ABO Genotype, “Blood-Type” Diet and Cardiometabolic Risk Factors

ABO Genotype, “Blood-Type” Diet and Cardiometabolic

Risk Factors

Jingzhou Wang

A thesis submitted in conformity with the requirements

for the degree of Master of Science (M.Sc.)

Graduate Department of Nutritional Sciences

University of Toronto

ABO Genotype, “Blood-Type” Diet and Cardiometabolic Risk Factors

Jingzhou Wang

Master of Science

Graduate Department of Nutritional Sciences

University of Toronto

2014

Abstract

The ‘Blood-Type’ diet advises individuals to eat according to their ABO blood group to optimize health without the support of science evidence. The objective of this study was to determine whether consumption of a diet in accordance with an individual’s ABO genotype is associated with various biomarkers of cardiometabolic health. Study subjects (n=1,455) were participants of the Toronto Nutrigenomics and Health study. Dietary intake was assessed using a one-month,

196-item food frequency questionnaire and a diet score was calculated to determine relative adherence to each of the four blood type diets. ABO blood group was determined by genotyping rs8176719 and rs8176746 in the ABO gene. The results show that adherence to the Type-A,

Type-AB, and Type-O diets were associated with favourable profile of certain cardiometabolic risk factors (P<0.05); however, these dietary effects were not dependent on someone’s ABO blood group. Therefore, the findings do not support the “Blood-Type” diet hypothesis.

Acknowledgments

I would not have reached this stage without the help of others. Here, I want to express my sincere gratitude to the people who made my life in the department truly enjoyable.

First of all, I want to thank my supervisor, Dr. Ahmed El-Sohemy for his excellent support throughout my Master’s training. His passion in the field of nutrigenomics attracts me to pursue this graduate degree and his brilliant ideas allow me to work on an interesting project.

With his approachability and patience, I was able to learn and improve my thinking, writing, speaking skills significantly. His generosity and flexibility has also enabled me to discover career opportunities and to develop my soft skills in other settings. It was truly my honour to be a student for Ahmed. I am very grateful for all the opportunities and knowledge he has offered. I cannot imagine a better mentor.

Second, I would like to thank Dr. Richard Bazinet and Dr. Elena Comelli for offering wonderful support as my committee members. Their critical feedback and constructive advice have helped me to further develop my critical thinking and to mature as a graduate student. It was also a privilege to have two of my most admiring undergraduate professors to serve on my advisory committee. I have also benefited tremendously from the teaching of Dr. Paul Corey and

Dr. Tony Hanley. Their high levels of expertise in statistics and epidemiology have helped me build a solid foundation in my research project. I also want to specifically thank Dr. Harvey

Anderson, who has played a significant role in my undergraduate training. Without his guidance,

I would not fall in love with research. His consistent support, even after I leave the lab, has been invaluable.

Third, I want to thank all the past and present members in the El-Sohemy team. The thesis would not be completed in a timely manner without the effort of all past students. I am also very iii

grateful to receive the support from Bibiana Garcia-Bailo, Daiva Nielsen, Ouxi Tian, Andre Dias,

Nanci Guest and Joseph Jamnik in both academic and social circumferences. The quality of my research project improved significantly thanks to their input. It was a pleasure to be a part of the big family.

Fourth, I would like to thank all the staff members and students in the department. Louisa

Kung and Emeliana D'Souza were always there to give an extra hand. Clara Cho and Diana

Sanchez-Hernandez were my first mentors in the research field and have taught me a significant amount of knowledge and skills that are useful in my future career. I am also equally grateful for the experiences and encouragement offered by my colleagues and teammates in the departmental student union and volleyball team. The assistances of all these fellows have no doubt facilitated the completion of my research project.

I also gratefully acknowledge the funding sources that made this thesis possible: the

Ontario Graduate Scholarship program and the Advanced Foods and Materials Network. In addition,

I was fortunate to be able to attend an international conference thanks to the generous support from

Dr. Ahmed El-Sohemy and School of Graduate Studies at University of Toronto.

My last gratitude goes to my family. My parents have offered invaluable guidance in my entire life. Without the support of my father, Song Wang, I would not have had the opportunity to study aboard. I want to specifically thank for my mother, Yu Liu, who passed away from liver cancer when I was 15 years old. Watching Animal Planet with my mother made me fall in love with biology when I was young; losing my closest person due to cancer led me to pursue a career in life sciences. To her I dedicate this thesis.

Thank you all!

“In a group of three people, there is always something I can learn from. Choose to follow the

strengths of others, use the shortcomings to reflect upon ourselves."

Confucious, 500 BC

Table of Contents

TABLE OF CONTENTS VI

LIST OF TABLES IX

LIST OF FIGURES X

LIST OF APPENDIX XI

Chapter 1: Introduction and Literature Review

1.1 INTRODUCTION 1

1.2 “BLOOD-TYPE” DIET 1

1.3 ABO BLOOD GROUP 2

1.3.1 ABO SYSTEM 2

1.3.2 ABO BLOOD GROUP AND DISEASES 3

1.3.3 BLOOD GROUPS AND NUTRITION 4

1.4 CARDIOMETABOLIC DISEASES 6

1.4.1 PREVALENCE AND ETIOLOGY 6

1.4.2 BIOMARKERS OF RISK 6

1.4.3 RISK FACTORS 9

1.5 SUMMARY AND RATIONALE 10

1.6 HYPOTHESIS AND ORGANIZATION OF THESIS 11

Chapter 2: Effect of ABO Genotype on Cardiometabolic Risk Factors

2.1 ABSTRACT 12

2.2 INTRODUCTION 13

2.3 METHODS 14

2.3.1 STUDY DESIGN AND PARTICIPANTS 14

2.3.2 ABO GENOTYP IDENTIFICATION 14

2.3.3 CARDIOMETABOLIC RISK FACTOR ASSESSMENT 15

2.3.4 STATISTICAL ANALYSIS 16

2.4 RESULTS 16

2.5 DISCUSSION 22

Chapter 3: Effect of "Blood-Type" Diet on Cardiometablic Risk Factors

3.1 ABSTRACT 23

3.2 INTRODUCTION 24

3.3 MATERIAL AND METHODS 27

3.3.1 STUDY DESIGN AND PARTICIPANTS 27

3.3.2 DIETARY ADHERENCE SCORE ASSESSMENT 27

3.3.3 CARDIOMETABOLIC RISK FACTOR ASSESSMENT 28

3.3.4 STATISTICAL ANALYSIS 28

3.4 RESULTS 29

3.5 DISCUSSION 38

Chapter 4: Effect of matching ABO genotype to "Blood-Type" Diet on Cardiometabolic Risk Factors

4.1 ABSTRACT 41

4.2 INTRODUCTION 42

4.3 MATERIAL AND METHODS 42

4.3.1 STUDY DESIGN AND PARTICIPANTS 42

4.3.2 ABO GENOTYPE IDENTIFICATION 43

vii

4.3.3 DIETARY ADHERENCE SCORE ASSESSMENT 43

4.3.4 CARDIOMETABOLIC RISK FACTOR ASSESSMENT 43

4.3.5 STATISTICAL ANALYSIS 43

4.4 RESULTS 44

4.5 DISCUSSION 54

Chapter 5: Overview and General Discussion

5.1 OVERVIEW 56

5.2 LIMITATIONS 58

5.3 FUTURE DIRECTIONS 60

5.4 IMPLICATION 61

REFERENCE 62

APPENDIX 70

viii

List of Tables

TABLE 2-1: SUBJECT CHARACTERISTICS BY ABO GENOTYPE 19

TABLE 3-1A: THE "TYPE-A " DIET CHARACTERISTICS 34

TABLE 3-1B: CARDIOMETABOLIC RISK FACTORS BY TERTILES OF “TYPE-A” DIET SCORE 34

TABLE 3-2A: THE "TYPE-AB" DIET CHARACTERISTICS 35

TABLE 3-2B: CARDIOMETABOLIC RISK FACTORS BY TERTILES OF “TYPE-AB” DIET SCORES 35

TABLE 3-3A: THE "TYPE-B" DIET CHARACTERISTICS 36

TABLE 3-3B: CARDIOMETABOLIC RISK FACTORS BY TERTILES OF "TYPE-B" DIET SCORES 36

TABLE 3-4A: THE "TYPE-O" DIET CHARACTERISTICS 37

TABLE 3-4B: CARDIOMETABOLIC RISK FACTORS BY TERTILES OF "TYPE-O" DIET SCORES 37

TABLE 4A: CARDIOMETABOLIC RISK FACTORS BY "TYPE-A" DIET SCORES AND ABO GENOTYPE 47

TABLE 4B: CARDIOMETABOLIC RISK FACTORS BY "TYPE-AB" DIET SCORES AND ABO GENOTYPE 50

TABLE 4C: CARDIOMETABOLIC RISK FACTORS BY "TYPE-B" DIET SCORES AND ABO GENOTYPE 52

TABLE 4D: CARDIOMETABOLIC RISK FACTORS BY "TYPE-O" DIET SCORES AND ABO GENOTYPE 53

List of Figures

FIGURE 2-1: ABO BLOOD TYPE ASSESSMENT BY GENOTYPING SINGLE NUCLEOTIDE POLYMORPHISMS 15

FIGURE 2-2: ABO BLOOD GROUP DISTRIBUTION IN THE ENTIRE TNH COHORT (A) AND BY ETHNOCULTURAL GROUPS (B) 17

FIGURE 2-3: LEVELS OF FASTING GLUCOSE IN INDIVIDUALS WITH DIFFERENT ABO BLOOD GROUPS 20

FIGURE 2-4: LEVELS OF FASTING INSULIN IN INDIVIDUALS WITH DIFFERENT ABO BLOOD GROUPS 21

FIGURE 3-1: TOTAL NUMBER OF RECOMMENDED ITEMS TO EAT (A) OR AVOID (B) IN FOUR MAJOR FOOD GROUPS LISTEDIN THE FFQ FOR EACH "BLOOD-TYPE" DIET 30

FIGURE 3-2: DIET SCORE DISTRIBUTION FOR EACH "BLOOD TYPE” DIET 31

FIGURE 4-1: NUMBER OF INDIVIDUALS WITH MATCHED AND UNMATCHED BLOOD GROUPS ACROSS THE TERTILE TERTILE OF EACH “BLOOD-TYPE” DIET SCORE 45

FIGURE 4-2: THE ASSOCIATION BETWEEN TERTILES OF “TYPE-A” DIET SCORES AND FASTING INSULIN IN SUBJECTS WITH BLOOD TYPE A AND THE OTHERS 48

FIGURE 4-3: THE ASSOCIATION BETWEEN TERTILES OF “TYPE-A” DIET SCORES AND FASTING GLUCOSE IN SUBJECTS WITH BLOOD TYPE A AND THE OTHERS 49

FIGURE 4-4: THE ASSOCIATION BETWEEN TERTILES OF “TYPE-AB” DIET SCORES AND FASTING GLUCOSE IN SUBJECTS WITH BLOOD TYPE A AND THE OTHERS 51

List of Appendix

TABLE: FOOD ITEMS INCLUDED IN THE "BLOOD-TYPE" DIET SCORE CALCULATION 65

Chapter 1: Introduction and Literature Review

1.1 Introduction

The “Blood-Type” diet is a set of nutritional guidelines that advocate people to eat according to their blood group. Although popular with the public, the diet remains controversial because no scientific research has examined its validity. The proposed study will be the first observational study to investigate whether consumption of a diet in accordance with one's blood group is associated with favourable health outcomes.

1.2 “Blood-Type” Diet

The connection between blood group and diet was first suggested by P.J. D’Adamo, a naturopathic physician, in his book “Eat Right For Your Type” published in 1996 [1]. The

“Blood-Type” diet theory has thereafter gained widespread attention from the public. With more than 7 million copies sold in over 60 languages, this book was a New York Times bestseller [2].

Based on his perceptions of the blood group evolution and the action of lectins, D’Adamo postulates that the ABO blood group reveals the dietary habits of our ancestors and people with different blood groups process food differently; thus, adherence to a diet specific to one’s blood group could improve health and decrease risk of chronic diseases such as cardiovascular disease.

Based on the “Blood-Type” diet theory, group O is considered the ancestral blood group in humans, so their optimal diet should resemble the high animal protein diets typical of the hunter- gatherer era. In contrast, those with group A should thrive on a vegetarian diet as this blood 1

2 group was believed to have evolved when humans settled down into agrarian societies. Following the same rationale, individuals with blood group B are considered to benefit from consumption of dairy products because this blood group was believed to originate in nomadic tribes. Finally, individuals with an AB blood group are believed to benefit from a diet that is intermediate to those proposed for group A and group B. The “Blood-Type” diet also proposes that lectins, which are sugar-binding proteins found in certain foods [3], could cause agglutination if they are not compatible with an individual’s ABO blood group. However, there has been no scientific research conducted to either support his explanations or validate the “Blood-Type” diet, making this popular theory controversial.

1.3 ABO Blood Group

1.3.1 ABO System

Although little is known about the effectiveness of the “Blood-Type” diet, the ABO blood group itself is well understood. A blood group is a classification of blood based on the presence or absence of surface antigenic substance on red blood cells. Depending on the blood group system, these antigens can be proteins, carbohydrates, glycoproteins or glycolipids. To date, 30 human blood groups are recognized by the International Society of Blood Transfusion [4].

Among them, the ABO blood group is the most important one in transfusion medicine and has been studied extensively. Group O, A, B, and AB are the four subgroups within the ABO system, which is estimated to account for 44%, 35%, 16% and 5% of global population, respectively [5].

The distinction of ABO blood group relies on the structural variation of a carbohydrate antigenic substance, called H-antigen, found on red blood cells. Such variation is a result of different

3 enzymatic activities determined by the alleles of the ABO gene locus. Specifically, the IA allele encodes a glycosyltransferase that attaches α-N-acetylgalactosamine to H-antigen, producing the

A antigen. The IB allele encodes a glycosyltransferase that bonds α-D-galactose to the H antigen, creating the B antigen. In contrast, the i allele does not produce any functional enzyme due to a frameshift mutation that results in premature termination of the translation, so H antigen remains intact in this case [6]. Because the IA and IB alleles are co-dominant to each other and both are dominant over allele i, different combinations of IA, IB and i can produce different ABH antigens on human red blood cells. For example, people with blood group AB will have genotype IAIB, whereas those with blood group A can have either IAIA or IAi [7]. The discovery of the ABO blood group not only was a major breakthrough in transfusion medicine, but also draws attention to the potential importance of blood group in health and disease risk management.

1.3.2 ABO Blood Group and Diseases

As one of the first recognizable genetic variants in humans, the ABO system has been studied for its association with a variety of diseases, including cardiometabolic syndromes, cancer and microorganism infections. As for cardiometabolic diseases, research has shown that people with group O on average had 25% lower level of von Willebrand factor (VWF), which is a critical protein involved in haemostasis [8]. Considering the fact that higher VWF levels could increase the risk of ischemic stroke [9], a lower expression of VWF may explain the observation that people with blood group O had a reduced risk of venous thromboembolism [10]. Although blood group O demonstrates some degree of protection against thrombosis, it may increase the risk of diabetes. One study has shown that blood group B was associated with a decreased risk of type 2 diabetes compared to blood group O (OR = 0.44, 0.27-0.70) [11], suggesting that the association between ABO blood group and cardiometabolic diseases is disease-specific.

Regarding its association with cancers, people with blood group O may have lower risks of pancreatic cancer [12] and gastric cancer [13], but a higher risk of non-melanoma skin cancer

[14]. Gates et al. have also found that the presence of the B antigen on the red blood cells was positively associated with ovarian cancer incidence [15]. Thus, certain ABO groups may be associated with different risks in different types of cancers. Besides its associations with cardiometabolic diseases and cancers, the ABO blood group may also alter the incidences of malaria and cholera. In a matched case-control study of 567 Malian children, Rowe et al. found that group O was correlated with a 66% reduction in the odds of developing severe malaria compared to the non-O blood groups. The difference in group O may be caused by the inhibition of Plasmodium falciparum resetting, which is a parasite virulence phenotype associated with severe malaria [16]. As for cholera, an observational study in Bangladesh found that both the incidence and severity of diarrhea were higher in group O, but lower in group AB [17]. Overall, the existing research suggests that ABO blood type may play a role in modifying an individual’s susceptibility to certain diseases.

1.3.3 Blood Groups and Nutrition

To date, research on the association between blood groups and nutrition is limited.

Nonetheless, the results from several published studies have suggested that blood groups might indeed alter the relationship between diet and physiological outcomes to some extent. For example, in a study on high-fat diet and intestinal phosphatase published in 1960s, Langman et al. found that people with group A not only had the lowest level of enzyme activity at baseline, but also had lowest increase of activity after giving a high-fat diet.[18] Because phosphatase negatively regulates fat absorption in the gut [19], a lower enzyme activity may explain why people with blood group A had higher levels of serum cholesterol in another study [20]. Although

5 the findings suggested that ABO blood group might alter a metabolic pathway related to nutrient absorption, the interpretation of the results is limited because no statistical analysis was performed in these two early studies. In addition to the potential to influence nutrient absorption, a Finnish group has demonstrated that ABO blood group might have an impact on the human gut microbiota composition. Compared to individuals with the B antigen (blood group B and AB), people with non-B antigens were shown to have lower diversity of the Eubacterium rectale-

Clostridium coccoides (EREC) and Clostridium leptum (CLEPT) –groups [21]. The linkage between ABO blood group and microbiota has been further substantiated in recent studies. For example, both Helicobacter pylori [22] and Lactobacillus spp. [23] have been shown to use ABO antigens as adhesion receptors. Furthermore, research has shown that different species may have different binding affinities to ABH antigen [23] and host ABO genotype can affect the enzyme secretion and metabolism of gut microbiota [24]. Considering the role of gut microbiota on nutrient breakdown and absorption [25], the results of these studies further imply that ABO blood group could be a genetic host factor that modulates the relationship between diet and health.

Besides the ABO system, research on another glycoprotein-based blood group system, known as

MNS, has found that individuals with blood group NN tended to have a greater reduction in low density lipoprotein (LDL) cholesterol in response to a low-fat diet compared to individuals with

MN and MM blood groups. [26] However, because ABO and MNS blood group systems are functionally distinct, the generalization of the MNS finding on the current project is limited.

Overall, several studies have suggested that a potential interaction may exist among blood groups, diet and health. However, the evidence, particularly for the ABO system, remains scarce.

1.4 Cardiometabolic Diseases

1.4.1 Prevalence and Etiology

Cardiometabolic diseases, including Type 2 diabetes (T2D) and cardiovascular diseases

(CVD), are the leading causes of mortality and morbidity globally [27]. Patients with this type of disease usually exhibit hypertension, hyperglycemia, dyslipidemia, inflammation and obesity

[28]. In Canada, over 3 million and 1.3 million of people suffer from T2D and CVD respectively

[29,30]. As a huge burden on the health care system, T2D is expected to cost the Canadian healthcare system $16.9 billion a year by 2020 [31], and CVD alone costs Canadian more than

20.9 billion each year [29].

The development of cardiometabolic diseases commonly involves excess accumulation of body fat and insulin resistance in adipose, hepatic and muscle tissues [32]. Visceral and central adiposity are known to increase the level of free fatty acids in the circulation [33,34]. High levels of free fatty acids can cause hepatic gluconeogenesis and lead to excess secretion of glucose, triglycerides and very low-density lipoprotein (VLDL). This process is usually accompanied by a reduction in high-density lipoprotein (HDL) cholesterol secretion and elevation of low-density lipoprotein (LDL) cholesterol. Eventually, these physiological states lead to dysglycemia and dyslipidemia, which are early signs of cardiometabolic disorders [35].

1.4.2 Biomarkers of risk

In order to prevent and manage T2D and CVD, a number of serum biomarkers along with the anthropometric measurements are usually monitored during clinical visits to indicate the risk of these two common diseases. These biomarkers are usually associated with glucose and fat metabolism that are important in the disease pathology. Regarding glucose metabolism, insulin 6

7 regulates the cellular uptake of glucose from blood when its level is high after a meal, and glucagon promotes gluconeogenesis and glycogenolysis in the liver when the level of glucose is low [36]. In patients with insulin resistance, higher levels of insulin are required to maintain glucose uptake in peripheral tissues [37]. Pancreatic β-cells produce and secrete more insulin in response, resulting in the individual becoming hyperinsulinemic [38]. This process continues until β-cells cannot meet the increased need of insulin secretion, which results in elevation of blood glucose

[39,40]. Therefore, high levels of glucose and insulin are strong predictors of the development of

T2D [36]. A number of approaches have been developed to monitor and indicate insulin resistance and β-cell dysfunction. In epidemiologic studies, the homeostasis model assessment of insulin resistance (HOMA-IR) and β-cell dysfunction (HOMA-Beta) is a method commonly used to indicate the function of glycemic regulation. HOMA is based on the levels of fasting insulin and glucose, which are indicators of the balance between insulin secretion and glucose maintenance in the liver [41]. The equation of calculating HOMA is shown below.

HOMA-IR = (fasting plasma insulin x fasting plasma glucose) / 22.5

HOMA-Beta = (20 x fasting plasma insulin) / (fasting plasma glucose - 3.5), where insulin is measured in mU/L and glucose is measured in mmol/L [42].

Results from the HOMA method have been proven to correlate well (r>0.7) with the traditional hyperglycemic clamp method [43,44].

As for biomarkers of cardiometabolic diseases in lipid metabolism, high levels of serum cholesterol and triglycerides are common indicators of dysregulation in lipid metabolism, which is a main feature of cardiometabolic diseases. Although serum total cholesterol level is associated with cardiometabolic risk, the lipoprotein carriers of cholesterol actually play a larger role in the pathologic process [45]. Lipoprotein particles consist of triglycerides, cholesterol, phospholipids 7

8 and apolipoproteins. Based on the composition of apolipoproteins and the relative density of triglycerides and cholesterol, lipoprotein particles can be classified into several categories [45].

Low-density lipoprotein (LDL) and high-density lipoprotein (HDL) are the two that are highly related to cardiometabolic risk. The main role of LDL is to transport triglycerides and cholesterol to tissues, whereas HDL is involved in returning cholesterol back to liver for excretion [46].

Elevated levels of LDL coupled with reduced levels of HDL are known to promote the formation of atherosclerotic plaques and thus the development of the CVD [45,46]. The atherosclerotic process begins when high amounts of LDL enter the vascular endothelium and become oxidized, which attracts macrophages that engulf the extra lipids and form foam cells. Meanwhile, the activated macrophages release high levels of pro-inflammatory cytokines, such as IL-6. This attracts platelets to attach to the site of injury and promotes the proliferation of smooth muscle cells. The process is also accompanied by an elevation of C-reactive protein (CRP) secretion, which plays a critical role in the phagocytosis by macrophages [47]. As a well-established biomarker of inflammation, CRP has been consistently associated with T2D and CVD [48].

Compared to healthy individuals, people with cardiometabolic diseases tend to have levels of

CRP that are two- to four-fold higher [48]. The effects of CRP and other pro-inflammatory cytokines all perpetuate the accumulation of foam cells and smooth muscle cells at the plaque. As a result, blood vessel narrows down and blood flow is diminished, which can lead to stroke and myocardial infarction [46]. In contrast, high levels of HDL can inhibit the deposition of cholesterol and prevent the formation of atherosclerotic plaques by transporting triglycerides and cholesterol back to the liver [49]. In addition to LDL and HDL, serum triglyceride (TG) concentration is another biomarker that strongly correlates with cardiometabolic diseases [50,51].

High levels of TG are known to both impede the scavenging capacity of HDL [52,53] and promote the formation of small, dense LDL particles, which are more easily oxidized and thus 8

9 more atherogenic than their normal-sized counterpart [54]. In summary, the levels of biomarkers involved in glucose and fat metabolism all play critical roles in the pathologic process of cardiometabolic diseases and thus are often used as surrogate indicators during disease screening

[55,56].

1.4.3 Risk Factors

Sex, age, family history, ethnicity are all known to alter the risk of cardiometabolic diseases [32,57]. For example, South Asian individuals have a higher prevalence of CVD in

Canada compared to Caucasian or East Asian [57]. In addition to these demographic factors, smoking, exercise and diet represent other risk factors that can be modified through lifestyle habits [58,59]. As one of the most important modifiable risk factors, diet has been consistently shown to alter the risk of cardiometabolic diseases. Regarding specific nutrients, consumption of foods rich in dietary fiber [60,61], magnesium [62,63], and vitamin D [64,65], may be beneficial in preventing the development of T2D and CVD. In contrast, diets with a high intake of dietary cholesterol and high ratio of saturated /polyunsaturated fat can increase the risk of both T2D and

CVD [66,67,68]. As for general dietary patterns, diets high in fruit and vegetable intakes are known to significantly reduce cardiometabolic disease risk. For example, in a recent landmark study. the Mediterranean diet, which is characterised by a high intake of olive oil, nuts, fruits and vegetables, a moderate intake of fish and poultry, and a low intake of dairy and red meat, has been shown to substantially reduce the incidence of major cardiovascular events by 30% [69].

Overall, research suggests that a healthy diet is a significant modifiable lifestyle factor in the prevention and management of cardiometabolic diseases.

1.5 Summary and Rationale

In summary, both ABO blood group and dietary habits have been shown to alter the risk of cardiometabolic diseases in previous research. The associations of these diseases with diet, as well as with blood group, may imply that a potential interaction exists between diet and blood group on subsequent cardiometabolic disease risk. Considering the popularity of the “Blood-

Type” diet and its lack of scientific evidence, it will thus be worthwhile to investigate whether the effects of dietary intake on diseases are dependent on an individual’s blood group. The results of this research has the potential to either substantiate or reject the health claims of the “Blood-

Type” diet, so that evidence-based recommendations can be made by health professionals in the future.

1.6 Hypothesis and Organization of Thesis

Hypothesis

Adherence to a diet specific to one's blood group is associated with a reduced disease risk profile.

Objectives

1. To determine the effect of ABO genotype on cardiometabolic risk factors.

2. To determine the effect of “Blood-Type” diet on cardiometabolic risk factors.

3. To determine the effect of matching ABO genotype and “Blood-Type” diet on

cardiometabolic biomarkers.

Chapter 2: Effect of ABO Genotype on Cardiometabolic Risk

Factors

2.1 Abstract

Background: Previous research has suggested that ABO blood group is associated with a variety of chronic diseases, such as cancer, thrombosis and Type 2 diabetes. However, the results on some of these linkages, such as the one with T2D risk, have largely remained inconsistent and inconclusive. The objective of this chapter is to examine the relationship between ABO blood group and cardiometabolic risk factors, which can further elucidate some of these linkages from a biomarker perspective.

Methods: Subjects (n=1,455) were participants of the Toronto Nutrigenomics and Health study.

ABO blood group was determined by genotyping rs8176719 and rs8176746 in the ABO gene.

Overnight 12-hour fasting blood samples were collected to measure serum cardiometabolic biomarkers by high performance liquid chromatography. ANCOVA, with age, sex, ethnicity, and energy intake as covariates, was used to compare anthropometric measurements and biomarkers of cardiometabolic health across different blood groups

Results: Subjects with blood group O had lower levels of serum insulin, HOMA-IR, and

HOMA-Beta compared to blood groups A and B (p<0.05). There was no difference in other cardiometabolic biomarkers across ABO blood groups.

Conclusion: ABO blood type is associated with certain biomarkers for Type 2 diabetes.

2.2 Introduction

The “Blood-Type” diet advises people to eat based on their blood group to optimize health outcome. For each blood group, the recommended diet represents as a distinct dietary pattern. For example, the Type-A diet is similar to a vegetarian diet, whereas the Type-O diet resembles the low-carbohydrate diet that has high animal protein intake. Before examining whether people with certain blood group may benefit more from following the corresponding

“Blood-Type” diet, it is critical to investigate the effect of different ABO genotypes on cardiometabolic risk factors.

As a well-known genetic variant among individuals, ABO blood group is not only critical in blood transfusion, but may also alter the risk of a variety of diseases, including cardiometabolic syndromes, cancers and microorganism infections [11,13,17]. Regarding cardiometabolic diseases, research has shown that people with group O may have a reduced risk of venous thromboembolism due to their lower level of von Willebrand factor [8], which is a critical protein involved in blood clotting [10]. Although blood group O may confer some degree of protection against thrombosis, it may increase the risk of diabetes. By using the data from the

Nurses’ Health Study, Qi et al. have shown that ABO blood group is not only associated with the serum level of E-Selectin, a well-known risk factor for T2D, but also found that blood group O was associated with an increased risk of Type 2 diabetes compared to blood group B [11].

However, the results on the some of these linkages, such as the one with T2D risk, have remained inconsistent. For example, the higher risk of T2D in people with blood group O was reported in two studies [11,70], but not replicated in others [71,72,73]. In order to address this inconsistency, this study intends to elucidate the linkage between ABO blood group and cardiometabolic disease risk in a young multiethnic population. This would allow us to further 13

14 investigate the etiological relationship upstream before disease onset and to study the role of ethnicity in the linkage between ABO blood group and cardiometabolic serum biomarkers.

2.3 Material and Methods

2.3.1 Study Design and Participants

Subjects (n=1,639) were participants of the Toronto Nutrigenomics and Health (TNH)

Study, which is a cross-sectional examination of young adults aged 20 to 29 years. All subjects were recruited between October 2004 and December 2010 and completed a general health and lifestyle questionnaire, which included information on age, sex, ethnocultural group and other subject characteristics. Subjects who were likely under-reporters (less than 800 kcal per day) or over-reporters (more than 3,500 kcal per day for females or 4,500 kilocalories per day for males) of energy intake were excluded from the analyses. Subjects were also excluded if they had missing data for any of the biomarkers of interest or ABO genotype (n=184). After exclusions,

1,455 subjects (993 women and 462 men) remained. Individuals were categorized into four major ethnocultural groups: White (n=703), East Asians (n=491), South Asians (n=155), and others

(n=106).

2.3.3 ABO Genotype Identification

The Sequenom MassArray® multiplex method was used to determine the blood group of study participants by genotyping two single nucleotide polymorphisms (SNPs)

(rs8176719Del>G; rs8176746A>C) in the ABO gene. The rs8176719 SNP indicates O-allele- specific 261delG while rs8176746 determines the galactose specificity of the encoded A/B

15 transferases and thus the expression of A and B alleles [7]. Previous studies have taken advantages of the genetic information to infer blood type from single nucleotide polymorphism

(SNP), which can be more accurate and efficient compared to the traditional blood typing that is based on antibody assay [11,74]. Figure 2.1 depicts the exact relationship between these SNPs and ABO blood type assessment.

Figure 2-1: ABO blood group assessment by genotyping SNPs.

2.3.3 Cardiometabolic Risk Factor Assessment

Anthropometric measurements including height, weight, blood pressure and waist circumference were determined as previously described [75]. Body mass index (BMI; kg/m2) was calculated and physical activity was measured by questionnaire and expressed as metabolic equivalent (MET)-hours per week, as described previously [75,76]. Overnight 12-hour fasting blood samples were collected to measure serum biomarkers of cardiometabolic disease, including triglycerides, free fatty acids, C-reactive protein, fasting glucose, fasting insulin, and total-, HDL- and LDL-cholesterol, as described previously [77]. After converting insulin concentrations from pmol/L to IU/mL, the homeostasis model of insulin resistance (HOMA-IR) was calculated by

16 using the formula: (insulin * glucose)/22.5, and the homeostasis model of beta-cell function

(HOMA-Beta) was calculated by using the formula: (20 * insulin)/(glucose - 3.5).

2.3.5 Statistical Analysis

Statistical analyses were performed using the Statistical Analysis Systems (SAS)

Software program (version 9.2; SAS Institute Inc., Cary, North Carolina). The a error was set at

0.05 and reported p-values are 2-sided. Subject characteristics were compared across different

ABO blood groups by using chi-square tests for categorical variables and analysis of variance

(ANOVA) for continuous variables. Analysis of covariance (ANCOVA) was used to compare means of cardiometabolic biomarkers among different ABO blood groups. Variables that were not normally distributed were either loge or square root transformed prior to analysis, but the mean values and standard errors are displayed without transformation to facilitate interpretation.

Means compared between groups were adjusted for multiple comparisons using the Tukey-

Kramer procedure. Age, sex, ethnicity were used as covariates in the ANCOVA analysis.

2.4 Results

As for blood group determination, SNPS rs8176719 and rs8176746 were genotyped to infer the ABO blood group of study subjects. Figure 2-2A shows that Group O, A, B and AB represented 37%, 38%, 19% and 6% of the population, respectively. The frequency of each blood group was different across ethnocultural groups (P<0.05). East Asian and South Asian individuals had a higher frequency of group B and a lower frequency of group A (Figure 2-2B).

B Caucasian East Asian

Others South Asian

Figure 2-2: ABO blood group distribution in the entire TNH cohort (A) and by ethnocultural groups (B).

Subject characteristics based on the ABO blood group are summarized in Table 2-1. After adjusting for age, sex, and ethnocultural group, subject characteristics were similar across ABO blood groups, except for insulin, HOMA-IR and HOMA-Beta. Group O individuals had lower levels of fasting insulin, HOMA-IR and HOMA-Beta compared to group A or B. Although the overall association between blood group and total cholesterol was significant (p=0.043), no difference was observed among specific ABO blood group. Figure 2-3 and Figure 2-4 respectively show the levels of fasting glucose and insulin in individuals with different ABO blood groups.

Table 2-1: Subject Characteristics by ABO Genotypea Genotype Characteristic O A B AB P value Subjects [n (% of total)] 543 (37) 544 (38) 277 (19) 91 (6) Age (y) 22.7 ± 2.5b 22.8 ± 2.5 22.4 ± 2.3 22.5 ± 2.6 0.13 Sex [n (%)] 0.31 Female 358 (36) 387 (39) 187 (19) 61 (6) Male 185 (40) 157 (34) 90 (19) 30 (7) Ethnocultural group [n (%)] <0.001c White 267 (38) 307 (44) 93 (13) 36 (5) East Asian 167 (34) 166 (34) 125 (25) 33 (7) South Asian 58 (37) 40 (26) 45 (29) 12 (8) Others 51 (48) 31 (29) 14 (13) 10 (10) Body mass index (kg/m2) 23.1 ± 0.2 22.8 ± 0.2 22.7 ± 0.2 23.6 ± 0.4 0.16 Systolic blood pressure (mm Hg) 114.7 ± 0.5 114.0 ± 0.5 113.0 ± 0.7 114.3 ± 1.2 0.73 Diastolic blood pressure (mm Hg) 70.0 ± 0.4 68.9 ± 0.4 69.1 ± 0.5 69.6 ± 0.9 0.25 Waist circumference (cm) 75.1 ± 0.4 73.9 ± 0.4 73.5 ± 0.6 74.9 ± 1.0 0.28 Glucose (mmol/L) 4.80 ± 0.02 4.78 ± 0.02 4.79 ± 0.02 4.79 ± 0.04 0.95 Insulin (pmol/L) 43.1 ± 1.5 48.5 ± 1.5 51.2 ± 2.1 45.5 ± 3.7 <0.001d HOMA-IR 1.29 ± 0.05 1.45 ± 0.05 1.56 ± 0.07 1.36 ± 0.12 <0.001d HOMA-Beta 96.8 ± 3.2 109.9 ± 3.2 112.4 ± 4.5 102.5 ± 7.9 <0.001d Total cholesterol (mmol/L) 4.19 ± 0.03 4.33 ± 0.03 4.21 ± 0.05 4.2 ± 0.08 0.043 HDL cholesterol (mmol/L) 1.51 ± 0.02 1.57 ± 0.02 1.52 ± 0.02 1.52 ± 0.04 0.16 LDL cholesterol (mmol/L) 2.24 ± 0.03 2.33 ± 0.03 2.25 ± 0.04 2.26 ± 0.07 0.06 Total/HDL cholesterol 2.93 ± 0.03 2.86 ± 0.03 2.9 ± 0.05 2.89 ± 0.08 0.91 Triglycerides (mmol/L) 0.98 ± 0.02 0.95 ± 0.02 0.95 ± 0.03 0.92 ± 0.05 0.53 hs-CRP (mg/L) 1.28 ± 0.11 1.31 ± 0.11 1.07 ± 0.16 1.68 ± 0.27 0.38 Free fatty acids (µmol/L) 481.0 ± 10.7 486.2 ± 10.7 479.0 ± 15.0 480.9 ± 26.2 0.92

a HDL, high density lipoprotein; LDL, low density lipoprotein; hs-CRP, high-sensitivity C-reactive protein. HOMA- IR, homeostasis model of insulin resistance; and HOMA-Beta, homeostasis model of beta-cell function. Differences across blood groups were assessed using ANOVA for continuous variables and a χ2 test for categorical variables. b Crude mean ± SE (all such values). c Overall comparison is significantly different after a Tukey-Kramer correction (P<0.05). d (Type A, Type B) > Type O after a Tukey-Kramer correction (P<0.05).

5 4.9 4.8 4.7 4.6 4.5 4.4 4.3

Fasng Glucose (mmol/L) 4.2 4.1 4 O A B AB

Figure 2-3: Levels of fasting glucose in individuals with different ABO blood groups.

60 b b 55 ab

50 a

Fasng Insulin (pmol/L) 35

30 O A B AB

Figure 2-4: Levels of fasting insulin in individuals with different ABO blood group

2.5 Discussion

Regarding the determination of ABO blood group, the reported frequency of blood group distribution in our study population is similar to the frequency reported previously for Canada

[78], in which group O and A being the most common two blood groups. The higher frequency of blood group B observed in people from Asian ethnocultural backgrounds is also consistent with previous findings [5]. The consistency of our results with existing knowledge therefore confirms the validity of our blood-typing method.

An examination of the relationship between the ABO blood group and levels of cardiometabolic risk factors showed no association for most of the biomarkers, except for fasting insulin and the derived assessment of insulin resistance and beta-cell function. The lower levels of these risk factors in subjects with blood group O were consistently shown in both men and women and across different ethnocultural groups (data not shown). This finding is in line with previous research suggesting that an association may exist between the blood groups and diabetes risk [11,70,71,72,73]. However, it is not clear which ABO blood group is at higher risk. Some studies reported that individuals with blood group O were at higher risk [11,70], but this was not replicated in others [71,72,73]. The inconsistent results from these studies may be partly explained by ethnicity, sample sizes and other unadjusted covariates. By investigating in a multiethnic population, our study indicates that the linkage between ABO blood group and diabetes risk exist consistently across ethnicities. The observed difference in fasting insulin levels in healthy young adults also indicates that may indeed play a role in the development of insulin resistance and the aetiology of T2D development and contribute to diabetes risk later in life.

Larger prospective studies will be helpful to further confirm the etiological relationship between

ABO blood group and T2D, which has the potential facilitate T2D screening and prevention.

Chapter 3: Effect of “Blood-Type” Diet on Cardiometabolic Risk

Factors

3.1 Abstract

Background: Each of the four “Blood-Type” diets represents a distinct dietary pattern.

Promoting high vegetable and fruits consumption, the Type-A diet resembles to a vegetarian diet.

Recommending moderate intake of fish and dairy consumption, the Type-AB and Type B diets are closely related to the well-known Mediterranean diet. As the only “Blood-Type” diet that allows high intake of animal proteins, the Type-O diet is similar to a low-carbohydrate diet.

Before examining whether people with certain blood group may benefit more from following the corresponding “Blood-Type” diet, it is critical to investigate the overall dietary effects in people regardless of their ABO blood groups, which is the objective of this chapter.

Methods: Subjects (n=1,455) were participants of the Toronto Nutrigenomics and Health study.

Results: Adherence to the Type-A diet was associated with lower BMI, waist circumference, blood pressure, serum cholesterol, triglycerides, insulin, HOMA-IR and HOMA-Beta (P<0.05).

Adherence to the Type-AB diet was also associated with lower levels of these biomarkers 23

(P<0.05), except for BMI and waist circumference. Adherence to the Type-O diet was associated with lower triglycerides (P<0.0001). No association was found for the Type-B diet.

Conclusion: Adherence to certain blood type diets is associated with favorable effects on some cardiometabolic risk factors in a healthy population of young adults.

3.2 Introduction

Diet is a well-known risk factor for many chronic diseases. Recent research has been increasingly focusing on the effect of an individual’s dietary pattern. Compared to studies on individual nutrients, the evaluation of an overall dietary pattern has the potential to capture the interactions and synergistic effects of different nutrients in foods and generate more comprehensive findings that can lead to more effective dietary recommendations.

The study of “Blood-Type” diet is essentially to examine whether a personalized dietary pattern exists for people with different ABO blood groups, one of the most well-known genetic variants in the general public. Recommending different food items to consume, each of the four

“Blood-Type” diets essentially represents a distinct dietary pattern. Promoting high vegetable and fruits consumption, the Type-A diet resembles to a vegetarian diet. Type-AB diet allows certain type of poultry, seafood and dairy consumption in addition to the recommendation on vegetable and fruits, so the Type-AB diet is similar to the well-known Mediterranean diet and may be more balanced than the Type-A diet. The Type-B diet can be seen as a less strict version of the

Mediterranean diet, as it allows more consumption of animal products, including eggs and cheese. As the only “Blood-Type” diet that recommends high intake of animal proteins as well as low intake of grain products, the Type-O diet shares similar features with the low-carbohydrate 24

25 diet. Therefore, each of the four “Blood-Type” diets is similar to some existing dietary patterns that have been shown to reduce the disease risk and promote health.

As for the Type-A diet pattern, vegetarian diets contain no meat, more grains, legumes, vegetables and fruits, which therefore is associated with lower dietary intake of total fat, saturated fat, and cholesterol and higher intake of fiber [79]. Consuming a vegetarian diet has been shown to reduce the risk of hospital admission [80], cataract [81], cancers [82,83], T2D [84,85], hypertension [86], ischemic heart disease [87], and all-cause mortality [88]. As for cardiometabolic diseases specifically, people who follow a vegetarian diet are known to have higher insulin sensitivity, lower blood pressure and LDL-cholesterol, which contribute to a lower risk of diabetes [89] and ischemic heart disease [80] compared to non-vegetarians.

As for dietary pattern that resembles to the Type-AB and Type-B diets, the Mediterranean diet is characterized by a high consumption of whole grains, fruits and vegetables, legumes and olive oil, and a moderate consumption of fish, dairy and wine [90]. This popular diet has recently received wide attention from the public as many studies have demonstrated its health-promoting effects [90]. Adherence to a Mediterranean diet has been shown to decrease the risk of preterm delivery [91], gestational diabetes [92], Alzheimer’s Disease [93], Parkinson’s Disease [94], and pancreatic [95], colorectal [96], breast [97,98], and esophageal cancers [98]. Regarding to cardiovascular diseases and diabetes, research has shown that adhering to the Mediterranean diet could have long-term beneficial effects on insulin sensitivity and blood pressure management, which could lead to lower risks of Type 2 diabetes and cardiovascular diseases [69,99]. It has been suggested that the observed beneficial effect may be attributed to higher intakes of anti- oxidants, n-3 fatty acids, and fibre from the Mediterranean diet [69]. Therefore, as evident from

26 the aforementioned studies, following a Mediterranean diet may be beneficial for preventing chronic diseases, such as cardiovascular diseases and Type 2 diabetes.

As the “Blood-Type” diet that restrains the intake of grains and promotes the consumption of animal products, the Type-O diet resembles to the low-carbohydrate diet and the high-protein diet. With carbohydrate content contribute to 15%-40% of total energy intake, low- carbohydrate diets have been shown to promote weight loss and enhance glycemic control

[100,101]. High-protein diets typically are characterized by 20-30% or more of total daily calories coming from proteins [102]. High intakes of proteins have been suggested to enhance appetite suppression [103] and insulin sensitivity [104]. Nonetheless, with its recommendation for animal protein consumption, especially red meat, the Type-O diet also shares some similarity with the high-fat diet, which is known to cause extra abdominal fat accumulation [105,106], hyperphagia [107], and insulin resistance [108]. Therefore, the Type-O diet may represent as the only non-healthy “Blood-Type” diet.

Based on findings of these previous studies on different dietary patterns, it is interesting to examine whether each of the four “Blood-Type” diets can influence serum biomarker profile in our study population. The examination of this association requires the quantification of the relative adherence to the “Blood-Type” diet based on the existing data from the Food Frequency

Questionnaire (FFQ). Considering the fact that the “Blood-Type” diet recommends the consumption or avoidance of specific food items, the diet adherence score system needs to be based on these items, instead of the general food group. Furthermore, because the recommendations of the “Blood-Type” diet does not specify the exact amount of consumption for each food item [1], the score system would assign points based on actual quantity of consumption for each food item and assumes that higher consumption indicates closer adherence. Based on the

27 food item and their quantities of consumption provided by the Food Frequency Questionnaire, the scoring system aimed to generate a continuously scaled grade to capture relative diet adherence of each individual. By examining the effect of “Blood-Type” diets in a population regardless of

ABO blood group, the goal of this chapter was to investigate whether following “Blood-Type” diets is associated with a healthier biomarker profile in the general population.

3.3 Material and Methods

3.3.1 Study Design and Participants

Refer to Chapter 2 (pages 14).

3.3.2 Dietary Adherence Score Assessment

Dietary intake was assessed by a one-month, Toronto-modified Willet 196-item semi- quantitative food frequency questionnaire (FFQ) as described previously [77]. Briefly, each subject was given instructions on how to complete the FFQ by using visual aids of portion sizes to improve the measurement of self-reported food intake. Subject responses to the individual foods were converted into daily number of servings for each item. In order to quantify the adherence to each of the four “Blood-Type” diets, four different diet scores were given to each subject regardless of his or her own blood group. Based on the food items listed in the “Blood-

Type” diet [1], subjects received one positive point for consuming one serving of each recommended food item and one negative point for consuming one serving of an item on the list of foods to avoid. Foods that are listed as “Neutral” were not included in the equation and do not contribute to the final score. The equation of calculating the diet score is attached below.

“Blood-Type” Diet Score = Σ (Recommended Food Groups to Eat * Servings) – Σ

(Recommended Food Groups to Avoid * Servings)

The lists of recommended foods to eat or avoid for each ABO blood group are shown in the

Appendix. Subjects were then grouped into tertiles based on their scores for each diet, with the top tertile representing those whose diet most closely resembles the corresponding blood type diet.

3.3.3 Cardiometabolic Risk Factor Assessment

Refer to Chapter 2 (page 15).

3.3.4 Statistical Analysis

Statistical analyses were performed using the Statistical Analysis Systems (SAS)

Software program (version 9.2; SAS Institute Inc., Cary, North Carolina). The a error was set at

0.05 and reported p-values are 2-sided. Subject characteristics were compared across ABO blood groups by using chi-square tests for categorical variables and analysis of variance (ANOVA) for continuous variables. Analysis of covariance (ANCOVA) was used to compare means of biomarkers of cardiometabolic disease risk across tertiles of diet scores. Variables that were not normally distributed were either loge or square root transformed prior to analysis, but the mean

29 values and standard errors are displayed without transformation to facilitate interpretation. Means compared between groups were adjusted for multiple comparisons using the Tukey-Kramer procedure. Age, sex, ethnocultural group and energy intake were used as covariates in the

ANCOVA analysis. Physical activity and smoking were also considered; however, they did not materially alter the results and were not included in the final model.

3.4 Results

Regarding the diet adherence assessment, the method used in this study was able to generate scores that are continuously scaled and have a normal distribution. Figure 3-1A and

Figure 3-1B demonstrate the total number of recommended items to eat or avoid respectively in four major food groups listed in the FFQ for each "Blood-Type” diet. Briefly, the Type-A diet recommends high consumption of grains, fruits, and vegetables. The Type-B diet recommends high intake of dairy products and moderate intakes of other food groups. The Type-AB diet is similar to the Type-B diet in terms of recommended food groups, but has more restrictions on specific food items. For example, only eggs and fish are recommended as sources of meat for group AB individuals. The Type-O diet encourages high consumption of meats and avoidance of grain products. In order to quantify the adherence to these four blood type diets, four different scores were calculated for each subject, regardless of his or her own ABO genotype. Figure 3-2 shows the score distribution. All four scores were normally distributed and did not require any transformation.

A 30

25 Fruit & Vegetables

20 Grains Dairy 15 Meat 10 Number of Food Items

0 Type-A Diet Type-B Diet Type-AB Diet Type-O Diet

25 Fruits & Vegetable

20 Grain

Dairy 15 Meat 10 Number of Food Items 5

0 Type-A Diet Type-B Diet Type-AB Diet Type-O Diet

Figure 3-1: Total number of recommended items to eat (A) or avoid (B) in four major food groups listed in the FFQ for each "Blood-Type” diet.

300 250 200 150 100 Frequency 50 0 -96 -88 -80 -72 -64 -56 -48 -40 -32 -24 -16 -8 0 8 16 24 32 40 48

Type-A Diet Score

350 300 250 200 150

Frequency 100 50 0 -39 -33 -27 -21 -15 -9 -3 3 9 15 21 27 33 39 45 51 57 63 Type-B Diet Score

300 250 200 150 100 Frequency 50 0 -40 -32 -24 -16 -8 0 8 16 24 32 40 48 56 64 72 80 88 96 Type-AB Diet Score

350 300 250 200 150

Frequency 100 50 0 -145 -135 -125 -115 -105 -95 -85 -75 -65 -55 -45 -35 -25 -15 -5 Type-O Diet Score

Figure 3-2: Diet score distribution for each "Blood-Type” diet.

Diet characteristics and cardiometabolic risk factor profile based on score tertile of each

“Blood-Type” diet are summarized from Table 3-1 to Table 3-4. The characteristics of Type-A diet are consistent with its recommendations. Shown in Table 3-1A, subjects in the highest tertile of the Type-A diet score consumed more fruits and vegetables and less meat and dairy products

(P<0.001). A higher adherence score was also associated with lower % energy from animal fat and higher dietary fiber intake (P<0.001). Table 3-1B shows the cardiometabolic risk factor profile associated with the Type-A diet score. With increasing adherence, subjects, regardless of their ABO blood group, had lower BMI, blood pressure, waist circumference, serum total cholesterol, triglycerides, insulin, HOMA-IR, and HOMA-Beta (P<0.05).

As for the two “Blood –Type” diets that recommend dairy consumption, high adherences to the Type-AB and Type-B diets were associated with higher intakes of dairy products (P<0.05)

(Table 3-2A and Table 3-3A). Higher adherence scores for these diets are also associated with lower intake of meat and higher intake of vegetables, fruits and grains, representing as a pattern that is similar to the Mediterranean diet. As for their associations with cardiometabolic biomarkers, Table 3-2B shows that adherence to the Type-AB diet was associated with lower blood pressure, serum total cholesterol, triglycerides, fasting insulin, HOMA-IR, and HOMA-

Beta (p<0.05). However, no association was found between these biomarkers and the Type-B diet in Table 3-3B. Although the overall association between the Type-B diet adherence and the level of HDL-cholesterol was significant (p=0.04), no difference was observed between each tertile of the diet score.

Regarding the last “Blood-Type” diet, the characteristics of the Type-O diet were also consistent with the diet’s recommendations, where subjects consumed more meat and less grain

33 products as they adhered more closely to the diet (P<0.001) (Table 3-4A). Thus, the energy intake from carbohydrate decreased and the percent of energy from fat increased as the Type-O diet adherence increased (P<0.05). Table 3-4B shows that adherence to the Type-O diet was associated with lower serum triglycerides (P<0.001). No association was found for other cardiometabolic risk factors for the Type-O diet.

Table 3-1A: The "Type-A" Diet Characteristics1 Type-A Diet

Score Diet Characteristics T1 T2 T3 P value Total energy intake (Kcal) 2163 ± 29a, 2 1857 ± 29b 1880 ± 29b <0.001 Fruit and vegetables, servings/day 6.1 ± 0.2a 6.3 ± 0.2a 8.3 ± 0.2b <0.001 Cereal, servings/day 3.8 ± 0.1a 3.3 ± 0.1b 3.5 ± 0.1b <0.001 Meat, servings/day 2.3 ± 0.1a 1.7 ± 0.1b 1.4 ± 0.1b <0.001 Dairy, servings/day 2.4 ± 0.1a 1.9 ± 0.1b 1.8 ± 0.1b <0.001 Energy from carbohydrate (%) 50.1 ± 0.4a 52.5 ± 0.4b 55.8 ± 0.4c <0.001 Energy from total fat (%) 31.3 ± 0.3a 30.1 ± 0.3b 28.4 ± 0.3c <0.001 Energy from animal fat (%) 17.6 ± 0.2a 15.0 ± 0.2b 11.4 ± 0.2c <0.001 Energy from vegetable fat (%) 13.8 ± 0.3a 15.1 ± 0.3b 17.0 ± 0.3c <0.001 Fiber (g) 21.0 ± 0.5a 21.3 ± 0.5a 28.2 ± 0.5b <0.001 1 Differences among tertiles of each diet score were assessed by analysis of variance. Values with different superscript letters in each diet are significantly different after a Tukey-Kramer correction (P<0.05). 2 Mean ± SE (all such values).

Table 3-1B: Cardiometabolic Risk Factors by Type-A Diet Score1 Type-A Diet

Score Cardiometabolic Risk Factors T1 T2 T3 P value Body mass index (kg/m2) 23.7 ± 0.2a, 2 23.6 ± 0.2ab 23.1 ± 0.2b 0.03 Systolic blood pressure (mm Hg) 117.6 ± 0.5a 117.1 ± 0.5a 115.4 ± 0.5b <0.001 Diastolic blood pressure (mm Hg) 70.4 ± 0.4ab 70.8 ± 0.4a 69.4 ± 0.4b 0.009 Waist circumference (cm) 77.0 ± 0.4a 76.6 ± 0.4ab 75.6 ± 0.4b 0.02 Fasting glucose (mmol/L) 4.86 ± 0.02 4.86 ± 0.02 4.84 ± 0.02 0.5 Fasting insulin (pmol/L) 52.7 ± 1.8a 53.3 ± 1.8a 46.3 ± 1.8b 0.002 HOMA-IR 1.61 ± 0.06a 1.63 ± 0.06a 1.41 ± 0.06b 0.002 HOMA-Beta 111.7 ± 3.7a 112.7 ± 3.9a 101.2 ± 3.8b 0.007 Total cholesterol (mmol/L) 4.26 ± 0.04a 4.23 ± 0.04ab 4.14 ± 0.04b 0.02 HDL cholesterol (mmol/L) 1.46 ± 0.02 1.47 ± 0.02 1.45 ± 0.02 0.35 LDL cholesterol (mmol/L) 2.34 ± 0.03 2.31 ± 0.03 2.28 ± 0.03 0.34 Total:HDL cholesterol 3.06 ± 0.04 3.05 ± 0.04 3.03 ± 0.04 0.64 Triglycerides (mmol/L) 1.00 ± 0.02a 0.99 ± 0.02ab 0.92 ± 0.02b 0.005 hs-CRP (mg/L) 1.53 ± 0.13 1.44 ± 0.14 1.07 ± 0.13 0.06 Free fatty acids (µmol/L) 458.7 ± 12.7 476.7 ± 13.2 457.7 ± 13.1 0.56 1 HDL, high density lipoprotein; LDL, low density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model of insulin resistance; and HOMA-Beta, homeostasis model of beta-cell function. ANCOVA adjusted for age, sex, ethnicity and energy intake was used to examine associations between levels of cardiometabolic risk factors and tertiles of adherence scores in each "blood type diet". Values with different superscript letters in each diet are significantly different after a Tukey-Kramer correction (p<0.05) 2 Mean ± SE (all such values).

Table 3-2A: The Type-AB Diet Characteristics1 Type-AB Diet

Score Diet Characteristics T1 T2 T3 P value Total energy intake (Kcal) 1964 ± 29ab 1901 ± 30a 2033 ± 29b 0.007 Fruit and vegetables, servings/day 5.2 ± 0.2a 6.5 ± 0.2b 9.1 ± 0.2c <0.001 Cereal, servings/day 3.5 ± 0.1ab 3.3 ± 0.1a 3.7 ± 0.1b 0.018 Meat, servings/day 2.2 ± 0.1a 1.8 ± 0.1b 1.5 ± 0.1c <0.001 Dairy, servings/day 1.9 ± 0.1a 2.0 ± 0.1a 2.2 ± 0.1b 0.002 Energy from carbohydrate (%) 50.6 ± 0.4a 52.6 ± 0.4b 55.3 ± 0.4c <0.001 Energy from total fat (%) 30.9 ± 0.3a 29.7 ± 0.3b 29.1 ± 0.3b <0.001 Energy from animal fat (%) 17.4 ± 0.2a 14.9 ± 0.2b 11.4 ± 0.2c <0.001 Energy from vegetable fat (%) 13.4 ± 0.3a 14.8 ± 0.3b 17.7 ± 0.3c <0.001 Fiber (g) 17.8 ± 0.5a 22.3 ± 0.5b 30.5 ± 0.5c <0.001 1 Differences among tertiles of each diet score were assessed by analysis of variance. Values with different superscript letters in each diet are significantly different after a Tukey-Kramer correction (P<0.05). 2 Mean ± SE (all such values).

Table 3-2B: Cardiometabolic Risk Factors by “ Type-AB” Diet Scores1 Type-AB Diet

Score Cardiometabolic Risk Factors T1 T2 T3 P value Body mass index (kg/m2) 23.7 ± 0.2 23.5 ± 0.2 23.2 ± 0.2 0.1 Systolic blood pressure (mm Hg) 117.3 ± 0.5a 117.1 ± 0.5a 115.5 ± 0.5b 0.006 Diastolic blood pressure (mm Hg) 70.8 ± 0.4a 70.2 ± 0.4ab 69.4 ± 0.4b 0.02 Waist circumference (cm) 76.8 ± 0.4 76.7 ± 0.4 75.7 ± 0.4 0.08 Fasting glucose (mmol/L) 4.88 ± 0.02 4.83 ± 0.02 4.84 ± 0.02 0.13 Fasting insulin (pmol/L) 56.1 ± 1.7a 49.8 ± 1.8b 45.6 ± 1.9c <0.001 HOMA-IR 1.72 ± 0.06a 1.52 ± 0.06b 1.39 ± 0.06c <0.001 HOMA-Beta 119.0 ± 3.7a 107.0 ± 3.8a 98.0 ± 3.9b <0.001 Total cholesterol (mmol/L) 4.27 ± 0.04a 4.19 ± 0.04ab 4.16 ± 0.04b 0.03 HDL cholesterol (mmol/L) 1.47 ± 0.02 1.45 ± 0.02 1.46 ± 0.02 0.72 LDL cholesterol (mmol/L) 2.35 ± 0.03 2.30 ± 0.03 2.28 ± 0.03 0.19 Total:HDL cholesterol 3.07 ± 0.04 3.05 ± 0.04 3.02 ± 0.04 0.31 Triglycerides (mmol/L) 1.01 ± 0.02a 0.96 ± 0.02b 0.93 ± 0.02b 0.004 hs-CRP (mg/L) 1.45 ± 0.13 1.48 ± 0.13 1.11 ± 0.14 0.08 Free fatty acids (µmol/L) 464.9 ± 12.5 470.4 ± 12.9 459.9 ± 13.5 0.76 1 HDL, high density lipoprotein; LDL, low density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model of insulin resistance; and HOMA-Beta, homeostasis model of beta-cell function. ANCOVA adjusted for age, sex, ethnicity and energy intake was used to examine associations between levels of cardiometabolic risk factors and tertiles of adherence scores in each "blood type diet". Values with different superscript letters in each diet are significantly different after a Tukey-Kramer correction (p<0.05) 2 Mean ± SE (all such values).

Table 3-3A: The Type-B Diet Characteristics1 Type-B Diet

Score Diet Characteristics T1 T2 T3 P value Total energy intake (Kcal) 1934 ± 30a 1890 ± 29a 2076 ± 29b <0.001 Fruit and vegetables, servings/day 5.3 ± 0.2a 6.4 ± 0.2b 9.1 ± 0.2c <0.001 Cereal, servings/day 3.5 ± 0.1 3.4 ± 0.1 3.5 ± 0.1 0.56 Meat, servings/day 2.1 ± 0.1a 1.7 ± 0.1b 1.7 ± 0.1b <0.001 Dairy, servings/day 1.8 ± 0.1a 2.0 ± 0.1b 2.3 ± 0.1b <0.001 Energy from carbohydrate (%) 50.4 ± 0.4a 53.1 ± 0.4b 54.9 ± 0.4c <0.001 Energy from total fat (%) 31.0 ± 0.3a 29.6 ± 0.3b 29.0 ± 0.3b <0.001 Energy from animal fat (%) 15.6 ± 0.3a 14.6 ± 0.3b 13.6 ± 0.2c <0.001 Energy from vegetable fat (%) 15.4 ± 0.3 15.1 ± 0.3 15.5 ± 0.3 0.57 Fiber (g) 19.7 ± 0.5a 22.3 ± 0.5b 28.6 ± 0.5c <0.001 1 Differences among tertiles of each diet score were assessed by analysis of variance. Values with different superscript letters in each diet are significantly different after a Tukey-Kramer correction (P<0.05). 2 Mean ± SE (all such values).

Table 3-3B: Cardiometabolic Risk Factors by the Type-B Diet Score1 Type-B Diet

Score Cardiometabolic Risk Factors T1 T2 T3 P value Body mass index (kg/m2) 23.5 ± 0.2 23.4 ± 0.2 23.5 ± 0.2 0.77 Systolic blood pressure (mm Hg) 117.3 ± 0.5 116.4 ± 0.5 116.4 ± 0.5 0.13 Diastolic blood pressure (mm Hg) 70.9 ± 0.4 70.0 ± 0.4 69.7 ± 0.4 0.06 Waist circumference (cm) 76.8 ± 0.4 76.1 ± 0.4 76.4 ± 0.4 0.54 Fasting glucose (mmol/L) 4.87 ± 0.02 4.85 ± 0.02 4.84 ± 0.02 0.56 Fasting insulin (pmol/L) 51.3 ± 1.8 51.4 ± 1.8 50.1 ± 1.8 0.35 HOMA-IR 1.57 ± 0.06 1.56 ± 0.06 1.53 ± 0.06 0.35 HOMA-Beta 108.3 ± 3.7 111.0 ± 3.8 107.5 ± 3.9 0.44 Total cholesterol (mmol/L) 4.25 ± 0.04 4.2 ± 0.04 4.17 ± 0.04 0.21 HDL cholesterol (mmol/L) 1.47 ± 0.02 1.47 ± 0.02 1.43 ± 0.02 0.04 LDL cholesterol (mmol/L) 2.33 ± 0.03 2.29 ± 0.03 2.31 ± 0.03 0.57 Total:HDL cholesterol 3.05 ± 0.04 3.01 ± 0.04 3.08 ± 0.04 0.29 Triglycerides (mmol/L) 0.99 ± 0.02 0.97 ± 0.02 0.96 ± 0.02 0.47 hs-CRP (mg/L) 1.39 ± 0.13 1.24 ± 0.13 1.45 ± 0.14 0.5 Free fatty acids (µmol/L) 467.9 ± 12.7 460.9 ± 12.9 466.9 ± 13.2 0.98 1 HDL, high density lipoprotein; LDL, low density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model of insulin resistance; and HOMA-Beta, homeostasis model of beta-cell function. ANCOVA adjusted for age, sex, ethnicity and energy intake was used to examine associations between levels of cardiometabolic risk factors and tertiles of adherence scores in each "blood type diet". Values with different superscript letters in each diet are significantly different after a Tukey-Kramer correction (p<0.05) 2 Mean ± SE (all such values).

Table 3-4A: The "Type-O" Diet Characteristics1 Type-O Diet

Score Diet Characteristics T1 T2 T3 P value Total energy intake (Kcal) 2247 ± 28a 1868 ± 28b 1784 ± 29b <0.001 Fruit and vegetables, servings/day 7.5 ± 0.2a 6.3 ± 0.2b 6.9 ± 0.2b <0.001 Cereal, servings/day 4.3 ± 0.1a 3.3 ± 0.1b 2.8 ± 0.1c <0.001 Meat, servings/day 1.6 ± 0.1a 1.7 ± 0.1a 2.1 ± 0.1b <0.001 Dairy, servings/day 2.7 ± 0.1a 2.0 ± 0.1b 1.5 ± 0.1c <0.001 Energy from carbohydrate (%) 54.1 ± 0.4a 52.8 ± 0.4b 51.5 ± 0.4c <0.001 Energy from total fat (%) 29.1 ± 0.3a 30.1 ± 0.3b 30.5 ± 0.3b 0.003 Energy from animal fat (%) 13.6 ± 0.3a 14.7 ± 0.2b 15.4 ± 0.2b <0.001 Energy from vegetable fat (%) 15.5 ± 0.3 15.4 ± 0.3 15.0 ± 0.3 0.52 Fiber (g) 27.1 ± 0.5a 21.9 ± 0.5b 21.7 ± 0.5b <0.001 1 Differences among tertiles of each diet score were assessed by analysis of variance. Values with different superscript letters in each diet are significantly different after a Tukey-Kramer correction (P<0.05). 2 Mean ± SE (all such values).

Table 3-4B: Cardiometabolic Risk Factors by Type-O Diet Scores1 Type-O Diet

Score Cardiometabolic Risk Factors T1 T2 T3 P value Body mass index (kg/m2) 23.6 ± 0.2 23.5 ± 0.2 23.4 ± 0.2 0.79 Systolic blood pressure (mm Hg) 116.8 ± 0.5 116.8 ± 0.5 116.6 ± 0.5 0.9 Diastolic blood pressure (mm Hg) 70.3 ± 0.4 69.9 ± 0.4 70.4 ± 0.4 0.67 Waist circumference (cm) 77.0 ± 0.4 76.5 ± 0.4 75.8 ± 0.4 0.12 Fasting glucose (mmol/L) 4.84 ± 0.02 4.85 ± 0.02 4.87 ± 0.02 0.6 Fasting insulin (pmol/L) 51.1 ± 1.9 50.9 ± 1.8 50.9 ± 1.8 0.93 HOMA-IR 1.56 ± 0.06 1.56 ± 0.06 1.56 ± 0.06 0.96 HOMA-Beta 110.0 ± 4.0 110.2 ± 3.8 106.7 ± 3.9 0.7 Total cholesterol (mmol/L) 4.28 ± 0.04 4.15 ± 0.04 4.2 ± 0.04 0.054 HDL cholesterol (mmol/L) 1.46 ± 0.02 1.46 ± 0.02 1.46 ± 0.02 0.83 LDL cholesterol (mmol/L) 2.36 ± 0.03 2.26 ± 0.03 2.33 ± 0.03 0.06 Total:HDL cholesterol 3.12 ± 0.04 3.00 ± 0.04 3.03 ± 0.04 0.09 Triglycerides (mmol/L) 1.04 ± 0.03a 0.96 ± 0.02b 0.91 ± 0.02b <0.001 hs-CRP (mg/L) 1.49 ± 0.14 1.41 ± 0.13 1.2 ± 0.14 0.14 Free fatty acids (µmol/L) 464.6 ± 13.7 452.1 ± 12.8 479.0 ± 13.2 0.39 1 HDL, high density lipoprotein; LDL, low density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model of insulin resistance; and HOMA-Beta, homeostasis model of beta-cell function. ANCOVA adjusted for age, sex, ethnicity and energy intake was used to examine associations between levels of cardiometabolic risk factors and tertiles of adherence scores in each "blood type diet". Values with different superscript letters in each diet are significantly different after a Tukey-Kramer correction (p<0.05) 2 Mean ± SE (all such values).

3.5 Discussion

The results from this study show that the scoring system was able to quantify the relative diet adherence based on a continuously scaled grade. The continuum for each diet score follows a normal distribution and allows us to divide the population into three equal proportions, where the top tertile has the highest score and thus the highest “Blood-Type” diet adherence. The range of score for each “Blood-Type” diet is different from each other and the score does not necessarily have a median value of 0. For example, the value of Type-O diet score is all within the negative range. This is mainly caused by the difference in the food items that are included in the score calculation. For example, as a low-carbohydrate diet, the Type-O diet recommends high consumption of meat and low intake of grain products. Thus, all grain products have a negative value in the Type-O diet score calculation. Considering that grain products are usually consumed with multiple servings per day for most subjects, the high quantity of consumption with a negative value will drive the entire score range to the negative zone. Nonetheless, the normal distribution of Type-O diet score still allows us to quantify relative diet adherence, where the top tertile represents people who follow a diet that is more similar to the Type-O diet.

The analysis of the “Blood-Type” diets and cardiometabolic risk factors in subjects with different ABO blood groups showed that adherence to certain diets was associated with healthier levels of certain biomarkers. Adherence to the Type-A diet generates the most favorable cardiometabolic risk profile. This observation is not surprising, considering this diet's emphasis on high consumption of fruits, vegetables and grains, and low consumption of meat products, which is a dietary pattern that has been associated with a lower risk of cardiovascular diseases

[109,110,111,112,113].

Compared to the Type-A diet, adherence to the Type-AB diet was also associated with favorable levels of certain risk factors, but not for BMI and waist circumference. The absence of significance for these two risk factors may be due, in part, to the recommendation to consume eggs and dairy products for group AB individuals, which may result in a higher intake of saturated fat [114]. Nevertheless, the Type-AB diet was still associated with health benefits and it may represent as a diet more balanced than the Type-A diet. In contrast, the Type-B diet also recommends consumption of dairy products; however, no associations were observed between this diet type and the levels of cardiometabolic risk factors. The absence of effect may be partially explained by the differences in recommendation for specific food items. For example, butter and margarine were on the “avoid” list for the Type-AB diet, but they were classified under the “neutral” category for the Type-B diet. Therefore, we expect that people with high adherence to the Type-B diet might consume more butter and margarine, which may have adverse effects on cardiometabolic risk factors [115,116]. From comparing diet characteristics of the Type-AB and Type-B diets, we also noticed that increasing level of the Type-AB adherence was associated with higher vegetable fat intake, but not for the Type-B diet. Considering the high concentration of monounsaturated and polyunsaturated fat found in vegetable fat and their protective role in promoting healthy serum biomarker profile [117], it is expected that adherence to the Type-AB diet led to a more favourable outcome than the Type-B diet.

As for the “Blood-Type” diet for group O individuals, the Type-O diet recommends high intake of animal proteins and avoidance of grain products. Therefore, this diet is similar to low- carbohydrate diets in terms of reduced energy from carbohydrate sources [118]. This similarity may explain why adherence to the Type-O diet was associated with low levels of serum triglycerides (TG), since low-carbohydrate diets have been shown to decrease circulating levels

40 of these lipids [118,119,120,121,122,123,124]. The lower level may be explained by reduced production of TG in the liver and/or increased cellular uptake in response to decreased carbohydrate intake [123]. Lowering triglyceride levels is known to have an overall cardiovascular benefit [125,126,127]; however, the overall health impact based on the result of only one cardiovascular biomarker is difficult to estimate, especially when no change of LDL and

HDL cholesterol were observed for the Type-O diet in this cohort. Nonetheless, the finding indicates that adherence to the Type-O diet may lead to a favorable change of certain cardiometabolic risk factor, which is likely attributed to its low carbohydrate content.

By investigating the “Blood-Type” diets in a population regardless of ABO genotypes, we found that adhering to the Type-A, Type-AB, and Type-O diets might have favorable effects on levels of certain biomarkers of cardiometabolic risk.

Chapter 4: The effect of matching the “Blood-Type” Diet to ABO

Genotype on Cardiometabolic Risk Factors

4.1 Abstract

Background: The “Blood-Type” diet claims that individuals will improve their health and wellness from following a diet according to their own ABO blood group. Therefore, the objective of this chapter was to determine whether matching the “Blood-Type” diet with the corresponding

ABO blood group generates a more favourable cardiometabolic risk factor profile.

Methods: Subjects (n=1,455) were participants of the Toronto Nutrigenomics and Health study.

ABO blood group was determined by genotyping rs8176719 and rs8176746 in the ABO gene.

Overnight 12-hour fasting blood samples were collected to measure serum cardiometabolic biomarkers through by high performance liquid chromatography. Dietary intake was assessed using a one-month, 196-item food frequency questionnaire and a diet score was calculated to determine relative adherence to each of the four “Blood-Type” diets. ANCOVA, with age, sex, ethnicity, and energy intake as covariates, was used to compare anthropometric measurements and biomarkers of cardiometabolic health by matching the blood group with the corresponding diet. For a certain “Blood-Type” diet, levels of cardiometablic risk factors from individuals with matched ABO blood group were compared with individuals with unmatched blood groups to determine whether matching makes a difference.

Results: Matching the “Blood-Type” diets with the corresponding blood group did not change the effect size of any associations between diet scores and cardiometabolic risk factors for all four

“Blood-Type” diets.

Conclusion: The effects of “Blood-Type” diets on levels of cardiometabolic risk factors are independent of an individual’s ABO blood group, so the findings do not support the “Blood-

Type” diet hypothesis.

4.2 Introduction

D’Adamo postulates that the ABO blood group reveals the dietary habits of our ancestors and adherence to a diet specific to one’s blood group can improve health and decrease risk of chronic diseases such as cardiovascular disease. Although it was shown in the previous chapter that certain “Blood-Type” diets might have beneficial effects on some of the cardiometabolic risk factors, the finding was based on a population with different ABO blood groups. In order to examine whether individuals would benefit more from following their own “Blood-Type” diet, this study compared the levels of cardiometabolic disease risk factors between individuals with the matched blood group and the unmatched group while sharing similar diet adherence.

4.3 Material and Methods

4.3.1 Study Design and Participants

Refer to Chapter 2 (pages 14).

4.3.2 ABO Genotype Identification

Refer to Chapter 2 (page 14).

4.3.3 Diet Adherence Score Assessment

Refer to Chapter 3 (page 25).

4.3.4 Cardiometabolic Risk Factor Assessment

Refer to Chapter 2 (page 15).

4.3.5 Statistical Analysis

Statistical analyses were performed using the Statistical Analysis Systems (SAS)

Software program (version 9.2; SAS Institute Inc., Cary, North Carolina). The a error was set at

0.05 and reported p-values are 2-sided. Variables that were not normally distributed were either loge or square root transformed prior to analysis, but the mean values and standard errors are displayed without transformation to facilitate interpretation. Means compared between groups were adjusted for multiple comparisons using the Tukey-Kramer procedure. Age, sex, ethnocultural group and energy intake were used as covariates in the ANCOVA analysis.

Physical activity and smoking were also considered; however, they did not materially alter the results and were not included in the final model. To determine whether matching the blood group

44 with the corresponding diet was associated with a more favorable cardiometabolic disease risk profile, we stratified the entire population into two groups; one with the matched blood group for the diet, and the other unmatched. We next examined the interaction between diet score and the matching status on levels of each cardiometabolic disease risk factor for each blood type diet by using the Tukey-Kramer correction. When a significant interaction effect was observed, we further compared the differences in the outcome between subjects with the matched blood group and the unmatched group in each of the tertiles of diet score.

4.4 Results

Figure 4-1 shows that there was no difference in terms of distribution between the matched and unmatched group across the tertiles of all four “Blood-Type” diets.

350 500 300 400 250 A 200 300 B 150 B/AB/O 200 A/AB/O 100 100 50 0 0 T1 T2 T3 T1 T2 T3 Type-A Diet Score Type-B Diet Score

500 350 300 400 250 AB 300 200 O

200 A/B/O 150 A/B/AB 100 100 50 0 0 T1 T2 T3 T1 T2 T3

Type-AB Diet Score Type-O Diet Score

Figure 4-1: Number of individuals with matched and unmatched blood groups across the tertile of each “Blood-Type” diet score.

Tables 4A, 4B, 4C, and 4D respectively show the associations between diet scores and cardiometabolic risk factors according to ABO blood group for each of the four “Blood-Type” diets. No significant interaction effect between diet score and blood group was observed for most of the risk factors, except for fasting glucose (p=0.02), insulin (p=0.02), and HOMA-IR (p=0.01) in the Type-A diet, and fasting glucose (p=0.02) in the Type-AB diet. When comparing the levels of fasting insulin and HOMA-IR between group A individuals and the other blood groups, a significant difference was observed in the second tertile, but not in the lowest or highest tertile of the Type-A diet score (Figure 4-1). No difference in fasting glucose was observed between the two groups in any tertile of the Type-A diet score (Figure 4-2). Similarly for fasting glucose in the Type-AB diet, no difference was observed between individuals with blood group AB and those with other blood groups in any tertile (Figure 4-3).

Table 4A: Cardiometabolic Disease Risk Factors by the Type-A Diet Scores and ABO Genotypea ABO Genotype A B/AB/O P values for ABO Type-A Diet Score Tertile T1 T2 T3 T1 T2 T3 * Diet interaction Number of Subjects (% of total) 177 (33) 182 (33) 185 (34) 307 (34) 295 (32) 309 (34) Body mass index (kg/m2) 23.1 ± 0.3b 23.6 ± 0.3 23.0 ± 0.3 24.0 ± 0.2 23.5 ± 0.2 23.1 ± 0.2 0.15 Systolic blood pressure (mm Hg) 117.1 ± 0.7 117.4 ± 0.7 115.8 ± 0.7 117.9 ± 0.6 116.9 ± 0.6 115.2 ± 0.6 0.48 Diastolic blood pressure (mm Hg) 69.6 ± 0.6 70.8 ± 0.6 68.6 ± 0.6 70.8 ± 0.5 70.7 ± 0.5 69.8 ± 0.5 0.39 Waist circumference (cm) 76.2 ± 0.6 76.7 ± 0.6 75.2 ± 0.6 77.5 ± 0.5 76.4 ± 0.5 75.8 ± 0.5 0.32 Glucose (mmol/L) 4.82 ± 0.03 4.91 ± 0.03 4.83 ± 0.03 4.88 ± 0.02 4.84 ± 0.02 4.84 ± 0.02 0.02c Insulin (pmol/L) 51.9 ± 2.8 60.5 ± 2.7 47.7 ± 2.8 54.4 ± 2.1 48.6 ± 2.2 45.6 ± 2.1 0.02d HOMA-IR 1.57 ± 0.09 1.87 ± 0.09 1.45 ± 0.09 1.68 ± 0.07 1.48 ± 0.07 1.39 ± 0.07 0.01d HOMA-Beta 114.6 ± 5.9 123.4 ± 5.8 105.2 ± 5.9 112.9 ± 4.5 105.8 ± 4.6 99.1 ± 4.5 0.26 Total cholesterol (mmol/L) 4.37 ± 0.06 4.34 ± 0.06 4.15 ± 0.06 4.21 ± 0.05 4.16 ± 0.05 4.14 ± 0.05 0.09 HDL cholesterol (mmol/L) 1.51 ± 0.03 1.47 ± 0.03 1.46 ± 0.03 1.44 ± 0.02 1.46 ± 0.02 1.44 ± 0.02 0.31 LDL cholesterol (mmol/L) 2.42 ± 0.05 2.4 ± 0.05 2.29 ± 0.05 2.3 ± 0.04 2.26 ± 0.04 2.28 ± 0.04 0.16 Total:HDL cholesterol 3.01 ± 0.06 3.10 ± 0.06 3.00 ± 0.06 3.09 ± 0.04 3.02 ± 0.05 3.04 ± 0.04 0.25 Triglycerides (mmol/L) 0.96 ± 0.04 1.01 ± 0.04 0.9 ± 0.04 1.02 ± 0.03 0.97 ± 0.03 0.93 ± 0.03 0.15 hs-CRP (mg/L) 1.48 ± 0.21 1.56 ± 0.2 1.02 ± 0.2 1.62 ± 0.16 1.34 ± 0.16 1.09 ± 0.16 0.25 Free fatty acids (µmol/L) 449.1 ± 19.9 481.6 ± 19.8 471.1 ± 19.9 474.3 ± 15.2 470.2 ± 15.8 448.5 ± 15.4 0.24 a HDL, high density lipoprotein; LDL, low density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model of insulin resistance; and HOMA-Beta, homeostasis model of beta-cell function. ANCOVA adjusted for age, sex, ethnicity and energy intake was used to examine the interaction effect between the ABO blood group and diet adherence on levels of cardiometabolic risk factors. The Tukey- Kramer procedure was used to adjust for multiple comparisons between groups within each ANCOVA. b Mean ± SE (all such values). c Overall interaction is significant after a Tukey-Kramer correction (P<0.05). d (T2 in Group A) > (T2 in Group B/AB/O) after a Tukey-Kramer correction (P<0.05).

5 4.9 4.8 4.7 4.6 4.5 A 4.4 B/AB/O

Glucose (mmol/L) 4.3 4.2 4.1 4 T1 T2 T3 Type-A Diet Score

Figure 4-2: Level of fasting glucose between individuals with blood group A and others by tertiles of Type-A diet score

30 A

Insulin (pmol/L) B/AB/O 20

0 T1 T2 T3 Type-A Diet Score

Figure 4-3: Level of fasting insulin between individuals with blood group A and others by tertiles of Type-A diet score.

Table 4B: Cardiometabolic Disease Risk Factors by the Type-AB Diet Scores and ABO Genotypea ABO Genotype AB A/B/O P values for ABO Type-AB Diet Score Tertile T1 T2 T3 T1 T2 T3 * Diet interaction Number of Subjects (% of total) 26 (29) 35 (38) 30 (33) 464 (34) 441 (32) 459 (34) Body mass index (kg/m2) 24.3 ± 0.7 25.1 ± 0.6 22.7 ± 0.6 23.6 ± 0.2 23.4 ± 0.2 23.2 ± 0.2 0.07 Systolic blood pressure (mm Hg) 116.4 ± 1.8 118.9 ± 1.6 114.9 ± 1.7 117.3 ± 0.5 116.9 ± 0.5 115.6 ± 0.5 0.39 Diastolic blood pressure (mm Hg) 71.4 ± 1.6 70.6 ± 1.4 69.4 ± 1.5 70.8 ± 0.4 70.2 ± 0.4 69.4 ± 0.4 0.96 Waist circumference (cm) 77.5 ± 1.6 78.9 ± 1.4 74.2 ± 1.5 76.8 ± 0.4 76.5 ± 0.4 75.8 ± 0.4 0.14 Glucose (mmol/L) 4.74 ± 0.07 4.84 ± 0.06 4.95 ± 0.06 4.89 ± 0.02 4.83 ± 0.02 4.84 ± 0.02 0.02c Insulin (pmol/L) 56.6 ± 6.8 48.4 ± 5.9 41.7 ± 6.3 56.0 ± 1.8 50.0 ± 1.8 45.9 ± 1.9 0.81 HOMA-IR 1.67 ± 0.22 1.48 ± 0.19 1.29 ± 0.20 1.72 ± 0.06 1.52 ± 0.06 1.4 ± 0.06 0.94 HOMA-Beta 133.6 ± 14.3 101.8 ± 12.4 82.7 ± 13.4 118.2 ± 3.8 107.4 ± 3.9 99.0 ± 4.1 0.15 Total cholesterol (mmol/L) 4.29 ± 0.15 4.14 ± 0.13 4.10 ± 0.14 4.27 ± 0.04 4.19 ± 0.04 4.16 ± 0.04 0.91 HDL cholesterol (mmol/L) 1.45 ± 0.07 1.42 ± 0.06 1.49 ± 0.06 1.47 ± 0.02 1.45 ± 0.02 1.46 ± 0.02 0.8 LDL cholesterol (mmol/L) 2.35 ± 0.12 2.32 ± 0.11 2.22 ± 0.12 2.35 ± 0.03 2.30 ± 0.03 2.28 ± 0.04 0.85 Total:HDL cholesterol 3.17 ± 0.14 3.00 ± 0.12 2.94 ± 0.13 3.07 ± 0.04 3.05 ± 0.04 3.02 ± 0.04 0.73 Triglycerides (mmol/L) 1.09 ± 0.09 0.87 ± 0.08 0.86 ± 0.08 1.01 ± 0.02 0.96 ± 0.02 0.94 ± 0.03 0.26 hs-CRP (mg/L) 1.68 ± 0.5 2.48 ± 0.44 1.03 ± 0.47 1.44 ± 0.13 1.39 ± 0.14 1.11 ± 0.14 0.21 Free fatty acids (µmol/L) 456.3 ± 48.9 529.4 ± 42.4 400.1 ± 45.7 465.7 ± 12.8 466.1 ± 13.3 464.3 ± 13.8 0.11 a HDL, high density lipoprotein; LDL, low density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model of insulin resistance; and HOMA-Beta, homeostasis model of beta-cell function. ANCOVA adjusted for age, sex, ethnicity and energy intake was used to examine the interaction effect between the ABO blood group and diet adherence on levels of cardiometabolic risk factors. The Tukey- Kramer procedure was used to adjust for multiple comparisons between groups within each ANCOVA. b Mean ± SE (all such values). c Overall interaction is significant after a Tukey-Kramer correction (P<0.05).

5.2

4.8

4.6 AB

A/B/O Glucose (mmol/L) 4.4

4.2

4 T1 T2 T3 Type-AB Diet Score

Figure 4-4: Level of fasting glucose between individuals with blood group A and others by tertiles of Type-AB diet score.

Table 4C: Cardiometabolic Disease Risk Factors by the Type-B Diet Scores and ABO Genotypea ABO Genotype B A/AB/O P values for ABO Type-B Diet Score Tertile T1 T2 T3 T1 T2 T3 * Diet interaction Number of Subjects (% of total) 87 (31) 103 (38) 87 (31) 395 (34) 383 (32) 400 (34) Body mass index (kg/m2) 23.9 ± 0.4 22.7 ± 0.4 23.8 ± 0.4 23.5 ± 0.2 23.5 ± 0.2 23.4 ± 0.2 0.05 Systolic blood pressure (mm Hg) 116.5 ± 1.0 115.4 ± 0.9 116.6 ± 1.0 117.5 ± 0.5 116.6 ± 0.5 116.3 ± 0.5 0.51 Diastolic blood pressure (mm Hg) 69.8 ± 0.9 69.8 ± 0.8 70.4 ± 0.9 71.1 ± 0.4 70.0 ± 0.5 69.5 ± 0.5 0.22 Waist circumference (cm) 76.7 ± 0.9 75.0 ± 0.8 76.7 ± 0.9 76.8 ± 0.4 76.4 ± 0.5 76.4 ± 0.5 0.34 Glucose (mmol/L) 4.85 ± 0.04 4.85 ± 0.04 4.82 ± 0.04 4.87 ± 0.02 4.84 ± 0.02 4.85 ± 0.02 0.81 Insulin (pmol/L) 54.1 ± 3.8 52.2 ± 3.5 57.9 ± 3.8 50.7 ± 1.9 51.1 ± 2.0 48.4 ± 2.0 0.9 HOMA-IR 1.67 ± 0.12 1.59 ± 0.11 1.78 ± 0.12 1.55 ± 0.06 1.55 ± 0.06 1.48 ± 0.06 0.93 HOMA-Beta 114.3 ± 8.1 112.4 ± 7.4 122.1 ± 8.1 106.8 ± 4.1 110.5 ± 4.2 104.3 ± 4.2 0.78 Total cholesterol (mmol/L) 4.21 ± 0.08 4.19 ± 0.08 4.13 ± 0.08 4.26 ± 0.04 4.21 ± 0.04 4.19 ± 0.04 0.96 HDL cholesterol (mmol/L) 1.43 ± 0.04 1.51 ± 0.04 1.42 ± 0.04 1.48 ± 0.02 1.46 ± 0.02 1.43 ± 0.02 0.09 LDL cholesterol (mmol/L) 2.32 ± 0.07 2.26 ± 0.06 2.26 ± 0.07 2.33 ± 0.04 2.30 ± 0.04 2.33 ± 0.04 0.88 Total:HDL cholesterol 3.10 ± 0.08 2.92 ± 0.07 3.11 ± 0.08 3.04 ± 0.04 3.04 ± 0.04 3.08 ± 0.04 0.2 Triglycerides (mmol/L) 1.01 ± 0.05 0.92 ± 0.05 0.99 ± 0.05 0.98 ± 0.03 0.98 ± 0.03 0.95 ± 0.03 0.53 hs-CRP (mg/L) 1.39 ± 0.28 1.02 ± 0.26 1.36 ± 0.28 1.40 ± 0.14 1.30 ± 0.15 1.48 ± 0.15 0.61 Free fatty acids (µmol/L) 490.3 ± 27.4 423.8 ± 25.1 471.6 ± 27.6 463.6 ± 13.8 471.5 ± 14.3 466.6 ± 14.2 0.28 a HDL, high density lipoprotein; LDL, low density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model of insulin resistance; and HOMA-Beta, homeostasis model of beta-cell function. ANCOVA adjusted for age, sex, ethnicity and energy intake was used to examine the interaction effect between the ABO blood group and diet adherence on levels of cardiometabolic risk factors. The Tukey- Kramer procedure was used to adjust for multiple comparisons between groups within each ANCOVA. b Mean ± SE (all such values).

Table 4D: Cardiometabolic Disease Risk Factors by the Type-O Diet Score and ABO Genotypea ABO Genotype O A/B/AB P values for ABO Type-O Diet Score Tertile T1 T2 T3 T1 T2 T3 * Diet interaction Number of Subjects (% of total) 182 (34) 182 (34) 179 (32) 300 (33) 319 (35) 293 (32) Body mass index (kg/m2) 23.8 ± 0.3 23.6 ± 0.3 23.3 ± 0.3 23.4 ± 0.2 23.4 ± 0.2 23.4 ± 0.2 0.83 Systolic blood pressure (mm Hg) 117.0 ± 0.7 117.2 ± 0.7 116.5 ± 0.7 116.6 ± 0.6 116.6 ± 0.6 116.6 ± 0.6 0.75 Diastolic blood pressure (mm Hg) 71.5 ± 0.6 70.4 ± 0.6 70.2 ± 0.6 69.6 ± 0.5 69.7 ± 0.5 70.4 ± 0.5 0.14 Waist circumference (cm) 77.4 ± 0.6 76.9 ± 0.6 76.3 ± 0.6 76.8 ± 0.5 76.2 ± 0.5 75.5 ± 0.5 0.93 Glucose (mmol/L) 4.86 ± 0.03 4.83 ± 0.03 4.88 ± 0.03 4.83 ± 0.02 4.86 ± 0.02 4.86 ± 0.02 0.39 Insulin (pmol/L) 48.0 ± 2.8 46.3 ± 2.7 47.5 ± 2.7 53.1 ± 2.3 53.9 ± 2.2 53.3 ± 2.2 0.79 HOMA-IR 1.46 ± 0.09 1.41 ± 0.09 1.45 ± 0.09 1.62 ± 0.07 1.65 ± 0.07 1.64 ± 0.07 0.72 HOMA-Beta 103.5 ± 5.8 99.2 ± 5.7 100.8 ± 5.8 114.3 ± 4.8 117.1 ± 4.6 111.2 ± 4.7 0.98 Total cholesterol (mmol/L) 4.26 ± 0.06 4.10 ± 0.06 4.13 ± 0.06 4.30 ± 0.05 4.18 ± 0.05 4.25 ± 0.05 0.76 HDL cholesterol (mmol/L) 1.45 ± 0.03 1.45 ± 0.03 1.42 ± 0.03 1.46 ± 0.02 1.46 ± 0.02 1.49 ± 0.02 0.43 LDL cholesterol (mmol/L) 2.33 ± 0.05 2.20 ± 0.05 2.30 ± 0.05 2.37 ± 0.04 2.29 ± 0.04 2.34 ± 0.04 0.85 Total:HDL cholesterol 3.16 ± 0.06 2.99 ± 0.06 3.06 ± 0.06 3.10 ± 0.05 3.00 ± 0.05 3.01 ± 0.05 0.68 Triglycerides (mmol/L) 1.07 ± 0.04 0.98 ± 0.04 0.91 ± 0.04 1.03 ± 0.03 0.95 ± 0.03 0.91 ± 0.03 0.65 hs-CRP (mg/L) 1.48 ± 0.20 1.25 ± 0.20 1.34 ± 0.20 1.50 ± 0.17 1.50 ± 0.16 1.12 ± 0.16 0.81 Free fatty acids (µmol/L) 470.7 ± 19.9 446.2 ± 19.2 480.8 ± 19.7 460.8 ± 16.4 455.5 ± 15.5 477.9 ± 16.0 0.89 a HDL, high density lipoprotein; LDL, low density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model of insulin resistance; and HOMA-Beta, homeostasis model of beta-cell function. ANCOVA adjusted for age, sex, ethnicity and energy intake was used to examine the interaction effect between the ABO blood group and diet adherence on levels of cardiometabolic risk factors. The Tukey- Kramer procedure was used to adjust for multiple comparisons between groups within each ANCOVA. b Mean ± SE (all such values).

4.5 Discussion

In order to examine whether individuals would benefit more from following their own

“Blood-Type” diet, the levels of cardiometabolic disease risk factors were compared between individuals with the matched and unmatched blood group while sharing similar diet adherence.

The number of subjects with the matched blood group for a certain diet (eg. Blood type A for the

Type-A diet) and with the unmatched group (eg. Blood type B, AB, and O for the Type-A diet) were evenly distributed across the diet tertiles. This finding indicates that people with certain

ABO blood groups did not have higher adherence to their own “Blood-Type” diet.

No significant interaction effects were observed between diet adherence and blood group for most of the risk factors, suggesting that effects of following “Blood-Type” diets are independent of an individual’s blood group. Although there were significant interaction effects for fasting glucose, insulin and HOMA-IR for the Type-A diet, and fasting glucose for the Type-

AB diet, those interactions may be due to chance since we did not apply the most conservative

Bonferroni post-hoc test to correct for multiple comparisons. Even if the interaction effects were not due to chance, those findings would not support the claim that matching the diet with the corresponding blood group results in more favorable effects. In the case of the Type-A diet, the significant interaction effects were mainly driven by higher levels of insulin and HOMA-IR in the second tertile for those with blood group A. Moving from low adherence to high adherence, group A individuals did not demonstrate more favorable changes in these biomarkers. As for fasting glucose levels with the Type-AB diet, subjects with blood group AB had slightly higher glucose concentrations as they adhered to the diet more closely, while the other blood groups showed no differences. These findings, therefore, demonstrate that matching the diet with the corresponding blood group was not associated with any additional benefits and may even be

55 associated with some adverse effects. However, the trend of adverse effects in AB individuals may be caused by a smaller sample size in this blood group.

For those in the unmatched blood group, we also tested whether each “Blood-Type” diet was associated with any of the outcomes by matching to each of the other blood groups (data not shown); however, no significant interactions were observed. Therefore, the associations observed with the “Blood-Type” diets were unrelated to any individual blood group.

Chapter 5: Overview and General Discussion

5.1 Overview

Our findings show that adherence to certain “Blood-Type” diets is associated with a favorable profile for certain cardiometabolic risk factors in young adults, but these associations were not related to an individual’s ABO blood group. To our knowledge, this is the first study to examine the association between the “Blood-Type” diets and biomarkers of cardiometabolic health, and the findings do not support the popular diet’s hypothesis.

An examination of the relationship between the ABO blood group and levels of cardiometabolic risk factors showed no association for most of the risk factors, except for fasting insulin and the derived assessment of insulin resistance and beta-cell function. The lower levels of these risk factors in subjects with blood group O were consistently shown in both men and women and across different ethnocultural groups (data not shown). This finding is in line with previous research suggesting that an association may exist between the blood groups and diabetes risk [11,70,71,72,73]. However, it is not clear which ABO blood group confers the best protection. Some studies reported that individuals with blood group O were at higher risk [11,70], but this was not replicated in other studies [71,72,73]. The inconsistent results may be partly explained by differences in sample sizes and other unadjusted covariates. Nevertheless, the observed difference in fasting insulin levels across different blood groups in the present study of young adults indicates that ABO may indeed contribute to diabetes risk later in life.

Regarding the association between “Blood-Type” diet and health outcome, the association between the Type-A diet adherence and favorable cardiometabolic risk profile is not surprising, considering this diet's emphasis on high consumption of fruits, vegetables and grains, and low

57 consumption of meat products, which is similar to a dietary pattern that has been recommended by various health agencies because of its association with a lower risk of cardiovascular diseases

[109,110,111,112,113]. Adherence to the Type-AB diet was also associated with favorable levels of several risk factors, which can be attributed to its similarity with the Mediterranean diet. No associations were observed between the Type-B diet and the cardiometabolic risk factors. This could be partially explained by its allowance on margarine and butter compared to the Type-AB diet. The Type-O diet is similar to low-carbohydrate diets [118], which may explain why adherence to this type of diet was associated with lower serum triglycerides, as previously observed for other low-carbohydrate diets [119,120]. By investigating the “Blood-Type” diets in a population with different ABO genotypes, we found that adhering to the Type-A, Type-AB, or

Type-O diets was associated with favorable effects on levels of certain biomarkers of cardiometabolic disease risk.

In order to examine whether individuals would benefit more from following their own

“Blood-Type” diet, the levels of cardiometabolic disease risk factors were compared between individuals with the matched blood group and the unmatched group while sharing similar diet adherence. However, no significant interaction effects were observed between diet adherence and blood group for most of the risk factors, suggesting that the effects of following “Blood-Type” diets are independent of an individual’s blood group. Although there were significant interaction effects for fasting glucose, insulin and HOMA-IR for the Type-A diet, and fasting glucose for the

Type-AB diet, those interactions may be due to chance since we did not apply the most conservative Bonferroni post-hoc test to correct for multiple comparisons. Even if not due to chance, those findings would not support the claim that matching the diet with the corresponding blood group results in more favorable effects. In the case of the Type-A diet, the significant interaction effects were mainly driven by higher levels of insulin and HOMA-IR in the second 57

58 tertile for those with blood group A. Moving from low adherence to high adherence, group A individuals did not demonstrate more favorable changes in these biomarkers. As for fasting glucose levels with the Type-AB diet, subjects with blood group AB had slightly higher glucose concentrations as they adhered to the diet more closely, while the other blood groups showed no differences. These findings, therefore, demonstrate that matching the diet with the corresponding blood group was not associated with any additional benefits and may even be associated with some adverse effects. For those in the unmatched blood group, we also tested whether each blood type diet was associated with any of the outcomes by matching to each of the other blood groups

(data not shown); however, no significant interactions were observed. Therefore, the associations observed with the blood type diets were unrelated to any individual blood group.

Several previous studies have questioned the validity of the blood type diets. Based on phylogenetic analysis of human ABO alleles, blood group A has been suggested to be the ancestral human blood group [128,129], rather than group O as postulated by D’Adamo [1]. As for the claim that certain food items contain lectins incompatible with an individual’s ABO blood group, studies to date suggest no ABO-specific agglutination [130]. The absence of scientific studies supporting the “Blood-Type” diets was reported by a recent systematic review [131]. In fact, D’Adamo mentioned in his book published in 1996 that several clinical trials were expected to be completed within 2 years. However, to date, their results have not been published [1].

5.2 Limitations

The present study has some limitations. Although we adjusted for age, sex, ethnicity and energy intake and tested physical activity and smoking as potential covariates, the observed

59 associations between blood type diet scores and cardiometabolic risk factors could be due to residual confounding. However, residual confounding is not likely to explain why there would be no differential association between ABO genotypes. Although we adjusted for multiple comparisons, it is still possible that some of the observed associations were due to chance. The study population consisted of an unequal distribution of different ethnocultural groups, which were shown to have a different prevalence of ABO blood groups [5] and might have different dietary patterns [75]. However, the associations between diet adherence and levels of biomarkers were still evident after adjusting for ethnicity. Since the scoring system in the present study only assessed relative adherence to each of the blood type diets, we could not determine the absolute number of people who strictly followed any of the diets. However, the observed associations between cardiometabolic risk factors and diet scores suggest that even relatively high adherence might be associated with a favorable cardiometabolic risk profile, but this was independent of

ABO genotype. Previous studies using diet scores have quantified relative adherence by deriving the score proportionally based on the recommended amount of consumption [132]. However, this approach may not be appropriate in quantifying the adherence to the “Blood-Type” diets, because it recommends people to eat or avoid specific food items without specifying the actual amount of consumption. Assigning points based on quantity of consumption may not give fair weight for each food item, in which products that are often consumed in large quantities, such as grains and cereals, may have larger contribution on the final score. This might explain why Type-O diet score is within the negative range, as all grain products contribute negatively. Nonetheless, without knowing the recommended quantity for each food item, assigning points based on actual quantity of consumption could be a relatively prudent method. Indeed, the results show that our scoring system is continuously scaled and the observed associations between the diet scores and the levels of cardiometabolic risk factors were consistent with previous studies 59

[84,86,90,100,101,102].

5.3 Future Directions

The present study was based on a cross-sectional study, so the findings could not assess the impact of dietary change on individual health. The generalizability of this study may also be limited as most study subjects were young and healthy. In order to address these limitations and further confirms the findings of the current study, we will test the same hypothesis in a different population, the Toronto Healthy Diet Study Cohort (NCT00516620). Study subjects are individuals with Type 2 diabetes aged from 18 to 82 and were randomized to either follow a low-

GI diet rich in fruits, vegetables and whole grains for 6 months or continue the existing diet.

Studying the “Blood-Type” diet hypothesis in this randomized clinical trial will allow us to investigate whether individuals of certain blood group, such as group A, may benefit more from adhering to a low-GI diet. The combination of results from two cohorts will provide solid evidence to test the validity of “Blood-Type” diet.

5.4 Implications

In summary, the present study is the first to investigate the validity of the “Blood-Type” diets and we showed that adherence to certain “Blood-Type” diets is associated with some favorable cardiometabolic disease risk profiles, which may explain anecdotal evidence supporting

61 these diets. However, the findings showed that the observed associations were independent of

ABO blood group and, therefore, the findings do not support the diet’s hypothesis. The results of this study, therefore, reject the health claims of blood type diets. These findings will be relevant to healthcare professionals who are often asked about the efficacy of these diets, as well as to those in the general public who are considering adopting a diet to improve their health.

References

1. D'Adamo P, Whitney C (1996) Eat Right 4 Your Type: The individualized diet solution to staying healthy, living longer & achieving your ideal weight. New York: Putnam. 2. (2012) EAT RIGHT FOR YOUR TYPE. OFFICAL WEBSITE OF DR. PETER D'ADAMO & THE BLOOD TYPE DIET. Hoop-La-Joop, LLC, Inc. 3. Ghazarian H, Idoni B, Oppenheimer SB (2011) A glycobiology review: carbohydrates, lectins and implications in cancer therapeutics. Acta Histochem 113: 236-247. 4. (October 2008) Table of blood group systems. International Society of Blood Transfusion. 5. Mourant AE, Kopec AC, Domaniewska-Sobczak K (1976) The Distribution of the Human Blood Groups and Other Polymorphisms. London: Oxford University Press. 6. Anstee DJ (2010) The relationship between blood groups and disease. Blood 115: 4635-4643. 7. Yamamoto F, Cid E, Yamamoto M, Blancher A (2012) ABO research in the modern era of genomics. Transfus Med Rev 26: 103-118. 8. Jenkins PV, O'Donnell JS (2006) ABO blood group determines plasma von Willebrand factor levels: a biologic function after all? Transfusion 46: 1836-1844. 9. Bongers TN, de Maat MP, van Goor ML, Bhagwanbali V, van Vliet HH, et al. (2006) High von Willebrand factor levels increase the risk of first ischemic stroke: influence of ADAMTS13, inflammation, and genetic variability. Stroke 37: 2672-2677. 10. Wu O, Bayoumi N, Vickers MA, Clark P (2008) ABO(H) blood groups and vascular disease: a systematic review and meta-analysis. J Thromb Haemost 6: 62-69. 11. Qi L, Cornelis MC, Kraft P, Jensen M, van Dam RM, et al. (2010) Genetic variants in ABO blood group region, plasma soluble E-selectin levels and risk of type 2 diabetes. Hum Mol Genet 19: 1856-1862. 12. Wolpin BM, Kraft P, Gross M, Helzlsouer K, Bueno-de-Mesquita HB, et al. (2010) Pancreatic cancer risk and ABO blood group alleles: results from the pancreatic cancer cohort consortium. Cancer Res 70: 1015-1023. 13. Aird I, Bentall HH, Roberts JA (1953) A relationship between cancer of stomach and the ABO blood groups. Br Med J 1: 799-801. 14. Xie J, Qureshi AA, Li Y, Han J (2010) ABO blood group and incidence of skin cancer. PLoS One 5: e11972. 15. Gates MA, Wolpin BM, Cramer DW, Hankinson SE, Tworoger SS (2011) ABO blood group and incidence of epithelial ovarian cancer. Int J Cancer 128: 482-486. 16. Rowe JA, Handel IG, Thera MA, Deans AM, Lyke KE, et al. (2007) Blood group O protects against severe Plasmodium falciparum malaria through the mechanism of reduced rosetting. Proc Natl Acad Sci U S A 104: 17471-17476. 17. Glass RI, Holmgren J, Haley CE, Khan MR, Svennerholm AM, et al. (1985) Predisposition for cholera of individuals with O blood group. Possible evolutionary significance. Am J Epidemiol 121: 791-796. 18. Langman MJ, Leuthold E, Robson EB, Harris J, Luffman JE, et al. (1966) Influence of diet on the "intestinal" component of serum alkaline phosphatase in people of different ABO blood groups and secretor status. Nature 212: 41-43. 19. Lalles JP (2010) Intestinal alkaline phosphatase: multiple biological roles in maintenance of intestinal homeostasis and modulation by diet. Nutr Rev 68: 323-332. 20. Langman MJ, Elwood PC, Foote J, Ryrie DR (1969) ABO and Lewis blood-groups and serum-cholesterol. Lancet 2: 607-609. 62

21. Makivuokko H, Lahtinen SJ, Wacklin P, Tuovinen E, Tenkanen H, et al. (2012) Association between the ABO blood group and the human intestinal microbiota composition. BMC Microbiol 12: 94. 22. Boren T, Falk P, Roth KA, Larson G, Normark S (1993) Attachment of Helicobacter pylori to human gastric epithelium mediated by blood group antigens. Science 262: 1892-1895. 23. Uchida H, Kinoshita H, Kawai Y, Kitazawa H, Miura K, et al. (2006) Lactobacilli binding human A-antigen expressed in intestinal mucosa. Res Microbiol 157: 659-665. 24. Hoskins LC, Boulding ET (1976) Degradation of blood group antigens in human colon ecosystems. I. In vitro production of ABH blood group-degrading enzymes by enteric bacteria. J Clin Invest 57: 63-73. 25. Wu GD, Chen J, Hoffmann C, Bittinger K, Chen YY, et al. (2011) Linking long-term dietary patterns with gut microbial enterotypes. Science 334: 105-108. 26. Birley AJ, MacLennan R, Wahlqvist M, Gerns L, Pangan T, et al. (1997) MN blood group affects response of serum LDL cholesterol level to a low fat diet. Clin Genet 51: 291-295. 27. Kraushaar LE, Kramer A (2009) Are we losing the battle against cardiometabolic disease? The case for a paradigm shift in primary prevention. BMC Public Health 9: 64. 28. Lee DS, Chiu M, Manuel DG, Tu K, Wang X, et al. (2009) Trends in risk factors for cardiovascular disease in Canada: temporal, socio-demographic and geographic factors. CMAJ 181: E55-66. 29. Foundation HaS (2013) Costs of cardiovascular diseases. 30. (2012) The prevalence and costs of diabetes. Canadian Diabetes Association. 31. Association CD (2013) The prevalence and costs of diabetes. 32. Leiter LA, Fitchett DH, Gilbert RE, Gupta M, Mancini GB, et al. (2011) Cardiometabolic risk in Canada: a detailed analysis and position paper by the cardiometabolic risk working group. Can J Cardiol 27: e1-e33. 33. Despres JP, Lemieux I, Bergeron J, Pibarot P, Mathieu P, et al. (2008) Abdominal obesity and the metabolic syndrome: contribution to global cardiometabolic risk. Arterioscler Thromb Vasc Biol 28: 1039-1049. 34. Despres JP, Lemieux I (2006) Abdominal obesity and metabolic syndrome. Nature 444: 881- 887. 35. Tchernof A, Despres JP (2013) Pathophysiology of human visceral obesity: an update. Physiol Rev 93: 359-404. 36. Stumvoll M, Goldstein BJ, van Haeften TW (2005) Type 2 diabetes: principles of pathogenesis and therapy. Lancet 365: 1333-1346. 37. Gastaldelli A (2011) Role of beta-cell dysfunction, ectopic fat accumulation and insulin resistance in the pathogenesis of type 2 diabetes mellitus. Diabetes Res Clin Pract 93 Suppl 1: S60-65. 38. Bremer AA, Mietus-Snyder M, Lustig RH (2012) Toward a unifying hypothesis of metabolic syndrome. Pediatrics 129: 557-570. 39. Prentki M, Nolan CJ (2006) Islet beta cell failure in type 2 diabetes. J Clin Invest 116: 1802- 1812. 40. Lencioni C, Lupi R, Del Prato S (2008) Beta-cell failure in type 2 diabetes mellitus. Curr Diab Rep 8: 179-184. 41. Wallace TM, Levy JC, Matthews DR (2004) Use and abuse of HOMA modeling. Diabetes Care 27: 1487-1495.

42. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, et al. (1985) Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28: 412-419. 43. Bonora E, Targher G, Alberiche M, Bonadonna RC, Saggiani F, et al. (2000) Homeostasis model assessment closely mirrors the glucose clamp technique in the assessment of insulin sensitivity: studies in subjects with various degrees of glucose tolerance and insulin sensitivity. Diabetes Care 23: 57-63. 44. Emoto M, Nishizawa Y, Maekawa K, Hiura Y, Kanda H, et al. (1999) Homeostasis model assessment as a clinical index of insulin resistance in type 2 diabetic patients treated with sulfonylureas. Diabetes Care 22: 818-822. 45. Brunzell JD, Davidson M, Furberg CD, Goldberg RB, Howard BV, et al. (2008) Lipoprotein management in patients with cardiometabolic risk: consensus statement from the American Diabetes Association and the American College of Cardiology Foundation. Diabetes Care 31: 811-822. 46. Steinberg D, Witztum JL (1990) Lipoproteins and atherogenesis. Current concepts. JAMA 264: 3047-3052. 47. Pepys MB, Hirschfield GM (2003) C-reactive protein: a critical update. J Clin Invest 111: 1805-1812. 48. Calder PC, Ahluwalia N, Albers R, Bosco N, Bourdet-Sicard R, et al. (2013) A consideration of biomarkers to be used for evaluation of inflammation in human nutritional studies. Br J Nutr 109 Suppl 1: S1-34. 49. Anderson TJ, Gregoire J, Hegele RA, Couture P, Mancini GB, et al. (2013) 2012 update of the Canadian Cardiovascular Society guidelines for the diagnosis and treatment of dyslipidemia for the prevention of cardiovascular disease in the adult. Can J Cardiol 29: 151-167. 50. Sarwar N, Danesh J, Eiriksdottir G, Sigurdsson G, Wareham N, et al. (2007) Triglycerides and the risk of coronary heart disease: 10,158 incident cases among 262,525 participants in 29 Western prospective studies. Circulation 115: 450-458. 51. Ninomiya JK, L'Italien G, Criqui MH, Whyte JL, Gamst A, et al. (2004) Association of the metabolic syndrome with history of myocardial infarction and stroke in the Third National Health and Nutrition Examination Survey. Circulation 109: 42-46. 52. Greene DJ, Skeggs JW, Morton RE (2001) Elevated triglyceride content diminishes the capacity of high density lipoprotein to deliver cholesteryl esters via the scavenger receptor class B type I (SR-BI). J Biol Chem 276: 4804-4811. 53. Skeggs JW, Morton RE (2002) LDL and HDL enriched in triglyceride promote abnormal cholesterol transport. J Lipid Res 43: 1264-1274. 54. Chait A, Brazg RL, Tribble DL, Krauss RM (1993) Susceptibility of small, dense, low- density lipoproteins to oxidative modification in subjects with the atherogenic lipoprotein phenotype, pattern B. Am J Med 94: 350-356. 55. Cummings J, Ward TH, Greystoke A, Ranson M, Dive C (2008) Biomarker method validation in anticancer drug development. Br J Pharmacol 153: 646-656. 56. (2001) Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharmacol Ther 69: 89-95. 57. Anand SS, Yusuf S, Vuksan V, Devanesen S, Teo KK, et al. (2000) Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE). Lancet 356: 279-284.

58. Chatterjee A, Harris SB, Leiter LA, Fitchett DH, Teoh H, et al. (2012) Managing cardiometabolic risk in primary care: summary of the 2011 consensus statement. Can Fam Physician 58: 389-393, e196-201. 59. Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, et al. (2004) Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet 364: 937-952. 60. Cho SS, Qi L, Fahey GC, Jr., Klurfeld DM (2013) Consumption of cereal fiber, mixtures of whole grains and bran, and whole grains and risk reduction in type 2 diabetes, obesity, and cardiovascular disease. Am J Clin Nutr 98: 594-619. 61. Ludwig DS, Pereira MA, Kroenke CH, Hilner JE, Van Horn L, et al. (1999) Dietary fiber, weight gain, and cardiovascular disease risk factors in young adults. JAMA 282: 1539- 1546. 62. Hruby A, Meigs JB, O'Donnell CJ, Jacques PF, McKeown NM (2013) Higher magnesium intake reduces risk of impaired glucose and insulin metabolism, and progression from prediabetes to diabetes in middle-aged Americans. Diabetes Care. 63. Del Gobbo LC, Imamura F, Wu JH, de Oliveira Otto MC, Chiuve SE, et al. (2013) Circulating and dietary magnesium and risk of cardiovascular disease: a systematic review and meta-analysis of prospective studies. Am J Clin Nutr 98: 160-173. 64. Pittas AG, Dawson-Hughes B, Li T, Van Dam RM, Willett WC, et al. (2006) Vitamin D and calcium intake in relation to type 2 diabetes in women. Diabetes Care 29: 650-656. 65. Giovannucci E, Liu Y, Hollis BW, Rimm EB (2008) 25-hydroxyvitamin D and risk of myocardial infarction in men: a prospective study. Arch Intern Med 168: 1174-1180. 66. Houston DK, Ding J, Lee JS, Garcia M, Kanaya AM, et al. (2011) Dietary fat and cholesterol and risk of cardiovascular disease in older adults: the Health ABC Study. Nutr Metab Cardiovasc Dis 21: 430-437. 67. Micha R, Mozaffarian D (2010) Saturated fat and cardiometabolic risk factors, coronary heart disease, stroke, and diabetes: a fresh look at the evidence. Lipids 45: 893-905. 68. Siri-Tarino PW, Sun Q, Hu FB, Krauss RM (2010) Saturated fat, carbohydrate, and cardiovascular disease. Am J Clin Nutr 91: 502-509. 69. Estruch R, Ros E, Salas-Salvado J, Covas MI, Corella D, et al. (2013) Primary prevention of cardiovascular disease with a Mediterranean diet. N Engl J Med 368: 1279-1290. 70. Jassim WE (2012) Association of ABO blood group in Iraqis with hypercholesterolaemia, hypertension and diabetes mellitus. East Mediterr Health J 18: 888-891. 71. Sidhu LS, Malhotra P, Singh SP (1988) ABO and Rh blood groups in diabetes mellitus. Anthropol Anz 46: 269-275. 72. Kamil M, Al-Jamal HA, Yusoff NM (2010) Association of ABO blood groups with diabetes mellitus. Libyan J Med 5. 73. Qureshi MA, Bhatti R (2003) Frequency of ABO blood groups among the diabetes mellitus type 2 patients. J Coll Physicians Surg Pak 13: 453-455. 74. Nakao M, Matsuo K, Hosono S, Ogata S, Ito H, et al. (2011) ABO blood group alleles and the risk of pancreatic cancer in a Japanese population. Cancer Sci 102: 1076-1080. 75. Garcia-Bailo B, Brenner DR, Nielsen D, Lee HJ, Domanski D, et al. (2012) Dietary patterns and ethnicity are associated with distinct plasma proteomic groups. Am J Clin Nutr 95: 352-361. 76. Ainsworth BE, Haskell WL, Leon AS, Jacobs DR, Jr., Montoye HJ, et al. (1993) Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc 25: 71-80. 65

77. Cahill L, Corey PN, El-Sohemy A (2009) Vitamin C deficiency in a population of young Canadian adults. Am J Epidemiol 170: 464-471. 78. (2010) Types & Rh System, Canadian Blood Services. Canadian Blood Services - Société canadienne du sang. 79. Haddad EH, Tanzman JS (2003) What do vegetarians in the United States eat? Am J Clin Nutr 78: 626S-632S. 80. Crowe FL, Appleby PN, Travis RC, Key TJ (2013) Risk of hospitalization or death from ischemic heart disease among British vegetarians and nonvegetarians: results from the EPIC-Oxford cohort study. Am J Clin Nutr 97: 597-603. 81. Appleby PN, Allen NE, Key TJ (2011) Diet, vegetarianism, and cataract risk. Am J Clin Nutr 93: 1128-1135. 82. Tantamango-Bartley Y, Jaceldo-Siegl K, Fan J, Fraser G (2013) Vegetarian diets and the incidence of cancer in a low-risk population. Cancer Epidemiol Biomarkers Prev 22: 286- 294. 83. Key TJ, Appleby PN, Spencer EA, Travis RC, Allen NE, et al. (2009) Cancer incidence in British vegetarians. Br J Cancer 101: 192-197. 84. Tonstad S, Stewart K, Oda K, Batech M, Herring RP, et al. (2013) Vegetarian diets and incidence of diabetes in the Adventist Health Study-2. Nutr Metab Cardiovasc Dis 23: 292-299. 85. Tonstad S, Butler T, Yan R, Fraser GE (2009) Type of vegetarian diet, body weight, and prevalence of type 2 diabetes. Diabetes Care 32: 791-796. 86. Pettersen BJ, Anousheh R, Fan J, Jaceldo-Siegl K, Fraser GE (2012) Vegetarian diets and blood pressure among white subjects: results from the Adventist Health Study-2 (AHS-2). Public Health Nutr 15: 1909-1916. 87. Crowe FL, Appleby PN, Allen NE, Key TJ (2011) Diet and risk of diverticular disease in Oxford cohort of European Prospective Investigation into Cancer and Nutrition (EPIC): prospective study of British vegetarians and non-vegetarians. BMJ 343: d4131. 88. Orlich MJ, Singh PN, Sabate J, Jaceldo-Siegl K, Fan J, et al. (2013) Vegetarian dietary patterns and mortality in Adventist Health Study 2. JAMA Intern Med 173: 1230-1238. 89. Kahleova H, Matoulek M, Malinska H, Oliyarnik O, Kazdova L, et al. (2011) Vegetarian diet improves insulin resistance and oxidative stress markers more than conventional diet in subjects with Type 2 diabetes. Diabet Med 28: 549-559. 90. Sofi F, Abbate R, Gensini GF, Casini A (2010) Accruing evidence on benefits of adherence to the Mediterranean diet on health: an updated systematic review and meta-analysis. Am J Clin Nutr 92: 1189-1196. 91. Mikkelsen TB, Osterdal ML, Knudsen VK, Haugen M, Meltzer HM, et al. (2008) Association between a Mediterranean-type diet and risk of preterm birth among Danish women: a prospective cohort study. Acta Obstet Gynecol Scand 87: 325-330. 92. Tobias DK, Zhang C, Chavarro J, Bowers K, Rich-Edwards J, et al. (2012) Prepregnancy adherence to dietary patterns and lower risk of gestational diabetes mellitus. Am J Clin Nutr 96: 289-295. 93. Scarmeas N, Stern Y, Tang MX, Mayeux R, Luchsinger JA (2006) Mediterranean diet and risk for Alzheimer's disease. Ann Neurol 59: 912-921. 94. Gao X, Chen H, Fung TT, Logroscino G, Schwarzschild MA, et al. (2007) Prospective study of dietary pattern and risk of Parkinson disease. Am J Clin Nutr 86: 1486-1494. 95. Bosetti C, Turati F, Dal Pont A, Ferraroni M, Polesel J, et al. (2013) The role of Mediterranean diet on the risk of pancreatic cancer. Br J Cancer 109: 1360-1366. 66

96. Bamia C, Lagiou P, Buckland G, Grioni S, Agnoli C, et al. (2013) Mediterranean diet and colorectal cancer risk: results from a European cohort. Eur J Epidemiol 28: 317-328. 97. Buckland G, Travier N, Cottet V, Gonzalez CA, Lujan-Barroso L, et al. (2013) Adherence to the mediterranean diet and risk of breast cancer in the European prospective investigation into cancer and nutrition cohort study. Int J Cancer 132: 2918-2927. 98. Jessri M, Rashidkhani B, Hajizadeh B, Jacques PF (2012) Adherence to Mediterranean-style dietary pattern and risk of esophageal squamous cell carcinoma: a case-control study in Iran. J Am Coll Nutr 31: 338-351. 99. Rossi M, Turati F, Lagiou P, Trichopoulos D, Augustin LS, et al. (2013) Mediterranean diet and glycaemic load in relation to incidence of type 2 diabetes: results from the Greek cohort of the population-based European Prospective Investigation into Cancer and Nutrition (EPIC). Diabetologia 56: 2405-2413. 100. Westman EC, Yancy WS, Jr., Mavropoulos JC, Marquart M, McDuffie JR (2008) The effect of a low-carbohydrate, ketogenic diet versus a low-glycemic index diet on glycemic control in type 2 diabetes mellitus. Nutr Metab (Lond) 5: 36. 101. Wheeler ML, Dunbar SA, Jaacks LM, Karmally W, Mayer-Davis EJ, et al. (2012) Macronutrients, food groups, and eating patterns in the management of diabetes: a systematic review of the literature, 2010. Diabetes Care 35: 434-445. 102. Ajala O, English P, Pinkney J (2013) Systematic review and meta-analysis of different dietary approaches to the management of type 2 diabetes. Am J Clin Nutr 97: 505-516. 103. Weigle DS, Breen PA, Matthys CC, Callahan HS, Meeuws KE, et al. (2005) A high-protein diet induces sustained reductions in appetite, ad libitum caloric intake, and body weight despite compensatory changes in diurnal plasma leptin and ghrelin concentrations. Am J Clin Nutr 82: 41-48. 104. Gannon MC, Nuttall FQ, Saeed A, Jordan K, Hoover H (2003) An increase in dietary protein improves the blood glucose response in persons with type 2 diabetes. Am J Clin Nutr 78: 734-741. 105. Flatt JP (1995) Use and storage of carbohydrate and fat. Am J Clin Nutr 61: 952S-959S. 106. Flatt JP, Ravussin E, Acheson KJ, Jequier E (1985) Effects of dietary fat on postprandial substrate oxidation and on carbohydrate and fat balances. J Clin Invest 76: 1019-1024. 107. Westerterp KR, Smeets A, Lejeune MP, Wouters-Adriaens MP, Westerterp-Plantenga MS (2008) Dietary fat oxidation as a function of body fat. Am J Clin Nutr 87: 132-135. 108. Wajchenberg BL (2000) Subcutaneous and visceral adipose tissue: their relation to the metabolic syndrome. Endocr Rev 21: 697-738. 109. Craig WJ (2009) Health effects of vegan diets. Am J Clin Nutr 89: 1627S-1633S. 110. Fung TT, Stampfer MJ, Manson JE, Rexrode KM, Willett WC, et al. (2004) Prospective study of major dietary patterns and stroke risk in women. Stroke 35: 2014-2019. 111. Fung TT, Willett WC, Stampfer MJ, Manson JE, Hu FB (2001) Dietary patterns and the risk of coronary heart disease in women. Arch Intern Med 161: 1857-1862. 112. Steffen LM, Kroenke CH, Yu X, Pereira MA, Slattery ML, et al. (2005) Associations of plant food, dairy product, and meat intakes with 15-y incidence of elevated blood pressure in young black and white adults: the Coronary Artery Risk Development in Young Adults (CARDIA) Study. Am J Clin Nutr 82: 1169-1177; quiz 1363-1164. 113. van Dam RM, Grievink L, Ocke MC, Feskens EJ (2003) Patterns of food consumption and risk factors for cardiovascular disease in the general Dutch population. Am J Clin Nutr 77: 1156-1163.

114. Keast DR, Fulgoni VL, 3rd, Nicklas TA, O'Neil CE (2013) Food sources of energy and nutrients among children in the United States: National Health and Nutrition Examination Survey 2003-2006. Nutrients 5: 283-301. 115. Patterson E, Larsson SC, Wolk A, Akesson A (2013) Association between dairy food consumption and risk of myocardial infarction in women differs by type of dairy food. J Nutr 143: 74-79. 116. von Ruesten A, Feller S, Bergmann MM, Boeing H (2013) Diet and risk of chronic diseases: results from the first 8 years of follow-up in the EPIC-Potsdam study. Eur J Clin Nutr 67: 412-419. 117. Willett WC (2012) Dietary fats and coronary heart disease. J Intern Med 272: 13-24. 118. Nordmann AJ, Nordmann A, Briel M, Keller U, Yancy WS, Jr., et al. (2006) Effects of low- carbohydrate vs low-fat diets on weight loss and cardiovascular risk factors: a meta- analysis of randomized controlled trials. Arch Intern Med 166: 285-293. 119. Hu T, Mills KT, Yao L, Demanelis K, Eloustaz M, et al. (2012) Effects of low-carbohydrate diets versus low-fat diets on metabolic risk factors: a meta-analysis of randomized controlled clinical trials. Am J Epidemiol 176 Suppl 7: S44-54. 120. Santos FL, Esteves SS, da Costa Pereira A, Yancy WS, Jr., Nunes JP (2012) Systematic review and meta-analysis of clinical trials of the effects of low carbohydrate diets on cardiovascular risk factors. Obes Rev 13: 1048-1066. 121. Halton TL, Willett WC, Liu S, Manson JE, Albert CM, et al. (2006) Low-carbohydrate-diet score and the risk of coronary heart disease in women. N Engl J Med 355: 1991-2002. 122. Samaha FF, Iqbal N, Seshadri P, Chicano KL, Daily DA, et al. (2003) A low-carbohydrate as compared with a low-fat diet in severe obesity. N Engl J Med 348: 2074-2081. 123. Stern L, Iqbal N, Seshadri P, Chicano KL, Daily DA, et al. (2004) The effects of low- carbohydrate versus conventional weight loss diets in severely obese adults: one-year follow-up of a randomized trial. Ann Intern Med 140: 778-785. 124. Yancy WS, Jr., Olsen MK, Guyton JR, Bakst RP, Westman EC (2004) A low-carbohydrate, ketogenic diet versus a low-fat diet to treat obesity and hyperlipidemia: a randomized, controlled trial. Ann Intern Med 140: 769-777. 125. Sacks FM, Tonkin AM, Craven T, Pfeffer MA, Shepherd J, et al. (2002) Coronary heart disease in patients with low LDL-cholesterol: benefit of pravastatin in diabetics and enhanced role for HDL-cholesterol and triglycerides as risk factors. Circulation 105: 1424-1428. 126. Rubins HB, Robins SJ, Collins D, Fye CL, Anderson JW, et al. (1999) Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of high-density lipoprotein cholesterol. Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group. N Engl J Med 341: 410-418. 127. Ginsberg HN (1997) Is hypertriglyceridemia a risk factor for atherosclerotic cardiovascular disease? A simple question with a complicated answer. Ann Intern Med 126: 912-914. 128. Saitou N, Yamamoto F (1997) Evolution of primate ABO blood group genes and their homologous genes. Mol Biol Evol 14: 399-411. 129. Calafell F, Roubinet F, Ramirez-Soriano A, Saitou N, Bertranpetit J, et al. (2008) Evolutionary dynamics of the human ABO gene. Hum Genet 124: 123-135. 130. Nachbar MS, Oppenheim JD (1980) Lectins in the United States diet: a survey of lectins in commonly consumed foods and a review of the literature. Am J Clin Nutr 33: 2338-2345. 131. Cusack L, De Buck E, Compernolle V, Vandekerckhove P (2013) Blood type diets lack supporting evidence: a systematic review. Am J Clin Nutr 98: 99-104. 68

132. Liese AD, Bortsov A, Gunther AL, Dabelea D, Reynolds K, et al. (2011) Association of DASH diet with cardiovascular risk factors in youth with diabetes mellitus: the SEARCH for Diabetes in Youth study. Circulation 123: 1410-1417.

Appendix

Table: Food items included in the "blood-type" diet score calculation Blood Type Food List A B AB O Skim milk - + / - Milk 1-2% - + / - Whole milk - + - - Soy milk + - / / Flavoured yogurt / + + - Plain yogurt / + + - Cottage cheese - + + - Cream cheese - / / - Other cheese / + + / Butter - / - / Grapes / + + / Prunes + / / + Bananas - + - + Cantaloupe / / / - Honeydew - / / - Watermelon / + + / Avocado / - - - Apple / / / / Apple juice / / / / Orange - / - - Orange juice - / - - Grapefruit + / + / Grapefruit juice + / / / Other fruit juices / + + / Strawberries / / / / Blueberries + / / + Peaches / / / / Tropical fruits - + + / Dried apricot + / / / Raisins / / / / Dried dates / / / / Cranberries + + + /

Tomatoes - - / / Tomato juice - - / / Tomato sauce - - / / Tofu + - + / String beans + / / / Broccoli + + + + Cabbage - + / / Cauliflower / + + - Brussels Sprouts / + / / Raw carrots + + / / Cooked carrots + + / / Corn / - - - Peas / / / / Mixed vegetables + + + + Beans + - / - Squash / / / / Zucchini / / / / Eggplant - + + / Yams - + + / Cooked spinach + / / + Raw spinach + / / + Kale + + + + Iceberg / / / / Romaine lettuce + / / + Celery + / + / Green peppers - + - / Red peppers - + - + Onions (Garnish) + / / + Onions + / / + Okra + / / + Cucumber / / + - Olives - - - - Beets / + + / Asparagus / / / / Egg whites / / / / Eggs / / / / Bacon - - - - Chicken sandwich / - - /

Chicken/Turkey with skin / - - / Chicken/Turkey without / - - / skin Hotdogs - / - + Chicken/Turkey hotdogs / - - / Processed meat sandwich - / - + Processed meat - / - + Lean hamburger - / - + Hamburger - / - + Sandwich - / - + Pork - - - - Beef - / - + Liver - / / + Chicken liver / - - / Other organ meats - / - + Canned tuna / / + / Fish cakes + + + + Shrimp - - - / Dark meat fish + + + / Other fish + + + + Cold cereal / - / - Cooked oatmeal + + + / Cooked oat bran + + + / Cooked breakfast cereal / - - - White bread / / / - Muffins - / / - Heavy bread + - + / Dark bread / - / / Bagel / / / - Biscuits / / / - Brown rice / / + / White rice / / + / Mac & cheese / / / - Pasta / / / - Whole grain pasta - - / - Barley / / / - Bulgur / - / - Other grains / / / -

Pancakes / / / - French fries - / / - Potatoes (scalloped) - / / - Potatoes (baked) - / / - Potato chips - / / - Corn chips / - - - Crackers / / / - Pizza / / / - Tortillas / / / - Skim milk - + / - Milk 1-2% - + / -

Notes: The food list was retrieved from the FFQ database of Toronto Nutrigenomics and Health Study. “+” signs indicate the foods that are recommended for the blood type. “-“ signs indicate the food to avoid for the blood type. “/” signs indicate the food that are neutral.