Introduction

The global burden of disease has shifted from communicable, maternal, perinatal and nutritional causes to non-communicable diseases (NCDs) such as those with atherosclerotic or metabolic sequelae, including ischaemic heart disease, arterial hypertension and diabetes mellitus (according to data from the WHO–Global Health Observatory1 and the Global Burden of Disease (GBD) study2). In 2010, the leading risk factors and diseases responsible for global deaths were tobacco smoking, high blood pressure, ischaemic heart disease and cerebrovascular disease, accounting for 55% of global mortality3,4,5. In 2019, the leading risk factors were high systolic blood pressure (causing 10.8 million global deaths; 19.2% of all deaths) and tobacco smoking (causing 8.7 million global deaths; 15.4% of all deaths)6. An overview of the leading causes of disease and death in 2019 is provided in Fig. 1, which shows disability-adjusted life years (DALYs) and global deaths, and highlights the proportion caused by cardiovascular diseases. According to the 2016 WHO status report, cardiovascular disease was the leading disease category contributing to NCDs, with a percentage of >44%, followed by cancer (22%), respiratory diseases (9%), diabetes (4%) and others (21%)7. The leading risk factors for death are similar for men and women6 (Fig. 1). However, the absolute numbers of deaths are higher for men, and tobacco and alcohol have a greater contribution to mortality and DALYs in men than in women. Three environmental risk factors — air pollution, unsafe water and sanitation, and especially child and/or maternal malnutrition — contribute disproportionately more to the burden of disease (DALYs) in the general population than other risk factors. Tobacco and alcohol use and occupational risks, such as injuries and exposure to particulates, carcinogens and noise at the workplace, further increase the environmental contribution to DALYs in men6 (Fig. 1). Accordingly, the environmental risks might contribute much more to the global burden of disease than previously estimated.

Fig. 1: Global number of deaths and DALYs in 2019 attributable to health risk factors according to cause and sex.
figure 1

The bar graphs show the global number of deaths and disability-adjusted life years (DALYs) in women and men in 2019 attributable to health risk factors. The pie charts show the attributable burden of disease for the leading risk factors. Cardiovascular disease (CVD) is the major category of sequelae for all leading risk factors, followed by respiratory disease, diabetes mellitus, cancer and infections. COPD, chronic obstructive pulmonary disease. Based on data from ref. 6.

Full size image

In 2018, 5,687 published genome-wide association studies provided insights into the influence of genetics on human health and disease8. For many complex human diseases, such as cancer, cardiovascular diseases, respiratory diseases and type 2 diabetes, genetic variation explains only a modest proportion of the disease risk. For example, for 28 chronic diseases, only a mean of 18.5% of the population-attributable fraction was related to genetic predisposition9. By contrast, a large part of the disease burden is probably attributable to environmental stressors and the interplay between the genes of an individual and the environmental burden6. Accordingly, the statement “genetics load the gun but environment pulls the trigger”10 (which is based on the concept proposed by Olden and Wilson11) perfectly characterizes the effect of the environment on health compared with the effect of genetic predisposition (Box 1). This postulate is supported by statistics on cardiovascular-associated and cancer-associated deaths, among which only 0.25 (16.4%) of 1.53 million deaths in 2000 in Western Europe could be attributed to genetic causes9. In addition, the substantial differences in disease incidence and prevalence between monozygotic twins (who share an identical genome) suggest that other factors, including the environment, determine the onset of chronic disease. Nevertheless, pollution-related diseases have not earned recognition in the Global Action Plan for the Prevention and Control of NCDs12 nor in the United Nations (UN) sustainable development goals (SDGs)13. Therefore, leading scientists in environmental health are calling for ‘environment-wide association studies’14,15 in analogy to genome-wide association studies16.

In this Review, we provide a detailed description of the exposome concept. We discuss the most important environmental physical and chemical exposures and their consequences for the burden of cardiovascular disease and summarize exposure assessment methods and the most important biomarkers for exposome research. We discuss the most relevant exposome studies related to cardiovascular diseases and highlight the relevance of the exposome concept for future health research and preventive medicine. Finally, we highlight the challenges and limitations of exposome studies.

The need for an exposome approach

Chemical pollution and global burden of disease

The Lancet Commission on pollution and health concluded that chemical pollution is currently the most important environmental cause of disease and premature death in the world12. Diseases caused by chemical pollution were responsible for an estimated 9 million premature deaths in 2015 (16% of all deaths worldwide), three times more deaths than those caused by AIDS, tuberculosis and malaria combined12. A main pollutant is ambient air pollution, which reduces the global average life expectancy by 2.9 years, a reduction in life expectancy that is strikingly more extensive than the reduction caused by the traditional cardiovascular risk factor of tobacco smoking (2.2 years)17. The WHO estimates that up to 12.6 million global deaths in 2012 were due to unhealthy environments18,19. These numbers will most probably increase further in future calculations given that a 2020 estimation indicates that 9 million premature global deaths per year are due to air pollution in the form of particulate matter with a diameter of ≤2.5 µm (PM2.5) alone20.

PM2.5 has high toxicity because these fine particulates can transmigrate directly from the lung epithelium into the bloodstream, causing endothelial dysfunction, inflammation and oxidative stress in the systemic vasculature21. PM2.5 can also enter the brain directly via the olfactory nerve, causing autonomic imbalance or activation of the hypothalamic–pituitary–adrenal axis21. The smaller the particle, the stronger its penetration capacity. These severe systemic adverse effects might explain why the excess mortality associated with PM2.5 exposure is around ten times higher than the excess mortality related to NO2 exposure and 20 times higher than that associated with exposure to ozone22. The role of PM2.5 as a major health hazard is also supported by the gradually increasing risk of all-cause and cardiovascular death associated with exposure to these air pollutants, with ozone being the lowest, followed by NO2, and PM2.5 being the highest23.

The role of air pollution as a leading health risk factor is supported by the dramatic drop in the rates of cardiovascular disease and even the extended life expectancy observed during the period of pollution-control measures before the 2008 Beijing Olympic Games24 or with the permanent lowering of diesel emissions by new restrictive laws in Tokyo25 and the USA26,27. Ambient air pollution and rising temperatures due to climate change are already ranking in the top 20 of the risk factors for DALYs and death described in the GBD study6 (Fig. 1). The GBD study estimated that diseases caused by all forms of pollution amount to 268 million DALYs28. However, the GBD study focuses on only a limited number of exposures for which worldwide data are available. Environmental risk in urban areas is also influenced by the excessive use of vehicles29 and the reduction of green spaces in favour of road infrastructure. Both of these factors prevent the development of heart-healthy cities10,30.

According to the data in Table 1, the primary chemical pollutants contributing to global excess deaths are air, soil, water and occupational pollution31. The hazardous effects of air pollution are mediated mainly by ambient PM2.5, those of water pollution by unsafe sources and the hazards at the workplace primarily by particulates. Health risks and deaths related to soil pollution are mainly due to heavy metals, deforestation, over-fertilization and pesticides, although nanoplastics and microplastics also make a substantial contribution32. Although lead toxicity is mainly caused by water and soil pollution, we present it separately in Table 1 because lead is a major environmental heavy metal that enters the water and food circulation as a consequence of its use in car batteries, water pipes and paints31, and is readily taken up by plants33 and, potentially, by farm animals and fish and other sources of seafood. A close relationship exists between water and soil pollution, given that polluted soils will contaminate surface and groundwater32. Heavy metals and metalloids are of specific concern for their contribution to cardiovascular sequelae because they can trigger oxidative stress, inflammation and sulfur-related toxicity by forming complexes with protein thiols34,35.

Table 1 Contribution of major pollutants to global deaths and cardiovascular disease
Full size table

Regarding the absolute number of excess deaths associated with pollutants, an apparent discrepancy exists between the more conservative calculations made using a set of extensive cohort studies or meta-analyses (usually from the GBD study, WHO, Global Health Observatory and European Environmental Agency) and the advanced calculations made using mostly up-to-date exposure–response functions and health data. The outcome of ‘progressive’ studies (those using the most current exposure–response functions or most pessimistic models) can easily double the conservative values, providing a more pessimistic picture of the environmental adverse effects on disease burden and mortality. Finally, almost all chemical pollutants are associated with adverse cardiovascular health effects such as hypertension and ischaemic heart disease (Table 1), stressing the importance of cardiovascular sequelae as the driver of the detrimental effects of the exposome.

Non-chemical pollution and global burden of disease

Transportation noise pollution

Not all established environmental risk factors are included in the GBD study, UN SDGs or Global Action Plan for the Prevention and Control of Non-Communicable Diseases. A striking example of an overlooked risk factor in this context is transportation noise, which has consistently been associated with a higher risk of cardiovascular diseases36,37,38. Traffic noise is a ubiquitous exposure (Fig. 2), with >20% of the population in the EU exposed to levels exceeding the EU guideline value of 55 dB (ref. 39). The European Environmental Agency estimated that traffic noise causes 12,000 premature deaths and 48,000 cases of ischaemic heart disease per year in Europe39, particularly in cities with a high noise burden40. These numbers are likely to be an underestimation given that the estimated number of people exposed to traffic noise has been based on areas with large agglomerations of people and that new studies have indicated that the effect of road traffic noise on the risk of cardiovascular disease starts at values below 55 dB (refs. 37,41,42). A notable reason for traffic noise not being included in the GBD study is that worldwide traffic noise exposure data are unavailable.

Fig. 2: Colocalization of environmental exposures in Denmark.
figure 2

The maps show the annual mean levels of particulate matter with diameter ≤2.5 µm (PM2.5) (part a) and ultrafine particles (UFP) (part b) in the air in Denmark in 2000, modelled by the Aarhus University using the Danish Eulerian Hemispheric Model and the Urban Background Model at 1 × 1 km spatial resolution186,187. Corine Land Cover map for Denmark in 2000, providing information for 44 classes of land use such as forest and arable land (part c). The European Environment Agency coordinates European Union Corine Land Cover maps. Map of road traffic noise at the most exposed facade at all residential addresses in Denmark in 2015, shown as the day–evening–night noise level (Lden), an indicator based on 24-h average sound levels (part d). Risk estimates for type 2 diabetes mellitus per interquartile range increase in exposure to air pollution, noise pollution and lack of green space in single-exposure and multi-exposure analyses based on a prospective study covering all of Denmark57 (part e). Parts a and b courtesy of Matthias Ketzel (unpublished data). Part c reprinted from ref. 188, CC BY 2.5. Part d courtesy of M.S. (unpublished data). Part e reprinted with permission from ref. 57, Elsevier.

Full size image

Light pollution

Light pollution is a novel, ubiquitous environmental risk factor, defined by the changes in natural night-time sky brightness induced by anthropogenic sources of light43,44, which are most evident in big cities and large metropolitan areas. According to a 2016 report, 83% of the world’s population and >99% of the US and European populations live under light-polluted skies43. A review published in 2009 provides an excellent overview of the substantial effect of light pollution on fauna and flora as well as on human NCDs45, which mainly result from poor sleep and altered circadian rhythms.

Lack of green spaces

The lack of green infrastructure in cities is considered a major environmental stressor. A study published in 2021 established that meeting the WHO recommendation of access to green space could prevent 42,968 deaths annually, representing 2.3% of the total deaths from natural causes and 245 years of life lost per 100,000 inhabitants per year46. The mechanisms for the benefits of access to green space include reducing harm (such as reducing exposure to air pollution, noise and heat), restoring capacities (such as restoration of attention and recovery from physiological stress) and building capacities (for example, encouraging physical activity and facilitating social cohesion)47.

Climate change

Climate change is a major environmental risk factor of concern and is associated with an increased relative risk of cardiovascular disease-related death48. Cardiovascular diseases induced by climate change are mediated by air pollution, increased ambient temperatures, vector-borne disease and mental health disorders49. Currently, the frequency of heat waves, which are responsible for more deaths than any other type of extreme weather in many parts of the world, is increasing in conjunction with mean temperatures50. From 2008 to 2017, extreme heat was associated with higher all-cause mortality than that seen in years with normal average temperature in the contiguous USA (which excludes Alaska and Hawaii), with a greater increase noted especially among older adults51. In patients with cardiovascular disease, particularly those with heart failure who take diuretics and β-blockers, the heat might result in severe volume depletion and, potentially, cardiogenic shock52. Accordingly, a worldwide study including data from 567 cities in 27 countries across 5 continents demonstrated that exposure to extreme temperatures (quantified as the 99th percentile for heat and the 1st percentile for cold) was associated with an increased risk of death from any cardiovascular cause as well as increased risk of death from ischaemic heart disease, stroke or heart failure compared with normal average temperatures52. In particular, among all cardiovascular-related causes of death, heart failure was associated with the highest rate of excess deaths related to extreme heat or cold temperatures52.

Multi-exposure to environmental risk factors

The number of environmental risk factors for cardiovascular disease is increasing, and these risk factors are rarely present in isolation, especially in densely populated and highly urbanized areas10. Although substantial medical and societal progress has been made during the past three decades to combat and prevent traditional cardiovascular risks factors, such as smoking, diabetes, hypercholesterolaemia, hypertension and obesity, on an individual basis, environmental stressors also need to be considered as crucial cardiovascular risk factors to reduce the global burden of cardiovascular disease. The potentially additive adverse health effects and disease burden of multi-exposure to environmental stressors might substantially add to previous estimations of the health effects of environmental risk factors that were assessed individually. For example, exposure maps of light, air and noise pollution and of the increase in concrete spaces and lack of green spaces in Western countries generally show a substantial colocalization of environmental risk factors, especially in highly urbanized and industrialized areas10. A similar co-exposure is seen for light and air pollution, with potential additive risk profiles for cardiovascular disease53. Preclinical studies on combined exposure to environmental pollutants and stressors are rare. However, one study in mice suggested additive cardiovascular damage after co-exposure to particulate matter and aircraft noise (which induce their primary adverse effects via lung inflammation and the brain stress response, respectively) mediated by activating partially synergistic pathological mechanisms54. In another study, co-exposure to particulate matter and light at night caused additive adverse effects on circadian clock components and cardiometabolic parameters55.

Epidemiological studies investigating adverse health effects of colocalized environmental risk factors from a multi-exposure perspective are becoming more frequent56. A comprehensive, large-scale example is a prospective cohort study from 2022 covering all of Denmark, which investigated the health effects of four air pollutants (PM2.5, ultrafine particles, elemental carbon and NO2), road traffic noise at both the most and least exposed residential facades, and lack of green space within 150 m and 1,000 m from the residence57 (Fig. 2). All exposures were estimated using state-of-the-art exposure models with high spatial (address-level) and temporal resolution. The study showed that combustion-related air pollution (ultrafine particles and NO2), road traffic noise and lack of green space were all independently associated with a higher risk of type 2 diabetes in a multi-exposure perspective (Fig. 2e). The estimated cumulative risk for all exposures was higher than for any single exposure. Similar results were found for the Danish population when investigating the effects of the three types of exposures in relation to cardiovascular disease (M.S., unpublished work). These results stress the importance of considering multiple environmental risk factors when planning cardiovascular disease prevention strategies.

The exposome concept

The term ‘exposome’ was introduced in 2005 by Wild58 and describes the totality of lifelong changes in human physiology and pathophysiology induced by the environment (that is, by exposure to environmental stressors16,59,60) (Fig. 3). These exposures comprise not only chemical and physical factors, such as air pollution, noise, ultraviolet (UV) radiation and climate change, but also socioeconomic and mental health determinants, including the social environment and capital, exposure to viral and bacterial pathogens, and psychosocial stress. These factors determine the general external environment and cannot be easily modified by the individual. The second component that defines the exposome comprises lifestyle or behavioural factors, including smoking habits, alcohol consumption, unhealthy diet, physical inactivity and use of consumer products such as cosmetics. These factors define the specific external environment, which the individual can modify more easily than, for example, reducing air pollution and noise exposure, controlling outside temperature, changing social capital or improving the climate. Notably, the exposome approach moves away from the one-exposure to a one-health-outcome type of analysis in favour of an analysis that considers that many exposures coincide and are associated with multiple health outcomes.

Fig. 3: The exposome concept.
figure 3

The exposome describes the totality of lifelong exposures to environmental risk factors, the induced pathological and mechanistic changes in central biochemical pathways, and the associated health effects58,59. Within the general external environment, the environmental factor with the largest effect on health is chemical pollution. Within the specific external environment, the environmental factors with the most substantial effect on health are tobacco smoking and unhealthy dietary habits. The general and specific external environment can have adverse effects on mental health (for example, by social isolation or work strain), but the contribution of environmentally triggered mental stress to the global burden of disease and global deaths might be underestimated owing to the lack of valid exposure–response relationship data for mental risk factors. Changes in the internal environment induced by exposures are most frequently quantified by changes in the transcriptome, epigenome, proteome and metabolome (and more rarely by analysis of the microbiome or adductome). Alterations in central biochemical pathways identified in exposome studies include dysregulation of circadian clock components that lead to impaired rhythmicity and phase shifts, the release of stress hormones (cortisol and catecholamines), production of reactive oxygen species by mitochondria and NADPH oxidase in activated immune cells, inflammation with tissue infiltration of activated immune cells, and oxidative damage in different organs (reviewed previously10,38). Environmental exposures can act synergistically or additively with genetic predisposition or pre-existing cardiovascular disease. Environmental exposures can exacerbate detrimental cardiovascular health outcomes such as atherosclerosis, vascular stenosis, myocardial infarction, heart failure and stroke.

Full size image

Landrigan et al. coined the word ‘pollutome’ to encompass all forms of pollution that have the potential to harm human health12. The pollutome includes chemical pollution, a crucial environmental health risk, but can also include non-chemical pollutants such as light and noise pollution. The exposome can be divided into multiple pollutomes related to specific periods of life (infancy, childhood, adolescence, adulthood and old age) or geographical locations12. Accordingly, pollutomes represent temporal and spatial exposures and their effect on human physiology and pathophysiology. Unfortunately, we understand only a minor part of the effects of different pollutomes on cardiometabolic health because multi-pollutant studies have not been conducted. Therefore, the current estimate of 9 million premature deaths per year attributable to all forms of chemical pollution might be just the tip of the iceberg12, particularly when considering that the estimated 8 million premature deaths attributable to air pollution alone are based on advanced calculations using novel exposure–response functions20. Most of the pollutome remains poorly understood and comprises emerging, unquantified effects of known pollutants and inadequately characterized health effects of emerging pollutants. This unknown and uncharacterized part of the pollutome is not included in the GBD study nor in other extensive, environmental health association studies.

Several exposome studies assessed the exposome divided into various subcategories (for example, the exposure component or the health-outcome component), which might facilitate the handling of data sets given that not all exposures and health effects will coincide. An example is the assessment of the occupational exposome, with detailed consideration of multiple toxic compounds that individuals are exposed to at work61. This study focused on occupational toxins and the subsequent development of cancer and also provided a detailed overview of how a specific exposome can be assessed61. Likewise, an obesity exposome model was developed to evaluate the environmental effect of climate, sex, ethnicity and socioeconomic factors on BMI, adiposity and cardiometabolic health outcomes62. Other exposome approaches have investigated the effect of environmental risk factors (external environment) and associated biomarkers (internal environment) on the development of diabetes and its cardiovascular sequelae14,63. Measuring the tumour exposome has been suggested as a robust basis for evaluating environmental exposures and the risk of cancer64. The term ‘urbanome’ describes the complexity of megacities with all the exposures and their effects on people living in them65, including socioeconomic factors, green space availability, stressors and pollutants. For example, the urbanome has been used to assess the effect of environmental exposures on pregnancy66.

The exposome can be assessed by two approaches: the top-down approach, which determines changes in the internal environment (biomarkers), and the bottom-up approach, which assesses the external environment67. In the top-down approach, the focus is on footprints of exposure in vivo by directly quantifying exogenous compounds, protein adducts and reactive metabolites as well as changes in biological response profiles in the host assessed by metabolomics, transcriptomics or proteomics. The bottom-up approach is characterized by the collection of comprehensive data on environmental exposures through surveys, sensors or trace analytical chemistry in environmental samples. The top-down approach enables the creation of new hypotheses for exposure–disease and exposure–response relationships. The bottom-up approach can help to generate new hypotheses on effects but does not necessarily investigate the effect. A complete exposome study will combine top-down and bottom-up approaches. A health and disease status of a person is determined by the combined effects of complex environmental exposures, associated biological responses and individual susceptibility over time68. Whereas the genomic composition of an individual is stable, the environmental exposures, how the individual responds to those exposures and how the responses ultimately manifest as health effects can vary over the life course of an individual. The exposome approach involves collecting relevant information across all variables, identifying meaningful biological consequences, and using the information for systems analysis and data-driven discovery.

As explained in detail in the Introduction, tobacco smoking and air, water and soil pollution are the most important environmental risk factors (exposures) for all-cause death and cardiovascular death6 (Fig. 3). Mental health and socioeconomic factors should be considered in future exposome studies given the observed association between these factors and cardiovascular risk69,70. However, environmental health studies have not characterized mental health and socioeconomic factors well. Genetic predisposition can increase the adverse health effects of environmental factors because of higher expression of susceptibility genes or suppression of genes related to resilience to exposure-triggered organ damage11 (Box 1). Individuals with pre-existing cardiovascular disease are at a higher risk of environmentally induced damage, as shown by field studies on traffic noise-induced cardiovascular complications in patients with coronary artery disease71,72. These findings from studies in humans are supported by preclinical data demonstrating that exposure to aircraft noise induces additive cardiovascular damage in mice with pre-established hypertension73. A study in China showed that the autonomic and vascular dysfunction induced by exposure to air pollution was higher in individuals with pre-existing heart disease and other risk factors such as inflammation and overweight74. Data from the MESA-Air cohort also support the presence of aggravated cardiovascular damage mediated by air pollution in patients with established heart disease and early signs of atherosclerosis and calcification75.

Moreover, the risk of plaque rupture has been shown to increase exponentially with higher air pollution levels76,77. The sum of environmental exposures, genetic predisposition and existing cardiovascular disease causes changes in the internal environment, including altered gene and protein expression, leading to dysregulation of central pathways such as stress hormone signalling, circadian rhythms, inflammation, oxidative stress and endothelial dysfunction, all of which increase the risk of major adverse cardiovascular events10. The analysis of all risk sources and the induced biological changes can be used for a precision medicine approach for the early diagnosis and treatment of NCDs78.

Environmental risk factors are crucial in morbidity (DALYs) and early-life and late-life mortality. The loss of healthy life years (or DALYs) induced by chemical pollution is more pronounced in infants12 (Fig. 4). By contrast, the effects of chemical pollution on premature death are more pronounced in older adults12 (Fig. 4). The high number of DALYs in older adults might be attributed to lead toxicity caused by soil pollution12. The high number of DALYs in infants primarily reflects the substantial amount of life years lost when one infant dies and, likewise, the high number of life years lived with a severe disability when one child develops a chronic disease in early life due to environmental pollution79. This high number of DALYs caused by environmental pollution in infants is primarily due to ‘untrained’ defence mechanisms in this population, which makes them less resistant or resilient to high doses of pollutants, for example, owing to lower body weight and epigenetic reprogramming at the neonatal stage. Conversely, the more pronounced effects of chemical pollution on mortality in old individuals can be attributed to the multimorbidity phenotype in older ages. People aged >60 years usually develop multiple comorbidities, including cardiovascular risk factors such as diabetes and hypertension80, which makes them a more susceptible group to the adverse health effects of environmental stressors. In addition, the lifelong accumulation of environmentally induced health damage also has a role in the higher mortality observed in older individuals81. Therefore, these groups have been selected as the primary study population for exposome studies, given that the environmentally triggered dynamic changes in biochemical pathways are most pronounced in these age groups. The SDGs track death from cardiovascular disease, cancer, diabetes and chronic respiratory diseases only in individuals aged 30–70 years. The SDGs do not track deaths in children and adolescents and/or the contribution of chronic diseases triggered by air pollution and other environmental factors82.

Fig. 4: Global deaths and DALYs for different pollutants and age groups.
figure 4

Estimated global deaths (part a) and disability-adjusted life years (DALYs) (part b) in 2015 that were related to exposure to different major pollution factors (occupational, soil, water and air pollution), classified by age at death. The global death and DALY values for the pollutants are presented cumulatively: occupational, occupational plus soil, occupational plus soil plus water, and occupational plus soil plus water plus air pollution. The major environmental cause of DALYs at younger ages is air pollution, whereas the major environmental cause of DALYs at older ages is soil pollution. This effect of soil pollution can largely be attributed to lead toxicity. Data are from the 2015 Global Burden of Disease study189.

Full size image

Assessment of the exposome

Assessment of environmental exposures

Several complementary methodologies, including measurements and mathematical exposure models, can be used to assess environmental exposure. These methods comprise geographical information systems with remote sensing, global positioning systems and geolocation technologies as well as stationary, portable and personal sensors, including mobile phone-based sensors and self-reported questionnaire assessments that record personalized external exposure estimates (Fig. 5). Large-area or global exposure maps for air pollutants, night-time light pollution43 and temperature are primarily based on satellite data (for example, Copernicus pollution data (7 × 3.5 km resolution) from the Copernicus Sentinel-5P satellite) supported by ground-based measurements that are primarily from governmental sources. In the past 10 years, mobile and personal sensors have been widely available. These devices can provide cross-regional exposure information by collecting data via wireless transmission from the network of phone devices. These data can be used to develop atmospheric models such as the high-resolution regional Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) for atmospheric chemistry processes83 and the low-resolution global ECHAM5/MESSy Atmospheric Chemistry (EMAC) model for atmospheric chemistry–climate interactions84. Both state-of-the-art models estimate concentrations of major pollutants (such as PM2.5 and ozone) by accounting for their atmospheric photochemistry, emissions (natural and anthropogenic) and sinks85,86.

Fig. 5: Assessment of environmental exposures and lifestyle factors.
figure 5

Personal or mobile sensors enable the continuous assessment of environmental exposures (such as air pollutants, noise, ultraviolet (UV) and artificial light at night) and health parameters (such as blood pressure, physical activity, and sympathovagal and emotional parameters) of an individual. Personal data can also be acquired via daily questionnaires submitted via app-based software. Improved exposure data from satellite, aircraft, drone and ground-based measurements provide a detailed spatial and temporal exposure history in relation to the global positioning system (GPS) location of the target individuals. In addition, land-use regression models allow accurate exposure estimations for air, noise or light pollution at the address level. A combination of all assessment methods of exposure allows the creation of exposure maps. Smart cities can facilitate intelligent sensing of exposures and can protect the individual from environmental hazards via early warnings sent to the individual’s mobile devices and through intelligent exposure management, for example, by advanced control of traffic, waste and nocturnal illumination and by the implementation of green spaces. The flow chart provides an example of a personalized sensor framework for continuous assessment of environmental exposures and lifestyle history for exposome studies. The exposure maps are reprinted with permission from the European Environment Agency (particulate matter with diameter ≤2.5 µm (PM2.5) levels in Europe), the UK Department for Environment, Food and Rural Affairs, © Crown 2004 (NO2 levels in London), US Department of Transportation, Bureau of Transportation Statistics, National Transportation Noise Map (noise pollution in the USA), and ref. 43, AAAS (global artificial light at night).

Full size image

Traffic-related air pollution is usually determined by a combination of ground-based detectors and land-use regression models comprising governmental traffic and digital ground model data as developed by the ESCAPE project using leave-one-out cross-validation87 or more advanced approaches using tenfold cross-validation88. These models consider the proximity of buildings to main roads, traffic streams, density and landscape characteristics. For the estimation of road traffic noise, address-specific calculations using a dedicated noise exposure model are necessary for obtaining high-quality exposure information, including information on 3D models of buildings, noise screens, traffic intensity and speed, and acoustic reflections. These data can be used to create metropolitan, national or worldwide exposure maps for different environmental exposures (Fig. 5). Given that all these data are usually stored for a long time, they allow inference of retrospective associations between exposure and health outcomes going back for decades and, in the best cases, provide a retrospective lifelong exposure history of populations.

Chemical exposures from soil, water and air pollution are assessed by sampling at multiple geographical positions, followed by detailed analysis (for example, mass spectrometry and atomic absorption spectroscopy) performed mainly by governmental laboratories. The exposure to these chemicals can also be assessed by analysis of human samples such as urine and blood16,89.

Non-chemical and non-physical exposures, such as mental stress (from work strain, mobbing, social isolation or profound grief), dietary habits, smoking, physical activity and other lifestyle factors, are assessed mainly by self-reported questionnaires16. Examples of questionnaires, for instance, for assessing noise-associated annoyance, sleep impairment or chronotype, and circadian rhythms, have been previously described and are available for use90,91.

Personal and mobile sensors

The rapidly evolving technology of small, wearable sensors is generating opportunities to measure the exposome, emotional states, behaviours and physiological responses on a sufficient scale for cardiovascular health research92,93,94. The use of wearable sensors and global positioning systems, for example, from everyday technology such as smartphones, can provide detailed information on the location, physical activity, mode of transport, environmental exposures (such as air pollution, noise and UV light), physiological responses (blood pressure and heart rate (variability), glucose levels, body temperature, or galvanic skin response) and emotional status of a person95,96,97,98,99,100. This information can be overlaid with information on the built environment that is derived, for example, from remote sensors, to obtain detailed insights into how multiple dimensions of the urban environment interact to shape the health of an individual92 (Fig. 5). The feasibility of such an approach was demonstrated by a study examining the use of a personal sensor set to continuously measure exposures over 24 h to assess part of the urban exposome and acute health responses97. Nevertheless, many challenges remain with such an approach, including implementing the methodology for an extended period and in larger populations, improving the ease of wear (for example, through miniaturization and extending battery life) and reducing costs101. In addition, the creation of smart cities will help to understand the exposome and to protect individuals from environmental hazards102 (Fig. 5 and Box 2).

Biomarkers of the exposome

In this section, we provide a short description of the main biomarkers for assessing exposure-triggered biochemical and functional changes in individuals (the ‘internal’ exposome or environment) as a detailed description of these biomarkers is beyond the scope of this Review. Direct determination of the presence and levels of pollutants in the blood (such as heavy metals and pesticides) allows the drawing of conclusions on exposure levels of the individual and also provides information on the potential biochemical reactions of these chemicals89. These biomarkers can be measured with omics methodologies used in genome, epigenome, transcriptome, proteome, metabolome, lipidome, microbiome and adductome research103,104. Multiple examples of omics-based characterization of cardiovascular disease as well as exposure-triggered adverse health effects have been reported previously105. In particular, redox-proteomics and phospho-proteomics provide mechanistic insights into cellular signalling; for example, as shown by a study demonstrating the strong effects of the redox proteome (sulfhydrome) on vascular function106 and a study that identified novel phosphorylation targets and kinases involved in the development or prevention of heart failure107. In addition, functional biomarkers are currently being established using biochip technology, allowing high-throughput cell culture-based assays such as receptor activity tests (for example, for androgen, oestrogen, G protein and aryl hydrocarbon signalling)104 and oxidative stress response reporter cell assays (such as for assessment of the activation of nuclear factor erythroid 2-related factor 2, a transcription factor that shows responsiveness to most environmental stressors108).

With a focus on specific metabolic, proteomic or transcriptomic signatures, exposure–health association studies measure stress hormones (catecholamines109 and cortisol110), indicators of disruption of circadian rhythms (melatonin111 and other factors112,113), markers of oxidative stress (malondialdehyde, 4-hydroxynonenal and 3-nitrotyrosine114,115) and inflammatory responses (cytokines and chemokines35,116). A study in humans demonstrated that a one-night exposure to railway noise caused changes in the plasma proteome (measured by Olink chip) and endothelial dysfunction117. These findings are supported by results from a study in mice showing that exposure to noise induces plasma proteome and lipidome changes (measured by liquid chromatography–mass spectrometry (LC-MS)) and microvascular dysfunction118. Both studies found that exposure to noise induced an inflammatory phenotype. An assessment and critical comparison of these approaches can be found in previous reviews of holistic exposome approaches103,104,119 and the blood exposome89. The hierarchy and combination of these approaches were also reviewed120. The most important current and potential future biomarkers are presented in Table 2.

Table 2 Present and future biomarkers of the exposome
Full size table

The different analytical work-up requirements for analysing the exposome with LC-MS depend on the water or lipid solubility of the analysed chemicals. In addition, the extremely wide concentration range of chemicals and metabolites in the blood — ranging from 160 fmol/l to 140 mmol/l, which represents a range of 11 orders of magnitude — is an analytical challenge89 (Fig. 6). Toxic chemicals of high concern are heavy metals and metalloids35 as well as common toxic compounds originating from industrial production processes such as bisphenols (bisphenol A), persistent organic pollutants (including pesticides such as dichlorodiphenyltrichloroethane), polychlorinated biphenyls, and perfluoroalkyl and poly-fluoroalkyl substances (PFASs)67. Current analytical methods are so sensitive that daily drug use in general populations can be monitored in wastewater by LC-MS121,122. Although untargeted, high-resolution mass spectrometry approaches can detect >30,000 small molecules in human serum, these methods cannot reliably measure blood concentrations of <0.1 μmol/l in a 50-μl sample89. Considering the wide concentration range of these small molecules in human blood (Fig. 6), untargeted analyses might miss about 90% of pollutants and 30% of endogenous and food chemicals, including hormones, carcinogens and endocrine disruptors. Databases on environmental toxins and their biochemical and health effects can help to promote exposome research (Box 3).

Fig. 6: Blood biomarkers of the exposome.
figure 6

a, Concentrations of small molecules and metals in human blood. Each line represents the cumulative distribution of concentrations of chemical compounds from a particular category (drugs: n = 49; food chemicals: n = 195; pollutants: n = 94; endogenous chemicals: n = 1,223). Within each compound category, concentrations cover a 107-fold range. Whereas the mean blood concentrations of drugs, food-derived chemicals and endogenous metabolites are 0.3–1.0 µmol/l, blood levels of pollutants are 1,000-fold lower. The cumulative percent refers to the proportion of each of the four distributions that are below a given blood concentration. For example, for pollutants (red line), approximately 50% of the chemicals had a blood concentration of <10–4 µmol/l, which is the blood concentration of hexachlorocyclohexane (S. M. Rappaport, personal communication). BDE 100, 2,2′,4,4′,6-pentabromodiphenyl ether; DDE, 1,1-bis-(4-chlorophenyl)-2,2-dichloroethene; OCDD, 1,2,3,4,6,7,8,9-octachlorooxanthrene; PCB 170, 2,2′,3,3′,4,4′,5-heptachloro-1,1′-biphenyl. Adapted with permission from ref. 89, EHP.

Full size image

For genomic analyses, arrays are now available to measure from 500,000 to 5 million single-nucleotide polymorphisms (SNPs) in a single analysis104, which is essential to determine the genotype across the whole genome (for example, in comparison to a reference genome, as explained in Box 1). A targeted approach to reveal the genotype related to specific biological pathways, such as the metabochip for SNPs in genes encoding metabolic proteins or the immunochip covering immune-related genes, would be cheaper than SNP-based arrays and sufficient for most exposome studies. Epigenetic analyses can be used to identify exposome-induced changes in DNA besides alterations in the DNA sequence, including DNA methylation patterns123,124,125, microRNAs (which regulate multiple cardiovascular disease pathways126) and histone modifications (such as methylation or acetylation, which have a high predictive potential for cardiovascular diseases127,128). The epigenome changes with age, and each cell type has its own epigenome. Therefore, analysing a mix of cell populations, rather than assessment with single-cell techniques, will provide information on the so-called metagenome.

Inflammation is an essential biomarker of exposure to chemical pollutants such as heavy metals or air pollution35. The EXPOsOMICS consortium showed that oxidative stress and inflammation markers are associated (via DNA methylation) with the adverse effects of air pollution on cardiovascular and cerebrovascular disease125. Cohort studies also found a significant association between plasma C-reactive protein levels (as a readout of systemic inflammation) and exposure to air pollution129. An analysis of the HELIX cohort showed that prenatal exposure to mercury is associated with an increased risk of non-alcoholic fatty liver disease and inflammation in childhood130. The EXPOsOMICS consortium established an association between air pollution and cardiovascular disease by demonstrating a substantial dysregulation of metabolic pathways using metabolomics131. Analysis of the microbiome104 and adductomics studies (that assess adducts of chemical pollutants with DNA or proteins132,133,134) can also be used to detect the biological effects of exposures. The SAPALDIA study established an association between arterial stiffness (a subclinical functional marker) and night-time noise exposure, which is a well-known risk factor for cardiovascular disease135. Exposure–health association studies, especially field studies, have also measured heart rate variability136, the degree of sympathovagal activation and endothelial function (for example, by flow-mediated dilatation) in response to noise exposure117,137 as well as markers of thrombosis and ischaemia, blood pressure, heart rate, and brachial artery vasodilatation after inhalation of diesel particulate matter138,139,140.

Exposome studies of cardiovascular outcomes

Selected exposome studies and relevance to cardiovascular risk prediction

Only a few studies conducted so far fulfil the criteria for a ‘real’ comprehensive exposome approach that includes analyses of environmental, lifestyle and socioeconomic exposures as well as biomarkers of exposures and outcomes (the internal environment; assessed, for example, using various omics approaches) and manifest cardiovascular diseases. These studies include the EXPOsOMICS project, a few studies from the HELIX and PACE projects, and the European Human Exposome Network (EHEN). Therefore, most of the available knowledge on environmental exposures in relation to the risk of cardiovascular disease originates from more traditional epidemiological studies.

Traditional epidemiological studies

These studies are based on few exposures (such as air pollution, traffic noise and green infrastructure) linked with either biomarkers (such as markers of inflammation (C-reactive protein)129, stress hormones (saliva cortisol)110, DNA methylation123 or subclinical cardiovascular markers such as arterial stiffness135) or manifest cardiovascular disease in analyses adjusted for socioeconomic status and lifestyle. Examples of important contributions to the research field using this approach include the ESCAPE and ELAPSE projects. These studies used harmonized estimations of various air pollutants (including PM2.5 and NO2) across several European cohorts. The results showed an association between high levels of these pollutants and a higher risk of cardiovascular outcomes, including cardiovascular mortality23,141,142, incidence of acute coronary and cerebrovascular events143,144,145, and hypertension146. Given that many of the cohorts in the ESCAPE and ELAPSE studies include biobank material, the mechanistic aspects (changes in the internal environment) can be measured at later stages, which would facilitate real exposome studies in the future. The SAPALDIA consortium reported a mechanistic role of DNA methylation (with pathway enrichment in the C-reactive protein cluster) for the association between traffic-related air pollution and noise and cardiovascular outcomes123. The consortium also found associations between traffic noise and arterial stiffness135. A list of larger cohorts that could fulfil the requirements of a complete exposome study by using available geocoding or biobanks for internal environment determination has been provided in the literature147.

Comprehensive exposome studies

More comprehensive exposome studies assessing the effects of the exposome on cardiovascular health are needed, but the following examples of exposome studies have reported some trends. The EXPOsOMICS project aimed to develop a new approach to assess environmental exposures, primarily focusing on air pollution and water contaminants. The consortium used a broad range of omics techniques and estimated biomarkers (such as oxidative stress, inflammatory cytokines and adverse DNA methylation patterns)124. Initial results of the EXPOsOMICS project indicated a mechanistic role of oxidative stress and inflammation in the association between air pollution and cardiovascular and cerebrovascular disease125. In addition, disturbance of metabolic pathways seems to be a pathological mechanism of air pollutant-mediated development of asthma and cardiovascular disease131. The EXPOsOMICS consortium is seeking to establish multiple collaborations with other exposome projects and studies to include more cardiovascular end points and biomarkers148.

In the HBM4EU-aligned cohorts, increased exposure to PFASs was associated with a lower BMI in teenagers149. The HELIX project focused on only early-life exposures150,151. The project defined several cardiovascular end points (including hypertension and cardiometabolic diseases) to be studied for potential associations with blood concentrations of chemical pollutants such as PFASs and pesticide metabolites152,153,154,155. The HELIX studies also found associations between air pollutants (NO2, PM10 and PM2.5) and higher plasma levels of IL-8 and hepatocyte growth factor, whereas only NO2 was associated with increased blood pressure156. The PACE consortium found an association between NO2 exposure during pregnancy and differential offspring DNA methylation in mitochondria-related genes and differential methylation and expression of genes involved in antioxidant defence pathways157. Prenatal exposure to PM10 and PM2.5 changed the DNA methylation pattern of genes related to respiratory health158. Another example is the Generation R study, which found that maternal urine bisphenol concentrations during pregnancy were associated with smaller carotid intima–media thickness (an early marker of arterial health) in the offspring during childhood159.

At present, the contribution of exposome studies to improving our understanding of the cardiovascular health effects and mechanisms of environmental exposures has been small, given that we already knew that oxidative stress, inflammation, impaired circadian rhythm, dysregulated metabolism and elevated levels of stress hormones promote cardiovascular disease. Nevertheless, hopefully, ongoing and future exposome studies will provide further insights, especially if the studies provide lifelong perspectives.

The European Human Exposome Network

The EHEN is the world’s largest network studying the effect of environmental exposures on health. This network is expected to provide substantial new knowledge in the coming years on the effects of the exposome on human health over the life course. EHEN was initiated in 2020 with a € 106 million grant from the European Commission. The project consists of nine large-scale research projects, of which four projects aim to investigate the effects of the exposome on cardiovascular outcomes, covering a plethora of external and internal exposures over the life course147,160,161,162. The ATHLETE project will gain information on the external environment (urban, chemical, behavioural and social exposures), the internal environment (including DNA methylation, transcriptomics and metabolomics) and early cardiovascular markers (such as cardiac and large-vessel imaging and blood pressure) to investigate the effects of the exposome longitudinally from pregnancy to adolescence, using 16 prospective birth cohorts161. The EXPANSE project will study the effect of the urban exposome on cardiometabolic disease by combining external exposome (such as air pollution, noise and temperature) and health data from 55 million adult Europeans, using omics data from >2 million of these individuals (from birth and adult cohorts), and performing exposome assessments for 5,000 people and screening for chemicals in 10,000 blood samples160. In birth and adult cohorts, the LongITools project will gain information on external exposures (air pollution, noise, green space and built environment), internal exposures (metabolomics, epigenomics and transcriptomics) and cardiovascular outcomes (such as repeated blood pressure and cardiac measurements) to elucidate molecular pathways underlying the associations between environmental exposures and cardiovascular health trajectories147. The EPHOR project will study the effects of the external occupational exposome (such as chemicals, particles, noise, strenuous physical activity, psychosocial strain and UV light) on the risk of cardiovascular disease, using a European cohort of >20 million workers162. Together, these four projects will provide an unprecedented data source to study the effects of the exposome on cardiovascular markers and diseases over the lifetime of individuals.

Opportunities, limitations and future directions

Information on the effect of environmental stressors that cause a specific disease or death will be available only when the total amount of temporally and spatially resolved environmental exposures can be quantified. These assessments include continuous measurement of the concentrations of various environmental toxins in air, water and soil to quantify chemical pollution, sound pressure levels for noise, temperature changes, and the intensity of UV radiation as well as repeated questionnaires, ideally throughout the entire lifespan of the individual, to record social status and mental stress conditions. Biological and biochemical parameters should be assessed simultaneously to correlate changes caused by the exposure with subsequent changes in the organism.

Wearable sensors, mobile applications and other technologies provide new opportunities to assess the external exposome and its association with health outcomes such as disease or premature death163. Studying thousands of individuals over their entire lifespan might exceed current data storage capacities and also overtax our current capacity to interpret and integrate the data to generate valuable insights. Therefore, clinical translation might be possible only with current systems biology or bioinformatics methods and technologies. Another limitation is the cost. The assessment of all biochemical changes induced by environmental exposures with the use of multiomics approaches (transcriptomics, proteomics, metabolomics and, potentially also, epigenomics and microbiomics) makes exposome studies very expensive.

Statistical challenges include adequate approaches for handling multiple testing in an exposome context with numerous exposures and outcomes as well as approaches for handling exposure misclassification of various degrees across the exposures in the exposome analysis, which, if not accounted for, will result in differential power to detect associations depending on the precision of the exposure estimation164. Disentangling these highly correlated exposures is difficult even when assessing only a few exposures such as noise and air pollutants. In the statistical approaches currently used in most studies, the best-estimated exposure will ‘win’, although, in real life, a combination of exposures is likely to be causing the disease. Even more advanced statistical models cannot account for one exposure being estimated more precisely than another. In addition, statistical power is often limited if many exposures are evaluated (such as with omics data), owing to multiple testing corrections and because, for most exposures, only low-to-moderate associations with the outcome are expected. A statistical approach to effectively deal with these issues might include using factor analysis techniques to detect intercorrelations between exposure variables and thus reduce dimensionality and increase clustering to a more limited number of underlying collective groups of co-exposures. This approach can also be accompanied by deep machine learning and artificial intelligence-supported techniques to maximize prediction performance. Lastly, there will be many ethical and data protection issues because geographical tracking will be required for optimal exposure assessment, making the included individuals vitreous by archiving crucial personal data on a server.

Conclusions

Determining the exposome is a promising approach to understanding the complex relationships between environment, behaviour, biology, genetics and disease phenotypes at the population level. The need to study the exposome is becoming especially important in assessing the relationship between air pollution and cardiovascular disease, the leading NCD category responsible for global deaths20. In the future, the limitations of exposome studies will be overcome by technological advances that enable us to handle big data, with a decrease in the costs of multiomics assays and bioinformatic analyses119. Nevertheless, we still must solve many justified ethical concerns. If we can solve these issues, a global and complete exposome study or environment-wide association studies will help to identify the effect of specific exposures on our health and thereby allow targeted prevention strategies to effectively mitigate the adverse health effects of environmental stressors. These prevention strategies, in turn, will be cost-effective for the global health system.

More important is identifying and quantifying the possible additive adverse health effects of co-exposures caused by the colocalization of environmental stressors, particularly in urbanized areas (Fig. 2) as demonstrated for cardiometabolic disease and type 2 diabetes57. The adverse health effects of co-exposure to environmental pollutants and stressors might surpass the worst-case scenarios. Therefore, we should focus on mitigating the adverse health effects of environmental exposures, which might synergize with other global environmental policy programmes. For example, a rapid phase-out from fossil fuels towards clean, renewable energy will save 3.6 million lives per year worldwide165, mainly owing to the prevention of NCDs such as ischaemic heart disease, stroke, diabetes and arterial hypertension20. Therefore, the acute phase-out of fossil fuels represents an effective strategy for the prevention of pollution-induced excess cardiovascular mortality while limiting global warming to 2 °C (ref. 165). This strategy would be an advantage even from an economic perspective, given that controlling ambient air pollution in the USA has yielded about $30 in benefits for every $1 invested. A similar economic benefit resulted from the removal of lead from fuel12.

The world population will reach 10 billion people by 2050. Considering that 75% of this population will live in urban areas, where they will contribute to 60–80% of the global energy use and be responsible for 70% of the greenhouse gas emissions, we need to develop new urban concepts to create a healthier city exposome10 such as the Superblocks model in Barcelona166, the 15-Minute City in Paris167 or the car-free city in Hamburg166. These heart-healthy city structures will reduce motorized traffic in streets; decrease heat-island effects; and provide access to green space, active travel space, and safer walking and cycling environments10. Another powerful way to create a healthier exposome would be to adhere to the new WHO air pollution guideline for PM2.5 levels, which was reduced to 5 µg/m3 in 2021, meaning that about 99% of all people worldwide live in areas with PM2.5 concentrations higher than those recommended by WHO. Of note, the EU limit of PM2.5 concentration of 25 µg/m3 is fivefold higher than the one recommended by WHO. This difference in thresholds might be crucial when considering that data on the relationship between exposure levels and cardiovascular events suggest no lower threshold limit, with studies demonstrating a strong relationship at levels below current regulatory limits76,168,169. However, there should be a reasonable compromise between population health benefits, socioeconomic costs and individual restrictions by air pollution legislation.

Therefore, in general, exposome research delivers an exciting approach to understanding the extent to which the environment is associated with the initiation, progression and prognosis of cardiovascular diseases, thereby delivering a framework to explain mechanisms and exposure pathways across the life course of the individual. We conclude with a quote from Gary W. Miller from the Mailman School of Public Health (Columbia University): “For years, the exposome has been a non-descript, fuzzy, and, at times, a rather unscientific idea. Today, we can say that the exposome represents a new science — a new way of approaching how the environment influences health. Truly a new discipline. As such, the field must establish and defend its identity, develop its core principles (i.e., the constellation), and forge a path forward (journal, society, regular meetings, websites, etc.)”170.