Current trend in air pollution exposure and stroke

Authors: Stephan Gabet, Laurent Puy

Abstract

Purpose of review

Stroke is the second leading cause of death worldwide, and exposure to particulate air pollution is now recognized as one of the major modifiable risk factors. However, air pollution can vary in terms of physicochemical composition and exposition specificities. Therefore, its relationships with stroke outcomes remain under intense investigation.

Recent findings

This review highlights, alongside particles, that short-term and long-term exposure to nitrogen dioxide (NO2) and ozone is likely to be also linked to stroke-related morbidity and mortality. Moreover, air pollution may increase the risk of transitioning from a healthy status to incident stroke and morbimortality after stroke. Additionally, relationships may vary depending on the air pollution mixture (e.g., particle-related components, pollutant interactions), pollutant sources (e.g., traffic-related or not), stroke etiology (ischemic or hemorrhagic), or exposed individual's characteristics (e.g., age, sex, genetic predisposition, weight status). Nonlinear dose-response functions and short-term effect lags have been reported, but these features need further refinement.

Summary

The relationship between stroke and air pollution is now well established. Nonetheless, future research should further consider the physicochemical properties of air pollutants, multiple exposures, and individual vulnerabilities. Moreover, advanced statistical methods should be more commonly used to better describe the relationship shapes.

INTRODUCTION

Stroke stands as the second foremost cause of death worldwide, with approximately 6.6 million deaths each year; it also ranks third among leading cause of mortality and disability combined, contributing to 5.7% of total disability-adjusted life years (DALYs) [1]. The burden of stroke has increased sharply worldwide over the past century to an annual incidence of 12.2 million cases, with an annual cost of approximately US$720 billion (0.66% of the global GDP) [1].

The main modifiable risk factors for stroke include high blood pressure, diabetes, dyslipidemia, malnutrition, smoking, and lack of physical activity. In addition, the global burden of air pollution on cardiovascular health is a compelling concern. Especially, short-term (hour to days) and long-term (weeks to years) exposure to particulate air pollution emerged as a recognized modifiable risk factor for stroke morbidity and mortality [2–4]. In 2019, air pollution was the fourth leading risk factor contributing to stroke death and disability combined [5].

This review summarizes the most recent advances in our understanding of stroke, due to short- and long-term exposure to air pollution. We address the impact of air pollution on stroke incidence and outcome, and discuss recent data which have deepened the understanding of the above relationships by considering exposure specificities or potential modifying and mitigating effects.

HOW TO DEFINE AIR POLLUTION?

Air pollution, which represents less than 0.1% of the total composition of the air, is a ubiquitous and complex mixture of solid and liquid particles and gases, both indoors and outdoors. Prominent components of atmospheric pollution include airborne particulate matter, nitrogen oxides (NOX) such as nitrogen dioxide (NO2), ground-level ozone (O3), and to a lesser extent, sulfur dioxide (SO2) and carbon monoxide (CO). In indoor air, we primarily find combustion products (e.g., particulate matter, NO2), volatile organic compounds (VOCs), and bioaerosols (allergens and bacterial and fungal compounds; of biological nature, they are not covered in this review focusing on chemical air pollution).

Particulate matter is commonly characterized by size, distinguished by aerodynamic particles less than 10 μm (PM10), which are small enough to be inhaled, particles less than 2.5 μm (PM2.5, also called fine particles), which can reach the pulmonary alveoli, and particles less than 1 μm (PM1), which are small enough to enter the bloodstream; particles less than 0.1 μm (PM0.1), part of the PM1 fraction, are called ultrafine particles (UFP). Particulate matter can be further characterized by its components such as black carbon, elemental carbon, organic carbon, and ions (e.g., nitrate, sulfate, ammonium).

Main anthropogenic sources of particles include heating and cooking equipment, road traffic, and industry. There are also nonanthropogenic particles such as pollen, sea spray, and volcanic ash. Regarding gases, major outdoor and indoor sources of NOX include, respectively, fossil fuel combustion (e.g., combustion engine vehicles, stationary power generation) and gas and coal heating and cooking equipment. Tobacco smoke is an additional indoor air pollution. Ozone is not emitted but formed by the conversion of NO2 and VOCs present in the atmosphere under the influence of ultraviolet rays. Ozone is therefore a secondary pollutant of particular concern in the context of climate change, which is leading to increase heat and sunshine.

The impact of air pollution on health is huge. An estimated 4.1 million (95% confidence 3.5–4.8) and 2.3 million (1.6–3.1) people die each year worldwide from outdoor and indoor fine particles, respectively. About one-fifth of them die from stroke, making air pollution the leading environmental risk factor for stroke, with an attributable fraction estimated at about 15–20% [1,6,7].

AIR POLLUTION AND STROKE: UNDERLYING PATHOPHYSIOLOGY

Multiple pathophysiological mechanisms have been implicated in the relationship between air pollution and stroke (Fig. 1). Air pollutants inhaled into the lungs trigger a local then systemic pro-inflammatory status. The release of pro-inflammatory and pro-oxidant mediators (increase of circulating levels of leukocytes and various inflammatory cytokines) is deleterious for the cardiovascular system, especially for endothelium integrity [8]. The smallest particles (PM1) can also directly affect the vasculature and circulating blood cells. They act by activating a proinflammatory response in endothelial cells, upregulating the expression of adhesion molecules, and generating reactive oxygen species, which are key mediators of accelerated atherosclerosis [9].

Exposition to PM can also induce a prothrombotic state by enhancing platelet activation and von Willebrand function, and also trigger an immunothrombosis reaction mediated by neutrophils extracellular traps (NETs) [10]. Data also suggest that particles may translocate to central regions at the level of the nasal passage, particularly the olfactory bulb where high doses of particles are deposited through the first airflow passage, raising the hypothesis of a direct neurotoxicity [11]. In addition, exposure to air pollution may induce cardiorespiratory dysautonomia and can also contribute to the development of common risk factors for cerebrovascular disease, including diabetes, hypertension, atrial fibrillation, and atherosclerosis, inducing a vicious cycle [12].

Finally, some recent epidemiological research sought to better discern underlying pathways, hence causal links [13,14]. Notably, Liang et al.[13], by retrieving genome-wide association study (GWAS) statistics for PM2.5 and CVDs, showed that exposure to fine particles may increase the risk of stroke diagnosis via several single nucleotide polymorphisms (SNPs).

EFFECTS OF LONG-TERM EXPOSURE TO AIR POLLUTION

Consequences of long-term exposure on stroke incidence

While published data are consistent and support the link between exposure to air pollution and the risk of new incident stroke, this association may vary according to the subtype of air pollutant. Indeed, Jin et al.[15] recently reported significant associations between air pollution and stroke incidence in over 30 million older US fee-for-service Medicare adjusted hazard ratio (HRa) 2.79% (95% confidence 2.74–2.84), 0.95% (0.93–0.96), and 0.26% (0.23–0.28) per increase of 1 μg/m^3^ in PM2.5, 1 part per billion (ppb) in NO2, and 1 ppb in warm-season O3, respectively. As well, Poulsen et al.[16^▪^] reported significant associations with stroke incidence in 2 million individuals aged over 50 years in Denmark: HRa 1.223 (1.175–1.273) per 5 μg/m^3^ increase in PM2.5, 1.095 (1.063–1.128) per 10 000 particles/cm^3^ increase in UFP, and 1.039 (1.023–1.056) per 5 mg/m^3^ increase in NO2. Consistent results were reported in several Chinese epidemiological studies between incident stroke and exposure to NO2[17] or ozone [18–20]. However, null associations were sometimes observed in another Chinese studies with particulate matter, NO2, O3, or CO [19,21]. Lastly, the study led by Riggs et al.[22] in nearly 27 000 participants of 45 years or above enrolled in the US REasons for Geographic And Racial Differences in Stroke (REGARDS) prospective cohort reported a significant association with PM10 HRa 1.07 (1.003–1.150) per 5.4 μg/m^3^ increase; no evidence of association was observed for PM2.5, O3, NO2, SO2, and CO.

The increased risk of a new stroke can also be observed in the rate of stroke-related hospitalization. Nethery et al.[23] showed that stroke-related hospitalizations were positively associated with PM2.5 in nearly 2 million US fee-for-service Medicare beneficiaries with acute or chronic cardiovascular HRa 1.016 (1.013–1.019) per 1 μg/m^3^ increase. Chen et al.[24^▪▪^] investigated the effects of various particulate matter fractions on ischemic stroke-related hospitalizations and showed that the finer the particulate matter fraction the higher the association. Furthermore, the effects on stroke-related hospitalizations seem to affect various age groups, genders, and stroke-subtypes. Thus, Kulick et al.[25] reported stronger associations with hospitalization for ischemic than hemorrhagic stroke in about 155 000 postmenopausal women from the US Women's Health Initiative (WHI) prospective HRa 1.15 (1.12–1.18) versus 1.13 (1.06–1.19) per 3.5 μg/m^3^ increment in PM2.5 and 1.05 (1.01–1.09) versus 1.02 (0.93–1.11) per 9.5-ppb increment in NO2.

Consequences of long-term exposure on the prognosis of stroke

Several Chinese studies reported significant associations of particulate matter exposure with stroke-related mortality [26^▪▪^,27]. For instance, Cai et al.[26^▪▪^] focused on the effects of PM2.5 in a retrospective cohort of over three million hospitalized patients for stroke, and reported significant associations, higher for ischemic than hemorrhagic adjusted odds-ratio (ORa) 1.158 (1.130–1.188) per 13.9 μg/m^3^ increment and 1.082 (1.047–1.119) per 21.6 μg/m^3^ increment, respectively. Qi et al.[28] unusually considered exposure to ambient UFP in an ecological study in the New York State and also showed positive relations.

Furthermore, some research has regarded not only stroke morbidity or mortality independently, but also the entire illness trajectory, from health status to incident stroke, poststroke CVDs, hospital readmissions, multimorbidity, and death.

Thus, Tran et al.[29] showed long-term exposure to NO2, PM2.5, CO, or SO2 was significantly associated with 30-day all-cause hospital readmission among nearly 450 000 US fee-for-service Medicare beneficiaries aged 65 years or older and primarily hospitalized for ischemic stroke. Similarly, Cai et al.[30] reported, in a cohort of more than one million participants in Sichuan (China), that higher short-term and long-term exposure levels to particulate matter increased the risk of 30-day stroke rehospitalization in stroke patients; higher risks were reported for long-term than short-term exposure.

Moreover, Zou et al. (2023) examined the contribution of long-term exposure to PM2.5, PM10, NO2, and NOX to cardiometabolic diseases (CMDs), including stroke, cardiometabolic multimorbidity (CMM), and subsequent death, in about 370 000 UK Biobank participants [31]. Increased exposure levels, and particularly to PM2.5, significantly increased the risks of transition from healthy status to CMD, and then to CMM and death. Moreover, Tian et al.[32] showed in the same UK Biobank cohort that increases in particulate matter and NOX levels were linked to more frequent transitions from healthy status to incident stroke, poststroke CVDs, and subsequent death.

EFFECTS OF SHORT-TERM EXPOSURE TO AIR POLLUTION

Consequences of short-term exposure on stroke incidence

Comparing to long-term exposure, the recent literature dealing with short-term exposure is broadly less extensive. Nonetheless, Toubasi and Al-Sayegh [33^▪▪^] conducted an outstanding meta-analysis to examine the acute effects of particulate matter, NOX, O3, SO2, and CO on ischemic stroke incidence. More than 18 million stroke cases were retrieved from 110 observational studies. Ischemic stroke incidence was significantly associated with many air meta-analytical relative risk (RR) 1.14 (1.12–1.16), 1.15 (1.13–1.17), and 1.09 (1.06–1.12) per 10 μg/m^3^ increase in PM10, PM2.5, and PM1, respectively; meta-analytical RR 1.30 (1.22–1.37), 1.05 (1.03–1.07), and 1.17 (1.12–1.21) per 10-ppb increase in NO2, O3, and SO2, respectively; and meta-analytical RR 1.26 (1.21–1.32) per 1-part-per-million (ppm) increase in CO. Nonetheless, these results should be interpretated with caution given the high heterogeneity between studies. Moreover, Badida et al.[34] performed a meta-analysis focusing on 152 studies led in low and middle-outcome countries (LMICs). Thus, they reported contrasted meta-analytical RR 1.0002 (0.9924–1.0080), 1.0044 (1.0007–1.0081), 1.0065 (0.9819–1.0318), and 1.0136 (1.0105–1.0167) per 10 μg/m^3^ increase in PM10, PM2.5, NO2, and SO2, levels, respectively; here again, heterogeneity between studies was particularly high.

Consequences of short-term exposure on the prognosis of stroke

Regarding mortality, Toubasi and Al-Sayegh [33^▪▪^] also reported increased risks of ischemic stroke mortality in their meta-analysis: meta-analytical RR 1.02 (1.00–1.04), 1.09 (1.04–1.15), and 1.09 (1.04–1.15) per 10 μg/m^3^ increase in PM10, PM2.5, and PM1, respectively; meta-analytical RR 1.33 (1.07–1.65), 1.26 (0.91–1.72), and 1.60 (1.05–2.44) per 10-ppb increase in NO2, O3, and SO2, respectively; and meta-analytical RR 1.10 (0.81–1.50) per 1-ppm increase in CO. As well, Badida et al.[34] reported significant risk associations in LMICs: meta-analytical RR 1.0064 (1.0038–1.0089), 1.0214 (1.0018–1.0413), 1.0219 (1.0048–1.0394), and 1.0173 (1.0074–1.0274) per 10 μg/m^3^ increase in PM10, PM2.5, NO2, and SO2, levels, respectively. Nonetheless, the heterogeneity between studies was high in the two meta-analyses; accordingly, results should be interpretated with cautious.

MODIFYING EFFECTS AND RELATIONSHIP SPECIFICATIONS

Many research works sought to specify the above associations by studying the potential modifying effects of various factors, mainly intrinsic. Thus, older people appeared in many studies to be more vulnerable to particulate matter [25,30,35,36] and ozone [18,20], although this trend was sometimes not reported [21,37]. On the contrary, the potential modifying effect of sex was still unclear, with stronger associations reported sometimes in men [20,36], sometimes in women, particularly for particulate matter [24^▪▪^,30,35], sometimes in neither, particularly for ozone [17,18,21,37,38]. Stronger associations between air pollution and stroke outcomes were consistently reported in individuals who were genetically predisposed [39,40] or affected [15,35,41] by cardiometabolic comorbidities such as diabetes and high blood pressure, but sometimes not [15,25]. In addition, overweight and obesity were suggested as a vulnerability to air pollution in some studies [17,20], but not all [25]. As well, a high physical activity level has been suggested has a protective modifying factor against particulate matter [24^▪▪^] or ozone [42] adverse effects. People with low education level [20,24^▪▪^], low socioeconomic status [41], or living within a deprived neighborhood [41] could be more affected by air pollution, but these interactions need also to be corroborated [20,25]. Finally, no conclusion can be drawn concerning the role of ethnic origin due to very heterogeneous results [15,23,25,28,30,35] while stratified studies on tobacco smoking are too scarce to be conclusive [25,42].

KNOWLEDGE GAPS AND FURTHER CONSIDERATIONS

Recent advances have led to a better understanding of the impact of exposure to air pollution on the incidence and outcomes of stroke, but some issues remain.

According to the literature, the shape of the curve of the relationship between air pollution and stroke is still understudied. Indeed, most studies do not rely on statistical models allowing such analysis, such as those incorporating a spline smoothing exposure term instead of a usual linear term. Thus, nonlinear dose--response functions were sometimes suggested, with steeper curves at lower particulate matter [24^▪▪^] or ozone [20] concentrations. On the contrary, steeper curves were elsewhere reported at higher particulate matter levels [35,43]; similarly, higher risk estimates were reported when focusing on periods of higher exposure levels, such as the cold season for particles [28,36] or ozone pollution peaks [38]. Moreover, nearly linear curves were also described for exposure to particulate matter [40,44], NO2[17,40], or ozone [18,21,38]. The curve patterns also appeared contrasted depending on the chronicity of the exposure [30], the particulate matter fraction [30,44], or particulate matter related components [26^▪▪^,44,45].

Concerning the latter particulate matter components, their relationships with stroke remain too rarely investigated. Although stroke-related outcomes, such as stroke incidence, ischemic stroke mortality, or poststroke functional disability, may be specifically linked to long-term exposure to particulate matter related EC [16^▪^], particulate matter related black carbon, organic carbon, and sulfate [26^▪▪^], or particulate matter related OC and sulfate [45], respectively. Cerebrovascular mortality was elsewhere related to short-term exposure to particulate matter related elemental carbon, organic carbon, and nitrate as well as to traffic-related air pollution (TRAP) [36]. Conversely, stroke risk was sometimes associated with non-TRAP only [16^▪^]. Haddad et al.[46^▪▪^] specifically meta-analyzed the relationships between long-term TRAP exposure and stroke incidence. They therefore reported some evidence of meta-analytical RR 1.22 (1.03–1.21) per 5 μg/m^3^ increase in traffic-related PM2.5 and 1.01 (0.96–1.06) per 10 μg/m^3^ increase in traffic-related NO2.

Furthermore, concerning specifically short-term effects, the plausible time delay after exposure remains not clearly determined, due in particular to sometimes heterogeneous results but especially to a limited literature. Nevertheless, the adverse effects are likely to occur in the first tens hours following the exposure, and not after a few days according to available data. Indeed, considering day-by-day exposure averages over the 7 days preceding stroke, the effects of NO2 appeared at their maximum on the day of exposure in a Chinese retrospective cohort [42]. Moreover, some research relied on advanced statistical methods, such as distributed lag nonlinear model (DLNM), to better assess the lag-response relationship specifically. Thus, the time series by Lv et al.[43], encompassing hourly exposure levels to PM2.5, PM10, NO2, and SO2, showed that risk for stroke-related emergency hospital admission was the strongest at lag 0 to 2 h. Another time series, based on PM2.5 daily exposure and an incident cerebrovascular disease registry, reported the highest risk estimates around a 4-day lag [47].

Moreover, although a growing body of evidence suggests that exposure to air pollution worsens the prognosis of stroke, the underlying reasons are unclear and are probably multiple. In particular, given the effect of air pollution on hemostasis, it could be argued that exposure could induce clot resistance leading to failure of recanalization procedure (intravenous thrombolysis and/or mechanical thrombectomy), a key predictor of mortality and functional outcome after ischemic stroke [48,49].

Finally, understanding of the interactions between environmental exposures has recently been deepened, notably thanks to multiple exposure approaches. Thus, Jiang et al.[42] showed that the association between acute NO2 levels and the risk of stroke was almost doubled in people exposed to higher PM2.5 concentrations, while no association was found in those exposed to lower concentrations. Moreover, joint effects of air pollutants on ischemic stroke incidence would be possibly overestimated through single-pollutant models compared with multipollutant models, underscoring the relevancy of multiexposure approaches [37]. In this line, de Bont et al.[50] performed a principal component analysis (PCA) to notably summarize the long-term exposure to a mixture of air pollutants (PM2.5, NO2, black carbon, and O3). They therefore reported significant associations with stroke incidence which were more consistent than single-pollutant model estimates. In another study, exposure to atmospheric pollutants (PM2.5, UFP, elemental carbon, and NO2), ambient noise, and surrounding greenness were all related to stroke in single-exposure analyses, but only PM2.5 and noise were independently associated with stroke in the multiexposure analysis [44]. Lastly, several works looked at the interactions between air pollution and other environmental components such as greenness. These studies highlighted a possible mitigation of the adverse effects of air pollution in people having access to green surfaces [51,52,53^▪^]; however, this phenomenon could depend on stroke etiology (ischemic or hemorrhagic).

The last point emphasizes the importance of embracing the problem of air pollution, not in isolation, but in an increasingly global and holistic manner, taking also into account the whole exposome and other public health contexts. And encouragingly, in light of this literature review, research seems to be clearly moving in this direction. Thus, while epidemiological studies were historically mainly conducted in high-income countries, the issue is now being addressed worldwide, including in LMICs where the health impact of air pollution is dramatic [5]. In addition, while particulate matter was long the focus of research, there is a diversification of the pollutants studied, alone or in mixtures, with an increasingly preponderant place for ozone which is of primary concern in the context of climate change. Moreover, COVID-19 pandemic-related measures such as lockdowns and curfews led to decreases in TRAP, mainly NO2 but to a lesser extent PM2.5 of which the residential and tertiary sectors are the main sources of emissions. Since among the effects of air pollution on stroke, only those linked to PM2.5 have been proven so far, the estimated benefits linked to the improvement of air quality during the pandemic remain limited [54,55]. Linking air pollutants other than PM2.5 to stroke and implementing long-term air quality public policies should broaden the range of levers for action and increase health gain estimates.

CONCLUSION

The stroke burden related to air pollution exposure is now well established. This relationship may vary depending on the air pollutant considered as well as the air pollution mixture, the pollutant sources or even individual's characteristics.

Preventive interventions are of utmost importance. The most effective way to reduce the health risks related to air pollution is to implement society-wide policies aiming to reduce emissions, main anthropogenic sources being heating, traffic, and industries [56]. Complementary, exposure can be further reduced through individual preventive measures, such as staying indoors or wearing appropriate air-purifying respirators during acute exposure periods as well as equipping the home with air cleaners at all times [57].

Future research should address gaps in the literature in understanding the potential influences of exposure specificities and individual vulnerabilities to design more adjusted preventive interventions, hence better supporting stakeholder decision-making and promoting individual-specific measures.

Acknowledgements

This review was conducted in line of the Mobility, related pollutions, and Cardiovascular Health (MobiCard) project, funded by the French National Research Agency (ANR) as part of its Young Researchers (JCJC) 2023 funding program.

Financial support and sponsorship

None.

Conflicts of interest

There are no conflicts of interest.