Université Libre de Bruxelles and Lebanese University The Doctoral School of Sciences and Technology

(1)

Co-supervision thesis

For the degree of Doctor

Université Libre de Bruxelles

and

Lebanese University

The Doctoral School of Sciences and Technology

Speciality: Public Health, Epidemiology

Presented and defensed by

NASSER Zeina

Outdoor air pollutants and cardiovascular diseases in Lebanon

Thesis Director (Belgium): Prof. LEVÊQUE Alain

Thesis Director (Lebanon): Prof. SALAMEH Pascale

Jury Members

(2)

LIST OF FIGURES

Figure 3.1. 1: Flowchart for identification and inclusion of relevant studies for the review .. 24

Figure 4.3. 1: ROC curves for Score 1 and Score 2 to predict cardiovascular diseases.

(5)

LIST OF TABLES

Table 4.1. 1: Annual Mean outdoor PM10 levels, major sources of air pollution and presence of association with cardiovascular diseases (CVD) in selected countries of Middle East.

Review of the published literature 2000-2013. ... 32

Table 4.2. 1: Baseline characteristics of the participants. ... 34

Table 4.2. 2: Subject's exposure to active, passive smoking and CVD ... 35

Table 4.2. 3: Exposure to outdoor air pollution and CVD ... 36

Table 4.2. 4: Adjusted odds ratios with their 95% confidence intervals from the logistic regression of CVD among cases and control... 37

Table 4.2. 5: Adjusted odds ratios with their 95% confidence intervals from the logistic regression of CVD among cases and control stratified by cigarette smoking status ... 37

Table 4.3. 1: Risk factor profile of the participants in the case control study ... 38

Table 4.3.2: Logistic Regression for predicting of cardiovascular disease events among the participants. Model 1= traditional risk factors (TRF) of CVD and Model 2 = TRF of CVD in addition to outdoor air pollution exposure assessment. ... 39

Table 4.3. 1: Risk factor profile of the participants in the case control study ... 38

(6)

ACKNOWLEDGEMENTS

This Thesis would never have been achieved without help and contribution of my committee members, friends and family. I wish to express my sincere gratitude to everyone who supported me during the last three years.

First of all, I would like to express my deepest appreciation to my principal advisor Professor Alain Lêveque for his continuous support, for guiding my research over the past three years and helping me to enrich and develop my background in epidemiology and public health. Thank you for being an excellent advisor. I hope that one day I would become as good an advisor to my students as Prof. Levêque has been to me.

My co-advisor, Professor Pascale Salameh, has been always there to listen and advise me. Thank you for your encouragement and friendship, for making it possible for me to accomplish this thesis, for always having your door open, I am deeply grateful to you for sharing your deep knowledge and expertise in research and teaching me how to conduct it.

My sincere thanks to Professor Catherine Bouland, president of my jury committee, thank you for your constructive criticisms throughout the different stages of my research, for commenting on my views and helping me understand and enrich my ideas.

(7)

I would like to thank Professor Nissaf Ben Alaya for her participation, encouragement and constructive criticism.

I would like to express my special appreciation and thanks to Professor Nadine Saleh President of my jury committee.

Most importantly, none of this would have been possible without the love and patience of my friend Linda Abou Abbas, who helped me to stay strong through these difficult years. Her support and care helped me overcome obstacles and stay focused on my graduate study. I greatly value her friendship and I deeply appreciate her belief in me. I love you.

(8)

ABSTRACT

Outdoor air pollution is increasingly considered as a serious risk factor for cardiovascular diseases (CVD). High levels of airborne particulate matter (PM) constitute the greatest international air pollution threat. The purpose of this thesis is to broaden our knowledge regarding the relationship between outdoor air pollution and cardiovascular diseases in the Middle Eastern countries, specifically in Lebanon. Moreover, we aimed to develop a scale as CVD screening tool among the Lebanese population. To achieve these goals, we conducted three studies. The first was a systematic review of the literature aiming to assess levels and sources of PM across the Middle East area and to search for an evidence of relationship between PM exposure and CVD (Paper I).The second manuscript was a multicenter case-control study investigating the association between outdoor pollutants and cardiovascular diseases among Lebanese adults (Paper II) while the third study was conducted to develop a score that can be used as a screening tool in clinical and epidemiological settings among the Lebanese adults (Paper III).

The annual average values of PM pollutants in the Middle East region are considered to be much higher than the WHO 2006 tolerated levels (PM2.5 = 10 µg/m3, PM10 = 20

(9)

highlight the importance of scale generation, which includes air pollution as predictive factor, as screening tool for patients at risk of CVD. This scale can foresee the cardiovascular disease outcomes better than the established score which use the traditional CVD risk factors (Paper III).

In conclusion this study brings new evidence regarding the effects of particulate matter on cardiac diseases, points out the harmful role of diesel exhaust on health and suggest a an important role of traffic exhaust particles in exacerbating heart diseases in the Middle East Region. The developed scale could detect persons at high risk for CVD in the clinical and epidemiological settings. In addition, it serves as an essential public health screening tool for the primary prevention of CVD.

(10)

RESUME

La pollution de l'air extérieur est de plus en plus considérée comme un facteur de risque important pour les maladies cardiovasculaires. Des niveaux élevés des matières particulaires constituent la plus grande menace de la pollution de l'air à l’échelle internationale. Le but de cette thèse est d'élargir nos connaissances sur la corrélation entre la pollution de l'air extérieur et les maladies cardio-vasculaires dans les pays du Moyen-Orient, en particulier au Liban. D'ailleurs, nous avons cherché à établir une échelle parmi la population libanaise comme outil de dépistage pour les maladies cardiovasculaires. Pour atteindre ces objectifs, nous avons mené trois études : La première a été une revue systématique des études visant à évaluer les niveaux et les sources des matières particulaires dans la zone du Moyen-Orient et à examiner l’association entre l'exposition aux particules et les maladies cardiovasculaires (Papier I). La deuxième a été une étude cas-témoins multicentrique pour évaluer l'association entre les polluants extérieurs et les maladies cardio-vasculaires chez les adultes libanais (Papier II) tandis que la troisième étude a été menée parmi les adultes libanais pour développer une échelle pouvant être utilisé comme un outil de dépistage épidémiologique et en clinique (Papier III).

Les valeurs moyennes annuelles de matières particulaires polluants dans la région du Moyen-Orient sont considérées comme beaucoup plus élevées que les niveaux tolérés par l’OMS 2006 (PM2,5 = 10µg / m3, PM10 = 20µg / m3). Nous avons découvert des

(11)

mortalité et d’hospitalisation (Papier I). Une augmentation du risque de maladies cardiovasculaire avec un rapport de cotes de 5,04, IC à 95% (de 4,44 à 12,85) si l’on vit près de l'autoroute occupée et de 4,76, IC à 95% (de 2,07 à 10,91) si l’on vit près du générateur diesel local a été remarquée parmi la population exposée à la pollution de l'air extérieur (Papier II). En outre, nos résultats mettent en évidence l'importance du développement d’une échelle incluant la pollution de l'air comme un facteur prédictif, comme un outil de dépistage pour les patients à risque de maladies cardiovasculaires. Cette échelle peut prédire mieux les évènements cardiaques que celle qui utilise les facteurs de risque traditionnels de maladies cardiovasculaires (Papier III).

En conclusion, cette étude apporte une nouvelle évidence concernant les effets des particules sur les maladies cardiaques. Elle souligne le rôle néfaste des gaz d'échappement diesel sur la santé et suggère le rôle important de particules de gaz d'échappement des voitures dans l'exacerbation de maladies cardiaques dans la région du Moyen-Orient. L'échelle développée de cette étude pourrait détecter les personnes à risque élevé de maladies cardiovasculaires en contexte cliniques et épidémiologiques. En plus, elle représente un outil de dépistage essentiel de la santé publique, pour la prévention primaire des maladies cardiovasculaires.

(12)

LIST OF ORIGINAL PAPERS

This thesis is based on the following published papers, which will be referred to in the text by their Roman numerals (I-II-III).

I. Zeina Nasser, Pascale Salameh, Wissam Nasser, Linda Abou Abbas, Elias Elias, Alain Leveque. Outdoor particulate matter (PM) and associated cardiovascular diseases in the Middle East: Review Paper. International Journal of Occupational Medicine and Environmental Health.2015; 28 (4) [PubMed].

II. Zeina Nasser, Pascale Salameh, Habib Dakik, Elias Elias, Linda Abou Abbas, Alain Leveque. Outdoor air pollution and cardiovascular diseases (CVD) in Lebanon: A case-control study. J Environ Public Health. 2015; 2015:810846 [PubMed].

(13)

LIST OF ABBREVIATIONS

APHEA Air Pollution and Health: A European Approach

ASSIGN Assessing cardiovascular risk using the Scottish

Intercollegiate Guidelines Network

BH Borj Hamoud

BMI Body Mass Index scale

CI Confidence Interval

CMB Chemical Mass Balance

CO Carbon monoxide

CRP C Reactive Protein

CUORE Cardiovascular Risk Assessment in Italy

CVD Cardiovascular disease

CVRF Traditional cardiovascular risk factors

EPA Environmental Protection Agency

FRS Framingham Risk Score

GC Greater Cairo

HDL High density lipoprotein concentration

HH Haret Hreik

HVAS High volume air sampler

IPFM Income-per-family-member

IRB Institutional Review Board

KSA Kingdom Saudi Arabia

LDL Low density lipoprotein concentration

MI Myocardial infarction

NAAQS National ambient air quality standard

NHANES National Health and Nutrition Examination Survey

NMMAPS National Mortality and Morbidity Air Pollution Study

NO2 Nitrogen dioxide

NPV Negative predictive value

O3 Ozone

OR Odds Ratio

P P value

PM Particulate Matter

PPV Positive predictive values

PROCAM Prospective Cardiovascular Munster

QRISK Cardiovascular disease risk score

ROC Receiver operator characteristics

ROS Reactive oxygen species

SBP Systolic blood pressure

SCORE Systematic Coronary Risk Evaluation

SD Standard deviation

Se Sensibility

SO2 Sulfur dioxide

Sp Specificity

SPSS Statistical Package for Social Sciences

TAQCC Tehran Air Quality Control Corporation

TF Tissue Factor

(14)

TRAP Traffic- related air pollutants

TSP Total suspended particles

UAE United Arab Emirates

WHO World Health Organization

XRF X-ray Fluorescence

(15)

THESIS STRUCTURE

(16)

INTRODUCTION

Outdoor air pollution, composed of a complex mixtures of gases (e.g., carbon monoxide (CO), nitrogen dioxide (NO2), ozone (O3), sulfur dioxide (SO2) and particulate matter (PM), is being increasingly considered a hazard on the cardiovascular system and has become an alarming health related problem as a result of the growing pace of urbanization and industrialization[1-5]. Cardiovascular diseases (CVD) are the leading cause of mortality and morbidity worldwide. This entity mainly consists of coronary heart disease, stroke and hypertension. Tobacco consumption, dyslipidemia, diabetes, physical inactivity and poor diet are considered as risk and contributing factors to CVD [6]. Over the past few years, a series of observational and epidemiological studies in the developed countries have explored the influence of environmental determinants on cardiovascular system [4, 7, 8]. As a result, traffic related air pollution was found to be a risk factor for myocardial infarction (MI) [9]. Moreover, living within vicinity of highways, up to a radius of 100 meters, was deemed another associated risk for acute myocardial events [10]. In addition, certain types of professions were more prone to cardiac events, such as Taxi or bus drivers[9], waste incinerator workers [11], smelters operatives [12], or chimney sweeps [13, 14]. Literature uncovered the fact that even shy amounts of exposure to pollutants can have deleterious effects on the human health system [2, 3, 15].

A study conducted by the NMMAPS (National Mortality and Morbidity Air Pollution Study) in the United States revealed that each increase by 20 μg/m3 of PM10 concentration

(17)

other hand, the effect of long term PM exposure on cardiovascular system has been studied by Laden et al. with an extended follow up reaching 28 years. Taking into account the PM2.5 levels, the authors measured its variant effect on cardiac system. They noted a

positive influence of PM on CV mortality: an increased risk of 1.28 per10 μg/m3 of PM2.5

and a relative risk of 0.69 once PM2.5 levels were decreased [17]. A different study

conducted by the American Cancer society II revealed a 12 % surge of death risks from Cardio Vascular pathologies for every 10 increase in PM2.5 exposure. Main cause of

demise was ischemic heart disease (18 % increase) while other reasons were arrhythmias and hear failures [4]. Another study conducted on large cohort of postmenopausal women among the different US cities found an association between elevated CV events and mortality for subjects with history of a 10 μg/M long term PM2.5 exposure [18].

(18)

(19)

CHAPTER 1 LITERATURE REVIEW

A literature review regarding the relationship between air pollution and cardiovascular diseases will be presented in this section of the thesis.

1.1 History of air pollution and cardiovascular disease

Past events such as the London fog in the 1952 [23] and Meuse Valley event in 1930 [24] in Belgium were associated with incidents of chest pain, respiratory complaints and increase in death toll and provided early observations concerning the harmful effects of air pollution on human health. . This initiated the first epidemiological studies approaching the association between outdoor air pollution and CVD.

1.2 Outdoor Air Pollutants

Researchers have raised concern over ambient air pollution, mostly emphasizing pollutants such as carbon monoxide, sulfur dioxide, oxides of nitrogen, ozone, and particulate matter.

1.2.1 Particulate Matter Air Pollution

In fact, the United Nations and the World Health Organization (WHO) have announced that particulate matter (PM) constitutes the greatest international air pollution threat [25]. Particulate matter consists of mixed solid and liquid aerosol particles that differ in size and chemical composition [26], and it is divided into: coarse (PM10–2.5 = 10–2.5 µm), fine

(PM2.5 < 2.5 µm), and ultrafine (PM0.1 < 0.1 µm) particles. Thousands of chemicals have

(20)

composition of PM determine the health risk it poses [26] these particles are inhaled deeply into the lungs where they can reach alveoli of the lungs, enter the pulmonary circulation and even pass into the systemic blood vasculature [2, 26, 27].

1.2.1.1 Biological Mechanisms of PM

Diverse biological pathways through which inhalation of PM into the lungs could be capable of triggering unfavorable effects on the cardiovascular system. The ultrafine particles and/or soluble PM constituents may rapidly enter pulmonary alveoli after inhalation and, subsequently, pass into the systemic blood circulation imposing its adverse inflammatory effects on the cardiovascular system. Another alternative pathway suggests that, following their inhalation, PM particles induce a pulmonary oxidative stress effect that might trigger systematic pro-oxidative and pro-inflammatory chain reactions and the release of pro-inflammatory mediators, activated leucocytes, cytokines (e.g., interleukin-6) and C-reactive protein from the pulmonary to the systemic circulation. These chemicals along with the pro-inflammatory mediators might, in turn, indirectly trigger adverse cardiovascular events summarized by endothelial dysfunction, cardiovascular oxidative stress, cardiovascular inflammation, acute myocardial infarctions, cardiac heart failure and chronic atherosclerosis [28-34].

1.3 Cardiac effects of air pollution exposure

1.3.1 Hospital admission

(21)

CO concentrations [36]. Moreover, a rise in ozone level along with nitrate and sulfate was noticed was linked to a rise in hospital admissions related to cardiac or respiratory problems [37].Other papers connected upsurges of particulate matter pollutants level to additional cardiopulmonary causes of hospital admissions [38]greater risk of Myocardial infarction [39], heart failure and arrhythmias among the elder population [40, 41].

1.3.2 Changes in Heart Rate and Cardiac Function

PM is responsible for variation in the physiological heart rate, potentially reflecting alterations in autonomic balance and/or cardiac electrical stability, which may cause arrhythmias and sudden death [4, 17]. Several studies showed that PM inhalation was associated with heart rate variability, within minutes to hours [2, 34, 42]. Moreover, other studies showed that traffic-related pollutants are a risk factor to alter heart rate variability particularly in older people [8, 43].

1.3.3 Homeostasis, coagulation profile, Vasomotor Tone Alterations and Hypertension

Generally speaking, PM has been found to cause rise in acute phase reactants, CRP (C Reactive Protein) plasma viscosity and inflammatory process. Moreover, it is linked to endothelial dysfunction and sparks abnormality and alteration in the cardiac autonomic system. Its effect as thrombus mediator and effect on the hemostasis was examined by several papers [44-46].

(22)

coupled with alteration in systolic blood pressure and largely affects the endothelial function of small capillaries[48].

1.3.4 Atherosclerosis

Air pollution can trigger vascular atherosclerosis especially when subjects endure Long exposure periods to pollutants [2, 4]. Precise explanation is missing but probable mechanisms have been speculated in the literature. The first hypothesis suggested a direct role of PM on the respiratory system, triggering an inflammatory response and subsequent chain reaction releasing inflammatory cytokines and prothrombic mediators [49]. The second explanation is about the translocation of PM particles (PM2.5 and PM0.1) into the

circulatory system bypassing the lungs and exerting their harmful effect directly on the cardiovascular system [50].

Allen et al. explored the role of residential proximity to major highways and PM2.5 exposure on calcification of abdominal aorta, a sign of systemic atherosclerosis. He used computer tomography as diagnostic modality on a large sample size from different major US mega cities. His team uncovered a higher risk of aortic calcification among residents who leaves near major roads. Moreover, the risk was sturdier among long term tenants. Hence, they concluded a strong connection between PM and systemic atherosclerosis [51].

1.4 Short term PM exposure studies

Not only extended exposure of PM can affect human health but even short term exposure could have its toll. A study conducted by the NMMAPS(National Mortality and Morbidity Air Pollution Study)revealed that each increase by 20 μg/m3 of PM10 concentration can

(23)

Registry) concluded that an exposure to only one hour of traffic was associated with a 2.92 odds ratio as risk for a heart attack [16].

1.5 Longer-term PM exposure studies

On the other hand, the effect of long term PM exposure on cardiovascular system has been studied by Laden et al. with an extended follow up reaching 28 years. Taking into account the PM2.5 levels, the authors measured its variant effect on cardiac system. They noted a

positive influence of PM on CV mortality: an increased risk of 1.28 per10 μg/m3 of PM2.5and a relative risk of 0.69 once PM2.5 levels were decreased [17].

A different study conducted by the American Cancer society II revealed a 12 % surge of death risks from Cardio Vascular pathologies for every 10 increase in PM2.5 exposure.

Main cause of demise was ischemic heart disease (18 % increase) while other reasons were arrhythmias and hear failures [4]. A third study conducted on large cohort of postmenopausal women among the different US cities found an association between elevated CV events and mortality for subjects with history of a 10 μg/M long term PM2.5

exposure [18].

1.6 Outdoor ambient air in the Middle East regions

(24)

Egypt

Cairo, the capital of Egypt with more than 15 million population [54], suffers from high concentrations of ambient pollutants, including: carbon monoxide, nitrogen oxides, sulfur dioxide, ozone and PM [55, 56].The phenomenon of thermal inversion further facilitates accumulation of atmospheric air pollutants emitted from different sources due to the slow wind movement and low temperature during cold seasons in Egypt [57]. Moreover, lack of rain and city layout characterized by narrow streets and tall buildings cause poor dispersion of pollutants over Cairo [54].

According to WHO, PM is the most common air pollutant in urban and industrial areas in Egypt [58]. In fact, almost 36% of atmospheric PM2.5 is attributed to on-road traffic where

more than 1 million cars, particularly older ones with poor technical specifications, circulate on the roads of Cairo. Other sources, namely open burning and residential combustion, contribute mainly to the remaining 54% of atmospheric PM2.5 [54].

Zakey has presented the mass concentrations of PM2.5 and PM10 and exposed their seasonal

variations from samples collected at 17 different sites, representing urban, industrial, residential and background (with a significant agriculture activity) sectors in the Greater Cairo (GC) over the period 2001–2002. The Egyptian law of environment has not yet established an annual limit for PM10 concentration, however, the concentrations of PM2.5

and PM10 measured in this study were generally high, with annual average values of 85±12

µg/m3 and 170±25 µg/m3[59], respectively, exceeding the WHO recommended guideline annual averages of 10 µg/m3 and 20 µg/m3 for PM2.5 and PM10 [60]. It is believed that the

combined effect of anthropogenic emissions (traffic and waste burning) and the transport of dust by wind from Moqattam hill and the desert to GC contribute to these high PM levels. The urban sector recorded the highest PM2.5/PM10 ratio (0.59) compared to the

(25)

spring season at the industrial sectors, whereas these concentrations were the lowest during the summer at the background sites [59] .

Another study was carried out to assess the source attribution of different ambient PM levels measured on 24-h basis at 6 sampling sites of Greater Cairo area representing industrial, traffic, residential and background conditions during 3 separate periods between February 1999 and June 2002. The samples were collected using the sampling protocol designed by Watson et al. [61]. Then, the Chemical Mass Balance (CMB) receptor model was used to allocate PM and its chemical constituents to their sources [62]. In this study, the PM10 mass concentration during the summer period of 2002 ranged between 99.2±5

µg/m3 in El-Zamalek (residential) to 175.3±9.1 µg/m3 in El-Maa’sara, an industrial site influenced by dust emissions from the surrounding cement factories. The PM2.5 mass

concentration varied between 34.7±1.9 µg/m3 in Kaha (background) to 60.7±3.2 µg/m3 in Shobra, an industrial location with numerous nearby lead smelters. This study has also revealed that PM2.5 is mainly attributed to a mobile source and open burning emissions. On

top of these sources, emissions from geological material were also a major contributor to PM10 [56] .

Iran

(26)

The annual average of PM10 concentrations in Tehran was 1.3 times the world’s average

(71 µg/m3) [67] and 4.5 times the WHO recommended guideline annual average values of 10 µg/m3 for PM2.5 and 20 µg/m3 for PM10 [60].

A study conducted in 2010 compared the mass and number concentrations of PM10, PM2.5,

and PM1.0 in the west-central parts of Tehran during 2 consecutive warm and cold seasons.

This study has shown that, in cold seasons, the PM10 mass concentration increased almost

2 folds and PM2.5 and PM1.0 nearly 3 times. The mean number concentration of these

particles was found to be almost 4.8 times in the same season[66]. This could be attributed to the Tehran’s geographical position being edged by mountains from the north, and to the fact that the wind is cold and slow, which makes air pollutants become trapped over Tehran [66-68].

A major finding of the study by Poursafa et al. in 2010 suggested PM to have impact on the platelet activation and atherosclerosis, which are associated with CVD [69]. In the same year, Davoodi et al. [70] examined the association between changes in the heart rate variability and exposure to air pollution in Tehran. During air pollution episodes, the maximum heart rate was significantly lower as compared to the clear air conditions (115.1±32.2 µg/m3 vs. 128.9±17.7 µg/m3). The occurrence of non-sustained supraventricular tachycardia, as a marker of cardiac dysfunction, was recorded in almost 43% of the participants during the polluted air conditions, whereas this arrhythmia was not reported during the non-polluted conditions (p = 0.001) [70].

(27)

tissue factor (TF) and thrombomodulin (TM) in children. The observed mean PM10 level

was notably high, exceeding the normal level (120.48 µg/m3 vs. 50μg/m3) [71].

Furthermore, a very recent study has provided quantitative estimates of the impact of short-term exposure to specific atmospheric pollutants on the health of population living in Tehran city from January 2010 to January 2011. The health impact of PM exposure on Tehran people’s health was assessed using the Air Quality Health Impact Assessment (Air Q 2.2.3) software created by the WHO European Centre for Environment Health, Bilthoven Division [72]. Air concentrations of PM pollutants in the capital were quantified by the Tehran Air Quality Control Corporation (TAQCC). According to this study, the annual mean PM10 level in Tehran was 90.58µg/m3. On average, 1367 extra total deaths

and 2580 extra cases admitted to hospitals due to CVD annually were attributable to the increase by 10 µg/m3 in PM10 levels [68].

Jordan

Air pollution in Jordan has become an increasing problem since the past 2 decades. This is partly attributed to the rapid growth of the Jordanian population where about 40% of the total inhabitants live in the capital Amman [73]. Nevertheless, the major source of air pollution in Amman is motor vehicle emissions with more than 80% of the 1 million registered vehicles in Jordan found in this city [74].

A cross-sectional study was conducted in order to examine the link between the traffic-related air pollutants (TRAP), particularly PM10 and total suspended particulates (TSP),

and vehicle traffic in Amman in 2010 in 2 high-polluted and 2 low-polluted random areas with relatively similar demographic and climatic characteristics [73]. Total suspended particulates and PM10 were measured as ambient air concentrations by the Jordanian Air

(28)

mean PM10 concentration in the high-polluted areas was significantly higher (164.9±58.7

µg/m3) than that in the low-polluted areas (90.9±34.4 µg/m3) exceeding the Jordanian standard for PM10 of 120 µg/m3 [73].

Kingdom of Saudi Arabia

Air pollution in the Kingdom of Saudi Arabia has recently become a public health concern due to the rapid population growth and the increased economic expansion associated with fuel over-consumption. According to Khodeir et al. [75], there is no published systematic research on the sources of atmospheric PM in Jeddah, the 2nd largest city and the most significant commercial center in the Kingdom of Saudi Arabia with an estimated population of 3.4 million. Desert storms are frequent in Jeddah and contribute to the most of PM in this area [75].

A study was conducted between June and September 2011 to evaluate mass concentrations of PM2.5 and PM10 and their sources in multiple sites in Jeddah using the nondestructive

X-ray Fluorescence (XRF) Spectrometer [76]. The results have revealed that the airborne particulate pollution was high in this city, and that the major sources of PM2.5 or PM10

included diverse industrial activities, fuel combustion and traffic emissions. The overall average mass concentrations were: 28.4±25.4 μg/m3 for PM2.5 and 87.3±47.3 μg/m3 for

PM10, exceeding the WHO annual average values [75].

Another study conducted from June 2006 to May 2007 in the capital Riyadh, using the same methodology as the one described above [76], has determined PM concentrations in the aerosol samples collected on rooftops of buildings from different sectors of the city. Concentration of PM10 was approximately 1 to 4 times higher than that of PM2.5. The

(29)

where mixed industrial activities, mostly ceramic, cement and stone cutting, were the major sources contributing to the local dust. According to this study, it is obvious that the environment of Riyadh was severely polluted [77].

A recent paper by Sun et al. in 2012 has found that more than 1.5-fold significant changes in genes related to oxidative stress, cholesterol and lipid synthesis pathways were triggered following the short- (1 day) and long-term (4-days) exposure of human bronchial epithelial cells to PM10 samples collected from the Kingdom of Saudi Arabia. These changes may contribute to respiratory diseases and CVD related to PM [78].

Kuwait

The environment of Kuwait, a desert country located on the Persian Gulf, was severely polluted from the 1990–1991due to Iraqi invasion and subsequent oil fires [79].

Alolayan et al. [80] have investigated other sources contributing to particulate air pollution in this country. Sand dust and oil combustion power plants were the first 2 major sources contributing to fine particles PM2.5 in Kuwait accounting for almost 54% and 18% of the

total atmospheric PM2.5, respectively. Meanwhile, the petrochemical industry, traffic and

transported emissions from the outside of this country contributed to the remaining 28% of the total PM2.5 [80].

A 12-month study was conducted to measure mean concentrations of PM10 and PM2.5 in 3

areas (2 urban-central and southern- and 1 remote desert-northern) in Kuwait during the period 2004–2005. The results showed that the annual mean PM10 concentration varied

between 66µg/m3 and 93µg/m3 across the 3 areas, exceeding the recommended WHO air quality guidelines. Moreover, the annual mean concentration of PM2.5 ranged from 37

(30)

During the same period, Al Salem [82] analyzed the ambient PM10 levels in Fahaheel, a

typical urban area in Kuwait, using the semi-empirical model described by Grivas et al. [83]. The annual mean concentration was 291µg/m3 and 289 µg/m3 in 2004 and 2005, respectively[82]. It was also mentioned in the published literature that so far there has been no specific study that would explore the link between deteriorating air quality and health problems in Kuwait [84].

Lebanon

(31)

or on top of the buildings. A study conducted in 2012 by Marc et al. [92]concerning the airborne carcinogens levels (particle-bound poly aromatic hydrocarbons, PPAH) when diesel generators were operating and when they were off, on the balconies of 20 residences located in the Hamra area of Beirut. They found that using the generators for a daily length of 3 hours was credited for 38% of the daily carcinogen exposure. This denotes a rise of around 60% in carcinogen levels compared to areas where no generators were found. This would have been elevated in locations where EDL power is cut for more than 3 hours per day [92].

Shaka et al.[93] measured PM concentrations in a costal site in Beirut, the capital of Lebanon, between February and May 2003. Particulate matter 10–2.5 and PM2.5 collected

on Teflon filters [94] were weighed using the Cahn microgram balance model 1500 and analyzed using the Nicolet AVATR Multibounce HATR 360 FTIR spectrometer. In this study, the 4-month average concentrations reported for PM2.5 and PM10 of 39.9 µg/m3 and

118.8 μg/m3, respectively, were very similar to those reported in the Eastern Mediterranean countries but higher than those in the Western Mediterranean ones [93].

Saliba et al. [90] have studied several years of PM measurements and their chemical composition in several sites of the Greater Beirut area (Haret Hreik HH: urban area, post war 2006 construction activities and inner-city site; Bourj Hammoud BH: urban area, close to Beirut harbor and a waste burning facility, and inner-city site; Abdl Aziz, Bliss and Seagate: urban areas and costal sites). PM concentrations were collected on Teflon filters and weighed using the Mettler-Toledo microgram balance (model UMX2).

The average concentrations in different sites of the Greater Beirut area from 2003 to 2007 ranged: 55.1–103.8 μg/m3 for PM10 and 27.6–41 µg/m3 for PM2.5 [90]. The reported levels

(32)

quantities of coarse particles detected in Haret Hreik could be attributed to the construction period that followed the July 2006 conflict [90].

In fact, it has been frequently reported in the published literature that construction activities are a major source of PM10 and not PM2.5 [56, 95-98]. More recently, high annual

concentrations of PM2.5 and PM10 of 20μg/m3 and 64 μg/m3, respectively, were recorded

in the capital Beirut between May 2009 and April 2010 [87].

Qatar

Expanding motor vehicle traffic and industrialization in the state of Qatar constitute the major source of air pollution [99]. The national average concentration values of PM10 in

Qatar had declined from 47 µg/m3 in 2000 to 31 µg/m3 in 2009 [100]. Nevertheless, in the capital Doha, the average PM10 levels in 2010 at some points exceeded the annual limit

recommended by the WHO [60] as a result of the combination of dust from desert storms and mega-constructions activities [100, 101].

A prospective cohort study population was conducted to assess the relationship between daily hospital admissions due to respiratory and CVD and the exposure to daily concentrations of air pollutants - carbon monoxide (CO), sulfur dioxide (SO2), nitrates, ozone (O3) and PM10 at different stations of Qatar during the period 2002–2005. The slight

(33)

Syria

Dry climate and the rapidly growing developing urban population (about 4–5 million) characterize the capital of Syria, Damascus. The main air quality problems in Damascus city are related to the increased use of motor vehicles, particularly ageing ones in service and the poor quality diesel. To report the air quality situation in Damascus within the framework of bilateral cooperation between Syria and Germany from 1999–2000, Meslmani examined 15 selected sites representing different areas of Damascus city. Particulate matter 10 was measured by the use of gravimetric methods using the high volume air sampler (HVAS), TRACERLAB Model MDS-170-257. The resulting 24-h average concentration of PM10 in Damascus varied between 44µg/m3 and 188µg/m3. This

study also reported that PM10 and TSP were the most effective pollutants in the air of

Damascus city [102].

Turkey

Istanbul, the highly populated capital of Turkey, with more than 12 million inhabitants, has demonstrated an excessive urban growth since 1970s [54, 103]. In fact, atmospheric PM is one of the serious concerns of Istanbul where road traffic contributes to the highest proportion of local emissions [97]. Transit route transport from Eastern Europe is of equal importance [104, 105].

The annual mean PM10 and PM2.5 levels collected on the gent stacked filter unit and

measured using gravimetric method [106] between the years 2001 and 2002 in Erdemli, a rural area in Turkey located in the Eastern Mediterranean region, were 36.4±27.8 µg/m3 and 9.7±5.9 µg/m3, respectively [97].

Karaca et al. [107] have measured the annual mean PM10 and PM2.5 concentrations from

(34)

at several Municipality stations in Istanbul. The recorded PM10 value of 47.1 μg/m3 and

PM2.5 value of 20.8 μg/m3 [107] were higher than the WHO recommended levels [60].

Another study by Yatkin and Bayram [108] in 2008 showed that the annual mean of PM2.5

and PM10 in urban Izmir in Turkey during the period 2004–2005 was 64 µg/m3 and 80

µg/m3, respectively. The PM10 and PM2.5 samples were determined using the chemical

mass balance model (CMB) [108].

Also, Theodosi et al. [109] have analyzed the complete chemical composition of different aerosol samples (water-soluble ions, trace metals, water-soluble organic carbon, organic and elemental carbon) collected at the Bogaziçi University sampling station in Bosphorus strait in the Greater Istanbul Area from November 2007 to June 2009. The measured PM10

concentration was 39.1 µg/m3[109].

A supplementary study by Koçak et al. [110] has examined the origin, source and potential impact of PM10 on surrounding regions over the Greater Istanbul Area for the same period.

This study indicates that 80% of PM10 originate from anthropogenic sources, largely fuel

oil combustion, refuse incineration and traffic emissions. The mean PM10 level collected

on the polycarbonate filters and analyzed using the gravimetric method [111], was the highest in the winter (44.5 µg/m3), lesser during the transitional period, and the lowest in the summer (29.8 µg/m3) [110].

United Arab Emirates

(35)

The 2 major urban cities, Abu Dhabi and Dubai, contribute to massive vehicular emissions and traffic congestion [112]. These 2 centers experience extremely high PM concentrations, particularly PM10 that originates mainly from windblown desert dust [115].

The average annual exposure level to urban outdoor PM2.5 in the UAE was estimated to be

80µg/m3 in 2010 [101]. A recent study has analyzed the daily PM10 mass concentrations collected from the Al Samha ambient air quality station, Abu Dhabi between 2007 and 2009. The mean daily PM10 concentration was 172±196 µg/m3 [116] .

According to Ying Li et al. [112], approximately 545 excess premature deaths (95% confidence interval (CI): 132–1224] were attributable to PM in the ambient polluted air in the UAE in the year 2007, accounting for nearly 7% of the total deaths that year [112]. Additionally, 200 annual deaths in the UAE were attributed to PM10 in 2009 [117].

In 2013, Macdonald et al. [118] examined the burden of diseases attributable to 6 environmental exposure routes in the entire UAE population. This study has found that 307 667 health-care facility visits due to CVD in 2008 were attributed to outdoor daily average PM10 with a relative risk of 1.003 (95% CI: 1.0024–1.0036) [118].

1.7 Risk scores for prediction of CVD

(36)

(37)

CHAPTER 2: OBJECTIVES

The purpose of this thesis is to broaden our knowledge regarding the relationship between outdoor air pollution and cardiovascular diseases in the Middle Eastern countries, specifically in Lebanon. Moreover, we aimed to develop a score scale for CVD screening among the Lebanese population.

The specific aims were:

• To assess the levels and sources of PM across the Middle East area and to search evidence for the relationship between PM exposure and CVD

• To investigate the association between outdoor pollutants and cardiovascular diseases among adults in Lebanon and to examine the possible moderator effect of cigarette smoking.

• To develop a score scale for CVD screening in clinical and epidemiological settings, stressing on the importance of adding outdoor air pollution to improve the validity of the scale among the Lebanese adult population

Significance of the research

(38)

CHAPTER 3: METHODS

The papers that constitute this thesis are based on three sources materials. The first is an extensive review of the published literature (2000–2013); conducted using PubMed, Medline and Google Scholar databases. The second was a multicenter case-control study conducted between October 2011 and October 2012. The third study was conducted using two case control studies for the creation of scale and for its clinical validation.

3.1

Outdoor particulate matter (PM) and associated cardiovascular

diseases in the Middle East (Paper I)

(39)

Figure 3.1. 1: Flowchart for identification and inclusion of relevant studies for the review

150 published and online articles, 22reports and 2 books reviewed using PubMed, Medline and Google

Scholar

39 articles and 4 reports excluded: •not carried out on human subjects e.g. animal studies

•pollutants other than PM

18 reports 2 books 111 articles

(40)

3.2

Association between outdoor air pollution and CVD: Case-control

study (Paper II)

3.2.1 Study design and population

It is a multicenter case control study, conducted between October 2011 and October 2012, in six hospitals in Lebanon (American University of Beirut Medical Center, Hotel Dieu Du France Hospital, Saint Joseph Hospital, Monla Hospital, Nabatieh governmental Hospital and Bekaa governmental Hospital), comparing CVD cases to a control group. Cases were defined as patients aged 40 years or above, hospitalized for cardiovascular disease, diagnosed with ST/non-ST elevation – myocardial infarction, stable/ unstable angina or heart failure, confirmed by a cardiologist based on their clinical presentation and laboratory exams [130]. The control group consisted of any subject aged 40 years or above admitted to the same hospitals in the same period for reasons excluding diabetes, hypertension, dyslipidemia, respiratory problems or cardiovascular diseases. All participants signed an informed consent before enrolment in our study. Only 3 cases and 7controls refused to take part in our research.

3.2.2 Data collection

(41)

categorized into quartiles (low, medium low, medium high and highest income ). Health and behavior habits questions were included. For instance, cases and controls were asked to evaluate their health condition: “how would you describe your current health status” on a 10 point scale. Smoking status was also assessed for cigarette and water pipe smoking separately. Due to the small number of previous cigarette smokers, this sub-goup was combined with current smokers. Passive smoking at home and at work was also weighed. On the subject of pollution, the traffic exposure indicator was assessed by the following question: “Are you living near a busy highway (<100 meter,>100 meters)?” The diesel emission exposure indicator was evaluated by the question: “Are you living close to local diesel generator”? : Diesel generators installed on balconies, in basements,over building entrances, on ground-level building service areas, over the side roads and in open parking lots). Traditional cardiovascular risk factors (CVRFs) were also collected from the hospital charts: hypertension, triglyceride, HDL (high density lipoprotein concentration), LDL (Low density lipoprotein concentration), family history of cardiovascular disease (CVD) as well as obesity according to Body Mass Index scale (BMI) "normal weight" (BMI 18.5– 24.9 kg/m2), "overweight" (BMI 25.0–29.9 kg/m2) and "obese" (BMI ≥ 30 kg/m2) [131].

3.2.3 Sample size calculation

Sample size calculation was done with a type I error of 5%, and a study power of 80%. In the absence of baseline data, the exposure of healthy Lebanese residence to pollution was considered to be equal to 50%. The minimal sample size necessary to show a twofold increase in risk of cardiovascular diseases consisted of 330 subjects, divided into 111 cases and 222 controls. In the present study the simple size was 340 divided into 121 cases and 219 controls.

3.2.4 Statistical analysis

(42)

analysis was performed using Statistical Package for the Social Sciences SPSS IBM (version 20.0). Means with their 95% confidence intervals, median with their interquartile ratio and percentages were used to describe continuous and categorical variables. Statistical bivariate analysis was performed. The Pearson chi-square (χ2) test was used for categorical variables. The student T-test and Mann-Whitney Utest were run for the continuous variables to compare their means. A p value < 0.05 was considered statistically significant. A multivariate analysis using logistic regression was carried out with CVD as the dependent variable. Adjusted odds ratios and their 95% confidence intervals were reported. The final logistic regression model was reached after ensuring the adequacy of our data using the Hosmer and Lemeshow test. Furthermore, due to the presumed relationship of cigarette smoking with both CVD [25] and exposure to outdoor air pollution [26], the interaction effect of this variable was tested (P value<0.001). The regression model was stratified according to cigarette smoking status.

3.3

Validation of screening scale for CVD in clinical settings (Paper III)

3.3.1 Study design and population

(43)

group included any subject aged 40 years or above hospitalized for reasons excluding diabetes, hypertension, dyslipidemia, respiratory problems or CVD. There were no significant differences in terms of age and gender between the two groups with P value > 0.05 in both samples.

3.3.2 Data collection

Subjects were interviewed by independent assistant after obtaining the informed consent. The ethical committee of our university waived the need for an official approval due to the observational nature of this study. We collected baseline data based on socio- demographic variables such as age, gender (male/female), smoking status (current/ never) and family history of cardiovascular disease (CVD). Hospital charts were used to collect variables such as systolic blood pressure (SBP), triglyceride, high density lipoprotein concentration (HDL) and Low density lipoprotein concentration (LDL). Regarding the outdoor air pollution matter, we published a recent study conducted in Lebanon revealing that outdoor air pollution such as living near busy highway (<100 meters, >100 meters) and living close to a local diesel generator (no/yes) was significantly associated with CVD [132].

3.3.3 Statistical analysis

(44)

(Are you living close to local diesel generator?) (no:0/yes:1) : Diesel generators installed on balconies, in basements, over building entrances, on ground-level building service areas, over the side roads and in open parking lots. Two multivariate analyses using logistic regression were carried out to evaluate predictors of the dependent variable. The second model included the outdoor air pollution variables while the first did not. Adjusted odds ratios and their 95% confidence intervals were reported. The Hosmer and Lemeshow goodness of fit was also calculated to assess the model fitting to data. The regressions of predictors served to generate two screening scales for CVD: the adjusted odds ratios (OR) obtained were rounded to the nearest units and used as coefficients in the generated scales.

Thus, we compared two risks screening models, the first composed of the following risk factors (Score1): class age (40-65/ >65), gender (male/ female), systolic blood pressure (no/yes), smoking status (current /never), triglyceride, HDL (no/yes), LDL (no/yes), and family history of CVD (no/yes). In the second model, we used the same risk factors of (Score1) in addition to risk factors related to outdoor pollution derived from previous study ( living near a busy highway and living close to local diesel generator) [132] (Score2). The calculated scores were applied for validation.

(45)

CHAPTER 4: RESULTS

Below is a summary of the main results from each study. Additional information and more comprehensive descriptions of the study results are given in paper (I, II and III).

4.1 Outdoor particulate matter (PM) and associated cardiovascular

diseases in the Middle East (Paper I)

PM levels in the Eastern Mediterranean region are much higher than in other regions, even when compared to the Western Mediterranean region. Although there is enough evidence on the association between the outdoor air pollution and CVD in the developed countries, the studies investigating this association in the Middle East are rather limited. Our findings enabled to point out only selected countries of the Middle East in this article review. The current review summarizes atmospheric PM levels and the major sources of air pollution affecting environment in the selected countries of the Middle East, and it highlights their potential association with CVD. Findings from this review manifest elevated PM concentrations in these countries, often exceeding the 2006 WHO annual average guidelines (PM2.5 10μg/m3, PM10 20 μg/m3). This could be explained by several factors

such as: high population density, rapid urbanization, dense traffic area, fossil fuel use, geographical settings of the region, frequent dusts outbreaks, temperature inversion during cold seasons, and the lack of rules and regulations concerning the reduction of emissions from anthropogenic (mobile and stationary) pollution sources.

(46)

middle-eastern countries (Egypt, Jordan, Kuwait, Lebanon, Syria and Turkey) disclose a shortfall in reporting any associations between PM and cardiovascular diseases.

The results of annual PM10 concentrations, main sources of the outdoor air pollution and

(47)

Table 4.1. 1: Annual Mean outdoor PM10 levels, major sources of air pollution and presence of association with cardiovascular diseases (CVD) in selected countries of Middle East. Review of the published literature 2000-2013.

Country Site Study _period Annual Mean _PM

10 (µg/m3)* Major sources of air pollution Presence of association with CVD

Egypt

Greater Cairo 2001-2002 170[135]

- Vehicle emission, traffic, industry, open burning , dust and sand storms [135]

- Thermal inversion [136]

Iran

Isfahan 2009-2010 120.48[71] _{Vehicle emission, temperature} inversion in cold seasons[137]

- Potential impact of PM on platelet activation and atherosclerosis [69]

- Changes in heart rate variability as a marker of cardiac autonomic dysfunction in polluted air conditions [70] - 1367 average extra deaths and 2580 extra cases admitted

to hospital due to CVD annually attributed to short exposure to PM10[68]

Tehran 2010-2011 90.58[68]

Jordan

Amman : high polluted

area 2010

164.9[73]

Vehicle emission[138] Amman : low polluted area 2010 90.9[73]

Kingdom of Saudi Arabia (KSA)

Jeddah 2011 87.3 [139] -Vehicle emission, traffic, dust sources, industry oil combustion _[139] - Development activities[140]

More than 1.5-fold changes in genes related to oxidative stress and cholesterol and lipid synthesis pathways were triggered following the short- and long- term exposure of human bronchial epithelial cells to PM10 [78]

South eastern Riyadh 2006-2007 597.2 [140] Kuwait Fahaheel 2005

289 [141] Traffic, oil combustion, petrochemical industry, power plant [80]

Lebanon 3 Greater Beirut areas: _{Haret Hreik, Borj} Hamoud& Bliss

2003-2007 55.1-103.8[142]

- Vehicle emission [89, 142] - Construction, waste burning[142] - Dust storms[87, 142]

Greater Beirut 2009-2010 64 [87]

Qatar National average 2009 31[72]

- - Vehicle emission,traffic, industrialization[99] -

- Duststorms, construction activities[74]

A slight increase in the concentration of air pollutants in 2005 accompanied by a slight increase in the daily admissions from the respiratory, ischemic heart diseases and CVD [99]

Turkey Erdemli 2001-2002 36.4 [143]

- Vehicle emission, traffic [143] - Transit route transport [144] - Refuse incineration, solid fuel [145] Istanbul 2002-2003 47.1 [146]

Urban Izmir 2004-2005 80 [147] United

Arab Emirates (UAE)

2010 - _- - Vehicle emission, traffic, industry[148] _{- Desert dust[67]} 307,667 health care visits due to CVD attributable to _{ambient daily average outdoor PM10[118]}

(48)

4.2 Outdoor air pollution and cardiovascular diseases (CVD) in Lebanon:

A case-control study (Paper II)

4.2.1Characteristics of the study sample

(49)

Table 4.2. 1: Baseline characteristics of the participants. Characteristics Controls n=219 (%) Cases n=121(%) P value Age 40-44 45-49 50-54 55-59 60-64 ≥65 43(19.7) 25(11.5) 16(7.3) 16(7.3) 28(12.8) 90(41.3) 7(5.9) 11(9.2) 18(15.1) 10(8.4) 20(16.8) 53(44.5) 0.007* Gender Males Females 116(53.0) 103(47.0) 63(52.5) 58(47.5) 0.902 Marital status Single Married Divorced or Widow 36(16.4) 157(71.7) 26(11.9) 1(0.8) 104(86.7) 15(12.5) <0.001* Education Illiterate Primary or less Complementary or less Secondary or less University degree 23(10.5) 82(37.4) 48(21.9) 37(16.9) 27(13.2) 11(9.7) 60(53.1) 25(22.1) 10(8.8) 7(6.2) 0.025*

Income per- family -member Low Med Low Med high Highest 58(27.0) 27(12.6) 76(35.3) 54(25.1) 35(30.2) 23(19.8) 35(30.2) 23(19.8) 0.221 BMI a Normal (BMI 18.5-24.9) Overweight(BMI 25-29.9) Obese (BMI≥ 30) 71(32.4) 102(46.4) 46(21.0) 22(19.0) 53(45.7) 41(35.3) 0.004* Family history of CVD b No Yes 143(65.9) 74(34.1) 62(51.7) 58(48.3) 0.012* Triglycerides mg/dl (Mean ±SD)c 146.5±43.6 202.2±67.0 <0.001* LDL mg/dld(Mean ±SD)c 86.4±46.3 128.9±38.3 <0.001* HDL mg/dle(Mean ±SD)c 51.7±15.6 43.3±14.1 <0.001* SBP mmHgf(Mean ±SD)c 122.5±15.04 129.9±19.1 <0.001* a

BMI Body mass index, b Cardiovascular disease, c Mean ± standard deviation, d LDL low density

(50)

4.2.2 Association between smoking and CVD

The results of the bivariate analysis for smoking exposure (active or passive) and cardiovascular diseases are presented in table 4.2.2. Current cigarette smokers had significantly higher risk of CVD than non- smokers with an Odds Ratio (OR) of 1.92 and a 95% Confidence Interval (CI) between 1.22 and 3.01. A positive association was also found with passive smoking at home OR 2.35, 95% CI (1.46-3.78). Regarding water pipe and passive cigarette smoking at work no significant association was pronounced.

Table 4.2. 2: Subject's exposure to active, passive smoking and CVD

Variable Controls

n=219 (%)

Cases n=121(%)

OR (95%CI) Cigarette smoking status

Non smokers Current smokers 137(63.1) 80(36.9) 57(47.1) 64(52.9) 1.92(1.22-3.01)

Passive cigarette smoking at home No Yes 106(51.2) 101(48.8) 37(30.8) 83(69.2) 2.35(1.46-3.78)

Passive cigarette smoking at work No Yes 21(32.3) 44(67.7) 15(33.3) 30(66.7) 0.95(0.42-2.14)

Water pipe smoking status Non smoker Current smoker 170(78.0)(9.6) 48(22.0) 100(82.6) 21(17.4) 0.74(0.42-1.31)

OR: Odds ratio, CI: Confidence interval

4.2.3 Association between outdoor pollution, cumulative exposure and CVD

(51)

Table 4.2. 3: Exposure to outdoor air pollution and CVD Variable Controls n=219 (%) Cases n=121(%) OR (95%CI) Highway proximity a >100 m <100 m 128(58.4) 91(41.6) 47(38.8) 74(61.2) 2.21(1.40-3.48)

Living duration near highwayb Never 1to 14 years 15to30 31years or more 143(65.3) 18(8.2) 26(11.9) 32(14.6) 53(4.38) 11(9.1) 28(23.1) 29(24.0) 1.00 1.14(0.73-3.72) 1.51(1.12-4.28) 2.28(1.62-5.77)

Local diesel generator proximityc No Yes 166(75.8) 53(24.2) 61(50.4) 60(49.6) 3.08(1.92-4.93)

Living duration near local diesel generator d Never

1 to 10 years 11 years and more

166(75.8) 28(12.8) 25(11.4) 61(50.4) 29(24.0) 31(25.6) 1.00 2.81(1.55-5.11) 3.37(1.84-6.16) a_{Are you living near a busy highway,}b_{Duration of living near a busy highway,}c _{Are you living close tolocal}

diesel generator,d Duration of living close to local diesel generator, OR: Odds ratio, CI: Confidence interval.

4.2.4 Multivariate Analysis

(52)

results revealed a clear difference between non- smokers and current smokers OR 4.6, 95% CI (1.10-19.25), OR 10.11, 95% CI (7.33-20.23) respectively. However, no interaction was found while examining any possible effect of cigarette smoking status on the association between living close to diesel generator and CVD (P for interaction: 0.24) (Table 4.2.5).

Table 4.2. 4: Adjusted odds ratios with their 95% confidence intervals from the logistic regression of CVD among cases and control

Global Model OR a 95% CI P value

Living near a busy highway <100m

>100m

1.0

5.04 4.44-12.85

<0.001*

Living close to local diesel generator

4.76 2.07-10.91 <0.001*

Dose-effect relationship** Duration of living near a busy highway

1.05 1.02- 1.07 0.016*

Duration of living close to local diesel generator

1.38 1.10-1.78 0.014*

Global Model: Adjusted for triglyceride, HDL, LDL, SBP, cigarette smoker status. ORa: adjusted odds ratio,

CI: confidence interval, *P value <0.05 statistically significant. ** Dose effect relationship: Adjusted for triglyceride, HDL, LDL, SBD, cigarette smoker status. ORa: adjusted odds ratio, CI: confidence interval, *P

value <0.05 statistically significant.

Table 4.2. 5: Adjusted odds ratios with their 95% confidence intervals from the logistic regression of CVD among cases and control stratified by cigarette smoking status

Exposure type

Cigarette smoking

Non smokers Current smokers

OR a 95% CI P value OR a 95%CI P value

Living near a busy highway <100m >100m 1.0 4.60 1.10-19.25 0.030* 1.0 10.11 7.33-20.23 <0.001*

Living close to local diesel generator

4.97 1.64-15.08 0.005* 5.02 1.52-11.93 0.004*

(53)

4.3 Outdoor air pollution improves the validity of a Screening Scale for

cardiovascular disease (CVD) in clinical settings (Paper III)

4.3.1 Construction of the scales (Model 1 and Model 2)

The baseline characteristics of our study population are listed in table 4.3.1. First sample analysis, included 340 patients: 219 (64.4%) controls and 121 (35.6%) cases. The means of the risk factors (hypertension, triglycerides and LDL) were significantly higher among cases than controls.

Table 4.3. 1: Risk factor profile of the participants in the case control study

characteristics Controls

n=219(64.4%)

Cases

n=121(35.6%)

P value Age years (Mean±SD) 59.94±14.53 62.94±13.05 0.06

Gender Male Female 116(53.0) 103(47.0) 63(52.5) 58(47.5) 0.902

Current cigarette smoker No Yes 137(63.1) 80(36.9) 57(47.1) 64(52.9) <0.001* Family history of CVDa No Yes 143(65.9) 74(34.1) 62(51.7) 58(48.3) 0.02* SBPb mm/Hg(Mean±SD) 122.5±15.0 129.9±19.1 <0.001* Triglycerides mmol/L(Mean±SD) 1.65±0.49 2.28±0.75 <0.001* HDLcmmol/L (Mean±SD) 1.33±0.40 1.12±0.36 <0.001* LDLdmmol/L (Mean±SD) 2.23±1.20 3.33±0.99 <0.001* Highway proximitye <100m >100m 128(58.4) 91(41.6) 47(38.8) 74(61.2) 0.001*

Local diesel generator proximityf No Yes 166(75.8) 53(24.2) 61(50.4) 60(49.6) <0.001*

*P value <0.005, (Mean±SD) Mean ±Standard deviation, aFamily history of cardiovascular diseases,

b

(54)

Table 4.3.2 shows two logistic regressions models predicting CVD in this case control study. Taking into account the adjusted OR and rounding to the nearest unit, Two different scores were computed as follows: Score 1= Class age*2 + cigarette smoker*2+ triglyceride*5 +LDL*5+SBP *3+ HDL) and Score 2 = Class age*2 + cigarette smoker *2 + triglyceride*6 +LDL*5+ SBP*3+ HDL + living near high way *2 + living close to a local power plant *3). The Hosmer and Lemeshow goodness of fit for both models suggest good calibration 0.64 and 0.94 respectively. The first score has a minimum of 2 and maximum of 20 and second score has a minimum of 2 and maximum of 26. In the first model, the mean of CVD individuals was 13.26 and its standard deviation 3.40, while in healthy controls, the mean was 7.83 and the standard deviation 4.1 (P <0.001). In the second model, the mean of CVD individuals was 16.75 and its standard deviation 3.98, while in healthy controls, the mean was 9.75 and the standard deviation 4.66 (P <0.001).

Table 4.3.2: Logistic Regression for predicting of cardiovascular disease events among the participants. Model 1= traditional risk factors (TRF) of CVD and Model 2 = TRF of CVD in addition to outdoor air pollution exposure assessment.

Model 1 Scale 1

Model 2 Scale 2

OR 95%CI P value OR 95%CI P value Class age 1.84 1.01-3.38 0.049* 2.19 1.13-4.24 0.020 SBPa 2.92 1.26-6.75 0.012* 3.25 1.33-7.97 0.010* Triglycerides 5.18 2.71-9.88 <0.001* 5.85 2.95-11.63 <0.001* HDLb 0.48 0.25-0.92 <0.027* 0.39 0.19-0.78 <0.001* LDLc 4.79 2.49-9.24 <0001* 5.02 2.47-10.18 <0.001* Cigarette smoker 2.19 1.20-3.98 0.01* 2.41 1.43-5.97 0.007* Highway proximityd 2.42 1.27-4.56 0.011* Generator proximitye 2.92 1.22-4.81 0.003*

*P value<0.05, OR: adjusted Odds Ratio, CI: Confidence Interval, aSystolic Blood Pressure,bHigh density lipoprotein concentration, cLow density lipoprotein concentration, dLiving near highway <100 meters,

(55)

4.3.2 Scales properties and thresholds

Receiver-operating characteristic (ROC) curves for CVD screening are shown in figure 1 comparing CVD patients with controls for the 2 scales. The areas under the curve were 0.737 (0.692-0.882; P < 0.001) and 0.864 (0.825-0.903; P<0.001) respectively.

According to the ROC curve of the first scale (without outdoor air pollution; figure 1), the threshold that gave the best sensibility and specificity was 9.5: Se= 80% and Sp= 60%. After applying this threshold, we obtained a concordance between CVD screening score and CVD patients confirmed by physician: Kappa = 0.412. Individuals with a positive score have a possibility of being a true CVD: OR= 2.81 [1.69 - 4.67].

Université Libre de Bruxelles and Lebanese University The Doctoral School of Sciences and Technology

Co-supervision thesis