The Role of Breakfast Consumption on Metabolism, Body Size and Chronic Disease Risk Amongst Healthy Adults

The role of breakfast consumption on metabolism, body size and chronic disease risk amongst healthy adults.

Angelica Quatela, BSc (Hons) Nut A thesis submitted for the degree of PhD (Nutrition and Dietetics) University of Newcastle, NSW, Australia 16 November 2017

This research was supported by an Australian Government Research Training

Program (RTP) Scholarship

Statements I hereby certify that the work embodied in the thesis is my own work, conducted under normal supervision.

The thesis contains published scholarly work of which I am a co-author. For each such work a written statement, endorsed by my supervisor, attesting to my contribution to the joint work has been included.

This thesis contains no material which has been accepted for the award of any other degree or diploma in any university or other tertiary institution and, to the best of my knowledge and belief, contains no material previously published or written by another person, except where due reference has been made in the text. I give consent to the final version of my thesis being made available worldwide when deposited in the University’s Digital Repository, unless an embargo has been approved for a determined period, subject to the provisions of the Copyright Act 1968.

Angelica Quatela 15-11-17

Dr Lesley MacDonald-Wicks 15-11-17

2 Acknowledgments I would like to express my gratitude to the following people for their contribution towards the completion of my PhD: • Firstly, I wish to thank the participants of the typical Aussie Bloke project for their time and for kindly volunteering to take part in the study. • I would like to deeply thank my dear supervisors, Dr Lesley MacDonald-Wicks, Dr Amanda Patterson and Prof Robin Callister. Thanks you for your precious guidance, patience and professionalism that allowed me to fulfil my dream to complete this PhD journey. I will be forever grateful to you all. • A special thank to A/Prof Mark McEvoy (HMRI) for his precious time and outstanding statistical support during the ‘Australian Longitudinal Study on Women’s Health longitudinal analyses. It has been a real pleasure working with you and thanks for everything you have taught me. • I wish to show my sincere gratitude to A/Prof Leanne Brown for her precious help in facilitating data collection for the Typical Aussie Bloke study in Tamworth Education Centre (Tamworth, NSW, Australia). • I would also like to acknowledge my deepest gratitude to: Prof Robert Callister (Head of the Faculty of Health of Medicine); A/Prof Shane Dempsey (Head of School, Health Sciences); Ms Shirley Savy (Research Training Officer). Also, thanks to all the School of Health Sciences Team, particularly to: Mr David Rambaldi, Mrs Alli Johns, Ms Brooke Allars, Ms Tara Magnay, Ms Sally Goodchap, Ms Fiona Whyte, Mrs Sharni Greenwood, Ms Clare Eley-Smith, Mrs Sandra Fitness and Mrs Ashely Gleeson for their administrative support over these years. Furthermore, thanks to all the technical officers team for their enduring support: Mrs Anna Bukey, Mrs Jessica Piotrowski, and Mr Philip Jacobson. • I particularly thank Debbie Booth (Senior Research Librarian, Academic Division, the University of Newcastle) for providing great assistance with the literature search for the Systematic review illustrated in chapter 5. • I wish to acknowledge the contribution of the research assistants Allison Brandt, Loren Stroud, and Kelly Rice, who kindly assisted with the systematic review process and data extraction illustrated in chapter 5.

3 • Chapters 3 and 4 analysed the Australian Longitudinal Study on Women’s Health surveys data. We are grateful to the Australian Longitudinal Study on Women’s Health for allowing us to access their survey data and to the women who took part at this study. I am also grateful to Prof Graham Giles (Cancer Epidemiology Centre of Cancer Council Victoria) for permission to use the ‘Dietary Questionnaire for Epidemiological Studies (Version 2),’ Melbourne: Cancer Council Victoria, 1996, for the purpose of these analyses. • I would also like to thanks my fellow RHD Students in HA06, for their kind understanding, help and support during the PhD journey. • Finally, I would like to deeply thank my family, especially my beloved husband (Andrea), for their constant support and for believing in me during this challenging and rewarding journey. Grazie ;-)

Angelica

4 Conflict of interest

Angelica Quatela reports no conflict of interest.

Publications and Presentations Arising from this Thesis

Manuscripts in Peer-Reviewed Journals: Published 1. Chapter 3: Quatela, A., R. Callister, A.J Patterson, M. McEvoy and L.K. MacDonald-Wicks (2017). "Breakfast Cereal Consumption and Obesity Risk amongst the Mid-Age Cohort of the Australian Longitudinal Study on Women’s Health." Healthcare 5(3): 49. 2. Chapter 5: Quatela, A., R. Callister, A.J. Patterson and L.K. MacDonald-Wicks (2016). "The Energy Content and Composition of Meals Consumed after an Overnight Fast and Their Effects on Diet Induced Thermogenesis: A Systematic Review, Meta-Analyses and Meta-Regressions." Nutrients 8(11): 670. 1. Chapter 4: Quatela, A., R. Callister, A.J. Patterson, M. McEvoy and L.K. MacDonald-Wicks (2017). “The protective effect of muesli consumption on diabetes risk: Results from 12 years of follow-up in the Australian Longitudinal Study on Women’s Health.” Nutrition Research 51: 12.

Manuscripts in Peer-Reviewed Journals: In the Process to be Resubmitted 1. Chapter 6: Quatela, A; A.J. Patterson; R. Callister; L.K. MacDonald-Wicks (2017) “Breakfast consumption habits of young Australian men from the ‘Typical Aussie Bloke’ study”. To be submitted to European Journal of Nutrition.

Conference abstracts in Conference Proceedings or Peer-Reviewed Journals: Published 1. Quatela A., A.J. Patterson, R. Callister, L.K. MacDonald-Wicks, “The ‘Typical Aussie Bloke study’: breakfast consumption habits of young Australian men.” Asia Pacific Conference on Clinical Nutrition (2017) in Adelaide, Australia (poster presentation by Quatela, A). 2. Quatela A., A.J. Patterson, R. Callister, M. McEvoy, L.K. MacDonald-Wicks, “The effects of breakfast cereal consumption on obesity risk over 12 years among mid-aged women in the Australian Longitudinal Study on Women’s Health”, in International Society of Behavioural Nutrition and Physical Activity (ISBNPA) in Victoria, Canada, (2017) (oral presentation by Patterson, A).

3. Quatela A., A.J. Patterson, R. Callister, M. McEvoy, L.K. MacDonald-Wicks, “Breakfast cereal consumption and incident obesity: 12 years analyses of the Australian Longitudinal Study on Women's Health”, Nutrition Society of Australia (NSA) Annual Scientific Meeting in Melbourne, Australia, (2016) (poster presentation by Quatela, A). 4. Quatela A., A.J. Patterson, R. Callister, M. McEvoy, L.K. MacDonald-Wicks, “Is breakfast cereal consumption an effective strategy to prevent diabetes for mid-age Australian women?” NSA Annual Scientific Meeting in Melbourne, Australia, (2016) (poster presentation by Quatela, A). 1. Quatela A., R. Callister, A.J. Patterson, M. McEvoy, L.K. MacDonald-Wicks, “Breakfast cereal consumption and incident Diabetes Mellitus: Results from 12 years of the Australian Longitudinal Study on Women’s Health”. Australian Longitudinal Study on Women’s Health (ALSWH) in Newcastle, Australia, (2016) (oral presentation by Quatela, A). 5. Quatela A., R. Callister, A.J. Patterson, L.K. MacDonald-Wicks, “What it is not known of the effect of fat intake at breakfast on DIT.” ISBNPA in Edinburgh, Scotland (2015) (poster presentation by Quatela, A). 6. Quatela A., R. Callister, A.J. Patterson, L.K. MacDonald-Wicks, “The effect of breakfast size and frequency on diet induced thermogenesis.” ISBNPA in Edinburgh, Scotland (2015) (poster presentation by Quatela, A).

Conference Abstracts in Conference Proceedings or Peer-Reviewed Journals: Accepted for Oral Presentation 1. Quatela A., A.J. Patterson, R. Callister, L.K. MacDonald-Wicks, “The ‘typical Aussie Bloke study’: The relationships of Habitual Breakfast consumption with mediators of obesity and chronic disease development amongst young Australian men.” Dietetic Association of Australia in Sydney, Australia, (2018) (accepted oral presentation by Quatela, A).

Glossary of Common Abbreviations AB = Above Baseline

ABS = Australian Bureau of Statistics

AIHW= Australian Institute of Health and Welfare

ALSWH = Australian Longitudinal Study on Women’s Health

BE = Breakfast Eating

BS = Breakfast Skipping

BM = Body Mass

BMI = Body Mass Index

CHO = Carbohydrate

CI = Confidence Interval

CHD = Coronary Heart Disease

CVD = Cardiovascular Diseases

CSANZ = Cardiac Society of Australia and New Zealand

DA = Diabetes Australia

AusDiab = Australia Diabetes

DAA = Dietitians Association of Australia

DIT = Diet Induced Thermogenesis

DQES = Dietary Questionnaire for Epidemiological Studies

DQES-FFQv2 = Dietary Questionnaire for Epidemiology Studies Version 2.

EBE = Early Breakfast Eaters

ECM = Energy Content of the Meal

EE = Energy Expenditure

F = Female

FE = Frequent Eaters

FFM = Fat Free Mass

FM = Fat Mass

HBE = Habitual Breakfast Eaters

HBS = Habitual Breakfast Skippers

HMRI = Hunter Medical Research Centre

HR = Hazard Ratio

ISBNPA = International Society for Behavioural Nutrition and Physical Activity conference

LBE = Late Breakfast Eaters

LCT = Long Chain Triglycerides

LFE = Less Frequent Eaters

M = Male

MBS: Medicare Benefits Schedule

MCT = Medium Chain Triglycerides

MET = Metabolic Equivalent Task

MIT = Meal Induced Thermogenesis

MUFA = Mono Unsaturated Fatty Acids

NCD-RisC = NCD Risk Factor Collaboration

NVDPA = National Vascular Disease Prevention Alliance

NHANES = National Health and Nutrition Examination Survey

NHMRC = National Health and Medical Research Council

N = Sample

NIH = National Institute of Health

NHLBI = National Heart, Lung and Blood Institute

NNS = National Nutrition Survey

NP = Not Provided

NPI = Not Provided Information

NS = Not Significant

NSA = Nutrition Society of Australia

OBE = Occasional Breakfast Eaters

OECD = Organisation for Economic Co-operation and Development

PA = Physical Activity

PAT = Physical Activity Thermogenesis

PBS = Pharmaceutical Benefits Scheme

PEE = Postprandial Energy Expenditure

PUFA = Poly Unsaturated Fatty Acids

RCT = Randomised Control Trial

RCOD= Randomised Cross Over Design

REE = Resting Energy Expenditure

RMR = Resting Metabolic Rate

RR = Relative Risk

RTE = Ready To Eat

RTEC = Ready to Eat Cereal

SD = Standard Deviation 10

SE = Standard Error

SEM = Standard Error of the Mean

SFA = Saturated Fatty Acids

SR = Systematic Review

TEF = Thermic Effect of Food

TRF = Thermic Response to Food

VO2 = Rate of Oxygen

WHO = World Health Organization

Table of Contents

Contents Statements ...... 2 Acknowledgments ...... 3 Conflict of interest ...... 5 Publications and Presentations Arising from this Thesis ...... 6 Manuscripts in Peer-Reviewed Journals: Published ...... 6 Manuscripts in Peer-Reviewed Journals: In the Process to be Resubmitted...... 6 Conference abstracts in Conference Proceedings or Peer-Reviewed Journals: Published ...... 6 Conference Abstracts in Conference Proceedings or Peer-Reviewed Journals: Accepted for Oral Presentation ...... 7 Glossary of Common Abbreviations...... 8 Table of Contents ...... 12 List of Tables ...... 18 List of Figures ...... 19 Abstract ...... 20 Chapter 1: Breakfast Consumption Habits in relation to Countries, Gender, and Socio-Economic Background in Adults ...... 24 1.1 Definitions...... 24 1.1.1 Energy Balance ...... 24 1.1.2 Energy Expenditure...... 24 1.1.3 Breakfast Cereal ...... 25 1.2 Breakfast ...... 25 1.2.1 Defining Breakfast ...... 25 1.2.2 Breakfast Eating and Breakfast Skipping ...... 27 1.2.3 Influences on Breakfast Consumption ...... 28 1.2.4 Food Selection ...... 30 Chapter 2: Critical Literature Review of Breakfast Consumption Habits in relation to Obesity and Chronic Diseases Risk in Adults ...... 36 2.1 Overview ...... 36 2.2 Is Breakfast Protective against Obesity? ...... 37 2.3 Mechanisms Proposed to Explain a Protective Effect of Breakfast on Obesity ...... 39 2.3.1 The Effect of Breakfast Consumption on Energy Intake ...... 40

2.3.2 The Effect of Breakfast Consumption on Diet Induced Thermogenesis ...... 42 2.3.3 The Effect of Breakfast Consumption on Physical Activity ...... 45 2.4 Is Breakfast Skipping Associated with Increased Risk of Chronic Diseases? ...... 47 2.5 Mechanisms behind a protective effect of Breakfast Consumption on Chronic Disease Risk ...... 48 2.5.1 The Effect of Breakfast Consumption on Blood Parameters ...... 49 2.5.2 Breakfast Consumptions and Nutrients Status ...... 50 2.6 Is Breakfast Cereal Consumption Protective against Chronic Disease and Obesity Risk?51 2.7 Gaps in the Literature ...... 53 2.8 Aims, Objectives and Hypothesis ...... 54 2.8.1 First Aim ...... 54 2.8.2 Second Aim: ...... 55 2.9 Studies ...... 56 2.10 Thesis Structure ...... 56 Chapter 3: Breakfast Cereal Consumption and Obesity Risk amongst the Mid-Age Cohort of the Australian Longitudinal Study on Women’s Health ...... 58 3.1 Overview ...... 59 3.2 Introduction ...... 59 3.3 Materials and Methods ...... 60 3.3.1 Participants ...... 61 3.3.2 Predictor Variables ...... 61 3.3.3 Outcome Variable ...... 62 3.3.4 Identification and Measurement of Confounding Factors...... 63 3.3.5 Statistical Analyses ...... 64 3.4 Results ...... 65 3.4.1 Participant Characteristics ...... 65 3.3.3 Breakfast Cereal Consumption and Risk of Obesity ...... 69 3.5 Discussion ...... 71 3.6 Conclusions ...... 74 3.7 Acknowledgments ...... 74 3.8 Author Contributions ...... 74 3.9 Conflicts of Interest ...... 75 Chapter 4: The protective effect of muesli consumption on diabetes risk: Results from 12 years of follow-up in the Australian Longitudinal Study on Women’s Health ...... 76 4.1 Overview ...... 77 13

4.2 Abstract ...... 77 4.3 Introduction ...... 77 4.4 Methods and Materials ...... 79 4.4.1 Participants ...... 79 4.4.2 Predictor Variables ...... 80 4.4.3 Outcome Variable ...... 81 4.4.4 Identification and Measurement of Confounding Factors...... 81 4.4.5 Statistical Analyses ...... 82 4.5 Results ...... 83 4.5.1 Participant Characteristics ...... 84 4.5.2 Breakfast Cereal Consumption and the Risk of Developing Diabetes ...... 84 4.6 Discussion ...... 85 4.7 Conflict of Interest ...... 88 4.8 Acknowledgements ...... 88 4.9 Financial Support ...... 89 4.10 Author contributions ...... 89 Chapter 5: The Energy Content and Composition of Meals Consumed after an Overnight Fast and Their Effects on Diet Induced Thermogenesis: A Systematic Review, Meta-Analyses and Meta-Regressions ...... 95 5.1 Overview ...... 96 5.2 Abstract ...... 96 5.3 Introduction ...... 96 5.4 Materials and Methods ...... 98 5.4.1 Search ...... 98 5.4.2 Eligibility Criteria ...... 99 5.5 Results ...... 102 5.5.1 Participant Characteristics ...... 103 5.5.2 Interventions ...... 104 5.5.3 Outcomes ...... 104 5.5.4 Comparison and Meta-Regression of the Effects of Higher and Lower Energy Intakes on DIT ...... 115 5.5.5 Influence of Macronutrient Composition on DIT ...... 117 5.5.6 Long Chain Triglycerides vs. Medium Chain Triglycerides...... 119 5.5.7 Monounsaturated Fat vs. Polyunsaturated Fat ...... 121 5.5.8 Structure of Fats ...... 122 14

5.5.9 Processed vs. Unprocessed Food ...... 122 5.5.10 One Bolus Event vs. Isocaloric Smaller Frequent Meals ...... 123 5.5.11 Fast vs. Slow/Normal Meal Consumption ...... 124 5.5.12 Palatable vs. Unpalatable ...... 125 5.6 Discussion ...... 135 5.6.1 Strengths of This SR ...... 140 5.6.2 Limitations ...... 140 5.6.3 Recommendations ...... 143 5.7 Conclusions ...... 144 5.8 Supplementary Materials ...... 144 5.9 Acknowledgments ...... 144 5.10 Author contributions ...... 145 5.11 Conflict of interests ...... 145 Overview for Chapter 6 and 7 ...... 146 Chapter 6: Breakfast Consumption Habits of Young Australian Men from the “Typical Aussie Bloke” study...... 148 6.1 Abstract ...... 148 6.2 Introduction ...... 149 6.3 Methods...... 150 6.3.1 Study Design ...... 150 6.3.2 Inclusion and Exclusion Criteria ...... 150 6.3.3 Survey ...... 150 6.3.4 Statistics ...... 152 6.4 Results ...... 152 6.4.1 Demographic Characteristics ...... 152 6.4.2 Comparison of Breakfast Consumption Patterns ...... 153 6.4.3 Habitual Breakfast Patterns among Habitual Breakfast Eaters ...... 153 6.4.4 Non-Habitual Breakfast Patterns among Habitual Breakfast Eaters ...... 153 6.4.5 Other Foods and Beverages consumed for Breakfast by Habitual Breakfast Eaters 154 6.4.6 Early vs Late Breakfast Consumption and its relation with Waking Habits ...... 154 6.4.7 Reasons for Consuming or Not Consuming Breakfast ...... 157 6.5 Discussion ...... 162 6.6 Conclusions ...... 165 6.7 Acknowledgments ...... 165

6.8 Conflict of Interest ...... 166 6.9 Authors’ contributions ...... 166 6.10 Ethics...... 166 Chapter 7: ‘Typical Aussie Bloke’ Part 2: Breakfast Consumption and Eating Patterns in relation to intermediate risk factors for Obesity and Chronic Disease Development...... 167 7.1 Introduction ...... 167 7.2 Methods...... 169 7.2.1 Online Survey ...... 169 7.2.2 Lab Measurement Session ...... 170 7.2.3 Statistics ...... 171 7.3 Results ...... 171 7.3.1 Habitual Breakfast Eaters (HBE) vs Occasional Breakfast Eaters (OBE) vs Habitual Breakfast Skippers (HBS) ...... 171 7.3.2 Early Breakfast Eaters (EBE) vs Late Breakfast Eaters (LBE) ...... 172 7.3.3 Frequent Eaters (FE) vs Less Frequent Eaters (LFE) ...... 172 7.4 Discussion ...... 173 7.5 Conclusion ...... 177 Chapter 8: Final Discussion and Recommendations for Future Research ...... 186 8.1 Overview ...... 186 8.2 Introduction ...... 186 8.2.1 Overall Aims ...... 186 8.3 Main Findings of this Thesis ...... 188 8.4 Implications for Future Research ...... 191 8.5 Strengths and Limitations ...... 192 8.6 Conclusions ...... 192 References ...... 194 Appendix 1: Abstracts for Chapter 3 ...... 212 Appendix 2: Authors Contribution for chapter 3 ...... 215 Appendix 3: Abstracts for Chapter 4 ...... 216 Appendix 4: Authors Contribution for chapter 4 ...... 218 Appendix 5: Abstracts for Chapter 5 ...... 219 Appendix 6: Authors Contribution for chapter 5 ...... 222 Appendix 7. Supplementary Table for Systematic Review – Chapter 2...... 223 Appendix 8: Abstract for Chapter 6 ...... 224

Appendix 9: Authors Contribution for chapter 6 ...... 225 Appendix 10. Recruitment Flier for ‘Typical Aussie Bloke’ study for Newcastle ...... 226 Appendix 11. Consent Form for ‘Typical Aussie Bloke’ study ...... 227 Appendix 12. Ethics Approval for ‘Typical Aussie Bloke’ Study ...... 228 Appendix 13: Online Questionnaire for ‘Typical Aussie Bloke’ study ...... 229 Acknowledgements ...... 255 References of acknowledgments ...... 255 Appendix 14: Conference Abstract for Chapter 7...... 257 Appendix 15: Authors Contribution for chapter 7 ...... 258 Appendix 16. Information Sheet for ‘Typical Aussie Bloke’ study ...... 259

List of Tables Table 1.1 Foods and beverages consumed for breakfast in Australia, America, China and Italy

Table 3.1. Characteristics of participants from the mid-age (2001) cohort of the Australian Longitudinal Study on Women’s Health at Survey 3 (n = 4143) by “any” breakfast cereal and ‘no’ breakfast cereal consumption.

Table 3.2 Characteristics of participants from the mid-age (2001) cohort of the Australian Longitudinal Study on Women’s Health at Survey 3 (n = 4143) by individual breakfast cereal consumption.

Table 3.3 Results of the logistic regression models examining the effects of consuming breakfast cereal at Survey 3 on the risk of developing obesity from Surveys 4–7.

Table 4.1 Characteristics of participants from the mid-age (2001) cohort of the Australian Longitudinal Study of Women’s Health at Survey 3 (n=8422); comparison of participants consuming ‘any’ cereal or ‘no’ cereal.

Table 4.2 Characteristics of participants from the mid-age (2001) cohort of the Australian Longitudinal Study of Women’s Health at Survey 3 (n=8422) by individual breakfast cereal consumption category.

Table 4.3. Logistic regression models with descrete time survival analyses of the effect of consuming breakfast cereal at S3 on the risk of developing diabetes at S4-7 amongst 8422 mid-age women.

Table 5.1 Participant characteristics and study protocols.

Table 5.2 Consumption of meals after an overnight fast and DIT

Table 6.1 Demographic characteristics of participants of the Typical Aussie Bloke study categorised by habitual breakfast eating patterns.

Table 6.2 Habitual and non-habitual food and beverage consumption for breakfast among Habitual Breakfast Eaters

Table 6.3 Demographic characteristics of HBE categorised by timing of first meal of the day consumption. 18

Table 7.1 Anthropometric and metabolic characteristics by habitual breakfast habits.

Table 7.2 Lifestyle characteristics by habitual breakfast habits.

Table 7.3 Anthropometric and metabolic characteristics by timing of breakfast consumption.

Table 7.4 Lifestyle characteristic by timing of breakfast consumption

Table 7.5 Anthropometric and metabolic characteristics by eating frequency.

Table 7.6 Socio-demographic and lifestyle characteristics by eating frequency.

List of Figures Figure 2.1. Thesis structure diagram.

Figure 4.1. Flow chart of participant selection

Figure 5.1. PRISMA Flow diagram (105) systematic search and review process.

Figure 5.2. Mixed Model Meta Regression: univariate association between energy intake (kJ) and DIT (kJ/h) (Model 1).

Figure 5.3. Meta-analysis with fixed effect of the mean differences in DIT between MCT and LCT.

Figure 5.4 Meta-analysis: mean differences in DIT between bolus vs. smaller frequent meals event (such as snacking).

Abstract Breakfast is often described as the most important meal of the day because it is believed to play an important role in preventing obesity and the development of chronic diseases such as diabetes and cardiovascular disease (CVD). However, there is insufficient and contradictory evidence to support this claim. This PhD describes four studies which further investigate the role of breakfast on metabolic rate, body composition and chronic disease risk markers.

The first two studies were longitudinal analyses of a large, representative sample of Australian women, investigating first the association between breakfast cereal consumption and incident obesity and secondly the risk of developing diabetes, in mid- age women from the Australian Longitudinal Study of Women’s Health (ALSWH). My analyses found that muesli consumption on its own or as a part of an oats based cereal group was significantly associated with a reduction in the risk of developing obesity and type 2 diabetes in women over a 12-year period. All-Bran was also found to be significantly protective against incident obesity in women over 12 years. No other breakfast cereal, including ‘any breakfast cereal’ (a variable that combined all breakfast cereal consumption) and ‘higher fibre’ (a variable used to combine the high fibre breakfast cereals) breakfast cereal consumption, were found to be significantly protective against obesity or diabetes risk.

A systematic review, meta-analysis and meta-regression of randomised cross over design studies was then conducted to examine the evidence available about the role of consuming breakfast (of varying macronutrient composition and/or energy composition) on diet-induced thermogenesis (DIT). This is of particular interest as even small changes in DIT may have significant effects on body weight and/or body composition over the longer term. The findings of the meta-regression indicated that the magnitude of the increase in DIT was influenced by the amount of energy ingested in a breakfast meal (for every 100 kJ increase in energy intake (EI), DIT increased by 1.1 kJ/h (p < 0.001)). DIT was also influenced by macronutrient composition; meals with a high protein or carbohydrate content had higher DIT than high fat meals although this effect was not always significant. Furthermore, DIT was affected by the eating pattern of the meal: consuming the same meal as a single bolus eating event compared to multiple smaller meals or snacks was associated with a significantly higher DIT (meta-analysis,

20 p= 0.02). While this analysis suggests that breakfast consumed as one meal may exert a protective effect on obesity, it must be interpreted with caution as it was based on short term (one day intervention) studies. A surprising finding of this systematic review was the lack of studies looking at DIT and breakfast over longer periods of time, therefore this SR could only make inferences based on short-term effects.

The heterogeneous nature of the breakfast meals found in the literature while completing the SR, and the lack of recent evidence regarding what constitutes an habitual breakfast meal in Australia, led me to determine that more research was needed on what currently constitutes breakfast in multicultural Australia. The ‘Typical Aussie Bloke’study was then formulated to address the lack of evidence about habitual breakfast habits among men and the fact that no data were available regarding habitual breakfast habits for an Australian population since the 1995 National Nutrition Survey. I investigated what foods and beverages currently constitute a typical breakfast amongst younger Australian men. A multi-site cross sectional study examined the breakfast habits of 112 young (18-44 y) Australian men, and found that the majority of men (83.5%) were Habitual Breakfast Eaters (HBE) (consumed breakfast ≥5 times/week) and 84% of them consumed breakfast between 5.01 to 8.00 am. A typical breakfast (≥5 times/week) for the majority of HBE was a combination of one or more of the following foods and/or beverages: coffee (40.4%), breakfast cereal (50.0%), milk for cereal (51.1%), fruit (28.7%), toast (13.8%), spreads (11.7%), and/or yogurt (12.8%). This typical breakfast may also include (1-4 times/week) eggs (58.5%), bacon (30.9%), juice (19.1%), and/or tea (17.0%).

In addition to examining breakfast habits in detail, the Typical Aussie Bloke Study investigated relationships between habitual breakfast consumption and the number of daily eating events, anthropometric measures, metabolic parameters and lifestyle characteristics of a sample of young Australian men. To the author’s knowledge, there have been no other studies the size of the TAB that have collected detailed data on anthropometric and metabolic measurements and compared them with breakfast habits data, timing of breakfast consumption or eating frequency. However, the proportion of Habitual Breakfast Skippers (HBS) was low among our sample, limiting our ability to compare health and lifestyle variables among HBE and HBS. I did ascertain that certain socio-demographic characteristics were related to breakfast patterns. HBE were 21 significantly more likely to have a university qualification (62.8%) than Occasional Breakfast Eaters (OBE) (28.6%) and HBS (20.0%). Early (before 8am) Breakfasts Eaters (EBE) were more likely to be older (p=0.0124), married (Early 59.4% vs Late 26.67p=0.023), have full time jobs (Early 69.8% vs Late 13.3%; p=0.01), earn ≥ AUD50K (Early 69.7% vs Late 26.7%; p=0.043) and have more dependent children (Early 35.6% vs Late 6.7%, p=0.032). However, in both groups the majority of men in this study had no dependent children.

It was also found that sleeping and waking habits did not differ significantly between HBE, OBE and HBS. However, differences were found amongst men consuming the breakfast meal Early vs Late, with a significantly higher proportion of EBE going to sleep before 11.00 pm and waking up before 8.00 am relative to Late Breakfast Eaters (LBE) (both p<0.001). Physical activity levels and fruit and vegetable consumption did not differ significantly between HBE, OBE and HBS, or between EBE and LBE. Although there was a significantly (p=0.015) higher percentage of HBE who consumed 5 or more daily eating events (59.6 %) in comparison to OBE (28.6%) and HBS (20%) who were instead more likely to eat 1-4 times/day. No differences in daily eating events were found amongst EBE vs LBE. Furthermore, differences in sleeping habits were found between men with different frequency of eating events. Men consuming 5 or more eating events per day (n=60; Frequent Eaters) were significantly more likely to go to sleep earlier than 11.00 pm (p=0.021) than men consuming 1-4 eating events per day (n=50; Less Frequent Eaters). Physical activity levels, fruit and vegetable consumption, and waking habits did not significantly differ between Frequent and Less Frequent Eaters. Metabolic and anthropometric parameters including BMI, waist, hip and chest circumferences, body composition, blood pressure, resting metabolic rate, blood glucose and lipid profiles did not differ between HBE, OBE and HBS; between EBE and LBE or between Frequent and Less Frequent Eaters.

Overall, this PhD has contributed to an increased understanding of breakfast consumption and skipping in men, and the associations of different types of breakfast and breakfast cereal with risk factors related to the development of obesity or diabetes in women. The findings of this PhD do not support the concept that consumption of any type of breakfast or breakfast cereal may be associated with a reduction in risk factors related to the development of obesity or chronic disease risk. Instead, it suggests that the

22 types of breakfast foods and/or breakfast composition may be important in any relationship between breakfast and health outcomes. This PhD work has also highlighted the lack of research examining the role of breakfast in longer term studies, suggesting the need to better investigate the role of breakfast in relation to total daily eating events, total daily energy intake and the overall quality of the diet. This would allow adjustment for other eating factors that may bias the association between breakfast consumption and health parameters. Finally, longer term intervention trials and cohort studies are needed to investigate the effects of breakfast consumption per se, different types of breakfast and breakfast cereal on daily eating patterns, on metabolism, body size and chronic disease risk in adults, using robust study designs and adequate sample sizes.

Chapter 1: Breakfast Consumption Habits in relation to Countries, Gender, and Socio-Economic Background in Adults

1.1 Definitions Before proceeding to discuss the literature and the metabolic effects of breakfast, it is necessary to first define some of the terms used.

1.1.1 Energy Balance Energy balance is defined as the balance between energy intake (EI), which is the energy (kJ) ingested in food and drinks, and energy expenditure (EE), which is the energy (kJ) burned by the body (National Heart, Lung and Heart Institute – National Institute of Health (NHLHI-NIH), 2014).

1.1.2 Energy Expenditure

1.1.2.1 Resting Metabolic Rate The quantity of energy used to maintain physiological function during resting conditions is defined as Resting Metabolic Rate (RMR) (Vandarakis, Salacinski et al. 2013); this is also referred to as Resting Energy Expenditure (REE).

1.1.2.2 Diet Induced Thermogenesis Diet Induced Thermogenesis (DIT) or the Thermic Effect of Food (TEF) is defined as the increase in RMR as a result of the consumption of a food or meal (Weststrate 1993, Reed and Hill 1996). Westerterp (Westerterp 2004) defined DIT as EE above fasting divided by the energy ingested from the food eaten, and it is generally expressed as a percentage (Westerterp 2004). DIT can be described by the following formula:

DIT = EE above fasting levels energy content of the food ingested

DIT is believed to account for approximately 10% of the total energy expended over 24 hours in energy balance conditions, i.e., when EI = EE (Westerterp 2004). Meal Induced Thermogenesis (MIT), the Thermic Response to Food (TRF), and Postprandial Energy Expenditure (PEE) above RMR are also terms used in this field, and they have the same meaning as DIT or TEF. In this thesis, DIT will be the term used.

1.1.3 Breakfast Cereal Breakfast cereal can be defined as a grain based food product usually made from oats, rice, wheat or corn, which may be minimally processed, such as drying and rolling the grain (eg. rolled oats), or cooked and flaked or puffed. Several grain varieties may be combined, and fruit and/or nuts added. It is often consumed with milk or yogurt, or in a dry state. Breakfast cereal is often eaten at breakfast, but it can also be consumed as a snack or at other meals during the day (Quatela, Callister et al. 2017).

1.2 Breakfast

1.2.1 Defining Breakfast A systematic review investigating the quality, content and context of breakfast amongst 24 non-interventional studies found that there is no universal definition of breakfast (Mullan and Singh 2010). Of these 24 studies, only eight defined breakfast, and these used differing definitions, which may have contributed to heterogeneous findings among studies. This issue of inconsistent definition has been echoed in a very recent review, which also summarised the heterogeneous definitions available in the literature (O'Neil, Byrd-Bredbenner et al. 2014). Some examples of the issues described in both reviews will be explained in this section.

The first area of disagreement in the definition of breakfast is whether breakfast can just be the ingestion of a beverage or if it has to also contain, or only contain, the consumption of food. Some authors considered breakfast to be any intake of food and/or beverages (Haines, Guilkey et al. 1996, Siega-Riz, Popkin et al. 1998, Aranceta, Serra- Majem et al. 2001) therefore, considering beverage an important part of breakfast. However, other authors do not consider beverages alone to be sufficient for breakfast. Some authors defined breakfast as the ingestion of any food (Wilson, Parnell et al. 2006) or at least one food group (O'Neil, Byrd-Bredbenner et al. 2014) or as an eating event in a certain pre-defined period of time (Affenito, Thompson et al. 2005, Barton, Eldridge et al. 2005, Albertson, Thompson et al. 2008, Albertson, Affenito et al. 2009).

The second main area of disagreement is the period of time during a day when a meal is defined as breakfast. Many studies differentiate the timing of breakfast consumption between weekends and/or holidays and weekdays (Aranceta, Serra-Majem et al. 2001, Affenito, Thompson et al. 2005, Barton, Eldridge et al. 2005, Albertson, Thompson et al. 2008, Albertson, Affenito et al. 2009). Some examples of this difference include 25 breakfast to be consumed between 5am and 10am on weekdays and 5am and 11am on weekends (Affenito, Thompson et al. 2005, Barton, Eldridge et al. 2005, Albertson, Thompson et al. 2008, Albertson, Affenito et al. 2009) or breakfast to be consumed between 6 and 10am during weekdays and 6 to 11am during the weekend days and/or holidays (Aranceta, Serra-Majem et al. 2001). Some definitions do not limit the time of day when breakfast should be consumed at all. For instance, some studies describe breakfast as any food or beverage ingested in a meal that the subjects considered as breakfast (Nicklas, Reger et al. 2000, Cho, Dietrich et al. 2003, Deshmukh-Taskar, Nicklas et al. 2010).

In some studies the authors define breakfast in terms of fasting or waking. For instance, one review defined breakfast as the first meal of the day that breaks the fasting state after the longest episode of sleep and is eaten within two to three hours after waking (O'Neil, Byrd-Bredbenner et al. 2014). It consists of food or beverage containing at least one food group and it can be eaten in any place (O'Neil, Byrd-Bredbenner et al. 2014). In another example, the authors defined breakfast as the first meal of the day consumed within two hours of waking, before or at the start of daily activities (for example working or travelling). It provides from 20% to 35% of total energy required daily and it is normally eaten by 10am (Timlin and Pereira 2007).

Other authors have described which foods should be consumed for breakfast. For example, Monteagudo et al. defined breakfast as the first meal of the day consumed before or at the start of the daily activities that provides 20% to 25 % of total energy required and includes foods such as healthy fats, cereals, fruit and dairy products (Monteagudo, Palacín-Arce et al. 2013). In other definitions, breakfast was considered to include a cereal product and at least one of the following foods: milk, eggs, meat or fish, fruit or fruit juice (Sjoberg, Hallberg et al. 2003). The difficulty of using food groups to define breakfast is that each culture has different foods they consume at breakfast (Grivetti 1995). In addition, it has been found that the food items that are considered to be part of a breakfast meal differ by gender, ethnicity, socio-demographic characteristics and education level (Grivetti 1995, Siega-Riz, Popkin et al. 2000).

It is clear that there is a need to develop a universally clear definition of breakfast with regards to the food type and the timing of consumption (Mullan and Singh 2010).

1.2.1.1 Breakfast Definition for this Thesis In this PhD, breakfast will be defined as the first meal of the day consumed within two hours of waking up before or at the start of the daily activities (such as working or travelling, etc.) either from 5 am to 10 am during weekdays or from 5 am to 11 am on weekend days with more than 837 kJ (200 kcal) ingested (Siega-Riz, Popkin et al. 1998, Affenito, Thompson et al. 2005, Barton, Eldridge et al. 2005, Timlin and Pereira 2007, Albertson, Thompson et al. 2008, Albertson, Affenito et al. 2009).

This definition includes only the consumption of food and or beverages which contain at least 837 kJ (200 kcal) because consuming less or equal to this amount is considered snacking by the American Dietetic Association (ADA) (2010).

1.2.2 Breakfast Eating and Breakfast Skipping There is also no universally agreed definition of ‘breakfast skipping’ (BS) to assist in the definition of ‘breakfast eating’ (BE). Some studies do not define these terms and where they are defined, there is no consistency. Some of the definitions summarised by O’Neil et al. (O'Neil, Byrd-Bredbenner et al. 2014) or found in the literature in this field are described below.

Sugiyama et al. (2009) and Dubois et al. (2012) defined BE as consuming breakfast every day of the week and BS as consuming breakfast less than 7 days per week (Dubois, Girard et al. 2009, Sugiyama, Okuda et al. 2012). Smith et al. (2013) defined BE consuming breakfast 5 or more times a week and BS as rarely or never consuming breakfast or eating breakfast less than or equal to twice per week (Smith, McNaughton et al. 2013). Keski-Rankonen et al. (2003) defined BS as not consuming breakfast at home (Keski-Rahkonen, Kaprio et al. 2003). The definitions that will be used in this thesis for regular breakfast eaters and regular breakfast skippers are the following:

1. Habitual Breakfast Skippers are those who eat breakfast rarely or never or up to two times per week 2. Habitual Breakfast Eaters are those who consume breakfast five or more times per week. 3. Occasional breakfast eaters as those who consumed breakfast three or four times per week.

1.2.3 Influences on Breakfast Consumption This section describes the Australian breakfast habits from the National Nutrition Survey (NNS) collected over 20 years ago (1995). This was the last updated large scale survey which collected and analysed habitual breakfast information in Australia. There has been another national health survey (The Australian Health Survey (AHS) conducted in 2011-12) in 2011/12 that contained national food data but this has not been analysed for breakfast consumption rates.

1.2.3.1 Gender Australians’ breakfast habits are described by the Australian Bureau of Statistics (ABS) in the results of the 1995 NNS (Australian Bureau of Statistics (ABS) 1997). This survey included 13,858 participants from 2 years old and over. Among those 19 years or older, 77.2 % were Habitual Breakfast Eaters and 16% were Habitual Breakfast Skippers (consuming breakfast rarely or never or 1 to 2 times per week) (ABS 1997). Another 6.3% consumed breakfast three to four days per week (ABS 1997). This survey also found gender differences: 73.5% of men aged 19 years or older were Habitual Breakfast Eaters and 19.1% were Habitual Breakfast Skippers whereas 80.8% of women of this age were breakfast eaters and 13.1% were breakfast skippers (ABS 1997).

1.2.3.2 Income From the secondary data analysis of the 1995 NNS carried out by Williams et al. (2002), habitual breakfast consumption (eating breakfast five or more times per week) differed in relation to income level (Breakfast eaters: Quintile 1 (lowest income) 83.9%, Quintile 5 (highest income) 79.7%; p<0.005) (Williams 2002). This study demonstrated that people with the lowest incomes had significantly higher breakfast consumption compared with people from all of the other higher income groups (Williams 2002).

1.2.3.3 Site of Breakfast Consumption The data from the 1995 NNS is summarised with respect to where breakfast was consumed in Australia. These data show that the majority of Australian adults consumed breakfast at home (90.6% males and 93.8% females aged ≥ 19 y). A minority of Australians consumed breakfast in a shop, restaurant or café (8.5% males and 4.8% females aged ≥ 19 y) with only a very small percentage of people consuming breakfast in other places (0.9% men and 1.4% women aged ≥ 19 y) (Williams 2002). It is

28 important to notice that these data from the 1995 NNS may not reflect the current situation, as eating habits are expected to change over time (OXFAM 2011). The ABS 2009-2010 Household Expenditure Survey reports that the expenditure for the consumption of meals out and fast foods have increased up to 50% in a six years period (ABS 2009-10). Australian spent almost a third of their annual household income (28%) on eating out and fast foods consumption in 2006 (ABS 2006). Therefore, considering the rising trend of consumption of meals outside home, it is possible that a higher percentage of people now may consume breakfast outside their homes.

1.2.3.4 Culture/Ethnicity Comparing Australian breakfast eaters with American trends, it appears that breakfast consumption was similar between the two nations in the 1990s. In particular, 74.8% of Americans aged 18 to 65 years consumed breakfast in 1989-1991 (Haines, Guilkey et al. 1996) which is similar to the 77.2% of Australian adults aged 19 years or older who were Habitual Breakfast Eaters (ABS 1997). However, comparing the Australian breakfast consumption habits with the Canadian habits, a higher proportion of Canadian consumed breakfast; 89% Canadians compared to 77.2% Australians (Barr et al. 2013). With regards to breakfast skipping, 17.3% of Americans and 11% of Canadians population skipped breakfast (Siega-Riz, Popkin et al. 2000), which is much higher than the 8.5% of Australians (ABS 1997). It is important to note that the American and Canadian data used were from 1 day 24 hour recalls (Haines, Guilkey et al. 1996), therefore, not allowing an assessment of the regularity of breakfast consumption. For example, American data was only a measure of the percentage of participants who ate breakfast from 5 to 9 am in the 24 hours prior to the recall day of the study (Haines, Guilkey et al. 1996).

A cross-sectional study conducted by Chen et al. (Chen, Cheng et al. 2014) in China in 2012-2013 included 24,159 Chinese aged 12-80 years old and found that 50.2% of people reported always consuming breakfast, 30.1% reported sometimes consuming breakfast, and 19.6% reported they rarely ate breakfast. These data show that in China the proportion of people who regularly consumed breakfast in 2012 was lower than in both America in 1989-1990 (74.8%) and Australia in 1995 (77.2%) (Haines, Guilkey et al. 1996, Williams 2002, Chen, Cheng et al. 2014). However, it is important to consider that the data published in China are more recent, and that the previous American and

Australian data may no longer represent the current breakfast habits in these two countries.

In Italy, a cross-sectional study was conducted in 168 undergraduate university students aged 18-25 years (Isa and Masuri 2011). This study found that 73.6% of students reported having breakfast and 26.4% reported not having breakfast on the day of data collection (Isa and Masuri 2011). The percentage of people consuming breakfast in Italy was similar to the Australian (77.2%) and American data (74.8%) (Haines, Guilkey et al. 1996, Williams 2002, di Giuseppe, Di Castelnuovo et al. 2012). However, this Italian survey represented the breakfast consumption of only a very young population group (18-25 years old) and therefore these data may not be representative of the entire adult population (Isa and Masuri 2011). Furthermore, it is important to note that this study did not assess the regularity of breakfast consumption but only the percentage of people having breakfast on the day of data collection (Isa and Masuri 2011).

1.2.4 Food Selection In Australia the 1995 NNS is the most updated evidence available about breakfast consumption. In author’s knowledge, no more recent papers about nutritional surveys have been published to assess the breakfast habits of Australian adults. However, a more recent survey (The AHS conducted in 2011-12) reported data about total breakfast cereal consumption from a 24 hours recall. Specifically, 38.6% of people aged 19 years over consumed ready to eat breakfast cereal and 6.2% consumed hot porridge style breakfast cereal. The food habits reported by the NNS in 1995, including breakfast habits have likely changed since 1995.’ The likelihood of foods habits changing with time is supported by The OXFAM international survey in 2011 which found that in Australia 62% of participants reported to have changed eating habits since two years before (2009) (OXFAM 2011).

The type of breakfast eaten differs between different countries (table 1) (Siega-Riz, Popkin et al. 2000, Williams 2002, di Giuseppe, Di Castelnuovo et al. 2012, Bai, McCluskey et al. 2014). Also, other factors including gender, ethnics groups, income, education is associated with breakfast habits (Siega-Riz, Popkin et al. 2000).

The difference in foods and beverages consumed for breakfast between different countries (Australia, America, China and Italy) are illustrated in Table 1. Australian breakfast habits are described in the secondary data analysis of the 1995 NNS (Williams 30

2002) which illustrates that milk (65.6 % of men and 68.7% of women), cereal products (including also breads) (76.2% of men and 81.2% of women) and fruit (14.3% of men and 19.9% of women) were the main foods consumed for breakfast by the majority of Australian adults.

The secondary data analysis conducted by Siega-Riz et al. of the 1994-1996 Continuing Survey of Food Intake by Individuals shows that the main foods and beverages consumed by Americans for breakfast were: bread (21.7% of people), coffee (15.1% of people), eggs with either cereal and/or cooked cereal and /or bread (15.3% of people) and cereals with either milk and/or bread and/or eggs (17.4% of people) (Siega-Riz, Popkin et al. 2000).

Howden et al. in his review described the traditional Chinese breakfast from unpublished data by Yap which consisted of rice porridge, dim sum and tea, if eaten at home and of rice porridge and noodle dishes if eaten out (Howden, Chong et al. 1993). Bai et al. conducted a survey investigating non-traditional breakfast products consumed by households in Beijing in 2007, in Nanjing in 2009 and in Chengdu in 2010 (Bai, McCluskey et al. 2014). This study found that 83% of the total households included non-traditional foods in their breakfast. In particular, 73.3% consumed fluid cow’s milk, 47.3 % consumed breads, 16.1% consumed cakes, 12.4% consumed yogurt and 8.5% consumed cereals. The findings of Bai et al. show that a large percentage of Chinese (83%) living in these three cities (Beijing, Nanjing and Chengdu) included at least one food group from non-traditional food sources in their breakfast; thus, showing a trend towards consuming a more westernised type of breakfast in China (Bai, McCluskey et al. 2014).

The Moli-sani project conducted in Italy between 2005-2009 reported amongst 18,177 subjects and reported by di Giuseppe et al. (2012) found that the main foods/beverages consumed by Italian adults for breakfast were coffee (93.3%), sugar (81.6%), crispbread/rusk (87.9%), milk (67.7%), biscuits (42.9%), yogurt (42.8%), brioche (38%), jam (37.3%) and tea (33.4%) (di Giuseppe, Di Castelnuovo et al. 2012). A smaller but still considerable proportion of the population consumed breakfast cereal (17%) (di Giuseppe, Di Castelnuovo et al. 2012).

1.2.4.1 Gender Gender differences were also found in breakfast consumption in American and Australian populations. The findings of Williams et al. (2002) in Table 1.1 illustrate that Australian males and females significantly differed with regards to fruits, eggs, sugar and honey consumption at breakfast (Williams 2002). The findings from Siega-Riz et al. (2000) showed that American men were more likely to consume eggs for breakfasts (18% of men while 13% of women, adjusted values) and less likely to consume bread than women (20% of men and 25% of women, adjusted values) (Siega-Riz, Popkin et al. 2000).

1.2.4.2 Culture/Ethnicity Siega-Riz et al. (Siega-Riz, Popkin et al. 2000) also suggested that breakfast habits differed between ethnic groups in the USA. African-American and Hispanic groups were more likely to eat eggs (African American 24%, Hispanic 18%, Anglos 14%, adjusted values) whereas Anglos were more likely to consume ready to eat (RTE) cereal (Anglos, 20%, African-American 12%, Hispanic 12%, adjusted values) (Siega-Riz, Popkin et al. 2000).

1.2.4.3 SES/Income Siega-Riz et al. (2000) also found differences between higher and lower socioeconomic classes (Siega-Riz, Popkin et al. 2000). Those with lower levels of income were more likely to consume eggs compared to people with higher incomes, who were more likely to eat cereal and bread (lowest income: 19.8% consumed eggs, 13.8% consumed RTE cereals and 18.7% consumed bread; highest income: 11.7% consumed eggs, 20.5% consumed RTE cereal and 24.5% consumed bread) (Siega-Riz, Popkin et al. 2000).

1.2.4.4 Education Education also appears to be an important factor for breakfast habits among Americans. Those people who attained an education level lower than grade 12, adjusting for all the other characteristics, had a 19% probability of eating eggs and a 14% probability of eating RTE cereal (Siega-Riz, Popkin et al. 2000). Participants with a college degree, adjusting for all the other factors, had a 10% probability of eating eggs and a 22% probability of eating RTE cereal (Siega-Riz, Popkin et al. 2000).

To summarise, different foods/beverages are consumed for breakfast between different countries (Siega-Riz, Popkin et al. 2000, Williams 2002, di Giuseppe, Di Castelnuovo et 32 al. 2012, Bai, McCluskey et al. 2014). Furthermore, the findings of Siega-Riz et al suggest that different ethnic groups are prone to different breakfast habits within same countries Many other factors such as gender, education and income can affect the food consumed for breakfast (Siega-Riz, Popkin et al. 2000).

Table 1.1 Foods and beverages consumed for breakfast in Australia, America, China and Italy.

Countries Study’s characteristic Population’s Foods and/or beverages consumed for breakfast % of people References characteristic Australia Secondary analysis of the 13,858 cereal products (cold cereal, hot cereal, breads, pastries, cakes and biscuits) 76.2% of men and 81.2% of women (Williams 1995 National Australian ≥ 2 2002) Nutritional Survey years old and milk 65.6 % of men and 68.7% of women (NNS) over. Data tea or coffee 57.5% of men and 64.1% of women displayed a from ≥19 fruit 14.3% of men and 19.9% of women years a sugar and honey 46.6% of men and 34.3% of women a eggs 7.65% of men and 4.9% of women cooked breakfast Less than 10% of all adults

USA Secondary analysis of the 15,641 cereals with either: milk or milk and bread or milk and cooked cereal or any other 17.4% (Siega-Riz, 1994-1996 Continuing Americans primary food (such as eggs or bread) Popkin et al. Survey of Food Intake by aged 18-65 y bread 21.7% 2000) Individuals coffee and/or soft drink and/or high fat dessert 15.1%

fruit 5.5% eggs with any other primary food groups (cereal and/or cooked cereal and/or 15.3% bread) miscellaneous breakfast foods (such as low fat milk items, soups or low-fat 3.3% dessert). China Survey investigating 770 cereal 8.5% (Bai, non-traditional breakfast households in McCluskey bread (including sweets buns) 47.3% products consumed by China et al. 2014) households in 2007 in cake (including pancake and pies) 16.1% Beijing, in 2009 in powder milk 2.9% Nanjing and in 2010 in Chengdu. fluid cow milk 73.3%

yogurt 12.4%

cheese and butter 1.3%

coffee 0.8%

juice 2.1%

soft drink 0.9%

burgers and hamburgers 1.8% sandwich 0.4%

sausage 7.3%

Italy 2005-2009 Moli-sani 18,177 breakfast cereals 17.0% (di Project subjects aged Giuseppe, biscuits 42.9% ≥ 35 y Di brioche 38.0% Castelnuovo et al. 2012) crispbread/rusk 87.9%

milk 67.7%

yogurt 42.8% tea 33.4%

coffee 93.9% sugar 81.6% honey 10.4% jam 37.3%

a= p<0.001comparing men and women.

Chapter 2: Critical Literature Review of Breakfast Consumption Habits in relation to Obesity and Chronic Diseases Risk in Adults

2.1 Overview It is estimated that 63% of Australian adults were either overweight or obese from the 2011- 2013 Australian Health Survey (ABS 2013). Being overweight or obese is one important risk factor for the development of chronic diseases such as Cardiovascular Disease (CVD) and diabetes (World Health Organisation (WHO) 2017). Chronic diseases affect a large proportion of the Australian population with 22% of the adult population diagnosed with CVD and 5.4% with diabetes (Australian Institute of Health and Welfare (AIHW) 2017). Breakfast is often described as the most important meal of the day (Brown, Bohan Brown et al. 2013) as it is believed to contribute to good health and nutrition by protecting against body weight gain (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013) and chronic diseases risk, especially diabetes and heart diseases (Cahill, Chiuve et al. 2013, Bi, Gan et al. 2015, Yokoyama, Onishi et al. 2016). However, the evidence available to support its protective effects is often limited and contradictory (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013, Casazza, Brown et al. 2015).

A number of mechanisms have been proposed to explain how breakfast may provide protective effects against obesity. Breakfast potentially contributes to an increase in total daily energy expenditure (EE) in the form of Diet Induced Thermogenesis (DIT) by stimulating the metabolism in the morning by breaking the hypo-metabolic morning fasting state (Casazza, Brown et al. 2015) but the evidence is limited (Farshchi, Taylor et al. 2005, Kobayashi, Ogata et al. 2014, Reeves, Huber et al. 2015). Another hypothesis is that the boost in energy provided by breakfast may increase the likelihood of performing physical activity (PA) in the morning (Betts, Richardson et al. 2014); however the majority of trials do not support this belief (Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017). Another possible mechanism of action is that breakfast consumption may result in decreased total energy intake (EI) throughout the day (Farshchi, Taylor et al. 2005) due to a reduction in snacking; however, the majority of the studies conducted on this topic do not support this hypothesis (Astbury, Taylor et al. 2011, Halsey, Huber et al. 2012, Levitsky and Pacanowski

2013, Betts, Richardson et al. 2014, Reeves, Huber et al. 2014, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017)

Among the other explanations for a protective role of breakfast on chronic disease development is that it may be associated with improved nutritional adequacy (Barr, DiFrancesco et al. 2013). Nutrient adequacy is believed to play a significant role in the prevention of chronic diseases (WHO 2017). Furthermore, although the evidence is limited and conflicting between cross sectionals studies and randomized cross overs trials, there is a suggestion that the prolonged fasting state of skipping breakfast may be detrimental to blood lipid or glucose profiles and therefore, breakfast consumption would be protective (Farshchi, Taylor et al. 2005, Astbury, Taylor et al. 2011, Deshmukh-Taskar, Nicklas et al. 2013, Kobayashi, Ogata et al. 2014, Meksawan, Pongthananikorn et al. 2014).

Overall, there is insufficient evidence to support the idea that breakfast plays a protective role in obesity and chronic disease prevention. It is not possible to clearly support any of the mechanisms of action (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013). Therefore, an aim of this PhD was to further investigate the effects of breakfast on metabolic parameters, body size and chronic disease development.

2.2 Is Breakfast Protective against Obesity? Casazza et al. (2013) stated that certain common beliefs regarding obesity are only presumptions, as they are not supported by evidence, and therefore there is a need for more research to ascertain their real effect (Casazza, Fontaine et al. 2013). The authors indicated that the belief that regularly eating (vs. skipping) breakfast protects against obesity (causative link) is based on conjecture rather than evidence that skipping breakfast leads to overeating later in the day. Brown et al. (2013) evaluated the evidence that regularly skipping breakfast leads to obesity using a meta-analysis of 88 studies undertaken in thirty countries on 5 continents and with mixed subpopulations. This analysis capitalised on a previous meta- analysis and three previous SRs in this area (Rampersaud, Pereira et al. 2005, Szajewska and Ruszczynski 2010, Horikawa, Kodama et al. 2011, Mesas, Munoz-Pareja et al. 2012). The meta-analysis found extensive evidence from observational studies to support a significant association between breakfast omission and weight gain (p=10-42) (Brown, Bohan Brown et

al. 2013); however, only limited and contradicting evidence was available from trials (Schlundt, Hill et al. 1992, Geliebter A 2000, Farshchi, Taylor et al. 2005). As observational studies cannot indicate causation, this meta-analysis can only provide evidence of an association between breakfast consumption and protection from risk of obesity.

Few trials (Schlundt, Hill et al. 1992, Geliebter A 2000, Farshchi, Taylor et al. 2005, Betts, Richardson et al. 2014, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017) have been conducted to investigate any causative role of breakfast consumption on weight change and these studies did not find consistent results. Furthermore, there was significant heterogeneity between these trials (Schlundt, Hill et al. 1992, Geliebter A 2000, Farshchi, Taylor et al. 2005, Betts, Richardson et al. 2014, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017). The trials have varied in participants’ age and weight, the length of interventions, and the type of breakfast provided. This has limited the capacity to draw any conclusions on any causative relationships between breakfast consumption vs skipping on obesity.

Interestingly, one trial in USA found that the effectiveness of a breakfast intervention on weight loss relied on previous breakfast eating (BE) habits (Schlundt, Hill et al. 1992). This study found that greater weight loss occurred when breakfast skippers were assigned to a BE intervention and when breakfast eaters were assigned to a breakfast skipping (BS) intervention in 52 obese women (p<0.06; breakfast eaters, mean (SD) BE 6.2 (3.3) kg, BS 8.9 (4.2) kg; breakfast skippers, BE 7.7 (3.3) kg, BS 6.0 (3.3) kg)). However, this study (Schlundt, Hill et al. 1992) only provided evidence of the effect of habitual breakfast habits in conjunction with weight loss strategies in an obese group.

Another recent Randomised Control Trial (RCT) in USA provided four-week BE and BS interventions to 49 women who were skippers and reported a significant increase in weight (mean (SD) 0.7 (0.8) kg; p < 0.01) in the BE, which was the opposite effect of the previous study (LeCheminant, LeCheminant et al. 2017). Two RCTs in the UK, The Bath Breakfast Projects, provided six week BE and BS interventions during free living conditions to 33 normal weight participants (26 were habitual breakfast eaters) (Betts, Richardson et al. 2014) or to 24 obese participants (14 were habitual breakfast eaters) (Chowdhury, Richardson et al. 2016). Both RCTs found no significant differences in body weight after 6 weeks of intervention (change from baseline: Betts et al. 2014 BE Change from baseline (95% 38

Confidence Interval (CI) -0.2 (-0.8, 0.4): BS -0.4 (-0.8, 20.1); Chowdhury et al. BE 1.0 (0.2, 1.7); BS 0.2 (-0.5, 1.0) (Betts, Richardson et al. 2014, Chowdhury, Richardson et al. 2016).

Brown et al. published a meta-analysis of eight RCTs in 2017 (Brown, Milanes et al. 2017). Of these eight trials, three provided the meals and five gave recommendations or meal plans; the breakfast interventions ranged from 2 to 16 weeks in duration. This meta-analysis does not support the hypothesis that eating breakfast is protective against obesity. Limitations were that the types and lengths of interventions were heterogeneous, but review indicates that limited data are available to answer this question.

Furthermore, neither of the meta-analyses by Brown et al. (Brown, Bohan Brown et al. 2013, Brown, Milanes et al. 2017) investigated the effects of different types of breakfast (e.g., breakfast cereal vs cooked breakfast) on obesity, therefore this is an area that still needs to be explored. It is possible that the type and quality of breakfast provided may make a substantial difference in body weight status rather than the consumption of any breakfasts per se. It is also possible that the heterogeneity in participants’ age, weight and previous breakfast eating habits in the trials were confounding the association of breakfast consumption with weight status. Also, the effect of breakfast on body weight may be minimal in the short term but may become substantial over the longer term, perhaps in years, and these trials could not find a significant effect because their investigations were limited to a relatively short term timeframe.

Therefore, there is a need for longer term trials, as well as trials that investigate the effects of different types of breakfasts, with more homogeneous participants (weight status, previous eating habits) to better explore whether there is a link between breakfast consumption and the risk of obesity.

2.3 Mechanisms Proposed to Explain a Protective Effect of Breakfast on Obesity The reviews of both Brown et al. (2013) and Clayton et al. (2016) suggest that breakfast may protect against obesity by creating a negative energy balance, either by decreasing EI or increasing EE, but the evidence available in this area is contradictory or unclear (Brown, Bohan Brown et al. 2013, Clayton and James 2016). One hypothesis is that eating breakfast may result in a lower cumulative EI during the rest of the day due to reduced snacking (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013); however, the majority of

the studies conducted in this area do not support this hypothesis (Astbury, Taylor et al. 2011, Halsey, Huber et al. 2012, Levitsky and Pacanowski 2013, Betts, Richardson et al. 2014, Reeves, Huber et al. 2014, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017). Another hypothesis is that eating breakfast may result in an increased EE due to foods/beverages ingested through DIT or due to an increase in PA levels (Clayton and James 2016); however, the evidence available for an effect of breakfast on DIT (Farshchi, Taylor et al. 2005, Kobayashi, Ogata et al. 2014, Reeves, Huber et al. 2015) and PA (Halsey, Huber et al. 2012, Betts, Richardson et al. 2014, Reeves, Huber et al. 2015, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017) is limited and contradictory. These mechanisms of action are fully explored below.

2.3.1 The Effect of Breakfast Consumption on Energy Intake A number of studies have investigated the hypothesis that eating breakfast may result in a lower cumulative EI during the rest of the day, possibly due to reduced snacking. The majority of these studies do not support this hypothesis, because they have either found an increase in EI in the breakfast consumption group (Levitsky and Pacanowski 2013, Betts, Richardson et al. 2014, Reeves, Huber et al. 2014, LeCheminant, LeCheminant et al. 2017) or did not find a significant difference between interventions (Astbury, Taylor et al. 2011, Halsey, Huber et al. 2012, Chowdhury, Richardson et al. 2016). Only one study supported this hypothesis, as it found an increase in daily EI in the BS group (Farshchi, Taylor et al. 2005).

Three studies using randomised cross-over designs (RCODs) found no significant differences on daily EI after BE and BS interventions. One trial provided each intervention for only one day to 20 regular breakfast eaters and normal weight men in the UK (Astbury, Taylor et al. 2011) and found a significantly higher EI at lunch in the BS group compared to the BE group (BS group consumed mean ± standard error of the mean (SEM) 5.76 ± 0.403MJ vs BE group consumed 4.90 ± 0.455MJ; p < 0.01) (Astbury, Taylor et al. 2011). However, the calculated combined EI from breakfast, preload and ad libitum test meal was not significantly different between the two groups (BS group 6.83 ± 0.403 MJ vs BE group 6.8 ± 0.455 MJ) (Astbury, Taylor et al. 2011). The other study (Halsey, Huber et al. 2012) provided the interventions for one week to 49 participants (26 F; 23 M; all breakfast eaters) in the UK and reported no significant difference in EI between the two interventions (p=0.131). The results were consistent both in males (3 days food records: BE mean ± SEM 8314 ± 448 kJ/d vs BS 7514 ±

368; p=0.127) and females (3 days food records: BE 7778 ± 410 kJ/d vs BS 7531 ± 536 kJ/d; p=0.567) (Halsey, Huber et al. 2012). Another RCT, the Bath Breakfast Project conducted in the UK, provided 6-week interventions during free living conditions to 24 obese subjects (14 were habitual breakfast eaters). This trial also did not find a significant difference in EI between the two interventions (difference: 1414 kJ/d; 95% CI: −313, 988) (Chowdhury, Richardson et al. 2016).

The other four trials reported a significant increase in overall daily EI in the BE trial, which was the opposite of the hypothesis. One RCOD provided one day intervention for both arms to 16 normal weight adults (13 F, 3 M; 11 regular breakfast eaters, 5 regular breakfast skippers) in the USA and reported that during the BE intervention participants consumed 450 calories more per day than during the BS intervention (p=0.01) (Levitsky and Pacanowski 2013). Reeves et al. (2014) recruited 37 participants, either normal weight or overweight in the UK (16 M, 21 F; 19 breakfast eaters, 18 breakfast skippers) and reported that during the BE trial participants consumed significantly more total daily EI compared to the BS intervention (7 days food diary: BE mean (SD) 8150 (4042) kJ vs BS 7477 (2159) kJ; p=0.03) (Reeves, Huber et al. 2014). A RCT, the Bath Breakfast Project in the UK, provided six-week BE and BS interventions during free living conditions to 33 normal weight participants of which 26 were habitual breakfast eaters (Betts, Richardson et al. 2014) and found that the BE group consumed significantly more energy per day compared to the BS group amongst normal weight subjects (BE group consumed mean (SD) 11422 (2397) kJ/d vs BS group consumed 9167 (2067) kJ/d, p=0.007). Another recent RCT conducted in the USA provided four-week BE and BS interventions to 49 women who were skippers and found the BE participants to consume 1113 (2075) kJ/day (mean (SD)) more than the BS participants (p<0.01) (LeCheminant, LeCheminant et al. 2017).

The only study by Farshchi, Taylor et al. (2005) supported the hypothesis of reduced overall daily EI in BE. The study provided 14 day breakfast consumption and skipping interventions to 10 healthy women in the USA, and reported the BE trial (mean (SD) 6970 (590) kJ/d) to have a significantly lower EI during the three days of food records than the BS trial (7350 (650) kJ/d, p=0.001) (Farshchi, Taylor et al. 2005).

In conclusion, the majority of the studies do not support the hypothesis that breakfast intake protects against obesity by decreasing total daily EI. The short term interventions and the 41

small sample sizes of the majority of the trials may have impacted on the ability to find an effect. In addition, the majority of the studies included mostly habitual breakfast eaters. Schlundt et al. (1992) suggest that the impact of breakfast consumption on weight status could vary depending on previous habitual breakfast habits (Schlundt, Hill et al. 1992). Therefore, longer term trials need to be conducted to better investigate this potential mechanism of action for the protective effect of breakfast on obesity amongst both habitual breakfast eaters and skippers.

2.3.2 The Effect of Breakfast Consumption on Diet Induced Thermogenesis Breakfast is believed to kick start the metabolism in the morning by breaking the hypo- metabolic fasting state (Casazza, Brown et al. 2015), thus increasing EE in the form of DIT. Even small differences on DIT every day could significantly impact body weight in the long term. Specifically, it has been suggested that a cumulative imbalance of 42–84 kJ/day on average can result in 0.5–1 kg of weight gain annually (Lean and Malkova 2016).

Only a few studies (Farshchi, Taylor et al. 2005, Kobayashi, Ogata et al. 2014, Reeves, Huber et al. 2015) have been conducted to investigate the effect of short term breakfast consumption vs skipping interventions on DIT. These studies had heterogeneous study designs, very small sample sizes and the breakfast interventions were only short term ranging from one day to 2 weeks. Therefore, the results were inconclusive.

A recent short term (one day) RCOD by Kobayashi et al. evaluated the effects of BE and BS on 24-hour DIT in laboratory conditions in a small sample of eight young habitual BE males in Japan (Kobayashi, Ogata et al. 2014). This study provided two interventions with each arm consuming the same calories (mean (SE) 9163 ± 519) kJ/day). The BS intervention group consumed this energy over two meals (lunch and dinner) whereas the BE group consumed the same energy over three meals (breakfast, lunch and dinner) (Kobayashi, Ogata et al. 2014). This trial found that 24-hour DIT did not change significantly with the BE intervention (1389 ± 184 kJ/24 h) compared to the BS intervention (1268 ± 385 kJ/24 h, p=0.76) (Kobayashi, Ogata et al. 2014) and did not support the hypothesis that BE increases the energy expended in the form of DIT in comparison to BS. It possible that the small difference in DIT (~120 kJ) if sustained daily could impact body weight in the long term. Also, the very small sample size in this study may have impacted on its ability to find a significant effect.

Farshchi et al. investigated adaptions to DIT due to short term breakfast consumption interventions and reported no significant difference in DIT following breakfast ingestion after 14 days of BE and BS interventions among 10 women (Farshchi, Taylor et al. 2005). Another RCOD in the UK (Reeves, Huber et al. 2015) compared DIT after a week of a BE intervention between habitual breakfast eaters and skippers (total 37 participants) in order to investigate whether previous breakfast eating habits would play a role. This study did not find a significant difference in DIT between the two groups (breakfast eaters mean (variation)1 147.87 (56.35) kJ vs skippers 156.50 (92.07) kJ over 150 min; P=0.74). Therefore, neither trial supports the hypothesis that short term breakfast consumption interventions significantly stimulate metabolism in the morning compared to skipping breakfast. However, it is important to consider that both studies provided only short-term interventions and had small sample sizes, and this could have impacted on the ability of the studies to find a true effect.

Furthermore, two studies with one day interventions compared DIT after consuming the same calories at breakfast vs afternoon and/or night time, in order to investigate if metabolic differences occur due to meal timing (Weststrate, Weys et al. 1989, Romon, Edme et al. 1993). In a RCOD by Romon et al. (1993), DIT measured by indirect calorimetry was significantly higher when the same snack (20% of total EI) was eaten in the morning compared to the afternoon or at night in nine men aged 28 ± 2 y (DIT as a percentage of energy content of meal (mean ± (SE): morning 15.9 % ± 1.6 %, afternoon 13.5 % ± 1.8 % and night 10.9 % ± 2.2 %; morning vs afternoon p=0.04, morning vs night p=0.002 and afternoon vs night p=0.06) (Romon, Edme et al. 1993). This study suggests that breakfast consumption may be protective against obesity by increasing the DIT in comparison to afternoon or night meals, due to the effect of timing of the meal event. It is important to note that the same 10 hour fasting conditions occurred between the three trials (morning, afternoon and night) (Romon, Edme et al. 1993). These findings suggest that the breaking of the long overnight fast by consuming breakfast may not be the reason for the increase in DIT, as the same long fast was applied during the afternoon and night sessions. Thus, other unknown factors are likely to be involved to increase DIT in the morning compared to afternoon and night when the same EI and the same fasting conditions are applied.

1 Unclear unit of measurement for variation 43

However, a RCOD conducted in the Netherlands by Weststrate et al. (1989) in 10 men aged 22 ± 0.5 years contradicts this idea, leading to uncertainty with regards to this possible mechanism of action (Weststrate, Weys et al. 1989). In the Weststrate et al. trial, the same meal (1906 kJ) was provided during morning and afternoon sessions and the findings demonstrated no significant difference in DIT between afternoon and morning interventions (DIT as a percentage of energy content measured for 4 hours: morning mean (SEM) 7.4 % ± 0.5 % and afternoon 6.9 % ± 0.9 %, 95 % CI: -1.8 % to 2.6 % of energy content of meal) (Weststrate, Weys et al. 1989).

An important aspect to consider is that these two studies (Weststrate, Weys et al. 1989, Romon, Edme et al. 1993) differ in study design and this could have impacted the findings. In particular, Romon et al. (1993) standardised the fasting conditions between each of the three trials while Weststrate et al. (1989) had different fasting conditions between morning and afternoon trials (12-14 hours overnight fasting for the morning session and just 6-7 hours for the afternoon trial) (Weststrate, Weys et al. 1989, Romon, Edme et al. 1993). Furthermore, Weststrate et al. (1989) repeated each intervention three times (morning and afternoon) for each subject whereas Romon et al. (1993) ran the intervention once for each of the morning, afternoon and night sessions. Again, there were small sample sizes in both studies.

The contradictory findings about the effects of BE vs BS and BE vs consuming the same meal for lunch or dinner on DIT could be the results of the study limitations and incongruences. These included the short-term nature of the studies conducted, the small sample sizes, and the poor comparability between studies due to different study designs and different energy and nutrient composition of the meals administered in interventions.

In conclusion, due to the uncertainty of the role of BE on stimulating metabolism in the morning in the form of DIT a systematic review (SR), meta-analyses and meta-regression was conducted as part of this PhD to investigate the role of consuming different macronutrient profiles and different EI for breakfast on DIT. This SR also investigated the impact of different breakfast eating patterns on DIT (Quatela, Callister et al. 2016). The results are presented in chapter 5 of this thesis.

2.3.3 The Effect of Breakfast Consumption on Physical Activity This section summarises the effects of consuming vs skipping breakfast on PA or PA thermogenesis (PAT). Only one of the trials that investigated the hypothesis that consuming breakfast increased PA levels supported this hypothesis (Betts, Richardson et al. 2014). The majority of the trials in this field do not support it (Halsey, Huber et al. 2012, Reeves, Huber et al. 2015, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017).

Findings from the Bath Breakfast Project among normal weight individuals suggest that PAT may be partially responsible for the proposed beneficial effect of breakfast on weight (Betts, Richardson et al. 2014). This study found that participants randomised to the six-week breakfast consumption intervention had a significantly higher PAT in the morning until 12 pm than the participants randomised to the six-week BS intervention (PAT: mean (SD) 2059 (950) kJ/d in the breakfast consumption intervention vs 1301 (519) kJ/d in the BS intervention group, p=0.01) (Betts, Richardson et al. 2014). This effect was also significant and of a larger magnitude when considering the total daily PAT ((BE group 6063 (2787) vs BS 4,213 (1,548); BE group had 1849 kJ/d more for PAT than the BS group, p=0.04). The findings of this study suggest that breakfast promotes a higher level of PAT throughout the entire day. However, findings from the same project conducted in obese subjects found a significantly higher PAT in the morning for the breakfast consumption group (BE mean (SD) 1820 (552) kJ/d vs BO 1033 (715) kJ/d; P = 0.03), but this effect was not sustained after 12.00 PM (3163 (565) kJ/d vs 2828 (2259) kJ/d; P =0.7) or over the entire day (5109 (1092) kJ/d vs 3971 (2966) kJ/d; P=0.3) (Chowdhury, Richardson et al. 2016). Together these two studies suggest that obese individuals may respond differently to BE and BS interventions compared to normal weight subjects, and this factor needs to be taken into consideration before providing breakfast recommendations to adults with different body weight. Another recent RCT (LeCheminant, LeCheminant et al. 2017) provided four weeks BE and BS interventions to 49 women who were BS and reported no significant compensation in PA levels measured by accelerometers in the BS group (28 days of accelerometer measurements: BE light mean (SD) 22.9 (60.7) min, moderate activity 3.2 (9.4) min, vigorous activity 0.2 (2.3) min, BS light 22.8 (47.1) min; moderate 0.28 (10.5) min, vigorous 1.0 (5.2)min; light p= 0.97; moderate p= 0.40, vigorous p=0.34).

Also, two short RCOD trials, both providing one week BE and BS interventions, did not find significant effects on PA levels. Halsey et al. (Halsey, Huber et al. 2012) found a significantly higher morning heart rate in the BE group (7 days morning heart rate measurement: BE mean (SEM) 82 ± 1.5 vs BO 77 ± 1.7 p<0.001) compared to the BS intervention. However, this effect of breakfast consumption on heart rate was not significant when taking the whole day into consideration (7 days heart rate measurement: BE 84 ± 1.5 vs 83 ± 1.6; p=0.235). Mean daily pedometer rate was also not significantly different between the two interventions (7 days pedometers rate: BE 18795 ± 1318 vs BO 17 915 ± 1231p=0.858). Reeves et al. (Reeves, Huber et al. 2015) also did not find a significant difference in pedometer counts over 7 days between BE vs BS interventions amongst 34 participants in the UK (p=0.57, no data provided) (Reeves, Huber et al. 2015).

From the five trials described here, only Betts et al. (2014) reported breakfast consumption to significantly increase PAT (Betts, Richardson et al. 2014); the other four trials did not find a significant effect. Trials were heterogeneous in study designs (RCOD vs RCT), length of interventions (ranging from 1 day to 6 weeks), participant characteristics (weight status, previous breakfast habits) and PA measurement techniques. Overall, these studies do not support the hypothesis.

To date there is no clear consensus on any mechanisms regarding an effect of breakfast on EI, DIT or PA, or whether BE per se helps to reduce obesity.

The majority of the studies recruited mainly habitual breakfast eaters and therefore there was not enough evidence to investigate if habitual breakfast skippers would respond differently to these interventions. Schlundt et al. (Schlundt, Hill et al. 1992) showed that breakfast interventions had different effects amongst habitual breakfast skippers compared to habitual breakfast eaters, suggesting that usual breakfast habits may be important factors to be considered when developing interventions (Schlundt, Hill et al. 1992). Future longer-term studies with larger sample sizes that compare habitual breakfast eaters and skippers are needed to better investigate these possible mechanisms of actions. Finally, much more research is needed in this field to try and unravel the mechanisms involved.

2.4 Is Breakfast Skipping Associated with Increased Risk of Chronic Diseases? A limited number of prospective studies have been conducted investigating an association between BE or BS and chronic disease risk outcomes but they have provided contradictory findings.

A prospective cohort study conducted by Cahill et al. reported a significant protective association of breakfast consumption on Coronary Heart Disease (CHD) risk. The study was conducted amongst 26,902 American men aged 45-82 years from the Health Professionals Follow-up Study, and investigated eating habits in 1992 and incidence of CHD over 16 years, using COX adjusted models (Cahill, Chiuve et al. 2013). This prospective analysis found that BS was associated with a 27% increased risk of CHD compared with breakfast consumption (Relative Risk (RR) =1.27, 95% CI: 1.06-1.53), suggesting a protective role for breakfast consumption against CHD risk (Cahill, Chiuve et al. 2013).

Another prospective study in Japan (Kubota, Iso et al. 2016) was conducted with 82,772 participants aged 45 to 74 years. Multivariate analyses showed that BS (breakfast consumption 0-2 times per week) was not significantly associated with CHD in this cohort (Hazard Ratio (HR) 0.96 (CI 0.73–1.25, p=0.974). However, this study found BS to be significantly associated with an increased risk of total CVD (HR 1.14, CI 1.01–1.27, p= 0.013), stroke (HR 1.18, CI 1.04–1.34, p = 0.007) and cerebral haemorrhage (HR 1.36, CI 1.10–1.70, p= 0.004) (Kubota, Iso et al. 2016), two outcomes not investigated by Cahill et al. (2015). Therefore, this study found an important role for breakfast consumption on protection against certain cardiovascular diseases, but not specifically CHD.

Another prospective study (Yokoyama, Onishi et al. 2016), also conducted in Japan, with 34,128 men and 49,282 women aged 40–79 years also investigated BS. The baseline survey data were collected from 1988 to 1990 and the follow up conducted up to the end of 2009 (median 19.4 years follow-up). This study found that BS was significantly associated with an increased risk of mortality from circulatory diseases (HR=1.42, CI 1.02–2.02, p=0.0485) in men, but not in women (HR 1.19, CI 0.71–1.05, p= 0.1828). Furthermore, BS was associated with all-cause mortality (HR=1.43, CI 1.21- 1.69; p<0.0001) both in men (HR=1.34; CI: 1.04– 1.73, p=0.0234) and in women (Yokoyama, Onishi et al. 2016). This study also investigated the association of BS on cancer risk, but did not find a significant association in either men

(HR 1.27, CI 0.98–1.65, p=0.0741) or women (HR 1.30, CI 0.87–1.94, p=0.1890) (Yokoyama, Onishi et al. 2016).

With regards to the role of BS on diabetes risk, Bi et al. (2015) conducted a recent meta- analysis of eight observational studies with a total of 106,935 participants; three of these studies were undertaken in the USA (Mekary, Giovannucci et al. 2012, Mekary, Giovannucci et al. 2013, Odegaard, Jacobs et al. 2013), two in Japan (Sugimori, Miyakawa et al. 1998, Nishiyama, Muto et al. 2009), two in China (Zhi 2007, Xiao, Wang et al. 2010) and one in Russia (Voronova, Nikitin et al. 2012). Four were prospective studies (follow up ranged from 6 to 18 y) (Sugimori, Miyakawa et al. 1998, Mekary, Giovannucci et al. 2012, Mekary, Giovannucci et al. 2013, Odegaard, Jacobs et al. 2013), three were cross sectional studies (Zhi 2007, Nishiyama, Muto et al. 2009, Voronova, Nikitin et al. 2012) and one was a case control study (Xiao, Wang et al. 2010). This meta-analysis (Bi, Gan et al. 2015) found a significant association between BS and increased risk of type two diabetes from both cross sectional (HR: 1.15; CI, 1.05, 1.24; heterogeneity test: P=0.770 and I2=0.0%) and cohort studies (HR:1.21; CI 1.12, 1.31; heterogeneity test: P=0.984; I2=0.0%) (Bi, Gan et al. 2015).

Overall, studies in different population groups have shown significant protective associations of breakfast on some CVDs but not consistently for the same CVD type. However, inconsistency between studies in the way confounding factors were adjusted is a limitation in the existing literature. A recent meta-analysis of eight observational studies reported evidence to support a BS association with higher risk of type 2 diabetes. Further studies are needed to investigate the effects of diet quality and different types of breakfast foods and macronutrient profiles vs breakfast skipping interventions on diabetes and CHD and CVD risks.

2.5 Mechanisms behind a protective effect of Breakfast Consumption on Chronic Disease Risk Breakfast consumption may be linked to a protective effect on risk of chronic diseases such as diabetes and heart disease, by multiple proposed mechanisms. Firstly, as obesity is associated with an increased risk of chronic disease (WHO 2016), the potential protective effect of breakfast consumption on obesity may play a significant role in chronic disease protection. Higher PA is a protective factor against chronic disease development (WHO 2017), therefore, the potential increase in PA due to breakfast consumption could be one mechanism of action

for the protective effect of breakfast against CHD development. However, as explained in the previous section, there is a need for long-term intervention trials to find more robust evidence about the effect of breakfast on obesity prevention to support this hypothesis (Brown, Bohan Brown et al. 2013).

One cross-sectional study (Deshmukh-Taskar, Nicklas et al. 2013) and five RCODs (Farshchi, Taylor et al. 2005, Astbury, Taylor et al. 2011, Betts, Richardson et al. 2014, Kobayashi, Ogata et al. 2014, Reeves, Huber et al. 2015) suggest that breakfast consumption may be associated with improved blood parameters such as lipid, glucose and/or insulin levels. Therefore, this would be associated with a reduction in the risk of chronic disease. Also, A large Canadian study found breakfast eaters to be associated with better nutrient adequacy than skippers. Nutrient adequacy is believed to play a significant role in the prevention of chronic disease (WHO 2017). These possible mechanisms of action against chronic disease development will be explored in the following section.

2.5.1 The Effect of Breakfast Consumption on Blood Parameters A number of interventional studies and one cross-sectional study have investigated the effect or association of consuming vs skipping breakfast on blood parameters. Their findings are limited due to the heterogeneous methods, small sample sizes and small interventions used.

Astbury et al. (Astbury, Taylor et al. 2011) provided one day BE and BS interventions. This study investigated blood parameters after a liquid meal was consumed in the morning following BE breakfast or BS interventions. This trial found that plasma free fatty acids (BE 0.2 ± 0.06 vs BS 0.5 ± 0.05 mmol/L) and plasma Glucagon-like peptide 1 (BE 15.8 ± 4.1vs BS 11.0 ± 1.8 pmol/L) were significantly higher in the breakfast omitting trial compared to the breakfast consumption intervention. However, no other plasma parameters including blood glucose (BE: 4.5 ± 0.2 vs BO 4.7 ± 0.1 mmol/L); serum insulin (BE 44.9 ± 7.1 vs BS 45.4 ± 7.6 pmol/L) and other blood parameters were found to be significantly different between the two groups.

Kobayashi et al. measured 24 hour blood glucose levels after BE vs BS interventions. This trial found that the BS intervention group had a significant increase in 24 hour blood glucose level compared to the BE intervention group (89 ± 2 mg/dl vs 83 ± 3, P < 0.05) (Kobayashi,

Ogata et al. 2014). This may suggest better daily glycaemic control in the breakfast consumption intervention group compared to the BS one (Kobayashi, Ogata et al. 2014).

Farshchi et al. (Farshchi, Taylor et al. 2005) measured blood values following breakfast consumption after 14 days of BE and 14 days of BS interventions. Peak postprandial serum insulin (BE 362 (166) pmol/L vs BS 400 (188) pmol/L insulin) and total cholesterol (BE 3.14 (0.41) mmol/L vs BS 3.40 (0.443) mmol/L) were significantly higher after 14 days of the BS trial compared to the 14 days BE trial (Farshchi, Taylor et al. 2005) suggesting that habitual BS causes detrimental effects on postprandial insulin and cholesterol levels, which may be linked over the long term to a higher chance of chronic diseases development.

Overall, there is limited evidence about the effect of BE or BS on blood parameters from the interventional studies. Considering the short term nature of the studies conducted, the conflicting findings and the heterogeneous study designs, longer term trials are needed to clarify research in this area.

One large cross sectional study was conducted in the USA called the National Health and Nutrition Examination Survey (NHANES) 1999–2006 study of 5316 young adults aged 20– 39 y (Deshmukh-Taskar, Nicklas et al. 2013). This study investigated breakfast habits, body size and cardio-metabolic parameters and found breakfast consumption to be significantly associated with lower body weight and signficantly improved cardiometabolic parameters including cholesterol (total cholesterol BS 192.7 ± 1.6 mg/dl; Ready to Eat Cereal (RTEC) 187.0 ± 1.8 mg/dl; other breakfasts 188.8 ± 0.9 mg/dl; p=0.017) and serum insulin levels (BS 12.1 ± 0.7 µU/ml; 10.3 ± 0.4 µU/ml; other breakfasts 11.0 ± 0.4 µU/ml; p=0.032). To the author’s knowledge, no other large databases are available in Australia to investigate breakfast habits and related them to metabolic parameters such as blood values. This suggests the need for a cross-sectional study to be conducted in an Australian population.

2.5.2 Breakfast Consumptions and Nutrients Status Among the other explanations for the protective effect of breakfast on chronic disease development is that it may improve nutritional profiles. Some observational studies have found that breakfast consumers, especially those consuming breakfast cereals, have better nutrients intakes across the day (Deshmukh-Taskar, Nicklas et al. 2010, Gibson and Gunn 2011). Barr et al. (2013) investigated BE and nutrient adequacy and found that breakfast

eaters, and especially breakfast cereal eaters, had improved nutrient adequacy than BS (Barr, DiFrancesco et al. 2013). Williams et al. (Williams 2014) used the National Health and Medical Research Council (NHMRC) guidelines to classify the level of evidence available about breakfast cereal intake and nutritional status (level A (highest) to level D (lowest))(Williams 2014) This SR reported that that there was some evidence (grade B) supporting regular breakfast cereal consumption to be associated with diets higher in vitamins and minerals and with lower fat intake. However, the evidence available was poor (grade C - based from 20 cross sectional studies) supporting the association of regular breakfast cereal with greater likelihood of meeting recommended nutrients intakes. Therefore, more research in this field is required.

2.6 Is Breakfast Cereal Consumption Protective against Chronic Disease and Obesity Risk? Breakfast cereal consumption has gained particular attention in this field as it is a commonly consumed type of breakfast in many countries such as the USA and Australia. In fact, two recent SRs (Williams 2014, Priebe and McMonagle 2016) summarised the evidence available from observational and interventional studies regarding associations between breakfast cereal intake and obesity, diabetes and CVD risks. Williams et al. (Williams 2014) used the NHMRC guidelines to classify the level of evidence available in this field (level A (highest) to level D (lowest) guidelines (NHMRC 2009, NHMRC 2011, Allman-Farinelli, Byron et al. 2014)) (Williams 2014).

The Williams SR reported a significant relationship between regular breakfast cereal consumption and a lower risk of being overweight or obese (Evidence grade B of the NHMRC guidelines) (Williams 2014). There was some evidence suggesting that high-fibre cereal and oat-based cereal had protective effects on weight gain, but there was insufficient evidence to fully support this relationship and further studies were considered needed (Williams 2014).

Two SRs (Williams 2014, Priebe and McMonagle 2016) reported evidence supporting an association between regular or frequent breakfast cereal consumption and BMI in children (p=0.020 in boys but not in girls (p=0.58) over 7.5 years (Albertson, Affenito et al. 2009), and p=0.001 in children over 3 years (Balvin Frantzen, Treviño et al.)); of BMI for age z scores

and risk of being overweight (p<0.01(Barton, Eldridge et al. 2005)); of adiposity in children (p=0.008 over 3-10 years (Albertson, Thompson et al. 2009)) and weight gain in men (p=0.007 over 13 years (Bazzano, Song et al. 2005)). The evidence from these studies has been gathered in men and children, thus suggesting the need to conduct more prospective studies in women.

Williams (Williams 2014) reported some evidence supporting an association of regular whole grain and high fibre breakfast cereal consumption with a reduction in diabetes risk (grade B of the NHMRC guidelines (NHMRC 2009, NHMRC 2011, Allman-Farinelli, Byron et al. 2014)). Similar findings were also supported by Priebe et al. (2016) who reported a significant association between high whole grain ready-to-eat cereal consumption and reduced risk of Type 2 diabetes. Specifically, the evidence came from two prospective studies: the Nurses’ Health study (Liu, Manson et al. 2000) and the Physicians Health study (Kochar, Djousse et al. 2007)). The Nurses Health study (Liu, Manson et al. 2000) found a significant association between whole grain breakfast cereal and a reduction in diabetes risk over 10 years among 75,521 women (fully adjusted models: ≤1 serving/week RR (95% CI) 0.81 (0.71, 0.93); 2 to 4 servings/week 0.70 (0.60, 0.81); 5-6 servings/week 0.71 (0.62, 0.82); ≤1 serving/day 0.66 (0.55, 0.80), p trend <0.0001) (Liu, Manson et al. 2000). The findings from the Physician Health Study (Kochar, Djousse et al. 2007) were that there was a significant association between breakfast cereal consumption and a reduction in the risk of diabetes over a mean of 19.1 years follow up among 21,152 US male physicians (fully adjusted models: HH (95% CI) ≤1 serving/week 0.83 (0.79 to 0.93); 2-6 servings/weeks 0.76 (0.67 to 0.86), ≥7 servings/week 0.69 (0.60 to 0.79); p for linear trend <0.0001). Whole grains and refined grains were both significantly associated with a reduction in diabetes risk, however, this relationship was stronger for whole grain consumption (fully adjusted models: RR (95% Confidence Interval) ≤1 serving/week 0.75 (0.64 to 0.88); 2-6 servings/weeks 0.76 (0.66 to 0.87); ≥7 servings/week 0.60 (0.50 to 0.71); p<0.001) in comparison to refined grains (fully adjusted models: ≤1 serving/week 0.88 (0.70 to 1.1); 2-6 servings/weeks 0.69 (0.53 to 0.90); ≥7 servings/week 0.95 (0.73 to 1.3); p for linear trend= 0.05).

Contrary to these two studies findings, Williams reported on a prospective study (Whitehall Study II) (McNaughton, Mishra et al. 2008) which analysed 7,339 participants aged 39–63 y for 12 years and found no significant relationship between medium or higher fibre breakfast

cereal consumption and diabetes risk. These conflicting results indicate the need for more studies to be conducted on this topic to further investigate the association of higher fibre cereal and whole grain on diabetes risk. Furthermore, the SR conducted by Williams found only weak evidence (grade D) to indicate that regular breakfast cereal consumption per se may reduce the risk of developing diabetes [9]. Therefore, again, more research is required in this field (Williams 2014, Priebe and McMonagle 2016).

Williams reported that there was a lower level of evidence (evidence grade C) supporting a relationship between high-fibre or whole grain breakfast cereal and a lower risk of CVDs (Williams 2014), and even more limited evidence (evidence grade D) to suggest an association between high-fibre and whole grain breakfast cereal and a reduction in the risk of hypertension, and a higher level of overall wellbeing (evidence of grade D) (Williams 2014). These SRs reported two papers investigating the effect of ready-to-eat cereal intake on CVD risk in the Physician’s Health Study. One paper reported any ready-to-eat cereal consumption to be associated with a reduction in the risk of developing hypertension (Kochar, Gaziano et al. 2012). The other paper found a significant reduction in incidence of heart failure associated with consumption of whole grain ready-to-eat cereal. This association was not significant for refined ready-to-eat cereal (Djousse and Gaziano 2007), again suggesting more research is needed in this field.

These two SRs concluded that there are still major gaps in the epidemiological literature with regards to breakfast cereal consumption in relation to obesity and chronic disease risk; thus, more prospective studies are required especially in other population groups, such as in women, to find more robust and clearer evidence in these topics (Williams 2014).

2.7 Gaps in the Literature

‘A large number of gaps in the literature were identified, and not all could be addressed within a single PhD thesis. Therefore, this thesis focussed on addressing the following gaps:

1. Identified GAP: There were still major gaps in the epidemiological literature with regards to an association between breakfast cereal consumption and obesity or chronic disease, especially amongst women (Williams 2014).

This PhD: Longitudinal examinations of data from the Australian Longitudinal Study on Women’s Health (ALSWH) were conducted to examine breakfast cereal consumption and incidence of obesity and diabetes. The two papers are presented in Chapters 3 and 4. 2. Identified GAP: The effect of breakfast consumption on DIT is unclear. This PhD: A SR was conducted to consolidate the evidence in this research area. The published paper is presented as Chapter 5. 3. Identified Gap: In authors’ knowledge, there are no studies of young Australian men that collected anthropometric measurements (BMI and circumferences), metabolic measurements (RMR, body composition values, blood parameters, blood pressure) and habitual breakfast habits data. Furthermore, the last available data collected on Australian men’s habitual breakfast habits was the NNS of 1995. This PhD: The ‘Typical Aussie bloke’ study aimed to explore the current breakfast habits of a sample of young Australian men and investigate any association between breakfast habits and metabolic or anthropometric parameters. These data are presented in chapters 6 and 7.

The overall aim of this PhD is to increase the evidence base with regards to the role of breakfast and breakfast cereal consumption on obesity and chronic disease, and to further understand any mechanisms of action.

2.8 Aims, Objectives and Hypothesis

2.8.1 First Aim To examine the role of breakfast and breakfast cereal consumption on the development of obesity and chronic disease risk.

2.8.1.2 Hypothesis Breakfast consumption will be associated with higher metabolic rates, better anthropometry, reduced metabolic risk factors, and lower incidence of chronic disease.

2.8.1.3 Objectives • To examine the effect and/or associations of consuming breakfast on measures of Resting Metabolic Rate (RMR) and DIT;

Studies addressing: Systematic Literature Review; Typical Aussie Bloke • To examine the associations of consuming breakfast and breakfast cereal on anthropometric parameters (height, weight, BMI, waist circumference and body composition) and obesity risk Studies addressing: ALSWH Obesity Study; Typical Aussie Bloke • To examine the associations of consuming breakfast on metabolic parameters (blood pressure, glucose, insulin and lipid profile). Studies addressing: Typical Aussie Bloke • To investigate the associations of consuming breakfast cereal on chronic disease outcomes. Studies addressing: ALSWH Obesity Study; ALSWH Diabetes Study

2.8.2 Second Aim: To describe typical Australian breakfast consumption habits of a sample of young Australian men and their determinants.

2.8.2.1 Hypothesis Breakfast consumption habits will be associated with socio-demographic parameters such as income and living arrangements.

2.8.2.2 Objectives • To describe the frequency of breakfast consumption among a sample of young Australian men and its association with socio-demographic variables, work and lifestyle habits. • Studies addressing: Typical Aussie Bloke • To describe the typical content of breakfast among a sample of young Australian men and its association with socio-demographic variables, work and lifestyle habits. • Studies addressing: Typical Aussie Bloke

2.9 Studies These aims and objectives were investigated by undertaking four separate studies:

1) A twelve year longitudinal investigation of breakfast cereal consumption and its relationship with the risk of developing obesity among the mid-age cohort of the Australian Longitudinal Study of Women’s Health (Chapter 3 – ALSWH Obesity Study) 2) A twelve year longitudinal investigation of breakfast cereal consumption and its relationship with the risk of developing diabetes among the mid-age cohort of the Australian Longitudinal Study of Women’s Health (Chapter 4 – ALSWH Diabetes Study); 3) A Systematic Review, meta-analysis and meta-regression exploring the effect of breakfast meals of varying energy contents and compositions consumed after an overnight fast on Diet Induced Thermogenesis (Chapter 5); 4) A multi-centre cross-sectional study investigating Habitual Breakfast consumption and its relationship with anthropometric and metabolic parameters, socio-demographic and lifestyle characteristics, such as work, sleep, physical activity and dietary intakes among young Australian men (Chapters 6 and 7).

2.10 Thesis Structure The thesis is composed of eight chapters as illustrated in Figure 2.1. Chapter 1 provides an introduction to the thesis. Chapter 2 summarises the background literature regarding breakfast consumption and its relationship with health outcomes. Chapters 3 (Quatela, Callister et al. 2017) and 4 (Quatela, Callister et al. 2018) describe the findings of the ALSWH longitudinal reporting the association of breakfast cereal consumption with obesity and diabetes risks amongst the mid-age Australian women. Chapter 5 (Quatela, Callister et al. 2016) describes the findings of the SR, meta-regression and meta-analysis investigating the effect of breakfast consumption on DIT. Chapter 6 (in the process to be submitted to European Journal of Nutrition) and chapter 7 describe the findings of the ‘Typical Aussie Bloke’ study investigating breakfast consumption habits in relation with anthropometric and metabolic parameters, physical activity and socio-demographic characteristics of young Australian men. Chapter 8 appraises the findings of this entire thesis.

CHAPTER 1: INTRODUCTION

CHAPTER 2: CRITICAL LITERATURE REVIEW

CHAPTER 4: CHAPTER 5: CHAPTER 7: CHAPTER 3: LONGITUDINAL SYSTEMATIC REVIEW, CHAPTER 6: CROSS META-ANALYSES AND CROSS SECTIONAL LONGITUDINAL STUDY SECTIONAL STUDY: STUDY META-REGRESSION STUDY: The protective effect of Breakfast Consumption ‘Typical Aussie Bloke’ Breakfast Cereal muesli consumption on The Energy Content and Part 2: Breakfast Consumption and Obesity diabetes risk: Results from Composition of Meals Habits of young Australian Consumption and Eating Risk amongst the Mid-Age 12 years of follow-up in the Consumed after an Overnight men from the ‘Typical Patterns in relation to Cohort of the Australian Australian Longitudinal Fast and Their Effects on Aussie Bloke’ study intermediate risk factors Longitudinal Study on Study on Women’s Health Diet Induced Thermogenesis: A Systematic Review Meta- for Obesity and Chronic Women’s Health Analyses and Meta- Disease Development. Regressions’.

CHAPTER 8: DISCUSSION AND RECOMMENDATIONS FOR FUTURE RESEARCH AND PRACTICE

Figure 2.1 Thesis structure diagram. This Figure illustrates the four studies conducted as part of this PhD and the eight chapters compromising this PhD thesis. 57

Chapter 3: Breakfast Cereal Consumption and Obesity Risk amongst the Mid-Age Cohort of the Australian Longitudinal Study on Women’s Health This paper was published in Health Care journal, the following is the citation.

Quatela, A., R. Callister, A.J. Patterson, M. McEvoy and L.K. MacDonald-Wicks (2017). "Breakfast Cereal Consumption and Obesity Risk amongst the Mid-Age Cohort of the Australian Longitudinal Study on Women’s Health." Healthcare 5(3): 49.

The work presented in the manuscript was presented at the International Society of Behavioural Nutrition and Physical Activity in June 2017, Victoria, Canada, (2017) (oral presentation by Patterson, A; Appendix 1) and at Nutrition Society of Australia (NSA), Melbourne (poster presentation by Quatela, A; Appendix 1)

The work presented in the manuscript was completed in collaboration with the co- authors (Appendix 2)

Quatela, A.1; Callister, R.2,3,5; Patterson, A.J.1,3,5, McEvoy M.4,5; MacDonald-Wicks, L.K.1,3,5*.

1 Discipline of Nutrition and Dietetics, School of Health Sciences, The University of Newcastle; Callaghan, NSW 2308, Australia

2; School of Biomedical Sciences and Pharmacy, The University of Newcastle, Callaghan, NSW 2308, Australia

3 Priority Research Centre for Physical Activity and Nutrition; The University of Newcastle, Callaghan, NSW 2308, Australia

4 Centre for Clinical Epidemiology & Biostatistics, School of Medicine and Public Health, The University of Newcastle, Callaghan, NSW 2308, Australia.

5 Hunter Medical Research Institute, New Lambton, NSW 2305, Australia

*Author to whom correspondence should be addressed.

Academic Editor: Sampath Parthasarathy

Received: 26 July 2017 / Revised: 16 August 2017 / Accepted: 25 August 2017 / Published: 30 August 2017

3.1 Overview This paper investigated the association of breakfast cereals, common types of breakfast foods, on the risk of developing obesity longitudinally over 12 years amongst a very large cohort of women of the Australian Longitudinal Study on women’s Health (ALSWH). These longitudinal analyses have provided good evidence of which types of common breakfast cereal are associated with obesity risk longitudinally.

3.2 Introduction In Australia in 2011–2012, 62.8% of the adult population was found to be either overweight (BMI 25.1–29.9 kg/m2) or obese (BMI ≥ 30 kg/m2) (ABS 2013) with the prevalence of obesity among women at 27.5% (WHO 2017). Worldwide, 39% of adults were overweight and 13% obese in 2014 (WHO 2017). Being overweight or obese is associated with unfavourable effects on blood cholesterol and triglycerides, insulin resistance, and blood pressure, which increase the risk of type 2 diabetes, cardiovascular diseases and ischemic stroke (WHO 2017). A higher BMI is also linked to a higher risk of some types of cancer (e.g., breast or colon cancer) (WHO 2017). A higher degree of overweight is associated with an increased risk of comorbidities related to noncommunicable diseases and higher mortality (WHO 2017). Obesity has an estimated cost of about 1–3% of total health expenditure in the majority of countries and the cost reaches 5–10% in the USA (Organisation for Economic Co-operation and Development (OECD) 2014)

Breakfast cereal is a grain based food product prepared from oats, corn, wheat or rice, and may undertake minimal processing, such as by drying and rolling the grain (e.g., rolled oats), or more substantial processing such as being cooked, and then flaked or puffed. Multiple grains may be mixed, and nuts and/or fruits added. Breakfast cereal is generally eaten with milk or yogurt but can be eaten in a dry state. Cereal is most frequently consumed at breakfast; however, it can also be eaten as a snack or at other meals during the day.

Breakfast cereal consumption has been proposed to be protective against development of obesity. Two systematic reviews (Williams 2014, Priebe and McMonagle 2016) found some evidence from prospective studies supporting the association of regular or frequent breakfast cereal consumption with BMI (Balvin Frantzen, Treviño et al. , Barton, Eldridge et al. 2005, Albertson, Affenito et al. 2009) or adiposity in children (Albertson, Thompson et al. 2009) and weight gain in men (Bazzano, Song et al. 2005). These systematic reviews examined studies in men and children, suggesting a need for more research to be conducted in women. Furthermore, Williams (Williams 2014) classified the evidence base available from observational and interventional studies conducted in this area using the National Health and Medical Research Council guidelines. Williams found there is some evidence (grade B of the National Health and Medical Research Council guidelines (NHMRC 2009, NHMRC 2011, Allman-Farinelli, Byron et al. 2014)) to support an association between regular breakfast cereal consumption and a lower risk of being overweight or obese. The mechanism of action responsible for this possible protective effect of breakfast cereal on weight status is unclear (Williams 2014) One mechanism, with some weak evidence (grade C), supports an association between high-fibre breakfast cereal intake and higher satiety and lower hunger levels, which may protect against weight gain (Williams 2014). Williams, however, concluded that there is not enough evidence to differentiate the effects of different types of breakfast cereal on body weight (Williams 2014).

It is clear that more studies are needed to investigate the effects of breakfast cereal, and different types of breakfast cereal, on the risk of developing obesity. The aim of this study was to investigate the effects of consumption of different categories of breakfast cereal on the risk of developing obesity over 12 years among participants of the mid-age cohort of the ALSWH. It is hypothesised that consumption of any breakfast cereal or consumption of high fibre breakfast cereal will be protective against the risk of developing obesity.

3.3 Materials and Methods The ALSWH collects data every 2–3 years from four age cohorts (14,247 women aged 18–23 in 1996; 13,714 women aged 45–50 in 1996; 12,432 women aged 70–75 in 1996 and 17,015 women aged 18–23 years in 2013) in Australia with a total of 58,000

participants (ALSWH 2017). More information regarding the ALSWH are provided elsewhere (Lee, Dobson et al. 2005).

The analyses reported in this paper were conducted in the mid-aged cohort only (women aged: 45–50 years at S1 in 1996; 47–52 years at S2 in 1998; 50–55 years at S3 in 2001; 53–58 years at S4 in 2004; 56–61 years at S5 in 2007; 59–64 years at S6 in 2010; 62–67 years at S7 in 2013). A food frequency questionnaire (FFQ) completed at S3 (2001) provided the dietary data (Giles and Ireland 1996 , Hodge, Patterson et al. 2000), that were used to categorise women who consumed breakfast cereal from a list of options. Obesity incidence was the outcome of interest because it is recognised to be associated with severe risk of comorbidities (WHO 2017). Obesity data were obtained at S4 (2004) to S7 (2013), 12 years after S3 (2001).

3.3.1 Participants This analysis was conducted in a representative sample of Australian women born between 1946–1951 who formed the mid-aged cohort of the ALSWH in 1996. These participants were randomly sampled from Medicare (the database of the Australian National Health Insurance Scheme), which includes all Australian citizens and permanent residents. The participants from remote and rural areas were sampled at twice the rate compared to participants in urban areas (Lee, Dobson et al. 2005). The ALSWH was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Human Research Ethics Committee of the University of Newcastle (approval numbers: H-076- 0795 and H-2012-0256) and University of Queensland (approval numbers: 2004000224 and 2012000950) (Lee, Dobson et al. 2005, alswh 2017). Written informed consent was obtained from all subjects (Lee, Dobson et al. 2005, alswh 2017). Permission to access the ALSWH data for the purpose of this investigation was provided on 19 January 2015.

3.3.2 Predictor Variables Breakfast cereal intakes were the predictor variables and were collected at S3 in 2001 using a validated FFQ: the Dietary Questionnaire for Epidemiology Studies Version 2 (DQES-FFQv2) developed by the Cancer Council Victoria (Giles and Ireland 1996 , Hodge, Patterson et al. 2000, ALSWH 2017). The Australian NUTTAB 95 database was used to analyse the dietary data (Lewis, Milligan et al. 1995). The DQES-FFQv2

was validated by Hodge, Patterson et al. (Hodge, Patterson et al. 2000) in 63 women of childbearing age who completed 7-day weighed food diaries and the DQES-FFQv2; the comparison between these two dietary methods found the DQES-FFQv2 to be a valid tool to collect dietary data in adult women (Hodge, Patterson et al. 2000).

The breakfast cereal question asked: “Over the last 12 months, on average, how often did you eat the following foods?” The following breakfast cereal options were listed: (1) All-Bran; (2) Sultana Bran, Fibre Plus, Branflakes; (3) Weet Bix, Vita Brits, Weeties; (4) Cornflakes, Nutrigrain, Special K; (5) muesli; (6) porridge. The frequency of consumption options allowed for categorisation into a dichotomous variable where yes referred to any consumption higher than “never” and no referred to no consumption (“never”).

In addition, the following variables were created. A higher fibre (or whole grain) breakfast cereal category was a dichotomous variable assigned yes when at least one of these five breakfast cereal categories was consumed: porridge; muesli; All-Bran; Weet Bix, Vita Brits and Weeties; Sultana Bran, Fibre Plus and Branflakes. The oat-based cereal variable was dichotomised as “yes” when porridge or muesli or both were consumed and “no” when neither of these breakfast cereals were consumed. The wheat cereal category was “yes” if any or multiple of Weet Bix, Vita Brits or Weeties; All- Bran; Sultana Bran, Fibre Plus, or Branflakes were consumed and “no” if none of these cereals were consumed. In addition, an “any breakfast cereal” variable was derived and assigned “yes” when at least one of the six breakfast cereal categories was consumed and “no” when no breakfast cereal was consumed.

3.3.3 Outcome Variable The outcome variable was obesity incidence based on data from S4–S7 (2004 to 2013) up to 12 years after S3 (2001) in women who had a BMI < 25 kg/m2 at S3. Women who were overweight (BMI ≥ 25–29.9 kg/m2) or obese (BMI ≥ 30 kg/m2) at S3 were excluded from the analyses. BMI was calculated from self-reported height and weight and provided as a continuous variable. The obesity variable was created by categorising BMI as: <30 kg·m−2 (non-obese); BMI ≥ 30 kg·m−2 (obese). For the mid-age cohort, self-reported height and weight have been validated by Burton et al. (Burton, Brown et al. 2010) in 159 mid-aged women; this validation study found substantial agreement

between the measured and self-reported height and weight and calculated BMI (Confidence Interval (CI) of the difference between the reported and the measured BMI: 0.12 (−0.13, 0.37)) (Burton, Brown et al. 2010).

3.3.4 Identification and Measurement of Confounding Factors Area of residence, managing on income, marital status, education level, smoking, physical activity, sedentary behaviour, diabetes, impaired glucose tolerance, heart disease, hypertension, and dietary intakes (daily energy intake and fibre intake) were considered as potential confounding factors. Data for potential confounding variables were obtained from S3 with the exception of education level, which was collected at S1.

Area of residence was a dichotomous variable categorised as either urban or non-urban (remote or rural) residence. Income was evaluated as the self-reported capacity to manage on income rather than the actual monetary income level. Participant responses were assigned to one of the following two categories: income easy (it is not too bad or it is easy) or income difficult (it is impossible, it is difficult all the time, or it is difficult some of the time). Smoking status was a categorical variable composed of never- smokers, ex-smokers, or smokers.

Physical activity status was evaluated using items from Active Australia’s 1999 National Physical Activity Survey (Armstrong, Bauman et al. 2000). Participants were asked to report the time spent in different categories of physical activity per week. The physical activity variable was expressed in Metabolic Equivalent Task (MET) minutes per week. A MET represents a unit of resting metabolic rate and is normally considered to be 3.5 mL oxygen/kg/min (Ainsworth, Haskell et al. 1993). Time in activities of different intensities were multiplied by intensity-specific coefficients for each level of physical activity as follow: 3.0 * X minutes of walking, 4.0 * X minutes of moderate intensity activities, and 7.5 * X minutes of vigorous activities; the three categories were then summed to calculate total MET minutes per week (Ainsworth, Haskell et al. 1993, ALSWH 2017). Physical activity levels were then assigned to the following four categories: Nil/sedentary for <40 MET min/week; low for 40 to <600 MET min/week; moderate for 600 to <1200 MET min/week; high for ≥1200 MET min/week (ALSWH 2017).

The presence or absence of hypertension, diabetes, impaired glucose tolerance or heart disease were derived from responses to this question ‘Have you ever been told by a doctor that you have’ followed by a list of medical conditions; responses were dichotomised (yes or no). The dietary factors (daily energy intake (kJ/day), and fibre (g/day) intake) were derived from the DQES-FFQv2 using the NUTTAB 95 database (Lewis, Milligan et al. 1995) as explained above.

3.3.5 Statistical Analyses The participant characteristics of breakfast cereal consumers or non-consumers were compared using a two sample independent t test (for parametric distributions), a Wilcoxon Rank Sum Test (for non-parametric distributions) or a two sample t test for proportions.

The relationship between breakfast cereal consumption (coded as yes or no consumption) at S3 and the risk of developing obesity between S4–S7 was investigated using multiple logistic regression with discrete time survival analysis models.

Confounding factors were identified as follows. A variable was a confounding factor if the p value for univariate regression analyses was ≤0.2 for the relationship between the potential confounding factor and both the predictor (breakfast cereals) and outcome (obesity incidence) variables. Variables that met these criteria were included in multivariate analyses. A variable “other breakfast cereal consumption” (coded as yes or no) was created and adjusted for in the models in order to adjust for the intake of breakfast cereals in addition to the cereal category being analysed.

Unadjusted and adjusted (for confounding factors) logistic regressions with discrete time survival analysis models were used to investigate the associations between different breakfast cereal consumption variables and subsequent incidence of obesity. Four statistical models were developed. The first (unadjusted, univariate) had only the breakfast cereal category of interest and a discrete measure of time (survey waves). The second model added non-dietary confounding factors: area of residency, income, smoking status, physical activity (METs) and hypertension. The third model adjusted for dietary confounding factors: daily energy intake, fibre, and consumption of other breakfast cereal. The fourth fully adjusted model included all confounding factors (dietary and non-dietary). Separate analyses were undertaken to investigate relationships 64

for each of the six categories of breakfast cereals listed in the FFQ, as well as for the oat-based breakfast cereals, for wheat-based breakfast cereals, and for higher fibre breakfast cereals. The HosmerLemeshow goodness-of-fit test was used to determine how well the logistic regression models fit the data. All analyses were completed using STATA version 13.

3.4 Results A total of 11,226 women completed S3, of which 10,629 completed the DQES-FFQv2 at S3. Five thousand six hundred and thirty-two (5632, 50.2%) women were excluded because they reported being overweight or obese at S3; 854 (7.6%) women were excluded because their daily energy intake at S3 was either <4500 or >20,000 kJ/day (Meltzer, Brantsæter et al. 2008, Dodd, Cramp et al. 2014). Four thousand one hundred and forty-three (4143, 36.9%) women were included in the analyses.

3.4.1 Participant Characteristics Of the 4,143 women included in the analyses, 3756 (90.7%) were breakfast cereal consumers and 387 (9.3%) were non-consumers. Of the 6486 women excluded, 5800 (89.4%) were breakfast cereal consumers and 686 (10.6%) were non-consumers. The participant characteristics of the women included in the analysis at S3 (2001) are described in Table 3.1 for all participants, breakfast cereal consumers, breakfast cereal non-consumers and in Table 3.2 for those consuming the individual types of breakfast cereals (e.g., muesli consumers). Of the total population in the analyses, approximately 60% were living in non-urban residency areas, 65% reported managing on income as either easy or not too bad, the majority of the population (62%) were never smokers, and 84% of these women engaged in some level of physical activity. Only 12% of the population reported having hypertension.

“Any” Cereal All Participants “Any” Cereal “No” Cereal vs. “No” Cereal p Value Sample size n = 4143 90.7% (n = 3756) 9.3% (n = 387) Mean Age (years) 52.4 ± 1.5 52.4 ± 1.5 52.4 ± 1.4 0.9322 Area of Residency Urban 38.9% 39.0% 38.2% 0.7699 1 Non-urban 60.4% 60.4% 60.5% 0.9913 Managing Income 2 Income difficult 34% 33.4% 40.1% 0.0084 3 Income easy 65.0% 65.6% 59.4% 0.0158 Smoking Status Never smokers 62.0% 63.2% 51.4% 0.0000 Ex-smokers 22.7% 22.4% 25.6% 0.1572 4 Current smoker 14.8% 14.1% 22.2% 0.0000 Physical Activity Sedentary 14.8% 14.1% 21.7% 0.0001 Low PA 31.2% 31.3% 20.2% 0.6555 Moderate PA 21.9% 22.3% 17.8% 0.0412 High PA 30.9% 31.2% 28.4% 0.2691 Hypertension no 88.2% 88.2% 88.4% 0.9351 yes 10.7% 10.7% 11.4% 0.6749 Diet * Energy intake from diet (kJ/day) 6623 ± 2465 6667 ± 2474 6052 ± 2188 0.0000 * Energy from alcohol (kJ/day) 150.4 ± 501.2 145.6± 475.1 202.3 ± (749.1) 0.0234 * Fibre (g/day) 20.0 ± 9.2 20.4 ± 9.3 16.5 ± 6.4 0.0000 This table summarises the data from 4143 women included in the analyses. * Data are presented as median ± interquartile range. The rest of the data are presented as mean ± Standard Deviation (SD) or % of participants. Abbreviations: PA = Physical Activity. 1 Non-urban: remote or rural. 2 Income difficult: managing income is impossible, difficult all or some of the time. 3 Income easy: managing income is not too bad or is easy. 4 Current smoker: smoking an indeterminate amount; smoking less than 10 cigarettes; smoking 10–19 cigarettes per day; smoking more than 20 cigarettes per day.

Table 3.2 Characteristics of participants from the mid-age (2001) cohort of the Australian Longitudinal Study on Women’s Health at Survey 3 (n = 4143) by individual breakfast cereal consumption.

Sultana Bran, Weet Bix, Vita Brits and Cornflakes, Nutrigrain and Muesli Porridge All-Bran Fibre Plus and Weeties Special K Branflakes 39.0% 50.6% 24.6% 30.9% 52.2% 41.3% Sample size (n =1614) (n = 2095) (n = 1019) (n = 1279) (n = 2161) (n =1709 ) Mean Age (years) 52.4 ± 1.4 52.4 ± 1.5 52.4 ± 1.4 52.4 ± 1.5 52.4 ± 1.5 52.4 ± 1.5 Area of Residency Urban 40.5% 38.3% 40.2% 38.9% 38.1% 40.4% 1 Non-urban 58.8% 61.1% 59.3% 60.8% 61.1% 59.1% Managing Income 2 Income difficult 28.7% 34.1% 33.3% 32.8% 35.2% 35.3% 3 Income easy 70.4% 65.1% 65.8% 66.1% 64.0% 63.6% Smoking Status Never smokers 66.8% 65.3% 66.3% 66.2% 63.1% 63.6% Ex-smokers 23.2% 28.3% 21.5% 20.4% 22.4% 20.2% 4 Current smoker 9.6% 12.5% 11.9% 12.8% 14.2% 15.9% Physical Activity Sedentary 8.7% 13.2% 10.7% 12.2% 13.9% 15.2% Low PA 32.3% 31.5% 30.4% 35.1% 32.2% 32.9% Moderate PA 23.4% 22.0% 23.0% 22.3% 23.0% 21.1% High PA 34.5% 32.5% 34.7% 29.1% 29.9% 29.4% Hypertension no 90.6% 88.2% 88.7% 87.9% 88.9% 88.6% yes 8.3% 10.5% 10.0% 11.2% 10.2% 10.2% Diet * Energy intake from diet 6757 ± 2404 6812 ± 2,481 6784 ± 2,498 6795 ± 2597 6,826 ± 2510 6,891 ± 2676 (kJ/day) * Energy from alcohol 202.6 ± 508.0 129.5 ± 401.7 181.3 ± 496.4 168.9 ± 484.0 138.8 ± 432.4 132.9 ± 458.1 (kJ/day) * Fibre (g/day) 21.6 ± 9.0 20.8 ± 9.3 23.3 ± 11.1 21.4 ± 9.7 20.6 ± 9.5 19.5 ± 9.4 67

This table summarises the data from 4143 women included in the analyses. * Data are presented as median ± interquartile range. The rest of the data are presented as mean ± SD or % of participants. Abbreviations: PA = Physical Activity. 1 Non-urban: remote or rural. 2 Income difficult: managing income is impossible, difficult all or some of the time. 3 Income easy: managing income is not too bad or is easy. 4 Current smoker: smoking an indeterminate amount; smoking less than 10 cigarettes; smoking 10–19 cigarettes per day; smoking more than 20 cigarettes per day.

The median (inter quartile range) energy intake was 6623 (2465) kJ/day obtained from diet and 150 (501) kJ/day from alcohol. Median fibre intake was 20.0 (9.2) g/day. The energy contribution from alcohol was compared with the Acceptable Macronutrient Distribution Ranges for Australians and New Zealanders. Median alcohol intake (2.2% of energy ingested) met the recommendation of <5% (NHMRC 2005). Median fibre intake (20.0 g/day), however, was lower than the Recommended Dietary Intake for Australian women of 28 g/day (NHMRC 2005).

A number of characteristics were significantly different between breakfast cereal consumers and non-consumers (Table 3.1). Women who did not consume breakfast cereal were significantly more likely to be smokers, experience difficulties in managing on their income, and be sedentary and less likely to engage in moderate levels of physical activity. Furthermore, those who did not eat breakfast cereal consumed less daily energy from their diet but more daily energy from alcohol and they consumed less fibre than cereal consumers.

3.3.3 Breakfast Cereal Consumption and Risk of Obesity A total of 308 (7.43%) incident cases of obesity were reported over 12 years of follow- up. Table 3.3 presents the results of the logistic regression with discrete time survival analysis models for all breakfast cereal categories both unadjusted and adjusted for confounding factors. Area of residence, income, smoking, physical activity, hypertension and dietary intakes (daily energy intake, daily fibre intake and other breakfast cereal consumption) were included in the multivariate analysis.

Table 3.3 Results of the logistic regression models examining the effects of consuming breakfast cereal at Survey 3 on the risk of developing obesity from Surveys 4–7.

Breakfast Cereal Mid 3 (Yes and No) Model 1 * Model 2 * Model 3 * Model 4 * Odds Ratio (CI) p value Odds Ratio (CI) p value Odds Ratio (CI) p value Odds Ratio (CI) p value Any breakfast cereal 0.76 (0.53, 1.09) 0.13 0.87 (0.59, 1.27) 0.46 0.87 (0.60, 1.25) 0.45 0.92 (0.63, 1.35) 0.68 Muesli 0.46 (0.35, 0.60) 0.00 0.55 (0.42, 0.73) 0.00 0.49 (0.31, 0.62) 0.00 0.57 (0.43, 0.75) 0.00 Porridge 0.77 (0.62, 0.97) 0.03 0.81 (0.64, 1.03) 0.08 0.80 (0.64, 1.01) 0.07 0.81 (0.64, 1.03) 0.09 All-Bran 0.56 (0.41, 0.76) 0.00 0.59 (0.43, 0.82) 0.00 0.65 (0.47, 0.89) 0.01 0.62 (0.44, 0.87) 0.01 Sultana Bran, Fibre Plus and Branflakes 0.86 (0.67, 1.10) 0.24 0.91( 0.70, 1.19) 0.50 0.93 (0.72,1.20) 0.57 0.94 (0 .72, 1.23) 0.68 Weet Bix, Vita Brits and Weeties 0.99 (0.79, 1.24) 0.91 1.04 (0.82, 1.32) 0.76 1.04 (0.82, 1.30) 0.77 1.07 (0.84, 1.36) 0.58 Cornflakes, Nutrigrain and Special K 1.35 (1.08 ,1.70) 0.01 1.27 (1.00, 1.61) 0.05 1.29 (1.03, 1.63) 0.03 1.26 (0.99, 1.60) 0.07 Oat-based breakfast cereal 0.62 (0.49, 0.78) 0.00 0.70 (0.55, 0.89) 0.00 0.66 (0.52, 0.83) 0.00 0.71 (0.55, 0.90) 0.01 Wheat-based breakfast cereal 0.89 (0.70, 1.13) 0.34 0.96 (0.75, 1.24) 0.77 0.99 (0.77, 1.27) 0.94 1.01 (0.78, 1.31) 0.92 Higher fibre (or whole grain) breakfast cereal 0.64 (0.48, 0.86) 0.00 0.77 (0.56, 1.06) 0.11 0.71 (0.52, 0.97) 0.03 0.79 (0.57, 1.10) 0.16 * Model 1 univariate model with predictor variable, outcome and a discrete measure of time (wave). Model 2 with predictor variable, outcome, discrite measure of time and adjusted for non dietary counfoundiing factors (smoking, managing income, area of residency, physical activity and hypertension). Model 3 with predictor variable, outcome, discrete measure of time and adjusted for dietary counfounding factors (daily energy intake, fibre and other breakfast cereals consumption). Model 4 with predictor variable, outcome, discrete measure of time and adjusted for dietary and non dietary counfouding factors.

The results of the fully adjusted models (adjusted for dietary and non-dietary confounding factors) are as follows. The consumption of any breakfast cereal was not associated with a significant reduction in the risk of developing obesity (OR: 0.92; CI: 0.63, 1.35; p = 0.68) whereas oat-based breakfast cereal was associated with a significant reduction in obesity risk (OR: 0.71; CI: 0.55, 0.90; p = 0.01). Wheat-based cereals (OR: 1.01; CI: 0.78, 1.31; p = 0.92) and higher fibre breakfast cereals (OR: 0.79; CI: 0.57, 1.10; p = 0.16) were not associated with a significant reduction in the risk of developing obesity. Among the six individual breakfast cereal categories, only muesli (OR: 0.57; CI: 0.43, 0.75; p = 0.00) and All-Bran (OR: 0.62; CI: 0.44, 0.87; p = 0.01) were significantly associated with a reduced risk of developing obesity. The other four breakfast cereal categories were not associated with a significant reduction in the risk of developing obesity.

3.5 Discussion This longitudinal study investigated the effect of breakfast cereal consumption on the risk of developing obesity (BMI ≥ 30 kg/m2) among a large cohort of mid-aged Australian women. The consumption of any breakfast cereal, regardless of type, or higher fibre (whole grain) breakfast cereal was not protective against obesity. However, muesli on its own or as a part of the oat-based cereal group and All-Bran were significantly protective against obesity development. These findings suggest that the type of breakfast cereal consumed matters with regards to its association with obesity risk, and that few breakfast cereals may be significantly associated with a reduction in obesity risk.

This analysis expands the evidence base in an area where data are limited. Two recent systematic reviews (Williams 2014, Priebe and McMonagle 2016) summarised the evidence available from cross sectional studies, prospective studies and Randomised Controlled Trials (RCTs) regarding the association between breakfast cereal intake and weight status. With regards to prospective studies, these two reviews found some evidence supporting the association of regular or frequent breakfast cereal consumption with BMI in children (p = 0.020 in boys but not in girls (p = 0.58) over 7.5 years [9], and p = 0.001 in children over 3 years (Balvin Frantzen, Treviño et al.)); of BMI for age z scores and risk of being overweight (p < 0.01(Barton, Eldridge et al. 2005)); of

adiposity in children (p = 0.008 over 3–10 years (Albertson, Thompson et al. 2009)) and weight gain in men (p = 0.007 over 13 years (Bazzano, Song et al. 2005)). Also, these two reviews identified the need for more prospective studies to be conducted in this area in other population groups. This paper has contributed to the understanding of the effect of breakfast cereal intake on obesity risk in mid-aged women in Australia. Furthermore, Williams (Williams 2014) found limited evidence available regarding the effect of different types of breakfast cereal on body weight. One of the strengths of our analyses was the ability to investigate the effects on obesity risk of different individual breakfast cereals and cereal categories, including categories based on fibre, oats or wheat content, thereby providing new data.

Our results found muesli to be significantly protective against obesity in mid-aged Australian women. The addition of porridge to form the oat-based cereal category weakened the protective effect and when porridge was analysed individually, it was not associated with a significant reduction in the risk of obesity. This suggests that something other than the oats in muesli, for example dried fruits, seeds or nuts, may exert a protective effect against obesity, or that there is something specific about muesli consumers that we have not been able to identify in our analysis. Another possibility is that the consumption of porridge in Australia is very seasonally based, commonly only consumed during the winter season. Therefore, the consumption of porridge may not be sufficiently consistent to provide protection against obesity. Another possibility is that porridge consumption may have not been captured to its full potential due to the season when the dietary data were collected.

A previous study conducted in an overweight and obese population reported that the consumption of 37.5 g of oat cereal twice a day for 12 weeks, replacing usual food intake for two eating events, (Chang, Huang et al. 2013) was significantly beneficial for improving weight status and BMI compared to another cereal type (weight change over 12 weeks: control + 0.52 ± 1.74 kg, oat cereal: −2.08 ± 2.05 kg) (Chang, Huang et al. 2013). However, this study provided evidence in an overweight and obese population only and the intake of 70 g/day may not be a realistic amount to be consumed regularly under normal circumstances. Our analyses of oat-based cereal (porridge and/or muesli) found that any intake of muesli on its own or as a part of an oat-based cereal category

was significantly associated with a protective effect against obesity over 12 years in a normal weight population.

All-Bran was also found to be significantly protective against obesity incidence in this cohort of women. The reason for this protective effect is unclear. More studies are needed to further investigate a mechanism of action and determine whether this protective effect is achieved in other populations. Although the high fibre content of All-Bran might be assumed to be the mechanism, this was controlled for in our analyses. Furthermore, higher fibre breakfast cereal was not significantly associated with a reduction in obesity risk in our cohort.

Our study has a number of strengths, the first of which is the large representative population sample analysed. Another strength is the longitudinal analysis of prospectively collected data, which lowers the risk of potential biases. Furthermore, the long period of follow-up (12 years) is an important distinction in comparison to other studies investigating breakfast cereal consumption. The robust statistical approach and the use of a validated DQES-FFQv2 are other major strengths of this study.

There are also a number of limitations to the analyses that need to be recognised. Principally this includes the use of self-reported data. However, there is evidence from the two validation studies, one validating the DQES-FFQv2 (Hodge, Patterson et al. 2000, Lee, Dobson et al. 2005) and the other the BMI (Burton, Brown et al. 2010), to suggest that self-reported data are adequately accurate. The analyses investigated whether there was a significant association between cereal intake collected at baseline and obesity incidence over a 12 year period. Therefore, our analyses could not account for any variations in dietary habits, such as cereal intake, over this period of time. The DQES-FFQv2 does not include information on how the breakfast cereal was prepared or when it was consumed. Similarly the categories of breakfast cereal in the DQES- FFQv2 do not capture rice-based breakfast cereal (such as “rice bubbles”) or high sugar varieties (such as ‘fruit loops’); therefore, it is not possible to evaluate whether these types of breakfast cereal would impact obesity risk.

It is important to be aware that the DQES-FFQv2 has a limited list of breakfast cereal choices, thus the bluntness of this FFQ could have impacted on our ability to evaluate breakfast cereal consumption. The list of options covers most of the common breakfast 73

cereal types from the viewpoint of nutrition professionals; however it is possible that if the completers of this FFQ did not find their brand name breakfast cereal listed in the FFQ, they might have ticked none. The brand names of breakfast cereal options listed by the FFQ are meant to be examples of the types of breakfast cereal that would fit in these options (e.g., Sultana Bran, FibrePlus and Branflakes are examples of wheat-based high fibre cereals of which there are a number of other brand names in Australia). However, there are no explicit instructions on how to complete the breakfast cereal questions in the DQES-FFQv2, including that these examples of breakfast cereal types should be used to enable completers to categorise the breakfast cereal they consume.

3.6 Conclusions Among mid-aged Australian women, muesli on its own, or as part of oat-based breakfast cereals, and All-Bran, but no other breakfast cereals, were associated with a reduction in obesity risk. This effect may be due to particular characteristics of these cereal eaters, but these relationships warrant further investigation

3.7 Acknowledgments The research on which this paper is based was conducted as part of the Australian Longitudinal Study on Women’s Health, the University of Newcastle and the University of Queensland. We are grateful to the Australian Government Department of Health for funding and to the women who provided the survey data. The authors thank Professor Graham Giles of the Cancer Epidemiology Centre of Cancer Council Victoria, for permission to use the Dietary Questionnaire for Epidemiological Studies (Version 2), Melbourne: Cancer Council Victoria, 1996. This work was supported by the University of Newcastle Research Training Program (RTP) scholarship.

3.8 Author Contributions All authors made substantial contributions to: the conception and design of the study and interpretation of data. Angelica Quatela and Mark McEvoy were primarily responsible for the statistical analysis. Angelica Quatela drafted the paper with support from Amanda Patterson, Lesley MacDonald-Wicks and Robin Callister. All authors were involved in critically revising the paper for intellectual content, for editing and provided approval of the final manuscript.

3.9 Conflicts of Interest The authors declare no conflict of interest

Chapter 4: The protective effect of muesli consumption on diabetes risk: Results from 12 years of follow-up in the Australian Longitudinal Study on Women’s Health This paper was published in the ‘Nutrition Research’ journal

Quatela, A., R. Callister, A.J. Patterson, M. McEvoy and L.K. MacDonald-Wicks (2018). The protective effect of muesli consumption on diabetes risk: Results from 12 years of follow-up in the Australian Longitudinal Study on Women’s Health. Nutrition Research 51: 12.

The work presented in the manuscript was presented at the Nutrition Society of Australia (NSA) conference, Melbourne (poster presentation by Quatela, A; Appendix 3) and at the Australian Longitudinal Study on Women’s Health (ALSWH) conference, Newcastle (oral presentation by Quatela, A; Appendix 3).

The work presented in the manuscript was completed in collaboration with the co-authors (Appendix 4).

Quatela, A.1;; Callister, R.2,3,5; Patterson, A.J.1,3,5, McEvoy M.4,5; MacDonald-Wicks, L.K.1,3,5*.

1 Discipline of Nutrition and Dietetics, School of Health Sciences, The University of Newcastle; Callaghan, NSW 2308, Australia

2; School of Biomedical Sciences and Pharmacy, The University of Newcastle, Callaghan, NSW 2308, Australia

3 Priority Research Centre for Physical Activity and Nutrition; The University of Newcastle, Callaghan, NSW 2308, Australia

4 Centre for Clinical Epidemiology & Biostatistics, School of Medicine and Public Health, The University of Newcastle, Callaghan, NSW 2308, Australia.

5 Hunter Medical Research Institute, New Lambton, NSW 2305, Australia

*Author to whom correspondence should be addressed.

Received 7 April 2017, Revised 14 December 2017, Accepted 18 December 2017, Available online 24 December 2017.

4.1 Overview This paper investigated the effect of breakfast cereals, common types of breakfast foods, on the risk of developing diabetes longitudinally over 12 years amongst a very large cohort of women. These longitudinal analyses provided good evidence of which types of common breakfast cereal are associated with diabetes risk longitudinally.

4.2 Abstract Diabetes affects 9.8% of Australian women. Breakfast cereal consumption is potentially protective against diabetes. This study investigated the effects of breakfast cereal consumption on the 12-year risk of developing diabetes among mid-aged participants of the Australian Longitudinal Study of Women’s Health (ALSWH). It was hypothesized that any breakfast cereal and higher-fiber breakfast cereals would be protective against the risk of developing diabetes. Data from Survey 3 (S3) to Survey 7 (S7) inclusive, from the 1946-51 ALSWH cohort were analyzed. Dietary data were obtained at S3 and the outcome was incident diabetes between S4-S7. Women were excluded if: they reported existing diabetes or impaired glucose tolerance at S3; dietary data were incomplete; or daily energy intake was <4,500 or >20,000kJ. Logistic regression with discrete time survival analyses investigated the association between breakfast cereal intake and incident diabetes. Models were adjusted for income, BMI, smoking, physical activity, education, and dietary intakes and included a measure of time. There were 637 incident cases of diabetes. Breakfast cereal intake per se was not associated with incident diabetes (OR: 1.00; p=0.98). Muesli consumption on its own (OR: 0.74; p=0.00) or as a part of oats-based cereal (OR: 0.84; p=0.047) was significantly associated with a decrease in the odds of developing diabetes. No other breakfast cereals were significantly associated with diabetes risk. Among mid-aged Australian women, muesli consumption was associated with a reduction in diabetes risk. This effect may be due to a particular profile of muesli eaters, but the relationship warrants further investigation.

Keywords: Breakfast; Edible Grain; Diabetes Mellitus; risk; Longitudinal Studies.

4.3 Introduction Diabetes mellitus is a major public health concern which affected 9.8% of Australian women in 2012 (AusDiab 2012). Worldwide, 8.5% of adults had diabetes in 2014 (WHO 2016). Diabetes has a significant impact on quality of life as it is the main cause of renal failure, lower limb amputation and blindness, and a major contributor to cardiovascular disease (WHO 2016). One and a half million deaths were due to diabetes in 2012 worldwide (WHO

2016). The World Health Organization (WHO) (WHO 2016) estimated that the cost of diabetes is US$827 million per year worldwide (Seuring, Archangelidi et al. 2015, Collaboration 2016).

Breakfast cereal can be defined as a grain-based food product usually made from oats, rice, wheat or corn, which may be minimally processed, such as drying and rolling the grain (eg. rolled oats), or cooked and flaked or puffed (Quatela, Callister et al. 2017). Several grain varieties may be combined, and fruit and/or nuts added. It is often consumed with milk or yogurt, or in a dry state. Breakfast cereal is often eaten at breakfast, but it can also be consumed as a snack or at other meals during the day.

It has been hypothesized that whole grain breakfast cereals might reduce the risk of diabetes because of their high fiber content and high nutrient density (phytochemicals, vitamins and minerals). The fiber content of wholegrain cereal products is hypothesized to improve the glycemic response to breakfast, and through this mitigate the development of Type 2 diabetes (Meyer, Kushi et al. 2000, Aune, Norat et al. 2013). Concomitantly, refined grains may increase the risk of developing diabetes due to low fiber content and subsequent high glycemic index (GI) or glycemic load (Aune, Norat et al. 2013).

The findings of a recent study conducted by Pastorino et al (Pastorino, Richards et al. 2016) found a significant increase in the risk of developing diabetes for 43 year old women who consumed a diet higher in fat, higher in GI and lower in fiber (p<0.01 adjusted for confounding factors). This suggests a protective effect for a low fat, low GI and high fiber diet in the development of diabetes for women of this age; however, similar analyses were not significant for men. The Pastorino et al study (Pastorino, Richards et al. 2016) investigated these characteristics for the diet generally, but the results suggest that the effect of varying glycemic load, fiber and fat content of breakfast cereals warrants investigation in the development of diabetes.

A study by Xu et al (Xu, Huang et al. 2016) reported that in the NIH-AARP Diet and Health Study a highly significant reduction in mortality from diabetes was found for breakfast cereal consumers compared with non-consumers in 367,442 subjects in the U.S.A. (p<0.05, quartile four (highest) OR 0.70, CI: 0.47,1.03) (Xu, Huang et al. 2016). These promising findings in a U.S. population support the need to investigate these relationships further in other populations.

In 2014 a systematic review (Williams 2014) conducted by Williams concluded there was limited evidence (grade B of the Australian National Health and Medical Research Council guidelines (NHMRC) (NHMRC 2009, NHMRC 2011, Allman-Farinelli, Byron et al. 2014)) supporting the protective effect of regular whole grain and high fiber breakfast cereal consumption regarding the development of diabetes (Williams 2014). These findings were supported by a recent systematic review that reported a significant association between high whole grain ready to eat cereal consumption and reduced risk of Type 2 diabetes risk (Priebe and McMonagle 2016). Furthermore, the systematic review conducted by Williams found only weak evidence (grade D) to indicate that regular breakfast cereal consumption per se may reduce the risk of developing diabetes (Williams 2014).

It is clear that further research to investigate the effect of breakfast cereal consumption on the risk of developing diabetes is warranted. It was hypothesized that consumption of any breakfast cereal and consumption of higher-fiber breakfast cereals would be protective against the risk of developing diabetes. This hypothesis was investigated by undertaking a longitudinal analysis assessing the effect of consuming any breakfast cereal, higher-fiber breakfast cereal and different types of breakfast cereal on the risk of developing diabetes in a large representative cohort of mid-aged Australian women over a 12-year period.

4.4 Methods and Materials The ALSWH is a longitudinal study of women in Australia (n=58,000) collected from four age cohorts (14,247 women aged 18-23 y, 13,714 women aged 45-50 y and 12,432 women aged 70-75 y in 1996, and 17,015 women aged 18-23 y in 2013). Data were prospectively gathered from 1996 to 2013. More details regarding the ALSWH are described elsewhere (Lee, Dobson et al. 2005).

For the purpose of these analyses, data were obtained from the mid-aged (women aged 45-50 y in S1 in 1996) cohort. Surveys were conducted every 2-3 years since 1996. Dietary data obtained from a food frequency questionnaire (Giles and Ireland 1996 , Hodge, Patterson et al. 2000) at S3 were used to identify participants who consumed breakfast cereal from a list of options. Diabetes incidence data were obtained at S4 (2004) to S7 (2013), up to 12 years after S3 (2001).

4.4.1 Participants Data from a representative sample of Australian women born between 1946-51 who formed the mid-aged cohort of the ALSWH (ALSWH 2017) were used. These women were

randomly sampled from the database of the national health insurance scheme, Medicare, which included all Australian permanent residents and citizens. The women from rural and remote areas were sampled at double the rate of women in urban areas (Lee, Dobson et al. 2005). Ethics approval for the ALSWH was provided from the Human Research Ethics Committees of the University of Newcastle and the University of Queensland (Lee, Dobson et al. 2005). Permission for access to these data for the purpose of these analyses was provided on the 19th of January 2015.

4.4.2 Predictor Variables The predictor variable was breakfast cereal consumption reported at S3. Dietary data were obtained from a validated food frequency questionnaire: the Dietary Questionnaire for Epidemiological Studies Version 2 (DQES-FFQv2). The DQES-FFQv2 was developed by the Cancer Council Victoria (Giles and Ireland 1996 , Hodge, Patterson et al. 2000, ALSWH 2017). The dietary data were analyzed using the Australian NUTTAB 95 database (Lewis, Milligan et al. 1995). The DQES-FFQv2 has been validated amongst childbearing age women who completed the DQES-FFQv2 and a 7-day weighed food diary; the comparison between these two dietary methods confirmed that the DQES-FFQv2 is a valid tool to assess dietary intake in adult women (Hodge, Patterson et al. 2000).

The DQES-FFQv2 asked the following question: ‘Over the last 12 months, on average, how often did you eat the following foods?’ The following breakfast cereal options were listed: 1) All-Bran; 2) Sultana Bran, Fiber Plus, Branflakes; 3) Weet Bix, Vita Brits, Weeties; 4) Cornflakes, Nutrigrain, Special K; 5) muesli; 6) porridge. Sultana bran, Fiber Plus and Branflakes are predominantly flaked wheat products that are high in insoluble fiber and may contain dried fruit. All-bran is a predominantly wheat cereal with a very high total fiber content. Weet Bix, Vita Brits and Weeties are wheat-based products in the form of biscuits. Cornflakes, Nutrigrain and Special K are breakfast cereals low in fibre. Muesli is a rolled- oats based cereal, which may or may not be toasted. It usually contains dried fruit and/or nuts and seeds. Porridge is a cooked cereal from rolled oats. It may be whole rolled oats or it may consist of oats that are finely chopped and fast cooked. The frequency of consumption options allowed categorization into a dichotomous variable where ‘yes’ referred to any consumption higher than ‘never’ and ‘no’ referred to no consumption (‘never’).

The ‘any breakfast cereal’ variable was a dichotomous variable assigned ‘yes’ when at least one of the six breakfast cereal categories was consumed and ‘no’ when no breakfast cereal

from these categories was consumed (all six breakfast cereal variables were equal to ‘never’)(Quatela, Callister et al. 2017). The oats-based cereal category was a dichotomous variable assigned ‘yes’ when muesli or porridge or both were consumed and ‘no’ when neither of these breakfast cereals were consumed (both these breakfast cereals were equal to ‘never’). The wheat cereal category was ‘yes’ if any or a multiple of Sultana Bran, Fiber Plus, Branflakes, All-Bran, Weet Bix, Vitabrits and Weeties were consumed, and ‘no’ if none of these cereals were consumed (i.e., all these breakfast cereals were equal to ‘never’). The ‘higher fiber’ (or whole grain) breakfast cereal was a dichotomous variable assigned ‘yes’ when at least one of the five breakfast cereal groups (muesli, porridge, All-Bran, Sultana Bran group, Weet Bix group) was consumed.

4.4.3 Outcome Variable The outcome variable was diabetes incidence from S4-S7. Women with pre-existing diabetes (S1-3) or impaired glucose tolerance (IGT) (S3) were excluded from the analyses. These outcomes were determined from the following questions: ‘Have you ever been told by a doctor that you have’, which was followed by a list of diseases including diabetes in S1 or insulin dependent diabetes and non-insulin dependent diabetes in S2. In the subsequent surveys (S3-S7) the question was: ‘In the past three years have you been diagnosed or treated for’ where the list of options included IGT, insulin dependent diabetes and non-insulin dependent diabetes (high blood sugar) in S3; or diabetes (high blood sugar) in S4 to S7. Diabetes and IGT variables were dichotomous variables (yes/no development of diabetes and/or IGT). Lowe et al (Lowe, Byles et al. 2010) compared 388 mid-aged women suffering from diabetes in one or more of these surveys (S1-4) with the Medicare (MBS) and Pharmaceutical Benefits Scheme (PBS) databases for the years 2002-2005 (Lowe, Byles et al. 2010). This study demonstrated self-reported diabetes to be a reliable assessment of diabetes incidence in this cohort (Lowe, Byles et al. 2010).

4.4.4 Identification and Measurement of Confounding Factors Marital status, income, body mass index (BMI), smoking, physical activity, education, area of residence, sedentary behavior, dietary intakes (fiber and daily energy) were considered as potential confounders. The potential confounding variables were obtained from S3 apart from education level which was collected at S1.

Income was considered in terms of the ability to manage on current income rather than consideration of the actual monetary income level. The ability to manage on current income

was categorized as: easy (it is not too bad or it is easy) or difficult (it is impossible; it is difficult all the time; it is difficult some of the time). BMI was calculated based on self- reported height and weight. Burton et al (Burton, Brown et al. 2010) validated the self- reported height, weight and BMI in 159 women from the mid-aged cohort and found substantial agreement between the measured and self-reported height, weight and BMI (Burton, Brown et al. 2010). BMI was provided as a continuous variable, however for the purpose of this analysis, this variable was categorized as: BMI <30 and BMI ≥30 kgm-2. Smoking status was categorized as: never-smokers, ex-smokers, or smokers.

Physical activity was determined using items from Active Australia’s 1999 National Physical Activity Survey (Armstrong, Bauman et al. 2000). Physical activity level was estimated in Metabolic Equivalent Task (MET) minutes. MET refers to a unit of resting metabolic rate

(RMR) and is typically considered to be 3.5mL O2/kg/minute (Ainsworth, Haskell et al. 1993). MET minutes were calculated using different coefficients for each intensity of physical activity as follows: 3.0* X minutes of walking, 4.0* X minutes of moderate intensity activities, and 7.5 * X minutes of vigorous intensity activities; activity was then summed to provide total MET minutes per week (Ainsworth, Haskell et al. 1993, ALSWH 2017). Physical activity was categorized as: Nil/sedentary for <40, low for 40 to <600, moderate for 600 to <1200, and high for ≥ 1200 MET minutes per week (ALSWH 2017).

Education was determined at Survey 1 and was categorized as the following three options: no formal qualifications; Intermediate Certificate (or equivalent) or Higher School or Leaving Certificate (or equivalent) or Trade/apprenticeship (eg. Hairdresser, Chef) or Certificate/diploma (eg. Child Care, Technician); University degree or University Higher degree (eg. Grad Dip, Masters, PhD).

The dietary factors, daily energy intake (kJ/d) and daily fiber intake (g/d) intake), were derived from the DQES-FFQv2 using the NUTTAB 95 database (Lewis, Milligan et al. 1995) as described above.

4.4.5 Statistical Analyses The characteristics of the women who did or did not report consuming breakfast cereal were compared using two sample t tests for proportions, two sample independent t tests (for parametric distributions), or Wilcoxon Rank Sum Tests (for non-parametric distributions).

Multiple logistic regression models were used to investigate the association between breakfast cereal intake category (coded as yes/no consumption) at S3 and the risk of developing diabetes between S4 to S7. Associations between breakfast cereal consumption and subsequent incidence of diabetes were determined using unadjusted and adjusted (for confounding factors) logistic regression with discrete time survival analysis models. Women were excluded if: DQES-FFQv2 at S3 was not completed; they reported existing diabetes at S1-3 or impaired glucose tolerance at S3; or if daily energy intake at S3 was ≤ 4,500 or ≥ 20,000 kJ/d (Meltzer, Brantsæter et al. 2008, Dodd, Cramp et al. 2014).

The following method was used to identify potentially confounding factors for the longitudinal analyses. A variable was considered a potential confounder when the p value of the regression analysis for the potentially confounding variable with both the predictor and outcome variables was ≤0.2 (Quatela, Callister et al. 2017). Variables that met these criteria were included in multivariate analyses. In order to account for the consumption of breakfast cereals in addition to the category being investigated, the ‘other breakfast cereal consumption’ variable (coded as ‘yes’ or ‘no’) was created and adjusted for in the analysis.

Four statistical models were produced. The first (unadjusted, univariate) had only the breakfast cereal variable of interest and a discrete measure of time (survey wave). The second model included non-dietary confounding factors (income, BMI, smoking status, physical activity and education). The third model included the dietary confounding factors (daily energy intake, fiber intake, and consumption of other breakfast cereal). The fourth fully adjusted model included all confounding factors (dietary and non-dietary). Separate analyses were undertaken to examine associations with consumption of any breakfast cereals; for each of the six categories of cereals listed in the survey; for any of the wheat-based cereals; for any of the oats-based cereals and for any of the higher fiber (whole grain) breakfast cereals. The Hosmer–Lemeshow goodness-of-fit test was used to determine how well the logistic regression models fit the data. All analyses were completed using STATA version 13.

4.5 Results A total of 11,226 women completed S3 of whom 10,629 completed the DQES (5.3% did not complete the DQES-FFQv2); 536 (4.8%) women were excluded because they reported existing diabetes at S1/S2/S3 and 60 (0.5%) women were excluded because they reported existing IGT at S3 (baseline); 1611(14.4%) women were excluded because their daily energy

intake at S3 was either ≤ 4500 kJ or ≥ 20,000 kJ/day. A total of 8422 (75%) women were included in the analyses (Figure 4.1).

4.5.1 Participant Characteristics The characteristics are presented for all participants, and for those who consumed any breakfast cereal, no cereal, and each individual breakfast cereal category (Tables 4.1 and 4.2). For the overall population, age was 52.5 (1.5) y (median (inter quartile range)). Most women (82.0%) were either married or in a defacto relationship. A large proportion of the study population (61.8%) found that managing on their income was either easy or not too bad. Forty two percent were healthy weight, 30.4% were overweight and 21.5% were obese. The majority of the participants (61.7%) were never smokers. Fifty percent of the population was sedentary or engaging in low levels of activity. The majority of the study population (68.8%) had a school qualification (Intermediate Certificate (or equivalent) or Higher School or Leaving Certificate (or equivalent) or Trade/apprenticeship (eg. Hairdresser, Chef) or Certificate/diploma (eg. Child Care, Technician)). The median and interquartile range for dietary factors were: energy intake from diet 6699 (2482) kJ/d, energy intake from alcohol 121.3 (435.2) kJ/d, and fiber intake 20.0 (9.2) g/d. When compared to the Nutrient Reference Values for Australians and New Zealanders, median alcohol intake provided 1.7% of energy ingested, which met the Australian recommendation of <5% (NHMRC 2005). Median fiber intake (20.0 g/d) was lower than the Australian recommendation of 28 g/d (NHMRC 2005).

Breakfast cereal consumers differed significantly from those who did not eat any breakfast cereal on a number of characteristics (Table 4.1). Those who did not eat cereal were more likely to be smokers, sedentary or less engaged in moderate levels of physical activity, and had lower education qualifications. Furthermore, those who did not eat breakfast cereal consumed lower daily energy from diet but more energy from alcohol and had lower fiber intakes than cereal consumers.

4.5.2 Breakfast Cereal Consumption and the Risk of Developing Diabetes During 12 years of follow-up, 637 (7.6%) incident cases of diabetes mellitus were reported. Table 4.3 presents the logistic regression models with survival analyses for all breakfast cereal categories both unadjusted and adjusted for confounding factors. Consumption of ‘any’ breakfast cereal (one or more cereal categories) was not significantly associated with the risk of developing diabetes (OR: 1.00; p=0.98). Muesli was the only individual breakfast cereal found to be significantly associated with a reduction in the risk of diabetes (OR: 0.74,

p=0.00). None of the other individual breakfast cereal categories were associated with a reduction in the risk of diabetes (Table 4.3). Oats-based cereal (either porridge or muesli or both) consumption was significantly associated with a reduction in the risk of developing diabetes (OR: 0.84; p=0.047). Wheat-based cereal (one or more of Sultana Bran, Fiber Plus, Branflakes; All-Bran; Weet Bix, Vita Brits, Weeties) consumption was not significantly associated with risk of developing diabetes (OR: 1.16; p=0.14). Higher fiber breakfast cereal (one or more of five breakfast cereal categories excluding Cornflakes, Nutrigrain and Special K) was not significantly associated with risk of developing diabetes (OR:0.82; p=0.12).

4.6 Discussion This study investigated the role of consuming breakfast cereal on the risk of developing diabetes in a large cohort of mid-aged Australian women. The majority of the breakfast cereal categories had no protective effect on developing diabetes over 12 years. Only muesli, consumed either by itself or as part of the oats-based cereal category, was protective against the development of diabetes. Also, consumption of higher fiber (whole grain) cereals did not provide protection from diabetes in these women. Therefore, our analysis did not support the hypothesized protective effect of consuming any breakfast cereal or higher-fiber cereals on diabetes risk.

Williams (Williams 2014) conducted a systematic review that investigated a number of research questions pertinent to this study. Although Williams (Williams 2014) found only limited evidence supporting the beneficial effect of regular breakfast cereal consumption on diabetes, our study found no significant relationship between breakfast cereal intake per se and risk. Xu et al (Xu, Huang et al. 2016) reported a highly significant reduction in mortality from diabetes for breakfast cereal consumers compared with non-consumers in a large cohort in the US. The odds ratio (0.70) of benefit was similar to the one obtained with muesli (0.75) in our study investigating diabetes incidence. In our cohort, although there was not a significant relationship between ‘any’ breakfast cereal intake and diabetes incidence, breakfast cereal eaters had significantly healthier lifestyle characteristics compared with non- consumers, specifically regarding smoking status, physical activity and dietary intake, which are all factors believed to be protective against the risk of developing diabetes.

The protective effect of oats-based breakfast cereal (muesli and porridge) in our study did not extend to porridge when analyzed individually. It is possible that this association of oats- based cereal with diabetes is being driven by muesli consumption, and that something other

than the oats in muesli is protective against diabetes, or that there is something particular about muesli eaters that we have not been able to adjust for. Also, porridge consumption in Australia is very seasonally based and is mostly consumed in the winter months. Therefore, porridge intake may not be sufficiently consistent to exert a protective effect against diabetes. Another option is that porridge intake may have not been assessed to its full potential due to the season when the dietary data were gathered.

Williams (Williams 2014) did not report on cohort studies that investigated muesli or oats- based cereal and diabetes, but did review short-term interventions in diabetic or normal populations investigating the effect of oats-based cereal, muesli or other types of breakfast cereals on glucose and insulin metabolism. In the majority of the studies conducted in diabetic populations, consumption of oats-based cereals or barley resulted in better glycemic control (Colagiuri, Miller et al. 1986, Golay, Koellreutter et al. 1992, Tappy, Gügolz et al. 1996, Tsihlias, Gibbs et al. 2000, Rendell, Vanderhoof et al. 2005). This suggests that the better glycemic control may be a possible mechanism of action for the protective effect of muesli or oats-based breakfast cereals on diabetes. However, this potential mechanism of action does not help to explain the differences we have observed in this study between muesli and porridge consumption.

The trials summarized in the systematic review by Williams (Williams 2014) were predominantly short-term studies and in populations diagnosed with diabetes. Long-term intervention studies and longitudinal analyses in prospective cohorts (like the analysis in this paper and the one conducted by Xu et al (Xu, Huang et al. 2016)) are needed in other populations to further investigate the effects of muesli, porridge and other breakfast cereals on diabetes risk to develop more robust evidence.

Two systematic reviews (Williams 2014, Priebe and McMonagle 2016) reported evidence supporting an association between regular or frequent whole grain or higher fiber breakfast cereal consumption and lower risk of diabetes from two prospective studies: the Physicians Health study [29] and the Nurses Health study (Liu, Manson et al. 2000). Specifically, Kocher et al (Kochar, Djousse et al. 2007) found that breakfast cereal consumption was associated with a significantly decreased risk of diabetes over a mean of 19.1 years follow up among 21,152 US males physicians (fully adjusted models: ≤1 serving/week 0.83 (0.79 to 0.93); 2-6 servings/week 0.76 (0.67 to 0.86), ≥7 servings/week 0.69 (0.60 to 0.79); p for linear trend <0.0001). This association was stronger for whole grain consumption (fully

adjusted models: ≤1 serving/week 0.75 (0.64 to 0.88); 2-6 servings/week 0.76 (0.66 to 0.87); ≥7 servings/week 0.60 (0.50 to 0.71); p<0.001) compared to refined grains (fully adjusted models: ≤1 serving/week 0.88 (0.70 to 1.1); 2-6 servings/week 0.69 (0.53 to 0.90); ≥7 servings/week 0.95 (0.73 to 1.3); p for linear trend= 0.05). The findings from the analyses of the Nurses Health study (Liu, Manson et al. 2000) also reported a significant protective effect of whole grain breakfast cereal on diabetes risk over 10 years among 75,521 women (fully adjusted models: ≤1 serving/week 0.81 (0.71, 0.93); 2 to 4 servings/week 0.70 (0.60, 0.81); 5- 6 servings/week 0.71 (0.62, 0.82); ≤1 serving/day 0.66 (0.55, 0.80), p trend <0.0001) (Liu, Manson et al. 2000). However, Williams’ review reported on another prospective study (Whitehall Study II), which found conflicting results (McNaughton, Mishra et al. 2008). This study analyzed 7,339 participants aged 39–63 y for 12 years and found no significant association between medium or higher fiber breakfast cereal consumption and diabetes risk. These conflicting results suggest the need for more studies to be conducted on this topic to further explore the effects of whole grain and higher fiber cereal on diabetes risk. One possible explanation for the difference between these studies is cultural differences in the foods eaten, that are an alternative to cereal consumption and their association with diabetes risk.

In our study, ‘any higher-fiber cereal intake’ was significant in unadjusted analyses or models adjusted for other dietary factors but not significant when adjusted for other lifestyle or demographic factors, such as smoking status, physical activity level, education and income. This suggests that the consumption of whole grain or higher-fiber cereals is associated with other positive lifestyle and demographic characteristics aligned with good health

Our study has a number of strengths, the first of which is the large representative sample of women used in the analysis. Furthermore, the longitudinal analysis of prospectively collected data, reducing potential bias, is a major strength of the study. The long period of follow-up (12 years) is a strength of this analysis in relation to other studies looking at breakfast cereal consumption. Additional strengths include the robust statistical approach used and the utilization of a validated FFQ.

In terms of limitations, our study relies completely on self-reported data. However, two validation studies, one validating the DQES-FFQv2 (Hodge, Patterson et al. 2000, Lee, Dobson et al. 2005) and the other diabetes incidence (Lowe, Byles et al. 2010), suggest that the reported data are adequately accurate. The DQESv2 FFQ does not allow completers to

specify when the breakfast cereal was consumed or how it was prepared. Also, rice-based breakfast cereals are not captured by the DQESv2, so we cannot comment on the role of these on diabetes risk. Whether diabetes incidence was Type 1 or Type 2 diabetes was not able to be determined, as the majority of the surveys did not provide this detail, however, considering the age group investigated, most will have developed Type 2 diabetes. This paper followed a rigorous methodology to establish the confounding factors to adjust for in this analysis. Though, it is possible that some unknown factors not collected in the surveys may have acted as confounders and they could have not been adjusted for. We acknowledge that the DQESv2 FFQ includes a limited list of breakfast cereal options, and the bluntness of this tool is likely to have influenced our ability to examine some aspects of breakfast cereal consumption. While the breakfast cereal categories do cover most common cereal types from the perspective of nutrition professionals, it is possible that if the women from the ALSWH cohort did not see their particular breakfast cereal listed in the FFQ, then they might have chosen none. The listed brand names within the FFQ are designed to be examples of the types of cereals that would fit in these categories (eg. Sultana Bran, Fiber Plus, Branflakes are examples of high fiber wheat-based cereals of which there are many other brand names in Australia). However, there is no specific direction on how to complete the breakfast cereal questions in the DQESv2 FFQ or that these examples should be used to enable people to categorize their cereal choices.

Finally, consumption of muesli by itself or as part of an oats-based cereal category in the eating pattern of Australian women was found to be protective against the development of diabetes. This effect may be due to a particular profile of muesli eaters, but these relationships warrant further investigation.

4.7 Conflict of Interest None

4.8 Acknowledgements The research on which this paper is based was conducted as part of the Australian Longitudinal Study on Women’s Health, the University of Newcastle and the University of Queensland. We are grateful to the Australian Government Department of Health for funding and to the women who provided the survey data.

The authors thank Professor Graham Giles of the Cancer Epidemiology Centre of Cancer Council Victoria, for permission to use the Dietary Questionnaire for Epidemiological Studies (Version 2), Melbourne: Cancer Council Victoria, 1996.

This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Human Research Ethics Committee of the University of Newcastle and University of Queensland. Written informed consent was obtained from all subjects.

4.9 Financial Support This work was supported by the Australian Government Research Training Program (RTP) Scholarship

4.10 Author contributions All authors (AQ; RC; AP; MM; LMW) have made substantial contributions to all of the following: conception and design of the study, analysis and interpretation of data, drafting the paper, critically revising the paper for important intellectual content and final approval of the version to be submitted.

11,226 completed S3 597 EXCLUDED as DQES not completed

536 EXCLUDED as reported existing diabetes at S1/S2/S3

60 EXCLUDED as 10,629 completed DQES reported existing IGT at S3

1,611 EXCLUDED as their daily energy intake at S3 was ≤ 4500 kJ or ≥ 20,000 kJ/day

8,422 were INCLUDED in the analyses

Figure 4.1 Flow chart of participant selection

All participants ‘Any’ ‘No’ No cereal vs cereal cereal cereal P value Sample size (n) 8422 91.5% (7702) 8.6% (720)

*Age (y) 52.5 (1.5) 52.5 (1.5) 52.5 (1.4) 0.7097 Managing income 1Difficult 37.2% 37.0% 40.4% 0.0659 2Easy 61.8% 62.0% 59.2% 0.1280 Education No qualification 15.0% 14.7% 17.8% 0.0288 3School 68.8% 68.8% 69.4% 0.7105 4University 15.5% 15.8% 12.1% 0.0088 Weight status 5Healthy weight 41.7% 41.5% 44.0% 0.1945 5Overweight 30.4% 30.6% 28.3% 0.2082 5Obese 21.5% 21.7% 19.7% 0.2150 Smoking status Never smokers 61.7% 62.7% 50.7% 0.0000 Ex-smokers 24.0% 23.8% 26.3% 0.1475 6Current smoker 13.9% 13.1% 22.6% 0.0000 Physical Activity Sedentary 17.2% 17.0% 22.1% 0.0005 Low PA 33.0% 33.1% 32.1% 0.5760 Moderate PA 20.9% 21.2% 16.9% 0.0068 High PA 27.6% 27.6% 27.4% 0.8719 Diet *Energy from diet (kJ/d) 6699 (2482) 6737 (2507) 6249 (2178) 0.0000 *Energy from alcohol (kJ/d) 121 (435) 120 (410) 153 (650) 0.0157 *Fiber (g/d) 20.0 (9.2) 20.4 (9.3) 16.7 (6.6) 0.0000 This table summarizes the data from the 8422 women included in the analyses.*Data are presented as median (interquartile range). The rest of the data are presented as mean (SD) or % of participants. 1Income difficult: Managing income is impossible, difficult all or some of the time 2Income easy: Managing income is not too bad or is easy 3School: Intermediate Certificate (or equivalent) or Higher School or Leaving Certificate (or equivalent) or Trade/apprenticeship (eg. Hairdresser, Chef) or Certificate/diploma (eg. Child Care, Technician) 4 University: University degree or University Higher degree (eg. Grad Dip, Masters, PhD) 5Healthy weight, overweight and obese classifications based on BMI. 6Current smoker: was defined as (smoker, an indeterminate amount; smoker, less than 10 per day; smoker, 10- 19 per day and smoker, 20 or more per day)

Muesli Porridge All-Bran Sultana Bran, Weet Bix, Vita Brits, Cornflakes, Fiber Plus, Branflakes Weeties Nutrigrain & Special K Sample size (n) 36.2% (3,051) 51.1% (4,306) 22.7% (1,910) 31.9% (2,685) 53.2% (4,478) 44.0% (3,707) *Age (y) 52.5 (1.4) 52.5 (1.5) 52.5 (1.4) 52.4 (1.5) 52.5 (1.5) 52.5 (1.5) Managing income 1Difficult 32.6% 37.9% 35.4% 37.1% 38.2% 38.9 2Easy 66.6% 861.4% 63.6% 62.0% 60.9% 60.1% Education No qualification 9.1% 14.3% 13.3% 14.5% 15.0% 15.8%

3School 66.9% 68.9% 69.4% 68.6% 70.2% 70.2% 4University 23.3% 16.0% 16.3% 15.8% 14.1% 13.3%

Weight status 5Healthy weight 47.2% 41.5% 46.0% 40.6% 40.9% 38.3% 5Overweight 29.6% 31.0% 29.8% 31.5% 30.5% 31.3% 5Obese 18.3% 21.5% 18.0% 21.9% 22.4% 23.3% Smoking status Never smokers 65.8% 64.3% 65.6% 65.1% 63.7% 63.2%

Ex-smokers 24.3% 23.8% 23.9% 22.2% 23.3% 22.7%

6Current smoker 9.6% 11.6% 10.2% 12.3% 12.8% 13.8% Physical Activity Sedentary 12.0% 16.1% 13.0% 14.5% 16.2% 17.6% Low PA 34.1% 33.4% 33.2% 35.9% 34.1% 34.7% Moderate PA 22.8% 21.3% 22.8% 21.6% 21.8% 20.9%

High PA 30.2% 28.3% 30.1% 27.2% 26.9% 25.8%

Muesli Porridge All-Bran Sultana Bran, Weet Bix, Vita Brits, Cornflakes, Fiber Plus, Branflakes Weeties Nutrigrain & Special K Diet *Energy from diet (kJ/d) 6782 (2450) 6859 (2592) 6813 (2528) 6818 (2601) 6857 (2545) 6904 (2644) *Energy from alcohol (kJ/d) 178 (483) 113 (387) 153 (462) 138 (437) 113 (400) 108 (398) *Fiber (g/d) 21.6 (9.0) 20.9 (9.4) 23.4 (11.3) 21.4 (9.7) 20.6 (9.4) 19.7 (9.2) This table summarizes the data from the 8422 women included in the analyses.*Data are presented as median (interquartile range). The rest of the data are presented as mean (SD) or % of participants. 1Income difficult: Managing income is impossible, difficult all or some of the time 2Income easy: Managing income is not too bad or is easy 3School: Intermediate Certificate (or equivalent) or Higher School or Leaving Certificate (or equivalent) or Trade/apprenticeship (eg. Hairdresser, Chef) or Certificate/diploma (eg. Child Care, Technician) 4 University: University degree or University Higher degree (eg. Grad Dip, Masters, PhD) 5Healthy weight, overweight and obese classifications based on BMI. 6Current smoker: was defined as (smoker, an indeterminate amount; smoker, less than 10 per day; smoker, 10-19 per day and smoker, 20 or more per day)

Table 4.3 Logistic regression models with descrete time survival analyses of the effect of consuming breakfast cereal at S3 on the risk of developing diabetes at S4-7 amongst 8422 mid-age women.

Breakfast cereal Model 1* Model 2* Model 3* Model 4*

Odd ratio (CI) P value Odd ratio (CI) P value Odd ratio (CI) P value Odd ratio (CI) P value

Any cereal 0.93 (0.70, 1.23) 0.60 1.01 (0.74, 1.38) 0.95 0.96 (0.73 , 1.28) 0.80 1.00 (0.73, 1.38) 0.98

Muesli 0.61 (0.51, 0.73) 0.00 0.74 (0.61, 0.90) 0.00 0.62 (0.52, 0.74) 0.00 0.74 (0.61, 0.90) 0.00

Porridge 0.89 (0.76, 1.04) 0.15 0.95 (0.80,1.13) 0.55 0.88 (0.75, 1.04) 0.13 0.93 (0.78, 1.11) 0.42

All-Bran 0.93 (0.77, 1.13) 0.46 1.01 (0.82, 1.25) 0.91 0.99 (0.81, 1.21) 0.93 1.01 (0.82, 1.26) 0.90 Sultana Bran/ Fiber Plus/ Branflakes 0.99 (0.84, 1.17) 0.91 1.00 (0.83, 1.20) 0.98 1.01 (0.86, 1.20) 0.87 0.99 (0.83, 1.20) 0.96 Weet Bix/ Vita Brits/ Weeties 1.13 (0.96, 1.32) 0.14 1.13 (0.95, 1.34) 0.16 1.14 (0.97, 1.34) 0.11 1.13 (0.95, 1.35) 0.16 Cornflakes/ Nutrigrain/ Special K 1.25 (1.07, 1.46) 0.01 1.16 (0.98 , 1.38) 0.08 1.23 (1.05, 1.44) 0.01 1.17 (0.99, 1.39) 0.07 Oats-based cereal 0.74 (0.63, 0.87) 0.00 0.86 (0.72; 1.02) 0.09 0.73 (0.62, 0.86) 0.00 0.84 (0.70, 1.00) 0.05**

Wheat-based cereal 1.12 (0.94, 1.33) 0.22 1.15 (0.95, 1.39) 0.17 1.18 (0.98, 1.41) 0.08 1.16 (0.95,1.41) 0.14 Higher fiber (or whole grain) cereal 0.76 (0.61, 0.95) 0.02 0.85 (0.67,1.08) 0.19 0.77 (0.61, 0.97) 0.02 0.82 (0.64,1.05) 0.12 *Model 1 univariate model with predictor variable, outcome and a discrete measure of time (wave) Model 2 with predictor variable, outcome, a discrete measure of time and adjusted for non dietary counfounding factors (income, education, BMI, smoking and physical activity ) Model 3 with predictor variable, outcome, a discrete measure of time and adjusted for dietary counfounding factors (daily energy intake, fiber, and other breakfast cereals consumption) Model 4 with predictor variable, outcome, a discrete measure of time and adjusted for dietary and non dietary counfouding factors. **p=0.047

Chapter 5: The Energy Content and Composition of Meals Consumed after an Overnight Fast and Their Effects on Diet Induced Thermogenesis: A Systematic Review, Meta-Analyses and Meta- Regressions This paper was published in Nutrients journal, the following is the citation.

Quatela, A., R. Callister, A. Patterson and L. MacDonald-Wicks (2016). "The Energy Content and Composition of Meals Consumed after an Overnight Fast and Their Effects on Diet Induced Thermogenesis: A Systematic Review, Meta-Analyses and Meta- Regressions." Nutrients 8(11): 670.

The work presented in the manuscript was presented at the International Society of Behavioural Nutrition and Physical Activity in June 2015, Edinburgh, Scotland (Two poster presentations by Quatela, A; Appendix 5).

The work presented in the manuscript was completed in collaboration with the co- authors (Appendix 6).

Quatela, A.1; Callister, R.2,3,4; Patterson, A.J.1,3,4, MacDonald-Wicks, L.K.1,3,4*.

1 Discipline of Nutrition and Dietetics, School of Health Sciences, The University of Newcastle; Callaghan, NSW 2308, Australia

2; School of Biomedical Sciences and Pharmacy, The University of Newcastle, Callaghan, NSW 2308, Australia

3 Priority Research Centre for Physical Activity and Nutrition; The University of Newcastle, Callaghan, NSW 2308, Australia

4 Hunter Medical Research Institute, New Lambton, NSW 2305, Australia

* Correspondence: [email protected] (L.M.-W); Tel.: +61-2-4921-6646

Received: 22 July 2016 / Revised: 30 September 2016 / Accepted: 18 October 2016 / Published: 25 October 2016

5.1 Overview This chapter investigates the evidence available on one hypothesised mechanism of action thought to be responsible for the protective effect of breakfast on obesity, the effect of breakfast consumption on stimulating the metabolism in the morning in the form of DIT. This chapter summarises the effect of breakfast meals varying on macronutrient composition, energy content and eating events on DIT.

5.2 Abstract

This systematic review investigated the effects of differing energy intakes, macronutrient compositions, and eating patterns of meals consumed after an overnight fast on Diet Induced Thermogenesis (DIT). The initial search identified 2482 records; 26 papers remained once duplicates were removed and inclusion criteria were applied. Studies (n = 27) in the analyses were randomized crossover designs comparing the effects of two or more eating events on DIT. Higher energy intake increased DIT; in a mixed model meta-regression, for every 100 kJ increase in energy intake, DIT increased by 1.1 kJ/h (p < 0.001). Meals with a high protein or carbohydrate content had a higher DIT than high fat, although this effect was not always significant. Meals with medium chain triglycerides had a significantly higher DIT than long chain triglycerides (meta- analysis, p = 0.002). Consuming the same meal as a single bolus eating event compared to multiple small meals or snacks was associated with a significantly higher DIT (meta- analysis, p = 0.02). Unclear or inconsistent findings were found by comparing the consumption of meals quickly or slowly, and palatability was not significantly associated with DIT. These findings indicate that the magnitude of the increase in DIT is influenced by the energy intake, macronutrient composition, and eating pattern of the meal.

Keywords: breakfast; meal; overnight fast; energy intake; macronutrient; diet-induced thermogenesis; thermic effect of food; meal-induced thermogenesis; resting metabolic rate

5.3 Introduction The meal consumed after an overnight fast, generally referred to as breakfast, is often described as ‘the most important meal of the day’(Brown, Bohan Brown et al. 2013) as it is believed to contribute to good health and nutrition by providing essential nutrients 96

early in the day (Rampersaud, Pereira et al. 2005). Skipping breakfast is associated with increased weight gain and obesity, suggesting that breakfast may be protective against weight gain (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013). Among the explanations for this protective effect of breakfast are that it stimulates the body’s metabolism because it breaks the overnight fast (DAA 2015), potentially contributing to increased total daily energy expenditure (EE). The extent of this effect would depend on the diet induced thermogenesis (DIT) response to the meal consumed. Evidence supporting this proposal is limited and contradictory (Betts, Richardson et al. 2014, Kobayashi, Ogata et al. 2014). Alternatively, eating breakfast may result in decreased energy consumption during the rest of the day, however the evidence available from previous trials in this area is also limited and contradictory(Brown, Bohan Brown et al. 2013).

Obesity is a major public health concern internationally with an estimated 13% and 39% of adults worldwide being obese or overweight respectively (WHO 2016), and 63% being either overweight or obese in Australia (AIHW 2016). Breakfast is often advocated as a strategy to prevent weight gain. However, a recent review (Casazza, Fontaine et al. 2013) and a separate meta-analysis (Brown, Bohan Brown et al. 2013) found that there is very limited evidence regarding the effects of breakfast on preventing weight gain (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013),and as most of the studies conducted have only been observational, there has been little investigation into the mechanisms by which breakfast may exert effects on preventing obesity (Brown, Bohan Brown et al. 2013). Both of these reviews found insufficient evidence to support consumption of breakfast for obesity prevention and suggested that further research in this area is required (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013).

Given the suggested importance of breakfast in the health arena, there are surprisingly few systematic reviews (SRs) consolidating the evidence of its effects on obesity prevention (Szajewska and Ruszczynski 2010, Brown, Bohan Brown et al. 2013, Williams 2014) and no SR and/or meta-analyses have investigated the effect of consuming breakfast on accelerating DIT, which may contribute to a reduced risk of weight gain. Most studies investigating the effects of food on DIT investigate these

effects after an overnight fast. Therefore these studies provide insights into the possible effects of breakfast on DIT although the meals used in these studies may not be typical of those eaten as breakfast and their goal may not have been to investigate the effects of breakfast.

No SRs have been conducted to compare the effects of meals after an overnight fast of varying macronutrient or micronutrient compositions, or different energy densities, on DIT. This suggests the need for a SR to be conducted in this area to address the lack of cohesive evidence of the role of these meals on DIT. This is of particular interest because even small changes in DIT may have significant effects on body weight and/or body composition over the longer term. Specifically, it has been suggested that an imbalance of 10–20 kcal/day can result in 0.5–1 kg of weight gain annually (Lean and Malkova 2016).

The primary question in this paper is whether there is any difference in the effects on DIT of meals consumed after an overnight fast of varying energy intake or macronutrient composition. Secondary questions are whether there is any difference in the effects on DIT if the same meal content is consumed using different eating patterns, such as a bolus meal versus repeated snacking, the effects of fast versus slow consumption of food, and whether food palatability of this meal has an effect on DIT. The outcomes of this SR will assist in better understanding the role of meal or snack consumption after an overnight fast on DIT and help to inform further research on the potential role of breakfast in health, as well as obesity prevention, treatment, and management.

5.4 Materials and Methods The SR protocol was published in ‘Prospero’ (CRD42014009030). Methodological decisions about the review process were made a priori.

5.4.1 Search With the assistance of a research librarian, four databases were searched: Cochrane (from 1992 to 14 May 2014), Cinahl (from 1937 to 14 May 2014), Embase (from 1947 to 14 May 2014), and Medline (from 1946 to 14 May 2014), and were then updated on the 23rd of February 2016 in order to capture the studies published between 14 May

2014 and 23 February 2016. This search used the following keywords: breakfast, morning meal, diet induced thermogenesis, thermogenesis, meal induced thermogenesis, thermic effect of food, resting metabolic rate, postprandial or post prandial metabolism, postprandial or post prandial metabolic rate, postprandial or post prandial energy expenditure, resting energy expenditure, postabsorptive or post absorptive energy expenditure, postabsorptive or post absorptive metabolic rate, basal metabolism, basal metabolic rate, and metabolic rate. For Embase, Medline, and Cinahl, limits were applied to include only human studies, those in English, and those conducted in adults. The Cochrane database did not allow these limits, however the word ‘adults’ was added as a keyword in order to limit the search to studies performed in adults. On the 23rd of February, this SR search was also expanded by the substitution of the keyword ‘morning meal’ by the keyword ‘meal’, in order to find studies that administered a meal after an overnight fast (breakfast) but did not use the terms morning meal or breakfast in the article. Also the keywords ‘thermic’ and ‘thermogenic’ were added to the search to ensure studies using slightly different language were not missed.

5.4.2 Eligibility Criteria

5.4.2.1 Inclusion and Exclusion Criteria For this review, only studies designated as level A evidence (randomized controlled trials (RCTs) and randomized crossover trials), as defined by the Academy of Nutrition and Dietetics, and with two or more eating events for comparison, were included. Studies were included if they provided a snack or a meal in the morning after participants fasted overnight. Studies were excluded if they provided infusions, injections, or capsules with the meal (e.g., saline or drug infusion, drug or placebo capsules). Interventions consisting of meals administered as enteral or intranasal or intra-gastric infusion or consisting of supplements instead of meals (e.g., protein or fat or sugar emulsions) or of meals supplemented with other components (e.g., addition of cellulose or pectin) including stimulants (e.g., caffeine, green tea, chilli, capsaicin, alcohol) were excluded. In these studies, the control meal (e.g., oral feeding or meal without stimulants) data were extracted if provided. When studies provided additional non-dietary interventions (e.g., exercise, sleep interventions), only data from the first meal consumed after an overnight fast intervention of the controlled arm were extracted.

Studies were included if they were published in English with male and/or female adult (≥18 years old) human participants. Data from populations such as children, adolescents, athletes or exercise-trained groups, patients with chronic or acute disease, obese individuals, pregnant or lactating women, or smokers were excluded. Studies with mixed populations of healthy weight, overweight, and obese populations were included; however, studies targeting only obese subjects or a mix of overweight and/or obese participants were excluded. Studies were also excluded when the majority of the participants were obese leading to a mean or median body mass index (BMI) ≥30 kg/m2. When studies compared specific populations (e.g., obese, pregnant women, athletes, smokers) to a control group, only data from the control group were included. The original search included all study designs, however, only RCT or randomized cross over designs were included in the analyses for this paper. Articles were excluded if they were expert opinion papers or if they described animal, in vitro, or in vivo experimental studies.

5.4.2.2 Outcome Measures (Dependent Variables) Diet induced thermogenesis measured by indirect calorimetry was the main outcome measure. Other outcomes of interest were indirect calorimetry fasting RMR and postprandial energy expenditure.

5.4.2.3 First Meal Consumed after an Overnight Fast (Independent Variables) The intervention was the first snack or meal of the day consumed in the morning after an overnight fast. Macronutrient compositions were described as percentages of the energy content of the meal. Energy was expressed in kJ.

5.4.2.4 Systematic Review Process Titles and abstracts were assessed for full text retrieval (Angelica Quatela (A.Q.)). Full text articles were assessed against the inclusion and exclusion criteria by two independent reviewers (A.Q. and Amanda Patterson (A.P.)). The quality criteria checklist for primary research of the Academy of Nutrition and Dietetics was used to assess the quality of the included studies by two independent reviewers (A.Q. and Lesley MacDonald-Wicks (L.M.-W.) or A.P.). The quality criteria tool assessed the studies for relevance and validity of the selected publications. A study was deemed positive if it met all the priority criteria, at least one of the validity criteria, and all of the relevance questions. A neutral rating indicated that most of the validity criteria were met 100

but the study may not have met one or more of the priority criteria and/or one or more of the relevance questions. A study was rated as negative if six or more of the validity and/or priority criteria were rated negative. Any discrepancies between reviewers at the full text and quality stage were assessed by a fourth reviewer (Robin Callister (R.C.)) until a consensus decision was reached.

5.4.2.5 Data Extraction The relevant data from the studies were extracted into tables (AQ) and evaluated for completeness (AP, RC, LMW). The following information was extracted: study design, significance, inclusion and exclusion criteria, country location, sample size, participant characteristics (intervention and comparator groups), recruitment, blinding used, intervention, statistical analysis, timing of measurements, dependent and independent variables, co-variates, length of follow up and results (key findings and other findings), and author conclusions.

5.4.2.5 Participant Characteristics Participant characteristics (age, gender, BMI, fat mass (FM), and fat free mass (FFM)) were extracted when provided or calculated from the data provided. BMI was calculated using the WHO criteria as illustrated in Supplementary Materials Table S1. FM and FFM were expressed as % of total body weight or in kg. If only individual participant data were provided, the mean, standard deviation (SD), and standard error (SE) were calculated with the formula described in Supplementary Materials Table S1. The percentage of males in the sample was calculated (100% indicated that only males were recruited and 0% only females).

5.4.2.6 Characteristics of the Meals The energy content of the meals was expressed in kJ. The conversion factor of 4.184 was used to convert kcal to kJ. When studies provided the macronutrient composition of meals only in grams, it was converted from g to % of energy using the two formulas described in Supplementary Materials Table S1.

5.4.2.7 Outcome Characteristics RMR, also known as Resting Energy Expenditure (REE), is defined as the quantity of energy used to maintain physiological function under resting conditions. DIT, also called the thermic effect of food, postprandial energy expenditure above baseline, or

101

meal-induced thermogenesis, is defined as the increase in RMR as a result of the consumption of food or a meal (Weststrate 1993, Reed and Hill 1996). DIT data were extracted in kJ and/or as the percentage of energy content of the meal (ECM) and/or as the percentage increase above baseline (AB) RMR. When the studies provided only the total postprandial energy expenditure and the RMR, the mean DIT was obtained from the difference of the total postprandial energy expenditure in kJ and the RMR in kJ for the same measure of time. DIT expressed in kJ or as percentage of energy content was divided by the number of hours that DIT was measured, or multiplied by 60 if it was provided in kJ/minute, in order to provide values as kJ per hour or percentage per hour. The formulas described in Supplementary Materials Table S1 were used to convert DIT from one unit of measurement to another when not provided by the authors.

5.4.2.8 Meta-Regressions The main outcome variable used in the meta-regression was DIT in kJ/h. Mixed model meta-regression was used to investigate the relationship between energy intake (kJ) after an overnight fast and DIT (kJ/h). The first model conducted was a univariate analysis, which only included DIT (kJ/h) and kJ intake. The second model also included four confounding factors (percentage of males, age, BMI and hours of DIT measurement). These meta-regression models were conducted using Stata/IC 13.1 (StataCorp LP, College Station, Texas, USA) and with consultant statistical support.

5.4.2.9 Meta-Analyses Fixed model meta-analyses were conducted in Review Manager (RevMan) to determine the mean difference in DIT (kJ/h) of pairs of comparisons. These meta-analyses were conducted with consultant statistical support.

5.5 Results A total of 2482 papers were identified from the four databases searched; 1756 papers remained after duplicates were removed and 351 full text articles were reviewed (Figure 5.1). Only 27 Level A evidence studies from 26 papers (one paper described two studies (Kasai, Nosaka et al. 2002)) were relevant to answer the review questions for this paper. Of the 26 papers, four were rated positive (Weststrate, Dopheide et al. 1990, Piers, Walker et al. 2002, Raben, Agerholm-Larsen et al. 2003, Clegg, Golsorkhi et al. 2013), none were rated negative, and the remaining 22 papers were rated neutral.

102

Table 5.1 summarizes the 27 studies for participants’ characteristics and study protocols. Table 5.2 summarizes the interventions and outcomes of the studies. Nine studies were conducted in the USA (Hill, Heymsfield et al. 1984, Segal, Edano et al. 1990, Tai, Castillo et al. 1991, Bennett, Reed et al. 1992, Bowden and McMurray 2000, Sawaya, Fuss et al. 2001, Thyfault, Richmond et al. 2004, Riggs, White et al. 2007, Barr and Wright 2010), five in Japan (Kasai, Nosaka et al. 2002, Nagai, Sakane et al. 2005, Hamada, Kashima et al. 2014, Toyama, Zhao et al. 2015), four in the UK (Kinabo and Durnin 1990, Kinabo and Durnin 1990, Blundell, Cooling et al. 2002, Clegg, Golsorkhi et al. 2013), two in Australia (Vaz, Turner et al. 1995, Piers, Walker et al. 2002), two in France (Martin, Normand et al. 2000, Allirot, Saulais et al. 2013), two in Denmark (Bendixen, Flint et al. 2002, Raben, Agerholm-Larsen et al. 2003), one in Germany (Petzke and Klaus 2008), one in Spain (Casas-Agustench, Lopez-Uriarte et al. 2009), and one in the Netherlands (Weststrate, Dopheide et al. 1990).The majority of the studies were not blinded, three were double blinded (Bendixen, Flint et al. 2002, Kasai, Nosaka et al. 2002), and two studies were single blinded (Piers, Walker et al. 2002, Riggs, White et al. 2007). One study provided intervention meals for two weeks for each arm (Martin, Normand et al. 2000) whereas all other studies provided only one day interventions.

5.5.1 Participant Characteristics A total of 350 participants were included. The participants’ characteristics are described in Table 5.1. The participants mean ages ranged from 20 to 69.4 years. Mean BMI ranged from 18.1 to 27.8 kg/m2. FM was mostly expressed in % and it ranged from 11.1% to 31.4%. FFM was only described in kg and it ranged from 41.7 to 66.8 kg. The sample size ranged from a minimum of four to a maximum of 29. The majority of the studies had a sample size <20; only four studies had ≥20 participants (Blundell, Cooling et al. 2002, Riggs, White et al. 2007, Casas-Agustench, Lopez-Uriarte et al. 2009, Allirot, Saulais et al. 2013). Fifteen studies had only males (Hill, Heymsfield et al. 1984, Segal, Edano et al. 1990, Bennett, Reed et al. 1992, Vaz, Turner et al. 1995, Martin, Normand et al. 2000, Sawaya, Fuss et al. 2001, Bendixen, Flint et al. 2002, Blundell, Cooling et al. 2002, Kasai, Nosaka et al. 2002, Piers, Walker et al. 2002, Thyfault, Richmond et al. 2004, Nagai, Sakane et al. 2005, Casas-Agustench, Lopez- Uriarte et al. 2009, Allirot, Saulais et al. 2013, Hamada, Kashima et al. 2014), eight studies had only females (Kinabo and Durnin 1990, Kinabo and Durnin 1990, Tai, 103

Castillo et al. 1991, Bowden and McMurray 2000, Kasai, Nosaka et al. 2002, Riggs, White et al. 2007, Petzke and Klaus 2008, Toyama, Zhao et al. 2015), and four studies had a mix of males and females (Weststrate, Dopheide et al. 1990, Raben, Agerholm- Larsen et al. 2003, Barr and Wright 2010, Clegg, Golsorkhi et al. 2013).

5.5.2 Interventions All studies required participants to attend the research setting in the morning after an overnight fast. The majority of the studies required arrival at the research center after a fasting period ranging from 10 to 14 h (Hill, Heymsfield et al. 1984, Kinabo and Durnin 1990, Segal, Edano et al. 1990, Tai, Castillo et al. 1991, Vaz, Turner et al. 1995, Bowden and McMurray 2000, Bendixen, Flint et al. 2002, Blundell, Cooling et al. 2002, Piers, Walker et al. 2002, Raben, Agerholm-Larsen et al. 2003, Thyfault, Richmond et al. 2004, Riggs, White et al. 2007, Petzke and Klaus 2008, Barr and Wright 2010, Allirot, Saulais et al. 2013, Hamada, Kashima et al. 2014, Toyama, Zhao et al. 2015). Many studies also required refraining from any exercise/physical activity or vigorous exercise either from the evening/dinner before (Vaz, Turner et al. 1995, Nagai, Sakane et al. 2005, Riggs, White et al. 2007, Hamada, Kashima et al. 2014), or from the day before (Bowden and McMurray 2000, Sawaya, Fuss et al. 2001, Allirot, Saulais et al. 2013, Clegg, Golsorkhi et al. 2013, Toyama, Zhao et al. 2015), or an even longer period of time (36 h to 3 days) (Hill, Heymsfield et al. 1984, Segal, Edano et al. 1990, Bennett, Reed et al. 1992, Bendixen, Flint et al. 2002, Piers, Walker et al. 2002, Raben, Agerholm-Larsen et al. 2003, Thyfault, Richmond et al. 2004). Fasting RMR was then measured and the meal administered. The interventions differed in energy intake and macronutrient composition between studies. Energy intakes for meals ranged from 418 to 6276 kJ. Carbohydrate (CHO) ranged from 0% to 90.4%, protein from 1.3% to 34.0%, and fat from 1.0% to 78.8% of energy intake. DIT was measured after the meal was consumed.

5.5.3 Outcomes The mean fasting RMR measured before administering the meals ranged from 191.6 to 375.3 kJ/h. The majority of studies measured DIT periodically rather than continuously over two (Vaz, Turner et al. 1995), three (Hill, Heymsfield et al. 1984, Segal, Edano et al. 1990, Toyama, Zhao et al. 2015), three and a half (Nagai, Sakane et al. 2005, Riggs, White et al. 2007), four (Martin, Normand et al. 2000, Allirot, Saulais et al. 2013), five

104

(Kinabo and Durnin 1990, Tai, Castillo et al. 1991, Bowden and McMurray 2000, Piers, Walker et al. 2002, Raben, Agerholm-Larsen et al. 2003, Barr and Wright 2010), or six hours (Kinabo and Durnin 1990, Sawaya, Fuss et al. 2001, Kasai, Nosaka et al. 2002, Petzke and Klaus 2008, Barr and Wright 2010, Clegg, Golsorkhi et al. 2013); only eight studies measured it continuously over one and a half (Hamada, Kashima et al. 2014), three (Blundell, Cooling et al. 2002), three and a half (Weststrate, Dopheide et al. 1990), four (Thyfault, Richmond et al. 2004) , five (Bendixen, Flint et al. 2002, Raben, Agerholm-Larsen et al. 2003, Casas-Agustench, Lopez-Uriarte et al. 2009), or six (Bennett, Reed et al. 1992) hours. Therefore, the duration of the DIT measurement period ranged from one and a half to six hours. The majority of studies measured DIT for five (Kinabo and Durnin 1990, Tai, Castillo et al. 1991, Bowden and McMurray 2000, Bendixen, Flint et al. 2002, Piers, Walker et al. 2002, Raben, Agerholm-Larsen et al. 2003, Casas-Agustench, Lopez-Uriarte et al. 2009, Barr and Wright 2010) or six hours (Kinabo and Durnin 1990, Bennett, Reed et al. 1992, Sawaya, Fuss et al. 2001, Kasai, Nosaka et al. 2002, Petzke and Klaus 2008, Barr and Wright 2010, Clegg, Golsorkhi et al. 2013). Mean DIT ranged from 4.5 to 99.4 kJ/h and from 0.77 to 4.3%/h (ECM), or from 1.3% to 41% above baseline (AB).

105

Figure 5.1 PRISMA Flow diagram (Prisma 2015) systematic search and review process.

106

Table 5.1 Participant characteristics and study protocols.

RMR (kJ/h, Sample (n) BMI (kg/m2) Gap between Intervention Reference Measured before Males (m) % FFM (% or kg) Protocol Meals and Types of Meals and Location Interventions a, b, Age (years) FM (% or kg) Provided c, etc.) Higher vs. lower energy intake Gap = each meal provided on different days. Kinabo and (a) high CHO, low Fat—lower Durnin BMI = 20.8 (0.2) † (a) 222 † (69.7) † Arrival at 8.00 AM; fasting energy (Kinabo and n= 16 FFM = 42.9 (3.1) † kg (b) 219 † (69.7) † from 8.00 PM. 30 min rest; (b) low CHO, high Fat—lower Durnin M = 0 † % FM = 240 (23.2) † g/kg body (c) 221.4 † (69.7) † RMR measured, meal energy 1990)—Paper 22 (5.8) † years weight (d) 208.2 † (69.7) † consumed within 10 min. (c) high CHO, low Fat—higher A †† energy UK (d) low CHO, high Fat—higher energy

n= 8 (4 low VO2max

group average is 43 High VO2max ml/kg/min—4 high BMI = 21.9 † High VO max Hill et al. 2 Gap = NPI VO2max group average FFM = 62.9 (3.6) † kg subjects 301.7 † 12 h fast; rest for 60–90 (Hill, Liquid meal (a) lower energy is 62 ml/kg/min) FM = 13.8 (2.8) † % (22.5) † min; then RMR measured; Heymsfield et intake M = 100 † % Low VO max Low VO max meals consumed within 10 al. 1984) 2 2 (b) medium level energy intake high VO max 20 (3.5) † BMI = 23.7 † subjects 294.6 † min. USA 2 (c) higher energy intake years FFM = 66.8 (6.9) † kg (40.5) † † † low VO2max 26 (5.2) FM = 16 (4.2) % years 14 days of intervention Gap = 28 days Martin et al. meals: meal consumed at (a) 14 days of intervention (Martin, n= 10 research centre daily meals: low energy, moderate fat Normand et M = 100 † % BMI = 22.2 (0.5) between 7.00 and 8.00 AM. breakfast al. 2000) 28 (2) years On day 15: arrive at 7.00 (b) 14 days of intervention France AM; overnight fast; RMR meals: high energy, low fat 107

RMR (kJ/h, Sample (n) BMI (kg/m2) Gap between Intervention Reference Measured before Males (m) % FFM (% or kg) Protocol Meals and Types of Meals and Location Interventions a, b, Age (years) FM (% or kg) Provided c, etc.) measured; meal consumed breakfast within 30 min at 8.00 AM. Bennet et al. n= 4 (untrained) Gap ≥ 24 h (Bennett, M = 100 † % (a) normal meal provided (25% BMI = 23 † (3) † Meal consumed at 8.30 Reed et al. 28 (8) years of energy intake of the total day) FM = 20.5 † (4.3) † % AM. 1992) Excluded trained (b) high fat meal: plus 50 g of fat USA subjects compared to the normal fat meal Segal et al. n= 11 12 h fast; 9.00 AM arrival; Gap = NPI BMI = 25.5 † (Segal, Edano M = 100 † % 30 min rest; RMR (a) 35% of each man 24 h RMR FFM = 66.1 (4.1) † kg 343.9 † (31.6) † et al. 1990) 31 (6.3) † years measured for three five- meal FM = 15.3 (2.2) † % USA Excluded obese subjects minute measurements. (b) 3013 kJ meal Meals varying in macronutrients composition Fast from 10.00 PM; arrival Nagai et al. at 7.30 AM; rest for 30 min; (Nagai, n= 14 Gap = NPI BMI = 21.3 (1.4) † 375.3 † (39.7) † continuous RMR Sakane et al. M = 100 † % (a) standard meal—low fat meal FM = 18.4 (3.6) † % 373.2 † (44.0) † measurement. (NP length); 2005) 23.6 (1.8) † years (b) standard meal—high fat meal meals consumed at 8.30 Japan AM within 15 min. n= 24 Blundell, M = 100 † % High fat consumers Cooling and Arrival at ~9.00 AM after High fat consumers (n = BMI = 21.2 (5.0) † High fat consumers King 12 h fast; 30 min of steady Gap = NPI 12) FM = 11.1 (4.3) † % 286.08 † (0.02) † (Blundell, RMR were measured; (a) high fat milkshakes drink 20.7 (1.6) † years Low fat consumers BMI = Low fat consumers Cooling et al. milkshake consumed within (b) high CHO milkshake drink Low fat consumers (n = 22.4 (2.0) † 259.58 † (0.03) † 2002) 5 min. 12) FM = 11.4 (4.3) † UK 21.6 (2.3) † years Bowden and n= 6 BMI = 21.7 (1.6) Arrival at lab at 6.30 AM; Gap = 2 days 208.2 † (12.6) † McMurray M = 0 † % FFM = 43.9 (3.3) Kg 10 h fast; RMR obtained (a) high CHO meal

108

RMR (kJ/h, Sample (n) BMI (kg/m2) Gap between Intervention Reference Measured before Males (m) % FFM (% or kg) Protocol Meals and Types of Meals and Location Interventions a, b, Age (years) FM (% or kg) Provided c, etc.) (Bowden and 33.8 (10.7) years FM = 21.4 (3.7) % over two 5 min periods; (b) high fat meal McMurray Excluded trained meal consumed within 20 2000) subjects min USA Thyfault et al. n= 12 Sedentary Arrival at 5.00 AM; 12 h (Thyfault, M = 100 † % fast; Gap = 7 days FM = 21.5 (4.3) † % 7262 (394.7) kJ †— Richmond et 24.8 (4.6) † years 30 min supine rest; RMR (a) high carbohydrate liquid meal FFM = 58.3 (2.7) † kg NP unit of time al. 2004) Excluded trained measured for 30 min; meal (b) moderate fat liquid meal USA subjects consumed within 10 min. Gap = ≥ 4 weeks and no more Raben et al. Arrival at 8.00 AM; 10 h than 8 weeks (Raben, n= 19 fast; 30 min supine rest; standard meal Agerholm- BMI = 22.1 (1.7) † M = 52.6 † % RMR measured for 45 min; (a) high protein meal Larsen et al. FM = 18.8 (4.7) † % 23.3 (2.1) † years meals consumed within 15 (b) high fat meal 2003) min at 9.45 AM. (c) high CHO meal Denmark Excluded high alcohol meal Petzke and 12 hr fast; RMR measured Klaus (Petzke n= 6 for 30min between 8 and 9 Gap = 2 days (a) 218 (12) and Klaus M = 0 † % BMI = 20.6 (2.5) AM; meal ingested between (a) low protein meal (b) 230 (13) 2008) 25.5 (2.6) years 9 and 9:30 AM and within (b) adequate protein meal Germany 10 min. n= 21 Overweight M = 0 † % BMI = 26.9 (1.7) † Riggs et al. 12 h fast; 10 min rest; RMR Overweight (n = 6) FFM = 48.4 (3.9) † kg Gap = 1 week to 2.5 months (Riggs, White measured between 7 and 22.8 (2.4) † years FM = 31.4 (2.7) † % (a) high protein, high fat bars et al. 2007) 8.00 AM; meal eaten within Normal weight (n = 12) Normal weight (b) high protein, low fat bars USA 15–20 min. 20.8 (2.6) † BMI = 21.1 (1.7) † Underweight (n = 3) FFM = 44.6 (4.0) † kg

109

RMR (kJ/h, Sample (n) BMI (kg/m2) Gap between Intervention Reference Measured before Males (m) % FFM (% or kg) Protocol Meals and Types of Meals and Location Interventions a, b, Age (years) FM (% or kg) Provided c, etc.) 20.7 (2.2) † years FM = 23.0 (2.9) † % Underweight BMI = 18.1 (1.5) † FFM = 41.7 (4.0) † kg FM = 19.5 (2.4) † % Arrived after an overnight Gap = minimum of four days Clegg et al. fast; rested for 30 min; (a) meal with bell pepper and (Clegg, n= 7 RMR measured between sunflower oil (18.4 g) Golsorkhi et M = 14.3 † % BMI = 21.9 † 7.30 AM to 9.00 AM at 1 (b) meal with bell pepper and al. 2013) 25.7 (3.6) years min intervals for 30 min; MCT oil (20.0 g) UK meal consumed within 15 Excluded two chilli meals min. Kasai et al.— study 1 Gap = NP n= 8 (a) 294.6 † (32.3) † Dinner at 9.00 PM; (Kasai, Liquid meal with (a) 10 g LCT M = 100 † % BMI = 22.7 (2.1) † (b) 280.3 † (29.9) † overnight fast; meal Nosaka et al. (b) 5 g MCT; 5 g LCT 26.8 (1.9) † years (c) 286.3 † (23.5) † consumed at 11.00 AM. 2002) (c) 10 g MCT Japan Gap = 1 to 2 day interval within the same week for each Kasai et al.— experimental session n= 8 (n = 7 for the two study 2 (a) 211.1 † (14.9) † (mayonnaise trials and margarine mayonnaise arms as one Dinner at 9.00 PM; (Kasai, (b) 206.2 † (22.5) † trials) drop out) BMI = 18.8 (1.1) † overnight fast; meal Nosaka et al. (c) 209.4 † (34.9) † Standard meal with (a) M = 0 † % consumed at 11.00 AM. 2002) (d) 198.2 † (32.8) † mayonnaise with 5 g LCT 28.1 (3.7) † years Japan (b) mayonnaise with 5 g MCT (c) margarine with 5 g LCT (d) margarine with 5 g MCT

110

RMR (kJ/h, Sample (n) BMI (kg/m2) Gap between Intervention Reference Measured before Males (m) % FFM (% or kg) Protocol Meals and Types of Meals and Location Interventions a, b, Age (years) FM (% or kg) Provided c, etc.) Casas- (a) 318.5 (95% CI Agustench et Arrived at 8.00 AM fast; 10 298.2—338.8) Gap = 1–11 days al. (Casas- n= 29 min rest; RMR measured (b) 318.9 (95% CI (a) standard meal rich in PUFA Agustench, M = 100 † % BMI = 24.1 (4.5) for 30 min; meal provided 298.5—339.3) (b) standard meal rich in MUFA Lopez-Uriarte 22 (4) years at 9.00 AM and eaten (c) 323.2 (95% CI (c) standard meal rich in SFA et al. 2009) within 30 min 302.3—344.1) Spain Piers et al. Arrival at 7.00–8.00 AM; n= 14 BMI = 27.8 (3.2) Gap = 7–14 days (Piers, Walker (a) 311 (40) 12–14 h of fast; 30 min M = 100 † % FFM = 62.7 (8.5) kg (a) meal with SFA et al. 2002) (b) 307 (36) rest; RMR measured for 35 38 (9) years FM = 29.5 (4.8) % (b) meal with MUFA Australia min. Gap = 14–28 days Standard meal with liquid test drink (a) conventional fat (rapeseed oil) Bendixen et (b) chemically structured fat (a) 290.4 † (28.5) † Fast ≥ 12 h; 30 min supine al. (Bendixen, n= 11 BMI = 22.5 (1.9) † (rapeseed oil and octanoic acid (b) 289.8 † (22.8) † rest; RMR measured for 45 Flint et al. M = 100 † % FM = 18 (3.2) † % by esterification with sodium (c) 292.2 † (26.6) † min; 2002) 25.1 (1.6) † years FFM = 62.9 (6.3) † kg methoxide) (d) 289.8 † (34.2) † meal consumed by 15 min. Denmark (c) lipase structured fat (rapeseed oil and octanoic acid by esterification with lipoxime IM) (d) physically mixed fat (blending rapessed oil and trioctanoate) Processed vs. unprocessed meals Barr and n= 17 in analyses BMI = 22.0 (2.2) † Fast for 12 h; 2 RMR Gap = on two consecutive days

111

RMR (kJ/h, Sample (n) BMI (kg/m2) Gap between Intervention Reference Measured before Males (m) % FFM (% or kg) Protocol Meals and Types of Meals and Location Interventions a, b, Age (years) FM (% or kg) Provided c, etc.) Wright (Barr M = 29.4 † % measurements ~30 min or not longer than a week apart. and Wright 25.5 (12.4) † years apart before and just before (a) whole-food meal as either 1½ 2010) consuming the meal; meals sandwich or 2 sandwiches USA consumed between 9.15 & (b) pre-prepared processed foods 11.15 AM and at as either 1½ sandwich or 2 approximately the same sandwiches. time for each measurement sessions. Bolus vs. smaller frequent meals Gap = 1 week Group A: Group A Arrival at 8.00 AM, at least (a) high carb-low fat meal; one Kinabo and n= 18 BMI = 21 (1.3) † Group A: 12 h fast; 30 min supine large meal Durnin M = 0 † % FFM = 42.4 (2.9) † kg (a) 226.8 † (27.0) † rest; RMR measured; meal (b) high carb-low fat meal; two (Kinabo and Group A (n = 8) FM = 23 (2.6) † % (b) 214.8 † (20.6) † consumed either as a large smaller meals Durnin 1990)- 24 (5.3) † years Group B Group B bolus meal within 20 min or Group B: Paper B Group B (n = 10) BMI = 21 (2.1) (c) 230.4 † (30.6) † as two smaller meals within (c) low carb-high fat meal; one Scotland, UK 20 (7.2) † years FFM = 43.6 (4.2) † kg (d) 221.4 † (14.4) † 10 min every 180 min. large meal FM = 23 (6) † % (d) low carb-high fat meal; two smaller meals Vaz et al. Gap = approximately 14 days BMI = 22.9 (1.8) † 12–14 h overnight fast; (Vaz, Turner n = 10 (a) standard meal-single meal FFM = 64 (5.7) † kg RMR measured after 30 et al. 1995) M = 100 † % (b) standard meal- three smaller FM = 16.6 (6.0) † % min rest. Australia meals Allirot et al. Arrival at 7.00 AM; fast Gap = at least 7 days n= 20 (Allirot, since 9.00 PM; RMR (a) one 20 min long episode M = 100 † % BMI = 22.0 (1.3) † Saulais et al. measured for 30 min; (b) 4 smaller meals in 10 min 27.1 (5.7) † years 2013) meals consumed either as a episodes

112

RMR (kJ/h, Sample (n) BMI (kg/m2) Gap between Intervention Reference Measured before Males (m) % FFM (% or kg) Protocol Meals and Types of Meals and Location Interventions a, b, Age (years) FM (% or kg) Provided c, etc.) France bolus event within 20 min or as smaller meals every hour within 10 min each. Gap = meals provided on RMR measured after 12–14 different days Tai et al. (Tai, h fast and minimum of 30 n= 7 (a) liquid meal taken in one Castillo et al. BMI = 20.8 (2.1) (a) 233.7 † (9.8) † min rest; meals consumed M = 0 † % eating event of 10 min long 1991) FM = 17.1 (5.4) % (b) 236.22 † (14.1) † as one bolus event within 26.7 (2.9) years (b) liquid meal in 6 equal smaller USA 10 min or as smaller meals meals at 30 min interval over every 30 min. 150 min. Fast vs. slow eating patterns Hamada et al. Fast since dinner (>10 h); Gap = NPI (Hamada, n= 10 BMI = 19.8 † 20 min semi-supine standard meal Kashima et al. M = 100 † % FM = 13 (2) % position rest; RMR (a) rapid eating 2014) 25 (1) years measured for 20 min. (b) slow eating Japan BMI Toyama et al. Dinner by 9.00 PM, fast (a) 21.3 (1.7) Gap = at least 7 days (Toyama, n= 9 until morning; arrival at (b) 21.3 (1.8) (a) 196.8 † (17.3) † Same meal provided Zhao et al. M = 0% 8.00 AM, 30 min supine FM (b) 191.6 † (17.6) † (a) fast eating (5 min) 2015) 22 (2.1) years rest; RMR measured; meal (a) 24.1 (3.8) % (b) regular eating (15 min) Japan consumed at 9.00 AM. (b) 24.0 (4.0) % Palatable vs. unpalatable meals n= 19 Old (a) young 319.8 † Sleep at university by 10.00 Sawaya et al. M = 100 † % BMI = 24.4 (0.9) (32.4) † PM, awakened at 6.30AM; Gap = 1 week interval (Sawaya, Fuss Old (n = 9) FFM = 55 (2.2) old 280.2 † (30.5) † 30 min rest; RMR (a) palatable meal et al. 2001) 69.4 (1.3) years FM = 26.2 (1.9) % (b) young 319.2 † measured for 30 min; meals (b) control meal USA Young (n = 10) Young (32.4) † consumed within 20 min.

113

RMR (kJ/h, Sample (n) BMI (kg/m2) Gap between Intervention Reference Measured before Males (m) % FFM (% or kg) Protocol Meals and Types of Meals and Location Interventions a, b, Age (years) FM (% or kg) Provided c, etc.) 23.4 (1) years BMI = 22.7 (0.5) old 268.8 † (23.8) † FFM = 64.1 (1.9) kg FM = 12 (1.3) % Weststrate et al. n= 12 Men (Weststrate, M = 50 † % Gap = At least 2 days FM = 12.0 (2.2) † % Dopheide et Men 22.7 (1.8) † years Overnight fast. (a) palatable meal Women al. 1990) Women (b) unpalatable meal FM = 29.0 (2.5) † % Netherlands— 21.2 (1.8) † years study 1 Data are described in mean (SD) unless otherwise indicated. † These data (mean and/or SD) were calculated or converted for one or more of these possible calculations or conversions (calculated the average and/or SD from individuals’ data; kcal converted to kJ; MJ converted to kJ; RMR kJ converted for unit of time; SE converted to SD, males’ percentage calculated from the total number of males in the sample). †† Kinabo et al. [31]—paper A is also part of results section: “meals varying in macronutrient composition”. a, b, c or d: refers to the different types of interventions provided as explained in the last column (gap between intervention meals and types of meals provided). BMI = Body Mass Index. CHO = Carbohydrate. FM = Fat Mass. FFM = Fat Free Mass. LCT = Long Chain Triglycerides. M = Males. MCT = Medium Chain Triglycerides. MUFA = Mono Unsaturated Fatty Acids. N = Sample. NP = Not Provided. NPI = Not Provided Information. PUFA = Poly Unsaturated Fatty Acids. RMR = Resting Metabolic Rate. SFA = Saturated Fatty Acids. VO2 = Rate of Oxygen consumption.

114

5.5.4 Comparison and Meta-Regression of the Effects of Higher and Lower Energy Intakes on DIT

Five studies (Hill, Heymsfield et al. 1984, Kinabo and Durnin 1990, Segal, Edano et al. 1990, Bennett, Reed et al. 1992, Martin, Normand et al. 2000) with the primary aim of comparing the effects of meals with different energy intakes on DIT were identified. Three studies (Hill, Heymsfield et al. 1984, Kinabo and Durnin 1990, Martin, Normand et al. 2000) found an increased DIT when a higher energy intake was consumed, although only one indicated statistical significance (Kinabo and Durnin 1990). Kinabo and Durbin (Kinabo and Durnin 1990) compared high CHO, low fat meals at two energy intake levels: 2520 kJ and 5040 kJ, and low CHO, high fat meals at the same two energy intake levels. This study found that a higher energy intake was associated with a significantly (p < 0.001) higher DIT, regardless of dietary composition. Higher energy intake (5040 kJ) resulted in a similar DIT for the high CHO, low fat (71.2 (15.5) kJ/h) and low CHO, high fat (68 (12.4) kJ/h) meals, and this was higher than the DIT for the lower energy intake (2520 kJ) high CHO, low fat (45.6 (9.3) kJ/h) and low CHO, high fat (45.6 (10.8) kJ/h) meals (Kinabo and Durnin 1990).

Hill et al. (Hill, Heymsfield et al. 1984) compared DIT following three meals of 2092 kJ, 4184 kJ, and 6276 kJ (Hill, Heymsfield et al. 1984). DIT was higher with higher energy intake: 2092 kJ meal DIT <10% above baseline RMR; 4184 kJ meal DIT 21% above baseline RMR, and 6276 kJ meal DIT 33.5% above baseline RMR, no p value provided (Hill, Heymsfield et al. 1984).

Martin et al. (Martin, Normand et al. 2000) compared two weeks of low energy, moderate fat meals (418 kJ) to two weeks of high energy, low fat meals (2929 kJ) and found a higher DIT after the high energy, low fat meals (low energy, moderate fat meals 4.5 (1.4) kJ/h; high energy, low fat meals 35.6 (2.6) kJ/h; no p value provided (Martin, Normand et al. 2000)).

Bennet et al. (Bennett, Reed et al. 1992) compared a high fat meal (kJ not provided) to a normal fat meal (kJ not provided). The high fat meal was 1881 kJ higher due to the addition of 50 g of fat compared to the normal fat meal. This study did not find any significant differences in DIT (high fat meal 1.2 (0.40) %/h, normal fat meal

115

1.3 (0.3) %/h; p > 0.05 for %/6 h ECM for all participants, including some trained individuals, for the statistical tests) (Bennett, Reed et al. 1992).

Segal et al. (Segal, Edano et al. 1990) compared consuming a meal with a fixed energy intake (3013 kJ) to a meal providing 35% of each individual’s 24 h RMR (caloric intake varying between participants, on average 2889 kJ intake) [22]. This study did not find any significant difference in DIT (fixed: 96.3 (17.6) kJ/h; 35% RMR: 89.3 (17.6) kJ/h, p > 0.05 for %/3 h ECM) but there was little difference in the energy intakes between the two meals (Segal, Edano et al. 1990).

In order to further resolve the effect of energy intake on DIT, mixed model meta- regression analyses were undertaken to investigate more broadly the relationship between energy intake (kJ) after an overnight fast and DIT (kJ/h). Two models were produced: the first one included only energy intake (kJ) and the outcome variable DIT (kJ/h); the second model also included four confounding factors (percentage of males, BMI, age, and hours of DIT measurement).

Figure 5.2 represents Model 1 (coefficient 0.011, standard error 0.0013, p < 0.001, 95% confidence interval, (CI) 0.0083; 0.014) conducted for 19 studies (Kinabo and Durnin 1990, Kinabo and Durnin 1990, Segal, Edano et al. 1990, Weststrate, Dopheide et al. 1990, Tai, Castillo et al. 1991, Vaz, Turner et al. 1995, Martin, Normand et al. 2000, Sawaya, Fuss et al. 2001, Bendixen, Flint et al. 2002, Blundell, Cooling et al. 2002, Kasai, Nosaka et al. 2002, Piers, Walker et al. 2002, Thyfault, Richmond et al. 2004, Nagai, Sakane et al. 2005, Riggs, White et al. 2007, Petzke and Klaus 2008, Allirot, Saulais et al. 2013, Clegg, Golsorkhi et al. 2013) with a total of 54 treatment arms. Eight studies could not be included in the meta-analyses because they had missing values for one or more of the variables investigated in the model. This model shows that DIT (kJ) increases significantly (p < 0.001) when the kJ content of meals increases, although this increase is of a small magnitude (coefficient 0.011). This model predicts that for every 100 kJ increase in energy intake, DIT increases by 1.1 kJ/h.

Model 2, adjusted for percentage of males, BMI, age, and hours of DIT measurement, also predicted a small but significant increase in DIT for every kJ intake (coefficient 0.012, standard error 0.0013, p < 0.001; CI: 0.0091; 0.014). This 116

model predicts that for every 100 kJ increase in energy intake, DIT increases by 1.2 kJ/h. In this model, 16 studies were included with a total of 48 arms. Three studies included in model 1 were not included in model 2 because they had missing values for one or more of the variables investigated (Weststrate, Dopheide et al. 1990, Vaz, Turner et al. 1995, Thyfault, Richmond et al. 2004). DIT accounted for 47.4% of the variance in Model 1 and 70.6% of the variance in Model 2.

Figure 5.2 Mixed Model Meta Regression: univariate association between energy intake (kJ) and DIT (kJ/h) (Model 1). The Figure is composed of circles and a regression prediction line (in red) representing the outcome (DIT); each circle represents the value of DIT (kJ/h) for an arm of a study, and the size of the circle is inversely proportional to the standard error (SE) of the study. The influence of each study on the model depends on the size of the SE. Specifically, a study arm with a large SE is represented in the figure by a small circle, which means that this study arm had a small influence on the model whereas a study arm with a small SE is represented by a large circle, which means that this study arm had a large influence on the model.

5.5.5 Influence of Macronutrient Composition on DIT

Six studies (Kinabo and Durnin 1990, Bowden and McMurray 2000, Blundell, Cooling et al. 2002, Raben, Agerholm-Larsen et al. 2003, Thyfault, Richmond et al. 2004, Nagai, Sakane et al. 2005) compared meals differing in macronutrient composition (fat vs. CHO and/or vs. protein). Five of these papers compared consuming a meal high in CHO with a meal high in fat. Nagai et al. (Nagai, Sakane 117

et al. 2005) reported a higher DIT with a high CHO meal (3255 (306.5) kJ) compared to an isocaloric meal high in fat (3255 (306.5) kJ). DIT was 43.1 (13.7) kJ/h for the high CHO meal and 32.6 (14.1) kJ/h for the high fat meal, p < 0.05 for %/3.5 h ECM (Nagai, Sakane et al. 2005). Blundell et al. (Blundell, Cooling et al. 2002) provided isocaloric comparisons (both meals contained 2092 kJ) and found a statistically significant effect on DIT (high CHO milkshake: habitually high fat consumers 38.2 (26.0) kJ/h and habitually low fat consumers 35.2 (15.6) kJ/h; high fat milkshake: habitually high fat consumers 27.5 (28.9) kJ/h and habitually low fat consumers 25.6 (14.5) kJ/h, p < 0.05 for kJ/day) (Blundell, Cooling et al. 2002). The other two studies provided meals with only small differences in energy content (high CHO meal 2068 kJ and high fat meal 2093 kJ) (Bowden and McMurray 2000); high CHO meal 3021 (1194.0) kJ and moderate fat meal 2996 (1167.4) kJ) (Thyfault, Richmond et al. 2004)), and DIT was as follows: high CHO meal 54.6 kJ/h, high fat meal 27.8 kJ/h (Bowden and McMurray 2000); high CHO meal 57.8 (19.1) kJ/h and moderate fat meal 49.8 (21.6) kJ/h (Thyfault, Richmond et al. 2004). No p values were provided for these comparisons; therefore, it is not known if these comparisons were statistically significantly different (Bowden and McMurray 2000, Thyfault, Richmond et al. 2004).

One study provided isocaloric comparisons and found no significant effect on DIT between high CHO, low fat meals and low CHO, high fat meals (Kinabo and Durnin 1990). This study (Kinabo and Durnin 1990), which was described in section 3.4 (higher energy vs. lower energy intake), provided the same group of subjects with high CHO, low fat meals of two different energy contents (2510 kJ and 5040 kJ), as well as low CHO, high fat meals of two different energy contents (2520 kJ and 5040 kJ) [31]. The DIT data were as follows: 5040 kJ high CHO, low fat meal 71.2 (15.5) kJ/h and 2520 kJ high CHO, low fat meal 45.6 (9.3) kJ/h vs. 5040 kJ low CHO, high fat meal 68 (12.4) kJ/h and 2520 kJ low CHO, high fat meal 45.6 (10.8) kJ/h, p > 0.05 for kJ/5h comparing high CHO, low fat meals with low CHO, high fat meals (Kinabo and Durnin 1990).

Additionally, only one study (Raben, Agerholm-Larsen et al. 2003) compared consuming isocaloric meals rich in protein vs. fat vs. CHO in participants of the same sex (females consumed 2500 kJ and males 3000 kJ). The high CHO and fat 118

meals had the same DIT, whereas the high protein meal had a higher DIT (CHO meal 39.2 kJ/h, fat meal 39.2 kJ/h, and protein meal: 45.9 kJ/h, p < 0.01 for %/5h ECM comparing four meals (an alcohol meal was excluded for the purpose of this SR)) (Raben, Agerholm-Larsen et al. 2003).

One study (Petzke and Klaus 2008) investigated the effect of consuming an adequate level of protein (3131 kJ) with a low level of protein (3114 kJ) in meals with similar energy contents. This study found a higher DIT when an adequate level of protein was consumed compared to a lower level (adequate protein meal 22.4 (5.7) kJ/h; low protein meal 7.8 (1.0) kJ/h, p = 0.001 for kJ/6 h and %/6 h) (Petzke and Klaus 2008).

Riggs et al. (Riggs, White et al. 2007) undertook isocaloric comparisons of meals differing in the amount of fat provided and found a higher DIT after consuming a moderate fat meal (1841 kJ) than an isocaloric low fat meal (1841 kJ), where both meals were high in protein, among normal weight participants (p < 0.005 for in %ECM) but not in overweight or underweight participants [24]. The DIT results were as follows; normal weight: moderate fat meal 43.1 (19.2) kJ/h vs. low fat meal or 31.0 (19.2) kJ/h; overweight: moderate fat meal 48.4 (20.0) kJ/h vs. low fat meal 46.3 (18.8) kJ/h; underweight: moderate fat meal 20.0 (16.4) kJ/h or vs. low fat meal 24.2 (15.6) kJ/h (Riggs, White et al. 2007).

5.5.6 Long Chain Triglycerides vs. Medium Chain Triglycerides

Three studies (Kasai, Nosaka et al. 2002, Clegg, Golsorkhi et al. 2013) compared meals containing medium chain triglycerides (MCT) with long chain triglycerides (LCT) and all found a statistically higher DIT with meals containing MCT rather than meals containing LCT. Clegg et al. (Clegg, Golsorkhi et al. 2013) provided two meals of the same energy content and macronutrient profiles but containing either MCT (20 g) or LCT (18.4 g) (1863 kJ) [17] and found a significantly higher DIT with the MCT meal (MCT: 29.4 (8.4) kJ/h vs. LCT: 21.9 (7.9) kJ/h, p < 0.005 for %/6 h ECM) (Clegg, Golsorkhi et al. 2013).

Kasai et al. (Kasai, Nosaka et al. 2002) conducted two studies comparing the effects of MCT vs. LCT. In study 1, three meals were administered (5 g of MCT and 5 g of LCT (1029 kJ) meal, 10 g MCT meal (1013 kJ), 10 g LCT meal (1046 kJ) (Kasai, 119

Nosaka et al. 2002). DIT was significantly increased when MCT meals were consumed vs. LCT (10 g MCT meal 19.5 (14.1) kJ/h vs. 10 g LCT meal 8.4 (4.6) kJ/h, p < 0.01 for % ECM). Furthermore, DIT was significantly higher for the meal with both MCT and LCT than the one containing only LCT (5 g MCT, 5 g LCT meal 17.7 (10.8) kJ/h vs. 10 g LCT meal 8.4 (4.6) kJ/h, p < 0.01 for %/6 h ECM) (Kasai, Nosaka et al. 2002).

In Study 2 (Kasai, Nosaka et al. 2002), four meals were administered containing: mayonnaise with 5 g MCT (1042 kJ), mayonnaise with 5 g LCT (1059 kJ), margarine with 5 g MCT (1004 kJ), or margarine with 5 g LCT (1020 kJ) [14]. This study found a significantly higher DIT with meals containing MCT as opposed to LCT (meal with mayonnaise and MCT: 14.0 (5.7) kJ/h vs. meal with mayonnaise and LCT: 8.2 (6.4) kJ/h, p < 0.05 for % ECM, meal with margarine and MCT 20.3 (15.7) kJ/h vs. meal with margarine and LCT 9.8 (8.2) kJ/h, p < 0.05 for %/6 h ECM (Kasai, Nosaka et al. 2002).

A fixed model meta-analysis was conducted with these three studies (Kasai, Nosaka et al. 2002, Clegg, Golsorkhi et al. 2013) to compare DIT (kJ/h) for the MCT vs. LCT arms. Because Kasai et al. (2002) in study 2 [14] administered two interventions for MCT (margarine or mayonnaise) and two interventions for LCT (margarine or mayonnaise) to the same people, two forest plots are presented (one with only the margarine interventions and the other one with only the mayonnaise intervention arms). This avoids the effects of repetition of the same participants in both the margarine and mayonnaise studies. Both analyses found a significantly higher DIT when MCT was consumed compared to LCT (p = 0.002; Figure 5.3a,b). For both models the heterogeneity is 0% with chi2 = 0.3 and p = 0.9 (Figure 5.3a), and chi2 = 0.7 and p = 0.7 (Figure 5.3b). The total sample size is 23 (for Figure 5.3a) or 22 (for Figure 5.3b) for each group of comparisons with the same people repeated for both interventions.

120

Figure 5.3 (a) Meta-analysis with fixed effect of the mean differences in DIT between MCT and LCT with the margarine trials of Kasai et al. Studies included in this meta-analysis are represented in the figure by symbols (green squares) and they are illustrated in the following order: Clegg et al. (Clegg, Golsorkhi et al. 2013)- 46.0% weight; Kasai et al. [14] - study 1 - 31.6% weight; Kasai et al. (Kasai, Nosaka et al. 2002) - study 2 margarine trial - 22.4% weight (Kasai, Nosaka et al. 2002); (b) Meta-analysis with fixed effect of the mean differences in DIT between MCT and LCT with the mayonnaise trials of Kasai et al. Studies included in this meta-analysis are represented in the figure by symbols (green squares) and they are illustrated in the following order: Clegg et al. (Clegg, Golsorkhi et al. 2013) - 28.6% weight; Kasai et al. (Kasai, Nosaka et al. 2002) - study 1 - 19.6% weight; Kasai et al. (Kasai, Nosaka et al. 2002) - study 2 - mayonnaise trial- 51.8% weight [14]. The % contribution of each study to the outcome is indicated as % weight.

5.5.7 Monounsaturated Fat vs. Polyunsaturated Fat

Two studies (Piers, Walker et al. 2002, Casas-Agustench, Lopez-Uriarte et al. 2009) compared meals containing monounsaturated fatty acids (MUFA), 121

polyunsaturated fatty acids (PUFA), or saturated fatty acids (SFA). Casas- Agustench et al. (Casas-Agustench, Lopez-Uriarte et al. 2009) found a significantly higher DIT after the consumption of meals containing PUFA (mean (95% CI) 2655 (2510–2799) kJ) or MUFA (mean (95% CI) 2608 (2428–2788) kJ) compared to the one containing SFA (mean (95% CI) 2599 (2421–2278) kJ). The DIT as mean (95% CI) was: PUFA meal 37.2 (29.5–44.8) kJ/h, MUFA meal 36.8 (30.5–43.0) kJ/h, and SFA meal 30.0 (24.2–35.8) kJ/h, p < 0.05 for kJ/5h amongst the three interventions (Casas-Agustench, Lopez-Uriarte et al. 2009).

Contrary to this finding, the study by Piers et al. (Piers, Walker et al. 2002) found no significant difference in DIT (SFA meal 29.6 (10) kJ/h vs. MUFA meal 28.4 (10) kJ/h, p > 0.05 for %/5 h ECM and p > 0.05 for kJ/5 h between meals containing MUFA or SFA (both meals: 2500 kJ) (Piers, Walker et al. 2002).

5.5.8 Structure of Fats

Bendixen et al. (Bendixen, Flint et al. 2002) compared consuming meals with either a conventional fat (sunflower oil) or a chemically structured fat (rapeseed oil and octanoic acid by esterification with sodium methoxide) or a lipase-structured fat (rapeseed oil and octanoic acid by esterification with lipoxime IM) or a physically mixed fat (blending rapeseed oil and trioctanoate) (Bendixen, Flint et al. 2002). The mean energy content of these four meals was 4698 (550.2) kJ (Bendixen, Flint et al. 2002). This paper found a significant effect of fat structure on DIT with the highest DIT associated with the meal containing a chemically structured fat and the lowest with the meal having the conventional fat (conventional fat meal 61.8 (15.2) kJ/h, chemically structured fat meal 72.8 (19.0) kJ/h, lipase-structured fat meal 69.2 (11.4) kJ/h, and physically mixed fat meal 65 (13.9) kJ/h, p = 0.005 for kJ/5 h) (Bendixen, Flint et al. 2002).

5.5.9 Processed vs. Unprocessed Food

Only one study (Barr and Wright 2010) compared consuming two meals with different levels of processing. One meal was composed of whole food (multi-grain bread and cheddar cheese either as one and a half sandwiches or two sandwiches) and the other was composed of processed foods (white bread and a processed cheese either as one and a half sandwiches or two sandwiches) (Barr and Wright 122

2010). Subjects could choose to consume either one and a half sandwiches (2520 kJ) or two sandwiches (3360 kJ), and this choice was kept constant for both trials, thus the two trials were isocaloric for the same participant. There was a highly significant increase in DIT after consuming the whole food meal compared to the more processed meal (whole food meal: 99.4 (40.7) kJ/h; processed meal: 63.9 (35.6) kJ/h, p < 0.001 for total kJ and p < 0.01 for total % ECM (Barr and Wright 2010)).

5.5.10 One Bolus Event vs. Isocaloric Smaller Frequent Meals

Four studies (Kinabo and Durnin 1990, Tai, Castillo et al. 1991, Vaz, Turner et al. 1995, Allirot, Saulais et al. 2013) compared administering a meal as a bolus event versus splitting the same meal into into two (Kinabo and Durnin 1990), three (Vaz, Turner et al. 1995), four (Allirot, Saulais et al. 2013) or six (Tai, Castillo et al. 1991) smaller equal meals or snacks to be consumed throughout the morning. The time between multiple meals was 180 min [32], 60 min [35], or 30 min [27,34]. Kinabo and Durbin (Kinabo and Durnin 1990) compared two eating patterns using two different meal compositions: high CHO, low fat and low CHO, high fat. All four studies had the same participants perform both interventions (total n = 55). The energy density was as follows: 5040 kJ or 2510 kJ × 2 either as high CHO and low fat meal or low CHO and high fat meal (Kinabo and Durnin 1990); 3150 or 1050 kJ × 3 [34]; 2823.4 kJ or 705.8 KJ × 4 meals [35]; and 3138 kJ or 523 kJ × 6 (Tai, Castillo et al. 1991).

Two studies (Kinabo and Durnin 1990, Vaz, Turner et al. 1995) found no significant difference in DIT between the bolus and the isocaloric smaller frequent meals event (high CHO, low fat meal as bolus 62.8 (13.2) kJ/h vs. smaller frequent meals event 63.5 (11.7) kJ/h, p > 0.05 for kJ/6 h; low CHO, high fat meal as bolus 59.3 (11.5) kJ/h vs. smaller frequent meals event 56.7 (8.0) kJ/h, p > 0.05 for kJ/6 h (Kinabo and Durnin 1990); bolus 71 (31.5) kJ/h vs. smaller frequent meals 52.3 (15.3) kJ/h, p > 0.05 for kJ/2 h (Vaz, Turner et al. 1995)).

The other two studies found a significantly higher DIT when the meals were consumed as a bolus event compared to smaller frequent meal events: bolus 43.8 (18.4) kJ/h vs. smaller frequent meals 33.2 (15.5) kJ/h, p < 0.05 for %/4 h ECM

123

[35]; bolus 48.2 (16.93) kJ/h vs. smaller frequent meals 34.9 (12.3) kJ/h, p < 0.05 for kJ/5 h (Tai, Castillo et al. 1991)).

In order to clarify these discrepancies, a meta-analysis of the mean differences between bolus and smaller frequent meal event trials with fixed effects was conducted in RevMan to find the overall effect on DIT [32]. For these analyses, DIT was compared in kJ/h in order to standardize the units between studies. The forest plot shows (Figure 5.4) the mean of the difference between bolus and smaller frequent meal event trials for each study. The overall mean of the difference is positive, which means that the DIT was lower in the smaller frequent meals event trials compared to the bolus trial. This overall effect on DIT was significant (p = 0.02). The heterogeneity was 14%, chi2 = 4.6 and p = 0.3.

Figure 5.4 Meta-analysis: mean differences in DIT between bolus vs. smaller frequent meals event (such as snacking). Studies included in this meta-analysis are represented in the figure by symbols (green squares) and they are illustrated in the following order: Kinabo and Durbin et al. [32] - Paper B low CHO, high fat meal - 37.6% Weight; Kinabo and Durbin et al. [32] - Paper B high CHO, low fat meal - 19.0% weight; Vaz et al. [34] - 6.0% weight; Allirot et al. [35] - 25.5% weight; Tai et al. [27] - 11.8% weight. Weight refers to amount of influence that the study exerts on the meta- analyses. The % contribution of each study to the outcome is indicated as % weight.

5.5.11 Fast vs. Slow/Normal Meal Consumption

Two studies (Hamada, Kashima et al. 2014, Toyama, Zhao et al. 2015) compared consuming the same isocaloric meal quickly or more slowly. One study (Hamada, Kashima et al. 2014) compared eating a meal (1255.2 kJ) as fast as possible to a 124

meal chewed as many times as possible until no lumps remained before swallowing. The other study (Toyama, Zhao et al. 2015) compared eating a meal (1464 kJ) in 15 min compared to 5 min. Both studies found a higher DIT when the meal was consumed by slower eating compared to fast eating (slower eating 502.1 (234.4) kJ/kg/h vs. fast eating 19.5 (142.2) kJ/kg/h, p < 0.05 for kcal/kg/90 min (Hamada, Kashima et al. 2014); slower eating 41.9 (14.6) kJ/kg/h vs. fast eating 31.6 (15) kJ/kg/h, p > 0.05 for cal/kg/180 min) (Toyama, Zhao et al. 2015)), although only one of the studies reached statistical significance (Hamada, Kashima et al. 2014).

5.5.12 Palatable vs. Unpalatable

Two studies (Weststrate, Dopheide et al. 1990, Sawaya, Fuss et al. 2001) compared consuming palatable vs. unpalatable isocaloric meals (2930 kJ (Sawaya, Fuss et al. 2001); 2000 kJ (Weststrate, Dopheide et al. 1990)) on DIT. There was no significant difference in DIT between these two approaches indicating palatability did not influence DIT. In the first study (Sawaya, Fuss et al. 2001), the effects of palatability were examined in both young and old participants (old participants: palatable meal 37.0 (15.9) kJ/h vs. unpalatable meal 46.4 (18.4) kJ/h; young participants: palatable breakfast 34.7 (13.6) kJ/h vs. unpalatable meal 39.5 (17.0) kJ/h, p > 0.05 for %/3 h ECM). In the second study there was also no difference in DIT with palatability (palatable meal 47.3 (14.2) kJ/h vs. unpalatable meal 52.9 (13.3) kJ/h, p > 0.05 for kJ/3/5 h and %/3.5 h (Weststrate, Dopheide et al. 1990)).

125

Table 5.2 Consumption of meals after an overnight fast and DIT.

Energy Ingested (kJ) and % Reference and DIT DIT % Energy DIT % above Energy from DIT (kJ) Conclusions Measurement Content of the Meal Baseline Macronutrients Higher vs. lower energy intake (a) 45.6 † (9.3) † kJ/h Kinabo and Durnin (b) 45.6 † (10.8) † kJ/h (Kinabo and Durnin (c) 71.2 † (15.5) † kJ/h No significant difference 1990) Paper A †† (a and b) 2520 kJ (d) 68 † (12.4) † kJ/h between meals differing on DIT = open circuit (c and d) 5040 kJ (a) 1.8 † (0.4) † %/h (a) 21 (4.3) †% NS (a and c vs. b and d macronutrient compositions. indirect calorimetry (a and c) 70% CHO, 11% (b) 1.8 † (0.5) † %/h (b) 21 (5.8) † % for kJ/5 h) Significantly higher DIT for using Douglas bag for 5 protein, 19% fat (c) 1.4 † (0.3) † %/h (c) 33 (7.7) † % *** (a and b vs. c and d meals with higher energy h; DIT measured for 10 (b and d) 24% CHO, 11% (d) 1.4 † (0.2) † %/h (d) 33 (6.2) † % for kJ or kcal /5 h) intake compared to lower min, every 10 min for 3 protein and 65% fat NS (all four meals energy intake. collections then every 20 compared for kJ or kcal min for 5 collections /5 h) (a) high and low

VO2 group: less than 10% Hill et al. (Hill, (a) 2092 kJ † (b) high VO max Heymsfield et al. 1984) (b) 4184 kJ † 2 An increase in meal size 23% DIT = indirect (c) 6276 kJ † increased DIT. Unclear if low VO max 19% calorimetry for 3 h; DIT 50% CHO, 16% protein, 34% 2 effect was significant. (c) high VO max measured every 30 min fat 2 41%, low VO2max 26% p value NP Martin et al. (Martin, Normand et al. 2000) (a) 418 kJ, 62% CHO, 34.4% No difference on DIT between (a) 1.1 (0.3) %/h DIT = ventilated hood fat, 3.6% protein (a) 4.5 † (1.4) † kJ/h low energy, moderate fat (b) 1.2 (0.1) %/h indirect calorimetry for 4 (b) 2920 kJ, 67% CHO, 24.6% (b) 35.6 † (2.6) † kJ/h meal, and high energy, low fat p value NP h; DIT measured every fat, 8.4% protein meal. hour 126

Energy Ingested (kJ) and % Reference and DIT DIT % Energy DIT % above Energy from DIT (kJ) Conclusions Measurement Content of the Meal Baseline Macronutrients (a) 1.3 † (0.3) † %/h NP kJ Bennet et al. (Bennett, (b) 1.2 † (0.40) † %/h No significant difference in (a) 55% CHO, 30% fat, 15% Reed et al. 1992) NS (a vs. b for all DIT between high fat meal protein (b) same meal as DIT = ventilated hood subjects (trained and and normal fat meal for the above plus 50 g of fat indirect calorimetry for 6 untrained) only overall subjects (trained vs. (addition of 1881 kJ compared h; continuously included untrained for untrained). to breakfast (a)) this SR) Segal et al. (Segal, Edano et al. 1990) (a) 35% of each man 24 h DIT = open circuit RMR (2889 *‡ kJ) NP (a) 3.2 † (0.8) † %/h (a) 89.3 † (17.6) † kJ/h (a) 11.9% indirect calorimetry for 3 macronutrients (b) 3.3 † (0.4) † %/h N.A. (b) 96.3 † (17.6) † kJ/h (b) 12.9% h; DIT measured for at (b) 3013 kJ †, 55% CHO, 24% NS (a vs. b %/3h) least 6 min periods every protein, 21% fat 30 min Meals varying in macronutrient composition Nagai et al. (Nagai, 3255 (306.5) † kJ, Sakane et al. 2005) (a) 1.3 † (0.4) † %/h (a) 1.7 (0.7) † % (a) 70% CHO, 10% protein, (a) 43.1 † (13.7) † kJ/h DIT was significantly higher DIT = open circuit (b) 1.0 † (0.4) † %/h (b) 1.3 (0.4) † % 20% fat (b) 32.6 † (14.1) † kJ/h in low fat meal compared to indirect calorimetry for * (a vs. b for %/3.5 h * (a vs. b for % (b) 20% CHO, 10% protein, * (a vs. b for kJ/3.5 h) high fat meal. 3.5 h; DIT measured for ECM) AB) 70% fat 6 min every 30 min (a) high fat consumers (a) high fat consumers (a) high fat 27.5 † (28.9) † kJ/h; low 1.3 † (1.4) † %/h; low consumers 10.2 † Blundell et al. (Blundell, The consumption of a high- 2092 kJ fat consumers 25.6 † fat consumers 1.2 † (10.8) † % Cooling et al. 2002) carbohydrate meal was (a) 19.9% CHO, 78.8% fat, (14.5) † kJ/h (0.7) † %/h low fat consumers DIT = ventilated hood significantly associated with 1.3% protein (b) 90.4% CHO, (b) high fat consumers (b) high fat consumers 9.9 † (5.6) † % indirect calorimetry for 3 an increased DIT compared to 1.3% fat, 8.3% protein 38.2 † (26.0) † kJ/h; low 1.7 † (1.2) † %/h; low (b) high fat h continuously a high fat meal fat consumers 35.2 † fat consumers 1.7 † consumers 14.2 † (15.6) † KJ/h (0.7) † %/h (9.7) † %

127

Energy Ingested (kJ) and % Reference and DIT DIT % Energy DIT % above Energy from DIT (kJ) Conclusions Measurement Content of the Meal Baseline Macronutrients * (a vs. b for kJ/day) low fat consumers 13.6 † (6.0) † % Bowden and McMurray (Bowden (a) 2068 kJ, 76% CHO, 5% No significant difference on and McMurray 2000) protein, 23% fat (a) 2.6 † %/h (a) 54.6 † kJ/h (a) 26.2 † % total energy expenditure DIT = open circuit (b) 2093 kJ (fixed amount), (b) 1.3 † %/h (b) 27.8 † kJ/h (b) 13.4 † % between high CHO and high spirometry for 5 h; DIT 21% CHO, 8% protein, 72% p value NP fat meal. measured for 10 min fat periods every 30 min Thyfault et al. (Thyfault, Richmond et al. 2004) (a) 57.8 † (19.1) † kJ/h or DIT = indirect 1.0 (0.3) † kJ/FFM/h or calorimetry with face (a) 3021 (1194.0) † kJ, 79% 0.7 (0.9) † kJ/BM/h (a) 1.9 † (0.6) † %/h mask for 4 h; DIT CHO, 20% protein, 1% fat (b) 49.8 † (21.6) † kJ/h or (b) 1.7 † (0.7) † %/h N.A. measured continuously (b) 2996 (1167.4) † kJ, 37% 0.8 (0.3) † kJ/FFM/h or p value NP with measurements CHO, 18% protein, 45% fat 0.6 (0.9) † kJ/BM/h averaged over 15 min p value NP periods for 1 h then for 30 min periods for the remaining hours Raben et al. (Raben, Agerholm-Larsen et al. 2500 kJ for f, 3000 kJ for m (a) 1.7%/h 2003) (a) 37.2% CHO, 31.8% (b) 1.4%/h Significant difference in DIT DIT = indirect protein, 31.1% fat (a) 45.9 † kJ/h (c) 1.4%/h between the different meal calorimetry with an (b) 23.9% CHO, 11.6% (b) 39.2 † kJ/h ** (a vs. b vs. c vs. also types administered. Protein open-circuit ventilated protein, 64.6% fat (c) 39.2 † kJ/h meal with alcohol had a higher DIT compared to hood system; (c) 65.4% CHO, 12.2% excluded for this SR fat and CHO meals. continuously for 5 h with protein, 23.7% fat for %/5 h) 5 min breaks every h if 128

Energy Ingested (kJ) and % Reference and DIT DIT % Energy DIT % above Energy from DIT (kJ) Conclusions Measurement Content of the Meal Baseline Macronutrients needed Petzke and Klaus (Petzke and Klaus 2008) DIT = indirect (a) 3114 kJ, 35.4% CHO, calorimetry ventilated- (a) 7.8 † (1.0) † kJ/h (a) 1.5 † (0.2) † %/h DIT was significantly higher 3.9% protein, 60.7% fat (a) 3.6 † (0.5) † % hood system for 6 h; 3 × (b) 22.4 † (5.7) † kJ/h (b) 4.3 † (1.1) † %/h in adequate protein meal (b) 3131 kJ, 27.8% CHO, (b) 9.7 † (2.5) † % 30 min measurements *** (a vs. b for kJ/6 h) *** (a vs. b for %/6 h) compared to low protein meal. 11.4% protein, 60.8% fat (first 5–10 min discarded) at 30, 150, and 270 min Overweight Overweight (a) 48.4 † (20.0) † kJ/h (a) 2.6 † (0.2) † %/h (b) 46.3 † (18.8) † kJ/h (b) 2.5 † (1.0) † %/h Riggs et al. (Riggs, Normal weight 1841 kJ † Normal weight White et al. 2007) (a) 43.1 † (19.2) † kJ/h (a) 23% CHO †††; 34% (a) 2.3 † (1.0) † %/h Significantly higher DIT for DIT = indirect (b) 31.0 † (19.2) † kJ/h protein, 43% fat (b) 1.7 † (1.0) † %/h the high protein, high fat meal calorimetry for 3.5 h; Underweight (b) 48% CHO †††; 28% Underweight in normal weight subjects. DIT measured every 30 (a) 20.0 † (16.4) † kJ/h protein, 24% fat (a) 1.1 † (0.9) † %/h min (b) 24.2 † (15.6) † kJ/h (b) 1.3 † (0.9) † %/h ** (a vs. b for normal ** (a vs. b for normal subjects for kcal/min/kg subjects for %/3.5 h) FFM) Clegg et al. (Clegg, Golsorkhi et al. 2013) 1863 kJ (a) 21.9 † (7.9) † kJ/h (a) 1.2 † (0.4) † %/h Pepper sunflower oil had a DIT = ventilated hood 35.5% CHO, protein 19.9%, (b) 29.4 † (8.4) † kJ/h (b) 1.6 † (0.5) † %/h significantly lower DIT than indirect calorimetry for 6 44.6% fat ** (a vs. b for kcal/6 h) ** (a vs. b for %/6 h) pepper MCT oil intervention. h; measured for 15 min every 30 min Kasai et al. (Kasai, (a) 1046 kJ † (a) 8.4 † (4.6) † kJ/h or (a) 0.8 † (0.5) † %/h (a) 2.8 † (1.7) † % Significant increase in DIT

129

Energy Ingested (kJ) and % Reference and DIT DIT % Energy DIT % above Energy from DIT (kJ) Conclusions Measurement Content of the Meal Baseline Macronutrients Nosaka et al. 2002) (b) 1029 kJ † 0.1 † (0.07) † kJ/kg/h (b) 1.7 † (1.1) † %/h (b) 6.3 † (3.9) † % when a liquid meal containing study 1 (c) 1013 kJ † (b) 17.7 † (10.8) † kJ/h or (c) 1.9 † (1.4) † %/h (c) 7.3 † (5.2) † % MCT was consumed DIT = indirect 43% CHO, 21% protein, 36% 0.3 † (0.2) † kJ/kg/h c) ** (a vs. b for %/5 h) compared to a meal with LCT. calorimetry Aeromonitor fat 19.5 † (14.1) † kJ/h or 0.3 ** (a vs. c for %/5 h) AE-300S; DIT measured † (0.2) † kJ/kg/h NS (b vs. c for kJ/5 h) for 6 h at 1 h intervals ** (a vs. b for cal/kg/6 h) ** (a vs. c for cal/kg/6 h) NS (b vs. c for cal/kg/6 h) (a) 8.2 † (6.4) † kJ/h or 0.2 † (0.1) † kJ/kg/h Kasai et al. (Kasai, (a) 1059 kJ † (b) 14.0 † (5.7) † kJ/h or (a) 0.8 † (0.6) † %h Nosaka et al. 2002) (b) 1042 kJ † 0.3 † (0.1) † kJ/kg/h c) (b) 1.3 † (0.5) † %h (a) 3.9 † (3.0) † % Significant increase in DIT in study 2 (c) 1020 kJ † 9.8 † (8.2) † kJ/h or 0.2 † (c) 1.0 † (0.8) † %h (b) 6.8 †(2.7) † % meals containing mayonnaise DIT = indirect (d) 1004 kJ † (0.2) † kJ/kg/h d) 20.3 † (d) 2.0 † (1.6) † %h (c) 4.7 † (3.9) † % or margarine with MCT calorimetry Aeromonitor 50% CHO, 10% protein, 40% (15.7) † kJ/h or 0.4 † (0.3) * (a vs. b for %/5 h) (d) 10.5 † (7.9) † % compared to LCT. AE-300S; DIT measured fat † kJ/kg/h * (c vs. d for %/5 h) for 6 h at 1 h intervals * (a vs. b for cal/kg/6 h) * (c vs. d for cal/kg/6 h) Mean (95% CI) a) 2655 Mean (95% CI) (2510–2799) kJ, 36.4 (35.9– Mean (95% CI) Casas-Agustench et al. (a) 12.3 (9.7– 36.7)% CHO, 11.7 (11.4– (a) 37.2 † (29.5–44.8) † (Casas-Agustench, 14.9)% 11.9)% protein, 51.9 (95% CI kJ/h Mean (95% CI): Lopez-Uriarte et al. (b) 11.8(9.7– 51.7–52.1)% fat (b) 36.8 † (30.5–43.0) † (a) 1.4 † (1.1–1.7) † %/h DIT was significantly higher 2009) 13.9)% (b) 2608 (2428–2788) kJ, 37 kJ/h (b) 1.4 † (1.2–1.7) † in PUFA and MUFA meals DIT = open circuit (c) 9.6 (7.7–11.4)% (95% CI 36.6–37.4)% CHO, (c) 30.0 † (24.2–35.8) † %/h compared to SFA meal. indirect calorimetry with * (a vs. c and b vs. 11.3 (95% CI 10.6–11.9)% kJ/h (c) 1.2 † (0.9–1.4) † %/h a canopy system for 5 h c for %AB) protein, 51.7 (95% CI 51.3– * (a vs. b vs. c and a vs. c continuously * (a vs. b vs. c for 52.0)% fat for kJ/5 h) % AB) (c) 2599 (2421–2278) kJ, 37.1 130

Energy Ingested (kJ) and % Reference and DIT DIT % Energy DIT % above Energy from DIT (kJ) Conclusions Measurement Content of the Meal Baseline Macronutrients (95% CI 36.4–37.7)% CHO, 11.2 (95% CI 10.8–11.6)% protein, 51.7 (95% CI 51.2– 52.1)% fat Piers et al. (Piers, Walker et al. 2002) DIT = open circuit 2500 † kJ (a) 29.6 † (10) † kJ/h (a) 1.2 † (0.4) † %/h No significant differences in (a) 9.5 † (3.2) † % ventilated hood canopy 42% CHO, 15% of energy (b) 28.4 † (10) † kJ/h (b) 1.1 † (0.4) † %/h DIT between SFA and MUFA (b) 9.3 † (3.2) † % system for 5 h; DIT from protein, 43% fat NS (a vs. b for kJ/5 h) NS (a vs. b for %/5 h) meals. measured for 30 min periods each hour (a) 61.8 † (15.2) † kJ/h (a) 1.3 † (0.3) † %/h Bendixen et al. (b) 72.8 † (19.0) † kJ/h (b) 1.5 † (0.3) † %/h (Bendixen, Flint et al. (c) 69.2 † (11.4) † kJ/h (c) 1.4 † (0.2) † %/h 2002) (d) 65 † (13.9) † kJ/h (d) 1.4 † (0.3) † %/h (a) 21.3 † (5.2) † % DIT was significantly higher DIT = indirect 4698 (550.2) † kJ, 34% CHO, * (a vs. b vs. c vs. d for * (a vs. b vs. c vs. d for (b) 25.1 † (6.6) † % in the three modified fat meals calorimetry with open 6% protein, 60% fat kJ/5 h) %/5 h) (c) 23.7 † (3.9) † % compared to the conventional circuit, ventilated hood ** (a vs. b for kJ/5 h) ** (a vs. b for %/5 h) (d) 22.4 † (4.8) † % fat meal. for 5 h continuously with NS (All other pairwise NS (all other pairwise 10 min breaks every hour comparison apart from a comparison apart from vs. b) a vs. b) Processed vs. unprocessed meals Barr and Wright (Barr 2520 or 3360 kJ and Wright 2010) (a) 39% fat, 40% CHO, 20% (a) 99.4 † (40.7) † kJ/h (a) 3.4 † (1.7) † %/h DIT = indirect protein Whole food meal showed a (b) 63.9 † (35.6) † kJ/h (b) 2.2 † (1.4) † %/h calorimetry using (b) ½ sandwich: 33% fat, 49% significant higher DIT *** (a vs. b kJ/5.8 and ** (a vs. b %/5.8 and spirometer and gas bags CHO, 15% protein or 2 compared to processed meal. 4.8 h) %/4.8 h) (a) 5.8 (0.11) h (b) 4.8 sandwiches: 33% fat, 50% (0.23) h CHO, 15% protein

131

Energy Ingested (kJ) and % Reference and DIT DIT % Energy DIT % above Energy from DIT (kJ) Conclusions Measurement Content of the Meal Baseline Macronutrients 2 min measurement by spirometer followed by 10 s exhalation for 5 or 6 breaths into a gas bag every hour Bolus vs. smaller frequent meals Kinabo and Durnin (Kinabo and Durnin Group A: 1990) - Paper B (a and c) 5040 kJ (a) 62.8 † (13.2) † kJ/h Group A Group A DIT = open circuit (b and d) 2520 × 2 kJ (b) 63.5 † (11.7) † kJ/h (a) 1.3 † (0.3) † %/h (a) 28 (6.9) † % No significant difference on indirect calorimetry (a and b) 70% CHO, 11% Group B: (b) 1.3 † (0.2) † %/h (b) 31 (5.0) † % DIT between meals consumed using Douglas bag protein, 19% fat (c) 59.3 † (11.5) † kJ/h Group B Group B as bolus vs. two smaller technique for 6 h; DIT (c and d) 24% CHO, 11% (d) 56.7 † (8.0) † kJ/h (c) 1.2 † (0.25) † %/h (c) 27 (7.8) † % frequent meals. measured for 10 min, protein, 65% fat NS (a vs. b and c vs. d (d) 1.1 † (0.2) † %/h (d) 28 (5.1) † % every 10 min for the first for kJ/6 h) 90 min and every 20 min for the last 90 min. Vaz et al. (Vaz, Turner DIT was lower in the small (a) 3150 † et al. 1995) (a) 71 † (31.5) † kJ/h frequent feeding regime (b) 1050 × 3 kJ † (a) 2.3 † (1.0) † %/h (a) 22.2%/h DIT = indirect (b) 52.3 † (15.3) † kJ/h compared to one bolus meal 53.3% CHO, 14.7% protein, (b) 1.7 † (0.49) † %/h (b) NP calorimetry for 2 h; DIT NS (a vs. b for kJ/2 h) event – but not significantly 32% fat measured every 30 min different. Allirot et al. (Allirot, DIT was significantly higher (a) 2823.4 kJ † Saulais et al. 2013) when the meal consumed as (b) total 2823.4 † divided in (a) 1.6 † (0.7) † %/h DIT = indirect (a) 43.8 † (18.4) † kJ/h one bolus event compared to 705.8 † kJ meals (b) 1.2 † (0.6) † %/h calorimetry for 4 h; DIT (b) 33.2 † (15.5) † kJ/h four smaller isocaloric meals 54.2 † % CHO, 6.3 † % * (a vs. b for %/4 h) measured for 30 min ingested over time in the Protein, 36.7 † % Fat periods morning. Tai et al. (Tai, Castillo et (a) one meal of 3138 kJ (a) 48.2 † (16.93) † kJ/h (a) 1.5 † (0.5) † %/h (a) 20.6 † (7.2) † % DIT was significantly higher

132

Energy Ingested (kJ) and % Reference and DIT DIT % Energy DIT % above Energy from DIT (kJ) Conclusions Measurement Content of the Meal Baseline Macronutrients al. 1991) (b) 6 meals of 523 kJ (b) 34.9 † (12.3) † kJ/h (b) 1.1 † (0.4) † %/h (b) 14.8 † (5.2) †% when the meal was consumed DIT = indirect 54.5% CHO, 14.0% protein, * (a vs. b for kJ/5 h) as a one bolus event compared calorimetry for 5 h; DIT 31.5% fat to six smaller isocaloric meals measured every 30 min ingested over time in the morning. Fast vs. slow eating patterns (a) 19.5 † (142.2) † Hamada et al. (Hamada, kJ/kg/h Kashima et al. 2014) Slowing eating was associated 1255.2 † kJ, 42% CHO, 8% (b) 502.1 † (234.4) † DIT = gas analyzer AE- with a significant increase in protein, 50% fat kJ/kg/h 310S for 1.5 h DIT compared to rapid eating. * (a vs. b for kcal/kg/90 continuously min) Toyama et al. (Toyama, Zhao et al. 2015) DIT = open-circuit (a) 31.6 † (15) † kJ/kg/h indirect calorimetry for 3 There was no significant 1464 kJ, 61.3% CHO, 16.4% (b) 41.9 † (14.6) † kJ/kg/h (a) 6.8 (4.8)% h; first hour difference in DIT between fast protein, 22.3% fat NS (a vs. b for (b) 8.5 (4.2)% continuously, second and eating and regular eating. kcal/kg/min) third hours measured for 15 min every 30 min intervals Palatable vs. unpalatable meals Sawaya et al. (Sawaya, (a) old 26.4 † (11.3) (a) old 37.0 † (15.9) † Fuss et al. 2001) (a) old 1.3 † (0.5) † %/h † % kJ/h DIT = ventilated hood young 1.2 † (0.5) † %/h young 21.7 † (8.5) † DIT did not significantly 2930 kJ †, 65% CHO, 12% young 34.7 † (13.6) † kJ/h indirect calorimetry, for (b) old 1.6 † (0.6) † %/h % differ between palatable and protein, 23% fat (b) old 46.4 † (18.4) † 6 h; DIT measured for 10 young 1.4 † (0.6) † %/h (b) old 34.6 † (13.7) unpalatable meals. kJ/h min with 5 min breaks NS (a vs. b %/3 h) † % young 39.5 † (17.0) † kJ/h for the 6 h young 24.8 † (10.6)

133

Energy Ingested (kJ) and % Reference and DIT DIT % Energy DIT % above Energy from DIT (kJ) Conclusions Measurement Content of the Meal Baseline Macronutrients † % Weststrate et al. (Weststrate, Dopheide et (a) 2.4 † (0.7) † %/h There was not a significant (a) 47.3 † (14.2) † kJ/h al. 1990) study 1 2000 kJ †, NP % energy from (b) 2.6 † (0.7) † %/h difference in DIT between (b) 52.9 † (13.3) † kJ/h DIT = ventilated hood macronutrient NS (a vs. b for %/3.5 palatable and unpalatable NS (a vs. b for kJ/3.5 h) indirect calorimetry for h) meals. 3.5 h continuously Data are described in mean (SD) unless otherwise described. a, b, c, etc. = these letters describe the types of meal interventions provided as illustrated in Table 5.1. † These data (mean and/or SD and/or 95% CI) were calculated or converted for one or more of these possible calculations or conversions (DIT % ECM calculated from DIT kJ, DIT % above baseline RMR calculated from DIT kJ, DIT kJ calculated from DIT % ECM, macronutrient % ECM calculated from grams, kcal converted to kJ, MJ converted to kJ, SE converted to SD, DIT % ECM or KJ or % above baseline RMR converted for unit of time, formulas described either in methodology or Supplementary Materials Table S1). †† Kinabo et al. [31]—paper A is also part of results section: “meals varying in macronutrient composition”. ††† calculated % CHO = 100 - (% energy from FAT + % of energy from protein). *‡ calculated 35% 24 h RMR = 35 * 8253.6 kJ (24 h RMR)/100. It is an average value. * p ≤ 0.05. ** p ≤ 0.01. *** p ≤ 0.001. BM = Body Mass. DIT = Diet Induced Thermogenesis. CHO = Carbohydrate. FFM = Fat Free Mass. LCT = Long Chain Triacylglycerol. MCT = Medium Chain Triacylglycerol. MUFA = Mono Unsaturated Fatty Acids. NS = Not Significant. NP = Not Provided. PUFA = Poly Unsaturated Fatty Acids. SFA = Saturated Fatty Acids. VO2 = Rate of Oxygen consumption.

134

5.6 Discussion

This review investigated the effects of meals consumed after an overnight fast that differed in energy content or macronutrient composition on DIT, as well as the effects of consuming the same meal as a single event or multiple small meals or snacks. Studies comparing the effects of differing energy intakes supported a conclusion that a higher energy intake resulted in a higher DIT. This finding was further supported by two meta- regressions (one unadjusted and one adjusted for confounding factors), which found that for every 100 kJ increase in energy intake, DIT increased by 1.1 (unadjusted) or 1.2 (adjusted) kJ/h. A number of studies compared the effects of meals differing in macronutrient composition. One study found that a meal high in protein resulted in a higher DIT than meals high in CHO or fat, and a number of studies suggested that a meal high in CHO resulted in a higher DIT than a meal high in fat. Medium chain triglyceride meals produced a higher DIT than long chain triglycerides, the effects of mono- and polyunsaturated fats compared to saturated fats were unclear, fat structure (e.g., sunflower oil compared to a chemically structured fat) influenced DIT, and the fat content of a meal had inconsistent effects on DIT. The DIT of meals consumed as two or three small meals did not differ to the DIT of the same meal consumed as a single meal, whereas meals consumed as four to six small meals had a lower DIT compared to the same meals consumed as a single meal. Together these findings indicate that meals consumed after an overnight fast result in a DIT and the magnitude of this DIT is influenced by the energy content, the macronutrient composition, and the eating pattern of the meal.

Five studies investigated the effects of consuming different energy intakes on DIT as a primary outcome (Hill, Heymsfield et al. 1984, Kinabo and Durnin 1990, Segal, Edano et al. 1990, Bennett, Reed et al. 1992, Martin, Normand et al. 2000). The study with the largest sample size found a significant increase in DIT with a higher energy intake (Kinabo and Durnin 1990). Two studies with much smaller sample sizes reported trends of a higher DIT with a higher energy intake (Hill, Heymsfield et al. 1984, Martin, Normand et al. 2000). Two other studies (Segal, Edano et al. 1990, Bennett, Reed et al. 1992) found no effect on DIT but the small sample sizes (eight and 11) could have impacted these findings. Additionally, one of these two studies provided little difference in energy intake between the two meals consumed (Segal, Edano et al. 1990). The meta- regressions subsequently undertaken to examine the effect of energy intake on DIT

135

across a much larger range of studies clearly support a conclusion that the energy content of meals consumed after an overnight fast influences DIT. Both the unadjusted meta- regression and the one adjusted for four confounding factors (percentage of males, age, BMI, and duration of DIT measurement) found similar significant relationships between a higher energy intake and a higher DIT. The magnitude of the increase in DIT was very small (1.1 or 1.2 kJ/h increase with each 100kJ increase in energy intake), and whether this increase is clinically meaningful may depend on the magnitude of the energy content of the meal. These findings are consistent with the conclusion of Westerterp that energy intake is a predictor of DIT (Westerterp 2004).

A number of studies compared the effects of meals consumed after an overnight fast differing in macronutrient composition. Five studies compared high fat vs. high CHO meals, and four of them found a higher thermogenic effect after the consumption of a high CHO meal compared to a high fat or moderate fat meal. The two studies that found significant effects had sample sizes of 24 males (Blundell, Cooling et al. 2002) and 14 males (Nagai, Sakane et al. 2005). The other two studies, which were conducted with smaller sample sizes (12 males (Thyfault, Richmond et al. 2004) and six females (Bowden and McMurray 2000)), showed trends for a higher thermogenic effect of high CHO meals, but they did not provide p values and this limited their conclusions (Bowden and McMurray 2000, Thyfault, Richmond et al. 2004). Furthermore, Thyfault et al. (Thyfault, Richmond et al. 2004) compared a high CHO meal with a moderate fat, moderate CHO meal (45% fat) and therefore, the moderate CHO content could have confounded the findings. The one study that reported no significant difference for this high fat vs. high CHO comparison was conducted in 16 females (Kinabo and Durnin 1990). Significant effects were found in the two studies conducted in males, whereas no significant effect was observed in the study in females, suggesting that males and females may respond differently following the consumption of CHO and fat meals. The differences in DIT between males and females may result from hormonal and/or body composition differences. Again, more research is needed to clarify these observations.

Two studies investigated the effects of protein on DIT; one found a significant increase in DIT after a high protein meal compared to high CHO or high fat meals in 19 participants of mixed gender (Raben, Agerholm-Larsen et al. 2003). The other study found a significant increase in DIT when a high protein meal was consumed compared to

136

an adequate protein meal even though this was a small study of only six females (Petzke and Klaus 2008). Both studies suggest a high thermogenic effect of protein, however due to the limited numbers of studies, more research is needed to further investigate the thermogenic effect of protein in meals in both males and females, as gender was suggested to have an effect on the studies comparing high CHO vs. high fat meals (Kinabo and Durnin 1990, Blundell, Cooling et al. 2002, Nagai, Sakane et al. 2005).

A review by Tappy et al. (Tappy 1996). supports the higher thermogenic effect of protein compared to fat; this review reported DIT to be 0%–3% for fat, 5%–10% for CHO, and 20%–30% for proteins (Tappy 1996). The different thermogenic effects of macronutrients are further reinforced by two studies comparing two diet interventions (Westerterp, Wilson et al. 1999, Lejeune, Westerterp et al. 2006). One study compared a high protein diet to an adequate protein diet for four days in 12 women and found a significantly higher DIT with the high protein diet (high protein diet 0.91 (0.25) MJ/d or 10.1 (2.7) % energy intake vs. high fat diet 0.69 (0.24) MJ/d or 7.6 (2.5) % energy intake, p < 0.05) (Lejeune, Westerterp et al. 2006). Another study compared a high protein diet to a high fat diet for 36 h in eight women and found a significantly higher 24 h DIT with the high protein diet intervention as opposed to the high fat diet (high protein diet 1295 (240) kJ/day or 14.6 (2.9) % energy intake vs. high fat diet 931 (315) kJ/day or 10.5 (3.8) % energy intake, p = 0.02) (Westerterp, Wilson et al. 1999). Therefore, the findings of this SR regarding the role of meals differing in fat, CHO and protein composition on DIT are consistent with other studies investigating the effect of diets varying on macronutrients compositions on DIT. With regard to possible mechanisms of action of the higher thermogenic effect of protein, Westerterp-Plantenga (Westerterp- Plantenga 2008) suggests that a higher protein diet may increase protein synthesis, which has a high energy cost, or if protein is oxidized the energy cost is higher than fat or CHO, and energy cost also varies with amino acid composition (Tessari, Kiwanuka et al. 2003, Westerterp-Plantenga 2008).

One study compared higher vs. lower levels of fat intake and found a significant increase in DIT following the consumption of a high fat meal compared to a low fat meal in female participants. This significant effect was only found in the 12 normal weight participants (Riggs, White et al. 2007) and not in the six overweight or three underweight participants, however the small numbers of participants in the overweight and

137

underweight groups would make finding effects in these groups difficult. Also, the different protein levels of the two meals could have impacted the findings of this study.

A number of studies compared the effects of different types of fats on DIT whereas no studies were identified that compared the effects of different types of proteins or carbohydrates on DIT. A significantly higher thermogenic effect was found when meals containing MCTs were compared to those containing LCTs; this finding was consistent in the three studies included (Kasai, Nosaka et al. 2002, Clegg, Golsorkhi et al. 2013). The meta-analysis combining these three papers confirmed this significant increase in thermogenesis following the consumption of a meal with MCTs compared to LCTs (p < 0.005). Two hypotheses have been proposed regarding possible mechanisms by which the higher thermogenic effect of MCTs vs. LCTs might be achieved. One suggests an important role for the liver. MCTs are transported directly to the liver by the portal vein whereas LCTs are transported by the lymphatic system to peripheral tissues (adipose tissue and muscle) (Bach and Babayan 1982). Also, LCTs need to bind to carnitine in order to pass through the mitochondrial membrane of the liver where B-oxidation occurs (Bach and Babayan 1982), whereas MCTs do not (Odle 1997). Therefore, MCTs being directly transported to the liver and that are easily able to pass through the mitochondrial membrane may be responsible for their higher DIT (Berry, Clark et al. 1985, Seaton, Welle et al. 1986).The second hypothesis suggests a role for the sympathetic nervous system. Dullo et al. (Dulloo, Fathi et al. 1996) found increased noradrenaline levels after MCT consumption and the authors suggested that sympathetic nervous system stimulation could therefore be responsible for the increase in energy expenditure of MCTs. Kasai et al. (Kasai, Nosaka et al. 2002) has indicated that more research is needed to support this proposed mechanism.

Inconsistent results were found between two studies that compared different saturation of fat on DIT. The study which found a significantly higher DIT by PUFA and MUFA meals compared to a SFA meal (Casas-Agustench, Lopez-Uriarte et al. 2009) had a much larger sample size (29 participants) than the study which did not find a significant difference between MUFA and SFA meals (14 participants) (Piers, Walker et al. 2002). More research is needed to clarify these findings. The effects of fat structure were also investigated. One study (Bendixen, Flint et al. 2002) found that fat structure, specifically meals containing a chemically structured fat, a lipase-structured fat, and physically

138

mixed fat was associated with a significantly higher DIT compared to a conventional type of fat, and the highest DIT was associated with the consumption of the chemically structured fat. These findings are clearly preliminary and more studies are needed to confirm these observations.

One study examined the thermogenic effect of a less processed (e.g., whole grain bread) meal compared to a more processed meal (Barr and Wright 2010) after an overnight fast and found a significantly higher DIT following consumption of the less processed meal. Although energy intake was consistent between trials in the same participant, the sample consisted of males and females, and there were two different caloric options within the study. It is unclear whether the choice of caloric options was controlled for in the analysis or whether there was a gender difference in the choice of caloric option. Furthermore, the macronutrient composition of these two meals differed and could have impacted the findings. Therefore, there is a need for further research comparing the effects of consuming more processed vs. less processed foods with the macronutrient content of the meals closely matched.

Four studies investigated the effects of consuming the same amount of calories and meal composition as a bolus event compared to a number of smaller meals during the morning. Two studies found a significant increase in DIT when the meal was consumed as one meal (bolus) instead of four (Allirot, Saulais et al. 2013) or six (Tai, Castillo et al. 1991) smaller frequent meals. Two other studies did not find a significant difference in DIT between bolus and two (Kinabo and Durnin 1990) or three (Vaz, Turner et al. 1995) smaller frequent meals regardless of macronutrient compositions (Kinabo and Durnin 1990, Vaz, Turner et al. 1995). The studies that did not find a significant difference provided less frequent meals for the snacking comparison, resulting in fewer meals with higher energy intake. The meta-analysis conducted on these four studies found that DIT was significantly higher when the meal was consumed as one bolus event. Together these results suggest that fewer larger meals result in a higher DIT than more frequent smaller meals.

Two studies (Hamada, Kashima et al. 2014, Toyama, Zhao et al. 2015) found a higher DIT following a meal eaten slowly compared to a meal eaten quickly, although these findings were only significant for one study (Hamada, Kashima et al. 2014). The study that found a significant effect was conducted in ten males whereas the study that did not

139

find a significant difference was conducted in nine females (Toyama, Zhao et al. 2015). It is possible that gender is a factor influencing these results. Furthermore, these two studies suggest that the time that is spent on chewing the food may influence the magnitude of DIT; however more research is needed to clarify this observation and to compare the effects between males and females. Only two studies (Weststrate, Dopheide et al. 1990, Sawaya, Fuss et al. 2001) have compared the effects of consuming a palatable versus an unpalatable meal on DIT in 19 males (Sawaya, Fuss et al. 2001) or 12 participants (6 males and 6 females) (Weststrate, Dopheide et al. 1990). Neither study found any significant differences in DIT. Although these findings suggest that palatability has no effect on DIT, the small number of studies limits the ability to draw any firm conclusions on this topic. Finally, it is important to note that no studies identified for this review have investigated the effects of differing micronutrients on DIT and this may potentially be another factor to influence DIT, which therefore warrants further investigation.

5.6.1 Strengths of This SR

This SR, including meta-analysis and meta-regression, is the first one to be conducted to investigate the effects of energy intake, macronutrient composition and eating events on DIT. This SR was able to identify and summarize the highest level of evidence available in this area and summaries where more research is needed in this field. The studies included were screened for quality/risk of bias and none of the studies had negative quality. Furthermore, the meta-regression was able to quantify the short-term effects of differing energy intakes after an overnight fast on DIT. In addition, the meta-analyses were able to quantify the influence of MCT vs. LCT and the role of consuming bolus eating events vs. smaller frequent meals in the morning on DIT. Considering the lack of evidence base regarding the role of meals consumed after an overnight fast on obesity prevention, this SR was able to provide evidence of the short-term effect of consuming different types of meals on DIT, which in the long term could play a significant role in obesity.

5.6.2 Limitations

There are a number of limitations regarding this SR. First, the included studies were very heterogeneous, differing in their research questions and the types of meals served after an overnight fast (as summarized in Tables 1 and 2). This heterogeneity limited the meta-

140

analyses that could be conducted and the confidence with which conclusions could be drawn. Furthermore, only meta-analyses with a minimum of three studies were included in this SR.

Secondly, the majority of studies investigating DIT are conducted after an overnight fast even if their primary aim is not to investigate the effect of breakfast per se. These studies met the criteria to be included because they administered meals/snacks (even if not a typical breakfast meal) after an overnight fast (definition of breakfast). Lacking are studies of the effects of meals representative of breakfast in specific cultures. Whether these representative breakfast meals would have a different effect on DIT is unclear.

Thirdly, the studies varied in the units of measurement used to report DIT, including kcal or kJ, % ECM, or % above baseline (AB), and this limited the direct comparability of the findings between the studies. In order to address this issue, whenever possible, DIT was converted into units of measurements that allowed direct comparisons (e.g., DIT kJ converted to % ECM or % AB). Also, the studies measured DIT for different lengths of time, ranging from one and a half to six hours, and the length of time that the DIT is measured affects the magnitude of DIT detected (Weststrate 1993, Reed and Hill 1996, Ruddick-Collins, King et al. 2013). In order to account for this limitation, the data were transformed into kJ/h allowing the results to be compared among studies. Furthermore, this confounding factor was adjusted for in the meta-regression model.

There is conflicting evidence about the length of time that DIT needs to be measured to provide accurate results. Two papers have recommended measuring DIT for at least three hours. One conducted a study in ten participants and concluded that 3 h DIT measurements are sufficient as 76% of DIT is obtained during this period (Ruddick- Collins, King et al. 2013). The other paper (Weststrate 1993) analyzed data from six studies with a total of 103 subjects and also recommended measuring DIT for three hours, as they found that the majority of the DIT was measured by three and a half hours when either high or low energy intakes (ranging from 1.3 MJ to 2.6 MJ) were consumed and in both men and women (Weststrate 1993). Another study (Reed and Hill 1996) conducted in 131 participants recommended measuring DIT up to six hours, as 3 h measurements underestimated the DIT response by 40% (Reed and Hill 1996). Therefore, it is possible that studies included in this review that measured durations of DIT shorter than six hours may have underestimated DIT. In addition, DIT was

141

measured differently in the studies, as some measured DIT for short durations at regular intervals (interval ranges also differed among studies) whereas other studies measured it continuously, which may have affected the magnitude of DIT detected. Specifically, Piers et al. (Piers, Soares et al. 1992) found a significantly (p < 0.01) lower DIT when it was measured at intervals compared to continuously in the same five subjects (Piers, Soares et al. 1992). Most of the studies included in this review measured DIT at intervals and therefore there is a risk that the DIT was underestimated.

The small sample size of the majority of the studies included is another limitation of this SR. Sample sizes ranged from four to 29 participants, which limits the statistical power and capacity to find small but possibly important effects, as well as limiting the generalizability of the findings. This makes similarity of study designs that can be included in meta-analyses and meta-regressions even more important.

The interventions provided in the studies included in the meta-regression on energy intake had substantial variations; for example, they had different macronutrient compositions, and in some instances the meals were administered differently (such as meals consumed as a bolus vs. smaller frequent meals). Furthermore, the studies included in the meta-analyses were few, and even then included some heterogeneity. For example, the studies included in the MCT and LCT meta-analysis had meals with different levels of LCT or MCT. Also, for the bolus vs. smaller frequent meals meta- analysis, the energy intakes differed between the studies included. Furthermore, the smaller frequent meals arm differed in the number of smaller meals provided between the studies (two, three, four and six). All these factors were not possible to be adjusted for and could have impacted the findings of the meta-regressions and meta-analyses.

Furthermore, the meta-analyses were conducted by considering the two groups for comparison as two different groups of people, even though in reality it was the same group of people repeated in cross-over designs. For this reason, the analyses conducted by RevMan used a more conservative approach, because it is more difficult to find statistical significance if two comparison groups consist of different people. However, the heterogeneity between studies is also expected to be underestimated for the same reason. Therefore, for the two meta-analyses conducted in this SR (MCT vs. LCT and meal vs. snacking), the heterogeneities are expected to be larger than the number provided and the p values are also expected to be even more significant (smaller) that the

142

ones provided (in both cases they are already highly significant). The CIs are also believed to be smaller than the ones provided.

Also, in the findings of this SR, meta-regression and meta-analyses were based primarily on studies rated neutral for quality, thereby limiting the confidence with which conclusions could be drawn.

Finally, the included studies were almost entirely short-term (single meal interventions; only one study investigated the effects of 2 weeks of the same meal daily prior to measurement following a single meal) (Martin, Normand et al. 2000), therefore, whether the effects observed reflect, at least in part, the novelty of unfamiliar meals consumed after an overnight fast on DIT is unclear.

It was not possible to draw any conclusions about the effect of a routine breakfast on DIT. Therefore larger, longer-term experimental studies are needed to draw conclusions about these topics. Specifically, there is the need to investigate the long-term effects of regular consumption of low energy intake vs. high energy intake meals and meals varying in macronutrient composition on DIT, and whether these factors ultimately affect total daily energy intake and DIT or the weight of participants. There is also a need to compare the effects of these influences on DIT between regular breakfast eaters and skippers, as this was found to be an important factor to consider by a previous trial conducted in this area (Schlundt, Hill et al. 1992).

5.6.3 Recommendations

Most of the studies included in this SR were rated neutral instead of positive quality because they had not provided enough or clear information about the selection of participants, recruitment, or inclusion and exclusion criteria. Therefore, it is recommended for future studies in this area to provide more information about recruitment, inclusion and exclusion criteria of participants, and any risk of biases in selection of the subjects.

This SR found heterogeneity between studies regarding the length of DIT measured. Therefore, this SR has identified the need for more clarity on how long DIT should be measured in order to provide accurate results, and in order to achieve more homogenous study designs in this field.

143

This SR also reports the heterogeneous ways DIT is reported between studies (kJ or kcal, % of ECM, or % above baseline RMR). It is recommended that future studies provide DIT in all of these three units of measure (kJ or Kcal, % ECM, and % above baseline RMR) in order to allow easier comparison with other studies conducted in this field and future meta-analyses.

It is also recommended for future studies to provide the data regarding baseline RMR and energy content provided by the meal, as this will allow a more comprehensive picture of the study design and results.

The majority of the studies had very small sample sizes. It is therefore recommended that future studies either increase the sample size in order to improve the statistical power of the studies, or provide evidence that the sample size used was adequate to detect an effect.

5.7 Conclusions

This systematic review has consolidated the current evidence regarding the effects of variations in energy intake, macronutrient composition, and the pattern of meals consumption after an overnight fast on DIT. It has also identified a substantial number of questions that remain to be answered, and the high level of uncertainty around many of the influences on the effects of meals on DIT. There is an enormous scope for future high quality studies in this field of research. Consensus on the duration of DIT measurement and larger sample sizes are just two ways in which research in this area could be improved. Comparisons of the effects of manipulations of meals consumed after an overnight fast on DIT in males and females, different age groups, and those who are healthy or have a range of obesity-related health conditions would also be informative.

5.8 Supplementary Materials

Table S1: (Appendix 7 of this thesis): Formulas used to calculate participants’ characteristics, macronutrients compositions, or DIT

5.9 Acknowledgments

This project was supported by the University of Newcastle Postgraduate Research Scholarship Central (UNRSC50:50). The authors wish to acknowledge the contribution of

144

three research assistants: Allison Brandt, Loren Stroud, and Kelly Rice who assisted with the systematic review process and data extraction. The authors wish to thank Debbie Booth, faculty librarian, University Library, the University of Newcastle, for providing assistance with the literature search.

5.10 Author contributions

All authors (A.Q.; R.C.; A.P., L.M.-W.) have made substantial contributions to all of the following: conception and design of the systematic review, systematic review process, acquisition of data, analysis and interpretation of data, drafting the paper, critically revising the review for important intellectual content, and final approval of the version to be submitted.

5.11 Conflict of interests

The authors declare no conflict of interest.

145

Overview for Chapter 6 and 7

In Chapters 3 and 4, we reported that consumption of certain breakfast cereals, including muesli and/or All Bran, were significantly associated with decreased incidences of diabetes and obesity among mid-aged women from the ALSWH. This demonstrated that breakfast cereal consumption may be protective against chronic disease development in large prospective cohort studies. Chapter 5 was a Systematic Review, meta-regression and meta- analysis of randomised cross over designs looking at the effect of a breakfast meal (ie after an overnight fast) on DIT. We found that the macronutrient and energy content of the meal had a significant impact on the absolute value of DIT, and we theorise that in the long term this may play a role in decreasing obesity risk. It was also determined that consuming a larger breakfast meal (as a bolus event) had significantly higher impacts on DIT than if the same foods were consumed as more frequent smaller meals during the morning. The heterogeneous nature of the breakfast meals found among the literature while completing the SR, lead us to determine that more research was needed on what actually constitutes breakfast.

Therefore, we designed a multi-centre cross-sectional study to examine breakfast consumption in detail among a young sample of Australian male. The ‘Typical Aussie Bloke’ study investigated associations between habitual breakfast consumption and metabolic, anthropometric parameters and lifestyle characteristics. We decided to focus on men as, while some limited research existed on breakfast consumption in females and mixed gender studies in Australia, there was an absolute dearth of information on men. Additionally, some international work had reported differences in breakfast consumption between genders. For example, a large cross sectional study conducted in America reported that a higher proportion of males were skipping breakfast in America (Deshmukh-Taskar, Nicklas et al. 2013). In support of this cross-sectional study examining what constitutes breakfast in Australia, was the fact that the latest evidence available regarding habitual breakfast habits was collected more than 20 years ago in the 1995 National Nutrition Survey.

A further benefit of the ‘Typical Aussie Bloke’ Study was that there are no other recent studies among young healthy Australian men with detailed anthropometric measurements (BMI and circumferences), metabolic measurements (RMR, body composition values, blood lipids, blood pressure) and breakfast consumption data.

Therefore the ‘Typical Aussie Bloke’ Study, described in Chapters 6 and 7, aimed to increase our understanding with respect to breakfast consumption habits, what constitutes a typical

146

Australian breakfast amongst a sample of young Australian males and the association of these with anthropometric and metabolic parameters, taking into account physical activity levels and socio-demographic differences.’

147

Chapter 6: Breakfast Consumption Habits of Young Australian Men from the “Typical Aussie Bloke” study. This paper is in the process to be submitted to the European Journal of Nutrition

Quatela, A; Patterson, A; Callister, R; MacDonald-Wicks, L. (2017) Breakfast consumption habits of young Australian men from the ‘Typical Aussie Bloke’ study. In the process to be submitted to European Journal of Nutrition.

The work presented in the manuscript was presented at the Asia Pacific Conference on Clinical Nutrition in Adelaide 2017 (poster presentation by Quatela, A, on 26-29 November 2017; Appendix 8).

The work presented in the manuscript was completed in collaboration with the co-authors (Appendix 9).

Quatela, A.1; Patterson, A.J.1,3,4; Callister, R.2,3,4; MacDonald-Wicks, L.K.1,3,4*.

1 Discipline of Nutrition and Dietetics, School of Health Sciences, The University of Newcastle; Callaghan, NSW 2308, Australia

2; School of Biomedical Sciences and Pharmacy, The University of Newcastle, Callaghan, NSW 2308, Australia

3 Priority Research Centre for Physical Activity and Nutrition; The University of Newcastle, Callaghan, NSW 2308, Australia

4 Hunter Medical Research Institute, New Lambton, NSW 2305, Australia

* Author to whom correspondence should be addressed.

6.1 Abstract Objective: Breakfast is often described as “the most important meal of the day”. Limited evidence is available regarding the foods/beverages currently constituting a typical breakfast. This study investigated current breakfast habits of young Australian men.

Methods: Men aged 18-44y were recruited from metropolitan and regional NSW Australia and completed an online survey.

Results: 112 men participated. The majority (83.5%) were Habitual Breakfast Eaters (≥5 times/week) and consumed this meal before 8am (84.0%). Breakfast for the majority of

148

habitual breakfast eaters consisted of one or more of the following foods and/or beverages eaten ≥5 times/week: coffee (40.4%), breakfast cereal (50.0%), milk for cereal (51.1%), fruit (28.7%), toast (13.8%), spreads (11.7%), and/or yogurt (12.8%). Breakfast may also include the following foods 1-4 times/week: eggs (58.5%), bacon (30.9%), juice (19.1%), and/or tea (17.0%).

Conclusion: Cereal, milk and fruit were the most common foods consumed by Australian men in this study and in the National Nutrition Survey (NNS) in 1995. There appear to be some differences in the percentages of consumption of these foods between this study and the 1995 NNS. However, it’s not possible to draw any conclusions about changes over time due to the differences in data collection and sample size.

Keywords: Breakfast, morning meal, Australia, males.

6.2 Introduction Breakfast is often referred to as the most important meal of the day (Brown, Bohan Brown et al. 2013) because is believed to be associated with better nutrient adequacy (Barr, DiFrancesco et al. 2013). Nutrient adequacy is believed to play a significant role in the prevention of chronic disease (WHO 2017) Breakfast is also possibly protective against weight gain (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013), although the evidence available to support these beliefs is limited and/or contradictory (Schlundt, Hill et al. 1992, Geliebter A 2000, Farshchi, Taylor et al. 2005, Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013, Betts, Richardson et al. 2014, Casazza, Brown et al. 2015, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017).

There is a lack of current data regarding the foods or beverages that constitute a typical breakfast in a multicultural country like Australia. Prior to this study, the most recent data regarding Australian breakfast habits were reported by Williams (Williams 2002) in a secondary data analysis of the National Nutrition Survey (NNS) in 1995 (Australian Bureau of statistics (ABS) 1997, Williams 2002). Williams reported that cereal products (76.2% of men and 81.2% of women), milk (65.6 % of men and 68.7% of women), and fruit (14.3% of men and 19.9% of women) were the main foods consumed for breakfast by the majority of Australian adults. Williams (Williams 2002) also reported that the foods and beverages consumed at breakfast differed between males and females, with a significantly lower percentage of men than women consuming fruit. As the 1995 NNS was conducted over 20 years ago, it is likely that the food habits of younger Australian men may have changed.

149

Therefore, a focus of the Typical Aussie Bloke study was to collect current information regarding the foods and beverages that constitute a breakfast meal for Habitual Breakfast Eaters among a sample of young (<45 y) Australian men. This study also examined the reasons for consuming or skipping breakfast and the timing of breakfast consumption.

6.3 Methods

6.3.1 Study Design Men aged 18-44y were recruited from metropolitan (Newcastle) and regional (Tamworth) areas in NSW Australia for a multicentre cross sectional study. The project was advertised using conventional (e.g., radio, television, magazines) and social (e.g.; university blogs, university Facebook) media, recruitment fliers (Appendix 10) and in person recruitment in key community locations (e.g., Callaghan campus, Tamworth Education centre, Hunter Medical Research Centre (HMRI), gyms). Participants completed an online survey and underwent a measurement session in the laboratory at the university’s Callaghan Campus or Tamworth Education Centre. Participants provided implied consent for the online survey and written consent (Appendix 11) for the laboratory measurement session. The protocol was approved by the University of Newcastle’s Human Research Ethics Committee (H-2015- 0199 – Appendix 12) and data were collected from September to December 2015.

6.3.2 Inclusion and Exclusion Criteria Participants were included if they were male and aged 18 to 44 years. The exclusion criteria were: being unable to attend a laboratory measurement session; having had cosmetic surgery that would change the shape of the body, having or being treated for a thyroid condition or insulin dependent diabetes, or claustrophobia. Participants were screened for eligibility by the first eight questions of the online survey.

6.3.3 Survey The online survey (Survey Monkey Inc., San Mateo, CA, USA) consisted of 56 questions (Appendix 13). Following the screening questions, eligible participants were directed to the remaining 48 questions which investigated lifestyle parameters such as demographic characteristics, breakfast consumption habits and waking habits.

Standard demographic and dietary questions were used where possible and most were consistent with questions from the Australian Longitudinal Study of Women’s Health (ALSWH) (ALSWH 2017) or Australian Bureau of Statistics (ABS 2011) surveys. The

150

question assessing habitual breakfast consumption was from the 1995 NNS (ABS 1997) and was the following: ‘How many days per week do you usually have something to eat for breakfast?’. The options for answers were: ‘Rarely or never; ‘1 to 2 days’; ‘3 to 4 days’; ‘5 or more days’; ‘Don't know/varies’(ABS 1997). Participants were defined as Habitual Breakfast Eaters (HBE) if they consumed breakfast 5 or more times during the week, Occasional Breakfast Eaters (OBE) if they consumed breakfast 3-4 times a week or Habitual Breakfast Skippers (HBS) if they consumed breakfast 1-2 times a week or rarely/never. Questions regarding reasons for consuming or not consuming breakfast were taken from Reeves et al. (Reeves, Halsey et al. 2013). The question was ‘What is the main reason you eat breakfast?’ with the following options: ‘It gives me energy’; ‘I want to lose weight’; ‘It helps prevent me from getting hungry before lunch time’; ‘I enjoy it’; ‘It helps me to wake up’; ‘It is what I always do’; ‘I am hungry’; ‘Eating breakfast makes it easier to control my weight’; ‘Other reasons’; ‘If other reasons was selected, please specify’. The question about reasons for skipping breakfast was the following: ‘On days that you do not have breakfast, what is the reason?’. The options for answers were: ‘Not enough time’; ‘I do not feel like eating first thing’; ‘I want to lose weight’; ‘Hung over’; ‘I have a cigarette instead’; ‘I do not have any food in the house’; ‘I do not have money for consuming breakfast’; ‘I rarely/never don’t eat breakfast’; ‘other reasons’; ‘If other reasons was selected, please specify:’.

Questions investigating the timing of waking, the timing of consumption of the first meal of the day, and the types of food and beverages consumed for breakfast were specifically developed for this study and were pilot-tested among staff at the University of Newcastle. The questions about waking habits were the following: ‘Please indicate the time you usually wake up on weekdays’ and ‘Please indicate the time you usually wake up on weekend days’ and the options were: ‘midnight to 5.00 am’; ‘5.01 to 6.00 am’; ‘6.01 to 7.00 am’; ‘7.01 to 8.00 am’; ‘8.01 to 9.00 am’; ‘9.01 to 10.00 am’; ‘10.01 to11.00 am’; ‘11.01 am to noon’.

The questions investigating timing of the first meal of the day were: ‘Please indicate the time you usually have the first meal of the day on weekdays’ and ‘Please indicate the time you usually have the first meal of the day on weekend days’. Participants were classified as Early Breakfast Eaters (EBE) if they consumed this meal between 5.01 and 8.00 am and Late Breakfast Eaters (LBE) if they consumed this meal between 8.01 am and noon. The answer options provided for these two questions were the same as those listed for the waking habits questions above.

151

The questions investigating foods and beverage consumed at breakfast were provided in the form of a short food frequency questionnaire (FFQ) and the questions were: ‘How many times per week do you usually have these foods for breakfast?’ and ‘How many times per week do you usually have these beverages for breakfast?’. Participates were asked to select the number of times they consumed the foods/beverages listed in this FFQ from the following options: ‘less than once per week’, ‘1’, ‘2’, ‘3’, ‘4’, ‘5’, ‘6’, ‘7 or more’. The foods listed in the FFQ included: ‘Fruits’; ‘Toast’; ‘Butter and/or margarine’; ‘Spread (e.g. jam, honey, peanut butter, Nutella, vegemite, etc.)’; ‘Cereals like All Bran’; ‘Cereals like Sultana BranTM, Fibre PlusTM, BranFlakesTM’; ‘Cereals like Weet BixTM ,Vita BritsTM, WeetiesTM’; ‘Cereals like ‘Corn Flakes, NutrigrainTM , Special KTM’; ‘Cereals like Porridge’; ‘Cereals like Muesli’; ‘Milk for cereal’; ‘Yogurt’; ‘Eggs’; ‘Bacon’; ‘Beans’; ‘Pancake/crepes’; ‘Other foods (please specify)’. The beverages listed in the FFQ were: ‘Coffee’; ‘Tea’; ‘Milk on its own’; ‘Hot chocolate milk/ milo’; ‘Juice’; ‘Smoothies’; ‘Other beverages (please specify)’.

Only participants who reported consuming breakfast at least 1-2 times per week were asked to complete the questions about the types of food and beverages consumed, and reasons for consuming breakfast.

6.3.4 Statistics Statistics were performed using Stata 13.1 (StataCorp LLC, College Station, Texas, USA). Descriptive statistics, such as means and distributions were calculated, and inferential statistics, including Chi Square or Fishers Exact tests, one way Anova or independent t test, were used to compare HBE versus OBE versus HBS, and for HBE consuming this meal early (EBE) versus late (LBE).

6.4 Results

6.4.1 Demographic Characteristics Data were obtained from 112 men. The majority of participants were born in Australia (87.5%) and only one (0.9%) was of Aboriginal or Torres Strait islander descendent. A majority of participants were married or in a de facto relationship (51.7%) and had a university qualification (56.3%). Furthermore, most (55.3%) were in full time employment, 25% were studying, and only a few men (2.7%) were on home duties or unemployed. The distribution of income was wide with slightly more (55.4%) in the lower income brackets (ranging from ≤AUS$25,000 to

152

participants were employed as professionals, paraprofessionals, managers or administrators, administrative assistants, sales or personal service workers (88.4%); only a minority worked in trades or as manual workers, machine operators or drivers (9.8%).

6.4.2 Comparison of Breakfast Consumption Patterns The majority of men were HBE (83.9%, n=94) consuming breakfast ≥ 5 times a week. Only a small proportion of participants were HBS (9.3%, n=10) consuming breakfast either never, rarely or 1-2 times per week. A minority of study participants were OBE (7.8%, n=7) who consumed breakfast 3-4 times a week; one person did not report his frequency of breakfast consumption.

Table 6.1 describes the characteristics of HBE (n=94), OBE (n=7) and HBS (n=10). These three groups were similar across all demographic characteristics except level of education. HBE were significantly more likely to have tertiary qualifications (62.8%), while a higher percentage of OBE (71.4%) and HBS (80.0%) had secondary school qualifications (p=0.010).

6.4.3 Habitual Breakfast Patterns among Habitual Breakfast Eaters Table 6.2 shows the habitual and non-habitual breakfast eating patterns amongst HBE. Habitual food and beverage intake for breakfast was defined as an intake of these items ≥ 5 times per week. This definition was used to identify the foods/beverages that compromised a typical breakfast for this cohort. Coffee was the beverage most habitually consumed by these men (40.4%). Breakfast cereal (50%) and milk for cereal (51.1%) were the food options consumed most regularly. The most popular types of breakfast cereals consumed were: muesli (16.8%); cereals like Weet Bix, Vita Brits, Weeties (15.9%); and porridge (12.1%). Fruits were consumed regularly by 28.6% of the study participants. Toast (13.8%), spreads (11.7%), yogurt (12.8%) and eggs (11.7%) were consumed habitually by smaller proportions of the study sample. With regards to the combination of foods, the highest frequency of consumption was breakfast cereal with milk and fruit (11.7%).

6.4.4 Non-Habitual Breakfast Patterns among Habitual Breakfast Eaters The most frequent foods and/or beverages consumed 1 to 4 times a week by HBE were eggs (58.5%), breakfast cereal (34.4%) and toast (47.9%). Butter or margarine was consumed by 41.5% and spreads by 36.2%, and bacon and fruits were each consumed by 30.9% of HBE. With regards to beverages, juice (19.1%) was the most popular drink, with coffee (18.1%) the second most popular beverage consumed 1 to 4 times a week.

153

With regards to combinations of foods consumed < 5 times a week, popular combinations were: eggs and toast (36.2%); eggs, toast and butter or margarine (29.4%); bacon and eggs (26.6%), and bacon, eggs and toast (17.0%). Another popular less frequent combination was breakfast cereal with yoghurt (14.9%).

6.4.5 Other Foods and Beverages consumed for Breakfast by Habitual Breakfast Eaters Six men reported consuming shakes (e.g., protein shakes, green infusions, meal shakes or unspecified type). Four participants reported consuming servings of meat (ham, steak or chicken) and six men reported consuming vegetables (such as mushroom, zucchini, spinach, carrots, capsicum, celery or onion). Five participants reported consuming almond or coconut milk and six reported consuming either oats or nuts. Two men reported consuming cheese and two reported consuming English muffins. The number of times these foods/beverages were consumed in a usual week and whether this reached habitual consumption is unknown.

6.4.6 Early vs Late Breakfast Consumption and its relation with Waking Habits The majority of participants reported waking between 5.01 and 8.00 am during the week (all participants: 86.6%; HBE: 77.6%) and between 6.01 and 9.00 am during the weekend (all participants: 78.6%; HBE: 79.8%). Most HBE consumed the first meal of the day between 6.01 and 8.00 am (76.6%) during the week and 7.01 and 9.00 am on weekends (65.9%). Most HBE consumed the first meal of the day early during the week (84.0%) compared to the weekend (54.6%). Also, most HBE consumed their first meal of the day within two hours of waking during the week (86.2%) and the weekend (87.2%); only a small proportion consumed this meal between two and five hours after waking during the week (13.8%) or on the weekend (12.8%).

Table 6.3 compares the characteristics of HBE who consume the first meal of the day Early (n=79) vs Late (n=15) during the week. These two groups differed significantly with regards to age, marriage, income, and number of dependent children. EBE were significantly older (p=0.0124) and more likely to be married (59.4%; p=0.023) compared to LBE. A significantly higher percentage of EBE (69.8%) had a full time job compared with LBE who were more likely to have a part time job (40.0%; p=0.01). EBE were more likely to earn ≥ AUD50K (69.7%) compared to LBE (26.7%; p=0.043). A higher percentage of EBE (35.6%) had one or more dependent children compared with LBE (6.7%, p=0.032), however, in both groups the majority of men had no dependent children.

154

Table 6.1 Demographic characteristics of participants of the Typical Aussie Bloke study categorised by habitual breakfast eating patterns.

Habitual Breakfast Occasional Habitual Breakfast P value Eaters (HBE) Breakfast Skippers (n=94) Eaters(OBE) (HBS) (n=7) (n=10) % (n) 83.9% (94) 6.3% (7) 8.9% (10) Age y (mean (SD)) 30 ± 7 25 ± 5 31 ± 6 0.1873 Height (mean ±SD ) 179.8 ± 6.6 177.0 ± 5.0 180.5 ± 5.3 0.4826 Weight (mean ±SD ) 82.5 ± 14.7 80 ± 14.4 84.2 ± 11.7 0.8395 BMI (mean ±SD ) 25.5 ± 4.2 25.5 ± 3.8 26.0 ± 4.3 0.9526 Country of birth Australia 86.2% 100.0% 100.0% 1Other countries 13.8% 0% 0% 0.465 ATSI descent Yes 0%* 0% 10.0% No 98.9%* 100.0% 90.0% 0.155 Marital status 2 Single 44.7%* 42.9% 70.0% Married or de facto relationship 54.3%* 57.1% 30.0% 0.349 Education 3High school or trade 36.2%* 71.4% 80.0% 4University or higher degree 62.8%* 28.6% 20.0% 0.010 Employment status Full time paid work 53.2% 57.1% 80.0% Part time paid work or casual paid 18.1% 14.3% work Studying 25.5% 28.6% 20.0% Home duties or Unemployed 3.2% 0% 0% 0.769 Annual household gross income

≤AUS$25,000 to 49,999 31.9%* 14.3% 20.0%

≥AUS$50,000 to 99,999 21.3%* 57.1% 50.0%

≥AUS$100,000 to 149,999 21.3%* 14.3% 20.0%

≥AUS$150,000 20.2%* 14.3% 10.0% 0.393

People living in the household Living alone 7.5%* 14.3% 20.0% 5Living with other people 90.4%* 85.7% 80.0% 0.233 Dependent children None 69.2% 100.0% 70.0% One or more 30.9% 0% 30.0% 0.255 Past/present or Future Occupation

Trades, manual workers, machine 8.5% 0% 30.0% operators or drivers Professional, Paraprofessional, 90.4% 100.0% 70.0% managers or administrator, administrative assistant, sales or personal service worker 155

Never had a paid job 1.1% 0 0 0.240

*The total number of subjects for this variable do not add up to 100% due to missing data reported (such as replied: ‘prefer not to answer’ or ‘don’t know’ Or don’t know/Varies’ or ‘don’t know or would rather not say’) Please also note: one participant was excluded from the data displayed in this table because it was not possible to classify his breakfast consumption habits as he answered ‘do not know/varies’ to the Breakfast Consumption Habits question. HBE = Habitual Breakfast Eaters who reported to consume breakfast 5 or more times per week OBE = Occasional Breakfast Eaters who reported to consume breakfast 3 to 4 times per week. HBS = Habitual Breakfast Skippers who reported to consume breakfast ‘rarely or never’ or 1 to 2 times per week.’ 1 Others country: New Zealand, Italy, Germany, United Kingdom, Netherlands, America, Canada, Bangladesh or South Korea or unknown. 2 Single includes: separated, divorced, widowed or never married. 3 School includes: Intermediate Certificate (or equivalent) or Higher School or Leaving Certificate (or equivalent) or Trade/apprenticeship (eg. Hairdresser, Chef) or Certificate/diploma (eg. Child Care, Technician) 4 University includes: University degree or University Higher degree (eg. Grad Dip, Masters, PhD) 5 Living with other people includes living with one, two, three, four or five other people ATSI – Aboriginal or Torres Strait Islander.

156

6.4.7 Reasons for Consuming or Not Consuming Breakfast Reasons for consuming breakfast were provided by participants (n=107) who reported consuming breakfast ≥ 1-2 times per week. The most common responses to the question ‘What is the main reason you eat breakfast?’ were: ‘I am hungry’ (30.8% (33)); ‘It is what I always do’ (24.3% (26)) or ‘It gives me energy’ (19.7% (20)) or ‘It helps prevent me from getting hungry before lunch time’ (13.1% (14)). Fewer participants replied with the following: ‘I enjoy it’ (7.5% (8)), ‘Eating breakfast makes it easier to control my weight’ (1.9% (2)); ‘It helps me to wake up’ (0.9% (1)) or ‘Other reasons’ (2.8% (3)). None of the participants replied ‘I want to lose weight’ as a reason for consuming breakfast. Other reasons for consuming breakfast included ‘it is the most important meal of the day’, ‘to be social’ or ‘because the kids must eat breakfast’.

A majority of study participants reported that ‘I rarely/never don’t eat breakfast’ (52.7% (59)) as their response to the question ‘On days that you do not have breakfast, what is the reason?’. Approximately 20% each of participants replied that: ‘I do not feel like eating first thing’ (18.0% (20)) or ‘Not enough time’ (17.0% (19)). A small number answered: ‘I do not have any food in the house’ (2.7% (3)); ‘I have a cigarette instead’ (1.8% (2)); ‘Hung over’ (0.9% (1)); ‘I want to lose weight’ (0.9% (1)) or other reasons (6.3% (7)). None of the participants replied ‘I do not have money for consuming breakfast’ as a reason for skipping breakfast. ‘Other reasons’ for not consuming breakfasts included ‘having slept in’, ‘not wanting to eat before exercising’ or ‘skipping breakfast to do something different’.

157

Table 6.2 Habitual and non-habitual food and beverage consumption for breakfast among Habitual Breakfast Eaters

Habitual Breakfast Eaters HBE consuming these (HBE) consuming these foods foods/beverages 1-4 times a and/or beverages ≥ 5 times a week (n=94) week (n=94) Foods Any breakfast cereal 50.0%1 34.4%2 • Cereals like muesli 19.2% 22.3% • Cereals like Weet Bix,Vita Brits, 17.0% 25.5% Weeties • Cereals like All Bran 6.4% 5.3% • Cereals like Cornflakes, Nutrigrain, 4.3% 14.9% Special K • Cereals like Sultana bran, Fiber Plus 2.1% 4.3% and Branflakes • Cereals like porridge 12.8% 24.5% Milk for cereal 51.1% 21.3% Fruit 28.7% 30.9% Yogurt 12.8% 24.5% Toast 13.8% 47.9% Spread (e.g. jam, honey, peanut butter, 11.7% 36.2% Nutella, vegemite, etc.) Butter and/or margarine 8.5% 41.5% Eggs 11.7% 58.5% Bacon 0 30.9% Beans 0 13.1% Pancakes 0 8.5% Beverages Coffee 40.4% 18.1% Tea 6.4% 17.0% juice 5.3% 19.1% Smoothies 4.3% 11.7% Milk on its own 1.1% 6.4% Combinations of foods and/or beverages • Cold breakfasts Any cereal & yogurt 10.6% 14.9% Fruit & yogurt 7.4% 12.8% Cereal with milk & fruit 11.7% 10.6% Cereal, fruit & yogurt 6.4% 8.5% Toast & butter or margarine & spread 5.3% 29.8% • Cooked breakfasts Toast & eggs 3.2% 36.2% Bacon & eggs 0 26.6% Toast, eggs, butter or margarine 0 29.4% Toast, bacon & eggs 0 17.0% Toast (and butter/margarine) with bacon 0 13.8% and eggs. HBE= Habitual Breakfast Eaters who reported to consume breakfast 5 or more times per week

158

1 Any breakfast cereal consumption habitually includes participants who have reported to consume one or more breakfast cereal types habitually. Please note that these subjects could also have consumed in addiction one or more breakfast cereal types 1-4 times a week. 2 Any breakfast cereal consumption 1-4 times includes the number of subjects who have consumed one or more breakfast cereals 1-4 times a week and did not consume one or more of the breakfast cereal types habitually.

159

Table 6.3 Demographic characteristics of HBE categorised by timing of first meal of the day consumption

Early Breakfast Late Breakfast P value Eaters (EBE)1 Eaters (LBE)2 (n=79) (n=15)

% (n) 84.0% (79) 16.0% (15) Age y (mean ±SD) 31 ± 7 26 ± 7 0.0124 Height (mean ±SD ) 179.8 ± 6.3 179.8 ± 8.0 0.9916 Weight (mean ±SD ) 81.9 ± 13.5 85.5 ± 20.4 0.3838 BMI (mean ±SD ) 25.4 ± 4.1 26.5 ± 4.8 0.3558 Country of birth Australia 86.1% 86.7% 3Other countries 13.9% 13.3% 1.000 ATSI descent Yes 0* 0 No 98.7%* 100.0% NA Marital status 4Single 39.2%* 73.3% Married or in de facto relationship 59.4%* 26.7% 0.023 Education 5School 31.6%* 60.0% 6University 67.1%* 40.0% 0.076 Employment status Full time paid work 60.8% 13.3% Part time paid work or casual paid work 13.9% 40.0% Studying 24.1% 33.3% Home duties or Unemployed 1.3% 13.3% 0.001 Annual household gross income (AUS$) ≤$25,000 to 49,999 26.6%* 60.0%* ≥$50.000 to 99,999 24.1%* 6.7%* ≥100.000 to 149,999 24.1%* 6.7%* $150,000 to 200,000 or more 21.5%* 13.3%* 0.043 People living in the household Living alone 6.3%* 13.3% 7Living with other people 91.1%* 86.7% 0.320 Dependent children None 64.6% 93.3% One or more 35.4% 6.7% 0.032 Past/present or Future Occupation Trades, manual workers, machine operators or 8.9% 6.7% drivers Professional, paraprofessional, managers or 89.9% 93.3% administrator, administrative assistant, sales or personal service worker Never had a paid job 1.3% 0% 1.000

* The total number of subjects for this variable not add up to 100% due to missing data reported (such as ‘replied prefer not to answer’ or ‘don’t know’ Or don’t know/Varies’ or ‘don’t know or would rather not say’). HBE= Habitual Breakfast Eaters who reported to consume breakfast 5 or more times per week

160

1EBE= Early Breakfast Eaters includes Habitual Breakfast Eaters who consumed the first meal of the day early on weekdays (between 5.01 and 8.00 am) 2LBE= Late Breakfast includes Habitual Breakfast Eaters who consumed the first meal of the day late on weekdays (between 8.01 am and noon) NA= Not Applicable 3 Others country: New Zealand, Italy, Germany, United Kingdom, Netherlands, America, Canada, Bangladesh or South Korea or unknown. 4 Single includes: separated, divorced, widowed or never married. 5 School includes: Intermediate Certificate (or equivalent) or Higher School or Leaving Certificate (or equivalent) or Trade/apprenticeship (eg. Hairdresser, Chef) or Certificate/diploma (eg. Child Care, Technician) 6 University includes: University degree or University Higher degree (eg. Grad Dip, Masters, PhD) 7 Living with other people includes living with one, two, three, four or five other people ATSI – Aboriginal or Torres Strait Islander

161

6.5 Discussion This study investigated the current breakfast habits of young Australian men by collecting current information regarding the foods and beverages that constitute a breakfast meal for Habitual Breakfast Eaters (HBE) among a sample of young (<45 y) Australian men. This study also examined the reasons for consuming or skipping breakfast and the timing of breakfast consumption. In this study the majority of the sample were HBE and consumed this meal earlier than 8.00 am. Those who habitually ate breakfast (HBE) were more likely to have a university qualification than Occasional Breakfast Eaters (OBE) and Habitual Breakfast Skippers (HBS). Early Breakfast Eaters (EBE) were more likely to have higher incomes, dependent children, full time jobs and were more likely to be married/defacto than Late Breakfast Eaters (LBE). Breakfast cereal and milk formed the most frequent foods consumed habitually and coffee was the most common beverage. The main reasons for consuming breakfast were feeling hungry, needing energy, and it being a habit, whereas not having enough time or not feeling hungry were the main reasons for skipping breakfast.

In the Typical Aussie Bloke study there was a higher percentage of men (83.9%) who reported consuming breakfast habitually than in the 1995 NNS (73.5% of men aged ≥19 y; 57.5% of men aged 19-24 y; 66.5% of men aged 25-44 y (ABS 1997)). The Typical Aussie Bloke study recruited a relatively small sample of young men living in Newcastle and Tamworth whereas the NNS represents the breakfast habits of a much larger sample from across Australia (13,858 participants aged two or more years old from urban and rural areas in all States and Territories). Thus, we would expect differences based on sampling but there is also the likelihood that breakfast habits have changed since 1995. This idea is supported by the findings of the OXFAM international survey in 2011 which reported that dietary changes occur over time. They found that in Australia, 62% of participants reported they had changed their eating habits in the two years before (OXFAM 2011).

In the Typical Aussie Bloke study, other than level of attained education, the HBE group did not differ significantly from OBE and HBS groups. This contradicts the findings of the NNS 1995 which found habitual breakfast consumption habits differed in relation to income level (Breakfast eaters: Quintile 1 (lowest income) 83.9%, Quintile 5 (highest income) 79.7%; p<0.005) (Williams 2002). The findings from the 1995 NNS found that people with the lowest incomes had a significantly higher proportion of HBE compared to people from all other income groups (Williams 2002). This difference could be due to the smaller sample size in the Typical Aussie Bloke study and the differences in sampling processes. There have been

162

no other investigations of habitual breakfast consumption and demographic data in Australian men reported since the 1995 NNS.

There was a higher proportion of Australian men in the Typical Aussie Bloke study who reported being HBE compared with adults in other countries. In an American population of men aged 18 y and older (surveyed 1989-1991) 74.8% were regular breakfast consumers (Haines, Guilkey et al. 1996) and in an Italian population 73.6% of students regularly consumed breakfast (Isa and Masuri 2011)). However, the proportion of Habitual Breakfast Eaters and Skippers amongst the ‘Typical Aussie Bloke’ was similar to a Canadian population where 89% of adults were breakfast eaters. The study of American men in 1989- 1991 may not represent current eating habits (Haines, Guilkey et al. 1996). Similarly, in the Italian study, although more recent, only undergraduate students aged 18-25 y were recruited and may not be representative of the Italian population as a whole (Isa and Masuri 2011). It is also important to note that these two studies did not assess the regularity of breakfast consumption but only the percentage of people consuming breakfast on the day that the 24 hour recalls were collected.

There are only a limited number of studies that have investigated the breakfast habits of Australia and other countries. Interestingly, the most common foods and beverages consumed by this sample are similar to the most frequent foods and beverages consumed in the 1995 NNS, which was conducted 20 years earlier than the ‘Typical Aussie Bloke’ study. The most frequently consumed foods/beverages in 1995 by Australian men over 19 years old were: cereal products (76.2%) including cold (44.3%) and hot cereals (4.9%), breads (44.0%) and pastries/cakes/biscuits (44.0%), milk (65.6 %) and tea or coffee (57.5%). Fruit was consumed by a smaller proportion of men (14.3%). Therefore, cereal, milk and fruit were the most common foods consumed in both sets of data. Though, there appear to be some differences in the percentages of consumption of these common foods, for example in fruits and milk consumption. However, we must recognise the inherent differences in data collection and design and not draw any strong conclusions about changes in breakfast consumption over time. Of course, the 1995 NNS assessed the foods/beverages consumed at breakfast on the day before data collection using a 24 hour recall whereas our study assessed the number of times men consumed these foods/beverages at breakfast in a usual week using an FFQ, and thus was trying to ascertain habitual breakfast foods habits.

163

The breakfast foods consumed by Australian men were similar to the American men in the Continuing Survey of Food Intake, which included bread (19.5%), eggs (17.5%), ready to eat cereal (16.7%), fruits (4.8%) and cooked cereals (4.0%) (Siega-Riz, Popkin et al. 2000). However, Australian food choices for breakfast were considerably different to the Italian adults’ breakfast foods which included coffee (93.3%), sugar (81.6%), crispbread/rusk (87.9%), milk (67.7%), biscuits (42.9%), yogurt (42.8%), brioche (38%), jam (37.3%) and tea (33.4%) (di Giuseppe, Di Castelnuovo et al. 2012).

Interestingly, significant demographic differences were found between HBE who consumed the meal early (n=79) and late (n=15) during the week. The findings show that EBE could be considered to have higher levels of responsibilities, including being married, having a full- time job, earning a higher income and having dependent children, which most likely requires an organised consistent schedule early in the morning. LBE were more likely to be younger and single and to have no dependent children and a lower income, which suggests that they had fewer responsibilities and therefore were less likely to follow a consistent schedule in the morning. Our study shows than amongst people who consumed breakfast habitually, breakfast was mostly consumed earlier than 8.00am (84%) during the week, while a considerable proportion of HBE consumed this meal later than 8.00 am on the weekend (45.6%). This suggests that the timing of breakfast consumption is related to their commitments. Participants have breakfast earlier during the week because they may either have to go to work or to take their children to school or they have to study whereas during the weekend the majority of participants have breakfast later than 8.00 am as they are free from these commitments.

Habit, hunger and energy needs were the main drivers for the consumption of breakfast whereas lack of time and not feeling like eating first thing in the morning were the main reasons for not consuming breakfast. These findings are similar to those of Reeves et al (2013) in the UK amongst 1068 subjects, as a large percentage of the participants replied ‘it gives me energy’ (70.2%), ‘I am hungry’ (66.1%) and/or ‘It is what I always do’ (57.4 %) as the reasons for consuming breakfast (Reeves, Halsey et al. 2013). Also, large proportions of the population replied ‘not enough of time’ (40.2%) and ‘I do not feel the need to eat in the morning’ (49.5%) as the reasons for skipping breakfast (Reeves, Halsey et al. 2013).

This study has a number of strengths. It is the first study to report data regarding the current habitual (2015) breakfast habits of men in Australia. Furthermore, this paper investigated

164

timing of breakfast consumption and its relation to waking habits, which has not previously been reported in Australia. For this study, we used a diverse range of recruitment strategies (fliers, conventional and social media) and multiple data collection sites including metropolitan (Newcastle) and regional areas (Tamworth). However, the sample that we managed to recruit was mostly homogenous in term of breakfast habits, as they mostly were Habitual Breakfast Eaters and early breakfast eaters, and in terms of occupation, they were mostly professional, paraprofessional, managers or administrator, administrative assistants, sales or personal service worker. It is possible that because the recruitment fliers indicated that the study was investigating breakfast and its relationships with health parameters, this has potentially biased the recruitment to subjects who habitually consumed breakfast and were interested in the health benefits of this behaviour. Moreover, the relatively small convenience sample does not allow the results to be generalized with confidence to the Australian male population. Furthermore, the evidence reported in this study is based on self- reported data. However, many questions, including the demographic and breakfast frequency questions, have been widely used in Australia by large national surveys such as ALSWH (ALSWH 2017), Australian Bureau of Statistics (ABS) in 2011 (ABS 2011) or NNS in 1995 (ABS 1997), and therefore, these questions can be considered well established methods to collect this self-reported information.

6.6 Conclusions This study found that in this sample of younger Australian men the majority consume breakfast most days of the week, and do so before 8am. The most commons foods consumed by these Australian men do not appear to have changed substantially since the NNS in 1995 as cereal, milk and fruit were the most common foods consumed at both times. There appear to be some differences in the percentages of these common foods consumed between these two studies, however, the differences in data collection and design/purpose do not allow us to draw any conclusions regarding changes over time.

6.7 Acknowledgments This work was supported by the Australian Government Research Training Program Scholarship (RTP). The authors wish to thank the participants of the typical Aussie Bloke study for their participation in the study. The authors also wish to thank A/Prof Leanne Brown for her kind help and support on facilitating data collection for this study in Tamworth Education Centre (Tamworth, NSW, Australia). The authors also wish to acknowledge Anna

165

Bukey, Jessica Plotrowski and Philip Jacobson for their kind help on ordering the consumables needed for data collection for this study.

6.8 Conflict of Interest None

6.9 Authors’ contributions All authors (AQ; AP; RC; LMW) have made substantial contributions to all of the following: conception and design of the study, analysis and interpretation of data, drafting the paper, critically revising the paper for important intellectual content and final approval of the version to be submitted.

6.10 Ethics This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Human Research Ethics Committee of the University of Newcastle. Written and implied informed consent was obtained from all subjects.

166

Chapter 7: ‘Typical Aussie Bloke’ Part 2: Breakfast Consumption and Eating Patterns in relation to intermediate risk factors for Obesity and Chronic Disease Development. The work presented in the manuscript was accepted for oral presentation at the Dietitians Association of Australia Conference in Sydney in May 2018 (Appendix 14).

The work presented in the chapter was completed in collaboration with the co-authors (Appendix 15).

7.1 Introduction Regular breakfast consumption may mitigate against chronic disease risk factors for diabetes and cardiovascular diseases though multiple mechanisms (Cahill, Chiuve et al. 2013, Bi, Gan et al. 2015, Yokoyama, Onishi et al. 2016). Breakfast is believed to minimise body weight gain (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013); however, the evidence available to support its protective effect is often limited or contradictory (Brown, Bohan Brown et al. 2013, Casazza, Fontaine et al. 2013, Casazza, Brown et al. 2015). Breakfast eating habits was also found to be associated with better nutrient adequacy than skippers in a large Canadian study (Barr, DiFrancesco et al. 2013). Nutrient adequacy is believed to play a significant role in the prevention of chronic disease (WHO 2017). Furthermore, One randomized controlled trial (RCT) (Betts, Richardson et al. 2014) found that regular breakfast consumption was associated with increased physical activity levels in the morning and increased daily energy expenditure, but other RCTs did not support this effect (Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017). Some researchers also suggest that regular breakfast consumption reduces overall daily energy intake (Farshchi, Taylor et al. 2005) by reducing snacking. However, the majority of the trials conducted on this topic do not support this belief (Astbury, Taylor et al. 2011, Halsey, Huber et al. 2012, Levitsky and Pacanowski 2013, Betts, Richardson et al. 2014, Reeves, Huber et al. 2014, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017). Some cross-sectional studies and randomized crossover trials (Farshchi, Taylor et al. 2005, Astbury, Taylor et al. 2011, Deshmukh-Taskar, Nicklas et al. 2013, Kobayashi, Ogata et al. 2014) suggest that skipping breakfast may have detrimental associations on metabolic blood biomarkers due to the prolonged fasting state; therefore, breakfast consumption may be protective by reducing fasting time.

167

A large cross-sectional study conducted in the USA using the National Health and Nutrition Examination Survey (NHNES) 1999-2006 (Deshmukh-Taskar, Nicklas et al. 2013) investigated associations among breakfast habits, body size and cardio-metabolic parameters, and found regular breakfast consumption to be significantly associated with lower body weight and better cardiometabolic parameters in young adults aged 20-39 years. Furthermore, some gender differences in breakfast consumption habits were noted in this study; for example, a higher percentage of men were skipping breakfast than women (26.1% male vs 21.2% female) (Deshmukh-Taskar, Nicklas et al. 2013). These findings suggest gender to be an important variable to consider when investigating breakfast habits. In this study we have focussed on investigating the breakfast habits of young Australian males. To the authors’ knowledge, no other studies of young men in Australia have collected comprehensive data on breakfast habits and mediators of obesity and chronic disease development. Relevant measures included in this study were anthropometric (Body Mass Index (BMI), circumferences and body composition), metabolic (Resting Metabolic Rate (RMR), blood glucose and lipid biomarkers, blood pressure) and habitual breakfast habits and the timing of breakfast consumption. Therefore, this chapter investigated associations between habitual breakfast consumption or the timing of breakfast and metabolic, anthropometric and lifestyle characteristics, such as fruit and vegetable consumption, Physical Activity (PA), sleeping and waking habits, among a sample of young Australian males (18-44 y). This has allowed further exploration of the potential mechanisms of action by which breakfast may be protective against obesity and intermediate risk factors of chronic disease development. As breakfast consumption is hypothesized to impact on the number of daily eating events, this chapter also investigated the association between a higher vs lower frequency of daily eating events on anthropometric, metabolic parameters and health behaviour characteristics.

The aims of this part of the ‘Typical Aussie Bloke’ study were:

1. To compare the anthropometric, metabolic, and health behaviour characteristics of a sample of young Australian men who are Habitual Breakfast Eaters, Occasional Breakfast Eaters and Habitual Breakfast Skippers 2. To compare the anthropometric, metabolic, and health behaviour characteristics of a sample of young Australian men who are Habitual Breakfast Eaters but differ in the timing of their breakfast consumption

168

3. To compare the anthropometric, metabolic, and health behaviour characteristics of a sample of young Australian men who differ in the frequency of their daily eating events

7.2 Methods

7.2.1 Online Survey The online survey in Survey Monkey (www.surveymonkeycom; San Mateo, CA) consisted of 56 questions, including the eight initial screening questions (appendix 13). Eligible participants were then directed to the remaining 48 questions which included lifestyle questions such as demographic, physical activity (PA), fruit and vegetable consumption, breakfast consumption habits, eating frequency, and waking and sleeping habits.

The Short Diet questions, including fruit and vegetable habits, were taken from the Australian Longitudinal Study on Women’s Health (ALSWH) surveys (for more information about these questions see Appendix 13) (ALSWH 2017). The Physical Activity (PA) questions were derived from the Active Australia’s 1999 National Physical Activity Survey (Armstrong, Bauman et al. 2000). The PA questions asked participants to report the amount of time (hours and minutes) spent doing each type of PA: leisure activities (walking, moderate activity and vigorous activity) and work (vigorous activity). The PA variables were then transformed into Metabolic Equivalent Task (MET) minutes (Ainsworth, Haskell et al.

1993). MET refers to a unit of RMR and is normally considered to be 3.5mL O2/kg/minute. MET minutes were determined using coefficients for each intensity of PA as follow: 3.0* X minutes of walking per week, 4.0* X minutes of moderate intensity activities per week, and 7.5 * X minutes of vigorous intensity activities per week; activity was then summed to provide total MET minutes per week. PA from leisure activities was then categorized as: Nil/sedentary for <40, low for 40 to <600, moderate for 600 to <1200, and high for ≥ 1200 MET minutes per week (ALSWH 2017).

The question assessing eating frequency was used by the National Nutrition Survey (NNS) in 1995 (ABS 1997). The question was as follows: ‘How many times do you usually have something to eat in a day (including snacks and evenings)?’ The answer options were: ‘once’; ‘2 to 4 times’; ‘5 to 6 times’; 7 or more times; ‘Don't know/varies’(ABS 1997). Participants were classified as frequent eaters (FE) if they consumed 5 or more eating events per day and less frequent eaters (LFE) if they consumed 1-4 eating events per day. Questions investigating the consumption of the first meal of the day, waking up and sleeping habits,

169

were specifically developed for this study and piloted amongst staff (Information about these questions can be found in Chapter 6 and/or in Appendix 13).

7.2.2 Lab Measurement Session Participants were instructed to attend the measurement session with a minimum of four hours fasting (only water was allowed) and to refrain from smoking; participants were also instructed to avoid any vigorous exercise for 24 hours prior to the measurement session. The measurement session was undertaken either at Callaghan Campus for the participants based in the Newcastle area or at Tamworth Education Centre for the participants based in the Tamworth area.

The lab session lasted approximately 80 min and consisted of the following measurements. Height and sitting height, body weight, and waist, hip and chest circumferences were measured respectively with a stadiometer, body mass scale and a metal tape measure, following the International Society for the Advancement of Kinanthropometry guidelines. The measurements were taken twice and a third measurement was taken if the first two differed by 0.5 kg or 0.5 cm or more, then the two closest measurements were averaged. Body composition was measured by bioelectrical impedance using the Inbody 720 (Inbody, Seoul, Korea). Peripheral blood pressure parameters were recorded using the Pulsecor Cardioscope BP+ (Pulsecor Ltd, Sydney, Australia). Participants were seated for five minutes before the first measurement, and a minimum of two measurements were obtained. The measurements were repeated to a maximum of five times if the first two measurements differed by more than 10mmHg for systolic peripheral pressures and 6mmHg for diastolic peripheral pressures, then the two closest measurements were averaged. RMR was measured for approximately 20 min with FitmateTM (Cosmed Asia-Pacific Ltd, Artarmon NSW, Australia), an indirect calorimeter with a canopy system, while the participants were lying on a bed. The first five minutes of RMR measurement were discarded and the RMR was considered stable when a minimum of five minutes measurement had a coefficient of variation less than or equal to 10%. The indirect calorimeter was calibrated with a calibration syringe at the beginning of each new measurement day. Capillary blood samples obtained from a finger were collected and used to measure lipid profiles (triglycerides, total cholesterol, HDL and LDL cholesterol) and glucose levels using the CardioCheck® PA (Point of Care Diagnostics Australia, Artarmon NSW, Australia). The quality check process to ensure the integrity of the strips used to measure the biomarker in the blood was conducted on a regular basis (monthly) as per guidelines (PTSDiagnostic 2017).

170

7.2.3 Statistics Statistical analyses were performed using Stata 13.1 (StataCorp LLC, College Station, Texas, USA). Descriptive statistics, such as measures of average and dispersion were completed. Inferential statistics were used to compare Habitual Breakfast Eaters (HBE), Occasional Breakfast Eaters (OBE) and Habitual Breakfast Skippers (HBS); HBE who consumed this meal Early (Early Breakfast Eaters (EBE)) vs Late (Late Breakfast Eaters (LBE)); and Frequent Eaters (FE) vs Less Frequent Eaters (LFE). The inferential statistical tests conducted were the following: Chi Square tests or Fisher Exact tests for categorical variables; Independent t tests (for parametric distributions) or Wilcoxon Rank sum tests (for non- parametric distributions) were performed for two independent group comparison for continuous variables; One Way Anova (for parametric distributions) or Kruskal Wallis tests (for non-parametric distribution) for three independent group comparisons with continuous variables.

7.3 Results

7.3.1 Habitual Breakfast Eaters (HBE) vs Occasional Breakfast Eaters (OBE) vs Habitual Breakfast Skippers (HBS) Demographic variables for HBE vs OBE vs HBS and EBE vs LBE are presented in Chapter 6. Tables 7.1 and 7.2 present the comparisons among HBE, OBE and HBS in relation to age, anthropometric, metabolic and other parameters. No significant differences were found for age between groups. There was a significant trend (p=0.015) for a higher percentage of HBE to consume 5 or more daily eating events (59.6 %) in comparison to OBE (28.6%) and HBS (20%) who were instead more likely to eat 1-4 times/day. PA levels (MET min/week) did not differ significantly (p=1.00) among groups with all three groups reporting high levels (>1200 MET min/week) of PA (HBE 68.1%; OBE 71.4%, HBS 80.0%). The majority of people in all three groups consumed on average 1-2 serves of fruit per day (HBE 57.4%; OBE 71.4%; 90% HBS) and 2-3 serves of vegetables per day (HBE 55.3%; OBE 85.7%; HBS 60%) with no significant differences between groups.

As illustrated in Tables 7.1 and 7.2, none of the anthropometric or metabolic parameters differed significantly among the three groups. On average the three groups met the guidelines for waist to hip ratio (Dietitians Association of Australia (DAA) 2015), waist circumference (National Heart Foundation (NHF) 2014), blood glucose levels (Diabetes Australia (DA) 2008), HDL cholesterol (NVDPA 2012), triglycerides (NVDPA 2012) and systolic and

171

diastolic blood pressure (NHF and AZSNZ 2005). However, the mean BMI of the three groups indicated they were slightly overweight and all groups had total cholesterol and LDL cholesterol concentrations higher than recommend by the National Vascular Disease Prevention Alliance (NVDPA) (NVDPA 2012). The guideline recommendations are included in the footnotes of Tables 7.1, 7.3 and 7.5.

7.3.2 Early Breakfast Eaters (EBE) vs Late Breakfast Eaters (LBE) Tables 7.3 and 7.4 compare EBE and LBE for age, anthropometric and metabolic parameters, PA, dietary habits, sleeping and waking habits. EBE and LBE were significantly different (p=0.0124) in age, with the majority of EBE (79.8%) being older than 24 y compared to LBE (46.7%; p=0.048). These two groups had significantly different times for self-reported waking (p<0.001) and going to sleep (p<0.001). All EBE (100%) reported waking earlier than 8.00 am and the majority of them (83.5%) went to sleep earlier than 11.00 pm during the week. The majority of LBE reported waking later than 8.00 am (53.3%) and going to sleep later than 11.00 pm (80.0%). As reported in Chapter 6, EBE were more likely to be married (EBE 59%, LBE 27%), have children (EBE 35%, EBE 7%, be employed full time (EBE 61%, LBE 13%), and to have an income ≥ AUD50K (69.7%) compared to LBE (40%; p=0.043). The number of daily eating events did not significantly differ between EBE and LBE (p=0.475). Also, PA levels and fruit and vegetable consumption did not differ significantly between these two groups.

As illustrated in Table 7.3, none of the anthropometric and metabolic parameters were significantly different between groups. On average, these two groups were meeting the guidelines for waist to hip ratio, waist circumference, glucose levels, HDL cholesterol and triglyceride concentrations, and blood pressure. On average, both groups were slightly overweight, and had total cholesterol and LDL cholesterol concentrations higher than the levels recommend by the NVDPA (NVDPA 2012) .

7.3.3 Frequent Eaters (FE) vs Less Frequent Eaters (LFE) Men consuming 1-4 eating events per day (LFE) were compared to men consuming 5 or more eating events per day (FE). Tables 7.5 and 7.6 present the comparisons of the characteristics of LFE vs FE. LFE and FE did not differ significantly in age. These two groups differed significantly for living arrangements with a higher proportion of FE (95%) living with other people (82%; p=0.041). Also, a higher percentage of FE (61.7%) were married compared to LFE (42.0%; p=0.05), and FE were more likely than LFE to have

172

children (p=0.082). A higher percentage of FE (80.0 %) went to sleep earlier than 11.00 pm compared to LFE (60.0%; p=0.021). No other measures were significantly different between the groups. As seen in Table 7.5, on average the two groups were meeting the guidelines for waist to hip ratio, waist circumference, glucose levels, HDL cholesterol and triglyceride concentrations, and blood pressure. However, all three groups were on average slightly overweight and had total cholesterol and LDL cholesterol concentrations higher than the NVDPA guidelines. None of the measures were significantly different between the groups.

7.4 Discussion This chapter investigated the relationships between habitual breakfast consumption, the timing of breakfast consumption, or daily eating frequency patterns and the anthropometric, metabolic and lifestyle characteristics of a sample of young Australian men. Overall, this study found that patterns of habitual breakfast consumption, timing of breakfast and daily meal frequency had little or no influence on these young men’s health characteristics. A significantly higher proportion of HBE ate 5 or more times per day in comparison to OBE and HBS who were more likely to eat 1-4 times per day.

The lack of an association between habitual breakfast consumption habits and health parameters in this study differs from the findings of a large (n=5316) cross sectional study of young US adults (20–39 years of age) using NHANES data from 1999–2006 (Deshmukh- Taskar, Nicklas et al. 2013). The US study found significant associations between breakfast consumption habits and body size and cardio-metabolic parameters. Deshmukh-Taskar et al. (2013) reported that breakfast eaters consuming ready to eat cereals, breakfast eaters consuming other types of breakfast and breakfast skippers were significantly less likely to be overweight, obesity, abdominal obesity, elevated blood pressure and elevated cholesterol (total, LDL and HDL), with breakfast eaters consuming ready to eat cereal being the group with the lowest lowest likelihood to be associated with those adverse health indicators (Deshmukh-Taskar, Nicklas et al. 2013).

Although our study was considerably smaller than the NHANES study, there were a number of other important differences. The number of men skipping breakfast in the US study was 26.1% compared to 8.9% in our study. As reported in Chapter 6, 40% of the men in our study consumed cereal (Ready to Eat Cereal (RTEC)) for breakfast compared to only 16% of men in the US study. Also, the men in our study were substantially more physically active than participants in the US study, with the majority of men in the current study undertaking

173

relatively high levels of PA per week (68.8% expended ≥1200 METs per week). Consequently, the men in our study represent a much ‘healthier’ sample relative to the findings of the US study, which may explain the lack of associations with health risk in our study. Additionally, the small proportion of our sample skipping breakfast means statistical power was reduced for comparison. Furthermore, our study could not differentiate between different types of breakfast consumption whereas the NHANES study was able to compare the associations of consuming different types of breakfast on outcome measures. Also, our study focussed on men, while the NHANES included both genders.

It is therefore possible that the smaller sample size of the current study and the low prevalence of breakfast skippers (8.9% compared to 26.1% in NHANES) may have impacted on our ability to find significant associations. It is also possible that in our study habitual breakfast habits and timing of breakfast consumptions were not associated with anthropometric and metabolic variables because this sample was composed mostly of healthy men. Although meeting many of the anthropometric and cardiovascular risk factor guidelines, on average these men were overweight and they did not meet the recommended cholesterol levels by the NVDPA in 2012 (NVDPA 2012), despite being relatively young. A more heterogeneous sample may have provided better opportunity to find a protective association of breakfast consumption on anthropometry and risk factor parameters. Also, although the men were overweight on average, their excess weight may have been muscle rather than body fat given their relatively high physical activity levels and healthy waist to hip ratios. It is also possible that age is a factor in the current study. All the men in the ‘Typical Aussie bloke’ study were relatively young (aged 18 – 44 years), and a greater age range with the inclusion of older men may have given us a better chance of finding significant associations, as adverse anthropometric and cardiovascular profiles are more prevalent amongst an older age groups. Therefore, this homogeneity within the sample outlines the need of more studies to be conducted in Australian men population with a larger and more heterogeneous sample with regards to BMI and health characteristics.

The lack of significant findings between habitual breakfast habits and timing of this meal consumption in our study also suggests that investigation of the association between breakfast habits, body size and metabolic parameters may need to include multiple years or time points in order to identify associations. This concept is supported by the findings of Chapters 3 (Quatela, Callister et al. 2017) and 4 (Quatela, Callister et al. 2018)of this thesis, which report that certain types of breakfast cereal consumption were significantly associated with a

174

reduction in diabetes and obesity risk over 12 years of follow-up in an older and very much larger group of women (ALSWH 1946-51 mid-age cohort). Taken together, this suggests the need to investigate the association of breakfast consumption over multiple years in male cohorts in order to properly ascertain the role of breakfast habits, timing of its consumption and breakfast cereal intake on long term risk of developing obesity and chronic diseases, and identify any gender differences.

Another possibility is that the type of breakfast consumed may have a more significant impact on anthropometric and metabolic parameters than just the consumption of breakfast per se. This study was unable to compare different food patterns because in this sample the majority of men consumed similar foods and beverages for breakfast (e.g., cereal with milk and/or fruit and/or coffee, please see Chapter 6). This idea is supported by the findings of Deshmukh-Taskar, Nicklas et al. 2013 who found that the breakfast eaters consuming ready to eat cereals were the ones with the lowest likelihood of developing obesity and with the best metabolic parameters in comparison to the other breakfast eaters and breakfast skippers (Deshmukh-Taskar, Nicklas et al. 2013). Therefore, further studies with large sample sizes and wider dietary patterns are needed to compare the impact of different breakfast compositions on anthropometric and metabolic parameters and chronic disease risks in Australian men.

In this study, a significantly higher proportion of HBE had 5 or more eating events per day in comparison to OBE and HBS who were more likely to have 1-4 eating events per day (p=0.015). This comparison indicated that HBE eat more times during the day than OBE and HBS, meaning that that HBE could potentially consume more energy in a day than OBE or HBS. Energy intake could not be ascertained in this study, however the majority of the trials investigating energy intake (EI) do suggest that eating breakfast increases EI during the day (Astbury, Taylor et al. 2011, Halsey, Huber et al. 2012, Levitsky and Pacanowski 2013, Betts, Richardson et al. 2014, Reeves, Huber et al. 2014, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017). This then leads to the suggestion that a lower daily EI may not be the mechanism for any protective role of breakfast in reducing the risk of obesity and chronic disease development. PA levels did not differ between HBE, OBE and HBS, nor between EBE and LBE in this study. There is conflicting evidence regarding the effect of breakfast on PA. Betts et al. recruited a mix of habitual and non-habitual breakfast eaters and normal weight men and this study found a significant increase in PA amongst the breakfast intervention group (Betts, Richardson et al. 2014). However, two other randomised

175

controlled trials did not find an effect on PA levels (Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017). One trial recruited obese men, the other normal weight women but in both trials the majority of whom were non-habitual breakfast eaters (Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017). These findings therefore suggest that gender or body size and/or habitual breakfast consumption or a combination of these factors may impact on the effect of breakfast consumption on PA. More studies are needed to clarify if a true significant effect exists between breakfast consumption and PA habits and if this ultimately impacts the risk of becoming obese and developing chronic disease.

Because some significant differences were found between HBE, OBE and HBS on the number of eating events consumed per day, this study also compared men having 1-4 eating events (LFE) with men having 5 or more eating event in a day (FE) with regards to body size measurements, metabolic parameters, PA, dietary habits and socio-demographic characteristics. This comparison showed that the number of eating events was not significantly associated with any of these variables. However, similar to the other groupings considered above, the author considers that adequately powered long term longitudinal studies are also needed to further explore this topic.

High fruit and vegetable intakes are well known to be associated with reduced risk of obesity and chronic disease (World Health Organization (WHO) 2017), therefore these variables were also analysed. Breakfast consumption habits, timing of its consumption and eating frequency were not associated with the number of serves of fruit and vegetables per day, suggesting that breakfast habits and eating frequency had little impact on daily servings of fruit and vegetables. This study has a number of strengths. First of all, to the author’s knowledge, this study is the first to report evidence about current habitual breakfast habits and timing of breakfast consumption in Australian men. The previous publication of data on habitual breakfast habits was based on the NNS in 1995 and this study provides recent data in an area where evidence is limited. Furthermore, this chapter investigated habitual breakfast consumption, timing of breakfast consumption, and eating frequency, in relation to a range of anthropometric and metabolic parameters which has not been investigated in Australia previously. This study utilised a diverse range of recruitment strategies (fliers, conventional and social media) and multiple data collection sites (Newcastle and Tamworth). However, the study also had some limitations with regards to the sample. Firstly, this study recruited a homogenous sample

176

regarding breakfast habits, as mostly were habitual breakfast eaters, and occupation, as most were professional, paraprofessional, managers or administrators, administrative assistants, sales or personal service workers. It is possible that because the recruitment fliers indicated that the study was investigating breakfast and its relationships with health parameters, this has potentially biased the recruitment to subjects who habitually consumed breakfast and were interested in the health benefits of this behaviour. Moreover, the relatively small convenience sample does not allow these results to be generalized with confidence to the Australian male population. Furthermore, some of the evidence collected was by means of a self-reported online survey. However, the majority of the questions, including demographics, short diet and PA questions, have been widely used in Australia by large national surveys such as ALSWH (ALSWH 2017), Australian Bureau of Statistics (ABS) in 2011 (ABS 2011) or NNS in 1995 (ABS 1997), and therefore, these questions can be considered well established methods to collect this self-reported information. This was a cross sectional study design and thus, it was only possible to investigate associations and not causation. In addition, measurement errors could have occurred during the measurement session. However, in order to reduce the likelihood of measurement errors, the measurements were all performed by the same researcher and were repeated multiple times until variation between measurements met pre-defined standards.

7.5 Conclusion Overall, this study found a significant association between habitual breakfast consumption and number of daily eating events. No other significant relationships were found between habitual breakfast consumption, timing of breakfast consumption or daily eating frequency, with intermediates of obesity and chronic disease risk amongst a generally healthy young sample of Australian men. Future studies are needed to investigate the long term association of habitual breakfast consumption and eating frequency amongst both young and middle aged Australian men in order to investigate the longer term relationships.

177

Table 7.1 Anthropometric and metabolic characteristics by habitual breakfast habits.

Habitual Occasional Habitual P value. Breakfast Breakfast Eaters Breakfast Eaters (HBE) (OBE) Skippers (HBS) (n=94) (n=7) (n=10)

Age y (mean (SD)) 30 ± 7 25 ± 5 31 ± 6 0.187 Body size and composition Height (cm) 179. 8 ± 6.6 177. 0 ± 5.0 180.5 ± 5.3 0.483 Weight (kg) 82.5 ± 14.7 80 ± 14.4 84.2 ± 11.7 0.840 BMI 25.5 ± 4.2 25.5 ± 3.8 26.0 ± 4.3 0.953 Chest circumference (cm) 97.5 ± 8. 1 97.9 ± 8.4 98.8 ± 7.8 0.876 Waist circumference (cm) 84.1± 10.0 84.8 ± 13.8 86.8 ± 11.2 0.734 Hip circumference (cm) 100.2 ± 7.6 100.8 ± 8.6 100.6 ± 7.1 0.976 Waist to hip ratio 0.8 ± 0.06 0.8 ± 0.07 0.9 ± 0.08 0.505 αFat mass (kg) *14 (7.1) 18.5 (24.4) *12.4 (18.8) 0.773 Fat mass (%) *17.8 ± 6.7 19.3 ±7.6 *17.7 ± 7.8 0.862 αVisceral fat area (cm2) *60.6 (38.1) 87.4 (73.7) *59.9 (76.4) 0.872 Skeletal muscle mass (kg) *38.4 ± 4.8 36.4 ± 3.4 *39.6 ± 3.3 0.376 Blood pressure Peripheral systolic BP 117 ± 12 122 ±14 117 ± 9 0.612 Peripheral diastolic BP 71 ± 7 74 ± 4 72 ± 6 0.415 Energy metabolism RMR kcal/d *2038 ± 627 *1932 ± 413 1965 ± 510 0.917 Finger stick blood analysis Glucose (mmol/l) 4.8 ± 0.5 4.9 ± 0.5 4.9 ± 0.4 0.710 LDL (mmol/L) *2.7 ± 0.9 2.4 ± 1.2 *3.1 ± 1.2 0.388 HDL cholesterol (mmol/L) 1.4 ± 0.3 1.4 ± 0.4 *1.3 ± 0.4 0.935 Total cholesterol (mmol/L) *4.6 ± 1.0 4.3 ± 1.1 *5.4 ± 1.1 0.071 αTriglycerides (mmol/L) *0.9 (0.5) 0. 8 (0.2) *1.1 (1.3) 0.172 Data are presented as mean ± SD unless otherwise indicated. HBE = Habitual Breakfast Eaters who reported to consume breakfast 5 or more times per week OBE = Occasional Breakfast Eaters who reported to consume breakfast 3 to 4 times per week. HBS = Habitual Breakfast Skippers who reported to consume breakfast ‘rarely or never’ or 1 to 2 times per week.’ αData presented as median (IQR) *Between 1 to 5 participants had missing values for this measurement Guidelines for anthropometric and metabolic variables: <0.9 waist to hip ratio (DAA 2015) <94 cm waist circumference (NHF 2014) 4-6 mmol/L glucose levels (DA 2008) >1.00 mmol/L HDL cholesterol (NVDPA 2012) triglycerides (<2.00 mmol/L) (NVDPA 2012) systolic≤120; diastolic≤80 mmHg blood pressure (NHF and CSAN 2005) <4.00 mmol/L Total cholesterol concentration (National Vascular Disease Prevention Alliance (NVDPA 2012) <2.00 mmol/L LDL cholesterol (NVDPA 2012)

178

Table 7.2 Lifestyle characteristics by habitual breakfast habits.

Habitual Occasional Habitual P value Breakfast Breakfast Eaters Breakfast Eaters (HBE) (OBE) Skippers (HBS) (n=94) (n=7) (n=10) Age groups (y) 18-24 y 25.5% 42.9% 20.0% 25-33 y 41.5% 57.1% 40.0% 34-44 y 33.0% 0 40.0% 0.387 BMI categories (kg/m2) Underweight BMI <20 4.3% 0 0 Normal weight BMI>=20 & BMI<25 44.7% 42.9% 50.0% Overweight BMI >=25 & BMI<3 41.5% 42.9% 30.0% Obese BMI >=30 9.6% 14.3% 20.0% 0.874 1Physical Activity level (MET min/week) Sedentary <40 2.1% 0 0 Low 40-599 7.5% 28.6% 0 Mod 600-1199 22.3% 0 20% High >1200 68.1% 71.4% 80% 0.423 Waking time Waking up ≤ 8.00 AM weekdays 91.5% 71.4% 90.0% Waking up >8.00 AM weekdays 8.5% 28.6% 10.0% 0.197 Sleeping time Going to sleep≤ 11.00 PM weekdays 73.4% 42.9% 70% Going to sleep >11.00 PM weekdays 26.6% 57.1% 30% 0.231 Eating events per day 1-4 times per day *38.3% 71.4% 80% 5 to 7 or more per day *59.6% 28.6% 20% 0.015 Fruits serves per day Don't eat fruit 2.1% 0 10.0% Less than one serve 9.6% 14.3% 0 1 serve 28.7% 42.9% 50.0% 2 serves 28.7% 28.6% 40.0% ≥3 serves 30.9% 14.3% 0 0.207 Vegetables serves per day Don't eat veggies 2.1% 0 0 Less than one serve 4.3% 0 0 1 serve 6.4% 0 30.0% 2 serves 27.7% 57.1% 30.0% 3 serves 27.7% 28.6% 30.0% ≥4 serves 31.9% 14.3% 10% 0.402 Data are presented as percentage. *The total number of subjects for this variable do not add up to 100% due to missing data reported (such as ‘replied prefer not to answer’ or ‘don’t know’ Or don’t know/Varies’ or ‘don’t know or would rather not say’). PA= Physical Activity HBE = Habitual Breakfast Eaters who reported to consume breakfast 5 or more times per week. OBE = Occasional Breakfast Eaters who reported to consume breakfast 3 to 4 times per week. HBS = Habitual Breakfast Skippers who reported to consume breakfast ‘rarely or never’ or 1 to 2 times per week.’ 1PA groups are based on total Mets min/week for leisure activities 179

Table 7.3 Anthropometric and metabolic characteristics by timing of breakfast consumption

Early Breakfast Late Breakfast P value Eaters (EBE)1 Eaters (LBE)2 (n=79) (n=15)

Age (y) 31 ± 7 26 ± 7 0.0124

Body size and composition Height (cm) 179.8 ± 6.3 179.8 ± 8.0 0.992 Weight (kg) 81.9 ± 13.5 85.5 ± 20.4 0.384 BMI (kg/m2) 25.4 ± 4.1 26.5 ± 4.8 0.356 Chest circumference (cm) 97.3 ±7.9 98.3 ± 9.2 0.652 Waist circumference (cm) 83.9 ± 9.3 84.9 ± 13.1 0.723 Hip circumference (cm) 99.9 ± 7.4 101.7 ± 8.6 0.415 Waist to hip ratio 0.8 ± 0.06 0.8 ± 0.06 0.673 αFat mass (kg) *14 (7.2) 14 (5.3) 0.816 Fat mass (%) *17.7 ± 6.7 18.8 ± 7.2 0.566 Visceral fat area (cm2) *65.5 ± 41.3 73.9 ± 53.2 0.497 Skeletal muscle mass (kg) *38.2 ± 4.6 39.2 ± 5.7 0.485 Blood pressure Peripheral systolic BP (mm Hg) 117 ± 12 116 ± 11 0.649 Peripheral diastolic BP (mm Hg) 71 ± 7 70 ± 8 0.780 Energy metabolism RMR (kcal/d) *2044 ± 629 * 2004 ± 641 0.920 Finger stick blood analysis Glucose (mmol/l) 4.8 ± 0.5 4.9 ± 0.3 0.191 LDL (mmol/L) *2.7 ± 0.9 *2.4 ± 0.8 0.272 HDL cholesterol (mmol/L) 1.4 ± 0.3 1.4 ± 0.5 0.636 Total cholesterol (mmol/L) 4.6 ± 1.0 *4.3 ± 1.0 0.223 αTriglycerides (mmol/L) *0.9 (0.5) 0.8 (0.4) 0.129 Data are presented as mean ± SD unless otherwise indicated. 1 EBE= Early Breakfast Eaters includes Habitual Breakfast Eaters who consumed the first meal of the day early on weekdays (between 5.01 and 8.00 am). 2 LBE= Late Breakfast Eaters includes Habitual Breakfast Eaters who consumed the first meal of the day late on weekdays (between 8.01 am and noon). α Data presented as median (IQR) *Between 1 to 5 participants had missing values for this measurement Guidelines for anthropometric and metabolic variables: <0.9 waist to hip ratio (DAA 2015) <94 cm waist circumference (NHF 2014) 4-6 mmol/L glucose levels (DA 2008) >1.00 mmol/L HDL cholesterol (NVDPA 2012) triglycerides (<2.00 mmol/L) (NVDPA 2012) systolic≤120; diastolic≤80 mmHg blood pressure (NHF and CSAN 2005) <4.00 mmol/L Total cholesterol concentration (National Vascular Disease Prevention Alliance (NVDPA 2012) <2.00 mmol/L LDL cholesterol (NVDPA 2012)

180

Table 7.4 Lifestyle characteristic by timing of breakfast consumption

Early Breakfast Eaters Late Breakfast Eaters P value (EBE)1 (n=79) (LBE)2 (n=15) Age groups (y) 18-24 y 20.3% 53.3% 25-33 y 44.3% 26.7% 34-44 y 35.4% 20.0% 0.048 BMI groups (kg/m2) Underweight BMI <20 5.1% 0 Normal weight BMI>=20 & BMI<25 43.0% 53.3% Overweight BMI >=25 & BMI<30 43.0% 33.3% Obese BMI >=30 8.9% 13.3% 0.734 3 Physical Activity level (MET min/week) Sedentary <40 2.5% 0 Low 40-599 8.9% 0 Mod 600-1199 21.5% 26.7% High >1200 67.1% 73.3% 0.805 Waking time Waking up ≤ 8.00 AM weekdays 100.0% 46.7% Waking up >8.00 AM weekdays 0 53.3% 0.000 Sleeping time Going to sleep≤ 11.00 PM weekdays 83.5% 20.0% Going to sleep >11.00 PM weekdays 16.5% 80.0% 0.000 Eating events per day 1-4 times per day *35.4% 53.3% 5 to 7 or more per day *62.0% 46.7% 0.218 Fruits serves per day Don't eat fruit 1.3% 6.7% Less than one serve 8.9% 13.3% 1 serve 27.9% 33.3% 2 serves 30.4% 20.0% ≥3 serves 31.6% 26.7% 0.502 Vegetables serves per day Don't eat veggies 1.3% 6.7% Less than one serve 5.1% 0 1 serve 6.3% 6.7% 2 serves 27.9% 26.7% 3 serves 24.1% 46.7% ≥4 serves 35.4% 13.3% 0.189 Data are presented as percentage. The total number of subjects for this variable do not add up to 100% due to missing data reported (such as ‘replied prefer not to answer’ or ‘don’t know’ Or don’t know/Varies’ or ‘don’t know or would rather not say’). PA = Physical Activity 1EBE = Early Breakfast Eaters includes Habitual Breakfast Eaters who consumed the first meal of the day early on weekdays (between 5.01 and 8.00 am). 2LBE =Late Breakfast Eaters includes Habitual Breakfast Eaters who consumed the first meal of the day late on weekdays (between 8.01 am and noon). 181

3PA groups are based on total Mets for leisure activities

182

Table 7.5 Anthropometric and metabolic characteristics by eating frequency.

Less Frequent Frequent Eaters P value Eaters (LFE)1 (FE) (n=60)2 (n=50)

Age (y) 29.3 ± 6.7 30.6 ± 7.0 0.331 Body size and composition Height (cm) 179.8 ± 5.9 179.5 ± 6.8 0.834 Weight (kg) 81.3 ± 15.1 83.0 ± 14.0 0.545 BMI (Kg/m2) 25.2 ± 4.4 25.7 ± 4.1 0.518 Chest circumference (cm) 96.3 ± 8.4 98.4 ± 7.8 0.184 Waist circumference (cm) 83.6 ± 11.1 84.7 ± 9.6 0.584 Hip circumference (cm) 99.6 ± 7.9 100.6 ± 7.5 0.472 Waist to hip ratio 0.8 ± 0.06 0.8 ± 0.06 0.787 αFat mass (kg) *14.2 (9.2) *13.0 (7.2) 0.657 Fat mass (%) *18.4 ± 7.0 *17.5 ± 6.8 0.526 αVisceral fat area (cm2) *59.9 (39.8) *58.9 (41.1) 0.764 Skeletal muscle mass (kg) *37.5 ± 4.7 *38.9 ± 4.7 0.131 Blood pressure Peripheral systolic BP (mm Hg) 116 ± 12 118 ± 12 0.487 Peripheral diastolic BP (mm Hg) 71 ± 7 71 ± 7 0.893 Energy metabolism RMR (kcal/d) *1964 ± 496 *2096 ± 679 0.268 Finger stick blood analysis Glucose (mmol/l) 4.8 ± 0.4 4.8 ± 0.5 0.946 LDL (mmol/L) *2.7 ± 1.0 *2.7 ± 0.9 0.983 HDL cholesterol (mmol/L) *1.4 ± 0.4 1.4 ± 0.3 0.517 Total cholesterol (mmol/L) *4.7 ± 1.1 4.6 ± 1.1 0.593 αTriglycerides (mmol/L) *0.9 (0.4) *0.9 (0.5) 0.562 Data are presented as mean ± SD unless otherwise indicated. α Data presented as median (IQR) * Between 1 to 5 participants had missing values for this measurement 1FE = Frequent Eaters who consumed 5 or more eating events per day 2LFE = Less Frequent eaters who consumed 1-4 eating events per day. Guidelines for anthropometric and metabolic variables: <0.9 waist to hip ratio (DAA 2015) <94 cm waist circumference (NHF 2014) 4-6 mmol/L glucose levels (DA 2008) >1.00 mmol/L HDL cholesterol (NVDPA 2012) triglycerides (<2.00 mmol/L) (NVDPA 2012) systolic≤120; diastolic≤80 mmHg blood pressure (NHF and CSAN 2005) <4.00 mmol/L Total cholesterol concentration (National Vascular Disease Prevention Alliance (NVDPA 2012) <2.00 mmol/L LDL cholesterol (NVDPA 2012)

183

Table 7.6 Socio-demographic and lifestyle characteristics by eating frequency.

Less Frequent Frequent Eaters P value Eaters (FE)2 (n=60) (LFE)1(n=50) Country of birth Australia 84.0% 90.0% 1Other countries 16.0% 10.0% 0.398 ATSI No 98.0%* 98.3% yes 0* 1.7% 1.000 Marital 2Single 56.0%* 38.3% Married/de facto 42.0%* 61.7% 0.050 Education 3School 44.0%* 40.0% 4University 54.0%* 60.0% 0.607 Income ≤$25,000 to 49.999 30.0%* 30.0%* ≥$50.000 to 99,999 30.0%* 21.7%* ≥100.000 to 149.999 12.0%* 26.7%* $150,000 to 200,000 or more 24.0%* 16.7%* 0.221 Employment status Full time paid work 58.0% 53.3% Part time paid work/casual paid work 16.0% 16.7% Studying 26.0% 25.0% Home duties/Unemployed 0 5.0% 0.555 Past/present or Future Occupation 5Trades, 10.0% 10.0% 6Professional,Paraprofessional, 90.0% 88.3% managers or administrator, administrative assistant, sales or personal service worker Never had a paid job 0 1.7% 1.000 Living arrangements Living alone 16.0%* 3.3%* 7Living with other people 82.0%* 95.0%* 0.041 Dependent children None 80.0% 65.0% One or more 20.0% 35.0% 0.082 Age groups 18-24 y 32.0% 21.7% 25-33 y 40.0% 45.0% 34-44 y 28.0% 33.3% 0.469 BMI groups (kg/m2) Underweight BMI <20 6.0% 1.7% Normal weight BMI>=20 & BMI<25 46.0% 46.7% Overweight BMI >=25 & BMI<30 34.0% 43.3%

Obese BMI >=30 14.0% 8.3% 0.442 8Physical Activity level (MET min/week) Sedentary <40 2.0% 1.7% 184

Less Frequent Frequent Eaters P value Eaters (FE)2 (n=60) (LFE)1(n=50) Low 40-599 8.0% 8.3% Mod 600-1199 26.0% 16.7% High >1200 64.0% 73.3% 0.713 Waking time Waking up ≤ 8.00 AM weekdays 84.0% 95.0% Waking up >8.00 AM weekdays 16.0% 5.0% 0.107 Sleeping time Going to sleep≤ 11.00 PM weekdays 60.0% 80.0% Going to sleep >11.00 PM weekdays 40.0% 20.0% 0.021 Fruits serve per day Don't eat fruit 2.0% 3.3% Less than one serve 12.0% 6.7% 1 serve 42.0% 23.3% 2 serves 24.0% 33.3% ≥3 serves 20% 33.3% 0.144 Vegetables serves per day Don't eat vegetables 2.0% 1.7% Less than one serve 4.0% 3.3% 1 serve 12.0% 3.3% 2 serves 34.0% 28.3% 3 serves 26.0% 28.3% ≥4 serves 22% 35% 0.411 Data are presented as percentage. *The total number of subject for this variable do not add up to 100% due to missing data reported (such as replied ‘prefer not to answer’ or ‘don’t know’ Or don’t know/Varies’ or ‘don’t know or would rather not say’) ATSI = Aboriginal or Torres strait islander descendent 1FE = Frequent Eaters who consumed 5 or more eating events per day 2LFE = Less Frequent eaters who consumed 1-4 eating events per day. PA= physical activity 1 Others country: New Zeeland, Italy, Germany, United Kingdom, Netherlands, America, Canada, Bangladesh or South Korea or unknown. 2 Single includes: separated, divorced, widowed or never married. 3 School includes: Intermediate Certificate (or equivalent) or Higher School or Leaving Certificate (or equivalent) or Trade/apprenticeship (eg. Hairdresser, Chef) or Certificate/diploma (eg. Child Care, Technician) 4 University includes: University degree or University Higher degree (eg. Grad Dip, Masters, PhD) 5 Trades includes: trades, manual workers, machine operators or drivers 6 Professional includes: professional, Paraprofessional, managers or administrator, administrative assistant, sales or personal service worker 7 Living with other people includes living with one, two, three, four or five other people 8 PA groups are based on total Mets for leisure activities

185

Chapter 8: Final Discussion and Recommendations for Future Research

8.1 Overview This chapter summarises the main findings of this PhD work in relation to the initial aims and hypotheses of the thesis. The findings of the studies are discussed in relation to the gaps in the literature, and the strengths and limitations of the included projects are reviewed. The chapter concludes by considering the implications of the findings for future research in this field.

8.2 Introduction The existing evidence regarding the purported protective association of breakfast consumption vs skipping on obesity risk and chronic disease prevention was found to be limited and inconclusive. This PhD therefore aimed to further explore the role of breakfast in the prevention of chronic disease. In order to do this, I utilised existing longitudinal dietary and health outcome data from the Australian Longitudinal Study of Women’s Health in epidemiological investigations, examined the existing literature on the impact of a breakfast meal on Diet Induced Thermogenesis (DIT) in a systematic literature review and meta-analyses, and collected detailed information on the breakfast habits and health and lifestyle variables in a sample of young adult males.

8.2.1 Overall Aims 8.2.1.1 First Aim To examine the role of breakfast and breakfast cereal consumption on the development of obesity and chronic disease risk.

8.2.1.1.1 Hypothesis Breakfast consumption will be associated with higher metabolic rates, better anthropometry, reduced metabolic risk factors, and lower incidence of chronic disease.

8.2.1.1.2 Objectives • To examine the effect and/or associations of consuming breakfast on measures of Resting Metabolic Rate (RMR) and DIT. Studies Investigating: Systematic Literature Review; ‘Typical Aussie Bloke’

186

• To examine the associations of consuming breakfast and breakfast cereal on anthropometric parameters (height, weight, BMI, waist circumference and body composition) and obesity risk. Studies Investigating: ALSWH Obesity Study; ‘Typical Aussie Bloke’ • To examine the associations of consuming breakfast on metabolic parameters (blood pressure, glucose, insulin and lipid profile). Studies Investigating: ‘Typical Aussie Bloke’ • To investigate the associations of consuming breakfast cereal on chronic disease outcomes. Studies Investigating: ALSWH Obesity Study; ALSWH Diabetes Study

8.2.1.2 Second Aim To describe typical Australian breakfast consumption habits of a sample of young Australian men and their determinants.

8.2.1.2.1 Hypothesis Breakfast consumption habits will be associated with socio-demographic parameters, such as income and living arrangements.

8.2.1.2.2 Objectives • To describe the frequency of breakfast consumption among a sample of young Australian men and its association with socio-demographic variables, work and lifestyle habits. Studies Investigating: ‘Typical Aussie Bloke’ • To describe the typical content of breakfast among a sample of young Australian men and its association with socio-demographic variables, work and lifestyle habits. Studies Investigating: ‘Typical Aussie Bloke’

These aims and objectives were investigated by undertaking four separate studies: 5) A twelve year longitudinal investigation of breakfast cereal consumption and its relationship with the risk of developing obesity among the mid-age cohort of the Australian Longitudinal Study of Women’s Health (Chapter 3 – ALSWH Obesity Study)

187

6) A twelve year longitudinal investigation of breakfast cereal consumption and its relationship with the risk of developing diabetes among the mid-age cohort of the Australian Longitudinal Study of Women’s Health (Chapter 4 – ALSWH Diabetes Study); 7) A Systematic Review, meta-analysis and meta-regression exploring the effect of breakfast meals of varying energy contents and compositions consumed after an overnight fast on Diet Induced Thermogenesis (Chapter 5); 8) A multi-centre cross-sectional study investigating habitual breakfast consumption and its relationship with anthropometric and metabolic parameters, socio-demographic and lifestyle characteristics, such as work, sleep, physical activity and dietary intakes among a sample of young Australian men (Chapters 6 and 7).

8.3 Main Findings of this Thesis This PhD has increased the evidence base regarding the associations of breakfast consumption per se, breakfasts of different types, and multiple types of breakfast cereal with risk factors related to the development of obesity and chronic disease risk. The overall findings of this PhD work do not support a beneficial association of breakfast consumption per se on DIT, BMI or chronic disease risk. However, all four studies suggest that the quality of the breakfast meal, in terms of types of foods (e.g. different breakfast cereals) and/or macronutrients composition (e.g. fat vs CHO vs protein), may be responsible for positive or negative effect of breakfast consumption on health outcomes. For example, the findings of the ALSWH analyses do not support the concept that any breakfast cereal consumption is protective against obesity and diabetes risk (Quatela, Callister et al. 2017, Quatela, Callister et al. 2018), but that certain types of breakfast cereal were found to be protective. Muesli consumed as part of an oat- based cereal category, or on its own, was found to be significantly associated with a reduction in both obesity and diabetes risk. Also, All-Bran was significantly related to a reduction in obesity risk but not diabetes risk. No other types of breakfast cereal were significantly associated with these two outcomes. It is important to note that this study only investigated ‘any’ breakfast cereal consumption vs ‘no’ consumption, and therefore did not give us information on other variations in habitual breakfast cereal consumption. These results suggest that the type of breakfast cereal consumed was an important determining factor for its role in protecting against body weight gain and 188

chronic disease risk. Therefore, the type of foods consumed at breakfast, and ultimately the quality of breakfast may be an important factor to consider when evaluating the role of breakfast consumption on obesity and chronic disease risk. The type of breakfast consumed may in fact be an important confounding factor in the research to date on the role of breakfast consumption on weight gain and other intermediate factors of obesity and chronic disease such as EI, EE and PA. This could at least in part explain some of the contradictory findings in the literature.

The type of breakfast consumption was also found to be important in the SR, meta- analysis and meta-regression (Quatela, Callister et al. 2016). This SR showed that a higher energy intake at breakfast significantly increased DIT (for every 100 kJ increase in energy intake, DIT increased by 1.1 kJ/h; p < 0.001). However, it is important to note that this increase was not clinically meaningful, suggesting that acute breakfast consumption does not play a significant role in increasing the metabolic rate in the morning. This finding is inconsistent with the long-held belief among health practitioners and the wider community of one of the benefits of breakfast. However, it was found that meals with certain macronutrient profiles did increase the DIT to a greater or lesser extent. For example, meals with a high protein or carbohydrate content had higher DIT than high fat meals, although this effect was not always significant (Quatela, Callister et al. 2016). These results indicate that the quality of the breakfast consumed could affect DIT, and higher DIT might well be an important outcome with regards to obesity prevention because even small differences in DIT every day could have a significant impact on weight gain. For instance, a cumulative imbalance of 42– 84 kJ/day on average can result in 0.5–1 kg of weight gain annually (Lean and Malkova 2016). However, the studies included in the SR provide limited evidence of the role of breakfast consumption in the longer term as they were all short-term interventions, and therefore can only explain the acute effect of consuming a meal after an overnight fast. There was a complete dearth of studies examining the effect of longer term habitual breakfast consumption on DIT and these are sorely needed to increase our understanding of the role of breakfast in weight gain and obesity risk. Therefore, we cannot say that our SR increased the evidence base regarding the effect of habitual breakfast consumption on DIT, and much more work in this area is needed.

189

Similarly, we were not able to show a protective association of habitual breakfast consumption on mediating factors for the development of chronic disease and obesity risk with the ‘Typical Aussie Bloke’ Study. While we were able to describe typical breakfasts for a sample of young adult men in Australia, the majority of our sample was Habitual Breakfast Eaters and the types of breakfasts they were consuming were relatively homogeneous. It was therefore not possible to determine whether different breakfast types and quality were associated with significant differences in health parameters because the majority of the participants were consuming similar breakfasts most of the time and the majority of them were in good health, with few parameters outside of the recommended ranges. This study was also unable to identify differences in PA for Habitual Breakfast Eaters, Occasional Breakfast Eaters and Habitual Breakfast Skippers, and therefore does not support the hypothesis that breakfast consumption may reduce the risk of developing obesity and chronic diseases by increasing energy expenditure in the form of PA. These findings therefore agree with the majority of the trials conducted in this field which have found no significant effects on PA following breakfast consumption and skipping interventions (Halsey, Huber et al. 2012, Reeves, Huber et al. 2015, Chowdhury, Richardson et al. 2016, LeCheminant, LeCheminant et al. 2017).

The first three studies of this PhD (ALSWH Obesity and Diabetes Study and systematic review) suggested that the quality of the breakfast meal or of breakfast cereal consumed may impact on obesity risk, diabetes risk and DIT. Therefore, it is possible that the type of breakfast consumed matters more than the consumption of any breakfast per se. This may in part explain why no significant differences on health outcomes, apart from frequency of daily eating events, were found between Habitual Breakfast Eaters, Occasional Breakfast Eaters and Habitual Breakfast Skippers in the ‘Typical Aussie Bloke’ study. Another possibility is that the relationship between breakfast consumption and health outcomes may only become apparent over multiple years. This possibility is supported by the findings of the ALSWH analyses which found significant associations for certain breakfast cereals consumed over a 12 year period. Another possibility is that the small number of Habitual Breakfast Skippers and Occasional Breakfast Eaters may have reduced the likelihood of finding an association. Furthermore, even though certain sociodemographic differences were found between Habitual Breakfast Eaters, Occasional Breakfast Eaters and Habitual Breakfast Skippers, the homogeneity of the 190

sample in terms of age, health and occupations, may have impacted in the ability to find significant associations between breakfast consumption and health characteristics. Therefore, future studies will need to aim to find a more heterogeneous population in order to better investigate the association of breakfast habits with health parameters.

8.4 Implications for Future Research The findings of the ALSWH analyses, the systematic review and the ‘Typical Aussie Bloke’ Study have indicated the need for significantly more high quality research on the role of breakfast consumption in obesity risk and health outcomes. In addition to more epidemiological investigations in longitudinal cohort studies, among men in particular, and with more detailed information on breakfast patterns, long-term intervention trials are needed. The systematic review identified a lot of very short term interventions studies looking at the effect of breakfast meals on DIT, but very little in the way of longer term studies. Breakfast interventions are needed in studies with reasonable sample sizes of breakfast skippers to elucidate whether changes in weight and health outcomes can be achieved with the introduction of regular breakfast consumption. Based on the findings of the SR and ALSWH analyses, it would seem that breakfast composition is important and future research needs to aim to measure breakfast components and quality in far more detail, in order to identify components of breakfast meals that may be protective. This area of research may be aided by the development and implementation of breakfast specific diet quality tools.

Summarising, this PhD suggests that future research will need to focus on:

• Longer term trials investigating the impact of different types of habitual breakfasts with varying food components, macronutrients, energy contents, etc. These need to occur in both Habitual Breakfast Eaters and Habitual Breakfast Skippers and good quality data collection needs to examine effects on body weight, daily total energy intakes, physical activity, blood parameters and measures of metabolic rate (RMR and DIT). These trials would need to consider previous breakfast habits, initial body weight status and gender in the sample population, as these factors were found to be heterogeneous in previous studies conducted and may have had impacts on findings in this field. • Multiple year longitudinal analyses of the associations of different types of breakfast on obesity and chronic disease risk are needed to build upon what we 191

have found in ALSWH women. Better elucidation of breakfast meal components, quality, amount consumed and frequency of consumption, will help to consolidate the findings in this area.

8.5 Strengths and Limitations The major strength of this PhD is that it managed to investigate in depth the two main aims of this thesis, specifically to examine the role of breakfast and breakfast cereal on the development of obesity and chronic disease risk and to describe typical breakfast consumption habits and their determinants in a sample of young Australian men. Another strength of this PhD is the robust statistical analyses that were applied, such as the longitudinal investigations of the ALSWH data, the meta-analysis and mixed model meta-regressions. This PhD addressed these aims in multiple population groups, differing in gender and age, and using multiple study designs, including robust longitudinal analyses in mid age Australian women, systematic review principles, meta- analyses and meta-regressions of level A evidence in both genders, and a multi-centre cross sectional investigation in a sample of young Australian men.

The main limitation of this PhD is that even though it has added to the evidence the role of breakfast in obesity and chronic disease risk, I have not been able to make any definitive conclusions about whether breakfast consumption per se is protective. I have found significant associations suggesting a role for oat based breakfast cereals, and perhaps All-Bran, and have determined that breakfast composition is an important determinant of the size of DIT, but was unable to identify any differences in anthropometric or health parameters for Habitual Breakfast Eaters vs Habitual Breakfast Skippers, except for number of daily eating events. These outcomes have been limited due to gaps in the existing literature, the bluntness of the dietary tools in the ALSWH, and the inability of the TAB study to recruit a heterogeneous sample. Identification of these limitations however has been extremely valuable in taking research in the area of breakfast consumption forward to plan more rigorous well designed studies and trials that can produce more definitive answers.

8.6 Conclusions In conclusion, this PhD has improved our understanding of the relationship between breakfast consumption, different types of breakfasts and risk factors associated with obesity and chronic disease risk. This PhD has not confirmed a protective role for 192

breakfast consumption per se, but has identified that breakfast composition is an important determinant in this relationship. My findings suggest that the types of breakfast foods and/or macronutrient profile may be the important factor in the role of breakfast in health, and this may be the reason for the contradictory results in the literature to date. Finally, future long term trials and longitudinal studies are needed to investigate the effect and/or associations of breakfast consumption, different types of breakfasts and breakfast cereal, on obesity and chronic disease risk in adults.

193

References Australian Bureau of Statistics (ABS (1997). National Nutrition Survey Selected Highlights. Canberra. ABS (2009-10). "Household Expenditure Survey Australia Summary of Results. 6530.0. 2011. Canberra, Australian Bureau of Statistics." ABS (2011). Census of Population and Housing: Household Form A. B. o. Statistics. Canberra, In: Statistics ABo, editor. ABS. (2013). "Obesity and chronic disease." Retrieved 14/12/2014, from http://www.abs.gov.au/ausstats/[email protected]/Lookup/4364.0.55.005Chapter1052011-12. ABS. (2013). "Overweight and obesity." Retrieved 05/06/2017, from http://www.abs.gov.au/ausstats/[email protected]/Lookup/33C64022ABB5ECD5CA257B82001 79437?opendocument. ABS ( 2006). Australian Social Trends. Australian Bureau of Statistics. Canberra. 2017. American Dietetic Association (ADA). (2010). "smart snacking for adults and teens." Retrieved 16-11-17, from https://nahnnet.org/docs/html/2014_Muevete_USA/Smart_Snacking_for_Adults_and- Teens-Muevete_2014.pdf. Affenito, S. G., D. R. Thompson, B. A. Barton, D. L. Franko, S. R. Daniels, E. Obarzanek, G. B. Schreiber and R. H. Striegel-Moore (2005). "Breakfast Consumption by African-American and White Adolescent Girls Correlates Positively with Calcium and Fiber Intake and Negatively with Body Mass Index." J Am Diet Assoc 105(6): 938- 945. Australian Institue of Health and Welfare (AIHW). (2016). "Overweight and obesity." Retrieved 19-07-16, from http://www.aihw.gov.au/overweight-and-obesity/. AIHW. (2017). "Chronic diseases." Retrieved 01-04-17, from http://www.aihw.gov.au/chronic-diseases/. Ainsworth, B. E., W. L. Haskell, A. S. Leon, D. R. Jacobs, H. J. Montoye, j. F. Sallis and R. S. Paffenbarger (1993). "Compendium of Physical Activities: classification of energy costs of human physical activities." Med Sci Sports 25(1): 71-80. Albertson, A. M., S. G. Affenito, R. Bauserman, N. M. Holschuh, A. L. Eldridge and B. A. Barton (2009). "The Relationship of Ready-to-Eat Cereal Consumption to Nutrient Intake, Blood Lipids, and Body Mass Index of Children as They Age through Adolescence." J Am Diet Assoc 109(9): 1557-1565.

194

Albertson, A. M., S. G. Affenito, R. Bauserman, N. M. Holschuh, A. L. Eldridge and B. A. Barton (2009). "The Relationship of Ready-to-Eat Cereal Consumption to Nutrient Intake, Blood Lipids, and Body Mass Index of Children as They Age through Adolescence." J Am Diet Assoc 109(9): 1557-1565. Albertson, A. M., D. Thompson, D. L. Franko, N. M. Holschuh, R. Bauserman and B. A. Barton (2009). "Prospective associations among cereal intake in childhood and adiposity, lipid levels, and physical activity during late adolescence." J Am Diet Assoc 109(10): 1775-1780. Albertson, A. M., D. Thompson, D. L. Franko, R. E. Kleinman, B. A. Barton and S. J. Crockett (2008). "Consumption of breakfast cereal is associated with positive health outcomes: evidence from the National Heart, Lung, and Blood Institute Growth and Health Study." Nutr Res 28(11): 744-752. Allirot, X., L. Saulais, K. Seyssel, J. Graeppi-Dulac, H. Roth, A. Charrie, J. Drai, J. Goudable, E. Blond, E. Disse and M. Laville (2013). "An isocaloric increase of eating episodes in the morning contributes to decrease energy intake at lunch in lean men." Physiol Behav 110-111: 169-178. Allman-Farinelli, M., A. Byron, C. Collins, J. Gifford and P. G. Williams (2014). "Challenges and lessons from systematic literature reviews for the Australian dietary guidelines." Aust J Prim Health 20 (3): 236-240. Australian Longitudinal study on Women's Health (ALSWH). (2017). "Australian Longitudinal Study on Women's Health: Surveys Women's Health Australia " Retrieved 05/06/17, from http://www.alswh.org.au/for-researchers/surveys. ALSWH. (2017). "Physical Activity – Survey 2 and later surveys." Retrieved 05/06/17, from http://www.alswh.org.au/images/content/pdf/InfoData/Data_Dictionary_Supplement/D DSSection2PA_S2andlater.pdf. ALSWH. (2017). "Privacy Policy – Australian Longitudinal Study of Women’s Health." Retrieved 05/06/17, from http://www.alswh.org.au/privacy-policy. ALSWH. (2017). "Progress." Retrieved 16-11-17, 2017, from http://www.alswh.org.au/about/progress. ALSWH. (2017). "surveys." Retrieved 05/06/17, from http://www.alswh.org.au/for- researchers/surveys.

195

Aranceta, J., L. Serra-Majem, L. Ribas and C. Perez-Rodrigo (2001). "Breakfast consumption in Spanish children and young people." Public Health Nutr 4(6A): 1439- 1444. Armstrong, T., A. Bauman and J. P. Davies (2000). Physical activity patterns of Australian adults: Results of the 1999 National Physical Activity Survey. Canberra, Australian Institute of Health and Welfare. Astbury, N. M., M. A. Taylor and I. A. Macdonald (2011). "Breakfast Consumption Affects Appetite, Energy Intake, and the Metabolic and Endocrine Responses to Foods Consumed Later in the Day in Male Habitual Breakfast Eaters." J Nutr 141(7): 1381- 1389. Aune, D., T. Norat, P. Romundstad and L. J. Vatten (2013). "Whole grain and refined grain consumption and the risk of type 2 diabetes: a systematic review and dose– response meta-analysis of cohort studies." Eur J Epidemiol 28(11): 845-858. Australia Diabetets (AusDiab). (2012). "The australian diabetes, obesity and lifestyle study." Retrieved April 2017. from https://www.bakeridi.edu.au/Assets/Files/Baker%20IDI%20Ausdiab%20Report_interac tive_FINAL.pdf. Bach, A. C. and V. K. Babayan (1982). "Medium-chain triglycerides: an update." Am J Clin Nutr 36(5): 950-962. Bai, J., J. J. McCluskey, H. Wang and S. Min (2014). "Dietary Globalization in Chinese Breakfasts." Can. J. Agr. Econ. 62(3): 325-341. Balvin Frantzen, L., R. P. Treviño, R. M. Echon, O. Garcia-Dominic and N. DiMarco "Association between Frequency of Ready-to-Eat Cereal Consumption, Nutrient Intakes, and Body Mass Index in Fourth- to Sixth-Grade Low-Income Minority Children." J Acad Nutr Diet 113(4): 511-519. Barr, S. B. and J. C. Wright (2010). "Postprandial energy expenditure in whole-food and processed-food meals: Implications for daily energy expenditure." Food and Nutr Res 54(5144): 1. Barr, S. I., L. DiFrancesco and V. L. Fulgoni, 3rd (2013). "Consumption of breakfast and the type of breakfast consumed are positively associated with nutrient intakes and adequacy of Canadian adults." J Nutr 143(1): 86-92. Barton, B. A., A. L. Eldridge, D. Thompson, S. G. Affenito, R. H. Striegel-Moore, D. L. Franko, A. M. Albertson and S. J. Crockett (2005). "The Relationship of Breakfast and

196

Cereal Consumption to Nutrient Intake and Body Mass Index: The National Heart, Lung, and Blood Institute Growth and Health Study." J Am Diet Assoc 105(9): 1383- 1389. Barton, B. A., A. L. Eldridge, D. Thompson, S. G. Affenito, R. H. Striegel-Moore, D. L. Franko, A. M. Albertson and S. J. Crockett (2005). "The Relationship of Breakfast and Cereal Consumption to Nutrient Intake and Body Mass Index: The National Heart, Lung, and Blood Institute Growth and Health Study." J Am Diet Assoc 105(9): 1383- 1389. Bazzano, L. A., Y. Song, V. Bubes, C. K. Good, J. E. Manson and S. Liu (2005). "Dietary intake of whole and refined grain breakfast cereals and weight gain in men." Obes Res 13(11): 1952-1960. Bendixen, H., A. Flint, A. Raben, C. E. Hoy, H. Mu, E. Xu, E. M. Bartels and A. Astrup (2002). "Effect of 3 modified fats and a conventional fat on appetite, energy intake, energy expenditure, and substrate oxidation in healthy men." Am J Clin Nutr 75(1): 47- 56. Bennett, C., G. W. Reed, J. C. Peters, N. N. Abumrad, M. Sun and J. O. Hill (1992). "Short-term effects of dietary-fat ingestion on energy expenditure and nutrient balance." Am J Clin Nutr 55(6): 1071-1077. Berry, M. N., D. G. Clark, A. R. Grivell and P. G. Wallace (1985). "The contribution of hepatic metabolism to diet-induced thermogenesis." Metabolism 34(2): 141-147. Betts, J. A., J. D. Richardson, E. A. Chowdhury, G. D. Holman, K. Tsintzas and D. Thompson (2014). "The causal role of breakfast in energy balance and health: A randomized controlled trial in lean adults." Am J Clin Nutr 100(2): 539-547. Bi, H., Y. Gan, C. Yang, Y. Chen, X. Tong and Z. Lu (2015). "Breakfast skipping and the risk of type 2 diabetes: a meta-analysis of observational studies." Public Health Nutr 18(16): 3013-3019. Blundell, J. E., J. Cooling and N. A. King (2002). "Differences in postprandial responses to fat and carbohydrate loads in habitual high and low fat consumers (phenotypes)." Br J Nutr 88(2): 125-132. Bowden, V. L. and R. G. McMurray (2000). "Effects of training status on the metabolic responses to high carbohydrate and high fat meals." Int J Sport Nutr Exerc Metab 10(1): 16-27.

197

Brown, A. W., M. M. Bohan Brown and D. B. Allison (2013). "Belief beyond the evidence: using the proposed effect of breakfast on obesity to show 2 practices that distort scientific evidence." Am J Clin Nutr 98 (5 ): 1298-1308. Brown, M. M. B., J. E. Milanes, D. B. Allison and A. W. Brown (2017). "Eating compared to skipping breakfast has no discernible benefit for obesity-related anthropometrics: systematic review and meta-analysis of randomized controlled trials." The FASEB Journal 31(1 Supplement): lb363. Burton, N. W., W. Brown and A. Dobson (2010). "Accuracy of body mass index estimated from self-reported height and weight in mid-aged Australian women." Aust N Z J Public Health 34(6): 620-623. Cahill, L. E., S. E. Chiuve, R. A. Mekary, M. K. Jensen, A. J. Flint, F. B. Hu and E. B. Rimm (2013). "Prospective study of breakfast eating and incident coronary heart disease in a cohort of male US health professionals." Circulation 128(4): 337-343. Casas-Agustench, P., P. Lopez-Uriarte, M. Bullo, E. Ros, A. Gomez-Flores and J. Salas-Salvado (2009). "Acute effects of three high-fat meals with different fat saturations on energy expenditure, substrate oxidation and satiety." Clinical Nutrition 28(1): 39-45. Casazza, K., A. Brown, A. Astrup, F. Bertz, C. Baum, M. B. Brown, J. Dawson, N. Durant, G. Dutton, D. A. Fields, K. R. Fontaine, S. Heymsfield, D. Levitsky, T. Mehta, N. Menachemi, P. K. Newby, R. Pate, H. Raynor, B. J. Rolls, B. Sen, D. L. Smith, Jr., D. Thomas, B. Wansink and D. B. Allison (2015). "Weighing the Evidence of Common Beliefs in Obesity Research." Crit Rev Food Sci Nutr 55(14): 2014-2053. Casazza, K., K. R. Fontaine, A. Astrup, L. L. Birch, A. W. Brown, M. M. Bohan Brown, N. Durant, G. Dutton, E. M. Foster, S. B. Heymsfield, K. McIver, T. Mehta, N. Menachemi, P. K. Newby, R. Pate, B. J. Rolls, B. Sen, D. L. Smith, Jr., D. M. Thomas and D. B. Allison (2013). "Myths, presumptions, and facts about obesity." N Engl J Med 368(5): 446-454. Chang, H. C., C. N. Huang, D. M. Yeh, S. J. Wang, C. H. Peng and C. J. Wang (2013). "Oat prevents obesity and abdominal fat distribution, and improves liver function in humans." Plant Foods Hum Nutr 68(1): 18-23. Chen, J., J. Cheng, Y. Liu, Y. Tang, X. Sun, T. Wang, Y. Xiao, F. Li, L. Xiang, P. Jiang, S. Wu, L. Wu, R. Luo and X. Zhao (2014). "Associations between breakfast

198

eating habits and health-promoting lifestyle, suboptimal health status in Southern China: a population based, cross sectional study." J. Transl. Med. 12(1): 348. Cho, S., M. Dietrich, C. J. Brown, C. A. Clark and G. Block (2003). "The effect of breakfast type on total daily energy intake and body mass index: results from the Third National Health and Nutrition Examination Survey (NHANES III)." J Am Coll Nutr 22(4): 296-302. Chowdhury, E. A., J. D. Richardson, G. D. Holman, K. Tsintzas, D. Thompson and J. A. Betts (2016). "The causal role of breakfast in energy balance and health: a randomized controlled trial in obese adults." Am J Clin Nutr 103(3): 747-756. Clayton, D. J. and L. J. James (2016). "The effect of breakfast on appetite regulation, energy balance and exercise performance." Proc Nutr Soc. 75(3): 319-327. Clegg, M., M. Golsorkhi and C. Henry (2013). "Combined medium-chain triglyceride and chilli feeding increases diet-induced thermogenesis in normal-weight humans." Eur J Nutr 52(6): 1579-1585. Colagiuri, S., J. J. Miller, J. L. Holliday and E. Phelan (1986). "Comparison of Plasma Glucose, Serum Insulin, and C-Peptide Responses to Three Isocaloric Breakfasts in Non-Insulin-Dependent Diabetic Subjects." Diabetes Care 9(3): 250-254. NCD Risk Factor Collaboration (NCD-RisC)(2016). "Worldwide trends in diabetes since 1980: a pooled analysis of 751 population-based studies with 4.4 million participants." The Lancet 387(10027): 1513-1530. Diabetes Australia (DA). (2008). "Blood glucose monitoring." Retrieved 16-11-17, 2017, from https://www.diabetesaustralia.com.au/Living-with-Diabetes/Type-1- Diabetes/Managing-Type-1-Diabetes/Blood-Glucose-Monitoring/. Dietitians Association of Australia (DAA). (2015). "Breakfast." Retrieved 19-07-16, from http://daa.asn.au/for-the-public/smart-eating-for-you/nutrition-a-z/breakfast/ DAA. (2015). "Overweight and obesity." Retrieved 19-07-16, from http://daa.asn.au/for-the-public/smart-eating-for-you/nutrition-a-z/overweight-and- obesity/. Deshmukh-Taskar, P., T. A. Nicklas, J. D. Radcliffe, C. E. O'Neil and Y. Liu (2013). "The relationship of breakfast skipping and type of breakfast consumed with overweight/obesity, abdominal obesity, other cardiometabolic risk factors and the metabolic syndrome in young adults. The National Health and Nutrition Examination Survey (NHANES): 1999-2006." Public Health Nutr 16(11): 2073-2082.

199

Deshmukh-Taskar, P., T. A. Nicklas, J. D. Radcliffe, C. E. O'Neil and Y. Liu (2013). "The relationship of breakfast skipping and type of breakfast consumed with overweight/obesity, abdominal obesity, other cardiometabolic risk factors and the metabolic syndrome in young adults. The National Health and Nutrition Examination Survey (NHANES): 1999–2006." Public Health Nutr 16(11): 2073-2082. Deshmukh-Taskar, P. R., T. A. Nicklas, C. E. O'Neil, D. R. Keast, J. D. Radcliffe and S. Cho (2010). "The Relationship of Breakfast Skipping and Type of Breakfast Consumption with Nutrient Intake and Weight Status in Children and Adolescents: The National Health and Nutrition Examination Survey 1999-2006." J Am Diet Assoc 110(6): 869-878. Di Giuseppe, R., A. Di Castelnuovo, C. Melegari, F. De Lucia, I. Santimone, A. Sciarretta, P. Barisciano, M. Persichillo, A. De Curtis, F. Zito, V. Krogh, M. B. Donati, G. de Gaetano and L. Iacoviello (2012). "Typical breakfast food consumption and risk factors for cardiovascular disease in a large sample of Italian adults." Nutr Metab Cardiovasc Dis 22(4): 347-354. Djousse, L. and J. M. Gaziano (2007). "Breakfast cereals and risk of heart failure in the physicians' health study I." Arch Intern Med 167(19): 2080-2085. Dodd, J. M., C. Cramp, Z. Sui, L. N. Yelland, A. R. Deussen, R. M. Grivell, L. J. Moran, C. A. Crowther, D. Turnbull, A. J. McPhee, G. Wittert, J. A. Owens and J. S. Robinson (2014). "The effects of antenatal dietary and lifestyle advice for women who are overweight or obese on maternal diet and physical activity: the LIMIT randomised trial." BMC Med 12(1): 1-19. Dubois, L., M. Girard, M. Potvin Kent, A. Farmer and F. Tatone-Tokuda (2009). "Breakfast skipping is associated with differences in meal patterns, macronutrient intakes and overweight among pre-school children." Public Health Nutr 12(01): 19-28. Dulloo, A. G., M. Fathi, N. Mensi and L. Girardier (1996). "Twenty-four-hour energy expenditure and urinary catecholamines of humans consuming low-to-moderate amounts of medium-chain triglycerides: A dose-response study in a human respiratory chamber." Eur J Clin Nutr 50(3): 152-158. Farshchi, H. R., M. A. Taylor and I. A. Macdonald (2005). "Deleterious effects of omitting breakfast on insulin sensitivity and fasting lipid profiles in healthy lean women." Eur J Clin Nutr 81(2): 388-396.

200

Farshchi, H. R., M. A. Taylor and I. A. Macdonald (2005). "Deleterious effects of omitting breakfast on insulin sensitivity and fasting lipid profiles in healthy lean women." Eur J Clin Nutr 81(2): 388-396. Geliebter A, Y. E., Haq S, Hashim SA. (2000). "Cholesterol and weight change following daily high or low fiber breakfast cereals." Obes Res 8(25S). Gibson, S. A. and P. Gunn (2011). "What's for breakfast? Nutritional implications of breakfast habits: insights from the NDNS dietary records." Nutr Bull 36(1): 78-86. Giles, G. and P. Ireland (1996 ). Dietary Questionnaire for Epidemiological Studies (Version 2),. Melbourne, Cancer Council Victoria Golay, A., B. Koellreutter, D. Bloise, J.-P. Assal and P. Würsch (1992). "The effect of muesli or cornflakes at breakfast on carbohydrate metabolism in type 2 diabetic patients." Diabetes Res Clin Pract 15(2): 135-141. Grivetti, L. E. P. D. (1995). "Morning Meals: North American and Mediterranean Breakfast Patterns: The European Anglo-Saxon Heritage." Nutr Today 30(1): 24-29. Haines, P. S., D. K. Guilkey and B. Popkin (1996). "Trends in Breakfast Consumption if US Adults Between 1965 and 1991." J Am Diet Assoc 96(5): 464-470. Halsey, L. G., J. W. Huber, T. Low, C. Ibeawuchi, P. Woodruff and S. Reeves (2012). "Does consuming breakfast influence activity levels? An experiment into the effect of breakfast consumption on eating habits and energy expenditure." Public Health Nutr 15(02): 238-245. Hamada, Y., H. Kashima and N. Hayashi (2014). "The number of chews and meal duration affect diet-induced thermogenesis and splanchnic circulation." Obesity 22(5): E62-E69. Hill, J. O., S. B. Heymsfield, C. McMannus, 3rd and M. DiGirolamo (1984). "Meal size and thermic response to food in male subjects as a function of maximum aerobic capacity." Metab. Clin. Exp. 33(8): 743-749. Hodge, A., A. J. Patterson, W. J. Brown, P. Ireland and G. Giles (2000). "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." Aust N Z J Public Health 24(6): 576-583. Horikawa, C., S. Kodama, Y. Yachi, Y. Heianza, R. Hirasawa, Y. Ibe, K. Saito, H. Shimano, N. Yamada and H. Sone (2011). "Skipping breakfast and prevalence of

201

overweight and obesity in Asian and Pacific regions: a meta-analysis." Prev Med 53(4- 5): 260-267. Howden, J. A., Y. H. Chong, S. F. Leung, L. B. Rabuco, M. Sakamoto, B. S. Tchai, K. Tontisiri, M. L. Wahlqvist, F. G. Winarno and M. Yap (1993). "Breakfast practices in the Asian region." Asia Pac J Clin Nutr 2(2): 77-84. Isa, K. A. and M. G. Masuri (2011). "The association of breakfast consumption habit, snacking behavior and body mass index among university students." Am. J. Food. Nutr, 1(2): 55-60. Kasai, M., N. Nosaka, H. Maki, Y. Suzuki, H. Takeuchi, T. Aoyama, A. Ohra, Y. Harada, M. Okazaki and K. Kondo (2002). "Comparison of diet-induced thermogenesis of foods containing medium-versus long-chain triacylglycerols." J Nutr Sci Vitaminol 48(6): 536-540. Keski-Rahkonen, A., J. Kaprio, A. Rissanen, M. Virkkunen and R. J. Rose (2003). "Breakfast skipping and health-compromising behaviors in adolescents and adults." Eur J Clin Nutr 57(7): 842-853. Kinabo, J. L. and J. V. Durnin (1990). "Effect of meal frequency on the thermic effect of food in women." Eur J Clin Nutr 44(5): 389-395. Kinabo, J. L. and J. V. G. A. Durnin (1990). "Thermic effect of food in man: Effect of meal composition, and energy content." Br J Nutr 64(1): 37-44. Kobayashi, F., H. Ogata, N. Omi, S. Nagasaka, S. Yamaguchi, M. Hibi and K. Tokuyama (2014). "Effect of breakfast skipping on diurnal variation of energy metabolism and blood glucose." Obes Res Clin Pract 8(3): e201-298. Kobayashi, F., H. Ogata, N. Omi, S. Nagasaka, S. Yamaguchi, M. Hibi and K. Tokuyama (2014). "Effect of breakfast skipping on diurnal variation of energy metabolism and blood glucose." Obes Res Clin Pract 8(3): e249-e257. Kochar, J., L. Djousse and J. M. Gaziano (2007). "Breakfast cereals and risk of type 2 diabetes in the Physicians' Health Study I." Obesity (Silver Spring) 15(12): 3039-3044. Kochar, J., J. M. Gaziano and L. Djousse (2012). "Breakfast cereals and risk of hypertension in the Physicians' Health Study I." Clin Nutr 31(1): 89-92. Kubota, Y., H. Iso, N. Sawada, S. Tsugane and J. S. Group (2016). "Association of Breakfast Intake With Incident Stroke and Coronary Heart Disease: The Japan Public Health Center-Based Study." Stroke 47(2): 477-481.

202

Lean, M. E. J. and D. Malkova (2016). "Altered gut and adipose tissue hormones in overweight and obese individuals: cause or consequence?" Int J Obes (2005) 40(4): 622-632. LeCheminant, G. M., J. D. LeCheminant, L. A. Tucker and B. W. Bailey (2017). "A randomized controlled trial to study the effects of breakfast on energy intake, physical activity, and body fat in women who are nonhabitual breakfast eaters." Appetite 112: 44-51. Lee, C., A. J. Dobson, W. J. Brown, L. Bryson, J. Byles, P. Warner-Smith and A. F. Young (2005). "Cohort Profile: The Australian Longitudinal Study on Women's Health." Int J Epidemiol 34(5): 987-991. Lejeune, M. P., K. R. Westerterp, T. C. Adam, N. D. Luscombe-Marsh and M. S. Westerterp-Plantenga (2006). "Ghrelin and glucagon-like peptide 1 concentrations, 24-h satiety, and energy and substrate metabolism during a high-protein diet and measured in a respiration chamber." Am J Clin Nutr 83(1): 89-94. Levitsky, D. A. and C. R. Pacanowski (2013). "Effect of skipping breakfast on subsequent energy intake." Physiol Behav 119(0): 9-16. Lewis, J., G. Milligan and A. Hunt (1995). NUTTAB95 Nutrient Data Table for Use in Australia. Canberra Liu, S., J. E. Manson, M. J. Stampfer, F. B. Hu, E. Giovannucci, G. A. Colditz, C. H. Hennekens and W. C. Willett (2000). "A prospective study of whole-grain intake and risk of type 2 diabetes mellitus in US women." Am J Public Health 90(9): 1409-1415. Lowe, J., J. Byles, X. Dolja-Gore and A. Young (2010). "Does systematically organized care improve outcomes for women with diabetes?" J Eval Clin Pract 16(5): 887-894. Martin, A., S. Normand, M. Sothier, J. Peyrat, C. Louche-Pelissier and M. Laville (2000). "Is advice for breakfast consumption justified? Results from a short-term dietary and metabolic experiment in young healthy men." Br J Nutr 84(3): 337-344. McNaughton, S. A., G. D. Mishra and E. J. Brunner (2008). "Dietary Patterns, Insulin Resistance, and Incidence of Type 2 Diabetes in the Whitehall II Study." Diabetes Care 31(7): 1343-1348. Mekary, R. A., E. Giovannucci, L. Cahill, W. C. Willett, R. M. van Dam and F. B. Hu (2013). "Eating patterns and type 2 diabetes risk in older women: breakfast consumption and eating frequency." Am J Clin Nutr 98(2): 436–443.

203

Mekary, R. A., E. Giovannucci, W. C. Willett, R. M. van Dam and F. B. Hu (2012). "Eating patterns and type 2 diabetes risk in men: breakfast omission, eating frequency, and snacking." Am J Clin Nutr 95(5): 1182-1189. Meksawan, K., S. Pongthananikorn and L. Tongyonk (2014). "Breakfast skipping is a risk factor for metabolic syndrome (641.2)." The FASEB Journal 28(1 Supplement). Meltzer, H. M., A. L. Brantsæter, T. A. Ydersbond, J. Alexander and M. Haugen (2008). "Methodological challenges when monitoring the diet of pregnant women in a large study: experiences from the Norwegian Mother and Child Cohort Study (MoBa)." Matern Child Nutr 4(1): 14-27. Mesas, A. E., M. Munoz-Pareja, E. Lopez-Garcia and F. Rodriguez-Artalejo (2012). "Selected eating behaviours and excess body weight: a systematic review." Obes Rev 13(2): 106-135. Meyer, K. A., L. H. Kushi, D. R. Jacobs, J. Slavin, T. A. Sellers and A. R. Folsom (2000). "Carbohydrates, dietary fiber, and incident type 2 diabetes in older women." Am J Clin Nutr 71(4): 921-930. Monteagudo, C., A. Palacín-Arce, M. d. M. Bibiloni, A. Pons, J. A. Tur, F. Olea- Serrano and M. Mariscal-Arcas (2013). "Proposal for a Breakfast Quality Index (BQI) for children and adolescents." Public Health Nutr 16(04): 639-644. Mullan, B. A. and M. Singh (2010). "A systematic review of the quality, content, and context of breakfast consumption." Nagai, N., N. Sakane and T. Moritani (2005). "Metabolic responses to high-fat or low- fat meals and association with sympathetic nervous system activity in healthy young men." J Nutr Sci Vitaminol 51(5): 355-360. National Heart Fundation (NHF). (2014). "waist circumference." from http://www.heartfoundation.org.au/healthy-eating/Pages/waist-measurement.aspx. National Heart, Lung and Heart Institute = National Health Institute (NHLBI-NIH). (2014). "Balance food and activity." Retrieved 15/03/2015, from https://www.nhlbi.nih.gov/health/public/heart/obesity/wecan/healthy-weight- basics/balance.htm. NHMRC. (2005). "Nutrient Reference Values for Australia and New Zealand Including Recommended Dietary Intakes." Retrieved 05/06/17, from https://www.nhmrc.gov.au/guidelines-publications/n35-n36-n37.

204

NHMRC. (2009). "NHMRC additional levels of evidence and grades for recommendations for developers of guidelines." Retrieved 05/06/17, from http://www.nhmrc.gov.au/_files_nhmrc/file/guidelines/stage_2_consultation_levels_and _grades.pdf. NHMRC. (2011). "How to asses the evidence." Retrieved 05/06/17, from https://www.nhmrc.gov.au/guidelines/publications/cp69. Nicklas, T. A., C. Reger, L. Myers and C. O’Neil (2000). "Breakfast consumption with and without vitamin-mineral supplement use favorably impacts daily nutrient intake of ninth-grade students." J Adolesc Health 27(5): 314-321. Nishiyama, M., T. Muto, T. Minakawa and T. Shibata (2009). "The combined unhealthy behaviors of breakfast skipping and smoking are associated with the prevalence of diabetes mellitus." Tohoku J Exp Med 218(4): 259-264. NVDPA (2012). Quick reference guide for health professionals Absolute cardiovascular disease risk management, NVDPA. O'Neil, C. E., C. Byrd-Bredbenner, D. Hayes, L. Jana, S. E. Klinger and S. Stephenson- Martin (2014). "The Role of Breakfast in Health: Definition and Criteria for a Quality Breakfast." J Acad Nutr Diet 114(12, Supplement): S8-S26. Odegaard, A. O., D. R. Jacobs, Jr., L. M. Steffen, L. Van Horn, D. S. Ludwig and M. A. Pereira (2013). "Breakfast frequency and development of metabolic risk." Diabetes Care 36(10): 3100-3106. Odle, J. (1997). "New Insights into the Utilization of Medium-Chain Triglycerides by the Neonate: Observations from a Piglet Model." J Nutr 127(6): 1061-1067. OECD. (2014). "Obesity update." Retrieved 05/06/17, from http://www.oecd.org/health/Obesity-Update-2014.pdf. OXFAM. (2011). "Rising food prices are changing what we eat reveals new global food survey." retrieved 15-11-17, from http://www.oxfam.org/en/pressroom/pressreleases/2011-06-15/rising-food-prices-are- changing-what-we-eat-reveals-new-global. Pastorino, S., M. Richards, M. Pierce and G. L. Ambrosini (2016). "A high-fat, high- glycaemic index, low-fibre dietary pattern is prospectively associated with type 2 diabetes in a British birth cohort." Br J Nutr 115(9): 1632-1642.

205

Petzke, K. J. and S. Klaus (2008). "Reduced postprandial energy expenditure and increased exogenous fat oxidation in young woman after ingestion of test meals with a low protein content." Nutr Metab 5 (1) (no pagination)(25). Piers, L. S., M. J. Soares, T. Makan and P. S. Shetty (1992). "Thermic effect of a meal. 1. Methodology and variation in normal young adults." Br J Nutr 67(2): 165-175. Piers, L. S., K. Z. Walker, R. M. Stoney, M. J. Soares and K. O'Dea (2002). "The influence of the type of dietary fat on postprandial fat oxidation rates: monounsaturated (olive oil) vs saturated fat (cream)." Int J Obes Relat Metab 26(6): 814-821. Priebe, M. G. and J. R. McMonagle (2016). " Effects of Ready-to-Eat-Cereals on Key Nutritional and Health Outcomes: A Systematic Review." PLoS ONE 11(10): e0164931. Prisma. (2015). "Prisma flow diagram." from http://www.prismastatement.org/PRISMAStatement/FlowDiagram.aspx. PTSDiagnostic. (2017). "CardioCheck PA guideline." Retrieved 16-11-17, 2017, from http://www.ptsdiagnostics.com/uploads/2/6/2/8/26289179/ps- 002461_en_rev._4_user_guide_cardiochek_pa.pdf. Quatela, A., R. Callister, A. Patterson and L. MacDonald-Wicks (2016). "The Energy Content and Composition of Meals Consumed after an Overnight Fast and Their Effects on Diet Induced Thermogenesis: A Systematic Review, Meta-Analyses and Meta- Regressions." Nutrients 8(11): 670. Quatela, A., R. Callister, A. Patterson, M. McEvoy and L. MacDonald-Wicks (2017). "Breakfast Cereal Consumption and Obesity Risk amongst the Mid-Age Cohort of the Australian Longitudinal Study on Women’s Health." Healthcare 5(3): 49. Quatela, A., R. Callister, A. J. Patterson, M. McEvoy and L. K. MacDonald-Wicks (2018). "The protective effect of muesli consumption on diabetes risk: Results from 12 years of follow-up in the Australian Longitudinal Study on Women’s Health." Nutr Res 51: 12-20. Raben, A., L. Agerholm-Larsen, A. Flint, J. J. Holst and A. Astrup (2003). "Meals with similar energy densities but rich in protein, fat, carbohydrate, or alcohol have different effects on energy expenditure and substrate metabolism but not on appetite and energy intake." Am J Clin Nutr 77(1): 91-100.

206

Rampersaud, G. C., M. A. Pereira, B. L. Girard, J. Adams and J. D. Metzl (2005). "Breakfast habits, nutritional status, body weight, and academic performance in children and adolescents." J Am Diet Assoc 105(5): 743-760; quiz 761-742. Reed, G. W. and J. O. Hill (1996). "Measuring the thermic effect of food." Am J Clin Nutr 63(2): 164-169. Reed, G. W. and J. O. Hill (1996). "Measuring the thermic effect of food." Am J Clin Nutr 63(2): 164-169. Reeves, S., L. G. Halsey, Y. McMeel and J. W. Huber (2013). "Breakfast habits, beliefs and measures of health and wellbeing in a nationally representative UK sample." Appetite 60(0): 51-57. Reeves, S., J. W. Huber, L. G. Halsey, Y. Horabady-Farahani, M. Ijadi and T. Smith (2014). "Experimental manipulation of breakfast in normal and overweight/obese participants is associated with changes to nutrient and energy intake consumption patterns." Physiol Behav 133: 130-135. Reeves, S., J. W. Huber, L. G. Halsey, M. Villegas-Montes, J. Elgumati and T. Smith (2015). "A cross-over experiment to investigate possible mechanisms for lower BMIs in people who habitually eat breakfast." Eur J Clin Nutr 69(5): 632-637. Rendell, M., J. Vanderhoof, M. Venn, M. A. Shehan, E. Arndt, C. S. Rao, G. Gill, R. K. Newman and C. W. Newman (2005). "Effect of a Barley Breakfast Cereal on Blood Glucose and Insulin Response in Normal and Diabetic Patients." Plant Foods Hum Nutr 60(2): 63-67. Riggs, A. J., B. D. White and S. S. Gropper (2007). "Changes in energy expenditure associated with ingestion of high protein, high fat versus high protein, low fat meals among underweight, normal weight, and overweight females." Nutr. J. 6: 40. Romon, M., J. L. Edme, C. Boulenguez, J. L. Lescroart and P. Frimat (1993). "Circadian variation of diet-induced thermogenesis." Am J Clin Nutr 57(4): 476-480. Ruddick-Collins, L. C., N. A. King, N. M. Byrne and R. E. Wood (2013). "Methodological considerations for meal-induced thermogenesis: measurement duration and reproducibility." Br J Nutr 110(11): 1978-1986. Sawaya, A. L., P. J. Fuss, G. E. Dallal, R. Tsay, M. A. McCrory, V. Young and S. B. Roberts (2001). "Meal palatability, substrate oxidation and blood glucose in young and older men." Physiol Behav 72(1-2): 5-12.

207

Schlundt, D. G., J. O. Hill, T. Sbrocco, J. Pope-Cordle and T. Sharp (1992). "The role of breakfast in the treatment of obesity: a randomized clinical trial." Am J Clin Nutr 55(3): 645-651. Seaton, T. B., S. L. Welle, M. K. Warenko and R. G. Campbell (1986). "Thermic effect of medium-chain and long-chain triglycerides in man." Am J Clin Nutr 44(5): 630-634. Segal, K. R., A. Edano, L. Blando and F. X. Pi-Sunyer (1990). "Comparison of thermic effects of constant and relative caloric loads in lean and obese men." Am J Clin Nutr 51(1): 14-21. Seuring, T., O. Archangelidi and M. Suhrcke (2015). "The Economic Costs of Type 2 Diabetes: A Global Systematic Review." Pharmacoeconomics 33(8): 811-831. Siega-Riz, A. M., B. M. Popkin and T. Carson (1998). "Trends in breakfast consumption for children in the United States from 1965-1991." Am J Clin Nutr 67(4): 748S-756S. Siega-Riz, A. M., B. M. Popkin and T. Carson (2000). "Differences in Food Patterns at Breakfast by Sociodemographic Characteristics among a Nationally Representative Sample of Adults in the United States." Prev Med 30(5): 415-424. Sjoberg, A., L. Hallberg, D. Hoglund and L. Hulthen (2003). "Meal pattern, food choice, nutrient intake and lifestyle factors in The Goteborg Adolescence Study." Eur J Clin Nutr 57(12): 1569-1578. Smith, K. J., S. A. McNaughton, V. J. Cleland, D. Crawford and K. Ball (2013). "Health, Behavioral, Cognitive, and Social Correlates of Breakfast Skipping among Women Living in Socioeconomically Disadvantaged Neighborhoods." J Nutr 143(11): 1774-1784. Sugimori, H., M. Miyakawa, K. Yoshida, T. Izuno, E. Takahashi, C. Tanaka, K. Nakamura and S. Hinohara (1998). "Health risk assessment for diabetes mellitus based on longitudinal analysis of MHTS database." J Med Syst 22(1): 27-32. Sugiyama, S., M. Okuda, S. Sasaki, I. Kunitsugu and T. Hobara (2012). "Breakfast habits among adolescents and their association with daily energy and fish, vegetable, and fruit intake: a community-based cross-sectional study." Environ Health Prev Med 17(5): 408-414. Szajewska, H. and M. Ruszczynski (2010). "Systematic review demonstrating that breakfast consumption influences body weight outcomes in children and adolescents in Europe." Crit Rev Food Sci Nutr 50(2): 113-119.

208

Tai, M. M., P. Castillo and F. X. Pi-Sunyer (1991). "Meal size and frequency: Effect on the thermic effect of food." Am J Clin Nutr 54(5): 783-787. Tappy, L. (1996). "Thermic effect of food and sympathetic nervous system activity in humans." Reprod Nutr Dev 36(4): 391-397. Tappy, L., E. Gügolz and P. Würsch (1996). "Effects of Breakfast Cereals Containing Various Amounts of β-Glucan Fibers on Plasma Glucose and Insulin Responses in NIDDM Subjects." Diabetes Care 19(8): 831-834. Tessari, P., E. Kiwanuka, M. Zanetti and R. Barazzoni (2003). "Postprandial body protein synthesis and amino acid catabolism measured with leucine and phenylalanine- tyrosine tracers." Am J Physiol Endocrinol Metab 284(5): E1037-E1042. Thyfault, J. P., S. R. Richmond, M. J. Carper, J. A. Potteiger and M. W. Hulver (2004). "Postprandial metabolism in resistance-trained versus sedentary males." Med. Sci. Sports Exerc 36(4): 709-716. Timlin, M. T. and M. A. Pereira (2007). "Breakfast frequency and quality in the etiology of adult obesity and chronic diseases." Nutr Rev 65(6 Pt 1): 268-281. Toyama, K., X. Zhao, S. Kuranuki, Y. Oguri, E. Kashiwa Kato, Y. Yoshitake and T. Nakamura (2015). "The effect of fast eating on the thermic effect of food in young Japanese women." Int J Food Sci Nutr 66(2): 140-147. Tsihlias, E. B., A. L. Gibbs, M. I. McBurney and T. M. Wolever (2000). "Comparison of high- and low-glycemic-index breakfast cereals with monounsaturated fat in the long-term dietary management of type 2 diabetes." Am J Clin Nutr 72(2): 439-449. Vandarakis, D., A. J. Salacinski and C. E. Broeder (2013). "A Comparison of Cosmed Metabolic Systems for the Determination of Resting Metabolic Rate." Res Sports Med 21(2): 187-194. Vaz, M., A. Turner, B. Kingwell, J. Chin, E. Koff, H. Cox, G. Jennings and M. Esler (1995). "Postprandial sympatho-adrenal activity: its relation to metabolic and cardiovascular events and to changes in meal frequency." Clin Sci 89(4): 349-357. Voronova, N. V., A. G. Nikitin, A. P. Chistiakov and D. A. Chistiakov (2012). "Skipping breakfast is correlated with impaired fasting glucose in apparently healthy subjects." Cent Eur J Med 7(3): 376-382. Westerterp-Plantenga, M. S. (2008). "Protein intake and energy balance." Regul Pept 149(1-3): 67-69. Westerterp, K. R. (2004). "Diet induced thermogenesis." Nutr Metab (Lond) 1(1): 5.

209

Westerterp, K. R., S. A. Wilson and V. Rolland (1999). "Diet induced thermogenesis measured over 24h in a respiration chamber: effect of diet composition." Int J Obes Relat Metab 23(3): 287-292. Weststrate, J. A. (1993). "Resting metabolic rate and diet-induced thermogenesis: a methodological reappraisal." Am J Clin Nutr 58(5): 592-601. Weststrate, J. A. (1993). "Resting metabolic rate and diet-induced thermogenesis: A methodological reappraisal." Am J Clin Nutr 58(5): 592-601. Weststrate, J. A., T. Dopheide, L. Robroch, P. Deurenberg and J. G. Hautvast (1990). "Does variation in palatability affect the postprandial response in energy expenditure?" Appetite 15(3): 209-219. Weststrate, J. A., P. J. Weys, E. J. Poortvliet, P. Deurenberg and J. G. Hautvast (1989). "Diurnal variation in postabsorptive resting metabolic rate and diet-induced thermogenesis." Am J Clin Nutr 50(5): 908-914. Wolrd Health Organization (WHO). (2016). "BMI classification." Retrieved 15-11-17, from http://apps.who.int/bmi/index.jsp?introPage=intro_3.html. WHO (2016). Global Report on Diabetes. WHO. (2016). "Overweight and Obesity." Retrieved 13/08/15, from http://www.who.int/gho/ncd/risk_factors/overweight_text/en/. WHO (2017). "10 facts on physical activity." WHO. (2017). "Diet, nutrition and the prevention of chronic diseases Report of the joint WHO/FAO expert consultation." Retrieved 13/08/15 from http://www.who.int/dietphysicalactivity/publications/trs916/summary/en/. WHO. (2017). "Noncommunicable diseases." Retrieved April 2017, from http://www.who.int/mediacentre/factsheets/fs355/en/. WHO. (2017). "obesity and overweight." Retrieved 05/06/17, from http://www.who.int/mediacentre/factsheets/fs311/en/. WHO. (2017). "Overweight and Obesity." Retrieved 05/06/17, from http://www.who.int/gho/ncd/risk_factors/overweight_text/en/. Williams, P. (2002). "What Australians eat for breakfast: an analysis of data from the 1995 National Nutrition Survey." Nutr Diet 59(2): 103-112. Williams, P. G. (2014). "The Benefits of Breakfast Cereal Consumption: A Systematic Review of the Evidence Base." Adv. Nutr. 5(5): 636S-673S.

210

Williams, P. G. (2014). "The Benefits of Breakfast Cereal Consumption: A Systematic Review of the Evidence Base." Adv Nutr 5(5): 636S-673S. Wilson, N. C., W. R. Parnell, M. Wohlers and P. M. Shirley (2006). "Eating breakfast and its impact on children's daily diet." Nutr Diet 63(1): 15-20. Xiao, H., J. Wang and X. Zhi (2010). "A case-control study on risk factors of type 2 diabetes mellitus in suburban of Tianjin." Chin J Dis Control Prev 14: 95-98. Xu, M., T. Huang, A. W. Lee, L. Qi and S. Cho (2016). "Ready-to-Eat Cereal Consumption with Total and Cause-Specific Mortality: Prospective Analysis of 367,442 Individuals." J Am Coll Nutr 35(3): 217-223. Yokoyama, Y., K. Onishi, T. Hosoda, H. Amano, S. Otani, Y. Kurozawa and A. Tamakoshi (2016). "Skipping Breakfast and Risk of Mortality from Cancer, Circulatory Diseases and All Causes: Findings from the Japan Collaborative Cohort Study." Yonago Acta Med 59(1): 55-60. Zealand, N. H. F. o. A. a. t. C. S. o. A. a. N. (2005). "Position statement on lipid management." Heart Lung Circ 14(4): 275-291. Zhi, X. Y. (2007). A study of prevalences and risk factors of type 2 diabetes and its complications in Tianjin. PhD Thesis, Tianjin Medical University.

211

Appendix 1: Abstracts for Chapter 3 • Abstract presented at the International Society of Behavioural Nutrition and Physical Activity (ISBNPA) in Victoria, Canada, 2017 (oral presentation by Patterson, A; Authors: Quatela A., A.J. Patterson, R. Callister, M. McEvoy, L.K. MacDonald-Wicks)

The effects of breakfast cereal consumption on obesity risk over 12 years among mid- aged women in the Australian Longitudinal Study on Women’s Health

Purpose: The obesity rate among Australian women was 27.5% in 2012. Breakfast cereal consumption is believed to be protective against obesity but the evidence available to support this belief is limited. Therefore, this longitudinal study aimed to investigate the effects of breakfast cereal consumption on the risk of developing obesity (BMI≥30kg/m2) over 12 years among women in the mid-age cohort of the Australian Longitudinal Study of Women’s Health (ALSWH).

Methods: Data from Survey 3 (S3) to Survey 7 (S7) inclusive, from the 1946-51 born ALSWH cohort were analysed. Dietary data (DQESv2 FFQ) were available at S3 and obesity incidence at S4-S7. Women were excluded if: dietary data were incomplete; they reported existing overweight and obesity cases; or if total energy intake was <4500 or >20,000kJ. Logistic regression models with survival analyses investigated the association between breakfast cereal intake (yes or no) and incident obesity over 12 years longitudinally. Models were adjusted for: income, area of residency, physical activity, smoking, hypertension, a discrete measure of time and dietary intakes (total energy intake (kJ/day), fibre intake (g/day) and other breakfast cereal consumption).

Results/findings: A total of 4143 women were included in the analyses. There were 308 (7.4%) incident cases of obesity. Breakfast cereal intake, regardless of type, was not associated with incident obesity (OR: 0.92; p=0.68; CI: 0.63, 1.35). All-Bran (0.67; p=0.02; CI: 0.48, 0.94), muesli (0.57; p=0.00; CI: 0.43, 0.75) and oat-based breakfast cereal (OR: 0.71; p=0.01; CI: 0.55, 0.90) consumption were associated with a strong and significant reduction in the risk of developing obesity. No other breakfast cereals were associated with a significant reduction in obesity risk.

212

Conclusions: Among mid-age Australian women, All-Bran, muesli and oat-based breakfast cereal consumption, but no other breakfast cereals intake, were associated with a significant decrease in the odds of developing obesity. This effect may be due to a particular profile of All-Bran, muesli and oat-based breakfast cereal consumers that we have not been able to fully adjust for, but these relationships warrant further investigation.

• Abstract presented at Nutrition Society of Australia in Melbourne, Australia 2016 (poster presentation by Quatela, A; Authors: Quatela A., A.J. Patterson, R. Callister, M. McEvoy, L.K. MacDonald-Wicks)

Breakfast cereal consumption and incident obesity: 12 years analyses of the Australian longitudinal study on women’s health

Background/Aims: The obesity rate among Australian women is 27.5%. Breakfast cereal consumption is thought to be protective against obesity. This study investigated the effect of breakfast cereal consumption on the risk of developing obesity (BMI≥30kg/m2) over 12 years among participants of the Australian Longitudinal Study of Women’s Health (ALSWH).

Methods: Data from Survey 3 (S3) to Survey 7 (S7) inclusive, from the 1946-51 ALSWH cohort were analysed. Dietary data (DQESv2 FFQ) were available at S3 and S7, obesity at S4-S7. Women were excluded if: dietary data were incomplete; they reported existing overweight and obesity cases; or if total energy intake was <4500 or >20,000kJ. Logistic regression models investigated the association between breakfast cereal intake (yes or no) and incident obesity. Models were adjusted for: education, income, physical activity, smoking, hypertension and dietary intakes.

Results: There were 255 (7.8%) incident cases of obesity. Total breakfast cereal intake was not associated with incident obesity (OR:0.79, p=0.284, CI:0.52, 1.21). There were no significant associations with most individual breakfast cereal types. Muesli consumption was associated with a strong and significant reduction in the risk of developing obesity (OR:0.68, p=0.014, CI:0.50, 0.92).

Conclusions: Among mid-age Australian women muesli consumption, but no other breakfast cereals, was associated with a reduction in obesity. This effect may be due to a

213

particular profile of muesli eaters that we have not be able to fully adjust for, but the relationship warrants further investigation.

Funding source(s): The University of Newcastle

214

Appendix 2: Authors Contribution for chapter 3 I attest that Research Higher Degree candidate Angelica Quatela contributed to the following published paper:

Angelica Quatela contributed to the study design, data analysis and manuscript preparation as first author. Prof Robin Callister, Dr Amanda Patterson, Dr Lesley MacDonald-Wicks, A/Prof Mark McEvoy contributed to the study design, data analysis and manuscript preparation within their capacity as PhD supervisors and/or co-authors.

Angelica Quatela 15-11-17

Dr Lesley MacDonald-Wicks 15-11-17

215

Appendix 3: Abstracts for Chapter 4 • Abstract presented at Nutrition Society of Australia in Melbourne, Australia, 2016 (poster presentation by Quatela, A; authors: Quatela A., A.J. Patterson, R. Callister, M. McEvoy, L.K. MacDonald-Wicks)

Is breakfast cereal consumption an effective strategy to prevent diabetes for mid-age Australian women?

Background/Aims: Diabetes Mellitus (DM) affects 9.8% of Australian women. Breakfast cereal consumption has been linked with better health outcomes, including for DM. This study investigated the effect of breakfast cereal consumption on the risk of developing DM among the Australian Longitudinal Study of Women’s Health (ALSWH), over 12 years.

Methods: Data from Survey 3 (S3) to Survey 7 (S7) inclusive, from the 1946-51 ALSWH cohort were analysed. Dietary data (DQESv2 FFQ) were available at S3 and S7, DM at S4-S7. Women were excluded if: dietary data were incomplete; they reported existing diabetes or IGT at S3; or if total energy intake was <4500 or >20,000kJ. Logistic regression models investigated the association between breakfast cereal intake (yes or no) and incident DM. Models were adjusted for: BMI, smoking, marital status, income, physical activity, and dietary intakes.

Results: There were 538 (8.1%) incident cases of DM. Total breakfast cereal intake was not associated with incident DM (OR:1.08, p=0.655, CI:0.76, 1.55). There were no significant associations with most individual breakfast cereal types, however women who consumed muesli had a strong and significant decrease in the odds of developing DM (OR:0.73, p=0.003, CI:0.59, 0.90).

Conclusions: Among mid-age Australian women muesli consumption, but no other breakfast cereals, was associated with a reduction in DM. This effect may be due to a particular profile of muesli eaters that we have not be able to fully adjust for, but the relationship warrants further investigation.

Funding source(s): The University of Newcastle

• Abstract presented at the Australian Longitudinal Study on Women’s Health Conference in Newcastle, Australia, 2016 (oral presentation by 216

Quatela, A; authors: Quatela A., R. Callister, A.J. Patterson, M. McEvoy, L.K. MacDonald-Wicks,)

Breakfast cereal consumption and incident Diabetes Mellitus: Results from 12 years of the Australian Longitudinal Study on Women’s Health

Objectives: Diabetes Mellitus (DM) currently affects 9.8% of Australian women. Breakfast cereal consumption has been linked with better health outcomes, including for DM. This study investigated the effect of breakfast cereal consumption on the risk of developing DM among women from the Australian Longitudinal Study of Women’s Health (ALSWH), over 12 years.

Methods: Data from Survey 3 (S3) to Survey 7 (S7) inclusive, from the 1946-51 ALSWH cohort were analysed. Dietary data (DQESv2 FFQ) were available at S3 and S7, and the outcome was incident DM at S4-S7. Women were excluded if: dietary data were incomplete; they reported existing diabetes or IGT at S3; or if total energy intake was <4500->20,000kJ. Logistic regression models were used to investigate the association between breakfast cereal intake (yes or no) at S3 and risk of developing DM. Models were adjusted for: BMI; smoking; marital status; income; dietary intakes (alcohol, fat, protein, carbohydrate, fibre, energy and other breakfast cereals).

Results: There were 538 incident cases of DM. Total breakfast cereal intake was not associated with incident DM (Odds Ratio=0.99, p value= 0.974, CI: 0.72; 1.37). There were no significant associations with most individual breakfast cereal types (All Bran; Sultana Bran, Fibre Plus & Branflakes;Weet Bix, Vita Brits & Weeties; Cornflakes, Nutrigrain & Special K; or Porridge), however, women who consumed muesli had a strong and significant decrease in the odds of developing DM (Odds ratio= 0.72, p value= 0.001, CI: 0.59; 0.88).

Conclusions: Among mid-age Australian women, total cereal consumption was not significantly associated with the development of DM over 12 years, however, the addition of muesli to their eating pattern appeared to be protective. This effect may be due to a particular profile of muesli eaters that we have not be able to fully adjust for, but the relationship warrants further investigation.

217

Appendix 4: Authors Contribution for chapter 4 I attest that Research Higher Degree candidate Angelica Quatela contributed to the following published paper:

Angelica Quatela 15-11-17

Dr Lesley MacDonald-Wicks 15-11-17

218

Appendix 5: Abstracts for Chapter 5 • Abstract presented at the ISBNPA in Edinburgh, Scotland, 2015 (poster presentation by Quatela, A; authors: Quatela A., R. Callister, A.J. Patterson, L.K. MacDonald-Wicks)

The effect of breakfast size and frequency on diet induced thermogenesis.

Purpose: A systematic review was conducted to investigate the effect of breakfast consumption on measures of resting energy expenditure (REE) amongst adults. This abstract focusses on one finding from this review: the effect of breakfast size and frequency on diet induced thermogenesis (DIT).

Methods: The Systematic Review protocol was registered in ‘Prospero'. The keywords and the inclusion and exclusion criteria were chosen a priori. Databases searched were Cochrane, Cinahl, Embase and Medline. English language intervention studies (RCT and quasi RCT) in healthy adult participants (no date limit) were included in this review. The quality criteria checklist for primary research of the Academy of Nutrition and Dietetics was used to assess the quality of included studies.

Results/findings: The initial search identified 473 records; 130 were retrieved after duplicate removals, and inclusion/exclusion criteria were applied by two reviewers independently (AQ & AP). Quality checks on 34 papers were undertaken by AQ & LMW. Of the final 34 papers included findings from two randomised crossover trials investigating the effect of breakfast size and frequency on DIT were extracted. Participants were healthy normal weight adults (20 men; 7 women respectively). Breakfast was consumed after overnight fast as either: one bolus amount over 20 or 10 minutes compared with 46 small meals for 10 minutes each consumed every hour over a four hour period (study 1), or every 30 minutes for 150 minute period (study 2). In both trials DIT was significantly higher when isocaloric meals were consumed as a bolus compared to small frequent events (study 1: DIT for 240 min was 6.2 (0.6) % vs 4.7 (0.5) % of energy content of meal (mean (SEM)); study 2: DIT for 5 hours was 241.00 (34.56) kJ/5 h vs 174.47 (25.10) kJ/5 h (mean (SEM)), both p value <0.05).

219

Conclusions: There was a significant increase in DIT associated with a bolus intake, such as breakfast, in a small time frame, compared to smaller, isocaloric, frequent events over a longer time period, such as snacking. This finding may be used to inform the study of weight gain prevention in a long term prospective trial.

• Abstract presented at the ISBNPA in Edinburgh, Scotland, 2015 (poster presentation by Quatela, A; authors: Quatela A., R. Callister, A.J. Patterson, L.K. MacDonald-Wicks )

What it is not known of the effect of fat intake at breakfast on DIT.

Purpose: Systematic literature review (SLR) to investigate effect of breakfast consumption on measures of REE amongst adults. This abstract reports one aspect of the review: effect of fat intake at breakfast on diet induced thermogenesis (DIT).

Methods: SLR protocol was registered in Prospero. Inclusion criteria and keywords were chosen a priori. Databases searched were Cochrane, Cinahl, Embase, Medline. English language intervention studies with healthy adults were included; no date limit applied. Inclusion criteria were applied by two reviewers. Quality of included papers was assessed by two reviewers using the Academy of Nutrition and Dietetics quality criteria checklist.

Results/findings: Initial search identified 473 records; 34 remained once duplicates removed and inclusion and quality criteria applied. Data from five trials investigating role of dietary fat consumption at breakfast on DIT were extracted. All studies were in men and used indirect calorimetry. The five studies were heterogeneous in all aspects of study design, one was rated positive quality, the remainders were neutral. Two studies compared type of fat: the positive quality study was a randomized cross over trial (RCOD) of an isocaloric (2.5MJ) high MUFA vs high SFA breakfast which found no effect on DIT (p=0.753); another RCOD found a greater increase in DIT with breakfast containing 30g medium chain (5.31MJ) cf. long chain triglycerides (5.44MJ) (p<0.05). Three studies compared amount of fat at breakfast: a RCOD found no difference in DIT after 15 days low energy/moderate fat breakfast (418kJ) cf. high energy/low fat breakfast (2920kJ); a RCOD study found no difference between normal fat breakfast 220

(25%TE) and high fat breakfast (25%TE + 50g fat)(p=0.40); a repeated measures study found greater increases in DIT with high CHO cf. high fat breakfast (both 2092kJ, p<0.05). Lack of consistency in study design meant no valid comparisons could be made.

Conclusions: This SLR was unable to draw any conclusions about the effect of fat intake at breakfast on DIT, due to the heterogeneity of included trials. Further research with consistent study design and isocaloric comparisons are needed to determine if caloric value or type and quality of fat is an important determinant of DIT.

221

Appendix 6: Authors Contribution for chapter 5 I attest that Research Higher Degree candidate Angelica Quatela contributed to the following published paper:

Quatela, A.; Callister, R.; Patterson, A.J.; MacDonald-Wicks, L.K. The Energy Content and Composition of Meals Consumed after an Overnight Fast and Their Effects on Diet Induced Thermogenesis: A Systematic Review, Meta-Analyses and Meta-Regressions. Nutrients 2016, 8, 670.

Angelica Quatela contributed to the study design, data analysis and manuscript preparation as first author. Prof Robin Callister, Dr Amanda Patterson, Dr Lesley MacDonald-Wicks, contributed to the study design, data analysis and manuscript preparation within their capacity as PhD supervisors and co-authors.

Angelica Quatela 15-11-17

Dr Lesley MacDonald-Wicks 15-11-17

222

Appendix 7. Supplementary Table for Systematic Review – Chapter 2. Angelica Quatela, Robin Callister, Amanda Patterson and Lesley MacDonald- Wicks

Table S1. Formulas used to calculate participants’ characteristics, macronutrients compositions, or DIT.

Formulas BMI BMI = body weight (kg)/height (m)2 (WHO 2016) SD SD = [SE × square root (sample size (n) − 1)] † Macronutrients (KJ) macronutrient (kJ) = macronutrient (g) × Atwater factor †† Percentage of energy from macronutrient (% of energy) = [(macronutrient (kJ)/kJ intake †††) macronutrients × 100] DIT (kJ) DIT (kJ) = [(DIT % ECM × kJ intake †††)/100] DIT % ECM DIT % ECM = [(DIT KJ/kJ intake †††) × 100] DIT % AB DIT % AB = [(DIT kJ/Fasting RMR) × 100] † Sample size minus one was used instead of only sample size in order to account for the small sample sizes. †† Atwater factors: 16 kJ for CHO, 17 kJ for protein and 37 kJ for fat. ††† kJ intake for breakfast. AB = Above Baseline. DIT = Diet Induced Thermogenesis. ECM = Energy Content of Meal. RMR = Resting Metabolic Rate.

References

1. WHO. BMI Classification. Available online: http://apps.who.int/bmi/index.jsp?introPage=intro_3.html (accessed on 24 October 2016).

223

Appendix 8: Abstract for Chapter 6 • Abstract presented at the Asia Pacific Conference in Clinical Nutrition in Adelaide, Australia, 2017 (poster presentation by Quatela, A; Authors: Quatela A., A.J. Patterson, R. Callister, L.K. MacDonald-Wicks).

The ‘Typical Aussie Bloke study’: breakfast consumption habits of young Australian men.

Abstract: Breakfast is thought to be important for long term weight maintenance and prevention of chronic disease development. However, little is known regarding the foods/beverages that currently constitute a typical Australian breakfast. This study aimed to describe the current breakfast habits of young Australian men (aged 18-44y). Recruitment was from metropolitan (Newcastle) and regional (Tamworth) NSW Australia for a multicentre cross sectional study. Participants completed an online survey investigating breakfast consumption habits and other lifestyle characteristics. The sample included 112 men. The majority of participants (84%) were habitual breakfast eaters (≥5 times/week) and consumed this meal early (84%; 5:01 to 8:00am). A typical breakfast for the majority of habitual breakfast eaters consisted of one or more of the following beverages and/or foods habitually (≥5 times/week): coffee (40.4%), breakfast cereal with milk (39.4%), fruit (28.7%), toast (13.8%), spreads (11.7%), and/or yogurt (12.8%). This breakfast may be augmented (1-4 times/week) to include: eggs (58.5%), bacon (30.9%), juice (19.1%), and/or tea (17.0%). It appears that the most common foods consumed by Australian men have not changed since the National Nutrition Survey (NNS) in 1995. Cereals, milk and fruit were the most common foods consumed in both studies. However, the proportions of foods are consumed appear to have changed since the NNS in 1995; for example fruits consumption appears to have increased and milk consumption decreased, suggesting that Australian male breakfast habits may have changed over time.

224 Appendix 9: Authors Contribution for chapter 6 I attest that Research Higher Degree candidate Angelica Quatela contributed to the following paper in the process to be submitted to European Journal of Nutrition:

Chapter 6: Quatela, A; Patterson, A.J.; Callister, R; MacDonald-Wicks, L.K. (2017) Breakfast consumption habits of young Australian men from the ‘Typical Aussie Bloke’ study. To be submitted to European Journal of Nutrition.

Angelica Quatela contributed to the study design, data analysis and manuscript preparation as first author. Prof Robin Callister, Dr Amanda Patterson, Dr Lesley MacDonald-Wicks, contributed to the study design, data analysis and manuscript preparation within their capacity as PhD supervisors and co-authors.

Angelica Quatela 15-11-17

Dr Lesley MacDonald-Wicks 15-11-17

225 Appendix 10. Recruitment Flier for ‘Typical Aussie Bloke’ study for Newcastle

226 Appendix 11. Consent Form for ‘Typical Aussie Bloke’ study

227 Appendix 12. Ethics Approval for ‘Typical Aussie Bloke’ Study

228 Appendix 13: Online Questionnaire for ‘Typical Aussie Bloke’ study

‘Typical Aussie Bloke’ study

A questionnaire by the University of Newcastle

229 Participants will be provided the information sheet in Survey Monkey and then they will be requested to reply to the following questions:

1. I have read the information statement and would like to participate in this study. If you select 'Yes,' this will be your informed consent to participate in the survey component of this study. Yes No

If participants do not answer to this question, this message will appear: This question requires an answer.

When participants answer No to this question they will be directed to the following messages:

‘You are no longer able to take part in this study because you have not provided consent’. ‘Thank you’.

They will then be directed to the exit page.

When Participants reply Yes, they will be directed to the following message:

‘The following seven questions will determine your eligibility to participate further in this study’

2 How old are you? Less than 18 years 18-44 years Over 44 years

(If they reply ‘less than 18 years’ or ‘over 44 years’, they are excluded)

230

If participants do not answer to this question, this message will appear: This question requires an answer.

3. What is your gender?

Male Female

(If female, excluded)

If participants do not answer to this question, this message will appear: This question requires an answer.

4. Have you had cosmetic surgery, which would change the shape of your body? Yes No

(If yes, excluded) If participants do not answer to this question, this message will appear: This question requires an answer.

5. Do you suffer from claustrophobia? Participation in this study requires the measurement of your resting metabolic rate using an indirect calorimetry hood. You will be asked to lie quietly in the laboratory while the amount of oxygen you use when resting is measured. You will have a clear plastic hood (shaped like a dome) you can see through over your head and you will be able to breathe normally, however if you suffer from claustrophobia you are advised not to participate. Yes No

(If yes, excluded)

231

If participants do not answer to this question, this message will appear: This question requires an answer.

6. Are you able to travel to the University of Newcastle Callaghan campus OR the University of Newcastle Department of Rural Health in Tamworth on one occasion for an 80 minute measurement session?

Yes – Callaghan (Newcastle) Yes - Tamworth No

(If no, excluded)

If participants do not answer to this question, this message will appear: This question requires an answer.

7. Do you currently have a thyroid condition or are you taking medication for a thyroid condition (e.g. hypo or hyperthyroidism)? Yes No

(If yes, excluded)

If participants do not answer to this question, this message will appear: This question requires an answer.

8. Do you currently use insulin for treatment of diabetes? Yes No

(If yes, excluded)

232

If participants do not answer to this question, this message will appear: This question requires an answer.

This message will appear to those participants found ineligible to take part in this study as soon as they meet an exclusion criterion:

‘Unfortunately, based on your answers you are not eligible to take part in this study. Thank you for your interest.’

This message will appear to those participants found eligible to take part in this study:

9. ‘Based on your answers, you are eligible to take part in this study. If you wish to continue please proceed and complete the following 20 minute questionnaire regarding your demographics, dietary intake, breakfast consumption, physical activity, and body satisfaction.’ Yes, I wish to participate. No, I no longer wish to participate in the ‘Typical Aussie Bloke’ study

If participants do not answer to this question, this message will appear: This question requires an answer.

If they reply ‘No, I no longer wish to participate in the ‘Typical Aussie Bloke’ study, they receive the following message:

“Thank you for your interest in this study.’

233

The following questionnaire asks questions about your demographics, sleeping and eating habits, body satisfaction, physical activity levels, and general health.

Demographics

10. What is your date of birth? ______

Day Month Year

If participants do not answer to this question, this message will appear: This question requires an answer.

This message will appear when an invalid date (older than 44 years old and younger than 18 years old) is entered: Please check the date. You must be aged 18-44 to participate in this study.

11 In what country were you born? Australia United Kingdom Italy Greece New Zealand Vietnam Other Prefer not to answer If “other” was selected, please specify: ______If participants do not answer to this question, this message will appear: This question requires an answer.

12. Are you of Aboriginal or Torres Strait Islander descent? No Yes, Aboriginal Yes, Torres Strait Islander Yes, both Aboriginal and Torres Strait Islander 234

Don't know Prefer not to answer

If participants do not answer to this question, this message will appear: This question requires an answer.

13. What is your current postcode?

If the postcode typed does not fall between 1000-3707, this message will appear: The comment you entered is in an invalid format. Please enter a NSW postcode.

If participants do not answer to this question, this message will appear: This question requires an answer.

14. What is your current marital status? Married Defacto (opposite sex) Defacto (same sex) Separated Divorced Widowed Single Prefer not to answer

If participants do not answer to this question, this message will appear: This question requires an answer.

15. What is the highest level of education you have completed? No formal qualifications School or intermediate certificate Higher school or leaving certificate Trade/apprenticeship Certificate/diploma

235

University undergraduate degree University postgraduate degree Prefer not to answer

If participants do not answer to this question, this message will appear: This question requires an answer.

16. Which of the following best describes your main current employment status?

Full time paid work Part time or casual paid work Work without pay (eg in a family business) Home duties Studying Unemployed Unpaid voluntary work Retired Unable to work due to sickness or injury Prefer not to answer

If participants do not answer to this question, this message will appear: This question requires an answer.

17. What is/was/will be your main occupation? (If you are a student, mark the type of occupation you are studying for)

Manager or administrator (eg personnel manager, managing supervisor) Professional (eg teacher, social worker, doctor, artist) Para-professional (eg welfare worker, technical officer, registered nurse) Trade (eg hairdresser, cook, mechanic) Administrative assistant (eg secretary, telephonist) Sales or personal service worker (eg sales assistant, bar attendant, child care worker, enrolled nurse)

236

Machine operator or driver (eg sewing machinist) Manual worker (labourer, cleaner, kitchen hand) Never had a paid job Other Prefer not to answer If “other” was selected, please specify:

If participants do not answer to this question, this message will appear: This question requires an answer.

18. What is/was/will be your partner/spouse’s main occupation? (If your partner/spouse is a student, select the occupation he/she is studying for)

No partner/spouse Manager or administrator (eg personnel manager, managing supervisor) Professional (eg teacher, social worker, doctor, artist) Para-professional (eg welfare worker, technical officer, registered nurse Trade (eg hairdresser, cook, mechanic) Administrative assistant (eg secretary, telephonist) Sales or personal service worker (eg sales assistant, bar attendant, child care worker, enrolled nurse) Machine operator or driver (eg sewing machinist) Manual worker (labourer, cleaner, kitchen hand) Never had a paid job Other Prefer not to answer If “other” was selected, please specify: ______

If participants do not answer to this question, this message will appear: This question requires an answer.

237

19. What is your current annual household income before taxes during the last 12 months? (This question refers to family household income. If you live in a share house or do not know your family income, just include your own income)

Don't know or would rather not say Less than $25,000 (less than $481 per week) $25,000 to $34, 999 ($481 to $673 per week) $35,000 to $49,999 ($673 to $962 per week) $50, 000 to $74, 999 ($962 to $1,442 per week) $75, 000 to $99,999 ($1,442 to $1,923 per week) $100,000 to $149, 999 ($1,923 to $2,885 per week) $150, 000 to $199,999 ($2,885 to $3,846 per week) $200,000 or more ($3,846 or more per week)

If participants do not answer to this question, this message will appear: This question requires an answer.

20. How many dependent children do you have in the age categories indicated? (Select as many as applicable) None One Two Three or more Prefer not to answers Under 2 years 2-5 years 6-10 years 11-15 years 16-18 years

If participants do not answer to this question, this message will appear: This question requires an answer. If you do not have any dependent children in a given age range, select "None" on each corresponding row. If you prefer not to answer this question, please select the “prefer not to answer” option on each corresponding raw.

238

21. In addition to yourself, how many people are currently living in your household? Live alone One other person Two other people Three other people Four other people Five or more other people Prefer not to answer

If participants do not answer to this question, this message will appear: This question requires an answer. Sleeping habits (page name)

The following questions are about your sleeping habits.

Sleeping habits (page name)

22. What time do you usually go to sleep on weekdays?

Before 9.00 pm 9.01 pm -10.00 pm 10.01 pm -11.00 pm 11.01 pm – 12.00 am 12.01 am -1.00 am After 1.00 am

If participants do not answer to this question, this message will appear: This question requires an answer.

23. What time do you usually go to sleep on weekend days?

Before 9.00 pm 9.01 pm -10.00 pm

239

10.01 pm -11.00 pm 11.01 pm – 12.00 am 12.01 am -1.00 am After 1.00 am

If participants do not answer to this question, this message will appear: This question requires an answer.

24. Please indicate the time you usually wake up on weekdays.

Midnight to 5.00 am 5.01 to 6.00 am 6.01 to 7.00 am 7.01 to 8.00 am 8.01 to 9.00 am 9.01 to 10.00 am 10.01 to 11.00 am 11.01 am to noon

If participants do not answer to this question, this message will appear: This question requires an answer.

25. Please indicate the time you usually wake up on weekend days.

Midnight to 5.00 am 5.01 to 6.00 am 6.01 to 7.00 am 7.01 to 8.00 am 8.01 to 9.00 am 9.01 to 10.00 am 10.01 to11.00 am 11.01 am to noon

240

If participants do not answer to this question, this message will appear: This question requires an answer.

Eating habits (page name) The following questions are about your eating habits on a typical day or in typical week.

Eating habits (page name) 26. Please indicate the time you usually have the first meal of the day on weekdays.

Midnight to 5.00 am 5.01 to 6.00 am 6.01 to 7.00 am 7.01 to 8.00 am 8.01 to 9.00 am 9.01 to 10.00 am 10.01 to11.00 am 11.01 am to noon

If participants do not answer to this question, this message will appear: This question requires an answer.

27. Please indicate the time you usually have the first meal of the day on weekend days.

Midnight to 5.00 am 5.01 to 6.00 am 6.01 to 7.00 am 7.01 to 8.00 am 8.01 to 9.00 am 9.01 to 10.00 am 10.01 to11.00 am 11.01 am to noon

241

If participants do not answer to this question, this message will appear: This question requires an answer.

28. How many times do you usually have something to eat in a day (including snacks and evenings)? Once 2 to 4 times 5 to 6 times 7 or more times Don't know/varies

If participants do not answer to this question, this message will appear: This question requires an answer.

29. How many days per week do you usually have something to eat for breakfast? Rarely or never 1 to 2 days 3 to 4 days 5 or more days Don't know/varies

If participants do not answer to this question, this message will appear: This question requires an answer.

30. How many times per week do you usually have these foods for breakfast?

(This question is only for those who replied they consume breakfast in question 29. The following options for times per week will be available in SurveyMonkey: less than once per week,1,2,3,4,5,6,7 or more).

Times per week Fruits Toast Butter and/or margarine

242 Spread (e.g. jam, honey, peanut butter, Nutella, vegemite, etc.) Cereals like All bran Cereals like Sultana BranTM, FibrePlusTM, BranFlakesTM Cereals like Weet BixTM ,Vita BritsTM, WeetiesTM Cereals like CornFlakes, NutrigrainTM , Special KTM Cereals like Porridge Cereals like Muesli Milk for cereal Yogurt Eggs Bacon Beans Pancake/crepes Other foods (please specify):

If participants do not answer to this question, this message will appear: This question requires an answer.

31. How many times per week do you usually have these beverages for breakfast?

Times per week Coffee Tea Milk on its own Hot chocolate milk/ milo Juice 243

Smoothies Other beverages (please specify):

If participants do not answer to this question, this message will appear: This question requires an answer.

32. What is the main reason you eat breakfast? (Choose only one option) (This question is only for those who replied they consume breakfast in question 29).

It gives me energy I want to lose weight It helps prevent me from getting hungry before lunchtime I enjoy it It helps me to wake up It is what I always do I am hungry Eating breakfast makes it easier to control my weight Others reasons If other reasons was selected, please specify:

If participants do not answer to this question, this message will appear: This question requires an answer.

33. On days that you do not have breakfast, what is the reason?

Not enough time I do not feel like eating first thing I want to lose weight Hung over I have a cigarette instead I do not have any food in the house I do not have enough money to have breakfast I rarely/never don’t eat breakfast

244

Other reasons If other reasons was selected, please specify:

If participants do not answer to this question, this message will appear: This question requires an answer.

34. How many serves of vegetables do you usually eat each day? A serving of vegetables is 75 grams (e.g. 1/2 cup of cooked vegetables, 1 cup of salad, or one medium sized potato). 1 serve 2 serves 3 serves 4 serves 5 serves 6 serves or more Less than one serve Don't eat vegetables

If participants do not answer to this question, this message will appear: This question requires an answer.

35. How many serves of fruit do you usually eat each day? A serving of fruit is 125 grams (e.g. one medium piece of fruit, two small pieces of fruit, one cup of chopped, frozen or canned fruit, or 2 tablespoons of dried fruit). 1 serve 2 serves 3 serves 4 serves 5 serves 6 serves Less than one serve

245

Don't eat fruit

If participants do not answer to this question, this message will appear: This question requires an answer.

246

Body satisfaction (page name)

The following questions are about your satisfaction with your body size and shape.

Body satisfaction (page name)

36. How important are your weight and shape to you? One of the most important things in my life Very important Somewhat important Not important at all

If participants do not answer to this question, this message will appear: This question requires an answer.

37. In the past month how dissatisfied have you felt about (select one of the following options for each line): Not at all Slightly Moderately Markedly Your weight Your shape

If participants do not answer to this question, this message will appear: This question requires an answer.

38. How would you describe yourself now? Very underweight Underweight Slightly underweight Average Slightly overweight Overweight Very Overweight Don’t know 247

If participants do not answer to this question, this message will appear: This question requires an answer.

39. How much would you like to weigh? Happy as I am 1-5kg more Over 5kg more 1-5kg less 6-10kg less Over 10kg less

If participants do not answer to this question, this message will appear: This question requires an answer.

40. Are you trying to lose weight now? Yes No

If participants do not answer to this question, this message will appear: This question requires an answer.

41. How often have you gone on a diet to lose weight during the last year? Never 1-4 times 5-10 times More than 10 times I am always on a weight-loss diet

If participants do not answer to this question, this message will appear: This question requires an answer.

42. Are you trying to put on weight now?

248

Yes No

If participants do not answer to this question, this message will appear: This question requires an answer.

43. How often have you changed your diet to put on weight during the last year? Never 1-4 times 5-10 times More than 10 times I am always on a weight-gain diet

If participants do not answer to this question, this message will appear: This question requires an answer.

44. How often have you taken supplements to help you lose weight? Never 1-4 times 5-10 times More than 10 times I am always using supplements

If participants do not answer to this question, this message will appear: This question requires an answer.

45. How often have you taken supplements to help you gain weight? Never 1-4 times 5-10 times More than 10 times I am always using supplements

249

If participants do not answer to this question, this message will appear: This question requires an answer.

46. How often do you compare your body to other men’s bodies? Never Rarely Sometimes Often

If participants do not answer to this question, this message will appear: This question requires an answer.

47. If money was NOT a problem, would you ever consider plastic surgery? Yes No If participants do not answer to this question, this message will appear: This question requires an answer.

Physical Activity (page title) The following questions are about your physical activity habits

Physical Activity (page title) The next three questions are about the amount of physical activity you did LAST WEEK. The types of activity we are interested in are: WALKING (fairly briskly, including walking to and from work or school); MODERATE leisure-time activities (like golf, social tennis, moderate exercise classes, recreational swimming or cycling, and gardening); and VIGOROUS leisure-time activities (the ones that make you puff and pant, e.g. vigorous

250

48. How many times did you do each type of activity during your leisure time LAST WEEK? (Only count the number of times when the activity lasted for 10 minutes or more)

a) Walking (briskly) times

times b) Moderate activity

times c) Vigorous activity

aerobics, competitive sport, vigorous cycling, running, swimming, etc.).

Please write ‘0’ in the box for each activity you DID NOT do’.

This message will appear when a participant will enter an invalid number or when do not answer this question: This question requires an answer. Please enter a whole number between 0 and 21. If you did not do a certain type of activity, enter "0."Option 21 maximum.

49. If you add up all the times you spent in each activity LAST WEEK, how much time did you spend ALTOGETHER doing each type of activity as part of your leisure time? Walking 3 times for 30 mins each time = 3 x 30 = 90 mins = 1 hour 30 minutes) (The following options for hours and minutes will be available in Survey Monkey – hours options to tick: from 0 to 40 - minutes options to tick: 00-05-10-15-20-25-30-35-40-45-50-55) a) Walking (briskly) hours mins

b) Moderate activity hours mins

c) Vigorous activity hours

mins

251

50. During the LAST WEEK, how much time did you spend ALTOGETHER in your WORK (paid or unpaid) doing VIGOROUS activity (that is, activity which made you puff or pant such as labouring, farm work, gardening, heavy work around the yard, heavy housework, etc.)?

TOTAL TIME in vigorous work-related activity last week hours mins This message will appear when the participant do not answer this question: This question requires an answer. “If you did not do a certain type of activity, select "0:00"

This message will appear when the participant do not answer this question: This question requires an answer. If you did not do vigorous activity in your work, enter "0:00."

51. Do you exercise to lose weight? Yes No

If participants do not answer to this question, this message will appear: This question requires an answer.

52. Do you exercise to put on muscle? Yes No

If participants do not answer to this question, this message will appear: This question requires an answer.

General Health (page title)

The following questions are about your health status and any medications you may be taking.

General Health (page title

252

53. Have you been diagnosed with a metabolic disorder (e.g. diabetes, liver or kidney disease)? Yes No

If participants do not answer to this question, this message will appear: This question requires an answer.

54. Are you currently taking any medication to control your metabolic disorder? (Only for who answered 'Yes' to Question 53) Yes No

If participants do not answer to this question, this message will appear: This question requires an answer.

55. Are you currently taking any medications that may affect your weight, resting metabolic rate, or fluid balance (e.g. corticosteroids, diuretics, weight loss medication, medication for the treatment of diabetes, beta-blockers, antipsychotics, sodium valproate, lithium, tricyclic antidepressants, anabolic steroids or human growth hormone)? Yes No

If participants do not answer to this question, this message will appear: This question requires an answer.

253

56. Please provide your contact details for booking your measurement appointment: Full Name: ______Address: ______Email Address: ______@______Preferred Phone: ______

If participants do not answer to this question, this message will appear: This question requires an answer.

57. How would you prefer to be contacted for the booking of your measurement appointment? Phone Email

If participants do not answer to this question, this message will appear: This question requires an answer.

Once participants finish completing the questionnaire, they will be directed to this message:

‘Thank you for completing the questionnaire. You will soon be contacted to book your measurement session’.

Exit page.

254

Acknowledgements Eligibility and medications questions The majority of the eligibility questions and questions about medications were most consistent with the Average Australian Women survey. Demographics Demographic data collected included age, marital status, education, employment, country of birth, number of children, and household composition. The type of information collected was most consistent with that collected in the Australian Bureau of Statistics Census of Population and Housing (ABS Census) (ABS 2011)and the ALSWH (ALSWH 2017), in order to determine representativeness of the sample.

Physical Activity The PA questions were the same developed by the ALSWH surveys (ALSWH 2017, ALSWH 2017) which were determined using items from Active Australia’s 1999 National Physical Activity Survey (Armstrong, Bauman et al. 2000) or were directly taken from the Active Australia’s 1999 National Physical Activity Survey (Armstrong, Bauman et al. 2000).

Dietary questions The short dietary questions including fruits and vegetable habits were most consistent with the ALSWH surveys (ALSWH 2017). The questions assessing habitual breakfast consumption habits and eating frequency were the same one used by the National Nutrition Survey in 1995 (ABS 1997). Questions regarding reasons for consuming or not consuming breakfast were most consistent with Reeves et al. study (Reeves, Halsey et al. 2013). Body image questions The body image questions were most consistent with the ALSWH surveys (ALSWH 2017) and the Average Australian Women surveys.

References of acknowledgments • ABS (1997). National Nutrition Survey Selected Highlights. • ALSWH. (2017). "Australian Longitudinal Study on Women's Health: Surveys Women's Health Australia " Retrieved 05/06/17, from http://www.alswh.org.au/for-researchers/surveys. 255

• ABS. 2011 Census of Population and Housing: Household Form In: Statistics ABo, editor. Canberra: Australian Bureau of Statistics 2011. • Average Australian Women (AAW) study. 2014 (Personal Correspondence) • Reeves, S., L. G. Halsey, Y. McMeel and J. W. Huber (2013). "Breakfast habits, beliefs and measures of health and wellbeing in a nationally representative UK sample." Appetite 60(0): 51-57.

256

Appendix 14: Conference Abstract for Chapter 7 • Abstract accepted for presentation at Dietitians Association of Australia conference in Sydney, Australia, May, 2018 (oral presentation by Quatela, A; Authors: Quatela A., A.J. Patterson, R. Callister, L.K. MacDonald- Wicks).

The ‘typical Aussie Bloke study’: The relationships of Habitual Breakfast consumption with mediators of obesity and chronic disease development amongst young Australian men.

Breakfast is often regarded as protective against obesity and chronic disease risk; however, the evidence to support this is limited and contradictory. This study explores the relationship between habitual breakfast consumption and mediators of obesity and chronic disease development in young men. This multicentre cross sectional study recruited men 18-44 y from metropolitan and regional NSW, Australia. Participants completed an online survey about breakfast habits and lifestyle characteristics, and attended a measurement session. Of the 112 men, 94 were Habitual Breakfast Eaters, 7 were Occasional Breakfast Eaters and 10 were Habitual Breakfast Skippers. Habitual Breakfast Eaters were more likely to have tertiary qualifications (62.8%); whereas a higher percentage of Occasional Breakfast Eaters (71.4%) and Habitual Breakfast Skippers (80.0%) had secondary school qualifications (p=0.010). No other demographic characteristics were found to significantly differ. Furthermore, men with different breakfast habits did not significantly vary for: body mass index; waist, hip and chest circumferences; body composition; blood pressure; resting metabolic rate; finger stick measurement of blood glucose and lipid profiles (Triglycerides; total, LDL and HDL Cholesterol); sleeping and waking habits; physical activity and fruit and vegetable consumption. However, Habitual Breakfast Eaters were more likely to consume ≥ 5 daily eating events (59.6%) than Occasional Breakfast Eaters (28.6%) and Habitual Breakfast Skippers (20%; p=0.015). Finally, except for number of daily eating events, no other significant relationships were found between breakfast habits and intermediates of obesity and chronic disease risk. Longitudinal studies are needed, with greater number of breakfast skippers, to investigate this more fully.

257

Appendix 15: Authors Contribution for chapter 7 I attest that Research Higher Degree candidate Angelica Quatela contributed to the chapter 7.

Chapter Title: ‘‘Typical Aussie Bloke’ Part 2: Breakfast consumption and eating patterns in relation to intermediate risk factors for obesity and chronic disease development.’

Angelica Quatela contributed to the study design, data analysis and chapter preparation as first author. Prof Robin Callister, Dr Amanda Patterson, Dr Lesley MacDonald- Wicks, contributed to the study design, data analysis and chapter preparation within their capacity as PhD supervisors and co-authors.

Angelica Quatela 15-11-17

Dr Lesley MacDonald-Wicks 15-11-17

258 Appendix 16. Information Sheet for ‘Typical Aussie Bloke’ study

Information Statement for the Research Project: Typical Aussie Bloke

(University of Newcastle, Callaghan Campus)

Document Version [3]; dated [18/08/2015]

You are invited to participate in the research project identified above, which is being conducted by Dr Amanda Patterson, Dr Lesley MacDonald-Wicks, Prof Robin Callister and Dr Leanne Brown from the Faculty of Health and Medicine at the University of Newcastle.

This research is part of Angelica Quatela’s PhD project at the University of Newcastle, supervised by Dr Lesley MacDonald-Wicks and Dr Amanda Patterson from the School of Health Sciences, and Prof. Robin Callister from the School of Biomedical Sciences and Pharmacy.

Why is the research being done?

The aim of this study is to collect detailed, high quality information about the body size, body shape, body composition, blood pressure and metabolic parameters (resting metabolic rate and metabolic health blood analysis) of Australian men. This project also aims to collect information on breakfast habits, in order to compare habitual breakfast eaters and habitual breakfast skippers on anthropometric data and metabolic parameters.

Who can participate in the research?

You can participate in this project if you:

• Are male and aged 18 to 44 years.

You cannot participate in this project if you:

• Have had cosmetic surgery which would change the shape of your body

• Have a known thyroid condition or take medication for a known thyroid condition

• Have insulin dependent diabetes

• Suffer from claustrophobia 259 • Are unable to attend the 80-minute measurement session on the University of Newcastle, Callaghan campus

What choice do you have?

Participation in this research is entirely your choice. Only those people who give their informed consent will be included in the project. Whether or not you decide to participate, your decision will not disadvantage you.

If you do decide to participate, you may withdraw from the project at any time without giving a reason, and have the option of withdrawing any data, which identifies you.

What would you be asked to do?

Firstly, you will be asked to complete an online questionnaire through SurveyMonkey to determine if you are eligible to participate in the study and gather some information regarding your demographics, dietary intake, breakfast consumption, physical activity, and body satisfaction. This questionnaire should take no more than 20 minutes to complete. If you are eligible and agree to participate, you will be asked to make an appointment for a measurement session.

Measurement Session

The following measurements will be taken by a researcher: height, sitting height, waist circumference, hip circumference, chest circumference. You will be asked to remove your T-shirt or shirt as part of the measurement procedure. Please wear lightweight shorts with an elastic waist (not heavy cotton shorts or jeans) to the measurement session.

Weight, body fat percentage and lean body mass will be measured using a bio-impedance device. This is similar to the bathroom scales that can measure body lean muscle and fat as a percentage, by sending a small electrical current through your body (you do not feel this current at all). You will need to stand on it with bare feet and hold some metallic handles in your hands while remaining very still for 90 seconds. Please let us know if you have any implanted electrical device like a pacemaker, as this effects whether you can be measured with the bio-impedance device.

A blood pressure device (Pulsecor Cardioscope) will be used to measure your systolic and diastolic blood pressure, as well as the stiffness or flexibility of your arteries and your pulse rate. A blood pressure cuff will be placed on your upper arm, inflated then slowly deflated. You will be asked to sit quietly for a number of these measurements (minimum of two and a maximum of five) to be obtained. Each measurement will take approximately 60 seconds.

Resting metabolic rate will be measured using an indirect calorimeter, a machine that measures the amount of oxygen you consume while you are lying quietly on a bed in the laboratory. A transparent hood will be placed over your head, chest and shoulders during this measurement, but you will be able to see 260

through this and you will be able to breathe normally. If you suffer from claustrophobia you may find this difficult to tolerate and may choose not to participate in this project.

A finger-prick blood sample will be collected for analysis using the CardioCheck PA equipment. A sterile disposable lancet will be used to puncture your finger (like that used by diabetics to test their blood sugar). 70µl (approximately three drops) of blood will be collected. Your finger may sting for a few seconds after the puncture. The blood collected will then be analysed by the Cardiocheck PA to determine your glucose, ketones and lipid profile.

For accurate measurements, you will be required to fast for 4 hours prior to attending your session. In this time you are free to drink water, but should avoid all eating, drinking anything other than water, and smoking. You should also avoid any strenuous exercise for 24 hours prior to your session.

Measurements will take place in the Nutrition and Dietetics Anthropometry Laboratory (HC57) in the Hunter Building, Callaghan campus.

How much time will it take?

• The online questionnaire will take no more than 20 minutes to complete. • Attendance at the measurement session will take no longer than 80 minutes.

What are the risks and benefits of participating?

All the measurements used in the study (questionnaire, weight, body composition, energy expenditure, finger sticks blood sample and blood pressure) have been widely used in research and are considered standard tools for their specific measurements; therefore, there are limited risks. Very minimal pain and/or discomfort may be experienced with the finger prick blood sample. The finger stick blood sample minimizes the potential side effects experienced with traditional blood sample (e.g. feeling light headed, fainting or bruising). In the unlikely scenario that you feel unwell during the procedure, you will be able to stop the measurement at any time without giving any reason.

Some questions in the questionnaire are of a sensitive nature (your feelings about your weight and shape, etc.). As with all information collected, your answers to these questions will be kept completely confidential and your name will not be stored alongside your responses. If you experience any feelings that are overwhelming or distressing while answering these questions, please seek help from your doctor or contact Lifeline on 131144. If you are a student of the University, you can access the University Counselling Service (4921 5801), or if you are staff member you can access the EAP (Employee Assistance Program) via the staff website. If you would like to address issues of weight or nutrition, you can contact the Dietitians Association of Australia (www.daa.asn.au) for advice on accessing an Accredited Practicing Dietitian.

261

For participating in this study, you will receive a copy of your results (body composition, body mass index, resting metabolic rate, physical measurements, blood pressure, and blood glucose and lipid levels) and their interpretation.

You will also be entered into a lottery to win a Fitbit (a wrist monitor). The measurement outcomes or responses provided in the questionnaire will not affect eligibility for the prize draw.

How will your privacy be protected?

Data collected through the online questionnaire

The questionnaire has been developed using Survey Monkey (www.surveymonkey.com). Questionnaire responses are sent to the researchers’ Survey Monkey account via a secure, encrypted connection. Only the researchers have access to the password protected Survey Monkey account. Please see the Survey Monkey Privacy Policy (http://www.surveymonkey.com/mp/policy/privacy-policy/) and Security Statement (http://www.surveymonkey.com/mp/policy/security/) for further information.

Data collected at the measurement session

Initially all data collected in non-electronic form will be stored in one of the investigator’s offices to ensure its security and the confidentiality of any identified data. Only the researchers and the chief investigators will have access to the raw data. The researchers will enter this raw data into a statistics program; at this point, all identifiers will be removed and replaced with a code. Once this information is entered on the data file, all raw data will be shredded and no person will be identifiable in the data files or published report. The data will be kept for a minimum of 5 years after the study in a password protected location at the University of Newcastle.

How will the information collected be used?

The results of this research will contribute to Angelica Quatela’s PhD thesis. Combined results will also be reported via national and international conferences and peer reviewed publications. Individual participants will not be identified in any reports arising from the study. Non-identifiable data may be also shared with other parties to encourage scientific scrutiny, and to contribute to further research and public knowledge, or as required by law. At the conclusion of the study a brief summary of the results will be available from the PhD researcher or Chief Investigator, and can be posted or emailed to you directly (please indicate on the written consent form prior to your measurement session).

What do you need to do to participate?

Please read this Information Statement and ensure you understand its contents before you continue your participation in this study. If there is anything you do not understand, or you have questions, contact the Chief Investigator, Dr Amanda Patterson on 49216420 or [email protected].

If you would like to participate, please go to the online questionnaire, available at: (https://www.surveymonkey.com/r/TypicalAussieBloke) 262

The questionnaire will first ask you to identify that you have read this information statement, and that by choosing the ‘Yes’ option, you will be consenting to participating in the questionnaire section of the study. The following seven questions will confirm whether you are eligible to participate in the study and if you are, it will allow you to complete the questionnaire. You will then be contacted to confirm a time convenient for you to attend the measurement session at the University of Newcastle, Callaghan Campus. At the start of the measurement session, you will be required to sign a written consent form for continued participation in the study.

Further information

If you would like further information, please contact the Chief Investigator: Amanda Patterson (Ph: (02) 4921 6420 or [email protected])

Thank you for considering this invitation.

Dr Amanda Patterson

Senior Lecturer, School of Health Sciences, Faculty of Health and Medicine

Angelica Quatela

BSc (Hons) Nutrition – PhD candidate, Nutrition and Dietetics Department, School of health and Science, Faculty of health and Medicine.

On behalf of the research team

Complaints about this research

This project has been approved by the University’s Human Research Ethics Committee, Approval No. H- 2015-0199.

Should you have concerns about your rights as a participant in this research, or you have a complaint about the manner in which the research is conducted, it may be given to the researcher, or, if an independent person is preferred, to the Human Research Ethics Officer, Research Office, The Chancellery, The University of Newcastle, University Drive, Callaghan NSW 2308, Australia, telephone (02) 49216333, email [email protected].

263