Death from Cardiovascular Diseases and Air Pollution in Shiraz , Iran ( March 2006-March 2012 )

Background: There is evidence that shows exposure to air pollution might be related to cardiovascular death. This study aimed to estimate the effect of ambient air pollutants on cardiovascular deaths in Shiraz, Iran. Methods: In this ecological study, we inquired data about the number of cardiovascular deaths by gender and age; and recorded air pollutants data including nitric oxide (NO), carbon monoxide (CO), nitrogen dioxide (NO2), nitrogen oxides (NOx), particulate matter (PM10), sulfur dioxide (SO2), ozone (O3), methane (CH4), total hydrocarbons (THC) and nonmethane hydrocarbons (NMHC) from 2006-2012. The association between air pollution and cardiovascular deaths was investigated by time series, cross correlations with and without lags with crude and adjusted negative binomial regression analysis through Minitab16 and Stata 11. Results: During this period, 17167 cardiovascular deaths occurred in Shiraz. After adjustment for confounding factors, NO and NOX were significantly related to total cardiovascular deaths; NO, NOX and SO2 were significantly related to deaths in women and SO2 was significantly related to deaths in the age of 18 to 60. NO and NOX pollutants were related to all cardiovascular deaths that happened one month later. Conclusion: The results of this study confirm the results of many other studies which show that air pollution can increase cardiovascular death.


Introduction
The effects of air pollution on human health have been a major environmental concern.Even before modern studies, events such as the sudden increase of death rates following a short interval of particulate matter and sulfur dioxide increase in London, in 1952 attracted much attention [1].Recent evidence indicates that long term exposure to air pollution from traffic may lead to preterm death [2] and coronary artery disease [3].The most important probable adverse acute and choric effects of air pollution include a variety of chronic heart diseases such as palpitation, changes in heart rate, blood pressure, artery constriction, blood coagulation and atherosclerosis [4].Some epidemiological evidences indicate that air pollution is related to cardiovascular disease exacerbation [5].Studies have shown that patients with congestive heart failure [6] or a history of myocardial infarction or diabetes [7] are at greater risk of death on days with higher levels of air pollution.Also survivors of myocardial infarction [8] and people with a history of myocardial infarction in exposure to air pollution are at more risk of reoccurrence when exposed to air pollution [9].
Numerous epidemiological studies have shown that exposure to fine particles such as PM 2.5 is related to cardiovascular diseases [10,11].In the six cities studies in America, there were significant positive associations between pulmonary and cardiovascular deaths with increased concentrations of ambient particulate matter [12].
Miller et al [13], in the Women's Health Initiative (WHI) study in America found a significant increase in the occurrence and fatality of cardiovascular outcomes by increase in exposure to PM 2.5 .Researchers using data from the Nurses' Health Study in America found that long term exposure to PM 2.5 is related to death due to coronary heart disease (CHD) [14].Chuang et al [15], in a cross sectional study on 1023 of elderly people in Taiwan reported that both systolic and diastolic blood pressure were sharply related with annual average levels of various pollutants (PM 10 , PM 2.5 , O 3 , NO 2 ), however some other studies did not report any relation [16].
According to studies done in Tehran, 70% of deaths in Tehran result from respiratory and cardiovascular problems and these problems are directly or indirectly related to air pollution.Central areas of Tehran have the highest pollution and death; and deaths from cardiovascular disease have shown more sensitivity to air pollution compared with other diseases [17].
Shiraz is one of the major cities of Iran, and has had a high population growth.It doesn't have modern transportation facilities such as subway lines.Private vehicles are the dominant mode of transport.There is a large number of vehicles and heavy traffic in some parts of the city.Also there are factories and industrial workshops that increase the amount and variety of pollutants in this city.Shiraz is located between mountains which block air exchange and is one of the eight major cities in Iran tackling air pollution issues [18].
Due to the increasing development of industry and increasing vehicles which have led to heavy traffic in cities, atmospheric pollutants are increasing in the world and its very apparent signs are destruction of the atmospheric ozone layer and global warming.Due to the importance of increasing cardiovascular disease death and its probable relation with air pollution researchers conducted this study in Shiraz, Iran.Few studies about this topic have been conducted in Iran [19].

Department of Health from March 2006 -March 2012 (six years).
Cardiovascular deaths included recorded deaths resulting from myocardial infarction, stroke, high blood pressure, arterial embolism, thrombosis, aortic aneurysm, dissecting aneurysm, pulmonary embolism, vascular diseases, other heart diseases, cardiovascular diseases, non-rheumatic mitral and aortic valve disorders, acute and sub-acute endocarditis, acute pericarditis, acute myocarditis, cardiomyopathy, heart failure and cardiovascular congenital malformations.The data was anonymous.
The health data was inquired from the Shiraz University Deputy of Health.The reason for death in each individual is recorded based on the death certificate signed by a physician.This information was released from the Shiraz University Deputy of Health after formal communication and acquiring permission from the authorities.All deaths in this city have to be recorded by law.The environmental data was measured by air pollution measuring stations working under the supervision of the Shiraz Province, Environmental Protection Agency.The environmental exposure data and the death data were all from the city of Shiraz.As this was an ecological study and the level of available data was limited, we were not able to match the exact place of residence with the air pollution data.
The authors of this study were not involved in measuring the pollutants.The pollutants are routinely measured and recorded by the Province Environmental Protection Agency (EPA).The air pollutants which were measured at the Shiraz Environmental Protection Agency included nitric oxide (NO), carbon monoxide (CO), nitrogen dioxide (NO 2 ), nitrogen oxides (NO x ), particulate matter with aerodynamic diameter ≤ 10 μm (PM 10 ), sulfur dioxide (SO 2 ), ozone (O 3 ), methane (CH 4 ), total hydrocarbons (THC) and nonmethane hydrocarbons (NMHC).
The mean daily temperature, relative humidity, rainfall, wind speed and direction were inquired from the Shiraz Meteorology Department from March 2006-March 2012 and were used to adjust for these confounding factors.

Statistical Analyses
The relationship between cardiovascular death and pollutants was investigated in six different gender and age groups which were; all people, men, women, children under 18 years old, adults from 18-60 years old and adults above 60 years old.The population group sizes used in the negative binomial regression were inquired from the Statistical Center of Iran based on the 2006 to 2011 (www.amar.org.ir)data and the years in between were calculated based on 0.03% population growth.
Descriptive results were reported.Also in order to determine the relationship between cardiovascular deaths and pollutants; negative binomial regression analysis was used and the incidence rate ratio (IRR) were computed in univariate and multivariate models.The multivariate model included the dependent variable (cardiovascular deaths), and the independent variables (contaminants and confounding variables temperature and humidity, rainfall, wind direction and speed).
The NO X variable in this study is the sum of NO and NO 2 , and the THC variable is the sum of CH 4 and NMHC and these variables were not used simultaneously in multivariate analyzes.
Air pollution is usually seasonable.There is more CO 2 pollution in the winter and more O 3 in the summer.We used the complete air pollution dataset for the mentioned years.We did not limit our analysis to specific months.
One month, lag times were computed for pollutants and their associations with cardiovascular mortality was determined by Pearson or Spearman correlation coefficients.The data was not normalized.We performed Pearson correlation for normal and Spearman correlation for non-normal data.However, most studies that tested the delayed health effects of contaminants observed the effects in delays of a few days and less than a month [20,21].Therefore, in this study we only examined the correlation of onemonth delay with cardiovascular deaths.Data were analyzed data through Minitab 16 and Stata 11.

Results
Table 1 shows the descriptive statistics of air pollutants under study in Shiraz during the years 2006-2012.Table 2 shows the number of cardiac deaths during 2006-2012.
The total number of cardiovascular deaths in Shiraz during March 2006 to March 2012 was 17167.Table 3 and 4 shows the results of crude and adjusted (for temperature, humidity, rainfall, wind speed and wind direction) negative binomial regression analysis investigating the relation between the monthly average of pollutants with total cardiovascular deaths in different gender and age groups.The NO X pollutant is a sum of NO and NO 2 and THC is the sum of CH 4 and NMHC.Due to the existence of over-dispersion (variance greater than mean) negative binomial regression analysis was used rather than Poisson regression to analyze the data in this study.Results of the adjusted analysis showed that nitrogen oxides remained significant in all people, while nitrogen oxides and sulfur dioxide were significant in women.
Results of multivariate analysis showed that dioxide sulfur remained significant in people 18-60, but there was not any significant relation between cardiovascular death and other pollutants.
Table 5 shows the correlation coefficient results between air pollutants and cardiovascular deaths happening one month later.
Results showed that there is a significant relation between increase of cardiovascular deaths and pollutants such as nitrogen oxides, CO, O 3 and SO 2 .

Discussion
Air pollution can affect the risk of cardiovascular diseases through mechanisms such as systemic inflammatory and oxidative stress responses, and can lead to increasing atherosclerosis and long-term effects.It can also cause consequences and side effects through changes in vascular function, coagulation, plaque stability and autonomic balance [5].It has been reported that exposure to air pollution causes increased plasma viscosity and shortened prothrombin time [22].Animal studies have also shown that exposure to particles increase atherosclerosis [23].However, in order to explain the positive statistical relation between death due to chronic ischemic heart disease and aerosol exposure, scientists require more information about the biological mechanism and their adverse effects [24].Some studies assume that nitrogen dioxide is a causal factor for thrombosis [25].In contrast, the known vasodilating effects of NO and mixtures of PM are highly variable and may have different physiological effects depending on the type of dominant compounds [26].On the other hand, comparing the relation of death outcomes and exposure to air pollution is difficult, in different studies.The wide disparity in dose-response in these studies is probably related to differences in design, condition, exposure assessment method, measurement scales [27], underlying health status of the population studied, particle composition, relative toxicity, measurement and control of potential confounding factors [28] or differences in exposure sensitivity to pollutants [29].
In an ecological study done by Maheswaran et al [3], in Sheffield, UK, a relation between PM 10 and increase death from coronary heart disease was found.A study in America has shown that exposure to particle matter (PM) is related to increase heart attacks and stroke [5].Also there was a relation between death from IHD and NOX and PM 10 exposure [28].In our study, there was not a significant relation between PM 10 in all people, men, women and other age groups with cardiac deaths.This result can be due to low levels of PM 10 in Shiraz.In this study only 370 days (17%) of the 2190 days studied in these six years had PM 10 more than the permissible 24 hour limit which is 150 μg/m 3 [30].In a similar study from Kerman, Iran although the median level of PM 10 was 118.1 μg/m 3 , no significant relation was observed between PM 10 and cardiovascular death either [19].In a study done in Sheffield, England, there was a significant association [rate ratio of 1.17 (95% CI 1.06-1.29)]between nitrogen oxides and excess risk of death from coronary heart diseases [3].In another study, the elderly were more susceptible to the effects of air pollution.The adjusted rate ratio (RR) for nitrogen dioxide in ischemic stroke in all ages was 1.11, in 40-64 year olds was 1.13 and in 65-79 year olds was 1.23.This result had borderline statistical significance in the 65-79 year age group and no significance in the other age groups [25].In a cohort study in Canada, a significant increase of 40% was observed in the risk of death due to circulatory diseases in exposure to nitrogen oxides [2].In the study by Raaschou-Nielsen et al [31], in Denmark, mortality rate ratios (MRRs) in cardiovascular diseases was equal to 1.26 and showed a significant 26% increase in cardiovascular death per doubling of NO 2 concentration after adjustment for confounding factors.However, Filleul et al's [32], study in France, showed no relation between death from cardiopulmonary diseases and NO 2 and the adjusted risk ratio was 1.00 (95% CI 0.98-1.02).In our study, the results of multivariate analysis indicated that nitrogen oxides had a positive relation with cardiovascular death in some groups.The reason might be that the average of Nitrogen oxides (NO X ) was higher in Shiraz than the healthy threshold which is 0.021 ppm [33].One of the main sources of NO X in Shiraz is motor vehicles.One of the other sources is natural gas [34] which is used for cooking and heating in Iran.Women might be more exposed and more sensitive to this pollutant, because they are more involved in cooking or using heating devices at home.In a study in California, the relation between death due to ischemic heart disease (IHD) and ozone was of borderline significance (HR, 1.06; 95%CI, 0.99-1.14)but when the analysis was limited to summer, the risk of death due to ischemic heart diseases increased significantly.However, there was evidence that showed the positive results for ozone and death from ischemic heart diseases may be due to the confounding effects of PM 10 [28].In other studies, the relation between cardiovascular admissions in hospitals with ozone in warm months was inconsistent and negative significant associations were observed in Hong Kong [35] and positive significant associations were observed in London [35].Another study in Five European Cities showed positive significant associations between myocardial infarction readmissions and ozone [36].Dadbakhsh et al's study in Shiraz, Iran showed that respiratory mortality in individuals over 60 years was positively related to ozone [37].In study of Hashemi et al's [19] in Kerman, Iran showed only significant correlation between ozone and female cardiovascular mortality (r = 0.31).However other studies didn't find an association between ozone and cardiovascular death [12,38,39].Also in our study, the results of multivariate analysis showed that the relation between ozone and cardiovascular death was not significant.These inconsistencies between studies require further investigation.The insignificance of ozone in the present study might be due to low levels of ozone (0.017 ppm) in Shiraz in comparison to the healthy threshold (0.05 ppm) [33].Studies reporting higher levels for ozone did not find a significant health effect [12,38,39].Europe's multi-center study on survivors after their first myocardial infarction showed that the rate ratio (RRs) of cardiac readmission for carbon monoxide was equal to 1.014 for each 200 μg/m 3 increase and statistically nonsignificant [36].In another ecological study in Sheffield, UK, there were borderline significant associations between carbon monoxide and death from coronary heart disease [3].However a study done in Shiraz, Iran showed that respiratory mortality in total, males, females, and individuals aged between 18 to 60 years was positively correlated with CO in ambient air [37].In our study, the results of multivariate analysis didn't show a significant relation between carbon monoxide and cardiovascular deaths which can be due to low levels of carbon monoxides in Shiraz.In this study in only 38 days from 2190 days (which is 1%) of the days under study, the daily CO level was above the 8 hour permissible level for CO which is 9 ppm [30].
In a study from Zhang et al [40], in Beijing, there was a positive significant relation between long term exposure to SO 2 and death due to all causes and death from cardiovascular diseases.However there was not a significant relation between death from cerebrovascular diseases and SO 2 .In a study from Filleul et al [32], in France, the adjusted hazard ratio for cardiovascular death from sulfur dioxide SO 2 was not significant either.In another study in California, in the adjusted analysis, the relation between sulfur dioxide and death due to cardiovascular diseases was reverse and not significant [41].In our study, sulfur dioxide had a positive significant relation with cardiovascular deaths in women and in the 18-60 age groups.The reason for the significant association between SO 2 and cardiovascular deaths in the present study is probably the higher levels of SO 2 .In this study the average monthly levels of SO 2 was 101.53 ppb, but in Chen et al's study which did not observe a significant association it was 4.5 ppb (Chen et al, 2005).
Another factor that seems to be important in studies of air pollution is gender.The study of Chen et al [41], in California showed that PM 2.5 was positively associated with fatal cardiovascular diseases in females, but not males.Also in another study in the US done only on males, no relation was observed between PM 2.5 and death from CHD [42].However the results of other studies did not show any difference between two genders [31,38,43].Recent studies in America have reported a relation between PM and higher risk of death due to CHD in women [13].In our study, in women, sulfur dioxide and nitrogen oxides had a significant relation with cardiovascular death but such a relation was not observed in men.In a study in Kerman, there was also a relation between respiratory deaths in men and ambient ozone and sulfur dioxide; but this relation was not observed in women [44].
In this study, pollutants such as nitrogen oxides, O 3 , SO 2 and CO were associated with cardiovascular deaths happening one month later.This association may be related to the delayed effects of these pollutants and needs further investigations.
As this study was an ecological study, a limitation was that the results cannot be directly inferred to the individual level.Also we were not able to adjust for migrations and population movements.

Conclusions
Evidence about the risk of death due to cardiovascular diseases and air pollution highlights the importance of interventions to reduce air pollution especially in areas with high air pollution.In most Iranian cities including Shiraz the main source of air pollution are motor vehicles.

Table 1 :
Level of pollutants in Shiraz during the years 2006 to 2012 ( a parts per billion, b Micrograms per Cubic Meter of air, c parts per million carbon).

Table 2 :
The number of cardiovascular deaths in Shiraz city during March 2006 to March 2012.

Table 3 :
Results of crude and adjusted negative binomial regression, and the effect of pollutants on cardiovascular deaths (ratio of increase in death in month per unit of increase in pollutants monthly average) in total, men and women.(*IRR=Incidence Rate Ratio; † Statistically Significant)

Table 4 :
Results of crude and adjusted negative binomial regression, and the effect of pollutants on cardiovascular deaths (ratio of increase in death in month per unit of increase in pollutants monthly average) in different age groups.(*IRR=Incidence Rate Ratio; † Statistically Significant).

Table 5 :
Correlation between air pollutants and cardiovascular deaths happening one month later ( † Statistically Significant (p < 0.05)).