APRIL 24, 1997
Subject: EPA's Proposed Air Quality Standard for Particulate Matter

My name is Carl Shy. I was a member of the Panel on Particulate Matter of EPA's Clean Air Science Advisory Committee (CASAC), and am currently a professor of epidemiology in the University of North Carolina School of Public Health. Over the past two years, I was an invited participant and discussant in EPA's workshops on the health effects of particulate matter, and, as a panel member of CASAC, a reviewer of the various drafts of EPA's criteria document for particulate matter and of the staff papers recommending an air quality standard for particulates.

I make this statement to urge the members of the United States Senate to support EPA's proposed revision of the air quality standard for particulate matter, and more specifically to support the proposal to establish a new standard for fine particulates, i.e. for particles less that 2.5 micrometers diameter (PM-2.5). I agree with EPA's proposal that the PM-2.5 standard be established at a concentration of 15 micrograms per meter cubed, annual average.

I suggest that there are three compelling reasons for EPA to establish the air quality standard for PM-2.5 as proposed:

(1) There is ample evidence that there is a causal relationship between population exposure to fine particulates, at concentrations now existing in the air environment of many cities in the U.S., and excess mortality, hospital admissions, respiratory symptoms in adults and children, and decreases in lung function of children. I will expand on my reasons for stating that the evidence for a causal relationship is compelling.

(2) Accepting the causal relationship between air particulates and excess mortality and morbidity in the population, the health burden of exposure to ambient air particulates at current levels in the U.S. is unacceptable, consisting of thousands of excess deaths, hospital admissions and respiratory disease episodes. This large health burden can be addressed by a concerted program to lower the concentration of ambient air particulates in those cities that exceed the proposed annual standard for PM-2.5. This is an achievable goal, one that will have a major health benefit for a majority of the U.S. population.

(3) The Clean Air Act of 1970, amended several times, requires the Administrator of EPA to establish national ambient air quality standards at a level that avoids unacceptable risks and protects public health with a margin of safety. The PM-2.5 standard proposed by EPA, of 15 micrograms per cubic meter annual average, provides a minimally acceptable margin of safety against the mortality and morbidity risks noted above. A number of well conducted epidemiologic studies demonstrate an increase in mortality and morbidity when PM-2.5 concentrations begin to exceed the proposed standard of 15 micrograms per cubic meter. In some cases, this excess mortality and morbidity is observed even when PM-2.5 concentrations reach 16 micrograms per cubic meter, a level that is less than 10 percent above the proposed standard. In the future, we may well find that the 15 microgram per cubic meter standard is not adequate to protect public health, but the proposed standard will at least move our country in the right direction of greatly minimizing the currently unacceptable health burden.

Since the issue of the causal relationship between air particulates and mortality/morbidity of the population is crucial to the three points I am presenting in this statement, I would like to expand on my reasons for asserting that the causal relationship has been established to a degree sufficient to require EPA to propose the new standard for PM-2.5. The criteria for causality I will briefly discuss are the same at those used by the Surgeon General of the U.S. in the 1965 report on the adverse health effects of cigarette smoking. These criteria are widely accepted in the scientific community as reasonable guidelines for assessing causality in the matter of disease risks to populations. These criteria are:

Consistency: Many studies, numbering more than 30, have observed a significant and meaningful relationship between population exposure to air particulates and excess mortality and morbidity (the latter being increased hospital admissions, increased respiratory symptoms, and decrease lung function in children). Many of these studies have been carefully scrutinized for potential sources of error, for confounding by weather factors and season of year, for limitations in statistical methods, and for inadequacies of measurement methods. None of these potential problems were found to diminish the consistently observed relationships between adverse health effects and particulate exposures. EPA's Criteria Document for Particulate Matter, which the CASAC carefully reviewed and approved, provides extensive discussion of these issues and reaches the conclusion that the evidence is fully compatible with a causal interpretation. In comparison with the 1987 major revision of the air quality standard for particulates, when a PM-10 standard was established, we now have a much larger body of evidence regarding adverse health effects of particulate air pollution.

Coherence: Among the more than 30 studies mentioned above, the same types of adverse health effects have been consistently observed. That is, the primary causes of excess deaths in mortality studies are deaths from heart and from lung diseases, and the primary causes of hospitalizations in morbidity studies are from diseases of the heart and lung. Other causes of death and hospital admissions do not show a consistent association with levels of air particulates. This agreement between the different types of studies strongly suggests a causal relationship, since the same disease endpoints are so consistently found when there is a common causal agent.

Exposure-Response Gradient: As levels of air particulates increased, there was a proportional increase in mortality or in morbidity in the above studies. The importance of finding an exposure-response gradient is that it becomes less and less likely that risk factors other than air pollution could explain the disease gradients. To do so, these other risk factors would also have to increase in nearly perfect step with changes in air quality, and there is no evidence that such risk factors exist, in spite of considerable effort to find them.

Cause Precedes Effect: The excess in mortality and morbidity consistently occurs either on the day of the elevated air particulate concentration or one to two days after the elevated concentration occurs. There is no evidence that excess mortality and morbidity precedes days with higher air particulates.

Strength of the Association: The relative increase in total mortality and morbidity associated with a 50% increase in air particulates is not large, being on the order of 5 to 10 percent above the mortality or morbidity observed on days with the lowest concentrations. From this point of view, the association between air particulates and mortality and morbidity is not nearly as strong as it is for tobacco smoking, excess alcohol consumption, or high fat diets. But, because a majority of the population is exposed on many days to air particulate concentrations above the proposed standard, the population burden of air particulate exposures is indeed large, with deaths and hospitalizations numbering in the several thousands during the course of one year. (Estimates range from 10.000 to 60,000 excess deaths alone, associated with days of higher air particulate levels in the U.S.)

Biological Plausibility: This criterion is satisfied when there is experimental evidence that air particulates can cause the type of human health responses that would lead to excess mortality and morbidity. We do not have firm evidence to satisfy this criterion, in part because, when animals or humans were experimentally exposed to particulates, the experimenters were unable to reproduce in the laboratory the complex mixture of particulates in the ambient atmosphere. Recently, these complex mixtures have been introduced into experimental studies, and the researchers report that they found physiological changes in the exposed animals that may indeed be precursors of mortality and morbidity. Furthermore, there is ample evidence from older studies that air pollution at considerably higher levels than now exist were causally related to excess mortality and morbidity in humans. Thus it is entirely reasonable that similar effects, but at lesser magnitude, are occurring at current elevated concentrations.

Specificity: This criterion requires that there be a unique relationship between exposure to the risk factor and the disease caused by this exposure, e.g. the nearly unique relationship that exists between asbestos exposure and mesothioloma of the lung. Very few causes of human disease manifest this unique relationship. Thus cigarettes cause lung cancer, but so do a number of chemicals that exist in the occupational environment. A lack of specificity does not argue against causality, nor does a lack of any one of the causal criteria negate the argument for causality, since each of these criteria are meant to serve as guidelines in making a judgment about a possible causal relationship.

Confounding by Other Air Pollutants: This is not one of the criteria addressed in the Surgeon General's report on cigarette smoking, but it is an issue frequently discussed at scientific meetings on the health effects of air pollution, and it was discussed in some depth at the CASAC meetings. The concern boils down to the question: can we conclude that fine particulates, as indexed by PM-2.5, are the causal agent responsible for the consistently observed excesses of mortality and morbidity? In my opinion, and in the opinion of other epidemiologists who are closely involved with research on the health effects of air pollution, the answer is , yes, it is reasonable to conclude that fine particulates, rather than any other regularly measured air pollutant, is our best measure of the causal agent. No other known air pollutant so consistently demonstrates an exposure-response relationship with mortality and morbidity across the many different studies reported in the literature. Although there are cities in which gaseous air pollutants, such as sulfur dioxide and ozone, are sometimes highly correlated with fine particle concentrations ( and thus prevent drawing conclusions as to the most likely causal component in the air pollution mixture), there are enough studies in which the other pollutants were either present in such low concentrations that adverse health effects from these levels would be unlikely, or there were studies in which the correlation between particles and other air pollutants was sufficiently small to enable the investigators to distinguish between the effects of the different pollutants. Moreover, there is sound biological evidence that fine and ultrafine particles penetrate deeply into the lung and produce an inflammatory response, whereas gaseous air pollutants do not reach the lower portions of the the lung except by becoming attached to fine particles.

There are many hazardous agents, often consisting of a complex of chemicals. that cause significant human health risks, for which we do not know the precise causal component, but which we clearly know is part of the causal chain leading to harm. Several prominent examples of this can be cited, such as cigarette smoke and high fat diets. There are more than 200 chemicals in cigarette smoke, a number of which may cause lung cancer, chronic lung disease, and heart disease, but we do not know precisely which of these chemicals are actually the causal agent for these consequences. The relationship between fine particulates and excess mortality/morbidity is analogous to the smoking-health effects association. Fine particulates are the best single indicator of that component of the air pollution mixture consistently associated with excess mortality and morbidity. From this point of view, fine particulates are the causal agent, in the sense that they are a very useful and predictable marker for the complex of chemical and physical factors that are causing the unacceptable health consequences, just a cigarettes are the causal agent for smoking-related deaths and disease. No other component of the air pollution mix is compatible with the diversity of findings on excess mortality and morbidity reported in the above studies.

For all of the above reasons, EPA has made the argument that a causal interpretation for fine particulates is a reasonable one. If we fail to take action on reducing population exposure to fine particulates, as proposed by EPA, we run the significant risk of continuing to allow thousands of excess deaths and diseases, with all of the associated societal costs, that could otherwise be prevented. It is not reasonable to wait until advances are made in scientific knowledge to prove conclusively that fine particulates are or are not the specific causal agent. We have sufficient evidence that fine particulate are a part of the causal chain, and that reduction in fine particulate levels is an efficient strategy to interrupt this chain.

Thank you for the opportunity to present this statement.


Carl M. Shy, MD, DrPHB Professor, Deparment of EpidemiologyB School of Public Health
University of North Carolina at Chapel Hill
Campus Box 7400
Chapel Hill, NC 27599