Introduction

Air pollution is a well-known risk factor for mortality and burden of disease [1]. Recent estimates from the Global Burden of Disease (GBD) study estimated that ambient air pollution caused 4.2 (95% confidence interval 3.7–4.8) million excess deaths (7.6% of total global mortality) for the year 2015 [2]. World Health Organization (WHO) in 2019, estimated that 99% of the global population lived in areas where air quality levels exceeded the air quality guideline values [3]. The major ambient air pollutants include particulate matter (PM), nitrogen dioxide (NO2), nitrogen oxide (NO), sulfur dioxide (SO2), carbon monoxide (CO), ozone (O3), and volatile organic compounds (VOCs) [4, 5]. PM is classified according to particle size. Fine particulate matter (PM2.5) is defined as particles with a diameter equal to or less than 2.5 μm and coarse particulate matter (PM10) is defined as particles with a diameter of 10 μm or less [6].

Long-term exposure to air pollution induces many health problems such as respiratory problems, cardiovascular disease, neurologic disorders, stroke, and cancer [6,7,8]. Air pollution is also recognized as the fourth largest risk factor for premature and lung cancer death [6,7,8,9]. Among air pollutants, PM2.5 and PM10 have been strongly connected to mortality and morbidity [10,11,12]. PM2.5 exposure in the long term increases the relative risk of all-cause mortality by 8% and cardiovascular events by as much as 10% [13,14,15]. The presence of NO2 and ground-level O3 have been reported as important contributors to mortality or morbidity due to respiratory and cardiovascular diseases [5, 16, 17]. NO2 is positively linked to increased mortality from cancer [18]. Evidence also documented that higher levels of O3 are connected with a higher risk of cognitive disorders, preterm birth, and reproductive health [19, 20].

Several studies are available that investigated the adverse health effects of air pollution [21], but the findings are inconsistent because air pollution is a complex mixture of pollutants from various sources. Some recent meta-analysis and cohort studies found an increased risk of mortality accompanied by air pollution [6, 22,23,24,25,26,27]. For instance, a cohort study by Peng et al. (2017) indicated that exposure to PM2.5 was significantly connected with mortality from all-cause (HR 1.30, 95% CI 1.19–1.42), respiratory diseases (HR 1.19, 95% CI 1.02–1.38), lung cancer (HR 1.72, 95% CI 1.36–2.19) and other cancers (HR 1.76, 95% CI 1.33–2.32) [23]. Kim et al. (2018) in a meta-analysis of cohort studies indicate that exposure to PM2.5, PM10, and NO2 were associated with increased mortality from all cancers [28]. Conversely, Tseng et al. (2015) in a cohort study found that exposure to PM2.5 was not significantly connected to all-cause (HR 0.92, 95% CI 0.72–1.17) and cardiovascular (HR 0.80, 95% CI 0.43–1.50) mortality [29].

Most existing meta-analysis focus on short-term effects of air pollution, especially for cardio-respiratory mortality [11, 30, 31], and the pooled associations between long-term exposure to air pollution with cardio-respiratory and lung cancer health is poorly understood. Further, some recent reports providing additional evidence of the associations from areas with low levels of air pollution, which found stronger associations between ambient air pollution, and all-cause, cardio-respiratory, and lung cancer mortality. Similarly, considering the gaseous air pollutants, the estimated associations between these pollutants (e.g. SO2, NO, NO2, and O3) with specific causes of mortality in long-term are still unclear. Thus, we decided to carry out a comprehensive literature review of cohort studies and perform a meta-analysis on the long-term association between particulate and gaseous air pollutants with all-cause and specific-cause of mortality. This review also systematically summarizes the effects of air pollution and mortality based on latest published evidence. The main goal of this systematic review and meta-analysis was to summarize the findings of cohort studies linked to air pollution with all-cause, cardiovascular, respiratory, and lung cancer mortality.

Method

Search of studies and selection

The current meta-analysis complies with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) methods (Table S1). The protocol of this study was registered on PROSPERO (https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=422945 with registration code = CRD42023422945). Electronic databases including Google Scholar, PubMed, Scopus, and Web of Science were independently searched by two researchers (B.K. and S.S.) to obtain the eligible studies up to 30 May 2022. The search for relevant literature was conducted with no restriction for regarding language, or publication date using appropriate keywords in the title and abstract, as well as Medical Subject Headings (MeSH). The search strategy is detailed in Appendix A (Supplementary Table). The reference lists of relevant studies were investigated for additional papers. We applied a combination of the following keywords: “air pollution”, “air pollutant*”, “particulate*”, “particle*”, "PM10", "PM2.5", “nitrogen oxide*”, “nitrogen dioxide”, “NO2”, “sulfur dioxide”, “SO2”, “black carbon”, “BC”, “ozone”, “O3”, “carbon monoxide”, “CO” linked with “mortality”, “all-cause mortality”, “cardiovascular mortality”, “respiratory mortality”, “lung cancer mortality”. The search was narrowed to prospective and retrospective cohort studies.

The title and abstract of the studies were screened and the full text of the selected studies was investigated for further assessment. The following inclusion criteria were applied for the extraction of information: (1) cohort studies explored the risk of mortality for all-cause, cardiovascular, respiratory, and lung cancer (2) studies investigated the exposure concentrations to PM10, PM2.5, BC, NO2, NO, SO2 and O3 (3) studies provided information about relative risk (RR), hazard ratio (HR), odds ratio (OR) and beta slope of regression models with 95% CIs (Confidence Intervals) which linked with air pollutants. All included studies clearly described the outcomes according to the International Classification of Diseases (ICD) codes including ICD-10 A00-R99 or ICD-9 001–779 to describe all-causes mortality, ICD-10 I10–I70 or ICD-9 400–440 to explain cardiovascular mortality, ICD-10 J00–J99 or ICD-9 460–519 to explain respiratory mortality and ICD-9 C34 to define lung cancer mortality. Review studies, letters to the editor, news articles, poster and conference abstracts were excluded. The detailed stepwise literature selection is presented in Fig. 1.

Fig. 1
figure 1

Summary of the study selection procedures by the PRISMA flow diagram

Full size image

Collection of data

Two investigators (B.K and S.S) separately extracted information including study location, study design, publication years, sample size, period of follow-up, method of exposure measurement, the concentration of air pollutants and their corresponding standard deviations, risk estimates of outcomes (RR, OR, HR) and their associated 95% CIs, mortality diagnosis by ICD codes, death rate, main findings and adjustment covariates (Table 1). We also reviewed the extracted data by the authors for quality control and assurance.

Table 1 Summary of selected studies characteristics
Full size table

Risk of bias assessment

The risk of bias and internal validity of studies was assessed by the Office of Health Assessment and Translation (OHAT) method as suggested by the National Institutes of Environmental Health Sciences-National Toxicology program [32]. Six domains including selection bias, confounders variable, exposure measurement bias, assessment of outcome, selective reporting bias, and missing data bias were evaluated. The risk of bias is categorized per domain as “low”, “probably low”, “probably high”, “high” and “not applicable” (Table 3 and see additional Table S2-S89). Furthermore, the Newcastle–Ottawa Quality Assessment Scale (NOQAS) method was also applied to evaluate the methodological quality of the selected studies. The methodological quality for each study is based on estimated scores categorized as “ ≥ 7, high”, “4–6, intermediate” and “ ≤ 3, low” (see Table S90- S91).

Data synthesis

Risk estimates obtained from included studies were expressed as RR, OR, HR, or beta (β) coefficients of regression. Estimates of OR, and HR were converted to RR by the following formulas:

$$RR=\frac{OR}{\left(1-r\right)+\left(r\times OR\right)}$$
(1)
$$RR=\frac{1-EXP(HR \times {\text{ln}}(1-r)}{r}$$
(2)

where r is the rate of death among the reference group. If r was not proved in the studies, the Human Mortality Database was used to obtain death rates based on gender, and age, and the year of study [28, 33]. The percent change of mortality is computed from the following equation (Percent change (%) = (RR-1) × 100%) [6, 33]. To obtain the RR of mortality from the beta (β) coefficients, we exponentiated the regression coefficient. Additionally, 95% Cl was estimated using an exponentiated regression coefficient with their standard error [β ± (1.96*SE)] [6, 34].

When exposure measurements were reported as ppb or ppm, these findings converted to μg/m3 as follows: NO2, 1 ppb = 1.88 μg/m3; O3, 1 ppb = 1.96 μg/m3; SO2, 1 ppb = 2.66 μg/m3; CO, 1 ppb = 1.15 μg/m3 [35]. If studies did not report the RR based on a unit of 10 μg/m3 increment in each pollutant, the estimates of RR across studies were standardized using the following formula [34]:

$${{\text{RR}}}_{{\text{Stdardized}}}={{\text{e}}}^{(\frac{\mathrm{Ln }({{\text{RR}}}_{{\text{Origin}}}}{{{\text{Increment}}}_{{\text{Origin}}}} \times {{\text{Increment}}}_{{\text{Standardized}}})}$$
(3)

Meta-analysis

Statistical analyses and forest plots were created by STATA12 and R version 3.6.1. The pooled effect (RR) of studies was computed using the random-effect and fixed-effect model based on the Mantel–Haenszel procedure. The presence of statistical heterogeneity between the estimated effect of studies was evaluated by I2 and Cochran’s Q-test (Significance level < 0.1). The I2 value equals 25% indicating a “low” degree of heterogeneity, 50% moderate, and I2 exceeds 75% suggesting a “high” degree of heterogeneity [36]. To assess the possible sources of heterogeneity, subgroup analyses were performed for sex, mean of age, and study location (Asia, Europe, Canada, and the United States). The presence of publication bias was evaluated by a funnel plot of the log RR against the standard error (SE). Funnel plot asymmetry was also tested by Egger's test with a significance level < 0.10. We applied the trim-and-fill method for detecting and adjusting the publication bias in our meta-analysis [6, 36]. Sensitivity analyses were achieved to investigate the robustness of our main analyses by analyzing the impact of excluding each study on the consistency of the results.

Results and discussion

Studies included

After searching databases, 3692 records were recognized. Eighteen publications were also added from the reference lists. Removing the duplicates and unrelated articles after the screening of abstracts, 127 studies were eligible to assess the title and full text. Finally, 88 articles were included in the meta-analysis and fulfilled the quality assessment criteria (Fig. 1) [14, 23,24,25,26,27, 29, 37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116]. Eighty-one studies involved an adult population with both genders, and five studies included only adult females [44, 67, 80, 86, 110] and two studies included only adult males [64, 75]. Among the included studies, 66 studies investigated all-cause mortality, 64 studies evaluated cardiovascular mortality, 37 studies assessed respiratory mortality, and 31 studies considered lung cancer mortality as an outcome. The included studies were carried out in Europe (31 studies), the United States (27 studies), Canada (8 studies), and Asia (22 studies) (Fig. 2A and B). The total number of participants in these studies was 270877792 individuals (Table 1).

Fig. 2
figure 2

A Geographic location of included studies and (B) the number of publications reporting the association between air pollutants with mortality

Full size image

The concentrations of pollutants and methods of exposure assessment are diverse between the studies. The overall mean concentration of PM2.5 derived from studies was 19.47 ± 4.26 µg/m3, with a mean concentration of 39.81 ± 0.78 µg/m3 in Asia, 14.92 ± 2.20 µg/m3 in Europe, 19.69 ± 3.33 µg/m3 in the USA, and 10.7 ± 2.08 µg/m3 in Canada, (Fig. 3, See more in Table S3). Pooled average concentrations of PM10, NO2, SO2, and O3 derived from studies were 45.3 ± 17.68, 23.54 ± 8.92, 16.1 ± 4.65 and 60.67 ± 14.96 μg/m3, respectively (Table 2). We found higher levels of air pollution in Asian developing countries (Fig. 3).

Fig. 3
figure 3

The pooled concentration of PM2.5 (A), PM10 (B), NO2 (C), and O3 (µg/m3) (D) stratified by location

Full size image
Table 2 Pooled concentrations of PM2.5, PM10, O3, NO2, SO2 and CO (µg/m3) derived from studies
Full size table

Methodological quality and risk of bias

The risk of bias rating based on investigated studies is presented in Table 3 (also see more detailed of risk of bias assessment in supplementary Table S4-S91). All individual studies had a high quality. The findings showed that the risk of bias for exposure measurement, assessment of outcome, and selective reporting bias differed among the studies. Selection bias for one study was rated as “probably high” risk [42] while other studies were “low” risk. For exposure measurement bias, 7 studies were rated as “probably high” risk because the sampling technique of these studies was not clearly described [27, 29, 42, 54,55,56,57]. Outcome assessment bias for 11 studies rated as “probably high” risk [41, 42, 51, 53, 57,58,59, 81,82,83,84]. The selective reporting bias of 10 studies rated as “probably high” risk [29, 41, 42, 48, 53,54,55, 57,58,59]. Results of the risk of bias assessment are displayed in Table 3, and summary tables justify for the judgments given for each bias domain presented in supplementary Table S4-S91.

Table 3 Risk of bias rating for each study
Full size table

Air pollutants and mortality

This is a meta-analysis investigating the relationship between exposure to air pollutants with the risk of mortality for all-cause, cardiovascular, respiratory, and lung cancer. We used data derived from 88 cohort studies carried out in 20 countries involving more than 270 million subjects. Using a random-effect model, exposure to PM2.5, PM10, NO2, and SO2 was strongly linked with mortality, whereas O3 was not significantly related. The pooled estimates of all-cause mortality with air pollutants are presented in Table 4. Positive and significant associations were observed between the percent changes of all-cause mortality with a 10 μg/m3 increment in PM2.5 (RR 1.08, 95% CI 1.07–1.09), PM10 (RR 1.10, 95% CI 1.06–1.14), and BC (RR 1.04, 95% CI 1.02–1.07) (Fig. 4a-c). The estimates of heterogeneity between studies were found to be high for PM2.5, PM10, and BC (Table 4, see Forest plot in Supplementary Fig. S1-7). Elevated risk of all-cause mortality significantly linked with a 10 μg/m3 increment in NO2 (RR 1.04, 95% CI 1.03–1.06), NOx (RR 1.02, 95% CI 1.01–1.04) and SO2 (RR 1.03, 95% CI 1.00–1.06) (Fig. 4a-c), but the association was not significant for O3 (RR 0.98, 95% CI 0.97–1.01) (Table 4, see Forest plot in Supplementary Fig. S1-7). Heterogeneities between the studies were found to be high for NOX, NO2, SO2, and O3.

Table 4 The estimated pooled RR of mortalities associated with air pollutants using the random effect model
Full size table
Fig. 4
figure 4figure 4

The pooled relative risk of all-cause mortality associated with PM2.5 (A), PM10 (B), BC (C), NO2 (D), SO2 (E), and O3 (F) exposure stratified by cause-specific mortality, gender (male and female) and age (< 64 years and ≥ 65 years)

Full size image

The highest elevated risk of cardiovascular mortality was significantly linked with a 10 μg/m3 increment in PM10 (RR 1.15, 95% CI 1.08–1.22), the second highest with NO2 (RR 1.06, 95% CI 1.04–1.08), followed by SO2 (RR 1.06, 95% CI 1.01–1.11), PM2.5 (RR 1.06, 95% CI 1.05–1.06), NOx (RR 1.03, 95% CI 1.01–1.06), and BC (RR 1.03, 95% CI 1.01–1.05) (Fig. 4a-e). A non-significant association was observed between the risk of cardiovascular mortality with exposure to O3 (RR 0.99, 95% CI 0.95–1.03) (Table 4, see Forest plot in Supplementary Fig. S8-14). The relationships between the risk of cardiovascular mortality with exposure to PM10, PM2.5, BC, NO2, NOx, SO2, and O3 were investigated by 46, 17, 13, 24, 5, 9, and 9 studies, with “high” levels of heterogeneities (I2 > 75%) (Table 5).

Table 5 The results of Egger’s and trim-and-fill test with the number of imputed studies to complete asymmetry in the Funnel plot
Full size table

Significant positive relations were observed between the risk of respiratory mortality with a 10 μg/m3 increment in PM2.5 (RR 1.066, 95% CI 1.034–1.097), PM10 (RR 1.196, 95% CI 1.114–1.279), BC (RR 1.048, 95% CI 1.025–1.07) and NO2 (RR 1.061, 95% CI 1.033–1.089), except for NOx (RR 1.026, 95%CI 0.998–1.055), SO2 (RR 1.041, 95% CI 0.964–1.118) and O3 (RR 0.971, 95% CI 0.944–0.998) (Table 4, Supplementary Fig. S15-21).

We observed significant relations between the risk of lung cancer mortality with a 10 μg/m3 increment in PM2.5 (RR 1.118, 95% CI 1.076–1.159, I2 77.40%, τ2 0.0042), PM10 (RR 1.127, 95%CI 1.029–1.224), BC (RR 1.048, 95% CI 1.025–1.07), NO2 (RR 1.067, 95%CI 1.039–1.095), NOx (RR 1.057, 95%CI, 1.005–1.11), and SO2 (RR 1.087, 95% CI 1.011–1.163). Applying the forest plot, the risk of lung cancer mortality was negatively related to exposure to O3 (RR 0.921, 95% CI 0.865–0.978) (Table 4, Supplementary Fig. S22-28).

Exposure to PM2.5 per 10 µg/m3 increment is associated with an elevated risk of mortality for all-cause (8%), cardiovascular (6%), respiratory (7%), and lung cancer (11.8%). Less or more the same associations have been reported between exposure to PM2.5 and all-cause mortality by two other studies [117, 118]. Our finding was higher than that (3.9%) obtained by Hart et al. [119], but considerably lower than those findings (53% and 17%) reported by studies conducted in China and the United States [37, 120]. The association between exposure to PM2.5 and cardiovascular mortality in the current study was comparable to that finding previously reported by Pope Lii et al. [90], but higher than the result obtained by a cohort study conducted in the United States [121]. The same but not significantly elevated risk was reported by Beelen et al. [69]. Consistent with our findings, several studies found an elevated relationship between exposure to PM2.5 with mortality due to respiratory disease and lung cancer [69, 80, 122, 123]. Air pollution related to traffic is likely to contribute as a source of PM2.5, which possibly leads to elevated deaths due to lung cancer [124].

The association between exposure to PM10 and all-cause mortality in the present study (10%) was considerably lower than the result reported by a previous study of meta-analysis (18%) [28], but the association was substantially higher than the estimated percent changes of 5% and 3.9% obtained two studies [36, 119]. The onset of mortality associated with exposure to PM2.5 and PM10 could be explained through numerous underlying mechanisms such as oxidative stress and systemic inflammation, which leads to direct neurotoxicity, hormonal dysregulation, promotion of cell turnover, epigenetic changes in the genome, suppression of DNA repair, DNA methylation, and consequently prompt carcinogenesis [125,126,127,128,129,130]. PM2.5 has also improved the production of inflammatory cytokines (interleukin (IL) -6 and IL-8) due to the mitochondrial generation of hydroxyl radical (•OH) as a reactive oxygen species (ROS) [126, 131,132,133,134,135].

The estimated percent changes of all-cause mortality per 10 μg/m3 increment in exposure to NO2 (4.5%), and SO2 (3.5%) in the present study were less or more similar to those findings obtained by meta-analyses studies [36, 136, 137]. In the present investigation, the significant relations between exposure to NO2 with all-cause (4.5%), cardiovascular (6.3%), and respiratory (6.1%) mortality were markedly higher than those findings reported by an earlier study of meta-analysis including all-cause (1.58%), cardiovascular (1.72%) and respiratory (2.05%) mortality [30]. Findings of a recent pooled analysis of 67 studies showed elevated associations between exposure to SO2 with respiratory (1.0067) and all-cause (1.0059) mortality, but the findings were lower than our results [31]. Our finding for all-cause mortality was substantially lower than the previous results obtained for NO2 (8.2% and 14%) and SO2 6.9% [79]. NO2 and SO2 can exacerbate the effects of oxidative stress and promote the progression of respiratory disease and lung cancers [138, 139].

No significant associations in the present study were found between exposure to O3 with all-cause (RR 0.99) and respiratory (RR 0.97). Similar results were reported by a study of meta-analysis for all-cause (RR 0.97) and respiratory (RR 0.99) mortality [136]. The same findings were also reported by a previous study [140].

Subgroup analyses

Figure 4A–E presents the pooled effects of air pollutants on cause-specific mortality with stratification for sex (male and female) and age (< 0–64 and ≥ 65). The RR of all-cause mortality related to PM2.5 exposure was highest in males (RR per 10 µg/m3 = 1.075, 95% CI = 1.030–1.09) compared to the female (1.039, 95% CI: 1.02–1.058) and individuals with age ≥ 65 years (RR per 10 µg/m3 = 1.046, 95% CI = 1.005–1.075), compared to the individuals with age < 64 years (1.028, 95% CI = 0.965–1.085) (Fig. 4). The RR of all-cause mortality due to PM10, NO2 and SO2 for male and individual age ≥ 65 years were also highest compared to the female and individuals with age < 64 years.

In sex-stratified analysis detected a significant association between exposure to PM2.5, PM10, BC NO2, and SO2, and mortality in males compared with females, but no significant association was found for O3. This finding could be related to diverse physiological functions in men and women. Moreover, individual characteristics (e.g., smoking, physical activity, alcohol consumption, work- exposures, etc.) are an important risk factor in air pollution-related mortality [122, 141]. Similarly, the risk of mortality from coronary heart disease, cardiorespiratory disease, and myocardial infarction, which is attributed to air pollution was higher in males and the elderly [142, 143].

Individual over 65 years old was more susceptible to ambient PM2.5, NO2, and SO2 exposure, while the younger (< 65 years old) age were more susceptible to BC and O3. Previous studies supported these findings [141, 142, 144, 145]. The elderly are typically more exposed to outdoor air pollution compared to the younger age [141, 146]. The physiological structures and body functions diminish with age, which might enhance the risks of air pollution-related mortality among the elderly [147].

Additional analyses

According to Egger’s test, we found significant publication bias for O3 (P-value 0.03), SO2 (P-value 0.005), and NOx (P-value 0.012) with all-cause mortality, but not for PM2.5 (P-value 0.52), PM10 (P-value 0.33) and BC (P-value 0.44). The sub-stratified analysis of the association between all-cause mortality with PM2.5 by location illustrated that the RR was higher in studies conducted in Canada (RR 1.17, 95% CI 1.15–1.20, I2 68.0%, p-value 0.008) compared with Asia (RR 1.04, 95% CI 1.01–1.08, I2 98.1%, p-value 0.000) and United states (1.08, 95% CI 1.06–1.09, I2 91.8%, p-value 0.000) (Supplementary Fig. S29-S35).

The trim and fill method result shows that about 7 records are essential to creating a complete asymmetry in the Funnel plot (P-value < 0.001) of All-Cause with PM2.5 (Table 5, Supplementary Fig. S36-S59). The graphical funnel plots appeared to be slightly asymmetrical for exposure to PM10, BC, NOx, O3, and SO2, suggesting the presence of publication bias for studies (p-value ≤ 0.05). Sensitivity analyses showed that the results were stable for the combination of pollutants with all-cause cardiovascular, and respiratory mortality, and the pooled RR did not alter when any individual record was excluded, indicating the robustness of the results (Table 5).

The current study has some limitations. First, we included only cohort studies, and other types of study designs such as case–control, cross-sectional, time-series, or case-crossover designs were not evaluated. Second, most of the included studies were performed in developed countries especially in urban areas of Europe and North America, while there was no study available for African countries. Third, the exposure measurements were conducted in outdoor environments, and indoor air pollution (e.g. home, school, office) was not considered. Fourth, we were also unable to perform meta-regression analyses to identify the sources of heterogeneity for some air pollutants due to low sample size.

Conclusion

The most considered air pollutants were associated with an increased risk of mortality due to cardiovascular and respiratory diseases, but the effects of PM10 and PM2.5 were stronger. Exposure to PM10 and PM2.5 is the predominant factor for mortality risk, contributing to RR 1.104 and 1.08 in all-cause mortality and RR 1.149 and 1.058 in cardiovascular mortality. The highest respiratory and lung cancer mortality was associated with exposure to PM10 (1.196 and 1.127), followed by PM2.5 (RR 1.066 and 1.118). The male and elderly adults seemed to be more susceptible to exposure to particle air pollution compared to the female and younger age groups. Thus, policymakers need to pay more attention to establishing new regulations and intervention strategies to enhance air quality. This subsequently leads to a diminishing of morbidity and mortality. Further population-based studies in this field are required to enhance the understanding of the adverse health effects of air pollution among vulnerable subgroups.