UvA-DARE (Digital Academic Repository)

Cardiovascular health in urban Suriname The Healthy Life in Suriname (HELISUR) study Diemer, F.S.

Publication date 2018 Document Version Final published version License Other Link to publication

Citation for published version (APA): Diemer, F. S. (2018). Cardiovascular health in urban Suriname: The Healthy Life in Suriname (HELISUR) study.

General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

Download date:05 Oct 2021

Cardiovascular health in urban Suriname:

The Healthy Life in Suriname (HELISUR) study Layout and printed by: Optima Grafische Communicatie (www.ogc.nl)

ISBN: 978-94-6361-181-7 Cardiovascular health in urban Suriname:

The Healthy Life in Suriname (HELISUR) study

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Universiteit van Amsterdam op gezag van de Rector Magnificus prof. dr. ir. K.I.J. Maex ten overstaan van een door het College voor Promoties ingestelde commissie, in het openbaar te verdedigen in de Agnietenkapel op donderdag 13 december 2018, te 10:00 uur door Frederieke Sophie Diemer geboren te De Bilt Promotiecommissie:

Promotor: prof. dr. R.J.G. Peters AMC - UvA

Copromotores: dr. L.M. Brewster Anton de Kom Universiteit van Suriname dr. L.M.W. Nahar-van Venrooij Anton de Kom Universiteit van Suriname

Overige leden: prof. dr. K. Stronks AMC - UvA prof. dr. A.E. Kunst AMC - UvA prof. dr. H.C.P.M. van Weert AMC - UvA prof. dr. V.W.V. Jaddoe Erasmus Universiteit Rotterdam dr. N.R. Bindraban AMC - UvA dr. E.P. Moll van Charante AMC - UvA dr. I.G.M. van Valkengoed AMC - UvA

Faculteit der Geneeskunde

Financial support by the Dutch Heart Foundation for the publication of this thesis is gratefully acknowledged. Contents

Chapter 1 General introduction and outline of the thesis 9

Part I Protocol and feasibility 25 Chapter 2 Exploring cardiovascular health: the Healthy Life in Suriname 27 (HELISUR) study. A protocol of a cross-sectional study BMJ Open 2014; 4: e006380 Chapter 3 Assessing population cardiovascular risk with advanced 39 hemodynamics: the Healthy Life in Suriname (HELISUR) feasibility study Pan American Journal of Public Health 2017; 41: e46

Part II Ethnic differences in cardiovascular health 53 Chapter 4 Hypertension and cardiovascular risk profile in urban Suriname: the 55 HELISUR study American Journal of Hypertension 2017; 30(11):1133-1140 Chapter 5 Body composition and cardiovascular risk in high-risk ethnic groups 71 Clinical Nutrition 2017; doi: 10.1016/j.clnu.2017.11.012 Chapter 6 Physical activity and obesity: is there a difference in the association 89 between the Asian- and African-Surinamese adult population? Ethnicity and Health 2017; 1:1-13 Chapter 7 Aortic pulse wave velocity in an Asian and African ancestry 109 population Submitted

Part III A comparison with Surinamese in the Netherlands 125 Chapter 8 Hypertension in Surinamese living in a middle- vs. high-income 127 country: the HELISUR vs. HELIUS studies Submitted

Chapter 9 General discussion 147

Appendix Summary 171 Nederlandse samenvatting 175 List of publications 179 Contributing authors 181 Author contributions 183 PhD portfolio 185 Dankwoord 187 About the author 191

Voor Suriname

Chapter 1 General introduction and outline of the thesis

Introduction 11

Introduction 1 Cardiovascular disease (CVD) is the leading cause of death worldwide.1 Annually, 17.7 million people die from CVD, which is equal to the entire population of the Nether- lands.1,2 However, the vast majority (>80%) of all CVD-related deaths occurs in low- and middle-income countries (LMIC)3, and, in contrast to the declining trend seen in high-income countries (HIC), the burden of CVD in LMIC continues to rise.4,5 In order to combat this epidemic, large-scale preventive measures need to be taken. Unfortunately, in many LMIC, the lack of data hampers the design and implementation of effective prevention strategies. Both the United Nations General Assembly as well as the World Heart Federation therefore recognized the importance of collecting reliable region- specific data on risk factors for cardiovascular disease.6,7 The work presented in this thesis is a response to these calls.

Well-established differences in cardiovascular risk and CVD exist between ethnic groups. People of Asian ancestry are disproportionately affected by coronary heart disease, whereas Africans have an increased risk of hypertension-related organ dam- age, such as left ventricular hypertrophy and stroke.8,9,10 The reasons behind these ethnic differences in CVD are complex and remain incompletely elucidated. In addition, most studies on ethnic differences in CVD are conducted in high-income countries in (descendants of) migrants from African, Asian, and Caribbean countries. They may not be applicable for LMIC where socioeconomic and environmental conditions as well as health care systems and policies are different.11

Suriname is a middle-income country in South-America with a multi-ethnic population of mainly African and Asian ancestry. Despite large similarities in socioeconomic status between the ethnic groups, substantial ethnic differences exist in cardiovascular mor- tality.12 Differences in underlying cardiovascular risk factors and asymptomatic organ damage might explain these ethnic disparities in CVD mortality. Yet, data on cardiovas- cular risk factors and asymptomatic organ damage were virtually absent at the start of our research.

The main aim of this thesis is to expand the evidence on cardiovascular health and related risk factors in urban Suriname, with a primary focus on ethnicity. Ethnicity is a multi-faceted concept that includes the different dimensions that defines a person’s identity. More specifically, members of an ethnic group have a shared history, ancestry, and identity, and share characteristics such as a geographical affiliation, culture and tra- ditions, language, and religion.13 This is in contrast to ‘race’, which merely refers to com- mon hereditary physical features, such as skin, hair and eye colour.14 Ethnic variations in 12 Chapter 1

cardiovascular risk may therefore relate to underlying differences in biology, behaviour, or the exposure to environmental factors. Therefore, collecting data on ethnicity might be important to monitor and understand differences in outcomes for different popula- tion groups. There are different ways to classify people into ethnic groups, for example, by self-reported ethnicity, observer-selected ethnicity, country of birth, nationality or surname.15 Self-report, which reflects how individuals view themselves, is the generally agreed-upon best way to define a person’s ethnic identity.16,17 In addition, self-report respects “individual dignity” by allowing an individual to decide how he or she classi- fies himself or herself as opposed to classification being assigned by another person.18 Therefore, in this thesis, self-reported ethnicity of the participant was used.

This introductory chapter is composed of two parts. In the first part, the study area and the study aims are outlined. The second part describes the overall cardiovascular risk profile and ethnic differences in cardiovascular health, with a focus on hypertension, obesity, and increased pulse wave velocity as a manifestation of early asymptomatic organ damage. In the third part, a comparison with Surinamese living in the Netherlands is made to gain insight into how to reduce the hypertension burden in urban Suriname. This chapter ends with a general outline of the thesis.

Part I | Protocol and feasibility

Suriname The Republic of Suriname is situated in the north-eastern part of South-America, and borders north to the Atlantic Ocean, south to Brazil, east to French Guyana, and west to Guyana. Ninety percent of the country is covered by tropical rainforest.19 The popula- tion consists of roughly 570,000 people, with almost half of its citizens residing in Para- maribo, the capital of Suriname.20,21 With a gross national income per capita of $6,990 US dollars in 2016, Suriname is classified as a middle-income country.20 However, since 2015, Suriname is in steep recession with inflation rates rising from 3.4% in 2014 to almost 60% by year-end 2016.22

The present demographic composition of Suriname is among the most varied in the world, as a result of historical migrations of people from different geographic areas. From the 17th to 19th century, enslaved people from predominantly West-Africa were brought to Suriname to work on coffee and sugar plantations. The working and living conditions were harsh and a large group of enslaved men and women fled the plantations and created new settlements in the rain forest. They continued to live in the interior and are now referred to as Maroons. Only recently Maroons are migrating to the urban areas for Introduction 13 work and educational opportunities. The descendants of African enslaved people who remained in the capital are referred to as the Creole. After the abolishment of slavery in 1 1863, people from India and the island of Java migrated to Suriname as indentured la- bourers.23 Nowadays, the population of Suriname is among the most varied in the world, with the capital being populated by 26% Creole, 23% South-Asians, 16% Maroons, and 10% Indonesians.21 The remaining 26% are mostly people of mixed ancestry, Chinese or Amerindians, the original inhabitants of Suriname.21

As a consequence of the independence of Suriname in 1975 and military rule in the 1980s, about half of the Surinamese population migrated to the Netherlands. The Suri- namese are one of the largest groups with a migration background in the Netherlands.24

Cardiovascular health in Suriname As most other LMIC, Suriname is going through an epidemiologic transition, from predominantly infectious diseases to predominantly non-communicable diseases.25,26 Nowadays, CVD are the leading cause of mortality among the Surinamese population, accounting for 27% of all deaths.27 The underlying factors for this transition are indica- tive of economic development and led to the adoption of a different, more westernized lifestyle, particularly in urban areas. Such lifestyle is perceived to be more desirable and modern and marked by an increase in the consumption of processed foods high in fat, salt, and sugar, a decrease in physical activity with sedentary lifestyles, increased tobacco smoking and harmful use of alcoholic beverages.26 These are in turn strongly linked to the development of cardiovascular risk factors, such as hypertension, diabetes mellitus type 2, dyslipidaemia, and obesity.26

To combat the high cardiovascular mortality in Suriname, it is imperative to develop an extensive evidence base, consisting of multiple large-scale studies. Therefore, we de- signed the Healthy Life in Suriname (HELISUR) study, an observational cross-sectional study on cardiovascular health among different ethnic groups in the capital of Suriname. The primary aim of the study was to assess the cardiovascular risk profile of the urban population of Suriname. As a secondary outcome, ethnic differences in cardiovascular risk were assessed. Figure 1 illustrates the natural history of CVD, together with the cardiovascular health data collected within the HELISUR study.

In the first part of this thesis, the protocol of the HELISUR study is illustrated, including the study aims, (Chapter 2) and the feasibility of the HELISUR study is tested (Chapter 3). 14 Chapter 1

CHD Established CVD Stroke Severe CKD Peripheral artery disease Deep vein thrombosis

Increased arterial stiffness Asymptomatic organ Left ventricular hypertrophy damage Microalbuminuria Asymptomatic CKD

Hypertension Diabetes mellitus Traditional risk Dyslipidemia factors Obesity (general and central) Adverse body composition

Prehypertension Prediabetes Primordial risk Borderline dyslipidemia factors Overweight

Physical inactivity Tobacco smoking Alcohol misuse Lifestyle Unhealthy diet Medication use

Non-modifiable Age, sex, ethnicity, risk factors family history of CVD

Figure 1. Schematic overview of the natural history of CVD. Variables on the right are collected within the HELISUR study. The variables in bold are explored in this thesis. CVD, cardiovascular disease; CHD, coronary heart disease; CKD, chronic kidney disease.

Part II | Ethnic differences in cardiovascular health

Marked ethnic differences exist in CVD mortality. People of Asian ancestry have a two- fold increased CHD risk compared to people of European ancestry. On the other hand, people of African ancestry enjoy significant protection of CHD, but have substantial higher stroke mortality rates. Reasons behind these ethnic differences in mortality are multifactorial and have not been fully explained yet.28,29,30

Overall and ethnic-specific cardiovascular risk profile In Suriname, substantial ethnic differences exist in cardiovascular mortality as well. For example, 37% of deaths in South-Asians were due to CVD, whereas this was only 10% in Maroons.12 Differences in cardiovascular risk factors may in part be responsible for the difference in CVD mortality. These risk factors are modifiable and can be controlled Introduction 15 or their effect may be reduced by making long-term lifestyle changes. However, little is known regarding the cardiovascular risk profile, including asymptomatic organ dam- 1 age, of urban Surinamese. In Chapter 4, the epidemiology of cardiovascular risk factors, asymptomatic organ damage, and established CVD in the urban population of Suriname is described, and ethnic differences in cardiovascular health status are assessed.

Obesity Obesity is an important risk factor for CVD, promoting atherosclerosis.31,32 Once consid- ered a problem only in HIC, obesity is now dramatically on the rise in LMIC, particularly in urban settings, as urban living is usually associated with a lower energy expenditure and a higher intake of energy-dense foods compared to rural life.33,34 There are different ways to measure obesity. Most frequently used is the body mass index (BMI), computed as height/weight2, and waist measures, such as waist circumference, waist-hip ratio, and waist-to-height ratio. Waist measures capture the distribution of fat rather than the overall amount of adipose tissue in the body (of which BMI is a proxy). In addition to measuring the fat distribution, one can also distinguish between fat mass and fat-free mass.

Previous studies showed that the fat distribution and body composition differs across ethnic groups. For example, at the same level of BMI, Asian populations have more total body fat and more abdominal fat, whereas African populations showed higher fat-free mass and body fat compared to Caucasians.35,36,37 As a consequence, lower cut-off values for BMI, waist circumference en waist-hip ratio were formulated that correspond to an increased risk for CVD.35,38,39

It is vital to accurately identify people at high CVD risk. A recent study in Suriname showed that waist circumference had more discriminatory power than BMI in the rela- tion with hypertension, diabetes, and an adverse cardiometabolic risk profile.40 How- ever, it remains unclear if more complex body composition measures such as fat mass or low fat-free mass together with a high fat mass (i.e. sarcopenic obesity41) are superior to the simple anthropometric measures, such as BMI and waist measures. Therefore, in Chapter 5, we examined whether body composition measures are superior to the simple anthropometric measures in the association with CVD risk.

Physical activity is a key element in the management of obesity.42 International guide- lines recommend that adults should do at least 150 minutes of moderate-intensity or 75 minutes of vigorous-intensity physical activity throughout the week.43 However, uncertainties remain regarding the importance of the individual physical activity char- acteristics, i.e. domain, duration or intensity that determines optimal weight control. In 16 Chapter 1

addition, although these recommendations apply to the population at large, irrespec- tive of sex or ethnicity, these recommendations are based on studies conducted in predominantly Caucasian populations living in HIC.44 Yet, several small studies showed that the response to physical activity interventions might be different across ethnic groups.45,46 Furthermore, evidence suggests that different exercise recommendations are required when different risk factors are targeted.47 Given that the prevalence of cardiovascular risk factors differs among ethnic groups further supports the potential need for ethnic-specific physical activity recommendations.48 In Chapter 6, the associa- tion between obesity and physical activity characteristics (i.e. domain, duration, and intensity) is explored within an Asian and African ancestry population.

Pulse wave velocity Evidence suggests that differences between ethnic groups are not fully explained by traditional cardiovascular risk factors.28,29,30,49 One way to address the issue of ethnic variability in CVD risk is to shift the focus from risk assessment to diagnosing early asymptomatic organ damage, for example in blood vessels.

Increased arterial stiffness, estimated non-invasively by pulse wave velocity (PWV), has emerged as an important predictor of cardiovascular events.50,51,52 It is a type of asymptomatic organ damage, characterized by a reduced arterial compliance.53,54 Previ- ous studies showed that arterial stiffness progresses with age and can be accelerated by different cardiovascular risk factors, hypertension being one of the most important.55 Given that African and Asian ancestry populations bear a disproportionately high car- diovascular risk factor burden, the association between PWV and cardiovascular risk factors may differ within these ethnic groups. Furthermore, there are conflicting results regarding ethnic differences in PWV after adjustment for cardiovascular risk factors.53 Therefore, in Chapter 7, we studied differences in PWV between African and Asian- Surinamese, taking traditional cardiovascular risk factors into account.

Part III | A comparison with Surinamese in the Netherlands

Hypertension is the main preventable cause of CVD mortality worldwide.53,56 With many effective and inexpensive treatments available, hypertension control and prevention of subsequent hypertension-related diseases should be achievable. Nevertheless, hy- pertension continues to rise in LMIC with little improvement in the levels of awareness, treatment, and control.57,58 In contrast to LMIC, the prevalence of hypertension slightly decreased and the level of hypertension control substantially increased in HIC.57,58 This indicates there is still considerable room for improvement for LMIC. A comparison with Introduction 17 a HIC, especially within populations with similar ethnicity, may give insight on where that “room” is, thereby providing a scientific basis for the development of public health 1 interventions.

The unprecedented mass migration of Surinamese to the Netherlands offers a unique opportunity to study differences in the prevalence, awareness, treatment, and control of hypertension between LMIC and HIC. Therefore, in Chapter 8, we compare hypertension prevalence, awareness, treatment and control between Surinamese living in Suriname and first and second generation Surinamese migrants living in the Netherlands. Differ- ences in the underlying determinants of hypertension between the two countries were explored as a secondary aim.

The HELISUR vs HELIUS study The HELISUR study was based on the Health Life in an Urban Setting (HELIUS) study, using similar methodology.59 The HELIUS study is conducted in Amsterdam, the Netherlands among six ethnic groups, including among individuals with a Surinamese background. Surinamese living in the Netherlands were identified based on their country of birth.60 A person was identified as Surinamese if he or she fulfils one of two criteria: he or she was born in Suriname and has at least one parent who was born in Suriname (first genera- tion); or he or she was born in the Netherlands and both parents were born in Suriname (second generation). Persons aged 18 to 70 years were randomly sampled from the municipality registry of Amsterdam. The registry contains data on the country of birth of the subject and his/her parents, which made stratification by ethnicity possible.59 In this paper, further classification into ethnic subgroup (i.e. African or South-Asian) was based on self-reported ethnicity. However, a person who reported to be of mixed/other origin which included an African ancestry term (e.g. Creole, Maroon, or African) was classified as African, whereas in other chapters presented in this thesis such a person would be classified as mixed/other.

Outline of the thesis

The first part of this thesis focuses on the HELISUR study protocol (Chapter 2) and the feasibility of the HELISUR study (Chapter 3).

The focus in the second part of the thesis is on ethnic differences in cardiovascular health. Chapter 4 presents the epidemiology of cardiovascular risk factors, asymp- tomatic organ damage, and established CVD in the HELISUR study population. It also highlights ethnic differences in the cardiovascular risk profile of urban Surinamese. 18 Chapter 1

Chapter 5 deals with ethnic differences in body composition and the association with cardiovascular risk. Chapter 6 aimed to describe ethnic differences in physical activity and its relation with obesity. In Chapter 7, ethnic differences in pulse wave velocity are studied, taking traditional cardiovascular risk factors into account.

In the third part of this thesis, a comparison with Surinamese living in the Netherlands is made to gain insight into how to reduce the hypertension burden in urban Suriname. In Chapter 8, we compared hypertension prevalence, awareness, treatment, and control in Surinamese from Suriname (HELISUR) with migrated Surinamese living in the Neth- erlands (HELIUS), and compared the underlying determinants of hypertension in both populations.

The general discussion of the main findings is presented in Chapter 9, which also high- lights the implications for practice and future research. Introduction 19

References 1. World Health Organization. Causes of death. 2008. Available at: http://www.who.int/ 1 healthinfo/ global_burden_disease/cod_2008_sources_methods.pdf. 2. Centraal Bureau voor de Statistiek (CBS). Bevolkingsteller. 2017. Available at: https://www. cbs.nl/nl-nl/visualisaties/bevolkingsteller. 3. World Health Organization. Global action plan for the prevention of noncommunicable diseases 2013-2020. Geneva: WHO; 2013. 4. Global Burden of Disease 2013 Mortality and Causes of Death Collaborators. Global, re- gional, and national age–sex specific all-cause andcause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015;385:117-171. 5. Mathers C, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Medicine. 2006;3(11):2011-2030. 6. United Nations General Assembly. Political declaration of the high-level meeting of the General Assembly on the prevention and control of non-communicable diseases. 2011. Available at: http://www.who.int/nmh/events/. 7. Yusuf S, Wood D, Ralston J, Reddy K. The World Heart Federation’s vision for worldwide cardiovascular disease prevention. Lancet. 2015;386(9991):399-402. 8. Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2095-2128. 9. Agyemang C, Kieft S, Snijder M, et al. Hypertension control in a large multi-ethnic cohort in Amsterdam, The Netherlands: The HELIUS study. Int J Cardiol. 2015;183:180-189. 10. Chaturvedi N. Ethnic differences in cardiovascular disease. Heart. 2003;89(6):681-686. 11. Mills A. Health care systems in low- and middle-income countries. The New England Journal of Medicine. 2014;370:552-557. 12. Punwasi W. Doodsoorzaken in Suriname 2002-2007. Paramaribo: Ministry of Health; 2007. 13. Simon P. Collecting ethnic statistics in Europe: a review. Ethnic and Racial Studies. 2012;35:1366-1391. 14. Senior P, Bhopal R. Ethnicity as a variable in epidemiological research. BMJ. 1994;309:327- 330. 15. Sandefur G, Campbell M, Eggerling-Boeck J. Chapter 2. Racial and ethnic identification, official classifications, and health disparities.Critical perspectives on racial and ethnic differ- ences in health in late life. Washington: National Academies Press (US); 2004. 16. Higgins P, Taylor E. Measuring racial and ethnic disparities in health care: efforts to improve data collection. Princeton, NJ: Mathematica Policy Research; 2009. 17. Institute of Medicine. Chapter 5: Improving data collection across the health care system. Race, ethnicity, and language data: standardization for health care quality improvement. Washington DC: The National Academies Press; 2009. 18. Office of Management and Budget. Recommendations from the Interagency Committee for the Review of the Racial and Ethnic Standards to the Office of Management and Budget con- cerning changes to the standards for the classification of federal data on race and ethnicity. Federal register. 1997a:36873-36946. 20 Chapter 1

19. Ministry of Labour, Technological Development and Environment. Suriname Second National Communication to the United Nations Framework Convention on Climate Change. Paramaribo 2013. 20. The World Bank Group. Suriname. 2016. Available at: https://data.worldbank.org/country/ Suriname. 21. General Bureau of Statistics Suriname. Resultaten van de achtste (8e) volks- en woningtelling in Suriname (Volume I) - Demografische en sociale karakteristieken en migratie. Paramaribo 2012. 22. MacDonald S. Suriname’s Economic Crisis. Center for Strategic and International Studies. April 20, 2017. Available at: https://www.csis.org/analysis/surinames-economic-crisis. 23. Mans D, Ganga D, Kartapawiro J. Meeting of the minds: traditional and herbal medicine in multiethnic Suriname. 2017. Available at: https://www.intechopen.com/books/aromatic- and-medicinal-plants-back-to-nature/meeting-of-the-minds-traditional-herbal-medicine- in-multiethnic-suriname. 24. Centraal Bureau voor de Statistiek. Bevolking naar migratieachtergrond. 2016. Available at: https://www.cbs.nl/nl-nl/achtergrond/2016/47/bevolking-naar-migratieachtergrond. 25. Lopez A, Mathers C, Ezzati M, Jamison D, Murray C. Global and regional burden of disease and risk factors, 2001: systematic analysis of population health data. Lancet. 2006;367(9524):1747-1757. 26. Yusuf S, Reddy S, Ounpuu S, Anand S. Global burden of cardiovascular diseases: part I: gen- eral considerations, the epidemiologic transition, risk factors, and impact of urbanization. Circulation. 2001;104:2746-2753. 27. Punwasi W. Doodsoorzaken Suriname 2010-2011. Paramaribo: Ministry of Health; 2012. 28. Tillin T, Hughes A, Mayet J, Whincup P, Sattar N, Forouhi N. The relationship between metabolic risk factors and incident cardiovascular disease in Europeans, South Asians, and African Caribbeans: SABRE (Southall and Brent Revisited) - a prospective population-based study. Journal of the American College of Cardiology. 2013;61(17):1777-1786. 29. Yusuf S, Hawken S, Ounpuu S, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet. 2004;364(9438):937-952. 30. Jain P, Kooner J, Raval U, Lahiri A. Prevalence of coronary artery calcium scores and silent myocardial ischaemia was similar in Indian Asians and European whites in a cross-sectional study of asymptomatic subjects from a U.K. population (LOLIPOP-IPC). Journal of Nuclear Cardiology. 2011;18(3):435-442. 31. van Rooy MJ, Pretorius E. Obesity, hypertension and hypercholesterolemia as risk factors for atherosclerosis leading to ischemic events. Current Medical Chemistry. 2014;21(19):2121- 2129. 32. Alpert M, Omran J, Bostick B. Effects of obesity on cardiovascular hemodynamics, cardiac morphology, and ventricular function. Current Obesity Reports. 2016;5(4):424-434. 33. NCD Risk Factor Collaboration. Worldwide trends in body-mass index, underweight, overweight, and obesity from 1975 to 2016: a pooled analysis of 2416 population- based measurement studies in 128·9 million children, adolescents, and adults. Lancet. 2017;390(10113):2627-2642. 34. Caballero B. The Global Epidemic of Obesity: An Overview. Epidemiologic Reviews. 2007;29(1):1-5. Introduction 21

35. National Institute for Health and Care Excellence. Assessing body mass index and waist circumference thresholds for intervening to prevent ill health and premature death among adults from black, Asian and other minority ethnic groups in the UK. 2013. Available at: 1 https://www.nice.org.uk/guidance/ph46. 36. Rao G, Powell-Wiley T, Ancheta I, et al. Identification of Obesity and Cardiovascular Risk in Ethnically and Racially Diverse Populations: A Scientific Statement From the American Heart Association. Circulation. 2015;132(5):457-472. 37. Heymsfield S, Peterson C, Thomas D, Heo M, Schuna J. Why are there race/ethnic differences in adult body mass index–adiposity relationships? A quantitative critical review. Obesity reviews. 2016;17(3):262-275. 38. World Health Organization. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363:157-163. 39. Alberti K, Zimmet P, Shaw J. International Diabetes Federation: a consensus on Type 2 diabetes prevention. Diabetes Medicine. 2007;24(5):451-463. 40. Krishnadath I, Toelsie J, Nahar-van Venrooij L, Hofman A, Jaddoe V. Ethnic and sex-specific cut-off values for adult obesity in the Suriname Health Study. Obesity Research and Clinical Practice. 2016:pii: S1871-403X(16)30111-9. 41. Prado C, Wells J, Smith S, Stephan B, Siervo M. Sarcopenic obesity: a critical appraisal of the current evidence. Clinical Nutrition. 2012;31(5):583-601. 42. Shook R, Hand G, Drenowatz C, et al. Low levels of physical activity are associated with dysregulation of energy intake and fat mass gain over 1 year. American Journal of Clinical Nutrition. 2015;102(6):1332-1338. 43. World Health Organization. Global recommendations on physical activity for health. Geneva 2010. 44. Gill J, Celis-Morales C, Ghouri N. Physical activity, ethnicity and cardio-metabolic health: does one size fits all? Atherosclerosis. 2014;232(2):319-333. 45. Celis-Morales C, Ghouri N, Bailey M, Sattar N, Gill J. Should physical activity recommen- dations be ethnicity-specific? Evidence from a cross-sectional study of South Asian and European men. PlOS One. 2013;8(12):e82568. 46. Misra A, Nigam P, Hills A, et al. Consensus physical activity guidelines for Asian Indians. Diabetes Technology & Therapeutics. 2012;14(1):83-98. 47. Hansen D, Dendale P, Coninx K, et al. The European Association of Preventive Cardiology Ex- ercise Prescription in Everyday Practice and Rehabilitative Training (EXPERT) tool: A digital training and decision support system for optimized exercise prescription in cardiovascular disease. European Journal of Preventive Cardiology. 2017;24(10):1017-1031. 48. Chiu M, Austin P, Manuel D, Tu J. Comparison of cardiovascular risk profiles among ethnic groups using population health surveys between 1996 and 2007. Canadian Medical As- sociation Journal. 2010;182(8):301-310. 49. Rabanal K, Meyer H, Tell G, et al. Can traditional risk factors explain the higher risk of cardio- vascular disease in South Asians compared to Europeans in Norway and New Zealand? Two cohort studies. BMJ Open. 2017;7(12):e016819. 50. Ben-Shlomo Y, Spears M, Boustred C, et al. Aortic pulse wave velocity improves cardiovascu- lar event prediction: an individual participant meta-analysis of prospective observational data from 17,635 subjects. J Am Coll Cardiol. 2014;63(7):636-646. 22 Chapter 1

51. Maroules C, Khera A, Ayers C, et al. Cardiovascular outcome associations among cardiovas- cular magnetic resonance measures of arterial stiffness: the Dallas heart study.J Cardiovasc Magn Reson. 2014;16:33-42. 52. Laurent S. Defining vascular aging and cardiovascular risk. J Hypertens. 2012;30(S3-8). 53. Faconti L, Nanino E, Mills C, Cruickshank K. Do arterial stiffness and wave reflec- tion underlie cardiovascular risk in ethnic minorities? JRSM Cardiovasc Disease. 2016(5):2048004016661679. 54. Cavalcante J, Lima J, Redheuil A, Al-Mallah M. Aortic stiffness: current understanding and future directions. J Am Coll Cardiol. 2011;57(14):1511-1522. 55. O’Rourke M, Hashimoto J. Mechanical factors in arterial aging: a clinical perspective. J Am Coll Cardiol. 2007;50(1):1-13. 56. Feigin VRG, Naghavi M, Parmar P, et al. Global burden of stroke and risk factors in 188 countries, during 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013. The Lancet Neurology. 2016;15(9):913-924. 57. Chow C, Teo K, Rangarajan S, et al. Prevalence, awareness, treatment, and control of hyper- tension in rural and urban communities in high-, middle-, and low-income countries. JAMA. 2013;310(9):959-968. 58. Mills K, Bundy J, Kelly T, et al. Global Disparities of Hypertension Prevalence and Con- trol: A Systematic Analysis of Population-Based Studies From 90 Countries. Circulation. 2016;134(6):441-450. 59. Stronks K, Snijder M, Peters R, Prins M, Schene A, Zwinderman A. Unravelling the impact of ethnicity on health in Europe: the HELIUS study. BMC Public Health. 2013;13:402. 60. Stronks K, Kulu-Glasgow I, Agyemang C. The utility of ‘country of birth’ for the classi- fication of ethnic groups in health research: the Dutch experience. Ethnicity and Health. 2009;14(3):255-269.

Part I Protocol and feasibility

Chapter 2 Exploring cardiovascular health: the Healthy Life in Suriname (HELISUR) study. A protocol of a cross-sectional study

Frederieke S. Diemer, Jet Q. Aartman, Fares A. Karamat, Sergio M. Baldew, Ameerani V. Jarbandhan, Gert A. van Montfrans, Glenn P. Oehlers, Lizzy M. Brewster BMJ Open 2014; 4: e006380 28 Chapter 2

Abstract

Introduction: Obesity, hypertension and diabetes are on a dramatic rise in low-income and middle- income countries, and this foretells an overwhelming increase in chronic disease burden from cardiovascular disease. Therefore, rapid action should be taken through preventive population-based programmes. However, in these regions, data on the population distribution of cardiovascular risk factors, and of intermediate and final end points for cardiovascular disease are scarce. The Healthy Life in Suriname (HELISUR) study is a cardiovascular population study in Suriname, which is part of the Caribbean Community. The HELISUR study is dedicated to provide data on risk factors and preva- lent cardiovascular disease in the multiethnic population, which is mainly of African and Asian descent.

Methods and analysis: In a cross-sectional, observational population-based setting, a random representative sample of 1800 citizens aged between 18 and 70 years will be selected using a cluster household sampling method. Self-reported demographic, socioeconomic and (cardio- vascular) health-related data will be collected. Physical examination will include the assessment of cardiovascular risk factors and prevalent cardiovascular disease. In addi- tion, we will study cardiovascular haemodynamics non-invasively, as a novel intermedi- ate outcome. Finally, fasting blood and overnight urine samples will be collected to monitor cardiometabolic risk factors. The main outcome will be descriptive in reporting the prevalence of risk factors and measures of (sub)clinical end organ damage, stratified for ethnicity and sex-age groups.

Ethics and dissemination: Ethical approval has been obtained from the State Secretary of Health. Data analysis and manuscript submission are scheduled for 2016. Findings will be disseminated in peer-reviewed journals, and at national, regional and international scientific meetings. Importantly, data will be presented to Surinamese policymakers and healthcare work- ers, to develop preventive strategies to combat the rapid rise of cardiovascular disease. HELISUR study protocol 29

Introduction

Cardiovascular disease (CVD) is rapidly increasing in low-income and middle-income countries (LMIC).1 At present, the number of cardiovascular deaths is responsible for the greatest proportion of the overall burden of disease, and projections on mortality 2 attributable to CVD forecast even higher death rates in 2030 (Figure 1). Awareness is 2 growing that preventive action based on country-specific risk factor data is imperative to combat this burgeoning epidemic.3

Figure 1. Cardiovascular death by World Bank’s income categories. The number of projected deaths in millions due to cardiovascular disease (y-axis) is depicted in high-income, middle-income and low-income countries for 2004, 2015 and 2030 (x-axis).2 This figure indicates that the projected increase in mortality is the highest in middle-income countries such as Suriname.

Suriname is a middle-income country located in the northeastern region of South America (Figure 2). The country is populated by different ethnic groups, including people of South Asian, African, Indonesian and Chinese origin, as well as a relatively small indigenous population of Amerindians.4 As in other LMIC, CVD is a rapidly ris- ing cause of premature mortality.5 However, akin to other countries with restrained resources, there is a paucity of data on the population distribution of risk factors, and of intermediate and final end points for CVD.

Taking the projections of a greater increase of CVD mortality in LMIC into account, we designed the Healthy Life in Suriname (HELISUR) study, a population-based study on cardiovascular health among different ethnic groups in an urban setting. The study aims to generate population data on established risk factors for CVD, such as obesity, hyper- tension, diabetes and hyperlipidaemia. In addition, we will collect data on intermediate 30 Chapter 2

end points, including non-invasive haemodynamics, and the prevalence of clinical cardiovascular end organ damage.

Figure 2. Map of Suriname, South America. Suriname is located in the northeastern part of South America. The country is bordered by French Guiana, Guiana and Brazil, but culturally, economically, and demographically it is part of the Caribbean Community (CARICOM). Urban areas include the capital Paramaribo in northern Suriname and the surrounding suburbs.

Methods

Study design HELISUR is a cross-sectional multiethnic population-based study of adults aged be- tween 18–70 years and living in Suriname. The study is scheduled from November 2013 to December 2016. Data analysis and manuscript submission are foreseen from 2016.

Study objectives The primary objective is to assess risk factors for CVD and prevalent CVD across sex, age and ethnic groups in the general population of Suriname. Secondary objectives are HELISUR study protocol 31 assessment of ethnic disparities in health behaviour (e.g. physical activity, nutrition) and cardiovascular morbidity. Furthermore, we will address the question to what extent established and novel risk factors can explain current CVD. These novel risk factors may include haemodynamic factors, migration to urban areas or novel biochemical risk fac- tors. 2 Study area Suriname is estimated to have 534 500 citizens, with 75% of the population members younger than 45 years of age.4 The current mean life expectancy at birth is 71 years. The vast majority of the population (70%) lives in urban areas, mainly the coastal area where the capital, Paramaribo, is located.4 Table 1 shows an overview of World Develop- ment Indicators, indicating the categorisation of Suriname in LMIC.

Table 1. Overview of world development indicators in Suriname Parameter Estimate Income level Upper middle-income Population 534 500 Life expectancy at birth, years 71 GNI per capita, $ 8480 GDP, $ 4.7×1015 GDP growth, % 4.5 Inflation, % 5.0 Footnote Table 1. World development indicators estimated by the World Bank. Data on mean life expectancy at birth are from 2011; all other data are from 2012. GDP, gross domestic product; GNI, gross national income.

Study population Non-institutionalised participants living in urban areas, aged 18–70 years, will be randomly sampled through the cluster household sampling method, as described in ‘Sampling scheme’. Pregnant women can be included but will be invited for physical examination 3 months after delivery. Participants who are unable to give informed consent or are deceased before the questionnaire is filled out will be excluded.

Sampling scheme Using the methods described by Lwanga7, the minimum necessary representative sam- ple size was calculated to be 1092 participants, with a statistical precision of at least 0.01 to estimate the risk factor prevalence at a 95% confidence level. The sample size is weighted to account for the complex sampling design. Furthermore, we estimated a 40–50% refusal or dropout rate, based on a literature review of population studies. Therefore, we decided to include 1800 participants. 32 Chapter 2

Household sampling will be used, and households will be selected by randomisation of geographical areas, taking the residence of different ethnic groups into account. This will ensure that the different ethnic groups are adequately represented in the sample. Paramaribo is divided into 1200 enumeration areas out of which 18 will be randomly selected by the General Bureau of Statistics in Suriname. In order to include 1800 participants, 100 individuals will be needed per enumeration area. A group of trained interviewers will each cover one enumeration area, visiting every house and all household members until 100 persons are included, meeting the inclusion criteria. To reduce inclusion bias, if household members are not at home, interviewers will revisit until every household member is included or has stated a refusal to participate. Par- ticipants will be invited thereafter for physical examination at the Academic Hospital in Paramaribo. Techniques to minimise no-shows include telephone and text message prompting.

Research procedures

Questionnaire Participants will fill out a questionnaire at home with the support of a trained inter- viewer (Table 2). At the day of the physical examination, the questionnaire is updated to match the day of the physical examination and ensure validity of the answers.

Physical examination Participants will visit the hospital for a physical examination including anthropometrics and blood and urine tests (Table 3). Resting brachial systolic and diastolic blood pres- sure (SBP and DBP) will be measured twice in the sitting position, with the occluding cuff at the heart level (WatchBP Home; Microlife A.G., Widnau, Switzerland). The ankle- brachial index will be assessed in the supine position by WatchBP Office ABI (Microlife A.G., Widnau, Switzerland), twice on both sides. Finally, a 12-lead electrocardiogram will be recorded (ECG-1200 Biocare; Collateral Medical, Mumbai, India).

Safar8 proposed that although the cyclic blood pressure curve is usually described exclusively as SBP and DBP, both dominating the basis of cardiovascular hypertensive epidemiology, the entire blood pressure curve should be taken into consideration to adequately assess cardiovascular damage. Pulsatile arterial haemodynamics are able to predict CVD in essential hypertension, renal failure, diabetes mellitus and ageing. Therefore, we will assess pulsatile arterial haemodynamics non-invasively. These novel indices in cardiovascular population risk factor assessment include central systolic pres- sure, central pulse pressure, augmentation index and pulse wave velocity, as measured and calculated by the Arteriograph (TensioMed, Budapest, Hungary). In addition, we will HELISUR study protocol 33

Table 2. Health indicators in the HELISUR questionnaire Theme Explanatory factors General Demographic factors Sex, age, marital status, household composition, urban status, educational level, occupational status Ethnicity Self-reported ethnicity of participant and (grand) parents 2 Proximal risk factors

Health-related behaviour: smoking, alcohol intake, salt intake, dietary pattern, physical activity, weight perception; Healthcare use and related factors: therapy compliance, GP quality perception; Working conditions: physical activity at work Cardiovascular History of high blood pressure, diabetes, hypercholesterolaemia, fainting, or health venous thrombosis;

Symptoms or history of end organ damage of heart or vessels (angina pectoris, myocardial infarction, claudicatio intermittens with the Rose questionnaire; and neurological or transient neurological attacks); Family history of high blood pressure, diabetes, hypercholesterolaemia, cardiovascular disease or sudden death Footnote Table 2. Overview of health indicators to be explored in the HELISUR questionnaire. The questionnaire addresses demographic factors, as well as factors concerning cardiovascular health, including the presence of cardiovascular risk factors and cardiovascular disease. GP, general practitioner.

Table 3. Physical examination and laboratory analyses Study procedure Variables Physical Anthropometry: height, weight, waist-hip circumference, thigh, arm and calf examination circumference Heart function measured by Nexfin: non-invasive measurement of cardiac output, stroke volume and systemic vascular resistance

Arterial stiffness measured by an Arteriograph: non-invasive measurement of pulse wave velocity, augmentation index and central blood pressure Ankle-brachial index: peripheral vascular disease Sitting blood pressure Electrocardiogram: left ventricular hypertrophy Collection of fasting blood sample and overnight urine sample Laboratory Blood: haemoglobin, HbA1c, glucose, triglycerides, total cholesterol, HDL analyses cholesterol, LDL cholesterol, creatine kinase Urine: pH, glucose, microalbuminuria Footnote Table 3. Overview of the cardiometabolic risk assessment with physical and laboratory examination, including non-invasive cardiovascular haemodynamics. HbA1c, glycated haemoglobin; HDL, high-density lipoprotein; LDL, low-density lipoprotein. 34 Chapter 2

estimate the cardiac output, left ventricular contractility and systemic vascular resis- tance, by means of continuous non-invasive finger arterial pressure using the Nexfin HD Monitor (BMEYE, Amsterdam, The Netherlands). Participants will receive an annotated summary of the main results.

Handling and storage of data and documents All data will be handled confidentially and anonymously. A participant identification code list will be used to link the data to the participant. This code will not be based on the patients’ initials and/or date of birth. The key to the code will be saved separately from the database and will be kept by the primary investigator and the data manager. The handling of personal data will comply with the Personal Data Protection Act. With the participant’s permission, biological samples will be stored during 15 years, anno- tated with an identification code.

Statistical methods The primary outcome is descriptive: the prevalence of cardiovascular risk factors will be provided, stratified for ethnicity, sex, age, socioeconomic status and/or other classifica- tions of interest. Means (continuous variables) or percentages (categorical variables) of risk factors (including biochemical factors) and potential confounders will be reported, classified for ethnicity, sex, age and/or other classifications of interest. To study associa- tions of potential risk factors or determinants with (intermediate) disease outcomes, linear (for continuous outcomes) or logistic regression analyses (for dichotomous outcomes) will be performed.

Risk assessment The overall cardiovascular risk will be estimated by the Framingham Risk Score for CVD (10-year risk). Logarithms of age, serum lipids, systolic blood pressure, smoking and diabetes mellitus will be used.9 The Rose angina questionnaire will be used to diagnose angina pectoris and classify participants into definite angina, possible and non-exertional chest pain.10 The medical outcome survey Short Form-12 is validated as a quality of life assessment tool for the general population.11 Physical activity will be assessed by means of the International Physical Activity Questionnaire, which classifies participants into three levels of physical activity (low, moderate or vigorous).12

Quality control We will ensure that quality controls will be executed throughout the conduct of the study, with regard to sample selection, data collection, data processing and reporting. The well-defined sampling design and prior estimation of sample size reduces sampling errors. Furthermore, we tested and validated the questionnaire, and included multiple HELISUR study protocol 35 cross-checked questions on the same topic to validate the results. Additionally, health care workers who collect the data will be well trained according to standard operating procedures. On every day of the data collection, we monitor and ascertain the perfor- mance of our measurement devices and check the questionnaire of the participants for answer’s correctness and completeness. Our database allows range checks for accuracy and checks for completeness of the entered data. We used independent double data 2 entry followed by matching and checking for data entry errors. Data cleaning will be performed according to expert consensus. Finally, we will check internal and external consistency of the analysed data before writing reports.13

Ethical considerations This study will be conducted according to the principles of the Declaration of Helsinki (59th WMA General Assembly, Seoul, October 2008) and in accordance with the Medical Research Involving Human Subjects Act (WMO). Participants will be required to give informed consent for the study participation, which includes the permission to store a blood and urine sample for future research; to link registries containing data relating to the participant’s health (e.g. pharmacy, hospital, general practitioner); and finally to link with mortality registries where applicable.

Discussion

CVD is generally viewed as a disease of affluent countries. However, it has become the main cause of disease burden and premature death in low-income and middle-income countries, killing more people than malaria, AIDS and tuberculosis together.1 The rapid increase in urbanisation in these countries creates lifestyle changes, with greater con- sumption of fat, salt and sugar, and less physical exercise. In line with these changes, health projections suggest that CVD burden will increase even further in these regions (from 33% in 2002 to 45% in 2030), with potentially devastating effects on personal and family incomes, and on the national economy.2

With this in mind, we designed the HELISUR population study to gather data on the age distribution and prevalence of risk factors for cardiovascular disease, and of CVD morbidity and mortality. To the best of our knowledge, for the first time in this- set ting, we will include detailed assessments of non-invasively measured cardiovascular haemodynamic parameters such as aortic pulse wave velocity and central aortic blood pressure as intermediate outcomes for cardiovascular disease. These data are expected to be of great value to further probe the cardiovascular health status of the participants. 36 Chapter 2

Limitations pertain to the cross-sectional design of the study, permitting only descrip- tive outcomes. Furthermore, we need to be wary of selective survival bias, as any risk factor that results in death might be underrepresented. Importantly, the HELISUR study will yield extensive cardiovascular data, much needed by healthcare workers and policymakers to design measures to prevent the tsunami of CVD expected within the coming decades. HELISUR study protocol 37

References

1. World Health Organization. Global status report on noncommunicable disease 2014. Geneva 2014. ISBN 978 92 4 156485 4. 2. Mathers C, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Medicine. 2006; 3(11): 2011-2030. 3. Pan American Health Organization and the Caribbean Community. Report of the Caribbean Commission on Health and Development. Kingston, Jamaica: PAHO; 2006. 2 4. General Bureau of Statistics Suriname. Resultaten van de achtste (8e) volks- en woningtelling in Suriname (Volume I) - Demografische en sociale karakteristieken en migratie. Paramaribo 2012. 5. Punwasi W. Doodsoorzaken Suriname 2010-2011. Paramaribo: Ministry of Health; 2012. 6. The World Bank Group. Suriname. 2016. Available at: https://data.worldbank.org/country/ Suriname. 7. Lwanga S, Lemeshow S. Sample size determination in health studies: a practical manual. Geneva: WHO; 1991. 8. Safar M. Pulse pressure, arterial stiffness and wave reflections (augmentation index) as cardiovascular risk factors in hypertension. Therapeutic Advances in Cardiovascular Disease. 2008; 2(1): 13-24. 9. D’Agostino R, Vasan R, Pencina M, et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008; 117(6): 743-753. 10. Rose G. The diagnosis of ischaemic heart pain and intermittent claudication in field surveys. Bulletin WHO. 1962; 27: 654-658. 11. Ware J, Kosinski M, Keller S. A 12-item short-form health survey: construction of scales and preliminary tests of reliability and validity. Medical Care. 1996; 34: 220-233. 12. Craig C, Marshall A, Sjostrom M, et al. International physical activity questionnaire: 12-coun- try reliability and validity. Medicine in Science and Sports and Exercise. 2003; 35: 1381-1395. 13. Alkerwi A, Sauvageot N, Donneau A, et al. First nationwide survey on cardiovascular risk fac- tors in Grand-Duchy of Luxembourg (ORISCAV-LUX). BMC Public Health. 2010; 10: 468-469.

Chapter 3 Assessing the feasibility of the Healthy Life in Suriname study: using advanced hemodynamics to evaluate cardiovascular risk

Jet Q. Aartman, Frederieke S. Diemer, Fares A. Karamat, Evelien Bohte, Sergio M. Baldew, Ameerani V. Jarbandhan, Gert A. van Montfrans, Glenn P. Oehlers, Lizzy M. Brewster

Pan American Journal of Public Health 2017; 41: e46 40 Chapter 3

Abstract

Objectives: To determine the feasibility of assessing population cardiovascular risk with advanced hemodynamics in the Healthy Life in Suriname (HELISUR) study.

Methods: This was a preliminary study conducted in May-June 2012 using the Technical- Economic-Legal-Operational-Scheduling (TELOS) method to assess the feasibility of the HELISUR, a large-scale, cross-sectional population study of cardiovascular risk factors and disease in Suriname. Suriname, a middle-income country in South America with a population of mostly African and Asian ethnicity, has a high risk of cardiovascular dis- ease. A total of 135 volunteers 18–70 years of age participated. A health questionnaire was tested in a primary health care center, and non-invasive cardiovascular evaluations were performed in an academic health center. The cardiovascular evaluation included sitting, supine, and standing blood pressure, and intermediate endpoints, such as cardiac output, peripheral vascular resistance, pulse wave velocity, and augmentation index.

Results: The TELOS testing found that communicating by cellular phone was most effective for appointment adherence, and that completion of the questionnaire often required assis- tance from a trained interviewer; modifications to improve the clarity of the questions are recommended. Regarding the extended cardiovascular assessments of peripheral and central hemodynamics, the findings showed these to be technically and operation- ally feasible and well tolerated by participants, in terms of burden and duration.

Conclusions: Findings of this feasibility assessment indicate that large-scale, detailed evaluations of cardiovascular risk, including a questionnaire and advanced central and peripheral hemodynamics, are feasible in a high-risk population in a middle-income setting. Feasibility of the HELISUR study 41

Introduction

Non-communicable diseases are a priority in the United Nations’ development agenda.1 Steps have been outlined to reduce non-communicable diseases, particularly in low- and middle-income countries (LMICs) where higher rates of non-communicable diseases are expected.2,3,4 The Caribbean Community (CARICOM) comprises many LMICs, where the rate of cardiovascular mortality is high and preventive action is urgently needed.4 A member of CARICOM, Suriname is a middle-income country with a population of mostly Asian and African ethnicity.5 Despite the high cardiovascular mortality, there is a pau- city of data on the population distribution of cardiovascular risk factors and subclinical 3 target organ ​damage caused by cardiovascular disease.6,7

To assist with planning and implementing a future large-scale, cross-sectional popula- tion study of cardiovascular disease - the Healthy Life in Suriname (HELISUR) study7 - the present study assessed the feasibility of a health questionnaire and a physical examination with non-invasive cardiovascular measurements.8

Methods

The Technical–Economic–Legal–Operational–Scheduling (TELOS) method was used to assess the feasibility of two procedures planned for the HELISUR study.9 The authors de- termined whether or not the questionnaire had been correctly completed and whether the electronic devices functioned properly.10,11 Furthermore, the costs of the tests and consumables were examined to determine if they were as expected, and whether there would be legal issues. In addition, it was determined whether the standard operation protocol functioned properly, and if the timelines would need adjustment. Finally, the feasibility of the assessment of intermediate cardiovascular endpoints with extensive non-invasive cardiovascular analyses was estimated, including sitting, supine, and standing blood pressure; cardiac output; peripheral vascular resistance; pulse wave velocity; and augmentation index. The outcomes of the questionnaire and the physical examination procedures are also reported as a secondary outcome.

Inclusion The questionnaire was tested in a primary health care center in the village of Lelydorp, a small city of about 20 000 residents, east of the capital. Patients and their family mem- bers were asked to volunteer. The procedures regarding the physical examination were tested at the Academic Hospital (Paramaribo, Suriname). All participants were 18–70 42 Chapter 3

years of age. To be included, volunteers had to speak Dutch, the national language. Ethnicity was self-defined.

Health questionnaire The previously validated health questionnaire was based on studies of Surinamese immigrants in the Netherlands of Surinamese immigrants: the Study on Ethnicity and Health (SUNSET12) and HEalthy LIfe in an Urban Setting (HELIUS13). As part of a related HELIUS study in which Surinamese were a major part of the study population, the inter- viewer was trained to explore unanswered questions with the use of pre-set alternative phrases as much as possible.13 The questions considered general health, nutrition, physical activity, income and education, risk factors for cardiovascular disease, and the use of prescription drugs. The percentage of the participants able to adequately answer each question (with or without help of the interviewer) was determined. Furthermore, the willingness of volunteers to participate, the clarity of the questions, and the time needed to complete the questionnaire were evaluated.

Physical examination The physical examination was performed at the Department of Cardiology at the Aca- demic Hospital in Paramaribo. The room temperature was 24 °C. Physical examination included anthropometry and blood pressure measurements at rest in the sitting posi- tion, with an appropriately adjusted cuff at heart level. The mean of two consecutive blood pressure measurements was used. Blood pressure categories were defined ac- cording to the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC7) guideline.14 Hypertension was defined as sitting blood pressure ≥ 140 mm Hg systolic and/or ≥ 90 mm Hg diastolic, or the use of antihypertensive drugs. Prehypertension was defined as 120−139 mm Hg systolic and/or 80−89 mm Hg diastolic, without antihypertensive drugs. Normoten- sion was defined as < 120 mm Hg systolic and < 80 mm Hg diastolic blood pressure. Furthermore, the Nexfin HD monitor (BMEYE, Amsterdam, the Netherlands) was used to assess cardiac output and peripheral vascular resistance by a continuous finger arterial blood pressure measurement.15 First, a 10-minute Nexfin measurement was performed in the supine position, followed by a 5-minute measurement in the standing position. Between these two Nexfin measurements, the pulse wave velocity and augmentation index profiles in the supine position was estimated using the Arteriograph (TensioMed, Budapest, Hungary).16 The participants received an annotated summary of the results.

Feasibility was determined in terms of the participants’ willingness and ability to un- dergo the physical examination and to complete the assessments, and in terms of the functionality of the standard operating procedures and technical devices. In addition, Feasibility of the HELISUR study 43 time scheduling of appointments and the maximum number of participants per day were determined. Biochemical assessments of blood and urine were not included in the feasibility study.

Ethics Ethical clearance was granted by the Ethics Committee of the Ministry of Health of Suriname, as a preliminary study for HELISUR (No. VG 021-2012). All participants gave oral informed consent that was witnessed by two investigators. All data were handled confidentially and anonymously.7 3 Statistical analysis Because feasibility was the primary outcome of the study, a formal sample size calcula- tion was not performed. Descriptive statistics were computed for both the persons that participated in the health questionnaire, as well as for those who were physically as- sessed. Furthermore, the non-invasive cardiovascular outcome data for all participants was analyzed. The study planned a priori to report outcome data classified by ethnicity. Data were analysed with Microsoft Excel™ (Microsoft Corp., Redmond, Washington, United States), IBM SPSS Statistics software, version 20 (SPSS Inc., an IBM company, Chicago, Illinois, United States), and GraphPad Prism Software version 5 (GraphPad Software Inc., San Diego, California, United States).

Results

Health questionnaire

Feasibility The researchers approached 88 participants of whom 17 were not included (14 declined to participate and 3 did not speak Dutch). Three participants were unable to finish the questionnaire due to either its duration and their time constraints, or an unwillingness to answer questions about their current disease. The remaining 68 participants were able to answer 58% of the questions independently and 94% with an interviewer’s assistance. The main problems with answering the questions appeared to be related to small differences in preferred phrases and interpretation of the Dutch language spoken in the Netherlands versus that of Suriname.

Outcomes The characteristics of the subjects participating in the health questionnaire feasibility study are reported in Table 1-A. There were 68 participants (23 men; 45 women) with 44 Chapter 3

a mean age of 46.3 years (Standard error [SE] 1.4 years). The mean years of education ranged from 7.3 (SE 0.7 years) in South Asians to 9.9 (SE 1.9 years) in other ethnic groups. Regarding the cardiovascular risk factors, 16.2% of the while 14.7% met “the fit standard,” corresponding to intensive exercise 3 times per week for at least20 minutes. Of the 31 participants with hypertension (45.6%), 5 were of African ethnicity (16.1%), 17 of South Asian (54.8%), 8 of Javanese (25.8%), and 1 of other ethnicity (3.2%). The majority used antihypertensive drugs (77.4%). Participants of African eth- nicity reported the highest proportion of hypercholesterolemia (40.0%) and diabetes (30.0%). Regarding cardiovascular disease, a history of myocardial infarction and stroke was reported by 5.9% and 1.5% of participants, respectively.

Table 1-A. Self-reported clinical characteristics of subjects participating in the feasibility study of administering a cardiovascular health questionnaire

Total African South Asian Javanese Other n = 68 n = 10 n = 28 n = 23 n = 7 Mena 23 (33.8) 5 (50.0) 10 (35.7) 6 (26.1) 2 (28.6) Age, yearsb 46.3 (1.4) 52.7 (4.6) 45.8 (2.2) 43.8 (2.5) 47.6 (4.1) Education, yearsb 8.1 (0.4) 9.2 (0.8) 7.3 (0.7) 8.1 (0.7) 9.9 (1.9) Smokersa 11 (16.2) 1 (10.0) 4 (14.3) 3 (13.0) 3 (42.9) Physically activea, c 10 (14.7) 1 (10.0) 4 (14.3) 3 (13.0) 2 (28.6) Hypertensiona 31 (45.6) 5 (50.0) 17 (60.7) 8 (34.8) 1 (14.3) Treateda 24 (35.3) 3 (30.0) 13 (46.4) 7 (30.4) 1 (14.3) Hypercholesterolemiaa 15 (22.1) 4 (40.0) 8 (28.6) 3 (13.0) 0 (0.0) Diabetes mellitusa 15 (22.1) 3 (30.0) 8 (28.6) 4 (17.4) 0 (0.0) Myocardial infarctiona 4 (5.9) 1 (10.0) 3 (10.7) 0 (0.0) 0 (0.0) Strokea 1 (1.5) 0 (0.0) 1 (3.6) 0 (0.0) 0 (0.0) Footnote Table 1-A. a Values are n (%). b Values are mean (SE). c Being physically active was defined as meeting the criteria of intensive exercise 3 times per week for at least 20 minutes, using the Short Questionnaire to Assess Health Enhancing (SQUASH) physical activity.17

Physical examination

Feasibility The extended cardiovascular assessments of peripheral and central hemodynamics appeared to be technically and operationally feasible. All devices functioned properly. The annotated summary of the results that participants received was much appreciated and named as an important incentive worthy of participating. Concerning organizational and scheduling feasibility, the participants were most successfully reached by cellular telephone, rather than by regular mail, e-mail, or wired phone. Making appointments through cell phones on short notice, in combination with text message prompting, Feasibility of the HELISUR study 45 appeared to give the least no-shows. The optimal number of participants was 4–6 individuals per day.

Outcomes The characteristics of the participants who were physically assessed are depicted in Table 1-B. There were 67 subjects (29 men; 38 women) with a mean age of 43.6 (SE 1.7 years). Mean body mass index was 28.1 (SE 0.6 kg/m2); African 29.0 (SE 1.0 kg/m2), South Asian 27.7 (SE 1.2 kg/m2), and other ethnicity 26.8 (SE 1.2 kg/m2). Only 31.3% were normotensive. Hypertensive and prehypertensive blood pressure levels were found in 38.8% and 29.9% of participants, respectively. Compared to participants 3 of self-defined South Asian and other ethnicity, participants of African ethnicity had higher systolic blood pressure levels (137 mm Hg in Africans versus 135 and 131 mm Hg in South Asians and people of other ethnicity, respectively) and generally more hypertension (41% vs. 35 and 40%) and prehypertension (41% vs. 22 and 20%).

Table 1-B. Clinical characteristics of subjects participating in the feasibility study on assessing non-invasive cardiovascular hemodynamics

Total African South Asian Others n = 67 n = 29 n = 23 n = 15 Mena 29 (43.4) 12 (41.4) 8 (34.8) 9 (60.0) Age, yearsb 43.6 (1.5) 44.4 (2.5) 45.2 (2.3) 39.7 (3.1) Body mass index (kg/m2)b 28.1 (0.6) 29.0 (1.0) 27.7 (1.2) 26.8 (1.2) Systolic BP, mmHgb 133 (2.0) 137 (2.7) 128 (3.0) 131 (5.3) Diastolic BP, mmHgb 82 (1.3) 83 (1.8) 79 (2.0) 85 (3.5) Heart rate, beats per minuteb 67 (1.3) 69 (2.1) 64 (1.7) 68 (2.7) Hypertensiona, c 26 (38.8) 12 (41.4) 8 (34.8) 6 (40) Prehypertensiona, d 20 (29.9) 12 (41.4) 5 (21.7) 3 (20) Footnote Table 1-B. a Values are n (%).b Values are mean (SE). c Hypertension was defined as blood pressure (BP) ≥ 140 mmHg systolic and/or ≥ 90 mmHg diastolic, or the use of antihyper- tensive drugs.14 d Prehypertension was defined as 120–139 mmHg systolic and/or 80–89 mmHg diastolic, without the use of antihypertensive drugs.14

Finally, non-invasive hemodynamic measurements were taken. Across the different age categories, higher pulse wave velocity and augmentation index profiles were found in South Asians compared to participants of African ethnicity (Figure 1-A and 1-B). Only in the oldest age category, participants of African ethnicity had slightly higher mean pulse wave velocities than those of South Asian ethnicity (11.3 vs. 11.2 m/s). Moreover, par- ticipants of Asian ethnicity had a lower mean cardiac output (6.1 [SE 0.3] in Asians vs. 6.7 [SE 0.3] L/min in Africans) and a higher systemic vascular resistance (1302.3 [SE 78.6] vs. 1198.9 [SE 85.3] dyn·s/cm5) compared to participants of African ethnicity (Table 2). 46 Chapter 3

10 9 8

(m/s) 7 a 6 22 - 36 years 5 velocity 37 - 52 years 4 53 - 70 years wave 3

Pulse 2 1 23 23 19 9 8 10 7 10 6 7 5 3 0 Total African South Asian Others

20

10 6 3 0

(%) 23 23 19 9 8 7 10 7 5 a -10 10

Index -20 22 - 36 years

-30 37 - 52 years 53 - 70 years -40

-50 Augmentation -60

-70 Total Africans South Asians Other

Figure 1. (A) Mean pulse wave velocity assessed by the Arteriograph device per age category in the total group and in different ethnic groups; (B) Mean augmentation index assessed by the Arterio- graph device per age category in the total group and in different ethnic groups. Arterial stiffness is considered increased when pulse wave velocity is higher than 10 m/s.18 A lower AIx indicates healthier arteries.

Upon standing, the increase in peripheral vascular resistance was higher in participants of African ethnicity, while the participants of South Asian ethnicity displayed a greater increase in cardiac parameters, in particular heart rate and left ventricular contractility. Feasibility of the HELISUR study 47 c c d d c % 9.9 25.0 13.7 11.6 48.9 -29.9 -22.4 (95% CI) (11 to 25.1) (3.7 to 30.3) (-2.3 to -0.7) (-0.6 to 15.7) (-37.7 to -19.9) (-49.5 to 194.1) (293.9 to 879.6) a b 8 Δ 18 17 -1.5 72.3 -28.9 20 586.7 8/12 African 41.7 ( 3.2) a 3 1785.6 (113.7) 77 (2.3) 90 (2.1) 5.2 (0.3) 141 (4.4) Standing 68.1 (3.0) 796.4 (52.9) a (85.3) p < 0.10 for indifference fractional increase or decrease in 1198.9 Supine 69 (2.6) 72 (1.7) 6.7 (0.3) 124 (3.8) 97.0 (4.0) 724.1 (46.4) 724.1 % -1.6 23.4 22.1 20.7 35.8 22.0 -17.8 One-sided d p < 0.05 for indifference fractional increase or decrease in cardiovascular (95% CI) (-0.9 to 0.4) (9.4 to 21.0) (9.9 to 22.2) (11.2 to 39.1) (2.5 to 571.0) (-23.5 to -8.6) One-sided c (136.0 to 350.8) a b Δ 16 16 25 -0.1 -16.0 243.4 286.8 18 7/11 44.1 (2.8) South Asian a 1589.1 (133.6) 81 (3.3) 89 (2.2) 6.0 (0.4) 146 (4.8) Standing 74.3 (2.6) 924.1 (44.4) Δ = standing – supine. b a (78.6) 1302.3 Supine 65 (1.8) 73 (2.0) 6.1 (0.3) 121 (3.4) 90.3 (4.0) 680.7 (44.0) ) Values are mean (SE). 5 a ootnote Table ootnote 2. Table Heart rate (beats/minute) Heart rate Diastolic BP (mm Hg) cardiovascular hemodynamic parameters between subjects of South Asian vs. African ethnicity. South Asian vs. African subjects of between hemodynamic parameters cardiovascular hemodynamic parameters between subjects of South Asian vs. African ethnicity. N Men/women Age in years Age Systolic BP (mm Hg) Stroke volume (mL) volume Stroke output (L/min) Cardiac contractility ventricular Left (dP/dt) Systematic vascular (dyn·s/cm resistance F Supine and standing hemodynamic parameters in subjects of South Asian and African ethnicity South Asian and African in subjects of 2. Supine and standing hemodynamic parameters Table Position 48 Chapter 3

Discussion

This study shows that it is feasible to study cardiovascular health in a middle-income​ country, with the use of an extensive health questionnaire and a physical examina- tion that includes assessment of intermediate cardiovascular endpoints through non-invasive peripheral and central hemodynamics. Moreover, our data indicate that the help of a trained interviewer and rephrasing of various questions was essential to successfully complete the questionnaire. It appeared that the physical examination also needed several logistic adjustments, such as making cell phones the preferred method of contacting participants.

Other findings that will be implemented are the time scheduling of appointments tak- ing transportation characteristics into account, a maximum of six participants per day, and adjustment to the order of the measurements. The preferred method of contacting subjects through mobile phone aligned with the findings of Hartzler and colleagues19, which showed that cell phones used in LMICs improved appointment adherence.

The results of the questionnaire showed a high burden of cardiovascular risk factors, including hypertension, diabetes, and lack of physical activity among participants. However, these results should be interpreted with some caution, as the sampling was non-random and the outcomes were based on self-reported data.

In the physical examination, we found evidence that participants of South Asian ethnic- ity showed higher age-adjusted pulse wave velocity and augmentation index values compared to those of African ethnicity; this potentially implicates ethnic differences in non-invasive hemodynamics. Furthermore, our data suggested a differential adaptation pattern to an orthostatic challenge between South Asian and African participants with, respectively, a predominantly cardiac vs. peripheral vascular response. A recent paper reported that individuals of African ethnicity had a greater response in total periph- eral resistance for a given change in muscle sympathetic nerve activity during tilting.20 These data should be considered hypothesis-generating and will be further explored in the final study.

The main strength of this study was the use of the TELOS method for assessing feasibility. It enabled us to collect more detailed data in a structured setting. These data will help us better design the final study and optimize the use of scarce resources in a middle- income​ setting as described above. Furthermore, Safar and colleagues21 have proposed that the entire blood pressure curve should be taken into consideration to evaluate cardiovascular risk, and that pulse and aortic pulse wave velocity are useful pulsatile Feasibility of the HELISUR study 49 hemodynamics to predict cardiovascular risk in essential hypertension, renal failure, diabetes mellitus, and aging. This is of particular importance in a high-risk population where more timely, preventive measures are needed. However, to our knowledge, the blood pressure curve has not been previously studied in a LMIC population setting. Our feasibility data indicate that it should be possible to assess non-invasive hemodynam- ics in this setting.

Limitations There were study limitations worth noting. A formal cost-analysis was not performed; it would have provided more complete information for the HELISUR study. Neverthe- 3 less, we did evaluate whether costs were as expected. Another limitation might be the interviewer’s help required by some participants - it might have influenced the results. However, we tried to avoid this as much as possible with the use of pre-set alternative phrases when questions remained unanswered. Finally, outcome data for the secondary objective should be considered hypothesis-generating only because volunteers were self-selected. In addition, the small sample size and the single measurement cycle of the (hemodynamic) variables in time preclude conclusions regarding differences between ethnic groups. However, our local data are similar to international trends in ethnic differences in cardiovascular risk.11,22

Conclusions The Pan American Health Organization recommends that more research on non-commu- nicable diseases and risk factors be conducted in CARICOM.4 Assessing the feasibility of cardiovascular population studies in the Caribbean, using questionnaires and costly devices, is particularly pertinent given the limited funding for research in LMICs. With a feasibility study, there is a greater promise of success in the final study23, and less chance of scarce funding being wasted by a failed one.8 In conclusion, cardiovascular mortality is the number one cause of death in LMICs and urgent preventive measures are needed. In order to provide data for prevention and intervention strategies, we as- sessed the feasibility of a health questionnaire and a physical examination including advanced central and peripheral hemodynamics in volunteers from a middle-income country. Although adaptations were necessary to optimize the data quality and quantity of the questionnaire and the physical examination, this feasibility study indicated that large-scale, detailed evaluations of cardiovascular risk are feasible in a middle-income setting, and that high-quality data can be collected to better prevent, detect, and treat cardiovascular disease in Suriname. 50 Chapter 3

References

1. United Nations General Assembly. Political declaration of the high-level meeting of the General Assembly on the prevention and control of non-communicable diseases. 2011. Available at: http://www.who.int/nmh/events/un_ncd_summit2011/political_declara- tion_en.pdf. 2. World Health Organization. Global status report on noncommunicable disease 2014. Geneva 2014. ISBN 978 92 4 156485 4. 3. Mathers C, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Medicine. 2006; 3(11): 2011-2030. 4. Pan American Health Organization and the Caribbean Community. Report of the Caribbean Commission on Health and Development. Kingston, Jamaica: PAHO; 2006. 5. The World Bank Group. Suriname. 2016. Available at: https://data.worldbank.org/country/ Suriname. 6. Punwasi W. Doodsoorzaken Suriname 2010-2011. Paramaribo: Ministry of Health; 2012. 7. Diemer F, Aartman J, Karamat F, et al. Exploring cardiovascular health: the Healthy Life in Su- riname (HELISUR) study. A protocol of a cross-sectional study. BMJ Open. 2014; 4: e006380. 8. Thabane L, Ma J, Chu R, et al. A tutorial on pilot studies: the what, why and how. BMC Medical Research Methodology. 2010; 10: 1. 9. Hall J. Accounting information systems. Mason, Ohio, United States: South Western Publish- ing Company; 2007. 10. Kewalbansing P, Ishwardat A, Brewster L, Oehlers G, van Montfrans G. Field testing in Suriname of two blood pressure-measuring devices for low-resource and middle-resource countries, according to a WHO protocol. Blood Pressure Monitoring. 2013; 18(2): 78-84. 11. Brewster L, Mairuhu G, Bindraban N, Koopmans R, Clark J, van Montfrans G. Creatine kinase activity is associated with blood pressure. Circulation. 2006; 114(19): 2034-2039. 12. Dekker L, Nicolaou M, van der A D, et al. Sex differences in the association between serum ferritin and fasting glucose in type 2 diabetes among South Asian Surinamese, African Surinamese, and ethnic Dutch: the population-based SUNSET study. Diabetes Care. 2013; 36(4): 965-971. 13. Stronks K, Snijder M, Peters R, Prins M, Schene A, Zwinderman A. Unravelling the impact of ethnicity on health in Europe: the HELIUS study. BMC Public Health. 2013; 13: 402. 14. National Heart, Lung, and Blood Institute. The seventh report of the Joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure (JNC 7). Bethesda: U.S. Department of Health and Human Services; 2003. 15. Eeftinck Schattenkerk D, van Lieshout J, van den Meiracker A, et al. Nexfin noninvasive continuous blood pressure validated against Riva-Rocci/Korotkoff. American Journal of Hypertension. 2009; 22(4): 378-383. 16. Baulmann J, Schillings U, Rickert S, et al. A new oscillometric method for assessment of arterial stiffness: comparison with tonometric and piezo-electronic methods. Journal of Hypertension. 2008; 26(3): 523-528. 17. Nationaal Kompas Volksgezondheid. 2014. Available at: https://www.volksgezond- heidenzorg.info/onderwerp/sport-en-bewegen/cijfers-context/huidige-situatie#!node- beweeggedrag-0. 18. Mancia G, Faqard R, Narkiewicz K, et al. 2013 ESH/ESC Guidelines for the management of arterial hypertension: the Task Force for the management of arterial hypertension of the Feasibility of the HELISUR study 51

European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens. 2013; 31(7): 1281-1357. 19. Hartzler A, Wetter T. Engaging patients through mobile phones: demonstrator services, success factors, and future opportunities in low and middle-income countries. Yearbook of Medical Informatics. 2014; 9: 182-194. 20. Okada Y, Galbreath M, Jarvis S, et al. Elderly blacks have a blunted sympathetic neural re- sponsiveness but greater pressor response to orthostasis than elderly whites. Hypertension. 2012; 60(3): 842-848. 21. Safar M. Pulse pressure, arterial stiffness and wave reflections (augmentation index) as cardiovascular risk factors in hypertension. Therapeutic Advances in Cardiovascular Disease. 2008; 2(1): 13-24. 22. Gunarathne A, Patel J, Potluri R, Gill P, Hughes E, Lip G. Secular trends in the cardiovascular 3 risk profile and mortality of stroke admissions in an inner city, multiethnic population in the United Kingdom (1997-2005). Journal of Human Hypertension. 2008; 22(1): 18-23. 23. Pena M, Bloomfield G. Cardiovascular disease research and the development agenda in low- and middle-income countries. Global Heart. 2015; 10(1): 71-73.

Part II Ethnic differences in cardiovascular risk

Chapter 4 Hypertension and cardiovascular risk profile in urban Suriname: the HELISUR study

Frederieke S. Diemer, Se-Sergio M. Baldew, Yentl C. Haan, Jet Q. Aartman, Fares A. Karamat, Lenny M. W. Nahar-van Venrooij, Gert A. van Montfrans, Glenn P. Oehlers, Lizzy M. Brewster

American Journal of Hypertension (2017); 30(11):1133-1140 56 Chapter 4

Abstract

Introduction: Hypertension is the leading risk factor responsible for premature death worldwide, but its burden has shifted to low- and middle-income countries. Therefore, we studied hypertension and cardiovascular risk in the population of Suriname, a middle-income country with a predominantly urban population of African and Asian ancestry.

Methods: A random sample of 1,800 non-institutionalized men and women aged 18–70 years was selected to be interviewed at home and examined at the local hospital for cardiovascu- lar risk factors, asymptomatic organ damage, and cardiovascular disease.

Results: The 1,157 participants examined (37% men) were mainly of self-defined Asian (43%) or African (39%) ancestry, mean age 43 years (SD 14). The majority of the population (71%) had hypertension or prehypertension, respectively, 40% and 31%. Furthermore, 72% was obese or overweight, while 63% had diabetes or prediabetes. Only 1% of the adult population had an optimal cardiovascular risk profile. Hypertension awareness, treatment, and control were respectively 68%, 56%, and 20%. In line with this, 22% of the adult population had asymptomatic organ damage, including increased arterial stiffness, left ventricular hypertrophy, microalbuminuria, or asymptomatic chronic kid- ney disease.

Conclusions: In this first extensive cardiovascular assessment in the general population of this middle-income Caribbean country, high prevalence of hypertension with inadequate levels of treatment and control was predominant. The findings emphasize the need for collaborative effort from national and international bodies to prioritize the implemen- tation of affordable and sustainable public health programs that combat the escalating hypertension and cardiovascular risk factor burden. Cardiovascular risk profile of the HELISUR study participants 57

Introduction

Hypertension is the main preventable cause of disability and premature death world- wide, but the prevalence is increasing in low- and middle-income countries (LMIC), in contrast to a slight decrease in high-income regions.1 At present, more than 80% of the global burden of hypertension occurs in LMIC, with half of the mortality occurring at a relatively young age between 30 and 65 years.2,3 The economic burden of hypertension and cardiovascular disease (CVD) in terms of mortality- and disability-adjusted life years in these resource-limited settings is high.1 Although there are ample data from migrants of African, Asian, and Caribbean descent in North America and Europe4,5,6, these data may not be applicable to populations in LMIC, where socioeconomic and environmental conditions as well as health systems and policies are different. Therefore, the United Na- tions emphasized the need for sufficient data on cardiovascular risk factors from these regions to combat the increasing cardiovascular mortality.7 Suriname is a multiethnic 4 middle-income country in South America, where CVD is the main cause of death among its population of predominantly African and Asian ancestry.8 However, there are scant data on the prevalence of hypertension and other cardiovascular risk factors. Therefore, we assessed hypertension and cardiovascular health status in a random sample of the general population.

Methods

Ethical clearance The study was conducted according to the principles of the Declaration of Helsinki (59th WMA General Assembly, Seoul, October 2008) and in accordance with the Medical Research Involving Human Subjects Act. Ethical clearance was obtained from the Ethics Committee of the Ministry of Health in Suriname in 2012 (Approval nr. VG021-2012). Written informed consent was obtained from all participants.

The HELISUR study The Healthy Life in Suriname (HELISUR) study is a cross-sectional population-based study conducted in the capital Paramaribo, the only main city in the country. The major- ity of the Surinamese population (73%) lives in an urban setting.9 Study procedures were published previously.10 A flowchart of the sampling procedure is depicted in Fig- ure 1. In brief, a random sample of 1,800 non-institutionalized urban men and women aged 18–70 years was selected to be interviewed at home (response rate: 72%). The interview included demographic, socioeconomic, dietary, and health-related questions. Subsequently, participants were invited to be examined in our research facility at the 58 Chapter 4

local Academic Hospital (response rate: 78%). The physical examination included anthropometric, medical, and physiological measurements including noninvasive he- modynamics, as well as laboratory tests administered by trained medical personnel.10 Main outcomes were the prevalence of hypertension, other cardiovascular risk factors, asymptomatic organ damage and CVD by sex, age, and ancestry.10

1200 enumeration areas

Random selection of 18 enumeration areas

Preliminary assessment n=1,800 persons Did not meet the age inclusion criteria (n=5) Died (n=9) Not retrieved (n=303) Eligible n=1,483 persons

Declined to participate (n=326)

Physical examination n=1,157 participants

Figure 1. Flowchart of the sampling procedure. The capital of Suriname is divided into 1,200 Cen- sus enumeration areas, consisting of 100–150 households. General Bureau of Statistics randomly selected 18 enumeration areas stratified by ancestry. Households were approached consecutively until 100 subjects per enumeration area were included (response rate: 72%). Preliminary assess- ment included the interviews administered at the subjects’ home. Eligible persons were subse- quently invited to visit the hospital for the examination (response rate: 78%).

Study procedures Participants were seen by a trained investigator in the morning in the fasting state. Questionnaires filled out at home were checked and updated. Blood pressure was mea- sured twice in the sitting position with an automated oscillometric device (WatchBP Office; Microlife AG, Widnau, Switzerland) and an appropriately adjusted cuff size on the left upper arm supported at heart level. Height and weight were measured twice to the nearest 0.1 cm and 0.1 kg, respectively, with participants wearing underwear. Body mass index was computed as mean weight (in kilograms) divided by the square of the mean height (in meters). We recorded 12-leads electrocardiograms, and estimated the aortic pulse wave velocity twice noninvasively in the supine position by analysis of the Cardiovascular risk profile of the HELISUR study participants 59 oscillometric pressure curves registered on the right upper arm, using the Arteriograph (TensioMed, Budapest, Hungary). Fasting venous blood samples and morning spot urine were collected for laboratory measurements.

Definitions and outcome parameters The population of the capital Paramaribo consists of persons of African (43%), Asian (33%), and other ancestry.9 Persons of African ancestry are locally considered to be Creole or Maroon. Maroons recently migrated to the capital and therefore tend to have less admixture with other populations than Creoles.11 Persons of Asian ancestry migrated from India or the island of Java to Suriname in the late 19th century and early 20th century.11,12 For the current analysis, we used self-defined ancestry.

We used the following definitions: hypertension, systolic blood pressure ≥140 mm Hg, or diastolic blood pressure ≥90 mm Hg, or receiving antihypertensive drug therapy; 4 prehypertension, systolic blood pressure 120–139 mm Hg or diastolic blood pressure 80–89 mm Hg, without antihypertensive drug therapy.13 Controlled hypertension in- cluded subjects with hypertension using antihypertensive medication with a systolic blood pressure <140 mm Hg and diastolic blood pressure <90 mm Hg.13 Furthermore, we used ethnic-specific cutoff values for body mass index: obesity ≥30 kg/m2 in African-Surinamese participants (i.e., Creole and Maroons) and ≥27.5 kg/m2 in Asian- Surinamese participants (i.e., Surinamese of South Asian and Indonesian ancestry).14,15 Overweight was defined as body mass index 25.0–29.9 kg/ 2m and 23.0–27.4 kg/m2, respectively, in African and Asian ancestry participants.14,15 We considered participants to be diabetic if HbA1c ≥6.5% or when glucose-lowering medication was used, while prediabetes was defined as HbA1c 5.7–6.4% without the use of antihyperglycemic drugs.16 Diabetes control was defined as HbA1c <7.0%.17 Finally, dyslipidemia was de- fined as having at least one of the following: total cholesterol ≥6.2 mmol/l, low-density lipoprotein cholesterol ≥4.1 mmol/l, high-density lipoprotein cholesterol <1.0 mmol/l, triglycerides ≥2.3 mmol/l, or the use of lipid-lowering drugs.18 Borderline dyslipidemia was defined as total cholesterol 5.2–6.2 mmol/l, low-density lipoprotein cholesterol 3.4–4.1 mmol/l, high-density lipoprotein cholesterol 1.0–1.6 mmol/l, or triglycerides 1.7–2.3 mmol/l without the use of lipid-lowering drugs.18 Smoking was defined as using tobacco on daily basis.

An adverse cardiovascular risk profile was defined as the presence of any of the risk factors hypertension, obesity, diabetes, dyslipidemia, or smoking. If prehypertension, overweight, prediabetes, or borderline dyslipidemia was present, this was considered a suboptimal cardiovascular risk profile. An optimal cardiovascular risk profile was the absence of any of these risk factors. 60 Chapter 4

Asymptomatic organ damage was defined as the presence of either increased arterial stiffness (pulse wave velocity > 10 m/s19), left ventricular hypertrophy (LVH, Sokolow–

Lyon criteria: S in V1 or V 2 + R in V5 or V6 [whichever is larger] ≥35 mm, or R in aVL ≥ 11 mm20), microalbuminuria (30−300 mg/dl in spot urine21), or asymptomatic chronic kidney disease (CKD), according to the European Society of Hypertension/European Society of Cardiology guidelines.13 Estimated glomerular filtration rate was calculated using the CKD-EPI equation, and asymptomatic CKD was defined as estimated glomeru- lar filtration rate 30−59 ml/min/1.73 2m .21,22

We defined CVD as history of coronary heart disease (history of myocardial infarction, or coronary bypass surgery), history of stroke, or the presence of severe CKD. Severe CKD was defined as estimated glomerular filtration rate <30 ml/ min/1.732 m .22

Finally, educational level was assessed through the highest completed level of educa- tion, and categorized into 2 groups according to the International Standard Classifica- tion of Education (ISCED): low (≤12 years of education, ISCED 1−2) or high (>12 years of education, ISCED ≥3).23

Statistical analysis Descriptive characteristics were computed for the total group (primary) and for African (with subgroups of Creole and Maroon) vs. Asian (with subgroups of South Asian and Indonesian) ancestry participants. We calculated the prevalence and the odds ratios of hypertension and other cardiometabolic risk factors, asymptomatic organ damage, and CVD for Asian vs. African ancestry participants, using the unadjusted and adjusted model for sex, age, body mass index, and educational level. Model fit was assessed with the appropriate goodness-of-fit test for logistic regression models. Statistical analyses were performed with the SPSS statistical software package for Windows, version 20.0 (SPSS, Chicago, IL).

Results

We analyzed data of 1,157 participants, 429 men and 728 women, with a mean age of 43 ± 14 years, as summarized in Table 1. Blood pressure was not optimal in 71% of the participants: 40% had hypertension and a further 31% was prehypertensive (Table 2, Figure 2). Hypertension awareness, treatment, and control were respectively 68, 56, and 20%. Hypertension prevalence did not differ by sex (P = 0.33) but men were more often prehypertensive than women (37 vs. 28%; P < 0.01). Figure 3 depicts hyperten- sion prevalence by age. In logistic regression analysis, Asian ancestry participants were Cardiovascular risk profile of the HELISUR study participants 61 being c ; 0.08 0.06 0.21 0.01 0.03 c <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 p -value and a * 14.4 62.8 5.6 ± 2.2 4.7 ± 1.1 3.3 ± 0.9 1.2 ± 0.3 1.3 ± 1.0 8.3 ± 2.6 27.8 ± 6.2 43.3 ± 13.8 81.3 ± 11.6 107.9 ± 22.2 129.8 ± 20.2 Total (n=1157) Total b b b b b b ac ab a d 14.3 66.7 being from significantly different 8.4 ± 2.6 b 5.1 ± 1.0 3.6 ± 0.9 1.7 ± 2.1 5.9 ± 2.2 1.2 ± 0.3 26.7 ± 4.7 80.2 ± 10.7 ; 47.6 ± 13.0 129.5 ± 19.2 c 100.5 ± 18.4 Values are means with SDs. are Values ‡ Indonesian (n=105)

and 4 b Asian b a b c b b b b c a 60.7 10.2 8.6 ± 2.5 6.2 ± 2.8 5.0 ± 1.0 3.6 ± 0.9 1.1 ± 0.2 1.5 ± 0.9 27.2 ± 5.6 81.3 ± 11.0 43.5 ± 13.2 Values are percentages. percentages. are Values 128.5 ± 19.1 104.0 ± 20.2 † South Asian (n=392) a c being from significantly different a a a a a a b b b 8.8 71.3 8.1 ± 2.4 4.4 ± 1.0 2.9 ± 0.9 1.2 ± 0.3 0.9 ± 0.5 4.8 ± 0.6 28.7 ± 6.8 80.6 ± 13.5 39.5 ± 13.8 129.3 ± 22.6 117.8 ± 24.0 aroons (n=216) M aroons African a ab ab a a a a a ab acd 19.2 59.2 Creole ; all at a level of P < 0.05, depicted in bold. of ; all at a level (n=245) b 8.4 ± 2.3 5.4 ± 1.8 4.5 ± 1.1 3.1 ± 1.0 1.2 ± 0.4 1.1 ± 0.7 83.1 ± 11.6 28.1 ± 6.6 132.7 ± 20.3 44.5 ± 13.5 112.7 ± 23.5 and a ‡ ‡ ‡ ‡ 2‡ ‡ Includes participants of other/mixed ancestry ( n = 199). ancestry other/mixed Includes participants of * ‡ ‡ † ‡ 2‡ ‡ † ootnote Table 1. ootnote Table Fasting glucose, mmol/L glucose, Fasting High education mmol/L cholesterol, Total mmol/L LDL-cholesterol, mmol/L HDL-cholesterol, mmol/L Triglycerides, m/s PWV, Diastolic BP, mmHg Diastolic BP, BMI, kg/m m eGFR, mL/min/1.73 significantly different from significantly different Women BMI, body mass index; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; LDL, low-density lipoprotein; PWV, pulse wave velocity. pulse wave PWV, lipoprotein; LDL, low-density rate; HDL, high-density lipoprotein; eGFR, estimated glomerular filtration index; BMI, body mass F Characteristics of the participants by ancestry the participants by of 1. Characteristics Table Systolic BP, mmHg Systolic BP, Age, years Age, Differences werefour tested the ancestry across groups with Differences 62 Chapter 4 0.29 0.12 0.48 0.41 0.55 0.80 0.28 0.95 0.28 0.40 0.05 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 p -value ; all at a level of of ; all at a level b * and a 0.4 2.5 3.4 2.8 1.4 4.7 2.0 39.5 21.6 58.8 54.4 40.9 22.3 16.6 18.0 36.8 68.0 55.9 20.6 (35.6) 35.7 (20.0) Total (n=1157) Total ac c c ac ac ab ab a a ab 0.0 1.0 2.9 1.9 1.0 1.9 19.0 22.9 38.1 67.4 60.5 1.0 24.8 61.5 61.5 46.7 41.0 14.3 38.5 (23.3) 31.3 (34.6) Percentage of subjects Percentage with hypertension/diabetes b Indonesian (n=105) being significantly different from being significantly different c ; c Asian a and a c c c c a b a b b 1.1 2.8 1.8 1.0 1.3 7.7 4.6 17.9 20.4 39.0 71.0 60.7 30.1 68.6 64.6 52.4 20.9 37.0 29.5 (18.1) 15.8 (26.3) South Asian (n=392) b b b b b b a a b b 0.5 0.5 3.7 0.9 1.4 0.5 7.9 9.3 15.7 25.9 38.9 72.8 51.9 37.7 11.1 33.3 16.7 27.3 33.3 (17.3) 50.0 (66.7) being significantly different from being significantly different aroons (n=216) M aroons b ; c African and b a a a a a a a Includes participants of ancestry other/mixed ( n = 199). a a ab a 3.3 0.0 2.9 1.6 1.2 3.3 15.9 22.9 36.7 63.8 53.4 4.5 37.0 22.9 22.0 47.7 40.7 40.7 Creole (n=245) 45.2 (24.1) 22.7 (46.3) † † ‡ † † ‡ LVH CHD Stroke Aware Aware Treated Treated Controlled Controlled Severe CKD Severe Microalbuminuria Asymptomatic CKD being significantly different from being significantly different a Percentage of treated subjects (percentage of all subjects with hypertension/diabetes mellitus). Differences were tested across the four ancestry ancestry four the across tested were Differences mellitus). hypertension/diabetes with subjects all of (percentage subjects treated of Percentage c Increased arterial stiffness Increased ootnote Table 2. are ootnote percentages. Values Table groups with groups Cardiovascular risk factors, asymptomatic organ damage, and CVD damage, asymptomatic organ risk factors, 2. Cardiovascular Table F mellitus. Dyslipidemia Smoking Early organ damage: Early organ CVD: Obesity Diabetes mellitus Hypertension P < 0.05, depicted in bold. hypertrophy. ventricular left disease; LVH, cardiovascular disease; CVD, kidney chronic heart disease; CKD, coronary CHD, Cardiovascular risk profile of the HELISUR study participants 63

Optimal Suboptimal Risk factor

100%

41%

77% 37% 50% 40% 22% Prevalence (%) Prevalence 54% 41% 35% 35% 29% 31% 28% 22% 1% 5% 0% No PreHT No Overweight No PreDM No Borderline Optimal Suboptimal and HT and obesity and DM and dyslipidemia and adverse

Hypertension Obesity Diabetes Dyslipidemia Cardiovascular risk profile 4 Figure 2. Cardiovascular risk profile of the general population. Bars are % of the total population. The majority of the population had an adverse cardiovascular risk profile, mainly due to a high hypertension and dyslipidemia prevalence. DM, diabetes mellitus; HT, hypertension; PreDM, prediabetes; PreHT, prehypertension.

100% ≤ 40 years > 40 years *

*

50% * * 87% * * Prevalence (%) Prevalence 57% 60% * 47% 42% 34% 29% 32% 32% 16% 17% 19% * 7% 9% 1% 8% 0% Hypertension Obesity Diabetes Dyslipidemia Smoking ≥1 risk factor Organ CVD damage Figure 3. Cardiovascular risk factors, asymptomatic organ damage, and CVD by age. Bars are % of the total population. Cardiovascular risk factors, asymptomatic organ damage, and CVD were more prevalent in participants above 40 years than in participants of 40 years and be- low. *Significantly higher in participants above 40 years (P < 0.01). CVD, cardiovascular disease.

less likely to have high blood pressure than participants of African ancestry (Figure 4). In participants of 40 years and younger, hypertension was more often seen in African ancestry than in Asian ancestry participants (21% vs. 14%; P = 0.06). 64 Chapter 4

OR (95% CI) Hypertension African (reference) Asian: unadjusted 0.81 (0.62–1.04) Asian: model I 0.72 (0.53–0.99) Obesity African (reference) Asian: unadjusted 1.05 (0.81–1.36) Asian: model I* 1.03 (0.78–1.35) Diabetes African (reference) Asian: unadjusted 1.92 (1.41–2.63) Asian: model I 2.09 (1.48–2.96) Dyslipidemia African (reference) Asian: unadjusted 2.19 (1.69–2.85) Asian: model I 2.30 (1.74–3.03) ≥1 risk factor African (reference) Asian: unadjusted 1.59 (1.18–2.16) Asian: model I* 1.40 (1.00–1.95) Organ damage African (reference) Asian: unadjusted 0.83 (0.61–1.12) Asian: model I 0.71 (0.51–0.99) CVD African (reference) Asian: unadjusted 2.20 (1.16–4.17) Asian: model I 2.03 (1.04–3.93)

0 2 4 6 Figure 4. Cardiovascular risk factors, asymptomatic organ damage, and CVD by ancestry. Values are odds ratios (OR) with 95% confidence intervals (CI). Reference group are African ances- try participants. African (n = 461), defined as Creole or Maroon; Asian (n = 497), defined as South Asian or Indonesian. Model I: adjusted for sex, age, BMI, and educational level (*not adjusted for BMI as the estimation of obesity was based on BMI). BMI, body mass index; CI, confidence interval; CVD, cardiovascular disease; OR, odds ratio.

The majority of the population (72%) was overweight (35%) or obese (37%). Four percent was underweight. A third of the population of ≤40 years was obese, and women were more often obese than men (46 vs. 21%; P < 0.01). There were no differences between the 4 ancestry groups or between African and Asian ancestry participants.

A smaller proportion of the population had diabetes, 22%, but 41% of the participants had prediabetes. The majority of the participants with diabetes was aware of their condition (59%) and received treatment (54%), but only 21% of those treated were controlled (11% of all diabetics). There were no differences in diabetes prevalence between men and women (P = 0.19), but Asian ancestry participants were more likely to have diabetes than participants of African ancestry (30 vs. 18%; P < 0.01). No ethnic differences existed in the proportion of participants ≤40 years with diabetes (9% Asian vs. 7% African; P = 0.53). Cardiovascular risk profile of the HELISUR study participants 65

Ninety-five percent of the population had dyslipidemia (41%) or borderline lipid ab- normalities (54%). Men (51%) and Asian ancestry participants (51%) had more dyslip- idemia than women (37%; P < 0.01) and African ancestry participants (33%; P < 0.01). Moreover, dyslipidemia was more often prevalent in young Asian ancestry participants ≤40 years (39%) compared to their African counterparts (27%; P = 0.01). Tobacco use was seen in 18% of the population. Smokers were more often men (35% vs. 8% in women; P < 0.01), but no differences were apparent between ancestry or age groups.

The majority of the population (77%) had at least one cardiovascular risk factor, and this was 62% in the population subgroup of 40 years and younger. The inclusion of prehypertension, overweight, borderline dyslipidemia, and prediabetes resulted in 1% of the population with optimal cardiovascular parameters (2% in the population ≤40 years). Men (83%) and Asian ancestry participants (82%) had a higher risk factor burden than women (74%; P < 0.01) or African ancestry participants (74%; P < 0.01). 4

We found asymptomatic organ damage in 22% of the participants (9% in the population ≤40 years). In line with the high prevalence of hypertension, the majority had increased arterial stiffness (17%), followed by LVH and microalbuminuria (both 3%) and asymp- tomatic CKD (1%). Despite the higher risk factor burden, Asian ancestry participants were less likely to have asymptomatic organ damage than their African counterparts (21% vs. 24%; P = 0.21). Moreover, asymptomatic organ damage was more often seen in African ancestry participants ≤40 years than in their Asian counterparts (13% vs. 5%; P < 0.01).

Established CVD was found in 5% of the participants, including stroke (3%), coronary heart disease (2%), and severe CKD (0.4%). CVD occurred more often in men than in women (7 vs. 3%; P < 0.01), mainly due to a higher prevalence of stroke (4 vs. 2% in re- spectively men and women; P < 0.01), and in Asian ancestry participants than in African ancestry participants (6 vs. 3%; P = 0.01), mainly due to a higher prevalence of coronary heart disease (4 vs. 1%; P < 0.01). Multiple imputations did not alter our findings.

Discussion

Our elaborate cardiovascular assessment indicates that the burden of hypertension and other cardiometabolic risk factors is very high in this middle-income country. The major- ity of the population had hypertension (40%) or prehypertension (31%). This is in line with the finding that stroke is the largest contributor to cardiovascular mortality in this country, accounting for 46% of all cardiovascular deaths.8 With only 1% of the adult 66 Chapter 4

population now having an optimal cardiovascular risk factor profile, the coexistence of multiple cardiovascular risk factors, also in the young, and the low levels of control found in this study are ingredients for an increase in CVD of epidemic proportions in the coming decades, unless drastic primary preventive measures are taken.

At present, the main focus of health care for CVD in many LMIC is secondary preven- tion.24 Patients enter treatment programs after becoming symptomatic, when costly high-technology interventions are needed. However, the large proportion of individuals with suboptimal cardiovascular parameters underscores the importance and potential benefit of intensified screening and reduction of risk factor burden at an earlier stage. In line with this, clinicians could extend their case-finding strategy to virtually all adult patients as very few participants in our random sample (<5%) appeared to have an optimal cardiovascular risk profile. Importantly, while participants of Asian ancestry had a higher cardiovascular risk factor burden, the participants of African ancestry had more organ damage (arterial stiffness, LVH, microalbuminuria, and CKD). This suggests that clinicians should be aware of asymptomatic organ damage in the African ancestry patients despite an apparently lower risk factor burden.

This study has several strengths. First, we assessed the full spectrum of cardiovascular risk factors, and added noninvasive methods to assess early cardiovascular damage, including arterial stiffness. However, we need to establish whether these noninvasive assessments may help stratify cardiovascular risk profiles. Second, in Suriname, people of African and Asian ancestry live under relatively similar socioeconomic circum- stances. This setting is a rather unique feature of our study as in previous studies these ancestry groups are predominantly compared with Caucasians with great differences in socioeconomic backgrounds.4,5,6 Lastly, we reported the complete cardiovascular risk factor profile, and thus provide high-quality cardiovascular data from a middle-income country in a region from which such data are scarce.25 A limitation is the assessment of hypertension and diabetes through the measurement of blood pressure and HbA1c within one visit. Although repeated tests are required for confirmation of the diagnosis, with epidemiological studies this is a common approach. Also, the diagnosis of LVH was based on ECG data rather than echocardiographic data. It should be noted that the use of Sokolow-Lyon criteria has a lower specificity and a greater sensitivity in people of African ancestry compared to Asians for detecting LVH.26 This might have overestimated the prevalence of LVH in the African ancestry group.

In conclusion, in order to provide data for preventive and intervention strategies, we conducted an extensive examination of cardiovascular risk factors, asymptomatic organ damage, and CVD in the population of Suriname. We found that only 1% of the popula- Cardiovascular risk profile of the HELISUR study participants 67 tion had an optimal cardiovascular risk profile. Our findings underscore the urgent need for the implementation of cost-effective preventive strategies in LMIC to reduce blood pressure and other cardiovascular risk factors and prevent a surge in CVD of epidemic proportions.

4 68 Chapter 4

References

1. Krishnamurthi R, Feigin V, Forouzanfar M, et al. Global and regional burden of first-ever ischaemic and haemorrhagic stroke during 1990-2010: findings from the Global Burden of Disease Study 2010. Lancet Global Health. 2013; 1(5): e259-281. 2. Yusuf S, Rangarajan S, Teo K, et al. Cardiovascular risk and events in 17 low-, middle-, and high-income countries. The New England Journal of Medicine. 2014; 371(8): 818-827. 3. Baingana F, Bos E. Changing patterns of disease and mortality in Sub-Saharan Africa: an overview. In: Jamison D, Feachem R, Makgoba M, et al., eds. Disease and mortality in Sub- Saharan Africa. 2nd ed. Washington DC: World Bank; 2006. 4. Tillin T, Hughes A, Mayet J, et al. The relationship between metabolic risk factors and incident cardiovascular disease in Europeans, South Asians, and African Caribbeans: SABRE (Southall and Brent Revisited) -- a prospective population-based study. Journal of the American Col- lege of Cardiology. 2013; 61(17): 1777-1786. 5. Agyemang C, Kieft S, Snijder M, et al. Hypertension control in a large multiethnic cohort in. Int J Cardiol. 2015; 183: 180-189. 6. Admiraal W, Holleman F, Snijder M, et al. Ethnic disparities in the association of impaired fasting glucose with the 10-year cumulative incidence of type 2 diabetes. Diabetes Research and Clinical Practice. 2014; 103(1): 127-132. 7. United Nations General Assembly. Political declaration of the high-level meeting of the General Assembly on the prevention and control of non-communicable diseases. 2011. Available at: http://www.who.int/nmh/events/un_ncd_summit2011/political_declara- tion_en.pdf. 8. Brewster L. Cardiovascular death in Surinamese. 2015. Available at: https://figshare.com/ articles/ Cardiovascular_Death_in_Surinamese_by_Country_and_Ethnicity/1509931. 9. General Bureau of Statistics Suriname. Resultaten van de achtste (8e) volks- en woningtelling in Suriname (Volume I) - Demografische en sociale karakteristieken en migratie. Paramaribo 2012. 10. Diemer F, Aartman J, Karamat F, et al. Exploring cardiovascular health: the Healthy Life in Su- riname (HELISUR) study. A protocol of a cross-sectional study. BMJ Open. 2014; 4: e006380. 11. Mans D, Ganga D, Kartapawiro J. Meeting of the minds: traditional and herbal medicine in multiethnic Suriname. 2017. Available at: https://www.intechopen.com/books/aromatic- and-medicinal-plants-back-to-nature/meeting-of-the-minds-traditional-herbal-medicine- in-multiethnic-suriname. 12. Sode C. Suriname: an Asian immigrant and the organic creation of the Caribbean’s most unique fusion culture. 2015. Available at: http://history.rutgers.edu/undergraduate/ honors-program/honors-papers-2015/1302-sode-honors-thesis-2015/file. 13. Mancia G, Faqard R, Narkiewicz K, et al. 2013 ESH/ESC Guidelines for the management of arterial hypertension: the Task Force for the management of arterial hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens. 2013; 31(7): 1281-1357. 14. NHLBI Obesity Education Initiative Expert Panel on the Identification, Evaluation, and Treat- ment of Obesity in Adult. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report. Bethesda: National Heart, Lung, and Blood Institute; 1998. Cardiovascular risk profile of the HELISUR study participants 69

15. National Institute for Health and Care Excellence. Assessing body mass index and waist circumference thresholds for intervening to prevent ill health and premature death among adults from black, Asian and other minority ethnic groups in the UK. 2013. Available at: https://www.nice.org.uk/guidance/ph46. 16. American Diabetes Association. Classification and diagnosis of diabetes. Diabetes Care. 2016; 39: S13-S22. 17. American Diabetes Association. Glycemic targets. Diabetes Care. 2016; 39: S13-22. 18. Expert panel on detection, evaluation, and treatment of high blood cholesterol in adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA. 2001; 285(19): 2486-2497. 19. Van Bortel L, Laurent S, Boutouyrie P, et al. Expert consensus document on the measurement of aortic stiffness in daily practice using carotid-femoral pulse wave velocity. Journal of Hypertension. 2012; 30(3): 445-448. 20. Sokolow M, Lyon T. The ventricular complex in left ventricular hypertrophy as obtained by unipolar precordial and limb leads. American Heart Journal. 1949; 37: 161-186. 21. National Kidney Foundation. KDIGO 2012 Clinical practice guideline for the evaluation and 4 management of chronic kidney disease. Kidney International. 2013; 3(1). 22. Levey A, Stevens L, Schmid C, et al. A new equation to estimate glomerular filtration rate. Annals of Internal Medicine. 2009; 150(9): 604-612. 23. United Nations Educational, Scientific, and Cultural Organization (UNESCO). International standard classification of education: ISCED . 2011 Quebec: UNESCO Institute of Statistics; 2012. 24. World Health Organization. Prevention of cardiovascular disease: Guidelines for assess- ment and management of cardiovascular risk. 2007. Available at: http://www.who.int/ cardiovascular_diseases/guidelines/Full%20text.pdf. 25. Forouzanfar M, Liu P, Roth G, et al. Global burden of hypertension and systolic blood pres- sure of at least 110 to 115 mm Hg, 1990-2015. JAMA. 2017; 317(2): 165-182. 26. Spencer C, Beevers D, Lip G. Ethnic differences in left ventricular size and the prevalence of left ventricular hypertrophy among hypertensive patients vary with electrocardiographic criteria. Journal of Human Hypertension. 2004; 18(9): 631-636.

Chapter 5 Body composition and cardiovascular risk in high-risk ethnic groups

Frederieke S. Diemer, Lizzy M. Brewster, Yentl C. Haan, Glenn P. Oehlers, Gert A. van Montfrans, Lenny M. W. Nahar-van Venrooij

Clinical Nutrition (2017); doi: 10.1016/j.clnu.2017.11.012 72 Chapter 5

Abstract

Background: Cardiovascular disease (CVD) is highly prevalent in Suriname, a middle-income country with predominantly people of African and Asian ancestry. We examined whether the more comprehensive body composition measures determined by bioelectrical imped- ance analysis (BIA) are superior to the more traditional BMI and waist measures in rela- tion to cardiovascular risk.

Methods: Data from the cross-sectional Healthy Life in Suriname (HELISUR) study were used to calculate BMI, waist-hip ratio, waist-to-height ratio, and waist circumference. BIA was used to estimate fat percentage, fat-free mass index, and fat-to-fat-free mass ratio. High cardiovascular risk was defined as 1) a 10-year Framingham coronary heart disease risk score ≥10% in African-Surinamese and ≥12% in Asian-Surinamese, and 2) an increased arterial stiffness (pulse wave velocity >10 m/s). Using logistic regression analysis, we pre-selected the strongest correlate (i.e. lowest p-value below 0.05) of all body com- position items for both outcomes of cardiovascular risk separately, and subsequently, used forward logistic regression modelling to determine whether other measures added value to the initial model with the strongest correlate (-2 log-likelihood (-2LL) of initial model minus -2LL of new model, c-square statistic >3.841, 1 df). Analyses were adjusted for sex, age and ethnicity.

Results: We examined 691 participants (65% women; 48% African-Surinamese) with a mean age of 42 (SD 14) years. Waist circumference was the strongest correlate for high 10- year CVD risk in the total group, in men and African-Surinamese. In Asian-Surinamese, fat-free mass index was the strongest correlate of high 10-year CVD risk. Increased arte- rial stiffness was most strongly related with waist-to-height ratio in the total group and in African-Surinamese, and with BMI in men. None of the measures were significantly associated in women (for both outcomes) and Asian-Surinamese (for increased arterial stiffness). Forward selection showed that only BMI added value next to waist-to-height ratio in the total group in relation to increased arterial stiffness.

Conclusions: Waist measures, in particular waist circumference and waist-to-height ratio, and BMI should be used in African and Asian-Surinamese to identify who is at increased car- diovascular risk. Overall, we found little advantage in using BIA measures rather than simple anthropometric measures. Body composition and cardiovascular risk 73

Introduction

Obesity is an important independent risk factor for cardiovascular events, promoting atherogenesis.1,2,3 Although obesity is often defined using the body mass index (BMI), a growing base of evidence suggests that waist measures, such as waist circumference and waist-hip ratio, are better predictors of cardiovascular disease (CVD).4,5 Further- more, body composition is known to vary between different ethnic groups.6,7 At the same level of BMI, Asian populations have more total body fat and more abdominal fat, whereas Afro-Americans showed higher fat-free mass and less body fat compared to Caucasians.6,7 High fat-free mass, a reflection of high muscle mass, has been linked to a lower risk for CVD, which might be mediated by physical activity.8 Moreover, a low fat- free mass together with a high fat mass (i.e. sarcopenic obesity) is proposed as a more sensitive predictor of cardiovascular risk, carrying the cumulative risk derived from each of the two individual body composition phenotypes.9 The use of body composition measures such as fat mass, fat-free mass or fat-to-fat-free mass ratio might therefore improve CVD risk prediction, particularly in multi-ethnic populations.

The burden of CVD in low- and middle-income countries (LMIC) has increased 5 greatly.10,11,12 In Suriname, a middle-income country in South America with an African and Asian ancestry population, CVD is the number one cause of mortality.13,14,15 Drastic measures should be taken to prevent further increases in CVD morbidity and mortality. Therefore, it is vital to accurately identify people at risk for CVD. A recent study in Suri- name showed that in most ethnic groups waist circumference had more discriminatory power than BMI in the relation with hypertension, diabetes, and an adverse cardio- metabolic risk profile.16 However, it remains unclear if more complex body composition measures determined by bioelectrical impedance analysis (BIA), such as fat percentage, fat-free mass index, and fat-to-fat-free mass ratio, could be useful in CVD risk assess- ment. Therefore, we examined whether BIA measures (fat percentage, fat-free mass index, fat-to-fat-free mass ratio) are superior to the simple anthropometric measures (BMI, waist circumference, waist-hip ratio, waist-to-height ratio) in the association with CVD risk in people of African and Asian ancestry.

Methods

Study population The study population consisted of participants of the Healthy Life in Suriname (HELI- SUR) study, a population-based observational study, as previously described.17 The HELISUR study was conducted according to the principles of the Declaration of Helsinki 74 Chapter 5

(59th WMA General Assembly, Seoul, October 2008) and in accordance with the Medical Research Involving Human Subjects Act. Ethical clearance was obtained from the Ethics Committee of the Ministry of Health in Suriname in 2012 (Approval nr. VG021-2012).

In brief, we randomly selected a representative multi-ethnic sample of 1800 non- institutionalized men and women aged 18-70 years living in the capital Paramaribo.17 Eligible subjects were interviewed at home and subsequently invited for an examina- tion at the local hospital.17 A total of 1157 subjects participated in both the interview and the physical examination.

Data collection

Questionnaire Information on demographic factors, tobacco smoking, use of medication, and CVD his- tory were collected by means of a questionnaire. Ancestry was self-defined.

Physical examination Participants were examined in the morning in a fasting state. Height and weight were measured in duplicate to the nearest 0.1 cm and 0.1 kg, respectively, with participants wearing no shoes and only light underwear. Waist circumference was obtained with a measuring tape, measured midway between the lower rib and the spina iliaca anterior to the nearest 0.1 cm. In more obese subjects for whom bone palpation was difficult, the level of the umbilicus was utilized to measure waist circumference. Hip circumference was measured at the level of the trochanter major to the nearest 0.1 cm. Repeat mea- surements were obtained for waist and hip circumferences, and a third measurement was made if the difference between the first two readings was more than 1.0 cm.

Body composition was assessed using bioelectrical impedance analysis (BIA) (Bodys- cout, Fresenius Kabi, ‘s Hertogenbosch, the Netherlands). BIA measures the electrical impedance of the body by introducing a small alternating electrical current (5-800 mA, 5 kHz-1 MHz) into the body. Adhesive electrodes were placed using the four-electrode method, as described by Lukaski et al.18 Impedance was measured on the left side of the participant in supine position. The measurement provides a rapid, non-invasive, and relatively reliable estimation of fat mass and fat-free mass with minimal intra- and inter-observer variability and <1% error on repeated measurements.18,19,20

Aortic pulse wave velocity was estimated non-invasively in duplicate in the supine po- sition by analysis of the oscillometric pressure curves registered on the right upper arm, using the validated Arteriograph device (TensioMed, Budapest, Hungary).21 Blood pres- Body composition and cardiovascular risk 75 sure was measured twice in the sitting position with an automated oscillometric device (WatchBP Office; Microlife AG, Widnau, Switzerland) and an appropriately adjusted cuff size on the left upper arm supported at heart level. Fasting blood samples were drawn to measure glucose and lipid spectrum.

Definitions Participants were mainly Asian (self-identified South Asian and Indonesian ancestry); or African-Surinamese (self-identified Creoles and Maroons), hitherto referred to as Afri- can and Asian.15 Persons of other ancestry (Native Surinamese, and persons of Chinese, European, and other ancestry) were excluded for the current analysis.

BMI was computed as weight divided by height squared. To capture abdominal obesity, we used waist-hip ratio (waist divided by hip), waist-to-height ratio (waist divided by height) and waist circumference. BIA measures included fat percentage (fat mass divided by weight x 100%), and fat-free mass index (fat-free mass divided by height squared), as a surrogate measure of muscle mass. Furthermore, we used fat-to-fat-free mass ratio (fat mass divided by fat-free mass) as a measure of sarcopenic obesity.9 5 To define high CVD risk, we used the Framingham Risk Score equation to calculate the 10-year coronary heart disease (CHD) risk.22 Subsequently, we used the lowered ethnic- specific thresholds as proposed by Cappuccio and colleagues of >12% in Asians and >10% in Africans to estimate the 10-year risk for combined CVD (i.e. CHD plus stroke).23 Additionally, we included pulse wave velocity (PWV), which is known as a predictor of CVD events, and defined increased arterial stiffness as a PWV >10 m/s.24

Statistical analysis We assessed differences in CVD risk and body composition between African and Asian men and women using the chi-square test and Student t-test. Subsequently, continu- ous data on different measures of body composition were transferred into categories based on international standards: BMI (African: <18.5, 18.5-25.0, 25.0-30.0, >30.0; Asian: <18.5,18.5-23.0, 23.0-27.5, >27.5), waist-hip ratio (men: ≥0.90; African women: ≥0.85; Asian women: ≥0.80), waist-to-height ratio (<0.50 or ≥0.50), waist circumference (African men: >94 cm; Asian men: >90 cm; women: >80 cm), fat percentage (men: ≤18, 18-25, >25; women: ≤25, 25-35, >35), fat- free mass index (men: <16.7; women: <14.6), and fat-to-fat-free mass ratio (quartiles: <0.38, 0.38-0.58, 0.58-0.81, >0.81). Five mod- els using logistic regression analysis were generated; for the total study population, for men, women, Asian and African. Categorized body composition measures were entered as independent variables and CVD risk as the dependent variable. First, using logistic regression analysis, we assessed the strongest correlate of CVD risk in pre-screening 76 Chapter 5

using the lowest p-value below p < 0.05 (i.e. highest Wald statistic). Subsequently, us- ing forward multivariable logistic regression modelling, we added the remaining body composition measures that were significantly associated with CVD risk in pre-screening one by one to the model with the strongest correlate. Due to multicollinearity issues, only one of the three waist measures was included in the multivariable model. To as- sess whether a body composition measure added value, we used the model chi-square statistic, which is equal to -2LL of the new model minus the value of the -2LL of the initial model. This value has a chi-square distribution, indicating that a significant better model should have a model chi-square statistic of >3.841 at the p = 0.05 level with 1 degree of freedom (df) (for 2 df >5.991; for 3 df >7.815, at the p = 0.05 level).25,26 All analyses were adjusted for sex, age, and ethnicity. Statistical analysis of the data was carried out using the SPSS version 20.0 (SPSS Inc., Chicago, IL, USA) software for Windows.

Results

Subjects Body composition was assessed in 733 participants. Of these, we excluded participants with a prior CVD event (n = 42). Thus, we analysed data of 691 participants (65% women; 48% African) with a mean age of 42.2 (SD 13.5) years (Table 1). High 10-year CVD risk was seen in 19.7% of the participants, with higher proportions in men than in women for both ethnic groups. Pulse wave velocity was measured in a subsample of 517 participants (56% women; 50% African). The overall prevalence of increased arterial stiffness was 14.5% and higher in women than in men for both ethnic groups.

Ethnic differences in body composition As depicted in Table 2, the majority of the population was overweight (33.6%) or obese (40.1%), with 71.6% up to 78.4% being abdominally obese depending on the specific measure used (waist-hip ratio, waist-to-height ratio or waist circumference). Abdominal obesity was less prevalent in African men (28.8% up to 50.5%; p < 0.01). Compared to African men, Asian men had higher waist-hip ratio/waist-to-height ratio/waist circum- ference, fat percent-age, and fat-to-fat-free mass ratio, and lower fat-free mass index (all p < 0.01), although BMI levels were similar (p = 0.73). Despite higher BMI levels in African women, Asian women had higher waist-hip ratio, lower fat-free mass index (all p < 0.01) but equal levels of fat percentage and fat-to-fat-free-mass ratio (both p > 0.05). Body composition and cardiovascular risk 77

Table 1. Characteristics of the total study population and by sex and ancestry African Asian Total Men Women Men Women (n=691) (n=111) (n=219) (n=132) (n=229) Framingham characteristics Age, years* 42.2 ± 13.5 40.7 ± 14.8 41.2 ± 13.3 43.0 ± 13.3 43.4 ± 13.1 Tobacco smoking, % 17.8 38.7 7.3§ 34.1 8.3§ Systolic blood pressure, mmHg* 128.6 ± 20.0 131.3 ± 19.0 128.9 ± 22.1 127.9 ± 17.2 127.3 ± 20.0 Cholesterol, mmol/L* 4.7 ± 1.1 4.1 ± 1.0 4.5 ± 1.0§ 4.9 ± 0.9║ 5.1 ± 1.1║ HDL, mmol/L* 1.2 ± 0.3 1.2 ± 0.3 1.3 ± 0.3§ 1.1 ± 0.2║ 1.2 ± 0.3§║ Diabetes mellitus, % 14.7 3.6 8.2 20.6║ 22.8║ CVD risk Framingham CHD risk score† 4.0 (1.0-10.0) 4.0 (2.0-10.0) 1.0 (1.0-7.0)§ 7.0 (3.0-13.0)║ 3.0 (1.0-10.0)§║ High 10-year CVD risk, % 19.7 27.0 17.4§ 25.8 14.8§║ Pulse wave velocity*‡ 8.3 ± 2.3 7.5 ± 1.8 8.6 ± 2.4§ 7.8 ± 1.9 9.0 ± 2.7§ Increased arterial stiffness, %‡ 14.5 9.9 15.5§ 10.6 17.9§ Footnote Table 1. *Mean with standard deviation. †Median with interquartile range. ‡Based on a different sample size: n = 518. §Significantly different from men with the same ancestry (p < 0.05). ║Significantly different from Africans of the same sex (p < 0.05). High 10-year CVD risk was defined as a 10-year risk of CHD ≥12% in Asians and ≥10% in Africans. Increased arterial stiffness was defined as a pulse wave velocity >10 m/s. 5 HDL, high-density lipoprotein; CVD, cardiovascular disease; CHD, coronary heart disease.

Table 2. Body composition measures of the total study population and by sex and ethnicity African Asian Women Women Total (n=691) Men (n=111) (n=219) Men (n=132) (n=229) BMI 27.9 ± 6.2 25.2 ± 5.1 30.4 ± 6.7* 25.5 ± 5.0 28.1 ± 5.6*† Underweight, % 4.1 4.5 2.7 7.6 3.1 Normal weight, % 22.3 49.5 19.6* 28.0† 15.3*† Overweight, % 33.6 28.8 24.7 44.7† 37.6† Obesity, % 40.1 17.1 53.0* 25.8 47.2* WHR 0.92 ± 0.10 0.90 ± 0.08 0.90 ± 0.10 0.96 ± 0.08† 0.93 ± 0.11*† High WHR, % 71.6 47.7 63.5* 75.0† 89.1*† WHtR 0.58 ± 0.10 0.50 ± 0.08 0.61 ± 0.11* 0.55 ± 0.08† 0.61 ± 0.10* High WHtR, % 78.4 50.5 84.9* 72.0† 89.5* WC 95.0 ± 15.5 88.1 ± 14.3 98.8 ± 16.7* 93.2 ± 13.8† 95.8 ± 14.8 High WC, % 71.8 28.8 84.9* 59.1† 87.3* BF% 35.5 ± 11.7 21.6 ± 9.2 40.9 ± 8.7* 27.5 ± 9.2† 41.5 ± 7.6* Low BF%, % 11.1 37.8 3.7* 16.7† 2.2* 78 Chapter 5

Table 2. Body composition measures of the total study population and by sex and ethnicity (con- tinued) African Asian Women Women Total (n=691) Men (n=111) (n=219) Men (n=132) (n=229) Elevated BF%, % 20.7 27.9 21.5 18.9 17.5 High BF%, % 68.2 34.2 74.9* 64.4† 80.3* FFMI 17.5 ± 2.6 19.4 ± 2.5 17.5 ± 2.5* 18.1 ± 2.1† 16.1 ± 2.3*† Low FFMI , % 17.1 9.0 11.0 19.7† 25.3† FM/FFM 0.60 ± 0.29 0.29 ± 0.16 0.73 ± 0.25* 0.40 ± 0.19† 0.74 ± 0.23* FM/FFM Q1, % 24.7 73.0 5.9* 50.0† 4.8* FM/FFM Q2, % 25.8 22.5 25.6 35.6† 21.8* FM/FFM Q3, % 25.0 3.6 30.1* 11.4† 38.4* FM/FFM Q4, % 24.5 0.9 38.4* 3.0 34.9* Footnote Tabel 2. Values are mean with standard deviation, except if stated otherwise. *Signifi- cantly different from men with the same ancestry (p < 0.05). †Significantly different from Africans of the same sex (p < 0.05). Underweight, normal weight, overweight, and obesity respectively, body mass index (BMI) < 18.5, 18.5-25.0, 25.0-30.0, >30.0 in Africans, <18.5, 18.5-23.0, 23.0-27.5, >27.5 in Asians; high waist- hip ratio (WHR), men ≥0.90, African women ≥0.85, Asian women ≥0.80; high waist-to-height ratio (WHtR), ≥0.50; high waist circumference (WC), African men >94 cm, Asian men >90 cm, women >80 cm; low, elevated, and high body fat percentage (BF%) respectively, men ≤18, 18-25, >25, women ≤25, 25-35, >35; low fat-free mass index (FFMI), men <16.7, women <14.6; fat-to-fat-free mass ratio (FM/FFM) quartiles (Q), Q1 <0.38, Q2 0.38-0.58, Q3 0.58-0.81, Q4 >0.81.

Body composition measures and cardiovascular risk Table 3A shows the pre-screening of body composition measures in the association with a high 10-year CVD risk, adjusted for sex, age, and ethnicity. Adjusted odds ratios (ORs) showed that overall participants with high waist measures had more often a high 10-year CVD risk. In line with this, the lowest p-value (i.e. strongest correlate) was seen for waist circumference in the total population, in men and in Africans. In Asians, fat-free mass index was the strongest correlate of a high 10-year CVD risk. In women, none of the body composition measures was significantly associated with the outcome. Forward logistic regression modelling showed that neither BMI, waist circumference, fat percentage nor fat-to-fat-free mass ratio added value over the strongest correlate (i.e. waist circumference in the total group, men and African; fat-free mass index in Asian) in the association with high 10-year CVD risk (p > 0.05 for all multivariable models; Table 4A).

Table 3B shows the pre-screening of body composition measures in association with an increased arterial stiffness. Adjusted ORs showed that overall participants with higher categories of BMI, waist-to-height ratio, waist circumference, fat percentage and Body composition and cardiovascular risk 79 p b 0.22 0.92 0.47 0.29 0.20 0.96 0.49 - 3 1 1 1 2 1 3 df (0.32 - 2.59) (0.37 - 3.20) (1.04 - 8.94) (0.89 - 72.69) (0.91 - 23.94) (1.07 - 45.95) reference reference reference reference reference Asian (n=261) - Adjusted O R (95%CI) Adjusted 0.91 1.09 8.04 4.66 7.02 3.05 Wald 4.412 0.010 0.516 1.111 3.271 0.003 2.422 p b 0.25 0.05 0.34 0.03 0.01 0.02 0.03 3 1 1 1 2 1 3 df (0.00 - 32.98) (0.73 - 16.80) (0.92 - 18.92) (0.54 - 35.85) (1.25 - 25.46) (1.67 - 31.35) (2.40 - 55.50) reference reference reference reference reference African (n=257) Adjusted O R (95%CI) Adjusted 0.27 3.50 4.16 4.40 5.64 7.24 Wald 1.309 5.892 0.930 8.716 5.982 5.513 9.270 11.55 p a 0.46 0.81 0.06 0.11 0.99 0.63 0.30 - - 3 1 1 1 2 1 3 df (0.28 - 4.25) (0.46 - 5.61) (0.59 - 7.26) (0.29 - 33.15) (0.49 - 44.56) reference reference reference reference reference omen (n=291)

W 5 - - Adjusted O R (95%CI) Adjusted 1.10 1.61 2.07 3.07 4.66 Wald 2.592 0.056 3.457 2.604 0.010 0.238 3.680 p 0.18 0.10 0.09 0.27 0.91 0.08 <0.01 a (0.01 - 7.20) (0.73 - 5.80) (0.76 - 7.96) (1.13 - 9.58) 3 1 1 1 2 1 3 (0.73 - 11.25) df (1.38 - 14.95) (1.68 - 11.48) Adjusted Adjusted reference reference reference reference reference O R (95%CI) M en (n=226) 0.29 2.05 2.46 2.86 4.54 3.29 4.39 Wald 1.799 2.776 2.970 2.658 0.014 6.660 15.167 p 0.31 0.10 0.64 0.01 0.01 0.02 0.05 3 1 1 1 2 1 3 df (0.01 - 2.18) (0.74 - 3.85) (0.99 - 5.14) (1.37 - 8.08) (1.28 - 8.90) (1.92 - 10.22) (1.31 - 16.88) reference reference reference reference reference Total (n=517) Total Adjusted O R (95%CI) Adjusted 0.13 1.68 2.25 3.33 3.38 4.43 4.69 Wald 1.051 4.523 0.214 6.016 5.155 7.843 11.330 Results of the pre-screening of candidate correlates of A) high cardiovascular disease risk and B) increased arterial stiffness, adjusted for sex, age, age, sex, for adjusted stiffness, arterial increased B) and risk disease cardiovascular high A) of correlates candidate of pre-screening the of Results High cardiovascular disease risk High cardiovascular Table 3. Table and ethnicity A) Measure BMI WHR WHtR WC BF% FFMI FM/FFM Category Normal weight Normal weight Underweight Overweight Obesity WHR Low High WHR Low WHtR Low High WHtR Low WC Low High WC Elevated BF% Elevated Low/normal BF% Low/normal 80 Chapter 5 p b 0.97 0.06 0.07 0.13 0.89 0.04 0.04 3 1 1 1 2 1 3 df (0.44 - 3.01) (0.40 - 3.55) (0.23 - 2.92) (0.30 - 2.90) (0.64 - 6.00) (0.10 - 0.92) (0.85 - 27.86) (0.34 - 25.68) reference reference reference Asian (n=261) Adjusted O R (95%CI) Adjusted Adjusted O R (95%CI) Adjusted 1.15 1.19 4.86 0.83 2.95 0.94 1.96 0.31 Wald 0.226 3.445 3.399 4.137 0.631 4.139 4.426 p b 0.07 0.07 0.93 0.05 0.04 0.02 <0.01 - 3 1 1 1 2 1 3 df (0.26 - 4.39) (0.56 - 7.22) (0.88 - 40.61) (2.51 - 66.71) (1.57 - 102.9) (1.71 - 76.56) (1.57 - 17.23) reference reference Adjusted Adjusted reference O R (95%CI) African (n=257) - Adjusted O R (95%CI) Adjusted 5.97 1.07 2.01 5.21 Wald 3.216 3.409 0.008 12.94 12.72 8.023 11.45 9.331 6.500 10.445 p a 0.75 0.26 0.36 0.18 0.83 0.24 0.33 - - 3 1 1 1 2 1 3 df (0.15 - 1.60) (0.28 - 2.32) (0.54 - 3.75) (0.78 - 98.25) (0.51 - 54.71) (0.55 - 56.93) reference reference reference omen (n=291) W - - Adjusted O R (95%CI) Adjusted Adjusted O R (95%CI) Adjusted 8.77 5.26 0.49 5.60 0.81 1.42 Wald 1.229 1.281 0.856 1.782 0.363 1.386 3.404 p 0.30 0.06 0.24 0.11 0.01 0.03 <0.01 a (0.74 - 4.54) (0.39 - 6.20) (0.14 - 1.63) 3 1 1 1 2 1 3 (0.54 - 11.55) df (1.30 - 16.54) (3.12 - 72.01) (1.54 - 714.95) (1.65 - 239.97) Adjusted Adjusted reference reference Adjusted Adjusted reference O R (95%CI) O R (95%CI) M en (n=226) 1.83 1.55 0.47 2.51 4.63 Wald 3.650 5.652 1.408 5.965 33.23 6.204 4.752 9.097 14.99 19.92 p 0.07 0.09 0.17 0.01 0.02 <0.01 <0.01 3 1 1 1 2 1 3 df (0.75 - 4.92) (0.85 - 6.28) (0.20 - 1.13) (0.56 - 2.93) (0.23 - 9.19) (1.10 - 5.54) (1.33 - 6.49) (1.64 - 17.49) reference reference reference Total (n=517) Total Adjusted O R (95%CI) Adjusted Adjusted O R (95%CI) Adjusted 1.92 2.32 0.48 1.28 1.45 2.47 5.36 2.94 Wald 6.936 2.856 5.033 7.043 6.069 7.735 12.162 High cardiovascular disease risk ( continued ) High cardiovascular arterial stiffness Increased A) Category B) Measure FM/FFM Q2 FM/FFM Q3 FM/FFM Q4 BMI FM/FFM Q1 High BF% High FFMI FFMI Low WHR WHtR WC BF% FFMI Overweight FM/FFM Obesity Category Normal weight Underweight Body composition and cardiovascular risk 81 (0.28 - 4.09) (0.39 - 2.68) (0.38 - 3.29) (0.43 - 6.31) (0.60 - 5.40) (0.40 - 4.34) (0.64 - 8.35) (0.41 - 34.21) (0.64 - 62.32) reference reference reference reference reference reference Adjusted O R (95%CI) Adjusted 1.07 3.75 1.03 1.11 6.33 1.64 1.80 1.32 2.32 (0.69 - 4.21) (0.09 - 2.24) (0.37 - 37.06) (1.75 - 35.74) (1.04 - 79.86) (1.27 - 14.68) (2.06 - 64.64) (2.40 - 81.40) (1.68 - 111.62) Adjusted Adjusted reference reference reference reference reference reference O R (95%CI) 1.70 0.46 3.70 7.91 9.12 4.32 2.06 2.40 13.70 - - (0.44 - 2.84) (0.29 - 2.11) (0.44 - 14.24) (0.76 - 16.86) (0.90 - 58.84) (0.44 - 13.20) (0.72 - 21.04) reference reference reference reference reference reference 5 - - Adjusted O R (95%CI) Adjusted 1.12 0.78 2.50 3.58 7.27 2.41 3.88 (0.65 - 9.88) (0.23 - 3.68) (0.72 - 5.72) (0.89 - 6.62) (0.82 - 12.13) (0.75 - 18.65) (0.63 - 20.27) (1.35 - 21.59) (0.36 - 170.62) Adjusted Adjusted reference reference reference reference reference reference O R (95%CI) 2.54 3.15 0.92 2.02 3.74 2.42 3.56 0.36 1.35 (0.70 - 3.17) (0.37 - 1.85) (0.98 - 5.42) (1.10 - 7.34) (1.14 - 5.64) (1.32 - 11.72) (1.13 - 24.17) (1.04 - 25.31) (1.53 - 11.04) reference reference reference reference reference reference Adjusted O R (95%CI) Adjusted 1.49 0.83 2.30 3.93 2.85 5.24 2.53 5.13 4.11 Increased arterial stiffness ( continued ) arterial stiffness Increased Significant associations are in bold. Values are odds ratios (OR) with 95% confidence intervals (CI), adjusted for sex, age, and ethnicity. age, sex, for intervals (CI), adjusted confidence ratios (OR) with 95% odds Values are in bold. are 3. Significant associations ootnote Tabel Adjusted for age and sex. age for Adjusted Adjusted for age and ethnicity. and ethnicity. age for Adjusted B) Category F BMI, body mass index; WHR, waist-hip ratio; WHtR, waist-to-height ratio; WC, waist BF%, circumference; body fat percentage; FFMI, fat-free mass index; quartiles. Q, ratio; mass FM/FFM, fat-to-fat-free a b Low WHR Low High WHR WHtR Low High WHtR High FFMI FFMI Low FM/FFM Q1 FM/FFM Q2 FM/FFM Q3 High BF% Low WC Low High WC BF% Low/normal BF% Elevated FM/FFM Q4 82 Chapter 5

fat-to-fat-free mass ratio were more likely to have increased arterial stiffness. In the pre-screening, increased arterial stiffness was most strongly associated with waist-to- height ratio in the total group and in the Africans, and with BMI in men. In women and Africans, none of the body composition measures were associated with increased arte- rial stiffness in the pre-screening. In forward logistic regression modelling, adding BMI to waist-to-height ratio increased the value of the model in the total group (χ-square statistic with 3 df = 8.935, p = 0.03; critical value of the χ-square statistic with 3 df = 7.815, at the p = 0.05 level; Tabel 4B).

Table 4. Forward logistic regression modelling to assess the strongest associated body composi- tion measures with A) high cardiovascular disease risk and B) increased arterial stiffness A) High cardiovascular disease risk Strongest Other significantly Initial New model Model χ2 correlate associated measures -2LLa -2LLb statistic df p Total (n=691) WC + BF% 359.125 356.562 2.563 2 0.28 Men (n=243) WC - - - - - Women (n=448) ------African (n=330) WC + FM/FFM 127.031 121.924 5.107 3 0.16 + BF% 127.031 125.461 1.570 2 0.46 + BMI 127.031 126.767 1.264 3 0.74 Asian (n=361) FFMI + WC 219.744 216.421 3.323 1 0.07 B) Increased arterial stiffness Strongest Other significantly Initial New model Model χ2 correlate associated measures -2LLa -2LLb statistic df p Total (n=518) WHtR + BMI 378.297 369.362 8.935 3 0.03 + FM/FFM 378.297 374.871 3.427 3 0.33 Men (n=226) BMI ------Women (n=291) ------African (n=257) WHtR + FM/FFM 167.572 163.832 3.740 3 0.29 + BMI 167.572 162.595 4.978 3 0.17 Asian (n=261) ------Footnote Tabel 4. a Initial -2 log-likelihood (-2LL) model included sex, age, ethnicity and the body composition measure that was most strongly related to the outcome in the pre-screening (for men/ women: not adjusted for sex; for African/Asian: not adjusted for ethnicity). b New -2LL model in- cluded the initial model plus a statistically significantly associated body composition measure. If the model χ2 statistic >3.84 with 1 degree of freedom (df) (i.e. p < 0.05), the new model added value over and above the initial model. The body composition measure or set of measures that were most strongly correlated with the outcome of interest are depicted in bold. BMI, body mass index; WHtR, waist-to-height ratio; WC, waist circumference; BF%, body fat per- centage; FFMI, fat-free mass index; FM/FFM, fat-to-fat-free mass ratio. Body composition and cardiovascular risk 83

Discussion

The main finding of this study is that cardiovascular risk in people of African and Asian ancestry overall is most strongly associated with waist measures, in particular waist circumference and waist-to-height ratio, independently of sex, age and ethnicity. BMI seemed to have additional value in the total group and in men in the association with increased arterial stiffness. Overall, we found little advantage in using BIA measures rather than simple anthropometric measures.

It is growingly accepted that waist measures are important determinants of obesity- related diseases; possibly more so than general obesity measures such as BMI.3,4,27,28 Correspondingly, we showed in an African and Asian population that overall BMI was inferior to waist measures in the association with CVD risk. One reason why waist measures best predict cardiovascular risk is the high correspondence of waist measures with the amount of visceral adiposity tissue. This is in turn strongly associated with a range of metabolic disturbances, such as impaired glucose tolerance and dyslipid- emia.29 Nevertheless, BMI was of additional value in the total group and in men using increased arterial stiffness as outcome for cardiovascular risk and, therefore, should not 5 be ruled out.

Our data implicate that more comprehensive BIA measures seem of little additional value to waist measures in cardiovascular risk prediction in people of African and Asian ancestry. This is in line with previous large-scale studies demonstrating that once waist- hip ratio is taken into account, fat percentage measured by BIA did not add to the pre- diction of cardiometabolic abnormalities30, increased arterial stiffness31, cardiovascular disease32, or mortality32. However, measuring the amount of visceral adipose tissue directly appeared to be a good predictor of CVD risk, and even improved risk prediction over BMI.31,33 This finding supports the hypothesis that the distribution rather than the overall amount of adipose tissue is important in CVD risk prediction.

Fat-free mass index showed benefits in Asian-Surinamese in the association with 10- year CVD risk. In contrast to studies finding no association8 or a negative association34, we found a positive association, indicating that persons with high fat-free mass index had more often a high CVD risk. Further exploration of the data showed that persons with high fat-free mass index also had higher waist-hip ratio, waist-to-height ratio or waist circumference. Second, persons with high fat-free mass index had higher systolic blood pressure levels, which is a determinant of the Framingham Risk Score equation.22 More research should elucidate if the use of fat-free mass index and fat-to-fat-free mass ratio aside waist measures improves the prediction of cardiovascular risk, especially as 84 Chapter 5

these measures require more resources and time compared to simple anthropometric measurements.

The present study also demonstrated ethnic differences in body composition suggest- ing that at similar BMI level Asians generally have more fat mass than Africans. This is in line with the international literature and supports the adoption of lower BMI and waist circumference thresholds for Asian populations compared to other ethnic popula- tions.6,16

Results from this study should be understood with a few limitations in mind. First, the cross-sectional nature of our study implies that causal associations cannot be estab- lished. Second, the Framingham risk score was based on an all-white cohort. We used the recommended ethnic-specific cut-off values for African and Asian populations23; how- ever, this was not validated for the Surinamese population in specific. Although some information bias might exist, this algorithm provides an initial heuristic for exploring risk patterns in Suriname. More importantly, we found similar results with pulse wave velocity, which is a more direct physical measure of damage of the vascular system.35 Nevertheless, prospective studies are needed to relate body composition measures with hard cardiovascular outcomes. Finally, we used bioelectrical spectroscopy to esti- mate fat percentage and fat-free mass index. It estimates fat-free mass based on values of total body water. Therefore, in obese subjects who have a relatively high amount of total body water, fat-free mass might be overestimated and consequently fat mass might be underestimated.18 Nevertheless, compared to single and multi-frequency bio- impedance analysis, bioelectrical spectroscopy relies less on assumptions that may be violated in disease states and is highly correlated with DEXA.19,36

In conclusion, CVD is the main cause of death in Suriname, and accurate identification of adults who are at high risk for obesity-related diseases is urgently needed. We found that overall waist measures are overall the strongest correlates of a high cardiovascular risk in this population of African and Asian ancestry. Furthermore, BMI should not be ruled out, but the more complex BIA measures seemed of no clinically relevant addi- tional value in the association with CVD risk. The use of waist measures and BMI can be well applied in resource-limited settings, where the health care infrastructure is less developed. Body composition and cardiovascular risk 85

References

1. McGee D. Body mass index and mortality: a meta-analysis based on person-level data from twenty-six observational studies. Annals of Epidemiology. 2005;15(2):87-97. 2. World Health Organization. Obesity and overweight. Geneva: WHO; 2015. 3. van Rooy MJ, Pretorius E. Obesity, hypertension and hypercholesterolemia as risk factors for atherosclerosis leading to ischemic events. Current Medical Chemistry. 2014;21(19):2121- 2129. 4. Staiano A, Reeder B, Elliott S, et al. Body mass index versus waist circumference as predictors of mortality in Canadian adults. International Journal of Obesity. 2012;36(11):1450-1454. 5. Srikanthan P, Seeman T, Karlamangla A. Waist-hip-ratio as a predictor of all-cause mortality in high-functioning older adults. Annals of Epidemiology. 2009;19(10):724-731. 6. National Institute for Health and Care Excellence. Assessing body mass index and waist circumference thresholds for intervening to prevent ill health and premature death among adults from black, Asian and other minority ethnic groups in the UK. 2013. Available at: https://www.nice.org.uk/guidance/ph46. 7. Rao G, Powell-Wiley T, Ancheta I, et al. Identification of Obesity and Cardiovascular Risk in Ethnically and Racially Diverse Populations: A Scientific Statement From the American Heart Association. Circulation. 2015;132(5):457-472. 8. Czernichow S, Bertrais S, Oppert J, et al. Body composition and fat repartition in relation to structure and function of large arteries in middle-aged adults (the SU.VI.MAX study). International Journal of Obesity. 2005;29(7):826-832. 5 9. Prado C, Wells J, Smith S, Stephan B, Siervo M. Sarcopenic obesity: a critical appraisal of the current evidence. Clinical Nutrition. 2012;31(5):583-601. 10. World Health Organization. Global status report on noncommunicable disease 2014. Geneva 2014. ISBN 978 92 4 156485 4. 11. Krishnamurthi R, Feigin V, Forouzanfar M, et al. Global and regional burden of first-ever ischaemic and haemorrhagic stroke during 1990-2010: findings from the Global Burden of Disease Study 2010. Lancet Global Health. 2013;1(5):e259-281. 12. Mathers C, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Medicine. 2006;3(11):2011-2030. 13. Brewster L. Cardiovascular death in Surinamese. 2015. Available at: https://figshare.com/ articles/Cardiovascular_Death_in_Surinamese_by_Country_and_Ethnicity/1509931. 14. Punwasi W. Doodsoorzaken Suriname 2010-2011. Paramaribo: Ministry of Health; 2012. 15. General Bureau of Statistics Suriname. Resultaten van de achtste (8e) volks- en woningtelling in Suriname (Volume I) - Demografische en sociale karakteristieken en migratie. Paramaribo 2012. 16. Krishnadath I, Toelsie J, Nahar-van Venrooij L, Hofman A, Jaddoe V. Ethnic and sex-specific cut-off values for adult obesity in the Suriname Health Study. Obesity Research and Clinical Practice. 2016:pii: S1871-403X(16)30111-9. 17. Diemer F, Aartman J, Karamat F, et al. Exploring cardiovascular health: the Healthy Life in Su- riname (HELISUR) study. A protocol of a cross-sectional study. BMJ Open. 2014;4:e006380. 18. Lukaski H, Johnson P, Bolonchuk W, Lykken G. Assessment of fat-free mass using bioelectri- cal impedance measurements of the human body. American Journal of Clinical Nutrition. 1985;41(4):810-817. 86 Chapter 5

19. Earthman C, Traughber D, Dobratz J, Howell W. Bioimpedance spectroscopy for clinical assessment of fluid distribution and body cell mass. Nutrition in Clinical Practice. 2007;22(4):389-405. 20. Segal K, Burastero S, Chun A, Coronel P, Pierson RJ, Wang J. Estimation of extracellular and total body water by multiple-frequency bioelectrical-impedance measurement. American Journal of Clinical Nutrition. 1991;54(1):26-29. 21. Horvath I, Nemeth A, Lenkey Z, et al. Invasive validation of a new oscillometric device (Arte- riograph) for measuring augmentation index, central blood pressure and aortic pulse wave velocity. J Hypertens. 2010;28(10):2068-2075. 22. Wilson P, D’Agostino R, Levy D, Belanger A, Silbershatz H, Kannel W. Prediction of coronary heart disease using risk factor categories. Circulation. 1998;97(18):1837-1847. 23. Cappuccio F, Oakeshott P, Strazzullo P, Kerry S. Application of Framingham risk estimates to ethnic minorities in United Kingdom and implications for primary prevention of heart disease in general practice: cross sectional population based study. BMJ Open. 2002;325(7375):1271. 24. Van Bortel L, Laurent S, Boutouyrie P, et al. Expert consensus document on the measurement of aortic stiffness in daily practice using carotid-femoral pulse wave velocity. Journal of Hypertension. 2012;30(3):445-448. 25. Petrie A, Sabin C. Medical statistics at a glance. 3rd Edition ed. Oxford: John Wiley & Sons; 2009. 26. Field A. Chapter 8: Logistic regression. Discovering statistics using SPSS. 3rd ed. London: Sage Publications Ltd; 2009. 27. Stern D, Smith L, Zhang B, Gordon-Larsen P, Popkin B. Changes in waist circumference relative to body mass index in Chinese adults, 1993-2009. International Journal of Obesity. 2014;38(12):1503-1510. 28. Klingberg S, Mehlig K, Lanfer A, Björkelund C, Heitmann B, Lissner L. Increase in waist cir- cumference over 6 years predicts subsequent cardiovascular disease and total mortality in nordic women. Obesity. 2015;23(10):2123-2130. 29. Van Gaal L, Mertens I, De Block C. Mechanisms linking obesity with cardiovascular disease. NAture. 2006;444(7121):875-880. 30. Borné Y, Hedblad B, Essén B, Engström G. Anthropometric measures in relation to risk of heart failure hospitalization: a Swedish population-based cohort study. European Journal of Public Health. 2014;24(2):215-220. 31. Strasser B, Arvandi M, Pasha E, Haley A, Stanforth P, Tanaka H. Abdominal obesity is associ- ated with arterial stiffness in middle-aged adults.Nutrition, Metabolism, and Cardiovascular Diseases. 2015;25(5):495-502. 32. Myint P, Kwok C, Luben R, Wareham N, Khaw K. Body fat percentage, body mass index and waist-to-hip ratio as predictors of mortality and cardiovascular disease. Heart. 2014;100(20):1613-1619. 33. Mahabadi A, Massaro J, Rosito G, et al. Association of pericardial fat, intrathoracic fat, and visceral abdominal fat with cardiovascular disease burden: the Framingham Heart Study. European Heart Journal. 2009;30(7):850-856. 34. Ferreira I, Snijder M, Twisk J, et al. Central fat mass versus peripheral fat and lean mass: opposite (adverse versus favorable) associations with arterial stiffness? The Amsterdam Growth and Health Longitudinal Study. Journal of Clinical Endocrinology and Metabolism. 2004;89(6):2632-2639. Body composition and cardiovascular risk 87

35. Cavalcante J, Lima J, Redheuil A, Al-Mallah M. Aortic stiffness: current understanding and future directions. J Am Coll Cardiol. 2011;57(14):1511-1522. 36. Leahy S, O’Neill C, Sohun R, Jakeman P. A comparison of dual energy X-ray absorptiometry and bioelectrical impedance analysis to measure total and segmental body composition in healthy young adults. European Journal of Applied Physiology. 2012;112(2):589-595.

5

Chapter 6 Physical activity and obesity: is there a difference in the association between the Asian- and African-Surinamese adult population?

Se-Sergio M. Baldew, Frederieke S. Diemer, Veronique Cornelissen, Glenn P. Oehlers, Lizzy M. Brewster, Jerry R. Toelsie, Luc Vanhees

Ethnicity and Health (2017); Jul 1:1-13 90 Chapter 6

Abstract

Background: The role of different physical activity (PA) characteristics, i.e. domain, duration and intensity in obesity prevention still requires investigation. Furthermore, ethnicity can modify the effect of PA on body composition. Therefore, we aim to describe the associa- tion between obesity and PA characteristics across the Asian-and African-Surinamese population, living in the capital of Suriname.

Methods: Between February 2013 and July 2015, we included 1157 healthy subjects, 18–70 years, from the Healthy Life in Suriname (HELISUR) study. We measured height, weight, hip and waist circumference and defined general and central obesity according to World Health Organization (WHO) recommendations. The International Physical Activity Ques- tionnaire was used to assess PA and to calculate the duration (minutes/week) and the total volume (METs-minutes/week) of activity. Ethnicity was self-reported.

Results: Out of 1157 participants we included 1079 (42.6% Asian-Surinamese, 40.1% African- Surinamese and 17.3% of other ethnicity), mean age 42.6 ± 13.6 years for analysis. Obesity prevalence ratio (PR) was significantly lower in participants meeting WHO PA recommendations [PR= 0.81 (0.68–0.97)], especially within the commuting [PR= 0.66 (0.47–0.91)] and leisure time domains [PR= 0.67 (0.47–0.94)], compared to participants that did not meet the recommendations. Active minutes/week and total volume of activity were inversely associated with obesity and waist circumference, in the overall (p < 0.05) and in the African-Surinamese population (p < 0.05), but not in the Asian- Surinamese population.

Conclusions: Meeting PA recommendations, particularly within the commuting and leisure time domains, is associated with lower obesity prevalence in the total population. Among the African- Surinamese population, PA within the leisure time domain, more active minutes/week and higher levels of total volume are associated with a lower obesity prevalence. This is not found in the Asian-Surinamese population. Physical activity and obesity 91

Introduction

Over one third of the world’s adult population is overweight and at least 13% is obese.1 Unfortunately, these numbers are increasing worldwide and at a faster rate in low- and middle-income countries, particularly in urban settings.2 Besides being the fifth leading cause of death3, obesity is an important cardiovascular risk factor and has also many adverse effects on other cardiovascular risk factors, including diabetes mellitus type 2, hypertension, and dyslipidemia.4,5 Therefore, there is an urgent need for effective treatment strategies.

Physical activity (PA) is a key element in the management of obesity6 and the American College of Sports Medicine recommends individuals to accumulate at least 150–250 min of moderate to vigorous PA per week with a targeted energy expenditure of 400 kcal per day.7 However, uncertainties remain with regard to the importance of each of the different PA characteristics8,9,10,11,12, i.e. domain, duration or intensity, that de- termines optimal weight control.13,14,15,16 Furthermore, these guidelines are based on data derived mainly from Caucasian populations, even though there is clear evidence that variations exist in body composition and cardiovascular risk factors between other ethnicities and Caucasians.17,18,19,20 Moreover, small studies suggest that the responses to PA interventions might be different across ethnic groups21,22, underlining the need for ethnic-specific PA recommendations for weight management. 6 Suriname is a middle-income country in the north-eastern part of South America, but culturally and economically has more similarities with the Caribbean. Due to historical migrations, the population of Suriname consists predominantly of people from Asian and African descent. This provides a unique opportunity to compare ethnic differences between these groups living within the same area. As such, the HEalthy LIfe in SURiname (HELISUR) study was designed to describe the prevalence of cardiovascular risk factors, intermediate endpoints, and cardiovascular diseases among the different ethnic groups living within the capital Paramaribo.23 The current report uses data from the HELISUR project to describe the association between obesity prevalence and PA characteristics (domain, duration and intensity) among the total population and across the Asian- and African-Surinamese population living in Paramaribo. 92 Chapter 6

Methods

Study population A detailed description of the HELISUR study design and participant recruitment has been published previously.23 In short, between February 2013 and July 2015 a total of 1800 randomly selected non-institutionalized individuals (18–70 years) living in the capital Paramaribo were contacted and interviewed at home by trained interviewers. From the recruited participants, 1157 subjects were further examined at the academic hospital in Paramaribo. Ethical clearance was obtained from the Ethics Committee of the Ministry of Health in Suriname (Approval nr. VG021-2012) and written informed consent was provided by all participants.

Ethnicity Participants were mainly Asian-Surinamese (self-identified South Asian and Indonesian descent); African-Surinamese (self-identified Creoles, that lived in Paramaribo for sev- eral generations and Maroons who migrated from the rural interior area to Paramaribo in the past 10–20 years) or other ethnicity (Chinese, Amerindians and Caucasians). In the current report we focused on the Asian and the African-Surinamese group, compris- ing the four main ethnic groups living in Paramaribo.

Assessment of PA The Dutch version of the International Physical Activity Questionnaire-long form (IPAQ- LF) was used to assess PA. This version has been validated previously24 and was pre- tested for reliability and face-validity in a small study sample. Participants were asked about their activities within the 1) working, 2) commuting, 3) domestic and garden, and 4) leisure time domain during the last seven days. Activities that were performed as part of paid or unpaid work were classified as activities within the working domain, whereas cycling and/or walking from one place to another were defined as activities within the commuting domain. Chores within and around the house were classified as activities within the domestic and garden domain whereas walking activities and/or activities that were not included within the former domains were classified within the leisure time domain. Within every domain the subjects were asked about the number of days and the amount of time they spend doing these activities and at what intensity. Activities could be performed at moderate or vigorous intensity or at a walking level. Moderate intensity was defined as experiencing a slight increase in heart rate and breathing rate, whereas vigorous intensity was defined as experiencing a significant increase in heart rate and breathing rate. Further-more, an activity within each domain and intensity had to be performed for at least 10 min continuously to be considered valid. Physical activity and obesity 93

The ‘Guidelines for data processing and analysis of the IPAQ: short and long form’25 were used to clean and analyze the data. These guidelines provide information on how to calculate the duration of activity per week (minutes/ week) for the domains separately and combined. This is done by multiplying the number of days by the amount of time spend per day being active for each domain and combined. Also, the PA volume was calculated for each participant by multiplying the duration by the assigned metabolic equivalent (MET) value for a specific activity within each domain (METs-minutes/week). The guidelines provide these MET values for the intensity of each activity, ranging from 3.0 to 8.0 METs. The total volume of PA is then calculated by combining the calculation of all the domains. Subsequently, participants who met at least one of the following recommendations were categorized as meeting the recommended level of PA that cor- responds to the WHO guidelines for PA: 1. Physical activity on at least 3 days of the week at a vigorous intensity for at least 20 min/ day 2. Physical activity on at least 5 days of the week at moderate intensity or walking for at least 30 min/day 3. Physical activity on at least 5 days of the week of any combination of walking, mod- erate or vigorous intensity activities with a minimum of 600 MET-minutes/week 4. Physical activity on at least 3 days of the week at a vigorous intensity achieving a minimum of at least 1500 MET-minutes/week 5. Physical activity of any combination of walking, moderate or vigorous intensity activity with a minimum of 3000 MET-minutes/week 6

Assessment of body composition During the hospital visit, data on body composition were assessed by a single investiga- tor.23 Waist circumference (WC) was measured at the midpoint between the lower mar- gin of the least palpable rib and the top of the iliac crest, whereas the hip circumference was measured at the broadest circumference below the waist, using a stretch resistant measuring tape. In order to calculate the body mass index (BMI: kg/m²), height and weight were measured using a stadiometer (SECA) and a scale (SECA 840), respectively. All four measurements were done in duplicate and a third measurement was made if the difference between the first two readings was >1.0 cm for the circumferences, >0.5 cm for height and >0.5 kg for weight. Mean values were used in the analysis. For the definition of overweight and obesity, we used the ethnic-specific BMI cut-off values that correspond to a high risk for undesirable health proposed by the WHO. For the African- Surinamese and other ethnicities overweight was defined as BMI ≥ 25 kg/ m² and obesity as BMI ≥ 30 kg/m², whereas for the Asian-Surinamese overweight was defined as BMI ≥ 23 kg/m² and obesity as BMI ≥ 27.5 kg/m².26,27 Central obesity was defined using the ethnic-specific criteria of the International Diabetes Federation.26,28 For the 94 Chapter 6

Asian-Surinamese participants, the cut-off values of ≥ 90 cm for men and ≥ 80 cm for women were used and for the African-Surinamese and other ethnicity participants the cut-off values of ≥ 94 cm for men and ≥ 80 cm for women were used.26

Statistical analysis Statistical analysis was performed using SPSS (version 21, SPSS Inc., Chicago, IL). Sen- sitivity analysis was done to compare the participants that were interviewed with the participants that were interviewed and completed the physical examination. Further- more, analyses were made for the total population and for each of the two main ethnic groups: Asian-Surinamese and African-Surinamese. Continuous variables were tested for normality by visual inspection of the histogram. Descriptive data were reported as means ± standard deviations, medians (range), or percentages; differences between both ethnic groups were tested with an ANOVA or chi-square test. For the total PA and for the separate characteristics (i.e. domain, duration and intensity), the percentage of participants meeting the recommended PA level were calculated for the total popula- tion and for each ethnic group.

In order to compare the obesity prevalence in participants meeting the recommended PA level and those not meeting the recommended level, we estimated prevalence ratios (PR) with corresponding 95% confidence intervals (95%CI) using Poisson regression models with robust variance29 and adjusted for sex and age.

The continuous variables duration (minutes/week), mean intensity (METs) and volume (METs-minutes/week) were categorized into tertiles. For the total population and for each ethnic group, the obesity PR (95% CI) within each tertile was calculated using the lowest tertile as the reference group. For BMI and WC, the mean values within each tertile of the duration, intensity and total volume were compared. This was done for the total population and for each ethnic group using an ANOVA test with the significance level set at a two-tailed p-value <0.05.

Results

From the 1157 participants that underwent physical examination, 78 were excluded for the final analysis, either because of missing (n = 19) or invalid PA data (n = 7), a history of coronary heart disease (n = 22) or stroke (n = 30), resulting in a sample of 1079 participants [42.6% Asian-Surinamese (n = 460), 40.1% African-Surinamese (n = 432) and 17.3% other ethnicity (n = 187)]. The other ethnicity group included mainly Physical activity and obesity 95 participants of mixed ethnicity (n = 164), but also a small number of Amerindians (n = 13), Chinese-Surinamese (n = 2), and people of other descent (n = 8) (Figure 1).

1,200 enumeration areas

Random selection of 18 enumeration areas

Preliminary assessment n=1,796 persons

Did not meet inclusion criteria (n=13) Died (n=9) Not retrieved (n=291)

Eligible n=1,483 persons

Declined to participate (n=326)

Physical examination n=1,157 participants

Missing data (n=19) Invalid PA data (n=7) History of coronary heart disease and/or stroke (n=52) Study population n=1,079 participants 6

Asian-Surinamese African-Surinamese Other ethnicity n=460 participants n=432 participants n=187 participants

Figure 1. Sampling scheme of the population. Participants from ‘other ethnicity’ were included in the total population analysis.

Table 1 shows the characteristics of the total study population and the two ethnic groups. The majority of the population was female (63.9%) with an equal distribution between both ethnic groups (p = 0.52). BMI was significantly higher in individuals of African-Surinamese origin (p < 0.01). However, based on the current definitions for overweight and central obesity, both were significantly more prevalent among the Asian-Surinamese population (both p < 0.01) whereas there was no difference with regard to the prevalence of obesity across both ethnic groups (p = 0.81). Overall, 78.5% of the participants met the PA recommendations, with the Asian-Surinamese being slightly more active compared to African Surinamese (82.2% in Asian-Surinamese vs. 76.6% in African-Surinamese, p < 0.05). 96 Chapter 6

Table 1. Characteristics of the total population and the two main ethnic groups. Total Asian- African- population Surinamese Surinamese (n=1079) (n=460) (n=432) p-value Sex (% women) 63.9 63.8 65.8 0.52 Mean age (years) 42.6 ± 13.6 43.7 ± 12.9 41.7 ± 13.9 0.02 Mean body mass index (kg/m²) 27.8 ± 6.3 27.1 ± 5.5 28.5 ± 6.8 <0.01 Overweight (%) 34.4 40.2 28.0 <0.01 Obesity (%) 36.7 39.1 38.4 0.81 Central obesity (%) 72.6 78.5 66.5 <0.01 Current smokers (%) 19.3 19.8 17.6 0.13 Mean hip circumference (cm) 102.9 ± 11.6 100.4 ± 10.1 105.1 ± 12.4 0.04 Mean waist circumference (cm) 94.9 ± 15.7 94.7 ± 14.2 95.1 ± 16.8 0.70 Mean waist-hip ratio 0.92 ± 0.10 0.94 ± 0.10 0.90 ± 0.09 0.04 Meeting the recommended PA levela (%) 78.5 82.2 76.6 0.05 Median weekly PA METs-minutes 3153(0-8774) 3808 (0-42684) 2583(0-8774) 0.24 Median PA METs 3.5(0-8) 3.5 (0-8) 3.5 (0-8) 0.70 Median weekly PA minutes 860 (0-8940) 1050 (0-8940) 763 (0-8550) 0.02 Footnote Table 1. Values are percentages (%), means ± standard deviation or medians (range). For the total population, participants from other ethnicities (n = 187) were also included. a The recommended level of physical activity according to the World Health Organization is as fol- lows: physical activity on at least 3 days of the week at a vigorous intensity for 20 min or achieving at least 1500 MET-minutes/week or physical activity at a moderate level or walking activity or an combination on at least 5 days of the week for at least 30 min or achieving a minimum of 600 MET-minutes/week or on at least 3 days of the week at a vigorous intensity achieving a minimum of at least 1500 MET-minutes/week or any combination of walking, moderate or vigorous intensity activity with a minimum of 3000 MET-minutes/week; values are the percentage of participants meeting the recommended level. p-value for comparison between ethnic groups. PA: Physical activity; METs: Metabolic equivalent values.

For the total sample, the age and sex-adjusted prevalence of obesity was 19% lower in participants who met the recommendations for PA (PR = 0.81 [95% CI 0.68–0.91]), as depicted in Table 2. Focusing on each of the domains of PA, the prevalence of obesity was 33–34% lower in participants who met the recommended level of PA within the com-muting or leisure time domain [for the commuting domain: PR = 0.66 (95% CI 0.47–0.91); for the leisure time domain: PR = 0.67 (95% CI 0.47–0.94)]. For the work and domestic and garden domain we did not find a significant association. Regarding the ethnic sub-group analysis, obesity was significantly less prevalent among African- Surinamese individuals meeting the PA recommendations within the leisure time domain [PR = 0.43 (0.20– 0.91)] compared to those that did not. This association could not be observed in the Asian-Surinamese population [PR = 0.78 (0.49–1.24)]. For the three remaining domains, we did not find a significant association in both the Asian- and the African-Surinamese population. Physical activity and obesity 97

Table 2. The association between obesity and physical activity for the total population and for each of the two main ethnic groups. Total Asian-Surinamese African-Surinamese population population population (n=1079) (n=460) (n=432) PR (95%CI) PR (95%CI) PR (95%CI) Total PA 0.81 (0.68–0.97) 0.75 (0.57–1.00) 0.89 (0.70–1.16) Per domain Work 0.90 (0.76–1.08) 0.88 (0.67–1.16) 0.89 (0.68–1.15) Commuting 0.66 (0.47–0.91) 0.58 (0.32–1.06) 0.70 (0.45–1.08) Domestic and garden 0.88 (0.75–1.03) 1.07 (0.82–1.40) 0.81 (0.64–1.02) Leisure time 0.67 (0.47–0.94) 0.78 (0.49–1.24) 0.43 (0.20–0.91) Footnote Table 2. For the total population, participants from other ethnicities (n = 187) were also included. The reference group was the participants who did not meet the recommended level of PA for the total volume and within each separate domain. Values are prevalence ratios and corre- sponding 95% confidence intervals. For total PA, adjustments were made for sex and age. For each domain, adjustments were made for sex, age and for PA in other domains. PA: physical activity; PR: prevalence ratio; CI: confidence interval.

As shown in Table 3, obesity was less prevalent in the overall sample among the par- ticipants being active for at least 1141 min/week [PR = 0.79 (0.66–0.94)] or performing more than 5281 METs-minutes/week [PR = 0.79 (0.66–0.95)]. Similar results could be observed for the subgroup of individuals from African-Surinamese origin, but not in the Asian-Surinamese individuals. No association between PA intensity and obesity prevalence could be established for the total population and for both ethnic groups. 6

Finally, as shown in Table 4, the mean BMI and WC decreased with increasing duration, intensity and volume for the overall population. For the African-Surinamese popula- tion both the BMI and WC decreased with increasing duration, intensity and volume, whereas for the Asian-Surinamese population only the intensity within the highest tertile significantly lowered the BMI, but not the WC, compared to the other tertiles.

As shown in Figure 1, from the 1800 recruited and interviewed participants, 1157 completed the physical examination. Sensitivity analysis showed that there were no differences between participants who were only interviewed and participants who were interviewed and completed the physical examination in sex, prevalence of smoking and meeting the WHO recommendations for PA (78.5% vs 81.4%, p > 0.05). However, participants who were also physically examined were significantly older (42.6 ± 13.6 vs 35.5 ± 14.0, p < 0.05) more often African-Surinamese (51.4% vs 40.1%, p < 0.05) and less often Asian-Surinamese (31.4% vs 42.6%, p < 0.05). 98 Chapter 6

Table 3. The prevalence ratio’s for obesity per tertile of each physical activity characteristic (dura- tion, intensity and volume) for the total population and per ethnic group. Tertile 1 Tertile 2 Tertile 3 p-value Duration (min/week) 0–495 496–1140 >1141 Total population 1 0.87 (0.72–1.05) 0.79 (0.66–0.94) 0.03 Asian-Surinamese population 1 1.11 (0.81–1.52) 1.05 (0.79–1.40) 0.82 African-Surinamese population 1 0.75 (0.57–1.00) 0.64 (0.49–0.85) 0.03 Intensity (METs) 0–3.24 3.25–3.84 >3.85 Total population 1 1.07 (0.91–1.27) 0.99 (0.80–1.23) 0.62 Asian-Surinamese population 1 1.20 (0.94–1.55) 0.86 (0.61–1.23) 0.90 African-Surinamese population 1 0.96 (0.75–1.23) 1.02 (0.76–1.37) 0.96 Volume (METs-min/week) 0–1778 1779–5480 >5281 Total population 1 0.88 (0.74–1.05) 0.79 (0.66–0.95) 0.05 Asian-Surinamese population 1 0.97 (0.73–1.30) 1.03 (0.78–1.37) 0.92 African-Surinamese population 1 0.85 (0.66–1.08) 0.60 (0.44–0.82) 0.04 Footnote Table 3.The total population (n = 1079) includes participants of the Asian-Surinamese population (n = 460), the African-Surinamese population (n = 432) and from other ethnicities (n = 187). Values are prevalence ratio’s with their corresponding 95% confidence interval between brackets. The range of the tertiles is given in italics for min/week, METs and METs-min/week sepa- rately. For duration and intensity, adjustments were made for sex, age, and for each other PA char- acteristic respectively. For volume, adjustments were made for sex and age. PA: physical activity; METs: metabolic equivalent values; min, minutes; METs-min/week, the volume calculated by multiplying the duration by the assigned METs for that activity.

Discussion

The results of this substudy of the HELISUR project show that 1) overweight and central obesity are more prevalent among the Asian-Surinamese compared to the African- Surinamese population living in urban Paramaribo; 2) obesity prevalence is similar in both ethnic groups; 3) meeting the PA recommendations and being active within the commuting and leisure time domain are all inversely associated with obesity prevalence in the total population; 4) prevalence ratios for obesity are lower in individuals spend- ing more time to PA and performing larger volumes both in the total population as well as within the African-Surinamese population; 5) no association could be established between PA and obesity prevalence in individuals of Asian origin; 6) body mass index and waist circumference were significantly lower with increasing duration, intensities and volume in the total population and in the African-Surinamese population; 7) only BMI was significantly lower in the Asian–Surinamese population in the highest tertile of intensity. Physical activity and obesity 99

Table 4. The mean body mass index and waist circumference for the total population and each ethnic group per tertiles of each physical activity characteristic (duration, intensity and volume).

Tertile 1 Tertile 2 Tertile 3 p-value Duration (min/week) 0–495 496–1140 >1141 Total population BMI 28.5 ± 6.9 27.8 ± 6.2 27.1 ± 5.71 0.03 WC 96.7 ± 17.1 94.3 ± 15.6 93.9 ± 14.41 0.03 Asian-Surinamese population BMI 26.9 ± 5.8 27.0 ± 5.7 27.3 ± 5.2 0.79 WC 94.8 ± 15.5 94.1 ± 15.4 94.9 ± 12.8 0.89 African-Surinamese population BMI 29.5 ± 7.3 28.2 ± 6.4 27.6 ± 6.41 0.03 WC 97.3 ± 17.8 94.9 ± 16.0 92.8 ± 15.81 0.05 Intensity (METs) 0–3.24 3.25–3.84 >3.85 Total population BMI 28.7 ± 6.8 27.9 ± 6.2 26.8 ± 5.72 0.05 WC 96.6 ± 16.2 94.8 ± 16.2 93.3 ± 14.51 0.02 Asian-Surinamese population BMI 27.4 ± 5.5 27.8 ± 5.9 26.2 ± 4.93 0.03 WC 95.1 ± 14.1 95.6 ± 15.6 93.3 ± 12.7 0.33 African-Surinamese population BMI 29.9 ± 7.2 28.5 ± 6.7 27.1 ± 6.11 0.01 WC 98.1 ± 17.4 94.8 ± 16.7 92.4 ± 15.81 0.01 Volume (METs-min/week) 0–1778 1779–5480 >5281 Total population BMI 28.5 ± 6.9 27.8 ± 6.2 27.1 ± 5.71 0.01 WC 96.5 ± 16.7 94.6 ± 15.8 93.6 ± 14.21 0.04 Asian-Surinamese population BMI 26.9 ± 5.5 27.0 ± 5.8 27.1 ± 5.5 0.66 WC 95.0 ± 14.9 93.8 ± 15.1 95.3 ± 12.9 0.59 African-Surinamese population BMI 29.4 ± 7.2 29.0 ± 6.7 27.9 ± 6.12 <0.01 6 WC 96.9 ± 17.4 96.6 ± 16.7 91.2 ± 15.32 <0.01 Footnote Table 4. The total population (n = 1079) includes participants of the Asian-Surinamese population (n = 460), the African-Surinamese population (n = 432), and from other ethnicities (n = 187). Values are the mean waist circumference in centimeters ± standard deviation. The range of the tertiles is given in italics for min/week, METs and METs-min/week. 1 Significantly different from tertile 1 at a p < 0.05. 2 Significantly different from tertile 1 and 2 at a p < 0.05. 3 Significantly different from tertile 2 at a p < 0.05. PA: physical activity; METs: metabolic equivalent values; min: minutes; METs-min, the volume cal- culated by multiplying duration by the assigned METs for that activity.

The results of our study are in line with previous studies showing that meeting the weekly recommended amount of PA, as formulated by the WHO, is inversely associated with obesity.30,31 This was especially true for being active within the commuting and lei- sure time domain.8,10 However, we could not establish such association for work-related and domestic and garden activities. There are likely various reasons for these different associations across the PA domains and obesity. First, we used self-reported PA data, which is sensitive to social desirability bias, recall bias, and overestimation.32,33,34 Foong et al.35 also reported that the adiposity and PA association was stronger for accelerom- 100 Chapter 6

eter-determined PA than questionnaire-determined PA. Therefore, an overestimation could have masked an existing association among the working and domestic and garden domain. For these two domains in particular, subjects find it very difficult to define the exact amount of time and the intensity of the activities.33,34 Commuting and leisure time activities are more planned activities allowing the participants to estimate the time and intensity spend in these domains more accurately.32

We also found a lower prevalence ratio of obesity among individuals of African origin meeting the recommended levels of PA in their leisure time, but not in individuals of Asian origin. Also, more time spend to PA and higher volumes were associated with lower prevalence ratio’s in the African-Surinamese but not in the Asian-Surinamese popula- tion. These results are in line with Kwon, Wang, and Hawkins (2013)36 who reported an inverse relation between leisure time PA and obesity among the African but not among the Asian population. The reason for this finding remains unclear and warrants further research. The use of categorical data leading to statistical power loss might be one of the reasons. However, we hypothesize that the Asian-Surinamese population is more sensitive for social desirable reporting37, leading to over-reporting of PA compared to individuals of African origin. To this end, the use of more objective instruments to quantify PA is necessary.

Among the African-Surinamese population the BMI and WC were consistently lower in the highest tertiles for duration, intensity and volume. Among the Asian-Surinamese only the BMI was significantly lower in the highest intensity tertile compared to the lowest tertile. Even though Lesser et al. (2012)38 did report that visceral adipose tissue was inversely associated to vigorous intensity PA among Asians, we did not find a lower WC for the highest intensity tertile compared to the lowest tertile. This significant lower BMI without the consistent lower WC among the Asian-Surinamese population points out that the adipose tissue might be less in other body areas except the abdominal area or they even have lower muscle mass. In order to accurately assess adiposity, its regional distribution and also lean mass, dual x-ray absorptiometry (DXA) would be a valid measuring tool.39 Unfortunately, we were limited to the use of BMI calculations and WC measurements to define obesity and central obesity using the WHO recom- mended ethnic specific cut-off values.27 However, these cut-off values have not been validated within our population.

The differences in association between body composition and PA between the Asian and African-Surinamese population could be caused by a difference in fat oxidation between ethnic groups. Compared to Caucasians, Asians have a lower level of fat oxida- tion at the same level of activity, whereas Africans show the same fat oxidation.18,40 Physical activity and obesity 101

Furthermore, compared to Caucasians, Asians need to engage in higher levels of PA for the same cardiovascular and metabolic profile.21,41 Therefore, Asians might benefit from exercise at higher intensity to affect visceral adipose tissue.

Unfortunately, insufficient dietary information precluded us from taking nutrition into account. Nutrition is at least as important in weight management.42,43,44 Previous studies have reported differences in dietary patterns between Asians, Africans, and Caucasians. Compared to the Caucasians, Asians had a higher energy intake, whereas Africans had similar energy intakes.45 South Asians also have a higher intake of staple food includ- ing rice, fried rice and noodles compared to Caucasians.46 The possible variety in diet between the Asian- and African-Surinamese population is definitively something which needs to be taken into consideration in future research.

Furthermore, because of the cross-sectional design, these findings are merely demon- strating associations and not necessarily prove causality. In this regard, it could be that obese participants tend to be less physically active within their leisure time and are more likely to use the car/bus for commuting to and from work, explaining the higher prevalence of obesity in individuals who were not physically active in the leisure time or commuting domain.

Finally, the combination of the four major ethnic groups, the South Asian, the Javanese, the Creole and Maroon population within respectively, the Asian- and African-Surinam- 6 ese population could also be a limitation. Due to the small number of participants within each major ethnic group we decided to make these two combinations based on the region of origin. However, we acknowledge that these groups have a different cultural and genetic background that could influence the differences in association between PA characteristics and body composition. For the definition of overweight, obesity and central obesity we used BMI and WC cut-off values that need further validation in our population. Therefore, future studies focusing on the ethnic differences should include sufficient number of participants of each group separately.

To conclude, our results show that meeting PA recommendations, especially within the commuting and leisure time domain, are associated with reduced levels of obesity in the total population. Furthermore, in the African-Surinamese population spending more time physically active and PA within the leisure time are associated with less obe- sity, whereas such an association could not be observed among the Asian-Surinamese population. Our results therefore underline the targets of the WHO to increase PA in prevention of obesity47 and emphasize the need for more in depth research into the role 102 Chapter 6

of PA characteristics using more objective tools in combating the obesity epidemic in general and in different ethnic groups. Physical activity and obesity 103

References

1. World Health Organization. Global health risks: mortality and burden of disease attributable to selected major risks. Geneva 2009. 2. Popkin B, Slining M. New dynamics in global obesity facing low- and middle-income coun- tries. Obesity reviews. 2013;14 (Suppl 2):11-20. 3. NCD Risk Factor Collaboration (NCD-RisC). Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19.2 million participants. Lancet. 2016;387(10026):1377-1396. 4. Bray G, Bellanger T. Epidemiology, trends, and morbidities of obesity and the metabolic syndrome. Endocrine. 2006;29(1):109-117. 5. Sharma A. Obesity and cardiovascular risk. Growth Hormone & IGF Research. 2003;13 (Suppl A):S10-17. 6. Shook R, Hand G, Drenowatz C, et al. Low levels of physical activity are associated with dysregulation of energy intake and fat mass gain over 1 year. American Journal of Clinical Nutrition. 2015;102(6):1332-1338. 7. Donnelly J, Blair S, Jakicic J, et al. American College of Sports Medicine Position Stand. Ap- propriate physical activity intervention strategies for weight loss and prevention of weight regain for adults. Medicine and Science in Sports and Excercise. 2009;41(2):459-471. 8. Abu-Omar K, Rütten A. Relation of leisure time, occupational, domestic, and commuting physical activity to health indicators in Europe. Preventive Medicine. 2008;47(3):319-323. 9. Du H, Li L, Whitlock G, et al. Patterns and socio-demographic correlates of domain-specific physical activities and their associations with adiposity in the China Kadoorie Biobank study. BMC Public Health. 2014;14:826. 10. Flint E, Cummins S, Sacker A. Associations between active commuting, body fat, and body mass index: population based, cross sectional study in the United Kingdom. BMJ Clinical 6 Research. 2014;349:g4887. 11. Wanner M, Götschi T, Martin-Diener E, Kahlmeier S, Martin B. Active transport, physical activ- ity, and body weight in adults: a systematic review. American Journal of Preventive Medicine. 2012;42(5):493-502. 12. Wanner M, Tarnutzer S, Martin B, et al. Impact of different domains of physical activity on cause-specific mortality: a longitudinal study. Preventive Medicine. 2014;62:89-95. 13. Dickie K, Micklesfield L, Chantler S, Lambert E, Goedecke J. Cardiorespiratory fitness and light-intensity physical activity are independently associated with reduced cardiovascular disease risk in urban black South African women: a cross-sectional study. Metabolic Syn- drome and Related Disorders. 2016;14(1):23-32. 14. Haskell W, Lee I, Pate R, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Associa- tion. Circulation. 2007;116(9):1081-1093. 15. Vanhees L, De Sutter J, Geladas N, et al. Importance of characteristics and modalities of physical activity and exercise in defining the benefits to cardiovascular health within the general population: recommendations from the EACPR (Part I). European Journal of Preven- tive Cardiology. 2012;19(5):1005-1033. 16. Vanhees L, Geladas N, Hansen D, et al. Importance of characteristics and modalities of physi- cal activity and exercise in the management of cardiovascular health in individuals with 104 Chapter 6

cardiovascular risk factors: recommendations from the EACPR. Part II. European Journal of Preventive Cardiology. 2012;19(5):1005-1033. 17. Carson A, Howard G, Burke G, Shea S, Levitan E, Muntner P. Ethnic differences in hypertension incidence among middle-aged and older adults: the multi-ethnic study of atherosclerosis. Hypertension. 2011;57(6):1101-1107. 18. Dulloo A, Jacquet J, Solinas G, Montani J, Schutz Y. Body composition phenotypes in path- ways to obesity and the metabolic syndrome. International Journal of Obesity. 2010;34 (Suppl 2):S4-S17. 19. Gasevic D, Ross E, Lear S. Ethnic Differences in cardiovascular disease risk factors: a system- atic review of North American evidence. Canadian Journal of Cardiology. 2015;31(9):1169- 1179. 20. Tillin T, Hughes A, Mayet J, et al. The relationship between metabolic risk factors and incident cardiovascular disease in Europeans, South Asians, and African Caribbeans: SABRE (Southall and Brent Revisited) -- a prospective population-based study. Journal of the American Col- lege of Cardiology. 2013;61(17):1777-1786. 21. Celis-Morales C, Ghouri N, Bailey M, Sattar N, Gill J. Should physical activity recommen- dations be ethnicity-specific? Evidence from a cross-sectional study of South Asian and European men. PlOS One. 2013;8(12):e82568. 22. White J, Jago R. Prospective associations between physical activity and obesity among adolescent girls: racial differences and implications for prevention. Archives of Pediatrics & Adolescent Medicine. 2012;166(6):522-527. 23. Diemer F, Aartman J, Karamat F, et al. Exploring cardiovascular health: the Healthy Life in Su- riname (HELISUR) study. A protocol of a cross-sectional study. BMJ Open. 2014;4:e006380. 24. Blikman T, Stevens M, Bulstra S, van den Akker-Scheek I, Reininga I. Reliability and validity of the Dutch version of the International Physical Activity Questionnaire in patients after total hip arthroplasty or total knee arthroplasty. Journal of Orthopaedic and Sports Physical Therapy. 2013;43(9):650-659. 25. IPAQ. Guidelines for data processing and analysis of the International Physical Activity Questionnaire. 2005. Available at: http://www.ipaq.ki.se/scoring.pdf. 26. National Institute for Health and Care Excellence. Assessing body mass index and waist circumference thresholds for intervening to prevent ill health and premature death among adults from black, Asian and other minority ethnic groups in the UK. 2013. Available at: https://www.nice.org.uk/guidance/ph46. 27. World Health Organization. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363:157-163. 28. World Health Organization. Waist circumference and waist-hip ratio: a report of a WHO Expert Consultation. Geneva: WHO; 2008. 29. Coxe S, West S, Aiken L. The analysis of count data: a gentle introduction to poisson regres- sion and its alternatives. Journal of Personality Assessment. 2009;91(2):121-136. 30. Ekelund U, Besson H, Luan J, et al. Physical activity and gain in abdominal adiposity and body weight: prospective cohort study in 288,498 men and women. American Journal of Clinical Nutrition. 2011;93(4):826-835. 31. Myers A, Gibbons C, Finlayson G, Blundell J. Associations among sedentary and active be- haviours, body fat and appetite dysregulation: investigating the myth of physical inactivity and obesity. British Journal of Sports Medicine. 2017;51(21):1540-1544. Physical activity and obesity 105

32. Fillipas S, Cicuttini F, Holland A, Cherry C. The international physical activity questionnaire overestimates moderate and vigorous physical activity in HIV-infected individuals com- pared with accelerometry. Journal of the Association of Nurses in AIDS Care. 2010;21(2):173- 181. 33. Hallal PC GLPDLFMJFARRPMSO. Lessons learned after 10 years of IPAQ use in Brazil and Colombia. Journal of Physical Activity and Health. 2010;7(Suppl 2):S259-264. 34. Sebastião E, Gobbi S, Chodzko-Zajko W, et al. The International Physical Activity Question- naire-long form overestimates self-reported physical activity of Brazilian adults. Public Health. 2012;126(11):967-975. 35. Foong Y, Aitken D, Winzenberg T, Otahal P, Scott D, Jones G. The association between physi- cal activity and reduced body fat lessens with age - results from a cross-sectional study in community-dwelling older adults. Experimental Gerontology. 2014;55:107-112. 36. Kwon S, Wang M, Hawkins M. Association between self-reported physical activity and obe- sity among White, Black, Hispanic, and Asian Americans: 2007 and 2009 Brfss. Ethnicity and Disease. 2013;23(2):129-135. 37. Johnson T, Shavitt S, Holbrook A. Survey response styles across cultures. In: Matsumoto D, Van de Vijver F, eds. Cross-cultural research methods in psychology. New York: Cambridge Unversity Press; 2011. 38. Lesser I, Yew A, Mackey D, Lear S. A cross-sectional analysis of the association between physical activity and visceral adipose tissue accumulation in a multiethnic cohort. Journal of Obesity. 2012:703941. 39. Toombs R, Ducher G, Shepherd J, De Souza M. The impact of recent technological advances on the trueness and precision of DXA to assess body composition. Obesity. 2012;20(1):30- 39. 40. Hall L, Moran C, Milne G, et al. Fat oxidation, fitness and skeletal muscle expression of oxida- tive/lipid metabolism genes in South Asians: implications for insulin resistance? PloS One. 6 2010;5(12):e14197. 41. Iliodromiti S, Ghouri N, Celis-Morales C, Sattar N, Lumsden M, Gill J. Should physical activity recommendations for South Asian adults be ethnicity-specific? Evidence from a cross-sectional study of South Asian and White European men and women. PloS One. 2016;11(8):e0160024. 42. Curioni C, Lourenço P. Long-term weight loss after diet and exercise: a systematic review. International Journal of Obesity. 2005;29(10):1168-1174. 43. de Munter J, Tynelius P, Magnusson C, Rasmussen F. Longitudinal analysis of lifestyle habits in relation to body mass index, onset of overweight and obesity: results from a large popu- lation-based cohort in Sweden. Scandinavian Journal of Public Health. 2015;43(3):236-245. 44. Unick J, Jakicic J, Marcus B. Contribution of behavior intervention components to 24-month weight loss. Medicine and Science in Sports and Exercise. 2010;42(4):745-753. 45. AS, Nightingale C, Owen C, et al. Nutritional composition of the diets of South Asian, black African-Caribbean and white European children in the United Kingdom: the Child Heart and Health Study in England (CHASE). British Journal of Nutrition. 2010;104(2):276-285. 46. Raza Q, Snijder M, Seidell J, Peters R, Nicolaou M. Comparison of cardiovascular risk fac- tors and dietary intakes among Javanese Surinamese and South-Asian Surinamese in the Netherlands. The HELIUS study. BMC Research Notes. 2017;10(23). 106 Chapter 6

47. World Health Organization. A guide for population-based approaches to increasing levels of physical activity: implementatino of the WHO global strategy on diet, physical activity and health. Geneva: WHO; 2007.

Chapter 7 Aortic pulse wave velocity in individuals of Asian and African ancestry: the HELISUR study

Frederieke S. Diemer, Se-Sergio M. Baldew, Yentl C. Haan, Fares A. Karamat, Glenn P. Oehlers, Gert A. van Montfrans, Bert-Jan H. van den Born, Ron J. G. Peters, Lenny M. W. Nahar-van Venrooij, Lizzy M. Brewster

Submitted 110 Chapter 7

Abstract

Introduction: Aortic pulse wave velocity has emerged as an important predictor of cardiovascular events, but data on ethnic differences in pulse wave velocity remain scarce. We explored differences in pulse wave velocity between people of Asian and African ancestry.

Methods: Data were used from the cross-sectional Healthy Life in Suriname (HELISUR) study. Pulse wave velocity was estimated oscillometrically with the Arteriograph. Main outcome of the study was the ethnic difference in pulse wave velocity, adjusted for age and blood pressure.

Results: We included 353 Asians and 364 Africans, aged respectively 44.9 (SD 13.5) and 42.8 (SD 14.1) years (p=0.05). Crude median PWV was higher in Asians than in Africans (8.1 [IQR 6.9 to 10.1] m/s vs 7.7 [IQR 6.5 to 9.3] m/s, p=0.03). After adjustment for age and MAP in multivariable linear regression, Asians had a 1.044 [95% CI 1.019 to 1.072] m/s higher PWV compared to Africans. Additional adjustment for sex, glucose, total cholesterol, HDL cholesterol, triglycerides, BMI, and waist circumference did not sub- stantially change the difference in pulse wave velocity between Asians and Africans (+1.044 [95% CI 1.016 to 1.074] m/s for Asians versus Africans).

Conclusion: Individuals of Asian ancestry have a higher pulse wave velocity than those of African ancestry. This persisted after adjustment for important cardiovascular risk parameters, including age and blood pressure. The higher PWV found in Asians is consistent with their increased coronary heart disease risk. Ethnic differences in PWV 111

Introduction

Well-established ethnic differences exist in the prevalence and incidence of cardio- vascular disease (CVD). People of Asian ancestry are disproportionately affected by coronary heart disease, whereas Africans have an increased risk of hypertension-related organ damage, such as left ventricular hypertrophy and stroke.1,2,3 Evidence suggests that these differences between ethnic groups are not fully explained by traditional cardiovascular risk factors.4,5 One way to address the issue of ethnic variability in CVD risk is to shift the focus from risk assessment to diagnosing early organ damage, for example in blood vessels.

Increased arterial stiffness, measured non-invasively by aortic pulse wave velocity (PWV), is one of the earliest detectable manifestations of adverse structural and func- tional changes of the vessel wall.6 With increasing age, changes in the vascular media occur, including increased collagen concentration and calcification, thereby reducing arterial compliance.7 When the arterial compliance decreases, the forward pulse wave travels faster, and the arterial waves reflected from the periphery reach the heart earlier during systole leading to an increase in systolic blood pressure (BP), which increases cardiac afterload. Conversely, diastolic BP decreases, resulting in increased pulse pres- sure.7,8 Thus, an increased PWV reflects three potential risk factors: increased systolic BP, widened pulse pressure, and altered vascular wall properties. In line with this, an increased arterial stiffness can be used to identify individuals at increased cardiovascu- lar risk and has emerged as an important predictor of cardiovascular events.9,10,11

Data regarding arterial stiffness in different ethnic groups are scarce and heteroge- neous in terms of populations, assessment techniques and statistical methods used 7 for data analysis.12,13,14,15,16,17,18 Moreover, the majority of these studies are conducted in high-income countries, in which other ethnic groups are predominantly compared to the Caucasian population as reference group. A study by Snijder and colleagues compared PWV in persons of African, South-Asian, and European ancestry.17 Compared to persons of European ancestry, persons of African or South-Asian ancestry had higher age-adjusted PWV values. After additional adjustment for BP and cardiovascular risk factors, the higher PWV in individuals of African ancestry disappeared, but persisted in South-Asians above the age of 40 years.17 Nevertheless, undetected confounding, including socio-economic status, may influence the association between ethnicity and PWV. Suriname, a middle-income country, is predominantly inhabited by persons of African and Asian ancestry, with large similarities in socioeconomic status.19,20 Our aim was to examine differences in PWV between Asian- and African-Surinamese, taking important cardiovascular risk parameters into account. 112 Chapter 7

Methods

Study population The study population consisted of participants of the Healthy Life in Suriname (HELI- SUR) study, a population-based observational study, as previously described.21 The HELISUR study was conducted according to the principles of the Declaration of Helsinki (59th WMA General Assembly, Seoul, October 2008) and in accordance with the Medical Research Involving Human Subjects Act. Ethical clearance was obtained from the Ethics Committee of the Ministry of Health in Suriname in 2012 (Approval nr. VG021-2012).

Briefly, we randomly selected a representative multi-ethnic sample of 1 800 non- institutionalized men and women aged 18 to 70 years living in the capital Paramaribo. Eligible subjects were interviewed at home and subsequently invited for an examina- tion at the local hospital. A total of 1 157 subjects participated in both the interview and the physical examination.22 For the current analysis, participants were excluded if data on PWV were unavailable or if they were of self-identified non-Asian or non-African ancestry (n=440).

Outcome measures Main outcome of the study was the difference in PWV between persons of Asian and African ancestry, adjusted for age and BP. As a secondary outcome, we adjusted for other continuous cardiovascular risk parameters and sex. Moreover, the association between PWV and ethnicity was further explored through stratification by traditional cardiovas- cular risk factor status, age, and sex.

Sample size calculation To calculate the required sample size, we used the findings by Rezai et al. who reported a 0.5 (SD 2.0) m/s higher PWV in persons of South-Asian ancestry compared to those of African ancestry, after adjustment for age, BP, and diabetes.16 Using Lehr’s formula 16 ( 2 ), we calculated that at least 267 participants in each group were (standardized difference ) required to have an 80% chance of detecting an ethnic difference in PWV at the 5% significance level.23

Data collection

Questionnaire Information on demographic factors and use of medication were collected by means of a questionnaire. Ethnic differences in PWV 113

Physical examination Participants were examined in the morning after an overnight fast. Height, weight, and waist circumference (WC) were measured in duplicate to the nearest 0.1 cm or 0.1 kg, with participants wearing no shoes and light underwear. Body mass index (BMI) was computed as mean weight divided by mean height squared. Aortic PWV was estimated twice in the supine position on the right upper arm, using the validated Arteriograph device (TensioMed, Budapest, Hungary).24 In short, after two conventional BP measure- ments, the Arteriograph produces a cuff pressure over the brachial artery that is 35 mmHg in excess of the systolic blood pressure measured. This suprasystolic pressure is used to analyze the pulse wave. To calculate the PWV, the distance from the jugulum to the symphysis is measured in a straight line using a tape measure, as a surrogate measure of the aortic length, and multiplied by two. The result is then divided by the difference in time between the beginning of the first wave and the beginning of the second (reflected) wave, resulting in the PWV in meters/second (m/s).24,25 The two conventional BP measurements of the Arteriograph were used to calculate the mean arterial pressure (MAP). For the definition of hypertension, BP was measured twice after a 5-minute rest in the sitting position with an automated oscillometric device (WatchBP Office; Microlife AG, Widnau, Switzerland) and an appropriately adjusted cuff size on the left upper arm supported at heart level. Blood samples were drawn after an overnight fast and analyzed for plasma glucose and serum lipid profile (UniCel DxC 600/800 System, Beckman Coulter, Brea, California, USA).

Definitions Participants were mainly of Asian (self-identified South Asian and Indonesian descent); or of African (self-identified Creoles and Maroons) ancestry.20 Traditional cardiovascu- lar risk factors included: 1) hypertension: systolic BP ≥ 140 mmHg, or diastolic BP ≥ 7 90 mmHg, or receiving antihypertensive medication (ATC codes C02, C03, C07, C08, and C09), 2) diabetes mellitus: glucose ≥ 7.0 mmol/L or glucose-lowering medication (ATC code A10), 3) dyslipidemia: total cholesterol ≥ 6.20 mmol/L, LDL cholesterol ≥ 4.1 mmol/L, HDL cholesterol < 1.0 mmol/L, triglycerides ≥ 2.3 mmol/L, or receiving cholesterol-lowering medication (ATC code C10), 4) obesity: BMI > 27.5 kg/m2 in Asians and BMI > 30.0 kg/m2 in Africans, and 5) abdominal obesity: WC: Asian men: ≥ 90 cm; African men: ≥ 94 cm; women: ≥ 80 cm.26,27,28,29

Statistical analysis Population characteristics for the Asian and African ancestry participants were com- pared using the Mann-Whitney U test (non-parametric), Student t-test (parametric), and Chi-square test (categorical). Multivariable linear regression analysis was performed to assess ethnic differences in log PWV (dependent), after adjustment for age and MAP 114 Chapter 7

(independent). Additional adjustments were made for sex, glucose, total cholesterol, HDL cholesterol, triglycerides, BMI, and WC.

To further explore the association between PWV and ethnicity, we plotted PWV against age and stratified by ethnicity and cardiovascular risk factor status. We assessed ethnic differences between those with and without cardiovascular risk factors using Mann- Whitney U tests. Subsequently, we stratified the analyses by the age cut-off value proposed by Snijder et al. (i.e. < 40 vs. ≥ 40 years) and, additionally, selected an age cut-off value most suitable for our study population in a post-hoc analysis.17 Finally, we assessed differences in PWV between men and women, and additionally stratified by ethnic subgroup.

In a sensitivity analysis, we evaluated differences between African vs Asian ancestry persons with missing PWV measurements who were excluded. We verified whether the assumptions of linear regression model were met, including normality, linearity, homoscedasticity, and no collinearity. Statistical analysis was carried out using the SPSS version 20.0 software for Windows (SPSS Inc., Chicago, IL, USA). Graphical presentation was based on GraphPad Prism version 7.0 (GraphPad Software, Inc., San Diego, CA).

Results

PWV was estimated in 717 subjects, 45% male, with a mean age of 43.8 (SD 13.9) years. The data of PWV were positively skewed (skewness: 1.078) with positive kurtosis (0.819). Conversion to base 10 logarithms improved the symmetry of the overall dis- tribution (skewness: 0.538; kurtosis: -0.405). Furthermore, lipid parameters (i.e. total cholesterol, LDL, HDL, triglycerides) were highly collinear (Variance Inflation Factor (VIF) > 10). However, LDL was calculated based on the other lipid parameters. After omitting LDL from the model, the assumption of multicollinearity was met (VIF of remaining lipid parameters below 2).

PWV by ethnicity Crude median PWV was higher in Asians than in Africans (8.1 [IQR 6.9 to 10.1] m/s vs 7.7 [IQR 6.5 to 9.3] m/s, p=0.03) (Table 1). However, Asians were older than Africans (44.9 [SD 13.5] years vs 42.8 [SD 14.1] years, p=0.05) and had a lower MAP (94.1 [SD 14.6] mmHg vs 97.4 [SD 17.1] mmHg, p<0.01). After adjustment for age and MAP in multivariable linear regression, Asians had a 0.019 [95% CI 0.008 to 0.030] higher log PWV than Africans, which corresponds to a 1.044 [95% CI 1.019 to 1.072] m/s higher PWV (Table 2). Ethnic differences in PWV 115

Table 1. Characteristics of the study population by ethnicity Asian African (n=353) (n=364) p-value Men, n (%) 175 (49.6) 150 (41.2) 0.02 Age, mean (SD) 44.9 (13.5) 42.8 (14.1) 0.05 PWV, median (IQR) 8.1 (6.9-10.1) 7.7 (6.5-9.3) 0.03 Systolic BP (supine), mean (SD) 127.6 (20.0) 132.4 (22.7) <0.01 Diastolic BP (supine), mean (SD) 77.4 (12.5) 79.9 (14.8) 0.01 MAP (supine), mean (SD) 94.1 (14.6) 97.4 (17.1) <0.01 Systolic BP (sitting), mean (SD) 129.7 (19.8) 132.2 (22.1) 0.12 Diastolic BP (sitting), mean (SD) 81.5 (11.1) 82.3 (13.2) 0.38 Hypertension, n (%) 152 (43.1) 163 (44.8) 0.64 Hypertension treatment, n (%) 101 (28.6) 90 (24.7) 0.24 Control in hypertensives, n (%) 37 (10.5) 33 (9.1) 0.52 Control in treated, n (%) 37 (36.6) 33 (36.7) 1.00 BMI, mean (SD) 26.5 (5.3) 27.5 (6.1) 0.02 Obesity, n (%) 121 (34.3) 121 (33.2) 0.77 Waist circumference, mean (SD) 94.3 (14.2) 93.4 (15.8) 0.41 Abdominal obesity, n (%) 264 (74.8) 226 (62.1) <0.01 Glucose, mean (SD) 6.2 (2.6) 5.2 (1.5) <0.01 Glucose-lowering treatment, n (%) 68 (9.3) 20 (5.5) <0.01 Diabetes mellitus, n (%) 83 (23.5) 29 (8.2) <0.01 Total cholesterol, mean (SD) 5.0 (1.1) 4.5 (1.1) <0.01 HDL, mean (SD) 1.1 (0.3) 1.3 (0.3) <0.01 LDL, mean (SD) 3.6 (0.9) 3.0 (1.0) <0.01 Triglycerides, mean (SD) 1.6 (1.4) 1.0 (0.6) <0.01 Cholesterol-lowering treatment, n (%) 21 (5.9) 6 (1.6) <0.01 Dyslipidemia, n (%) 192 (54.7) 115 (31.6) <0.01 Footnote Table 1. SD, standard deviation; IQR, interquartile range; BMI, body mass index; PWV, pulse wave velocity; BP, blood pressure; MAP, mean arterial pressure; HDL/LDL, high/low-density lipoprotein. 7

Table 2. Ethnic differences in log pulse wave velocity, adjusted for cardiovascular risk parameters Log PWV Estimated differences p-value Asian vs African (reference) 0.017 (0.000 to 0.033) 0.048 Model 1: age, MAP 0.019 (0.008 to 0.030) 0.001 Model 2: age, MAP, sex 0.023 (0.012 to 0.034) 0.000 Model 3: age, MAP, glucose 0.018 (0.006 to 0.029) 0.003 Model 4: age, MAP, total cholesterol, HDL, triglycerides 0.018 (0.006 to 0.030) 0.004 Model 5: age, MAP, BMI 0.020 (0.009 to 0.031) 0.000 Model 6: age, MAP, WC 0.018 (0.007 to 0.030) 0.001 Model 7: age, MAP, sex, glucose, total cholesterol, HDL, 0.019 (0.007 to 0.031) 0.002 triglycerides, BMI, WC (fully adjusted) Footnote Table 2. Values are regression coefficients with 95% confidence interval (CI). MAP, mean arterial pressure; HDL, high-density lipoprotein; BMI, body mass index; WC, waist cir- cumference. 116 Chapter 7

Further adjustment for sex, glucose, total cholesterol, HDL cholesterol, triglycerides, BMI, and WC did not attenuate the ethnic difference in log PWV. The higher PWV in Asians compared to Africans persisted in the fully-adjusted model: 0.019 [95% CI 0.007 to 0.031], which resembles a 1.044 [95% CI 1.016 to 1.074] m/s higher PWV.

PWV by ethnicity and traditional cardiovascular risk factors PWV stratified by ethnicity and traditional cardiovascular risk factor status is depicted in Figure 1. We found no ethnic difference in crude median PWV in the presence of tra- ditional cardiovascular risk factors. However, in those without hypertension or without obesity, Asians had substantially higher PWV values (no hypertension: 7.2 [IQR 6.3 to 8.5] m/s in Asians vs 6.8 [IQR 6.3 to 7.9] m/s in Africans, p=0.03; no obesity: 7.6 [IQR 6.6 to 9.7] m/s in Asians vs 7.2 [IQR 6.4 to 8.5] m/s in Africans, p=0.02).

PWV by ethnicity and age We found no ethnic difference in crude median PWV between Asians and Africans aged 40 years and above (9.2 [IQR 7.6 to 10.8] m/s in Asians vs 8.9 [IQR 7.6 to 10.8] m/s in Africans, p=0.33). Further exploration of the data showed that a cut-off value of 50 years was more suitable in this study population (point of intersection in Figure 1A). Asians ≥ 50 years had a significantly higher PWV compared to Africans ≥ 50 years (10.1 [IQR 8.7 to 11.8] m/s in Asians vs 9.1 [IQR 7.9 to 11.3] m/s in Africans, p<0.01). Restricting the age- and MAP-adjusted analysis to persons 50 years and above resulted in a 0.053 [95% CI 0.032 to 0.074] higher log PWV in Asians compared to their African counterparts, which can be translated in a 1.13 [95% CI 1.08 to 1.19] m/s higher PWV. The ethnic difference in Asians and Africans aged 50 years and above remained after additional adjustment for sex, glucose, cholesterol, HDL cholesterol, triglycerides, BMI, and WC (1.14 [95% CI 1.08 to 1.21] m/s).

PWV by ethnicity and sex Women overall had a higher crude median PWV than men in both ethnic groups (Asian: 8.8 [IQR 6.9 to 11.0] m/s in women vs 7.5 [IQR 6.8 to 9.0] m/s in men; African: 8.2 [IQR 6.9 to 10.0] m/s in women vs 7.1 [IQR 6.3 to 8.5] m/s in men; both p<0.01). These sex differences in PWV persisted in the fully adjusted model (1.11 [IQR 1.07 to 1.15] m/s higher PWV in Asian women vs Asian men; 1.13 [IQR 1.08 to 1.17] m/s higher PWV in African women vs African men).

Additional stratification by ethnicity showed that Asian men had substantially higher crude median PWV values than African men (7.5 [IQR 6.8 to 9.0] m/s vs 7.1 [IQR 6.3 to 8.5], p=0.01). In women, no significant ethnic difference in crude PWV was seen (8.8 [IQR 6.9 to 11.0] in Asian women vs 8.2 [IQR 6.9 to 10.0] in African women, p=0.10). However, ETHNIC DIFFERENCES IN PWV 117

A B PWV by ethnicity PWV by ethnicity and hypertension 12 12 ) 10 ) 10 s s / / m m ( (

V 8 V 8 W W P P

6 6

20 30 40 50 60 70 20 30 40 50 60 70 Age (years) Age (years) Asian Asian - Hypertension African - Hypertension African Asian - No hypertension African - No hypertension

C D PWV by ethnicity and diabetes mellitus PWV by ethnicity and dyslipidemia 12 12

) 10 ) 10 s / s / m m (

(

V V 8

8 W W P P

6 6

20 30 40 50 60 70 20 30 40 50 60 70 Age (years) Age (years) Asian - Dyslipidemia Asian - Diabetes African - Diabetes African - Dyslipidemia Asian - No dyslipidemia African - No dyslipidemia Asian - No diabetes African - No diabetes

E PWV by ethnicity and obesity F PWV by ethnicity and abdominal obesity 12 12

) 10 ) 10 s s / / m m ( (

V V 8

8 W W P P 7

6 6

20 30 40 50 60 70 20 30 40 50 60 70 Age (years) Age (years) Asian - Abdominal obesity African - Abdominal obesity Asian - Obesity African - Obesity Asian - No abdominal obesity African - No abdominal obesity Asian - No obesity African - No obesity figure 1. Pulse wave velocity (PWV), stratifi ed by ethnicity and traditional cardiovascular risk fac- tor status. Data points depict the pulse wave velocity (PWV) with standard error of the mean (SEM) for that specifi c age. 118 Chapter 7

after adjustment for age and MAP, both Asian men and women had significantly higher PWV values than African men and women, respectively (Asian men: +1.07 [95%CI 1.03 to 1.11] m/s; African women: +1.04 [95%CI 1.00 to 1.08] m/s). The higher PWV in Asian men and women persisted in the fully-adjusted model (Asian men: +1.05 [95%CI 1.01 to 1.09] m/s; African women: +1.04 [95%CI 1.00 to 1.08] m/s).

Compared to those with a valid PWV measurement (n=717), Asians and Africans with no PWV measurement (n=242) were more often female with higher BMI and WC values. The distribution of age, MAP, glucose, and lipid spectrum was similar across the two groups (data not shown). Compared to excluded Asians (n=145), excluded Africans (n=97) were younger, but other important determinants of PWV, such as MAP and sex, were similar. Therefore, if data on PWV were available in those excluded, it is likely that the ethnic difference in PWV remained.

Discussion

The main finding of this study is that persons of Asian ancestry had a substantially higher PWV of approximately 1 m/s compared to those of African ancestry, after ad- justment for age and BP. This corresponds to a 15% increased risk of CVD mortality.11 The ethnic difference in PWV persisted after additional adjustment for sex and other cardiovascular risk parameters, including glucose, lipid spectrum, BMI, and WC.

Few studies have compared PWV in Asian and African ancestry populations. The study by Snijder et al. compared PWV values estimated by the Arteriograph in different an- cestry groups living in the Netherlands (aged 18-70 years).17 Interestingly, this study also found higher PWV values in individuals of South-Asian ancestry after adjustment for cardiovascular risk parameters, including age and MAP. This is in accordance with our finding and with two UK-based studies. The Southall And Brent REvisited (SABRE) study compared arterial stiffness across older South Asians, Africans, and White Europeans in the UK (mean age 70 years).15 In addition to measuring the carotid-femoral PWV, they calculated the elasticity coefficient (the ratio of central pulse pressure to stroke volume, cPP/SV, i.e. the inverse of total arterial compliance) to assess arterial stiffness. Although no ethnic differences in PWV were detected, South Asians had a higher cPP/SV than White Europeans and Africans, which persisted after adjustment for traditional cardio- vascular risk factors. The study by Rezai and colleagues was conducted in a relatively small cohort of 68 South Asians, 67 Africans, and 63 White Europeans. PWV, measured with the Arteriograph, was significantly higher (on average + 0.5 [SE 0.2] m/s) in South Ethnic differences in PWV 119

Asians compared to both Africans and White Europeans, after adjustment for age, BP, and diabetes.16

Consistent with the literature, we found a clear age-dependent rise in PWV in both ethnic groups.7,30 Furthermore, differences in PWV between Asians and Africans increased with age, with significantly higher PWV values in Asians aged 50 years and above compared to their African counterparts. In the study by Snijder et al., the ethnic difference in PWV became apparent at a lower age cutoff value of 40 years, which may be explained by the different reference group used (i.e. Caucasians in the study by Snijder and colleagues17 vs Africans in our study). The higher PWV in individuals of Asian ancestry compared to those of African or Caucasian ancestry points towards early vascular ageing.31 Sex- stratified analyses demonstrated higher PWV values in women than in men from both ethnic groups. Whether PWV differs in men and women is not yet conclusive. Several studies found no sex differences in aortic PWV.30,32,33 Yet, sex differences in PWV that become apparent with increasing age are more consistently reported and linked to menopausal-related changes.32,33,34,35 In our study, sex differences in PWV marginally increased with age in Asians but decreased in Africans (data not shown). Moreover, a previous validation study showed that the Arteriograph measures relatively high PWV values in women when PWV is higher (> 8 m/s) compared to the Sphygmocor.36 However, these findings should be viewed with caution, as both age- and sex-specific analyses in PWV were not the primary endpoints of the study and hampered by limited sample sizes.

It remains largely unknown why a higher PWV is found in Asians, even when adjusted for or in the absence of cardiovascular risk factors. Previous studies found that an in- crease in aortic PWV by 1 m/s corresponds to an age-, sex-, and risk factor-adjusted risk 7 increase of 15% in CVD mortality.11 The higher PWV of 1 m/s found in Asians compared to Africans corresponds to the higher coronary heart disease risk seen in Asian popula- tions in Suriname and in other regions.3,4,37,38,39 Although it remains of major importance to detect and control traditional cardiovascular risk factors, additional screening for arterial stiffness may improve cardiovascular risk assessment, particularly in persons of Asian ancestry. Future studies should evaluate the clinical potential of measuring the PWV in the assessment of cardiovascular disease risk.

Several limitations need to be addressed. First, the inferences that can be made are limited by the cross-sectional design of the study. Second, PWV was estimated using the Arteriograph device which is, though extensively validated, not identical to the current golden standard carotid-femoral PWV.24,25,40 Furthermore, although the analyses were adjusted, these confounders measured in a cross-sectional setting may not be 120 Chapter 7

appropriate surrogates for the lifetime exposure to these factors. For example, MAP will only partially capture the lifetime exposure to fluctuating MAP levels in an individual. Also, hypertension can be confounded by the use of medication or unknown factors that are associated with long-term hypertension. However, adjustment for hyperten- sion treatment resulted in a similar difference in PWV between people of Asian and African ancestry (data not shown). Finally, other confounders, including vitamin D status and inflammatory markers, such as C-reactive protein and interleukin-6, are known to be associated with ethnic differences in PWV, but were not measured in the current study.16,41,42

In conclusion, we compared PWV estimates in an Asian and African ancestry population and found substantial higher PWV values in Asians. This persisted after adjustment for important cardiovascular risk parameters, including age and BP. The higher PWV in Asians correspond to their increased coronary heart disease risk. Ethnic differences in PWV 121

References

1. Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2095-2128. 2. Agyemang C, Kieft S, Snijder M, et al. Hypertension control in a large multi-ethnic cohort in Amsterdam, The Netherlands: The HELIUS study. Int J Cardiol. 2015;183:180-189. 3. Chaturvedi N. Ethnic differences in cardiovascular disease. Heart. 2003;89(6):681-686. 4. Tillin T, Hughes A, Mayet J, et al. The relationship between metabolic risk factors and incident cardiovascular disease in Europeans, South Asians, and African Caribbeans: SABRE (Southall and Brent Revisited) - a prospective population-based study. Journal of the American Col- lege of Cardiology. 2013;61(17):1777-1786. 5. Rabanal K, Meyer H, Tell G, et al. Can traditional risk factors explain the higher risk of cardio- vascular disease in South Asians compared to Europeans in Norway and New Zealand? Two cohort studies. BMJ Open. 2017;7(12):e016819. 6. Cavalcante J, Lima J, Redheuil A, Al-Mallah M. Aortic stiffness: current understanding and future directions. J Am Coll Cardiol. 2011;57(14):1511-1522. 7. O’Rourke M, Hashimoto J. Mechanical factors in arterial aging: a clinical perspective. J Am Coll Cardiol. 2007;50(1):1-13. 8. Laurent S. Defining vascular aging and cardiovascular risk. J Hypertens. 2012;30(S3-8). 9. Ben-Shlomo Y, Spears M, Boustred C, et al. Aortic pulse wave velocity improves cardiovascu- lar event prediction: an individual participant meta-analysis of prospective observational data from 17,635 subjects. J Am Coll Cardiol. 2014;63(7):636-646. 10. Maroules C, Khera A, Ayers C, et al. Cardiovascular outcome associations among cardiovas- cular magnetic resonance measures of arterial stiffness: the Dallas heart study.J Cardiovasc Magn Reson. 2014;16:33-42. 11. Vlachopoulos C, Aznaouridis K, Stefanadis C. Prediction of cardiovascular events and all- cause mortality with arterial stiffness: a systematic review and meta-analysis. J Am Coll Cardiol. 2010;55(13):1318-1327. 12. Baldo M, Cunha R, Ribeiro A, et al. Racial differences in arterial stiffness are mainly de- termined by blood pressure levels: results from the ELSA-Brasil study. J Am Heart Assoc. 7 2017;6(6):e005477. 13. de Lima Santos P, de Oliveira Alvim R, Ferreira N, et al. Ethnicity and arterial stiffness in Brazil. Am J Hypertens. 2010;24(3):278-284. 14. Goel A, Maroules C, Mitchell G, et al. Ethnic difference in proximal aortic stiffness: an obser- vation from the Dallas Heart Study. JACC: Cardiovasc Imaging. 2017;10(1):54-61. 15. Park C, Tillin T, March K, et al. Adverse effect of diabetes and hyperglycaemia on arterial stiffness in Europeans, South Asians, and African Caribbeans in the SABRE study. J Hyper- tens. 2016;34(2):282-289. 16. Rezai M, Wallace A, Sattar N, Finn J, Wu F, Cruickshank J. Ethnic differences in aortic pulse wave velocity occur in the descending aorta and may be related to vitamin D. Hypertension. 2011;58(2):247-53. 17. Snijder M, Stronks K, Agyemang C, Busschers W, Peters R, van den Born BJ. Ethnic differences in arterial stiffness the Helius study. Int J Cardiol. 2015;191:28-33. 122 Chapter 7

18. Faconti L, Nanino E, Mills C, Cruickshank K. Do arterial stiffness and wave reflec- tion underlie cardiovascular risk in ethnic minorities? JRSM Cardiovasc Disease. 2016(5):2048004016661679. 19. The World Bank Group. Suriname. 2017. Available at: http://data.worldbank.org/country/ suriname. 20. Algemeen Bureau voor de Statistiek. Resultaten achtste (8e) volks- en woningtelling in Suriname: ressorten van de districten Paramaribo en Wanica naar etnische groep. 2012. Available at: http://www.statistics-suriname.org/index.php/statistieken/downloads/ category/30-censusstatistieken-2012. 21. Diemer F, Aartman J, Karamat F, et al. Exploring cardiovascular health: the Healthy Life in Su- riname (HELISUR) study. A protocol of a cross-sectional study. BMJ Open. 2014;4:e006380. 22. Diemer F, Baldew S, Haan Y, et al. Hypertension and cardiovascular risk profile in a middle- income setting: the HELISUR study. Am J Hypertens. 2017;30(11):1133-1140. 23. Lehr R. Sixteen S-squared over D-squared: a relation for crude sample size estimates. Statis- tics in Medicine. 1992;11:1099-1102. 24. Horvath I, Nemeth A, Lenkey Z, et al. Invasive validation of a new oscillometric device (Arte- riograph) for measuring augmentation index, central blood pressure and aortic pulse wave velocity. J Hypertens. 2010;28(10):2068-2075. 25. Baulmann J, Schillings U, Rickert S, et al. A new oscillometric method for assessment of arterial stiffness: comparison with tonometric and piezo-electronic methods. J Hypertens. 2008;26(3):523-528. 26. Mancia G, Faqard R, Narkiewicz K, et al. 2013 ESH/ESC Guidelines for the management of arterial hypertension: the Task Force for the management of arterial hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens. 2013;31(7):1281-1357. 27. American Diabetes Association. Classification and diagnosis of diabetes. Diabetes Care. 2016;39:S13-S22. 28. Expert panel on detection, evaluation, and treatment of high blood cholesterol in adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA. 2001;285(19):2486-2497. 29. The National Institute for Health and Care Excellence. Body mass index and waist circumfer- ence thresholds for intervening to prevent ill health among black, Asian and other minority ethnic groups. 2013. Available at: http://www.nice.org.uk/advice/lgb13/resources/body- mass-index-thresholds-for-intervening-to-prevent-ill-health-among-black-asian-and- other-minority-ethnic-groups-60521144747461. 30. Mitchell G, Parise H, Benjamin E, et al. Changes in arterial stiffness and wave reflection with advancing age in healthy men and women: the Framingham Heart Study. Hypertension. 2004;43(6):1239-1245. 31. Nilsson P, Lurbe E, Laurent S. The early life origins of vascular ageing and cardiovascular risk: the EVA syndrome. Journal of Hypertension. 2008;26(6):1049–1057. 32. McEniery C, Yasmin, Hall I, et al. Normal vascular aging: differential effects on wave reflec- tion and aortic pulse wave velocity: the Anglo-Cardiff Collaborative Trial (ACCT). JACC. 2005;46(9):1753-1760. 33. Logan J, Barksdale D. Pulse wave velocity in Korean American men and women. Journal of Cardiovascular Nursing. 2013;28(1):90-96. Ethnic differences in PWV 123

34. Lee H, Oh B. Aging and arterial stiffness. Circulation Journal. 2010;74(11):2257-2265. 35. Suzuki H, Kondo K. Pulse wave velocity in postmenopausal women. Pulse. 2013;1(1):4-13. 36. Ring M, Eriksson M, Zierath J, Caidahl K. Arterial stiffness estimation in healthy subjects: a validation of oscillometric (Arteriograph) and tonometric (SphygmoCor) techniques. Hypertension Research. 204;37(11):999-1007. 37. Punwasi W. Doodsoorzaken Suriname 2010-2011. Paramaribo 2012. 38. Bahall M, Seemungal T. An audit of cardiac mortality due to acute myocardial infarction at a tertiary institution in the southwestern region of Trinidad and Tobago. West Indian Medical Journal. 2015;2(2):60-65. 39. Harding S, Rosato M, A T. Trends for coronary heart disease and stroke mortality among migrants in England and Wales, 1979–2003: slow declines notable for some groups. Heart. 2008;94(4):463-470. 40. Van Bortel L, Laurent S, Boutouyrie P, et al. Expert consensus document on the measurement of aortic stiffness in daily practice using carotid-femoral pulse wave velocity. J Hypertens. 2012;30:445-448. 41. Raed A, Bhagatwala J, Zhu H, et al. Dose responses of vitamin D3 supplementation on arterial stiffness in overweight African Americans with vitamin D deficiency: A placebo controlled randomized trial. PLoS One. 2017;12(12):e0188424. 42. Cruickshank J, Silva M, Molaodi O, et al. Ethnic Differences in and Childhood Influences on Early Adult Pulse Wave Velocity. Hypertension. 2016;67(6):1133-1141.

7

Part III A comparison with Surinamese in the Netherlands

Chapter 8 Hypertension prevalence, awareness, treatment, and control in Surinamese in a middle- versus high-income country: the HELISUR and HELIUS studies

Frederieke S. Diemer, Marieke B. Snijder, Charles A. Agyemang, Yentl C. Haan, Fares A. Karamat, Ron J. G. Peters, Gert A. van Montfrans, Glenn P. Oehlers, Lizzy M. Brewster*, Karien Stronks*

* Both authors contributed equally Submitted 128 Chapter 8

Abstract

Objective: We studied hypertension prevalence, awareness, treatment, control and determinants related to hypertension among persons living in a middle-income country (MIC) com- pared with those of similar ethnicity living in a high-income country.

Methods: We used data of the HELISUR and HELIUS studies among Surinamese and Surinamese migrants living in the Netherlands (aged 18-70 years), respectively. Groups were formed based on country and self-defined ethnicity (African/South-Asian), and stratified by sex. Age-adjusted odds ratios (OR) with 95% confidence intervals (CI) were calculated and multivariable logistic regression was used to assess determinants of hypertension.

Results: We included 1,000 Surinamese and 6,971 first and second generation Surinamese migrants in the Netherlands. Crude hypertension prevalence was similar in Suriname (ranging from 40 to 44% between subgroups) and the Netherlands (36-47%), with similar control rates (Suriname: 34-44%; the Netherlands: 34-49%). After adjustment for age, no significant differences in hypertension prevalence, awareness, treatment and control between countries were seen in men. However, women in Suriname were substantially more often hypertensive with lower levels of awareness and control than Surinamese women in the Netherlands (African: OR 1.54 [95%CI 1.19, 2.00]; South- Asian: 1.90 [1.35, 2.67]; awareness: 0.62 [0.43, 0.88] in African women; control: 0.48 [0.28, 0.84] in South-Asian women). In both countries, age and waist circumference were the most important determinants of hypertension.

Conclusions: Particularly women in Suriname bear a relatively high hypertension burden with lower levels of control. Early blood pressure screening and obesity prevention may reduce the hypertension burden in urban regions of MIC. Hypertension in Surinamese living in Suriname and the Netherlands 129

Introduction

Hypertension is the leading risk factor for mortality worldwide, and is more common in low- and middle-income countries (LMIC) than in high-income countries (HIC).1,2,3 With many effective and inexpensive treatments available, hypertension control and preven- tion of subsequent hypertension-related diseases should be achievable.2 Nevertheless, hypertension control rates are suboptimal in LMIC, and no significant improvements were observed in the past decade.2,3 In contrast to LMIC, the prevalence of hyperten- sion slightly decreased and the level of hypertension control substantially increased in HIC.2,3 This indicates that there is still considerable room for improvement for LMIC. A comparison with a HIC, especially within populations with similar ethnicity, may give insight on where that “room” is, thereby providing a scientific basis for the develop- ment of public health interventions.

Suriname, a middle-income country (MIC) in South America, bears a high burden of hypertension.4,5,6 About half of the population migrated to the Netherlands between 1970 and 1990. This unprecedented mass migration offers a unique opportunity to study differences in the prevalence, awareness, treatment, and control of hypertension between MIC and HIC. This comparison could shed light on where significant improve- ments can be made and what populations should be targeted in order to reduce the high hypertension burden in Suriname. Therefore, we compared the Healthy Life in Suriname (HELISUR) study in urban Suriname with the Healthy Life in an Urban Setting (HELIUS) study, which included Surinamese who migrated to the Netherlands; both studies using similar methodology.7,8,9 Our main objective was to compare hypertension prevalence, awareness, treatment and control levels between Surinamese living in Suriname and Surinamese living in the Netherlands. As a secondary objective, we evaluated whether the underlying determinants of hypertension differed between the two countries.

Methods 8

Study population and study design We used cross-sectional data from the HELISUR and HELIUS studies, which were con- ducted in the capitals of Suriname and the Netherlands, respectively. Data from both studies were collected with similar questionnaires and physical and laboratory assess- ments. Ethical clearance was obtained from the respective ethics committees and par- ticipants gave written informed consent prior to the enrolment in the studies. The full details of both studies have been described elsewhere.7,8,9 In short, the HELISUR study was carried out between 2013 and 2015.7 The capital Paramaribo is divided into 1,200 130 Chapter 8

enumeration areas out of which 18 were randomly selected by the General Bureau of Statistics. In order to include 1,800 participants, trained interviewers were sent to the different enumeration areas, visiting every house and all household members until 100 persons were included. To reduce selection bias, if household members were not at home, interviewers revisited the address with a maximum of three times. The response rate was 72%. Subsequently, participants were invited for a physical examination at the local Academic Hospital. Of the 1,800 participants, 1,485 were eligible for physical examination (n=9 died; n=3 outside age range after re-evaluation; n=303 could not be retrieved (e.g. changed phone number or moved to a different address)). Of the 1,485 eligible individuals, 328 declined to participate (participation rate: 78%). A total of 1,157 persons participated in the interview as well as the physical examination.

In the Netherlands, the HELIUS study was conducted from 2011 to 2015 in Amsterdam among six ethnic groups, including people with a Surinamese background.8,9 The Dutch municipal registry was used to identify Surinamese subjects according to the country of birth of the participant or their parents.10 More specifically, a person was defined as Surinamese if he/she fulfilled one of two criteria: (1) he/she was born in Suriname and has at least one parent born in Suriname (first generation), or (2) he/she was born in the Netherlands and both parents were born in Suriname (second generation). Participants were randomly selected from the municipal registers, stratified by ethnicity, and were sent invitation letters to participate in the study. Of those invited, 62% of the Suri- namese origin participants were contacted. Of them, 51% agreed to participate in the study.9 Overall, 7,694 participants with a Surinamese background filled in the HELIUS questionnaire and underwent a physical examination at one of the research locations.

Participants from both studies were classified as African or South-Asian according to their self-reported ethnicity, in order to explore potential ethnic differences in the outcome measures. We refer to people of African and South-Asian ancestry who live in Suriname as African-SU and South-Asian-SU, whereas those Surinamese migrated or born from parents migrated from Suriname to the Netherlands will be indicated with African-NL and South-Asian-NL.

Participants’ flow of both studies is depicted in Figure 1. Surinamese of non-African or non-South-Asian ancestry were excluded from the current analysis (n=151 in HELISUR; n=500 in HELIUS). We additionally excluded participants with missing data on any of the variables (n=6 in HELISUR; n=223 in HELIUS) and assumed that these data were missing at random, resulting in a complete-case-only analysis with 1,000 Surinamese living in Suriname and 6,971 Surinamese living in the Netherlands. Hypertension in Surinamese living in Suriname and the Netherlands 131

Suriname (HELISUR)

Total participated N=1,157 Non-African/non-South Asian n=151

African/South Asian n=1,006

Missing data n=6 Complete cases n=1,000

Men Women n=375 n=625

African-SU South-Asian-SU African-SU South-Asian-SU n=216 n=159 n=386 n=239

The Netherlands (HELIUS)

Total participated n=7,694 Non-African/non-South Asian n=500

African/South Asian n=7,194

Missing data n=223 Complete cases n=6,971 8

Men Women n=2,899 n=4,072

African-NL South-Asian-NL African-NL South-Asian-NL n=1,556 N=1,343 n=2,447 N=1,625

Figure 1. Flowchart of both study populations 132 Chapter 8

Measurements Educational level was based on the highest qualification obtained. Those with no edu- cation or primary school only were considered as low educational level. Alcohol intake in the past 12 months (yes vs no) and smoking status (yes vs no or ex-smoker) were also obtained by questionnaires. Physical activity was estimated using the Interna- tional Physical Activity Questionnaire-long form (IPAQ-LF) in HELISUR, and the SQUASH questionnaire in HELIUS.11,12 The duration (min/week physically active) was divided into quartiles, and the first quartile (least physically active) was classified as low physical -ac tivity. All participants were asked to bring their prescribed medications to the research locations and antihypertensive agents were identified using Anatomical Therapeutic Chemical (ATC) classification system, i.e. ATC-codes C02 (antihypertensive agents), C03 (diuretics), C07 (beta-blocker), C08 (calcium antagonists), and C09 (ACE-inhibitors or angiotensin antagonists) were considered antihypertensive medication.

Height, weight and waist circumference (WC) were measured in duplicate to the nearest 0.1 cm or 0.1 kg, with participants wearing no shoes and light clothing only. A third measurement was made if the difference between the first two readings was more than 0.5 cm (for height), 0.5 kg (for weight), or 1.0 cm (for WC). Mean height, weight and WC were calculated by taking the mean of the two measurements (the two closest values if a third measurement was obtained). Body mass index (BMI) was computed as weight in kilograms / height in meters squared. For the definition of obesity and abdominal obesity, we used the respective ethnic-specific BMI and WC cut-off values that corre- spond to a high cardiovascular risk.13,14 In specific, obesity was defined as BMI ≥30 kg/ m2 for persons of African ancestry, and as BMI ≥27.5 kg/m2 for those of South-Asian ancestry.13,14 Abdominal obesity was defined as WC ≥94 cm in African men, ≥90 cm in South-Asian men, and ≥80 cm in women.13,14

Blood pressure (BP) was measured twice in the sitting position after a 5 minute rest with an automated oscillometric device (WatchBP Office; Microlife AG, Widnau, Switzerland), using an appropriately adjusted cuff size on the left upper arm supported at heart level. Hypertension was defined as a systolic BP ≥140 mm Hg, or a diastolic BP ≥90 mm Hg, or being on antihypertensive medication.1 Awareness of hypertension was defined as the proportion of hypertensive individuals who self-reported any prior diagnosis of hypertension by a health care professional. Treatment of hypertension was defined as the proportion of hypertensive individuals who were receiving prescribed antihyper- tensive medication for high BP management. BP control was defined as the proportion of hypertensive individuals on antihypertensive medication with systolic BP <140 mm Hg and diastolic BP <90 mm Hg.1 Hypertension in Surinamese living in Suriname and the Netherlands 133

Statistical analysis Characteristics of the study population were expressed as proportions or means with standard deviations (SD). Because of significant differences between men and women in the ethnic differences in hypertension, we stratified our analyses by sex. To assess potential differences in the age distribution of hypertension across the subgroups, the following age categories were formed: 20-29, 30-39, 40-49, 50-59, and 60-69 years. We calculated age-adjusted odds ratios (OR) with 95% confidence intervals (CI) by means of logistic regression analysis to study differences in hypertension prevalence, awareness, treatment and control across the subgroups. Subsequently, univariable logistic regression analysis was performed to explore the underlying determinants of hypertension stratified by country and ethnicity. We focused on risk factors of hyperten- sion, including age, low educational level, smoking, alcohol consumption, low physical activity, BMI, and WC. Per ethnic group in each country, multivariable logistic regression models were generated using forced entry. In case of multicollinearity, the variable with the highest Wald-statistic (i.e. lowest p-value) was added to the multivariable model. Additionally, within each ethnic group, we tested for effect modification by country by adding interaction terms (determinant x country) to the regression models. Model fit was assessed with the Hosmer-Lemeshow goodness-of-fit test for logistic regression models. Statistical analysis of the data was carried out using the SPSS version 20.0 (SPSS, Inc., Chicago, IL, USA) software for Windows. Graphical presentation was based on GraphPad Prism version 7.0 (GraphPad Software, Inc., San Diego, CA).

Results

Characteristics of the study population For both sexes and ethnicities, participants living in Suriname were generally younger, had more often a low educational level and less physically active than those living in the Netherlands (Table 1). In men, no large differences were observed in the prevalence of smoking, alcohol consumption, obesity, and abdominal obesity between the two 8 countries. In women, however, the prevalence of smoking and alcohol consumption was significantly lower, and the prevalence of obesity and abdominal obesity was signifi- cantly higher, in Suriname compared to the Netherlands.

No significant differences in crude BP levels were seen between the two countries, except for a higher diastolic BP in South-Asian women living in Suriname compared to their counterparts living in the Netherlands (Table 1). Regarding ethnic differences in BP within the Netherlands, we found significantly higher crude systolic and diastolic BP levels in African men and women compared to South-Asian men and women (all 134 Chapter 8

p<0.01). In Suriname, only African men had higher crude systolic BP levels than South- Asian men.

Table 1. Characteristics of both study populations MEN South-Asian- South-Asian- African-SU African-NL SU NL n=216 n=1,556 n=159 n=1,343 Age, mean (SD) 43.0 (15.1) 48.0 (12.9) 43.4 (13.5) 44.7 (13.6) Low educational level, % 28.2 6.4 40.3 12.6 Tobacco smoking, % 38.0 42.8 42.1 39.8 Alcohol consumption (past 12 months), % 78.7 78.9 70.4 67.5 Physical activity, % < 1140 min/week (Q1) 56.3 20.2 56.0 24.8 1140–2340 min/week (Q2) 20.9 21.1 21.4 21.8 2340–3420 min/week (Q3) 12.6 28.3 8.8 28.5 > 3420 min/week (Q4) 10.2 30.3 13.8 24.9 Body mass index, mean (SD) 25.5 (5.0) 26.3 (4.4) 25.7 (5.0) 25.8 (4.1) Obesity, % 17.6 17.2 28.3 27.7 Waist circumference, mean (SD) 90.3 (14.2) 92.2 (12.5) 95.3 (14.1) 93.8 (12.1) Abdominal obesity, % 37.5 40.6 66.7 62.0 Systolic BP, mean (SD) 133.9 (20.0) 134.3 (17.3) 129.5 (16.6) 130.5 (16.3) Diastolic BP, mean (SD) 83.7 (12.8) 84.5 (10.7) 83.1 (11.0) 82.5 (10.1) South-Asian- South-Asian- WOMEN African-SU African-NL SU NL n=386 n=2,447 n=239 n=1,625 Age, mean (SD) 42.3 (13.7) 47.8 (12.2) 43.4 (13.0) 46.0 (13.3) Low educational level, % 34.2 4.7 46.4 15.3 Tobacco smoking, % 9.1 24.4 7.1 19.0 Alcohol consumption (past 12 months), % 42.7 62.2 39.3 47.6 Physical activity, % < 1140 min/week (Q1) 68.5 18.5 50.8 18.8 1140–2340 min/week (Q2) 20.5 27.0 33.2 28.8 2340–3420 min/week (Q3) 5.8 27.0 9.2 27.0 > 3420 min/week (Q4) 5.2 27.5 6.7 25.5 Body mass index, mean (SD) 29.9 (7.0) 28.8 (5.9) 28.2 (5.7) 26.7 (5.3) Obesity, % 46.1 37.4 46.9 39.0 Waist circumference, mean (SD) 98.2 (16.9) 93.4 (14.8) 97.1 (14.5) 90.0 (13.5) Abdominal obesity, % 85.2 81.0 90.8 76.9 Systolic BP, mean (SD) 129.6 (21.2) 129.9 (18.4) 127.9 (20.4) 126.4 (19.6) Diastolic BP, mean (SD) 80.6 (11.6) 80.2 (10.6) 80.2 (10.8) 77.5 (10.2) Footnote Table 1. Values are mean with standard deviations (SD) or proportions (%). Q1-4, quartile 1 to 4; BP, blood pressure. Hypertension in Surinamese living in Suriname and the Netherlands 135

Prevalence of hypertension The age distribution of hypertension in men and women is depicted in Figure 2. Across all age categories (except <30 years in men), crude hypertension was more prevalent in participants living in Suriname than in those living in the Netherlands. These differ- ences were more pronounced in the younger age categories (<30, 30-39 and 40-49 years) and more in women than in men.

Men 100 90 80 * 70 60 50 40

Prevalence Prevalence (%) 30 20 10 0 20-29 30-39 40-49 50-59 60-69 Age category

African-SU African-NL South-Asian-SU South-Asian-NL Figure 2. Hypertension by 10-year age groups in (A) men and (B) women. *significantly higher than African-NL; †significantly higher than South-Asian-NL.

Crude hypertension prevalence in both men and women did not vary between the two countries (Figure 3). In the Netherlands, crude hypertension prevalence was higher in African ancestry men and women than in respectively South-Asian ancestry men and women (both p<0.01). This ethnic difference in the prevalence of hypertension was not seen in Suriname. After adjustment for age, no differences were seen in the prevalence of hypertension in men from Suriname vs the Netherlands (Figure 3-A). In 8 women, the prevalence of hypertension was significantly higher in Suriname than in the Netherlands (Figure 3-B). This could be explained by the fact that women in Suriname were more abdominally obese. After adjustment for WC, the difference in hypertension prevalence in women across both countries disappeared (data not shown).

Awareness, treatment and control of hypertension In men, hypertension awareness, treatment, and control did not vary substantially between the two countries (Figure 3-A). This was seen both in African and South-Asian 136 Chapter 8

A) MEN n/N OR (95%CI)

SU NL SU vs NL (ref)

African Hypertension 95/216 (44.0) 724/1,556 (46.5) 1.30 (0.94 to 1.80)

Awareness 51/95 (53.7) 366/724 (50.6) 1.24 (0.79 to 1.94)

Treatment 41/95 (43.2) 303/724 (41.9) 1.18 (0.74 to 1.89)

Control 14/41 (34.1) 102/303 (33.7) 1.08 (0.54 to 2.16)

South-Asian Hypertension 68/159 (42.8) 549/1,343 (40.9) 1.29 (0.88 to 1.89)

Awareness 40/68 (58.8) 284/549 (51.7) 1.58 (0.92 to 2.71)

Treatment 39/68 (57.4) 278/549 (50.6) 1.77 (1.00 to 3.15)

Control 18/39 (46.2) 112/278 (40.3) 1.24 (0.63 to 2.46)

B) WOMEN n/N OR (95%CI)

SU NL SU vs NL (ref)

African Hypertension 164/386 (42.5) 1,107/2,447 (45.2) 1.54 (1.19 to 2.00)

Awareness 90/164 (54.9) 775/1,107 (70.0) 0.62 (0.43 to 0.88)

Treatment 102/164 (62.2) 706/1,107 (63.8) 1.18 (0.82 to 1.69)

Control 45/102 (44.1) 333/706 (47.2) 0.87 (0.57 to 1.32)

South-Asian Hypertension 95/239 (39.7) 578/1,625 (35.6) 1.90 (1.35 to 2.67)

Awareness 56/95 (58.9) 360/578 (62.3) 1.07 (0.67 to 1.70)

Treatment 72/95 (75.8) 363/578 (62.8) 2.50 (1.47 to 4.24)

Control 25/72 (34.7) 177/363 (48.7) 0.48 (0.28 to 0.84)

Figure 3. Differences in hypertension prevalence, awareness, treatment and control between Suri- name and the Netherlands in men (A) and women (B), adjusted for age. Values are proportions (%) or age-adjusted odds ratios (OR) with 95% confidence intervals (CI).

men. African ancestry women living in Suriname were significantly less often aware of their high BP than their counterparts living in the Netherlands (Figure 3-B). This lower awareness persisted after adjustment for age. Treatment and control levels in African women were similar between the two countries. Among South-Asian women, awareness did not differ between the two countries. However, compared to their counterparts living in the Netherlands, South-Asian ancestry women living in Suriname were substantially more often treated, but less often controlled. After adjusting for age, these differences in treatment and control remained.

Determinants of hypertension Age, BMI, and WC were positively and significantly associated with hypertension preva- lence in men across all four subgroups in univariable analysis (p<0.01) (Supplemental Table 1). Low physical activity was associated with higher odds for hypertension in men, in both countries and ethnic groups. Moreover, a statistically significant associa- Hypertension in Surinamese living in Suriname and the Netherlands 137 tion with hypertension was seen for alcohol consumption (African-SU men), smoking (African-NL men), and low educational level (South-Asian-NL men). Age, BMI, and WC were also positively associated with hypertension prevalence in women across all four subgroups (p<0.01). In addition, in the Netherlands, women who had a self-report of smoking and consumed alcohol were less likely to have hypertension, and those with a low educational level or low physical activity were more likely to have hypertension. This was seen in both ethnic groups.

The results from the multivariable logistic regression analysis are shown in Table 2. Both in men and women, age and WC were the determinants that remained significantly associated with hypertension, in both countries and ethnic groups. Additionally, in

Table 2. Multivariable analysis of determinants of hypertension prevalence and control in men and women MEN African-SU African-NL South-Asian-SU South-Asian-NL (n=216) (n=1,556) (n=159) (n=1,343)

OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI) Age 1.08 (1.05, 1.11) 1.06 (1.05, 1.07) 1.08 (1.04, 1.11) 1.09 (1.08, 1.10) Low educational 0.70 (0.33, 1.52) 1.17 (0.74, 1.85) 0.52 (0.24, 1.13) 0.80 (0.54, 1.17) level Smoking 1.48 (0.75, 2.92) 0.82 (0.65, 1.03) 0.80 (0.37, 1.69) 0.98 (0.75, 1.28) Alcohol 0.58 (0.25, 1.34) 1.07 (0.81, 1.41) 2.00 (0.85, 4.72) 1.12 (0.84, 1.48) consumption Low physical 1.26 (0.64, 2.49) 1.15 (0.87, 1.53) 1.06 (0.50, 2.23) 1.12 (0.82, 1.52) activity WC 1.05 (1.02, 1.08) 1.04 (1.03, 1.05) 1.03 (1.00, 1.06) 1.06 (1.04, 1.07) WOMEN African-SU African-NL South-Asian-SU South-Asian-NL (n=386) (n=2,447) (n=239) (n=1,625)

OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI) Age 1.09 (1.06, 1.11) 1.09 (1.08, 1.10) 1.11 (1.08, 1.15) 1.12 (1.10, 1.14) Low educational 1.03 (0.61, 1.73) 0.96 (0.62, 1.50) 0.51 (0.25, 1.02)* 1.43 (1.01, 2.03) level 8 Smoking 0.86 (0.37, 2.02) 0.92 (0.74, 1.15) 0.74 (0.18, 2.98) 0.77 (0.55, 1.09) Alcohol 0.82 (0.49, 1.35) 0.84 (0.69, 1.02) 1.09 (0.55, 2.14) 0.76 (0.58, 0.98) consumption Low physical 0.68 (0.40, 1.15) 1.18 (0.93, 1.51) 1.49 (0.77, 2.90) 1.03 (0.74, 1.44) activity WC 1.04 (1.03, 1.06) 1.04 (1.03, 1.05) 1.08 (1.05, 1.11)* 1.02 (1.00, 1.04) Footnote Table 2. Values are odds ratios (OR) with 95% confidence intervals (CI). Significantly associated determinants after adjustment are depicted in bold. *Association is significantly differ- ent compared to the association in Surinamese living in the Netherlands (p-value interaction term < 0.05) BMI, body mass index; WC, waist circumference. 138 Chapter 8

South-Asian women living in the Netherlands, low educational level and consumption of alcohol were associated with, respectively, higher and lower odds for hypertension. Only the association between low educational level and hypertension in Suriname dif- fered significantly from the association in the Netherlands (p-value for interaction for country with educational level <0.05). Low educated persons in Suriname were less likely to have hypertension, whereas in the Netherlands this association was reversed.

Discussion

Key findings Our findings show that hypertension was highly prevalent in men and women living in Suriname, with suboptimal levels of awareness, treatment and control. While in men these estimates were similar to their counterparts living in the Netherlands, in women hypertension prevalence in Suriname was substantially higher and started at an earlier age compared to their counterparts living in the Netherlands. Also, the levels of aware- ness and control were lower in women in Suriname compared to those in the Nether- lands. Underlying behavioural determinants of hypertension in Suriname were similar to those in the Netherlands, with age and WC being associated with hypertension.

Limitations of the study Several limitations need to be addressed. First, although similar methods were used in both countries, the sampling procedure had to be adapted to the local situation. In Suriname, it was not feasible to randomly select persons using the municipal register. However, we randomly selected enumeration areas and minimized selection bias (i.e. persons who were not at home were less likely to be included than those who were at home) by revisiting every address to a maximum of three times until all household mem- bers were included or stated a refusal. Second, there were differences in sample size, with a smaller sample size in Suriname. This might have resulted in fewer significant as- sociations between the determinants and hypertension. Nevertheless, the distribution between ethnic-sex groups across both countries was similar. As in many epidemiologi- cal studies, the definition of hypertension was based on BP measurements within one visit, which might have overestimated the prevalence estimates. Also, antihypertensive treatment was based on the medication that participants brought with them to the research location. Persons who forgot to bring their medication and had normal blood pressures were misclassified as normotensives. Nevertheless, it is unlikely that these limitations modify the comparability between the subgroups, as these limitations ap- ply to all groups. Another limitation is the use of different measures to assess physical activity, i.e. with the IPAQ questionnaire in HELISUR and with the SQUASH questionnaire Hypertension in Surinamese living in Suriname and the Netherlands 139 in HELIUS. To improve the comparability between the two questionnaires, we used the overall duration of physical activity and divided it into quartiles to define those least physically active. However, we cannot state with confidence that inter-country differ- ences in physical activity are attributed to the use of different measures or are, in fact, true differences. Finally, we could not measure the association between hypertension and several important determinants of hypertension such as diet, salt intake, and psy- chosocial stress due to absence of the data in the study.

Discussion of the key findings

Prevalence of hypertension We found a higher prevalence of hypertension in women living in Suriname (LMIC) compared to their counterparts living in the Netherlands (HIC). This is consistent with earlier studies, observing a higher prevalence of hypertension in LMIC than HIC. How- ever, these studies were conducted among heterogeneous populations. Studies among more homogenous groups that migrated from a LMIC to a HIC show different results with regards to hypertension prevalence. For example, a recent study among African migrants (Ghanaians) living in Western Europe and non-migrants living in rural and urban Ghana found a clear gradient in the prevalence of hypertension.15 Hypertension prevalence increased from rural through urban Ghana to Europe.15 In our comparison between Surinamese living in a LMIC versus HIC, we found that this gradient of hyper- tension from LMIC to HIC does not seem to apply. In fact, it was the opposite for some groups, particularly women. In addition, hypertension was more prevalent at earlier ages in Suriname. The presence of hypertension in relatively young people might play a role in the high premature mortality rates observed in LMIC, where people often die in their most productive years.16 Importantly, the higher prevalence of hypertension in relatively young people in Suriname compared to the Netherlands clearly illustrates the magnitude of the hypertension burden in urban centers in LMIC.

Awareness, treatment, and control of hypertension 8 Awareness, treatment, and control rates were suboptimal in both countries. The low- est control levels were seen in African men (34% of those treated, 15% of those with hypertension; both in African-SU men as well as African-NL men). Interestingly, men in Suriname had similar awareness, treatment, and control levels as Surinamese men in the Netherlands, suggesting that for men the access and quality of the health care system of Suriname are similar to that of their counterparts in the Netherlands. In con- trast, women living in Suriname were less aware and, despite higher treatment levels, still less controlled for hypertension than their counterparts living in the Netherlands. Reasons behind this pattern remain unclear, but might relate to gender inequalities that 140 Chapter 8

women in LMIC are still facing to date.17 Gender inequalities in education, income, or employment may limit the ability of women from LMIC to protect their health.17 In Su- riname, for example, women have substantially lower health literacy levels than men.18 Consequently, they may have more difficulties navigating the health care system, understanding medical instructions, and being compliant to their medication regime, putting them at risk for adverse (cardiovascular) health outcomes.19,20,21 Focusing on patient education is therefore recommended to improve awareness of hypertension and adherence to antihypertensive treatment. However, it is important to recognize culturally and ethnically different perspectives on hypertension and antihypertensive treatment. A study in the Netherlands showed that particularly patients with a migra- tion background, including those with a Surinamese background, reported lowering or leaving off their prescribed antihypertensive dosage.22 Explanations for altering prescribed dosage were: disliking chemical medications, fear of side effects, preference for alternative treatment and, in Surinamese men, worries about the negative effects on their sexual performance.22 Future studies should focus on how cultural and ethnic fac- tors influence hypertension management in Suriname, but culturally appropriate health education may be beneficial for Surinamese patients with hypertension.

Determinants of hypertension prevalence The determinants associated with hypertension were similar in Suriname and in Suri- namese living in the Netherlands, with one exception for educational level. Low edu- cation was associated with hypertension in persons living in the Netherlands, but not among those living in Suriname. Nowadays, in HIC, persons with a low socioeconomic status (SES) generally experience a higher burden of risk factors.23 However, in the be- ginning of the 20th century, the association of SES and risk factors used to be reverse, which classified CVD as a ‘disease of affluence’.24 This epidemiological transition, in which the social gradient in risk factors changes overtime, is currently ongoing in many LMIC. In line with this, studies from these regions often demonstrated that increasing wealth and education were associated with unhealthy behaviour and adverse health outcomes in LMIC.25,26,27 These differences in the social gradient of risk factors between LMIC and HIC offer an attractive explanation for the contrasting association between education and hypertension seen in Suriname vs the Netherlands.

In both countries, age, BMI and WC were the most important determinants for hyperten- sion prevalence, which is consistent with the literature.28 This suggests that the high prevalence of hypertension in Surinamese is likely due to aging of the population and risk factors accompanied by westernization, including an unhealthy diet and sedentary lifestyle, which are important contributors for an elevated BMI or WC. The association between (abdominal) obesity and hypertension is well-established throughout numer- Hypertension in Surinamese living in Suriname and the Netherlands 141 ous studies and highlights the pivotal role of obesity in the hypertension pandem- ic.29,30,31 The association between age and hypertension is explained by the age-related structural and functional changes in blood vessels, which eventually cause narrowing of the vascular lumen and reduction of the arterial compliance. This, in turn, increases BP.32 It also explains the high proportion (>75%) of people with hypertension of 60 years and older.

We found a lower prevalence of hypertension among women in the Netherlands who smoke and consumed alcohol. Further exploration of the data demonstrated that those women were substantially younger and less (abdominally) obese, which could provide an explanation for this unexpected finding.

Both in Surinamese living in a MIC and in a HIC, hypertension is highly prevalent and BP levels are poorly controlled. Yet, those living in Suriname, particularly women, bear a higher burden of hypertension with lower levels of control. Major efforts are needed to reduce the high burden of hypertension and concomitant diseases in urban regions of LMIC. With age and WC being important behavioural determinants of hypertension, early screening for hypertension and interventions aimed at (abdominal) obesity pre- vention may reduce the high hypertension burden in this middle-income population.

In conclusion, we compared hypertension prevalence, awareness, treatment, control, and determinants related to hypertension in a MIC with a HIC within populations with the same ethnicity. We found that those living in Suriname, particularly women, bore the highest burden of hypertension with lower levels of control compared to their counterparts living in the Netherlands. Early BP screening and interventions aimed at (abdominal) obesity prevention are urgently needed to reduce the high burden of hypertension in urban regions in LMIC.

8 142 Chapter 8

References

1. Mancia G, Faqard R, Narkiewicz K, et al. 2013 ESH/ESC Guidelines for the management of arterial hypertension: the Task Force for the management of arterial hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). Journal of Hypertension. 2013;31:1281–1357. 2. Chow C, Teo K, Rangarajan S, et al. Prevalence, awareness, treatment, and control of hyper- tension in rural and urban communities in high-, middle-, and low-income countries. JAMA. 2013;310(9):959-968. 3. Mills K, Bundy J, Kelly T, et al. Global Disparities of Hypertension Prevalence and Con- trol: A Systematic Analysis of Population-Based Studies From 90 Countries. Circulation. 2016;134(6):441-450. 4. The World Bank Group. Suriname. 2017. Available at: http://data.worldbank.org/country/ suriname. 5. Algemeen Bureau voor de Statistiek. Resultaten achtste (8e) volks- en woningtelling in Suriname: ressorten van de districten Paramaribo en Wanica naar etnische groep. 2012. Available at: http://www.statistics-suriname.org/index.php/statistieken/downloads/ category/30-censusstatistieken-2012. 6. Diemer F, Baldew S, Haan Y, et al. Hypertension and cardiovascular risk profile in a middle- income setting: the HELISUR study. Am J Hypertens. 2017;30(11):1133-1140. 7. Diemer F, Aartman J, Karamat F, et al. Exploring cardiovascular health: the Healthy Life in Su- riname (HELISUR) study. A protocol of a cross-sectional study. BMJ Open. 2014;4:e006380. 8. Stronks K, Snijder M, Peters R, Prins M, Schene A, Zwinderman A. Unravelling the impact of ethnicity on health in Europe: the HELIUS study. BMC Public Health. 2013;13:402. 9. Snijder M, Galenkamp H, Prins M, et al. Cohort Profile: the Healthy Life in an Urban Setting (HELIUS) study. BMJ Open. 2017;):e017873. 10. Stronks K, Kulu-Glasgow I, Agyemang C. The utility of ‘country of birth’ for the classi- fication of ethnic groups in health research: the Dutch experience. Ethnicity and Health. 2009;14(3):255-269. 11. Blikman T, Stevens M, Bulstra S, van den Akker-Scheek I, Reininga I. Reliability and validity of the Dutch version of the International Physical Activity Questionnaire in patients after total hip arthroplasty or total knee arthroplasty. Journal of Orthopaedic and Sports Physical Therapy. 2013;43(9):650-659. 12. Wendel-Vos G, Schuit A, Saris W, Kromhout D. Reproducibility and relative validity of the short questionnaire to assess health-enhancing physical activity. Journal of Clinical Epide- miology. 2003;56:1163–1169. 13. WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363(9403):157-163. 14. National Institute for Health and Care Excellence. Assessing body mass index and waist circumference thresholds for intervening to prevent ill health and premature death among adults from black, Asian and other minority ethnic groups in the UK. 2013. Available at: https://www.nice.org.uk/guidance/ph46. 15. Agyemang C, Nyaaba G, Beune E, et al. Variations in hypertension awareness, treatment, and control among Ghanaian migrants living in Amsterdam, Berlin, London, and nonmigrant Ghanaians living in rural and urban Ghana - the RODAM study. J Hypertens. 2018;36(1):169- 177. Hypertension in Surinamese living in Suriname and the Netherlands 143

16. World Health Organization. Global status report on noncommunicable disease 2014. Geneva 2014. ISBN 978 92 4 156485 4. 17. Filippi V, Ronsmans C, Campbell O, et al. Maternal health in poor countries: the broader context and a call for action. The Lancet. 2006;368(9546):1535–1541. 18. Diemer F, Haan Y, Nannan Panday R, van Montfrans G, Oehlers G, Brewster L. Health literacy in Suriname. Social Work in Health Care. 2017;56(4):283-293. 19. Donovan-Kicken E, Mackert M, Guinn T, Tollison A, Breckinridge B, Pont S. Health literacy, self-efficacy, and patients’ assessment of medical disclosure and consent documentation. Health Communication. 2012;27(6):581-590. 20. McCarthy D, Waite K, Curtis L, Engel K, Baker D, Wolf M. What did the doctor say? Health literacy and recall of medical instructions. Medical Care. 2012;50(4):277-282. 21. Ngoh L. Health literacy: a barrier to pharmacist-patient communication and medication adherence. Journal of the American Pharmacists Association. 2009;49(5):e132-146. 22. Beune E, Haafkens J, Agyemang C, Schuster J, Willems D. How Ghanaian, African-Surinamese and Dutch patients perceive and manage antihypertensive drug treatment: a qualitative study. Journal of Hypertension. 2008;26(4):648-656. 23. Stringhini S, Sabia S, Shipley M, et al. Association of socioeconomic position with health behaviors and mortality. JAMA. 2010;303(12):1159-1166. 24. Wilkinson R. The epidemiological transition: from material scarcity to social disadvantage? Daedalus. 1994;123(4):61-77. 25. Subramanian S, Perkins J, Özaltin E, Davey Smith G. Weight of nations: a socioeconomic analysis of women in low- to middle-income countries. American Journal of Clinical Nutri- tion. 2011;93(2):413-421. 26. Rehkopf D, Dow W, Rosero-Bixby L. Differences in the association of cardiovascular risk factors with education: a comparison of Costa Rica (CRELES) and the USA (NHANES). Journal of Epidemiology and Community Health;64(9):821-828. 27. Allen L, Williams J, Townsend N, et al. Socioeconomic status and non-communicable disease behavioural risk factors in low-income and lower-middle-income countries: a systematic review. Lancet Global Health. 2017;5:e277-289. 28. Sarki A, Nduka C, Stranges S, Kandala NB, Uthman O. Prevalence of hypertension in low- and middle-income countries: a systematic review and meta-analysis. Medicine. 2015;94(50):e1959. 29. Patel S, Ali M, Alam D, et al. Obesity and its relation with diabetes and hypertension: a cross-sectional study across four low- and middle-income country regions. Global Heart. 2016;11(1):71-79 e4. 8 30. Strumpf E. The obesity epidemic in the United States: causes and extent, risks and solutions. New York, NY 2004. 31. Hall J. The kidney, hypertension, and obesity. Hypertension. 2003;41:625-633. 32. Pinto E. Blood pressure and ageing. Postgrad Med J. 2007;83:109–114. 144 Chapter 8

Supplemental table 1. Univariable associations of potential determinants with hypertension in (A) men and (B) women A) mEN African-SU African-NL South-Asian-SU South-Asian-NL (n=216) (n=1,556) (n=159) (n=1,343)

OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI) Age 1.09 (1.06, 1.12) 1.07 (1.06, 1.08) 1.08 (1.05, 1.11) 1.10 (1.08, 1.11) Low educational 0.93 (0.51, 1.68) 1.38 (0.92, 2.07) 0.62 (0.33, 1.19) 1.64 (1.18, 2.26) level Smoking 1.48 (0.85, 2.58)* 0.79 (0.65, 0.97) 0.68 (0.36, 1.29)* 0.90 (0.72, 1.12) Alcohol 0.47 (0.24, 0.91) 0.88 (0.69, 1.12) 1.48 (0.73, 2.98) 0.86 (0.68, 1.08) consumption Low physical 1.56 (0.90, 2.69) 1.34 (1.05, 1.72) 1.51 (0.80, 2.87) 1.43 (1.11, 1.84) activity BMI 1.13 (1.07, 1.20) 1.13 (1.10, 1.16) 1.10 (1.03, 1.18) 1.16 (1.12, 1.19) WC 1.07 (1.04, 1.09) 1.06 (1.05, 1.07) 1.05 (1.02, 1.08) 1.07 (1.06, 1.09) B) womEN African-SU African-NL South-Asian-SU South-Asian-NL (n=386) (n=2,447) (n=239) (n=1,625)

OR (95%CI) OR (95%CI) OR (95%CI) OR (95%CI) Age 1.10 (1.08, 1.12) 1.10 (1.09, 1.12) 1.10 (1.07, 1.14) 1.13 (1.12, 1.15) Low educational 1.52 (0.99, 2.32) 1.76 (1.21, 2.57) 1.22 (0.73, 2.06)* 4.00 (3.01, 5.31) level Smoking 1.02 (0.50, 2.05) 0.72 (0.59, 0.86) 0.82 (0.29, 2.28) 0.57 (0.43, 0.75) Alcohol 0.62 (0.41, 0.93) 0.54 (0.46, 0.64) 0.97 (0.57, 1.66)* 0.48 (0.39, 0.59) consumption Low physical 0.68 (0.45, 1.05)* 1.49 (1.21, 1.83) 0.99 (0.59, 1.67) 1.57 (1.22, 2.03) activity BMI 1.12 (1.08, 1.16) 1.13 (1.11, 1.14) 1.16 (1.10, 1.22) 1.15 (1.12, 1.17) WC 1.06 (1.04, 1.07) 1.06 (1.05, 1.06) 1.07 (1.05, 1.10) 1.07 (1.06, 1.08) Footnote Supplemental Table 1. Values are odds ratios (OR) with 95% confidence intervals (CI). p-values <0.05 are in bold. *Association is significantly different compared to the association in Surinamese living in the Neth- erlands (p-value interaction term < 0.05) BMI, body mass index; WC, waist circumference

Chapter 9 General discussion

Discussion 149

The general aim of this thesis was to expand the evidence on cardiovascular health and related risk factors in the urban middle-income population of Suriname, with a primary focus on ethnic differences in cardiovascular risk. Such data will support the develop- ment of effective preventive and intervention strategies aimed at reducing the high cardiovascular disease (CVD) burden in Suriname.

This final chapter presents the key findings of this thesis. First, the main findings are summarized. Second, reflections on the main findings are made, followed bysome methodological considerations. Finally, the implications of the findings and recommen- dations for future research are discussed.

Summary of the main findings

Our data showed that more than three-third (77%) of this urban Surinamese adult popu- lation had at least one cardiovascular risk factor: hypertension (40%), diabetes mellitus (22%), dyslipidemia (41%), obesity (37%) or tobacco smoking (18%) (Chapter 4). An additional 22% was prehypertensive, prediabetic, overweight or had borderline dyslip- idemia, resulting in a mere 1% of the population having an optimal cardiovascular risk factor profile. In those aged 40 years and younger, 62% had at least one cardiovascular risk factor and only 2% had an optimal cardiovascular risk factor profile (Chapter 4).

Marked differences were seen in the prevalence of traditional cardiovascular risk factors across the 4 ethnic groups (Chapter 4). Crude hypertension prevalence was highest in Creole (48%), which significantly differed from that of Maroons (38%) and South-Asians (37%). South-Asians had the highest crude prevalence of diabetes and dyslipidemia (respectively, 30% and 52%), followed by Indonesians (25% and 47%) and Creole (22% and 37%). In Maroons, the prevalence of diabetes and dyslipidemia was relatively low (respectively, 11% and 27%). For general obesity, no crude differ- ences were seen between the 4 ethnic groups.

After dichotomizing the ethnic groups into an African and Asian ancestry population, crude hypertension prevalence did not differ between Africans and Asians (Chapter 4, 7, 8). However, after adjustment for sex, age, body mass index (BMI), and educational 9 level, those of African ancestry were more likely to have hypertension compared to their Asian counterparts (Chapter 4). Diabetes mellitus was more often seen in persons of Asians ancestry, both with HbA1c criteria (Chapter 4) and with fasting plasma glucose criteria (Chapter 7). Also, dyslipidemia was more prevalent in persons of Asian ancestry than in those of African ancestry (Chapter 4 and 7). For general obesity, no differences 150 Chapter 9

were seen between African and Asian ancestry participants (Chapter 4, 6, 7). However, abdominal obesity was substantially more prevalent in Asians compared to Africans (Chapter 5 and 6). Body composition assessments showed that Asians had lower fat- free mass and a higher fat percentage at similar or lower BMI levels as their African counterparts (Chapter 5). In addition, Chapter 5 suggested that of body composition measures, measures for abdominal obesity such as waist circumference should be used to accurately identify persons at risk for CVD. Regarding behavioural risk factors, the prevalence of smoking was similar in Asians and Africans (Chapter 6). Asians were slightly more physically active than Africans (Chapter 6). Meeting the physical activity recommendations was associated with lower obesity prevalence in the total population (Chapter 6). Also, ethnic differences existed in the association between obesity and physical activity characteristics, with leisure time activity and the overall duration of the activity being associated with lower odds for obesity in the African population but not in the Asian population (Chapter 6). In the Asian population, no association could be established between different physical activity characteristics and obesity (Chapter 6).

In addition to traditional cardiovascular risk factors, we assessed early organ damage, which was prevalent in 22% of the population. The majority of these individuals had increased arterial stiffness (77%) (Chapter 4). Moreover, we found that increased arte- rial stiffness varied by ethnic group (Chapter 7). Individuals of Asian ancestry had a substantially higher pulse wave velocity (PWV) of approximately 1 m/s compared to those of African ancestry, after adjustment for age and blood pressure (BP) (Chapter 7). This ethnic difference in PWV could not be explained by differences in other car- diovascular risk parameters, including fasting glucose, lipid spectrum, BMI, and waist circumference (Chapter 7).

Our findings showed that in this urban Surinamese population hypertension awareness, treatment, and control were suboptimal, with no differences between the ethnic groups (Chapter 4 and 8). We compared our data with those on Surinamese living in the Nether- lands (Chapter 8). This comparison demonstrated that in both countries, hypertension is highly prevalent and BP levels poorly controlled. Yet, those living in Suriname, particu- larly women, bore a higher burden of hypertension with lower levels of control. Age, waist circumference, and BMI were important determinants of hypertension. Therefore, we recommend the design and implementation of preventive programs aimed at the early detection of hypertension and (abdominal) obesity prevention in order to reduce the high hypertension burden in urban Suriname (Chapter 8). Discussion 151

Reflection on the main findings

Urban Suriname bears a high cardiovascular risk factor burden In many low- and middle-income countries (LMIC), the lack of data on cardiovascular risk factors hampers the design and implementation of effective prevention strategies.1 This thesis provides important information on the status quo of cardiovascular risk factors and early organ damage in urban Suriname. Our extensive cardiovascular assessment indicates that the burden of cardiovascular risk factors is high in this middle-income population, with only 1% of the adult population having an optimal cardiovascular risk profile, the coexistence of multiple cardiovascular risk factors, also in the young, and the suboptimal levels of control. An increase in CVD of epidemic proportions may be anticipated, unless drastic preventive measures are taken.

The prevalence of cardiovascular risk factors in Suriname has nearly doubled over the past 35 years.2,3,4 It is important to bear in mind that the HELISUR study was conducted before the economic crisis in 2015.5 Due to excessive inflation rates, the availability and affordability of medicines has decreased in Suriname. Furthermore, by definition, a person from a LMIC has only a limited amount of money available that can be spent on health care, such as medicines.6 If medicine prices exceed the budget, persons may discontinue their treatment regime. Also, a study in Suriname found that the most im- portant determinant of adequate food intake was the level of income.7 A limited budget, therefore, threatens not only pharmacological treatment, but also the possibility to live a healthy lifestyle. We hypothesize that the current cardiovascular health status of the urban population might be even worse.

The high hypertension burden in Suriname is further illustrated by the comparison with Surinamese living in a high-income country (HIC). We found a higher prevalence of hypertension in women living in Suriname compared to their counterparts living in the Netherlands. This is consistent with earlier studies, observing a higher prevalence of hypertension in LMIC than HIC.8,9,10 However, these studies were conducted among het- erogeneous populations in terms of ethnicity. Studies among more homogenous groups that migrated from a LMIC to a HIC show different results with regards to hypertension prevalence. For example, a recent study among African migrants (Ghanaians) living in Western Europe and non-migrants living in rural and urban Ghana found a clear gradi- 9 ent in the prevalence of hypertension.11 Hypertension prevalence increased from rural through urban Ghana to Europe.11 In our comparison between Surinamese living in a LMIC versus HIC, we found that this gradient of hypertension from LMIC to HIC does not seem to apply. In fact, it was the opposite for some groups, particularly women. If major 152 Chapter 9

efforts are not made to prevent rapid increases in hypertension in LMIC, the current gradient seen in many populations may disappear, as already observed in Suriname.

Comparing the prevalence of 4 major cardiovascular risk factors with other stud- ies conducted in LMIC shows that our prevalence was generally within the reported range.8,9,12,13,14,15 Recently, the Suriname Health Study, based on the WHO STEPwise approach, was conducted in Suriname.16 Overall, a lower prevalence was reported of hypertension, diabetes (fasting plasma glucose criteria), and general obesity of respec- tively 26%, 13%, and 26%.17,18,19 Differences in the study design and methods may explain the different prevalences found. First, rural and semirural areas were included in the Suriname Health Study. In those areas, the adoption of a western lifestyle may be less pronounced, resulting in a lower prevalence of cardiovascular risk factors.20,21,22 This is also reflected in the findings of the Suriname Health study. For example, persons living in the rural interior were significantly less likely to have obesity and diabetes compared to those living in urban areas.18,23 The lower prevalence of cardiovascular risk factors might be further explained by the fact that participants of the Suriname Health Study were on average younger, possibly due to the lower age inclusion criteria (15-65 years compared to 18-70 years in HELISUR).16 Participants of the Suriname Health Study had a median age of 35 years, whereas in HELISUR the median age was 44 years.17,24 Also, a different distribution of ethnic groups might have accounted for a varying preva- lence of cardiovascular risk factors. The HELISUR study included more South-Asians and Creole but less Maroons than the Suriname Health Study. Although these differences in ethnic distribution adequately reflect the population of interest (i.e. relatively more South-Asians and Creole and less Maroons live in the capital compared to the rural interior)25, it may contribute to the overall higher prevalence of risk factors seen in the HELISUR study, as South-Asians and Creole generally had a higher cardiovascular risk factor burden than Maroons. Finally, BP measurements were repeated three times in the Suriname Health Study instead of two times in the HELISUR study. This may added to the lower prevalence of hypertension in the Suriname Health Study compared to HELISUR.26

Cardiovascular risk varies between ethnic groups Our data highlight differences in cardiovascular risk between ethnic groups. Individuals of Asian ancestry had substantially more often diabetes and dyslipidemia than those of African ancestry. Furthermore, they had a higher PWV than individuals of African ancestry, even when adjusted for (or in the absence of) cardiovascular risk factors. For hypertension, the difference between ethnic groups was less evident. Crude prevalence of hypertension was substantially higher in Creole than in South-Asians and, remark- ably, than in Maroons. However, Maroons were also substantially younger and the Discussion 153 difference in hypertension between Creole and Maroons disappeared after adjustment for age. Comparing the pooled African subgroup with the pooled Asian subgroup, those of African ancestry were more often hypertensive after adjustment for confounding variables.

Ethnic disparities in cardiovascular risk factors parallel the ethnic differences seen in established CVD, with hypertension-related organ damage and stroke being more prevalent among African subgroups and coronary heart disease being more prevalent among Asian subgroups.27,28,29,30

Ethnic differences in cardiovascular risk are not easily explained. Nevertheless, several factors have been suggested to act as contributors, which can largely be classified as biological contributors, non-biological contributors and (epi)genetic factors.

Biological contributors: Asian subgroups exhibit a greater tendency towards visceral fat deposition than African or other subgroups. This is in turn strongly associated with a range of metabolic disturbances, such as impaired glucose tolerance and dyslipid- emia.32 In the current thesis, we found that at similar or lower BMI, Asians had greater waist measures and were more often abdominally obese than Africans. This is in line with the international literature and supports the adoption of lower waist circumfer- ence thresholds for Asian populations compared to other ethnic populations.33,34,35 In addition, populations of Asian ancestry have a higher degree of insulin resistance and lower β-cell insulin secretion, independent of adiposity, which might explain the high prevalence of diabetes.36 In African ancestry populations, greater salt sensitivity37, ab- normalities in salt transport38, blunted nocturnal dipping37, and enhanced vasoconstric- tion39 have been proposed as contributors to the higher hypertension burden.

Non-biological contributors: Besides biological variations between ethnic groups, health behaviour and non-biological contributors, such as the social environment and ethnic discrimination, are major determinants for ethnic differentials in cardiovascular health. Since mid-20th century, cardiovascular risk factors have escalated worldwide, with clear differences in rural and urban areas, suggesting cardiovascular risk factors are a result of industrialization and determined by lifestyle choices.27 Furthermore, widely varying prevalence of cardiovascular risk factors were seen within populations of the same 9 ancestry living in different environments.11,40 Also, in our study, we found differences in hypertension between individuals of African descent living in urban Suriname com- pared to those living in urban areas in the Netherlands, which point towards important and modifiable components of risk. Similarly, in the Suriname Health Study in Suriname, the highest and lowest prevalence of hypertension was found in respectively Creole 154 Chapter 9

and Maroons, which are both considered persons of African ancestry.17 This suggests a significant impact of environment and/or lifestyle on the development of cardiovascu- lar risk.

The majority of the studies on ethnic differences are conducted in HIC, in which other ethnic groups are predominantly compared to the Caucasian population as reference group. Differences in socio-economic status and perceived ethnic discrimination may obscure the relationship between ethnicity and cardiovascular risk factors. For example, African populations with lower socioeconomic status living in HIC, such as the United States, develop hypertension more frequent than is anticipated based on their anthropometric and measurable socioeconomic risk factors.41 This demonstrates the importance of the social environment in the assessment of (hypertension) risk. In Suriname, however, it could be speculated that the differences in socioeconomic status and ethnic discrimination between Surinamese of African and Asian ancestry are less pronounced. Yet, ethnic differences in cardiovascular risk were found.

(Epi)genetics: To date, genetic studies do not support the hypothesis of excess heritable risk among Asian or African subgroups compared to Caucasian subgroups, and the gene variants that have been associated with BP and insulin resistance have similar effects in all ethnic groups.42,43 However, epigenetic variation in response to environmental exposures can influence the phenotype, but also depends on the genotype.44 This might contribute to inter-individual variability in gene expression, offering an attractive explanation for ethnic differences in cardiovascular risk. In the current study, however, information on (epi)genetic factors was not available.

Placing our results in the context of the existing knowledge, we cannot state with con- fidence that African and Asian subgroups in Suriname are either more susceptible or more exposed to factors that influence the development of cardiovascular risk factors. Most likely, it is a combination of both that drives ethnic differences in cardiovascular risk. However, the extent to which each of the factors (i.e. biological, non-biological, (epi)genetic) contribute to ethnic differences remains unclear and might vary for each cardiovascular risk factor.

Association between physical activity and obesity differs by ethnicity Although physical activity should be promoted in the total Surinamese population in order to reduce obesity, we found ethnic differences in the association between obesity and physical activity characteristics. Leisure time activity and the overall duration of activity were more important in Africans than in Asians in the relation with obesity. Previous studies demonstrated ethnic differences in physical activity in comparison Discussion 155 with Caucasian populations.45,46,47,48,49 For example, compared to Caucasians, Asians need to engage in higher physical activity levels to have the same cardio-metabolic risk profile, while Africans need to engage in longer periods of light intensity activity to improve their body composition.45,46 Furthermore, it has become evident that there is a need for specific physical activity recommendations to tackle different CV risk fac- tors.50,51 Taken together, our results underscore the need for ethnic- and cardiovascular risk factor-specific recommendations for physical activity.

Hypertension awareness, treatment, control are suboptimal in urban Suriname With many effective and inexpensive treatments available, hypertension control and prevention of subsequent hypertension-related diseases should be achievable.26 Yet, awareness, treatment, and control levels of hypertension were suboptimal in urban Suriname (respectively, 68%, 56%, and 20%). Women were more often aware, treated, and controlled for hypertension than men. This is consistent with previous studies and might be related to a higher health-seeking behaviour.52,53,54 However, comparing women from Suriname to Surinamese women living in the Netherlands showed that women living in Suriname were less aware and, despite higher treatment levels, still less controlled for hypertension than their counterparts living in the Netherlands. This was in contrast to men living in Suriname, who had similar awareness, treatment, and control levels as Surinamese men in the Netherlands. Reasons behind this pattern re- main unclear, but might relate to gender inequalities that women in LMIC are still facing to date.55 Gender inequalities in education, health literacy, income, or employment may limit the ability of women from LMIC to protect their health, for example through a decreased awareness on how to maintain a good (cardiovascular) health or an inability to afford (antihypertensive) medication.55,56,57,58 Further studies are needed to guide interventions aimed at improving these patterns.

No ethnic differences in the awareness, treatment, and control for hypertension were found. However, this result should be viewed with caution, as the ethnic differences in hypertension awareness, treatment, and control were not the primary outcomes of the study and hampered by limited sample sizes. 9 Methodological considerations

Each chapter in this thesis provides the limitations relevant to that specific study. In this section, we discuss several methodological considerations of general importance. 156 Chapter 9

Bias Bias can be defined as a systematic difference between the results of a study and the true state of affairs.59 It can create a spurious association (i.e. an overestimation of an effect) or mask a real one (i.e. an underestimation of an effect). In health studies, bias can broadly be categorized as either selection bias or information bias.59

Selection bias Selection bias occurs when participants included in the study are not representative of the population to which the results will be applied, e.g. persons who agreed to participate in a study may differ from those who did not agree to participate.59 As it was not feasible to use the municipality register and randomly select persons, we had to adapt the selection procedure to the local situation. Therefore, we used randomly selected enumeration areas, assigned trained interviewers a starting point and a di- rection of walking using a door-to-door approach. Data collected with a door-to-door approach avoids selection bias affecting hospital data in LMIC where universal health coverage is uncommon.60 Furthermore, to reduce selection bias, the interviewers had to visit every household up to a maximum of three times until all household members between the ages of 18-70 years were included or stated a refusal. To warrant the cross- sectional nature of the study, we updated the questionnaires at the day of the physical examination. To minimize no-shows, we made appointments through cell phones on short notice, in combination with text message prompting. Response rates may be considered adequate: 72% for the questionnaires and 78% for the physical examina- tion. Unfortunately, we do not have information on differences between responders and non-responders at this stage. However, we can use the Census data of Paramaribo (i.e. the population of interest) to evaluate whether our sample is representative for the population of Paramaribo (Table 1).25 This comparison reveals that participants of the HELISUR study were more likely to be women, which is commonly observed in surveys, and more likely to be South-Asian and less likely to be Creole, compared to the general population of Paramaribo. This selection towards more women and South-Asians may have led to an overestimation of crude prevalence estimates of those risk factors that were associated with female sex and South Asian ancestry, such as obesity and diabetes mellitus. This may affect the generalisability of our findings to the urban population. Subjecting the data to a weighing procedure would enable making inferences to the entire population, but was not done in the present study. In the Suriname Health Study, data were subjected to a weighting procedure, transforming the dataset into a nation- ally representative sample.16 The prevalence of obesity and diabetes mellitus in the urban areas was lower in this study compared to our data.18,24 Discussion 157

Table 1. Comparison of sex and ethnic distribution between the population of interest and the study population Census 2012 – Population of HELISUR – Questionnaire and Paramaribo HELISUR – Questionnaire physical examination n=240,924 n=1,800 n=1,157 Sex Men 49% 38% 37% Women 51% 62% 63% Ethnicity Creole 28% 22% 21% Maroon 17% 22% 19% South-Asian 25% 31% 34% Indonesian 11% 8% 9% Other/mix 20% 17% 17% Footnote Table 1. Values are proportions. Comparisons are made between the sex and ethnic dis- tribution of the population of Paramaribo25 and that of the HELISUR study population.

When we compared data from the participants who were interviewed but not examined with the participants who were examined, we found no differences in the distribution of sex. However, participants who were interviewed but not examined were more often Maroon (179 [29%] vs. 216 [19%], p<0.01), less often South Asian (152 [24%] vs. 393 [34%], p<0.01), and more likely to be younger (35 vs. 42 years, p<0.01).

Although the vast majority of the Surinamese (73%) lives in an urban area,25 it is im- portant to keep in mind that the urban population may not be representative for the entire country. Urban residents had a worse cardiovascular risk profile than their rural counterparts, regarding diabetes and obesity.18,23 Therefore, we emphasize that the data of the HELISUR study cannot be extrapolated to the entire population of Suriname and only with caution to the population of Paramaribo.

Information bias Information bias is a bias that arises in a study because of misclassification of the exposure or outcome measurements.59 For example, the presence of hypertension and other cardiovascular risk factors was assessed through measurements within one visit, while repeated tests are required for confirmation of the diagnosis.26 The prevalence 9 of cardiovascular risk factors might therefore be an overestimation. Nevertheless, with epidemiological studies this is a common approach.

Information on several confounding variables (e.g. educational level, medication use, and physical activity) was self-reported and recall bias might have occurred. Persons 158 Chapter 9

who forgot to mention their medication and had normal blood pressures were classified as normotensives instead of controlled hypertensives. The same applied to antihyper- glycemic and antihyperlipidemic medication and the respective proportions of diabet- ics and persons with dyslipidemia. A type of bias that can arise with self-reported data is social desirability bias, in which participants give answers in the direction they perceive are of interest to the researcher or under-report socially unacceptable behaviours.59 This could be relevant for the assessment of tobacco smoking or physical activity, which may have resulted in a respective under- and overestimation of the proportion of smok- ers and physically active persons. Nevertheless, as these misclassifications applied to all groups, it is unlikely that these limit the comparability between the ethnic groups.62

Confounding Confounding occurs when we find a spurious association between a potential risk factor and a disease outcome or miss a real association between them because we have failed to adjust for any confounding variables.59 In this thesis, we dealt with confounding us- ing stratification and adjustment.

Although information on many confounders was available in this study, residual confounding may still be present, as in any observational study. For example, single measurements of confounding variables would be poorly representative of lifetime exposure.63 Furthermore, information on several potential confounding variables was not collected within the HELISUR study. Early life influences, for example in utero or childhood, may play a strong independent role in determining adult cardiovascular risk, but were not examined.64,65 Also, food intake and dietary salt intake are important determinants of cardiovascular risk factors, but were beyond the scope of this paper. A recently published paper of the Suriname Health Study suggested that only 5% of the Surinamese had an overall healthy food intake (i.e. adequate and not excessive).7 Food intake also differed across ethnic groups with Maroons and Amerindians (mainly from the inlands) consuming less often an adequate food intake but also less often an exces- sive food intake.7 Moreover, next to kitchen salt, Surinamese meals contained mostly products high in salt. Given that the majority of the HELISUR participants consumed more than one hot meal per day66, there is little doubt that the population salt consumption far exceeds the recommended maximum of 5 g per person per day.67 As food and salt intake are important determinants of hypertension and a potential confounder in the association between ethnicity and hypertension, this may be a limitation of our study. Overall, residual confounding may have resulted in an overestimation of the estimates. Discussion 159

Note on ethnicity In this thesis, self-reported ethnicity was used. Although self-reported ethnicity is considered the best variable, we recognize that these categories are arbitrary and that heterogeneity exists within each ethnic group.68,69

Not all analyses in this thesis were stratified for the 4 major ethnic groups. The reason for this was that the analyses were hampered by limited sample sizes and thus would have reduced power to detect a significant effect. Especially, the Indonesian subgroup was small (n=105; 9%), although this accurately reflects their relative small repre- sentation within the urban Surinamese population (11%, Table 1). The pooling of the ethnic groups may have led to an underestimation of the estimated difference between African and Asian ancestry populations. For example, the unadjusted prevalence of hypertension was 10% lower in Maroons compared to Creole (38% vs 48%). The same holds true to a lesser extent for the South-Asian and Indonesian ancestry subgroups. Combining the two ethnic groups might have diluted the differences between African and Asian ancestry groups. We acknowledge that this is a limitation for the analyses presented in the chapters.

Implications and future directions

Based on the findings of the current thesis and the available literature, several recom- mendations for research and policy can be made to achieve a Healthier Life in Suriname.

There is need for population-based interventions The main focus of health care for CVD in many LMIC is secondary prevention.70 Patients enter treatment programs after becoming symptomatic, when costly high-technology interventions are needed. However, primary CVD prevention is relatively cheap and has a much greater return (e.g. less premature deaths, reduced economic losses).71 The World Health Organization has identified a set of “best buy interventions” that mainly focuses on improvements in salt intake, physical activity, tobacco use, and alcohol consumption.71,72 These population-based interventions have proven to be very cost- effective and are appropriate to implement in LMIC such as Suriname, where the health care infrastructure is less developed.71,72 In Suriname, a first step in the right direction 9 is made by introducing the Health in All Policies (HiAP) approach in 2017.73 The HiAP is a collaborative approach across all levels and sectors of government to improve health, for example, through nutritional labeling on foods high in fat, sugar, and salt or spatial planning to promote physical activity.73 Although the formulation of these policies is 160 Chapter 9

promising, the effective implementation and monitoring of these policies might still be a challenge, especially with the limited resources available in Suriname.

There is a need for intensified screening for and control of hypertension With hypertension being responsible for almost half of the CVD deaths, interventions aimed at lowering BP or improving awareness, treatment, and control are of great im- portance.70 In urban Suriname, hypertension was highly prevalent and started already at a young age (25% of people below the age of 40 years had hypertension). In addition, effective and low-cost drug therapy is globally available. Yet, the levels of awareness and control are disappointing. Routine opportunistic screening for hypertension during regular health care visits provides a simple but reliable way to increase hypertension awareness in LMIC.74,75 This screening for high BP should start already early in life. Fur- thermore, focusing on education can improve hypertension prevalence, awareness, and control levels. However, it is important to recognize culturally and ethnically different perspectives on hypertension and treatment thereof.76 To date, such information is lack- ing in Suriname, and, therefore, future studies should focus on how cultural and ethnic factors influence management of hypertension in Suriname.

General and abdominal obesity were the most important determinants of hypertension. This highlights the pivotal role of obesity in the hypertension pandemic and warrants the implementation of interventions aimed at (abdominal) obesity prevention. Our results on the association between physical activity and obesity underscore the impor- tance of physical activity in the prevention of obesity. In addition, we found that waist measures were the most suitable body composition measure to assess cardiovascular risk and therefore advocate the use of a waist measure in clinical practice to assess who is at increased risk for CVD.

There is a need for ethnic-specific research and prevention programs Although interventions aimed at early BP screening and obesity prevention should target the population as a whole, a more ethnic-specific approach might be warranted in the prevention of diabetes and dyslipidemia. Persons of Asian ancestry were twice as likely to have diabetes and dyslipidemia compared to those of African ancestry, sug- gesting that the “one size fits all” approach may not be appropriate for these specific risk factors. Therefore, routine opportunistic screening for diabetes and dyslipidemia may be useful in Surinamese of Asian ancestry.

Given the ethnic differences in cardiovascular risk, it may be useful to register ethnicity in the routine health care data registry. This may be culturally charged due to potential risks (e.g. discrimination or stigmatisation), yet, the appropriate use of ethnicity may Discussion 161 provide valuable information on differences in cardiovascular health of subgroups that warrant further investigation and intervention.77

Because we did find differences between the 4 ethnic groups, future research in Suri- name should study these ethnic groups separately with a large enough sample size so that the underlying causes of ethnic inequalities in health can be explored. This would also enable the identification of cut-off values specific for the ethnic groups living in Suriname instead of using those validated in white populations or in ethnic groups from other countries.

There is a need for prospective data and ongoing public health surveillance A cross-sectional study design is a correct way to assess the current health status of a population, especially in a low-resource setting.62 However, the cross-sectional nature of our study implies that causal associations cannot be established. Data presented in this thesis would ideally be the baseline data for a longitudinal study, in which these baseline measurements are repeated during follow-up studies to enable longitudinal analyses on the relationship between ethnicity and incident CVD. For example, continu- ation of the HELISUR study could elucidate whether Asians with an increased arterial stiffness but without hypertension have a higher risk of developing CVD. Also, it would be valuable to relate body composition measures to hard cardiovascular outcomes instead of CVD risk scores and surrogate intermediate endpoints.

Cardiovascular risk profiles change overtime, in a negative way (e.g. due to the epide- miological transition or an economic crisis) or in a positive way (e.g. due to implementa- tion of effective interventions). To follow trends in cardiovascular risk, it is important to introduce public health surveillance in Suriname, in which a cross-sectional survey is repeated every 5 years.78 This will also enable the evaluation and adaptation of public health interventions.

General conclusion

Cardiovascular disease is rapidly rising in many low- and middle-income countries. Yet, the lack of data on cardiovascular risk factors hampers the design of effective preven- 9 tion strategies. The current thesis describes the status quo of cardiovascular risk factors and early organ damage in the middle-income population of urban Suriname. The cardiovascular risk factor burden in urban Suriname is alarmingly high, with only 1% of the adult population having an optimal cardiovascular risk profile, the coexistence of multiple cardiovascular risk factors, also in the young, and the suboptimal levels of 162 Chapter 9

control. An increase in CVD of epidemic proportions may be anticipated, unless drastic preventive measures are taken. Preventive strategies at the population level, such as screening for hypertension and obesity prevention, together with ethnic-specific approaches for diabetes and dyslipidemia in Asian ancestry populations, should be implemented in order to protect the cardiovascular health of urban Surinamese. Discussion 163

References

1. United Nations General Assembly. Political declaration of the high-level meeting of the General Assembly on the prevention and control of non-communicable diseases. 2011. Available at: http://www.who.int/nmh/events/un_ncd_summit2011/political_declara- tion_en.pdf. 2. Schaad J. De prevalentie van diabetes melltius bij drie bevolkingsgroepen in bepaalde plat- telandsgebieden van Suriname. Paramaribo: Proefschrift Universiteit van Suriname; 1979. 3. Schaad J, Oehlers G. Onderzoek naar het voorkomen van coronaire hartziekten en hun risicofactoren bij drie bevolkingsgroepen in 1975-1976 op het platteland van Suriname. Paramaribo. 4. Schaad J, Terpstra J, Oemrawsingh I, Nieuwenhuizen Kruseman A, Bouwhuis-Hoogenderf M. Diabetes prevalence in the three main ethnic groups in Surinam (South America): a popula- tion survey. Netherlands Journal of Medicine. 1985;28:17-22. 5. MacDonald S. Suriname’s Economic Crisis. Center for Strategic and International Studies. April 20, 2017. Available at: https://www.csis.org/analysis/surinames-economic-crisis. 6. Niëns LBW. Measuring the affordability of medicines: Importance and challenges. Health Policy. 2013;112(1-2):45-52. 7. Nahar-van Venrooij L, Marhe E, Antonius- Smits C, Krishnadath I. Adequate and excessive food consumption in Suriname, a multi-ethnic middle-income country. International Jour- nal of Public Health (in revision). 2018. 8. Sarki A, Nduka C, Stranges S, Kandala NB, Uthman O. Prevalence of hypertension in low- and middle-income countries: a systematic review and meta-analysis. Medicine. 2015;94(50):e1959. 9. Chow C, Teo K, Rangarajan S, et al. Prevalence, awareness, treatment, and control of hyper- tension in rural and urban communities in high-, middle-, and low-income countries. JAMA. 2013;310(9):959-968. 10. Mills K, Bundy J, Kelly T, et al. Global Disparities of Hypertension Prevalence and Con- trol: A Systematic Analysis of Population-Based Studies From 90 Countries. Circulation. 2016;134(6):441-450. 11. Agyemang C, Nyaaba G, Beune E, et al. Variations in hypertension awareness, treatment, and control among Ghanaian migrants living in Amsterdam, Berlin, London, and nonmigrant Ghanaians living in rural and urban Ghana - the RODAM study. J Hypertens. 2018;36(1):169- 177. 12. Dagenais G, Gerstein H, Zhang X, et al. Variations in diabetes prevalence in low-, middle-, and high-income countries: results from the Prospective Urban and Rural Epidemiological study. Diabetes Care. 2016;39(5):780-787. 13. Salas A, Acosta D, Ferri C, et al. The prevalence, correlates, detection and control of diabetes among older people in low and middle income countries. A 10/66 Dementia Research Group Population-Based Survey. PloS One. 2016;11(2):e0149616. 9 14. Global Health Observatory (GHO) data repository. Raised total cholesterol. Geneva: WHO; 2008. 15. Rosengren A, Teo K, Rangarajan S, et al. Psychosocial factors and obesity in 17 high-, middle- and low-income countries: the Prospective Urban Rural Epidemiologic study. International Journal of Obesity. 2015;39(8):1217–1223. 164 Chapter 9

16. Krishnadath I, Smits, CC, Jaddoe V, Hofman A, Toelsie J. A national surveillance study on non-communicable disease risk factors: Suriname Health Study protocol. JMIR Research Protocols. 2015;4(2):e75. 17. Krishnadath I, Jaddoe V, Nahar-van Venrooij L, Toelsie J. Ethnic differences in prevalence and risk factors for hypertension in the Suriname Health Study: a cross sectional population study. Population Health Metrics. 2016;14(33). 18. Krishnadath I, Nahar-van Venrooij L, Jaddoe V, Toelsie J. Ethnic differences in pre- diabetes and diabetes in the Suriname Health Study. BMJ Open Diabetes Research & Care. 2016;4:e000186. 19. Krishnadath I, Toelsie J, Hofman A, Jaddoe V. Ethnic disparities in the prevalence of meta- bolic syndrome and its risk factors in the Suriname Health Study: a cross-sectional popula- tion study. BMJ Open. 2016;6:e013183. 20. Htet A, Bjertness M, Sherpa L, et al. Urban-rural differences in the prevalence of non- communicable diseases risk factors among 25–74 years old citizens in Yangon Region, Myanmar: a cross sectional study. BMC Public Health. 2016;16:1225. 21. Houehanou Y, Lacroix P, Mizehoun G, Preux PM, Marin B, Houinato D. Magnitude of cardio- vascular risk factors in rural and urban areas of Benin: findings from a nationwide STEPS survey. PloS One. 2015;10(5):e0126441. 22. Miranda J, Gilman R, Smeeth L. Differences in cardiovascular risk factors in rural, urban and rural-to-urban migrants in Peru. Heart. 2011;97(10):787-796. 23. Nahar-Van Venrooij L, Baldew S, Krishnadath I. Prevalence and risk factors of obesity in the Suriname Health Study. Paper in progress. 2018. 24. Krishnadath I. Chronic disease risk factor surveillance: databook for Suriname. Paramaribo: Department of Public Health, Anton de Kom University of Suriname; 2017. 25. General Bureau of Statistics Suriname. Resultaten van de achtste (8e) volks- en woningtelling in Suriname (Volume I) - Demografische en sociale karakteristieken en migratie. Paramaribo 2012. 26. Mancia G, Faqard R, Narkiewicz K, et al. 2013 ESH/ESC Guidelines for the management of arterial hypertension: the Task Force for the management of arterial hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens. 2013;31(7):1281-1357. 27. Tillin T, Hughes A, Mayet J, et al. The relationship between metabolic risk factors and incident cardiovascular disease in Europeans, South Asians, and African Caribbeans: SABRE (Southall and Brent Revisited) -- a prospective population-based study. Journal of the American Col- lege of Cardiology. 2013;61(17):1777-1786. 28. Park C, March K, Ghosh A, et al. Left-ventricular structure in the Southall And Brent REvisited (SABRE) study: explaining ethnic differences. Hypertension. 2013;61(5):1014-1020. 29. Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2095-2128. 30. Chaturvedi N. Ethnic differences in cardiovascular disease. Heart. 2003;89(6):681-686. 31. Cappuccio F, Oakeshott P, Strazzullo P, Kerry S. Application of Framingham risk estimates to ethnic minorities in United Kingdom and implications for primary prevention of heart disease in general practice: cross sectional population based study. BMJ Open. 2002;325(7375):1271. Discussion 165

32. Van Gaal L, Mertens I, De Block C. Mechanisms linking obesity with cardiovascular disease. NAture. 2006;444(7121):875-880. 33. Alberti K, Zimmet P, Shaw J. International Diabetes Federation: a consensus on Type 2 diabetes prevention. Diabetes Medicine. 2007;24(5):451-463. 34. National Institute for Health and Care Excellence. Assessing body mass index and waist circumference thresholds for intervening to prevent ill health and premature death among adults from black, Asian and other minority ethnic groups in the UK. 2013. Available at: https://www.nice.org.uk/guidance/ph46. 35. World Health Organization. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363:157-163. 36. Dagogo-Jack S. Ethnic disparities in type 2 diabetes: pathophysiology and implications for prevention and management. Journal of the National Medical Assocation. 2003;95(774):779- 789. 37. Ortega L, Sedki E, Nayer A. Hypertension in the African American population: a succinct look at its epidemiology, pathogenesis, and therapy. Nefrologia. 2015;35(2):139-145. 38. Worthington M, Wendt M, Opie L. Sodium transport in hypertension: assessment of membrane-associated defects in South African black and white hypertensives. Journal of Human Hypertension. 1993;7(3):291-297. 39. Brewster L, Clark J, van Montfrans G. Is greater tissue activity of creatine kinase the genetic factor increasing hypertension risk in black people of sub-Saharan African descent? Journal of Hypertension. 2000;18:1537-1544. 40. Cooper R, Rotimi C, Ataman S, et al. The prevalence of hypertension in seven populations of west African origin. American Journal of Public Health. 1997;87(2):160-168. 41. Cooper R, Forrester T, Plange-Rhule J, et al. Elevated hypertension risk for African-origin populations in biracial societies: modeling the epidemiologic transition study. Journal of Hypertension. 2015;33(3):473-481. 42. Kooner J, Saleheen D, Sim X, et al. Genome-wide association study in individuals of South Asian ancestry identifies six new type 2 diabetes susceptibility loci. Nature Genetics. 2011;43(10):984-989. 43. Franceschini N, Fox E, Zhang Z, et al. Genome-wide association analysis of blood-pressure traits in African-ancestry individuals reveals common associated genes in African and non- African populations. American Journal of Human Genetics. 2013;93(3):545-554. 44. Relton C, Davey Smith G. Epigenetic epidemiology of common complex disease: prospects for prediction, prevention, and treatment. PloS One. 2010;7(10):e1000356. 45. Celis-Morales C, Ghouri N, Bailey M, Sattar N, Gill J. Should physical activity recommen- dations be ethnicity-specific? Evidence from a cross-sectional study of South Asian and European men. PlOS One. 2013;8(12):e82568. 46. Gill J, Celis-Morales C, Ghouri N. Physical activity, ethnicity and cardio-metabolic health: does one size fits all? Atherosclerosis. 2014;232(2):319-333. 47. Kwon S, Wang M, Hawkins M. Association between self-reported physical activity and obe- 9 sity among White, Black, Hispanic, and Asian Americans: 2007 and 2009 Brfss. Ethnicity and Disease. 2013;23(2):129-135. 48. Misra A, Nigam P, Hills A, et al. Consensus physical activity guidelines for Asian Indians. Diabetes Technology & Therapeutics. 2012;14(1):83-98. 166 Chapter 9

49. White J, Jago R. Prospective associations between physical activity and obesity among adolescent girls: racial differences and implications for prevention. Archives of Pediatrics & Adolescent Medicine. 2012;166(6):522-527. 50. Hansen D, Dendale P, Coninx K, et al. The European Association of Preventive Cardiology Ex- ercise Prescription in Everyday Practice and Rehabilitative Training (EXPERT) tool: A digital training and decision support system for optimized exercise prescription in cardiovascular disease. European Journal of Preventive Cardiology. 2017;24(10):1017-1031. 51. Vanhees L, Geladas N, Hansen D, et al. Importance of characteristics and modalities of physi- cal activity and exercise in the management of cardiovascular health in individuals with cardiovascular risk factors: recommendations from the EACPR. Part II. European Journal of Preventive Cardiology. 2012;19(5):1005-1033. 52. Irazola V, Gutierrez L, Bloomfield G, et al. Hypertension prevalence, awareness, treatment, and control in selected communities of nine low- and middle income countries: results from the NHLBI/UHG Network of Centers of Excellence for Chronic Diseases. Global Heart. 2016;11(1):47-59. 53. Patel S, Ali M, Alam D, et al. Obesity and its relation with diabetes and hypertension: a cross-sectional study across four low- and middle-income country regions. Global Heart. 2016;11(1):71-79 e4. 54. Pinto E. Blood pressure and ageing. Postgrad Med J. 2007;83:109–114. 55. Filippi V, Ronsmans C, Campbell O, et al. Maternal health in poor countries: the broader context and a call for action. The Lancet. 2006;368(9546):1535–1541. 56. Diemer F, Haan Y, Nannan Panday R, van Montfrans G, Oehlers G, Brewster L. Health literacy in Suriname. Social Work in Health Care. 2017;56(4):283-293. 57. Dégano I, Marrugat L, Grau M, et al. The association between education and cardiovascular disease incidence is mediated by hypertension, diabetes, and body mass index. Scientific Reports. 2017;7:12370. 58. Wang YCJ, Wang K, Edwards C. Education as an important risk factor for the prevalence of hypertension and elevated blood pressure in Chinese men and women. Journal of Human Hypertension. 2006;20:898-900. 59. Petrie A, Sabin C. Medical statistics at a glance. 3rd Edition ed. Oxford: John Wiley & Sons; 2009. 60. Modesti P, Agostoni P, Agyemang C, et al. Cardiovascular risk assessment in low-resource settings: a consensus document of the European Society of Hypertension Working Group on Hypertension and Cardiovascular Risk in Low Resource Settings. Journal of Hypertension. 2014;32(5):951-960. 61. Snijder M, Stronks K, Agyemang C, Busschers W, Peters R, van den Born BJ. Ethnic differences in arterial stiffness the Helius study. Int J Cardiol. 2015;191:28-33. 62. Gordis L. Epidemiology. 4th ed. Philadelphia: Saunders Elsevier; 2009. 63. Tirosh A, Shai I, Afek A, et al. Adolescent BMI trajectory and risk of diabetes versus coronary disease. New England Journal of Medicine. 2011;364(14):1315-1325. 64. Gluckman P, Hanson M, Cooper C, Thornburg K. Effect of in utero and early-life conditions on adult health and disease. New England Journal of Medicine. 2008;359:61-73. 65. Bhargava S, Sachdev H, Fall C, et al. Relation of serial changes in childhood body-mass index to impaired glucose tolerance in young adulthood. New England Medical Journal. 2004;350:865-875. Discussion 167

66. Nahar-van Venrooij L. An analysis of the nutrition situation in Suriname: identification of those problems important for the development of Surinamese Food-Based Dietary Guidelines. Paramaribo: PAHO; 2016. 67. World Health Organization. Guideline: sodium intake for adults and children. Geneva: WHO; 2012. 68. Higgins P, Taylor E. Measuring racial and ethnic disparities in health care: efforts to improve data collection. Princeton, NJ: Mathematica Policy Research; 2009. 69. Institute of Medicine. Chapter 5: Improving data collection across the health care system. Race, ethnicity, and language data: standardization for health care quality improvement. Washington DC: The National Academies Press; 2009. 70. World Health Organization. Global atlas on cardiovascular disease prevention and control. Geneva: WHO; 2011. 71. World Health Organization. From burden to “best buys”: reducing the economic impact of non-communicable diseases in low- and middle-income countries. Geneva: World Economic Forum; 2011. 72. World Health Organization. Global action plan for the prevention of noncommunicable diseases 2013-2020. Geneva: WHO; 2013. 73. Panamerican Health Organization (PAHO). The Health in All Policies package 2017-2018: intersectoral collaboration for health and development. Paramaribo: PAHO; 2017. 74. Maurer J, Ramos A. One-year routine opportunistic screening for hypertension in formal medical settings and potential improvements in hypertension awareness among older per- sons in developing countries: evidence from the Study on Global Ageing and Adult Health (SAGE). American Journal of Epidemiology. 2015;181(3):180-184. 75. Nerenberg K, Zarnke K, Leung A, et al. Hypertension Canada’s 2018 guidelines for diagnosis, risk assessment, prevention, and treatment of hypertension in adults and children. Cana- dian Journal of Cardiology. 2018;34(5):506-525. 76. Beune E, Haafkens J, Agyemang C, Schuster J, Willems D. How Ghanaian, African-Surinamese and Dutch patients perceive and manage antihypertensive drug treatment: a qualitative study. Journal of Hypertension. 2008;26(4):648-656. 77. Ellison G, Smart A, Tutton R, Outram S, Ashcroft R, Martin P. Racial categories in medicine: a failure of evidence-based practice? PLoS Medicine. 2007;4(9):1343-1346. 78. World Health Organization. STEPwise approach to surveillance (STEPS). Geneva: WHO; 2014.

9

APPENDICES Summary Nederlandse samenvatting List of publications Contributing authors Author contributions Portfolio Dankwoord About the author

APPENDICES 171

Summary

Cardiovascular disease (CVD) is the leading cause of death worldwide, but the vast ma- jority of CVD-related deaths occur in low- and middle-income countries (LMIC). While a declining trend is seen in high-income countries (HIC), the burden of CVD continues to rise at an alarming pace in LMIC and drastic preventive measures need to be taken. Unfortunately, in many LMIC the lack of data hampers the design and implementation of effective prevention strategies.

Well-established differences in cardiovascular risk and CVD exist between ethnic groups. The reasons behind these ethnic differences are complex and remain incompletely elucidated. Most studies on ethnic differences in cardiovascular risk are conducted in HIC in (descendants of) migrants from African, Asian, and Caribbean countries. They may not be applicable for LMIC where socioeconomic and environmental conditions as well as health care systems and policies are different.

Suriname is a middle-income country in South-America with a multiethnic population of mainly African and Asian ancestry. Despite large similarities in socioeconomic sta- tus between the ethnic groups, substantial ethnic differences exist in cardiovascular mortality. Differences in underlying cardiovascular risk factors and asymptomatic organ damage might explain these ethnic disparities in CVD mortality. Yet, data on cardiovas- cular risk factors and asymptomatic organ damage are virtually absent.

The main aim of this thesis is to strengthen the evidence base for cardiovascular health data in urban Suriname, with a primary focus on ethnic differences in cardiovascular risk. Such insights will assist the development of effective preventive and intervention strategies aimed at reducing the high CVD burden in Suriname.

To address the main aim of this thesis, we designed the Healthy Life in Suriname (HELI- SUR) study, an observational cross-sectional study on cardiovascular health among different ethnic groups in the capital of Suriname. The primary aim of the study was to assess the cardiovascular risk profile, including cardiovascular risk factors, early organ damage, and established CVD, of the urban population of Suriname. As a secondary outcome, ethnic differences in cardiovascular risk were assessed. In the first part of this thesis, the study protocol of the HELISUR study was described (Chapter 2). In Chapter 3, the feasibility of the HELISUR questionnaire and the physical examination was assessed in a convenience sample. Although adaptations were necessary to optimize the data A quality and quantity of the questionnaire and the physical examination, the findings 172 APPENDICES

showed that large-scale evaluations of cardiovascular risk, including advanced non- invasive hemodynamics, were feasible in this middle-income setting.

The second part of this thesis focused on the overall and ethnic-specific cardiovascular risk profile of this urban Surinamese population. Chapter 4 described the cardiovascu- lar risk profile of the 1,157 participants. We found that more than three-third (77%) of the participants had at least one of the following traditional cardiovascular risk factors: hypertension (40%), diabetes mellitus (22%), dyslipidemia (41%), obesity (37%) or tobacco smoking (18%). An additional 22% was prehypertensive, prediabetic, over- weight or had borderline dyslipidemia, resulting in a mere 1% of the population having an optimal cardiovascular risk factor profile. In those aged 40 years and younger, 62% already had at least one cardiovascular risk factor and 2% had an optimal cardiovas- cular risk factor profile. Aside from traditional cardiovascular risk factors, we assessed early organ damage, which was prevalent in 22% of the population. The majority of these persons had an increased arterial stiffness (77%). Established CVD was found in 5% of the participants, including stroke (3%), coronary heart disease (2%), and severe chronic kidney disease (0.4%). Marked ethnic differences were seen in the prevalence of traditional cardiovascular risk factors. Crude hypertension prevalence was highest in Creole (48%), which significantly differed from that of Maroons (38%) and South-Asians (37%). South-Asians had the highest prevalence of diabetes and dyslipidemia (respec- tively, 30% and 52%), followed by Indonesians (25% and 47%) and Creole (22% and 37%). In Maroons, the prevalence of diabetes and dyslipidemia was relatively low (re- spectively, 11% and 27%). For obesity, no differences were seen between the 4 ethnic groups. After dichotomizing the ethnic groups into an African and an Asian ancestry population, crude hypertension prevalence did not differ between Africans and Asians. However, after adjustment for several confounders, those of African ancestry were more likely to have hypertension compared to their Asian counterparts. Diabetes mellitus and dyslipidema were more prevalent in persons of Asian ancestry than in those of African ancestry. No ethnic differences were seen in general obesity.

Hypertension awareness and treatment were respectively 68% and 56%. Control of hypertension was seen in 36% of those treated for hypertension (20% of all hyperten- sives). A slight majority of the participants with diabetes was aware of their condition (59%) and received treatment (54%), but only 21% of those treated were controlled (11% of all diabetics).

Obesity is an important risk factor for CVD, promoting atherosclerosis. However, it remains unclear what obesity measure is most accurate for identifying persons at high CVD risk. In Chapter 5, we examined whether more complex body composition mea- APPENDICES 173 sures such as fat mass or low fat-free mass together with a high fat mass (i.e. sarcopenic obesity) are superior to the simple anthropometric measures, such as body mass index (BMI) and waist measures, in the association with CVD risk. We found that cardiovascu- lar risk in people of African and Asian ancestry is most strongly associated with waist measures, in particular waist circumference and waist-to-height ratio, independent of sex, age or ethnic group. BMI seemed to have additional value in the total group and in men in the association with increased arterial stiffness. Overall, we found little advan- tage in using body composition measures rather than simple anthropometric measures. The study also demonstrated ethnic differences in body composition, suggesting that Asians generally have greater waist measures and more fat mass than Africans at similar BMI level.

Physical activity is a key element in the management of obesity. We explored the as- sociation between obesity and physical activity within an Asian and African ancestry population in Chapter 6. The results of this study showed that meeting the physical activity recommendations and being active within the commuting and leisure time domain are all inversely associated with obesity in the total population. In the Asian population, no association could be established between physical activity and obesity. In the African population, spending more time on physical activity and performing larger volumes of physical activity was associated with lower obesity prevalence ratios. Our results underscore the need for ethnic-specific recommendations for physical activity.

Evidence suggests that differences in CVD risk between ethnic groups are not fully explained by traditional cardiovascular risk factors. One way to address the issue of ethnic variability in CVD risk is to shift the focus from risk assessment to diagnosing early asymptomatic organ damage, for example in blood vessels. In Chapter 7, we studied ethnic differences in pulse wave velocity (PWV), a type of early organ damage characterized by reduced arterial compliance, taking traditional cardiovascular risk fac- tors into account. The main finding of the study was that persons of Asian ancestry had a substantially higher PWV of approximately 1 m/s compared to those of African ancestry, after adjustment for age and blood pressure. The ethnic difference in PWV persisted after additional adjustment for other cardiovascular risk parameters, including fasting glucose, lipid spectrum, BMI, and waist circumference. It remains largely unknown why a higher PWV is found in Asians, even when adjusted for or in the absence of cardio- vascular risk factors. However, the higher PWV in Asians corresponds to their known increased coronary heart disease risk. A In the third part of this thesis, we compared hypertension prevalence, awareness, treatment, control, and determinants related to hypertension in urban Suriname with 174 APPENDICES

Surinamese living in a HIC (participants of the Health Life in an Urban Setting (HELIUS) study conducted in Amsterdam, the Netherlands). The findings of Chapter 8 showed that hypertension was highly prevalent in men and women living in Suriname, with suboptimal levels of awareness, treatment and control. While in men these estimates were similar to their counterparts living in the Netherlands, in women hypertension prevalence in Suriname was substantially higher and started at an earlier age compared to their counterparts living in the Netherlands. Also, the levels of awareness and control were lower in women in Suriname compared to those in the Netherlands. Underlying behavioural determinants of hypertension in Suriname were similar to those in the Netherlands, with age and waist circumference being associated with hypertension. This suggests that preventive measures should prioritize the early detection of hyper- tension and prevention of (abdominal) obesity.

Finally, Chapter 9 reflected on the main findings, in the light of several methodological considerations. Also, the implications of the findings and recommendations for future research were discussed. Two main conclusions arose from this thesis: 1. Urban Suriname bears a high cardiovascular risk factor burden, with only 1% of the adult population having an optimal cardiovascular risk profile, the coexistence of multiple cardiovascular risk factors, also in the young, and the suboptimal levels of control. An increase in CVD of epidemic proportions may be anticipated, unless drastic preventive measures are taken. Preventive strategies at the population level, such as screening for hypertension and obesity prevention, are urgently needed to reduce the high burden of risk factors and concomitant CVD in Suriname. 2. Ethnic differences in cardiovascular risk exist in urban Suriname. Persons of Asian ancestry had substantially more often diabetes and dyslipidemia than those of Afri- can ancestry. Furthermore, they had a higher PWV than persons of African ancestry, even when adjusted for or in the absence of cardiovascular risk factors. Those of African ancestry were more often hypertensive after adjustment for confounding variables. Ethnic disparities in cardiovascular risk factors parallel the ethnic differ- ences seen in established CVD and warrant the implementation of ethnic-specific preventive strategies. APPENDICES 175

Nederlandse samenvatting

Hart- en vaatziekten (HVZ) zijn wereldwijd doodsoorzaak nummer één. Het overgrote deel van de sterfgevallen door HVZ gebeurt in lage- en middeninkomenslanden (LMIL). Terwijl er in hoge inkomenslanden (HIL) een neerwaartse trend gaande is, is de ziekte- last door HVZ in LMIL juist in een hoog tempo aan het stijgen en moeten er dringend preventieve maatregelen worden genomen. Helaas zorgt het tekort aan gegevens in veel LMIL voor een stagnatie in de ontwikkeling en implementatie van effectieve interventies.

Er zijn duidelijke verschillen in HVZ tussen etnische groepen. De oorzaken van deze etnische verschillen zijn complex en nog onvoldoende belicht. De meeste onderzoeken naar etnische verschillen in cardiovasculair risico gebeuren in HIL bij (afstammelingen van) migranten uit Afrika, Azië en de Caraïben. Deze resultaten hoeven niet direct ex- trapoleerbaar te zijn naar LMIL, waar de socio-economische condities, de omgeving, de gezondheidssystemen en de politiek anders zijn.

Suriname is een middeninkomensland in Zuid-Amerika met een multi-etnische populatie van voornamelijk Afrikaanse en Aziatische afkomst. Ondanks grote overeenkomsten in de sociaal-economische status tussen de etnische groepen, zijn er substantiële etnische verschillen in HVZ sterfte. Verschillen in onderliggende risicofactoren en vroegtijdige orgaanschade kunnen wellicht de etnische verschillen in HVZ sterfte verklaren, maar gegevens waren bij de start van ons onderzoek nauwelijks beschikbaar in Suriname.

Het centrale doel van dit proefschrift is het vergroten van de kennis over het cardiovas- culair risicoprofiel van Surinamers. Daarnaast zal de focus liggen op etnische verschil- len in het cardiovasculair risico. Zulke gegevens zullen bijdragen aan de ontwikkeling van effectieve preventieve en interventie strategieën gericht op het reduceren van de hoge HVZ ziektelast in Suriname.

Om het centrale doel te behalen, hebben we de Healthy Life in Suriname (HELISUR) studie opgericht. Dit is een observationele cross-sectionele studie over de cardio- vasculaire gezondheidsstatus van de verschillende etnische groepen die leven in de hoofdstad van Suriname. Het hoofddoel van deze studie was het vaststellen van het cardiovasculair risicoprofiel, inclusief de risicofactoren, vroegtijdige orgaanschade en HVZ. Daarnaast onderzochten we de etnische verschillen in het cardiovasculair risico. A In het eerste deel van deze thesis worden het HELISUR studieprotocol beschreven (Hoofdstuk 2). In Hoofdstuk 3 wordt de haalbaarheid van de vragenlijst en het licha- 176 APPENDICES

melijk onderzoek getest in een groep vrijwilligers. Ondanks dat er enkele aanpassingen nodig waren om de kwaliteit van de data te optimaliseren, liet dit haalbaarheidsonder- zoek zien dat het mogelijk was om op grote schaal in een middeninkomensland deze cardiovasculaire metingen, inclusief de geavanceerde noninvasieve hemodynamische meetmethoden, te doen.

De focus in het tweede deel lag op het algemene en etnisch specifieke risicoprofiel van de stedelijke Surinamers. Hoofdstuk 4 beschrijft het cardiovasculair risicoprofiel van de 1,157 deelnemers tussen de 18-70 jaar (gemiddeld 43 jaar). We vonden dat meer dan driekwart (77%) van de deelnemers minstens een van de volgende risicofactoren hadden: hypertensie (40%), diabetes mellitus (22%), dyslipidemie (41%), obesitas (37%) of roken (18%). Daarnaast was 22% prehypertensief, prediabeet, had overge- wicht of bordeline dyslipidemie. Dit resulteerde in slechts 1% van de populatie die een optimaal cardiovasculair risicoprofiel had. Van de jonge mensen onder de 40 jaar had 62% een risicofactor en maar 2% van de mensen een optimaal risicoprofiel. Naast de traditionele cardiovasculaire risicofactoren hebben we ook gekeken naar vroegtijdige orgaanschade, wat bij 22% van de deelnemers voorkwam. De meerderheid van deze mensen had een verhoogde vaatstijfheid (77%). Zelfgerapporteerde HVZ kwam voor in 5% van de deelnemers: herseninfarct (3%), hartinfarct (2%), en ernstige nierschade (0.4%). Ook waren er duidelijke etnische verschillen te zien in het vóórkomen van tra- ditionele cardiovasculaire risicofactoren. Hypertensie was ongecorrigeerd het hoogste in de Creolen (48%), wat significant verschilde van Marrons (38%) en Hindostanen (37%). Op hun beurt hadden de Hindostanen de hoogste prevalentie van diabetes en dyslipidemie (respectievelijk, 30% en 52%), gevolgd door de Javanen (25% en 47%) en de Creolen (22% en 37%). Bij de Marrons kwamen diabetes en dyslipidemie relatief weinig voor (11% en 27%). Obesitas verschilde niet tussen de 4 etnische groepen. Daarna werd de populatie verdeeld in een groep van Afrikaanse afkomst (Creolen en Marrons) en van Aziatische afkomst (Hindostanen en Javanen). Het bleek dat hyperten- sie ongecorrigeerd niet verschilde tussen de groep van Afrikaanse afkomst en die van Aziatische afkomst. Echter, na corrigeren voor verschillende factoren, waren de mensen van Afrikaanse afkomst vaker hypertensief dan zij van Aziatische afkomst. Diabetes mellitus en dyslipidemia kwamen juist significant vaker voor bij de Aziatische groep. Er werden geen verschillen gezien in obesitas tussen de groep van Afrikaanse afkomst en die van Aziatische afkomst.

Achtenzestig procent (68%) van de mensen met hypertensie was zich hiervan bewust en 56% werd hiervoor behandeld. Een goed ingestelde bloeddruk was aanwezig in 36% van zij die behandeld werden voor hoge bloeddruk (20% van alle hypertensieven). Een meerderheid van de diabeten was zich bewust hun ziekte (59%) en kreeg medicijnen APPENDICES 177

(54%), maar slechts 21% van zij die behandeld werden, waren goed ingesteld (11% van alle diabeten).

Obesitas bevordert atherosclerose en is daardoor een belangrijke risicofactor voor HVZ. Echter, het blijft onduidelijk welke vorm van obesitas meting nu het meest geassocieerd is met een hoog risico op HVZ. In Hoofdstuk 5 vergeleken we de meer gecompliceerde lichaamssamenstellingsmaten, zoals vetmassa of een lage vet vrije massa samen met een hoge vetmassa (sarcopenic obesity), met de simpele antropometrische metingen, zoals de body mass index (BMI) of middelomtrekken, in de associatie met HVZ risico. We vonden dat het cardiovasculair risico bij personen van Afrikaanse en Aziatische afkomst het sterkste geassocieerd was met middelomtrekmaten, in het bijzonder de middelom- trek en de middel-lengte ratio, onafhankelijk van geslacht, leeftijd, of etniciteit. BMI was van toegevoegde waarde in de totale groep en bij mannen in de associatie met een verhoogde vaatstijfheid. Concluderend, het gebruik van lichaamssamenstellingsmaten is van weinig toegevoegde waarde naast de simpele antropometrische metingen. De studie liet ook zien dat er etnische verschillen zijn in lichaamssamenstelling, namelijk dat Aziaten over het algemeen een grotere middelomtrek en meer vetmassa hebben bij dezelfde BMI als Afrikanen.

Fysieke activiteit draagt in belangrijke mate bij aan het beperken van obesitas. We heb- ben gekeken naar de relatie tussen obesitas en fysieke activiteit in een populatie van Aziatische en Afrikaanse afkomst in Hoofdstuk 6. De resultaten van deze studie lieten zien dat het voldoen aan de aanbevolen hoeveelheid fysieke activiteit en actief zijn binnen het transport en vrije tijd domein negatief geassocieerd zijn met obesitas in de totale populatie. In de Afrikaanse populatie was een langere duur en meer volume geassocieerd met minder obesitas. Dit zou kunnen pleiten voor etnisch specifieke aanbevelingen voor fysieke activiteit.

Etnische verschillen in het risico op HVZ worden niet volledig verklaard door verschil- len in cardiovasculaire risicofactoren. Een manier om de etnische variatie in HVZ risico te onderzoeken is door de focus te verleggen van risico-inschatting naar het di- agnosticeren van vroegtijdige orgaanschade in bijvoorbeeld bloedvaten. Een type van vroegtijdige orgaanschade is een verhoogde polsgolfsnelheid. In Hoofdstuk 7 hebben we de etnische verschillen in de polsgolfsnelheid (PWV) onderzocht. De belangrijkste bevinding was dat mensen van Aziatische afkomst een substantieel hogere PWV van 1 m/s hadden, na correctie voor leeftijd en bloeddruk. Het etnische verschil in PWV persisteerde na additionele correctie voor glucose, lipiden, BMI en middelomtrek. Het A is nog niet duidelijk waarom mensen van Aziatische afkomst een hogere PWV hebben, 178 APPENDICES

zelfs na correctie voor cardiovasculaire parameters. Echter, de hogere PWV in mensen van Aziatische afkomst past bij hun hogere risico op ischemische hartziekten.

In het derde deel van deze thesis hebben we de prevalentie, het bewustzijn, de behande- ling, het halen van de streefwaarden en de determinanten van hypertensie vergeleken tussen Surinamers die wonen in Suriname en eerste- en tweede-generatie Surinaamse migranten die wonen in een HIL (deelnemers van de Healthy Life in an Urban Setting (HELIUS) studie in Amsterdam, Nederland). Hoofdstuk 8 beschrijft dat in beide groepen hypertensie veelvoorkomend was met suboptimale aantallen van bewuste, behandelde en gecontroleerde mensen. Terwijl mannen in Suriname even vaak hypertensief waren als Surinaamse mannen in Nederland, waren vrouwen in Suriname vaker hypertensief en reeds op een jongere leeftijd dan Surinaamse vrouwen in Nederland. Ook waren vrouwen in Suriname zich vaker niet bewust van hun hoge bloeddruk en minder vaak behandeld. De onderliggende determinanten van hypertensie in Suriname waren gelijk aan die van Nederland, met leeftijd en middelomtrek als de belangrijkste determinan- ten van hypertensie. De vroegtijdige detectie van hypertensie en het voorkomen van (centrale) obesitas moeten dan ook worden geprioriteerd.

Ten slotte is in Hoofdstuk 9 gereflecteerd op de belangrijkste bevindingen, daarbij rekening houdend met de methodologische beperkingen. Ook is er uiteengezet wat de bevindingen impliceren en zijn er aanbevelingen gedaan voor toekomstig onderzoek. De bovenstaande hoofdstukken hebben geleid tot de volgende twee conclusies: 1. De cardiovasculaire gezondheidsstatus van de stedelijke Surinamers was verre van optimaal: slechts 1% had een optimaal risicoprofiel. Om te voorkomen dat er een stijging zal plaatsvinden in het aantal cardiovasculair gerelateerde sterfgevallen, zullen er drastisch preventieve maatregelen moeten worden genomen op popu- latieniveau, zoals vroegtijdig screenen voor hypertensie en het voorkomen van obesitas. 2. Er zijn etnische verschillen in het cardiovasculair risico van de stedelijke Surinamers. Mensen van Aziatische afkomst hadden significant vaker diabetes en dyslipidemie dan personen van Afrikaanse afkomst. Daarnaast hadden mensen van Aziatische af- komst een hogere polsgolfsnelheid, zelfs na correctie voor en in de afwezigheid van risicofactoren. Aan de andere kant hadden mensen van Afrikaanse afkomst vaker hypertensie na correctie voor andere oorzakelijke factoren. De etnische verschillen in het cardiovasculair risico passen bij de etnische verschillen in HVZ en vragen om een etnisch specifieke aanpak van preventie. APPENDICES 179

List of publications

Publications presented in this thesis Diemer FS, Brewster LM, Haan YC, Oehlers GP, van Montfrans GA, Nahar-van Venrooij LMW. Body composition measures and cardiovascular risk in high-risk ethnic groups. Clin Nutr. 2017, pii: S0261-5614(17)31410-3. doi: 10.1016/j.clnu.2017.11.012.

Diemer FS, Baldew SM, Haan YC, Aartman JQ, Karamat FA, Nahar-van Venrooij LMW, van Montfrans GA, Oehlers GP, Brewster LM. Hypertension and cardiovascular risk profile in a middle-income setting: the HELISUR study. Am J Hypertens. 2017, 30(11):1133-1140. doi: 10.1093/ajh/hpx105.

Baldew SM, Diemer FS, Cornelissen V, Oehlers GP, Brewster LM, Toelsie JR, Vanhees L. Physical activity and obesity: is there a difference in association between the Asian- and African- Surinamese adult population? Ethn Health. 2017; 1-13. doi: 10.1080/13557858.2017.1346187.

Aartman JQ, Diemer FS, Karamat FA, Bohte E, Baldew SM, Jarbandhan AV, van Montfrans GA, Oehlers GP, Brewster LM. Assessing population cardiovascular risk with advanced hemodynamics: the Healthy Life in Suriname (HELISUR) feasibility study. Revista Pana- mericana de Salud Pública/Pan American Journal of Public Health 2017; 41:e46.

Diemer FS, Aartman JQ, Karamat FA, Baldew SM, Jarbandhan AV, van Montfrans GA, Oehlers GP, Brewster LM. Exploring cardiovascular health: the Healthy Life in Suriname (HELISUR) study. A protocol of a cross-sectional study. BMJ Open 2014, 4(12);1-6.

Publications not related to this thesis Diemer FS, Haan YC, Nannan Panday RV, van Montfrans GA, Oehlers GP, Brewster LM. Health literacy in Suriname. Social Work in Health Care 2017; 56(4):283-293.

Haan YC, Oudman I, Diemer FS, Karamat FA, van Valkengoed IG, van Montfrans GA, Brew- ster LM. Creatine kinase as a marker of obesity in a multi-ethnic population. Molecular and Cellular Endocrinology 2017; 442:24–31.

Jarbandhan AV, Hoozemans MJM, Buys R, Diemer FS, Baldew SM, Aartman J, Veeger DHEJ. Prevalence of self-reported stroke in association with ethnic background within a multi-ethnic population in Paramaribo, Suriname: Results from the HeliSur study. A Neuroasia 2016; 21(4):303–310. 180 APPENDICES

Haan YC, Diemer FS, van der Woude L, Van Montfrans GA, Oehlers GP, Brewster LM. The risk of hypertension and cardiovascular disease in women with uterine fibroids. Journal of Clinical Hypertension 2018; 20(4):718-726. doi:10.1111/jch.13253. APPENDICES 181

Contributing authors

Jet Aartman Veronique Cornelissen Formerly: Department of Vascular Medi- Department of Rehabilitation Sciences, cine Research Center for Cardiovascular Academic Medical Center, University of Rehabilitation, KU Leuven, Amsterdam, Leuven, Belgium Amsterdam, the Netherlands Yentl Haan Charles Agyemang Department of Vascular Medicine Department of Public Health Academic Medical Center, University of Academic Medical Center, University of Amsterdam, Amsterdam, Amsterdam, the Netherlands Amsterdam, the Netherlands Ameerani Jarbandhan Se-Sergio Baldew Department of Physical Therapy, Anton Department of Physical Therapy, Anton de Kom University of Suriname, de Kom University of Suriname, Paramaribo, Suriname Paramaribo, Suriname Fares Karamat Evelien Bohte Department of Vascular Medicine Formerly: Department of Vascular Medi- Academic Medical Center, University of cine Amsterdam, Academic Medical Center, University of Amsterdam, the Netherlands Amsterdam, Amsterdam, the Netherlands Gert van Montfrans Department of Internal Medicine Bert-Jan van den Born Academic Medical Center, University of Department of Vascular Medicine Amsterdam, Academic Medical Center, University of Amsterdam, the Netherlands Amsterdam, Amsterdam, the Netherlands Lenny Nahar-van Venrooij Department of Public Health, Anton de Lizzy Brewster Kom University of Suriname, Amsterdam Institute for Global Health Paramaribo, Suriname and Development (AIGHD), University of Amsterdam, A Amsterdam, the Netherlands 182 APPENDICES

Rani Nannan Panday Karien Stronks Department of Vascular Medicine Department of Public Health Academic Medical Center, University of Academic Medical Center, University of Amsterdam, Amsterdam, Amsterdam, the Netherlands Amsterdam, the Netherlands

Glenn Oehlers Jerry Toelsie Department of Cardiology, Academic Department of Physiology, Anton de Kom Hospital Paramaribo, University of Suriname, Paramaribo, Suriname Paramaribo, Suriname

Ron Peters Luc Vanhees Department of Cardiology Department of Rehabilitation Sciences, Academic Medical Center, University of Research Center for Cardiovascular Reha- Amsterdam, bilitation, KU Leuven, Amsterdam, the Netherlands Leuven, Belgium

Marieke Snijder Department of Public Health Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands APPENDICES 183

Author contributions

Chapter 2: Exploring cardiovascular health: the Healthy Life in Suriname (HELISUR) study. A protocol of a cross-sectional study GPO, GAvM and LMB conceived the study, wrote the protocol and obtained funding. GPO sought ethical approval. FSD and LMB drafted the manuscript. All authors read, edited draft versions and approved the final manuscript.

Chapter 3: Assessing population cardiovascular risk with advanced hemodynamics: the Healthy Life in Suriname (HELISUR) feasibility study GPO, GAvM and LMB conceived the study, wrote the protocol and obtained funding. GPO sought ethical approval. JQA, FSD and EB collected the data. JQA, FSD, and FAK analysed the data. JQA, FSD, and LMB drafted the manuscript. All authors read, edited draft versions and approved the final manuscript.

Chapter 4: Hypertension and cardiovascular risk profile in urban Suriname: the HELI- SUR study GPO, GAvM and LMB conceived the study and wrote the protocol. GPO sought ethical approval. FSD, SMB, and JQA were responsible for the recruitment of the participants and the collection of the data. FSD and SMB conducted the statistical analyses. FSD, SMB, and LMB interpreted the data and drafted the manuscript. All authors read, edited draft versions, and approved the final manuscript.

Chapter 5: Body composition and cardiovascular risk in high-risk ethnic groups GPO, GAvM and LMB conceived the study and wrote the protocol. LMWN had the primary responsibility for the article. FSD collected and analysed the data. LMWN oversaw the statistical analysis and made substantial contributions in the interpretation of the data. All authors read, edited draft versions and approved the final manuscript.

Chapter 6: Physical activity and obesity: is there a difference in the association be- tween the Asian- and African-Surinamese adult population? GPO, GAvM and LMB conceived the study and wrote the protocol. GPO sought ethical approval. SMB and FSD analysed the data. SMB, FSD, VC and JRT drafted the manuscript. All authors read, edited draft versions and approved the final manuscript.

Chapter 7: Aortic pulse wave velocity in an Asian and African ancestry population GPO, GAvM and LMB designed the study. FSD collected the data. FSD and YCH performed A the data analysis. FSD, SMB and LMB drafted the manuscript. All authors critically re- viewed the manuscript, edited draft versions and approved the final manuscript. 184 APPENDICES

Chapter 8: Hypertension in Surinamese living in a middle- versus high-income coun- try: the HELISUR and HELIUS studies FSD drafted the manuscript and conducted the statistical analyses. KS and LMB made substantial contributions in the interpretation of the data and in revising the manu- script. All authors critically reviewed the manuscript. APPENDICES 185

Portfolio

PhD candidate: Frederieke Sophie Diemer PhD period: May 2015 – September 2018 PhD supervisor: prof. dr. Ron J. G. Peters PhD daily supervisors: dr. Lizzy M. Brewster, dr. Lenny M. W. Nahar-van Venrooij Institution: Department of Cardiology, AMC-UvA

Courses Academic Medical Center Graduate School, University of Amsterdam: - Practical Biostatistics, 2015 - Pubmed (e-learning course), 2015 - Clinical Epidemiology: Evaluation of Medical Tests, 2018 - Advanced Topics in Biostatistics, 2018 - Basiscursus Regelgeving en Organisatie voor Klinische onderzoekers (BROK), 2018 Erasmus Summer Programme, Erasmus Medical Center Rotterdam: - Conceptual Foundation of Epidemiologic Study Design, 2017 - Causal Inference, 2017 - Causal Mediation Analysis, 2017 - Markers and Prediction Research, 2017 Tulane University, New Orleans: - Introduction to Epidemiology, 2018 - Research Methods in Epidemiology, 2018

Oral presentations - HELISUR Final Symposium, Paramaribo, 2018 - 27th European Meeting on Hypertension and Cardiovascular Protection, Milan, 2017 - Symposium Scientific Research Center Suriname, Paramaribo, 2017 - 26th Scientific Meeting of the International Society of Hypertension, Seoul, 2016 - 26th European Meeting on Hypertension and Cardiovascular Protection, Paris, 2016 - Caribbean Public Health Agency (CARPHA) Conference, Turks and Caicos, 2016 - Surinamedag 2014, Den Haag, 2014 - Health Conference, Paramaribo, 2014

Poster presentations - HELISUR Final Symposium, Paramaribo, 2018 - 27th European Meeting on Hypertension and Cardiovascular Protection, Milan, 2017 A - 26th European Meeting on Hypertension and Cardiovascular Protection, Paris, 2016 186 APPENDICES

- 7th Annual Conference of the Consortium of Universities in Global Health, San Fran- cisco, 2016 - Joint Scientific Meeting of the Belgian, Dutch and Swiss Hypertension Societies, Antwerp, 2015 - Surinamedag 2015, Hoofddorp, 2015 - EuroPRevent 2015, Lisbon, 2015 - 25th European Meeting on Hypertension and Cardiovascular Protection, Milan, 2015 - Surinamedag 2014, Hoofddorp, 2014 - 23rd European Meeting on Hypertension and Cardiovascular Protection, Milan, 2015.

Organizing - Training days for HELISUR interviewers, Paramaribo, 2012 - HELISUR final symposium, Paramaribo, 2018

Coordinating Coordinating of the HELISUR study among 1,157 Surinamese living in Paramaribo

Supervising Master thesis, Lisa van der Woude, Medicine, AMC, 2013-2014 Master thesis, Iteke van Seventer, Medicine, AMC, 2014-2015 Master thesis, Ramona Budike, Public Health, University of Suriname, 2015-2017 Master thesis, Kimberly Tsen, Public Health, University of Suriname, 2015-2017 Master thesis, Saskia Castelen, Public Health, University of Suriname, 2015-2017 Master thesis, Astrid van Engel, Public Health, University of Suriname, 2015-2017 Master thesis, Indira Madhuban, Public Health, University of Suriname, 2015-2017 Master thesis, Suleta Monsels, Public Health, University of Suriname, 2015-2017 Master thesis, Chiquita Mijnals, Public Health, University of Suriname, 2015-2017 Master thesis, Sushma Harlal, Public Health, University of Suriname, 2015-2017

Reviewer of submission to Scientific Reports Nature, 2015

Funding - Catharina van Tussenbroek Fonds, Travel Grant for Young Women Academics, 2015 - Travel and Accommodation Grant European Society of Hypertension, 2016 - Spinozafonds, Universiteit van Amsterdam, 2017

Other Journal club, Scientific Research Center Suriname, 2016-2018 APPENDICES 187

Dankwoord

Het is eindelijk zover: mijn proefschrift is af! Terwijl ik mijn dankwoord schrijf, dringt het tot me door wat er in 5 jaar kan gebeuren. We hebben from scratch een groot onder- zoek opgezet waarbij ik alle tussenstappen heb mogen organiseren en meemaken: van studieprotocol naar pilotstudie, van interviews aan huis naar lichamelijke onderzoeken in het Academisch Ziekenhuis Paramaribo, van de data analyseren naar aan tafel bij de Surinaamse Minister van Volksgezondheid om onze resultaten te overhandigen. Eén ding is zeker: zonder de steun, de hulp, de wijsheid, de ervaringen en de motiverende woorden van de mensen om mij heen was dit zeker niet gelukt. Een aantal personen wil ik graag in het bijzonder bedanken.

Prof. dr. R.J.G. Peters, beste Ron, wat was ik blij toen je antwoordde dat je me wilde begeleiden. Helaas was de begeleiding maar van korte duur, want ik heb deze als zeer prettig ervaren! Je relativerende blik en oplossingsgerichte aanpak waren precies hetgeen ik nodig had om dit promotietraject tot een goed einde te maken. De veelbe- sproken fles wijn kan nu eindelijk worden leeggedronken!

Dr. L.M. Brewster, beste Lizzy, jij hebt mij wegwijs gemaakt in de wereld van de weten- schap. Ik bewonder je enorme gedrevenheid als onderzoeker. Daarnaast wil ik je bedan- ken voor je kritische blik bij het meedenken en meelezen van mijn stukken. Overigens, als jij nooit op het idee was gekomen om HELISUR op te zetten, had mijn leven er nu heel anders uit gezien. Dank daarvoor!

Dr. L.M.W. Nahar-van Venrooij, beste Lenny, jij bracht structuur en richting aan in een nogal ongestructureerde wereld. Ik heb ontzettend veel van je geleerd. Over jouw schouder heb ik mogen meekijken in de wereld van de epidemiologie en het was dan ook door jou dat ik enthousiast werd voor dit vak. Daarnaast hebben we elkaar persoon- lijk goed leren kennen. Hoe jij werk en privé combineert is bewonderenswaardig. Dank voor de motivatie, het vertrouwen en de betrokkenheid.

Hooggeachte leden van de promotiecommissie, prof. dr. K. Stronks, prof. dr. A.E. Kunst, prof. dr. V.W.V. Jaddoe, prof. dr. H.C.P.M. van Weert, dr. N.R. Bindraban, dr. E.P. Moll van Charante en dr. I.G.M. van Valkengoed, hartelijk dank voor het lezen en beoordelen van dit proefschrift en de bereidheid om plaats te nemen in mijn promotiecommissie.

Daarnaast wil ik prof. dr. Karien Stronks bedanken voor de inspirerende gesprekken. Je A hebt me altijd gesteund en welkom laten voelen bij de Helius onderzoeksgroep. Mijn registratie tot epidemioloog heb ik grotendeels aan jou te danken. 188 APPENDICES

Een onderzoek als HELISUR zet je niet in je eentje op. Ten eerste, alle deelnemers van de HELISUR studie, bijzonder veel dank voor jullie deelname en vertrouwen. Ten tweede, alle 50 interviewers die door de hete zon hebben gelopen om de HELISUR interviews af te nemen. Door jullie inzet hadden we binnen 2 maanden 1,800 respondenten! Daar- naast ook de data-entry medewerkers van de Medische Faculteit van de ADEK en de research assistenten die hun wetenschappelijke stage binnen HELISUR hebben gedaan. Super bedankt dat jullie ons wilde ondersteunen! Daarnaast wil ik alle lieve zusters en cardiologen van het Academisch Ziekenhuis Paramaribo bedanken voor het opnemen van deze vreemde eend in de bijt.

Drs. G.P. Oehlers, beste Glenn, zonder uw inzet was HELISUR nooit van de grond geko- men. Binnen no time waren zaken geregeld als u hielp. Daarnaast konden Sergio en ik met alles bij u terecht en stond u altijd achter ons.

Lieve Jet, samen werden wij naar Suriname ‘uitgezonden’ en ik had geen betere mede- stagiaire kunnen treffen. Je hulp is van onschatbare waarde geweest! Jij had de hands on mentaliteit en ik de… SPSS data mentaliteit. De werk-privé balans hebben we in ieder geval altijd goed weten te houden!

De hypertensiegroep, dank voor alle samenwerking en nieuwe wetenschappelijke inzichten. Gert, dat je naar Suriname bent gevlogen voor het HELISUR congres was een grote eer! Yentl en Fares, de samenwerking was altijd prettig en vruchtbaar. Op de congressen hebben we elkaar nog veel beter leren kennen. Nu is het jullie beurt!

Lieve Sergio, Fenna, Marlies en Eva, wat ik ben ik blij dat ik jullie heb leren kennen. Het duurde even (he Fenna?) maar dan heb je ook wat! Bedankt voor alle gezellige en bovenal lekkere (geen oplos-) koffies op de Medische Faculteit. We konden altijd bij elkaar terecht met een lach en een traan (die laatste was in ieder geval nooit van Eva). Nu ik niet meer in Suriname ben, zal de vriendschap een andere vorm aannemen, maar ik heb er alle vertrouwen in dat die altijd gaat blijven bestaan. Sergio, dat jij helemaal uit Suriname komt om mijn paranimf te zijn, maakt deze dag sowieso fantastisch.

Lieve Puck, zonder jou was Suriname toch een stuk minder leuk geweest. Super blij dat ik bij jou in de tandartsstoel kroop. Dat we nu praktisch familie zijn, is de kers op de taart.

Lieve Pili’s, vriendinnen vanaf de Intreeweek and still going strong! Ooit samen in de collegebanken en ’s avonds in de Havelaar en nu ieder zijn eigen (carrière)pad. Trots op APPENDICES 189 jullie! Lynne en Britt, ik koester de leuke vakanties in Suriname. Snel een keer met zijn allen! Of wordt het toch eerst een reis naar Australië?

Lieve Tomaatjes, vrienden vanaf de middelbare school en nog steeds elk jaar op vakan- tie met elkaar. Tomatenjongens, jullie humor is de beste van de wereld. Na één vakantie met jullie heb ik weer genoeg gelachen voor de rest van het jaar. Marij, Kiry, Nienke en Nathalie, mijn slimme, knappe vriendinnen! Met jullie heb ik het gevoel dat ik de hele wereld aan kan. Ik weet zeker dat we op ons 80e nog steeds samen zijn.

Lieve Carla, dank voor je warmte en zorgzaamheid. We gaan proberen je lekkere Su- rinaamse recepten te evenaren, maar dat is natuurlijk per definitie een verloren zaak. Hopelijk kom je snel deze kant op om ons te bezoeken en anders hebben we een goed excuus om lekker vaak naar Suriname te gaan.

Lieve papa, mama, Dorine, Pieter, Daan en Marit, ik weet dat ik altijd met alles bij jul- lie terecht kan! De vakanties naar Suriname zullen bij mij altijd een speciaal plekje in mijn hart hebben. Papa en mama, dank dat jullie er altijd voor me zijn. Niets is teveel gevraagd en een betere opa en oma zouden wij ons niet kunnen wensen! Dorine, dank dat jij op deze bijzondere dag aan mijn zijde wil staan.

Lieve Gavin, rots in de branding. Wat ben ik ontzettend blij dat ik jou heb ontmoet. Hoe jij in het leven staat is bewonderenswaardig. Bedankt voor het adviseren, motiveren, maar – vooral – accepteren de afgelopen jaren. Op naar een nieuw avontuur samen!

Lieve kleine Aaron, wat word je al een groot en wijs mannetje! Je humor en je nieuws- gierigheid werken aanstekelijk. Bedankt voor het elke dag herinneren waar het in het leven omdraait.

Bedankt - Gran tangi,

Frederieke

A

APPENDICES 191

About the author

Frederieke Sophie Diemer was born on July 16th 1989 in De Bilt, the Netherlands. After graduating in 2007 at the Christelijk Gymnasium Utrecht, she moved to Amsterdam and started her training in Biomedical Sciences at the University of Amsterdam. She received her Bachelor of Science (BSc) degree in Biomedical Sciences in 2010 and a Master of Science (MSc) degree (cum laude) in Biomedical Sciences in 2012. During her MSc training, Frederieke did a specialisation in endocrinology and cardiovascular research and wrote a thesis on the association between the glycocalyx thickness and cardiovascular risk. During her master internship, she was trained in the HELIUS study procedures and subsequently went to Suriname to test the feasibility of the HELISUR study.

In 2013, Frederieke moved to Suriname and started her job as the research coordinator and data manager of the HELISUR study at the Academic Hospital Paramaribo. She had a fruitful collaboration with researchers from the Faculty of Medicine of the University of Suriname and tutored several master Public Health students during their master thesis. In 2015, Frederieke officially started as a PhD candidate at the Academic Medi- cal Center, University of Amsterdam, under supervision of Professor dr. R. Peters, dr. L. Brewster and dr. L. Nahar-van Venrooij, of which this thesis is the result. During her PhD training, Frederieke followed several (post)graduate epidemiology courses and will be registered as an epidemiologist B at the Netherlands Epidemiological Society. Frederieke will start this year with her new job at the National Health Care Institute (Zorginstituut Nederland).

Frederieke is married and has a two-year-old son.

A