U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
PB-289 126
Investigation of Selected Correlations
Between Industrial Activity and
Community Disease
Enviro Control, Inc, Rockville, MD
Prepared for
Environmental Protection Agency, Washington, DC Office of Toxic Substances
Aug 78
-------
United Stoto«
Etwironmontal Protection
Agency
Office of
Toxic Subttoncet
Weihington Of 20460,
August 1978
Toxic Su
c/EPA
Investigation of Selected
Correlations Between
Industrial Activity and
Community Disease
-------
TECHNICAL REPORT DATA
(Please read InttrucHont on the rtvefu before completing)
1. REPORT NO.
EPA—560/6-78-004
2.
3. R
4. TITLE AND SUBTITLE
Investigation of Selected Correlations Between
Industrial Activity and Community Disease
B. REPORT DATE
August 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR
-------
EPA-560/6-78-004
INVESTIGATION OF SELECTED CORRELATIONS
BETWEEN
INDUSTRIAL ACTIVITY AND COMMUNITY DISEASE
Contract No. 68-01-4304
Or. Vincent J. DeCarlo
Sxxpervisor, Special Actions Group
Office of Toxic Substances
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
Prepared for
Environmental Protection Agency
Office of Toxic Substances
Washington, D.C. 20460
August 1978
-------
NOTICE
This report has been reviewed by the Office of Toxic Substances,
Environmental Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the views and
policies of the Environmental Protection Agency. Mention of trade names
or commercial products is for purposes of clarity only and does not con-
stitute endorsement or recommendation for use.
11
-------
PREFACE
This investigation was conducted by Enviro Control, Inc., under con-
tract with the U.S. Environmental Protection Agency's Office of Toxic
Substances (Contract No. EPA 68-01-4304).
Each chapter of this report was the responsibility of one of the six
principal authors and, for the most part, prepared by that author, with
contributions as needed from the others. The chapters and their prin-
cipal authors are as follows:
Chapter Principal Author
I Introduction
II Summary
Willard W. Perry
III Conclusions and Recommendations
IV Methodology Robert Goldsmith
V Copper Smelting Case Study Margaret E. Mattson
VI Steel Manufacturing Case Study Mary A. Urbanek
VII Bituminous Coal Mining Case Study Carl B. Bailey
VIII Viscose Rayon Case Study Bertram W. Berney
The authors wish to acknowledge the many technical and editorial con-
tributions of John D. Morton who served as Program Manager throughout the
study. Although not listed as one of the principal authors, Dr. Tee L.
Guidotti contributed several important pieces on the etiology of the
diseases of interest. These research papers are included in the appen-
dices under his name.
The authors are indebted to the many individuals and organizations
who provided data and information for use in the study. For their
extra efforts and cooperativeness, special mention is made of:
Tom Brian, Larry Jackson; Tennessee Center for Health Statistics
Dr. Frank Zsoldas, Ms, Bowers; Wyoming General Hospital
Clyde Bridger; Illinois Center for Health Statistics
Barry Evans; EPA Environmental Photographic Interpretation Complex
Albert Evans; Consultant
Dolores Fintress, Larry Finnegan, Roger Forstensen; Census Bureau
Ed Digon; Pennsylvania Division of Epidemiologic Research
Ase Wijmuller; Hendrickson Corporation
Dr. Joseph Lyon; Utah Cancer Registry
Justice Manning; EPA Research Triangle Park
John Wilson, Janet VanSwearinger; Montana Bureau of Records and
Statistics
David Bigler; U.S. Steel Corporation
iii
-------
John Brockert; Utah Bureau of Health Statistics
Dr. J. Wanless Southwick; Utah Department of Environmental Health
Dennis Ha'ddow; Montana Bureau of Air Pollution
Ann Wesemann; Illinois Department of Public Health
C. T. Mitchell; Enka Company
Robert Foster; Tennessee Department of Public Health
Dennis Lohman, John Clark, Robert Yuhnke; Pennsylvania Department
of Environmental Resources
Col. McCall; West Virginia Department of Public Health
Wayne Montney; Illinois EPA
Michael Bursic; Bureau of Mines
John Bater; U.S. Geological Survey
John Rakes; Wyoming County Public Health Office
Dr. Alexander Kelter; Arizona Department of Health Services
Lester Foote; Arizona Office of Vital Statistics
Ray Kopische; Gila County Air Quality Officer
iv
-------
TABLE OF CONTENTS
Chapter Page No.
INTRODUCTION 1-1
A. Study Purposes and Objectives 1-1
B. Study Approach 1-3
C. Background 1-4
D. Outline of Remainder of Report 1-7
II SUMMARY II-l
A. Summary of Case Studies and Findings II-l
B. Discussion of Results 11-29
III CONCLUSIONS AND RECOMMENDATIONS III-l
.. A. Conclusions and Recommendations Pertaining
to Industry-Disease Relationships III-l
B. Conclusions and Recommendations Pertaining
to the Multiple Regression Analysis III-3
C. Conclusions and Recommendations Pertaining
to Methodology for Investigating Industry-
Disease Relationships III-5
IV STUDY METHODOLOGY IV-1
A. Overview IV-1
B. Initial Correlations IV-1
C. Screening and Validation - IV-2
D. Hypothesis Development and Test IV-17
E. .Site Selection IV-20
F. Data Collection IV-21
G. Industry Characterization IV-22
H. Pathway Analysis IV-22
I. Etiology Analysis IV-23
J. Air and Water Quality Analysis IV-24
K. Morbidity/Mortality Analysis IV-25
-------
TABLE OF CONTENTS
(continued)
Chaptei-
Page No.
COPPER SMELTING CASE STUDY
A. Introduction and Summary
B. Background
C. Approach
D. Results
E. Conclusions and Recommendations
V-l
V-l
V-4
V-20
V-29
V-43
VI
STEEL MANUFACTURING CASE STUDY VI-1
A. Introduction and Summary VI-1
B. Background VI-5
C. Study Description (Activities and Methods) VI-24
D. Results of Analysis VI-58
E. Conclusions and Recommendations VI-69
Vll
BITUMINOUS COAL MINING CASE STUDY
A. Introduction
B. Hypothesis Development
C. Hypothesis Test
D. Conclusions and Recommendations
VII-1
VII-1
VII-3
VII-38
VII-57
VIII VISCOSE RAYON INDUSTRY CASE STUDY
A. Introduction
B. Development of Hypothesis
C. Test of Hypothesis
D. Conclusions
VIII-1
VIII-1
VIII-2
VIII-10
VIII-27
vi
-------
TABLE OF CONTENTS
(continued)
Appendix > Page No.
A Atmospheric Transport A-l
B Copper Smelting Industry Characterization B-l
Cl Etiology of Diseases Associated with Copper
Smelting Cl-1
C2 Selected Annotated Bibliography of Literature
Relating Digestive Diseases and Copper Smelting C2-1
D Maps of the Copper Study Areas D-l
E History of Area Demography and Industrial
Activity in Gila and Pinal Counties of Arizona E-l
F Mortality and Pollution Data in Support of
Copper Smelting Case Study F-l
G Selected Disease Categories and Included ICDA
Codes for Copper Smelting Case Study G-l
H Disease Associations in American Steel Mining
Communities H-l
I Aerial Photo Analysis Performed by Environmental
Photographic Interpretation Complex (EPIC) 1-1
J Mortality Rate Computation Methodology and
ECI, SSI and ICDA Cause of Death Code Conversion
Tables J-l
Kl Etiology of Cardiovascular Diseases Associated
with Bituminous Coal Mining Kl-1
K2 , Etiology of Respiratory Diseases Associated
with Bituminous Coal Mining K2-1
L Hospital Record Abstract Forms L-l
M Effects of Exposure to Carbon Disulfide and
Hydrogen Sulfide M-l
vii
-------
Chapter I
INTRODUCTION
A. STUDY PURPOSE AND OBJECTIVES
This report documents the methods and results of an investigation of
selected correlations between industrial activity and community disease.
It constitutes the second phase of an EPA-sponsored effort to develop
systematic and objective methodology for investigating the possible in-
dustrial origins of chronic community disease. The first phase/ per-
formed by System Sciences, Inc.(SSI), was a multiple regression analysis
of county mortality and industry data to identify statistical correla-
tions between specific industries and diseases [1]. The purpose of the
present study is to investigate these statistically derived industry-
disease correlations to determine if they have a plausible physical and
etiological basis. By so doing, the study provides a measure of the
effectiveness of the multiple regression analysis for generating valid
associations between industry and community disease.
In this light, the ultimate objective of the study is methodologi-
cal, i.e., to develop and evaluate methods for analyzing and investi-
gating regression-derived industry-disease correlations to assess their
physical and etiological plausibility. These methods vary in complexity
from simple screening analyses and literature surveys to in-depth epi-
demiological field investigations. To provide the basis for methodology
development, the specific objectives of the study are:
[1] System Sciences, Inc., Investigation Into the Industrial Correlates
of Environment Related Mortality, 1976.
1-1
-------
(1) To screen the regression-derived correlations to eliminate
those that are weak and the likely result of confounding
factors, and to identify those that are statistically
sound and promising candidates for true associations.
(2) To perform field studies of a selected number of industry-
disease correlations to investigate the possible existence
t of causal relationships by the verification (or lack of
verification) of: (a) an identifiable toxic substance
emitted by the industry, (b) a feasible pathway by which
the toxic substance can reach a population at risk,
(c) the presence of the toxic agent in the environment
to which the population at risk is exposed, (d) elevated
mortality or morbidity in the population at risk, and
(e) a known etiological relationship between the disease
arid the toxic agent.
(3) To review the results of (1) and (2) above and evaluate the
effectiveness of the multiple regression approach for identi-
fying possible industrial sources of community disease.
Most of the study effort was devoted to the field investigations
which were conducted over a 12-month period. The objective of these
investigations was not to prove or disprove that the industry was the
cause of the associated disease. Rather, it was to determine if a
rational causal hypothesis could be derived from and supported by readily
available field data. The degree of "proof" that can be associated with
these hypotheses depends upon the amount and quality of field data and
the evidence they provide, which is generally quite limited. Thus,
the field investigations are a first step in the more extensive effort
required to establish convincing proof of the existence of true rela-
tionships between the industries and the associated diseases. They are,
in reality, investigations to assess the plausibility of cause-and-
eff*ct associations identified by other means. Their purpose is to pro-
vide the basis for an intermediate decision, before the commitment of
the much more extensive effort required to establish proof.
1-2
-------
B. STUDY APPROACH
To satisfy the study objectives, in-depth case studies of four
industry-disease correlations were conducted. Three of these correla-
tions were selected by us from those generated by the SSI multiple re-
gression analysis after we had made a thorough screening and validation
of the regression results. The fourth correlation was a subjective cor-
relation arising from community complaints and from an independent exam-
ination of county mortality data. The fourth correlation was included
in the investigations to test an alternative method of selection and to
broaden the chance of useful field studies if the regression method
failed.
The four industry-disease correlations for which in-depth field in-
vestigations were conducted were:
. • Copper Smelting -Digestive Diseases
•, Steel Manufacturing -Cancer of the Digestive Organs
• Bituminous Coal Mining -Respiratory and Cardiovascular Diseases
• Viscose Rayon Manufacturing -Cardiovascular Diseases
The field investigations consisted of case studies of specific sites,
«
each containing a suspect industry and a population at risk. To deter-
mine if disease-specific mortality was greater than expected in the
population at risk, comparisons were made with demographically similar
communities not containing the industry. Where mortality was elevated
in the population at risk, data collected on industrial processes, emis-
sions and effluents, air and water quality,, meteorology and hydrology,
demographics, etiology, and any other pertinent factors provided the
basis for interpreting the apparent industry-disease association in
terms of possible causal hypotheses. When available, morbidity data
were collected and analyzed. For each of the four case studies, at
least two separate sites were investigated to provide a measure of
replication consistency. In general, the first site was used for devel-
opment of a set of causal hypotheses that were then tested for applica-
bility at a second site.
1-3
-------
Due to the wide variation in the location of the study sites and in
the type and amount of data available in the different states and re-
gions, the case studies differed considerably in the specific methods
and analyses used. This turned out to be particularly useful in the
development of methodology, since it offered the study team the oppor-
tunity to experiment with and evaluate a variety of techniques. In
general though, the investigation of whether or not the population at
risk experienced higher-than-expected mortality was accomplished through
comparisons of age-adjusted, disease-specific mortality rates with those
of demographically similar (matched-control) areas.
The study was conducted over a 16-month period, in which the first
four months' effort consisted of the selection of industry-disease cor-
relations to be investigated and the detailed planning of the case
studies. The results of this initial planning phase have been docu-
j
mented earlier [2]. The four case studies, documented in the present
report, were carried out over the 12-rmonth period following the initial
planning phase.
C. BACKGROUND
For the most part, EPA's motivation for investigating potential
environmental health problems and sources has been laboratory and occu-
pational health findings or catastrophic pollution incidents which indi-
cate that a particular substance is harmful. Since there are a great
many potential problems and a limited budget for investigations, EPA can
pursue only a small fraction of the possible environmental health prob-
lems that are posed by the numerous researchers and interest groups. In
many instances, it is the enthusiasm and persistence of the proponents
of the investigation which determine the priority of a proposed study.
In the search for more objective methods for identifying critical
environmental health problems that should be studied, EPA initiated the
methodological development and evaluation effort of which this study is
[2] Enviro Control, Inc., Plan for the Investigation of Selected Corre-
lations Between Industrial Activity and Community Disease, Volumes
I and II, February 1977.
1-4
-------
a part. Objectivity is achieved by relying on statistical correlations
to indicate the industrial sources that are potentially the most haz-
ardous to community health. The correlations are made between the level
of industrial activity for U.S. counties and the county mortality rates,
for various industries and diseases. By this objective means, the par-
ticular industries that are most strongly correlated with specific dis-
eases are identified. These correlations are then investigated epidemi-
ologically, to determine if a reasonable causal hypothesis can be de-
veloped to explain them. In the development of this basic methodology,
SSI performed the statistical work to establish the industry-disease
correlations, and this present study describes the work by Enviro Con-
trol, Inc. (ECI) to investigate the causal basis for the correlations.
It has been recognized that the success of the regression method-
ology is not a foregone conclusion. There are a number of reasons why
a multiple regression analysis of county industry and disease data may
,. i
not be very effective in identifying industrial sources of community
diseases. First, the industry categories are quite broad and contain
many different kinds of industry. Thus, a particular industry that is
strongly contaminating the community may represent only a small part of
the industrial category, and therefore show up as a weak correlation.
Second, the emissions of harmful substances could be largely a plant
management problem which occurs in a small portion of the plants com-
prising a given industry, again resulting in a weak correlation. Third,
the measure of industrial activity for a county may be a poor measure of
the amount of disease-causing agents emitted by the industry. Finally,
the polluting industrial facility in a county may affect only a small
portion of the county population or it may affect the populations in ad-
jacent counties, and the increase in the overall mortality rate for a
county with the industry may be too small to detect.
For these reasons, it was by no means clear that the regression
analysis would yield valid correlations. True industry-disease effects
could be lost in the noise, and any correlations that were found could
be the result of other unknown factors. Consequently, the principal
purpose of the investigation reported here is to determine if the cor-
1-5
-------
relations that were found have a plausible causal basis. In this way,
this study is an evaluation of the feasibility of generating meaningful
correlations by means of a multiple regression analysis of county mor-
tality and industry data.
Although the total methodology consisting of the county data corre-
lation analysis together with the epidemiological investigation is
unique, a number of studies have been conducted which are relevant to
this approach—either to the correlation analysis or to the epidemio-
logical investigation—which lends credence to the basic concept.
Studies based on data from Cancer by County [3] are examples of the
correlation analysis approach. The methodology used in these types of
studies is exemplified by the publication of Blot et al., entitled
"Cancer Mortality in U.S. Counties with Petroleum Industries" (Science
295:51-53, 1977). Counties with employment in the petroleum industry
over a certain -level were selected and compared with a group of "con-
trol" counties of comparable demography and geography. Ratios of age-
adjusted mortality rates were then compared for 23 cancer sites. This
methodology does not consider occupation, location or duration of resi-
dence, or lifestyle (e.g., diet and smoking) of the decedents and,
therefore, is more useful as a screening tool.
The type of study which we have denoted as a "epidemiological in-
vestigation" attempts to clarify an association by considering smaller
geographic units, location of residences with respect to the industrial
pollution pathway, and specific compounds in the industrial emissions
or effluents. An example of this approach is a recent study by Mata-
noski et al. of cancer mortality near an arsenical pesticide plant in
Baltimore, Maryland (NTIS, PB-255-019, May 1976). In this case, census
tracts near the plant were compared with tracts at some distance from
the .plant relative to cancer mortality. Death certificates were ob-
tained, and decedents were categorized by place of last residence. Age-
adjusted mortality rates were computed and compared for statistical
significance.
[3] Mason, T. J., and F. W. McKay, U.S. Cancer Mortality by County: 1950-
1969, DHEW Publ. No. (NIH) 75-780, U.S. Government Printing Office,
Washington, D.C. 1973.
1-6
-------
D. OUTLINE OF REMAINDER OF REPORT
The balance of the report is organized into seven chapters as fol-
lows.
Chapter JJ (Summary)—This is a complete summary of the study con-
sisting of a description of each of the four case studies and a discus-
sion of the results. In the latter, the results of all four case stud-
ies are discussed in the light of the study objectives. This consti-
tutes the only integrated discussion of the results of the four cases
in this report.
Chapter III (Conclusions and Recommendations)—This presents and
discusses the overall conclusions and recommendations derived from the
study.
Chapter TV (Study Methodology)—This chapter describes the case
study methodology. Also included is a description of the methodology
for screening, validating, and selecting the industry-rdisease correla-
tions for field study.
Chapters V, VI, VII, and VIII (Case Studies)—These chapters pro-
vide detailed documentation of the four case studies. Each chapter con-
stitutes a separate case study, presented in the order of: Copper
Smelting, Steel Manufacturing, Bituminous Coal Mining, and Viscose Rayon
Manufacturing.
1-7
-------
Chapter II
SUMMARY
A. SUMMARY OF CASE STUDIES AND FINDINGS
1. Case Study Selection
To select industry-disease correlations for field investigation, the
results of the regression analysis were first subjected to a fairly
rigorous screening and validation process. Weak correlations and cor-
relations unfit for study were eliminated from consideration. A pro-
gressive series of statistical tests were performed on the remaining
correlations to identify those that could be validated by independent
statistical means. The final test was a matched-control test in which
the mortality rates for the top ten counties in industrial activity were
compared to the mortality rates for a set of control counties which were
demographically similar to the top ten. For comparison purposes,
industry-disease associations were identified by means other than the
regression analysis (e.g., occupational associations) and subjected to
the same type of validation testing. The correlations that survived
these screening and validation procedures are listed in Table II-l.
Four industry-disease correlations were selected from this list for
in-depth field investigation. These are listed below.
Industry
Copper
Smelting
Steel
Manufacturing
Bituminous
Coal Mining
Viscose Rayon
Disease
Digestive Disease
Cancer of Digestive Organs
Cardiovascular Disease,
Respiratory Disease
Cardiovascular Disease
Source of
Initial Correlation
Multiple 'Regression
Analysis
Multiple Regression
Analysis
Multiple Regression
Analysis
Other
Chapter
V
VI
VII
VIII
Each of these case studies is described and discussed in detail in a
separate chapter of this report as noted. This present section gives a
brief summary of each investigation and its findings.
II-l
-------
Table II-1
CANDIDATE INDUSTRY-DISEASE ASSOCIATIONS FOR FIELD STUDY
SIC
INDUSTRY
DISEASE
3357
102R
121R
3312
2321
3331
3421
34RK
2879
2823
Non-Ferrous Wire Drawing
Copper Ore Mining*
Bituminous Coal Mining*
Blast Furnaces*
Men's Shirts & Nightwear
Primary Copper*
Cutlery
Fabricated Metal
Products
Pesticides & Agricul-
tural Chemicals
Viscose Rayon*
Neoplasm, digestive organs
Neoplasm, intestine
Acute myocardial infarction
Respiratory diseases
Acute interstitial & bronchopneu-
monia
Bronchitis, emphysema & asthma
Emphysema
Asthma
Other respiratory diseases
Major cardiovascular disease
Ischemic heart disease
Acute myocardial infarction
Chronic ischemic heart disease,
angina pectoris
Diseases of veins, lymphatics &
circulatory system
Neoplasm, digestive organs
Major cardiovascular disease
Ischemic heart disease
Alcoholic cirrhosis of liver
Digestive diseases
Respiratory disease
Hypertensive disease
Respiratory diseases
Cardiovascular Disease
*Associations selected for field investigation.
II-2
-------
2. Copper Smelting Case Study
a. Basis for Correlation
The multiple regression analysis showed a relatively strong corre-
lation between the copper mining industry and respiratory, cardiovascu-
lar, and digestive diseases in white females. Upon examination and
validation of the correlation data, it was found that the copper
smelting industry is generally located in the same counties as the
mines and is the more logical source of the associated diseases. The
matched-control county test supported this view when it resulted in
stronger correlations for copper smelting than for copper mining. The
strongest of the disease correlations with copper smelting was digestive
disease.
b. Study Site Selection
For the purpose of causal hypothesis development, three sites were
selected in the Arizona counties of Gila and Final. These are shown in
Table II-2. Note that one of the sites (Hayden) contained two copper
smelters; hence, the initial field study involved the investigation of
four copper smelters. The criteria for selection were high industrial
activity, elevated mortality rates in digestive diseases, absence of
other confounding industries, and the ready availability of industry
data, air and water quality data, and mortality data. Particular fea-
tures of the mortality data in Arizona were the inclusion of length of
residence information on the death certificates and the availability of
most of the death certificate information on computer tape.
Anaconda, Montana, was selected as the primary hypothesis test site.
Of all the candidate copper smelting sites, it possessed the largest
population at risk for which acceptable mortality data were available,*
a characteristic that was felt to be very important since the investi-
gation of the Arizona sites had shown that larger exposed populations
* Tacoma, Washington, had a larger population but mortality data were
not available for the level of subdivisions needed, i.e., for census
tracts.
II-3
-------
were necessary. Also, there were several populations in the vicinity
of Anaconda that appeared to be demographically similar to Anaconda's
and hence collld 'serve as control populations.
Table II-2. COPPER SMELTING SITES INVESTIGATED
' *•
HYPOTHESIS
DEVELOPMENT
HYPOTHESIS
TEST
SITE
Miami
Hayden
San Manuel
Anaconda
Magna
Douglas
Ajo
Morenci
COUNTY/STATE
Gila/Arizona
Gila/Arizona
Pinal/Ar izona
Deer Lodge/Montana
Salt Lake/Utah
Cochi s e/Ar izona
Pima/Ar izona
Greenlee/Ar izona
SMELTING COMPANY
Inspiration Consolidated
Copper Company
Kennecott Copper Corp . ;
ASARCO
Magma Copper Company
Anaconda Company
Kennecott Copper Corp.
Phelps-Dodge Corporation
Phelps-Dodge Corporation
Phelps;-Dodge Corporation
In the investigation of the Gila and Final County sites, mortality,
demographic, health, and air quality data were acquired for the entire
state. Consequently, it was possible to investigate the three other
copper smelting sites in Arizona with little extra effort. These were
the Phelps-Dodge plants at Ajo, Morenci, and Douglas. In addition, the
data collected in Utah for the steel manufacturing case study permitted
the ready investigation of the copper smelting site at Magna, Utah. All
told, a total of eight copper smelter sites (nine copper smelters) were
investigated: three for hypothesis development and five for hypothesis
test. These are listed in Table I1-2. This large number of sites was
a feature of the copper smelting investigation that was not repeated
for any of the other case studies (each of the other three case studies
investigated two sites only).
11-4
-------
c. Industry Characterization
Detailed descriptions of both general and site-specific aspects of
the copper smelting industry were prepared, including the production and
emission histories and characteristics. The key emissions were deter-
mined to be sulfur dioxide and particulates containing numerous trace
. metals including lead, arsenic, cadmium, and copper. The industry
generally recycles its water so that there are no obvious effluent
problems, although indirect contamination of water supplies may be pos-
sible from tailings ponds or surface leachates which contain industry-
related pollutants.
i
i
d. Etiological Basis
! Research into the etiology of diseases likely to be associated with
| copper smelting was conducted through literature searches and consulta-
i
I tions with medical experts (see Appendix C). Toxic agents produced by
I the copper industry which have been linked to diseases are summarized
I .:
I below:
I
i • Sulfur Oxides—Aggravation of respiratory diseases including
| asthma, chronic bronchitis, emphysema; eye irritation; sus-
I pected in respiratory and gastrointestinal cancers; dystrophic
liver functions suspected.
• Arsenic—Poorly differentiated epidermoid bronchogenic car-
j • cinoma; skin cancer; possibly colon cancer and/or neurological
disorders.
• Lead—Lead colic; anemia; psychotic states; peripheral neu-
ropathy; may be nephrotoxic.
• Cadmium—Correlated with hypertension, liver disease, bone
disease, kidney impairment, spleen and thyroid dysfunction.
• Manganese, Zinc—Congenital abnormalities.
• Nickel — Dermatological irritation; cancer of the sinuses;
'. hemophagic pneumonia (nickel carbonyl).
• Copper—Orally causes nausea, vomiting and diarrhea; inhaled
. causes copper fume fever.
In general, there does not appear to be a strong linkage between the
emitted toxic agents and digestive diseases, although both sulfur dioxide
and cadmium have been implicated to some extent in liver disease.
II-5
-------
e. Collection of Data
Industry Data—Age, production, and emission characteristics were
collected for all smelters. Emission data included stack height and
control history and the amount of sulfur dioxide (802) and particulate
matter discharged. Most plants installed acid plants for the control
and recovery of S02 emissions in the early 1970's. However, one plant
(ASARCO) does not have any controls and uses the technique of curtail-
ment and an extremely tall stack (.1,000 feet) to satisfy state pollu-
tion regulations.
Air and Water Quality Data—Air quality data were obtained from the
State Air Quality Reports for all smelter towns back to 1969 or 1968.
These data consisted of SOj and particulate matter concentrations. The
measured constituents of the particulate matter were benzene solubles,
nitrates, sulfates, and most of the trace metals, of which arsenic, cad-
mium, copper, 'and lead were the most significant. The key problem with
the air quality data was the lack of data for most of the nonsmelter
towns of interest, i.e., the control towns. Water quality data were
collected for the Gila and Final County sites in Arizona from the State
Water Control Bureau. Water chemistries were obtained for the public
wells for 20 to 30 years at 3-year intervals. The data included the con-
centrations of minerals and trace metals, but in general it was spotty.
No water quality data was obtained from smelter sites other than those
in Gila and Final Counties in Arizona.
Meteorology/Hydrology Data—Data on wind patterns and inversion
characteristics were obtained for the several sites from state and local
officials. Qualitative hydrologic characteristics were obtained for
each area, including information on water "sources for the potable water
systems.
Health Data—Mortality data for all diseases were collected for all
eight sites investigated. The usual form of data was magnetic tape;
however, for the hypothesis development sites in Ginal and Final
Counties of Arizona, hard copies of the death certificates were also
obtained and coded. The hard copy was needed to acquire occupational
information and to obtain data prior to 1968. The years of mortality
I-I-6
-------
data obtained were as followst
1957-1975: Miami, Hayden, San Manuel (Arizona)
' 1968-1975: Ajo, Morenci, Douglas (Arizona)
1968-1975: Anaconda (Montana)
1966-1975: Magna (Utah)
No quantitative morbidity data were acquired. In Gila and Final.
Counties, the hospitals were all owned or influenced by the copper com-
panies and would not release the health information until stronger evi-
dence could be provided that the smelters were the source of community
disease. No attempt was made to acquire morbidity data in the other
sites investigated, due to a lack of sufficient time to process these
data.
Qualitative health information was provided by hospital and state
health officials. For example, it was observed by at least two hospital
officials that alcohol abuse was a problem in the smelter towns in Gila
and Final Counties. This was attributed to the ready availability of
money supplied by steady copper smelting employment and the lack of cul-
tural activities or other means of entertainment. No health officials
had observed any unusual incidence of digestive diseases other than
alcoholic cirrhosis of the liver.
Body burden data on lead, arsenic, and cadmium in Arizona were
available from recent research performed by the Center for Disease Con-
trol. These data were used in developing and testing the causal hypoth-
eses in Arizona.
Demographic Data—Data on population and socioeconomic character-
istics were obtained for Arizona, Montana, and Utah from the Census
Bureau. Fifth count, file C, tapes from the 1970 Census, which provided
the data at the level of enumeration district, were used. Enumeration
district maps were acquired to enable the correlation of the street ad-
dress and town of residence to enumeration district.
11-7
-------
f. Analysis of Data
The basic method of analysis consisted of dividing the mortalities
into 34 disease categories, and then comparing the age-adjusted mortal-
ity rates for the towns hypothesized as being exposed to the rates for
the towns assumed to be unexposed, i.e., the controls. Since most towns
are a considerable distance apart and, in general, there is no inter-
vening population to speak of, it was a relatively simple matter to
separate the population into potentially exposed and unexposed cate-
gories. Because the exposed populations for each site were relatively
small, various aggregations of towns were examined to facilitate statis-
tical significance testing. The Mantel-Haenszel chi-sguare test was
used to test the statistical significance of the relative risk values
for the exposed population.
The mortality data were examined in the light of air and water
quality data or body burden data to associate excess mortalities with
particular chemicals found in the air, water, or body. In this way,
hypotheses were developed at particular sites and tested at others.
Attempts were made to examine the effect of duration of residence; how-
ever, there were insufficient mortality data to permit any significant
findings. To avoid possible confounding by occupational exposure, the
principal findings were based ori"mortalities of females; the copper
smelting industry has not employed women in the smelters until the past
two or three years.
In Arizona, it was difficult to find suitable control populations
whose socioeconomic characteristics matched those of the population at
risk. The smelter towns had a different socioeconomic structure than
the nonsmelter towns. For this reason, the population of the entire
state was used as the control.
i
Due to the lack of suitable control areas, for which air and water
quality or body burden data were available, a dose-response type of
analysis was performed with the Arizona data. The approach here was to
correlate the concentration of a particular pollutant with the mortality
rates for the smelter towns and for the towns near smelters for which
II-8
-------
data were available. To assist in this correlation and to provide
assurance that the resulting correlations were not due to socioeconomic
•\
confounding, a factor analysis was performed with all the data for the
towns consisting of mortality rates, socioeconomic characteristics, air
quality data, and industry characteristics such as age, production, and
!
emission characteristics.
g. Summary of Findings for Copper Smelting Case Study
(1) Causal Hypothesis Development
From the study of the eight copper smelter sites, four causal hy-
potheses were developed between the industry and four separate diseases,
namely, respiratory disease, digestive disease, cerebrovascular disease,
and cancer of the colon and pancreas. Of these four diseases, the first
two had been correlated with the copper industry by means of the multiple
regression analysis; the last two had not.
(2) Respiratory Disease Findings
A strong correlation was found between the ambient levels of SC>2 and
particulates and acute respiratory conditions such as pneumonia, influ-
enza, pleurisy, pulmonary congestion, and lung edema. This correlation
was demonstrated in two different ways. First, in the matched-control
analysis for all smelter sites in Arizona, Montana, and Utah, excessive
age-adjusted mortality rates in females were associated with excessive
levels of SC>2 and particulates, and lower mortality rates were associ-
ated with lower levels of S02 and particulates. Second, for 11 towns
in Arizona (including some nonsmelter towns) for which air quality and
mortality data were available, a factor analysis of the data combined
with'a multiple regression analysis showed that the mortality rate for
acute respiratory diseases in females was strongly correlated with the
!
level of SO? and the age of the smelter. The same analysis showed that
the disease was not correlated with socioeconomic factors and that no
other disease was correlated with S02•
These findings, coupled with the well-established relationship
between SO2 and respiratory disease, suggest that SO2 emissions from
II-9
-------
the copper smelters cause excessive mortalities from acute respira-
tory diseases.* It appears that particulate emissions could also be a
key factor; however, there is no way to determine at this time the rela-
tive importance of SC>2 and particulates in contributing to the deaths,
since the two occur together for the sites examined (this is a feature
of several 50% air pollution epidemiological studies).
The chronic respiratory diseases of bronchitis, emphysema, and
asthma were not found to be associated with Sty and particulates.
This is an interesting finding that deserves some explanation, since
the etiology of respiratory diseases has demonstrated that SC>2 is an
etiologic agent for these chronic respiratory diseases. A possible
explanation is that individuals afflicted with chronic bronchitis,
emphysema, or asthma leave the copper smelting areas when the disease
develops to a serious level.
(3) Digestive Disease Findings
Digestive and respiratory diseases were the two strongest correla-
tions with the copper industry resulting from the original regression
analysis. However, the digestive disease correlation is of special
interest because it survived the top 10 county matched-control test and
the respiratory disease correlation did not.
Mortality rates in females for diseases of the liver, pancreas, and
gallbladder were found to be significantly elevated in the Arizona and
Montana smelter towns but not in Utah. No conclusive evidence could be
developed that linked the elevated mortalities to a particular pollutant
from the copper smelting industry. The most reasonable explanation is
that the most prevalent digestive disease (liver cirrhosis) is a result
of excessive consumption of alcohol in the copper smelting towns. The
excessive use of alcohol is a problem recognized by health officials in
Arizona. It is a common malady in rural communities whose residents
have relatively high and steady incomes and limited means of entertain-
ment. This relationship is an indirect socioeconomic relationship with
*There are no other large sources of SO2 in the region.
11-10
-------
the industry that doesn't involve an industrial pollutant. This hypoth-
esis is supported by the following:
(a) The health officials in Arizona smelter areas have observed
that there is excessive consumption of alcohol by smelter
workers and their families.
(b) The smelter counties in Arizona and Montana consume a
higher than average quantity of alcoholic beverages.
(c) In Utah, where digestive diseases were not significantly
elevated, the majority of residents are nondrinking Mormons.
Lead, cadmium, and arsenic are known to accumulate in and cause
damage to the liver. The levels of these heavy metals are elevated in
some smelter areas; however, the levels cannot be correlated with
digestive disease prevalence to any significant degree. Although a
direct effect or a synergistic effect with alcohol cannot be ruled out,
it appears that copper smelter pollutants are not a major factor in the
elevated digestive diseases found in many copper smelter communities.
(4) Findings for Other Diseases
Mortality rates in females for cerebrovascular disease and cancer
of the colon and pancreas were significantly elevated in Anaconda in
Montana but not in Arizona or Utah. Similarly, levels of cadmium and
arsenic were very high in Anaconda but not in the average Arizona smel-
ter town (data for Utah was not available). Considering the etiology
of the diseases, these findings could be explained by the following hy-
potheses:
- (a) The high levels of cadmium and arsenic (a suspect human
carcinogen) emitted from the smelter in Anaconda are a
possible cause of excess mortalities from cancer of the
•• • colon and pancreas.
(b) The high levels of cadmium are a possible cause of the
excess mortalities from cerebrovascular disease.
It is noted that both the disease et'ioldgies and the supporting data
are not sufficiently strong to provide convincing evidence of a possible
relationship between the diseases and the industry. These relationships
are highly speculative at this time, and considerable additional evi-
dence is needed to provide a credible causal hypothesis.
11-11
-------
3. Steel Manufacturing Case Study
a. Basis for Correlation
The multiple regression analysis found a fairly strong correlation
between an industrial grouping that contained steel manufacturing, and
cardiovascular diseases and digestive neoplasms in females. Close in-
spection of the data disclosed that steel manufacturing was the main
component of the industrial group. Upon screening the correlation using
a matched-control test for the top eight steel manufacturing counties,
the cardiovascular correlation disappeared in females. However, the
correlation with digestive neoplasms in females remained strong. Since
steel plants were not expected to employ females in hazardous areas, it
was assumed that the validated correlation between steel manufacturing
activity and digestive neoplasms in females was an indication of a com-
munity health problem and, consequently, this correlation was selected
for investigation.
b. Study Site Selection
Johnstown, Pennsylvania, was selected as the site of the initial
field investigation for the purpose of developing a causal hypothesis.
There were several reasons for selecting Johnstown:
• high industrial index for steel manufacturing;
• high county mortality rate for digestive neoplasms in
females;
• relatively large population at risk;
• excellent data available, including mortality data on
tape, morbidity data from a county registry, and fairly
complete air quality data;
• no other large industries that might confound the findings.
Prbvo, Utah, was selected as the hypothesis test site on the basis
of high industrial index, little industrial clutter, large population at
risk, and good availability of data including morbidity data on tape
from the cancer registry.
It is emphasized here that, in selecting sites for the steel in-
vestigation and, indeed, in selecting the sites for all the investi-
11-12
-------
gations, a prime consideration was the avoidance of the potentially
confounding factors associated with large cities such as multiple
sources of pollutants and extensive urbanization. Hence, large metro-
politan areas were not considered in the search for study sites. In the
case of steel, this meant that the large steel cities of Pittsburgh,
Buffalo, and Birmingham were not considered suitable sites for the pres-
ent investigation.
c. Industry Characterization
A full description of the general and site-specific characteristics
of the steel industry pertaining to possible community health problem
was prepared. In general, a steel plant is comprised of four major
processing units: coke ovens which produce coke for firing the furnaces,
sintering plants which process fugitive metal fines into reusable form,
blast furnaces which produce iron from ore, and steel-making furnaces
(open hearth, electric arc, or basic oxygen) which produce steel from
pig iron and scrap. Coke ovens emit particulates and organics including
benzofa]pyrene (BaP). Sintering plants emit hydrocarbons and particu-
lates.'. Blast furnaces and steel furnaces emit large quantities of par-
ticulates and oxides.
d. Etiological Basis
The organics, hydrocarbons, and particulates produced by the various
steel processes have a recognized potential for causing disease. Labora-
tory and occupational health research has associated a number of diseases
with the emissions from steel plants. Summarized below are the findings
related to etiological agents and associated diseases:
• Organics (includes BaP)— malignant neoplasms (including
digestive neoplasms); respiratory disease; nephritis;
gastroenteritis; liver disease.
.• Sulfur Dioxide—respiratory disease; digestive and
respiratory neoplasms; liver disease.
• ParticuJates—digestive and respiratory neoplasms;
respiratory diseases.
11-13
-------
• Lead—gastrointestinal and respiratory diseases; central
nervous 'system diseases.
• Cadmium—hypertension and other cardiovascular disease;
liver disease; nephritis.
As. is evident from this listing, there is a substantial etiological
basis for a cause-and-effect relationship between the steel industry and
(, • • ' ' •
digestive neoplasms. Organics, sulfur dioxide, and particulates pro-
duced by steel processes have been related to neoplasms of the digestive
organs.
e. Collection of Data
Industry Data—Data on the age, production history, and total emis-
sions of criteria pollutants were assembled for the steel plants under
study. An interesting method of obtaining and verifying information on
the growth of the steel industry in Johnstown was utilized. By analyz-
ing several sets of aerial photographs taken over the past 40 years, the
EPA Environmental Photographic Interpretation Complex (EPIC) in Warren-
ton, Virginia, was able to determine when new plants were built or
operations were expanded. Also, EPIC was able to verify that there were
no large point sources of pollution in the Johnstown Air Basin other
than the steel industry.
Air and Water Quality Data—Excellent air quality data for Johnstown
were obtained from the Pennsylvania Department of Environmental Re-
sources. Several years of data for criteria pollutants, trace metals,
BaP, and benzene-soluble organics (BSO) were available from monitoring
stations in the Johnstown Air Basin. The air quality data for Provo,
Utah, and vicinity were not as complete as we had originally been led
to believe. The principal data obtained was for total suspended par-
ticulates (TSP). For both Utah and Johnstown, some data were obtained.
from an EPA study which provided BaP and TSP measurements. It was
determined that communities in the vicinity of both Johnstown and Provo
obtained their potable water from sources too remote to be affected by
effluents or emissions from the steel plants.
11-14
-------
Meteorology Data—Air basin characteristics for Johnstown in Penn-
sylvania were obtained from Pennsylvania State University studies of the
area.
1 Health Data—For both the Johnstown and Provo sites, mortality data
for all diseases were obtained on magnetic tapes for 11 years, from 1965
through 1975. The number of mortality records obtained is given in
Table II-3. These data did not contain duration of residence or occupa-
tional information. Morbidity and mortality data from 1966 to 1975 were
also obtained for the Provo site from the Utah Cancer Registry. A prob-
lem with these data was that they provided incidence and prevalence by
county and not by town or census unit. We had planned to obtain and
analyze cancer registry information for Johnstown also, but the flood
of July 1977 prevented this. An interesting characteristic of the Provo
site is that the majority of the population is Mormon; thus, health
problems related to alcohol or smoking are less prevalent.
Table II-3. MORTALITY RECORDS ACQUIRED FOR STEEL INVESTIGATION
Johnstown Site
Provo Site
Number of Deaths from All Diseases for All Ages
Exposed Areas
WM
4,000
4,000
WF
3,300
3,300
Control Areas
WM
5,300
4,400
WF
3,700
3,300
WM = white males; WF = white females
Demographic Data—The fifth count, file C, tapes from the 1970 Census
were used for population and socioeconomic data. This file provides data
at the enumeration district level or the census tract level.
From the analysis of aerial photographs, EPIC determined key popu-
lation shifts over a 30-year period for the Johnstown area. Using
meteorological and topographic data, EPIC developed estimates of the
population at risk as a function of time for the Johnstown Air Basin.
11-15
-------
f. Analysis of Data
To develop a causal hypothesis at the Johnstown site, exposed and
control populations were selected on the basis of air transport and
I
socioeconomic considerations. The work of EPIC was used to assist in
the definition of the exposed population. A control population that
matched selected socioeconomic variables to within 10% was selected.
As in the case of the copper smelter case study, mortalities were
divided into 34 disease categories, and the age-adjusted mortality rates
of the exposed population were compared to those of the control popu-
lation. The significance of differences in mortality rates was computed
using the Mantel-Haenszel x2 test. Causal hypotheses were developed by
correlating differences in mortality rates between exposed and control
populations with differences in air quality measurements, and identi-
fying those correlations having sound etiological basis.
Although both male and female mortality rates were computed, to pre-
clude bias due to occupational exposure, the hypotheses developed were
based only on the data for females. Male mortality rates were examined
to provide verification where possible, and to see if unusual health
problems of possible occupational origin were indicated by the data.
At Provo, Utah, this process was repeated with special attention
given to the diseases and specific emissions identified from the Johns-
town analysis. The exposed population consisted of industrial towns or
<
municipalities, which were examined separately rather than by aggrega-
tion. Control populations were found that matched the selected socio-
economic variables to within 15% or less.
g. Summary of Findings for Steel Manufacturing Case Study
(1) At the Johnstown site, there are a number of diseases with sig-
nificantly elevated age-adjusted mortality rates in females. These are:
• Malignant neoplasms
— oral cancers (males & females)
—respiratory cancers (males & females)
— breast, urinary, genital cancers (males & females)
—central nervous system cancers (males & females)
—peritoneal and other digestive cancers (females only)
11-16
-------
• Cardiovascular disease (males s females)
• Respiratory disease (females only)
• Urinary disease (females only)
For all other diseases, there were no significant differences between
the exposed and control death rates in females.
(2) No conclusive evidence was obtained which linked a specific pol-
lutant to any of the elevated diseases. However, on the basis of eti-
ological considerations and of the measured high levels of pollutants
in the exposed areas compared to national standards, the following
hypotheses were indicated:
Disease Etiologic Agent
Malignant neoplasms (oral, respiratory, Organics, particulates
breast, CNS, peritoneal)
Cardiovascular disease Cadmium
Respiratory disease Organics, particulates,
SO2, lead
Urinary disease Organics, cadmium
The organics emissions (e.g., BaP) are the predominant etiologic agent
in these hypotheses taken as a whole. Hence, the primary causal hypoth-
esis developed is the relationship between organic emissions from the
coking processes and a variety of diseases in the community, principally
cancers. This is in agreement with the original correlation from the
multiple regression analysis which correlated digestive cancers with the
steel industry.
(3) At Johnstown, a number of additional diseases were found to be
significantly elevated in males only, indicating that they possibly were
of occupational origin. These diseases are: stomach cancers; leukemia;
blood diseases; hypertension; cerebrovascular diseases; bronchitis,
emphysema, asthma; liver diseases. No checks could be made to verify
that these diseases were occupationally derived; data and time did not
permit this.
(4) Only one of the causal hypotheses developed at Johnstown was
supported by the Provo data. Mortality rates for malignant neoplasms
of the breast, urinary, and genital organs and of the central nervous
11-17
-------
system were found to be significantly elevated at Provo as well as at
Johnstown. Other neoplasms and cardiovascular, respiratory, and urinary
diseases were not significantly elevated in Provo.
(5) It was determined that the pollutant concentrations of organics
and particulates at the Utah site are at least an order of magnitude
less than the concentrations at Johnstown. This is due to differences
in plant emission concentrations, proximity of the exposed population to
the pollutant sources, and meteorological and topographic characteris-
tics. The latter is a particularly strong reason for the difference:
the Johnstown site is in a narrow and well-defined air basin, which
traps the emissions. At Provo, the population is located in a very
wide valley where emissions are not confined. Another contributing
factor to the difference is the fact that the Johnstown Steel Works are
at least 50 years older than the Geneva Works; hence, the Johnstown
population has had a longer exposure to pollutants than has the Provo
population. These factors suggest that some of the causal hypotheses
identified in Johnstown may be valid, even though the death rates for
the diseases are not elevated in Provo. For example, respiratory can-
cers at Provo were found to be elevated in males but not in females.
The high'rates in males could be due to occupational exposure, and the
normal rates in females could indicate that the community exposure is
too low to induce a measureable effect.
(6) In both sites, the death rates from endocrine, metabolic, and
blood diseases were found to be elevated in males and not in females.
This may indicate an occupational problem possibly caused by exposure
to trace metals within the plant.
11-18
-------
4. Bituminous Coal Mining Case Study
a. Basis for Correlation
The strongest correlation produced by the multiple regression
analysis was between bituminous coal mining and cardiovascular and re-
spiratory diseases. The correlation was strong in both males and fe-
males, indicating a potential community health problem. The correlation
passed all of the screening tests, including the matched-control test
for the top 10 coal mining counties.
b. Study Site Selection
Wyoming County, West Virginia, was selected as the causal hypothesis
development site because of high industrial activity, lack of confound-
ing industry, high mortality rates for the diseases of interest, and
availability of morbidity data.
For the hypothesis test site, the coal mining area of southern
Illinois (counties of Saline, Williamson, Franklin, Randolph, and Perry)
was selected. This area was chosen principally because of high indus-
trial activity and the potential of relatively large exposed and control
populations. Also, there was no appreciable industrial clutter and
mortality data were readily available.
c. Industry Characterization
The three principal elements of coal mining operations are the mine,
the coal preparation plant (tipple), and the coal transportation system
(i.e., track conveyors, trucks, and railroad cars). In Wyoming County,
underground mining is predominant; in southern Illinois, both surface
and underground mining are prevalent. Coal preparation plants clean and
size the coal. Usually water is used in the cleaning process, and the
coal is dried using shakers, centrifuges, or thermal dryers. Thermal
drying, accomplished by heated air, is the predominant drying process
in Wyoming County and southern Illinois.
All coal operations produce pollutants to some degree. The three
principal pollutants are acid mine drainage, coal dust emissions, and
11-19
-------
coal combustion products. Mines and coal preparation plants discharge
water that is often highly acidic (acid mine drainage) and capable of
contaminating community water supplies with solubilized toxic minerals.
Coal dust is discharged by mines and coal preparation plants, princi-
pally thermal dryers. Also, coal handling, storage, and transportation
in and around coal preparation plants produce significant quantities of
fugitive coal dust emissions. Coal combustion products are produced
principally by thermal dryers which burn coal to generate the hot flue
gases for drying. Also, coal combustion gases are emitted from burning
piles of waste (gob piles) discharged by the coal preparation plants.
In recent years (since the late 1960's), the thermal dryers have
incorporated highly efficient control devices to reduce the particulate
emissions. However, combustion gases are not controlled and EPA be-
lieves that the amount of NOx and SOx emitted is too small to require
control. Although coal dust emissions from coal handling, storage, and
transportation could largely be controlled by enclosing the operations
or covering the coal with tarpaulins, there does not appear to be much
of this in practice.
d. Etiological Basis
The health hazards associated with coal mining occupations are well
recognized and documented. There are four major occupational respira-
tory diseases of coal miners that have been linked with coal dust and
related emissions:
• Coal Workers' Pneumoconiosis (simple CWP)—this results
from the accumulation of inhaled coal dust in discrete
macules in the lung and can be associated with some
changes in respiratory function; rarely if ever disabling.
• Progressive Massive Fibrosis (PMF, or "complicated" CWP)—
this is an advanced, uncommon form of CWP which results in
obstructive (emphysematous) and restrictive pulmonary
disease and may be life threatening.
• Chronic Bronchitis—this probably represents chronic
bronchial irritation from dust, but it has not been ex-
tensively studied.
• Silicosis—this occurs in workers who drill rock and sink
shafts through free silica (SiO2) deposits or who drive the
transport cars and can be life threatening in severity.
11-20
-------
With regard to cardiovascular diseases associated with coal mining,
an excess mortality from heart disease is a predictable result of a high
prevalence of lung disease. The pulmonary and cardiovascular systems
are intimately related, and there exist several mechanisms by which lung
dysfunction can result in heart failure. Other environmental factors
such as low water hardness have been associated with cardiovascular dis-
ease; however, mine drainage would tend to increase rather than decrease
water hardness.
e. Collection of Data
Industry Data—Data on location, age, production, and other charac-
teristics of the special mining operations at the sites were obtained
from Keystone Coal Buyers Manuals. Source emission data were obtained
for the counties from the National Emissions Data System (NEDS).
In Wyoming County, West Virginia, there are many underground mines
and several preparation plants, five of which have thermal dryers. In
southern Illinois, there are both surface and underground mines and ten
preparation plants, five of which have thermal dryers. Actual numbers
of mines and preparation plants vary considerably over time. Because
of the relatively small population that may be exposed to an individual
mine or preparation plant, it was necessary to examine a number of mines
or plants at each site in order to provide large enough exposed popula-
tions. In effect, this treated the industry as an area rather than as
a point source.
Air and Water Quality Data—No air quality data were available for
either site; air quality had not been monitored in any of the counties
by the states or counties. In the absence of air quality measurements,
dispersion models were used to estimate concentrations of pollutants in
communities. Water chemistry data for Wyoming County potable water
supplies were obtained from the County Health Department.
Meteorology/Hydrology Data — The two sites investigated are very
different in topography and weather. The Wyoming County site consists
of narrow valleys which confine the emissions and effluents to rela-
tively small regions. The southern Illinois area is flat in comparison
(rolling hills), enabling wide dispersal of the emissions and effluents.
11-21
-------
The needed meteorological data for analysis were obtained from the
closest weather stations in both areas. General hydrologic information
was obtained from state officials; site-specific hydrologic information
was obtained by on-site observations.
Health Data—Death certificates for respiratory and cardiovascular
diseases for Wyoming County in West Virginia for the years 1968-1972
were supplied by the West Virginia Vital Statistics Office; death cer-
tificates for all diseases for the years 1965-1975 were ordered but
never received. Mortality data for the Illinois site were obtained for
all diseases for the years 1974-1976.
Morbidity records for all illnesses were obtained from Wyoming
General Hospital in Wyoming County. These data were abstracted by local
medical record abstractors. Diagnostic and demographic data were ob-
tained for all diseases and the cardiovascular and respiratory cases
were also'staged for severity, using a form designed for this purpose.
The plan.had'been to obtain mortality records from two other hospitals
in the area which, together with Wyoming General, would have provided
an estimated 75% of the admissions from Wyoming County; however, ap-
provals could not be obtained in time to acquire these data. Conse-
quently, the morbidity records that were collected from Wyoming General
are of limited value due to admission selectivity.
Some qualitative observations related to health problems in Wyoming
County were offered by hospital staff members. They observed that coal
miners are heavy tobacco users, smoking cigarettes when they are out of
the mines and chewing tobacco in the mines where smoking is not allowed.
It was estimated that better than 90% of the miners use tobacco exten-
'sively, and this presently appears to be extended to their wives. This
indicates that tobacco use is a factor to be considered in any findings.
Demographic Data—Population and socioeconomic characteristics were
'Obtained from the fifth count, file C, 1970 Census tapes, which provide
data at the enumeration district level. Enumeration district maps were
acquired to enable the correlation of the street address and town of
residence to enumeration district.
11-22
-------
f. Analysis of Data
Correlating etiological findings with the mining-related emission
and effluent characteristics, several conceptual causal hypotheses were
derived for Wyoming County. These include:
• Respiratory/cardiovascular diseases caused by coal dust
emissions from mining operations, principally the opera-
tions associated with coal preparation plants utilizing
the thermal drying process.
• Respiratory/cardiovascular diseases caused by coal com-
bustion products from thermal dryers and burning waste
piles.
• Cardiovascular diseases caused by the toxic effects of
acid mine drainage (e.g., solubilized cadmium).
By analyzing the water quality and the transport of emissions,
assessments were made of the plausibility that Wyoming County communi-
ties would be exposed to hazardous levels of hypothesized etiologic
agents. Age-adjusted mortality rates for cardiovascular and respiratory
diseases were then computed for potential populations at risk and com-
pared to control populations. To assist in determining the part that
terrain and socioeconomics play in the relationship between industry
and disease, a regional analysis was made of the county-by-county mor-
tality and industry data that had been used as the data base for the
initial regression analysis.
The hypothesis developed by integrating the results of the emission
transport, water quality, and mortality analyses was then tested by two
separate investigations: first, by the analysis of morbidity data
(hospital patient records) in Wyoming County; and second, by a mortality
analysis at the southern Illinois site. In both cases, disease patterns
were examined to determine if they supported the hypothesis. At the
Illinois site, an emission transport analysis was performed to estimate
the level of exposure to the population at risk.
11-23
-------
g. Summary of Findings for Bituminous Coal Mining Case Study
(1) In Wyoming County and southern Illinois, the respiratory and
cardiovascular mortality rates for communities in the proximity of
thermal dryers exceeded the rates for distant populations. The amount
of excess was statistically significant for all cases except for respi-
ratory diseases in Illinois, where the rate ratio of 1.5 was not statis-
tically significant.
(2) Based on the water quality analysis, hazardous levels of toxic
minerals were not found in Wyoming County water systems. The evidence
indicates that acid mine drainage is not a health problem in Wyoming
County.
(3) The analysis of Wyoming General Hospital records in Wyoming
County disclosed no unusual morbidity patterns. This result is not par-
ticularly significant, since only approximately one-fourth of the mor-
bidities in Wyoming County were analyzed and the hospital selectivity
bias for those examined is unknown.
(4) The results of the regional mortality analysis show that, when
the mortality rates for respiratory and cardiovascular diseases are
compared for coal-mining and non-coal-mining counties in similar terrain,
the rate ratios are higher in mountainous terrain than in flat terrain,
indicating that terrain is a factor in the relationship between coal
mining and respiratory/cardiovascular diseases.
(5) Although there is no conclusive evidence of the existence of
hazardous levels of coal dust in the coal mining communities, the re-
sults of the investigation point towards the hypothesis that coal dust
from operations in and around the thermal dryer coal preparation plant
is the cause of the elevated respiratory and cardiovascular disease.
The etiology is sound and the results of the site-specific mortality
analyses show a correlation between the proximity to thermal dryer
coal preparation plants arid mortality rates. The results of the re-
gional mortality study can be explained by the fact that the narrow
valleys in mountainous terrain tend to confine the coal dust to the
11-24
-------
populated areas. The results of the morbidity study do not support the
coal dust hypothesis; however, as previously mentioned, these results
are discounted due to the unknown hospital selectivity bias.
. (6) Some consideration was given to the socioeconomic hypothesis
that heavy smoking by coal miners and their families and the use of coal
for home heating are the causes of excess respiratory and cardiovascular
diseases. Although these hypotheses are not supported by the evidence
produced, the contrary evidence is not sufficiently strong to rule them
out.
11-25
-------
5. Viscose '< Ha yon Case* Study
a.- Basis for Correlation
The basis for the correlation between the viscose rayon industry
and community disease, principally coronary heart disease, differs from
the other case studies. The viscose rayon correlation is not a result
of the multiple regression analysis, as were the others; rather, it is
a result of community complaints regarding an odor problem associated
with a particular plant. Upon investigation, it was ascertained that
potential toxicants (hydrogen sulfide and carbon disulfide) are emitted
in substantial amounts from the industry and that there is documented
evidence that these toxicants may be harmful to viscose rayon workers.
A comparison of the mortality rates between counties with viscose rayon
plants and control counties showed that coronary heart disease was sig-
nificantly elevated in some of the exposed counties.
jb. Study Site Selection
•Two sites were selected for field investigation on the basis of the
history and characteristics of the viscose rayon plant, the size and
proximity of the population at risk, the absence of other industry and
pollution sources, the availability of emission and air quality data for
the-pollutants of interest, and the availability of health data. The
two sites were Elizabethton, Tennessee, and Lewistown, Pennsylvania.
c. Etiological Basis
A search of relevant medical and occupational health literature dis-
closed that there is considerable evidence that elevated coronary heart
disease and death due to cardiovascular disease in viscose rayon workers
are associated with exposure to carbon disulfide (CS2) and hydrogen
sulfide (H2S) in workplace air. A study of rayon manufacturing processes
and plants disclosed that relatively large quantities of €82 and H-.S are
emitted to the atmosphere during the manufacturing process. Thus, the
potential exists for chronic exposure of nearby populations to these
etiologic agents.
11-26
-------
d. Collection of Data
Various industry, environmental, and health data were collected for
the two sites. These data consisted of:
• Plant process characteristics and production history
• Emission estimates and point source data
• Meteorological data
• Air and water quality data
• Demographic data for exposed and control populations
— fifth count, file C, 1970 Census
• Mortality data for all diseases
—Elizabethton, Tennessee; 1960-1975
(total deaths, 2,190; deaths 35-74, 1,203 records)
—Lewistown, Pennsylvania; 1965-1975
(total deaths, 2,992; deaths 35-74, 1,546 records)
e. Analysis of Data
A transport analysis was performed to estimate populations exposed
to various levels of CS2 and F^S. Two types of controls were used to
test the hypothesis at each site. One set of controls consisted of the
county residual which consisted of the population of the county living
outside the city that contained the rayon plant and considered not to
be at potential risk. The other set of controls consisted of an aggre-
gate of towns which were demographically and socioeconomically similar
to the study site. Age-adjusted mortality rates for 34 disease cate-
gories were computed for males and females and compared for exposed and
control populations. Differences in mortality rates were tested for
significance by means of the Mantel-Haenszel x test.
£. Summary of Findings for Viscose Rayon Case Study
(I) No di.-ath rntes (for any of the 34 disease categories) were found
to be significantly cJcvated in the exposed populations at either Eliza-
bethton or Luwistown. That is, a significant community health effect
associated with viscose rayon plants could not be found from the analy-
sis of the mortality data.
11-27
-------
(2) Ambient CS2 concentrations for the exposed populations were es-
timated to be in the range of 25 to 100 ppb. The plants in both sites
have been1 operating since before 1930. These findings indicate that
chronic exposure of populations to these ambient concentrations of CS2
will not induce premature death.
11-28
-------
B. DISCUSSION OF RESULTS
1. Validity of Causal Hypotheses
In three of the four industry-disease investigations, plausible
causal hypotheses were derived from the site-specific data and analysis.
The principal relationships are summarized in Table II-4.
This section discusses the validity of these hypotheses in terms of
five specific measures of merit:
coherency of explanation
strength of association
specificity of association
time sequence of variables
consistency of association
These five measures of merit are summarized for each of the relation-
ships in Table II-5 and are discussed in the paragraphs below.
a. Respiratory Diseases Associated with Copper Smelting
This is the most well-founded relationship developed during the
study. It scores satisfactory marks on all five measures of merit.
Particular features are the strength of the relationship borne out by
both a matched control and a dose-response type test, the consistency
of the association in several study sites, the temporal relationship
with the age of the smelter, and the lack of a correlation of SO2 with
other diseases and socioeconomic factors. The key weaknesses are the
lack of verification of the duration of exposure for the decedents and
the deficiency in strength due to the small exposed populations and the
small number of observations. Additional work that is needed would con-
sist of the following:
• Verification of the exposure of decedents by establishing
the duration of residence from state and local records.
.• Collection and analysis of hospital records to further
strengthen the association and investigate the hypothesis
that the incidence of obstructive pulmonary diseases is
also elevated but does not result in mortalities because
those with the disease move away from the smelters and
11-29
-------
Table II-4. PRINCIPAL CAUSAL HYPOTHESES
u>
O
INDUSTRY
Copper
Smelting
Steel
Manufacturing
Bituminous
Coal Mining
DISEASE
Acute Respiratory Diseases
(pneumonia, influenza,
pleurisy, etc.)
1
Digestive Diseases
(primarily cirrhosis
of the liver)
2
Malignant Neoplasms
3
Respiratory/Cardiovascular
Diseases
2 emitted from smelters
Particulates may be con-
tributing agent
Alcohol principally
Cadmium emitted from smelters
a possible contributing
agent
Organic emissions (e.g., BaP)
from coking and smelting
plants
Particulates from various
processes also a contribut-
ing agent
Coal dust emitted from coal
mining operations, princi-
pally those in and around
thermal dryer coal prepara-
tion plants
REMARKS
Chronic inhalation of SC>2 and
particulates damages pulmonary
system which decreases resis-
tance to acute respiratory dis-
eases
Excessive consumption of alcohol
in copper smelting community
causes high prevalence of liver
cirrhosis mortalities
Inhalation and ingestion of
organics and particulates
induce malignant neoplasms in
various organs
Inhalation of coal dust leads to
pulmonary dysfunction which
leads to cardiac involvement
-------
Table II-5. SUMMARY EVALUATION OF PRINCIPAL DERIVED CAUSAL HYPOTHESES
ASSOCIATION
Acute diseases
of respiratory
•ysteej associ-
ated with
copper smelt-
ing through
St>7 ft parti-
culate emis-
sions .
Digestive
disease*.
primarily
alcoholic
cirrhosis of
the liver.
associated
with copper
smelting
Cancers of
oral, r«-
geiiitouruidry
*yt»temA Asso-
ciated with
»tfcel maiiu-
f actur ing
through
organic
emisittona fium
diking plants.
Cardiovascular
ft respiratory
diseases asso-
ciated with
bituminous
coal mining.
through coal
du*t •missions
from thtrrsul
Jry«r coal
preparation
COHERENCY
UP EXPLANATION
Wvll-established etiology
between SO] t respiratory
diseases. Reasonable ex-
planation can be given
for why acute diseases
were found to be associ-
ated ft cnconic respira-
tory diseases were not.
HD coherent explanation
why copper smelting emis-
sions or effluents would
induce high digestive
disease mortality rates.
particularly foe liver
cirrhosis. Evidence
that cadmlue emitted by
•melters could be a con-
tributing cause i bow-
evex , the saost logical
explanation, supported by
•Icohol consumption data
« health observations, Is
the abuse of aloohol in
copper smelting areas.
Well-ektabliUM*! Btiulouy
relating malignant ueo~
by coking «nd ulntttr lug
plants.
Uood Coherent explanation
etiology connecting coal
dust to pulmonary dis-
uafces.
STRiNuTH pr ^ASSOCIATION
e .Matched control test showed
statistically strong corre-
lation between disease ft
^ndustry.
e Dose- r esponse test showed
strong correlation between
SOj concentrations, age of
industry ft disease.
e Strong correlation between
disease ft Industry, based
OM matched -control test.
except In Utah.
e Recognised (but not veri-
fied} correlation between
alcohol con sumption ft
copper industry, except in
Utah.
* Strong correlation between
diseases ft Industry based
Jonnstown.
e Measurement data snuw ex-
posed population in Johns-
town exposed to levels of
organic* much higher than
nation*! lavultt.
e Littl* correlation between
aiseascts ft industry at
Provo, utati.
e HeatturttSMnt data show ur-
uauiu concentration* at
Pcovo are much lower than
e Correlation between die-
tsaoe ptevaleucn ft induutiy
relatively weak becaun* of
small wxposMd population.
e no verificatiou that tlw
populations *r« expustad to
bazarduuB Itivtflw of coal
dust.
, SPECIFICITY Or ASSOCIATION
e Specific disease ft specific
etlologlc agent have been
identified.
e Other diseases not corre-
lated with agent.
e Rlevated mortality rates
found In nunoccupatlonal
portion of population
(females) indicating com-
munity disease.
e Soc Loeconomic factors not
correlated with mortality
rates.
e Specific disease ft specific
etiology Identified (alco-
holic cirrhosis of liver).
e Disease prevalence occur-
ring In females. I.e.. non-
occupatlonal association
with copper smelters.
e Rather broad set of dis-
eases associated with a
broad category of emissions.
• Disease prevalence occur.-
ring In females, i.e., non-
occupational association
with steel industry.
e Diseases other than asso-
ciated diseases ate closely
matched between exposed 4
control areas.
e Specific disease ft specific
etiologic aueitt identified.
e Specificity is weak because
other diseases went not
examined.
TIME SEQUENCE ]
or VARIABLES
e esxlmated period of
exposure consistent
with mortality data.
e Emission measurement
data consistent, with
exposure period.
e Mortality rates
shown to be corre-
lated with age of
smRlter* •
e could not determine
period of exposure
of individuals wtto
died of disease.
e Disease prevalence
data consistent with
age of pleat.
e Ho specific data or
analysis showing
teaxpQial trends ft
time sequences.
e Estimated period of
exposure consistent
with disease preva-
lence data.
e Emission measure-
ment data consis-
tent, with exposure
period.
e Mo specific date or
analysis showing
temporal trends ft
time sequences.
e ttstimated period of
exposure consistent
with mortality date.
e Mo specific date or
analysis •howing
t Major • I trends ft
time, sequences.
CONSISTENCY
Of ASSOCIATION
• Association vali-
dated by both
mstched -control
and dose -response
type of test.
e Association con-
sistent betwees.
copper smelting
sites.
e Association con-
sistent between
sites.
e Association Dot
strongly veri-
fied at eecond
sitei although
consistency has
not been demon-
strated, the
lack oc associ-
ation at second
site is not
necessarily in-
consistent ft can
be explained by
tne low concen-
trations ob-
served.
e Association con-
sistent foe two
sites but rela-
tively w*ax In
both cases.
-------
die elsewhere. In addition, the health records would
provide data on smoking history, which could be used to
check the effect that smoking has on the causal rela-
tionship.
• Investigation of additional copper smelting sites to
further strengthen the relationship. A particularly
valuable site would be Tacoma, Washington, where the
smelter is located within a city and there are large
populations at risk. Analysis of mortality data at the
census tract level would permit a dose-response type of
analysis as well as a matched-control analysis.
b. Digestive Diseases Associated with Copper Smelters
This is a less well-founded relationship between the disease (liver
cirrhosis) and a socioeconomic factor (alcohol consumption) associated
with the industry. The association is founded principally on the
correlation between the prevalence of liver cirrhosis and the presence
of the industry. A specific agent (alcohol) has been identified, and
some evidence has been acquired which associates excess alcohol consump-
tion with copper smelter towns. However, further work is needed to pro-
vide more convincing evidence, namely:
• Collection and analysis of alcohol consumption data for
copper smelting and control towns.
• Analysis of hospital records to provide verification of
the incidence of alcoholic cirrhosis.
• Investigation of additional sites to strengthen the
association (see previous remarks on Tacoma, Washington).
• Determination of body burden of cadmium in digestive
disease morbidities or mortalities if possible.
11-32
-------
c. Malignant Neoplasms Associated with Steel Plants
This appears to be a fairly sound relationship based on etiological
considerations and the strength of association found in Johnstown, where
the concentrations of the etiologic agents were very high. However, the
relationship needs confirmation in other sites, since the investigation
in Provo, Utah, did not support it strongly, owing perhaps to much lower
concentrations of toxicants. Thus, additional work is needed to test
and validate the relationship:
• Investigation of additional sites, with and without coke
plants to verify that they are the primary contributor to
neoplasms. The large steel plants in the large cities
would be good sites because of high concentrations and
good population characteristics.
• Verification of exposure of deceased individuals by es-
tablishing duration of residence from state and local
records.
• Collection and analysis of hospital and cancer registry
records to obtain and analyze smoking histories of
patients.
d. Respiratory and Cardiovascular Diseases
Associated with Bituminous Coal Mining
Although the investigation provided a good coherent explanation of
the association, the strength of the association is not satisfactory due
to small exposed populations and a lack of measurements to verify the
existence of hazardous concentrations of coal dust in the community.
Further evidence is needed to validate the relationship, namely:
• Air quality measurements in exposed and control areas to
verify differences in coal dust concentrations.
• Collection and analysis of mortality data for all diseases,
and for as many years as possible, to investigate the
specificity of the association.
11-33
-------
• Collection and analysis of hospital records for all hos-
pitals in region.
• Investigation of additional sites.
e. Summary of Causal Hypothesis Evaluation
The four case studies have developed and tested a number of hypoth-
eses with the following results:
• One well-founded and verified hypothesis was developed
(respiratory diseases-copper smelting).
• One reasonably well-founded hypothesis was developed
which needs verification (malignant neoplasms-steel
plants [coking process]).
• Two weak hypotheses were developed which need consider-
ably more data and analysis to provide satisfactory
strength and confirmation (digestive diseases-copper
smelting; respiratory/cardiovascular diseases-bituminous
coal mining).
• One negative investigation (cardiovascular disease-
viscose rayon plants) which found no site-specific evi-
dence of a relationship between community disease and the
industry.
2. Validity of Regression Analysis
What can be said about the validity of the multiple regression
approach in light of the results of the four case studies? Table II-6
is a comparison of the results of the original multiple regression
analysis with those of the case studies. For all the correlations
identified by the regression analysis, plausible causal hypotheses were
established from the analysis of site-specific data. For the one cor-
relation that was identified from considerations other than the regres-
sion analysis, no causal relationship could be developed from the field
data.
On the face of it, these results tend to give considerable credence
to the multiple regression analysis of county mortality and industry
11-34
-------
Table 11-6. COMPARISON OF RESULTS OF MULTIPLE REGRESSION ANALYSIS TO CASE STUDY RESULTS
Correlations
identified
by
Multiple
Regression
Analysis
Correlation
identified
by other means
ORIGINAL CORRELATIONS
Industry
Copper Mining
{copper smelting)
Steel Manufacturing
Bituminous Coal
Mining
Viscose Rayon
Associated Diseases
Respiratory Diseases*
Digestive Diseases
Digestive Neoplasms
Cardiovascular Diseases*
Respiratory Diseases
Cardiovascular Diseases
Cardiovascular Diseases
RESULTS OF CASE STUDIES
Was Causal
Hypothesis
Developed?
Yes
Yes
Yes
Yes
Yes
Yes
No
Strength of
Evidence
Good
Fair
Fair
Weak
Weak
Weak
Good
These diseases did not appear to be correlated on the basis of top 10 county matched-
control test, however they were correlated in the original regression analysis.
-------
data as a means for identifying sources of community disease. However,
this conclusion must be tempered with consideration of the limitations
of both the case studies and the regression analysis.
In the first place, the results of the case studies are by no means
conclusive. Although evidence was found in all multiple regression
cases of plausible causal relationships supported by coherent explana-
tions, in general the evidence is quite weak. Further investigation is
needed to provide conclusive evidence in all cases.
Secondly, the negative result for the non-regression correlation
may not be particularly significant. It shows that our choice of a
potential community health problem was not a good one. Further, it
shows that our validation methods based on a matched-control test for
the top 10 counties are not particularly effective. As far as the
multiple regression analysis is concerned, the only significance that
can be attached to the negative result for viscose rayon is that the
regression analysis appears to be more effective than our method in
identifying' real industry-disease associations.
We should also examine the results in the light of what we would
expect from the multiple regression analysis. The basic concept, upon
which the multiple regression analysis of county data is founded, is
that, for communities exposed to disease-causing substances from indus-
try, there exists a relationship between the mortality rate in the com-
munity and the chronic dose received. Figure Il-la depicts this rela-
tionship. If the dose that the population received and the mortality
rates could be observed in many communities, then a regression analysis
of these data would show a strong correlation between the industry and
the disease, because the variance about the regression line (the mor-
tality rate-dose relationship) is much less than the total variance in
mortality rates giving no consideration to the dose received.
However, the data on community mortality rates and dosage do not
exist; instead, we must use county mortality rates as an estimate of
community rates, and industrial index as a measure of dose. To see
what this will do to the correlation, let us first examine the effect
of county mortality rates.
11-36
-------
Flgura
Mortality
Rate in
Exposed
Community
/ Excess Mortality Rate
/ ^-Contributed by Industry
' ^-Background Mortality Rate
~>
Dosage to Exposed Community
Figure ri-lb
County
Mortality
Rate
Figure II-lc
Dosage to Exposed Community
Industrial Index
Figure II-l
Correlation Between Mortality Rate, Dosage, and Industrial Index
11-37
-------
When county mortality rates arc used instead of those of the exposed
community, the increases in mortality rate due to the community exposure
will be diluted by being included with the entire population of the
county. This has the effect of decreasing the slope of the relation-
ships between mortality rate and dose. The amount of the decrease will
depend on the ratio of the exposed population to the county population.
Since this will vary from county to county, this introduces an uncer-
tainty as shown in Figure Il-lb. Here, the variance about the regres-
sion line is large compared to the total variance, and therefore the
correlation is much weaker than in the ideal case of Figure Il-la.
The surrogate measure for dose is the industrial index.* Dose is a
product of time of exposure and concentration, where concentration is
a function of source strength, transport phenomena, and distance of
exposed community from industry. The only one of these factors that is
related to industrial index is the source strength; hence county-to-
county variations in time of exposure (age of plant and length of resi-
dence) , transport phenomena, and distance from the plant will introduce
equivalent uncertainties in the relationship between industrial index
and dose. Furthermore, the relationship between industrial index and
source strength is both weak and uncertain. Variations in type of
plant, processes, plant management, and controls introduce uncertain-
ties. The relationship is also diluted in the same way as mortality
rates due to the fact that the industrial index is a measure of the in-
dustrial activity in a four-digit SIC code, which often is a broad in-
dustrial category containing several different types of industries in
addition to the type which is producing the harmful substance.
Figure II-lc illustrates how the relationship will be affected by
using industrial index as a measure of dose. Here, the variance about
the regression line is nearly as large as the total variance; hence, a
very weak correlation would result from the regression analysis.
From these considerations, we would expect the correlation coeffi-
cients between county mortality rates and industrial indices to be
quite small, and indeed they were. However, this alone would not pre-
clude an effective multiple regression analysis, because the analysis
^Industrial index as defined by SSI = (number of employees)2/county work
force. .
-------
of a large number of observations (3,000 counties) could easily pull out
the correlation even though there is a great deal of dilution and uncer-
tainty.
. The problem with weak correlations is that they are easily con-
founded by socioeconomic relationships between industries and between
diseases. These socioeconomic or geographical relationships can pro-
duce stronger correlations than true industry-disease correlations.
Thus, the problem becomes one of differentiating between a few true
correlations and many confounding correlations.
Looking at it in this light, we can define the characteristics of
industries that would be expected to produce high industry-disease
correlations (if a real effect exists) and, therefore, would be more
likely to be differentiated from socioeconomic confounding. Industries
that are'the sole or primary element of an SIC category will minimize
the effect of industrial dilution and therefore provide stronger cor-
relations. Likewise, industries that are generally located in rural
counties—where the exposed population is a large fraction of the county
population—would minimize the effect of county population dilution and
maximize the correlation coefficient. Finally, industries that are very
intense and concentrated and tend to group together in large towns or
cities will produce a large exposed population which will reduce the
dilution problem and provide stronger correlations.
It is extremely interesting to observe that the industries that were
selected from the regression analysis for study possessed characteristics
which provided the best chance for obtaining large correlation coeffi-
cients. All three industries (copper mining, steel manufacturing, and
bituminous coal mining) were the sole or principal component of their
4-digit SIC categories. Copper mining and bituminous coal mining were
rural industries which minimized the county dilution problem. Steel
manufacturing is a highly concentrated industry which provides large
exposed populations and therefore reduces the county dilution.
Thus, if true disease associations existed for the copper, steel
and coal mining industries, it is entirely reasonable that the multiple
regression analysis would identify them. Thus, the fact that the field
11-39
-------
investigations found some evidence of causal relationships for these
industries is compatible with what we would expect, and supports the ..
validity and worth of the multiple regression analysis for limited types
and characteristics of industries,
We conclude this discussion with the observation that screening and
validation of the results is a very important part of the multiple re-
gression analysis. That is, the true associations do not stand out by
themselves; it is through screening and statistical validation tests
that the valid associations are identified. For example, to identify
the copper, steel, and bituminous correlations, we screened almost 800
correlations and statistically validated 39.
3: Evaluation of Case Study Methodology
One of the objectives of the study is to develop methods for in-
vestigating relationships between industries and diseases. Specific
methodological findings and problems that resulted from the four case
studies are discussed here.
a. Size of Exposed Population
For two of the case studies, copper smelting and bituminous coal
mining, the investigation was hampered by the small size of the exposed
populations. For the most part, the small populations were a charac-
teristic of the industry. Both copper smelters and coal preparation
plants are generally located in rural areas where only relatively small
communities are highly exposed.
The problem with small communities is that there are too few deaths
to provide a statistically significant comparison between exposed and
control areas. The problem is particularly acute for diseases with low
mortality rates. Figure II-2 shows the relationship between exposed
population, mortality rate, and relative risk. For a given effect
(relative risk) in the exposed population, the figure graphs the age-
and sex-specific population required to show a statistically significant
difference between the exposed and control populations as a function of
the mortality rate. For a mortality rate of 10, a relative risk of 2,
11-40
-------
100,000
• Exposed and control populations
of equal size
• Mortality rates based on aggregation
of three years of records
• Significant increase defined'as
two standard deviations
20
40 100 200 400
Mortality Rate in Control Population
1,000
Figure II-2
Magnitude of Exposed Population Necessary
To Show a Statistically Significant Increase in Mortality
for a Given Time and Relative Risk
11-41
-------
and 3 years aggregation of death records, an age- and sex-specific popu-
lation of 40,000 is required to show significance. This required popu-
lation is reduced to 7,500 if the relative risk is as high as 4.
When restricted to small populations—as in the case of bituminous
coal mining—there are only two possible solutions to the problem:
(1) a.ggregate a large number of exposed populations or (2) aggregate a
large number of years of data (Figure II-3 shows the effect of the num-
ber of years of data). In both-cases, certain effects are lost in the
aggregation. In the former, it is the effect of differences in site
characteristics and plant processes, and in the latter it is the tem-
poral effects. In -both the copper smelting and the bituminous coal
mining investigations, both .types of aggregation were used since the
age- and sex-specific populations ,of individually exposed areas were of
.the order of a thousand or so.
In the case of steel manufacturing and viscose rayon, sizable popu-
lations were available that were relatively clear of industrial clutter.
E,or steel, large cities could,have been studied (e.g., Pittsburgh, Bir-
mingham); however, the decision was made early in the study not to se-
lect large cities because of urban pollution and socioeconomic problems.
Now that we have experienced the difficulty of obtaining health
records over many years (discussed below) and have seen the value of
being able to show effects of stratifications, we feel that larger
cities with large exposed populations should be investigated. This is
especially true for intense, highly concentrated industries such as
steel manufacturing. Although multiple sources of pollution are a prob-
lem with large cities, examination of health records at the census-tract
level should be sufficiently fineTgrained to disclose disease patterns
associated with the industry. Further, the large exposed populations
will enable a significant result to be obtained as a function of time,
thus showing temporal trends with industrial activity and control
.measures.
II-T42
-------
! 10.0,000
I
4J
04
•8
W
8,
I
•1-1
g.
40,000
20,000 -
10,000
4,000 —
2,000 —
1,000
400 —
• Exposed & control populations
of equal size
• Relative risk = 2
• Exposed mortality rate exceeds
control mortality rate by two
20 40 100 200 400
Mortality Rate in Control Population
1,000
Figure II-3
Effect on Exposed Population Requirements
of Aggregating Mortality Records Over Time
11-43
-------
£. Health Data
Magnetic tapes of death records provide easy, low-cost access to
many records; however, we found several problems with them:
• Limited number of years of data on tape (most states
started computerizing their records during the last few
years, so prior records do not exist on tape).
• Lack of certain information on tape such as occupation or
specific address (the latter is a particularly serious
problem in urban areas, since the tape generally provides
city or town of residence and not census tract or enumera-
tion district).
• They can contain a significant number of errors, and there
is no way to check data without going back to death cer-
tificates. This had to be done for Johnstown, where
errors in town of residence were discovered.
Thus, until states put place of residence such as zip code on tape,
hard copies of the death certificates are needed for epidemiological
investigations in urban areas. For rural areas, magnetic tapes can be
used; however, death certificates are needed for occupational informa-
tion and for providing data prior to the years for which tape data are
available.
It would also be extremely useful for epidemiologic studies if dura-
tion of residence were included as well as place of residence. This has
been done only recently in a few states. Acquisition of residence in-
formation by other means is very time-consuming, requiring on-site in-
spection of local records such as voter registration, property tax, etc.
We did not find central data banks of morbidity records to be useful
in the sites investigated. The cancer registries in Utah did not pro-
vide addresses by town or zip code (except for the last three years).
Likewise, PAS data for the sites investigated lacked place of residence.
11-44
-------
As for health records from hospitals, we found some hospitals eager
to help us and some that wanted no part of the study. Hospitals that
were associated with the industry being investigated offered to partici-
pate if they could be convinced that a health problem existed.
A final note about health data collection—we found, in communicat-
ing with vital statistics departments and health registries, that it was
a difficult matter to find out precisely what data were available and in
what form. Statements would be made about the number of years of infor-
mation available that would turn out to be incorrect. In several cases,
the precise facts about the data were not known until an official re-
quest was made for the data. The turnaround time required to fill a re-
quest for data was rather variable, ranging from a few weeks to several
months.
c. Control Areas
Two problems with the matched-control methodology occurred through-
out the study:
• Difficulty in finding suitable control populations in
i
the rural areas.
• Lack of air quality measurements for control areas.
The first problem is due to the fact that the industry under investiga-
tion affects the socioeconomic characteristics of the area significantly,
particularly in rural areas. Hence, towns in the region that do not
have a similar industrial base are not similar socioeconomically. In
the more urbanized areas, finding suitable control areas is not as dif-
ficult.
The second problem constitutes a serious difficulty with the matched-
control methodology. EPA and the states are effective at taking air
quality measurements in areas containing polluting industries; however,
in most areas without pollution sources, few if any measurements are
being taken. Thus, it was not possible in most cases to verify that the
air quality in the control areas was better than that in the exposed
area.
11-45
-------
In the present study, this problem could not be solved by taking
measurements during the course of the study. In most cases, the
industry had' incorporated control devices in the past few years; we
needed to compare the air quality prior to emission control, since we
were studying populations that were exposed prior to emission control.
However, in future studies of causal relationships by EPA, the focus
should be on the health problems that the industry is causing currently
or in the past several years since controls were adopted. Hence, the
problem with the lack of air quality measurements in control areas can
be taken care of by going out and measuring the air quality.
d'. Exposure Confirmation
It had been planned to confirm the exposure of decedents by deter-
mining the length of residence from a check of state and local records.
This was attempted in Wyoming County with tax records and found to be a
time-consuming and expensive procedure. For lack of time and resources,
it was not,carried through. In Arizona, the length of residence in the
state was indicated on the death certificate; however, the populations
were too small to effectively analyze the mortalities by length of
residence. Furthermore, the length of residence in the state was not
a good indicator of the exposure (i.e., length of residence in exposed
towns).
>
Any further validation of the industry-disease relationships should
confirm and analyze the length of residence. This information would be
particularly useful where the exposed populations are sufficiently large
to permit examination of temporal effects. To reduce the cost of this
effort, a 10% random sample would provide a sufficiently strong indica-
tion of the exposure of the total decedents.
e. Missing, Incomplete, and Incorrect Data
A troublesome problem with most sites was missing and incomplete
data. This was particularly true for water and air quality data which
was often spotty. In some cases, the measurements were obviously incor-
rect. This points out the importance of the site selection effort.
Extra effort should be spent in checking the accuracy and completeness
of the data available prior to selecting the site.
11-46
-------
f. Smoking Histories
Other than qualitative information, no data on smoking habits of the
mortalities and morbidities were obtained. The matched-control method-
ology should in general provide a reasonable match of smoking habits,
unless the presence of the industry influences the habit—as might be
the case with bituminous coal mining. Cigarette consumption data can be
developed by county, but probably not by community. Although these data
were not developed in this study, it would be useful in further work to
assist in the control matching. To fully pursue the influence of smok-
ing on the causal relationships, smoking histories of the decedents in
both the control and the exposed populations would be required. This
is an extensive effort that should be performed only if maximum "proof"
of a relationship is required.
11-47
-------
Chapter III
CONCLUSIONS AND RECOMMENDATIONS
A. CONCLUSIONS AND RECOMMENDATIONS PERTAINING
TO INDUSTRY-DISEASE RELATIONSHIPS
(I) In three of the four cases studied, substantial evidence was
f.ound which indicates that pollutants from the industry in question
could be the source of disease in the community. In no case was the
evidence sufficient to be conclusive and, in general, represented only
a small fraction of the epidemiological evidence needed to establish and
confirm cause-and-effeet relationships. This was to be expected con-
sidering the exploratory nature of the work, principally concerned with
investigating methodologies. With these reservations in mind, the
following conclusions and observations regarding industry-disease
relationships are made:
(a) SC>2 and particulate emissions from copper smelters may
be instrumental in causing excessive mortalities from
respiratory diseases in the exposed communities.
(b) Abuse of alcohol is the probable cause of excessive
digestive diseases in copper smelter communities.
(c) Organic emissions from coking and sintering operations
associated with steel manufacturing are the likely cause
of excessive mortalities from cancers of the oral,
respiratory, digestive, and genitourinary systems in
communities exposed to high concentrations of steel
plant emissions.
(d) Coal dust emissions from thermal-dryer preparation
• plants and coal transport vehicles are a possible cause
of elevated respiratory and cardiovascular diseases in
bituminous coal mining communities.
(e) Emissions or effluents from viscose rayon plants do not
appear to cause a significant community health problem.
III-l
-------
(2) Further study is required to provide conclusive evidence of the
industry-disease relationships identified for the copper smelting, stteel,
and coal mining industries (items (a), (c), and (d) above). The follow-
ing studies are recommended:
(a) A continuation of the investigation of the association
between copper smelting and respiratory/digestive
diseases, to include:
• field investigation of additional sites, in
particular, Tacoma, Washington;
• collection and analysis of morbidity records;
• verification of the exposure of mortalities/
morbidities by determining length of residence;
• collection and analysis of alcohol consumption
. data.
(b) A continuation of the investigation of the association
between the steel industry and various diseases, to
include:
• investigation of additional sites, in par-
ticular, Pittsburgh, Buffalo, Detroit, and
Birmingham;
• verification of exposure by determining
length of residence;
• collection and analysis of hospital and
cancer registry records.
(c) A continuation of the investigation of the association
between the bituminous coal mining industry and respira-
tory/cardiovascular diseases, to include:
• investigation of additional sites;
• collection and analysis of additional mortality
records for as many years as possible;
• collection and analysis of hospital records;
• air quality measurements to verify differences
in coal dust concentrations between exposed
and control areas.
111-2
-------
B. CONCLUSIONS AND RECOMMENDATIONS PERTAINING
TO THE MULTIPLE REGRESSION ANALYSIS
• • ', i •
(1) For certain types of industry-disease associations, the results
of the study have demonstrated that the multiple regression analysis of
county mortality and industry data is effective in identifying associ-
ations having a plausible cause-and-effect basis founded on etiological
considerations and mortality-pollutant patterns at the local level.
The types of industry-disease associations that appear to be identi-
fiable through the multiple regression analysis are those in which
(1) the industry is the sole or principal component of the four-digit
SIC category and (2) the industry is sufficiently intense or sufficiently
rural to expose a large portion of the county population within which it
resides. This conclusion is based on the results of the three case
studies selected from the multiple regression analysis. Each of the
industry-disease associations studied satisfied both of the criteria
above, and each of the case studies provided some evidence of possible
causal relationships between the industry and associated diseases.
(2) The value of the multiple regression approach for identifying
true associations for industries which are not the principal component
of the SIC category or which are not sufficiently intense or rural to
expose a large portion of the county population is unknown. To explore
this capability of the regression approach, further case studies need
to be performed.
(3) The results of the study with regard to the effectiveness and
value of the multiple regression approach are positive and encouraging.
To further develop the capabilities of the approach and to further
understand its limitations, the following recommendations are offered:
(a) Continue the field studies of copper smelting, steel
manufacturing, and bituminous coal mining to provide
stronger evidence that true relationships exist and,
thereby, provide confirmation that the multiple re-
gression approach is effective. (See Section A(2)
above for specific recommendations for these studies.)
III-3
-------
(b) Investigate other industry-disease correlations iden-
tified by the multiple regression analysis. The char-
acteristics of the industries selected for these new
case studies should vary with regard to SIC category
dominance and capability for exposing large fractions
of the county population. In this way, these studies
will provide a means for assessing the limitations of
the regression approach.
(c) Perform a study to develop an industrial index that
will be a better estimator of dosage to exposed commun-
ities. The study should examine (1) the value and
feasibility of developing different indices for dif-
. ferent classes of industries; (2) the feasibility of
incorporating age of industry to provide a measure of
duration of exposure; and (3) the feasibility of
utilizing indices that correlate closely with point
source emission estimates for a given class of industry.
(d) With the improved industrial index (see (c) above),
rerun the multiple regression analysis with recent mor-
tality data (1972-1977) to identify industry-disease
correlations more in line with current waste control
policies and procedures.
-------
C. CONCLUSIONS AND RECOMMENDATIONS PERTAINING TO METHODOLOGY
FOR INVESTIGATING INDUSTRY-DISEASE RELATIONSHIPS
The basic methodology used for investigating industry-disease rela-
tionships consisted of a matched-control approach in which health and
environmental data are compared for exposed and matched-control areas
to identify diseases whose incidence or prevalence can be associated
with emissions or effluents produced by the industry. Based on the
results of the four case studies, certain conclusions and recommenda-
tions can be made regarding this methodology.
(a) The investigation of multiple sites is critical to es-
tablishing and confirming true relationships. Vari-
ation in effect of industry and difficulty in finding
matched-control populations make multiple site investi-
gations a necessity. Two sites were found to be insuf-
ficient to provide satisfactory evidence.
(b) Sites with small populations should be avoided, even
though they are ideal from other points of view (e.g.,
lack of industrial clutter and other sources of pollu-
tion) . The difficulty in gathering health data for a
significant number of years and the weakening of the
analysis due to inability to examine stratifications of
the data make these sites relatively ineffective for
investigation of causal relationships. Although large
city sites were not studied, it is believed that the
advantages of large exposed populations would outweigh
the disadvantages; i.e., the added strength of analysis
associated with large exposed populations would more
than compensate for the problems with industrial clutter
and urban pollution.
(c) A weakness in the matched-control methodology is the
general unavailability of good air quality data for the
control areas. Taking measurements of the current air
quality is not necessarily the answer, since these
^ measurements do not indicate the air quality during
prior years.
III-5
-------
(d) Effective use of time trends could not be made, due to
the'need for aggregating the health data over time and
due to the difficulty of obtaining information on expo-
sure duration (length of- residence) for the affected
individuals. Hence, no conclusions can be made about
the value of temporal trends for investigating
industry-disease relationships.
(e) Neither centralized morbidity data systems nor records
acquired from hospitals could be used effectively. The
former generally lacked needed place-of-residence in-
formation, and the latter could not be obtained for a
meaningful percentage of the hospitals in the study
areas. Based on our findings, we conclude that putting
the effort into investigation of the mortalities at
additional sites is more valuable and has a greater
chance of success than putting the same amount of effort
into collecting and analyzing hospital morbidity records
for a given site. When centralized sources of morbidity
data include adequate address information, morbidity
records should prove useful in the investigation of
causal relationships. It is recommended that steps be
taken by EPA to encourage state and local disease regis-
tries to include census tract or town of residence in-
formation on their morbidity records.
(f) Interpretation of aerial photography for previous years
was useful in providing historical evidence of changes in
population and pollution point sources (i.e., stacks, con-
trol devices, or effluent discharges). However, for the
most part, this information was more readily available
from other sources to the accuracy required in the study.
Based on this experience, we feel that retrospective
photointerpretation is valuable only when corroborative
evidence is required or in special cases where the in-
formation is not available elsewhere. Although not studied
III-6
-------
here, we feel that photointerpretation would have greater
value in present-day or prospective studies, where the data
can be collected as desired and use can be made of various
remote sensing techniques to provide more quantitative in-
formation on the pathways and the dispersion patterns of
emissions and effluents.
III-7
-------
Chapter IV
STUDY METHODOLOGY
A. OVERVIEW
This chapter discusses the various methods and procedures used in
all phases of this study. The initial phase of the study, performed
under a separate EPA contract, identified correlations between broad
industry groups (as defined by Standard Industrial Classification [SIC]
codes) and categories of disease mortality. These correlations were
then screened and validated to choose those most significant and best
suited for further investigation.
Following this, study sites were selected and hypotheses were de-
veloped and tested in separate sites. Data on mortality and morbidity,
air and water quality, etiology, and industry characteristics were col-
lected, analyzed, and integrated to arrive at the conclusions presented
.in this report.
While each industry-disease correlation selected for study was in-
vestigated using a somewhat different protocol, the general methodology
was essentially the same. This general methodology and the specific
procedures comprising it are discussed in this chapter.
B. INITIAL CORRELATIONS
The initial industry-disease correlations were generated primarily
by a multiple regression analysis of disease-specific, age-adjusted mor-
tality rates by county (for years 1968-1972) on indices of industrial
activity for some 460 industries, based on 1959 data. These industries
were, in some cases, actually groups of industries (sometimes unrelated
with respect to type or process) which were consistently located in the
same counties. The index of industrial activity was a function of the
number of employees in the industry and the total work force in the
county. This analysis was done separately for males and females and for
whites and nonwhites. The nonwhite regression had some calculation
errors which could not be corrected in time for inclusion in the later
study phases, so the investigation considered only the associations
between industrial activity and mortality in whites.
IV-1
-------
In addition, the counties were divided by EPA randomly into two
groups for the regression analysis. Therefore, the separate regressions
for each half of the total counties and for each sex provided a measure
of the consistency of the derived correlations. However, even a strong
association might not be reflected as correlations in all subsamples
since males are more likely to be exposed occupationally, and either
subset of counties might not necessarily contain a particular industry.
A total of 791 industry-disease correlations resulted from the re-
gression approach. Other sources of potential associations were used to
augment those found through the regression analysis and to provide a
means for the screening procedures to determine the strongest testable
associations regardless of source. The multiple regressions constituted
a completely objective source providing the largest number of possible
associations. The other potential associations were generated by manual
inspection of county industry and mortality data, through literature
searches and from observations made by State health officials on en-
vironmental health problems. Thirty-six associations were identified in
this manner.
G. SCREENING AND VALIDATION
Correlations from the regression analysis were ranked and then
grouped into six main categories containing 791 associations. These
categories were based on the Values of R and Ar in the regression and
whether or not the derived correlations were present in both subsamples
of counties and in males and females. Table IV-1 describes the six
categories and Table IV-2 lists the correlations by category. Also
listed in Table IV-2 are the 37 associations identified by means other
than the regression analysis.
By a series of statistical tests and screening procedures outlined
in Figure IV-1, a sma.ll number of the strongest correlations were se-
lected from the 791 associations identified by the regression analysis.
These served as the candidate associations for field study. The proce-
dure by which the 791 associations were boiled down to a few is described
below.
IV-2
-------
Correlations Identified By
SSI Regression Analysis
Definition of
6 Categories of Associations
(Table iv-l)
791 Associations
(Table IV-3)
Preliminary Statistical Teiti
and Screening Procedure!
Ranking *nd teleotion of top four
categories
Elimination of unfit industries
and Infectloui diseases
Selection of Induetrlei associated
with two or more diseases
39 Associations Remaining
(Table IV-3)
Statistical Significance Tasting
1. Strength of relatlonihlp between
mortality rates and Industrial
activity
2. Significance of differences
between mortality rates for top 10
counties and national means
3. Significance of differences
between mortality rates for cop 10
counties and for matched control
counties
(Removed counties with industrial
clutter and populations > 200,000.
Split copper into two categories.)
8 Industries and Associated Diseases*
(Table II-U
•Candidate associations
for fioU Jtudy.
Associations Identified By
Other Sources
37 Associations
(Table IV-})
2 Industries and Associated Diseases*
[Table II-l)
4 Industries and
Associated Diseases
figure IV-1. Procedures for Screening and Selecting
Industry-Disease Associations for Case Study
IV-3
-------
Table IV-1
SCREENING CRITERIA FOR"INDUSTRY^DISEASE ASSOCIATIONS
SOURCE
I. Regression
Data*
II. Data
Inspection
III. Basic
Principles
IV. Literature
V. Health
Officials
GROUP
DESIGNATION
a) Upper
MT-AB
b) Lower
MF-AB
c) Upper
M or F-AB
d) Lower
M or F-AB
e) Upper
MF-A or B
f) Upper
M or F-A or B
Inferred
Associations
Basic
Associations
Recorded
Associations
Field
Associations
SCREENING CRITERIA
Correlation in both males and females
Correlation in both sample A and sample 8
Upper R and fir2 class*
Correlation in both males and females
Correlation in both sample A and sample B
Lower R and Ar2 classc
Correlation in either t&ales or females
Correlation in both sample A and sample B
Upper R and Ar2 class'1
Correlation in either males or females
Correlation in both sample A and sample B
Lower R and Ar2 class0'3
Correlation in both males and females
Correlation in either sample A or sample B
Upper R and Ar2 class11
Correlation in either males or females
Correlation in either sample A or sample B
Upper R and Ar2 class*
Association between high industrial activity
and excess mortality
Industrial processing of known disease-
causing agents
Recorded occupational associations
Problems identified by officials
Supportive epidemiological evidence
NUMBER OF
ASSOCIATIONS
25
31
15
t
548
16
156
(791)
24
5
7
1
(37)
aAll data is uncorrected, untruncated, weighted, and white.
''Upper R and 4r2 class (R > 0.5 and Ar2 > 0.01 in at least one subsample of males
and females) .
cLower R and Ar2 class (residue from upper MF-AB, i.e., R < 0.5 or Ar2 < 0.01).
*260 potential associations for males and 288 for females were identified in this
category. The listing is not included in this report.
Nomtnclatura: N • male: F - femalei A • sample A of U.S. counties!
B • sample B of U.S. Counties t R • multiple correlation
coefficient; Ar • change in correlation coefficient
IV-4
-------
Table IV-2. POTENTIAL INDUSTRY-DISEASE ASSOCIATIONS
SOURCE
CROUP
DESIGNATION,
DISEASE*
POTENTIAL ASSOCIATION
'INDUSTRY
I. Regression
Data
a) Upper MF-AB
06 - Neoplasm, digestive organs s
peritoneua
08 - Neoplasm, snail c large
intestines
09 - Neoplasm, rectum
27 - Major cardiovascular disease
28 - Rheumatic heart disease
29 - Hypertensive disease
32 - Chronic ischemic heart disease
I angina poctoris
38 - Diseases of the respiratory
systen
39 - Acute Interstitial t broncho-
pneumonia
40 - Bronchitis, emphysema s asthma
43 - Emphysema
45 - other diseases of the respira-
tory systen
46 - Diseases of the digestive
system
47 - cirrhosis of liver
48 - Alcoholic cirrhosis of liver
49 - other & unspecified cirrhosis
of liver
P030 - Jewelry, toys, wire drawing
PO30 - Jewelry, toys, wire drawing
PO30 - Jewelry, toys, wire drawing
2321 - Hen's dress shirts & nightwear
121R - Bituminous mining
4932 - Gas & other services combined
34RR - Fabricated metal products
P001 - SS big city industries
41RR - Local passenger transportation
PO32 - River transportation & petroleum
102R - Copper mining
316R - Luggage
P001 - SS big city industries
102R - Copper mining
102R - Copper mining
L21R - Bituminous mining
41RR
P001
Local passenger transportation
SS big city industries
P001 - SS big city industries
P018 - Aircraft
PO01 - SS big city industries
2084 - Wines, brandy & brandy spirits
41RR - Local passenger transportation
3273 - Ready mixed concrete
PO41 - Transportation
b) Lower MF-AB
08 - Neoplasm, small & large
intestines
11 - Neoplasm, esophagus, liver,
gallbladder, peritoneum &
other digestive organs
27 - Major cardiovascular disease
29 - Hypertensive disease
30 - Ischemic heart disease
31 - Acute nyocardial infarction
32 - Chronic ischemic heart disease
s angina pector is
34 - Cerebrovascular disease
38 - Diseases of the respiratory
system
PO36 - Motors & generators
P001 - 55 big city industries
POS2 - River transportation & petroleum
PO61 - footwear
2321 - Men's dress shirts & nightwear
121R - Bituminous mining
POS2 - River transportation & petroleum
P047 - steel
32PB - Glass
121R - Bituminous mining
P061 - Footwear
2321 - Men's dress shirts s nightwear
121R - Bituminous mining
327R - Concrete, gypsum & plaster
PO69 - Phonograph records s transmis-
sions
121R - Bituminous mining
These are Jisease codes (or 56 disease categories used by SSI
in the regression analysis U«a Appendix J).
(continued)
IV-5
-------
Table IV-2 (continued)
SOURCE
GROUP
DISICNATIOtl'
DISEASE
I. ASSOCIATION
INDUSTRY
I. Regression
Data
(cent.)
b) Lover Nf-AB
(cent.)
c) Upper
H orf-AB
Females
d) lower
N orP-AB
39 - Acute interstitial c bronoho-
pneunonia
40 - Bronchitis, eaphysena c asthma
41 - Bronchitis
42 - Emphysema
46 - Diseases of the digestive
system
47 - Cirrhosis of livu
48 - Alcoholic cirrhosis of liver
49 - Other C unspecified cirrhosis
of liver
121R - Bituminou* mining
121R - Bituminous mining
121R - Bituninous raining
P006 • Transportation, aircraft,
pottery, «tc.
102R - Copper mining
2951 - Paving mixtures t blocks
2084 - Mine*, brandy s brandy spirita
41RR - Local passengar transportation
2084 - Nines, brandy c brandy spirits
102R - Copper mining
102R - Copper mining
1273 - Ready mixed concrete
POO1 - SS big city industries
2291 - Pelt goods, N.E.C.
327R - Concrete, gypsum S plaster
06 - Neoplasm, digestive organs t
peritoneum
11 - Neoplasm, esophagus, liver,
gallbladder, peritoneum C
other digestive organs
13 - Neoplasm, trachea, bronchus
& lung
38 - Diseases of the respiratory
system
48 - Alcoholic cirrhosis of liver
15 - Neoplasm, breast
27 - Major cardiovascular disease
28 - Rheumatic heart disease
31 - Acute myocardial infarction
49 - Other £ unspecified cirrhosis
of liver
NOTKi In the following correlations.
P047 - Steel
41SR - School buses
POO3 - Foodo, machinery & coinminica-
tions
PO52 - River transportation s petroleum
209R - Miscellaneous food preparations
PO59 - Cigars & anthracite mining
316R - Luggage
POO1 - 55 big city industries
P047 - Steel
279R - Printing trades services
P047 - Steel
PO06 - Transportation, aircraft,
pottery, etc.
r2 > 0.01, but R <0.5.
55 - Diseases of the skin t subcu-
taneous tissue
16 - Neoplasm, female genital
organs
37 - Diseases of veins, lymphatics
6 other diseases of circu-
latory system
4171 - Bus terminal facilities
POO3 - Foods, machinery & comunica-
tions
121R - aituoinous mining
(548 potential associations were identified in this category. Because of
the large number and the weakness of the correlations, the listing is not
Included here.)
(continued)
IV-6
-------
Table IV-2 (continued)
SOURCE
GROUP
DESIGNATION
DISEASE
ASSOCIATION
I. Regression
Data
(cont.)
•) Upper NP-
* orB
08 - Neoplasm, mall «
large intestine*
37 - Major cardiovascular disaaae
31 - Acute myocardlal infarction or
other acute ischealc heart
diaaasa
32 - Chronic Ischanic heart disease
£ angina pectoris
38 - Diaeaaea of the reapiratory
ayatam
39 - oiaaaaaa of the digestive
ayatan
47 - Cirrhoaia of liver
48 - Alcoholic cirrhoaia of liver
49 - Unspecified cirrhoaia of liver
P030 - Jewelry, toys, wire drawing
2649 - Paper t paper board products
P030 - Jewelry, toys, wire drawing
2S9R - Misc. furniture t fixtures
102R - Copper mining
3273 - Ready mixed concrete
279R - Printing trade services
3421 - Cutlery
102R - Copper mining
327R - Concrete, gypsum & plaster
327R - Concrete, gypsum t plaster
PO38 - Motors, tool c die, metal
Industries
259R - Misc. furniture & fixtures
2084 - Wines, brandy & brandy spirits
3421 - Cutlery
3273 - Ready mixed concrete
f) Upper A orB
Males
OS - Neoplasm, buccal cavity s,
pharynx
06 - Neoplasm, digestive organs
07 - Neoplasm, stomach
08 - Neoplasm, small s large
intestines
09 - Neoplasm, rectum
3672 - Cathode ray picture tubes
PO42 - Ind. gases, syn. rubber,
cooperage
2591 - Venetian blinds s shades
349R - Fabricated metal products
279R - Printing trades services
2951 - Paving mixtures & blocks
PO18 - Aircraft
2393 - Textile bags
3672 - Cathode ray picture tubes
279R - Printing trades services
P047 - steel
' 3421 - Cutlery
PO30 - Jewelry, toys, wire drawing
349R - Fabricated metal products
PO27 - Asphalt, asbestos, coal tar
3623 • Welding apparatus
P047 - Steel
29S1 - Paving^ mixtures £ blocks
P030 - Jewelry, toys, wire drawing
38PA - Optics
PO61 - Footwear, tanning, lamps, misc.
textile
P047 - steel
3672 - Cathode ray picture tubes
3297 - Non-clay refractories
PO30 - Jewelry, toys, wire drawing
PO27 - Asphalt, gaskets, etc.
3623 - Welding apparatus
349R - Fabricated metal products
3841 - Surgical & medical instruments
2032 - Canned specialties
279R * Printing trades services
P047 - Steel
3334 - Primary aluminum
37 2R - Aircraft (. parts
P030 - Jewelry, toys, wire drawing
3554 - Paper industries machinery
349R - Fabricated metal products
229S - Coated fabric, non-rubber
354Q - Machine tools, metal working
machinery
3251 - Brick s struct, clay tile
(continued)
IV-7
-------
Table IV-2 (continued)
SOURCE
I. Regression
Data
(coat.)
GROUP
DESIGNATION
f) Upper A ore
Ma lei
(cant.)
POTENTIAL ASSOCIATION
DISttSB
11 - Neoplasm, esophagus, liver.
gallbladder, peritoneum t
other digestive organs
13 - Neoplasm, trachea, bronchus
s lung
14 - Neoplasm, other respiratory
organs
24 - Diabetes nellltus
27 - Major cardiovascular disease
31 - Acute myocardial infarction
32 - Chronic lachemlc heart disease
s angina pectoris
33 - Other heart disease
34 - Cerebrovascular disease
.
11 - Arteriosclerosis
38 - Diseases of respiratory system
INDUSTRY
3672 - Cathode ray picture tubes
. 349R - Fabricated metal products
29S1 - Paving mixtures £ blocks
209R - Miscellaneous food preparations
PO42 - Ind. gases, syn. rubber, etc.
2591 - Venetian blinds & shades
36PO - Batteries, wiring devices
121R - Bituminous mining
229S - Coated fabric, non-rubber
29SR - Paving S roofing materials
202R - Dairy products
29S1 - Paving mixtures t blocks
3672 - Cathode ray picture tubes
2591 - Paving mixtures & blocks
P061 - Footwear, tanning, etc.
349R - Fabricated metal products
PO3B - Motors, etc., metals industries
2295 - Coated fabric, non-rubber
366P - Communication equipment
2591 - Venetian blinds & shades
2821 - Plastics materials
PO30 - Jewelry, toys, wire drawing
3623 - Welding apparatus
3273 - Ready mixed concrete
36PC - Electrical equipment
3273 - Ready mixed concrete
259R - Misc. furniture & fixtures
3SSR - Special ind. machinery
22PC - Rugs, synthetic weaving
349R - Fabricated metal products
PO43 - Semiconductors, tubes.
scientific instruments
2S9R - Misc. furniture 6 fixtures
38PB - Mechanical instruments, etc.
3822 - Automatic temperature controls
3273 - Ready nixed concrete
3631 - Household cooking equipment
2282 - Yarn throwing £ winding S mills
36PD - Batteries, wiring devices
2295 - Coated fabric, rubberized
3672 - Cathode ray picture tubes
36PB - Glee. meas. instru., etc.
366P - Communications equipment
2951 - Paving mixtures & blocks
3672 - Cathode ray picture tubes
2042 - Prepared animal feed
2091 - Cotton seed oil mills
261R - Pulp mills
3662 - Radio, tv communication equip.
314R - Footwear
2393 - Textile bags
361P - Transformers
22PB - Weaving mills, cotton
366P - Communication equipment
PO30 - Jewelry, toys, wire drawing
22PC - Carpets, weaving, synthetics
3273 - Ready mixed concrete
2299 - Textiles
2951 - Paving mixtures t blocks
3672 - Cathode ray picture tubes
121R - Bituminous mining
316R - Luggage
333R - Nonferrous metals
3333 - Zinc
(continued)
IV-8
-------
Table IV-2 (continued)
SOURCE
GROUP
DESIGNATION
POTENTIAL ASSOCIATION
DISEASE
INDUSTRY
I. Regression
Data
(cont.)
f) Upper A ore
Males
(oont.)
Females
47 - Cirrhosis of liver
48 - Alcoholic cirrhosis of liver
49 - Other fi unspecified cirrhosis
of liver
SO - Other digestive diseases
08 - Neoplasm, mall ( large intee.
12 - Neoplasm, respiratory tract
13 - Neoplasm, trachea, etc.
27 - Major cardiovascular disease
28 - Chronic rheumatic heart
disease
29 - Hypertensive disease
30 - Ischenic heart disease
31 - Acute myocardlal infarction
& other acute iacheaic heart
disease
32 - Chronic ischenic heart disease
39 - Acute interstitial £ brbncho-
pneunonia
47 - cirrhosis of liver
48 - Alcoholic cirrhosis of liver
314R - Footwear, except rubber
P017 - Residuals in metal ores, etc.
PO18 - Aircraft engines & props
2951 - Paving mixtures & blocks
P018 - Aircraft engines I props
38PA - Optics
314R - Footwear, except rubber
336R - Nonferrous foundries
3672 - Cathode ray picture tubes
3421 - Cutlery
2063 - Beet sugar
316R - Luggage
3591 - Venetian blinds t shades
279R - Printing trades services
P038 - Machines, hardware, etc.
PO47 - steel
P030 - Jewelry, toys, wire drawing
2951 - Paving mixtures & blocks
3822 * Mechanical measuring devices
279R - Printing trades services
3273 - Ready nixed concrete
3672 - Cathode ray picture tubes
3273 - Ready mixed concrete
P047 - Steel
3442 - Metal doors, sash & trim
2649 - Paper s paper board products
279R - Printing trades services
3421 - Cutlery
PO30 - Jewelry, toys, wire drawing
3423 - Bdge tools
PO30 - Jewelry, toys, wire drawing
3672 - Cathode ray picture tubes
2295 - Coated fabric, non-rubber
P047 - Steel
3442 - Metal doors, sash & trim
P038 - Motors, etc., metals industries
PO30 - Jewelry, toys, wire drawing
2649 - Paper & paper board products
4932 - Gas & other services combined
PO47 - Steel
3442 - Metal doors, sash £ trio
PO33 - Motors, etc., metals industries
2649 - Paper & paper board products
4932 - Gas s other services combined
279R - Printing trades services
341R - Fabricated metal parts
P030 - Jewelry, toys, wire drawing
2S9R - Misc. furniture & fixtures
2064 - Wines, brandy S brandy spirits
3672 - Cathode ray picture tubes
279R - Printing trades services
327R - Concrete, plastur & gypsum
102R - Copper mining
(continued)
IV-9
-------
Table IV-2 (continued)
SOURCE
II. Data
• Inspection
III. Basic
Principle*
GROUP
DESIGNATION
•Inferred
Association!
Basic
Associations
POTENTIAL ASSOCIATION
DISEASE
OS - Neoplasm, buccal cavity
06 - Neoplasm, digestive
organs
18 - Neoplasm, bladder
21 - Leukemia
24 - Diabetes mellltus
31 - Myocardial infarction
32 - Other heart disease
35 • Arteriosclerosis
39 - Pneumonia
40 - Bronchitis, emphysema,
asthna
41 - Bronchitis
52 - Nephritis/nephrosis
Systemic poisoning;
respiratory s skin diseases
Congenital malformations
Mesotheliomai asbestosis
Respiratory disease
Leukemia
INDUSTRY
2421 - Sawmills
2421 - Sawmills
2649 - Paper & paper board products
241R - Logging
2327 - Men's trousers
3442 - Metal doors, sash & trim
2394 - Canvas
2421 - Sawmills
2421 - Sawmills
2421 - Sawmills
2421 - Sawmills
121R - Bituminous mining
121R - Bituminous mining
2321 - Men's dress shirts & nightvrear
2341 - Clothing
2326 - Men's clothing
22 Textiles
31 Leather
287J,- fertilizers, mixing
2421 - Sawmills
2421 - Sawmills
2327 - Trousers
2328 - Men's clothing
241R - Logging
Pesticide s herbicide use & manufacturing
(highly toxic, long-lasting agents which can
enter through several bodily routes) .
Pesticides, industrial processes involving
molybdenum, manganese, zinc, mercury & lead.
(Since the developing fetus is susceptible
to external agents, a study of congenital
malformations in industrialized areas may
reveal new interactions of environmental
teratogens . )
Production of asbestos. (Highly toxic in
small dosesi one exposure is often sufficient
to produce disease. Minute fibers can linger
in air and lungs for long periods. Contami-
nation of communities near asbestos production
is likely.)
Smelters and foundries
Cathode ray picture tubes; nuclear power plants
(continued )
IV-10
-------
Table IV-2 (concluded)
SOURCE
CROUP
DESIGNATION
DISEASE
POTENTIAL ASSOCIATIONS
INDUSTRY
IV. Literature
Recorded
Associations
Neoplasty respiratory tr»ctj
Petrochemicals, sulfur, sulfuric acid,
synthetic rubber, liquid hydrocarbon*,
magnesium, fertilisers (Houston Ship
Channel Area).*
'Citationi MacDonald, Preprint (1976)
Neoplasm, bladder, lung t
liver in nalee
Chemical industry, i.e., dyes, cosmetics,
Pharmaceuticals, soaps, rubber (139 indus-
trialised counties).*
•Citationi Hoover and Frauaeni, Environ. Res. 9:196 (1975)
Neoplasm, nasal t bladder | Leather finishing.*
•Citationi Acheson. Brit. Ned. J. 1:385 (1970)
Cole, Brit. J. Ind. HedT 11^ (1954)
Liver angiosarcoma | Vinyl chloride.*
•Citationi Creech, J. Oecup. Ned. 16:150 (1974)
Neoplasm, lung in males
•Citationi Blot, An. J.
I Paper, chemical, petroleum t transportation.*
pidemiol. 103(6)1539 (1976)
Neoplasm, lung I Uranium mining.*
'Citationi Cassarett, Toxicology. HacHillan (1975)
Cardiovascular diseases I Viscose rayon production.*
•Citationi schilling, Am. Heart J. SOU (1970)
V. Health t
Pollution
Officials
Field
Associations
Cardiovascular disease
Viscose rayon products. (A site in Warron
County, Virginia, with an exces* cardio-
vascular mortality and heavy viscose rayon
production has been identified. Preliminary
profiles of the industrial process and
popualtion at risk have been compiled.)
IV-11
-------
The six categories of regression data were ranked independently by
seven of the ECI staff on the basis of the potential of the various
groups to distinguish viable correlations. The top four categories
determined by consensus were (a), (b), (c), and (e) which contained 87
associations.
From this group of 87 associations we eliminated industry categories
obviously unfit for the study, such as hon-point-source industries
(transportation) and large conglomerate industry groupings (55 big city
industries). Because of its infectious etiology, rheumatic heart dis-
ease was eliminated. Fifty-seven associations remained after this
process.
We then selected the 39 associations in which industries were asso-
ciated with two or more similar diseases. This does not imply that non-
specificity was used as a criterion for selecting associations. Rather,
it aimed at industries which appeared to exert their effects on several
sites within an organ system. These associations are listed in Table
IV-3.
The remaining associations were then examined to determine which
were statistically significant, unconfounded by nonindustrial variables,
and likely to be related causally. For this we used a step-wise process
involving three sequential tests.
The first test consisted of determining the statistical significance
of mortality rates for the counties with high industrial activity, and
then examining the correlation between the degree of significance and the
industrial activity. The second test consisted of aggregating the mor-
tality rates for the top 10 or so counties with the highest amount of the
industry in question, and testing the statistical significance between
the aggregated mortality rates and the national mean mortality rates for
the associated diseases. The third test consisted of a mortality con-
trol test, in which the aggregated mortality rates for the top 10 coun-
ties were compared to the aggregated mortality rates for a set of con-
trol counties, which did not contain the industry in question but whose
socioeconomic characteristics matched those of the industrial counties.
IV-12
-------
The last test was by far the most meaningful test, but involved con-
siderably more effort than the others and, therefore, was applied only
to those associations that survived the first two tests.
Table IV-3
SCREENED INDUSTRY-DISEASE ASSOCIATIONS FROM REGRESSION DATA
ASSOCIATED WITH MORE THAN ONE SIMILAR DISEASE
sic»
INDUSTRY
DISBA8B CATBCORY
SPECIFICS
327R
3273
2951
P030
102R
121R
P047
2321
2084 '
Concrete, gypsum c plaster
Ready mixed concrete
Paving mixture! I block*
Jewelry, toy*, wire
drawing
Copper mining
Bituminous coal mining
Steel
Hen's shirta t nightwear
Nlnea, brandy t brandy
spirits
Cardiovascular diseases
Digestive diseases
Cardiovascular diseases
Digestive diseases
Digestive diseasss
Cardiovascular diseases
Digestive diseases
Respiratory diseases
Cardiovascular diseases
Digestive diseases
Respiratory diseases
Cardiovascular diseases
Cardiovascular diseases
Digestive diseases
Cardiovascular diseases
Digestive diseases
32 - Chronic ischemic heart disease (la)*
47 - Cirrhosis (Ie)
49 - Other cirrhosis (Ib)
46 - Diseases of digestive system (Is)
31 - Myocardial infarction (Ie)
48 - Alcoholic cirrhosis (Ib)
49 - Other cirrhosis
46 - Diseases of digestive system (Ib)
27 - Major cardiovascular diseases (Ie)
06 - Neoplasm, digestive system (la)
08 - Neoplasm, small e. large intestines (la.Ie)
09 - Neoplasm, rectum (la)
38 - Diseases of respiratory system (Id)
40 - Bronchitis, emphysema, asthma (la)
42 - Enphysema (la)
31 - Myocardial infarction (Ie)
47 - Cirrhosis (Ib)
48 - Alcoholic cirrhosis (Ib)
46 - Diseases of digestive system (Ib)
38 - Diseases of respiratory system (la)
39 - Acute interstitial s bronchopneumonia (Ib)
40 - Bronchitis; emphysema, asthma (Ib)
41 - Bronchitis (Ib)
45 - Other respiratory diseases (la)
27 - Major cardiovascular diseases (la)
30 - Ischemic heart disease (la)
31 - Myocardial infarction (la)
32 - Chronic ischemic heart disease (la)
37 - Other circulatory diseases (Ic, females)
27 - Major cardiovascular diseases (Ic, females)
30 - Ischemic heart disease (la)
31 - Myocardial infarction (Ic, females)
06 - Neoplasm, digestive system (Ic)
27 - Major cardiovascular diseases (la)
30 - Ischemic heart disease (la)
31 - Myocardial infarction (Ib)
46 - Diseases of digestive system (Ib)
47 - Cirrhosis (Ib)
48 - Alcoholic cirrhosis (la)
Letters in parentheses refer to Table IV-2 grouping.
IV-13
-------
In all three tests, a major concern was the effect of county popula-
tion on the statistical significance of mortality rate differences. The
race- and sex-specific populations of the counties used in the regres-
sion analysis varied over several orders of magnitude, from less than
one hundred for rural counties to over a million for large city counties.
Obviously, a high mortality rate for a population of 100 is much less
Significant than the same mortality rate for a much larger population.
The basic statistical approach used to test the significance between
mortality rates for different size populations was used for all three
screening and validation methods. It is statistically and mathemati-
cally equivalent to the chi-square test of the difference between two
proportions or percentages.
The first screening technique used was graphical display to test the
strength of the relationship between mortality rate and industrial ac-
tivity. Plots were made for a number of the associations using several
different measures of industrial activity. No strong correlations ap-
peared, although some associations showed trends in a positive direction.
The 'cases which showed promising trends eventually were validated by
another screening procedure, but the graphical method by itself was not
a suitable measure of the strength of association.
For the next screen, sample counties were chosen by taking the 10
with the highest industrial index, [(number of employees)^/total county
work force] for a particular industry, among those with populations less
than 200,000.* The population cutoff was made in an attempt to remove
counties where there would be extensive industrial activity and to avoid
the complications in disease etiology introduced by urban living. Mean
age-adjusted mortality rates were calculated; those rates which were at
least two standard deviations above the national mean remained as indi-
cators of potential investigative associations.
Two categories were removed without calculating aggregated rates be-
cause of the nature of the industry. Printing appeared in 62 counties;
In some instances, there were not 10 counties having populations less
than 200,000, and only the counties that did were included.
IV-14
-------
however, only one county had a population less than 200,000, and the
association was more likely with urbanization than with printing. A
miscellaneous furniture category was also omitted because all 17 coun-
ties had populations above the cutoff. The industrial index gave an
indication of this distribution: all printing counties had indices less
than 0.3, and all furniture counties fell below 0.1.
The remaining industry-disease associations involved 12 industries.
The counties included in the aggregates were then examined for compli-
cating industrial activity. Eliminated from the aggregate were counties
with mining, smelting, or chemical industries, as well as those with any
industry with an SIC code having the same first two digits as the indus-
try of interest. As a result, the industry grouping "anthracite coal
mining and cigars" (PO59) was eliminated because eight of the ten sample
counties contained other industries considered to be potential complica-
tions, i.e., inorganic chemical manufacture, blast furnaces, and bitumi-
nous coal. Removal of those eight counties lowered the number of stan-
dard deviations above the U.S. mean from 33.0 to 0.1. It was also noted
that the industry being examined was mainly cigars (in all 10 counties)
since anthracite mining was present in only two. A similar but not as
dramatic a change occurred with "men's shirts and nightwear." Removal
of two counties containing coal mining dropped female rates for three
heart disease categories from values of more than ±10 sigma to less than
±2. Nevertheless, "men's shirts and nightwear" was retained as a candi-
date association for further testing, since its significance was still
appreciable. This also added some weight to our association for bitumi-
nous coal mining.
It was decided to split one industry (copper ore mining) into two
components because the smelting operation is so frequently associated
withrmining, introducing an obvious complication. Counties containing
mining operations without smelters were taken as one set, and a new
group for copper smelting was established. After the composites had
been narrowed in this manner, new aggregated mortality rates were cal-
culated. Only three of eight disease associations held up for mining,
whereas six met the test for smelting.
IV-15
-------
The final test of the associations was to determine if the mortality
rate for the sample counties was significantly higher than that calcu-
lated for a control set. The method consisted of matching each county
in the sample set as closely as possible with a county in which the
industry was not present. The matching criteria included: same state
(if possible); median family income; percent foreign stock (matching
country of origin if the number was greater than a few percent); percent
living in urban areas; highest level of school completed; median age;
population per square mile; and net migration. The control set thus
contained the same number of counties as the sample group, located in
the same or adjacent states and having approximately the same socioeco-
nomic characteristics. The comparison between mean aggregated rates for
the samples and their controls led to a final set of eight industries
having a total of 21 diseases associated with them. For all of these,
the sample was at least two standard deviations above the control for
one sex and, in most cases, for both sexes.
Associations derived from nonregression sources were submitted to
similar screening procedures. Very few of them held up under this
method, which was not altogether surprising since the data used for the
testing (i.e., county mortality rates) was the same data that had been
used to generate the regression lists. It would be expected that these
associations would have shown up in the regression analysis, if the mor-
tality rates for counties where the industry was located were higher
than average. It was possible, however, that the method of comparison
with matched controls could show up a relative difference without
absolute values being abnormally high.
It was necessary to use somewhat modified procedures because of the
nature of some of our industry definitions. Viscose rayon production,
for example, is only one component of SIC code 2823. Plants were located
using industrial contacts. A county not containing a plant but adjacent
to the county where the plant was located was also included when it was
observed that mortality rates there were among the highest found for
that disease and after learning that the plant was near the county bor-
der. Thus it is an advantage to know plant sites at this stage rather
than knowing only that the county contains the industry somewhere.
IV-16
-------
For uranium mining (SIC 1094), mortality rates for diseases 05
through 56 were computed for eight counties. Those diseases having
rates greater than two sigma above the national average were tested in
matched control counties.
After the initial examination of sawmills (SIC 2421), that category
was eliminated because too many counties contained it; also, it was be-
lieved that the original association came about because of the abundance
of sawmills rather than because of the existence of a true relationship,
since the diseases associated with the industry were quite varied.
We also attempted to determine potential associations by examining
the 10 counties with the highest mortality rates for arteriosclerosis
and congenital anomalies. Arteriosclerosis was studied by taking the
10 counties having populations above 10,000 that had the highest total
white mortality rate for the disease. Industrial activity in these
counties was examined to determine if any SIC code appeared consistently,
but none was found. The same is true for congenital anomalies.
Two of the associations derived from nonregression sources survived
the screening and validation procedures, making a total of 10 industries
with their associated diseases appropriate for further study. The final
group of screened and validated associations is listed in Table II-l in
Chapter II of this report. The selection of four industries was thought
to be the maximum number for adequate study under existing time and
budget constraints.
D. HYPOTHESIS DEVELOPMENT AND TEST
The basic case study approach for each industry-disease association
involved two major steps: causal hypothesis development and causal hy-
pothesis testing. The hypothesis development consisted of identifying a
feasible industry-to-disease linkage comprised of a suspect industrial
pollutant or etiological agent, a plausible pathway from the industry to
the population, increased disease-specific mortality, and a credible
disease etiology. The hypothesis testing involved a series of epidemi-
ological analyses at specific sites to provide evidence that the hypoth-
esis is valid—or invalid, as the case may be. For the most part, these
IV-17
-------
analyses consisted of retrospective examinations of mortality and air
and water quality patterns in relation to the source of the hypothesized
toxic agent (i.e., the industrial facility), giving consideration to
temporal as well as spatial associations.
As originally planned, the hypothesis development was to evolve from
a detailed examination of the industry, the environment, and the disease
etiology for specific sites and for the specific diseases under consider-
ation. The hope was that this would lead to identification of an etio-
logic agent emitted from the industry, a feasible pathway to an exposed
population, and a credible disease etiology. Early on it was found that
this type of examination would, in general, lead to several hypotheses—
any one of which could be valid. In order to determine which of the
hypotheses was indicated by the particular site situation and to verify
the possibility of an industry-related community disease problem, it was
decided to include the analysis of mortality data in the hypothesis
development process. By examining mortality patterns, this analysis
indicated whether or not there was an association between disease and
industry for the particular site under study and pointed to the
hypothesis which best explained the association.
In general, both the hypothesis development and test approaches in-
volved the same key elements, the primary difference being in the choice
of sites for study. This will be discussed in greater detail in the
following section; but, essentially, hypothesis development sites were
located in counties with high mortality rates, while test sites were
selected with no consideration given to mortality rates.
Detailed industrial characterizations identified potentially hazard-
ous emissions or effluents. Pathways were determined from knowledge of
the pollutants involved and from meteorologic, hydrologic, demographic,
topographic, and aerial photography data. Etiologic relationships were
determined from occupational diseases, literature searches, and sugges-
tions by expert consultants.
The synthesis of information about the elements of the hypothesis
resulted in the formulation of several alternative hypotheses as they
IV-18
-------
related to specific populations 'at risk. Each of these hypotheses was
then evaluated by examination of mortality, morbidity, and air and water
quality data.
Confirmation of the etiological hypothesis resulted from the exposed
area having increased mortality rates for the disease being studied,
while rates from other diseases are reasonably equivalent to those in
the unexposed area. Thus, examination of mortality rates from each of
several disease categories demonstrates, if the hypothesis is tenable,
that only the disease that has been etiologically implicated causes
increased mortality in the exposed area. If several unrelated diseases
are found to cause increased mortality, the etiologic hypothesis would
not be reinforced due to the lack of specificity in elevated mortality
rates.
Air and water quality data reflect the validity of the proposed
pathway. It is evident that these data provide a confirmation of the
existence of a plausible pathway. Together, health and environmental
quality data determine which hypotheses are consistent with the actual
data available at a particular site.
The supporting analyses to the hypothesis formulation and validation
process discussed above are: (1) the industry analysis to identify
emissions and effluents; (2) etiology analysis to identify possible eti-
ologic agents; (3) pathway analysis to identify possible pathways and
exposed populations; (4) air and water quality analysis to confirm
probability of population exposure .to toxicants; and (5) mortality and
morbidity analysis to determine disease prevalence patterns. These five
analyses each contributed to the hypothesis synthesis and validation;
the magnitude of their contribution was considered as criteria for
acceptance or rejection of a hypothesis.
It should be noted that each investigation differed slightly in the
hypothesis development and testing procedures, although they all fol-
lowed the same general procedure outlined above. These differences re-
sulted from variations in the type, quality, and availability of data at
each study site.
IV-19
-------
B. SITE SELECTION
The major assumption governing the selection of sites was that the
inclusion of. large metropolitan areas would substantially confuse, con-
found, and bias the study's results. It was felt that the possible ef-
fects of urban living, nonindustrial pollution, and industrial clutter
would seriously compromise the study's efficacy. Furthermore, given the
existence of suitable demographic data for nonmetropolitan areas, it
really was not necessary to include large cities in the study. The
ultimate effect of this decision has been discussed in Chapter III of
this report.
For hypothesis development, study sites were selected on the basis
of both high mortality rates and high industrial activity. After a
hypothesis, consisting of an industry-originated etiologic agent and a
plausible pathway, was developed, then study sites were selected for
hypothesis testing on the basis of industrial characteristics only.
Thus, the complete process required two successive site selections, with
the second being performed only after a plausible hypothesis was deter-
mined at the first site.
The general criteria used in the selection of the hypothesis devel-
opment sites were:
• High industrial activity for the study industry.
• Existence of sufficiently large population at risk.
• Lack of industrial clutter in or near the site.
• For the disease of interest, age-adjusted, sex-specific
county mortality rates statistically significantly
higher than the national average.
The criteria used in the selection of the hypothesis test sites
were:
• For the study industry, presence of high industrial
activity and emission characteristics that coincide
with the hypothesis being tested.
• Existence of sufficiently large population at risk.
• Lack of industrial clutter in or near the site.
IV-20
-------
In addition, to provide a reasonable comparison for the study areas
designated as "exposed," similar nearby areas, presumably unexposed,
were chosen as controls. Control sites were selected to match the study
sites -in terms of population size, demographic characteristics, and
location.
The analytical procedure for selecting the final matched control
sites from the group of preliminary sites used detailed census data—
i.e., age, family income, residence in 1965, race, education, sex, na-
tivity, heating fuel, and number of persons per room in housing units—
to match study and control areas on the basis of socioeconomic charac-
teristics. The procedure utilized a weighted average of least squares
deviations to compare the census statistics for study and control sites.
The site which was the closest match, over all weighing schemes, was
selected.
F. DATA COLLECTION
A major effort in this project has been to identify and acquire the
necessary secondary data. The types of data acquired were mortality and
morbidity, air and water quality, industrial processes, and demographic.
In general, mortality data were obtained from State Vital Statistics
Offices, either as hard copy or on computer tape. While computerized
data were obtained in less time and required less effort to prepare for
analysis, they also provided less information than did copies of death
certificates.
Air and water quality data were obtained from State Air Quality
Reports, EPA's AEROS Data Systems, State Departments of Health, EPA's
STORET System, local engineers and sanitarians, regional EPA offices,
and individual industrial facilities.
The best source of industrial data was the particular company itself.
About half of the companies we contacted provided process descriptions.
The EPA and U.S. Department of Interior libraries supplemented this
information.
IV-21
-------
Demographic data were readily available for all study areas from
the Census Bureau. We obtained Fifth Count tapes from the 1970 Census
for all the areas under study, and we also obtained the enumeration
district maps from the Census Bureau which enabled us to correlate the
street address or town of residence to a specific population at risk.
G: INDUSTRY CHARACTERIZATION
A principal effort in each of the case studies was to characterize
the specific industries under investigation in terms of processes and
pollution sources, with particular emphasis on historical trends. In-
formation was sought from various sources including trade literature,
trade associations, federal, state and local agencies, and the plants
themselves. Of particular value were EPA literature and data bases,
e.g., NEDS, arid EPA reports concerned with developing point-source
guidelines for specific industries. In acquiring information from spe-
cific industrial plants, some plants were very helpful while others
simply ignored our requests. State and local health departments were
cooperative in supplying site-specific information, but generally the
information was limited. In some cases the local Chambers of Commerce
were particularly useful in providing information on plant history.
ff. PATHWAY ANALYSIS
In the pathway analysis a mechanism was postulated for each poten-
tial association of industrial pollutant and health effect, by which the
pollutant might have been transported from the source to the affected
population. Airborne or waterborne transport was considered, according
to the type of emission.
For airborne transport, some evidence was available from air quality
rionitoring in the presumed area of population exposure. Generally, how-
ever, it was necessary to estimate the pathway and the intensity of
exposure from historical meteorologic data. The most useful data were
from the STAR (STability ARray) analyses prepared by the National
Weather Center in Asheville, North Carolina. These data present wind
IV-2 2
-------
speed and direction for various classes of atmospheric stability and can
therefore be used, with a diffusion model, to estimate ground level con-
centration patterns in conditions favoring air pollution. STAR data are
available off-the-shelf for some observing stations and are prepared to
order for others. Unfortunately they are available only for major re-
porting stations (normally commercial airports), and the nearest station
to a study site might be 50 to 100 miles distant and in different ter-
rain. In addition, source strengths have usually to be estimated from
very scanty data, especially for periods well in the past. In such
circumstances it is not permissible to attach any more weight to calcu-
lated exposures than an indication whether significant airborne exposure
can be plausibly hypothesized or not. Appendix A describes the atmos-
pheric dispersion methodology used in the case studies to provide a
basis for estimating the plausibility of exposure.
For waterborne transport, there are two types of pathways that can
sometimes be very clearly defined: in surface streams and in under-
ground aquifers. These can furnish unequivocal evidence for or against
a potential pathway: for example, industrial effluent into a river can
be assumed not to affect an upstream potable water intake but likely to
affect one downstream. If underground water is withdrawn by individual
private wells the picture may be much less clear, with source and path-
way in doubt and usually no useful evidence from sampling. Where ap-
propriate we obtained information on stream and other water sampling,
and on geologic characteristics, from the U.S.' Geological Survey or from
State organizations.
J. ETIOLOGY ANALYSIS
The etiologic basis for industry-disease associations was investi-
gated .by means of literature searches and communications with medical
experts. Relevant literature on the toxicology, epidemiology, and
pathophysiology of diseases of interest was identified through computer-
ized and manual searches and photocopied for reference throughout the
study period. Additional information on emission characteristics of the
suspect industries and occupational exposure within them was also col-
lected.
IV-23
-------
J. AIR AND WATER QUALITY ANALYSIS
In analyzing air quality, our first approach was to establish tem-
poral and spatial trends in ambient concentrations of pollutants for as
many years back as the monitoring stations in the areas of interest had
covered. A visual screen of the data indicated any extraordinary levels
of pollutants. For those substances identified, graphing of the concen-
trations was done to set up visual aids for the purposes of hypothesis
development.
In taking this approach, we were aware of two inherent potential
difficulties. First, air quality data come from monitoring stations
that are not necessarily located near the populations. It is difficult
to establish from several stations whether they represent high, low, or
average concentrations of the pollutants of interest, since meteorologic
factors have not been taken into account. Second, monitoring station
data may not reflect the concentration to which the population at risk,
if any, is being exposed. For example, monitors located several miles
from a plant may be measuring substances in the stack plume, while the
population surrounding the plant may be exposed mainly to fugitive emis-
sions that are not related to stack emissions.
The second approach taken in the air quality analysis was the es-
tablishment of a scenario describing how the atmospheric condition af-
fects the dispersal of pollution from the stacks. This scenario was
based on wind speed and direction data and atmospheric stability infor-
mation which was obtained from the National Weather Service.
Ideally, one might have hoped to obtain a quantitative figure for
the dosage of a particular pollutant experienced by the population at
risk over a 30- to 50-year exposure period. This was not feasible for
many reasons—foremost among them being the paucity of data available.
Concentrations and absolute values of pollutants going up a stack were
obtainable from stack samples and/or emission inventories. Evaluating
their behavior in the environment is far more complex, due to variabil-
ity of the atmosphere. The monitoring stations in the population cen-
ters we investigated, if present at all, had been established within the
last several years and thus were not capable of providing us with data
on earlier levels.
IV-24
-------
Investigation of water quality consisted primarily of analysis of
chemical water quality of the potable water sources. The source of
potable water (surface or underground) was identified, and then data on
water quality were obtained for as far back in time as was available.
We then determined the magnitude of the parameters and examined the data
for temporal and spatial trends.
For the analysis of community water supplies coming from underground
sources, the initial step in the analysis of the well water quality data
was to visually scan the data for each well and evaluate whether there
had been any temporal trend for eacn chemical parameter. Data were
tabulated for the water quality parameters measured for each well. The
wells were grouped by the following priorities: (1) town or population
served, (2) proximity to one another, and (3) proximity to industrial
operations. For those wells that had a systematic variation for certain
parameters, graphing was done and slopes of the linear least squares
line determined. Due to seasonal variations in concentration, the
month(s) in which the data were collected was included on these graphs.
These values were compared with linked parameters (such as TDS and sul-
fates) from the same wells, and values from other wells in the same well
field, other sources supplying the same town, and other wells in indus-
trial and control areas.
In order to link the wells with the industry and the population
served, the following information was obtained:
Length of time the well had been in operation.
Draw depth of the well for the time period of interest.
Type of underlaying strata for each well.
Extent and type of casing.
Importance of each well in supply distribution system.
Service area of population served, present and historical.
K. MORBIDITY/MORTALITY ANALYSIS
Morbidity data were available at two sites. These data were col-
lected from hospital records and cancer registries and analyzed for dif-
ferences in number of disease-specific admissions between study and con-
trol areas.
IV-2 5
-------
The mortality data analysis considered principally the strength of
the association between exposure and mortality. The strength of the
association was assessed by examining the magnitude of age-specific
crude mortality rates and summary statistics, such as standardized rate
ratios. In addition, statistical significance testing of differences
between standardized rates and the Mantel-Haenszel procedure for ob-
taining a summary relative risk over age categories were used. We also
calculated the number of deaths that might be expected in a study area
based on its age distribution and the age-specific mortality rates in
the unexposed area.
For the purposes of determining disease specificity, the case studies
examined the mortalities for 34 cause-of-death categories. These 34
categories include nearly all of the 56 cause-of-death categories exam-
ined in the SSI regression analysis, but they represent a different
method of disease categorization. Reference is made to Appendix J for-
the relationship between the cause-of-death categories and the ICDA
disease codes.
IV-2 6
-------
Chapter V
COPPER SMELTING CASE STUDY
A. INTRODUCTION AND SUMMARY
1. Study Objectives
The objective of this case study is to perform field investigations
to explore the relationship between the copper smelting industry and
community disease, in particular, digestive, respiratory, and cardio-
vascular diseases. The association between these diseases and the cop-
per smelting industry was generated by the multiple regression analysis
which correlated industry data with mortality data for U.S. counties [1].
The results of this present case study will assist in the evaluation of
the effectiveness of the multiple regression analysis for identifying
true associations between industry and community disease.
We wish to determine if the diseases originally associated with the
industry in the regression analysis might be caused by the industry or
might be due to other factors. We also wish to determine if there are
other health problems in the community which were not identified by the
regression analysis. Finally, we wish to evaluate methodology for use
in the objective identification and exploration of community health
problems caused by^ industrial pollution.
2. Study Approach
The approach consisted of selecting study sites and collecting the
health and environmental data necessary to test the hypotheses. In this
particular investigation, we tried to evaluate as many primary copper
smelting sites as possible in order to provide replication and overcome
the problems posed by the small populations of smelter towns. We also
investigated diseases other than those found by the regression analysis
to assist in determining the specificity of the original associations.
[1] System Sciences, Inc., Investigation Into the Industrial Correlates
of Environment-Related Mortality, 1976.
V-l
-------
Thus, we collected mortality records for 34 disease groupings in order
to permit identification of diseases missed by the regression approach.
Only the data for white females were used, since many of the males in
the smelter towns work in the industry and hence are exposed to an en-
tirely different environment for one-fourth of the time. Mortality
rates, actual numbers and expected numbers of deaths, standardized mor-
tality ratios (SMRs), relative risks as compared to control areas, and
statistical significance were examined. Mortality trends were then
integrated with the pollution data in order to determine if relation-
ships between the copper smelting industry and community disease existed.
The eight sites studied were:
Hayden, Arizona (Gila County)
American Smelting and Refining Company
Kennecott Copper Corporation
Ajo, Arizona (Pima County)
Phelps-Dodge Corporation
Douglas, Arizona (Cochise County)
Phelps-Dodge Corporation
Morenci, Arizona (Greenlee County)
Phelps-Dodge Corporation
San Manuel, Arizona (Pinal County)
Magma Copper Company
Miami-Inspiration, Arizona (Gila County)
Inspiration Consolidated Copper Company
Anaconda, Montana (Deer Lodge County)
Anaconda Company
Magna, Utah (Salt Lake County)
Kennecott Copper Corporation
These sites represent eight of the total of 15 primary copper smel-
ters 'in the United States. The objective of choosing several sites for
*
this particular study was an attempt to achieve greater reliability
through replication. The ideal situation, of course, would have been
to examine data for all 15 sites, an impossibility under the scope of
the present study.
The selection of the eight sites was governed more by practical than
by technical considerations. In this study, our starting point is
V-2
-------
already existing data; thus, the sites most feasible to examine are
those for which the necessary data are readily available. Thus, in the
case of six of the sites, the central data collection was done by the
Arizona.Department of Health Services. Acquisition of information on
six sites was in this case much simplified. Anaconda was chosen because
of the ready availability to us of a large body of health and pollution
data collected by state, local, and federal agencies. Magma in Utah was
included because data on it were available as a result of the steel case
study.(see Chapter VI).
3. Key Findings
Four hypotheses pertaining to industry-disease relationships were
formulated after the integration of mortality and pollution data. They
are:
• Elevated mortality from acute respiratory diseases in
Arizona smelter towns appears to be related to a combina-
. tion of high sulfur dioxide and high particulates.
• High arsenic and cadmium from smelters in Anaconda may
account for the excess mortality from cancer of the colon
and pancreas.
• Excess mortality from cerebrovascular diseases in Anaconda
can be explained by high levels of ambient cadmium.
• Digestive diseases, in particular cirrhosis of the liver,
are most likely a result of excess alcohol consumption,
although the synergistic effects of cadmium, lead, and
arsenic cannot be ruled out.
V-3
-------
B. BACKGROUND
1. Basis for Selection of the Copper Smelting Industry
In the original regression analysis performed by System Sciences,
Inc. (SSI), copper ore mining (SIC 102R) was found to be associated with
several disease categories:
• Respiratory diseases—bronchitis, emphysema, asthma, and
acute respiratory diseases such as pneumonia, influenza,
pleurisy, edema, and bronchiectasis (ICDA* 460-519,490-493)
• Cardiovascular diseases—myocardial infarction (ICDA 410-411)
• Liver diseases—cirrhosis of the liver and alcoholic cir-
rhosis (ICDA 571)
• Digestive diseases—nonmalignant diseases of the digestive
system such as diseases of the esophagus, stomach, duodenum,
intestine, peritoneum, gallbladder, and pancreas (ICDA 520-
577)
Although the association was for copper mining, not smelting, we
questioned the validity of using the copper mining industry as a start-
ing point for a study of community pollution for several reasons:
• It has no obvious toxic emissions or effluents, other than
ambient dust.
• Mining operations are frequently located some distance from
population centers.
• Mining is very often associated with nearby smelting opera-
tions, which are notorious producers of pollution and hence
a more logical source of community health problems.
Furthermore, in the original regression methodology, copper smelting had
been grouped into a large, heterogeneous category of industries and had
not been analyzed as a separate industry. Therefore, it was quite
likely that mortalities actually associated with the smelting industry
were obscured due to the aggregation, and that the mortalities observed
for mining were actually related to the smelting operation. This rea-
soning prompted us to focus on the copper smelting industry.
ICDA codes from the Eighth Revision International Classification of
Diseases.
V-4
-------
2. Overview of the Copper Industry
Copper occurs free in nature as native copper, the predominant ore
being the sulfide. The ore is mined by both open-pit and underground
means. Since mined ore contains at most 6% copper, extensive extraction
procedures are required. The processes used vary, depending on the
nature of the particular ore. In general, they involve crushing, con-
centration by flotation, roasting at 600° to 800°C to remove part of
the sulfur, and finally smelting at 1100° to 1600°C to produce matte.
At this stage, the matte is about 30% copper sulfide and iron sulfide.
The sulfur is removed by air treatment in a converter, and the iron is
removed as oxide slag by addition of a siliceous flux. The blister
copper is then further refined to remove traces of sulfur, iron, and
precious metals in a reverberatory furnace or by electrolyte refining.
1
A schematic of the processes is given in Figure V-l. A more detailed
description of the industry is given in Appendix B of this report.
CONV1
ATIHC
-
FI8Z-
imSING
-
ttECTROimC
SXTutuK
-1
Figure V-l. Primary Copper Processes
V-5
-------
During the process of smelting, sulfur dioxide and participates con-
taining heavy metals (such as lead, cadmium, and arsenic) are released.
Until the 1970's, pollution was largely uncontrolled; however, recent
regulations (ca. 1970) have required the industry to control the levels
of emissions by the use of antipollution equipment such as electrostatic
precipitators and sulfuric acid plants. The amount of uncontrolled
emissions produced is huge. For example, the Kennecott smelter at
Hayden produces almost 200 tons of sulfur dioxide per day and about 1
ton of particulates per day [2]. For every ton of copper produced, 0.07
pounds of cadmium and 5 pounds of arsenic are released [3]. A typical
smelter can produce hundreds of tons of copper a day.
3. Health Hazards Associated with Emissions from the Industry
There are no "diseases of copper smelters," per se, although epi-
demiological and toxicological data suggest the plausibility of a cause-
and-effeet relationship between certain emissions and diseases.
Some specific adverse effects are summarized below. A detailed and
documented review of the literature is given in Appendix C.
a. Sulfur Dioxide
Sulfur dioxide combines with particulates, moisture and/or oxidants
to form acid sulfates, which are known to aggravate respiratory diseases
such as asthma and chronic bronchitis and to cause reduced lung function
and irritation of the respiratory tract. Much literature exists that
reports associations between sulfur dioxide pollution and increased mor-
tality from respiratory diseases. The health effects appear to be re-
lated to length of exposure.
[2] EPA, Research Triangle Park, N.C., Office of Air Quality Planning &
Standards, Emission Factors for Trace Substances, EPA-450/2-73-001,
Research Triangle Park, N.C., December 1973.
[3] Effluent Guidelines Division, Office of Water and Hazardous Materi-
als, EPA, Development Document for Interim Final Effluent Limitations
Guidelines and Proposed New Source Performance Standards for the
PRIMARY COPPER SMELTING SUBCATEGORY AND THE PRIMARY COPPER REFINING
SUBCATEGORY OF THE COPPER Segment of the NONFERROUS METALS MANUFAC-
TURING Point Source Category, EPA 440/1-75-032-b, Washington, D.C.,
February 1975..
V-6
-------
The action of sulfur dioxide appears to be related to the presence
of particulates. In a statement prepared for Congress, the National
Academy of Sciences said that:
"Sulfur dioxide acting alone is unlikely to be physio-
logically important, except during periods of severe
inversion coupled with grossly inadequate emission con-
trol. However, sulfur oxide concentrations that are
within the range of ambient concentrations not infre-
quently experienced in polluted areas of the United
States may produce effects by one or more of the fol-
lowing mechanisms:
1. A direct effect of ambient sulfuric acid.
2. A direct effect of ambient suspended par-
ticulate sulfates.
3. An effect of sulfur dioxide potentiated
by adsorption onto inert particles of
respirable size.
4. An effect of sulfur dioxide potentiated
by metallic particles that catalyze its
' oxidation in the airway.
5. An effect of sulfur dioxide potentiated
by simultaneous inhalation of ozone.
Inasmuch as the various sulfur oxides tend to be present
together in polluted air, the effects described above may
well occur simultaneously in an additive, or perhaps
synergistic, manner." [4]
The synergism between particulate matter and sulfur dioxide in chronic
effects on the lung is thought to be mediated via adverse effects on
physiological clearance mechanisms [5].
At annual mean concentrations of 0.04 ppm, the frequency of lung
disease may increase and mortality from bronchitis and lung cancer may
occur. At concentrations of 0.2 to 0.3 ppm, patients with chronic lung
disease may experience accentuation of symptoms, and the general popula-
[4] Air Quality and Stationary Source Emission Control, A Report by the
Commission on Natural Resources, National Academy of Sciences, Na-
tional Academy of Engineering, National Research Council, Prepared
for the Committee on Public Works, U.S. Senate, Serial No. 94-4,
March 1975.
[5] Proceedings of the Conference on Health Effects of Air Pollutants,
Assembly of Life Sciences, National Academy of Sciences-National
Research Council, October 3-5, 1973, Prepared for the Committee on
Public Works, U.S. Senate, -Serial No. 93-15, November 1973.
V-7
-------
tion may experience increased mortality [6]. At concentrations greater
than 1 ppm, sulfur dioxide causes nausea and coughing, even in persons
without preexisting respiratory disease. The taste threshold is at 0.3
ppm. The relationship between length of exposure and concentration of
S02 is shown in Figure V-^2 [7] . The air quality standards for the three
states studied—Arizona, Utah, Montana—are shown in Table V-l.
10 Years
ui
c
a
in
a
UL
O
X
O
ui
1 Year -
I Month -
« O»v« -
I
1 O»y -
8 Hours -
1 Hour _
5 Min. _
30 See. -
' J\! Hi/ ., 'I ' 1—171 T—i l l i '
-j Increased CjrdlovOSculJ' Morbidity.
~f— Increased HospitJl Admissions ',
3 Sec. I
,,,-si
Range of concentrations and
exposure tirr^i in winch cca'.ni
h;v; ss-n roorted i.i excels
o! no/mjl ex^5ceo(ion.
Rang* of concentrjtions jnd
exposure tiinsj in v/hich s!<)-
nific*nt h?s!in e"cctsh3v*
been reported.
Ran;s of concentrations and
excosure times in vvnic*" fJO
heditn effects are suspected.
Conversion factor: (|ig/m3) (3.75
0.01 0.02 0.10 0.2 0.5 1.0 2.0 5.0
SULPHUR DIOXIDE (P.P.M.J,*
iO.O
Potentially haraful doses of SOj:
SO uq/m3 (.02 ppm) for 1 year
240 ug/m3 (.09 ppm) for 1 month
, BOO pg/m3 (.30 ppn) for 1 day
ppm
Figure V-2, Health Effects of Sulfur Dioxide Pollution [7]
[6] HEW, Consumer Protection and Environmental Health Service, National
Air Pollution Control Administration, Air Quality Criteria for
Sulfur Oxides, NAPCA Publ. No. AP-50, January 1969.
[7] Barringer, Health effects of sulfur dioxide pollution, in Correla-
tion Spectrometry Applied to Earth Resources, Princeton University
Conference, June 1970.
V-8
-------
Table V^l
AMBIENT AIR QUALITY STANDARDS: SULFUR DIOXIDE & PARTICULATES—
ARIZONA, UTAH, MONTANA, FEDERAL
ARIZONA STANDARDS'
POLLUTANT
Sulfur Dioxide
Particulates
AVERAGING PERIOD
Annual
72-hour
24-hour
1-hour
Annual
AMBIXNT AIR QUALITY STANDARDS
50 ug/m3
120 ug/m3
250 ug/m3
850 ug/m3
70 ug/m3
UTAH STANDARDS^
POLLUTANT
Sulfur Dioxide
Particulates
AVERAGING PERIOD
Annual
24-hour
3-hour
Annual
24-hour
AMBIENT AIR QUALITY STANDARDS
PRIMARY
80 ug/m3 (0.03 ppo)
365 ug/m3 (0.14 ppm)
None
75 ug/m3
260 ug/m3
MONTANA STANDARDS0
POLLUTANT
Sulfur Dioxide
Suspended
Particulates
Settled
Particulates
AVERAGING PERIOD
Annual
24-hour
1-hour
Annual
24-hour
3-month
SECONDARY
None
None
1300 ug/m3 (0.5 ppm)
60 ug/m3
150 ug/m3
AMBIENT AIR QUALITY STANDARDS
0.02
0.10
0.25
PP»
ppm
ppm
75 ug/m3
200 ug/m3
15 T/mi2
(residential area)
FEDERAL STANDARDS
POLLUTANT
Sulfur Dioxide
AVERAGING PERIOD
Annual
24 -hour
3-hour
AMBIENT AIR QUALITY STANDARDS
PRIMARY
80 ug/m3
365 ug/m3
(to Standard
SECONDARY
60 ug/m3
260 ug/m3
1300 ug/m3 .
*Arisona Department of Health, Division of Air Pollution Control, July 1972.
''Utah State Division of Health.
°Annual Air Quality Data Sunaary for Montana, 1976, Department of Health and
environmental Science*, Air Quality Bureau.
V-9
-------
b. Arsenic
Adverse effects of arsenic include anemia, gastric disturbances,
renal symptoms, ulceration, carcinogenesis (poorly differentiated epi-
dermqid bronchogenic carcinoma), and liver and common bile duct degene-
ration in animals. .Arsenic may exert a synergistic effect with alcohol,
and may be taken into the body through air, water, and food.
Maximum permissible atmospheric concentrations have not been adopted
in. the United States for arsenic. Water standards, however, have been
adopted, and concentrations of 2 to 4 mg per liter are reported not to
interfere with the self-purification of streams. The Public Health Ser-
vice Drinking Water-Standards state that the concentration of arsenic in
drinking water should not exceed 0.01 mg/liter, and concentrations in
excess of 0.05 mg/liter are grounds for rejection of the supply [3].
. The magnitude of the arsenic pollution (and other heavy metals) pro-
duced by smelters is enormous. For example, it has been determined that
1 to 4 milligrams of arsenic are deposited per square meter per month
within a 1-mile radius of a zinc smelter in East Helena, Montana [8].
The potential to spread to the surrounding community has also been
demonstrated by the finding of a gradient of urinary arsenic in residents
living downwind from a smelter. Levels were 0.3 ppm at a distance of
0.0 to 0.4 miles, and 0.02 ppm at a distance of 2.0 to 2.4 miles.
Samples of vacuum cleaner dust declined from 1300 ppm at a distance of
0.0 to 0.4 miles, to 70 ppm at a distance of 2.0 to 2.4 miles [9].
NIOSH has recommended that no worker be exposed to a concentration
of arsenic in excess of 2 yg/m3 of air as determined by a 15-minute
sampling period' [10]. Calculations based on estimates of 8-hour time-
[8] EPA, Office of Air Programs, Research Triangle Park, Helena Valley
Montana Area Environmental Pollution Study, AP-91, Research Tri-
angle Park, N.C., 1972.
[9] Milham, S., and T. Strong, Human arsenic exposure in relation to a
copper smelter, Environ. Res. 7:176-182, 1974.
[10] HEW, NIOSH, Occupational Exposure to Inorganic Arsenic: New Cri-
teria, DHEW Publ. No, (NIOSH) 75-149, NIOSH, Washington, D.C., 1975.
V-10
-------
weighted-average arsenic exposure and actual deaths from respiratory
cancer mortality in exposed workers have shown an apparent dose-response
relationship between arsenic exposure and respiratory cancer mortality
(see Table V-2). Based on a 40-year working life, these calculations
suggest that respiratory cancer mortality was twice that expected at a
dosage equivalent to exposure for 40 years at 3.0 ug As/m3 on an 8-hour
TWA basis, while the observed-expected ratio was 0.6 at the equivalent
to 1.0 U9
Table V-2. RESPIRATORY CANCER DEATHS BY EXPOSURE TO ARSENIC CATEGORY
ARSENIC EXPOSURE
Average
Dosage*
(mq)
3.74
4.84
5.53
6.04
6.68
7.35
8.17
8.78
10.30
Projected
8-Hour
TWA -(yq)
1.0
3.0
6.0
10.0
20.0
40.0
90.0
160.0
740.0
Total
Deaths
(n=173)
26
17
24
22
27
18
13
13
13
RESPIRATORY CANCER DEATHS
Observed
(n=28)
1
2
4
3
3
2
3
5
5
Expected
1.77
1.01
1.38
1.36
1.70
0.97
0.77
0.79
0.72
0/E
0.6
2.0
2.9
2.2
1.8
2.1
3.9
6.3
7.0
Source: Reference [10].
*Total dose based on a 40-year working life.
c. Cadmium
Cadmium is a highly toxic element affecting many systems. It is
apparently absorbed without regard to the bodily level of cadmium, indi-
cating a lack of homeostatic mechanisms for the control of cadmium
levels. Adverse effects include pulmonary emphysema, renal dysfunction
(proteinuria), anemia, and bone and tooth disease. It may also have
carcinogenic [11] and gastrointestinal (including liver) effects. Its
implication in hypertensive disease has been shown through animal experi-
ments and epidemiological evidence; that is, high levels of cadmium have
been found in body tissues of those who died from hypertensive disease.
[11] Sunderman, F. W., Carcinogenic effects of metals, Fed. Proc. 37(1)
40-46, 1978.
V-ll
-------
High serum levels were also found in residents of a steel city suffering
from cardiovascular disease. In a study on the effects of chronic ex-
posure of rats to cadmium, rats received, from time of weaning, drinking
water containing 5 ppm of cadmium. Hypertension began to appear after
about a year and the incidence, which was greater in females, increased
with age. Both sexes had decreased median life spans compared to
cadmium-free controls. In this study, the renal concentration of cad-
mium found in the rat kidney was less than that reported for human kid-
ney. Thus, one form of hypertension in man may be related to the ac-
cumulation of cadmium in the kidney, probably in the renal cortex, where
cadmium is firmly chelated to a zinc- and cadmium-containing protein [8].
Maximum permissible atmospheric concentrations for cadmium have not
been adopted in the United States. In 1963, the National Air Sampling
Network reported that the nation's air contained an average 24-hour con-
centration of 0.002 to 0.037pg/m3 in urban areas and from 0.0004 to
0.026 in nonurban areas. The highest value in the survey was 0.050
ug/m3 in Covington, Kentucky. Standards have, however, been set for
water. According to the Public Health Service Drinking Water Standards,
a drinking water supply containing in excess of 0.01 mg of cadmium per
liter should be rejected.
A summary of the effects of varying levels of cadmium exposure is
given in Table V-3 [12].
d. Lead
Adverse effects of lead include weakness, anemia, neurological and
psychiatric disturbances, and gastrointestinal dysfunctions. It may be
taken in through air, water, or food routes.
The effects of chronic exposure of rats to lead were studied by
giving rats 5 ppm of lead in drinking water from weaning until death.
Lead exerted a continuous adverse effect at all ages and in both sexes
[12] HEW, Center for Disease Control, NIOSH, Criteria for a Recommended
Standard. Occupational Exposure to Cadmium, HEW Publ. No. (NIOSH)
76-192, Washington, D.C., August 1976.
V-12
-------
Table V-3. SUMMARY EFFECTS OF CADMIUM EXPOSURE
SPECIES ROUTE
DOSE or
CONCENTRATION
Occupational Exposure in Humans*
Humana Inhalation
Huaana Inhalation
Humana Inhalation
Human* Inhalation
Humana Inhalation
Humans Inhalation
Humana Inhalation
Humans Inhalation
Humans Inhalation
Humans Inhalation
Humans Inhalation
Humans Inhalation
Humans Inhalation
Humans Inhalation
Humana Inhalation
Humans Inhalation
Humans Inhalation
Humans Inhalation
. Humans Inhalation
Humans Inhalation
2500-2900 mg -min/k1
0.1-2S mg/m3
0.1-24 nq/B*
0.04-19 mg/m3
3-15 mg/m3
0.14 mg/m3
0.6-2.8 mg/m3
0.028-2.8 mg/m3
0.33-1.9 mg/m3
20-700 wg/m3
60-680 ug/m3
80-450 ug/m3
74-210 ug/m3
134 ug/m3
125 wg/m3
Below 100 ug/m3
Below 90 ug/m3
66 ug/m3
31 ug/m3
16-29 ug/m3
Exposure in Animal Studies
Rats po> water
Rats po, diet
Rabbits sc
Rats, mice po, water
Rats po, water
Rats ip
Rats sc
Rats po, water
Nice sc
Hamsters iv
Rats ip
Rats Inhalation
50 ppm
45-135 ppm
650 pg/kg
5 ppm
0.2-200 ppm
1 mg/kg
850 ug/kg
5 ppm
630 ugAg
880 ug/kg
1.8 mg/kg
2.8 mg/m3
EXPOSURE
Onoe
Years
Years
Years
Years
Years
Years
Years
Years
Years.l hr/day
Years
Years
Years
1-20 years
Years
Years
Years
21-40 years
1-12 years
Years
3 months
6 months
10 weeks
6-12 weeks
45 days
Once
Up to 4 years
Once
Once
Once
7 months
COMMENTS
ratal pulmonary edema
Rickets and dental problems in
offspring of workers
General complaints, cnronaxy changes
Slight anemia, yellow fringe on teeth
Anosmia, fatigue, renal dysfunction.
hepatic dysfunction, emphysema.
yellow fringe on teeth
Emphysema, proteinuria
Anosmia, proteinuria
Anosmia, proteinuria
Proteinuria, aminoaciduria
Proteinuria
Gastrointestinal and respiratory
symptoms, cartoue teeth
Emphysema, proteinuria
Reduced pulmonary function, proteinuria
Proteinuria
Anemia, proteinuria
Questionable effects
Reduced pulmonary function, proteinuria
Reduced pulmonary function, proteinuria
No effects
Improvement in conditions of workers
exposed at 125 wg/m3, no effects in
new workers
Anemia
Anemia, bleaching of incisor teeth
Anemia, proteinuria
Hypertension after 1 year
Vascular changes in kidneys
Increased adrenal activity
Testicular necrosis
No significant increase in tumors
Petal abnormalities
Fetal abnormalities
Petal abnormalities
Altered estrus, decreased weights of
offspring
•Some investigators noted deaths of cadmium workers
necessarily attributing them to expos-ire
Source: Reference (12).
from various causes, including cancer, without
V-13
-------
as evidenced by reduced life spans and increased mortality rates. Con-
centration's of lead in the rat organs were similar to those reported for
man [8].
Lead can be seriously injurious to health as a result of accumula-
tions in the body. Long-term daily intake of less than 0.6 mg by
healthy adults may cause small increases in body burden, but no clinical
.disease. An intake in excess of 0.6 mg per day may result in the accumu-
lation during a lifetime of a dangerous quantity of lead in the body.
It has been shown that the lead content of dustfall, and conse-
quently of soil, vegetation, and outdoor dust, decreases exponentially
with distance from smelters and that children living in contaminated
areas absorb excessive amounts of lead [13]. The National Air Sampling
Network reported a range of 0.1 to 2.3 yg/nr of lead in urban areas, and
from 0.002 to 0.15 yg/m3 in nonurban areas. The highest city in the
1964 NASN survey was Minneapolis with 2.7 yg/m3. Maximum permissible
atmospheric concentrations for lead have not yet been established in the
United States.
The lead concentration in drinking water supplies ranged from traces
to 0.04 mg per liter, averaging 0.01 mg per liter. At concentrations of
0.1 mg per liter, bacterial decomposition of organic matter is inhibited
and some fish are susceptible to lead poisoning. Adults consume 1 to 3
liters of drinking water per day. The Public Health Service Drinking
f.
Water Standards state that 0.05 mg of lead per liter constitutes grounds
for rejection of the water supply.
[13] Parkinson, D. K., Lead contamination around secondary smelters:
Estimation of dispersal and accumulation by humans, Science 186:
1120-1123, 1974.
V-14
-------
e. Fluoride*
The role of fluoride in the production of cancer is controversial
[14]., A small amount of epidemiological and toxicological research
points to its possible involvement. Experiments with mice have shown
that fluoride stimulates cancer growth at concentrations as low as 1 ppm,
whether taken in water or by subdermal injection or even in in vitro
suspensions. Also, when fluoride is incorporated into the carcinogen
dimethylaminobenzene, its cancer-producing ability is greatly enhanced.
High cancer rates in workers exposed to fluoride have been reported,
although the results are not clear-cut because radiation and carcinogen
exposure were also present. In a steel city in Ontario, correlations
were found between ambient fluoride levels, increased cancer rates, and
skeletal fluorosis [15]. Pollution data indicate that standard maximum
levels (5 ppb) for fluoride were exceeded in the early 1970's in the
Canadian steel city. Levels in home grown food were 17 to 130 times the
maximum permissible level. Atmospheric levels of fluoride of up to 1180
mg F/100 cm2/30 days were observed in the Ontario steel city in 1972, a
value 30 times higher than for the rest of Ontario.
* Although not an emission from copper smelters, fluoride is dis-
cussed here because one of the areas studied has a problem with
fluoride contamination as a result of another pollution source.
[14] Editorial: Can fluoride cause cancer?, Fluoride 5(4):169-170,
October 1972.
[15] Cecilioni, V. A., Further observations on cancer in a steel city,
Fluoride 7 (3):153-165, July 1974.
V-15
-------
4.. Description of the Study Sites
Eight sites involving nine copper smelters were studied:
• Ajo, .Arizona (Pima County)
—Phelps-Dodge Corporation
• Morenci, Arizona (Greenlee County)
—Phelps-Dodge Corporation
• Douglas, Arizona (Cochise County)
-—Phelps-Dodge Corporation
• Hayden, Arizona (Gila County)
—American Smelting and Refining Company
— Kennecott Copper Corporation
• San Manuel, Arizona (Final County)
—Magma Copper Company
• Miami-Inspiration, Arizona (Gila County)
— Inspiration Consolidated Copper Company
• Anaconda, Montana (Deer badge. County)
—Anaconda Company
• Magna, Utah (Salt Lake County)
— Kennecott Copper Corporation
Maps of these areas are given in Appendix D.
AjOf established in 1864, is located in southern Arizona, west of
Tuscon. A large percentage of the labor force is employed in the mining
and smelting operations of the New Cornelia Mine owned by the Phelps-
Dodge Corporation directly east-southeast of town. Other occupations
are largely in service and commercial trade. From 1960 to 1970, the
population decreased from 7,049 to 5,881, or -16%.
Clifton-Morenci, on the eastern border of the state, was established
in the 1880"s and today is dominated by the copper smelter owned by
Phelps-Dodge. Trade and service occupations are also represented as
well as governmental activities in Clifton, the county seat. From 1960
to 1970, the population increased from 7,248 to 8,140, a 12% change.
Douglas is the largest city in Cochise County and is situated on the
U.S.-Mexico border. The principal occupations are: mining and smelting,
apparel manufacture, government employment, cattle transport, and agri-
culture. The smelter is located about 2.5 miles west of town.
V-16
-------
Hayden is located one mile north of the junction of the Gila and San
Pedro Rivers at the southern end of the county. Over 1,400 workers are
employed in the town's two copper smelters. A small amount of retail
trade exists and agriculture is not prominant. The population has de-
clined 27% from 1960 to 1970, going from 1,760 to 1,283.
San Manuel, a recently established town, started in the 1950's. The
population has been stable, changing only 2% from 1960 to 1970 (1,913 to
1,953). The mining and smelting facilities of the Magma Copper Company
employ 3,000 people. The area also supports agriculture, cattle ranches,
and guest farms.
Miami,* established in the 1900's, is located in the Final Mountains
of east central Arizona, between the walls of a steep canyon. Inspira-
tion consolidated Copper Company, Miami Copper Company, and Ranchers
Exploration and Development operate open-pit and underground copper
mines and leaching operations. Inspiration also operates a smelter,
refinery, and rod mill. The copper industry in the Miami area employs
3,500 workers. Other industries are a lime quarry, cattle ranching, and
a sawmill. The area was for many years a producer of high-grade as-
bestos which was mined in the area from 1913 on. In the early days
most material was hand-processed, but from the mid-1940's to the mid-
1970 's several mills were operated in the town of Globe, about 7 miles
southeast of Miami. All plants except one closed by 1974 as a result
of federal safety regulations.
A detailed history of the demography and industrial activity in Gila
and Final Counties is given in Appendix E.
Anaconda is in the Deer Lodge Valley of Montana. The valley is 35
miles long and 8 to 10 miles wide, sloping in a northerly direction from
Anaconda at an elevation of a little over 5,000 feet to Garrison at an
elevation of approximately 4,500 feet. The smelter at Anaconda employs
about 1,700 men and processes copper and zinc ore mined in Butte,
* A smelter was in operation from 1924 to 1971 in Superior, 10 miles
northwest of Miami.
V-17
-------
approximately 27 miles east. The majority of the population is engaged
either directly or indirectly with the copper industry. The smelter is
two miles east-southeast from the center of town.
Magna, Utah, is located on the floor of the Salt Lake Valley, 4,220
feet above sea level and about five miles from the smelter. It is be-
tween the Wasatch Mountains to the east and the Oquirrh Mountains to the
west. • ;
The major processes and emissions from the copper smelting industry
were outlined in Section B-2, and a detailed description of the industry
is given in Appendix B. In Table V-4 we list the features specific to
the sites under study [16].
[16] EPA, Office of Air and Waste Management, Office of Air Quality
Planning and Standards, Background Information for New Source
Performance Standards: Primary Copper, Zinc, and Lead Smelters,
Volume I: Proposed Standards, EPA-450/2-74-002a, Research Triangle
Park, N.C., October 1974.
V-18
-------
Table V-4. DESCRIPTIONS OF SELECTED COPPER SMELTERS*
CONPMR
mm/
ASARCO/
KagH/
tan Manual,
XUona
Phelfo-oedge/
Dooglaa.
Arlaona
Phelpe-oodgo/
Noiancl,
Phalpa-oodgo/
Ajo, Arlaona
Inaplratlon/
Miaul,
Wliona
Kannaoott/
Haydan.
Arltona
Reiwecott/
Nagna. Utah
Anaconda/
Anaconda,
Montana
AOB OP rum
riatr
1911
1996
1910
1941
1990
1919
1907
1*06
LAST
NOOIM-
CATION,
1971
(Haw eon
vartat
aold
plant)
1971
(New EEP
1964
(Roaatar
aold
plant)
1971
(DMA •
acid
plant.)
1971
(Blaot.
furn. ,
avph-
conv.t
acid pit
undar
eonat.)
1968
(riuld
bad
roaatar
and aold
plant)
196*
(Ranovad
roaatar a
and con-
vartad
to graanr
faad
reverb*)
(Haw acl<
plant)
MATERIALS
COMC.
ft/dj
1,000
1.700
1,160
1.111
no
•40
1,050
1.100
1,710
lUSttR
corm
166
110
lAnoda
Co)
MS
(Anoda
CO)
470
{Anoda
lAnoda
Co)
300
110
790
900
lAnoda
Cul
PRODUCTION EOUirMEl
ROASTERS
To?
OUR/
tTEf
Nona
17/7
1/0
Nona
Nona
1/0
Nona
Nona
OM S1WIN
carraoL
EQUIPMENT
Roaatar gaaaa
Join ravarb
gaaaa prior
to traabaant.
Roaatar gaaaa
traatad fay
ESP and than
Join ravarb
gaaaa.
All roaatar
gaaaa to
Paraona 790
t/d acid
plant.
All roaatar
gaaaa pass
thru oyolonaa
• ISP (90%),
thin Join Sot
of Convartar
gaaaa a paaa
thru 196*
tonaanto 750
r/4 aoM
Plant.
Roaatar a
nlat, but
for drying
eoncantrata
only.
"«JV
1 ol
RTBY
1/0
1/0
1/0
4/0
1/0
1/0
1/0
1/0
J/l
IMC rVRHACES
GAS STREAM
COMTROI.
EQUIPMENT
toverb gaa thru
HIBa and water
•pray chanbar,
than Jolnad
»lth roaatar
gaaaa to ESP
(M.1%) and
100-lt atack.
Ravarb gaaaa
traatad In ma
ind ESP (S9«).
than out 919-ft
Hack.
All ravarb
geeea traatad
In NIB. than
Join roaatar
aaeee and paaa
out S44-ft
Mart.
Ravarta gaa
traatad In two
KS'a In paral-
lal (7t.9«).
Ravarb gaaaa
Join Convartar
gaaaa, than to
1ST and 160-ft
atack.
Ravarb gaaaa
paaa thru NRBa,
flua, and 17S-
ft .tack
Ravarb gaaaa
traatad by ma.
alaad with SOI
of convartax
gaaaa, paaaad
thru ESP (991)
and 600-ft
atack.
Ravarb gaaaa
traatad by MB
and ESP (SOT),
then out two
410-ft atacka.
Ravarb gaaaa
paaa thru watar
apray chanter
ard flua, then
oln converter
laaaa.
Fof
stay
4/1
1/1
B/O
1/0
1/1
7/1
XHIVSRTEM
CAS OTREAH
COMTROI.
BUU^PMIHT
All eonvartai
gaaaa to 197O
Ruat Engineer-
ing 7*0 t/d
acid plant.
Converter gaeen
paaaad thru ESP
I95«l , then to
590-ft etack.
Convartar gaaea
traatad In BSP.
than paaaad out
561-ft atack.
Converter gaaaa
traatad In ESP
(99. 5V.
Converter gaaaa
Join reverb
gaaaa, then to
P.SP and 360-tt
•tack.
Converter gaaaa
treated by bal-
loon flua to
200-ft atack
Convartar gaaaa
pass thru bal-
loon flue, them
S0« of gaaaa
•Uad with ra-
varb gaaaa, and
S0« of gaaea
traatad In ESP
(9S«) and addad
to roaatar
gaaaa
All converter
gaeak enter ESP
I60-7M), then
97. 5» of gaaea
traatad In five
acid planta
(1.400 t/d
•,•0,1, and CM
flva acid plant
atacke.
Convartar gaaaa
Join reverb
gaaaa, paaa
thru aettllng
flue and ESP
(16. 9«), then
out atack.
NUMBER
1
Iroaat.
furnace
1
Itall
gaa)
(furn.)
Iconv.)
1
Iroaat.
turn.)
1
Iconv.)
1 Itall
nee,
(urn.)
1
(conv.)
1
(furn.,
conv.)
1
(turn.)
1
Iconv. 1
1
(tall
gaa)
1
(turn.,
conv.)
1
(furn.)
S
(tall
gaa)
1
(turn.,
conv.)
STACK DATA
IWT
Kll
100
10
SIS
550
544
556
600
600
360
179
100
100
600
406
95,
»9,
S7,
Bl,
126
915
riM RATB
(eclp)
401.000
at, loo
179,000
117.000
470.000
164,000
110,000
406.000
197.000
160.BOO
161.400
7S.OOO
311. OOO
1.1S7.0DO
259.000
1 .950.000
SO]
COHC.
2,101)
15,000
54,000
15,700
15.200
14,000
15.400
7,000
3.900
19,100
1.300
7,190
1.500
l.OOO
3.300
PART1C.
(t/d)
1.5
0
1.79
0.54
39
9
14.7
o.e
0.7
Qbk.
get.
0
1.1
4.13
0.46
(acid
Plaal
»*•*»•.•
float
11.9
so]
It/dl
390
11. B
111
767
902
492
SK.
761
3BO
7B
ISt
u
117
11)
12
790
'four?*i taf«r*nc«
V-19
-------
C. APPROACH
1. General Methodology
a. Analysis of Mortality Data
Data interpretation was based on the results of the mortality
analyses. We began with the objective of determining whether or not
there was an excess of mortalities in the study areas. This involved
establishing criteria to screen rates and designate them as elevated or
not elevated. Because of the small populations involved in many of the
towns of interest, we did not look at mortality rates alone. Since mor-
tality rates are standardized to a population base of 100,000, it is
possible for a particular rate to be misleadingly high when, in fact,
6nly a very few deaths (e.g., two or three) are involved. We therefore
identified disease categories having elevated rates by a combination of
several measures. We considered a rate to be elevated if it met the
following criteria:
• higher than the national average and one or more comparison
areas;
• standardized mortality ratio (SMR) of over 100; and
• p <0.05 as determined by the Mantel-Haenszel test of sig-
nificance [17]} that is, relative risk >1 and x2 >3.85.
The rates analyzed were those of white females.* We based our con-
clusions on female rates because, in the towns studied, a large propor-
tion of the males are workers at the smelter. Smelter workers are ex-
posed, for a fourth of the time, to a situation where the pollution con-
ditions are very different from those outside the plant. Males, there-
fore, represent a mixed-exposure situation and were not considered in
this approach.
[17] Mantel, N., and W. Haenszel, Statistical aspects of the analysis of
data from retrospective studies of disease, J. Natl. Cancer Inst.
22(4):719-748, 1959.
* The demographic category "white" may include those of Hispanic
stock. To separate those of Hispanic origin would require evalua-
tion of country of origin and/or surname data from death certifi-
cates; this information is not coded on the mortality tapes pro-
vided by the states.
V-20
-------
Furthermore, even in towns where, the majority of males are employed
in the smelting operation, they do not represent as homogeneous a study
population as do females. It is frequently the case that men employed
in the copper industry have had previous employment histories in other
types of industry with potential exposure to hazardous substances. This
history would represent a potential confounding factor in the study and
could not be known without conducting a case study of individual resi-
dents.
Although the original regression analysis identified a group of
seven diseases, we obtained and analyzed the rates for some 34 disease
groupings. We felt that an analysis of all rates would enable us to
verify the results of the original regression and help in an evaluation
of its usefulness. This did not entail any extra cost or effort in the
acquisition of data, since the rates for all diseases for a particular
area were included on the same tape. It did, of course, entail more
effort in the actual analysis and etiology research.
, The original plan submitted for this program called for the inves-
tigation of only two sites—one for hypothesis development and the other
for hypothesis testing. In the copper smelting study, we wanted to
examine the rates for as many sites as possible to permit replication,
and as a compensation for the small populations frequently found in com-
pany towns. It was ambitious to evaluate eight sites in a time frame
originally designed for two sites. Consequently, .it was necessary to
sacrifice depth for breadth in certain aspects of the study. We felt at
this stage that it was more important to determine if a number of copper
smelter towns had elevated mortalities for the same disease rather than
to scrutinize one town in great depth. After a beginning of this type,
we envisaged that the next step in the long-range plan would be to study
each town in detail.
Because of the small populations and few deaths, it was necessary
in some parts of the analysis to aggregate disease groups and towns in
order to have an adequate number of data points. Therefore, in Arizona
the only town large enough to analyze separately was Douglas. First,
the rates for individual diseases for Douglas were analyzed separately.
V-21
-------
Second, the number of categories was decreased in order to produce more
data points per category. Thus, the original 34 disease groups were
collapsed into 6 and the deaths from all 6 towns were combined in one
group. The aggregated data were screened by comparing deaths per popu-
lation for the smelter town aggregate to the same ratio for a control
area. This method is crude and not adjusted for age but is believed to
be suitable to indicate broad trends.
b. Comparison Areas
The original intention was to select as many control sites as fea-
sible, so that conclusions would not be based on comparison to only one
other figure. Comparison areas were chosen which did not have smelting
or another industry that is a major source of pollution, and which were
reasonably similar to the study areas in terms of socioeconomic vari-
ables.
For Anaconda, five areas were identified: Billings, Bozeman, Havre,
Livingstone, and Great Falls. For Magna, three were chosen: Logan,
Cedar City, and Price. Conversations with local health and pollution
officials in Montana and Utah confirmed the suitability of these areas.
In Arizona, the selection of control areas was more of a problem.
Much of the area is dominated by nonwhite populations from Indian reser-
vations. Furthermore, towns with significant white populations often
were found to differ from study areas in socioeconomic characteristics,
such as education, median age, and income. Other areas were primarily
retirement areas or.college towns. There are wide differences in cli-
mate throughout the state, which may have an effect on certain health
problems (such as respiratory diseases) or air quality (such as the
amount of particulates). Other issues were the presence of other pol-
luting industries such as coal-fired power plants. It was finally
decided to use the State of Arizona as a comparison area.* In this way,
it was felt that local differences would average out.
The summary demographic statistics for the study sites and compari-
son areas are given in Table V-5.
* Actual data used were for the State of Arizona minus Gila and Final
Counties.
V-22
-------
Table V-5
SUMMARY DEMOGRAPHIC STATISTICS: ARIZONA SMELTER TOWNS AND CONTROL AREA;
ANACONDA AND CONTROLS; MAGMA AND CONTROLS
STATE
I
1
|
1
1
DBfOGRAPHIC AREA
AJo (S)
Douglas (S)
Morenci (S)
Bayden (S)
San Manuel (S)
Hind (S)
Aggregated smelter towns
Arizona1 (C)
Anaconda (S)
Billings (c)
Bozenan (C)
Great Falls (C)
Havre (C)
Livingston (C)
Magna (S)
Cedar City (C)
Kearns (C)
Logan (C)
Price (C)
DQtOGRAPHIC VAPT&niJts
Median
Age
29.5
23.0
29.5
21.0
18.0
23.0
23.0
29.5
29.5
29.5
21.0
29.5
23.0
39.5
21.0
21.0
16.0
23.0
29.5
Median
Family
Income
9.500
7,500 .
9,500
9,500
11,000
7,500
8,500
9,500
8,500
9,500
8,500
9,500
9,500
7,500
9,500
8.500
9,500
7,500
8,500
% in
Same House
as in 1965
41.8
57.5
51.1
70.3
32.2
63.1
52.2
41.0
59.4
48.3
29.5
48.8
48.2
52.5
66.0
46.4
64.3
39.9
53.7
%
White
89.9
98.4
97.2
100.0
97.9
99.1
96.7
91.4
98.5
98.0
99.2
97.2
96.8
99.6
99.4
98.6
99.6
97.9
98.7
% With
Less Than
1 Person
per Room
83.4
83.5
87.5
71.1
77.3
83.8
82.4
88.1
92.1
94.4
95.2
92.7
91.6
94.7
84.3
90.6
77.0
93.9
92.4
%
Completing
High
School
49.5
43.3
52.0
41.9
63.2
41.8
47.1
58.9
53.2
68.2
77.2
66.1
66.1
53.2
56.4
78.4
53.9
80.4
57.5
%
Female
51.0
52.7
51.2
50.6
48.7
51.1
51.4
50.9
51.8
52.0
49.2
51.6
50.1
52.1
50.4
50.0
50.1
50.2
51.0
% Native-
Born,
Native
Parents
72.3
43.3
83.6
68.6
85.9
71.3
61.1
83.2
72.5
82.6
87.1
81.2
80.6
78.3
88.9
92.1
92.4
86.3
79.1
KJ
Ul
(S) denotes towns with smelter.
(C) denotes control town or area
*Data used was for Arizona minus Gila and Final Counties.
-------
2. Data Integration
After identifying elevated, mortalities by the method described above,
the next step was to determine if the effect could be explained by pol-
lution patterns. The question was: Does the excess mortality occur in
areas which have pollution of a type implicated in the etiology of the
disease?
.The nature of the pollution data available had an obvious and im-
portant effect upon the type of analysis that could be undertaken.
Several problems with the available pollution data were recognized.
•There is a lack of information on exactly how ambient environmental
measurements relate to exposure of individuals. This is a result of
(1) unanswered questions on :the absorption, metabolism, and excretion
of low doses of pollutants over a long period and (2) lack of detailed
knowledge on the fate, transformation, interaction, and dispersal of
pollutants in the atmosphere. Further complicating the issue is the
influence of personal lifestyle,such as how long the individual has been
exposed, hi.s general health, diet, smoking habits, and level of health
care. These gaps in the present knowledge are reflected in the fact
that, for many of the pollutants released by the smelting industry, no
atmospheric standards exist.
Another major drawback is the general lack of pollution data on
control areas. In a rigorous analysis, one would want to compare mor-
tality and pollution patterns in a study area to the same data in a
control area. Unfortunately, much of the detailed pollution data which
is presently available is from large urban areas or unpopulated ex-
panses, both of which are unsuitable as controls.
3. Data Collection and Data Bases
For each of the sites examined, it was necessary to collect health,
pollution, industry, and demographic information. This was a major
effort and required (1) identification of the sources of the data;
(2) consultations with officials to determine the cost, availability,
and quality of the information; and (3) acquisition of the data, either
as hard copy, magnetic tape, or by manual abstracting on site. In some
V-24
-------
cases there were gaps or redundancies among the various sources; we
often found that measurements for a particular time, location, or pollu-
tant, which would have been useful in the analysis, simply did not exist.
The following subsections summarize the major data sources compiled
for the study.
a. Mortality Data Bases
ARIZONA 1968-1975
YEAR
1968
1969
1970
1971
1973
1974
1975
COUNTIES
All
Arizona
Counties
CAUSES
OF
DEATH
All
PLACE OF
RESIDENCE
OF DECEASED
Yes
LENGTH OF
RESIDENCE
ARIZONA
Yes
TOWN
Yes
No
1
i
Note: The core of this data base was acquired on tape from
the Arizona Department of Health Services (ADHS). Note that
Arizona listed length of residence in the town on death cer-
tificates for only three years, although length of residence
in the state is listed for all years. ,
GILA & FINAL COUNTIES 1957-1975
CAUSES
YEAR COUNTIES OF
DEATH
1957 Gila & All
1960 Final
1963
1966
1969
1972
1975 \
PLACE OF
RESIDENCE
OF DECEASED
Yes
-
LENGTH OF
RESIDENCE OCCUPATION
ARIZONA
Yes
-
TOWN
No Yes
•
•
Note: These data were obtained as hard copy from the ADHS. All
data contained on the certificates was coded and punched in-house
and put on tape. If the case of 1969-1975, it was necessary to
code and punch occupational information only since the other
variables are available on the ADHS tape.
V-25
-------
GILA & FINAL COUNTIES 1959-1967
CAUSES
YEAR COUNTIES OF
DEATH
1959 Gila & Digestive
1961 Final Diseases
1962
1964
1965
1967 -1
•
PLACE OF
RESIDENCE
OF DECEASED
Yes
•
LENGTH OF
'RESIDENCE OCCUPATION
ARIZONA
Yes
TOWN
No Yes
1
Note: These data were available as hard copy. Since it involved
only 144 certificates, it was not put on tape and could easily be
manipulated by hand if the preliminary analysis indicated the need
to examine digestive mortalities from this early time period.
MONTANA 1968-1975
CAUSES PLACE OF
YEAR COUNTIES OF RESIDENCE
DEATH OF DECEASED
LENGTH OF
RESIDENCE
MONTANA
TOWN
1968
1969
1974
1975
All
Montana
Counties
All
Yes
No
Mo
UTAH 1965-1975
YEAR
COUNTIES
CAUSES
OF
DEATH
PLACE OF
RESIDENCE
OF DECEASED
LENGTH OF
RESIDENCE
UTAH
TOWN
1965
1966
• • • •
• • • •
1974
1975
All All No No No
Utah
Counties
I 1
-
V-26
-------
b. Demographic Data Bases
Census data were used in the selection of comparison areas and in
the determination of denominators for mortality rates. These data are
from the fifth count of the 1970 U.S. Census of Population and were made
available on computer tape by the Bureau of the Census. The data are
grouped by enumeration districts, which are administrative divisions
devised by the Bureau of the Census that respect boundaries of towns and
contain a population of about 800 each. The vast majority of the data
points used are based on a 15% sample and are subject to some chance
variability due to sampling. Nevertheless, the variability can be es-
timated , and its importance is lessened by the aggregation of several
enumeration districts into exposed and unexposed areas. Census tapes
were obtained for the states of Arizona, Montana, and Utah.
In addition to the Census tapes, other demographic and background
information was obtained from local chambers of commerce, banks, and
other public affairs offices. Conversations with local individuals
during site visits were also useful.
c. Air Pollution Data Bases
• Arizona
— Arizona Air Quality Reports 1969-1975
— Estimated Impact of Smelter Operations on SC-2 Concen-
trations in the Rocky Mountain Area
—SAROAD and NEDS data
/
• Montana
— Helena Valley Area Environmental Pollution Study (EPA)
—Montana Air Quality Bureau Reports
— CHESS study (EPA)
— SAROAD and NEDS data
• Utah
— Utah Ambient Air Pollution Reports
— A Summary of Air Pollution Source Emission Calculations
for Utah
— SAROAD and NEDS data
—CHESS study (EPA)
V-27
-------
d. Industrial Data
Data on the industry were obtained through standard textbooks, con-
versations with officials, brochures from the companies, and from re-
ports prepared by EPA.
e. Morbidity Data
Due to the cost and effort involved in the acquisition of hospital
morbidity data, it was used in only one of the studies in this program
(bituminous coal mining). However, during our visit to Arizona, we
interviewed the administrators of three hospitals* in order to obtain
an overview of the community health problems and find out if there were
any readily available reports or records that might be useful to our
stbdy.
Two of the hospitals visited were owned and operated by the industry.
They provided useful background information, but said that they could
not release any more specific information until the case for involvement
of the smelters in community disease was more clear-cut. They referred
to a very valid point: the finding of a high mortality rate in a county
with a large industrial operation sometimes leads to a hasty indictment
of the industry before all the facts are in. Since the company towns
constitute only a small fraction of the total population of a county,
it must first be established whether the deaths in the town itself are
excessive.
Apart from the question of hospital cooperation, it was determined
that any collection of morbidity data from these hospitals would be a
difficult job. Records are not indexed by town of residence of the pa-
tient, so the name file would have to be examined card by card to iden-
tify patients from the towns of interest. Retrieval by residence is
theoretically more feasible for those hospitals which subscribe to auto-
mated records systems such as the Professional Activity Study (PAS) and
the Hospital Utilization Project (HUP), since the file can be searched
on zip code. Unfortunately for retrospective studies such as ours,
these systems are relatively new and the inclusion of zip code patient
identifiers has only been in practice for the last few years.
*Miami-Inspiration Hospital; Kennecott Copper Corporation Hospital;
Pi Dal General Hospital.
V-28
-------
D: RESULTS
The mortality rates for 34 disease groups for white females in the
smelter towns were analyzed as described in Section C of this chapter.
The rates for the diseases which were elevated in at least one of the
three towns of Douglas, Anaconda, and Magna are shown in Table V-6. The
supporting data for this set (SMRs, relative risk, x2 values, and con-
trol data) are given in Tables F-l through F-3 in Appendix F. The mor-
tality rates for all other diseases for these towns are included for
reference in Appendix F, Tables F-15 through F-17.
Table V-6
ELEVATED MORTALITY RATES IN'DOUGLAS, ANACONDA, AND MAGNA
EC I
»
5
7
10
12
14
18
19
20
27
31
CAUSE Of DEATH
DISEASE CATEGORIES'
Cancer of tha colon S rectum
Cancer of tha pancreas
Cancer of the genitourinary
organs except bladder
Other cancers except
leukemia
Blood, endocrine,
nutritional ft
metabolic> diseases
Cardiovascular diseases
Cerebrovascular diseases
Acute respiratory diseases
Diseases of liver, gall-
bladder s pancreas
Diseases of the skin c
musculoskeletal system
& ill-defined diseases
AGE-ADJUSTED MORTALITY RATES*
Douglas ,
Arizona
33
S
63.
110
57
382
79
6£
SI
60.
Anaconda ,
Montana
56
39_
45
98
24
443
133
45
56
21
Magna,
Utah
33
0
54
131
77.
379
70
11
33
57
National
Averages
N.A.
12
45
90
• 34
384
84
26
N.A.
N.A.
'The specific diseases involved in these categories Jure given in Appendix
G.
^Number of deaths per 100,000; white females; ages 35-74: 1968-1975 (SMRs,
relative risks, and X* are given in Appendix F). The age-adjusted mor-
tality rate was calculated for deaths occurring in the 35-74 age group.
Underlined rates are significantly higher than at least one control area
at, at least, the p < 0.05 level.
N.A. • National averages not available for these disease categories.
V-29
-------
From Table V-6 we can make the following observations:
• No one disease is significantly elevated in all three smelter
towns, although two diseases are elevated in two of the towns.
Disease 12 (cancers of the glands and lymph nodes, secondary
neoplasms, and benign neoplasms) are high in Douglas and
Anaconda. Also elevated in Douglas and Anaconda is disease
27 (diseases of liver, gallbladder, and pancreas).
• Colon, rectum, and pancreas cancer are high in Anaconda.
• Genitourinary cancer, acute respiratory diseases, and skin/
musculoskeletal disorders are high in Douglas.
• Cardiovascular diseases and metabolic/nutritional disorders
are high in Magna.
Table V-7 shows the results of the analysis of combined disease
groupings for the aggregate of six Arizona smelter towns; the actual
number of deaths involved in these groupings is shown in Table F-4 in
Appendix F. The mortality rates for the individual towns included in
the aggregate are given in Appendix F, Tables F-15 through F-17.
Table V-7
MORTALITY RATIOS FOR ARIZONA SMELTER TOWN AGGREGATE
CAUSE OF DEATH (ECI ft)
Organ cancers (2-11)
Other cancers
except leukemia (12)
Neoplasms of the
lymphatic* (13)
Vascular diseases (17-19)
Respiratory diseases (20, 21)
Digestive diseases (22-27)
Genitourinary diseases (28, 29)
MORTALITY* RATIOS
Smelter
Towns*
.0053
.0036
.0015
.0307
.0050
.0037
.00125
Control
Area0
.0051
.0016
.0008
.0333
.00049
.001
.0007
Smelter Town Ratio
Control Area Ratio
1.03
2.26
1.88
0.92
10
3.76
1.79
*White females, all ages, 1968-1975.
h.. ,, . number of HF deaths in smelter towns
"Mortality ratio
Mortality ratio
sex-, race-specific population in smelter towns
number of WF deaths in control area
sex-, race-specific population in control area
From Table V-7 we see that, in the smelter town aggregate, acute
respiratory diseases show the greatest elevation.
V-30
-------
Our next step, after identifying the diseases which are elevated in
the smelter areas, was to attempt to explain the excess mortality in
terms of environmental pollution. The following sections describe the
three disease-pollution hypotheses which resulted from the integration
of environmental and mortality data. No environmental explanation for
the remainder of the elevated diseases was found.
1. Sulfur Dioxide and Particulates
Data on the pre-emission control levels of five major smelter pol-
lutants (sulfur dioxide, particulates, cadmium, lead, and arsenic) for
the eight sites were collected and analyzed. Table V-8 summarizes the
findings for sulfur dioxide and particulates for the sites. This table
represents the condensation of a great deal of data 'from several sources.
Tables F-5 through F-9 and Figure F-l in Appendix F tabulate in detail
the figures from which this summary was derived.
Table V-8
SUMMARY OF SULFUR DIOXIDE AND PARTICULATE POLLUTION
FOR ARIZONA, UTAH, AND MONTANA*
AREA
Arizona
Aggregate
Magna,
Utah
Anaconda ,
Montana0
TIME
PERIOD
1969-1971
1969-1971
1965-1966
1961-1962
SULFUR DIOXIDE3 (ug/ra3)
Range
107-151
61-103
40-91
Avq ±0
135 i 24
83 t 20
60 t 17
Category
High
Medium
Low
PARTICULATES* (yg/m3)
Range
111-141
55-70
10-339
Avq ±o
122 t 16
63 + 7
89d
Category
High
Low
Medium
"Detailed data for this summary is given in Appendix F, Tables F-5 through
F-9 and Figure F-l.
aAnnual standard: 50 ug/m3. ^Annual standard: 70 ug/m3.
"Data for 1969-1971 not available. Data for the three three areas is com-
parable, insofar as it is all from the pre-eoission control period.
^Standard deviation not available from data source.
Based on the available data and the existing standards for sulfur
dioxide and particulates, the Arizona, Utah, and Montana sites fall into
different exposure categories. The Arizona sites are high in both sul-
fur dioxide and particulates; Magna has an intermediate level of expo-
V-31
-------
sure from'sulfur dioxide and is low in particulates. Anaconda is inter-
mediate in particulates and low in sulfur dioxide. Table V-9 shows in
greater detail the lengths of time that sulfur dioxide and particulates
in Arizona'were in excess. The range for sulfur dioxide was from 15 to
193 days per year in excess of the state's daily standard of 250 yg/m3.
Douglas exceeded the standard for 138 days. Ajo, Claypool, Douglas, and
Hayden exceeded both the annual and the daily standard for particulates;
Douglas and Hayden were by far the highest. Figure V-3 indicates that
both Douglas and Hayden exceeded the maximum daily standard of 100 yg/m3
96% of the time in 1969; Ajo slightly exceeded the annual standard of
70 yg/m3 in 1969.
!
Table V-9
DETAILED SULFUR DIOXIDE MONITORING FOR ARIZONA SITES (\ig/m3)
Monitoring Period,
Percent Data Recovery
Maximum Hourly Average
Percent of Hours > 850 ug/m3
(State Hourly Standard)
Maximum Daily Average
_
Percent of Days > 250 vg/m3
(State Dally Standard)
Annual Average
AJO
3/19-4/21
7/7-12/31
77
3,928
1.5
1,105
5.0
-------
l'T I 111 I l~T I I I I II I I I
Figure V-3. Arizona: Percent of Time Concentrations of Particulates
Less Than Stated Value in 1969 (from Arizona Department of. Health,
Division of Air Pollution Control, September 1970)
V-33
-------
Table V-10
SUMMARY DATA FOR ACUTE RESPIRATORY DISEASES HYPOTHESIS
AREA
ARIZONA
SHELTER
AGGREGATE
DOUGLAS
ANACONDA
MAGNA
MORTALITY RATE FROM
ACUTE RESPIRATORY
DISEASE: WHITE FEMALES
EXCESSIVE
10-fold excess over
control area.
EXCESSIVE
60«
NOT EXCESSIVE
45*
NOT EXCESSIVE
11*
SULFUR DIOXIDE EXPOSURE
HIGH
Annual average is 135
ug/m3j standard is 50.
S02 levels > 250 ug/m3
for 15 to 193 day* in
1969.
HIGH
SO2 levels > 2SO ug/m3
for 138 days in 1969.
LOW
Annual average: 60.
MEDIUM
Annual average: 83.
PARTICULATES EXPOSURE
HIGH
Annual average is
122 ug/m3 ; standard
is 70.
HIGH
Exceeded maximum
daily standard of 100
ug/m3 96\ of the time
in 1969.
MEDIUM
Annual average: 89.
LOW
Annual average: 63.
Significantly higher than controls at p < 0.05.
*Not significantly different from controls.
PHYSIOLOGICAL DATA:
• Particulates and sulfur dioxide in combination have deleterious effects upon
the respiratory system.
• Levels of sulfur dioxide > 240 ug/m3 for more than one month are deleterious.
• Particulate levels of 70 ug/m3 (annual) or 100 ug/m3 (daily) are considered
unhealthy.
The omission of corresponding pollution data for the control areas
is unavoidable because such measurements have not been done on a routine
basis for areas suitable as controls in studies of this type. In our
conversations with the local pollution officials, the comment was fre-
quently made that there is a definite need to establish monitoring net-
works in nonproblem areas which could serve as potential control sites
in pollution-disease correlation. In the case of Arizona, for example,
the only data available for even general comparison purposes were re-
cently released by the Arizona Department of Health Services in its 1976
Air Quality Data for Arizona. The annual average data indicate that
hazardous levels of SO2 are not found in nonsmelter areas such .as Tucson
and Phoenix. Even in 1976 when pollution control was strictly enforced,
it was evident that smelter towns still exceed nonsmelter towns in
V-34
-------
pollution from sulfur dioxide (see Table V-ll). This rank order can
reasonably be expected to have been more pronounced in the pre-1970 era
focused on by this study. Hence, although the data in Table V-10 lack
the rigorousness of a comparison of exposed and nonexposed pollution
data, it is reasonable to assume that the SOa levels in the control
areas were low enough to make the comparison implied in Table V-10 a
valid one.
Table V-ll
1976 ANNUAL AVERAGES FOR SULFUR DIOXIDE (\lg/m* )
Smelter Towns
Ajo (5 sites)
Douglas (10 sites)
Hayden (6 sites)
Miami (2 sites)
Morenci (10 sites)
San Manuel (5 sites)
Nonsmelter Cities
Phoenix (5 sites)
Tucson (7 sites)
16
34
103
38
49
44
7
5
± 10
± 11
± 29
± 24
± 52
± 20
± 0.5
± 2
Source; Arizona Department of Health Services.
(Raw data for individual sites is given in
Table F-18, Appendix F.)
2. Heavy Metals
Table V-12 summarizes the levels of cadmium, lead, arsenic, and
fluoride in the Arizona and Montana smelter sites. The available data
for Utah is in units which cannot be compared to that for Arizona and
Montana. The records used to generate these summaries are detailed in
Tables F-10 through F-13 in Appendix F. Since there is much less basic
toxicological data (and consequently no environmental standards) for
these elements than for SOa, it is difficult to designate areas as
"high" or "low." Furthermore, there has not been routine monitoring for
V-35
-------
heavy metals for the comparison areas as is the case with the sulfur
dioxide analysis discussed above. With these limitations in mind, we
can draw the following conclusions about heavy metal pollution in smel-
ter towns.
Table:V-12
SUMMARY OF HEAVY METAL POLLUTION FOR ARIZONA AND MONTANA
AREA
CADMIUM
Arizona Smelter Towns
Arizona Town* Without Smelters
Anaconda
LEAD
-Arizona Smelter Towns
Arizona. Towns Without Smelters
Anaconda
ARSENIC
Arizona Smelter Towns
Arizona Towns Without Smelters
Anaconda
FLUORIDE
Anaconda (West)
YEARS
1969
1969
1971-72
1969
1969
1961-62
1969
1969
1961-62
1965
POLLUTION
Atmospheric
(pg/m3 )
Average ±o
.006 ± .005
.004 t .005
.03 ± .02
1.2 ± 1.2
0.18 i 0.2
0.54
.0065 ± .0027
.0042 ± .0042
0.45
Nonatmospheric
3.6 pptn (soil)
•33 ppm (soil)
12.6 ppm (grass)
(1) Contamination by cadmium is less of a problem in Arizona as a
wjhole than it is in Anaconda. The levels in Anaconda approach the upper
limit of those measured in a nationwide survey of nonurban areas; these
values were 0.0004 to 0.026 yg/m3. Based on a sample taken in 1977,
levels of soil cadmium in a residential street adjacent to the plant are
3ft least three times average soil levels. The town of Hayden in Arizona
has a level from 10 to 100 times higher than the other five Arizona
smelter towns.
(2) Contamination by arsenic is much less of a problem in all Ari-
zona smelter towns than in Anaconda. Anaconda has the highest arsenic
V-36
-------
level in Montana and is also higher than New York City and Chicago.
Based on a sample taken in 1977, levels of soil arsenic are about four
times higher than average soil levels.
(3) Lead levels in the Arizona smelter towns were high in comparison
to the nationwide survey results of 0.002 to 0.15 yg/m3 in nonurban
areas. Anaconda is lower than Arizona, but still above the nationwide
levels.
Furthermore, fluoride, although not emitted from copper smelters,
is present in high levels in West Anaconda.
Table F-14 in Appendix F shows the heavy metal body burden data from
two studies. One is from Anaconda and Bozeman done on hair in 1969; the
other was done with hair, urine, and blood samples from children in 1975.
The body burden data support the conclusion that individuals who live in
smelter areas take up and retain lead, cadmium, and arsenic and have
bodily levels significantly different from individuals living in non-
smelter towns. The body burden data do not appear to follow the same
rank order as the environmental pollution levels, although such a com-
parison may not be warranted due to the difference in time periods of
the two data sets and the use of children as subjects in the Baker
study [18].
When compared to mortality rates from the control areas, death rates
from cancers of the colon and pancreas and from cerebrovascular diseases
were elevated in Anaconda but not in Douglas. The excessive rates from
the cancers in Anaconda may be explained by the presence of high levels
of suspected carcinogens (cadmium, arsenic, fluoride) in that area. The
excessive rate from cerebrosvascular disease may be accounted for by the
high cadmium levels in Anaconda. The evidence for this hypothesis is
summarized in Table V-13.
[18] Baker, E. L., C. G. Hayes, P. J. Landrigan, J. L. Handke, R. T.
Leger, W. J. Housworth, and J. M. Harrington, A nationwide survey
of heavy metal absorption in children living near primary copper,
lead and zinc smelters. Am. J. Epidemiol 106(4):261-273, 1977.
V-37
-------
Table V-13
SUMMARY DATA FOR HEAVY METAL HYPOTHESIS
AREA
Anaconda
Douglas
CAUSE OF DEATH
Cancer of colon
Cancer of pancreas
Cerebrovascular disease
Cancer of colon
Cancer of pancreas
Cerebrovascular disease
ACE-ADJUSTED
MORTALITY
RATE i W
56
39
133
33
5
79
CADMIUM
(ug/m3)
.03
.011
ARSENIC
(ug/m3)
0.45
.004
FUIORIDE
(PP». 9)
12.6
HA
Underscored rates are significantly higher than control values.
HA • Data not available.
National averages (WF) (mortality/100,000)i Cancer of colon 4 26
Cancer of pancreas 39
Cerebrovascular disease 84
PBYSIOU3GKAL DATA:
e Recent toxicological and apidemiological evidence indicate that
cadmium, arsenic, and fluoride are all implicated as possible
carcinogens. Cadmium has been linked with vascular and
hypertensive disease as well. (See Section B above.)
As pointed out above, because there is insufficient heavy metal data
for the comparison areas and a lack of environmental standards, we can
only make a rank order comparison. Thus, the available data point to
the fact that in Anaconda there are high levels of cancer of the colon
and pancreas as well as cerebrovascular disease which occur in conjunc-
tion with elevated environmental levels of pollutants that have been
associated with these diseases. In Douglas, rates for the diseases are
not elevated with respect to control values, and the levels of suspect
pollutants are low.
V-38
-------
3. Digestive Diseases
Diseases of the liver, gallbladder, and pancreas (disease 27) are
elevated in Douglas, Anaconda, and the Arizona smelter town aggregate.
A breakdown of individual diseases included in this grouping is given .
in Table V-14. From these data we see that 70 percent (24 out of 34)
of the deaths are from cirrhosis of the liver; 38 percent (9 out of 24)
of the cirrhosis cases were listed on the death certificate as cirrhosis
due to alcoholism. In Arizona, it appears that most of the deaths in
this category are due to alcoholic cirrhosis. The diagnosis of "unspe-
cified cirrhosis" may indicate that the disease is not related to alco-
hol consumption or it may reflect a local tendency to omit mention of
alcoholism on death certificates in deference to family sensitivity on
the issue, a practice which has been reported to occur. Although lead,
cadmium, and arsenic have all been described as either accumulating in
and/or actually damaging the liver, the evidence is not sufficient to
suggest a role for these metals in the aggravation of alcoholic cirrho-
sis.
V-39
-------
Table V-13
DISEASES OF LIVER, GALLBLADDER AND PANCREAS: NUMBER OF DEATHS (WHITE FEMALES, 1968-1975)
•H
n
5
M
0
n
1
vi
VI
•H
V
I
DISEASES OF
LIVER, GAJULBLADD5P & PANCREAS
Alcoholic cirrhosis of liver
Other specified cirrhosis of
liver (without alcoholism)
Unspecified cirrhosis of
liver (without alcoholism)
Acute & subacute necrosis of
liver
Suppurative hepatitis & liver
abscess ,
Other diseases of liver
Cholelithiasis
Cholecystitis & cholangitis
Other diseases of gallbladder
& biliary ducts
Diseases of pancreas
Column Totals
NUMBER OF DEATHS
Miani
0
0
0
0
0
0
0
0
0
0
0
Hayden
0
0
0
0
0
0
0
0
0
0
0
Ajo
2
0
0
0
-
0
0
0
0
0
0
2
Douglas
3
1
3
1
0
0
0
2
1
2
13
San
Manuel
0
0
0
0
0
0
0
0
0
0
0
Horenci
2
0
0
0
0
0
0
0
o-
0
2
Anaconda
0
2
6
1
0
0
0
0
0
2
11
Magna
2
1
2
0
0
0
_
0
0
0
1
6
Row
Totals
9
4
11
2
0
0
0
2
1
5
34
All of
Arizona
471
80
128
16
2
66
34
31
15
65
908
All of
Montana
58
31
99
30
10
15
10
10
10
26
299
All of
Utah
105
35
96
31
14
22
16
14
14
39
386
I
£»
3
-------
4. Factor Analysis
Specific data on the industry, environment, and demographics were
assembled for 11 Arizona towns in the proximity of copper smelters. The
towns range from 0.3 to 9 miles from smelters. Table V-15 lists the
towns and available data. These data, together with disease-specific
mortality rates in white females for the towns, were subjected to an
iterative factor analysis to attempt to find the commonalities among the
data, i.e., to identify groups of variables that are associated with
common factors. The purpose of the analysis was to explore the patterns
in the data to see if certain diseases appear to be associated with spe-
cific variables, either socioeconomic, industrial, or environmental in
nature. Patterns and associations in the data involving disease vari-
ables were then tested for strength of association by means of a mul-
tiple regression analysis in which the disease mortality rate was taken
to be the dependent variable.
This analysis resulted in only three strong associations. These are
shown in the following table.
DISEASE
Acute Respiratory
Disease
(pneumonia , etc . )
Digestive Disease
(primarily liver
cirrhosis)
Cancer of Genito-
urinary Organs
ASSOCIATED VARIABLES
• Ambient S02 concentration
• Age of smelter
• Percent population white
• Tons per day of concentrates
• Percent not completing high
school
• Number of persons per room
• Tons per day of concentrates
STEPWISE
MULTIPLE
REGRESSION
COEFFICIENT (R)
0.84
0.88
0.44
0.64
0.91
0.74
0.81
(AR)2
0.70
0.07
0.02
0.21
0.43
0.55
0.11
The clearest association is that between respiratory disease and
ambient S02 concentrations. To a much lesser degree, the disease is
associated with the age of the smelter. These two variables together
V-41
-------
Table V-15
DEMOGRAPHIC, ENVIRONMENTAL AND INDUSTRIAL DATA
FOR ARIZONA TOWNS IN VICINITY OF COPPER SMELTERS
ARIZONA TOWNS IN VICINITY OF COPPER SMELTERS
Specific 'Population4
Median Age
Median Income
% in Same Housa*
% White"
% with <1 person
per room -
% Completed High School
% Female'
« Native-Born Parents
Age of Smelter (years)
Distance to Nearest
Smelter (miles)
Ambient Arsenic (wg/m3 )
Ambient .Lead (ug/m3)
Ambient Cadmium (yg/m3)
Ambient SOj (ug/m3)
Particulatesd (yg/m3)
Concentrates (tons/day)
Blister Copper
(tons/day)
Stack Height (feet)
SOj Concentration (ppm)
Particulates Emitted
(tons/day)
SOj Quitted (tons/day)
209
21
9,500
70
100
71
42
51
69
60
0.3
™«
—
.016
367*87
224
2,000
366
300
5,900
1.5
240
482
18
11,000
32
98
77
63
49
as
16
1
.01
±.002
0.15
±.07
.003
±001
102*40
—
1,700
310
515
15,000
1.79
313
638
23
7,500
63
99
83
41
51
71
57
0.85
—
—
.01
±.01
80±47
—
840
300
275
2,500
"
78
790
23
8,500
61
98
81
37
50
72
47
0.5
.01
±.01
0.58
±0.10
.01
±.01
1581132
142
~
~
--
--
—
--
453
30
7,500
70
95
84
45
51
85
57
0.85
.01
±.003
0.93
±0.31
.01
±.01
80±47
—
840
300
275
2,500
~
78
2,017
30
8,500
58
98
90
49
51
85
57
2.5
—
—
—
—
—
840
300
275
2,500
--
78
448
23
11,000
48
96
93
66
49
85
60
9
~
—
—
—
—
2,000
366
300
5,900
1.5
290
ISO
20
9,500
48
100
81
38
50
72
60
0.6
.004
±.00?
3.0
±1.4
.004
35±17
—
2,000
366
300
5,900
1.5
290
979
30
9,500
42
90
83
SO
51
72
25
0.5
.004
±.002
0.27
±0.21
.002
±.018
95±67
78
680
197
360
7,000
0.7
380
663
30
9,500
51
97
88
52
51
84
33
0.5
.0045
±.003
1.2
±1.4
.005
±.001
115*20
—
2,113
470
600
14,000
14.7
526
2,400
23
7,500
58
98
84
43
53
43
65
0.7
.02
±.02
1.13
±0.69
.002
±.002
80±40
~
2,260
365
556
15,200
9
492
All demographics based on 1970 Census.
Since last census.
Nonvhite includes Indians and blacks.
Insufficient data for use in factor
— • data missing; row mean used in
analysis.
factor analysis.
V-42
-------
are strongly correlated with the disease mortality rate in females
(R = 0.88). Further, no appreciable correlation was found between re-
spiratory disease and the socioeconomic variables. This result provides
evidence that S02 is the principal etiologic agent contributing to ele-
vated respiratory disease mortalities.
The other two associations found appear to be principally socio-
economic relationships. In both cases, the only industry-related vari-
able is the production variable of tons per day of concentrates. It is
not clear what this variable represents. It could be a surrogate mea-
sure of some emission or effluent which was not present in the data, or
it could represent more of a socioeconomic measure related to total
number of employees and total income generated by smelter workers.
B. CONCLUSIONS AND RECOMMENDATIONS
1. Conclusions
• Three hypotheses resulted from this study which link a specific
industry emission to an excess mortality pattern. They are discussed
in detail in Section D.
• The regression analysis, using county-level data, successfully
pointed out elevated mortalities which proved also to be high when
examined on a town-by-town basis. In the case of respiratory disease,
the original regression category was very broad. The present study nar-
rowed down the diseases involved considerably. In the case of digestive
diseases, the most likely cause is not industry pollution, but rather
excess alcohol consumption. The usefulness of the regression technique
to objectively detect community health problems is limited by quantita-
tive factors such as the ratio of town-to-county populations, preva-
lence of the disease, and influence of socioeconomic factors.
V-43
-------
2. Recommendations
• The regression methodology can be made more specific by the use
of mortality figures from the towns actually at risk from a particular
industrial pollutant, rather than the rates for the entire county which
may considerably dilute the town effect. Screening for town-level ef-
fects can be accomplished by the approach used here, i.e., use of data
obtained from state registries. It would be highly desirable if all
states included in their mortality certificates and tapes the length of
the decedent's residence in the town and his or her occupation. This
would be a valuable epidemiological tool.
• There is a need for more environmental data for control, non-
industrial areas.
• The relationship between ambient pollution levels and individual
exposure needs further evaluation. It would be useful to perform pilot
studies in towns such as those described here, using personal monitors
on groups of residents to more precisely determine their exposure.
• The work described here represents a study of about half of the
primary copper smelters in the United States. Three specific industry-
dj.sease associations have been identified. It would be useful to deter-
mine if these associations occur in the exposed populations of the re-
maining smelters in the country.
V-44
-------
Chapter VI
STEEL MANUFACTURING CASE STUDY
A. INTRODUCTION AND SUMMARY
1. Study Objectives
The objective of this case study is to explore the relationship
between the steel industry and community disease, in particular,
digestive cancers and cardiovascular diseases. The correlations
between these diseases and the steel industry were generated by the
multiple regression analysis of county industry and mortality data.
This case study is a field test to attempt to ascertain the nature of
the relationship between the industry and the diseases. The results
of this investigation are to assist in the evaluation of the effec-
tiveness of the multiple regression analysis for identifying true
associations between industry and community disease. Thus, the ultimate
.purpose of this case study is to assist in the development of method-
ology for identifying sources of harmful substances affecting the health
of the community.
The regression analysis correlated the presence of the steel indus-
try (SIC 3312) in counties with elevated mortalities from digestive
neoplasms. Testing this correlation involved finding a plausible cause-
and-effect hypothesis at the sites studied. We chose to study Johnstown,
Pennsylvania, as our hypothesis development site and Provo, Utah, for
hypothesis test. After determining which steel processes {and therefore
which emissions) were present at the selected sites, which emissions had
a plausible etiology for the relevant diseases, and what the ambient air
levels of these emissions were, we were able to delineate a population
potentially at risk. We were then able to analyze mortality data in
terms of potentially exposed and unexposed areas and determine signi-
ficant elevations in disease mortality rates and to associate them with
specific etiologic agents.
Several methodologies were developed in the course of these analyses
to utilize the data available. Transport analyses were devised by our
VI-1
-------
staff as well as by EPA's Environmental Photographic Interpretation Com-
plex (EPIC) to characterize pollution patterns and populations at risk.
EPIC data were used to get a historical perspective on changes in both
industrial activity and population by comparing aerial photographs from
as far back as 1939 to the present. The availability of cancer registry
data made it possible to analyze incidence rates and other factors be-
sides mortality. Finally, several statistical analyses were developed
and used, depending on data type and availability, to determine signifi-
cant differences between test and control values.
2. Study Approach
The field study was conducted at two sites. The first site consti-
tuted the hypothesis development site and the second, the hypothesis
test site. Both sites were chosen using the basic assumptions that the
study county should be no larger than 200,000, should have a high indus-
trial index for the SIC code 3312, and should contain no other industry
that might produce confounding emissions. These conditions being met,
the hypothesis development site (Johnstown, Pennsylvania) was chosen
because that county showed elevated mortality rates for the diseases of
interest. The hypothesis test site (Provo, Utah), on the other hand,
was chosen using only the basic assumptions. Of course, in both cases
data availability was an important factor in choosing the site.
For each site and its control, we obtained the following data:
• Mortality and census data—used to calculate age-adjusted
mortality rates' for each site.
• Air quality data—to determine population exposure.
• Site-specific industry data—production figures and
emission control data to determine historical exposure
as well as emissions present.
The data for Johnstown were analyzed, plausible etiologies developed in
light of those causes of death found elevated, and hypotheses developed.
The Provo data were analyzed in terms of the hypotheses developed at
Johnstown. Possible explanations for any differences found in the two
sites were explored.
VI-2
-------
3. Key Findings
. The hypotheses developed in Johnstown are summarized below:
In both males and females—
• Various malignant neoplasms (ICDA codes* 140-151 oral
cancers, 158-163 digestive cancers, 170-174 breast
cancers, 190-239 central nervous system cancers and
leukemias) caused by. high-level exposure to organic
emissions, notably BaP.
• Cardiovascular diseases (ICDA codes 400-458) caused by
exposure to high cadmium levels.
• Respiratory diseases (ICDA codes 460-519) caused by
high exposure to particulates, S02, and trace metals.
In males only (possibly occupational)— ,
• Endocrine, metabolic, and blood diseases (ICDA codes
240-289) due to exposure to trace metals.
• Liver diseases (ICDA codes 570-577) due to exposure to
organics and trace metals.
In females only—
• Urinary diseases (ICDA codes 580-599) from trace metal
exposure.
We then tested these hypotheses using the Provo data. The following
hypotheses were well supported by the Provo data:
In both males and females—
• Various malignant neoplasms (ICDA codes 170-174 breast
cancers, 190-199 central nervous system cancers, 210-
239 benign neoplasms of the central nervous system)
causes by exposure to organic emissions such as BaP.
In males only (possibly occupational)—
• Endocrine, metabolic, and blood diseases (ICDA codes
240-289) due to trace metal exposure.
Several other hypotheses were to a lesser degree supported by the
Provo data, i.e., mortality rates were elevated but not significantly.
*Note: All ICDA codes refer to the 8th Revision.
VI-3
-------
The causes of death in these cases included respiratory cancers, where
in Utah the mortality rates were elevated only in males, and there not
significantly. The cardiovascular diseases, elevated only in Utah women,
were not significantly higher. While leukemias were significantly higher
in males in Johnstown, they were only slightly elevated in males in
.Utah. The same was true for respiratory diseases in females.
By comparing the pollutant levels in' Provo to those in Johnstown,
'the lower death rates in Provo can be readily explained. The pollutant
.levels at the Utah sites are roughly an order of magnitude lower than
at Johnstown. Thus, assuming the Johnstown hypotheses are valid, there
is a logical explanation why the mortality rates for some of the dis-
eases were not significantly:elevated in Utah. This rationale for the
Utah results does not "prove" that the Johnstown causal hypotheses were
correct—it simply provides a reason why the Utah results may not
"disprove" the Johnstown results.
'4. Conclusions and Recommendations
Some of the hypotheses developed at Johnstown withstood the test at
Utah, notably malignant neoplasms caused by exposure to BaP. The reason
that other hypotheses developed at Johnstown did not withstand the test
at Utah could have been due to the much lower exposure levels at Provo.
Before definite conclusions can be drawn, more detailed studies must
be performed to determine the effect of such variables as occupation,
tobacco and alcohol consumption, and duration of residence on the mor-
taility rate patterns. Also, more detailed air quality data should be
obtained, especially for control sites.
VI-4
-------
fl. BACKGROUND
1. Basis for Selecting Associations
Potential industry-disease associations identified from the regres-
sion data were examined statistically to determine those likely to be
causally related. The 10 counties having the highest industrial index
for steel (SIC code P047) were aggregated and the mean sex-specific age-
adjusted mortality rates for the specific diseases indicated by the
regression were calculated. These were compared to the mean United
States mortality rates for the diseases. The results of this comparison
are found in Table VI-1.
Table VI-1
AGGREGATED MORTALITY RATES—TOP 10 INDUSTRIAL COUNTIES
SIC CODE P047 COMPARED TO U.S. MEAN
Disease
Cardiovascular Disease
Ischemic Heart Disease
Myocardial Infarction
Digestive Neoplasms
Aggregated Mean
Mortality Rates
MM
1,481
701
542
110
WF
695
298
198
69
Number of Sigma
from U.S. Means
WM
+8.5
+ 5.3
+11.8
+5.9
WF
+11.5
+7.5
+9.6
+ 2.1
Since many of these counties were highly industrialized, in an
effort to exclude effects from other polluting sources or "urbanization"
we also examined eight uncluttered counties with high industrial indices
for SIC code 3312 (a component of PO47). The results of this analysis
are shown in Table VI-2.
VI-5
-------
Table VI-2
AGGREGATED MORTALITY RATES—UNCLUTTERED COUNTIES
SIC CODE 3312 COMPARED TO U.S.i MEAN
Disease
Cardiovascular Disease
Ischemia Heart Disease
Myocardial Infarction
Digestive Neoplasms
Aggregated Mean
Mortality Rates
WM
1,435
677
519
112
WF
685
292
193
69
Number of Sigma
from U.S. Means
WM
+4.6
+2.4
+8.0
+5.9
WF
+ 9.4
+ 5.9
+7.8
+ 2.0
As the tabled data shows, the rates for counties containing 3312
were significantly higher than the U.S. mean, so a further analysis was
done to compare these rates to a similar mean rate for an aggregate of
matched control counties not containing the industry (see Table VI-3).
The only association in which the mortality rates for the industrial
counties were significantly higher than the U.S. mean and the mean for
the aggregated control counties was that for digestive neoplasms. This
association was therefore selected for further study.
Table VI-3
AGGREGATED MORTALITY RATES—UNCLUTTERED COUNTIES
SIC CODE 3312 COMPARED TO MATCHED CONTROL COUNTIES
I*\ -i r*r* ^ r**\
Cardiovascular Disease
Ischemic Heart Disease
Myocardial Infarction
Digestive Neoplasms
Number of Sigma
WM
+ 2.4
+ 1.4
-0.6
+ 5.3
WF
-3.2
-2.8
-1.3
+3.2
VI-6
-------
2. Description of Steel Industry
From iron ore, scrap metal, oxygen, coal and limestone, the steel
industry produces hundreds of different steels, each having unique
physical properties. Basically, iron ore is converted into iron and the
content of other materials in the iron is suitably adjusted by the re-
moval of some materials and the addition of others. Reaction vessels
hold up to 300 tons of material and their chemical contents are con-
trolled to less than one percent.
The operation of making steel from iron ore with coal and oxygen is
shown in Figure VI-1. Initially, cleaned and sized coal is reduced by
heating in coke ovens to coke (almost pure carbon). Next, coke, lime-
stone and iron ore are mixed in a blast furnace with enough 02 to burn
the coke. This reaction releases enough heat to reduce the ore to
molten iron. The next step refines the iron to steel by oxidation in
one of three types of furnaces: open hearth, basic oxygen, or electric
arc. This process creates the steel alloys with their special proper-
ties. Usually, selectively separated scrap and other materials are added
to the iron in the furnace, while recarbonizing materials are added to
the molten steel as it leaves the furnace.
The sinter plant is basically a recycling process. Small iron-
bearing fines are collected from the steel plant and agglomerated to a
size suitable for blast furnace feed. The fines cannot be directly fed
to the furnace, because the high rate of gas flow through the furnace
would carry them out.
A wide range of pollutants are generated by the diverse processes
described. Water is used in a steel plant to cool hot steel, to clean
airstreams, to cool hot coke, etc. In the actual steelmaking process,
water comes in direct contact with coke in the quenching process, causing
hydrocarbons to be carried into the wastewater. The furnace off gases
carry particulate matter which enters the plant's wastewater when water
is used to scrub the gases. Consequently, the same pollutants in air
emissions are also in wastewater effluents when water scrubbers are used.
VJ-7
-------
CD
IAir
Emissions
Coal
OVEN
Coke
Coal
Distillation
Products
*
By-Product
Recovery
IAir
Emissions
Air
Plant Fines
SINTER
PLANT
Air
Iron Ore
Flux
Sinter
IAir
Emiss
BLAST
FURNACE
ions
IAir
Emissi
O2
Iron
Scrap
Flux
BASIC
OXYGEN
OPEN
HEARTH
ELECTRIC
ARC
JSlaq
ons
Steel
Slag
Figure VI-1. Steel Production
-------
A broad range of pollutants from various sources are emitted into the
atmosphere: hydrocarbons and oxides, gases and particulates, fugitive
arid stack emissions. Coke ovens reduce the complex hydrocarbons that
make up coal, resulting in a large list of known and potentially toxic
emissions, both particulate and gaseous. The sintering operation emits
particulates, but they are hard to qualify due to the diverse nature of
the feed. In blast furnaces, impurities are separated from iron ore that
is reduced, creating gaseous and particulate emissions. Steel furnaces
generate large amounts of respirable-sized particulates which are gener-
ally oxides.
a. Coke Ovens
Coke has been produced in beehive and in by-product, enclosed slot-
type ovens. The beehive oven does not recover the by-products of de-
structive distillation of coal, while the enclosed slot-type does. About
99% of all metallurgical coke today is produced by the enclosed slot-type
ovens. The beehive oven allows limited amounts of air into the chamber
to burn with the volatile products distilled from the coal. The burning
matter generates heat for further distillation. The enclosed slot-type
is heated by external combustion of fuel gases around the oven walls.
A by-product coke oven is typically made of silica brick and is rec-
tangular in shape (approximately 1.5' x 20' x 40'). Ovens are arranged
side by side and are called batteries; as many as 100 ovens may be in a
battery. Gases burning in flues in the walls between adjacent ovens,
each of which holds 16 to 20 tons of coal, heat the coal up to 1100°F.
Coking time varies from 16 to 24 hours, depending on oven size, coal
quality, degree of coking achieved, etc.
Special cars (lorries) filled with coal, riding on rails on the roof
of the battery, charge each oven with coal through holes (ports) in the
top of the oven. The lorry cars discharge from the bottom and carry a
measured weight of coal. A collection system, which runs the length of
the battery, collects the gases and volatile matter distilled from the
coal during the coking process under suction through pipes in the top of
VI-9
-------
each oven. At the end of the coking process, the doors at each end of
the oven are removed and a ram from a pushing machine which travels the
length of the battery pushes the coke into a quenching car. The car
carries the red-hot coke to a quenching tower where it is cooled by water
sprays.
The gases and volatile materials (by-products) collected during the
coking process may be recovered by various processes including tar
decanting, recovery of phenol, distillation and absorption of ammonia,
crystallization of ammonium sulfate, and fractionation, recovery and
refining of light oils. The size of the plant and economics determine
the extent of the recovery process employed; often the material is used
as fuel inside the plant.
Emissions from the coking process are produced during coking, pushing
and quenching (see Table VI-4). Their composition is varied depending
on the coal used, degree of coking, degree of by-product recovery, etc.
These emissions and their status with respect to health are further
elucidated in Table VI-5. The particle size distribution is shown in
Table VI-6.
Table VI-4. EMISSION FACTORS FOR COKING PROCESSES
pollutant
Particulate
CO
Hydrocarbon3
N02
Ammonia
kg/metric ton of Coal Charged
Coking Cycle
0.05
0.3
0.75
0.005
0.03
Discharging
0.3
0.035
0.1
—
0.05
Expressed as methane.
VI-10
-------
table VI-5. HAZARDOUS CHEMICALS PRESENT IN EMISSIONS FROM COKE PRODUCTION
Known Hazardous Chemicals
Chemical Component
a s 6 naphthyl amine
4-aminobiphenyl
carbon Monoxide
pyridine
benzene
toluene
xylene
phenol
o-,m-,p-cresol
pyrene
chrysene
benzo (a) pyrene
benzo (e) pyrene
d Ibenzo (a , h ) anthracene
dibenzo(a,g)fluorene
nickel carbonyl
thiophenes
selenium
arsenic
Ivtriim
•JSdUBI
cadmium
lead
mercury
tar
soot
hydrogen cyanide
ammonium cyanide
Class
amines
combustion gases
heterocyclics
hydrocarbons
phenols
polynuclear
organometal 1 ics
sulfur compounds
trace elements
fine participates
cyanides
Suspected Hazardous Chemicals
Chemical Component
benzole acid
hydroxybenzoic acid
hydrochloric acid
sulfuric acid
ammonia
aniline
methylanilines
amnonium sulfate
formaldehyde
acetaldehyde
paraldehyde
methylchrysenes
benzo (a) anthracene
dine thy Ibenzanthracene
methyl mercaptan
ethyl mercaptan
antimony
beryllium
silver
vanadium
ammonium thiocyanate
Class
acids and anhydrides
amines
inorganic salts
carbonyl compounds
polynuclear
sulfur compounds
trace elements
cyanides
-------
Table VI-6
PARTICLE SIZE DISTRIBUTION FOR A COKE OVEN EMISSION SAMPLE
Size, microns
13.5
8.6
. 5.6
4.0
2.5
1.3
0.8
0.5
% by Weight of Total Sample
Collected on Plates
31.3
27.7
12.3
9.1
7.3
7.3
3.8
1.0
b. Sintering
The sintering process agglomerates fine iron-containing particles
into a material suitable for charging into the blast furnace. Particles
used in the process include fine ore, blast furnace flue dust, mill
scale, turnings, coke fines, limestone fines, and other miscellaneous
fines from the plant. They are placed on a grate and heated from above
with a downdraft. At 2300 to 2700°F the fines agglomerate. The end
process crushes, screens, and sizes the hot sinter.
Due to the material being handled—fine dust particles—sintering
generates a large amount of air pollution. Handling generates fugitive
emissions, and hydrocarbon fumes may be evolved when oily scrap is used.
Also dependent on the feed material is the sulfur content of the gas
stream which may reach 2,000 ppm. Particulate emissions are estimated
to be 22 Ib/ton of sinter. Table VI-7 describes the particle size dis-
tribution.
Controls on the sinter process include cyclones, electrostatic pre-
cipitators, baghouses, and scrubbers. Typically, a dry cyclone by itself
will reduce emissions to about 2 Ib/ton of sinter; while a dry cyclone in
series with an electrostatic precipitator further reduces emissions to
about 1 Ib/ton of sinter.
VI-12
-------
Table VI-7. SINTERING EMISSIONS
Particle Size Distribution
Size, microns
44
30
20
10
5
Percent
Than By
98
93
87
72
25
Less
Weiqht
c. Blast Furnaces
A blast furnace is roughly cylindrical in shape and often measures
100 feet high and 20 feet in inside diameter. Overall height including
the superstructure is often 150 feet. The structure has a steel shell
lined with heat-resistant brick. Operating parameters range from 10 to
30 psi and from 400°F at the top of the furnace to over 3000°F at the
bottom.
Generally, iron ore, coke, and flux are charged at the top of the
furnace and hot air at the bottom. The air flows countercurrent to the
falling charge. Carefully weighed coke, flux (limestone and others), and
iron ore (including sinter pellets, mill scale, and iron or steel scrap)
are carried to the top of the furnace in skip cars and charged to the
furnace through a lock hopper. Air is compressed, then passed through a
preheating vessel(cleaned furnace off-gas is used to heat the vessel),
and introduced into the bottom of the furnace above the molten iron and
slag. Oil or coal sludge is often injected along with the air.
As the coke descends through the furnace, it is preheated and finally
burned. The heat of the reaction, combined with the hot air, reduces the
iron ore (usually Fe2O3, FesO.,, FeCO3, and FeS) to Fe and then to molten
iron. Fluxes combine with impurities to form a molten material, slag.
Slag is lighter than the molten iron and floats on it at the bottom of
the furnace.
VI-13
-------
Gases exit from the top of the furnace while the iron and the slag
are periodically drawn off. The slag is usually handled in one of two
ways: 'the air-cooling method pours the slag into dry pits in the ground
to become one large solid; the other process rapidly cools the slag with
high-pressure water streams and results in fine, brown sand-like par-
ticles. The molten iron is transferred by railcar to molds for pig iron,
or to a holding vessel, or directly to a steel furnace. Gases exit at
400°F at a rate of 5 tons of gases/ton of iron. Table VI-8 describes
the off-gas composition. An estimate of particle loading in the gas is
200 Ib/ton of iron before control. Particulate composition is described
in Table VI-9. Typically, particulates have been cleaned from the gas
stream, because the hot steam is used to preheat the air entering the
furnace—as well as in other processes involved in steelmaking—and
particulates would clog the heat exchanger. Common controls on the gas
stream include dry cyclones, web scrubbers, and electrostatic precipi-
tators.
Table VI-8. BLAST FURNACE GAS COMPOSITION
Gas
CO 2
02
CO
H2
CHi,
N2
H20
Percent of
Total Gases Emitted
12.8
<0.1
26.5
3.5
0.2
56.9
<0.1
VI-14
-------
Table VI-9
CHEMICAL COMPOSITION OF DRY, BLAST-FURNACE FLUE DUST
Component
Weight Percent
(range)
Iron
Ferrous oxide
Silicon dioxide
Aluminum oxide
Magnesium oxide
Calcium oxide
Sodium oxide
Potassium oxide
Zinc oxide
Phosphorus
Sulfur
Manganese
Carbon
36.5 - 50.3
N.A.
8.9 - 13.4
2.2 - 5.3
0.9 - 1.6
3.8 - 4.5
N.A.
N.A.
N.A.
0.1 - 0.2
0.2 - 0.4
0.5 - 0.9
3.7 - 13.9
N.A. = not available.
VI-15
-------
d. Open Hearth
An open-hearth furnace is a shallow, rectangular refractory basin,
which typically measures 30" x 90' and is lined with refractory material.
Often several furnaces are operated in a line. Water-cooled doors are in
the front, fuel jets are in opposite sides, an oxygen lance may be in the
ceiling, and the taphole is in the lower back of the furnace.
Blast furnace iron, steel scrap, or equal portions of both are
charged into the furnace in addition to a flux via the doors. Initially,
flux (usually limestone) and scrap are dumped through the doors by the
charging machine. Oil, tar, gas, etc., are burned from either end of the
furnace. When the charge is partially melted, molten iron is poured into
the furnace. Oxygen lances have been incorporated in newer furnaces and
retrofitted onto many older ones for faster heating times. As the charge
heats, carbon, manganese, phosphorus, sulfur, and silicon are removed from
the iron, and they make up the slag along with the limestone which floats
on the molten steel. When the steel has reached a predetermined carbon
content at about 2900°F, the furnace is tapped. A large ladle is placed
under the taphole and the plug in the taphole is blown out. As the
steel flows into the ladle, other materials can be added to create the
desired steel. Because the initial charge was carefully measured, the
ladle fills with steel and the slag overflows into a smaller ladle along-
side.
The flue gases pass through a heat exchanger which preheats the air
needed for combustion ; its particle loading is described in Figure
VI-2 and Table VI-10. Waste gases are composed of air, carbon dioxide,
water vapor, sulfur oxides, nitrogen oxides, zinc oxides, and fluorides.
The zinc oxide quantities are dependent on the use of galvanized steel
scrap. Fluorides can be either gaseous or particulates and are generated
when fluorspar (CaFa) is used as a flux. Emission controls include
electrostatic precipitators, venturi scrubbers, and baghouses, with the
precipitator the device of choice.
VI-16
-------
100
90 4
80
70
60
§
M 50
33
40 f
30
A
A
x
x
X
o
20 f O
10
A
fi
ft
o Basic Oxygen
A Electric Arc
x Open Hearth
10
20 30
Particle Size, y
40
50
Figure VI-2. Steel Furnaces: Particle Sizes by Percent by Weight Less Than Stated Sizes
-------
Table VI-10. STEEL FURNACE PARTICULATE EMISSIONS, UNCONTROLLED
Type of Furnace
Emission Factor
(Ib/ton of steel)
Chemical
Composition
Open Hearth
Not Oxygen-Lanced
Fe203
Si02
A1203
CaO
MnO
Oxygen-Lanced
21
Fluorides
Fe203
FeO
Si02
A1203
CaO
MnO
MgO
S
Basic Oxygen
40
Fe203
FeO
Mn
SiO2
A1203
CaO
MgO
P205
C
S
Electric Arc
10
Fe203
FeO
Cr203
SiO2
A1203
CaO
MgO
MnO
ZnO
CuO
NiO
PbO
C
S
P
VI-18
-------
e. Basic Oxygen
The basic oxygen furnace (EOF) is pear shaped, lined with refractory
material, and open at the top. The entire furnace tilts for charging and
discharging. The furnace is first charged with scrap steel, immedi-
ately followed by molten iron (65% to 80% of the charge is molten iron).
After being charged and when the furnace is in a vertical position, an
exhaust hood is lowered to cover the top and an oxygen lance is lowered
through the hood into the furnace to about 6 feet above the charge.
Oxygen is injected through the lance at flows up to 6,000 cfm and at
pressures up to 160 psi. Later, fluxes are added and the steel is refined
at about 5000°P (3000°C). As the impurities are oxidized, they combine
with the flux to create the floating slag. After 45 minutes, the content
of the steel is tested and the steel is poured through the taphole into a
ladle. The taphole holds back the slag, which is later emptied for dis-
posal. As the steel is tapped, alloys are added to the ladle.
The off-gases from the BOP process carry a very heavy particulate
load. Figure VI-2 and Table VI-10 describe the concentration and size
distribution of the particulates. Gaseous emissions include about 0.2 lb/
ton of product (0.1 kg/MT) of fluorides. All EOF processes have particu-
late emission controls; they are electrostatic precipitators, venturi
scrubbers, and baghouses. Excess air is often used to cool the gases
before they reach the controls.
VI-19
-------
f. Electric Arc
An electric-arc furnace is a refractory-lined cylinder, capable of
handling from 20 to 400 tons. On the larger, newer furnaces the roofs
are removable for charging, and doors in the front are for small addi-
tions and charging the smaller furnaces. Three carbon electrodes are in-
serted through the roof to just above the charge, and an arcing electric
current supplies the heat. Sometimes oxygen is lanced into the furnace.
Tapholes for discharging the steel are either in the bottom for gravity
draining or above the charge requiring the furnace to be tilted.
Once the roof of the furnace is swung out of the way, it is charged
with scrap and pig iron. The roof is then replaced and the electrodes
are lowered. Electricity arcs between metal and the electrodes, creating
an intense heat. After the charge melts, the flux is added and an oxi-
dizing slag is formed. During this process, oxygen may be lanced across
the furnace. High-quality steel requires two slag-forming periods. Con-
sequently, the initial slag is tapped off and more flux is quickly added
to form a second slag under a slightly positive pressure. The steel is
not oxidized further in the second slag period. When the analysis of the
steel is correct, the furnace is tipped and the steel poured out.
The emission gases include oxygen, carbon dioxide, carbon monoxide,
nitrogen and fluorides. Particulate emissions are described in Figure
Vl-2 and Table VI-10. Baghouses are the control device of choice for the
small particles generated by the process.
VI-20
-------
3. Prior Evidence of Disease Associated with Industry
Several investigators have linked the steel industry to various
health effects, including neoplasms. These studies have considered both
occupational and environmental exposure. Several studies will be dis-
cussed here, while a more encompassing review may be found in Appendix H.
Cecilioni has performed several studies relating cancer mortalities
to the presence of the steel industry in Hamilton, Ontario [1,2]. The
cancers studied were respiratory, gastrointestinal, and genitourinary.
He found that:
"A review of the mortality rates for cancer for the years
1966 to 1970 in Hamilton reveals a considerably higher
death rate from cancer in Hamilton than in the less
industrialized city of Ottawa. The highest rate (65 per
100,000) occurred in the proximity of the steel mills,
compared with the death rates (23 and 12 per 100,000)
farther distant." ([2], p. 153)
Since the steel industry produces fluoride contamination and several
studies have implicated fluoride in cancer, Cecilioni analyzed the area
for fluoride. He found fluoride in vegetation, especially in the steel
mill area. In addition, skeletal fluorosis in cancer patients was 3 to
5 times the normal level.
Although Cecilioni's work does not unequivocally prove that the
fluoride from steel mills causes cancer in the adjacent community, his
findings are very strong and fulfill several important criteria in
establishing causality in suspect industry-disease relationships. These
criteria are: (1) dose-response (higher mortality rates around steel
mills); (2) identification of a specific industrial agent (fluoride);
(3) measurement of excessive levels of the emission in the environment;
(4) elucidation of a potential pathway to the affected population; and
(5) impact on the affected population (e.g., skeletal fluorosis).
[1] Gecilioni, V. A., Lung cancer in a steel city, its possible relation
to fluoride emissions, Fluoride 5(4):172-187, 1972.
[2] Cecilioni, V. A., Further observations on cancer in a steel city,
Fluoride 7 (3):153-156, 1974.
VI-21
-------
Lloyd et al. studied 50,072 steelworkers. in seven U.S. plants and
analyzed the relationship between work in specific areas of industry and
death from all causes during the period 1953 to 1961 [3-8]. The areas
were studied for observed and expected deaths from malignant neoplasms,
vascular CNS lesions, and heart diseases. The authors conclude:
"Two exceptions which did not appear to be due to selection
for health were the excess mortality from vascular lesions
of the central nervous system for men employed in sheet
finishing and shipping and excess mortality from malignant
neoplasms of the respiratory system for nonwhite coke plant
workers. It is suggested that these differences may be due
to factors in the working environment."([6], p. 157)
Using the steelworker mortality data of Lloyd et al., Mazumdar et
al. calculated an index of cumulative exposure for coke oven workers [9].
They concluded that the excess of respiratory cancer observed in this
group is related, as expected, to both level of exposure and length of
time exposed. They conclude that:
"Even though the specific carcinogenic element or elements
in the coke oven environment are not known at the present
time, increasing levels of exposure to coal tar pitch
volatiles are related to an increased risk of dying from
malignant neoplasms. Further coal tar pitch volatile
measurements need to be made and analyses designed which
will specify the dose-response relationship more precisely."
[3] Lloyd, J. W., and A. Ciocco, Long-term mortality study of steel-
workers. I. Methodology, J. Occup. Ned. 11(6):299-310, 1969.
[4] Robinson, H., Long-term mortality study of steelworkers. II.
Mortality by level of income in whites and non-whites, J. Occup.
Ned. 11(8):411-416, 1969.
[5] Redmond, C. K., E. M. Smith, J. W. Lloyd, and H. W. Rush, Long-term
mortality study of steelworkers. III. Follow-up, J. Occup. Med. 11
(10):513-521, 1969.
[6] Lloyd, J. W., F. E. Lundin, Jr., C. K. Redmond, and P. B. Geiser,
Long-term mortality study of steelworkers. IV. Mortality by work
area, J. Occup. Med. 12(5):151-157, 1970.
[7] Lloyd, J. W., Long-term mortality study of steelworkers. V. Respira-
tory cancer in coke plant workers, J. Occup. Med. J3(2):53-68, 1971.
[8] Redmond, C. K., A. Ciocco, J. W. Lloyd, and H. W. Rush, Long-term
mortality study of steelworkers. VI. Mortality from malignant neo-
plasms among coke oven workers, J. Occup. Med. 14(8):621-629, 1972.
[9] Mazumdar, S., C. Redmond, W. Sollecito, and N. Sussman, An epidemi-
ological study of exposure to coal tar pitch volatiles among coke
oven workers, J. Air Pollut. Control Assoc. 25 (4):382-389, 1975.
VI-22
-------
In the largest iron and steel works in Japan, six cases of lung can-
cer were found among coke oven workers from 1947 to 1971. An epidemi-
ological survey [10] indicated that the lung cancer incidence was higher
among retired blast furnace and coke oven workers in the iron and steel
works than in the general Japanese male population. The ratio of lung
cancer cases to malignant neoplasia, of all sites was also greater for
blast furnace and coke oven workers.
Benzo[a]pyrene, a known carcinogenic agent, was measured by Tanimura
.[11] and found in quantities from 10- to 60-fold higher at steel produc-
tion sites than in the air of a neighboring industrial town. These
findings document the occurrence of environmental contamination by car-
cinogenic agents involved in iron and steel production.
[10] Sakabe, H., K. Tsuchiya, and N. Takekura, Lung cancer among coke
oven workers. A report to Labor Standard Bureau, Ministry of Labor,
Japan, Ind. Health 13(1-2):57-68, 1975.
[11] Tanimura, H., Benzo[a]pyrene concentrations at Japanese iron and
steel works, Arch. Environ. Health 17(2):172-177, 1968.
VI-23
-------
C. STUDY DESCRIPTION (ACTIVITIES AND METHODS)
1. General Methodology (Hypothesis Development and Test)
As mentioned previously, a two-stage methodology was used for this
study. The first stage consisted of developing hypotheses linking the
steel industry with elevated community mortality at Johnstown, Penn-
sylvania, a site chosen specifically for this purpose. The second stage
involved studying Provo, Utah, a site chosen primarily to test the hy-
potheses previously developed.
The general criteria used in the selection of the hypothesis de-
velopment site are as follows:
• high industrial activity for the steel industry,
• lack of industrial clutter,
• existence of sufficiently large population at risk, and
• significantly higher mortality rates for digestive neoplasms.
The first three criteria above are those used in the selection of the
hypothesis test site. Table VI-11 shows how Johnstown and Provo meet
these criteria.
Table VI-11
SITE SELECTION CRITERIA FOR THE STEEL INDUSTRY
HYPOTHESIS DEVELOPMENT AND TEST SITES
Criterion k
1. High Steel Activity
2. Lack of Industrial
Clutter
3. Sufficient Popula-
tion
4. Elevated Mortalities
(Digestive Neoplasms)
Johnstown
Industrial index: 3082
Employs: 14,000
Some clutter: small firms
Coal outaide city in
Cambria County
42 , 000
Age-adjusted mortality
rates: t« - 120
MF - 75
Provo
Industrial index: 634
Employs: 5,000
Only a few small firms
53,000
Not relevant to test
site selection
VI-2 4
-------
The following sections describe the methods used to establish the
hypotheses in Johnstown and then test them in Provo. A site description
and site-specific industry characterization are included for each site
as well as data collection and analyses.
2. Hypothesis Development—Johnstown, Pennsylvania
a. Site Description
Johnstown, Pennsylvania, a city of 42,000 people, is located in a
narrow Y-shaped valley formed by the confluence of the Stonycreek and
Little Conemaugh Rivers in Cambria County. The downtown business sec-
tion of the community lies in this valley at the point of the confluence
and the residential areas have grown up along the branches of the rivers
and on the plateaus above the town. The city is 70 miles east of
Pittsburgh, 38 miles west of Altoona, and 75 miles north of Cumberland,
Maryland. Johnstown's average annual precipitation is 51 inches. The
normal temperature range is from 10° to 60°F in the winter to 50° to
90°F in the summer. The elevation in the valley is about 1,200 feet,
while elevations on the ridges are around 2,000 feet. Figure VI-3
shows the layout of the Johnstown Air Basin Region.
Johnstown, originally the site of an Indian village, was founded in
1800 by Joseph Johns. In 1834 it was made a borough and was incorporated
as a city in 1889. Johnstown is most noted for its floods. First inun-
dated in 1808, Johnstown has since been flooded 18 times. In 1889 the
Great Flood occurred when the South Fork Dam breached completely and 500
million cubic feet of water inundated the valley, causing the loss of
2,200 lives and $10 million worth of property. In 1936 the rivers rose
and caused a flood that, while killing only 40 people, brought about $40
million worth of property damage. Most recently, in July of 1977,
another flood caused by heavy rains inundated the valley. The damage
from the most recent flood included the closing of the steel plants,
perhaps permanently.
VI-25
-------
ranklln /
CONEMAUGH /
I UPPER YODER
/ TOWNSHIP
Figure VI-3. Communities in the Johnstown Air Basin
VI-26
-------
b. Site-Specific Industry Characterization
The steel industry has a long history in Johnstown due to its de-
posits of iron ore, coal, and limestone. In 1976, about 14,000 people
were employed in the steel industry, 12,000 at Bethlehem and 2,000 at
U.S. Steel. Figure VI-4 shows the location of the steel industry in the
air basin.
The first forge was built in Johnstown in 1809; the first large-
scale ironworks in 1840. Steel was first made by the Cambria Iron Co.
in 1852. The plant changed hands throughout the early years, finally
being purchased by Bethlehem Steel Co. in 1923. The main product of
the plant in the early years was steel rails. The first by-product
works, designed primarily to make coke for blast furnaces, was erected
in 1895 at Johnstown. The Franklin Works, the largest plant in Johns-
town, was built around 1898.
Table VI-12 describes the production capacity history of the Beth-
lehem and U.S. Steel plants in Johnstown.
A steel facilities inventory of the Bethlehem and U.S. Steel plants
in the Johnstown area before July 1977 is found in Table VI-13. Emis-
sion controls for these furnaces and the sinter plants came on line in
the 1960's and 1970's. Although control efficiences indicated in Table
VI-13 were high, overall emissions were excessive as shown in Table VI-14.
The Bethlehem plant in Johnstown discharges alarmingly large amounts
of toxic heavy metals, especially in its water outfalls, where about 26
Ib/day of cadmium, 39 Ib/day of arsenic, and 912 Ib/day of lead are dis-
charged. Although the population of Johnstown is not exposed via a
water route, the Bethlehem Johnstown Works has exceeded Council on Eco-
nomic Priorities criteria for discharges of 18 out of 20 pollutants,
including all the potentially poisonous heavy metals and cyanide. U.S.
Steel's Johnstown Works, a much smaller plant, is not a major polluter;
Bethlehem's Johnstown Works emits 82 times as much soot and dust as the
U.S. Steel plant.
An analysis of industrial polluting sources, including the steel
works, done by Barry Evans of EPIC for the Johnstown area may be found
in Appendix I. This study relates the presence and changes in emission
sources from 1939 to 1977.
-------
BETHLEHEM STEEL
(FRANKLIN WORKS)
River
300-ft contour
Figure VI-4. Location of Steel Industry in Johnstown Air Basin
VI-28
-------
Table VI-12. JOHNSTOWN STEEL INDUSTRY—PRODUCTION CAPACITY HISTORY
Year
1901
1952
1956
1953
1962
1963
1965
1970
U.S. STEEL
STEEL FURNACES
Furnaces
Type
N.O.
OH
EA
OH
EA
OH
EA
OH
EA
N.D.
OH
EA
EA
No.
N.D.
2
2
2
2
2
2
2
2
N.D.
2
2
3
Annual
Capac.*
N.D.
18.900
5,500
19.000
6.0OO
19,000
6.000
19,000
6,000
N.D.
N.D.
N.D.
N.D.
BETHLEHEM STEEL
COKE PLANT
Ovens
Type
BP
BH
BP
BP
BP
BP
BP
BP
BP
No.
160
920
436
436
316
316
316
316
316
Annual
Capac.*
N.D.
2,268,000
2,268,000
1,836.000
1,836,000
N.D.
N.D.
N.D.
BLAST FURNACES
No.
6
N.D.
N.D.
7
6
6
6
5
Annual
Capac.a
65O.OOO
1.674.00O
1.664,000
2. 046,000
2,OOO,OOO
2.000,000
N.D.
N.D.
STEEL FURNACES
Furnaces
Type
BS
OH
OH
OH
OH
N.D.
OH
OH
OH
No.
N.D.
N.D.
24
24
21
N.D.
20
15
8
Annual
Capac.a
600,OOO
225,OOO
2,028,000
2,280,000
2,400.00
2.400,000
2.400,000
N.D.
N.D.
ROLLING 6 PLATE MILLS
Machines
Type
N.D.
N.D.
R
PL
R
PL
N.D.
R
PL
N.D.
R
PL
R
PL
No.
N.D.
N.D.
15
2
16
2
N.D.
N.D.
N.D.
N.D.
14
2
14
2
Annual
Capac.*
300,OOOb
324,000°
4.573,000
39O.OOO
5.740.OOO
40O.OOO
N.D.
5.385.OOO
49O.OOO
N.D.
4.955.OOO
490,000
5,664.500
5OO.OOO
I
to
vO
'Annual Capacity in net tons.
hsteel rails.
Structural steel.
N.D. - No Data BP = By-Product BS = Bessemer BH - Beehive
R - Rolling EA = Electric Arc PL = Plate
OH - Open Hearth
-------
Table VI-13. STEEL FACILITIES INVENTORY FOR JOHNSTOWN, PENNSYLVANIA
U>
O
FACILITY
COKE
OVENS
BLAST
FURNACES
STEEL
FURNACES
SINTER
PLANT
CONTROLS
Number
Tear Installed
Number
Year Installed
Humber /Type/Size
Year Installed
Number
Year Installed
Year Installed
Location s Type
(% Efficiency)
BETHLEHEM STEEL JOHNSTOWN WORKS
316
1922
5
Prior to 190O
8/Open Hearth, Oj-Lanced/180 tons each
2
1973 Blast Furnace - Baghouse
1968 Sinter Plant - ESP (98%)
1972 Sinter Plant - Scrubber
1964 Open Hearth - ESP (98.2%)
1969 Open Hearth - ESP (98.2%)
U.S. STEEL JOHNSTOWN WORKS
0
-
O
-
I/Electric Arc/30 tons
2/Electric Arc/3 tons each
1966 (30 ton)
1924 (3 ton)
1922 (3 ton)
0
-
1966 30-ton Electric Arc - Baghouse (95.0%)
1968 Both 3-ton Electric Arc - to above
Baghouse
-------
Table VI-14. STEEL INDUSTRY: AIR EMISSION INVENTORY (tons/year)
Emission
Particulates
S02
NOX
Johnstown, Pennsylvania
Bethlehem Steel
8,600
13,700
190
-*
U.S. Steel
100
0
120
c. Data Collection and Analysis
. As seen in Table VI-15, the population of the City of Johnstown has
been declining over the years, while the population of the metropolitan
area has increased. The decrease in population is due, in part, to the
lack of industrial diversity which causes young people to move elsewhere
to seek employment. This has resulted in an older population in Johns-
town. Tables VI-16 and VI-17, obtained from 1970, file C, fifth count
census tapes, show the age/race/sex distribution of the population in
Johnstown and some summary demographic statistics for Cambria County.
Table VI-15
TOTAL POPULATION OF JOHNSTOWN, PENNSYLVANIA
AND ITS METROPOLITAN AREA
POPULATION
Year
1920
1930
1940
1950
1960
1970
Johnstown
67,327
66,993
66,668
63,232
53,949
42,476
Metropolitan Area
105,000
112,641
262,822
VI-31
-------
Table VI-16
AGE/RACE/SEX DISTRIBUTION OF 1970 POPULATION OF JOHNSTOWN, PENNSYLVANIA
POPULATION DISTRIBUTION BY AGE, RACE AND SEX
Age Group
0-2
3-14
15-24
25-34
35-44
45-54
55-64
65-74
75 and over
Total
White Male
787
3,957
2,727
1,615
1,885
2,698
2,204
1,340
877
18,090
White Female
762
3,852
3,384
1,864
2,213
3,211
2,863
2,245
1,259
21,653
Nonwhite Male
97
459
253
92
95
103
158
51
55
1,363
Nonwhite Female
75
331
281
157
138
166
133
45
44
1,370
All ages, WM + WF + NWM + NWF = 42,476
The areas used for hypothesis development, those labeled Exposed and
Unexposed on Table VI-17, were Johnstown, Dale, East Conemaugh, and
Franklin compared to Cambria County outside the air basin. The demo-
graphic indices for these areas matched within 10%, except for median
age. The difference in ages should not be a confounding factor in our
analysis, since we are using age-adjusted mortality rates for our com-
parison.
Development of a plausible hypothesis for health effects requires a
plausible pathway for transport of atmospheric pollutants from the plant
to the population at risk. This was essentially a meteorological prob-
lem, for which two approaches will be discussed. The first approach
relied on investigations published in the scientific literature. The
second approach made use of available modeling methodologies to calcu-
late a theoretical average concentration of various pollutants at various
distances from the factory.
VI-32
-------
Table VI-17. SUMMARY DEMOGRAPHIC STATISTICS, 1970 CENSUS, FOR CAMBRIA COUNTY, PENNSYLVANIA
AREA
Johnstown
Browns town
Daisytown
Dale
E. Conemaugh/Franklin
East Taylor Township
Ferndale
Geistown
lora in/Stony Creek Township
Lower Voder Township
Middle Taylor Township
Southnont
Upper Yoder Township
Hestfflont
Exposed'
Unexposed^
Median
Age
39.5
29.5
29.5
29.5
39.5
29.5
39.5
29.5
29.5
29.5
29.5
39.5
29.5
39.5
39.5
29.5
Median
Family
Income
8,500
8,500
8,500
8,500
8,500
8,500
9.500
11,000
9,500
8,500
8,500
ll.OOO
11,000
13,500
8,500
8,500
% In
Same House
as in 1965
67.1
79.9
76.7
66.2
73.3
69.6
70.2
70.3
65.5
84.2
81.4
75.6
61.7
68.0
67.5
71.4
%
White
93.6.
99.4
99.8
98.9
89.9
97.8
100.0
100.0
99.9
96.9
96.9
99.4
97.8
100.0
93.5
99.6
% With
Less Than
1 Person
per ROOB
94.7
89.5
91.1
97.4
91.6
94.1
97.7
96.8
94.3
86.3
87.9
95.7
96.6
98.9
94.7
91.9
* Com-
pleting
High
School
4O.1
44.8
33.5
48.4
38.2
46.5
60.4
59.0
53.3
36.9
34.2
74.6
69.5
75.8
40.4
42.1
t
Female
54.2
52.9
47.1
54.4
52.2
50.5
54.4
51.6
51.5
50.6
50.4
53.0
51.6
53.3
54.1
50.8
% Native-
Born of
Native
Parents
74.9
66.0
74.7
87.1
70.6
80.7
85.7
81.3
85.1
67.9
75.6
81.2
89.0
75.4
75.2
80.7
I
Ul
Ul
'Exposed - Johnstown, Dale, E. Conemaugh/Franklin
^Unexposed - Cambria County, except Johnstown Air Basin
(Johnstown plus boroughs and townships above = Air Basin)
-------
The two reports in the scientific literature [12,13] were parts of
an integrated study performed by Pennsylvania State University's Depart-
ment of Meteorology. The study attempted to correlate meteorological
variables with urban air pollution data in Johnstown. Air quality data
on particulates and meteorological parameters (wind speed and direction)
were collected during the initial phase of the study.
In the first study reported by Panofsky and Prasad, ,a diurnal vari-
ation in both vertical mixing potential (measured by fluctuation of the
bivane) and pollution (as measured by soiling tape transparencies) were
reported for clear skies, with greatest particulate concentrations oc-
curring at night. This pollution was attributed primarily to low-level
sources. Concerning the correlation of pollution from the Franklin
Works with meteorological data, the authors note that,
"...it is likely that, on a clear day with east or
northeast winds, fumigation of particulates from the
Franklin Works may become important. However, east-
erly winds are usually accompanied by thick clouds,
often with precipitation." [12]
A mathematical model invoked by Panofsky and Prasad demonstrates,
however,
"...that the huge Franklin Works do significantly
affect the air pollution at Johnstown, but only in
the case of the relatively infrequent east winds. Of
course, the low-cost residential areas to the north-
east and east of.the Franklin Works encounter serious
air pollution problems." [12]
The second investigation was carried out by a student at Penn State
University who performed a meteorological experiment using a zinc cadmium
sulfide tracer to chart the wind flow in the Stony Creek and Conemaugh
River Valley [13] . The purpose of the tracer study was to attempt to
correlate meteorological variables with urban air pollution.
[12] Panofsky, H. A., and B. Prasad, The effect of meteorological factors
on air pollution in a narrow valley, J. Appl. Meteorol. 6:493-499,
1967.
[13] Smith, D. B., Tracer study in an urban valley, J. Air Pollut. Con-
trol Assoc. Jfl (9):600-604, 1968.
VI-34
-------
In a previous study [14] it was concluded that air pollution was
greatest at night when the sky was clear, winds were calm, and vertical
ventilation was weakest. Therefore, the subsequent tracer study was
performed on calm, clear autumn nights. As the map in this chapter
shows (see Figure VI-4), Johnstown is located at the confluence of the
Little Conemaugh River and Stony Creek where they merge to form the
Conemaugh River. The predominant nocturnal winds under the stated
conditions tend to follow typical nocturnal drainage flow from Stony
Creek to the north over Johnstown. Thus, the point of release for
the tracer was in the middle of Stony Creek, south of the confluence of
the rivers. Figure VI-5 illustrates the approximate tracer concentra-
tion isopleths as reported by Smith [13]. As can be noted in the
abrupt stoppage of tracer flow toward the northeast, the Franklin Works
forms an effective heat island which becomes a barrier to tracer flow
as material is carried aloft by thermal effluent from the steel plant.
The implications of these two studies, from the standpoint of the
present epidemiologic investigation, are twofold. First, the effect of
pollution from the Franklin Works should be most severely felt by resi-
dents of the Boroughs of East Conemaugh and Franklin, located east and
northeast of the plant. In addition, emissions from the Franklin Works
would be expected to significantly affect air quality in the City of
Johnstown to the west and south.
The air quality data available for Johnstown were quite complete
(see Table VI-18). The Pennsylvania Department of Environmental Re-
sources had air quality reports for trace metals as well as criteria
pollutants from six sampling stations throughout the Johnstown Air Basin.
In addition, a special study they had done of the area [15] produced
data on benzo[a]pyrene (BaP) and benzene-soluble organics (BSO) through-
out the air basin.
[14] Pennsylvania State University, Department of Meteorology, Meteoro-
logical Effects on Air Pollution at Johnstown, Pa., Interim Report
1964-1965.
[15] Pennsylvania Department of Environmental Resources, Johnstown Air
Basin Air Quality Study, Johnstown, Pa.
VI-35
-------
Source: Reference [13]
Figure VI-5
Tracer Concentration Jsop2eths
VI-36
-------
Table VI-18
AIR QUALITY DATA FOR THE JOHNSTOWN, PENNSYLVANIA, AIR BASIN (ARITHMETIC MEANS & STANDARD 'DEVIATIONS)
I
u>
PARAMETER
(unit): years data collected
TSP
(Ug/a3):1972-1976
Dnstfall
(ton/Ml2 /«o) : 1972-1976
Sal fate
(Ug/B3): 1972-1976
Nitrate
(Ug/n3): 197 2-1975
Fluoride
(Ug/B3) 11972,1975
Chloride
(Mg/B3):1973
BSO
(Vg/B3)il975
BaP
(ng/B3) 1 1975-1976
Lead
-------
The basic trend for all pollutants seemed to be that the areas high
on the valley walls and ridges, as exemplified by Westmont, were not
exposed to high emission levels; while the areas closest to the Franklin
Works, Franklin and East Conemaugh, were most highly exposed. The rest
of the valley seemed to have uniformly distributed pollutant levels.
This is consistent with a transport analysis done for this study by
Barry Evans of EPIC. A more detailed description of this analysis plus
concentration estimates may be found in Appendix I.
Water quality data were not considered for this analysis, not be-
cause there was no pollution caused by the steel industry, but because
the water supply for the Johnstown area and the downstream communities
comes from sources unaffected by steel industry effluents.
Mortality data from 1965 through 1975 were used for the Johnstown
analyses. They were obtained on tape from the Pennsylvania Department
of Health.* The following items were coded on the tape:
State file number
Place where death occurred
Residence of decedent
Hospital where death occurred
Cause of death
Place of accident (if accidental death)
Date of death
Sex of decedent
Race of decedent
Date of birth of decedent
Age of decedent
If an autopsy was performed
Conversations with Mr. Ed Digon of the Pennsylvania Department of
Health, Division of Epidemiologic Research, revealed a possible resi-
dency-reporting problem on death certificates from the Johnstown Air
*Basic data for use in this study were supplied by the Pennsylvania
Department of Health, Harrisburg, Pennsylvania. Any analysis, inter-
pretation, or conclusion based on these data is solely that of Enviro
Control, Inc., and the Pennsylvania Department of Health specifically
disclaims responsibility for any such analysis, interpretation, or con-
clusion.
VI-38
-------
Basin. A study done by them for a single year of data showed an 18%
overreporting of Johnstown as the place of residence of the decedent.
To correct for this potential problem, we obtained actual street ad-
dresses from about 900 hard copy death certificates which were selected
by cause of death. After using an R. F. Polk Directory and detailed
street maps to verify actual residence, we found the error to be only
9% over the 11 years of data we examined.
We had hoped to obtain morbidity data from the County Tumor Registry
from the Conemaugh Valley Memorial Hospital. Unfortunately, before our
request had been processed the area was flooded and the Tumor Registry
staff was unable to provide the requested information. Due to these
unfortunate circumstances, all mortalities from malignant neoplasms were
calculated from death certificate data. No incidence or severity data
were available.
Average annual age-specific mortality rates for all age groups and
average annual age-adjusted mortality rates for ages 35-74 were calcu-
lated per 100,000 population, as described in Appendix J. The age-
adjusted mortality rates for the exposed (Johnstown, Dale, East Cone-
maugh and Franklin) and unexposed (Cambria County except the Johnstown
Air Basin) areas were compared using the Mantel-Haenszel test [16]. The
results of these calculations can be seen in Table VI-19.
[16] Mantel, N., and W. Haenszel, Statistical aspects of the analysis of
data from retrospective studies of disease, J. Nat. Cancer Inst.
22(4):719-748, April 1959.
VI-39
-------
Table VI-19.
COMPARISON OF AGE-ADJUSTED MORTALITY RATES FOR JOHNSTOWN, PENNSYLVANIA
(using the Mantel-Haenszel x2 test)
H
I
CAUSE OF DEATH (ECI 1)
Infectious t parasitic diseases (1)
Cancer, oral, pharynx £ esophagus (2)
Cancer, stomach (3)
Cancer, duodenum £ snail intestine (4)
Cancer, colon £ rectum (5)
Cancer, liver t bile ducts (6)
Cancer, pancreas (7)
Other digestive cancers (8)
Cancer, respiratory organs (9)
Cancer, genitourinary organs except bladder (10)
Cancer of the bladder (11)
Other cancers except leukemia (12)
Neoplasms of lymphatics(13)
Endocrine, nutritional, metabolic t blood disease (14)
Mental £ nervous system diseases (IS)
Rheumatic heart disease (16)
Hypertensive diseased?)
Cardiovascular diseases (18)
Cerebrovascular diseases (19)
Respiratory diseases (20)
Bronchitis, emphysema & asthma (21)
Pneumoconiosis(22)
Other digestive diseases(23)
Diseases of stooach £ duodenum(24)
Appendicitis, hernia £ others(2S)
Gastroenteritis, noninfectious(26)
Diseases of liver, gallbl&dder t pancreas (27)
Diseases of urinary systea(28)
Diseases of the genital organs(29)
Diseases of pregnancy £ childbirthUO)
Diseases of the skin (, nusculoskeleton £ ill-
defined diseasesOD
Congenital anomalies(32)
Perinatal diseases(33)
Accidents, poisons £ violence (34)
White Hales
Mortality Rate
Exposed
29
33
4O
0
64
10
32
0
199
40
17
73
52
78
19
32
36
1,501
194
126
73
36
3
21
22
0
88
33
8
0
19
5
O
116
Unex posed
23
20
26
O
49
8
20
0
94
30
14
48
30
44
13
20
13
946
116
116
45
155
2
15
19
1
56
23
3
0
12
3
0
120
Relative
Risk
1.19
1.74
1.60
—
1.31
1.31
1.59
~
2.13
1.40
1.23
1.49
1.65
1.82
1.43
1.59
2.68
1.87
1.72
1.07
1.60
—
1.36
1.32
1.14
~
1.58
1.44
2.82
—
1.62
1.61
—
—
Level of
Signif.
NS
.025
.05
—
MS
NS
MS
—
.005
NS
NS
.025
.025
.005
NS
NS
.005
.005
.005
NS
.005
—
NS
NS
NS
—
.005
NS
NS
—
NS
NS
—
—
White Females
Mortality Rate
Exposed
10
6
11
1
42
14
9
6
26
44
3
111
25
61
10
36
27
645
127
40
12
0
O
5
20
1
34
18
1
0
12
4
1
38
Unexposed
8
0
7
O
34
9
9
I
12
49
2
85
23
76
12
27
19
482
105
23
9
0
1
5
14
0
27
7
2
1
11
4
0
29
Relative
Risk
1.34
11.93
1.55
—
1.20
1.42
—
4.26
2.22
—
1.18
1.30
1.17
—
—
1.30
1.38
1.40
1.21
1.64
1.31
—
—
—
1.43
—
1.28
2.58
—
—
1.10
—
—
1.34
Level of
Siqnif .
NS
.01
NS
—
NS
NS
~
.05
.005
—
NS
.025
NS
—
—
NS
NS
.005
NS
.025
NS
—
—
"
NS
—
NS
.005
~
—
NS
—
—
NS
NS - Hot significant.
-------
3. Hypothesis Test—Prove, Utah
a. Site Description
Prove, Utah, a city of 53,000 people, is. located in the Utah Valley
on the eastern shore of freshwater Utah Lake in Utah County. Provo is
the largest city in the county, but several other smaller cities are
scattered along the eastern shore of the lake. The city of Provo is 45
miles south of Salt Lake City. Provo's average annual precipitation is
11 inches. The normal temperature range is from 19° to 38°F in the
winter, to 55° to 90°F in the summer. Utah Valley is bordered on the
east by the Wasatch Mountains (11,000 feet) and on the west by the Lake
Mountains (10,000 feet). The elevation in the valley is about 4,500
feet. Figure VI-6 shows the layout of the Provo area.
Provo was founded in 1849 by Mormon settlers from Salt Lake City.
At first agriculture was the primary industry, with the good water
sources from the eastern mountains being of tremendous assistance for
irrigation. Shortly after its founding, Provo began to start manufac-
turing various products, mostly connected with agriculture. The Knight
Woolen Mill was one ,of the largest firms at this time. In 1875 Brigham
Young Academy (now Brigham Young University) was founded. The univer-
sity has been and remains today one of the major contributors to
Provo's economy.
b. Site-Specific Industry Characterization
The steel industry has a long history in the State of Utah. Shortly
after the first Mormon pioneers arrived in the valley of the Great Salt
Lake in 1847, an exploring expedition discovered significant iron and
coal deposits in the area which now surrounds Cedar City, in the south-
ern part of Utah.
Plans were laid immediately for starting an iron industry. A colon-
izing company left Provo in 1850 and journeyed by ox team to the area
near what is now Cedar City, where they built Utah's first blast furnace.
In 1852,the first pig iron made west of the Missouri River was produced
in this furnace.
VI-41
-------
(::¥#**: MOUNTAINS ::tettmttmtt
Figure VI-6. Utah Valley
VI-4 2
-------
In the meantime, private enterprise organized the Deseret Iron Com-
pany to provide capital for development of the iron industry. The
Deseret Iron Company purchased the original Utah blast furnace and began
to enlarge facilities. But despite their efforts the venture was not a
financial success.
Subsequent attempts to start an iron industry in southern Utah were
made in 1854 and 1868. The latter continued in production for 15 years/
and the pig iron was used to make stoves, grates, iron pots, frying pans,
flat irons, buckets, and other castings. But this operation, like those
before it, was not a financial success and ceased operation in 1883.
The first financially successful operation to use southern Utah iron
ores was begun in 1922 when the Columbia Steel Corporation started con-
struction of a 450-ton blast furnace at Ironton,-south of Provo. The
first iron was produced at Ironton on April 30, 1924. Pig iron pro-
duced at the Ironton furnace was intended primarily for consumption in
Columbia Steel Corporation's steel plants at Pittsburg and Torrance in
California. The balance of the output was sold to foundries throughout
the western states. This arrangement continued following the purchase
of the Columbia Steel Corporation by the United States Steel Corpora-
tion in 1930.
In 1941-1942, the Federal Government decided to build additional
steel production facilities in the West as a precaution against possible
closing of the Panama Canal by enemy attacks.
A new steel mill was designed and built without charge or fee by
U.S. Steel's Columbia Steel Company. The new mill was constructed in
Utah because of adequate transportation; minimum distance from sources
of iron ore, coal, limestone, and dolomite; ready access of sufficient
fresh water; and safe distance from the vulnerable Pacific Coast. This
mill was named the Geneva Plant after the small summer resort of Geneva
on the shore of Utah Lake. The first pig iron was smelted at the new
plant in January of 1944, followed by the first open-hearth steel a
month later.
VI-4 3
-------
At the close of World War II, the War Assets Administration offered
to sell the Geneva Plant to private industry. U.S. Steel bid for the
plant and took over operation of the Utah facilities on June 19, 1946.
Since 1946 many facilities have been added to the Geneva Works, nearly
doubling its annual steelmaking capacity to 2,500,000 ingot tons per
year, and adding a variety of new products such as hot-rolled strip and
sheets, welded steel pipe, and synthetic nitrogen chemicals.
Located six miles northwest of Provo, the Geneva Works is one of the
larger steelmaking facilities in the west. Geneva's principal products
are plates, hot-rolled sheets and coils, structural shapes, welded steel
pipe, pig iron, coke, blast furnace and open-hearth slag products, coal
chemicals, and nitrogen products for fertilizer and industrial use.
A steel facilities inventory of the Geneva Works is found in Table
VI-20. Emission controls for the plant, begun in the 1950"s (see Table
VI-20), included electrostatic precipitators (ESP) and new coke oven
doors and jambs to reduce fugitive emissions. The emissions for the
plant are summarized in Table VI-21.
Air pollution control measures were not taken without public pres-
sure, however. In the late fifties, a group of farmers complained to
the Utah Department of Agriculture that some of their livestock were
afflicted with swollen joints and loss of teeth. An investigation by
the State discovered that this condition was due to a type of fluoride
poisoning caused by excessive airborne fluoride emissions coming from
the blast furnaces and open hearths at the Geneva Wprks. U.S. Steel
eventually agreed to pay damages to the farmers and undertook the
building of a wet scrubbing system at the furnaces to supplement the
electrostatic precipitators that had been operating since 1956.
Geneva's water pollution control program is quite good; its dis-
charges exceed Council on Economic Priorities (CEP) criteria for only
6 out of 20 pollutants.
VI-44
-------
Table VI-20. STEEL FACILITIES INVENTORY FOR PRQVO, UTAH
FACILITY
' COKE '
OVENS
BLAST
FURNACES
STEEL
FURNACES
SINTER
PLANTS
CONTROLS
Number
Year Installed
Number
Year Installed
Number/Type/Size
Year Installed
Number
Year Installed
U.S: STEEL GENEVA WORKS
252
1944
3
1944
10/Open Hearth/340 tons each
1944
2
1941
1946 Blast Furnace — Inertial Dust Catcher,
Counter Current Scrubber, ESP
1954 Sinter Plant - ESP
1955 Blast Furnace - Orifice Scrubber
1955 Open Hearth - ESP
1962 Sinter Plant - Scrubbers
1962 Open Hearth — Wet Scrubbers
1971 Coke Oven - New Doors, Jambs, etc.
1974 Open Hearth - More Power to ESP
1975 Open Hearth - Collector Modified
1976 Coke Oven — Practices Modified
Table VI-21. STEEL INDUSTRY, PROVO, UTAH: AIR EMISSION INVENTORY
(tons/year)
EMISSION
Particulates
S02
NOX
U.S. STEEL, PROVO,
UTAH
4,700
320
—
VI-4 5
-------
c. Data Collection and Analysis
As seen in Table VI-22, the population of Utah County has been in-
creasing steadily over the years. This has been the trend since its
founding. Utah, in general, has rapid population growth, and Utah
County reflects this trend. Table VI-23 shows the age/race/sex distri-
bution of the population in Provo as determined from 1970, file C, fifth
count census tapes. Summary demographic statistics for Utah County com-
munities and the control site, Ogden, which is in Weber County, are
shown in Table VI-24.
Table VI-22. UTAH COUNTY: HISTORICAL AND FORECAST POPULATIONS
COMMUNITY
American Fork
Lehi
Lindon
Or em
Pleasant Grove
Provo
Spanish Fork
Springville
Utah County (uninc.)
Other Communities
TOTALS
POPULATION
1940
3,333
2,733
587
2,914
1,941
18,071
4,167
4,796
11,001
7,839
57,382
1950
5,126
3,627
801
8,351
3,195
28,937
5,230
6,475
11,592
8,578
81,912
1960
6,373
4,377
1,150
18,394
4,772
36,047
6,742
7,913
12,056
9,437
106,991
1970
7,713
4,659
1,644
25,729
5,327
53,131
7,284
8,790
12,583
10,916
137,776
1975
10,803
5,597
2,161
35,292
6,566
59,729
8,404
10,764
14,830
13,851
167,997
1980
Est.*
12,689
6,729
2,914
46,767
8,127
67,103
9,539
12,891
17,568
17,900
202,227
Projections based on the 208
Association of Governments.
Water Quality Study for Mountainlands
The study area demographics matched those of Ogden within 15% except
for three indices in Provo that were outside this limit. These indices-
median family income, % in same house as in 1965, and % completing high
school—all reflect the presence of Brigham Young University, whose
26,000 enrollment is bound to have an effect on such factors. The use
of health data from decedents aged 35 to 74 only should balance this at
VI-46
-------
least in part. The other Utah County communities studied were more
closely matched to Ogden demographically.
Table VI-23. AGE/RACE/SEX DISTRIBUTION OF 1970 POPULATION OF PROVO, UTAH
POPULATION DISTRIBUTION BY AGE, RACE AND SEX
Age Group
0-2
3-14
15-24
25-34
35-44
. 4.5-54
55-64
65-74
75 & over
All Ages
White Male
1,785
4,136
10,502
3,429
1,450
1,337
1,298
806
396
25,139
White Female
1,562
4,029
12,880
2,570
1,441
1,664
1,427
998
690
27,261
Nonwhite Male
23
40
130
113
9
11
3
6
0
335
Nonwhite Female
36
17
226
69
26
0
13
0
0
387
All Ages, WM -I- WF + NWM + NWF = 53,122
Table VI-24. SUMMARY DEMOGRAPHIC STATISTICS, 1970 CENSUS
FOR COMMUNITIES IN UTAH S WEBER COUNTIES, UTAH
TOWN
American Fork
Lehi
Lindon
Or em
Pleasant Grove
Provo
Spanish Fork
Springville
Ogden
Median
Age
23.0
21.0
20.0
20.0
21.0
21.0
23.0
23.0
23.0
Median
Family
Income
6,500
8,500
9,500
9,500
8,500
7,500
8,500
8,500
9,500
% In Same
House As
in 1965
63.3
62.9
62.7
53.0
70.0
30.2
62.7
62.4
52.0
%
White
100.0
100.0
100.0
99.1
100.0
98.6
99.9
99.5
95.5
% With
Less Than
1 Person
per Room
89.5
86.4
76.6
86.8
85.4
89.5
90.9
89.4
92.6
% Com-
pleting
High
School
71.8
70.3
64.5
77.6
64.4
78.7
68.3
70.5
60.7
»
Female
51.1
48.9
49.5
50.1
50.4
52.0
50.6
51.0
51.3
« Native-
Born of
Native
Parents
94.5
93.1
95.3
91.4
93.1
87.1
92.6
92.3
86.2
VI-47
-------
Meteorology in the Geneva-Provo-Orem area is regulated primarily by
the mountains to the north and east and Utah Lake which extends 10 miles
to the west. Winds tend to blow out of the mountains and over the lake
as is evidenced by the wind rose for Provo presented in Figure VI-7.
Data on atmospheric stability were not readily available for Provo or
Or em.
Lines indicate percent
of time wind blew from
indicated direction
Figure VI-7. Wind Rose for Provo, Utah (1966-1969 data)
A transport analysis using equation (4) in Appendix A was attempted
to delineate population at risk and estimate dosage of particulate emis-
sions to which they have been exposed. Since neither STAR data nor
stability information was readily available for Orem or Provo, a STAR
tabulation from Salt Lake City was used. During plotting of isopleths
of calculated long-term concentrations of particulate matter, it became
obvious that Salt Lake City's meteorology was inappropriate. The moun-
tains and lake in Salt Lake are oriented differently from those near
Orem, such that prevailing winds in Salt Lake blow at right angles to
those in Provo-Orem. Thus, the modeling effort was abandoned for this
site.
VI-48
-------
The air quality data for Provo was not as complete as that available
for Johnstown. The State of Utah apparently does not have the extensive
monitoring programs that Pennsylvania does. We were able, however,
through the efforts of Justice Manning at EPA, Research Triangle Park,
to obtain BaP data from a special study of the Geneva Works [18]. The
data available are presented in Table IV-25.
Table IV-25. AIR QUALITY DATA FOR RELEVANT SITES IN UTAH
SITE
Geneva
Lehi
Lindon
Or em
Pleasant Grove
Provo
Ogden
TSP (uq/mj)
Ann. Geo. Mean Max. 24 -hr Avg.
73 330
67 302
83 480
78 374
70 531
88 481
79.3 378
SITE & LOCATION WITH REGARD
TO COKE BATTERIES
1 2.0 km northwest
2 2.7 km northwest
3 2.4 km northwest
4 1.8 km north
5 1.3 km northeast
6 2.4 km southeast
7 4.0 km northwest
8 2.6 km south
9 30.0 km south N
,«.,«„, ^ fControls
10 20.0 km north /
BaP
(ng/m3)
2.1
3.8
3.2
2.4
3.1
1.6
2.1
1.5
0.1
0.8
NO2
(ppm)
.028
.025
TSP
(ug/m3 )
124
235
180
187
136
275
139
N.D.
N.D.
N.D.
[18] Stanford Research Institute, Human Population Exposures to Coke
Oven Atmospheric Emissions, prepared for EPA Contract No. 68-01-4314,
by SRI, August 1977.
VI- 49
-------
Utah Valley is a broad valley prone to inversions, ' particularly in
the winter months. While the emissions tend to be more dispersed here
than in the narrow Johnstown Valley, there is still a noticeable haze
in the valley during the inversions.
Mortality data from 1965 through 1975 were used for-the Provo analy-
ses. They were obtained on tape from the Utah Division of Health.* The
following items were coded on the tape:
Sex of decedent
Date of death
Race of decedent
Date of birth
Location of death
Marital status of decedent
Residence of decedent
Cause of death
If an autopsy was performed
If a violent death
We were able to obtain, on tape, the cancer registry data for Utah**
from 1966 through 1975 from Dr. Joseph L. Lyon of the University of
Utah', Department of Family and Community Medicine. We were informed
that this tape contained residence information by zip code, which would
have enabled us to examine the incidences and mortalities at a more de-
tailed level than the county. Unfortunately, after obtaining the tape,
we found that zip code had been coded only for two years of data. There
was not enough data for us to analyze successfully. We were thus forced
to look at these data on a county level only, comparing Utah County to
Weber County; however, we were able to compare incidence and mortality
in bhese two counties.
Average annual age-specific incidence and mortality rates for all
age groups and for ages 35-74 were calculated per 100,000 population as
*Basic data used in this study were supplied by the Utah Division of
Health, Salt Lake City, Utah. Any analysis, interpretation, or con-
clusion based on these data is solely that of Enviro Control, Inc.
**These data were supplied by the Utah State Cancer Registry. Any
analysis, interpretation, or conclusion based on these data is solely
that of Enviro Control, Inc.
VI-50
-------
described in Appendix J using both the mortality and cancer registry
tapes.
,The age-adjusted incidence and mortality rates for the test sites
were compared to the control areas using the Mantel-Haenszel test as
previously described. The results of these calculations can be seen in"
Tables VI-26 through VI-29.
VI-51
-------
Table VI-26. COMPARISON OF AGE-ADJUSTED MORTALITY RATES* FOR WHITE FEMALES IN UTAH
(Using the Mantel-Haenszel X* Test)
Ul
to
CAUSE
Or DEATH
ItKl lift
1
2
3
4'
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
RATE IN
OGDEN
4
3
9
0
22
4
7
0
13
32
2
70
21
41
23
40
10
302
67
30
12
0
2
5
15
0
34
16
1
1
23
1
O
55
AMERICAN PORK
Rate
7
0
13
0
47
0
55
0
21
39
0
101
10
64
11
42
18
404
133
32
11
0
0
0
11
0
42
11
0
O
116
0
0
63
Relative
Risk
2.15
«
1.95
—
2.19
—
7.25
—
1.34
1.14
—
1.38
—
1.64
—
0.92
2.00
1.34
2.08
0.99
—
—
—
~
~
—
1.O4
—
—
—
4.48
—
—
1.16
Level of
Siqnif .
US
—
NS
—
MS
—
.005
—
NS
NS
—
NS
~
NS
—
NS
NS
NS
.025
NS
—
—
- —
—
—
—
NS
—
—
—
.005
—
~
MS
LEHI
Rate
26
0
13
0
26
12
26
0
12
26
0
105
SO
24
0
35
O
281
104
30
0
0
0
0
0
0
0
0
O
O
77
0
0
27
Relative
Risk
7.20
—
1.45
—
1.16
3.04
3.67
—
—
—
—
1.31
2.49
—
—
—
—
—
1.56
~
—
—
—
~
—
—
—
—
—
—
3.57
—
—
—
Level of
Siqnif.
: NS
~
NS
~
NS
NS
NS
—
—
—
~
NS
NS
—
—
—
—
—
NS
—
—
—
~
—
—
—
--
—
--
—
0.01
—
—
—
LINOON
Rate
,0
0
.0
0
0
0
0
0
0
0
0
0
0
64
0
0
0
110
46
64
0
O
0
0
0
0
0
0
0
0
O
0
0
0
Relative
Risk
—
—
—
—
—
—
—
—
~
~
—
--
—
1.22
—
—
—
—
—
1.66
—
—
—
—
—
--
—
—
~
.
—
—
—
—
Level of
Siqnif.
—
—
—
—
—
—
—
—
—
—
—
—
—
NS
—
~
—
~
—
NS
—
—
—
«
—
—
~
—
—
—
—
—
~
—
OREM
Rate
3
2
5
0
21
0
15
0
0
31
5
81
11
21
20
22
3
314
84
9
9
O
0
5
17
0
25
13
0
O
19
4
O
24
Relative
Risk
—
—
— '
—
—
—
—
—
~
—
2.54
1.13
—
—
—
—
—
1.04
1.28
—
—
—
--
--
1.00
—
—
—
—
—
—
4.33
—
—
Level of
Siqnif.
—
• —
' —
—
—
—
~
—
—
—
NS
NS
—
—
--
—
—
NS
NS
—
—
—
—
—
NS
—
—
—
—
~
—
NS
—
—
Data for 1965 through 1975 froa Mortality tape.
See Table VI-19 for definitions of causes of death by ECI I.
(continued)
-------
Table VI-26 (cont.). COMPARISON OF AGE-ADJUSTED MORTALITY RATES* FOR WHITE FEMALES IN UTAH
M
(Jl
Ul
OUJSE
Of DEATH
-------
Table VI-27. COMPARISON OF AGE-ADUSTED MORTALITY RATES* FOR WHITE MALES IN UTAH
(Using Mantel-Haenszel x* Test)
CAUSE
OP DEATH
(ECI »)t:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IS
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
RATE IN
OGDEN
14
15
IS
0
33
3
19
3
75
30
13
31
36
34
56
25
16
778
101
107
67
2
2
23
20
1
64
21
1
0
94
2
u
159
AMERICAN FORK
Rate
47
0
0
0
43
10
15
0
60
35
0
102
18
60
40
7
7
750
88
82
60
0
0
15
30
O
0
27
0
0
121
0
0
175
Relative
Risk
3.83
—
—
—
1.35
2.68
—
—
—
1.19
—
3.21
—
1.80
—
—
—
—
—
—
—
—
—
—
1.69
—
--
1.23
—
—
1.25
—
—
1.09
Level of
Siqnif.
.01
—
—
—
US
US
—
—
—
NS
."boS
—
NS
—
—
—
—
~
—
—
—
—
—
NS
—
—
NS
—
—
NS
—
—
NS
LEHI
Rate
.0
0
0
0
25
0
0
13
108
16
0
48
16
63
29
0
0
712
70
95
143
0
0
0
9
16
41
45
0
0
109
22
0
165
Relative
Risk
—
—
—
„
«
—
—
5.11
1.44
—
--
1.40
—
1.74
—
—
—
—
--
—
2.10
—
—
—
—
16.97
—
2.09
—
—
1.10
11.75
—
1.06
Level of
Siqnif .
—
—
—
—
—
—
—
NS
NS
—
—
NS
—
NS
—
—
—
—
—
—
NS
—
—
—
—
NS
—
NS
—
—
NS
.025
—
NS
LINDON
Rate
0
0
0
O
0
O
O
0
0
0
0
0
0
0
0
O
O
74
74
0
0
0
0
0
0
O
74
0
0
0
0
O
0
0
Relative
Risk
—
—
—
—
—
—
—
—
~
—
—
—
—
—
~
--
—
—
—
~
~
—
—
—
—
—
.75
—
—
—
—
—
—
—
Level of
Siqnif.
—
—
—
—
—
—
—
—
—
—
~
—
—
—
—
—
—
—
—
—
~
~
—
—
—
—
NS
—
--
—
—
—
-
—
OREH
Rate
13
19
7
O
21
0
32
0
76
31
0
54
37
22
18
19
13
598
99
60
51
4
0
15
18
0
26
15
0
O
39
7
0
87
Relative
Risk
~
1.27
~
—
—
—
1.43
~
0.98
1.02
1.77
0.90
—
—
—
—
—
—
~
—
2.88
—
--
—
--
~
—
--
—
—
2.41
—
—
Level of
Siqnif.
. —
NS
—
—
—
—
NS
—
NS
NS
~
NS
NS
—
—
--
~
~
—
—
~
NS
—
~
--
—
—
—
—
—
—
NS
—
—
Data for 1965 through 1975 from mortality tape.
See Table VI-19 for definitions of causes of death by ECI I.
(continued)
-------
Table VI-27 (cont.). COMPARISON OF AGE-ADJUSTED MORTALITY RATES* FOR WHITE MALES IN UTAH
<
M
I
CAUSE
Or DEATH
-------
Table VI-28. AGE-ADJUSTED INCIDENCE AND MORTALITY RATES FOR WHITE FEMALES IN UTAH*
CAUSE Of DEATH
Site of. NeoplasM
Oral, pharynx, esophagus
Stomach
Snail intestine
Colon
Rectum
Liver
Gallbladder, pancreas
Other digestive
Nasal , larynx
Trachea, bronchus, lung
Other respiratory t intrathoracic organs
Henatopoietic & reticuloendothelial systems
Bone, connective tissue, skin
Breast
Genitourinary
Bladder
Eye, nervous system
Endocrine glands
Ill-defined
Lymph nodes
Unknown
INCIDENCE RATES
Utah
County
4
6
2
34
12
3
13
3
1
10
0
15
16
164
155
14
9
17
0
16
15
Weber
County
10
13
4
43
16
3
18
2
2
10
2
16
25
160
174
7
14
11
0
15
16
Relative
Risk
'
—
--
—
~
—
—
1.17
—
—
—
—
—
1.03
~
2.08
—
1.48
—
1.09
—
Level of
Signif .
~
—
—
~
~
—
—
NS
—
—
--
--
—
NS
~
.05
~
NS
—
NS
—
MORTALITY RATES
Utah Weber
County County
3 4
5 10
1 3
20 20
7 6
3 3
12 14
3 2
0 1
8 10
0 1
11 12
4 7
60 49
52 51
6 3
9 12
3 2
0 0
8 10
13 15
Relative
Risk
~
—
~
—
1.16
—
—
1.17
—
—
—
—
—
1.23
1.02
2.57
~
1.17
~
—
--
Level of
Signif.
~
— -
~
—
NS
--
—
NS
—
~
~
~
—
NS
NS
NS
—
NS
—
—
—
Utah Cancer Registry Data by County, 1966-1975.
-------
Table VI-29. AGE-ADJUSTED INCIDENCE AND MORTALITY RATES FOR WHITE MALES IN UTAH*
i
ui
-j
CAUSE OF DEATH
Site of Neoplasm
Oral, pharynx, esophagus
Stomach
Snail intestine
Colon
RectuB
Liver
Gallbladder, pancreas
Other digestive
Nasal, larynx
Trachea, bronchus, lung
Other respiratory s intrathoracic organs
Henatopoietic & reticuloendotheltal systems
Bone, connective tissue, skin
Breast
Genitourinary
Bladder
Eye, nervous system
Endocrine glands
Ill-defined
Lymph nodes
Unknown
INCIDENCE RATES
Utah
County
45
17
2
36
17
2
20
3
10
ei
2
22
20
3
161
24
15
5
0
27
24
Weber
County
49
17
3
50
21
0
15
1
19
79
3
22
14
3
145
37
9
3
0
17
19
Relative
Risk
---
~
~
—
~
" —
1.22
2.85
~
1.00
~
~
1.40
~
1.10
~
1.75
1.47
—
1.62
1.27
Level of
Signif .
~
~
—
~
~ ' '
~
NS
MS
—
NS
~
—
NS
—
NS
~
NS
NS
~
.05
NS
MORTALITY RATES
Utah
County
19
14
0
22
10
2
19
2
7
76
2
17
7
1
65
11
11
2
0
19
23
Weber
County
24
15
2
30
13
0
15
1
7
71
3
19
5
1
75
20
7
0
0
11
18
Relative
Risk
~
--
--
~
--
~
1.17
1.17
~
1.05
~
—
1.40
—
--
~
1.67
— •
--
1.73
1.25
Level of
Signif.
~
~
—
—
- —
—
NS
NS
~
NS
~
—
NS
~
~
—
NS
—
--
NS
NS
Utah Cancer Registry Data by County for 1966-1975.
-------
D. RESULTS OF ANALYSIS
1. Basis for Initial Causal Hypotheses
The data from Johnstown, Pennsylvania, were analyzed in order to
develop hypotheses relating the steel industry emissions, specifically
airborne emissions, to increased community mortality. We found deaths
from several causes to be significantly elevated; these are listed in
Table VI-30.
Table VI-30
CAUSES OF DEATH WHICH ARE SIGNIFICANTLY* ELEVATED
IN THE EXPOSED AREAS OF THE JOHNSTOWN AIR BASIN
CAUSE OF DEATH
(BCI «)|ICDA Code*
Oral cancers
.(2) (140-150
Stomach cancers
(3), 151
Peritoneal & unspecified digestive cancers
(8) 1 158-159
Respiratory cancer
(9)> 160-163
Breast & other cancers
(12) i 170-174, 190-199, 210-239
Leukemia
(13); 200-209
Endocrine, nutritional, metabolic S blood
disease (14) ; 240-289 '
Hypertensive disease
(17) (400-404
Cardiovascular diseases
(18) )410-458
Cerebrovascular diseases
(19)i430-438
Respiratory diseases
(20) 1460-486, 500- 508, 510-514, 516-519
Bronchitis, emphysema & asthma
(21); 490-493
Diseases of liver, gallbladder £ pancreas
(27); 570-577
Diseases of urinary system
(28); 580- 599
Significantly Elevated?
MM
Yes
yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
No
HF
Yes
No
Yes
Yes
Yes
1 No
No
No
Yes
No
Yes
No
No
Yes
Tp < 0.05
*ICDA Codes (8th Revision)
Vl-58
-------
Five of the six air sampling stations in the air basin were within
the exposed area. Data from these five sites were averaged to give a
"mean" pollution level in the exposed area. The data from the Westmont
site were used as representative values for the unexposed area on the
recommendation of State officials and due to the lack of air quality
data for control areas (a common problem throughout the nation). These
data are shown in Table VI-31.
Table VI-31
AIR QUALITY DATA FOR EXPOSED AND
UNEXPOSED AREAS OF JOHNSTOWN SITE
POLLUTANT
BaP
BSO
Lead
Cadmium
TSP
Sulfate
UNITS
ng/m3
yg/m3
ug/m3
yg/m3
yg/m3
yg/m3
LEVELS
Exposed
10.4
5.7
.890
.005
101
16.0
Unexposed
2.3
2.2
.193
.003
42
11.7
Six of the fourteen causes of death which were elevated (as shown
in Table VI-30) were cancers. As the air quality data show, the levels
of BaP and BSO in the exposed area are two and a half to almost five
times higher than in the unexposed area. The level of particulars was
also elevated two and a half times in the exposed area. BaP is a known
potent carcinogen. The composition of the BSO may be varied, depending
on the coal used for coking, the amount and type of oily scrap fed
through the sintering plant, etc. As discussed earlier in this chapter,
however, many known hazardous and carcinogenic organics are present in
the steel industry emissions, particularly from coking processes. Par-
ticulates have also been shown to be associated with stomach cancer.
j
Sulfur dioxide, particulates, organics, and several trace metals
have been implicated in respiratory diseases including bronchitis.
These pollutants are also elevated in the exposed area. The liver and
VI-59
-------
urinary diseases have been linked with exposure to lead, other trace
elements, and sulfur dioxide. The hypertensive, cardiovascular, and
cerebrovascular diseases have been linked to elevated cadmium levels;
although the cadmium elevation we show between exposed and unexposed
may not be great enough to explain the elevations of these diseases
found' in Johnstown.
From the above analyses, we can postulate the following hypoth-
eses:
For males and females:
• Oral, respiratory, breast, and genitourinary cancers
caused by high levels of BaP and other organics.
• Cardiovascular disease caused by excess cadmium.
For males only (possibly occupational):
• Stomach cancers and leukemia caused by high levels of
BaP and other organics plus particulates.
• Endocrine, metabolic, and blood diseases from elevated
trace metal concentrations.
• Hypertensive and cerebrovascular disease caused by
excess cadmium.
• Bronchitis, emphysema, and asthma caused by elevated
S02 levels.
• Liver disease caused by exposure to organics and trace
elements.
For females only:
• Some digestive cancers caused by exposure to organics.
• Respiratory diseases due to SO2 exposure.
• Urinary diseases due to trace metal exposure.
Nearly all causes of death are elevated in the female population at
risk compared to the control population. Where deaths in the control
populations exceed those in the population at risk, the differences are
negligible. Conversely, the population at risk shows several causes of
VI-60
-------
death which are more frequent than in the control population and which
have more than twofold relative risks. These include cancers of the
oral cavity, pharynx and esophagus, respiratory system, peritoneum and
retroperitoneum and other nonspecific digestive system sites, and dis-
eases of the urinary system.
In males, only two causes of death are more frequent in the control
population than in the population at risk, and both can be reasonably
attributed to employment in,the coal mining industry. In light of
previous studies it is not surprising that respiratory cancer rates are
higher in the population at risk; an indication of probable occupation-
ally related disease. The same reasoning would seem to apply to other
respiratory and cardiovascular diseases. Two relative risks achieve
statistical significance and exceed 2.0. These are for cancers of the
respiratory tract and for hypertensive disease.
2. Hypothesis Test Results
The hypotheses (p. VT-60) were tested using the data from the Provo
site. After analyzing the mortality data for Provo, we found that sev-
eral causes of death were elevated, but generally not significantly.
These are listed in Table VI-32. Since we analyzed data for eight com-
munities exposed to the steel emissions at this site, elevated rates are
based on the number of communities which showed elevations as well as
the number which had significantly elevated rates.
Table VI-33 shows the results of comparing the elevated causes of
death in Johnstown with those causes of death which were elevated in
Utah. We found that there were, indeed, causes of death which were ele-
vated in Johnstown that were also high in Utah. Specifically, deaths
from breast and other cancers were high in both males and females in
Johnstown and also in Utah. Looking at the cancer registry data from
Utah showed this elevation to be due, at least at a county level, to
excess deaths from neoplasms of the breast and endocrine glands in fe-
males and from neoplasms of the nervous system and endocrine glands in
males.
VI-61
-------
Table VI-32
CAUSES OF DEATH WHICH ARE ELEVATED
IN THE UTAH COMMUNITIES NEAR THE GENEVA WORKS
CAUSE OF DEATH
(EC I »);ICDA Code*
Pancreatic cancer
(7); 157
Respiratory cancer
(9) (160-163
Breast & other cancers
(12), -170-174, 190-199, 210-239
Leukemia
(13) ; 200-209
Endocrine, nutritional, metabolic
& blood diseases (14) ,-240-289
Cardiovascular diseases
(18); 410-458
Cerebrovascular diseases
(19); 430-430
Respiratory diseases
(20); 460-486, 500-508, 510-514, 516-519
Skin, musculoskeletal & ill-defined
diseases (31) ,-680-739,780-796
Congenital anomalies
(32), -740-759
NUMBER OF
White Males
# Elevated
-
3
7
3
4
-
-
-
-
4
1 Siqnif.
-
0
2
0
3
-
-
-
-
1
COMMUNITIES
White Females
1 Elevated
3
-
6
3
4
4
6
3
4
-
1 Siqnif.
1
-
0
0
0
0
2
0
3
-
ICDA Codes (8th Revision)
Table VI-33
ELEVATED CAUSES OF DEATH IN JOHNSTOWN WHICH WERE ALSO ELEVATED IN UTAH
ELEVATED CAUSES Of DEATH IN JOHNSTOWN
UTAH RESULT
WHITE MALES & FEMALES I
Cancer, oral, pharynx & esophagus
Cancer, respiratory organs
Breast s other cancers
Cardiovascular diseases
Not elevated
elevated only in NH (not significantly)
Elevated in KM 6 W (significant only in WO
Elevated only in WF (not significantly)
WHITE HALES:
Stomach cancers
Leukemia
Endocrine, nutritional, metabolic 6 blood diseases
Hypertensive disease
Cerebrovascular diseases
Bronchitis, emphysema & asthma
Diseases of liver, gallbladder 6 pancreas
Not elevated
Elevated (not significantly)
Significantly elevated
Not elevated
Not elevated
Not elevated
Not elevated
WHITE FEMALES:
Peritoneal t unspecified digestive cancers
Respiratory diseases
Diseases of urinary system
Not elevated
Elevated (not significantly)
Not elevated
VI-62
-------
Respiratory cancer was significantly higher in Johnstown, in both
males and females. In Utah it was higher only in males, and not signi-
ficantly higher there. This may be due to decreased exposure to BaP and
BSD in Utah, where the community levels are much lower. If this is true,
the increased rate in males might be due to occupational exposure to
higher levels. Another important factor in this instance is the compli-
cating influence of tobacco use. In the Utah communities studied, 60%
to 80% of the population are Mormon. Since most Mormons do not smoke,
the rates in the exposed and control populations could be greatly re-
duced due to the factor. This, then, could indicate that the combina-
tion of smoking and high BaP levels in Johnstown causes the excessive
deaths.
Cardiovascular disease, which was elevated in both men and women in
Johnstown was elevated in women only in Utah, and not significantly.
This might suggest a different etiology than previously presented, since
our initial hypothesis in Johnstown was based on cadmium, and the cad-
mium levels were only slightly elevated in the exposed areas. Possibly
the linkage of cardiovascular heart failure with respiratory dysfunction
(see Etiology Section in Chapter VII, Bituminous Coal Mining Case Study)
is the cause for elevation of the disease in Johnstown.
In Johnstown, deaths from endocrine, metabolic, and blood diseases
were significantly elevated in males only. This was again true in Utah
where females showed rates that were elevated, but not significantly so.
This may be occupational exposure, or it could reflect the difference
in reaction to the relevant pollutant between males and females.
Leukemia was significantly elevated in males in Johnstown and ele-
vated, but not significantly so, in Utah. The same was true for respi-
ratory diseases in females.
In general, the effects of the steel industry on Utah communities
would be expected to be less than in Johnstown due to the lower levels
of pollutants. Figure IV-8 shows the difference in BaP levels with re-
spect to distance from the plants in Johnstown and in Utah. As can be
seen from the figure, concentrations at the same distance from the plant
VI-63
-------
are much higher in Johnstown, due to both the amount of plant emissions
tf
and the geography. The difference in pollutant levels should have a
profound effect on community health and is the probable reason why we
found fewer causes of death elevated in Utah than in Johnstown.
100
o
•H
4J
nj
-u
0)
o
0
O
o
•H
VH
a)
ft
-------
3. Methodological Findings
One of the primary methodologies developed in this study was the use
of computerized mortality and cancer registry data. Given the scope of
the project and the time limitations, use of hard copy data was not
possible. The data available on tape had their own limitations—lack
of occupational, length of residence, and religious (important in Utah)
information. Another factor was residency reporting. Giving residency
by town rather than by census tract or enumeration district was fine in
some cases; however, in Johnstown we would have liked to analyze the
areas in the City of Johnstown closer to the Franklin Works separately
from those areas to the south. In most cases, more detailed residency
reporting would have made dose-response analyses possible.
The residency information on the Utah Cancer Registry tape was par-
ticularly disappointing. Only two years of data with detailed residency
information were available, and this was by zip code which is not easily
converted to populations. We were forced to analyze the bulk of the
registry data by county, which is quite limiting for an analysis of this
sort.
After deciding how to prepare the basic data, i.e., how to group
ages, how to age-adjust and for which ages, which demographics to exam-
ine, etc., we had to determine which statistical methods were valid as
applied to our data. Several methods were considered, including the
Cochran test, Chiang's method, and tests using the Poisson distribution.
The method finally decided upon was the Mantel-Haenszel x2 test.
Another methodology developed in this study involved the use of
interpretation of aerial photography. Barry Evans of EPIC assisted us
in this .by tracing the presence, growth, and change of polluting sources
in Johnstown from 1939 to the present (see Appendix I). Besides the
data presented in the appendix, topographic maps were prepared with
overlays displaying polluting sources and residential areas for each of
the years examined. This graphic representation of the historical resi-
dential and industrial patterns was very useful to us in analyzing the
data from this site.
VI-65
-------
Any analysis of the kind we performed depends on the availability of
air quality data for the pollutants of interest. We found that these
data varied greatly,not only from state to state but also from test site
to control site. Pennsylvania had data for criteria pollutants plus
heavy metals, ions, BaP, BSO, etc., for several years in the Johnstown
Air Basin. For the rest of Cambria County, however, there were no data.
Utah data, except for a special study done by EPA, were strictly for
criteria pollutants. Here again, data for the control site were scarce.
One of the points to be developed in this analysis was how our data,
analyzed on a town level, compared with the previously done regression
data which had been analyzed on a county level. Table VI-34 shows how
the basic data from each method compares in Johnstown. An important
factor in any analysis of this sort is sufficient population at risk.
This criterion is more easily met using county .data, although the steel
industry investigation had sufficient populations on a town level also.
If the population immediately surrounding the source is small, the
county unit may not show increased mortality. Conversely, factors
other than the source of pollution studied in the county might increase
countywide mortality rates. Another important consideration is a pos-
sible development of a dose-response trend as exposure levels decrease
with distance from the source.
VI-66
-------
Table VI-34.
AGE-ADJUSTED MORTALITY RATES, WHITE FEMALES
(ages 35-74, per 100,000, annualized)
CAUSE OP DEATH
SSI I
1
4
5
6
7
8-9
10
11
12
16,17.19
16
15,20,23
21,22
24-26
27
28
•29
31-37
34
38
40
46
50
ECI «
l-34a
2-13
2*
3-8*
3
4-5
7
6,8*
9
10
11
12
13
14
16-18*
16
17
18
19
20-22
21
23-27
28-29
1965-1975
Cambria County
County
1,102
274
2
69
10
33
8
13
20
51
4
102
27
66
589
34
22
S3 4
111
37
11
54
13
Exposed
1,273
278
6
63
11
43
9
20
26
44
3
111
25
61
708
36
27
645
127
52
12
60
19
Unexposad
988
231
0
60
7
34
9
10
12
49
2
85
23
76
528
27
19
482
105
32
9
47
9
1968-1972
SSI Data Canbria County
County
3,617
252
1
75
10
43
5
15
16
43
3
85
23
58
494
25
6
500
107
34
8
62
34
National
883
253
4
63
8
35
12
9
26
45
3
90
22
3,4
398
15
9
384
84
37
11
45
23
"Except 19.
•Some difference in 1CDA numbers, SSI vs. ECI.
VI-67
-------
4. Evaluation of Investigative Approach
One of the important evaluations which must be made from this study
concerns the value of the original regression approach for the steel
industry. The steel industry indicated by the regression was a very
broad SIC code containing many sources and types of pollutants. This
makes isolating the pertinent factor(s) with health effects difficult.
Another drawback, particularly in the steel industry, was the industrial
clutter found in counties with the steel industry. Our findings in Utah
seem to indicate that there is a minimum pollutant level, below which
few health effects are seen. In steel, it is difficult to find counties
with high pollutant levels and little industrial clutter. Where other
polluting sources are present, the matter of implicating steel industry
emissions is much more complicated.
The approach we selected for studying industrial activity and caus-
ally related community disease was a two-phase effort, first hypothesis
development and subsequent hypothesis testing. Its implementation in
Johnstown and Provo brought to light several possible limitations of its
use. Site selection is a problem. If one selects a test site to avoid
other sources of pollution, as we selected Provo, it is possible that
the industrial activity and the resultant pollution levels are too low
to produce a measureable effect. Thus, care must be taken to select
sites with sufficient industrial activity and emissions. This may force
the investigation to the larger metropolitan areas where general urban
pollution becomes a problem.
There are, however, several strong points in the approach we se-
lected. First, it was an objective approach, at least within the limits
of data availability. It was also industry-specific, that is, defining
exactly what processes and, therefore, what pollutants were present at
the sites studied. We studied two fully integrated steel mills and
characterized emissions by processes. Finally, the use of data more
specific than county data allowed us to see mortality patterns with re-
spect to actual pollution levels at the sites.
VI-68
-------
E. CONCLUSIONS AND RECOMMENDATIONS
Some of the hypotheses that were developed in Johnstown were sup-
ported by the Utah data. Notable among these, especially since it
applied to both males and females, was malignant neoplasms caused by
increased exposure to organic emissions, especially BaP. Even more
convincing, perhaps, is the fact of this finding even though exposure
levels in Utah were low and, perhaps due to religious practices, cancer
mortalities are low.
Of course, some hypotheses developed in Johnstown did not seem to
hold up well in Utah. These could also be explained by religious and
exposure differences, by climatological and geographic differences, or
by ethnic and demographic differences. It seems likely that there sim-
ply was not a long or high enough exposure in Utah to cause noticeable
health effects.
Several aspects of this study could be expanded to give a greater
insight into exactly what emissions and at what levels are health
hazards. Determining occupation, religion, and length of residence of
the decedents, while requiring a time-consuming study of hard copy data,
would do much to characterize mortality patterns.
Obtaining more complete air quality data is also important. Since
the parameters we are interested in are often not measured, this might
involve actual sampling and assay, especially in designated control
areas where the present monitoring networks have no samplers.
Finally, a detailed, in-depth study of a large industrial city with
high industrial activity with respect to steel, such as Pittsburgh or
Detroit, would give us a large enough population at risk to make a
detailed analysis of mortality patterns, especially with regard to a
dose-response relationship for specific pollutants.
VI-69
-------
Chapter VII
BITUMINOUS COAL MINING CASE STUDY
A. INTRODUCTION
This chapter documents the field investigation of the association
between bituminous coal mining and respiratory/cardiovascular diseases.
This association was clearly the strongest correlation found by the
multiple regression analysis of county industry and mortality data [1].
It occurred in both males and females and in both county samples, where
each sample contained roughly half of the 3,000 U.S. counties. The cor-
relation was convincingly validated by the top 10 county matched-control
test. Most of the component diseases .that comprise respiratory and
cardiovascular diseases were found to be correlated with the industry.
Because of the unusual strength of the correlation, particularly in fe-
males, we felt that it was the most likely of all correlations to repre-
sent a true cause-and-effect association between an industry and com-
munity disease.
The listing below, extracted from the Standard Industrial Classifi-
cation (SIC) Manual [2] describes the several components of the three-
•.
digit SIC code (SIC 121) correlated with respiratory and cardiovascular
diseases. The various diseases comprising respiratory/cardiovascular
diseases are listed in Table VII-1.
Group Industry
No. No.
121 BITUMINOUS COAL AND LIGNITE MINING
1211 Bituminous Coal and Lignite
Establishments primarily engaged in producing bituminous coal or lignite or In
developing bituminous coal or lignite mines. This Industry includes underground
mining, auger mining, strip mining, culm bank mining, and coal cleaning, crushing,
screening, and sizing plants, whether or not operated In conjunction with the minei
served.
[1] System Sciences, Inc., Investigation Jnto the Industrial Correlates
of Environment-Related Mortality, 1976.
[2] Office of Management and Budget, Standard Industrial Classification
Manual, 1972.
VII-1
-------
121S Bituminous Coal and Lignite Mining Services
Establishments primarily engaged In performing bituminous coal and lignite miulni
services for others on a contract, fee, or similar basis. Establishments which have
complete responsibility for operating mines for others on a contract, fee, or similar
basis ere classified according to the product mined rather than as bituminous coal and
' lignite mining services.
Table VII-1
CORRELATED CAUSES OF DEATH—RESPIRATORY AND CARDIOVASCULAR DISEASES
Respiratory/Cardiovascular Diseases Components
ICDA Number
(8th Rev.)
Respiratory. Diseases
Bronchopneumonia
Acute interstitial pneumonia
Emphysema
Asthma
Acute nasopharyngitis
Acute sinusitis
Acute .laryngitis and tracheitis
Acute pharyngitis
Acute tonsillitis
Acute upper respiratory infection of multiple or unspecified sites
Acute bronchitis and bronchiolitis
Influenza> unqualified
Influenza with pneumonia
Influenza with other raspiratory manifestations
Influenza with digestive manifestations
Viral pneumonia
Pneumococcal pneumonia
Other bacterial pneumonia
Pneumonia, unspecified
Hypertrophy of tonsils and adenoids
Peritonsillar abscess
Chronic pharyngitis and nasopharyngitis
Chronic sinusitis
Chronic laryngitis
Other diseases of upper respiratory tract
Spontaneous pneumothorax
Pulmonary congestion and hypostasis
Other pneumoconioses and related diseases
Bronchiectasis
Other diseases of respiratory system
Cardiovascular Diseases
Acute myocardial infarction
Other acute and subacute forms of ischemia heart disease
Chronic ischemic heart disease
Angina pectoris
Asymptomatic ischemic heart disease
Pulmonary embolism and infarction
Phlebitis and thrombophlebitis
Portal vein thrombosis
Other venous embolism and thrombosis
Varicose veins of lower extremities
Varicose veins of other sites
Noninfective disease of lymphatic channels
Other diseases of circulatory system
485
484
492
493
460
461
464
462
463
465
466
470
471
472
473
480
481
482
486
500
501
502
503
506
508
512
514
516
518
519
410
411
412
413
414
450
451
452
453
454
456
457
458
VII-2
-------
To investigate the relationship between bituminous coal mining and
respiratory/cardiovascular diseases in the community, field studies were
conducted at two separate sites: Wyoming County, West Virginia (see
Figure VII-1) and the bituminous coal mining region in southern Illinois
consisting of a 5-county area (see Figure VII-2). Wyoming County was
studied first, for the purpose of developing a plausible causal hypoth-
esis. The possible sources of pollution in the county were investigated,
including their history, magnitude, distribution with respect to the
population, and possible toxic effects. Based on these data and inter-
county and intracounty mortality patterns, plausible hypotheses were
constructed.
Testing of the hypotheses was accomplished by two methods: by a
morbidity study in Wyoming County and by a mortality study at the
southern Illinois site. The latter site was selected principally be-
cause of industrial characteristics befitting the hypothesis being
tested and .the ready availability of mortality data.
B. HYPOTHESIS DEVELOPMENT
1. Approach
To attempt to construct a plausible causal hypothesis for the asso-
ciation between bituminous coal mining and respiratory/cardiovascular
disease, an investigation was carried out in Wyoming County, West Vir-
ginia, consisting of five separate but related research efforts. The
first two, which were conducted more or less in parallel, consisted of
(1) a thorough study of the coal mining industry in general, and in
Wyoming County in particular, to fully characterize the mining-related
sources of pollution in the community, and (2) etiology research of re-
spiratory and cardiovascular diseases in coal miners and coal mining
communities to identify medical evidence of connections between toxic
substances from mining operations and disease. At this point, several
candidate hypotheses were postulated involving sepecific sources of
toxic agents produced by the mining operations. These hypotheses were
examined in -light of the next two research efforts to determine which
of the candidates were plausible, giving consideration to the site-
VII-3
-------
M
M
I
WEST VIRGINIA
Figure VII-1
Wyoming County, West Virginia:
First Study Site for Bituminous
Coal Mining Case Study
-------
Figure VII-2
Southern Illinois Bituminous
Coal Mining Region: Second
Study Site for Bituminous
Coal Mining Case Study
VII-5
-------
generated evidence. These efforts consisted of: (3) a pathway analysis
to explore evidence of the suspect toxicants being transported to com-
munity populations, and (4) a mortality study to determine if the
disease-specific mortality patterns provided support for any of the hy-
potheses . At this juncture, questions still remained regarding the
possible role of certain factors (particularly socioeconomic factors)
in the industry-disease relationship. Hence, a fifth effort was under-
taken which consisted of a detailed examination of intercounty disease
patterns in Appalachia and elsewhere, using the SSI county mortality
data base. The sections immediately following describe the results of
these five investigations.
2. Characterization of the Bituminous Coal Mining Industry
The two principal operations in the coal mining industry are mining
and coal preparation. Wyoming County, which was the third highest coal-
producing county in the country during the 1950's and 1960's, has had as
many as 90 mines and 15 coal preparation plants (tipples) operating at
one time [3]. In general, the number of mines is decreasing and their
size is increasing; however, the county is still a highly ranked coal-
producing county [4].
The operations of mining and coal preparation are each described
below. Following this is a discussion of the environmental pollutants
produced by these mining operations. In addition to mining and coal
preparation, coal transportation is discussed as an operation which is
a source of community pollution.
a. Mining Operations
There are three general methods used to mine coal: surface, augur,
and underground. The technique used depends on the seam's accessibil-
ity, width, and depth. Strip mining is employed when the seam is near
[3) U.S. Bureau of Mines and West Virginia Geological and Economic Sur-
vey, The Mineral Industry of West Virginia,
[4] Mining Information Services, Keystone Coal Industry Manual series,
VII-6
-------
the surface. The auger technique is used on seams exposed on hillsides,
when the seam is narrow, or the overburden is unstable for underground
mining. Auger mining involves a large auger drill (16 to 84 inches in
diameter) that bores horizontally into the coal seam where exposed.
Underground mining is classified by the mine's access to the seam.
Drift mines enter the coal directly on exposed hillsides. A slope shaft
angles down from the mine entrance to the coal seam. When the access is
a vertical shaft to a seam far below the surface, it is called a shaft
mine.
In Wyoming County most mines (-90%) are the underground type. Sur-
face mining in the area (-7%) is the contour technique which follows a
coal seam outcropping around a mountainside. Because seams are usually
horizontal, the overall effect of a mining operation is a contour line
around a mountain. Auger mining (-3%) usually follows the contour strip
mining [4].
Coal is transported from the mines by train, truck, or conveyor to
preparation plants or directly to a consumer. When coal preparation
plants (tipples) are close at hand, conveyor systems carry coal directly
to the tipple raw-coal storage. Small operations usually truck coal
out, while large mines may load directly onto a train.
b. Coal Preparation Operations
Coal preparation plants, commonly called tipples, perform three
functions: the crushing of coal lumps to specific sizes, the separation
of impurities (dirt, rock, etc.), and the removal of excess moisture.
A series of crushers, grinders, and screens are used to reduce and
separate the coal to uniform sizes. The finer the coal is crushed, the
more impurities, which are physically trapped in the coal, can be re-
moved in the process. The crushing and screening processes are for the
most part enclosed, resulting in limited fugitive coal dust emissions.
Impurities are separated by two cleaning processes: washing with
water and specific-gravity separation. Large coal sizes are sprayed
with water to rinse off clinging dirt particles. The separation pro-
cesses for smaller sizes of coal exploit the differences in specific
VII-7
-------
gravity between coal and impurities (coal is lighter). Usually water
and sometimes air is used as the medium. Effluents and emissions from
these processes include dirt, rock, and coal particles.
Drying of the coal is done by natural draining, screening, centrifu-
gation, filtering, and thermal drying. Effluents from these processes
again contain coal, rock, and dirt particles. Emissions are generally
limited due to the wet nature of the material. However, in the case of
thermal drying, large quantities of coal dust are entrained by the hot
flue gases used for drying the coal. The majority of coal preparation
plants in the United States use thermal dryers. Table VII-2 lists the
types and numbers of thermal dryers in the United States.
;
Table VII-2
THERMAL DRYERS IN THE UNITED STATES IN 1973*
Type
Fluidized Bed
Multi-louver
Rotary
Screen
Suspension or Flask
Vertical Tray & Cascade
Number
66
16
36
12
31
1
* Source: EPA, CoaJt Preparation Environmental
Engineering Manual, Publ. No. EPA-6002-
76-138, May 1976.
Open coal storage piles are not actual processes of tipples but are
associated with them. Uncovered piles of coal are sources of fugitive
-.emissions and leachate from rainwater. Also, the handling of coal by
trucks or conveyors around tipples generates fugitive coal dust emis-
sions.
In Wyoming County there are five thermal dryers currently in opera-
tion; Figure VII-3 is a map of Wyoming County showing the location of
these tipples.
VII-8
-------
KOPPERSTON
•/••' Kopperston
Oceana
V (Hatcher) ©
Coal Preparation Plants
Figure VII-4. Map of Wyoming County: Location of Coal Preparation Plants
-------
c. Environmental Pollution from Mining Operations
Pollutants from the coal industry include mine seepage, tipple solid .
waste, tipple process water, coal combustion products, and coal particu-
late emissions. Active and abandoned mines often seep acid water into
surfa'be waters and aquifers. Tipples dump the waste material separated
from -coal into large piles around the plant. Water used in the separa-
tion processes is sometimes acidic and is usually released to surface
water . Because coal is burned to supply energy for tipple processes
and hot flue gas for thermal dryers, coal combustion products are emitted
to the atmosphere. Burning tipple waste piles also emit coal combustion
produces. Coal particles are entrained and emitted into the air by mine
ventilation systems, transport systems, tipple operations, and thermal
dryers.
Acid Mine Drainage — Due to the sulfur content of the coal in Wyoming
County, it is possible for mine drainage to be highly acidic. If the
drainage has a pathway to the potable water supply, the various minerals
that can be solubilized by the acid en route can constitute a signifi-
cant health problem. However, in Wyoming County there is a considerable
amount of limestone in the substrata which will neutralize the acid
drainage to some extent and reduce the health hazard. The chemistry of
acid formation and neutralization is postulated as follows.
Once pyritic material in coal and overburden is exposed to oxygen
(in" air) and water (in vapor or liquid phase), the sulfur will be oxi-
dized to sulfite which in turn is hydrated to form sulfuric acid [5] :
FeS2 + H2O + 7/2 O2 - >• FeSO^ + H2SOi,
Acid formation drops the pH to 2 to 4.5 which tends to solubilize min-
erals (iron, manganese, etc.) at a greater rate than normal.
Upon reaching the surface, several reactions occur which result in
'the formation of bright-orange ferric hydroxide precipitate:
[5] Grim, E. C., and R. D. Hill, Environmental Protection in Surface
Mining of Coal, EPA-670/2-74-093, EPA, Cincinnati, Ohio, October
1974.
VII-10
-------
2 FeSOi, + 1/2 02 + H2SOI+ - »- Fe2(SOi,)3 + H20
Fe2 (59^)3 + 6 H2O — r*- 2 Fe(OH)3 + 3 H2SOi<
• Neutralization of acidic mine drainage occurs as native limestone
(CaC03 ) reacts with sulfuric acid to form carbonic acid and bicarbonate:
CaC03 — »- C02 + H2O + Ca+
C02 + H20 - >- H2C03
CaCOs + 2 H2CC>3 - >- 2 HCOs + CO2 + Ca** + H2O
The end result of neutralized acid mine drainage is an excess amount
of minerals and trace elements and excess total dissolved solids, es-
pecially sulfates and bicarbonates. These latter compounds can exert
laxative effects on humans. Trace metals in sufficient quantity may
pose a number of public health threats. High quantities of divalent
ions (Ca+^Mg"1"1") cause elevation in water hardness. The public health
significance of water hardness is not clear, but researchers have at-
tempted to correlate many chronic diseases, particularly cardiovascular
diseases (heart disease, hypertension, and stroke) with low water hard-
ness.
Coal Dust Emissions — Coal dust emissions originate from mine venti-
lation exhaust, transport vehicles, preparation plant operations (fugi-
tive) , and thermal dryers [6] .
It is, conceivable that the coal dust carried out of underground
mines by ventilation systems could create environmental pollution. How-
ever, significant amounts of pollution have never been shown to arise
from ventilation exhausts. These systems do not actively entrain or
heat coal dust; they only move the dust out of the mines. Ventilation
shaft locations change over time because they follow the working face
[6] Seeley Mudd Series, Coal Preparation.
VII-11
-------
of the mine; consequently, pollution from shafts would generally be
short term. Also, the exhausts are usually located a considerable dis-
tance from the populated areas. Consequently, mine ventilation exhausts
are not considered a major air pollution source.
Uncovered transport vehicles (railroad cars and trucks) disperse
coal dust throughout their routes. Railroad tracks are historically
stable, while truck routes vary with mine locations. Accurately defined
truck routes for the past 40 years are not available. The magnitude and
characteristics of this source of coal dust are extremely difficult to
estimate.
Fugitive emissions from tipples occur from the handling, storage,
and operations of the tipple. The loading and unloading of ROM and
clean coal, respectively, create fugitive emissions. Uncovered con-
veyors are also sources of windblown dust. Within the plant, the fol-
lowing operations may create fugitive emissions:
• sizing
• prewetting
• dewatering
• mechanical drying
Emissions from thermal dryers typically have gas volumes from 85,000
to 425,000 cubic meters per hour, temperatures qf about 93°C, and dust
concentrations of 230 to 690 g/m3 [7]. Table VI1-3 gives the character-
istics of the five thermal dryers in Wyoming County and shows the total
particulate matter emitted from each as compiled by NEDS [8]. All emis-
sions are determined without collectors or controls. Table VII-4 dis-
plays the characteristics of particulate emissions, particle size dis-
tribution and emission factors; also emission estimates for a 500-ton/hr
unit a're shown for an 8-hr day. There are ^1012 respirable particles/m3
emitted from a fluidized-bed unit with a dust loading of 450 g/m3
[7] TRW Systems Group and Resources Research, Inc., Air Pollutant Emis-
sion Factors—Supplement, DHEW Publ. No. PB-206 923, August 1970.
[8] EPA, National Emissions Data System (NEDS), Point Source Listings,
EPA, Research Triangle Park, North Carolina.
VII-12
-------
Table VII-3. THERMAL DRYER (FLUID-BED TYPES) CHARACTERISTICS: WYOMING COUNTY, WEST VIRGINIA*
PLANT/LOCATION
Keystone t2/
Herndon
Itaann/
Itaann
Kopperston/
Kopperston
Haben/
Maben
National Pocahontas/
Hullensville
MAXIMUM RATE
(tons of coal
dried/hr)
300
500
410
150
360
CONTROLS
Type
of
Device
Centrifugal &
Wet Scrubber
Centrifugal 6
Met Scrubber
Centrifugal s
Wet Scrubber
Centrifugal S
Wet Scrubber
Gravity
Efficiency for
Particulates
(%)
99.3
99.5
98.0
99.1
85
STACK PARAMETERS
Height
(ft)
93
136
114
30
84
Diameter
(ft)
5.5
9.9
17
4
5.7.
Temperature
(°F)
130
110
130
145
150
Flow Rate
(ACFM x
103)
130
142.8
130
67
180
Calculated
" Particulate
Emissions
(ton/vr)
243
107
80
24
67
*Source: NEDS, reference (8].
-------
(specific gravity of coal ^1.4). Table VI1-5 shows the results of an
analysis of the particulate emissions from a thermal dryer [9].
•Table VII-4. THERMAL DRYER PARTICULATE EMISSIONS CHARACTERIZATION*
Thermal Dryer
Emission
Ranges
Typical Particulate Size Distribution
by % Weight Less Than Stated Sizes
2y
0.1 to 2.5
, 10M
0.5 to 10
20)1
5 to 42
40g
14 to 50
60 u
20 to 70
Uncontrolled
Emission Factors
(Ib participates/
ton coal dried)
16 to 25
Particulate
Emissions'1
(tons/day)
Total
32 to 50
Respirable
0.16 to 5
;Source: Reference |7).
For a 500-ton/hr unit and 1 shift (8-hr)/day.
Table'VII-5
ANALYSIS OF PARTICULATE EMISSIONS FROM A THERMAL DRYER*
ELEMENT
Be
Cd
As
V
Nn
Ni
Sb
Cr
2n
Cu
Pb
Se
B
F
Li
Ag
Sn
Fe
CONCENTRATION
(ppmw)<»
1
«50
<100
50
50 to 100
20 to 30
<50
30
<100
30
<.10
—
10
—
<10
<1
<50
5000
ELEMENT
Sr
Na
K
Ca
Si
Mg
Bi
Co
Ge
Mo
Ti
Te
Zr
Ba
Al
Cl~
so£
CONCENTRATION
(ppmw)a
100
300
1000 to 2000
3000
1.5%
1000
<10
<10
<30
<10
500
<100
10
200
1.0%
40 to 118
1040 to 3920
Source: Reference [9).
Parts per million by weight.
[9] EPA, Background Information for Standards of Performance: Coal
Preparation Plants, Vol. I, II, III, Publ. No. EPA 450/2-74-021.
VII-14
-------
The control of transport vehicle and fugitive tipple emissions is
accomplished by enclosing the source. Open storage piles are replaced
with silos, conveyors are covered, and crushing and screening units are
r ' I ; •
enclosed with treated vents. Vehicles cover their loads with tarpaulins.
Consequently, fugitive emissions are controlled by enclosing the coal
and restricting its exposure to wind.
Controls on thermal dryer emissions are primary dry centrifugal re-
covery and secondary wet collectors. Cyclones collect up to 95% by
weight of the entrained particulates [7,10] and return them to the prod-
uct coal. All secondary controls are of the wet venturi type [10].
High-efficiency, high-pressure drop scrubbers can achieve 99% efficien-
cies by weight. Relative efficiencies of cyclone and venturi collectors
by particle size are shown in Figure VII-4. Typical grain loadings from
various types of control equipment are as follows [7]:
Outlet Grain
Type of Control Loading, g/scf
Cyclones (product recovery) 6-9
Multiple Cyclones (product recovery)' ! 2.5 -3.5
Venturi Scrubber* (pressure drop 5 in. water) 0.1 -0.2
Venturi Scrubber* (pressure drop 20 in. water) 0.02-0.03
*Following product recovery systems on the thermal dryer
effluent.
The efficiency of venturi scrubbers is dependent on particle size
distribution, and a large pressure drop is required to remove the smal-
ler particles. Consequently, high power inputs are required for high-
effiency collection of respirable particles [11] . Venturi scrubbers
were not incorporated at Wyoming County dryers until 1969.
[10] Nunenkamp, D. C., Coal Preparation Environmental Engineering Manual,
EPA Publ. No. EPA-600/2-76-138, May 1976.
[11] Vandegrift, A. E., et al., Particulate Pollutant System Study, Vol.
III. Handbook of Emission Properties, Publication PB-203 522, May
1971.
VII-15
-------
rum scaueBC* I«-IKCM THBOAT. JO-INCH ««TE» GAUGE)
i SPRAY To«ea ni'foor
(oarIL:CI
CTROSTATIC PRECIPITATOI) IMBCONO CONTACT TIJȣI
MULTIPLE CYCLONES IU-IHCM DIAUETE* 'USES!
» / SIUPLE CTCLOHC (4->ooT OI
NtSTIAL COLLECTCS
JO 19 30
PARTICLE size. «•>«>•«•
Figure VTI-4. Composite Grade (Fractional) Efficiency
Curves Based on Test Silica Dust
Coal Combustion Products—The principal sources of coal combustion
products are the burning waste (gob) piles and the burning of coal in
the thermal dryers to create the flue gas for drying coal. The particu-
late matter produced by the coal burning in the thermal dryer has been
included in the coal dust emission estimates discussed in the previous
section. These emissions consist principally of coal dust entrained in
the hot air but also include the particulates from the coal combustion
process. The gaseous combustion products emitted by the thermal dryer
consist principally of nitrogen and sulfur oxides. Because of the rela-
tively low rate of fuel burned (only 5 tons per hour in the largest
thermal dryer), the gaseous emissions are quite small and EPA has not
found it necessary to set standards for SOV and NOV emission controls.
A A
Further, NEDS, published by EPA, does not include point source estimates
of SOX and NOX emissions for thermal dryers (it actually publishes
values of zero). EPA has established that the gaseous emissions from
thermal dryers are much less than controlled emissions from small coal-
burning power plants [9].
VII-16
-------
Neither actual measurements nor estimates of emission rates were
found for burning gob piles. Due to the low coal composition of gob
•piles and the inaccessibility of oxygen in the piles, actual burning is
very slow and emission rates are very low. A cqunt of 15 burning gob
piles .in Wyoming County was made by the U.£>. Bureau of Mines [12].
3. Etiology
The part that coal dust plays in chronic respiratory disease in coal
miners is well established. It is conceivable, therefore, that coal
dust can be the etiologic agent for excessive respiratory disease in
coal mining communities. Furthermore, a reasonable case can be made for
an indirect connection between cardiovascular disease and coal dust,
where the respiratory dysfunction caused by chronic coal dust inhalation
accelerates or induces heart disease. The etiological basis for associ-
ating coat dust with respiratory and cardiovascular disease is illus-
trated in Figure VII-5. The etiology is similar to that for coal
workers' pneumoconiosis (CWP) in which coal dust is the toxic agent
entering the body by the respiratory route, causing chronic respiratory
disease which then leads to cardiovascular disease through increased
resistance of the pulmonary arterial bed and poor oxygen exchange. In
community disease, the exposure would be of long term and low level, in
contrast to the shorter term, high-level exposure causing CWP in miners.
The paragraphs below summarize first, the etiology of respiratory
disease and second, the etiology of cardiovascular disease associated
with coal mining, where the principal etiologic agent is considered to
be coal dust. Following this is a brief discussion of the etiology of
respiratory and cardiovascular diseases associated with other principal
sources of pollution from the coal industry, i.e., acid mine drainage
and coal combustion products. Appendix K presents a more thorough
treatment of the health effects of the possible etiologic agents asso-
ciated with mining operations.
[12] McNay, L. N., Coal Refuse Fires, An Environmental Hazard, U.S.
Bureau of Mines Information Circular No. 1C 8515, 1971.
VII-17
-------
EXPOSURE TO RESPIRABLE COAL DUST
• Coal particles penetrate
the alveoli
RESPIRATORY DISEASES
• Increased resistance of
pulmonary arterial bed
• Poor oxygen exchange
> CARDIOVASCULAR DISEASES
Bronchopneumonia; acute interstitial pneumonia;
emphysema; asthmai pneumoconiosis from silicates;
other pneumoconioses
Malignant neoplasms of: nose; nasal cavities;
glottis; true vocal cordi trachea; bronchus;
and pleura
JMyocardial infarction) ischemic heart disease;
\angina pec tor is
Figure VII-5. Etiology: Coal Dust Particles
a. Respiratory Disease Etiology
Coal miners themselves are predisposed to a variety of illnesses,
chiefly respiratory. There are four major occupational respiratory
diseases'of coal miners:
• Coal Workers' Pneumoconiosis (CWP) — this results from the
accumulation of inhaled coal dust in discrete macules in
the lung and can be associated with some changes in re-
spiratory function, rarely if ever disabling.
• Progressive Massive Fibrosis (PMF, or "complicated CWP) —
this is an advanced, uncommon form of CWP which results in
obstructive (emphysematous) and restrictive pulmonary dis-
ease and may be life threatening.
• Chronic Bronchitis—this probably represents chronic
bronchial irritation from dust, but has not been exten-
sively studied.
• Silicpsis—this occurs in miners who drill rock, sink
shaft, or drive the transport cars ("motormen") and can be
life threatening in severity.
There is a limited amount of work*that suggests increased respira-
tory disease in nonminer residents of coal mining communities, but this
is not consistent between communities.
*See Appendix K2 for references.
VII-18
-------
b. Cardiovascular Disease Etiology
Although there appears to be a moderate elevation in mortality in
coal miners from "heart disease" in general and "myocardial infarction"
in particular, these data must be interpreted very cautiously. An ex-
cess mortality from "heart disease" is a predictable result of a high.
prevalence of lung disease, because the pulmonary and cardiovascular
systems are intimately related. Mechanisms for this interdependence
are primarily the following:
(1) Progressive destruction of lung tissue also reduces the
cross-sectional area of the pulmonary arterial bed. Re-
duced oxygen tensions also induce vascular constriction in
the pulmonary artery. These phenomena result in reduced
compliance, therefore increased resistance of the pulmonary
arterial bed, which is normally a very low resistance hydro-
dynamic system. This increased resistance puts a markedly
increased work load on the right heart; over time, the right
ventricle hypertrophies (increases its muscle mass) to ac-
commodate the need to generate excessive pressure. If the
right ventricle cannot sustain the additional work, it will
fail, a condition known as "right heart failure" which has
several distinct features from the more common left heart
failure. In addition, the hypertrophic right ventricle may
demand more oxygen than its overextended coronary arteries
are capable of supplying, producing angina pectoris. The
phenomenon of right ventricular hypertrophy in association
with lung disease is known as "cor pulmonale" and has been
well documented to occur in coal miners with complicated
coal workers' pneumoconiosis.
(2) Distension of the right atrium by the same mechanism de-
scribed above and hypoxemia due to poor oxygen exchange by
severe lung disease sensitize the heart to arrhythmias. In
severe acid/base disturbances, which are not uncommon in
chronic lung disease, the heart is also made more irritable.
Foci in the cardiac conduction system and myocardium are
thus more susceptible to spontaneous depolarization and may
escape from electrical domination by the sinus node. Thus,
irregularities in rhythm are among the most common conse-
quences of moderate to severe lung disease. At the extreme,
these arrhythmias can be fatal.
VII-19
-------
(3) Severe lung disease is a chronic process of symptomatic im-
provement followed by exacerbation in a stepwise downhill
course. It is rarely a smooth history of decline, as car-
diac disease more commonly may be. An intercurrent illness,
such as influenza, may be all that is needed to acutely de-
compensate the patient and precipitate a cardiac problem.
Unless a physician is attuned to this in his record-keeping,
the interrelation may be lost in chart review.
(4) When a patient ceases to breathe as a result of lung dis-
ease and its attendant disturbance in gas exchange, the
next event to occur (after brain damage, which does not
declare itself immediately) is cardiac arrest. Undoubt-
edly, many deaths from lung disease are recorded as deaths
from cardiac arrest.
.The cardiovascular diseases which might be found in excess in coal
mining communities would consist primarily of ICDA numbers 410 (myocar-
dial infarction), 411 (acute/subacute ischemic heart disease), 412
(chronic ischemic heart disease), and 413 (angina pectoris). These are
different manifestations of the same basic disease process: athero-
sclerotic coronary artery disease. If coal mining community residents
are predisposed to accelerated atherosclerosis, these diseases should
be uniformly elevated in. risk.
c. Effects of Trace Metals
Certain minerals that can be solubilized by acid mine drainage
(such as trace metals) have been linked to cardiovascular diseases by
Schroeder [13]:
"Cadmium is a causal factor in hypertension. A replica
of the human disease was developed in rats and cured
by removing the causal factor by chelation. Zinc
deficiency could also play a part in hypertension."
Trace metals may be taken in, among other ways, through contaminated
drinking water. For example, it is known that soft water, which is more
corrosive than hard, can leach cadmium from water pipes and conduits.
[13] Schroeder, H. A., The role of trace elements in cardiovascular
disease, Med. Clin. North Am. 58(2):381-396, 1974.
VII-20
-------
d. Effects of Coal Combustion Products
The adverse health effects of sulfur dioxide are well known and have
been discussed in the copper smelting case study.
Nitrogen dioxide also exerts deleterious effects on the respiratory
system. Animal experiments have shown that it provokes obstruction of
the bronchi and rupture of the alveoli. For example, rats exposed
intermittently for a lifetime (about 2.5 years) to 15 ppm of N02 de-
veloped emphysema, a chronic obstructive lung disease, and exhibited an
attendant loss of oxygen-absorbing surface area in the lung. This re-
sulted from destruction of tissue, with a loss of about two-thirds of
the air sacs (alveoli). Stagnant air was retained in the enlarged, ob-
structed lungs following their removal from the chest. The main impact
of injury initially was in the smaller airways that conduct gases im-
mediately to and from the alveoli. These features constitute a reason-
able facsimile of human emphysema [14].
Nitrogen dioxide likewise reduces resistance of the respiratory
tract to infection. Prolonged exposure to considerable amounts of
nitrogen oxides contained in tobacco smoke leads to chronic bronchitis.
At the levels found in the atmosphere in certain cities, nitrogen di-
oxide is associated with an increase in acute respiratory problems among
people of all ages. This is especially true of influenza epidemics.
After two or three years of exposure to this risk, infections of the
lower respiratory tract, bronchitis, and repeated respiratory infections
become more frequent in children. The danger levels of nitrogen dioxide
reported in winter are 160 yg/m3 on the average with maximum levels ex-
ceeding 190 yg/m3, 40 days out of 100 [15].
[14] Workshop on "Possible Opportunity for Preventive and/or Therapeutic
Medical Payoff from Grant-Supported Studies on 03 and N(>2," spon-
sored by NCI, April 18, 1977.
[15] Gervois, M., and C. Dubois, Epidemiologic study of respiratory
manifestations related to nitrogen oxides in urban air pollution,
Rev. Epidemiol. Med. Soc. 22(8):710-713, 1974.
VII-21
-------
4. Pathway Analysis
The topography of Wyoming County is very distinctive. The entire
county is crisscrossed by mountain ridges and narrow valleys, such that
•there.is a total absence of open spaces or flatland. The ridges are
steep and the valleys are narrow, giving them a canyon-like appearance.
The sides of the "canyons" are roughly 1,800 feet high. Nearly all
the population lives along the valley floors (in the hollows) which are
•never very wide—many times only a few hundred feet. Rivers and streams
flow along most of the valley floors. The major coal preparation plants
'(thermal dryers) are located in the same valleys as the major population
centers of Mullens, Pineville, Oceana, Kopperston, and Corrinne (see
Figure VII-3).
The narrow valleys clearly define and channel the three main routes
by which mine pollutants can reach the community, i.e., via water, air,
and vehicles. Consequently, the three routes coincide, which makes it
difficult to determine which route is responsible for an effect. In the
case of the vehicular route, we clearly understand the pathway but there
is no way to estimate concentrations or exposures to the populations at
risk. For the water route, identifying the particular pathway by which
the water from the mines or tipples reaches the population is difficult.
However, we have been able to examine the quality of the water supplies
in Wyoming County for evidence of industry-related pollutants, e.g.,
minerals solubilized by acid mine drainage.
a. Air Quality Analysis
In the case of the air route, no air quality measurements were
available, so we cannot verify concentrations of either coal dust or
coal combustion products. Taking current measurements would not solve
the problem since the measurements would not represent conditions during
the desired exposure period, i.e., the years prior to 1970 when emission
controls were not in effect. In the absence of air quality data, some
thought was given to point source dispersion modeling to estimate down-
wind concentrations. However, due to the peculiarities of the terrain
and the lack of meteorological data for the region, it was felt that the
VII-22
-------
modeling would not be meaningful. Consequently, we were not able to
find or develop useful quantitative estimates of the amounts of toxi-
cants arriving by the air route.
• t
!•
b. Water Quality Analysis
During site visits to West Virginia, sanitarians at the state and
local levels provided water quality analyses of raw and treated munici-
pal water supplies in Wyoming County. Most public water supplies in
Wyoming County use well water. The averages of the annual analyses were
compared against the Public Health Service (PHS) standards and the new
EPA standards to determine if any violations of the standards existed
which might pose an obvious threat to public health. Compiled data are
presented in Table VII-6. Table VII-7 presents the standards against
which the data were compared. Finally, Table VII-8 summarizes how the
water supplies of Kopperston, Mullens, Oceana, Pineville, and Corrinne
adhered to the standards shown in Table VII-7.
Judging from the high iron and manganese content of the Wyoming
County municipal water supplies, it appears as if Wyoming County does
suffer from some acid mine drainage (AMD) which has a tendency to solu-
bilize iron and manganese, thus producing the high readings for those
elements. High sulfate and total dissolved solids readings, also re-
sults of AMD, are not seen in the county, thus indicating that the
problem is not as acute as in other coal mining regions.
VII-23
-------
Table VII-6. WATER' QUALITY 'ANALYSIS: WYOMING COUNTY, WEST'-VIRGINIA
l
NJ
PARAMETER
Alkalinity (phth)
Alkalinity (m.o. )
Hardness
Cadmium
Chromium
Fluoride
Lead
Silver
Chloride
Copper
Iron
Manganese
Nitrate (as NO3)
Sulfate
TDS
Zinc
Calcium
Magnesium
Sodium
pH (s.u.)
Turbidity (s.u.)
Potassium
Arsenic
Nitrate (as N)
Mercury
AVERAGE FINISHED & RAW WATER QUALITIES, rag/* (BY TOWNS)
Kopperston
Finished
0.0
4.0
22.0
0.0
0.0
0.12
0.0
0.0
7.0
0.006
0.09
0.007
4.0
6.0
43.5
0.05
2.23
1.98
0.98
5.8
-
0.96
-
-
-
Mullens
Finished .
0.0
127.5
131.5
0.0
0.0
0.33
0.0
0.0
19.25
0.025
0.52
0.02
2.3
71.25
380.6
0.20
23.52
12.62
41.63
7.45
2.24
4.23
0.005
0.045
0.00001
Raw
0.0
77.66
82.0
0.0
0.0
0.15
0.0
0.0
19.0
0.0
6.51
0.27
0.87
71.33
304.5
0.037
18.27
1.87
52.4
6.8
-
7.17
0.005
0.19
-
Oceana
Finished
1.5
67.926
104.5
0.0
0.0
0.56
0.0
0.0
34.0
0.0
0.39
0.05
1.3
-
304.5
0.08
21.50
4.61
8.5
7.9
3.5
1.58
-
-
-
Raw
0.0
26.66
48.0
0.0
0.0
0.12
0.0
0.0
9.0
0.10
3.84
0.123
2.5
66.0
130.5
0.05
10.44
9.60
31.57
6.9
-
2.00
-
-
-
Pineville
Finished
0.0
97.3
64.0
0.0
0.0
0.40
0.0
0.0
5.0
0.01
0.19
0.0026
1.6
39.5
261.0
0.12
11.0
4.6
39.1
7.2
2.8
-
0.005
-
-
Raw
0.0
90.0
57.3
0.0
0.0
0.15
0.0
0.0
18.6
O.O..
18.57
0.58
1.5
46.25
230.55
0.13
25.7
2.78
51.0
6.58
-
4.42
0.0055
0.34
-
Corrinne.
Finished
0.0
149.5
76.0
0.0
0.0
0.47
0.0
0.0
18.05
0.0
2.93
0.46
2.0
12.0
348.0
0.05
8.75
4.14
55.37
7.1
5.75
1.07
-
-
-
Raw
O.O
122.75
134.0
0.03
0.0
0.085
0.0
0.0
7.75
0.27
19.8
0.5
1.78
72.75
435.0
8.54
21.34
16.76
91.17
7.65
0.9
1.74
0.019
0.14
-
-------
Table VII-7. DRINKING WATER QUALITY STANDARDS
Substance
Mineral Salts
IDS
Hardness
Calcium
Magnesium
Sodium
Alkalinity (phenolphthalein)
Alkalinity (methyl orange)
Chlorides
Fluorides
Nitrates
Sulfates
Trace Metals
Arsenic
Cadmium
Chromium
Copper
Iron
Lead
Manganese
Mercury
Selenium
Silver
Zinc
USPHS (1962)
(mg/A)
500
*
-
-
-
-
-
250
0.9-1.7
45
250
0.01
0.01
0.05
1.0
0.3
0.05
0.05
-
0.01
0.05
5.0
USEPA (1977)
(mg/A)
-
*
.
-
-
-
-
250
1.4-2.4
10
250
0.05
0.01
0.05
1.0
0.3
0.05
0.05
0.002
0.01
0.05
5.0
There is no standard for hardness. The following Classifica-
tion of Water by Hardness Content Chart is provided by the EPA
in Quality Criteria for Water (EPA 440/19-76-023):
Concentration
mg/& CaC03
0-75
75-150
150-300
300+
Description
soft
moderately hard
hard
very hard
VII-25
-------
Table VII-8
RESULTS OF COMPARISON OF AVERAGE CHEMICAL HATER QUALITY OF
WYOMING COUNTY WATER SUPPLIES WITH PHS AND EPA STANDARDS3
PARAMETER
Cadmium
Chromium
Fluoride .
Lead
Silver
Chloride
Copper
Iron
Manganese
Nitrate
Sulfate
Total Dissolved
Solids
Mercury
Arsenic
Zinc
TOWN*'
C
C,M,P,O,K
C,M,P,0,K
C,M,P,0,K
C,M,P,0,K
C,M,P,0,K
C,M,P,0,K
P
C,M,0
M,P,
0,C
C,M,P,0,K
C,M,P,O,K
C,M,P,O,K
C,M,P,0,K
C
C
TYPE OF
WATER
Raw
Raw/Finished
Raw/Finished
Raw/Finished
Raw/Finished
Raw/Finished
Raw/Finished
Raw
Raw/Finished
Raw
Raw/Finished
Raw/Finished
Raw/Finished
Raw/Finished
Raw/Finished
Raw
Raw
COMMENT
Exceeds all standards
All in compliance
All in compliance
All in compliance
All in compliance
All in compliance
All in compliance
Exceeds all standards
Exceeds all standards
All in compliance
All in compliance
All in compliance
All in compliance
Exceeds PHS standards
Exceeds all standards
If supply is not mentioned, it is assumed to be in compliance;
no raw water was analyzed for Kopperston.
Town code: C = Corrinne; M = Mullens; P = Pineville; O = Oceana;
K = Kopperston.
VII-26
-------
5. Candidate Hypotheses
Based on the identified pollutants, the etiology of respiratory and
cardiovascular diseases, and the pathway analysis, candidate causal
hypotheses were constructed. These were then examined in light of the
evidence already developed to assess the plausibility of the hypotheses.
The candidates and the results of the preliminary evaluation are pre-
sented in Table VII-9.
The etiology connecting coal dust and particulate matter to respira-
tory disease is well established, and there appears to be a plausible
connection between cardiovascular disease and chronic respiratory dis-
ease. The population in Wyoming County is certainly exposed to coal
dust from the tipples and fugitive vehicular emissions; however, in the
case of' mine emissions, the population is generally remote to the mine
ventilation shafts. The principal difficulty with any of the coal dust
hypotheses, is the lack of reliable evidence of any kind on the magnitude
of exposure. Observations on site visits disclosed that the communities
are uncommonly dusty, even in this day when considerable care is taken
to control the dust and particulate emissions from thermal dryers and
vehicles. The narrow valleys help contain the dust to the populated
regions, and the wind and vehicular traffic constantly stirring up ac-
cumulations of the dust in the valleys adds to the ambient coal dust
concentrations. At this point, for purposes of hypothesis development,
we will assume that ambient dust concentrations in the communities have
been sufficiently high in the past to be instrumental in causing chronic
respiratory and heart diseases. Since the thermal dryers are a major
source of coal dust, it is reasonable to conjecture that populations
living close to thermal dryers will be exposed to higher dust concen-
trations than remote populations. Hence, we would expect to find more
prevalence of disease in the vicinity of thermal dryers than elsewhere.
Thus, to explore the plausibility of the coal dust hypothesis, we will
examine respiratory and cardiovascular disease patterns as described in
the next section.
VII-27
-------
Table VII-9. PRELIMINARY EVALUATION' OF CANDIDATE HYPOTHESES
I
K)
00
HYPOTHESIS
Etiologic
Agent
Coal Dust
&
Particulates
Coal
Combustion
Products
Acid Water
(Solubillzed
Minerals)
Source
Mines
(Ventilation Shafts)
Tipples (Principally
Thermal Dryers)
Vehicles
Gob Piles
Tipples
(Thermal Dryers)
Mines
(Drainage)
Tipples
(Coal Hashing)
Gob Piles
(Leachate)
EVALUATION CRITERIA
Plausible
Etiology
Yes
Yes
Yes
Yes
Yes
Partly
(Cardiovascular Dis.)
Partly
(Cardiovascular Dis.)
Partly
(Cardiovascular Dis.)
Suitable
Population
At Risk
No
Yes
Yes
Yes
Yes
Possibly
Possibly
Possibly
Sufficient
Concentration
Unknown
Unknown
Unknown
No
No
No
(Hater Quality Analysis)
No
(Hater Duality Analysis)
No
(Hater Quality Analysis)
PLAUSIBILITY
OF
HYPOTHESIS
Heak
Possible
Possible
Heak
Heak
Heak
Heak
Heak
-------
In the case of coal combustion products—in particular, the gaseous
products from.gob piles and thermal dryers—the emissions appear to be
too low to provide a reasonable possibility of harmful effects in the
community. • Hence, these hypotheses are not considered to be plausible.
Likewise, the hypotheses based on a water route are considered to
be too weak to be worthy of further consideration. The water quality
analysis of public water supplies disclosed no hazardous levels of toxic
agents. Although there is some evidence of acid mine drainage in the
water system (e.g., elevated levels of manganese and iron), it is clear
that this is not a major problem in the area due to low concentrations of
sulfates.
Two additional hypotheses have been posed which are not directly
related to the coal mining industry but which have an indirect connec-
tion. The first of these is the hypothesis that coal burning in the
home is the source of excessive respiratory and cardiovascular diseases
in the community. This hypothesis was not supported by the 1970 census
information which disclosed that better than 90% of the households in
Wyoming County use gas and oil for home heating. The second indirect
hypothesis is that coal miners and their families are heavy smokers.
This was verified by a local tobacco distributor and by hospital offi-
cials who felt that more than 90% of the miners smoked and most of their
, t
wives did likewise. In general, the miners chew tobacco when in the
mines, where smoking is not allowed, and smoke cigarettes when out of
the mine. Some further consideration is given to this hypothesis in the
remainder of the case study.
6. Mortality Analysis
On the basis of the above analysis, the most plausible hypothesis
for a causal relationship between industry and disease is the hypothesis
that coal dust emissions from the preparation plants (principally ther-
mal dryers) and other sources (coal transport vehicles and mine venti-
lation systems) are instrumental in causing respiratory/cardiovascular
disease. To examine this hypothesis in light of mortality patterns, two
VII-29
-------
separate investigations were made. In the first case, it was assumed
that the thermal dryer is the major source of the coal dust and that
communities 'close to the source would be exposed to a greater extent
than distant' communities. The major population centers in Wyoming
County, which were all demographically very similar, were divided into
two groups (those close to thermal dryers and those far from thermal
dryers) and the mortality rates were compared.
In the second mortality investigation, it was assumed that the coal
dust was an area problem associated with and accentuated by the particu-
lar mountainous terrain found in much of Appalachia. That is, it was
assumed that coal operations in the narrow valleys create coal dust that
is trapped in the valleys and pervades the entire population living in
the valleys. In this case, the mortality patterns of counties with and
without significant coal mining operations were comparecl for both moun-
tainous and flat regions.
The results of these two mortality investigations are discussed
below.
a. Mortality Study for Thermal Dryer Hypothesis
The major population centers in Wyoming County and the distance from
the nearest thermal dryer are listed below:
Distance from Nearest
Town Population Thermal Dryer (miles)
Kopperston 906 1
Mullens 2,985 3
Pineville 1,135 6
Oceana (Hatcher) 1>642 6
Corrinne 948 6
Because of their proximity to thermal dryers, the first two were con-
sidered to be the "exposed" population, and the last three were grouped
as the "control" population
Copies of death certificates for Wyoming County for the years 1968
through 1972 were obtained from the Director of Registration and Public
VII-30
-------
Area Town
Exposed •
Control •
' Mullens
^ Kopperston
Pineville
Oceana (Hatcher)
Corrinne
Health Statistics, Health Department, West Virginia, for respiratory and
cardiovascular diseases (see Table VII-1). The distribution of deaths
from these diseases between exposed and control towns is given below:
Number of Deaths
32
2
9
26
7
Because of the small number of deaths, we decided to treat the two
diseases together rather than look at them separately. It is to be
noted that we ordered death certificates for additional years and dis-
eases, however this order was never filled due, presumably, to work
overload of the Registrar and his staff.
To create the age-adjusted mortality rates, 1970 census population
figures were used. Wyoming County is divided into 29 enumeration dis-
tricts (EDs) of which the following districts match town boundaries:
Town ED
Kopperston 1
Mullens 18,19,20,26
Corrinne 17,25
Pineville 13
Oceana (Hatcher) 2
The remaining EDs are rural and their boundaries follow streams or moun-
tain ridges. As a result, small towns such as Herndon, Rock View, and
Baileyville can be divided by two or more EDs. Also, the rural EDs are
so large that they include both exposed and unexposed populations.
Using the 1970 census information, the demographic characteristics
of the exposed study population were compared to those of the control
population. Table VII-10 presents this comparison. It is evident that
most of the socioeconomic factors considered are quite similar for the
two groups. The significant differences occur in "% finished high
school" and "% using coal for heating." In both these cases, the values
for the control group are biased towards higher disease incidence,
thereby increasing the suitability of the population for control pur-
poses.
VII-31
-------
Table VII-10. SUMMARY OF DEMOGRAPHIC CHARACTERISTICS OF
ENUMERATION DISTRICTS IN WYOMING COUNTY, WEST VIRGINIA
CHARACTERISTICS
Median Male Age Group
Median Female Age Group
% Female
% Native of Native Parents
% White
% < 1.00 Persons per Room
% Electric Home Heat
% Coal Home Heat
% High School Grad
% At Same Residence as 1965
Median Income, $* 1,000
EXPOSED
25-34
25-34
52
97
95
95
1.8
1.5
52
66
8-9
CONTROL
25-34
25-34
53
98
96
87
3.7
10
40
53
7-8
To compare the respiratory/cardiovascular disease prevalence in the
exposed and control areas, age-adjusted mortality rates were computed
for white females (95% of the population is white). Since females do
not work in mines, any differences in mortality rates for females should
indicate a community health problem rather than an occupational problem.
The significance in differences between exposed and control areas was
tested using the Cochran test. The results are presented in Table
VII-11.
Table VII-11. MORTALITY STUDY RESULTS: WYOMING COUNTY*
Population
Mortalities
Age-Adjusted Mortality Rate/105
p (Cochran test)
EXPOSED
1,867
19
584
0
CONTROL
1,845
15
381
.04
*Based on mortalities in white females, ages 0-74,
during 1968-1972, from respiratory and cardiovascular
diseases.
VII-32
-------
Although the number of deaths is small, the difference between the
age-adjusted rates for the two areas is statistically significant. The
probability is less than 0.04 that the two values would occur in the
same population. This result provides some evidence in support of the
thermal dryer coal dust hypothesis.
It is to be noted that this result tends to work against the two
socioeconomic hypotheses/ in which smoking and home heating by coal are
assumed to be the cause of disease in coal mining communities. In the
case of smoking, there are no apparent reasons why the exposed area
should be any different from the control, since they are both located
in the center of coal mining operations and occupationally and socio-
economically they should be very similar. So a difference in mortality
rates cannot be readily blamed on tobacco usage.
In the case of use of coal for home heating, 10% of households in
the control population used coal as opposed to 1.5% of the exposed.
Hence, if home use of coal is the cause, then the control area should
exhibit higher mortality rates than the so-called exposed area, which
wasn't the case.
b. County. Mortality Pattern Analysis
To shed light on the factors that might be important in the associ-
ation between bituminous coal mining and respiratory/cardiovascular dis-
ease, we performed a regional analysis of the SSI county mortality data.
For the mountainous regions of West Virginia, Kentucky, and Tennessee,
we divided all the counties into three categories with respect to the
amount of coal mining in the county: high production, low production,
and no production. For each state, we computed the average county mor-
tality rates in each category for two disease groups: respiratory dis-
ease and cardiovascular disease. We then did the same thing for the
flat regions of Ohio, Illinois, and Kentucky. In both cases, we elimi-
nated all counties with cities and all counties with populations less
than 5,000, so that all counties used had populations in the range of
5,000 to 27,000 for the 35-74 age bracket. The results are presented
in Tables VII-12 and VII-13.
VII-33
-------
Table VI1-12
EFFECT OF COAL MINING ON COUNTY MORTALITY RATES IN
MOUNTAINOUS TERRAIN (WEST VIRGINIA, TENNESSEE, KENTUCKY)
State
West
Virginia
Tennessee
Kentucky
Total for
Three
States
Coal
Production
None
Low
High
None
Low
High
None
Low
High
None
Low
High
No. of
Counties
12
10
8
18
5
0
2
4
5
32
19
13
National
All. U.S. Counties
Average:
AGE-ADJUSTED MORTALITY RATES '
in White Females, ages 35-74
Respiratory
Disease
39
51
~ 51
48
46
—
33
38
40
44
47
47
37.5
Cardiovascular
Disease
451
485
519
403
414
—
440
438
515
423
456
517
397
For mountainous terrain, these data show a clear trend between the
amount of industrial activity and the average county mortality rates for
cardiovascular diseases. This trend is pronounced in each state and in
the aggregation of all states.
In flat terrain, we see a similar trend, but in this case it is much
less pronounced. For example, the relative risk between high-production
and no-production counties is 1.24 for mountainous terrain and only 1.07
for flat terrain. The difference between these relatives rates is quite
significant considering the number of counties involved and the regional
similarities between the "exposed" and control counties. The fact that
for counties with no coal production the aggregated mortality rates for
the two types of terrain are fairly close adds to the significance of
the difference in rate ratios between mountainous and flat terrains.
VII-34
-------
Table VII-13
EFFECT OF COAL MINING ON COUNTY MORTALITY RATES IN
FLAT TERRAIN (KENTUCKY, OHIO, ILLINOIS
' ' ' ( •
State
•
Kentucky
Ohio
Illinois .
Total for
Three
States
Coal
Production
None
Low
High
None
Low
High
None
Low
High
None
Low
High
No. of
Counties
7
0
2
7
2
1
5
4
1
19
6
4
National •
All U.S. Counties
Average:
ACE-ADJUSTED MORTALITY RATES
in White Females, ages 35-74
Respiratory
Disease
34
—
28
43
41
56
40
38
23
39
39
34
37.5
Cardiovascular
Disease
363
—
407
430
459
466
469
446
489
416
450
442
397
A possible inference from these data is that the type of terrain is
a factor in the association between industry and cardiovascular disease.
If we put this finding together with our coal dust hypothesis, a plaus-
ible explanation is that the valleys in mountainous terrain tend to con-
fine the coal dust to the communities; whereas in flat terrain, the coal
dust is more widely dispersed.
Looking at respiratory diseases, we see that the mortality rates
tend to be correlated with coal production in mountainous terrain, but
the trend is not as strong as in cardiovascular disease. Furthermore,
in flat terrain there is no clear association between respiratory dis-
ease and-coal; mining production. Although this can be interpreted as
refuting the coal dust hypothesis, it can be explained as being com-
VII-35
-------
patible with the hypothesis. If we assume that chronic respiratory
disease can lead to death from cardiovascular stress prior to death from
respiratory disease, then this can explain why the increase in cardio-
vascular disease is more pronounced. This is a reasonable explanation
which is hard to prove without examination of autopsy records to see if
there was a higher incidence of cardiovascular decedents with respira-
tory problems in coal mining regions.
The results of this regional county analysis tend to negate the
smoking hypothesis, i.e., the hypothesis that coal mining operations en-
courage heavy use of tobacco which extends to families and communities.
If smoking were the cause, there should be no difference in relative
rates between mountainous and flat terrains.
7. Discussion of Hypothesis Development Results
Although the evidence is not strong, it appears to point to the hy-
pothesis that coal dust from coal mining operations is the source of
chronic respiratory and.cardiovascular diseases in coal mining communi-
ties. A major source of the coal dust is the coal preparation plants,
particularly the large thermal dryers which are the focal points of coal
handling and transport operations and are located in the same valleys
as the major population centers. This hypothesis is supported by the
analysis of respiratory and cardiovascular mortalities in the county,
which shows that the mortality rates in females are higher in the vicin-
ity of thermal dryers than elsewhere. A contributing factor may be the
canyon-like valleys which confine the coal dust to populated areas. The
analysis of regional county mortality patterns shows that the relative
mortality rates between coal mining and non-coal mining counties are
higher for mountainous terrain than for flat terrain. One of the
strongest arguments in favor of the coal dust hypothesis is the etiology
of respiratory disease in coal miners, which establishes a definite
causal connection.between coal dust and respiratory disease.
The principal weakness in the coal dust hypothesis is the lack of
ambient air quality data to show that the communities are indeed exposed
VII-36
-------
to hazardous levels of respirable size coal dust. In present-day opera-
tions, the thermal dryers use highly efficient control devices which
collect as much as 99% of the coal dust and particulate emissions. No
measurements are available to show the ambient conditions prior to 1969
when the present-day control devices were installed.
Other possible hypotheses appear to lack one or more key elements.
The combustion products from thermal dryers and other sources are not
produced in sufficient quantities to be hazardous. Source strengths
are much less than found in metropolitan areas throughout the county.
Water quality is generally good in Wyoming County, and there are no
hazardous levels of harmful substances that are by-products of a severe
acid mine drainage problem. Furthermore, water pollution from mine
operations cannot be connected with respiratory diseases.
Home heating with coal does not appear to be the source of the ex-
cess diseases, because a very high percentage of the households use gas
or oil heating and the mortality analysis shows a lower mortality rate
in the area with highest usage of coal for home heating. However, we
may be seeing the latent health effects of the exposure that occurred
years ago when coal for home heating was more popular. So this hypoth-
esis cannot be completely disregarded. Neither can we convincingly dis-
regard the smoking hypothesis. Although the mortality patterns tend to
work against this hypothesis, there is insufficient data to make a
strong case either way.
There is always the possibility that migration can cause the mortal-
ity patterns, particularly at the county level. In coal mining areas in
general and in Wyoming County in particular, the populations have been
fairly stable. The general trend during the 1950's and 1960's has been
an emigration of young people, such that most coal mining areas show a
negative net migration during the period of most interest; in Wyoming
County, the net migration between 1960 and 1970 was -24.4%. Even if we
assume that enough young people move out such that a severe change in
age distribution occurs, the age-adjusted mortality rate would not
change. The type of migration that affects epidemiological analysis is
a selective migration of sick or healthy people. To effect an increase
VII-37
-------
in mortality with negative migration, the emigrants must be more
healthy than those remaining. There is no evidence that this has been
the case for coal mining regions.
C. HYPOTHESIS TEST
1. Objective
To test the coal dust hypothesis, two independent investigations
were carried out. The first was the collection and analysis of hospital
patient records in Wyoming County, West Virginia, to determine morbidity
patterns. The second was a field investigation (primarily a mortality
pattern study) at a new site. For this site, a five-county area in
southern Illinois was selected. Besides being a heavy coal producer
with a number of large thermal dryers, the area appeared to have good
population-at-risk characteristics. Furthermore, mortality data were
readily accessible, which had not been the case in Wyoming County. The
results of these two hypothesis test investigations are discussed below.
2. .Wyoming County Morbidity Study
a. Approach
.There are six hospitals in the vicinity of Wyoming County; however,
patient flow information supplied by the West Virginia State Department
of Health showed that three hospitals admit the large majority of Wyom-
ing County patients. The hospitals and the estimated percent of study
site total admissions served by the hospitals are listed below.
Percent of Total
. Hospital Town Study Site Patients
Stevens Clinic Welch 23
Appalachian Regional Beckley 20
Wyoming General Mullens 32
On the basis of this information, it was decided to analyze morbid-
ity records for Wyoming County patients from the three, principal hospi-
tals serving Wyoming County. Although all three hospitals initially
voiced support for our study and gave tentative approval to supply us
VII-38
-------
with patient data or records, only one (Wyoming General Hospital) could
obtain official approval and could provide the patient data on a time
scale and in a manner compatible with our study requirements. Although
this severely reduced the effectiveness of the morbidity analysis be-
cause of the unknown and indeterminable hospital-selection biases of the
1 : , ' " ; . •
patients, we decided to proceed with the Wyoming General Hospital data
since the hospital treats a significant percentage of the Wyoming Counizy
patients and, further, we felt we had ways to analyze the data that
would eliminate much of the selection bias.
Due to the expense of data collection and to the fact that we were
primarily interested in environmental as opposed to occupational dis-
eases, we collected hospital records for females only. We felt that
the disease stage might provide valuable information for assessing
toxic effects, hence staging was performed on 50% of the records col-
lected. .
;-"
The objective of the morbidity data analysis was to identify disease
residence^ patterns, e.g., to determine if disease incidence or severity
is a function of the distance of residence from a thermal dryer. To
accomplish this, the prevalence of the diseases of interest (respira-
tory and cardiovascular) was determined as a function of town of resi-
dence. However, to reduce the effect of selectivity bias, we used a
measure of prevalence consisting of the ratio of the disease-specific
cases to the total number of cases from all diseases less the disease-
specific cases. If, in a given community, the probability that a pa-
tient will select Wyoming General Hospital is the same for respiratory
and cardiovascular diseases as it is for other diseases, then the ratio
gives an unbiased measure of the relative prevalence between communities.
However, if there is appreciable disease selectivity and the patient
selects different hospitals for different diseases, then the results
will be biased and less valid. We do not know at this point the magni-
tude of disease selectivity; hence, the results must be interpreted in
the light of this unknown.
The data abstracting included histories and other lifestyle infor-
mation; however, too little of the data was available on the records to
permit a meaningful analysis.
VII-39
-------
b. Morbidity Data Collection
The morbidity data collection at Wyoming General Hospital in Mullens,
West Virginia, was carried out using seven local individuals who per-
formed abstracting work and clerical support such as pulling and re-
filing charts.
Medical records were abstracted for white females in the towns of
Corrinne, Kopperston, Mullens, Oceana (Hatcher), and Pineville. Records
date back to the late 1950's. All diseases (except respiratory and
cardiovascular) were abstracted on the short form. Respiratory and car-
diovascular diseases were abstracted on the long form. (See Appendix L
which presents the Hospital Record Abstract Forms.)
The entire hospital file was screened in order to identify the pa-
tients of interest. After 50% of the records had been abstracted, the
use of the long form was discontinued and the short form was used for
respiratory and cardiovascular patients. This was done in the interest
of time and expense, and because it was believed that a 50% sample using
the long form would be adequate to assess the usefulness of the staging
data.
The following conditions were not abstracted: routine surgeries
(such as tonsils and hemorrhoids), obstetrical cases, accidents, tooth
extractions,, and fractures. Although these conditions were not ab-
stracted, their numbers were recorded.
The procedure employed involved: (1) identifying the charts of
interest, (2) pulling and screening of charts to eliminate routine sur-
geries, obstetrical cases, etc., (3) abstracting, and (4) refiling the
charts.
In step 1, the clerk manually screened the chart number index cards
which were 'alphabetical by last name and included the address, race,
chart number, dates of admissions, in- or out-patient status, and active
or inactive status. The sex was determined by the first name, and the
chart was pulled when the sex was questionable which was rather frequent
in that area. The chart number index consisted of nine drawers of 3-by-
5 file cards; each drawer had three columns. The first column in all of
VII-40
-------
the drawers was screened first, then the second and third columns. This
"random" sampling method was used in case the study had to be stopped
short of completion due to time and/or funds limitations.
As the chart numbers were identified, they were listed on the coding
forms (see AppendixL). An "I" was added to the chart number for the
inactive patients (those not hospitalized in the last five years) be-
cause they were stored separately from the active files.
The abstractors located and pulled each identified chart, then
quickly scanned it. Charts with nondisease conditions (accidents, etc.)
were refiled immediately and a notation made beside the code number.
The remaining charts of interest were collected and abstracted.
Respiratory and cardiovascular diseases were abstracted using the
long form. For repeated diagnoses, staging forms were used only for the
first three admissions. All other diseases were abstracted using the
short forms (see Appendix L). When a file required a long form but not
all admissions were respiratory or cardiovascular, only the staging
forms were used. After abstracting the records were refiled.
An evaluation of the data collection was made on the third day. Ap-
proximately one full day of abstracting had been completed by two ab-
stractors. More than 50% of all cases were out-patients, and their
files were .not coded with an HICDA number, were incomplete and unsum-
marized; as a result, these were omitted from the study. About 50% of
the remaining cases had minor surgery, fractures, accidents, etc., and
created record-handling problems greater than their value; consequently,
.their numbers were recorded (during the record scans) but they were not
abstracted. In the first full day of abstracting, about 40 records were
processed—one-third of which required the long form.
The short forms required about two to three minutes, and the long
forms about 5 to 20 minutes.
VI1-41
-------
The following listing summarizes the data collected:
• Charts pulled 2,603
• Charts, of miscellaneous diseases 379
•••". Charts of respiratory diseases 97
• Charts of cardiovascular diseases 108
• •Total charts abstracted 584
c. Results of Morbidity Analysis
After the elimination of a small number of unusable records, a total
of 565 medical record histories remained which represented 565 separate
individuals. These had been coded into respiratory and cardiovascular
disease categories by the abstractors according to the category defin-
tions shown in Table VII-1. Table VII-14 gives the number of admissions
by major disease category and town.
Table VII-14
NUMBER OF HOSPITAL ADMISSIONS BY DISEASE CATEGORY AND TOWN
Town
kopperston
Corrinne
Mullens
Oceana
Pineville
Total by
Disease
Category
Respiratory
Diseases
2
4
54
15
20
95
Cardiovascular
Diseases
1
6
54
22
24
107
Other
Diseases
8
26
190
44
95
363
Total
by Town
11
36
298
81
139
565
To test the thermal dryer hypothesis, the admissions were grouped
in the hypothesized exposed and control areas where Kopperston and Mul-
lens are considered to be the exposed towns. To eliminate the hospital
selectivity bias, the "other diseases" were taken as control diseases
which would exhibit the same selectivity as the respiratory and cardio-
vascular diseases. Therefore, the ratio of respiratory or cardiovascu-
VII-42
-------
lar diseases to other diseases would represent the relative disease
prevalence which can be compared between exposed and control areas.
Table VII-15 presents these results.
Table VII-15
COMPARISON OF RELATIVE DISEASE PREVALENCE*
BETWEEN EXPOSED AND CONTROL AREAS
Town
Exposed3
Control2'
DISEASES
RESPIRATORY
Number
of
Admis .
56
39
Relative
Disease
Prevalence
0.28
0.24
CARDIOVASCULAR
' Number
of
Admis.
55
52
Relative
Disease
Prevalence
0.28
0.32
OTHER
Number
of
Admi s .
198
165
Relative
Disease
Prevalence
1
1
Relative Disease Prevalence =
aKopperston, Mullens
Number of Admissions
Number of "Other Diseases" Admissions
^Corrinne, Oceana (Hatcher), Pineville
It is immediately apparent that there is no appreciable difference
in relative disease prevalence between the postulated exposed and con-
trol regions, for either respiratory or cardiovascular diseases. Hence,
these results are negative with regard to the thermal dryer hypothesis.
As a check of this result, this comparison was made by an indepen-
dent method. In this case, the results of a patient flow study of
Wyoming General Hospital admissions were used to estimate the expected
distribution of diseases between exposed and control regions. On the
basis of the results of two surveys of admission patterns, it was de-
termined that out of 64 total admissions during the survey period, 60%
were persons with residences in the exposed area. Thus, in order for
the exposed area to indicate a higher prevalence of disease, the ratio
of the number of admissions for the exposed to that for the control
must be greater than 1.5 (60%/40%). Clearly this is not the case in
Table VllrlS; for respiratory diseases the ratio is 1.43 (56/39) and
for cardiovascular diseases it is 1.05 (55/52).
VII-43
-------
Data from the National Health Survey [16] were used to estimate the
fraction of hospital admissions with respiratory or cardiovascular dis-
ease. This was then compared with similar fractions for Wyoming County.
Table VII-16 gives the results. These data show that the fraction of
admissions is higher for Wyoming County. This could indicate that the
prevalence of both respiratory and cardiovascular diseases is higher
than normal.
Table VII-16
COMPARISON OF WYOMING COUNTY HOSPITAL ADMISSION DATA
TO THAT OF SOUTHERN UNITED STATES*
Number of Admissions
Fraction of All Diseases:
Wyoming County
Fraction of All Diseases:
Southern United States3
Respiratory
Diseases
95
0.17
0.12
Cardiovascular
Diseases
107
0.19
0.11
*Source: Reference [16] .
aBased on a total of 7,295,000 hospital admissions in 1972.
To complete the exploration of the Wyoming County admission data,
a detailed examination of the data was performed by disease stage, by
specific ICDA number, and by age. The attempt was to ascertain if there
were any.patterns in the data that would be illuminating. This detailed
examination did not reveal anything new and showed the general lack of
significant differences between exposed and control areas for specific
diseases, stages, and ages.
116] In-Patient Utilization of Short-Stay Hospitals by Diagnosis -
United States - 1972, U.S. Department of Health, Education and
Welfare, November 1975.
VII-44
-------
3. Southern Illinois Field Study
a. General Site Characteristics
,' The bituminous coal mining district of southern Illinois (Randolph,
Perry, Franklin, Williamson, and Saline Counties) was selected for a
test site. Reasons for the selection were the presence of extensive
coal mining (both surface strip and underground), associated tipples,
and the potential exposure of large populations (e.g., ^-13,000 in
Harrisburg, Illinois, vs. ^3,000 in Mullens, West Virginia). Also,
mortality data by town of residence, cause of death, age, race, sex,
and year of death were readily available from the Illinois Department
of Public Health, State Center for Health Statistics.
The terrain consists of gently rolling hills, much less distinctive
and confining than Wyoming County. Figure VII-6 is a map of the area
showing the major towns and coal-mining operations.
There are 10 major tipples in the study area (see Table VII-17).
Nine have operating histories dating back to the 1940's under various
company names as the plants were bought and sold. Thermal dryers are
operated by five of the tipples. These five are located on the map
shown in Figure VII-6. Only three of the dryers are sufficiently close
to towns to constitute a potential problem to the towns. Table VII-18
lists these three tipples and associated populations at risk (towns) and
distance between the tipples and towns.
VII-45
-------
I
£.
en
0*^*-t*\~
risburg*
!
ier( I
Saline Bounty^
Preparation Plants with Thermal Dryers
Towns/Cities
Figure VII-6. Map of Southern Illinois Study Area: Major Towns and Coal-Mining Operations
-------
Table VI1-17. PREPARATION PLANTS: SOUTHERN ILLINOIS*
Preparation Plant (Company)
Thermal Dryer
Closest Town
Captain (Southwestern Illinois)
Streamline (Southwestern Illinois)
Orient #4 (Freeman)
Orient #5 (Freeman
Central (Sahara)
Zeigler #4 (Bell & Zoller)
Delta (Ayshire)
Old Ben #21 (Old Ben)
Old Ben #26 (Old Ben)
Old Ben #27a (Old Ben)
Yes
No
No
No
Yes
Yes
Yes
Yes
No
Percy
Harrisburg
Johnston City
Source: Reference [4] .
New plant.
Table VII-18
:. RELATIVE LOCATION OF TOWNS FROM THERMAL DRYERS: SOUTHERN ILLINOIS
Tipple (Company)
Direction of/Distance
to Town from Tipple
Min & Max, miles
Town/Population
Central
(Sahara)
Captain
(Southwestern Illinois)
Zeigler
(Bell & Zoller)
E/2.3 - 4.6
S/3.0 - 3.6
WNW/6.0 - 6.5
WSW/1.7 - 3.1
Harrisburg/13,000
Carrier Mills/2,000
Percy/1,000
Johnston City/4,000
VII-47
-------
b. Transport Analysis .-^-.;
No air quality information was available for the towns'.of interest
in the southern Illinois study site. Consequently, calculations were
made using a stack dispersion model to estimate a measure of coal dust
exposure to the populations in the vicinity of the thermal dryers. It
is felt that the results of this type of modeling have much greater
applicability to the type of terrain in Illinois than to that in
Wyoming County, West Virginia.
, Turner's calculation method described in Equation (4) of Appendix A
was used to estimate exposure of two southern Illinois communities to
emissions from thermal dryers at coal preparation plants. These com-
munities were Harrisburg, Illinois, 5 to 7 km due east of the Sahara
mine and preparation plant, and Johnston City, 5 to 7 km west-southwest
of the Zeigler mine and preparation plant. This report presents input
parameters and results for the model, methods of constructing isopleths,
applicability of results, and a short critique of the applicability of
the Turner model to these sites.
Input data for the Turner equation were obtained from the Illinois
EPA's Total Air System File (Springfield, Illinois) and from the Na-
tional Climatic Center's STAR data files (Asheville, North Carolina).
The point source data for the two sites and some meteorologic informa-
tion are summarized in Table VII-19. Plume rise was calculated accord-
ing to the method of Holland (see Appendix A). The STAR data used were
obtained from observing stations in Evansville, Indiana, 80 miles east
of the study sites. This STAR program was based on observations from
January I960' to December 1964, 24 observations per day. It was run for
six Pasquill stability classes—three unstable, two stable, and one
neutral. '
Estimated values for long-term concentration of thermal dryer emis-
sions (X) were calculated for hypothetical receptors at 0.25 to 10 km'
from each plant in 16 directions (points of the compass). Topographic
maps of the study sites were obtained (P.S. Geological Survey, Reston,
Virginia) and the calculated concentrations placed on a radial grid at
VII-46
-------
appropriate distances and directions. In constructing concentration
isopleths, high values to the northwest and southeast were smoothed out
since they represented characteristics of Evansville meteorology rather
than meteorology of the study areas. Although this determination alters
the shape of the isopleth, it should not alter the meaning of the calcu-
lated values since stability categories in Evansville are similar to
those in southern Illinois. Isopleths are drawn around the Sahara and
Zeigler point sources in Figures VII-7 and VII-8, respectively.
Table VII-19
INPUT DATA FOR ESTIMATED LONG-TERM CONCENTRATIONS OF THERMAL DRYER
EMISSIONS IN THE VICINITY OF HARRISBURG AND JOHNSTON CITY, ILLINOIS
Parameter (units)
Inside stack diameter (m)
Physical stack height (m)
Atmospheric pressure (millibars)
Ambient temperature (°K)
Stack gas temperature (°K)
Stack gas exit velocity (m sec~* )
Wind speed (m sec" )
Effective stack height (m)
Emission rate (g/sec)
Preparation Plant Data
Zeigler
4.57
27.4
1013
286.8
360.8
2.08
1.5
52.88
38.1
Sahara
1.5
21
1013
286.8
324.7
15.5
1.5
51.55
156.2
Uncontrolled.
Applicability of the Turner analysis must be addressed. The only
ambient air quality data for southern Illinois is from the towns of
Effingham, Carbondale, Mt. Vernon, Marion, and Metropolis. The average
of the annual geometric means for total suspended particulates (1976) in
towns for which statistically significant data were reported (Effingham,
Mt. Vernon, Marion) was 56 mg/m3 [17] . Air quality of this magnitude is
among the best in the state.
[17) Illinois Environmental Protection Agency, Division of Air Pollution
Control, 1976 Annual Air Quality Report, Springfield, Illinois,
1977.
VII-49
-------
EsOmated Annual Averages of Concentration of Thermal Drier Emissions
In (he Vicinity of Harrtsburg. ffinois Resulting from Operation of
Sahara Coal Preparation Plant ( jig/m3)
Figure VII-7. Concentration Isopleths^ Sahara Coal Preparation Plant and Harrisburg, Illinois
-------
M
M
I
Estimated Annual Averages of Concentration ol Thermal Drier Emissions
in the Vicinity of Johnston City, Illinois Resulting from Operation of .
Ziegler Coal Mine and Preparation Plant (Ug/m3)
-.v-TC^ / \ //
^ 4 \ /r
Figure VII-8. Concentration Isoplethsi Zeigler Coal Preparation Plant and Johnston City, Illinois
-------
The computed concentrations shown in Figure VII-7 and VII-8 are
yearly average values and do not represent concentrations to be expected
when the tipples are operating and the meteorological conditions favor
heavy town pollution. Further, the concentrations do not include fugi-
tive coal dust from other sources in and around the tipple or from reen-
trainment of accumulations of dust in the community. Based on these
considerations, we would expect the actual concentrations to be con-
siderably higher than the computed values; however, there is no way to
estimate how much higher they would be. As it is, the computed values
shown in the figures are quite low compared to concentrations found in
mines (tens of micrograms per cubic meter compared to hundreds of micro-
grams per cubic meter). Thus, the transport analysis has not provided
verification that the population at risk was exposed to harmful levels
of coal dust. At best we can say that the results are inconclusive and
that air quality measurements must be found if a more conclusive answer
is to be provided.
It is to be noted that the concentrations were computed for the case
of uncontrolled emissions, which was largely the situation prior to
1969. In recent years, EPA has adopted emission standards for tipples
and the mining companies have incorporated control devices which are
highly efficient. Thus, we would expect a less significant relationship
between tipples and community disease in today's environment.
c. Mortality Analysis
To further explore the validity of the thermal dryer hypothesis,
mortality patterns at the southern Illinois site were studied. The
towns in closest proximity to thermal dryers were considered to be the
populations at risk, and mortality rates were compared between these
towns and a set of control towns which were remote to coal preparation
operations. The study and control towns are as follows:
VII-52
-------
Study Towns: Percy, Johnston City, Carriers Mills, Harrisburg
Control Towns: Marion, Herrin
Both the study and control populations are located in the same geo-
graphical area and have similar socioeconomic characteristics. Table
VII-20 shows how closely the two areas are matched in race distribution
and in number of persons per room, a key measure of lifestyle.
Table VII-20
STUDY AND CONTROL SOCIOECONOMIC INDICATORS: SOUTHERN ILLINOIS
Indicator
Race
White
Negro
Other
Total
Persons per Room
1.00 or less
1.01 to 1.50
1.51 or more
Study Area
Number
15,913
495
34
16,443
140
14
65
%
97
3
0
100
64
6
30
Control Area
Number
20,993
321
33
21,347
208
13
104
%
98
2
0
100
64
4
32
The numbers of mortalities, by year of death (1974-1976), cause of
death, residence, age, race, and sex, were acquired from the Illinois
Department of Public Health, State Center for Health Statistics. Age-
adjusted mortality rates in females were computed for respiratory,
cardiovascular, and "all other diseases," and differences in rates were
tested for significance by the Cochran test.
The results, presented in Table VII-21, show that both respiratory
and cardiovascular diseases are elevated in the population at risk, but
only the cardiovascular difference is statistically significant. The
mortality rates for all other diseases are similar for the study and the
control populations. This provides evidence of disease specificity.
VII-53
-------
Table VI1-21
MORTALITY STUDY RESULTS: SOUTHERN ILLINOIS
Age-Adjusted
Mortality
Rate/105
White Females
Relative Risk
p (Cochran test
Study
Control
:)
Respiratory
Diseases
94
62
1.5
0.21
Cause of Death
Card iovascular
Diseases
479
359
1.3
0.01
All Other
Diseases
611
691
0.9
—
4. Discussion of Hypothesis Test Results
Of the two tests performed on the coal dust hypothesis, the result
of one was negative and the other, positive. The morbidity study in
Wyoming County provided the negative result, which was the lack of a
disease pattern associated with the thermal dryers. The mortality study
in southern Illinois provided the positive result, which was the eleva-
tion of respiratory and cardiovascular disease in the vicinity of ther-
mal dryers.
The morbidity result is not considered particularly significant
because of the hospital selectivity problem. Due to the limited sam-
pling of the hospitals in the area, we had no way of compensating for
dis"ease-specific differences in patient flow from the study and control
areas. Probably the only thing that can be concluded from the morbidity
analysis is that there is not a very large difference in respiratory and
cardiovascular disease prevalence between the exposed and control areas.
That is, if a relative risk does exist between the two areas, it is not
outstanding.
The southern Illinois investigation provided significant evidence
tha-t females living near thermal dryers have a higher risk of death from
cardiovascular disease than those living elsewhere. They also seem to
have a higher risk of respiratory disease; however, the evidence is not
VII-54
-------
quite as strong. The problem with this result is that we cannot produce
concrete evidence that connects the excess mortalities to the thermal
dryer. Air quality measurements are not available, and modeling of the
coal dust dispersion indicates that the coal dust concentrations in the
communities at risk might be too low to be harmful. Although we believe
that the socioeconomic characteristics of the control and "exposed"
areas are very similar, there is always the chance that the differences
in the mortalities are caused by a difference in lifestyle.
At this point it is fair to state that the results of the hypothesis
test effort are inconclusive. Examining all the evidence, pro and con,
that has been developed to date, the balance sheet looks like this with
regard to the coal dust hypothesis:
Pro
Counties with bituminous coal mining have higher mortality
rates in respiratory and cardiovascular diseases than do
counties with no coal mining (both the regression analysis
and the regional county pattern analysis give strong con-
firmation of this).
In Wyoming County and in southern Illinois, communities
close to thermal dryers have higher mortality rates in
respiratory and cardiovascular diseases than do distant
communities (these are statistically significant results
except for respiratory diseases in southern Illinois),
Coal dust is a known etiologic agent for respiratory
disease and its connection to cardiovascular disease is
medically plausible.
The relationship between county mortality rates and coal
mining activity appears to be stronger in mountainous
regions than in flat regions. (This provides indirect
support to the coal dust hypothesis; i.e., the explanation
is that the coal dust concentrations are enhanced by the
confinement of the valleys and lessened by the dispersal
over flat land.)
VII-55
-------
• The percentage of hospital patients with respiratory and
cardiovascular diseases in the Wyoming County study site
is higher than average for Southern United States.
• No other reasonable explanation for the correlation be-
tween coal mining and respiratory/cardiovascular disease
has been found which is supported by the evidence at hand.
Con
The exposure of coal mining communities to high levels of
respirable size coal dust cannot be verified. Measurements
are unavailable and dispersion model computations do not
confirm the existence of significantly high levels of dust
concentration in the population at risk.
Morbidity patterns in Wyoming County do not confirm that
disease 'prevalence is higher in the vicinity of thermal
dryers (this negative result is not particularly signifi-
cant due to insufficient data to derive conclusive re-
i
suits).
Duration of exposure (duration of residence) of decedents
in the exposed population was not confirmed.
Other socioeconomic or environmental causes of excessive
mortality cannot be ruled out, such as excessive use of
tobacco or home heating with coal. There is some evidence
against these causes, but it is not sufficiently strong to
preclude them.
Respiratory diseases are less strongly correlated with the
coal mining industry than cardiovascular diseases, for both
the intercounty and intracounty analyses. This tends to
work against the respiratory route as the primary source of
disease.
VII-56
-------
D. CONCLUSIONS AND RECOMMENDATIONS
There is little doubt that cardiovascular and respiratory diseases
in the community are somehow associated with the bituminous coal mining
industry. However,-the nature of that association is open to question
and the results of this case study are inconclusive. The most plausible
causal hypothesis is that respirable size coal dust is the etiologic
agent which arrives in the community from various sources including
emissions from coal preparation plants, coal transport operations, and
mine ventilation systems. There is evidence that links the elevation
in community disease with the proximity of the community to the thermal
dryer. However, because the stack emissions alone do not appear to
provide sufficient community pollution, we surmise that the source of
the harmful pollution is the coal dust emissions from all the coal mining
operations that center on the coal preparation plant including coal
transportation and handling operations.
We have found two other possible hypotheses that link industry pol-
lution directly with respiratory or cardiovascular diseases; however,
there is no significant evidence to support these hypotheses. We con-
clude that these hypotheses which involve the pollutants of coal com-
bustion products and acid mine drainage .may be contributors but are not
the principal cause.
In terms of study methodology, the major difficulty with the study
approach was the inability to verify population exposure, either in terms
of ambient pollutant concentrations or in terms of time of exposure
(i.e., duration of residence). Because of the many sources of the hy-
pothesized etiological agent and the difficulty of modeling pollutant
dispersions over complex terrains, the only solution for verification
of ambient concentrations is on-site measurements. In the case of
retrospective studies, such as ours, the present-day measurements would
have to be extrapolated to previous source conditions. Since essen-
tially no health records include duration of residence information on
them, to verify duration of exposure requires a tedious and time-
VII-57
-------
consuming search of state and local records. This was too expensive
an operation for the present methodological investigation; however, it
is necessary for establishing the existence of a causal relationship.
A major difficulty with the investigation was the small populations
found in coal mining communities. There were not enough deaths to per-
mit the useful examination of the effects of age, year of death, speci-
fic disease, etc. Aggregation of populations over time and disease
categories was necessary to be able to show statistical significance,
and these aggregations obscured many of the effects that could illumi-
nate the nature of the relationship.
With regard to further investigating the bituminous coal mining-
respiratory/cardiovascular disease relationship, we recommend:
(1) Mortality studies at several additional sites to obtain
a measure of consistency. These studies are fairly easy
to perform and, in the absence of larger populations
which can show the effect of significant factors, the
multiple-site approach is a feasible alternative.
(2) A concerted effort to verify coal dust exposure at prom-
ising sites for which there is a significant mortality
pattern. Extensive measurements and modeling are re-
quired to verify concentrations, and a thorough record
search of decedents or phone follow-up of relatives of
decedents is necessary to establish duration of exposure.
(3) A detailed study of the differences in the smoking, diet,
and heating habits between coal mining and non-coal mining
communities. This requires interviews and questionnaires.
With regard to general methodology for epidemiological investigation
of environmental relationships, the following recommendations are made:
(1) Attempts should be made to standardize death certificates
to include zip code, length of residence at present address,
and occupation. This information should be recorded on
tape to facilitate the use of the data in mortality studies.
VII-58
-------
(2) NEDS should be expanded to include the dates that pro-
cesses and controls came on-line. This is important
; , information for determining historical exposure. In
addition, it would be helpful if process descriptions
were provided in more detail.
(3) Centralized morbidity systems such as cancer registries
and PAS should include the zip code and place of resi-
dence on patient records. Without this information, the
records are not particularly useful in environmental
epidemiology. Also, the occupation and duration of resi-
dence of the patient would be a great help.
VII-59
-------
Chapter VIII
VISCOSE RAYON INDUSTRY CASE STUDY
A. INTRODUCTION
During the selection of industry-disease associations for study, it
was felt desirable to include one case study selected by a different
method from that initially prescribed for the investigation. One reason
was that the basic method, regression analysis by counties, might fail
entirely to select useful associations. For example, the "real" associ-
ations might all be obscured because a county is too large a unit for
study, or it might be that strong associations were, found only in areas
of intense industrialization and hence of multiple exposure, with poor
prospects of disentangling the effects of specific emissions from one
industry. Another reason for testing alternative methods of selection
was that EPA regarded methodological development as an objective of
equal stature to the search for possible specific causal relationships.
The regression analysis was a purely objective method of selection,
with no preconceived associations. Accordingly, for the pressent study
we sought an industry-disease combination where a specific association
might be expected. We also sought one that offered a simple exposure
to toxicants, free from confounding pollutants.
A br.ief. review of industries and contact with public health offi-
cials spon identified the viscose rayon industry as a promising subject
for study. It has an uncomplicated air pollution potential (carbon
disulfide and hydrogen sulfide only); about half of the plants are in
small towns or rural areas; and there is a lot of evidence about occu-
pational health hazards, including cardiovascular disease. Community
exposure was indicated by complaints about odor. Finally, a comparison
of counties with viscose rayon plants and control counties without,
showed elevated mortality from coronary heart disease in some of the
counties with viscose rayon industry. Accordingly, we had a causal
hypothesis, suitable for testing, that involved coronary heart disease
and other cardiovascular effects.
VIII-1
-------
Since the hypothesis was developed in this way, the viscose rayon
study is different from the other three in that there was no field sur-
vey for hypothesis development. Instead, we proceeded directly to test
the hypothesis at two sites.
B. DEVELOPMENT OF HYPOTHESIS
1. Selection of Viscose Rayon Industry
The first indicator that viscose rayon might be a suitable subject
for study was the community complaint of odor commonly associated with
the industry. It was noted that although hydrogen sulfide was undoubt-
edly the main cause of complaint, it was known to be accompanied by sub-
stantial amounts of carbon disulfide, possibly a greater health hazard
to the community; hydrogen sulfide has a much lower odor threshold.
On closer examination of the industry, it became apparent that it
offered excellent conditions for an uncluttered study of community
health in the vicinity of industrial activity. About half of the U.S.
viscose rayon plants are located outside of metropolitan areas, in small
towns (populations 10,000 to 20,000) or in rural areas.
Selection of the industry was finally determined by the existence
of considerable evidence about occupational disease and some indication
of an environmental hazard; see Sections B.3 and B.4 below.
2. Description of the Viscose Process
The manufacture of viscose rayon fiber begins with sheets of bleached
cellulose sulfate (from wood pulp or cotton wastes) delivered from a
pulping mill. The raw cellulose is placed in a steeping press and
soaked in 17% to 20% sodium hydroxide for one hour, after which excess
liquor is drained off, carrying with it impurities such as cellulose
degradation products.
The soft alkali cellulose remaining is shredded to flakes in a
shredder, and the resulting crumbs are aged for 24 to 48 hours at 75°F
in large steel vessels.
VIII-2
-------
The aged crumbs are placed in large churns to begin the xanthation
process in which the crumbs are churned with carbon disulfide under re-
duced pressure for two hours to produce small balls of cellulose xan-
thate:
Rcell ONa + CS2^t Rcell °CSSNa (Xanthation Reaction)
(where Rcell represents a cellulosic carbohydrate) .
The xanthate balls are then dissolved in dilute sodium hydroxide
solution to produce the viscose solution. This reaction takes two to
three hours. The viscose is ripened in a series of tanks under continu-
ous blending and deaeration during which viscose modifiers (amines and
ethylene oxide polymers) may be added. Ripening is performed to adjust
the concentration of sodium cellulose xanthate to proper specifications
for the spinning operation:
Rcell OCSSNa *-— - Rcell ONa + c$2* (Ripening Reaction)
The viscose solution is then injected through a spinneret die con-
taining 750 to 2,000 minute holes into an acid spin bath containing sul-
fur ic acid and zinc sulfate. The acid bath does the following:
• It coagulates the viscose, forming a skin around the fila-
ment which regenerates to restore cellulose. The overall
reaction is as follows:
Rcell OCSSNa + H2SOi» — »• Rceii OH + Na2SOlt + CS2f
• It yields intermediate formation of a zinc xanthate which
allows more time for stretching, aligning, and compacting
regenerated cellulose.
• It decomposes by-products of the xanthation process. Among
the decomposition reactions are the following which give
rise to significant quantities of hydrogen sulfide and
carbon disulfide:
Na2CS2 + H2SOit - >• Na2SOlf + H2S+ + CS2t
Na2S + H2S(\ - >• Na2SOJt
The fibers thus produced are stretched, cut, finished, wound, and
shipped .
A flow diagram of the entire viscose process is presented in Figure
VIII-1.
VIII-3
-------
Cellulose
sheet!
Sleeping;
pressing
Shredding
Aging
Regeneration
FIBER PRODUCTION
Stretching Cutting Finishing
Casting, washing, bleaching, etc
FILM PRODUCTION
Drying
Acid biith
Figure VIII~1. Flow Diagram of the Viscose Process
This process description indicates process areas from which the pol-
lutants of primary concern would emanate. Carbon disulfide is exhausted
from ripening tanks and the spinning room, and hydrogen sulfide is ex-
hausted primarily from the spinning room. Emission rates of these
materials are discussed later. Although most of the gases emanating
from the processes are exhausted to the atmosphere through a fume stack,
some fugitive emissions would also be anticipated.
Aqueous effluents from the process result from the disposal of
wastes from the recycled spin bath chemicals and from washing and other
final treatment of the fiber. Untreated effluent contains zinc sulfates,
sulfuric acid, sodium sulfate, carbon disulfide, hydrogen sulfide, and
traces of cellulosic material lost from the process stream. Concentra-
tions of these compounds in the effluent from the plant depend on the
efficiency of the spin bath recovery cycle and the amount of treatment
that wastewater receives before being discharged. Amounts of these
chemicals discharged per ton of fiber produced are not available at this
time; however, for every pound of regular rayon produced, 0.05 pounds of
zinc sulfate and 0.84 of pure sulfuric acid are used.
VIII-4
-------
The finishing of rayon fiber involves the use of large quantities
of water for washing, rinsing, and bleaching. The bleach process may
add low concentrations of sodium hypochlorite, hydrogen peroxide, or
chlorine dioxide to the wastewater stream.
3. Evidence of Occupational Disease
There is substantial evidence of increased incidence of various dis-
eases among workers in the viscose rayon industry. The occupational
subgroups most affected are workers in the spinning room and in the cut-
ting, washing, and drying areas, where exposures to carbon disulfide and
hydrogen sulfide are greatest. Few studies on rayon workers have made
an effort to distinguish between exposures to the two agents since the
agents occur together in job areas where concentrations are greatest.
A recent publication by the National Institute for Occupational Safety
and Health critically evaluates a number of retrospective, prospective,
and clinical epidemiologic studies done on viscose rayon workers [1].
These studies are summarized in Table M-l in Appendix M which is adapted
from that document, and reports the number of workers studied, their
mean age, concentrations of toxicant, mean duration of exposure, and
effects noted. Table M-2 outlines animal studies of the toxicity of
carbon disulfide.
The CS2-H2S mixture induces cardiovascular effects, reproductive
system effects, neurologic effects, effects on the eyes, and other ef-
fects. Development of atherosclerotic lesions at relatively early ages
has been reported among rayon workers [2-5]; this is thought to lead to
[1] National Institute for Occupational Safety and Health, Criteria for
a Recommended Standard: Occupational Exposure to Carbon Disulfide,
Washington, D.C., 1977.
[2] Tiller, J. R., R. S. F. Schilling, and J. N. Morris, Occupational
toxic factor in mortality from coronary heart disease, Br,. Med. J.
4:407-411, 1968.
[3] Attinger, E., Chronisch Schwefelkohlen stoffvergiftung unter dem
"scheinbar ungewbhnlichen" Bilde einer schweren Gefasskrankheit,
Schweiz. Med. Wochenschr. 75:667, 1948.
[4] Nunziante, C. A., J. Clin. Med. 54:731, 1953.
[5] Browning, E., Toxicity and Metabolism of Industrial Solvents,
Elsevier Publishing Company, Amsterdam, 1965.
VIII-5
-------
increased cardiovascular and coronary artery disease [6]. A recent re-
view article [7] addresses toxicologic and physiologic manifestations
of chronic and acute carbon disulfide exposures of animals and man.
Neurologic, psychiatric, and ocular manifestations are some of the more
prominent problems relating to CS2 intoxication which are addressed.
Mancuso reports on C£>2 as a cause of suicide in man [8] .
Two sets of studies reported in the literature in recent years lay
the foundation for the viscose rayon-cardiovascular disease association.
Tiller et al. and Schilling [2,6] have reported on a retrospective mor-
tality study in England and Wales. Hernberg and colleagues have pub-
lished a series of reports [9-14] on a retrospective study and an on-
going prospective study in Finland, begun in 1967 and continuing to the
present. Both sets of studies report significantly elevated mortality
due to coronary heart disease in the CS2/H2S-exposed viscose rayon
workers, compared with nonprocess workers or other sets of controls.
[6] Schilling, R. S. F., Coronary heart disease in viscose rayon
workers, Am. Heart J. 80:1-2, 1970.
[7] Davidson, M., and M. Feinleib, Carbon disulfide poisoning: A re-
view, Am. Heart J. 83:100, 1972.
[8] Mancuso, T. F., and B. Z. Locke, Carbon disulfide as a cause of
suicide—Epidemiological study of viscose rayon workers, J. Occup.
Med. 14:595-606, 1972.
[9] Hernberg, S., T. Partanen, C. H. Nordman, and P. Sumari, Coronary
heart disease among workers exposed to carbon disulphide, Br. J.
Ind. Med. 27:313-325, 1970.
[10] Hernberg, S., C. H. Nordman, T. Partanen, V. Christiansen, and
P. Virkola, Blood lipids, glucose tolerance and plasma creatinine
in workers exposed to carbon disulfide, (fork, Environ. Health 8:
11-16, 1971.
[11] Hernberg, S., M. Nurminen, and M. Tolonen, Excess mortality from
coronary heart disease in viscose rayon workers exposed to carbon
disulfide, Work, Environ. Health 10:93-99, 1973.
[12] Nurminen, M., Survival experience of a cohort of carbon disulphide
exposed workers from an eight-year prospective follow-up period,
Int. J. Epidemiol. 5:179-185, 1976.
[13] Tolonen, M., S. Hernberg, M. Nurminen, and K. Tiitola, A follow-up
study of coronary heart disease in viscose rayon workers exposed
to carbon disulphide, Br. J. Ind. Med. 32:1-10, 1975.
[14] Hernberg, S., M. Tolonen, and M. Nurminen, Eight-year follow-up of
viscose rayon workers exposed to carbon disulfide, Work, Environ.
Health 2:27-30, 1976.
VIII-6
-------
In the study by Tiller et al. [2], death records were combed for
factory workers who were grouped as former process workers or former
nonprocess workers. A set of controls consisting of deceased nonrayon
workers was selected from the same death registers. The death certifi-
cates were classified by cause, age at death, and occupation. Percent-
age of deaths from specific causes was the dependent variable. Results
showed that percentage of deaths due to coronary heart disease was sig-
nificantly higher (p < 0.001) in rayon process workers (42%) than in
nonprocess rayon workers (24%), other local men (17%), and British men
as a whole (14%). Certain defects in the study were noted, such as
vagueness of the term "process" worker, lack of data on extent and
length of exposure, unsuitability of control population, etc. Neverthe-
less, the study does show an association between rayon workers and coro-
nary heart disease.
Hernberg and co-workers did a retrospective mortality study and a
prospective morbidity and mortality study on a cohort of 343 Finnish
rayon workers. For the retrospective cohort, the researchers reported
that 52% of the deaths among rayon workers were attributable to coronary
heart disease, while the expected percentage based on national mortality
experience was 31.7%. This difference was significant at the p < 0.002
level for a two-tailed test [9]. Clinical parameters in the prospec-
tive study included ECG following exercise, chest radiography, blood
pressure, blood lipids, glucose tolerance, and plasma creatinine and
cholesterol. Of these, only blood pressure and creatinine concentration
were significantly different from matched controls [10,11]. Finally,
the follow-up of the cohorts of workers has led to Nurminen's life
table study [12] which, based on the 8-year cumulative incidence rates
of coronary heart disease for exposed and control groups, gave a rela-
tive risk of 2.2, significantly larger than 1.0.
These, two sets of studies are the main basis for looking at coronary
heart disease in residents near viscose rayon plants. As already noted,
the tables in Appendix M summarize the range of physiologic and patho-
logic conditions which have been reported as possibly being related to
exposure to CS2 or CS2/H2S mixtures.
VIII-7
-------
4. Evidence of Environmental Exposure
Three studies in the United States monitored hydrogen sulfide con-
centrations near viscose rayon plants and a Russian study monitored hy-
drogen sulfide and carbon disulfide. The results are summarized in
Table VIII-1. The Russian study is the only one we found that measured
carbon disulfide; unfortunately, sampling protocol, analytical methods,
mean concentrations, and statistical analysis were not reported. For
other studies, carbon disulfide concentrations were estimated on the
basis of an assumed ratio to hydrogen sulfide in the emissions, as de-
scribed later in the site studies (Section C).
5. Hypothesis
On the basis of the foregoing evidence, we decided to test the fol-
lowing hypothesis:
Emission of carbon disulfide and hydrogen sulfide (or one of them)
from viscose rayon plants causes increased incidence of coronary heart
disease and other cardiovascular disease in the exposed population.
(The site studies looked for associations also with other diseases,
for two reasons: in case all diseases, or an unrelated group, were
elevated by a nonspecific factor; and to look for increased incidence
in other diseases known to be associated with the specific toxicants.)
VIII-8
-------
Table VIII-1. AMBIENT CS2 AND H2S CONCENTRATIONS IN VICINITY OF VISCOSE RAYON PLANTS
LOCATION
ELIZABETHTON, Tennessee
Site 1 : 610 m M of plant
Site 2: 1070 m ENE of plant
Site 1
Site 2
LOWLAND, Tennessee
Site 1: 610 m SW of plant
Site 2: 1070 m HE of plant
Site 1
Site 2
OSSR: 910 m from factory
ENKA, North Carolina
1.6 fan M of plant
1.6 km ESE of plant
8 tan.H of plant
SUBSTANCE
H2S
H2S
CS2
CSj
H2S
H,S
CS2
CS2
H2S
CSj
MEAN
CONCENTRATION
)
14.8
1.2
-844
-68
8.6
15.3
-86
-153
:
RANGE
(wg/m3)
0.3-82.5
<1-2.1
~™
0.6-25.7
1.1-43.5
__
1.4-49.5
5.4-222
[H2S] > IS iig/n3 234 days/365 days
[H2S] > 15 (ia/m3 46 days/365 days
[H2S] > 15 ug/m3 151 days/365 days
METHOD
APHA bubbler method, 24-hr samples
Extrapolation based on
CS2:B2S ratio of 57:1
APHA bubbler method, 24-hr samples
Extrapolation based on
CS2:H2S ratio of 10:1
Lead Acetate Tape
REFERENCE
1
2
1
3
4
5
References;
1 Mincer, R. A., and K. E. Noll, University of Tennessee, Knoxville, Regulations Governing Specific Odorous
Compounds, 1974.
2 Source emission data obtained through Tennessee Department of Public Health.
3 NIOSH, Occupational Exposure to Carbon Disulfide, 1977.
4 Pushkina, N. N., G. I. Korotkova, and Z. F. Chekanovskaya, Hygiene and Sanitation 35(9):476-477, 1970.
5 Resources Research, Inc., Hydrogen Sulfide Measurements Near Canton and Enka, N.C., December 1967 through
November 1968.
-------
C. TEST OF HYPOTHESIS
1. Selection of Sites
Sites were selected for hypothesis testing by surveying all towns
that have had viscose rayon plants for a period of 30 or more years.
Criteria for selection were:
• Elevated county mortality rates for the diseases involved
in the hypothesis.
• Little or no presence of other industry which might lead
to confounding emissions and/or effluents.
• Sufficient population at risk from plant emissions for
an epidemiclogic study.
• -Existence of emission and/or ambient air data on the
pollutants of interest.
• Available history of industrial activity (date of original
plant construction, duration of production of viscose rayon,
production capacity, etc.).
Table VIII-2 summarizes data for 15 locations.
The final selection was made by site visits to assess geographic,
topographic, and demographic characteristics. Four sites were visited:
Enka, North Carolina; Elizabethton, Tennessee; Lowland, Tennessee; and
Lewistown, Pennsylvania. Elizabethton and Lewistown were chosen as the
best sites because of the close proximity of the population to the emis-
sion source. Lewistown's plant was closed by flooding in 1972 and has
not reopened. This was not a serious disadvantage because the period of
operation (1921-1972) was appropriate to our study. It also offered the
possibility of detecting a downtrend in mortality or morbidity; however,
this could not be done with the available data.
VIII-10
-------
Table VIII-2. VISCOSE RAYON-PRODUCING TOWNS AND THEIR APPROPRIATENESS FOR EPIDEMIOLOGIC STUDY
TOWN, STATE
Enka,
North Carolina
Elizabethton,
Tennessee
Lowland,
Tennessee
Front Royal,
Virginia
Lewis town,
Pennsylvania
Childersberg,
Alabama
Fairhaven,
Vermont, New York
Painesville,
Ohio
Utica,
New York
Preder icksburg ,
Virginia
Nitro,
West Virginia
Parkersburg,
West Virginia
LeMoyne ,
Alabama
COUNTY
Buncombe
Carter
Hamblen/Green
Warren
Mifflin
Talladega
Shelby
Rutland (Vt.)
Washington, (N.Y.)
Lake
Oneida/Herkimer
Stafford/
Spotsylvania
Putnam/Kanawha
Kanawha
Mobile
7 / / / / */***/ * / « / *«*
No
Yes
Yes
Yes
Yes
No
No
No
Yes
Yes
Yes
No
No
Yes
No
Yes
Yes
No
No
Yes
Yes
Yes
Medium
Low
Medium
Low
Low
Medium
Medium
Low
Low
Medium
High
High
High
High
High
60,000
12,000
>1,000
8,000
11,000
3,000
22,000
90,000
15,000
8,000
44,000
190,000
Yes
Yes
Yes
Yes
No
Suburban
Urban
Rural
Urban
Urban
1929
1925
1948
1941
1921
1949
1969
1936
1927
1952
t WMMR = White male mortality rate for myocardial infarction
NMMR = National mean mortality rate for myocardial infarction
WFMR *> White female mortality rate for myocardial infarction
-------
2. Elizabethton Study
a. Site Description
The Beauni.t Fibers plant in Elizabethton, Tennessee, commenced
operation in the mid-1920's. It has produced rayon by the viscose pro-
cess from that time up to the present. A neighboring yarn plant (Bem-
berg) emitted detectable quantities of ammonia into the atmosphere from
the cuprammonium process. This plant was shut down in 1975.
Elizabethton is located in Carter County, Tennessee, which is in the
northeast corner of the state. It is a town of about 12,000 population
located at the confluence of the Doe and Watauga Rivers. The Beaunit
and Bemberg plants are west of the center of town. Two to three hundred
homes in Elizabethton are 600 to 2000 meters south of the plant. The
remainder of the population lives 0.8 to 4 kilometers east of the plant.
The Beaunit and Bemberg plants are located in the Watauga River
Vhlley about 1500 feet above sea level (see Figure VIII-2). Hills sur-
rounding the plants range from 300 to 600 feet above the plant. Data
in Section b below indicate that annual average concentrations of carbon
disulfide would be on the order of 130 to 450 ug/m^ in Elizabethton.
The entire population of Elizabethton is postulated to be at risk as
evidenced by the transport analysis in the following section.
Jb. Population Exposure
Exposure of the Elizabethton population to plant emissions can be
estimated in two ways:
• From field measurements of hydrogen sulfide (Mincer and
Noll, 1974; see Table VIII-1 above). These showed means
of 14.8 and 1.2 pg/m3 in the vicinity of the plant. Ap-
plying a ratio of CS2:H2S = 57:1 (see below), this implies
mean carbon disulfide concentrations of 68 to 844 ng/m3
(in the conditions of the field tests).
• From estimated emission rates, used as source strengths in
a diffusion model. Estimated rates are shown in Table
VIII-3. The ratio CS2:H2S ranges from 2:1 to 57:1, and we
chose to use the figure from the Elizabethton plant, i.e.,
57:1.
VIII-12
-------
I
M
CO
^. / II - I •- =•
Figure VJJJ-2. Map of" Elizabethton, Tennessee
-------
Table VIII-3. CS2 AND H2S EMISSION INFORMATION FROM VISCOSE RAYON PLANTS
SITE/ SOURCE
Elizabethton, Tennessee
(Beaunit Fibers)
Enka, North Carolina
(American Enka)
EPA
NIOSH
EMISSIONS (Ib/ton rayon)
CS2
960
551
300
55
H2S
16.75
120
54
6
CS2:H2S
RATIO
57:1
5:1
6:1
9:1
2:1
to
10:1
METHOD
Engineering Calculation
Engineering Calculation
Analysis of Emission
Points
Engineering Calculation
(reliability rating: E)
REFERENCE
1
2
3
4
<
M
M
M
I
References;
1 Source emission data obtained through Tennessee Department of Public Health; also contained
in EPA's National Emissions Data System.
2 Personal correspondence, C. T. Mitchell, American Enka Company.
3 EPA, Air Pollutant Emission Factors, AP-42
4 NIOSH, Criteria for a Recommended Standard: Occupational Exposure to Carbon Disulfide,
p. 23, 1977.
-------
The diffusion calculations were developed as described in Appendix
A, which presents a method of modeling long-term concentrations of pol-
lutants based on emission characteristics of the plant and local mete-
orology. Turner's model [15], equation (4) in the appendix, was calcu-
lated using the input data presented in Table VIII-4. The plume rise
equation according to Holland was used to calculate H (see Appendix A,
p. A-3). The relative frequency function was obtained from a STAR pro-
gram (National Climatic Center,Asheville, North Carolina) run for Tri-
City Airport, 13.5 miles from Elizabethton, and based on observations
from January of 1966 to December of 1970, 8 observations per day. This
particular STAR was run for only five Pasquill stability classes: three
unstable conditions, the neutral condition, and an aggregate of all
stable conditions.
Long-term concentrations of CS2 were calculated for hypothetical
receptors at 0.25, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 4, 5, and 10 kilometers
from the plant in 16 directions (points of the compass)(see Table VIII-5)
Radial grid lines were superimposed on a topographic map of Eliza-
bethton and the calculated concentrations placed on the grid at appro-
priate distances and directions. Analysis centered on concentrations
in the eight primary directions (N, NE, E, SE, S, SW, W, NW) since there
is slight bias in reporting of wind direction that is partial to these
directions. Topographic considerations were next addressed. The pol-
lutant concentration was assumed to be greater due to wind flow patterns
in valleys and less on mountains and ridges. The resulting concentra-
tion isopleths can be seen in Figure VII1-3. It must be emphasized that
the intention is only to show that significant exposure is not incon-
sistent with available data. An extensive field study, with a grid of
sampling stations and on-site meteorological observations over a year
or more, would be needed for reliable quantitative estimates of popula-
tion exposure. However, it appears that the calculated estimates may
be of the right order of magnitude. Table VIII-1 shows a mean H2S
[15] Turner, D. B., Workbook of Atmospheric Dispersion Estimates, PHS
Publication 999-AP-26, 1969.
VIII-15
-------
Table VIII-4
INPUT PARAMETERS FOR HOLLAND PLUME RISE EQUATION AND TURNER MODEL
Parameter
Physical stack height
Stack gas exit velocity
Inside stack diameter
Stack gas temperature
Ambient temperature
Average wind speed
Atmospheric pressure
Effective stack height
Relative frequency function
CS2 emission rate
Value
12.2 m
15.24 m/sec
0.66 m
302°K
287. 1°K
1 . 5 m/sec
1.013 x 103 millibars
22.85 m
STAR routine
(Tri-City Airport, TN)
112 g/sec
Concentrations determined at distances of 0.25, 0.5,
0.75, 1, 1.5, 2, 2.5, 3, 4, 5, and 10 kilometers
Standard deviation of vertical component of plume
determined for appropriate distance for appropriate
stability class
Reference
1
1
1
1
2
-
-
3
4
1
5
References:
Tennessee Department of Public Health, Source Emission Data
for Beaunit Plant, Elizabethton, Tenn., 1975
U.S. Department of Commerce, Environmental Science Services
Administration, Climatic Atlas of the United States, 1978.
Calculated from above data after the method of Holland
(see Appendix A).
STAR Tabulation, Bristol, Tenn., National Climatic Center,
Asheville, N.C., 1971.
Turner, D. B., Workbook of Atmospheric Dispersion Estimates,
PHS Publication 999-AP-26, 1969.
VIII-16
-------
Table VIII-5
CALCULATED LONG-TERM CONCENTRATIONS OF CARBON DISULFIDE EMISSIONS AT
VARIOUS DISTANCES FROM BEAUNIT FIBERS PLANT, ELIZABETHTON, TENNESSEE
Wind
Direction
N
HUB
HE
ENE
E
ESE
SB
SSE
S
SSN
SH
HSH
W
WNW
NW
HNH
Estimated Long-Term Concentrations of CS2 (ug/m3)
Distance (km)
0.25
199
215
346
404
578
276
152
99
237
158
322
330
423
277
305
110
0.5
223
275
578
596
649
274
138
106
234
169
390
374
460
307
244
102
C.75
170
221
491
489
503
207
101
82
172
130
299
275
335
226
166
75
1
127
169
333
375
378
154
74
61
126
97
223
201
244
165
117
55
1.5
76
103
238
230
227
92
44
37
75
59
134
119
143
97
66
32
2
51
70
162
156
153
62
29
25
50
39
90
79
95
65
43
22
2.5
37
51
118
114
110
44
21
18
36
28
65
57
68
47
31
16
3
28
39
91
88
85
34
16
14
28
22
50
43
52
36
23
12
4
19
26
60
58
56
22
11
9
18
14
33
28
34
23
15
8
5
13
19
44
42
40
16
8
6
13
10
23
20
24
17
11
6
10
5
7
16
15
14
6
3
2
5
4
8
7
8
6
4
2
-------
I
(-•
CD
V-J ) '.«.
Estimaled Annual Averages til Concentration of Carbon Disullkte
In the Vicinity of Elizabethlon. Tennessee ( a g/m*)
O
Figure VIJJ-J. Concentration Isopleths: Elizabethton, Tennessee
-------
concentration of 14.8 yg/m3 at 610 m W of the plant, and 1.2 yg/m3 at
1070 m ENE. Applying the CS2:H2S ratio of 57:1 (see Table VIII-3), the
estimated CS2 concentrations are 844 yg/m3 and 68 yg/m3. Interpolation
in Table VIII-5 gives concentrations of 600 yg/m3 and 200 yg/m;5 at these
points.
c. Mortality
Age-adjusted mortality rates (standardized to the 1970 U.S. popula-
tion distribution) were computed for deaths occurring between 1965 and
1975. Table VIII-6 presents these data and rate ratios for Elizabethton
white males and two sets of controls (i.e., (1) Carter County excluding
Elizabethton and (2) Athens, Newport, and Greenville). Table VIII-6
presents the same data for white females.
It is apparent from Table VIII-6 that cardiovascular disease mor-
tality is not elevated in Elizabethton as compared with its controls.
In fact, among males, only cancer of the urinary bladder appears ele-
vated in Elizabethton when compared to both sets of controls. The only
other causes of death which show appreciable elevation compared to
either control are cancers of the oral cavity, pharynx, esophagus,
stomach, and pancreas, and other diseases of the urinary system. None
of these rate ratios exceeds 2.4, and their significance, both statis-
tically and practically, is dubious.
Among white females, diseases of the liver, gallbladder and pan-
creas; bronchitis, emphysema and asthma; and cancer of the colon or
rectum are elevated in Elizabethton compared to both controls. Rate
ratios for these diseases range from 2.0 to 4.0. Cardiovascular mor-
tality, with which we are particularly interested, shows no elevation
in Elizabethton and, if anything, appears in excess in the controls.
VIII-19
-------
Table VIII-6. RGB-ADJUSTED MORTALITY RATES (PER 100,000 POPULATION), 1965-1975,
ELIZABETHTON AND CONTROLS: WHITE MALES, WHITE FEMALES
I
KJ
O
CAUSE OF DEATH
Infectious « parasitic diseases
Cancer, oral, pharynx t. esophagus
Cancer , stcoach
Cancer, duodenum & small intestine
Cancer, colon & rectu»
Cancer, liver t bile ducts
Cancer, pancreas
Other digestive cancers
Cancer , respiratory organs
Cancer, genitourinary organs except bladder
Cancer of the bladder
Other cancers except leukemia
Neoplasms of lymphatics
Endocrine, nutritional, metabolic t blood diseases
Mental & nervous system diseases
Rheumatic heart disease
Hypertensive disease
Cardiovascular diseases
Cerebrovascular diseases
Respiratory diseases
Bronchitis, emphysema t, asthma
Pneumoconiosis
Other digestive diseases
Diseases of stomach & duodenum
Appendicitis, hernia & others
Gastroenteritis, noninfectious
Diseases of liver, gallbladder 6 pancreas
Diseases of urinary system
Diseases of genital organs
Diseases of pregnancy t childbirth
Diseases of skin & musculoskeleton & ill-defined
diseases
Congenital anomalies
Perinatal diseases
Accidents, poisons & violence
WHITE MALES MORTALITY RATES
Elizabethton
12
28
22
0
37
3
15
0
12S
44
18
42
33
38
34
.4
8
881
148
87
45
7
0
16
7
0
28
42
0
0
15
0
0
136
Control
• 1
7
6
9
0
30
5
7
0
113
26
1
53
39
35
22
15
19
895
134
87
68
5
0
20
9
0
34
20
8
0
15
2
0
139
Rate
Ratio
1.7
4.7
2.4
-
1.2
0.6
2.1
-
1.1
1.7
18.0
0.8
0.8
1.1
1.5
0.3
0.4
1.0
1.1
1.0
0.7
1.4
-
0.8
0.4
-
0.8
2.1
0
-
1.0
0
-
1.0
Control
12
16
36
25
2
21
6
26
2
140
26
3
47
35
52
23
5
56
1263
174
87
42
0
2
21
16
2
40
28
6
0
39
3
0
142
Rate
Ratio
0.8
o.a
0.9
0
1.8
0.5
0.6
0
0.9
1.7
6.0
0.9
0.9
0.7
1.5
0.8
0.1
0.7
0.9
1.0
1.1
0
0
0.8
0.4
0
0.7
1.5
0
-
0.4
0
-
1.0
WHITE F.EJlftLES MORTALITY RATES
Elizabethton
3
3
6
0
54
9
0
0
20
26
3
73
18
27
3
13
3
349
107
6
24
0
0
6
3
0
34
24
0
0
25
7
0
28
Control
»1
T
9
14
1
IB
6
4
2
18
6T
6
60
15
53
9
15
11
4O1
143
38
12
0
0
6
6
0
9
13
0
0
18
7
O
54
Rate
Ratio
1.0
0.3
0.4
0
3.0
1.5
0
0
1.1
0.4
O.5
1.2
1.2
0.5
0.3
0.9
0.3
0.9
0.7
0.7
2.0
-
-
. 1.0
0.5
-
3.8
1.8
-
-
1.4
1.0
-
0.5
Control
»2
5
9 _
8
1
23
5
14
O
21
51
5
73
14
48
22
14
32
450
108
28
6
0
1
3
10
1
15
16
0
0
15
3
0
34
Rate
Ratio
0.6
0.3
0.8
0
2.3
1.8
0
-
1.0
0.5
0.6
1.0
1.3
0.6
0.1
0.9
0.1
0.8
1.0
0.2
4.0
-
0
2.0.
0.3
0
2.3
1.5
-
-
1.6
2.3
-
0.8
Control Hi: carter County
Control 12: Athens, Newport, Greenville
-------
3. Lewistown Study
a. Site Description
Viscose rayon production was brought to the United States; by the
British textile firm of Courtaulds & Co., Ltd., in the 1910's. The
American company, called the American Viscose Corporation, built its
third plant in Lewistown, Pennsylvania. This plant began producing
rayon fiber in August of 1921. During the Second World War, the Ameri-
can Viscose Corporation was nationalized and sold to American investors
by the British Government. FMC Corporation obtained control of the five
or so plants formerly owned by American Viscose in the late 1940's.
In 1975, FMC sold its plants to Avtex Fibers, Inc.
The production of viscose rayon at the Lewistown plant was abruptly
ended in the summer of 1972 when Hurricane Agnes caused the Juniata
River to crest above the plant grounds, ruining machinery and stopping
production. Since the demand for rayon was not great at that time, FMC
decided not to reopen the plant.
Lewistown is located at a bend in the Juniata River. The dormant
American Viscose plant is directly across the river from the center of
town. The plant is less than a mile from the majority of the 11,000
residents of Lewistown who live along the bend in the river (see map,
Figure VIII-4). Major topographic features influencing the air movement
in the Juniata Valley include Shade Mountain to the east, Jack's Moun-
tain to the north, and a small peak directly behind the plant.
EPA's Environmental Photographic Interpretation Complex (EPIC) pro-
vided some input into the Lewistown study. EPIC personnel surveyed
aerial photographs of Lewistown taken in 1938 and 1964. Overlays were
constructed, outlining residential and commercial centers in both of
those years. Factories, stacks, and sewage treatement facilities were
also delineated and noted on maps at the appropriate time intervals.
There was no indication of substantial changes in the plant or in popu-
lation distribution, suggesting that the Lewistown situation has been
fairly stable over a long period.
VIII-21
-------
H
H
NJ
Figure VIII-4. Map of Lewistown, Pennsylvania
-------
b. Population Exposure
It was not possible at Lewistown to establish a plausible estimate
of air. pollutant transport and population exposure as for Elizabethton.
No ambient air monitoring had been done, and the meteorological data
were judged to be quite inadequate to support theoretical estimates.
Since it is now impossible to measure local concentrations of CS2 and
H2S (unless the plant is reactivated), we are dependent on general ap-
praisal of the likelihood of significant population exposure. The
higher exhausting of fumes at Lewistown (350-foot stacks) would con-
siderably reduce exposure close in to the plant, but would not be ex-
pected to protect more distant residents from some exposure. In fact,
a 1966 report [16] on a study of pollution from the American Viscose
factory states that, "You can smell them [the plant] every once in a
while." Since the odor threshold for H2S, the more odorous of the two
compounds, is approximately 30 ug/m3 [17], the implication is that dur-
ing these episodes the 082 concentration was on the order of a few hun-
dred yg/m3.
c. Mortality
As was done for Elizabethton, age-adjusted mortality rates for 34
disease categories for Lewistown were compared with two sets of controls.
In one analysis, rates from 1965-1972 were compared to those for Mifflin
County excluding Lewistown (Control #1). In the second analysis, rates
for 1965-1975 were compared with those for the aggregate population of
Berwick, Shamokin, Mt. Carmel, Chambersberg, and Sunbury. Age-adjusted
mortality rates and rate ratios are presented in Table VIII-7 for white
males and in Table VIII-8 for•white females.
It can be readily seen from Table VIII-7 that mortality from cardio-
vascular disease is not elevated in Lewistown white males as compared to
its controls. In fact, for congenital anomalies only do Lewistown males
[16] Premerlani, R., A Case Study of American Viscose Division of FMC
Corporation of Lewistown, Pa., December 1966.
[17] National Institute for Occupational Safety and Health, Criteria for
a Recommended Standard: Occupational Exposure to Hydrogen Sulfide,
Washington, D.C., May 1977.
VIII-23
-------
Table VIII-7
AGE-ADJUSTED MORTALITY RATES (PER 100,000 POPULATION),
LEWISTOtfN AND CONTROLS: WHITE MALES
CAUSE OF DEATH
Infectious c parasitic disease*
Cancer, oral, pharynx t esophagus
Cancer, stomach
Cancer, duodenum S small Intestine
Cancer, colon s rectum
Cancer, liver & bile ducts
Cancer, pancreas
Other digestive cancers
Cancer, respiratory organs
Cancer, genitourinary organs except bladder
Cancer of the bladder
Other cancers except leukemia
Neoplasms of lymphatics
Endocrine, nutritional, metabolic s blood diseases
Mental 6 nervous system diseases
Rheumatic heart disease
Hypertensive disease
Cardiovascular diseases
Cerebrovascular diseases
Respiratory diseases
Bronchitis, emphysema s asthma
Pneumoconiosis
Other digestive diseases
Diseases of stomach S duodenum
Appendicitis, hernia £ others
Gastroenteritis, noninfectious
Diseases of liver, gallbladder c pancreas
Diseases of urinary system
Diseases of genital organs
Diseases of pregnancy s childbirth
Diseases of skin t musculoskeleton t ill-defined
diseases
Congenital anomalies
Perinatal diseases
Accidents, poisons 6 violence
MORTALITY RATES
Lewistown
1965-1972
11
11
11
0
52
4
27
0
111
30
0
45
33
53
26
a
23
974
136
67
22
8
0
12
IS
0
24
19
0
0
4
4
0
106
Control
• 1
2
10
7
2
39
5
12
5
62
14
3
28
29
25
7
16
12
702
84
74
20
7
0
4
9
0
6
18
2
0
7
0
0
81
Rate
Ratio
5.5
1.1
1.6
0
1.3
0.8
2.3
0
1.8
2.1
0
1.6
1.1
2.1
3.7
0.5
1.9
1.4
1.6
0.9
1.1
1.1
-
3.0
1.7
-
4.0
1.1
0
-
0.6
-
-
1.3
MORTALITY RATES
Lewis town
1965-1975
11
11
11
0
63
4
31
0
157
41
0
64
40
60
26
15
23
1266
173
100
47
8
0
12
23
0
31
19
0
0
7
4
0
120
Control
12
23
27
16
1
65
9
18
1
137
32
8
64
29
42
27
29
30
1197
167
117
49
121
0
14
28
0
67
20
7
0
12
1
0
92
Rate
Ratio
0.5
0.4
0.7
0
1.0
0.4
1.7
0
1.1
1.3
0
1.0
1.4
1.4
1.0
0.5
0.8
1.1
1.0
0.9
1.0
0.1
-
0.9
0.8
-
0.5
1.0
0
-
0.6
4.0
-
1.3
Control'Hi Nlfflin County (excluding
Boroughs of Lewistown and
Juniata Terrace and Derry
Township)) 1965-1972.
Control 12i
Berwick, Shanokln, Nt.
1965-1975.
Camel, Chambersberg, Sunburyi
VIII-24
-------
Table VIII-8
AGE-ADJUSTED MORTALITY RATES (PER 100,000 POPULATION),
LEWISTOWN AND CONTROLS: WHITE FEMALES
CAUSE OF DEATH
Infectious £ parasitic diseases
Cancer, oral, pharynx t, esophagus
Cancer , stomach
Cancer, duodenum £ small intestine
Cancer, colon & rectum
Cancer, liver & bile ducta
Cancer, pancreas
Other digestive cancers
Cancer , respiratory organs
Cancer, genitourinary organs except bladder
Cancer of the bladder
Other cancers except leukemia
Neoplasms of lymphatics
Endocrine, nutritional, metabolic c blood diseased
Mental & nervous system diseases
Rheumatic heart disease
Hypertensive disease
Cardiovascular diseases
Cerebrovascular diseases
Respiratory diseases
Bronchitis, emphysema I asthma
Pneumoconioais
Other digestive diseases
Diseases of stomach t duodenum
Appendicitis, hernia & others
Gastroenteritis, noninfectioua
Diseases of liver, gallbladder & pancreas
Diseases of urinary system
Diseases of genital organs
Diseases of pregnancy t childbirth
Diseases of skin t musculoskeleton & ill-defined
diseases
Congenital anomalies
Perinatal diseases
Accidents, poisons & violence
MORTALITY RATES
Lewistown
1965-1972
3
9
0
0
28
0
8
0
14
45
0
93
17
25
3
12
11
302
49
15
0
0
0
5
2
0
17
IB
2
0
6
2
0
35
Control
• 1
13
2
0
0
24
9
6
O
7
25
2
63
17
19
8
13
22
452
91
28
5
0
0
2
16
• o
6
8
0
0
4
2
0
21
Rate
Ratio
0.2
4.5
-
-
• 1.2
0
1.3
-
2.0
1.8
0
1.5
1.0
1.3
0.4
0.9
0.5
0.7
0.5
0.5
0
-
-
2.5
0.1
-
2.8
2.3
0
-
1.5
1.0
-
1.7
MORTALITY RATES
Lewistown
1965-X975
3
-------
exceed both sets of controls. Among the causes for which Lewistown
white males appreciably exceed either control are infectious and para-
sitic diseases, cancers of the pancreas and genitourinary system, mental
and nervous system diseases, and diseases of the liver, gallbladder, and
pancreas.
Deaths from cancers of the oral cavity, pharynx, or esophagus among
white females are more frequent in Lewistown than in either control area.
In both analyses, the rate ratio was 4.5. Also elevated when compared
to both controls was the category "diseases of the urinary system," with
rate ratios of 2.3. Mortality from diseases of the stomach, duodenum,
liver, gallbladder, and pancreas was elevated in Lewistown as compared
to the rest of Mifflin County, although the rate ratio did not exceed
2.8.
4. DiscussiQn of Results
Ultimately, the plausibility of an industry-community disease hy-
pothesis is dependent on the "exposed" community exhibiting higher-than-
expected mortality from diseases which have been etiologically and mete-
orologically associated with the industry.
One can immediately see that mortality from cardiovascular disease,
which has been linked etiologically to the viscose rayon industry, is
not elevated in'either study site for either sex. This result in males
is somewhat surprising, although the occupational mortality (probably
depend.ent on heavy exposure) would not be expected to contribute much
to the mortality in Elizabethton or Lewistown.
Turning to other diseases, it appears that there is an excess of
diseases of the liver, gallbladder, and pancreas among females in both
study sites. Cirrhosis would seem to predominate in this disease group,
and it is unlikely that the excess is environmental in nature. It may
be reasonable to investigate further the apparent bladder cancer excess
found in Elizabethton, as the disease has a strong occupational and
environmental component.
On the basis of the industrial, demographic, and environmental
characteristics of Elizabeth and Lewistown, we would expect both to show
VIII-26
-------
similar excesses of industry-related mortality if any did, in fact,
exist. These patterns are not evident, however, particularly with re-
spect to cardiovascular disease. It is, therefore, unlikely that the
viscose rayon industry can be hypothesized as contributing to cardio-
vascular disease mortality in communities containing the industry.
D, CONCLUSIONS
The following conclusions have been derived from the two field in-
vestigations of the viscose rayon industry.
(1) Carbon disulfide and hydrogen sulfide are emitted from viscose
rayon plants and are present in the ambient air of nearby populated
areas.
(2) A hypothesis linking emissions (CS2/ H2S) from viscose rayon
plants to cardiovascular disease mortality in nearby populations is not
supported by the mortality data available for 1965-1975.
(3) Further study should be made of bladder cancer in males working
in the Elizabethton plant and of diseases of the liver, gallbladder, and
pancreas in females in both Lewistown and Elizabethton.
VIII-27
-------
Appendix A
ATMOSPHERIC TRANSPORT
-------
Appendix A
ATMOSPHERIC TRANSPORT
INTRODUCTION
By far the most satisfactory way of estimating population exposure
to air pollutants from industrial activity is from actual long-term
sampling at stations throughout the area. These measurements are rarely
available, and there are few to support our case studies. It is still
possible to make order of magnitude estimates to test the plausibility
of a hypothesized pathway from source to population at risk, by using
one of the established dispersion models. This is what we have done,
and this Appendix discusses the procedures.
We wish to emphasize here the limitations that we discuss at the
end of the Appendix. Accurate estimates of ground level concentration
from source strength and meteorological data are not easy to make, even
in uncomplicated terrain. They require frequent source sampling and
long-term micrometeorological observations at several stations. We have
had to rely in general on guesses for source strength and on meteor-
ological data from one or two stations 10 to 200 miles away.. The best
that can be expected from such calculations is an indication that one's
hypothesis is not inconsistent with climatological data.
Equations used to estimate long-term average concentrations from
point sources are generally similar in form. Several of the models
which are used in the case studies use the long-term concentration
estimates from equations similar to those developed in this Appendix.
The models then go on to perform other manipulations on the data.
Input data were acquired and equations were solved to give estimates
of exposure. Input parameters for the long-term concentration equation
are source-specific stack and emission data and meteorological data.
The meteorological data have been manipulated in a STability ARray
(STAR) program as discussed by Doty and Wallace [1]. The output of
the STAR program is a frequency of occurrence and relative frequency
distribution of various winds by wind direction, wind speed, and
stability.class. The relative frequency is inputted into each of the
various dispersion models. Models which are used in various chapters
of the report include the following:
Air Quality Display Model (AQDM) [2]
[1] Doty, S. R., and B. L. Wallace, A Climatological Analysis of Pas-
quill Stability Categories based on 'STAR' Summaries, NOAA, National
Climatic Center, April 1976.
[2] TRW Systems Group, Air Quality Display Model, PB 189 194. November
1969.
A-l
-------
• Climatological Dispersion Model (COM) [3]
• Long-term average concentrations calculated after
the work of Turner [4] and Gifford [5]
The AQDM and COM are similar models which are used by EPA for long-
term pollution potential; the COM is an updated version of the AQDM.
All of the dispersion models above contain common features which are
discussed in this appendix.
DEVELOPMENT OF EQUATIONS USED IN REPORT
The Pasquill-Gifford plume formulation is the basis for the afore-
mentioned models. This formulation is based on the dispersion of a
plume as it spreads out from a stack. The diagram below illustrates
the principles used in this type of dispersion modeling.
plume center! ine
Source
Receptor
[3] Busse, A. D., and J. R. Zimmerman, User's Guide for the Climato-
logical Dispersion Model, National Environmental Research Center,
EPA-R4-73-024, December 1973.
[4] Turner, D. B., Workbook of Atmospheric Dispersion Estimates, PHS
Publication 999-AP-26, 1969.
[5] Gifford, F. A., Ch. 3, pp. 65-116, in U. S. Atomic Energy Commission,
Meteorology and Atomic Energy, 1968.
A-2
-------
Several features should be noted about the source-receptor relation-
ship indicated in the above diagram:
• The source and any point in the plume may be identified by
a set of x,y,z Cartesian coordinates with their origin at
the base of the stack.
• The effective height of emission (H) is the height at which
the plume centerline becomes horizontal. It is the stack
height (h) plus an incremental factor (Ah) related to the
stack parameters, buoyancy and vertical momentum of the
stack effluent. The incremental factor used to determine
effective height of emission in most of the case studies
in this report was determined according to the method of
Holland [6]:
vsd ( _3 T - T
Aft = -a- 1.5 + 2.68 x 10 3 p
Ts
where:
Aft = the rise of the plume above the stack, m
d = the inside stack diameter, m
vs = stack velocity, m sec"1
; u = wind speed, m sec
p = atmospheric pressure, mb
Ts = stack gas temperature, °K
Ta = air temperature, °K
and 2.68 x io~3 is a constant having units of mb~1m~1.
The study on copper smelters used the plume rising according
to Briggs [7].
Within the plume, a Gaussian distribution of stack effluent
is assumed. The limits of the plume dispersion at any dis-
tance, x, are represented by one standard deviation in the
y- and z-directions (Oy,oz) at that distance. At some point,
L, a limit of dispersion in the z-direction is reached; this
is called the depth of the mixing layer and represents the
limiting value of 0Z.
[6] Holland, J. Z., A meteorological survey of the Oak Ridge Area,
pp. 554-559, in Atomic Energy Commission Report ORO-99, Washington,
D.C., 1953.
[7] Briggs, G. A., Plume Rise, USAEC Critical Review Series, TID-25075,
National Technical Information Service, Springfield, Virginia, 1969.
A-3
-------
• Wind speed (u) and atmospheric turbulence (S) are important
parameters affecting the plume.
• The concentration of material from the plume (X) affecting
. a receptor on the ground at distance x and on the plume axis
may be determined by solving the equation
X(x,y,z;H) =
for y,z =0,
exp
IJL
2 Oy
exp
1 H
(1)
X(x,0,0;tf) =
exp
(2)
The long-term effect of wind direction frequency distribution is
dealt with; as follows: The circumference at distance X is divided
into 16 sectors, each corresponding to a wind vector which blows toward
it. The sector width is then 2irx/16. If it is assumed that the popu-
lation is, on the average, located along the centerline of the plume,
the following equation describes the long-term average concentration
to which it is exposed:
X =
JfiL
v/27 UO- (2TTX/16)
exp
z
(3)
The frequency function, f, is determined by the amount of time the wind
was blowing toward the sector, e.g. from a wind rose.
Concentration of a pollutant at a given distance from a source will
also depend on atmospheric stability and wind speed. D. B. Turner [8]
has described the application of Pasquill stability categories. Since
these categories are used in many of the theoretical workups, they are
defined in Table A-l. The stability classifications which are deter-
mined from STAR programs are ground-level meteorological observations
(surface wind speed, cloud cover, ceiling) supplemented by solar ele-
vation data (latitude, time of day, time of year) as inputs. No ver-
tical sounding data are used.
[8] Turner, D. B., A diffusion model for an urban area,
Weteoro2. J(l):83-9l, February 1964.
J. Appl,
A-4
-------
Table A-l. STABILITY CLASSES AND DEFINITIONS*
Pasquill
Stability Definition
Class
A Extremely unstable
B Unstable
C Slightly unstable
D Neutral (day/night)
E Slightly stable
F,G Stable to extremely stable
*
The STAR program can be run for up to eight stability classes,
including a breakdown of the neutral category into day and
night divisions.
The modification in Equation (3) which results from the use of dir-
ection-specific stability frequencies is that f now becomes a function
of stability, direction, and wind speed [f = f(Q;S,N), where N repre-
sents a given wind speed class]. A direction- and stability-specific
frequency is the output of the STAR routine. In order to determine
the long-term ground-level concentration, the following formula after
the work of Turner [4] is used to give a more accurate picture:
S N
X<*,8) = - " > l&f-J&t&JD
azs
2KsJ
(4)
Thus, at a given distance for a particular wind direction, the long-
term concentration is determined by summating each wind speed (N) within
each stability (S), and then summating the concentrations for the range
of stabilities. Note that az varies with each stability and that the
wind speed, u, is determined by the midpoint of the wind speed class
which is being suramated. The frequency function f(Q,S,N) is output
from the STAR program. (It must be remembered that a northerly wind
will expose a population south of the plant.)
When determined for a series of distances and the 16 directions
for which STAR data are available, enough data become available to plot
long-term concentration isopleths; this was done in the copper smelting,
yiscose rayon, coal, and steel studies. The Air Quality Display Model
was used by EPA to develop theoretical SO* concentration isopleths in
the vicinity of Rocky Mountain smelters {9]. Results from this report
are presented in the chapter on copper smelting. Since elevation rela-
tive to the stack has a large effect on ground-level concentration,
this variable was included in that study.
[9J Estimated Impact of Smelting Operations on S02 Concentrations in
the Rocky Mountain Area, Staff Report, Air Quality Management
Branch, GAP, EPA, August 1972.
A-5
-------
For our studies in the rugged hill and valley terrain of Appalachia,
two different techniques were used to estimate concentration. In the
Johnstown study, published research was used to give an indication of
pollutant transport patterns in the Conemaugh Valley under very specific
meteorologic conditions.
Dispersion of coal dust from Wyoming County coal preparation plants
was modeled by using Equation (3) without STAR data. A representative
wind speed was selected, stability was assigned as neutral, sector
widths were determined by detailed topographic analysis, and the wind
direction was assumed to be entirely regulated by the orientation of
the river valley, flowing upvalley and downvalley with equal frequency.
This simplification of the Turner model provides the epidemiologist
with a high estimate of long-term dosage for coal dust in the ambient
air.
CAUTIONARY NOTES
As indicated at the outset of this Appendix, the models and equa-
tions presented in the theoretical sections of the chapters in the body
of the report are subject to a number of qualifying conditions. More
of these variables are taken into account in the full models than in
the instances when equations were used. However, results from all data
manipulations are roughly comparable, so the transport analysis does
serve its purpose from an epidemiological standpoint.
Shortcomings in the various estimates are outlined below:
Emission Inventory — Emission inventories are crude at best, espe-
cially in view of the long term pollution potentially leading to chronic
disease that was being investigated in this study. Little or no infor-
mation could be obtained on the history of emissions from the plants
under study. One method used was the emission factor without any
pollution controls as the historical, worst-case assumption. Changes
in process, production rate, and location make this a questionable
estimate. Reliable data on production history of the industries was
very hard to find.
Emission inventories are subject to error. Also, they use constant
average emissions which are not reflective of plant outflow, since pol-
lution type and quantity vary significantly on a daily, weekly, and
seasonal basis.
Topographic Considerations — Dispersion estimates and models are
worked out for a basically flat terrain. Although several sites are on
flat terrain, most have confounding topographic features within five
miles of the sites. Others are located in mountainous or hill-and-valley
terrain. In the chapter on copper smelting, a modified AQDM based on
terrain was used. For the rugged valleys of Wyoming County discussed in
the coal mining chapter, valley character and orientation were considered
to be the sole determinant of wind flows. Where terrain is more or less
flat (distant mountain, rolling hills), Equation (4) was applied using
available STAR data. The results are correspondingly more reliable.
A-6
-------
Climatologic Data — In most cases where STAR data were; used, it
had to be obtained from observing stations 10 to 200 miles away.
Mixing-Layer Depths — These are difficult to estimate from
ground-level observations. Rawinsonde measurements are more appro-
priate but are not generally available for sites in this study.
Nature of Study — Since the study is one of chronic diseases, we
are interested in the long-term levels of pollutants. This is more a
problem with the emission source (Q) than with the meteorological data
since the latter can be averaged over a multiple-year period and an
average taken. In addition, equations defining long-term concentrations
are worked out from short-term, point source measurements. Although we
do use point sources in this study, the shortcomings in use of these
equations for long-term averages must be considered.
SYMBOLS USED
d = inside stack diameter
f = frequency with which wind flows in given direction
f(Q,S,N) = frequency with which wind flows in given direction; output
of STAR routine (relative frequency function)
H = effective stack height
h = physical stack height
Ah = rise of plume above stack
L = height of the stable layer aloft (mixing height)
p '•= atmospheric pressure
Q = emission rate
S = stability class
ay = standard deviation of lateral component of plume
°z = standard deviation of vertical component of plume
°z5 = standard deviation of vertical component of plume as
function of stability class S
T = ambient temperature
• . «
TS = stack gas temperature
u = wind speed
u = average wind speed of wind speed class N
vs = stack gas exit velocity
X = long-term average concentration
x,y,z = distances in Cartesian coordinate system
A-7
-------
Appendix B
COPPER SMELTING INDUSTRY CHARACTERIZATION
-------
Appendix B
COPPER SMELTING INDUSTRY CHARACTERIZATION*
GENERAL DESCRIPTION OF THE INDUSTRY
The primary copper industry includes mining of ore, preparing ore
for smelting, smelting, and refining.
Ores
The primary copper industry is characterized by the type of ore being
processed. The bulk of most ore deposits are copper sulfides; however,
the upper portions of deposits are generally oxides (see Table 1). Gold,
silver, nickel, zinc, arsenic, and lead are often associated with copper
ore. The average grade of ore presently being mined is 0.7% copper.
The process physically separates copper ore from worthless rock, gangue,
and chemically separates copper ore from other elements. Different pro-
cesses are used to separate copper from the two general ore types (oxide
and sulfide). See Figure 1.
Table 1
COPPER ORES2
Sulfide:
Oxide;
Chalcocite
Covellite
Chalcopyrite
Bornite
Enargite
Tetrahedrite
Cuprite
Tenor ite
Malachite
Azurite
Chalcanthite
Brochantite
Cu2S
CuS
CuFeS2
Cu5FeSi»
CuaAsSi,
Cu3SbS2
Cu20
CuO
CuC03»Cu(OH)2
2CuC03'Cu(OH)2
CuSOi»«5H20
CuSO.,'3Cu(OH)2
This appendix was prepared by Carl B. Bailey.
B-l
-------
OXIDES
CONCENTRATING
ELECTROWINNING
MARKETABLE PRODUCT
CONVERTING
-».
FIRE REFINING
-*
ELECTROLYTIC
REFINING
1
MARKETABLE PRODUCT
Figure 1. PRIMARY COPPER PROCESSES
Mining
There are two techniques used to mine copper—underground and open
pit mining. Ore deposits containing a high percentage of copper and
located far below the surface are mined by the underground method. How-
ever, most high-grade (high-percentage) ore deposits have been depleted.
Low-grade ore, which requires the mining of large quantities of ore
(25,000 tons per day) to be profitable, is mined by the open pit method.
Copper Processes
Leaching: Oxide ores are treated with sulfuric acid which dissolves
the oxide and leaves the gangue unaffected. The ore is agitated and
soaked with the leach solution for several days. The copper is recovered
from the solution by combinations of various methods of washing, filter-
ing, settling, or electrowinning.
Electrowinning; The copper, now in the form of copper sulfate, can
be recovered from the leach solution by electrolysis. An insoluble anode
and copper cathode are placed in the solution creating cells through
which an electric current is passed. The copper is chemically released
from the sulfate and deposited on the cathode. Copper in this form is
98% pure and is marketed or further refined.
B-2
-------
Concentrating; Sulfide ores may be processed by various combinations
of the following processes: concentrating, roasting, smelting, convert-
ing, and refining. After the mined ore is crushed to a fine size, it is
violently agitated with air in a concentrator. Water and reagents in the
concentrator create a froth that adheres selectively to the sulfide ore.
As a result, the gangue sinks to the bottom of the tank and the ore (20%
to 30% Cu) floats to the surface and is removed.
Roasting: The roasting process heats the concentrate in one of two
types of roasters—multiple hearth or fluid bed. The copper-sulfur ratio
is adjusted for optimum results by removing sulfur as sulfur dioxide
after oxidation. Also, volatile material is expelled. The resulting
material, calcine, is charged while hot to a reverberatory or electric
furnace and smelted.
Smelting: The first part of the smelting process produces a molten
material of iron sulfides and copper sulfides, called matte, in a rever-
beratory (reverb) or electric furnace. In the molten state the matte is
heavier than the gangue and they separate. When the gangue is heated,
oxides which are called slag are formed. A flux, limestone or silica,
is added which aids the separation of slag from the matte. At this
point, the slag and matte are removed separately, and the matte enters
the converter.
Converting; In the converter, a stream of air is forced through the
matte. The iron and sulfur react with oxygen to release large amounts
of heat. As air flows through the molten matte, a flux (silica) is added
to form an iron-silicate slag, and removes all iron from the matte.
'2 CuFeS2 + 402 + 2 Si02 -»• 3S02 + 2 FeSi03 + Cu2S + A
.(matte) (air) (flux) (sulfur (slag) (cuprous
dioxide) sulfide)
The slag is removed and air continues to oxidize the sulfur.
Cu2S + 02 -»• 2 Cu + S02 + A
At the end of this reaction, all sulfur is oxidized and copper, some
copper oxides, and a few impurities remain. This material is 98.5%
copper, called blister copper.
Fire Refining: The blister copper is transferred to a small rever-
beratory furnace in a molten state or is melted in one. Again, air is
forced through the material to oxidize all impurities which rise to the
top and are removed. The ends of green logs (poling) or reducing gases
are then passed through the material to reduce the copper oxides to
copper metal. The molten metal is then cast into slabs (tough-pitch
copper) and sold or cast into anodes for electrolytic refining.
B-3
-------
Electrolytic Refining: The anodes are placed in cells with an elec-
trolyte (copper sulfate and sulfuric acid) and a pure copper cathode
starting sheet; then an electric current passes through the cell. Copper
ions migrate from the anode to the electrolyte and are deposited on the
cathode. The electrolytically refined copper contains about 99.9% copper.
SITE-SPECIFIC INDUSTRY CHARACTERIZATIONS
The plant processes, age, products, and air pollution controls of
the study sites are summarized in Table 2. The Kennecott and Magma
plants are relatively new, and the Inspiration plant is the only one in
the United States that uses the leaching and electrowinning processes.
EFFLUENTS AND EMISSIONS
Effluents
The plants in the study counties do not have distinct discharges.
Because of their geographic location, the sites in Arizona have an acute
water problem; consequently, no water is wasted by the industry. How-
ever, material from tailings ponds associated with concentrators can
leach into the ground. Rain can carry materials into the environment
from the mines and plants.
Surface water and rainfall are scarce in Arizona; consequently, water
loss at discharges is uneconomical. For instance, in the Miami-Hayden-
San Manuel area, the mean relative humidity is less than 40%: the mean
annual number of days with a maximum temperature of 90°F and above is
greater than 90 days; and the mean annual precipitation is only 18
inches. Process water used for acid plant blowdowns, cooling of anodes,
etc., is recycled. Operations that use a water stream to cool and granu-
late hot slag are not utilized by the study plants. Water is lost mainly
through evaporation, and there are no point source discharges.
Tailings pond associated with concentrators may be a source of con-
tamination to ground water resources. Water may leach the tailings if
they are soluble. Again, evaporation is probably the greatest method
of water loss. Also, the soil permeability will affect the amount of
leaching.
The mean annual 18 inches of rain in the study area may carry con-
taminants to water resources. Rain runoff from the plants may carry
plant materials as process chemicals and spills to water resources.
However, the annual rainfall does not seem to be a major or continual
agent contributing to pollution.
B-4
-------
Table 2
STUDY SITE PLANT DESCRIPTIONS3
COMPANY/LOCATION
ASARCO
Hay den, Arizona
Kennecott
Hayden, Arizona
Magrca
San Manuel ,
Arizona
Inspiration
Miami, Arizona
Magma
Superior, Arizona
,Phelps-Dodge
Douglas, Arizona
Phelps-Oodge
AJo, Arizona
Phelps-Oodge
Morend, Arizona
Anaconda
Anaconda, Montana
Kennecott
Magna. Utah
AGE
1912
1958
1955
1911
Closed
1971
1904
1917
1942
1903
1916
ELECTROLYTIC
REFINING
No
No
Yes
Yas
No
No
No
NO
Yes
PRODUCTS
Blister copper
180.000 t/yr.
(Anodes)
Blister copper
80.000 t/yr.
(Anodes)
Electrolytlcally
Refined
57,000 t/yr.
Fire-Refined
183,000 t/yr.
Electrolytlcally
Refined
75,000 t/yr.
Blister copper
118,000 t/yr.
Blister copper
4,100 t/yr.
F1 re-Refined
142,000 t/yr.
Blister copper
55,400 t/yr.
Fire-Refined
182,000 t/yr.
Fire-Refined
202.000 t/yr.
Ft re-R (fined
190.00J t/yr.
PROCESS DESCRIPTION
• Roasting (multiple
hearths)
• Smelting
t Converting
• Roasting (fluid
bed)
• Smelting
• Converting
• Smelting
• Converting
• F1re Refining
Refining
• Smelting
• Converting
t Fire Refining
• Smelting (Electric
Furnace)
• Converting
• F1re Refining
* Leaching
• Electrowlnnlng
• Roasting (multiple
hearth)
t Converting
t Fire Refining
• Smelting
• Converting
« Roasting (fluid
bed)
• Smelting
• Converting
t Fire Refining
• Drying (multiple
hearth)
• Smelting
• Converting
• Fire Refining
• Smelting
« Converting
• Fire Refining
B-5
-------
Emissions
There are three major sources of emissions in a copper smelter.
Tables 3 and 4 give the general ranges of S02 emissions associated with
each process. Other miscellaneous sources of emissions are fugitive
emissions. Such sources include:5 hot calcine transfer points, reverb
feed and discharge points, converter rollout, and leaks in flues and
ducts. Fugitive emissions are difficult to quantify and have been esti-
mated between 0 and 12%5 of the total amount of SOa generated by the
smelter. Table 5 lists volatile metals found in domestic copper ore
concentrates that may become emissions. Table 6 gives particulate and
SOa emission estimates for the copper smelters in Arizona. Since size
distribution of particles was not available, the unit "tons per day"
does not indicate the amount of respirable size (0.3 p to 5.0 y aerody-
namic diameter) emitted.
Table 3 •
RANGES OF S02 MISSIONS, PROCESS-SPECIFIC7
PROCESS
Roasting
Smelting
Converting
% TOTAL SO
With
Roasting
20-50
10-30
40-50
EMISSIONS
Without
Roasting
20-40
60-80
Table 4
EMISSION FACTORS3FOR PRIMARY COPPER SMELTERS WITHOUT CONTROLS6
TYPE OF OPERATION
Roasting
Smelting (Reverberatory)
Converting
Refining
Total Uncontrolled
PARTICULATES^
Ib/ton kg/Mt
45
20
60
10
135
22.5
10
30
5
67.5
SULFUR
Ib/ton
60
320
879
-
1250
OXIDES
kg/Mt
30
160
435
-
625
Emission factors expressed as units per unit weight of concen-
trated ore produced.
Electrostatic precipitators have been reported to reduce
emissions by 99.7 percent.
B-6
-------
Table 5
TYPICAL LEVELS OF VOLATILE METALS IN DOMESTIC COPPER ORE CONCENTRATES7
PLANT
ASARCO
Kennecott
Magma
Inspiration
Phelps-Dodge
Douglas
Phelps-Dodge
Mo rend
Phelps-Dodge
AJo.
Anaconda
LEAP
ZINC
ARSENIC
CADMIUM
BERYLLIUM
VANADIUM
ANTIMONY
TIN
Table 6.
PROCESS
Roast ing-Smelt ing
Converting
Roasting
(SOX Converting)
Smelting
(SOX Converting)
Smelting
Converting
Converting3
(Acid Plant)
Smelting
Converting
Roast ing-Smelt ing
Converting
Roast ing-Smelt ing
Converting
Smelt ing-Convert ing
Drying- Smeltlng-
Convertlng
CONCENTRATION LEVEL
<5000 ppm
5000 - 20,000 ppm
_> 20,000 ppm
10,000 ppm
10,000 - 40,000
i«o,ooo
200 - 1000 ppm
1000 - 10,000 ppm
>10,000
<1000 ppm
<10 ppm
<100 ppm
<200 ppm
>200 - 5000 ppm
>5000 ppm
<1000 ppm
Z CONCENTRATES SURVEYED
96Z
2Z
2Z
67%
32Z
1Z
88Z
10Z
2Z
100Z
100Z
100Z
97Z
3Z
-------
Control of emissions has been practiced only in recent years, and
'then with various degrees of efficiency. In general, many smelters use
•balloon flues and cyclones for particulate removal; however, these de-
vices are very inefficient in removing respirable size particles, as
shown in Figure 2. Electrostatic precipitators (ESP), on the other hand,
(are capable of removing respirable particles (Figure 2). For the removal
of S02, many problems exist and are well documented in the literature.
Proposed systems for the primary copper industry are:7
Sulfuric Acid Plants
Elemental Sulfur Plants
Calcium-Based Scrubbing Systems
Dimethylaniline (DMA) Scrubbing
Ammonia Scrubbing
Sodium Sulfite-Bisulfite Scrubbing
i 6AO PILTERHOUSE
.,; VENTURI SCRUBIER «-INCH THROAT. JO-INCM «ATER OAUGE)
I (MAT TOWER (M-POOT OIAUIT(R)
I DRY ELECTROSTATIC PRECIPITATOR IMCCONO CONTACT TIME)
(MULTIPLE CYCLONES (IMNCH DIAMETER TUBISI
{ SIMPLE CYCLONE (4-FOOT DIAMETER)
( INERTIAL COLLECTOR
10 40
(•ARTICLE SIZC. «il
so
10
Figure 2
COMPOSITE GRADE (FRACTIONAL) EFFICIENCY CURVES
BASED ON TEST SILICA DUST7
Often these processes require a constant, uniform gas stream; conse-
quently, the plant operation procedures may be modified. For example:5
• A minimum and maximum number of converters blowing at one time.
• Converters finish flowing one at a time.
In addition, collection devices are needed around emission sources in the
plant to reduce fugitive emissions, which would otherwise not be treated.
B-8
-------
In general, emissions were not controlled until the 1970's. Now,
particulates are usually well controlled and SOa is poorly controlled,
if|at all. The theory that "the solution to pollution is dilution" is
still practiced—ASARCO increased their converter stack height from 300
to 1,000 feet.
Figures 3 through 10 are schematic respresentations of site-specific
emission control devices.
Smelting Converting
"T i
WHB"1" Balloon Flue
Flue
I
Stack
fWHB = Waste Heat Boiler
Stack
Figure 3. INSPIRATION EMISSION CONTROL DEVICES
Roasting Smelting Converting
WHB 1971
Rust Engineering
Acid Plant 750 t/d
Water Spray
Chamber
ESP (98.3%)
Stack
Stack
ESP = Electrostatic Precipitator (Efficiency)
Figure 4. ASARCO EMISSION CONTROL DEVICES7
B-9
-------
Roasting Smelting Converting
Cyclones
WHB Balloon Flue
50%
ESP (50%) ESP (95%)
1968
Monsanto
Acid Plant
50%
(95%)
Stack Stack
Figure 5. KENNECOTT EMISSION CONTROL DEVICES7
Smelting Converting
ESP (98%) ESP (98%)
0-100%
Stack
Humidifying Tower
Mist Precipitator
I _
Converter
I
Absorbing Tower
Gas Cleaning
Acid Plant
Stack Stack
Figure 6. MAGMA EMISSION CONTROL DEVICES5
B-10
-------
Roasting Smelting Converting
I T T
ESP WHB ESP
I I
I i'
Stack Stack
Figure 7. PHELPS-DODGE/DOUGLAS EMISSION CONTROL DEVICES
Roasting Smelting Converting
r i
Parsons
Acid Plant ESP (78.5%) ESP (98.5%)
750 T/D
I
I
Stack Stack
Figure 8. PHELPS-DODGE/MORENCI EMISSION CONTROL DEVICES
Smelting Converting
ESP
I
Stack
Figure 9. PHELPS-DODGE/AJO EMISSION CONTROL DEVICES
B-ll
-------
Drying Smelting Converting
I
Water Spray
Flue
I
ESP (50%)
Stack
Figure 10. ANACONDA EMISSION CONTROL DEVICES
MAGMA: PROCESS DESCRIPTION & HISTORY5
The Magma Copper Company operates an underground mine, concentrator,
smelter, electrolytic refinery, and continuous rod-casting operation in
San Manuel, Arizona. Smelting of copper concentrates began in 1956 and
has had two expansions—in 1965 and 1971. Ore production increased from
30,000 to 40,000 tons per day (t/dj arid then to 62,500 t/d, respectively.
In 1975 Magma employed 4,151 people and produced 187,427 tons of copper.
The San Manuel Mine ore deposit is chiefly chalcopyrite with a copper
content of about 0.75%. Metallic minerals include pyrite, chalcopyrite;
chalcocite, mblygdenite, silver, and gold. The economically feasible
(from a mining standpoint) deposit was initially fishhook-shaped and up
to 400 feet thick. The average depth of overburden was 670 feet. The
daily capacity is 62,500 tons, and block caving is the mining technique
employed.
A train carries the ore from the mine to the plant six miles away.
The train of 36 100-ton-capacity cars makes 18 round trips daily.
Crushing, grinding, and flotation concentrate the ore. Primary,
secondary, and tertiary crushing reduce the ore to less than an inch in
size. Grinding in rod mills and then in ball mills precedes flotation.
Retention time in the flotation cells (1,056 cells of 40-cubic-foot
capacity, and 63 cells of 300-cubic-foot capacity) is nine minutes.
Concentrate from the 300-cubic-foot cells is reground in the ball mill
and recirciilated to the 40-cubic-foot cells. Molybdenum can be recovered
by flotation in another series of cells. The gangue from the cells is
now called tailings and is thickened and discharged to tailings ponds.
B-12
-------
The copper concentrate is fed to three reverb furnaces (102 feet
long). The concentrate is about 28% copper and 10% moisture. The side-
feed reverbs are oil-fired or natural gas-fired. Smelting occurs at
about .270P°F and combustion air is preheated. Slag is dumped close to
the plant. The matte is 32% to 35% copper.
The matte is transported to one of six Fierce-Smith type converters
by 300 cubic feet of ladles. An 18-hour period is required to produce
blister copper. Slag is returned to the reverbs.
Ladles again transport the molten copper to four holding furnaces
for fire refining, 99.7% copper. Anodes weighing 820 pounds are cast
and cooled by water sprays.
The electrolytic refinery came on line in 1971. Anodes are sus-
pended in refining cells, 4 inches apart, with starter sheets between
the anodes. Electric current, at 21,000 amps per cell and DC volts at
0.23 per cell, provide a current density of 24 amps per square foot.
After 28 days, the 820-Ib anode is 130 pounds and recycled to the con-
verter. Cathodes are 345 pounds.
The electrolyte circulates through cells at a rate of 4 to 5 grams
per minute. It contains 48 g/fc copper and 185 g/& sulfuric acid, and
the temperature is controlled at 150°F. Copper in solution is removed
in an electrolytic purification (liberator) section which has an insol-
uble lead anode and starter cathodes. Copper ions leave solution and
desposit on a low-qual(ity liberator cathode which is recycled to the
.smelter. Impurities, which settle out into the slime, are collected,
acid-leached, filtered, and dried. This material is then shipped to a
precious metal refinery for recovery of gold, silver, and selenium. The
slime is washed out of the refining cells every 28 days. Some cathodes
are melted and cast into 5/16" and 13/32" diameter rods.
Emission Controls
To date (1 July 1977), Magma is the only company to provide any
information after being contacted. Figure 6 gives a description of
their emissions by a comprehensive plant operation program and ambient
monitoring system. Converter operations are synchronized to provide a
constant, uniform offgas to the acid plant for greater SOa efficiency.
In addition, their eight ambient air quality monitoring stations' data
are analyzed with meteorology reports. Consequently, guidance informa-
tion about possible adverse weather conditions resulting in air pollution
episodes can be anticipated and plant operations may be curtailed.
B-13
-------
INSPIRATION CONSOLIDATED COPPER COMPANY; HISTORY1»°»9
1916 — First production of copper from the underground mine and flota-
> ition process.
1926 — Began production of ferric-sulfate leaching of mixed oxide-
sulfide ore and of ectrolytic copper.
1933 — By May, all mining and smelting stopped due to the depression.
1935 — Production began again, in September.
1948 — Production from open-pit mining began and the leaching plant was
enlarged.
1954 — Underground mining was'stopped and completely replaced by pit
mining.
1958 — Molybdenum recovery began.
1960— The International Smelting and Refining Smelter was purchased
and a new refinery built.
1962 — The "Christmas" underground mine began to produce.
1966 — Treated ore production increased' from 16,500 to 20,000 t/d.
1968 — Capacity increased to 25,000 t/d of ore and the Christmas mine
was converted to an open pit, and the "Ox-Hide" mine was devel-
oped. Copper fabrication and rod plant.
1972 — Pollution control construction began—electric copper furnace
and acid plant.
ANACONDA PROCESS DESCRIPTION11
The Anaconda Company smelting complex at Anaconda, Montana, consists
of two plants—the pyrometallurgical smelter and the hydrometallurgical
smelter. The two plants can produce 35 million and 6 million pounds of
copper per month; however, legal agreements limit the combined output to
35 million pounds per month.
Major processes at this complex consist of four reverberatory fur-
naces, the electric furnace, six converters, two refining furnaces, a
baghouse, an electrostatic precipitator, a sulfuric acid plant, two lime
kilns, a boiler house, the Arbiter plant, concentrate drying and blend-
ing area, tailings ponds, and slag pile.
The Anaconda Company currently has variances for particulate emis-
sions from the main stack and for the 90% S02 control rule which applied
B-14
-------
to the smelter. Compliance with both of these rules is related to com-
pletion of new smelting facilities followed by expansion of the sulfuric
acid plant. Compliance with the particulate emission rule will be
achieved for the main stack and matte tapping hoods when the reverbera-
tory furnaces are shut down (probably in 1975). Compliance with an
interim 75% control rule will be achieved in early 1977, followed by a
6-month review of the applicable standards.
HISTORY OF THE ANACONDA SMELTING OPERATION11
1883 — Smelter construction started. A concentrator building was com-
pleted.
1883 — Upper works built.
1884 — Smelter construction was completed. It has a capacity of 10
million pounds per month. The refinery had a capacity of 30
million pounds per month and another of equal capacity was
started. Custom smelting was done from the start.
1887 —- The Great Northern Railroad was extended to Butte so that ores
and copper could be shipped to Great Falls.
1887-1888 — Lower works built when the upper works were found to be too
small.
1889 — The lower works burned in March. They were rebuilt before winter
with a larger capacity. In 1892 the smelter output was 100 mil-
lion pounds of the metal.
1894 —On January 1, the Butte, Anaconda and Pacific Railroad went into
operation.
1896 — Sulfuric acid manufacture from the flue gases started. Leaching
of tailings was carried on successfully. Electrolytic refining
of copper started.
1896 — The company sold 107 million pounds of copper, 4.5 million ounces
of silver, and 14,400 ounces of gold.
1898 — Half of the company's yearly output of 124 million pounds (10.03
x 106 Ib/mo) of copper was electrolytically refined.
1899 — Amalgamated bought Anaconda Copper Mining Co.
1902 — The Washoe Smelter was first used on January 25. Included in
the facilities was the original tall stack and settling chamber.
1903 — The Anaconda Copper Mining Co. paid $330,000 in damages to local
ranchers for animal loss.
B-15
-------
1903 — A 300-foot stack and the existing settling chamber were installed.
1910 — Mining options were first obtained in Chile by an American firm.
1914 — Electrolytic refining of zinc was developed.
1918 — The present main stack and electrostatic precipitators were in-
stalled to remove solid and liquid matter from the flue gas.
1919 — The main stack was first used.
1919 — The company constructed a plant at the reduction works for pro-
duction of high-quality fertilizer.
1923 — The Anaconda Co. bought the Chile Copper Co. which had started
operations at Chuquicamata, Chile, in 1915.
1931 — Use of natural gas at the smelter was started. All coal burning
apparatus was replaced with gas burners.
1941 — The manganese concentrator and nodulizing plants were completed.
This was at the request of the U.S. Government to alleviate a
shortage created by World War II.
1955 — The Anaconda Reduction Department produced 252.3 million pounds
of copper, 153.6 million pounds of zinc, 145,300 tons of sulfuric
acid, 89,500 tons of fertilizer, 73,800 tons of manganes nodules,
and large quantities of arsenic.
1958 — An ammonium phosphate plant was put into operation.
1959 — The zinc crusher, zinc concentrator, and manganese flotation
plant were closed.
1961 — The phosphate plant was .shut down.
1964 — The arsenic plant was shut down.
1966 — A fifth converter was added.
1967 — The Montana Legislature enacted the "Clean Air Act of Montana."
1969 — The zinc leaching and purification plants were closed.
1969 — Ambient air standards governing sulfur oxides at smelters, regu-
lation 90-015, was passed by the Montana State Board of Health
in November.
1970 — On April 30, 1970, regulation 90-008 was adopted. It became
effective on June 30, 1973.
B-16
-------
1971 — A Johnson-Marsh Chem-Jet Dust Suppression system was installed
at the East Anaconda Crusher.
197i — The new flue system removing off-gases from the reverberatory
furnaces and converters was completed.
1971 — Anaconda's Chilean holdings were nationalized.
1971 — On November 17, the Anaconda Co. was issued a construction permit
for "roasters, smelting furnaces, and converters" to be located
at Anaconda, Montana.
1971 — On December 15, the Anaconda Company's request for a change.in
the State regulations concerning sulfur oxide emissions from
primary non-ferrous smelters was denied.
1972 — In January, Governor Anderson submitted the "Implementation Plan
for Control of Air Pollution in Montana" to the Environmental
Protection Agency without the section on S02 control.
i '
1972 — The zinc refining operations at Great Falls ended.
i . '
l 1972 — On July 27, EPA proposed a regulation to control the emission ,
j of sulfur oxides at the Anaconda Company smelter and to hold
public hearings on the proposed regulation 30 days later.
j 1972 — On July 28, a construction permit for "water-cooled converter
I . hoods, converter uptakes and flues, gas scrubber, fiberglass
flue to acid plant, 660 TPD acid plant with storage and loadout
I facilities" was approved.
1972 — On August 30, EPA held a public hearing in Helena pursuant to
proposed regulations for control of sulfur oxide emissions from
smelters.
1972 — On September 26, the Anaconda Company filed a civil suit against
William D. Ruckelshaus (EPA Administrator) and others, asking
., that EPA be required to hold an adjudicative hearing and prepare
an Environmental Impact Statement prior to promulgating rules
for control of sulfur oxide emissions from smelters.
1972 — On December 19, the federal district court in Denver rendered
its decision in favor of the Anaconda Company in its suit against
EPA (see 1972, September 26).
1973 — On February 8, Notices of Violation and Order To Take. Corrective
Action were issued by the Air Quality Bureau for excessive vis-
ible emissions from the two lime kiln stacks and the foundry
cupola furnace. Both lime kiln emissions were corrected by June
1. Approval of control equipment for the foundry has been ap-
proved. Installation is expected shortly.
B-17
-------
1973 — On February 12, the Anaconda Company requested a variance from
regulation 16-2.14(1)-S1470(2), formerly 90-008.
1973 — Use of the new water-cooled hoods and flue system was started.
1973 — The seventh and eighth converters were added.
'1973 — Operation of the 660 t/d sulfuric acid plant was started in June.
•!• ' '• .
1973 —"' In June, the U.S. Supreme Court upheld rulings by lower courts
in a lawsuit filed by the,Sierra Club against EPA that EPA must
disapprove all State implementation plans which did not contain
''>"' provisions to prevent degradation of existing air quality in
areas that had cleaner air than allowed by the national standards.
1973 — In July, the Anaconda Company requested and was granted a vari-
ance from regulation 90-004 governing particulate emissions from
the smelter. This variance expires in July 1974.
1973 — In July, the Anaconda Company started construction of a baghouse
"to remove particulate matter from the main flue gases. Expected
completion date is August 1, 1974. This baghouse will not treat
all gases going to the main stack until the new electric furnace
replaces all but one of the reverberatory furnaces and the flue-
solids roaster is operational.
1973 — In July, the U.S. Court of Appeals, Tenth Circuit, ruled that
the federal district court was not the proper one to hear the
case of the Anaconda Company vs. William Ruckelshaus et al. (see
September 26 and December 19, 1972), that the impact statement
was not necessary and that EPA had acted in a proper manner in
establishing rules for sulfur oxide controls.
1973 — On July 26, the Anaconda Company initiated action to classify
certain new equipment at their plant as air pollution control
equipment and, as such, be eligible for a class seven taxation
rate.
1973 — In September, EPA held one of four scheduled public hearings (in
Denver) to try to determine the best of four possible approaches
to achieving "no significant deterioration" of air quality.
B-18
-------
ADDENDA
• Map of the Kennecott copper works at Magna
• Production figures for Magna (1940-1971)
• List of all primary copper smelters in the
United States
• Annual capacity of four Arizona copper
smelters (1927-1973)
B-19
-------
KENNECQTT COPPER COMPANY, MAGNA, UTAH
INTERNATIONAL
AIRPORT
NORTH TEMPUE ,
OUT AH SMELTER
O UTAH REFINERY
ARTHUR
CONCENTRATOR
MAGNA
CONCENTRATOR
\ BONNEVIULE
CONCENTRATOR
Kennecott Copper Corporation
UTAH COPPER DIVISION
and Visitor Guide
O BINGHAM CANYON MINE
MAINTENANCE SHOT
O PRECIPITATION
PLANT
The copper production cycle starts at the Bingham
Canyon/ Mine Q, where,.106,500 tons of ore and over
300.00Q t°ns of overburden, or waste material, are
removed daily. Because overburden contains trace
amounts of copper, it is treated with a solution which
is subsequently processed at the precipitation plant
O. Ore hauled by train from the mine Is processed
through three concentrator plants Q where it is
crushed, ground and concentrated to nearly 30 percent
copper. The smelter O takes the concentrate and
produces blister copper, 98.6 percent pure; then
copper anodes 99.5 percent pure. At the electrolytic
refinery Q the copper is further purified to 99.96
percent and cast into various shapes for Kennecott
customers.
Furnished by:
Communications Department
Utah Copper Division
Kennecott Copper Corporation
P. 0. Box 11299
Salt Lake City, Utah 84111
B-20
-------
ANNUAL PRODUCTION (103 tons) OF COPPER BY THE SMELTER
AT MAGNA, UTAH, 1940-1971
Year
1971
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
Copper Production
260
296
297
228
169
265
259
200
203
218
213
219
145
189
238
251
Year
1955
1954
1953
1952
1951
1950
1949
1948
1947
1946
1945
1944
1943
1942
1941
1940
Copper Production
233
2.12
269
283
271
279
197
] 227
267
114
226
302
326
318
271
249
Source: Health Consequences of Sulfur Oxides: A Report
from CHESS. 1970-1971. U.S. EPA, Office of R&D,
National Environmental Research Center, Research
Triangle Park, N.C.
B-21
-------
PRIMARY COPPER SMELTERS IN THE UNITED STATES
American Smelting and Refining Company
El Paso, Texas
Hayden, Arizona
Tacoma, Washington
The Anaconda Company
Anaconda, Montana
Cities Service Corporation
Copperhill, Tennessee
Inspiration Consolidated Copper Company
Miami. Arizona
Kennecott Copper Corporation
Garfield, Utah
Hayden, Arizona
Hurley, New Mexico
McGill, Nevada
Magma Copper Company
San Manuel, Arizona
Phelps-Dodge Corporation
Ajo, Arizona
Douglas, Arizona
Morenci, Arizona
White Pine Copper Company
White Pine, Michigan
B-22
-------
Tovt »f
Off.
Jtt.tt*
140401
l*elo»t
C»«, tut
CO
to
U)
S»o,t>oo
100,000
L*0.<»0
If 9,000
COPPER SMELTING — ANNUAL CAPACITY
^, A
^^_« »__« •—• • —-JM
/
-* • * • • • • »—*
A A
ifi3
H73
SOURCE: American Bureau of Health Statistics, Yearbook.
NOTE: Magna Co. Smelter at San Manual started production in 1962, A-A-A; Kennecott started
in 1958, o-oo . All oE the data are based on the capacities of the smelter reverbatories,
except for the data for the Old Dominion Smelter, which had no reverbatory. Data for the
Old Dominion is based on blast furnace capacity.
The capacity of copper smelting works is given as estimated by the respective proprietors,
stated in tonnage of capacity for smelting materials that yield a product. Capacity in
terms of copper product varies with the grade of the ore charged. Ore and concentrates
are metallurgically synonymous terms.
-------
REFERENCES
> • i . I
1. Mining, (October 1973), Arizona Highways Vol. XLIX, No. 10.
2. Copper Products and Production, Encyclopedia Britannica.
3. Development Document for Interim Final Effluent Limitations Guidelines
and Proposed Nev Source Performance Standards for the Primary Copper
Smelting Subcategory and the Primary Copper Refining Subcategory of
the Copper Segment of the Nonferrous Metals Manufacturing, (1975),
U.S. Environmental Protection Agency.
4. Roger's Industrial Chemistry, C. C. Furnas, (1946) D. Van Nostrand
Co. Inc., New York.
5. Magma Copper Company correspondence.
6. Compilation of Air Pollutant Emission Factors, (1976)., U.S. Environmental
Protection' Agency.
7. Background Information for New Source Performance Standards: Primary
Copper, Tins, and Lead Smelters, (1974), U.S. Environmental Protection
Agency.
8. Inspiration Consolidated Copper Company conversation.
9. Inspiration Consolidated Copper Company correspondence.
10. Observation.
11. Montana Air Pollution Bureau.
B-24
-------
Appendix Cl
ETIOLOGY OF DISEASES ASSOCIATED WITH COPPER SMELTING
Appendix C2
SELECTED ANNOTATED BIBLIOGRAPHY OF LITERATURE
RELATING DIGESTIVE DISEASES AND COPPER SMELTING
-------
Appendix Cl
ETIOLOGY OF DISEASES ASSOCIATED WITH COPPER SMELTING*
COPPER SMELTING AS A SOURCE OF TOXIC POLLUTANTS
Studies of many mining operations or heavy industry have identified
diseases that are directly correlated with the primary product itself,
i.e., coal mining and coal miners' pneumoconiosis or fluorescent lights
and beryllium pneumonitis. Copper smelting plants, on the other hand,
emit a large number of trace and heavy metals, many of which are neces-
sary in living organisms. Their mere presence in slightly elevated
levels with certain diseases does not necessarily make a case for disease
causation. One of the difficulties with copper smelters is the extreme
variety of materials given off, much of which in gross excess may be as-
sociated with disease. Almost no long-term, low-dose animal experiments
have been done to anticipate the types of disease to expect. Most as-
sumptions of the environment's impact are extrapolated from high-dose,
long-/short-term exposures; these data derive from workers in mines and
plants.
IMPORTANT VARIABLES IN MONITORING COPPER SMELTER DISEASES
Unlike coal mines, copper smelters are unique for several reasons:
1) Some of the trace metals given off are necessary to human and
animal biologic functions; therefore, slight aberrations may not be de-
tectable for years or, in fact, generations. Since effects are often de-
layed and may manifest themselves after exposure has ceased, studies must
be conducted on stable populations (1,26).
2) The scope of investigation will of necessity need to be much
larger if a seripus evaluation of a copper smelters' effect on the envi-
ronment is to be appreciated. Because of materials released, more subtle
diseases may become manifest (vide infra) and need to be monitored (17).
3) Not only should diseases of excess be considered, but also those
of deficiencies. With multiple trace metals in competition for binding
sites in the gastrointestinal system, absorption of necessary elements
may be eliminated leading to decreasing levels of these elements. Pollu-
tants from smelters may be acting like fertilizers that have already mod-
ified the mineral content of the soil and its products, creating a man-
ganese-deficiency in grazing animals and a copper-deficiency in grains
and plants (27).
4) If a "specific industry-disease association" is to be made, it
will be necessary to monitor more extensively the food sources. Arsenic,
which is produced by smelters, is concentrated by shellfish but is also
*
This appendix was prepared by Tee L. Guidotti, M.D., in collaboration
with Philip Smith, M.D.
-------
used in insecticides extensively and can be foi'md in apple cider and wine.
Therefore, the real source of arsenic contamination may be missed if only
samples of water and air are monitored. Furthermore, a report from Hop-
kins soon to be released will report that one of the largest sources of
coppe'r in water is copper pipes in common households.
5) All previous studies involving copper smelters, including the
International Journal of Cancer report on 839 copper smelters, pay only
lip service to smoking but do not control for this variable in evaluating
lung cancer. In this particular report 10 out of 11 subjects smoked,
which is an inordinately high percentage and would in itself be associ-
ated with increased incidence of lung cancer. Of further importance is
the known exacerbation of smoking and the inhalation of toxic fumes—the
cigarette acting as the catalyst (1).
6) Arsenic and possibly other pollutants in excess can lead to pa-
resthesia, muscle weakness, and other types of neurologic deficits. Al-
though many neurologic signs and symptoms are subtle and tend to be over-
looked, they often serve as an early marker for toxic effects (16). Man-
ganese and lead toxicity can produce acute self limited psychiatric dis-
turbances. In-the case of manganese intoxication, the syndrome very
closely reseiiibles schizophrenia; on the other hand, manganese deficiency
may'produce a Parkinson-like state. It has been postulated that manga-
nese and possibly other trace metals may modulate the dopaminergic ap-
paratus of the brain (17).
7) As has been mentioned, when evaluating the respiratory system it
cannot be overemphasized that the impact of smoking (vide supra) must be
carefully controlled. When looking at data from a smelter community, not
only must the type of smoking be recorded (cigar/pipe/cigarette/marijuana)
but so too the duration and sequence of smoking. Sequence refers to
whether cigarette smoking occurred before the institution of cigar smok-
ing. This is important because individuals who smoke cigarettes first
tend to inhale pipe smoke and, therefore, may have as much lung disease
as cigarette smokers. Many medical record departments do not consider
pipe smokers.
8) Because manganese, copper, zinc, and selenium are needed for en-
zyme systems and hormonal balance, diseases of the endocrinologic system
must be considered. Deficiencies in manganese may cause sterility in
males. Copper is used in intrauterine devices to prevent conception, and
in rats copper deficiency causes infertility. There is also some modula-
tion of the adrenal coricosteroids by copper and zinc, although no direct
link has been made to specific diseases such as Gushing's or Addison's
diseases (17).
9) The teratologic potential of trace metals in man has not yet been
investigated. Manganese-deficient chicks have offspring with chonodro-
dystrophy; copper-deficient lambs have offspring with enzootic ataxia
(neurological dyscoordination). Zinc-deficient rats have multiple con-
genital anomalies and malformations involving the face and limbs. There-
Cl-2
-------
fore, when surveying the smelter community, careful obstetrical histories
must be sought. This should include any spontaneous abortions and pathol-
ogy of aborted fetuses.
10); Acute infections are accompanied by well-documented changes in
serum concentrations of iron, zinc, and copper. However, serum concentra-
tions during illnesses do not necessarily reflect body stores; whether or
not these elements are necessary for host defense is unproven.
11) Dental abnormalities are often signs of high-dose exposure to
sulfuric acid and other acids in aerosolized form. These changes are ini-
tially reversible, but with chronic high-dose exposure lead to permanent
damage (14).
TOXICOLOGY OF DISEASES OF COPPER SMELTER WORKERS
There are no specific diseases per se of copper smelter workers. As
has been mentioned, the variety of pollutants is extensive although the
amounts are generally small. The following represent the most likely dis-
ease associations that one might anticipate, i.e., pollutants that are
both present in higher concentrations and have the best scientific evi-
dence of disease association.
Cadmium
There is good evidence that cadmium is correlated with hypertension.
Whether cadmium causes hypertension in humans is controversial, but animal
studies do demonstrate an increased incidence of hypertension with long-
term, low-dose exposure. In a recent report on 27 copper smelter workers,
an increased incidence
-------
Sulfuric Acid/Sulfur Dioxide
Sulfuric acid and sulfur dioxide are both toxic respiratory pollu-
tants, the former perhaps more so than the latter. In one of the best
long-term low-dose experiments in dogs, SOa and other auto emission prod-
ucts were studied (22). SO2 alone caused not only pathologic changes in
the lung (emphysema) but also early changes in pulmonary function. Amdur
has reported the respiratory effects in guinea pigs of both sulfur dioxide
and sulfuric acid in mist form. In combination, the pollutants decreased
growth and damaged the lung more than either of the two did alone (8).
Lead
The toxic effects of lead have been well documented (13), and low-
dose exposure is most likely to result in the following: 1) peripheral
neuropathies such as paresthesia and hyperesthesia, also possible are
subtle mental disorders such as depression; 2) nephropathy—manifested as
an .Interstitial nephritis; 3) anemia; and 4) gastrointestinal symptoma-
tology—abdominal colic, constipation, and weight loss. Even though toxic
symptoms are not usually present until blood levels reach 80 yg/100 m£,
occupational exposure that results in absorption of 15 g daily can achieve
a positive lead balance. Therefore, lead levels can be misleading and at
present are based on high-level industrial exposure (16). Cooper reported
in 1974 on 7,032 lead smelter and battery plant workers, who were, in some
cases exposed to high doses of lead, documenting no differences in stan-
dardized mortality ratio (SMR) for all causes or life expectancy (24).
Nickel
High exposure to nickel has three relatively distinct hazards associ-
ated with it: 1) contact dermatitis that develops when exposed to solu-
tions of nickel; 2) hemorrhagic pneumonia when nickel carbonyl is inhaled;
and 3) cancer of the upper respiratory tract (nasal sinuses) and lung if
nickel dust is inhaled. Since arsenic is often mixed in, it is unclear
which is more at fault. As with other trace metals, there are no long-
term low-dose experiments with nickel. There is evidence that nickel in
the water (NiSOi,) is more readily absorbed than is nickel bound to food.
SUMMARY
Copper smelters release toxic agents but also trace elements essen-
tial for biologic function. To evaluate the impact of a smelter on com-
munity health, the following should be considered:
• Monitoring of as many food sources as possible, to avoid
missing other significant sources of pollutants.
• In-depth evaluation of human systems beyond respiratory,
cardiac, and gastrointestinal systems to include endocrino-
logic, neurologic, and renal systems.
Cl-4
-------
• Teratology and abortion histories of females.
• As detailed a smoking history as possible.
The following represent a list of the diseases most suspect for com-
munity 'tisk:'
• respiratory—laryngitis, asthma, bronchitis, emphysema,
fibrosis, pneumonia
• renal—hypertension, proteinuria, stones
• neurologic—paresthesia, hyperesthesia, neuralgia,
paralysis, mental disorders (depression)
• gastrointestinal—granulomas in the liver
• oncologic—cancer of lung and upper respiratory tract
(nasopharynx), skin, gastrointestinal tract; leukemia
• endocrinologic—infertility
• dental etching of enamel, caries
• skin hyperpigmentation, keratosis, dermatitis
• reproductive—birth defects, abortions
REFERENCES
1. Tokudome, S., et al., A cohort study on mortality from cancer and
other causes among workers at a metal refinery, Int. J. Cancer 17:310-
317, 1976.
2. Pinto, S., W. Tacoma et al., Effect of arsenic troxide exposure on
mortality, Arch. Environ. Health 7(5):583-591, 1963.
3. Milham, S., and T. Strong, Human arsenic exposure in relation to a
copper smelter, Environ. Res. 7:176-182, 1974.
4. Greenwaldj Effects of inhalations of low concentrations of sulfur di-
oxide upon men and other mammals, Arch. Ind. Hyg. Occup. Ned. 10(6):
455, 1954.
5. Amdur, M., R. Schutz, and P. Drinker. Toxicity of sulfuric acid mist
to guinea pigs, Arch. Ind. Hyg. Occup. Med. 5:318, 1952.
6. Anfield, B., and C. G. Warner, A study of industrial mists containing
sulfuric acid, Ann. Occup. Hyg. 2:185-194.
7. Lewis, T., et al., Toxicity of long-term exposure to oxides of sulfur,
Arch. Environ. Health 26:16-21, 1973.
8. Amdur,.M., Aerosols formed by oxidation of sulfur dioxide, Arch.
Environ.. Health 23:459-468, 1971.
9. Amdur, M., L. Silverman et al., Inhalation of sulfuric acid mist by
human subjects, Arch. Ind. Hyg. Occup. Med. 6:305-313, 1952.
Cl-5
-------
10. Goldman, A., and W. Hill, Chronic broncopulmonary disease due to sul-
furic acid fumes, Arch. Ind. Hyg. Occup. Ned. 8(3):205-211, 1953.
11. Villar, T., Vinyard sprayers lung, Am. Rev. Respir. Dis. 110:545-555,
1974.
12. Smith, T., et al., Chemistry of sulfur and arsenic in airborne copper
smelter particulates, Bull. Environ. Contain. Toxicol. 15(6):651-659,
1976.
13. Occupational Exposure to Inorganic Lead, NIOSH HSM 99-73-11010, 1972.
14. Occupational Exposure to Sulfuric Acid, NIOSH HSM 99-72-116, 1974.
15. Occupational Exposure to Inorganic Arsenic, NIOSH HSM 99-72-127, 1973.
16. Industrial Health Foundation, Proceedings of the Symposium on Toxicol-
ogy of Metals, June 5-6, 1974.
17. Symposium on trace elements, Wed. Clin. North Am. 60(4):831, July
1976.
18. Chowhury, P., and D. Louria, Influence of cadmium and other trace
metals on human a\ antitrypsin, Science 191:480-481, 1975.
19. Harvey et al;., Measurement of liver cadmium concentrations in pa-
tients and industrial workers by neutron-activated analysis, Lancet,
p. 1269, June 7, 1975.
20. Scott et al., Clinical and biochemical abnormalities in coppersmiths
exposed to cadmium, Lancet, p. 394, August 21, 1976.
21. National Academy of Sciences, Copper - Medical and biologic effects
(ISBN 0-309-02536-2), 1977.
22. Hyde, D. M., W. S. Tyler, D. Duneworth, and Orthoefer, Pulmonary
lesions in beagle dogs chronically exposed to air pollutants, Am. Rev.
Respir. Dis. 115(4):22, 1974.
23. National Research Council, Lead: Airborne Lead in Perspective, Na-
tional Academy of Sciences, Washington, D.C., 1972.
24. Cooper, C., Mortality of workers in lead smelters and lead battery
plants, paper #59, International Symposium on Recent Advances in the
Assessment of Health Effects on Environmental Pollution, Paris, June
24-28, 1974.
25. Newman et al., Histologic types of bronchogenic carcinoma of copper-
mining and smelting community, Ann. N.Y. Acad. Sci. 271:260, 1976.
26. Blot, W. Y., Arsenical air pollution and lung cancer, Lancet 2(7926):
142-144, 1975.
27. Volsin, A., Why two hundred years of scientific research on cancer
have been a disillusionment, Rev. Pathol. Gen. 63(749):631-672, 1963.
28. Schwartz, D., and R. Flamant, Statistical studies of the etiology of
bronchogenic cancer, Rev. Tuberc. Pneumol. (Paris) 26(2):5-15, 1963.
Cl-6
-------
Appendix C2
SELECTED ANNOTATED BIBLIOGRAPHY OF LITERATURE
RELATING DIGESTIVE DISEASES AND COPPER SMELTING
Copper and Brass Research Association. Copper and health.
6483:11-20, 1948.
After an extensive historical review of the occupational hazards
of working with copper, the review concludes, "Today, so far as
we know, copper is negligible as an industrial poison and the
ill-health that is found among copper miners or workers with the
metal is caused by something else connected with their work—
heat or overexertion, or irritating dust, or the presence of
lead or arsenic mixed with the copper." [p. 20]
McCaull, Julian. Building a shorter life. Environment 23(7):2-15,
38-41, 1971.
Cadmium'pollution of the environment is reviewed with respect
to basic characteristics, emission sources, uses, concentra-
tion levels, and effects on human health. Cadmium dust, fumes,
and mist are emitted during the refining of zinc, copper, and
lead, as well as during extraction of cadmium. Cadmium is
correlated with hypertension, liver damage, bone disease,
emphysema in industrial workers, cancer, and kidney impair-
ment.
. - i
Fulkerson, S. Cadmium—the dissipated element—revisited, Environmental
Aspects of.Trace Contaminants Program and Atomic Energy Commission
Interagency Agreement AG 389. Oak Ridge National Laboratory, Oak Ridge,
Tennessee, Energy Division of National Science Foundation Research
Applied to National Needs, 1975, 28 p.
i
Cadmium air pollution is a serious concern near nonferrous
metal smelters, especially zinc and lead smelters. Emissions
from these operations can lead to soil and water contamination.
Emissions from steel mills, incinerators, or coal burning
plants are probably insignificant sources of air pollution,
or can be reduced to insignificance by proper abatement prac-
tices.' Estimates of the rate of accumulation of cadmium in
i the kidney and of the critical concentration (200 ppm) in the
renal cortex necessary to cause renal tubular dysfunction have
been used to calculate the maximum permissible chronic expo-
j sure by various environmental pathways. Cigarettes are a more
i important source of inhaled cadmium to the general public than
are most industrial emissions. Chronic cadmium exposure has
caused hypertension in laboratory animals.
C2-1
-------
Sterekhova,. N. P., A. A. Maratkanova, and I. K. Brodskaya. Experimental
data on the combined activity of sulfur dioxide and dust at a copper
smelting plant. Gig. Tr. Prof. Zabol. 16(3):52-54, 1972.
Respiratory and digestive disorders are frequent occurrences in
the-copper smelting industry. The effects of sulfur dioxide
and dust from industrial ash containing silicon dioxide, copper,
iron, sulfur, zinc, calcium oxide, and lead were studied in
rats. After four hours of daily exposure to sulfur dioxide in
concentrations of 143.0 mg/m for four months, a toxic effect
manifested by dystrophic alterations in almost all systems and
organs, particularly the liver, was noted. After exposure to
the dust for the same period, sclerotic alterations developed
in the bronchi and lungs. After exposure to combined activity,
sclerotic processes in the lungs were more pronounced than in
either of the former tests.
Ponteva, M., and L. Hansson. Lead and common gastrointestinal symptoms.
Duodecim 90(5):335-341, 1974.
The blood concentrations of lead, zinc, and copper in Finnish
military recruits with gastrointestinal symptoms, and of con-
trols who did not have such symptoms, were determined. The
lead content in the blood of the group with the gastrointes-
tinal symptoms was significantly different from the symptom-
free group. No significant difference existed with respect
to zinc and copper levels.
Babushkina, L. G., N. P. Sterekhova, and F. S. Kuzmina. Some indices
of lipid metabolism in the blood of workers engaged in copper works
[Russian].' Ter. Arkh. 43 (9):100-104, 1971.
Indices of lipid metabolism were determined in the blood of
94 copper smelter workers with toxicochemical hepatitis
pneumosclerosis (in the presence of absence of hepatitis
developing due to a chronic action of sulfur dioxide), and
in the blood of silicosis patients and donors. All patients
showed a significant increase of total lipids content com-
pared to donors.
Colten, S. R. A review of the health hazards from copper exposure.
J. Occup. Med. 26:621-624, 1975.
Exposure to copper dusts and fumes can result in respira-
tory and/or dermatological effects. The prompt emetic
effect of copper usually limits its toxicity when taken
orally. However, when ingested in elemental or salt form,
gastrointestinal and liver effects are observed (the
C2-2
-------
author suggests that high enough dust concentrations would
probably cause similar effects). Some adverse reactions are
diarrhea, nausea, vomiting, and abnormal liver functioning.
Biopsy of hepatic tissues from one group of copper-salts-
poisoned patients showed tissue mucous which was attributed
. to acute copper toxicity.
Hale, Ellen. Use of 'hazardous1 herb widespread. Tucson Daily Citizen,
Tuscon, Arizona, Wednesday, May 4, 1977.
Use of a toxic herb mistaken for a medicinal tea plant may be
very widespread in Arizona and could account for some of the
state's high incidence of liver disease. Two Tuscon special-
ists, Dr. Alfred E. Stillman and pharmacologist Ryan J.
Huxtable have found several close associations between adults
who have drunk the herb in a tea and who now suffer from
liver disease. Arizona has long had unexplainable levels
of cirrhosis, about 10 percent higher than the national
average, according to Huxtable.
COMMENTS:
Copper smelter activity has been associated with adverse health'
effects, both among workers and in the community. Copper itself does
not appear to be the toxic agent, but rather trace contaminants and
gaseous wastes (in a few reported cases, ingestion of copper itself—
in the form of salts—has been associated with liver abnormalities).
By-products and effluents of the industry which have appeared in the
community and have been associated with disease are: arsenic (respira-
tory system cancer); cadmium (liver damage, cancer); and sulfur dioxide
(liver disease). Furthermore, high levels of lead have been found in
persons with gastrointestinal symptoms.
C2-3
-------
Appendix D
MAPS OF THE COPPER STUDY AREAS
• Arizona
• Ajo, Arizona
• Morenci, Arizona
• Douglas, Arizona
• San Manuel, Arizona
• Miami-Inspiration, Arizona
• Hayden, Arizona
• Montana
• Anaconda, Montana
• Magna-Salt Lake Area, Utah
• Magna, Utah
-------
FREOONIA
PAGE
COCONINO
MOHAVE
I KING MAN
> BULLHEAD
CITY
L« LAKEHAVASU
CITY
WILLIAMS
FLAGSTAFF
YAVAPAI SEDONA
CLARKDALE
•• COTTONWOOD
BAGDAD I
PRESCOTT • HUMBOLDT.-OEWEY
MAYER
r* PARKEfl
• PAYSON
WICKENBURG <
NAVAHO
APACHE
WINSLOW
• HOL8ROOK
SNOWFLAKE • ST. JOHNS •
SHOW LOW .
« LAKESIDE .
• PINETOP
SPRINGERVILLE
GILA
MARICOPA
/
YUMA
YUMA
BUCKEYE . pHQ
GILA BEND
•
/ • MIAMI * • UL-1-
ENIX I \4 .
J SUPERIOR/
4FLORENCEKE/N
.COOLIDGE/.^
. CASA GRANDE
' ELOY . M
PINAL ORACLE. *s
J
•*
jf AJ° . TUCSON
*-^^^ PIMA BE
'°^^ GREENL
^ GRAHAM
CLIFTON-
Y MORENCI*
EN \j/f
(LEMAN « PIMA 0o^CAN
THATCHER * • ^ •
AMMOTH SAFrCPD
AN MANUEL
• wiucox
NSON
GREEN VALLEY •
/ SHELTER TOWN
SANTA CRUZ
ATAGONIA.
. NOGALES
COCHISE
Slt'iRA VISTA
•
8IS3 = = «
ON;
.DOUGLAS
MAP OF ARIZONA
AS OF FEBRUARY 1. :S72
10 0 10 23 30
SCALE-UILES
D-l
-------
O
ro
. ;••••*- : : •-,-.,>.,v .-.,,
!••'•'' ,.«W ,'| f1 '.} f*-'?l'.*^[ ;"•// // .1 «»•.**, !
>*•* -» — -"• XV ^i t—'..^•. *^i . it'Vi ••^l /^ 7^ ** *
,p^^Qfe-^;. ^/K^i
x t^. ">~; "•••^Lr" •,<;«w>^.'">" v.-w&zrffi
^ -^^r.^fe-ol
r*^>^ -:'^tj ^ ^ r^ \Ax^
'!
AJOS ARIZONA
SCALE 1.62500
• MILES
CONTOUR INTERVAL 40 FEET
rrtO LINES RtPRESENT 20 FOOT CONTOURS
DATUM IS MEAN SEA LEVEL
-------
o
U)
V '^^*'-v^'N^K**g*Sr382a
^•* '$ ^ls»l?5iif?P
«w3s
;-• •Oj*..^^.'.^
«Miil
Wa--W''W(Sw'tSf-->N •' ?'
u
* '1- ••'' "
MORENCI, ARIZONA
SCALE 1:62500
CONTOUR INTERVAL 80 FEET
DATUM IS MEAN SEA LEVEL
-------
X
x O
/Douglas Municipal
• Pumping Plant
1000
1000 MOO 3000 «000 5000 6000 7000 FEET
1 KILOMETER
CONTOUR INTERVAL 20 FEET
D*JV.MJS_M£AN SEA LEVEL
DOUGLAS, ARIZONA
D-4
-------
•• s.
"V .
SAN MANUEL, ARIZONA
SCALE 1.P4000
o
...... 1
1000
o 1000_ ??°9_ -'to0 40K ;a'° 605° I000 FEET
! 5 0 1 KILOMSTER
D-5
-------
:• ^^J^^^l^^^^^^^^^^^
^SSg^^^S^^y/»WM^.^^^'<-i^^
MIAMI-INSPIRATION ARIZONA
"'v/?j>, - S'-^-J ,.-£&:.{?**:>***^f.^JjP^ //
1 MILE
600C 7000 FEET
Zl
Contour interval 25 feet
-------
Ll££%y.^
HAYDEN, ARIZONA
SCALE 1:24000
o
TUCSON 71 Ml.
(CENTRAL)
1 MILE
I'.W
1000
?000 3000 4000 5000 6000 7000 FEET
l KILOMETER
CONTOUR INTERVAL 40 FEET
DOTTED LINES REPRESENT 20-FOOT CONTOURS
DATUM IS MEAN SEA LEVEL.
-------
'WMI« Sutptiui I'" •' ""«>
Spring |. ,. «
f *
Cnu
MONTANA
D-8
-------
O
vo
arm c,.«i ' x . '•
ANACONDA
ANACONDA, MONTANA
SCALE 1:62500
C(JNroUf< INII.NVM. SO FLF1
OUTFID IINL*, ktfhfslNI 40 FOO' CONTOURS
UAIUM is MIAN bfA I ivLL
-------
s.
-' \
TRAVKRSC MOV.VJM/.VS
KIIES
0 i 10 IS' X R
o s 10 is n n jo»
KIIOKTERS
MAGNA-SALT LAKE AREA
D-10
-------
, * p.
O UTAH KFINEW r>ii
AMHUH
.CONCENT..,*,
Ktiuwntl CoppM Corpwitlm
UTAH COPPER DIVISION
MAGNA. UTAH
Th» copter production cyclt ittrtt it the Binghim
Ctnyon Mint O. where 106.600 toni of ore and over
300,000 ton* of overburden, or waste material, are
removed dally. Because overburden contains trace
amounts of copper, it is treated with a solution which
s subsequently processed at the precipitation plant
O. Ore hauled by train from the mine Is processed
through three concentrator plants O where it is
crushed, ground and concentrated to nearly 30 percent
copper. The smelter O takes the concentrate and
produces blister copper, 98.8 percent pure: then
copper anodes 99.6 percent pure. At the electrolytic
refinery Q the copper Is further purified to 99 96
percent and cast into various shapes for Kennecott
customers.
D-ll
-------
Appendix E
HISTORY OF AREA DEMOGRAPHY AND INDUSTRIAL ACTIVITY
IN GILA AND; FINAL COUNTIES OF ARIZONA
-------
Appendix E
HISTORY OF AREA DEMOGRAPHY AND INDUSTRIAL ACTIVITY
IN GILA AND FINAL COUNTIES OF ARIZONA*
GENERAL
Gila and Final Counties of Arizona constitute one of the study areas
for the investigation of the relationship between the copper smelting
industry and community health effects. There are presently four copper
smelters in the area.
Copper mining and the related production of silver, gold, zinc, and
molybdenum have been the basic forms of mineral resource utilization and
the major impetus for the growth and development of the communities within
the study area. Within Gila and Final Counties in general, growth patterns
have been greatly influenced and determined not only by the mining activi-
ties but also by the fact that a significant amount of land is unavailable
for residential development. A combination of vast expanses of government-
controlled land, the extensive Indian reservations, and the large private
land holdings controlled by the mining and ranching interests have helped
channel area growth. Today, as in years past, populations within the study
area are in close proximity to mining and smelter activities and are clus-
tered into discrete towns separated by stretches of uninhabited desert.
With the exception of Winkelman, all of the study area communities were
founded originally as camps associated with either silver or copper mining
activities. Today the strong influence of copper mining is evident through-
out the study area; i.e., the distribution and patterns of employment, the
characteristics of the labor force, the types of commercial service es-
tablishments, and the kinds of trade goods offered' for sale are reflective
of an area with one majpr economic base. All of the urban communities are
oriented towards servicing this mining base either directly or indirectly.
As a result of this total dependence upon mining, the size of the various
communities fluctuates in response to the economic status of the local
industry. This association was particularly evident in the early 1930"s
during the depression, when several mines were shut down and many miners
were forced to leave their homes or find jobs in the few area mines that
remained open. During that period most of the study area communities
decreased in population, as indicated in Table 1 which presents population
statistics for the study area for the past fifty years.
Extensive amounts of historical information concerning the communities
are not readily available. The primary sources used to obtain community
and industrial historical data were as follows:
Chamber of Commerce for each community
Arizona Office of Economic Planning and Development
U.S. Department of the Interior, Bureau of Mines
Library of Congress
U.S. Department of the Interior, Geological Survey
U.S. Department of Housing and Urban Development
"This appendix was prepared by Susan A. Perlin.
E-l
-------
Table 1. POPULATION CHARACTERISTICS OF STUDY AREA
W
to
SMELTER TOMNS .
Inspiration
Ha yd en
San Manuel
Superior
Miami
TOWNS WITHIN 10
MILES OP SMELTER
Claypool
Globe
Central Heights
Kearny
Hinkelman
Oracle
Mammoth
POPULATION
1976"
N.D.
1,300-1, 50O
f
5,400-5,700
3,650-3,900
N.D.
6,450-6,600
N.D.
2,800-3.000
1,100-1,300
/
/
1970
200
1,283
4,332
4,975
3,394
2,245
7,333
2,289
2,829
974
2,237
1,953
1960
400*
1,760
4,524
4,875
3,350
2,505
6,217
2,486
902
1,123
500*
1,913
1950
soo"7
1,494
(-)
4,500C
4,329
2,000C
6,419
N.D.
(-)
548
500C
500"
1940
600**
3.000
(-)
4,000"'
4,722
l.SCK/*
6,141
N.D.
(-)
524
SOO1*
650d
1930
500e
3,500"
(-)
5.500°
7,693
2,OOOe
7,157
N.D.
(-)
729
200s
ISO6
1920
18e
4,000e
(-)
3.000e
6,689
N.D.
7,704
N.D.
(-)
573
ISO*
4SOe
CHANGE IN POPULATION (A*)
I960- 1970
-27.1
-4.2
+2.1
+ 1.3
-10.4
+ 18.0
-7.9
+213.6
-13.3
+347 . 4
+ 2.1
195O-1960
194O-195O
-20.0
+17.8
+8.3
-22.6
+ 25.3
-3.1
—
+1O4.9
0.0
+282.6
-16.7
-50.2
—
+ 12.5
-8.3
+ 11.1
+4.S
—
+ 4.6
0.0
-:j.i
1930-1940
+ 20.0
-14.3
—
-27.3
-38.6-
-10.0
-14.2
-
-28.1
+150.0
+333.3
192D-193O
+2677.8
-12.5
—
+83.3
+ 15.0
—
-7.1
—
»27.2
+ 33.3
-C6.7
^Population estimates from the Arizona Office of Economic Planning and Development, Hid-2976 Population KxtJmates for Nnn-Met.ropolitan
Arizona Communities (no date).
fcArizona, The Grand Canyon State,- A State Guide, Revised by Joseph Miller. Hastings
Arizona Business and Professional Directory, Vol. 17, 1951-53, Baldwin and Mullin-
<*Busincss and Professional Directory of Arizona, 1941-43, Arizona Directory Company,
eArizona state Business Directory, 1920 and 1930 editions. Gazetteer Publishing and
1976 population estimate fov Mamnoth/Oracle/San Manuel: 9,700-9,900 people.
(-) = town not in existence.
N.D. ° no data found.
House (American Guide Series), Now York,
-Kille Company, Phoenix, 1951.
1941.
Printing Company, nrnvrr, 192O, 113O.
-------
• U.S. Securities and Exchange Commission
• Arizona State University (University Library), Tempo, Arizona
• State of Arizona, Department of Library, Archives and Public
Records
Table 2 highlights the chronology of important events concerning the
founding of communities and the operation of copper industries within the
study area. As indicated in this table, all communities (with the excep-
tion of Kearny and San Manuel) have had a very long history of exposure
to emissions and effluents associated with the various phases of copper
mining and processing.
Table 2. CHRONOLOGY OF EVENTS
1867 — Mammoth eatabllahed aa mining camp for area gold mines.
1870'a — Hlnkelman founded as ranchers nova Into area. Economy originally based on serving
the area agricultural sector.
1676 -- City of Globe established to serve the area silver mines.
1880 — Oracle established as a supply center for several mines in nearby Santii Catalina
Mountains.
ca. 1880 — Superior founded as mining camp near site of Silver Queen silver nine.
1881 — Old Dominion Copper s Smelting Company otarta operations at Globe and builds a "small
smelter" (2).
-- Small copper furnace starts operations at present site of Miami (10).
1883 — Some copper mining started at Ray deposits, near present site of Hayden (10).
1898 — Old Dominion Company builds new smelter at Globe, and for the next 12 years the Old
Dominion mine remains one of the world's "greatest copper mines" (2).
1900 — Smelter built at Kelvin to treat copper ores from Ray and Miami nines. Processed
over 16,000 tons of ore and later shut down (11).
1907 — Extensive development of Ray copper deposits (10).
— TownsIte of Miami established as a copper camp (2).
1909 — Hayden founded as milling and smelter camp for area copper mines.
— Inspiration Mine Company starts constructing huge reduction plant at Miami.
1911 — Copper production starts at the Miami Copper Company deposits.
— ASARCO builds copper smelter at Hayden to treat Ray copper concentrates and custom
ores. '
— Ray mines copper production starts on large sealei concentrating mill completed with
8,000 ton per day capacity.
-'.Copper production starts at Magma Copper Company Superior location (10).
1912 -- Miami Copper Co. complete* concentrator, bringing capacity to 1,000 tons of ore per
day (11)
— Discovery of rich copper ore at Superior in the mines which for several previous
years had been producing silver (11).
1915 — Flotation process introduced at Inspiration and becomes the first large-scale copper
flotation plant In U.S.A., with capacity of 14,000 tons of ore per day (10).
— International Smelting Company erects new smelter at Miami to handle the
Increased load of copper concentrates resulting from completion of two new mills at
Miami and Inspiration (11).
(continued)
E-3
-------
Table 2 (continued). CHRONOLOGY OF EVENTS
1923 — Miami Copper Co. increases mill capacity from 3,000 to 12,000 tons of ore per day
and improves block caving methods to increase ore production (11).
1924 ~ Smelter facility completed by Magma Copper Company at Superior (10).
~ Development started of low-grade ore at Miami.
1926 — Large-scale leaching and precipitation operations started at Inspiration—built
9,000 ton leaching plant (11).
1927 . -- Miami Copper Co. increases mill capacity from 12,000 to 18,000 tons of ore per
day (11).
ca. 1930 -- Depression of early 1930's has terrific impact on all copper mines in Arizona, with
many forced to close for various lengths of time (11).
-- Old Dominion mine and smelter at Globe shuts down (2).
1939 -- Advent of World War II causes copper prices to rise and increases production in
Arisona.
1943 — Open-pit copper production starts at Castle Dome mine, owned by Miami Copper Co. and
located six miles west of Miami. Built concentrator with 12,000-ton per day capacity.
(Venture backed by U.S. Government for war needs.) (11)
1948 -- Open-pit development is completed at Inspiration and Ray mines (10).
— Underground exploration starts at San Manuel (10).
1950 ~ Open-pit production starts at Ray mines (10). Prior to this, production was all
underground, with 79 million tons of ore mined from 1911-1954 (11).
1951 -- Xenneeo'tt installs pit-crushing plant at Ray and expands its milling facilities at
Haydan in order to maintain ore production and treatment at 15,000 tons per day.
ASARCO's Haydan plant smelts the concentrate.
1953 ~ Castle Dome ore reserves exhausted) mine closed after producing 257,000 tons of
copper plus gold and silver (11).
— Development started of Miami1* low-grade ore (10).
— Asbestos purchase depot opened by U.S. Government at Globe (10).
— Copper Cities deposit (three miles north of Miami and mined by Copper Cities Mining
Co., a subsidiary of Miami Copper Co.) comes into production (11).
ca. 1954 -- Community of San Manuel established as a new mining camp associated with development
of the San Manuel copper ore deposits. •
1954 — Miami starts production from low-grade copper ore body (10).
~ Concentrator moved from Castle Dome to Copper Cities mine site.
-- Ray mine stops underground mining and becomes completely open-pit operation. To
handle increased ore production, the concentrator capacity is increased to 15,000
tons per day (11).
1956 — San Manuel starts underground mining, flotation concentration and smelting of
concentrates (19). Flotation concentrator has 30,000-ton capacity (11)
1960 ~ Inspiration Consolidated Copper Co. purchases International Smelting and Refining Co.
smelter located at Miami. At time of purchase, smelter already treating Inspiration
concentrates, cement copper and material from toll customers (18).
ca. 1960 -- Community of Kearny built for the Ray Mines Division to house workers at neaby open-
pit mines and reduction plant.
~ Capacity of Ray mine's concentrator increased to 22,500 tons per day (11).
1962 — Copper Cities mine, since opening in 1953, has produced over 27 million tons of ore
at a daily rate of 12,000 tons (11).
~ Inspiration's Christinas mine to start production in mid-1962.
ca. 1963 ~ Miami Copper Co. has exhausted mineable grade ore in their underground Miami mines.
To recover remaining copper, the entire underground network was flooded with a
leaching solution (notei roughly 10 years later, in 1973, copper was still being
recovered by this leaching process) (20).
(continued)
E-4
-------
Table 2 (continued). CHRONOLOGY OF EVENTS
1967 . •— Bluebird mine, between the Inspiration and Castle Domes mines at Miami, produced
4,100 tons of ceoent copper. Mine owned by Ranchers Exploration and Development
Co. (11).
1969 — Magma Copper Co. starts program to increase ore production at San Manuul underground
Dine from 40,000 to over 60,000 tons per day by 1972. Plans started for electro-
lytic refinery at San Manuel with capacity to extract 200,000 tons of refined copper
per year starting 1972 (21).
— Magma Copper Co. starts program to increase ore production at Superior from 1,500 to
2,500 tons per day by 1974 by expanding mine and mill facilities. Plans call for
ultimate closing of Superior smelter and shipment of concentrates to Sim Manuel (21).
— Xennecott Copper Corp. Ray Mines Division completes electrowinnlng plant.to produce
refined cathode copper fron solutions obtained at its newly-built silicate ore, vat
leaching plant at Ray. Electrowinning plant designed to treat 10,000 tons of ore
daily also includes contact acid plant with designed capacity of 750 tons per day (21).
— Inspiration consolidated Copper Co. starts production of copper rod at its new mill
in Miami.
1971 — ASARCO installs sulfurie acid plant at Hayden smelter to reduce SOa emissions.
1973 ~ Inspiration Consolidated Copper Co. completes air pollution control project (to be
fully operational in 1974), consisting of building a new smelter with the world's
largest electric smelting furnace, new siphon-type converter•, a waste gas cleaning
plant and a sulfurie acid plant. Copper production will eventually rench annual rate
of 10 million pounds (20).
— Ksnnecott completes installation at their Hayden saelter of an SOj air quality control
system having an expanded 900 ton per day sulfurie acid plant, improved converter
hoods and gas collector, and a complex computer control system (14). tlhen the system
is fully operational in 1974, it ia expected to bring smelter into compliance with all
ambient alx quality standards (22).
'— Magma's San Manuel mine was largest producer in Ariiona.
— At San Manuel smelter, installed new water-cooled converter hoods as port of air
quality control program. Made plans for neutralization plant to dlspone of sulfurie
acid which will be a by-product of the new sulfurie acid plant, designed to convert
SOj gas from the smelter into acid. Acid plant projected to be completed in summer
' of 1974, with daily production of 2,000 tons of sulfurie acid at full production.
This smelter chosen by the Smelter Control Research Association as a tnst site for
proposed experimental plant control SOj generated by reverberatory furnaces (20).
-- Magma's Superior mine reached daily production of 3,000 tons from deepor high-grade
ore body of 4.2% copper (20).
— ASARCO smelter at Hayden modified by addition of anode easting plant to eliminate
casting of blister copper cakes that previously were remelted and cast into anodes at
the refinery. Also built 1,000-foot stack to discharge weak SOj gases that cannot be
handled by the acid plant (14). Anticipated that when stack is fully operational in
1974, ambient air quality will improve (22). ASARCO alao installing "closed loop"
SO] monitoring system at smelter that monitors ambient air quality and indicates when
SOj emissions exceed allowable limits so that smelting can be curtailed to adequately
control emissions (22). This system is jointly owned by ASARCO and Kennecott (23).
~ Kennecott initiates plans to expand Ray nines silicate leach plant from 10,000 to
14,000 tons of ore per day, with 14 months scheduled for completion (22).
1974 -- Cities Service's Pinto Valley open-pit mine (eight miles west of Miami) starts pro-
duction with expected daily capacity of 40,000 tons of ore by year's end. Capacity
at start-up is 20,000 tons of ore per day. The facility has a mill and sends con-
centrates to the Inspiration smelter (20).
-- Kennecott Copper Corp. Ray Mines Division received the first unconditional smelter-
operation permit from the Air Pollution Control Diviaion as a result of removing over
90* of smelter particulates and SOa emissions.
E-5
-------
HISTORY OF COPPER COMMUNITIES
Globe
Globe is located in a steep canyon, just south of the rugged Salt
River Canyon, at an elevatipn of 3,500 feet above sea level. A number of
hills and mountain ranges with elevations from 4,000 to 8,000 feet surround
the Globe-Miami area.
The community, today the County Seat for Gila County, was once a typical
example of the boom and bust company mining towns. Originally settled in
1876 as the result of a silver strike, the community really came into its
greatest wealth and prosperity starting in the late 1880's after the silver
had been exhausted and the rich copper deposits were discovered. The Old
Dominion Copper Company was established in the late 1880's with a "small
smelter," but this was soon replaced by a new, larger unit in 1898. For
the next 12 years, or until around 1923, the Old Dominion was "one of the
greatest copper mines in the world." Other mines developed in the a^rea
and there was extensive cattle grazing in Gila County, with much of the
income from these two businesses spent or invested in Globe. The town
flourished and called itself the "Capital City of the County with a Copper
Bottom" (2).
, The decline of Globe started in the early 1900's as a result of miners"
strikes, drought, overgrazing of the range, and the establishment of new
copper mines and a new camp in nearby Miami. The 1931 Depression closed
many area mines including the Old Dominion operations, and for two years
most of Globe's men worked for the WPA (Work Projects Administration). The
Old Dominion never reopened because of outdated equipment, very low grade
ore, and flooding of the mine. When operations started up again in Miami,
many of Globe's miners found work there. Globe has remained a quiet town
ever since.
Today Globe is a trade and service center for the surrounding areas of
southern and central Gila County.
A description of the community written in the early 1950's indicated
that the mining and smelting activities had had a long history of impact
on the area. At the northern end of town "black slag dumps and weathered
tailings dumps (gigantic accumulations of copper ore residue)" were the
remains of a once great copper mine (2). The waste material pushed out on
the flat plateaus and into the gulch of Final Creek and formed man-made
hills along the road.
Apparently by the 1950's the town's appearance had not changed signifi-
cantly from the original old-fashioned western mining town. Now, as in the
past, the town attracts many Apache Indians from nearby San Carlos Reserva-
tion. Even in the early days when the community was just a mining camp,
the population was a mixture of many nationalities including English,
Mexicans, Italians, and Slavs (2).
E-6
-------
Presently, Globe, along with many of the other communities in the
study area, faces the problem of lack of adequate housing available for
residents. A 1970 survey found that slightly over half of the town's
2,899 housing units were deteriorating or dilapidated, and that approxi-
mately 300 families were living in substandard housing and were unable to
afford anything better (6).
Miami-Inspiration/Central Heights
Miami is situated near Globe, at an elevation of 3,500 feet above sea
level, in the foothills of the Final Mountains. As noted in the descrip-
tion of Globe, Miami lies between the walls of a fairly steep canyon.
In 1907 the camp of Miami was established approximately seven miles
west of Globe beside low-grade copper deposits that were much more exten-
sive than those of Globe's Old Dominion mine. Originally the huge deposit
was mined by two companies—i.e., the eastern part by the Miami Copper
Company and the western part by the Inspiration Consolidated Copper (ICC)
Company. In 1912 the ICC Co. began construction of its huge reduction
plants, and from 1915 to 1920 the Miami district was producing to the
limit of its capacity. The community spread out into the suburbs of Lower
Miami and neighboring Claypool and supported a population that was growing
larger than that of its rival, Globe.
Then, as now, the town's business section was on the valley floor, and
most residences were located in the hills and canyons surrounding the
floor. For a long time the area supported a mixture of flimsy homes,
typical of mining camps, and some better residences. On the hills to the
north of town, company villages and the huge cppper plants have been exist-
ing for over 30 years. A very graphic description of the area, written in
the early.1950's, vividly portrays the industry's effect on the town:
"The smelter's coal-black slag dumps and cement-colored
tailings from the concentrators stretch for miles over the
heights and down into the flat alongside the town. At night
when the smelter is working, the mounds of slag glow red with
streams of molten rock dumped from the furnaces." (2)
In the 1920's Miami was a rowdy community with over 40 gambling halls
and brothels. Today it is a far more quiet place with roughly half the
population size. A 1970 survey found 54 percent of the town's 2,775 hous-
ing units deteriorated or dilapidated, with some in such bad condition
that clearance and redevelopment were suggested as the most feasible means
of providing safe, decent housing. The survey also found the majority of
commercial buildings in the business district were vacant, deteriorating
and/or obsolete (6). For many years now there has been little construc-
tion and residential growth in Miami.
Up to 1971 the most recent development in the Miami area had taken
place in the Central Heights and Corbe Valley districts. Little Acres,
Inspiration, and Miami Gardens were predominantly company housing projects
E-7
-------
(built and maintained by Inspiration Consolidated Copper Co.) and offered
hosr.es that were in better than average condition (6) .
In the 1950"s Inspiration was noted as a well-kept company-owned vil-
lage for the mining officials, supervisors, and skilled employees. Even
then, the huge vats of the ICC leaching plant could be seen in the gulch
below the community (2).
Prior to 1907 Miami was almost completely uninhabited by whites. As
the town prospered, more than a dozen different nationalities were repre-
sented, with the Mexicans and Slavics forming the largest groups. After
the postwar depression of 1921, the racial composition started shifting
towards having a majority of Mexicans (2).
Mining activity across Arizona had been severely curtailed during the
econosnic depressions of 1921 and 1931. Unlike some other mines in the
area, those of Miami were able to resume full production in the post-
depression periods. During the 1931 slump Miami Coppar Co. curtailed
production and ICC shut down almost completely, putting almost the entire
town's population out of work.
During Korld War II, although hampered by strikes, Miami district
mines produced copper to full capacity. To help the war effort, the Castle
Doms mine, to the west of town, was brought into production and kept opera-
ting until the ores were exhausted in the 1950's.
Today Miami is the scene of extensive copper mining and processing
activities (see Table 2). The ICC Co. has three open-pit mines, and
ssieiter and rod fabrication plants. Ranchers Exploration and Development
Corp. has open-pit and underground operations, plus leaching and concen-
trating facilities at their Bluebird mine. Cities Service Co., mining in
the area since 1911, has a newly-developed open-pit facility, Pinto Valley,
and a new underground mine, Miami East, in addition to its Miami mine.
This firm also operates two concentrator and leaching facilities and an
electrowinning plant.
The copper industry is the major employer, and other area activities
such as cattle ranching, asbestos mining and processing, and sawmill opera-
tions offer a limited number of jobs to Miami residents.
ttaydan and Winkelman
Hayden and Winkelman are located in the San Pedro Valley at the con-
fluence of the San Pedro and Gila Rivers. The terrain is rolling desert
with an elevation of approximately 2,000 feet above sea level.
Winkelman was originally founded in the 1870's as an agricultural com-
munity serving the agricultural businesses of the San Pedro and Gila River
valleys, and also served as a shipping point for cattle raised in these
valleys. Later, as the copper industry moved into the area, Winkelman be-
came a trading center and residential area for employees of the mines and
shelters (1).
S-8
-------
Hayden was originally founded in 1909 as a milling and smelter camp
for the area mines and remained a "company town" until 1956 when it was
incorporated and became a self-governing community.
. Productive mining in this area started in the early 1900's with the
development of the Ray mines, now owned by Kennecott Copper Corp. Also in
the early 1900's American Smelting and Refining Co. built a smelter in
Hayden to treat concentrates from the Ray mines. In 1958 Kennecott also
constructed a smelter at Hayden. Presently the mining and ore processing
activities of these two companies dominate the economies of both communi-
ties.
For many years now, growth and development of these communities have
been severely restricted since the mining companies own extensive land-
holdings in this area. With the exception of land presently occupied by
homes and businesses, most of the remaining area is owned by the copper
companies (6).
A 1970 survey of these two communities indicated that 59 percent of
the housing stock (i.e., total of 741 units) were sound. The remainder
were deteriorating or dilapidated, and many had excessive deterioration (6)
San Manuel and Mammoth
San Manuel and Mammoth are both located in the broad San Pedro River
Valley at elevations ranging from 4,500 feet (i.e., Mammoth) to approxi-
mately 7,000 feet (i.e., San Manuel) above sea level.
San Manuel was built in 1954 by the San Manuel Copper Corp., a wholly
owned subsidiary of the Magma Copper Company, as a mining camp for the
employees of the newly established San Manuel mine, mill, smelter and re-
fining complex. Originally the camp had approximately 7,000 residents and
was described as a completely modern town with 2,000 homes plus parks,
schools, churches, etc., that had sprung up in the desert overnight (2).
In the'1950's, when the mining was getting under way, the San Manuel
mine was noted as one of the largest copper producers of the time. The
company was then planning to spend over $100 million to create a producing
capacity of over 70,000 tons of copper annually (2).
Mammoth, unlike San Manuel, was founded in the late 1800's as a gold
mining camp and site of a stamp mill for gold taken from the area mines.
The community is located near the St. Anthony deposit which was developed
in the late 1800's as three separate properties—Collins Mine, Mammoth
Mine, and Mohawk Mine. From its establishment in the 1800's to its ex-
haustion 'around 1953, this cluster of mines produced two million tons of a
variety of ores including gold, silver, vanadium, molybdenum, lead, and
zinc. In the early 1880's Mammoth was one of the "busiest and most lusty"
of mining camps in the country. When the gold mill was moved out of town,
and much later when the ores were depleted, Mammoth survived. The town has
E-9
-------
been sustained by residents in the valley, the stockmen, the agricultural
sector, and by the fact that much money has been spent there by the San
Manuel Copper Corp. and its employees for supplies, services, and recrea-
tion (1).'
Iri the 1950's Mammoth was a quiet one-street town extending about one-
half mile and having many homes and stores built of adobe, with many al-
ready falling into ruin (2).
When the St. Anthony deposit was active, the community of Tiger was
one of the area's largest towns but its existence was totally dependent on
the mines. When the ores were depleted in the 1950's, most of Tiger's
residents were forced to move and most of them went to Mammoth (see Table 1)
Today Tiger consists of a warehouse and storage area for some mining equip-
ment (1).
Oracle
This community was founded in the 1880's as a supply center for several
area mines. It is located in the San Pedro River Valley at an elevation
of 4,500 feet.
Oracle provides homes for many of the area miners and for many years
also has been a health resort and guest ranch community. Many rustic homes
and guest ranches are in the hills surrounding Oracle.
Superior
Attention was first attracted to the Superior area in 1875 with the
discovery of the rich Silver King mine. Subsequent opening of the nearby
Silver Queen mine caused a great influx of people to the area. In 1910 the
Magma Copper Company (today a subsidiary of Newmont Mining Corp.) formed to
take over the Silver Queen properties, and 14 years later built a large
smelter to process all of its ore plus hundreds of carloads of ore from
smaller mines. The copper ore from the Magma mine was so rich that the
mines and smelter were kept operational even during the slump of 1929-1933,
when many mines throughout the country were shut down (2). The smelter at
.Superior was eventually shut down in 1971.
The community of Superior originally was established as a mining camp
to house workers from Magma's underground mine. The commercial and service
sectors primarily serve the mine employees in the immediate area.
Kearny
Kearny, located at an elevation of approximately 2,000 feet, is a com-
pletely planned community developed about 17 years ago for the Ray Mines
Division of Kennecott to house workers at their nearby open-pit mine and
reduction plant. Ranching, farming, and home building have a great influ-
ence on the economy of the community; however, the copper industry dominates
E-10
-------
the financial structure. The community is described as having a completely
modern architectural design and no substandard housing (19).
The Ray mines and ASARCO's smelter at nearby Hayden provide most of the
employment for Kearny residents. The commercial and service Electors pro-
vide a limited number of jobs.
HISTORY OF COPPER COMPANIES
Inspiration Consolidated Copper Corp.
This company owns and operates several mines including the Thornton,
Live Oak, Red Hill, Black Copper, and Ox Hide mines around Inspiration and
the Christinas open-pit and underground mines near Christmas. By-products
of the mining operations include molybdenum, silver, gold, and selenium.
The Ox Hide mine is a heap-leaching operation. Ore is mined, placed
on pads with impermeable bottoms, and leached with recirculated sulfuric
acid solutions. Copper is recovered from solution by precipitation on
scrap iron and then is trucked to Inspiration for smelting and refining.
\ 1 ' •
The smelter,located at Inspiration treats copper-bearing materials
from the various Inspiration mines, plus custom and toll materials from
outside customers. New copper-bearing materials treated in 1970 totaled
313,431 tons as compared with 316,591 tons in 1969.
At the Christmas mine, processing facilities include crushing and con-
centrating equipment, with the concentrates being sent to Inspiration for
smelting and-refining. At the remaining mines around Inspiration, facili-
ties include crushing and leaching plants, concentrators, precipitation,
electrowinning and electrorefining plants, smelter, and rod fabrication
plants.
Around.1972 all the mines and plants were operating at a full daily
capacity rate of roughly 42,000 tons of ore. At the copper content for
that year, it was estimated that total company mine production would have
been approximately 70,000 tons of copper per annum. At that time it was
also estimated that the electrorefining plant and smelter were operating
at capacity and producing approximately 45,000 tons of copper for return
to toll customers. (Source: Company Annual Report, no date on our Xerox
copy, but probably 1972.)
In 1951 the company owned 2,825 acres at Inspiration and had a 10,000-
ton leaching plant and 15 unit concentrator with sliming department (15).
Tables 3 and 4, extracted from references 15 and 16, give copper production
figures for years 1948 through 1951 and years 1958 through 1961, respec-
tively.
E-ll
-------
Table 3
SMELTER PLANT PRODUCTION, 1948-1951 (15,16)
(tons of copper)
From Ores:
• Electrolytic .copper from
cathode & starting sheets
• 'Blister copper in excess
of starting sheet require-
ments
From Leaching-Iri-Place:
• Blister copper
1948
31,317
7,034
--
1949
26,264
4,959
—
1950
29,446
6,906
2,089
1951
30,297,
4,680
4,147
Table 4
SMELTER PLANT PRODUCTION, 1958-1961 (15,16)
(tons of copper)
, , 1
Prom Ores:
• Electrolytic copper
•• Blister copper
• Cement copper
From Leaching-In-Place:
• Electrolytic copper
• Blister copper
1958
37,511
1,152
—
2,379
778
1959
i
41,659
1,534
—
1,133
2,685
1960
35,324
2,460
75
2,430
112
1961
36,616
212
397
1,940
— —
E-12
-------
In 1960 the company bought the International Smelting and Refining Co.
smelter located at Miami. At the time of purchase the smelter already was
treating Inspiration concentrates, cement copper, and material from toll
customers. By the end of 1960 the company also had plans to expand their
electrolytic refinery at Inspiration. This enlargement plus the smelter'
purchase was to enable the company to ship the entire output of its mines
and a part of the smelter's toll and custom ore, in the form of refined
copper cathodes (18).
In 1962 the company owned 3,034 acres at Inspiration, including the
Live Oak Pit and Thornton Pit. The Christmas mine was scheduled to come
into production in mid-1962. Also in 1962 the facility had a 15,500-ton-
per-day capacity dual-process leaching plant and flotation concentrator
at Inspiration, and a 4,000-ton-per-day capacity flotation concentrator at
Christmas (16).
In 1962 the copper smelter at Inspiration had a 1,200-ton-per-day
capacity furnace, a 260-ton-per-day capacity anode plant, and a 140-ton-
per day capacity electrolytic refinery (16).
In 1969 mill capacities were as follows: at Inspiration, a leaching
plant and flotation concentrator at 25,000 TPD; and at the Christmas mine,
a flotation concentrator of 5,500 TPD (17).
In 1969 the Inspiration smelter had a 1,300 TPD furnace capacity; an
anode plant of 260 TPD capacity; an electrolytic refinery of 190 TPD capacity;
and a continuous-casting rod plant of 250 TPD (17).
ASARCO
Table 5 summarizes the material processed and the blister copper pro-
duced at ASARCO1s copper smelter located at Hayden.
Table 5
PRODUCTION AT ASARCO'S HAYDEN SMELTER (9)
Year
1970
1971
1972
1973
1974
1975
1976
Material Processed (tons)
716,000
632,000
755,000
746,000
614,000
522,000
523,000
Blister Copper Produced (tons)
148,000
134,000
160,000
147,000
129,000
105,000
102,000
E-13
-------
The Hayden smelter principally treats copper concentrates and precipi-
tate^ produced 'by mines in the southwestern United States. ASARCO operates
two additional-smelters, with one in Texas and the other in Washington.
For all three smelters combined, roughly one-third of the incoming copper
is received from ASARCO1s mines, approximately one-half is treated for
other companies on a toll basis, and the remainder is derived from pur-
chased materials.
To comply with pollution control standards over the past five years,
the company installed a sulfuric acid plant to reduce SOa emissions.
Magma Copper Co.
In 1951 the Magma Copper Co. owned 2,478 acres at Superior.
In 1961 the company's Magma mill at Superior was an 1,800-ton-per-day
(TPD) flotation mill, and the one at San Manuel was a 35,000 TPD flotation
mill. Both locations employed a total of 3,500 workers in 1961 (16).
The Superior underground mine produces copper, gold, and silver, while
the San Manuel -mine produces (by underground block caving) copper, gold,
silver, and molybdenite.
In 1969 the Superior smelter produced 17,618 tons of copper and the San
Manuel smelter produced 95,722 tons of copper (17). In the early 1970's
the Superior smelter was permanently shut down and concentrates from the
mill were sent to the San Manuel smelter for processing.
Kenhecott Copper Corp.
Kennecott owns and operates four open-pit copper mines in western
United States. One of these, the Ray mine, is located approximately 10
miles northwest of Kearny. Molybdenum, gold, silver, and other nonferrous
metals are contained in the copper ore and are extracted as by-products.
Kennecott also owns and operates a concentrating mill which converts
the ore into copper concentrates by a process of crushing, grinding, and
flotation. Kennecott operates a smelter at Hayden to process its copper
concentrates into blister copper. The blister copper is then shipped to
one of Kennecott's plants for eleetrorefining (i.e., at the electrore-
fineries at Salt Lake City, Utah or Baltimore, Maryland) or fire refining
(i.e., near Silver City, New Mexico).
Two types of ore are mined at Ray: 1) a sulfide ore which is processed
at the concentrating and smelting facilities located at Hayden (approxi-
mately 20 miles south of the mine), and 2) a silicate ore which is sub-
jected to a hydrometallurgical process at a plant located adjacent to the
mine site.
Waste rock removed at the mine is hauled to a waste dump and subjected
to a leaching process. Water is pumped to the top of the dump and allowed
E-14
-------
to percolate downward, leaching out the soluble copper. The resulting
copper-bearing solution is treated in a nearby precipitation plant to pro-
duce precipitate copper which also is shipped to the Hayden smelter.
In 1951 the company had an 8-unit concentrator-flotation plant at
Hayden. At"that time a large part of the mining activities were being
converted from underground to open-pit operation. To maintain ore produc-
tion and .the .treatment of 15,000 tons per day, a pit-mine crushing plant
was installed at Ray and milling facilities were expanded at Hayden. At
that time ASARCO's Hayden smelter was smelting the copper concentrates
from Ray. The company employed 600 men in 1951 (15).
In 1969 the open-pit mine at Ray covered about 2,100 acres (17). In
the same year.the company completed an electrowinning plant to produce
refined copper cathode from solutions from the silicate ore, vat leaching
plant (design capacity 10,000 tons of ore daily) located at Ray. The com-
pany also completed a contact acid plant with a 750-ton-per-day capacity
(21).
HISTORY OF OTHER INDUSTRIES IN THE STUDY AREA
Asbestos Mining and Milling
Gila County, in the vicinity of Globe, is the only area in the United
States that-produces high-grade, low-iron-content chrysotile asbestos. The
chrysotile deposits, first discovered in this area around 1878, were not
exploited until 1913. This started a rush that lasted through World War I
to locate and develop deposits father northwest in the county. Additionally,
this activity was extended to the eastern part of the county subsequent to
the opening in 1921 of the Fort Apache and San Carlos Indian Reservations to
noninetallic mining. In 1920 annual production reached a peak of 1,200 tons,
a rate not duplicated again until 1927-1928., The depression of the 1930's
virtually stopped the asbestos mining through the middle of the decade.
Mining 'increased again in the late 1930's, and by 1940 annual production
reached 1,200 tons. With the exception of a few years at the end of World
War II, annual production has consistently matched or exceeded this level
(11).
In 1952 the General Services Administration established a depot at
Globe for the purchase of strategic grades of chrysotile. This stimulated
the local industry from the postwar decline, and at the peak of activity
at least 16 deposits were producing asbestos or were being developed. Each
year since 1962, when the purchase program was terminated, mining has been
limited between two and four sites. Since 1948, annual production with few
exceptions has exceeded 2,000 tons and in several years has considerably
exceeded 3,000 tons (11).
E-15
-------
Figure 1 indicates the locations of asbestos deposits in Gila County.
Mo'st of the deposits are in two areas that cover about 100 square miles in
the northern part of the county, approximately 25 to 35 miles northeast of
Globe.
Prior to 1941, asbestos processing mills were built at several of the
mine sites. These mills were used only briefly, and all but a small por-
tion of the material produced was actually hand-cobbed fiber. Since the
post-World War II era, construction of mills has concentrated in Globe,
where there are adequate power facilities (11). In the 1950's there were
several mills in Globe, including those of Jaquays Mining Corporation,
Metate Asbestos Corporation, and Le Tourneau Asbestos Company.
By 1970 chrysotile was being produced from only one underground mine
(owned by Jaquays). From 1973 to 1974, the quantity of asbestos produced
dropped 69 percent due to a loss of market caused by new EPA regulations
on consumer use and handling procedures (23). As a result of these rulings
arid other air quality control regulations, all mills except the Jaquays
facility went out of business. In 1974 the Jaquays plant, consisting of
crushing, screening, and air-separation units, underwent a $100,000 modi-
fication program to comply with air pollution requirements for asbestos
processing (23).
Other Mining Activities in Gila and Final Counties (12)
Because of the rich variety of mineral resources throughout Arizona
and the study area, there are numerous types of mining activities in addi-
tion to the copper industry. Some of these other concerns are as fol-
lows:
The Arizona Gypsum Company, Final Mammoth Gypsum Company, and National
Gypsum Company produce gypsum from open-pit mines in the lower San Pedro
Valley deposits near Winkelman.
There are three open-pit mines and processing plants for crude perlite
in the Miami and Superior areas. One company operates an open-pit mine,
crusher, and furnace for mercury deposits in Gila County.
In addition to the copper mined by the subject industries, molybdenum,
gold, and silver are also produced at the mines operated by Inspiration
Consolidated Copper Company, Kennecott Copper Company, Cities Service
Company (Miami Division), and Magma Copper Company's San Manuel and
Superior Divisions (there is no molybdenum mining at the Superior mine).
E-16
-------
OPUMTIN
I'M «f «*Mlt*
O
[Source; U.S. Geological Survey, Arizona Bureau of
Mines, U.S. Bureau of Reclamation, "Mineral and Water
Resources of Arizona," Committee on Interior and
Insular Affairs, U.S. Senate, 1969.]
Figure 1. Asbestos Deposits in Gila County
E-17
-------
REFERENCES
1. Cteasey, S. C., "Geology of the San Manuel Area, Final County,
Arizona," Geological Survey Professional Paper #471, 1965.
2. Miller, Joseph, "Arizona - The Grand Canyon State - A State Guide,"
compiled by Workers of the Writers' Program for the WPA, 1940 (first
published), 1956 reprint.
3. Arizona Department of Mineral Resources, "Historical Data and Special
Articles Relating to the Copper Industry in Arizona, the United States
and the World Showing the Need for a Copper Tariff," October 1957.
4. Arizona Office of Economic Planning and Development, "Kearny-Arizona
Community Profile." (no .date)
5. Arizona Office of Economic Planning and Development, "Globe-Arizona
Community Profile." (no date)
6. Ferguson, Morris and Assoc., "General Development Plan, Gila County,
Arizona, 1971."
7. Arizona Department of Mineral Resources, "Copper Industry Statistics
for 1965 Compared With Other Years, Arizona, United States and Wprld,"
August 1966.
8. Securities and Exchange Commission, Washington, D.C., 10-K Form,
Annual Report Persuant to Section 13 or 15(d) of the Securities and
Exchange Act of 1934, for Fiscal Year Ending December 31, 1976,
Commission File #1-1369, Kennecott Copper Corp.
9. Same as reference 8» for ASARCO.
10. Tuck, Frank J., "Fifty Years of Mining in the State of Arizona 1912-
1962," Arizona Department of Mineral Resources (no date).
11. U.S. Geological Survey, Arizona Bureau of Mines, U.S. Bureau of
Reclamation, "Mineral and Water Resources of Arizona," Committee on
Interior and Insular Affairs, U.S. Senate, 1969.
12. U.S. Department of the Interior, Preprint from U.S. Bureau of Mines
Yearbook, "The Mineral Industry of Arizona," 1970.
13. Arizona Department of Mineral Resources, "Mining in Arizona - Its
Past - Its Present - Its Future," 1964.
14. Furness, J. W., "History of the Development of the Copper Industry of
the World," U.S. Bureau of Mines; from "Copper Resources of the World,"
vol. XVI, International Geological Congress, 1933.
E-18
-------
15. "Mines Register - Successor to the Mines Handbook and the Copper
Handbook, vol. XXIV - Describing Non-Ferrous Metal Mining Companies
in the Western Hemisphere," Atlas Publishing Co., 1952.
i,
16. Nightingale, Geoffrey (ed.), "Mines Register," vol. XXVI, American
Metal Market Co. Publishers, 1962-1963.
17. Ranriells, A. Karl (ed.), "Mines Register," vol. XXIX, American
Metal Market Co. Publishers, 1971.
18. Shillings Mining Review, vol. XLVIII, #50, March 12, 1960.
19. Chamber of Commerce Data for the Community of Kearny, Arizona, no
date.
20. Arizona Department of Mineral Resources, "Copper Industry Statistics
for 1973 Compared With Other Years, Arizona, United States and World,
February 1975.
21. Arizona Department of Mineral Resources, "Copper Industry Statistics
for 1969 Compared With Other Years, Arizona, United States and World,
September 1970.
22. Preprint from the 1973 Bureau of Mines Yearbook, "The Mineral Indus-
try of Arizona," U.S. Department of the Interior.
23. Preprint from the 1974 Bureau of Mines Yearbook, "The Mineral Indus-
try of, Arizona," U.S. Department of the Interior.
E-19
-------
Appendix F
MORTALITY AND POLLUTION DATA
IN SUPPORT OF COPPER SMELTING CASE STUDY
Table F-l Statistical Data for Douglas and Control
Table F-2 Statistical Data for Anaconda
Table F-3 Statistical Data for Magna and Controls
Table F-4 Data for Aggregated Arizona Smelter Towns
Table F-5 Annual Average Sulfur Dioxide Levels in Arizona Smelter
Towns
Table F-6 Sulfur Dioxide Levels in Anaconda
Table F-7 Emissions from Smelter and Estimated Sulfur Dioxide
Concentration for Magna, Utah
Table F-8 Particulate Levels for Anaconda
Figure 1 Total Suspended Particulate Concentrations Versus Time
for Utah CHESS Communities, 1940-1971
Table F-9 Summary of Particulate Data for Arizona Smelter Towns
Table F-10 Environmental Cadmium Data
Table F-ll Environmental Lead Data
Table F-12 Environmental Arsenic Data
Table F-13 Fluoride Content of Random Grass Samples
Table F-14 Body Burden Data
Table F-15 Age-Adjusted Mortality Rates for White Females (per
100,000) for Anaconda and Controls
Table F-16 Age-Adjusted Mortality Rates for White Females (per
100,000) for Arizona Smelter Towns and Controls
Table F-17 Age-Adjusted Mortality Rates for White Females (per
100,000) for Magna and Controls
Table F-18 1976 Sulfur Dioxide Data Summary
-------
TABLE F-l
STATISTICAL DATA FOR DOUGLAS AND CONTROL
Observed _ , ^ .
, • Relative o
Number _. , X
, _ .. • . Risk
of Deaths
(a) «>)
Cause of Death 10
Douglas
Control
Cause of Death 12
Douglas
Control
Cause of Death 20
Douglas
Control
Cause of Death 27
Douglas
Control
Cause of Death 31
Douglas
Control
17
978 2.1 8.6
32 32.9
1,267 2.9
31 153.7
386 8.4
16 7.2
1,000 2.0
15 11.9
757 2.6
Population Population
35-74 WF All WF
2,400 6,455
300,341 779,118
(a) All ages, WF, 1968-1975.
(£>) Note: The relative risk and x2 were calculated for the 35-74 age group to determine the influence
of the older age group on the significance of the results. R.R. and x2 decreased in all cases,
however, all diseases remained significantly elevated relative to the control.
Disease
10
12
20
27
31
R.R.
2.1
3.2
5.7
2.0
3.3
x2
6.4
30.1
40.0
5.5
16.9
-------
TABLE F-2
STATISTICAL DATA FOR ANACONDA
«g
i
..-
Cause of Death 5
Anaconda
Controls :
Great Falls
Havre
Billings
Bozeman
Livingston
Cause of Death 7
Anaconda
Controls :
Great Falls
Havre
Billings
Bozeman
Livingston
Observed
Number
of Deaths
(a)
17
40
8
35
22
11
10
24
4
16
5
2
Expected
. Number
of Deaths
(a)
11
47
9
48
13
9
5
21
4
21
6
4
SMR
160
86
86
73
165
121
209
114
99
75
85
50
Relative
Risk
(b)
2.9
7.9
6.6
0.7
1.7
1.3
1.9
9.5
4.3
5.7
X2
(c)
7.0
4.3
5.87
0.24
0.6
0.26
0.48
17.9
3.6
2.66
Population Population
35-74 WF All WF
2,188 4,984
10,749 30,093
1,766 5,079
11,475 31,414
2,522 9,118
1,630 3,660
-------
I
to
TABLE
F-2
(continued)
Cause of Death 19
Anaconda
Controls :
Great Falls
Havre
Billings
Bozeman
Livingston
Cause of Death 27
Anaconda
Controls:
Great Falls
Havre
Billings
Bozeman
Livingston
Observed
Number
of Deaths
(a)
74
245
37
243
72
57
14
36
5
39
5
4
Expected
Number
of Deaths
(a)
53
235
51
248
71
49
7
31
6
32
8
5
SMR
140
104
72
100
101
117
208
117
88
122
61
74
Relative
Risk
(b)
1.5
2.3
1.8
2.3
1.6
1.7
4.2
4.2
2.1
4.0
X2
(c)
3.0
4.2
6.0
5.4
1.4
1.6
3.0
3.2
3.7
4.0
fa; All ages
(b) 35-74
(c) X2 > 3.85 indicates p < 0.05
-------
TABLE F-3
STATISTICAL DATA FOR MAGMA AND CONTROLS
Cause of Death 12
Magna
Controls:
Logan
Price
Cedar City
Cause of Death 14
Magna
Controls:
Logan
Price
Cedar City
Cause of Death 18
Magna
Controls :
Logan
Price
Cedar City
Observed
Number
of Deaths
(a)
22
33
18
14
14
21
10
12
139
353
122
123
Expected
Number
of Deaths
(a)
18
43
16
15
10
30
10
9
123
402
124
102
SMR
123
76
110
93
138
70
101
136
113
88
98
120
Relative 2
Risk X
(b) (c)
1.9 4.2
1.2 0.2
1.6 1.3
3.5 5.4
3.0 2.7
1.6 .008
1.8 9.8
1.4 2.1
1.1 0.15
Population Population
35-74 WF All WF
1,557 4,975
2,803 11,058
1,295 3,202
1,262 4,462
(a) All ages
(b) 35-74
(c) x2 > 3.85 indicates p < .05
-------
TABLE F-4
DATA FOR AGGREGATED ARIZONA SMELTER TOWNS (Number of Deaths, White Females, All Ages, 1968-1975)
Ul
— ~___^_^ POPULATION
___^ (WP, all ages)
DISEASE CATEGORY ' .^TOWN^
Organ Cancers:
Cancer, oral, pharynx and esophagus
Cancer, stomach
Cancer, duodenum and small intestines
Cancer, colon and rectum
Cancer, liver and bile ducts
Cancer, pancreas
Cancer, respiratory organs
Cancer, genitourinary organs except bladder
Cancer, bladder
Total
Other cancers except leukemia
Neoplasms of lymphatics
Vascular Diseases:
Hypertension disease
Cardiovascular diseases except stroke
Cerebrovascular
Total
Respiratory Diseases:
Acute respiratory diseases
Bronchitis, emphysema and asthna
Total
Digestive Diseases:
Other digestive diseases
Diseases of stomach and duodenum
Apendicitus, hernia and others
Gastroenteritis; ncnir.fectious
Diseases of liver, gallbladder and pancreas
Total
Urinary and Genital Diseases:
Diseases of urinary system
Diseases of the genital organs
Total
672
Hayden
o.
0
0
0
0
0
0
0
0
0
1
2
0
12
0
12
S
0
5
0
0
0
0
1
1
1
0
1
1,717
Miami
0
0
0
4
0
2
1
4
1
12
7
1
2
87
0
89
10
1
11
0
0
6
0
S
11
6
0
6
2,050
San Manuel
0
0
0
0
0
0
1
1
0
2
4
1
0
11
0
11
2
0
2
0
O
0
0
0
0
I
0
1
1,582
Morenci
0
0
0
0
0
2
0
1
4
8
5
2
0
28
7
35
2
2
4
0
0
0
0
4
4
1
O
1
6,455
Douglas
1
4
1
13
1
2
5
17
3
47
32
8
4
206
52
262
31
6
37
0
1
$
0
16
26
8
0
8
2,693
Ajo
0
1 .
1
1
1
0
2
5
0
12
7
2
1
42
13
56
13
1
14
0
1
2
0
2
5
1
0
1
15,169
Town Total's
81
56
16
465
76
47
IB
Control3
779,118
Arizona
4,040
1,267
684
25,916
386
854
587
aData are from the State of Arizona minus r.ila and Final Counties.
-------
TABLE F^5
ANNUAL AVERAGE SULFUR DIOXIDE LEVELS
IN ARIZONA SMELTER TOWNS (pg/tn3)
Location 1969 1970 1971
Ajo
Claypool
Clifton
Douglas ESE3
Douglas ENE4
Hayden
San Manuel
Average
Average for
1969-1971: 135
98
147
90
64
109
377
147
147
189
48
135
44
51
481
113
151
50
46
110
57
55
336
101
107
Source: Arizona Department of Health Services
F-6
-------
TABLE F-6
SULFUR DIOXIDE LEVELS IN ANACONDA
(lead peroxide method)
Date of Collection S02 (yg/m3)
5/9/66 - 6/2/66 50.8
4/11/6G - 5/9/66 39.3 (low)
3/14/66 - 4/11/66 48.3
2/9/66 - 3/14/66 42.6
1/12/66 - 2/9/66 47.5
11/17/66 - 12/16/66 63.1
10/15/66 - 11/17/66 70.5
9/9/65 - 10/15/66 81.1
,8/12/65 - 9/9/65 73.8
7/15/65 - 8/12/65 90.2 (high)
Average 61 ±17
Source: Montana Air Pollution Bureau
F-7
-------
TABLE F-7
EMISSIONS FROM SMELTER AND ESTIMATED
SULFUR DIOXIDE CONCENTRATION FDR MAGNA, UTAH
SO2Emissions
Year Tons/Day S02(|jg/m3)
1971 193 .- 61.8a
1970 261 Q4b
1969 322 I03b
aWest-Gaeke Method
b
Values estimated by CHESS
Source: Health Consequences of.Sulfur Oxides:
A Report from CHESS. 1970-1971. U.S. EPA,
Office of R&D, National Environmental
Research Center, Research Triangle Park, N.C.
F-8
-------
TABLE F-8
PARTICULATE LEVELS FOR ANACONDA
Total Suspended
Year
1961-1962
1961-1962
1962
1962
1962
Season
Year
Dec . -Feb .
Mar. -May
Jurie-Aug.
Sep. -Nov.
No. of
Samples
123
30
31
32
30
Particulates (ug/nr)
Max.
339
181
339
166
225
Min.
10
11
10
51
20
Avg.
89
62
115
104
73
Source: Montana Air Pollution Bureau
F-9
-------
FIGURE 1
TOTAL SUSPENDED PARTICULATE CONCENTRATIONS
VERSUS TIME FOR UTAH CHESS COMMUNITIES, 1940-1971
(Partibulate Levels for Magna)
I I I I II I I I I I
I I I I I I I I II I I I I I
40GDEN
o SALT LAKE CITY
OKEARNS
oMAGNA
ANNUAL AIR QUALITY STANDARD
I I I I I I I I I I I I I I I I I I I I I I
1940
1945
1950
1955
TIME, years
1965
1970 1971
Source: Health Consequences of Sulfur Oxides: A Report from CHESS.
1970-1971. U.S. EPA, Office of R&D, National Environmental
Research Center, Research Triangle Park, N.C.
F-10
-------
TABLE F-9
SUMMARY OF PARTICULATE DATA FOR ARIZONA SMELTER TOWNS
(Annual Geometric Mean in yg/m^)
Location 1969 1970 ..1971
Ajo
Douglas NNE
Hayden
San Manuel
Clifton
East Plantsite
Claypool
Average ± 0
Average for
1969-1971: 122 ± 16
87
240
224
43
51
51
117
i ± 83
83
264
*
55
51
51
165
111 ± 86
75
303
*
*
43
*
144
141 ± i
*no data available
Source: Arizona Department of Health Services
F-ll
-------
TABLE F-10
ENVIRONMENTAL CADMIUM DATA
Arizona (1969 data, yg/m3)
Sme.lter Towns
Ajo
claypool
Douglas
E. Plantsite
Hayden
San Manuel
.005
.02
.011
.006
.321
.002
Non-Smelter
Davis Dam
Florence
Organ Pipe
Page
Average ± a
Towns
.002
.013
.002
.002
.004 ± .005
Average ± o .006 ± . 005
Anaconda (1971-1972 data, pg/m3)
American Junior High School .02
Kirkeby Residina .02
Farmers Union 0
S.&N Cement .06
Johnson's Residence .04
Bailey's Residence .04
Average ± a
Soil Analysis'
. 03' ± . 02
Cd' (ppm) = 3.6
Magna (Nov. 1970 - Feb. 1972 data, mg/m2/month)'
Nov. '
Dec.
Jan. '
Feb.
March
April
70
71
0.
0.
0.
0.
0.
0.
17
00
05
00
19
20
May
June
July
Aug.
Sept.
Oct.
0
0
0
0
0
-
.14
.05
.07
.16
.23
Nov.
Dec.
Jan.
Feb.
1971
'72
0
0
0
0
Average
.07
.04
.05
.02
± «
- .11± .08
Comment; There are no environmental standards for cadmium. The 1963
National Air Sampling Network found a range of 0.004 to 0.026 pg/m3
in non-urban areas. The highest value in the survey was 0.050 pg/m3
in Covington, Kentucky. Standard soil level is < 1 ppm.
Arizona Department of Health Services .
b
Montana Air Pollution Bureau.
<•«•
"Sample taken by Enviro Control site visit team, November 1977.
CHESS study (This data is the only cadmium data for Magna found; it is
in units not directly comparable to the Arizona and Montana data. )
F-12
-------
TABLE F-ll
ENVIRONMENTAL LEAD DATA
• " ^ a
Arizona (1969 data, yg/nr)
Smelter Towns Non-Smelter Towns
A jo .1 Davis Dam , .1
Claypool 1.2 Florence .5
Douglas .5 Organ Pipe .1
E. Plantsite 2.2 Page .04
Hayden 3.2 , , ,« ^ -
r. « i n* Average ± o .18 ± .2
San Manuel .04 *
Average - a 1.2 * 1.2
Anaconda (1961-1962 data, pg/m3)
Min. Avc
.2 .70
.1 .53
0 .48
.03 .46
0 .54
>' Magna (Nov. 1970 - Feb. 1972 data, mg/m2/mo.)C
Nov. '70 22.98 May - Nov. 2.41
Period
1961-1962
1962
1962
1962
1961-1962
Dec. -Feb.
Mar. -May
June-Aug .
Sep.r-Nov.
Year
No. of Samples Max,
30
31
32
30
123
1.7
1.7
2.1
3.2
3.2
Dec.
Jan. '71
Feb.
March
April
1.84
3.21
1.95
7.75
6.08
June
July
Aug.
Sept.
Oct.
4.96
1.26
5.46
4.02
4.82
Dec.
Jan. '72
Feb.
Average ±
2.02
2.30
2.53
a =3.89
Comment: There are no environmental standards for lead. The National
Air Sampling Network reported a range of .002 to .15 pg/m of lead in
non-urban areas. The highest in the survey was 2.7 ug/im in Minneapo-
lis.
a
Arizona Department of Health Services
b
Montana Air Pollution Bureau
c
CHESS study (This data is the only lead data for Magna found. It
is in units which are not directly comparable to the Arizona
and Montana data.)
F-13
-------
TABLE F-12
ENVIRONMENTAL ARSENIC DATA
Arizona (1969 data, yg/m3)
•Smelter Towns
Non-Smelter Towns
Ajo
Claypool
Douglas
E. Plantsite
Hayden
San Manuel
.003
.01
.004
.007
.006
.009
Davis Dam
Florence
Organ Pipe
Page
Average ± o
.001
.01
.005
.001
.0042 ± .0042
Average ± a .0065 ± .0027
Anaconda (1961-1962 data,
Period
1961-1962
1962
1962
1962
1961-1962
>d
Dec. -Feb.
Mar. -May
June-Aug.
Sep. -Nov.
Year
No. of Samples
30
31
32
30
123
Max.
2.5
1
1.9
1.8
2.5
Min.
d
Soil Analysis : November 1977, 33 ppm
Avg.
2.5
1
1.9
1.8
2.5
0
0
0
0
0
.54
.3
.52
.5
.45
Comment; There is no ambient standard for arsenic. However, a
study by the Montana State Board of Health found that the value
for Anaconda was higher than any other city in the state as well
as New York and Chicago. Normal soil levels are about 8 or 9 ppm.
No data available on Magna.
Arizona Department of Health Services.
**
'Montana Air Pollution Bureau.
Sample taken by Enviro Control site visit team, November 1977.
F-14
-------
TABLE F-13
FLUORIDE CONTENT OF RANDOM GRASS SAMPLES
(Deer Lodge Valley)
Location
Anaconda (west)
Anaconda (north)
Opportunity (east)
/
Anaconda (east)
Deer Lodge (east)
Galen (west)
Deer Lodge (west)
Galen (east)
Date of Concentration Date of Concentration
Collection in ppm Collection in ppm
7/15/65
11/3/65
11/3/65
11/3/65
8/19/65
8/19/65
8/11/65
8/19/65
12.6
3.7
5.0
3.7
2.2
2.0
2.0
3.2
6/30/66 4
6/30/66 2
Source: Montana Air Pollution Bureau
F-15
-------
TABLE F-14
BODY BURDEN DATA
Heavy Metal Levels in Hair. Blood and l'rinej
- Mean 1 s. E. (number of samples) -
Location
S s
& *
IH
Carlson
owns
U
Ajo
Anaconda
Douglas
Hayden
Miami
Horenci
San Manuel
Albuquerque
PerrYvilleMO
Safford AZ
Hair Lead
-------
TABLE F-15
AGE-ADJUSTED MORTALITY RATES FOR WHITE FEMALES
(per 100,000) FOR ANACONDA AND CONTROLS
CAUSE OF DEATH (ECI #)
Infectious fi parasitic diseases (1)
Cancer, oral, pharynx & esophagus (2)
Cancer, stomach (3)
Cancer, duodenum S small intestine (4)
Cancer, colon & rectum (5)
Cancer, liver & bile ducts (6)
Cancer, pancreas (7)
Other digestive cancers (8)
Cancer, respiratory organs (9)
Cancer, genitourinary organs except
bladder (10)
Cancer of the bladder (11)
Other cancers except leukemia (12)
Neoplasms of lymphatics (13)
Endocrine, nutritional, metabolic
& blood disease (14)
Mental & nervous system diseases (15)
Rheumatic heart disease (16)
Hypertensive heart disease (17)
Cardiovascular diseases (18)
Cerebrovascular diseases (19)
Respiratory diseases (20)
Bronchitis, emphysema S asthma (21)
Pneumoconiosis (22).
Other digestive diseases (23)
Diseases of stomach s duodenum (24)
Appendicitis, hernia £ others (25)
Gastroenteritis, noninfectious (26)
Diseases of liver, gallbladder £
pancreas (27)
Diseases of urinary system (28)
Diseases of the genital organs (29)
Diseases of pregnancy S child-
birth (30) . .
Diseases of the skin s musculoskeleton
£ ill-defined diseases (31)
Congenital anomalies (32)
Perinatal diseases (33)
Accidents, poisons £ violence (34)
AGE-ADJUSTED
Anaconda
0
0
6
0
56
0
39
0
27
45
0
98
20
24
18
29
0
443
133
45
19
0
8
10
24
0
56
11
0
0
21
0
0
25
MORTALITY
RATES
Controls
Great Falls
4
11
7
0
19
8
21
0
30
45
1
100
21
34
20
15
8
308
86
23
19
0
0
3
9
0
35
10
1
0
66
3
0
53
Havre
0
0
8
0
14
0
21
0
27
30
0
84
34
7
7
20
7
253
55
34
0
0
0
0
14
0
13
7
0
0
20
0
0
43
Billings
3
e
5
0
21
7
7
0
20
37
1
71
18
21
14
20
8
280
74
31
21
0
0
2
14
0
27
6
1
0
42
3
0
58
Bozeman
9
0
4
0
T4
0
10
0
9
31
4
96
IB
20
3
14
4
228
519
22
10
0
«
-------
TABiE F-16
AGE-ADJUSTED MORTALITY RATES FOR WHITE FEMALES
(per 100,000) FOR ARIZONA SMELTER TOWNS AND CONTROLS
CAUSE OP DEATH (ECI 1)
Infectious s parasitic diseases (1)
Cancer, oral, pharynx S esophagus (2)
Cancer, stomach (3)
Cancer, duodenum & small intestine (4)
Cancer, colon & rectum (S)
Cancer, liver S bile ducts (6) .
Cancer, pancreas (7)
Other digestive cancers (8)
Cancer, respiratory organs (9)
Cancer, genitourinary organs except
bladder (10)
Cancer of the bladder (11)
Other cancers except leukemia (12)
Neoplasms of 'lymphatics (13)
Endocrine, nutritional, metabolic
& blood disease "(14)
Mental & nervous system diseases (15)
Rheumatic heart disease (16)
Hypertensive disease (17)
Cardiovascular diseases (18)
Cerebrovascular diseases (19)
Respiratory diseases (20)
Bronchitis, emphysema & asthma (21)
Pneumoconiosis (22)
Other digestive diseases (23)
Diseases of stomach & duodenum (24)
Appendicitis, hernia S others (25)
Gastroenteritis, noninfectious (26)
Diseases of liver, gallbladder s
pancreas (27)
Diseases of urinary system (28)
Diseases of the genital organs (29)
Diseases of pregnancy & child-
birth (30)
Diseases of the skin £ musculoskeleton
£ ill-defined diseases (31)
Congenital anomalies . (32)
Perinatal diseases (33)
Accidents, poisons fi violence (34)
AGE-ADJUSTED MORTALITY RATES
Arizona Smelter Towns
Miami
0
0
0
0
0
0
0
1 °
23
56
0
131
23
17
0
62
18
910
0
57
18
0
0
0
36
0
62
18
0
0
17
0
0
143
Hayden
0
0
0
0
0
0
0
0
0
0
0
102
41
102
0
0
0
529
0
188
0
0
0
0
0
0
85
0
0
0
84
0
0
53
San Manuel
0
0
0
0
0
0
0
0
72
0
0
95
0
0
0
36
0
314
0
0
0
0
0
0
0
0
0
0
0
0
15
0
0
35
Morenci
0
0
0
0
0
157
0
0
12
43
0
86
46
78
0
0
0
487
187
31
0
0
0
0
0
0
25
31
0
0
16
0
0
94
Douglas
11
5
15
0
33
6
S
0
IS
63
0
110
20
57
17
9
10
382
79
60
19
0
0
5
29
0
67
15
0
0
61
0
0
61
Ajo
13
0
12
16
0
0
0
13
26
58
0
49
30
29
0
0
16
239
62
69
16
0
0
0
12
0
24
0
0
0
0
11
0
38
Controls
75
33
7
0
29
2
12
0
33
31
0
35
17
19
10
12
1
323
77
10
0
0
0
6
12
0
33
9
1
0
18
2
0
45
F-18
-------
TABLE F-17
AGE-ADJUSTED MORTALITY RATES FOR WHITE FEMALES
(per 100,000) FOR MAGNA AND CONTROLS
CAUSES OF DEATH (ECI «)
Infectious s parasitic diseases (1)
Cancer, oral, pharynx & esophagus (2)
Cancer, stomach (3)
Cancer, duodenum s small intestine (4)
Cancer, colon & rectum (S)
Cancer, liver & bile ducts (6)
Cancer, pancreas (7)
Other digestive cancers (6)
Cancer, respiratory organ* (9)
Cancer, genitourinary organs except
bladder (10)
Cancer of the bladder (11)
Other cancers except leukemia (12)
Neoplasms of lymphatics (13)
Endocrine,, nutritional, metabolic
£ blood disease (14)
Mental s nervous system diseases (15)
Rheumatic heart disease (16)
Hypertensive disease (17)
Cardiovascular diseases (18)
Cerebrovascular diseases (19)
Respiratory diseases (20)
Bronchitis, emphysema C asthma (21)
Pneumoconiosis (22)
Other digestive diseases (23)
Diseases, of stomach s duodenum (24)
Appendicitis, hernia t others (25)
Gastroenteritis, noninfectious (26)
Diseases of liver, gallbladder &
pancreas (27)
Diseases of urinary system (28)
Diseases of the genital organs (29)
Diseases of pregnancy 6 child-
birth (30)
Diseases of the skin S musculoskeleton
s ill-defined diseases 131)
Congenital abnormalities (32)
Perinatal diseases (33)
Accidents, poisons c violence (34)
AGE-ADJUSTED
Hagna
5
9
0
0
33
0
0
0
0
54
0
131
15
77
21
25
9
379
70
11
0
0
0
20
15
0
33
20
0
0
57
10
0
43
MORTALITY
RATES
Controls
Logan ,
0
0
7
0
13
5
10
0
12
9
0
63
12
24
16
42
10
206
70
23
4
0
0
2
11
0
21
10
0
0
11
0
0
41
Price Cedar City
9
0
0
0
20
6
23
0
16
12
0
98
15
26
26
35
26
265
52
38
13
0
0
0
6
0
29
0
0
0
14
0
0
50
15
0
0
0
30
0
7
0
7
29
0
75
23
59
23
44
15
322
60
22
1)
0
0
1!>
0
0
15
7
0
0
22
8
0
45
F-19
-------
TABLE F-18
1976 SULFUR DIOXIDE DATA SUMMARY (in
Site
Smelter Towns i
AJO*
AjO
AjO
AjO
AjO
Average ±o
Douglas
Douglas
i •
Douglas
Douglas
Douglas
i
Douglas
Douglas
Douglas
Douglas
r
Douglas
Average ±o
i
Hayden
Hayden
Hayden
Hayden
Hayden
Hayden
Average to
Miami
Miami
Average ±a
Morenci
Morenci
Morenei
Morenci
Morenei
Morenci
L
Morenei
Morenei
Morenci
Morenci
Average ±0
Location
Town Square, approx. 0.5 mi, N54*W from smelter stack
Oxidation Pond, approx. 1.2 mi, N8aE from smelter stack
S. Tailings Dam, approx. 1.3 mi, N85°E from smelter stack
Caoelback Mtn., approx. 1.2 mi, N88°E from smelter stack
AZ DOT Well Rd.
Northwest, approx. 1.22 mi, N30°W of smelter stack
Pirtleville, 1.3 mi, N22.5°E of smelter stack
2 km NNW of stack
U.S. 666, 2 km NNE of stack
North, 0.75 mi, N15°E of smelter stack
Curtis, approx. 2.65 mi, N32°W of smelter stack
Fairgrounds, approx. 3 mi, N55'E of smelter stack
Fir, approx. 14 mi, N47°E of smelter stack
1012 G. Avenue (Mercantile)
Martin, approx. 4.2 mi, N20°E of smelter stack
Approx. 1 mi North of the junction of Highway 77 and 177
Hayden Fire Station, 640 Canyon Drive
Hayden Jet. - Highway 177
Montgomery Ranch
Montgomery Ranch
Old Town Jail - Canyon Drive
Jones Ranch
Lower Miami School
Cadillac Point - 5500, 1.75 mi, N51°W of smelter stack
Clifton Fina, 1.75 mi, S69*E of smelter stack
East Plantsite, 2.5 mi, S35«E of smelter stack
Metcalf, approx. 3.9 mi, N28*W from smelter stack
Oroville, approx. 2.75 mi, N48'E of smelter stack
Star go, 1.5 mi, N63°W of smelter stack
Eagle Creek Road, approx. 4.75 mi, S84TW of smelter stack
Fairbanks, approx. 2.1 mi, S26*E of smelter stack
Stargo, Highway 666
U.S. 666, 4.5 km NH of stack
Operator
PD
PD
PD
PD
State
PD
PD
State
State
PD
PD
. PD
PD
PD
PD
JCC
JCC
JCC
State
JCC
State
State
State
PD
PD
PD
PD
PD
PD
PD
PD
State
State
Method
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
Coul
1976
Annual
Average
12
33
11
7
21
16 ±10
51
43
43
39
39
31
17
33
31
14
34 ±11
135
105
57
127
183
113
103 ±29
56
21
38 ±24
105
34
7
83
9
109
16
13
130
108
49 ±52
Source/ 1976 Air Quality Data for Arizona, reported by the
Arizona Department of Health Services
(continued)
F-20
-------
TABLE F-18 (continued)
Site
Smelter Towns
San Manuel
San Manuel
San Manuel
San Manuel
San Manuel
Average ±O
Location
(continued)
Upper Shopping Center, 2 tan SSW of smelter stacks
Golf Course, 3 k» WNW of smelter stacks
East Peppersauce Wash, 5.8 km ESE of smelter stacks
IDS Church
Mine site, 10.5 km NW of smelter stacks
Operator
Magma
Magma
Magma
State
Magma
Method
Coul
Coul
Coul
Coul
Coul
1976
Annual
Average
45
75
17
42
42
44 ±20
Nonsmelter Cities
Phoenix
Phoenix
Phoenix
Phoenix
Phoenix
Average to
Tucson
Tucson
Tucson
Tucson
Tucson
Tucson
Tucson
Average ±o
1845 E. Roosevelt
241 N. Central, Valley Bank Center, 29th Floor
IS E. Monroe, Valley Bank Annex, 3rd Floor
1740 W. Adams
1845 E. Roosevelt
32 N. Stone, Home Federal Tower, 6th Floor
151 W. Contress
Florence Hwy/McGee Rd.
Sahuaro Monument
1721 N. Tanque Verde Loop
22nd/Craycroft
University of Arizona
Maricopa
PD
PO
State
State
PD
Pima
Pima
Pima
Pima
Pima
State
Coul
Coul
Coul
Bubbler
Bubbler
Coul
Flame
Coul
Coul
Coul
Coul
Bubbler
7
8
8
7
7
7 ±.5
7
7
8
4
6
5
6_
5 ±2
F-21
-------
Appendix G
SELECTED DISEASE CATEGORIES AND INCLUDED ICDA CODES
FOR COPPER SMELTING CASE STUDY
G>
-------
Table G-l
SELECTED DISEASE CATEGORIES AND INCLUDED ICDA CODES
ECI
Disease
Number
3
7
10
Disease Group
Cancer of colon and rectum
Cancer of pancreas
Cancer of genitourinary organs
ICDA Codes
(8th Rev.)
153,154
157
180
Cancer of colon and rectum
Cancer of pancreas
Cancer of cervix uteri
except bladder
12
181
182
183
184
185
186
187
Other cancers except leukemia 170-174
190-199
210-239
14
18
Endocrine, nutritional,
metabolic, and blood diseases
Vascular diseases
19
20
Cerebrovascular diseases
27
Diseases of liver, gallbladder
and pancreas
31
240-289
410-458
430-438
Acute respiratory diseases , 460-486
500-508
. . 510-519
Disease of skin and musculo-
skeletal system and ill-defined
diseases
570
571
571.0
572
573
574
576
577
680-739
780-796
Chorionepithelioma
Other malignant neoplasms of uterus
Cancer of ovary, fallopian tube and broad ligament
Cancer of other and unspecified female genital organs
Cancer of prostate
Cancer of testis
Cancer of other and unspecified male genital organs
Malignant neoplasm of bone
Cancer of other and unspecified sites (eye, brain,
thyroid, and other endocrine gland. Lymph nodfis in
head, face and neck, secondary neoplasms)
Benign neoplasms of digestive and respiratory systems,
bone, skin, connective tissue, musculature, gonito-
urinary organs, nervous and endocrine systems
Diseases of thyroid gland; other endocrine glands
(pituitary, adrenal, gonadal dysfunction)) avitaminoses
and other nutritional deficiencyi other metabolic
diseases (congenital disorders of amino-acid, carbo-
hydrate and lipid metabolism); other diseases such as
plasma protein, porphyrin, etc., disorders) diseases
of blood and blood-forming organs
Ischemic heart disease (myocardial infarction, chronic
ischemic heart disease with and without hypertensive
disease)) diseases of pericardium, endocardium, and
myocardium) pulmonary heart disease; congestive
heart failure) disorders of cardiac rhythm) corebro-
vascular diseases (including cerebral hemorrhage,
embolism and ischemia)) arteriosclerosis; phlobitis
Hemorrhage,
ischemia
with and without hypertension; thrombosis;
Acute'nasopharyngitis, acute bronchitis and broncnioli-
tis> influenza) pneumonia
Chronic pharyngitis, sinusitis) larynx disorders
Pleurisy; pneumothorax) pulmonary congestion and hypo-
staslsi pneumoconioeisi interstitial pneumonii;
bronchiectasis; pulmonary collapse; edema of the
lung; emphysema
Acute and subacute necrosis of liver
Cirrhosis of liver
Alcoholic cirrhosis of.liver
Suppurative hepatitis and liver abscess
Other diseases of liver
Cholelithiasis
Other diseases of gallbladder and biliary ducts
Diseases of pancreas
Boils and carbuncles, cellulitis, acute lymphadenitis,
skin infections; other inflammatory conditions of
skin and subcutaneous tissue; other diseases of skin
and subcutaneous tissue (disease of sweat and sebaceous
glands)
Arthritis and rheumatism, other bone and joint diseases
(osteomyelitis, etc.); diseases of the musculoskeletal
system (nyosltis, myastheoia gravis); connective tissue
disorders* congenital disorders of the nervous system
(spina bifida, hydrocephalus); circulatory, respiratory,
gastrointestinal, urogenital, chromosomal abnormalities)
perinatal diseases) disorders of pregnancy) birth
injuries; anorexia; prematurity
G-l
-------
Appendix H
DISEASE ASSOCIATIONS IN AMERICAN STEEL MINING COMMUNITIES
-------
Appendix H
DISEASE ASSOCIATIONS IN AMERICAN STEEL MILL COMMUNITIES*
There is little direct information on disease risk among the resi-
ents of communities that are economically dominated by steel mills.
What information exists is concerned almost exclusively with cancer mor-
tality. One must therefore extrapolate from the known emission and ef-
fluent products of this industry and from occupational data.
CHARACTERISTICS OF STEEL COMMUNITIES
In general, steel cities are less stereotyped in their cultural and
ethnic profile than coal mining communities. They may be discreet com-
munities such as Bethlehem, Pennsylvania; metropolitan areas such as
Gary, Indiana; or districts of a much larger metropolis such as the
Sparrow's Point/Dundalk area of Baltimore, Maryland. They often have
large ethnic minorities; this vastly complicates interpretation of cer-
''<£ , tain diseases rates, such as cancer of the stomach, which are thought
to be strongly influenced by dietary or cultural factors. Steel com-
munities are often on the receiving end of large population migrations,
especially from Appalachia and the deep South, so that a subset of the
population may reflect the risk profile of a distant, very different
locale. In short, steel communities are not as simple in their demo-
graphic and epidemiologic characteristics as other industrial cities may
be.
f
HAZARDS ASSOCIATED WITH STEEL MILL OPERATION
The hazards associated with blast furnace operation (as outlined in
Volume II of the Plan for Investigation) require some medical comment:
t
© Gaseous Emissions from the Blast Furnace—Although oxides
of nitrogen are undoubtedly present as a consequence of the
high temperatures achieved in blast furnaces, none of the
communities cited by the Environmental Protection Agency
as exceeding ambient air quality criteria for NOX in the
period 1971-1973 were steel cities except for Baltimore,
tr ' where there is also a large concentration of other indus-
tries. On the other hand, it is quite plausible that
oxides of sulfur are significant as a health hazard in
! these communities, with the associated risk of chronic
j bronchitis and chronic obstructive pulmonary disease.
*This report was prepared for Enviro Control, Inc., by Tee L. Guidotti,
M.D., medical consultant on this project.
H-l
-------
• Particulate Matter—It is not entirely clear what agents
may exist in the ash of blast furnaces. The ash of coke
ovens, however, is of great concern due to its content
of potential carcinogens; there may be 1% benzo(a)pyrene
(BaP) in such particles.
• Leachates Into the Water Table—These may carry a variety
of metal contaminants, including arsenic, lead, and cad-
mium. Water used for cooling purposes, on the other hand, .
is likely to be fairly free of toxic metals.
• Gaseous Emissions from Metal Processing; Flue Dust Fines
after Cyclone Separation; Metal Fines from Furnace
Processes—This group is much more likely to pose an
occupational threat to workers than to be a risk to the
outside population.
These hazards are unique to the blast furnace. Hazards associated
w£th other basic steel processes include:
• Coke Oven Fumes—As will be shown, these constitute a major
occupational hazard. There is as yet no clear evidence
that they are a risk beyond the plant gates, but this
probably depends on the local siting of the coke ovens.
These emissions are prime candidates for implication in
any community cancer excess.
• Silicates from Firebrick—These particles from crubling
firebrick may aerosolize in the respiratory range and are
the cause of silicosis in workers at the blast furnace.
They are unlikely to be implicated in community risk.
• Asbestos—This is used in heat shields and in various uses
around the plant. Its ubiquity makes it an important facr-
tor in assessing cancer risk. Furthermore, it is already
demonstrated to be significant in conjugal and community
exposure.
• Heat—The intense heat generated by the blast furnace may
be more than an occupational hazard. It may be significant
in atmospheric reactions which change the composition of
air pollution (analogous to its role in producing the
oxides of nitrogen). Furthermore, it may volatilize metal
or other normally particulate compounds and render them
airborne.
• Machine Oils—There is a well-documented excess of lung and
oropharyngeal cancer among machinists, and there is emerging
evidence that the agent may be the fine petroleum oils with
which the piece and the machinery are lubricated. This may
be a factor in finishing operations; it is unlikely to be
significant in basic steel production.
H-2
-------
DISEASES OF STEELWORKERS
By and large, steelworkers tend to be healthy and have overall
cause-specific mortality rates substantially lower than the general
population. Their causes of death are parallel to those of the general
population, except that they have a disproportionate rate of accidents.
A large and elegant body of work has been accumulated on cancer mor-
tality among steelworkers. It is now well established that workers in
the coke ovens, who constitute a small fraction of steelworkers as a
whole, are at enormously high risk for lung cancer. This association
is exposure- and cumulative-dose-related. There is a lesser excess of
cancer of the rectum. There are highly suggestive, but not definitive,
associations for cancer of the kidney (hypernephroma—an uncommon but
not rare tumor) and bladder, and a possible association with cancer of
the pancreas, in workers associated with other, mostly finishing pro-
cesses. This pattern of organ involvement strongly suggests a carcino-
gen which (1) is inhaled, (2) is excreted by the kidney and concentrated
in urine, and (3) may be incidentally ingested as well. Thus, there may
be more to an excess risk of cancer among steelworkers than lung cancer
in coke oven workers—and it should still be emphasized that steel-
workers in general have a lower mortality from cancer at all sites than
the general population.
Metalworkers who work with steel, but who are not necessarily steel-
workers, show an interesting pattern of cancer mortality in Britain.
They have a significant excess mortality from lung cancer and question-
able excesses for thyroid and bone cancers. The latter two are rela-
tively unusual in the general population, and one wonders whether there
might be a carcinogenic agent which is handled metabolically in much
the same way as either iodine or calcium. Since the mortality rates for
cancers at all sites (aggregated) are lower among steelworkers than in
the population at large, rates for cancers at specific sites which do
not achieve statistical significance for excess when compared to the
general population may be strikingly more significant if compared to an
occupational control cohort.
Heart disease as a cause of death was seen in excess only in certain
finishing operations and among janitors and general laborers (whites
only—blacks did not share this excess risk). Nonetheless, maintenance
workers, who also sustain the intermittent, variable exposure which is
similar to that suffered by the community at large, had a remarkably low
mortality from heart disease. Cerebrosvascular disease and stroke were
elevated to significance in one finishing operation, but not in basic
steel processes. An exhaustive study of cardiovascular disease and myo-
cardial infarction among British steelworkers operating the blast fur-
naces could document no excess risk, despite elevated blood levels of
carbon monoxide.
H-3
-------
DISEASE PATTERNS IN STEEL COMMUNITIES
There is little information on disease incidence in steel communi-
ties beyond cancer mortality.
The wives of boilermakers, an occupation which involves steel fabri-
cating but not exposure to machine oils, have a slightly increased mor-
tality from cancer at all sites (aggregate) in Britain and, specifically,
elevated mortality from lung cancer. Although the standardized mortality
ratio is not striking (132), there may be a factor at work in conjugal
exposure or in overall elevated community risk.
County cancer mortality for those parts of the country that have
concentrations of steel mills disproportionate to other concentrations
of industry, such as Pittsburgh, Birmingham, and Gary, show elevated
rates at all sites (aggregate) for both males and females, further sug-
gesting a community risk. Cancer of the stomach and of the lung and,
interestingly, of the rectum appear at higher risk in both sexes in
Pittsburgh (Allegheny County). Cancer of the oropharynx (mouth and
throat), larynx, and esophagus—all uncommon and very likely related to
environmental exposure—are at excess in men only (corrected for the
known male predominance of these tumors anyway). This suggests but can-
not prove a local occupational exposure. As pointed out in Volume II of
the Plan for Investigation, cancer rates in general were found to be in
excess for residents of a Canadian steel city, Hamilton, and attempts
have been made to associate this with fluoride emissions. On a biologi^-
cal basis, this is not particularly convincing, although it is possible
that fluorides are a marker for some other biologically active pollutant
which is released in a similar manner.
The following, therefore, emerge as the diseases most likely to be
found in excess in steel communities:
Cancer of the lung
Cancer of the stomach
Cancer of the rectum
Cancer of the pancreas
Cancer of the bladder
Cancer of the kidney (hypernephroma)
Cancer of the thyroid (it would be useful to identify the
tissue type)
Chronic bronchitis and chronic obstructive pulmonary disease
Upper and lower respiratory tract infections (unlikely to be
a major cause of death, however)
• Cancer of the head and neck
There is at present no evidence for an excess of cardiovascular dis-
ease, although one cannot truly generalize from basically health steel-
workers to the effect on an entire population.
H-4'
-------
Several important disease processes are overlooked in current stud-
ies. The opportunity exists to complete this omission by supplying the
missing elements in the profile of steel community health. These in-
clude :
• Diseases of the liver and biliary tract (bearing in mind
that high rates of alcqholic cirrhosis may exist)
• Congenital malformations
• Spontaneous abortions and miscarriages
SOURCES OF INFORMATION
Adelstein, A. M. Occupational mortality: Cancer. Ann. Occup. Hyg. 15:
53-57,. 197 2.
Cecilioni, V. A. Lung cancer in a steel city, its possible relation to
fluoride emissions. Fluoride 5:172-181, 1972.
Foehr, R. Evaluation bioclimatique d'un poste de fondeur de haut-
fourneau. Arch. Mai. Prof. 30:144-146, 1969.
Jones, J. ,G., and A. Sinclair. Arterial disease amongst blast furnace
workers. Ann. Occup. Hyg. 18:15-20, 1975.
Lloyd, J. W., and A. Ciocco. Long-term mortality study of steelworkers.
I. Methodology. J. Occup. Med. 11:299-310, 1969.
Lloyd, J. W., et al. Long-term mortality study of steelworkers. IV.
Mortality by work area. J. Occup. Med. 22:151-157, 1970.
Lloyd, J. W. Long-term mortality study of steelworkers. V. Respiratory
cancer in coke plant workers. J. Occup. Med. 23:53-68, 1971.
Mason, T. J., et al. Atlas of Cancer Mortality for U.S. Counties: 1950-
1969, DHEW Publ. No. 75-780. U.S. Government Printing Office, Washing-
ton, D.C;, 1975.
Radford, E. P. Cancer mortality in the steel industry. Paper presented
to the Conference on Occupational Carcinogenesis, New York Academy of
Sciences, New York, 27 March 1975.
Redmond, C. K., B. R. Strobino, and R. H. Cypess. Cancer experience
among coke by-product workers. Ann. N.Y. Acad. Sci. 272:102-115, 1975.
Redmond, C. K., et al. Long-term mortality study of steelworkers. VI.
Mortality from malignant neoplasms among coke oven workers. J. Occup.
Med. 24:621-629, 1972.
Tokudome, S., and M. Kuratsune. A cohort study on mortality from cancer
and other causes among workers at a metal refinery. Int. J. Cancer 17:
310-317, 1976.
U.S. Department of Labor, Bureau of Labor Statistics. Occupational
Injuries and Illnesses in the United States, by Industry, 1973. U.S.
Government Printing Office, Washington, D.C., 1975.
H-5
-------
Appendix I
AERIAL PHOTO ANALYSIS PERFORMED BY
ENVIRONMENTAL PHOTOGRAPHIC INTERPRETATION COMPLEX (EPIC)
-------
Appendix I
AERIAL PHOTO ANALYSIS PERFORMED BY
ENVIRONMENTAL PHOTOGRAPHIC INTERPRETATION COMPLEX (EPIC)
Methodology for Johnstown and Lewi8town. Pennsylvania
Air Photo Analyses
EPIC's primary tasks performed for this project included (1) the identification
of changes in residential patterns in the Johnstown area from 1939 to 1972;
(2) the identification and location of significant point and non-point sources
of pollution in the area; (3) the analysis of the development: of industry in
the area from 1939 to the present. In addition, an estimate was made (for
Johnstown only) of the area containing the highest ground level concentration
j of air pollution (principally sulfur dioxide and partlculate matter) based
i upon a simple "box" diffusion model.
• Aerial photography was used as (he data source for most of the required work,
i and black and white photography dating back to 1939, at a scale of 1:24,000,
was utilized in the interpretation and analysis process. Some recent (1977)
color aerial coverage was also used in analyzing industrial activity in the
i area. For the concentration estimates, emissions and meteorological data
| were obtained from Enviro Control and EPA/Research Triangle Park, North Caro-
lina. Additional data on diffusion estimates for the Johnstown area is yet
to be received from TRP.
The finished product consisted of several 7.5 minute USGS map sheets with
i acetate overlays showing residential changes, point and non-point sources,
I and the estimated area of maximum ground level concentration. A written
i text was prepared describing industrial activity in the area from 1939 to the
{ present with respect to changes in size, new construction, and addition of
i air pollution control equipment. In addition, a photo-mosaic with an overlay
| depicting the principal industries was also prepared.
An almost identical study was also done for the Lewistown, Pennsylvania area.
In this instance, however, an air pollution concentration and area estimation
1 was not undertaken, and the only aerial photography used was from the years
1938 to 1964. Table 1-1 delineates the aerial photo coverage and the over-
lays prepared for Johnstown and Lewistown, Pennsylvania.
1-1
-------
Table 1-1
'DATES OF AERIAL PHOTO COVERAGE FOR JOHNSTOWN AND LEWISTOttN, PENNSYLVANIA
GEOGRAPHICAL
AREA
8'
<->
8
o
•I
Lewistpwn
Burnham
Johnstown
Ge is town
Hoover svijle
Windber
Nanty Glo
Vintondale
DATE
Topography
Map
1966
1966
1964
1964
1971
1971
1964
1964
Overlay
#1
1938
1938
1938
1938
1938
1938
1938
1938
#2
1964
1964
1962
, 1962
1962
1962
1962
1962
#3
—
1972
1972
1972
1972
1972
Photo-
Revision
1973
.
1972
—
1972
1972
1-2
-------
Industrial Activity and Development in the
Johnstown, Pennsylvania Area
(see Figure 1-1)
Site A
This medium-sized industrial plant is engaged in heavy fabrication, probably
structural steel, arid was built between 1939 and 1955. It is very probably
associated with Bethlehem Steel or U.S.Steel. There is only one medium-
sized .stack associated with the plant, and no plume is visible from it. The
plant has only expanded once, by about 30%, between 1972 and 1977. No air
pollution control equipment is presently attached to the stack.
Site B
This site is actually composed of two brick kiln complexes located directly
across the river from each other. Both of these were built prior to 1939.
One kiln had been removed from the west bank complex between 1939 and 1955.
Between 1939 and 1955, three kilns had been added, and other facilities ex-
panded at the east bank complex. All of the kilns on the west bank had been
removed between 1962 and 1972, and all operations had been terminated. The
complex on the east bank appears to have gone out of operation during this
time period also, and all of the facilities were physically removed between
1972 and 1977.
Site C
This large fabricating (probably structural steel) plant is associated with
Bethlehem Steel, and was built before 1939. There are two small stacks, and
several very small stacks or "chimneys" on the roof.which have been in place
since before 1939. No stack plumes, or air pollution control equipment, are
visible on; 1977 aerial photography. The plant was expanded by approximately
20% between 1939 and 1955, and water clarifiers or mixing tanks were installed
between 1972 and 1977. Other than that, no changes have taken place since 1955.
Site D
,— i
Two complexes (Da and Db) comprise this large Bethlehem Steel operation.
Complex Da is engaged in coke production, and complex Db is engaged in coke
and coke by-products prodution, and heavy fabrication (probably structural
steel). In 1939 there were four medium-sized stacks serving Da, and three
medium-sized stacks serving Db. Portions of the fabricating plants in complex
Db were rebuilt between 1955 and 1962, and one medium-sized stack had been
removed. There has been no change in the number of coke ovens at either
1-3
-------
^p^eci.-^.
m^p^Hi".-.:
saiLs*.^^"^^- .•
J '-" ' *••/, V*..- v • —-:•—-^ ^-.rvv»Ti\ -ttf" : r" >,*/ ' x1 . .
' y/'-'i'5?^ ".-"-* •,'{:&•: IT' "- '' fs^- -•
«: ^fr^ ^^n^2%r*-
^l;'/'^, ;^4(vR^A--V]
- • . /^ / • • i~ I •- \ i • *\ ••" *»*>•. \^ -J
_J^ 'JLLl— * \ m -1 • 1- V ' A'V^f *>^
1-4
-------
complex since 1939. Another medium-sized stack was removed from Db between
1962 and 1972. Two medium-sized stacks were removed from, and two small
stacks were added to the Da complex between 1962 and 1972. No stack plumes
or air pollution control equipment were visible on 1977 aerial photography.
Site E
This large Bethlehem Steel complex appears to be engaged in structural steel
fabrication, and was built prior to 1939. Approximately ten small stacks were
visible at the site in 1939. Several of the smaller stacks had been removed
and replaced with two medium-sized stacks between 1962 and 1972,. No other
changes have taken place since 1939, and no stack plumes or air pollution
control equipment were visible on 1977 aerial photography.
Site F
This very large complex is the Franklin Works (iron smelter and steel plants),
which is owned and operated by Bethlehem Steel Corporation. It has many
small and medium-sized stacks, and contributes approximately 80% of the
ambient sulfur dioxide and particulate matter loading in the Johnstown area.
Some open hearth furnaces had been removed, and some blast furnaces had been
added between 1955 and 1962. Some other plant facilities were also removed
between 1962 and 1972. Electrostatic precipitators were installed between
1972 and 1977 to handle many of the dust-producing processes.
Site G
This Bethlehem Steel plant appears to be engaged in structural steel fabrica-
tion. It was built prior to 1939, and has only one small stack. The plant
expanded approximately 10% between 1939 and 1955, and by another 10% between
1962 and 1972.
Site H
This site is a small light industrial plant, with one medium-sized stack,
built before 1939. It was replaced by an office building between 1962 and 1972.
Site I
This stack is associated with a small power plant for a junior high school.
Site J
U.S.Steel owns and operates this heavy fabrication/structural steel complex.
It was built prior to 1939, and has approximately 6 or 7 small stacks. One
plant in the complex was rebuilt between 1962 and 1972. Three of the taller
1-5
-------
stacks appear to have some type of pollution control equipment installed
on them. •
Site K
The stack is for the heating plant connected to a complex of greenhouses.
Site L
The stack is also for a heating plant connected to a complex of greenhouses.
Site M
The medium-sized stack depicted here is for a small power plant associated
with the State Rehabilitation Center.
Site N
The stack annotated here is medium-sized, and associated with a small
industrial power plant.
1-6
-------
Johnstown, Pennsylvania Concentration Estimates
o Must assume uniform distribution in the valley (area delineated on map;
see Figure 1-1)
o Average valley width in the delineated area is approx. 12,000 ft or 3658 m
o Mixing depth is approximately 550 ft or 168 m
Pollutant Emissions:
Normal Peak
Total Sulfur Dioxide 423 g/sec 667 g/sec
Total Particulate Matter 212 g/sec 217 g/sec
u D' vy
= Ground Level Concentration
Q = Total Emissions
u = Average Wind Speed
D = Mixing Depth
y = Average Valley Width
4.23 x 10
SO2 Concentration Estimates
8 ug/sec / 1 \ I I \
\ 3658 m/ U68 m'
6 m/sec
(7.05-x 107)(2.7 x 104) (6.0 x 103)
11.4 x 10 or 114 ug/m (normal)
6.67 x 108
6
/ 1 ] / 1 }
A 3658 / U68/
180 ug/m3 (peak)
Particulate Matter Estimates
57 (normal)
•y '
(.= 58.
2 H9/m" (peak)
1-7
-------
Appendix J
MORTALITY RATE COMPUTATION METHODOLOGY
and
ECr, SSI AND ICDA CAUSE OF DEATH CODE CONVERSION TABLES
T
-------
Appendix J
MORTALITY RATE COMPUTATION METHODOLOGY
AND
EC I, SSI AND ICDA CAUSE OF DEATH CODE CONVERSION TABLES
FOR SITES OF INTEREST
1. Print the number of deaths, populations, age-specific mortality
1 rates, and age-adjusted mortality rates for these places as follows:
a. White males; white females; nonwhite males; nonwhite females
b. Age groups:
0- 2
3-14
15-24
25-34
35-44
45-54
55-64
65-74
75 and over
Note; Age-adjusted mortality
rates to be calculated
using data from 35 to
74 year age groups
only.
c. Causes of death:
ECI CAUSE-OF-DEATH CODES
COMPARED TO 8TH REVISION (1968 AND AFTER) ICDA CODES
ECI *
1
' 2
3
4'
5
6
7
B
9
10
11
12
13
14
15
16
17
IS
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
ICDA Code (8th Revision)
000-136
140-150
151
152
153-154
155-156
157
158-159
160-163
180-187,189
188
170-174,190-199,210-239
200-209
240-289
290-380
390-398
400-404
410-458 .
430-438
460-486 , 500-508 , 510-514 , 516-519
490-493'
SIS
520-530
531-537
540-560,562-569
561
570-577
580-599
600-629
630-678
680-739,780-796
740-759'
760-779
800-1000
Cause of Death
Infectious s parasitic diseases
Cancer, oral, pharynx s esophagus
Cancer, stomach
Cancer, duodenum fi small intestine
Cancer, colon fi rectum
Cancer, liver fi bile ducts'
Cancer, pancreas
Other digestive cancers
Cancer, respiratory organs
Cancer, genitourinary organs except bladder
Cancer of the bladder
Other cancer* except leukemia
Neoplasms of lymphatics
Endocrine, nutritional, metabolic fi blocd diseases
Mental c nervous ' system diseases
Rheumatic heart disease
Hypertensive disease
Cardiovascular diseases
Cerebrovascular diseases
Respiratory diseases
Bronchitis, emphysema fi asthma
Pneumoconiosls
Other digestive diseases
Diseases of stomach fi duodenum
Appendicitis, hernia fi others
Gastroenteritis, noninfectious
Diseases of liver, gallbladder S pancreas
Diseases of urinary system
Diseases of genital organs
Diseases of pregnancy fi childbirth
Diseases of skin fi musculoskeleton fi ill-defined diseases
Congenital anomalies
Perinatal diseases
Accidents, poisons fi violence
J-l
-------
ECI CAUSE-OF-DEATH CODES
COMPARED TO 7TH REVISION (BEFORE 1968) ICDA CODES
ECI *
ICDA Code (7th Revision)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
000-138,696-697,764,767-769
140-150
151
152
153-154
155
157
158-159
160-164
171-180
181
156,165,170,190-199,210-219,222-239
200-205
250-299,772
300-326,340-390,765
400-416
441-447
330-334,420-434,450-468
330-334
470-500,510-527,763
240-241,501-502,527
523
530-539
540-545
550-570,572-578
571
580-587
590-609
610-637
640-689
221,242-244,690-695,698-749,766,780-795
220,750-759
760-762,769-776
800-1000
CAUSE-OF-DEATH CODES FOR CANCER REGISTRY DATA
ECI#
C-l
C-2
C-3
04
05 .
C-6
0?
C-8
09
C-10
Oil
012
C-13
014
015
016
017
018
019
C-20
021
ICD #'s
400-509
510-519
520-529
530-539
540-549
550-559
560-579
580-599
600-619
620-629
630-659
690-699
700-739
740-759
790-899
880-889
900-929
930-949
950-959
960-969
990-999
Site of Neoplasm
Oral, pharynx, esophagus
Stomach
Small intestine
Colon
Rectum
Liver
Gallbladder, pancreas
Other digestive
Nasal, larynx
Trachea, bronchus, lung
Other respiratory and intrathoracic organs
Hematopoietic and reticuloendothelial systems
Bone, connective tissue, skin
Breast
Genitourinary
Bladder
Eye, nervous system
Endocrine glands
Ill-defined
Lymph nodes
Unknown
J-2
-------
SSI CAUSE-OF-DEATH CODES COMPARED TO 8TH REVISION ICDA CODES
SSI Coda
ICOA Code (8th Revision)
DiMasa Categories
4
5
6
7
8
9
10
11
12
13
14
IS
16
17
18
19
20
21
22
23
24
25
26
000^999
000-136,290-389,630-679,710-739,760-E999
140-289,390-629,680-709,740-759
140-239
140-149
150-159
151
.152-153 ._
154
157
150,155-156,158-159
160-163
162
160-161,163
174
180-184
185-187
183
189
191-192
204-207
200-203,208,209
170-173,190,193-199,210-239
250 .
240-246,251-279
280-289
All deaths
Infective and parasitic diseases) mental disorders;
diseases of the nervous system and sense organs;
complications of pregnancy, childbirth and the
puarperium; diseases of the nusculoskeletal system
and connective tissue; certain causes of mortality
in early infancy; symptoms and ill-defined condi-
tions; and accidents, poisonings and violence
All other deaths, including neoplasms; endocrine,
nutritional and metabolic diseases} diseases of the
blood and bloodforming organs; diseases of the cir-
culatory system; diseases of. the respiratory system;
diseases of the digestive system; diseases of the
genitourinary system; diseases of the skin and sub-
cutaneous tissue; congenital anomalies
All neoplasms
Buccal cavity and pharynx
Digestive organs and peritoneum
Stomach
Small and large intestines
Rectum
Pancreas
Esophagus, liver, gallbladder, peritoneum and
other digestive organs
Respiratory system
Trachea, bronchus and lung
Other parts of the respiratory system
Breast
Female genital organs
Hale genital organs
Bladder
Other urinary organs
Brain and other parts of the nervous system
Leukemia
Other neoplasms of lymphatic and hematopoietic
tissues
Other malignant neoplasms and benign neoplasms
Diabetes mellitus
Other endocrine, nutritional and metabolic diseases
Diseases of blood and bloodforming organs
(continued)
J-3
-------
SSI CAUSE-OF-DEATH CODES COMPARED TO 8TH REVISION ICDA CODES
(concluded) ;
SSI Code
27
28
29
30
31
32
33
34
35
36
37
,38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
ICDA Code (8th Revision)
390-448
390-398
400-404
410-414
410-411
412-414
420-429
430-438
440
441-448
450-458
460-519
484-485
490-493
490-491
292
493
510-511.513,517
460-483 ,486 , 500-S08 , 512 , 514-516 , 518 , 519
520-577
571 '
571.0
571.8-571.9
520-570,572-577
580-629
580-584
590 .
591-629
680-709
741-759
Disease Categories
Major cardiovascular disease
Chronic rheumatic heart disease and active
rheumatic fever
Hypertensive disease
Ischemia heart disease
Acute myocardial Infarction and other acute or
subacute forms of ischemic heart disease
Chronic ischemic heart disease and angina
pectoris
Other forms of heart disease
Cerebrovascular disease
Arteriosclerosis
Other diseases of arteries, arterioles and capil-
laries
Diseases of veins, lymphatics and other diseases of
circulatory system
Diseases of the respiratory system
Acute interstitial and bronchopneumonia
Bronchitis , emphysema and asthma
Bronchitis
!
Emphysema !
Asthma
Empyema, pleurisy, and other chronic interstitial
pneumonia
Other diseases of the respiratory system
Diseases of the digestive system
Cirrhosis of liver
Alcoholic cirrhosis of liver
Other and unspecified cirrhosis of liver
Other diseases of the digestive system
Diseases of the genitourinary system
Nephritis and nephroais
Infections of kidney
Other diseases of the genitourinary system
Diseases of the skin and subcutaneous tissue
Congenital anomalies
J-4
-------
2. Calculations:
a. Age-adjusted mortality rates using only the 35-74 year age: groups
(see note oh page J-l) for each place or aggregate:
A(t) = £ S(i.t) P(i)
where:
S(i,t ) = the age-specific death rate in age group i
over t years
100.000
x
- the age-adjusted death rate over t years
P(i,t) = the number of deaths in age group i over t years
P(i) = the number in age group i in the U.S. standard
population
t = the number of years observed
i = a particular age group
n » the number of age groups
b. U.S. standard population:
Age Number
35-44
45-54
55-64
65-74
20,327,701
20,846,499
16,811,212
11,281,137
29.35
30.10
24.27
16.29
3. Population Data: 1970 fifth count, file C, census tapes for sites
j of interest.
4. Mortality Data: Death'.certificate tape for sites of interest.
J-5
-------
5. Summary Demographic Statistics
For places of interest, summary demographic statistics from 1970
census tapes as-follows:-.
median age
. .' median income
% in same house in 1965
% white
% 41.00 persons per room
% .finishing high school
% female
% native of native parents
- %,-electricity used 'for heating
J-6
-------
Appendix Kl
ETIOLOGY OF CARDIOVASCULAR DISEASES
ASSOCIATED WITH BITUMINOUS COAL MINING
Appendix K2
ETIOLOGY OF RESPIRATORY DISEASES
ASSOCIATED WITH BITUMINOUS COAL MINING
-------
Appendix Kl
ETIOLOGY OF CARDIOVASCULAR DISEASES
ASSOCIATED WITH BITUMINOUS COAL MINING*
The incidence data for cardiovascular disease in coal mining communi-
ties' is misleading in several respects when one attempts to generalize
risk data to a broader population than miners alone. The greatest source
of error is the intrinsic interrelation of pulmonary and cardiovascular
disease, such that most patients with advanced pulmonary disease will as a
matter of course develop heart disease. Another source of confusion is
the absence of data on uncomplicated atherosclerotic cardiovascular dis-
ease alone in coal mining communities, so that no population model exists
for estimating risk in coal regions in general. Finally, miners are pre-
disposed by their working conditions to develop peripheral venous disease
(thrombophlebitis, varicose veins), and this should not be allowed to com-
plicate consideration of arterial disease.
HEART DISEASE IN COAL MINERS
Although there appears to be a moderate elevation in mortality from
"heart disease" in general and "myocardlal infarction" in particular,
these data must be interpreted very cautiously. An excess mortality from
"heart disease" is a predictable result of a high prevalence of lung dis-
ease, because .the pulmonary and cardiovascular systems are intimately re-
lated. Mechanisms for this interdependence are primarily the following:
a) Progressive destruction of lung tissue also reduces the
cross-sectional area of the pulmonary arterial bed. Re-
duced oxygen tensions also induce vascular constriction in
the pulmonary artery. These phenomena result in reduced
compliance, therefore increased resistance of the pulmonary
arterial bed, which is normally a very low resistance hydro-
dynamic system. This Increased resistance puts a markedly
increased work load on the right heart; over time, the right
ventricle hypertrophies (increases its muscle mass) to accom-
modate the need to generate excessive pressure. If the right
ventricle cannot sustain the additional work, it will fail, a
condition known as "right heart failure" which has several
distinct features from the more common left heart failure.
In addition, the hypertrophic right ventricle may demand more
oxygen than its overextended coronary arteries are capable of
supplying, producing angina pectoris. The phenomenon of
right ventricular hypertrophy in association with lung dis-
ease is known as "cor pulmonale" and has been well documented
to occur in coal miners with complicated coal workers' pneu-
moconiosis.
*This report was prepared for Enviro Control, Inc., by Tee L. Guidotti,
M.D., medical consultant on this project.
Kl-l-A
-------
b) Distension of the right atrium by the same mechanism de-
scribed above and hypoxemia due to poor oxygen exchange by
severe lung disease sensitize the heart to arrhythmias. In
severe acid/base disturbances, which are not uncommon in
chronic lung disease, the heart is also made more irritable.
.Foci in the cardiac conduction system and myocardium are
thus more susceptible to spontaneous depolarization and may
escape from electrical domination by the sinus node. Thus,
irregularities in rhythm are among the most common conse-
quences of moderate to severe lung disease. At the extreme,
these arrhythmia can be fatal.
c) Severe lung disease is a chronic process of symptomatic im-
provement followed by exacerbation in a stepwise downhill
course. It is rarely a smooth history of decline, as cardiac
disease more commonly may be. An intercurrent illness, such
as influenza, may be all that is needed to acutely decompen-
sate the .patient and .precipitate a cardiac problem. Unless
a physician is attuned to this in his record-keeping, the
interrelation may be lost in chart review.
d) When a patient ceases to breathe as a result of lung dis-
ease, and its attendant disturbance in gas exchange, the next
event to occur (after brain damage, which does not declare
itself immediately) is cardiac arrest. Undoubtedly, many
deaths from lung disease are recorded as deaths from cardiac
arrest.
It should also be noted that primary cardiac disease can have dra-
matic effects on the lungs, in the form of pulmonary edema, congestive
heart failure, pulmonary emboli, hemoptysis, and so forth. As a general
rule, however, these are less likely to be mistaken for primary lung dis-
ease in records review.
The incidence of cardiac complications of lung disease in coal mining
communities is not the primary question of a disease risk survey, although
it may be interesting from the standpoint of clinical management. The
fundamental problem is the risk of lung disease per se and the risk of
primary cardiac disease per se.
Of overriding importance in assessing disease risk is the question of
predisposition .to atherosclerotic coronary artery disease. Data on aggre-
gate "-heart disease" does not help to approach this problem because of the
reservations outlined above. One study that looked at the incidence of
peripheral arterial disease in another arterial bed (cerebral) is described
below and .does not :suggest any particular .predisposition.
Kl-2
-------
ARTERIAL DISEASE IN COAL MINERS
Stroke is the result of a disturbance in the arterial blood supply .to
the brain. By far the most common causes are complications of atheroscle-
rosis and hypertension. (Congenital deformities are much less common,
usually o'ccur in younger patients, and tend to be clinically distinguish-
able. Traumatic causes of brain infarction are not considered "strokes.")
Hypertension and .atherosclerosis are closely associated, and it is gener-
ally not useful to separate them for epidemiological purposes. The inci-
dence of stroke can, be assumed to reasonably reflect the incidence of se-
vere peripheral vascular disease in general.
The best information on the incidence of stroke in coal mining com-
munities comes from a study of coal miners conducted in 1967 by the United
Mine Workers in an attempt to audit their health services. The population
consisted of bituminous coal miners in Western Pennsylvania, and the diag-
nostic criteria used were clinical and comparable to other studies. They
derived an incidence of 7.9 strokes per 1,000 program-eligible miners per
year, a figure comparable to the population at large in several other
studies, there is thus no evidence for an increased incidence of stroke
among bituminous coal miners.
It would appear unlikely that coal miners are significantly predis-
posed to peripheral arterial disease and, by extension, such a predisposi-
tion would be even more unlikely in the nonmining population of coal
mining communities.
VENOUS DISEASE IN COAL MINERS
This is briefly discussed in Appendix K2. The basic disease process
of concern is venous distension and thrombosis caused by sluggish blood
flow. Coal miners sit or crouch for long periods in awkward positions,
or may stand for long periods at the face without moving around ve'ry much.
These are well known to cause stasis of venous flow, and the high inci-
dence of venous disease in miners comes as no surprise.1 When superficial
veins dilate as a result, they become typical varicose veins; however
cosmetically undesirable and troublesome to the patient, they are trivial
as a cause.of.death. Much more serious is In situ thrombosis and thrombo-
phlebitis. The clot which forms on the wall of the vein may break off and
travel in the venous circulation to the lung (through the right heart);
this is pulmonary embolism, which is a major cause of death from vascular
disease. Pulmonary embolism usually results from thrombophlebitis of the
deep veins of the legs, especially in men.
It is highly unlikely that venous disease in coal miners has anything
to do with coal dust exposure; the high incidence is easily explicable by
the working conditions in the mines. There is no reason to believe that
there is any increased risk whatsoever in nonminer residents of mining
communities.
Kl-3
-------
Portal vein thrombosis is a very serious, rare complication of cir-
rhosis and occasionally of infections involving organs that drain into the
portal vein. The excess in coal miners probably simply represents the
high prevalence of cirrhosis. There is very little reason to believe that
it has anything to do with coal dust exposure, and it is so infrequent as
to be unrewarding in search.
There is no reason to suspect any elevation in risk for peripheral
venous disease in nonminer residents of coal mining communities.
CARDIOVASCULAR DISEASE WHICH MIGHT BE FOUND
IN EXCESS IN COAL-MINING COMMUNITIES
ICDA numbers 410 (myocardial infarction), 411 (acute/subacute ischemic
heart disease), 412 (chronic ischemic heart disease), and 413 (angina pec-
toris) are different manifestations of the same basic disease process:
atherosclerotic coronary artery disease. If coal mining community resi-
dents are predisposed to accelerated atherosclerosis, these diseases should
be uniformly elevated in risk. This data is easily confused by the high
prevalence of lung disease. If chart review data is to be used to derive
a risk estimate, every effort must be made to exclude cardiac complica-
tions of lung disease. The etiology of such complications is fundamentally
different than that of primary cardiac disease, and lung disease will in
effect be counted twice if it cannot be separated out. The best available
data, though limited, suggests that mining populations have no particular
predisposition to accelerated atherosclerosis.
ICDA number 414 (asymptomatic ischemic heart disease) is not par-
ticularly useful in view of the high frequency of relatively trivial
abnormalities on electrocardiogram; the chart data available will not
be sufficient to support the diagnosis consistently from patient to
patient.
ICDA numbers 450 (pulmonary embolism), 451 (thrombophlebitis), 452
(portal vein thrombosis), 453 (other venous disease), 454 (lower extremity
varices), and 456 (varices, other sites) have satisfactory alternative ex-
planations. The chance that a community population risk could be found is
exceedingly small.
ICDA number 457 (noninfective disease of lymphatic channels) is most
likely a reflection of edema and probably overlaps with 454. There is no
reason to suspect any relationship between coal and any form of extrapul-
inonary lymphatic disease. (Lymphoma, malignancy of the lymphatic system,
is coded under other designations.)
ICDA number 458 (other diseases of the cardiovascular system) is es-
sentially so vague as to be useless and undoubtedly overlaps with the more
specific codes cited above.
Kl-4
-------
CONCLUSIONS
Searching for an elevated risk from cardiovascular disease in coal
mining communities will probably be low in yield. If any excess were
found, it'would most likely reflect cardiac complications of primary pul-
monary disease. It is possible, but unlikely, that atherosclerotic coro-
nary artery disease could be increased in risk; it would appear in several
ICDA classifications and should also be reflected in incidence data for
stroke. Venous disease and lymphatic disease are so unlikely to be ele-
vated in community risk that they may be essentially ignored.
BIBLIOGRAPHY
F/jlk, L. A., J. P. Zimmerman, and C. H. Bisdee. Stroke among a coal
mining population. Johns Hopkins tied. J. 120:380-392, 1967.
Lee, D. H. K. Coal workers' pneumoconiosis—State of knowledge and re-
search. J. Occup. Med. 13:183-188, 1971.
Naeye,. R. L. Soft coal workers' pneumoconiosis: A quantitative, post-
mortem study. J. Occup. Wed. 1386-88, 1971.
Kl-5
-------
Appendix K2
ETIOLOGY OF RESPIRATORY DISEASES
ASSOCIATED WITH BITUMINOUS COAL MINING*
1. DISEASE INCIDENCE IN AMERICAN BITUMINOUS COAL MINING COMMUNITIES
Studies of disease incidence in coal mining communities have concen-
trated on the coal miner himself. There has been little attention paid to
the nonminer residents of the regions where coal is mined, and what stud-
ies exist come mostly from the United Kingdom. One finds a dearth of in-
formation on the effect of this industry on community health in the United
States.
Characteristics of American Coal Mining Communities
Coal-mining communities in the United States differ from other settle-
ments in several ways which may affect epidemiologic trends.
a) Active bituminous coal mining in the United States is concentrated
in one geographic region, Appalachia, which has had a high prevalence of
infectious pulmonary disease. Historically, tuberculosis has been rela-
tively more prevalent and, until recently, less well controlled than in
other regions. Histoplasmosis is endemic to the area as well.
b) Coal mining communities, especially in Appalachia, have been sub-
ject to wildly fluctuating economic circumstances, which have led to a
skewed age distribution in the population. Some Appalachian counties lost
virtually their entire healthy male cohort aged -20 to 40 during the peak
years of emigration in the 1950's. Thus, the age structure of individual
communities in Appalachia may confuse estimates of the incidence of age-
related, especially chronic, diseases. There may also be a bias toward
overrepresent ing chronic disease presenting in youth or in the complica-
tions of institutionalization or mental deficiency, since such conditions
will occur disproportionately in the adult male population which did not
emigrate, as the result of selection pressures.
c) Poverty and inadequate health services, especially in the most rural
areas of Appalachia, may affect the quality of incidence data, since seri-
ous disease may be underreported. Depending on the continuity of local
medical care, diagnostic data may not be accurate in the distant past.
d) Coal mining communities tend to be relatively homogeneous in socio-
economic class and ethnic origin (there are certain important exceptions
to this).
*This report was prepared for Enviro Control, Inc., by Tee L. Guidotti,
M.D., medical consultant on this project.
K2-1
-------
Diseases of Bituminous Coal Miners
Coal miners themselves are predisposed to a variety of illnesses,
chiefly respiratory. There are four major occupational respiratory dis-
eases of coal miners:
• Coal Workers' Pneumoconiosis, simple (CWP); this results
from the accumulation of inhaled coal dust in discrete
macules in the lung and can be associated with some changes
in respiratory function, rarely if ever disabling.
• Progressive Massive Fibrosis (PMF, or "complicated CWP);
this is an advanced, uncommon form of CWP which results in
obstructive (emphysematous) and restrictive pulmonary dis-
ease and may be life threatening.
• Chronic Bronchitis; this probably represents chronic bron-
chial irritation from dust, but has not been extensively
studied.
• Silicosis; this occurs in miners who drill rock, sink
shafts, or drive the transport cars ("motormen") and can be
life threatening in severity.
Lung cancer in coal miners is a controversial issue. American coal
miners have been reported to have excess mortality from lung cancer, but
this has not been found in the United Kingdom. (The British data may be
diluted by other job categories.) Furthermore, coal miners have been re-
ported to have excess cancer of the upper respiratory tract (pharynx, pos-
sibly larynx), which is much less frequent than cancer of the lung. If
there were a weak carcinogenic substance in coal dust, there might be an
elevated community risk for cancer of the lung and upper respiratory tract
in settlements with heavy dust fallout or in the families of miners. It
is therefore interesting to note that there is indeed an excess of cancer
of the oropharynx in coal mining counties in Appalachia, but not of lung
cancer.
Tuberculosis is reportly more prevalent in U.S. coal miners than in all
U.S. working men, but again this has not been found in the U.K. One must
recall that TB was poorly controlled in Appalachia in general until re-
cently, so that a comparison with all U.S. working men is not satisfactory.
Miners in general have excess mortality from pneumonia and from pul-
monary embolism. The latter surely reflects their prolonged awkward pos-
ture underground, which predisposes to venous stasis and deep vein throm-
bosis.
Nonrespiratory diseases to which coal miners appear to be at excess
risk include cancer of the stomach, gastric ulcer, hepatic cirrhosis, and
syphilis. The last three are perhaps better explained by factors of life-
style than by occupation. Miners also have high rates of fatal and dis-
abling accidents on the job, and of homicide.
K2-2
-------
Disease Patterns in Bituminous Coal Mining Communities
There is data quoted in earlier work on this project to suggest in-
creased respiratory disease in nonminer residents of coal mining communi-
ties, but this did not prove consistent between communities. In Appalachia
there is a shared risk of tuberculosis and of histoplasmosis complicating
these findings.
Relatively few cancers seem to cluster in Appalachian coal mining
counties; these include cancer of the cervix, upper respiratory trace
(oropharynx), skin (all types), and bone (all types). One intriguing asso-
ciation from the U.K. suggests excess mortality from leukemia among the
wives of coal miners.
The disease experience of coal miners which may plausibly result from
an exposure shared by nonminer residents of the community can be added to
the scant data available on coal mining communities themselves to yield a
list of conditions most likely to be found in excess in such settlements.
.The following emerge as the most likely diseases to be found in excess
in bituminous coal mining communities:
• Chronic Obstructive Pulmonary Disease (emphysema, chronic
bronchitis)
• Tuberculosis
• Cancer of the Lung
• Cancer of the Head and Neck
• Pneumonia
• Skin Cancer
• Leukemia
Bibliography
Adelstein, A. M. Occupational mortality: Cancer. Ann. Occup. Hyg. 15;
53-57, 1972.
British Registrar-General's Decennial Supplement for England and Wales:
Occupational Mortality Tables. H.M.S.O., London, 1972.
Milham, S., Jr. Occupational Mortality in Washington State, 1950-1971.
Washington State Department of Social and Health Services, 1974, pri-
vately distributed.
Morgan, W. K. C., and A. Seaton. Occupational Lung Diseases. W. B.
Saunders, Philadelphia, 1975.
K2-3
-------
2. COAL WORKERS' PNEUMOCONIOSIS IN THE UNITED STATES
The American coal miner works at a hazardous occupation, and his
hazards are many. The respiratory problems which attend coal mining are
complex; despite a massive accumulation of data, they remain highly contro-
versial. Since the controversy is oVer substantial issues of pathophysi-
oiogy and risk factors, it makes little sense to avoid it. Rather, in
this review the best available understanding of coal workers' pneumo-
coniosis will be presented in an attempt to come somewhat closer to the
reality, if not to a consensus.
Coal miners are subject to the same respiratory illness as the popu-
lation at large; in addition, they have at least three well-defined occu-
pational respiratory diseases (1):
• coal workers pneumoconiosis, the subject of this review;
• silicosis, especially among mine construction crews, roof
bolters, and face workers at narrow seams; and
• bronchitis, usually chronic, related to acute dust exposure.
A discussion of the health pf coal miners is incomplete without notes
on the industry and miners' daily activities.
Coal Mining in the United States
Coal mining is far from the experience and background of most physi-
cians. The American industry is heavily capital-intensive and has a long
history of political controversy, economic reversal, and labor unrest.
The major centers of coal mining technology and development are remote
from all but a few university centers of medical research. Furthermore,
a tradition of mutual cooperation between medical researchers and either
labor or management has been absent (2), and outsiders are simply not wel-
come in the coalfield, by either side. For these reasons, physicians in
general, even those with a professional interest in the pneumoconioses,
seldom have a detailed understanding of the industry's operations and jobs.
This is unfortunate, because the jobs are highly specialized and imply
very different exposure risks and dangers.
The following summary of underground coal mining operations is meant
to outline the workings of a typical modern, mechanized American coal
mine (3). Details of the operations and machinery vary somewhat between
mines, and are generally not applicable to either strip mining or under-
ground metals mining.
Coal mining is fundamentally a process of dislodging a layer of vege-
table matter, which has attained the consistency of soft rock, from be-
twe'en layers of hard rock. The coal seam lies horizontally in most of
West Virginia and Kentucky but slopes upward and is exposed at the surface
in parts of Pennsylvania, due to geological fractures. Mines are named
for the town or company and numbered according to the geological stratum,
K2-4
-------
from the bedrock to where the coal seam lies; e.g., various mines in one
area with a deep seam may be called "Peabody No. 9" or "Oak Creek No. 11."
Strip mining; is an efficient method of extracting coal but is suitable
only for more shallow seams and larger operations; it carries a substan-
tial additional cost for reclamation of the land stripped off the seam and
replaced—the "overburden." Occasionally, strip and underground mines co-
exist in the same coalfield. Underground mines are constructed by first
drilling a vertical shaft, then a horizontal corridor. Perpendicular to
the corridor, shallow "panels" are cut at intervals. The panels are en-
larged into "rooms" when the exposed "face" of the seam is to be "cut," or
mined, at that point. The floor plan of a mine is therefore a rectilinear
array of rooms coming off a main corridor and numerous accessory corridors,
branching from the main corridor. A few rooms, having served their pur-
pose in mining, are turned into crude workshops or garages for machinery.
This room-and-pillar arrangement may be the structurally safest and most
efficient floor plan* but still leaves 65% of the coal unmined between the
rooms, serving as pillars to support the roof.
Access to a mine is through a "portal," usually an electric elevator
in the vertical shaft. Electric shuttlecars run by "raotormen" on rails
negotiate the main corridor. In the past, sand was sprinkled on the steel
rails to provide traction; this resulted in clouds of airborne pulverized
silica and led to a high incidence of silicosis among motormen and brake-
men.
At the face of the seam, an operative with a motorized drilling ma-
chine drills about 10 holes, each 10 feet deep, into the face and inserts
two sticks of dynamite in each. Electronic detonation is triggered by a
"shotfireman" after the nearby rooms are evacuated; the area is reentered
about five minutes later, after the fumes from the explosion have been al-
lowed to dissipate. The atmosphere is tested with a lantern burning a low
flame, called a "canary" for reasons of historical whimsy. Although acrid
and smoky, the area is normally thought safe to reenter shortly. In the
past, this process known as "shotfiring" exposed many miners to toxic con-
centrations of nitrogen dioxide when they reentered prematurely (4).
The next step is "undercutting," in which a six-ton machine with a
cutting tool like that of a chain saw cuts a slot about 12 cm high, two
meters wide, and over a meter deep into the base of the seam, which is
usually 50 to 70 cm high in Appalachia. The largest machine in the mines,
the continuous miner, then chews into the seam and conveys the broken
chunks of coal into a bin, where it is loaded into a loading machine for
transport. These operations at the face of the mine are extremely dusty,
and these workers sustain the heaviest exposure to coal dust** (5).
*An alternative plan, called "longwalling," is the predominant system in
Europe and is used by some American mines.
**Masks are very seldom worn, as they are very uncomfortable, hot, and
impair vision.
K2-5
-------
The coal is brought to a conveyor belt, a V-shaped belt travelling
about a meter per second through a small accessory shaft to the surface,
carrying loads of coal 15 to 20 cm deep. Although against safety regula-
tions, it is quite common for men to hitch a fast ride to the surface by
jumping on the conveyor belt. At the surface, the coal is dumped in a
"main pit" where it is soon loaded onto another copveyor belt, the "slope
belt" which lifts it high in the air and deposits it on the top of a huge
pile, the "tipple." The coal is stored there until it is loaded onto
trucks or railroad cars, by being lifted again by conveyor belt into a
huge funnel, the "cyclone." The mechanics and the surface workers, who
are often older underground miners retiring from work at the face, are
cbnsidered active "miners" but have much less exposure to coal dust than
other workers.
As the face is cut and the coal is removed, the room is deepened by a
cutting machine. This is a potentially very hazardous operation, since
the integrity of the ceiling or "roof" cannot be guaranteed. A special-
ized operative called the "roofbolter" drills dozens of 2-meter holes
perpendicularly into the roof and secures it with long reinforcement bolts
with expandible heads and plates. Drilling into solid rock exposes the
roofbolter, like the shaft driller on the construction crew, to the risk
of silicosis.
Roof falls are the most common fatal accidents in underground mining,
so that great care is taken to reinforce the corridors with oak or steel
beams, to roofbolt the rooms, and the sound the roof for "kettle bottoms."
These are large, semicylindrical'petrified tree trunk fragments lying over
the seam. (When the coal was still a swamp, these would have been float-
ing or mired logs.) They may drop precipitously when undercut and are
exceedingly dangerous.
Each underground miner carries his light on his helmet wired to a bat-
tery on his belt, his survival pack, and his lunch pail. They work in
teams of nine, with a foreman, mechanic (or electrician), motorman, two
roofbolters, and operatives for each machine. An underground miner's
average workday in the mine proper is only 6.5 hours; the rest of his time
is spent in transit and setting up the machinery.
„ Sources of first-hand information on coal-mining operations are not
limited to industry and union representatives. Trade journals such as
Coal Age and Coal Mining and Processing, publications of the U.S. Bureau
of Mines and the U.S. Mine Enforcement and Safety Administration (MESA),
and university departments of engineering are helpful, once one becomes
familiar with the jargon. Although operating mines discourage visitors,
there are exhibits of coal mining in museums; perhaps the best known is in
th6 Deutsche Museum in Munich. Finally, the city of Beckley, West Vir-
ginia, has acquired a small exhausted mine, outfitted it with representa-
tive if outdated equipment, and provides tours with lectures on the his-
tory and technology of coal mining.
K2-6
-------
Epidemiology of Coal Mining Populations and Prevalence of CWP
Discussions of coal workers' pneumoconiosis (CWP) without reference to
the epidemiology of coal mining populations in general are liable to dis-
tort the significance of respiratory disease in this community.
Appalachia is a cultural and historical unity* with geographic bound-
aries that also map the distribution of coal. The same mountains that en-
forced isolation when Appalachia was the American frontier in 1825 bore
the coal that was mined since the 1830's. The coal, 250 billion tons of
it, thereupon became a powerful shaping force upon the region. After the
frontier passed beyond Appalachia, the area was geographically isolated
from the mainstream of American economic life and trade and there devel-
oped an intensely provincial, introspective regional culture. This life-
style and world view proved inadequate to cope with the sudden push to
exploit the region's coal resources in the late 19th century, backed pri-
marily by remote Eastern and British financial backers. ,The profound
social and economic dislocations which followed have persisted to the
present day (19,20).
Demographically, these factors of history and geography produced a
marked and sustained loss of population from the region, in the largest
internal mass movement of population within the United States up to that
time. In the 1950's alone, the State of West Virginia lost 417,000 citi-
zens in net migration, and only 251,000 were born to replace them. The
entire region lost over a million residents, a fifth of its population, in
1950. Well-established paths of migration led to Baltimore, Pittsburgh,
Cleveland, Cincinnati, and ultimately to Detroit and Chicago. Most of the
emigrants were young (52% were 18 to 34 years old, compared to 34% of the
area population), men (7% excess over women), and more highly educated
than those they left behind (but not as highly as those to whom they mi-
grated) (19).
These trends resulted in large part from diminished demand for coal,
lack of alternative economic opportunities, and a regional inability to
mobilize community resources (20). Automation and the mechanization of
coal mining also reduced the work force, from some 800,000 in the 1920's
to somewhat more than 112,000 (the present membership of the United Mine
Workers of America) (21). In 1974, with the increasing shortage of petro-
leum available to the United States from foreign suppliers, the price of
coal abruptly more than doubled** and Appalachia is once again experienc-
ing an inflow of capital (21). This has meant the entry of many more new
*Appalachia is comprised of parts of southwestern Pennsylvania, parts of
eastern Ohio, all but the northern panhandle of West Virginia, western
Maryland, western Virginia, eastern Kentucky, eastern Tennessee, western
North Carolina, northern Georgia, and northeastern Alabama. It is
slightly larger than England, Scotland, Wales, and Ireland combined and
has a population roughly the same as that of Belgium (19).
**To $24 per ton, utility-grade bituminous.
K2-7
-------
young coal miners in the work force, which has resulted in an age distri-
bution skewed to both older middle age and younger early adulthood, with a
demographic gap in between (22).
The age-adjusted mortality rates for Appalachia, when compared with
the nation as a whole, have shown that heart disease and cancer* have been
the leading causes of death in Appalachia, the latter at a significantly
lower rate than for the entire United States (23,24). Mortality from pul-
monary disease has been increasing relative to the rest of the country,v
however, as has been infant mortality. The area, as a whole, was rela-
tively, late In controlling tuberculosis (23).
Although mining catastrophes, such as the Blacksville No. 1 (West Vir-
ginia) disaster of 1972 (6) and, most recently, Black Mountain No. 1
(Kentucky) in- 1976, are the most visible in the media, coal mining as an
industry has an exceedingly high rate of nonfatal accidental injuries on
the job (2,6), and coal miners have a high mortality from accidents (2,25).
For nonaccident causes of d'eath, the data are somewhat confusing. Al-
though coal miners in general have been shown to have a norma-1 life expec-
tancy (26), they have demonstrably high standardized mortality ratios (SMR)
in certain causes of death (25). The answer to this paradox is found in
the so-called "healthy worker effect." This principle of occupational epi-
demiology points ..out that the industrial work force is composed of largely
healthy individuals with much less risk of death or serious illness than
the population at large (27).. Thus the cohort of coal miners, who are
under particular selection pressure due to the strenuous demands of their
work, should survive well beyond the "normal" life span and should have
low SMRs.
To the contrary, miners have an elevated risk for a wide range of dis-
eases, compared to the U.S. male population at large, particularly tuber-
culosis, cancer in general, lung cancer, respiratory disease in general,
gastric ulcer, and syphilis (25). On an intuitive basis, gastric ulcer
and syphilis., like the lesser excess of deaths from cirrhosis and the high
excess from homicides (25), can be attributed to life-style in a hard-
working rural working-class white society (22). The remainder are less
easily explained.
i
The elevated mortality for tuberculosis may reflect the historical
prevalence In Appalachia and its delayed control in the area. The British
experience does not confirm suspicions of an elevated risk for TB among
coal miners (28).
For lung cancer, the situation is somewhat different. There is no
reas'on to> suspect than an American coal miner smokes more than his age-
matched nonmining male paired counterpart, and he probably smokes less
since he cannot indulge his habit underground. Nonetheless, the best data
Unusually high concentrations of cancer deaths in Appalachia are few;
they include cancer of the cervix, lip, and oropharynx in women, of skin,
and bone (24).
K2-8
-------
available show an SMR of 192 (meaning a risk factor of 1.92) (25) compared
to 107 for his British counterpart (1). This discrepancy is complicated
by differences in record-keeping: the British data aggregate several job
categories and may have diluted the true figure (29). Thus there is no
satisfactory answer to the question of whether coal miners have a truly
elevated risk of lung cancer, but there is enough evidence to suspect this
in th'e United States (30).
. The prevalence of coal workers' pneumoconiosis (CWP) in the United
States has been a matter of dispute, due to differences in methodology and
criteria for radiographic diagnosis. The last, best study in the U.S. was
the second "round" of the Interagency Study administered by the National
Institute for Occupational Safety and Health. It yielded a prevalence of
simple CWP of 10% in a carefully selected representative population of
U.S. coal miners (n=9,000), with 0.4% showing evidence of complicated CWP
(progressive massive fibrosis or PMF) (1). The prevalence of simple CWP
and of complicated CWP were both highest in anthracite miners (eastern
Pennsylvania) and lowest in the Rocky Mountains (14), and higher in face
miners than in surface workers. These rates are somewhat less, in general,
than those for British and Welsh coal miners (1,14). These figures almost
surely reflect exposure spanning at least two decades before the transect.
Autopsy series are clearly biased, because the patient obviously had a
terminal'condition. Nonetheless, it is of some use to recognize that the
prevalence of complicated CWP at necropsy rose with the time spent under-
ground, and that PMF was present in roughly a quarter of West Virginia
bituminous miners with CWP who came to autopsy before 1966. Almost all of
these miners with PMF had evidence of cor pulmonale. Silicotic nodules
were present, almost always as an incidental finding, in 13% of miners
with CWP, with or without PMF (30).
• • ' \
Diagnosis, Screening, and Disability Evaluation
The diagnosis of CWP has become more a matter of legal definition than
of medical opinion. This is obviously due to the regulations promulgated
under the Federal Coal Mine Health and Safety Act of 1969, as amended in
1972 (6,.7).
From 1969 to 1972, under the Social Security Administration, which ad-
ministered "black lung" compensation, the diagnosis of CWP required two
criteria: radiologic evidence of pneumoconiosis, and one of three objec-
tive .indices of functional impairment—radiographic or autopsy evidence of
progressive massive fibrosis, objective evidence of ventilatory impairment
(by pulmonary function testing), or objective evidence of impaired gas
exchange (by arterial blood gas desaturation or reduced diffusion capac-
ity) . In 1972 the Act was amended, and its implementation became the re-
sponsibility of the Department of Labor. Since then, the requirement for
radiographic proof of pneumoconiosis has been dropped and has been re-
placed by physical examination and a questionnaire on symptoms (based on
the similar standardized format of the U.K. Medical Research Council,
which emphasizes chronic bronchitis). These changes have been sharply
K2-9
-------
attacked as a regression from objective to increasingly subjective cri-
teria (8).
An apolitical, scientifically derived diagnosis of CWP begins with a
working definition, in pathologic terms. Applying and slightly modifying
the International Labor Organization's definition of pneumoconiosis to
coal provides: coal workers' pneumoconiosis is the accumulation of coal
dust, .produced by both mining and handling, in the lungs and the tissue
reaction to its presence.
Radiology. There is no practical method of detecting coal dust in the
lung of the living patient other than by X-ray. Sputum production of coal
pigment is compatible with a recent acute exposure and says nothing about
cumulative lung burden.
The interpretation of the chest film in CWP relies on the occupational
history, since there is no pathognomonic sign for CWP and there is more
similarity than difference between the pneumoconioses. It is even not
clear why CWP should project a radiographic shadow in the first place,
since coal is mostly elemental carbon: a density of 0.4 g/cm2 is required
for coal dust to be barely visible on the chest film, and the total lung
burden of dust is demonstrably lower than that required to produce shadows
by density of dust alone (9). This strongly suggests that the origin of
the "interstitial" pattern of small nodular opacities characteristic of
CWP is the local tissue reaction, superimposed on local concentrations of
coal mine dust in macules such as those demonstrated for tantalum powder
(11,12). There is even some evidence for endogenous iron deposition (13,
14). The standard system for communicating the morphology, nodularity,
and severity of radiographic changes in cases of suspected pneumoconiosis
is the UlCC/Cincinnati* Classification of the Radiographic Appearance of
the Pneuoconioses (15b). This necessarily involved system uses a 12-point
scale which markedly lowers interobserver variation in interpretation and
variation in radiological technique, which have been major problems in
pneumoconiosis radiology (1,16). Using this system, it is nonetheless
difficult to detect progression of CWP among most miners, even over a
decade's time, once the diagnosis is established (1,16).
A characteristic pattern of "pinhead" multiple nodules, usually with
smooth edges, is called "simple" CWP. When any opacity is greater than 1
cm in diameter, it is termed "complicated" CWP, or progressive massive
fibresis. .
Pulmonary Function. The pathologic picture of CWP, involving bron-
chiole, arteriole, and alveolar wall, would suggest that simple CWP has a
protean effect on ventilation, perfusion, and gas exchange. In point of
fact, it usually does not (1,14-17).
In general, coal miners have slightly lower forced vital capacities
(FVC) and FEVj fractions than do nonminers; these changes do not become
*UICC is 1'Union International Contre le Cancer.
K2-10
-------
statistically significant until the miners have worked underground 30 years
or more. The FVC and FEVi and lowest, on the average, for anthracite
miners and less reduced for bituminous miners, but are not markedly re-
duced in any large population sample of working miners (1,14-16). The geo-
graphic distribution suggests that miners in the far West have essentially
no impairment at all (1,8).
Total iung capacity (TLC) and derived residual volume (FVC substracjted
from a radiographically derived TLC) are both increased among bituminous
miners, but the degree of hyperinflation does not correspond well to the
UlCC/Cincinnati radiographic classification. Where there was evidence of
preexisting obstructive airway disease .(FEVj/FVC <70%), more hyperinfla-
tion was present; but this is a natural consequence of obstruction per se.
The most reasonable hypothesis explaining this is that a mild traction
emphysema results from peribronchiolar fibrosis at the coal macule (1).
Studies of pulmonary mechanics have shown that static compliance is
normal in most miners with simple CWP but may be reduced in complicated
CWP. On the other hand, dynamic compliance is reduced consistently, and is
frequency-dependent in most, implying the presence of small airways dis-
ease (1,16,17).
Closing volumes are slightly reduced in nonsmoking miners with simple
CWP, compared to nonsmoking controls. Cigarette smoking obliterates any
differences in the smoking groups. Thus the degree of small airways dis-
ease associated with CWP is overwhelmed by the changes due to smoking (1).
Diffusion capacity is also much reduced by cigarette smoking in all
groups, but is within normal limits for nonsmoking miners with simple CWP.
This index of gas exchange surface area and permeability is reduced only
in complicated CWP (1,17).
Arterial blood oxygen tension in miners with simple CWP is slightly
desaturated compared to nonminer controls, but increases to normal with
exercise. By contrast, miners with complicated CWP desaturate somewhat
with exercise although they may remain nonetheless within entirely normal
limits. The alveolar-arterial oxygen gradient, (A-a)P02» is extremely
sensitive to smoking, but is increased compared to smoking-matched non-
miners and increases slightly with exercise and correlates this with
higher UlCC/Cincinnati classification. Even so, the alterations are mini-
mal (1).
Pulmonary hypertension is not a feature of simple CWP, nor of compli-
cated CWP at rest. Pulmonary arterial pressure may increase more with
exercise in miners with complicated CWP (1,17), but total pulmonary vascu-
lar resistance decreases normally (17). Perfusion radionuclide; scanning,
on the other hand, is normal in simple CWP but is abnormal uniformly in
complicated CWP, with perfusion defects (1).
In summary, simple coal workers' pneumoconiosis is associated with
functionally insignificant small airways disease and gas exchange
K2-11
-------
impairment, which are vastly outstripped by the effects of smoking. Com-
plicated coal workers' pneumoconiosis is associated with obstructive air-
,ways disease of varying severity, with a restrictive component and 'altera-
tions ,in pulmonary perfuslon. The functional significance of this impair-
,ment is highly variable between patients.
,Qisability Evaluation. Having radiographic evidence of abnormality
.describes a condition. Demonstrating functional aberrations describes a
disease. Neither contributes an adequate.description of the individual's
Ability to cope with,the physical demands placed upon him by his work and
.recreation (1.8).
The Devaluation of.miners'for purposes of compensation is based, at
present, on impairment—on the.demonstrable .reduction in respiratory (both
.ventilatory and.oxygenation) capacity—and not truly on disability,
-------
working miners began their habit and were well into their
occupational dust exposure.
• The coal mining community is not particularly receptive to
medical arguments to quit smoking: their work is unrelieved
and heavy, their historical experience with physicians has
• been"unsatisfactory, and the peer-group pressure to smoke
(and drink) is high.
• When a near majority of the adult U.S. population smoked for
several decades, combined exposure to cigarette smoke and to
occupational hazards was the rule, not the exception, and
compensation should be treated accordingly; smokers of that
era should not be penalized.
• . Coal miners are not immune from the same smoking-related
respiratory diseases that afflict the population at large;
when it is impossible to unequivocally attribute respiratory
impairment to an occupational exposure, at least it makes
sense to associate the two by proving the presence of CWP
before according preferential treatment.
Sources of Variability in Coal Dust Deposition
The four types of coal, in decreasing order of density and economic
value, are: anthracite, bituminous, subbituminous, and lignite (which is
barely more dense than dried peat). Each type is graded for useful char-
acteristics by "rank."
Coal dust is composed of elemental carbon predominantly, with small
quantities of pyrite (FeSa), magnetite (FeaOi,), calcite (CaC03), dolomite
(CaMg(003)2), sulfur, water, and traces of aluminum (30). Quartz is also
present in a proportion that varies with the mining operation; more quartz
particles are airborne with rock drilling and cutting operations. In gen^
eral,.however, the airborne respirable particles tend to be more carbon and
less mineral-than the parent seam (presumably more buoyant in air) (30).
i , I
The size distribution of respirable particles has been extensively
studied in experimental models, and much is known about the aerodynamic
behavior of such particles. It has long been known that particles above
10 ym aerodynamic diameter tend to be deposited in larger airways due to
inertial impact ion. Furthermore, particles of less than 0.1 Mm diameter
mostly remain airborne due to the Brownian movement phenomenon. Thus, the
particles in the mid-range are most likely to penetrate to the alveoli and,
once there, to sediment and deposit on the airspace wall (31). The veloc-
ity of this sedimentation is expressed by the following equation:
where d is the aerodynamic diameter of the particle, £ is the vector of
gravitational acceleration, y is the viscosity of air, and p is the
K2-13
-------
density of the particle. Thus the velocity of sedimentation depends not
only on particle size but also on density. One may therefore conceive of
the distribution of particles from a mixed coal mine aerosol as a sedimen-
tation gradient, with lighter coal particles penetrating deeper and sedi-
ment ing more slowly.
Another factor producing separation and heterogenous deposition of a
mixed particle dust is the anatomy of the airways, which is a series of
bifurcations. Secondary flows and eddies result in local nonuniform flow,
slowing of forward velocity, and permit foci of Increased deposition at
the bifurcation. Thus the behavior'of particles in the 0.5 to 2.0 ym
range is governed by boundary-layer and laminar-flow phenomena, resulting
in disproportionate deposition at bronchial and bronchiolar bifurcations.
These phenomena dictate fairly uniform deposition for particles below 1.0
Mm but increasing nonuniformity of deposition with increasing inspiratory
velocity for larger particles up to 5 JJm (32).
Thus the pattern deposition of coal dust in the lungs of any individual
miner at any one exposure is a complex function of the geologic character-
istics of the coal seam, the coal mining operation being performed, the
ventilation of the room where the mining is taking place, and the miner's
breathing pattern at the time.
Pathology and Pathophysiology of CWP
The pathophysiology of CWP is closely linked to patterns of deposition
of the dust and to mechanisms of clearance.
Simple CWP. The essential, defining pathologic finding in CWP is the
coal dust macule, a discrete (occasionally confluent) nonindurated focus
of black pigment microscopically visible under the visceral pleura and on
the cut surface of the lung parenchyma.
Microscopically, the coal dust macules are typically in various stages
of development. The pigment is mostly intracellular in macrophages, whose
persistence has been noted over 10 years after cessation of exposure. The
macrophages are concentrated in the peribronchial and periarterial inter-
stitium, where local fibrosis is apparent. Limited dilation of the respi-
ratory bonchioles is common, in a pattern compatible with centrilobular
emphysema but without evidence of bronchiolitis. The distribution of
macules follows the known patterns of peribronchial and periarterial lym-
phatic drainage,* and draining nodes typically show heavy accumulations of
pigment (14,30,33,34).
*The lymphatic drainage of the lung consists of local "bronchus-associated
lymphoid tissue" draining into "Macklin's sump" at the terminal bronchi-
oles, thence following lymphatics in the pulmonary arterial rays back to
the hilum, or following the pulmonary venous lymphatics back to the hilum
(35).
K2-14
-------
Acutely, the dust inhaled is rapidly cleared by the highly effective
centripetal mucoclllary stream, which removes 3 prodigious burden. None-
theless, a quantity is left behind for phagocytosis by alveolar macro-
phages. The macrophages, having engulfed the individual particles, may
migfa.te dn the alveolar surface or interstitially to the edge of the muco-
ciliary escalator (this process is not well understood), or may migrate
interstitially. Although interstitial migration probably drains into lym-
phatics for the most part, a significant amount of dust is trapped in local
concentrations at the coal dust macules, where presumably draining is im-
paired, overwhelmed, or an anatomic "dead end." These transport mechan-
isms are impaired by inflammatory respiratory disease and by simultaneous
exposure to irritant exogenous agents which may be ciliostatic (36), and
by exposure to, other particulates which may saturate the phagocytic capac-
ity of the macrophage (37).
The pathophysiology of the focal fibrosis and focal emphysema is 'of
some interest, although of limited practical importance because this de-
gree of emphysema in a centrilobular pattern is not clinically significant
(33,34). Although one customarily thinks of phagocytosis as a fairly
"clean" process, it is not. Phagocytosis of inert particles, such as
latex beads, by.alveolar macrophages is enough to release small quantities
of lysosomai enzymes into the surrounding milieu. These include collage-
nase, elastase, plasminogen activator, and lysosomai hydroxylases (38).
Thus, alveolar macrophages may not be activated by ingesting inert parti-
cles such as carbon (39), but they are,stimulated to release, inadvertently
as it were, potentially locally destructive enzymes (40). Teleologically,
this is reasonable in a system evolved to dispatch microorganisms, but it
is counterproductive when dealing with inert particles. Furthermore,
partially denatured protein is chemotactic, so that the cellular reaction
perpetuates itself (40). The schema below is a plausible mechanism for
formation and extension of the coal macule.
Aerodynamic .
filtration
Removal to muco- .
ciliary escalator
Lymphatic drainage ^
Dust exposure
Dust penetration to alveoli
t
— Macrophage phagocytosis
Interstitial carriage
Interstitial accumulation
.Chemotaxls
t
Coal dust macule
* Proteolytic enzyme
release, local
. Focal fibrosis
\ *
^•Disturbed lymphatic
drainage
Ther.e are certainly opportunities for individual variation implied by
this schema (33,36,41). Miners with otj-antitrypsin deficiency, for exam-
ple, conceivably might have more extensive local fibrosis than otherwise.
Those lacking dynein arms in their tracheobronchial cilia, a defect re-
K2-15
-------
cently described in Kartagener's syndrome, may have difficulty with muco-
ciliary clearance (42). The avidity and efficiency of macrophages in
clearing mycobacteria is a genetic variable in rabbits (43) and may have a
genetic counterpart in man.
Complicated CWP, or Progressive Massive Fibrosis. Progressive massive
fibrosis (PMF), or complicated CWP, is an occasionally fulminant, usually
slowly progressive complication of the advanced grades of simple CWP. In
contrast to simple CWP, there is good evidence that PMF shortens life and
causes respiratory disability (not just barely detectable impairment (1,
14-16). It is diagnosed radiographically, in which context it is usually
referred to as complicated CWP, because PMF generally has been used as a
pathological description at autopsy. The criterion is any opacity greater
than 1 cm in diameter on X-ray but any nodule greater than 3 cm in diame-
ter at autopsy (41).
Pathologically, PMF arises ,as a fibrous mass or masses, usually in the
upper lobes and posterior segments, heavily pigmented and with a smooth
contour. Only rarely do these masses cavitate of calcify, despite their
occasionally enormous size; when present, these findings suggest tubercu-
losis. The nodules are much like gigantic coal dust macules, somewhat.
more tightly packed. They do not have the whorled pattern so characteris-
tic of silicosis, nor are they stellate as asbestos.nodules. Bronchioles
are obliterated in the fibrotic mass and vascular structures are invaded
by fibrpblasts, resulting in an obliterative endarteritis. As the mass
expands, larger vessels become involved and in situ thrombosis further
compromises pulmonary perfusion (30,34). As the disease progresses, the
compliance of the pulmonary hypertension develops, with cor pulmonale and
progressive congestive failure (30,41).
The pathophysiology of PMF has been hotly debated; confusion in this
difficult areas is made worse by the recognition of at least seven differ-
ent patterns of fibrosis and simplification in the lungs of coal miners
with CWP (30,33,41):
• the coal dust macule (with collagen deposition)
• the silicotic nodule (present most frequently in mine con-
struction crews and roofbolters)
• chronic bronchitis with bronchiolitis
• focal emphysema (often related to chronic bronchitis with
bronchiolitis)
• tuberculosis
• rheumatoid(Caplan's) nodules
• histoplasmosis (endemic to much of Appalachia)
Many attempts to associate PMF with either silica or tuberculosis have
not proven the point. In both cases, the agents (either silica or acid-
fast bacilli) cannot be consistently demonstrated in the PMF lesions or
K2-16
-------
surrounding lung parenchyma, and antituberculosis therapy fails to arrest
the process. Although the hypotheses that PMF may result from concomitant
exposure to silica or tubercle bacillus are very attractive, based on the
known cytotoxicity and fibrosis-inducing behavior of the agents, respec-
tively, the evidence supports neither. It is entirely plausible that in-
frequent individual cases may be due to mixed CWP and either tuberculosis
or silicosis, but these processes cannot be invoked for the majority of
cases of PMF on present evidence (30,41). PMF is not related to the total
lung or even local burden of dust (1,34).
Other theories of pathophysiology have employed immunologic mechanisms
on the model of Caplan's syndrome (to be discussed later). These concepts
are still in their germinal stages and await a more complete phenomenolog-
ical description of the rheumatology of coal miners.
Individual variations in susceptibility to fibrosis have not been well
studied for any population, with the exception of dj-antitrypsin defi-
i ciency, which has not been widely investigated in coal miners with PMF.
! This is a significant omission from the literature and a valid direction
| for systematic research. A representative sample of miners with simple
CWP and PMF should be screened with the following techniques:
I • Oi-antitrypsin activity and P^ type
| • fibroblast in vitro culture
• hydroxyproline synthesis studies
I • history of abnormal fibrosis (e.g., keloid formation,
| premature wrinkling, family history of such)
• . history of exposure to asbestos or oxidizing irritative
gases (e.g., N02, 03)
1 • collagen structure and crystallization (by X-ray dif-
I fraction).
These studies should clarify the question of variations in host response,
even if the results are negative.
Finally, simultaneous exposure to other agents inducing fibrosis, be-
sides mycobacteria, must be considered. Simultaneous or sequential expo-
sure to nitrogen dioxide (NOa) with carbon particles works synergistically
to modestly increase the severity of centrilobular emphysema which results
from NOa alone (43), but does not seem to induce nodular fibrosis. N02 is
a well-recognized mining hazard, generated mostly from welding and shot-
firing (4).
The most obvious nonoccupational exposure in miners is cigarette smok-
ing, and there is ample evidence that smoking impairs particulate clear-
ance (44) and predisposes to secondary infection (36,44). Furthermore,
the smoking miner is at greater risk for abnormalities in pulmonary func-
tion and for chronic bronchitis than his nonsmoking colleague (1,14,15,17).
Whether smoking contributes to the development of PMF is problematical,
K2-17
-------
but it is certainly a confusing factor which raises morbidity from respi-
ratory disease in any population. There is little reliable published data
on the smoking histories of miners with simple CWP compared to those-with
PMF.
Rheumatologic and Immunologic Aspects of CWP. Caplan's syndrome is
the association of CWP, often quite minimal, pulmonary rheumatoid nodules,*
and clinical rheumatoid arthritis (RA). The features of the syndrome
rarely can be seen in other pneumoconioses. This is the best-defined
clinical syndrome and probably represents one end of a spectrum of pneumo-
coniotic-immunologic disturbances (1,13,44).
Welsh coal miners with PMF have an increased prevalence of RA, but not
miners in general. This led to the hypothesis that a population with
rheumatoid diathesis may be at increased risk for both PMF and for rheuma-
toid complications of CWP. The only available markers for such a diathe-
sis are serological, rheumatoid factor (actually several related anti-
immunoglobin antibodies) and antinuclear antibody. (These may reflect the
early disease state rather than a diathesis, however.) Surveys demon-
strated that rheumatoid factor was present in sera of miners-with Caplan's
syndrome with or without PMF, but not at any unusual frequency in miners
with PMF alone, simple CWP alone, or in coal miners at large. On the
other hand, antinuclear antibody was present in miners at large, particu-
larly where CWP was prevalent (anthracite much more frequently than bi-
tuminous), and at high frequency in miners with either Caplan's syndrome
or PMF (1). Thus, the serologic abnormalities appear to indicate the
presence of an active immunological process rather than a population at
risk.
Antilung antibodies to connective tissue elements have been recognized
for some years, but are seen in other chronic lung diseases as well (45).
Only Goodpasture's syndrome, where antibody to lung and kidney basement
membrane is involved, provides an accepted model for an autoimmune pulmon-
ary disease, and its pathology is quite distinct. It has been proposed
that protein adsorption onto coal dust particles may favor induction of
rheumatoid factor (46), but the pathophysiological role of this remains
obscure.
In short, the immunological aspects of CWP are still at the stage of
phenomenological description.
Conclusions
Medical concern over CWP is best directed toward the relatively small
fraction of miners with PMF, which is a potentially life-threatening occu-
pational disease. The treatment of advanced PMF is the treatment of cor
*These peripheral, well-circumscribed, occasionally calcified nodules are
quite distinctive: they have fibroblasts in palisades, may centrally
necrose, and are surrounded by lymphocytes; they may resemble silicotic
nodules, and an endarteritis may be present.
K2-18
-------
pulmonale and restrictive lung disease and falls within the competence of
any good chest physician. At the present state of the art, the fibrosis
cannot be reversed and cannot even be arrested in fulminant cases.* Even
removal from the dust exposure has no effect on the progression of the
disease.
The key to decreasing the risk of PMF is to reduce the exposure to all
occupational inhalants, including coal mine dust and irritant gases. Re-
duction of smoking would be desirable, but it is naive to expect this soon.
Proposals to employ prophylactic chemical treatments to prevent a pneumo-
coniosis (47) are questionable to say the least; the experience with alu-
minum powder as prophylaxis against silicosis is a recent warning that
such agents may compound a problem they were designed to correct (48).
Although the incidence of CWP is probably on the decline due to im-
proved dust control, the technology is imperfect and cumbersome. At the
moment, the best that can be done is to enforce existing safety regula-
tions, ensure adequate ventilation, limit time spent in more hazardous
jobs, and support the technical development of better methods.
A description of the problems of Appalachia appeared recently in the
medical literature; like so many, it was oversimplified, distorted, and
failed to recognize the substantial progress the region has made since the
1950's. No mention was made of the historically dominant force for medi-
cal progress in the region, the United Mine Workers of America,** which is
again falling:into confusion just as the region as a whole advances. Par-
ticularly galling to citizens of Appalachia, and to those who have worked
in the region, was the patronizing remark: "...the hillbilly is the most
uncomplaining patient in North America" (49); he certainly is not, nor is
there any reason why he should be.
*As knowledge of the pathophysiology of fibroblast injury and prolifera-
tion advances, it might be reasonable to design a prospective study as
follows: suitable tuberculosis-free patients with fulminant PMF (or
silicosis) would be randomly assigned to receive intrabronchial cortico-
sterpids, penicillamine, or aminobenzoate, and followed for arrest of
their progression. Such a study should only be conducted by experienced
clinical investigators at a major medical center.
**This many-faceted, often difficult institution is responsible for imple-
menting most advances in mine safety and for the indispensable Appala-
chian Regional Hospital network.
K2-19
-------
References Cited
1. Morgan, W. K. C., and N. L. Lapp, Respiratory disease in coal miners,
Am..Rev. Respir. Dis. 123:531-559, 1976.
2. Dessauer, P., Development of patterns of coal workers' pneumoconi-
osis and its association with respiratory impairment, Ann. N.Y. Acad.
Sci. 200:220-251, 1972.
3. This summary was compiled from personal experience supplemented by
documents and discussions with representatives of the U.S. Mining
Enforcement and Safety Administration, the Appalachian Laboratory
for Occupational Respiratory Diseases, the United Mine Workers, the
Beckley (West Virginia) Exhibition Mine, the Island Creek Coal Com-
pany, and the Ziegler Coal Company. I have personally visited coal-
fields, both underground and surface facilities, in West Virginia,
Kentucky, and Arizona-..
4. Kennedy, M. C. S., Nitrous fumes and coal-miners with emphysema,
Ann. Occup. Hyg. 15:285-300, 1972.
5; Schlick, D. P., and N. L. Fannick, Coal in the United States, pp.
13-26, in* M. M. Key, L. E. !Ker.r, and M. Bundy (eds.), Pulmonary Re-
actions ,CD Coal Dust: A Review of U.S. Experience, Academic Press,
New York, 1971.
6, Some of the legal information presented is drawn from the transcript:
Hearing before the Subcommittee on labor of the Committee on Labor
and Public Welfare, U.S. Senate, 92nd Congress, on Implementation of
the Federal Coal Mine Health and Safety Act, 1972, Government Print-
ing Office, Washington, D.C., 1972. This source deals mainly with
• safety; however.
7. , I participated in the clinical evaluation of new applicants for com-
„ pensation presenting to the Pulmonary Clinic of the West Virginia
University Medical Center, July-September 1974.
8. (a) Morgan, W. K. C., Respiratory disease in coal miners, J. Am. Ned.
Assoc. 231:1347-1348, 1975. (b) Phelps, H. W., Response to commen-
tary by W. Keith C. Morgan (letter), J. Am. Ned. Assoc. 234:915,
1975. (c) Morgan, W. K. C., Respiratory problems in coal miners
(letter), J. Am. Ned. Assoc. 235:1324-1325, 1976,
9. Worth, G., Is coal mine dust visible in X-rays of the lungs?, pp.
273-280, in H. A. Shapiro (ed.), Pneumoconiosis, Proceedings of the
International Congress, Johannesburg, 1969, Oxford University Press,
London, 1970.
11. Friedman, P. J., and G. M. Tisi, "Alveolarization" of tantalum pow-
der in experimental bronchography and the clearance of inhaled par-
ticles from the lung, Radiology, 104:523-525, 1972.
12. Rossiter, C. E., Relation of lung dust content to radiological
changes in coal workers, Ann. N.Y. Acad,. Sci. 200:465-477, 1972.
13. Liddell, D., Validation of classifications of pneumoconiosis, Ann.
N.Y. Acad. Sci. 200:527-551, 1972.
K2.-20
-------
14. Morgan, W. K. C., Coal workers' pneuraoconiosis, pp. 161-209, in Occu-
pational Lung Diseases, W. B. Saunders, Philadelphia, 1975.
15. (a) Lapp, N. L., and A. Seaton, Pulmonary function, pp. 153-177, in
Pulmonary Reactions to Coal Dust: A Review of U.S. Experience, op.
cit..
15. (b) UICC Committee, UlCC/Cincinnati classification of the radio-
graphic appearance of the pneumoconioses: A cooperative study, Chest
58:57-67, 1970.
16. (a) Lapp, N. L., and A. Seaton, Lung mechanics in coal workers' pneu-
moconiosis, Ann..N.Y. Read. Sci. 200:433-464, 1972.
16. (b) Liddell, F. D. K., Assessment of radiological progression of sim-
ple pneumoconiosis in individual miners, Br. J. Ind. Med. Ji:186-195,
1974.
17. Ulmer, W. T., and G. Reichel, Functional impairment in coal workers'
pneumoconiosis, Ann. N.Y. Acad. Sci. 200:405-432, 1972.
18. Gaensler, E. A., and G. W. Wright, Evaluation of respiratory impair-
ment, Arch. Environ. Health 12:146-189, 1966.
19. Brown, J. S., and G. A. Hillery, Jr., The great migration, 1940-1960,
pp; 54-78, in T. R. Ford (ed.), The Southern Appalachian Region: A
Survey, University of Kentucky Press, Lexington, 1962.
20. Caudill, H. M., Night Comes to the Cumberlands, Atlantic Monthly
Press; Boston, 1963.
21. The year of the coal miner, Bus. Week, n.v.:44-50, August 31, 1974.
22. Ross, M. H., The Appalachian coal miner: His way of living, working,
and relating to others, Ann. N.Y. Acad. Sci. 200:184-196, 1972.
23. Hamilton, C. H., Health and health services, pp. 219-224, in The
Southern Appalachian Region: A Survey, op. cit.
24. National Cancer Institute, Atlas of Cancer Mortality by U.S. Counties:
1950-1969, Government Printing Office, Washington, D.C., 1975.
25. Enterline, P. E., A review of mortality data for American coal miners,
! Ann. N.Y. Acad. Sci. 200:260-272, 1972.
26. Ortmeyer, C. E., et al., The mortality of Appalachian coal miners,
1963-1971, Arch. Environ. Health 29:67-71, 1974.
27. Goldsmith, J. R., What do we expect from an occupational cohort?, J.
Occup'. Med. 17:126-127, 1975.
28. Dick, J. A., Tuberculosis in coal miners (letter), Lancet J:44, 1977.
29. (a) Mooriey, F. S., Coal workers' pneumoconiosis and carcinoma of the
lung (letter), Lancet 1:43, 1975. (b) McLintock, J. S., (reply),
Lancet i:224-225, 1975. (c) Mooney, F. S., (reply), Lancet 1:390,
1975.
30. (a) Laquer, W. A., Final Report on the Pathology of Coal Workers'
Pneumoconiosis and Related Investigation (PHS Grant 64-65-9), Appala-
chian Regional Hospital, Beckley, West Virginia, 1966.
K2-21
-------
30. (b) DeNee, P. B., Mine dust characterization using the scanning elec-
tron microscope, Am. Ind. Hyg. Assoc. J. n.v.:645-660, 1972.
31. Morrow, P. E., Models for the study of particle retention and elimi-
nation in the lung, pp. 103-119, in Inhalation Carcinogenesis, U.S.
.Atomic Energy Commission Symposium Series No. 18, 1970.
32. Bell, K. A., and S. K. Friedlander, Aerosol deposition in models of a
human lung bifurcation, Staub-Reinhalt. Luft 33:n.p., 1973.
33. Heppleston, A. G., The pathological recognition and pathogenesis of
emphysema and fibrocystic disease of the lung with special reference
to coal workers, Ann. N.Y. Acad. Sci. 200:347-369, 1972.
34. Parkes, W. R., Occupational Lung Disorders, pp. 217-269, Butterworth,
London, 1974.
35. Dannenberg, A. J., Influence of environmental factors on the respira-
tory tract, J. Reticuloendothelial Soc., in press, 1977.
36. Green, G. M., Lung defense mechanisms, Med. Clin. North Am. 57:547-
562, ,1973.
37. Sorokin, S. P., and J. D. Brain, Pathways of clearance in mouse lungs
exposed to iron oxide aerosols, Anat. Rec. 182:581-626, 1975.
38. Unanue, E. R., Secretory function of mononuclear phagocytes, Am. J.
Patho2. 83:396-417, 1976.
39. Dannenberg, A. M., et al., Macrophage activation in tuberculous
lesions, Excerpta Med. Int. Congr. Ser. 325:223-235, 1973.
40. Dannenberg, A. M., Macrophages in inflammation and infection, N. Engl.
J. Med. 293:489-493, 1975.
41. Naeye, R. L., Types of fibrosis in coal workers' pneumoconiosis, Ann.
N.Y. Acad. Sci. 200:381-400, 1972.
42. Afzelius, B. A., A human syndrome caused by immotile cilia, Science
293:317-319, 1976.
43. (a) Lurie, M. B., and A. M. Dannenberg, Macrophage function in infec-
tious disease with inbred rabbits, Bacteriol. Rev. 29:466-476, 1965.
43. ,(b) Boren, H. G., Sequence of exposure to environmental agents as a
determinant of pulmonary injury, pp. 396-410, in Proceedings of the
Tenth Aspen Emphysema Conference, Government Printing Office, Washing-
ton, D.C., 1967.
44. Green, G. M., Pathophysiology of the alveolar macrophage system, pp.
463-480, in Proceedings of the Tenth Aspen Emphysema Conference, op.
cit.
45. Burrel, R., Immunological aspects of coal workers' pneumoconiosis,
Ann. .N.Y. Acad. Sci. 200:94-105, 1972.
46. Wagner, J. C., Etiological factors in complicated coal workers' pneu-
moconiosis, Ann. N.Y. Acad. Sci. 200:401-404, 1972.
K2-22
-------
47. Ulmer, W. T., and W. Weller, Aspects of silicosis prophylaxis, pp.
503-507, in Pneumoconiosis, Proceedings of the International Congress,
Johannesburg, 1969, op. cit.
48. Guidotti, T. L., Pulmonary aluminosis—A review, Bull. Soc. Pharmacol.
Environ* Pathol. 3:16-18, 1975.
49. (a) The exploitation of Appalachia, Lancet JJ:899, 1976. (b) Whis-
nant, D. E., (reply), Lancet JJ.-1358, 1976. < (c) Morgan, W. K. C.,
(reply), Lancet 1:197, 1977.
K2-23
-------
Appendix L
HOSPITAL RECORD ABSTRACT FORMS
Short Form (Non-Respiratory & Non-Cardiovascular Diseases)
Long Form (Respiratory & Cardiovascular Diseases)
Code List
-------
.-\aST.1AC7 FORM
Hospital
Abstractor's Name_
Date Abstracted
SHORT FORM
(NON-RESPIRATORV & NON-CARDIOVASCULAR DISEASES)
PATIENT'S E.C.I. IDENTIFICATION NUMBER -1
(1-7)
Personal/Demographic Data (Write in "unknown" if data unavailable)
1. Patient's Present Residence:
Town
(8-9)
lip Code
(10-14)
Length of Residence (Years)
(13-16)
2. Patient's Previous Residence:
_• Town & State
(17-18)
Length of Residence (Years)
(19-20)
3. Date of Birth (Month/Oay/Year) / /•
(21-22)(23-24)(25-26)
«. Sex (enter code) M » 1, F «2
(29)
5. Race (enter code) Cauc • 1, Negro • 2
Hispanic • 3, Other • 4
.! (30)
6. Pate of Host Recent Admission (Month/Day/Year) /
(8-9) (10-11X12-13)
Final Diagnoses: 1) _ H-ICDA Code _ ; __
2) . (14-17)
~~
(18-21)
(22-25)
Condition at discharge: dead • 1, alive • 2
Family history of the disease or predisposing ' 6
conditions: Yes • 1 , No • 2 __
If yes, specify _ (27)
7. Date of last previous admission (Month/Day/ Year) _ / __ /
(29-30) (31-32) (33-34)
Final Diagnoses: 1) _ H-ICDA Code _
. (35-38)
(39-42)
Family history of the disease or predisposing (43-46)
conditions: Yes • 1, No • 2
If yes, specify _
(47)
3. Date of last previous admission (Month/Oay/Year)
(49-50)(51-52)(53-54)
Final Diagnoses: 1) H-ICDA Code
2) (55-58)
3) . (59-62)
(63-66)
Family history of the disease or predisposing
conditions? Yes • 1, No • 2
If yes, spec1fy:__
L-l
-------
HOSPITAL RECORD ABSTRACT FORM
Hospital
Abstractor's Name
Date Abstracted
LONG FORM
(CARDIOVASCULAR AND RESPIRATORY DISEASES)
PATIENT'S E.C.I. IDENTIFICATION'NUMBER -1
(1-7)
Pejrsonal/Dgmoigraphlc Data (Write 1n "unknown" 1f 'data unavailable)
•1. Patent's Present Residence:
Town ... - -
(8-9)
Zip Code
(10-14)
Length of Residence (Years)
(15-16)
2. Patient's Previous Residence:
Town & State
(17-18)
Length of Residence (Years)
(19-20)
3. Date of Birth (M6nth/Oay/Year) /
(21-22)(23-24) (25-26)
4. Sex (enter code) M « 1, F *2
(29)
5. Race (ente> code) Cauc • 1, Negro « 2-
Hispanic - 3, Other • 4
(30)
6.. Religion (enter code) Prot » 1. Cath • 2
Jewish • 3, None • 4,
Other • 5 (specify) . (3D
7. Marital Status (enter code) Married • 1, Never
• Married • 2, Widowed • 3, Divorced - 4
(32)
8. Education (enter code) None « 1, Elefti * 2.
3ome~H.S. • 3, Grad H.S. • 4, Tech'/Voc • 5
Grad College - 7, Grad School • 8 :
Insurance Carrier (enter code) UMW « 1, Blue
Cross/Blue Shield » 2, Medicare/ Medicaid • 3
Other • 4 (specify) - ...
(34)
L-2
-------
LONG FORM (cont.)
PATIENT NO.
10. Occupational History
[OCCUPATION (OR DETAILED JOB DES.)| DATES EMPLOYED | NO. YEARS
.PRESENT.
PREVIOUS
PREVIOUS
Comments:
History at expoiurg noc Benclonad abovi:
FROM TO
FROM TO
FROM TO
(36-50)
11. Alcohol Consumption History
PRESENT CONSUMPnni<
(Drinks/Day)
HOW MANY YEARS
PAST CONSUMPTION
(Drinks/Day)
ROW MANY YEARS
BEER
WINE
HARD LIQUOR
CooBtnti:
(S1-6S)
12. Tobacco Use History
CinWENT AKT. USED/DAY
HOW KANT YEARS
PREVIOUS AKT. USED/DAY
HOW MANY YEARS
CICABETTES
CICARS
PIPE
CHEWING TOBACCO
SNUFF
Brand:
Fllcir or non-Miter
(66-80)
L-3
-------
LONG FORM (cont.)
PATIENT ,'IQ. -2
(1-7)
ADMISSION INFORMATION
[•
Directions: Record all admissions in reverse chronological order. For all
•dmissions involving respiratory or cardiovascular disease, complete a
staging form in addition to admission information.
13. - Date of Most Recent Admission (Month/Cay/Year) / /
(8-9) (10-11) (12-13)
final Diagnoses: 1) H-ICDA Code
2) ., (14-17)
31. (1S-21)
(22-25)
Condition at discharge: dead » 1, alive • 2
Family history of the disease or predisposing <26)
conditions: Yes • .1, No « 2
If yes, specify (27)
14. Date of last previous admission (Honth/Day/Year) / /
(29-30) (31-32) (33-34)
Final Diagnoses: 1), ,__H-ICDA Code
2)' » (35-38)
3) ., (39-42)
Family history of the disease or predisposing (43-46)
conditions: Yes » 1, No • 2
If yes, specify
(47)
15. Date of last'previous admission (Month/Oay/Year)
(8-9)(10-11)(12-13)
Final Diagnoses: 1) H-ICDA Code
2) -
(13-21)
(22-25)
Family history of the disease or predisposing
conditions? Yes • 1, No • 2
If yes; specify:
(26)
16. Date of last previous admission (Month/Oay/Year)
(29-30) (31-32) (33-34)
Final Diagnoses: 1) H-ICDA Code
2) (35-38)
3) (3^2'
' (43-46)
Family history of the disease or predisposing
conditions? Yes - 1, No • 2
If yes, specify:
(47)
17. Date of last previous admisstlon (Month/Oay/Year)
(49-50) (51-52) (53-54)
Final Diagnoses: 1) _ H-ICDA Code _
(55-58)
(59-62)
(63-66)
Family history of the disease or predisposing
conditions? Yes • 1 . No • 2
If yes, specify: _ _
TeTT"
L-4
-------
LONG FORM (cont.)
CARDIOVASCULAR DISEASE STAGING FORM
ADMISSION DATE (Mo/Day/Yr)_
(35-36) (37-3(1) (39-40)
CRITERIA
DISEASE ABSENT
DISEASE PRESENT
DISEASE SERIOUS
NYHA Class
Clinical
Diastollc BP
Chest X-Rjy
EKG
Oiaonoses
I
(41)
Nona
(42)
<90
(41)
Normal
(44)
Normal or NSSHWC'S
(41)
None
T«T
II
(47)
Pretlbljl (dent
TwT
90-110
(49)
CT rat(o >1.5
(50)
Any other abnormality
(51)
Congestive failure
TsjT
MI-V
153)
Angina req. medication,
dyjpnea on exertion,
anasarca, arrhythmias
other than PVCs 100
?110 '
(55)
Pleural effusion, Kerley 8
lines, vascular conges-
tion, pulmonary edema
(56)
(Per physician's Inter-
pretation)
(57)
MI, stroke, TIA. gangrene,
claudicatlon, aneurysm
'(581
L-5
-------
LONG FORM (cont.)
PATIENT NO.
(1-7
ADMISSION DATE (Mo/Day/Yr) / /
(8-9) (10-ll) (
RESPIRATORY DISEASE STAGING FORM
ror each of th« criteria below, check those that apply:
CKITERI/1
Clinical
DISEASE ABSENT
No mention of
coughing,
wheezing,
dyspnea or
hemoptysis
DISEASE PRESENT
Occasional wheezing
and/or coughing
DISEASE SESIOUS
Chronic wheeling, chronic or
purulent coughing, chronic
shortness of breath, exertion
dyspnea, hemoptysis
(U)
121)
Chest X-Ray Normal
(15)
Hyperinflation, reduced
diaphragmatic excursion,
inc. retrosternal space,
chronic obstructive air-
way disease, coal
workers' pneumoconiosls
TITT
(Per radiologist's assessment)
Prog, massive flbrcsls, severe
coal workers' disease, adv.
pneumoconlosis incl. sllicosls
lung cancer, cavitating lesions
honeycombing/sever? interstitial
fibrosls.
(29)
PaOj (torr) >80
(16)
65-85
(21)
60
(17)
40-60
(24)
<40
(11)
>2.0
(18)
(2S)
Diagnosis
None
(19)
Diagnosis of emphysema, bron-
chitis, asthma, bronchopneu-
monla, acute Interstitial
pneumonia
(26)
Pulmonary edema, bronchlectasis,
cor pulmonate, lung abscess
lung cancer
(33)
Treatment
None
Amnopfiyllira. Ttdral,
Mom=, antibiotic prn,
IPPB before surgery
T7BT
TT7T
Any sterold-prednlsone, corti-
sone, prednisolone, dexametna-
sone, or other; terbutaline,
Uoproterenol (Ituprtll. cromo-
lyn flntal, Asi-znt>, mucomist,
fsocth/irlnc (.•'.••••i'-.:r.-Z I
T3IT
L-6
-------
CODE MST
EHVIROA'YOi-tlNG GENERAL PATIENT CODE LIST
ENVIRO
i
2
3
4
5
6
7
S
9
10
11
12
1}
14
15
16
17
ia
19
20
21
22
23
24
25
26
27
28
29
30
91
94
97
100
WYCXIIIG GEN.
ENVIRO
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
92
95
98
r
WYOMING GEN.
ENVIRO
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83.
84
85
86
87
88
89
90
93
96
99
WYOMING GEN.
L-7
-------
Appendix M
EFFECTS OF EXPOSURE TO
CARBON DISULFIDE AND HYDROGEN SULFIDE
-------
Table M-l
EFFECTS OF OCCUPATIONAL EXPOSURE TO CARBON DISULFIDE AND HYDROGEN SULFIDE
Age:
No. of Mean or
Workers Range
Concentration Duration:
(mg/cu m)* Mean or
Range
(years)
CS2
H2S
Effects
Ref-
erence
100
43
107
53
32
185 25-35
450-1,000
30-1,500
200-400
62-174
100
125
33
350
116
39
47
22
33
50
31-137
ca 124
40-81
19-65
>62
10
13
2
1-6 6-10
>5
28 44 ca 62
38 51 29-118** ***
100 48 29-118** ***
Polyneuritis in 88%, gas- 31
trie disturbances in 28%
21 Encephalppathy 31
1-9 Ophthalmic pressure. 138/ 68
110, vs 115/87 in controls
ca 5 Eye burning in 96% of 69
rayon-production workers,
44% of cell-fiber workers;
pupillary light reactions
abnormal
Psychomotor and psycho- 63
logic disturbances
Coronary heart disease in 48
5.6%, vs 1.2% in controls
Asthenospermia, hypo- 58
spermia, teratospennia
Periodontal changes 10
Coronary heart disease in 49
16.5%, vs 2.7% in controls
13 Coronary heart disease in 48
3.6%, vs 1.2% in controls
20 Ocular vascular rigidity 66
15 Ophthalmic circulation 65
slowed
Source: NIOSH, Criteria for a Recommended Standard: Occupational Exposure
to Carbon Disulfide, 1977.
M-l
-------
Table M-l (continued)
No. of
Workers
Age:
Mean or
Range
Concentration
(mg/cu m)*
CS2 H2S
Duration:
Mean or
Range Effects
(years)
Ref-
erence
118
343
45
29-118** ***
29-118** ***
15
11
343
45
29-118** ***
11
319
45
29-118** ***
322
45
29-118** ***
343 25-72
29-118** ***
Polyneuropathy, abnormal 56
EEC's
Angina in 17%, vs 11% in 8
controls; blood pressure
140/91, vs 136/85 in con-
trols; coronary heart
disease cause of 52% of
deaths, vs 31.7% nationally
Fasting glucose levels 43
increased with longer ex-
posures; plasma glucose
levels higher than in
controls
Coronary heart disease 44
mortality 5.6 times that
in controls; total mor-
tality 2.7 times controls
Coronary heart disease 45
more frequent than in con-
trols: fatal infarctions
4.8, total infarctions
3.7, nonfatal infractions
2.8, angina 2.2, "coronary
ECG's" 1.4 times higher than
controls
Life expectancy decreased
0.9-2.1 years, depending
on age, during 8-year
followup
47
M-2
-------
Table M-l (continued)
Age:
No. of Mean or
Workers Range
Concentration
(mg/cu m)*
Duration:
Mean or
Range Effects
Ref-
erence
(years)
CS2
H2S
36
165
138
94
189
42
397 35-64
53
630 20-40
<50
18
<30
29-118** ***
29-118** ***
<59** ***
31-50
22-44
12-31
ca 28
>6 Peripheral nerve and CNS 55
damage; conduction velo-
cities slowed; EMG's
abnormal
Coronary heart disease 41
cause of 42% of deaths in
highly exposed workers,
24% in moderately exposed,
14% nationally
<10 Coronary mortality during 17
8-year followup 5.8%, vs
2.6% in controls; total
mortality 10.2% vs 6.7%
in controls
Immunologic reactions de- 75
creased; job absenteeism
increased
<10 Arteriosclerotic changes 33
in 30.4%, hypertension in
23.2%
<1 Hypotension, nervous sys- 54
tern excitability
>3 Spontaneous abortions in 61
14.3%, vs 6.8% in con-
trols; premature births
in 8.6%, vs 2.8% in con-
trols
M-3
-------
Table M-l (continued)
Age:
No. of Mean or
Workers Range
Concentration Duration:
(mg/cu m)* Mean or
Range
(years)
CS2
H2S
Effects
Ref-
erence
209 20-40
60
25
500 18-60
500 20-40
94
18
>22
ca 16
<9
3-9
Menstruation Irregular, 60
painful, abundant, and
prolonged
Muscular power diminished, 57
reflexes slowed
1-30 Retinal degeneration, 67
conjunctival inflammation,
temporary corneal opaci-
ties, color-vision dis-
turbances
Menstruation abundant, 60
painful, prolonged
<1 Hypotension; nervous sys- 54
tern excitability
*1mg/cu m = 0.321 ppm
**Thes« studies are based on the same cohort of workers, exposed to carbon
disulfide plus hydrogen sulfide at concentrations averaging 29-88 mg/cu m
irt the 1960's, 59-118 mg/cu m in the 1950's, and higher before 1950.
***Hydrogen sulfide concentrations are included in those given for carbon
disulfide and were estimated to be about 10% of the total.
M-4
-------
Table M-2
EFFECTS OF EXPOSURE TO CARBON DISULFIDE
OR TO CARBON DISULFIDE PLUS HYDROGEN SULFIDE ON ANIMALS
Route of
Exposure
Exposure
Species Concentration*
Exposure
Duration
Effects
CS2
H2S
12
**
1.0 0.1
1.0 0
0.1 0.1
0.1 0
70-110 d
before
mating and
during
pregnancy
160 d
Ref-
erence
Inhalation Rat 2,330
2,330
11 " 2,000
" " 1,500
0 6 hr/d
5 d/wk
10 wk;
then
3 d/wk
12 wk
0 6 hr/d
5 d/wk
2-5 wk
0 2 hr/d
throughout
pregnancy
0 5 hr/d
6 d/wk
1-15 mon
Lethargy, loss of
motor control,
slowed MCV's with
no recovery in 12
wk
Lethargy, slowed
but reversible
MCV's
Increased fetal
mortality, de-
creased fertility
Weakness, paraly-
sis, myelln and
neuron degenera-
tion, weight loss
78
78
81
79
Increased fetal
mortality, terata
Inflammation of
bronchi, weight
changes, increased
serum asparatate
aminotransferase
and blood cholin-
esterase activi-
ties; most severe
with combined ex-
posures
88
85
Source; NIOSH, Criteria for a Recommended Standard: Occupational Exposure
to Carbon Disulfide, 1977.
M-5
-------
Table M-2 (continued)
Route of
Exposute
Species
Exposure
Concentration*
Exposure
Duration
Effects
Ref-
erence
CS2
H2S
Inhalation. Mouse 2,000 0
11 Rabbit 780- 0
2,330
2 hr/d
throughout
pregnancy
6 hr/d
5 d/wk
38 wk
Increased fetal
mortality, de-
creased fertility
Paralysis, CNS
damage, slight
liver damage,
weight loss
81
84
" , " 930 140 30 min/d Abnormalities of 86,
930 0 120 d bone marrow, kid- 87
0 140 neys, spleen; de-
creased spermato-
;' genesis, loss of
appetite, blood
changes; most se-
vere with combined
exposure
ip . Rat 78 0 4 mon Testicular lesions, 80
(every no spermatogenesis
other d)
" " 78 0 2 mon Decreased number 80
(every of spermatozoa;
other d) blood vessels en-
gorged, walls
thickened
" " 39 0 " No effects 80
^Concentration given in mg/cu m for inhalation exposures, mg/kg for injec-
tions; 1 mg/cu m - 0.321 ppm
**Hydrogen sulfide concentration included in that for carbon dlsulfide
M-6
-------
BIBLIOGRAPHY ON EFFECTS OF OCCUPATIONAL EXPOSURE
"TO CARBON DISULFIDE OR HYDROGEN SULFIDE
(Including References Mentioned in Tables M-l and M-2)
1. Properties and Essential Information for Safe Handling and Use of
Carbon Disulfide, Chemical Safety Data Sheet SD-12, rev. Washington
DC, Manufacturing Chemists' Association, Inc, 1967, 16 pp
2. Weast RC (ed): Handbook of Chemistry and Physics, ed 55. Cleveland,
Chemical Rubber Co, 1974, p C-234
\
3. Folkins HO: Carbon dlsulfide, in Kirk-Othmer Encyclopedia of
Chemical Technology, ed 2 rev. New York, Interscience Publishers,
1964, vol 4, pp 370-85
4. Synthetic Organic Chemicals—United States Production and Sales of
Miscellaneous Chemicals, 1974 (Preliminary). US Dept of Commerce, US
International Trade Commission, 1976, p 9 '
5. Klapproth E: Carbon disulfide, in Chemical Economics Handbook.
Menlo Park, Calif, Stanford Research Institute, 1976, pp 625.5030A to
625.5030E
6.. Plant observation reports and evaluations for carbon disulfide.
1 Menlo Park, Calif, Stanford Research Institute, February, 1977, 46 pp
(submitted to NIOSH under Contract No. CDC-99-74-31)
7. Geiger E: [Permissible limits for carbon disulfide and hydrogen
sulfide in work rooms under special consideration of the viscose
rayon industry.] Text Rundsch 7:185-87, 1952 (Ger)
8. Hernberg S, Partanen T, Nordman CH, Sumari P: Coronary heart disease
among workers exposed to carbon disulphide. Br J Ind Med 27:313-25,
1970
9. Rosensteel RE, Shama SK, Flesch JP: FMC Corporation—American
Viscose Division—Nitro, West Virginia, Health Hazard Evaluation
Determination report No. 72-21-91. Cincinnati, US Depi: of Health,
Education, and Welfare, National Institute for Occupational Safety
and Health, Hazard Evaluation Services Branch, Division of Technical
Services, 1973, 40 pp
10. Gondzik W, Jarzynka W, Pinska E: [The examinations of oral cavity
and teeth in man-made fibers plant workers exposed to carbon
disulfide.] Med Pracy 20:78-83, 1969 (Pol)
11. Gafafer WM (ed): Occupational Diseases—A Guide to Their
.Recognition. Public Health Service bulletin No. 1097. US Dept of
Health, Education, and Welfare, Public Health Service, 1964, pp 107-
08
M-7
-------
BIBLIOGRAPHY (continued)
12. Simpson JY: Notes on the anesthetic effects of chloride of
hydrocarbon, nitrate of ethyle, benzln, aldehyde, and bisulphuret of
carbon. Mon J Med Sci 8:740-45, 1848
13. Survey of Carbon Disulphide and Hydrogen Sulphide Hazards in the
Viscose Rayon Industry, bulletin No. 46. Harrisburg, Pa, Dept of
Labor and Industry, Occupational Disease Prevention Division, 1938,
69 pp
14. Hamilton A: Carbon disulphid, CS2, in Industrial Poisons in the
United States. New York, Macmillan, 1925, pp 360-70
15. Delpech A: [Incidents which develop in rubber workers-r-Inhalation of
carbon disulfide vapors.] Union Med 10:265, 1856 (Fre)
16. Bryce Til: VII—Synopsis of a case of chronic poisoning by bisulphide
of carbon. Edinburgh Med J 32:140-41, 1886
17. Piorry M: [Intoxication by carbon disulfide.] Gaz Hop 61:241, 1858
(Fre)
18. Ross J: Two cases of chronic poisoning by bisulphide of carbon. Ked
Chron 5:257-69, 1886
19. Indiarubber—Dangers incidental to the use of bisulphide of carbon
and naphtha, in Oliver T (ed): Dangerous Trades. New York, EP
Button and Co, 1902, pp 470-74
20. Foreman W: Notes of a fatal case of poisoning by bisulphide of
carbon. Lancet 2:118-19, 1886
21. Peterson F: Three cases of acute mania from inhaling carbon
bisulphide. Boston Med Surg J 127:325-26, 1892
22. Bard CL: Cases In practice—Malignant pustule—And insanity due to
bisulphide of carbon. South Cal Pract 7:476-85, 1892
23. Hamilton A: Industrial Poisons Used in the Rubber Industry, bulletin
No. 179. US Dept of Labor, Bureau of Labor Statistics, 1915, pp 5-64
24. Jump HD, Cruice JM: Chronic poisoning from bisulphide of carbon.
Univ Pa Med Bull 17:193-96, 1904
25. Francine AP: Acute carbon bisulfid poisoning. Am Med 9:871, 1905
26. Hamilton A: The making of artificial silk in the United States and
some of- the dangers attending it, in Discussion of Industrial
Accidents and Diseases, bulletin No. 10. US Dept of Labor, Division
of Labor Standards, 1937, pp 151-60
M-8
-------
I
I
I . • .
! BIBLIOGRAPHY (continued)
27. Rubin HH, Arieff AJ: Carbon disulfide and hydrogen sulfide clinical
study of chronic low-grade exposures. J Ind Hyg Toxicol 27:123-29,
1945
i
28. Barthelemy HL: Ten years' experience with industrial hygiene in
j connection with the manufacture of viscose rayon. J Ind Hyg Toxicol
^ 21:141-51, 1939
i 29. Carbon disulfide—Its toxicity and potential dangers. Public Health
! Rep 56:574-81, 1941
30. American Standards Association: Allowable Concentration of Carbon
! Disulfide, ASA Z37.3-1941. New York, ASA, 1941, 7 pp
| 31. Viglianl EC: Carbon dlsulphlde poisoning in viscose rayon factories.
Br J Ind Med 11:235-44, 1954
i 32. Lilis R, Gavrilescu N, Roventa A, Dumitriu C, Nestorescu B: [Kidney
function studies (plasma flow) of workers with a history of long-term
exposure to carbon disulfide.] Stud Cercet Ig Sanat Publica, pp 95-
104, 1967 (Rum)
I
33. Gavrilescu N, Lilis R: Cardiovascular effects of long-extended
carbon disulfide exposure, in Brieger H, Teisinger J (eds):
Toxicology of Carbon Disulfide. Amsterdam, Excerpta Medica
Foundation, 1967, pp 165-67
34. Ehrhardt W: Experiences with the employment of women exposed to
carbon disulfide, in Brieger H, Teisinger J (eds): Toxicology of
i Carbon Disulphlde. Amsterdam, Excerpta Medica Foundation, 1967, p
; 240
I
; 35. Finkova A, Simko A, Jindrichova J, Kovarik J, Preiningerova 0,
! Kllmova A, Korisko F: [Gynecologic problems of women working in an
' environment contaminated with carbon disulfide.] Cesk Gynekol
38:535-36, 1973 (Cze)
36. Jindrichova J: [Health status of workers in the cord fiber
industry.] Prac Lek 9:10-17, 1957 (Cze)
1 37. Baranowska B: [Evaluation of the skin as an absorption route for
1 carbon disulfide.] Int Arch Gewerbepath Gewerbehyg 21:362-68, 1965
(Ger)
38. Soucek B, Pavelkova E: [Absorption, metabolism and effects of carbon
disulfide in the organism—Part IV—Long-term experiments (Czech).]
Prac Lek 5:181-91, 1953 (Cze)
j 39. Harashima S, Masuda Y: Quantitative determination of absorption and
| elimination of carbon disulfide through different channels in human
body. Int Arch Gewerbepath Gewerbehyg 19:263-69, 1962
M-9
-------
BIBLIOGRAPHY (continued)
40. Mack T, Freundt KJ, Henschler D: Inhibition of oxidative N-
demethylation in man by low doses of inhaled carbon disulphide.
Biochem Pharmacol 23:607-14, 1974
41. Tiller JR, Schilling RSF, Morris JN: Occupational toxic factor in
mortality from coronary heart disease. Br Med J 4:407-11, 1968
42. Jacobs MB: Carbon disulfide, in The Analytical Chemistry of
Industrial Hazards, Poisons, and Solvents, ed 2. New York,
Interscience Publishers, 1949, vol 1, pp 323-26
43. Hernberg S, Nordman CH, Partanen T, Christiansen V, Virkola P: Blood
lipids, glucose tolerance and plasma creatlnine in workers exposed to
carbon disulphide. t^ork Environ Health 8:11-16, 1971
44. Hernberg S, Nurmi-nen M, Tolonen M: Excess mortality from coronary
heart disease in viscose rayon workers exposed to carbon disulfide.
Work Environ Health 10:93-99, 1973
45. Tolonen M, Hernberg S, Nurminen M, Tiitola K: A follow-up study of
coronary heart disease in viscose rayon workers exposed to carbon
disulphide. Br J Ind Med 32:1-10, 1975
46. Hernberg S, Tolonen M, Nurminen M: Eight-year follow-up of viscose
rayon workers exposed to carbon disulfide. Work Environ Health 2:27-
30, 1976
47. Nurminen M: Survival experience of a cohort of carbon disulphide
exposed workers from an eight-year prospective follow-up period. Int
J Epidemiol 5:179-85, 1976
48. Cirla AM, Villa A, Tomasini M: [Investigation of the incidence of
coronary disease in workers exposed to carbon disulfide in a viscose-
rayon industry.] Med Lav 63:431-41, 1972 (Ita)
49. Locati G, Cirla AM, Villa A: Prevalence of coronary heart disease
among 253 viscose rayon workers admitted to the Cllnica del Lavoro of
Milan from 1947 to 1969. Med Lav 61:442-46, 1970
50. Goldberger E: Heart Disease—Its Diagnosis and Treatment, ed 2.
Philadelphia, Lea and Febiger, 1955, pp 597-608
51. Lieben J, Menduke H, Flegel EE, Smith F: Cardiovascular effects of
CS2 exposure. J Occup Med 16:449-53, 1974
52. Salant W, Kleitman N: Pharmacological studies on acetone. J
Pharmacol Exp Ther 19:293-306, 1922
53. Dautov FF: [Health measures to improve work conditions during on-
the-spot industrial training of students-operators of an occupational
technical school preparing the attendant staff for petrochemical
industries.) Gig Tr Prof Zabol 6:8-11, 1971 (Rus)
M-10
-------
BIBLIOGRAPHY (continued)
54. Kramarenko IB, Yakovleva IN, Grishko FI, Litvinova YA: Hygienic
evaluation of vocational training of young workers in spinning shops
of a viscose mill. Hyg Sanit 36:379-84, 1971
55. Seppalainen AM, Tolonen M, Karli P, Hanninen H, Hernberg S:
Neurophysiological findings in chronic carbon disulfide poisoning—A
descriptive study. Work Environ Health 9:71-75, 1972
56. Seppalainen AM, Tolonen MT: Neurotoxicity of long-term exposure to
carbon disulfide in the viscose rayon industry—A neurophysiological
study. Work Environ Health 11:145-53, 1974
57. Vasilescu C: Motor nerve conduction velocity and electronyogram in
carbon disulphide poisoning. Rev Roum Neurol 9:63-71, 1972
58. Lancranjan I, Popescu HI, Klepsch I: Changes of the gonadic function
in chronic carbon disulphide poisoning. Med Lav 60:566-71, 1969
59. Lancranjan I: Alterations of spermatic liquid in patients
chronically poisoned by carbon disulphide. Med Lav 63:29-33, 1972
60. Vasilyeva IA: [Effect of low concentrations of carbon disulfide and
hydrogen sulfide on the menstrual function of women and the estrual
cycle in an experiment.] Gig Sanit 7:24-27, 1973 (P.us)
61. Petrov MV: [Some data on the course and termination of pregnancy in
female workers of the viscose industry.] Pediatr Akush Ginekol 3:50-
52, 1969 (Rus)
62. Bezvershenko AS: [Some data on the functional condition of the
sexual glands in female workers subjected to the influence of carbon
disulfide.] Gig Truda 17:191-95, 1965 (Rus)
63. Hanninen H: Psychological picture of manifest and latent carbon
disulphide poisoning. Br J Ind Med 28:374-81, 1971
64. Tuttle TC, Wood GD, Grether CB: Behavioral and Neurological
Evaluation of Workers Exposed to Carbon Disulfide (CS2). Unpublished
report submitted to NIOSH by Westinghouse Electric Corporation,
Behavioral Services Center, Columbia, Md, 1976, 156 pp
65. Raitta C, Tolonen M, Nurminen M: Microcirculation of ocular fundus
in viscose rayon workers exposed to carbon disulfide. Albrecht von
Graefes Arch Klin Exp Ophthalmol 191:151-64, 1974
66. Raitta C, Tolonen M: Ocular pulse wave in workers exposed to carbon
disulfide. Albrecht von Graefes Arch Klin Exp Ophthalmol 195:149-54,
1975
67. Szymankowa G: [Observations on the effects of carbon disulfide on
vision in workers engaged in the manufacture of synthetic fibers.]
Klin Oczna 38:41-44, 1968 (Pol)
M-ll
-------
BIBLIOGRAPHY (continued)
68. Maugeri U, Cavalier! A, Visconti E: [Ophthalmodynamography in
occupational carbon disulfide poisoning.] Med Lav 57:730-40, 1966
(Ita)
69. Savic SM: , Influence of carbon disulfide on the eye. Arch Environ
Health 14:325-26, 1967
70.' Goto S, Hotta R: The medical and hygienic prevention of carbon
disulphide poisoning in Japan, in Brieger H, Teisinger J (eds):
Toxicology of Carbon. Disulphide. Amsterdam, Excerpta Medica
Foundation^ 1967, pp 219-30
71. Goto S, Hotta R, Sugimotd K: Studies on chronic carbon disulfide
poisoning—Pathogeriesis of retinal microarteurysm due to carbon
disulfide, with special reference to a subclinical defect of
carbohydrate metabolism. Int Arch Arbeitsmed 28:115-26, 1971
72. Goto S, Suglmbto K, Hotta R, Fujioka Y, Graovac-Leposavic L, Savic
SM, Jovicic M: Retinal microaneurysm in carbon disulfide workers in
Yugoslavia. . Prac Lek 24:66-70, 1972
73. Hotta R, Sugimoto K, Goto S: [Retlnopathia sulfocarbonica and its
natural history.) Acta Soc Ophthalmol Jpn 76:1561-66, 1972 (Jpn)
74. Toldnen M: Chronic subclinical carbon disulfide poisoning. Work
Environ Health 11:154-61, 1974
i
75. Kashin LM: Overall imniuriological reactivity and morbidity of workers
exposed to carbon disulfide. Hyg Sanlt 30:3317-35, 1965
• . ' • ., .. . i .
76.' Joffe VI: [Some results of the study of general immunological
reactivity of the organism in its clinical and epidemiological
aspects.] Sbornik Trudov Mezhinstitutski Nauchnoi Konferentsii 3:21-
36, 1954 (Rus)
i
77. Mancuso TF, Locke BZ: Carbon disulphide as a cause of suicide—
Epidemiological study of viscose rayon workers". J Occup Med 14:595-
606, 1972
78. Seppalainen AM, Linnoila I: Electrophyslological findings in rats
With experimental" carbon disulphide neuropathy. Neuropathol Appl
Neurobiol 2:209-16, 1976
79. Szendzlkowskl S, Stetklewicz J, Wronska-Nofer T, Zdrajkowska I:
Structural aspects of experimental carbon disulfide neuropathy—I.
Development of neurohistological changes in chronically intoxicated
rats. Int Arch Arbeitsmed 31:135-49, 1973
80. Gondzik M: His'to'logy and histochemistry of rat testicles as affected
' by carbon disulfide.' Pol Med J 10:133-39, 1971
M-12
-------
BIBLIOGRAPHY (continued)
81. Yaroslavskiy VK: [Toxic effect of carbon disulfide on the
reproductive function and intensification of the effect of
tryptophan.] Byull Eksp Biol Med 68:88-91, 1969 (Rus)
82. Petrun KM: [Effect of carbon disulfide on certain biochemical
parameters of the state of the organism when entering the body
through the skin.] Gig Tr Prof Zabol 11:50-53, 1967 (Rus)
83. Cohen AE, Paulus HJ, Keenan RG, Scheel LD: Skin absorption of carbon
disulfide vapor in rabbits—I. Associated changes in blood protein
and zinc. AMA Arch Ind Health 17:164-69, 1958
84. Cohen AE, Scheel LD, Kopp JF, Stockell FR, Keenan RG, Mountain JT,
Paulus HJ: Biochemical mechanisms in chronic carbon disulfide
poisoning. Am Ind Hyg Assoc J 20:303-23, 1959
85. Misiakiewicz Z, Szulinska G, Chyba A: [Effect of the mixture of
carbon disulfide and hydrogen sulfide in air on white rats under
conditions of continuous exposure for several months.] Rocz Panstw
Zakl Hig 23:465-75, 1972 (Pol)
86. Wakatsuki T, Hlgashikawa H: [Experimental studies on CS2 and H2S
poisoning—The histological changes in hematopoietic organs and .other
main internal organs.] Shikoku Igaku Zasshi 14:549-54, 1959 (Jpn)
87. Wakatsuki T: [Experimental study on the poisoning by carbon
disulphide and hydrogen sulphide.] Shikoku Igaku Zasshi 15:671-700,
1959 (Jpn)
88. Barilyak IR, Vasilyeva IA, Kalinovskaya LI: [Effect of small
concentrations of carbondisulfide and hydrogensulfide on intrauterlne
development in rats.] Arkh Anat Gistol Etnbriol 68:77-81, 1975 (Rus)
89. Flesch JP, Lucas JB: Olin Corporation—Film Division—Pisgah Forest,
North Carolina, Health Hazard Evaluation Determination report No.
73-8-132. Cincinnati, US Dept of Health, Education, and Welfare,
National Institute for Occupational Safety and Health, Hazard
Evaluation Services Branch, 1974, 10 pp
90. Oppl L: [Methods for prevention of carbon disulfide pollution of the
air in the production of synthetic fibers] in Brieger H, Teisinger J
(eds): Toxicology of Carbon Disulphide. Amsterdam, Excerpta Medica
Foundation, 1967, pp 245-48 (Fre)
91. Viles FJ: Field determinations of carbon disulfide in air. J Ind
Hyg.Toxicol 22:188-96, 1940
92. Reece GM, White B, Drinker P: Determination and recording of carbon
disulfide and hydrogen sulfide in the vicose-rayon industry. J Ind
Hyg Toxicol 22:416-24, 1940
M-13
-------
Proceedings of the
Seventeenth Annual Hanford Biology Symposium
at Richland, Washington, October 17-19, 1977
Editors D. Dennis Mahlum Patricia L. Mackeft
Melvin R. Sikov Floyd D. Andrew
Sponsored by
Division of Biomedical and Environmental Research
Department of Energy
and
Pacific Northwest Laboratories
Battelle Memorial Institute
There is a growing awareness that prenatal or neo-
natal mammals are often more sensitive to chemical
and physical agents than adults. Strongly associated
with developing technologies for new or enhanced
energy sources are the prospects of dangerous en-
vironmental pollutants. Developing these energy tech-
nologies in ways that contain the spread of pollutants
requires a knowledge of the effects of such pollutants
on mammals. To disseminate this knowledge, several
recent conferences have dealt with the toxicology of
energy-related pollutants, with emphasis on the adult
mammal. Developmental Toxicology of Energy-
Related Pollutants focuses on the effects of such
pollutants on the immature mammal, systematically
considering viewpoints in many disciplines. Pollutants
discussed include radionuclides and external ionizing
radiation, electric and magnetic fields, microwaves,
heavy metals, hydrdcarbons, and aromatic nitrogen
compounds. Methods are presented for assessing
damage and for improving the extrapolation from
values obtained with mammals to the human. One
topic of current high visibility, illustrating the strong
interdisciplinary nature of the participation, was the
panel discussion "Women in the Workplace," which
brought out many facets in the protection of the
unborn. The conference left the participants with
deepened insights, broadened perspective, and an
awareness of information still needed—qualities
that will extend, it is believed, to readers of the
proceedings as well.
Published as part of the DOE Symposium Series
by the Technical Information Center,
U. S. Department of Energy, 1978
660 pages, 6 by 9 in., paper binding
Library of Congress Catalog Card
Number 78-606139 (CIP)
Available as CONF-771017 for $12.50
from the National Technical Information Service
(see order form)
Gonads and Gametes
Vulnerability of Female Germ Cells in Developing Mice and Monkeys to Tritium,
Gamma Rays, and Polycyclic Aromatic Hydrocarbons
Genetic Differences in Polycyclic-Aromatic-Hydrocarbon Metabolism and Their Effects
on Oocyte Killing in Developing Mice
Depressed Food Intake and Reduced Reproduction in the Japanese Quail After a
Single Dose of Prudhoe Bay Crude Oil
Altered RNA Metabolism in Cultured Mammalian Cells Exposed to Low Levels of
Cadmium2'1': Correlation of the Effects with Cadmium2 + Uptake and Intracellular
Distribution
Interspecific Comparison of the Effects of Continuous Ionizing Radiation on the
Primitive Mammalian Stem Germ Cell
Effects of Methyl Methanesulfonate and Dimethylnitrosamine on Spsrm
Production in Syrian Hamsters
DNA Variability of Abalone Sperm as a Measure of Copper Toxicity
Developmental Aspects of Epoxide-Metabolizing Enzyme Activities in Adrenals,
Ovaries, and Testes of the Rat
Entry of Circulating Mutagens into the Rete Testis Fluid of Rats
Organic Pollutants
Developmental Toxicology of Organic Pollutants
Radiosensitivity of the Mouse Embryo and Dominant Lethal Mutation Induction by
Ionizing Radiation or Methyl Methanesulfonate Analyzed In Vitro
Inhibition of Mouse Embryo Development In Vitro by Benzolalpyrene
Characteristics of a New Approach in Teratology: In Vitro Differentiation of Early Mouse
Embryos Beyond the Preimplantation Period After Cyclophosphamide Treatment In Vivo
Application of a Sensitive In Vivo Teratological System to the Testing of Benzolalpyrene
Transplacental Toxicology of the Polychlorinated and Polybrominated Biphenyls
Radiations—External
Prenatal Cobalt-60 Irradiation Effects on Early Postnatal Development of the Squirrel
Monkey Offspring
Effects of Low-Dose X-Radiation on the Matrix Cells in the Telencephalon of Mouse
Embryos
Adrenal Activity of Adult Rats Irradiated with 150 rads During Fetal Life
Hematopoietic Recovery of Pigs After Chronic Prenatal Irradiation
Reproductive Function in Mice Exposed to Ancestral and Direct Irradiation
Neutron Irradiation of Rat Embryos In Utero
Adult Behavioral Deficit in Rats Exposed Prenatally to 918-MHz Microwaves
Influence of Microwave Exposure on Functional Maturation of the Rat
Responses of Domestic Chicks to 60-Hz Electrostatic Fields
Mediations—Internal
Age-Related Variation in Thyroidal Exposures from Fission-Produced Radioiodines
Distribution of2 39 Pu in Neonatal Beagles
Late Effects of Perinatally Administered Plutonium
Methodology and Human Implications
Developmental Toxicity Evaluations: An Overview of Current Animal Models and
Methods
Toxic Effects of Metals on DNA Synthesis
A Multidisciplinary Approach to the Assessment of Developmental Toxicity
Associated with Chronic Lead Exposure
Behavioral Assessment in the Developmental Toxicology of Energy-Related
Industrial Pollutants
Developmental Toxicology as Input to the Methodology for Human Studies of
Energy-Related Pollutants
Women in the Workplace: A Panel Discussion
Some Problems of Women in the Workplace: An Introduction
Women in the Workplace
Reproductive Hazards on the Job: A Holistic Perspective
Considerations for Protecting the Unborn
Legal Implications
Inorganic Pollutants
Teratogenic Potential of Sulfur Dioxide and Carbon Monoxide in Mice and Rabbits
Effect of Carbon Monoxide on Glucose Metabolism and Growth of Rat Embryos
Contrasting Embryopathy Produced by Acute High and Chronic Low Doses of
Methyl Mercury
Lead Distribution and Effects During Development in the Rat
Effects of Lead Exposure on the Developing Rat Parietal Cortex
Arsenic as a Teratogen
Absorption of Cadmium in the Newborn and Juvenile Guinea Pig
Use of Perfusion in the Study of Placental Transport of Mercury and Cadmium
Strain and Tissue Differences in Cadmium-Binding Protein in Cadmium-Treated Mice
Cadmium Toxicity in the Pregnant Rat
Cadmium Toxicity in Preimplantation Mouse Embryo Development
Cadmium-Induced Teratogenicity and Embryotoxicity in the Mouse
Prenatal Cadmium Exposure: Effects on Essential Trace Metals and Behavior in Rats
Author Index
Subject Index
-------
Send order and remittance to:
National Technical Information Service
U. S. Department of Commerce
Springfield, Virginia 22161
(Purchasers outside the North American continent
should write NTIS for prices.)
Enclosed is $_
.for
No. of
copies
$12.50
Price
per copy
Developmental Toxicology of Energy-Related Pollutants
Title
CONF-771017
Report No.
DCheck DNTIS Deposit Account No.
D American Express Card Account No.
Signature
Card Expiration Date
Name.
Address.
>
0
rr
UJ
2
UJ
u.
4° c«
2 UJ E
"- ^ UJ
§H^
loc 5
.§< D
-O O- rn
ago
» •%
S9
g *. o s-
•t c w to
4 | §C
•8 e s s
5^-50,
8*|-E
<* Q C ^
-ZIP-
8
£
CD
C
;o
^
sdioisotopas in tAot
1968, $6.00
oc
0
m
o
_
1 1
0
U
1
is
ns of the Nuel
.9
€
U
1
1
'§"
5
00
<0
n
S
_m
1
4
•o
§8
n d
£5
J!§
*• en
1969, $6.00
adiation Biology o1
(CONF-690501), 1
OC
#
8
CD
V)
o s
&I
1-1
f— (0
11
inesis (CONF-E
isorders of An
970, $9.00
ihalation Carcinoga
yeloproliferative D
(CONF-680529), 1
i S
*
fi
O
UJ
•o
c
ra
0)
1
es: Radiation I
970, $6.00
ledical Radionuclid
(CONF-691212), 1
5
.a
i
s
Carcinc
o
S
'&
Is
"3 CD
g«»
go
e r*
'C 0)
!orphology of Exps
(CONF-700501 ),1
5
8
CO
0
8; 2
-. 3
5&
CO UJ
8 E
" =
LL '*
Z 0
o.|
y§
0*8
®CCD
= •8**.
g>g£
•5 o en
•ecipitation Scaven
iomedical Implicati
(CONF-710201 ),1
Q. Off
1
£
1
C
Q
.2
ra
IS
30
— w
SCN-
onuclides: Crh
:-711101), 19:
linical Uses of Radi
Techniques (CONF
0
o
m °
<2 CD
C>
co"
r~ -
en n o
«- r~ ID
_'2co
in ..-
O —
in o .
o .- M-
CN m i—
^82
u. £: *^
agenesis (CON
phere (CONF-;
ONF-730505),
adionuclide Carcini
arbon and the Bios
harmal Ecology (C1
K(jt-
*
in
8
m
r^.
li.
Z
O
y
>
Ari
Frimuni
lignancy and 1
c
•§
S«
"5
SrS
e en
t~
*
in
r~
en
8
n
r^
u.
O
U
(A
o
obas and Probl
&.
t/i
i
c
a
Is
l£
s
O)
f-
CM
CM
11 .
1 o ii
c O O
£ 3 UJ
740513),
LL
z
o
CJ
osy stems
u
UJ
c
Mineral Cycling in Southeaster
1975, $13. 60
en
o
CO
8
^r
r^-
Li_
Z
o
u
I
Radiation and the Lymphatic !
$9.00
$
0
•o 8
S<2
g
S<6
Ife
6"-
^ ^T
•*; CN
28
Atmosphere— Surface Exchang
Pollutants (1974) (CONF-74
Evaluation
6.75
c**
c
15
>o
00
Environmantal Chemistry and
(CONF-760429), 1978, $15.
JS
§
S
T^
r^
li.
Z
o
CJ
c
O
1
Scientific Problems of Coal Uti
$9.00
c
+*
J3
1
•o
8
ra
CC
k
io
uj in
Developmental Toxicology of
(CONF-771017), 1978, $12.
*
i
------- |