DietandCVDriskfactorsinBotswana Kwape PhDthesis2012

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/305481199
Diet and cardiovascular disease risk factors in Botswana_PhD thesis
Book · January 2012
CITATION READS
1 2,579
1 author:
Lems Kwape
National Food Technology Research Center
8 PUBLICATIONS 99 CITATIONS
SEE PROFILE
All content following this page was uploaded by Lems Kwape on 21 July 2016.
The user has requested enhancement of the downloaded file.

Diet and cardiovascular disease risk factors in Botswana
Lemogang Daniel Kwape
BS. Dietetics, Texas Southern University (Houston, Texas USA)
MSc. Nutrition & Health (Public Health & Epidemiology) Wageningen, The
Netherlands.
A thesis presented for the degree of
Doctor of Philosophy
in the Public Health Nutrition Research Group, Institute of Applied Health

Sciences, University of Aberdeen
June 2012
Declaration
Declaration
I declare that the work in this thesis has been undertaken by myself, and has not been
submitted in support for a degree from any other university. All work was carried out
by myself with assistance from individuals and organisations acknowledged here.
Bantle Modibedi and Tapologo Ramotswaiso were involved in recruitment and
measurement of participants along with myself. Blood samples were collected and
analysed by staff at Diagnofirm Medical Laboratories in Gaborone Botswana. The Food
Frequency Questionnaire was analysed by Dr Maria Jackson, University of Jamaica at
Mona. All quotation marks and all sources of information have been acknowledged.
...............................................
Lemogang Daniel Kwape
June, 2012
i|P a ge
Table of Contents
Table of Contents
Declaration i
Abstract vi
Acknow ledgements vii
List of abbreviations viii
List of figures ix
List of tables x
Summary xiii
1 Chapter 1 2
1.1 Introduction 2
1.1.1 Definitions 2
1.1.2 Morbidity and mortality 4
1.2 Risk factors fo r cardiovascular disease 6

1.2.1 Sex 8
1.2.2 Blood pressure 9
1.2.3 Diabetes mellitus 11
1.2.4 Blood lipids 13
1.2.5 Physical activity 16
1.2.6 Obesity 17
1.2.7 Alcohol 19
1.2.8 Smoking 22
1.3 Nutrition transition 23
1.4 Diet and risk of CVD 23

1.4.1 Dietary fat 24
1.4.2 Dietary fibre 27
1.4.3 Sodium and potassium 29
1.4.4 Antioxidant nutrients 30
1.4.5 Fruit and vegetable consumption 31
1.4.6 Dietary patterns 32
1.5 Antiretroviral medication and ca rdiovascular diseases 33
1.6 Diet and CVD in Botswana 34
1.7 Aims and objectives 40
2 Methodology 42
2.1 Introduction to Bo tswana 42
2.2 Roles in the study 44
2.3 Ethical considerations 45

ii | P a g e
Table of Contents
2.4 Recruitment centres 45
2.5 Study participants 46
2.6 Sample size 47
2.7 Questionnaire 48
2.7.1 Socioeconomic status 48
2.7.2 Education 48
2.7.3 History of cardiovascular diseases 49
2.7.4 Smoking 49
2.7.5 Physical activity questionnaire 51
2.7.6 Use of hormone replacements or birth control 52
2.7.7 Food frequency questionnaire (FFQ) 53
2.8 Biophysical measurements 54

2.8.1 Body weight and height 54
2.8.2 Waist and hip circumference 55
2.8.3 Body composition 56
2.8.4 Blood pressure and pulse rate 58
2.9 Collection of biological samples and initial bio chemistry 58

2.9.1 Quality control 59
2.10 Statistical Analyses 61

2.10.1 Descriptive statistics 61
3 Diet and CVD risk factors 64
3.1 Introduction 64
3.2 Statistical methods 65

3.2.1 Descriptive statistics 65
3.2.2 Dietary patterns 68
3.3 Socio-demographic characteristics 70

3.3.1 Socio-demographic characteristics by sex 70
3.3.2 Socio-demographic characteristics by age group 75
3.4 Anthropometric and physical chara cteristics of participants 79
3.5 Lipids, Lipoproteins and glucose measurements of participants 87
3.6 Dietary Data 91

3.6.1 Under and over reporting 91
3.6.2 Macronutrient intake 93
3.6.3 Proportion of individuals commonly consuming specific foods 99
3.6.4 Intakes of commonly consumed foods 110
3.6.5 Dietary pattern analyse 120
3.7 Discussion 126

3.7.1 Obesity 126
3.7.2 Hypertension 131
3.7.3 Dyslipidaemia 132
3.7.4 Diet and dietary patterns 134
4 Associations between diet and CVD risk factors 138
iii | P a g e
Table of Contents
4.2 Correla tions between diet and CVD risk factors 141
4.3 Association between total fa t intake and CVD risk factors 143
4.4 Association between SFA intake and CVD risk fa cto rs 149
4.5 Association between PUFA intake and CVD risk factors 155
4.6 Association between PUFA:SFA ratio and CVD risk fa cto rs 160
4.7 Association between dietary fibre intake and CVD risk factors 165
4.8 Association between alcohol intake and CVD risk factors 170
4.9 Association between dietary patterns and CVD risk factors 176
4.9.1 Association between dietary patterns and CVD risk factors in women 176
4.9.2 Association between dietary patterns and CVD risk factors in men 181
4.10 Association between frui t and vegetables intake and blood pressu re 193
4.11 Discussion 198

4.11.1 Saturated fat intake and CVD risk factors 198
4.11.2 Polyunsaturated fatty acids intake and CVD risk factors 200
4.11.3 Fibre intake and CVD risk factors 201
4.11.4 Alcohol intake and CVD risk factors 202
4.11.5 Dietary patterns and CVD risk factors 204
4.11.6 Fruit and vegetable intake and blood pressure 205
5 Diet and risk of CVD/diabetes/hypertension 208
5.1 Introduction 208
5.3 Comparison between participants with CVD, hypertension or diabetes 214
5.4 Characteristics of “healthy” and “diseased ” participants 219

5.4.1 Socio-demographic characteristics 219
5.4.2 Physical and anthropometric measurements 221
5.4.3 Macronutrient intake 223
5.4.4 Lipids, lipoproteins and glucose 224
5.5 Risk of CVD/diabetes/hypertension by dietary factors 226

5.5.1 Carbohydrate intake 226
5.5.2 Fibre intake 226
5.5.3 Total fat intake 227
5.5.4 Saturated fatty acid intake 227
5.5.5 Polyunsaturated fatty acids intake 227
5.5.6 Polyunsaturated fatty acids and saturated fatty acid ratio (PUFA:SFA ratio) 227
5.5.7 Alcohol use 227
5.6 Discussion 236
6 Discussion 241
6.1 Strengths 242
6.2 Limitations 243
6.3 Major findings 245

6.3.1 CVD risk factors 245
6.3.2 Association between diet and CVD risk factors 248
iv | P a g e
Table of Contents
6.3.3 Diet and disease 254
6.4 Future research wo rk focus 257
6.5 Public health perspective 261
6.6 Conclusion 263
Appendix 1 Ethics permit i
Appendix 2 Information sheet (English) i
Appendix 2 Information sheet (Setswana) ii
Appendix 3 Questionnaire iii
Appendix 4 Description of food items listed in the FFQ xxviii
Appendix 5 Publication xxxii
v|P a ge
Abstract
Abstract
Kwape LD: Diet and cardiovascular disease risk factors in Botswana
Cardiovascular disease (CVD) is the leading cause of mortality and morbidity wo rldwide.
In Sub-Saharan Africa, rates of CVD are increasing rapidly, but there is little evidence
about the potential determinants of CVD risk in this population.
This thesis investigated CVD risk factors in Gaborone, capital city of Botswana, by (i)
documenting CVD risk factors in this population, (ii) investigating the association between
diet and CVD risk factors and (iii) assessing the association between diet and risk of CVD.
787 adults were recruited. Of these 566 were generally “healthy” with no his tory of CVD,
while 221 (“diseased”) had at least one reported CVD condition, hypertension or
diabetes. The median (interquartile range) age was 27 (23, 32) and 52 (42, 62) years for
healthy and diseased participants respectively. All participants completed an interview
administered questionnaire, including a food frequency questionnaire. Height, weight,
waist circumference and blood pressure were measured, and a non-fasting blood sample
was obtained for analysis of lipids, lipoproteins and glucose.
A high prevalence of overweight and obesity (36.8%), particularly in women (50.0%), and
low HDL cholesterol (<1.0 mmol/L men and <1.3 mmol/L women) (62.6%) was found.
High levels of triglycerides, LDL cholesterol, glucose and high blood pressure were also
found in this population of young adults in Gaborone.
Total fat and/or saturated fat intake (as percentage energy) was significantly linearly
associated with increased LDL cholesterol (p=0.017), triglycerides (p=0.048), glucose
(p=0.044) and with decreased HDL cholesterol (p=0.021). However, fibre, polyunsaturated
fatty acids and dietary patterns were not independently associated with CVD risk factors.
Carbohydrates intake was significantly associated with increased risk of disease.

Unexpectedly, saturated fat intake was associated with reduced disease risk, but
weakened after nutrients adjustment.
CVD risk factors are relatively high in this population. These results suggest a need for
further research on CVD in Botswana.
vi | P a g e
Acknowledgements
Acknowledgements
Many people contributed in many ways to the success of the research work described
in this thesis.
First I would like to thank the participants who volunteered their time and blood in the
advancement of science.
I am grateful to Professor Geraldine McNeill for her guidance and encouragement

throughout my PhD. Thank you for always seeing the glass half full.
Dr Lindsey Masson, thank you for your attention to details and for your guidance and
encouragement.
Dr Nadeem Sarwar, many thanks for your guidance and invaluable contribution in the
formulation and initiation of the study.
Dr Lorna Aucott, thank you for your advice on analysis of the data.
Thank you to Professor Kiran Bhagat and the staff at Heart Foundation of Botswana for
your support during planning and data collection of this study.
Phlebotomists at DML, thank for your services during the data collection. It was not
easy but we made it.
Thank you to the nurses and support staff at Cardiac, Extension 2 and Nkoyaphiri
clinics.
Bantle Modibedi and Tapologo Ramotswaiso thank you for your assistance during the
data collection.
I am immensely grateful to the Department of Research Science and Technology,

University of Aberdeen, Diagnofirm Medical Laboratory & the Dennis Burkitt fellowship
for providing funding for my PhD.
Staff and fellow PhD students (past and present) in the Public Health Nutrition
Research Group, thank you for your assistance during my PhD. You shared, you cared
and the group will remain close to my heart.
Finally to my family, thank you for your immeasurable support. I owe my success to
you and with this thesis I humbly thank you.
vii | P a g e
Abbreviations
List of abbreviations
AIDS Acquired Immune Deficiency Syndrome
ANOVA Analysis of Variance
ARV Anti-Retro Viral
BFHS Botswana Family Health Survey
BIA Bio-Impedance Analysis
BMI Body Mass Index
BMR Basal Metabolic Rate
BWP Botswana Pula (1 BWP~ £0.11)
CDE Carbohydrates-Deficiency Transferrin
CHD Coronary Heart Disease
CSO Central Statistics Office
CVD Cardiovascular disease
DBP Diastolic Blood Pressure
DM Diabetes Mellitus
DALY Disability Adjusted Life Years
DML Diagnofirm Medical Laboratories
FBG Fasting Blood Glucose
FFQ Food Frequency Questionnaire
GGT Gamma-Glutamyl Transferase
GLM General Linear Model
HC Hip Circumference
HDL High Density Lipoprotein
HIV Human Immuno Virus
IHD Ischaemic Heart Disease
IQR Inter Quartile Range
LDL Low-Density Lipoprotein
LMIC Low Middle Income Countries
MI Myocardial Infarction
MRC Medical Research Council
MRFIT Multi Risk Factors Intervention Trial
MUFA Monounsaturated Fatty Acids
OR Odds Ratio
PAI Physical Activity Index
PAL Physical Activity Level
PURE Prospective Urban and Rural Epidemiology
RCT Randomised Controlled Trials
RR Relative Risk
SACN Scientific Advisory Committee on Nutrition
SBP Systolic Blood Pressure
SD Standard Deviation
SFA Saturated Fatty Acids
SPSS Statistical Package for Social Sciences
PCA Principal Component Analysis
P:S ratio PUFA:SFA ratio
PUFA Polyunsaturated Fatty Acids
THUSA Transition and Health during Urbanisation in South Africa
WHO World Health Organisation
WHR Waist-to-Hip ratio
viii | P a g e
Abbreviations
List of figures
Figure 1.1 Complex lesion and thrombosis in atherosclerosis 3

Figure 1.2 CVD mortality trends for males and females (United States 1979-2007) 4
Figure 1.3 Global mortality and risk factors 6
Figure 1.4 Conceptual framework of development of CVD 7
Figure 1.5 CVD mortality for men and women in the USA 8
Figure 1.6 Prevalence of CVD risk factors by gender in Tanzania. 9
Figure 1.7 Prevalence of hypertension (≥ 140/90mmHg) in Sub-Saharan Africa 10
Figure 1.8 Age- and sex-adjusted prevalence of hypertension (HTN), smoking, elevated
cholesterol, obesity, and diabetes among CVD cases in the early versus the later ti me
period of the Framingham study. 12
Figure 2.1 Map of Botswana showing population distribution 43
Figure 2.2 Sample of smoking history questionnaire 50
Figure 2.3 Sample of the FFQ 53
Figure 2.4 Measuring waist and hip circumference 56
Figure 2.5 Measuring body composition with bio-impedance analysis 57
Figure 3.1 Age distribution of participants 71
Figure 3.2 Income categories of participants by sex 74
Figure 3.3 Income of participants by age group 78
Figure 3.4 Prevalence of underweight, overweight and obesity by BMI 81
Figure 3.5 Prevalence of abdominal obesity 82
Figure 3.6 Prevalence of hypertension 85
Figure 3.7 Proportion of physical activity levels of participants 86
Figure 3.8 Prevalence of low HDL & high atherogenic index 89
Figure 3.9 Proportion of commonly consumed food items 100
Figure 3.10 Fat and oils used in food preparation 119
Figure 5.1 Distribution of participants by disease status 208
Figure 5.2 Age distribution of participants by disease status 210
Figure 5.3 Proportion of participants with overweight and obesity 222
Figure 5.4 Proportion of participants with hypertension 222
Figure 5.5 Proportion of participants with high lipids, lipoproteins and glucose 225
ix | P a g e
List of tables
List of tables
Table 1.1 BMI categories (WHO, 2000) 18
Table 1.2 WC categories for risk of metabolic complications (WHO, 2000) 18
Table 1.3 Description of Mediterranean, prudent western dietary and oriental patterns
32
Table 1.4 Literature search strategy for diet and CVD in Botswana 35
Table 1.5 Studies measuring CVD risk factors and diet in Botswana 37
Table 2.1 Physical activity status and weight factors 51
Table 2.2 Physical activity levels for different lifestyles 52
Table 3.1 Socio-demographic characteristics by sex: n (%) 71
Table 3.2 Lifestyle characteristics of participants by sex: n (%) 73
Table 3.3 Socio-demographic characteristics of participants by age group: n (%) 76
Table 3.4 Lifestyle characteristics of participants by age group: n (%) 77
Table 3.5 Mean (SD) anthropometric measurements by age and sex 80
Table 3.6 Mean (SD) of physical measurements 84
Table 3.7 Mean (SD) of lipids, lipoproteins and glucose (mmol/L) 88
Table 3.8 Number (%) of participants classed as under and over reporters 92
Table 3.9 Variation of selected variables by energy cut-off points: Mean (SD) 93
Table 3.10 Daily energy, carbohydrate, fibre and protein intake: Median (P25, P75) 95
Table 3.11 Daily fat and alcohol intake (grams) of participants: Median (P25, P75) 96
Table 3.12 Mean (SD) percentage contribution of macronutrients daily to energy 98
Table 3.13 Proportion n (%) of individuals who commonly consumed cereals and
starches by age group and sex 103
Table 3.14 Proportion (%) of individuals who commonly consumed meat and dairy
products by age group and sex 105
Table 3.15 Proportion n(%) of individuals who commonly consumed fruit and
vegetables by age group and sex 107
Table 3.16 Proportion n (%) of individuals who commonly consumed sugar sweets and
condiments by age group and sex 109
Table 3.17 Median (P25, P75) intake of commonly consumed cereals and starchy
foods (g/day) by sex and age group 111
Table 3.18 Median (P25, P75) intake of commonly consumed dairy and meat products
(g/day) by sex and age group 112
Table 3.19 Median (P25, P75) intake of commonly consumed vegetables (g/day) by
sex and age group 114
Table 3.20 Median (P25, P75) intake of commonly consumed fruit (g/day) by sex and
age group 115
Table 3.21 Median (P25, P75) intake of commonly consumed sugar, sweets,
condiments (g/day) by sex and age group 117
Table 3.22 Median (P25, P75) intake of commonly consumed beverages (ml/day) by
sex and age group 118
Table 3.23 Dietary patterns for all participants 121
Table 3.24 Dietary patterns for all women 122
Table 3.25 Dietary pattern for all men 123
Table 3.26 Results from WHO steps survey in selected sub-Saharan countries 127
Table 3.27 Participants characteristics by BMI classification 130
Table 4.1 Correlations between diet and CVD risk factors 142
x|P a ge
List of tables
Table 4.2 Association between lipid, lipoprotein and glucose levels mean (SE) and total
fat intake (%E) in all participants 144
Table 4.3 Association between lipid, lipoprotein and glucose levels mean (SE) with
total fat intake (%E) in men and women 145
Table 4.4 Association between physical measurements mean (SE) and total fat intake
(%E) in all participants 147
Table 4.5 Association between physical measurements mean (SE) and total fat intake
(%E) in men and women 148
Table 4.6 Association between lipid lipoprotein and glucose levels mean (SE) and SFA
intake (%E) in all participants 150
intake (%E) in men and women 151
Table 4.8 Association between physical measurements mean (SE) and SFA (%E) in all
participants 153
Table 4.9 Association between physical measurements mean (SE) and SFA (%E) in men
and women 154
Table 4.10 Association between lipid lipoprotein and glucose levels mean (SE) and
PUFA (%E) for all participants 156
PUFA (%E) for men and women 157
Table 4.12 Association between physical measurements mean (SE) and PUFA (%E) in all
participants 158
Table 4.13 Association between physical measurements mean (SE) and PUFA (%E) in
men and women 159
PUFA:SFA ratio in all participants 161
PUFA:SFA ratio in men and women 162
Table 4.16 Association between physical measurements mean (SE) and PUFA:SFA ratio
in all participants 163
Table 4.17 Association between physical measurements mean (SE) and PUFA:SFA ratio
in men and women 164
dietary fibre (g/1000kcal) in all participants 166
dietary fibre (g/1000kcal) in men and women 167
Table 4.20 Association between physical measurements mean (SE) and dietary fibre
(g)/1000kcal in all participants 168
(g)/1000kcal in men and women 169
Table 4.22 Association between lipids, lipoproteins and glucose mean (SE) and alcohol
use (g) in all participants 171
Table 4.23 Association between lipids, lipoproteins and glucose mean (SE) and alcohol
use (g) in men and women 172
Table 4.24 Association between physical measurements mean (SE) and alcohol use (g)
in all participants 174
Table 4.25 Association between physical measurements mean (SE) and alcohol use (g)
in men and women 175
xi | P a g e
List of tables
Table 4.26 Association between lipids, lipoproteins and glucose and the “high sweets”
pattern 177
Table 4.27 Association between physical measurements and the “high sweets” pattern
178
Table 4.28 Association between lipids lipoproteins and glucose and the “high fruit”
pattern 179
Table 4.29 Association between physical measurements and the “high fruit” pattern
180
Table 4.30 Association between lipids, lipoproteins and glucose and the “mixed”
pattern 182
Table 4.31 Association between physical measurements and the “mixed” pattern 183
Table 4.32 Association between lipids, lipoproteins and glucose and the “convenience”
pattern 185
Table 4.33 Association between physical measurements and the “convenience”
pattern 186
Table 4.34 Association between physical measurements and the “traditional” pattern
188
189
Table 4.36 Association between lipids, lipoproteins and glucose and the “high
vegetable” pattern 191
Table 4.37 Association between physical measurements and the “high vegetable”
pattern 192
Table 4.38 Association between fruit and vegetables and blood pressure 194
Table 4.39 Association between fruit and vegetables and blood pressure 195
Table 5.1 Pearson correlation coefficient for nutrients, physical and biochemical
measurements 213
Table 5.2 Socio-demographic characteristics of participants by disease status: n (%) 216
Table 5.3 Mean (SD) physical and anthropometric measurements by disease status 217
Table 5.4 Daily mean (SD) macronutrient intake (%E) of participants by disease status
217
Table 5.5 Mean (SD) biochemical indicators (mmol/L) by disease status 218
Table 5.6 Socio-demographic characteristics of participants by disease status: n (%) 220
Table 5.7 Mean (SD) physical and anthropometric measurements by disease status 221
Table 5.8 Daily mean (SD) nutrient intake (%E) of participants by disease status 223
Table 5.9 Median (P25, P75) biochemical indicators (mmol/L) by disease status 224
Table 5.10 OR (95% CI) for risk of CVD/diabetes/hypertension according to
carbohydrate (%E) intake 228
Table 5.11 OR (95% CI) for risk of CVD/diabetes/hypertension according to fibre
(g/1000kcal) intake 229
Table 5.12 OR (95% CI) for risk of CVD/diabetes/hypertension according to total fat
(%E) intake 230
Table 5.13 OR (95% CI) for risk of CVD/diabetes/hypertension by SFA (%E) intake 231
Table 5.14 OR (95% CI) for risk of CVD/diabetes/hypertension by PUFA (%E) intake 232
Table 5.15 OR (95% CI) for risk of CVD/diabetes/hypertension by PUFA:SFA ratio 233
Table 5.16 OR (95% CI) for the likelihood of having CVD/diabetes/hypertension by
alcohol use 234
xii | P a g e
Summary
Summary
Cardiovascular disease (CVD) is the leading cause of mortality and morbidity
worldwide. In Sub-Saharan Africa, rates of CVD appear to be increasing rapidly, yet
there is little evidence about the potential determinants of CVD risk in the region. This
thesis therefore sought to investigate CVD risk factors in Gaborone, capital city of
Botswana, by (i) documenting CVD risk factors in the population, (ii) investigating the
association between diet and CVD risk factors and (iii) assessing the association
between diet and CVD.
In this study 787 individuals aged ≥18 years were recruited. Of these, 566 were
generally “healthy” with no history of CVD, while 221 had at least one CVD condition,
hypertension or diabetes. All participants completed an interview administered
questionnaire, including a food frequency questionnaire, measurement of height,
weight, waist circumference and blood pressure, and a non-fasting blood sample was
obtained for analysis of lipids, lipoproteins and glucose.
In chapter 3 CVD risk factors in healthy 566 “healthy” participants were described.
Obesity measured by defined by BMI≥30 was high in women (6.9% in < 25 years old,
20.6% in 25-35 years old and 46.7% in >35 years). In men the prevalence of obesity
was <6% across all age groups and was lowest at 1% in men aged <25 years. At least 1
in 5 women aged >25 years had abdominal obesity and for those women aged >35
years almost 1 in 2 had abdominal obesity, while only 1 in 10 men aged 25-35 years
had abdominal obesity. Conversely, at least 15% of men 35 years or younger were
underweight (BMI <18.5 kg/m2). The prevalence of overweight and obesity was
significantly higher in older and married participants as well as in participants with a
xiii | P a g e
Summary
primary education or less, but was not significantly different by income or physical
activity level although physical activity level was significantly lower in women than in
men (p ≤ 0.001). About 1 in 4 men and women aged >35 years had high blood pressure
(SBP ≥ 140 mmHg or DBP ≥ 90 mmHg). Men had higher SBP compared to women (p ≤
0.001) and both SBP and DBP increased significantly with age.
Sixty-three percent of participants had a low HDL cholesterol level (< 1.0 mmol/L men
& < 1.3 mmol/L women), and more women than men had low HDL cholesterol (74.4%
versus 44.7%, p<0.001). The proportion of participants with low HDL cholesterol
increased with age in both sexes. Around 29% of men and women in the oldest age
group had a high atherogenic index (total cholesterol/HDL cholesterol > 5) and more
women than men had a high atherogenic index (11.3% versus 7.5%, p<0.001).
The proportion of participants with a high triglyceride level (≥1.7 mmol/L) was 7.4%,
and was significantly higher in men than women (12.1% versus 5.0%, p<0.001).
Conversely, the percentage of participants with high LDL cholesterol level (≥3.8
mmol/L) was significantly higher in women than me n (7.6% versus 6.5%, p<0.001). The
percentage of participants with high a random glucose level (≥6.1 mmol/L) was 4.6%
and was similar for both sexes. The relatively high prevalence of CVD risk factors in this
population are similar to those reported in other Sub-Saharan countries.
Reported energy intake (as measured by FFQ) was unexpectedly high in this
population, possibly due to overestimation. However the percentage contribution to
energy from the macro-nutrients was within recommended levels.
xiv | P a g e
Summary
In chapter 4 the association between diet and CVD risk factors in “healthy”
participants was investigated. SFA (%E) intake was associated with increased LDL
cholesterol (p for trend = 0.017), triglycerides (p for trend = 0.048), glucose (p for trend
= 0.044) and was also associated with decreased HDL cholesterol (p for trend = 0.021)
after adjustment for potential confounders (age, sex, physical activity level, education,
income, tobacco use and alcohol use). PUFA intake was not associated with CVD risk
factors. Dietary fibre was inversely associated with random glucose only in men (p for
trend = 0.046), while alcohol intake was significantly positively associated HDL
cholesterol (p for trend <0.001) and inversely associated with BMI (p for trend = 0.020)
only in women.
In Chapter 5 the effect of diet on risk of disease (CVD/hypertension/diabetes) was
investigated. Individuals in the highest tertile of carbohydrate intake had increased risk
of disease compared to those in the lowest tertile of intake (OR 3.53 95% CI 1.45-8.60),
but the OR was no longer significant after adjustment for fibre and total fat.
Unexpectedly, high total fat and SFA intakes were associated with reduced risk of
disease but the association also weakened after adjustment for other nutrients. PUFA,
fibre and alcohol were not associated with disease risk. In this study however
“diseased” participants were older, their time of diagnosis of disease was not
verifiable, and they had a significantly lower intake of energy and fat, but a higher
carbohydrate intake compared to “healthy” participants. Therefore it is likely that
following diagnosis the “diseased” participants were advised on dietary changes (e.g.
replacing fat with carbohydrates) and weight loss.
In summary, this study found a relatively high prevalence of overweight and obesity
and of low levels of HDL cholesterol in young adults in Gaborone. High levels of
xv | P a g e
Summary
triglycerides, LDL cholesterol, glucose and high blood pressure were also found. The
results from the FFQ support literature suggesting that dietary patterns are shifting
from traditional diet to diets high in sugar, refined starches and meat. It is therefore
necessary for government to intensify CVD prevention strategies by promotion of
healthy eating and physical activity, and to monitor CVD risk factors through research
and surveillance.
xvi | P a g e
Chapter 1: Introduction
Chapter 1
Introduction
“If we knew what it was we were doing, it wouldn’t be called resea rch, would it?” Albert Einstein
1 |P age
1 Chapter 1
1.1 Introduction
1.1.1 Definitions
Cardiovascular disease (CVD) covers a number of disorders, amongst them diseases of
the cardiac muscles and the vascular system supplying the heart, brain and other vital
organs (Mendis et al., 2011).
CVD morbidity and mortality is predominantly due to coronary heart disease (CHD) and
cerebrovascular diseases (stroke). CHD is clinically manifested as ischemic heart
disease (IHD). It develops when the arteries supplying the blood to the heart become
partially or wholly blocked. This is usually caused by fatty deposits building up inside of
the arteries. Symptoms of CHD are chest pain which is temporary and treatable;
however, if blood supply to the heart is interrupted for a long time it is characterised
by severe chest pain (angina) and damage to the heart muscles resulting in a heart
attack (myocardial infarction). Stroke is caused by a disruption in the flow of blood to
part of the brain either because of narrowing of blood vessel (ischemic stroke) or
rupture of a blood vessel (Hemorrhagic stroke) (Sanders & Emery, 2003).
CVD is usually associated with atherosclerosis, an inflammatory disease characterised
by the accumulation of lipids and fibrous elements in the intima of the large to
medium arteries (Lusis, 2000). Atherosclerosis is primarily initiated by the
accumulation of low density lipoprotein (LDL) in the sub-endothelial matrix. Oxidation
of LDL is the most significant modification that gives rise to pro-inflammatory activity.
The accumulation of oxidised LDL triggers the recruitment of monocytes and
lymphocytes to the arterial wall and stimulates endothelial cells to produce pro-
2 |P age
inflammatory factors (adhesion molecules and growth factors), and at the same time
LDL can inhibit nitric oxide production which has anti-atherogenic properties. Highly
oxidised LDL is then taken up by macrophages to form foam cells.
Accumulations of these cholesterol-engorged macrophages or foam cells are
described as a fatty streak: the first visible sign of atherosclerosis. In humans, such
'fatty streak' lesions can usually be found in the aorta in the first deca de of life, the
coronary arteries in the second decade, and the cerebral arteries in the third or fourth
decades. Because of differences in blood flow dynamics, there are preferred sites of
lesion formation within the arteries such as branches and curvatures. Several risk
factors contribute to the damage of the endothelium and pro-inflammation including
hypertension, haemodynamic factors, sex hormones, diabetes, infections, elevated
homocysteine and hyperlipidaemia (Ross, 1999; Lusis, 2000).
Accumulation of lipids and smooth muscle cells and monocyte derived
macrophages lead to the formation of a fibrous plaque. Although a fibrous
plague can grow sufficiently large to block blood flow, the most important clinical
complication is an acute occlusion due to the formation of a thrombus or blood
clot (figure 1.1), resulting in myocardial infarction or stroke (Lusis, 2000).
Figure 1.1 Complex lesion and thrombosis in atherosclerosis (Lusis, 2000)
3 |P age
1.1.2 Morbidity and mortality
In 2002, almost one third (16.7 million) of global deaths were due to CVD, with low
income and middle income countries (LMIC) accounting for more than 86% of the
global CVD disease burden (WHO, 2002). In 2008 17.3 million people died from
CVD, and more than 17% of the deaths occurred in people younger than 60 years
(WHO, 2010). Over the last 20 years, deaths from CVD have been declining in high
income countries (WHO, 2010). In the United States of America, for example, the
death rate from CVD has been steadily declining for both men and women (figure
1.2). However, 80% of CVD mortality occurs in LMIC countries and continues to
increase rapidly; this disparity in CVD mortality has been largely attributed to better
combination of treatment regimes and a reduction in risk factors in developed
countries (WHO, 2010).
Figure 1.2 CVD mortality trends for males and females (United States 1979-2007),
(Roger et al., 2011)
4 |P age
In sub-Saharan Africa an estimated US$84 billion national income of some LMIC
countries will be lost due to CVD diseases (Abegunde et al., 2007). CVD is the major
contributor to the global burden of disease among the non-communicable diseases,
and it is projected that it will remain amongst the world’s leading killers until at
least 2030. Consequently, CVD is also projected to rank in the top 10 leading causes
of disability adjusted life years (DALY) in the same time period (Mathers & Loncar,
2006).
5 |P age
1.2 Risk factors for cardiovascular disease
Major risk factors for CVD are behavioural (tobacco use, physical inactivity,
unhealthy diets and harmful use of alcohol) and metabolic (hypertension, diabetes,
obesity and raised blood lipids), and other risk factors include age, gender poverty,
genetic predisposition and psychological factors (Mendis et al., 2011). In 2009 WHO
published a global health risk report which attributed more than 15 million (~60%)
deaths in middle income countries to CVD risk factors such as hypertension,
tobacco, alcohol use, overweight and obesity, low fruit and vegetable intake and
low physical activity (figure 1.3). (WHO, 2009).
Figure 1.3 Global mortality and risk factors (WHO, 2009)
6 |P age
Some of the risk factors described in figure 1.4 leading to atherosclerosis and
ultimately CVD include; elevated and modified low density lipoproteins (LDL), free
radicals caused by smoking, hypertension, genetic predisposition, infectious micro
organisms, obesity, low high density lipoprotein (HDL), and a combination of these
and other factors (Berse et al.,1992; Ross, 1999).
Lifestyle risk factors such a smoking, diet, physical inactivity and heavy alcohol
consumption are modifiable through lifestyle change whereas sex, age, race and
familial history (genetic and shared environment) are non-modifiable. Antiretroviral
medication used to treat human immune-deficiency virus (HIV) have also been
reported to increase risk of myocardial infarction (DAD study group, 2003).
Modifiable risk factors/lifestyle

risk factors:
Non modifiable risk
factors:
 Smoking
 Heavy alcohol
o Sex
consumption
o Age
 Physical inactivity
o Race
 Diet o Familial
o Saturated fat history
o Salt
o Cholesterol
o Energy
Intermediate conditions:
o Hypertension
o Obesity
o High low density cholesterol
o Low high density cholesterol
Anti- o Diabetes
retrov iral
treatment
Cardiovascular
disease
o CHD
o Stroke
Figure 1.4 Conceptual framework of development of CVD (modified from Wong,
Black & Gardin, 2005)
7 |P age
1.2.1 Sex
A review of 33 cohort, case-control and nested studies that examined the incidence,
prevalence or mortality of CVD in men and women concluded that, globally, men
have a higher prevalence of CVD than women (Pilote et al., 2007), although the
latest data from the United States report prevalence of CVD to be 39.9% and 42.7%
in men and women respectively (Roger et al., 2012), and since the late 1970s CVD
deaths have been declining more rapidly in men than in women (figure 1.5)
(Rosamond et al., 2008).
Figure 1.5 CVD mortality for men and women in the USA (Rosamond et al., 2008)
A study in urban Tanzania observed a higher prevalence of CVD risk factors,
including obesity, abdominal obesity, dyslipaedemia, diabetes mellitus and
metabolic syndrome in women than in men (figure 1.6) (Njelekela et al., 2009).
Similarly, The Heart of Soweto study reported that men were significantly less likely
than women to have one or more risk factors for heart dis ease (OR 0.3, 95% CI 0.2
to 0.4) in a subgroup of 714 participants who denied having a pre-existing risk factor
for heart disease (Tibazarwa et al., 2007).
8 |P age
Figure 1.6 Prevalence of CVD risk factors by gender in Tanzania. (Njelekela et al.
2009)
1.2.2 Blood pressure
High blood pressure or hypertension is currently defined by the seventh report of
the Joint National Committee on prevention, detection, evaluation, and treatment
of high blood pressure as a systolic blood pressure of ≥ 140 mmHg and a diastolic
blood pressure of ≥90mmHg (Chobanian et al., 2003). Globally 40% of the adults
have raised blood pressure, and the highest proportion is in Africa (46%) while the
lowest is in the Americas (35%). The WHO attributes 7.1 million (13%) deaths
globally, and 64.3 million DALY to hypertension (WHO, 2002) and globally, 51% of
stroke and 45% of IHD deaths are attributable to high systolic blood pressure (WHO,
2011). The Multi Risk Factor Intervention Trial (MRFIT) in the US showed that
hypertension was associated with more than 50% increased risk of death from CHD,
9 |P age
and the risk was similar in both black and white diabetes patients (Vaccaro et al.,
1998).
With regards to hypertension in Sub-Saharan Africa, a systematic review of studies
on hypertension reported a higher prevalence of hypertension in urban (47%)
compared to rural areas (11%) (Addo et al., 2007). The same review also showed a
steady increase in the prevalence of hypertension from the youngest to the oldest
age group (figure 1.7).
Figure 1.7 Prevalence of hypertension (≥ 140/90mmHg) in Sub-Saharan Africa by
age and sex (Addo et al. 2007)
The Heart of Soweto study (a cross-sectional study in a South Africa township),
reported that of the 1691 volunteers (99% black) who were screened, one-third of
both men and women had either elevated / systolic(≥ 140 mmHg) or diastolic (≥ 90
10 | P a g e
mmHg) blood pressure (Tibazwara et al., 2008). Another cross-sectional study in
South Africa (The South African Stroke Prevention Initiative (SASPI)) confirmed that
high blood pressure was a major risk factor for CVD and reported that only 57%
(95% CI 51.8-62.3) had normal blood pressure (Thorogood et al., 2007). In
metropolitan Ghana a community based survey also reported that more than 54%
of women had elevated blood pressure (Duda et al., 2007).
One of the earliest studies on blood pressure in Bushmen in Botswana did not
record any of the 152 participants as having hypertension. In fact, blood pressure
did not increase with age (Truswell et al., 1972). Truswell and colleagues attribute
these normal blood pressures reading in Bushmen to several factors such as low salt
intake, lack of obesity, and freedom from stress. More than 20 years later a
descriptive study in Botswana involving 337 participants (> 60years) found that 19%
(95% CI 15-23%) of people aged > 60 years had a blood pressure reading above
160/95mmHg (Clausen et al., 2000).
1.2.3 Diabetes mellitus
Diabetes mellitus (DM) is defined as elevated fasting plasma glucose ≥7.0mmol/l
(126mg/dl) or a 2-hour plasma glucose (venous plasma after ingestion of 75g oral
glucose load) ≥.11.1mmol/l (200mg/dl) (WHO/IDF, 2006). Globally ~10% of people
aged ≥ 25 years have diabetes, and the highest prevalence is in the Eastern
Mediterranean region and the Americas (11%) while the lowest is in Europe (WHO,
2011). It is estimated that by 2030 about 439 million (7.7%) people worldwide will
have diabetes, and the African region is expected to have the largest proportional
increase of 98% (from 12.1 million to 23.9 million) in adult type 2 diabetes
numbers by 2030 (Shaw et al., 2010).

11 | P a g e
Results from the Framingham study comparing participants examined in two time
periods (1952-1974 and 1974-1988) showed that when comparing CVD risk factors
over the two time periods, the prevalence of hypertension, smoking and high
cholesterol significantly declined, but the prevalence of DM (adjusted for age and
sex) increased almost 2 fold (figure 1.8) among CVD cases (8.1% versus 14.6%,
P=0.0009) (Fox et al., 2007).
Figure 1.8 Age- and sex-adjusted prevalence of hypertension (HTN), smoking,
elevated cholesterol, obesity, and diabetes among CVD cases in the early versus
the later time period of the Framingham study. Bars represent 95% CIs. (Fox et al.
2007)
Limited data on type II diabetes prevalence in Africa suggest that there has been a
steady increase in diabetes between the 1980s and 1990s with higher rates
12 | P a g e
observed in urban compared to rural areas (Mbanya et al., 2010). Data from the
INTERHEART-Africa study suggest that diabetes is one of five risk factors that
account for almost 90% of initial myocardial infarction, and suggest that
“uncontrolled major CVD risk factors will have a larger impact on the burden of CVD
in Africa than elsewhere” (Steyn et al., 2005).
A recent systematic review and meta-analysis of health examination survey and
epidemiological studies of 2.7 million participants in 199 countries and territories
from 1980-2008 reported a global age-standardised mean fasting plasma glucose
(FPG) to have risen by 0.07 mmol/L and 0.09 mmol/L per decade for men and
women respectively. The prevalence of diabetes also increased from 8.3% and 7.5%
to 9.8% and 9.2% for men and women respectively. The highest increase in diabetes
and FPG was found in Oceania and the smallest increase in Sub-Saharan Africa
(Danaei et al., 2011). In Botswana the same study shows that FPG remained at
around 5mmol/L for both men and women, but diabetes prevalence from 1980-
2008 increased from ~7% to ~10% in women while in men it remained at ~7%
during the same period.
1.2.4 Blood lipids
1.2.4.1 Triglycerides
Triglycerides are used for energy storage and are formed by the liver through a
process called lipogenesis. Triglycerides are known to be influenced by obesity,
diabetes, intake of alcohol and a high carbohydrate diet (Garrow & James, 2000). An
increased secretion or delayed clearance of triglyceride rich lipoproteins has been
13 | P a g e
associated with increased risk of CHD especially when HDL cholesterol is low (<1.0
mmol/L) (Jacobson et al., 2007, Kannel & Vasan, 2009).
An updated meta-analysis of prospective studies in 29 western countries involving
262, 525 participants including 10, 158 CHD cases suggested that raised circulating
triglycerides are associated with CHD risk. In the same study adjustment for
established CHD risk factors, especially HDL cholesterol, weakened the association.
Nonetheless, a combined odds ratio for CHD after adjustment for risk factors
remained significant at 1.72 (95% CI, 1.6-1.9) for participants in the top tertile
compared to the bottom tertile of triglyceride levels, with a similar impact in men
and women (Sarwar et al., 2007).
A cross-sectional study of metabolic syndrome and obesity among 150 hospital
workers at Kanye Seventh Day Adventist hospital in Botswana reported that 14% of
participants had high triglyceride levels (≥ 1.7mmol/L) (Garrido et al., 2009).
Mengesha (2007) reported that almost 40% of diabetic patients attending
Gaborone city council clinics in Botswana had high triglycerides.
1.2.4.2 Low Density Lipoprotein (LDL) cholesterol
LDL cholesterol is required for the building of cell membrane and for synthesis of
steroid hormones. When LDL cholesterol is oxidised, scavenger receptors present
on the surface of macrophages recognise it triggering macrophages to become lipid-
laden into foam cells. Foam cell formation is considered as an early step towards
the development of atherosclerosis (Lusis, 2000). A high plasma LDL (≥ 1.30
mmol/L) has been reported to be independently associated with increased risk of
14 | P a g e
carotid atherosclerosis OR 6.9 (95% CI. 1.3-37.8) among women with metabolic
syndrome in Italy (Montalcini et al., 2005).
A recent systematic review and meta-regression analysis of 108 randomised
controlled trials (comparing lipid modifying agents or diet) involving almost 300 000
participants at risk of cardiovascular events followed for at least six months
reported that LDL cholesterol reduction of 10% was associated with a relative 10%
decrease in CHD (Briel et al., 2009).
1.2.4.3 High Density Lipoprotein (HDL) cholesterol
HDL cholesterol is responsible for the elimination of excess cholesterol from
peripheral tissues to the liver through a process called reverse cholesterol transport
Furthermore, HDL cholesterol inhibits movement of macrophages into the arterial
cell wall and is also important for maintaining endothelial cell lining as well as
inhibition of LDL cholesterol oxidation (Plump et al., 1994).
Cooney et al (2009), in their data of over 90,000 men and women without pre-
existing CHD pooled from 7 European prospective studies (SCORE project), reported
a significant inverse association between HDL cholesterol and CVD mortality in men
(multivariate adjusted Hazard Ratio 0.76 (95%CI. 0.70-0.83) in men and Hazard
Ratio 0.60 (95%CI. 0.51-0.69) in women per 0.5 mmol/L increase in HDL cholesterol.
Furthermore, evidence from a systematic re-examination of five major
epidemiological studies concluded that there was an independent inverse
association between HDL cholesterol levels and CHD rates (2-3% decrease in CHD
for a 1mg/dl increment in HDL cholesterol) even after adjusting for multiple
covariates (age, smoking, BMI, systolic blood pressure and LDL- cholesterol).
15 | P a g e
However, this association was attenuated when non-HDL cholesterol was adjusted
for (Gordon et al., 1989). These findings are however not supported by a recent
systematic review and meta-regression of 108 randomised control trials which
reported that an increase in HDL- cholesterol was not associated with reduced risk
of CVD morbidity or total mortality (Briel et al., 2009).
Data on the association between plasma lipids and vascular disease in Sub-Saharan
Africa are limited. Nonetheless, a cross-sectional study in Cameroon reported weak
inverse correlations (r=-0.19, p=0.054) between HDL-cholesterol and BMI (Achidi
and Tangoh, 2010).
1.2.5 Physical activity
Regular moderate physical activity such as walking and participating in sports for 30
minutes or more for at least 5 days a week has benefits for health. In 2008, 31% of
adults aged ≥15 years globally had insufficient physical activity with the highest
prevalence of insufficient physical activity found in the Americas and the Eastern
Mediterranean regions where more than 50% of women are insufficiently active.
Insufficient physical activity also increased with level of countries’ income (WHO,
2011). In 2004 physical inactivity was estimated to be the fourth leading cause of
death worldwide with over 3 million deaths and 2.1% of global DALY attributable to
it (WHO, 2009).
Analysis of data from the US Behavioural Risk Factor Surveillance System survey
with almost 300, 000 participants found that patients with CHD are less likely to
comply with physical activity recommendations (Zhao et al., 2008). However, a
16 | P a g e
recent meta-analysis of 26 prospective cohort studies followed up for 4-25 years
reported that physical activity offered a significant protection against CHD in
individuals who performed moderate [RR 0.88 (95% CI. 0.83-0.93)] and high [RR
0.73 (95% CI. 0.66-0.80)] levels of physical activity (Sofi, 2008).
The INTERHEART study, a global case control study including 27 098 participants
from 52 countries reported that physical activity was one of the modifiable risk
factors associated with reduced risk of myocardial infarction in women [OR 0.40
(95%CI 0.41-0.57)] and men [OR 0.77 (95%CI 0.71- 0.83)] (Anand et al., 2008).
Furthermore, a study in Tanzania also confirmed that physical activity was higher in
rural compared to urban area, and people in rural areas had a more favourable lipid
profile than urban dwellers (Mbalilaki et al., 2007). In communities undergoing rural
to urban transition CVD risk factors such as systolic blood pressure, total cholesterol
and LDL cholesterol were lower in overweight and obese women and men who had
a moderate level of physical activity compared to those who were inactive. (Kruger
et al., 2003).
1.2.6 Obesity
Overweight and obesity are defined as abnormal or excessive fat accumulation that
presents a risk to health. A crude population measure of obesity is the body mass
index (BMI); a person’s weight (in kilograms) divided by the square of his or her
height (in metres). BMI classification is provided in table 1.1. The worldwide
prevalence of overweight in adults aged ≥20 years was 35% in 2008, and the
prevalence of obesity was 10% and 14% for women and men respectively. The
prevalence of overweight and obesity was highest in the Americas (62% overweight
17 | P a g e
and 26% obesity) and lowest in South-East Asia (14% overweight and 3% obesity).
Women were more likely to be overweight and obese, and in LMIC countries,
women had twice the obesity prevalence of men (WHO, 2011).
Table 1.1 BMI categories (WHO, 2000)

2
BMI (cat egory) BMI (kg/(m)
Underweight <18.5
Normal 18.5-24.9
Overweight 25-29.9
Obese 30-34.99
Morbidly obese ≥35
Waist circumference (WC), hip circumference (HC) and waist-hip ratio (WHR)
measure abdominal obesity which is not captured by BMI (table 1.2).
Table 1.2 WC categories for risk of metabolic complications (WHO, 2000)
Men Women
WC
Increased risk 94-101.9cm (37-40 inches) 80-87.9cm(31.5-34.6 inches)
Substantially increased risk ≥ 102cm (=>40 inches) ≥ 88cm (=>34.6 inches)
WHR
Increased risk ≥ 0.90 ≥ 0.85
WHR has been found to be a better predictor of heart attack than BMI in men and
women, across all ages and ethnic groups, as well as in those with or without
18 | P a g e
dyslipaedemia, diabetes, or hypertension (Yusuf et al., 2005) or in post menopausal
women (Van Pelt et al., 2001). However a WHO expert consultation on WC and
WHR concluded that there was insufficient data to suggest giving other measures
any priority and that the joint use of measurement where possible is desirable
(WHO expert consultation, 2011).
Overweight and obesity are major risk factors for a number of chronic diseases,
including diabetes (Kahn et al., 2006) and CVD (Kannel et al., 1991). There is
evidence that central obesity as measured by WHR or waist circumference are the
simplest ways of measuring obesity (Kannel et al., 1991) in men (Rexrode et al.,
2001) and women (Roxrode et al., 1998). Overweight and obesity have also been
associated with CVD and CHD mortality in men independent of baseline CHD (Batty
et al., 2005). Once considered a problem only in high income countries, overweight
and obesity are now dramatically on the rise in low and middle-income countries,
particularly in urban settings (Puoane et al., 2002).
In Botswana while there is a dearth of data on CVD (Bhagat, 2001), high levels of
obesity have been reported. Data from the Botswana Family Health Survey IV (BFHS
IV) indicated that 40% and 16.1% of women and men aged 20-49 years respectively
were either overweight or obese. (Letamo, 2010). A cross-sectional study of 150
health workers in Kanye Botswana reported that more than 50% of women were
either overweight or obese (Garrido et al., 2009).
1.2.7 Alcohol
Excessive alcohol consumption is associated with many health problems including ,
but not limited to, cancers, liver disease, CVD, gastrointestinal disorders,
19 | P a g e
complicated pregnancy, and psycho-social problems such as injuries and violence
(Agarwal, 2002; NICE, 2010). An estimated 4.5% of the global burden of disease is
caused by harmful use of alcohol with cancers, CVD and liver cirrhosis responsible
for a quarter of the burden (WHO, 2011).
However, a publication reporting lower mortality rates from CHD in France
compared to the USA or UK despite high intake of saturated fat presented a
paradoxical situation (“French Paradox”) which was in part attributed to a
protective effect of high wine consumption (Renaud & Longeril, 1992).
Subsequently, studies have suggested moderate (1-2 units/day) alcohol
consumption to be protective against CVD, reduce insulin resistance and blood
pressure (Rimm, 2000; Agawal, 2002 and Ronksley et al., 2011). Rimm et al (1999)
reported in the their meta-analysis of 42 experimental studies on moderate alcohol
use and risk of CHD that there was an overall 24.7 % decrease in risk of CHD
associated with an intake of 30g/day of alcohol and this protective effect has been
attributable to the link with increased HDL cholesterol and apolipoprotein A1, and
reduced concentration of fibrinogen.
A recent systematic review and meta-analysis of 63 intervention studies on the
effect of alcohol consumption on biological markers reported a significant dose-
response between alcohol consumption and levels of HDL cholesterol,
apolipoprotein A1, adiponectin and fibrinogen (Brien et al., 2011).
In the data from the Women’s Health Study in the US Djoussè et al (2009) also
reported that, moderate alcohol drinking accounted for an estimated 86% reduction
in risk of CVD compared with non-drinkers in women. This protective effect of
20 | P a g e
moderate alcohol consumption was attributed to alcohol’s beneficial effect on
lipids, glucose metabolism, inflammatory/haemostatic factors and blood pressure.
In South Africa, alcohol consumers in the Transition and Health during Urbanisation
in South Africa (THUSA) study were reported to have higher HDL cholesterol, blood
pressure and serum iron compared to non-consumers (Gopane et al., 2010). Two
biological markers of alcohol consumption (percentage carbohydrates-deficient
transferrin (%CDT) and gamma-glutamyl transferase (GGT)) have been reported to
be positively associated with HDL cholesterol and negatively associated with BMI in
the Prospective Urban and Rural Epidemiology (PURE) study in South Africa (Pisa et
al., 2010). In the same study GGT was also positively associated with blood
pressure, triglycerides and total cholesterol (Pisa et al., 2010).
It has been suggested that it may the type of alcoholic beverage and not the alcohol
per se (Renaud & Longeril 1992) that is protective, however other arguments point
to the ethanol in drinks (Rimm et al., 1999) and the drinking patterns (Van de Wiel
& de Lange, 2008) but not the type of drinks. Pletcher et al (2005) in their findings
from the Coronary Artery Risk Development in Young Adults (CARDIA) study of over
3000 participants (aged 33-45 years) reported that coronary calcification increased
with alcohol consumption, and binge drinking (≥ 2 binges/month) was associated
with increased risk of coronary calcification [OR 2.2 (95% CI 1.7-2.9)] compared to
non-drinkers after adjustment for potential confounders; notably in this study
moderate drinking in black men was associated with increas ed risk of coronary
calcification. Rimm & Stampfer (2002) in their editorial “Wine, beer, and spirits: Are
they really horses of a different colour” argue that ethanol in beer and wine is
similar; therefore the effects of the two beverages on lipid and hae mostatic factors
21 | P a g e
will be similar. However, they caution that residual confounding by diet, physical
activity and other socio-demographic characteristics need to be carefully addressed.
1.2.8 Smoking
According to WHO’s global health risk report almost 6 million people die (6%
women and 12% men) annually from tobacco use and exposure (WHO, 2009). The
highest prevalence of smoking is in Europe while the lowest is in Africa and more
men than women smoke in all regions of the world (Mendis et al., 2011). More than
4000 components are contained in cigarettes. However, nicotine and carbon
monoxide in particular are responsible for reduced myocardial oxygen supply;
nicotine raises blood pressure and heart rate predisposing smokers to myocardial
infarction (Lakier, 1992; Smith et al., 2000). Cigarette smoking has been suggested
to play a mediatory role in oxidative stress and athero-thrombotic disease (Ambrose
& Barua, 2004)
The association between smoking and vascular diseases and all cause mortality has
been known for some time. For example in the Seven Countries study involving
more than 12,000 men followed up for 25 years a clear dose response relationship
between smoking and CHD and stroke was reported (Jacobs et al., 1999). A meta-
analysis of 9 cohort studies with up to 39 years follow-up found exposure to
environmental tobacco smoke (passive smoking) to increase risk of death from
heart disease in women RR 1.15 (95% CI. 1.03 -1.28) (Kaur et al., 2004).
A systematic literature review of tobacco use among adults 15 years and older in
Sub-Saharan Africa reported a prevalence of low tobacco use in Nigerian female
general medical patients (9%) and high prevalence in coloured male in South Africa
22 | P a g e
(79%), tobacco use was also high in middle aged men (30-49 years) compared to
women (Townsend et al., 2006).
1.3 Nutrition transition
The phenomenon called the “nutrition transition” is characterised by rapid
urbanisation, economic growth and improving incomes, technological change,
increased processed foods (refined starch, high fat and high sugar) and emerging
obesity (Popkin 2001, Popkin and Gordon-Larsen, 2004).
A review of evidence on the nutrition transition and CVD risk factors in the Middle
East and North African countries has reported a steady and consistent increase in
energy, protein and fat per capita, and a decrease in fruit and vegetables and
physical activity (Mehio et al., 2010). The nutrition transition is also fast becoming
evident in urban setting of Sub Saharan countries where obesity is associated with
high energy intake in Cameroonian men (Jackson et al., 2006), body image in The
Gambia (Sievo et al., 2005) and high socio-economic status as well as low physical
activity.
1.4 Diet and risk of CVD
Evidence on the association between diet and CVD is increasing, however due to
varied methodologies different conclusions and/or recommendations have been
made. A systematic review of evidence from prospective cohort studies and
randomised controlled trials (RCT) investigating a causal link between dietary
factors and coronary heart disease found varying degree of evidence using the
Bradford hill criteria (Mente et al., 2009). Mente et al created a score based on the
23 | P a g e
first 4 of 9 Bradford hill guidelines which systematically evaluates evidence strength,
consistency, temporality and coherence and the highest score of 4 indicates the
strongest evidence.
The strongest evidence of a causal relationship for protection against CHD in
prospective cohort studies was found in Mediterranean diet (high intake of
vegetables, fruit, legumes, nuts, whole grains, cheese or yoghurt, fish, and mono-
unsaturated fatty acids), nuts, high quality diet and high vegetable diet. Intake of
fish, omega 3 fatty acids, folate, whole grains, fruit and vegetables, fibre, alcohol,
dietary vitamin E, dietary beta carotene, and dietary vitamin C showed moderate
evidence of a causal relationship for protection against CHD in prospective cohort
studies. Although the Mediterranean dietary pattern displayed a significant inverse
association with CHD in RCTs [OR 0.32 (95% CI 0.15-0.48)], it was the only dietary
pattern investigated. Trans-fatty acids and high glycaemic index foods had strongest
evidence for association with increased risk for CHD in cohort studies, whereas
evidence on the causal association of saturated fatty acids, meat, eggs, and milk
with CHD was weak (Mente et al., 2009).
1.4.1 Dietary fat
Total dietary fat in general has been associated with increased risk of CVD.
However, evidence is increasingly suggesting that it may be the type of fat rather
than total fat that is predictive of CVD risk (Hu et al., 2001).
1.4.1.1 Saturated fatty acids (SFA)
SFA are the main fatty acids found in meat and dairy products. Palmitic acid (16:0) is
the most commonly occurring dietary SFA present in palm oil and meat. Stearic acid
24 | P a g e
(18:0) is found mainly in meat and cocoa. Short chain SFA (2-6 carbons long)
produced by bacterial fermentation of carbohydrates are present in cow’s milk and
butter while medium chain SFA (8-14 carbon long) such as lauric acid ((12:0) and
myristic (14:0) are found in some tropical oils such as palm and coconut oils
(Sanders & Emery, 2003).
SFA intake has been associated with increased risk of CVD probably due to
cholesterol raising characteristics, increased insulin resistance and increased
platelets aggregation (Geissler & Powers, 2005). However, evidence on the
association of SFA and CVD risk is at best equivocal (Ravnskov, 1998). Recently a
meta-analysis of prospective epidemiological studies reported that there was no
significant evidence for concluding that high SFA intake compared to low intake is
associated with increased risk of CHD [RR 1.07 (95% CI 0.96-1.19)], stroke [RR 0.81
(95% CI 0.62-1.05)] or total CVD 1.00 (95% CI: 0.89-1.11), but suggested that CVD
risk factors may be influenced by othe r nutrients that replace dietary saturated fat
(Siri-Tarino et al., 2010).
1.4.1.2 Monounsaturated fatty acids (MUFA)
MUFA are found in red meat, whole milk, nuts (ground, cashew, macadamia) and
oils (olive oil, safflower, almond, avocado, corn, sesame grape seed rapeseed
(canola) and others). MUFA have been suggested to be the neutral in terms of their
effect on lipoprotein profile (Hayes 2002). MUFA have also suggested to be
inversely associated with reduced risk of CHD when replaced (isocaloric) for SFA (Hu
et al., 1999). However, a recent pooled-analysis of 11 cohort studies has found no
25 | P a g e
association between MUFA and CHD when 5%E from SFA was replaced with MUFA
(Jakobsen et al., 2009).
1.4.1.3 Polyunsaturated fatty acids (PUFA)
PUFA are required in the diet to promote normal physiological functions linked to
membrane integrity and regulatory cell signals . PUFAs are mainly found in
safflower, sunflower flaxseed and fish oils. Linoleic acid (LA, 18:2n-6) is the major
essential fatty acid and has a cholesterol lowering effect (Sanders & Emery, 2003).
The cholesterol-lowering effect of PUFA has been shown to be particularly
protective against CHD (Hu et al., 1997; Oh et al., 2005).
A recent review by the American Heart Association on omega-6 fatty acids and the
risk of CVD concluded that “Aggregate data from randomized trials, case-control
and cohort studies, and long term animal feeding experiments indicate that the
consumption of at least 5% to 10% of energy from omega-6 PUFAs reduces the risk
of CHD relative to lower intake” (Harris et al., 2009).
Alpha-linolenic acid (LNA, 18:3n-3) is the other primary essential fatty acid. There is
evidence of dietary alpha-linolenic acid intake and its metabolites eicosapentaenoic
acid (EPA) and docosahexaenoic acid (DHA) being cardio-protective (Wijendran &
Hayes, 2004). The possible mechanisms are not clear but may be linked more with
cardiac functions than with plasma lipids (Wijendran & Hayes, 2004). Furthermore,
supplementation with fish oil high in long chain n-3 PUFA (1.4g/day) compared to
sunflower oil (3.6g/day linoleic acid) and a control (blend of palm and soybean oil)
in a randomised controlled trial of patients awaiting carotid plaque removal was
26 | P a g e
found to increase EPA and DHA in the carotid plaque fractions, and plaque stability
was also improved (Thies et al., 2003).
1.4.1.4 PUFA: SFA ratio
Recent evidence from cohort studies (Jacobsen et al., 2009) and randomised
controlled trials (Mozaffarian et al., 2010) suggest that replacing SFA intake with
PUFA intake significantly reduces risk of CHD. However, it has also been suggested
that a balance of 1:1.3:1 for PUFA: MUFA: SFA is critical in improving LDL/HDL ratio
(Hayes, 2002).
1.4.1.5 Trans-unsaturated fat
Most unsaturated fatty acids in nature have their double bonds in the cis
configuration (Sanders & Emery, 2003), however when overheated the hydrogen
atoms can jump from cis to trans positions, and the molecule behaves like a SFA.
Dietary intake of trans fatty acids have been associated with high risk of CHD (Willet
et al., 1993; Pietinen et al., 1996; Hu et al., 1997; Oomen et al., 2001 and Oh et al.,
2005) possibly due to increased LDL cholesterol and decreased HDL cholesterol (de
Roos et al., 2001).
1.4.2 Dietary fibre
Dietary fibre, defined as the remnants of indigestible portions of plant food (Trowell
1972), is primarily found in whole grain cereals, legumes, pulses and certain fruit
and vegetables. There are two types of fibre, water soluble and water insoluble
fibre. Insoluble fibre consists of mainly lignin, cellulose and hemicelluloses part of
the plant, while soluble fibre comprises gel forming or viscous fibres such as pectins,
27 | P a g e
gums, mucilage and hemicelluloses. Soluble fibre has been found to reduce total
cholesterol and LDL cholesterol (Brown et al., 1999) possibly by sequestration of
bile acids and increased faecal output (Trowell, 1972) as well reduction in
postprandial lipid levels (Cara et al., 1992).
Intake of fibre, particularly soluble fibre found in cereals and fruit, has been found
to be inversely associated with risk of CHD in cohort studies (Pereira et al., 2004).
Some prospective studies have reported fibre to be protective against CHD in
women (Wolk et al., 1999), and CHD death in elderly (Streppel et al., 2008) or men
who smoke (Pietinen et al., 1996).
Similarly, the National Health and Nutrition Examination Survey I in the US (NHANES
I) reported an inverse association between intake (lowest versus highest quartile) of
soluble fibre and risk of subsequent CHD and CVD incidence (p for trend=0.01) and
mortality (p for trend=0.03) in a nationally representative sample (Bazzano et al.,
2003). Cereal fibre intake in particular has been reported to reduce the risk of CHD
in among women [RR 0.66 (95% CI 0.49-0.8)] (Wolk et al., 1999) and the elderly [RR
0.79 (95% CI 0.62-0.99)] in the highest quintile compared to the lowest quintile
(Mozaffarian et al., 2003). Recently Tighe et al (2010) in their randomised controlled
dietary trial of middle aged healthy individuals reported that 3 daily servings of
whole grain food (wheat or wheat and oats) significantly reduced SBP and pulse
pressure compared to a control (refined grain) after 6 weeks for oats and wheat and
12 weeks for both groups.
28 | P a g e
1.4.3 Sodium and potassium
Sodium and potassium are the major solutes in body fluids, and are responsible for
maintaining osmolarity and volume of intra and extracellular fluid. Sodium and
chloride are also responsible for transportation of small molecules across cell
membranes as well as absorption of nutrients from the gut (Sanders & Emery,
2003).
Increased sodium intake has been associated with elevated blood pressure possibly
due to fluid retention and alteration of ion transport in vascular smooth muscle
(Garrow & James, 2000). A recent Cochrane review and meta-analysis of 7
randomised controlled trials with at least 6 months follow-up on reduced dietary
salt (salt reduction or advice to reduce) for the prevention of CVD concluded that
salt reduction was associated with a reduction of between 1 and 4 mmHg in SBP but
not CVD morbidity or mortality in normotensive participants (Taylor et al., 2011).
The major dietary sources of potassium include fruit and vegetables, meat and meat
products. There is evidence for an inverse association between potassium intake
and risk of stroke (Geissler & Powers, 2005). A recent meta-analysis of 11
prospective studies (1966-2009) with nearly 250,000 participants investigating the
relationship between dietary potassium intake and risk of stroke and cardiovascular
outcome found that increasing potassium intake by 1.64g/day was associated with
a 21% reduction in risk of stroke, and was suggestive of protection against risk of
CHD and CVD (D’Elia et al., 2011). Joint effects of long term sodium and potassium
intake measured by 24 h-hour urinary excretion have also been investigated in the
Trial of Hypertension Prevention follow-up study among pre-hypertension adults
29 | P a g e
aged 30-54 years, and a higher sodium to potassium excretion ratio was significantly
associated with increased risk of subsequent CVD [RR 1.24 (95% CI 1.05-1.46)]
compared to low sodium and potassium excretion (Cook et al., 2009).
1.4.4 Antioxidant nutrients
Antioxidants may play a role in preventing CVD by inhibiting oxidative reactions
caused by free radicals through several mechanisms including (i) scavenging free
radicals and thus reducing the damage to the endothelium, (ii) binding to transition
metal ion catalysts such as copper and iron to prevent generation of free radicals.
Antioxidants needed in the diet are vitamin E, vitamin C, carotenoids, selenium and
flavonoids, and diets high in antioxidants are rich in fruit and vegetables, vegetable
oils, nuts and legumes (Garrow & James, 2000).
Dietary intakes high in certain antioxidants have been associated with lower CVD
risk (Joshipura et al., 2001; Buijsse et al., 2008). However evidence from a meta-
analysis of 68 randomised controlled trials with over 200 000 participants showed
that treatment with high doses of orally administered beta carotene (mean 17.8
mg/day), vitamins A (mean 20,219 IU/day) E (mean 569 IU/day) and C (mean 488
mg/day) did not lower the risk of CVD and may increase mortality (Bjelakovic et al.,
2007). Similarly, Cook et al (2007) in their randomised factorial trial of vitamin C
(500 mg/day) and E (600 IU/day) and beta carotene (50 mg/day) with >9 years
follow up reported no overall effect of either vitamin C (RR, 1.02, 95% CI 0.92-1.13),
vitamin E (RR, 0.94 95% CI 0.85-1.04) or beta-carotene (RR 1.02 95% CI 0.92-1.13)
on CVD outcomes of women at high risk of CVD.
30 | P a g e
1.4.5 Fruit and vegetable consumption
Fruit and vegetable consumption has been associated with reduced risk for CVD (Liu
et al., 2000; Joshipura et al., 2001; Bazzano et al., 2002) through their protective
components such as fibre, antioxidants, phenolic compounds and flavonoids (Van
Duyn et al., 2000).
Hung et al (2004) examined the association between fruit and vegetable
consumption and risk of major chronic diseases in a group of participants including
over 71 000 females in the Nurses’ Health Study and more than 37, 000 males in the
Health Professionals’ follow-up study. The authors reported a significant inverse
association between consumption of fruit and vegetables and CVD. The associations
were also observed in pooled multivariate analysis which found a strong inverse
association between higher fruit and vegetable intake (≥ 8 serving compared to <
1.5 serving relative risk and CVD [RR 0.70 (95% CI 0.55-0.89)], and the inverse
association was stronger in current smokers and non vitamin users (Hung et al.,
2004).
Correspondingly, a meta-analysis of 9 major cohort studies (7 US and 2 Finland)
with a follow-up period of 5-18 years found an inverse association between the risk
of CHD and an additional portion of fruit and vegetable [RR 0.96 (95% CI 0.93-0.99)]
and fruit [RR 0.93 (95%CI 0.89-0.96)] although authors stated publication bias as a
limitation because they excluded some key studies because of insufficient data in
some and no assessment of fruit and vegetable in others (Dauchet et al., 2006).
31 | P a g e
1.4.6 Dietary patterns
There is evidence of association between CHD rates and intake of specific foods or
nutrients (Geissler & Powers, 2005). However, due to the correlations between
nutrients it is difficult to be certain which dietary component is responsible, or
whether there is an interplay of several dietary components that confers protection
against CVD. Several studies have assessed the association of different dietary
patterns (table 1.3) with CHD risk.
Table 1.3 Description of Mediterranean, prudent western dietary and oriental
patterns
Dietary pattern Foods
Mediterranean High: Monounsaturated to saturated fat ratio, legumes, fruit,

vegetables and cereal
(Trichopoulou et al., 2003)
Moderate: Alcohol, milk and milk products
Low: Meat and meat products
Prudent (Fung et al., 2001) High: Legumes, fruit, vegetables and cereal, fish and poultry
Moderate: Alcohol, milk and milk products
Low: Meat and meat products, beverages with added sugar
Western (Fung et al., 2001) High: Red meat, processed meats, eggs, refined grain, sweets,
desserts, French fries and high fat dairy products.
Oriental (Iqbal et al., 2008) High: Tofu, soy
32 | P a g e
High scores of Prudent or Mediterranean diet have been associated with lower risk
of CHD and CHD mortality (Hu et al., 2000; Fung et al., 2001) in men [RR 0.75 (95%
CI 0.59-0.95)] and women [RR 0.76 (95% CI 0.60-0.98)], and all-cause mortality [HR
0.75 (95% CI 0.64-87)] (Trichopoulou et al., 2003) as well as lower mortality in the
elderly (Knoops et al., 2004). Conversely, the highest quintile of the Western diet
has been associated with an increased risk of CHD compared with the lowest
quintile in both men and women independent of other risk factors (Hu et al., 2000;
Fung et al., 2001). Preliminary results from a sub study of a multicentre randomised
primary prevention trial of CVD furthermore suggests that a Mediterranean diet
with additional olive oil or nuts improves the lipid profile and reduces blood
pressure compared to a low fat diet (Estruch et al., 2006). The INTERHEART study
found an inverse association between the Prudent diet and acute MI, while a U-
shaped association was observed with Western diet. H owever, no association was
observed with the Oriental diet (Iqbal et al., 2008).
1.5 Antiretroviral medication and cardiovascular diseases
UNAIDS (2010) estimated that more than 33 million people globally were living with
HIV in 2009. Of these 22.5 million (68% of global total) lived in Sub-Saharan Africa.
With an estimated 5.4 million people living with HIV South Africa has the largest
number of people living with HIV in the world. Swaziland has the highest HIV
prevalence in the world at 25.4% while approximately 1 in 5 people in Botswana are
HIV positive. Antiretroviral (ARV) medication coverage has improved in the region
and Botswana has one of the highest coverage at more than 90% (UNAIDS, 2010).
ARV medication used to treat HIV induces peripheral lipodystrophy, a condition
characterised by fat wasting of the face, limbs and upper trunk but not the
33 | P a g e
abdomen. ARVs and protease inhibitors which have high affinity for the catalytic
site of HIV=1 protease may cause lipodystrophy by binding, inhibiting and impairing
human protein involved in lipid metabolism (Carr et al., 1998; Carr et al., 1998),
potentially making patients more susceptible to CVD risk factors such as high
triglycerides and total cholesterol. An observational study of 11 cohorts comprising
of more than 23, 000 HIV-1 infected patients in Europe, the United States and
Australia reported a 26% increased risk of myocardial infarction per year of
exposure during the first four to six years of ARV therapy (DAD study group, 2003).
A follow up (> 6years) study confirmed the increased risk of MI during ARV therapy
(DAD study group 2007). While the full mechanism by which protease inhibitors
may lead to increased rate of MI is still not fully understood, impaired lipid
metabolism, smoking, diabetes, hypertension, previous CVD and dyslipaedemia
could be some of the reasons (DAD study group., 2007; Pere et al., 2008).
Therefore alteration to lifestyle factors such as increased physical activity, smoking
cessation, controlling hypertension and improving dietary habits could possibly
reduce CVD risk in patients on ARV therapy (Mallon & Sattar, 2008). There are
indications that adherence to a Mediterranean diet (Turcinov et al., 2009) and high
quality diet rich in fibre (Hendricks et al., 2003) may be protective against
lipodystrophy in HIV patients on ARV treatment. Furthermore a prescribed low fat
diet has been associated with decreased lipid levels in people with HIV (Barrios et
al., 2002).
1.6 Diet and CVD in Botswana
Data on diet and CVD in Botswana is limited, searches were conducted in electronic
database (Ovid Medline) in process & other non-indexed citations and Ovid medline
34 | P a g e
(R) from 1946-2011. References from papers and reports were also searched, and
contact was made with researchers and nutrition experts in Botswana for any
unpublished papers or reports. Table 1.4 shows the number of article from the
literature search strategy.
Table 1.4 Literature search strategy for diet and CVD in Botswana
Search History: 1946 to 2011

1 cardiovascular disease$.mp. or Cardiovascular Diseases/ 85386
2 Heart disease$.mp. or Heart Diseases/ 166109
3 Stroke/ or Stroke.mp. 149136
4 blood pressure$.mp. or Blood Pressure/ 333749
5 Diabetes Mellitus/ 82957
6 Cholesterol, HDL/ or Cholesterol/ or Cholesterol, LDL/ or 121483
Cholesterol, VLDL/ or Cholesterol, Dietary/
7 Obesity/ or Obesity, Morbid/ or Obesity, Abdominal/ or 157218
Obesity$.mp.
8 1 or 2 or 3 or 4 or 5 or 6 or 7 961056
9 Diet, Protein-Restricted/ or Diet, Gluten-Free/ or Diet, Atherogenic/ 456729
or Diet, Vegetarian/ or Diet/ or Ketogenic Diet/ or Diabetic Diet/ or
Diet Therapy/ or Diet, Reducing/ or Diet, Fat-Restricted/ or Diet,
Carbohydrate-Restricted/ or Diet, Macrobiotic/ or Diet Records/ or
Diet, Sodium-Restricted/ or Diet Surveys/ or Diet, High-Fat/ or Diet,
Cariogenic/ or diet$.mp. or Diet, Mediterranean/ or Diet Fads/
10 fat$.mp. or Dietary Fats, Unsaturated/ or Dietary Fats/ or Fats, 617601
Unsaturated/ or Fats/
11 Dietary Fiber/ 11579
12 alcohol$.mp. 273040
13 Dietary Carbohydrates/ or Carbohydrates/ or carbohydrate$.mp. 148831
14 protein$.mp. or Dietary Proteins/ 3018685
15 fruit$.mp. or Fruit/ 568228
16 vegetable$.mp. or Vegetables/ 34680
17 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 4162862
18 subsaharan africa.mp. or "Africa South of the Sahara"/ 6874
19 botswana.mp. or Botswana/ 1331
20 18 or 19 8065
21 8 and 17 and 19 8
35 | P a g e
8 papers were identified (Jackson et al., 2010; Mengesha, 2007; Clausen et al., 2006;
Walker & Walker, 2001; Clausen et al., 2000; Edmunds, 1979; Truswell, 1977;
Truswell et al., 1972). 3 of the papers (Jackson et al., 2010; Walker & Walker, 2001;
Edmunds, 1979) were not related to diet and CVD and one was a symposium paper
(Truswell, 1977). 5 additional papers (Letamo, 2011; Maruapula et al., 2011; Garrido
et al., 2009; Maruapula et al., 2007; WHO Botswana STEPS survey, 2007) were also
found through other references and contacts with researchers in Botswana (table
1.5).
As shown in table 1.5 there is limited research on CVD risk factors in Botswana. All
studies are cross-sectional with varied representation of the population groups. For
example, Garrido et al (2009) focused his investigation on hospital staff , Mengesha
(2007) looked at CVD risk factors in diabetes patients , Clausen et al (2000 & 2006),
studied the elderly. Only two studies looked at CVD risk factors in a representative
sample of the population (WHO Botswana STEPS surveys, 2007 and Letamo (2011).
36 | P a g e
Table 1.5 Studies measuring CVD risk factors and diet in Botswana
Study Design Study No. Of Age(yr) Findings

population participants Obesity Blood pressure Other
Blood pressure of Kung Descriptive Men and 152 (73 F, 79 M) 15-83 Systolic blood pressure – 117-126 mmHg Sodium chlorid e exc ret ion – 30mEq (~2g/day)
bushmen in Northern study women (Kung Diastolic blood pressure – 64-73 mmHg
Botswana. Bushmen)
Trusswell, (1972)
Mobidity and he althc are Descriptive Elderly ( men & 337 (228 F, 109 M) ≥ 60 Sev erely malnourish ed (BMI <16) – 7% High blood pressure preval enc e - 19%, 95% Smoking – 11% (70% men)
utilisation amon g elderly in study women) Moderat ely malnourished (BMI 16-18.5) – 14% CI. (15-23) Alcohol use – 56% (80% men)
Mman kgod i village. Normal (BM I 18.5-29.9) – 71% Mean seru m cholesterol – 4.19 mmol/L 95%
Clausen et al (2000) Obese (BMI ≥30) – 7 % CI.(4.07-4.31)
Mean seru m triglycerides – 1.58 mmol/L 95% CI.
(1.50-1.60)
Nutritional status , tob acco Descriptive Elderly ( men & Underwei ght (BMI< 18.5) – men 20.1% - Alcohol consumption – 34%
use and alcohol women) Overweight (BM I 25-29.9 – women – 21.3% Tobacco use – 39%
consumption of older Obesity (BMI ≥ 30) – women – 27.9%
people (abstract) C lausen
et al (2006)
Health and dietary p atterns Cross sectional Elderly ( men & 1086 (560 F, 19,M) 60-99 - - Widely consumed foods; tea (91%), sorghu m
of the elderly. women) (82%) and maiz e meal (63%)
Maru apula et al (2007) Dietary pat terns; B eer, M eat /fruit ,
Veget able/bread, Seasonal produc e
Milk/t ea/candy patt erns
WHO Botsw ana STEPS Cross sectional Adults 4003 25-64 Overweight – M en 22.1%, Women 53.4% Raised SBP ≥ 140mmHg and DBP ≥ 90mmHg Current smokers – M en 32.8%, Women 7.8%
survey. 2007 Obesity - Men 5.6%, Wo men 24.6% – Men 28.8 %, Wo men 37.0% Alcohol consumer (current) Men 30.3% ,
Women 8.8 %
% not engagin g in v igorous physic al activit y -
Men 59.5%, Women 84.2%
Hypertension and related Cross sectional Adults 401 (287 F, 114 M) 30-70 - 56.4% obese (BMI ≥ 30) Hypert ension 33.5% high cholest erol
risk fac tors in type 2 27.2% - overweight (BM I 25-29.9) Men - 53.1% 38.9% high tri glyc erid es
diabetes mellitus in Women – 64.4%
Gaborone city clin ics
Botswana
Mengesha. (2007)
Metablolic syndro me and Cross sectional Adults 150 (106F, 44 M) 22-65 Metabolic syndro me – 34% Hypert ension - 44% Low HDL – 80%
obesity among wo rkers at Obesity – 28.7%,
Kanye SD A hospital. Overweight – 27.3%.,
Garrido et al. (2009) High waist circumf erence – 42%
The prevalence of and cross sectional Adults 9023 (4916 F, 4107 20-49 Underwei ght (BMI <18.5) – M en 19.1%, Wo men
factors associated with M) 11.5%
overweight and obesity in Overweight (BMI 25-29.9) – Men 12.9%, Wo men
Botswana. 23.3%
Letamo (2011) Obesity (BMI ≥ 30) - Men 3.2 %, Women 15.4%
*Ov erwei ght and obesity increased with age,
37 | P a g e
urbanisation, education
Socioeconomic status an d Cross sectional Adolescents 704 (432 F, 272 M) 14.9 Suggestive ev idenc e of High SES associated with :
urbanis ation linked to (mean) -More snacks serving
snacks and obesity in -Fewer s ervin g of traditional di et
adolescents in Botswan a. - High overweight and obesit y
Maru apula et al. (2011)
Chapter 1: Aims and objectives
Recent medical research and health policies in sub-Saharan Africa have,
understandably, predominantly focused on efforts to tackle the communicable
diseases such as the HIV pandemic in the region, with strategies to prevent and control
chronic diseases, including CVD, being largely neglected. Although rates of CVD in Sub-
Saharan Africa appear to be increasing rapidly, there is little evidence about the
potential determinants of CVD in the region, especially among Black populations. It has
been suggested that populations in sub-Saharan Africa are in economic and cultural
transition (Vorster et al, 2005), with suspected increases in rates of obesity (Tibazarwa
et al., 2008; Vorster et al, 2007) and hypertension (Tibazarwa et al., 2008; Schutte et
al., 2005) possibly reflecting changes in diet and lifestyle (Vorster, 2002). Furthermore,
it has been suggested that a marked increase in the occurrence of CHD in urban
African population will be inevitable in the next generation (Walker et al., 2002).
Previous studies (Truswell et al., 1972; Truswell, 1977, Clausen et al., 2000), including
the landmark pan-global INTERHEART study (Yusuf et al, 2004; Steyn et al., 2005), have
provided some insight into the relevance of established risk factors to CVD in sub-
Saharan Africa and have demonstrated the feasibility of conducting such investigations
in the region. Although the INTERHEART study is one of the largest studies of CVD to
be conducted in Africa, it involved a total of only 157 MI cases and 369 controls who
were Black Africans (Yusuf et al., 2004), including only 16 cases and 30 controls from
Botswana (Steyn et al., 2005): too few to reliably characterize the nature of any
associations. An important challenge, therefore, is to establish resources for CVD
research of appropriate scale in a rigorous yet cost-effective manner in sub-Saharan
populations where there is unmet scientific and public health need. The established
38 | P a g e
medical and research infrastructure together with the economic and political stability
of Botswana provide an ideal foundation upon which to build such a resource.
National public health policies based on evidence from rigorously conducted research
into the causes and determinants of CVD in Botswana, therefore, could help deliver
successful and regionally appropriate strategies to tackle CVD.
39 | P a g e
1.7 Aims and objectives
The study’s overall aim was to systematically investigate risk factors for CVD and
their association with diet, lifestyle and environmental factors in the most
populated urban city Gaborone in Botswana.
Specific objectives:
o To measure and document CVD risk factors such as diet, biochemical
indicators and lifestyle in Gaborone, Botswana
o To investigate the association between diet (fat, fibre, alcohol, dietary
patterns and food groups) and CVD risk factors
o To assess the association between diet and risk of disease.
40 | P a g e
Chapter 2 Methodology
Chapter 2
Methodology
41 | P a g e
2 Methodology
This study the Botswana “Pelo” Initiative (“Pelo” means heart in Setswana) was a
cross-sectional assessment of risk factors for CVD and their associations with dietary
variables in participants recruited from the clinics and workplace wellness programs
in Gaborone the capital city Botswana.
2.1 Introduction to Botswana
Botswana is a landlocked country sharing borders with South Africa to the south,
Namibia to the west, Zimbabwe in the north east and Zambia in the north (figure
2.1). It has a surface area of more than 580 000 square kilometres and a population
of about 2 million inhabitants. The capital and the most populated city is Gaborone
with over 200 000 inhabitants. More than 35% of the population are younger than
15 years old, about 60% are aged between 15-64 years and 5% are 64 years or
older. The median age of the population is 21.9 years (22.7 years females versus
20.7 years males), and females out-number males with 92 males for every 100
females. The fertility rate in Botswana stands at 3.2/1000 women with rural areas
having double the fertility rate compared to cities and towns, and infant mortality
rate stands 51/1000 while the population crude death rate stands at 11.2/1000
largely due to the high prevalence of HIV/AIDS in the general population reported at
17.6% (Central Statistics Office, 2009).
42 | P a g e
Gaborone
Figure 2.1 Map of Botswana showing population distribution
(www.botswanatourism.us)
43 | P a g e
The population distribution in Botswana is high in the eastern part where soils are
fertile and rainfall patterns are better (Figure 2.1). Almost 60% of the population
live in urban cities and villages. It is projected that the population of Botswana will
increase modestly to over 2.4 million by 2031 (Central Statistics Office, 2001). These
projections are based on the 2001 census, and the mortality effects of HIV/AIDS
have been factored in (Central Statistics Office, 2004).
Economically, Botswana has experienced some of the world’s fastest growth with
an average growth of >6% in the 1970s through 1990s mainly due to diamond
revenues. Consequently, there have been major infrastructure and human
developments. As a result, 97% of the population has access to safe drinking water,
almost 90% of the people live within 15 km of a health facility, and care is accessed
at a nominal fee of BWP5 (£0.50). Botswana has literacy rates of more than 80%.
Poverty rates have also been steadily declining, from 59% in the 1980s to a current
level of 23.4% in 2004. The greatest burden of poverty is borne more by the rural
population (Central Statistics Office, 2009).
2.2 Roles in the study
The author conceptualised, designed and sought funding for the study with support
from supervisors. He also trained and supervised the research assistants , and was
also responsible for the overall management of the study. He recruite d, interviewed
and carried out physical measurements of participants with the assistance of the
research assistants. The author carried out all statistical analyses.
44 | P a g e
Blood samples were drawn by phlebotomists from Diagnofirm Medical Laboratories
(DML) in laboratory depots nearest to the recruitment sites. Blood sample analyses
and storage was carried out at DML. All data entry was carried out by research
assistants supervised by the author and a biomedical statistitian (Dr Stoffel Moeng)
from the University of Botswana. Formatting and analysis of data from the FFQ was
undertaken by Dr Maria Jackson a nutritionist at the University of Jamaica Mona
campus.
2.3 Ethical considerations
Ethical approval for the pilot phase was sought from the Botswana Ministry of
Health Research and Development Division (protocol number 00551) (Appendix 1).
This investigation conformed to the principle of the Declaration of Helsinki (2008).
The research objectives of the study were explained to all participants in English or
Setswana depending on participants’ preference and participation on a voluntary
basis was highlighted. No payment was made to participants, however
refreshments were offered at the end of measurements. Participants gave informed
verbal and written consent, and a copy of the information sheet (appendix 2) with
contact details of the author was given to the participants.
2.4 Recruitment centres
The recruitment of participants was carried out in Gaborone Botswana. Gaborone is
the capital city and also the most populated city in the country with over 200,000
inhabitants, accounting for about 14% of the country’s population. Recruitment of
participants without known CVD (“healthy”) was carried out in workplaces including
45 | P a g e
offices (public or private), retailers, and open market vendors in the city of
Gaborone. Participants with known CVD/hypertension/diabetes (“Diseased”) were
recruited in one private clinic (cardiac clinic) and two public clinics (Extension 2 &
Nkoyaphiri). One public clinic (Extension 2) is in the centre of t he city while the
other (Nkoyaphiri) is located in the outskirts of Gaborone city. The cardiac clinic is
for paying clients and has been operating since 2002 and specialises in CVD,
diabetes and general care. The public clinics provide a wide range of gene ral care
including general consultation, wound dressing, injections, counselling, maternal
and child care as well as routine check up for patients such as hypertension and
HIV/AIDS patients.
Individuals attending the clinics were approached and invited to participate in the
study during their appointment days. Identification, recruitment and survey of all
study participants were carried out by the author assisted by two trained research
assistants fluent in both English and Setswana. All aspects of the recruitment
process were overseen by author.
2.5 Study participants
Participants were eligible for inclusion if they (i) were ≥18 years (ii) had a
previous/known history of cardiovascular disease (defined as any of myocardial
infarction, stroke, heart failure, rheumatic heart disease or coronary heart disease);
or (iii) had reported type 2 diabetes or hypertension, (iv) were essentially “healthy”
with no known history of CVD or diabetes and hypertension. Eligible participants
were included in the study if they provide d informed written consent. Individuals
who did not provide such consent or were pregnant were excluded.
46 | P a g e
“Healthy” participants were recruited from people attending clinics for non CVD
related sickness, or those accompanying them, and from workplace wellness
program around Gaborone city. Workplace wellness programs are programs within
the workplace created initially to raise awareness of HIV/AIDS, but are now used to
promote general health and wellbeing of employees. The program was initiated by
government and the private sector has since embraced it. Workplaces such as
offices (public and private), retailers, and street vendors were approached and
objectives of the study were explained to management and permission to talk to
staff sought. Participants recruited from the workplace wellness programs were
considered to have no reported CVD or diabetes or hypertension unless otherwise
established in the selection criteria.
2.6 Sample size
The study aimed to recruit a total of 1000 participants including about 200 (20%)
individuals with pre-existing confirmed CVD, about 200 (20%) individuals with type
2 diabetes or hypertension and about 600 (60%) “Healthy” individuals as defined
earlier. Sample sizes were estimated based on the proportion of the general
population estimated to be hypertensive (8.18%) (Central Statistics Office, 2007). A
total of 1000 study participants in cross-sectional analyses are expected to provide
>99% power to detect an additional 5% in the variance explained by a variable of
interest (in this case dietary variables e.g., fibre intake) at a significance level of 0.01
in a multiple regression with 10 covariates which together account for 50% of the
variance of a risk factor of interest (e.g., serum cholesterol). A subgroup of 200
47 | P a g e
study participants would provide >90% power to detect an additional 5% of the
variance explained at a significance level of 0.01 for the analysis described above.
2.7 Questionnaire
A structured questionnaire was used to collect data on socio-demographic factors
(age, sex, ethnicity, and place of birth, education, occupation, and household
income), lifestyle (smoking habits, physical activity, and alcohol consumption),
personal and familial medical history and diet (appendix 3). The questionnaire was
developed by the author and was administered by face-to-face interview by the
author and two research assistants fluent in English and Setswana.
2.7.1 Socioeconomic status
Household estimated income was assessed using the same categories used in the
Botswana Food Consumption Survey (Jackson, 2010). T he categories of household
income in Botswana Pula (BWP), (1 BWP ~ 0.1Pound sterling) were as follows < 600
Pula, 600 – 3000, ≥ 3000 – 10 000, and > 10 000 per month. Households earning <
600 Pula were considered to be living below the poverty datum line.
2.7.2 Education
Education level of participants was defined as (1) no formal education, (2) Primary
education (1-7 years), (3) Secondary education (8-12 years) and (4) Tertiary
education (> 12 years) which included vocational training, college and university
48 | P a g e
qualification. Participants were classified in these respective categories even if they
had not completed them.
2.7.3 History of cardiovascular diseases
Participants were asked about their history of CVD and approximate age of
diagnosis. Participants were also asked the same questions about their family
(mother, father, brother, sister, son or daughter). It was not possible to ascertain
diagnosis of CVD therefore the CVD/diabetes/hypertension status is participant
reported. But given that most of the patients were recruited at clinics where they
were being reviewed for the same conditions it is likely that the reported statuses
were reasonably accurate.
2.7.4 Smoking
Participants’ smoking habits were assessed using a series of questions which
included asking about whether they have smoked at least one cigarette per day for
at least one year. Length of smoking was assessed by further asking about the age
the participants started smoking and their smoking habits in the time thereafter.
(figure 2.2)
49 | P a g e
26) Did you smoke at the following ages? If yes, how many cigarettes did you smoke each day?
Under 20 non-smoker  smoker  cigarettes each day

Age 20 non-smoker  smoker  cigarettes each day
Figure 2.2 Sample of smoking history questionnaire
Smoking was expressed in total grams of tobacco per week (1 cigarette = 1 gram).
One pack year was defined as one packet, or 20 g tobacco, smoked each day over a
course of 1 year (Prignot, 1987).
For example, pack years of a 34 year old man who smoked 5 cigarettes/day
between the ages of 13 and 20 years, 7 cigarettes/day at age 20-30 years, and 12
cigarettes/day at age 30-34 years was calculated as follows;

Pack years = 5 * (20-13) +7 * (30-20) + 12 * (34-30) = 7.25 pack years
20
50 | P a g e
2.7.5 Physical activity questionnaire
Physical activity was assessed using a questionnaire based on the Framingham
physical activity index (PAI) (Kannel & Sortie, 1979). Participants were asked to
estimate their physical activity in the last 24 hours, specifically the number of hours
spent sleeping, doing sedentary, light, moderate and heavy physical activities at
work and during their leisure time. The physical activity index composite score was
calculated by summing the number of hours spent in each activity intensity level
and multiplying by a respective weight factor derived from the estimated oxygen
consumption requirement for each intensity level. Table 2.1 shows physical activity
status and weight factor (Kannel & Sortie, 1979).
Table 2.1 Physical activity status and weight factors
Physical activity status Weight

factor
Sleep 1.0
Sedentary 1.1
Light activity 1.5
Moderate level of activity 2.4
Heavy activity 5.0
51 | P a g e
For example, a person who sleeps 10 hours spends 10 hours doing moderate level
activity and 4 hours doing heavy activity had a physical activity index of:
PAI = Sleep (10hours x 1) + mod. activity (10 hours x 2.4) + heavy activity (4 hours x 5.0) = 54
Physical activity level (PAL) was calculated by dividing the PAI by 24 hours. All
calculations were carried out in a template made in excel.
PAL = Sleep (10hours x 1) + mod. activity (10 hours x 2.4) + heavy activity (4 hours x 5.0) = 2.25
24hrs
Participants PAL were then classified according to categories as shown in table 2.2
(FAO/WHO/UNU, 2004).
Table 2.2 Physical activity levels for different lifestyles
Lifestyle Example PAL
Extremely inactive No exercise <1.40
Sedentary Office work, light household chores, little or no exercise 1.40-1.69
Moderate Construction worker or person running 1 hour a day 1.70-1.90
Vigorous active Agriculture wo rker (manual), swimming, walking long distances 2.00-2.40
Extremely active Co mpetitive sport e.g cycling >2.40
2.7.6 Use of hormone replacements or birth control
Female participants were asked about their use of birth control pills or injection.
Age when they started and stopped was recorded. Menstruation in the last 12
months was recorded and if menstruation had stopped age of last menstruation
was recorded. Furthermore information on use of hormone replacement therapy,
length of use and type used was collected.
52 | P a g e
2.7.7 Food frequency questionnaire (FFQ)
Long term dietary intake was assessed using a pre-piloted and validated (Jackson et
al, submitted) 122 item FFQ (appendix 4) developed by the National Food
Technology Research Centre, University of Botswana, Ministry of health and
University of the West Indies to assess intake in the Botswana National Food
consumption Survey (Jackson et al., 2010). Food items are categorised into main
food groups including cereals; peas, beans and nuts; milk and milk products; meat;
fruit and vegetables; and beverages. Participants were asked to specify their
frequency of consumption of each of the food items in the last 3 months as “almost
never”, “once a month”, “2-3 times a month”, “once a week”, “2-4 times a week”,
“5-6 times a week”, “once a day” or “more than twice a day”. Figure 2.3 shows a
sample from the FFQ.
Foods from Almost Once/ 2-3 Once/ 2-4 5-6 Once/ 2+

Animals never mth / week / / day / Portion size
mth week week day
Eggs
Chicken with skin
Chicken without
skin
Figure 2.3 Sample of the FFQ
Household measuring utensils and a picture of foods were used to help participants
estimate their usual portion size. Data entry and nutrient analysis were performed
using South Africa’s Medical Research Council (MRC) food finder 3 adapted for
Botswana’s indigenous food items and the data was exported to SPSS 19.
53 | P a g e
Nutrient content of food items was calculated using the South African food-finder
software based on the South African food composition tables (Langenhoven et al.,
1991) and food composition of some indigenous Botswana foods. The average
weight of the food items was used for the determination of portion weights in
grams. Daily nutrient intakes were calculated from the questionnaire by multiplying
the frequency of intake by the nutrient composition specified for each food item
and its portion weight. Nutrients from all foods were summed to obtain a total
nutrient intake for each participant.
2.7.7.1 Alcohol drinking
Information on alcohol use was assessed by asking participants about their drinking
habits, type of drink and number of drinks per day, as well as age which they started
drinking and age when they stopped drinking. Alcohol intake was calculated from
the FFQ.
2.8 Biophysical measurements
2.8.1 Body weight and height
Weight was measured using an electronic measuring scale (Seca gmbh @ co
Germany model 8821321138). Prior to measurement, scales were calibrated using a
standard calibration weight and placed on a hard and flat surface. Participants were
requested to remove shoes and heavy clothing, empty their pockets and stand on
the scale with weight evenly distributed between both feet, and head facing
forward. Weight was measured twice and recorded to the nearest 0.05kg. Average
of the two measurements was used in analysis.
54 | P a g e
To measure height, participants were asked to take off any hair ornaments that
could obstruct height measurement. Height was measured using a portable
stadiometer (Seca gmbh @ co Germany). The participants were asked to stand with
heels together, arms to the side, legs straight, shoulders relaxed and the head
looking straight forward. The researcher ensured that heels, buttocks, scapulae and
the back of the head were against the vertical surface of the stadiometer. The
participants were asked to inhale deeply and hold their breath whilst maintaining
an erect position and the headboard was lowered to the highest point with enough
pressure to compress hair. Height was measured twice and read to the nearest
0.1cm and an average of the two measurements was used in analysis.
Body mass index (BMI) was calculated as weight (kg) / height (m) 2. BMI was
categorised as underweight (<18.5 kg/m2), normal (≥ 18.5 < 25.0 kg/m2), overweight
(≥25.0 < 30 kg/m2) and obese (≥ 30 kg/m2) (WHO, 2000).
2.8.2 Waist and hip circumference
Waist circumference (WC) was measured to evaluate participants’ central adiposity.
Prior to measurement, participants were asked to remove any heavy outer clothing.
WC was measured to the nearest 0.1 cm twice midway between the lowest rib and
the superior border of the iliac crest using a non-stretchable standard measuring
tape with measurement taken at end of normal expiration. The researcher was
positioned at the side of the participant such that they could observe the tape at
eye level.
Hip circumference was measured at the level of most protrusion as shown on figure
2.4. HC was measured twice and recorded to the nearest 0.1 cm and an average of
55 | P a g e
the two measurements was used in analysis. Waist-hip ratio was calculated as waist
(cm)/hip (cm). Visceral obesity was defined as WC ≥ 102 cm in men and ≥ 88cm in
women, and waist-to-hip ratio (WHR) was defined as high when ≥ 0.9 in men and
≥ 0.85 in women (WHO, 2000).
Figure 2.4 Measuring waist and hip circumference
2.8.3 Body composition
Body composition refers to the amount of lean body mass, bone and body fat that
makes up total body weight. Body composition was measured by bioelectric
impedance analysis (BIA) (Bodystat 1550 version 2/02 UK). BIA works under the
principle that a lean compartment of the body consisting of 60-75% electrolytic
water conducts electricity better than adipose tissue which is very low in water
content (5-10%). The impedance measurement therefore reflects the degree of
resistance to the flow of current in the body (University of Vermont,
http//nutrition.uvm.edu accessed 8 April 2009).
There are conditions that can interfere with BIA body composition measurements,
so participants are usually be advised to avoid the following:

56 | P a g e
 Vigorous exercise 8-12 hours before measurement
 Alcohol 24-48 hours before measurement
 Eating or drinking at least 2 hours before measureme nt
However we were unable to give this advice prior to measurement. At
measurement participants were asked to remove all metallic objects, shoes and
socks. Participants were asked to lie on a non-conducting surface for at least 5
minutes in a lying position, ensuring that participant’s arms did not touch the rest of
the body, the thighs were not touching and the feet were at least 20cm apart. The
participant’s wrist, hand, ankle and foot were cleaned with an alcohol swab, and
self-adhesive disposable electrodes were placed on the right hand and foot, as
shown in Figure 2.5.
Figure 2.5 Measuring body composition with bio-impedance analysis
Adapted from http://www.integrativetherapieswellness.com/bioimpedance-testing.html (accessed 7
April 2011)
A battery generated signal (500 micro Amps @ 50kHz) was then passed through the
body to measure impedance. Age, height, weight and gender were entered when
57 | P a g e
prompted. The procedure was done twice and readings of impedance, total body
water, fat free mass and fat mass were recorded as displayed in the BIA monitor.
2.8.4 Blood pressure and pulse rate
Systolic and diastolic blood pressures were measured using an electronic blood
pressure machine (Microlife BP A 100 Windau/Switzerland). Participants were
requested to remove outer garments and all other tight clothes. The cuff was
placed on the bare right arm while the participant remained in the supine position.
The electronic BP machine was then switched on to start measuring. When the
electronic reading was complete, measurements of BP and pulse rate were
recorded as displayed. The procedure was repeated for a second measurement and
the average of the two was used in analysis. Hypertension was defined as systolic
blood pressure reading ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg
(WHO/ISH, 2003).
2.9 Collection of biological samples and initial biochemistry
Non-fasting blood samples were drawn by phlebotomists from each participant.
Blood samples were drawn in a sitting position to limit the possibility of systematic
differences (e.g. plasma volumes may be higher in the supine position than when
sitting). A total of 35ml of blood was drawn from each participant in 3 x serum
tubes (red top) , 3 x EDTA tubes (purple top) and 1 fluoride oxalate (grey top).
Participants’ ID number and date and time of sample collection were recorded on
water resistant labels adhered to tubes. Samples were stored in ice before being
transferred by a courier to the Diagnofirm Medical Laboratories (ISO 17025 quality
management system certified) where samples were processed within 24 hours.

58 | P a g e
Biochemistry measurements included measurement of total cholesterol (Abbott
Clinical Chemistry cholesterol 7D62-20 30-3126/R3), low density lipoprotein-
cholesterol (Abbott Clinical Chemistry LDL 7D74-20 30-3140/R3), high density
lipoprotein- cholesterol (Abbott Clinical Chemistry HDL 3K33-20 30-4005/R4),
triglycerides (Abbott Clinical Chemistry triglycerides 7D74-20 30-3140/R3), and
glucose (Abbott Clinical Chemistry glucose 3L82-20 and 3L82-40 30-3941/R4). All
analyses were performed using Achitech CI 8200 automated analyser (Abbott
laboratories). LDL cholesterol (LDL-C) was calculated from HDL and total cholesterol
(LDL-C = total cholesterol - (HDL-C) – Trigycerides/5). The rest of the blood samples
were stored at -70 degrees Celsius.
Low HDL-cholesterol was defined as < 1.0 mmol/L (40 mg/L) for men and < 1.3
mmol/L (50 mg/L). High triglyceride level was defined as ≥1.7 mmol/L (150 mg/L),
and High LDL-cholesterol as ≥4.2 mmol/L (160 mg/L). Atherogenic index (AI) = total
cholesterol/HDL cholesterol and AI >5 was considered high risk (NCEP, 2002).
2.9.1 Quality control
Quality control measures during the pilot phase included: (i) a 2 day initial training
session for study staff to ensure understanding of the study protocol and standard
operating procedures; (ii) routine (monthly) checking and calibration of measuring
equipment, (iii) Entry of information from questionnaires, physical and biochemical
measurements onto a central database by trained research assistants supervised by
a statistitian and the principal investigator with random checks of data entry for
accuracy and completeness. The entered data were monitored every 2 weeks by
59 | P a g e
the principal investigator to track recruitment rates and identify potential
recruitment issues, (iv) 2 monthly refresher and feedback sessions for all study staff
to ensure consistency of methods and (v) questionnaires and consent forms were
kept in a locked cabinet (at the National Food Technology Research Centre in
Botswana) and access to the key was restricted to the principal investigator or a
person assigned by him only. Entered data was/is kept in a password-protected
computer.
60 | P a g e
2.10 Statistical Analyses
Statistical analyses were performed using SPSS windows version 19 (SPSS Inc,
Chicago Illinois). All analyses was two tailed with significance set at α = 0.05.
2.10.1 Descriptive statistics
All variables were tested for normality by visually inspecting histograms . Data on
socio-demographics, lifestyle, dietary intake, biophysical and biochemical
measurements were calculated as means ± standard deviation (SD) for normally
distributed variables and median and inter-quartile range for skewed variables.
Categorical variables were presented as frequencies or percentages.
In Chapter 3, socio-demographic characteristics of “Healthy” participants were
compared by sex (male & female) using an independent T-test and Mann-Whitney
test for normal and skewed data respectively. Variation between age groups (<25
yrs, ≥ 25<35 years and ≥ 35 years) was tested using ANOVA and Kruskal-Wallis test
for normal and skewed data distributions respectively. General Linear Models
(GLM) were also used to compare variation of continuous variables by sex and
across age groups. To assess where differences between age groups occurred,
Scheffe’s post hoc test was used. Pearson’s Chi-squared test was used to compare
categorical variables between the age groups and sex.
In chapter 4, associations between tertiles (thirds) of dietary % energy (%E)
variables (Total fat, saturated fat, polyunsaturated fat, fibre) alcohol (g/1000kcal),
dietary pattern, fruit and vegetable groups), and CVD risk factors were estimated by
GLM analysis with adjustments for established CVD risk factors.
61 | P a g e
In chapter 5, comparisons of CVD risk factors and socio-demographic characteristics
were made between “healthy” and “diseased” participants. The relationship of
dietary variables and disease status was explored using logistic regression analyses
with adjustments for risk factors and other dietary variables.
Statistical analyses methods are explained in detail at the beginning of each results
chapter.
62 | P a g e
Chapter 3: Diet and CVD risk factors
Chapter 3
Diet & CVD risk factors
63 | P a g e
3 Diet and CVD risk factors
3.1 Introduction
This chapter describes socio-demographic, anthropometric and dietary characteristics as well
as biochemical indicators of participants with no known CVD, hypertension or diabetes
(“healthy”) in the capital city of Botswana, Gaborone. This chapter also presents comparisons
of participants’ characteristics by sex and age group. A full list of descriptions of food items in
the FFQ is given in appendix 4. The numbers of participants presented in the tables vary due
to missing data. However, where there is variation, the numbers of participants are clearly
indicated.
64 | P a g e
3.2 Statistical methods
3.2.1 Descriptive statistics
3.2.1.1 Socio-demographic and lifestyle characteristics
Socio-demographic and lifestyle characteristics were presented as frequencies and
percentages. Data was presented by sex and age group (<25, ≥25≤35 and >35 years). The
Pearson Chi-squared test was used to compare socio-demographic variables by sex and by
age group. Age, the number of people living in the household and pack years of smoking was
not normally distributed. As a result medians and inter-quartile ranges (IQR) are presented.
The non parametric Mann-Whitney and Kruskal-Wallis tests were used to compare whether
the age of participants and number of people living in the household were significantly
different by sex and age groups respectively.
3.2.1.2 Anthropometric and physical characteristics
Means (SD) and medians (IQR) where appropriate were computed for height, weight, BMI,
HC, WC, WHR, body fat mass, SBP, DBP, heart rate and PAL. To conduct analysis of variance,
univariate General Linear Models (GLM) was used with age group and sex set as fixed factors.
A post hoc Scheffe’s test was used to assess the pair-wise comparison of group means
between each pair of factor levels for age group which had more than two levels. The
prevalence (%) of underweight (BMI<18.5 kg/m2), overweight (BMI ≥25<30 kg/m2) and
obesity (BMI≥30 kg/m2), abdominal obesity WC (>102cm men and >88cm women), WHR
65 | P a g e
(>0.9 men and 0.85 women) and hypertension (SBP>140mmHg/DBP>90mmHg) was
presented by sex and age group.
3.2.1.3 Biochemical indicators
Means (SD) are presented for total cholesterol, HDL cholesterol, LDL cholesterol, random
glucose and atherogenic index (total cholesterol/HDL cholesterol). Triglyceride
measurements were not normally distributed and so medians (IQR) were presented. Analysis
of variance was conducted using GLM with age group and sex as fixed factors. Post hoc
analysis was carried out using Scheffe’s test. The prevalence of low HDL cholesterol (<1.0
mmol/L for men & <1.3 mmol/L for women) and high atherogenic index (>5.0) were also
calculated and presented by sex and age group.
3.2.1.4 Nutrient intake
Under and over reporting of nutrient intake were assessed using three methods. Black et al
(1996) recommend as acceptable physical activity level [Energy intake (EI)/ Basal Metabolic
rate (BMR)] range of 1.2-2.5; where 1.2 represents a PAL of an individual who is inactive and
2.5 represents a PAL of a person undertaking heavy manual work. Willett (1998) suggests
acceptable energy intakes from a FFQ to range from 800kcal-4000kcal (500-3500) kcal for
women & 800-4000 kcal for men). A study on diet and obesity in populations of African origin
in Cameroon, Jamaica and UK used energy cut-offs of 800kcal-5500kcal for acceptable
reporting (Jackson et al., 2006). To determine whether over reporting may have a ffected the
results, socio-demographic and physical characteristics of participants with energy intakes of
66 | P a g e
<5500 and >5500kcal were compared and no significant difference between the groups was
found (table 3.9). Since the three cut-offs excluded a substantial number of participants
65.2%, 55.2% and 30% respectively and no significant differences in characteristics between
over and under-reporters were found, all participants were included in the analysis.
Medians (IQR) are presented for all nutrients which were not normally distributed. The
percentage contribution of nutrients to total energy (energy density) was calculated for all
macronutrients (e.g. 1g 9kcal/g for fat) and fibre intake was calculated per 1000kcal. The
ratio of PUFA and SFA (PS ratio) was also calculated. To conduct analysis of variance
univariate General Linear Models (GLM) were used with age group and sex set as fixed
factors. A post hoc Scheffe’s test was used to assess pair-wise comparison of group means
between age groups.
The percentage of individuals who commonly consumed foods (≥2 times per week by at least
25% of the participants) and their respective median (IQR) intakes are presented. The
definition of commonly consumed foods was based on that used in a food consumption
survey carried out in Botswana (Jackson, 2010). Chi-squared test for difference was used to
determine the differences in commonly consumed foods by sex and age group. To conduct
analysis of variance univariate General Linear Models (GLM) was used with age group a nd sex
were set as fixed factors. A post hoc Scheffe’s test was used to assess pair-wise comparison
of food between age groups.
67 | P a g e
3.2.2 Dietary patterns
Dietary patterns were identified using principal component analysis (PCA). PCA identifies
patterns in data and highlights the similarities and differences in the data. Daily intakes in
grams for all the 122 food items in the FFQ were used in PCA. Due to possible differences in
consumption patterns in men and women (Newby & Turker., 2004) attributable to sex
influenced eating behaviours (Randall et al., 1990), analyses were carried out for all
participants and separately by sex. The number of components (food patterns) was selected
using the scree plot in which eigen values were plotted against each component. The number
of components was chosen based on the point where the scree plots curve begins to level off
(Cartell, 1966). Where the number of components was not clear, straight lines that
summarised the vertical and horizontal part of the plot were drawn and the point of inflexion
was marked where the two straight lines met, and only factors to the left of the point of
inflexion were included (Fields, 2009).
Varimax rotation was then applied to enable easy interpretation of factors. Varimax
orthogonal rotation (which allow factors to be rotated independently) ascertain that food
items are loaded by the factor they relate to most, hence after rotation the loading of
variables are maximised onto one factor and minimised on the remaining factors (Fields,
2009 ). Factor loadings, which are correlations of each food item with dietary pattern were
set at < -0.3 and > 0.3 because that is what is recommended (Fields A, 2009). Comrey & Lee
(1992) have suggested that a sample size of about 200 is far and 300 is good and more than
500 is very good. Given the potential differences in dietary pattern between men and women
68 | P a g e
and the study sample size of 564 (336 women and 229 men), dietary patterns were
generated for all individuals and by sex.
Factor scores were generated for all participants for each dietary pattern. To calculate the
factor score, participants’ standardised score for each food item was multiplied by the
corresponding factor loadings of the food item and summed across food types.
Patterns were named quantitatively (Newby & Turker., 2004) by tallying food items with food
groups in the FFQ and composing a name that suited the dietary pattern.
69 | P a g e
3.3 Socio-demographic characteristics
3.3.1 Socio-demographic characteristics by sex
Socio-demographic characteristics of the study participants are shown in table 3.1. There
were 559 participants of whom 335 (60%) were women and 224 (40%) were men. The
median (IQR) age of the participants was 27 (23, 32) years. 36.3% of the participants were
aged below 25 years, the majority (43.4%) were aged 25-35 years, while 13.3% and 6.9%
were aged 35-45 years and > 45 years respectively. Women were older than men (27 versus
25 years, p=0.001). Figure 3.1 shows that almost half of the men were younger than 25 years
compared with 30% of women.
Only 2.1% of participants had no formal education, and 10% attained primary education only
while the rest had a secondary school education or better. There was no significant
difference in education level between men and women.
Over 87% of the participants were single, while 11% reported being married; the rest of the
participants were divorced, widowed or separated. About 5% more men were single
compared to women. Conversely, proportionally more women were married compared to
men. However, the differences between the sexes in marital status were not statistically
significant.
Religious beliefs were varied amongst the study participants, with the majority being
Christians (women 81.3% versus men 58.9%) and the rest being either of traditional beliefs
(including beliefs in ancestors and taboos), Muslims or holding no religious belief.
70 | P a g e
Table 3.1 Socio-demographic characteristics by sex: n (%)
All Men Women p-value for

difference *
n 559 224 335
Age (years)† 27 (23, 32) 25 (22, 30) 27 (24, 33) 0.001 ‡
Education
No formal (0 years) 12 (2.1) 4 (1.8) 8 (2.4)
Primary (0-7 years) 56 (10.0) 19 (8.3) 37 (11.0) 0.517
Secondary (8-12 years) 367 (65.7) 155 (69.2) 212 (63.3)
Tertiary (> 12 years) 124 (22.2) 47 (20.5) 78 (23.3)
Marital status
n 565 229 336
Single 495 (87.6) 207 (90.4) 288 (85.7) 0.115
Married 63 (11.2) 22 (9.6) 41 (12.2)
Divorced/widowed/separated 7 (1.2) 0 7 (2.1)
Religion
n 551 224 327
Christianity 398 (72.2) 132 (58.9) 266 (81.3) < 0.001
Traditional beliefs 60 (10.9) 41 (18.3) 19 (5.8)
Islam/other 7 (1.3) 5 (2.2) 2 (0.6)
None 86 (15.6) 46 (20.5) 40 (12.2)
*Pearson Chi-squared test, ‡P-value Mann-Whitney test, †Median (IQR)
Men Women
<25 years ≥ 25 <35 years ≥ 35 years <25 years ≥ 25 <35 years ≥ 35
16% 23% 30%

45%
39% 47%
Figure 3.1 Age distribution of participants
71 | P a g e
As shown in table 3.2, marginally more women reported owning a house than men (41.2%
versus 33.5%, p=0.065). Household possessions such as car, cellular phone, television, radio,
and computer were not significantly different between men and women. However, men
were significantly more likely than women to own livestock (55.3% versus 36.9%) and a
bicycle (15.7% versus 9.6%).
In total 17.3% of participants reported being current smokers, and proportionally more men
than women smoked (35.4% versus 5.0%, p < 0.001). The median number of pack years (0.55
pack years) was relatively low, and men had a significantly higher number of pack years than
women (0.75 versus 0.15). The median number of people living in the household was 3 in
both men and women.
42.6% of participants had a family history of CVD (Stroke, MI, diabetes, hypertension or
sudden death). Women were significantly more likely than men to have a family history of
CVD (47.8% versus 34.9%), hypertension (41.2% versus 25.3,) and diabetes (9.6% versus
6.1%).
72 | P a g e
Table 3.2 Lifestyle characteristics of participants by sex: n (%)
All Male Female p-value †
n 562 227 335
Household ownership
House 214 (38.1) 76 (33.5) 138 (41.2) 0.065

Car 108 (19.1) 39 (17.2) 69 (20.8) 0.289
Bicycle 66 (12.1) 35 (15.7) 31 (9.6) 0.031
Cellular phone 542 (96.4) 219 (96.5) 323 (96.4) 0.971
Television 404 (71.9) 159 (70.0) 245 (73.1) 0.424

Radio 431 (76.8) 175 (77.4) 256 (76.4) 0.780
Computer 119 (21.5) 45 (20.1) 74 (22.4) 0.511
Livestock 247 (44.3) 125 (55.3) 122 (36.9) < 0.0001
Smoking
Current Smoker 98 (17.3) 81 (35.4) 17 (5.0) < 0.001
Pack years * (smokers only) 0.55 0.75 0.15 <0.001‡

(0.15, 2.08) (0.25, 2.33) (0.10, 1.00)
n 564 228 336

No. Living in household* 3 (2, 5) 3 (1, 4) 3 (2, 5) 0.001‡
Family history (mother/father)
Stroke, MI, diabetes, hypertension 241 (42.6) 80 (34.9) 161 (47.8) 0.002
or sudden death
Hypertension 197 (34.8) 58 (25.3) 139 (41.2) <0.001
Diabetes 46 (8.2) 14 (6.1) 32 (9.6) 0.147
†Pearson Chi-squared test for difference between men and women * Median (IQR), ‡ Mann-Whitney test
26% of participants had a monthly household income below the Botswana poverty datum
line of <600 Botswana Pula (BWP) per month (CSO BIDPA, 2004), [BWP600 ~ £60 (US100)],
73 | P a g e
and 54.9% earned between BWP600 and BWP3000, while the rest (19.2%) of the participants
had a monthly household income of more than BWP3000 (figure 3.2). There was no
significant difference in income between men and women.
70
60 p=0.41 for difference between men

59.1 and women
54.9
50 52.1
40
%
All
30
Men
27.7
26.0 Women
20 23.6
17.0 18.2
15.1
10
2.2 2.2 2.1
0
<600 >600≤3000 >3000≤10000 >10000
Income (BWP)/month
Figure 3.2 Income categories of participants by sex
74 | P a g e
3.3.2 Socio-demographic characteristics by age group
Table 3.3 shows socio-demographic characteristics of the participants by age group. As
indicated in table 3.1, females and males represented 60% and 40% of all participants
respectively. However, there was an equal proportion of men and women in the youngest
age group (<25 years), whereas in the other two age groups about two thirds of participants
were women while the rest were men (p=0.001).
Education level was significantly different (p<0.001) by age group: more than 90% of
participants younger than 35 years had a secondary school education or better, whereas only
65% of participants aged 35 years or older had the same level of education.
All participants younger than 25 years were single, and only 9.8% of participants aged ≥25<35
years were married while 33.3% of those 35 years and older were married (p<0.001).
About 70% of participants younger than 35 years held Christian beliefs while 83.3% of
participants older than 35 years held the same belief (p=0.040). One third of participants
younger than 25 years had no religion or simply subscribed to traditional beliefs.
75 | P a g e
Table 3.3 Socio-demographic characteristics of participants by age group: n (%)
<25 years ≥25<35 years ≥35 years p-value *

n (%) 205 (36.4) 245 (43.4) 114 (20.2)
Men 103 (50.2) 88 (35.9) 37 (32.5) 0.001
Women 102 (49.8) 157 (64.1) 77 (67.5)
Education
n 200 244 113
No formal (0 years) 4 (2.0) 3 (1.2) 5 (4.4)
Primary (0-7 years) 6 (3.0) 16 (6.6) 34 (30.1) <0.001
Secondary (8-12 years) 160 (80) 160 (65.6) 47 (41.6)
Tertiary (> 12 years) 30 (15.0) 65 (26.6) 27 (23.9)
Marital status
n 205 244 114
Single 205 (100) 219 (89.8) 70 (61.4)
Married 0 24 (9.8) 38 (33.3) <0.001
Divorced/Widowed 0 1 (0.4) 6 (5.2)
/separated
Religion
n 201 241 108
Christianity 136 (67.7) 172 (71.4) 90 (83.3) 0.040
Traditional beliefs 20 (10.0) 31 (12.9) 8 (7.4)
Islam/Other 2 (1.0) 5 (2.0) 0
None 43 (21.4) 33 (13.7) 10 (9.3)
* Pearson Chi-squared test for difference between age group
House ownership was more common in the oldest age group (52.2%) than in participants
younger than 25 years (35.8%) and those aged ≥25<35 years (33.3%) (p=0.002) (table 3.4).
Participants in the oldest age group were also more likely to own a car (p<0.001). Other
household possessions such television and radio were significantly more likely to be owned
by ≥25 years than those aged <25 years. However, ownership of bicycle, cellular phone,
computer and livestock were not significantly different between the age groups.
The proportion of current smokers was not significantly different between the three age
groups. However, as expected, there was a significant increase in pack years with age
(p<0.001). The median number of people living in the household was highest in the oldest
76 | P a g e
age group followed by the youngest and middle age group (p=0.03). Family history of
diseases was significantly higher in the oldest group for CVD and hypertension but not
diabetes.
Table 3.4 Lifestyle characteristics of participants by age group: n (%)

n 205 241 114
Household ownership
House 73 (35.8) 81 (33.3) 59 (52.2) 0.002
Car 24 (11.8) 47 (19.4) 36 (32.1) <0.001
Bicycle 30 (14.9) 26(11.0) 10 (9.3) 0.270
Cellular phone 195 (95.6) 239 (98.4) 106 (93.8) 0.070
Television 120 (58.8) 194 (79.8) 88 (77.9) <0.001
Radio 134 (65.7) 206 (85.1) 89 (78.8) <0.001
Computer 35 (17.3) 56 (23.4) 27 (24.3) 0.210
Livestock 77 (37.9) 194 (79.8) 55 (49.1) 0.069
Smoking
Smokers n (%) 35 (17.1) 44 (18.0) 18 (15.8) 0.880
Pack years 0.25 1.1 2.1 < 0.001‡
(Smokers only) (0.1,0.5) (0.4,2.2) (0.3,6.6)
No. Living in household 3 (2, 5) 2 (1, 4) 4 (2, 5) 0.030‡
Family history
(mother/father)
Stroke, MI, diabetes, 66 (32.2) 110 (44.9) 63 (55.3) <0.001
hypertension or sudden
death
Hypertension 55 (26.8) 91 (37.1) 49 (43.0) 0.008
Diabetes 11 (5.4) 24 (9.8) 11 (9.6) 0.195
* Pearson Chi-squared test for difference between age groups,†Median (P25, P75) ‡ Kruskal-Wallis test,
As shown in Figure 3.3, 31.5 % of participants age below 25 years, 24.1 % aged ≥25<35 year
and 20.1% aged 35 years and older reported a monthly income below the Botswana poverty
datum line (BWP 600). Over half of all age groups earned a monthly income of BWP
77 | P a g e
>600≤3000. Around 10% of participants younger than 25 years earned more than BWP
3000/month compared to more than 20% in the two older groups (p<0.001).
70 P<0.001 for difference

between age group
60
58.1
<25 years
50 52.7
53.2 ≥25<35 years
40
≥35 years
%
30 31.5
20 24.1
22.9
20.7
20.2
10
9.4
1 2.5 3.7
0
<600 >600≤3000 >3000≤10000 >10000
Income (BWP)/month
Figure 3.3 Income of participants by age group
78 | P a g e
3.4 Anthropometric and physical characteristics of participants
Table 3.5 shows the anthropometric characteristics of participants. Men were taller than
women (p<0.001). No significant difference was found across all age groups; however the
men in the youngest group were about 2cm taller than the other age groups. Weight was
highest in the oldest group (72.4 kg) in women and in men (70.9 kg), and lowest in the
youngest age group for both men and women, but there was no significant difference in
weight by sex.
BMI, HC, WC and waist-to-hip ratio tended to increase with age. BMI and HC were
significantly greater in women than men, whereas waist-to-hip ratio was significantly higher
in men than women. WC did not differ by sex.
79 | P a g e
Table 3.5 Mean (SD) anthropometric measurements by age and sex
Age groups <25 years ≥25<35y years ≥35y years p-value (GLM)
Male Female Male Female Male Female Sex Age group
n 101 101 85 156 36 75

*, a,b
Height (cm) 172.1 (7.3) 160.3 (6.3) 170.4 (7.1) 160.9 (5.7) 170.4 (6.9) 158.9 (6.4) <0.001 NS
*, a,b,c
Weight (kg) 61.0 (9.4) 60.5 (10.4) 64.9 (12.7) 67.1 (12.8) 70.9 (12.4) 72.4 (13.1) NS <0.001
2 *, a,b,c
BMI (kg/m ) 20.6 (2.5) 23.5 (3.9) 22.3 (4.0) 25.9 (4.9) 24.4 (3.8) 28.7 (5.2) <0.001 <0.001
*, a,b
HC (cm) 91.1 (6.9) 102.1 (9.1) 96.0 (8.9) 105.9 (11.4) 97.6 (8.5) 108.5 (11.9) <0.001 <0.001
*, a,b,c
WC (cm) 72.4 (6.1) 73.3 (7.7) 78.1 (10.2) 79.5 (10.3) 85.9 (9.3) 85.0 (11.5) NS <0.001
Waist-to-hip r atio 0.80 (0.1) 0.72 (0.1) 0.81 (0.1) 0.75 (0.1) 0.88 (0.1) 0.78 (0.08) <0.001 <0.001
General linear model (GLM) with age group and sex as fixed factors, Scheffe test a 18- <25 vs. 25-<35, b18-<25 vs. ≥35, c>25-<35 vs. ≥35,*P < 0.05 (for scheffe test), NS-Not significant, SD-Standard
deviation, BMI-Body Mass Index, HC-Hip-Circumference, WC-Waist-Circumference.
80 | P a g e
18.6 % and 5.9% of men and women respectively in the youngest age group, and 15.3% and
1.9% of participants in the middle age group were underweight (BMI < 18.5 kg/m2), but no
one in the oldest age group was underweight (figure 3.4). The prevalence of overweight
including obesity was 20% and 53.5% in men and women of the middle age group
respectively, and 47.3% and 70.7% in men and women in the oldest group respectively.
Figure 3.4 shows that the prevalence of obesity was much higher in women than in men in all
age groups, and increased with age in women.
100% 1.0
2.9 5.9 5.6 6.9
90% 20.6
14.1
80% 22.8
46.7
41.7
70%
32.9
Prevalence (%)
60%
77.5
50% Obese
64.7
24.0 Overweight
40% 64.4
Normal weight
30%
52.8 44.5 Underweight
20%
29.3
10% 18.6 15.3
5.9
0% 1.9
<25 ≥25≤35 >35 <25 ≥25≤35 >35
Men Women
Age group (years) & Sex
Figure 3.4 Prevalence of underweight, overweight and obesity by BMI
81 | P a g e
Furthermore as shown in figure 3.5 the prevalence of abdominal obesity [as defined by WC
(>102 cm in men & > 88cm in women)] was highest in women in the oldest age group at
42.7% compared to 21.3% and 5.1% in the middle and youngest age group respectively. In
men high WC was less than 4% in the three age groups, and there was a significant difference
by sex (p<0.001). Prevalence of abdominal obesity as defined by a high waist-to-hip ratio
(>0.9 for men & >0.85 women) was highest in the oldest group (42.9% and 14.9% for men
and women respectively). The middle age group had a prevalence of about 10% for both men
and women (p<0.001). WC increased with age in women but not men, while waist-to-hip
ratio increased with age in men but not women.
50.0 High WC (>102cm men & > 88cm women) p<0.001
High WHR (>0.9 men & > 0.85 women) p<0.001

45.0 42.7
40.0 42.9
35.0
Prevalence (%)
30.0
25.0 21.3
20.0
14.9
15.0
10.6 9.7
10.0
5.1 4.1
5.0 2.9 3.5 2.9
1.0
0.0
<25 ≥25≤35 >35 <25 ≥25≤35 >35
Men Women
Figure 3.5 Prevalence of abdominal obesity
82 | P a g e
As shown in table 3.6, mean body fat measured by bio impedance analysis was highest
in women in the oldest age group (40.6kg) followed by women in the middle age
group (36.4 kg) and lowest in men in the youngest group (15.8 kg). Body fat was
statistically different between men and women (p<0.001), and across age groups
(p<0.001). However, the interaction between sex and age group was also statistically
significant (p=0.008).
Systolic blood pressure (SBP) was highest in men (128.4mmHg) in the oldest age group,
and lowest in women in the youngest group (115.6 mmHg), and was significant
different by sex and age groups (p<0.001). SBP was not statistically different between
the youngest and middle age groups however the other age group comparisons were
significantly different (p<0.001).
Diastolic blood pressure (DBP) was highest in women in the oldest age group (80.0
mmHg) and lowest in women in the youngest age group (70.5 mmHg). There was no
significant difference in DBP between men and women, but there was a significant
difference by age group (p<0.001).
Heart rate was highest in the oldest age group in women (78.3beats/min) and lowest in
the youngest age group in men (66.1 beats/min). Heart rate was significantly different
between sexes (p<0.001) and across age groups (p<0.001). Median physical activity
level (PAL) was lowest in the women (1.31) in the youngest age group and highest in
men (1.90) in the oldest age group; PAL was significantly different by sex (p<0.001) but
not age group.
83 | P a g e
Table 3.6 Mean (SD) of physical measurements
Age groups <25 years ≥25<35 years ≥35 years p-value (GLM) †
n 93 96 71 146 32 70
Fat mass (kg) 15.8 (3.5) 31.9 (6.4) 20.9 (5.7) 36.4 (6.7) 24.2 (5.3) 40.6 (6.5) <0.001 <0.001*, a,b,c
n 101 101 85 156 36 75
SBP (mmHg) 122.5 (12.7) 115.6 (10.7) 125.4 (14.9) 119.6 (12.8) 128.4 (14.4) 127.2 (17.1) <0.001 <0.001*, b,c
DBP (mmHg) 70.7 (11.2) 70.5 (9.0) 76.7 (10.1) 74.4 (10.9) 78.1 (11.5) 80.0 (11.4) NS <0.001*, a,b,c
Heart rate 66.1 (9.2) 77.5 (10.2) 67.2 (9.8) 78.1 (10.0) 70.8 (12.6) 78.3 (10.7) 0.001 <0.001*, b,c
(beats/minutes)
Physical activity 1.60 1.31 1.45 1.35 1.90 1.49 <0.001 NS

level (median
P25, P75) (1.26, 2.23) (1.15, 1.69) (1.20, 2.20) (1.16, 1.72) (1.24, 2.27) (1.20, 1.98)
†General linear model (GLM) with age group and sex as fixed factors Scheffe’s test a18-25 vs >25-35, b 18-25 vs >35, c>25-35 vs >35,*P < 0.05,
. SD-Standard Deviation, SBP-Systolic Blood Pressure, DBP-Diastolic Blood Pressure, NS-Not significant.
84 | P a g e
As shown on figure 3.6, prevalence of hypertension (SBP > 140mmHg) was measured in
9.9% and 3.0% of men and women in the youngest group, 15.3% and 7.7% in men and
women in the middle age group and 22% and 20% for men and women respectively in the
oldest group. The prevalence of hypertension increased with age in both men and women.
More men than women had high SBP, but there was little difference between the sexes in
the prevalence of high DBP.
30 High SBP(>140mmHg)
High DBP (> 90mmHg)

25 SBP > 140mmHg or DBP > 90mmHg
25.0
24.0
22.2
20
Prevalence (%)
20.0
17.3
15
15.3 15.3
10 10.8 11.1 10.9

9.9 9.4 9
7.7
5
5.0
3.9 4.0
3.0
0
<25 ≥25≤35 >35 <25 ≥25≤35 >35
Men Women
Figure 3.6 Prevalence of hypertension
85 | P a g e
Figure 3.7 shows proportion of participants at different level of physical activity. More
than two third of all women were categorised as sedentary. Proportionally more women
in the oldest age group compared to the younger age groups were classified as having a
vigorous or extreme physical activity. Proportions of men doing sedentary activity were
40%, 63% and 55% in the oldest, middle and youngest age groups respectively. Proportion
of men engaged in extreme activity was similar in men in the middle and oldest age group
while vigorous activity was highest in men in the oldest age group.
80 Sedentary activity (PAL<1.70)
moderate activity (PAL1.70-1.99) 75

70 73
Vigorous activity (PAL 2.00-2.40) 68
60 63 Extreme activity (>2.40)
PAL Frequency (%)
50 55
40
40
30
27
20
20 19
18
15 15 16
10 13 14 14
10 9
7 7 3 8 4 7
0
<25 ≥25≤35 >35 <25 ≥25≤35 >35
Men Women
Figure 3.7 Proportion of physical activity levels of participants
86 | P a g e
3.5 Lipids, Lipoproteins and glucose measurements of participants
Table 3.7 shows lipid, lipoprotein and glucose measurements of participants by age group
and sex. Mean/median levels were generally within recommended levels except for HDL
cholesterol in women which was below the recommended level (1.3 mmol/L) across all
age groups (NCEP, 2002).
LDL cholesterol was higher in women than men, but triglycerides were higher in men than
women. There was no sex difference in levels of total or HDL cholesterol, glucose or
atherogenic index.
Total and LDL cholesterol, triglycerides, glucose and atherogenic index increased with age,
but HDL cholesterol was similar across age groups.
87 | P a g e
Table 3.7 Mean (SD) of lipids, lipoproteins and glucose (mmol/L)
<25 years ≥25<35 years ≥35 years p-value(GLM) †
n 94 93 75 142 31 67
Total Cholesterol 3.32 (0.72) 3.68 (0.73) 3.83 (0.85) 3.87 (0.81) 4.19 (1.23) 4.24 (0.97) NS 0.001*, a,b,c
HDL Cholesterol 1.07 (0.24) 1.19 (0.27) 1.09 (0.37) 1.15 (0.36) 1.18 (0.86) 1.05 (0.29) NS NS
LDL Cholesterol 2.10 (0.58) 2.37 (0.68) 2.51 (0.78) 2.55 (0.74) 2.69 (1.11) 2.96 (0.94) 0.012 <0.001*,a,b,c
Triglycerides 0.74 0.56 0.84 0.74 1.27 1.02 <0.001 <0.001*, a,b,c
Median (P25, P75) (0.51, 0.95) (0.46, 0.72) (0.62, 1.31) (0.54, 0.93) (0.79, 1.69) (0.63, 1.34)
Random Glucose 4.43 (0.65) 4.57 (0.63) 4.59 (0.87) 4.55 (0.77) 4.71 (1.06) 4.97 (0.74) NS 0.018*, b,c
Atherogenic index † 3.19 (0.64) 3.21 (0.81) 3.76 (1.15) 3.55 (1.04) 4.21 (1.52) 4.30 (1.35) NS <0.001*,a,b,c
†General linear model (GLM) with age group and sex as fixed factors Scheffe test a18-25 vs >25-35, b 18-25 vs >35, c>25-35 vs >35,*P < 0.05, †Atherogenic index= (total cholesterol/HDL
cholesterol), NS- not significant
88 | P a g e
Sixty-three percent of participants had a low HDL level (< 1.0 mmol/L men & < 1.3 mmol/L
women, NCEP, 2002), and more women than men had low HDL cholesterol (74.4% versus
44.7%, p<0.001) (figure 3.8). Proportion of participants with low HDL cholesterol increased
with age in both sexes. Around 29% of men and women in the oldest age group had a high
atherogenic index (total cholesterol/HDL cholesterol >5) and more women than men had a
high atherogenic index (11.3% versus 7.5%, p<0.001).
90
Low HDL P<0.001 for sex difference
80 P=0.001
82.1
High Atherogenic P<0.001 for sex difference
70 Index 75.7
P=0.001
67.7
60
Prevalence (%)
50
46.7 48.4
40
41.5
30
29.0 28.4
20
10
8.0 1.1 9.9
0
<25 ≥25≤35 >35 <25 ≥25≤35 >35
Men Women
Figure 3.8 Prevalence of low HDL & high atherogenic index
89 | P a g e
The proportion of participants with high a triglyceride level (≥1.7 mmol/L) was 7.4%, and was
significantly higher in men than women (12.1% versus 5.0%, p<0.001). Conversely, the
percentage of participants with high LDL cholesterol level (≥3.8 mmol/L) was significantly
higher in women than men (7.6% versus 6.5%, p<0.001). The percentage of participants with
high a random glucose level (≥6.1 mmol/L) was 4.6% and was similar for both sexes.
90 | P a g e
3.5.1 Diet
3.5.2 Under and over reporting
Various methods have been suggested to identify individuals who have under or over
reported their dietary intake on FFQ. Willet (1998) recommends that acceptable energy
intakes from FFQs can range from 800 – 4000 kcal and 500-3500kcal for men and women
respectively. Black et al (1996) suggest a plausible physical activity level [Energy intake (EI)
/Basal Metabolic Rate (BMR)] of 1.2- 2.5.However, a recent paper (Jackson et al., 2006)
investigating the relationship of diet to overweight and obesity across different populations
of African origin in Cameroon, Jamaica and the United Kingdom used energy cut-off points of
<800 and >5500kcal to indicate under and over reporting. The same cut-offs were also used
in the Botswana Food Consumption Survey (Jackson 2010). Table 3.8 shows the number (%)
of participants classed as under or over reporters using these 3 different sets of cut-offs.
Based on these cut-off points, 6.2% of participants are classified as under reporting and
58.9% as over reporting when using Black et al (1996) cut-offs, 0.2% and 55.0% would be
excluded based on Willet’s (1998) under and over reporting cut-off points respectively, and
Jackson et al (2006) cut-off for under and over reporting would exclude 0.2% and 29.8%
respectively.
91 | P a g e
Table 3.8 Number (%) of participants classed as under and over reporters using different
methods
n=551 Under reporting Over reporting Normal ranges
EI/BMR <1.14 >2.5 >1.14≤2.5

(Goldberg et al., 1991)
Black et al 1996)
N (%) 34 (6.2) 325 (58.9) 192 (34.8)
Kcal/day <500 >4000 ≥400≤4000

(Willet, 1998)
N (%) 1 (0.2) 303 (55.0) 247 (44.8)
Kcal/day) <800 >5500 ≥800≤5500

(Jackson et al 2006)
N (%) 1 (0.2) 164 (29.8) 386 (70.0)
Since FFQs are designed to rank nutrient intakes rather than obtain absolute values (Willet,
1998), and given that the main aim of the study is to investigate association of diet (based on
ranking) with CVD risk factors all participants were used in the analyses. As shown in table
3.9 over reporters did not differ significantly from those with energy intakes < 5500kcal in
age, sex, level of education, BMI or waist circumference. Furthermore overreporting was not
significantly different across BMI categories (table 3.9). However interpretation of the
median intakes of nutrients should be interpreted with caution.
92 | P a g e
Table 3.9 Variation of selected variables by energy cut-off points: Mean (SD)
Energy*
<5500kcal ≥5500kcal p-value†
n=396 n=164
1 ‡
Age (years) 27 26 NS
(22, 33) (23, 31)
Sex
Male n (%) 161 (70.6) 67 (29.4)
Female n (%) 237 (71.0) 97 (29.0) NS
Level of education
No formal education n (%) 7 (58.3) 5 (41.7)
Primary n (%) 45 (80.4) 11(19.6) NS
Secondary n (%) 256 (70.0) 110 (30.0)
Tetiary n (%) 85 (70.2) 36 (29.8)
Income
<600 n (%) 103 (72.5) 39 (27.5)
≥600<300 n (%) 212 (69.7) 92 (30.3)
≥3000<10000 n (%) 64 (68.8) 29 (31.3) NS
≥ 10000 n (%) 11 (91.7) 1 (8.3)
2 ¥
BMI (kg/m ) 24.4 (5.0) 23.9 (4.7) NS
Underweight n (%) 29 (69.0) 13 (31.0)
Normal n (%) 215 (69.6) 94 (30.4)
Overweight n (%) 88 (73.3) 32 (26.7)
¥
Obese n (%) 59 (72.0) 23 (28.0) NS
¥
Waist circumference (cm) 78.1(10.7) 78.0 (9.5) NS
*energy cut-off (Jackson et a.,l 2006),1 Median (P25, P75), 2 Mean (SD), †Person Chi-squared,¥ Independent t-test, ‡Mann-
Whitney test
3.5.3 Macronutrient intake
Table 3.10 shows the daily macronutrients intake of participants. There was no statistically
significant difference in energy intake by age group or sex. The reported median energy
intakes are rather high and should be interpreted with caution because of possible over
estimation of intake measured by FFQ.
There was no significant difference in carbohydrates intake (g/day) by age group or sex. No
significant difference was found by sex or age group for fibre intake when expressed as either
g/day or g/1000kcal/day.
93 | P a g e
Median fibre intakes appear to be within the highest recommended levels of 25-30g/day
(Shikany & White, 2000; National Research Council, 2005) in the United States and higher
than UK’s individual maximum intake of non-starch polysaccharides (24g/day) (SACN, 2011).
However, given the high daily median energy intake possibly due to over estimation of
intake, total fibre intake for all participants was lower than the US recommended intake of
10-13g/1000kcal (Shikany & White, 2000; National Research Council, 2005).
There was no significant difference in median protein intake by sex. However, a significant
difference was found by age group (p=0.023), and the post-hoc test attributed the difference
to the youngest and middle age groups.
No significant differences were found across age groups or sex for absolute intakes of total
fat, MUFA, PUFA, or SFA, and PS ratio was also not different by sex or age group (table 3.11).
The percentage of participants who reported taking alcohol ranged between 61-72% for men
and 41-57% for women. Among consumers of alcohol, there was a significant difference
between men and women with men reporting higher alcohol intake than women (p=0.002),
but not between age group. Out of 149 men who drank, 41% consumed more than 15g (1
unit) of alcohol per day compared to only 20% of women. At least 25% of alcohol consumers
in all age groups consumed more than 1 unit of alcohol per day.
94 | P a g e
Table 3.10 Daily energy, carbohydrate, fibre and protein intake: Median (P25, P75)
<25 years ≥25<35 years ≥35 years p-value

†
(GLM)
n 103 102 88 154 36 77
Energy (kcal) 3784.7 4369.8 4407.9 4410.0 3872.5 3767.5 NS NS
(2591, 6361) (3532, 6098) (3243,6664) (3207, 6122) (2934, 6972) (2774, 5236)
Carbohydr ate 530.8 638.0 576.3 650.0 503.9 528.7 NS NS
(g) (363.3, 892.2) (430.5, 852.9) (400.0, 872.9) (458.1, 854.7) (357.1, 825.9) (383.1, 849.3)
Total fibre(g) 31.9 36.1 32.4 34.5 31.7 28.3 NS NS
(20.5, 48.4) (24.8, 48.2) (22.5, 45.5) (23.5, 53.4) (19.4, 54.2) (22.0, 44.0)
Total fibre 7.47 7.71 7.43 7.88 7.87 8.24 NS NS
(g/1000kcal) (6.22, 9.89) (6.11, 9.42) (5.90, 9.44) (6.51, 9.88) (7.04, 9.10) (6.73, 10.72)
a
Protein (g) 137.3 147.4 165.9 146.9 148.9 130.2 NS 0.023*
(88.5, 221.2) (102.7, 192.5) (114.1, 277.8) (97.0, 217.8) (96.9, 259.0) (96.1, 184.7)
†
General linear model (GLM) with age group and sex as fixed factors, * p<0.05 Scheffe test a 18-25 vs >25-35
95 | P a g e
Table 3.11 Daily fat and alcohol intake (grams) of participants: Median (P25, P75)

(GLM) *
n 103 102 88 154 36 77
Total Fat 115.2 132.6 135.2 134.4 120.0 100.5 NS NS

(71.8, 173.8) (94.2, 200.4) (81.2, 188.9) (85.5, 185.4) (84.7, 197.1) (73.0, 167.2)
MUFA 37.5 44.5 47.1 44.2 40.7 32.5 NS NS
(22.9, 63.0) (30.7, 65.7) (28.1, 67.7) (28.4, 63.7) (29.5, 70.6) (21.6, 57.8)
PUFA 24.7 30.9 26.4 31.3 23.0 26.2 NS NS
(16.9, 46.9) (23.3, 49.7) (17.8, 45.5) (19.9, 45.3) (15.5, 37.9) (16.3, 43.1)
SFA 34.6 45.2 42.0 40.4 36.8 29.2 NS NS
(21.7, 50.4) (30.7, 62.7) (25.0, 63.4) (26.1, 60.4) (25.6, 62.9) (19.2, 50.9)
PS ratio 0.7 (0.5, 1.0) 0.7 (0.5, 1.0) 0.6 (0.5, 1.1) 0.8 (0.5, 1.1) 0.6 (0.5, 1.0) 0.9 (0.6, 1.3) NS NS
Consumers of 63 (61.2) 51 (50.0) 63 (71.6) 88 (57.0) 23 (63.0) 32 (41.6) <0.001 NS

alcohol n (%)
Alcohol 8.00 3.1 11.64 2.18 3.8 1.10 0.002 NS

consumers (1.44, 45.11) (0.71, 13.73) (2.00, 45.36) (.73, 11.39) (0.40, 70.50) (0.88, 5.78)
only
*
General linear model (GLM) with age group and sex as fixed factors
96 | P a g e
Table 3.12 shows the percentage contribution of the macronutrients to daily total energy
intake across age group and sex. All mean intakes (%E) were within the WHO
recommended level (WHO, 1990) across all sex and age groups, except for men aged 25
years and older and women over 35 years who had a higher percentage of energy derived
from protein compared to the WHO’s highest recommendation of 15%. There was a
significant difference in the %E contribution of CHO and protein by both age group and
sex. The percentage energy contribution from total fat, SFA and MUFA was significantly
different by age group but not sex.
97 | P a g e
Table 3.12Mean (SD) percentage contribution of macronutrients daily to total energy intake
*
Recommendation <25 years ≥25<35 years ≥35 years p- value
WHO (1990) UK Dept of Male Female Male Female Male Female Sex Age group
Health(1991)
%E
Total 55-75 50 55.6 (11.2) 55.2 (10.0) 52.9 (10.5) 57.2 (10.4) 54.5 (14.2) 59.3 (12.7) 0.001 <0.001* b,c
CHO
Protein 10-15 - 14.0 (3.8) 13.9 (3.7) 15.5 (3.9) 14.3 (3.9) 15.4 (3.4) 15.0 (3.8) 0.029 0.001* c
Total Fat 15-30 35 26.3 (8.2) 29.0 (8.0) 26.5 (8.0) 26.9 (7.1) 26.8 (7.7) 26.0 (9.2) NS <0.001* b,c
SFA <10 11 8.3 (3.1) 9.3 (3.2) 8.3 (3.0) 8.5 (2.8) 8.5 (2.8) 7.9 (3.4) NS <0.001* b,c
PUFA 3-7 6.5 6.3 (2.7) 7.0 (2.6) 6.3 (2.9) 6.7 (3.0) 6.1 (3.1) 6.6 (2.5) NS NS
MUFA - 13 9.0 (3.5) 9.8 (3.4) 9.1 (3.2) 9.0 (2.7) 9.2 (2.8) 8.8 (3.8) NS <0.001* b,c
*General linear model (GLM) with age group and sex as fixed factors * p<0.05 Scheffe test b 18-25 vs >35, c>25-35 vs >35, *P < 0.05 , CHO – Carbohydrates, MUFA-Mono Unsaturated
Fatty Acids, PUFA- Poly Unsaturated Fatty Acids, SFA-Saturated Fatty Acids,
98 | P a g e
3.5.4 Proportion of individuals commonly consuming specific foods
Figure 3.9 shows the percentages of participants who commonly consumed various foods
according to the FFQ. Food items were considered “commonly consumed” if consumed ≥ 2
times a week by at least 25% of the people (Jackson, 2010). Out of 122 food items, only 35
food items were consumed ≥ 2 times per week by at least 25% of participants. The most
commonly consumed food items were tomato (79%), sugar (67%), beef (63%), rice, white
bread (60%), brown bread (59%), mayonnaise (55%), mealie (maize) meal (54%), carrots
(54%) and sweets/candy (54%). Chicken with skin, whole milk and banana were consumed by
between 50-52% of participants. For the subsequent analysis only the 35 food items in figure
3.9 were used.
99 | P a g e
Yoghurt 25
Pumpkin 25
Peanut butter 26
Cookies 27
Chicken without skin 27
Powdered milk 31
Fruit juice 32
Beet root 33
Naatjee 33
Tea 34
Fat cakes 34
Soda/carbonated drinks 35
Pasta 35
Sorghum meal 37
Cabbage 38
Food items consumed
Margarine 38
Orange 40
Potato chips 41
Eggs 41
Apple 42
Pepper(green) 47
Coffee 50
Banana 52
Chicken with skin 52
Whole milk 52
Sweets/candy 54
Carrot 54
Mealie meal 54
Mayonnaise 55
Rice 59
Brown bread 59
White bread 60
Beef 63
Sugar 67
Tomato 79
0 10 20 30 40 50 60 70 80 90
% consuming food items ≥ 2 per week
Figure 3.9 Proportion of commonly consumed food items
100 | P a g e
When commonly consumed food were stratified by sex, 33 food items were consumed by at
least 25% of men, while 40 food items were commonly consumed by the same proportion of
women. 31/33 items consumed by men were also consumed by women. Baked beans and
Baked/mashed potato were the only two food items consumed by ≥ 25% of men and not
women. Food items commonly consumed by women but not men were cheese snacks,
breakfast cereal, cookies, pear, pumpkin, yoghurt, peanut butter, fruit drink, spinach and
chicken without skin.
Participants in the middle age group had the highest number of commonly consumed food
(40 items), followed by the youngest age group with 36 items and the oldest age group had
the least number of items (32). There were 27 food items that were consumed in all age
groups. Foods commonly consumed in the youngest age group but not in other groups were
baked beans and baked/boiled/mashed potato (middle age group), and beetroot, baked
beans, yoghurt, baked/boiled/mashed potato, powdered milk, breakfast cereal and pasta
(oldest age group).
Items consumed in the middle age group but not in other groups were peanut butter,
beetroot, baked beans, yoghurt, baked/boiled/mashed potato, powdered milk, breakfa st
cereal and pasta (oldest age group), and pumpkin, pear, fruit drink and peanut butter
(youngest age group. Herbal tea and bean leaves were the only two food items that were
consumed in the oldest age group but not in the other two age groups.
101 | P a g e
Table 3.13 shows the proportion (%) of individuals who commonly consumed cereals and
starchy foods by age group and sex. The percentage of participants’ commonly consuming
white bread was highest in the youngest age group for both men (70.2%) and women
(69.6%), and there was a significant difference across the age group but not sex. The
consumption of brown bread and fatcakes (deep-fat fried bread) were also highest in the
youngest age group but no statistical significance was reached by age group or sex. Rice,
mealie-meal (maize/corn meal) and pasta consumption were significantly different by age
group but not sex. Only the proportion of individual commonly consuming sorghum meal was
significantly different across age group and sex. The proportion of individuals’ commonly
consuming all cereals and starchy foods tended to be lowest in the oldest age group except
for sorghum meal which was highest in the oldest group.
102 | P a g e
Table 3.11 Proportion n (%) of individuals who commonly consumed cereals and starches by age group and sex
<25 year ≥25<35 year ≥35 year p-value*

Male Female Male Female Male Female Age group Sex Age group/sex
White bread 73 (70.2) 71 (69.6) 47 (53.4) 98 (62.8) 17 (45.9) 31 (40.3) <0.001 NS <0.001
Brown Bread 64 (62.1) 64 (62.7) 54 (61.4) 92 (60.0) 17 (46.0) 43 (55.8) NS NS NS
Rice 64 (62.1) 68 (66.7) 50 (56.8) 94 (60.3) 19 (51.4) 38 (49.4) 0.043 NS NS
Mealie-meal 48 (46.6) 46 (45.1) 53 (60.2) 88 (56.4) 24 (64.9) 46 (59.7) 0.009 NS NS
Sorghum 23 (22.3) 31 (30.3) 26 (29.5) 65 (41.7) 22 (59.5) 40 (51.9) <0.001 0.022 <0.001
Pasta 37 (35.9) 39 (38.2) 32 (36.4) 63 (40.4) 11 (29.7) 17 (22.1) 0.024 NS NS
Fatcakes 37 (35.9) 40 (39.2) 29 (33.0) 55 (35.3) 10 (27.0) 21 (27.3) NS NS NS
*P-value; Chi 2 for difference, NS; not significant

103 | P a g e
The proportions of individuals’ commonly consuming meat and dairy products are shown in
table 3.14. The proportion of individuals who commonly consumed whole milk was highest in
the oldest age group for both men (59.5%) and women (70.1%), and was significantly
different by age-group but not by sex. Powdered milk consumption was only significantly
different by age group, whereas, yoghurt was significantly different by age group and sex.
Beef of all the meats, was consumed by the highest percentage of individuals, and the
highest percentage of consumers was in the middle age group in men (73.9%). The
proportion of Individuals consuming beef was significantly higher in men compared to
women, but was not significantly different by age group. Men and women in the middle age
group had the highest proportion of consumers of chicken with-skin, and consumption was
significantly different by age-group but not sex.
104 | P a g e
Table 3.12 Proportion (%) of individuals who commonly consumed meat and dairy products by age group and sex

Male Female Male Female Male Female Age group Sex Age group/sex
Whole-milk 42 (40.8) 50 (49.0) 51 (58.0) 75 (48.1) 22 (59.5) 54 (70.1) 0.001 NS 0.020
Powder-milk 31 (30.1) 30 (29.4) 34 (38.6) 54 (34.6) 5 (13.5) 21 (27.3) 0.036 NS NS
Yoghurt 21 (20.4) 36 (35.3) 20 (22.7) 47 (30.1) 6 (16.2) 11 (14.3) 0.020 0.037 0.008
Beef 64 (62.1) 63 (61.8) 65 (73.9) 98 (62.8) 26 (70.3) 38 (49.4) NS 0.041 0.042
Chicken with 45 (43.7) 48 (47.1) 56 (68.2) 86 (55.1) 20 (54.1) 38 (49.4) 0.024 NS NS

skin
Chicken 28 (7.2) 25 (24.5) 31 (35.2) 37 (23.7) 8 (21.6) 25 (32.5) NS NS NS
without skin
Eggs 44 (42.7) 41 (40.2) 42 (47.7) 65 (41.7) 18 (48.6) 22 (28.6) NS NS NS
P-value; Chi2 for difference, NS; not significant

105 | P a g e
The proportion of individuals who commonly consumed fruit and vegetables varied by age
group and sex as shown in table 3.15. Tomato was the most commonly consumed food and
the proportion of consumers increased with age and was significantly different by sex. The
proportion of individuals who commonly consumed carrots increased with age in women but
not men and was significantly different by sex and age group. The proportion of individual
commonly consuming pumpkin was also significantly different by sex age group although the
increase with age was observed in men only.
The percentage of people who consumed peppers, apples, oranges and naatjee (mandarin)
was significantly higher in women than men but not across age group. Conversely, the
proportion of individuals who commonly consumed cabbage and beet root was significantly
different age group but not sex. Banana was commonly consumed by about half of all the
individuals and no significant difference was observed by age group or sex.
106 | P a g e
Table 3.13 Proportion n(%) of individuals who commonly consumed fruit and vegetables by age group and sex
Male Female Male Female Male Female Age Sex Age

group group/sex
Tomato 71 (68.9) 80 (78.4) 63 (71.6) 131 (84.0) 34 (91.9) 68 (88.3) 0.004 0.004 0.002
Carrots 44 (42.7) 51 (50.0) 38 (43.2) 101 (64.7) 15 (40.5) 54 (70.1) 0.022 <0.001 <0.001
Pepper 40 (38.8) 53 (52.0) 35 (39.8) 72 (46.2) 13 (35.1) 51 (66.2) NS 0.001 0.002
Cabbage 31 (30.1) 28 (27.5) 35 (39.8) 66 (42.3) 16 (43.2) 36 (46.8) 0.003 NS 0.037
Beetroot 26 (25.2) 26 (25.4) 29 (33.0) 67 (42.9) 14 (37.8) 22 (28.6) 0.007 NS 0.019
Pumpkin 15 (14.6) 21 (20.6) 19 (21.6) 59 (37.8) 13 (35.1) 15 (19.5) 0.002 0.031 <0.001
Banana 49 (47.6) 60 (58.8) 42 (47.7) 92 (59.0) 17 (45.9) 31 (40.3) NS NS 0.048
Apple 42 (40.8) 47 (46.1) 33 (37.5) 70 (44.9) 10 (27.0) 34 (44.2) NS 0.055 NS
Orange 31 (30.1) 47 (46.1) 30 (34.1) 64 (41.0) 16 (43.2) 38 (49.5) NS 0.011 NS
Naatjee/ 22 (21.4) 41 (40.2) 24 (27.3) 55 (35.3) 15 (40.5) 30 (40.0) NS 0.010 0.029

Naraki

107 | P a g e
Table 3.16 shows the frequency of individuals who consumed sugar sweets and condiments.
Two thirds of individuals reported to commonly consume sugar and proportion of consumers
was similar by sex and age group. Similarly, the proportions of individuals who commonly
consumed peanut butter and soda/sweetened carbonated beverage were not significantly
different by age group or sex.
The proportion of women who commonly consumed candies/sweets and margarine was
significantly higher than men’s across all age groups, and decreased with age in women.
However, consumption of candies/sweets and margarine was not different by age group for
all participants. The proportion of individuals commonly consuming cookies and mayonnaise
was significantly different by age group but not by sex. The highest percentage of individuals
who commonly consumed potato chips was in women and there was a significant difference
by age group but not sex. Percentage of participants who commonly consumed coffee was
highest the middle age group and lowest in the oldest age group and there was a significant
difference by age group but not sex; conversely, tea was highest in the oldest age group and
consumption increased from the youngest to the oldest age group.
108 | P a g e
Table 3.14 Proportion n (%) of individuals who commonly consumed sugar sweets and condiments by age group and sex
<25 years ≥25<35 years ≥35 years p-value*

Male Female Male Female Male Female Age group Sex Age
group/sex
Sugar 68 (66.0) 66 (64.7) 68 (77.3) 104 (66.7) 20 (54.1) 52 (67.5) NS NS NS
Candies/sweets 51 (49.5) 67 (65.7) 38 (43.2) 94 (60.3) 13 (35.1) 39 (50.6) NS 0.001 0.002
Cookies 29 (28.2) 32 (31.4) 20 (22.7) 56 (35.9) 4 (10.8) 12 (15.6) 0.002 NS 0.003
Mayonnaise 56 (54.4) 63 (61.8) 49 (55.7) 94 (60.3) 17 (45.9) 32 (41.6) 0.013 NS NS
Margarine 27 (26.2) 50 (49.0) 26 (29.5) 71 (45.5) 13 (35.1) 28 (36.4) NS <0.001 0.003
Peanut butter 21 (20.4) 26 (25.5) 23 (26.1) 50 (32.1) 9 (24.3) 18 (23.4) NS NS NS
Potato chips 31 (30.1) 60 (58.8) 37 (42.0) 70 (44.9) 14 (37.8) 18 (23.4) 0.008 NS <0.001
Coffee 48 (46.6) 51 (50.0) 49 (55.7) 85 (54.5) 16 (43.2) 31 (40.3) 0.048 NS NS
Tea 23 (22.3) 20 (19.6) 36 (40.9) 59 (37.8) 21 (56.8) 32 (41.6) <0.001 NS <0.001
Soda/carbonated 32 (31.1) 42 (21.2) 29 (33.0) 59 (37.8) 16 (43.2) 19 (24.7) NS NS NS

drink
Fruit juice 28 (27.2) 38 (37.3) 22 (25.0) 65 (41.7) 7 (18.9) 22 (28.6) NS 0.011 0.014

109 | P a g e
3.5.5 Intakes of commonly consumed foods
Table 3.17 shows median (P25, P75) intake of cereals and starches consumed of the
participants. The median intake for both white and brown bead was 40g/day, and intake was
significantly higher in men than in women for both white and brown bread, but there was no
significant difference by age group. The median intakes of rice, mealie meal, sorghum meal
pasta and fat cakes were not significantly different by either by age group or sex except for
fatcakes where there was a marginally significant difference by age group (p=0.049).
Table 3.18 shows median daily intakes of dairy and meat products. There was a significant
difference (highest in the youngest age group) in milk intake by age group (p=0.026).
However, there were no significant differences in median intakes of the rest of meat and milk
products by age group or sex.
110 | P a g e
Table 3.15 Median (P25, P75) intake of commonly consumed cereals and starchy foods (g/day) by sex and age group
†
All <25 years ≥25<35 years ≥35 years p-value
Male Female Male Female Male Female Age Sex
group
White bread
n 338 73 71 47 98 17 31
40 60 40 60 40 40 32 NS 0.033
(30, 80) (36, 80) (30, 60) (32, 80) (30, 70) (26, 70) (20, 40)
Brown bread
n 335 64 64 54 92 17 43
40 (30, 80) 62 (31, 80) 40 (30, 75) 72 (32, 80) 40 (20, 60) 50 (26, 80) 30 (16, 40) NSc* <0.00
1
Rice
n 334 64 68 50 94 19 38
79 (72, 144) 72 (72, 177) 72 (72, 138) 96 (70, 180) 84 (72, 144) 96 (72, 144) 72 (70, 122) NS NS
Mealie meal
n 305 48 46 53 88 24 46
400 400 340 400 400 572 400 NS NS
(280, 800) (400, 593) (280, 563) (280, 1000) (280, 738) (400, 1000) (280, 800)
Sorghum meal
n 207 23 31 26 65 22 40
500 400 500 500 500 500 500 NS NS
(250, 750) (250,600) (200, 750) (238, 813) (250, 750) (300, 750) (300, 750)
Pasta
n 199 37 39 32 63 11 17
180 180 180 204 204 204 204 NS NS
(136, 340) (136, 260) (136, 385) (136, 385) (136, 280) (172, 544) (154, 374)
Fatcakes
n 193 37 40 29 55 10 21 0.049 NS
111 | P a g e
84 (56, 140) 112 (56, 152) 63 (56, 140) 96 (56, 175) 112 (56, 140) 56 (56, 74) 70 (63, 140)
†
General linear model (GLM) with age group and sex as fixed factors, * p<0.05 Scheffe test c>25-<35 years vs ≥35 years
Table 3.16 Median (P25, P75) intake of commonly consumed dairy and meat products (g/day) by sex and age group
†
Male Female Male Female Male Female Age group Sex
Whole-milk
n 294 42 50 51 75 22 54
Median (ml/day) 250 325 250 250 250 250 250 0.026a* NS
(200, 500) (200, 500) (230, 500) (200, 500) (200, 500) (200, 425) (200, 500)
Powdered milk
n 175 31 30 34 54 5 21
10 (6, 12) 12 (5, 26) 12 (8, 17) 10 (6, 12) 8 (6, 12) 8 (5, 10) 10 (8, 12) NS NS
Yoghurt
n 142 21 36 20 47 6 11
80 (62, 175) 100 (70, 418) 70 (40, 175) 90 (70, 280) 100 (40, 175) 70 (70, 88) 70 (40, 70) NS NS
Beef
n 354 64 63 65 98 26 38
128 (96, 240) 128 (96, 240) 136 (96, 240) 128 (96, 240) 132 (96, 240) 128 (96, 240) 144 (70, 240) NS NS
Chicken with skin
n 293 45 48 56 86 20 38
75 (37, 115) 80 (47, 128) 60 (32, 114) 80 (46, 124) 66 (34, 115) 64 (46, 111) 75 (44, 121) NS NS
Chicken without skin
n 155 28 25 31 37 8 25
56 (44, 110) 58 (42, 133) 67 (44, 110) 56 (42, 122) 44 (41, 86) 111 (46, 165) 56 (38, 110) NS NS
Eggs
n 232 44 41 42 65 18 22 NS NS
42 (36, 90) 54 (36, 98) 54 (36, 95) 55 (36, 100) 40 (28, 70) 46 (36, 75) 38 (36, 75)
†General linear model (GLM) with age group and sex as fixed factors * p<0.05 Scheffe test a <25 years vs >25-<35 years
112 | P a g e
Table 3.19 shows the median daily intake of vegetables. For vegetables, tomato had the
highest median daily intake of 60g, followed by carrot at 50g, and cabbage beetroot and
pumpkin at 48g. No significant difference by age and sex were found for the vegetables
except cabbage where the youngest age group had the highest median intake (72g) and the
oldest group had the lowest intake (48g) (p=0.047). There was also a significant interaction
(Sex*Age group) for beetroot (p=0.026).
Median daily intake of fruit was highest for orange at 120g, followed by apple at 80g and
banana and naatjee both at 50g/day (table 3.20). No significant difference in intake was
found by age or sex for fruit.
113 | P a g e
Table 3.17 Median (P25, P75) intake of commonly consumed vegetables (g/day) by sex and age group
*
Male Female Male Female Male Female Age Sex
group
Tomato
n 448 71 80 63 131 34 68
60 (32, 80) 40 (30, 80) 64 (33, 80) 50 (30, 80) 60 (32, 80) 80 (38, 80) 60 (40, 80) NS NS
Carrot
n 305 44 51 38 101 15 54
50 (40, 80) 50 (28, 50) 50 (40, 100) 50 (20, 50) 50 (40, 100) 50 (40, 100) 50 (50, 65) NS NS
Pepper
n 264 40 53 35 72 13 51
30 (24, 40) 31 (16, 40) 30 (26, 45) 30 (16, 40) 30 (21, 40) 40 (20, 40) 30 (24, 40) NS NS
Cabbage
n 213 31 28 35 66 16 36
48 (48, 96) 72 (48, 120) 72 (48, 120) 72 (48, 96) 60 (48, 120) 48 (24, 57) 48 (24, 57) 0.047 NS
Beetroot
n 184 26 26 29 67 14 22
48 (24, 60) 48 (24, 60) 60 (24, 120) 60 (36, 96) 48 (24, 60) 40 (24, 61) 48 (24, 60) NS NS
Pumpkin
n 143 15 21 19 59 13 15
48 (40, 100) 72 (48, 120) 48 (40, 96) 60 (48, 96) 48 (36, 100) 48 (48, 96) 60 (48, 120) NS NS
*
114 | P a g e
Table 3.18 Median (P25, P75) intake of commonly consumed fruit (g/day) by sex and age group
*
Banana
n 291 49 60 42 92 17 31
50 (30, 80) 60 (30,135) 50 (30, 75) 60 (40, 100) 50 (30, 100) 30 (20, 55) 50 (30, 75) NS NS
Apple
n 237 42 47 33 70 10 34
80 (56, 146) 80 (32, 153) 80 (60, 150) 80 (32, 110) 80 (63, 144) 60 (32, 68) 80 (63, 150) NS NS
Orange
n 226 31 47 30 64 16 38
120 (48, 160) 120 (60, 224) 120 (64, 240) 96 (48, 120) 120 (64, 180) 80 (48, 120) 64 (48, 120) NS NS
Naatjee
n 188 22 41 24 55 15 30
50 (40, 100) 55 (40, 105) 40 (25, 100) 65 (40, 100) 80 (40, 100) 40 (40, 80) 65 (40, 65) NS NS
*
115 | P a g e
As shown in table 3.21, daily median intakes of sugar, candy/sweets and cookies were 75g,
90g and 60g respectively. No significant differences by sex or age group were found for sugar
and cookies, but women had a significantly higher intake of candy/sweets than men.
Mayonnaise, margarine, and peanut butter had daily median intakes of 8g, 7g, and 10g
respectively. A significant difference was found for mayonnaise by age group (p=0.045), but
no significant difference was found for margarine by sex or age group. The daily median
intake of peanut butter was not significantly different by age group, but was significantly
different by sex (p=0.028) with women recording higher intakes than men. Potato chips
median daily intake for all participants for potato chips was 120g, and women had
consistently higher median intake than men (p=0.011).
The median daily intake of coffee and tea was 250ml for both (table 3.22), and while there
was no significant difference in intake found for coffee by age group or sex, there was a
significant difference found for tea by sex (p=0.037). The median daily intake of soda was
330ml and women had a significantly higher intake than men. Higher intake of soda was
significantly higher in the youngest group compared to oldest group.
116 | P a g e
Table 3.19 Median (P25, P75) intake of commonly consumed sugar, sweets, condiments (g/day) by sex and age group
*
Sugar
n 380 68 66 68 104 20 52
75 (25, 100) 44 (19, 100) 75 (24, 100) 75 (30, 100) 75 (24, 100) 75 (14, 100) 50 (24, 95) NS NS
Candy/sweets
n 303 51 67 38 94 13 39
90 (48, 150) 90 (50, 180) 90 (60, 150) 48 (30, 120) 120 (60, 150) 30 (30, 75) 90 (42, 150) NS 0.026
Cookies
n 153 29 32 20 56 4 12
60 (30, 144) 60 (24, 147) 78 (45, 160) 47 (26, 80) 54 (30, 113) 35 (20, 146) 53 (32, 173) NS NS
Mayonnaise
n 311 56 63 49 94 17 32
8 (4, 16) 8 (5, 15) 10 (8, 16) 10 (6, 20) 8 (4, 16) 4 (4, 9) 8 (4, 10) 0.045 NS
Margarine
n 215 27 50 26 71 13 28
7 (4, 10) 8 (3, 10) 8 (5, 10) 7 (4, 10) 7 (4, 10) 6 (3, 9) 10 (4, 10) NS NS
Peanut butter
n 140 21 26 23 50 9 18
10 (4, 10) 8 (4, 10) 10 (7, 20) 8 (4, 10) 10 (4, 16) 8 (4, 10) 9 (8, 20) NS 0.028
Potato Chips
n 231 31 60 37 70 14 18
120 (60, 150) 120 (60, 150) 150 (60, 150) 80 (60, 150) 120 (60, 150) 60 (60, 120) 123 (60, 200) NS 0.011
†
117 | P a g e
Table 3.20 Median (P25, P75) intake of commonly consumed beverages (ml/day) by sex and age group
†
Coffee
n 281 48 51 49 85 16 31
250 250 250 250 250 250 250 NS NS
(200, 350) (250, 350) (200, 350) (140, 250) (200, 350) (200, 2500 (250, 350)
Tea
n 192 23 20 36 59 21 32
250 250 250 250 250 250 250 NS 0.037
(200, 350) (140, 350) (200, 350) (200, 350) (200, 350) (225, 315) (250, 388)
Soda/sweetened
carbonated drink
n 197 32 42 29 59 16 19
330 174 330 264 330 132 264 0.014c** 0.014
(132, 330) (132, 330) (264, 503) (132, 330) (264, 330) (132, 283) (132, 330)
Fruit juice
n 183 28 38 22 65 7 22
250 200 200 200 278 200 225 NS NS
(132, 348) (132, 346) (132, 330) (132, 349) (200, 348) (132, 330) (132, 400)
†
General linear model (GLM) with age group and sex as fixed factors ,**p<0.01 Scheffe’s test c>25-<35 versus ≥35
118 | P a g e
Figure 3.10 shows the fat/oils used during cooking. The most commonly used fat/oil was
sunflower oil (84.1%) followed by margarine, butter, Olive oil and holsum (solid fat made
from palm oil) respectively.
90
80 84.1
70
% of participants
60
50
40 48.1
30
32.4
20
10
9.7 1.4
0
Sunflower oil Margarine Butter Olive oil Holsum
Type of fat/oil used in food preparation
Figure 3.10 Fat and oils used in food preparation
119 | P a g e
3.5.6 Dietary pattern analyse
Dietary patterns identified by PCA are presented in table 3.23, table 3.24 and table 3.25
dietary patterns composed of “high vegetables” containing mainly vegetables, sweets and
beverages , “Mixed” composed of a variety of foods and “traditional” characterised by
refined starchy foods and meat were observed for all participants.
Among women two patterns were identified: “high sweets” composed mainly of sweets and
vegetables and “high fruit” containing a lot of fruit and some meat , while men had four
dietary patterns; “mixed”, “convenience” high in fruit, beverages and alcohol, “traditional”
high in meat and starchy foods and “high vegetable diet” high in vegetables and some meat.
Men uniquely showed a pattern suggestive of high alcohol, meat and starchy foods, while
women’s diet was characterised by sweets/puddings, vegetables fruit and meat.
In all participants the three patterns accounted for 16.2% of the variance, and in women the
two patterns explained 14.8% of variance, while in men 21.1% of the variance was explained
by the 4 patterns.
Dietary patterns with a high factor score of high vegetable, mixed and high fruit diet are
expected to be protective against CVD risk factors while those with high factor score of
traditional diet, high sweets and convenience diet are expected to increase risk factors for
CVD.
120 | P a g e
Table 3.21 Dietary patterns for all participants
Components High vegetable , sweets Composite diet Refined starchy

and beverage foods and meat
(prudent diet)
Variance 5.9 5.7 4.6
explained (%)
Pears Pumpkin Madombi
Cucumber Lamb Eggs
Somusa Strawberry Samp
Mageu Goat meat Rice
Sweets/candy Jelly Sour milk
Pumpkin Salad dressing Green beans
Carrot Cheese Pasta
Lettuce Soy milk Magwinya
Avocado Black eye beans Canned fish
Beetroot Jam Mealie meal
Potato chips Fresh fish Brown bread
Fruit juice Mushroom Green peas
Peppers(green/red/yellow) Sweet potato Liver
Grapes Tripe Canned corned
beef
Yoghurt Dried fruit Beef
Soda/carbonated drink Custard Chicken with skin
Cookies/biscuits Jugo beans Minced meat
(beef)
Margarine Buns/scones/muffins Chicken without
skin
Cauliflower Cream cheese Canned beans in
tomato sauce
Mayonnaise Pork Whole milk
Rum/whisky/gin Millet dintshu
Fruit drink Tswana beans
Cabbage Grape fruit
Salad dressing Madombi
Mushroom
Cake
121 | P a g e
Table 3.22 Dietary patterns for all women
Components 1 High Sweets Factor 2 High fruit (fruit and Meat) Factor
(sweets and vegetables) loading loading
Variance explained (%) 7.9 6.9
Pumpkin 0.614 Goat meat 0.668
Beetroot 0.589 Lamb 0.644
Potato chips 0.566 Canned fruit in syrup 0.631
Salad dressing 0.549 Fresh fish 0.604
Cucumber 0.530 Dried salted fish 0.602
Mushroom 0.511 Pawpaw 0.582
Sweet/candy 0.481 Tripe 0.563
Carrot 0.466 Grapefruit 0.563
Cake 0.454 Madombi (boiled bread) 0.534
Coffee 0.445 Jam 0.485
Cabbage 0.442 Custard 0.460
Mayonnaise 0.441 Sweet potato 0.451
Custard 0.438 Pork 0.449
Strawberry 0.430 Minced meat (beef) 0.437
Canned fish 0.414 Diet soda 0.430
Liver 0.414 Dried fruit 0.404
Avocado 0.411 Mushroom 0.392
Cheese 0.408 Tswana bean 0.384
Pepper(green/r ed/yellow) 0.404 Jugo bean 0.380
Cornflakes 0.397 Blackeye beans 0.375
Buns/scones/muffins 0.387 Samp (maize without bran) 0.370
Jelly 0.377 Strawberry 0.369
Beef 0.376 Magwinya (bread deep fried in 0.369
oil)
Sugar 0.351 Jelly 0.364
Makgomane (indigenous 0.336
squash)
Dried fruit 0.350 Grapes 0.320
Tomato 0.342 Melon 0.319
Seswaa (pounded meat) 0.340 seswaa 0.305
Fruit drink 0.333 Pear 0.302
Margarine 0.333 Water melon 0.301
Ice-cream 0.331
Kidney 0.325
Gizzards 0.321
Spinach 0.321
Pear 0.320
Apple 0.318
Lettuce 0.316
Chicken with skin 0.312
Savory meat 0.308
Peanut butter 0.304
Pasta 0.304
Green peas 0.301
Fresh fish 0.302
122 | P a g e
Table 3.23 Dietary pattern for all men
Components Mixed diet Factor Convenience diet (Fruit and Factor Traditional diet (Refined High Factor
loading beverage + alcohol) loading starchy foods and meat) vegetable diet loading
Variance 6.2 5.7 4.8 4.4
explained (%)
Lamb 0.922 Samusa 0.859 Madila 0.579 Cauliflower 0.633
Cheese 0.872 Mageu (traditional non alcoholic 0.824 Canned fish 0.566 Cabbage 0.587
brew)
Jelly 0.866 Banana 0.811 Maize without bran (samp) 0.542 Pumpkin 0.585
Strawberry 0.792 Pear 0.779 Green beans 0.505 Margarine 0.580
Soymilk 0.784 Apple 0.684 Green peas 0.504 Peanut butter 0.543
Salad dressing 0.737 Grapes 0.647 Eggs 0.479 Green pepper 0.517
Ditloo (jugo beans) 0.667 Wine 0.560 Mealie meal 0.470 Carrot 0.494
Traditional sour cream 0.591 Rum/whisky 0.526 Boiled bread (madombi) 0.462 Broccoli 0.489
(mayere)
Baked/boiled potato 0.531 Yoghurt 0.523 Rice 0.441 Lettuce 0.471
Black eye beans 0.496 Sweets 0.494 Minced meat 0.439 Mushrooms 0.407
Goat meat (nama ya pudi) 0.488 Diet soda 0.454 Chicken without skin 0.426 Liver 0.400
Canned corned beef 0.418
Jam 0.407 Soda/carbonated drinks 0.444 Mophane worm (phane) 0.408 Kidney 0.397
Indegenous squash 0.381
(makgomane)
Millet 0.379 Pineapple 0.423 Beef 0.402 Cucumber 0.392
Buns/scones/muffins 0.378 Avocado 0.415 Canned beans in tomato sauce 0.399 Condensed milk 0.353
Chicken without skin 0.344 Whole milk 0.397 Bean leaves 0.349 Fruit juice 0.339
Lettuce 0.353 Cookies/biscuits 0.384 Goat meat 0.367 Cornflakes 0.314
Traditional beer commercial brew 0.346 Phaphata 0.337
Magwinya 0.332
Brown bread 0.328
Tripe 0.315
Naatjee 0.309
Whole milk 0.306
123 | P a g e
Key findings
Socio-demographic characteristics
 In total 564 healthy participants were r ecruited of which 40.4% were men and 59.6% wer e
women
 The median (p25, p75) age for all participants was 27 (23, 32) years and men were on average 2
years younger than women (p=0.001)
 More than 50% of participants reported a household income of BWP600 -3000. 31.5% of
participants <25 years reported household income below the poverty datum line (BWP600)
compared to 24.1% and 21.2% in older age groups respectively.
 42.6% of participants (47.8% Women versus 34.9%,) reported a family history (mother/father) of
CVD (stroke, MI, diabetes, hypertension)
Lifestyle char acteristics

 35.4% and 5% of men and women respectively reported to be current smokers (p<0.001 )
 PAL was highest in men (1.90) in the oldest age group and lowest in women (1.31) in the youngest
age group.
Anthropometric and physical measurements
 18.6% and 15.3% of men <25years and 25-35 years respectively were underweight (BMI < 18.5
2
kg/m ). 14.1% and 41.7% of men 25-35 years and >35 years respectively were overweight (BMI ≥
2
25.0 < 30.0 kg/m ).
 32.9% and 24.0% of women 25-35 years and >35 years respectively were overweight. 20.6% and
46.7% respectively were obese.
 21.3% and 42.7% of women aged 25-35 years and >35 years respectively had a high WC (>88cm).
3.5% and 2.9% of men in the same age groups had a high WC (>102cm).
 10.6% and 42.9% of men ≥25≤35 years and >35 years respectively had a high WHR (>0.9), 9.7% and
14.9% of women in the same age groups had a high WHR (0.85). WHR was significantly different by
sex and age group.
 SBP was highest in the oldest male group (128.4 mmHg) and lowest in the youngest female group
(115.6 mmHg), and was significantly different by sex and age groups.
 DBP was highest in the oldest female group and lowest in the youngest female group, and was
significantly different between age groups but not sex.
 The prevalence of hypertension (SBP>140 &DBP >90mmHg) was highest in the men in oldest age
group (25%) and lowest women in the youngest age group (5%) (p=0.001).
Lipids, lipoproteins and glucose measurements

 All mean/median lipid, lipoprotein and random glucose levels were within the recommended
optimum levels except mean HDL cholesterol which was low in women across all age groups .
 The mean total cholesterol level was highest in the oldest women group (4.24 mmol/L)
and lowest in the youngest men group (3.32 mmol/L) and was significant different
between age groups but not sex.
 The mean LDL cholesterol level was lowest the youngest men group (2 .10 mmol/L) and
highest in the oldest women group (2.96 mmol/L). Significant differences were found
between age group and sex.
124 | P a g e
 The median triglycerides level were highest in the oldest men group (1.27 mmol/L) and
lowest in the youngest women group (0.56 mmol/L). Median triglycerides level was
significantly higher in men and increased with age.
 Mean random glucose level increased with age (p=0.018) and was highest in women in the oldest
age group (4.97 mmol/L) and lowest in men in the youngest age group (4.43 mmol/L).
 Low HDL cholesterol was highest in women in the oldest group (82.1%), followed by women in th e
middle (75.7%) and youngest age (67.7%). In men 41.5%, 46.7% and 48.4% from the lowest to the
highest age group respectively had low HDL cholesterol, and significant differences were found by
sex.
 High atherogenic index was highest in men and women in the oldest age group (men 29.0%,
women 28.4%) and lowest in the youngest age group (men 0 %, women 1.1%) and a significant
difference was found between age group but not sex.
Dietary measurements
 29.8% and 0.2% of participants reported daily energy intake >5500 kcal and <500 kcal respectively
 The median energy intake for all participants was 4182 (3038, 5990) kcal/day. Highest energy
intake was found in women in the middle age group and lowest in the women oldest age group but
no significant differences by age group or sex were found.
 The mean %E intake from CHO, protein and fat (total fat, SFA, PUFA) were found to be within WHO
recommended levels. CHO and protein %E intake were significantly different by sex and age group.
 10 top commonly consumed foods (≥ 2 times a week by at least 25% of participants) were;
tomatoes, sugar, beef, white br ead, brown bread, rice, mayonnaise, mealie meal, carrots and
sweets/candy.
 As shown in table below, several dietary patterns were found for men and women
All Men Women
-High vegetable (vegetable, -Mixed diet -Sweets and vegetables (high

sweets and beverage) sweets)
-Convenience diet (fruit,
-Mixed diet beverage + alcohol) -Meat and fruit (high fruit)
- Traditional (high refined -Traditional diet (high refined

starchy foods and meat ) starchy foods and meat
-High vegetable diet
125 | P a g e
3.6 Discussion
This study to our knowledge is the first to comprehensively investigate CVD risk factors
and their association with diet in Botswana. In this study, conducted in the capital city of
Botswana a high prevalence of CVD risk factors particularly obesity, abdominal obesity,
hypertension, dyslipidemia and physical inactivity was found. A shift in commonly
consumed foods, from traditional diets high in starchy foods (sorghum and maize meal) to
dietary patterns with sugars and meats was also evident. These main findings are briefly
discussed below.
3.6.1 Obesity
Obesity measured by BMI was much higher in women (20.6% and 46%) compared to men
at (5.6% and 5.9%) in participants aged 25-<35years and >35 years respectively.
Abdominal obesity was similarly high in women (21.3% and 42.7%) in the respective age
groups, but very few men (~3%) were abdominally obese in both age groups. These high
rates of obesity in women are consistent with results from a nationally representative
Botswana Family Health Survey IV of 2007 where more than 1 in 4 women older than 30
years were reported to be obese, and obesity was higher in women in cities/towns
(17.6%) compared to those in rural areas (13.7%) (Letamo, 2010). Similarly, the WHO
Botswana steps survey (2007) reported that 1 in 4 women aged 25-64 years were obese
compared to 1 in 20 men. Table 3.26 show a pattern of high prevalence of obesity in
women but not men in selected Southern African countries (WHO steps surveys, available
at http://www.who.int/chp/steps accessed Jan 2012).
126 | P a g e
Table 3.24 Results from WHO steps survey in selected sub-Saharan countries
Botswana (2007) Zambia (2008) Cameroon (2004) Malawi (2009)* Zimbabwe (2005)
(n=4003) (n= 1928) (n=10011) (n=5206) (n=3081)
(66% F, 33% M) (60% F, 40% M) (68% F, 32% M) (75% F, 25% M)
% Men Women Men Women Men Women Men Women Men Women
2
BMI ≥30 kg/m 5.6 24.6 5.1 18.6 7.5 21.2 2.0 7.3 3.9 19.4
High Waist-to-Hip ratio - - - - - - - - 9.5 23.4
High BP 28.8 37.0 11.0 13.2 25.6 23.1 37.2 29.2 23.2 29.0
High total cholesterol - - 12.8 17.3 - - 6.3 11.0 20.2 21.3
Low HDL cholesterol - - - - - - - - 13.7 11.5
Diabetes/elevated FBG - - 2.1 3.0 5.8 5.5 6.5 4.7 9.8 10.2
Low physical activity 26.7 41.7 35.3 32.6 - - - - 54.6 57.0
Current smoker 32.8 7.8 17.5 1.5 8.0 1.0 25.9 2.9 33.4 5.0
Alcohol consumers 30.3 8.8 43.5 17.7 89 82 30.1 4.2 83.1 66.2
*Msyamboza et al (2011), high blood pressure(SBP≥140 mmHg and/or DBP ≥90 mmHg), High WHR (Men >0.95 , Women >0.8 5), high total cholesterol(≥5
127 | P a g e
mmol/L), elevated fasting blood glucose (FBG) ((≥7 mmol/L), low HDL cholesterol (men <1.0 mmol/L, women <1.3 mmol/L), lo w ph ysical activity (<600 MET
minutes/week)
Puoane et al (2002) in the South African demographic health survey reported 31.8%
and 43.4% of black African women were obese by BMI and WC. Furthermore, other
studies in sub-Saharan Africa have also reported high levels of obesity particularly in
women (Tibazarwa et al., 2008, Njelekela et al., 2009, Njelekela et al., 2001, Sievo et
al., 2006). A study on Transition and Health during Urbanisation in South Africa (THUSA
study) found obesity to be highest in upper class urban dwellers but still 1 in 4 women
in rural traditional villages were obese (Vorster et al., 2005).
The obesity level of women in this urban setting in Botswana are similar to those
reported in some developed countries; for example Health surveys in England and
Scotland reported obesity prevalence to be 1 in 4 for men and women (Aresu et al.,
2010, Bromley et al., 2011). However obesity levels reported in this thesis are lower
than those reported in the United States of America where 1 in 3 persons are obese
(Flegal et al., 2010).
Urbanisation and rapid nutritional transition have been highlighted as a possible
reason for high obesity levels (Abubakari et al., 2007, Vorster et al., 2005). In addition,
it has been suggested that developing countries including those in sub-Saharan Africa
have experienced a rapid shift in dietary pattern resulting in increased consumption of
fat and added sugar consistent with a Western diet (Popkin, 2001, Popkin & Gordon-
Larsen 2004) that potentially predisposes people to development of CVD risk factors
(Bourne et al., 2002). High levels of obesity in sub-Saharan Africa have also been
attributed to perception of body weight (Sievo et al 2006, Puoane et al., 2002). Sievo
128 | P a g e
et al (2006) in their study investigating distribution of overweight and obesity in
relationship to socio-economic and behavioural factors reported 68% of overweight
and 37% of and obese subjects did not perceive themselves as “bigger” (overweight).
In this study more than two thirds of women and about half of men were found to be
inactive (sedentary physical activity). Younger participants in particular were found to
be proportional less active than older participants in both sexes. Unexpectedly,
overweight and obese men had higher median physical activity level (PAL) compared to
those who were neither overweight nor obese (1.8 versus 1.6) although the difference
was not significant. However, in women PAL was similar in obese, overweight and in
those who were neither overweight nor obese (table 3.27). Possible reasons for the
unexpected results could be due to (i) overweight and obese participants in this study
engage in more vigorous physical activity such as exercising after they have been
advised to reduce weight due to underlying conditions or body image issues, (ii)
physical activity could be task driven and traditionally men tend to do more physically
demanding activity than women.
Kruger et al (2003) reported in the THUSA study that women who are physically
inactive had a higher BMI compared to moderately active and most active women. In
addition, Jackson et al suggests that “energy expenditure may be a more important
determinant of weight status than diet in developing countries and that diet becomes
more important where the majority of the population has a sedentary lifestyle”
(Jackson et al., 2006).
129 | P a g e
Table 3.25 Participants characteristics by BMI classification
Women Men
¥
Neither overweight Overweight obese p-value* Neither overweight or Overweight or obese p-value*
nor obese obese
n 166 91 74 185 38
Age (years) 25 (22, 30) 27 (25, 32) 34 (27, 43) <0.001† 24 (22, 28) 33 (28, 38) <0.001‡
PAL 1.4 (1.2, 1.8) 1.3 (1.2, 1.7) 1.4 (1.2, 1.7) NS 1.6 (1.2, 2.2) 1.8 (1.4, 2.3) NS
Marital status
Single 157 (54.7) 78 (27.2) 52 (18.1) <0.001 173 (86.1) 28 (13.9) 0.001
Married 8 (21.1) 12 (31.6) 18 (47.4) 12 (54.5) 10 (45.5)
Education
Primary 16 (35.6) 10 (22.2) 19 (42.2) 0.018 18 (78.3) 5 (21.7) 0.015
education
or less
Secondary 109 (51.4) 61 (28.8) 42 (19.8) 134 (87.6) 19 (12.4)
Tertiary 39 (53.4) 21 (28.8) 13 (17.8) 29 (69.0) 13 (31.0)
Income
<600 49 (53.8) 25 (27.5) 17 (18.7) NS 45 (86.5) 7 (13.4) NS
600-3000 90 (52.9) 41 (24.1) 39 (22.9) 107 (82.9) 22 (17.1)
130 | P a g e
>3000 24 (37.5) 24 (37.5) 16 (25.0) 31 (81.6) 7 (18.4)

† ‡ ¥
*Chi-squared test, Kruskal-Wallis test, Mann-Whitney test, Overweight and obesity in men were co mbined due to small nu mbers.
3.6.2 Hypertension
Early research on blood pressure in Botswana reported normal blood pressure in men
and women (152 Bushmen aged 15-83 years). Researchers attributed these normal
blood pressures to low salt intake, lack of obesity and associated illnesses, freedom
from the stress of civilisation and high magnesium intake (Truswell et al., 1972). In the
generally healthy group of participants in this study 1 in 4 men and women >35 years
were hypertensive. These results are consistent with the WHO Steps survey in
Botswana Steps survey which reported a prevalence of hypertension of 28.8% and
37.0% in men and women aged 25-64 years (Botswana steps survey, 2007). The
hypertension rate is also similar to that found in the general public age 15-64 years in
Cameroon and Zimbabwe, but lower than in Malawi (Msyamboza et al., 2011) as
shown in Table 3.24 (WHO steps surveys, available at http://www.who.int/chp/steps
accessed Jan 2012).
Similar and higher levels of hypertension have also been found elsewhere in Sub-
Saharan Africa; in Tanzania for example hypertension was reported in 38.8% and 52.2%
of rural and urban populations respectively (Edwards et al., 2000). Duda et al (2007)
also reported elevated blood pressure in 43% of women with no history of
hypertension in Accra Ghana. Furthermore, a systematic review of population based
studies on hypertension including at least 400 participants aged 15 years or more in
Sub-Saharan Africa found hypertension level to range from 11% in rural Ghana to 47%
in females in South Africa (Addo et al., 2007). Hypertension was also found to be higher
131 | P a g e
in urban compared to rural areas, and also increased steadily with age for both men
and women (Addo et al., 2007).
Recently, a cross sectional study (Millennium Village Project) evaluating the prevalence
of hypertension in poor rural villages in east and south east Africa (Malawi, Rwanda
and Tanzania) reported an overall prevalence of 22% with higher prevalence in men
than women (24% versus 20%) (Steward de Ramirez et al., 2010).
The high prevalence of hypertension in this study are also similar to those found in
developed countries, the Scottish Health survey of 2009 for example reported that
34.6% and 30.4% of men and women aged 16 years and older were estimated to be
hypertensive (Corbett et al., 2010). Data from the National Health and Nutrition
Examination survey report prevalence of hypertension in the USA to have remained
steady from 1999-2008 at about 30% and was higher for non Hispanic blacks (~40%)
compared to the rest of the population (Yoon et al., 2010).
3.6.3 Dyslipidaemia
In this study we found varying degrees of dyslipaedemia. These results are consistent
with those found in the INTERHEART African study (Steyn et al., 2005). A higher
prevalence of low HDL cholesterol was found in women than men (74.4% versus
44.7%). These results are comparable to results found among volunteer workers of an
observational cross-sectional study at the Seventh Day Adventist hospital in Kanye
Botswana where 79.2% and 20.8% of women and men respectively were found to have
132 | P a g e
low HDL cholesterol (Garrido et al., 2009). Similar results have also been found
elsewhere in Sub-Saharan Africa, for example in Cameroon 52% of 313 workers at the
University of Buea had low HDL cholesterol (Achidi & Tangoh, 2010). Njelekela et al
(2009) in Tanzania reported that women had 3.4 times greater odds of having low HDL
(OR 3.4, 95%CI: 1.7-7.0) compared to men. The current low levels of HDL cholesterol in
Sub-Saharan Africa are in contrast to earlier studies in black Africans where favourable
HDL cholesterol levels, high physical activity and a frugal traditional diet were
suggested to be responsible for the low levels of CHD in black Africans (Walker &
Walker 1978). Low HDL has been attributed to several factors including genetic factors,
overweight and obesity, high carbohydrate diet (>60 %E) and physical inactivity (NCEP,
2002).
The mean levels of total cholesterol, LDL cholesterol and triglycerides were within
recommended levels for other populations (NCEP, 2002) and comparable to those
reported in northwest South Africa (Oosthuizen et al., 2002). However, a considerable
number of participants had high triglyceride and LDL cholesterol levels, and this
emerging pattern is consistent with finding from other countries in sub-Saharan Africa
(Njelekela et al., 2001, van de Sande et al., 2000). The blood lipid profiles of Batswana
living in this urban city indicate a relatively lower risk for cardiovascular disease.
Possibly due to low energy intake from fat (<30% of total energy) and use of
polyunsaturated sunflower oil during cooking which was reported by more than 84% of
participants. However, the high levels of obesity particularly in women may in the near
future markedly increase levels of dyslipaedemia in this population.
133 | P a g e
The percentage of participants with high random glucose is consistent with findings of
the WHO Steps report for countries in Sub-Saharan Africa as shown in Table 3.24
(WHO steps surveys, available at http://www.who.int/chp/steps accessed Jan 2012).
These findings support suggestions that diabetes prevalence may be increasing rapidly
in the region (Mbanya et al., 2010). A recent systematic analysis of epidemiological
studies and surveys of 2.7 million participants with 370 country-years reported rising
fasting plasma glucose (FPG) worldwide. In Botswana although fasting plasma glucose
levels were relatively low (<5%), prevalence of diabetes in 2008 was marginally higher
in women than men (Danaei et al., 2011). Given that obesity is a risk factor for
hyperglycaemia, most women in this study may be at an even greater risk of diabetes.
3.6.4 Diet and dietary patterns
In this study we found rather high estimated energy intake across all age and sex
groups. However, energy contributions from carbohydrate and fat (MUFA, PUFA, SFA)
in this study were within the WHO recommendations (WHO, 1990) and similar to those
found in a previous study in Botswana (Jackson, 2010) and in South Africa (Vorster et
al., 2005).
Interpretation of absolute energy and nutrient intakes therefore should be made with
caution. Nonetheless high energy intakes in excess of 3500 kcal have been observed in
similar populations; Jackson et al (2006) reported energy intake (FFQ) of over 3000
kcal/day in rural and urban Cameroon after excluding under reporters (<800 kcal/day
and over reporters (>5500 kcal/day). Energy intakes reported in this study are higher
134 | P a g e
than the estimated caloric intake of 2245-2618 kcal/day for Africa (van Wesenbeeck et
al., 2009) and the latest estimated intake for the United Kingdom (SACN, 2011). While
under reporting has been associated with women who are obese or on special diets in
developing and developed countries (Mendez et al., 2004, Johansson et al., 2001), in
this study women’s reported energy intake energy was not significantly higher than
men, suggesting that body image awareness did not play a role in estimating energy
intake, however it is possible that participants may have responded in a manner that
enhanced social status or self-presentation ( impression management) as reported
elsewhere (Bienias, 1999). Furthermore, the FFQ as an instrument is subject to
memory errors and perception of portion sizes (Willett, 1998).
Commonly consumed foods were varied in this urban setting, probably indicating an
ongoing nutrition transition in Botswana. The top 5 commonly consumed items
included tomatoes, sugar, beef, white bread and brown bread. This is a shift compared
to an earlier report on dietary patterns of the elderly in Botswana that ranked staple
foods (sorghum and maize) in the top 3 (Maruapula & Chapman-Novakofski, 2007).
Dietary patterns in this study are indicative of transitions from a diet composed largely
of staple foods (e.g. sorghum and maize) to a diet with more refined starches, snacks,
sugar and meat. This shift from a staple diet to a “western diet” has been observed in
other Sub-Saharan countries; in Ouagadougou Burkina Faso for example, “snacking”
and “modern foods” dietary patterns were typical in urban areas and the latter was
associated with a higher risk of overweight (OR 1.19, 95 CI; 1.03-1.36, p=0.018)
(Becquey et al., 2010). In South Africa (THUSA study) dietary pattern in urban setting s
135 | P a g e
were found to be changing. For example intake of maize meal was significantly lower in
urban compared to rural areas, but meats, fruit and vegetables intakes were
significantly higher in urban compared to rural areas (Vorster et al., 2005). Keding et al
(2011) has also reported an emergence of a “modern nutritional food pattern” in
women in rural Tanzania. Evidence from these findings lends more support to
increasing evidence that the Sub-Saharan region is rapidly undergoing nutrition
transition.
136 | P a g e
Chapter 4: Association between diet and CVD risk factors
Chapter 4
Association between diet and CVD
risk factors
137 | P a g e
Chapter 4 Association between diet and CVD risk factors
4 Associations between diet and CVD risk factors
This chapter describes the association between dietary variables (Total fat, SFA, PUFA,
PUFA:SFA ratio, alcohol, fibre, dietary patterns and fruit and vegetables) and CVD risk factors
in “healthy” participants. Descriptions of participants’ characteristics are detailed in chapter
3. The numbers of participants presented in tables vary due to missing data. However, where
there is variation, the numbers of participants are clearly indicated in the tables.
138 | P a g e
Spearman and Pearson correlation coefficients were produced to evaluate crude associations
between nutrients (total fat, SFA, PUFA, PUFA:SFA ratio, fibre and alcohol) and physical
measurements (BMI, WC, WHR, and body fat percentage) and biochemical indicators (total
cholesterol, triglycerides, LDL cholesterol, HDL cholesterol, random glucose and atherogenic
index (total cholesterol/HDL cholesterol). Pearson and Spearman correlations were
undertaken for variables with normal and skewed distributions respectively.
The association between nutrients [total fat, SFA, PUFA, PUFA:SFA ratio, fibre, alcohol],
dietary patterns and food groups (fruit and vegetables) with CVD risk factors were estimated
by computing general linear models and adjusting least squares mean of CVD risk factors by
categories of dietary factors. Total fat, SFA and PUFA were categorised as tertiles (thirds) of
percentage of total energy (%E), PUFA:SFA ratio and fibre (g/1000kcal) as tertiles.
Fruit and vegetable intake were adjusted for energy by computing unstandardised residuals
computed in a regression model with total energy as the independent variable and fruit/
vegetable/fruit and vegetable as the dependant variables. The unstandardised residuals were
then divided into tertiles.
Using the UK recommended safe limits of 2-3 drinks per day for women and 3-4 drinks per
day for men (Department of Health, available at
http://www.dh.gov.uk/en/Publichealth/Alcoholmisuse/DH_125368 accessed 07/10/11) for
alcohol, three categories of alcohol consumption were created: category 1 non consumers,
category 2 moderate alcohol consumers (≤3 units/24g of alcohol per day) and the last
139 | P a g e
category was for participants who consumed more than 3 units of alcohol per day. 3 units
were used as the cut-off point for both sexes since it was an upper limit for women and a
lower limit for men. 1 unit of alcohol is equivalent to ½ a pint of ordinary (3-4%) beer or 25ml
of spirits (40% alcohol by volume) or 50ml of sherry (20% alcohol by volume).
Factors scores of participants were categorised into tertiles (low medium high) to assess the
association between dietary patterns and CVD risk factors at different levels of factor score.
As indicated in chapter 3, dietary patterns with a high factor score for “high vegetable”,
“mixed” and “high fruit” are expected to be protective against CVD risk factors, and those
with high factor score of “traditional diet”, “high sweets” and “convenience” are expected to
increase CVD risk factors.
Dependant variables were BMI, waist circumference, waist-to-hip ratio, % fat, SBP, DBP total
cholesterol, HDL-cholesterol and LDL-cholesterol, triglycerides, atherogenic index (total
cholesterol/HDL-cholesterol) and random glucose. Since CVD risk factors are widely known to
increase with age and sex, model 1 was adjusted for age (years) and sex. Model 2 was
additionally adjusted for other risk factors (confounders), smoking history (pack years) and
physical activity level (continuous variable), Education (no formal, primary, secondary and
tertiary education), income (<BWP 600, ≥BWP600<BWP3000, ≥BWP3000<BWP10000,
≥BWP10000) were included in the model as fixed factors and the rest of the variables were
entered in the model as covariates.
In chapter 3 women had significantly higher LDL cholesterol and BMI than men while men
had significantly higher triglycerides than women. Both these parameters could have been
140 | P a g e
influence by diet, therefore in this chapter analyses was carried out for all participants
combined and by sex.
Tertiles for all variables were created for all individuals (“diseased” and “healthy”), therefore
in this chapter numbers of participants in each tertile are not always balanced.
4.2 Correlations between diet and CVD risk factors
Table 4.1 shows the correlation coefficients between diet and CVD risk factors. There was no
significant correlations between total fat (%E) and CVD risk factors in women but there were
significant positive associations with total cholesterol and LDL cholesterol in men (r=0.15 for
both). SFA (%E) was inversely and significantly correlated with waist-to-hip ratio in women
(r= -0.11) but not in men. PUFA (%E) was only significantly inversely correlated with body fat
in women (r=-0.13), however no significant correlations were found between PUFA (%E) and
CVD risk factors in men. PUFA:SFA ratio was significantly correlated with total cholesterol
(r=0.14) and LDL cholesterol (r=0.11) only. Total fibre intake (g/1000kcal) was only
significantly inversely correlated with random glucose in men (r=-0.16) but not women.
Alcohol intake was significantly and negatively correlated to LDL cholesterol (r=-0.12) and
atherogenic index (r=-0.16, p<0.05), but was positively correlated with HDL cholesterol
(r=0.16, p<0.01) in women. However in men no statistical significance in correlations
between alcohol and CVD risk factors was reached.
141 | P a g e
Table 4.1 Correlations between diet and CVD risk factors
Total fat (%E) ∆ SFA (%E) ∆ PUFA (%E)* PUFA:SFA ratio Fibre (g/1000kcal) Alcohol (g)*
∆
M F M F M F M F M F M F
BMI (kg/m2) 0.13 0.05 0.05 0.06 -0.02 -0.08 -0.07 0.02 -0.05 -0.06 0.04 -0.09
WC (cm) 0.04 -0.02 0.02 -0.03 -0.01 -0.04 -0.02 0.05 -0.05 -0.02 0.10 -0.09
Waist-to-hip ratio 0.01 -0.11 0.01 -0.11† -0.03 -0.06 0.03 0.07 -0.05 0.02 0.01 -0.05
Body fat (%) 0.08 -0.02 0.05 -0.01 0.02 -0.13† 0.08 0.02 -0.04 0.001 0.10 -0.08
Total cholesterol 0.15† -0.08 0.08 -0.10 0.03 -0.01 -0.03 0.14† -0.07 -0.04 0.01 -0.05
(mmo/L)
Triglycerides 0.06 -0.04 0.13 -0.05 0.08 -0.06 0.02 0.04 0.01 -0.02 0.10 -0.09
(mmol/L)
LDL-cholesterol 0.15† -0.03 0.11 -0.06 0.07 0.00 0.002 0.11† 0.03 -0.04 -0.03 -0.12†
(mmol/L)
HDL cholesterol -0.03 -0.09 -0.02 -0.10 0.004 0.07 -0.04 0.09 -0.13 -0.00 -0.07 0.16‡
(mmol/L)
Random glucose 0.08 0.06 0.09 0.05 -0.01 0.01 0.02 0.06 -0.16† -0.03 -0.05 -0.10
(mmol/L)
Page
142 |
Atherogenic index 0.13 -0.02 0.07 -0.02 0.09 -0.05 0.05 0.02 0.08 -0.01 -0.03 -0.16‡
∆
*Spearman co rrelation, Pearson correlation, †p<0.05, ‡p<0.01, ¥p<0.001.
4.3 Association between total fat intake and CVD risk factors
Table 4.2 shows the association between total fat (%E) intake with blood lipids, lipoproteins
and glucose in all participants. After adjustment for age and sex (model 1) LDL cholesterol
increased from the lowest to the highest tertile of total fat intake (p for trend =0.035) and
the association remained statistically significant in model 2 (p for trend =0.007). The
atherogenic index also increased with increasing intake of total fat in model 1 (p for trend
=0.032), however the association weakened in model 2 (p for trend =0.053). No significant
association was found between intake of total fat and other biochemical indicators.
In men, there was a steady increase in LDL cholesterol from the lowest to highest tertile of
total fat intake in model 1 (p for trend =0.089), and the association strengthened in model 2
(p for trend =0.044). Other biochemical indicators did not show any significant difference or
linear trend across tertiles of total fat intake in men (table 4.3).
In women HDL cholesterol was higher in the lowest tertile of total fat intake compared to
highest and there was a steady and significant decrease from the highest to the l owest tertile
even after full adjustment in model 2 (p for trend=0.032). Conversely, atherogenic index
increased with increasing tertile of total fat intake (p for trend =0.029), and the trend
remained strong (p=0.030) in model 2. (table 4.3).
143 | P a g e
Table 4.2 Association between lipid, lipoprotein and glucose levels mean (SE) and total fat
intake (%E) in all participants
All
T1 (low) T2 (medium) T3 (high) P trend
Fat (%E) Mean (SD) 17.46 (3.58) 25.46 (2.17) 34.98 (5.03)
n 127 178 195
Total chole sterol Unadjusted 3.77 (0.08) 3.75 (0.07) 3.83 (0.06) 0.558
(mmol/L) Model 1 3.70 (0.07) 3.74 (0.06) 3.86 (0.06) 0.102
Model 2 3.66 (0.12) 3.73 (0.11) 3.91 (0.13) 0.109
Trigylcerides Unadjusted 0.96 (0.05) 0.86 (0.05) 0.91 (0.04) 0.442

(mmol/L) Model 1 0.92 (0.05) 0.89 (0.04) 0.97 (0.04) 0.427
Model 2 0.90 (0.07) 0.92 (0.07) 1.00 (0.08) 0.380
LDL cholesterol Unadjusted 2.44 (0.07) 2.45 (0.06) 2.55 (0.06) 0.246
a
(mmol/L) Model 1 2.38 (0.07) 2.44 (0.06) 2.57 (0.06) 0.035
a
Model 2 2.33 (0.11) 2.39 (0.10) 2.67 (0.12) 0.007
HDL cholesterol Unadjusted 1.12 (0.03) 1.14 (0.03) 1.10 (0.03) 0.657
(mmol/L) Model 1 1.12 (0.03) 1.14 (0.03) 1.10 (0.03) 0.666
Model 2 1.15 (0.05) 1.18 (0.05) 1.14 (0.06) 0.547
Random glucose Unadjusted 4.59 (0.07) 4.57 (0.06) 4.65 (0.06) 0.447
(mmol/L) Model 1 4.56 (0.07) 4.55 (0.06) 4.66 (0.06) 0.262
Model 2 4.62 (0.11) 4.58 (0.10) 4.83 (0.12) 0.427
Atherogenic index Unadjusted 3.56 (0.10) 3.52 (0.08) 3.68 (0.08) 0.352
a
Model 1 3.48 (0.09) 3.54 (0.08) 3.74 (0.08) 0.032
Model 2 3.29 (0.15) 3.44 (0.14) 3.60 (0.16) 0.053
a
High v Low (P<0.05)
Model 1 [age (years), sex]
Model 2 [Age (years), sex, education , smoking history (pack years), physical a ctivity level (PA L), inco me]
144 | P a g e
Table 4.3 Association between lipid, lipoprotein and glucose levels mean (SE) with total fat intake (%E) in men and women
Men Women
T1 (low) T2 (medium) T3 (high) P trend T1 (low) T2 (medium) T3 (high) P trend
Total fat (%E) Mean (SD) 17.28 (3.99) 25.62 (2.11) 35.16 (4.78) 17.63 (3.21) 25.34 (2.21) 34.87 (5.19)
n 58 73 69 69 105 126
Total chole sterol Unadjusted 3.55 (0.12) 3.56 (0.11) 3.82 (0.11) 0.092 3.97 (0.10) 3.89 (0.08) 3.84 (0.08) 0.320
(mmol/L) Model 1 3.56 (0.11) 3.58 (0.10) 3.80 (0.10) 0.121 3.84 (0.10) 3.89 (0.08) 3.90 (0.07) 0.644
Model 2 3.55 (0.20) 3.57 (0.19) 3.79 (0.21) 0.119 3.92 (0.14) 3.97 (0.13) 3.99 (0.13) 0.598
Trigylcerides Unadjusted 1.01 (0.08) 0.94 (0.07) 1.09 (0.08) 0.458 0.92 (0.07) 0.80 (0.06) 0.81 (0.05) 0.186
(mmol/L) Model 1 1.01 (0.08) 0.95 (0.07) 1.07 (0.07) 0.616 0.82 (0.06) 0.81 (0.05) 0.86 (0.05) 0.636
Model 2 0.89 (0.14) 0.82 (0.13) 0.93 (0.14) 0.693 0.80 (0.09) 0.81 (0.08) 0.86 (0.09) 0.495
LDL cholesterol Unadjusted 2.25 (0.10) 2.28 (0.09) 2.50 (0.10) 0.069 2.61 (0.10) 2.57 (0.08) 2.57 (0.07) 0.784
(mmol/L) Model 1 2.26 (0.10) 2.29 (0.09) 2.49 (0.09) 0.089 2.48 (0.09) 2.58 (0.07) 2.63 (0.07) 0.169
a
Model 2 2.36 (0.18) 2.39 (0.17) 2.63 (0.18) 0.044 2.51 (0.14) 2.61 (0.12) 2.68 (0.12) 0.162
HDL cholesterol Unadjusted 1.06 (0.06) 1.12 (0.05) 1.10 (0.05) 0.602 1.16 (0.04) 1.16 (0.03) 1.10 (0.03) 0.156
a
(mmol/L) Model 1 1.06 (0.06) 1.12 (0.05) 1.10 (0.05) 0.648 1.19 (0.04) 1.16 (0.03) 1.09 (0.03) 0.030
a
Model 2 0.99 (0.120 1.06 (0.10) 1.00 (0.11) 0.897 1.24 (0.06) 1.20 (0.05) 1.13 (0.05) 0.032
Random glucose Unadjusted 4.58 (0.11) 4.40 (0.10) 4.63 (0.10) 0.762 4.59 (0.09) 4.69 (0.07) 4.67 (0.07) 0.477
(mmol/L) Model 1 4.59 (0.12) 4.40 (0.09) 4.62 (0.10) 0.872 4.51 (0.09) 4.69 (0.07) 4.70 (0.07) 0.093
Model 2 4.93 (0.20) 4.72 (0.19) 4.92 (0.20) 0.948 4.50 (0.13) 4.69 (0.12) 4.69 (0.12) 0.108
Atherogenic index Unadjusted 3.48 (0.14) 3.46 (0.13) 3.74 (0.13) 0.189 3.63 (0.14) 3.56 (0.11) 3.65 (0.10) 0.909
a
(mmol/L) Model 1 3.50 (0.14) 3.48 (0.12) 3.71 (0.12) 0.245 3.41 (0.09) 3.57 (0.10) 3.76 (0.09) 0.029
a
Model 2 3.21 (0.25) 3.18 (0.23) 3.47 (0.25) 0.176 3.30 (0.18) 3.45 (0.16) 3.65 (0.17) 0.030
a
High v Low (P<0.05), Model 1 [age (years) adjusted], Model 2 [Age (years), educa tion, smoking history (pack years), physical activity level (PA L), inco me]
145 | P a g e
As shown in table 4.4 WHR appeared to decrease with increasing total fat intake in the crude
model (p for trend =0.019), however the trend was weakened in model 1 and model 2. No
clear association between BMI, WC, body fat, SBP and DBP with total fat intake was found in
all participants.
In men no significant associations were found between physical measurements and total fat
intake (table 4.5). However, body fat percentage was significantly higher in the lowest tertile
compared to the middle tertile of total fat intake in model 1 and model 2 (p=0.024).
In women, BMI increased with increasing tertile of total fat intake (%E) and there was a
significant linear trend observed (p=0.039) after age adjustment which also remained strong
in model 2 (p=0.028). WC was significantly higher in the medium tertile compared with the
lowest tertile of total fat intake (p=0.033). WHR was highest in the medium tertile and lowest
in the highest tertile of total fat intake (p=0.034). No significant trends were observed for
WC, WHR and Body fat. A significant test for trend was found in the crude model for SBP (p
for trend =0.029) and DBP (p for trend =0.009), however the strength of the trend was not
significant in models 1 and 2 (Table 4.5).
146 | P a g e
Table 4.4 Association between physical measurements mean (SE) and total fat intake (%E)
in all participants
All
Total fat (%E) Mean (SD) 17.46 (3.58) 25.46 (2.17) 34.98 (5.03) Fat (%E)
n 145 197 210
2
BMI (kg/m ) Unadjusted 23.93 (0.41) 24.05 (0.35) 24.61 (0.34) 0.203
Model 1 23.52 (0.35) 23.79 (0.30) 24.20 (0.30) 0.138
Model 2 22.76 (0.57) 24.36 (0.54) 23.46 (0.63) 0.260
WC (cm) Unadjusted 78.22 (0.86) 78.31 (0.74) 77.79 (0.72) 0.701

Model 1 77.53 (0.77) 78.15 (0.67) 78.12 (0.67) 0.560
Model 2 74.75 (1.27) 77.92 (1.22) 75.63 (1.43) 0.857
a
WHR Unadjusted 0.79 (0.01) 0.78 (0.01) 0.77 (0.01) 0.019
Model 1 0.79 (0.01) 0.79 (0.01) 0.78 (0.01) 0.431
Model 2 0.77 (0.01) 0.77 (0.01) 0.77 (0.01) 0.361
Body fat % Unadjusted 29.11 (0.94) 28.67 (0.80) 30.18 (0.78) 0.382
b
Model 1 27.50 (0.52) 27.30 (0.45) 27.89 (0.45) 0.576
Model 2 28.11 (0.80) 28.86 (0.78) 27.33 (0.92) 0.828
SBP (mmHg) Unadjusted 123.29 (1.16) 121.11 (1.01) 121.69 (0.97) 0.291
Model 1 122.95 (1.13) 121.62 (0.97) 123.11 (0.96) 0.915
Model 2 121.76 (1.81) 118.27 (1.73) 121.29 (2.03) 0.297
DBP (mmHg) Unadjusted 75.78 (0.91) 73.91 (0.79) 73.87 (0.76) 0.109
Model 1 75.26 (0.88) 73.96 (0.77) 74.42 (0.76) 0.471
Model 2 74.78 (1.45) 71.99 (1.39) 72.12 (1.63) 0.515
a b
High v Low (P<0.05), Mediu m v Lo w (P<0.05)
Model 2 [Age (years), sex, education , smoking history (pack years), physical a ctivity level (PA L), inco me]
147 | P a g e
Table 4.5 Association between physical measurements mean (SE) and total fat intake (%E) in men and women
Men Women
Total fat(%E) Mean (SD) 17.28 (3.99) 25.62 (2.11) 35.16 (4.78) 17.63 (3.21) 25.34 (2.21) 34.87 (5.19)
n 65 83 75 81 114 135
2
BMI (kg/m ) Unadjusted 21.88 (0.45) 21.54 (0.40) 22.17 (0.42) 0.640 25.58 (0.56) 25.88 (0.47) 25.96 (0.44) 0.587
a
Model 1 21.91 (0.42) 21.60 (0.37) 22.08 (0.39) 0.763 24.95 (0.52) 25.89 (0.43) 26.32 (0.40) 0.039
a
Model 2 20.67 (0.76) 20.19 (0.70) 20.69 (0.77) 0.971 25.46 (0.78) 26.49 (.072) 26.94 (0.73) 0.028
WC (cm) Unadjusted 77.33 (1.19) 75.70 (1.06) 77.31 (1.11) 0.990 78.93 (1.19) 80.20 (1.01) 78.07 (0.94) 0.569
Model 1 77.33 (1.04) 75.84 (0.93) 77.16 (0.97) 0.905 77.54 (1.09) 80.22 (0.92) 78.82 (0.85) 0.361
b
Model 2 73.88 (1.90) 71.77 (1.76) 72.90 (1.93) 0.494 76.90 (1.63) 80.00 (1.52) 78.44 (1.54) 0.279
a,c
WHR Unadjusted 0.82 (0.01) 0.81 (0.01) 0.82 (0.01) 0.887 0.76 (0.01) 0.76 (0.01) 0.74 (0.01) 0.031
Model 1 0.82 (0.01) 0.81 (0.01) 0.82 (0.01) 0.808 0.75 (0.01) 0.76 (0.01) 0.74 (0.01) 0.310
c
Model 2 0.82 (0.01) 0.81 (0.01) 0.81 (0.01) 0.482 0.74 (0.01) 0.75 (0.01) 0.73 (0.01) 0.588
Body fat % Unadjusted 19.74 (0.76) 17.90 (0.66) 19.57 (0.70) 0.867 36.08 (0.85) 36.09 (0.71) 35.79 (0.66) 0.786
b
Model 1 19.82 (0.65) 18.16 (0.57) 19.21 (0.60) 0.496 34.95 (0.76) 36.10 (0.63) 36.46 (0.59) 0.118
b
Model 2 18.36 (1.16) 16.72 (1.07) 17.71 (1.18) 0.467 35.42 (1.11) 36.47 (1.02) 37.03 (1.04) 0.102
a
SBP (mmHg) Unadjusted 123.50 (1.73) 123.40 (1.54) 126.79 (1.61) 0.165 123.12 (1.53) 119.46 (1.30) 118.85 (1.20) 0.029
Model 1 123.56 (1.70) 123.54 (1.51) 126.59 (1.58) 0.193 121.93 (1.47) 119.47 (1.23) 119.65 (1.13) 0.223
b
Model 2 118.82 (2.95) 123.02 (2.68) 126.66 (3.61) 0.117 123.06 (2.29) 114.30 (2.27) 117.21 (2.17) 0.906
a
DBP (mmHg) Unadjusted 74.10 (1.40) 73.15 (1.25) 75.27 (1.30) 0.540 77.10 (1.20) 74.45 (1.02) 73.08 (0.94) 0.009
Model 1 74.16 (1.34) 73.30 (1.20) 75.05 (1.24) 0.627 76.23 (1.17) 74.46 (0.98) 73.62 (0.90) 0.081
Model 2 73.76 (2.32) 72.18 (2.11) 72.54 (2.84) 0.998 75.26 (1.89) 71.73 (1.87) 71.82 (1.80) 0.343
a b c
High v Low (P<0.05), Mediu m v Lo w (P<0.05), High v Medium (P<0.05)
Model 1 [age (years) adjusted], Model 2 [Age (yea rs), education , smoking history (pack years), physical a ctivity level (PA L), income]
148 | P a g e
4.4 Association between SFA intake and CVD risk factors
The associations between dietary SFA (%E) intake and levels of blood lipids, lipoproteins and
random glucose in all participants are shown in table 4.6. No clear direction was observed in
the association between SFA and total cholesterol. Triglycerides showed a positive significant
linear association with SFA intake in model 2 (p for trend =0.048). LDL cholesterol was
positively and significantly associated with SFA in models 1 and 2. Similarly, random glucose
was found to have a positive significant linear association with SFA in model 1 (p for trend
=0.020) and model 2 (p for trend =0.044). Conversely, HDL cholesterol was found to have a
negative significant linear association with SFA intake in model 2 (p for trend =0.021).
In men, LDL cholesterol increased significantly with tertiles of SFA intake in model 1 (p for
trend =0.038) and model 2(p for trend =0.024). Atherogenic index increased with increasing
tertiles of SFA intake, however this did not reach statistical significance (p for trend =0.081).
There was no significant association between SFA intake and total cholesterol, triglycerides,
HDL cholesterol or random glucose in men (table 4.7).
In women, HDL cholesterol decreased with increasing SFA intake (p trend=0.009) and the
strength of linear association remained strong in model 2 (p=0.012). The atherogenic index in
model 1 (p for trend =0.030) and model 2 (p for trend =0.034), and glucose level in model 1
(p for trend=0.027) and model 2 (p for trend=0.031) increased significantly with SFA intake
(table 4.7).
149 | P a g e
intake (%E) in all participants
All
SFA (%E) Mean (SD) 4.90 (1.19) 7.76 (0.72) 11.63 (1.96)
n 128 174 198
(mmol/L) Model 1 3.68 (0.07) 3.80 (0.06) 3.82 (0.06) 0.173
Model 2 3.76 (0.11) 3.88 (0.11) 3.87 (0.11) 0.239

(mmol/L) Model 1 0.86 (0.05) 0.94 (0.04) 0.97 (0.04) 0.089
a
Model 2 0.81 (0.08) 0.90 (0.07) 0.94 (0.07) 0.048
b a
(mmol/L) Model 1 2.33 (0.07) 2.51 (0.06) 2.53 (0.06) 0.022
b a
Model 2 2.39 (0.10) 2.60 (0.10) 2.60 (0.10) 0.017
(mmol/L) Model 1 1.17 (0.03) 1.09 (0.03) 1.10 (0.03) 0.084
a,c
Model 2 1.20 (0.05) 1.11 (0.05) 1.10 (0.05) 0.021
b a
(mmol/L) Model 1 4.50 (0.07) 4.53 (0.06) 4.71 (0.06) 0.020
a
Model 2 4.61 (0.11) 4.65 (0.10) 4.79 (0.10) 0.044
Atherogenic Unadjusted 3.50 (0.10) 3.67 (0.09) 3.59 (0.08) 0.455

b a
index Model 1 3.39 (0.09) 3.67 (0.08) 3.68 (0.08) 0.018
Model 2 3.76 (0.11) 3.88 (0.11) 3.87 (0.11) 0.239
a b c
High v Low (P<0.05), Mediu m v Lo w (P<0.05), High v Medium(P<0.05)
Model 2 [Age (years), sex, education , smoking history (pack years), physical a ctivity level, inco me]
150 | P a g e
Table 4.7 Association between lipid lipoprotein and glucose levels mean (SE) and SFA intake (%E) in men and women
Men Women
SFA (%E) Mean (SD) 4.89 (1.25) 7.81 (0.70) 11.51 (1.81) 4.91 (1.16) 7.72 (0.75) 11.72 (2.06)
n 59 66 75 68 104 121
(mmol/L) Model 1 3.55 (0.11) 3.61 (0.11) 3.76 (0.10) 0.146 3.82 (0.10) 3.97 (0.08) 3.85 (0.07) 0.831
Model 2 3.55 (0.20) 3.56 (0.20) 3.74 (0.20) 0.208 3.90 (0.14) 4.04 (0.13) 3.94 (0.13) 0.769
(mmol/L) Model 1 0.95 (0.08) 1.04 (0.07) 1.04 (0.07) 0.341 0.76 (0.06) 0.83 (0.05) 0.88 (0.05) 0.167
Model 2 0.79 (0.14) 0.89 (0.14) 0.92 (0.14) 0.239 0.75 (0.10) 0.83 (0.09) 0.88 (0.09) 0.127
a b
(mmol/L) Model 1 2.20 (0.10) 2.33 (0.09) 2.47 (0.09) 0.038 2.44 (0.09) 2.68 (0.07) 2.58 (0.07) 0.245
Model 2 2.29 (0.17) 2.44 (0.17) 2.59 (0.17) 0.024 2.48 (0.13) 2.70 (0.12) 2.63 (0.12) 0.237
(mmol/L) Model 1 1.14 (0.06) 1.05 (0.05) 1.10 (0.05) 0.548 1.22 (0.04) 1.13 (0.03) 1.09 (0.03) 0.009
a
Model 2 1.11 (0.10) 0.95 (0.10) 1.01 (0.10) 0.252 1.26 (0.06) 1.18 (0.05) 1.13 (0.05) 0.012
a
(mmol/L) Model 1 4.50 (0.11) 4.41 (0.10) 4.67 (0.09) 0.229 4.50 (0.09) 4.65 (0.07) 4.75 (0.07) 0.027
a
Model 2 4.81 (0.20) 4.71 (0.20) 4.95 (0.20) 0.361 4.50 (0.13) 4.67 (0.12) 4.75 (0.12) 0.031
b a
Model 1 3.40 (0.13) 3.60 (0.13) 3.65 (0.12) 0.166 3.34 (0.13) 3.71 (0.10) 3.69 (0.09) 0.030
b a
Model 2 3.07 (0.24) 3.32 (0.24) 3.38 (0.24) 0.081 3.25 (0.18) 3.59 (0.17) 3.60 (0.17) 0.034
151 | P a g e
a b
High v Low (P<0.05), Mediu m v Lo w (P<0.05),
Model 1 [age (years) adjusted], Model 2 [Age (yea rs), education , smoking history (pack years), physical a ctivity level, inco me]
Table 4.8 shows the association between SFA intake (%E) and physical measurements. In all
participants BMI had a positive linear association with intake of SFA in model 1 (p for trend
=0.007) and model 2 (p for trend =0.003). No clear linear associations were found between
SFA and other physical measurements.
In men BMI was highest in the high tertile of SFA intake, however, the linear trend observed
was weak (p=0.061) in model 1 and remained so in model 2 (p=0.064). SBP was significantly
and positively associated with SFA intake in model 2 although the association was weaker in
the preceding models. No associations were observed between WC, WHR, body fat
percentage and DBP with SFA intake.
No associations between SFA intake and physical measurements were observed in women,
except for BMI which was significantly higher in the high tertile of SFA intake compared to
the low tertile (p=0.027) in model 2. The mean BMI for women in all tertiles of SFA intake
was >25kg/m2 (overweight).
152 | P a g e
Table 4.8 Association between physical measurements mean (SE) and SFA (%E) in all
participants
All
SFA (%E) Mean (SD) 4.90 (1.19) 7.76 (0.72) 11.63 (1.96)
n 145 192 215
2
a
Model 1 23.15 (0.35) 23.91 (0.31) 24.37 (0.29) 0.007
a
Model 2 23.20 (0.55) 24.00 (0.53) 24.56 (0.54) 0.003
WC (cm) Unadjusted 77.61 (0.86) 78.71 (0.75) 77.86 (0.71) 0.828

Model 1 76.58 (0.77) 78.62 (0.69) 78.41 (0.65) 0.073
Model 2 75.56 (1.22) 77.16 (1.18) 77.17 (1.20) 0.118
WHR Unadjusted 0.78 (0.01) 0.78 (0.01) 0.77 (0.01) 0.172

Model 1 0.78 (0.01) 0.79 (0.01) 0.78 (0.01) 0.764
Model 2 0.78 (0.01) 0.78 (0.01) 0.78 (0.01) 0.965
Model 1 27.52 (0.52) 27.10 (0.46) 27.98 (0.44) 0.500
Model 2 27.38 (0.80) 26.95 (0.78) 28.11 (0.79) 0.293
Model 1 121.86 (1.11) 121.85 (0.99) 123.53 (0.94) 0.254
Model 2 119.07 (1.77) 119.42 (1.71) 120.66 (1.73) 0.285
Model 1 74.66 (0.88) 74.64 (0.79) 74.15 (0.74) 0.663
Model 2 72.56 (1.39) 72.73 (1.35) 72.06 (1.36) 0.668
a
High v Medium (P<0.05)
Model 2 [Age (years), sex, education , smoking history (pack y ears), physical a ctivity level, inco me]
153 | P a g e
Table 4.9 Association between physical measurements mean (SE) and SFA (%E) in men and women
Men Women
SFA (%E) Mean (SD) 4.89 (1.25) 7.81 (0.70) 11.51 (1.81) 4.91 (1.16) 7.72 (0.75) 11.72 (2.06)
n 65 74 84 80 118 119
2
a
Model 1 21.17 (0.42) 22.03 (0.39) 22.22 (0.38) 0.061 24.93 (0.52) 25.75 (0.42) 26.48 (0.40) 0.021
a
Model 2 19.84 (0.73) 20.72 (0.73) 20.89 (0.74) 0.064 25.58 (0.77) 26.35 (0.72) 27.09 (0.73) 0.027
b
WC (cm) Unadjusted 75.10 (1.18) 78.33 (1.11) 76.55 (1.05) 0.359 79.63 (1.20) 78.94 (0.99) 78.70 (0.95) 0.544
Model 1 74.95 (1.03) 77.90 (0.97) 77.06 (0.92) 0.129 78.02 (1.11) 79.07 (0.91) 79.52 (0.87) 0.292
Model 2 71.27 (1.84) 73.80 (1.85) 72.88 (1.87) 0.257 77.82 (1.62) 78.80 (1.54) 79.14 (1.56) 0.357
WHR Unadjusted 0.81 (0.01) 0.82 (0.01) 0.82 (0.01) 0.714 0.76 (0.01) 0.75 (0.01) 0.74 (0.01) 0.101
Model 1 0.81 (0.01) 0.82 (0.01) 0.82 (0.01) 0.430 0.75 (0.01) 0.75 (0.01) 0.75 (0.01) 0.670
Model 2 0.81 (0.01) 0.81 (0.01) 0.81 (0.01) 0.826 0.74 (0.01) 0.74 (0.01) 0.74 (0.01) 0.796
Model 1 19.27 (0.64) 18.71 (0.61) 19.01 (0.58) 0.762 35.53 (0.78) 35.34 (0.61) 36.80 (0.59) 0.201
Model 2 17.54 (1.13) 17.12 (1.14) 17.39 (1.15) 0.861 36.09 (1.10) 35.70 (1.03) 37.37 (1.04) 0.203
a
Model 1 122.50 (1.69) 123.59 (1.58) 127.08 (1.50) 0.098 120.80 (1.48) 120.03 (1.21) 119.86 (1.15) 0.878
a
Model 2 118.94 (2.99) 120.88 (2.99) 124.47 (3.04) 0.048 118.03 (2.18) 117.37 (2.07) 117.34 (2.09) 0.928
a
Model 1 73.25 (1.34) 74.36 (1.26) 74.66 (1.18) 0.719 75.95 (1.18) 74.76 (0.96) 73.50 (0.92) 0.257
Model 2 70.66 (2.38) 72.58 (2.38) 73.23 (2.41) 0.362 73.95 (1.74) 73.03 (1.65) 71.58 (1.67) 0.278
154 | P a g e
a b
High v Low (P<0.05), Mediu m v Lo w (P<0.05)
4.5 Association between PUFA intake and CVD risk factors
There was no evidence for an association between tertiles of PUFA intake and total
cholesterol, triglycerides, LDL cholesterol, HDL cholesterol and random glucose in all
participants (table 4.10). Conversely, atherogenic index was positively associated with PUFA
(p for trend 0.027).
As shown in table 4.11 significant linear associations were found between tertiles of PUFA
with lipids, lipoproteins and glucose in all participants. However in men, triglycerides levels
were marginally higher (p=0.069) in the high tertile of PUFA intake compared to the medium
tertile, and the test for linear trend was borderline significant after adjustment for age
(p=0.062) but did not improve in model 2. Mean LDL cholesterol was significantly higher in
the highest tertile of PUFA intake compared to the medium tertile and the lowest but the
linear association was not statistically significant. Atherogenic index was significantly higher
in the highest tertile of PUFA compared to the medium but trends were not statistically
significant.
As show in table 4.12 WHR was linearly association with tertiles of PUFA intake in the crude
model (p for trend =0.020) but this association disappeared in model 1 and model 2. No
association was found between tertiles of PUFA intake (%E) and the rest of the physical
measurements. After stratifying by sex (table 4.13), the only significant linear association
found was between PUFA intake and WHR in the crude model in women only (p for trend
=0.047).
155 | P a g e
Table 4.10 Association between lipid lipoprotein and glucose levels mean (SE) and PUFA
(%E) for all participants
All
PUFA (%E) Mean (SD) 3.71 (0.73) 5.93 (0.63) 9.41 (2.52)
n 137 176 187
(mmol/L) Model 1 3.74 (0.07) 3.74 (0.06) 3.84 (0.06) 0.274
Model 2 3.81 (0.12) 3.68 (0.12) 3.91 (0.11) 0.280

(mmol/L) Model 1 0.93 (0.05) 0.85 (0.04) 1.01 (0.04) 0.195
Model 2 0.87 (0.08) 0.78 (0.08) 1.02 (0.07) 0.241
(mmol/L) Model 1 2.42 (0.06) 2.42 (0.06) 2.56 (0.06) 0.099
Model 2 2.50 (0.11) 2.38 (0.11) 2.64 (0.10) 0.087
(mmol/L) Model 1 1.11 (0.03) 1.14 (0.03) 1.10 (0.03) 0.704
Model 2 1.13 (0.06) 1.14 (0.05) 1.09 (0.05) 0.308
(mmol/L) Model 1 4.58 (0.07) 4.58 (0.06) 4.62 (0.06) 0.680
Model 2 4.61 (0.11) 4.70 (0.11) 4.68 (0.10) 0.807

index Model 1 3.57 (0.09) 3.50 (0.08) 3.73 (0.08) 0.168
c
Model 2 3.46 (0.15) 3.39 (0.14) 3.74 (0.13) 0.027
b c
Medium v Lo w (P<0.05), High v Medium (P<0.05)
156 | P a g e
Table 4.11 Association between lipid lipoprotein and glucose levels mean (SE) and PUFA (%E) for men and women
Men Women
PUFA (%E) Me an(SD) 3.68 (0.79) 5.89 (0.58) 9.40 (2.56) 3.74 (0.68) 5.96 (0.66) 9.42 (2.52)
n 67 68 65 70 107 121
Total chole sterol Unadjusted 3.61 (0.11) 3.59 (0.11) 3.75(0.12) 0.363 3.97 (0.10) 3.84 (0.08) 3.87 (0.08) 0.447
(mmol/L) Model 1 3.60 (0.11) 3.60 (0.10) 3.75 (0.11) 0.313 3.87 (0.10) 3.86 (0.08) 3.92 (0.07) 0.722
Model 2 3.60 (0.21) 3.50 (0.19) 3.78 (0.20) 0.155 3.94 (0.15) 3.94 (0.13) 3.99 (0.13) 0.671
c b
c
(mmol/L) Model 1 0.97 (0.07) 0.91 (0.07) 1.16 (0.07) 0.062 0.89 (0.06) 0.77 (0.05) 0.86 (0.05) 0.686
Model 2 0.83 (0.14) 0.78 (0.13) 0.98 (0.14) 0.178 0.86 (0.10) 0.77 (0.09) 0.85 (0.08) 0.942
c
(mmol/L) Model 1 2.31 (0.09) 2.24 (0.09) 2.50 (0.09) 0.146 2.51 (0.09) 2.58 (0.07) 2.62 (0.07) 0.369
c
Model 2 2.41 (0.18) 2.28 (0.17) 2.65 (0.17) 0.064 2.53 (0.14) 2.60 (0.13) 2.64 (0.12) 0.345
HDL cholesterol Unadjusted 1.07 (0.05 1.17 (0.05) 1.05 (0.05) 0.818 1.16 (0.04) 1.11 (0.03) 1.15 (0.03) 0.872
(mmol/L) Model 1 1.07 (0.05) 1.17 (0.05) 1.05 (0.05) 0.824 1.18 (0.04) 1.10 (0.03) 1.14 (0.03) 0.482
Model 2 0.98 (0.11) 1.08 (0.10) 0.98 (0.11) 0.949 1.22 (0.06) 1.16 (0.05) 1.19 (0.05) 0.469
(mmol/L) Model 1 4.56 (0.10) 4.52 (0.10) 4.52 (0.10) 0.738 4.59 (0.09) 4.63 (0.07) 4.72 (0.07) 0.265
Model 2 4.91 (0.21) 4.79 (0.19) 4.82 (0.20) 0.568 4.56 (0.13) 4.62 (0.12) 4.70 (0.12) 0.211
c
Model 1 3.52 (0.12) 3.32 (0.12) 3.85(0.13) 0.061 3.58 (0.13) 3.67 (0.10) 3.59 (0.09) 0.927
c
Model 2 3.23 (0.25) 3.01 (0.23) 3.57 (0.24) 0.067 3.44 (0.18) 3.53 (0.17) 3.48 (0.16) 0.799
b c
Medium v Lo w (p<0.05), High v Medium (p <0.05),
157 | P a g e
Table 4.12 Association between physical measurements mean (SE) and PUFA (%E) in all
participants
All
PUFA (%E) Mean (SD) 3.71 (0.73) 5.93 (0.63) 9.41 (2.52)
n 153 193 189
2
Model 1 24.10 (0.34) 23.76 (0.30) 23.86 (0.31) 0.597
Model 2 24.22 (0.59) 23.55 (0.56) 24.42 (0.53) 0.485
WC (cm) Unadjusted 78.99 (0.83) 77.32 (0.74) 78.16 (0.74) 0.456

Model 1 78.24 (0.75) 77.40 (0.68) 78.28 (0.70) 0.967
Model 2 77.13 (1.32) 76.28 (1.25) 78.34 (1.18) 0.316
b a
WHR Unadjusted 0.79 (0.01) 0.77 (0.01) 0.77 (0.01) 0.020
Model 1 0.79 (0.01) 0.78 (0.01) 0.78 (0.01) 0.632
Model 2 0.78 (0.01) 0.77 (0.01) 0.78 (0.01) 0.711
Model 1 27.73 (0.50) 27.56 (0.46) 27.45 (0.46) 0.679
Model 2 29.63 (0.85) 27.72 (0.84) 29.37 (0.75) 0.844
Model 1 122.76 (1.09) 122.84 (0.98) 121.91 (0.99) 0.564
Model 2 121.69 (1.91) 120.64 (1.80) 119.78 (1.70) 0.254
Model 1 73.63 (0.85) 75.39 (0.77) 74.22 (0.79) 0.611
Model 2 73.50 (1.50) 73.96 (1.42) 72.68 (1.34) 0.235
a b
High v Low (P<0.05), Mediu m v Lo w (P<0.05),
158 | P a g e
Table 4.13 Association between physical measurements mean (SE) and PUFA (%E) in men and women
Men Women
PUFA (%E) Mean(SD) 3.68 (0.79) 5.89 (0.58) 9.40 (2.56) 3.74 (0.68) 5.96 (0.66) 9.42 (2.52)
n 77 76 70 78 125 127
BMI Unadjusted 22.11 (0.41) 21.42 (0.42) 22.04 (0.43) 0.904 26.56 (0.57) 25.80 (0.45) 25.42 (0.45) 0.115
2
(kg/m ) Model 1 22.03 (0.39) 21.52 (0.39) 22.01 (0.40) 0.973 26.04 (0.53) 25.90 (0.41) 25.65 (0.41) 0.565
Model 2 20.72 (0.77) 20.18 (0.72) 20.61 (0.75) 0.848 26.56 (0.79) 26.33 (0.74) 26.26 (0.71) 0.658
WC (cm) Unadjusted 77.36 (1.09) 75.41 (1.11) 77.40 (1.14) 0.978 80.58 (1.21) 78.45 (1.00) 78.59 (1.00) 0.201
Model 1 76.97 (0.96) 76.00 (0.98) 77.21 (1.00) 0.862 79.31 (1.12) 78.68 (0.88) 79.09 (0.88) 0.881
Model 2 73.09 (1.94) 71.95 (1.80) 73.22 91.88) 0.930 78.86 (1.68) 78.04 (1.57) 78.88 91.50) 0.988
a
WHR Unadjusted 0.82 (0.01) 0.81 (0.01) 0.82 (0.01) 0.889 0.77 (0.01) 0.75 (0.01) 0.75 (0.01) 0.047
Model 1 0.82 (0.01) 0.81 (0.01) 0.82 (0.01) 0.786 0.76 (0.01) 0.75 (0.01) 0.75 (0.01) 0.308
Model 2 0.81 (0.01) 0.81 (0.01) 0.82 (0.01) 0.652 0.75 (0.01) 0.74 (0.01) 0.74 90.01) 0.414
Model 1 19.07 (0.60) 18.72 (0.60) 19.19 (0.62) 0.611 36.19 (0.75) 36.24 (0.62) 35.56 (0.60) 0.518
Model 2 17.49 (1.19) 17.19 (1.10) 17.49 (1.15) 0.100 36.69 (1.12) 36.56 (1.06) 36.05 (1.01) 0.514
SBP Unadjusted 124.66 (1.60) 125.07 (1.63) 123.89 (1.66) 0.769 121.86 (1.57) 119.91 (1.25) 119.24 (1.24) 0.190
(mmHg) Model 1 124.49 (1.57) 125.29 (1.60) 123.93 (1.63) 0.835 120.70 (1.50) 120.15 (1.17) 119.81 (1.17) 0.895
Model 2 120.65 (3.17) 121.95 (2.94) 120.16 (3.08) 0.736 117.69 (2.26) 117.43 (2.10) 117.59 (2.02) 0.991
DBP Unadjusted 73.01 (1.29) 75.52 (1.30) 73.93 (1.34) 0.622 75.21 (1.24) 74.66 (0.98) 74.03 (0.98) 0.454
(mmHg) Model 1 72.81 (1.23) 75.80 (1.25) 73.86 (1.28) 0.226 74.31 (1.20) 74.83 (0.94) 74.43 (0.93) 0.926
Model 2 70.18 (2.47) 73.56 (2.29) 70.89 (2.40) 0.142 72.46 (1.80) 73.14 (1.68) 72.84 (1.61) 0.911
a
High v Low (P<0.05)
159 | P a g e
Model 1 [age (years) adjusted], Model 2 [Age (yea rs), education , smoking history (pack years), alcohol intake (g/day), physical activity level , income]
4.6 Association between PUFA:SFA ratio and CVD risk factors
Tables 4.14-4.17 show the associations between PUFA:SFA ratio and CVD risk factors. No
significant association was found between risk factors and PUFA:SFA ratio except body fat %
in women which marginally decreased with increasing PUFA:SFA ratio in age adjusted model
(p for trend=0.049).
160 | P a g e
PUFA:SFA ratio in all participants
All
PUFA:SFA r atio Mean (SD) 0.45 (0.11) 0.80 (0.11) 1.43 (0.42)
n 167 186 147
(mmol/L) Model 1 3.75 (0.06) 3.76 (0.06) 3.82 (0.07) 0.491
Model 2 3.72 (0.11) 3.87 (0.11) 3.87 (0.11) 0.283

(mmol/L) Model 1 0.92 (0.04) 0.95 (0.04) 0.92 (0.05) 0.911
Model 2 0.83 (0.08) 0.98 (0.07) 0.93 (0.07) 0.746
(mmol/L) Model 1 2.46 (0.06) 2.46 (0.06) 2.50 (0.06) 0.662
Model 2 2.43 (0.11) 2.53 (0.10) 2.55 (0.10) 0.513
(mmol/L) Model 1 1.10 (0.03) 1.12 (0.03) 1.13 (0.03) 0.493
Model 2 1.12 (0.05) 1.16 (0.05) 1.14 (0.05) 0.443
(mmol/L) Model 1 4.56 (0.06) 4.58 (0.06) 4.55 (0.07) 0.233
Model 2 4.73 (0.11) 4.50 (0.10) 4.70 (0.10) 0.971

index Model 1 3.59 (0.08) 3.66 (0.08) 3.54 (0.09) 0.644
Model 2 3.44 (0.15) 3.56 (0.14) 3.48 (0.14) 0.613
161 | P a g e
Table 4.15 Association between lipid lipoprotein and glucose levels mean (SE) and PUFA:SFA ratio in men and women
Men Women
PUFA:SFA ratio Mean(SD)
n 74 72 54 93 114 93
(mmol/L) Model 1 3.67 (0.10) 3.61 (0.10) 3.66 (0.12) 0.957 3.81 (0.08) 3.88 (0.08) 3.96 (0.08) 0.212
Model 2 3.69 (0.16) 3.29 (0.21) 3.64 (0.20) 0.446 3.82 (0.15) 4.08 (0.13) 4.04 (0.15) 0.111
(mmol/L) Model 1 0.99 (0.07) 1.04 (0.07) 1.01 (0.08) 0.806 0.85 (0.05) 0.84 (0.05) 0.81 (0.05) 0.640
Model 2 0.84 (0.11) 0.87 (0.15) 1.05 (0.14) 0.106 0.83 (0.09) 0.90 (0.08) 0.85 (0.09) 0.937
(mmol/L) Model 1 2.36 (0.09) 2.33 (0.09) 2.35 (0.10) 0.982 2.55 (0.08) 2.58 (0.07) 2.62 (0.08) 0.489
Model 2 2.44 (0.15) 2.22 (0.19) 2.33 (0.18) 0.733 2.52 (0.14) 2.66 (0.12) 2.70 (0.14) 0.333
(mmol/L) Model 1 1.09 (0.05) 1.11 (0.05) 1.08 (0.06) 0.900 1.10 (0.03) 1.13 (0.03) 1.18 (0.03) 0.099
Model 2 1.09 (0.08) 0.98 (0.11) 1.08 (0.10) 0.820 1.14 (0.06) 1.24 (0.05) 1.18 (0.06) 0.086
(mmol/L) Model 1 4.65 (0.09) 4.48 (0.10) 4.44 (0.11) 0.133 4.64 (0.08) 4.68 (0.07) 4.65 (0.08) 0.961
Model 2 4.71 (0.16) 4.23 (0.21) 4.78 (0.19) 0.427 4.62 (0.13) 4.61 (0.12) 4.72 (0.13) 0.916
Model 1 3.48 (0.12) 3.68 (0.12) 3.52 (0.14) 0.853 3.68 (0.11) 3.63 (0.10) 3.53 (0.11) 0.311
Model 2 3.35 (0.20) 3.43 (0.26) 3.36 (0.24) 0.772 3.57 (0.19) 3.51 (0.16) 3.53 (0.19) 0.318
162 | P a g e
Model 1 [age (years) adjusted], Model 2 [Age (years), education, smoking history (pack years), physical a ctivity level, inco me
Table 4.16 Association between physical measurements mean (SE) and PUFA:SFA ratio in
all participants
All
PUFA:SFA r atio Mean (SD)
n 190 203 160
2
Model 1 24.33 (0.31) 23.61 (0.30) 23.77 (0.34) 0.207
Model 2 24.58 (0.57) 23.27 (0.53) 24.17 (0.52) 0.305
WC (cm) Unadjusted 78.15 (0.76) 77.53 (0.73) 78.71 (0.82) 0.608

Model 1 78.45 (0.68) 77.47 (0.67) 78.03 (076) 0.683
Model 2 77.41 (1.27) 76.05 (1.18) 77.87 (1.16) 0.949
WHR Unadjusted 0.78 (0.01) 0.77 (0.01) 0.78 (0.01) 0.931

Model 1 0.78 (0.01) 0.78 (0.01) 0.78 (0.01) 0.995
Model 2 0.77 (0.01) 0.77 (0.01) 0.78 (0.01) 0.140
Model 1 28.07 (0.46) 27.22 (0.45) 27.50 (0.50) 0.404
Model 2 28.71 (0.84) 27.10 (0.78) 29.21 (0.77) 0.401
Model 1 123.02 (0.98) 122.33 (0.96) 121.66 (1.09) 0.356
Model 2 118.73 (1.83) 121.49 (1.70) 120.42 (1.66) 0.723
Model 1 73.92 (0.78) 74.57 (0.76) 74.74 (0.87) 0.483
Model 2 71.97 (1.44) 74.41 (1.34) 74.32 (1.31) 0.577
163 | P a g e
Table 4.17 Association between physical measurements mean (SE) and PUFA:SFA ratio in men and women
Men Women
PUFA:SFA Mean(SD)
ratio
n 85 80 58 105 123 102
BMI Unadjusted 22.01 (0.40) 21.73 (0.41) 21.80 (0.48) 0.735 26.33 (0.49) 25.35 (0.45) 25.91 (0.50) 0.551
2
(kg/m ) Model 1 22.01 (0.37) 21.75 (0.38) 21.75 (0.45) 0.649 26.59 (0.45) 25.37 (0.42) 25.64 (0.46) 0.143
Model 2 21.48 (0.70) 21.16 (0.65) 21.47 (0.61) 0.980 26.90 (0.84) 25.37 (0.72) 25.94 (0.78) 0.260
WC (cm) Unadjusted 76.93 (1.05) 76.32 (1.08) 76.97 (1.25) 0.981 79.13 (1.06) 78.32 (0.98) 79.73 (1.07) 0.690
Model 1 77.09 (0.92) 76.36 (0.95) 76.69 (1.09) 0.780 79.57 (0.97) 78.35 (0.89) 79.15 (0.97) 0.763
Model 2 74.00 (1.79) 74.01 (1.65) 76.24 (1.54) 0.623 79.83 (1.77) 77.64 (1.53) 79.05 (1.65) 0.761
WHR Unadjusted 0.81 (0.01) 0.82 (0.01) 0.82 (0.01) 0.843 0.75 (0.01) 0.75 (0.01) 0.76 (0.01) 0.611
Model 1 0.81 (0.01) 0.82 (0.01) 0.81 (0.01) 0.971 0.75 (0.01) 0.75 (0.01) 0.75 (0.01) 0.947
Model 2 0.79 (0.01) 0.80 (0.01) 0.82 (0.01) 0.191 0.75 (0.01) 0.74 (0.10) 0.76 (0.01) 0.339
b a
Model 1 18.76 (0.58) 18.86 (0.58) 19.46 (0.67) 0.433 37.21 (0.66) 35.41 (0.60) 35.36 (0.67) 0.049
Model 2 17.99 (1.08) 18.50 (0.95) 19.86 (0.86) 0.570 36.80 (0.18) 36.19 (1.02) 35.32 (1.16) 0.180
SBP Unadjusted 126.76 (1.52) 123.41 (1.57) 123.03 (1.81) 0.116 118.52 (1.36) 120.93 (1.26) 120.77 (1.38) 0.246
(mmHg) Model 1 126.75 (1.49) 123.49 (1.54) 122.93 (1.78) 0.101 119.19 (1.29) 120.95 (1.18) 120.15 (1.29) 0.598
Model 2 124.21 (2.91) 123.07 (2.69) 121.58 (2.51) 0.246 114.76 (2.35) 118.96 (2.02) 119.78 (2.18) 0.129
DBP Unadjusted 74.48 (1.23) 73.41 (1.27) 74.65 (1.47) 0.929 72.81 (1.07) 75.47 (0.99) 75.22 (1.08) 0.112
(mmHg) Model 1 74.49 (1.18) 73.50 (1.41) 74.53 (1.41) 0.983 73.20 (1.03) 75.49 (0.94) 74.80 (1.03) 0.274
Model 2 73.35 (2.25) 74.53 (2.08) 74.06 (1.94) 0.514 70.85 (1.92) 73.82 (1.65) 74.51 (1.78) 0.199
164 | P a g e
a b
High v Low (p<0.05), Mediu m v Lo w (p<0.05) Model 1 [age (years) adjusted], Model 2 [Age (yea rs), education, smoking history (pack years), physical activity level,
income]
4.7 Association between dietary fibre intake and CVD risk factors
Table 4.18 shows the association between dietary fibre intake (g/1000kcal) and blood lipids,
lipoproteins and glucose levels. Total cholesterol had an inverse linear association with
tertiles of fibre intake (p for trend =0.041) as did random glucose (p for trend =0.007) in
model 1 only. No significant associations were found with fibre intake for the other lipids and
lipoproteins.
In men mean glucose levels increased linearly across tertiles of dietary fibre intake in model 1
(p=0.025) and model 2 (p=0.046). An unexpected inverse linear trend of HDL cholesterol with
tertiles of fibre intake was observed however the strength of the association was weak (table
4.19).
In women no clear association between dietary fibre intake and lipids, lipoproteins or
glucose was found, however there was suggestion of a decrease in total cholesterol with
increasing fibre intake (p for trend=0.085).
As shown in table 4.20 and table 4.21, no significant relationship between tertiles of dietary
fibre intake and physical measurements was found.
165 | P a g e
Table 4.18 Association between lipid lipoprotein and glucose levels mean (SE) and dietary
fibre (g/1000kcal) in all participants
All
Fibre g/1000g Mean (SD) 5.71 (1.11) 8.10 (0.67) 11.83 (2.13)
n 190 169 141
a
(mmol/L) Model 1 3.87 (0.06) 3.74 (0.07) 3.68 (0.07) 0.041
Model 2 3.88 (0.11) 3.70 (0.12) 3.66 (012) 0.392

(mmol/L) Model 1 0.96 (0.04) 0.92 (0.04) 0.90 (0.05) 0.358
Model 2 0.98 (0.07) 0.96 (0.08) 0.83 (0.07) 0.507
(mmol/L) Model 1 2.51 (0.06) 2.44 (0.06) 2.45 (0.07) 0.481
Model 2 2.58 (0.10) 2.45 (0.11) 2.38 (0.11) 0.372
(mmol/L) Model 1 1.15 (0.03) 1.11 (0.03) 1.07 (0.03) 0.088
Model 2 1.11 (0.05) 1.12 (0.06) 1.10 (0.06) 0.896
a
(mmol/L) Model 1 4.70 (0.06) 4.60 (0.06) 4.46 (0.07) 0.007
Model 2 4.62 (0.10) 4.75 (0.12) 4.48 (0.11) 0.258

index Model 1 3.58 (0.08) 3.59 (0.08) 3.63 (0.09) 0.702
Model 2 3.66 (0.14) 3.41 (0.16) 3.39 (0.15) 0.314
a
High v Low(P<0.05)
Model 2 [Age (years), sex, education , smoking history (pack years), , physical activity level, inco me]
166 | P a g e
Table 4.19 Association between lipid lipoprotein and glucose levels mean (SE) and dietary fibre (g/1000kcal) in men and women
Men Women
Fibre g/1000kcal mean (SD) 5.65 (1.16) 8.00 (0.65) 11.54 (1.84) 5.76 (1.07) 8.17 (0.68) 12.01 (2.28)
n 79 66 55 109 103 86
Total cholesterol Unadjusted 3.69 (0.10) 3.69 (0.11) 3.54 (0.13) 0.353 3.91 (0.08) 3.85 (0.08) 3.88 (0.09) 0.769
(mmol/L) Model 1 3.74 (0.10) 3.62 (0.11) 3.53 (0.12) 0.166 3.99 (0.08) 3.83 (0.08) 3.82 (0.09) 0.144
Model 2 3.66 (0.19) 3.61 (0.21) 3.54 (0.20) 0.461 4.10 (0.14) 3.93 (0.13) 3.90 (0.13) 0.085
(mmol/L) Model 1 1.04 (0.07) 0.95 (0.07) 1.04 (0.08) 0.947 0.87 (0.05) 0.87 (0.05) 0.75 (0.06) 0.122
Model 2 0.85 (0.13) 0.75 (0.15) 0.93 (0.14) 0.480 0.85 (0.09) 0.87 (0.09) 0.75 (0.09) 0.232
(mmol/L) Model 1 2.35 (0.09) 2.30 (0.09) 2.38 (0.10) 0.829 2.66 (0.07) 2.56 (0.07) 2.51 (0.08) 0.162
Model 2 2.39 (0.17) 2.43 (0.19) 2.50 (0.18) 0.404 2.71 (0.13) 2.60 (0.12) 2.53 (0.13) 0.103
(mmol/L) Model 1 1.14 (0.05) 1.13 (0.05) 1.00 (0.06) 0.071 1.16 (0.03) 1.09 (0.03) 1.16 (0.04) 0.995
Model 2 1.08 (0.10) 1.06 (0.11) 0.94 (0.11) 0.069 1.21 (0.06) 1.14 (0.05) 1.21 (0.05) 0.887
a
a
(mmol/L) Model 1 4.68 (0.09) 4.50 (0.10) 4.36 (0.11) 0.025 4.71 (0.07) 4.68 (0.07) 4.56 (0.08) 0.162
a
Model 2 4.96 (0.19) 4.80 (0.21) 4.65 (0.20) 0.046 4.69 (0.13) 4.71 (0.12) 4.54 (0.12) 0.187
Model 1 3.52 (0.12) 3.49 (0.13) 3.70 (0.14) 0.346 3.62 (0.10) 3.69 (0.10) 3.53 (0.11) 0.553
Model 2 3.18 (0.23) 3.18 (0.26) 3.38 (0.25) 0.289 3.48 (0.18) 3.58 (0.17) 3.38 (0.17) 0.505
a
High v Low (P<0.05, Model 1 [age (years) adjusted], Model 2 [Age (years), educa tion, smoking history (pack years), physical activity level, income]
167 | P a g e
(g)/1000kcal in all participants
All
Fibre g/1000g Mean (SD) 5.71 (1.11) 8.10 (0.67) 11.83 (2.13)
n 211 189 153
2
Model 1 24.14 (0.29) 23.88 (0.31) 23.53 (0.34) 0.177
Model 2 25.11 (0.54) 23.40 (0.62) 23.29 (0.60) 0.159
WC (cm) Unadjusted 77.85 (0.72) 78.46 (0.76) 77.97 (0.84) 0.914

Model 1 78.57 (0.65) 77.94 (0.69) 77.16 (0.77) 0.163
Model 2 78.51 (1.12) 75.96 (1.40) 75.40 (1.36) 0.173
WHR Unadjusted 0.78 (0.01) 0.78 (0.01) 0.77 (0.01) 0.667

Model 1 0.79 (0.01) 0.79 (0.01) 0.78 (0.01) 0.145
Model 2 0.77 (0.01) 0.77 (0.01) 0.76 (0.01) 0.211
Model 1 27.84 (0.45) 27.64 (0.46) 27.10 (0.51) 0.279
b a
Model 2 30.21 (0.79) 27.78 (0.89) 27.81 (0.86) 0.261
Model 1 121.60 (0.94) 123.24 (1.00) 122.70 (1.12) 0.451
Model 2 119.58 (1.74) 120.31 (2.01) 120.47 (1.96) 0.731
Model 1 73.95 (0.74) 75.10 (0.79) 74.29 (0.88) 0.773
Model 2 72.59 (1.39) 73.53 (1.60) 72.97 (1.55) 0.595
Model 2 [Age (years), sex, education , smoking history (pack years), , physical activity level, inco me]
168 | P a g e
Table 4.21 Association between physical measurements mean (SE) and dietary fibre (g)/1000kcal in men and women
Men Women
Fibre mean (SD) 5.65 (1.16) 8.00 (0.65) 11.54 (1.84) 5.76 (1.07) 8.17 (0.68) 12.01 (2.28)
n 89 75 59 122 114 94
BMI Unadjusted 21.87 (0.39) 21.88 (0.42) 21.79 (0.47) 0.900 25.72 (0.46) 26.20 (0.47) 25.55 (0.52) 0.807
2
(kg/m ) Model 1 22.08 (0.36) 21.62 (0.39) 21.80 (0.44) 0.626 26.12 (0.42) 26.13 (0.43) 25.12 (0.48) 0.122
Model 2 20.51 (0.72) 20.24 (0.78) 20.42 (0.74) 0.886 26.71 (0.76) 26.64 (0.72) 25.72 (0.75) 0.137
WC (cm) Unadjusted 76.81 (1.02) 77.19 (1.12) 76.00 (1.25) 0.616 78.62 (0.99) 79.28 (1.01) 79.20 (1.11) 0.698
Model 1 77.49 (0.89) 76.44 (0.98) 75.91 (1.09) 0.266 79.43 (0.90) 79.16 (0.92) 78.20 (1.02) 0.370
Model 2 73.33 (1.81) 72.09 (1.97) 71.95 (1.86) 0.552 79.00 (1.62) 78.7 (1.52) 78.16 (1.59) 0.550
WHR Unadjusted 0.82 (0.01) 0.82 (0.01) 0.81 (0.01) 0.296 0.75 (0.01) 0.75 (0.01) 0.75 (0.01) 0.527
Model 1 0.82 (0.01) 0.82 (0.01) 0.81 (0.01) 0.111 0.75 (0.01) 0.75 (0.01) 0.75 (0.01) 0.640
Model 2 0.82 (0.01) 0.81 (0.01) 0.80 (0.01) 0.250 0.74 (0.01) 0.74 (0.01) 0.74 (0.01) 0.798
Model 1 19.70 (0.57) 18.44 (0.60) 18.69 (0.66) 0.616 35.79 (0.62) 36.70 (0.63) 35.31 (0.69) 0.613
Model 2 11.19 (1.08) 10.36 (1.18) 10.71 (1.14) 0.612 36.34 (1.08) 37.05 (1.07) 35.51 (1.07) 0.390
SBP Unadjusted 123.00 (1.48) 126.17 (1.61) 124.98 (1.83) 0.393 118.64 (1.26) 121.00 (1.31) 120.95 (1.43) 0.227
(mmHg) Model 1 123.40 (1.46) 125.67 (1.59) 124.97 (1.80) 0.560 119.57 (1.20) 120.81 (1.23) 120.09 (1.36) 0.767
Model 2 120.78 (2.96) 122.02 (3.22) 121.11 (3.05) 0.852 116.62 (2.16) 118.36 (2.04) 117.26 (2.14) 0.615
DBP Unadjusted 73.03 (1.20) 74.79 (1.30) 75.02 (1.47) 0.296 73.56 (0.99) 76.00 (1.03) 74.07 (1.12) 0.730
(mmHg) Model 1 73.55 (1.15) 74.18 (1.26) 75.02 (1.41) 0.719 74.21 (0.95) 75.87 (0.98) 73.39 (1.08) 0.210
Model 2 71.56 (2.33) 71.82 (2.54) 72.63 (2.40) 0.846 72.31 (1.72) 74.26 (1.62) 71.43 (1.70) 0.142
169 | P a g e
4.8 Association between alcohol intake and CVD risk factors
Table 4.22 shows the relationship between blood lipids, lipoproteins and random glucose
and alcohol intake for all participants. LDL cholesterol was inversely linearly associated with
alcohol intake (p for trend =0.049) in the crude model but the association disappeared in
subsequent models. HDL cholesterol was lowest in non consumers compared to moderate
and high alcohol consumer in the crude model (p for trend =0.036) and model 1 (p for trend
=0.004), but the linear association was no longer significant in model 2. Atherogenic index
was also highest in non alcohol consumers and lowest in high alcohol consumers and the test
for trend was significant in the crude model (p for trend =0.032) but not in models 1 or 2.
In men atherogenic index was higher in participants with a moderate alcohol intake and
lower in those with a high alcohol intake (p=0.037), however the test for trend was not
statistically significant in any model. The other blood parameters did not show any significant
association with alcohol intake in men (table 4.23).
In women, HDL cholesterol was highest in the highest alcohol intake category and lowest in
no consumers (p for trend <0.001) in all models (table 4.23). The other blood parameters did
not show any significant relationship with categories of alcohol use in women, although
atherogenic index decreased with alcohol consumption.
170 | P a g e
Table 4.22 Association between lipids, lipoproteins and glucose mean (SE) and alcohol use
(g) in all participants
All
T1 (none) T2 (moderate) T3 (high) P trend
Alcohol (g) 0 >0≤24 >24
n 217 213 70
(mmol/L) Model 1 3.76 (0.06) 3.78 (0.06) 3.89 (0.11) 0.296
Model 2 3.92 (0.11) 3.74 (0.11) 3.90 (0.17) 0.934

(mmol/L) Model 1 0.89 (0.04) 0.98 (0.04) 0.86 (0.07) 0.756
Model 2 0.90 (0.07) 1.01 (0.07) 0.96 (0.11) 0.964
a
(mmol/L) Model 1 2.49 (0.05) 2.48 (0.05) 2.43 (0.10) 0.639
Model 2 2.64 (0.10) 2.38 (0.10) 2.58 (0.16) 0.857
a
a,c
(mmol/L) Model 1 1.09 (0.03) 1.11 (0.03) 1.25 (0.05) 0.004
Model 2 1.11 (0.05) 1.16 (0.05) 1.20 (0.08) 0.206
(mmol/L) Model 1 4.61 (0.06) 4.60 (0.05) 4.58 (0.10) 0.817
Model 2 4.57 (0.11) 4.71 (0.10) 4.57 (0.17) 0.934
a
c
Model 1 3.59 (0.07) 3.68 (0.07) 3.36 (0.13) 0.135
Model 2 3.62 (0.14) 3.50 (0.13) 3.26 (0.22) 0.101
a c
High vs. None (p<0.05), High vs. Moderate (p <0.05)
171 | P a g e
Table 4.23 Association between lipids, lipoproteins and glucose mean (SE) and alcohol use (g) in men and women
Men Women
T1 (none) T2 (moderate) T3 (high) P trend T1 (none) T2 (moderate) T3 (high) P trend
Alcohol (g) 0 >0≤24 >24 0 >0≤24 >24
n 71 82 47 145 131 23
Total cholesterol Unadjusted 3.61 (0.11) 3.70 (0.10) 3.62 (0.14) 0.973 3.93 (0.07) 3.81 (0.07) 4.04 (0.18) 0.576
(mmol/L) Model 1 3.59 (0.10) 3.73 (0.10) 3.59 (0.12) 0.996 3.91 (0.07) 3.81 (0.07) 4.15 (0.17) 0.187
Model 2 3.57 (0.20) 3.68 (0.20) 3.57 (0.22) 0.992 4.00 (0.12) 3.87 (0.13) 4.28 (0.20) 0.110
c
Triglycerides Unadjusted 0.96 (0.08) 1.05 (0.07) 1.02 (0.09) 0.613 0.83 (0.05) 0.87 (0.05) 0.61 (0.12) 0.084
(mmol/L) Model 1 0.94 (0.07) 1.08 (0.07) 1.00 (0.08) 0.591 0.82 (0.04) 0.88 (0.05) 0.70 (0.11) 0.324
Model 2 0.79 (0.13) 0.95 (0.14) 0.83 (0.15) 0.787 0.80 (0.08) 0.89 (0.08) 0.67 (0.13) 0.303
(mmol/L) Model 1 2.31 (0.09) 2.44 (0.08) 2.23 (0.11) 0.599 2.65 (0.06) 2.50 (0.07) 2.63 (0.16) 0.890
Model 2 2.40 (0.17) 2.53 (0.17) 2.34 (0.19) 0.698 2.68 (0.12) 2.50 (0.12) 2.68 (0.19) 0.960
a,c
HDL cholesterol Unadjusted 1.07 (0.05) 1.10 (0.05) 1.11 (0.06) 0.604 1.11 (0.03) 1.13 (0.03) 1.38 (0.07) <0.001
a,c
(mmol/L) Model 1 1.07 (0.05) 1.10 (0.05) 1.11 (0.06) 0.608 1.11 (0.03) 1.13 (0.03) 1.37 (0.07) <0.001
a
Model 2 1.01 (0.10) 1.03 (0.10) 1.04 (0.11) 0.731 1.17 (0.05 1.17 (0.05) 1.44 (0.08) <0.001
(mmol/L) Model 1 4.49 (0.10) 4.61 (0.09) 4.47 (0.12) 0.912 4.72 (0.06) 4.58 (0.06) 4.69 (0.15) 0.843
Model 2 4.75 (0.19) 4.90 (0.19) 4.79 (0.21) 0.802 4.70 (0.11) 4.57 (0.12) 4.68 (0.19) 0.944
a
b
Model 1 3.43 (0.12) 3.78 (0.11) 3.39 0.813 3.72 (0.09) 3.55 (0.09) 3.33 (0.21) 0.094
c
(0.15)
b
Model 2 3.12 (0.24) 3.48 (0.24) 3.12 (0.26) 0.993 3.58 (0.16) 3.43 (0.16) 3.13 (0.26) 0.059
a b c
High vs. None (p<0.05), Moderate vs. None (p<0.05), High vs. Moderate (p <0.05)
172 | P a g e
Model 1 [age (years) adjusted], Model 2 [Age (yea rs), energy (kcal), education , smoking history (pack years), physical a ctivi ty level, in come]
Table 4.24 shows the association of physical measurements with alcohol intake. Significant
inverse linear association between alcohol use and BMI was found in the crude model (p for
trend <0.001) but not in the adjusted models. An inverse linear association between alcohol
use and body fat percent was highly significant in the crude model (p for trend <0.001) and
model 2 (p for trend =0.032) but not model 1. WHR increased with higher use of alcohol in
the crude model (p for trend =0.006) but the test for trend was no longer significant in
models 1 and 2. No clear association between alcohol use and WC, SBP or DBP was found in
all participants.
In men, no significant association was found between alcohol use and BMI, WC , body fat
percentage, SBP or DBP table 4.25. However, WHR was higher in high consumers compared
to moderate consumers (p<=0.033), although no significant test for trend were observed.
In women, BMI was significantly lower in high alcohol users compared to moderate users of
alcohol (p=0.014), and the test for trend was significant in model 1 (p=0.018) and model 2
(p=0.020). WC decreased with increasing alcohol use, but was not significant. Body fat
percentage also decreased significantly with alcohol use in model 1 (p for trend=0.044), but
the trend was not significant in model 2.
The association between alcohol use and CVD risk factors in women should be interpreted
with caution since there was a small (<30) sample size in women in the high alcohol use
category.
173 | P a g e
Table 4.24 Association between physical measurements mean (SE) and alcohol use (g) in all
participants
All
T1 (none) T2 (moderate) T3 (high) P trend
Alcohol (g) 0 >0≤24 >24
n 239 235 79
2 a,c
BMI (kg/m ) Unadjusted 24.67 (0.32) 24.44 (0.32) 22.28 (0.55) <0.001
c
Model 1 23.84 (0.29) 24.07 (0.28) 22.78 (0.50) 0.067
a
Model 2 24.50 (0.56) 23.69 (0.55) 22.22 (0.88) 0.115
WC (cm) Unadjusted 78.51 (0.68) 77.89 (0.68) 77.42 (1.17) 0.421

Model 1 77.80 (0.64) 77.80 (0.62) 77.31 (1.13) 0.710
Model 2 78.52 (1.27) 76.57 (1.22) 75.70 (1.97) 0.246
a,c
WHR Unadjusted 0.77 (0.01) 0.77 (0.01) 0.80 (0.01) 0.006
Model 1 0.78 (0.01) 0.78 (0.004) 0.79 (0.01) 0.794
Model 2 0.78 (0.01) 0.77 (0.01) 0.79 (0.01) 0.596
a,c
Body fat % Unadjusted 31.08 (0.71) 29.47 (0.71) 23.26 (1.29) <0.001
Model 1 27.66 (0.43) 27.62 (0.41) 26.22 (0.78) 0.108
b a
Model 2 30.19 (0.78) 27.51 (0.77) 24.88 (1.32) 0.032
Model 1 122.06 (0.93) 123.38 (0.89) 120.70 (1.62) 0.466
Model 2 119.72 (1.81) 121.11 (1.74) 117.69 (2.81) 0.256
Model 1 74.25 (0.73) 74.91 (0.71) 73.43 (1.28) 0.580
Model 2 73.98 (1.44) 73.72 (1.39) 73.86 (2.24) 0.663
a b c
High vs. None (p<0.05), Moderate vs. None (p<0.05), High vs. Moderate (p <0.05)
174 | P a g e
Table 4.25 Association between physical measurements mean (SE) and alcohol use (g) in men and women
Men Women
T1 (none) T2 (moderate) T3 (high) P trend T1 (none) T2 (moderate) T3 (high) P trend
Alcohol (g) 0 >0≤24 >24 0 >0≤24 >24 Alcohol (g)
n 78 91 54 161 144 25
2 a,c
a,c
Model 1 21.62 (0.38) 21.99 (0.36) 21.96 (0.46) 0.579 25.99 (0.36) 26.06 (0.38) 23.62 (0.92) 0.018
a
Model 2 20.32 (0.72) 20.44 (0.74) 20.61 (0.80) 0.670 26.45 (0.69) 26.59 (0.71) 24.05 (1.12) 0.020
a,c
WC (cm) Unadjusted 76.20 (1.10) 75.91 (1.00) 78.84 (1.30) 0.123 79.61 (0.85) 79.16 (0.89) 74.23 (2.19) 0.022
Model 1 76.11 (0.96) 76.16 (0.87) 78.54 (1.14) 0.101 79.27 (0.78) 79.21 (0.81) 75.75 (2.00) 0.103
Model 2 72.30 (1.80) 71.88 (1.85) 74.39 (2.05) 0.214 79.00 (1.46) 78.60 (1.51) 75.25 (2.40) 0.084
c
WHR Unadjusted 0.82 (0.01) 0.81 (0.01) 0.83 (0.01) 0.232 0.75 (0.01) 0.75 (0.01) 0.73 (0.02) 0.221
c
Model 1 0.82 (0.01) 0.81 (0.01) 0.83 (0.01) 0.226 0.75 (0.01) 0.75 (0.01) 0.74 (0.01) 0.555
Model 2 0.81 (0.01) 0.80 (0.01) 0.83 (0.01) 0.273 0.74 (0.01) 0.74 (0.01) 0.73 (0.02) 0.536
a,c
Body fat% Unadjusted 19.16 (0.70) 18.68 (0.63) 19.29 (0.86) 0.906 36.44 (0.59) 36.12 (0.62) 31.58 (1.57) 0.004
a,c
Model 1 18.97 (0.60) 18.88 (0.54) 19.20 (0.73) 0.808 36.17 (0.53) 36.18 (0.56) 33.11 (1.41) 0.044
Model 2 17.44 (1.12) 17.14 (1.12) 17.61 (1.25) 0.873 36.44 (0.97) 36.55 (1.01) 33.52 (1.68) 0.062
SBP (mmHg) Unadjusted 123.44 (1.59) 126.45 (1.46) 123.05 (1.90) 0.873 120.88(1.10) 119.84 (1.16) 116.76 (2.79) 0.171
Model 1 123.28 (1.56) 126.65 (1.43) 122.94 (1.86) 0.172 120.66 (1.03) 119.85 (1.09) 118.57 (2.64) 0.713
Model 2 120.03 (2.91) 122.79 (3.02) 120.88 (3.30) 0.418 119.18 (1.96) 117.21 (2.03) 114.81 (3.20) 0.484
Model 1 73.45 (1.23) 75.33 (1.13) 73.13 (1.47) 0.392 74.89 (0.83) 74.35 (0.87) 73.56 (2.10) 0.800
Model 2 71.40 (2.30) 72.94 (2.38) 72.01 (2.60) 0.652 73.39 (1.57) 72.33 (1.62) 71.76 (2.56) 0.623
a c
High vs None (p<0.05), High vs Moderate (p<0.05), Model 1 [age (years) adjusted], Model 2 [Age (years), energy (kcal), education, smoking history (pack
years), physical activity level, income]
175 | P a g e
4.9 Association between dietary patterns and CVD risk factors

4.9.1 Association between dietary patterns and CVD risk factors in women
4.9.1.1 “high sweets” pattern
Table 4.26 shows the association between “high sweets” dietary pattern and tertiles
expressed as (Low, medium and high) of factor score, and levels of lipids, lipoproteins and
glucose in women. Participants with a high factor score are more likely to eat most of the
food items in a particular dietary pattern compared to those with a medium or low factor
score. Significant inverse linear associations between tertiles of factor score for the “high
sweets” pattern and total cholesterol (p for trend =0.001), LDL cholesterol (p for trend
=0.003) and atherogenic index (p for trend =0.014) were found in the crude model, however
subsequent models were not statistically significant.
As shown in table 4.27, a significant inverse linear association between the “high sweet”
pattern and body fat percent was found in the crude model (p for trend =0.047) but not in
model 1 or model 2. No significant association was found between this dietary pattern and
the rest of the physical measurements.
4.9.1.2 “High fruit” pattern
Tertiles of factor score of the “high fruit” dietary pattern in women were significantly linearly
associated with total cholesterol in the crude model (p for trend =0.032) but not in model s 1
or 2. No significant association between this dietary pattern and the rest of the lipids
lipoproteins, glucose or physical measurements was found as shown in table 4.28 and table
4.29. However, WC and WHR in the medium tertile of the “high fruit” pattern were
significantly lower than the low and high tertile of “high fruit” pattern.
176 | P a g e
Table 4.26 Association between lipids, lipoproteins and glucose and the “high sweets”
pattern
Tertiles of high sweets dietary pattern factor scores (Women)

T1 (Low) T2 (Medium) T3 (High) P for trend
n 99 96 105
a,c
Total cholesterol Unadjusted 4.06 (0.08) 3.96 (0.09) 3.67 (0.08) 0.001
(mmol/L) Model 1 3.97 (0.08) 3.94 (0.08) 3.77 (0.08) 0.088
Model 2 4.02 (0.14) 3.99 (0.13) 3.89 (0.13) 0.306
c
Triglycerides Unadjusted 0.87 (0.06) 0.91 (0.06) 0.74 (0.06) 0.108
(mmol/L) Model 1 0.79 (0.53) 0.89 (0.05) 0.83 (0.05) 0.606
Model 2 0.77 (0.09) 0.87 (0.09) 0.81 (0.009) 0.646
a,c
(mmol/L) Model 1 2.65 (0.08) 2.61 (0.08) 2.51 (0.08) 0.193
Model 2 2.66 (0.13) 2.61 (0.13) 2.56 (0.13) 0.450
(mmol/L) Model 1 1.15 (0.03) 1.16 (0.03) 1.10 (0.03) 0.259
Model 2 1.19 (0.06) 1.20 (0.05) 1.16 (0.06) 0.603
(mmol/L) Model 1 4.66 (0.08) 4.65 (0.07) 4.66 (0.07) 0.981
Model 2 4.59 (0.13) 4.60 (0.12) 4.70 (0.12) 0.388
a
Model 1 3.65 (0.11) 3.60 (0.11) 3.60 (0.10) 0.731
Model 2 3.56 (0.18) 3.48 (0.17) 3.46 (0.17) 0.565
a c
High vs. Low (p<0.05), High vs. Medium (p <0.05)
Model 1 [age (years) adjusted]
Model 2 [Age (years), energy (kcal), educa tion, smoking history (pa ck years), physical a ctivity level (PAL),
income]
177 | P a g e
Table 4.27 Association between physical measurements and the “high sweets” pattern
Tertiles of high sweets dietary pattern factor scores (Women)

n 109 109 108
2
Model 1 25.85 (0.45) 25.34 (0.44) 26.31 (0.45) 0.478
Model 2 26.34 (0.79) 25.70 (0.74) 26.90 (0.75) 0.458
WC (cm) Unadjusted 79.91 (1.03) 78.73 (1.03) 78.29 (1.04) 0.269

Model 1 78.53 (0.95) 78.28 (0.94) 80.06 (0.96) 0.270
Model 2 78.34 (1.68) 77.70 (1.57) 79.59 (1.59) 0.441
WHR Unadjusted 0.76 (0.01) 0.76 (0.01) 0.74 (0.01) 0.085

Model 1 0.75 (0.01) 0.75 (0.01) 0.75 (0.01) 0.924
Model 2 0.74 (0.01) 0.75 (0.01) 0.74 (0.01) 0.900
a
Model 1 35.82(0.64) 35.84 (0.65) 35.27 (0.66) 0.636
Model 2 35.66 (1.12) 36.02 (1.07) 37.01 (1.07) 0.216
Model 1 119.43 (1.27) 120.25 (1.25) 120.66 (1.28) 0.504
Model 2 118.69 (2.48) 116.00 (2.50) 116.01 (2.54) 0.774
Model 1 73.95 (1.02) 74.72 (1.00) 74.86 (1.02) 0.535
Model 2 72.49 (2.03) 70.91 (2.05) 73.33 (2.08) 0.533
a
High vs. Low (p<0.05)
Model 2 [Age (years), energy (kcal) , educa tion (, smoking history (pa ck years), physical activity level (PAL),
income]
178 | P a g e
Table 4.28 Association between lipids lipoproteins and glucose and the “high fruit” pattern
Tertiles of high fruit dietary pattern factor scores (Women)

n 98 97 99
b a
Total cholesterol Unadjusted 3.72 (0.09) 3.98 (0.08) 3.97 (0.08) 0.032
(mmol/L) Model 1 3.77 (0.08) 3.97 (0.08) 3.93 (0.08) 0.175
Model 2 3.85 (0.13) 4.03 (0.14) 4.00 (0.13) 0.228

(mmol/L) Model 1 0.83 (0.05) 0.83 (0.05) 0.85 (0.05) 0.778
Model 2 0.82 (0.09) 0.83 (0.09) 0.83 (0.09) 0.889
(mmol/L) Model 1 2.51 (0.08) 2.65 (0.08) 2.60 (0.08) 0.379
Model 2 2.54 (0.12) 2.67 (0.13) 2.61 (0.12) 0.531
(mmol/L) Model 1 1.10 (0.03) 1.17 (0.03) 1.14 (0.03) 0.418
Model 2 1.14 (0.05) 1.21 (0.06) 1.20 (0.05) 0.269
(mmol/L) Model 1 4.71 (0.07) 4.70 (0.07) 4.55 (0.07) 0.132
Model 2 4.69 (0.12) 4.69 (0.13) 4.55 (0.12) 0.182
Model 1 3.61 (0.10) 3.61 (0.10) 3.63 (0.10) 0.907
Model 2 3.51 (0.17) 3.53 (0.17) 3.45 (0.17) 0.660
a b
High vs. Low (p<0.05), Medium vs. Lo w (p<0.05)
Model 2 [Age (years), energy (kcal), educa tion, smoking history (pa ck year s), physical a ctivity level (PAL),
income]
179 | P a g e
Table 4.29 Association between physical measurements and the “high fruit” pattern
Tertiles of high fruit dietary pattern factor scores (Women)

n 112 110 109
2
Model 1 26.28 (0.44) 25.19 (0.44) 26.04 (0.45) 0.693
Model 2 26.64 (0.73) 25.78 (0.77) 26.45 (0.74) 0.771
WC (cm) Unadjusted 79.51 (1.01) 77.03 (1.03) 80.39 (1.03) 0.544

Model 1 80.63 (0.92) 76.60 (0.93) 79.57 (0.94) 0.420
Model 2 80.06 (1.52) 76.07 (1.61) 78.96 (1.55) 0.418
c
WHR Unadjusted 0.75 (0.01) 0.74 (0.01) 0.76 (0.01) 0.387
Model 1 0.76 (0.01) 0.74 (0.01) 0.76 (0.01) 0.763
Model 2 0.75 (0.01) 0.72 (0.01) 0.75 (0.01) 0.804
Model 1 36.48 (0.64) 35.58 (0.64) 35.84 (0.65) 0.482
Model 2 36.70 (1.04) 36.07 (1.10) 36.11 (1.05) 0.532
Model 1 118.60 (1.24) 121.68 (1.24) 120.07 (1.26) 0.410
Model 2 115.34 (2.21) 122.13 (2.64) 116.93 (2.07) 0.499
Model 1 73.01 (0.99) 75.97 (0.99) 74.57 (1.01) 0.273
Model 2 71.51 (1.81) 73.50 (2.15) 73.20 (1.69) 0.667
income]
180 | P a g e
4.9.2 Association between dietary patterns and CVD risk factors in men
4.9.2.1 “Mixed” pattern
Table 4.30 shows the association between the factor score of the “mixed diet” pattern with
lipids, lipoproteins and glucose in men. Total cholesterol, triglycerides , HDL and random
glucose were not associated with the “mixed diet” pattern. However, there was a trend of a
positive linear association between LDL cholesterol and the “mixed diet” pattern in model 2
(p for trend=0.071). Furthermore, atherogenic index was significantly positively associated
with the “mixed diet” pattern in the crude model (p for trend =0.031) but not in models 1 or
2. No significant association between the “mixed diet” pattern and physical measurements
was found (table 4.31).
181 | P a g e
Table 4.30 Association between lipids, lipoproteins and glucose and the “mixed” pattern
Tertiles of mixed dietary pattern factor scores (Men)

n 66 71 63
(mmol/L) Model 1 3.59 (0.11) 3.74 (0.10) 3.61 (0.11) 0.898
Model 2 3.50 (0.21) 3.67 (0.20) 3.61 (0.20) 0.478

(mmol/L) Model 1 0.97 (0.07) 1.03 (0.07) 1.04 (0.07) 0.478
Model 2 0.81 (0.14) 0.85 (0.14) 0.89 (0.14) 0.438
(mmol/L) Model 1 2.21 (0.09) 2.43 (0.09) 2.39 (0.10) 0.171
Model 2 2.26 (0.18) 2.48 (0.17) 2.50 (0.17) 0.071
(mmol/L) Model 1 1.15 (0.05) 1.12 (0.05) 1.01 (0.06) 0.069
Model 2 1.07 (0.11) 1.07 (0.10) 0.95 (0.10) 0.134
(mmol/L) Model 1 4.53 (0.10) 4.61 (0.10) 4.44 (0.10) 0.532
Model 2 4.85 (0.20) 4.85 (0.20) 4.75 (0.20) 0.480
a
index Model 1 3.39 (0.13) 3.59 (0.12) 3.71 (0.13) 0.082
Model 2 3.04 (0.25) 3.21 (0.24) 3.40 (0.24) 0.058
a
High vs. Low (p<0.05)
Model 2 [Age (years), energy (kcal), educa tion, smoking history (pa ck years), physical a ctivity level, inco me]
182 | P a g e
Table 4.31 Association between physical measurements and the “mixed” pattern
Tertiles of mixed dietary pattern factor scores (Men)

n 74 74 74
2
Model 1 22.08 (0.40) 21.62 (0.39) 21.86 (0.39) 0.682
Model 2 20.70 (0.77) 20.26 (0.75) 20.42 (0.73) 0.612
WC (cm) Unadjusted 77.06 (1.12) 75.53 (1.12) 77.57 (1.12) 0.747

Model 1 77.25 (0.98) 75.96 (0.97) 76.95 (0.98) 0.830
Model 2 73.35 (1.94) 72.19 (1.87) 72.62 (1.83) 0.608
WHR Unadjusted 0.81 (0.01) 0.82 (0.01) 0.82 (0.01) 0.196

Model 1 0.81 (0.01) 0.82 (0.01) 0.82 (0.01) 0.348
Model 2 0.81 (0.01) 0.81 (0.01) 0.81 (0.01) 0.658
Model 1 19.08 (0.62) 19.11 (0.58) 18.76 (0.63) 0.714
Model 2 17.39 (1.18) 17.84 (1.13) 16.93 (1.13) 0.606
Model 1 125.33 (1.60) 124.42 (1.59) 123.98 (1.61) 0.553
Model 2 123.95 (2.94) 122.51 (2.36) 121.81 (2.58) 0.806
Model 1 74.17 (1.26) 74.03 (1.25) 74.25 (1.26) 0.965
Model 2 73.29 (2.28) 71.65 (1.84) 73.97 (2.01) 0.755
Model 2 [Age (years), energy (kcal), educa tion, smoking history (pa ck years), physical a ctivity level, inco me]
183 | P a g e
4.9.2.2 “Convenience” pattern
As shown in table 4.32 significant association was only found between blood lipids,
lipoproteins and glucose and the “convenience diet” pattern in the crude model. However,
mean HDL-cholesterol was lower in the medium tertile of convenience diet pattern
compared to the lower low tertile (p=0.041).
Table 4.33 shows the mean physical measurements according to tertiles of factor scores of
the “convenience diet” pattern. Mean BMI increased with increasing tertile of factor scores
of the “convenience diet” pattern, and the association was suggestive of a positive
association in model 1 (p=0.061) but not model 2. WHR had a significant linear inverse
association with the factor score of the “convenience diet” pattern in the crude model (p for
trend =0.008) but not in the subsequent models. A significant association was also found
between the “convenience diet” pattern and DBP in the crude model (p for trend =0.012). No
association between factor score and WC, body fat and SBP were found.
184 | P a g e
Table 4.32 Association between lipids, lipoproteins and glucose and the “convenience”
pattern
Tertiles of convenience dietary pattern factor scores (Men)

n 62 64 66
(mmol/L) Model 1 3.74 (0.11) 3.50 (0.11) 3.69 (0.10) 0.766
Model 2 3.69 (0.19) 3.40 (0.21) 3.63 (0.21) 0.725
b a
(mmol/L) Model 1 1.07 (0.07) 0.98 (0.07) 0.99 (0.07) 0.439
Model 2 0.91 (0.13) 0.80 (0.14) 0.80 (0.15) 0.357
(mmol/L) Model 1 2.33 (0.10) 2.25 (0.10) 2.45 (0.09) 0.371
Model 2 2.41 (0.17) 2.32 (0.18) 2.56 (0.19) 0.340
(mmol/L) Model 1 1.17 (0.06) 1.02 (0.06) 1.09 (0.05) 0.332
Model 2 1.09 (0.10) 0.92 (0.11) 1.01 (0.11) 0.398
(mmol/L) Model 1 4.57 (0.10) 4.62 (0.10) 4.41 (0.01) 0.304
Model 2 4.84 (0.19) 4.91 (0.20) 4.70 (0.21) 0.409
Model 1 3.55 (0.13) 3.54 (0.13) 3.59 (0.13) 0.819
Model 2 3.23 (0.24) 3.20 (0.26) 3.30 (0.26) 0.759
a b
Model 2 [Age (years), energy (kcal), educa tion, smoking history (pa ck years), physical a ctivity level (PAL), in come
185 | P a g e
Table 4.33 Association between physical measurements and the “convenience” pattern
Tertiles of convenience dietary pattern factor scores (Men)

n 74 74 75
2
Model 1 21.26 (0.40) 21.96 (0.39) 22.33 (0.39) 0.061
Model 2 20.00 (0.73) 20.58 (0.76) 20.96 (0.78) 0.138
WC (cm) Unadjusted 78.63 (1.11) 75.41 (1.12) 76.12 (1.10) 0.110

Model 1 76.76 (1.00) 76.05 (0.99) 77.05 (0.98) 0.836
Model 2 73.23 (1.83) 72.37 (1.91) 71.90 (1.97) 0.412
b a
WHR Unadjusted 0.84 (0.01) 0.81 (0.01) 0.81 (0.01) 0.008
Model 1 0.83 (0.01) 0.81 (0.01) 0.81 (0.01) 0.203
Model 2 0.82 (0.01) 0.81 (0.01) 0.80 (0.01) 0.063
Model 1 18.73 (0.62) 19.08 (0.63) 19.15 (0.61) 0.649
Model 2 17.33 (1.11) 17.46 (1.20) 17.39 (1.20) 0.953
Model 1 124.98 (1.64) 125.46 (1.63) 123.32 (1.60) 0.476
Model 2 123.44 (2.48) 121.56 (2.94) 126.31 (3.26) 0.735
a,c
c
Model 1 74.62 (1.28) 75.86 (1.27) 72.01 (1.25) 0.151
Model 2 74.09 (1.90) 74.73 (2.49) 73.67 (2.49) 0.660
a b c
High Vs Low (p<0.05), Medium vs. Low (p <0.05), High vs. Mediu m (p<0 .05)
income]
186 | P a g e
4.9.2.3 “Traditional diet” pattern
As shown in table 4.34 no significant association was found between the “traditional diet”
and blood lipids, lipoproteins and glucose. However HDL-cholesterol was significantly lower
in the highest tertile than the middle tertile in model 2 (p=0.043). Conversely, the
atherogenic index was significantly higher in the highest tertile compared to the middle
tertile in model 2 (p=0.025). No significant relationship between physical measurements and
the “traditional diet” pattern was found (table 4.35).
187 | P a g e
Tertiles of traditional dietary pattern factor scores (Men)

n 64 70 66
(mmol/L) Model 1 3.68 (0.11) 3.66 (0.10) 3.60 (0.11) 0.617
Model 2 3.59 (0.20) 3.65 (0.20) 3.59 (0.21) 0.962

(mmol/L) Model 1 1.04 (0.07) 0.97 (0.70) 1.02 (0.07) 0.842
Model 2 0.84 (0.14) 0.77 (0.14) 0.94 (0.14) 0.419
(mmol/L) Model 1 2.39 (0.10) 2.25 (0.09) 2.41 (0.09) 0.855
Model 2 2.41 (0.18) 2.35 (0.18) 2.54 (0.18) 0.404
(mmol/L) Model 1 1.08 (0.05) 1.16 (0.05) 1.04 (0.05) 0.516
Model 2 1.03 (0.11) 1.12 (0.11) 0.95 (0.11) 0.420
(mmol/L) Model 1 4.58 (0.10) 4.55 (0.10) 4.47 (0.10) 0.446
Model 2 4.78 (0.20) 4.82 (0.20) 4.84 (0.20) 0.704
Model 1 3.56 (0.13) 3.44 (0.12) 3.69 (0.13) 0.459
Model 2 3.15 (0.24) 3.06 (0.24) 3.49 (0.25) 0.109
income]
188 | P a g e
Tertiles of traditional dietary pattern factor scores (Men)

n 74 74 75
2
Model 1 22.11 (0.39) 21.43 (0.39) 22.02 (0.39) 0.872
Model 2 20.68 (0.75) 19.88 (0.75) 20.61 (0.76) 0.917
WC (cm) Unadjusted 76.76 (1.13) 75.97 (1.11) 77.43 (1.11) 0.675

Model 1 77.03 (0.99) 75.62 (0.97) 77.52 (0.97) 0.725
Model 2 72.99 (1.89) 71.39 (1.89) 73.25 (1.92) 0.873
WHR Unadjusted 0.82 (0.01) 0.81 (0.01) 0.82 (0.01) 0.940

Model 1 0.82 (0.01) 0.81 (0.01) 0.82 (0.01) 0.980
Model 2 0.81 (0.01) 0.80 (0.01) 0.81 (0.01) 0.715
Model 1 18.61 (0.61) 19.22 (0.60) 19.12 (0.61) 0.553
Model 2 17.17 (1.15) 17.77 (1.16) 17.28 (1.17) 0.916
Model 1 126.09 (1.61) 121.85 (1.57) 125.85 (1.57) 0.914
Model 2 123.92 (2.81) 118.24 (3.16) 126.05 (3.04) 0.858
Model 1 75.68 (1.26) 72.73 (1.25) 74.07 (1.25) 0.365
Model 2 76.64 (2.19) 70.10 (2.47) 74.53 (2.38) 0.559
income]
189 | P a g e
4.9.2.4 “High vegetable” pattern
As shown on table 4.36 total-cholesterol was positively associated with “high vegetable diet”
pattern in the crude model (p for trend =0.023) and model 1 (p for trend=0.027), but not in
model 2. A similar but non-significant association was found for LDL-cholesterol. No
significant association relationship was established with triglycerides, random glucose, HDL-
cholesterol and atherogenic index. However, HDL cholesterol in the middle tertile was
significantly higher than the lowest tertile.
While a weak positive linear trend was found between BMI and “vegetable diet” pattern in
model 1 (p=0.090), this trend remained non significant in model 2. No significant associations
were found between the “high vegetable diet” and the rest of physical measurement as
shown in table 4.37.
190 | P a g e
Table 4.36 Association between lipids, lipoproteins and glucose and the “high vegetable”
pattern
Tertiles of high vegetable dietary pattern factor scores (men)

n 62 66 64
b a
Tot. cholesterol Unadjusted 3.39 (0.11) 3.81 (0.11) 3.75 (0.11) 0.023
(mmol/L) Model 1 3.42 (0.11) 3.76 (0.10) 3.75 (0.10) 0.027
Model 2 3.44 (0.20) 3.69 (0.20) 3.68 (0.20) 0.137

(mmol/L) Model 1 1.03 (0.07) 0.94 (0.07) 1.07 (0.07) 0.692
Model 2 0.88 (0.14) 0.77 (0.14) 0.91 (0.14) 0.760
(mmol/L) Model 1 2.24 (0.09) 2.33 (0.09) 2.47 (0.09) 0.090
Model 2 2.37 (0.18) 2.41 (0.18) 2.52 (0.18) 0.322
b
b
(mmol/L) Model 1 1.02 (0.05) 1.18 (0.05) 1.07 (0.05) 0.494
Model 2 0.98 (0.11) 1.10 (0.11) 1.00 (0.10) 0.766
(mmol/L) Model 1 4.46 (0.10) 4.58 (0.10) 4.58 (0.10) 0.477
Model 2 4.73 (0.20) 4.86 (0.20) 4.84 (0.20) 0.505

Index Model 1 3.54 (0.13) 3.49 (0.13) 3.65 (0.13) 0.533
Model 2 3.21 (0.25) 3.20 (0.25) 3.31 (0.25) 0.597
a b
Model 2 [Age (years), energy (kcal), educa tion, smoking history (pa ck years), physical a ctivity level (PAL), in come
191 | P a g e
Table 4.37 Association between physical measurements and the “high vegetable” pattern
Tertiles of high vegetable dietary pattern factor scores (Men)

n 72 76 75
2
Model 1 21.37 (0.40) 21.86 (0.39) 22.31 (0.39) 0.092
Model 2 19.99 (0.75) 20.53 (0.75) 20.78 (0.75) 0.189
WC (cm) Unadjusted 75.33 (1.14) 76.66 (1.10) 78.10 (1.10) 0.082

Model 1 76.00 (0.99) 76.03 (0.96) 78.11 (0.96) 0.127
Model 2 72.00 (1.87) 72.13 (1.88) 73.64 (1.88) 0.281
WHR Unadjusted 0.81 (0.01) 0.81 (0.01) 0.82 (0.01) 0.274

Model 1 0.82 (0.01) 0.81 (0.01) 0.82 (0.01) 0.416
Model 2 0.81 (0.01) 0.80 (0.01) 0.82 (0.01) 0.440
Model 1 18.45 (0.61) 19.26 (0.60) 19.24 (0.61) 0.362
Model 2 16.91 (1.15) 17.81 (1.15) 17.44 (1.15) 0.578
Model 1 123.81 (1.62) 126.53 (1.57) 123.32 (1.59) 0.832
Model 2 123.34 (3.51) 124.41 (2.73) 121.82 (2.70) 0.454
Model 1 73.20 (1.28) 75.12 (1.24) 74.08 (1.25) 0.624
Model 2 72.00 (2.74) 75.86 (2.13) 72.53 (2.11) 0.951
income]
192 | P a g e
4.10 Association between fruit and vegetables intake and blood pressure
As shown in table 4.38 SBP and DBP had a significant positive linear association with tertiles
of residuals of vegetable intake in the crude model and model 1 but not in model 2. No
significant association between residuals of fruit/fruit and vegetable, and blood pressure
were found in all participants.
When the analysis was stratified by sex; the significant association between residuals of
vegetable intake and blood pressure was found only men in the crude model and model 1 for
both SBP and DBP No significant linear associations were found between residuals of fruit or
fruit and vegetable intake in men or women (table 4.39).
193 | P a g e
Table 4.38 Association between fruit and vegetables and blood pressure
All Tertiles of energy-adjusted fruit and vegetables intake

Fruit n 191 159 185
Model 1 121.64 (0.95) 123.54 (1.06) 122.54 (1.04) 0.525
a,c
Model 2 121.43 (1.70) 123.28 (1.90) 116.42 (1.81) 0.134
Model 1 74.20 (0.76) 75.35 (0.84) 73.87 (0.82) 0.767
Model 2 74.12 (1.36) 73.70 (1.52) 71.41 (1.45) 0.258
Vegetables n 218 186 149

a
a
Model 1 121.01 (0.91) 122.93 (1.02) 124.80 (1.21) 0.013
Model 2 120.29 (1.46) 122.32 (2.00) 121.97 (1.97) 0.269
b a
a
Model 1 73.23 (0.72) 74.97 (0.81) 76.31 (0.96) 0.011
Model 2 72.11 (1.18) 74.45 (1.61) 74.45 (1.59) 0.136
Fruit & veg 198 169 169

Model 1 120.97 (0.95) 121.82 (1.01) 123.17 (1.02) 0.117
Model 2 119.90 (1.62) 120.90 (1.90) 119.39 (1.73) 0.967
Model 1 73.55 (0.75) 74.80 (0.80) 74.95 (0.81) 0.205
Model 2 72.02 (1.29) 74.40 (1.51) 72.99 (1.37) 0.641
a b c
High vs. Low (p<0.05), Medium vs. Lo w (p<0.05), High vs. Mediu m (p<0.05)
Model 1 [age (years) and sex adjusted]
income]
194 | P a g e
Table 4.39 Association between fruit and vegetables and blood pressure
Men Women
Tertiles of energy-adjusted fruit and vegetables intake
T1 (Low) T2 (Medium) T3 (High) P trend T1 (Low) T2 (Medium) T3 (High) P trend
Fruit n 95 66 61 102 99 130
Model 1 123.07 (1.40) 126.73 (1.69) 124.60 (1.76) 0.496 120.05 (1.30) 120.23 (1.32) 120.16 (1.16) 0.947
Model 2 124.47 (2.39) 122.34 (2.90) 119.67 (3.29) 0.138 118.41 (2.41) 122.87 (2.53) 114.81 (2.10) 0.603
DBP (mmHg) Unadjusted 73.92 (1.16) 75.52 (1.39) 73.05 (1.43) 0.638 74.63 (109) 75.41 (1.10) 73.83 (0.96) 0.684
Model 1 73.77 (1.11) 75.30 (1.33) 73.55 (1.38) 0.899 74.49 (1.04) 75.28 (1.05) 74.05 (0.92) 0.751
Model 2 74.40 (1.90) 72.61 (2.30) 71.74 (2.61) 0.389 73.69 (1.96) 71.36 (1.72) 71.36 (1.72) 0.465
Vegetables n 116 66 42 104 120 107

b a
SBP (mmHg) Unadjusted 121.58 (1.28) 127.66 (168) 127.86 (2.11) 0.012 118.57 (1.37) 119.58 (1.27) 122.22 (1.35) 0.058
b a
Model 1 122.00 (1.28) 127.26 (1.67) 127.35 (2.09) 0.031 119.80 (1.29) 118.45 (1.20) 122.41 (1.26) 0.149
Model 2 121.90 (2.06) 130.77 (3.32) 123.02 (3.12) 0.897 118.35 (2.06) 114.92 (2.40) 121.34 (2.55) 0.133
b a
a
Model 1 72.28 (1.01) 75.29 (1.31) 77.41 (1.65) 0.009 74.04 (1.04) 74.51 (0.96) 75.10 (1.02) 0.463
Model 2 72.41 (1.62) 76.04 (2.63) 74.25 (2.44) 0.531 71.59 (1.70) 72.94 (1.99) 74.56 (2.11) 0.147
Fruit & veg n 102 55 57 97 119 115

b
Model 1 122.67 (1.32) 127.39 (1.79) 125.25 (1.79) 0.246 120.36 (1.33) 118.51 (1.20) 121.68 (1.22) 0.465
Model 2 122.55 (2.03) 123.15 (3.12) 123.06 (2.90) 0.758 117.70 (2.51) 116.86 (2.45) 117.79 (2.16) 0.768
Model 1 72.58 (1.04) 76.01 (1.41) 75.14 (1.40) 0.143 74.58 (1.07) 74.22 (0.96) 74.87 (0.98) 0.836
Model 2 72.08 (1.60) 74.80 (2.45) 73.59 (2.28) 0.895 71.90 (2.03) 73.23 (1.98) 72.80 (1.74) 0.606
195 | P a g e
a b
High vs. Low (p<0.05), Medium vs. Lo w (p<0.05), Model 1 [age (yea rs) adjusted], Mod el 2 [Age (years), educa tion, smoking history (pack years), physical
activity level (PAL), inco me]
Key Fi ndings
Total fat intake (% energy)
 Intake of total fat was positively linearly associated with LDL cholesterol level in all participants (p for
trend =0 .007). After stratifying by sex this significant association was restricted to men only.
 HDL cholesterol level was inversely associated with total fat intake in women (p for trend=0.032) but not
men.
 Intake of total fat had a linear positive association with atherogenic index in women (p for trend =0.030)
but not men .
 BMI was not associated with total fat intake in all participa nts, but after stratifying by sex total dietary fat
had a significant positive association with BMI of women (p for trend=0.039) but not men.
SFA (% energy)
 In all participants intake of SFA had a linear positive and significant association with levels of triglycerides
(p for trend =0 .048), LDL cholesterol (p for trend =0.017) and random glucose [model 2 (p for tr end
=0.044), and an inverse linear association with HDL chol esterol (p for trend =0.021).
 Intake of SFA was positively associated with LDL cholesterol level in men (p for trend=0.038) but not
women.
 SFA intake was inversely associated with HDL cholesterol (p for trend=0.012), and positively associated
with random glucose (p for trend =0.031) and atherogenic index (p for trend =0 .034) in women but not
men.
 Intake of SFA was positively associated with BMI in all pa rticipants (p for trend =0.003), but after
stratifying by sex this association was restricted to women only (p for trend=0.027).
PUFA (% energy)
 PUFA intake had a positive linear association with atherogenic index (p for trend =0.027) in all
participants. After stratifying by sex the association was no longer significant.
 PUFA intake had a negative linear association with WHR in the unadjusted model only in all participants (p
for trend=0.020). After stratifying by sex the negative association was restricted to women only.
PUFA:SFA r atio
 PUFA:SFA ratio was inversely associated with body fat in age adjusted model only and the association was
restricted to women only.
Fibre (g/100kcal)
 Fibre intake had an inverse association with random glucose after adjustment for age and sex (p for
trend=0 .007). After stratifying by sex this association was seen in men (p for trend=0.046) only.
196 | P a g e
Alcohol use
 Alcohol intake had a significant inverse association with LDL cholesterol and atherogenic index in the
unadjusted model only for all participants.
 Alcohol intake had a significant positive association with HDL cholesterol after age and sex adjustment but
the association weakened in the full model. After stratifying by sex alcohol intake had a significant
association with HDL cholesterol level in women (p for trend <0.001) but not men.
 Alcohol intake had a significant inverse linear association with body fat percent even after full model
adjustment (p for trend =0.032) in all participants. After stratifying by sex the linear association was
restricted to women although weakened.
 Alcohol intake had a significant inverse linear association with BMI in women only (p for trend=0 .020).
Dietary patterns
“High sweets” pattern (women)
 The “high sweets” dietary pattern was inversely associated with total cholesterol (p for trend =0.001), LDL
cholesterol (p for trend =0.003), atherogenic index (p for trend =0.014) and body fat percent (p for trend
=0.047) in the unadjusted models only.
“High fruit” pattern (women)

 The “High fruit” dietary pattern had a positive linear association with total cholesterol (p for trend =0.032)
in the unadjusted model only.
“Mixed” pattern (men)

 The “mixed” dietary pattern had a significant positive linear association with Atherogenic index in the
unadjusted model (p for trend =0.031) only.
“Convenience” pattern (men)

 A significant linear association was found between the “convenience” dietary pattern and total
triglycerides in the unadjusted model (p for trend =0.031) only.
 Significant inverse linear association was found between this pattern and WHR (p for tr end =0.008) and
DBP (p for trend =0.012) in the unadjusted models only.
“Traditional” pattern (men)

 The “traditional” dietary pattern had no significant association with any CVD risk factors.
“High vegetable” pattern (men)

 A significant linear association was found between the “vegetable” dietary pattern and total cholesterol in
the unadjusted model (p for trend =0.023) and age adjusted model (p for trend =0.027).
Fruit and vegetables consumption and blood pressure

 A Significant positive association between vegetable consumption and SBP and DBP was found in the age
and sex adjusted model in all participants . After stratifying by sex the significant association with SBP and
DBP was restricted to men only.
197 | P a g e
4.11 Discussion
In this cross-sectional study, higher intakes total fat (%E) and SFA (%E) were associated
with higher LDL cholesterol, higher atherogenic index, BMI and random glucose as well
as lower HDL cholesterol. Intake of PUFA had a significant association with atherogenic
index only. Higher intake of fibre (g/1000kcal) had a significant inverse association with
random blood glucose. Higher alcohol use (>24g/day) was associated with lower BMI
and higher HDL cholesterol in all participants and the association was significant in
women but not men. Associations between all dietary patterns and CVD risk factors
were not significant. However, higher intake of energy adjusted vegetables was
associated with higher SBP and DBP.
4.11.1 Saturated fat intake and CVD risk factors
SFA intake was positively associated with LDL cholesterol and atherogenic index but
was inversely associated with HDL cholesterol. These results are consistent with
Hegsted et al (1993) findings in their analysis of the combined published data of
metabolic studies and field trial; they concluded that “saturated fatty acids increase
serum cholesterol and they are the most powerful predictors of serum cholesterol
concentration”, and LDL cholesterol roughly parallel the changes in serum cholesterol.
Similarly, meta-analysis of 395 dietary experiments of metabolic studies of solid food
diets in healthy volunteers reported that in multivariate analysis, isocalori c increase in
saturated fat was significantly associated with increased LDL cholesterol (Clarke et al.,
198 | P a g e
1997). Another meta-analysis of 60 controlled trials of healthy adults older than 17
years reported that isocaloric replacement of carbohydrates (1%) with SFA increased
total cholesterol, LDL and HDL cholesterol but not atherogenic index (Mensink et al.,
2003). In the 35 out of 60 studies that reported intakes of individual SFA, lauric acid
(12:0) increased LDL, HDL and total cholesterol, but decreased atherogenic index
suggesting that there was a proportionally higher HDL cholesterol increase by lauric
acid. Myristric (14:0) and palmitic (16:0) acids increased total and LDL cholesterol but
reduced triglycerides, while stearic acid (18:0) decreased atherogenic index.
Conversely, a recent meta-analysis of prospective cohort studies evaluating the
association of saturated fat with CVD found no significant association with risk of CHD
RR 1.07 (95% CI; 0.96-1.19), stroke RR 0.81 (95% CI; 0.62-1.05)or total CVD RR 1.00
(95% CI; 0.89-1.11) (Siri-Tarino et al., 2010) for highest versus low intake of SFA.
In this study while only ~ 7% of participants had high LDL cholesterol, more than half of
all the participants had low HDL cholesterol, possibly due to an unbalanced PUFA:SFA
ratio. The association between individual SFA and CVD risk factors were not
investigated in this population. Therefore future research may provide more insight
into the extent to which individual SFA may be hypercholesterolemic, as has been
suggested by Hayes & Kholsa (1992) in their analysis of data from feeding trials of
cebus monkeys and humans.
199 | P a g e
4.11.2 Polyunsaturated fatty acids intake and CVD risk factors
In this study significant associations were found between PUFA intake and atherogenic
index but not other CVD risk factors despite reported intakes above 6%E as advised
(Harris et al., 2009). Main dietary contributors to PUFA intake in this population were
vegetable oils which were used by more than 80% of participants in cooking. Possible
reasons for failure to find any association between PUFA and other CVD risk factors are
(i) the majority of participants in this study were relatively young (median age <35
years) and generally healthy, therefore the association of PUFA intake and the CVD risk
factors may not be apparent (ii) even though the total fat intake was <30%E the
PUFA:SFA ratio was less ~0.7 which is less than the recommended balanced ratio of
1:1 (NCEP, 2002) (iii) the threshold of effective PUFA intake may possibly be high in
this population, given the potential influence of other variables such as genetics and
obesity (iv) vegetable oils contain a small amount of n-3 fatty acids and consumption
of foods high in n-3 PUFA such as fish was scarce in this population therefore n-3 PUFA
daily intake could also be potentially lower than the suggested 0.5-1% %E (Wijendran &
Hayes, 2004).
Other studies have reported that substituting PUFA (largely n-6) for carbohydrates was
associated with reduced LDL cholesterol, trigylcerides and atherogenic index, and
increased HDL cholesterol (Mensink et al., 2003). A prospective cohort study of 3277
healthy Danish men and women free of ischaemic heart disease (IHD) reported a high
intake of n-3 long chain PUFA (0.78 g/day) to be significantly cardio-protective
compared to a low intake (0.11 g/day) and this cardio-protection was restricted to
200 | P a g e
women (HR 0.62: 95% CI 0.40-0.97) (Vedtofte et al., 2011). A 20 years follow-up of the
Nurses’ Health study also reported that higher intake of PUFA (7.4 %E) was associated
with a 25% reduction in risk of CHD compared to low intake (4.1 %E) (Oh et al., 2005),
and substituting SFA with PUFA was associated with decreased risk of CHD in women
(Hu et al., 1997).
It has been suggested that n-6 PUFA compete with n-3 PUFA, and high n-6 PUFA may
offset the cardiovascular benefits of n-3 PUFA by contributing to the formation of
thrombi and atheroma (Simopoulos, 1999). However, in the Health Professionals
cohort study of more than 50 000 men; high or low n-6 PUFA intake was found not to
appreciably counteract or augment CHD benefit of high intake of long and
intermediate chain n-3 PUFA (Muzzafarian et al., 2005).
4.11.3 Fibre intake and CVD risk factors
In this study higher fibre intake was associated with lower random glucose and total
cholesterol after adjustment for age and sex. These finding are consistent with the
recent Cochrane review of trials investigating the effect of whole grain diet on risk
factors for CHD which reported that intake of oatmeal significantly reduced total
cholesterol -0.19 mmol/L (95% CI -0.30 to -0.08) and LDL cholesterol -0.18 mmol/L
(95% CI -0.28 to -0.09) compared to intake of refined grain diets (Kelly et al., 2009).
Dietary fibre, particularly soluble fibre from cereals, has also been reported to reduce
risk of CHD in epidemiological studies. In the Nurse’s Health Study of more than 68 000
201 | P a g e
women aged 37-64 years, intake of cereal fibre (5g/day increase) was associated with a
37% reduction in risk of CHD (Wolk et al., 1999). Similarly, in the NHANES I
epidemiological follow-up study, intake of 5.9g/day of soluble fibre compared to
0.9g/day of soluble fibre had a 15% and 24% lower risk of CHD and CVD respectively
(Bazzano et al., 2003).
In this study the median fibre intake of 8.24 g/1000kcal in the highest fibre tertile of
intake was lower than the recommended 10-15g/100kcal (National Research Council,
2005). This relatively low intake of fibre could be an indication that people in this
population are increasingly consuming more refined foods and animal sourced foods
consistent with a population in nutrition transition (Popkin & Gordon-Larson, 2004).
4.11.4 Alcohol intake and CVD risk factors
Higher alcohol use was associated with lower BMI and body fat %, and higher HDL
cholesterol compared to non-alcohol use in women but not men. Findings of a
beneficial effect of alcohol intake on HDL cholesterol in this study are consistent with
reported benefits of moderate alcohol consumption (up to 30g/day or 40g/day) on
biological markers (higher HDL and adeponectin and lower levels of fibrinogen) (Rimm
et al., 1999, Brien et al., 2011).
In the THUSA study in South Africa Gopane et al (2010) reported that mean alcohol
intake of 30.2g/day in men and 11.4 g/day in women was associated with significantly
higher HDL cholesterol compared to non-drinkers, but, alcohol drinkers compared to
202 | P a g e
non-drinkers also had significantly higher triglycerides, blood pressure and serum
ferritin which are potentially detrimental to health. Still in South Africa, Pisa et al
(2010) reported that self-reported alcohol intake of >26.6g/day and >4.6 g/day in men
and women respectively was related to significantly lower BMI and significantly higher
HDL cholesterol and blood pressure compared to non-drinkers. Individuals in the
highest quartile of gamma-glutamyl transferase [GGT (biological marker of alcohol
intake)] level also had a significantly lower BMI, and higher HDL cholesterol, total
cholesterol and triglycerides compared to the lowest quartile. Pisa et al (2010)
however concluded that the “... cardioprotective effect of alcohol possibly may
disappear because the increase in blood pressure offset the benefits of increase in HDL
cholesterol”.
A recent systematic review and meta-analysis of 84 prospective studies mostly in
developed countries, reported that alcohol consumption at 2.5-14.9 g/day was
associated with 14-25% reduction in the risk of total mortality, CHD incidence and
mortality and stroke incidence and mortality compared to non-drinkers. However,
intakes of more than 60 g/day increased risk of stroke mortality and incidence by more
than 40% (Ronksley et al., 2011). The mechanism for the benefit of moderate alcohol
consumption includes higher HDL cholesterol thus improving reverse cholesterol
transport and possible inhibition of blood clotting and platelet aggregation as well as
improved glucose metabolism (Agarwal, 2001, Djoussè et al., 2009). It has been
suggested that the benefits of alcoholic beverages are largely from the alcohol rather
than other components (such as antioxidants) in the drink (Rimm, 2000) and a pattern
203 | P a g e
of alcohol drinking characterised by irregular heavy drinking may increase the risk of
cardiovascular outcomes (van de Wiel & de Lange, 2008). In this study, <30 women
had a reported alcohol intake >24g/day (3 units/day), therefore the results should be
interpreted with caution.
4.11.5 Dietary patterns and CVD risk factors
It was expected that a “high vegetable”, “mixed” and “high fruit” diet would be
protective against CVD risk factors, and “traditional diet”, “high sweets” and
“convenience” diet would be positively associated with CVD risk factors. However, no
associations between dietary patterns and CVD risk factors were found. The lack of
association could be due to (i) participants were generally healthy and therefore their
risk profile may be low (ii) the possible lack of power in the study, for example women
had a larger sample size of about 300 which is considered good while the men’s sample
size was about 200 which is considered fair (Fields, 2009). Furthermore, dietary
patterns were difficult to name due to a variety of foods with high factor loading, for
example, “high sweets” pattern had sweets as the major foods, but meat and
vegetable also featured in the pattern too. Therefore the patterns did not clearly
identify healthy or unhealthy patterns.
Elsewhere in sub-Saharan Africa the association of some dietary patterns and CVD risk
factors have been reported. A study of more than 1000 adults (15-65 years) in
Ouagadougou, Burkina Faso found a “modern foods” pattern high in meat, cereal,
beans and eggs to be associated with a high prevalence of overweight (OR 1.19, 95%
204 | P a g e
CI: 1.03-1.36) after adjustment for well known confounders (Becquey et al., 2010). In
Cameroon, an explorative cross-sectional study of dietary patterns and hypertension in
541 members of the defence force (mostly men) reported a dietary pattern high in fruit
and vegetables to confer 59% protection against hypertension (Nkondjock & Bizome,
2010). In developed countries, a higher factor score for the “Western” dietary pattern
(high in processed meats, red meat and refined cereals) has been positively associated
with higher BMI, glycated haemoglobin, and lower HDL cholesterol (Kerver et al.,
2003). Some dietary patterns such as the Mediterranean diet (Estruch et al., 2006),
Prudent diet/healthy (Hu et al., 2000; Fung et al., 2001; Berg et al., 2008) have been
inversely associated with CVD and risk factors while other diets including the Western
diet have been associated with increased risk of CVD (Hu et al., 2000; Fung et al.,
2001).
4.11.6 Fruit and vegetable intake and blood pressure
A high intake of vegetables (energy adjusted) was unexpectedly associated with higher
systolic and diastolic blood pressure in men after adjusting for age, but this associati on
was no longer significant after full model adjustments. This unexpected association
may be due to the addition of salt at the table and to vegetables (tomato, pepper,
carrots) particularly when they have been prepared as soup/sauce or with meat dishes
to accompany starchy foods. Therefore, vegetable intake may be acting as a marker for
salt intake.
205 | P a g e
The South Africa Food based dietary guidelines (FBDG) advice on limiting addition of
salt during the cooking processes and at the table (Vorster et al., 2001). Increased
Intake of fruit and vegetables has been reported to appreciably reduce SBP (2.8 mmHg)
and DBP (1.1 mmHg) in DASH (Dietary Approach to Stop Hypertension) randomised
feeding trial (Appel et al., 1997), and in a dietary advice intervention where a decrease
of 2 mmHg and 1.6 mmHg for SBP and DBP respectively was observed over a 6 months
period in Oxfordshire United Kingdom (John et al., 2002).
206 | P a g e
Chapter 5 Diet and risk of disease
Chapter 5
Diet & risk of disease
207 | P a g e
5 Diet and risk of CVD/diabetes/hypertension

5.1 Introduction
This chapter describes the association between diet and risk of CVD/diabetes
/hypertension by comparing participants with no known CVD/diabetes/hypertension
(“healthy”) and participants with reported CVD/diabetes/hypertension (“diseased”). As
shown in figure 5.1, 566 “healthy” participants and 221 “diseased” participants were
recruited. It has been reported that people with diabetes attending clinic in Gaborone
Botswana suffered from co-existing CVD risk factors such as hypertension (Mengesha,
2007) and in this thesis as shown in figure 5.1 more than half of the participants with
diabetes were also hypertensive (59%) and 22% reported having a cardiovascular
condition.
Healthy (566) Diseased (221)
Hypertensive (140)
50
566
51
25
14
35 10 36
CVD (84) Diabetes (111)
Figure 5.1 Distribution of participants by disease status
208 | P a g e
“Diseased” participants were generally older that the “healthy” participants, and in an
attempt to reduce the confounding effect of age, “healthy” participants > 30 years old,
and “diseased” participants younger than 65 years old were selected for the
comparison as shown in figure 5.2. Before including all the “diseased” in one group, a
comparison of socio-demographic characteristics, physical measurements, nutrient
intakes and biochemical indicators between the three diseased groups
(CVD/diabetes/hypertension) was undertaken.
209 | P a g e
50
45
43.3 Healthy Diseased
40
35
Selected “diseased”
30
Percentage (%)
25 26.4
Selected "healthy"
20
15 16.6
14.1 14.1
12.6
10 11.5
8.5 9
7.6 7 7.5 7
5
4.5 4.1 2.7 2 1.1 0.5 0
0
18-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70
Age group (years)
210 | P a g e
Figure 5.2 Age distribution of participants by disease status

The aim of this chapter was to assess the effect of diet on risk of disease. Socio-
demographic characteristics, anthropometric measurements, nutrient intakes and
biochemical indicators between “diseased” participants and “healthy” participants
were compared to determine potential confounders.
Comparison of the characteristics between the 3 diseased groups was carried out using
Pearson’s Chi-squared test for categorical variables, and ANOVA for normally
distributed variables, and the Kruskal-Wallis test for skewed variables. Comparison
between “healthy” and “diseased” participants was carried out using Pearson’s Chi-
squared test for categorical variables, and the independent t-test for normally
distributed variables, while Mann-Whitney U-test was used for skewed variables.
Multiple logistic regression analysis was used to estimate odds ratios (95% CI) for the
risk of disease (CVD/diabetes/hypertension) associated with intake of nutrients
[tertiles of total fat, SFA and PUFA as %E, fibre as g/1000 kcal and alcohol (categories of
alcohol use]. To preserve power, logistic regression analysis was carried out for both
sexes combined. Models were adjusted for age (years) and sex (model 1); age (years),
sex, education (no formal, primary, secondary and tertiary), income (<600, ≥600<3000,
≥3000<10000, ≥10000 BWP), family history of CVD (yes/no), smoking (smoker/non-
smoker), alcohol use (drinker/former drinker/non drinker) (model 2); the variables
included in model 2 plus BMI (continuous), WC (continuous), physical activity level
(continuous), triglycerides (continuous), glucose (continuous) (model 3); the variables
211 | P a g e
included in model 3 plus total, SFA (%E), total fat (%E) and fibre (g/1000kcal) (model 4).
Nutrient intakes were mutually adjusted for as recommended by Willett (1998). The
test for trend was carried out by entering the variable for tertile of nutrients intake in
the logistic regression model as a continuous variable.
Pearson’s correlation coefficients (table 5.1) were calculated to assess the associations
between the different for nutrients, physical and biochemical measurements for all
participants. Spearman correlation coefficients were produced for associations with
energy and triglycerides which were skewed. Only covariates that were not highly
correlated (r <0.4) in each group were adjusted for in the models.
212 | P a g e
Table 5.1 Pearson correlation coefficient for nutrients, physical and biochemical measurements
Nutrients
Total energy CHO (%E) Protein (%E) Total fat SFA (%E) PUFA (%E) Fibre Alcohol (g)
intake (kcal) (%E) (g/1000kcal)
Total energy† 1 -0.20** 0.013 0.13** 0.15** 0.03 -0.23** 0.45**
intake (kcal)
CHO (%E) - 1 -0.53** -0.74** -0.69 -0.33** 0.26** -0.20** - -
Protein (%E) - - 1 0.37** 0.34** 0.01 0.09 -0.15** - -
Total fat (%E) - - - 1 0.87** 0.60** -0.31** -0.17** - -
SFA (%E) - - - - 1 0.18** -0.41** -0.11** - -
PUFA (%E) - - - - - 1 -0.03 -0.15** - -
Fibre - - - - - - 1 -0.19** - -
(g/1000kcal)
Alcohol (g)
Physical measurements and biochemical indicators
(mmol/L) Total Triglycerides LDL HDL Random BMI WC (cm) WHR SBP DBP
2
cholesterol cholesterol cholesterol glucose (kg/m ) (mmHg) (mmHg)
Total chole sterol 1 0.44 0.91** 0.27** 0.23** - - - - -

Triglycerides† 1 0.41** -0.21** 0.32** - - - -
LDL cholesterol - - 1 -0.07 0.24** - - - - -
HDLcholesterol - - - 1 -0.08* - - - - -
Glucose - - - - 1 - - - - -
2
BMI (kg/m ) - - - - - 1 0.78** 0.17** 0.32** 0.35
WC (cm) - - - - - - 1 0.57** 0.37** 0.35**
WHR - - - - - - - 1 0.29** 0.22**
SBP (mmHg) - - - - - - - - 1 0.79*
DBP (mmHg) - - - - - - - - - 1
213 | P a g e
†
Spearman correlation, ** (p<0.01), * (p<0.05),
5.3 Comparison between participants with CVD, hypertension or
diabetes
Table 5.2 shows the socio-demographic characteristics of 121 of the 221 “diseased”
participants who had either CVD only, hypertension only, or diabetes only. Participants
with 2 conditions or more were excluded in this analysis to avoid overlap of subjects
between groups.
Participants with CVD only were more than 10 years younger on average than
participants with diabetes or hypertension only (p<0.001). More than 60% of
participants with diabetes only were married compared to 28% and 14.3% of
participants with hypertension only and CVD only respectively (p=0.004). The
proportion of individuals who had attained tertiary education was lower in participants
with CVD only (11.4%) compared to the other two groups (p=0.001) and the proportion
owning livestock was highest in the CVD only group (p=0.049).
Participants with CVD only had a lower WHR than those with hypertension only or
diabetes only (p=0.006). As expected, SBP (p=0.001) and DBP (p=0.002) were highest in
participants with hypertension only compared to those with CVD only or diabetes only.
BMI, WC, body fat percent and physical activity level were not significantly different
between the three groups (table 5.3).
214 | P a g e
Total energy intake was lower in the diabetes group compared to the other two groups
(p<0.001), and fibre intake was highest in the diabetes only group (p=0.01). The rest of
the dietary variables (%E) were not statistically significantly different between diseased
groups (table 5.4).
Total cholesterol, triglycerides and LDL cholesterol were significantly lower in
participants with CVD only compared with the other two groups. As expected, blood
random glucose level was highest in the diabetes only group compared to the CVD only
and hypertension only groups (p<0.001). Atherogenic index was also highest in the
diabetes only group (p=0.009). HDL cholesterol was lowest in the diabetes only group
(p=0.033) (table 5.5).
In tables 5.3 and 5.5 the diabetes only group had <30 participants, therefore the results
should be interpreted with caution.
215 | P a g e
Table 5.2 Socio-demographic characteristics of participants by disease status: n (%)
CVD only HPT only Diabetes only p value†

n 35 50 36
Male 6 (17.1) 12 (24.0) 12 (33.3) 0.283
Female 29 (82.9) 38 (76.0) 24 (66.7)
Age mean SD (years) 37.6 (15.7) 49.0 (10.5) 49.9 (12.8) <0.001*
Marital status
Married 5 (14.3) 14 (28.0) 22 (61.1) 0.004
Education
No formal education 3 (8.6) 4 (8.0) 3 (8.3)
Primary education 4 (11.4) 25 (50.0) 13 (36.1)
Secondary education 24 (68.6) 10 (20.0) 13 (36.1) 0.001
Tertiary education 4 (11.4) 11 (22.0) 7 (19.4)
Income/month (BWP ~£0.1)
<600 6 (17.1) 5 (10.0) 1 (2.8)
≥600<3000 20 (57.1) 23 (47.0) 15 (42.9)
≥3000<10000 9 (25.7) 18 (36.7) 13 (37.1) 0.056
≥10000 0 3 (6.0) 6 (16.7)
Ownership
House 18 (51.4) 31 (62.0) 22 (64.7) 0.408
Livestock 21 (60.0) 24 (51.1) 11 (31.4) 0.049
Car 7 (20.0) 17 (34.0) 17 (47.2) 0.053
Computer 10 (28.6) 20 (42.6) 14 (38.9) 0.420
Television 27 (77.1) 36 (76.6) 31 (86.1) 0.513
Cellular phone 32 (94.1) 44 (93.6) 35 (97.2) 0.741
Family history of CVD 27 (77.1) 35 (70.0) 20 (55.6) 0.137
Smoking 2 (5.7) 5 (11.3) 2 (6.1) 0.580
Current alcohol drinker 7 (21.2) 13 (26.5) 7 (19.4) 0.225
†Pearson Chi-squared test, * A NOVA
216 | P a g e
Table 5.3 Mean (SD) physical and anthropometric measurements by disease status
1
CVD only HPT only Diabetes only p value
n 35 43 29
2
BMI (kg/m ) 28.17 (6.67) 29.75 (7.11) 27.54 (4.96 0.303
Body fat (%) 36.09 (9.67) 40.32 (9.54) 35.03 (11.01) 0.081
WC (cm) 84.53 (12.99) 90.93 (14.62) 88.69 (10.96) 0.116
Waist-to-hip r atio 0.77 (0.08) 0.83 (0.10) 0.85 (0.08) 0.006
SBP (mmHg) 127.16 (15.67) 143.90 (24.86) 127.71 (20.80) 0.001
DBP*(mmHg) 78.75 89.75 77.25 0.002†
(74.13, 87.13) (78.38, 97.63) (72.13, 86.00)
Physical activity level 1.48 (0.55) 1.30 (0.57) 1.45 (0.56) 0.299
1
ANOVA, *Median (p25, p75), †Kruskal-Wallis non parametric test
Table 5.4 Daily mean (SD) macronutrient intake (%E) of participants by disease status
1
CVD only HPT only Diabetes only P value
n 35 47 36
Energy (kcal) * 3989 3035 2548 <0.001†
(2376, 4410) (2285, 4083) (1950, 38.64)
Carbohydr ate 60.31 (12.27) 59.60 (11.56) 61.19 (14.19) 0.849
Protein 13.39 (4.01) 14.41 (3.57) 15.40 (4.47) 0.113
Total fat 25.09 (6.35) 24.31 (8.68) 23.09 (9.81) 0.602
SFA 7.46 (2.32) 7.28 (3.04) 6.09 (3.09) 0.089
PUFA* 6.73 5.63 5.34 0.430†
(4.40, 8.84) (4.19, 7.43) (4.09, 7.81)
Fibre (g/1000kcal)* 7.74 8.40 9.82 0.010†
(5.69, 9.51) (6.97, 9.70) (7.76, 12.30)
1
ANOVA, *Median (p25, P75), †Kruskal-Wallis non parametric test
217 | P a g e
Table 5.5 Mean (SD) biochemical indicators (mmol/L) by disease status
1
CVD only HPT only Diabetes only P value
n 33 43 29
Total chole sterol 4.23 (0.77) 4.76 (0.83) 4.69 (0.98) 0.021
Triglycerides* 0.76 (0.52, 1.11) 1.20 (0.96, 1.85) 1.18 (0.80, 2.41) <0.001†
LDL cholesterol 2.88 (0.69) 3.35 (0.73) 3.31 (0.88) 0.022
HDL cholesterol 1.15 (0.26) 1.15 (0.35) 0.99 (0.20) 0.033
Random glucose * 4.43 (4.01, 5.01) 4.81 (4.50, 5.19) 6.74 (5.60, 10.67) <0.001†
Atherogenic index* 3.54 (3.01, 4.61) 4.15 (3.42, 5.34) 4.66 (3.70, 5.62) 0.009†
1
ANOVA, *median (P25, P75), †Kruskal-Wallis non para metric test
Comparison of socio-demographic characteristics, physical and anthropometric
measurements as well as lipids, lipoproteins and glucose were generally different
between the three disease groups, however dietary variables which are the key
variables of interest were only significantly different for total energy intake and dietary
fibre intake. The three groups were combined into one group and appropriate
adjustments were made in the model to account for possible confounding.
Nevertheless, caution must be exercised when interpreting the results. Botswana has a
dearth of data on diet and CVD, and this work is in part hypothesis generation.
218 | P a g e
5.4 Characteristics of “healthy” and “diseased” participants
5.4.1 Socio-demographic characteristics
Table 5.6 shows the socio-demographic characteristics of the “healthy” and “diseased”
participants. At least two thirds of participants in the “healthy” and “diseased” groups
were female. Although an attempt was made to reduce the difference in age between
the two groups, the “diseased” participants’ median age remained relatively higher
than the “healthy” participants’ median age (50 versus 37 years, p<0.001). “Diseased”
participants were also more likely to earn higher income (p=0.002), own a house
(p<0.001) or car (p=0.010). Conversely, “healthy” participants were more likely to be
current drinkers (p=0.014), and have a secondary education or higher (p=0.002). No
significant differences between groups were found in participants’ smoking habits,
ownership of livestock, cellular phone and television.
Only 45.6% of the “diseased” female participants reported having had a menstruation
in the last 12 months compared to 80.4% of the “healthy” participants (p<0.001). 13
(11.6%) healthy participants reported being post menopausal compared to 52 (39.1%)
diseased participants.
219 | P a g e
Table 5.6 Socio-demographic characteristics of participants by disease status: n (%)
“Healthy” “Diseased” P-value*

n 171 185
Gender Male 52 (30.4) 43 (23.2) 0.127
Female 119 (69.6) 142 (76.8)
†
Age 37 (33, 43) 50 (37, 56) <0.001
Median(P25, P75)
Marital status Married 50 (29.4) 67 (36.2) 0.156
Education No formal education 5 (2.9) 21 (11.4)
Primary education 38 (22.4) 64 (21.6)
Secondary education 85 (50.0) 70 (37.8) 0.002
Tertiary education 42 (24.7) 30 (16.2)
Income/month‡ <600 32 (19.5) 17 (11.0)
≥600<3000 86 (52.4) 81 (46.8)
≥3000<10000 42 (25.6) 65 (38.0) 0.002
≥10000 4 (2.4) 14 (7.9)
Ownership House 79 (46.7) 123 (68.0) <0.001
Livestock 78 (46.4) 98 (54.4) 0.207
Car 44 (26.2) 71 (39.0) 0.010
Computer 40 (24.1) 60 (33.3) 0.058
Television 133 (81.1) 151 (83.4) 0.259
Cellular phone 162 (95.9) 176 (96.7) 0.675
Smoking Current smoker 20 (12.6) 12 (7.0) 0.088
Alcohol use Current drinker 55 (32.9) 35 (19.2) 0.014
Former drinker 35 (21.0) 44 (24.2)
History of CVD 96 (56.1) 119 (64.3) 0.115
@
Menstruation Yes 90 (80.4) 61 (45.9) <0.001
Post-menopause 13 (11.6) 52 (39.1) <0.001
@
*Pearson’s Chi-squared test, †Mann-Whitney U test, ‡ Inco me (1BWP=£0.1), women only
220 | P a g e
5.4.2 Physical and anthropometric measurements
Table 5.7 describes the physical and anthropometric measurements by disease status.
“Diseased” participants had significantly higher BMI (p=0.003), body fat percent (p=0.007),
WC (p<0.001), WHR (p=0.001), SBP (p<0.001) and DBP (p<0.001) compared to “healthy”
participants. Only physical activity level was significantly higher in “healthy” participants
compared to “diseased” (p<0.001).
Figure 5.3 and figure 5.4 show that “diseased” participants were significantly more likely
to have a high WC, high WHR, high SBP and high DBP. More than 29% and more than 35%
of “healthy” and “diseased” participants were overweight and obese respectively.
However, proportions of overweight and obese participants were not significantly
different between the two groups.
Table 5.7 Mean (SD) physical and anthropometric measurements by disease status
“Healthy” “Diseased” P-value†

n
2
BMI (kg/m ) 26.77 (5.13) 28.75 (6.67) 0.003
Body fat % 35.05 (9.48) 38.22 (10.47) 0.007
Waist circumference (cm) 84.06 (10.81) 89.38 (13.58) <0.001
Waist-to-hip r atio (WHR) 0.80 (0.09) 0.84 (0.11) 0.001
Blood pressure (mmHg)
SBP 126.51 (15.55) 137.87 (23.39) <0.001
DBP 78.74 (11.12) 86.12 (16.52) <0.001
Physical activity level 1.62 (0.58) 1.41 (0.56) <0.001
†Independent t-test
221 | P a g e
45 "Healthy"
p=0.004 p<0.001
"diseased"
40 41.9
NS 41.2
NS
Proportion of participants (%) 35
35.3 35.9
30
29.3 29.3
25 26.6
20
20.1
15
10
0
Overweight (BMI≥25 Obese (BMI≥30 kg/m2) High WC (women high WHR (women ≥
and < 30 kg/m2) ≥88cm, men ≥102cm) 0.85, men ≥0.90)
Figure 5.3 Proportion of participants being overweight and obese
40 P<0.001 P<0.001
35 "Healthy"
Proportion of participants (%)
35.9 35.2
30 "Diseased"
25
20
18.7
15 15.7
10
5
0
SBP ≥ 140 mmHg SBP ≥ 90 mmHg
High Blood Pressure
Figure 5.4 Proportion of participants with hypertension

222 | P a g e
5.4.3 Macronutrient intake
Table 5.8 shows that energy intake was significantly higher in “healthy” compared to
“diseased” participants (p<0.001). “Healthy” participants also obtained a significantly
higher proportion of their energy intake from total fat (p=0.002) and SFA (p<0.001)
compared to the “diseased” participants. However the “diseased” group had a
significantly higher proportion of energy from carbohydrates (p=0.006) and a hi gher fibre
intake (p=0.014).
Table 5.8 Daily mean (SD) nutrient intake (%E) of participants by disease status
n 169 181
Energy (kcal) 3903 3035 <0.001*
median (P25, P75) (2819, 5407) (2272, 4193)
Carbohydr ates (%E) 56.89 (12.34) 60.44 (11.71) 0.006
Fibre (g/1000kcal) 8.63 (2.70) 9.42 (3.30) 0.014
Protein (%E) 14.99 (3.75) 14.80 (3.82) 0.636
Total fat (%E) 26.10 (8.06) 23.33 (8.27) 0.002
SFA (%E) 8.04 (3.07) 6.84 (2.98) <0.001
PUFA (%E) 5.94 (4.40, 7.65) 5.54 (3.99, 7.46) 0.082*

Median (IQR)
†Independent sample t-test, *Mann-Whitney U test
223 | P a g e
5.4.4 Lipids, lipoproteins and glucose
Table 5.9 describes blood lipid, lipoprotein and glucose levels of participants. Generally
average levels of lipids, lipoproteins and glucose were within the optimal levels in both
groups, except for HDL cholesterol which was below the optimal level (<1.00 mmol/L for
men and <1.30 mmol/L for women) for men in both groups (NCEP, 2002). Average
concentrations of total cholesterol, triglycerides, and LDL cholesterol and glucose were
significantly higher in “diseased” participants compared to “healthy” participants. Median
HDL and atherogenic index were not significantly different between the groups.
Table 5.9 Median (P25, P75) biochemical indicators (mmol/L) by disease status
n 147 166
Total chole sterol mean (SD) 4.15 (1.00) 4.64 (0.88) <0.001*
Triglycerides 1.01 (0.70, 1.43) 1.21 (0.81, 1.99) <0.001
LDL cholesterol mean (SD) 2.81 (0.93) 3.21 (0.79) <0.001*
HDL cholesterol 1.00 (0.84, 1.24) 1.09 (0.90, 1.26) 0.070
Random glucose 4.67 (4.21, 5.30) 5.13 (4.52, 6.74) <0.001
Atherogenic index 3.90 (3.19, 5.04) 4.13 (3.49, 5.17) 0.085
† Mann-Whitney U test,*Independent sample t-test
224 | P a g e
As shown in figure 5.5 the “diseased” group had a significantly higher proportion of
participants with high triglycerides (p<0.001) and glucose (p<0.001) compared to
“healthy” participants. The percentage of participants with high total cholesterol, LDL
cholesterol, atherogenic index and low HDL cholesterol were not significantly different
between the two groups.
80
- High total chol ≥ 6.2 mmol/L NS
- High triglycerides ≥ 1.7 mmol/L 73.9 Healthy Diseased
70
71.9
- High LDL chol ≥ 3.8 mmol/L
60 - Low HDL chol (Men < 1.0 mmol/L,

women < 1.3 mmol/L)
50 -Impaired glucose (≥ 6.1 mmol/L)

Percentage (%)
- High risk atherogenic index > 5.0
40 p<0.001
p<0.001 NS
30 33.4 32.3 31.4
NS
26
20 21.3
15.8
10 13.7
NS
7.6
3.4 3.8
0
Total chol Triglycerides LDL chol Low HDL chol Glucose Atherogenic
index
Figure 5.5 Proportion of participants with high lipids, lipoproteins and glucose
225 | P a g e
5.5 Risk of CVD/diabetes/hypertension by dietary factors
Tables 5.10 – 5.15 show the crude and adjusted odds ratios (95% CI) for risk of disease
associated with thirds of intake of carbohydrate, fibre, total fat, SFA, PUFA (as %E),
PUFA:SFA ratio and alcohol respectively. The ORs for all variables in each model are also
shown.
Overall, age was associated with a significant increased risk of disease in models 1, 2 and 3
but not in model 4. Family history of CVD, income (>BWP3000), blood glucose and systolic
blood pressure were also significantly associated with increased risk of disease. Only
tertiary education was found to be consistently and significantly associated with an
inverse risk of disease in models 3 and 4.
5.5.1 Carbohydrate intake
Table 5.10 shows the association between carbohydrate intake and the risk of
CVD/diabetes/hypertension. Compared to the low tertile of carbohydrate intake, the
highest tertile was significantly associated with disease risk in the crude model (OR 2.03;
95% CI, 1.21-3.40, p for trend=0.006) and model 3 (OR 3.53; 95% CI, 1.45-8.60, p for
trend=0.006), but not in models 2 or 4.
5.5.2 Fibre intake
Fibre was not associated with disease risk in any of the models (table 5.11)
226 | P a g e
5.5.3 Total fat intake
Table 5.12 shows the ORs for disease risk associated with intake of total fat (%E). A
significantly lower risk was found in the categories of medium (OR 0.47, 95% CI, 0.28-0.78)
and high intake (OR 0.44, 95 CI, 0.26-0.75,) relative to the low intake category in the
unadjusted model (p for trend=0.002). Similarly ORs were significant in models 2 and 3 as
well as in the highest intake of model 4, but ORs were not significant in age and sex
adjusted model.
5.5.4 Saturated fatty acid intake
In table 5.13 a high SFA intake was unexpectedly significantly associated with reduced
disease risk compared to low intake in the crude model, models 1, model 2 and model 3
(OR 0.26, 95% CI, 0.11-0.62, p for trend=0.002 in model 3). In model 4 the association was
no longer significant.
5.5.5 Polyunsaturated fatty acids intake
PUFA was not significantly associated disease risk (table 5.14) in any model.
5.5.6 Polyunsaturated fatty acids and saturated fatty acid ratio (PUFA:SFA ratio)
The PUFA:SFA ratio was not associated with disease risk (table 5.15) in any model.
5.5.7 Alcohol use
As shown in table 5.16 a high intake of alcohol was associated with reduced disease risk in
the unadjusted model, but this association was not significant after further adjustments.
227 | P a g e
Table 5.10 OR (95% CI) for risk of CVD/diabetes/hypertension according to carbohydrate (%E) intake
Tertile (mean) n Crude Model 1* Model 2* Model 3 * Model 4 *

(healthy/Diseased)
Low (44.91) 56/77 1.00 1.00 1.00 1.00 1.00
Medium (57.69) 53/49 1.10 (0.65-1.88) 0.90 (0.51-1.61) 1.16 (0.60-2.26) 1.51 (0.64-3.57) 1.21 (0.44-3.39)
High (70.99) 60/55 2.03 (1.21-3.40) 1.42 (0.81-2.53) 1.91 (0.98-3.72) 3.53 (1.45-8.60) 2.44 (0.67-8.89)
p for trend 0.006 0.194 0.052 0.005 0.147
Age (years) 1.07 (1.05-1.10) 1.06 (1.03-1.09) 1.01 (0.97-1.06) 1.01 (0.97-1.06)
Sex (male) 0.76 (0.45-1.28) 1.16 (0.59-2.29) 1.14 (0.47-2.76) 1.19 (0.48-2.96)
Family History (CVD) 1.90 (1.10-3.29) 2.04 (1.01-4.12) 2.09 (1.03-4.25)

Income (3000-10000 BWP) 3.90 (1.61-9.45) 3.10 (1.06-9.07) 2.84 (0.95-8.48)
Income (>10000 BWP) 9.07(1.94-42.48) 5.80 (0.89-37.97) 5.42 (0.81-36.14)
Education (tertiary) 0.18 (0.05-0.69) 0.19 (0.03-1.05) 0.19 (0.03-1.06)
Current smokers 1.68 (0.82-3.15) 1.48 (0.44-5.02) 1.60 (0.47-5.46)
Former Drinker 1.61 (0.82-3.15) 1.45 (0.62-3.37) 1.42 (0.61-3.32)
Current drinker 1.10 (0.53-2.27) 1.15 (0.45-2.91) 1.06 (0.41-2.74)

Physical activity level 0.54 (0.28-1.03) 0.56 (0.29-1.07)
WHR (high) 1.57 (0.67-3.66) 1.41 (0.60-3.35)
2
BMI (≥25<30kg/m ) 1.67 (0.75-3.72) 1.61 (0.72-3.61)
BMI (≥30kg/m2) 0.64 (0.28-1.49) 0.64 (0.27-1.48)
Triglycerides (mmol/L) 1.00 (0.66-1.52) 1.01 (0.66-1.53)
Glucose (mmol/L) 1.53 (1.14-2.05) 1.55 (1.15-2.10)
SBP (mmHg) 1.05 (1.02-1.07) 1.05 (1.02-1.07)
Fibre (g/1000 kcal) 0.96 (0.84-1.09)
Total fat (%E) 0.98 (0.91-1.04)
*Subjects with missing data excluded, Model 1, adjusted for age and sex
228 | P a g e
Model 2, adjusted for age, sex, Family history of CVD, income, education, smoking, alcohol
Model 3, adjusted for age, sex, Family history of CVD, income, smoking education, alcohol, Physical activity, BMI, WHR, glucose, triglyceride, SBP
Model 4, adjusted for age, sex, Family history of CVD, income, smoking education, alcohol, Physical activity, BMI, WHR, glucose, triglyceride, SBP, Fibre, total fat
Table 5.11 OR (95% CI) for risk of CVD/diabetes/hypertension according to fibre (g/1000kcal) intake

(healthy/dise ased)
Low (5.82) 52/45 1.00 1.00 1.00 1.00 1.00
Medium (8.23) 61/57 1.08 (0.63-1.85) 0.71 (0.40-1.29) 0.77 (0.40-1.48) 0.75 (0.33-1.70) 0.69 (0.29-1.57)
High (12.08) 56/79 1.63 (0.96-2.76) 1.05 (0.59-1.87) 1.10 (0.56-2.12) 0.77 (0.33-1.78) 0.56 (0.23-1.36)
P for trend 0.058 0.739 0.716 0.568 0.205
Age (years) 1.08 (1.05-1.10) 1.06 (1.03-1.09) 1.02 (0.98-1.07) 1.02 (0.97-1.06)
Sex (male) 1.39 (0.83-2.32) 1.31 (0.68-2.53) 1.43 (0.62-3.31) 1.36 (0.58-3.21)
Family history (CVD) 1.90 (1.10-3.28) 2.20 (1.10-4.39) 2.20 (1.09-4.45)

Income(3000-10000 BWP) 3.67(1.51-8.88) 2.40 (0.85-6.77) 2.37 (0.81-6.81)
Income (>10000 BWP) 8.97(1.94-41.59) 5.81 (0.89-37.92) 5.71 (0.83-39.49)
Current smoker 1.58 (0.60-4.18) 1.47 (0.44-4.94) 1.76 (0.52-5.94)
Former drinker 1.56 (0.80-3.06) 1.51 (0.66-3.44) 1.62 (0.68-3.83)
Current drinker 1.01 (0.49-2.06) 0.89 (0.36-2.20) 1.00 (0.38-2.64)
Physical activity 0.59 (0.32-1.10) 0.59 (0.32-1.11)

WHR (high) 1.49 (0.65-3.43) 1.48 (0.64-3.41)
BMI (25<30 kg/m2) 1.58 (0.72-3.49) 1.54 (0.69-3.44)
BMI (≥30 kg/m2) 0.62 (0.27-1.41) 0.62 (0.27-1.42)
Triglycerides (mmol/L) 1.05 (0.69-1.59) 1.08 (0.71-1.64)
Glucose (mmol/L) 1.43 (1.09-1.89) 1.53 (1.13-2.09)
SBP (mmHg) 1.04 (1.02-1.06) 1.04 (1.02-1.07)
Total fat (%E) 0.94 (0.90-0.98)
229 | P a g e
Model 4, adjusted for age, sex, Family history of CVD, income, smoking education, alcohol, Physical activity, BMI, WHR, glucose, triglyceride, SBP, total fat
Table 5.12 OR (95% CI) for risk of CVD/diabetes/hypertension according to total fat (%E) intake

(healthy/dise ased)
Low (16.41) 47/83 1.00 1.00 1.00 1.00 1.00
Medium (25.24) 64/53 0.47 (0.28-0.78) 0.62 (0.36-1.08) 0.41 (0.22-0.79) 0.48 (0.21-1.08) 0.46 (0.20-1.04)
High (34.43) 58/45 0.44 (0.26-0.75) 0.61 (0.34-1.07) 0.44 (0.23-0.86) 0.37 (0.16-0.85) 0.35 (0.15-0.81)
P for trend 0.002 0.077 0.015 0.017 0.429
Age (years) 1.07 (1.05-1.10) 1.05 (1.02-1.09) 1.01 (0.97-1.05) 1.01 (0.96-1.05)
Sex (male) 1.38 (0.83-2.29) 1.31 (0.67-2.55) 1.41 (0.60-3.31) 1.43 (0.60-3.38)

Income (3000-10000 BWP) 4.18 (1.69-10.34) 2.95 (1.01-8.60) 2.90 (0.98-8.61)
Income(>10000 BWP) 11.35 (2.32-55.55) 7.38(1.06-51.39) 7.37 (1.2-53.11)
Current smoker 1.83 (0.68-4.90) 1.56 (0.47-5.20) 1.73 (0.51-5.90)
Former drinker 1.69 (0.85-3.34) 1.52 (0.65-3.57) 1.71 (0.71-4.12)
Current drinker 1.06 (0.51-2.19) 0.99 (0.39-2.52) 1.35 (0.48-3.77)

WHR (high) 1.41 (0.61-3.24) 1.33 (0.57-3.14)
BMI (25<30 kg/m2) 1.57 (0.70-3.49) 1.48 (0.66-3.33)
BMI (≥30 kg/m2) 0.64 (0.28-1.47) 0.59 (0.25-1.37)
Triglyceride(mmol/L) 1.06 (0.70-1.61) 1.04 (0.68-1.57)
Glucose (mmol/L) 1.52 (1.12-2.05) 1.51 (1.11-2.05)
SBP (mmHg) 1.04 (1.02-1.06) 1.05 (1.02-1.07)
Fibre (g/1000kcal) 0.96 (0.84-1.09)
230 | P a g e
Model 4, adjusted for age, sex, Family history of CVD, income, smoking education, alcohol, Physical activity, BMI, WHR, glucose, triglyceride, SBP, dietary fibre
Table 5.13 OR (95% CI) for risk of CVD/diabetes/hypertension by SFA (%E) intake

(healthy/dise ased)
Low (4.49) 53/86 1.00 1.00 1.00 1.00 1.00
Medium (7.76) 62/57 0.57 (0.35-0.93) 0.74 (0.43-1.26) 0.50 (0.27-0.93) 0.48 (0.22-1.05) 0.52 (0.20-1.39)
High (11.39) 54/38 0.43 (0.25-0.74) 0.52 (0.30-0.93) 0.47 (0.24-0.91) 0.26 (0.11-0.62) 0.28 (0.07-1.10)
P for trend 0.002 0.026 0.018 0.002 0.068
Age (years) 1.07 (1.05-1.10) 1.06 (1.02-1.09) 1.01 (0.96-1.05) 1.01 (0.96-1.05)
Sex (male) 1.39 (0.83-2.32) 1.33 (0.68-2.61) 1.51 (0.64-3.59) 1.46 (0.61-3.50)

Income (3000-10000 BWP) 3.98 (1.63-9.74) 2.77 (0.95-8.08) 2.53 (0.86-7.42)
Income(>10000 BWP) 10.40 (2.18-49.53) 7.38(1.08-50.63) 6.84 (0.99-47.40)
Current smoker 1.72 (0.64-4.58) 1.64 (0.49-5.50) 1.80 (0.54-6.06)
former drinker 1.59 (0.81-3.12) 1.45 (0.62-3.35) 1.57 (0.67-3.68)
Current drinker 1.05 (0.51-2.17) 1.02 (0.40-2.59) 1.10 (0.43-2.84)

WHR (high) 1.45 (0.63-3.38) 1.35 (0.57-3.16)
BMI (25<30 kg/m2) 1.73 (0.77-3.90) 1.67 (0.73-3.80)
BMI (≥30 kg/m2) 0.66 (0.28-1.52) 0.61 (0.26-1.43)
Glucose (mmol/L) 1.54 (1.13-2.09) 1.53 (1.02-2.10)
SBP (mmHg) 1.05 (1.02-1.07) 1.04 (1.02-1.07)
Total fat (%E) 0.99 (0.93-1.07)
Fibre (g/1000kcal) 0.99 (0.98-1.01)
*Subjects with missing data excluded, , Model 1, adjusted for age and sex
231 | P a g e
Model 4, adjusted for age, sex, Family history of CVD, income, smoking education, alcohol, Physical activity, BMI, WHR, glucose, t riglyceride, SBP, total fat, dietary fibre
Table 5.14 OR (95% CI) for risk of CVD/diabetes/hypertension by PUFA (%E) intake

(healthy/diseased)
Low (3.76) 47/83 1.00 1.00 1.00 1.00 1.00
Medium (5.96) 64/53 0.67 (0.40-1.13) 0.78 (0.44-1.37) 0.84 (0.44-1.60) 0.81 (0.37-1.78) 1.07 (0.47-2.47)
High (9.62) 58/45 0.67 (0.40-1.10) 0.94 (0.54-1.62) 0.76 (0.41-1.42) 0.69 (0.32-1.49) 1.30 (0.52-3.28)
P for trend 0.107 0.801 0. 391 0.340 0.581
Age (years) 1.08 (1.05-1.10) 1.06 (1.03-1.10) 1.02 (0.97-1.06) 1.01 (0.97-1.06)
Sex (male) 1.44 (0.86-2.40) 1.30 (0.67-2.53) 1.44 (0.63-3.32) 1.39 (0.59-3.31)

Income (3000-10000 BWP) 3.59 (1.50-8.62) 2.56 (0.91-7.18) 2.43 (0.85-7.00)
Income(>10000 BWP) 9.16 (1.97-42.61) 6.16 (0.95-40.21) 6.14 (0.90-41.73)
Smoking (smoker) 1.68 (0.63-4.46) 1.49 (0.45-5.00) 1.66 (0.50-5.55)
former drinker 1.59 (0.82-3.10) 1.54 (0.68-3.51) 1.62 (0.69-3.79)
Current drinker 0.98 (0.48-1.99) 0.93 (0.38-2.28) 1.08 (0.42-2.75)

WHR (high) 1.44 (0.63-3.28) 1.34 (0.58-3.11)
BMI (25<30 kg/m2) 1.62 (0.74-3.57) 1.56 (0.70-3.49)
BMI (≥30 kg/m2) 0.62 (0.27-1.42) 0.58 (0.25-1.35)
Glucose (mmol/L) 1.47 (1.10-1.96) 1.53 (1.12-2.09)
SBP (mmHg) 1.04 (1.02-1.06) 1.04 (1.02-1.07)
Total fat (%E) 0.94 (0.89-0.99)
Fibre (g/1000kcal) 0.99 (0.98-1.01)
232 | P a g e
Model 4, adjusted for age, sex, Family history of CVD, income, smoking education, alcohol, Physical activity, BMI, WHR, glucose, triglyceride, SBP, total fat, dietary fibre
Table 5.15 OR (95% CI) for risk of CVD/diabetes/hypertension by PUFA:SFA ratio

(healthy /diseased)
Low (0.46) 50/54 1.00 1.00 1.00 1.00 1.00
Medium (0.81) 64/46 0.67 (0.39-1.14) 0.72 (0.40-1.29) 0.72 (0.38-1.40) 0.68 (0.29-1.55) 0.64 (0.28-1.49)
High (1.55) 55/81 1.36 (0.40-1.10) 1.47 (0.84-2.56) 1.45 (0.77-2.73) 1.59 (0.72-3.59) 1.38 (0.61-3.13)
P for trend 0.175 0.127 0. 199 0.220 0.414
Age (years) 1.08 (1.05-1.10) 1.07 (1.03-1.10) 1.02 (0.98-1.07) 1.02 (0.97-1.06)
Sex (male) 1.35 (0.81-2.26) 1.28 (0.65-2.49) 1.46 (0.62-3.41) 1.39 (0.59-3.31)

Income (3000-10000 BWP) 3.68 (1.53-8.82) 2.59 (0.92-7.31) 2.50 (0.86-7.27)
Income(>10000 BWP) 9.05 (1.92-42.54) 6.90 (1.04-45.52) 6.43 (0.92-44.75)
Smoking (smoker) 1.49 (0.56-3.96) 1.34 (0.40-4.99) 1.59 (0.47-5.34)
former drinker 1.62 (0.83-3.17) 1.69 (0.74-3.86) 1.72 (0.73-4.06)
Current drinker 1.01 (0.49-2.08) 0.98 (0.39-2.47) 1.11 (0.43-2.87)

WHR (high) 1.39 (0.60-3.23) 1.26 (0.53-2.95)
BMI (25<30 kg/m2) 1.77 (0.79-3.96) 1.68 (0.75-3.77)
BMI (≥30 kg/m2) 0.61 (0.27-1.40) 0.57 (0.24-1.34)
Glucose (mmol/L) 1.43 (1.09-1.88) 1.52 (1.12-2.05)
SBP (mmHg) 1.04 (1.02-1.06) 1.04 (1.02-1.07)
Total fat (%E) 0.95 (0.91-0.99)
Fibre (g/1000kcal) 0.99 (0.98-1.01)
233 | P a g e
Model 4, adjusted for age, sex, Family history of CVD, income, smoking education, alcohol, Physical activity, BMI, WHR, glucose, triglyceride, SBP, total fat, dietary fibre
Table 5.16 OR (95% CI) for the likelihood of having CVD/diabetes/hypertension by alcohol use
Tertile (median) n Crude Model 1* Model 2* Model 3 * Model 4 *

(healthy/diseased)
Non user (0g) 81/106 1.00 1.00 1.00 1.00 1.00
Moderate intake (0.5g) 69/67 0.74 (0.48-1.16) 0.86 (0.54-1.39) 0.80 (0.46-1.40) 0.55 (0.28-1.09) 0.51 (0.25-1.04)
High intake (60.0g) 19/8 0.32 (0.13-0.77) 0.53 (0.20-1.36) 0.35 (0.10-1.23) 0.67 (0.18-2.45) 0.59 (0.16-2.17)
P for trend 0.011 0.213 0.174 0.156 0.114
Age (years) 1.07 (1.05-1.10) 1.05 (1.02-1.09) 1.02 (0.98-1.06) 1.01 (0.97-1.05)
Sex (male) 1.25 (0.73-2.13) 1.11 (0.59-2.08) 1.22 (0.56-2.69) 1.10 (0.48-2.48)

Income (3000-10000 BWP) 3.40 (1.45-7.97) 2.73 (1.00 -7.46) 2.64 (0.94-7.43)
Income(>10000 BWP) 9.44 (2.00-44.62) 5.21 (0.80-33.82) 5.39 (0.78-37.04)
Smoking (smoker) 1.50 (0.59-3.86) 1.56 (0.48-5.05) 1.86 (0.57-6.04)

WHR (high) 1.38 (0.62-3.07) 1.29 (0.58-2.91)
BMI (25<30 kg/m2) 1.54 (0.71-3.34) 1.50 (0.68-3.31)
BMI (≥30 kg/m2) 0.57 (0.26-1.27) 0.54 (0.24-1.23)
Glucose (mmol/L) 1.44 (1.10-1.90) 1.56 (1.15-2.12)
SBP (mmHg) 1.04 (1.02-1.06) 1.05 (1.02-1.07)
Total fat (%E) 0.94 (0.90-0.98)
Fibre (g/1000kcal) 0.99 (0.98-1.01)
234 | P a g e
Model 4, adjusted for age, sex, Family history of CVD, income, smoking education, Physical activity, BMI, WHR, glucose, t riglyceride, SBP, total fat, dietary fibre
Diet and risk of disease
Key findings
 Risk of disease was significantly higher in participants with a high carbohydrate intake versus a low intake
(OR 3.53, 95% CI, 1.45-8.60, p for trend =0.005) in model 3, but after adjustment for nutrients the
association was no longer significant.
 Compared to individuals with a low total fat intake, those with a high total fat intake had a significantly
reduced risk of disease (OR 0.37, 95% CI, 0.16-0.85, p for trend =0.017) model 3, but the association was
not significant in model 4.
 Risk of disease was significantly reduced in participants with a high SFA intake (OR 0.26, 95% CI, 0.11 -0.62,
p for trend =0.002) in model 3 but after adjustment for nutrients the association was no longer significant.
 High alcohol intake had a significant protective effect in the unadjusted model (OR 0.32, 95% CI, 0.13-
0.77, p for trend =0.011), but after adjusting for potential confounders the association was no longer
significant.
 There was no significant association between PUFA, PUFA:SFA rati o or fibre intake and the risk of disease
235 | P a g e
5.6 Discussion
In this study no significant association between diet and disease risk was found after
adjusting for other nutrients. However, a higher intake of carbohydrate (%E) was
associated with increased risk of having CVD/diabetes/hypertension, and higher intake
of total fat, SFA and alcohol intake was associated with a reduced disease risk.
Plausible explanations for the unexpected inverse association between SFA intake and
disease risk could include (i) the “diseased” group had a significantly lower (p<0.001)
SFA (%E) intake than the “healthy” group probably due to standard health advice on
fat intake reduction given after diagnosis, (ii) the mean levels of SFA intake 11.30 %E in
the highest category is almost within the recommended intake of < 10%E (WHO, 1990)
and intakes were not heterogenous enough to produce the expected associations (iii)
other unknown mechanisms may be responsible for the unexpected protective effect.
Notwithstanding, in this study higher intake of SFA was associated with significantly
higher LDL cholesterol in “healthy” men (chapter 3). A recent meta-analysis of
prospective epidemiologic studies (16 CHD end points and 8 stroke end points)
involving more than 300 000 participants aged ~30-89 years showed no evidence for an
association between SFA intake and risk of CHD (RR 1.07: 95 CI: 0.96 – 1.19) or CVD
(RR 1.00: 95 CI: 0.89 – 1.11) (Siri-Tarino et al., 2010).
In this study we also found that “diseased” participants had a significantly higher intake
of carbohydrate compared to “healthy” participants (p=0.006). Given that “diseased”
participants had a significantly higher intake of carbohydrates and significantly lower
236 | P a g e
intake of total fat compared to “healthy” participants, it is therefore possibly that
“diseased” participants may be replacing fat with carbohydrate in their diet. Recently,
Kuipers et al (2011) in their review of current scientific data concluded that SFA
accumulates under eucaloric conditions in normal weight subjects who consume a
carbohydrate rich diet with high glycaemic index and under hypocaloric condition in
subjects with metabolic syndrome and non-alcoholic fatty liver disease who consume
carbohydrate rich diets. The lack of favourable effect of replacing fat with
carbohydrate in the diet on CVD health have been shown in cohort studies and
controlled trials; eucaloric replacement of SFA with carbohydrates has been cautioned
in the prevention of heart disease since it does not improve atherogenic index
(total:HDL cholesterol ratio) (Mensink et al., 2003). Jacobsen and colleagues (2009) in
their pooled analysis of 11 cohort studies from America and Europe reported an overall
significant direct association between substituting carbohydrates for fat and the ris k of
coronary events (HR: 1.07; 95% CI: 1.01-1.14), however they caution that the effect of
substitution of carbohydrates may vary depending on the type and quality consumed.
In the same review replacing SFA with PUFA was found to be significantly protective
against coronary events and coronary death.
It is well known that high fibre intake, particularly soluble fibre from grains and
legumes improves blood cholesterol and possibly confers protection against CVD
(Flight & Clifton, 2006, Bazzano, 2008). However, in this study higher intake of fibre did
not confer any protection against being “diseased”. In the Nurses’ Health study a
prospective cohort of more than 68 000 women, Wolk et al (1999) found that after
237 | P a g e
multivariate analysis women in the highest compared to the lowest category of long
term cereal fibre intake had a 34% lower risk of total CHD, but intake of fibre from fruit
and vegetables was not appreciably associated with risk of total CHD. This study
investigated total fibre intake and did not look at different types of fibre, therefore
given the relatively high absolute fibre intake it is possible that most of the dietary
fibre may have been provided from non cereal sources, especially given that staple
cereals sorghum and maize which are usually refined were commonly consumed by
only 37% and 54% of all participants respectively.
In this comparison of “diseased” and “healthy” participants, there was a non-significant
trend of a protective effect of alcohol. Only 8 “diseased” participants were taking >24g
of alcohol per day therefore there was insufficient power to detect any significant OR.
Self reported alcohol drinking has been reported to improve some and not other CVD
risk factors. A meta-analysis of 42 experimental studies carried out between 1968 and
1998 that assessed effects of moderate alcohol consumption on biological markers
predicted a 24% reduction in risk of CHD with an intake of 30g of alcohol/day (Rimm,
1999). Pisa et al (2010) in their prospective urban and rural epidemiology (PURE) study
in South Africa found that self reported drinkers had a favourable lipid profile and BMI
but not blood pressure. An explorative case-control study of the association between
alcohol consumption and coronary artery disease (CAD) in 1476 middle aged men
undergoing angiography in China reported that moderate to heavy drinking increased
risk of CAD, however the study was not representative of the general population and
238 | P a g e
alcohol consumption was self reported; furthermore participants were relatively old
and already susceptible to CAD (Zhou et al., 2010).
239 | P a g e
Chapter 6: General discussion
Chapter 6
General discussion
240 | P a g e
6 Discussion
The main objectives of this thesis were (i) to measure and document CVD risk
factors in adults in Gaborone, Botswana, (ii) to investigate the associations between
diet (nutrients, dietary patterns and food groups) and CVD risk factors, and (iii) to
assess the association between diet and risk of disease
(CVD/hypertension/diabetes). This chapter discusses the strengths and limitations
of the study and summarises the main findings on CVD risk factors and their
associations with diet in Botswana, along with suggestions for future research work
on diet and CVD. Finally, a public health perspective on CVD risk reduction and
prevention is explored.
241 | P a g e
6.1 Strengths
The main strengths of the study are that (i) it is the first study to comprehensively
measure a wide range of CVD risk factors in Botswana with a reasonably large
sample size and (ii) detailed information on risk factors and potential confounders
were collected thus making it possible to measure CVD risk factors and investigate
their association with lifestyle factors.
All measurements in this study were carried out by the author and trained research
assistants. Blood samples were analysed within 24 hours of collection in an
accredited laboratory. All remaining blood samples were also stored for future
analysis. Furthermore, the quality and completeness of data collected was high.
The FFQ used in this study was validated against four 24 hr recalls administered
over 12 months in 79 adults aged 18-75 years. FFQs were administered at the
beginning and end of a one year validation study. Energy adjusted Pearson’s
correlation coefficients between 24 hour recall and FFQ were moderate to good for
most nutrients, ranging from 0.30 for beta-carotene to 0.76 for alcohol.
Misclassification of participants into opposite quartiles was ≤10% in total energy,
CHO, protein total fat, dietary fibre and alcohol although misclassification was
higher than 10% for micronutrients and some mineral intake (Jackson et al .,
submitted). There was also a reasonable classification (weighted Kappa >0.40) of
nutrients into quartiles between the two methods as desired in epidemiological
studies (Masson et al., 2003).
242 | P a g e
6.2 Limitations
While the study had some limitations recognition of these limitations would help
improve the design of future studies. First, this study was a cross-sectional study,
so it was not possible to assess disease incidence or make any causal inference.
Second, the study was undertaken in one city therefore it is possible that it may not
be representative of the general population. Third, while some of the healthy
participants were recruited from the clinics, the study targeted workplace wellness
program and this may have influenced the study results because of their possible
high socio-economic status. Finally, while all efforts were made to recruit across all
age groups the “healthy” participants were generally young compared to the
“diseased” group, although efforts were made to match participants by age.
Therefore, future studies extending from this study will be informed by such
observations and should aim to recruit participants from a range of settings that
would improve the generalisability of findings.
Blood samples collected from participants were non-fasting samples and were
collected at different times of the day. However, it has been reported that levels of
lipids, lipoproteins and apolipoproteins change minimally (not clinically significant)
in response to normal food intake compared to those taken >8hrs since the last
meal in individuals aged 20-95 years in the Copenhagen general population
(Langsted et al., 2008).
Information on dietary intake was collected using a FFQ, a tool that depends on
memory and participants’ perception of portion sizes, therefore the results are
243 | P a g e
likely to be subject to bias (Willet, 1998). To minimise errors in this study, the FFQ
was interview administered by the author and trained research assistants, and
pictures of food and life size examples of utensils such as plates, spoons and cups
were used to help participants estimate their usual intake.
FFQs have a tendency to overestimate dietary intake (Rodriguez et al., 2002;
Masson et al., 2003). Correspondingly, in this study participants appeared to have
overestimated their intakes, however no significant differences in sex, income,
education level, BMI and WC were found between over-reporters and the rest of
the participants. Furthermore, nutrient intakes were ranked in tertiles of %E in all
analyses which has the advantage of adjusting for total energy.
Dietary patterns were assessed using principal components analysis. In the future
other methods of dietary pattern analysis such as factor analysis which can
estimate underlying factors could be explored. Furthermore, score d based diet
score based methods such as the Mediterranean Diet Score which has been
reported to be inversely associated with CVD risk (Panagiotakos et al., 2006) need
to be investigated in African populations.
244 | P a g e
6.3 Major findings
6.3.1 CVD risk factors
6.3.1.1 Obesity
The prevalence of obesity measured by BMI was high in women: 6.9%, 20.6% and
46.7% from the youngest to the oldest age group respectively. In men the
prevalence of obesity was <6% across all age groups and was lowest at 1% in men
aged <25 years. At least 1 in 5 women aged >25 years had abdominal obesity and
for those women aged >35 years almost 1 in 2 were abdominally obese, while only
1 in 10 men aged 25-35 years were abdominally obese. Conversely, at least 15% of
men 35 years or younger were underweight (BMI <18.5 kg/m2). Only 5.9% of
women in the youngest age group were underweight.
The prevalence of overweight and obesity was significantly higher in older and
married participants as well as in participants with only primary education or less
but was not significantly different by income or physical activity level although
physical activity level was significantly lower in women than in men (table 3.27).
These results present a scenario of a double burden of overweight and underweight
in this population comparable to South Africa (Puoane et al., 2002). The high
prevalence of overweight and obesity in women compared to men is also similar to
studies from other countries in Sub-Saharan Africa (Sievo et al., 2006; Sodjinou et
al., 2008; Achidi & Tangoh, 2010; Msyamboza et al., 2011).
245 | P a g e
6.3.1.2 Hypertension
In this relatively healthy group 1 in 4 men and women aged >35 years had high
blood pressure (SBP ≥ 140mmHg or DBP≥ 90mmHg) and at least 1 in 10 men had
high blood pressure. The WHO Botswana steps survey (2007) also reported that
about 1 in 3 people surveyed had raised blood pressure. This raises concern that
there could be a considerable proportion of Batswana with undiagnosed
hypertension possibly predisposing them to risk of stroke.
6.3.1.3 Dyslipidaemia
The prevalence of low HDL cholesterol (<1.0 mmol/L men or < 1.3 mmol/L women)
was most common in women in the oldest group (82.1%), followed by women in
the middle (75.7%) and youngest age (67.7%). In men 41.5%, 46.7% and 48.4% from
the youngest to the oldest age group respectively had low HDL cholesterol, with
significant differences found by sex. 7.4% of the “healthy” participants had high
triglycerides (≥1.7 mmol/L) and men had significantly higher triglycerides compared
to women (12.1% versus 5.0%). Conversely, high LDL cholesterol (≥ 3.8 mmol/L) was
significantly more common in women compared to men (7.6% versus 6.5%). Due to
low HDL cholesterol almost 1 in 3 participants had a high atherogenic index (total
cholesterol/HDL cholesterol) although average levels of blood lipids were all within
normal ranges.
246 | P a g e
6.3.1.4 Diet
The most commonly consumed foods (consumed ≥ 2 times a week by at least 25%
of participants) foods were tomatoes followed by sugar, beef, white and brown
breads, rice, mayonnaise, mealie (maize) meal, carrots and sweets/candy.
The national staple food sorghum was ranked only 22 nd compared to a rank of 2nd in
elderly in Botswana in 1998 (Maruapula & Chapman-Navakofski, 2007). High sugar
and meat and refined starches now feature prominently as commonly consumed
foods. Correspondingly, dietary patterns in this group shows that women followed
a “high sweets” and a “high fruit” pattern while men followed “mixed”
“convenience”, “traditional” and a “high vegetable” diet. Recently Maruapula et al
(2010) reported that overweight and obese adolescents in Botswana consumed
high servings of snacks (high fat/sugar) compared to underweight or normal BMI
adolescents, and were more likely to be resident in town and cities as well as being
of a high socio-economic suggesting a link to westernised lifestyle.
Emergence of overweight and obesity, the co-existence of overweight and
underweight (particularly in men), hypertension, low HDL cholesterol, increasing
blood glucose, a shift from traditional staples to snacks and refined starches and
high physical inactivity in this population are indicative of a population undergoing
rapid nutritional transition (Popkin, 2002; Omran 1971). The WHO global risk report
(2009) ranks all these risk factors in the top 10 risk factors accounting for almost 13
million deaths in middle income countries. While mean levels of blood lipids and
glucose were within optimal levels in this population, the high proportion of
participants with low HDL cholesterol, high atherogenic index as well as high levels
247 | P a g e
of obesity suggests that this population may in the not so distant future experience
a rapid increase in the prevalence of hypertension, diabetes and other CVD.
6.3.2 Association between diet and CVD risk factors
6.3.2.1 Fatty acids
In chapter 4 SFA (% E) intakes was associated with increased LDL cholesterol,
triglycerides, glucose, atherogenic index and BMI, but was associated with
decreased HDL cholesterol even after adjustment for physical activity level, income,
education and smoking history. The strong association between SFA and CVD risk
factors suggest that in this population intake of SFA may be a strong contributor to
CVD risk. However, the intake of SFA was below the recommended intake of <10 %E
(WHO, 1990). Therefore SFA intake of around 8-9 %E in this population may be high
enough to increase CVD risk, especially if the individual SFA are composed of
medium chain SFA such as lauric (12:0) myristic acid (14:0) and palmitic (16:0) acids
which are known to be cholesterolemic (Zock et al., 1994). For example, beef is the
third most commonly consumed food in this population and the most common SFA
in beef is palmitic acid (Ministry of Agriculture, Fisheries and Food, 1998) which has
cholesterol raising properties.
Trans fatty acids were not investigated in this study, however their consumption is
reported to have adverse effects on serum lipids, promote inflammation and cause
endothelial dysfunction (Muzaffarian et al., 2006) and thus the intake of trans fatty
acids and the association with CVD risk factors in this population warrants
investigation.
248 | P a g e
Furthermore, when PUFA intake is low or the ratio of PUFA:SFA is < 1 as is the case
in this population CVD risk factors such as LDL and HDL cholesterol may be
negatively affected (Hayes, 2002). It has also been suggested that “if
cholesterolemic response depends in part on the effect of myristic (14:0) and
linoleic (18:2) acids, it is probable that any other factor (e.g. dietary cholesterol,
genetic variability, anthropometric variable such as obesity) that down-regulates
the LDL cholesterol receptors activity would affect dietary fatty acids threshold”
(Hayes & Khosla, 1992) . Therefore taken together the significant positive linear
association between SFA and blood lipids (LDL cholesterol, triglycerides), the
considerable proportion of overweight and obese particularly in women, and the
possibility of a propensity to rapid weight gain (adipogenic) due a “thrifty
genotype” as suggested by Prentice et al (2005), it is plausible that optimal dietary
intake of SFA may be lower in this population than recommended. Therefore a
reduction in intake of SFA could potentially benefit this population group if it wa s
reduced to <7%E, and PUFA intake increased to up to 10%E as recommended by the
National Cholesterol Education Program’s Adult Treatment panel III (NCEP, 2002).
6.3.2.2 Fibre
Intake of fibre had a linear inverse association with total cholesterol and random
blood glucose after age and sex adjustment, but the association with glucose was
strongest in men after adjustment for conventional risk factors for CVD (p for trend
=0.046). No significant association between fibre intake was found with other CVD
risk factors, possibly because CVD risk factors, in particular blood lipids and
lipoproteins, were on average within normal ranges and the participants were
249 | P a g e
generally healthy. The Baltimore longitudinal study on aging reported significant
inverse associations between whole grain and cereal fibre and physical
measurements (weight, BMI, WC), plasma lipids (total and LDL cholesterols) and
glucose (Newby et al., 2007), however participants were older (>50 years), and a 7-
day diet recall was used to assess intake. Furthermore the CVD status of subjects
was not indicated. A recent meta-analysis of RCTs on the relationship between
wholegrain and CHD risk factors found that total cholesterol (-0.20 mmol/L; 95% CI,
-0.31 to -0.10) and LDL cholesterol (-0.18 mmol/L; 95% CI, -0.28 to -0.09) were
significantly reduced in subjects on oatmeal compared to low cereal fibre controls,
however most of the participants in these studies already had CHD risk factors
(overweight, elevated serum cholesterol, or high blood pressure) (Kelly et al., 2009).
Evidence from observational studies and RCTs also support the idea that soluble
dietary fibre lowers cholesterol and reduces some risk factors for CHD (Bazzano,
2008). In this study the lack of significant association between fibre intake and
other CVD risk factors may be because of fibre composition. Only total fibre was
measured in the study, therefore it was not possible to investigate association of
CVD risk factors with different types of fibre. However it is possible that soluble
fibre from cereals may be low given that most of the cereals (bread, sorghum, rice
and maize meal) have lower soluble fibre or virtually no soluble fibre in the case of
rice (Slavin, 2003), and removal of fibre by milling may offset the health benefits of
whole grain (Harris & Kris-Etherton, 2010). Furthermore other bioactive
components responsible for protective effect such as vitamins, minerals, phytates
250 | P a g e
and phytosterols may have been removed during processing thus further
diminishing the synergistic benefits derived from fibre and its constituents.
In summary people in this population would benefits from:
 Increased soluble fibre intake to about 10-25 g/day
 Consumption of whole grain cereal with minimum processing e.g. rolled
grain instead of milled
 Increased fruit and vegetable intake to at least 5 portions a day
6.3.2.3 Alcohol
Intake of alcohol was inversely associated with LDL cholesterol, atherogenic index,
BMI, and body fat %, but was positively associated with HDL cholesterol in
unadjusted and age and sex adjusted models. After multivariate adjustment the
inverse association was limited to BMI and body fat% and was significant in women
only, and the positive linear association between alcohol intake and HDL
cholesterol was also significant in women only. The positive benefits of moderate
alcohol intake (15g/ 1 unit per day for women and 30g/2units per day for men) on
CVD biomarkers (HDL cholesterol, apolipoprotein, fibrinogen and adiponectin) have
been established (Rimm et al., 1999; Brien et al., 2011) and there is compelling
evidence that different types of beverage (wine, beer and spirits) confer similar
effects on biomarkers (Rimm & Stampfer, 2002). Alcohol intakes of 40g/day or
more have been reported to increase risk of hypertension and stroke (Chen et al.,
2008; Pisa et al., 2010; Ronskey et al., 2011) and irregular heavy drinking may also
increase risk of CVD outcomes (van de Wiel & de Lange, 2008)
251 | P a g e
In this study significantly more men than women drank alcohol, however, there was
a suspiciously low consumption of alcohol for both men and women, possibly due
to under reporting of alcohol intake. While no empirical data exist on alcohol use
and patterns of drinking in Botswana, the recent amendment of the liquor
regulation (Republic of Botswana, 2008) imposing a 30% levy on alcoholic
beverages in an attempt to curb alcohol abuse may be an indication that alcohol
use and abuse in Botswana may be higher than reported in this study. Therefore
recommendations for alcohol intake in this population must be accompanied by
education on the potential harm of alcohol abuse. Nonetheless overwhelming
evidence suggests that moderate alcohol intake (8-15g/day) 1-2units for women
and up to 30g/day (1-3 units) for men is beneficial to health (Brien et al., 2011 and
Ronksley et al., 2011).
6.3.2.4 Dietary patterns
Six dietary patterns were identified, two (high sweet and high fruit) in women and
four (mixed, convenience, traditional and high vegetable) in men. None of these
dietary patterns were associated with CVD risk factors after adjustment for age,
education, income, physical activity level and smoking history. These results were
unexpected and are contrary to other studies in Sub-Saharan Africa, which
investigated the association of dietary patterns and CVD risk factors. For example,
recently Becquey et al (2010) found that an increase of 1 unit of “ modern foods”
score (high in meat, cereal, beans and eggs) to be associated with a high prevalence
of overweight (OR 1.19, 95% CI: 1.03-1.36) in over 1000 adults in Burkina Faso, and
in Cameroon high dietary pattern scores for fruit and vegetables conferred 59%
252 | P a g e
protection against hypertension compared to a low dietary pattern score
(Nkondjock & Bizome, 2010). Possible explanations for the lack of association
between dietary patterns and CVD risk factors in this study include (i) analyses in
this study were based on generally healthy participants with a low CVD risk
however it is possible that the patterns’ association with CVD risk factors may be
different in people with chronic diseases (ii) dietary patterns were difficult to name
due to a variety of foods with high factor loading, for example, “high sweets”
pattern had sweets as the major foods, but meat and vegetable also featured in the
pattern too (iii) Possible over estimation of nutrients intake by FFQ could have
affected the association between dietary patterns and CVD risk factors and (iv)
dietary patterns consider overall diet rather than individual nutrients, and thus take
into account possible complicated interactions among nutrients. However,
“because there are many potential differences in nutrients between dietary
patterns, this approach cannot be specific about the particular nutrients
responsible for the observed differences in disease risk, and thus it may not be very
informative about biological relations between dietary components and disease
risk” (Hu et al., 2000).
6.3.2.5 Fruit and vegetables
Intake of fruit and vegetables has been shown to significantly reduce both SBP and
DBP in a randomised feeding trial (Appel et al., 1997) and in a dietary advice
intervention (John et al., 2002). However, in this study an unexpected positive
association between energy adjusted intake of vegetables and blood pressure (SBP
& DBP) was found. It is conceivable that this relationship may be due to the
253 | P a g e
common practice of adding salt at the table or adding salt to vegetables when
preparing soup to accompany starchy foods (Gokah & Gumpo, 2010). Therefore,
vegetable intake, particularly those vegetables (tomato, pepper, carrots) that are
used to prepare soup may be acting as a marker for salt intake. However we did not
adjust for salt intake in the analysis because we could not estimate it from the FFQ
and no urine was collected in this study. Furthermore, other unknown mechanisms
may be responsible for the association. Therefore, the association between salt and
vegetable intake merit further investigation.
6.3.3 Diet and disease
In chapter 5 the associations between diet and diagnosed disease was explored.
This was carried out to generate hypotheses in a population that has limited data
on CVD risk factors and their relationship with lifestyle factors. The “diseased”
individuals were on average 13 years older than the “healthy” individuals, and the
disease status was self reported. Furthermore the “diseased” were not newly
diagnosed. Therefore this analysis is limited.
High intake of carbohydrates (%E) was associated with increased disease risk and
high intakes of total fat, SFA and alcohol were found to be suggestive of protection
against disease. However, all odds ratios were not significant after mutual
adjustment for nutrients suggesting that the nutrient of interest may have been
acting as a marker for another nutrient. In this study however “diseased”
participants were older, their time of diagnosis of disease was not verifiable, they
had significantly lower energy intake and fat intake, but higher carbohydrates
(possibly from sugar) intake compared to “healthy” participants. Therefore it is

254 | P a g e
conceivable that following diagnosis the “diseased” participants were advised on
dietary changes and weight loss. It is also possible that “diseased” participants may
have replaced fat with carbohydrates.
While a recent meta-analysis of prospective studies suggested that there is no
evidence for an association between SFA intake and increased risk of CVD (Siri-
Tarino et al., 2010), a recent pooled analysis of cohort studies found that
substitution of fat with carbohydrates (particularly high GI) is associated with
increased risk of coronary events (Jakobsen et al., 2009).
In this study fibre intake was not associated with disease risk. Current evidence
from RCT suggest that only fibre from whole grain oats is protective against CVD
risk factors (Tighe et al., 2010; Kelly et al., 2009) and prospective studies support
evidence that high intake of dietary fibre particularly soluble fibre from cereal and
fruit reduce the risk of CHD (Pereira et al., 2004; Bazzano et al., 2003; Wolk et al.,
1999). Soluble fibre decreases serum cholesterol by binding to bile acids resulting in
increased bile secretion thus depleting bile liver stores and increasing uptake from
the circulation thereby protecting against CHD (Tungland & Meyer, 2002). In this
study only data on total dietary fibre was available, therefore different type of fib re
were not investigated. Furthermore, more than 90% of the population in Botswana
consumed less than the recommended amount of fruit and vegetable (5portion a
day) (WHO Botswana steps survey, 2007). Additionally, given that the fibre
intake/1000kcal in both groups was lower than 10-15g/100kcal as recommended
(National Research Council, 2005) it is possible that fibre intake relative to total
255 | P a g e
energy intake may be low, or the insoluble fibre: soluble fibre intake ratio may not
be in the recommended 3:1 ratio (Shikany & White, 2000).
A protective effect of moderate alcohol consumption on biological markers
associated with risk of CHD (Brien et al., 2011) and cardiovascular outcome
(Ronksley et al., 2011; Rimm et al., 1999) is well-established, however there are
concerns about the blood pressure raising effect of alcohol even at a moderate
level of consumption (Pisa et al., 2010). In this study intake of alcohol >24 g/day
was protective in the unadjusted model, but the effect was no longer significant
after adjustment for age and other confounders. This was possibly due to the low
power of the study particularly in the highest alcohol intake group in which there
were only 8 participants.
In this population it is possible that that there are a myriad of factors which
influence the relationship between diet and health in Botswana. For example some
Batswana cultural norms encourage women to have big bodies because it is
perceived as beautiful. It is also common practice in Botswana that women cook
and as a result they are possibly more likely to over eat. Furthermore, some people
hold beliefs that being fat is a sign of wealth and health (especially since the advent
of HIV/AIDS). It is therefore possible that these perceptions may be one of the
barriers to fully understanding the dynamic relationship between diet and health in
Botswana. Given that there is limited research on the influence of cultural norms
and other factors on health and diet, in the future further investigation is
warranted.
256 | P a g e
6.4 Future research work focus
In this thesis a first attempt has been made to measure and document diet and risk
factors for CVD and associations between them. However, the results from this
thesis give only a snapshot of risk factors in Gaborone, and therefore it would be
unwise to make inferences to the general population of Botswana. Therefore there
is a need to undertake national studies that are prospective in design in order
monitor levels of these risk factors, and to investigate their association with lifestyle
and environmental factors.
In the short term further investigation of the causes of the high prevalence of
obesity and low HDL cholesterol in women and high blood pressure in this
population need to be undertaken. There is also a need to investigate intakes of
different types of carbohydrates (refined and unrefined), fibre (soluble and
insoluble), PUFA (n-3 and n-6) and SFA (lauric, myristic, palmitic and stearic) so that
(i) the intake in the population can be quantified, and (ii) the associations with CVD
risk factors can be investigated.
Concurrently other biological markers of CVD such as C-reactive protein (CRP),
interleukin-6 (IL-6), homocysteine, and fibrinogen levels and the association with
diet need investigation because they may potentially highlight why some
established dietary risk factors are not showing relationships with CVD risk factors
in this population.
The high prevalence of low HDL may be indicative of familial hypercholesterolemia
due to a possible defective Apoliporotein A1 (Dastani et al., 2006). Apolipoproteins
257 | P a g e
are integral to the structural integrity of lipoproteins and enzyme regulation. They
have been reported to be better predictors of CVD, for example upper versus lower
tertiles of apolipoprotein B predicted an increased risk of CVD, and apolipoprotein
were found to have a higher ability than LDL cholesterol in predicting CVD (Benn et
al., 2007). Therefore documentation of apolipoproteins in Batswana will be
instrumental in mapping CVD risk factors.
Furthermore, non-invasive measurements of arterial stiffness such as pulse wave
analysis (PWA) and ankle brachial pressure index (ABPI) could add value towards
understanding risk of CVD in Botswana.
While the use of dietary patterns for investigating associations between diet and
CVD risk factors has advantages over focussing on individual nutrients, no
independent associations were found between dietary patterns and CVD risk
factors using principal components analysis. Therefore future work may need to
explore other methods of dietary pattern analysis such as factor analysis which can
estimate underlying factors, furthermore, scored based diet methods such as the
Mediterranean Diet Score which has been reported to be inversely associated with
CVD risk (Panagiotakos et al., 2006) need to be investigated in this African
population.
The research work on this thesis is the first undertaken after validation of the FFQ,
and given the relatively high estimated energy intake in this study there may be a
need for re-evaluation and adaptation of the questionnaire to reflect intakes based
on local food composition tables rather than on the South African food composition
tables.
258 | P a g e
Furthermore, understanding of variability and susceptibility of individuals to obesity
by identifying candidate genes for “thrifty phenotype” in Botswana may in the
future pave way to better interventions against CVD and other nutrition related non
communicable diseases as had been suggested by Prentice et al (2005).
In the medium and long term, prospective studies with a nationally representative
sample size would help better understand the interplay between CVD risk factors
and lifestyle and environment.
There has been only one case-control study on risk factors associated with CVD
(myocardial infarction) in Botswana (Yusuf et al, 2004) and the author is not aware
of any trial or intervention of any nature on the association between diet and CVD
or CVD risk factors in Botswana. Therefore in the future there is a need to explore
well-designed nutrition based intervention trials. For example, nutrition
intervention trials on the benefits of indigenous foods such as mophane worm
(imbrassia belina) which is high in protein and unsaturated fat (Speth, 2010),
morama nut (Tylosema esculentum, fabaceae) high in MUFA and PUFA (Bower et
al., 1988; Jackson et al., 2010) and Marula fruit (sclerocarya birrea) high in vitamin C
(Wehmeyer, 1966;) and others would in the future provide valuable critical
information on how best to address CVD using food based approaches in Botswana
and the region.
A recent cross-sectional study of 179 patients with cardiomegaly referred for
echocahardiography in Botswana reported a strong association between HIV
infection and cardiomyopathy and pericanditis (Schwatz et al., 2012). Given the
high prevalence rate of HIV in Botswana and the subsequent high coverage of anti
259 | P a g e
retroviral medication (UNAIDS/WHO, 2008) which has been reported to be
associated with increased risk of CVD in other countries (DAD study group, 2003),
there is need to also explore the benefits of a healthy diet in people who are on
antiretroviral medication and CVD risk factors in Botswana. (Barrios et al., 2002;
Turcinov et al., 2005; Tsiodras et al., 2009).
Migration to developed parts of countries (rural-urban migration) and to developed
countries in a quest for a better life has been ongoing the world over, and Sub-
Saharan Africa is no exception. Available evidence supports an increase in CVD risk
factors in the migrant communities (Luke et al., 2001; Forouhi & Sattar, 2006).
These migrations continue to raise questions about the aetiology of CVD and how
Africans respond to lifestyle change in relation to CVD and other non communicable
diseases. Therefore, there is a need to undertake collaborative regional studies in
sub Saharan Africa and the rest of the world to better understand CVD and its
association with lifestyle and the environment. Findings from studies such as global
The INTERHEART and THUSA study in South Africa have given insights into CVD risk
factors and would be helpful in planning multi centre studies on CVD in the region.
An attempt lead by the School of Medicine at the University of Botswana is also
underway to establish a regional study to investigate how genetic factors are
associated with hypertension in the South African Development Community (SADC)
(Mokone George, personal communication). A repeat of the WHO Steps surveys
would also be worthwhile in the surveillance of CVD risk factors in Botswana and
other countries in the region.
260 | P a g e
6.5 Public health perspective
It is evident that Botswana is experiencing an emergence of risk factors associated
with CVD and other non-communicable diseases. These risk factors include
overweight and obesity, low HDL cholesterol, high blood pressure, elevated glucose
levels, physical inactivity, consumption of refined food and consumption of foods
high in fat.
The World Health Assembly (2008) action plan for the global strategy for the
prevention and control of noncommunicable diseases (NCD) 2008-2013 reaffirms
its commitment to a global strategy for the prevention and control of NCDs to
reduce premature mortality and improve quality of life through implementation of
tobacco control, and strategies on diet, physical activity, and evidence based
intervention to reduce public health problems caused by harmful use of alcohol
(WHA, 2008). In Botswana the introduction of acts and regulations such as (i)
control of smoking act of 1992 (Botswana government, 1992) which prohibited
smoking in public spaces as well as advertising, and (ii) the recent amendment of
the liquor regulation (Republic of Botswana, 2008) which imposes a 30% levy on
alcoholic beverages will hopefully be instrumental in realising the global strategy
for the prevention and control of NCD. The national plan of action for nutrition
(2005-2010), which aimed to enhance multi-sectoral collaboration and cooperation
in improving the nutrition situation in Botswana has also stimulated interest in diet
and CVD, and work described in this thesis was in part encouraged by the plan.
261 | P a g e
Because CVD is the leading cause of death in most countries, the relation of diet to
CVD should figure prominently in dietary recommendations (Astrup et al., 2011).
Therefore, there is a need to build comprehensive evidence-based food-based
dietary guidelines as recommended by WHO (WHO, 2010) with specific reference
and recommendations to CVD and other non communicable diseases.
Furthermore, given the high literacy rate and availability of different media (cellular
phone, radio, television and computers) of communication as reported in this
thesis, government and non-governmental organisations need to exploit these
resources to get the nutrition messages through to the public. Simple messages
such as “five a day” could be sent via text messages to promote consumption of
fruit and vegetables, and physical activity could be encouraged by providing bicycle
routes on the roads and other recreational facilities in communities and schools as
well as encouraging people to participate in sports.
The government, through effective enforcement of laws, policies and
intensification of education, should continue to encourage the food industry and
food retailer to help market healthier foods to Batswana. In addition, there is also a
need to (i) elevate nutrition education to a core subject in the schools’ curriculum
(ii) start nutrition education early and continue with it throughout the first 12 years
of school and (iii) encourage and support more young people to train as dieticians
and nutritionists. This in turn would help improve better understanding of nutrition
at an early age and hopefully reverse the imminent CVD epidemic in Botswana.
262 | P a g e
6.6 Conclusion
The prevalence of low HDL cholesterol levels, overweight and obesity (particularly in
women) is high in the population of younger adults in Gaborone. High levels of
triglycerides, LDL cholesterol, high blood pressure and glucose levels are also of
concern as they were recorded in at least 5% of the ‘healthy’ participants. Estimated
energy intake is also high in this group although overestimation by the FFQ used is
suspected. The 10 most commonly consumed foods were tomatoes, sugar, beef, rice,
white bread, brown bread, mayonnaise, mealie meal, carrots and sweets/candy with
sorghum, the traditional staple, ranking only 22 nd.
Total fat and SFA intake had positive linear associations with LDL cholesterol,
atherogenic index and BMI, and had an inverse association with HDL cholesterol, even
though the mean intakes were below recommended levels as a percentage of energy.
Unexpectedly PUFA and fibre intakes were only inversely associated with glucose levels
possibly due to non-fasting blood samples. Alcohol intake was inversely associated with
body fat % and BMI, and was positively associated with HDL cholesterol. Unlike studies
in other populations in Africa and elsewhere, dietary patterns were not associated with
CVD risk factors after adjustment for confounding variables. Also unexpectedly,
vegetable intake was found to be positively associated with blood pressure.
Further research is required to investigate diet and CVD risk factors prospectively in a
nationally representative sample in Botswana. Furthermore, public health policy
should intensify prevention strategies for CVD and other non communicable diseases
263 | P a g e
by promoting healthy eating (through use of food-based dietary guidelines), physical
activity, and monitoring CVD risk factors through a surveillance system to reduce the
burden of CVD in Botswana in the future.
“The only way we are going to reduce disease, is to go backward to the diets and
lifestyles of our ancestors.” - Denis Burkitt
264 | P a g e
References
References
265 | P a g e
References
A
Abegunde, DO., Mathers, CD., Adam, T., Ortergon, M., Strong, K. 2007, "The burden
and costs of chronic diseases in low-income and middle-income countries ", Lancet,
vol. 370, pp. 1929-1938.
Achidi, EA & Tangoh, DA. 2008, "Risk assessment of cardiovascular disease among staff
of the University of Buea, South western Cameroon", Journal of Public Health and
Epidemiology, vol. 2, no. 9, pp. 251-261.
Abubakari, A.R., Lauder, W., Agyemang, C., Jones, M., Kirk, A. & Bhopal, R.S. 2008,
"Prevalence and time trends in obesity among adult West African populations: a
meta-analysis", Obesity Reviews, vol. 9, no. 4, pp. 297-311.
Addo, J., Smeeth, L. & Leon, D.A. 2007, "Hypertension In Sub-Saharan Africa: A
Systematic Review", Hypertension, vol. 50, no. 6, pp. 1012-1018.
Agarwal, D.P. 2002, "Cardio-protective effects of light–moderate consumption of

alcohol: a review of putative mechanisms", Alcohol and Alcoholism, vol. 37, no. 5,
pp. 409-415.
Ambrose, J.A. & Barua, R.S. 2004, "The pathophysiology of cigarette smoking and
cardiovascular disease: An update", Journal of the American College of Cardiology,
vol. 43, no. 10, pp. 1731-1737.
Anand, S.S., Islam, S., Rosengren, A., Franzosi, M.G., Steyn, K., Yusufali, A.H., Keltai, M.,
Diaz, R., Rangarajan, S., Yusuf, S. & INTERHEART Investigators 2008, "Risk factors for
myocardial infarction in women and men: insights from the INTERHEART study",
European heart journal, vol. 29, no. 7, pp. 932-940.
Appel, L.J., Moore, T.J., Obarzanek, E., Vollmer, W.M., Svetkey, L.P., Sacks, F.M., Bray,
G.A., Vogt, T.M., Cutler, J.A., Windhauser, M.M., Lin, P., Karanja, N., Simons -
Morton, D., McCullough, M., Swain, J., Steele, P., Evans, M.A., Miller, E.R. & Harsha,
D.W. 1997, "A Clinical Trial of the Effects of Dietary Patterns on Blood Pressure",
New England Journal of Medicine, vol. 336, no. 16, pp. 1117-1124.
Aresu, M., Chaudhury, M., Diment, E., Fuller, E. & et al. 2010, Health Survey for
England, NHS information centre for health and social care, Leeds England.
Astrup, A., Dyerberg, J., Elwood, P., Hermansen, K., Hu, F.B., Jakobsen, M.U., Kok, F.J.,
Krauss, R.M., Lecerf, J.M., LeGrand, P., Nestel, P., Risérus, U., Sanders, T., Sinclair,
A., Stender, S., Tholstrup, T. & Willett, W. 2011, "The role of reducing intakes of
266 | P a g e
References
saturated fat in the prevention of cardiovascular disease: where does the evidence
stand in 2010?", The American Journal of Clinical Nutrition, vol. 93, pp. 684-688..
B
Barrios, A., Blanco, F., Garcia-Benayas, T., Gomez-Viera, J.M., de la Cruz, J.J., Soriano, V.
& Gonzalez-Lahoz, J. 2002, "Effect of dietary intervention on highly active
antiretroviral therapy-related dyslipemia", AIDS (London, England), vol. 16, no. 15,
pp. 2079-2081.
Batty, G.D., Shipley, M.J., Jarrett, R.J., Breeze, E., Marmot, M.G. & Davey Smith, G.
2006, "Obesity and overweight in relation to disease-specific mortality in men with
and without existing coronary heart disease in London: the original Whitehall
study", Heart (British Cardiac Society), vol. 92, no. 7, pp. 886-892.
Bazzano, L.A. 2008, "Effects of soluble dietary fibre on low density lipoprotein
cholesterol and coronary heart disease risk", Current Atherosclerosis Reports, vol.
10, pp. 473-477.
Bazzano, L.A., He, J., Ogden, L.G., Loria, C.M., Vupputuri, S., Myers, L. & Whelton, P.K.
2002, "Fruit and vegetable intake and risk of cardiovascular disease in US adults:
the first National Health and Nutrition Examination Survey Epidemiologic Follow-up
Study", American Journal of Clinical Nutrition, vol. 76, no. 1, pp. 93-99.
Bazzano, L.A., He, J., Ogden, L.G., Loria, C.M. & Whelton, P.K. 2003, "Dietary Fiber
Intake and Reduced Risk of Coronary Heart Disease in US Men and Women: The
National Health and Nutrition Examination Survey I Epidemiologic Follow-up
Study", Archives of Internal Medicine, vol. 163, no. 16, pp. 1897-1904.
Becquey, E., Savy, M., Danel, P., Dabire, HB., Tapsoba, S., Martin-Prevel, Y. 2010,
“Dietary patterns of adults living in Ouagadougou and their association with
overweight”, Nutrition Journal, vol. 9, no. 13.
Benn, M., Nordestgaard, B.G., Jensen, G.B. & Tybjærg-Hansen, A. 2007, "Improving
Prediction of Ischemic Cardiovascular Disease in the General Population Using
Apolipoprotein B", Arteriosclerosis, Thrombosis, and Vascular Biology, vol. 27, no. 3,
pp. 661-670.
Berg, C.M., Lappas, G., Strandhagen, E., Wolk, A., Torén, K., Rosengren, A., Aires, N.,
Thelle, D.S. & Lissner, L. 2008, "Food patterns and cardiovascular disease risk
factors: The Swedish INTERGENE research program", The American Journal of
Clinical Nutrition, vol. 88, no. 2, pp. 289-297.
267 | P a g e
References
Berse, B., Brown, L., Van de Water, L., Dvorak, H. & Senger, D. 1992, "Vascular
permeability factor (vascular endothelial growth factor) gene is expressed
differentially in normal tissues, macrophages, and tumors", Molecular biology of
the cell, vol. 3, no. 2, pp. 211-220.
Bhagat, K. 2001, “A situational analysis of the chronic non-communicable diseases in

Botswana”, Report, Ministry of Health, Gaborone, Botswana.
Bienias, JL. 1999 “Under and over reporting in food frequency questionnaire” Rush
Institute for Health and aging
Bjelakovic, G., Nikolova, D., Gluud, L.L., Simonetti, R.G. & Gluud, C. 2007, "Mortality in
Randomized Trials of Antioxidant Supplements for Primary and Secondary
Prevention", The Journal of the American Medical Association, vol. 297, no. 8, pp.
842-857.
Black, AE., Coward, WA., Cole, TJ., Prentice, AM. 1996, “Human energy expenditure in
affluent societies: an analysis of 574 doubly-labelled water measurements”,
European Journal of Clinical Nutrition, vol. 50, pp.72-92.
Botswana STEPS survey, 2007, “Chronic disease risk factors surveillance report”
Ministry of Health, Gaborone, Botswana.
Bourne, LT., Lambert, EV., Steyn, K. 2002, “Where does the black population of South
Africa stand on the nutrition transition?”, Public Health Nutrition, vol.5, no. 1A, pp.
157-162
Bower, N., Hertel, K., Oh, J. & Storey, R. 1988, "Nutritional Evaluation of Marama Bean
(Tylosema esculentum, Fabaceae): Analysis of the Seed ", Economic Botany, vol. 42,
no. 4, pp. 433-540.
Briel, m., Ferreira-Gonzalez, I., You, JJ., Kananicolas, PJ., et al. 2009, "Association
between change in high density lipoprotein cholesterol and cardiovascular disease
morbidity and mortality: systematic review and meta-regression analysis", British
Medical Journal, vol. 338, b92.
Brien, S.E., Ronksley, P.E., Turner, B.J., Mukamal, K.J. & Ghali, W.A. 2011, "Effect of
alcohol consumption on biological markers associated with risk of coronary heart
disease: systematic review and meta-analysis of interventional studies", BMJ, vol.
342, d671.
Bromley, Catherine et al. 2011. The Scottish Health Survey 2010, Vol. 1: Main report.
The Scottish Executive.
268 | P a g e
References
Brown, L., Rosner, B., Willett, W.C. & Sacks, F.M. 1999, "Cholesterol-lowering effect of
dietary fiber: A meta-analysis", Am.J.Clin.Nutr., vol. 69, pp. 30-42.
Brown, B.G., Zhao, X., Chait, A., Fisher, L.D., Cheung, M.C., Morse, J.S., Dowdy, A.A.,
Marino, E.K., Bolson, E.L., Alaupovic, P., Frohlich, J., Serafini, L., Huss-Frechette, E.,
Wang, S., DeAngelis, D., Dodek, A. & Albers, J.J. 2001, "Simvastatin and Niacin,
Antioxidant Vitamins, or the Combination for the Prevention of Coronary Disease",
N Engl J Med, vol. 345, no. 22, pp. 1583-1592.
Buijsse, B., Feskens, E.J.M., Kwape, L., Kok, F.J. & Kromhout, D. 2008, "Both alpha- and
β-Carotene, but Not Tocopherols and Vitamin C, Are Inversely Related to 15-Year
Cardiovascular Mortality in Dutch Elderly Men", The Journal of nutrition, vol. 138,
no. 2, pp. 344-350.
Buijsse, B., Feskens, E.J., Schulze, M.B., Forouhi, N.G., Wareham, N.J., Sharp, S., Palli, D.,
Tognon, G., Halkjaer, J., Tjonneland, A., Jakobsen, M.U., Overvad, K., van der A, D.L.,
Du, H., Sorensen, T.I. & Boeing, H. 2009, "Fruit and vegetable intakes and
subsequent changes in body weight in European populations: results from the
project on Diet, Obesity, and Genes (DiOGenes)", American Journal of Clinical
Nutrition.
C
Cara, L., Dubois, C., Borel, P., Armand, M., Senft, M., Portugal, H., Pauli, A., Bernard, P.
& Lairon, D. 1992, "Effects of oat bran, rice bran, wheat fiber, and wheat germ on
postprandial lipemia in healthy adults", American Journal of Clinical Nutrition, vol.
55, no. 1, pp. 81-88.
Carr, A., Samaras, K., Burton, S., Law, M., Freund, J., Chisholm, D.J. & Cooper, D.A.
1998, "A syndrome of peripheral lipodystrophy, hyperlipidaemia and insulin
resistance in patients receiving HIV protease inhibitors", AIDS (London, England),
vol. 12, no. 7, pp. F51-8.
Carr, A., Samaras, K., Chisholm, D.J. & Cooper, D.A. 1998, "Pathogenesis of HIV-1-
protease inhibitor-associated peripheral lipodystrophy, hyperlipidaemia, and insulin
resistance", The Lancet, vol. 351, no. 9119, pp. 1881-1883.
Cattell, RB. 1966, “The scree test for the number of factors”, Multivariate Behavioral
Research, vol. 1, pp. 245-276.
Central Statistics Office 2009, Botswana AIDS Impact Survey III: 2008, Government
press, Gaborone, Botswana.
269 | P a g e
References
Central Statistics Office 2007, Health statistics: 2005-2006 [Homepage of Central

Statistics Office], [Online]. Available: www.cso.gov [2009, april].
Central Statistics Office 2001, Population projections for Botswana: 2001-2030,

Goverment press, Gaborone, Botswana.
Central Statistics Office. 2004, Botswana AIDS Impact Survey II, Government press,
Gaborone Botswana.
Chen, L. Davey Smith, G., Harbord, R. & Lewis, S. 2008, "Alcohol Intake and Blood
Pressure: A Systematic Review Implementing a Mendelian Randomization
Approach", PLoS Med, vol. 5, no. 3.
Chobanian, A.V., Bakris, G.L., Black, H.R., Cushman, W.C., Green, L.A., Izzo, J.L., Jones,
D.W., Materson, B.J., Oparil, S., Wright, J.T., Roccella, E.J. & and the National High
Blood Pressure Education Program Coordinating Committee 2003, "The Seventh
Report of the Joint National Committee on Prevention, Detection, Evaluation, and
Treatment of High Blood Pressure", JAMA: The Journal of the American Medical
Association, vol. 289, no. 19, pp. 2560-2571.
Clarke, R., Frost, C., Collins, R., Appleby, P. & Peto, R. 1997, "
Dietary lipids and blood cholesterol: quantitative meta-analysis of metabolic ward
studies", British Medical Journal, vol. 314, no. 112.
Clausen, F., Sandberg, E., Ingstad, B. & Hjortdahl, P. 2000, "Morbidity and health care
utilisation among elderly people in Mmankgodi village, Botswana", Journal of
epidemiology and community health, vol. 54, no. 1, pp. 58-63.
Clausen, T., Charlton, K.E., Gobotswang, K.S.M. & Holmboe-Ottesen, G. 2005,

"Predictors of food variety and dietary diversity among older persons in Botswana",
Nutrition, vol. 21, no. 1, pp. 86-95.
Comrey, A. & Lee, H. 1992, A first course in factor analysis Hillsdale, New Jersey.
Cook, N.R., Albert, C.M., Gaziano, J.M., Zaharris, E., MacFadyen, J., Danielson, E.,
Buring, J.E. & Manson, J.E. 2007, "A Randomized Factorial Trial of Vitamins C and E
and Beta Carotene in the Secondary Prevention of Cardiovascular Events in
Women: Results From the Women's Antioxidant Cardiovascular Study", Archives of
Internal Medicine, vol. 167, no. 15, pp. 1610-1618.
Cook, N.R., Obarzanek, E., Cutler, J.A., Buring, J.E., Rexrode, K.M., Kumanyika, S.K.,
Appel, L.J., Whelton, P.K. & for the Trials of Hypertension Prevention Collaborative
Research Group, 2009, "Joint Effects of Sodium and Potassium Intake on
270 | P a g e
References
Subsequent Cardiovascular Disease: The Trials of Hypertension Prevention Follow-

up Study", Archives of Internal Medicine, vol. 169, no. 1, pp. 32-40.
Cooney, M.T., Dudina, A., De Bacquer, D., Wilhelmsen, L., Sans, S., Menotti, A., De
Backer, G., Jousilahti, P., Keil, U., Thomsen, T., Whincup, P. & Graham, I.M. "HDL
cholesterol protects against cardiovascular disease in both genders, at all ages and
at all levels of risk", Atherosclerosis, vol. In Press, Corrected Proof.
Corbett, J., Given, L., Leyland, A., MacGrogor, A., Marryat, L., Miller, M. & Reid, S. 2009,
The Scottish Health Survey 2008, Scottish Government, Edingburgh.
D
DAD Study Group, 2007, "Class of Antiretroviral Drugs and the Risk of Myocardial
Infarction", The New England journal of medicine, vol. 356, no. 17, pp. 1723-1735.
DAD Study Group, 2003, "Combination Antiretroviral Therapy and the Risk of
Myocardial Infarction", The New England journal of medicine, vol. 349, no. 21, pp.
1993-2003.
Dalal, S., Beunza, J.J., Volmink, J., Adebamowo, C., Bajunirwe, F., Njelekela, M.,
Mozaffarian, D., Fawzi, W., Willett, W., Adami, H. & Holmes, M.D. 2011, "Non-
communicable diseases in sub-Saharan Africa: what we know now", International
journal of epidemiology, vol. 40, no. 4, pp. 885-901.
Danaei, G., Finucane, M.M., Lu, Y., Singh, G.M., Cowan, M.J., Paciorek, C.J., Lin, J.K.,
Farzadfar, F., Khang, Y., Stevens, G.A., Rao, M., Ali, M.K., Riley, L.M., Robinson, C.A.
& Ezzati, M. 2011, "National, regional, and global trends in fasting plasma glucose
and diabetes prevalence since 1980: systematic analysis of health examination
surveys and epidemiological studies with 370 country-years and 2·7 million
participants", The Lancet, vol. 378, no. 9785, pp. 31-40.
Dastani, Z., Dangoisse, C., Boucher, B., Desbiens, K., Krimbou, L., Dufour, R., Hegele,
R.A., Pajukanta, P., Engert, J.C., Genest, J. & Marcil, M. 2006, "A novel nonsense
apolipoprotein A-I mutation (apoA-IE136X) causes low HDL cholesterol in French
Canadians", Atherosclerosis, vol. 185, no. 1, pp. 127-136.
Dauchet, L., Amouyel, P., Hercberg, S. & Dallongeville, J. 2006, "Fruit and Vegetable
Consumption and Risk of Coronary Heart Disease: A Meta-Analysis of Cohort
Studies", Journal of Nutrition, vol. 136, no. 10, pp. 2588-2593.
Declaration of Helsinki, 2008, 59th WMA General Assembly, Seoul, Korea.
271 | P a g e
References
de Roos, N.M., Bots, M.L. & Katan, M.B. 2001, "Replacement of Dietary Saturated Fatty
Acids by Trans Fatty Acids Lowers Serum HDL Cholesterol and Impairs Endothelial
Function in Healthy Men and Women", Arteriosclerosis, Thrombosis, and Vascular
Biology, vol. 21, no. 7, pp. 1233-1237.
D'Elia, L., Barba, G., Cappuccio, F.P. & Strazzullo, P. 2011, "Potassium Intake, Stroke,
and Cardiovascular Disease: A Meta-Analysis of Prospective Studies", Journal of the
American College of Cardiology, vol. 57, no. 10, pp. 1210-1219.
Department of Health 2011, 14th Feb 2011-last update, Alcohol advice [Homepage of
Department of Health UK], [Online]. Available:
http://www.dh.gov.uk/en/Publichealth/Alcoholmisuse/DH_125368 [accessed,
October 2011].
Djousse, L., Lee, I., Buring, J.E. & Gaziano, J.M. 2009, "Alcohol Consumption and Risk of
Cardiovascular Disease and Death in Women: Potential Mediating Mechani sms",
Circulation, vol. 120, no. 3, pp. 237-244.
Duda, R.B., Kim, M.P., Darko, R., Adanu, R.M.K., Seffah, J., Anarfi, J.K. & Hill, A.G. 2007,
"Results of women health study of Accra: Assessment of blood pressure in women",
Int J Cardiol, vol. 117, pp. 115-122.
E
Eckel, R.H., Kahn, R., Robertson, R.M. & Rizza, R.A. 2006, "Preventing Cardiovascular
Disease and Diabetes: A Call to Action from the American Diabetes Association and
the American Heart Association", Circulation, vol. 113, no. 25, pp. 2943-2946.
Edmunds, A. 1979, "Idiopathy cardiomyopathy in Botswana", Journal of Tropical

Medicine & Hygiene, vol. 82, no. 1.
Edwards, R., Unwin, N., Mugusi, F., Whiting, D., Rashid, S. & Kissima, J. 2000,
"Hypertension prevalence and care in an urban and rural area of Tanzania", J
Hypertens, vol. 18, pp. 145-52.
Estruch, R., Martinez-Gonzalez, M.A., Corella, D., Salas-Salvado, J., Ruiz-Gutierrez, V.,
Covas, M.I., Fiol, M., Gomez-Gracia, E., Lopez-Sabater, M.C., Vinyoles, E., Aros, F.,
Conde, M., Lahoz, C., Lapetra, J., Saez, G., Ros, E. & for the PREDIMED Study
Investigators*, 2006, "Effects of a Mediterranean-Style Diet on Cardiovascular Risk
Factors: A Randomized Trial", Annals of Internal Medicine, vol. 145, no. 1, pp. 1-11.
272 | P a g e
References
F
FAO/WHO/UNU Expert Consultation. 2004, Human energy requirements: Energy
requirements of adults, UN/WHO/FAO, Rome Italy.
Fields, A. 2009, “Discovering statistics using SPSS” 3 rd Edition SAGE publications London
Flegal, K.M., Carroll, M.D., Ogden, C.L. & Curtin, L.R. 2010, "Prevalence and Trends in
Obesity Among US Adults, 1999-2008", JAMA: The Journal of the American Medical
Flight, I & Clifton, P. 2006, “Cereal grains and legumes in the prevention of coronary
heart disease and stroke: a review of literature”, European Journal of Clinical
Nutrition, Vol. 60, pp. 1145-1159.
Forouhi, NG & Sattar, N. 2006, “CVD risk factors and ethnicity - a homogenous
relationship”, Atheroscler. Suppl, vol. 7, no. 1.
Fox, C.S., Coady, S., Sorlie, P.D., D'Agostino, R.B.,Sr, Pencina, M.J., Vasan, R.S., Meigs,
J.B., Levy, D. & Savage, P.J. 2007, "Increasing Cardiovascular Disease Burden Due to
Diabetes Mellitus: The Framingham Heart Study", Circulation, vol. 115, no. 12, pp.
1544-1550.
Fung, T.T., Rimm, E.B., Spiegelman, D., Rifai, N., Tofler, G.H., Willett, W.C. & Hu, F.B.
2001, "Association between dietary patterns and plasma biomarkers of obesity and
cardiovascular disease risk", The American Journal of Clinical Nutrition, vol. 73, no.
1, pp. 61-67.
Fung, T.T., Willett, W.C., Stampfer, M.J., Manson, J.E. & Hu, F.B. 2001, "Dietary Patterns
and the Risk of Coronary Heart Disease in Women", Archives of Internal Medicine,
vol. 161, no. 15, pp. 1857-1862.
G
Garrido, RA., Semeraro, MB., Temesgen, SM., Simi, MR. 2009, "Metabolic syndrome
and obesity among workers at Kanye Seventh-Day Adventist Hospital, Botswana ",
South African Medical Journal, vol. 99, no. 5.
Garrow, J., James, W. & Ralph, A. 2000, "Human Nutrition and Dietetics", Churchhill
Livingstone, London, 10th edition.
Geissler, C., & Powers, H. 2005, "Human Nutrition ", Elsevier Churchhill Livingstone,
London, 11th edition.
273 | P a g e
References
Gokah, TK & Gumpo, 2010, "Enabling and empowering – the need for an intergrated
approach to address hypertension among African adults ", Health Education
Research, vol. 25, no. 3 pp. 510-518.
Gopane, R., Pisa, P., Vorster, H., Kruger, A. & Margetts, B. 2010, "Relationshins of
alcohol intake with biological health outcomes in an African population in
transition: THUSA study", S Afr J Clin Nutr, vol. 23, no. 3 (supplement).
Gordon, D., Probstfield, J., Garrison, R., Neaton, J., Castelli, W., Knoke, J., Jacobs, D.,Jr,
Bangdiwala, S. & Tyroler, H. 1989, "High-density lipoprotein cholesterol and
cardiovascular disease. Four prospective American studies", Circulation, vol. 79, no.
1, pp. 8-15.
H
Harris, K & Kris-Etherton, PM. 2010, “Effects of Whole Grains on Coronary Heart
Disease Risk”, Current Atherosclerosis Reports, vol. 12 , no. 6, pp 368-376
Harris, W.S., Mozaffarian, D., Rimm, E., Kris-Etherton, P., Rudel, L.L., Appel, L.J., Engler,
M.M., Engler, M.B. & Sacks, F. 2009, "Omega-6 Fatty Acids and Risk for
Cardiovascular Disease: A Science Advisory From the American Heart Association
Nutrition Subcommittee of the Council on Nutrition, Physical Activity, and
Metabolism; Council on Cardiovascular Nursing; and Council on Epidemiology and
Prevention", Circulation, vol. 119, no. 6, pp. 902-907.
Hayes, K. & Khosla, P. 1992, "Dietary fatty acid thresholds and cholesterolemia", The
FASEB Journal, vol. 11(suppl), pp. 394-400.
Hayes, KC. 2002, "Dietary fat and heart health: in search of the ideal fat", Asia Pacific J
Clin Nutr, vol. 6, no. 8, pp. 2600-2607.
Hegsted, D., Ausman, L., Johnson, J. & Dallal, G. 1993, "Dietary fat and serum lipids: an
evaluation of the experimental data [published erratum appears in Am J Clin Nutr
1993 Aug;58(2):245]", The American Journal of Clinical Nutrition, vol. 57, no. 6, pp.
875-883.
Hendricks, K.M., Dong, K.R., Tang, A.M., Ding, B., Spiegelman, D., Woods, M.N. &
Wanke, C.A. 2003, "High-fiber diet in HIV-positive men is associated with lower risk
of developing fat deposition", American Journal of Clinical Nutrition, vol. 78, no. 4,
pp. 790-795.
274 | P a g e
References
Hooper, L., Summerbell, C., Thompson, R., Sills, D., Roberts, F., Moore, H. & Davey
Smith, G. 2011, "Reduced or modified dietary fat for preventing cardiovascular
disease (Review)", The Cochrane Collaboration. The Cochrane Library Published by
John Wiley & Sons, Ltd., no. 7.
Hu, F.B., Manson, J.E. & Willett, W.C. 2001, "Types of Dietary Fat and Risk of Coronary
Heart Disease: A Critical Review", Journal of the American College of Nutrition, vol.
20, no. 1, pp. 5-19.
Hu, FB., Stampfer, MJ., Manson, JE., Rimm, E., Colditz, GA., Rosner, BA., Hennekens,
CH., Willett, WC. 1997, “Dietary fat intake and the risk of coronary heart disease in
women”, New England Journal of Medicine, vol. 337, pp. 1491-9
Hu, F.B., Rimm, E.B., Stampfer, M.J., Ascherio, A., Spiegelman, D. & Willett, W.C. 2000,
"Prospective study of major dietary patterns and risk of coronary heart disease in
men", American Journal of Clinical Nutrition, vol. 72, no. 4, pp. 912-921.
Hung, H., Joshipura, K.J., Jiang, R., Hu, F.B., Hunter, D., Smith-Warner, S.A., Colditz,
G.A., Rosner, B., Spiegelman, D. & Willett, W.C. 2004, "Fruit and Vegetable Intake
and Risk of Major Chronic Disease", JNCI Journal of the National Cancer Institute,
vol. 96, no. 21, pp. 1577-1584.
I
Ingelsson, E., Massaro, J.M., Sutherland, P., Jacques, P.F., Levy, D., D'Agostino, R.B.,
Vasan, R.S. & Robins, S.J. 2009, "Contemporary Trends in Dyslipidemia in the
Framingham Heart Study", Archives of Internal Medicine, vol. 169, no. 3, pp. 279-
286.
Iqbal, R., Anand, S., Ounpuu, S., Islam, S., Zhang, X., Rangarajan, S., Chifamba, J., Al-
Hinai, A., Keltai, M., Yusuf, S. & on behalf of the INTERHEART Study Investigators,
2008, "Dietary Patterns and the Risk of Acute Myocardial Infarction in 52 Countries:
Results of the INTERHEART Study", Circulation, vol. 118, no. 19, pp. 1929-1937.
J
Jackson, J., Duodu, K., Holse, M., de Faria, M., Jordaan, D., Chingwaru, W., Hansen, A.,
Cencic, A., Kandawa-Schultz, M., Mpotokwane, S., Chimwamurombe, P., de Kock, H.
& Minnaar, A. 2010, " The Morama Bean (Tylosema esculentum) A Potential Crop
for Southern Africa. . ", Advances in Food & Nutrition Research, vol. 61, pp. 187-246.
Jackson, M., Walker, S., Cruickshank, JK., Sharms, S., Cade, J., Mbanya, JC., Younger, N.,
Forester, TF., Wilks, R. 2006, " Diet and overweight and obesity in populations of
275 | P a g e
References
African origin: Cameroon, Jamaica and the UK. ", Public Health Nutrition, vol. 10,
pp. 12-130.
Jackson, M. 2010 “Dietary Intakes of adults in Botswana and association with weight
status,” Kanye Botswana.
Jackson, M., Motswagole, BS., Kwape, L., Kobue-Lekalake, R., Mongwaketse, T.,
Mokotedi, M., Rakgangtswana, B., Jackson, J. “Validation and reproducibility of
Food Frequency Questionnaire for use among adults in Botswana,” (submitted)
Jacobs, D.R.,Jr, A, Adachi, H., Mulder, I., Kromhout, D., Menotti, A., Nissinen, A. &
Blackburn, H. 1999, "Twenty-five-Year Follow-up of the seven countries Study",
Arch Intern Med, vol. 159, pp. 733-740.
Jacobson, T.A., Miller, M. & Schaefer, E.J. 2007, "Hypertriglyceridemia and

cardiovascular risk reduction", Clinical therapeutics, vol. 29, no. 5, pp. 763-777.
Jakobsen, M., O’Reilly, E., Heitmann, B., Pereira, M., Balter, K., Fraser, G., Goldbourt,
U., Hallmans, G., Knekt, P., Liu, S., Pietinen, P., Spiegelman, D., Stevens, J., Virtamo,
J. & Willett, W. 2009, "Major types of dietary fat and risk of coronary heart disease:
a pooled analysis of 11 cohort studies", Am J Clin Nutr, vol. 89, no. 1, pp. 425–32.
Johansson, L., Solvoll, K., Bjørneboe, G. & Drevon, C. Am J Clin Nutr 1998;68:266–74.,
"Under- and overreporting of energy intake related to weight status
and lifestyle in a nationwide sample", American Journal of Clinical Nutrition, , no.
68, pp. 266–74.
John, J., Ziebland, S., Yudkin, P., Roe, L. & Neil, H. 2002, "Effects of fruit and vegetable
consumption on plasma antioxidant concentrations and blood pressure: a
randomised controlled trial", The Lancet, vol. 359, no. 9322, pp. 1969-1974.
Joshipura, K.J., Hu, F.B., Manson, J.E., Stampfer, M.J., Rimm, E.B., Speizer, F.E., Colditz,
G., Ascherio, A., Rosner, B., Spiegelman, D. & Willett, W.C. 2001, "The Effect of Fruit
and Vegetable Intake on Risk for Coronary Heart Disease", Annals of Internal
Medicine, vol. 134, no. 12, pp. 1106-1114.
K
Kahn, S.E., Hull, R.L. & Utzschneider, K.M. 2006, "Mechanisms linking obesity to insulin
resistance and type 2 diabetes", Nature, vol. 444, no. 7121, pp. 840-846.
Kannel, W.B. & Sorlie, P. 1979, "Some health benefits of physical activity. The
Framingham Study", Archives of Internal Medicine, vol. 139, no. 8, pp. 857-861.
276 | P a g e
References
Kannel, W.B., Adrienne Cupples, L., Ramaswami, R., Stokes, J., Kreger, B.E. & Higgins,
M. 1991, "Regional obesity and risk of cardiovascular disease; the Framingham
study", Journal of clinical epidemiology, vol. 44, no. 2, pp. 183-190.
Kannel, W.B. & Vasan, R.S. 2009, "Triglycerides as vascular risk factors: New
epidemiological Insights", Curr Opin Cardiol, vol. 24, no. 4, pp. 345-350.
Kaur, S., Cohen, A., Dolor, R., Coffman, C. & Bastian, L. 2004, "The impact of
environmental tobacco smoke on women's risk of dying from heart disease: A
meta-analysis", Journal of Women's Health, vol. 13, no. 8, pp. 888-898.
Keding, G., Msuya, J., Maass, B. & Krawinkel, M. 2011, "Dietary patterns and nutritional
health of women: the nutrition transition in rural Tanzania.", Food Nutr Bull., vol.
32, no. 3, pp. 218-26.
Kelly, S., Summerbell, C., Brynes, A., Whittaker, V. & Frost, G. 2009, "
Wholegrain cereals for coronary heart disease (Review)", The Cochrane
Collaboration and published in The Cochrane Library, , no. 1.
Knoops, K.T.B., de Groot, L.C.P.G.M., Kromhout, D., Perrin, A., Moreiras-Varela, O.,
Menotti, A. & van Staveren, W.A. 2004, "Mediterranean Diet, Lifestyle Factors, and
10-Year Mortality in Elderly European Men and Women: The HALE Project", JAMA:
The Journal of the American Medical Association, vol. 292, no. 12, pp. 1433-1439.
Kruger, HS., Venter, CS., . Vorster, HH. 2003, " Physical activity as a risk factor for
cardiovascular disease in communities undergoing rural to urban transition: The
THUSA study", Cardiovascular Journal of South Africa, vol. 14, no. 1, pp. 16-23.
Kuipers, R., . de Graaf, D., . Luxwolda, M., . Muskiet, M., Dijck-Brouwer, D. & Muskiet, F.
2 0 1 1, " Saturated fat, carbohydrates and cardiovascular disease", The Netherlands
Journal Medicine, vol. 6 9, no. 9, pp. 372-378.
L
Lakier JB. 1992, “Smoking and Cardiovascular Disease”, The American Journal of
Medicine, vol. 93 (suppl 1A).
Langenhoven, M., Kruger, M., Gouws, E. & Faber, M. (eds) 1991, MRC Food
Consumption tables, 3rd edn, MRC, South Africa.
Langsted, A., Freiberg, J.J. & Nordestgaard, B.G. 2008, "Fasting and Non-fasting Lipid
Levels", Circulation, vol. 118, no. 20, pp. 2047-2056.
277 | P a g e
References
Letamo, G. 2011, "The prevalence of, and factors associated with, overweight and
obesity in Botswana", Journal of Biosocial Science, vol. 43, no. 01, pp. 75.
Liu, S., Manson, J.E., Lee, I., Cole, S.R., Hennekens, C.H., Willett, W.C. & Buring, J.E.
2000, "Fruit and vegetable intake and risk of cardiovascular disease: the Women's
Health Study", American Journal of Clinical Nutrition, vol. 72, no. 4, pp. 922-928.
Luke, A., Cooper, RS., Prewitt, TE., Adeyemo, AA., Forrester, TE. 2001, “Nutrition
Consequences of the African diaspora” Annu. Rev. Nutr, vol. 21, pp. 47-71
Lusis, A.J. 2000, "Atherosclerosis", Nature, vol. 407, no. 6801, pp. 233-241.
M
Mallon, PW & Sattar, N. Chapman-Novakofski, K. 2008, "Cardiovascular Risk in the HIV-
infected Patient ", touch briefings.
Mann, J.I. & Cummings, J.H. 2009, "Possible implications for health of the different
definitions of dietary fibre", Nutrition, metabolism, and cardiovascular diseases :
NMCD, vol. 19, no. 3, pp. 226-229.
Maruapula, S.D., Chapman-Novakofski, K. 2007, "Health and Dietary Patterns of the

Elderly in Botswana", J Nutr Educ Behav, vol. 19, pp. 311-319.
Maruapula, S.D., Jackson, J.C., Holsten, J., Shaibu, S., Malete, L., Wrotniak, B., Ratcliffe,
S.J., Mokone, G.G., Stettler, N. & Compher, C. 2011, "Socio-economic status and
urbanization are linked to snacks and obesity in adolescents in Botswana", Public
health nutrition, vol. 14, no. 12, pp. 2260.
Masson, L., MCNeill, G., Tomany, J., Simpson, J., Peace, H., Wei, L., Grubb, D. & Bolton-
Smith, C. 2003, "Statistical approaches for assessing the relative validity of a food-
frequency questionnaire: use of correlation coefficients and the kappa statistic",
Public health nutrition, vol. 6, no. 03, pp. 313.
Mathers, C.D. & Loncar, D. 2006, "Projections of global mortality and burden of disease
from 2002 to 2030", PLoS Medicine, vol. 3, no. 11, pp. 2011-2030.
Mbalilaki, J.A., Hellenius, M.L., Masesa, Z., Hostmark, A.T., Sundquist, J. & Stromme,
S.B. 2007, "Physical activity and blood lipids in rural and urban Tanzanians",
Nutrition, metabolism, and cardiovascular diseases : NMCD, vol. 17, no. 5, pp. 344-
348.
278 | P a g e
References
Mbanya, J., Minkoulou, E., Salah, J. & Balkau, B. 1998, "The prevalence of hypertension
in rural and urban Cameroon", International journal of epidemiology, vol. 27, no. 2,
pp. 181-185.
Mbanya, J.C.N., Motala, A.A., Sobngwi, E., Assah, F.K. & Enoru, S.T. "Diabetes in sub-
Saharan Africa", The Lancet, vol. 375, no. 9733, pp. 2254-2266.
Mehio, S.A., Nasreddine, L., Mokdad, A., Adra, N., Tabet, M. & Hwalla, N. 2010, "
Nutrition Transition and Cardiovascular Disease Risk Factors in Middle East and
North African countries: Reviewing the evidence.", Ann Nutr Metab, vol. 57, pp.
193-203.
Mendez, M.A., Monteiro, C.A. & Popkin, B.M. 2005, "Overweight exceeds underweight
among women in most developing countries", The American Journal of Clinical
Mendis, S., Puska, P. & Norrving, B. 2011, Global Atlas on Cardiovascular disease
prevention and control, World Health Organisation in collaboration with World
Heart Federation and World Stroke organisation, Geneva Switzerland.
Mengesha, A. 2007, " Hypertension and related risk factors in type 2 diabetes mellitus
(DM) patients in Gaborone City Council (GCC) clinics, Gaborone, Botswana.",
African Health Sciences, vol. 7, no. 4.
Mensink, R.P., Zock, P.L., Kester, A.D. & Katan, M.B. 2003, "Effects of dietary fatty acids
and carbohydrates on the ratio of serum total to HDL cholesterol and on serum
lipids and apolipoproteins: a meta-analysis of 60 controlled trials", The American
Journal of Clinical Nutrition, vol. 77, no. 5, pp. 1146-1155.
Mente, A., de Koning, L., Shannon, H.S. & Anand, S.S. 2009, "A systematic review of the
evidence supporting a causal link between dietary factors and coronary heart
disease", Archives of Internal Medicine, vol. 169, no. 7, pp. 659-669.
Ministry of Agriculture Fisheries and Food. “Fatty acids: supplement to the composition
of foods”, Royal Society of Chemistry and Ministry of Agriculture, Fisheries and
Foods. London.
Montalcini, T., Gorgone, G., Feridico, D., Ceravalo, R., Emanuele, V., Sesti, G., Perticone,
F. & Pujia, A. 2005, "Association of LDL cholesterol with carotid atherosclerosis in
menopausal women affected by metabolic syndrome", Nutrition, Metabolism and
Cardiovascular Diseases, vol. 15, pp. 368-372.
279 | P a g e
References
Mozaffarian, D., Micha, R. & Wallace, S. 2010, "Effects on coronary heart disease of
increasing polyunsaturated fat in place of saturated fat: A systematicreview and
meta-analysis of randomized controlled trials", PLoS Medicine, vol. 7, no. 3.
Mozaffarian, D., Ascherio, A., Hu, F.B., Stampfer, M.J., Willett, W.C., Siscovick, D.S. &
Rimm, E.B. 2005, "Interplay Between Different Polyunsaturated Fatty Acids and Risk
of Coronary Heart Disease in Men", Circulation, vol. 111, no. 2, pp. 157-164.
Mozaffarian, D., Katan, M.B., Ascherio, A., Stampfer, M.J. & Willett, W.C. 2006, "Trans
Fatty Acids and Cardiovascular Disease", N Engl J Med, vol. 354, no. 15, pp. 1601-
1613.
Mozaffarian, D., Kumanyika, S.K., Lemaitre, R.N., Olson, J.L., Burke, G.L. & Siscovick,
D.S. 2003, "Cereal, Fruit, and Vegetable Fiber Intake and the Risk of Cardiovascular
Disease in Elderly Individuals", JAMA: The Journal of the American Medical
Msyamboza, KP; Ngwira, B; Dzowela, T; Mvula, C; Kathyola, D;Harries, AD; Bowie, C;

(2011) “The Burden of Selected Chronic Non-Communicable Diseases and Their Risk
Factors in Malawi: Nationwide STEPS Survey”, PLoS One, vol 6, no. 5.
Murray, C.J. & Lopez, A.D. 1997, "Mortality by cause for eight regions of the world:
Global Burden of Disease Study", The Lancet, vol. 349, no. 9061, pp. 1269-1276.
N
National Institute for Health and Clinical Excellence (NICE) 2010, "Prevention of
cardiovascular disease at population level." NICE, London UK..
National Research Council 2005, " Dietary Reference Intakes." Dietary Reference
Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and
Amino Acids (Macronutrients). The National Academies Press Washington, DC.
NCEP 2002, "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, vol. 106, no. 25, pp. 3143-
3421.
Newby, P., Maras, J., Bakun, P., Ferrucci, L. & Turker, K.L. 2007, "Intake of whole grains,
refined grains, and cereal fiber measured with 7-d diet records and the association
with risk factors for chronic disease", Am J Clin Nutr, vol. 86, pp. 1745-53.
280 | P a g e
References
Newby, P.K. & Tucker, K.L. 2004, "Empirically Derived Eating Patterns Using Factor or
Cluster Analysis: A Review", Nutrition reviews, vol. 62, no. 5, pp. 177-203.
Newby, P., Muller, D., Hallfrisch, J., Qiao, N., Andres, R. & Tucker, K.L. 2003, "Dietary
patterns and changes in body mass index and waist circumference in adults", The
American Journal of Clinical Nutrition, vol. 77, no. 6, pp. 1417-1425.
Njelekela, M., Mpembeni, R., Muhihi, A., Mligiliche, N., Spiegelman, D., Hertzmark, E.,
Liu, E., Finkelstein, J., Fawzi, W., Willett, W. & Mtabaji, J. 2009, "Gender-related
differences in the prevalence of cardiovascular disease risk factors and their
correlates in urban Tanzania", BMC Cardiovascular Disorders, vol. 9, no. 1, pp. 30.
Njelekela, M., Negishi, H., Nara, Y., Tomohiro, M., Kuga, S., Noguchi, T., Kanda, T.,
Yamori, M., Mashalla, Y., Jian Liu, L., Mtabaji, J., Ikeda, K. & Yamori, Y. 2001,
"Cardiovascular risk factors in Tanzania: a revisit", Acta Tropica, vol. 79, no. 3, pp.
231-239.
Nkondjock, A & Bizome, E. 2010, “Dietary patterns associated with hypertension

prevalence in the Cameroon defence forces”, European Journal of Clinical Nutrition,
vol. 64, pp.1014-1021.
O
Oh, K., Hu, F.B., Manson, J.E., Stampfer, M.J. & Willett, W.C. 2005, "Dietary Fat Intake
and Risk of Coronary Heart Disease in Women: 20 Years of Follow-up of the Nurses'
Health Study", American Journal of Epidemiology, vol. 161, no. 7, pp. 672-679.
Omran, RA. 1971 “ The Epidemiologic Transition: A Theory of the Epidemiology of Population
Change”, The Milbank Memorial Fund Quarterly, vol. 49, no. 4, pp. 509-538
Oomen, C.M., Ocke, M.C., Feskens, E.J., van Erp-Baart, M.A., Kok, F.J. & Kromhout, D.
2001, "Association between trans fatty acid intake and 10-year risk of coronary
heart disease in the Zutphen Elderly Study: a prospective population-based study",
Lancet, vol. 357, no. 9258, pp. 746-751.
Oosthuizen, W.; Voster, HH.; Kruger, A.; Venter, CS.; Kruger, HS.; de Ridder, JH. 2002,
"Impact of urbanisation on serum lipid profiles- the THUSA survey ", S Afr Med J.
vol. 92, no. 9, pp. 723-728.
281 | P a g e
References
P
Panagiotakos, D.B., Pitsavos, C. & Stefanadis, C. 2006, "Dietary patterns: A
Mediterranean diet score and its relation to clinical and biological markers of
cardiovascular disease risk", Nutrition, Metabolism and Cardiovascular Diseases,
vol. 16, no. 8, pp. 559-568.
Pere, D., Ignacio, S.L., Ramon, T., Fernando, L., Alberto, T., Pompeyo, V., Juan, G., M
Jose, G., Paloma, G., Antonio, V., Jaime, C., Esteban, R., Bernardino, R., M Luisa,
G.A., Trinitario, S., Ferran, T., Juan Ramon, L. & Myriam, G. 2008, "Dyslipidemia and
Cardiovascular Disease Risk Factor Management in HIV-1-Infected Subjects Treated
with HAART in the Spanish VACH Cohort", The open AIDS journal, vol. 2, pp. 26-38.
Pereira, M.A., O'Reilly, E., Augustsson, K., Fraser, G.E., Goldbourt, U., Heitmann, B.L.,
Hallmans, G., Knekt, P., Liu, S., Pietinen, P., Spiegelman, D., Stevens, J., Virtamo, J.,
Willett, W.C. & Ascherio, A. 2004, "Dietary Fiber and Risk of Coronary Heart
Disease: A Pooled Analysis of Cohort Studies", Archives of Internal Medicine, vol.
164, no. 4, pp. 370-376.
Pietinen, P., Rimm, E.B., Korhonen, P., Hartman, A.M., Willett, W.C., Albanes, D. &
Virtamo, J. 1996, "Intake of Dietary Fiber and Risk of Coronary Heart Disease in a
Cohort of Finnish Men: The Alpha-Tocopherol, Beta-Carotene Cancer Prevention
Study", Circulation, vol. 94, no. 11, pp. 2720-2727.
Pilote, L., Dasgupta, K., Guru, V., Humphries, K.H., McGrath, J., Norris, C., Rabi, D.,
Tremblay, J., Alamian, A., Barnett, T., Cox, J., Ghali, W.A., Grace, S., Hamet, P., Ho,
T., Kirkland, S., Lambert, M., Libersan, D., O'Loughlin, J., Paradis, G., Petrovich, M. &
Tagalakis, V. 2007, "A comprehensive view of sex-specific issues related to
cardiovascular disease", Canadian Medical Association journal, vol. 176, no. 6, pp.
S1-44.
Pisa, P., Kruger, A., Vorster, H.H., Margetts, B. & Loots Du, T. 2010, "Alcohol
consumption and cardiovascular disease risk in an African population in transition:
The Prospective Urban and Rural Epidemiology (PURE) study", South African Journal
of Clinical Nutrition, vol. 23, no. 3 (supplement).
Pletcher, M.J., Varosy, P., Kiefe, C.I., Lewis, C.E., Sidney, S. & Hulley, S.B. 2005, "Alcohol
Consumption, Binge Drinking, and Early Coronary Calcification: Findings from the
Coronary Artery Risk Development in Young Adults (CARDIA) Study", American
Journal of Epidemiology, vol. 161, no. 5, pp. 423-433.
282 | P a g e
References
Plump, A., Scott, C. & Breslow, J. 1994, "Human Apoliprotein A-1 gene expression
increases high density lipoprotein and suppresses atherosclerosis in the
apolipoprotein E-deficient mouse", Proc. natl. Acad. Sci, vol. 91, pp. 9607-9611.
Popkin, B.M. 2002, "Part II. What is unique about the experience in lower-and middle-
income less-industrialised countries compared with the very-high-income
industrialised countries?", Public health nutrition, vol. 5, no. 1a, pp. 205.
Popkin, B.M. 2001, "The Nutrition Transition and Obesity in the Developing World",
The Journal of nutrition, vol. 131, no. 3, pp. 871S-873S.
Popkin, B.M & Gordon-Larsen. 2004, "The Nutrition Transition: worldwide obesity
dynamics and their determinants ", International Journal of Obesity, vol. 28, , pp.
S2-S9.
Prentice, A.M. February 2006, "The emerging epidemic of obesity in developing

countries", International journal of epidemiology, vol. 35, no. 1, pp. 93-99.
Prentice, A.M., Rayco-solon, P. & Moore, S.E. 2005, "Insights from the developing
world: thrifty genotype and thrifty phenotypes", Proceedings of the Nutrition
Society, vol. 64, pp. 153-161.
Prignot, J. 1987, “Quantification and chemical markers of tobacco smoke”, Eur J Respir
Dis, vol 70, no.1, pp. 1-7.
Pryer, J.A., Nichols, R., Elliott, P., Thakrar, B., Brunner, E. & Marmot, M. 2001, "Dietary
patterns among a national random sample of British adults", Journal of
epidemiology and community health, vol. 55, no. 1, pp. 29-37.
Puoane, T., Steyn, K., Bradshaw, D., Laubscher, R., Fourie, J., Lambert, V. & Mbananga,
N. 2002, "Obesity in South Africa: the South African demographic and health
survey", Obesity research, vol. 10, no. 10, pp. 1038-1048.
R
Randall, E., Marshall, J., Graham, S. & Brasure, J. 1990, "Patte rns in food use and their
associations with nutrient intakes", The American Journal of Clinical Nutrition, vol.
52, no. 4, pp. 739-745.
Rasmussen, B.M., Vessby, B., Uusitupa, M., Berglund, L., Pedersen, E., Riccardi, G.,
Rivellese, A.A., Tapsell, L., Hermansen, K. & for The KANWU Study Group, 2006,
"Effects of dietary saturated, monounsaturated, and n-3 fatty acids on blood
283 | P a g e
References
pressure in healthy subjects", American Journal of Clinical Nutrition, vol. 83, no. 2,
pp. 221-226.
Ravnskov, U. 1998, "The Questionable Role of Saturated and Polyunsaturated Fatty

Acids in Cardiovascular Disease", J Clin Epidemiol, vol. 51, no. 6, pp. 443–460.
Renaud, S. & de Lorgeril, M. 1992, "Wine, alcohol, platelets, and the French paradox
for coronary heart disease", The Lancet, vol. 339, no. 8808, pp. 1523-1526.
Republic of Botswana. 2008, Control of goods (intoxicating liquor (levy)) (amendment)

regulations , S.I. No.111 Government Gazette Extraordinary, XLVI (84) edn,
Government press, Gaborone.
Rexrode, K.M., Buring, J.E. & Manson, J.E. 2001, "Abdominal and total adiposity and
risk of coronary heart disease in men", International journal of obesity and related
metabolic disorders : journal of the International Association for the Study of
Obesity, vol. 25, no. 7, pp. 1047-1056.
Rexrode, K.M., Carey, V.J., Hennekens, C.H., Walters, E.E., Colditz, G.A., Stampfer, M.J.,
Willett, W.C. & Manson, J.E. 1998, "Abdominal Adiposity and Coronary Heart
Disease in Women", JAMA: The Journal of the American Medical Association, vol.
280, no. 21, pp. 1843-1848.
Rimm, E.B., Ascherio, A., Giovannucci, E., Spiegelman, D., Stampfer, M.J. & Willett, W.C.
1996, "Vegetable, fruit, and cereal fiber intake and risk of coronary heart disease
among men", JAMA: The Journal of the American Medical Association, vol. 275, no.
6, pp. 447-451.
Rimm, E. 2000, "Alcohol and Cardiovascular disease", Current Atherosclerosis Reports,

vol. 2, pp. 529-535.
Rimm, E.B. & Stampfer, M.J. 2002, "Wine, Beer, and Spirits: Are They Really Horses of a
Different Color?", Circulation, vol. 105, no. 24, pp. 2806-2807.
Rimm, E.B., Williams, P., Fosher, K., Criqui, M. & Stampfer, M.J. 1999, "Moderate
alcohol intake and lower risk of coronary heart disease: meta-analysis of effects on
lipids and haemostatic factors", BMJ, vol. 319, no. 7224, pp. 1523-1528.
Rodrıguez, M., Mendez, H., Torun, B., Schroeder, D. & Steyn, A. 2002, "
Validation of a semi-quantitative food-frequency questionnaire for use among
adults in Guatemala ", Public Health Nutr, vol. 5, no. 5, pp. 691-698.
Roger, V.L., Go, A.S., Lloyd-Jones, D.M., Adams, R.J et al. 2011, "Heart Disease and
Stroke Statistics—2011 Update. ", Circulation, vol. 123, no. 4, pp. e18-e209.
284 | P a g e
References
Ronksley, P.E., Brien, S.E., Turner, B.J., Mukamal, K.J. & Ghali, W.A. 2011, "Association
of alcohol consumption with selected cardiovascular disease outcomes: a
systematic review and meta-analysis", BMJ, vol. 342.
Rosamond, W., Flegal, K., Furie, K., Go, A., Greenlund, K., Haase, N., Hailpern, S.M., Ho,
M., Howard, V., Kissela, B., Kittner, S., Lloyd-Jones, D., McDermott, M., Meigs, J.,
Moy, C., Nichol, G., O’Donnell, C., Roger, V., Sorlie, P., Steinberger, J., Thom, T.,
Wilson, M. & Hong, Y. 2008, "Heart Disease and Stroke Statistics —2008 Update",
Circulation, vol. 117, no. 4, pp. e25-e146.
Ross, R. 1999, "Atherosclerosis -- An Inflammatory Disease", The New England journal

of medicine, vol. 340, no. 2, pp. 115-126.
S
Sanders, T. & Emery, P. 2003, "Molecular basis of human nutrition", Taylor and Francis,
london, .
Sarwar, N., Danesh, J., Eiriksdottir, G., Sigurdsson, G., Wareham, N., Bingham, S.,
Boekholdt, S.M., Khaw, K. & Gudnason, V. 2007, "Triglycerides and the Risk of
Coronary Heart Disease", Circulation, vol. 115, no. 4, pp. 450-458.
Schutte, A.E., Kruger, H.S., Wissing, M.P., Underhay, C. & Vorster, H.H. 2005, "The
emergence of the metabolic syndrome in urban obese African women: The POWIRS
study", South African Journal of Science, vol. 101, no. 1-2, pp. 61-67.
Schwartz, T., Magdi, G., Steen, T. & Sjaastad, I. 2012, "

HIV as a risk factor for cardiac disease in Botswana: a cross-sectional study",
International Health, vol. 4, no. 1, pp. 30-37.
Scientific Advisory Committee on Nutrition (SACN). 2011, , Dietary recommendations

for energy [Homepage of SACN], [Online]. Available:
http://www.sacn.gov.uk/reports_position_statements/reports/sacn_dietary_reco
mmendations_for_energy_report.html [2012, march].
Shaw, J.E., Sicree, R.A. & Zimmet, P.Z. 2010, "Global estimates of the prevalence of
diabetes for 2010 and 2030", Diabetes research and clinical practice, vol. 87, no. 1,
pp. 4-14.
Shikany, JM & White, GL. 2000, “Dietary Guideline for Chronic Disease Prevention”,
Southern Medical Journal, vol. 93, no. 12.
285 | P a g e
References
Siervo, M., Grey, P., Nyan, O.A. & Prentice, A.M. 2006, "A pilot study on body image,
attractiveness and body size in Gambians living in an urban community", Eat
Weight Disord, vol. 11, no. 2, pp. 100-109.
Simopoulos, A.P. 1999, "Essential fatty acids in health and chronic disease", The
American Journal of Clinical Nutrition, vol. 70, no. 3, pp. 560S-569S.
Siri-Tarino, P., Sun, Q., Hu, F. & Krauss, R. 2010, "Saturated fat, carbohydrate, and
cardiovascular disease ", Am J Clin Nutr., vol. 91, no. 3, pp. 502–509.
Slavin, J. 2003, "Why whole grains are protective: biological mechanisms", Proceedings
of the Nutrition Society, vol. 62, pp. 129-134.
Smith, CJ., Fischer, TH., Sears, SB. 2000, “Environmental Tobacco Smoke,
Cardiovascular Disease, and the Nonlinear Dose-Response Hypothesis”,
Toxicological Sciences, vol. 54, pp. 462-472
Sodjinou, R., Agueh, V., Fayomi, B. & Delisle, H. 2008, "Obesity and cardio-metabolic
risk factors in urban adults of Benin: Relationship with socio-economic status,
urbanisation, and lifestyle patterns", BMC Public Health, vol. 8, no. 84.
Sofi, F., Capalbo, A., Cesari, F., Abbate, R. & Gensini, G.F. 2008, "Physical activity during
leisure time and primary prevention of coronary heart disease: an updated meta -
analysis of cohort studies", European journal of cardiovascular prevention and
rehabilitation: official journal of the European Society of Cardiology, Working
Groups on Epidemiology & Prevention and Cardiac Rehabilitation and Exercise
Physiology, vol. 15, no. 3, pp. 247-257.
Speth, J.D. 2010, The paleo-anthropology and archaeology of big game hunting: protein
fat or politics? Springer, New York.
Srinath Reddy, K. & Katan, M.B. 2004, "Diet, nutrition and the prevention of
hypertension and cardiovascular diseases", Public health nutrition, vol. 7, no. 1A,
pp. 167-186.
Stevens, J., Ahn, K., Juhaeri, , Houston, D., Steffan, L. & Couper, D. 2002, "Dietary Fiber
Intake and Glycemic Index and Incidence of Diabetes in African-American and
White Adults: The ARIC Study", Diabetes care, vol. 25, no. 10, pp. 1715-1721.
Stewart de Ramirez, S., Enquobahrie, D., Nyadzi, G., Mjungu, D., Magombo, F., Ramirez,
M., Ehrlich-Sachs, S. & Willett, W. 2010, "Prevalence and correlates of
hypertension: a cross-sectional study among rural populations in sub-Saharan
Africa", Journal of Human Hypertension, vol. 24, pp. 786-795.
286 | P a g e
References
Steyn, K., Sliwa, K., Hawken, S., Commerford, P., Onen, C., Damasceno, A., Ounpuu, S.,
Yusuf, S. & INTERHEART Investigators in Africa 2005, "Risk factors associated with
myocardial infarction in Africa: the INTERHEART Africa study", Circulation, vol. 112,
no. 23, pp. 3554-3561.
Streppel, M.T., Ocke, M.C., Boshuizen, H.C., Kok, F.J. & Kromhout, D. 2008, "Dietary
fiber intake in relation to coronary heart disease and all-cause mortality over 40 y:
the Zutphen Study", American Journal of Clinical Nutrition, vol. 88, no. 4, pp. 1119-
1125.
T
Taylor, R., Ashton, K., Moxham, T., Hooper, L. & Ebrahim, S. 2011, " Reduced dietary
salt for the prevention of cardiovascular disease: a meta-analysis of randomized
controlled trials (Cochrane review).", Am J Hypertens., vol. 24, no. 8, pp. 843-53.
Thies, F., Garry, J.M., Yaqoob, P., Rerkasem, K., Williams, J., Shearman, C.P., Gallagher,
P.J., Calder, P.C. & Grimble, R.F. 2003, "Association of n-3 polyunsaturated fatty
acids with stability of atherosclerotic plaques: a randomised controlled trial", The
Lancet, vol. 361, no. 9356, pp. 477-485.
Thorogood, M., Connor, M., Tollman, S., Lewando Hundt, G., Fowkes, G. & Marsh, J.
2007, "A cross-sectional study of vascular risk factors in a rural South African
population: data from the Southern African Stroke Prevention Initiative (SASPI)",
BMC public health, vol. 7, pp. 326.
Tibazarwa, K., Ntyintyane, L., Sliwa, K., Gerntholtz, T., Carrington, M., Wilkinson, D. &
Stewart, S. 2009, "A time bomb of cardiovascular risk factors in South Africa: results
from the Heart of Soweto Study "Heart Awareness Days"", International journal of
cardiology, vol. 132, no. 2, pp. 233-239.
Tighe, P., Duthie, G., Vaughan, N., Brittenden, J., Simpson, W.G., Duthie, S., Mutch, W.,
Wahle, K., Horgan, G. & Thies, F. 2010, "Effect of increased consumption of whole-
grain foods on blood pressure and other cardiovascular risk markers in healthy
middle-aged persons: a randomized controlled trial", The American Journal of
Clinical Nutrition, vol. 92, no. 4, pp. 733-740.
Townsend, L., Flisher, A.J., Gilreath, T. & King, G. 2006, "A systematic literature review
of tobacco use among adults 15 years and older in sub-Saharan Africa", Drug and
alcohol dependence, vol. 84, no. 1, pp. 14-27.
287 | P a g e
References
Trichopoulou, A., Costacou, T., Bamia, C. & Trichopoulos, D. 2003, "Adherence to a

Mediterranean Diet and Survival in a Greek Population", The New England journal
of medicine, vol. 348, no. 26, pp. 2599-2608.
Trowell, H. 1972, “ Ischemic heart disease and dietary fiber”, The American Journal of
Clinical Nutrition, vol. 25, pp. 926-932
Truswell, A.S. 1977, "Diet and nutrition of hunter-gatherers", Ciba Foundation

symposium, vol. 49, pp. 213-21.
Truswell, A.S., Kennelly, B.M., Hansen, J.D.L. & Lee, R.B. 1972, "Blood pressures of
!Kung bushmen in Northern Botswana", American Heart Journal, vol. 84, no. 1, pp.
5-12.
Tsiodras, S., Poulia, KA., Yannakoulia, M., Chimienti, SN., Wadhwa, S. Karchmer., AW.,
Mantzoros, CS. 2009, “Adherence to Medittereanean diet is favourably associated
with metabolic parameters in HIV positive patients with the HAART-induced
metabolic syndrome and lipodystrophy”, Metabolism, vol. 58, no. 6, pp. 854-859.
Tungland, B. & Meyer, D. 2002, "Non digestible oligo and polysaccharrides (Dietary
fiber):The physiology and role in human health and food", Comprehensive reviews
in food science and food safety, vol. 3.
Turcinov, D., Stanley, C., Rutherford, G.W., Novotny, T.E. & Begovac, J. 2009,
"Adherence to the Mediterranean diet is associated with a lower risk of body-shape
changes in Croatian patients treated with combination antiretroviral therapy",
European journal of epidemiology, vol. 24, no. 5, pp. 267-274.
U
UNAIDS. 2010, Report on the global AIDS epidemic. Geneva Switzerland.
University of Vermont. Department of Nutrition and Food Sciences "Methods of body

composition analysis", University of Vermont available at
www.nutrition.uvm.edu/bodycomp/bia (accessed August 2009).
V
Vaccaro, O., Stamler, J. & Neaton, J.D. 1998, "Sixteen-year coronary mortality in black
and white men with diabetes screened for the Multiple Risk Factor Intervention
Trial (MRFIT)", International journal of epidemiology, vol. 27, no. 4, pp. 636-641.
288 | P a g e
References
van der Sande MAB, Milligan PJM, Nyan OA, et al. 2000, “Blood pressure patterns and
cardiovascular risk factors in rural and urban Gambian communities”, J Hum
Hypertens, vol.14, pp.489–496.
van de Wiel, A. & de Lange, D. 2008, "Cardiovascular risk is more related to drinking
patterns than the type of alcoholic drinks", The Netherlands Journal Medicine, vol.
66, no. 11, pp. 467-473.
Van Duyn, M.A. & Pivonka, E. 2000, "Overview of the health benefits of fruit and
vegetable consumption for the dietetics professional: selected literature", Journal
of the American Dietetic Association, vol. 100, no. 12, pp. 1511-1521.
Van Pelt RE, Evans EM, Schechtman KB, Ehsani AA, Kohrt WM. 2001 “Waist
circumference vs body mass index for prediction of disease risk in postmenopausal
women.” Int J Obes Relat Metab Disord vol. 25, pp. 1183–1188
van Wesenbeeck, C., Keyzer, M. & Nubé, M. 2009, "Estimation of undernutrition and
mean calorie intake in Africa: methodology, findings and implications.", Int J Health
Geogr, vol. 8, no. 37.
Vedtofte, M.S., Jakobsen, M.U., Lauritzen, L. & Heitmann, B.L. 2011, "Dietary α-
linolenic acid, linoleic acid, and n–3 long-chain PUFA and risk of ischemic heart
disease", The American Journal of Clinical Nutrition, vol. 94, no. 4, pp. 1097-1103.
Vorster, H. 2001, "South African food based dietary guidelines", S Afr J Clin Nutr
(Supplement), vol. 14, no. 3, pp. 1-80.
Vorster, H.H. 2002, "The emergence of cardiovascular disease during urbanisation of

Africans", Public health nutrition, vol. 5, no. 1 A, pp. 239-243.
Vorster, H.H., Venter, C.S., Wissing, M.P. & Margetts, B.M. 2005, "The nutrition and
health transition in the North West Province of South Africa: A review of the THUSA
(Transition and Health during Urbanisation of South Africans) study", Public Health
Vorster, H.H., Kruger, A., Venter, C.S., Margetts, B.M. & Macintyre, U.E. 2007,
"Cardiovascular disease risk factors and socio-economic position of Africans in
transition: the THUSA study", Cardiovascular journal of Africa, vol. 18, no. 5, pp.
282-289.
289 | P a g e
References
W
Walker, A.R., Walker, B.F., Sci, D.D. & Adam, F. 2002, "Variations in occurrences of
nutrition-related diseases in Sub-Saharan Africans in stages of transition: what of
the future?", Nutrition (Burbank, Los Angeles County, Calif.), vol. 18, no. 1, pp. 71-
74.
Walker, A.R.P. & Walker, B.F. 1978, "High high-density-lipoprotein cholesterol in

African
children and adults n a population free of coronary heart disease ", British Medical
Journal, vol. 2, pp. 1336-1338.
Wehmeyer, A. 1966, "Nutrient composition of some edible wild fruit found in the
transvaal", South African Medical Journal, vol. 40, no. 16, pp. 1102-1104.
WHO. 1990. Diet, nutrition and the prevention of chronic diseases. Technical Series
Report 797. Geneva Switzerland.
WHO. 2002, World Health Report - 2002, WHO press, Geneva Switzerland.
WHO expert consultation 2011, Waist circumference and waist-hip ratio, WHO press,
Geneva Switzerland.
WHO. 2010, Procedural Manual for the Development of Regional and Country-Specific
Food-Based Dietary Guidelines (FBDGs), WHO press, Geneva Switzerland.
WHO. 2009, Global Health risk: mortality and burden of disease attributable to selected
major risks, WHO press, Geneva Switzerland.
WHO. 2008, 2008-2013 Action plan for the Global Strategy for the Prevention and
Control of Noncommunicable Diseases, WHO press, Geneva Switzerland.
WHO/International Diabetes Federation (IDF). 2006, “Definition and diagnosis of

diabetes mellitus and intermediate hyperglycemia”, WHO press, Geneva
Switzerland.
WHO. 2000, "Obesity; prevention and managing the global epidemic. Report of WHO
consultation", WHO technical report series;894 .
WHO/International Society of Hypertension (ISH). 2003, "Statement on management of

hypertension", Journal of hypertension, vol. 21, pp. 1983-1992.
WHO STEPS surveys. available at http://www.who.int/chp/steps accessed Jan 2012).
290 | P a g e
References
Wijendran, V. & Hayes, K. 2004, "Dietary n-6 and n-3 fatty acid balance and
cardiovascular health ", Annual Review of Nutrition, vol. 24, pp. 597-615.
Willett, WC. 1998, “Nutritional Epidemiology”, Oxford University press.
Willett, W.C., Stampfer, M.J., Manson, J.E., Colditz, G.A., Speizer, F.E., Rosner, B.A.,
Sampson, L.A. & Hennekens, C.H. 1993, "Intake of trans fatty acids and risk of
coronary heart disease among women", Lancet, vol. 341, no. 8845, pp. 581-585.
Wolk, A., Manson, J.E., Stampfer, M.J., Colditz, G.A., Hu, F.B., Speizer, F.E., Hennekens,
C.H. & Willett, W.C. 1999, "Long-term Intake of Dietary Fiber and Decreased Risk of
Coronary Heart Disease Among Women", JAMA: The Journal of the American
Medical Association, vol. 281, no. 21, pp. 1998-2004.
Wong ND, Black HR and Gardin JM, " Preventive cardiology: a practical approach",
Chicago: McGraw-Hill.
Y
Yoon, S., Ostchega, Y. & Louis, T. 2010, , Recent trends in the prevalence of blood
pressure and its treatment and control, 199-2008 - data brief, no.. 48. [Homepage
of US department of health and human services], [Online]. Available:
http://www.cdc.gov/nchs/data/databriefs/db48.pdf [2012, March].
Yusuf, S., Hawken, S., Ôunpuu, S., Bautista, L., Franzosi, M.G., Commerford, P., Lang ,
C.C., Rumboldt, Z., Onen, C.L., Lisheng, L., Tanomsup, S., Wangai Jr, P., Razak, F.,
Sharma, A.M. & Anand, S.S. 2005, "Obesity and the risk of myocardial infarction in
27 000 participants from 52 countries: a case-control study", The Lancet, vol. 366,
no. 9497, pp. 1640-1649.
Z
Zhao, G., Ford, E.S., Li, C. & Mokdad, A.H. 2008, "Are United States adults with
coronary heart disease meeting physical activity recommendations?", The American
Journal of Cardiology, vol. 101, no. 5, pp. 557-561.
Zhou, X., Li, C., Xu, W., Hong, X. & Chen, J. 2010, "Relation of alcohol consumption to
angiographically proved coronary artery disease in Chinese men", The American
Journal of Cardiology, vol. 6, no. 12.
Zock, P.L., de Vries, J.H. & Katan, M.B. 1994, "Impact of myristic acid versus palmitic
acid on serum lipids and lipoprotein level in healthy women and men", Arterioscler
Thromb Vasc Biol, vol. 14, pp. 567-575.
291 | P a g e
Appendices
Appendices
292 | P a g e
Appendix 1 Ethics permit

Appendix 2 Information sheet
Appendix 2 Information sheet (English)

INFORMATION TO PARTICIPANTS
You are invited to take part in a major health research project called the “Botswana Pelo study”. The aim of this study is to hel p
understand the causes of cardiovascular diseases (such as heart disease and stroke) in Botswana. It is the largest such study ever
done in Botswana.
Cardiovascular diseases (including heart diseases) are the leading cause of adult death worldwide, and deaths due to these
diseases are expected to increase substantially in Botswana and the rest of Africa over the next few years. Research has shown
that cardiovascular diseases are caused by a combination of factors, such as lifestyle, environment and diet. We therefor e in vite
you to participate in this study to help us better understand the causes of cardiovascular diseases in Botswana .
If you choose to participate in this research, you will be asked some questions about your medical history, lifestyle and die t. Your
blood pressure, weight, waist circumference and body fat will also be measured, and about 35ml (2 table spoons) of blood will be
collected from a vein in your arm. This blood will be used for measuring levels of lipids, lipoproteins and glucose in your blood to
investigate cardiovascular disease risk factors. Additionally we would like to access your medical records for your medical history
such as your current prescriptions if any. All these activities will take approximately 45 minutes.
Your name and information collected will be kept confidential at all times. Your participation or refusal to participate in t his
research will not by any means affect your treatment at the clinic . You will not be paid or be asked to pay any additional money if
you agree to participate in our research. The results of the research will be circulated in a brief summary that will be circ ulated to
you and the public at the end of the r esearch.
This project is being led by Lemogang Kwape who is a Nutritionist at National Food Technology Research Centr e in Botswana and
a PhD student at the University of Aberdeen in the UK.
This study has been approved by the Health Research Committee in the Ministry of Health in Botswana.
If you have any questions about the study, please contact Lemogang Kwape at the National Food Technology Research Centre.
Phone number 5440 441
I confirm that I have been fully informed about the study. I have also had the opportunity to ask questions and clarifications.
I agree  I disagree 
I also understand that my participation is voluntary and that I am free to stop at any time without giving any reason.
I am aware that data collected may be used in scientific and other publications for the advancement of scientific and public
knowledge. I agree  I disagree 
I understand my identity will be kept confidential, and that I would not benefit financially in any way from taking part.
I give permission for long term (12 months) storage and use of my blood samples for cardiovascular disease research (even aft er
my incapacity or death), and I relinquish all rights to these sample that I am donating to the study.
I agree to take part in the Botswana Pelo initiative I agree  I disagree 
Name Signature Date
For the Researcher:
The above statements were read to the participant. He/she agreed to each of th e above sta tements and has agreed to participate
in the study
Name Signature date

Appendix 2. Information sheet & consent (Setswana)
Appendix 2 Information sheet (Setswana)

KITSISO GO BATSAAKAROLO (Information to participants)
O lalediwa go tsaya karolo mo dipatlisisong tsa botsogo tse di bidiwang “Botswana Pelo study”. Maikemisetso a dipatlisiso tse ke
go leka go tlhaloganya tseo tse di ka tswang di amanngwa le malwetse a pelo le ditshika mono Botswana.
Lefatshe ka bophara go lemotshega ha malwetsi a pelo le a ditshika ele one a eteletseng pelo dipalo tsa dintsho. Se se tshwen yang
ke gore dipalo di lebega di akanyediwa go oketsega thata mo ngwageng tse masome mabi tse di tlang, bogolo jang mo mahatshing
a a fahatlhogang jaaka la Botswana le Afrika ka bophara. Re dira dipatlisiso ka malwetsi a, gore re itse re bo re tlhaloganye kafa
dijo, tikologo le tse dingwe di ka tswana di amana le malwetsi a ka teng.
Fa o dumela go tsaa karolo o tlaa bodiwa dipotso ka botsogo jwa gago, ka dijo, le tse dingwe jaaka go goga motsoko, go itshi dila
mmele, le tsa matshelo a selegae. O tlaa lekanngwa bokete le boleele jwa mmele, ga m mogo le seemo sa mafura le nama mo
mmeleng. Re tla lekanya le gore seemo sa gago sa madi a magolo se eme jang, mme ere ko bofelong re tsee madi (selekanyo sa
leso lele jang ga bedi) go sekaseka seemo sa mafura, sukuri le tse dingwe tse di ka amanang le malwetsi a pelo le ditshika mo
mmeleng. Re kopa gape tetla ya go sekaseka ditso tsa botsogo jwa gago ga mmogo le melemo mo dikarateng ya gago.
Ditshekatsheko tse tsotlhe ditlaa tsaa nako ya metsotso e le masome mane le botlhano (45 minutes).
Leina kana sepe se seka go supang se tlaa tsewa ele sephiri. Ga o patelediwe go tsa karolo, ebile go gana go tsaa karolo gago kake
ga ama ka gope kalafi ya gago mo kokelong. Ga tsaa karolo ga go duelelwe ebile ga o tlhoke go duela sepe go tsaa karolo. Ko
bofelong lo tlaa itsisiwe ka maduo a dipatlisiso tse ka mekwalo le dikitsiso tsa dikgang.
Dipatlisiso tse di eteletswe pele ke Lemogang Kwape yo eleng moitseanape wa dijo ebile a batlisisa ka malwetsi a pelo ko Nati onal
Food Technology Research centre. O tlaabo a thusiwa ke ba Unibesithi ya Aberdeen ko Scotland ko a ithutelang dithuto tsa gagwe
tsa PhD. Mme madi le tse dingwe di tswa kwa lephateng la Dipatlisiso Boranyane le Boitseanape mono Botswana.
Ba lephata la botsogo mono Botswana ba sekasikile dipatlisiso tse babo ba difa tetla gore di tswelele.Ha o na le dipotso kana
dikakgelo ka dipatlisiso tse o ka itshwaraganya le Lemogang Kwape ko National Food Technology Research Centre (NFTRC) mo
mogaleng o: 5440 441
Ke tlhaloseditswe ka botlalo ka dipatlisiso tse, ebile ke filwe sebaka sa go botsa le go tlhalosediwa fa keneng ke sa tlhaloganye teng
kea dumalana  Ga ke dum alane 
Ke tlhaloganya gape gore dikarabo tsame ga mmogo le tsedingwe tse di lekangwang di ka helela di dirisitswe mo mekwalong ya
dipatlisiso ga mmogo le mo go ruteng sechaba.
Ke tlhaloseditswe e bile ke tlhaloganya gore leina lame ka na sepe fela sa bosupi jwa me se tlaa tsewa ele sephiri. Ke tlhaloganya
gape gore ga kena go duelelwa go tsaa karolo.
Kefa tetla ya go boloka (kgwedi tse lesome le bobedi)) dirisa madi ame mo dipatlisisong tsa malwetse a pelo le ditshika le morago
ga leso lame.
Ke dumela go tsaa karolo mo dipatlisisong tse.
Leina Monw ana Letsatsi
Mmatlisisi:
Ke file mo tsaakarolo tlholoso tsotlhe ka go tsaa karolo. O dumalane.
Leina Monwana Letsatsi

Appendix 3. Questionnaire
Appendix 3 Questionnaire
ID. NO.
Entered by:
Date:
Botswana Pelo Study

(Diet and Cardiovascular diseases risk factors in Botswana)
QUESTIONNAIRE
Initial screening criteria
History of Cardiovascular disease: Yes No
Myocardial infarction  
Stroke  
Angina  
Rheumatic heart disease  
Coronary heart disease  
Heart failure  
Other: specify ______________________________
No history of above CVD but known history: Yes No
Diabetes  
Healthy participants: Yes No
Wellness screening  
Routine check up  
Other: specify ______________________________
Pregnant : Yes No
 
Socio-demographic characteristics
1) Study ID
2) Name ……………….. ……… ………….. …………………
3) Address ……………………………………………………………..
4) Tel. Landline………………mobile……............. alternative phone …………..
5) Gender 1) Male  2) Female 

6) How old are you?  years
Day Month Year
7) When were you born?
8) Where were you born? ____________ (village/town/city) ___________(country if not Botswana)
9) Which of the following languages do you speak?
1) Setswana  2) English  3) Kalanga 

4) Herero  5) Seyei  6) Sekgalagadi/Sengologa 
7) Other (specify) _____________
10) What is your Religion? 1) Christianity  2) Islam  3) Traditional beliefs 

4) None  5) Other (specify) ______________
11) What is your level of Education?
1) No formal education 
2) Primary  1. No formal education
3) Secondary  2. Primary:1- 7 years
3. Secondary: 8-12 years

4) Tertiary  4. Tetiary: >12 years
12) What is your Occupation?
1) Professional  2) Artisan  3) Business  4) Homemaker 

5) Farmer  6) Labourer  7) Unemployed  8) Retired 
9) Other (specify) __________________________________
13) What is your marital status?
1) Single  2) Married  3) Divorced  4) Widow  5) Separated 

6) Other
14) What is your estimated household monthly income (Pula)?
1) < 600  2) <3000  3) ≥3000 - <10,000  3) ≥10,000 

15) How many people live in the household? 
Family History of CVD

16) Do you or the following members of your family ever been diagnosed with the following?
High Approx. age Diabetes Approx. age Insulin Approx. age Angina Approx. age Heart Approx. age Stroke Approx. age Sudden Age of
attack
Blood Of diagnosis Of diagnosis Y/N Of diagnosis Of diagnosis Of diagnosis Of diagnosis Death death
pressure
Participant - -
Mother
Father
Sister(s)
Brother(s)
Son (s)
Daughter(s)
vi
YES NO
17) Do you feel pain in your lower extremities while walking?  

(If no, please go to question 21)
18) Does this pain go away at rest immediately?  

19) If yes how long does it take to go away? Hrs  Min
20) Does the pain occur every time you walk?  
21) Does your heart beat increase or become irregular  
When you are resting?
22) Have you ever been diagnosed with arterial fibrillation  

(irregular heart beat)
23) Past surgical history (most recent)  

23a) Procedure performed ________________
23b) Year of surgery 
Smoking history
24) Have you ever smoked as much as one cigarette each day for as long as a year?
Yes  No 
25) If yes, how old were you when you started smoking regularly?  Years
26) Did you smoke at the following ages? If yes, how many cigarettes did you smoke each day?
Under 20 non-smoker  smoker  cigarettes each day

Age 20 non-smoker  smoker  cigarettes
each day

each day

each day

each day

each day

each day

each day
27) Do you smoke now? Yes  No 
If yes, how man cigarettes do you smoke each day? Cigarettes each day
If no, how old were you when you gave up?  Years
28) Have you ever snuffed tobacco regularly? Yes  No 

If yes, for how long have you snuffed?   Years
If yes, how much do you snuff each week? snuffs/week
29) Have you ever smoked a pipe regularly? Yes  No 
If yes, for how long have you smoked one/did you smoke one?  Years
If yes, how many ounces of tobacco do/did you smoke each week ounces/
week
30) Are you exposed to other people’s smoke? Yes  No 

31) If yes where are you exposed?
1) Work  2) Home  3) Public area  4) Friends house 
Alcohol drinking history
32) What best describes your drinking habits?
1) I drink  2) I used to drink  3) never drank 

If never drank go to 36
33) What do/did you drink?
1) Beer  2) Cider  3) Whisky  4) Traditional brew  4) Wine 

5) Other
34) a) Number of drinks per day  drinks/day
35) Age started drinking  years

36) If used to drink, age stopped drinking  years
Physical Activity
Rest
37) Average hours of sleep  Hours

Job
How many days per week do you usually do paid work?  Days
38) What do you do? _____________________________
39) How many hours per day do you spend in paid work?  Hours
40) In each day you do paid work, how many hours do you usually spend in:
41) Sedentary activities  Hours

(e.g. standing, sitting, reading, writing, working at computer)
42) Light activities  Hours

(e.g. general lab work, general office work, driving, unhurried walking)
43) Moderate activity  Hours

(e.g. light lifting or carrying, moderate walking, painting, decorating)
44) Heavy activity  Hours

(e.g. heavy lifting or carrying, hurried walking, going upstairs or ladders, digging, strenuous exercise)
Leisure activities
In your other activities, including housework and travelling, please estimate how many hours per
day you spend in each type of activity, and then indicate how many days each week you usually
do this type of activity
Any time you do not list below will be assumed to be spent in sedentary activities.
45) Light activities  Hours

(e.g. washing, dressing, shaving, ironing, cooking, washing dishes, shopping, driving, unhurried walking)
46) Moderate activity  Hours

(e.g. gardening, golf, cricket, playing pool, cycling, child care, light lifting or carrying, moderate walking,
painting, decorating)
47) Heavy activity  Hours

(e.g. competitive sports, football, swimming, playing tennis, squash, hockey, running, aerobics, dancing,
going upstairs or ladders, digging, strenuous exercise)
Household ownership
Does your household own any of the following?
YES NO YES NO
48) House   52) Television  

49) Car   53) Radio  
50) Bicycle   54) Computer  
51) Cell phone   55) Livestock  
Stress assessment (Perceived Stress)

The questions in this scale ask you about your feelings and thoughts during the last month.
0 = Never 1 = Almost Never 2 = Sometimes 3 = Fairly Often 4 = Very Often
56) In the last month, how often have you been upset because of something that happened
unexpectedly?
0 1 2 3 4
57) In the last month, how often have you felt that you were unable to control the important
things in your life?
0 1 2 3 4
58) In the last month, how often have you felt nervous and “stressed”?
0 1 2 3 4
59) In the last month, how often have you felt confident about your ability to handle your
personal problems?
0 1 2 3 4
60) In the last month, how often have you felt that things were going your way?
0 1 2 3 4
61) In the last month, how often have you found that you could not cope with all the things that
you had to do?
0 1 2 3 4
62) In the last month, how often have you been able to control irritations in your life?
0 1 2 3 4
63) In the last month, how often have you felt that you were on top of things?
0 1 2 3 4
64) In the last month, how often have you been angered because of things that were outside of
your control?
0 1 2 3 4
65) In the last month, how often have you felt difficulties were piling up so high that you could
not overcome them?
0 1 2 3 4
For women only
66) Have you ever used birth control pills or injections? Yes  No 
67) If yes, Age when began  years Aged stopped  years
68) Have you had menstrual period in the last 12 months? Yes  No 
69a) If no, at what age did they stop?  Years
69b) Why did they stop? 1) Menopause  2) Hysterectomy  3) Radiation 
Other specify ______________
70) Have you ever used female hormone replacement therapy? Yes  No 
71a) If yes, are you still taking them? Yes  No 
71b) How long have you taken them? Months  Years

71c) Which type have you used most? Estrogen  Estrogen + progesterone
Food Frequency Questionnaire
We are interested in finding out about your usual eating pattern. Please tell us how
often and how much of each of the following foods you eaten during the last 3 months.
Cereals Almost Once 2-3 Once 2-4 5-6 Once 2+

/Starches never
/ / / / / / / Portion size
mth mth week week week day day
White bread
Brown bread
Phaphatha/baked
bread
Mapakiwa /
Lebaka
Fat cakes
Rice
Pasta
Sorghum meal
Lebelebele/millet
Mealie meal
Baked,boiled or
mashed potato
Sweet potato
Tsabana/fortified
sorghum meal
Samp
Mealie rice
Madombi
Break fast cereals
Peas Beans and Almost Once 2-3 Once 2-4 5-6 Once 2+
Nuts never
Baked beans
Black eye beans
Tswana beans
Bambara nuts
(ditloo)
Peanuts
(manoko)
Green beans
Green peas
Soy/mince/chunk
s) somosa
Milk and Milk Almost Once 2-3 Once 2-4 5-6 Once 2+ Portion size
products never
/ / / / / / /
Whole milk
Skimmed milk
Powdered milk
Condensed milk
Soy milk
Madila/sour milk
Mayere/sour
cream
Cheese
Ice cream
Yoghurt
Infant formula
Foods from Almost Once 2-3 Once 2-4 5-6 Once 2+
Animals never
Eggs
Chicken with skin
Chicken without
skin
Beef
Beef biltong
Mutton (Lamb)
Nama ya podi
Fresh fish
Pork
Minced meat
Serobe
Seswaa/ pounded
meat
Liver
Kidney
Gizzards
Corned beef
sausage/ pastaram
i/polony
Almost Once 2-3 Once 2-4 5-6 Once 2+ Portion size

never
/ / / / / / /
Phane
Canned fish
(Lucky Star)
Dried salted fish
Tuna fish
(canned)
Vegetables Almost Once 2-3 Once 2-4 5-6 Once 2+

never
Beetroot
Tomato
Carrot
Pumpkin/butternut
Maraka /
makgomane
Lerotse/melon
Spinach
Morogo wa
dinawa
Broccoli
Cauliflower
Lettuce
Cucumber
A vocado
Pepper
Mushroom
Cabbage
Tswii
Fruit Almost Once 2-3 Once 2-4 5-6 Once 2+

never
Orange
Naraki
Grapefruit
Pawpaw
Mango
(seasonal)
Banana
Almost Once/ 2-3 Once 2-4 5-6 Once 2+

never
mth / / / / / / Portion size
mth week week week day day
Apple
Peach
Pears
Watermelon
(seasonal)
Grapes
Morula
(seasonal)
Dried fruit
Canned fruit
Strawberry
Pineapple
Mmilo (seasonal)
Moretlwa
(seasonal)
Other, specify
What type of fat do you use most often during cooking?
1) Butter  2) Margarine  3) sunflower oil  4) Olive oil  5) Holsum 
Sweets, fats, Almost Once 2-3 Once 2-4 5-6 Once 2+

oils, and never
condiments
Sweets/candy
Sugar (e.g. added

to tea)
Cookies/biscuit
Cake
Buns/scones/muffi
ns
Peanut butter
Margarine
Jam
Cream cheese
Salad dressing
Mayonnaise
Cheese snacks
Almost Once 2-3 Once 2-4 5-6 Once 2+ Portion size

never
/ / / / / / /
Potato chips
(fries)
Savory meat pies
Jelly
Custard
Beverages Almost Once 2-3 Once 2-4 5-6 Once 2+

never
Sweetened
beverages
Diet sodas
Fruit Juic es
Fruit drinks
Beer
Rum/Whiskey/Gi
n
Wine
Chibuku
Coffee
Tea (Five roses)
Herbal Tea
(Rooibos)
Choc olate drink
Mageu
74) Cardiovascular disease diagnosis
Myocardial Angina Rheumatic Coronary Heart failure

infaction heart disease heart disease
Electrocardiogram
    
Echocardiogram
    
Ultrasoun d
    
Angiography
    
Troponin
    
Nuclear scan
    
_________________________________________________________________________
75) Do you or any of your family members been diagnosed with the following
conditions (from chart/self reported)
Diagnosi s Date of diagnosi s 1 Self reported
2 Chart
(WHO version 2007) month year
i) Hypertension (I10)
 
ii) Angina Pectoris (I20)
 
iii) Acute myocardial infarction (I21)
 
iv) Subsequent Myocardial infarction (I22)
 
v) Ot her acute IHD (I24)
 
vi) Subarachnoid haemorrhage (I60)
 
vii) Intracerebral haemorrhage (I61)
 
viii) Cerebral infarction (I63)
 
.
ix) Stroke not specified as haemorrhagic
or infar tion(I64)  
76) Prescription medication (from chart/Self reported)
Medication Duration 1 YES
Month Year 2 NO
i) Cholesterol ↓
 
ii) Diuretics
 
v) Nitrates
 
vi) Insulin
 
vii) Oral Hypoglycemics
 
viii) Digoxin
 
ix) Thiazides
 
x) Warfarin
 
ix) Ca Channel Blocker
 
x) Beta Blocker
 
xi) Platelet antagonist
Homeopathic  
xii) Heparin
 
Physical Measurements
Time : hrs
Blood Pressure First reading Second reading
77) Systolic
78) Diastolic
79) Heart rate
80) Ankle
81) Brachial
82) Pulse wave Analysis
Anthropometric measurement
83) Weight (kg)
84) Height (cm)
85) Waist circumference (cm)
86) Hip circumference (cm)
87) BIA Impedance
88) Fat mass
89) Lean mass
90) Total body water (litre)
91) Visible lipoatrophy/ lipohyperatrophy in the extremities Yes  No 

If yes specify where___________________________________________________
Blood sample Information
92) Blood samples collected 1) YES  2) NO 

Day Month Year
93) Sample ID#:  Date sample taken
Day Month Year

94) Date and time of last meal :  Hr/Min
95) Number of EDTA tubes  96) Number of serum tubes 
Diabetics Only
96) How long have you been diagnosed with diabetes? years
YES NO
97) Are you taking any diabetes medication?  

98) How long have you been on medication? years
YES NO
99) Have you been advised to change your diet?  

100) Are you adhering to the diet advice?  
101) Do you have trouble seeing?  
102) Do you have any wounds that are not healing?  
Appendix 4. Description of food items listed in the FFQ
Appendix 4 Description of food items listed in the FFQ

Food group Local name Descripti on
Starches and cereals White Bread White bread (loaf)
Bro wn Bread Whole wheat bread (loaf)
Phaphata Flat bun (English muffins)
Mapakiwa/ Lebaka Ho me-made Buns
Fat cake Deep-fat fried (doughnut)
Rice White rice boiled
Pasta Pasta shells boiled
Sorghum meal Sorghum flour
Lebelebele/Millet Millet flour
Mealie meal Maize/ Corn meal
Potato (baked/ boiled/ mashed Irish potato
Sweet Potato White fleshed sweet potato
Tsabana Sorghum-soya infant food
Samp Maize coarse grits
Mealie rice Maize fine grits
Madombi Du mplings
Breakfast cereals Cornflakes

Peas beans and nuts Baked beans Canned beans in tomato sauce
Black eye peas Black eye peas boiled
Tswana Beans Local black eye peas boiled
Bambara Nuts (ditloo) Bambara beans boiled
Peanuts (Manoko) Peanuts salted/ roasted
Green beans Beans boiled
Texturized vegetable protein Texturized Soy

Milk and milk products (di ary)
Whole Milk Dairy milk (fu ll fat)
Skimmed M ilk Dairy milk (no fat)
Powdered whole milk Dairy milk (fu ll fat, dried)
Sweetened Condensed Milk Dairy milk, condensed with sugar
added
Soy Milk Soya bean milk
Madila Sour milk (thick)
Mayere Buttermilk
Cheese Cheddar cheese
Ice Cream Vanilla ice cream
Yoghurt Flavoured / plain s mooth yoghurt
Infant Formula Nestle infant formu la

Foods from ani mals (meat and meat
products)
Eggs Bro wn eggs
Chicken with Skin Chicken meat with skin
Chicken without skin Chicken meat without skin
Beef Cow meat
Beef Biltong Beef jerky
Mutton/ Lamb Young goat / sheep meat
Nama ya Podi Goat meat
Fresh Fish River fish (Chobe Bream)
Pork Pig meat
Minced Meat Ground beef
Serobe Tripe and intestines
Seswaa Shredded meat (beef/goat)
Liver Cow liver
Kidney Cow kidney
Gizzards Chicken gizzards
Corned Beef Canned corned beef
Processed meats Beef / chicken/ pork processed
(Sausage/pastrami/polony) meat
Phane Mophane worm
Canned Fish Canned sardines
Dried salted Fish Fresh water / sea water fish,
dried and salted
Tuna fish (canned) Canned tuna fish

Vegetables
Beetroot Beet, boiled
Tomato Tomato cooked in sauce
Carrot Carrots boiled
Green peas Green peas, boiled
Pump kin/butternut Pump kin/ butternut, boiled
Maraka / Makgomane Traditional /Melon boiled with added
salt/milk
Lerotse Traditional Melon boiled
Spinach Spinach leaves boiled
Rape/Cho molia Brassica oleracea leaves
Morogo wa dinawa Cowpea leaves boiled/ dried /boiled
Broccoli Broccoli boiled
Cauliflower Cauliflower boiled
Lettuce Iceberg lettuce salad
Cucu mber Cucu mber raw, sliced
Avocado Avocado, raw, sliced
Pepper (Green/red/yellow) Peppers sliced, cooked
Mushroom Oyster mushroom cooked
Cabbage Cabbage shredded, boiled
Tswii Water lily tuber

Frui t
Orange Orange fru it eaten fresh
Naraki/ Naatjee Mandarins eaten fresh
Grapefru it Grapefru it eaten fresh
Pawpaw Pawpaw fru it eaten fresh
Mango Mango fruit eaten fresh
Banana Banana eaten fresh
Apple Apple eaten fresh
Peach Peach eaten fresh
Pears Pears eaten fresh
Watermelon Watermelon eaten fresh
Grape Grapes eaten fresh
Morula Traditional fru it (fro m maru la
tree) eaten fresh
Dried Fruit Mixture of raisins, prunes
Canned Fruit mix of diced or sliced fruit and
sometimes syrup
Strawberry Strawberry fru it eaten fresh
Pineapple Pineapple Fruit eaten fresh
Mmilo Wild fruit eaten dried
Moretlwa Wild berries

Sweets fats oils and condi ments
Sweets/candy confection made fro m a
concentrated solution of sugar
Sugar (e.g. added to tea) White or brown sugar
Cookies/biscuits Small, flat, baked treat
Cake sweet and enriched baked dessert
Buns/scones/muffins Small British quick bread of
Scottish origin
Peanut Butter Ground peanut paste
Margarine Butter substitute made fro m
Vegetable fat
Jam Concentrated preparation of fruits
and sugar
Cream cheese Soft, mild-tasting, white cheese
Salad Dressing Sauce for salads
Mayonnaise stable emulsion of o il, egg yolk
and either vinegar or lemon ju ice
Cheese snacks
Potato chips (fries) Irish potato deep fried
Savory meat pies Meat pastry baked
Jelly Desserts made with sweetened
and flavoured gelatin
Custard cooked mixture of milk or cream
and egg yolk used as dessert or
dessert sauce

Beverages
Soda/Sweetened beverages Beverages with added sugar or
carbonated
Diet Sodas Carbonated beverages with
sweetener
Fruit juices Juices naturally fro m the fru it
Fruit drinks Liquids made fro m fru it flavors
Beer Alcoholic beverage made fro m
brewing and fermentation of
sugars, mainly derived fro m
malted cereal grains
Ru m/Whiskey/Gin Alcoholic beverages made fro m
sugarcane by-products (ru m),
grain mash (wh iskey) and
juniper berries (g in)
Wine Alcoholic beverage made fro m
fermented grapes
Chibuku Opaque beer made fro m barley
or sorghum
Coffee Brewed beverage with a dark,
slightly acidic flavour prepared
fro m roasted coffee beans
Tea (Five Roses) Aromatic beverage prepared
fro m the cured leaves by
combination with hot or boiling
water or b lack tea
Herbal Tea (Rooibos) Infusions of oxid ized rooibos
leaves with hot or boiling water
Chocolate drin k Hot beverage made of shaved
chocolate, melted chocolate or
cocoa powder, heated milk or
water, and sugar
View publication stats

DietandCVDriskfactorsinBotswana Kwape PhDthesis2012

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

DietandCVDriskfactorsinBotswana Kwape PhDthesis2012

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Diet and cardiovascular disease risk factors in Botswana_PhD thesis

Book · January 2012

The user has requested enhancement of the downloaded file.

Lemogang Daniel Kwape

BS. Dietetics, Texas Southern University (Houston, Texas USA)

A thesis presented for the degree of

in the Public Health Nutrition Research Group, Institute of Applied Health

by myself with assistance from individuals and organisations acknowledged here.

Bantle Modibedi and Tapologo Ramotswaiso were involved in recruitment and

analysed by staff at Diagnofirm Medical Laboratories in Gaborone Botswana. The Food

Frequency Questionnaire was analysed by Dr Maria Jackson, University of Jamaica at

Lemogang Daniel Kwape

Acknow ledgements vii

List of abbreviations viii

1.2 Risk factors fo r cardiovascular disease 6

1.3 Nutrition transition 23

1.4 Diet and risk of CVD 23

1.5 Antiretroviral medication and ca rdiovascular diseases 33

1.6 Diet and CVD in Botswana 34

1.7 Aims and objectives 40

2.1 Introduction to Bo tswana 42

2.2 Roles in the study 44

2.3 Ethical considerations 45

2.5 Study participants 46

2.6 Sample size 47

2.8 Biophysical measurements 54

2.9 Collection of biological samples and initial bio chemistry 58

2.10 Statistical Analyses 61

3 Diet and CVD risk factors 64

3.2 Statistical methods 65

3.3 Socio-demographic characteristics 70

3.4 Anthropometric and physical chara cteristics of participants 79

3.5 Lipids, Lipoproteins and glucose measurements of participants 87

3.6 Dietary Data 91

3.7 Discussion 126

4 Associations between diet and CVD risk factors 138

4.1 Statistical methods 139

4.11 Discussion 198

5 Diet and risk of CVD/diabetes/hypertension 208

5.1 Introduction 208

5.2 Statistical methods 211

5.3 Comparison between participants with CVD, hypertension or diabetes 214

5.4 Characteristics of “healthy” and “diseased ” participants 219

5.5 Risk of CVD/diabetes/hypertension by dietary factors 226

5.6 Discussion 236

6.1 Strengths 242

6.2 Limitations 243

6.3 Major findings 245

6.4 Future research wo rk focus 257

6.5 Public health perspective 261

6.6 Conclusion 263

Appendix 1 Ethics permit i

Appendix 2 Information sheet (English) i

Appendix 2 Information sheet (Setswana) ii

Appendix 3 Questionnaire iii

Appendix 4 Description of food items listed in the FFQ xxviii

Appendix 5 Publication xxxii

Carbohydrates intake was significantly associated with increased risk of disease.

I am grateful to Professor Geraldine McNeill for her guidance and encouragement

I am immensely grateful to the Department of Research Science and Technology,

Figure 1.1 Complex lesion and thrombosis in atherosclerosis 3