Air Pollution Study of 10 Million People Links Fine Particles to Heart Hospitalizations

In one of the largest investigations of its kind ever attempted, researchers have traced how even modest levels of airborne fine particulate matter, known as PM2.5, push people into hospitals with heart attacks, strokes and other cardiovascular emergencies. Drawing on the health records of more than ten million residents of the Netherlands, the study reconstructed the air pollution history of each individual over more than a decade, then asked a deceptively simple question: does the danger come from pollution breathed in this week, or from pollution accumulated over years? The answer, published in the journal Environmental Advances, is that both matter, but in strikingly different ways, with different rhythms in the body and different victims across cardiovascular conditions.

The scale of the undertaking is what sets this research apart. Led by Eugenio Traini of the Netherlands Organisation for Applied Scientific Research and colleagues, the team assembled a cohort of 10,156,661 people aged 30 and older who were registered at a Dutch address on 1 January 2012. Rather than relying on sparse monitoring stations, the researchers used LOTOS-EUROS, an open-source chemistry transport model that simulates atmospheric chemistry, transport and deposition to estimate PM2.5 concentrations at a resolution of roughly two by two kilometres, every hour, across the entire country from 2006 to 2019. Those gridded estimates were aggregated to postal code areas and linked to each participant’s residential history, allowing the team to follow people as they moved through landscapes of shifting air quality.

PM2.5 refers to airborne particles smaller than 2.5 micrometres across, roughly a thirtieth of the width of a human hair. These particles are emitted directly from burning solid fuels, road transport and industry, or formed secondarily in the atmosphere from precursor emissions such as ammonia from agriculture and sulphur compounds from shipping. Because of their tiny size, they evade the lungs’ defences, penetrate deep into the respiratory tract, and can enter the bloodstream, where they trigger systemic inflammation, oxidative stress and vascular dysfunction. In the Netherlands, average annual concentrations fell from about 20 micrograms per cubic metre in 2006 to 13 micrograms per cubic metre in 2019, a decline reflecting European and national clean-air policies, yet the study suggests that these remaining, lower levels continue to exact a measurable toll.

The researchers tracked first-time hospitalisations for three cardiovascular outcomes: ischaemic heart disease, myocardial infarction, and stroke, together with a composite outcome capturing the first hospitalisation for any cardiovascular cause. Over the follow-up period ending in 2019, the cohort recorded 437,089 incident ischaemic heart disease hospitalisations, 205,663 strokes and 151,345 heart attacks, totalling 592,259 first cardiovascular hospitalisations. Because modelling time-varying exposure across ten million people is computationally punishing, the team used a nested case-control design: each person hospitalised was matched with four controls of the same sex and exact age who were still event-free at the moment of the case’s admission, a technique known to yield estimates comparable to full cohort analyses.

The short-term findings show how quickly dirty air can tip vulnerable patients into crisis. In fully adjusted models, a 10 microgram per cubic metre increase in PM2.5 averaged over the six days preceding admission was associated with a 5.8 percent higher risk of hospitalisation for ischaemic heart disease, a 5.3 percent higher risk for myocardial infarction, and a 4.7 percent higher risk for a first cardiovascular hospitalisation overall. Even exposure on the day of admission itself, lag zero, carried elevated odds. Stroke was the exception: its estimates pointed in a positive direction but the confidence intervals crossed the null, meaning the acute association could not be confirmed statistically in this population.

The most sophisticated part of the analysis went beyond simple averages. Using distributed lag models, a statistical framework that estimates how risk is distributed across each individual day or year of past exposure, the researchers uncovered distinct temporal fingerprints for each condition. For ischaemic heart disease, the risk signal was clearest in the one to six days immediately before hospitalisation, consistent with the idea that pollution can acutely trigger events in people whose arteries are already compromised. For myocardial infarction, the effect emerged beginning one day after exposure. For stroke, however, a suggestive increase in risk appeared two to three weeks before admission, hinting that cerebrovascular events may follow more prolonged inflammatory and coagulation pathways rather than an immediate trigger.

The long-term results were even more striking in magnitude. A 10 microgram per cubic metre increase in annual PM2.5 averaged over the two years spanning the hospitalisation year and the year before was associated with a 39.8 percent higher risk of ischaemic heart disease admission, a 22.2 percent higher risk of stroke, a 13.4 percent higher risk of heart attack, and a 32.5 percent higher risk of a first cardiovascular hospitalisation overall. Distributed lag analysis showed that risk rose with cumulative exposure up to roughly three years before hospitalisation for ischaemic heart disease and the composite outcome, then plateaued or declined. For heart attack, the cumulative effect peaked around four years before admission. These estimates are larger than those reported in some earlier European studies, possibly because the time-varying exposure assignment reduced misclassification compared with the common practice of anchoring exposure to a single baseline year.

Several mechanisms and caveats frame how these numbers should be read. Effect estimates were consistently slightly higher in men than women, and higher among people aged 65 and older than those aged 30 to 64, though the differences were modest. Sensitivity analyses in a subset of about 305,000 participants with lifestyle data on smoking, alcohol, body mass index and education broadly supported the main results, although with wider uncertainty due to the smaller sample. The study adjusted extensively for individual socioeconomic factors such as income, civil status and country of origin, plus neighbourhood-level indicators including urbanisation and deprivation, and estimates attenuated with adjustment while remaining consistently positive. The authors note that their exposure model explained 72 percent of the variance in monitored concentrations, meaning residual prediction error could blur some lag-specific findings, and that hospitalisation registry completeness declined over part of the study period, a form of misclassification that would likely push the true effects down rather than up.

The researchers also caution that the analysis focused on PM2.5 alone and did not disentangle correlated co-pollutants such as nitrogen dioxide and ozone, which would require multipollutant mixture methods still under active methodological development. Nor did it capture exposures outside the home, at workplaces or during commuting, which would tend to understate rather than exaggerate the true associations. Looking ahead, the LOTOS-EUROS framework offers a promising capability the study did not yet exploit: source apportionment, or attributing particulate mass to specific emitters such as agriculture, industry and traffic, which could reveal which sources drive the greatest cardiovascular harm.

The message that emerges is sobering for a country that has largely cleaned up its air. Average Dutch PM2.5 levels were moderate by European standards and fell steadily throughout the study, yet both short bursts and years of accumulated exposure were consistently linked to hospitalisations at concentrations well below the World Health Organization’s historical guideline limits, echoing pooled European evidence that found no clear threshold below which cardiovascular effects vanish. The outcome-specific timing, acute for coronary events, more drawn-out for stroke, argues for epidemiological studies that treat exposure as a living, changing history rather than a single snapshot. For policymakers, the findings suggest that further reductions in fine particulate pollution, even from already low baselines, would translate directly into fewer heart attacks, strokes and cardiac hospital beds filled.

Subject of Research: Associations between short- and long-term exposure to ambient PM2.5 and cardiovascular disease hospitalisation in a nationwide Dutch cohort

Article Title: Short- and long-term exposure to ambient PM 2.5 and risk of cardiovascular disease hospitalisation: A nationwide cohort study in the Netherlands

Article References: Traini, E., Shen, Y., de Crom, T. O., Pekel, F., Henzing, B., Zhang, J. J., Pseftogkas, A., Ge, C., Timmermans, R., & Pronk, A. (2026). Short- and long-term exposure to ambient PM2.5 and risk of cardiovascular disease hospitalisation: A nationwide cohort study in the Netherlands. Environmental Advances, 26, Article 100756. https://doi.org/10.1016/j.envadv.2026.100756

Image Credits: AI Generated

DOI: 10.1016/j.envadv.2026.100756

Keywords: PM2.5, air pollution, cardiovascular disease, heart attack, stroke, ischaemic heart disease, hospitalisation, Netherlands, distributed lag models, LOTOS-EUROS, cohort study, environmental health