Effects of Airborne Particles on Health

16. In contrast to the gaseous pollutants, with which it has often been possible to carry out controlled exposure of volunteers and of animals and thus reach reasonable conclusions about concentrations at which harm is likely to occur, no similar studies with mixtures of particulate pollutants characteristic of ambient PM₁₀ have to date been technically possible. Thus, all the evidence which the Panel has considered has come from epidemiological studies of populations. Most have considered short-term effects on health and these have been of two types: first, analyses of health events, such as patterns of mortality or episodes of hospitalisation, relating these in time to episodes of air pollution, and, secondly, analyses over longer periods of the inter-relations between health events and routinely recorded concentrations of air pollutants. Several problems arise in such population studies. Most important is the weather, which is not only partly responsible for the occurrence of pollution episodes but is also strongly related itself to patterns of ill-health. In the United Kingdom, for example, cold weather in winter is associated both with urban pollution episodes and also with substantial increases in death rates and hospitalisations, making it very difficult to disentangle what are generally quite large effects of the weather (and sometimes associated viral epidemics) from the rather less striking effects of accompanying air pollution. A second important problem is the fact that air pollutant measurements have generally been made on an area and intermittent basis, and do not adequately represent the actual exposures of individuals in the population. Indeed, there may be a very wide range of individual exposures in people living in an urban area when the PM₁₀ sampler shows just one value, some of these people being exposed to considerably higher concentrations. This one value is of necessity used in epidemiological studies to represent the exposures of individuals collectively. It is important to bear in mind that this technique, though perfectly acceptable scientifically, means that the true effect of low concentrations of PM to on individuals cannot be determined with confidence from such studies.
17. The most usual objective of studies of the effects of particulate air pollution has been to investigate the possibility of associations between measurements of the pollutant and events such as death, changes in symptoms or lung function, episodes of hospitalisation, or doctor consultations, and to estimate the likelihood of any associations being due to chance. Such studies need to take account of other factors, such as weather, that may affect concentrations of air pollutants and that may also affect health. Individual epidemiological studies can usually describe the strength of any associations between pollution and health; the likelihood of these being cause and effect depends also on such factors as the consistency of the findings across many studies and the demonstration of relationships between intensity of exposure and the effects. In addition, it is considered desirable (though not essential) that any association should be understandable in terms of known biological mechanisms before being accepted as causative.
18. The Panel have reviewed a large number of published studies investigating the association of particulate air pollution with excess mortality. The original London studies from the 1950s to the 1970s showed a relationship between rises in Black Smoke/sulphur dioxide and excess numbers of deaths from heart and lung diseases, although the absolute concentrations of pollutants then were very much greater than those occurring in the United Kingdom today. Subsequent studies in the United States have shown that these associations can still be demonstrated at the lower concentrations found in a number of cities in that country. An analysis of eight studies in diffeient United States cities has calculated that a rise in PM₁₀ of about 10 µg/m³ (as a 24-hour average) may be associated with an increase in daily mortality of about 1%. Such an analysis of published studies has allowed an expert group of the World Health Organization to calculate the likely excess numbers of deaths associated with different concentrations of PM₁₀, and these are shown in Table 3. In four of the eight United States studies a breakdown of individual causes of death was given. Death from heart diseases, which was responsible for 45% of all deaths, showed an increase of 1.4% in relation to a rise of 10 µg/m³, while death from lung diseases, which caused 5% of all deaths, rose by 3.4%. The strongest association was between death and average PM₁₀ exposure over the preceding five days.
19. The Panel considered whether such excess death rates represented either more people dying overall, that is an increase in absolute mortality, or rather the deaths of already ill people being brought forward, perhaps by only a few days, and therefore fewer dying over the subsequent period. There is as yet little direct evidence on this, but the excess deaths are most clearly seen among older people, and are caused by acute worsening of conditions, such as coronary artery disease and chronic lung disease, that are most unlikely to have arisen as a direct result of a recent pollution episode. On the assumption that the demonstrated associations are indeed causative, we concluded on the basis of available evidence that PM₁₀ pollution episodes are most likely to exert their effects on mortality by determining the time of death of those rendered susceptible by preexisting disease. The Panel was not, however, able to dismiss the possibility that prolonged exposure to air pollution may contribute to the development of these diseases. One long-term study of six United States cities with contrasting levels of pollution has shown significantly higher mortality rates (overall and for diseases of the heart and lung combined) in the most polluted city. In addition, recent data from a large cohort in the United States give some support to the hypothesis that long-term exposure to particulate pollution in the past may have increased risks of lung cancer as well as showing an effect on mortality from heart and lung diseases. We remain uncertain as to whether the confounding effects of social class have been adequately controlled for in these studies, but since the highest exposures to air pollution in cities commonly go hand-in-hand with other confounding factors such as cigarette smoking, poverty, poor housing, and unemployment, it is unlikely that any practicable study in the near future will produce more reliable results. Therefore, we conclude that long-term health effects remain a possible consequence of exposure to particulate pollution. In public health terms, however, any such effects of the levels of air pollution currently occurring in the United Kingdom are likely to be very small compared to those of the better-recognised social determinants of mortality mentioned above.
20. Since, as pointed out in paragraph 10 above, diesel exhaust is an important contributor to PM₁₀, in urban areas, the Panel have considered the evidence associating lung cancer with exposure to diesel exhaust. Studies in two different strains of rat have shown that sustained, long-term exposure to high concentrations of inhaled diesel engine exhaust is associated with an increased incidence of benign and malignant lung tumours, the increase related to the exposure concentrations. Epidemiological studies have suggested an increased risk among heavily exposed workers, although this has not been a consistent finding in all investigations. The International Agency for Research on Cancer has classified diesel engine exhaust as a probable human carcinogen, in that there is limited evidence of its carcinogenicity in humans and sufficient evidence of carcinogenicity in experimental animals. The Panel considers that extrapolation from the observations made in people with heavy industrial exposure indicates that any risk of lung cancer from the concentrations found in the streets of the United Kingdom is likely to be exceedingly small. We have therefore given greater weight to other health effects when discussing the basis of an Air Quality Standard.
21. It has been suggested that the demonstrated relationships between particulate air pollution and deaths are unlikely to represent cause and effect, since it is not plausible that such low concentrations of particles could cause people to die of heart and lung disease and stroke. However, the relationships are remarkably consistent between different studies, and results similar to those in the United States and the United Kingdom have recently been found in cities of other countries such as Greece, Germany and China. There is also evidence of a relationship between the magnitude of the effect and the concentration of particles to which the population has been exposed. Such statistical associations increase the likelihood that the relationship is casual. It should be remembered, as mentioned in paragraph 16 above, that within a population exposed to particles measured at a fixed point, there will be some individuals who are exposed to either lower or much higher concentrations than those recorded centrally. Thus, not only will there be a range of susceptibility in the population but also there will be a range of exposures, and it is likely that some of those who are the most susceptible will be exposed to relatively high concentrations (compared with those monitored centrally) during any pollution episode.
    Furthermore, the Panel consider it plausible that inhalation of particles of the physico-chemical types characteristic of urban air pollution episodes at appropriate concentrations could cause lung inflammation and this in turn could precipitate episodes of cardiovascular and pulmonary illness in susceptible individuals. We have therefore taken the view that episodes of particulate air pollution are responsible for causing excess deaths among those with pre-existing lung and heart disease.
22. Associations between ambient concentrations of particles and other indices of ill-health have also been reported. Again, the confounding effects of weather need to be taken into account, but there is consistent evidence that rises in concentrations of PM₁₀ may be associated with increased numbers of admissions to hospital, increases in reported symptoms, and decreases in lung function. While not all studies have shown identical results, the coherence of the overall pattern is consistent with the hypothesis that particulate pollution is able to cause temporary worsening of already existing lung disease. The Panel believe that these results are also plausible evidence of a causative relationship, and consistent with the findings discussed above. This has been examined in detail by the Department of Health's Committee on the Medical Effects of Air Pollutants, whose conclusions have been made available to us. The expert group of the World Health Organization also estimated the approximate effects of different concentrations of PM₁₀ on some indices of ill-health (see Table 3).
23. There has been a well-documented rise in the prevalence³ of such allergic disorders as asthma, hay fever and eczema in the United Kingdom and other countries with a western lifestyle. This topic also has been examined in detail by the Department of Health's Committee on the Medical Effects of Air Pollutants, whose conclusions have been made available to us. The Panel have considered the possibility that outdoor air pollution, including particles, might have been in part responsible for this rise in allergic disease since, if so, such an effect would need to be considered in recommending an air quality guideline. We have concluded that there is no clear evidence that pollution at the concentrations found in the outside air is able to cause asthma (as opposed to provoking attacks in people who already have the disease) and we are of the view that outdoor air pollution is very unlikely to have contributed to the observed increase to any significant degree. It should be noted that this increase in allergic disease has occurred especially in young children over a period during which ambient concentrations of particles in the air of towns and cities in the United Kingdom have decreased.

Table 3 Summary of short-term exposure-response relationship of PM₁₀ with different health effect indicators

Health effect indicator	Estimated change in daily average PM10 concentration needed for a given effect (in µg/m3)*
Daily mortality: 5% change 0% change 20% change	50 100 200
Hospital admissions for respiratory conditions: 5% change 10% change 20% change	25 50 100
Numbers of asthmatic patients using extra bronchodilators: 5% change 10% change 20% change	7 14 29
Numbers of asthmatic patients noting exacerbation of symptoms: 5% change 10% change 20% change	10 20 40

*   Adapted, with permission, from WHO Regional Office for Europe, 1995. The original document indicates the published studies on which these estimates are based 

24. Since the pioneering work in London from the 1950s to the 1970s, relatively few investigations of the effects of particulate air pollution on populations have been carried out in the United Kingdom. The Panel have however reviewed two recent studies, in London and Birmingham. During an air pollution episode in London in December 1991, daily concentrations of Black Smoke rose to 228 µg/m3 along with hourly nitrogen dioxide concentrations up to 423 ppb.⁴ This episode was associated with an increase in overall mortality of about 10% and smaller increases in hospital admissions for lung disease and doctor consultations for upper respiratory symptoms, these adverse effects being confined to older adults. These findings are comparable with what might have been predicted on the basis of studies conducted on the effects of particles in the United States and elsewhere, but the relative contributions of the two pollutants remain unclear. The Panel commissioned a special study, in Birmingham, of hospital admissions over a 2-year period, from April 1992, in relation to measurements of PM₁₀ in that city. The analysis took account also of temperature and concentrations of sulphur dioxide, nitrogen dioxide and ozone. Associations between PM₁₀ and lung and stroke admissions were found, though the latter association was weak; no association was found with hospital admissions for heart disease. On the basis of these analyses, a rise in average daily PM₁₀ of 10 µg/m³ would be predicted to result in an increase of about one admission to hospital every other day for lung illness in Birmingham, a city of about 1 million people.
25. We have agreed that well-conducted studies have shown a relationship between concentrations of PM₁₀ and health effects, such that the higher the concentration of particles, the greater the effect on the health of the population and, conversely, the lower the concentration, the smaller the effect. Such studies, however, have not been able to show whether there is a threshold concentration below which effects do not occur. The Panel considers that this may in part be a consequence of the fact, mentioned in paragraph 21, that measurements of particles made at a central monitoring site in such studies do not reliably reflect the exposures of all individuals in the area, which may have been higher or lower depending on their proximity to sources of pollution and their patterns of activity. Thus, where epidemiological studies have suggested an effect on health of very low concentrations of particles, this may be attributable to episodes of illness occurring in susceptible people who had in fact been exposed to higher concentrations. The concept of a threshold is particularly important; if a threshold does not exist there is no theoretical basis by which a standard can be fully justified in terms of improvement of health, since it could be argued that any figure chosen would have a smaller benefit than any lower figure, eventually down to zero. It is however arguable, since the risks to the population become smaller as the concentrations of PM₁₀ fall, that there will come a point at which pollution would affect only a tiny minority of the population and further decreases in concentration would become ineffective in terms of public health benefits. We have therefore taken the view with respect to particles that there is likely to be a concentration at which adverse health effects are of only trivial public significance, even though that concentration is not at present known.
26. The Panel have noted that there is likely to be an irreducibly low concentration of particles in the air, due to natural processes. Indeed, even if all man-made particle generation ceased in the United Kingdom, because of the long-term persistence of the smallest particles this natural background concentration would be increased by drift from other countries, just as currently particles generated in this country will contribute to the world-wide concentration.

⁴   1 part per billion (ppb) is one part, by volume, in one thousand million, or 1 in 10⁹; 1 ppb of nitrogen dioxide is equivalent to 1.88 µg/m³ at 25"C and 1013 millibars.