next
previous
items

Briefing

Harm to human health from air pollution in Europe: burden of disease 2023

Briefing Published 24 Nov 2023 Last modified 24 Nov 2023
17 min read
Photo: © Fani Papageorgiou, ZeroWaste PIX/ EEA
Air pollution is currently the most important environmental health risk factor in Europe. It remains an important cause of poor health and contributes in particular to respiratory and cardiovascular diseases. This briefing presents information for 2021 of the estimated harm to human health caused by three key air pollutants: fine particulate matter, nitrogen dioxide and ozone. This year’s assessment also presents an estimation of the health impacts associated with specific diseases to which air pollution contributes. Such impacts are expressed using burden of disease metrics, namely ‘morbidity’ (the state of having a disease or disability) and ‘mortality’ (deaths that have occurred due to a specific disease or a group of diseases).

Key messages

  • Air pollutant concentrations in 2021 remained well above the levels recommended by the World Health Organization (WHO) in its air quality guidelines. Reducing air pollution to these guideline levels would prevent a significant number of attributable deaths in EU Member States (EU-27); 253,000 from exposure to fine particulate matter (PM2.5); and 52,000 from exposure to nitrogen dioxide (NO2). Furthermore, reducing the short-term exposure to ozone (O3) would have avoided 22,000 attributable deaths.
  • Between 2005 and 2021, the number of deaths in the EU attributable to PM2.5 fell by 41%.
  • For specific diseases, the greatest harm to human health (burden of disease) is from ischemic heart disease for PM2.5 and diabetes mellitus for NO2.
  • For each specific air pollution-related disease, the relative contribution to poor health (the burden of disease) from mortality and morbidity can vary significantly. For instance, mortality is by far the dominant contributor for ischemic heart disease and lung cancer, while for asthma it is morbidity. This highlights the importance of considering morbidity to avoid underestimating the harm to human health.

Harm to human health from air pollution in Europe, burden of disease 2023 is part of the Air quality in Europe 2023 package which includes the briefing on Europe’s air quality status 2023 and the briefing on Air pollution in Europe.

Measuring impacts on health: the burden of disease

‘Burden of disease’ describes the impact of a health outcome (e.g. a disease) measured by different indicators, i.e. mortality and morbidity. It is often quantified in terms of Disability-adjusted life years (DALYs). DALYs combine the impacts of both mortality and morbidity in one indicator.

Mortality refers to the number of deaths that have occurred because of a specific disease or group of diseases. Mortality is expressed either as attributable deaths or years of life lost:

  • Attributable deaths are deaths statistically attributable to exposure to a risk factor, e.g. air pollution. The attribution is based on the evidence from studies for the causal link between a risk factor and a health outcome leading to death. Attributable deaths are considered preventable if their cause can be eliminated.
  • Years of life lost (YLL) is defined as the years of potential life lost due to death. YLL is an estimate of the average number of additional years that people in a population would have statistically lived if they had not died before reaching a certain statistical life expectancy. The YLL measure considers the age at which deaths occur; therefore, the contribution to the total number of lost life years is higher for a death occurring at a younger age and lower for a death occurring at an older age. Mortality data at the national level are the baseline indicators to estimate attributable deaths and YLL. The YLL per 100,000 inhabitants is also used in this briefing as an indicator that is comparable across countries.

Morbidity is the state of having a disease or disability, measured by, for instance, the prevalence of a disease in a population. In this briefing, morbidity is expressed as years lived with disability (YLD), meaning years lived in reduced health due to a particular health outcome. YLD is estimated by multiplying the number of prevalent cases of a particular health outcome with a disability weight, a factor indicating the severity of the health outcome on a scale from 0 (full health) to 1 (most severe health state).

Disability-Adjusted Life Year (DALY) is one lost year of a healthy life due to disease, injury or risk factor. DALYs are obtained by adding YLL and YLD for the same disease or group of diseases. The burden of disease is the sum of these DALY values in a population. Therefore, DALYs standardise health effects by expressing in one number the number of people affected and the duration and severity of the health effects.

Figure 1. Burden of disease as sum of years lived with disability and years of life lost

How air pollution harmed health in Europe in 2021

Total mortality

This section presents the mortality related to all natural causes (i.e. excluding accidental and other non-natural causes) attributable to key air pollutants.

In 2021 in the EU-27:

  • 253,000 deaths were attributable to exposure to PM2.5 concentrations above WHO’s guideline level of 5 µg/m3 (micrograms per cubic metre of air).
  • 52,000 deaths were attributable to exposure to NO2 concentrations above WHO’s guideline level of 10 µg/m3.
  • 22,000 deaths were attributable to short-term exposure to Oconcentrations above 70 µg/m3.

In addition to the EU-27 member countries, a larger set of EEA countries were also assessed: 40 for PM2.5 and 41 for NO2 and O3. The results are presented in Table 2 and Table 3, and summarised below for all countries:

  • 293,000 deaths were attributable to exposure to PM2.5 concentrations above WHO’s guideline level of 5 µg/m3.
  • 69,000 deaths were attributable to exposure to NO2 concentrations above WHO’s guideline level of 10 µg/m3.
  • 27,000 deaths were attributable to short-term exposure to O3 concentrations above 70 µg/m3.

Both groups of countries experienced a slight increase in the mortality attributable to PM2.5 and NO2 and a decrease of the mortality attributable to O3 in 2021 compared to 2020.

These differences are not significant since they are within the uncertainty intervals of the calculations. However, the slight increase could be explained by slight differences in the total exposure (higher for PM2.5 and NO2 and lower for O3) and by a slight increase in total European mortality, mainly due to the impact of the COVID-19 pandemic.

SENSITIVITY ANALYSIS OF ATTRIBUTABLE DEATHS DUE TO AIR POLLUTION

It is important to note that the attributable deaths presented above do not capture possible additional deaths caused by exposure to PM2.5 and NO2 concentrations below the WHO’s 2021 guideline levels. The scientific evidence is less certain for exposures below WHO’s guideline levels than for health impacts above them.

However, there is no evidence of a threshold below which air pollution does not impact health. With this in mind, EEA has also performed a sensitivity analysis of attributable deaths for exposure to PM2.5 and NO2 at concentrations above 0 µg/m3. Deaths attributable to O3 in the sensitivity analysis were calculated for exposure to concentrations above 20 µg/m3 as previously recommended by the WHO. The level of uncertainty around these estimates is higher than for the estimates presented above.

Table 1 presents the potential additional attributable deaths and potential total attributable deaths by pollutant for the EU-27 and for the larger number of European countries included in the assessment.

The EEA table Air Quality Health Risk Assessments presents estimates from 2005 to 2021 of the deaths attributable to exposure to PM2.5, NO2 and O3  for countries and the different groups of countries. 

Table 1. Potential additional attributable deaths and potential total attributable deaths in 2021 when lower concentrations (below WHO guideline levels) of key pollutants are considered

Static Content listing
Geographical scope Pollutant Potential additional attributable deaths Potential total attributable deaths 
EU-27 PM2.5                             179,000                               432,000
NO2                               90,000                               142,000
O3                               86,000                               108,000
40 European countries PM2.5                             194,000                               487,000
41 European countries NO2                             106,000                               175,000
O3                             100,000                               127,000

Note: Türkiye is not included in the PM2.5 estimations since the number of background monitoring stations from which data is available was too low to produce concentration maps for fine particles

Table 2 presents the total country population, the population-weighted mean concentrations (as an indication of exposure) and the estimated number of attributable deaths by country in 2021.

Table 3 presents the YLL and YLL per 100,000 inhabitants associated with exposure to PM2.5, NO2 and O3 concentration levels in 2021.

The sections below provide information at country level on the mortality due to all natural causes that are attributable to key pollutants.

 

  • PM2.5
  • NO2
  • O3

For PM2.5, the highest absolute numbers of attributable deaths in 2021 occurred in Poland, Italy and Germany (in order of decreasing rank). When considering YLL per 100,000 inhabitants, the highest relative numbers were observed in south-eastern European countries. The highest concentrations of PM2.5 were also observed in the same region, namely, in North Macedonia, Bosnia and Herzegovina, Serbia, Montenegro and Kosovo under UNSCR 1244/99 (in order of decreasing rank).

The lowest relative health impacts due to PM2.5 exposure occurred in countries situated in the north and north-west of Europe, including Iceland, Finland, Sweden, Norway and Estonia.

Map 1. Mortality due to exposure to PM2.5, 2021

Note: The map presents, for each country, attributable deaths, YLL and YLL/100,000 due to exposure to PM2.5 above the 2021 WHO AQ guideline level of 5 µg/m3 in 2021.

The legend and the number included on the map for each country represent the YLL/100,000 inhabitants.

For NO2, the highest absolute numbers of attributable deaths in 2021 occurred in Türkiye, Italy and Germany (in order of decreasing rank). When considering YLL per 100,000 inhabitants, the highest relative numbers were observed in Bulgaria, Romania, Serbia, Cyprus and Türkiye (in order of decreasing rank).

The lowest relative health impacts due to NO2 exposure occurred in Iceland, Sweden, Finland, Estonia and Denmark.

Map 2. Mortality due to exposure to NO2, 2021

Note: The map presents, for each country, attributable deaths, YLL and YLL/100,000 due to exposure to NO2 above the 2021 WHO AQ guideline level of 10 µg/m3 in 2021.

The legend and the number included on the map for each country represent the YLL/100,000 inhabitants.

The countries with the highest absolute numbers of attributable deaths in 2021 due to short-term exposure to O3 were Italy, Germany and Türkiye, in order of decreasing rank. The highest rates of YLL per 100,000 inhabitants occurred in Albania, Montenegro, Bosnia and Herzegovina, Greece and Kosovo (in order of decreasing rank).

The countries with the lowest relative impacts were Norway, Iceland, Ireland, Finland and Sweden.

Map 3. Mortality due to short-term exposure to O3, 2021

Note: The map presents, for each country, attributable deaths, YLL and YLL/100,000 due to short-term exposure to O3 above 70 µg/m3 (SOMO35) in 2021.

The legend and the number included on the map for each country represent the YLL/100,000 inhabitants.

 

Estimating the harm of air pollution to human health

Since 2014, the EEA has been estimating mortality by all natural causes (i.e. not for each specific disease individually) due to exposure to air pollution, with support from our European Topic Centres. Mortality is the most serious health outcome of exposure to air pollution, with robust and extensive scientific evidence of causality.

Exposure to air pollution causes or aggravates a range of specific diseases, and this is why this year’s report includes an individual estimation of the burden of disease for these diseases.

This new analysis allows better assessment of the link between mortality and morbidity. As a result, the share of mortality and morbidity in the total burden of disease for selected diseases can be combined and compared.

Concentration-response functions are used to attribute a health risk to exposure to air pollution. These functions are based on epidemiological studies and estimate the increase in risk per unit of concentration of a certain air pollutant. As an example, the concentration-response function for mortality by all natural causes from exposure to PM2.5 used in this briefing assumes a linear increase in the relative risk of mortality of 8% per 10 µg/m3 increase in the annual mean concentration of PM2.5 as recommended by WHO (2021).

In this briefing the EEA applies  the concentration-response functions for mortality due to all natural causes that can be attributable to air pollution as set out in WHO’s 2021 air quality guidelines

The main analysis presented in this briefing for mortality due to all natural causes calculates mortality only for concentrations above 5 μg/m3 for PM2.5 and above 10 μg/m3 for NO2. These values correspond to the 2021 air quality guideline levels, below which the level of uncertainty surrounding the concentration-response functions is higher.

Applying this approach provides an estimate of deaths attributable to levels of PM2.5 and NO2 above the recommended guideline levels. In other words, these are the attributable deaths that could have been avoided if in 2021 the WHO recommended levels had been reached. Regarding O3, the concentration from which premature deaths are calculated is 70 µg/m3, following WHO’s 2013 recommendations.

Note that the estimates of attributable deaths presented here may differ from others calculated at national and European levels due to differences in methodological approaches. The kinds of methodological differences that can affect estimates of attributable deaths include: 

  • the use of different concentration maps, where concentrations are estimated at different scales;
  • the use of more detailed demographic data. For instance, the EEA assumes a unique national value of all-cause mortality, while countries might have specific regional or local data;
  • the use of different relative risks or different concentration-response functions;
  • the use of different assumptions in the range of concentrations considered. For example, the main analysis carried out by the EEA for mortality by all natural causes attributable to PM2.5 considers concentrations above WHO’s 2021 guideline level (5 µg/m3), while a sensitivity analysis has considered all the range of concentrations, that is, above 0 µg/m3.

The EEA uses concentration maps with the same resolution and applies the same assumptions for all of Europe, enabling a robust comparison across European countries.

For morbidity, years lived with disability (YLD) were estimated for nine pollutant-disease pairs. For these pairs, scientific evidence of an association between exposure to a specific air pollutant and the disease is robust and enough data for the estimations were available. For PM2.5, morbidity was estimated for:

  • asthma in children (<15 years);
  • chronic obstructive pulmonary disease (in adults ≥25 years);
  • ischemic heart disease (in adults ≥25 years);
  • lung cancer (in adults ≥25 years);
  • diabetes mellitus (in adults ≥25 years);
  • stroke (in adults ≥25 years).

 

For NO2, morbidity was estimated for:

  • asthma (in adults ≥15 years);
  • diabetes mellitus (in adults ≥25 years);
  • stroke (in adults ≥25 years).

For O3, estimates were calculated for the number of hospital admissions (in adults ≥65 years) due to respiratory diseases. No study was identified linking the incidence or prevalence of diseases to exposure to O3, and the number of hospital admissions due to respiratory diseases was used as a proxy for an approximation of the morbidity burden resulting from exposure to O3.

Additionally, mortality was also estimated for the same six pollutant-disease pairs in the case of PM2.5 and for the same three in the case of NO2, both as attributable deaths and years of life lost (YLL).

In both cases (morbidity and mortality for specific causes), the impacts (burden of disease) were estimated for the same range of concentrations as for all causes of natural mortality: above 5 µg/m3 for PM2.5, above 10 µg/m3 for NO2, and above 70 µg/m3 for O3.

Finally, adding the respective YLD and YLL, we obtained the total burden of disease per pollutant-disease pair.

New terminology in this briefing

In previous reports and briefings, the EEA has used the terms ‘health risk assessment’ and ‘premature deaths’, which, without being wrong, can be replaced by more accurate ones.

In this briefing we use ‘environmental burden of disease’ instead of ‘health risk assessment’. The term ‘environmental burden of disease’ relates to the methodology to estimate the share of the disease burden attributed to the exposure to air pollution (the environmental risk factor we are considering). The term ‘health risk assessment of air pollution’ is a wider term that also includes, for instance, an estimate of the risks of past, current or future exposure to air pollution and of changes in exposure that may result from planned policies or other modifications of air quality (WHO, 2016).

The term ‘premature deaths’ is defined as deaths that occur before a person reaches an expected age. This expected age is typically the life expectancy for a country, stratified by sex and age. But every death is premature, in the sense that for all ages there is always a remaining life expectancy. Therefore, when using the environmental burden of disease approach, it is more correct to use the term ‘attributable death’, since it represents a death statistically attributable to the exposure to a specific risk factor (in this case, again, air pollution). Nevertheless, we keep the term ‘premature deaths’ in connection with the political goal set in the Zero Pollution Action Plan. Finally, we still use the remaining life expectancy to estimate the years of life lost due to the attributable deaths.

 

Specific causes of mortality and morbidity

To estimate the total harm of air pollution on health, we use the burden of disease concept, which combines in the indicator DALYs, the impacts of both mortality (deaths due to a specific disease, expressed as years of life lost or YLL) and morbidity (state of having a disease or disability, expressed as years lived with disability or YLD). This section presents the individual estimates of the burden of disease for several specific diseases for which the scientific evidence of an association between the disease and exposure to an air pollutant is robust. 

PM2.5

For the six specific diseases considered for PM2.5 (Figure 2), the highest impact corresponds to ischemic heart disease (759,303 DALYs), followed by stroke, diabetes mellitus, chronic obstructive pulmonary disease, lung cancer and asthma.

In the cases of ischemic heart disease and lung cancer, mortality drives the total burden of disease: it accounts for 99% and 98%, respectively, of the impact, making the relative contribution of morbidity to the total burden of disease almost insignificant for those cases.

For the other four diseases, the contribution of morbidity is much more relevant, representing 20% for stroke, 24% for diabetes mellitus, 50% for chronic obstructive pulmonary disease and 99% for asthma.

This emphasises the importance of considering morbidity impacts to avoid underestimating the overall health impact of air pollution.

Figure 2. Burden of disease for PM2.5 and NO2, 2021

Note: For PM2.5 the data relates to six specific identified diseases and covers 40 European countries, while for NO2 the data relates to three specific identified diseases and covers 41 European countries. 

Click here for different chart formats and data

Focusing only on the number of deaths attributable to each disease, the six diseases considered for PM2.5 resulted in a total of 232,000 attributable deaths, below the 293,000 estimated when using all-cause natural mortality in 40 European countries (Figure 3).

Using all-cause mortality might lead to an overestimation of the total mortality, as it includes some deaths which have a probability of not being associated with air pollution (infectious diseases, for instance). On the other hand, using specific causes of mortality might lead to an underestimation of the total mortality, as it only considers health outcomes with strong evidence for its association with air pollution, and could exclude outcomes that are presently unknown.

Other factors can also influence in the different results, although to a minor extent: the burden of disease for specific causes was estimated at country level, while for all natural causes it was estimated at grid level (using the concentration maps and population density maps grids) and then aggregated. Furthermore, for all-cause mortality, and according to the concentration-response functions, adults above 30 years are considered, while for the specific causes, different age groups are considered, as specified in the section 'Estimating the harm of air pollution to human health'.

Figure 3. Mortality due to exposure to PM2.5 and NO2, 2021

Note: Mortality due to exposure, expressed as attributable deaths, to PM2.5 and NO2. The inner circle shows mortality for all natural causes. The outer circle shows mortality for six (PM2.5) and three (NO2) specific causes and 40 and 41 European countries, respectively, 2021. 

Click here for different chart formats and data

NO2

For NO2 (Figure 2), the highest impact on health is due to diabetes mellitus, with 314,574 DALYs, out of which 57% correspond to mortality and 43% to morbidity. This is followed by stroke (204,723 DALYs), where mortality represents 75% and morbidity 25%. And, finally, for asthma (115,425 DALYs), mortality represents 7% and morbidity 93%.

When considering only the mortality part of the burden of disease of the three diseases included, we estimated 37,000 deaths, well below the 69,000 deaths attributed to all-natural causes (Figure 3). Here the underestimation seems more clear, probably due to the fact that only three specific diseases were considered.

Finally, the estimated total burden of disease for NO2 (634,721 DALYs, considering three diseases) is almost 4 times lower than for PM2.5, (2,528,363 DALYs, considering six diseases).

O3

In the case of O3, the burden of disease is estimated using the indicator, ‘hospital admissions for respiratory diseases’ since no study was identified linking the incidence or prevalence of diseases to exposure to O3. Around 16,000 admissions of adults aged 65 years or above were estimated in 2021 in the 41 European countries considered.

Information at country level of the burden of disease can be found in the Eionet ETC HE Report 2023/7.

 

Policy context

For several decades, the European Union (EU) has had air quality standards in place for key air pollutants in the ambient air quality directives. These values, based on WHO’s recommendations at the time, also reflected the technical and economic feasibility of their attainment across EU Member States and are in general less demanding than the WHO’s 2005 air quality guidelines.

In 2021, WHO updated its air quality guidelines for the first time since 2005, lowering the recommended levels for PM, NO2 and O3. This update is based on systematic reviews of the latest scientific evidence outlining how air pollution affects human health. On 26 October 2022 the European Commission adopted a proposal for a revised Ambient Air Quality Directive, aiming, among others, to align air quality standards more closely with WHO’s updated recommendations.

Under the European Green Deal, the Zero Pollution Action Plan sets a 2030 target of improving air quality, with a focus on PM2.5, in order to reduce the number of attributable premature deaths caused by air pollution in the EU by a minimum of 55%, relative to those in 2005. EEA’s indicator Premature deaths due to exposure to fine particulate matter in Europe shows that premature deaths in the EU attributable to PM2.5 exposure above the WHO guideline level of 5µg/m3 fell by 41% between 2005 and 2021. The Zero Pollution Action Plan target is likely to be surpassed if EU policies on air, climate and energy are adequately implemented.

The analysis in this briefing refers to 41 countries, which are the 27 Member States of the European Union (EU-27) and additional European countries. Türkiye is not included in the PM2.5 estimations since the number of background monitoring stations from which data is available was too low to produce concentration maps for fine particles. This leaves 40 countries for which PM2.5 estimations were made.

The burden of disease estimations are made individually for the respective air pollutants. They cannot be added together as they exhibit some degree of correlation. This is the case especially for the burden of disease due to fine particles and NO2.

Identifiers

Briefing no. 23/2023
Title: Harm to human health from air pollution in Europe: burden of disease 2023
EN HTML: TH-AM-23-026-EN-Q - ISBN: 978-92-9480-614-7 - ISSN: 2467-3196 - doi: 10.2800/721439

Disclaimer

The country assessments are the sole responsibility of the EEA member and cooperating countries supported by the EEA through guidance, translation and editing.

Permalinks

Geographic coverage

Temporal coverage

Topics

Document Actions