Effects of Dietary Pulses on Factors of · Effects of Dietary Pulses on ... DASH‐ Dietary Approaches to Stopping Hypertension ... of low‐glycemic index8, Dietary Strategies to

Effects of Dietary Pulses on

Lipid Risk Factors of Cardiovascular Disease and Oxidative Stress

by

Vanessa Ha

A thesis submitted in conformity with the requirements

for the degree of Master of Science

Graduate Department of Nutritional Sciences

University of Toronto

© Copyright by Vanessa Ha 2013

ii

Effects of Dietary Pulses on Lipid Risk Factors of CVD and Oxidative Stress

Vanessa Ha

Master of Science

Department of Nutritional Sciences

University of Toronto

2013

ABSTRACT

The objective was to conduct a systematic review and meta‐analysis of randomized feeding

trials to assess the effect of dietary pulses (beans, chickpeas, lentils, peas) on established lipid targets of

cardiovascular disease (CVD) and perform a secondary analysis of our randomized feeding trial to assess

whether dietary pulses as a means of lowering the glycemic index offer further CVD protection by

reducing oxidative stress. The meta‐analysis of 26 trials (n=1013) found dietary pulse interventions

significantly lowered LDL‐C compared with isocaloric control interventions (mean difference=‐

0.17mmol/L [95% CI: ‐0.25, ‐0.09]; p<0.0001). No treatment effects were observed for Apo‐B and non‐

HDL‐C. Our feeding trial found no significant differences between the high‐dietary pulse diet and high‐

fibre control diet on markers of oxidative stress, including thiobarbituric acid reactive substances

(TBARS), conjugated dienes (CDs), and protein thiols. Overall, the results suggest dietary pulses reduce

LDL‐C but not oxidative stress as a means of reducing cardiovascular risk.

Abstract Word Count: 150

iii

TABLE OF CONTENTS Title Page Number ABSTRACT ..................................................................................................................................................... ii

TABLE OF CONTENTS ................................................................................................................................... iii

LIST OF ABBREVIATIONS ............................................................................................................................ vii

LIST OF FIGURES ........................................................................................................................................... x

LIST OF TABLES ............................................................................................................................................ xi

CHAPTER I‐ INTRODUCTION ................................................................................................................. 1

CHAPTER II‐ LITERATURE REVIEW ........................................................................................................ 4

2.1 CARDIOVASCULAR DISEASE .................................................................................................................... 5

2.1.1 PREVALENCE ............................................................................................................................ 5

2.1.2 COMPLICATIONS AND MORTALITY ......................................................................................... 5

2.1.3 RISK FACTORS .......................................................................................................................... 6

2.1.3.1 CORONARY HEART DISEASE .................................................................................. 6

2.1.3.2 STROKE .................................................................................................................. 6

2.2 LIPIDS AND CARDIOVASCULAR DISEASE ................................................................................................. 6

2.2.1 GUIDELINES ............................................................................................................................. 7

2.2.1.1 NATIONAL CHOLESTEROL EDUCATION PROGRAM‐ ADULT TREATMENT PLAN III8

2.2.1.2 CANADIAN CARDIOVASCULAR SOCIETY .............................................................. 8

2.2.1.3 CANADIAN DIABETES ASSOCIATION .................................................................... 8

2.2.1.4 AMERICAN DIABETES ASSOCIATION .................................................................... 9

2.3 OXIDATIVE STRESS AND CARDIOVASCULAR DISEASE ............................................................................. 9

2.4 DIET AND CARDIOVASCULAR DISEASE .................................................................................................. 10

2.4.1 DIET AND DRUGS ................................................................................................................... 10

2.4.2 DIETARY PATTERNS AND CARDIOVASCULAR DISEASE .......................................................... 11

2.4.2.1 PROSPECTIVE COHORT STUDIES ......................................................................... 11

2.4.2.2 DIETARY INTERVENTION TRIALS ......................................................................... 12

2.5 PLANT‐PROTEIN BASED DIET AND CARDIOVASCULAR DISEASE ........................................................... 13

2.5.1 PROSPECTIVE COHORT STUDIES ........................................................................................... 14

2.5.2 DIETARY INTERVENTION TRIALS ............................................................................................ 15

2.6 DIETARY PULSES AND CARDIOVASCULAR DISEASE ............................................................................... 15

2.6.1 DIETARY PULSES IN GUIDELINES ........................................................................................... 16

iv

2.6.2 DIETARY PULSES AND CARDIOVASCULAR DISEASE ............................................................... 17

2.6.3 DIETARY PULSES IN DIETARY PATTERNS ............................................................................... 17

2.6.4 DIETARY PULSES AND ESTABLISHED LIPID TARGETS OF CARDIOVASCULAR DISEASE .......... 18

2.6.4.1 EVIDENCE FROM META‐ANALYSES ..................................................................... 18

2.6.4.2 EVIDENCE FROM RANDOMIZED FEEDING TRIALS IN HUMANS .......................... 18

2.6.5 DIETARY PULSES AND OXIDATIVE STRESS ............................................................................. 19

2.6.6 DIETARY PULSES AND OTHER CARIOVASCULAR RISK FACTORS ............................................ 20

CHAPTER III‐ HYPOTHESIS, OBJECTIVES, AND RATIONALE ................................................................... 21

3.1 HYPOTHESIS ............................................................................................................................. 22

3.2 OBJECTIVES .............................................................................................................................. 22

3.3 RATIONALE ............................................................................................................................... 22

CHAPTER IV‐ THE EFFECT OF DIETARY PULSES ON ESTABLISHED LIPID THERAPEUTIC TARGETS OF

CARDIOVASCULAR DISEASE: A SYSTEMATIC REVIEW AND META‐ANALYSIS OF RANDOMIZED

CONTROLELD TRIALS .......................................................................................................................... 24

4.1 CITATIONS ............................................................................................................................................. 25

4.2 ABSTRACT .............................................................................................................................................. 27

4.3 INTRODUCTION ..................................................................................................................................... 28

4.4 METHODS .............................................................................................................................................. 29

4.4.1 STUDY STRATEGY SELECTION ................................................................................................ 29

4.4.2 ELIGIBILITY CRITERIA ............................................................................................................. 29

4.4.3 DATA EXTRACTIONS .............................................................................................................. 30

4.4.4 TREATMENT OF MISSING DATA ............................................................................................ 31

4.3.5 STATISTICAL ANALYSIS .......................................................................................................... 31

4.5 RESULTS ................................................................................................................................................. 32

4.5.1 SEARCH RESULTS ................................................................................................................... 32

4.5.2 STUDY CHARACTERISTICS ...................................................................................................... 32

4.5.3 GASTRO‐INTESTINAL SYMPTOMS ......................................................................................... 34

4.5.4 EFFECT ON LDL‐C ................................................................................................................... 34

4.5.5 EFFECT ON NON‐HDL‐C ......................................................................................................... 35

4.5.6 EFFECT ON APOLIPROTEINS B ............................................................................................... 35

4.5.7 PUBLICATION BIAS ................................................................................................................ 36

4.6 DISCUSSION ........................................................................................................................................... 36

v

4.6.1 RESULTS IN RELATION TO OTHER STUDIES ........................................................................... 36

4.6.2 INTER‐STUDY HETEROGENEITY ............................................................................................. 38

4.6.3 LIMITATIONS ......................................................................................................................... 39

4.6.4 CONCLUSIONS ....................................................................................................................... 40

4.7 ACKNOWLEDGMENTS ........................................................................................................................... 41

4.8 SOURCE OF FUNDING ........................................................................................................................... 41

4.9 CONFLICTS OF INTEREST/DISCLOSURE ................................................................................................. 41

4.10 TABLES ................................................................................................................................................. 44

4.11 FIGURES ............................................................................................................................................... 50

CHAPTER V‐ EFFECT OF A HIGH DIETARY PULSES, LOW GLYCEMIC INDEX DIET ON OXIDATIVE STRESS

MARKERS IN TYPE 2 DIABETES ........................................................................................................... 57

5.1 ABSTRACT .............................................................................................................................................. 59

5.2 INTRODUCTION ..................................................................................................................................... 60

5.3 METHODS .............................................................................................................................................. 60

5.3.1 PARTICIPANTS ....................................................................................................................... 61

5.3.2 PROTOCOL ............................................................................................................................. 61

5.3.3 DIETARY INTERVENTION ....................................................................................................... 62

5.3.4 BIOCHEMICAL ANALYSIS ....................................................................................................... 62

5.3.5 STATISTICAL ANALYSIS .......................................................................................................... 62

5.4 RESULTS ................................................................................................................................................. 63

5.4.1 OXIDATIVE STRESS MARKERS ................................................................................................ 64

5.4.2 GLYCEMIC EXCURSION .......................................................................................................... 64

5.5 DISCUSSION ........................................................................................................................................... 64

5.5.1 RELATIONS TO OTHER STUDIES............................................................................................. 65

5.5.2 LIMITATIONS ......................................................................................................................... 66

5.6 CONCLUSIONS ....................................................................................................................................... 67

5.7 SOURCE OF FUNDING ........................................................................................................................... 67

5.8 CONFLICTS OF INTEREST/DISCLOSURE ................................................................................................. 68

5.9 TABLES ................................................................................................................................................... 70

5.10 FIGURES ............................................................................................................................................... 75

CHAPTER VI‐ OVERALL DISCUSSION AND LIMITATIONS ...................................................................... 77

6.1 OVERALL DISCUSSION ........................................................................................................................... 78

vi

6.2 LIMITATIONS ......................................................................................................................................... 79

6.2.1 SYSTEMATIC REVIEW AND META‐ANALYSIS ......................................................................... 79

6.2.2 DIETARY INTERVENTION TRIAL ............................................................................................. 79

6.3 MECHANISM OF ACTION ....................................................................................................................... 80

6.4 CLINICAL IMPLICATIONS ........................................................................................................................ 81

6.5 FUTURE DIRECTIONS ............................................................................................................................. 81

CHAPTER VII‐ CONCLUSIONS .............................................................................................................. 83

CHAPTER VIII‐ REFERENCES ................................................................................................................ 85

vii

LIST OF ABBREVIATIONS

CVD‐ Cardiovascular Disease

LDL‐C‐ low‐density lipoprotein cholesterol

CHD‐ Coronary Heart Disease

NHS‐ Nurses’ Health Study

AHS‐ Adult Health Study

USDA‐ United States Department of Agriculture

DASH‐ Dietary Approaches to Stopping Hypertension

ox‐LDL‐C‐ oxidized‐LDL‐C

Apo‐B‐ apolipoprotein‐B

Non‐HDL‐C‐ non‐high density lipoprotein cholesterol

NHANES‐ National Health and Nutrition Examination Survey

BP‐ blood pressure

TC‐ total cholesterol

PAR‐ population attributable risk

FHS‐ Framingham Heart Study

CTT‐ Cholesterol Treatment Trialists

NCEP‐ATP III‐ National Cholesterol Education Program‐ Adult Treatment Plan III

TG‐ Triglyceride

T2DM‐ type 2 diabetes mellitus

CCS‐ Canadian Cardiovascular Society

CDA‐ Canadian Diabetes Association

ADA‐ American Diabetes Association

FRS‐ Framingham Risk Score

NHS‐ Nurses’ Health Study

viii

HPFS‐ Health Professional Follow‐Up Study

HDL‐C‐ high‐density lipoprotein cholesterol

FAO/WHO‐ Food and Agriculture Organization of the United Nations/World Health Organization

AHA‐ American Heart Association

EASD‐ European Association for the Study of Diabetes

FDA‐ Food and Drug Association

EFSA‐ European Food Safety Authority

SCFA‐ short chain fatty acids

n‐ sample size

MDA‐ malendialdehyde

TBARS‐ thiobarbituric acid reactive substances

CD‐ conjugated dienes

CI‐ confidence intervals

PRISMA‐ Preferred Reporting Items for Systematic Review and Meta‐Analyses

MQS‐ Methodological Quality Score

MD‐ mean difference

% E‐ percent energy

g/d‐ grams per day

SD‐ standard deviation

Ob‐ Obese

OW‐ Overweight

HC‐ Hypercholesterolemia

N‐ normal/healthy

Pre‐MS‐ Pre‐Metabolic Syndrome

IR‐ Insulin Resistant

ix

IS‐ Insulin Sensitive

M‐ Men

W‐ Women

C‐ Crossover Design

P‐ Parallel Design

IP‐ inpatient

OP‐ outpatient

UR‐ unclear risk of bias

LR‐ low risk of bias

HR‐ high risk of bias

AHS‐ Adult Health Study

USDA‐ United States Department of Agriculture

GI‐ glycemic index

Haemoglobin A1c‐ HbA1c

CV‐ coefficient of variation

FBG‐ fasting blood glucose

ORAC‐ Oxygen Radical Absorbance Capacity

SFA‐ saturated fatty acid

MUFA‐ monounsaturated fatty acid

PUFA‐ polyunsaturated fatty acid

PREDIMED‐ Prevención con Dieta Mediterrànea

x

LIST OF FIGURES

CHAPTER II

Figure 1. Schematic Diagram of the Formation of Atheromatous Plaques via ox‐LDL‐C.

CHAPTER IV

Figure 1. Flow of the literature search.

Figure 2. Forest plot of trials investigating the effect of isocaloric exchange of dietary pulses for control diets on LDL‐C. Figure 3. Forest plots of subgroup analyses for dichotomous variables investigating the effect of isocaloric exchange of dietary pulses for other adequate comparators on LDL‐C in all participants. Figure 4. Forest plot of trials investigating the effect of isocaloric exchange of dietary pulses for control diets on non‐HDL‐C. Figure 5. Forest plots of subgroup analyses for dichotomous variables investigating the effect of isocaloric exchange of dietary pulses for other adequate comparators on non‐HDL‐C in all participants. Figure 6. Forest plot of a trial investigating the effect of isocaloric exchange of dietary pulses for control diet on Apo‐B. Figure 7. Funnel plot investigating publication bias and effect of small study effects in clinical trials with isocaloric exchange of dietary pulses for other adequate comparators on LDL‐C (7A) and non‐HDL‐C (7B) in all participants. CHAPTER V

Figure 1. Flowchart of Participants through the Trial.

Figure 2. Correlation of Change in Pulse intake (∆Pulse Intake) and Change in Conjugated Dienes (∆CDs).

xi

LIST OF TABLES

CHAPTER II

Table 1. Summary of Lipid Targets by Major Chronic Disease Guidelines in Adults.

Table 2. Summary of Lipids Targets set by NCEP‐ATP III.

Table 3. Summary of Lipid Targets set by CCS (2012).

Table 4. Summary of Lipid Targets set by CDA (2013).

Table 5. Summary of Lipid Targets set by ADA (2013).

CHAPTER IV

Table 1. Search Strategy for Studies Assessing the Effect of Dietary Pulses on Lipids in Randomized, Controlled Feeding Trials.

Table 2. Study Characteristics Table.

Table 3. Study Quality Assessment by Using the Heyland MQS. Table 4. Risk of Bias Assessment by Using the Cochrane Risk of Bias Tool.

Table 5. Subgroup analyses by a priori dichotomous categorization for LDL‐C and non‐HDL‐C.

CHAPTER V

Table 1. Definitions of quadrant groups for glucose excursion analysis.

Table 2. Baseline Characteristics of Participants.

Table 3. Daily Nutritional Profile at Baseline and End of Study. Table 4. Mean study measurements and significance of treatment differences on oxidative stress markers. Table 5. Associations between changes in oxidative stress markers and changes in study outcomes.

INTRODUCTION

Page 1 of 93

CHAPTER I‐ INTRODUCTION

INTRODUCTION

Page 2 of 93

1. INTRODUCTION

Abnormal lipids have been identified as one of the most modifiable risk factors for cardiovascular

disease (CVD), contributing to almost 50% of the population’s risk1. While statins are effective in

reducing blood LDL‐C2, major health organizations maintained that the initial and essential approach to

CVD prevention and management is to modify dietary and lifestyle patterns3‐6. Studies have shown that

82% of coronary heart disease (CHD) may be prevented with healthier dietary and lifestyle choices7.

Dietary pulses (beans, chickpeas, peas, and lentils) have received particular attention for their

ability to reduce CVD risk as part of low‐glycemic index8, Dietary Strategies to Stop Hypertension

(DASH)9, high‐fibre10, Mediterranean11, and vegetarian12 dietary patterns. They are high in vegetable

protein, fibre and have a low glycemic index13, 14, all of which are associated with improved CHD

outcomes15‐17. Dietary pulses are also high in 7S globulin, a protein which has been shown to increase

uptake and degradation of LDL‐C in human hepatic cells18. Small controlled feeding trials have

demonstrated that diets which emphasize dietary pulses improved lipid levels19‐30. Taken together these

data suggest that high dietary pulse intake may improve serum lipids.

Dietary pulses may offer further cardiovascular protection through reducing oxidized‐LDL‐C (ox‐

LDL‐C). Dietary pulses are a good source of vegetable protein which has been shown to reduce the

oxidation of LDL‐C31, 32. Reducing levels of ox‐LDL‐C is hypothesized to reduce the atherogenicity of LDL‐C

since ox‐LDL‐C is taken up more rapidly by the scavenger receptors of macrophages in the

subendothelial arterial wall33‐35. Recent prospective cohort studies have reported the benefits of higher

plant protein intakes on CVD risk36, 37 and dietary patterns which may include dietary pulses such as low

glycemic index8, DASH9, and Mediterranean11 have also been shown to reduce markers of oxidative

stress. However except for one, few studies have directly assessed a high dietary pulse intake on

oxidative damage32.

INTRODUCTION

Page 3 of 93

Despite these emerging benefits of dietary pulses, most major chronic disease prevention

guidelines have not made specific recommendations for dietary pulse intake to reduce cardiovascular

risk. Rather guidelines have encouraged the consumption of dietary pulses as part of a low‐glycemic

index8, DASH9, high‐fibre10, Mediterranean11, or vegetarian12 diet. There is a need for higher quality

evidence to support the lipid‐lowering and oxidative stress mitigating effects of dietary pulses for

guidelines development. To provide higher quality evidence on the effects of dietary pulses on lipids to

reduce cardiovascular risk, the following thesis work will present: 1) the results of a systematic review

and meta‐analysis of randomized feeding trials of the effects of dietary pulse interventions on

established lipid targets of CVD risk (LDL‐C, Apo‐B, and non‐HDL‐C) and 2) the results of a secondary

endpoint analysis of a randomized feeding trial of the effects of a high dietary pulse diet as a means to

lower the glycemic index compared to a high‐fibre control diet on markers of oxidative stress.

LITERATURE REVIEW

Page 4 of 93

CHAPTER II‐ LITERATURE REVIEW

LITERATURE REVIEW

Page 5 of 93

2.1 CARDIOVASCULAR DISEASE 2.1.1 PREVALENCE

Cardiovascular disease (CVD) is a major public health concern. CVD is a class of diseases relating

to the heart and vascular system including coronary heart disease (CHD), stroke, peripheral vascular

disease, heart failure, rheumatic and congenital heart disease. According to the National Health and

Nutrition Examination Survey (NHANES: 2007–2010), 83.6 million Americans (35.3%) were suffering

from CVD in 201038and 1.3 million Canadians were diagnosed in 200739. Specifically the 2 most prevalent

types of CVD, CHD and stroke, were diagnosed in 15.4 million (6.4%) and 6.8 million (2.4%) individuals,

respectively38. Although the rate of CHD has declined over the decades (NHANES: 1971‐2006), the

prevalence of CHD is still high38. Similar trends have also been reported for Canada39. Despite recent

advances in the understanding of CVD prevention, the prevalence of CVD still remains high.

2.1.2 COMPLICATIONS AND MORTALITY

CVD has serious long‐term consequences. The pathogenesis of CVD is usually characterized by

the development and progression of atheromatous plaques in the arterial wall40. Although

atherogenesis begins as early as adolescence, CVD is considered a preventable disease7. Individuals with

CHD have lower life expectancy because they are at high risk for myocardial infarctions (i.e. heart

attacks), stroke, and heart failure, all of which increases the risk of long‐term disability and mortality. In

2009, the overall death rate attributable to CVD was 236.1 per 100, 000 or 1 in 3 deaths, and specifically

CHD and stroke accounted for 1 in 6 deaths and 1 in 19 deaths, respectively38. Most individuals survive

their first heart attacks38, however, re‐occurring heart attacks can weaken the heart muscle and

decrease the quality of life. It has been estimated that approximately every 34 seconds, 1 American has

a coronary event, and approximately every minute, an American will die of one38. Taken altogether, the

data suggest that the burden of CVD remains high.

LITERATURE REVIEW

Page 6 of 93

2.1.3 RISK FACTORS

2.1.3.1 Coronary Heart Disease

Risk factors of CHD are characterized by the 10‐year CHD Framingham Risk Score (FRS)4. The Risk

Score predicts the risk of CHD based on non‐modifiable risk factors including sex and age, and

modifiable risk factors including diabetes, smoking, blood pressure (BP), total cholesterol (TC), and LDL‐C

levels4. It has been estimated that 80 to 90% of patients with CHD have at least 1 of the 4 modifiable risk

factors of CHD and unhealthy dietary and lifestyle habits have been shown to account for more than

80% of coronary heart events7. The INTERHEART Study, a case‐control study of 15 152 acute myocardial

infarction cases and 14 820 controls from 52 countries, reported that diabetes contributed 9.9% of the

population attributable risk (PAR), smoking at 35.7% (PAR), hypertension at 17.9% PAR, and abnormal

lipids at 49.2% PAR1. In addition to these major risk factors, obesity, poor dietary patterns, physical

inactivity, family history of CHD, and small LDL‐C particles have been associated with an increase in CHD

risk3, 4.

2.1.3.2 Stroke

Risk factors of stroke are characterized by the 10‐year Stroke FRS41. Similar to the CHD risk

score, risk factors for stroke include non‐modifiable risk factors including sex and age, and modifiable

risk factors including systolic BP, diabetes, smoking, CVD and atrial fibrillation status, left ventricular

hypertrophy, and hypertension. Consequences of stroke have serious debilitating effects and it is one of

the leading causes of long‐term disability. Similar to those of CHD, other risk factors associated with

stroke include obesity, poor dietary patterns, physical inactivity, and family history of stroke41.

2.2 LIPIDS AND CARDIOVASCULAR DISEASE

Abnormal lipids have been recognized as one of the most preventable risk factors for CVD. The

pathogenesis of CVD is usually characterized by the development and progression of atheromatous

plaques in the arterial wall40. The Framingham Heart Study (FHS) was the first to correlate high lipid

LITERATURE REVIEW

Page 7 of 93

levels with CHD incidence42, and the INTERHEART Study has identified that almost 50% of acute

myocardial infarctions can be prevented by modifying abnormal lipid levels1. Further the success of

statins, a class of drugs used to lower cholesterol levels by inhibiting the hepatic enzyme HMG‐CoA

reductase, have shown to lower CHD and stroke risk and 43, 44 the Cholesterol Treatment Trialists’ (CTT)

Collaboration have also reported that every 1 mmol/L reduction of LDL‐C can reduce CHD risk by 20%45.

As a result of the overwhelming data supporting the role of abnormal lipids in heart health, major

chronic disease guidelines including the National Cholesterol Education Program‐ Adult Treatment Plan

III (NCEP‐ATP III), Canadian Cardiovascular Society (CCS), Canadian Diabetes Association (CDA), and

American Diabetes Association (ADA) have recognized the importance of lipids in heat health and have

set lipid targets to lower the risk of CVD3, 4, 6, 46.

2.2.1 GUIDELINES

Table 1. Summary of Lipid Targets by Major Chronic Disease Guidelines in Adults.

Guideline Primary Target Secondary Target

NCEP‐ATP III (2004)4 LDL‐C Non‐HDL‐C

CCS (2012)3 LDL‐C Apo‐B

Non‐HDL‐C CDA (2013)46 LDL‐C ‐ ADA (2013)6 LDL‐C ‐

Several chronic disease guidelines have set lipid targets for individuals to meet to lower

cardiovascular risk. NCEP‐ATP III, CCS, CDA, and ADA have set LDL‐C as the primary treatment target to

prevent and manage CVD risk. In addition, NCEP‐ATP III has identified non‐HDL‐C as a secondary target

and similarly, CCS recognize both Apo‐B and non‐HDL‐C as secondary targets.

LITERATURE REVIEW

Page 8 of 93

2.2.1.1 National Cholesterol Education Program‐ Adult Treatment Plan III (NCEP‐ATP III)4

Table 2. Summary of Lipids Targets set by NCEP‐ATP III.

Primary Target Secondary Target

LDL‐C <2.59mmol/L Individuals with TG levels ≥2.26mmol/L,

non‐HDL‐C <3.37 mmol/L

NCEP‐ATP III has set absolute lipid targets for all adults (≥20 years) regardless of their health

status and number of CVD risk factors. An LDL‐C level of <2.59mmol/L (100mg/dL) or is considered

optimal. Individuals with triglyceride (TG) levels ≥2.26mmol/L (200mg/dL) or should also set their non‐

HDL‐C targets at <3.37 mmol/L (130mg/dL).

2.2.1.2 Canadian Cardiovascular Society (CCS)3

Table 3. Summary of Lipid Targets set by CCS (2012).

CCS has recommended different lipid targets depending on one’s 10‐ year CHD Framingham Risk

Score (FRS). At both the highest (FRS ≥20%) and intermediate (10‐19%) risk levels, CCS recommends the

LDL‐C target to be set at ≤2 mmol/L (77.22mg/dL) or ≥50% decrease from baseline and the secondary

targets for Apo‐B is set at ≤ 0.8 g/L and non‐HDL‐C at ≤2.6 mmol/L (100.39mg/dL). At the lowest risk

level (FRS <10%), the LDL‐C target is set at ≥ 50% reduction from baseline with no defined targets for the

secondary endpoints.

2.2.1.3 Canadian Diabetes Association (CDA)46

Table 4. Summary of Lipid Targets set by CDA (2013).


LDL‐C <2.0mmol/L Apo‐B <0.9g/L

Non‐HDL‐C <2.6mmol/L

Risk Level Primary Target Secondary Target

High (FRS ≥ 20%) LDL‐C ≤ 2 mmol/L or ≥ 50% decrease

Apo B ≤ 0.8 g/L Non HDL‐C ≤ 2.6 mmol/L

Intermediate (FRS 10‐ 19%) LDL‐C ≤ 2 mmol/L or ≥ 50% decrease

Apo B ≤ 0.8 g/L Non HDL‐C ≤ 2.6 mmol/L

Low (FRS <10%) LDL‐C ≥ 50% decrease ‐

LITERATURE REVIEW

Page 9 of 93

CDA recognizes fasting for >8 hours may be inappropriate for individuals with diabetes,

especially if insulin is part of treatment. Under these circumstances, non‐HDL cholesterol or Apo‐B

measurements should be used instead. The target for the primary endpoint, LDL‐C, is set at ≤2.0 mmol/L

(77.22mg/dL) for individuals with diabetes. Secondary lipid targets for individuals that cannot fast

include a Apo‐B target of <0.9 g/L and/or a non‐HDL‐C target of <2.6mmol/L (100.39mg/dL).

2.2.1.4 American Diabetes Association (ADA)6

Table 5. Summary of Lipid Targets set by ADA (2013).


Overt CVD, LDL‐C <1.8mmol/L Non‐overt CVD, LDL‐C<2.59mmol/L

‐

ADA recommends that individuals with diabetes should measure their fasting lipid profile at

least once a year, but those with a low‐risk lipid profile at least every 2 years. A low risk lipid profile is

defined as having levels of LDL‐C <2.59mmol/L (100 mg/dL), HDL‐C >1.30mmol/L (50 mg/dL), and TG

1.68mmol/L (150 mg/dL). For individuals without overt CVD, the target for LDL‐C is set at <2.59mmol/L

(100 mg/dL). For individuals with overt CVD, the target for LDL‐C is set at <1.8 mmol/L (70 mg/dL). No

secondary lipid target has been set.

2.3 OXIDATIVE STRESS AND CARDIOVASCULAR DISEASE

Oxidative damage to serum LDL‐C may play a potentially important role in the etiology of CHD

risk. It has been proposed that ox‐LDL‐C is more likely to be taken up by macrophages via scavenger

receptors and deposited as atheromatous plaques as the macrophage dies and becomes foam cells,

leading to increased incidence of cardiovascular events (Figure 1)47.

LITERATURE REVIEW

Page 10 of 93

Figure 1. Schematic Diagram of the Formation of Atheromatous Plaques via ox‐LDL‐C. Reprinted with permission from the Annual Review of Nutrition, Vol 25 © 2005 by Annual

Reviews www.annualreviews.org.47 Abbreviations: LDL, low‐density lipoprotein; MM‐LDL, minimally modified‐LDL; ox‐LDL, oxidized‐LDL; ROS, reactive oxygen species; SR‐A scavenger receptor A; ROS,

reactive oxygen species; RNS, reactive nitrogen species; M‐CSF, macrophage colony stimulating factor

2.4 DIET AND CARDIOVASCULAR DISEASE

2.4.1 DIET AND DRUGS

Many medications have become available for the prevention, management, and treatment of

CVD events. While many are effective and safe, recent examples have not delivered the anticipated

benefits. Statins have been shown to reduce CVD events, however these drugs are not well‐tolerated by

all for whom they are indicated48, and a recent meta‐analysis found statins did not reduce CVD

mortality2, 44. Further fibrates, a class of hypocholesterolemic medications, was demonstrated in a meta‐

analysis to increase rates of all‐cause mortality with no significant benefits for myocardial infarctions

and stroke49. Finally, a recent clinical drug trial of 25, 673 high risk individuals for CHD given extended

release of niacin vs. placebo found individuals given the drug treatment did not have lower rates of

myocardial infarctions and stroke, but rather increased risk of developing diabetes complications50. The

LITERATURE REVIEW

Page 11 of 93

lack of anticipated effects by drugs has reinforced the position of major health organizations that diet

and lifestyle change remain the cornerstone of CVD prevention and management.

2.4.2 DIETARY PATTERNS AND CARDIOVASCULAR DISEASE

Diet has been recognized as an essential and preventive method of reducing cardiovascular risk3‐

6. Since Ancel Key’s landmark publication of the Seven Countries Study which identified the influence of

nutrition and diet on CHD risk51, there has been a renewed interest on understanding the effects of

dietary and lifestyle patterns on heart health. In the past, nutrition studies have focused on a single

nutrient component or food on cardiovascular risk, but more recently, the attention has shifted to

focusing on the totality of the diet52. Diets such as DASH for blood pressure reduction9, Portfolio for lipid

improvements53, and Mediterranean for CHD risk reduction54 have all shown greater improvements in

CVD risk factors and events than any single nutrient component or food. In addition, the analysis of

overall dietary patterns is more reflective of real‐life conditions as individuals are likely to eat foods in

the context of a meal and substitute one food for another rather than simply adding another food

component to their diet55. For example, current recommendations advise to reduce total saturated fat

intake to lower CVD. However, a majority of studies have reported that saturated fat intake is harmful

only when substituting for polyunsaturated fatty acids (PUFAs), but there is no conclusive evidence of

harm when substituting for carbohydrates or monounsaturated fatty acids (MUFAs)56. Food choices

appear in clusters in diets and they reflective an overall pattern in the diet (eg. individuals with high

saturated fat intake are less likely to have high PUFA intake)55. Given that there is strong collinearity

between food intakes, focusing on encouraging a “healthy” dietary pattern may have more impact on

the prevention and management of CVD risk3‐6.

2.4.2.1 Prospective Cohort Studies

Prospective cohort studies have provided evidence that dietary patterns have important

implications for CVD prevention and management. Two “basic” dietary patterns have been identified: 1)

LITERATURE REVIEW

Page 12 of 93

Western pattern characterized by high consumptions of red and processed meat, butter, potatoes,

refined grains, and high‐fat dairy; 2) prudent pattern characterized by high consumptions of vegetables,

fruits, legumes, seafood, and whole grain consumption57. Although the Western diet is more prevalent

in some developed countries, more developing countries are adapting to this dietary habit as their

economic development becomes more integrated with the Western economy52. In the Health

Professional Follow‐Up Study (HPFS: 1986‐1994), after coronary risk factors were adjusted, the prudent

diet was associated with a decreased risk of 24% for CHD compared to the lowest‐quantile characterized

as having the lowest adherence to the prudent diet. In contrast, the Western diet was associated with

an increased risk of 46% for CHD when the extreme quantile was compared to the lowest55. The

difference in risk demonstrates the prudent diet offers an overall of 50% CHD risk reduction compared

to the Western diet55. Similar findings have also been reported by other prospective cohorts for both the

prudent and Western diets58, 59. These findings are not surprising as the prudent diet is made up of foods

and nutrient components that have been well associated with improved cardiovascular outcomes37, 52, 57.

There is evidence from dietary intervention trials suggestive that the combination of foods that are

known to offer cardio‐protection have additive effects when taken together53. Other benefits of health

have also been positively associated with greater adherence to the prudent diet including lower risks of

CVD events, cancer, and total mortality60.

2.4.2.2 Dietary Intervention Trials

Intervention trials have provided further support to strengthen the benefits of dietary patterns

on cardiovascular risk. Studies of dietary patterns have shown greater reductions in CVD risk factors

than isolated nutrients or foods. The Portfolio diet, for example, has 4 primary interventions: plant

sterols, viscous fibre, soy protein, and nuts. Individually each of these foods has been recognized by the

FDA to lower LDL‐C53. The combination of these foods in the same diet, however, has found significant

serum LDL‐C reductions to levels equivalent to the effects of early statin drug trials after 6 months of

LITERATURE REVIEW

Page 13 of 93

intervention, an outcome greater than any of the foods consumed singly53. The Dietary Approach to

Stop Hypertension (DASH) diet characterized by a low salt diet by increasing consumptions of fruits and

vegetables, nuts, whole grains, reduced low‐fat dairy and red meat intakes has also shown similar

benefits of dietary patterns over isolated foods but on blood pressure9. More importantly, the

Prevención con Dieta Mediterrànea (PREDIMED), a recent randomized feeding trial with a dietary

pattern intervention design, has shown benefits on clinical cardiovascular outcomes54. The PREDIMED

study compared the control diet (dietary advice to reduce fat intake) to 2 intervention arms, a

Mediterranean diet supplemented with either extra‐virgin olive oil or mixed nuts. Combined together,

the Mediterranean intervention diets showed a significant 30% reduction in the composite of

myocardial infarction, stroke, and death from cardiovascular‐cause events, after a median 4.8 years of

follow‐up. There is strong support demonstrating the effects of dietary patterns on cardiovascular risk

factors and clinical hard‐endpoints. These beneficial dietary pattern effects have been recognized by

guidelines in the prevention and management CVD3‐6.

2.5 PLANT‐PROTEIN BASED DIET AND CARDIOVASCULAR DISEASE

There is a growing interest in preventing cardiovascular risk using foods high in plant protein57,

61. Sources of high plant protein include fruits and vegetables, whole grains, soy, and dietary pulses.

Diets high in plant‐protein have been commonly been characterized in vegetarian and vegan diets, both

of which have demonstrated significant reductions of both cardiovascular risk factors and overall risk in

large prospective cohort studies such as the 62, 63. Higher plant protein intake may offer cardio‐protective

effects because it displaces refined carbohydrate consumption, and is associated with higher fibre and

lower saturated fat intakes37, 57, 64. It is also likely that the different sources of plant protein may have an

unique protein composition that further offers cardiovascular risk protection. For example, dietary

pulses and soy are high in 7S globulin, a protein that has been shown to have hypocholesterolemic

effects in humans18.

LITERATURE REVIEW

Page 14 of 93

2.5.1 PROSPECTIVE COHORT STUDIES

Prospective cohort studies have primarily focused on the analysis of vegetable proteins and

have reported a protective association against cardiovascular events. A study of a Finnish cohort (n=

5133) found an inverse association between vegetables and CHD risk with the multivariate relative risks

for CHD death at 0.66 (95% CI: 0.46, 0.96) for men and 0.66 (95% CI: 0.35‐1.23) in women65. Two of the

largest prospective cohorts, the Nurses’ Health Study (NHS: 1976‐1994) and Health Professionals’

Follow‐up Study (HPFS: 1986‐1994) have also shown reduced CVD events with higher vegetable and fruit

intakes66. Vegetables are high in many cardio‐protective nutrients including fibre, potassium, flavonoids,

and antioxidants57, but are also a good source of protein which may have important cardiovascular

implications. The Iowa Women’s Health Study of almost 30 000 post‐menopausal women, reported a

significant inverse relationship with vegetable protein intake and CHD deaths with greater displacement

of carbohydrates or animal protein by vegetable protein intake17. Contrary to this finding, the Adult

Health Study (AHS)67 and a Japanese prospective cohort study68, both of which did not assess for

displacement of nutrients like the Iowa Women’s Health Study, found no significant associations with

total vegetable protein intake and stroke. These findings appear to suggest the cardio‐protective effects

of vegetable proteins is greatest when it displaces carbohydrates or animal protein intakes; however

there is still the possibility that high vegetable protein intake is protective against of CHD mortality but

not stroke events. If the former is true, however, more studies are needed to better understand the

association of vegetable protein intake and stroke risk when vegetable protein displaces other nutrient

components in the diet.

Although soy protein has received particular attention for its hypolipidemic effects69, the

association of soy protein on cardiovascular events from prospective cohort studies have not been well‐

studied. To our knowledge, only one prospective cohort in Chinese women (n= 75 000) has been

published assessing soy protein intake and CHD events, which reported a 75% CHD event reduction with

LITERATURE REVIEW

Page 15 of 93

~18 grams per day of soy protein compared to ~5 grams per day70. Most of the cardio‐protective effects

of soy have been studied through dietary trials, however there is an urgent need for more studies to

improve the existing evidence for the association of soy protein intake and hard endpoints such as

myocardial infarction, stroke, and mortality.

2.5.2 DIETARY INTERVENTION TRIALS

With the exception of soy protein, there is a limited number of dietary intervention trials that

have assessed the effects of other sources of plant protein on cardiovascular risk factors. The

hypocholesterolemic effects of soy protein have been well documented69 and as a result, both Canada

and US have issued health claims to encourage the consumption of soy protein to improve lipid levels to

reduce CVD risk. Only one study has been identified to assess the effects for each of wheat gluten

protein31, barley protein71, and total vegetable protein72 on cardiovascular risk factors. Both trials of

wheat gluten and vegetable proteins have shown significant favourable on lipids with wheat gluten

protein also showing improvements in ox‐LDL‐C when compared to a high fibre diet31, 72. In contrast,

barley protein consumption did not show favourable effects on lipids when compared to casein;

however they may still be used to increase the overall intake of protein71. These studies offer support

for different sources of plant protein as a method of decreasing CVD risk; however there is a need for

more studies to improve the existing evidence to support the effects of plant protein on CVD risk

factors.

2.6 DIETARY PULSES AND CARDIOVASCULAR DISEASE

The Food and Agriculture Organization of the United Nations/World Health Organization

(FAO/WHO) define dietary pulses as: “dry seeds of leguminous plants which are distinguished from

leguminous oil seeds by their low fat content.73” They do not include soybeans or peanuts because both

are high in fat content, or green beans or peas because these include both the seeds and pods of the

LITERATURE REVIEW

Page 16 of 93

legume. They are high in plant protein, fibre, and have a low glycemic index, which have all been shown

to improve cardiovascular outcomes8, 13, 19.

2.6.1 DIETARY PULSES IN GUIDELINES

Major health organizations have encouraged dietary pulse intake as part of a strategy to lower

the risk of chronic diseases but none have directly made recommendations to increase dietary pulse

consumption based on lipid‐lowering effects and CVD prevention. Public health guidelines from Health

Canada and the FAO/WHO encourage dietary pulse intake as a method of increasing fibre intake, and

decreasing the risks of obesity, T2DM, and CVD74, 75. Chronic disease association guidelines from the

American Heart Association (AHA)76, as well as the American (ADA)6, Canadian (CDA)5, and European

(EASD)77 Diabetes Associations encourage dietary pulse consumption as a means for improving

cardiometabolic outcomes through reducing postprandial glycemia, reducing energy intake, lowering

the GI, increasing dietary fibre and displacing cholesterol. In all cases, the evidence on which these

recommendations is based has been assigned a low‐grade, while the FAO/WHO (minimum of 20‐g/d)

and AHA (2‐3‐cups/wk) are the only associations to find sufficient evidence to quantify intake targets74,

76. The Canadian Cardiovascular Society (CCS) did not make recommendations for dietary pulses;

however it does recognize the LDL‐C lowering effects of soy3. Neither Health Canada nor its American

(FDA) or European (European Food Safety Authority [EFSA]) counterpart permits any pulse related

health claims, and none of the major Canadian or American hypertension78, 79 and dyslipidemia3, 4

guidelines address dietary pulses in their recommendations. This situation, in which diabetes and heart

association have made recommendations for dietary pulses but have assigned the recommendation as

based on a low‐grade of evidence and the lack of data to support health claims and other

cardiometabolic recommendations, serves as a call for stronger evidence.

LITERATURE REVIEW

Page 17 of 93

2.6.2 DIETARY PULSES AND CARDIOVASCULAR DISEASE

The association of dietary pulses and cardiovascular risk is difficult to ascertain. Prospective

cohort and case‐control studies have grouped dietary pulses with soybean and peanut products

(collectively referred to as legumes) in the analyses. Collectively legumes have shown significant

cardiovascular risk reduction with greater intakes80, 81; however unlike soybeans and peanuts, dietary

pulses have a low fat content and are very likely to act differently than soybeans and peanuts in their

cardio‐protective effects. More likely, dietary pulses lower the risk of CVD through other nutritional

components. The NHANES (1999‐2002) study found that Americans. who consumed approximately ½

cup dry beans or peas had higher intakes of fibre, vegetable protein, folate, zinc, iron and magnesium

and lower intakes of saturated and total fat82, all of which are cardio‐protective. Dietary pulses are also

high in 7S globulin which has been shown to have hypocholesterolemic effects18. However to date, there

has been no prospective cohort study that directly assessed dietary pulses only and CVD risk. There is a

need for observational studies to better understand the relationship between high dietary pulse intake

and heart health.

2.6.3 DIETARY PULSES IN DIETARY PATTERNS

Dietary pulses have been included in several dietary patterns. These diets have been shown to

improve serum lipid levels, including the Portfolio53, low‐GI53, DASH9, high fibre10, and vegetarian12 diets.

Further the Mediterranean diet has been shown to improve the risk for composite of myocardial

infarctions, stroke, and cardiovascular‐cause deaths54. However as dietary pulses were included with

other dietary interventions in these studies, it is difficult to ascertain the independent effects of dietary

pulses on cardiovascular risk factors and outcomes. Dietary intervention trials specific to assessing

dietary pulses on cardiovascular risk are needed to better qualify and quantify its benefits.

LITERATURE REVIEW

Page 18 of 93

2.6.4 DIETARY PULSES IN INTERVENTION TRIALS

2.6.4.1 Evidence from Meta‐Analyses

Two previous meta‐analyses have been conducted investigating the effects of dietary pulses

(beans, chickpeas, lentils and peas) on lipids13, 83. Both studies involving 15 trials estimated higher

dietary pulse intakes for 2 to 16 weeks compared to control diets showed significant improvements in

total cholesterol (TC) (≈0.30 mmol/L, p<0.001; or ≈7.12%, p<0.05) and LDL‐C levels (≈0.20 mmol/L,

p<0.001; or ≈6.16%, p<0.05) at doses of 50‐440 grams per day (i.e. 1 to 10 175‐g servings/week), while

Anderson et al. also reported a significant TG reduction (16.6%, p<0.05)83. A review of controlled feeding

trials since these meta‐analyses were published shows the results have remained largely consistent8, 20,

25. Although both meta–analyses reported the effects of dietary pulses on LDL‐C, neither considered

other established lipid outcomes including Apo‐B and non‐HDL‐C. Controlled feeding trials, however,

have shown that consumption of dietary pulses reduce TC and VLDL‐C levels22, 23, 25, reductions which

would be reflected in a lower non‐HDL‐C level. Future meta‐analyses need to consider these established

lipid targets to provide higher level of evidence for the development of clinical practice guidelines for

the recommendations of dietary pulses.

2.6.4.2 Evidence from Randomized Feeding Trials in Humans

Clinical studies have shown regular consumption of dietary pulses can reduce serum TC and LDL‐

C but have little or no effect on HDL‐C 19‐28, 30. Few studies, however, have reported the effects of dietary

pulses on other established lipid targets such as Apo‐B and non‐HDL‐C.

The mechanism by which dietary pulses lower cholesterol levels is unclear; however, the effect

can be broadly divided into 2 pathways: extrinsic and intrinsic. The extrinsic pathway proposes that

dietary pulses reduce serum cholesterol by displacing saturated and trans fat intake. Americans who

consumed approximately ½ cup dry beans or peas daily had lower intakes of saturated and total fat82.

While the effects of saturated and total fat on cardiovascular risk are conflicting, many human clinical

LITERATURE REVIEW

Page 19 of 93

trials have shown a beneficial effect when saturated fat is replaced with polyunsaturated or

monounsaturated fatty acid intake (PUFA or MUFA, respectively)84. Further, dietary pulses are high in

soluble fibre which is known to bind to dietary cholesterol in the intestine and prevent its absorption85,

86. Possibly more importantly, dietary pulses are high in 7S globulins, a peptide which has been shown to

increase uptake and degradation of LDL‐C in human hepatic cells, and is also hypothesized to confer the

favourable lipid effects of soybeans18. Of the intrinsic pathway proposes that in addition to binding to

dietary cholesterol, the soluble fibre in dietary pulses can bind to bile acids to prevent its re‐absorption.

Consequently, there is an increase in the production of bile acids by the liver, which decreases the liver

pool of cholesterol and increases uptake of serum cholesterol by the liver thereby decreasing circulating

cholesterol in the blood22. Further fermentation of fibre in the colon produces short chain fatty acids

(SCFA) such as propionate and butyrate, which are associated with inhibition of cholesterol synthesis19,

23.

2.6.5 DIETARY PULSES AND OXIDATIVE STRESS

There is suggestion that a high dietary pulse intake may reduce levels of ox‐LDL‐C. Foods high in

plant protein such as wheat gluten have been shown to improve levels of oxidative stress in a dietary

intervention trial31. Dietary pulses are also high in plant protein; however, only one trial of 3 weeks in

duration (n= 30) has assessed the effects of dietary pulses, at an intake of 130 grams per day, on

oxidative stress, reporting no significant treatment differences when compared to an average American

diet32. Despite these findings, the study still reported a significant reduction of ox‐LDL‐C when

comparing the 2 treatment arms (p= 0.035)32. The findings of this study suggest that high dietary pulse

intake may still reduce levels of ox‐LDL‐C through a mechanism that is independent of oxidative stress.

Possibly dietary pulses lower the levels of serum LDL‐C, thereby reducing the availability of LDL‐C for

oxidation. Indeed several meta‐analyses have demonstrated high dietary pulse intake lowers LDL‐C

LITERATURE REVIEW

Page 20 of 93

levels13, 83. However given the short duration and small sample size of this study, larger, longer, and high‐

quality clinical trials are needed to further understand the effects of dietary pulses on oxidative stress.

2.6.6 DIETARY PULSES AND OTHER CARIOVASCULAR RISK FACTORS

Regular dietary pulse consumption has also been shown to improve other risk factors of CVD

including glycemia, body weight, and blood pressure. A meta‐analysis of 11 trials reported that dietary

pulses resulted in significant fasting blood glucose (standardized mean difference=‐0.82 [95% confidence

interval: ‐1.36 to ‐0.27]; p=0.03) and insulin (standardized mean difference= ‐0.49 [95% confidence

interval= ‐0.93 to ‐0.04]; p=0.03) reduction87. Early studies have shown high consumption of dietary

pulses resulted in exceptionally low glycemic responses when fed to healthy volunteers14, and in later

experiments were demonstrated to possess a carbohydrate profile that was more slowly digested than

that of other foods such as cereals88. This property of slower absorption, which is common to the α‐

glucosidase inhibitor class of medications such as acarbose suggest that dietary pulses are an important

means of lowering the glycemic index of the diet and, by analogy with acarbose, CHD risk89.

The review of the available evidence shows the data from experimental trials in humans are

largely inconclusive for body weight, with the only meta‐analysis published 10 years ago (n=11 trials)

finding no effect of dietary pulses on body weight.83 However, since the meta‐analysis has been

published, several dietary intervention studies have reported significant weight loss with higher dietary

pulse intake8, 90, suggesting that perhaps a beneficial body weight effect may have been missed in the

meta‐analysis due to the limited available studies. Further, a review of the few randomized trials

conducted to‐date is suggestive of a tendency for modest blood pressure‐lowering effect8, 20, 25, 90.

However, as a good source of fibre, dietary pulses would be expected to show a blood pressure lowering

effect91.

HYPOTHESIS, OBJECTIVES, AND RATIONALE

Page 21 of 93

CHAPTER III‐ HYPOTHESIS, OBJECTIVES, AND RATIONALE


Page 22 of 93

3. HYPOTHESIS, OBJECTIVES, RATIONALE 3.1 HYPOTHESIS

A diet high in dietary pulses will improve lipid risk factors of cardiovascular disease.

1. Diets emphasizing dietary pulses will improve levels of established lipid targets of cardiovascular

risk including LDL‐C, Apo‐B, and non‐HDL‐C compared to control diets in a systematic review and

meta‐analysis of randomized feeding trials

2. A high‐dietary pulse diet as a means to lower the glycemic index of the diet will reduce oxidative

stress markers including thiobarbituric acid reactive substances (TBARS), conjugated dienes

(CD), and protein thiols compared to a high‐fibre control diet in a secondary endpoint analysis

of a randomized, controlled, parallel trial of 12 weeks in T2DM

3.2 OBJECTIVES

Overall objective: To investigate the effects of dietary pulses (beans, chickpeas, lentils, and peas)

on lipid risk factors associated with increased CVD risk.

Specifically:

1. To assess the effects of dietary pulses on established lipid targets of cardiovascular disease in a

systematic review and meta‐analysis of randomized feeding trials.

2. To determine whether a diet high in dietary pulses, as a means to lower the GI of diet,

compared to a high fibre diet will improve oxidative damage to serum lipids and proteins, as

assessed by TBARS and CD in the LDL fraction and protein thiols in serum in a secondary

endpoint analysis of a 12‐week randomized feeding trial in T2DM.

3.3 RATIONALE

Although most major chronic disease prevention guidelines encourage the consumption of

dietary pulses, few guidelines have made recommendations based on the direct lipid lowering or

cardiovascular risk reduction benefits of dietary pulses. While 2 previous meta‐analyses have assessed


Page 23 of 93

the effects of dietary pulses on lipids13, 83, neither assessed the effects on established lipid targets nor

did they include studies that were at least 3 weeks in duration, a requirement to satisfy the FDA criteria

used in the scientific evaluation of lipid‐lowering health claims92. In all cases, the evidence on which the

recommendations for dietary pulses are based on has been assigned a low‐grade. Therefore, there is a

need for higher quality evidence to support the lipid‐lowering effects of dietary pulses for guideline

development. Therefore a systematic review and meta‐analysis assessing the effects of dietary pulses on

established lipid targets of CVD was undertaken.

Dietary pulses may offer further cardiovascular protection through reducing oxidized‐LDL‐C (ox‐

LDL‐C). Dietary pulses are a good source of vegetable protein and this nutritional property may provide

further protection from CVD by reducing the oxidation of LDL‐C31. Further like soy, they are high in

phenolics which may have beneficial effect on oxidative stress. Few feeding trials in humans have

assessed the effects of dietary pulses on oxidative stress except for one small trial of 30 individuals32.

Therefore a human feeding trial comparing the effects of a high dietary pulse diet as a means to lower

the GI to a high fibre comparator diet on oxidative stress was undertaken.

DIETARY PULSES AND LIPIDS

Page 24 of 93

CHAPTER IV‐ THE EFFECT OF DIETARY PULSES ON ESTABLISHED LIPID THERAPEUTIC TARGETS OF CARDIOVASCULAR DISEASE: A SYSTEMATIC REVIEW AND META‐ANALYSIS OF RANDOMIZED

CONTROLELD TRIALS


Page 25 of 93

4.1 CITATIONS

The abstract of this work has been published in the following journals:

1. Ha V, Jayalath VH, de Souza RJ, Sievenpiper JL, Mirrahimi A, Chiavaroli L, Beyene J, Kendall CWD,

Jenkins DJ. The Effects of Dietary Pulses on Established Lipid Therapeutic Targets of Cardiovascular

Disease: A Systematic Review and Meta‐analysis of Randomized Controlled‐Feeding Trials. Journal of

the American College of Nutrition 2012; 31(3): 218.

2. Ha V, Sievenpiper JL, de Souza RJ, Jayalath VH, Mirrahimi A, Blanco Mejia S, Beyene J, Kendall CWD,

Jenkins DJ. Dietary Pulse Intake May Improve Levels of LDL‐C and Non‐HDL‐C: A Systematic Review

and Meta‐analysis. Canadian Journal of Diabetes 2012; 36 (5): S10.

3. Ha V, de Souza RJ, Sievenpiper JL, Jayalath VH, Beyene J, Jenkins DJA, Kendall CWC. The Effect of

Dietary Pulses on Lipids in Controlled Feeding Trials: A Systematic Review and Meta‐Analysis. FASEB J

2012; 26:117.4.


Page 26 of 93

THE EFFECT OF DIETARY PULSES ON ESTABLISHED THERAPEUTIC LIPID TARGETS OF CARDIOVASCULAR

DISEASE: A SYSTEMATIC REVIEW AND META‐ANALYSIS OF RANDOMIZED CONTROLLED TRIALS

Vanessa Ha MSc Candidate1,4, John L Sievenpiper Knowledge Synthesis Lead and Resident Physician1,7,

Russell J de Souza Post‐doctoral Fellow1,4,8, Viranda H Jayalath Research Student1,4, Arash Mirrahimi

Research Coordinator1, Arnav Agarwal Research Student,9 Laura Chiavaroli PhD Candidate1,4, Sonia

Blanco‐Mejia Research Coordinator1,4, Frank M Sacks Professor of Cardiovascular Disease

Prevention13,14,15, Marco Di Buono Vice President, Science and Research10, Adam M Bernstein Director of

Research13,16, Lawrence A Leiter Professor of Medicine and Nutritional Sciences1,2,3,4,5, Penny M Kris‐

Etherton Distinguished Professor of Nutrition17, Vladimir Vuksan Professor of Nutritional Sciences1,2,4,

Richard P Bazinet Associate Professor of Nutritional Sciences4, Robert G Josse Professor of Medicine and

Nutritional Sciences1,2,3,4,5, Joseph Beyene Associate Professor of Clinical Epidemiology and

Biostatistics6,8,11, Cyril WC Kendall Research Associate and Adjunct Professor of Nutritional Sciences1,4,12,

David JA Jenkins University Professor of Medicine and Nutritional Sciences1,2,3,4,5,

1Clinical Nutrition and Risk Factor Modification Centre, 2Keenan Research Centre of the Li Ka Shing Knowledge

Institute, and 3Division of Endocrinology, St. Michael’s Hospital, Toronto, ON, CANADA; 4Departments of

Nutritional Sciences, and 5Medicine, and 6The Dalla Lana School of Public Health, Faculty of Medicine, University of

Toronto, Toronto, ON, CANADA; 7Departments of Pathology and Molecular Medicine, 8Clinical Epidemiology &

Biostatistics, 9Health Sciences, Faculty of Health Sciences, McMaster University, Hamilton, ON, CANADA; 10American Heart Association, Houston, Texas, USA; 11The Hospital for Sick Children Research Institute, Toronto,

ON, CANADA; 12College of Pharmacy and Nutrition, University of Saskatchewan, Saskatoon, Saskatchewan,

CANADA; 13Department of Nutrition, Harvard School of Public Health, and 14the Channing Laboratory, Department

of Medicine, Brigham and Women’s Hospital and 15Harvard Medical School, Cambridge, MA, USA; 16Wellness

Institute of the Cleveland Clinic, Lyndhurst, OH, USA; 17Department of Nutritional Sciences, Pennsylvania State

University, University Park, PA, USA.

Corresponding author: John L Sievenpiper MD, PhD Toronto 3D Knowledge Synthesis and Clinical Trials Unit, St. Michael's Hospital 61 Queen Street East, Toronto, ON, M5C 2T2, CANADA Tel: 416 867 7475 Fax: 416 867 7495 Email: [email protected]


Page 27 of 93

4.2 ABSTRACT

Background: Evidence from controlled feeding trials supports dietary pulses (beans, peas, chickpeas,

and lentils) as a dietary intervention to improve dyslipidemia, but heart health guidelines have stopped

short of ascribing specific heart‐health benefits to dietary pulses or have graded the beneficial evidence

as low. To improve the evidence on which dietary guidelines are based, we conducted a systematic

review and meta‐analysis to assess the effect of dietary pulses on established therapeutic lipid targets

for cardiovascular risk reduction: LDL‐C, Apo‐B, and non‐HDL‐C.

Methods and Results: We searched MEDLINE, EMBASE, Cochrane, and CINAHL databases through Feb

11 2013 and included randomized controlled feeding trials of ≥3‐weeks duration that compared a diet

emphasizing dietary pulse intake with an isocaloric diet without pulses. Three independent reviewers

extracted relevant data with disagreements resolved by consensus. Data were pooled by the generic

inverse variance method using random effects models and expressed as mean differences (MD) with

95% confidence intervals (CI). Heterogeneity was assessed (Chi2) and quantified (I2). Study quality and

risk of bias were assessed. Twenty‐six isocaloric trials (n= 1013) satisfied the inclusion criteria. Diets

emphasizing dietary pulses at a median dose of 130g/d (~1.5 servings) significantly lowered LDL‐C

compared with isocaloric control diets (MD= ‐0.17 mmol/L [95% CI: ‐0.25, ‐0.09]; p<0.0001). No

treatment effects were observed for Apo‐B and non‐HDL.

Conclusions: Pooled analyses demonstrated that dietary pulses significantly improve LDL‐C. The majority

of trials, however, were short in duration and poor quality. There is a need for larger and higher quality

trials

Clinical Trial Registration: NCT01594567

Abstract word count: 247

Key words: systematic review, meta‐analysis, dietary trials, legumes, dietary pulses, dyslidpidemia,

cardiovascular disease


Page 28 of 93

4.3 INTRODUCTION

Abnormal blood concentrations of lipids and lipoproteins are one of the most important

modifiable risk factors for cardiovascular disease (CVD). Although statins are effective in reducing blood

LDL‐C2, major health organizations maintained that the initial and essential approach to CVD prevention

and management is to modify dietary and lifestyle patterns3‐6.

Dietary non‐oil‐seed pulses are foods which have received particular attention for their ability to

reduce CVD risk as part of low‐glycemic index (GI)8, Dietary Strategies to Stop Hypertension (DASH)9,

high‐fibre10, Mediterranean11, and vegetarian12 dietary patterns. Dietary non‐oil‐seed pulses, including

beans, peas, chickpeas, and lentils, have a low GI, are high in viscous soluble fibre, vegetable protein,

and various bioactive compounds13, 86. Observational studies have shown legumes including dietary

pulse consumptions are associated with CVD reduction80, 81 and small controlled feeding trials8, 19‐23, 25‐30,

90, 93, 94 have demonstrated that diets which emphasize dietary pulses improve LDL‐C levels. Although

most major chronic disease prevention guidelines encourage the consumption of dietary pulses as part

of a dietary strategy3, 74, 77, 95, 96, none of the guidelines have made recommendations based on the direct

lipid lowering or cardiovascular risk reduction benefits. The focus of guidelines, such as those of the

American Heart Association which recommends ≥4‐5 servings/week96, has been on the effects of dietary

pulses to improve the diet quality in the management of diabetes and CVD risk5, 6, 74, 77, 95, 96. In all cases,

the evidence on which recommendations are based has been assigned a low‐grade5, 6, 74, 77, 95, 96.

Dyslipidemia guidelines including those from the National Cholesterol Education Program (NCEP) adult

treatment panel III (ATP III) and the Canadian Cardiovascular Society (CCS) do not address dietary pulses

intake directly3, 4. There is a need for higher quality evidence to support the lipid lowering effects of

dietary non‐oil‐seed pulses for guidelines development.

To improve the evidence on which dietary guidelines are based, we conducted a systematic

review and meta‐analysis of randomized controlled feeding trials on the effect of dietary non‐oil‐seed


Page 29 of 93

pulses on established therapeutic lipid targets for cardiovascular risk reduction including LDL‐C, Apo‐B,

and non‐HDL‐C.

4.4 METHODS

The Cochrane Handbook for Systematic Reviews of Intervention was used as a guideline for this

meta‐analysis97. Reporting of results followed the Preferred Reporting Items for Systematic Reviews and

Meta‐Analyses (PRISMA) guidelines98. The review protocol is available at ClinicalTrials.gov (registration

no. NCT01594567).

4.4.1 STUDY STRATEGY SELECTION

We searched MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, and CINAHL

through Feb 11 2013, to identify randomized controlled dietary trials of dietary non‐oil‐seed pulses.

Table 1 shows the search strategy. Manual searches of references cited by published studies also

supplemented the database search. No restriction was placed on language.

4.4.2 ELIGIBILITY CRITERIA

We included randomized trials that investigated the effect on LDL‐C, Apo‐B and non‐HDL‐C of

diets high in dietary non‐oil‐seed pulses compared to control or usual diets matched for energy for a

period of ≥3 weeks, a duration which satisfies the minimum follow‐up requirement of the FDA used in

the scientific evaluation of lipid‐lowering health claims92. Studies that only examined dietary non‐oil‐

seed pulses (beans, chickpeas, lentils, and peas) were included; peanuts and soybeans were excluded

because of their high oil content. We included interventions in which dietary pulses were not the sole

intervention, provided that dietary pulses were the dominant intervention used to achieve intervention

goals (n=3)8, 10, 99. Studies providing only dietary pulse extracts, such as bean extracts or isolated fibre

supplements, were not eligible. Both control and treatment arms must have been matched for total

energy (i.e. an isocaloric comparison). The selected endpoints included only those which have been


Page 30 of 93

identified as therapeutic lipid targets in the NCEP‐ATP III4, 2012 CCS3, 2012 American Diabetes

Association (ADA)6, and 2013 Canadian Diabetes Association (CDA) guidelines3.

The study by Anderson et al.85 was quasi‐randomized. We attempted to reduce bias induced by

non‐randomization by re‐analyzing this study by randomly assigning participants to either treatment or

control, stratified by baseline total cholesterol and age, and treating the post‐control and post‐

treatment arms as a parallel study design.

4.4.3 DATA EXTRACTIONS

Studies that met the inclusion criteria had their study characteristics and results extracted by 3

independent reviewers (VH, VHJ, and RJdS). As well, a 10‐year CHD risk score was calculated for each

study’s population using the Framingham risk equation3. If information was missing for CHD risk

prediction, participants were assumed to be at higher risk for that particular dimension (eg. assumed to

be smokers, hypertensive, low HDL‐C, and/or have family history of CHD).

The quality of each study was assessed using Heyland’s Methodological Quality Score (MQS)100.

Studies could receive a maximum score of 13 points. Studies with scores of ≥8 were considered high

quality. Points were awarded based on the quality of study methods, sample selection and follow‐up,

and intervention.

Studies were assessed for risk of bias using the Cochrane Risk of Bias Tool. Domains of bias

assessed were sequence generation, allocation concealment, blinding, outcome data, and outcome

reporting97. Studies were marked as, high risk of bias when a methodological flaw was likely to have

affected the true outcome, low risk of bias if the flaw was deemed inconsequential to the true outcome,

and unclear risk of bias when insufficient information was provided to permit judgment. All

disagreements were resolved by consensus.


Page 31 of 93

4.4.4 TREATMENT OF MISSING DATA

When non‐HDL‐C was not reported, they were calculated from aggregate data using the

formula: non‐HDL =TC‐HDL‐C. A formula was used to calculate SD for non‐HDL‐C:

; N denotes the total sample size; S denotes standard deviation; and as

a reliable covariance ( ) could not be estimated, it was assumed to be zero. Trials that did not

report either change‐from‐baseline differences within or between treatments, or end‐differences

between treatments had these calculated from the available study data using standard formulae97.

Authors were contacted, when possible, to request additional information (n= 3)20, 30, 90.

4.3.5 STATISTICAL ANALYSIS

Data were analyzed using Review Manager (RevMan) 5.0.25. Pooled analyses for isocaloric

dietary pulse feeding trials were conducted using the Generic Inverse Variance method using random

effects weighting. Mean endpoints of LDL‐C, Apo‐B and non‐HDL‐C were compared between dietary

pulse arms and comparator arms. Data were expressed as mean differences (MD) with 95% CI. To

mitigate the unit‐of‐analysis error from including trials with multiple intervention arms, we combined

arms to create single pair‐wise comparisons (n=1)94. To impute standard deviations for between‐

treatment differences in crossover trials, correlation coefficients between baseline and end‐of‐

treatment values within each individual trial were derived from the reported within and between

treatment SD according to a published formula101. These correlation coefficients were transformed into

z‐scores ± SD, meta‐analyzed using inverse‐variance weighing, and back transformed to derive a pooled

correlation coefficient. A pooled correlation of 0.72 was used for LDL‐C analysis but a correlation

coefficient of 0.5 was assumed for non‐HDL‐C, owing to a lack of data, with sensitivity analyses done at

0.25 and 0.75. A two‐sided p‐value <0.05 was set as the level of significance for comparisons of MD.


Page 32 of 93

The Q‐statistic (Chi2) assessed and I² quantified the inter‐study heterogeneity with significance

set at p<0.10. An I² ≥50% indicated “substantial” heterogeneity and ≥75% indicated “considerable”

heterogeneity97. Sources of heterogeneity were explored using a priori subgroup analyses according to

baseline cholesterol values, pulse dose and type, duration of follow‐up, the difference in fibre content

and saturated fat between the dietary pulses and control diet, study design (crossover or parallel), and

study quality (MQS). A priori subgroup analyses were not conducted for Apo‐B because only one study

was included. A post‐hoc analysis by sex (percentage of males) was also conducted. To determine if any

single trial exerted an undue influence on the overall results, a sensitivity analysis was undertaken in

which each individual study was removed from the meta‐analysis and the effect size was re‐calculated

with the remaining studies.

Possible publication bias was assessed by visual inspection of funnel plots, and formally tested

using Begg’s and Egger’s tests, with a P value <0.05 considered evidence of small‐study effects.

4.5 RESULTS

4.5.1 SEARCH RESULTS

Figure 1 shows the flow of literature. The search identified 3002 reports, 2940 of which were

determined to be irrelevant on review of the titles and abstracts. The remaining 62 reports were

reviewed in full. A total of 22 reports providing data for 26 randomized trials were identified8, 10, 19‐30, 85, 86,

90, 93, 94, 99, 102, 103.

4.5.2 STUDY CHARACTERISTICS

Table 2 shows the characteristics of the 26 randomized trials (n=1013). Eight trials (31%) were

conducted in hyperlipidemic participants, 3 trials (12%) in normolipidemic participants, and 15 trials

(58%) in a combination of both. Median age of participants was 50.2 years with approximately equal

number of men and women. Median baseline LDL‐C was 3.50 mmol/L and non‐HDL‐C was 4.34 mmol/L.

Studies had a median of 3 risk factors associated with increased CHD risk as defined by the Framingham


Page 33 of 93

risk equation, implicating a moderate CHD risk (i.e. 10‐year CHD risk ≤20%)4. Three trials were rated as

CHD risk equivalent as according to NCEP‐ATP III guidelines4 because they were conducted in individuals

with Type 2 Diabetes8, 29, 99. All trials explicitly excluded individuals on lipid‐lowering medications with

the exception of 3 trials where it was unclear. Participants were from 6 countries: 11 trials were from

the United States (42%), 6 trials from Australia (23%), 5 trials from Canada (19%), 2 trials from Spain

(8%), and one trial each from Ireland and Iran (4%). Dietary pulse type included: beans (14 trials [54%]),

peas (2 trials [8%]), chickpeas (2 trials [8%]), lentils (1 trial [4%]), and mixed pulses (8 trials [31%]).

Dietary pulses were administered as flour in 3 trials (12%), as whole foods in 18 trials (69%), and mixed

format (flour and/or whole foods) in 3 trials (12%) at a median dose of 130g/day (range: 50‐377g/d).

Two trials did not report the form in which dietary pulses were administered as. The background diet

consisted of 39‐65% energy (E) from carbohydrate, 10‐35% E protein, and 20‐41% E fat with a median

fibre and saturated fat intake of 20g/d (range: 13‐47g/d) and 11%E (range: 5‐15%E), respectively, in the

comparator diets, and 26g/d (range: 17‐53g/d) and 11%E (range: 5‐15%E), respectively, in the dietary‐

pulse enriched diets. The method of increasing dietary pulses while maintaining caloric balance between

arms differed across protocols: 15 trials (58%), replaced non‐dietary pulse carbohydrates (eg. bread

products, canned spaghetti, oat bran), specifically 7 trials replaced fibre, 5 trials (19%) replaced animal

protein, 3 trials (12%) emphasized dietary pulses to achieve a low‐glycemic diet, and 3 trials (12%) did

not specify. Three trials (12%) were weight loss interventions designed to reduce total caloric intake by

30 or 35%. Five trials (19%) were metabolically‐controlled (eg. all foods were provided), 17 trials (65%)

were partially metabolically controlled (eg. test foods were provided), and 4 trials (15%) were not

metabolically controlled (eg dietary advice was offered). Thirteen trials (50%) had a crossover design.

Twenty‐two trials (84%) were conducted in an outpatient setting, 2 trials (8%) in an inpatient setting,

and 2 trials (8%) in a combination of the two. Median follow‐up was 6 weeks (range: 3‐52 weeks).


Page 34 of 93

By the Heyland MQS, seven trials (27%) were of high quality (MQS ≥8). Poor description of

randomization, absence of double‐blinding, lack of intention‐to‐treat statistical analysis, and high

dropout rates contributed to lower scores (Table 3). The Cochrane Risk of Bias Tool showed that 17

(77%) reports had unclear risk of bias and 5 reports (23%) had low risk of bias for sequence generation.

Sixteen reports (73%) had unclear risk of bias and 6 reports (27%) had low risk of bias for allocation

concealment. Nine reports (41%) had unclear risk of bias, 12 reports (54%) had low risk of bias, and 1

report (4%) had high risk of bias for blinding. Twenty‐one reports (95%) were scored with low risk of bias

for incomplete outcome data. Twelve reports (54%) scored unclear risk of bias for selective outcome

reporting (Table 4). Funding of all trials was from: agency alone (50%); agency‐industry sources (27%);

industry alone (15%); and 2 trials were unclear of their funding sources.

4.5.3 GASTRO‐INTESTINAL SYMPTOMS

Eleven trials provided self‐reported gastrointestinal (GI) symptoms data. Four trials (36%)

reported participants with upset stomach, 7 trials (64%) with flatulence, 6 trials (54%) with bloating, 1

trial with diarrhea and constipation each (9%), and 3 trials (27%) with increased stool frequency. Two

trials compared GI symptoms to the comparator group but found no statistical differences across all

symptoms except for increased flatulence in the comparator for 1 trial28. Most studies reported that

symptoms tended to improve over the course of the dietary pulse treatment. Only 2 trials had 1 to 2

participants reporting GI symptoms as a reason for study withdrawal24, 28.

4.5.4 EFFECT ON LDL‐C

Twenty‐one reports involving 25 comparisons (n= 686 intervention, 684 control) provided data

on the effects of dietary pulses on LDL‐C (Figure 2). Diets emphasizing dietary pulses significantly

reduced LDL‐C compared to control diets (MD= ‐0.17 mmol/L [95% CI: ‐0.25, ‐0.09]; p<0.0001) but

showed significant evidence of inter‐study heterogeneity (I2= 80%; p≤0.00001). Sensitivity analyses in


Page 35 of 93

which each individual study was systematically removed and the effect size recalculated did not alter

conclusions.

A priori subgroup analyses were carried out to test for possible modification of the effect of

dietary pulses on LDL‐C by study design characteristics. None of the subgroup analyses showed

significant effects. In a post‐hoc analysis by sex, we found greater LDL‐C reduction as the ratio of males

to females increased (β= ‐0.0036 [95% CI: ‐0.006, ‐0.001]; p= 0.012) (Figure 3 and Table 5).

4.5.5 EFFECT ON NON‐HDL‐C

Twenty reports involving 22 comparisons (n= 605 intervention, 605 control) provided data on

the effects of dietary pulses on non‐HDL‐C (Figure 4). Diets emphasizing dietary pulses did not have a

significant effect on non‐HDL‐C (MD= ‐0.09 mmol/L [95% CI: ‐0.19, 0.00]; p=0.06) but showed significant

evidence of inter‐study heterogeneity (I2= 98%; p≤0.00001). Sensitivity analyses showed when Abete et

al.90 Anderson et al. (1984)19, Belski et al.20, Gravel et al. 24, Hermsdorff et al.25, or Marinangeli et al. was

removed, the pooled effect size became significant. The use of a correlation coefficient of 0.25 did not

alter conclusions but a correlation coefficient of 0.75 resulted in a significant reduction in non‐HDL‐C.

A priori subgroup analyses were carried out to determine if the effect of dietary pulses on non‐

HDL‐C differed according to study design characteristics. We found higher fibre intake in the dietary

pulse compared to the control arm showed a significantly greater non‐HDL reduction. None of the other

subgroup analyses showed significance (Figure 5 and Table 5). Post‐hoc analysis by sex also did not

shown significant (Table 5).

4.5.6 EFFECT ON APOLIPROTEINS B

One trial met the inclusion criteria for Apo‐B analysis. Isocaloric exchange of dietary pulses for

another diet did not significantly alter Apo‐B (MD= 0.02 [95% CI:‐0.04, 0.08]) (Figure 6).


Page 36 of 93

4.5.7 PUBLICATION BIAS

Inspection of funnel plots for evidence of publication bias revealed asymmetry favouring small

studies with LDL‐C reducing effects (Figure 7A). Egger’s test detected significant evidence of publication

bias in LDL‐C trials (p= 0.003) but Begg’s test did not show significant evidence of bias. Non‐HDL‐C

showed no significant evidence of publication bias by visual inspection of funnel plot, Egger’s and Begg’s

tests (Figure 7B).

4.6 DISCUSSION

We conducted a systematic review and meta‐analysis of 26 randomized, controlled trials of the

effect of dietary non‐oil‐seed pulses (beans, chickpeas, lentils and peas) on established therapeutic lipid

targets for cardiovascular risk reduction in 1013 predominantly middle‐age, normolipidemic or

hyperlipidemic adults at moderate CHD risk, most of whom were not taking lipid lowering medications.

Pooled analyses showed a significant LDL‐C‐lowering‐effect of 0.17 mmol/L with a median dose of

130g/d of dietary pulses (~1.5 servings per day) over a median follow‐up of 6‐weeks. There was no

significant effect of dietary pulses on Apo‐B and non‐HDL‐C. Although some studies reported increased

GI symptoms initially from the dietary pulse treatment, most reported that the effects subdued over the

course of the study.

4.6.1 RESULTS IN RELATION TO OTHER STUDIES

This is the first systematic review and meta‐analysis to report the effect of dietary pulse intake

on all established lipids and lipoprotein CHD risk factors including LDL‐C, Apo‐B, and non‐HDL‐C. The

observed LDL‐lowering effect of dietary pulses is consistent with that of 2 previous meta‐analyses13, 83;

however, our analysis was limited to randomized controlled trials with at least 3 weeks of follow‐up

duration in conformity with FDA guidelines92.

We found a non‐significant decreasing effect on non‐HDL‐C despite a significant reduction in

LDL‐C, suggesting a rise in VLDL‐C. This contradictory result may be explained by elevated triglyceride


Page 37 of 93

(TG) levels. Dyslipidemic guidelines suggest that when TG level is greater than 2.24 mmol/L

(~200mg/dL)4 or 1.50 mmol/L (~133mg/dL)3, LDL‐C and non‐HDL‐C no longer correlate well with each

other. We found 13 trials (52%) in our analysis reported higher TG levels (≥2.24 mmol/L) and in a post‐

hoc analysis, individuals with normal levels of TG had greater reductions in LDL‐C and non‐HDL‐C than

those with high levels of TG (data not shown). However, we were unable to conduct a correlation

analysis to examine the association of LDL‐C and non‐HDL‐C stratified by TG levels; future analyses

should consider this association.

The mechanism by which dietary pulses lower cholesterol levels is unclear; however there are

several hypotheses that have been generated to explain this benefit. Dietary pulses reduce serum

cholesterol by displacing saturated and trans fat intake. The 1999‐2002 NHANES study found that

Americans, who consumed approximately ½ cup dry beans or peas daily had lower intakes of saturated

and total fat82. While the effects of saturated and total fat on cardiovascular risk are conflicting, many

clinical trials have shown a beneficial effect when saturated fat is replaced with monounsaturated and

polyunsaturated fatty acid intake (MUFA and PUFA, respectively)84. Further, dietary pulses are high in

viscous fibre. Viscous fibre has been shown to trap bile acids in the intestine and prevent their

absorption85, 86, and human studies have demonstrated significant increases in fecal bile acid output

related to LDL‐C reduction104. Consequently, there is an increase in the production of bile acids by the

liver, which decreases the liver pool of cholesterol and increases uptake of LDL‐C by the liver thereby

decreasing circulating LDL‐C in the blood22. Further fermentation of fibre in the colon produces SCFA

such as propionate and butyrate, which have been associated with inhibition of cholesterol synthesis in

the liver19, 23. Possibly of greater importance is the presence of a 7S globulin fraction, common to dietary

pulses, which is consistent with this cholesterol reduction property18. It has been shown that peptides

digested from this fraction, which may be absorbed intact, have a markedly inhibiting effect on Apo‐B

synthesis, the lipoprotein which influences hepatic cholesterol secretion.


Page 38 of 93

4.6.2 INTER‐STUDY HETEROGENEITY

We reported significant evidence of inter‐heterogeneity in our LDL‐C analysis; however none of

our a priori subgroup analyses could explain the source of inter‐study heterogeneity. To further

investigate possible explanations, we conducted a post‐hoc subgroup analysis by sex, where we treated

the ratio of males to females from each study as a continuous variable. We found that studies

conducted in all males tended to show a greater difference in LDL‐C levels with dietary pulse intake

compared to control10, 19, 21, 22, 85, 90, than in the one study conducted in females only24. Possible reasons

why males may have shown a greater LDL‐C reduction was because: 1)males tend to have poorer dietary

patterns and respond more favourably to risk factors of CVD when dietary habits are healthier105;

2)males of similar age have higher levels of LDL‐C than pre‐menopausal and post‐menopausal women

on hormone replacement therapy (HRT) and therefore LDL‐C is more likely to be reduced in men4; 3)

compared to the total cholesterol fraction, males have higher LDL‐C whereas females have higher HDL‐

C; the former is more responsive to dietary changes than the latter4 and; 4)as there is only one all‐

female study included, there may not be enough female‐only studies for comparison. Although the

subgroup analysis by sex showed reduction of inter‐study heterogeneity, the level is still moderate and

future analyses are needed to better understand the sources of heterogeneity.

The effect of dietary pulses on LDL‐C should not be underestimated despite the high level of

inter‐study heterogeneity. The LDL‐C effect translated to a ~5% reduction from baseline. This reduction

in LDL‐C equates to ~5‐6% risk reduction in major vascular events (major coronary event, stroke, and

coronary revascularization)45. This is important especially in hypercholestrolemic patients who prefer

dietary approaches to managing cholesterol levels or for those who cannot tolerate statin therapies.

Further, the MDs of the majority of trials (23 of 25) fell within the 95% CI of the pooled estimate for LDL‐

C suggesting robustness in our data and increasing confidence in our conclusions. Therefore, while inter‐


Page 39 of 93

study heterogeneity was high in our pooled analysis of LDL‐C, the effect estimate may still have

meaningful clinical applications.

Our non‐HDL‐C findings were complicated by significant evidence of inter‐study heterogeneity.

In a previous systematic review and meta‐analysis on the effect of dietary pulses on glycemic control,

we reported significant effect modification by diabetes status, pulse type, dose, follow‐up duration,

study design, study quality, and baseline metabolic control87. In the current analysis, we only found a

significant reducing non‐HDL‐C effect when the dietary pulse arm had greater fibre intake than the

comparator. Diets high in fibre have been shown to reduce non‐HDL‐C106 and be inversely associated

with CHD risk15. However, a substantial amount of inter‐study heterogeneity remains unexplained. To

further address inter‐study heterogeneity, we conducted sensitivity analysis, where the removal of

Abete et al.90, Anderson et al. (1984)19, Belski et al.20, Gravel et al.24, Hermsdorff et al.25, or Marinangeli

et al.102 resulted in a significant non‐HDL‐C reduction. Each of these studies favoured the effect of the

control treatment, but there was no common characteristics common amongst them that would lead us

to believe that there is any bias in these analyses. Other sources of heterogeneity across trials need to

be investigated when more trials are published.

4.6.3 LIMITATIONS

There are limitations to our work. First, the majority of studies were of low quality (MQS<8),

short duration (<3 months), and did not report enough data to judge risk of bias. These observations

reinforce the need for longer, better‐designed clinical trials that report data more systematically and

transparently to better understand the effect of dietary pulses on serum lipids. Second, only one trial

reported on non‐HDL‐C values; therefore we had to calculate non‐HDL‐C and imputed its SD. To assess

whether our method of SD imputation was reasonable, we conducted a sensitivity analysis, where we

pooled reported SDs and substituted the calculated SD for trials that did not report these values. Our

findings for non‐HDL‐C indicate that the second method of SD imputation did not alter our primary


Page 40 of 93

conclusions, suggesting that our method of SD imputation is unlikely to have biased our conclusions.

Third, only one trial provided Apo‐B data. These deficiencies highlight the need for future trials to

examine these endpoints as they are important biomarkers of cardiovascular health. Lastly, publication

bias remains a possibility. We observed plot asymmetry favouring small studies with LDL‐C reducing

effects which was confirmed by a significant Egger’s test.

4.6.4 CONCLUSIONS

Our findings have implications for cardiovascular health. Dietary pulses resulted in a modest yet

clinically meaningful reduction in the primary therapeutic lipid risk factor, LDL‐C, showing a 0.17 mmol/L

reduction (equivalent to ~5% reduction from baseline). This is important especially when combined as

part of diets that have shown clinically impactful lipid‐lowering that can potentially achieve upwards of

30% reductions in LDL‐C53 and may be particularly helpful for those who prefer dietary approaches to

managing cholesterol levels or for those who cannot tolerate statin therapies. The median pulse intake

of ≈1.5 servings/day required to achieve this benefit supports a higher intake target than that proposed

by the AHA of ≥4‐5 servings/week. Achieving a high level of intake may prove a challenge in some

Western countries, given that current U.S. level of intake is 0.2 serving/day. However, a dietary pulse

intake of 130g/d is very reasonable as this level is currently consumed by many cultures without reports

of side effects that would limit consumption. As the majority of these trials were conducted on a

background heart‐healthy NCEP‐like diet, including >20‐25g/d of fibre and <10%E of saturated fat, these

results can be considered in addition to the 5‐10% LDL‐C lowering expected from these diets alone53.

Whereas guidelines have largely focused on dietary pulses as part of a healthy dietary pattern in the

prevention and management of diabetes and CVD, these data support lipid‐lowering and cardiovascular

risk reduction recommendations. As dietary pulses may have beneficial effects on other cardiometabolic

risk factors including body weight, blood pressure, and glucose control87, future systematic reviews and


Page 41 of 93

meta‐analyses should evaluate the effects of dietary pulses on these endpoints and others to address

these factors that contribute to cardiovascular risk.

4.7 ACKNOWLEDGMENTS

We would like to thank Xavier Thobaut Gusmini and Júlio Medeiros Rego for their technical

support.

Aspects of this work was presented at the 2012 Experimental Biology Conference, San Diego, CA,

USA, April 21‐25 2012, 15th Annual Canadian Society for Endocrinology and Metabolism/Canadian

Diabetes Association (CSEM/CDA) Professional Conference and Annual Meetings, Toronto, ON, Canada,

October 10‐13 2012., and 53rd Annual Conference at the American College of Nutrition, Morristown, NJ,

USA, November 14‐17 2012.

4.8 SOURCE OF FUNDING

This work was funded by a Canadian Institutes of Health Research (CIHR) Knowledge Synthesis

Grant (Funding Reference Number: 119797). VH was supported by an Ontario Graduate Scholar (OGS)

award. RJD was funded by a Canadian Institutes for Health Research (CIHR) Postdoctoral Fellowship

Award. DJAJ was funded by the Government of Canada through the Canada Research Chair Endowment.

The funders had no role in study design, data collection and analysis, decision to publish, or preparation

of the manuscript.

4.9 CONFLICTS OF INTEREST/DISCLOSURE

JLS has received several unrestricted travel grants from The Coca‐Cola Company to present

research at meetings and is a co‐investigator on an unrestricted research grant from The Coca‐Cola

Company. JLS has also received travel funding and honoraria from Abbott Laboratories, Archer Daniels

Midland, and the International Life Sciences Institute (ILSI) North America and Brazil; and research

support, consultant fees, and travel funding from Pulse Canada. Spouse of JLS works for Unilever


Page 42 of 93

Canada. RJD, JB, and CWCK are co‐investigators on an unrestricted grant from The Coca‐Cola Company.

VV holds American (No. 7,326,404 B2) and Canadian (No. 2,410,556) patents for the use of viscous fibre

blend in diabetes, metabolic syndrome and cholesterol lowering. VV currently holds grant support for

ginseng research from the Canadian Diabetes Association, Canada and the National Institute of

Horticultural & Herbal Science, RDA, Korea. CWCK has served on the scientific advisory board, received

research support, travel funding, consultant fees, or honoraria from Pulse Canada, Barilla, Solae,

Unilever, Hain Celestial, Loblaws Inc., Oldways Preservation Trust, the Almond Board of California, the

International Nut Council, Paramount Farms, the California Strawberry Commission, the Canola and Flax

Councils of Canada, and Saskatchewan Pulse Growers. CWCK also receives partial salary funding from

research grants provided by Unilever, Loblaw’s, and the Almond Board of California. DJAJ has served on

the Scientific Advisory Board of Sanitarium Company, Agri‐Culture and Agri‐Food Canada (AAFC),

Canadian Agriculture Policy Institute (CAPI), California Strawberry Commission, Loblaw Supermarket,

Herbal Life International, Nutritional Fundamental for Health, Pacific Health Laboratories, Metagenics,

Bayer Consumer Care, Orafti, Dean Foods, Kellogg’s, Quaker Oats, Procter & Gamble, Coca‐Cola, NuVal

Griffin Hospital, Abbott, Pulse Canada, Saskatchewan Pulse Growers, and Canola Council of Canada;

received honoraria for scientific advice from Sanitarium Company, Orafti, the Almond Board of

California, the American Peanut Council, International Tree Nut Council Nutrition Research and

Education Foundation and the Peanut Institute, Herbal Life International, Pacific Health Laboratories,

Nutritional Fundamental for Health, Barilla, Metagenics, Bayer Consumer Care, Unilever Canada and

Netherlands, Solae, Oldways, Kellogg’s, Quaker Oats, Procter & Gamble, Coca‐Cola, NuVal Griffin

Hospital, Abbott, Canola Council of Canada, Dean Foods, California Strawberry Commission, Haine

Celestial, Pepsi, and Alpro Foundation; has been on the speakers panel for the Almond Board of

California; received research grants from Saskatchewan Pulse Growers, the Agricultural Bioproducts

Innovation Program (ABIP) through the Pulse Research Network (PURENet), Advanced Food Materials


Page 43 of 93

Network (AFMNet), Loblaw, Unilever, Barilla, Almond Board of California, Coca‐Cola, Solae, Haine

Celestial, Sanitarium Company, Orafti, International Tree Nut Council Nutrition Research and Education

Foundation and the Peanut Institute, the Canola and Flax Councils of Canada, Calorie Control Council,

Canadian Institutes of Health Research, Canada Foundation for Innovation, and the Ontario Research

Fund; and received travel support to meetings from the Solae, Sanitarium Company, Orafti, AFMNet,

Coca‐Cola, The Canola and Flax Councils of Canada, Oldways Preservation Trust, Kellogg’s, Quaker Oats,

Griffin Hospital, Abbott Laboratories, Dean Foods, the California Strawberry Commission, American

Peanut Council, Herbal Life International, Nutritional Fundamental for Health, Metagenics, Bayer

Consumer Care, AAFC, CAPI, Pepsi, Almond Board of California, Unilever, Alpro Foundation,

International Tree Nut Council, Barilla, Pulse Canada, and the Saskatchewan Pulse Growers. Dr Jenkins'

wife is a director, VV is the vice‐president and part owner, and LC is a clinical study coordinator of

Glycemic Index Laboratories, Toronto, Ontario, Canada. VH, VHJ, AA, SBM, MD, FMS, AMB, PKE, RPB,

RGJ, and LAL have no declared conflicts of interest related to this paper.


Page 44 of 93

4.10 TABLES

Table 1. Search Strategy for Studies Assessing the Effect of Dietary Pulses on Lipids in Randomized, Controlled Feeding Trials*. DATABASE SEARCH PERIOD SEARCH

MEDLINE 1948 to week 2 of February 2013

1. (Fabaceae or lentil$ or chickpea$ or bean$ or pea$ or legume$ or leguminous) and (lipid$ or cholesterol$ or apolipoprotein B or hyperlipidemia or lipaemia). mp

2. Limit to animals 3. 1 not 2 4. Limit to Clinical Trials, Clinical Trial, ALL 5. Limit to Clinical Trial 6. Limit to Controlled Clinical Trial 7. Limit to Randomized Controlled Trial

EMBASE Classic and EMBASE 1947 to Week 6 of 2013

1. (Fabaceae OR lentil$ OR chickpea$ OR bean$ OR pea$ OR legume$ OR leguminous) AND (lipid$ OR cholesterol$ OR apolipoprotein B OR hyperlipidemia OR lipaemia). mp

2. Limit to Animals and Animal Studies 3. 1 not 2 4. Limit to Clinical Trial 5. Limit to Randomized Controlled Trial 6. Limit to Controlled Clinical Trial

The Cochrane Library 1991 to February 13 2013

1. (Fabaceae OR lentil$ OR chickpea$ OR bean$ OR pea$ OR legume$ OR leguminous) AND (lipid$ OR cholesterol$ OR apolipoprotein B OR hyperlipidemia OR lipaemia). mp

CINAHL 1982 to February 13 2013

1. (lentil$ OR chickpea$ OR bean$ OR pea$ OR legume$ OR leguminous) AND (lipid$ OR VLDL OR apolipoprotein B OR hyperlipidemia OR lipaemia)

* The original search was conducted on May 9 2011 for MEDLINE, EMBASE, and CINAHL, and June 13 2011 for the Cochrane Library; updated searches for all databases were performed on December 2 2011, April 9, October 2 and 29, November 13 2012, and February 13 2013.


Page 45 of 93


STUDY PARTICIPANTS*AGE (SD

or Range), yBASELINE (SD)† CHD‐RISK

SCORE‡

LIPIDMEDICATION

USEDESIGN§ SETTING║ FEEDING

CONTROL¶ENERGY

BALANCE***PULSE DOSE

(grams/day)**

Abete et al. 2009 [90]18 Ob

(18M:0W)~37.1±8.0

3.01 (0.75), 4.33(0.72)‐

~3.53, ~4.91

2‐

No POP

SpainNon‐Metab

Negative (‐30% E)

~90

Abeysekara et al. 2012 [86] ║║ 87(30M:57W)

59.7±6.32.96 (0.86), 2.93 (0.84)

‐~3.29, ~3.39

33

No C OP

CanadaPartial Neutral 250

Anderson et al. 1984 [19]20 HC

(20M:0W)~54±8.4

4.92 (0.31), 5.72 (0.36) ‐

~6.50, ~7.04

5‐

No PIP

United StatesMetab Neutral 101

Anderson et al. 1990‐ I [85]6 HC

(6M:0W)64±2.4

‐‐‐

4‐

No POP

United StatesPartial Neutral ~113

Anderson et al. 1990‐ II [85]9 HC

(9M:0W)57±9

‐‐‐

4‐

No POP


Anderson et al. 1990‐ III [85]9 HC

(9M:0W)54 ± 9

‐‐‐

4‐

No POP


Belski et al. 2011 [20] ║║ 93 OW/Ob (52M:41W)

~46.6±103.33 (0.76), 3.29 (0.77)

‐~3.99, ~3.85

43

No POP

AustraliaPartial Negative (‐35% E) 50

Cobiac et al. 1990 [21]20 HC

(20M:0W)29‐65

4.69 (0.67)‐

~5.13

4‐

No C OP

AustraliaPartial Neutral ~377

Duane et al. 1997 [22] 9 N

(9M:0W)58 (41‐78)

3.58 (0.83), 3.23 (0.66)‐

4.45 (0.63), 4.24 (0.62)

5‐

‐ C IP

United StatesMetab Neutral ~251

Finley et al. 2007‐ N [23]40 N

(20M:20W)~37.4±11

‐‐

~2.88, ~2.98

32

No POP

United StatesPartial Neutral 130

Finley et al. 2007‐Pre‐MS [23]40 Pre‐MS(20M:20W)

~42.4±10‐‐

~3.66, ~3.66

32

No POP

United StatesPartial Neutral 130

Gormley et al. 1979 [103] 53 Nmajority of participants

30‐50

‐‐

~5.50, ~5.49

33

‐ POP

IrelandPartial Neutral ~59

Gravel et al. 2010 [24] ║║ 114 Pre‐MS(0M:114W)

~51.2±8.63.82 (0.87), 3.91 (0.83)1.09 (0.23), 1.11 (0.23)

~4.40, ~4.45

‐2

No POP

CanadaPartial Neutral ~81

Hermsdorff et al. 2011 [25]30 Ob

(17M:13W)36 ± 8

3.32 (0.60), 3.68 (1.06)‐

3.18, 4.30

33

No POP

SpainNon‐Metab Negative (‐30% E) ~198

Hodgeson et al. 2010 [30]74 OW/Ob(26M:48W)

~57.9±7.93.25 (0.91), 3.34 (0.92)

‐~3.85, ~3.96

32

Yes POP

AustraliaPartial Neutral ~64

Jenkins et al. 2012 [8]121 T2D

(61M:60W)~59.5±12.8

2.36 (1.01), 2.18 (0.75)‐

3.00 (1.03), 2.95 (0.96)

CHD RiskEquiv

Yes POP

CanadaNon‐Metab Neutral 196

Jimenez‐Cruz et al. 2004 [99]8 T2D

‐51±3

4.27 (0.70), 5.02 (0.06)‐

~4.98, ~5.53

CHD RiskEquiv

No C OPUSA

Non‐Metab Neutral _

Mackay et al. 1992 [26]39 HC

(22M:17W)~47 (28‐66)

4.19 (0.84)‐

~5.13

44

‐ C OP

New ZealandPartial Neutral 80

Marinangeli et al. 2011 [102] ║║ 23 OW/Ob, HC (7M:16W)

~52.0±10.63.81 (0.62), 3.78 (0.62)

‐~4.77, ~4.80

32

No C OP

CanadaMetab Neutral ~138

Pittaway et al. 2006 [27]47

(19M:28W)53 ± 9.8

‐‐‐

54

No C OP

AustraliaPartial Neutral 140

Pittaway et al. 2007 [28]27

(10M:17W)50.6 ± 10.5

‐‐‐

54

No C OP

AustraliaPartial Neutral 140

Shams et al. 2008 [29] 30 T2D 50.2±3.83.69 (0.44), 3.74 (0.34)

‐~4.76, ~4.70

CHD RiskEquiv

No C OPIran

Partial Neutral 50

Winham et al. 2007 [93] 23 HC

(10M:13W)45.9 ± 10.6

3.47 (0.37), 3.50 (0.50)‐

~4.01, ~4.09

11

No C OP


Winham et al. 2007 [94]16 Mildly IR(7M:9W)

43 ± 123.37 (0.62), B: 3.57 (0.83), P: 3.34 (0.72)

‐~4.00, B: ~4.25, P: ~4.01

21

No C OP


Zhang et al. 2010‐ IS [10]36 IS

(36M:0W)53.8 ± 7.6

3.13 (0.62), 3.21 (0.52)‐

~3.65, ~3.78

4‐

No C IP/OP


Zhang et al. 2010‐ IR [10]28 IR

(28M:0W)55.5 ± 8

3.34 (0.96), 3.32 (0.82)‐

~4.25, ~4.17

4‐

No C IP/OP


*Ob= Obese; OW= Overweight; HC= Hypercholesterolemia; T2D= Type 2 Diabetes; N= normal/healthy; Pre‐MS= Pre‐Metabolic Syndrome; IR= Insulin Resistant; IS= Insulin Sensitive; M= Men; W= Women; "‐"not reported; "~" calculated values †Values listed in the following order: LDL‐C (mmol/L), Apo‐B (g/L), non‐HDL‐C (mmol/L); Value for the control arm is reported first followed by value for the treatment arm; some endpoints combined baseline control and treatment arms together and therefore, reported only one value; For Winham et al.94, B denotes Bean Arm and P Pea Arm ‡Framingham 10‐year CHD Risk Factor Score Assessment4. For a more conservative approach, participants were assumed to be smokers, hypertensive, have low HDL‐C, and/or have family history of CHD if not reported. Risk score for men is reported first followed by risk score for women. §C= Crossover Design; P= Parallel Design; The study by Winham et al.94 had a crossover design with one control arm and 2 treatment arms (beans and peas). To mitigate unit‐of‐analysis error, we combined the 2 groups to create a single pair‐wise comparison, which we conservatively analyzed as a parallel trial for the overall analysis.


Page 46 of 93


PULSE FORM¶¶

TYPE OF

PULSE††COMPARATOR

DIET (CHO: PROTEIN: FAT % E)

FIBRE (g/d)‡‡

SATURATED FAT(%E)‡‡ FOLLOW‐UP MQS§§ FUNDING

Whole Mixed No Pulses 52:19:3220±1726±15

~5~5

8 weeks 7 Agency/Industry

Whole Mixed No Pulses 44‐49:15‐16:37‐3822±1030±15

1110


Whole Beans Oat‐Bran 43‐44:20:36‐3747±444±8

~14~15


Whole Beans No Pulses 43:19:381323

‐12



‐14



‐11


Flour Beans Wheat 39‐44:19‐22:30‐3225±739±12

‐‐

1 year 8 Agency

Whole Beans Spaghetti 46‐47:17‐18:3211±222±2

~13~14

4 weeks 8 Industry

‐ Mixed No Pulses 60:17:231924

‐‐

6‐7 weeks 8 Agency

Whole Beans Chicken Soup 47‐51:14‐16:32‐35~18±9~18±4

‐‐

12 weeks 6 Agency

Whole Beans Chicken Soup 41‐48:15‐17:36‐41~18±7~17±4

‐‐

12 weeks 6 Agency

Whole Peas Cornflakes ‐ ‐ ‐ 6 weeks 3 ‐

‐ Mixed No Pulses 48‐49:17‐18:3318±923±10

~11~11

16 weeks 6 Industry

Whole Mixed No Pulses 50‐51:19:31‐3318±5 26±6

5±0.95±0.3

8 weeks 8 Agency

Flour Beans White Bread 39‐40:20‐24:31‐3424±936±10

11±3.812±3.4

16 weeks 7 Agency

Whole Mixed High Fiber 46‐49:22‐23:29‐31~27~38

9±3.38±2.8


Whole Beans High GI 51‐54:26‐27:18‐233053

99

3 weeks 9 Agency

Whole/Flour Beans Low‐fiber 53‐54:16‐17:2827±629±7

10±39±3

6 weeks 5 Agency

Flour Peas White Flour 55:15:301919

88

4 weeks 6 Industry

Whole/Flour Chickpea Whole Wheat 44‐46:18‐19:32‐3328±14.930±9.8

~13~12

~5.5 (5‐6) weeks

5 Agency

Whole/Flour Chickpea Whole Wheat 43‐44:17‐18:3429±10.828±7.5

~14~14

5 weeks 5 Agency

Whole Lentils No Pulses 48‐52:18‐19:29‐3123±629±3

‐‐

6 weeks 7 ‐

Whole Beans Carrots 50:17:31‐3421±826±8

~11~10

8 weeks 6 Industry

WholeBeansPeas

Carrots 50‐53:15‐16:31‐3221±821±2

911

8 weeks 7 Agency

Whole Beans Chicken 50‐51:18:34~18~42

11±112±1

4 weeks 3 Agency

Whole Beans Chicken 50‐51:18:34~18~42

11±112±1

4 weeks 3 Agency

║IP= inpatient; OP= outpatient ¶ metab denotes all foods were provided; partial if test food or some meals were provided; and non‐metab if no foods were provided **based on cooked weight; dry weight was converted to wet weight by multiplying 2.75 ††Mixed refers to more than one type of dietary pulses being studied ‡‡ Value for the control arm is reported first followed by value for the treatment arm §§ MQS= Heyland Methodological Score; Score of ≥8 denotes higher quality100 ║║For Abeyesekara et al. analysis was for n= 8086; Belski et al. age is based on n=13120; Gravel et al. n=13224; Marinangeli et al. n=29102 ¶¶Dietary Pulses were provided in 1 of 2 forms: 1) Whole, where intact dietary pulses were consumed; 2) Flour, where dietary pulses were grinded to a powder form and was incorporated into baked foods ***Neutral denotes weight‐maintaining diet; Negative weight‐reduction diet (% energy restricted from baseline diet reported in brackets)


Page 47 of 93

Table 3. Study Quality Assessment by Using the Heyland MQS100.

MQSRandomization

(n/2)Blinding

(n/1)Analysis

(n/2)Selection

(n/1)Compatability

(n/1)Follow-up

(n/1)Protocol

(n/1)Co-interventions

(n/2)Crossovers

(n/2) (n/13)

Gormley et al. 1979 [103] 1 0 0 0 1 0 0 1 0 3Anderson et al. 1984 [19] 1 0 2 0 1 1 1 2 0 8Anderson et al. 1990 [85] 1 0 0 1 1 0 1 2 0 6Cobiac et al. 1990 [21] 1 0 2 0 1 1 1 2 0 8Mackay et al. 1992 [26] 1 0 0 0 1 0 1 2 0 5Duane et al. 1997 [22] 1 0 2 0 1 1 1 2 0 8Jimenez-Cruz et al. 2004 [99] 1 0 2 1 1 1 1 2 0 9Pittaway et al. 2006 [27] 1 0 0 0 1 0 1 2 0 5Pittaway et al. 2007 [28] 1 0 0 0 1 0 1 2 0 5Finley et al. 2007 [23] 1 0 0 1 1 0 1 2 0 6Winham et al. 2007 [93] 2 0 0 1 1 0 1 2 0 7Winham et al. 2007- COM [94] 1 0 0 1 1 0 1 2 0 6Shams et al. 2008 [29] 1 0 2 0 1 1 0 2 0 7Abete et al. 2009 [90] 1 0 2 1 1 1 0 1 0 7Gravel et al. 2010 [24] 2 0 0 1 1 0 0 2 0 6Hodgeson et al. 2010 [30] 2 0 0 1 1 0 1 2 0 7Zhang et al. 2010 [10] 1 0 0 0 1 0 1 2 0 5Belski et al. 2011 [20] 2 1 0 1 1 0 1 2 0 8Hermsdorff et al. 2011 [25] 1 0 2 0 1 1 1 2 0 8Marinangeli et al. 2011 [102] 1 0 0 1 1 0 1 2 0 6Abeysekara et al. 2012 [86] 2 0 0 1 1 0 1 1 0 6Jenkins et al. 2012 [8] 1 0 2 1 1 0 1 2 0 8

StudyMethods† Sample‡ Intervention§

*The Heyland MQS assigns a score of 0 or 1 or from 0 to 2 over 9 categories of quality related to study design, sampling procedures, and interventions, for a total of 13 points. Trials that scored ≥8 were considered to be of higher quality100.

†Randomization was scored 2 points for being randomized with the methods described, 1 point for being randomized without the methods described, or 0 points for being neither randomized nor having the methods described. Blinding was scored 1 point for being double‐blind or 0 points for “other.” Analysis was scored 2 points for being intention‐to‐treat; all other types of analyses scored 0 points.

‡ Sample selection was scored 1 point for being consecutive eligible or 0 points for being preselected or indeterminate. Sample comparability was scored 1 point for being comparable or 0 points for not being comparable at baseline. Follow‐up was scored 1 point for being 100% or 0 points for <100%.

§ Treatment protocol was scored 1 point for being reproducibly described or 0 points for being poorly described. Co‐interventions were scored 2 points for being described and equal, 1 point for being described but unequal or indeterminate, or 0 points for not being described. Treatment crossovers (where participants were switched from the control treatment to the experimental treatment) were scored 2 points for being ≥10%, 1 point for being <10%, and 0 points for not being described.


Page 48 of 93

Table 4. Risk of Bias Assessment by Using the Cochrane Risk of Bias Tool*.

Random sequence generation (selection bias)

Allocation concealment (selection bias)

Blinding of participants and personnel (performance bias)

Incomplete outcome data (attrition bias)

Selective reporting (reporting bias)

Other bias

0% 25% 50% 75% 100%

Low risk of bias Unclear risk of bias High risk of bias

*Studies were rated UR if insufficient information was given to assess risk; LR if the study design is likely to have little influence over the true outcome; HR if the study design is likely to have an influential effect on the true outcome. †Random Sequence Generation assessed whether the method of randomization was described. ‡Allocation concealment assessed whether investigators could tell which treatment participants were going to be randomized to. §Blinding of participants and personnel assessed whether the study was blinded to investigators and/or participants. ||Incomplete outcome data assessed whether missing outcome data effected true outcome. ¶Selected outcome reporting assessed whether investigators pre‐registered trial and/or specified primary and secondary outcomes. ** Other bias assessed other sources of bias that are not listed in the Cochrane Risk of Bias Tool. Only Abeyesekara et al.86 is listed as the dietary pulse arm was given both food and dietary advice throughout the study, but the control arm was simply told to keep following their usual dietary habits.

†

‡

§

¶

Random sequence generation (selection bias)†

Allocation concealment (selection bias)‡

Blinding of participants and personnel (performance bias)§

Incomplete outcome data (attrition bias)||

Selective reporting (reporting bias)¶

Other bias**


Page 49 of 93

Table 5. Subgroup analyses by a priori dichotomous categorization for LDL‐C and non‐HDL‐C.

Subgroup Range No. of Trials N β (95% CI) Residual I2 p‐valueLDL‐CA Priori SubgroupBaseline LDL 2.27‐5.32 mmol/L 17 773 ‐0.10 (‐0.42, 0.21) 83.6 0.482Dose 50‐ 377 grams/day 24 949 0.00 (0.00, 0.00) 65.4 0.316Difference in Fiber Intake* ‐3‐ 24 grams/day 25 992 ‐0.01 (‐0.02, 0.01) 80.4 0.256Difference in Saturated Fat Intake† ‐1‐ 2 % Energy 17 724 ‐0.01 (‐0.22, 0.20) 78.8 0.920Duration 3 weeks‐ 1 year 25 960 0.01 (0.00, 0.02) 80.3 0.240Post‐hoc SubgroupPercentage of Males 0‐ 100% 23 946 ‐0.004 (‐0.01, ‐0.00) 53.1 0.012

Non‐HDL‐CA Priori SubgroupBaseline non‐HDL‐C 2.99‐ 6.70 mmol/L 16 762 0.01 (‐0.31, 0.28) 98.8 0.929Dose 50‐377 grams/day 21 861 0.00 (0.00, 0.00) 98.1 0.519Difference in Fiber Intake* ‐3‐ 23 grams/day 21 816 ‐0.03 (‐0.06, 0.00) 98.3 0.024Difference in Saturated Fat Intake† ‐1‐ 2 % Energy 15 660 0.05 (‐0.23, 0.34) 97.9 0.706Duration 3 weeks‐ 1 year 22 869 0.01 (‐0.01, 0.03) 97.6 0.289Post‐hoc SubgroupPercentage of Males 0‐ 100% 19 783 0.00 (0.00, 0.01) 98.1 0.446

Abbreviations: N denotes the number of participants in each subgroup; residual I2 was reported as a percent value and significance as a P‐value with p < 0.01 as significant.

*Difference in fibre intake compared fibre intake of the dietary pulse arm with control at the end of the study. If fibre intake at the end of the study was not provided, intended fibre intake (as specified in the study protocol) for each arm was used for calculation.

† Difference in saturated fat intake compared saturated fat intake of the dietary pulse with the control at the end of the study. If saturated fat intake at the end of the study was not provided, intended saturated fat intake (as specified in the study protocol) for each arm was used for calculation.


Page 50 of 93

4.11 FIGURES

Figure 1. Flow of the literature search.

61 reports reviewed in full

22 reports (26 trials) included in the meta‐analysis (n= 1013)

39 reports excluded 1 review paper 3 studies with no dietary pulse intervention 2 studies with dietary pulse extracts 7 studies with unsuitable endpoints 1 hypocaloric study 3 acute or short‐term studies 12 co‐intervention trials where dietary pulses were not emphasized 1 study with soy in control group 4 studies were not randomized 2 studies lacked control group 1 study reported in unusable statistics 1 cannot be retrieved 1 provided incorrect data and authors could not be

contacted

3001 reports identified 1006 MEDLINE (through Feb 11 2013) 1060 EMBASE (through Feb 11 2013) 881 Cochrane Library (through Feb 11 2013) 35 CINAHL (through Feb 11 2013)

2940 reports excluded on basis of title and/or abstract

1266 duplicate reports 14 animal or in vitro studies 1 book chapter 5 commentaries or editorials 82 case reports 151 review papers 256 observational studies 1112 studies with no dietary pulse intervention 8 studies with dietary pulse extracts 10 studies with unsuitable endpoints 8 acute or short‐term studies 27 co‐intervention trials where dietary pulses

were not emphasized


Page 51 of 93

84909090909297040607070707070809101010111111111212

-2 -1 0 1 2Favours Dietary Pulses Favours Control

Study Pulse, n Control, nMean Difference (95% CI)

in LDL‐C (mmol/L) Anderson et al. 1984 [19] 10 10 ‐0.25 [‐0.86, 0.36]Anderson 1990‐ III [85] 9 9 ‐0.76 [‐2.15, 0.63]Anderson 1990‐ I [85] 6 6 ‐0.17 [‐1.89, 1.55]Cobiac et al. 1990 [21] 20 20 ‐0.02 [‐0.27, 0.23]Anderson 1990‐ II [85] 9 9 ‐0.43 [‐1.61, 0.75]Mackay et al. 1992 [26] 39 39 0.05 [‐0.15, 0.25]Duane et al. 1997 [22] 9 9 ‐0.31 [‐0.56, ‐0.06]Jimenez‐Cruz 2004 [99] 8 8 ‐1.77 [‐2.42, ‐1.12]Pittaway et al. 2006 [27] 47 47 ‐0.18 [‐0.28, ‐0.08]Winham et al.2007‐COM[94] 23 23 ‐0.16 [‐0.34, 0.02]Pittaway et al. 2007 [28] 27 27 ‐0.20 [‐0.36, ‐0.04]Winham et al. 2007 [93] 16 16 ‐0.13 [‐0.31, 0.05]Finley et al 2007‐ N [23] 20 20 ‐0.17 [‐0.31, ‐0.03]Finley 2007‐ Pre‐MS [23] 20 20 ‐0.22 [‐0.38, ‐0.06]Shams et al. 2008 [29] 30 30 0.02 [‐0.02, 0.06]Abete et al. 2009 [90] 8 10 ‐0.88 [‐1.17, ‐0.59]Gravel et al. 2010 [24] 60 54 0.15 [‐0.07, 0.37]Zhang et al. 2010‐IR [10] 36 36 ‐0.21 [‐0.41, ‐0.01]Zhang et al. 2010‐ IS 28 28 ‐0.26 [‐0.46, ‐0.06]Marinangeli 2011 [102] 23 23 0.13 [‐0.18, 0.44]Hodgson et al. 2011[30] 37 37 ‐0.03 [‐0.28, 0.22]Hermsdorff et al.2011[25] 15 15 ‐0.34 [‐0.58, ‐0.10]Belski et al. 2011[20] 46 47 0.03 [‐0.13, 0.19]Abeysekara et al.2012[86] 80 80 ‐0.23 [‐0.43, ‐0.03]Jenkins et al. 2012 [8] 60 61 ‐0.06 [‐0.20, 0.08]

Total (95% CI) 686 684 ‐0.17 [‐0.25, ‐0.09]Heterogeneity: Tau² = 0.03; Chi² = 117.42, df = 24 (P < 0.00001); I² = 80%Test for overall effect: Z = 4.03 (P < 0.0001)

Figure 2. Forest plot of trials investigating the effect of isocaloric exchange of dietary pulses for control diets on LDL‐C. A pooled effect estimate is shown as a diamond. Paired analyses were applied to all crossover trials. Data are mean differences (MD) with 95% CI, where p‐values are for Generic Inverse Variance random effects models. Inter‐study heterogeneity was tested by Cochrane’s Q (I2) at a significance level of P<0.10 and quantified by I2, where I2 ≥ 50 % is considered to be evidence of substantial heterogeneity and ≥75%, considerable heterogeneity.


Page 52 of 93

Mean Difference (95% CI) in LDL‐C, mmol/L

Favours ControlFavours Dietary Pulses

Between SubgroupsWithin SubgroupsResidual I2 P ValueSubgroup Level

Total

Trials Participants

25 960 ‐0.17 (‐0.25 to ‐0.09)

11Baseline24

<3.37 mmol/L

≥3.37 mmol/L

6 432

341

‐0.13 (‐0.46 to 0.20)

‐0.22 (‐0.48 to 0.04)

Type

Beans (1)

Chickpeas (2)

1

15 461

30

‐0.21 (‐0.39 to ‐0.03)

0.02 (‐0.54 to 0.58)

Dose

Crossover

<100 grams/day

12

MQS

83.1 0.643‐0.09 (‐0.51 to 0.32)

80.0 <0.00001

2 74 ‐0.19 (‐0.60 to 0.22)

Lentils (3)

Peas (4) 2 39 0.10 (‐0.34 to 0.54) Mixed (5) 6 379 ‐0.26 (‐0.51 to ‐0.01) 0.58976.5See Figure Legend

≥100 grams/day

8 407

16 542

‐0.08 (‐0.22 to 0.05)

‐0.18 (‐0.29 to ‐0.08) ‐0.10 (‐0.27 to 0.07) 60.8 0.239

DesignParallel

386

‐0.19 (‐0.39 to 0.01) 13 574

‐0.18 (‐0.34 to ‐0.01)

<8

≥8

19 659 ‐0.16 (‐0.30 to ‐0.01)

6 301 ‐0.25 (‐0.49 to 0.00)

0.01 (‐0.25 to 0.27)

‐0.09 (‐0.38 to 0.19)

79.4 0.925

80.1 0.511

-1.25 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25

Saturated Fat Intake ≥7% E

<7% E67615 ‐0.14 (‐0.29 to 0.02)

‐0.59 (‐1.03 to ‐0.14) 482

0.45 (‐0.03 to 0.93) 71.5 0.062

Dietary FibreIntake

<28 grams/day 13 392 ‐0.20 (‐0.38 to ‐0.01) ≥28 grams/day 12 610 ‐0.16 (‐0.34 to 0.01)

0.03 (‐0.22 to 0.29) 77.5 0.781

Figure 3. Forest plots of subgroup analyses for dichotomous variables investigating the effect of isocaloric exchange of dietary pulses for other adequate comparators on LDL‐C in all participants. Point estimates for each subgroup level (diamonds) are the pooled effect estimates. The dashed line represents the pooled effect estimate for the overall (total) analysis. The residual I2 value indicates the inter‐study heterogeneity unexplained by the subgroup. Pair‐wise between‐subgroup mean differences with 95% CI (mmol/L) for comparator were as follows: (1 vs. 2) ‐0.23 (‐0.78, 0.31), (1 vs. 3) ‐0.44 (‐1.11, 0.22), (1 vs. 4) ‐0.52 (‐1.09, 0.04), (1 vs. 5) ‐0.16 (‐0.60, 0.28), (2 vs. 1) 0.23 (‐0.31, 0.78), (2 vs. 3) ‐0.21 (‐0.90, 0.48), (2 vs. 4) ‐0.29 (‐0.89, 0.31), (2 vs. 5) 0.07 (‐0.41, 0.55), (3 vs. 1) ‐0.07 (‐0.41, 0.55), (3 vs. 2) 0.44 (‐0.22, 1.11), (3 vs. 4) ‐0.08(‐0.79, 0.63), (3 vs. 5) 0.28 (‐0.33, 0.89), (4 vs. 1) 0.52 (‐0.04, 1.09), (4 vs. 2) 0.29 (‐0.31, 0.89), (4 vs. 3) 0.08 (‐0.63, 0.79), (4 vs. 5) 0.36 (‐0.15, 0.87), (5 vs. 1) 0.16 (‐0.28, 0.60), (5 vs. 2) ‐0.07 (‐0.55, 0.41), (5 vs. 3) ‐0.28 (‐0.89, 0.33), (5 vs. 4) ‐0.36 (‐0.87, 0.15).


Page 53 of 93

Study Pulses, n Control, nMean Difference (95% CI) in non‐HDL‐C (mmol/L)

Gormley et al. 1979 [103] 23 25 ‐0.50 [‐0.60, ‐0.40]Anderson et al. 1984 [19] 10 10 0.48 [0.28, 0.68]Cobiac et al. 1990 [21] 20 20 0.01 [‐0.07, 0.09]Anderson 1990‐ II [85] 6 6 ‐0.99 [‐1.87, ‐0.11]Anderson 1990‐ III [85] 9 9 ‐0.60 [‐1.93, 0.73]Anderson 1990‐ I [85] 9 9 ‐0.60 [‐2.78, 1.58]Mackay et al. 1992 [26] 39 39 0.01 [‐0.03, 0.05]Duane et al. 1997 [22] 9 9 ‐0.17 [‐0.31, ‐0.03]Jimenez‐Cruz 2004 [99] 8 8 ‐1.47 [‐1.78, ‐1.16]Pittaway et al. 2006 [27] 47 47 ‐0.21 [‐0.27, ‐0.15]Winham et al.2007‐COM[94] 23 23 ‐0.13 [‐0.21, ‐0.05]Winham et al. 2007 [93] 16 16 ‐0.26 [‐0.32, ‐0.20]Pittaway et al. 2007 [28] 27 27 ‐0.22 [‐0.32, ‐0.12]Shams et al. 2008 [29] 30 30 ‐0.34 [‐0.38, ‐0.30]Abete et al. 2009 [90] 8 10 0.49 [0.31, 0.67]Gravel et al. 2010 [24] 60 54 0.19 [0.15, 0.23]Hodgson et al. 2011[30] 37 37 0.01 [‐0.03, 0.05]Hermsdorff et al.2011[25] 15 15 0.36 [0.26, 0.46]Marinangeli 2011 [102] 23 23 0.06 [0.00, 0.12]Belski et al. 2011[20] 46 47 0.12 [0.10, 0.14]Jenkins et al. 2012 [8] 60 61 ‐0.11 [‐0.27, 0.05]Abeysekara et al.2012[86] 80 80 ‐0.21 [‐0.23, ‐0.19]

Total (95% CI) 605 605 ‐0.09 [‐0.19, 0.00]Heterogeneity: Tau² = 0.04; Chi² = 1325.60, df = 21 (P < 0.00001); I² = 98%Test for overall effect: Z = 1.92 (P = 0.06)

ear979984990990990990992997004006007007007008009010011011011011012012

IV, Random, 95% CI

-2 -1 0 1 2Favours Dietary Pulses Favours Control

Figure 4. Forest plot of trials investigating the effect of isocaloric exchange of dietary pulses for control diets on non‐HDL‐C. A pooled effect estimate is shown as a diamond. Paired analyses were applied to all crossover trials. Data are mean differences (MD) with 95% CI, where p‐values are for Generic Inverse Variance random effects models. Inter‐study heterogeneity was tested by Cochrane’s Q (I2) at a significance level of P<0.10 and quantified by I2, where I2 ≥ 50 % is considered to be evidence of substantial heterogeneity and ≥75%, considerable heterogeneity.


Page 54 of 93

-1.25 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 1.25

Mean Difference (95% CI) in non‐HDL‐C, mmol/L

Favours ControlFavours Dietary Pulses

Between SubgroupsWithin SubgroupsResidual I2 P ValueSubgroup Level

Total

Trials Participants

22 869 ‐0.09 (‐0.19 to 0.00)

10Baseline24<4.14 mmol/L

≥4.14 mmol/L

6 421

341

‐0.02 (‐0.42 to 0.39)

‐0.11 (‐0.42 to 0.20)

Type

Beans (1)

Chickpeas (2)

1

11 317

30

‐0.21 (‐0.51 to 0.08)

‐0.34 (‐1.20 to 0.52)

Dose

Crossover

<100 grams/day

11

MQS

98.9 0.698‐0.09 (‐0.60 to 0.42)

98.0 0.060

2 74 ‐0.21 (‐0.83 to 0.40)

Lentils (3)

Peas (4) 3 92 ‐0.16 (‐0.66 to 0.34) Mixed (5) 6 379 0.09 (‐0.26 to 0.44) 0.66697.9See Figure Legend

≥100 grams/day

9 460

12 401

‐0.05 (‐0.25 to 0.15)

‐0.05 (‐0.24 to 0.15) 0.00 (‐0.28 to 0.28) 98.1 0.992

DesignParallel

322

0.05 (‐0.22 to 0.32) 11 454

‐0.25 (‐0.49 to ‐0.01)

<8

≥8

15 568 ‐0.14 (‐0.38 to 0.10)

7 301 ‐0.09 (‐0.42 to 0.24)

‐0.30 (‐0.66 to 0.06)

0.05 (‐0.36 to 0.46)

96.6 0.103

97.7 0.799

Saturated Fat Intake ≥7% E

<7% E61213 ‐0.13 (‐0.38 to 0.11)

0.42 (‐0.20 to ‐1.05) 482

‐0.56 (‐1.23 to 0.12) 97.8 0.099

Dietary FibreIntake

<28 grams/day 11 277 ‐0.01 (‐0.30 to 0.27) ≥28 grams/day 10 539 ‐0.18 (‐0.45 to 0.09)

‐0.17 (‐0.56 to 0.23) 98.4 0.387

Figure 5. Forest plots of subgroup analyses for dichotomous variables investigating the effect of isocaloric exchange of dietary pulses for other adequate comparators on non‐HDL‐C in all participants. Point estimates for each subgroup level (diamonds) are the pooled effect estimates. The dashed line represents the pooled effect estimate for the overall (total) analysis. The residual I2 value indicates the inter‐study heterogeneity unexplained by the subgroup. Pair‐wise between‐subgroup mean differences with 95% CI (mmol/L) for comparator were as follows: : (1 vs. 2) ‐0.21 (‐1.06, 0.64), (1 vs. 3) ‐0.09 (‐1.12, 0.96), (1 vs. 4) ‐0.27 (‐1.04, 0.50), (1 vs. 5) ‐0.52 (‐1.20, 0.17), (2 vs. 1) 0.21 (‐0.64, 1.06), (2 vs. 3) 0.12 (‐0.93, 1.18), (2 vs. 4) ‐0.06 (‐0.85, 0.73), (2 vs. 5) ‐0.30 (‐1.01, 0.40), (3 vs. 1) 0.09 (‐0.96, 1.13), (3 vs. 2) ‐0.12 (‐1.18, 0.93), (3 vs. 4) ‐0.18 (‐1.18, 0.81), (3 vs. 5) ‐0.43 (‐1.36, 0.50), (4 vs. 1) 0.27 (‐0.50, 1.04), (4 vs. 2) 0.06 (‐0.73, 0.85), (4 vs. 3) 0.18 (‐0.81, 1.18), (4 vs. 5) ‐0.25 (‐0.86, 0.37), (5 vs. 1) 0.52 (‐0.17, 1.20), (5 vs. 2) 0.30 (‐0.40, 1.01), (5 vs. 3) 0.43 (‐0.50, 1.36), (5 vs. 4) 0.25 (‐0.37, 0.86).


Page 55 of 93

Study Pulses, n Control, nMean Difference (95% CI)

in Apo‐B (g/L)Gravel et al. 2010 [24] 60 54 0.02 [‐0.04, 0.08]

Test for overall effect: Z = 0.67 (P = 0.50)

CI8]

8]

IV, Random, 95% CI

-0.1 -0.05 0 0.05 0.1Favours Dietary Pulses Favours Other Diet

Figure 6. Forest plot of a trial investigating the effect of isocaloric exchange of dietary pulses for control diet on Apo‐B. An effect estimate is shown as a diamond. The datum is mean difference (MD) with 95% CI, where the P value is for Generic Inverse Variance random effects model.


Page 56 of 93

A.

B.

Figure 7. Funnel plot investigating publication bias and effect of small study effects in clinical trials with isocaloric exchange of dietary pulses for other adequate comparators on LDL‐C (7A) and non‐HDL‐C (7B) in all participants. The solid line represents the pooled effect estimate expressed as the weighted mean difference for each analysis. Dashed lines represent pseudo‐95% confidence limits.

Egger’s P‐value= 0.003 Begg’s P‐value= 0.888

Egger’s P‐value= 0.676 Begg’s P‐value= 1.000

DIETARY PULSES AND OXIDATIVE STRESS

Page 57 of 93

CHAPTER V‐ EFFECT OF A HIGH DIETARY PULSES, LOW GLYCEMIC INDEX DIET ON OXIDATIVE STRESS MARKERS IN TYPE 2 DIABETES


Page 58 of 93

EFFECT OF A HIGH DIETARY PULSES, LOW GLYCEMIC INDEX DIET ON OXIDATIVE STRESS MARKERS IN

TYPE 2 DIABETES

Vanessa Ha HBSc1,4, John L Sievenpiper MD PhD1,8, Russell J de Souza RD ScD1,9, Arash Mirrahimi1,11,

Richard P Bazinet PhD4, Robert G Josse MBBS1,2,3,4,5, Cyril WC Kendall PhD1,4,10, Joseph Beyene6,7,9,

David JA Jenkins MD PhD DSc1,2,3,4,5,

1Clinical Nutrition and Risk Factor Modification Centre, 2Keenan Research Centre of the Li Ka Shing

Knowledge Institute, and 3Division of Endocrinology, St. Michael’s Hospital, Toronto, ON, CANADA; 4Departments of Nutritional Sciences, and 5Medicine, 6Dalla Lana School of Public Health, Faculty of

Medicine, University of Toronto, Toronto, ON, CANADA; 7The Hospital for Sick Children Research

Institute, Toronto, ON, CANADA; 8Departments of Pathology and Molecular Medicine, 9Clinical

Epidemiology & Biostatistics, McMaster University, Hamilton, ON, CANADA; 10College of Pharmacy and

Nutrition, University of Saskatchewan, Saskatoon, SK, CANADA; 11School of Medicine, Faculty of

Medicine, Queen’s University, Kingston, ON, CANADA

Corresponding author:

David JA Jenkins MD PhD DSc University Professor Canada Research Chair in Nutrition and Metabolism University of Toronto 150 College St., Toronto, Ontario, M5S 3M2, CANADA Tel: 416 867 7475 Fax: 416 867 7495 Email: [email protected]


Page 59 of 93

5.1 ABSTRACT

Background: Diets high in vegetable protein such as dietary pulses (beans, chickpeas, lentils, and peas)

have been inversely associated with coronary heart disease (CHD). Whether dietary pulses reduce CHD

by reducing oxidative stress is unclear. We conducted a secondary analysis of a randomized controlled

clinical trial to assess the effect of a diet emphasizing dietary pulses on oxidative damage in type 2

diabetes (T2DM).

Methods and Results: This secondary analysis compared 113 T2DM participants who had serum

available and completed either 12 week dietary intervention trial high in dietary pulses, which was used

to lower the glycemic index of the diet, or high in wheat fibre diets in a randomized parallel design.

Markers of oxidative stress were measured using thiobarbituric acid reactive substances (TBARS) and

conjugated dienes (CDs) in the LDL fraction, and protein thiols in serum. There were no treatment

differences in oxidative stress markers, although pooling data from both treatments showed a

significant inverse correlation between changes in dietary pulse intake with changes in CDs (r=‐0.225;

p=0.016), suggesting a higher dietary pulse intake is associated with improved oxidative stress status.

Estimated glucose excursions, however, were not significantly associated with markers of oxidative

damage.

Conclusions: This secondary analysis of a randomized controlled trial did not detect treatment

differences in markers of oxidative damage although dietary pulse intake tended to be inversely related

to one marker of oxidative stress. Appropriately powered intervention trials of dietary pulses on

oxidative damage as the primary outcome are warranted.

Clinical Trial Registration: NCT01063361

Abstract Word Count: 237

Keywords: clinical trial, legumes, dietary pulses, oxidative stress, oxidative damage, cardiovascular

disease


Page 60 of 93

5.2 INTRODUCTION

Diets emphasizing vegetable proteins have been shown to reduce coronary heart disease (CHD)

mortality risk when substituting for carbohydrates or animal protein in the Iowa Women’s Study17.

Similarly, human dietary intervention trials have shown high plant protein intake that replaced

carbohydrate improve lipid levels31, 72. These findings are suggestive that foods that are high in vegetable

proteins may offer cardio‐protective effects. However other than soy proteins64, and a study on wheat

gluten31 and barley71, few studies have explored other food sources of plant protein, or assessed the

mechanism by which vegetable protein may have to reduce cardiovascular risk.

Dietary non‐oil‐seed pulses (beans, peas, chickpeas, and lentils) are a good source of vegetable

protein and they have been shown to reduce CHD risk factors, including LDL‐C and systolic blood

pressure (SBP) 8, 25. One cup of dietary pulses, on average, can provide 16.3 grams of the 50 grams of

daily protein intake as recommended by United States Department of Agriculture (USDA)107. Similar to

soy, dietary pulses contain the 7S globulin peptide which has been shown to increase uptake and

degradation of LDL‐C in human hepatic cells18. However, the effect of dietary pulses on ox‐LDL‐C, an

emerging risk factor of CVD, is unclear. Dietary patterns such as Dietary Strategies to Stop Hypertension

(DASH)108, low glycemic index (GI)109, and Mediterranean11 diets, which may contain dietary pulses, have

been shown to reduce markers of oxidative stress. We have therefore assessed the effect of a low‐GI

diet specifically emphasizing dietary pulses on oxidative stress markers in a secondary analysis of

participant serum from a trial of pulse consumption in type 2 diabetes (T2DM).

5.3 METHODS

This trial has been registered at clinicaltrials.gov (number, NCT01063361). Details of the

methods and results of the original study of 121 participants have been reported previously8.


Page 61 of 93

5.3.1 PARTICIPANTS

Participants were recruited from newspaper and public transport advertisements as well as

hospital clinics. A total of 131 eligible participants were randomized (Figure 1). Eligible participants had

an existing diagnosis of T2DM for at least 6 months, were taking a stable dose of oral anti‐hyperglycemic

agents for at least the previous 2 months, and had hemoglobin A1c (HbA1c) values that were between

6.5% to 8.5% of total hemoglobin both at initial screening and at the visit 1 week prior to commencing

the study. No participants had clinically significant cardiovascular, renal, or liver disease or a history of

cancer. The present analysis focuses on the 113 participants who completed the study, and had serum

samples available for analysis for oxidative damage markers at weeks 8, 10, or 12.

5.3.2 PROTOCOL

The study followed a randomized, parallel design with 2 treatment arms of 3 months’ duration

consisting of 1) a low‐GI diet emphasizing dietary pulse consumption (high dietary pulse diet), and 2) a

high wheat fibre diet (high fibre diet). Eligible participants were randomized by a statistician based on

sex and HbA1c value (HbA1c value ≤7.1% or >7.1% of total hemoglobin). Neither the dietitians nor the

participants were blinded, but equal emphasis was placed on healthiness of both diets.

Participants attended the research centre for screening and at week −1, baseline (week 0), and

weeks 2, 4, 8, 10, and 12 of the study. Participants were considered as completers of the study if they

made their last visit at week 8, 10, or 12. At each visit, anthropometric outcomes (body weight and waist

circumference) were measured and a fasting blood sample was taken. If HbA1c values rose above 8.5%

of total hemoglobin on 2 successive occasions, and could not be explained by specific circumstances,

participants were withdrawn from the study and referred back to their physician.

The study was approved by the research ethics board of St. Michael’s Hospital and the

University of Toronto, Toronto, Ontario, Canada. Written consent was obtained from all participants.


Page 62 of 93

5.3.3 DIETARY INTERVENTION

The emphasis of the treatment diet was to consume low GI foods as achieved primarily through

increased consumption of dietary pulses. The target dietary pulse consumption was 1 cup per day

(approximately 190 g per day, or 2 servings per day) of cooked beans, chickpeas or lentils. While a high

fibre diet was achieved by consumption of whole wheat and whole grain carbohydrate foods (whole

wheat breakfast cereals, breads, brown rice, etc). Participants were provided with a checklist of 15‐

grams carbohydrate portions of recommended foods and the quantities they were expected to consume

daily. Seven‐day food records covering the week prior to each visit were discussed with the dietitian and

assessed for adherence to prescribed dietary interventions.

5.3.4 BIOCHEMICAL ANALYSIS

Fasting HbA1c, glucose, and lipids were measured a fresh sample in the hospital routine clinical

laboratory. Measures of oxidative damage were made in batches at the end of the study on serum from

fasting blood samples stored at ‐70°C. Oxidative stress was measured as thiobarbituric acid reactive

substances (TBARS), (CV=5.0%) and conjugated dienes (CDs) (CV=3.6%) in the LDL‐C fraction by the

methods of Jentzsch et al.110 and Ahotupa et al.111, respectively. Total protein thiols were also measured

on stored serum as a marker of oxidized plasma proteins by the method of Hu et al.112 (CV=2.3%).

Diets were assessed for macronutrients, fatty acids, cholesterol, fibre and glycemic index using

ESHA food processor software (version 10.9.0), a program based on USDA data and international GI

tables.

5.3.5 STATISTICAL ANALYSIS

The primary outcome of the original study was HbA1c8. The present analysis focused on oxidized

products in serum of participants who had sufficient serum at both zero and week 8, 10, or 12 (n=113).

The treatment differences in markers of oxidative stress were assessed using an unpaired 2‐

tailed Student’s t‐test. Pearson correlations and partial correlations were undertaken on pooled data


Page 63 of 93

from both treatment arms to determine the relation of change in oxidized products to measures of

study outcomes and change in dietary pulse intake.

We indirectly assessed the impact of postprandial glycemic excursions on oxidative damage by

stratifying participants into 4 groups based on their changes in fasting blood glucose (FBG) and HbA1c

(See Table 1 for classification of participants). If participants showed no changes in both FBG and HbA1c,

then they were excluded from the analysis (n= 2). A 2‐sample t‐test compared participants in the lowest

glucose excursion quadrant with those in the highest glucose excursion quadrant, and with those in the

other 3 quadrants combined.

All analyses were carried out using SAS software (version 9.3). Results are expressed as absolute

mean changes from baseline with 95% confidence intervals (CI), unless otherwise noted. Significance

was set at p< 0.05.

5.4 RESULTS

Fifty‐eight individuals (91%) in the high fibre diet and 55 individuals (82%) in the high dietary

pulse diet completed the study and provided blood samples for oxidative stress measurements (Figure

1). There were no significant differences between the 2 treatment arms at baseline (Table 2). Dietary

variables were not significantly different between treatment groups at baseline. There were, however,

significant treatment differences in the dietary variables as change from baseline (Table 3). Notably, the

dietary pulse arm had greater changes in dietary pulse, percent energy from dietary protein, vegetable

protein, and fibre intake than the fibre diet (p<0.05). There was a greater reduction and lower GI at the

end of the study on the dietary pulse than the high‐fibre arm. In contrast, the dietary pulse diet had

smaller changes in percent energy from saturated fat (SFA) and polyunsaturated fatty acid (PUFA)

intakes, although SFA was still lower and PUFA still higher at the end of the study than the fibre diet.

HbA1c was reduced significantly in the treatment compared to control as was SBP. The effects

have been repeated previously8.


Page 64 of 93

5.4.1 OXIDATIVE STRESS MARKERS There was no significant baseline or treatment difference in markers of oxidative stress: TBARS,

CDs, and protein thiols (Table 4). However in the pooled analysis, change in dietary pulse intake was

inversely correlated with change in CDs (Figure 2 and Table 5).

Pooled data from both treatments showed significant positive associations between changes in

TBARS and CDs with changes in LDL‐C, TC:HDL‐C and TG (Table 5). A positive correlation was also found

between change in CDs and change in diastolic blood pressure (DBP) and mean arterial pressure (MAP),

while change in protein thiols was positively correlated with heart rate reduction. Partial Pearson

correlations adjusting for change in body weight did not change results except for the correlation

between change in protein thiols with change in LDL‐C, changing from non‐significant to significant after

adjustment (Table 5).

5.4.2 GLYCEMIC EXCURSION

Both treatment arms were pooled to assess the effects of estimated postprandial glycemic

excursions on markers of oxidative stress. Within each quadrant, we did not find significant changes in

any of the oxidative stress markers from baseline except for quadrant C where it showed improvements

of protein thiol level (p= 0.0014).

We further assessed whether participants with different levels of glycemic excursion would

effect levels of oxidative stress markers. Comparing participants with the lowest estimated glycemic

excursions (Group D) with those with the highest estimated glycemic excursions (Group A), there were

no significant differences on TBARS, CDs, and protein thiols. Nor was there a significant effect on

oxidative stress markers when participants with the lowest glycemic excursion was compared to those in

all other groups combined (Group D vs. combined Groups A, B, and C).

5.5 DISCUSSION

Our results indicated a diet high in plant protein from dietary pulses (beans, chickpeas, lentils,

peas) has no greater reduction on markers of oxidative stress than a high fibre comparator diet in 113


Page 65 of 93

T2DM participants after 12 weeks of intervention. However, an inverse correlation was found between

change in dietary pulse intake and change in CDs, suggesting higher intakes of dietary pulses improved

levels of CDs. Lower estimated glycemic excursions did not significantly affect levels of oxidative stress

markers.

5.5.1 RELATIONS TO OTHER STUDIES

To our knowledge, no randomized trials have assessed the effects of diets high in plant protein

on oxidative stress. Most dietary intervention studies have focused on investigating specific food

sources of plant protein such as soy31, 113, wheat gluten31, and barley71. In comparison to these results,

our analysis of dietary pulses did not show improvements to oxidative damage. However, we did

observe a significant inverse relationship between dietary pulse intake and CDs. Although a high dietary

pulse diet may not reduce oxidative stress, there is still evidence suggesting a high dietary pulse intake

may reduce levels of ox‐LDL‐C32. A previous weight‐loss trial in 30 obese individuals found a diet high in

dietary pulses compared to a standard American diet significantly reduced levels of ox‐LDL‐C without

reducing levels of oxidative stress as measured by malondialdehyde (MDA) and 8‐isoprostane F2α32.

These results suggest dietary pulses may reduce levels of ox‐LDL‐C by reducing the number of LDL‐C

particles that are available for oxidation. Indeed, 2 previous meta‐analyses have shown that diets

emphasizing dietary pulses significantly reduce LDL‐C levels compared to control diets13, 83. Furthermore,

dietary pulses are high in 7S globulin, a protein that has also been shown to increase uptake and

degradation of LDL‐C in humans18. Consequently, lower levels of LDL‐C available for oxidation after

dietary pulse consumption may reduce CVD risk by lowering LDL‐C particle size.

We found no effect of estimated glucose excursion on markers of oxidative stress in our pooled

analysis of both treatment arms. The lack of an effect is unexpected. Dietary pulses have a low‐GI and

some are high in protein and fibre, both of which lower postprandial blood glucose, a factor which has

been strongly associated with oxidative stress109, 114, 115. Previous randomized dietary trials of low‐GI


Page 66 of 93

diets have shown reductions in oxidative stress markers109. It is possible that our study did not find

significant changes in oxidative stress markers with glycemic excursion reductions because our

measurements were calculated and indirect; we used a combination of changes in FBG and HbA1c levels

as a proxy for changes in postprandial glycemia. In addition, the use of a healthy control group that is

high in fibre (i.e. a positive control group), may have minimized the opportunity to see treatment

differences in our analysis since whole wheat lignans and other phenolics have antioxidant properties116.

5.5.2 LIMITATIONS

There are several limitations to our work. First and most importantly, the present analysis on

oxidative stress levels represents a secondary analysis. The trial was designed to provide good power for

measures of HbA1c8 which may not have had sufficient power to detect significant treatment

differences in markers of oxidative damage. Second, we did not adjust for antioxidants found in foods

and supplementations. The use of Oxygen Radical Absorbance Capacity (ORAC) values to measure the

antioxidant capacity of foods is discouraged by the USDA, citing reasons that the values were tested in

vitro and the antioxidant effects cannot be confidently extrapolated to human physiology117. The use of

antioxidant supplements was not adjusted for in our analysis because the frequency of use was not

different between the 2 treatment arms at baseline and at the end of the trial (data not shown). Third,

there are limitations to the methods we used to measure TBARS, CDs, and protein thiols. Each of these

oxidative stress markers has its own inherent limitations including measurement and may not be

representative of overall oxidative stress levels because they are rapidly degraded in vivo110‐112. Further

our methods of oxidative stress measurement may not be sensitive enough to measure levels of

oxidative stress markers in participants’ serum. For examples, while we reported TBARS values in the

range between 0.217‐ 0.294 μmol/L and measurements can be made between 0.1 to 2.1 μmol/L,

measurements of TBARS are most accurate at levels >0.3μmol/L110. Taking together all these issues

associated with measuring oxidative stress, we tried to increase the sensitivity of our analysis by using 3


Page 67 of 93

different methods of measurement of oxidative stress. We found no treatment differences on oxidative

stress regardless of the methods used. Last, although our original study did not show a significant LDL‐C

reduction, we still reported a significant non‐HDL reduction8, suggesting that a high dietary pulse intake

may still improve cardiovascular risk by reducing serum atherogenic lipids.

5.6 CONCLUSIONS

A diet high in vegetable protein from dietary pulses (beans, chickpeas, lentils, and peas) of

approximately 213 grams per day does not appear to reduce markers of oxidative damage when

compared to a high fibre diet in individuals with T2DM. Although our secondary analysis did not show

treatment differences in measures of oxidative stress, meta‐analyses have reported high dietary pulse

intake can still reduce CVD risk by improving levels of LDL‐C13, 83. As our analysis was secondary in nature

and the methods of measuring oxidative stress were relatively insensitive, we cannot rule out a

potential oxidative stress benefit of high vegetable protein diets. Therefore further large, long‐term,

high‐quality trials powered to assess the effect of high vegetable protein diets on oxidative stress are

needed. To identify research gaps and provide a more precise estimate of the effect of vegetable

proteins on these endpoints, our group has undertaken a series of systematic reviews and meta‐

analyses of randomized trials. Results from this project will guide our understanding of the role of

vegetable protein in cardiovascular risk reduction.

5.7 SOURCE OF FUNDING

This work was supported by ABIP through the PURENet and the Saskatchewan Pulse Growers.

VH was supported by an Ontario Graduate Scholar (OGS) award. RJD was funded by a Canadian

Institutes for Health Research (CIHR) Postdoctoral Fellowship Award. DJAJ was funded by the

Government of Canada through the Canada Research Chair Endowment. None of the sponsors had a

role in any aspect of the present study, including design and conduct of the study; collection,


Page 68 of 93

management, analysis, and interpretation of the data; and preparation, review, or approval of the

manuscript.

5.8 CONFLICTS OF INTEREST/DISCLOSURE

JLS has received several unrestricted travel grants from The Coca‐Cola Company to present

research at meetings and is a co‐investigator on an unrestricted research grant from The Coca‐Cola

Company. JLS has also received travel funding and honoraria from Abbott Laboratories, Archer Daniels

Midland, and the International Life Sciences Institute (ILSI) North America and Brazil; and research

support, consultant fees, and travel funding from Pulse Canada. Spouse of JLS works for Unilever

Canada. RJD, JB, and CWCK are co‐investigators on an unrestricted grant from The Coca‐Cola Company.

CWCK has served on the scientific advisory board, received research support, travel funding, consultant

fees, or honoraria from Pulse Canada, Barilla, Solae, Unilever, Hain Celestial, Loblaws Inc., Oldways

Preservation Trust, the Almond Board of California, the International Nut Council, Paramount Farms, the

California Strawberry Commission, the Canola and Flax Councils of Canada, and Saskatchewan Pulse

Growers. CWCK also receives partial salary funding from research grants provided by Unilever, Loblaw’s,

and the Almond Board of California. DJAJ has served on the Scientific Advisory Board of Sanitarium

Company, Agri‐Culture and Agri‐Food Canada (AAFC), Canadian Agriculture Policy Institute (CAPI),

California Strawberry Commission, Loblaw Supermarket, Herbal Life International, Nutritional

Fundamental for Health, Pacific Health Laboratories, Metagenics, Bayer Consumer Care, Orafti, Dean

Foods, Kellogg’s, Quaker Oats, Procter & Gamble, Coca‐Cola, NuVal Griffin Hospital, Abbott, Pulse

Canada, Saskatchewan Pulse Growers, and Canola Council of Canada; received honoraria for scientific

advice from Sanitarium Company, Orafti, the Almond Board of California, the American Peanut Council,

International Tree Nut Council Nutrition Research and Education Foundation and the Peanut Institute,

Herbal Life International, Pacific Health Laboratories, Nutritional Fundamental for Health, Barilla,

Metagenics, Bayer Consumer Care, Unilever Canada and Netherlands, Solae, Oldways, Kellogg’s, Quaker


Page 69 of 93

Oats, Procter & Gamble, Coca‐Cola, NuVal Griffin Hospital, Abbott, Canola Council of Canada, Dean

Foods, California Strawberry Commission, Haine Celestial, Pepsi, and Alpro Foundation; has been on the

speakers panel for the Almond Board of California; received research grants from Saskatchewan Pulse

Growers, the Agricultural Bioproducts Innovation Program (ABIP) through the Pulse Research Network

(PURENet), Advanced Food Materials Network (AFMNet), Loblaw, Unilever, Barilla, Almond Board of

California, Coca‐Cola, Solae, Haine Celestial, Sanitarium Company, Orafti, International Tree Nut Council

Nutrition Research and Education Foundation and the Peanut Institute, the Canola and Flax Councils of

Canada, Calorie Control Council, Canadian Institutes of Health Research, Canada Foundation for

Innovation, and the Ontario Research Fund; and received travel support to meetings from the Solae,

Sanitarium Company, Orafti, AFMNet, Coca‐Cola, The Canola and Flax Councils of Canada, Oldways

Preservation Trust, Kellogg’s, Quaker Oats, Griffin Hospital, Abbott Laboratories, Dean Foods, the

California Strawberry Commission, American Peanut Council, Herbal Life International, Nutritional

Fundamental for Health, Metagenics, Bayer Consumer Care, AAFC, CAPI, Pepsi, Almond Board of

California, Unilever, Alpro Foundation, International Tree Nut Council, Barilla, Pulse Canada, and the

Saskatchewan Pulse Growers. Dr Jenkins' wife is a director of Glycemic Index Laboratories, Toronto,

Ontario, Canada.


Page 70 of 93

5.9 TABLES

Table 1. Definitions of quadrant groups for glucose excursion analysis.

Quadrant Group Definition Implications

A ‐ no change in FBG and an increase in HbA1c, or ‐ decrease in FBG and no change in HbA1c, or ‐ decrease in FBG and an increase in HbA1c

High postprandial glucose response

B ‐ increase in FBG and an increase HbA1c Intermediate postprandial glucose response

C ‐ decrease in FBG and a decrease in HbA1c Intermediate postprandial glucose response

D ‐ no change in FBG and a decrease in HbA1c, or ‐ increase in FBG and no change in HbA1c, or ‐ increase in FBG and a decrease in HbA1c

Low postprandial glucose response


Page 71 of 93

Table 2. Baseline Characteristics of Participants.

High Wheat Fibre Diet(n= 58)

Low‐GI Legume Diet(n= 55) p‐value

Age, mean (SD), y 62 (7.7) 58 (10.2) 0.070Sex Male 27 (47) 31 (56) Female 31 (53) 24 (44) 0.348Race/ethnicity European African 10 (17) 6 (11) 0.422 East Indian 16 (28) 15 (27) 1.000 European 16 (28) 13 (24) 0.671 Far Eastern 4 (7) 6 (11) 0.521 Other whites/white 5 (9) 10 (18) 0.170 Others 7 (12) 5 (9) 0.763Weight, mean (SD), kg 82.4 (17.0) 87.0 (20.4) 0.194BMI, mean (SD) 30.0 (5.6) 31.7 (6.8) 0.141Current smokers 1 (2) 0 (0) 1.000HbA1c value, % of total hemoglobin, mean (SD) ≤ 7.1% 30 (52) 26 (47) >7.1% 28 (48) 29 (53) 0.708Duration of DM, mean (SD) 8.6 (6.6) 9.3 (6.5) 0.557LDL‐C, mean (SD), mmol/L 2.4 (1.0) 2.1 (0.7) 0.136Non‐HDL‐C, mean (SD), mmol/L 2.0 (1.0) 1.6 (0.3) 0.016TC:HDL‐C, mean (SD) 3.5 (0.9) 3.7 (1.0) 0.229Triglycerides, mean (SD), mmol/L 1.3 (0.7) 1.6 (1.1) 0.043Medication use Metformin 56 (97) 51(93) 0.430 Sulfonylurea 22 (38) 17(31) 0.553 Thiazolodinedione 7 (12) 10 (18) 0.435 DPP‐4 Inhibitor 8 (14) 5(9) 0.559 Meglitinides (nonsulfonylurea) 3 (5) 3(5) 1.000 α‐Glucosidase inhibitor 0 (0) 2(4) 0.235Cholesterol Lowering medications 38 (66) 42(76) 0.222Blood pressure medications 35(60) 41 (75) 0.115

No. of Participants (%)


Page 72 of 93

Table 3. Daily Nutritional Profile at Baseline and End of Study†.

Baseline End Baseline End P‐value‡Energy, kcal/d 1605 (1502‐1708) 1442 (1340‐ 1545)* 1742 (1638 ‐1845) 1540 (1442‐ 1638)* 0.554Pulse intake, g/d 45.4 (32.1‐ 58.8) 14.4 (8.62‐ 20.3)* 55.4 (38.8‐ 72.0) 212.9 (183.9‐ 242.0)* <0.0001Total fat, %E 32.9 (31.4‐ 34.4) 28.5 (26.9‐ 30.0)* 33.1 (31.5‐ 34.7) 30.0 (28.3‐ 31.6)* 0.364SFA, % 9.4 (8.8‐ 10.0) 8.7 (8.1‐ 9.4) 10.3 (9.6‐ 11) 8.0 (7.4‐ 8.7)* 0.004MUFA, %E 12.7 (11.8‐ 13.5) 10.5 (9.8‐ 11.3)* 12.7 (11.9‐ 13.5) 11.7 (10.8‐ 12.5) 0.139PUFA, %E 7.2 (6.7‐ 7.7) 5.8 (5.3‐ 6.2)* 6.7 (6.2‐ 7.2) 6.6 (6.0‐ 7.1) 0.006Diet Cholesterol, mg/1000kcal 142.9 (127.3‐ 158.4) 154.4 (133.8‐ 175.0) 141.0 (126.2‐ 155.8) 148.7 (125.1‐ 172.3) 0.742Protein, %E 19.5 (15.2‐ 17.8) 21.2 (20.3‐ 22.2)* 19.8 (19.0‐ 20.6) 23.0 (22.0‐ 24.1)* 0.038Plant Protein, %E 8.12 (7.59‐ 8.66) 7.60 (7.14‐ 8.06) 7.99 (7.48‐ 8.51) 10.40 (9.81‐ 10.99)* 0.0002Carbohydrates, %E 46.4 (44.6‐ 48.1) 48.6 (46.7‐ 50.5)* 46.0 (46.7‐ 50.5) 45.6 (43.8‐ 47.4) 0.067Fibre, g/1000kcal 16.5 (15.2‐ 17.8) 18.4 (17.2‐ 19.6)* 15.9 (14.4‐ 17.4) 25.6 (23.3‐ 28.0)* <0.0001Alcohol, %E 2.5 (1.3‐ 3.7) 3.7 (2.2‐ 5.3) 2.2 (0.8‐ 3.6) 3.8 (1.5‐ 6.1) 0.824Glycaemic Index§ 78.6 (77.2‐ 80.0) 81.7 (80.5‐ 82.8)* 80.0 (78.4‐ 81.6) 65.8 (64.0‐ 67.6)* <0.0001Glycaemic Load 145.0 (134.7‐ 155.3) 140.9 (131.9‐ 151.0)* 161.8 (148.4‐ 175.3) 115.0 (105.8‐ 124.1)* 0.059

High‐Fibre Diet(n=58)

Low‐GI Pulse Diet(n=55)

Abbreviations: SFA denotes saturated fatty acid, MUFA monounsaturated fatty acid, PUFA polyunsaturated fatty acid *Significant difference within treatment by 2‐tailed unpaired t test (p<0.05) †Values are expressed as means with their 95% confidence intervals ‡P values represent the significance of differences between treatment as changes from baseline by 2‐tailed paired t test (p<0.05) §Based on the GI bread scale


Page 73 of 93

Table 4. Mean Study Measurements and Significance of Treatment Differences on Oxidative Stress Markers‡.

Baseline End Mean difference Baseline End Mean difference P‐value†TBARS (μmmol/mmol LDL‐C) 0.254 (0.227, 0.282) 0.242 (0.216, 0.268) ‐0.012 (‐0.030, 0.006) 0.249 (0.219, 0.279) 0.261 (0.227, 0.294) 0.012 (‐0.012, 0.035) 0.292CD (μmmol/mmol LDL‐C) 29.013 (26.205, 31.821) 28.873 (25.827, 31.920) ‐0.139 (‐1.320, 1.040) 31.074 (27.481, 34.667) 30.508 (26.682, 34.334) ‐0.566 (‐2.160, 1.030) 0.672Protein Thiols (μM) 390.64 (373.59, 407.68) 422.82 (405.67, 438.98) 32.18 (13.00, 51.37)* 396.78 (376.10, 417.46) 413.58 (389.08, 438.08) 16.80 (‐10.01, 43.69) 0.359

High‐Fibre Diet(n=58)

Low‐GI Pulse Diet(n=55)

Abbreviations: TBARS denote thiobarbituric acid reactive substances; CD conjugated dienes *Significant difference within treatment by 2‐tailed paired t test (p<0.05) †P values represent the significance of differences between treatment as changes from baseline by 2‐tailed paired t test (p<0.05) ‡ Values are expressed as means with their 95% confidence intervals


Page 74 of 93

Table 5. Associations between Changes in Oxidative Stress Markers and Changes in Study Outcomes.

∆TBARS(μmmol)

∆CD(μmmol)

∆Thiols(μM)

∆TBARS(μmmol)

∆CD(μmmol)

∆Thiols(μM)

∆Pulse Intake (grams/day) ‐0.018 ‐0.225* ‐0.142 ‐0.016 ‐0.226* ‐0.142∆SBP (mm Hg) ‐0.084 0.099 0.012 ‐0.057 0.142 0.048∆DBP (mm Hg) 0.006 0.238* 0.061 0.028 0.287* 0.096∆HR (beats per min) 0.154 0.135 0.232* 0.129 0.104 0.202*∆WC (cm) 0.084 0.031 0.002 0.077 0.044 0.008∆FBG (mmol/L) 0.017 0.062 ‐0.014 0.009 0.06 ‐0.024∆HbA1c (%) ‐0.012 0.035 0.167 ‐0.031 ‐0.006 0.164∆LDL‐C (mmol/L) 0.250* 0.297* 0.186* 0.218* 0.249* 0.249*∆TC:HDL‐C 0.253* 0.307* ‐0.007 0.223* 0.295* ‐0.020∆TG (mmol/L) 0.306* 0.400* ‐0.032 0.281* 0.434* ‐0.080∆TBARS (μmmol) ‐ 0.109 0.027 ‐ 0.060 0.006∆CD (μmmol) 0.109 ‐ ‐0.021 ‐0.082 ‐ ‐0.082∆Thiols (μM) 0.027 ‐0.021 ‐ 0.006 ‐0.082 ‐

Pearson Correlations Partial Correlations†

*Significance by Pearson and Partial correlations (p<0.05) †Partial Correlations adjusted for change in body weight


Page 75 of 93

5.10 FIGURES

Figure 1. Flowchart of Participants through the Trial.


Page 76 of 93

Figure 2. Correlation of Change in Pulse intake (∆Pulse Intake) and Change in Conjugated Dienes (∆CDs). Analysis was conducted with Pearson’s Correlation (p=0.016).

R=‐0.225

∆Pulse Intake

∆CD

s (mmol/L)

OVERALL DISCUSSION AND LIMITATIONS

Page 77 of 93

CHAPTER VI‐ OVERALL DISCUSSION AND LIMITATIONS


Page 78 of 93

6.1 OVERALL DISCUSSION

Two studies have been undertaken to assess the effects of dietary pulses (beans, chickpeas,

lentils, and peas) on oxidative stress and lipid measures associated with cardiovascular risk. Our

systematic review and meta‐analysis found that high dietary pulse consumption compared to control

diets reduced LDL‐C by 0.17 mmol/L or 5% from baseline, but found no significant effects on secondary

lipid targets of CVD risk including Apo‐B and non‐HDL‐C. Moreover, although our intervention trial,

which emphasized dietary pulse intake as a means to lower glycemic index, did not significantly reduce

oxidative stress markers compared to a high fibre control diet, an inverse correlation was observed

between dietary pulse intake and one of the oxidative stress markers. Taken together, our results

suggest that a diet high in dietary pulses may improve cardiovascular risk through lowering levels of LDL‐

C.

Our results support previous findings from other studies. Our systematic review and meta‐

analysis of 26 trials found dietary pulse intervention compared to control diets reduced LDL‐C by 0.17

mmol/L. These results are in agreement with 2 previous meta‐analyses assessing dietary pulse intakes

which have also reported significant LDL‐C reductions13, 83. Similarly, our results for oxidative stress

markers are congruent with previous findings. A previous human feeding trial of 30 obese participants

on a high dietary pulse diet compared to a typical American diet had also found no differences in

oxidative damage using measurements of malendialdehyde (MDA) and isoprostane‐F2α. Despite these

findings, the investigators still reported significant reductions of serum ox‐LDL‐C32. Our feeding trial

suggests that high dietary pulse intake may not directly reduce levels of oxidative stress, however, they

may still reduce levels of ox‐LDL‐C by lowering the number of LDL‐C particles available for oxidation as

suggested by our meta‐analysis. Consequently, lower levels of LDL‐C available for oxidation after dietary

pulse consumption may reduce CVD risk by lowering LDL‐C particle size.


Page 79 of 93

Several hypotheses have been generated in understanding the mechanism of action with

respect to the lipid‐lowering effects of dietary pulses. Dietary pulses are high in fibre and have a low‐

glycemic index, both of which are associated with reduced lipid levels and CHD risk15, 118. Possibly, more

important is that dietary pulses and common to all legumes including soy and peanuts, is that they are

all high in a protein called 7S globulin. This protein fraction has been considered responsible for the

hypolipidemic effect of soy proteins18. Prospective cohort studies have shown high intakes of legumes

(dietary pulse and soy/peanuts) are associated with reduced risk of CHD and stroke80, 81 but prospective

cohort study has been conducted to date to assess the association of dietary pulse intake on

cardiovascular events. More studies are needed to understand whether higher pulse intake can lower

the risk of CVD hard endpoints such as myocardial infarctions, stroke, and CVD‐related mortality.

6.2 LIMITATIONS

6.2.1 SYSTEMATIC REVIEW AND META‐ANALYSIS

Limitations of our systematic review and meta‐analysis include: 1) the majority of studies were

of low quality (MQS<8), short duration (<3 months), and did not report enough data to judge risk of bias;

2) the majority the non‐HDL‐C values had to be calculated and SDs imputed. 3) only one trial provided

Apo‐B data. These observations reinforce the need for longer, better‐designed clinical trials that report

data more systematically and transparently to better understand the effect of dietary pulses on serum

lipids as they are important biomarkers of cardiovascular health.

6.2.2 DIETARY INTERVENTION TRIAL

Limitations of our dietary intervention trial include: 1) the analysis on oxidative stress levels

represents a secondary analysis which was not sufficiently powered to detect significant treatment

differences in markers of oxidative damage; 2) adjustments for antioxidants found in foods and

supplementations were not possible. The use of Oxygen Radical Absorbance Capacity (ORAC) values to

measure the antioxidant capacity of foods is discouraged by the USDA, citing reasons that the values


Page 80 of 93

were tested in vitro and the antioxidant effects cannot be confidently extrapolated to human

physiology117. The antioxidant supplements was not adjusted for in our analysis because the frequency

of use was not different between the 2 treatment arms at baseline and at the end of the trial (data not

shown); 3. we did not directly measure levels of ox‐LDL‐C but used markers of oxidative stress (TBARS,

conjugated dienes, and protein thiols) as surrogates. As these markers are rapidly degraded in vivo, they

may not be representative of the level of ox‐LDL‐C in circulation. There is a need for future trials to

improve the existing evidence on the effects of dietary pulses on ox‐LDL‐C.

6.3 MECHANISM OF ACTION

The mechanism by which dietary pulses lower cholesterol levels is unclear; however, the effect

can be broadly divided into 2 pathways: extrinsic and intrinsic. The extrinsic pathway proposes that

dietary pulses reduce serum cholesterol by displacing saturated and trans fat intake57, 82. High saturated

and trans fat intake has been associated with animal protein intake which has also been associated with

increased cardiovascular risk119. Possibly more importantly, dietary pulses are high in 7S globulins, a

peptide which has been shown to increase uptake and degradation of LDL‐C in human hepatic cells, and

is also hypothesized to confer the hypocholesterolemic effects of soybeans18. Dietary pulses are also

high in soluble fibre which is known to bind to dietary cholesterol in the intestine and prevent its

absorption85, 86. Of the intrinsic pathway proposes that in addition to binding to dietary cholesterol, the

viscous soluble fibre in dietary pulses can bind to bile acids to prevent its re‐absorption. Consequently,

there is an increase in the production of bile acids by the liver, which decreases the liver pool of

cholesterol and increases uptake of serum cholesterol by the liver thereby decreasing circulating

cholesterol in the blood22. Further fermentation of fibre in the colon produces short chain fatty acids

(SCFA) which have been associated with inhibition of cholesterol synthesis19, 23.


Page 81 of 93

6.4 CLINICAL IMPLICATIONS

Dietary patterns that emphasize dietary pulses can lower the risk for CVD. Although we did not

show a significant difference on oxidative stress marker with higher dietary pulse intake, our systematic

review and meta‐analysis found a significant LDL‐C reduction in diets that emphasize dietary pulse

intake. Our data also reinforce the evidence from and support existing recommendations for heart

healthy dietary patterns of which pulses are a key component: Portfolio, DASH, and Mediterranean3‐6.

The median dietary pulse intake in our meta‐analyses was of 130g/day (~1.5 servings/day), a higher

intake target than that proposed by the AHA of ≥4‐5 servings/week. Achieving a high level of intake may

prove a challenge in some Western countries, given that current U.S. level of intake is 0.2 serving/day82.

However, a dietary pulse intake of ~1.5 servings/d is very reasonable as this level is currently consumed

by many cultures without reports of side effects that would limit consumption. As the majority of these

trials were conducted on a background heart‐healthy NCEP‐like diet, including >20‐25g/d of fibre and

<10%E of saturated fat, these results can be considered in addition to the 5‐10% LDL‐C lowering

expected from these diets alone53. Whereas guidelines have largely focused on dietary pulses as part of

a healthy dietary pattern in the prevention and management of diabetes and CVD, these data support

lipid‐lowering and cardiovascular risk reduction recommendations.

6.5 FUTURE DIRECTIONS

Future research is needed. The reducing effects of dietary pulses on LDL‐C has been shown in

our and 2 previous meta‐analyses13, 83; therefore future analyses should focus on assessing the effects of

dietary pulses on other established lipid endpoints. More importantly, human feeding trials measuring

Apo‐B is needed. Our analysis had only identified one trial that met the inclusion criteria24; therefore it is

likely we were under‐powered to accurately detect the effects of dietary pulses on Apo‐B. Secondly,

there is an urgent need for further large, long‐term, high‐quality trials to assess the effect of high

vegetable protein diets on oxidative stress and other major cardiovascular risk factors including body


Page 82 of 93

weight, glycemic control, lipids, and blood pressure. To identify research gaps and provide a more

precise estimate of the effect of vegetable proteins on these endpoints, our group has undertaken a

series of systematic reviews and meta‐analyses of randomized trials. Results from this project will guide

our understanding of the role of vegetable protein in cardiovascular risk reduction. Lastly, although

dietary intervention trials have shown favourable effects of dietary pulses on biomarkers of CVD risk, no

prospective cohort studies have assessed specifically the association of dietary pulses on hard endpoints

of CVD. Most observational studies have reported high legume intakes which include dietary pulses and

soy or peanuts, have favourable effects on CVD risk13, 83, but none have assessed legume intake

independent of soy or peanuts. More studies are needed to understand the effects of dietary pulses on

CHD or stroke risk. Lastly, the role of ox‐LDL‐C in human cardiovascular health is uncertain. Although

animal and cell culture studies have supported the hypothesis of ox‐LDL‐C and CVD risk35, large human

clinical trials supplementing dietary patterns with anti‐oxidants have reported no improvements on

incidence of CHD and stroke120, 121. Consequently, no major dyslipidemic guidelines including NCEP‐ATP

III and Canadian Cardiovascular Society (CCS) have recognized and set target goals for serum ox‐LDL‐C or

have made recommendations for anti‐oxidant supplements for reducing cardiovascular risk3, 4. The role

of ox‐LDL‐C needs to be more directly assessed in relation to cardiovascular outcomes.

CONCLUSIONS

Page 83 of 93

CHAPTER VII‐ CONCLUSIONS

CONCLUSIONS

Page 84 of 93

7.1 CONCLUSIONS Dietary patterns that emphasize dietary pulses (beans, chickpeas, lentils, and peas) improve

cardiovascular risk by reducing LDL‐C levels but not oxidative stress.

This thesis work demonstrated the following:

1. In a systematic review and meta‐analysis of 26 randomized trials (n= 1013), diets

emphasizing dietary pulses significantly lowered LDL‐C compared with isocaloric control

diets (mean difference= ‐0.17 mmol/L [95% CI: ‐0.25, ‐0.09]; p<0.0001). No treatment

effects were observed for Apo‐B and non‐HDL‐C.

2. In a secondary endpoint analysis of a feeding trial in 113 participants, there were no

significant differences between a high dietary pulse diet as a means to lower the glycemic

index to a high fibre comparator diet on markers of oxidative stress including thiobarbituric

acid reactive substances (TBARS), conjugated dienes (CDs), and protein thiols.

REFERENCES

Page 85 of 93

CHAPTER VIII‐ REFERENCES

REFERENCES

Page 86 of 93

8.1 REFERENCES

1. Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, McQueen M, Budaj A, Pais P, Varigos J, Lisheng L, Investigators IS. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the interheart study): Case‐control study. Lancet. 2004;364:937‐952

2. Law MR, Wald NJ, Rudnicka AR. Quantifying effect of statins on low density lipoprotein cholesterol, ischaemic heart disease, and stroke: Systematic review and meta‐analysis. Bmj. 2003;326:1423

3. Anderson TJ, Gregoire J, Hegele RA, Couture P, Mancini GB, McPherson R, Francis GA, Poirier P, Lau DC, Grover S, Genest J, Jr., Carpentier AC, Dufour R, Gupta M, Ward R, Leiter LA, Lonn E, Ng DS, Pearson GJ, Yates GM, Stone JA, Ur E. 2012 update of the canadian cardiovascular society guidelines for the diagnosis and treatment of dyslipidemia for the prevention of cardiovascular disease in the adult. The Canadian journal of cardiology. 2013;29:151‐167

4. National Cholesterol Education Program Expert Panel on Detection E, Treatment of High Blood Cholesterol in A. Third report of the national cholesterol education program (ncep) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel iii) final report. Circulation. 2002;106:3143‐3421

5. Dworatzek PD AK, Gougeon R, Husein N, Sievenpiper JL, Williams SL. Canadian diabetes association 2013 clinical practice guidelines for the prevention and management of diabetes in canada: Nutrition therapy. Canadian Journal of Diabetes. 2013;37:S46‐S55

6. Association AD. 2013 clinical practice recommendations. Diabetes Care. 2013;36:S1‐ S110 7. Stampfer MJ, Hu FB, Manson JE, Rimm EB, Willett WC. Primary prevention of coronary heart

disease in women through diet and lifestyle. The New England journal of medicine. 2000;343:16‐22

8. Jenkins DJ, Kendall CW, Augustin LS, Mitchell S, Sahye‐Pudaruth S, Blanco Mejia S, Chiavaroli L, Mirrahimi A, Ireland C, Bashyam B, Vidgen E, de Souza RJ, Sievenpiper JL, Coveney J, Leiter LA, Josse RG. Effect of legumes as part of a low glycemic index diet on glycemic control and cardiovascular risk factors in type 2 diabetes mellitus: A randomized controlled trial. Archives of internal medicine. 2012;172:1653‐1660

9. Sacks FM, Svetkey LP, Vollmer WM, Appel LJ, Bray GA, Harsha D, Obarzanek E, Conlin PR, Miller ER, 3rd, Simons‐Morton DG, Karanja N, Lin PH, Group DA‐SCR. Effects on blood pressure of reduced dietary sodium and the dietary approaches to stop hypertension (dash) diet. Dash‐sodium collaborative research group. The New England journal of medicine. 2001;344:3‐10

10. Zhang Z, Lanza E, Kris‐Etherton PM, Colburn NH, Bagshaw D, Rovine MJ, Ulbrecht JS, Bobe G, Chapkin RS, Hartman TJ. A high legume low glycemic index diet improves serum lipid profiles in men. Lipids. 2010;45:765‐775

11. Ambring A, Friberg P, Axelsen M, Laffrenzen M, Taskinen MR, Basu S, Johansson M. Effects of a mediterranean‐inspired diet on blood lipids, vascular function and oxidative stress in healthy subjects. Clinical science. 2004;106:519‐525

12. Barnard ND, Cohen J, Jenkins DJ, Turner‐McGrievy G, Gloede L, Green A, Ferdowsian H. A low‐fat vegan diet and a conventional diabetes diet in the treatment of type 2 diabetes: A randomized, controlled, 74‐wk clinical trial. The American journal of clinical nutrition. 2009;89:1588S‐1596S

13. Bazzano LA, Thompson AM, Tees MT, Nguyen CH, Winham DM. Non‐soy legume consumption lowers cholesterol levels: A meta‐analysis of randomized controlled trials. Nutrition, metabolism, and cardiovascular diseases : NMCD. 2011;21:94‐103

REFERENCES

Page 87 of 93

14. Jenkins DJ, Wolever TM, Taylor RH, Ghafari H, Jenkins AL, Barker H, Jenkins MJ. Rate of digestion of foods and postprandial glycaemia in normal and diabetic subjects. British medical journal. 1980;281:14‐17

15. Pereira MA, O'Reilly E, Augustsson K, Fraser GE, Goldbourt U, Heitmann BL, Hallmans G, Knekt P, Liu S, Pietinen P, Spiegelman D, Stevens J, Virtamo J, Willett WC, Ascherio A. Dietary fiber and risk of coronary heart disease: A pooled analysis of cohort studies. Archives of internal medicine. 2004;164:370‐376

16. Mirrahimi A, de Souza RJ, Chiavaroli L, Sievenpiper JL, Beyene J, Hanley AJ, Augustin LS, Kendall CW, Jenkins DJ. Associations of glycemic index and load with coronary heart disease events: A systematic review and meta‐analysis of prospective cohorts. Journal of the American Heart Association. 2012;1:e000752

17. Kelemen LE, Kushi LH, Jacobs DR, Jr., Cerhan JR. Associations of dietary protein with disease and mortality in a prospective study of postmenopausal women. American journal of epidemiology. 2005;161:239‐249

18. Lovati MR, Manzoni C, Corsini A, Granata A, Fumagalli R, Sirtori CR. 7s globulin from soybean is metabolized in human cell cultures by a specific uptake and degradation system. The Journal of nutrition. 1996;126:2831‐2842

19. Anderson JW, Story L, Sieling B, Chen WJ, Petro MS, Story J. Hypocholesterolemic effects of oat‐bran or bean intake for hypercholesterolemic men. The American journal of clinical nutrition. 1984;40:1146‐1155

20. Belski R, Mori TA, Puddey IB, Sipsas S, Woodman RJ, Ackland TR, Beilin LJ, Dove ER, Carlyon NB, Jayaseena V, Hodgson JM. Effects of lupin‐enriched foods on body composition and cardiovascular disease risk factors: A 12‐month randomized controlled weight loss trial. International journal of obesity. 2011;35:810‐819

21. Cobiac L, McArthur R, Nestel PJ. Can eating baked beans lower plasma cholesterol? European journal of clinical nutrition. 1990;44:819‐822

22. Duane WC. Effects of legume consumption on serum cholesterol, biliary lipids, and sterol metabolism in humans. Journal of lipid research. 1997;38:1120‐1128

23. Finley JW, Burrell JB, Reeves PG. Pinto bean consumption changes scfa profiles in fecal fermentations, bacterial populations of the lower bowel, and lipid profiles in blood of humans. The Journal of nutrition. 2007;137:2391‐2398

24. Gravel K LS, Asselin G, Dufresne A, Lemay A, Forest JC, Dodin S. Effects of pulse consumption in women presenting components of the metabolic syndrome: A randomized controlled trial. Mediterranean Journal of Nutrition and Metabolism. 2010;3:143‐151

25. Hermsdorff HH, Zulet MA, Abete I, Martinez JA. A legume‐based hypocaloric diet reduces proinflammatory status and improves metabolic features in overweight/obese subjects. European journal of nutrition. 2011;50:61‐69

26. Mackay S, Ball MJ. Do beans and oat bran add to the effectiveness of a low‐fat diet? European journal of clinical nutrition. 1992;46:641‐648

27. Pittaway JK, Ahuja KD, Cehun M, Chronopoulos A, Robertson IK, Nestel PJ, Ball MJ. Dietary supplementation with chickpeas for at least 5 weeks results in small but significant reductions in serum total and low‐density lipoprotein cholesterols in adult women and men. Annals of nutrition & metabolism. 2006;50:512‐518

28. Pittaway JK, Ahuja KD, Robertson IK, Ball MJ. Effects of a controlled diet supplemented with chickpeas on serum lipids, glucose tolerance, satiety and bowel function. Journal of the American College of Nutrition. 2007;26:334‐340

29. Shams H TF, EntezariM, Abadi A. Effects of cooked lentils on gycemic control and blood lipids of patients with type 2 diabetes. ARYA Atherosclerosis Journal. 2008;4:1‐5

REFERENCES

Page 88 of 93

30. Hodgson JM, Lee YP, Puddey IB, Sipsas S, Ackland TR, Beilin LJ, Belski R, Mori TA. Effects of increasing dietary protein and fibre intake with lupin on body weight and composition and blood lipids in overweight men and women. International journal of obesity. 2010;34:1086‐1094

31. Jenkins DJ, Kendall CW, Vidgen E, Augustin LS, van Erk M, Geelen A, Parker T, Faulkner D, Vuksan V, Josse RG, Leiter LA, Connelly PW. High‐protein diets in hyperlipidemia: Effect of wheat gluten on serum lipids, uric acid, and renal function. The American journal of clinical nutrition. 2001;74:57‐63

32. Crujeiras AB, Parra D, Abete I, Martinez JA. A hypocaloric diet enriched in legumes specifically mitigates lipid peroxidation in obese subjects. Free radical research. 2007;41:498‐506

33. Albertini R, Moratti R, De Luca G. Oxidation of low‐density lipoprotein in atherosclerosis from basic biochemistry to clinical studies. Current molecular medicine. 2002;2:579‐592

34. Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztum JL. Beyond cholesterol. Modifications of low‐density lipoprotein that increase its atherogenicity. The New England journal of medicine. 1989;320:915‐924

35. Steinbrecher UP, Zhang HF, Lougheed M. Role of oxidatively modified ldl in atherosclerosis. Free radical biology & medicine. 1990;9:155‐168

36. Halton TL, Willett WC, Liu S, Manson JE, Albert CM, Rexrode K, Hu FB. Low‐carbohydrate‐diet score and the risk of coronary heart disease in women. The New England journal of medicine. 2006;355:1991‐2002

37. Hu FB, Willett WC. Optimal diets for prevention of coronary heart disease. JAMA : the journal of the American Medical Association. 2002;288:2569‐2578

38. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Borden WB, Bravata DM, Dai S, Ford ES, Fox CS, Franco S, Fullerton HJ, Gillespie C, Hailpern SM, Heit JA, Howard VJ, Huffman MD, Kissela BM, Kittner SJ, Lackland DT, Lichtman JH, Lisabeth LD, Magid D, Marcus GM, Marelli A, Matchar DB, McGuire DK, Mohler ER, Moy CS, Mussolino ME, Nichol G, Paynter NP, Schreiner PJ, Sorlie PD, Stein J, Turan TN, Virani SS, Wong ND, Woo D, Turner MB, American Heart Association Statistics C, Stroke Statistics S. Executive summary: Heart disease and stroke statistics‐‐2013 update: A report from the american heart association. Circulation. 2013;127:143‐152

39. Canada PHAo. 2009 tracking heart disease and stroke in canada. 2009 40. Munro JM, Cotran RS. The pathogenesis of atherosclerosis: Atherogenesis and inflammation.

Laboratory investigation; a journal of technical methods and pathology. 1988;58:249‐261 41. D'Agostino RB, Wolf PA, Belanger AJ, Kannel WB. Stroke risk profile: Adjustment for

antihypertensive medication. The framingham study. Stroke; a journal of cerebral circulation. 1994;25:40‐43

42. Castelli WP, Anderson K, Wilson PW, Levy D. Lipids and risk of coronary heart disease. The framingham study. Annals of epidemiology. 1992;2:23‐28

43. Nissen SE, Wolski K. Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes. The New England journal of medicine. 2007;356:2457‐2471

44. Ray KK, Seshasai SR, Erqou S, Sever P, Jukema JW, Ford I, Sattar N. Statins and all‐cause mortality in high‐risk primary prevention: A meta‐analysis of 11 randomized controlled trials involving 65,229 participants. Archives of internal medicine. 2010;170:1024‐1031

45. Cholesterol Treatment Trialists C, Baigent C, Blackwell L, Emberson J, Holland LE, Reith C, Bhala N, Peto R, Barnes EH, Keech A, Simes J, Collins R. Efficacy and safety of more intensive lowering of ldl cholesterol: A meta‐analysis of data from 170,000 participants in 26 randomised trials. Lancet. 2010;376:1670‐1681

46. Mancini GBJ HR, Leiter LA. Canadian diabetes association 2013 clinical practice guidelines for the prevention and management of diabetes in canada: Dyslipidemia. Canadian Journal of Diabetes. 2013;37:S110‐S116

REFERENCES

Page 89 of 93

47. Singh U, Devaraj S, Jialal I. Vitamin e, oxidative stress, and inflammation. Annual review of nutrition. 2005;25:151‐174

48. Blaier O, Lishner M, Elis A. Managing statin‐induced muscle toxicity in a lipid clinic. Journal of clinical pharmacy and therapeutics. 2011;36:336‐341

49. Saha SA, Kizhakepunnur LG, Bahekar A, Arora RR. The role of fibrates in the prevention of cardiovascular disease‐‐a pooled meta‐analysis of long‐term randomized placebo‐controlled clinical trials. American heart journal. 2007;154:943‐953

50. Group HTC. Hps2‐thrive randomized placebo‐controlled trial in 25 673 high‐risk patients of er niacin/laropiprant: Trial design, pre‐specified muscle and liver outcomes, and reasons for stopping study treatment. European heart journal. 2013;34:1279‐1291

51. Keys A, Menotti A, Karvonen MJ, Aravanis C, Blackburn H, Buzina R, Djordjevic BS, Dontas AS, Fidanza F, Keys MH, et al. The diet and 15‐year death rate in the seven countries study. American journal of epidemiology. 1986;124:903‐915

52. Hu FB. Globalization of food patterns and cardiovascular disease risk. Circulation. 2008;118:1913‐1914

53. Jenkins DJ, Jones PJ, Lamarche B, Kendall CW, Faulkner D, Cermakova L, Gigleux I, Ramprasath V, de Souza R, Ireland C, Patel D, Srichaikul K, Abdulnour S, Bashyam B, Collier C, Hoshizaki S, Josse RG, Leiter LA, Connelly PW, Frohlich J. Effect of a dietary portfolio of cholesterol‐lowering foods given at 2 levels of intensity of dietary advice on serum lipids in hyperlipidemia: A randomized controlled trial. JAMA : the journal of the American Medical Association. 2011;306:831‐839

54. Estruch R, Ros E, Salas‐Salvado J, Covas MI, Corella D, Aros F, Gomez‐Gracia E, Ruiz‐Gutierrez V, Fiol M, Lapetra J, Lamuela‐Raventos RM, Serra‐Majem L, Pinto X, Basora J, Munoz MA, Sorli JV, Martinez JA, Martinez‐Gonzalez MA, Investigators PS. Primary prevention of cardiovascular disease with a mediterranean diet. The New England journal of medicine. 2013;368:1279‐1290

55. Hu FB, Rimm EB, Stampfer MJ, Ascherio A, Spiegelman D, Willett WC. Prospective study of major dietary patterns and risk of coronary heart disease in men. The American journal of clinical nutrition. 2000;72:912‐921

56. Astrup A, Dyerberg J, Elwood P, Hermansen K, Hu FB, Jakobsen MU, Kok FJ, Krauss RM, Lecerf JM, LeGrand P, Nestel P, Riserus U, Sanders T, Sinclair A, Stender S, Tholstrup T, Willett WC. The role of reducing intakes of saturated fat in the prevention of cardiovascular disease: Where does the evidence stand in 2010? The American journal of clinical nutrition. 2011;93:684‐688

57. Hu FB. Plant‐based foods and prevention of cardiovascular disease: An overview. The American journal of clinical nutrition. 2003;78:544S‐551S

58. Osler M, Heitmann BL, Gerdes LU, Jorgensen LM, Schroll M. Dietary patterns and mortality in danish men and women: A prospective observational study. The British journal of nutrition. 2001;85:219‐225

59. Fung TT, Rimm EB, Spiegelman D, Rifai N, Tofler GH, Willett WC, Hu FB. Association between dietary patterns and plasma biomarkers of obesity and cardiovascular disease risk. The American journal of clinical nutrition. 2001;73:61‐67

60. Heidemann C, Schulze MB, Franco OH, van Dam RM, Mantzoros CS, Hu FB. Dietary patterns and risk of mortality from cardiovascular disease, cancer, and all causes in a prospective cohort of women. Circulation. 2008;118:230‐237

61. Rao V, Al‐Weshahy A. Plant‐based diets and control of lipids and coronary heart disease risk. Current atherosclerosis reports. 2008;10:478‐485

62. Key TJ, Appleby PN, Rosell MS. Health effects of vegetarian and vegan diets. The Proceedings of the Nutrition Society. 2006;65:35‐41

REFERENCES

Page 90 of 93

63. Jenkins DJ, Kendall CW, Marchie A, Jenkins AL, Augustin LS, Ludwig DS, Barnard ND, Anderson JW. Type 2 diabetes and the vegetarian diet. The American journal of clinical nutrition. 2003;78:610S‐616S

64. Hu FB. Protein, body weight, and cardiovascular health. The American journal of clinical nutrition. 2005;82:242S‐247S

65. Knekt P, Reunanen A, Jarvinen R, Seppanen R, Heliovaara M, Aromaa A. Antioxidant vitamin intake and coronary mortality in a longitudinal population study. American journal of epidemiology. 1994;139:1180‐1189

66. Hung HC, Joshipura KJ, Jiang R, Hu FB, Hunter D, Smith‐Warner SA, Colditz GA, Rosner B, Spiegelman D, Willett WC. Fruit and vegetable intake and risk of major chronic disease. Journal of the National Cancer Institute. 2004;96:1577‐1584

67. Sauvaget C, Nagano J, Hayashi M, Yamada M. Animal protein, animal fat, and cholesterol intakes and risk of cerebral infarction mortality in the adult health study. Stroke; a journal of cerebral circulation. 2004;35:1531‐1537

68. Iso H, Stampfer MJ, Manson JE, Rexrode K, Hu F, Hennekens CH, Colditz GA, Speizer FE, Willett WC. Prospective study of fat and protein intake and risk of intraparenchymal hemorrhage in women. Circulation. 2001;103:856‐863

69. Anderson JW, Johnstone BM, Cook‐Newell ME. Meta‐analysis of the effects of soy protein intake on serum lipids. The New England journal of medicine. 1995;333:276‐282

70. Zhang X, Shu XO, Gao YT, Yang G, Li Q, Li H, Jin F, Zheng W. Soy food consumption is associated with lower risk of coronary heart disease in chinese women. The Journal of nutrition. 2003;133:2874‐2878

71. Jenkins DJ, Srichaikul K, Wong JM, Kendall CW, Bashyam B, Vidgen E, Lamarche B, Rao AV, Jones PJ, Josse RG, Jackson CJ, Ng V, Leong T, Leiter LA. Supplemental barley protein and casein similarly affect serum lipids in hypercholesterolemic women and men. The Journal of nutrition. 2010;140:1633‐1637

72. Jenkins DJ, Wong JM, Kendall CW, Esfahani A, Ng VW, Leong TC, Faulkner DA, Vidgen E, Greaves KA, Paul G, Singer W. The effect of a plant‐based low‐carbohydrate ("eco‐atkins") diet on body weight and blood lipid concentrations in hyperlipidemic subjects. Archives of internal medicine. 2009;169:1046‐1054

73. Nation WHOFaAOotU. Cereals, pulses, legumes and vegetable proteins, 1st ed. Rome, Italy; 2007.

74. Organization WH. The world health report 2010. 2010 75. Canada HMtQiRo. Research gaps and priorities related to health claim validation in: Agriculture

and agri‐food canada. 2010 76. American Heart Association Nutrition C, Lichtenstein AH, Appel LJ, Brands M, Carnethon M,

Daniels S, Franch HA, Franklin B, Kris‐Etherton P, Harris WS, Howard B, Karanja N, Lefevre M, Rudel L, Sacks F, Van Horn L, Winston M, Wylie‐Rosett J. Diet and lifestyle recommendations revision 2006: A scientific statement from the american heart association nutrition committee. Circulation. 2006;114:82‐96

77. (EASD) DaNSGDotEAftSoD. Evidence‐based nutritional approaches to the treatment and prevention of diabetes mellitus. Nutrition, Metabolism & Cardiovascular Diseases. 2004;14:373‐394

78. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jr., Jones DW, Materson BJ, Oparil S, Wright JT, Jr., Roccella EJ, Joint National Committee on Prevention DE, Treatment of High Blood Pressure. National Heart L, Blood I, National High Blood Pressure Education Program Coordinating C. Seventh report of the joint national committee on prevention, detection, evaluation, and treatment of high blood pressure. Hypertension. 2003;42:1206‐1252

REFERENCES

Page 91 of 93

79. CHEP H. 2013 chep recommendations: Lifestyle management. 2013 80. Bernstein AM, Sun Q, Hu FB, Stampfer MJ, Manson JE, Willett WC. Major dietary protein sources

and risk of coronary heart disease in women. Circulation. 2010;122:876‐883 81. Bazzano LA, He J, Ogden LG, Loria C, Vupputuri S, Myers L, Whelton PK. Legume consumption

and risk of coronary heart disease in us men and women: Nhanes i epidemiologic follow‐up study. Archives of internal medicine. 2001;161:2573‐2578

82. Mitchell DC, Lawrence FR, Hartman TJ, Curran JM. Consumption of dry beans, peas, and lentils could improve diet quality in the us population. Journal of the American Dietetic Association. 2009;109:909‐913

83. Anderson JW, Major AW. Pulses and lipaemia, short‐ and long‐term effect: Potential in the prevention of cardiovascular disease. The British journal of nutrition. 2002;88 Suppl 3:S263‐271

84. Siri‐Tarino PW, Sun Q, Hu FB, Krauss RM. Saturated fat, carbohydrate, and cardiovascular disease. The American journal of clinical nutrition. 2010;91:502‐509

85. Anderson JW, Gustafson NJ, Spencer DB, Tietyen J, Bryant CA. Serum lipid response of hypercholesterolemic men to single and divided doses of canned beans. The American journal of clinical nutrition. 1990;51:1013‐1019

86. Abeysekara S, Chilibeck PD, Vatanparast H, Zello GA. A pulse‐based diet is effective for reducing total and ldl‐cholesterol in older adults. The British journal of nutrition. 2012;108 Suppl 1:S103‐110

87. Sievenpiper JL, Kendall CW, Esfahani A, Wong JM, Carleton AJ, Jiang HY, Bazinet RP, Vidgen E, Jenkins DJ. Effect of non‐oil‐seed pulses on glycaemic control: A systematic review and meta‐analysis of randomised controlled experimental trials in people with and without diabetes. Diabetologia. 2009;52:1479‐1495

88. Jenkins DJ, Ghafari H, Wolever TM, Taylor RH, Jenkins AL, Barker HM, Fielden H, Bowling AC. Relationship between rate of digestion of foods and post‐prandial glycaemia. Diabetologia. 1982;22:450‐455

89. Hanefeld M, Cagatay M, Petrowitsch T, Neuser D, Petzinna D, Rupp M. Acarbose reduces the risk for myocardial infarction in type 2 diabetic patients: Meta‐analysis of seven long‐term studies. European heart journal. 2004;25:10‐16

90. Abete I, Parra D, Martinez JA. Legume‐, fish‐, or high‐protein‐based hypocaloric diets: Effects on weight loss and mitochondrial oxidation in obese men. Journal of medicinal food. 2009;12:100‐108

91. Streppel MT, Arends LR, van 't Veer P, Grobbee DE, Geleijnse JM. Dietary fiber and blood pressure: A meta‐analysis of randomized placebo‐controlled trials. Archives of internal medicine. 2005;165:150‐156

92. FDA. Evidence‐based review system for the scientific evaluation of health claims. 2009 93. Winham DM, Hutchins A.M. Baked bean consumption reduces serum cholesterol in

hypercholesterolmeic adults. Nutrition Research. 2007;27:380‐386 94. Winham DM, Hutchins AM, Johnston CS. Pinto bean consumption reduces biomarkers for heart

disease risk. Journal of the American College of Nutrition. 2007;26:243‐249 95. USDA. Dietary guidelines for americans. Washington D.C.: US Government Printing Office; 2010. 96. Association AH. How do i follow a healthy diet? 2010;2013 97. JPT H. Cochrane handbook for systematic review of interventions Great Britain: Antony Rowe

Ltd; 2012. 98. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic

reviews and meta‐analyses: The prisma statement. Bmj. 2009;339:b2535

REFERENCES

Page 92 of 93

99. Jimenez‐Cruz A TW, Bacardi‐Gascon M, Rosales‐Garay P. A high‐fiber, moderate‐glycemic‐index, mexican style diet improves dyslipidemia in individuals with type 2 diabetes. Nutrition Research. 2004;24:19‐27

100. Heyland DK, Novak F, Drover JW, Jain M, Su X, Suchner U. Should immunonutrition become routine in critically ill patients? A systematic review of the evidence. JAMA : the journal of the American Medical Association. 2001;286:944‐953

101. Elbourne DR, Altman DG, Higgins JP, Curtin F, Worthington HV, Vail A. Meta‐analyses involving cross‐over trials: Methodological issues. International journal of epidemiology. 2002;31:140‐149

102. Marinangeli CP, Jones PJ. Whole and fractionated yellow pea flours reduce fasting insulin and insulin resistance in hypercholesterolaemic and overweight human subjects. The British journal of nutrition. 2011;105:110‐117

103. Gormley TR KJ, O'Donnell B, McFarlane R. Effect of peas on serum cholesterol levels in humans. Irish Journal of Food Science and Technology. 1979;3:101‐109

104. Jenkins DJ, Wolever TM, Rao AV, Hegele RA, Mitchell SJ, Ransom TP, Boctor DL, Spadafora PJ, Jenkins AL, Mehling C, et al. Effect on blood lipids of very high intakes of fiber in diets low in saturated fat and cholesterol. The New England journal of medicine. 1993;329:21‐26

105. WH C. Key determinants of the health and well‐being of men and boys. International Journal of Men's Health. 2003;2:1‐30

106. Reyna‐Villasmil N, Bermudez‐Pirela V, Mengual‐Moreno E, Arias N, Cano‐Ponce C, Leal‐Gonzalez E, Souki A, Inglett GE, Israili ZH, Hernandez‐Hernandez R, Valasco M, Arraiz N. Oat‐derived beta‐glucan significantly improves hdlc and diminishes ldlc and non‐hdl cholesterol in overweight individuals with mild hypercholesterolemia. American journal of therapeutics. 2007;14:203‐212

107. Association NPG. Pulses: The perfect food. 2012;2013 108. Miller ER, 3rd, Erlinger TP, Sacks FM, Svetkey LP, Charleston J, Lin PH, Appel LJ. A dietary pattern

that lowers oxidative stress increases antibodies to oxidized ldl: Results from a randomized controlled feeding study. Atherosclerosis. 2005;183:175‐182

109. Botero D, Ebbeling CB, Blumberg JB, Ribaya‐Mercado JD, Creager MA, Swain JF, Feldman HA, Ludwig DS. Acute effects of dietary glycemic index on antioxidant capacity in a nutrient‐controlled feeding study. Obesity. 2009;17:1664‐1670

110. Jentzsch AM, Bachmann H, Furst P, Biesalski HK. Improved analysis of malondialdehyde in human body fluids. Free radical biology & medicine. 1996;20:251‐256

111. Ahotupa M, Ruutu M, Mantyla E. Simple methods of quantifying oxidation products and antioxidant potential of low density lipoproteins. Clinical biochemistry. 1996;29:139‐144

112. Hu ML. Measurement of protein thiol groups and glutathione in plasma. Methods in enzymology. 1994;233:380‐385

113. Jenkins DJ, Kendall CW, Jackson CJ, Connelly PW, Parker T, Faulkner D, Vidgen E, Cunnane SC, Leiter LA, Josse RG. Effects of high‐ and low‐isoflavone soyfoods on blood lipids, oxidized ldl, homocysteine, and blood pressure in hyperlipidemic men and women. The American journal of clinical nutrition. 2002;76:365‐372

114. Ceriello A, Esposito K, Piconi L, Ihnat M, Thorpe J, Testa R, Bonfigli AR, Giugliano D. Glucose "peak" and glucose "spike": Impact on endothelial function and oxidative stress. Diabetes research and clinical practice. 2008;82:262‐267

115. Ceriello A, Esposito K, Piconi L, Ihnat MA, Thorpe JE, Testa R, Boemi M, Giugliano D. Oscillating glucose is more deleterious to endothelial function and oxidative stress than mean glucose in normal and type 2 diabetic patients. Diabetes. 2008;57:1349‐1354

116. RH L. Whole grain phytochemicals and health. Journal of Cereal Science. 2007;46:207‐219 117. USDA. Oxygen radical absorbance capacity (orac) of selected foods, release 2 (2010). 2012

REFERENCES

Page 93 of 93

118. Goff LM, Cowland DE, Hooper L, Frost GS. Low glycaemic index diets and blood lipids: A systematic review and meta‐analysis of randomised controlled trials. Nutrition, metabolism, and cardiovascular diseases : NMCD. 2013;23:1‐10

119. Pan A, Sun Q, Bernstein AM, Schulze MB, Manson JE, Stampfer MJ, Willett WC, Hu FB. Red meat consumption and mortality: Results from 2 prospective cohort studies. Archives of internal medicine. 2012;172:555‐563

120. Omenn GS, Goodman GE, Thornquist MD, Balmes J, Cullen MR, Glass A, Keogh JP, Meyskens FL, Valanis B, Williams JH, Barnhart S, Hammar S. Effects of a combination of beta carotene and vitamin a on lung cancer and cardiovascular disease. The New England journal of medicine. 1996;334:1150‐1155

121. Lonn E, Bosch J, Yusuf S, Sheridan P, Pogue J, Arnold JM, Ross C, Arnold A, Sleight P, Probstfield J, Dagenais GR, Hope, Investigators H‐TT. Effects of long‐term vitamin e supplementation on cardiovascular events and cancer: A randomized controlled trial. JAMA : the journal of the American Medical Association. 2005;293:1338‐1347

Documents

Effects of Dietary Pulses on Factors of · Effects of Dietary Pulses on ... DASH‐ Dietary Approaches to Stopping Hypertension ... of low‐glycemic index8, Dietary Strategies to