^A/456/R-97/001
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-456/R-97-001
September 1995
AjLr_
SUMMARY AND ANALYSIS OF
AVAILABLE AIR TOXICS
HEALTH EFFECTS DATA
FINAL DRAFT
Prepared for:
Information Transfer and
Program Integration Division
EPA Work Assignment Manager: Julie Andresen
-------
TABLE OF CONTENTS
Abstract iii
1.0 Introduction 1
2.0 Methods Used to Identify and Review Studies of Air Toxics Impacts 3
2.1 Discussions with EPA, Other Federal And State Agencies and
Academic Investigators 3
2.2 Literature Search Methodology 3
2.2.1 Search Strategies and Databases 3
2.2.2 Document Procurement 6
2.3 Procedures for Screening Articles and Documents 6
M)
4 2.4 Detailed Review Methods 7
3.0 Background 9
3.1 Approaches to Establishing a Qualitative Link Between Air
Pollution and Health Effects 9
3.1.1 Definition of Toxic Air Pollutant 11
3.1.2 General Approaches to Establishing a Qualitative
Relationship Between Air Toxics Exposures and Human
Health Effects 11
3.2 Past Efforts to Confirm the Relationship Between Air Toxics
Exposures and Adverse Health Effects 16
3.2.1 Types of Health Effects That Have Been Studied 16
3.2.2 Large-Scale Geographic Studies of Health Effects 20
3.2.3 Exposure Studies 21
3.2.4 Biomarker Studies 23
3.2.5 Risk Assessment Studies 25
4.0 Summary and Results of Study Reviews 27
4.1 Epidemiologic Studies Evaluating the Link Between Exposure to
HAPS and Adverse Health Effects 27
4.1.1 Overview of Studies Evaluated 27
4.1.2 Evaluation of Data Relating Adverse Health Effects and
Exposure to HAPs 47
4.1.3 Directions of Recent and Ongoing Studies 50
4.2 Impact of Recent Studies on Establishing the "Qualitative Relationship" 50
-------
5.0 Major Data Gaps and Opportunities for Further Research 55
5.1 Major Data Gaps 55
5.2 Potential Opportunities for Additional Research 56
6.0 References Cited 59
Appendix A. References Identified Relating to Air Toxics Exposure
and Health Effects A-l
Appendix B. Summaries and Analyses of Selected Studies B-l
-------
LIST OF TABLES
Table 2-1 EPA, Other Federal and State Air Toxics Contacts
and Academic Experts 4
Table 2-2 List of Data Bases Searched for Air Toxics-Adverse
Health Effects Information 5
Table 2-3 Format for Detailed Review of Air Toxics Health Effects Studies 8
Table 3-1 Hazardous Air Pollutants as Listed in Section 112(b) of the
Clean Air Act Amendments 12
Table 4-1 Exposures and Effects Summary for Studies Investigating Health
Effects of Ambient Exposure to Air Toxics 28
Table 4-2 Categorization of Studies Investigating Health Effects of
Ambient Exposures to Air Toxics 40
111
-------
ABSTRACT:
The purpose of this work assignment was to identity and review recent information regarding
the qualitative relationship between exposure to toxic air pollutants and adverse human health effects.
At the direction of USEPA OAQPS staff, ICF identified and contacted experts in air pollution
epidemiology, and conducted limited computer searches to identify published literature and technical
reports related to this issue. Two literature searches were performed, the first aimed at identifying
recent (1993 to present) epidemiologic studies of air pollution adverse effects, and the second aimed
at identifying studies by key researchers in the field identified through discussions with EPA and
through the results of the first literature search. Each of the literature searches identified over 400
bibliographic citations; these lists were screened and those documents which were determined to be
the most relevant were designated for procurement.
Among the articles and reports provided by EPA and those identified in the computer
searches, over 165 were designated for procurement. In the time allocated to the project, about 125
documents were retrieved from libraries or other sources. These documents were further screened to
identify those which provided information relevant to the elucidation of the relationships between air
toxics exposures and adverse health effects. Thirty-four articles meeting this criterion were reviewed
and their results and limitations briefly summarized. Detailed summaries were prepared of twelve of
these studies which either described effects of ambient air toxics exposures which had not previously
been documented, described effects in geographic regions or populations which had not previously
been studied, or which employed novel approaches for assessing toxic air pollutant effects.
The studies which were identified included reports of adverse health effects of toxic air
pollutant exposures in both U.S. (14 studies) and foreign populations (30 studies). Cancer and
respiratory toxicity were the most commonly-studied effects; some studies addressed reproductive or
neurological endpoints, or total and cause-specific mortality. Arsenic and metals were the most
common pollutants addressed (primarily from metal smelting operations), followed by volatile organic
chemicals and polynuclear aromatic hydrocarbons (PAHs). In a few cases, specific pollutants were
not identified. Relatively few studies used both measured pollutant concentrations and observed
health outcomes. Often exposures were estimated based on modeled air concentrations or using
proxies for exposure measurements, such as residential location or duration. A number of studies
used measured or modeled pollutant exposures as input to risk assessment calculations. Some of these
risk assessment studies were reviewed, but the major focus of this report is on epidemiologic
approaches to assessing air toxics risks.
The findings of health effects related to air toxics exposures near metals smelting operations
generally confirm previously-identified relationships. Novel findings among the studies included
neurobehavioral effects and elevated levels of chromosomal aberrations in populations living near
uncontrolled hazardous waste disposal facilities, and a possible relationship between adverse
respiratory effects and exposures to volatile organic pollutants at levels below those previously
thought to be associated with such effects. In addition, several studies found relationships between
"biomarkers" (measured tissue levels of pollutants or metabolites, or physiologic or biochemical
changes due to chemicals) and emissions or exposures to toxic air pollutants. These studies included
measurements of tissue levels of mercury, cadmium, and chlorinated pesticides, proteinuria resulting
from cadmium exposure, and the quantification of DNA-pollutant chemical reaction products
(adducts) due to PAH exposures.
IV
-------
1.0 INTRODUCTION
This Draft Report summarizes work done by ICF Kaiser International for the U.S.
Environmental Protection Agency Office of Air Quality Planning and Standards, Program Review
Group, under task 2-44 of EPA contract 68-D2-0189. The report provides a summary of recent
evidence that links toxic air pollutant exposures to the occurrence of adverse human health effects.
The study supports EPA's efforts to develop information for Congress, regarding "the public health
significance of estimated remaining risks [associated with hazardous air pollution exposure]" and
"concrete examples of populations that have experienced adverse health effects as a result of exposure
to hazardous air pollutants.", which is required under the Clean Air Act, Section 112(f).
The scope of this project involved the following activities:
• Discussions with OAQPS project staff and ORD and regional EPA personnel to
identify studies of air pollution health effects and populations exposed to toxic air
pollutants;
• Discussions with ATSDR staff, experts at state and local regulatory and health
agencies, and academic institutions, for the same purpose;
• Performance of computerized literature searches focusing on recent epidemiological
studies related to air toxic health impacts and articles by key investigators in
environmental epidemiology;
• Procurement, screening, and review of articles and other references identified by
EPA, the air toxics experts, and through the computerized literature searches; and,
• Preparation of this draft report summarizing the results of the literature searches,
identifying important data gaps, and suggesting directions for further research.
In keeping with direction from OAQPS, this effort has retained a narrow focus on recent
studies of authenticated adverse human health effects of toxic air pollution exposures. Most of the
information sources reviewed are articles published in refereed journals, although a number of
technical reports prepared by or for EPA and other government agencies have also been reviewed.
Resource limitations did not allow us to independently investigate reports of disease clusters, or to
review raw data sources (disease registries, death certificates, or hospital records), except to the
extent that these data had been independently evaluated by other investigators and reported in the
published literature or technical reports.
Similarly, resource limitations did not allow us to provide detailed reviews of more than a
representative portion of the large number of studies which were identified. (A complete listing of
these studies is provided in Appendix A.) Instead, detailed reviews are provided (in Appendix B) of
only a limited number of recent epidemiologic studies which provide novel or provocative findings, or
which illustrate recently developed methods for investigating air pollutant-adverse effect relationships.
The remainder of this report is organized as follows:
Air Toxics Report 1 Final Draft
-------
Section 2 provides a listing of the air toxics experts identified and contacted, and a
detailed description of the literature search methods used to identify and obtain studies
for review. The screening criteria used to select studies for detailed review, as well
as the review procedure itself, are also described.
Section 3 provides a review of background information related to the relationship
between toxic air pollutant exposures and adverse health effects. This section explains
the various techniques that have been used to establish a link between toxic air
pollutant exposures and health effects, and provides a brief historical overview of past
efforts to establish this linkage. This general information establishes the context for
the focused discussion of the more recent studies that is provided in Section 4.
Section 4 provides a summary of the results of the literature searches, screening, and
review process. The most recent useful studies are broken down into categories
related to types of adverse effect, the pollutants examined, and the state or country
where the affected populations were located. Brief summaries are provided of a small
number of studies, and general appraisal of other relevant studies is also given.
Section 5 discusses the data gaps identified in our literature survey, and identifies a
number of areas for further research to address the data gaps.
Air Toxics Report 2 Final Draft
-------
2.0 METHODS USED TO IDENTIFY AND REVIEW STUDIES OF AIR TOXIC IMPACTS
This section provides a brief summary of the procedures which were used to identify
potentially relevant studies, to screen them, and to review them and summarize their results.
2.1 Discussions with EPA, Other Federal and State Agencies, and Academic Investigators
As noted in Section 1, many discussions of potentially useful studies were carried out with the
EPA WAM and technical staff, and with experts from other EPA offices, other regulatory agencies,
and academic experts in air toxics and epidemiology. In addition to the individuals identified through
discussions with OAQPS, the literature searches discussed below also identified individuals actively
working in the air toxics field.
EPA OAQPS staff provided a number of documents for review, as well as assistance in
identifying and following up with experts inside and outside the agency. Table 2-1 provides a list of
the individuals identified by ICF and OAQPS, indicating those who were successfully contacted and
from whom information concerning the relationship between air toxics and adverse effects was
obtained.
Contacts with experts were made both by ICF and OAQPS staff. Detailed interviews were
not conducted with everyone on this list. In some instances, only exchanges of telephone messages
were possible within the time frame of the study. When individuals were contacted, they were asked
both for general insights and for specific studies concerning the relationship between air toxics and
health in their geographical area or area of technical expertise.
2.2 Literature Search Methodology
In addition to the sources identified by the experts, several computer literature searches were
also carried out to identify further studies for review. OAQPS provided the results of a February
1995 DIALOG literature survey on air toxics issues which served as a starting point for ICF's efforts.
2.2.1 Search Strategies and Data Bases Searched
Two distinct search strategies were employed. The first strategy was narrowly focused on the
recent epidemiologic literature (1993 and later). The general intent of this search was to identify
quite recent studies of air pollution effects that had not previously been reviewed by EPA. The focus
was on studies which provided both exposure and effects data, although the initial search was broadly
based, using a large number of key words to assure good coverage. A mixture of U.S. and a small
number of major foreign data bases was searched, as listed in Table 2-2. Initial broad screening
searches were performed by the ICF Information Services Group, and lists of titles were provided to
technical staff for review. Based on this review, searches were fine-tuned, a final list of data bases
was developed, and the final search was completed, providing full bibliographic information.
The second search strategy was intended to broaden the range of documents identified in the
initial search, by looking further back in time. In addition, the search was keyed to the names of
epidemiologists and scientists identified by EPA and ICF as being active in the field of air
Air Toxics Report 3 Final Draft
-------
TABLE 2-1 EPA, OTHER FEDERAL AND STATE AIR TOXICS CONTACTS AND
ACADEMIC EXPERTS
USEPA-OAQPS
Julie Andresen( 1,2)
Eric Ginsberg(l)
Margaret Round(l)
Dennis Pagano(l,2)
USEPA Other Offices
Ila Cote (ORD HERL)(1,2)
Joellyn Lewtas (ORD HERL)
Robert Fegley (ORD)
John Vandenberg (ORD HERL)(1)
Jane Metcalfe (ORD)
David Otto (HERL)(1)
USEPA Regional Offices
Linda Anderson-Carnahan (Region IV)
Bryan Bateman (Region IX) (1)
CDC
Godfrey Oakley (Analytical Epidemiology Division) (1)
ATSDR
Rick Canady (1)
State Regulatory Agencies
Joann Held NJDEQE (1)
Bliss Higgins LADEQ (1)
Academic Air Toxics Experts
Carl Shy (UNC SPH)
John Speizer (Harvard SPH)
Notes:
(1) Indicates individuals who provided studies or other information
(2) Reviewers of Draft of this Report
Air Toxics Report 4 Final Draft
-------
Table 2-2. LIST OF DATA BASES SEARCHED FOR AIR TOXICS-ADVERSE HEALTH
EFFECTS INFORMATION
MEDLINE
Toxline
Cancerlit
Pascal
EMBASE
Env.Bib
Pollution Abs
Enviroline
NTIS
Energy SciTec
Ei Compendex*Plus
D-S/IAPV/INCIDENCE & PREVALENCE (IPD)
Air Toxics Report 5 Final Draft
-------
pollution adverse effects and epidemiology. The same databases were searched using this strategy as
in the first search, omitting a few foreign sources which had not been productive.
2.2.2 Document Procurement
As noted previously, EPA OAQPS and other offices provided a number of documents to ICF
for review. In addition, both of the computer literature searches identified very large numbers of
documents (greater than 400 in each case) potentially related to toxic air pollution adverse effects.
The results of the searches (study titles and abstracts) were reviewed by ICF technical staff and the
studies which appeared to be relevant were identified for procurement. Articles were obtained from
the ICF library, through interlibrary loans, or by sending runners to nearby libraries. A few
government technical reports were ordered from NTIS. A complete list of documents which were
identified for procurement is provided in Appendix A. This list also identifies articles that had been
procured at the time this Draft Report was written. At that time, all but approximately 40 of over
165 requested articles and reports had been obtained.
2.3 Procedures for Screening Articles and Documents
When articles and other documents were obtained, each was screened by ICF staff to
determine the level of review it would receive. After scanning the abstract and text of the article, we
placed them into the following general categories:
Not Relevant. Studies in this category did not contain any significant information bearing on
the relationship between air toxics exposures and adverse effects. Studies which fell into this category
included those dealing with, for example, the environmental epidemiology of a specific adverse effect
which did not address its relationship to air toxics exposures, or studies which were simply
descriptions of air toxics exposure measuring methods, etc. These studies received no further review.
Background Studies. The studies which fell into this category were generally those which
provided general background information on the relationship between air toxics and health effects, but
which did not provide original exposure and effects data for any specific population. As discussed in
Section 3, studies in this category included general historical reviews, pure methodology discussions,
disease mapping studies which did not directly link observed patterns to air toxics, exposure
evaluations, some biomarkers studies (if they were not accompanied by exposure or effects data), and
risk assessments for air toxics exposures. These studies were reviewed in varying levels of detail
(depending on their level of relevance), and provided reference material for the background discussion
in Section 3.
Exposure-Effects Studies. Any study which provided information on both air toxics
exposures and the occurrence of adverse effects in a specific population was included in this category.
To "make the cut", the studies had to address specified "air toxics", as defined in Section 3.1.
Studies of "criteria" pollutants (other than lead, and those studies which identified specific paniculate
toxic pollutants), and studies which did not identify specific pollutants also went into the
"background" category. As noted in Section 2.1, the bias in identifying these studies was focused on
recent articles, so some older epidemiologic studies with exposure and effects data were also used as
background.
Air Toxics Report 6 Final Draft
-------
All recent exposure effects studies which we identified and obtained (about 40) were read and
reviewed by ICF. The results of this review are summarized in Sections 4.1 and 4.2. A subset of
these studies was selected for especially thorough review, and detailed-two page summaries (see
Appendix B) were prepared for each of these studies. The criteria used to select studies for
summarization included:
• The study provided information on a source category, pollutant, or effect not
previously characterized;
• The study provided information on toxic air pollution adverse effects in a population
or region not previously studied;
• The study employed novel or emerging methods for characterizing air pollutant-
adverse effect relationships, or, in a few cases;
• The study illustrated key limitations and problems in evaluating the adverse effects of
toxic air pollutants.
Limited reviews were performed of the remainder of the exposure-effects studies. Section 4.1
presents a tabular description of these studies, and a brief overview of their major findings.
2.4 Detailed Review Methods
The reviews of selected studies involved the development of detailed itemized summaries by
senior technical personnel. A standardized format for the detailed summaries was developed, and is
shown in Table 2-3. The areas addressed in the summaries include the identities of the pollutants
studied, the exposure measures which were used, the study methods, the nature of the observed
effects, and the overall study conclusions. At the end of each summary, the reviewers also provide
comments on the methods, quality, and conclusions of the study. Striking findings in key studies are
often the subject of debate or criticism by other workers in the field, and to the extent that the studies
provoke comments or technical criticism, these are also addressed in this final summary section.
Air Toxics Report 7 Final Draft
-------
TABLE 2-3. FORMAT FOR DETAILED REVIEW OF AIR TOXICS HEALTH EFFECTS
STUDDSS
STUDY SUMMARY: (author, veart
CITATION:
SOURCES OF EXPOSURE;
HEALTH EFFECTS EVALUATED;
CHEMICALS EXPOSED TO:
EXPOSURE MEASURES:
CONFOUNDING FACTORS ASSESSED;
STUDY METHODS:
STUDY RESULTS;
AUTHOR'S CONCLUSIONS;
COMMENTS;
Air Toxics Report 8 Final Draft
-------
3.0 BACKGROUND
This section presents a brief discussion of the technical issues associated with establishing a
qualitative relationship between exposure to toxic air pollutants and the development of adverse effects
in humans. In section 3.1, historical background on the earliest efforts to establish air pollutant-
adverse effects relationships is provided, and basic definitional issues are addressed (what is a "toxic
air pollutant" for purposes of this assessment?). In addition, the major scientific approaches to
evaluating the relationship between air toxics exposure and health effects (geographic pattern analysis,
exposure/biomarker studies, risk assessment, and formal epidemiological analysis) are described.
Section 3.2 provides a summary discussion of modern efforts to evaluate the effects of air toxics on
human health, identifying the types of effects which have been evaluated, and placing in context past
efforts at geographical pattern analysis, exposure and biomarker studies, and risk analysis. This
discussion sets the stage for the treatment of the recent epidemiological studies of the relationship
between air toxics exposure and adverse health effects which is presented in Section 4.
3.1 Approaches to Establishing a Qualitative Link Between Toxic Air Pollution and Health
Effects.
The suspicion that man-made air pollution contributes to poor health is at least 300 years
old1. However, the systematic investigation of the relationship between air pollution and ill health
has only been undertaken in this century, primarily since 1950. The earliest efforts were focused on
characterizing the relationship between obvious and acute effects (respiratory irritation, exacerbation
of asthma, other respiratory and cardiovascular disease and death) and short-duration incidents ("air
pollution episodes") of high exposures to combustion products, primarily coal smoke2. In air
pollution episodes, the fact of high exposure was easy to establish (if only qualitatively), the spatial
and temporal relationships between exposures and adverse effects were clear, and the nature of the
adverse effects (serious incapacitation and mortality, in some cases) were easy to detect and quantify.
Because of the clear relationship between exposures and effects, early attempts at controlling air
pollution exposures were aimed primarily at these products of combustion, which in the U.S. are
known for regulatory purposes as "criteria pollutants"3. With the exception of lead4, the criteria
pollutants have the common characteristics that their major, if not exclusive, adverse effects at near-
normal ambient exposure concentrations are respiratory or ocular irritation, or exacerbation of
1 Suggestions of adverse effects air pollution associated with air pollution episodes in London were made as
early as the 17th century.
2 For example, an air pollution incident in Donora, PA, in 1948 resulted in an estimated 350 deaths from
exposures to very high levels of mixed particulate, SO2, CO, and other combustion products. Similar episodes
of acute mortality in London, England (in 1952) and the Meuse Valley in Belgium (1930) have been recorded in
this century.
3 The criteria pollutants defined in the Clean Air Act are inhalable particulates (PM10), S02, nitrogen oxides
(NOX), carbon monoxide (CO), and lead.
4 Lead is a criteria pollutant; "lead compounds" are HAPs. Thus lead falls on the borderline between the
criteria and toxic pollutants. Studies of the adverse effects of lead exposure on human health are not included
as a major focus of this study, however, because the relationship between lead exposures and adverse effects is
so well-established, and because the literature on lead adverse effects is so large.
-------
cardiovascular stress, and the technical basis for regulating these pollutants is primarily related to
these effects5.
The relationship between exposures to "toxic" air pollutants and adverse health effects is not
so clear. Except for a few well-known cases (the sudden release of a large volume of methyl
isocyanate in Bhopal, India, for example) where extremely high exposures to accidental releases of
industrial chemicals resulted in severe acute health effects, the affects of adverse exposures to
industrial chemicals are generally much harder to detect, for a number of reasons.
First, except in cases of accidental releases, exposure concentrations of toxic air pollutants in
outdoor ambient air are generally much lower than exposure levels to the criteria pollutants in similar
settings, since the criteria pollutants tend to be emitted in much larger amounts, at least in urban area.
With the exception of those chemicals that have specific irritant properties, acute systemic effects such
as are seen in occupational settings are less likely to be detected in the general population exposed to
normally-occurring levels of "toxic" air pollutants, although as will be discussed later, at least one of
the studies we reviewed shows evidence of adverse effects of toxics exposures (perhaps in especially
sensitive individuals) at concentrations far below those previously thought to be hazardous. Often, the
effects of the usually low chronic exposures to toxic air pollutants may be subtle, and may develop
slowly over time in response to cumulative exposures (chronic effects), or may not develop until long
after exposures occur Gatent effects). The effects of exposure may include the same kinds of
respiratory effects seen with exposure to criteria pollutants, or they may include effects on other
organ systems in the body which are not so easily linked to inhalation exposures.
Thus it is not easy to directly characterize the risks associated with general population
exposures to toxic air pollutants under conditions of chronic low-level exposures. Nonetheless, there
is a high level of suspicion that such effects are occurring. The bases for this suspicion include the
fact that many toxic air pollutants are suspected or known human carcinogens and even low levels of
exposure could theoretically cause increased cancer risks. In a smaller number of cases, animal or
controlled human studies indicate that noncarcinogenic effects might be expected to occur at exposures
near ambient levels. In some instances, allergic sensitization may results in adverse effects in a small,
especially sensitive subset of the exposed population. In addition, some lexicologists have also put
forward theories that very low-level pollutant exposures (much lower than the no-effect levels
indicated by laboratory studies) may cause subtle adverse effects by mechanisms which are not yet
understood6.
A final issue is the often-observed coincidence between exposures to criteria pollutants and
exposures to toxic air pollutants. This phenomenon arises because both types of pollution are
associated with areas of high population density and industrial development. As discussed below,
many epidemiologic studies simply use measures of one or other of the criteria pollutants
5 Some criteria pollutants may cause chronic systemic effects in humans, however. Particulate pollutant
exposures, SO,, and acid aerosols have been linked to elevated cancer risks and other cause-specific mortality in
some instances, and exposures to lead are regulated primarily based on its neurobehavioral effects.
6 An example of such effects would include "multiple chemical sensitivity syndrome." The nature and
causal mechanism of this syndrome has not been identified.
10
-------
(participates, usually) as the sole measure of exposure, and use it as a proxy for all "air pollution".
Couple this with the fact that many of the toxic air pollutants (metals and low-volatility organics) are
often present as part of paniculate air pollution (especially near industrial point sources), and the issue
of adequate exposure characterization becomes a key to evaluating potential relationships between air
pollution and adverse health effects. The following sections discuss these various issues.
3.1.1 Definition of Toxic Air Pollutant
The primary focus of this report has been on locating and reviewing studies of exposures and
adverse effects of substances which Congress has designated as "Hazardous Air Pollutants" (HAPs).
The definition of "Hazardous Air Pollutant" (Clean Air Act Section 112(c)) includes 189 chemical
compounds and families of compounds (Table 3.1). On the list are a wide range of organic
compounds, the compounds of twelve metallic elements and metalloids, "all radionuclides" (including
radon), and "fine mineral fibers". Included also are the broad categories of "coke oven emissions"
and "polycyclic organic matter" (POMs), both of which include many high-molecular weight organic
compounds, such as polycyclic aromatic hydrocarbons (PAHs).
This is an extremely inclusive list, covering the vast majority of the mass (greater than 99
percent) of all of the chemicals identified through Toxics Release Inventory (TRI) reporting as being
released to air from industrial facilities in 1992. As part of this effort, we performed chemical-
specific literature searches on the top 15 chemicals from the TRI list, but not on the other HAPs. No
studies of any kind were found for most of the chemicals on the list; in fact, the great majority of the
useful studies were limited to a relatively few (less than ten) pollutants, mostly metals. (See Section
4.) We did not specifically exclude studies of chemicals not on the list (except for criteria pollutants),
but we identified no useful epidemiologic studies of non-occupational populations which related
human health effects to any toxic chemicals which were not HAPs.
3.1.2 General Approaches to Establishing a Qualitative Relationship Between Air
Toxics Exposures and Human Health Effects
At some level of certainty, no one seriously questions the fact that there is a relationship
between exposures to toxic air pollutants and human health. The key questions are: At what levels of
exposure do these affects occur, and what is their nature and magnitude? Thus, studies which attempt
to establish a qualitative link between air toxics and adverse effects will inevitably involve, to some
extent, quantitative measurement or estimation of both levels of exposure and the magnitude of
effects. The level of certainty with which it can be ascertained that an effect is, in fact, due to an
exposure is closely tied to the extent to which the relationship between exposures and effects can be
established quantitatively (often through statistical hypothesis testing). Thus the qualitative and
quantitative relationship between air toxics exposures and health effects are inextricably linked.
Table 3-1 HAZARDOUS AIR POLLUTANTS AS LISTED IN SECTION 112(b) OF THE
CLEAN AIR ACT
11
-------
Table 3-1 (continued)
12
-------
Table 3-1 (continued)
13
-------
Table 3-1 (continued)
14
-------
There are five general types of approach that are employed to establish a relationship between
air toxic exposures and health effects:
Studies of Laboratory or Occupational Exposures. Adverse health effects of exposures to
specific pollutants are often first discovered through observations of adverse effects in
workers exposed to high levels of the pollutants, or in laboratory studies of animal or human
toxicity. These studies do not directly address the potential for adverse effects occurring in
the general population at lower exposure levels.
Studies of Geographic Patterns of Disease Incidence or Mortality. It is often the case that
studies of vital statistics, disease incidence, or mortality disclose geographic patterns of
adverse health effects that are suggestive of a relationship to specific pollutants or pollutant
sources (see section 3.2.2). Unfortunately, If such studies are not supplemented by exposure
data, and are not controlled for confounding factors other than pollutant exposures, it is not
possible to support inferences of causation associated with pollutant exposures.
Studies of General Population Exposures. Exposure Indices, and Biomarkers. There have
been a number of environmental monitoring studies and data gathering exercises that have
been used to estimate human exposures to pollutants and draw inferences about potential
adverse effects (see Section 3.2.4). This information is often used, in conjunction with
toxicity data, to conduct risk assessments. In some instances, measurable indices of exposures
(biomarkers of exposures), such as body burdens or tissue concentrations of pollutants, can be
used to document exposures and evaluate the potential for adverse effects. Other biomarkers
(biomarkers of effects) can include altered enzyme or metabolite levels, DNA-chemical
adducts, or chromosome abnormalities. Not all biomarkers are indicative of clinical effects
(NAS/NRC 1989).
Risk Assessments. In a risk assessment, information about exposures (which may reflect
actual measured exposures or exposures estimated using emissions and environmental models)
is combined with toxicity information (from occupational or laboratory studies) to develop
predictive estimates of the frequency or severity of occurrence of adverse effects in human
populations. Risk assessment is a commonly used approach to evaluate the potential for
adverse effects from air toxics exposure, but risk assessments may be subject to a high degree
of uncertainty due to imprecision in exposure estimates, and uncertainties in dose-response
information, especially at low doses, as discussed in section 3.2.5.
Formal Epidemiological Investigations. A "formal" epidemiological study involves systematic
investigation of the relationship between an observed pattern of adverse health effects and
exposures to one or more agents. The analysis of actual (as opposed to estimated) health
outcome information is what distinguishes an epidemiological study from a risk assessment or
a biomarkers study. Systematic efforts to control for confounding factors (factors other than
toxics exposures which may be responsible for the observed effects) are what distinguish a
formal ("analytical") epidemiologic study from a simple "descriptive" summary of geographic
patterns of disease incidence. Recent epidemiological studies are the primary focus of our
data gathering and review efforts, since they would provide the strongest direct evidence of
air toxics-adverse effect associations. These studies are discussed in detail in Section 4.
15
-------
Formal epidemiological studies, are the "gold standard" in establishing a relationship between
air toxics exposures and adverse health effects. If adverse effects are unambiguously observed to
occur in a distinctive pattern (in time or space, or in a given population), and this pattern can be
shown to be significantly associated with estimated (or, preferentially, with measured) exposure
patterns, and if other potential causes for the observed pattern in affects can be controlled for, then a
strong inference can be drawn that the observed effect is associated with the agent in question.
Support for a relationship between exposures and effect is further strengthened if the effect is
consistent with the known toxic properties of the agent in question. Unfortunately, few studies
measure up to this "gold standard". As will be seen in the following sections, many suggestive
indications of possible relationships of air toxics exposures to health effects derived from geographic
studies, exposure patterns, or risk assessment cannot be followed up due (most commonly) to a lack
of comprehensive or reliable health outcome data, or due to a lack of exposure information. Many
other suggestive findings from geographic studies do not stand up to detailed investigation; factors are
found other than air toxics exposure which can adequately explain the observed patterns of health
effects.
EPA has been investigating the relationship between air toxics and adverse health effects for
many years, using all of the methodologies mentioned above. As will be discussed in more detail
below, a major controversy currently surrounding the issue of air toxics health effects is the apparent
inconsistency arising from the relatively high estimates of chronic health impacts which have been
derived from exposure studies and risk assessments, compared to the relatively limited success that
has been achieved in confirming the existence of these relationships through epidemiologic studies of
exposed populations. This controversy has helped to focus our data gathering effort on the
identification of recent epidemiological studies which can directly confirm the relationship between air
toxics and adverse health effects. Our literature review also identified a number of risk assessment
studies, but our major focus has been on epidemiological studies.
3.2 Past Efforts to Confirm the Relationship Between Air Toxics Exposures and Adverse
Health Effect
In this section we provide a brief summary of major past efforts at characterizing the
relationship between exposure to toxic air pollutants and adverse health effects. Included in this
section are a discussion of the types of health effects which have been studied in conjunction with air
toxics exposures, and summaries of how the current state of understanding of the major issues in the
area has evolved over time in response to key findings from geographic studies, exposure studies, risk
assessments, and epidemiologic investigations. The material in this section is intended to provide a
brief historical overview, rather than a comprehensive treatment of the subject, and focuses on
developments through approximately 1993; more recent studies are discussed in more detail in Section
4.
3.2.1 Types of Health Effects That Have Been Studied
Attempts to establish associations between air toxics exposures and health effects have
focused on four major types of health effects: cancer, respiratory irritation and other respiratory
16
-------
toxicity, neurobehavioral toxicity, and reproductive and developmental effects. Other effects are also
occasionally studied, but much less frequently. Some of the toxic metals represent special cases, each
having their own unique pattern of noncancer effects. The renal effects of cadmium exposures,
neurodevelopmental impacts of lead, and the reproductive toxicity of mercury exposures are the most
well-studied examples. In addition, a few studies use total mortality, or cause-specific mortality, as
endpoints.
Among all of the data sources which we have examined, particularly the earlier studies,
cancer is by far the most frequently-studied endpoint. Epidemiologic studies of air toxics have
focused on cancer for two major reasons. First, there are established and easily accessible databases
of cancer mortality and, to a lesser extent, incidence, at national and regional levels. In the U.S., the
National Cancer Institute (NCI) has assembled and maintains a database of cancer mortality
information at the county level starting from 1950, based on information provided by the National
Center for Health Statistics (Fraumeni 1987). NCI also maintains a cancer incidence database
covering a statistically representative weighted sample of about 12 percent of the U.S. population for
the same period. Many states maintain similar cancer registries, as do many foreign countries.
Another reason why cancer appears frequently as an epidemiological endpoint is its high mortality
rate, so that studies can be based on analyses of death certificate files, which are also maintained by
most states.
The second reason that cancer is a major focus of air toxics studies is that many toxic air
pollutants, as noted previously, are suspect or confirmed human carcinogens7. Laboratory and
occupational toxicity data suggest that many high-volume air toxics (metals, PAHs, benzene, some of
the chlorinated solvents and organic monomers) can cause cancer at ambient exposures levels. This
has been borne out by the results of a number of epidemiologic studies of populations exposed to
metals and PAH air pollution from metal smelters and combustion sources. These carcinogenic
pollutants also are convenient subjects for environmental studies because they are persistent in air and
soil-water systems, and exposures can thus can be readily measured.
Respiratory toxicity8 is a natural subject for epidemiologic investigation of airborne
pollutants, since inhalation is the major route of exposure for these pollutants, and the respiratory
tract is the initial site of pollutant contact and absorption. Also, inhalation endpoints have been the
traditional subject of large-scale studies of the adverse effects of air toxics exposures to the criteria
pollutants. Unlike cancer, however, there is no uniform national database or register of respiratory
diseases, although national and state-level statistics on the incidence of some of the major noncancer
respiratory diseases (asthma, emphysema, etc.) are available. This is partially because of the
multiplicity of respiratory diseases, the difficulty of diagnosing them without physical examination and
7 In EPA's proposed rulemaking under Section 112(g) of the Clean Air Act, 120 of the 189 HAPs were
classified as known, probable, or possible human carcinogens (59 Federal Register 15504 April 1, 1994)
8 Folinsbee (1993) identifies the confirmed respiratory adverse effects of air pollution exposure as,
"responses ranging from reversible changes in respiratory symptoms and lung function, changes in airway
reactivity, and inflammation, structural remodeling of pulmonary airways, and impairment of pulmonary host
defenses, to increased respiratory morbidity and mortality."
17
-------
laboratory testing, and the wide range in severity of many of the disorders thought to be associated
with exposure to toxic air pollutants.
Indications of the potential magnitude of adverse effects of toxic air pollutants can be found in
numerous studies of the relationship between paniculate exposures and acute and chronic health
effects. In many of these studies, (usually where there is no clearly-identified dominant source of
paniculate air pollutants) it is not known which chemical constituents of particulates contribute to the
observed increases in risk, and it is therefore difficult to attribute any given fraction of these effects to
toxic air pollutants, as they are defined in this report. While toxic metals and organic pollutants may
contribute significantly to these effects, acid aerosols and acid salts (which are not HAPs) bound to
the particulates are thought to be major causative agents in the observed effects in many of the more
recent studies.9 Because of the known heterogeneity in paniculate composition, studies of
"paniculate" air pollutants which do not identify specific toxic agents have not been a major focus of
our effort. In addition to metals and PAHs, exposures to volatile organic HAPs have been implicated
in a number of studies of adverse health effects10.
As with respiratory effects, there are no national or regional registries of neurological disease.
Also, the characteristics of many neurological disorders dictate the use of sophisticated testing and/or
physiological monitoring methods to confirm diagnosis or evaluate the severity of impairments (Shy
1993). Thus, until recently, there have been relatively few studies of the association between
neurological symptoms and exposure to toxic air pollutants. Exceptions include studies of Parkinson's
disease, whose geographical patterns of mortality have been found to be correlated with specific
industrial activities, but not specifically with air pollution (Rybicki et al 1993).
The existing literature on neurobehavioral effects of toxic air pollutants is dominated by
discussions of the adverse effects of lead (which is a toxic criteria pollutant, as well as a "HAP") on
intellectual and behavioral indices in children. These studies generally describe decrements in
performance as a function of biomarkers of lead exposure, such as blood lead concentrations or heme
metabolite levels. An important limitation of many of these studies is that there is often little
information on the sources of lead exposures, and lead from deteriorating paint and in pipes and
solders used for drinking water distribution can contribute significantly to total exposures. As noted
above, one of the important characteristics of inorganic lead compounds is their persistence in the
environment, which results in multi-pathway exposures to particulate materials deposited onto surface
soil. Even this material could be counted as an impact of toxic air pollution, since the ultimate source
of exposure was deposited air particulate.
Information related to the relationship of organic toxic air pollutant exposures to neurological
effects is quite limited. Much of the effort in this area has been directed at characterizing the etiology
9 Recent reviews on this subject include Dockery and Pope (1993) on the acute effects of particulates,
Folinsbee (1993) for a general overview of particulate health impacts, and Reichardt (1995) for an informal
review of the health impacts of very fine particulates
10 An example is the Kanawha Valley study (Ware et al 1993), reviewed in Section 4. Other instances
where volatile organics have been implicated in the epidemiologic studies of respiratory toxicity include studies
of populations near hazardous waste disposal facilities. Several of these studies are also discussed on Section 4.
18
-------
of the physical and behavioral symptoms of "sick building syndrome" (Otto et al 1992), and
investigating neurological impacts of exposures near toxic waste disposal sites. Some of these latter
studies are discussed in Section 4.
The potential reproductive and developmental11 impacts of exposure to air toxics have also
been explored in a number of studies. The principal sources of data concerning the incidence and
nature of reproductive effects include "state birth defects registries,....birth certificates, hospital and
clinic records" (Shy 1993). There is, however, currently no comprehensive national registry of birth
defects and other reproductive outcomes, and only about half of the states have birth defects
registries12. Compared, however, to neurological disorders, data concerning major birth defects are
relatively easy to obtain.
Epidemiologic studies of toxic air pollution impacts on reproduction in the U.S. have focused
on a very few specific agents. Among these is vinyl chloride, whose teratogenic activity in animals
prompted at least two epidemiological investigations of populations living near PVC manufacturing
plants13. Another major toxic pollutant whose exposures have been strongly linked to adverse
reproductive outcomes in mercury. The preponderance of the data concerning mercury is related to
exposures to wastewater discharges rather than to mercury emitted into the air, however.
Reproductive endpoints have also been investigated in populations residing near a number of
hazardous wastes sites contaminated with volatile organic chemicals (Geschwind et al 1992). In some
of these studies, air monitoring results support the possible relationship between inhalation exposures
and the observed effects.
A few other endpoints have been used in attempts to link toxic air pollution to adverse health
effects. One of the most important has been total and cause-specific mortality other than cancer.
Many early studies of paniculate air pollution impacts were based on developing statistical
relationships between geographical exposure patterns and mortality. One of the most recent studies
using this approach is the "Six Cities Study" (Dockery et al. 1993), although similar studies have
been conducted of single cities in the U.S. and smaller geographical regions14, and in eastern Europe
11 Note that reproductive and developmental toxicity are distinct types of adverse effect. They have been
lumped together in this discussion in keeping with EPA's risk assessment guidelines.
i; The Centers for Disease Control has operated the Birth Defects Monitoring Program (BDMP) since the
1970s. This registry of birth outcome data currently covers only about 12 percent of the live births in the U.S.,
and CDC is working with state registry systems to improve coverage (G. Oakley, CDC Analytical
Epidemiology Division, telephone interview).
13 Neither Edmonds et al. (1978), nor Rosenman et al. (1989) found evidence for an association between
vinyl chloride emissions and the incidence of central nervous system malformations. Theriault et al. (1983),
however found a consistent, but not statistically significant relationship between CNS birth defects and distance
of residence from a PVC plant in New Jersey.
14 Kaldor and Harris (1984), for example examined the relationship between air toxics exposure and major-
cause mortality in an industrialized area near San Francisco. This study is discussed in more detail in Section
4, Schwartz and Dockery (1992) reported on the relationship between daily pollution levels and acute mortality
in Philadelphia.
19
-------
and the ex-Soviet Union15. Again, it is important to note that some or all of the effects of mortality
seen in these studies may be associated with criteria air pollutants, rather than toxic air pollutants.
Finally, a number of studies which we identified have attempted to establish a relationship
between exposure to toxic pollutant and cell mutations or chromosomal abnormalities (Klemens et al.
1995, Lakhanisky et al. 1993, Laurent et al. 1993). There have also been efforts to correlate air
monitoring results for PAHs and levels of PAH-DNA adducts (biochemical reaction products of the
PAH with DNA) in exposed populations (Binkova et al. 1995). These studies represent "biomarkers
of effects", although the ultimates physiological and health significance of these endpoints is not clear.
3.2.2 Large-Scale Geographic Studies of Health Effects
Among the earliest studies in the attempt to link environmental factors (including toxic air
pollution) to adverse health effects were a number of large-scale mapping studies of cancer mortality
and incidence data in the United States. As noted above, these analyses were facilitated by the
availability of national-level cancer mortality data which was evaluated and summarized by NCI
starting in the early 1970s. This analysis culminated in the publication of a national atlas of county
levels of organ-specific cancer mortality data.16 Based on the maps produced during this efforts,
NCI conducted a large number of epidemiological analyses of regional cancer hot-spots throughout
the U.S.17 These studies confirmed the existence of a gradient of increased cancer risk in cities
compared to rural areas, associated primarily with differences in smoking habits and occupational
chemical exposures. In addition, connections between observed cancer hot-spots were found to
employment in World War II shipyards (associated with asbestos), to the textile and chemical
industries, and to occupational exposures to pesticides in certain farming regions. Links to ethnicity,
dietary habits, nutritional status, and to alcohol consumption were also found for a number of groups.
Two possible links to air pollution were also found in these followup studies; one was an
increased incidence of lung cancer in residents of counties where metal smelting operations were
located (Blot and Fraumeni 1975), and increased cancer incidence in counties (mostly in the gulf coast
and New Jersey) with high concentrations of petrochemical and chemical industries. In neither case
was a direct association with air toxics exposure established (cancer patterns on the gulf coast appear
to be correlated with drinking water contamination, however), but the existence of such a linkage
remains plausible. These findings spurred a number of studies of the relationship between residence
near individual metal smelting operations and chemical plants and adverse health effects. Some of
these studies are discussed in more detail in Section 4.
Similar large-scale mapping studies of geographic variations in cancer incidence have been
conducted in Europe18. These studies generally confirm the urban-rural pattern of cancer incidence,
13 For a review, see Jedrychowski (1995)
16 Pickle et al. (1987) provide county-level maps of cancer mortality for the period 1950-1980.
17 For a summary of these studies, see Fraumeni (1987).
18 For reviews, see Doll (1992), Tomatis (1991), and Hemminki and Pershagen (1995).
20
-------
the strong dependence of cancer mortality on smoking habits and occupation, and suggested testable
hypotheses regarding the relationship between toxic air pollution from industrial sources. A number
of subsequent epidemiological investigations have been undertaken of the contribution of emissions
from specific sources to cancer risks, as discussed in Section 4.
As noted previously, there are no readily accessible published data sources which provide
comprehensive information on geographic variations in the incidence or mortality from diseases other
than cancer, at the same level of detail as provided in the NCI cancer atlases19. Therefore, there has
been only limited large-scale analysis of the geographic patterns of diseases other than cancer. One
data source that has been used in the past is the Birth Defects Monitoring Program, which was an
effort funded by the Centers for Disease Control in the 1970s, and which involved the collection of
incidence data for major birth defects at a number of participating hospitals throughout the U.S. In
recent years, the coverage of this program has declined so that is now no longer representative of
overall U.S. patterns in birth defects20. In 1992, 25 states had active birth defects monitoring
programs, which varied in degree of coverage and methods of data collection and organization
(Greenwood 1993). Some states, including New Jersey, Louisiana, California, and Washington, have
developed maps of birth defects incidence, and performed analyses of possible relationships between
geographic patterns of birth defects and industrial facilities or hazardous waste disposal sites. Some
of these studies are discussed in Section 4.
During the course of this project, we did not identify any other comprehensive national
resources on geographic patterns of disease incidence which could be easily adapted for use in
establishing the connection between toxic air pollution and cancer, nor did we review any mapping or
geographic studies of other diseases. In theory, geographic patterns could be mapped and studied by
using a sample of hospital admissions or discharge data, and many epidemiologic studies of potential
air pollution affects have been conducted based on hospital data. We did not investigate the potential
utility of this data source further.
3.2.3 Exposure Studies
A major limitation on the investigation of the relationship between air toxics exposure and
adverse health effects is the shortage of air monitoring data for most toxic pollutants. In contrast to
the situation for criteria pollutants, there is no national monitoring system for air toxics which can
provide even general information on the urban and rural concentration patterns of these pollutants in
ambient air. Epidemiological studies which seek to establish an association between air toxics
exposure and health effects must therefore either include a monitoring program, find a local
monitoring program which can provide data on the exposures of interest, or use some proxy for
measured exposures. Such proxies may include modeled pollutant levels, pollutant emissions, or, as
noted above, residential distance from an air pollution source. EPA and other organizations have
19 The National Center for Health Statistics maintain age-, sex-, and cause-specific mortality records at the
county level for the entire U.S. As noted by Greenwood (1993), a number of academic institutions, including
the University of North Carolina School of Public Health, are actively pursuing investigations of geographic
patterns of disease mortality, including respiratory diseases, potentially related to toxic air pollutants.
20 See note 12.
21
-------
undertaken a number of efforts to fill the gap in knowledge concerning exposure levels to toxic air
pollutants. These efforts have ranged from the establishment of centralized data bases of air toxics
information to geographically focused efforts at exposure (and risk) estimation. The following
paragraphs briefly review some of these efforts.
EPA currently has no comprehensive national monitoring system or data base of toxic air
pollutant monitoring data21. The two national data bases on air pollutant levels, the Aerometric
Information Retrieval System (AIRS) and the Air Pollution Enforcement Data System (APEDS) do
not include information on toxic air pollutant levels other than lead, which is included by virtue of its
status as a criteria pollutant (Greenwood 1993). These databases also contain information on
paniculate levels, but this is not speciated with regard to chemical composition.
During the 1980s, EPA attempted to investigate the connection between air toxics and adverse
health effects by a number of local and regional exposure estimation, exposure modeling, and risk
assessment studies. Some of these investigations were undertaken as part of the Integrated
Environmental Management Program (IEMP). This program was initiated in 1985 in an attempt to
characterize multipathway exposure to pollutants in a number of urban areas and to evaluate strategies
for their control22. In almost all cases, the air toxics exposure estimates were developed based on
emissions inventory information and air quality models, and only occasionally validated by air
monitoring information (an exception is the Kanawha Valley, as discussed in Section 4). The
estimated exposures were used, along with toxicological models, to estimate risk associated with
toxics exposure, as will be discussed below.
During and after the IEMP, EPA made several attempts to characterize air toxics exposures at
the regional23 and national level through large air modeling studies based on emissions data. As
early as 1985, EPA published a national exposure and cancer risk estimation study for air toxics
(Haemisseger 1985). This and three other detailed exposure modeling and risk estimation exercises
conducted by EPA are discussed in detail in Section 3.2.4.
For the most part, attempts to measure air toxics exposures to evaluate their relationship to
spatial patterns in adverse health effects have been opportunistic and highly localized. Where
monitoring data related to pollutant releases from specific sources are available, they have
occasionally been used to support epidemiologic studies, as discussed in Section 4. By far the
dominant approach to quantitatively estimating potential exposures to air toxics, however, is to use
emissions information and air modeling data as was done by EPA in the IEMP discussed above.
With the passage of Emergency Planning and Community Right-to-Know Act (EPCRA) and the
21 EPA developed a data base of modeled (not measured) air toxics concentrations for the IEMP cities and
some other locations, and distributes it through the National Air Toxics Information Clearinghouse (Cote and
Bayard 1988).
22 The areas addressed were Baltimore, Philadelphia, the Santa Clara Valley, the Kanawha Valley, and
Denver.
23 Studies similar to the IEMP air modeling studies were performed in the San Francisco Bay Area (Kaldor
and Harris 1984) and Chicago (USEPA 1989a).
22
-------
promulgation of other requirements for pollution sources to provide emissions data, it has become
easier to get emissions data for use in air modeling studies; in addition, exposure and risk
characterization methods based on emissions modeling are required components of the permitting
process for many industrial facilities.
A limited number of intensive efforts have been made, however, to comprehensively evaluate
air toxics exposure patterns for small areas or narrowly defined populations. Among these have been
EPA's TEAM (Total Exposure Assessment Methodology) studies of individuals daily patterns of
indoor and outdoor air pollutant exposures, and the Airborne Toxic Element and Organic Substance
(ATEOS) study in New Jersey. In the TEAM studies (Wallace 1991, Wallace et al. 1984a,b) EPA
undertook to develop and employ personal monitoring technologies to evaluate patterns of exposure to
toxics substances during the course of daily activities of study populations in a number of U.S. cities.
The most important finding of these studies, for purposes of this effort, was that indoor air exposures
appeared to contribute more to total air toxics exposures than did outdoor air. Often indoor air toxics
concentrations were higher than outdoor exposures by a factor of ten or more. These studies did not,
however, determine the specific sources of indoor (or outdoor) pollutant exposures, and the degree to
which the findings of relative indoor-outdoor exposure relationships can be generalized to larger
populations has not been established.
The ATEOS (Lioy and Daisey 1988) study evaluated exposures to airborne organic and
inorganic pollutants at three urban and one suburban site in New Jersey for four 39-day periods
during two consecutive summers and winters, and developed a detailed characterization of the
temporal variability in pollutant levels during the sampling periods. In addition, detailed chemical
speciation of the paniculate air pollutants was performed, and source apportionment modeling was
used to identify the contributions of major classes of pollutant sources to the observed exposures.
The results of ATEOS were also used to develop quantitative risk estimates for exposure to air toxics
at the various sites. A smaller scale effort, also in New Jersey (Butler et al. 1993), studied the
relative contributions of ambient air pollution and other exposure pathways to the total intake of
benzo(a)pyrene (a typical carcinogenic PAH compound) for typical exposed individuals.
3.2.4 Biomarker Studies
Biomarker studies attempt to evaluate and characterize exposures to pollutants through
measuring changes in tissue levels or other indicators of exposure in exposed receptors (in this case,
people). Biomarkers can range from tissue concentrations of pollutants and changes in enzyme or
metabolite levels caused by pollutant exposures, to analyses of pollution-induced reaction products of
genetic material. As noted previously, NAS/NRC (1989) divides biomarkers into two categories,
"biomarkers of exposure" and "biomarkers of effect". The former term refers to tissue
concentrations of pollutants and metabolites, while the latter terms refers to metabolic or physiologic
changes that can be regarded as the first manifestation of adverse effects. (Some authors also view
some of the subtle neurobehavioral effects of chemical exposures as "biomarkers", but we reserve this
term for measurable changes in biochemical and physiological markers, consistent with the NAS
definition.)
In the case of toxic air pollutants, biomarker studies have been used both to provide indices of
the occurrence and severity of exposures, and as predictors of potential health risks. Among the roost
well-studied biomarkers of toxic pollutant exposures are the changes in blood lead concentrations,
23
-------
tooth lead concentrations, and heme metabolites, associated with exposure to lead compounds in air,
water, soil, and diet. Although such measurements can be used to reconstruct air exposure levels for
the general population, it is almost always the case that direct air exposure contributes a relatively
small fraction of total exposures compared to diet or exposures to contaminated soils or drinking
water2*. Where the indirect pathways are dominant, however, it is usually clear that the original
source of these indirect exposures is air pollution. EPA uses children's blood lead concentrations as
an indicator of potential adverse neurodevelopmental effects of lead exposures.
Blood or tissue concentrations of other metals are also occasionally used as indicators of
exposure and potential adverse effects for airborne toxics. Among the studies that use biomarkers of
exposures studies which we located through our search of recent literature are evaluations of tissue,
hair, and urine cadmium levels in a population near heavily industrialized cities in the ex-Soviet
Union (Busteva et al. 1994), and a study of arsenic concentrations in hair, urine and blood of children
living near a smelter (Benko et al. 1977). Mercury is another metal whose concentrations in tissue or
urine are often used as indices of exposure and potential adverse effects.
Recently, biomarkers which provide indices of exposure to mutagenic substances have become
more prominent. In addition to the studies of chromosome aberrations (biomarkers of effects) around
hazardous waste disposal sites which were noted previously, at least two other families of biomarker
techniques have been increasingly applied to characterize potential exposures to mutagenic and
possibly carcinogenic toxic air pollutants. The measurement and study of DNA adducts, or chemical
reaction products of pollutants with the DNA (i.e., genetic material) is a techniques that has recently
been used. Our survey of recent literature identified two studies (Binkova et al. 1995, Perera et al.
1990) of the relationship between DNA-PAH adducts and PAH exposures in ambient air. Both
studies found strong relationships between measured PAH levels in ambient air and PAH adduct
concentrations in peripheral leukocytes, suggesting that this method may be a useful indicator of
recent exposure levels. In addition, because DNA adducts are premutagenic lesions, these data
suggest that exposures to ambient levels of PAHs may lead to mutagenesis, and ultimately, to
carcinogenesis. The results in both of these studies were not affected by confounding factors such as
cigarette smoking.
All of these biomarker methodologies discussed above support, in one way or another, the
qualitative relationship between air toxics exposures and adverse health effects. They can show, first
of all, that exposures to given agents have occurred, and that sufficient amounts of the agents have
been taken into the body to cause changes in tissue levels or metabolism, or to cause adverse health
effects. They can also be correlated with exposure levels, to some extent, and can be used to
document (as in the case of blood lead levels) when potentially harmful levels of exposure have
occurred. Finally, some of the biomarkers can be considered precursors to clinically significant
adverse effects.
24 In the commonly-occurring case of exposures resulting from metal smelting operations or waste disposal,
the original source of exposure may have been air releases, although the actual exposure medium is soil or
house dust. The other most common sources of lead exposure are particulates from automobile exhaust (less
common in recent years, with the phase-out of lead in gasoline), leaching from pipes or solder in water
distribution systems, and ingestion of lead paint.
24
-------
3.2.5 Risk Assessment Studies of Toxic Air Pollutant Impacts
Because of the difficulty in detecting and measuring toxic air pollution effects, EPA has often
relied on predictive risk assessment methods to estimate their magnitude, in the absence of actual
health outcomes data. Risk assessment, in the context of this paper, means the development of
quantitative estimates of the probability or severity of adverse effects using measured or estimated
exposure information, and lexicological models and parameter values describing the relationship
between exposures and effects25.
The technical controversies surrounding the application of risk assessment methods constitute
the main motivation for this study of recent epidemiologic studies. Risk assessment studies, cannot,
in and of themselves, prove the existence of a relationship between air toxics exposures and adverse
effects. If the exposure measures or estimates used as inputs to the risk estimate are incorrect, or if
the postulated relationship between exposure levels and risk is incorrect, little confidence can be
placed in the risk assessment results. Nonetheless, risk assessment provides a relatively inexpensive
method for developing risk information in support of policy decisions, where direct measures of
effects are not available. For this reason, EPA, and others have undertaken a number of risk
assessment efforts to characterize the relationship between air toxics exposure and adverse effects.
As noted above, an early EPA effort to quantify the adverse effects of air toxic exposures on
human health was a cancer risk assessment undertaken by Haemissegger et al. (1985). This analysis,
which used emissions data, air quality models, and EPA cancer dose-response parameter estimates to
evaluate risks from a relatively small number of pollutants, estimated that between 1,800 and 2,400
cases of cancer per year were caused by air toxics. The aggregate totals in this analysis were driven
to a large extent by very high risks calculated for areas near a relatively few point sources.
EPA developed more refined estimates in the "Controllability Study" (1989b) and in the
evaluation of Outdoor Exposure to Air Toxics (1990a). The first study, building on the results of the
IEMP efforts discussed above, estimated cancer risks due to air pollutant exposures in the five IEMP
cities. The total cancer risk from toxic air pollutants across all five cities was found to range from 2
to 10 cases per million individuals per year, implying comparable or somewhat lower national
population cancer risks than the previous study. This study found that the chemicals which accounted
for the greatest risk were POM (27 percent), 1,3-butadiene (19 percent), formaldehyde (18 percent),
and hexavalent chromium (17 percent). The source categories which were associated with the greatest
risks were vehicles (55 percent), chrome plating operations (9 percent), solvent use (7 percent), and
wood smoke (6 percent). Emissions from chemical manufacturing accounted for approximately 1
percent of the total estimated risks. A subsequent review of these estimates (Portney 1991), used the
same data derived by EPA to develop much lower estimates of nation-wide air toxics-related cancer
risks.
25 In a typical risk assessment for toxic air pollutant, individual cancers risk are estimated by multiplying
modeled or measured air concentrations of pollutants by the Unit Risk values for those pollutants. For
noncarcinogens, air concentrations may be compared to, or divided, by Reference Concentrations, to generate
an index of the potential for adverse noncancer effects.
25
-------
EPA's 1990 study estimated the relative contribution of a number of chemicals to the total
cancer risks due to air toxics. They estimated that the largest contributor to total cancer risks was
PIC (products of incomplete combustion) at 35 percent, followed by 1,3-butadiene (12 percent),
hexavalent chromium (9 percent), benzene (8 percent), and formaldehyde (6 percent). EPA (OPPE)
is currently conducting a national air toxics exposure study, (as part of the "Cumulative Exposure
Project") which will also be used to develop risk estimates to toxic air pollutant exposures. This
project is still in its initial stages.
Similar efforts to use exposure and risk assessment methods to develop population cancer risk
estimates due to toxic air pollutants have been performed in Sweden (Tornqvist and Ehrenberg 1994,
Hemminki and Pershagen 1995). In one study, the lifetime cancer risks associated with exposures to
typical urban air pollution levels was estimated to be on the order of 85 per 100,000, with the major
contributors being POM (primarily from vehicle exhausts, about 58 percent), 1,3-butadiene (about 24
percent), and all radionuclides (about 15 percent). The other study estimated lower lifetime cancer
risks, of about 17.5 per 100,000 at average Swedish ambient exposure levels, with the major
contributors being PAH, asbestos, benzene, and arsenic.
These estimates of cancer risks associated with toxic air pollutants remain extremely
controversial. This is partly because of the uncertainty associated with the modeling procedures used
to estimate exposures, and also because of the very large uncertainties associated with cancer dose-
response estimation at very low dose levels. There are major controversies associated with low-dose
cancer risk modeling for all of the major contributors to cancer risks which were identified in the
above studies.
In addition to the risk assessments for cancer just discussed, EPA (1990b) has conducted a
large-scale risk assessment study of the potential for adverse noncancer health effects of air toxics
exposures. The approach taken in this study was to compare monitoring data or modeled pollutant
concentrations to health reference levels (inhalation RfDs) or LOAELs (lowest observed adverse effect
levels, usually from controlled laboratory studies in animals or occupational studies; they are almost
always higher than reference doses). The analysis found that there were numerous instances where
measured or modeled concentrations of pollutants exceeded reference concentrations. The
concentrations of about one-third of all of the pollutants for which health reference levels were
available exceeded the reference levels at 25 percent or more of the exposure locations.
Concentrations for two percent of the pollutants examined exceeded LOAELs at one or more
exposure locations. As was the case for the cancer risk assessment, these results are consistent with a
high degree of suspicion that exposure to at least some toxic pollutants is resulting in adverse health
effects. However, they do not provide any direct confirmation that adverse effects of the types
modeled in the study actually are occurring.
26
-------
4.0 SUMMARY AND RESULTS OF STUDY REVIEWS
This section summarizes the results of some recent epidemiology studies that we identified in
our research that assessed both exposure to HAPs (or a surrogate of exposure, such as distance to a
major emitter), along with observed health effects associated with the exposures. Section 4.1.1
provides an overview of the studies reviewed, the types of exposures that were evaluated by these
studies, the types of endpoints that were evaluated, and the methodologies used. Section 4.1.2
provides more detailed descriptions of these studies and an evaluation of the overall knowledge base
related to associations between exposures to HAPs and adverse health effects. Section 4.1.3
addresses the "bottom line" from these studies and their impact on establishing a link between air
toxics exposure and adverse health effects. The reader is also referred to Appendix B for detailed
reviews of selected studies.
4.1 Epidemiology Studies Evaluating the Link Between Exposure to HAPs and Adverse
Health Effects
Based on the literature searches and interviews with experts described in Section 2, 34 studies
were identified and obtained that attempted to associate some measure of exposure to HAPs with
observed adverse health effects. As discussed in Section 2, although an attempt was made to cover
the literature, the reviewed studies do not constitute an exhaustive list. For example, because the
purpose of this study was to identify specific geographical areas where health effects may be
occurring as a result of HAP exposure, negative studies were de-emphasized. Furthermore, a number
of published and unpublished studies were identified, but could not be obtained and reviewed, due to
time constraints. For certain well-studied classes of exposures and effects (most notably, exposure to
arsenic from smelters and lung cancer), time constraints again prevented a thorough documentation of
the available literature. Finally, due to the large number of reviews covering the scientific literature
prior to 1990 (see Section 3), emphasis was placed on the literature since 1990.
4.1.1 Overview of Studies Evaluated
Table 4-1 summarizes the 34 retrieved studies evaluating health effects and using
measurements and/or surrogate markers for levels of exposure to HAPs. It should be emphasized that
the summaries provided in this table do not represent a detailed analysis of the methods, results, and
limitations of each study. Instead, the table lists the key facts about the study, including the most
important conclusions and main limitations. Additional limitations, including important confounders
that were not accounted for, may also exist.
It is also important to note that most study authors only assessed exposure (or measured an
exposure surrogate) to one air pollutant or a small group of air pollutants. Exposure to many other
air pollutants (and to pollutants in other environmental media) probably occurred in most of the study
populations. The degree to which these other exposures may confound the study conclusions depends
on the type of other pollutant and the endpoint evaluated. In situations where there may be
confounding ambient exposure to HAPs other than the one measured, the observed effect may be
erroneously attributed to a specific pollutant, but this will not affect the overall conclusion regarding
an association between HAP exposure and adverse health effects.
27
-------
H
Comments
8
v^
§
00
~4>
w
3
CA
S
(A
*-*
1
U
£
3
e3
toM
S
£
3
O
O.
X
u
u
f
4)
g-
fc:
>»
T3
B
on
Increased prevalence of
bronchitis, asthma,
angina, skin rashes, ear
infections. Possible
recall bias,
toxicological
mechanism unclear.
11
*°|*
* £*3 IS
0 .£ o > «
t/T e3 JZ ~ "S
c o > 2 . o
0 '£ - g fr 8
& & t* 5 2 .b 8
iiiiil
£ -0 S W)N^ E
"eS O
C« "Jys
4) *"
S E
c o
1*
c *?
e i>
= o
8 g
S *5
4, T3
O *j
C C
4) o
S fe
si
Oi T3
-1
a. 4>
Z "H
fe
* S i* ~
8.§ 79
i i ^ U
4> C ^
* I2* t
"S | a §
E 8 '« U
Tl-
ON
o\ •«
. O
"53 '«
2 ^
^ 4)
4) s«
1- M
4> w
•a 2
CD U
Significant increase in
hearing loss in exposed
group.
| £
'S '«
•5 c
^l§
*^ *Zi
2 S
S g
00 CP
c ^
§ 1
S S
TO
£
re
"8
C/3
O
O.
X
4>
C
8
c
4)
12
'£5
£
c
4>
^ _^
2 C3
IE °
o M
c c «
III!
too £ Q. «
•S< fe*
^ o
CQ a fe 3j
« 2 o S
a c E.U
r~
t~
Ox
— "a
^ §
ts'«
4)
o «
^ M
U (A
C O
4) u.
CQ U
Increased DNA adducts
correlated significantly
with increased exposure
toPAHs. Well
conducted, controlled
for smoking.
1
§J5
re —
4) 4)
1 W
«, •§
0 0
= i
•o -°
•a 4>
"t"j
g^
Q .S
42
0
4-| 3
C -o
4> .« -0
£ 15 «
« E <
0 § S
CL Z
x h£ D
*> c
"5 'C 4>
c o ^
g-H 1
111
c
O 1/5 y
0 o '•=
IlSl
4> C 5 •«=
^^ ^3
HSfl
<*. "0 . •
o "S !2 •*?
« * « §
& $ * 8
!C Z o «;?
« Q S £
in
rt\
5
•a
« 8
« K
> i
O eo
^ >
.S £
CQ CM
Arsenic exposure
resulted in significantly
increased risk.
Controlled for smoking
and occupational
exposure.
..
£>
"J3
C
0 W5
E S
^^ ^
^ 4>
u^ ,25
S M | .
« S 'S & -s M "8
g S fe C u. c «5
^ .2 ^ 3 o .5 o
8li = |H
^ c — o e « o
^ D O. S C to 0
r_
3
JB"-rf
N -5.3 ««
ro ^S ** ^
4) « W5 •—
c S < 2 c
8-isll-l
M S « JR rS >
P, u S -c W ^.
M^ If •• -
"l §.°.wi g
3 C X X) ' 4>
j S u £ D cz
s
GO
OS
•a o
*-* *—
o
t> 1
O 4)
CQ rt
* U
oo
-------
f
I
I
i
8
'•5
I
•o
4
Oi
S
ents
1
o
£
I
ffi
CO
«
1
u
O
a.
X
T5
3
1/5
Illll
S
e IS
60 S
o
*o . •
II
M
tl
t-> D-
I 8
ro
00
?.§ c
u £ o
'•S 60 4)
*- P fc
o 5 X
u o 8
c
i
cr
E
2
Q.
E
c to w
.1 -§ '=
'•5
i-O -
"8 §'i
S^
ON
(N
*>
S 2
S §
-a g
c .5
8 «
.a o
^ 3 fe e
>. exZ o
c •c p _, tS
O C O
4-« C >X
o. o h
»^> Wa ^J
on CL<
ON
73 .2
-------
•8.8
I
E
8 « I -8
ill S
ill 1i.i
sft
if*•>
T3
S
en
•3 1 E'-s 8
.^-i c •— O ?*
° «? > -S '-5
HM C/J ^w .W
f, C rt C ^5
0 .2 E .2 o
-r an V-, or, C _
>% wa O vi rt •=
.= 'E S 'E fa e3
;" C /^ C **• —
?* 4> TO 4> T3 O-
">. _
Congenital CNS
malformation near vin
chloride plant. U.S.:
Kanawha valley, Wesi
Virginia
« -3
•»-> c
-------
b
Ih
a <2
««__
'•3 -a
€ S
'if
•§ S
§ •§
H 5
8
a
ili
ll&l
•R » 2 =
sl£l
III!
g
v>
I
8U "t3
«- O
C _c, N
•- «
•a
O
.fc P o
C
o
4> _ O
111
^Jl Q} U*
O
O 4)
« '«;
o v
12 o
<*- c
O O
o __'
5 JS
"c
•=
4J u O • —
S S ° g.
8 B-S#
•s a a .
b § .E ^
g S o 2
• S m
r- U
5 *«
O CQ
*Z2 60
O 3
C «
ry fu
near
U.S.:
ato
ts
spir
iden
Re
es
re
factory.
Hawaii
•81
.E '•« v-
^ E 1
2 0- §
« ." T3 W
•1111
fil.sC!
^i
Os
ON
•a •§
4-t •— C
0 -S .2
J^.* C W) flJ
•o c Q* c:
I oo 8
re!
s «
I - f
I ON £
* — ft.
t~-
00
-------
Comments
C/3
I
a
tfi
a
X
u
'5.
•o
s
00
'
alt
12|
w « o
• — rt c«
u. SS ° ^
SI'i i
5 Sa 1
V.J «5 * o 5 O .ti
3 on c S S <«
=5 s^e 1 S
Cu 0= J2 « C T3
S -a v*. O
•5 S ° -6 ™
O J8 t« e<3 S
-
o -o o> - c «
-s -S a 6 a u
. c
•S »S
als
s
e
bo ^
5 -a
o o
bo- "5
| 8-5
1 S £
CD C 0.
o
S «
O O
C Q.
^^
D'S
tj T3
O C
£ _ rt
511 .
S g ^S
'S c c 'S
rt e .;;
•S o S
. U
Loui
Neurob
populati
reproce
Superfund ste
Baton Rouge
C £ -p
i- rt S
S -5 J=
cs
m
-------
Comments
§
I*
o
to
S
3
s
1
O)
1-
3
M
_
Cll
2J
tn
a
X
UJ
0
1
P-
?">
^
T3
S
Lung cancer mortality
significantly correlated
with degree of
urbanization. In people
^65 years old,
significant correlation
with SO2 after
controlled for smoking.
o
(«
S £•>
81" '5b
c "
o 73
60 o
_c "^^
*^ o
*£ 0)
Sj C^ ^
*— ^ 15
£J O
^N S Q
o S *+•*
+-»
t/3 *• cc
"p 2 £f g
~" ^ EJ P
S 8 s "a
c
°«i &
4) =
= t
S 0
i?»
•O 4) T3
_ O C
e« C «
C « ~
O ° 4>
'5> w-S
OS 5 2
ro
ON
*"" ig
73^ =
•So
« w"'^
(/5 60 4)
C Q U
30^
ii O o
High quality, but SCEs
indicate repair of
damage, rather than
persistent damage.
12
'S
g «
•g |
V* C^
11
en 8
uo
C
0
\c E
c "c ^
4) ca cQ
C »^- ^ ^
8c *^- 5
«. M .2
i- 5? *J «5
•!- te * S3
«-M O "O t
g -5 § ^ "S
l "3 a s> s
S> > E •— **
C
60
rt
•g cL
x c
•S 1
rt 73
o «5
v. 3
111
en c CQ
ON
2
• 73
73 c
.*-
•w W
4) "2
S3 ««
3
8 -s *^ *§
alii
.
M <^~
•t! "S **
^" ?
fl3 "T1! T2i
*— t U C3
O. eg ^
fa E *
^ X *5ii
c i3 ^ E
*> «^ «f C
o k. 4) fc
2 4-J („> O
1 1 &'i
o; o cc c£
Ui * I
~; * C
P JS
•^ . &*
1§I
O O, '
e v
= 73 g
fl) .^M C
o p «^-
CQ f* CQ
U 73 CQ
00
Ox
73
•Q 73
.- .y c
ll^ J= O
e« CX •£;
* 8 t
SO o
Decreased birth weight
at ^ 10 km from
smelter. Not due to
difference in parity, but
no control for smoking,
occupational exposure.
O M
- 0
fen S
1^
•*-»
•S "5.
t. t«
5^
«- E
CQ O
c *fc
•2 g
60 fi
4) C
*•" e«
c «->
4) T3
U ^
C C v.
!2 fcl
Bl T3 en
^
£
60 60
'5 .E
S t! C
"P G£
•^ « ^
(/5 C ***
1 8.0
Oca.
73
tS 73
E '~* ^
£ °r'^
|rS ||
z-o 8
en
-------
S3
«
• Mi
«
_6JD
I
•
55
1
H
Comments II
£3
O
or>
g
3
4)
S
U
1
1
4)
1
O
cx
£
u
'i
4>
B
f-
* —
T3
5
Significantly increased
cancer and
cardiovascular disease.
Attributed to SO2 and
nickel.
8
09 *5
| S
O '•? C
•g 8|
C9 r" H
£> w *o
•3 *O TO
eo w *j
|| |
^
2
-e 2
.S= ..TO
2 d"^
C 00 C
.2 i- .2
•f «2 j«
— c5
1 2 «
O ,^S ^-«
c* 1 E
&
^
•41 0
Si
c '35
I* 60
S 3
c "*
ti
^-* —
P
•c <»
ON "§
— '^ 0
rp" S« "*^
^ kx"§
i 8 g
208
Decreased performance
on objective tests
(finger tapping, hand-
eye coordination) with
increased hair Hg.
However, hair Hg
higher in control area.
Unpublished.
a
S
13
VM
>
O f
|"s
i §
CO c
» ».
"*5 T!
> o
M"§
K1!
4> °
'C c
.3 .-
1 g
^ *o
'« '«
ac 2
•s"§d
§ g c/o "g
— 3 O O "O * O
£ c S- .S < g- .a
I'o |l°"s^l.
^ E "S s ^ * *
°
U
4>
•O
'o
C
it
a w)
u fi
o _
T3 2
o c «3
a .2 «
alg.
^ ^ ts c
4) T3 C .2
3 C •— '5
w *o e
S4) .3
c Si b" E
c 5 o cd
. ^^ *? L-i
3 -II I
*-<«••
3 g V5
«SD
o u
S "t; «> w
4> O * , S
^ « 3 4) §
o c _. ex 55
.S g ^ 5 2
w s ^ ^ "i
4) C t *" W
il|l|
U Q 5 p S
1
.
13 -g
t ~
?r^ O
§ X
o
O M
Significantly increased
risk in exposed group.
Controlled for
occupational exposure,
smoking.
E
&
•S
^) t^
(A
oj &£
(j 4)
C<3 --3
o w
&.1
~J C
nf
4>
W< *13
rt o3
55 go
0 S
g-E
C *W
8 -
si
*O ,S-
*S -O
-------
Signifi
for all
pathwa
expos
levels.
no eff
-
I
't*
g
§
I
u
«fl
8,-S
II
•c t> o
TO 'r- ^
III
CD -H- w
"Si
2 §
C/3 .
S
«
S
S
o
o.
X
UJ
4)
•5 u LL] -E
lagji
^"™ yj CM C
"c5 ^ c^ O
P 'S _- E
"s S «
v- O O ;«
lllls
.O e« g « «
H J5 >= T3 T3
bo £
2 •£
.
. S i "8
1
.
O. cs
.E-2
II
-5
CJ ^- ^
'^* O *I«3
2 c a
4J ro o
& E -
•' P £
bet
bio
Relationship
releases and bi
human exosu
_w
a>
V.
.52 M
4i C
a .2
e 60
S «
S «-
= oo
g.D
n
W)
a -§• s
11 •: § ^
'S 1 S .S3
S
o •- «
j: £3 s
K O fc- «! .-
ll
U
-
3 «?
oa ^i
4>
TJ
s
on
OS ^
ON •—« C
all
S bo a>
ill
•53
s
V
o °^
* —
8
°
"8
-------
WD
1
I
i
I
S
•e
i
I
4
V
3
a
o
1
v.
4-J
1
U
i
V
s
i,
£'
*
3
in
on found
No assoc
4)
O
s
is
11
M
«
.2
<4_>
I
0
o o
S3 t
O^3
w
V
and
S
Leuma
consumptio
and Eastern
countries)
ke
on
•o
§
c
II
2 S
u i/-, bO 0>
TO >•*, O **
* & r« 8
on — w o
Increase in m
types of birth
correlated with
variation but no
distance to plan
Well-controlled.
O.-O 10
£§ E
"e c '?
S £ |
Si 2 €
8 "2 "".
Ǥ S a
S.S
Ml
1 g>
IS
e o
- E
'« c
c -2
.1 52
1 ^
C «5
< '-5
> i
•O T3
si
CO "5
CO
oo
Os
« .2
>
181
•o ."§ £-.!?'« .S
c
CT1
E
2
a.
>.
on
C C
O tS
c
3
!> I
la 8
o
'E
c
rt
'C O 13
c t?
OJ c
S5 4) o
g E o.
iu s -g
•^ o S
**• U* TO
1
g
O ^
O o •
•2 v« C
cs S O
•- o -a
x * 5
« -S =:
on S O
•§g-^
3 C
-q cc
C 73
'c: s
0 C/5
O<
o\
•S I
S-'TO
... bO 4>
CO
-------
"
1 1
E S£ 2 8 £
relation of BAP
Poo
ty, little
odology
orr
ith
ali
eth
for
qual
m
Amb
affec
lyso
expo
gene
j; v-4 v*> __ .^ & r?
&!S « E !£.£ S
0 „ C >- 4= 2 T3
M -s s « 1 s I
.« 2 a. « S o. t
> 8 E «£ •*= E g.
o
U Hi
I
a
u
rt
O
00
3
1
'
u
11
£. «
II si *.§
1 1 1 i * s
VO C o O to 4>
Exposure
o
3 o
C c
•™ rs 4_»
E £ 3
if
3 C C
S 0 —
g£
u
O c
lie
? 3 O
B
3 *
C § ^
."2
J5 w
o w
•r .=
i^v
E «=
> .2
*-*
^5 ^3
C k-
1 I
rt O
«> C
".
£
o
o.
.E o
™ 'Sf
«-* O
l
ac {a
s
g
1
3
c .5
•- «
ee
60 O
— _ C>
s|-|
s •§ s
"5 o"—,
^M C^ ^^
e3 c
| 1 •>
111?
u
0
J=
£ u
«
•o
S
C/5
CN
CO
S
1
8
"a s
•
3-a
C^ O
5 -S °
^ 2'i
"S 60 *4>
s g fc
>• o 8
OS **3
ON ^*
v— C
O
<3 "5
o %
S S
N U
-------
4)
v> __ O
S SS 0
3
C/5
•a §
*'S
3 vi
§ 2
i s
N U
o
fc>
o
C/3
S
3
W
S
a
1
w
4)
w
3
1
4>
W
3
««
8.
X
tu
%
V)
O
4-*
.&
*u«
o
ss
<^ a
M O
5 t
0 S
«T M £?
13 c >
S» 2 '«
- 5 %
c „- >>
4> 13 ~
1*8 •§
1 6 S
•a § §
*5 -2 u
fe M 0
11 =
< w s
II ° U H
i
^
^
a. . .
30 ° i 5 &
K •& < s ^
o
N
S *> v> «"
B5 «0 « M
i- a S C
CO
CO
-------
Because high exposure to HAPs is generally limited to a localized area around point sources, such
situations would generally be limited to studies conducted in highly industrialized areas, such as the
Kanawha valley, with several different industries emitting different HAPs in a localized area.
Exposure to HAPs in indoor air presents a special case of potential confounding. Most of the
population spends a majority of the time indoors, and personal exposure monitoring data indicate that
HAP concentrations in indoor air can be much higher than those outside (Wallace 1991, Wallace et al
1984a,b). Very few of the studies that we evaluated used personal exposure monitoring or controlled
for indoor air exposure to HAPs. Due to the wide variation in indoor exposure, it is difficult to
design studies that assess the effects of HAPs in ambient air, while effectively controlling for indoor
exposures.
When the confounding exposure is to criteria pollutants, the potential for confounding depends
on the endpoint being studied. If the exposure to criteria pollutants is sufficiently high, it may be
impossible to determine whether certain effects (e.g., acute respiratory effects, eye irritation) resulted
from exposure to criteria pollutants or HAPs. By contrast, other effects, such as objective
neurological deficits, are rarely attributed to criteria pollutants, except for lead. Thus, if exposure to
lead is low, exposure to criteria pollutants is less likely to confound any observed association
between neurological effects and HAP exposure.
Table 4-2 categorizes the studies screened by study type, the type of pollutant that was
evaluated (keeping in mind that other exposures may have occurred), the exposure measure or
surrogate used, the endpoints that were evaluated, and the location of the study. As shown, cancer
was the most commonly evaluated endpoint, accounting for 14 studies. General systemic effects,
including overall mortality, were evaluated in 11 studies, while neurological endpoints were evaluated
in 9 studies and reproductive/developmental endpoints were evaluated in 7 studies. Noncancer
respiratory effects were assessed in 7 studies, and 3 studies considered genotoxicity endpoints.
(Because several studies evaluated multiple endpoints, the breakdowns discussed here do not add up to
the total of 34 studies reviewed.)
A variety of different exposure measures or surrogates was used in the reviewed studies.
Measured ambient outdoor levels were used in 8 studies. The most common measure (9 studies) was
simply whether or not the subject lived in the exposed area (indicated as "residence" in Table 4-2).
An additional 8 studies used distance (or inverse distance) from a source as a surrogate for exposure,
often in an attempt to establish a dose-response relationship. Dispersion models that took into account
such factors as wind speed and direction, topography, and distance to emitters were also used to
estimate exposure levels in 3 studies. Eight studies used other exposure surrogates, often an exposure
index that combined such factors as distance to the emitter, duration of exposure, and amount of
exposure during different time periods. Biomarkers of exposure and/or effect were used in 4 studies.
Several studies also reported ambient pollutant concentrations or emission levels, but did not use these
data to evaluate the epidemiological data or in a dose-response analysis.
The most common type of study design (13 studies) was the geographical correlation study,
which simply compares the incidence of an effect in an exposed area (or areas with different degrees
of exposure) with that in the control area. Cross sectional studies accounted for 12 studies. In these
studies, a selected group of the exposed population was matched or
39
-------
I
1
S
S
I
13
.£?
|
e
1
'•S
o
1
I
3
4
V
»2
m
0
c
'o
ex
Exposure Measure
Pollutant type
Evaluated
«>
1?
^
s
C/3
^3
3
C/3
!S
2 .'5
•; c«S
*? §73
POO
4>* ?A .
Cancer, reproductiv
mortality, respirator
CNS, general (othei
systemic)
Residence
Multiple (Metals
organics, not
further specified
73
c
g
M
£
o
•t
2
73
c3
4)
•a
CQ
.a
!
A
5
Biomarker (DNA
adducts)
Personal exposure
monitoring, ambient
measure, DNA adduct
| levels
CX3
ac
u
'i
$
«
u
§
73
0)
£
73
•4-*
>
O
C ON
5 —
e.s
2 S
C3 ^
u ^
•; g
S«£
Cancer mortality
Other (measure based on
emissions, distance from
plant, weighting based on
soil samples)
„
Metals (arsenic,
cadmium, lead,
manganese, zinc
S02
8
o
V
8
U
00
—
73
4>
£
O
&
cd
W5
•a
Systemic (kidney
function, as measun
/3-2-microglobulin
excretion)
Ambient measure,
biomarker (urinary and
hair cadmium levels)
1
"3
1
73
c
.2
«
««
(A
£
0
73
t3
>
3 Tf
<^ o\
CO
T8
t
o
Residence (emissions
were reported)
4) c^
sf
**3
•J §
ca4'*3
w ^5
8 u
rn ^
oo
73
\-
'•5
8
1
§
a
•: S
21
•a
Respiratory,
neurological, generz
(cardiovascular,
gastrointestinal)
Distance, other (measure
based on distance to site
(1/d2) and concentration
of chemicals at that site)
oT
Organics
(benzene, toluen
tetrachloro-
ethylene, others)
73
c
_o
B5
CA
£
o
•n
2
73
a3
73
Q
-------
Location
CQ
ea
'=
i
*
"
*s
2 <-*•
Is
o
o.
T3
O
"o
C T3
oc
U
wo
El
o £ c
it 5 .o
o ^.2
Hi
o
2
Exposure
•*
E
0
Cu fT1
12
_o
o
o>
CD
4-*
I
s
en
o
1
o
•o
3
t/2
«!
8
t/5
2
o
H
«5 c
«« o
M U
s«
o
U
es
s
C/3
00
t~-
00
o\
TO
OJ
£
s
00
-------
I
v>
g.
I
1
Ml
«^
a
•
.2
o
5
s
a
T3
w
Exposure Measure
Pollutant type
Evaluated
P.
£4
T3
a
•^
B
en
c
P3
ex
en
O 1—
§ C
•^3 4i
c5 o
0 E
M E g
•g 5-c
•^ 4j ^i
o s _y
Ambient air moniti
other (biomarker--!
levels of organochl
compounds)
Organic
(hexachloro-
benzene)
"rt
o _,
U
cQ £3
bQ Aj
J
•a
4_*
P.
Si™*^ ?2
Respiratory
function tes
radiograph)
Residence
1
w
.0
in
1
-
« 4> "S O
b s* o°'§
i 8^ g=5
O >iiQ. S ex
s^
d" JS c
-j 4>
^ W /^«S
U-< ™"™ ^^
-2 5
*«
•— c
Ctf ^Q
O ^*
O ^
s
2
•a
*-•
-------
I
g
I
VI
a
i
I
I
Ml
1
1
"•5
35
"8
B
o
If
5
o
ta
o
3
c
'5
O.
T3
8
&
TO
4)
£
7
1
X
Pollutant type
Evaluated
•o
3
oo
-o
3
en
|
'3>
1
ll
vIS -g
& aj
•S •«
X *3
o 2
II
SI
1
i
i
c
'J5
v ^
— !2 "^
u T, « 2 8
"e * c ° O
sU 8*1
6a>£f '£
—
c
M
8
u
ON
ON
"TO"
«
1
TO
8"
o
1
*s ^3
CO c
• ^^\
. 03
u-
4)
M
•S
S
•o
1
TO
U
Residence
^
1
&
"TO
0)
S
HH
TO
o
Geograph
correlatiol
13
S
.-
to
O
TO —
TO 00
cx
8
w_
§ s
S *
^5S
e
3
c
CX _,
o *^
4> *4)
Q ^
u
c
TO
Residence, dis
o"
o
If
•— —
s _
Geograph
correlatiol
.
"w
4)
E
o
c
tft
"P oo
IS
^*
I
§
S
5
9
c
b£)
£?
"•i
t;
o
^
_c
o
o
Ambient air m
emissions data
o" s~^.
••• U*
C 4)
4) O.
5£ „- o.
« E o
. .s „
"i "« £ -S
| « 1 1 o
<; c t3 o c/3
^.^
2
IS g
11
S
ON
i
<*>
o
Z
• *^
S
1
8
£
13
.2
">
TO
1
1
1
2~
u 5
fi3
•J|
c c
4) TO
II
/^
1
4)
2
4>
"TO
c
0
i
(/>
V)
8
u
ON
ON
13
4-J
I
O
H »
D .*§
o^ ex a
DOS
4)
U
4)
•o
I
U
i
u
O <4«i
'C o
rll
Distance, othe
based on distai
direction, and
exposure)
2
Organics, met
"5
w
0
V
U
13
t>
0
§
"8
^
e«
1
t-i
O
u
1
"u
c
!£
Residence; am
measured
^»
4>
I
1
"o
k.
Case-cont
oo
ON
C
4)
bo
TO
V.
1
-------
e
o
1
'o
ex
T3
U
s^
Exposure Measui
Pollutant type
Evaluated
o
•o
3
en
•a
5
2
en
c
o
'5b
w
p
Exposure biomarker
only (no effect
biomarker)
qj
4) '"•
e3 w
'e
0 ° «
'«5 JS2 2
•1 5.1-
JU-s
m
.s if «* B?
aiis-sn
6SII-I11
73
•^- C
"a. -2
bb"S
O 8
f*xl
Os
ON
^^
C/3
a.
^
j=
a.
.2
'5
•a
Developmental
(premature delivery,
birth defects, child
growth), respiratory a
neurologic disease,
cancer, general
morbidity
c g
*~* 73
§c
_0
SJ 'E?
E .52
o c
1/5 5
11-s
ISs
•S 73 w
S > w
{2 JS E
v." «* S O
s.:slu s
& o =s x" -o
e^ 8,Q--g
« O .2 §
S "5 b Q 2
•a"
.a I «
^s o o
bb73 "*
rw O V-
U U CX
ID
ON
.C
.M
>
^
02
c
o
T^fy
C fe
~ o
a*
Sal
ts
•o
en
Z
U
Distance
51
> -g
,
cd
O
Modified
control
73
"
bo
g
3
Other (Gasoline
consumption as s
6
-------
c
.2
S
5
'c
'5
a.
c
£
3
en
taM
S
5i
O
D.
X
LD
Pollutant type
Evaluated
8.
c*
>,
T3
3
c/o
^_^
"O
00
ca
*c
fe ^™*
c ^*
53 t^
^ »• cd
&\G
• C *Kft
Sa>
^
Respiratory and irritath
symptoms
4)
^
i
i
c
0>
IE
<
fl> x— s
S "
U3 «"
81
> X
73
c
0
'1
1/5
00
8
u
ON
ON
75
4_*
S-
03
•o
s
4-»
8
to
•a
Lung cancer, mortality
(general), development
(sex ratio)
u
o
T3
'i
c^
wT „ to
3-2 * .
ft) C ••• • t/3
E 1> — T u CO
<« 4) S r!
^^ I- ^ W5 O
"cL S3 "S .— •• *"
•S M « O O- T3
w—4
o
•£..1
cd ^
bb "w
8 t-
u 8
73
4-t
W5
E
cd
i-**" i/^
SI
2-
u
s
C w
4) M
nC ^. Cd
* O bo
D§>
Respiratory,
neurotoxicity, general
c
*v5
73
V
^
<— <
!!!
4)
73
•n
O f^
II
4)
.S
cd
S
u
u
I
« ^
V« ^^
3 cd
i'l
€ >>
4) 2
x> '5
E g
< 0.
1
^o.
'E o o
§ | £
6 eix
^— •
cd
'1.2
O3 ^rt
bb "5
8 t
^3 o
•
73
«
03
O
JC
II
03
C
S
« S
Genotoxicity (e.g.,
chromosome aberration
sister chromatid
exchanges), other (seru
lysozyme)
73
€
15
1
i
2
i?
gi
>• o
73
c
.2
4—*
W5
un
i
u
ON
ON
"
75
•g
0
CO
N
8
1
Other (drug
prescriptions as
surrogate for general
health effects)
1
4-1
c ^_
Q^ flj
11
W c
•g-i
=i £2
3
O
^-
1
N
-------
"g
I
I
o
1
00
U
O
o
o
•o
^
u
'i
o
J3
u
"o
>>
u
>>
"o
O.
II
I
<
Cu
• *s
U
§
g
u
U
I
•s
o
•o
U
-------
otherwise compared with a comparable group in an unexposed area, and the prevalence of adverse
effect was compared between the two groups. This type of study is more rigorous than geographical
correlation studies, because it is possible to control for confounding factors, such as age, ethnicity,
income, and smoking habits. We evaluated 7 case-control studies; these are considered the most
rigorous of the study types evaluated. Case-control studies identify a group of people with the disease
or effect of interest (the cases) with a comparable group that does not have the effect (controls), and
then compare the exposure histories of the two groups. Four prevalence studies (which only reported
an effect, and did not include control groups) were also included. Several studies included multiple
phases, with different types of studies methodologies applied in the different phases.
As shown in Table 4-2, we located similar numbers of studies evaluating exposure to metals
and to organic compounds. There were 14 studies in the former category and 17 studies in the latter
category. Several of these were studies that evaluated exposure to both metals and organic
compounds; 7 studies evaluated both organic and inorganic HAPs or did not specify the exposure
type. An additional study attributed the observed effects to silica fibers. We did not review any
studies that specifically evaluated effects of ambient exposure to radionuclides, although radiation
exposure also occurred in the study of Revich (1995).
In our identification of articles to review, we emphasized literature focusing on the U.S.; 14
of the studies evaluated U.S. populations. Since air quality standards in Canada and western Europe
are generally comparable to those of the U.S., results from the 11 studies of populations in those
countries are also likely to be applicable to the U.S. population. Due to the relatively high level of
air pollution in eastern european countries and parts of China, much recent research has focused on
these countries. However, because of the likelihood of concomitant exposure from contamination of
water or food, and the high level of criteria pollutants in many of these areas, it is often more
difficult to attribute any observed effects reliably to HAP exposure. In addition, HAP exposure levels
may be higher than those observed in the U.S. We reviewed 5 studies from Eastern Europe and one
study from China.
4.1.2 Evaluation of Data Relating Adverse Health Effects and Exposure to HAPs
This section provides brief summaries of the results of the studies of greatest interest,
emphasizing novel exposures and endpoints, and studies that lend themselves to generalization of
exposure-affects relationships for HAPs. Detailed summaries of selected studies and evaluations of
the study conclusions are provided in Appendix B. In preparing this Appendix, we have not only
summarized the key findings reported by the study authors, but also provided an independent analysis
of the degree to which the data which are presented support the authors' conclusions regarding
adverse effect-exposure relationships.
Key factors in such an analysis are the qualitative and quantitative biological plausibility of the
reported effect. Qualitative plausibility addresses whether the reported effects are consistent with the
known toxic endpoints of the chemical(s) to which the population was exposed, based on data from
laboratory animals and occupational studies at higher exposure levels. Quantitative plausibility
addresses whether exposure level at which the effect is observed is consistent with concentration-
response relationships from other sources. Studies that do not meet the test of qualitative plausibility
should be viewed with suspicion, because it is very unusual for an organ or system that is a target at
low exposure concentrations to not be a target at higher levels. Addressing the issue of quantitative
47
-------
plausibility is more complicated, because concentration-response data may not be available for human
exposure to the chemical of interest. While extrapolation from animal data is routinely conducted for
risk assessment purposes, such extrapolations are conducted using methods that generally err on the
side of conservatism, and which may be accurate only within a very broad range of uncertainty
(perhaps spanning two or more orders of magnitude). However, because most epidemiological
studies of air toxics have been conducted with high-volume chemicals, occupational exposure has also
usually been elevated. Thus, at least some human dose-response data usually exists for the endpoint
of interest, although no-effect levels may not always have been determined.
Generally it is difficult to interpret studies that do not meet the test of biological plausibility in
terms of establishing a simple relationship between exposure to the agent being studied and the
occurrence of the observed adverse effects. Very often the explanation for the observed association
must be sought in the presences of some confounding factor that has not been included in the analysis.
Occasionally, the additional factor may be an unknown biological mechanism of the agent being
studies. However, strong data are needed in order to support the proposal of a new mechanism, and
such data are rarely available from epidemiological investigations.2* The following discussions
address these issues in evaluating the results of the epidemiologic studies.
One of the most extensively investigated connections between exposure to HAPs and health
effects is that between lung cancer and exposure of populations near smelters to arsenic. Several
studies addressing this relationship are summarized in Table 4-1 and 4-2 (Brown et al. 1984, Frost et
al. 1987, Pershagen 1985). These studies tend to show increased risk associated with exposure (or
exposure surrogates, such as distance), although the apparent increase is not statistically significant in
all cases. For example, Frost et al. (1987) found increased risk of lung cancer (of borderline
statistical significance) in a case-control study of a population near an arsenic-emitting smelter that
was conducted with women only in order to reduce confounding from occupational exposure.
However, there was no control for smoking and no effect was seen in the cross-sectional phase of
their study. Pershagen (1985) analyzed lung cancer data in residents near an arsenic-emitting smelter,
controlling for smoking status and occupational exposure. In the group that was not occupationally
exposed, there was an increased relative risk for both nonsmokers and smokers, but the increase
reached statistical significance only among the smokers. Hughes et al. (1988) reviewed over 10
studies investigating health effects (primarily lung cancer) in communities near arsenic-emitting
industries. They noted that about half of the studies reported significant adverse effects, while about
half of them reported no effect or decreased risk in the exposed populations. However, they noted
that many of the studies (particularly those that observed no effect) lacked sufficient statistical power
to detect the small increases in risk that would be expected, and they suggested that some small
increase in risk is likely. An association between inhalation-pathway arsenic exposure and lung
cancer meets the criterion for biological plausibility, in light of the carcinogenic activity of arsenic
demonstrated in occupational studies.
Nordstrom et al. (1978) found decreased birth weight in babies born to mothers who lived
close to an arsenic-emitting smelter. However, it is unclear if the magnitude of the decrease was
26 An occasional exception to this rule is in the case of very rare (low-probability) effects of toxic pollutant
exposures. In such cases, large-population epidemiologic studies may be the only way to detect adverse effects
that are not apparent in animal or occupational studies.
48
-------
clinically significant (Hughes et al. 1988). It appears likely that some increased risk of lung cancer
occurs near arsenic-emitting smelters, but the size of the increased risk remains unclear.
Several studies investigated a possible association between vinyl chloride emissions and
central nervous system birth defects (Edmonds et al. 1978, Rosenman et al. 1989, Theriault et al.
1983). While all three studies reported some positive result, each is limited by internal
inconsistencies or potential confounding that precludes strong inferences about causality from being
drawn. Two of the studies (Edmonds et al. 1978, Rosenman et al. 1989) did not control for the
mother's (or father's) smoking or drinking habits, exposures that increase the risk of birth defects.
Edmonds et al. (1978) observed a significant decrease in birth defects incidence in populations near a
PVC factory during a period of generally decreasing emissions, but the decrease in incidence
preceded the actual decrease in emissions from the plant being studied. Theriault et al. (1983) found
that the increased incidence of birth defects correlated with seasonal variation in exposures, but not
with distance to the plant. Rosenman et al. (1989) reported increased birth defects at exposure levels
several orders of magnitude lower than those reported to cause developmental toxicity in mice (John
et al. 1981), raising questions about the biological plausibility of the observation, or possibly about
the reported exposure levels. Overall, these studies provide insufficient data to conclude that there is
a relationship between ambient exposure to vinyl chloride and central nervous system birth defects27.
Several other studies were reviewed that investigated the relationship between ambient HAP
exposures and neurological or neurobehavioral symptoms. The most notable of these were two recent
studies conducted of populations exposed near Superfund sites (Dayal et al. 1995, Kilburn and
Warshaw 1995). In a well-controlled cross-sectional study, Kilburn and Warshaw (1995) found lower
scores on objective neurophysiological tests related to balance and response to visual stimuli, and on
neuropsychological tests related to pattern recognition and manual skill tasks. They were unable to
establish an exposure-response relationship, but this may have been due to uncharacterized variability
in the meteorological conditions which affected the relationship between distance from the site and
exposure. A variety of organic chemicals was reported at the site, including aromatics, aliphatics,
chlorinated compounds, and PCBs. Dayal et al. (1995) evaluated the self-reported incidence of
symptoms in a population near a Superfund site, comparing the incidence in the most-exposed groups
with the incidence in the least-exposed group. There was no objective assessment of health effects
and no control for confounding factors. Nonetheless, significant differences were observed between
the two exposure extremes, most notably in the prevalence of learning difficulties, unusual fatigue,
nervousness, and numbness in extremities. Exposure to chemicals that migrated off-site into drinking
water may also have occurred in both of these studies.
In an example of the state of the art in exposure monitoring and the use of biomarkers,
Binkova et al. (1995) measured PAH DNA adducts in a group of women in the Czech Republic who
worked outdoors for about 8 hours per day. Personal exposure monitoring was used, allowing both
indoor and outdoor exposure to be evaluated; exposures to both respirable particles (<2.5 j*m) and to
PAHs were measured. Levels of DNA adducts in white blood cells were increased immediately after
days of high measured PAH exposure. This study demonstrated that DNA adducts can be used as
biomarkers of exposure, reflecting short-term exposure levels. In addition, DNA adducts can be used
as biomarkers of effect, because, if unrepaired, they can lead to gene mutations, which can ultimately
27 It should be noted, however, that these studies by no means rule out to possibility of such a relationship.
49
-------
lead to cancer. However, due to the multiple steps from gene mutation to tumor development, DNA
adducts and gene mutations are best viewed as indicating carcinogenic potential, rather than indicating
actual risk of cancer. As discussed in Section 3.2.4, biomarker studies, especially of DNA adducts,
are a fertile current field of research. Only a few primary studies on this topic were obtained and
reviewed for this project.
In another biomarker study, Bustueva et al. (1994) measured ambient exposure to airborne
cadmium in three Russian cities, as well as using hair and urinary cadmium levels as biomarkers of
exposure. Urinary cadmium is a reliable indicator of recent cadmium exposure, as shown by several
occupational studies. The presence of (3-2-microglobulin in urine (termed proteinuria) was used as a
biomarker of effect, indicating impaired kidney function. Although no significant effect was seen in
the general population, this was probably due to the small sample size and resulting low statistical
power of the study. Proteinuria (/3-2-microglobulin levels above 250 ngfL) was observed only in the
exposed population. Collecting biological samples and conducting laboratory testing, as in this study,
is more labor-intensive than doing epidemiological investigations using disease registries. However,
because proteinuria is a well-characterized effect of cadmium exposure, and both exposure and effect
biomarkers can be monitored by urinalysis, this is an area of potentially fertile future investigation
where high exposure to cadmium is expected.
4.1.3 Directions of Recent and Ongoing Studies
As discussed in the preceding sections, studies of the relationship between toxic air pollution
exposure and adverse health effects are proceeding on a number of fronts. In section 3, we discussed
how the earlier efforts to address this question employed techniques such as follow-up studies of
geographic patterns of disease (particularly cancer), emissions inventories, exposure and risk
assessment studies, biomarkers studies of selected pollutants, and focused studies of particular classes
of toxic air pollutant sources. Much attention was paid initially to the well-studied and common
metallic pollutants cadmium, mercury, and to lead and the other criteria pollutants, or other general
indicators of air quality. In addition, U.S. studies are being followed up by similar risk assessments
in Europe, particularly in Sweden, Denmark, and in eastern Europe and the ex-Soviet Union.
Individually, these various studies have provided data which contribute to the growing understanding
of the relationship between air pollution exposure and adverse affects, on both the qualitative and
quantitative level.
Classes of Sources Being Studied
Recently, the pattern of investigations of adverse effects of air toxics have taken a slightly
different course than earlier studies, and new avenues of investigation have been pursued. As noted
in Sections 4.1.1 and 4.1.2, studies of disease patterns around metals smelting operations are still
being performed, both because the association between these sources and adverse effects has been
well-established, and because of the relative ease with which metals pollution can be measured in air
and soils. Recent advances in this area have included, for example, the detection of adverse
reproductive outcomes (Nordstrom et al. 1978), and elevated cancer risks (Frost et al. 1987), among
women living near smelters. These findings help to rule out occupational exposures as the sole
determining factor in the occurrence of adverse effects. Also, studies of metals processing operations
(and other pollutant sources) are coming out of eastern Europe and the ex-Soviet Union (Mizernitskiy
et al. 1993, Norseth 1994, Bustueva 1994), helping to broaden the geographic base of these findings.
50
-------
The other major class of sources which is showing up frequently in the recent literature is
hazardous waste disposal sites. Beginning in the late 1980's, studies of adverse health effects near
such sites began to appear, including U.S. studies conducted by ATSDR and others, as well as a
number of foreign studies. In many instances, the major classes of pollutants present at the sites are
volatile organic chemicals, and in some cases these studies include data on both adverse health
outcomes and air concentrations of pollutants. Several of these studies, which were reviewed in
Section 4.1.1, have found an association between measures of air pollutant exposures and health
outcomes of various types.
Surprisingly, the amount of epidemiological information available concerning the relationship
between air toxics releases from chemical and petrochemical industry operations and adverse health
effects is not very great, in relation to the number of such facilities which are operating in the U.S.
Attempts to confirm the relatively high risks predicted to be associated with emissions from these sites
using epidemiologic studies have generally not been successful. One of the most well-studied
examples is PVC polymerization plants, where three studies have all found only equivocal evidence of
a connection between vinyl chloride releases and increased incidence of birth defects (Rosenmann et
al. 1989, Theriault et al 1983, and Edmonds et al. 1978). Similarly, Zhao et al. (1994) found
biomarkers of exposures (altered enzyme levels) in populations exposed to vinyl chloride in China,
but no other adverse effects. The positive results of two other studies of the relationship between
industrial chemical emissions and adverse effects from multiple sources (the Kanawha Valley study of
Ware et al. (1993) and the mortality study of an industrialized area near San Francisco by Kaldor et
al. (1984)), have been seriously questioned by other investigators (Collins et al. 1994, and Wong and
Bailey 1993, respectively). Two additional studies of potential health effects near petrochemical
facilities are included in Appendix B. The first, by Kilburn and Warshaw (1995) quantifies the
neurobehavioral deficits found in a population living near an oil reprocessing facility which is also a
Superfund site. There is no indication in this study whether emissions were due to plant operations or
to improper waste disposal. The other study, by Woodall et al. (1992) documents increases in the
frequency of emergency room visits in a hospital near a petrochemical facility on days when "upsets"
(increased emissions) occurred, but there are no measured pollutant levels to support this finding, and
the major pollutants emitted from the facility are "criteria" pollutants and not HAPs.
Emissions from mobile sources have also been evaluated as a potential cause of adverse health
effects, primarily from a risk assessment perspective. Releases of POMs and 1,3-butadiene from
mobile sources have contributed significantly to the aggregate cancer risks in EPA's major risk
assessment studies, as discussed in Section 3.2.
Another source of toxic air pollutant exposure which has been subject to a large amount of
research effort recently is domestic wood smoke. Efforts have been directed both at characterizing
exposures (Larson and Koenig 1994) and at evaluating the toxic and mutagenic potential of this source
class (Lewtas and Lewis 1992, Cupitt et al. 1994), which is a major contributor to the total mass of
air pollutants generated in some airsheds. Evaluation of this source class is still apparently at the
exposure and risk assessment stage, however, and we were not able to find any published studies
which contained both exposure data and information on adverse effects.
Types of Effects Bein£ Studied
51
-------
As noted in section 4.1.1, a number of studies have recently been published in two areas
which had not previously been addressed. In the area of neurobehavioral assessment, a number of
newly developed and previously standardized techniques for assessing neurological and psychological
impairment have been applied to the issue of air toxics. Previously, these methods were used mostly
in the clinical or occupational setting. More recently, the results of a number of these studies, for
example Kilburn and Warshaw (1995) and Dayal et al. (1995), have suggested that adverse
neurological effects of toxic air pollutant exposures are occurring in populations residing near some
toxic waste disposal sites.
The other types of effects which are becoming more widely studied are cytogenetic and
mutagenic impacts. The studies we have identified include the use of mutagenesis screening tests to
characterize the mutagenic activity of paniculate air pollutants, measurement of DNA-pollutant
adducts in populations exposed to toxic air pollutants, and the measurement of chromosomal
aberrations in populations exposed to volatile organic pollutants. None of these efforts has yet
formally linked genetic damage to clinically detectable adverse effects, such as increased cancer risks
or reproductive effects, so these studies still fall into the category of "biomarkers".
4.2 Summary: Impact of Recent Studies on Establishing the "Qualitative Relationship"
None of the studies reviewed during this effort significantly changes the general picture of the
relationship between exposure to toxic air pollutants and the occurrence of adverse human health
effects. None of the recent epidemiologic studies provides a comprehensive picture of population
risks on the national level, even for one pollutant, let alone for all pollutants. Rather, the
epidemiologic studies which were reviewed focus on individual aspects of the problem (one or a few
pollutants, effects in relatively small regions or populations) and, like the earlier studies, contribute
incrementally to a broader understanding of air toxics health effects. Aggregate estimates of
population risk must continue to rely on the relatively early national studies of cancer incidence
patterns, and on risk assessment-based estimation procedures, as discussed in Section 3.2
The studies we have reviewed, however, do shed additional light on the qualitative
relationship between air toxics exposures and adverse health effects by (1) providing confirmatory
examples of adverse effects of pollutants and sources previously identified, (2) by characterizing some
new effects of air toxics exposure which had not previously be measured, and (3) by identifying some
source classes which had not previously been associated with confirmed adverse effects. Perhaps the
most significant impact of the studies which we reviewed has been to provide additional support for
the idea that certain types of noncancer adverse effect (e.g. neurological and behavioral impairment,
cytogenetic damage), previously seen only in occupational studies, also could be measured in the
general population following exposures to lower levels of toxic pollutants, primarily volatile organics.
The studies reviewed also provided an additional information related to air toxics adverse
effects, particularly on the international level. Some of the important findings of our reviews of these
studies include:
• The most successful recent attempts to document the relationship between HAP
exposures and adverse health effects in the U.S. and western Europe have generally
been either case-control studies, or cross-sectional studies focused on smaller
populations near discrete air pollution sources. Studies of geographic patterns of
52
-------
disease incidence over large areas are becoming less common, both because of the
unavailability of disease incidence data and because of the lack of HAP exposure data
for large areas. Studies of larger areas may be of limited value in most instances
because the increased statistical power gained from evaluation large populations is
offset by reduced ability to detect very low response rates in the less-exposed
segments of the population. Regional studies in the U.S. (the IEMP studies, including
the Kanawha Valley Study), have had only limited success in demonstrating exposure-
effect relationships, primarily because of a lack of exposure data.
• In eastern Europe and the ex-Soviet Union, by contrast, large population and regional
studies are still being undertaken. The comparative success of these studies often
reflects the higher levels of pollutants in these areas, and the correspondingly higher
levels of adverse effects. While these studies help to confirm general air pollution-
disease relationships, the relatively high pollution levels studied, and the highly
variable quality of the studies themselves, has limited their generalizability to U.S.
populations.
• On the whole, the epidemiologic case for a relationship between measurable adverse
effects and exposures to HAPs has been made more successfully for toxic inorganic
pollutants than for organic pollutants. This situation arises both out of the
comparative ease in measuring exposure levels to the inorganic pollutants, and from
the fact that arsenic and cadmium (which are among the metals most frequently
studied) cause cancer, the incidence of which in a given population can be followed
relatively easily through tumor registries and death certificates28. This situation has
been changing recently, through the development of refined exposure measurement
techniques for organics, the measurement of biomarkers of organic exposures, and
through the application of testing methods which address neurobehavioral effects of
exposures to organic chemicals which had not previously been measurable.
• The studies reviewed vary in the extent to which they address potential confounding
factors, such as lifestyle (smoking, drinking), occupation, socioeconomic factors,
sensitive populations, and exposures to multiple pollutants. The quality of the best
studies (of which there are only a few) is such that they can usually make a strong
case by themselves for air pollution-adverse effect relationships, where such
relationships are found. In addition, a few agents (toxic metals) have been subject to
multiple studies which together show a generally consistent relationship. Indoor air
pollution is a major confounding factor which is rarely addressed in studies of the
adverse effects of ambient air toxics exposures. As noted previously, indoor
exposures have been shown to contribute a large proportion of the total air toxics
exposures for typical individuals, and they could thus potentially obscure or distort
apparent relationships between ambient air toxics exposures and adverse effects.
28 While risk assessment studies (Section 3.2.5) have predicted that a substantial proportion of all of the
cancer risks associated with HAP emissions (particularly from mobile sources) can be attributed to organic
pollutants, so far epidemiologic studies have not been able to confirm this prediction in the general population.
53
-------
Taken together, these observations point to a gradually changing paradigm for epidemiological
studies of the relationships between air pollution-adverse effects and adverse health effects. If recent
trends continue, studies in the future will tend to be focused on small exposed populations or
geographical areas, and on single sources or small numbers of sources. Case-control studies may
become more prominent than in the past as a tool for following up suggestions of causality, and more
studies may incorporate personal monitoring or microenvironmental modeling of exposures to account
for indoor and outdoor exposures. Exposure studies may be supported to a greater extent by the
measurement of biomarkers in the exposed populations. Studies incorporating these features would
seem to have the most promise in elucidating the relationships between hazardous air pollutant
exposures and adverse effects for individual sources or populations.
Whatever the methodologies which are employed, an integrative framework will be needed
that incorporates the existing knowledge base into a comprehensive picture of the relationship between
air toxics exposures and adverse effects. The elements of this integrative approach will continue to
include direct epidemiologic studies of populations exposed to air toxics (the subject of this report), as
well as toxicologic and dose-response considerations from animal and occupational studies,
refinements in exposure assessment and modeling methods, and risk assessment methods for
extrapolating the results of individual studies to larger populations.
54
-------
5. MAJOR DATA GAPS AND OPPORTUNITIES FOR FURTHER RESEARCH
This section summarizes what we found to be some the major data gaps associated with
characterizing the relationship between air toxic air pollutant exposures and human health effects, and
suggests approaches to filling these gaps.
5.1 Major Data Gaps
As discussed throughout this report, the ability to document and characterize the relationship
between air toxic exposures has always been limited by shortages of certain kinds of information and
limitations on our understanding of the physiochemical and biological mechanisms through which air
toxic adversely affect human health. Our review of recent studies sheds light on the potential for
addressing some of these data gaps, which have historically limited our ability to evaluate air toxics-
adverse effect relationships, including the following:
Lack of Integrated Exposure and Health Data
As noted in Section 3.0, both exposure data and health data status or outcome are needed to
support epidemiologic studies of the relationship between air toxics exposures and adverse health
effects. The shortage of exposure data and health data in general are both major obstacles to such
studies on the national level. There are no national data bases of HAP monitoring data (Section
3.2.3) nor are there any comprehensive national data bases on the potential non-cancer health impacts
(Section 3.2.2 and see below). Thus, for investigators without the resources to gather health or
ambient concentration data, studies of the health impacts of HAP exposure are limited to those
pollutants, areas and populations for which health status, biomarkers or exposure data are fortuitously
available, or for which they can be obtained cheaply. The recent studies in the literature for the most
part reflect this pattern, with most of the recent studies being limited to single exposure sources, or to
relatively small exposed populations.
The lack of integrated exposure and effects data will continue to be a major limitation of the
ability to identify adverse effects of HAPs for the foreseeable future. The reasons for this include
both the cost and difficulty of tracking the health status of large populations, and the difficulty and
costs of evaluating exposures to the large number of HAPs. This difficulty can only be overcome to
some extent by the biomarker techniques which are beginning to be applied to studies of HAP-health
effect relationships.
Lack of Indoor Air Pollutant Exposure Information
Another key data gap limiting the ability to relate air toxics exposures to health effects is the
lack of indoor exposure data in most cases. Since the majority of individuals spend most of their time
(usually 80 percent or more) indoors, and since pollutant concentrations in indoor air tend to be quite
different from (and higher than) those outdoors, studies which do not take indoor air quality into
account will have difficulty in elucidating the true relationship between air toxics exposures and
effects. It is not clear that there is a simple general solution to this problem. Instead, indoor-outdoor
pollutant relationships need to be addressed in the context of individual health effects studies.
55
-------
Lack of Non-Cancer Disease Registries
The most important data gap repeatedly identified in the literature, and obvious from the
nature of the studies which were reviewed, is the lack of nationally-standardized and comprehensive
disease registries for adverse effects of toxics exposures other than cancer. Particularly mentioned as
key data gaps are the lack of information on birth defects and reproductive outcomes (Shy 1993).
Currently, studies of these impacts require investigators to obtain access to local or state health status
information, whose availability is highly variable from state to state, or to obtain information from
hospital or other medical records. This difficulty is less of a concern for case-control studies, but can
severely limit the ability to do large-population cohort analyses or cross-sectional studies.
For some diseases (neurological and behavioral disorders, and less severe acute effects such as
respiratory irritation), it is unlikely that any kind of nationally standardized registry will ever be
practical, and thus other approaches to addressing these impacts must be found. Some of the studies
reviewed described approaches to addressing these issues, but most require some form of survey or
testing approach to support each study of adverse effects.
Data Gaps Related to Exposure-Response Relationships and Toxicological Mechanisms
There are a number of other "data gaps," which are more gaps in mechanistic knowledge than
gaps in information. These include the uncertainties related to potential adverse effects of organic
pollutant exposures in very sensitive individuals (Neutra et al. 1991), and whether adverse effects
occur in "normal" individuals at exposures below generally accepted "safe" levels, as suggested by at
least one of the studies reviewed (Ware et al. 1993). If low-exposure effects turn out to be a general
phenomenon, and the biochemical mechanism for such effects can be found, then the case for the
occurrence of adverse effects at observed ambient pollutant concentrations becomes much stronger,
and potential routes to avoidance of adverse effects in sensitive individuals may be identified.
Another technical issue which can only be addressed by fundamental advances in scientific
knowledge is the nature of the relationship between specific cytogenetic biomarkers (DNA adduct
formation, for example) and clinical adverse effects (for example, increased cancer risks). The
potential utility of genotoxic endpoints (e.g. chromosome abnormalities) as predictors of cancer risks
also needs to be explored. If such relationships can be established, then the biomarkers studies
become much more convincing as indicators of adverse health effects. The same is true of any
studies which advance the state of knowledge concerning dose-response relationships at low exposures
levels for carcinogenic and noncarcinogenic chemicals. Uncertainty in this area is a major reason
why risk assessment studies of air toxics adverse effects are subject to question.
5.2 Potential Opportunities for Additional Research
This section identifies a number of practical steps for improving the nature of the information
available related to toxic air pollutant adverse health effects. The topics discussed below are limited
to issues that can be addressed by OAQPS without substantial additional commitment of resources.
They are aimed primarily at continued exploration of easily-accessible information, and not at the
actual performance of epidemiological, monitoring, or laboratory studies. The latter types of
investigation may be desirable in and of themselves, but recommending priorities in these areas are
beyond the scope of this review. The following activities may be fruitful.
56
-------
Complete the Review of Previously Identified Literature Studies. The limited resources
available for this study did not allow all of the sources of information which were identified to be
fully reviewed, or, in some cases, even to be procured. There are a number of additional studies that
could be obtained and reviewed:
• Additional ATSDR documents related to health effects from hazardous wastes sites;
• A number of recent studies of neurobehavioral and cytogenetic impacts of air toxics
that were identified as secondary references in the existing bibliography;
• A number of other studies in various areas identified late in the effort.
Procuring and reviewing these studies, and incorporating them into the report as appropriate,
would improve the overall coverage and completeness of our effort.
Finish Efforts to Interview Government and Academic Experts. As pointed out in Section 2,
in the limited time allocated to this project, we were not able to contact all of the Federal and State
regulatory personnel who were identified as experts, and we talked to only a small proportion of the
academic experts identified. The individuals who we were able to contact generally supplied us with
useful information on local efforts, and unpublished and ongoing studies of air pollution adverse
effects. Continuing this effort could provide a more comprehensive picture of current research.
Identify and Review Potential Health Outcome Data From NCHS. The National Center for
Health Statistics (NCHS) has provided data to support a number of efforts to characterize the
relationship between air toxics exposures and adverse health outcomes, notably the NCI cancer
mapping efforts. Information from previous EPA studies (primarily the 1990 Screening Evaluation)
indicated that NCHS has made some efforts in the past, on an ad hoc basis, to develop detailed
information on the geographic patterns of respiratory disease and mortality, and on other air toxics-
related adverse effects. It might be worthwhile to get more information on these and more current
NCHS efforts which could provide useful information on air toxics-disease relationships.
Follow-Up of Birth Defects and Reproductive Outcomes Data Sources. We did not
systematically evaluate the quality of national and state data sources related to birth defects
reproductive outcomes. Getting more information on the states who have birth defects registries,
could greatly improve our coverage of this issue. This would fit in with the strategy of continuing to
pursue academic data sources, as the UNC - School of Public Health group (Professor Shy) is
reportedly quite active in this area, and may have developed or used some of these data bases.
Review Literature and Databases Related to Health Outcomes from Hospitals. HMOs.
Insurers, and Other Health Care Providers. We did not systematically investigate these potential data
sources. It would probably be useful to do a more focused literature survey in the public health
databases, and to contact industry groups, such as the American Hospital Association, insurance
groups, and other health care organizations. This effort would help to greatly expand our knowledge
of past efforts to link data from these sources to environmental factors, and potential data sources for
doing so in the future.
57
-------
Review Selected Studies of Indoor Air Toxics and Indoor-Outdoor Air Pollutant
Relationships. The current study has touched only indirectly on indoor-outdoor air pollutant
relationships, and how they bear on the issue of establishing the relationship between air toxics
exposures and adverse health effects. A modest effort could help firm up our understanding in two
key areas; first, how studies of "sick building syndrome" and related health issues has affected the
understanding of the toxicology of toxic air pollutants and second, how studies of indoor air quality
can clarify the relationship between outdoor ambient exposures and health effects.
Conduct a Symposium on the State of Knowledge Regarding the Adverse Effects of Toxic Air
Pollutants. When this report is finalized it could serve as discussion document for academic,
government and industry professionals who are studying air toxics-adverse effects relationships. A
symposium built around response to this report could address topics such as how the recent studies
change the prevailing view of the nature or magnitude of the health impacts of air toxics exposures,
and how the relationships could best be further elucidated and how data gaps could be addressed.
58
-------
6.0 REFERENCES CITED
Baker, DE, S Greenland, J Mendlein, and P Harmon. 1988. A health study of two communities
near the Stringfellow waste disposal site. Arch Environ Health 43:325-34.
Bencko, V, K Symon, V Chladek, and J Pihrt. 1977. Health aspects of burning coal with a high
arsenic content, II. Hearing changes in exposed children. Environ. Res. 13: 386-395.
Binkova, B, J Lewtas, I Miskvova, J Lenicek, and R Sram. 1995. DNA adducts and personal air
monitoring of carcinogenic polycyclicaromatic hydrocarbons in an environmentally exposed
population. Carcinogenesis (ENGLAND) 16(5): 1037-46.
Blot, WJ and JF Fraumeni Jr. 1975. Arsenical air pollution and lung cancer. Lancet 2:142-4.
Brown, LM, LM Pottern, and WJ Blot. 1984. Lung cancer in relation to environmental pollutants
emitted from industrial sources. Environ. Res. 34:250-261.
Bustueva, KA, BA Revich, and LE Bezpalko. 1994. Cadmium in the environment of three Russian
cities and in human hair and urine. Arch. Environ. Health (USA). 49(4): 284-288.
Butler, JP, GB Post, PJ Lioy, JM Waldman, and A Greenberg. 1993. Assessment of carcinogenic
risk from personal exposure to Benzo(a)pyrene in the Total Human Environmental Exposure Study
(THEES). J. Air Waste Manage. Assoc. 43(7): 970-7.
Collins, JJ and JM Ramlow. 1994. Respiratory and irritant health effects of ambient volatile organic
compounds. The Kanawha County Health Study's Letter. Am. J. Epidemiol. 140(1): 72-75.
Cordier, S, G Theriault, and H Iturra. 1983. Mortality patterns in a population living near a copper
smelter. Environ. Res. 31:311-322.
Cupitt, LT, WG Glen, and J Lewtas. 1994. Exposure and risk from ambient particle-bound
pollution in an airshed dominated by residential wood combustion and mobile sources. Environ.
Health Perspect. 102, Suppl 4: 75-84.
Dayal, H, S Gupta, N Trieff, D Maierson, and D Reich. 1995. Symptom clusters in a community
with chronic exposure to chemicals in two superfund sites. Arch. Environ. Health 50(2): 108-11.
Dockery, DW, CA Pope III, X Xu, JD Spengler, JH Ware, ME Fay, BG Ferris, and FE Speizer.
1993. An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med.
329(24): 1753-9.
Edmonds, LD, CE Anderson, JW Flynt Jr., and LM James. 1978. Congenital central nervous
system malformations and vinyl chloride monoma exposure: a community study. Teratology 17: 137-
142.
59
-------
Frost, F, L Harter, S Milham, R Royce, A Smith, J Hartley, and P Enterline. 1987. Lung cancer
among women residing close to an arsenic emitting copper smelter. Arch. Environ. Health 42: 148-
152.
Fraumeni, JF Jr. 1987. Etiologic insights from cancer mapping. Keynote lecture. Princess
Takamatsu Symp 18: 13-25.
Geschwind, S, J Stolwijk, M Bracken, E Fitzgerald, A Stark, C Olsen, and J Melius. 1992. Risk of
congenital malformations associated with proximity to hazardous waste sites. Am J Epidemiol 135:
1197-1207.
Greaves, WW, WN Rom, JL Lyon, G Varley, DD Wright, and G Chiu. 1981. Relationship
between lung cancer and distance of residence from non-ferrous smelter stack effluent. Am J Ind
Med 2:15-23.
Greenwood, D. Research Triangle Institute. 1993. Summary and analysis of available air toxics
health effects data: final report. Research Triangle Park, NC: Research Triangle Institute. Prepared
for U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards. RTI
Report Number 94U-5510-020.
Grimalt, JO, J Sunyer, V Moreno, OC Amaral, M Sala, A Resell, JM Anto, and 3 Albaiges. 1994.
Risk excess of soft-tissue sarcoma and thyroid cancer in a community exposed to airborne
organochlorinated compound mixtures with a high hexachlorobenzene content. Int. J. Cancer 56(2):
200-3.
Haemisegger et al. 1985. The air toxics problem in the United States: an assessment of cancer risks
for selected pollutants. US Environmental Protection Agency, Washington, DC, May 1985.
Hallenborg, CP and N Marsh. 1991. Toxic substances in the environment affecting respiratory
function of people in Hawaii. Hawaii Med J 50(3): 101-7.
Hertzman, C, M Hayes, J Singer and J Highland. 1987. Upper Ottawa Street landfill site health
study. Environ Health Perspect 75: 173-95.
Hughes, JP, L Polissar, and G Van Belle. 1988. Evaluation and synthesis of health effects studies of
communities surrounding arsenic producing industries. Int J Epidemiol 17:407-413.
Jarup, L., C.G. Elinder and G. Spang. 1988. Cumulative blood-cadmium and tubular proteinuria: a
dose-response relationship. Int Arch Occup Environ Health 60: 223-229.
John, JA, FA Smith, BKJ Leong, and BA Schwetz. 1977. The effects of maternally inhaled vinyl
chloride on embryonal and fetal development in mice, rats, and rabbits. Toxicol Appl Pharmaco 39:
497-513.
Kaldor, J and J Harris. 1984. Statistical association between cancer incidence and major cause
mortality & estimated residential - Exposure to Air Emissions from Petroleum and Chemical Plants.
Environmental Health Perspectives 54: 319-332.
60
-------
Kilburn, KH and RH Warshaw. 1995. Neurotoxic effects from residential exposure to chemicals
from an oil reprocessing facility and superfund site. Neurotoxicol. Teratol. 17(2): 89-102.
Klemans, W, C Vleminckx, L Schriewer, I Joris, N Lijsen, A Maes, M Ottogali, A Pays, C
Planard, G Rigaux, Y Ros, M Vanderiviere, J Vandenvelde, L Verschaeve, P Deplaen, and T
Lakhanisky. 1995. Cytogenetic biomonitoring of a population of children allegedly exposed to
environmental pollutants - phase 2 - results of a three-year longitudinal study. Mutation Research:
Genetic Toxicology. 342(3-4): 147-156
Kunst, AE, CWN Looman, and JP Mackenbach. 1993. Determinants of regional differences in lung
cancer mortality in the Netherlands. Soc Sci Med (United Kingdom) 37(5): 623-631.
Lakhanisky, T, D Bazzoni, P Jadot, I Joris, C Laurent, M Ottogali, A Pays, C Planard, Y Ros, and
C Vleminckx. 1993. Cytogenic monitoring of a village population potentially exposed to a low level
of environment pollutants. Phase I, SCE analysis. Source: Mutation Research 319(4): 317-323.
Larson, TV and JQ Koenig. 1994. Wood Smoke: Emissions and Non-cancer Respiratory Effects.
Annu Rev Public Health 15: 133-56.
Laurent, C, T Lakhanisky, P Jadot, I Joris, M Ottogali, C Planard, D Bazzoni, JM Foidart, and Y
Ros. 1993. Increased sister chromatid exchange frequencies observed in a cohort of inhabitants of a
village located at the boundary of an industrial dumping ground: phase I. Cancer Epidemiol
Biomarkers Prev 2(4): 355-62.
Lewtas J, C Lewis, R Zweidinger, R Stevens, and L Cupitt. 1992. Sources of genotoxicity and
cancer risk in ambient air. Pharmacogenetics (England) 2(6): 288-96.
Lioy, P and JM Daisey. 1987. Toxic air Pollution: a comprehensive study of non-criteria air
pollutants. Chelsea, Michigan: Lewis Publishers.
Matanoski, G, E Landau, J Tonascia, C Lazar, EA Elliott, W McEnroe, and K King. 1981. Cancer
mortality in an industrialized area of Baltimore. Environ Res 25:8-28.
Milham, S. 1977. Studies of morbidity near a copper smelter. Environ Health Perspect. 19:131-
132.
Mizernitskiy, YL, MY Drozhzhev, MV Kostyuchyenko, and BA Revich. 1993. Pregnancy outcome
in an ecologically unfavorable industrial region. Reprod Toxicol 7(5): 510.
NAS/NRC. 1989. Biologic markers in reproductive toxicity. National Academy of Sciebces/
National Research Council. Washington, DC: National Academy Press, 15-35.
Neutra, R, J Lipscomb, K Satin, and D Shusterman. 1991. Hypotheses to explain the higher
symptom rates observed around hazardous waste sites. Environ Health Perspect 94: 31-8.
Nordstrom, S, L Beckman, and I Nordenson. 1978. Occupational and environmental risks in and
around a smelter in northern Sweden. I. Variations in birthweight. Hereditas 88:43-46.
61
-------
Nordstrom S, L Beckman, and I Nordenson. 1979. Occupational and environmental risks in and
around a smelter in northern Sweden. V. Spontaneous abortion among female employees and
decreased birth weight in their offspring. Hereditas 90:291-296.
Norseth, T. 1994. Environmental pollution around nickel smelters in the Kola Peninsula (Russia).
Sci. Total Environ. (Netherlands) 148(2-3): 103-8.
Otto, D, KH Hudnell, D House, L Molhave, and WS Counts. 1992. Exposure of Humans to a
Volatile Organic Mixture: I. Behavioral Assessment. Arch Environ Hlth 47:23-30.
Otto, D, I Skalik, P Subrt, J Ratcliffe, D House, and R Sram. 1995. Association of mercury
exposure with neurobehavioral performance of children in Bohemia. Submitted.
Ozonoff, D, A Aschengrau, and P Coogan. 1994. Cancer in the vicinity of a department of defense
superfund site in Massachusetts. Toxicology & Industrial Health. 10(3): 119-141.
Perera, FP and AM Jefferey. 1990. Molecular Epidemiology and Cancer Prevention, Cancer detect
Prev 14(6): 639-45.
Pershagen, G. 1985. Lung cancer mortality among men living near an arsenic emitting smelter. Am
J Epidemiol 122: 684-694.
Phillips, LJ. 1992. Regional Relationships between Toxin Releases and Environmental and Human
Exposure to Toxic Substances. Bull Environ Contam Toxicol 48(6): 795-802.
Revich, BA. 1995. Public health and ambient air pollution in arctic and subarctic cities of Russia.
Sci Total Environ (Netherlands) 160-161: 585-92.
Rosenman, KD, IE Rizzo, MG Cononmos, and GJ Halpin. 1989. Central nervous system
malformations in relation to polyvinyl chloride production facilities. Arch Environ Health 44(5):
279-82.
Rybicki, BA, CC Johnson, J Uman, and JM Gorell. 1993. Parkinson's disease mortality and the
industrial use of heavy metals in Michigan. Mov Disord 8(1): 87-92.
Sabsay, MI, FK Tetelyutina, NM Kanunikova, VI Zhdanova, and ZS Kornyaeva. 1993. Ecological
situation and congenital defects in Udmurtiya. Reprod Toxicol 7(5): 517.
Shy, CM. 1993. Epidemiological studies of neurotoxic, reproductive, and carcinogenic effects of
complex mixtures. Environ Health Perspect 101(Suppl. 4): 183-186.
Swaen, GMH and JJM Slangen. 1995. Gasoline consumption and leukemia mortality and morbidity
in 19 European countries: An ecological study. International Archives of Occupational and
Environmental Health (Germany). 67(2): 85-93.
Theriault, G, H Iturra, and S Gingras. 1983. Evaluation of the association between birth defects and
exposure to ambient vinyl chloride. Teratology 27: 359-370.
62
-------
US EPA. 1989a. US Environmental Protection Agency. Analysis of Air Toxics Emissions,
Exposures, Cancer Risks, and Controllability in Five Urban Areas. Vol I, Base Year Analysis &
Results. Document no. EPA-450/2-89-012a. Research Triangle Park, NC: US EPA Office of Air
and Radiation, Office of Air Quality Planning and Standards.
US EPA. 1990a. US Environmental Protection Agency. Cancer risk from outdoor exposure to air
toxics, Vol. 1. Final Report. Research Triangle Park, NC: US EPA Office of Air Quality Planning
and Standards.
US EPA. 1990b. US Environmental Protection Agency. Toxic Air Pollutants & Noncancer Health
Risks: Screening Studies. Research Triangle Park, NC: US EPA Office of Air Quality Planning and
Standards.
Wallace, LA. 1991. Personal exposure to 25 volatile organic compounds EPA's 1987 team study in
Los Angeles, CA. Toxicol Ind Health 7(5-6): 203-208.
Wallace, L, E Pellizzari, T Hartwell, M Rosenzweig, M Erickson, C Sparacino, and H Zelon. 1984.
Personal exposure to volatile organic compounds in breathing-zone air, drinkng water, food and
exhaled breath. Environ Res 35: 293-319.
Wallace, L, E Pellizzari, T Hartwell, H Zelon, C Sparacino, and R Whitmore. 1984. Analyses of
exhaled breath of 355 urban residents for volatile organic compounds. Indoor Air: Chemical
Characterization and Personal Exposure 4: 15-20.
Ware, JH, JD Spengler, LM Neas, JM Samet, GR Wagner, D Coultas, H Ozkayanak, and M
Schwab. 1993. Respiratory and irritant health effects of ambient volatile organic compounds : the
Kanawha county health study. American journal of epidemiology 137(12): 1287-1301.
Williams, FL, SA Ogston, and OL Lloyd. 1995. Sex ratios of births, mortality, and air pollution:
can measuring the sex ratios of births help to identify health hazards from air pollution in industrial
environments? Occup Environ Med (England) 52(3): 164-9.
Wong, O and WJ Bailey. 1993. Cancer incidence and community exposure to air emissions from
petroleum and chemical plants in Contra Costa County, California: A critical epidemiological
assessment. J Environ Health 56(5): 11-17.
Woodall, G, Jr, LH Kuller, EO Talbott, J Ramlow and B Salthouse. 1992. Epidemiological study
for the Kenova and Ceredo, West Virginia and Catlettsburg, Kentucky area, (Phase II). Final Report.
Prepared for USEPA Region IV, Air Compliance Branch. March 1992.
Yansheva, NY, IS Kireeva, and IA Chernichenko. 1993. The oncoepidemiologic investigation
under chemical carcinogens' regulation. Eksperimental'naya Onkologiya 15(5): 16-20.
Zhao, MY, CJ Ying, N Shao, Y Yang, CF Yang, L Shi, and WQ Liu. 1994. The study of health
effects of vinyl chloride air pollution on population. Biomedical and Environmental Sciences 7(2):
136-143.
63
-------
Zmirou, D, A Deloraine, P Saviuc, C Tillier, A Boucharlat, and N Maury. 1994. Short-term health
effects of an industrial toxic waste landfill: a retrospective follow-up study in Montchanin, France.
Archives of environmental health. 49(4): 228-238.
64
-------
APPENDIX A
REFERENCES IDENTIFIED
RELATING TO AIR TOXICS EXPOSURE
AND ADVERSE HEALTH EFFECTS
A-l
-------
* Andersen, O, H Autrup, AD Dayan, J Fenger, J Forslund, P Grandjean, P Gron, FP Jensen,
L Keiding, S Kjaergaard, JC Larsen, J Lewtas, E Lynge, H Moller, L Moller, PA Nielsen, N
Ostenfeldt, H Pilsgaard, C Portier, e. Poulsen, S Rastogi, T Skov, AS Thomsen, K Ulbak,
and A Osterlind. 1993. Report from the Working Group on Cancer and the
Non-Occupational Environment. Denmark, Pharmacol. Toxicol. Suppl. (Denmark) 72(1):
167-171.
* Austin, DF, V Nelson, B Swain, L Johnson, S Lum and P Flessel. 1984. Epidemiological
Study of the incidence of cancer as related to industrial emisssions in Conta Costa County,
California. Report to the Environmental Protection Agency. Pub. No. EPA-600-S1-84-008.
From Wong and Bailey, 1993.
* « ATSDR. 1986-1988. Biennial Report to Congress. Vol I & Vol II.
* ATSDR. 1993. Study of Symptom and Disease Prevalence. Caldwell Systems Hazardous
Waste Incinerator, Caldwell, North Carolina.
* ATSDR. 1995. Final Report Analytic Study to Evaluate Associations Between Hazardous
Waste Sites and Birth Defects.
* ATSDR. 1995. End Stage Renal Disease Study (Draft).
* ATSDR. 1995. Southbend Subdivision Health Outcomes Study. Harris County, Texas.
(Draft final report)
4 ATSDR. 1995. Disease and Symptom Prevalence Survey. Tucson International Airport Site,
Tucson, Arizona.
* Axelsson, G and R Rylander. 1980. Environmental chromium dust and lung cancer
mortality. Environ. Res. 23:469-476.
* Baker, DE, S Greenland, J Mendlein, and P Harmon. 1988. A health study of two
communities near the Stringfellow waste disposal site. Arch Environ Health 43:325-34.
* Barbone, F, M Bovenzi, F Cavallieri, and G Stanta. 1995. Air pollution and lung cancer in
Trieste, Italy. American Journal of Epidemiology (USA). 141(12): 1161-1169.
* Bartsch, H and L Tomatis. 1982. Mutagenecity and carcinogenecity of environmental
chemicals. Regul. Toxicol. Pharmacol. 2(2): 94-105.
* Bateman, B. Relative contributions of stationary & mobile sources to individual cancer risk
in the San Fransicso Bay Area.
Bencko, V and K Symon. 1977. Health aspects of burning coal with a high arsenic content.
I. Arsenic in hair, urine and blood in children residing in a polluted area. Environ. Res. 13:
378-385.
* = Reference was retrieved.
* = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-2
-------
Bencko, V, K Symon, V Chladek, and J Pihrt. 1977. Health aspects of burning coal with a
high arsenic content, II. Hearing changes in exposed children. Environ. Res. 13: 386-395.
Binkova, B, J Lewtas, I Miskvova, J Lenicek, and R Sram. 1995. DNA adducts and
personal air monitoring of carcinogenic polycyclicaromatic hydrocarbons in an
environmentally exposed population. Carcinogenesis (ENGLAND) 16(5): 1037-46.
Blair, A, R Saracci, PA Stewart, PB Hayes, and C Shy. 1990. Epidemiologic evidence on
the relationship between formaldehyde exposure and cancer. Scand. J. Work Environ. Health
(Finland) 16(6): 381-93.
Blair, A, SH Zahm. 1992. Clues to cancer etiology from studies of farmers. Scand J Work
Environ Health (Finland) 18(4): 209-15.
Blot, WJ, LM Brown, LM Pottern, and JF Fraumeni, Jr. 1986. Lung cancer mortality
among men living near an arsenic-emitting smelter. Am. J. Epidemiol. 124(1): 154-5.
Blot, WJ and JF Fraumeni Jr. 1975. Arsenical air pollution and lung cancer. Lancet
2:142-4.
Blot, WJ and JF Fraumeni Jr. 1994. Arsenic and lung cancer. In: Samet, J. M. (Ed.).
Lung biology in health and disease, Vol. 74. Epidemiology of lung cancer. XVI+543p.
New York, New York, USA; Basel, Switzerland: Marcel Dekker, Inc., p207-218.
Braback, L, A Breborowicz, S Dreborg, A Knutsson, H Pieklik, and B Bjorksten. 1994.
Atopic sensitization and respiratory symptoms among Polish and Swedish school children. In:
Clinical and experimental allergy. 24(9): 826-835.
Brooks, JH, KK Renz. 1993. Linking environmental and health care databases, assessing the
health effects of environmental pollutants. Int. J. Biomed. Comput. (England) 32(3-4): 279-
88.
Brown, LM, WJ Blot, SH Schuman, VM Smith, AG Ershow, RD Marks, and JF Fraumeni
Jr. 1988. Environmental factors and high risk of esophageal cancer among men in coastal
South Carolina. J. Natl. Cancer. Inst. 80(20): 1620-5.
Brown, LM, LM Pottern, and WJ Blot. 1984. Lung cancer in relation to environmental
pollutants emitted from industrial sources. Environ. Res. 34:250-261.
Bustueva, KA, BA Revich, and LE Bezpalko. 1994. Cadmium in the environment of three
Russian cities and in human hair and urine. Arch. Environ. Health (USA). 49(4): 284-288.
Butler, JP and TJ Kneip. 1987. An investigation of interurban variations in the chemical
composition and mutagenic activity of airborne paniculate organic matter using an integrated
chemical class; Bioassay System. Atmos. Environ. 21(4): 883-892.
* = Reference was retrieved.
4 = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-3
-------
Butler, JP, GB Post, PJ Lioy, JM Waldman, and A Greenberg. 1993. Assessment of
carcinogenic risk from personal exposure to Benzo(a)pyrene in the Total Human
Environmental Exposure Study (THEES). J. Air Waste Manage. Assoc. 43(7): 970-7.
Castilla, EE. 1994. Economic activity data as indicator of environment. Teratology
Cohen, Y, GE Anderson, LR Chinkin, G Pascoe, CE Schmidt, and A Winer. 1994. The
Valdez Air Health Study. RCAC Valdez Air Study Review Committee Detter]. Risk Anal.
(UNITED STATES) 14(6): 887-93.
* Cohen, MA and PB Ryan. 1991. Source receptor study of volatile organic compounds and
paniculate matter in the Kanawha Valley, (West Virginia, USA), II. Analysis of factors
contributing to VOC and particle exposures. Atmos. Environ. Part B Urban Atmos. 25(1):
95-108.
* Cohen, MA and PB Ryan. 1991. Source receptor study of volatile organic compounds and
paniculate matter in the Kanawha Valley, (West Virginia, USA), I. Methods and descriptive
statistics. Atmos. Environ. Part B Urban Atmos. 25(1): 79-94.
* Collins, JJ and JM Ramlow. 1994. Respiratory and irritant health effects of ambient volatile
organic compounds. The Kanawha County Health Study's Letter. Am. J. Epidemiol.
140(1): 72-75.
* Connelly, RR, R Spirtas, MH Myers, CL Percy, and JF Fraumeni Jr. 1987. Demographic
patterns for mesothelioma in the United States. J. Natl. Cancer Inst. 78(6): 1053-60.
* Corbo, GM, F Forastiere, V Dell'Orco, R Pistelli, N Agabiti, B De Stefanis, G Ciappi, and
CA Perucci. 1993. Effects of environment on atopic status and respiratory disorders in
children. Journal of allergy and clinical immunology. 92(4): 616-623.
* Cordier, S, G Theriault, and H Iturra. 1983. Mortality patterns in a population living near a
copper smelter. Environ. Res. 31:311-322.
* Cote, I and S Bayard. 1988. National Air Toxics Information Clearinghouse: qualitative
and quantitative carcinogenic risk assessment. Government Reports Announcements & Index
(GRA&I), Issue 03.
* Cupitt, LT, WG Glen, and J Lewtas. 1994. Exposure and risk from ambient
particle-bound pollution in an airshed dominated by residential wood combustion and mobile
sources. Environ. Health Perspect. 102, Suppl 4: 75-84.
* Dayal, H, S Gupta, N Trieff, D Maierson, and D Reich. 1995. Symptom clusters in a
community with chronic exposure to chemicals in two superfund sites. Arch. Environ. Health
50(2): 108-11.
* = Reference was retrieved.
4 = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-4
-------
Dayal, HH, S Khuder, R Sharrar, and N Trieff. 1994. Passive smoking in obstructive
respiratory diseases in an industrialized urban population. Environmental research: (New
York, NY). 65(2): 161-171.
Devesa, SS, WJ Blot, and JF Fraumeni, Jr. 1989. Declining lung cancer rates among young
men and women in the United States: a cohort analysis. J. Natl. Cancer Inst. 81(20):
1568-71.
Devesa, SS, WJ Blot, and JF Fraumeni, Jr. 1991. Declining lung cancer rates among young
men and women in the United States: a cohort analysis. J. Natl. Cancer Inst. 83(5): 368-70.
Devesa, SS, WJ Blot, BJ Stone, BA Miller, RE Tarone, JF Fraumeni Jr., et al. 1995.
Recent cancer trends in the United States. J. Natl. Cancer Inst. 87(3): 159-60, 175-82.
Devesa, SS, DT Silverman, JL Young, ES Pollack, CC Brown, JW Horm, CL Percy, MH
Meyers, FW McKay, and JF Fraumeni, Jr. 1987. Cancer incidence and mortality trends
among whites in the United States. J. Natl. Cancer Inst. 79(4): 701-70.
De vos Irvine, H, DW Lament, DJ Hole, and CR Gillis. 1993. Asbestos and lung cancer in
Glasgow and the west of Scotland. Br. Med. J. 306(6891): 1503-1506.
Dobrowolski, JW and B Smyk. 1993. Environmental risk factors of cancer and their
primary prevention. J. Environ. Pathol. Toxicol. Oncol. 12(1): 55-7.
Dockery, DW and CA Pope III. 1994. Acute respiratory effects of particulate air pollution.
Annu. Rev. Public Health. 15: 107-32.
Dockery, DW, CA Pope III, X Xu, JD Spengler, JH Ware, ME Fay, BG Ferris, and FE
Speizer. 1993. An association between air pollution and mortality in six U.S. cities. N.
Engl. J. Med. 329(24): 1753-9.
Doll, R. 1992. Are we winning the war against cancer? A review in memory of Keith
Durrani. Clin. Oncol. (R Coll Radiol) (England) 4(4): 257-66.
Doll, R. 1992. Health and the environment in the 1990s. Am J Public Health 82(7): 931-
32,933-41,941-43.
Doll, R. 1991. Progress against cancer: an epidemiologic assessment. The 1991 John C.
Cassel Memorial Lecture. Am J Epidemiol 134(7): 675-88.
Doll, R. 1991. Urban and rural factors in the aetiology of cancer. Int J Cancer 47(6):
803-10.
Doll, R and R Peto. 1981. The causes of cancer: quantitative estimates of avoidable risks of
cancer in the United States today. J Natl Cancer Inst 66(6): 1191-308.
* = Reference was retrieved.
* = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-5
-------
* Dosemeci, M, RN Hoover, A Blair, LW Figgs, S Devesa, D Grauman, and JF Fraumeni.
1994. Farming and prostate cancer among African-Americans in the southeastern, United
States. J Natl Cancer Inst 86(22): 1718-9.
* Dyomin, SN, LA Buldacov, IA Ternovsky, ZB Tokarskaya, TP Fomina, FD Tretyakov, GN
Ivanova, VN Shevchenko, VA Kolmogortsev, AG Uralshin, and I Suslova. 1994. Factors
effecting the morbidity in populations living in the vicinity of atomic industry plants. Sci.
Total Environ, (Netherlands) 142(1-2): 105-9.
* Edmonds, LD, CE Anderson, JW Flynt Jr., and LM James. 1978. Congenital central
nervous system malformations and vinyl chloride monoma exposure: a community study.
Teratology 17: 137-142.
* Ekbom, A, JK McLaughlin, BM Karlsson, O Nyren, G Gridley, HO Adami, and JF
Fraumeni, Jr. 1994. Pancreatitis and pancreatic cancer: a population-based study. Nyren J
Natl Cancer Inst 86(8): 625-7.
* Engel, RR, C Hopenhayn-Rich, O Receveur, and AH Smith. 1994. Vascular effects of
chronic arsenic exposure, a review. Epidemiologic Reviews 16(2): 184-209.
* Falk, RT, LW Pickle, ET Fentham, SD Greenberg, HL Jacobs, P Correa, and JF Fraumeni,
Jr. 1992. Epidemiology of bronchioloalveolar carcinoma. Cancer Epidemiol Biomarkers
Prev 1(5): 339-44.
* Folinsbee, LJ. Human health effects of air pollution. 1993. Environ Health Respect
100:145-56.
* Fraumeni, JF Jr. 1987. Etiologic insights from cancer mapping. Keynote lecture. Princess
Takamatsu Symp 18: 13-25.
* Frost, F, L Harter, S Milham, R Royce, A Smith, J Hartley, and P Enterline. 1987. Lung
cancer among women residing close to an arsenic
emitting copper smelter. Arch. Environ. Health 42: 148-152.
Geschwind, S, J Stolwijk, M Bracken, E Fitzgerald, A Stark, C Olsen, and J Melius. 1992.
Risk of congenital malformations associated with proximity to hazardous waste sites. Am J
Epidemiol 135: 1197-1207.
* Goldstein, B, J Gibson, R Henderson, J Hobbie, P Landrigan, D Mattison, F Perera, E
Pfitzer, E Silbergeld, and G Wogan. 1987. Biological markers in environmental health
research. In: Symposium on the Role of Biomarkers in Reproductive and Developmental
Toxicology, Washington, D.C., USA, January 12-13, 1987. Environ Health Perspect, 74(0):
3-10.
* Gray, GM and JD Graham. 1991. Risk assessment and clean air policy. Jorunal of Policy
Analysis and Management 10(2): 286-295.
* = Reference was retrieved.
4 = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-6
-------
Greater Tokyo Bureau of Hygiene. 1989. [Report concerning the effect of chromium in a
health survey (10-year survey)]. Report, Greater Tokyo, Bureau of Hygiene, Tokyo, Japan.
(Japanese)
Greaves, WW, WN Rom, JL Lyon, G Varley, DD Wright, and G Chiu. 1981.
Relationship between lung cancer and distance of residence from non-ferrous smelter stack
effluent. AmJIndMed 2:15-23.
Greenwood, D. Research Triangle Institute. 1993. Summary and analysis of available air
toxics health effects data: final report. Research Triangle Park, NC: Research Triangle
Institute. Prepared for U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards. RTI Report Number 94U-5510-020.
Grimalt, JO, J Sunyer, V Moreno, OC Amaral, M Sala, A Rosell, JM Anto, and J Albaiges.
1994. Risk excess of soft-tissue sarcoma and thyroid cancer in a community exposed to
airborne organochlorinated compound mixtures with a high hexachlorobenzene content. Int.
J. Cancer 56(2): 200-3.
Haemisegger et al. 1985. The air toxics problem in the United States: an assessment of
cancer risks for selected pollutants. US Environmental Protection Agency, Washington, DC,
May 1985.
Hallenborg, CP and N Marsh. 1991. Toxic substances in the environment affecting
respiratory function of people in Hawaii. Hawaii Med J 50(3): 101-7.
Hemminki, K and G Pershagen. 1994. Cancer risk of air pollution. Epidemiological
Evidence - Environ Health Respect 102(Suppl 4): 187-192.
Harris, RH, JH Highland, JV Rodricks, and SS Papadoupulos. 1984. Adverse health effects
at a Tennessee hazardous waste disposal site. Hazardous Waste 1:183-204.
Hertzman, C, M Hayes, J Singer and J Highland. 1987. Upper Ottawa Street landfill site
health study. Environ Health Perspect 75: 173-95.
Hoel, DG and DL Davis. 1992. Trends in cancer mortality in 15 industrialized countries
1969-86. J Natl Cancer Inst 84(5): 313-20.
Hughes, K, ME Meek, and R Burnett. 1994. Inorganic arsenic: evaluation of risks to health
from environmental exposure in Canada. J Environ Sci Health Part C Environ Carcinog
Ecotoxicol Rev (USA) 12(2): 145-159.
Hughes, K, M Meek, PKL Chan, J Shedden, S Bartlett, and LJ Seed. 1994. Nickel and its
compounds: Evaluation of risks to health from environmental exposure in Canada. J Environ
Sci Health Part C Environ Carcinog Ecotoxicol Rev (USA). 12(2): 417-433.
* = Reference was retrieved.
* = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-7
-------
Hughes, JP, L Polissar, and G Van Belle. 1988. Evaluation and synthesis of health effects
studies of communities surrounding arsenic producing industries. Int J Epidemiol 17:407-
413.
Jarup, L., C.G. Blinder and G. Spang. 1988. Cumulative blood-cadmium and tubular
proteinuria: a dose-response relationship. Int Arch Occup Environ Health 60: 223-229.
Jedrychowski, W. 1995. Review of recent studies from Central and Eastern Europe
associating respiratory health effects with high levels of exposure to 'traditional' air
pollutants. Environmental Health Perspectives (USA) 103(Suppl. 2): 15-21.
John, JA, FA Smith, BKJ Leong, and BA Schwetz. 1977. The effects of maternally inhaled
vinyl chloride on embryonal and fetal development in mice, rats, and rabbits. Toxicol Appl
Pharmaco 39: 497-513.
Kabat, GC. 1993. Recent Developments in the Epidemiology of Lung Cancer. Semin Surg
Oncol 9(2): 73-79.
Kaldor, J and J Harris. 1984. Statistical association between cancer incidence and major
cause mortality & estimated residential - Exposure to Air Emissions from Petroleum and
Chemical Plants. Environmental Health Perspectives 54: 319-332.
Kantor, AF, P Hartge, RN Hoover, and JF Fraumeni, Jr. 1985. Familial and environmental
interactions in bladder cancer risk. Int J Cancer 35(6): 703-6.
Kelly, TJ, R Mukund, CW Spicer, and AJ Pollack. 1994. Concentrations and
transformations of hazardous air pollutants. Environmental Science & Technology 28(8):
378A-387A.
Kilburn, KH and RH Warshaw. 1995. Neurotoxic effects from residential exposure to
chemicals from an oil reprocessing facility and superfund site. Neurotoxicol. Teratol. 17(2):
89-102.
Klemans, W, C Vleminckx, L Schriewer, I Joris, N Lijsen, A Maes, M Ottogali, A Pays, C
Planard, G Rigaux, Y Ros, M Vanderiviere, J Vandenvelde, L Verschaeve, P Deplaen, and
T Lakhanisky. 1995. Cytogenetic biomonitoring of a population of children allegedly
exposed to environmental pollutants - phase 2 - results of a three-year longitudinal study.
Mutation Research: Genetic Toxicology. 342(3-4): 147-156
Kleinjans, JC, JM van Maanen, and FJ van Schooten. 1993. Human respiratory disease:
environmental carcinogens and lung cancer risk. Ciba Found Symp. (Netherlands) 175:
171-8; discussion 178-81.
Kotesovec, F, J Brynda, J Pilatova, J Nozicka, and M Leixner. Mortality in the extremely
air polluted areas and its variability during a year. Czech-Republic.
* = Reference was retrieved.
* = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-8
-------
* Kraut, A, E Chan, PJ Lioy, FB Cohen, BD Goldstein, and PJ Landrigan. 1991.
Epidemiologic investigation of a cancer cluster in professional football players. Environ Res
56(2): 131-43.
* Kunst, AE, CWN Looman, and JP Mackenbach. 1993. Determinants of regional differences
in lung cancer mortality in the Netherlands. Soc Sci Med (United Kingdom) 37(5): 623-631.
* Lakhanisky, T, D Bazzoni, P Jadot, I Joris, C Laurent, M Ottogali, A Pays, C Planard, Y
Ros, and C Vleminckx. 1993. Cytogenic monitoring of a village population potentially
exposed to a low level of environment pollutants. Phase I, SCE analysis. Source: Mutation
Research 319(4): 317-323.
* Larson, TV and JQ Koenig. 1994. Wood Smoke: Emissions and Non-cancer Respiratory
Effects. Annu Rev Public Health 15: 133-56.
Laurent, C, T Lakhanisky, P Jadot, I Joris, M Ottogali, C Planard, D Bazzoni, JM Foidart,
and Y Ros. 1993. Increased sister chromatid exchange frequencies observed in a cohort of
inhabitants of a village located at the boundary of an industrial dumping ground: phase I.
Cancer Epidemiol Biomarkers Prev 2(4): 355-62.
* Lawson, AB and FL Williams. 1993. Applications of extraction mapping in environmental
epidemiology. Stat Med (England) 12(13): 1249-58.
Lechner, JF. 1994. Urban air carcinogens and their effects on health. Report No.:
CONF-9410236-2. 13p. Health and environment meeting, Mexico City (Mexico), 20-21 Oct
1994. Sponsored by Department of Energy, Washington, DC.
* Leuenberger, P, U Ackermann-Liebrich, J Schwartz, AP Pernuchoud, O Brandli, C
Schindler, and Sapaldia-team. 1994. Swiss study on air pollution and lund disease in adults.
Source: Meeting of the European Respiratory Society (ERS), Nice, France, October 1-
October 5, 1994. European Respiratory Journal 7(Suppl. 18): 178s.
* Lewtas, J. 1993. Airborne carcinogens. Source: Conference on environment and cancer,
prevention of cancer Aarhus, Denmark, May 10-13, 1992, Pharmacol Toxicol, 72(Supply 1):
S55-S63.
* Lewtas J. 1989. Assessment of Exposure and Potential Risk from Mutagenic Air Pollutants.
Environ Mol Mutagen 14(Suppl 15): 237.
Lewtas, J. 1993. Complex mixtures of air pollutants: Characterizing the cancer risk of
polycyclic organic matter. Environ Health perspect (USA). 100:211-218.
* Lewtas, J. 1989. Emerging methodologies for assessment of complex mixtures: application
of bioassays in the Integrated Air Cancer Project. Toxicol Ind Health 5(5): 839-50.
* = Reference was retrieved.
* = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-9
-------
Lewtas, J. 1994. Human exposure to complex mixtures of air pollutants. Toxicol Lett
(Netherlands) 72(1-3): 163-9.
Lewtas, J, JE Gallagher, and DM DeMarini. 1991. Environmental Mutagens and Risk
Assessment. Source: Govt Reports Announcements & Index (GRA&I), Issue 13.
Lewtas J, C Lewis, R Zweidinger, R Stevens, and L Cupitt. 1992. Sources of genotoxicity
and cancer risk in ambient air. Pharmacogenetics (England) 2(6): 288-96.
Lewtas, J, MG Nishioka, and BA Petersen. 1986. Identification and Comparative Risk
Assessment of Airborne Carcinogens from Combustion Sources. In: Govt Reports
Announcements & Index (GRA&I), Issue 10.
Lioy, P and JM Daisey. 1987. Toxic air Pollution: a comprehensive study of non-criteria air
pollutants. Chelsea, Michigan: Lewis Publishers.
Lippmann, M and J Lioy. 1985. Critical issues in air pollution epidemiology. Environ
Health Respect 62: 243-58.
Liteplo, RG and M Meek. 1994. E.3,3'-Dichlorobenzidine: Evaluation of risks to health
from environmental exposure in Canada. J Environ Sci Health-Environ Carcinog Ecotoxicol
Rev C12(2): 287(6).
Matanoski, G, E Landau, J Tonascia, C Lazar, EA Elliott, W McEnroe, and K King. 1981.
Cancer mortality in an industrialized area of Baltimore. Environ Res 25:8-28.
Maynard, RL, KM Cameron, R Fielder, A McDonald, and A Wadge. 1995. Setting air
quality standards for carcinogens: An alternative to mathematical quantitative risk assessment -
Discussion paper. Human and Experimental Toxicology (United Kingdom). 14(2): 175-186.
Meek, ME and PKL Chan. 1994. Xylenes: Evaluation of risks to health from environmental
exposure in Canada. J environ Sci Health Part C Environ Carcinog Ecotoxicol Rev (USA).
12(2): 545-556.
Meek, ME, PKL Chan, and S Bartlett. 1994. Polycyclic aromatic hydrocarbons: Evaluation
of risks to health from environmental exposure in Canada. J Environ Sci Health Part C
Environ Carcinog Ecotoxicol Rev (USA). 12(2): 443-452.
Meek, ME, M Giddings, and R Gomes. 1994. Monochlorobenzene: Evaluation of Risks to
Health from Environmental Exposure in Canada. J Environ Sci Health-Environ Carcinog
Ecotoxicol Rev CI2(2): 409(7).
Milham, S. 1977. Studies of morbidity near a copper smelter. Environ Health Perspect.
19: 131-132.
* = Reference was retrieved.
4 = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-10
-------
* Mizernitskiy, YL, MY Drozhzhev, MV Kostyuchyenko, and BA Revich. 1993. Pregnancy
outcome in an ecologically unfavorable industrial region. Reprod Toxicol 7(5): 510.
* Moller, H. 1993. Occurrence of carcinogens in the external environment: Epidemiological
investigations. Pharmacol Toxicol Suppl (Denmark) 72(1): 39-45.
* Moeller, L, D Schuetzle, et al. 1994. Future research needs associated with the assessment
of potential human health risks from exposure to toxic ambient air pollutants. Source:
Environmental Health Perspectives 102(Suppl 4): 193-210.
* Moolgavkar, SH, DW Dockery, and CA Pope III. 1994. Air pollution and mortality (3),
New Engl J Med 330(17): 1237-1238
* Mueller, NE and JF Fraumeni Jr. 1992. The emerging epidemic of non-Hodgkin's
lymphoma: current knowledge regarding etiological factors. Research directions and general
discussion. Cancer Res 52(Suppl 19): 5571s.
* NAS/NRC. 1989. Biologic markers in reproductive toxicity. National Academy of
Sciebces/ National Research Council. Washington, DC: National Academy Press, 15-35.
* Neutra, R, J Lipscomb, K Satin, and D Shusterman. 1991. Hypotheses to explain the higher
symptom rates observed around hazardous waste sites. Environ Health Perspect 94: 31-8.
* Newhook, R, G Long, ME Meek, RG Liteplo, P Chan, J Argo, and W Dormer. 1994.
Cadmium and its compounds: Evaluation of risks to health from environmental exposure in
Canada. J Environ Sci Health Part C Environ Carcinog Ecotoxicol Rev (USA). 12(2):
195-217.
* Ng, TP and WC Tan. 1994. Epidemiology of allergic rhinitis and its associated risk factors
in Singapore. International Journal of Epidemiology 23(3): 553-558.
* Nordstrom, S, L Beckman, and I Nordenson. 1978. Occupational and environmental risks in
and around a smelter in northern Sweden. I. Variations in birthweight. Hereditas 88:43-46.
* Nordstrom S, L Beckman, and I Nordenson. 1979. Occupational and environmental risks in
and around a smelter in northern Sweden. V. Spontaneous abortion among female
employees and decreased birth weight in their offspring. Hereditas 90:291-296.
* Norseth, T. 1994. Environmental pollution around nickel smelters in the Kola Peninsula
(Russia). Sci. Total Environ. (Netherlands) 148(2-3): 103-8.
* Osborne, JS 3rd, CM Shy, and BH Kaplan. 1990. Epidemiologic analysis of a reported
cancer cluster in a small rural population. Am J Natl 132(Suppl 1): S87-95.
* Otto, D. 1992. Assessment of Neurobehavioral Response in Humans to Low Level Volatile
Organic Compound (VOC) Sources. Ann. NY Acad Sci 641: 248-260.
* = Reference was retrieved.
* = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-l 1
-------
* Otto, D, KH Hudnell, D House, L Molhave, and WS Counts. 1992. Exposure of Humans
to a Volatile Organic Mixture: I. Behavioral Assessment. Arch Environ Hlth 47:23-30.
* Otto, D, I Skalik, P Subrt, J Ratcliffe, D House, and R Sram. 1995. Association of mercury
exposure with neurobehavioral performance of children in Bohemia. Neurotoxicology.
Submitted 1/95.
* Ozonoff, D, A Aschengrau, and P Coogan. 1994. Cancer in the vicinity of a department of
defense superfund site in Massachusetts. Toxicology & Industrial Health. 10(3): 119-141.
Ozonoff, D, ME Cohen, A Cupples, T Heeren, A Schatzkin, T Mangione, M Dresner, and T
Colton. 1987. Health problems reported by residents of a neighborhood contaminated by a
hazardous waste facility. Am J Ind Med 11: 581-97.
* Paustenbach, DJ, RD Bass, and P Price. 1993. Benzene toxicity and risk assessment,
1972-1992: implications for future regulation. Environ Health Perspect (USA) 101(Suppl.
6): 177-200.
* Perera, FP and AM Jefferey. 1990. Molecular Epidemiology and Cancer Prevention, Cancer
detect Prev 14(6): 639-45.
* Pershagen, G. 1985. Lung cancer mortality among men living near an arsenic emitting
smelter. Am J Epidemiol 122: 684-694.
* Pershagen, G and L Simonato. 1990. Epidemiological evidence on air pollution and cancer.
In: Air Pollution and Human Cancer (Tomatis L, ed.). Heidelberg: Springer-Verlag, p.63-
74.
* Phillips, LJ. 1992. Regional Relationships between Toxin Releases and Environmental and
Human Exposure to Toxic Substances. Bull Environ Contain Toxicol 48(6): 795-802.
* Pickle, LW, TJ Mason, N Howard, R Hoover, and JF Fraumeni, Jr. 1987. Atlas of US
cancer Mortality among whites: 1950-1980. DHHS publ. No. (NIH) 87-2900. Washington,
DC: US Government Printing Office.
* Reichhardt, Tony. 1995. Weighing the health risks of airborne particulates: recent
epidemiological studies link an increase in mortality and sickness to particulates, but do the
data justify a tighter air quality standard? Environ Sci & Tech 29(8): 360A-364A.
* Revich, BA. 1995. Public health and ambient air pollution in arctic and subarctic cities of
Russia. Sci Total Environ (Netherlands) 160-161: 585-92.
* Rosenman, KD, JE Rizzo, MG Conomos, and GJ Halpin. 1989. Central nervous system
malformations in relation to polyvinyl chloride production facilities. Arch Environ Health
44(5): 279-82.
* = Reference was retrieved.
* = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-12
-------
Rybicki, BA, CC Johnson, J Uman, and JM Gorell. 1993. Parkinson's disease mortality and
the industrial use of heavy metals in Michigan. Mov Disord 8(1): 87-92.
Sabsay, MI, FK Tetelyutina, NM Kanunikova, VI Zhdanova, and ZS Kornyaeva. 1993.
Ecological situation and congenital defects in Udmurtiya. Reprod Toxicol 7(5): 517.
Samet, JM and FE Speizer. 1993. Assessment of health effects in epidemiologic studies of
air pollution. Environ Health Perspect 101(Suppl 4): 149-54.
Schwartz, J and DW Dockery. 1992. Increased mortality in Philadelphia associated with
daily air pollution concentrations. Am Rev Respir Dis 145(3): 600-4.
Shy, CM. 1989. Review, discussion, and summary of epidemiological studies. Environ
Health Perspect 79: 187-90.
Shy, CM. 1993. Epidemiological studies of neurotoxic, reproductive, and carcinogenic
effects of complex mixtures. Environ Health Perspect 101(Suppl. 4): 183-186.
Spengler, JD and J Ware. 1990. Kanawha County West Virginia USA Health Study.
Source: World Conference on Lung Health, Boston Massachusetts, USA May 20-24, 90.
Am Rev Respir dis, 141(4 part 2): A77.
Stevens, RK, CW Lewis, TG Dzubay, LT Cupitt, and J Lewtas. 1990. Sources of
mutagenic activity in urban fine particles. Toxicol Ind Health 6(5): 81-94.
Summerhays, J. 1989. Estimation and Evaluation of Cancer Risks attributable to air
pollution in South East Chicago, Septl989.
Swaen, GMH and JJM Slangen. 1995. Gasoline consumption and leukemia mortality and
morbidity in 19 European countries: An ecological study. International Archives of
Occupational and Environmental Health (Germany). 67(2): 85-93.
Tashkin, DP, R Detels, M Simmons, H Liu, AH Coulson, J Sayre, and S Rokaw. 1994.
The UCLA population studies of chronic obstructive respiratory disease: XI. Impact of air
pollution and smoking on annual change in forced expiratory volume in one second. AM. J.
Respir. CRTT. Care Med. 149(5): 1209-1217.
Theriault, G, H Iturra, and S Gingras. 1983. Evaluation of the association between birth
defects and exposure to ambient vinyl chloride. Teratology 27: 359-370.
Tomatis, L. 1984. Exposures associated with cancer in humans. J Cancer Res Clin Oncol
108(1): 6-10.
Tomatis, L. 1988. Environmental cancer risk factors: a review. Acta Oncol 27(5): 464-
472.
* = Reference was retrieved.
* = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-13
-------
* Tomatis, L. 1991. Outdoor Air Pollution and Lung Cancer. Ann Oncol (Netherlands) 2(4):
265-7.
* Tomatis, L. 1995. How much of the human disease burden is attributable to environmental
chemicals? Toxicolgy Letters 77(1-3): 1-8.
* Tornqvist, M and L Ehrenberg. 1994. On cancer risk estimation of urban air pollution.
Environ Health Perspect 102(Suppl. 4): 173-182.
* US EPA. 1989a. US Environmental Protection Agency. Estimation and Evaluation of
Cancer Risks Attributable to Air Pollution in Southeast Chicago. Chicago, IL: US EPA Air
and Radiation Division, Region V.
* US EPA. 1989b. US Environmental Protection Agency. Analysis of Air Toxics Emissions,
Exposures, Cancer Risks, and Controllability in Five Urban Areas. Vol I, Base Year
Analysis & Results. Document no. EPA-450/2-89-012a. Research Triangl Park, NC: US
EPA Office of Air and Radiation, Office of Air Quality Planning and Standards.
* US EPA. 1990a. US Environmental Protection Agency. Cancer risk from outdoor exposure
to air toxics, Vol. 1. Final Report. Research Triangle Park, NC: US EPA Office of Air
Quality Planning and Standards.
* US EPA. 1990b. US Environmental Protection Agency. Toxic Air Pollutants & Noncancer
Health Risks: Screening Studies. Research Triangle Park, NC: US EPA Office of Air Quality
Planning and Standards.
* Utell, MJ, JM Samet, DV Bates, MR Becklake, DW Dockery, BP Leaderer, JL Mauderly,
and FE Speizer. 1988. Air pollution and health. Am Rev Respir Dis 138(4): 1065-8.
* Viren, JR and A Silvers. 1994. Unit risk estimates for airborne arsenic exposure: an
updated view based on recent data from two copper smelter cohorts. Regulatory toxicology
and pharmacology. 20(2): 125-138.
* Vitnerova, N and D Hartsman. Respiratory status of children living in polluted and control
area. Czech-Republic.
* Wallace, LA. 1991. Personal exposure to 25 volatile organic compounds EPA's 1987 team
study in Los Angeles, CA. Toxicol Ind Health 7(5-6): 203-208.
* Wallace, L, E Pellizzari, T Hartwell, M Rosenzweig, M Erickson, C Sparacino, and H
Zelon. 1984. Personal exposure to volatile organic compounds in breathing-zone air,
drinkng water, food and exhaled breath. Environ Res 35: 293-319.
* Wallace, L, E Pellizzari, T Hartwell, H Zelon, C Sparacino, and R Whitmore. 1984.
Analyses of exhaled breath of 355 urban residents for volatile organic compounds. Indoor
Air: Chemical Characterization and Personal Exposure 4: 15-20.
* = Reference was retrieved.
4 = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-14
-------
* Ware JH, Dockery DW, A Spiro III, FE Speizer, and BG Ferris, Jr. 1984. Passive
smoking, gas cooking, and respiratory health of children living in six cities. Am Rev Respir
Dis 129(3): 366-74.
* Ware, JH, JD Spengler, LM Neas, JM Samet, GR Wagner, D Coultas, H Ozkayanak, and M
Schwab. 1993. Respiratory and irritant health effects of ambient volatile organic compounds
: the Kanawha county health study. American journal of epidemiology 137(12): 1287-1301.
* Watts, R, Fitzgerald B, G Heil, H Garabedian, R Williams, S Warren, L Fradlein, and J
Lewtas. 1989. Use of bioassay methods to evaluate mutagenicity of ambient air collected
near a municipal waste combustor. JAPCA 39(11): 1436-9.
* Weetman, DF. 1994. Current knowledge of health effects from volatile organic chemicals in
the environment: Report of the International Meeting organised by Indoor Air International
in London on 27-28 October 1993. Cent Eur J. Public Health (Czech Republic). 2(2):
130-131.
* Wilcosky, TC. 1993. Biological markers of intermediate outcomes in studies of indoor air
and other complex mixtures. Environ Health Perspect 101(Suppl. 4): 193-7.
* Williams, FL, SA Ogston, and OL Lloyd. 1995. Sex ratios of births, mortality, and air
pollution: can measuring the sex ratios of births help to identify health hazards from air
pollution in industrial environments? Occup Environ Med (England) 52(3): 164-9.
* Wilson, R. 1988. Measuring and comparing risks to establish a de minimis risk level, Regul
Toxicol Pharmacol 8(3): 267-82.
* Wong, O and WJ Bailey. 1993. Cancer incidence and community exposure to air emissions
from petroleum and chemical plants in Contra Costa County, California: A critical
epidemiological assessment. J Environ Health 56(5): 11-17.
* Woodall, G, Jr, LH Kuller, EO Talbott, J Ramlow and B Salthouse. 1992. Epidemiological
study for the Kenova and Ceredo, West Virginia and Catlettsburg, Kentucky area, (Phase II).
Final Report. Prepared for USEPA Region IV, Air Compliance Branch. March 1992.
* Xu, YZ, WJ Blot, HP Xiao, A Wu, YP Feng, BJ Stone, J Sun, A Ershow, BE Henderson,
and JF Fraumeni. 1989. Smoking, air pollution, and the high rates of lung cancer in
Shenyang China. JNCI 81:1800-1806.
* Xu, ZY, WJ Blot, G Li, JF Fraumeni, Jr., DZ Zhao, BJ Stone, Q Yin, A Wu, BE
Henderson, and BP Guan. 1991. Environmental determinants of lung cancer in Shenyang,
China. IARC Sci Publ (France) 105: 460-5.
Xu, ZY, WJ Blot, HP Xiao, A Wu, YP Feng, BJ Stone, J Sun, AG Ershow, BE Henderson,
and JF Fraumeni, Jr. 1989. Smoking, air pollution, and the high rates of lung cancer in
Shenyang, China. J Natl Cancer Inst 81(23): 1800-6.
* = Reference was retrieved.
4 = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-15
-------
Yansheva, NY, IS Kireeva, and IA Chernichenko. 1993. The oncoepidemiologic
investigation under chemical carcinogens' regulation. Eksperimental'naya Onkologiya 15(5):
16-20.
Zhang, J and X Li. 1987. Chromium pollution of soil and water in Jinzhou. Journal of
Chinese Preventive Medicine 21: 262-264.
Zhao, MY, CJ Ying, N Shao, Y Yang, CF Yang, L Shi, and WQ Liu. 1994. The study of
health effects of vinyl chloride air pollution on population. Biomedical and Environmental
Sciences 7(2): 136-143.
Zmirou, D, A Deloraine, P Saviuc, C Tillier, A Boucharlat, and N Maury. 1994.
Short-term health effects of an industrial toxic waste landfill: a retrospective follow-up study
in Montchanin, France. Archives of environmental health. 49(4): 228-238.
* = Reference was retrieved.
* = Full citations of ATSDR documents are not available. No attempt was made to obtain them due
to time constraints. A-16
-------
APPENDIX B
SUMMARIES AND ANALYSES
OF SELECTED STUDIES
B-l
-------
STUDY SUMMARY; Brown et al. 1984
CITATION;
Brown, LM, LM Pottern, and WJ Blot. 1984. Lung cancer in relation to environmental
pollutants emitted from industrial sources. Environ. Res. 34:250-261.
SOURCES OF EXPOSURE;
Zinc smelter and steel plant in eastern Pennsylvania
HEALTH EFFECTS EVALUATED;
Lung cancer mortality
CHEMICALS EXPOSED TO;
Arsenic, cadmium, copper, lead, manganese, zinc
EXPOSURE MEASURES:
Air monitoring of metals and sulfur dioxide, soil samples for metals provided general
information. An exposure surrogate was also used in the analysis, which was based on
residential distance to the smelter or steel plant, weighted by emissions data.
CONFOUNDING FACTORS ASSESSED:
Occupational exposure, smoking, amount of smoking, age
STUDY METHODS;
This was a case-control study of 335 white male lung cancer deaths in 1974-1977 and 332
white male control deaths in eastern Pennsylvania. Cases and controls were identified from
state mortality records, and next of kin were interviewed with the interviewers blinded as to
whether they were interviewing case or control relatives. As an exposure measure, maps
were divided into 1 km2 grids, which were classified as "heavy" or "light" exposure based on
soil and air sampling and prevailing wind direction. The surrogate individual exposure
measure was based on distance of residence to the plants and an emissions weighting for the
grid of residence.
STUDY RESULTS:
Increased relative risks of lung cancer were reported for heavy versus light exposure areas for
arsenic and cadmium. When cases and controls were broken down by proximity to the zinc
smelter (taking into account the prevailing wind direction), there was a significantly increased
relative risk, adjusted for smoking and usual employment in the steel or zinc industries, based
on residence in 1950. There was also an apparent trend with distance, although no trend test
was conducted. Based on usual residence, relative risk was increased in those near the zinc
Air Toxics Study Summary g.2 Brown et al. 1984
-------
smelter (with or without adjustment for smoking and occupation), but the increase was not
significant.
AUTHOR'S CONCLUSIONS;
The study authors concluded that the results suggest an increased lung cancer risk associated
with ambient exposure to heavy metals, including arsenic and cadmium. However, they noted
that the small sample size and limited exposure data preclude a causal conclusion.
COMMENTS;
The study was well-conducted and controlled for the major confounding factors of smoking
habits, occupational exposure, and age. The data suggest increased risk of lung cancer
following environmental exposure to inhaled arsenic or cadmium. It is limited by the lack of
detailed exposure data and the small number cases and controls in the "near" proximity group
(< 10-16 km).
Air Toxics Study Summary B-3 Brown et al. 1984
-------
STUDY SUMMARY: Bustueva et al. 1994
CITATION;
Bustueva, KA, BA Revich, and LE Bezpalko. 1994. Cadmium in the environment of three
Russian cities and in human hair and urine. Arch. Environ. Health (USA). 49(4): 284-288.
SOURCES OF EXPOSURE:
Alkaline storage battery factory, cadmium production plant, cadmium-dye plant (each in a
separate city)
HEALTH EFFECTS EVALUATED;
Beta-2-microglobulin (/3-2m) excretion in urine (proteinuria) as an indicator of kidney function
CHEMICALS EXPOSED TO;
Cadmium
EXPOSURE MEASURES;
Ambient exposure measure of cadmium in air. Levels of cadmium in hair and urine were
also used as biomarkers of exposure. Soil levels of cadmium were also measured.
CONFOUNDING FACTORS ASSESSED:
None
STUDY METHODS;
The study included 153 adult women and 430 children exposed to atmospheric cadmium.
However, not all assays were conducted with all subjects. The total number of controls is
unclear, but 34 adults were used as controls for the determination of /3-2m. For the children,
only levels of cadmium in urine and hair were reported. The cadmium concentration in air
was measured at various distances from the cadmium plant, and urinary /3-2m was reported
for the corresponding group of subjects. Levels of /3-2-microglobulin in urine were measured
as an indicator of kidney function. A separate study was conducted on the effects of
occupational exposure in the cadmium plants.
STUDY RESULTS;
Although average levels of (3-2m were not elevated compared to controls in the population
exposed to environmental cadmium, the maximum levels were generally higher. In addition,
only the groups exposed to environmental cadmium had individuals with j3-2m levels above
250 uglL. However, only a total of 4 subjects had this level of increased /3-2m excretion,
and no statistical tests were conducted. None of the children had elevated /3-2m excretion
(defined as >250 /ig/L). Daily cadmium intake from air was estimated at 0.0035 jtg,
Air Toxics Study Summary B_4 Bustueva et al. 1994
-------
compared with daily intake from water of 0.0001 /ig and intake from food of 0.113 /ig.
However, absorption appeared to have been taken into account only for the calculation of air
and water intake, but not for the calculation of food intake. Cadmium levels in food were
elevated, presumably due to air deposition. In the occupational study, the concentration of
cadmium in air correlated with the percent of subjects with proteinuria.
AUTHORS CONCLUSIONS:
The study authors concluded that the observed cadmium levels in air resulted in elevated
cadmium levels in hair and urine of exposed children and adults. They also suggested that
cadmium exposure from air, together with cadmium intake in food, can result in proteinuria,
which is indicative of diminished kidney function.
COMMENTS:
A major strength of this study is that it measured exposure and effect in the same population.
In addition, the study authors attempted to determine total cadmium intake from air, food, and
water, to determine the degree to which the observed effect was from inhalation exposure.
However, the small size of the control group makes it difficult to determine the background
level of proteinuria in the control population. The observed endpoint is plausible, since
cadmium is a known kidney toxin, and proteinuria occurs after several years of occupational
exposure leads to the accumulation of cadmium in the kidney. However, the cumulative
exposure necessary to result in proteinuria above background levels is generally reported at
much higher levels. For example, Jarup et al. (1988) reported proteinuria at a cumulative
exposure of 691 yr x ng/m3, while the highest ambient exposure reported in this study was
0.085 /ig/m3. This suggests three possible explanations for the observed findings: (1)
Increased proteinuria did occur as a result of cadmium exposure, and a major source of
cadmium intake was food contaminated with cadmium from air deposition. (2) Increased
proteinuria did occur, and this study population is more sensitive than that studied by Jarup et
al. (3) There was no cadmium-related effect on the incidence of proteinuria, and the observed
deviations result from statistical fluctuations. The available data are insufficient to
differentiate among these possibilities.
Reference:
Jarup, L., C.G. Elinder and G. Spang. 1988. Cumulative blood-cadmium and tubular
proteinuria: a dose-response relationship. Int Arch Occup Environ Health 60: 223-229.
Air Toxics Study Summary B-5 Bustueva et al. 1994
-------
STUDY SUMMARY: Daval et al. 1995
CITATION;
Dayal, H and S Gupta. 1995. Symptom clusters in a community with chronic exposure to
chemicals in two superfund sites. Arch. Environ. Health 50(2): 108-11.
SOURCES OF EXPOSURE;
Two National Priority List sites located east of Houston, Texas
HEALTH EFFECTS EVALUATED;
Neurological, respiratory, cardiovascular, gastrointestinal, and miscellaneous symptoms
CHEMICALS EXPOSED TO;
Potential exposure to: benzene, toluene, naphthalene, tetrachloroethylene, trichloroethylene,
polychlorinated biphenyls, dichlorodiphenyltrichloroethane, 1,1,2-trichloroethane, 1,2-
dichloroethane, heptachlor, fluorene, pyrene, benzo(a)anthracene, benzo(a)pyrene, chrysene,
lead, copper, chromium, nickel, and zinc.
EXPOSURE MEASURES;
An exposure measure was calculated by calculating (c/d2) for each residence of the test subject
(where c is the estimated amount of chemicals present at the site at the time of residence and
d is the distance to the site), and weighting by the duration of residence at that location. For
each test subject, this value was calculated for both sites and the two exposure ratings were
added together. Amounts of chemicals present at the two sites were determined for period
from August 1962 to February 1990.
CONFOUNDING FACTORS ASSESSED:
None.
STUDY METHODS:
Health effects and residence data for 6209 African-American residents of a community near
two Superfund sites were gathered by a mail-in questionnaire. The response rate was 73%.
The questionnaire also evaluated smoking and drinking history, but data supplied on these
factors were limited. Cumulative exposures for study participants were calculated from the
residence and chemical site data. Reported symptoms in the subjects from the top 10% of the
exposure distribution (321 subjects) were compared with reported symptoms in the bottom
30% of the exposure distribution (351 subjects).
STUDY RESULTS:
Air Toxics Study Summary' g_6 Dayal et al. 1995
-------
There were significant increases in the prevalence of several of the symptoms, especially those
affecting the nervous system. Significant increases of at least 50% between the low- and
high-exposure groups were observed in prevalence of learning difficulties, unusual fatigue,
decreased sense of smell, and peculiar odor/taste. Significant increases were also observed
for several other symptoms, including nervousness, numbness in fingers/toes, chest pains,
irregular heartbeat, and skin rashes. The individuals in the high-exposure group were also
twice as likely to report five or more neurological symptoms than individuals in the limited
exposure group.
AUTHOR'S CONCLUSIONS:
The authors noted the following study limitations: possible self-selection, response, and
differential recall biases, subjective symptom data, no chemical analysis of site wastes, no
evaluation of probable exposure route(s), the lack of determination of actual exposure based
on prevailing wind information, and the lack of evaluation of confounding factors, such as
occupational exposure, smoking, and drinking. However, the reporting of effects on mostly
one system (the nervous system) supports the apparent relationship between exposure and
effects. Study results are also supported by the fact that the chemicals to which residents
would be exposed are known from animal studies to cause neurological effects. The study
authors concluded that their study indicates that an analytical epidemiological investigation of
this population is warranted.
COMMENTS:
Although there are several limitations to the study, as identified by the study authors, this
study does provide suggestive evidence that neurotoxic effects resulted from exposure to
chemicals at the NPL site. Because no information on exposure route was available, it is
unclear whether the predominant exposure was oral (presumably from drinking water
contamination) or via inhalation. However, the evidence is sufficient to warrant better
characterization of actual exposures and exposure route(s).
Air Toxics Study Summary 3.7 Dayal et al. 1995
-------
STUDY SUMMARY: Edmonds et al. 1978
CITATION;
Edmonds, LD, CE Anderson, JW Flynt Jr., and LM James. 1978. Congenital central
nervous system malformations and vinyl chloride monomer exposure: a community study.
Teratology 17: 137-142.
SOURCES OF EXPOSURE;
Polyvinyl chloride polymerization plant, other chemical plants in the Kanawha valley
HEALTH EFFECTS EVALUATED;
Central nervous system (CNS) birth defects
CHEMICALS EXPOSED TO;
Vinyl chloride, 100 other compounds emitted by chemical plants in the Kanawha valley
EXPOSURE MEASURES:
Mean hourly vinyl chloride monomer emissions, number of abnormal vinyl chloride emission
incidents, distance/direction from plant
CONFOUNDING FACTORS ASSESSED;
Maternal age, parental occupation, history of previous pregnancies and birth defects,
education, socioeconomic status
STUDY METHODS;
Due to concerns about occupational exposure to vinyl chloride being linked to birth defects,
CNS birth defects were evaluated in U.S. counties with polyvinyl chloride polymerization
(PVC) plants. A detailed study was conducted in the Kanawha valley, West Virginia, which
had a CNS birth defect rate above the national average during the early 1970s. All infants
(n=47) with central nervous system (CNS) defects born in Kanawha County in 1970-1974
were identified using the national Birth Defects Monitoring Program (BDMP), which
contained data on 95% of all reported births in the county, and state vital statistics records
noting CNS defects recorded on birth or death certificates. Controls were matched for race,
paternal education, maternal age within 2 years, and parental residence in the county at the
time of conception. Families of 46 cases and the corresponding controls were interviewed to
determine the previous pregnancy, miscarriage, and birth defect history, and occupational
history of both parents. The study authors did not control for smoking or alcohol use.
STUDY RESULTS:
Air Toxics Study Summary B-8 Edmonds et al. 1978
-------
CNS birth defect incidence rates for Kanawha County in 1970-1972 were found to be 1.5-2
times higher than the national average, based on the BDMP data. There was no significant
difference between cases and controls with regard to previous reproductive experience of the
parents, occupational parental exposures to VC, or any of the other control variables. Using
the matched-pairs, signed-ranks test, there was no significant difference between cases and
controls with regard to distance of residence from the plant. There was apparent clustering of
cases living less than three miles from the plant in the direction to the northeast of the plant.
However, analyses of the locations of cases and controls, when grouped by distance from the
plant, found no consistent differences between the geographical distribution of case residences
and control residences.
The total number of cases of CNS defects were relatively constant in 1970-1972 (14, 15, and
12 cases, respectively), but decreased to 5 cases in 1973 and 1 case in 1974. Emissions data
indicated that average emissions in the years 1967-1973 were relatively stable at 235-270
pounds/hour, decreasing to 180 pounds/hour in 1974; the number of large releases also
decreased during this period. Exposure monitoring was limited to a single release incident,
when vinyl chloride monomer concentrations were 0.1-0.2 ppm downwind (to the northeast)
of the plant. The decrease in the reported incidence of CNS defects did not coincide with the
decrease in average emissions.
AUTHOR'S CONCLUSIONS;
The study authors stated that the data did not allow any conclusions to be drawn regarding the
relationship between air pollution and birth defects in the community. They noted that if
there were such an association for vinyl chloride, decreased birth defect rates would be
expected to follow decreased atmospheric vinyl chloride levels after an ~9-month lag.
However, the largest decrease in birth defect rates preceded the decline in vinyl chloride
emissions. Similarly, the lack of a strong geographic pattern of cases relative to the plant,
and the presence of numerous other sources of chemical emissions were identified as
weakening the possible relationship between the plant VC emissions and CNS defects.
COMMENTS:
The study authors have drawn appropriately conservative conclusions based on the data
quality and small study size. The data in this study, in and of themselves, cannot explain the
elevated rate of CNS birth defects in this county compared to the national average as being
related to vinyl chloride monomer exposure. Similarly, however, exposures to vinyl chloride,
or the combined effects of the numerous HAPs present in the air in the valley, cannot be
ruled out as a factor contributing to birth defects incidence in the study area.
This study is limited by two major factors. First, the small sample size and short duration
make it very susceptible to statistical fluctuations. For example, the CNS birth defect rate
(per 10,000 births) dropped from about 43 in 1971 to about 3 in 1974. There was no similar
decrease in vinyl chloride emissions preceding this drop; emissions dropped slightly from
1973 to 1974, and more strongly in 1975. It is difficult to evaluate the significance of these
data without emissions and birth defect data from 1975 and later years. This study is also
limited by the failure of the authors to assess or control for smoking and drinking among
cases and control parents, both of which could affect the incidence of birth defects.
Air Toxics Study Summary B-9 Edmonds et al. 1978
-------
STUDY SUMMARY: Frost et al. 1987
CITATION:
Frost, F, L Barter, S Milham, R Royce, A Smith, J Hartley, and P Enterline. 1987. Lung
cancer among women residing close to an arsenic emitting copper smelter. Arch. Environ.
Health 42: 148-152.
SOURCES OF EXPOSURE:
Copper smelter and arsenic refinery in Tacoma, Washington
HFALTH EFFECTS EVALUATED;
Lung cancer mortality
CHEMICALS EXPOSED TO;
Arsenic
EXPOSURE MEASURES:
Distance from smelter (cross-sectional study). For the case-control study, a cumulative
exposure index was used, based on years x 1/distance, multiplied by a weighting factor that
accounted for varying estimated emissions levels.
CONFOUNDING FACTORS ASSESSED;
Age, race
STUDY METHODS;
Two study methods were employed to evaluate the possible increase in lung cancer deaths in
women from 1935 to 1969 due to arsenic exposure from the copper smelter and arsenic
refinery. In the cross-sectional study, three exposure groups were defined. The first group
resided within a 3-mile radius of the facility, the second group resided between 3 and 9 miles
from the site, and the third group resided between 6 and 40 miles from the plant. The
expected numbers of lung cancer deaths were calculated and the observed rates of lung cancer
deaths were determined from death certificate information. The second investigation was a
case-control study. Each lung cancer death case was matched to a female control according to
age and year of death. A total of 156 pairs were used in the analysis.
STUDY RESULTS:
In the first investigation, the total numbers of observed lung cancer deaths did not exceed the
expected numbers of lung cancer deaths. In fact, the relative risk (observed/expected, O/E)
was < 1 in all groups, with risk significantly lower than expected in the lowest exposure
group. This decrease was attributed to the lower level of urbanization in this group. In the
Air Toxics Study Summary B-10 Fr°st et al. 1987
-------
case-control study, the cumulative exposure index was 27% higher for cases than for controls
(p=0.10); the p value dropped to 0.07 after adjusting for a 20-year latency. When the
exposure levels were divided into quintiles of exposure, the odds ratios ranged from 1 to 1.6,
and a trend test fell just short of significance (p=0.07).
AUTHOR'S CONCLUSIONS:
The study authors offered several explanations for the mixed results. They suggested that the
national rates used to calculate the number of expected deaths due to lung cancer may not be
valid for the study area. Inverse distance was used as part of a surrogate measure of
exposure, but the measure used may not give an accurate picture of exposure. The unknown
period of latency may prevent detection of significant trends in increased risk of lung cancer
from low-level arsenic exposure. Confounding factors, such as movement of residence,
socioeconomic status, and smoking habits, were not considered. Because the observed
increases were so small, the study authors concluded that the data indicate that any increased
risk would be small and limited to those who lived very close to the smelter for over 20
years. They also noted that the community 24-hour average ambient arsenic level during
1980-1985 exceeded the 8-hour workplace standard of 10 /ig/m3. (The source of this
statement is unclear, since the mean ambient level reported for 1969-1980 ("the only period
for which such data are available") was 0.8 jig/m3 at 2100 feet from the smelter.)
COMMENTS;
This study indicated that any increased risks resulting from arsenic exposure near this smelter
are likely to be small. All increases in relative risk were small, and no statistical test was
significant at a p value of 0.05. However, increased risk of lung cancer is plausible in light
of the unit risk for inhaled inorganic arsenic (4.3E-3 per pg/m3). The lack of control for
smoking makes the results of this study very hard to interpret.
Air Toxics Study Summary B-ll Frost et al. 1987
-------
STUDY SUMMARY: Hallenborg et al. 1991
CITATION;
Hallenborg, CP and N Marsh. 1991. Toxic substances in the environment affecting
respiratory function of people in Hawaii. Hawaii Med J 50(3): 101-7.
SOURCES OF EXPOSURE:
Air emissions from sugar cane factory
HF.AI/TH EFFECTS EVALUATED;
Standard pulmonary function variables
CHEMICALS EXPOSED TO:
Paniculate emissions from sugar cane burning and processing; evidence is presented that fine
silicate particles are a key pollutant contributing to health effects
EXPOSURE MEASURES;
Residential distance from factory, duration of residence (no direct measurements)
CONFOUNDING FACTORS ASSESSED;
Pulmonary function was compared to literature control values matched for age, weight, and
sex
STUDY METHODS;
Spirometry and diffusing capacity measurements were performed on 14 adult residents who
have lived near the plant for 4 to 40 years. Histories were taken and physical examinations
were also performed.
STUDY RESULTS;
A high proportion of all of the residents had reduced diffusing capacity (79 percent), forced
expiratory volume (14 percent), forced expiratory fraction measures (93-100 percent), total
lung capacity (83 percent), vital capacity (43 percent) or elevated residual volume (93
percent). Self-reporting and physical examinations also disclosed a high frequency of
respiratory symptoms, but no radiological or other evidence of asbestosis or other frank
disease related to inorganic fiber exposures. Decreased diffusing capacity was found to be
strongly related to length of residence near the plant, but the other symptoms were not.
AUTHOR'S CONCLUSIONS;
Air Toxics Study Summary B-12 Hallenborg et al. (1991)
-------
The authors conclude that the findings in this study are consistent with "the presence of some
lung disease in this population." They recommend long-term studies to confirm whether
sugar cane burning and processing are adversely affecting the respiratory health of nearby
residents.
COMMENTS;
The lack of actual exposure measurements make it hard to determine whether the observed
respiratory symptoms were related to sugar cane processing, and there is no information on
occupational exposure histories or how "normal" control values of the various pulmonary
function measures were derived. A matched control group would have been desirable. The
hypothesis that the silicate component of air emissions may be responsible for the observed
effects is plausible, and supported by other studies of sugar cane processing emissions and
health impacts, but not directly confirmed by this study. These results illustrate the need for
better characterization of unconventional sources of toxic air pollution.
Air Toxics Study Summary B-13 Hallenborg et al. (1991)
-------
STUDY SUMMARY: Kilburn and Warshaw 1995
CITATION:
Kilburn, KH, and RL Warshaw. 1995. Neurotoxic Effects From Residential Exposure to
Chemicals from and Oil Reprocessing Facility and Superfund Site. Neurotoxicology and
Teratology. 17(2): 89-102.
SOURCES OF EXPOSURE;
Emissions from an oil and chemical reprocessing facility and superfund site in Louisiana
HEALTH EFFECTS EVALUATED:
Self-reported neurological/psychological symptoms, mood indices from questionnaire, results
from objective neurophysiological and neuropsychological testing
CHEMICA1
Primarily organic chemicals (aliphatic, aromatic and chlorinated, including PCBs) and some
metals; major site operations were poorly-controlled burning and distillation of wastes
EXPOSURE MEASURES;
Length of residence in area affected by air emissions (based on screening-level models),
distance from residence to site (no direct measurements of exposure)
CONFOUNDING FACTORS ASSESSED;
Control groups from a distant location were compared in terms of age, sex, race, education,
occupational chemical exposures, smoking status, drug use, trauma, history of neurological,
psychological disease
STUDY METHODS;
The study population consisted of 77 women and 55 men living within 2.4 kilometers of the
site, including both self-selected and randomly-selected individuals. The control group was
randomly selected from a town 35 kilometers away from the site. Questionnaires related to
physical and psychological symptoms and mood variables were administered to the study
group and to the control population, along with a battery of neurophysiological and
neuropsychological tests. Questionnaire and test results were compared between the control
and study groups, between the self-selected and randomly-selected individuals in the study
group, and as a function of residential distance from the site and residential duration.
STUDY RESULTS:
Air Toxics Study Summary B-14 Kilburn and Warshaw 1995
-------
The exposed group scored significantly lower on four of the neurophysiological tests (related
to balance and response to visual stimuli), and showed a significant impairment in several of
the neuropsychological tests (pattern recognition and manual skill tasks), compared to the
control group. The exposed group also scored higher than did the control groups on indices
of mood disturbance consistent with depression. There was no difference in symptoms as a
function of distance from the site or length of residence, nor between self-selected and
randomly-selected subjects.
AUTHOR'S CONCLUSIONS;
The authors conclude, "Subjects exposed residentially for up to 17 years to chemicals
dispersed from a waste oil reprocessing plant showed neurophysiological and
neuropsychological impairment."
COMMENTS;
This is generally a well-controlled study that takes great care to control for major confounding
factors. While the control and study populations did vary on a few variables (level of
education, for instance), there is no evidence that these differences accounted for the observed
differences in test results. The lack of an apparent relationship between exposure measures
and response within the exposed group somewhat weakens support for a link between
chemical exposures and adverse effects, but plausible explanations for this phenomenon
(nonuniform prevailing winds for lack of correlation with distance, short latency period for
lack of correlation with duration of exposure) are advanced. In light of the reported off-site
migration of chemicals in water, exposure in drinking water may also have occurred. This
study provides a good illustration of the current state of the art in the epidemiological
investigation of neurophysiological and neuropsychological effects of chemical exposures.
Air Toxics Study Summary B-15 Kilburn and Warshaw 1995
-------
STUDY SUMMARY: Laurent et al. 1993
CITATION;
Laurent, C, T Lakhanisky, P Jadot, I Joris, M Ottogali, C Planard, D Bazzoni, JM Foidart,
and Y Ros. 1993. Increased sister chromatid exchange frequencies observed in a cohort of
inhabitants of a village located at the boundary of an industrial dumping ground: phase I.
Cancer Epidemiol Biomarkers Prev 2(4): 355-62.
SOURCES OF EXPOSURE;
Industrial chemical disposal site near a village in Belgium, which released chemicals into the
air due to containment failure
HEALTH EFFECTS EVALUATED;
The frequency of sister chromatid exchanges (SCEs) in peripheral white blood cells of village
residents
CHEMICALS EXPOSED TO;
Aliphatic, aromatic, and chlorinated hydrocarbons, including the known carcinogens/mutagens
benzene, vinyl chloride, vinylidene chloride, trichloroethylene (TCE), tetrachloroethylene
(perchloroethylene, PCE)
EXPOSURE MEASURES;
Analyses of gases released from landfill, short-terms measurements of ambient air
concentrations at several locations in the village
CONFOUNDING FACTORS ASSESSED;
Control groups were matched for age, sex, and smoking habits
STUDY METHODS:
SCE levels in lymphocytes were assayed using standard cytogenetic methods in 51 village
residents (11 children, 30 nonsmoking adults, and 10 smoking adults), randomly selected
from a group of volunteers. A control group was selected from a village 400 kilometers away
where chemical exposures were not thought to have occurred. SCE frequencies were also
compared to literature values.
STUDY RESULTS:
All three groups (children, nonsmokers, smokers) from the village next to the disposal site
had significantly elevated frequencies of SCE compared to the corresponding control groups.
SCE levels were elevated approximately two-fold in exposed children, and slightly less in
Air Toxics Study Summary B-16 Laurent et al. 1993
-------
both adult smokers and nonsmokers, compared to the control groups. SCE frequencies in the
control groups were within the normal range expected from previous studies.
AUTHOR'S CONCLUSIONS;
The authors concluded that there is an association between residence in the village and
elevated frequencies of SCEs. They infer that the observed increase is related to chemical
exposures, and point out that, despite the relatively low levels of exposure measured in
ambient air, the SCE frequencies in residents are comparable to those seen in occupational
groups receiving much higher exposures. They are cautious to point out that no firm
conclusions can be drawn regarding the relationships between elevated SCE rates and
increased risk of cancer or genetic defects, and elect to characterize the observed SCE
increase merely as an indicator of exposures.
COMMENTS;
This study seems to have been very well-designed and executed, and the relatively sudden
failure of the containment at the dump site, with the corresponding rapid release of large
volumes of volatile contaminants to the air near a village, provided an excellent opportunity to
investigate the relationship between volatile chemical exposures and cytogenetic abnormalities.
The use of both literature controls and a separate control group strengthens the conclusion that
SCE levels were abnormally elevated in the exposed village residents. The significance of
SCE frequency elevation, in terms of increased cancer or other health risks, remains
unresolved at this time, as appropriately noted by the authors. However, it should be noted
that, unlike gene mutations, which are passed on in later cell divisions, SCEs are indicative of
repair of damage to DNA, and thus are not directly passed on in later cell divisions.
Air Toxics Study Summary g.17 Laurent et al. 1993
-------
STUDY SUMMARY: Pershagen 1985
CITATION;
Pershagen, G. 1985. Lung cancer mortality among men living near an arsenic-emitting
smelter. Am J. Epidem. 122: 684-694.
SOURCES OF EXPOSURE;
Arsenic-emitting smelter in northern Sweden
HEALTH EFFECTS EVALUATED:
Lung cancer deaths
CHEMICALS EXPOSED TO;
Arsenic, various metals (e.g., copper, lead, zinc), sulfur dioxide
EXPOSURE MEASURES:
Limited measurements of arsenic concentrations in ambient air and soils. Residence in
exposed areas was used as an exposure surrogate.
CONFOUNDING FACTORS ASSESSED;
Occupational exposure, smoking, indoor radon exposure (based on housing type), age
STUDY METHODS;
This was a case-control study of 212 men with pulmonary carcinoma and 424 matched
controls in northern Sweden. Cases were identified in the county containing the exposed and
reference areas from cancer registries. Ambient air levels were measured at distances up to 7
km from the smelter. Controls were extracted from a national death register and had died in
the study area during the same time period, but without a diagnosis of lung carcinoma. Data
on occupation, smoking and residence were obtained by questionnaire and interviews of next
of kin.
STUDY RESULTS:
Although ambient air concentrations of metals in suspended dust were measured for several
years, only highly summarized data were presented. Monthly average arsenic and lead
concentrations were mostly <0.5 ^ig/m3 and a few measurements exceeded 1 /tg/m3.
However, no data were available for earlier periods when emissions were substantially higher.
Excluding smelter workers and miners, the relative risk for lung cancer was 2.2 for
nonsmokers in the exposed area versus nonsmokers in the reference area, and 2.3 for smokers
in the exposed area versus smokers in the references area. The increase observed for smokers
Air Toxics Study Summary B-18 Pershagen 1985
-------
was statistically significant, but increase for nonsmokers was not. Most of the difference in
occupational distribution between the exposed and referent populations was accounted for by
increased employment at the smelter in the exposed area and increased farming, forestry, and
fishing work in the referent area. The age- and smoking-standardized relative risk was 2.0
for residents in the exposed area (95% confidence interval 1.2-3.4). Significant increases in
age- and smoking-standardized relative risk were also observed in miners and smelter
workers. Daily tobacco consumption of the cases in the reference area tended to be higher
than among cases in the exposed area; no information was available on age at start of
smoking. Analysis of housing characteristics for associations with increased levels of indoor
radon did not explain the increased risk.
AUTHOR'S CONCLUSIONS;
There was an increased relative risk for lung cancer in the exposed area that remained after
standardization for age, smoking habits, and occupation. The increased risk could also not be
explained by differences in tobacco consumption or in housing characteristics. The study
authors also noted that additional studies were conducted based on duration of residence in the
exposed area. These studies tended to show increased risk with longer residence in the
exposed area, but the numbers were too small for meaningful conclusions.
COMMENTS:
This was a well-conducted and well-controlled study that provides strong suggestive evidence
of an association of increased lung cancer deaths with exposure to emissions from a smelter.
The significantly increased relative risk remained after control for the major likely
confounding factors. While it would be useful to have more quantitative exposure-response
information and data on exposure levels when emissions were higher, an effect is plausible in
light of the unit risk for inhaled inorganic arsenic (4.3E-3 per /xg/m3) and the higher emissions
that occurred prior to monitoring. Information on other exposures (criteria pollutants and
other metals) would have been helpful.
Air Toxics Study Summary B_19 Pershagen 1985
-------
STUDY SUMMARY: Rosenman et al. 1989
CITATION;
Rosenman, KD, JE Rizzo, MG Conomos, and GJ Halpin. 1989. Central nervous system
malformations in relation to polyvinyl chloride production facilities. Arch Environ Health
44(5): 279-82.
SOURCES OF EXPOSURE;
Two polyvinyl chloride production facilities in New Jersey. Additional potential exposure
sources in the area that were not evaluated included foundries, an industrial gas manufacturer,
an acrylate manufacturer, a plastics manufacturer, and an electrical generation plant.
HEALTH EFFECTS EVALUATED;
Birth defects, central nervous system (CNS) birth defects
CHEMICALS EXPOSED TO;
Vinyl chloride
EXPOSURE MEASURES:
Distance from plant. In addition, annual stack and fugitive emissions were reported.
CONFOUNDING FACTORS ASSESSED:
Parental occupation, maternal age
STUDY METHODS:
This was a modified case-control study of the relationship between birth defects in the years
1977-1980 and distance from two polyvinyl chloride (PVC) production facilities in New
Jersey. Birth defect information was taken from Burlington County's and adjoining counties'
delivery room hospital logs. Residence information was taken from birth certificates. For
New Jersey residents who delivered in Pennsylvania, birth defect and residence data were
taken from computer printouts of birth certificate information. In general, 3 controls were
matched to each CNS birth defect case according to maternal age, month and year of birth,
race, and sex of child. Parents were interviewed to confirm the presence of the birth defect
and to determine whether the parents worked in the PVC facilities; only CNS birth defect
cases in which the parents could be interviewed were included in the analysis. Odds ratios
were calculated for birth defects and, specifically, for CNS birth defects.
STUDY RESULTS:
The odds ratio for CNS birth defects showed a trend of increasing odds ratio with increasing
proximity to the higher-emitting plant. Although the odds ratio never achieved statistical
Air Toxics Study Summary B-20 Rosenman et al. 1989
-------
significance, an odds ratio > 1 (indicating increased risk) was observed at all distance
intervals where cases were observed. No statistical analysis for trend was conducted. A
weaker trend was seen in analysis of the second facility; again, all odds ratios were > 1.
When the data was adjusted for the elevated trend in CNS birth defects, no correlation was
seen between birth defects and distance from the facilities. The CNS birth defects could not
be attributed to occupational exposure to vinyl chloride or ethylene oxide.
AUTHOR'S CONCLUSIONS;
The study authors concluded that the data suggest an increased probability of CNS birth
defects with exposure to vinyl chloride. They noted that, due to the small sample size, the
study has only a 54% statistical chance of detecting a 4-fold increased risk. Therefore, the
lack of statistical significance of the observed increases is not surprising. However, the
highest ambient exposure to vinyl chloride reported by the EPA (actual value was not
reported) is much lower than the 2500 ppm reported by John et al. (1981) to cause skeletal
and urethral effects in rats. No comparison was made to the vinyl chloride concentration
found to cause fetotoxicity in mice (500 ppm), but vinyl chloride levels near PVC facilities
tend to be < 10 ppm.
COMMENTS;
This study did not consider many potential confounding factors, such as pregnancy and birth
defect history, smoking, drinking, education, or socioeconomic status. Although the observed
trend with distance is suggestive, the conflict with animal toxicity data on the levels at which
vinyl chloride is teratogenic increases the likelihood that the observed effect may be due to a
confounding factor. Because of the strong drop-off of apparent effect with distance from the
plant, this study could not be expanded by increasing the size of the study population.
However, it could be expanded by studying CNS birth defects over a longer time period.
Reference:
John, J.A., Smith, F.A., Schwetz, B.A. 1981. Vinyl chloride: inhalation teratology study
in mice, rats, and rabbits. Environ. Health Perspect. 41: 171-177
Air Toxics Study Summary B-21 Rosenman et al. 1989
-------
STUDY SUMMARY: Ware et al. 1993
CITATION:
Ware, JH, JD Spengler, LM Neas, JM Samet, GR Wagner, D Coultas, H Ozkayanak, and M
Schwab. 1993. Respiratory and irritant health effects of ambient volatile organic compounds
: the Kanawha county health study. American Journal of Epidemiology 137(12): 1287-1301.
SOURCES OF EXPOSURE;
11 chemical manufacturing plants in the Kanawha valley
HEALTH EFFECTS EVALUATED:
Chronic respiratory symptoms (wheezing, cough, asthma, bronchitis) and acute irritation of
eyes, nose, throat, or chest in elementary school children grades 3-5
CHEMICALS EXPOSED TO:
Outdoor vapor badge samples of 15 volatile organic compounds (VOCs) were taken at fixed
locations (schools). Substances analyzed for included toluene, xylenes, benzene, decane,
1,1,1-trichloroethane, carbon tetrachloride, chloroform, and 1-butanol. Subjects were also
exposed to other air pollutants.
EXPOSURE MEASURES:
Continuous sampling for 15 VOCs at 75 elementary schools; maximum, median, mean, and
standard deviation were reported. Criteria pollutants were also measured and reported in
separate studies.
CONFOUNDING FACTORS ASSESSED;
Parental smoking, socioeconomic status, age, criteria pollutants, air moisture in homes
STUDY METHODS:
Health questionnaires were completed for 8,549 children, a 97% return rate. Because
children in special instruction units were excluded, data from 7,796 children were analyzed.
Schools were categorized by proximity to chemical industry complexes into four groups: in
valley, near industry; in valley, far from industry; outside of valley, but near valley; outside
of valley, and far from valley. The incidence of health outcomes was adjusted for parental
smoking and socioeconomic status. Dampness of homes was strongly related to outcomes,
but was not related to proximity to industry, so no adjustment was made for this variable.
STUDY RESULTS:
There was a significant trend with increasing proximity in the adjusted health measures for
persistent wheezing, asthma, and acute eye irritation. The strongest difference was between
Air Toxics Study Summary B-22 Ware et al. 1993
-------
children attending school in the valley and those attending school outside the valley.
However, there was little correlation between incidence of effect and proximity to industry
within the valley. For asthma, the odds ratio (OR) was 1.27 (95% confidence interval 1.09-
1.48), and for the composite indicator of chronic lower respiratory effects, the OR was 1.17
(confidence interval 1.04-1.31). A correlation between concentration of VOCs and several of
the health effects was also observed.
AUTHOR'S CONCLUSIONS:
The study authors concluded that children attending schools closer to chemical plants and
schools with higher measured concentrations of VOCs were more likely to report respiratory
symptoms and eye irritation. They further stated that regional air pollutants, such as
particulates, nitrogen dioxide, and sulfuric acid, were not potential confounders, although the
possibility remains that the VOCs were a marker for other pollutants (likely HAPs). The
study authors noted that the monitoring method used (vapor monitoring badges) may have
underestimated exposure, compared to levels measured using an active monitoring method.
COMMENTS;
The authors of this study draw a strong inference that exposures to VOCs from industrial
sources are causally associated with the observed gradients in respiratory symptoms in
children, even though the measured levels of volatile organics are far lower (in some cases by
many orders of magnitude) than levels which have been shown to be associated irritant effects
in laboratory and occupational studies. A subsequent review of this study (Collins, et. al
1994) questioned the grounds for postulating a relationship between outdoor VOC exposures
and symptoms, pointing out the low levels of measured VOCs in ambient air, the failure to
include consideration of indoor air exposures and other confounding factors, and
methodological concerns regarding the way the epidemiological data were analyzed. In
response to these comments, the authors of the original study agreed that it would have been
very desirable to have indoor air monitoring data, since a pilot study of homes in the region
indicated that indoor organic exposures were often higher than outdoor exposures. In light of
these factors, it does not appear that a clear causal relationship between exposures to ambient
VOC pollutants and respiratory symptoms has been demonstrated.
Reference:
Collin, J.J., J.M. Ramlow, and M.J. Teta. 1994. Respiratory and Irritant Health Effects of
Ambient Volatile Organic Compounds: The Kanawha Valley Study Getter). American Journal of
Epidemiology, 140: 72-75.
Air Toxics Study Summary B-23 Ware et d. 1993
-------
STUDY SUMMARY: Woodall et al. 1992
CITATION;
Woodall, G, Jr., L.H. Kuller, E.G. Talbott, J. Ramlow and B. Salthouse. 1992.
Epidemiological study for the Kenova and Ceredo, West Virginia and Catlettsburg, Kentucky
area, (Phase II). Final Report. Prepared for USEPA Region IV, Air Compliance Branch.
March 1992.
SOURCES OF EXPOSURE;
Oil refinery complex
HEALTH EFFECTS EVALUATED:
General health symptoms, including respiratory, cardiovascular, dermatological, and affective
(e.g., anxiety) disorders
CHEMICALS EXPOSED TO:
Not reported, except that the oil refinery is a major point source of criteria pollutants
EXPOSURE MEASURES:
Days of high emissions (> 10% higher than baseline for historically similar operating
conditions for any pollutant) from oil refinery complex. These were defined as "upset days."
Wind direction on upset days was also taken into account in a reanalysis of the data.
CONFOUNDING FACTORS ASSESSED;
Pre-existing chronic conditions (e.g., bronchitis, stroke, allergy, asthma) and smoking status
were reported, but not included in the analysis for the first phase of the study. In the cross-
sectional study, confounding factors assessed were smoking, age, sex, education level, and
residence in an air-conditioned home.
STUDY METHODS:
The study was initiated as a result of public complaints about air pollution and related health
problems (cancer, skin irritation, lethargy, memory loss, joint pain, and breathing difficulties)
in the area. The first part of the study compared the number of visits to health care facilities
during 30 "upset days" (defined as emissions > 10% higher than baseline for any pollutant)
with visits during non-upset days. A total of 636 emergency room and urgent care facility
visits were examined, excluding those unlikely to be related to air quality, such as trauma,
laboratory test requests, and worker's compensation releases. In the reanalysis, upset days
included only those days on which the prevailing wind direction was from die refinery toward
the study area; 20 upset days were identified and compared with 40 non-upset days. In the
second part of the study, 50 randomly selected complainants were interviewed regarding the
observed health symptoms, the types of observed pollution, and any non-health effects (e.g.,
Air Toxics Study Summary B-24 Woodall et al. 1992
-------
effects on plants). An attempt was made to verify reported symptoms by comparison with
medical records, but an insufficient number of consent forms were obtained. In the third
part of the study, 67 people in the exposed area and 66 people in a control unexposed area
were interviewed regarding the frequency and severity of specific symptoms.
STUDY RESULTS:
In the first phase of the study, there were a total of 337 visits that met that inclusion criteria
on upset days, compared with 299 on the non-upset days. The difference was borderline
significant only for those seen in the emergency room and discharged (p=0.06). There was
no significant difference between upset and non-upset days in the distribution of symptoms or
in the initial diagnosis. In the reanalysis (taking into account wind direction), there was a
significant increase in emergency room visits resulting in discharge on upset days compared to
nonupset days (p=0.031). There were no significant differences in reported symptomology
or initial diagnosis, but the statistical power of the study was reduced by the reduced number
of upset days. In the second phase of the study, respiratory complaints were most common,
followed by skin and eye irritation and neurologic complaints, primarily headaches and
fatigue. The major types of pollution reported were dust fallout and odor. In the third part
of the study, analysis of underlying chronic medical conditions showed there was a borderline
significantly higher prevalence of nervous disorders (p=0.06) and a significantly higher
prevalence of "other conditions" in the exposed area. The reported incidence of chronic lung
disease was nonsignificantly increased. However, several persistent respiratory and other
symptoms, including cough, breathing difficulties, headache, sore throat, and sinusitis, were
significantly increased in the exposed area. There were no significant differences between the
exposed and unexposed areas in demographics, cigarette smoking, or percent of population
with an air-conditioned home.
AUTHOR'S CONCLUSIONS;
The study authors concluded that "there is sufficient presumptive evidence of significant
health effects occurring in the Kenova/Ceredo area to warrant further prospective study,"
including ambient air monitoring and prospective surveillance of respiratory disease. The
study authors suggested that levels of specific pollutants, including hydrogen sulfide, sulfur
dioxide, and carbon monoxide, be monitored. They also suggested that objective medical
examination, including pulmonary function testing, could be conducted to reduce the
subjective nature of the data.
COMMENTS:
The study was well-designed and used two different study designs to determine whether acute
symptoms were correlated with pollutant exposure. In the first phase of the study, the study
population served as its own control. Therefore, control for confounding factors was not
necessary. The cross-sectional study was appropriately controlled and indicated that
respiratory symptoms and headache are more common in the exposed area, suggesting a
connection to air pollutant exposure. A major limitation of this study is the lack of any
monitoring data or reporting of the air pollutants to which the population was exposed. The
respiratory and irritative symptoms reported in the cross-sectional study are consistent with
exposure to criteria pollutants, the only pollutants specifically mentioned as being released.
Air Toxics Study Summary B-25 Woodall et al. 1992
-------
However, the study authors also suggest that levels of hydrogen sulfide, a HAP, be
monitored. Thus, although it would be difficult (even with monitoring data) to determine the
degree to which the observed effects can be attributed to HAP exposure, it is possible that
HAP exposure makes a contribution to the observed effect.
Air Toxics Study Summary B-26 Woodall et al. 1992
-------
TABLE 3-1. HAZARDOUS AIR POLLUTANTS DEFINED BY THE CLEAN AIR
ACT SECTION 112(b)
000050000
000051285
000051796
000053963
000056235
000056382
000057147
000057578
000057749
000058899
000059892
000060117
000060344
000060355
000062533
000062737
000062759
000063252
000064675
000067561
000067663
000067721
000068122
000071432
000071556
Formaldehyde
2 ,4-Dinitrophenol
Ethyl carbamate
2-Acetylaminofluorene
Carbon tetrachloride
Parathion
1 , 1 -Dimethyl hydrazine
beta-Propiolactone
Chlordane
Lindane
N-Nitrosomorpholine
Dimethyl aminoazobenzene
Methyl hydrazine
Acetamide
Aniline
Dichlorvos
N-Nitrosodimethylamine
Carbaryl
Diethyl sulfate
Methanol
Chloroform
Hexachloroethane
Dimethyl formamide
Benzene
1,1,1 -Trichloroethane
000106445
000106467
000106503
000106514
000106887
000106898
000106934
000106990
000107028
000107051
000107062
000107131
000107211
000107302
000108054
000108101
000108316
000108383
000108394
000108883
000108907
000108952
000110543
000111422
000111444
p-Cresol
1 ,4-Dichlorobenzene
p-Phenylenediamine
Quinone
1 ,2-Epoxybutane
Epichlorohydrin
Ethylene dibromide
1,3-Butadiene
Acrolein
Allyl chloride
1 ,2-Dichloroethane
Acrylonitrile
Ethylene glycol
Chloromethyl methyl ether
Vinyl acetate
Methyl isobutyl ketone
Maleic anhydride
m-Xylene
m-Cresol
Toluene
Chlorobenzene
Phenol
Hexane
Diethanolamine
Bis(2-chloroethyl)ether
-------
000050000
000072435
000072559
000074839
000074873
000074884
000075003
000075014
000075058
000075070
000075092
000075150
000075218
000075252
000075343
000075354
000075445
000075558
000075569
000076448
000077474
000077781
000078591
000078875
000078933
000079005
Formaldehyde
Methoxychlor
p,p'-DDE
Bromomethane
Chloromethane
lodomethane
Ethyl chloride
Vinyl chloride
Acetonitrile
Acetaldehyde
Dichloromethane
Carbon disulfide
Ethylene oxide
Bromoform
1 , 1 -Dichloroethane
Vinylidene chloride
Phosgene
2 -Methyl aziridine
Propylene oxide
Heptachlor
Hexachlorocyclopentadiene
Dimethyl sulfate
Isophorone
1 ,2-Dichloropropane
Methyl ethyl ketone
1 , 1 ,2-Trichloroethane
000106445
000114261
000117817
000118741
000119904
000119937
000120809
000120821
000121142
000121448
000121697
000122667
000123319
000123386
000123911
000126998
000127184
000131113
000132649
000133062
000133904
000140885
000151564
000156627
000302012
000334883
p-Crcsol
Baygon
Bis(2-ethylhexyl)phthalate
Hexachlorobenzene
3 ,3-Dimethoxybenzidine
3, 3 '-Dimethyl benzidine
Catechol
1 ,2,4-Trichlorobenzene
2,4-Dinitrotoluene
Triethylamine
N ,N-Dimethylaniline
1 ,2-Diphenylhydrazine
Hydroquinone
Propionaldehye
1 ,4-Dioxane
Chloroprene
Tetrachloroethylene
Dimethyl phthalate
Dibenzofurans
Captan
Chloramben
Ethyl acrylate
Ethylene imine (aziridine)
Calcium cyanamide
Hydrazine
Diazomethane
13
-------
000050000
000079016
000079061
000079107
000079118
000079345
000079447
000079469
000080626
000082688
000084742
000085449
000087683
000087865
000088062
000090040
000091203
000091225
000091941
000092524
000092671
000092875
000092933
000094757
Formaldehyde
Trichloroethylene
Acrylamide
Acrylic acid
Chloroacetic acid
1 , 1 ,2,2-Tetrachloroethane
Dimethyl carbamoyl chloride
2-Nitropropane
Methyl methacrylate
Pentachloronitrobenzene
Dibutyl phthalate
Phthalic anhydride
Hexachlorobutadiene
Pentachlorophenol
2,4, 6-Trichlorophenol
o-Anisidine
Naphthalene
Quinoline
3,3' -Dichlorobenzidene
Biphenyl
4-Aminobiphenyl
Benzidine
4-Nitrobiphenyl
2,4-D
000106445
000463581
000510156
000532274
000534521
000540841
000542756
000542881
000584849
000593602
000624839
000680319
000684935
000822060
001120714
001319773
001330207
001332214
001336363
001582098
001634044
001746016
007550450
007647010
p-Cresol
Caibonyl sulfide
Chlorobenzilate
2-Chloroacetophenone
4, 6-Dinitro-o-cresol
2,2,4-Trimethylpentane
1 ,3-Dichloropropene
Bis(chloromethyl)ether
2,4-Toluene diisocyanate
Vinyl bromide
Methyl isocyanate
Hexamethylphosphoramide
N-Nitroso-N-methylurea
Hexamethylene- 1 , 6-diisocyanate
1,3-Propane sultone
Cresol
Xylenes
Asbestos
Polychlorinated biphenyls
(a) 4-chlorobiphenyl
(b) decachlorobiphenyl
Trifluralin
Methyl tert butyl ether
2,3,7, 8-Tetrachlorodibenzo-p-
dioxin
Titanium tetrachloride
Hydrochloric acid
14
-------
000050000
000095476
000095487
000095534
000095807
000095954
000096093
000096128
000096457
000098077
000098828
000098862
000098953
000100027
000100414
000100425
000100447
000101144
000101688
000101779
000105602
000106423
Formaldehyde
o-Xylene
o-Cresol
o-Toluidine
2,4-Toluene diamine
2,4,5-Trichlorophenol
Styrene oxide
1 ,2-Dibromo-3-chloropropane
Ethylene thiourea
Benzotrichloride
Cumene
Acetophenone
Nitrobenzene
4-Nitrophenol
Ethylbenzene
Styrene
Benzyl chloride
4,4-Methylene
bis(2-chloroaniline)
Methylene diphenyl diisocyanate
4 ,4'Methylenedianiline
Caprolactam
p-Xylene
000106445
007664393
007723140
007782505
007803512
008001352
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
000000000
p-Cresol
Hydrogen fluoride
Phosphorus
Chlorine
Phosphine
Toxaphene
Antimony Compounds
Arsenic Compounds
Beryllium Compounds
Cadmium Compounds
Chromium Compounds
Cobalt Compounds
Coke Oven Emissions
Cyanide Compounds
Glycol ethers
Lead Compounds
Manganese Compounds
Mercury Compounds
Fine mineral fibers
Nickel Compounds
Polycyclic Organic Matter
Radionuclides
Selenium Compounds
15
-------
TECHNICAL REPORT DATA
(PLEASE READ INSTRUCTIONS ON THE REVERSE BEFORE COMPLETING)
1. REPORT NO. 2.
EPA-456/R-97-001
4. TITLE AND SUBTITLE
SUMMARY AND ANALYSIS OF AVAILABLE AIR TOXICS HEALTH
EFFECTS DATA
7. AUTHOR(S)
1CF KAISER INTERNATIONAL
FAIRFAX, VA 22031
9. PERFORMING ORGANIZATION NAME AND ADDRESS
12. SPONSORING AGENCY NAME AND ADDRESS
OFFICE OF AIR QUALITY PLANNING AND STANDARDS
ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
THIS REPORT REVIEWS AND ANALYZES RECENT INFORMATION REGARDING THE QUALITATIVE LINK
BETWEEN EXPOSURE TO TOXICS AIR POLLUTANTS AND ADVERSE HUMAN HEALTH EFFECTS. IT
DESCRIBES THE TYPES OF HEALTH EFFECTS, SUMMARIZES FINDINGS ON THE MOST IMPORTANT
STUDIES, AND POINTS OUT MAJOR DATA GAPS AND NEEDS FOR ADDITIONAL RESEARCH.
1 7. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
18. DISTRIBUTION STATEMENT
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI FIELD/GROUP
21. NO. OF PAGES
22. PRICE
------- |