of Dr. Ronald E. Wyzga to
the Commi ttee
on Environment and Public Works United
October 2, 2002
I am Dr. Ronald E. Wyzga. I work for the Electric Power Research
Institute (EPRI), in Palo Alto, California.
EPRI, a voluntarily-funded 501(c
I began my research career working
on the relationship between health
There are a large number of scientific studies that report a link between air pollution and human health. I have personally been involved in some, and EPRI has supported many more. The majority of these studies link particulate matter with health effects; however, some of these studies do not show an association with health, and other studies implicate gaseous pollutants in addition to, or in place of particulate matter. In any consideration of the health and air pollution issue, it is important to keep in mind that air pollution is a complex mixture of many different types of gases and particles. Discerning specific causative agents is a challenge we in the scientific community are working to address. Today I want to highlight some of the work that EPRI has recently been involved in to provide you with some of our latest results.
There have been several major facets to our research:
It is important to
understand which specific components of air pollution are associated with
health impacts. In studies
undertaken to date, the strongest associations between air pollution and health
are with particulate matter. In
studies which include particulate matter (PM) and other pollutants, such as
ozone and carbon monoxide (CO), in their analyses, PM is most consistently
associated with health responses; there are, however, some exceptions where
other pollutants, especially carbon monoxide, are most highly associated with
health responses. Very few studies have considered a comprehensive set of the
pollutants, especially the different chemical constituents of particulate
matter, in their analyses.
2.) The EPRI ARIES (Aerosol Research Inhalation Epidemiology Study) project was designed to examine the toxicity of the various components of PM and air pollution. This study is unique in terms of the number of air quality parameters measured and the number of health effects examined. This study, undertaken in Metropolitan Atlanta in conjunction with several universities, US Department of Energy, and others, characterized the air quality on a daily or more frequent basis for over one hundred air quality variables. This characterization, accompanied by a suite of epidemiological studies, allowed us to examine the influence of the various components of air pollution on a variety of health outcomes.
with PM-2.5 (particles 2.5 microns in size and smaller) and carbon
monoxide. However, the only
fraction of PM-2.5 that showed any association with the cardiovascular effects
were PM2.5 fraction
that has been at the center of attention as the potential cause of negative
health impacts. For total mortality,
the pollutants most consistently associated with premature death are oxygenated
hydrocarbons, substances that to date have had limited study.
EPRI has initiated smaller ARIES-like studies in
Baltimore and St. Louis to determine whether the results from Atlanta can be
A major toxicology effort will start soon in which the
effects of coal combustion emissions will be investigated by exposing animals
5.) EPRI has also been active in trying to understand the implications of alternative statistical methods used in the analyses of epidemiological data. Given the recent discovery that the applications of statistical software have led to erroneous results in some pollution health studies, the EPA is delaying its review of particulate matter health effects. Other statistical analyses require judgments that can impact their outcome. It is important to understand these potential impacts.
EPRI has undertaken studies to understand the nature of exposure to the various constituents of air pollution, including particulate matter and its major constituents. We have found that there appears to be a better association between personal exposure to particulate matter and outdoor measured levels than there is for many of the gaseous pollutants.
What is particularly important is that recent results suggest that
there are short periods of time (in specific environments) when personal
exposures to pollutants are much higher (by factors of 5 for PM and over 50 for
carbonaceous particles) than the levels that we measure at our monitoring
also need to
establish whether these short-term peak exposures are related to health
7.) Our joint study with Washington University of some 50,000 Veterans was designed to answer the question of whether there are long-term (chronic) effects associated with air pollution. In this study we found that after adjusting for many other factors Veterans who lived in cities with higher levels of nitrogen dioxide and very high ozone levels died earlier than those living in cleaner cities. We could find no such effect, however, when we examined particulate matter.
There is a clear association
between air pollution and health in the US at pollution levels we have
experienced in the 1990s and earlier.
pollutants We now have some limited data on the differences between personal
exposures and outdoor measurements. These differences are not the same for
every pollutant measured, leading to possible statistical impacts on the
results of the analyses of the relationships between air pollution and health.
There is as yet no accepted
biological explanation for the link between the levels of pollution found in
the US today and observed health responses. Past research has focused on epidemiological
studies—observational studies on humans going about their normal activities
. Laboratory research, which has been limited to date,
can focus on establishing the underlying biological mechanisms that can cause
negative health effects. Several
possible biological explanations have been put forth to explain the results
from epidemiological studies, and recent laboratory results support some of
these hypotheses. For example, one study appeared to show that blood clotting
can increase with exposure to higher levels of fine particulates. If this
occurs, it could be an explanation for why some heart disease effects are
related to fine particulate levels in epidemiological studies.
Particulate matter is a complex
mixture and its composition varies over time and place.
There are limited data on the
toxicity of the different components of particulate matter.
. Particle size, specifically the ultrafine fraction,
may also be important in the development of health effects. A number of studies have investigated the
effects of ultrafine particles and have found lung inflammation and other
respiratory effects, although it appears that chemical composition may play a
key role in the responses observed.
Cardiovascular and systemic effects of ultrafine particles have been
investigated to only a limited extent. Bioaerosols
are not considered to account for the reported health effects of ambient PM as
their concentrations are very low and health effects can occur at times when
bioaerosol concentrations are low.
Finally, the organic compounds associated with PM have been little-studied from a toxicological perspective, although they represent a substantial portion of the mass of ambient PM (10-60% of total dry mass). Other fractions of PM including sulfates and nitrates appear to be of less concern.
In a recent draft report, the Netherlands Aerosol Programme concluded: “Based upon current toxicological and human clinical knowledge: water, sea salt, ammonium sulfate, ammonium nitrate, and probably non-crystalline crustal material too, can be considered an inert part of PM-10 at the ambient concentrations in the Netherlands.” This report has not yet been finalized, and the conclusions are still under discussion.
In order to more fully understand
which components of PM are responsible for the health effects observed,
additional toxicological studies must be conducted. Studies
, which examine the toxicity of
emissions from various sources of pollution , can be informative
in identifying those pollutants (and sources) most highly associated with
The EPRI ARIES study was
designed to examine the toxicity of the various components of PM and air
pollution. This study is unique in terms of the number of air quality
constituents measured and the number of health effects examined. The best way to increase our understanding of
the types of PM and air pollution that may be responsible for the health
effects observed in other studies is to undertake a study in which all of the
potentially relevant fractions of PM are measured
. Traditionally we only measure what is required because of
local, state or federal regulations. On occasion a research study may measure a
larger array of air pollutants, but it is rare to have a large number of
constituents measured systematically over an extended period of time. ARIES addresses this need through detailed
air quality characterization for a period of over 2 years and
through undertaking several epidemiological studies to relate air quality
characteristics to health effects.
Appendix A provides further details about ARIES , as well as provisional
Extensive daily - and in some cases
continuous - measurements were made for all of the particle size fractions and
constituents about which concerns have been raised. At the same time several epidemiological studies were
undertaken to examine the potential health effects of the various constituents.
Initial results from the analytical team focused
upon the subset
of air pollution measures tied to the major existing hypotheses about the
pollution 2 years worth of data are now under
preparation for peer review. The draft
results are very informative, and I would like to share them with you.
These results are complex and reflect a methodology that examined pollutants individually. Analyses which consider several pollutants simultaneously are planned and may help identify the pollution components that are of greatest concern.
pollutants are statistically significantly associated with
of those over 65 years old; they include PM-2.5, PM-10, CO (carbon monoxide),
and oxygenated hydrocarbons. When
alternative statistical models were applied the results were most consistent for
oxygenated hydrocarbons, a measure that has not previously been
considered in air pollution health studies.
Results are available for several morbidity (disease) measures
including emergency room admissions to Atlanta area hospitals, unscheduled
physician visits to a health maintenance organization (HMO), and responses of
defibrillator devices implanted in patients with erratic heart rhythms. Preliminary analyses of heart rate
variability considered only PM-2.5 and not its components nor gases. Based on
these limited data, PM-2.5 was found to be associated with
statistically-significant changes in heart rate
Lung and respiratory problems were
related to PM-10 and to pollutant gases including ozone, nitrogen dioxide, and
· Heart disease responses were much more likely to be related to PM-2.5, carbon monoxide, and nitrogen dioxide.
· Organic compounds were associated with several cardiovascular effects.
· When the components of PM-2.5 were considered, the only ones found to be significant were elemental and organic carbon.
· There was little evidence of any health effects tied to acid aerosols.
· No associations were found between any health effect and total soluble metals; additional analyses are planned to look at individual metals.
associations were found with ultrafine particles
· No cardiovascular or respiratory effects were associated with sulfates.
ARIES did not look at sources of pollution directly. We did, however, undertake a source-attribution analysis of the organic compounds in Atlanta. Cardiovascular effects were found in the winter months only in this study. In the winter months organic compound concentrations were tied principally to wood smoke, although diesel emissions were also a contributor. Diesel emissions were also a major contributor to organic compounds in the summer months when no cardiovascular effects were related to these compounds.
There is a great need for additional studies which focus upon the specific components of particulate matter and examine their relationship to human health. The ARIES study will provide an important piece of evidence in understanding which fractions of PM and of air pollution are the most important in affecting human health. ARIES results are from one metropolitan area, Atlanta. Atlanta is a logical place for a study; it has high pollution levels, many sources of pollution, and no unique sources of pollution that would yield a unique result. Nevertheless it is important to undertake similar studies in other metropolitan areas. We are now engaged in similar, although more limited, studies in St. Louis and Baltimore, where detailed monitoring is underway. Much of this monitoring is funded by EPA’s supersites monitoring program. Undertaking such studies is expensive because the air quality monitoring itself is costly; hence governmental resources to undertake such studies are critical.
Secondly more laboratory studies are needed
which examine specific fractions of particulate matter and its toxicity. Since it would be very costly and time-consuming to test all
specific compounds rigorously in laboratories, special protocols should be
considered which examine the mixture of pollutants associated with specific
sources. For example, studies are now
underway at the National Environmental Respiratory Center examin
ing the toxicity
An ongoing committee of the National Research Council, of which I am a member, will issue a report next year identifying the highest priority research needs to inform particulate matter-health policy issues.
The implications of the
statistical methods used to investigate the relationship between health and air
pollution need to be fully understood.
use of one particular statistical approach and its related software . Fortuitously at a meeting
of EPRI researchers with our advisors, it was decided to use alternative
statistical methods in our research, and we have examined these methods
thoroughly. We have found that on occasion ARIES
results, especially in the mortality analyses, can be influenced by changes in
the statistical approach even when the alternative approaches are judged
reasonable by statisticians. For
example, carbon monoxide (CO) was found to be statistically significantly
associated with daily deaths of those over 65 years
old with one approach but not with the other.
Fortunately most results were similar across the various approaches, but there are
some differences . It is important
to articulate and understand these differences.
1. Air pollution likely impacts the health of individuals in the US today.
2. Particulate matter is a likely candidate to explain these impacts.
3. Not all fractions of particulate matter appear to be equally toxic.
When health effects are associated with fine particles,
our research points strongly to particles
contain carbon as the agent of concern; in most US cities,
carbon-containing particles are also the largest particle component by weight.
5. Gaseous pollutants are still of concern and cannot be ignored.
6. There is a strong need to identify with more certainty those specific components of air pollution which cause health effects.
7. We need to understand in more detail the personal exposure of susceptible individuals to the various air pollution components. In particular, we need to identify when and where peak exposures occur and whether these peaks are important to health.
8. There is a great need to apply alternative statistical methods in analyzing data and to understand the influence of a specific method.
Decreasing the non-toxic part of
particulate matter will not reduce health risks.