairy Science 1984;67(2):255-62.

92. Nestel PJ, Pally S, MacIntosh GL, Greeve MA, Middleton S, Jowett J, Meikle PJ. Circulating inflammatory and atherogenic biomarkers are not increased following single meals of dairy foods. European Journal of Clinical Nutrition 2012;66:25-31.

93. Neyestani TR, Djazayery A, Shab-Bidar S, Eshraghian MR, Kalayi A, Shariatsadeh N, Khalaji N, Zahedirad M, Gharavi A, Houshiarrad A, et al. Vitamin D receptor Fok-I polymorphism modulates diabetic host response to vitamin D intake. Diabetes Care 2013;36:550-6.

94. Pereg D, Kotliroff A, Gadoth N, Hadary R, Lishner M, Kitay-Cohen Y. Probiotics for patients with compensated liver cirrhosis: a double-blind placebo-controlled study. Nutrition 2011;27:177-81.

95. Pitnus S, Murru E, Carta G, Cordeddu L, Batetta B, Accossu S, Pistis D, Uda S, Ghiani ME, Mele M, et al. Sheep cheese naturally enriched with a-linolenic, conjugated linolenic and vaccenic acids improves the lipid profile and reduces anandamide in the plasma of hypercholesterolaemic subjects. British Journal of Nutrition 2013;109:1453-62.

96. Ranganathan N, Friedman EA, Tam P, Rao V, Ranganathan P, Dheer R. Probiotic dietary supplementation in patients with stage 3 and 4 chronic kidney disease: a

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 158

6-month pilot scale trial in Canada. Current Medical Research and Opinion 2009;25(8):1919-30.

97. Rossouw J, Burger E, Van der Vyver P, Ferreira J. The effect of skim milk, yoghurt, and full cream milk on human serum lipids. The American Journal of Clinical Nutrition 1981;34(3):351-6.

98. Selinger CP, Bell A, Cairns A, Lockett M, Sebastian S, Haslam N. Probiotic VSL#3 prevents antibiotic-associated diarrhoea in a double-blind, randomized, placebo-controlled clinical trial. Journal of Hospital Infection 2013;84:159-65.

99. Seppo L, Juahanian T, Korpella R. A fermented milk high in bioactive peptides has a blood pressure-lowering effect in hypertensive subjects. american Journal of Clinical Nutrition 2003;77:326-30.

100. Shab-Bidar S, Neyestani TR, Djazayery A, Eshraghian MR, Houshiarrad A, Gharavi A, Kalayi A, Shariatzadeh N, Zahedirad M, Khalaji N, et al. Regular consumption of vitamin D-fortified yogurt drink (Doogh) improved endothelial biomarkers in subjects with type 2 diabetes: a randomized double-blind clinical trial. BMC Medicine 2011;9(125).

101. Sialvera TE, Koutelidakis AE, Richter DJ, Yfanti G, Kapsokefalou M, Micha R, Goumas G, Diamantopoulos E, Zampelas A. Phytosterol supplementation does not affect plasma antioxidant capacity in patients with metabolic syndrome. International Journal of Food Sciences and Nutrition 2013;64(1):21-7.

102. Trautvetter U, Ditscheid B, Kiehntopf M, Jahreis G. A combination of calcium phosphate and probiotics beneficially influences intestinal lactobacilli and cholesterol metabolism in humans. Clinical Nutrition 2012;31:230-7.

103. Tripolt NJ, Leber B, Blattl D, Eder M, Wonish W, Scharnagl H, Stajakovic T, Obermayer-Pietsch B, Wascher TC, Pieber TR, et al. Effect of supplementation with Lactobacillus casei Shirota on insulin sensitivity, B-cell function, and markers of endothelial function and inflammation in subjects with metabolic syndrome: a pilot study. Journal of Dairy Science (89-95).

104. Troost FJ, van Baarlen P, Lindsey P, Kodde A, de Vos WM, Kleerebezem M, Brummer RJM. Identification of the transcriptional response of human intestinal mucosa to Lactobacillus plantarum WCFSI in vivo. BMC Genomics 2008;9:374-.

105. van der Zaner K, Bots ML, Bak AAA, Koning MMG, de Leeuw PW. Enzymatically hydrolyzed lactotripeptides do not lower blood pressure in mildly hypertensive subjects. american Journal of Clinical Nutrition 2008;88:1697-702.

106. van Mierlo LAJ, Koning MMG, van der Zander K, Draijer R. Lactotripeptides do not lower ambulatory blood pressure in untreated whites: results from 2 controlled multicenter crossover studies. american Journal of Clinical Nutrition 2009;89:617-23.

107. Widhalm HK, Herkner K, Kiefer I, Seemann R, Rieder A, Kunze M, Widhalm K. Effect of daily intake of yoghurt and bread enriched with biologically active

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 9: Page 159

substances on blood lipids and vitamin A in adolescents and young adults. International Journal of Food Sciences and Nutrition 2011;62(1):52-9.

Chapter 9: Probiotic fermented milk or isolated probiotic bacteria for primary prevention of cardiovascular disease in adults (Review) Chapter 10: Page 1

Health benefits of non-nutritive food components

Chapter 10: Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 2

10.1 FOREWORD

Chapters 6-9 report the results of the randomised controlled trial detailed in Appendix A, and Chapter 10 summarised results of 23 randomised controlled trials exploring the effect of probiotic bacteria on serum lipid concentrations. This chapter now returns to epidemiological investigations of diet-disease relationships to explore the association between habitual yoghurt intake and common carotid intima-media thickness; a marker of atherosclerotic vascular disease progression.

One theme of this thesis has been to explore the importance of non-nutritive food components as distinct from major dietary sources. Chapter 2 of this thesis investigated the association of flavonol intake from tea and non-tea sources with atherosclerotic vascular disease mortality. This chapter extends this theme to yoghurt consumption as distinct from non-fermented dairy consumption by exploring the association of dairy product intake, from fermented and unfermented sources, with common carotid artery intima-media thickness; a risk factor for atherosclerotic vascular disease pathogenesis.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 3

10.2 ABSTRACT

American Journal of Clinical Nutrition (2011, 94:1, 234-9) Kerry L Ivey, Lewis JR, Hodgson JM, Zhu K, Dhaliwal SS, Thompson PL, Prince RL

Association between yogurt, milk, and cheese consumption and common carotid artery intima-media thickness and cardiovascular disease risk factors in elderly women

Background: Despite the contribution of dairy foods to total dietary saturated fat intake, available data indicate that dairy consumption may lower the risk of cardiovascular disease.

Objective: The objective of this study was to investigate the relation between consumption of milk, cheese, and yogurt and common carotid artery intima-media thickness (CCA-IMT) in a cohort of elderly women.

Design: Dairy consumption was assessed with a validated food frequency questionnaire in 1080 participants randomly selected from ambulant white women aged >70 y living in Perth, Western Australia. CCA-IMT was assessed by using B-mode carotid ultrasound 3 y later. Cardiovascular disease risk factors, including serum lipids and blood pressure, were assessed at baseline.

Results: Total dairy product, milk, and cheese consumption was not associated with CCA-IMT (P > 0.05), whereas yogurt consumption was negatively associated with CCA-IMT (unadjusted standardized b = -0.081, P = 0.008; baseline risk factor–adjusted standardized b = -0.075, P = 0.015). Participants who consumed < 100 g yogurt/d had a significantly lower CCA-IMT than did participants with lower consumption (unadjusted = -0.024 mm, P = 0.002). This relation remained significant after adjustment for baseline, dietary, and lifestyle risk factors (multivariable analysis = -0.023 mm, P = 0.003).

Conclusion: Increased consumption of yogurt, but not of other dairy products, is associated with a lower CCA-IMT, independent of other risk factors.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 4

10.3 INTRODUCTION

Elevated common carotid artery intima-media thickness (CCA-IMT) has been shown to be predictive for stroke, coronary events including myocardial infarction, and coronary deaths (1-3). Recognized risk factors for cardiovascular disease which have been identified as contributing to CCA-IMT include lifestyle, blood pressure and serum lipids

(4, 5), however investigations examining the link between the intake of whole foods and

CCA-IMT are scarce.

Although the World Health Organization recommends limiting intake of cholesterol, myristic acid and palmitic acid from dairy fat sources to achieve reductions in cardiovascular disease risk (6), beneficial associations between dairy consumption and cardiovascular disease risk factors and outcomes have been identified in several studies

(7, 8). As far back as the early 1970’s, benefits of regular consumption of fermented dairy products on atherosclerotic vascular disease (ASVD) and cardiovascular health have been suggested (9) with evidence of beneficial effects in randomized control trials

(10-12), including a potential reduction in arterial stiffness in formulated fermented milk products containing tripeptides (13, 14)

This study aimed to investigate the association between intake of different dairy sources and CCA-IMT in a population based study of older women from the Calcium Intake

Fracture Outcome Study (CAIFOS) cohort. Potential mediators of ASVD were also explored by assessing the relationships of dairy consumption with recognized risk factors for cardiovascular disease.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 5

10.4 SUBJECTS AND METHODS

10.4.1 Participants

Participants were recruited in 1998 to a 5-year prospective, randomized, controlled trial of oral calcium supplements to prevent osteoporotic fractures (15). Participants were recruited from the Western Australian general population of women aged over 70 years by mail using the electoral roll - a requirement of citizenship. Over 99% of Australians of this age are registered on the roll. Of the 5,586 women who responded to a letter inviting participation, 1,510 women were willing and eligible. Participants were ambulant and did not have any medical conditions likely to influence 5-year survival.

They were excluded if they were receiving bone-active agents, including hormone replacement therapy. From this group 1,500 women were recruited, and of these, 1,080 had complete food frequency questionnaire and carotid ultrasound data. Participants were similar in terms of disease burden and pharmaceutical consumption to the whole population of this age, but they were more likely to be from higher socio-economic groups (16). Informed consent was obtained, and the Human Ethics Committee of the

University of Western Australia approved the study.

10.4.2 Baseline vascular disease risk assessment

The participants provided their previous medical history, and current medications were verified by their general practitioner. These data were coded using the International

Classification of Primary Care – Plus (ICPC-Plus) method (17), allowing aggregation of different terms for similar pathologic entities as defined by the ICD-10 coding system.

These data were then used to determine the presence of pre-existing vascular disease; including ischemic heart disease, heart failure, arrhythmia, stroke (excluding hemorrhagic stroke), and peripheral vascular disease (K74000-99011). The presence of

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 6 baseline vascular disease risk factors were recorded including pre-existing diabetes

(T89001-90009). Smoking status was coded as non-smoker or ex-smoker/current smoker if they had smoked more than 1 cigarette per day for more than 3 months at any time in their life. Cardiovascular medications included beta-blockers, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers,

3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors and anti-platelet agents. Physical activity was assessed using a previously validated questionnaire in which participants reported the time in hours of involvement in up to four sports, recreational activities and other forms of regular physical activity including walking, that were undertaken in the past 3 months. Energy expenditure (kJ/d) for these activities was calculated with the use of published energy costs (18, 19).

10.4.3 Assessment of dairy consumption

A validated semi-quantitative food frequency questionnaire developed by the

Anti-Cancer Council of Victoria was used to assess baseline dietary intake; including milk, cheese and yogurt consumption (20). The process of collection was identical, whereby a research assistant supervised the completion of the questionnaire in small groups. Food models, cups, spoons and charts for frequency were provided. Energy and nutrient intakes were estimated based on frequency of consumption and an overall estimate of usual portion size (21). Frequency of consumption was represented as serves per day, specifically they reported average consumption of milk, cheese and yogurt over the preceding 12 months, using 200 grams as the yogurt serve size, 250 grams as the milk serve size, and 30 grams to represent the average cheese serve size.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 7

10.4.4 Clinical measurements

Common carotid artery intima-media thickness and focal carotid plaques were measured in 2001 (Year 3) using a B-mode carotid ultrasound machine (8.0 MHz linear array transducer fitted to an Acuson Sequoia 512 ultrasound machine) and a standard image acquisition protocol (22). The far walls of the distal 2 cm of the left and right common carotid arteries were examined, and images were taken from 3 different angles

(anterolateral, lateral and posterolateral) to account for asymmetrical wall thickening.

End-diastolic images were recorded and a semi-automated edge-detection software program was used for image analysis. The same technician performed off-line analysis of all images. The mean, maximum, minimum and standard deviation of measurements for each of the three views on both sides was collected. The CCA-IMT from each of the six images was averaged to give an overall mean CCA-IMT. Plaque was defined as a clearly identified area of focal increased thickness ( 1 mm) of the intima-media layer.

A short-term precision study was undertaken using the same combination of technicians. Twenty non-trial subjects were selected and repeat CCA-IMT measurements made between 0 and 31 days apart (mean 10.3 days). The coefficient of variation for the repeat measures, calculated using the root-mean-square method, was

5.98% (23).

Baseline weight was assessed using digital scales with participants wearing light clothes and no shoes. Baseline height was assessed using a stadiometer, and the body mass index was calculated in kg/m2. Serum lipid profiles were obtained for 1,002 (92.8%) of the participants at baseline. Total cholesterol, high-density lipoprotein cholesterol

(HDLC) and triglyceride concentrations were determined using a Hitachi 917 auto analyzer (Roche diagnostics). Low-density lipoprotein cholesterol (LDLC) was

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 8 calculated using Friedewald’s method (24). Blood pressure was measured on the right arm with a mercury column Manometer using an adult cuff after the patient had been seated and resting for at least 5 minutes; the average of 3 such measurements was obtained.

10.4.5 Statistics

Before commencing statistical analysis, a pre-specified analytical protocol was produced. Baseline ASVD risk factors were tested for differences using ANOVA or chi-squared test where appropriate. The relationships between total dairy, milk, cheese and yogurt consumption and CCA-IMT, HDLC, LDLC and systolic and diastolic blood pressures were examined in regression analysis. Unadjusted, age adjusted and baseline risk factor-adjusted models. The baseline risk factor-adjusted model included age, body mass index, energy intake, energy expended in physical activity, use of vascular medication, diabetes, and history of both vascular disease and smoking (4) .

Yogurt consumption was then dichotomized into 3 groups for further analysis by analysis of variance (ANOVA) with Tukey’s test or a chi-squared test where appropriate. Post hoc comparisons were only made after the main effect of the factor was found to be significant in the ANOVA in the age-adjusted and multivariable analysis. Stepwise linear regression was then used to identify covariates which were associated with CCA-IMT. These covariates were then included in the ANCOVA model in Table 5.

Atherosclerotic plaque odds ratios (OR) and 95% confidence intervals (CI) were obtained using binary logistic regression. The data was analyzed using SPSS (version

15; SPSS Inc, Chicago, IL) and SAS (Version 9, SAS Institute Inc., Chicago, IL).

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 9

10.5 RESULTS

The baseline characteristics of participants are shown in Table 1. The mean CCA-IMT was 0.780 ± 0.128 mm, ranging from 0.423 to 2.078mm.Total milk cheese and yogurt consumption ranged from 0 to 1,223 g/day (mean 555 ± 239 g/day), with milk on average representing 82.7% ( mean 459 ± 220 g/day), yogurt representing 14.6%

(mean 81 ± 97 g/day) and cheese the remaining 2.7% (mean 15 ± 13 g/day). The average fat content of milk consumed was 0.73% fat. Of the 1,033 subjects consuming cheese, 26.7% consumed low fat cheese.

10.5.1 Dairy product intake and CCA-IMT (Table 2)

High yogurt consumption, but not milk or cheese consumption, was associated with lower CCA-IMT, and higher HDLC. After adjustment for pre-specified baseline risk factors the effect on HDLC was not significant. There were no associations with other cardiovascular risk factors including LDLC, systolic and diastolic blood pressure.

Total dairy intake was not significantly associated with CCA-IMT by linear regression in unadjusted or adjusted analyses. However total dairy intake was significantly associated with HDLC in unadjusted and adjusted analyses; ß = 0.073, P = 0.025 and ß

= 0.069, P = 0.032, respectively.

10.5.2 Yogurt and CCA-IMT (Figure 1, Table 3-5)

The effect of yogurt was explored further by dividing the participants into 3 groups; low

(< 100 g/day), moderate (100 – 199 g/day) and high yogurt consumption (≥ 200 g/day).

In an unadjusted analysis, the CCA-IMT of moderate and high yogurt consuming participants was 3.1% and 3.0%, respectively, lower than participants with a low yogurt intake (Figure 1).

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 10

Participants with yogurt consumption greater than 100 g/day (moderate and high yogurt consumers) had significantly lower CCA-IMT than those with yogurt consumption of less than 100 g/day (unadjusted mean difference = - 0.024 mm, P = 0.002; multivariable analysis mean difference = - 0.023 mm, P = 0.003). Another cardiovascular risk factor selected in the preplanned analysis that was associated with yogurt consumption was

HDLC (Table 3). High yogurt consumers also consumed significantly more protein, carbohydrate, fiber, fish, vitamin E, and fruit (Table 4).

The relationship between yogurt consumption and CCA-IMT was examined in stepwise linear regression models to account for potential covariance of independent variables.

The multivariable candidate variables included age, body mass index, energy intake, energy expended in physical activity, use of vascular medication, diabetes, history of both vascular disease and smoking, HDLC, consumption of protein, carbohydrate, fiber, fish, vitamin E, and fruit at baseline. The results of ANCOVA analysis (Table 5) showed that in addition to yogurt intake (P = 0.019), which remained highly significant, the most parsimonious model consisted of age (P = 0.002), smoked ever (P = 0.033) and energy expended in physical activity (P = 0.044).

Dividing the participants into tertiles of yogurt intake (lower tertile: 0 – 4 g/day, middle tertile: 4 – 100 g/day, upper tertile: 100 - 600 g/day) produced similarly significant relationships between yogurt intake and CCA-IMT (data not shown).

10.5.3 Carotid atherosclerotic plaques

The presence of carotid atherosclerotic plaques in the low yogurt consumption group was not significantly different between yogurt consumption groups (data not shown).

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 11

10.6 DISCUSSION

Yogurt consumption was negatively associated with CCA-IMT in this cohort of older women, whilst total dairy, milk and cheese consumption showed no significant association. These findings are supportive of the prospective study by Tavani et al. (25) who found that yogurt consumption, but not milk or cheese consumption, was associated with lower risk of acute myocardial infarction; a clinical outcome of ASVD associated with CCA-IMT (1).

Consumption of more than 100 grams of yogurt per day was associated with a significantly lower CCA-IMT. This relationship was sufficiently robust to remain following adjustment for identified baseline, dietary and lifestyle risk factors. Compared with low yogurt consumers, subjects with moderate yogurt consumption had a 3.1% lower CCA-IMT (0.024 mm difference), with no additional benefit from higher yogurt intakes. This difference is likely to be of clinical significance, as the long-term epidemiology cohort study by Hodis et al. (2) found that a 0.03 mm increase in

CCA-IMT had a relative risk of coronary event of 3.1.

A controlled trial by Kiessling et al. (26) found that daily yogurt consumption increased

HDLC levels. This is consistent with the observed positive association between yogurt consumption and HDLC in the current study, indicating that the food frequency questionnaire was sufficiently accurate to detect this association.

The inclusion of HDLC in linear regression models did not account for the relationship between yogurt consumption and CCA-IMT. Although this may be due to a type two statistical error, another possibility is that the association between yogurt consumption and CCA-IMT may be mediated by other mechanisms in addition to an increased

HDLC. However, as HDLC has been found to be associated with bulbar intima-media

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 12 thickness, but not with CCA-IMT (3), it is possible that HDLC does not mediate the relationship between yogurt consumption and CCA-IMT.

Although an association between yogurt consumption and HDLC was identified, this did not extend to LDLC, triglyceride or blood pressure. This is surprising in view of the fact that antihypertensive compounds in the probiotic species Lactobacillus casei have been identified (27), and fermented milk has been found to significantly lower plasma triglycerides and systolic blood pressure in healthy adult men (28). There was no significant association between yogurt consumption and LDLC which is consistent with the findings of Kiessling et al. (26) who found that despite yogurt consumption improving HDLC levels, there was no significant effect on total cholesterol or LDLC levels.

It should be noted that the causality of the relationship between yogurt consumption and

CCA-IMT cannot be established due to the observational nature of the study. Also, despite the inclusion of baseline, dietary and lifestyle risk factors into statistical models, residual or unmeasured confounders cannot be ruled out. Although the majority of yogurt products consumed in Australia contain probiotic cultures of Lactobacillus acidophilus, Bifidobacteria and Lactobacillus casei (29), identification of causality is further limited by the complexity of yogurt composition, in particular the concentration and species of viable probiotic bacteria. However, the strength of the association is such that despite these factors, the association remains significant even after adjustment for baseline, dietary and lifestyle risk factors.

In this cohort of elderly women, yogurt intake was beneficially associated with

CCA-IMT, suggesting that increased consumption of yogurt, but not other dairy products, may prevent thickening of the common carotid artery intima-media. As the

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 13 relationship remains significant following adjustment for known risk factors, these results do not appear to be due to co-correlation with other baseline, dietary and lifestyle risk factors. Through its role in reducing CCA-IMT, prolonged daily yogurt consumption of at least 100 g/day may play a role in stroke and atherosclerosis prevention. Ultimately, in order to make public health recommendations regarding yogurt intake, further observational and clinical trials are necessary to establish whether moderate to high yogurt consumption can prevent or even ameliorate intima-media thickening of the common carotid artery.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 14

10.7 CHAPTER 10 REFERENCES

1. Bots ML, Hoes AW, Koudstaal PJ, Hofman A, Grobbee DE. Common carotid intima-media thickness and risk of stroke and myocardial infarction: the Rotterdam Study. Circulation 1997;96(5):1432-7. 2. Hodis HN, Mack WJ, LaBree L, et al. The role of carotid arterial intima-media thickness in predicting clinical coronary events. Annals of Internal Medicine 1998;128:262-9. 3. Ebrahim S, Papacosta O, Whincup P, et al. Carotid plaque, intima media thickness, cardiovascular risk factors, and prevalent cardiovascular disease in men and women : the British Regional Heart Study. Stroke 1999;30(4):841-50. 4. Dobs AS, Nieto FJ, Szkio M, et al. Risk Factors for Popliteal and Carotid Wall Thicknesses In the Atherosclerosis Risk in Communities (ARIC) Study. American Journal of Epidemiology 1999;150(10):1055-67. 5. Alagona C, Soro A, Westerbacka J, et al. Low HDL cholesterol concentration is associated with increased intima-media thickness independent of arterial stiffness in healthy subjects from families with low HDL cholesterol. European Journal of Clinical Investigation 2003;33(6):457-63. 6. Joint WHO / FAO Expert Consultation on Diet Nutrition and the Prevention of Chronic Diseases. Diet, nutrition and the prevention of chronic diseases: report of a joint WHO / FAO expert consultation. World Health Organ Tech Rep Ser 916. Geneva: World Health Organisation, 2003. 7. Abbott RD, Curb JD, Rodriguez BL, Sharp DS, Burchfiel CM, Yano K. Effect of dietary calcium and milk consumption on risk of thromboembolic stroke in older middle-aged men : the Honolulu Heart Program. Stroke 1996;27(5):813-8. 8. Oshaug A, Bugge KH, Refsum H. Diet, an independent determinant for plasma total homocysteine. A cross sectional study of Norwegian workers on platforms in the North Sea. European Journal of Clinical Nutrition 1998;52(1):7-11. 9. Mann GV, Spoerry A. Studies of a surfactant and cholesteremia in the Maasai. American Journal of Clinical Nutrition 1974;27(5):464-9. 10. Anderson JW, Gilliland SE. Effect of fermented milk (yogurt) containing Lactobacillus acidophilus L1 on serum cholesterol in hypercholesterolemic humans. Journal of the American College of Nutrition 1999;18(1):43-50. 11. Aihara K, Kajimoto O, Hirata H, Takahashi R, Nakamura Y. Effect of powdered fermented milk with Lactobacillus helveticus on subjects with high-normal blood pressure or mild hypertension. Journal of the American College of Nutrition 2005;24(4):257-65. 12. Hata Y, Yamamoto M, Ohni M, Nakajima K, Nakamura Y, Takano T. A placebo-controlled study of the effect of sour milk on blood pressure in hypertensive subjects. American Journal of Clinical Nutrition 1996;64(5):767-71. 13. Jauhiainen T, Rönnback M, Vapaatalo H, Wuolle K, Kautiainen H, Korpela R. Lactobacillus helveticus fermented milk reduces arterial stiffness in hypertensive subjects. International Dairy Journal 2007;17(10):1209-11.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 15

14. Jauhiainen T, Ronnback M, Vapaatalo H, et al. Long-term intervention with Lactobacillus helveticus fermented milk reduces augmentation index in hypertensive subjects. Eur J Clin Nutr 2010;64(4):424-31. 15. Prince RL, Devine A, Dhaliwal SS, Dick IM. Effects of calcium supplementation on clinical fracture and bone structure: results of a 5-year, double-blind, placebo-controlled trial in elderly women. Archives of Internal Medicine 2006;166(8):869-75. 16. Bruce DG, Devine A, Prince RL. Recreational Physical Activity Levels in Healthy Older Women: The Importance of Fear of Falling. Journal of the American Geriatrics Society 2002;50(1):84-9. 17. Britt H. A new coding tool for computerised clinical systems in primary care-ICPC plus. Australian Family Physician 1997;26:S79. 18. McArdle WD, Katch FI, Katch VL. Energy, nutrition and human performance. Philadelphia: Lea & Febiger, 1991. 19. Pollock ML, Wilmore JH, Fox SM. Health and fitness through physical activity. New York: Wiley, 1978. 20. Allison H, Amanda JP, Wendy JB, Paul I, Graham G. The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Australia New Zealand Journal of Public Health 2000;24(6):576-83. 21. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pacific Journal of Clinical Nutrition 1994;3:19-31. 22. Salonen JT, Salonen R. Ultrasound B-mode imaging in observational studies of atherosclerotic progression. Circulation 1993;87(3 Suppl):II56-65. 23. Bonnick SL, Johnston CC, Jr., Kleerekoper M, et al. Importance of precision in bone density measurements. Journal of Clinical Densitometry 2001;4(2):105-10. doi: JCD:4:2:105 [pii]. 24. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clinical Chemistry 1972;18(6):499-502. 25. Tavani A, Gallus S, Negri E, La Vecchia C. Milk, dairy products, and coronary heart disease. Journal of Epidemiology and Community Health 2002;56(6):471-2. 26. Kiessling G, Schneider J, Jahreis G. Long-term consumption of fermented dairy products over 6 months increases HDL cholesterol. European journal of clinical nutrition 2002;56(9):843-9. 27. Sawada H, Furushiro M, Hirai K, Motoike M, Watanabe T, Yokokura T. Purification and characterization of an antihypertensive compound from Lactobacillus casei. Agric Biol Chem 1990;54(12):3211-9. 28. Kawase M, Hashimoto H, Hosoda M, Morita H, Hosono A. Effect of administration of fermented milk containing whey protein concentrate to rats and healthy men on serum lipids and blood pressure. J Dairy Sci 2000;83(2):255-63. 29. Shortt C. The probiotic century: historical and current perspectives. Trends in Food Science and Technology 1999;10(12):411-7. Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 16

10.8 TABLES AND FIGURES

Chapter 10, Table 1: Baseline demographics and cardiovascular risk factors of participants1

Characteristics

Age (years) 75.1 ± 2.7

BMI (kg/m2) 27.0 ± 4.4

Energy intake (kJ/day) 7273 ± 2355

Physical activity (kJ/day) 595 ± 658

Total vascular medications 644 (59.6)

Previous CVD2 136 (12.6)

Diabetes at baseline 51 (4.7)

Smoked ever 374 (34.6)

HDLC (mmol/L)3 1.45 ± 0.38

LDLC (mmol/L)4 3.67 ± 0.99

Total cholesterol (mmol/L) 5.84 ± 1.09

Triglyceride (mmol/L) 1.56 ± 0.72

Mean CCA-IMT (mm)5 0.78 ± 0.13

Maximum CCA-IMT (mm)5 0.92 ± 0.15

Systolic blood pressure 137.4 ± 18.2 (mmHg) Diastolic blood pressure 73.2 ± 11.0 (mmHg) 1 Results are the mean ± SD or number (%) (n = 1,080). 2 CVD: Cardiovascular disease. 3 HDLC: High density lipoprotein cholesterol (n = 1,002) 4 LDLC: Low density lipoprotein cholesterol (n = 1,002) 5 CCA-IMT: Common carotid artery intima-media thickness.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 17

Chapter 10, Table 2: Relationship of baseline dairy intake (g/day) to cardiovascular disease risk factors1

Milk consumption Cheese consumption Yogurt consumption

Standardized P Standardized P Standardized P ß value ß value ß value

CCA-IMT2 Unadjusted 0.032 0.306 0.060 0.056 -0.081 0.008 Age-adjusted 0.026 0.396 0.057 0.064 -0.080 0.008 Multivariate adjusted3 0.039 0.226 0.062 0.057 -0.075 0.015

HDLC4 Unadjusted 0.046 0.153 0.006 0.849 0.075 0.018 Age-adjusted 0.047 0.151 0.006 0.847 0.075 0.018 Multivariate adjusted3 0.046 0.149 0.017 0.607 0.058 0.059

LDLC4 Unadjusted -0.023 0.484 -0.010 0.763 -0.042 0.181 Age-adjusted -0.021 0.510 -0.009 0.773 -0.043 0.178 Multivariate adjusted3 -0.035 0.300 -0.010 0.760 -0.053 0.102

Systolic blood pressure Unadjusted 0.017 0.586 0.045 0.156 -0.058 0.059 Age-adjusted 0.015 0.627 0.044 0.162 -0.058 0.059 Multivariate adjusted3 0.011 0.725 0.038 0.243 -0.044 0.157

Diastolic blood pressure Unadjusted -0.001 0.962 0.054 0.088 -0.020 0.524 Age-adjusted 0.001 0.979 0.055 0.083 -0.020 0.519 Multivariate adjusted3 -0.003 0.927 0.052 0.112 -0.015 0.631

1Results are presented as standardized ß value by unadjusted, age-adjusted and baseline risk factor-adjusted linear regression (n = 1,080). 2 CCA-IMT: Common carotid artery intima-media thickness 3 Multivariate adjusted for age, body mass index, energy intake, energy expended in physical activity, use of vascular medication, history of cardiovascular disease, diabetes and history of smoking. 4 HDLC: High density lipoprotein cholesterol. LDLC: Low density lipoprotein cholesterol (n = 1,002).

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 18

Chapter 10, Table 3: Baseline, lifestyle and cardiovascular risk factors by yogurt consumption group1

Low Moderate High consumption consumption consumption < 100 g/day 100 - 199g/day ≥ 200 g/day P value

Number (% of all subjects) 628 (58) 245 (23) 207 (19)

Age (years) 75.1 ± 2.7 74.9 ± 2.6 74.9 ± 2.6 0.460

BMI (kg/m2) 27.2 ± 4.5 27.0 ± 4.5 26.5 ± 3.9 0.087

Energy intake (kJ/day) 7324 ± 2307 7200 ± 2650 7203 ± 2132 0.714

Physical activity (kJ/day) 597 ± 685 630 ± 608 548 ± 630 0.417

Total vascular medications 378 (60) 147 (60) 119 (58) 0.782

Previous CVD2 84 (13) 29 (12) 23 (11) 0.641

Diabetes at baseline 27 (4) 14 (6) 10 (5) 0.673

Smoked ever 218 (35) 82 (34) 74 (36) 0.868

HDLC (mmol/L)3 1.4 ± 0.4 1.4 ± 0.4 1.5 ± 0.4 0.009

LDLC (mmol/L) 4 3.7 ± 1.0 3.6 ± 0.9 3.7 ± 0.9 0.089

Total cholesterol (mmol/L) 5.9 ± 1.1 5.7 ± 1.1 5.9 ± 1.0 0.116

Triglycerides (mmol/L) 1.6 ± 0.7 1.6 ± 0.8 1.5 ± 0.6 0.538

Systolic blood pressure (mmHg) 138.3 ± 18.9 136.0 ± 17.0 136.6 ± 17.3 0.180

Diastolic blood pressure (mmHg) 73.2 ± 11.2 73.0 ± 10.6 73.1 ± 11.2 0.964

1 Results are the mean ± SD or number (%) tested by ANOVA or chi squared test where appropriate (n = 1,080). 2CVD: Cardiovascular disease. 3HDLC: High density lipoprotein cholesterol in 1,002 participants. 4LDLC: Low density lipoprotein cholesterol in 1,002 participants.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 19

Chapter 10,Table 4: Baseline dietary factors by yogurt consumption group1

Low Moderate High consumption consumption consumption P value < 100 g/day 100 - 199g/day ≥ 200 g/day

Number (%) 628 (58) 245 (23) 207 (19)

Energy intake (kJ/day) 7324 ± 2307 7200 ± 2650 7203 ± 2132 0.714

Protein (g/day) 77.0 ± 27.8 81.5 ± 25.1 90.2 ± 28.4 <0.001

Carbohydrate (g/day) 187.6 ± 61.9 196.6 ± 58.5 211.6 ± 58.7 <0.001

Fiber (g/day) 22.5 ± 7.8 23.7 ± 8.2 24.8 ± 7.6 <0.001

Total fat (g/day) 65.8 ± 25.3 62.3 ± 24.8 66.8 ± 23.9 0.101

Saturated fat (g/day) 26.2 ± 12.3 24.2 ± 11.6 26.6 ± 11.1 0.052

Vitamin E (mg/day) 6.1 ± 2.1 6.2 ± 2.2 6.5 ± 2.1 0.036

Fish (serves/day) 0.35 ± 0.36 0.43 ± 0.45 0.44 ± 0.36 0.002

Fruit (serves/day) 2.0 ± 1.0 2.3 ± 0.9 2.3 ± 0.9 <0.001

Vegetable (serves/day) 2.8 ± 1.0 2.9 ± 1.0 2.7 ± 0.9 0.368

1 Results are mean ± SD tested by ANOVA (n = 1,080).

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 20

Chapter 10, Table 5: Common carotid artery intima-media thickness (mm) according to yogurt consumption group1

Model Low Moderate High consumption P consumption consumption > 200 g/day value < 100 g/day 100 - 199g/day

Number (%) 628 (58) 245 (23) 207 (19)

Unadjusted 0.789 ± 0.005a 0.764 ± 0.007b 0.765 ± 0.008b 0.008

Age-adjusted 0.789 ± 0.005a 0.765 ± 0.008b 0.766 ± 0.009b 0.012

Multivariable-analysis2 0.788 ± 0.005a 0.766 ± 0.008b 0.765 ± 0.009b 0.013

1 Results are the estimated mean ± SEM tested by ANCOVA (n = 1080). a, b Denotes significantly different values (P < 0.05) by Bonferroni test. 2 Multivariable analysis adjusted for age, smoked ever and energy expended in physical activity.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 10: Page 21

Chapter 10, Figure 1: Proportional reduction in common carotid artery intima-media thickness (CCA-IMT) between low (< 100 g/day) n = 628, moderate (100-199 g/day) n = 245 and high (≥ 200 g/day) yogurt consumption, n = 207.

Results are mean and standard error tested by ANOVA.

* Denotes statistically significantly different by Tukeys post-hoc test from low yogurt consumption at P < 0.05.

Chapter 10:Association between yogurt, milk and cheese consumption and common carotid artery intima-media thickness and cardiovascular risk factors in elderly women Chapter 11: Page 1

Health benefits of non-nutritive food components

Chapter 11: Concluding statements

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Chapter 11: Concluding statements Chapter 11: Page 2 11.1 FOREWORD

This thesis set out to explore the health benefits of consuming specific non-nutritive food components; flavonoids and probiotics. Despite a body of evidence to suggest that non-nutritive food components may add to the health promoting effects of foods, current dietary guidelines focus on nutrient profile of foods and diets, and do not include non-nutritive food components. As such, it was anticipated that the results of this thesis would clarify whether foods high in flavonoids and probiotics should be endorsed as important contributors to a health promoting diet.

This thesis, in accordance with the University of Western Australia Higher Degree by

Research Rules Act, is presented as a series of scientific papers that resulted from the

PhD candidature. Therefore, discussions of study limitations and key findings in relation to relevant literature have been provided in chapters 2 to 10. This thesis chapter will relate study findings to research questions and thesis aims, and outline the theoretical and practical implications of this research.

Chapter 11: Concluding statements Chapter 11: Page 3 11.2 FINDINGS IN RELATION TO RESEARCH QUESTIONS AND THESIS

AIM

This thesis has shown that particular non-nutritive food components are associated with reduced risk of mortality and improved vascular health, and can elicit health benefits. In achieving this aim, specific research questions were posed, and hypotheses to answer these questions were carefully derived. Table 1 summarises specific hypotheses and the research questions to which they pertain.

Flavonols from both tea and non-tea sources were associated with reduced risk of atherosclerotic vascular disease mortality (Chapter 2). Because tea is a major dietary contributor to total flavonol intake, with demonstrated vascular benefits, the findings from this thesis clarified the beneficial association of flavonol consumption beyond the major food source from which it was obtained.

Continuing the exploration of the vascular benefits of flavonoid consumption, the beneficial relationship between proanthocyanidin consumption and renal outcomes was demonstrated in Chapter 3. When compared to those with low proanthocyanidin intake, elderly women with high habitual proanthocyanidin consumption had better renal function and lower risk of adverse renal associated events. The kidneys are highly vascularised organs, and the hypothesis that through improving vascular health proanthocyanidins may improve renal health, was raised.

Following establishment of beneficial associations of flavonoid class intake with atherosclerotic vascular disease mortality and renal events, and the finding that the benefits of flavonols may extend beyond their major dietary contributor, the relationship of total-flavonoid intake with all-cause mortality was explored in Chapter 4. This thesis has been the first to investigate and report that total-flavonoid intake from all classes was beneficially associated with reduced risk of all-cause mortality, irrespective of the

Chapter 11: Concluding statements Chapter 11: Page 4 database used to derive flavonoid intake estimates. In fact, when the United States

Department of Agriculture and Phenol-Explorer flavonoid food composition databases were compared in an epidemiological context (Chapter 5), it was concluded that the databases yielded highly correlated intake estimates for total-flavonoids, flavanols , flavanones and anthocyanidins. However, the poorer correlation between flavonol and flavone intake estimates were due to differences in USDA and PE methodologies.

In addition to epidemiological studies, this thesis also presented data from the randomised controlled trial described in Appendix A, which explored the effect of daily consumption of probiotics from yoghurt and capsules for 6 weeks. The results of this study were complex and varied. It was observed that probiotics from yoghurt were effective at improving bacterial profile of the feaces, whereas probiotics from capsules were not (Chapter 6). Conversely the probiotic therapy appeared to have a small detrimental effect on glycaemic control (Chapter 7), and had no effect on blood pressure or cholesterol concentrations. (Chapter 8).

Despite this clinical trial reporting no effect of probiotics on serum lipid profile, the results of the Cochrane approved systematic review and meta-analysis (Chapter 9) indicate that probiotic bacteria, particularly when supplemented in capsule form, improve total cholesterol and low density lipoprotein cholesterol concentrations. The benefits of probiotic consumption appear to extend beyond the clinical trial setting. In a population based study of elderly postmenopausal women, habitual high probiotic yoghurt consumption was associated with improved common carotid artery intima media thickness; a risk factor for atherosclerotic vascular disease.

Chapter 11: Concluding statements Chapter 11: Page 5 11.3 IMPLICATIONS OF THESIS FINDINGS FOR RESEARCH

This thesis described the beneficial the association of proanthocyanidin intake with renal function and events in elderly postmenopausal women (Chapter 3). Despite strong in vitro and animal model data to support the renoprotective effects of proanthocyanidins (1-5), there is a need for this relationship to be investigated in human clinical trials.

The finding of the beneficial association between total-flavonoid consumption and all-cause mortality (Chapter 4) was supported by mechanistic and human trial data

(6-13). However, results of other population studies using older flavonoid composition databases to explore associations of flavonoid intake with total mortality are less convincing (14, 15). However, with the development of two comprehensive and up-to-date flavonoid food composition databases (USDA and PE databases) which produce highly correlated total-flavonoid intake estimates (Chapter 5), replication studies are indicated to further elucidate the associations of flavonoids with all-cause mortality.

Randomised controlled trials from this thesis (Chapter 7) and others (16-18) have demonstrated no beneficial effect of probiotic bacteria on glycaemia in normoglycaemic individuals. Despite this, animal and human studies suggesting acute and long-term hypoglycaemic effect of probiotic bacteria in physiological states of hyperglycaemia and insulin resistance (19-21). With limited and conflicting human trial data, future replication studies, particularly in diabetic patients, are indicated in order to clarify the role of probiotic strains on glycaemic control.

There is very little data describing the cardiovascular benefits of probiotics in population settings. Beneficial associations of habitual probiotic yoghurt consumption with atherosclerotic vascular disease risk factors (Chapter 10) and outcomes (22).

Chapter 11: Concluding statements Chapter 11: Page 6 However, they are yet to be replicated or explored in different cohorts. Therefore, in order to further our understanding of the atherosclerotic benefits of probiotic yogurt intake, further observational and clinical trials are indicated.

Chapter 11: Concluding statements Chapter 11: Page 7 11.4 IMPLICATIONS OF THESIS FINDINGS FOR CLINICAL PRACTICE

The finding that flavonols are associated with reduced risk of atherosclerotic vascular disease (Chapter 2), is supported by clinical trial and epidemiological data showing cardioprotective effects of flavonols (6, 7, 23, 24). Data from this thesis has contributed substantially to the scientific dossier regarding the vascular benefits of flavonoid consumption, and adds further support for the incorporation of flavonoid recommendations into current evidence-based dietary approaches for vascular disease prevention.

Despite a long history of safe use and the incorporation of probiotics into current therapeutic guidelines (25, 26), the ability of probiotics to colonize the gastrointestinal tract and the appropriate mode of administration in order to maximize probiotic efficacy remains uncertain. Clinical trial data from this thesis (Chapter 6) and others (27, 28) suggest that probiotic yoghurt, and not probiotic capsules, is an effective mode of administration for inducing beneficial changes to fecal BB12. However, other clinical trial data does not support this hypothesis (29). Given the little data regarding colonisation efficacy of probiotics, it is surprising probiotics feature in current therapeutic guidelines for the prevention and treatment of pouchitis (25) and management of bacterial overgrowth in bariatric surgery patients (26).

This thesis is likely to have greatest importance to clinical practice in the area of probiotics and cholesterol. There is now high quality evidence that probiotic supplementation, particularly probiotics in isolated and not fermented milk form, lower total cholesterol and low density lipoprotein cholesterol (Chapter 9). Probiotics are less likely to have beneficial effects when administered to adults with low to moderately low cholesterol concentrations (Chapter 8).

Chapter 11: Concluding statements Chapter 11: Page 8 11.5 CHAPTER 11 REFERENCES

1. Yanarates O, Guven A, Sizlan A, et al. Ameliorative effects of proanthocyanidin on renal ischemia/reperfusion injury. Renal Failure 2008;30(9):931-8. 2. Nakagawa T, Yokozawa T, Satoh A, Kim HY. Attenuation of renal ischemia-reperfusion injury by proanthocyanidin-rich extract from grape seeds. Journal of nutritional science and vitaminology 2005;51(4):283-6. 3. Avramovic V, Vlahovic P, Mihailovic D, Stefanovic V. Protective effect of a bioflavonoid proanthocyanidin-BP1 in glycerol-induced acute renal failure in the rat: renal stereological study. Ren Fail 1999;21(6):627-34. 4. Stefanovic V, Savic V, Vlahovic P, Cvetkovic T, Najman S, Mitic-Zlatkovic M. Reversal of experimental myoglobinuric acute renal failure with bioflavonoids from seeds of grape. Renal Failure 2000;22(3):255-66. 5. Lee YA, Kim YJ, Cho EJ, Yokozawa T. Ameliorative effects of proanthocyanidin on oxidative stress and inflammation in streptozotocin-induced diabetic rats. Journal of Agricultural and Food Chemistry 2007;55(23):9395-400. 6. Edwards RL, Lyon T, Litwin SE, Rabovsky A, Symons JD, Jalili T. Quercetin reduces blood pressure in hypertensive subjects. Journal of Nutrition 2007;137(11):2405-11. 7. Egert S, Bosy-Westphal A, Seiberl J, et al. Quercetin reduces systolic blood pressure and plasma oxidised low-density lipoprotein concentrations in overweight subjects with a high-cardiovascular disease risk phenotype: a double-blinded, placebo-controlled cross-over study. British Journal of Nutrition 2009;102(07):1065-74. 8. Brown AL, Lane J, Coverly J, et al. Effects of dietary supplementation with the green tea polyphenol epigallocatechin-3-gallate on insulin resistance and associated metabolic risk factors: randomized controlled trial. British Journal of Nutrition 2009;101(06):886-94. 9. Loke WM, Hodgson JM, Proudfoot JM, McKinley AJ, Puddey IB, Croft KD. Pure dietary flavonoids quercetin and (-)-epicatechin augment nitric oxide products and reduce endothelin-1 acutely in healthy men. American Journal of Clinical Nutrition 2008;88(4):1018-25. 10. Schroeter H, Heiss C, Balzer J, et al. (–)-Epicatechin mediates beneficial effects of flavanol-rich cocoa on vascular function in humans. Proceedings of the National Academy of Sciences 2006;103(4):1024-9. 11. Loke W, Hodgson J, Croft K. The biochemistry behind the potential cardiovascular protection by dietary flavonoids. Edtion ed. In: Fraga CG, ed. Plant phenolics and human health: biochemistry, nutrition and pharmacology. New Jersey: John Wiley & Sons, 2009. 12. Hodgson JM, Croft KD. Tea flavonoids and cardiovascular health. Molecular Aspects of Medicine 2010;31(6):495-502. 13. Rice-Evans C. Flavonoid antioxidants. Current medicinal chemistry 2001;8(7):797-807. 14. Mink PJ, Scrafford CG, Barraj LM, et al. Flavonoid intake and cardiovascular disease mortality: a prospective study in postmenopausal women. American Journal of Clinical Nutrition 2007;85(3):895-909.

Chapter 11: Concluding statements Chapter 11: Page 9 15. Cutler GJ, Nettleton JA, Ross JA, et al. Dietary flavonoid intake and risk of cancer in postmenopausal women: The Iowa Women's Health Study. International Journal of Cancer 2008;123(3):664-71. 16. Laitinen K, Poussa T, Isolauri E. Probiotics and dietary counselling contribute to glucose regulation during and after pregnancy: a randomised controlled trial. British Journal of Nutrition 2009;101(11):1679-87. 17. Yadav H, Jain S, Sinha PR. Antidiabetic effect of probiotic dahi containing Lactobacillus acidophilus and Lactobacillus casei in high fructose fed rats. Nutrition 2007;23(1):62-8. 18. Tabuchi M, Ozaki M, Tamura A, et al. Antidiabetic effect of Lactobacillus GG in streptozotocin-induced diabetic rats. Bioscience Biotechnology and Biochemistry 2003;67(6):1421-4. 19. Bukowska H, Pieczul-Mróz J, Jastrzebska M, Chełstowski K, Naruszewicz M. Decrease in fibrinogen and LDL-cholesterol levels upon supplementation of diet with Lactobacillus plantarum in subjects with moderately elevated cholesterol. Atherosclerosis 1998;1998(137):2. 20. Naruszewicz M, Johansson M-L, Zapolska-Downar D, Bukowska H. Effect of Lactobacillus plantarum 299v on cardiovascular disease risk factors in smokers. American Journal of Clinical Nutrition 2002;76(6):1249-55. 21. Sanggaard K, Holst J, Rehfeld J, Sandstrom B, Raben A, Tholstrup T. Different effects of whole milk and a fermented milk with the same fat and lactose content on gastric emptying and postprandial lipaemia, but not on glycaemic response and appetite. British Journal of Nutrition 2004;92:447-59. 22. Tavani A, Gallus S, Negri E, La Vecchia C. Milk, dairy products, and coronary heart disease. J Epidemiol Community Health 2002;56(6):471-2. 23. Huxley RR, Neil HA. The relation between dietary flavonol intake and coronary heart disease mortality: a meta-analysis of prospective cohort studies. European Journal of Clinical Nutrition 2003;57(8):904-8. 24. Perez-Vizcaino F, Duarte J, Andriantsitohaina R. Endothelial function and cardiovascular disease: Effects of quercetin and wine polyphenols. Free Radical Research 2006;40(10):1054-65. 25. Mowat C, Cole A, Windsor A, et al. Guidelines for the management of inflammatory bowel disease in adults. Gut 2011;60(5):571-607. 26. Mechanick JI, Kushner RF, Sugerman HJ, et al. American Association of Clinical Endocrinologists, the Obesity Society, and American Society for Metabolic and Bariatric Surgery medical guidelines for clinical practice for the perioperative nutritional, metabolic and nonsurgical support of the bariatric surgery patient. Obesity 2009;17(S1):S3-S72. 27. Savard P, Lamarche B, Paradis M-E, Thiboutot H, Laurin É, Roy D. Impact of Bifidobacterium animalis subsp. lactis BB-12 and, Lactobacillus acidophilus LA-5-containing yoghurt, on fecal bacterial counts of healthy adults. International Journal of Food Microbiology 2011;149(1):50-7. 28. Anderson ADG, McNaught CE, Jain PK, MacFie J. Randomised clinical trial of synbiotic therapy in elective surgical patients. Gut 2004;53(2):241-5. 29. Saxelin M, Lassig A, Karjalainen H, et al. Persistence of probiotic strains in the gastrointestinal tract when administered as capsules, yoghurt, or cheese. International Journal of Food Microbiology 2010;144(2):293-300.

Chapter 11: Concluding statements Chapter 11: Page 10 11.6 TABLES Chapter 11, Table 1: Thesis research questions and hypotheses used to explore them

Research question: Can the benefits of flavonoid consumption be separated from the benefits attributable to their major dietary whole food sources?

Hypothesis: High habitual intake of flavonols from tea and non-tea sources will both be associated with a reduced risk of atherosclerotic vascular disease mortality in a population of elderly women

Research question: Given strong evidence for the vascular benefits of flavonoid consumption, is flavonoid intake associated with better functioning of highly vascularised organs?

Hypothesis: High habitual proanthocyanidin intake is associated with better renal function and reduced risk of clinical renal outcomes in a population of elderly women.

Research question: Given the well documented benefits of flavonoid consumption, is the consumption of flavonoids from any class associated with death from any cause?

Hypothesis: Using two comprehensive food composition databases to assess flavonoid intake, total-flavonoid consumption will be inversely associated with risk of 5-year all-cause mortality.

Research question: Given a lack of gold standard biomarker to reflect flavonoid intake, is flavonoid intake able to be adequately assessed in population settings?

Hypothesis: There will be strong agreement between flavonoid intake estimates derived from United States Department of Agriculture and Phenol-Explorer flavonoid food content databases.

Research question: Is supplementation of the diet with probiotic bacteria able to alter faecal microflora profile?

Hypothesis: Daily probiotic supplementation, from capsules and the whole food (yoghurt) form, will improve faecal counts of Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis BB12.

Research question: Are probiotic bacteria able to improve glycaemic control?

Hypothesis: The probiotic bacteria L. acidophilus La5 and B. animalis subsp lactis Bb12, supplemented in a whole food (yoghurt) or isolated (capsules) form, will improve biomarkers of glycaemic control.

Research question: Using an appropriately powered study design, does daily supplementation with probiotic bacteria improve cardiovascular disease risk factors? Hypothesis: Six week supplementation with Lactobacillus acidophilus La5 and Bifidobacterium animalis subsp lactis Bb12, provided in either yoghurt or capsule form, will improve home blood pressure and serum lipid profile men and women with features of the metabolic syndrome Research question: Is it possible to draw conclusions about the cardioprotective role of probiotic bacteria?

Hypothesis: Using a Cochrane Library approved review of literature, supplementation with probiotic fermented milk and isolated probiotic bacteria will improve serum lipid profile in adults.

Research question: Is probiotic yoghurt beneficially associated with cardiovascular health in the population?

Hypothesis: Habitual intake of yoghurt, but not other dairy products, will be beneficially associated with common carotid artery intima-media thickness.

Chapter 11: Concluding statements Appendix A: Page 1 Health benefits of non-nutritive food components

Appendix A: Randomised controlled trial of yoghurt and its probiotics: study protocol

Kerry Ivey

Bachelor of Science (Nutrition) Curtin University, School of Public Health

Postgraduate diploma (Dietetics) Curtin University, School of Public Health

Supervisors: Professor Richard L. Prince, Professor Jonathan M Hodgson, Associate Professor Deborah Kerr

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 2 A.1 FOREWORD

Chapters 6 to 8 have reported results of a randomised controlled tiral exploring the health benefits of 6 week daily supplementation of probiotic yoghurt and probiotic capsules. This thesis was presented as a series of papers, and the structure of journal articles do not allow for detailed presentation of clinical trial methods. As such, this appendix provides a more detailed summary of the prespecified study protocol, that was implemented between February 2012 and February 2013. The rationale and background for study outcomes has been presented in thesis chapters. However, this appendix does explain the rationale for decisions regarding study design.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 3 A.2 STUDY DESIGN RATIONALE

The study described in this appendix is a is a block randomised, factorial, parallel, double blinded parallel study investigating the effects of Lactobacillus acidophilus La5 and Bidobacterium animalis subsp. lactis Bb12 from capsules and yoghurt. The following provides a rationale for features of this study design.

A.2.1 Recruitment procedure

To ensure a population based cohort, and minimise recruitment bias, a random sample of names and address of current electors in Western Australia were selected from the electoral roll.

A.2.2 Randomisation procedure

Participants for this study were not recruited at one time point. Recruitment letters were sent out continuously, and participants were recruited at various time points during study implementation. To avoid chronological bias of intervention group assignment, a block randomisation method was implemented (1).

A.2.3 Study design

A.2.3.1 Factorial study design

Yoghurt is a dairy product produced by the bacterial fermentation of milk. The World

Health Organisation (WHO) (2) defines probiotics as live microorganisms which, when administered in adequate amounts, confer a health benefit on the host. This definition encompasses both the bacteria and food products containing bacteria, but it remains unclear whether probiotic food products possess greater probiotic action than the isolated probiotic bacteria.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 4 Biologically active peptides released from proteins during the bacterial fermentation of milk are thought to contribute the metabolic benefits associated with probiotic consumption and aids in gastrointestinal colonisation of the probiotic species (3).

However there are limited intervention data investigating this hypothesis.

In order to assess the independent and additive effects of probiotic bacteria

(Lactobacillus acidophilus La5 and Bidobacterium animalis subsp. lactis Bb12) from either yoghurt or capsules, a factorial study design was implemented (4).

A.2.3.2 Parallel study design

There is data to suggest the metabolic effects of probiotics may continue after probiotic supplementation has ceased (5). Therefore, a parallel rather than crossover study design was implemented (6).

A.2.4 Test articles

A.2.4.1 Control test articles

Yoghurt is made from the bacterial fermentation of milk. Some yoghurt products contain probiotic bacteria, whilst others are made using traditional yoghurt starter cultrures which are not commonly considered as probiotic. Fermentation of milk which results in the alteration of the free amino acid and carbohydrate composition of the milk

(3, 7-9). These metabolites of probiotic bacteria, such as bioactive peptides, are thought to contribute to the beneficial effects of probiotics (3).

This study aimed to investigate and contrast the effects of isolated probiotic bacteria

(from capsules) with probiotic bacteria and bacterial fermentation products (from yoghurt). Therefore, placebo capsules not containing probiotic bacteria were considered an appropriate control for the probiotic capsules. Milk with a similar nutritional profile

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 5 was considered an appropriate control for the probiotic yoghurt as it did not contain any probiotic bacteria or products of probiotic fermentation, whilst controlling for any effects ascribable to consumption of dairy foods.

A.2.4.2 Probiotic test articles

The strains Lactobacillus acidophilus La5 and Bidobacterium animalis subsp. lactis

Bb12 were chosen because efficacy of demonstrated capacity to survive the harsh environment of the human gastrointestinal tract (10-12), adhere to hydrocarbons (13,

14), and exert metabolic benefits (15).

A.2.4.3 Timing of test article consumption

In order for probiotic bacteria to colonise the gastrointestinal tract, they must survive the harsh acidic and digestive conditions of the upper gastrointestinal tract. In order to minimise inter and intraindividual variation of gastrointestinal function at time of test article administration, test articles were consumed 30 minutes prior to the first meal of the day.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 6 A.3 PARTICIPANTS

A.3.1 Participants

All participants met the following criteria for inclusion in the study:

1. Ambulant men and postmenopausal women aged above 55 years

2. Minimal usual probiotic intake (consuming less than 400 g yoghurt per week, and

not taking probiotic supplements).

3. Overweight or obese (body mass index ≥ 25 kg/m2).

4. Participants will be at increased risk of metabolic syndrome and diabetes,

indicated by the presence of at least one of the following features associated with

increased risk of metabolic syndrome:

a. Waist circumference greater than 94 cm in men and 80cm in women

b. Blood pressure ≥120/80 mmHg

Participants were excluded if they fulfilled any of the following criteria:

1. Unlikely to complete the 6 week study;

2. Intolerance to dairy foods;

3. Unable to attend the study centre on 6 occasions over 9 weeks;

4. Current antibiotic use;

5. Patients on immunosuppressive treatments;

6. Type two diabetes requiring pharmaceutical treatment.

Children, mentally ill individuals and persons in dependant relationships were not be recruited for this study.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 7 A.3.2 Sample size

A sample size of 68 in each main effect dairy treatment arm (probiotic yoghurt or no yoghurt) was deemed sufficient to detect a 3.6% change in LDLC (~0.13 mmol/L), with

80% power at P=0.05. This number was increased to 76 per group (total of 152) to allow for a predicted drop-out from treatment of around 10%.

With P=0.05, 68 participants per main effect dairy treatment arm will also provide >80% power to detect: 1) a 3.5 mm Hg difference in mean systolic blood pressure (measured twice daily over 7 days; and 2) a 5% change (~0.3 mmol/L) in fasting glucose concentrations.

A.3.3 Recruitment

A sample of 8,000 names and addresses of current electors aged above 55 years in the

Perth metropolitan area were selected from the West Australian electoral roll by the

West Australian Data Linkage System (WADLS). This sample was cross-referenced against death registrations to remove any one who had died.

The names and addresses were encrypted and burnt to disc at the WADLU and delivered to Quickmail (Kenwick, AUSTRALIA); an approved mail consulting and management company. Quickmail posted recruitment letters to the 8,000 selected names. People who wished to take part in the study contacted the study centre.

Following a return contact by post or telephone, 887 respondents underwent initial telephone screening. Those that were eligible were sent a study outline and informed consent documentation, and will be booked for a screening clinic visit for final determination of eligibility for the study.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 8 A.4 STUDY DESIGN

A randomised, controlled, parallel, factorial design study of 6 weeks duration (Figure 1) was performed.

Appendix A, Figure 1: Study plan

Probiotic capsule (group A)

Placebo capsule (group B) Probiotic yoghurt

Non-probiotic Probiotic capsule (group C) milk

Placebo capsule (group D)

-3 -2 -1 0 1 2 3 4 Timeline Lead in period Intervention period (weeks)

Screening assessment Endpoint assessment Blood pressure monitor collection

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 9 Participants were randomised to one of four groups (Table 1). All test articles were consumed 30 minutes prior to the first meal of the day, every day, for 6 weeks. The four test articles used in this study were:

1. 200g/d Probiotic yoghurt: containing a minimum L. acidophilus La5 count of

1.41x108 CFU/g and a minimum B. lactis Bb12 count of 3.8x107 CFU/g.

2. 250mL long-life milk: containing no probiotic bacteria and energy matched to

the yoghurt test article.

3. 3x probiotic capsules/d: each capsule contains1.0x109 CFU of both L.

acidophilus La5 and B. lactis Bb12.

4. 3 x placebo capsules/d: energy matched to the probiotic capsules, and not

containing probiotic bacteria.

Appendix A, Table 1: Outline of intervention groups in the full factorial study design

Dairy supplementation

Probiotic Yoghurt Milk Capsule supplementation Probiotic capsule Group A Group B Placebo capsule Group C Group D

Participants were provided with a 3 week supply of dairy test articles (yoghurt or milk) at baseline and week 3, and were provided with a 6 week supply of capsules (either probiotic or placebo) at the baseline visit.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 10 A.4.1 Screening assessment

After signing the informed consent document, volunteers undertook the screening studies to determine if potential participants fulfilled the criteria for inclusion in the study.

1. Screening questionnaire: used to determine usual probiotic intake, demographic

information such as education, past occupation and smoking history, health

history, medications and co-morbidities.

2. Anthropometry

a. Standing height was measured by a wall-mounted stadiometer to the

nearest 0.1cm.

b. Body weight was measured by an electronic scale to the nearest 0.1 kg.

c. Waist girth was measured by a tape measure to the nearest 0.1 cm.

Participants were asked to raise shirt above the waist, stand erect, with

feet together and abdomen relaxed. Waist circumference was assessed at

the narrowest part of the torso when standing behind participant, with

tape in a horizontal plane and pulled lightly until appropriate tension is

achieved.

3. Clinic blood pressure: Following relaxation in a seated position for 5 minutes,

two blood pressure measurements were performed, 1 minute apart. The blood

pressure cuff was positioned on bare skin of the right arm, and located 2-3 cm

above the antecubital position. The arm was rested in a lateral position with the

cuff in line with the heart.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 11 A.4.2 Lead in period

All eligible volunteers completed a 3 week lead-in (washout) period prior to randomisation. During this period, participants were advised to maintain their normal diet, whilst avoiding consumption a small number of dietary products containing probiotic bacteria, including probiotic supplements and probiotic fortified foods such as yoghurts and fermented milk drinks. These dietary restrictions were maintained throughout the 6 week intervention period.

A.4.3 Randomisation

Participants were randomly assigned to one of 4 intervention groups (1:1:1:1) using computer generated random numbers (generated by a biostatistician who was not involved in the conduct of the study) sealed in opaque envelopes. All study personnel and participants were blinded to treatment assignment for the duration of the study. A senior investigator not involved in trial implementation held the randomisation code in a password protected folder, which was not broken until the trial had been completed and the analytical protocol had been finalised.

A.4.4 Intervention period

Following randomisation, participants were be required to consume either 200 mL yoghurt or 250mL reduced fat milk, and three capsules (either probiotic or placebo) every day for the duration of the intervention period. During the 6 week intervention, participants attended the clinic three times (baseline, week 3 and week 6) to receive test articles and to complete questionnaires. All major outcome measurements were performed at baseline and 6 weeks. Participants also attended the clinic at week -1 and week 5 to collect blood pressure monitors for home blood pressure monitoring.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 12 Compliance was monitored by a compliance diary and the return of empty containers at the end of 6 weeks.

All questionnaires were administered either in hard copy, or electronically using the electronically via either the Survey Monkey or Anti-Cancer Council of Victoria electronic systems. When using electronic survey systems, participants only entered de-identified data, including study number, initials and date of birth.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 13 A.5 PRIMARY OUTCOME VARIABLES

Primary outcome variables were assessed at baseline and week 6.

A.5.1 Biochemical measurements

Venous blood samples were collected in the morning after an overnight fast from 10 pm.

Measurements were be performed using routine laboratory techniques or commercially available kits at Pathwest, Royal Perth Hospital (Perth, Australia). The biochemical markers assessed as primary outcomes include:

a. Fasting serum lipids: Fasting total, LDL and HDL cholesterol, triglyceride

concentrations.

b. Markers of blood glucose control: Fasting blood glucose level, glycated

haemoglobin and insulin concentration.

A.5.2 Blood pressure measurements

A.5.2.1 Home blood pressure

Participants performed bidaily home blood pressure measurements during the 7 days preceeding both the baseline and week 6 visits. A fully automated home blood pressure monitor (UA-767PC, A&D, Japan) was provided to each participant at visits held one week prior to baseline, and one week prior to week 6. During these visits, participants were guided through the correct measurement procedure and were asked to perform a blood pressure reading in front of the study coordinator. This was to ensure the reliability of the blood pressure data from home monitoring.

Specifically, participants recorded blood pressure twice daily; approximately 1 hour after breakfast and approximately 1 hour after the evening meal. In a quiet room,

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 14 participants relaxed in a seated position for 5 minutes, and then performed 2 measurements over the following 5 minutes. The blood pressure cuff was positioned on bare skin, and located 2-3 cm above the antecubital position, and the arm was rested in a lateral position with the cuff in line with the heart.

Systolic blood pressure, diastolic blood pressure, pulse pressure and heart rate measurements were automatically recorded on the home blood pressure monitor, and were downloaded when monitors were returned to the study centre at the baseline and week 6 visits. In addition, participants also kept a diary where they manually recorded results of each home blood pressure measurement.

A.5.2.2 Clinic blood pressure

Following relaxation in a seated position for 5 minutes, two blood pressure measurements were performed, 1 minute apart. The blood pressure cuff was positioned on bare skin of the right arm, and located 2-3 cm above the antecubital position. The arm was rested in a lateral position with the cuff in line with the heart.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 15 A.6 SECONDARY OUTCOME VARIABLES PRESENTED IN THESIS

Secondary outcome variables were assessed at baseline, week 3 and week 6 visits, unless otherwise stated.

A.6.1 Fecal bacterial content

Fecal samples were collected at the end of the washout (baseline) and at the end of the

6-week intervention period (week 6). Participants were instructed to collect stool samples from the bowel motion preceding their baseline and week-6 visits, and stored at

-40C in fecal specimen jars, prior to delivery to the study center at their next visit.

Participants were instructed to avoid contamination of the sample with urine during the collection procedure.

A.6.1.1 DNA extraction

Genomic DNA was extracted from 200 mg of each faecal sample using a PowerSoil

DNA purification Kit (MolBiol Carlsbad, California) according to the manufacturer’s instructions with a slight modification (prior to the DNA extraction, the samples were subjected to three cycles of freezing-thawing to ensure complete lysis of all bacteria). DNA was also extracted from individual probiotic capsules by diluting the contents of each capsule in 200 µl of water and then extracting the DNA using the

PowerSoil kit as described above. Extractions on probiotic capsules were performed in triplicate. The DNA concentration of all samples were then analysed (NanoDrop,

ND-1000) according to the user’s manual.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 16 A.6.1.2 Quantitation of bacteria numbers in probiotic capsules using droplet digital

PCR

The absolute numbers of BB12, LA5, Bifidobacteria and total bacteria in each probiotic capsule was determined using droplet digital PCR (ddPCR). ddPCR was repeated on serial dilutions of capsule DNA from 1:10 to 1:10,000. This data was used as the standard curve for analog qPCR.

The primers and probes used for ddPCR quantitation of the concentrations of BB12, LA5,

Bifidobacteria and total bacteria in each capsule were as previously described (12), with the exception that the probe for LA5 was labelled with Joe at the 5’ end and BHQ1 at 3’ end. The probes for BB12, Bifidobacteria and total bacteria were all labelled with FAM at the 5’ end and BHQ2 at the 3’ end.

Four separate ddPCR runs were conducted in triplicate using a QX100TM droplet digital

PCR system (BioRad, Gladesville, NSW, Australia) according to the manufacturer’s instructions. Briefly, the ddPCR reaction mixture consisted of 12.5 μl of a 2 × ddPCR master mix (Bio-Rad), 2 μl of primer/probe mix (12.5mM each of the primer and probe), 1 μl of

DNA (adjusted to 50 ng/µl) and 9.5 μl of H2O to make a final volume of 25 μl. Droplets were generated using the Droplet Generator (DG) with 70 µL DG Oil per well with a DG8 cartridge and cartridge holder, 25 µL PCR reaction mix, and DG8 gasket. Droplets were dispensed into the 96-well PCR plate by aspirating 40 µL from the DG8 cartridge into each well. The PCR plate was then heat-sealed with a foil seal and the sealed plate was placed in the PCR thermocycler. Cycling consisted of 95°C for 10 min, followed by 45 cycles of 94°C for 30 s and 58°C for 45 s, 1 cycle of 98°C for 10 min with a 12°C hold. After the reaction, the droplets were read using the Droplet Reader, and QuantaSoft software converted the data into the number of template copies per μl of PCR mixture. The number of copies in 1 μl of DNA solution was then calculated.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 17 A.6.1.3 Quantitation of bacteria numbers in faecal samples using conventional qPCR

Analog (conventional) qPCR testing was used to quantitate the numbers of BB12, LA5,

Bifidobacteria and total bacteria in faecal samples using the standard curve generated from ddPCR (above).

The same primers and probes used in the ddPCR were also used for qPCR, with the exception that the probe for Bifidobacteria was labelled with Cy5 at the 5’ end and the probe for total bacteria was labelled with Rox at the 5’ end. The probes for BB12 and LA5 were labelled with FAM and Joe at their 5’ ends, respectively.

This enabled multiplex detection of all four reactions on a Rotor-Gene 6000 (Qiagen,

Victoria, Australia). The multiplex qPCR reaction mixture consisted of 10 μl of a 2× PCR master mix (Roche, Castle, NSW Australia), 4 x 1 μl of each primer/probe mix (containing a

6.25 mM concentration of each primer and probe), 1 μl of faecal DNA extract (adjusted to 50 ng/µl) and H2O to a final volume of 20 μl. Cycling conditions were as follows: 10 min at

95°C, followed by 45 cycles of 95°C for 20 seconds and 60°C for 45 seconds. A standard curve and a negative control were included on all runs.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 18 A.7 SECONDARY OUTCOMES NOT PRESENTED IN THIS THESIS

A.7.1 Metabolomic profile

Gas chromatography mass spectrophotomorty analysis was undertaken to determine areas under the curve for water soluble metabolites present in baseline and week 6 serum samples.

A.7.2 Gastrointestinal symptoms

Gastrointestinal symptoms, including pain, bloating, defecation frequency and stool consistency, were assessed using the Irritable Bowel Syndrome Severity Scale (16); a tool appropriate for assessing improvements in irritable bowel syndrome symptoms in clinical trials (17).

A.7.3 Functional health and wellbeing, physical and mental health

Functional health and wellbeing, physical and mental health was assessed using the

SF-36 Health Survey.

A.7.4 Quality of life

Quality of life was assessed using the World Health Organisation Quality of Life-BREF questionnaire.

A.7.5 Biochemical measurements related to bone metabolism and inflammation

Parathyroid hormone, serum calcium, phosphate, albumin, creatinine, procollagen type I

N-terminal propeptide, osteocalcin, and estradiol.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 19 A.7.6 Inflamatory markers

High sensitivity C-reactive protein, interleukin-6 and 8, monocyte chemoattractant protein 1 and tumour necrosis factor-alpha.

A.7.7 Fasting urine sample

Urine was collected after an overnight fast at the same time as the blood sample at baseline and week 6 visits. The urine samples will be analysed for creatinine, phosphorus, calcium, and N-terminal telopeptide.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 20 A.8 MEASUREMENTS OF EXPOSURE AND CONFOUNDING EFFECTS

A.8.1 Anthropometry

The following variables were assessed at baseline, 3 and 6 weeks:

a. Standing height measured by a wall-mounted stadiometer to the nearest 0.1cm.

b. Body weight measured by an electronic scale to the nearest 0.1 kg.

c. Waist girth measured by a tape measure to the nearest 0.1 cm. Participants were

asked to raise shirt above the waist, stand erect, with feet together and

abdomen relaxed. Waist circumference was be assessed at the narrowest part

of the torso when standing behind participant, with tape in a horizontal plane

and pulled lightly until appropriate tension is achieved.

d. Hip girth was be measured by a tape measure to the nearest 0.1 cm.

Participants were be asked to stand erect, with feet together and abdomen

relaxed. Hip circumference was assessed at the point of maximum extension

of the buttocks, with tape in a horizontal plane and pulled lightly until

appropriate tension is achieved.

A.8.2 Dietary intake

Dietary intake was assessed at Baseline and week 6 using a validated semi-quantitative food frequency questionnaire developed by the Anti-Cancer Council of Victoria (18-20).

Energy and nutrient intakes were estimated based on frequency of consumption and an overall estimate of usual portion size (21). A beverage questionnaire was used to assess average tea and coffee consumption over the past 12 months (22).

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 21 A.8.3 Physical activity

At baseline, week 3 and week 6, physical activity level was assessed by the

International Physical Activity Questionnaire.

A.8.4 Compliance

A.8.4.1 Empty container collection

Throughout the 6 week intervention, participants were be instructed to retain all test article containers and remaining product. This was returned at the week 6 clinic visit.

A.8.4.2 Compliance diary

At the baseline visit, participants were provided with a diary, and were instructed to complete daily recordings of time and amount of test article consumed. The diary was returned at the week 6 clinic visit.

A.8.5 Medication use

At screening, participants were asked to bring all prescribed and non-prescribed medications, supplements and remedies. The commencement date, reason for use, name, dose, mode of administration and frequency was recorded. Where necessary, medications were cross-referenced with the complete list of product and generic medication names in the electronic Monthly Index of Medical Specialties.

At all subsequent visits, participants were asked to report any changes to their medication regimens, and where necessary, to bring new medication packages to the clinic.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 22 A.8.6 Co-morbidities, past medical history and adverse events

At screening, participants were asked whether they are currently diagnosed with any medical conditions by a physician. The date of diagnosis and medications prescribed as a result of the condition were recorded. Participants were also asked to provide dates and details of significant past medical history or procedures. At each subsequent visit, participants were asked to report any changes to co-morbidities.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 23 A.9 APPENDIX A REFERENCES

1. Kao LS, Tyson JE, Blakely ML, Lally KP. Clinical research methodology I: Introduction to randomized trials. Journal of the American College of Surgeons 2008;206(2):361. 2. Joint FAO/WHO Expert Consultation on Evaluation of Health and Nutritional Properties of Probiotics in Food. Health and nutritional properties of probiotics in food including powder milk with live lactic acid bacteria. Argentina, 2001. 3. Meisel H. Overview on milk protein-derived peptides. International Dairy Journal 1998;8(5-6):363-73. 4. Piantadosi S. Factorial Designs in Clinical Trials. Edtion ed. Encyclopedia of Biostatistics: John Wiley & Sons, Ltd, 2005. 5. Kalliomäki M, Salminen S, Poussa T, Arvilommi H, Isolauri E. Probiotics and prevention of atopic disease: 4-year follow-up of a randomised placebo-controlled trial. The Lancet 2003;361(9372):1869-71. 6. Mills E, Chan A-W, Wu P, Vail A, Guyatt G, Altman D. Design, analysis, and presentation of crossover trials. Trials 2009;10(1):27. 7. Frengova GI, Simova ED, Beshkova DM, Simov ZI. Production and monomer composition of exopolysaccharides by yogurt starter cultures. Canadian Journal of Microbiology 2000;46(12):1123-7. 8. Lamoureux L, Roy D, Gauthier SF. Production of oligosaccharides in yogurt containing bifidobacteria and yogurt cultures. Journal of Dairy Science 2002;85(5):1058-69. 9. Lourens-Hattingh A, Viljoen BC. Yogurt as probiotic carrier food. International Dairy Journal 2001;11(1):1-17. 10. Alander M, Mättö J, Kneifel W, et al. Effect of galacto-oligosaccharide supplementation on human faecal microflora and on survival and persistence of Bifidobacterium lactis Bb-12 in the gastrointestinal tract. International Dairy Journal 2001;11(10):817-25. 11. Fukushima Y, Kawata Y, Hara H, Terada A, Mitsuoka T. Effect of a probiotic formula on intestinal immunoglobulin A production in healthy children. International Journal of Food Microbiology 1998;42(1):39-44. 12. Savard P, Lamarche B, Paradis M-E, Thiboutot H, Laurin É, Roy D. Impact of Bifidobacterium animalis subsp. lactis BB-12 and, Lactobacillus acidophilus LA-5-containing yoghurt, on fecal bacterial counts of healthy adults. International Journal of Food Microbiology 2011;149(1):50-7. 13. Schillinger U, Guigas C, Heinrich Holzapfel W. In vitro adherence and other properties of lactobacilli used in probiotic yoghurt-like products. International Dairy Journal 2005;15(12):1289-97. 14. Collado MC, Jalonen L, Meriluoto J, Salminen S. Protection mechanism of probiotic combination against human pathogens: in vitro adhesion to human intestinal mucus. Asia Pacific Journal of Clinical Nutrition 2006;15(4):570-5. 15. Ataie-Jafari A, Larijani B, Alavi Majd H, Tahbaz F. Cholesterol-lowering effect of probiotic yogurt in comparison with ordinary yogurt in mildly to moderately hypercholesterolemic subjects. Annals of Nutrition and Metabolism 2009;54(1):22-7.

Randomised controlled trial of yoghurt and its probiotics: study protocol Appendix A: Page 24 16. Francis CY, Morris J, Whorwell PJ. The irritable bowel severity scoring system: a simple method of monitoring irritable bowel syndrome and its progress. Alimentary Pharmacology & Therapeutics 1997;11(2):395-402. 17. Irvine EJ, Whitehead WE, Chey WD, et al. Design of treatment trials for functional gastrointestinal disorders. Gastroenterology 2006;130(5):1538-51. 18. Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Australia New Zealand Journal of Public Health 2000;24(6):576-83. 19. Xinying PX, Noakes M, Keogh J. Can a food frequency questionnaire be used to capture dietary intake data in a 4 week clinical intervention trial? Asia Pacific Journal of Clinical Nutrition 2004;13(4):318-23. 20. Ambrosini GL, van Roosbroeck SAH, Mackerras D, Fritschi L, de Klerk NH, Musk AW. The reliability of ten-year dietary recall: implications for cancer research. Journal of Nutrition 2003;133(8):2663-8. 21. Ireland P, Jolley D, Giles G, et al. Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pacific Journal of Clinical Nutrition 1994;3:19-31. 22. Hodgson JM, Devine A, Puddey IB, Chan SY, Beilin LJ, Prince RL. Tea intake is inversely related to blood pressure in older women. J Nutr 2003;133(9):2883-6.

Randomised controlled trial of yoghurt and its probiotics: study protocol