Produced by the Royal College of Physicians of Edinburgh and Royal College of Physicians and Surgeons of Glasgow

Air pollution and stroke: is a causative association plausible? (update)

  • Professor A Seaton, Emeritus Professor of Environmental and Occupational Medicine, University of Aberdeen and Senior Consultant, Institute of Occupational Medicine, Edinburgh, Scotland


The links between air pollution and respiratory disease are well established. Research has now suggested that air pollution could also have a causative effect in stroke. How plausible is this association? Prof Anthony Seaton reviews the evidence.

Key Points

  • Air quality standards have improved dramatically in UK cities over the last 50 years.
  • In spite of improvements, epidemiologists can relate death rates from chest and heart disease to apparently trivial levels of pollution particles.
  • Very small pollution particles which reach the alveoli (tiny air sacs where gasses exchange with the blood) may explain this continuing association. These particles can cause inflammation in the alveoli (lung disease) and alter blood clotting factor to make blood vessel clots more likely (heart disease).
  • Recent studies in the Netherlands and in Taiwan suggest small pollution particles may also relate to the occurrence of strokes.
  • Nanometre- (small-) sized pollution particles come increasingly from diesel engines and their relation to health will need further research.

Declaration of interests: No conflict of interests declared

I well remember the air pollution episodes of my childhood in the 1940s and 1950s: cold, still winter days when coal was used for all domestic heating and the smoky fog obscured the houses on the opposite side of the street. It did not take a great leap of imagination to believe that this was in part responsible for the huge numbers of patients we saw in the wards with end-stage bronchitis and emphysema as students and young doctors in the later 1950s and 1960s. Some half a century later, in 1992, I was asked to chair a committee responsible for recommending Air Quality Standards to the United Kingdom Government. Over those 50 years there had been dramatic reductions in air pollution in the cities of the UK, and yet epidemiologists were still able to demonstrate associations between what appeared to be trivial levels of particulate pollution, of the order of a few tens of micrograms per cubic meter, and death rates from respiratory and cardiovascular disease. Yes, cardiovascular disease! And when we looked at the data it became apparent that although the relative risk of death from lung disease was greater than that from heart disease, the absolute risks (since many more people died of heart attacks) were greater for cardiovascular disease. Did it make sense? Could about a milligram of soot, inhaled over 24 hours, cause people to die from heart disease?

A possible explanation of this association came to me in 1994 on picking up my copy of the Lancet and reading a paper that reported the variations in clotting factors in the blood of elderly people in relation to season.1 With colleagues, I proposed the hypothesis that the nanometer-sized particles that largely comprise combustion-generated air pollution cause alveolar inflammation sufficient to alter blood coagulability and thus increase risks of adverse cardiovascular events.2 Since then, evidence has accumulated in support of this hypothesis.3 We have taken it further by suggesting that since the lung’s defences are directed at eliminating microorganisms with the potential to invade and multiply, it makes teleological sense that the lung should respond similarly to other small inhaled particles, both by local alveolar inflammation and by triggering a systemic, acute phase reaction.4 It is important that any hypothesis should explain the known facts, and another observation that is now generally accepted is that not only are air pollution episodes associated with acute episodes of cardiovascular disease but also lifetime experience of pollution modifies long-term risks of heart disease.5 In other words, if you are brought up in a polluted city, you are more likely to develop heart disease than if you have lived in the country. Alterations in risk factors, such as fibrinogen concentrations in the blood, would explain this. But if this mechanism were responsible for the observed associations, air pollution would be expected to influence other diseases where blood coagulability or platelet aggregation is of pathogenic importance. Is this the case?

A problem in investigating these effects of air pollution is that they are weak, and hard to demonstrate except in very large populations. The confidence with which one can state that there is an epidemiological association depends in part on the frequency of the event one is studying, and such conditions as stroke and pulmonary embolism are very much less common causes of death than heart attack. Nevertheless, a number of studies have now shown such associations. To take two which seem reasonably convincing, in the Netherlands, a 4% increase in risk of death from both stroke and thromboembolism has been shown to accompany a rise of 40 μg/m3 in exposures to particulate pollution; the association with stroke was statistically significant, while that for thromboembolism was not quite.6 In Taiwan, a recent study has shown associations between stroke, both ischaemic and haemorrhagic, and particulate pollution, nitrogen dioxide and carbon monoxide.7 These associations were present only in the hot season, except that with carbon monoxide, which persisted in the cooler weather.

Air pollution research is made complex by the presence of multiple factors that vary simultaneously - temperature, atmospheric pressure, wind speed and different sources and mixtures of particles and gases. There is still much discussion about the plausibility of these associations representing a cause-and-effect relationship, part of the difficulty being the extremely low doses of pollutants being associated with serious health effects. Current research therefore is directed at understanding mechanisms and ingenious epidemiological designs are necessary to link this with real-life experience of individuals. My own view is that the most likely mechanism is increase in blood coagulability and/or decrease in fibrinolytic ability together with changes in endothelial and platelet function as a response to alveolar inflammation. I am not disturbed by fact that some reports note associations of disease with particles, some with gases such as NO2 and CO, and some with all of these. All simply represent the same pollutant, combustion exhaust, mainly now from diesel vehicles in the UK, and all are confounded by the same hitherto unmeasured pollutant: nanometer-sized particle numbers.4 For the past couple of years these particle numbers have been counted in major UK cities and it should soon be possible to relate such counts to health end-points. When I tell you that on a moderately polluted day, about 100 μg/m3 in the air means that some 1,000 billion of these particles are deposited in our alveoli in 24 hours, compared to the 50 billion that the lung is used to, you may see why such reactions are possible.

In 1964-5 I spent a year in Stoke-on Trent as a medical registrar, and was impressed with the frequency of aspirin-taking in the patients I saw there with gastric haemorrhage. When I asked why they did so, the answer was always the same - ‘to keep off the dust’. This was the dust from the potteries, notorious for causing silicosis, but Arnold Bennett’s Five Towns had also been extremely polluted from the coal-fired kilns. Decades later, I began to wonder if these patients had a point.


Since writing this article two additional pieces of information have come to light. First, a meta-analysis of published papers investigating associations of air pollution and cardiovascular disease has been published by the Department of Health’s Committee on the Medical Effects of Air Pollution.8 A significant increase in the risk of cerebrovascular disease was detected in relation to exposures to particles (measured as PM10) and to nitrogen dioxide. I have previously discussed the possibility that associations with the latter gas are likely to be confounded with small particles (nanoparticles) from traffic pollution.4

A second study, as yet only presented orally, has investigated the associations of air pollution, measured by a number of different physical and chemical criteria (mass at different diameters, number, gases) and cardiovascular episodes over several years in London. This should soon indicate whether or not nanoparticles are the critical factor in determining these effects.


  1. Woodhouse PR, Khaw K-T, Meade TW et al. Seasonal variations of plasma fibrinogen and factor VII activity in the elderly: winter infections and death from cardiovascular disease. Lancet 1994; 343:435-9.
  2. Seaton A, MacNee W, Donaldson K et al. Particulate air pollution and acute health effects. Lancet 1995; 345:176-8.
  3. Schwartz J. Air pollution and blood markers of cardiovascular risk. Environ Health Perspect 2001; 109(Suppl 3):405-9
  4. Seaton A, Dennekamp M. Ill-health associated with low concentrations of nitrogen dioxide; an effect of ultrafine particles? Thorax 2003; 58:1012-15.
  5. Pope CA, Burnett RT, Thun JM et al. Lung cancer, cardiopulmonary mortality and long-term exposure to fine particulate air pollution. J Am Med Assoc 2002; 287:1132-41.
  6. Hoek G, Brunekreef B, Fischer P et al. The association between air pollution and heart failure, arrhythmia, embolism, thrombosis, and other cardiovascular causes of death in a time series study. Epidemiol 2001; 12:355-7.
  7. Tsai S-S, Goggins WB, Chiu H-F et al. Evidence for an association between air pollution and daily stroke admissions in Kaohsiung, Taiwan. Stroke 2003; 34:2612-16.
  8. Committee on the Medical Effects of Air Pollution. Cardiovascular disease and air pollution. London: Department of Health; 2006.