Low-level pollution - health impacts: evidence review

4 Discussion

This section discusses the evidence, starting with studies identified in the preliminary and supplementary searches in order of publication.

4.1 Health Effects Institute (HEI) reports

Along with the Dominski review, the preliminary searches identified three significant reports produced by the Health Effects Institute (HEI) assessing the health impacts of low-level air pollution. Whilst these reports are not peer-reviewed, they represent a scale and level of evidence that warrants their inclusion in this report; additionally, the countries of focus are broadly comparable to Scotland in terms of pollution levels and income (Gross National Income [GNI] per capita).

In 2014, the HEI issued RFA 14-3, ‘Assessing Health Effects of Long-Term Exposure to Low Levels of Ambient Air Pollution’. The three health effects studies included all-cause mortality, cause-specific mortality and morbidity endpoints (Health Effects Institute, 2014). Three teams were responsible for the research led by three principal investigators:

4.1.1 European HEI study, Dr Bert Brunekreef (Utrecht University)

The ELAPSE study analysed the health outcomes related to exposure to low ambient air pollution concentrations below international guidelines in Europe (based upon U.S. EPA National Ambient Air Quality standards (NAAQs), and WHO Air Quality Guidelines for annual average concentrations). The study developed new exposure models for Europe and identified four pollutants of interest, PM_2.5, BC, NO₂, and O₃, and analysed four categories of physical health outcomes, including natural and cause-specific mortality, coronary and cerebrovascular events, lung cancer incidence, and asthma and chronic obstructive pulmonary disease (COPD) incidence. They analysed data from 15 study (pooled) cohorts and seven large administrative cohorts, producing rich individual-level data for ~325,000 participants and a large sample size of ~28 million. Almost all participants (in 2010) had annual average exposures below the European Union limit values for PM_2.5 (25 μg/m³) and NO₂ (40 μg/m³), and about 14% had exposures below the NAAQs for PM_2.5 (12 μg/m³)^[2]. Within the pooled cohort the average pollutant exposures were 15 μg/m³ PM_2.5, 1.5 × 10^-5/m BC, 25 μg/m³ NO₂, and 67 μg/m³ O₃. Among the administrative cohorts, mean concentrations of PM_2.5 ranged from 12 to 19 μg/m³, except for the Norwegian cohort (8 μg/m³).

The study found significant associations between PM_2.5, BC, and NO₂ exposure and natural-cause, cardiovascular, respiratory, and lung cancer mortality, as well as stroke, asthma, and COPD hospital admissions. The study also reported significant associations between NO₂ and acute coronary heart disease and between PM_2.5 and lung cancer incidence. The shape of the associations between exposure and natural-cause mortality showed steeper slopes at lower exposures, indicating increased risks for mortality at even the lowest observed concentrations; furthermore there were no concentration levels where associations were not found for PM_2.5, BC, and NO₂ (Brunekreef et al., 2021).

Note: this study did not include data from Scotland. Furthermore, there is limited data from England due to complicated data privacy laws (Brunekreef et al., 2021).

4.1.2 Canadian HEI study, Dr Michael Brauer (University of British Columbia)

The MAPLE study examined whether exposure to PM_2.5 at concentrations below the current U.S. air quality standards (12 μg/m³) was linked to an increased risk of nonaccidental death in 7.1 million Canadian adults. Satellite data, air quality monitoring data, and atmospheric modelling were combined to estimate outdoor PM_2.5 exposures across Canada from 1981 to 2016. In a large representative sample of Canadian adults, comprehensive epidemiological analyses were conducted to evaluate the risk of death at different PM_2.5 exposure levels and identify the lowest concentration at which associations with health effects could be detected. The results of the study found that long-term outdoor PM_2.5 exposures as low as 2.5 μg/m³ were linked to an increased risk of death, with variation across different geographical regions and with smaller effects when adjusted for O₃ concentrations. The study identified associations with health effects at PM_2.5 concentrations below the current ambient air quality standard of 12 μg/m³, suggesting that lowering the standard could yield further health benefits (Brauer et al., 2022).

4.1.3 American HEI study, Dr Francesca Dominici (Harvard University)

Dominici et al. (2019) examined the risk of mortality associated with exposure to low levels of ambient air pollution in 68.5 million older Americans (aged 65 and over). They created annual exposure models for three pollutants: PM_2.5, NO₂, and O₃, using a spatial resolution of 1 km x 1 km for the United States from 2000 to 2016. They used three causal inference approaches and two traditional regression approaches to analyse the data. They found that PM_2.5 was associated with an increased risk of all-cause mortality of 6% to 8% per 10 µg/m³, which was consistent across all five approaches. The effect estimates were larger in a low-exposure sub-cohort. The consistency of the associations across the methods suggests that long-term exposure to PM_2.5 is likely to have a causal effect on mortality, providing stronger evidence than previous studies (Dominici et al., 2019).

4.2 Dominski mapping review

The preliminary literature search found an essential piece of literature which formed the basis for the main search strategy described in the methods. Dominski et al. (2021) provided a comprehensive mapping review of systematic reviews and meta-analyses that assessed the relationship between air pollution and human health. They identified 240 relevant papers published globally, with most of the selected studies carried out in North America (n = 93), followed by Europe (n = 85), and Asia (n = 85). The majority of these geographies are broadly similar to Scotland in levels of air pollution. The total sample size considered in the reviews was 118,113,670 (123 reviews reported), with a mean of 960,273 (minimum of 376 and maximum of 28,215,394 people) and ages ranged from 0 to 120 years.

The Dominski review investigated five research questions (RQs):

1. RQ1. How many systematic reviews (SRs) and meta-analyses (MAs) have been published on the effects of air pollution on health? Is there any temporal trend? What is the geographical distribution of the authors?

2. RQ2. What methodological characteristics did the SRs and MAs present?

3. RQ3. What were the most commonly investigated health outcomes?

4. RQ4. Which air pollutants and environments were most commonly investigated?

5. RQ5. What is the relationship between health outcomes and air pollutants?

Regarding RQ5, concerning the relationship between health outcomes and air pollutants, 75% of reviews (180/240) showed a positive association, 53/240 (22%) studies were classified as ambiguous and only seven (3%) showed a negative association or no harmful effect(s). Of these seven the health outcomes they analysed were: asthma, childhood cancer, congenital heart defects, pneumonia, stroke, telomere length, and venous thrombosis.

Overall, the review found that exposure to air pollution is associated with a range of adverse health outcomes. The studies suggest that even low levels of PM exposure can increase the risk of respiratory and CVD, cancer, and premature death. Exposure to PM_2.5 was the most widely studied pollutant and an association was found with 8/10 health outcomes evaluated. The most frequently studied health outcome was CVD (32 reviews).

4.3 Discussion of results by health outcome

The order of the following sections is based upon the volume of evidence identified by the search strategy in this review. Therefore, this order does not necessarily reflect the body of evidence before 1 January 2020, nor does it reflect the potential public health impact or importance.

4.3.1 Cardiovascular disease (CVD)

As described in section 3, the results of the literature search only produced one paper that specifically looked at CVD, however this is likely an artefact of the search timeframe. The paper identified by the search strategy was So et al. (2022), which analysed the data from all Danish residents aged 30 and above (n=3,083,227). They found that long-term exposure to PM_2.5 (mean: 12.4 µg/m³), NO₂ (20.3 µg/m³), and/or BC (1.0 × 10^-5/m) was statistically significantly associated with all studied mortality outcomes (except for chronic kidney disease). Furthermore, their analysis showed that a 5 µg/m³ increase in PM_2.5 was associated with higher mortality due to CVD (1.09; 1.07-1.12) (So et al., 2023). This mirrors the evidence collated by Dominski et al. (2020) which presented several systematic reviews and meta-analyses, all of which showed that an increase in the concentration of particulate matter is associated with CVD and mortality, in the short- and long-term. The most commonly reported associated outcomes were ischemic heart disease, heart failure, ischemic stroke, and arrhythmia. Additionally, Brunekreef et al. (2021), Dominici et al. (2019) and Brauer et al. (2022) all report a significant positive association between air pollution, CVD, and mortality. This significant association on a combined sample of over 103 million shows the overwhelming consensus of evidence that air pollution is associated with both CVD and mortality.

Whilst the three HEI reports, the Dominski review, and the evidence identified by this review all provide significant evidence for the association and impact of air pollution on cardiovascular health globally, the current evidence available in Scotland is not conclusive. There are three published research papers that analyse the health impacts of low-level air pollution in Scotland – Yap et al. (2012), Willocks et al. (2012) and Lee et al. (2019), the latter two papers found no association between air pollution and CVD, which contrasts with most of the evidence globally.

Yap et al. (2012), investigated the association between exposure to black smoke (BS) air pollution and mortality in two Scottish cohorts over a 25-year period. The results show consistent associations between BS and mortality and more specifically significant associations with all-cause mortality, cardiovascular mortality, ischaemic heart disease and respiratory mortality. None of these associations were affected by adjustment for area-level deprivation but all were associated with inverse-distance monitoring, meaning the further the individual lived from the air pollution source the smaller the association. These results are consistent with the global evidence that air pollution has negative effects on human health, including CVD.

In contrast to Yap et al. (2012), Willocks et al. (2012) and Lee et al. (2019) found no association between air pollution and CVD in Scotland. Willocks et al. (2012) used an ecological time-series design, and an over dispersed Poisson log-linear model to examine if there is evidence of an association between short-term exposure to particulate matter (PM₁₀) and hospital admissions due to CVD, in Edinburgh and Glasgow during 2000 to 2006. The pollution data were lagged, relative to the admissions data, to ensure that the exposure occurred before the response. The results of this analysis showed no consistent associations between PM₁₀ concentrations and cardiovascular hospital admissions as all the estimated relative risks were close to one, and all but one of the associated 95% confidence intervals contained the null risk of one.

However, the absence of evidence of an association in this study does not prove that there is no association. The wider literature, from countries similar to Scotland, establishes a strong association between CVD and air pollution, both in the short- and long-term (Brunekreef et al., 2021; Dominski et al, 2020). The absence of an association found by Willocks et al. (2012) may be due to several limitations in their methodology. Firstly, they only collected data on PM₁₀, therefore the full spectrum of air pollutants that are harmful to human health and linked to CVD are not considered. For example, PM_2.5 has been shown to be more strongly associated with adverse health outcomes than PM₁₀ (UKHSA, 2022). This in part is because PM_2.5 has a smaller diameter and can therefore penetrate deeper into the lungs and cross into the bloodstream. Additionally, BC, NO₂, and O₃ have all been strongly associated with adverse health outcomes including CVD. In addition to the absence of data on other important pollutants, the data on PM₁₀ is incomplete. On 12.9% (Edinburgh) and 1.1% (Glasgow) of days no PM₁₀ measurements were available. As a result, Willocks et al. (2012) only used data for days where at least one pollution measurement was available. The monitoring data was downloaded from the Scottish Air Quality website from sites throughout Scotland. Whilst the incomplete data poses issues around spatial representivity of the monitoring data compared to the population data, incomplete air pollution monitoring data is a common occurrence throughout the wider literature (Chastko and Adams, 2019), and does not negate the relevance of the research. The authors also discuss the possibility that the study design did not provide enough statistical power to detect a pollution-health relationship. The number of CVD admissions was relatively low – interquartile range (IQR) 4 to 8 cases per day in Edinburgh, and between 8 and 13 cases per day in Glasgow. Finally, the study only analysed hospitalisations and the long-term effects of air quality on CVD were not discussed, despite being widely accepted and documented elsewhere in the literature (Willocks et al., 2012).

Lee et al. (2019) reviewed the epidemiological evidence quantifying the health impact of air pollution in Scotland. They analysed the relationship of four pollutants (PM₁₀, PM_2.5, NO₂, and NOx) against cardio-respiratory disease and total non-accidental mortality outcomes. The study used data from mainland Scotland across a two-year period (2015-2016) with the study region spatially partitioned into 1,252 Intermediate Zones (IZ). The results show that all pollutants exhibit significant associations with respiratory disease but not CVD.

When evaluating the absence of the association with CVD it is important to consider how the authors analysed the effects of confounding factors. Lee et al. (2019) discuss smoking, which has been widely evidenced to affect cardio-respiratory disease incidence. Data on smoking prevalence was not available at IZ scale so their model used the Scottish Index of Multiple Deprivation (SIMD) score as a proxy for smoking. The justification for this decision is a paper by Kleinschmidt et al. (1995), however the current (2019) association between smoking and socio-economic status was not discussed and therefore this may have affected the validity of the analysis. The Scottish Health Survey (2021) shows that the age-standardised prevalence of current smoking status was higher among adults living in more deprived areas than among those living in less deprived areas (24% and 5%, respectively). This pattern has been evident since 2003 in the Scottish Health Survey data, and for longer elsewhere, however, all categories of deprivation have seen a reduction in smoking rates over time. Whilst the association between smoking status and socio-economic status cannot be refuted, the strength of the association has changed over time. There is no evidence that CVD or total non-accidental mortality are associated with any of the four pollutants, as the 95% credible intervals contain the null risk of one, and the estimated risks are mostly very close to one. However, it is possible that the adjustment for confounders may be oversensitive due to the historic data used to quantify the association. The confounders, including SIMD, were retained in all models but the significance of these variables is not stated (Lee et al., 2019).

In conclusion, whilst Willocks et al. (2012) and Lee et al. (2019) did not find an association between air pollution and CVD, this is likely an artefact of the study design and data. These limitations explain in part why the results are different from the earlier research by Yap et al. (2012) and the vast consensus of global research and data.

4.3.2 Mental health and well-being

As stated, the most evaluated health outcome in our search results was mental health and well-being, and this is likely reflective of the search strategy timescale. The evidence on the physiological health impacts of air pollution has been established for several years, therefore the majority of these papers would have been published prior to 1 December 2020. The timescale of this search evidences the increased focus on research investigating the relationship between other health outcomes, including mental health and well-being, in recent years. There is a good body of evidence on the association between air pollution and mental health and well-being and studies have consistently shown that exposure to air pollutants is linked to various negative mental health outcomes.

The search identified three studies which analysed the association between air pollution and mental health and well-being, specifically in individuals aged 18 or younger. Mok et al. (2021) analysed a cohort of individuals born in Denmark between 1 January 1979 and 31 December 2006 (n = 1,424,670), with information on estimated daily exposures to PM_2.5 and NO₂ at residence from birth to their tenth birthday. They found that higher exposure to PM_2.5 and NO₂ during childhood increased the risk of self-harm. Specifically, the risk of self-harm was elevated by 1.45-fold for individuals exposed to 17–19 μg/m³ of PM_2.5 on average per day and by 1.59-fold for those exposed to higher levels (>19 μg/m³), when compared with a mean daily exposure of <13 μg/m³. In addition, higher mean daily exposure to NO₂ was associated with increased self-harm risk, but the dose-response relationship observed was less clear than for PM_2.5 (Mok et al., 2021). Two other studies (Reuben et al., 2021; Latham et al., 2021), used data from the Environmental-Risk Longitudinal Twin Study, a population-based cohort study of 2,232 children born between 1 January 1994 and 4 December 1995 across England and Wales to analyse the impact of air pollution on mental health and well-being. Reuben et al. (2021) discovered that higher levels of outdoor NOx exposure during adolescence were associated with a greater impact on psychopathology during the transition to adulthood. This suggests that NOx may be a nonspecific risk factor for the development of psychopathology (Reuben et al., 2021). Latham et al. (2021) identified a higher risk of major depressive disorder in adolescents with greater exposure to NOx and PM_2.5. The risk of major depressive disorder was highest for participants within the top quartile of annual exposure to NOx (adjusted OR = 1.43) and PM_2.5 (adjusted OR = 1.35). This indicated the potential role for childhood ambient air pollution exposure in the development of adolescent major depressive disorder (Latham et al., 2021).

In conclusion, this evidence suggests that air pollution has significant implications for mental health and well-being in individuals aged 18 or younger. Whilst further research is needed to better understand the mechanisms underlying these associations and inform effective interventions, exposure to air pollution during childhood/adolescence is associated with the development of mental health conditions and there is a need to mitigate the adverse effects.

This association has also been investigated in adults. Li et al. (2023) used data collected between March 2006 and October 2010 from 354,897 participants aged 37 to 73 years from the UK Biobank. During a median follow-up of 9.7 years, they observed that PM_2.5 and NOx were associated with an increased risk of major depressive disorder. Specifically, each 5 μg/m³ increase in PM_2.5 was associated with a hazard ratio of 1.16, and each 20 μg/m³ increase in NOx was associated with a hazard ratio of 1.02 for major depressive disorder. There was also a significant interaction between genetic susceptibility and air pollution, indicating that individuals with high genetic risk and high PM_2.5 exposure had the highest risk of developing major depressive disorder. Additionally, there was an interaction between PM_2.5 exposure and an unhealthy lifestyle, with participants having the highest major depressive disorder risk when they had both the least healthy lifestyle and high air pollution exposures (D. Li et al., 2023). Petrowski et al. (2021) investigated the effects of air pollution, specifically PM₁₀, on determinants of mental health and well-being, measured by life satisfaction, stress resilience, anxiety, depression, and self-esteem, in a representative sample of 3,020 German adults. A multivariate linear regression analyses show that higher life satisfaction, more self-esteem and higher stress resilience are predicted by less air pollution (PM₁₀). Confounders including individual income, age, and gender were adjusted for (Petrowski et al., 2021). Similarly, outcomes from "Understanding-Society: The-UK-Household-Longitudinal-Study" were linked to air pollution data and showed that with every 10 μg/m³ increase NO₂, SO₂, PM₁₀ and PM_2.5 the likelihood of poor mental well-being increased. Additionally, living in more polluted local authorities (SO₂, PM₁₀, and PM_2.5 ) as well as belonging to a non-UK born ethnic group increased the odds of poor mental well-being (Al Ahad et al., 2022)_. In conclusion, the research identified also provides positive evidence of the association between air pollution and mental health and well-being in adults. These findings underscore the importance of addressing air pollution as a potential risk factor for mental health issues in both adults and children.

There is also evidence for short-term associations between air pollution and mental health and well-being. Newbury et al. (2021) used a retrospective cohort study to assess the longitudinal association between air pollution exposure and mental health service use among individuals with first presentations of psychotic and mood disorders. Between 2008 and 2012, 13,877 individuals (≥15 years) had first contact with the South London and Maudsley NHS Foundation Trust for psychotic and mood disorders. High-resolution estimates of NO₂, NOx, PM_2.5 and PM₁₀ were linked to the individual’s residential address. Interquartile range increases in NO₂, NOx and PM_2.5 were associated with 18%, 18% and 11% increased risk for in-patient days respectively. Similarly, interquartile range increases in NO₂, NOx, PM_2.5 and PM₁₀ were associated with 32%, 31%, 7%, and 9% increased risk respectively for community mental health service events after 1 year, and these associations persisted after 7 years. This analysis indicates that there is an association between residential air pollution exposure and increased mental health service use among people recently diagnosed with psychotic and mood disorders (Newbury et al., 2022). Qiu et al. (2022) also analysed the effects of short-term increases in air pollution (PM_2.5 and NO₂) on acute psychiatric hospital admissions, but in adults aged 65 years and older in the USA. They observed that short-term increased exposure is associated with increased risk of acute hospital admissions for psychiatric disorders. For every 5 μg/m³ increase in PM_2.5, there is a corresponding increase in hospital admission rates of 0.62% for depression, 0.77% for schizophrenia, and 1.19% for bipolar disorder. Similarly, for every 5 parts per billion (ppb) increase in NO₂, there is an increase in hospital admission rates of 0.35% for depression and 0.64% for schizophrenia (Qiu et al., 2022). A study analysing air pollution and psychiatric emergency room visits in Sweden also demonstrated a significant association between air pollution and an increased number of psychiatric emergency room visits, specifically PM_2.5 and NO₂. Furthermore, the study identified individuals with pre-existing mental health conditions, males, and individuals aged 65 or older as groups with increased susceptibility (Muhsin et al., 2022). Following coal mine fires in Australia, Carroll et al. (2022) looked specifically at the impacts of PM_2.5 levels on the utilisation of ambulance and hospital services for mental health conditions in the local population. They found that PM_2.5 exposure was associated with an increase in access for mental health services. Specifically, a 10 μg/m³ increase in daily PM_2.5 was associated with a 38% rise in ambulance attendances for anxiety and a 26% rise in emergency department presentations for depression. This suggests that health services should expect to have an increase in the number of people seeking assistance for mental health conditions during extreme air pollution events (Carroll et al., 2022).

Alongside the general link between air pollution and mental health and well-being, evidence indicates a clear link between air pollution and short-term mental health and well-being. Studies examining different populations, including individuals with first presentations of psychotic and mood disorders, older adults, air pollution events, and psychiatric emergency room visits, consistently show associations between residential air pollution exposure and increased mental health service use.

While causality cannot be assumed, the evidence clearly indicates an association between air pollution and negative effects on mental health and well-being in a wide variety of populations and settings. These findings emphasize the importance of addressing air pollution as a significant environmental factor affecting mental health and well-being.

4.3.3 Dementia

Multiple studies have provided evidence linking air pollution to the development and exacerbation of dementias. In the French population-based cohort known as the 3C Study, an association was found between PM_2.5 exposure and various forms of dementia, including all-cause dementia, Alzheimer’s disease, and vascular/mixed dementia. Specifically, for every 5 µg/m³ increase in PM_2.5, the hazard ratio was 1.20 for all-cause dementia, 1.20 for Alzheimer's disease, and 1.33 for vascular/mixed dementia. However, no significant association was observed between NO₂ or BC exposure and dementia risk in this study (Mortamais et al., 2021). Similar findings were reported in Sweden by Kriit et al. (2021), who estimated that 5% of annual dementia cases in Sweden were attributable to PM_2.5 exposure, equating to 820 cases per year.

In the UK Biobank data analysis, which included a large sample size of 187,194 individuals, consistent associations were observed between PM_2.5 exposure and increased risk of all-cause dementia and Alzheimer's disease. Specifically, each increase in PM_2.5 was associated with a hazard ratio of 1.17 for both all-cause dementia and Alzheimer's disease. PM_2.5 is not the only air pollutant which is associated with dementia. NO₂ exposure was found to be associated with a higher risk of any incident dementia, Alzheimer's disease, and vascular dementia (n= 187,194) (Parra et al., 2022).

Further evidence from the Betula project, a longitudinal study on aging, memory, and dementia in Sweden, demonstrated that every 1 μg/m³ difference in annual mean PM_2.5 concentration was associated with a hazard ratio of 1.23 for dementia after adjusting for factors including age, sex, apolipoprotein E (APOE) gene (a risk factor for Alzheimer’s disease), Scandinavian Odor Identification Test (SOIT), cardiovascular diseases and education (Andersson et al., 2023).

As well as the association between air pollution and dementia incidence, air pollution has been evidenced to exacerbate symptoms of Parkinson's disease. An analysis of hospital admission data from 18 French areas showed a small but significant influence of PM₁₀, PM_2.5 and NO₂ on all-ages Parkinson's disease hospital admissions, indicating that air pollution contributes to the worsening of Parkinson's disease symptoms (Goria et al., 2021).

Zhao et al. (2021) assessed the relationship between O₃ exposure and mortality attributable to four neurological diseases, including Parkinson’s disease and dementia. Their analysis utilised data from the 2001 Canadian Census Health and Environment Cohort, comprising 3.5 million adults with a combined follow-up of 51,045,700 person-years. The results demonstrated positive associations between O₃ exposure and mortality from Parkinson’s disease and dementia, even after adjusting for demographic, socioeconomic, environmental, and contextual factors. Specifically, each interquartile range increase in O₃ (10.1 ppb) was associated with hazard ratios of 1.09 for Parkinson's disease and 1.08 for dementia. Between 2001 and 2016, 5.7% and 5.0%, of deaths from Parkinson's disease and dementia, respectively, were attributable to O₃ exposure (N. Zhao et al., 2021). The other neurological outcomes assessed are discussed in section 1.4.5..

In summary, multiple studies identified in this review provide consistent evidence linking air pollution, particularly PM_2.5, to an increased risk of dementia and potentially the exacerbation of Parkinson's disease symptoms. These findings underscore the importance of addressing air pollution as a public health concern and mitigating the burden of these neurodegenerative conditions.

4.3.4 Cognition

This review identified several publications that demonstrate the detrimental effects of air pollution on cognition across various populations and age groups.

Thompson et al. (2023) conducted a systematic review and meta-analysis that identified and analysed eighty-six studies assessing the effects of air pollution on cognition including attention, memory, language skills, and academic performance. Whilst the literature identified was contradictory there is moderate evidence for harmful associations between PM_2.5 and general cognition in adults over 40 years old. Additionally, they found associations between PM_2.5, NOx, and PM₁₀ exposure with executive function, particularly working memory, in children. Furthermore, O₃ was associated with general cognition in adults aged 40 and above, while NOx showed links with reasoning/IQ in children. Moreover, a 1 μg/m³ increase in PM_2.5 was significantly associated with lower verbal fluency and decreased performance in executive function tasks (Thompson et al., 2023).

As with dementia, the harmful effects of air pollution on cognition and dementia are evident. Wu et al. (2022) use data from the Swedish National Study on Aging and Care in Kungsholmen (SNAC-K) (n= 1,987) to demonstrate the harmful effects of long-term exposure to ambient air pollutants incidence of cognitive impairment, not dementia (CIND). Their analysis showed that a 1 μg/m³ increase in PM_2.5 exposure was associated with a 75% increased risk of incident CIND, while weaker associations were found for PM₁₀ (hazard ratio for 1-μg/m³ = 1.08) and NOx (hazard ratio for 10 μg/m³ = 1.18). Interestingly, they demonstrate that air pollution is a risk factor for the progression of CIND to dementia. Grande et al. (2021) also analysed SNAC-K data and found an association between air pollution, particularly PM_2.5 up to 8.6 μg/ m³, and cognitive decline. They also observed that the presence of cerebrovascular diseases further amplified the risk associated with air pollution by 6% (Grande et al., 2021).

Examining the combined effects of air pollution and deprivation on cognition, Christensen et al. (2022) analysed data from the Emory Healthy Aging Study (EHAS) comprising 11,897 participants aged 50 years and above. They observed that individuals residing in areas with higher air pollution concentrations and higher deprivation levels exhibited higher cognitive function instrument (CFI) scores, indicating increased perceived memory and cognitive decline. Notably, even after accounting for confounding factors related to deprivation, there were clear synergistic effects of air pollution and deprivation on cognition (Christensen et al., 2022).

The detrimental effects of air pollution on cognition extend to prenatal exposure. Zhang et al. (2022) examined 348 mother-child dyads from New York City and Boston, assessing the associations between prenatal PM_2.5 exposure and childhood cognition. The study found associations between PM_2.5 exposure during the third trimester of pregnancy and cognition scores in children, with variations observed based on location and child sex. The pooled analysis also indicated that maternal education and urbanicity may modify the association between prenatal PM_2.5 exposure and childhood cognition (Zhang et al., 2022).

Like deprivation, residential greenness was found to play a role in the relationship between air pollution and cognition. Wang et al. (2023) utilised data from the Longitudinal Study of Australian Children, involving 6,220 adolescents and 2,623 mid-life adults to assess the effects of air pollution and greenness on cognitive function, working memory and executive function. Within the adolescents, an IQR increase of NO₂ was associated with increased odds (19–24%) of having poorer executive functions, associations weren't observed between air pollution and other outcomes. For the adults, high NO₂ exposure predicted poorer cognitive function across all outcomes, while high PM_2.5 predicted poorer attention only. Increased levels of residential greenness were associated with better executive function in the minimally adjusted models for both adolescents and adults. However, these associations disappeared when fully adjusted for other factors, indicating that while residential greenness may have some positive effects, it does not eliminate the harmful association between air pollution and cognition (Wang et al., 2023).

In summary, the papers identified support the harmful effects of air pollution on cognition across various populations, including adults, children, and individuals exposed prenatally. Studies consistently demonstrate associations between air pollution exposure and cognitive impairments, including attention, memory, language skills, and academic performance. This emphasises the need for effective strategies to reduce air pollution and mitigate impact on cognitive health across population groups.

4.3.5 Neurological

As discussed in section 1.4.3, alongside cognitive and dementia health outcomes, the wider neurological effects of air pollution have been assessed. Zhao et al. (2021) assessed the relationship between O₃ exposure and mortality attributable to neurological diseases. Their analysis utilised data from the 2001 Canadian Census Health and Environment Cohort, comprising 3.5 million adults with a combined follow-up of 51,045,700 person-years. The results demonstrated positive associations between O₃ exposure and mortality from four neurological diseases (Parkinson’s disease, dementia, stroke and multiple sclerosis), even after adjusting for demographic, socioeconomic, environmental, and contextual factors. Specifically, each interquartile range increase in O₃ (10.1 ppb) was associated with hazard ratios of 1.06 for stroke and 1.35 for multiple sclerosis. Notably, the significance of these associations was unaffected by most covariates, except for adjustment for Canadian regions. Between 2001 and 2016 3.8% and 19.1% of deaths from stroke and multiple sclerosis, respectively, were attributable to O₃ exposure (N. Zhao et al., 2021).

Research also suggested that air pollution, including CO, NO_2, SO₂, O₃ and PM_2.5, are associated with an increased number of emergency department visits for nervous system disorders. The analysis used data from a large Canadian city between 2004 and 2015 and demonstrated that increased levels of ambient CO, PM_2.5, and SO₂ were linked to an elevated risk of emergency department visits for episodic and paroxysmal diagnoses, particularly migraines. This association was observed at various time lags between 0-15 days, with significant effects observed on days 1, 6, and 7 following exposures (Lukina et al.., 2022). Similarly Min et al. (2023) analysed data from New York and found that a 1 μg/m³ increase in PM_2.5 (lagged 0-1 years) was associated with an increased risk of headache and convulsion by 1.06 and 1.04, respectively (Min et al., 2023). This evidence highlights the short-term impact of ambient air pollution on neurological health.

Other neurological conditions linked to air pollution include autism spectrum disorder and attention deficit hyperactivity disorder. Li et al. (2023) analysed data from a longitudinal study of 2,750 children in the Netherlands and found evidence that higher levels of exposure to PM were associated with more severe autism spectrum disorder and attention deficit hyperactivity disorder symptoms. While no other associations were observed, this research provides some evidence about associations between PM and neurodevelopmental diseases among adolescents (Y. Li et al., 2023).

The evidence presented in this section provides further evidence of the link between air pollution and Parkinson's disease, dementia and stroke, and evidence of an association with other neurological outcomes including multiple sclerosis, emergency department visits for nervous system disorders, autism spectrum disorder and attention-deficit/hyperactivity disorder in children. These findings highlight the detrimental effects of air pollution on neurological health, both in terms of mortality and the incidence and severity of neurodevelopmental and neurological disorders.

4.3.6 Respiratory

Numerous studies have investigated the relationship between air pollution and respiratory health. The Dominski review notes that most of the research has focussed on respiratory health outcomes, with harmful effects being described and evidenced in a wide range of populations and for a range of respiratory outcomes. This report identified four additional studies investigating this association.

Anenberg et al. (2022) considered the effects of air pollution on paediatric respiratory health. Using existing global annual average NO₂ concentration datasets for 2010-2012 they linked and compared these to population and baseline asthma rates to estimate NO₂ attributable paediatric asthma incidence. Their analysis estimated that 1.85 (95% CI 0.93-2.80) million new paediatric asthma cases were attributable to NO₂ globally in 2019, with two thirds of these occurring in urban areas. The proportion of paediatric asthma cases attributable to NO₂ in urban areas decreased from 19.8% in 2000 to 16.0% in 2019. Variations in the decline of urban attributable fractions were observed across different regions, with decreases in high-income countries, Latin America and the Caribbean, central Europe, eastern Europe, central Asia, southeast Asia, east Asia, and Oceania, and increases in south Asia, sub-Saharan Africa, and north Africa and the Middle East. The changes in NO₂-attributable paediatric asthma rates were influenced by NO₂ concentrations, paediatric population size, and asthma incidence rates, which varied regionally (Anenberg et al., 2022). Zhao et al. (2021) also analysed respiratory impacts in children by analysing data from 915 children within the GINIplus and LISA birth cohorts from Munich (n = 181) and Wesel (n = 734) Germany. They found that per interquartile range increase in air pollutants during the first year of life, forced expiratory volume in one second (FEV1) z-scores declined annually by -0.012 for PM_2.5 to -0.023 for the PMcoarse fraction^[3]. Effect estimates of infancy exposure for certain air pollutants were higher for groups with asthma, older maternal age, and breastfeeding <12 weeks than their counterparts. These results indicate that infancy exposure to higher air pollution may reduce lung function development up to adolescence (15 years) (Q. Zhao et al., 2021). The studies conducted by Anenberg et al. (2022) and Zhao et al. (2021) provide valuable insights into the detrimental effects of air pollution on paediatric respiratory health, adding to the large consensus identified in the Dominski review.

Soarez and Silva (2022) reviewed the effects of ground-level O₃ on respiratory health. They identified 59 eligible studies, of which 83% presented significant correlations between O₃ with asthma, COPD, or acute respiratory distress syndrome (ARDS). The eligible studies that reported negative non-significant associations discussed a lack of data or topographic differences as the main issue with these divergent results (Soares and Silva, 2022).

It is widely accepted that air pollution may have harmful effects on pre-existing respiratory conditions. One study aimed to investigate personal exposure to air pollution and its effects on the respiratory health of COPD patients in London. A small number of patients (n=115) carried a personal monitor specifically designed to measure temperature, NO_2, O₃, NO, CO, PM_2.5 and PM₁₀. Individuals also recorded daily information on respiratory symptoms and measured peak expiratory flow ( They found that gaseous air pollutants were associated with a deterioration in COPD patients’ health. They observed an increase of 16.4%, 9.4% and 7.6% in the odds of exacerbation for an interquartile range increase in NO₂, NO and CO, respectively. O₃ was observed to have adverse associations with peak expiratory flow and breathlessness. No association was observed between particulate matter and any outcome (Evangelopoulos et al., 2021). This evidence highlights the importance of considering the effects of air pollution on respiratory health for both individuals with diagnosed conditions, as well as those who may develop health outcomes due to exposure.

In conclusion, the Dominski review provided a strong evidence base for the harmful effects of air pollution on respiratory health. The studies reviewed here are consistent with previous findings. This evidence highlights the need for comprehensive measures to mitigate air pollution and protect respiratory health, considering both pre-existing conditions and potential future health risks for individuals exposed to pollutants.

4.3.7 Development

As discussed, the harmful effects of air pollution on cognition were observed in studies looking at prenatal exposure (Zhang et al., 2022). In addition, prenatal exposure to air pollution has been associated with harmful effects on a child’s neuropsychological functioning. Iglesias-Vasquez et al. (2022) analysed and modelled data from 473 mother-child pairs within the ECLIPSES study in Catalonia, Spain. Their model, once adjusted for maternal biological, sociodemographic and lifestyle characteristics, area deprivation index, and amount of greenness around the home’s address, showed that all air pollutants assessed, except PM_2.5 absorbance, were associated with lower motor function in children, although no association between prenatal exposure to air pollution and cognitive and language functions was observed (Iglesias-Vázquez et al., 2022).

However, not all research on pre-natal air pollution has found evidence of harmful associations. Starling et al. (2022) calculated the average PM_2.5 and O₃ in each trimester of pregnancy to estimate the effects on DNA methylation detectable at birth. Whilst some differentially methylated regions were identified, no statistically significant associations were observed (Starling et al., 2022).

In addition to prenatal exposure, there is evidence for the harmful effects of air pollution exposure during early life and childhood. Ahmed et al. (2022) used data from the Mothers and their Children’s Health (MatCH) study, a 2016/17 sub-study from a prospective longitudinal study, the Australian Longitudinal Study on Women’s Health, and found that children who were exposed to moderate and high levels of PM_2.5 had higher odds of emotional and behavioural problems, and gross motor delay, compared with children with low levels of exposure. Children’s lifetime exposure to ‘moderate levels’ of PM_2.5 (5.9–7.1 µg/m³) was associated with 1.27-fold higher odds of emotional/behavioural problems. There was not enough evidence to demonstrate an association between NO₂ exposure or living within 200 m of major roads (Ahmed et al., 2022).

In conclusion, the effects of prenatal exposure to air pollution on prenatal and childhood development presented in these studies is inconclusive and more comprehensive studies are needed to better understand the complex relationship.

4.3.8 Mortality

Three studies investigated the association between air pollution and mortality. Bai et al. (2022) investigated the relationship between outdoor PM_2.5 exposure and mortality. Their analysis showed that every 10 μg/m³ increase in PM_2.5 is associated with approximately two deaths, of which 31.7% were attributable to diabetes and major cardiovascular events. Specifically, 4.5% were explained by PM_2.5‐induced diabetes, 22.8% by PM_2.5‐induced major cardiovascular events, and 4.5% through their interaction (Bai et al., 2022). A large Danish study analysed data from over 3 million residents aged 30 or older from 2000 to 2017, alongside estimated annual mean concentrations of PM_2.5, NO₂, BC and O₃. Their findings showed that long-term exposure to PM_2.5, NO₂ and/or BC was significantly associated with mortality from various causes. An increase of 5 µg/m³ in PM_2.5 was associated with higher mortality rates for all-natural causes, cardiovascular disease, respiratory disease, lung cancer, diabetes, dementia, psychiatric disorders, asthma, and acute lower respiratory infection. The associations with long-term exposure to O₃ were generally negative, meaning the more O₃ the less harmful effects on health. The study’s results remained significant even after adjusting for demographic and socioeconomic factors, as well as indirect adjustment for smoking and body mass index (So et al., 2022).

So et al. (2023) also used this data set to evaluate the associations between mortality and long-term exposure to eight PM_2.5 elemental components (copper, iron, zinc, sulphur, nickel, vanadium, silicon, and potassium). Their analysis found significant positive associations between all-natural mortality with silicon and potassium with most causes of mortality, excluding chronic kidney disease and potassium, and diabetes and silicon. The strongest associations were for psychiatric disorder mortality and in addition, iron was relevant for mortality from respiratory diseases, lung cancer, chronic kidney disease, and psychiatric disorders; zinc with mortality from chronic kidney disease, respiratory disease, and lung cancer, and nickel and vanadium with lung cancer mortality (So et al., 2023).

In conclusion, the three studies provide evidence for the association between air pollution and mortality. Bai et al. (2022) found a significant relationship between PM_2.5 exposure and mortality, with a substantial proportion of deaths attributable to diabetes and major cardiovascular events. The Danish study by So et al. (2022) revealed that long-term exposure to PM_2.5, NO₂, and BC was associated with increased mortality from various causes, while O₃ showed a generally negative association. Furthermore, So et al. (2023) identified significant positive associations between mortality and exposure to specific PM_2.5 elemental components. These findings highlight the adverse health effects of air pollution on mortality, in addition to the specific health outcomes discussed in this report.

4.3.9 Cancer

The Dominski review identified two papers assessing the relationship between air pollution and cancer and a relative lack of literature in comparison to other health outcomes. Additionally, the review identified no evidence associating air pollution and childhood cancer. However, the two papers identified demonstrated an association between PM_2.5 and PM₁₀ and breast cancer mortality; namely each 10 μg/m³ of PM_2.5 was associated with 1.17-fold increase in risk and each 10 μg/m³ of PM₁₀ was associated with 1.1-fold increased risk (Dominski et al., 2021). The search conducted in this report identified two additional papers focusing on the association between cancer and air pollution, both of which demonstrated a positive association. Yim et al. (2022) analysed global trends in lung cancer incidence over a 22-year period, from 1990 to 2012, and investigated the association between air pollution and lung cancer rates alongside the risk from smoking. Their analysis shows that the global decrease in lung squamous cell carcinoma incidence is associated with reduced tobacco consumption whereas the increase in lung adenocarcinoma incidence is associated with air pollution. More specifically there is evidence of the potential involvement of black carbon and sulphate in the pathogenesis of lung adenocarcinoma (Yim et al., 2022).

Interestingly, whist the Dominski review found no evidence of synthesis reviews that analysed the association between air pollution and childhood cancer, Huls et al. (2023) analysed the environmental, social and behavioural risk factors associated with spatial clustering of childhood cancer incidence across 159 counties in Georgia (USA). Their analysis demonstrated consistent associations of environmental (pesticide exposure) and social/behavioural stressors (low socioeconomic status, alcohol) with spatial clustering of paediatric cancer class II (lymphomas and reticuloendothelial neoplasms), but not for other cancer classes. Whilst an association between childhood cancers was demonstrated in the analysis, a direct link with air pollution was not established, as the environmental data on air pollution was clustered with other environmental risk factors (Hüls et al., 2023).

The association between air pollution and cancers is evidenced in the literature, however the strength of the association is not as well established as it is for other health outcomes. Furthermore, the relationship more specifically with childhood cancers has limited evidence and needs further assessment.

4.3.10 Neonatal

Recent research has begun to consider the effects of air pollution during neonatal development. Ren et al. (2023) investigated the spatiotemporal changes in neonatal disease burden caused by PM_2.5 at a global level, considering the socio-demographic index (SDI), gender and nationality. Using data from the Global Burden of Disease Study database, they found that the burden of neonatal diseases associated with PM_2.5 has increased since 1990, particularly in regions with lower SDI, such as South Asia and Sub-Saharan Africa. Neonatal preterm birth was identified as the leading cause of preterm death. The global age-standardized mortality rate was 2.09 per 100,000 population in 2019. The study highlights PM_2.5 as a significant environmental hazard for new-born diseases (Ren et al., 2023). Zhao et al. (2023) also used the Global Burden of Disease Study database to examine the spatiotemporal trends in the burden of neonatal disorders caused by ambient and household PM_2.5 at global, regional, and national levels from 1990 to 2019. In 2019, approximately one-fifth of global neonatal disorders was attributed to PM_2.5 exposure, with ambient PM_2.5 accounting for 7.5% and household PM_2.5 for 13.2%. The burden of neonatal disorders related to household PM_2.5 decreased between 1990 and 2019, while that related to ambient PM_2.5 increased, particularly in regions with lower SDI. South Asia and East Asia had the highest rates and population attributable fractions (PAFs) of ambient PM_2.5-related neonatal disorders DALYs in 2019. There was an inverted V-shaped relationship between rates and population PAFs of ambient PM_2.5-related neonatal disorders DALYs in 2019, as well as their corresponding estimated annual percentage change (EAPCs), with SDI, whereas rates and PAFs of household PM_2.5-related neonatal disorders DALYs in 2019 showed a highly negative correlation with SDI. Overall, the burden of ambient PM_2.5-related neonatal disorders has significantly increased over the past three decades, particularly in regions with lower SDI, while household PM_2.5-related burden has decreased but still represents a substantial portion of the overall burden of PM_2.5-related neonatal diseases (Zhao et al., 2023).

In conclusion, recent studies have shed light on the concerning impact of air pollution, particularly PM_2.5, on neonatal health, however further evidence is needed to understand the extent of this impact and the underlying biological mechanisms.

4.3.11 Type-2 Diabetes

Air pollution has emerged as a significant environmental risk factor associated with various health conditions, including type-2 diabetes. A systematic review and meta-analysis by Yang et al. (2020) compiled the evidence showing significant associations of PM_2.5, PM₁₀ and NO₂ with type-2 diabetes incidence and prevalence. Also, whilst the data was too heterogeneous for meta-analysis the majority of studies on glucose-homoeostasis markers also showed increased risks with higher air pollutants levels (Yang et al., 2020). Liu et al. (2021) evaluated the spatio-temporal changes in type-2 diabetes incidence and prevalence attributed to PM_2.5 from 1990 to 2019 in 204 countries and regions. Their analysis found that overall the global burden of type-2 diabetes attributable to PM_2.5 increased significantly since 1990, particularly within specific demographic groups including the elderly, males, as well as individuals living in Africa, Asia and low-middle SDI regions (Liu et al., 2021). However, declining trends were observed in North America, South America, Europe, Australia, and other regions with high SDI. Bai et al. (2022) investigated the relationship between outdoor PM_2.5 exposure and mortality. Of over 200,000 individuals analysed in Ontario, Canada, from 1996 to 2014, they estimated that each 10 μg/m³ increase in PM_2.5 was associated with approximately two incident cases of diabetes. Further to this, every 10 μg/m³ increase in PM_2.5 is associated with approximately two deaths, of which 31.7% were attributable to diabetes and major cardiovascular events. Specifically, 4.5% were explained by PM_2.5‐induced diabetes, 22.8% by PM_2.5‐induced major cardiovascular events, and 4.5% through their interaction (Bai et al., 2022).

The Dominski review, as well as the evidence discussed, supports a significant association between air pollution, specifically PM_2.5, and the incidence and prevalence of type-2 diabetes.

4.3.12 Ocular

Only one study from the search analysed the effects of air pollution on ocular outcomes such as visual impairment and age-related eye disease. This study utilised data from the Canadian Longitudinal Study on Aging, including 30,097 adults aged 45 to 85 years, alongside annual mean PM_2.5 data estimated from satellite data to residents’ home address. The analysis showed that increased PM_2.5 levels were associated with visual impairment, glaucoma, and visually impairing age-related macular degeneration. In the single pollutant models, increased PM_2.5 levels (per interquartile range) were associated with visual impairment (odds ratio [OR] = 1.12), glaucoma (OR = 1.14), and visually impairing age-related macular degeneration (OR = 1.52) after adjustment for sociodemographic variables and co-morbidities. There was also limited evidence of a week association (after adjustment) with cataract (OR = 1.06). Whilst based upon one study, there is an observed associations between PM_2.5 and ocular outcomes, suggesting the need for further studies to confirm these associations and explore potential mechanisms (Grant et al., 2021).

4.3.13 Primary care healthcare service use

One study analysed the effects of air pollution on short term primary and pharmaceutical care usage. Following several mine fires between February and March 2014 births, general practitioner (GP) presentations and prescription dispensing for children born in the Latrobe Valley, Australia. They observed that exposure to fire related PM_2.5 during pregnancy was associated with an increase in the dispensing of systemic steroids, while exposure during infancy was linked to antibiotic dispensing. Additionally, even at relatively low levels, exposure to ambient PM_2.5 during infancy was associated with an increase in antibiotics and GP presentations, independent of exposure to the fire. Sex-specific differences were observed, with stronger associations in girls for GP presentations and stronger associations in boys for steroid skin cream dispensing. These findings highlight the potential health effects of PM_2.5 exposure during critical developmental periods and the importance of considering sex differences in susceptibility (Ziou et al., 2023).

4.3.14 Contradictory evidence

This report identified two studies (Cortes et al. 2023; Kusters et al. 2022) which did not report any positive association between air pollution and the studies’ health outcome.

Firstly, Cortes et al. (2023) used a time-stratified case-crossover study design with individual-level mortality data to estimate the short-term association between exposure to PM₁₀ and O₃, and cardiovascular and respiratory mortality in Rio de Janeiro, Brazil, between 2012 and 2017. They observed no consistent associations between pollutant and mortality outcomes. However, this study has several limitations that should be considered. Firstly, the study relies on an ecological design, which can limit causal inferences at the individual level. Secondly, the analysis only considers two pollutants and does not account for potential confounding factors such as socio-economic status, smoking prevalence, or access to healthcare. These factors could therefore explain the absence of observed associations.

Secondly, Kusters et al. (2022) examined the association between exposure to air pollution during pregnancy and childhood and cognitive function, as well as emotional and behavioural problems, in adolescents. They analysed data from a birth cohort in Rotterdam (n=5,170). Air pollutant concentrations were estimated using land use regression models. Various cognitive domains and emotional and behavioural problems were assessed through self- and parent-reported measures. There was no observed association between air pollution exposure during pregnancy and childhood and lower cognitive function, or an increased prevalence of emotional and behavioural problems in adolescents. However, the authors recognise that the absence of any observed associations was likely due to factors such as residual confounding, selection bias, or chance, and that further research is needed to understand if a causal relationship may exist.

These studies do not negate the level of evidence for the negative health impacts of air pollution; however, they do highlight the need for further research using more robust methodologies and comprehensive confounding adjustments to better understand the complex relationship between air pollution and health outcomes and importantly the best way of evaluating the associations and potential causality.