EPA-650/1-74-004
May 1974
HEALTH CONSEQUENCES
OF SULFUR OXIDES:
A Report
from
CHESS, 1970-1971
U S Environmental Protection Agency
Office of Research and Development
Notional Environmental Research Center
Research Triangle Park N C 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, Environmental Protection
Agency, have been grouped into five series. These five broad categories were estab-
lished to facilitate further development and application of environmental technology.
Elimination of traditional grouping was consciously planned to foster technology transfer
and a maximum interface in related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5 . Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL HEALTH EFFECTS RESEARCH
series. This series describes projects and studies relating to the tolerances of man
for unhealthful substances or conditions. This work is generally assessed from a
medical viewpoint, including physiological or psychological studies. In addition to
toxicology and other medical specialties, study areas include biomedical instrumen-
tation and health research techniques utilizing animals - but always with intended
application to human health measures.
DISTRIBUTION STATEMENT
Document is available to the public for sale through the Superintendent of
Documents, U.S. Government Printing Office, Washington, D.C. 20402.
-------
EPA-650/1-74-004
HEALTH CONSEQUENCES
OF SULFUR OXIDES:
A Report from CHESS, 1970-1971
Human Studies Laboratory
Program Element 1A1005
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
NATIONAL ENVIRONMENTAL RESEARCH CENTER
RESEARCH TRIANGLE PARK/NORTH CAROLINA 27711
May 1974
-------
EPA REVIEW NOTICE
This report has been reviewed by the Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
11
-------
ABSTRACT
Epidemologic studies of the U.S. Environmental
Protection Agency's Community Health and Environ-
mental Surveillance System (CHESS) program pro-
vide dose-response information relating short-term
and long-term pollutant exposures to adverse health
effects. This report presents results of studies con-
ducted in CHESS communities in New York and the
Salt Lake Basin during 1970-1971. In addition,
studies conducted in Idaho-Montana, Chicago, and
Cincinnati, in which health indicators similar to those
used in CHESS were employed, are included. In this
report, attention is focused on the health effects
associated with sulfur oxides, but the relative con-
tribution of various air pollutants, especially sulfur
dioxide, total suspended particulates, and suspended
sulfates, to observed disease frequencies is considered.
Health indicators of long-term pollution effects em-
ployed in these studies included increased prevalence
of chronic bronchitis in adults, increased acute lower
respiratory infections in children, increased acute
respiratory illness in families, and subtle decreases in
ventilatory function of children. Health indicators for
short-term pollution effects were aggravation of
cardiopulmonary symptoms and of asthma. Thresh-
old estimates were developed for the effects of the
pollutants considered. These estimates support
existing National Primary Air Quality Standards for
long-term exposures, insofar as these standards could
be assessed in terms of the health indicators men-
tioned. With regard to short-term exposures, the
studies indicated that adverse effects were being
experienced even on days below the National Primary
Standard for 24-hour levels of sulfur dioxide and
total suspended particulates. These adverse health
effects, however, appear to be associated with
suspended sulfate levels rather than to the observed
concentrations of sulfur dioxide and total suspended
particulates, as evidenced by the consistency of the
relationship between symptom aggravation and sul-
fate levels and the lack of consistency for this
relationship with other pollutants.
ill
-------
PREFACE
This is a report on a series of extensive epidemi-
ologic investigations on the health consequences of
sulfur oxides being carried out under the Community
Health and Environmental Surveillance System
(CHESS) conducted by the Environmental Protection
Agency in cooperation with local and state govern-
ments, universities, and private research organi-
zations.
The studies are presented in monograph form in
order to provide the opportunity to assess the
consistency of the findings from multiple studies. The
health effects of pollution exposure may often be
subtle rather than overt, and a single study may only
suggest an adverse health effect. A series of studies
showing a consistent pattern, however, will provide
evidence to incriminate or absolve exposures to
various levels of given pollutants as antagonistic to
health. At the expense of some duplication, each
study is presented in detail so that it can be read
independently of the other studies.
CHESS health indicators focus on measures of
selected acute and chronic disease, altered physi-
ology, and pollutant burdens. Acute health effects
attributable to short-term pollutant exposure are
quantified by comparing response frequency against
daily variations in pollution levels. Effects attrib-
utable to long-term exposure are identified by con-
trasting disease prevalence in High and Low exposure
communities.
These studies describe the wide and complex range
of subtle and overt adverse health effects that may be
attributed to pollutant exposures but leave a number
of problems unanswered that can only be partially
resolved at this point. Examples of these unanswered
problems include the relative health effects of sulfur
dioxide and suspended sulfates, the relative impor-
tance of repeated peak short-term exposures
compared to elevated annual average exposures, and
the relative significance of indoor air pollutants.
Studies included in this monograph are the "state
of the art" summary of our findings to date. As new
techniques and areas of study appropriate to CHESS
are identified, they will be incorporated into the
CHESS program. Developmental work is now in
progress to deploy more sensitive and objective health
indicators, particularly physiologic and biochemical
responses, of short-term variations in air quality. The
questions raised by these present studies and the
potential need for modifying existing short-term
primary standards or establishing standards for other
pollutants will be validated or rejected by future
reports from ongoing CHESS and other studies.
CHESS studies on effects attributable to oxides of
sulfur, suspended sulfates, and other particulates are
programmed through 1975.
Studies reported in this monograph were con-
ducted in only a portion of the current CHESS
program. Results are given for three CHESS commu-
nities in New York and four CHESS communities in
the Salt Lake Basin. In addition, the monograph
includes findings from studies conducted in Idaho-
Montana, Chicago, and Cincinnati, in which health
indicators similar to those used in CHESS were
employed. Air measurements from these areas were
obtained from local agencies and other Federal
networks, however, and were somewhat dissimilar to
standard CHESS air monitoring techniques.
IV
-------
CONTRIBUTING AUTHORS
Ferris B. Benson,B.A.
Robert M. Burton, B.S.
Dorothy C. Calafiore, Dr.P.H.
Robert S. Chapman, M.D.
Arlan A. Cohen, M.D.
Anthony V. Colucci, Sc.D.
John P. Creason, M.S.
Wave E. Culver, M.S.
Richard C. Dickerson, B.S.
Thomas D. English, Ph.D.
John H. Farmer, PhD.
John F. Finklea, M.D., Dr.P.H.
Jean G. French, Dr.P.H.
Harvey E. Goldberg, M.D.
Julius Goldberg, Ph.D.
Douglas I. Hammer, M.D., M.P.H.
Victor Hasselblad, Ph.D.
Carl G. Hayes, Ph.D.
L. Thomas Heiderscheit, M.S.
Marvin B. Hertz, Ph.D.
David O. Hinton, B.S.
Dennis E. House, M.S.
Robert G. Ireson, M.S.
Gory J. Love, Sc.D.
Gene R. Lowrimore, Ph.D
Kathryn E. McClain
Cornelius J. Nelson, M.S.
William C. Nelson, Ph.D.
Vaun A. Newill, M.D., S .M.Hyg.
Lyman J. Olsen, M.D.
Blaine F. Parr
Mimi Pravda, B.S.
Peggy B. Ramsey, B.S.
Wilson B.Riggan,Ph.D.
Charles R. Sharp, M.D., M.P.H.
Carl M. Shy, MJD.,Dr.P.H.
J. Wanless Southwick, Ph.D.
G. Wayne Sovocool, Ph.D.
Walter B. Steen, B.S.
Jose M. Sune,M.A.
Donald H. Swanson, B.S.
Lawrence A. Truppi, M.S.
-------
-------
CONTENTS
Section Page
CHAPTER 1. INTRODUCTION 1-1
1.1 An Overview of CHESS 1-3
CHAPTER 2. SALT LAKE BASIN STUDIES 2-1
2.1 Human Exposure to Air Pollutants in Salt Lake Basin Communities 2-3
2.2 Prevalence of Chronic Respiratory Disease Symptoms in Adults: 1970
Survey of Salt Lake Basin Communities 2-41
2.3 Frequency of Acute Lower Respiratory Disease in Children: Retrospective
Survey of Salt Lake Basin Communities, 1967-1970 2-55
2.4 Aggravation of Asthma by Air Pollutants: 1971 Salt Lake Basin Studies . 2-75
CHAPTER 3. ROCKY MOUNTAIN STUDIES 3-1
3.1 Human Exposure to Air Pollutants in Five Rocky Mountain Communities,
1940-1970 3-3
3.2. Prevalence of Chronic Respiratory Disease Symptoms in Adults: 1970
Survey of Five Rocky Mountain Communities 3-19
3.3. Frequency of Acute Lower Respiratory Disease in Children: Retrospective
Survey of Five Rocky Mountain Communities, 1967-1970 3-35
CHAPTER 4. CHICAGO-NORTHWEST INDIANA STUDIES 4-1
4.1. Human Exposure to Air Pollutants in the Chicago-Northwest Indiana
Metropolitan Region, 1950-1971 4-3
4.2. Prevalence of Chronic Respiratory Disease Symptoms in Military Recruits:
Chicago Induction Center, 1969-1970 4-23
4.3. Prospective Surveys of Acute Respiratory Disease in Volunteer Families:
Chicago Nursery School Study, 1969-1970 4-37
CHAPTER 5. NEW YORK STUDIES 5-1
5.1. Human Exposure to Air Pollution in Selected New York Metropolitan
Communities, 1944-1971 5-3
5.2. Prevalence of Chronic Respiratory Disease Symptoms in Adults: 1970
Survey of New York Communities 5-33
5.3. Prospective Surveys of Acute Respiratory Disease in Volunteer Families:
1970-1971 New York Studies 5-49
5.4. Aggravation of Asthma by Air Pollutants: 1970-1971 New York Studies . 5-71
5.5. Frequency and Severity of Cardiopulmonary Symptoms in Adult Panels:
1970-1971 New York Studies 5-85
5.6. Ventilatory Function in School Children: 1970-1971 New York Studies .5-109
CHAPTER 6. CINCINNATI STUDY 6-1
6.1. Ventilatory Function in School Children: 1967-1968 Testing in Cincinnati
Neighborhoods 6-3
CHAPTER 7. SUMMARY AND CONCLUSIONS 7-1
7.1. Health Consequences of Sulfur Oxides: Summary and Conclusions Based
upon CHESS Studies of 1970-1971 7-3
APPENDICES A-l
A. CHESS Measurement Methods, Precision of Measurements, and Quality
Control A-3
B. NASN Laboratory Methodology B-l
C. Questionnaires Used in the CHESS Studies C-l
D. Abbreviations and Conversion Factors D-l
E. Bibliography E-l
vu
-------
CHAPTER 1
INTRODUCTION
1-1
-------
1.1 AN OVERVIEW OF CHESS
Carl M. Shy, M.D., Dr. P.H., Wilson B. Riggan, Ph.D.,
Jean G. French, Dr. P.H., William C. Nelson, Ph.D.,
Richard C. Dickerson, B.S., Ferris B. Benson, B.A.,
John F. Finklea, M.D., Dr. P.H., Anthony V. Colucci, Sc.D.,
Douglas I. Hammer, M.D., M.P.H., and
Vaun A. Newill, M.D., M.S.Hyg.
1-3
-------
INTRODUCTION
CHESS STUDY STRATEGIES
CHESS, an acronym for the Environmental Pro-
tection Agency's (EPA) Community Health and
Environmental Surveillance System, is a National
program of standardized epidemiologic studies de-
signed to measure simultaneously environmental qual-
ity and sensitive health indicators in sets of communi-
ties representing exposure gradients to common air
pollutants. The purpose of the CHESS program is to
evaluate existing environmental standards, obtain
health intelligence for new standards, and document
the health benefit of air pollution control. In order to
accomplish this mission, it is necessary to provide, as
far as possible, consistency in design and in methods
of collecting, processing, and analyzing data among
the various studies. For this reason, the program is
coordinated by a single organization in cooperation
with local public health agencies, universities, and
private research institutes. The elements of the
CHESS program are described in following sections.
A CHESS set consists of a group of communities
selected to represent an exposure gradient for desig-
nated pollutants. Each CHESS set generally includes
High, Intermediate, and Low exposure communities
selected to evaluate existing air quality standards for
particulates, sulfur oxides, nitrogen oxides, and pho-
tochemical oxidants. Current CHESS area sets (as of
March 1973) are shown in Figure 1.1.1. The possi-
bility of detecting pollutant interactions is enhanced
by selecting CHESS sets with exposure to a single
pollutant alone or in combination with other pollu-
tants.Because effects of short-term carbon monoxide
exposures are more precisely studied in controlled
exposure chambers, a CHESS area set was not
established to measure carbon monoxide effects.
In addition to satisfying the pollutant exposure
gradient, other bases for community selection within
a CHESS set were that the communities be as close to
one another as possible to minimize climatalogical
NEW YORK/NEW JERSEY
(SOX AND PARTICULATES)
SALT LAKE BASIN
-------
variation and be made up of predominantly white,
middle-class populations that were matched, as far as
possible, for socioeconomic and other demographic
factors.
The samples chosen for study within a community
are not intended to be representative of the general
population. Air quality standards are being establish-
ed to protect particularly vulnerable segments of the
population, so that for some studies we deliberately
recruit asthmatics and subjects with preexisting heart
and lung disease. Middle-class residential neighbor-
hoods containing three or four elementary schools
and a junior and senior high school are also chosen
because they represent a large proportion of the
population, have a more homogeneous family and
social class distribution, and represent neighborhoods
whose characteristics change more slowly than those
of central city neighborhoods.
CHESS programs will operate from 3 to 5 years in
most areas. Measurement of sensitive health indi-
cators during an interval of improving air quality is an
optimal way to quantitate the health benefits of
pollutant controls. Studies may be extended to detect
time lags between air quality improvement and
anticipated health benefit.
Environmental pollutants can affect the health of
individuals over a broad spectrum of biological
response, as shown in Figure 1.1.2. More severe
effects such as death and chronic disease will be
manifested in relatively smdl proportions of the
population, and the role of environmental pollutants
in the mortality and chronic disease experience of a
community is difficult to quantify because so many
other determinants of death and disease cannot be
adequately measured. The lower strata in the re-
sponse spectrum of Figure 1.1.2 are subclinical
manifestations of pollutant exposure. At a point in
time, many more individuals in an exposed com-
munity will respond with altered physiology or
pollutant burdens (tissue residues of pollutants that
accumulate in the body) than will die or develop
chronic disease. Furthermore, lower levels of the
response spectrum can be more readily quantified and
measured objectively. Because responses at lower
levels are more rapidly manifested, they may be more
useful to demonstrate an immediate health benefit of
pollution control.
CHESS health indicators focus on measures of
selected acute and chronic diseases, altered physiol-
ogy, and pollutant burdens. These indicators may be
conveniently divided into two categories: health
effects attributable to short-term pollutant exposure
A f
1RTALITY I
MORTALITY
MORBIDITY
ADVERSE
HEALTH
EFFECTS
PATHOPHYSIOLOGIC
CHANGES
I
PHYSIOLOGIC CHANGES OF
UNCERTAIN SIGNIFICANCE
POLLUTANT BURDENS
i-*-PROPORTION OF POPULATION AFFECTED-*-
Figure 1.1.2. Spectrum of biological
•response to pollutant exposure.
and effects attributed to long-term exposure. Sample
sizes and response frequencies employed for each
CHESS health indicator are shown in Table 1.1.1.
Acute health effects are observed by following panels
of subjects who have been systematically preenrolled
and comparing response frequency against daily
variations in pollutant levels. Various statistical anal-
yses are employed to isolate the effects of tempera-
ture and season on response frequency. By making
simultaneous observation in Low exposure communi-
ties, effects of environmental covariates such as
temperature can be quantified, and pollutant-
temperature interactions can be detected. These
health indicators are related to exposures of 1 to 96
hours, depending upon the nature of the response.
Effects attributable to long-term exposure are
identified by contrasting disease prevalence in High
and Low exposure communities. An acute effect,
such as excess acute respiratory disease, may be a
manifestation of chronically impaired resistance to
disease. Persistence of illness excess or of altered
physiology in a High exposure community provides a
means to discriminate between effects attributable to
short-term and to long-term exposure. Disentangling
the effects of dose rate, that is, large doses in short
intervals versus repeated small doses r>ver long peri-
ods, is difficult in community studies and generally
requires controlled experimentation.
Introduction
1-5
-------
Table 1.1.1. CHESS HEALTH INDICATORS
Exposure
Indicator
Sample size
per
CHESS community
Frequency
of
response
Short-term
«4 days)
Long-term
(usually
>1 year)
Frequency of asthma attacks
Aggravation of chronic
respiratory symptoms
Aggravation of cardiac
symptons
Acute irritation symptoms
during episodes
Daily mortality
Pollutant burdens
Impairment of lung function
Incidence of acute respiratory
disease in families
Frequency of acute lower
respiratory disease in
children
Prevalence of chronic
respiratory disease
50 to 75
50 to 75
50 to 75
400 to 1000
Variable
800 to 1200
1500 to 2000
1000
1500 to 3000
1500 to 3000
Daily
Daily
Daily
3 to 4 times
yearly
Daily
Once in
2 years
3 times yearly
Once every
2 weeks
Once in
2 years
Once in
2 years
Many factors contribute to community differences
in the distribution of diseases associated with air
pollution exposure. Factors that, along with environ-
mental pollution, codetermine the quantitative level
of a health indicator in a community are called
"covariates." Covariate information obtained in the
CHESS program includes data on other environ-
mental factors, such as temperature and humidity,
and personal or family characteristics, such as age,
sex, race, education of parents, occupational dust and
fume exposure, cigarette smoking habits, geographical
migration, and previous illness experience. Covariates
may be kept constant across study groups by careful
selection of participants, or they may be measured
and then appropriately adjusted through application
of statistical procedures.
ENVIRONMENTAL MONITORING
Air monitoring stations are sited in each CHESS
community to provide credible estimates of pollutant
exposure for the study population. The large majority
of study subjects live within 1.5 miles of the
monitoring station. Stations and sample inlets are
placed to be as representative of community-wide
exposure as possible. Topography, land use adjacent
to the station, emission sources, and population
distribution are considered in selection of CHESS
monitoring sites.
At the inception of the CHESS program in 1969,
manual instruments were operated 7 days each week
to monitor 24-hour integrated concentrations of total
suspended particulates, suspended sulfates, suspended
1-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
nitrates, sulfur dioxide, and nitrogen dioxide;
monthly concentrations of dustfall and trace metals
in dustfall were also determined. Since that time,
instrumentation has been added to measure respirable
suspended particulates, carbon monoxide, ozone, and
total hydrocarbons, and automated or refined anal-
ysis procedures have been incorporated in the system.
Beginning in January 1972, automated sampling
methods listed in Table 1.1.2 were introduced into
some CHESS stations, and the entire monitoring
Table 1.1.2. CHESS ENVIRONMENTAL MONITORING
Measurement device3
High-volume sampler
Cyclone
Cascade impactor
Tape sampler
Dustfall bucket
Bubbler train (TCM)
Autoanalyzer
Bubbler train (sodium
hydroxide)
Autoanalyzer
Autoanalyzer
Autoanalyzer
Autoanalyzer
Autoanalyzers
Type of
measurement
Manual
Manual
Manual
Automated
Manual
Manual
Automated
Manual
Automated
Au tomated
Automated
Automated
Automated
Sampling
period
24 hours
24 hours
24 hours
2 hours
Monthly
24 hours
Continuous
24 hours
Continuous
Continuous
Continuous
Continuous
Continuous
Pollutant
Total suspended
paniculate
Suspended sulfates
Suspended nitrates
Trace metals
Respirable suspended
paniculate (<5 /jm)
Total suspended
particulate
Respirable suspended
particulate
Suspended particulates
Settleable particulates
Trace metals
Sulfur dioxide
Sulfur dioxide
Nitrogen dioxide
Nitrogen dioxide
Carbon monoxide
Ozone
Total hydro-
carbons
Temperature,
humidty,
wind speed,
and wind
direction
Analysis method
Gravimetric
Turbidimetric
Methylthymol blue
Hydrazine sulfate-
copper sulfate
Copper-cadmium
Atomic absorption
Gravimetric
Gravimetric
Gravimetric
Optical density
Gravimetric
Atomic absorption
West-Gaeke
Flame-photometric
Jacobs- H och he ise r
Chemiluminescent
Nondispersive
infrared
Ch em i lu m in escen t
Flame emission
Weather Bureau
standard
Period
employed
Since 1969
1969-1971
Since 1971
1969-1971
Since 1971
Since 1969
1972-1973
Since 1970
Since 1970
1970-1972
Since 1970
Since 1969
Since 1969
Since 1974
Since 1969
Since 1974
Since 1974
Since 1974
Since 1974
Since 1974
aNot all measurement devices are employed at each CHESS site.
Introduction
1-7
-------
system is now being automated. Automated stations
will telemeter real-time pollutant measurements to a
central processing station, providing data to relate
short-term environmental variations to health indi-
cators of acute response. Further details on measure-
ment methods and instrumentation are given else-
where in this monograph.'
Reagents and filters used in manual instruments
are prepared centrally and periodically checked at
designated quality control points before shipment to
field stations. Instruments are also calibrated and
maintained centrally. Environmental samples ob-
tained in the field are processed at a central labora-
tory in which systematic checks of procedures are
made. Duplicate side-by-side monitors are operated
intermittently at selected field stations to determine
reproducibility of individual sampling trains. Auto-
mated instruments, having on-call telemetric output,
permit daily routine instrument performance checks
from the central processing point. Computer pro-
grams are presently being developed to monitor daily
output of automated instruments and to flag devia-
tions from expected performance. CHESS quality
control features are more fully described elsewhere.1
Stationary monitors have inherent drawbacks as
estimators of human exposure. Within short-term
frames, individuals frequently change between indoor
and outdoor environments, from one neighborhood
to another, and from residential to occupational
exposures. Quiet, acceptable, small-scale instruments
for personal monitoring have not been developed,
though the need has been evident for years. Likewise,
quiet indoor air monitors having desirable response
characteristics have yet to be marketed. These defi-
ciencies require reliance on stationary monitors to
estimate current environmental exposures of com-
munities.
Attempts to relate chronic disease to long-term
exposures are fraught with methodologic difficulties.
Chronic disease is likely to result from cumulative
exposures of several years to several decades. Quanti-
tative air monitoring data for a given neighborhood or
city usually are available only for the past few years.
Attributing area differences in chronic disease preva-
lence to current pollutant exposures frequently will
lead to underestimates of the true exposure associ-
ated with illness. During the 1940's and 1950's, when
coal was widely consumed for domestic heating and
industrial control measures were less prevalent, air
pollutant concentrations tended to be considerably
higher than values measured in 1970 or 1971. This is
especially true in cities where stationary sources are
principal contributors to community air pollution.
Therefore, in order to assess accurately the magnitude
of the dose-response relationship, estimates of past
exposure are required and have been made in CHESS
chronic disease studies. In communities dominated by
a single point source, these estimates are based on
past emissions. Emissions data and meteorologic
information are entered into a meteorologic disper-
sion model, yielding estimates of ambient exposure.
These results are calibrated to current ambient air
monitoring data. In multiple-source urban areas,
adequate past emissions data seldom are available,
and dispersion models tend to cope inadequately with
the complex variability of urban exposures. In these
situations, local suspended particulate or dustfall
measurements dating back to the early 1950's, when
available, proved to be a crude but necessary baseline
for extrapolating from current air monitoring data to
past exposure.
In addition to long-term exposure estimates, some
CHESS studies attempt to discriminate between
chronic disease prevalence of long-time residents and
recent immigrants to a High exposure area. After
discounting the relatively high illness rates of most
recent immigrants (residence duration of 1 year or
less), disease prevalence is compared in residents of 2
to 4 years duration from High and Low exposure
areas. Other things being equal, excess disease found
among recent residents of High exposure areas may
be attributed to more recent exposure. If the effect is
cumulative, area differences in disease prevalence will
tend to be amplified among long-term residents and
will confirm the findings among recent immigrants.
These approaches provide one means to assess the
impact of exposure duration on chronic disease
prevalence.
DATA ANALYSIS
Since sample sizes for community studies neces-
sarily are large, the computer is an essential part of
the data processing operation. Not only does it
perform the standard storage and retrieval function
for health and environmental exposure data and
perform statistical analyses, but it also aids exten-
sively in the study logistics. The computer is used to
select candidates in priority order for the repetitive
panel studies, to print required identifying informa-
tion on the questionnaires, and to prepare mailing
labels for health collection materials. To lessen an
enormous amount of data coding and keypunching,
1-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
optically sensed questionnaire forms are used.
Although the information on these forms is read
directly onto magnetic tapes, much data editing is
still necessary. For each type of questionnaire, edit
computer programs are utilized to identify errors.
As with the environmental monitoring samples, all
questionnaire development, data processing, and sta-
tistical analysis are performed at a central facility;
hence, uniform control of these important functions
is maintained.
Statistical analyses of CHESS data sets require
several different techniques. Standard analyses are
not always possible since the dependent health
variables are confounded by discrete design variables
and continuous covariates. These analyses had to
allow for unbalanced designs, missing data, repeated
measurements, serial correlation, and nonnormality.
Most of the quantitative health variables, such as
pulmonary function tests, were analyzed using a
general linear regression model as described by
Graybill.2 These models allow for both continuous
and discrete covariates and do not assume that these
covariates are linear (the word "linear" in linear
model, refers only to the coefficients of these
effects). Most of the categorical responses, such as
chronic respiratory or acute lower respiratory disease
responses, were analyzed using a general linear model,
as described by Grizzle et al.3 This model is also a
least squares procedure and is very similar to the
regression model. This procedure produces marginal
means adjusted for the other factors or covariates and
partitions Chi Squares in a manner similar to the
partitioning of sums of squares in regression. The
procedure is robust (i.e., moderate deviations from
underlying assumptions will not alter conclusions
about the statistical significance of observed differ-
ences), except for small cell sizes. In most analyses,
small cell sizes were not a problem. These analyses
permit testing and estimating of area differences after
accounting for the previously mentioned covariates.
SUMMARY
This paper has described the structure and ration-
ale of the CHESS program, the methods employed to
relate environmental exposure to human health ef-
fects, and the various study strategies used to
implement this program. Subsequent papers in this
monograph will present detailed exposure-effects
information derived from the CHESS program in its
first full year of operation. Papers will focus on
effects related to sulfur oxide exposures.
REFERENCES FOR SECTION 1.1
1. CHESS Measurement Methods, Precision of Meas-
urements, and Quality Control. In: Health Conse-
quences of Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Publication
No. EPA-650/1-74-004, 1974.
2. Graybill, F. An Introduction to Linear Statistical
Models (Vol. 1). New York, McGraw-Hill Book
Company, Inc., 1961.
3. Grizzle, I.E. , C.F. Starmer, and G.G. Koch.
Analysis of Categorical Data by Linear Models.
Biometrics. 25(3):489-504, September 1969.
Introduction
1-9
-------
CHAPTER 2
SALT LAKE BASIN STUDIES
2-1
-------
2.1 HUMAN EXPOSURE TO AIR POLLUTANTS
IN SALT LAKE BASIN COMMUNITIES, 1940-1971
Marvin B. Hertz, Ph. D., Lawrence A. Truppi, M.S.,
Thomas D. English, Ph.D., G. Wayne Sovocool, Ph. D.,
Robert M. Burton, B.S., L. Thomas Heiderscheit, M.S.,
and David O. Hinton, B.S.
2-3
-------
INTRODUCTION
The Salt Lake Valley offers a unique setting in
which to observe the relationships between the health
parameters of the Community Health and Environ-
mental Surveillance System (CHESS) program and a
gradient of human exposure to sulfur dioxide.
Previous epidemiological studies investigating sulfur
dioxide effects have been confounded by mixed
exposures in urban areas to total suspended particu-
lates and to other pollutants. Four communities in or
near the Salt Lake Valley were chosen to represent an
exposure gradient to sulfur dioxide, which is emitted
primarily from a large copper smelter west of the
study communities. It was also expected that the
exposure levels of the communities to other pol-
lutants would be below the National Ambient Air
Quality Standards and that concentrations of
suspended particulates in the communities would be
similar.
Emission controls have already been added to the
copper smelter near the Utah CHESS communities,
and further controls are planned. In fact, the primary
sulfur dioxide emission from the smelter in 1971 was
less than one-fifth the emission of 1940.1 Thus, the
Utah CHESS study allows health effects to be
observed as the level of sulfur dioxide decreases.
The purpose of this report is to provide long-range
pollutant exposure information that can be used for
studies of human health effects within these Utah
communities. These health effects studies relate both
chronic and acute respiratory diseases and other
health effect indicators to human air pollution
exposures.
Community Descriptions
TRAVERSE WO(/,\T/l;YS
MILES
0 5 10 15 20 25
5 10 15 20 25 30 35
KILOMETERS
Figure 2.1.1. Salt Lake Valley CHESS
communities.
mean sea level (Figure 2.1.1). Great Salt Lake lies to
the northwest of the communities in the valley, the
nearest point of the lake being about 12 miles from
Salt Lake City. The ridges of the Wasatch Mountains,
which rise to a height of 11,300 feet, and the ridges
of the Oquirrh range, which rise to a height of 10,000
feet, form a formidable barrier to east-west move-
ment of air masses.
The Utah CHESS study areas are situated in the
north central part of the state of Utah and include
Ogden, Salt Lake City, Kearns, and Magna (Figure
2.1.1). These communities were selected on the basis
of meteorological information, emissions inventories,
and previous air quality measurements to provide an
exposure gradient for sulfur dioxide with similar
levels of other pollutants, which were expected to be
below National Ambient Air Quality Standards. All
four communities have primarily white, middle-class
populations.
Except for Ogden, the study areas are located on
the floor of the Salt Lake Valley, 4220 feet above
Because of the topographic features of the area,
local climatology plays an important role in air
pollution considerations. In particular, wind flow
around Salt Lake City, for the most part, is a
function of the proximity of the Salt Lake Valley to
the canyons on the west slopes of the Wasatch
Mountains. Differential heating by the sun results in
up-slope winds in the afternoon and evening and
down-slope winds at night and in the morning.
An inspection of the wind rose for the Salt Lake
City Airport for 1951-1960 (Figure 2.1.2) reveals
that about 50 percent of the winds blow from the
south to southeast and 25 percent from west-
2-4
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
134
0 5 10 15
PERCENT OCCURRENCE
1-3. 4-7 312 13-17 18-23 >23
WIND SPEED, mph
Figure 2.1.2. Surface climatologic wind rose,
Salt Lake City Airport.
northwest to north, there are 5 percent calms, and
the remaining winds (20 percent) are spread around
the other directions. Winds from the south to
southeast represent nighttime down-slope winds,
which usually persist about 15 hours, 9 p.m. to 12
noon, and result from cool air flowing down the
slopes of the Salt Lake Valley to the south and the
Wasatch canyons to the east and southeast of Salt
Lake City. With daylight, heating of the mountain
slopes causes an up-valley flow (WNW-N) to start
about noon and continue during the afternoon and
evening. Mean velocity of the up-valley winds is about
9 mph, and that of the down-valley winds is about 7
mph.
A mountain-valley wind pattern such as this has
important implications for the long-term air pollution
situation around Salt Lake City. Pollutant sources to
the north and northwest or those to the south and
southeast would affect the city about 75 percent of
the time. There are two major oil refineries to the
northwest, both possible sources of sulfur dioxide,
and it is reasonable to expect that emissions from
both sources would flow directly into Salt Lake City
with the winds that persist from noon to 9 p.m.
almost every day. However, the greatest source of
sulfur dioxide around Salt Lake City and the CHESS
areas is the copper smelter near Garfield, Utah. This
large point source lies 13 miles due west of the city,
and from that location winds that would carry
emissions directly to the city would occur only 4
percent of the time.
Two CHESS communities, Magna and Kearns, are
located much closer to the smelter source, ap-
proximately 5 miles and 8 miles from the smelter,
respectively, on the floor of the Salt Lake Valley
between Salt Lake City and the smelter. It was
thought these sites would receive a high exposure of
sulfur dioxide with the up-valley (WNW-N) winds in
the Salt Lake Valley.
The remaining CHESS site, Ogden, lies out of the
Salt Lake Valley 35 miles due north of Salt Lake City
in the foothills of the Wasatch Mountains. Inspection
of summarized wind data for Ogden (Figure 2.1.3)
reveals a pattern different from that around Salt Lake
City. Down-slope winds out of Weber Canyon and
13 47 3-12 13-17 1823 >23
WIND SPEED, mph
Figure 2.1.3. Surface climatologic wind rose,
Ogden (Hill Air Force Base).
Salt Lake Basin Studies
2-5
-------
other large canyons to the east result in prevailing
easterly winds from the Wasatch Mountains. A
southerly flow, which might carry sulfur dioxide
emissions from Salt Lake City, is evident.
Climatological data from the radiosonde station at
Salt Lake City Airport rank this region as one of the
poorest in the nation for total hours of inversion
conditions and also for total number of days during
which inversion occurs. Inversions are observed on
about 80 percent of the mornings throughout the
year. The highest frequency of morning inversions
occurs in summer, with a frequency of from 90 to 95
percent. Although winter inversions average only 75
to 80 percent in the morning, they tend to persist
during the day, particularly with stagnant conditions.
In warmer months, rapid heating tends to dissipate
the more frequent morning inversions. Computation
of morning urban mixing depths reveals no monthly
average above 400 meters; the critical morning mixing
depth for high air pollution advisories is 500 meters.
Another criterion for high air pollution is average
wind speeds aloft of 4 m/sec through the urban
mixing depth, and Salt Lake City averages less than
that in 6 of the 12 months of the year.
Mnnitorina Human Exposure to Air
Pollutants
The strategy and methods employed in the CHESS
program for monitoring human exposure to environ-
mental pollutants have been described elsewhere.2'3
For the Utah CHESS communities, the air monitoring
sites were located within an approximate 1.5-mile
radius of the center of the human population under
surveillance. These air monitoring sites were located
to represent the air quality exposure of the respective
study populations in residential neighborhoods.
Sulfur dioxide, nitrogen dioxide, and total suspended
particulate concentrations, coefficient of haze levels,
and dustfall have been measured at the sites since
December 1970. Data on sulfur dioxide and total
suspended particulate concentrations were also
available from the Utah State Department of Health.
Location and Description of the Monitoring Sites
The CHESS air monitoring sensors in Magna are
located on the roof of the Brockbank Junior High
School. The school is one-story with a roof about 20
feet above ground. There are no tall trees or other
obstructions to airflow within several hundred yards
of the school. The area around the school building is
a grass-covered yard with a small asphalt parking lot.
The school is located in the residential neighborhood.
A moderately busy thoroughfare is located two
blocks south of the school.
The air monitoring sensors for the Kearns com-
munity are located on the roof of the Kearns Junior
High School, a one-story building except for the
auditorium, which extends slightly higher. The air
monitoring equipment is located at the one-story
height approximately 20 feet above ground. The
immediate area surrounding the school is grass-
covered except for a parking lot north of the school.
The school is located in the residential neighborhood
with no busy thoroughfares nearby.
In Salt Lake City, the air monitoring sensors are
on top of the two-story City-County Health Complex
approximately 35 feet above ground. To the north
and east of the building are fairly busy commercial
streets. Parking lots are located to the south and west
of the building. The buildings near the Health
Complex are one story, and no tall trees are present.
Air monitoring sensors in Ogden are located on the
roof of the one-story City-County Health Depart-
ment. The roof is about 15 feet above ground. A
50-foot-tall tree is located approximately 40 feet east
of the building. A fire department building to the east
of the building has a smoke stack approximately 40
feet high about 200 feet from the monitoring
equipment. The stack emits from the gas-fired boiler
in the fire department. The streets on two sides of the
building are busy commercial thoroughfares.
Pollutant Measurement Methods (CHESS)
Sulfur Dioxide - The 24-hour sulfur dioxide gas
bubbler system is composed of relatively simple parts.
From an ambient air inlet, air is drawn through a
6-inch-long bubbler stem of 6-mm tubing. The tubing
is drawn to an inner tip diameter of 0.025 inch to
control bubble size. The air then bubbles through 35
ml of a 0.1 -M potassium tetrachloromercurate (TCM)
solution contained in a 164- by 32-mm polypropy-
lene tube. Exhaust air passes through a glass wool
moisture trap. The air then passes through a 23-gauge
hypodermic needle used as a critical orifice to control
the flow rate at approximately 0.5 liter/min. A
second moisture trap is placed between the needle
and the vacuum source. Moisture traps are used to
reduce corrosion for both the vacuum pump and the
needle. The traps also provide a container for
overflow of TCM solution if the train is assembled
incorrectly. Flow rate is measured at the beginning
2-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
and end of each sample period. Chemical analysis is
performed using the modified West-Gaeke proce-
dure.4"6 Results are reported in micrograms per cubic
meter (jug/m3).
Nitrogen Dioxide - Nitrogen dioxide concentrations
were determined from 24-hour gas bubbler samples
using the Jacobs-Hochheiser7 technique. The nitrogen
dioxide data are being reported in this paper although
the method has recently come under much criticism.
It is widely held that the Jacobs-Hochheiser method
does not give accurate results because of several
factors. Two factors include, first of all, a fluctuating
absorption efficiency and, secondly, probable inter-
ference from other environmental pollutants.
Although at the time of data collection the Jacobs-
Hochheiser method was the Federally accepted
method of measuring nitrogen dioxide concentration,
these data should only be used in the light of the
recent findings mentioned previously.
The 24-hour nitrogen dioxide gas bubbler system
is very similar to the sulfur dioxide bubbler. The air
bubbles through two 164- by 32-mm polypropylene
tubes, each containing 35 ml of 0.17V sodium
hydroxide. Otherwise, components and operation are
the same as described above for the sulfur dioxide
bubblers.
fraction is determined using a reduction diazo
coupling method with automated analysis.10 Results
are reported in micrograms per cubic meter (/ug/m3).
Dustfall Paniculate - The measurement of atmos-
pheric particulate dustfall provides an approximate
gauge of the amount of environmental contaminants
that precipitate out of the atmosphere onto the
respective CHESS neighborhoods. Specifically, deter-
minations are made of total dustfall and of lead, zinc,
and cadmium in dustfall. These data are useful for
estimating the proportions of pollutants entering via
the respiratory and gastrointestinal systems.
The dustfall unit is an open-top, cylindrical
container, which collects the larger and more dense
fractions of atmospheric particulates. Made of
durable, nonreactive polyethylene, the container is
supplied with a press-on lid. The entire unit is
therefore suitable for shipment. When in use, the
bucket is suspended on an aluminum mast at a height
at least 6 feet above ground for a 1-month period.
The particles collected are emptied into a glass
beaker, water is evaporated, and the weight of the
remaining matter is determined.11 Residue is also
analyzed for trace metals by atomic absorption.12
Results are reported in grams or milligrams per square
meter per month (g/m2/mo).
Total Suspended Particulate - The high-volume air
sampler is used to collect fairly large quantities of
particulate matter on filters. The sampler passes
ambient air through an 8- by 10-inch glass fiber filter
at a starting flow rate of approximately 60 ft3/min.
The flow rate is sufficiently large to allow collection
of suspended particulates between 90 and 0.1 ;um in
diameter.
The sampler and rotameter are routinely calibrated
as a unit for accurate flow rate determinations. Each
unit is routinely inspected and repaired every 25
calendar days. A shelter of standard design protects
the sampler from adverse weather conditions and
vandalism. The shelter is located sufficiently high
above ground to provide unobstructed airflow.
The total weight of particulate matter collected on
each filter is determined gravimetrically following the
Environmental Protection Agency standard reference
method.8-9
For the sulfate fraction of suspended particulate, a
turbidimetric method is used, ^and turbidity is
measured using a spectrophotometer.4 ° The nitrate
Coefficient of Haze (COH) - The A.I.S.I. Tape
Sampler is designed to provide continuous short-term
measurements of particulate matter. The principal
components of the tape sampler are a vacuum pump,
sampling nozzle, automatic timer, and sampling tape
(filter). The vacuum pump draws ambient air through
a filter held in the sampling nozzle. Particulates are
deposited upon the tape for a period of 2 hours, then
the tape is automatically advanced to the next clean
area. The sampler airflow, which is measured by a
built-in flowmeter, is approximately 0.25 ft3/min.
The unit is housed in a metal carrying case, which
both protects the instrument and prevents soiling of
the tape outside the sample spot. The light trans-
mittance of each sampled area is determined at a
central laboratory. Specific methods for the deter-
mination of COH units have been previously
described.13 Results are reported as COHs per 1000
linear feet of air (COHs/1000 lin ft).
Precision of Measurements - The CHESS program
conducted tests to determine the precision of environ-
mental measurements.3 These tests were done mainly
in the southeastern CHESS cities, Birmingham,
Salt Lake Basin Studies
2-7
-------
Charlotte, and Greensboro. Duplicate sensors were
operated at air monitoring sites within these cities on
a daily basis for 8 months in Birmingham and for 6
months in Charlotte and Greensboro. Comparisons of
the regular samples with duplicate samples were used
to statistically determine the precision of CHESS
measurements for sulfur dioxide, nitrogen dioxide,
total suspended particulate, and particulate dustfall.
The arithmetic mean1 percentage errors were as
follows:
1. Sulfur dioxide, ±27 percent.
2. Nitrogen dioxide, ±16 percent.
3. Total suspended particulate, ±6 percent.
4. Particulate dustfall, ±25 percent.
The uncertainty associated with the nitrogen dioxide
determinations using the Jacobs-Hochheiser tech-
nique should be noted once again. As mentioned in
the section on nitrogen dioxide measurement
method, the data should only be considered in the
light of the recent findings concerning the ques-
tionable validity of the method.
Sampling Times and Frequency - Table 2.1.1
compares the measurement methods, sampling times,
and sampling frequencies used by CHESS and the
Utah State Division of Health. CHESS measurements,
except for particulate dustfall and COHs, are taken
daily for each 24-hour period. The 24-hour period
normally begins and ends between 8 a.m. and 12
noon. The COH measurements provide a 2-hour value
representing the even hours of the day, for example,
8 to 10 a.m., 10 a.m. to 12 noon, etc. Particulate
dustfall measurements are taken over a 1-month
period.
Table 2.1.1. COMPARISON OF CHESS AND UTAH STATE DIVISION OF HEALTH METHODS,
SAMPLING TIMES, AND SAMPLING FREQUENCIES
Measurement by
CHESS
Utah State
Division of
Health
Pollutant
Sulfur dioxide
Nitrogen dioxide
Total suspended
particulate
Total suspended
particulate
Dustfall
particulates
Sulfur oxides
Total suspended
particulate
Method
Bubbler (24-hour
sample; West-Gaeke
analysis)
Bubbler (24-hour
sample; Jacobs -
Hochheiser analysis)
High-volume
sampler
A. I.S.I. Tape
Sampler
Dustfall bucket
Automatic
conductivity
High-volume
sampler
Sampling frequency
Daily
Daily
Daily
Every 2 hours
Monthly
Continuously
Daily
Sampling time
24 hours
24 hours
24 hours
2 hours
1 month
24 hours
(1963-69)
1/2 hour
(1970-71)
24 hours
2-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Pollutant Measurement Methods (Utah)
The Utah State Division of Health measures sulfur
oxides on a continuous basis using the automatic
conductivity method with hydrogen peroxide as the
absorbing solution.14 If acidic gases are present in the
ambient air sample, they can be expected to interfere
with this method. These measurements are sum-
marized into 24-hour periods representing daily
values for the years 1963-1969. Half-hour averages
are available for 1970-1971. Total suspended particu-
late concentrations are determined using a high-
volume air sampler similar to that used by CHESS.
RESULTS AND DISCUSSION
Annual Concentrations
The Utah State Division of Health has monitored
suspended particulate in Salt Lake City, Ogden,
Magna, and Kearns since 1963, 1965, 1968, and
1971, respectively (Figure 2.1.4). The annual
geometric mean decreased between 1963 and 1967
for Salt Lake City and between 1965 and 1966 for
Ogden. Beginning in 1966, however, the annual
geometric means for Salt Lake City, Magna, and
Ogden appear to be relatively stable. Since only the
1971 annual mean for Kearns is known, no trend can
be determined.
Sulfur oxide concentrations have been determined
for Salt Lake City and Ogden since 1965 and for
Magna and Kearns since 1970 and 1971, respectively
(Figure 2.1.5). The annual arithmetic mean for Salt
Lake City dropped considerably in 1967. Between
1967 and 1971 the means increased yearly. However,
the annual mean for 1971 was still considerably less
than the 1965 mean. The means for Ogden fluctuated
slightly between 1965 and 1970. The mean for 1971,
however, showed a 4-fold increase over the value for
1970. Magna showed a slight decrease between 1970
and 1971, although no trend could be determined.
Annual averages based on the Utah State Division
of Health data are tabulated in the Appendix, Tables
2.1.A.land2.1.A.2.
130
120
110
100
p 90
o
o
3
o
CE
Q_
70
60
50
40
1963
40GDEN
OSALT LAKE CITY
OKEARNS
0MAGNA
1964
1965
1966
1967
1968
1969
1970
1971
TIME, years
Figure 2.1.4. Total suspended particulate concentrations versus time, Utah State Division of
Health, 1963-1971.
Salt Lake Basin Studies
2-9
-------
130
120
110
100
*?= 90
UJ
o
o
o
X
o
70
40GDEN
OSALT LAKE CITY
OKEARNS
OMAGNA
1965
1966
TIME, years
Figure 2.1.5. Sulfur oxide concentrations versus time, Utah State Division of Health, 1966-1971.
Table 2.1.2 summarizes the Utah CHESS 1971
annual averages for all pollutants. The National
Primary Ambient Air Quality Standards are shown
for comparison. In no case are these standards
exceeded.
Table 2.1.3 compares Utah State Division of
Health results with Utah CHESS results. Since Utah
CHESS measurements were begun in December 1970,
only the 1971 annual averages for each area are
shown. For both total suspended particulate and
2-10
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.1.2. SUMMARY OF UTAH CHESS 1971 ANNUAL AVERAGES
FOR POLLUTANT CONCENTRATIONS"
Pollutant
Sulfur dioxide, /ug/m3
Suspended sulfate,0 jUg/m3
Total suspended
paniculate,0 jug/m3
Nitrogen dioxide, M9/m3
Suspended nitrate,0 jug/m3
Particulate,
COHs/1000linft
Dustfall paniculate.
g/m2/mo
Cadmium in dustfall.
mg/m2/mo
Lead in dustfall.
mg/m2/mo
Zinc in dustfall.
mg/m2/mo
Ogden
8.0
4.8
66.5
37.4
2.0
0.3989
4.34
0.08
6.70
4.26
Salt Lake
City
15.3
6.1
70.9
40.6
2.4
0.5654
3.56
0.11
8.21
3.88
Kearns
21.7
6.4
38.2
22.3
1.7
0.2026
2.44
0.17
4.10
3.68
Magna
61.8
9.6
53.9
20.8
1.3
0.2776
4.61
0.11
3.99
6.02
Standard11
80.0
-
75.0
100.0
-
—
—
—
_
—
Concentrations are annual arithmetic mean values unless otherwise noted.
bNational Primary Ambient Air Quality Standard.
cAnnual geometric mean values.
Table 2.1.3. COMPARISON OF UTAH STATE DIVISION OF HEALTH DATA WITH
UTAH CHESS DATA, 1971 ANNUAL AVERAGE POLLUTANT CONCENTRATIONS
Pollutant
Total
suspended
paniculate,8
M9/m3
Sulfur
dioxide,b pg/m3
Sulfur oxides,b
Mg/m3 s
Measured by
CHESS
Utah
CHESS
Utah
Ogden
66.5
78.0
8.0
10.5
Salt Lake City
70.9
94.0
15.3
21.0
Kearns
38.2
51.0
21.7
34.1
Magna
53.9
71.0
61.8
107.4
aAnnual geometric mean.
bAnnual arithmetic mean.
Salt Lake Basin Studies
2-11
-------
sulfur dioxide (sulfur oxides), the Utah CHESS and
the Utah Sti. t Division of Health results showed the
same pollutant gradients across the four communities,
but sulfur oxide levels from the State of Utah were
considerably higher than those obtained by Utah
CHESS. As mentioned previously, The Division of
Health uses an automatic conductivity method that
measures sulfur oxides and is subject to interference,
whereas the West-Gaeke method employed by CHESS
is specific for sulfur dioxide.
Daily Pollutant Cumulative Frequency
Distribution
Frequency distributions and summaries of daily
pollutant concentrations for the Utah CHESS com-
munities are given in Tables 2.1.A.3 through2.I.A.7.
Cumulative frequency distributions of daily sulfur
dioxide concentrations in 1971 exhibited the
expected differences across communities (Figure
2.1.6). In Magna and Kearns, which are near the
smelter, concentrations were highly dependent on
wind direction, as reflected in the steep slopes in
Figure 2.1.6. Smelter emissions would usually be
quite diluted and probably overshadowed by local
emission sources in Salt Lake City and Ogden,
resulting in lower peak exposures and in lower peak
sulfur dioxide concentrations in these cities. The
gradual slopes in Figure 2.1.6 for Salt Lake City and
Ogden reflect internal sources rather than smelter
emissions.
Annual average concentrations in all four com-
munities are below the National Annual Primary
Standard (80 jug/m3). Magna, which also has the
highest annual average sulfur dioxide concentration
(61.8 /Lig/m3), exceeded the National Daily Air
Quality Standard (365 Mg/m3) on 1 percent of the
days.
The total suspended particulate frequency dis-
tribution is shown in Figure 2.1.7. Magna, Ogden, and
Salt Lake City rarely exceeded the daily standard
(260 jUg/m3); Kearns did not approach the daily
standard. The pattern of community exposure to
particulates differed from the pattern of exposure to
sulfur dioxide. Particulate concentrations were
highest in Salt Lake City and Ogden, while sulfur
dioxide concentrations were highest in Magna and
Kearns.
Gaseous sulfur dioxide reacts with atmospheric
components to form suspended sulfates. A cumula-
tive frequency distribution of daily suspended sulfate
concentrations is shown in Figure 2.1.8. The gradient
of daily community exposure from highest to lowest -
Magna, Kearns, Salt Lake City, and Ogden - is the
same as that shown for sulfur dioxide in Figure 2.1.6.
Monthly Concentrations
Monthly arithmetic mean sulfur dioxide con-
centrations (Figure 2.1.9) showed, as expected, that
Magna had the highest sulfur dioxide concentration,
with Kearns, Salt Lake City, and Ogden following in
that order. Due to seasonal differences in inversion
frequency, winter months generally had higher sulfur
dioxide concentrations than summer months. A strike
at the smelter during the month of July was reflected
in an extremely low level of sulfur dioxide in Magna
and Kearns.
Monthly geometric mean suspended sulfate con-
centrations (Figure 2.1.10) revealed a community
gradient similar to that of sulfur dioxide. Again
summer had lower levels than winter. The effect of
the strike in July was again manifested in Magna and
Kearns.
Maximum daily sulfur dioxide concentrations for
each month are shown in Figure 2.1.11. Comparison
of Figure 2.1.9 and 2.1.11 reveals that although Salt
Lake City, Ogden, and Kearns had similar monthly
arithmetic mean sulfur .dioxide concentrations,
Kearns had considerably higher peak values. Peak
values for Kearns, however, were considerably lower
than those for Magna.
Salt Lake City had the highest monthly total
suspended particulate concentration (Figure 2.1.12).
Ogden, Magna, and Kearns followed in that order.
The higher values in Ogden and Salt Lake City can
possibly be explained by the location of monitoring
sites in areas of fairly heavy automotive traffic.
Summer months again showed lower levels of total
suspended particulate.
Salt Lake City had the highest monthly nitrogen
dioxide concentration (Figure 2.1.13), with Ogden
following close behind. Kearns and Magna had similar
but considerably lower concentrations. The higher
concentrations in Salt Lake City and Ogden can
possibly be explained by site locations in areas of
heavy motor vehicle traffic. Nitrogen dioxide con-
centrations fell markedly during the summer months.
2-12
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
400 —
DAILY AIR QUALITY STANDAR
ANNUAL AIR QUALITY STANDARD
O JS
60GDEN
OSALT LAKE CITY
OKEARNS
OMAGNA
95
99
PERCENT LESS THAN INDICATED CONCENTRATION
Figure 2.1.6 Cumulative frequency of sulfur dioxide concentrations for Utah CHESS
communities.
Salt Lake Basin Studies
2-13
-------
300
200
1
1 100
90
4
oc.
o
o
o
oc
•X.
o.
70
60
50
40
30
20
10
DAILY AIR QUALITY STANDARD
ANNUAL AIR QUALITY STANDARD
..—£-
&OGDEN
OSALT LAKE CITY
OKEARNS
DMAGNA
10 30 50 70 90
PERCENT LESS THAN INDICATED CONCENTRATION
99
Figure 2.1.7. Cumulative frequency of total suspended particulate concentrations for Utah
CHESS communities.
2-14
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
A OGDEN
O SALT LAKE CITY
O KEARNS
D MAGNA
10 30 50 70 90
PERCENT LESS THAN INDICATED CONCENTRATION
Figure 2.1.8. Cumulative frequency of suspended sulfate concentrations for Utah
CHESS communities.
Salt Lake Basin Studies
2-15
-------
&OGDEN
OSALT LAKE CITY
OKEARNS
OMAGNA
DEC JAN
197o}l971
DEC JAN
1971J1372
TIME, months
Figure 2.J.9. SulfurjHoxide concentrations versus time for Utah CHESS communities.
DecembeM970-April 1972.
2-16
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
20
^ 15
^4
O
t-
ft:
u
o
o
oo
s
o
LU
Q-
oo
10
60GDEN
OSALT LAKE CITY
OKEARNS
DMAGNA
DEC JAN
197011971
JUN
TIME, months
DEC JAN
197111972
Figure 2.1.10. Suspended sulfate concentrations versus time for Utah CHESS communities,
December 1970 - April 1972.
Salt Lake Basin Studies
2-17
-------
o
o
X
o
Q
o:
AOGDEN
oSALT LAKE CITY
OKEARNS
DMAGNA
DEC JAN
1970'1971
DEC JAN
197111972
TIME, months
Figure 2.1.11. Maximum 24-hour sulfur dioxide concentrations versus time for Utah CHESS
communities, December 1970 -April 1972.
2-18
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
160
120
o
£
o
o
O
40
60GDEN
OSALT LAKE CITY
OKEARNS
OMAGNA
DEC JAN
1970il971
JUN
DEC JAN
1971 {1972
TIME, months
Figure 2.1.12. Total suspended particulate concentrations versus time for Utah CHESS
communities, December 1970 - April 1972.
Salt Lake Basin Studies
2-19
-------
AOGDEN
OSALT LAKE CITY
OKEARNS
OMAGNA
DEC JAN JUN DEC JAN
19 70 11971 197111972
TIME, months
Figure 2.1.13. Nitrogen dioxide concentrations versus time for Utah CHESS communities,
December 1970 - April 1972.
The gradient and seasonal patterns for suspended
nitrates (Figure 2.1.14) were the same as those for
nitrogen dioxide. Lowest suspended nitrate concen-
trations, however, were recorded in February and
March.
Monthly arithmetic mean COH levels (Figure
2.1.15) followed the same gradient as total suspended
particulates, with Salt Lake City having the highest
level. COH levels were considerably lower during
summer months.
No consistent area or seasonal patterns for month-
ly dustfall or for cadmium, lead, and zinc in dustfall
(Tables 2.1.A.8 through 2.1.A.ll)could be observed.
2-20
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
£OGDEN
OSALT LAKE CITY
OKEARNS
QMAGNA
DEC JAN
197111972
APR
TIME, months
Figure 2.1.14. Suspended nitrate concentrations versus time for Utah CHESS communities,
December 1970 - April 1972.
Interrelationships among Pollutants
Prior studies have been conducted to determine
the relationship between various pollutant concen-
trations. In order to systematically examine these
relationships for the Utah communities, a set of
correlation matrices was calculated (Tables 2.1.A.12
and 2.1.A. 13). These matrices present the correlation
coefficients between a specific pollutant at a given
site and another pollutant and site.
To determine the relationship between suspended
sulfates and sulfur dioxide, regression lines and
correlation coefficients were calculated for daily
averages of suspended sulfate against corresponding
values of sulfur dioxide. The calculations for the Utah
CHESS communities (Table 2.1.4) demonstrate
reasonably high correlation coefficients between
suspended sulfates and sulfur dioxide in Magna,
Kearns, and Ogden (0.67, 0.58, and 0.46, respective-
ly). A low correlation (0.31) was found in Salt Lake
City, which had the largest number of small sulfur
dioxide emission sources.
Similar calculations were made for suspended
sulfates and total suspended participates (Table
2.1.4). Correlations were quite similar for all four
communities (0.50, 0.53, 0.57, and 0.48 for Magna,
Kearns, Salt Lake City, and Ogden, respectively). The
reasons for the observed relationships are not clear.
One possible explanation is that a large point
emission source of sulfur dioxide also generates a
large amount of suspended sulfate, resulting in a high
correlation between the two. In an area like Salt Lake
City, where there are multiple urban sources of sulfur
dioxide, suspended sulfate concentrations may
become a complex function of sulfur oxide and
suspended particulate concentrations.
Salt Lake Basin Studies
2-21
-------
1
o
o
1.25
1.00
2 0.75
O
o
O
UJ
-------
from 1940-1969. The ratio of sulfur dioxide concen-
tration for Kearns for 1971 obtained by the State of
Utah to the 1971 smelter emission allowed another
estimate of past exposure for Kearns to be made. The
conductivity method used by the State of Utah yields
higher sulfur dioxide results because salts and other
electrolytes affect instrument response. Both sets of
estimates were compared with estimates obtained by
employing a mathematical diffusion model developed
by the National Environmental Research Center's
Division of Meteorology. The diffusion model made
use of smelter emission data, wind roses, and
topographical characteristics of the area to compute
contours of sulfur dioxide concentrations in the
vicinity of Magna.
Although production of copper has increased from
1940 to 1971, emissions have decreased from 970
tons/day in 1940 to 193 tons/day in 1971 (Table
2.1.A. 14). Two factors account for this reduction.
First, the number of sulfuric acid plants utilizing
sulfur recovered from emissions have increased from
one in 1940 to seven in 1971. Second, air pollution
control devices in the form of baghouses, scrubbers,
cyclones, and mist eliminators have been installed.
Theoretical sulfur dioxide contours around the
smelter source (Figure 2.1.16) were determined by
applying a mathematical diffusion model, the Air
Quality Display Model (AQDM).15 As suspected, the
up-valley and down-valley wind pattern in the region
resulted in maximum downwind concentrations to
the north-northwest. Reversal of wind direction to
up-slope in the afternoon caused high concentrations
in the south to southeast. These wind patterns
resulted in an elongated pattern of sulfur dioxide
contours. Fortunately, no populated areas are located
near the centers of concentrations since Great Salt
Lake lies to the northwest and the Oquirrh Mountains
to the south-southeast. The pattern displayed in
Figure 2.1.16 shows Magna located on the fringe of
the contours with a theoretical concentration of 41
20
30 40 50
50 40
30
Figure 2.1.16. Sulfur dioxide concentration (ug/m3) contours based on mathematical
diffusion model for 1971 smelter emission rate of 193 tons/day.
Salt Lake Basin Studies
2-23
-------
Mg/m3 for the 1971 emission rate of 193 tons/day.
Measurements of sulfur dioxide by the CHESS
program revealed a concentration of 62 Mg/m3 f°r the
same year. The excess of observed over predicted
values may be accounted for by emissions from
sulfuric acid plants near Magna. No emissions data
were available from these plants, and thus they could
not be factored into the diffusion model. The
diffusion model was useful in demonstrating that
annual exposure estimates obtained from the ratio of
1971 observed air quality to 1971 emissions were not
unreasonably high or low.
Annual average total suspended particulate data
for Salt Lake City from the National Air Surveillance
Network are available back to 1953, whereas cor-
responding sulfur dioxide data are available only back
to 1962. In order to obtain another estimate of the
sulfur dioxide concentrations for the years 1953 to
1961, an average ratio of sulfur dioxide to total
suspended particulate for the available Salt Lake data
was used.
To estimate total suspended particulate concen-
trations in Magna and Reams, the ratios of Magna's
and Kearn's 1971 observed concentrations (66.2 and
44.5 ;ug/m3) to the 1971 production of copper for
the Garfield smelter (260,000 tons) were calculated.
These ratios were multiplied by the production figure
for each year back to 1940 to give annual estimates
of particulate exposure. Annual production figures
are given in Table 2.1.A.15.16
Actual and estimated air quality data for the Utah
CHESS communities are compiled in Table 2.1.A.16.
Table 2.1.5 summarizes the estimates of long-term
pollutant exposure given in Table 2.1.A.16 and
identifies the communities according to the exposure
gradient established for sulfur dioxide and suspended
sulfates. Clearly, pollution gradients of sulfur oxide
exposure are evident between each of the four
locations and between the three decades in time.
The graphs of sulfur dioxide and suspended sulfate
concentrations shown in Figures 2.1.17 and 2.1.18
Table 2.1.5. ANNUAL ARITHMETIC AVERAGE POLLUTANT CONCENTRATIONS FOR
UTAH CHESS COMMUNITIES, 1940-1971
Pollutant
Sulfur dioxide
Total suspended
particulate
Suspended
sulfates
Community
Low (Ogden)
Intermediate I
(Salt Lake City)
Intermediate II
(Kearns)b
High
(Magna)
Low
Intermediate I
Intermediate II
High
Low
Intermediate I
Intermediate II
High
Concentration,3 jug/m3
1940-49
-
—
-
(234)
-
_
—
(63)
—
—
—
(27.7)
1950-59
-
-
(50)
(142)
-
151
(40)
(60)
-
(11.4)
(10.4)
(19.5)
1960-69
—
(20)
(33)
(95)
100
112
(38)
(57)
4.8
(7.5)
(8.8)
(15.3)
1965-70
-
17
(33)
(93)
97
89
(43)
(64)
4.0
5.7
(8.7)
(15.0)
1971
8
15
22
62
78
81
45
66
5.6
7.3
7.8
12.4
aParentheses indicate estimated values.
bKearns, the Intermediate II community, was established in 1953.
2-24
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
£OGDEN
o SALT LAKE CITY
O KEARNS
O MAGNA
ANNUAL AIR QUALITY STANDARD
1940
1970 1971
Figure 2.1.17. Sulfur dioxide concentrations versus time for Utah CHESS communities,
1940-1971.
indicate that previous sulfur dioxide concentrations
have been considerably in excess of the present
National Primary Air Quality Standard. Figure 2.1.19
indicates that total suspended particulate values in
Salt Lake City and Ogden have been considerably
above the present National Primary Air Quality
Standard. The historical values of total suspended
particulate for Magna and Kearns have been relatively
low.
SUMMARY
The position of the CHESS communities in the
Salt Lake Basin relative to the major copper smelter
and the local meteorological pattern provided an area
gradient of exposure to sulfur oxides. The exposures
decreased with distance from the point source.
Primary Air Quality Standard for sulfur dioxide (80
jUg/m3) during more than 75 percent of the past 32
years on an annual basis. Increased utilization of
sulfur oxides by the copper smelter for the manufac-
ture of sulfuric acid and installation of control
devices resulted in a downward trend in emissions and
ambient air concentrations of sulfur dioxide at
Magna.
Kearns, Salt Lake City, and Ogden, Utah, in that
order, yielded a descending exposure gradient to
sulfur oxides. Magna and Kearns experienced a
greater frequency of short-term high concentrations
of sulfur dioxide since they were closer to the smelter
source. Correlations between sulfur dioxide and
suspended sulfates were higher in Magna and Kearns
than in Salt Lake City and Ogden.
The city of Magna, Utah, located in close
proximity to the smelter, had high exposure to sulfur
oxides. Although the prevailing winds transport
source emissions away from Magna most of the time,
concentrations were estimated to exceed the National
Air quality measurements made in the Utah
CHESS residential areas in 1971 did not exceed the
National Annual Ambient Air Quality Standards.
CHESS air monitoring sites were purposely located in
close proximity to the population under study and
Salt Lake Basin Studies
2-25
-------
I I I I I I I I I ! I
60GDEN
OSALT LAKE cm ~
OKEARNS
QMAGNA -
1945
1950
1955
TIME, years
1960
1965
19701971
Figure 2.1.18. Suspended sulfate concentrations versus time for Utah CHESS communities,
1940-1971.
200
o 100
o
o
o
oc
<
D_
60GDEN
o SALT LAKE CITY
O KEARNS
Q MAGNA
— ANNUAL AIR QUALITY STANDARD
1940
1945
1950
1965
1955 1960
TIME, years
Figure. 2J.19. Total suspended particulate concentrations versus time for Utah CHESS
communities, 1940-1971.
1970 1971
2-26
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
therefore represented average exposure of the
neighborhoods.
Estimates of past exposures to sulfur dioxide
dating to 1940 were derived on the basis of annual
smelter emissions. These estimates indicated that past
exposures were considerably higher (about 120
percent higher in the 1950's) than current levels.
Although past exposure estimates are admittedly
crude, they do offer some guidance in determining
past human exposure to air pollutants.
REFERENCES FOR SECTION 2.1
1. Heaney, R.T., J.R. Fletcher, and A.W. Huffaker.
Evaluation of Sulfur Dioxide Injury to Economic
Crops. Utah Copper Division, Kennecott Copper
Corporation, Salt Lake City, Utah. (Presented at
63rd Annual Meeting Air Pollution Control
Association, St. Louis. June 14-19, 1970.)
2. Shy, C.M., W.B. Riggan, J.G. French, W.C.
Nelson, R.C. Dickerson, F.B. Benson, J.F.
Finklea, A.V. Colucci, D.I. Hammer, and V.A.
Newill. An Overview of CHESS. In: Health
Consequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
3. CHESS Measurement Methods, Precision of
Measurements, and Quality Control. In: Health
Consequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
4. West, P.W. and G.C. Gaeke. Fixation of Sulfur
Dioxide as Sulfitomercurate III and Subsequent
Colorimetric Determination. Anal. Chem.
25:1816-1819,1956.
5. Scaringelli, PP., B.E. Saltzman, and S.A. Frey.
Spectrophotornetric Determination of Atmos-
pheric Sulfur Dioxide. Anal. Chem.
39:1709-1719, December 1967.
6. U.S. Environmental Protection Agency. National
Primary and Secondary Ambient Air Quality
Standards; Reference Method for the Determina-
tion of Sulfur Dioxide in the Atmosphere
(Pararosaniline Method). Federal Register.
56(84):8187-8190, April 30, 1971.
7. Jacobs, M.B. and S. Hochheiser. Continuous
Sampling and Ultramicro Determination of
Nitrogen Dioxide in Air. Anal. Chem.
50:426428,1958.
U.S. Environmental Protection Agency. National
Primary and Secondary Ambient Air Quality
Standards; Reference Method for Determination
of.Suspended Particulates in the Atmosphere
(High Volume Method). Federal Register
J6(84):8191-8194, April 30,1971.
9. McKee, H.C., R.E. Childers, and 0. Saenz.
Collaborative Study of Reference Method for the
Determination of Suspended Particulates in the
Atmosphere (High Volume Method). Southwest
Research Institute. San Antonio, Texas. Contract
CPA 70-40, SwRI Project 21-2811. June 1971.
10. Selected Methods for the Measurement of Air
Pollutants. Division of Air Pollution, Public
Health Service, U.S. Department of Health,
Education, and Welfare. Cincinnati, Ohio. PHS
Publication No. 999-AP-ll. 1965. p. 1-1 to 14
and J-l to J4.
11. Collection and Analysis of Dustfall (Settleable
Particulates). In: Annual Book of ASTM Stand-
ards. The American Society for Testing and
Materials. Philadelphia, Pa. ASTM Test Method
D1739-70.
12. Colucci, A.V., T. Hinners, and J. Kent. Analysis
Methods for Trace Metals. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Unnumbered Intramural Report.
13. Air Quality Criteria for Particulate Matter.
National Air Pollution Control Administration,
Public Health Service, U.S. Department of
Salt Lake Basin Studies
2-27
-------
Health, Education, and Welfare, Durham, N.C.
NAPCA1 ubiication No. AP-49. January 1969.
14. Air Quality Criteria for Sulfur Oxides. National
Air Pollution Control Administration, Public
Health Service, U.S. Department of Health,
Education, and Welfare, Durham, N.C. NAPCA
Publication No. AP-50. January 1969.
15. Martin, D.O. An Urban Diffusion Model for
Estimating Long Term Average Values of Air
Quality. J. Air Pollut. Contr. Assoc. 21(1): 16-19,
1971.
16. Yearbook of the American Bureau of Metal
Statistics. American Bureau of Metal Statistics.
New York.
APPENDIX
Table 2.1.A.I.
TOTAL SUSPENDED PARTICULATE
CONCENTRATION FROM UTAH
STATE DIVISION OF HEALTH,
1963-1971
Table 2.1.A.2. SULFUR OXIDE
CONCENTRATION FROM UTAH STATE
DIVISION OF HEALTH, 1965-1971
Community
Ogden
Salt Lake City
Kearns
Magna
Year
1965
1966
1967
1968
1969
1970
1971
1963
1964
1965
1966
1967
1968
1969
1970
1971
1971
1968
1969
1970
1971
Particulate
concentration,3 /K|/m3
102
84
92
78
84
69
78
121
112
104
95
79
85
90
84
94
51
70
65
70
71
Concentrations are annual geometric mean values for daily
24-hour averages. Data since 1970 have been corrected for
annual average ambient barometric pressure and daily aver-
age ambient temperature at the sampling site and apply at
standard temperature and pressure.
Community
Ogden
Salt Lake City
Kearns
Magna
Year
1965
1966
1967
1968
1969
1970
1971
1965
1966
1967
1968
1969
1970
1971
1971
1970
1971
Sulfur oxide
concentration,3
jug/m3
5.2
2.6
5.2
2.6
2.6
2.6
10.5
49.8
41.9
5.2
7.9
10.5
15.7
21.0
34.1
120.5
107.4
Concentrations are annual arithmetic mean values. Averages
for 1965-1969 are from the National Aerometric Data Bank
printouts based on Utah State Division of Health Data sub-
mitted to the Mitre Corporation in 1969. Data since 1970
have been corrected for average ambient barometric pressure
at the sampling site.
2-28
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
CO
DC
>- r*.
O <»
LU x
LU <
^ s
LL ^
li
V) *~
o
(J
x
O
QC
CO
<
Q
00
<
•
oi
—
.Q
(0
I/I
0
•M •
>
i_
O)
<§
ro
E
01
c
+J
«
l_
+J
c
OJ
O
0
O
&*
:na
;?
u «
v •o
j c:
u +•>
t/i
«
O
+J
3
_Q
4->
t/1
-o
>>
3
cr
t.
L)
'El
4-> d
U "C
O 0
<_
St;
is a
JZ t
•i— -4-
< i
n
3
i
a
u
,c
c
c
T
C
s
a
o
CM
TJ
4->
C/l
-
•o
+J
LO
CT>
tr>
en
0
Ot
o
r-.
o
UJ
co
E
0
4->
m
3 >
* QJ
1 -O
J
- c
J p
J C
J i"
E
3
X
3 >
- I-
J ui
- ^l
O
Community
OOOOOOO OOOOOOO r^OOOOC3C3 •— O ro O O CO f—
^OO^O^^ O^OOOOCi CMOCDOOOO OOOi — O O O in
ooooooo ooooffir^co cNjr^csjcucvj'^JCNj CMromojorooo
OOOOOOO OCDOOCSJCD^" i^VOi — i — CT^«*i — COCTiCDi — CSJOOCM
•— •— ^- c\j ^- *cj*o
^•cvjrocooo^CNj vDor-~.ur>r^oco o^moocri«*-o • — oo««*-vooo< — co
v£) i£) c\j oJiot-^CT>CT»Lnro i — \o i — or-^« — r-^
r- r~ (M ,— r- •— .— r- <*• CJ CO COCO^-
*ir>Lr)iocriLr>LO in< — ^-i — r^cvj(£) c\JLni^.i — ^DLDC\J cvjcocooQi — • — r^
m^-coc3cOi — csj kooiroo^o^co^D oomoroi — «^O^ intoi*€^oo^po
rocvjt— •— CXJCMCO in ro ro •— in <* «5 ,— •— ^o^-or^Ot r-cnr*-.mr-.eDin
,— r- ^ ^-^,_ CSJ.— CO
cvjoo^oai-— r*-^o o^-oyDOiouo r-cvjr--c\iioU3csj r--^i-ocsjCT»coco
MI — i — c\ji — CM *j-roroi — ^j-rotn ot-DT^-exjr^Lnoo fOLnm^-co'^-CTt
r— CO r— f— r- .— CsJ
o^r*-«Dp — cvj^-tx) LnLT)CT*en^-uD«i- c\jPOu-)Ln«a-i — oo LOIDCOI — mr^.^-
cococot£jir)oim ojior^oiotDO cvr^r^-fooio^uo ro o> 10 i — ^-r^rs.
r— r— •— CM-— »— cvii— oj*d-r-i— i— roesj^-cop^ovtMor^in
r~.LncOi — o^r^oo o^incO'y-oovoo^ oJOtn^-CTir^CTt mr^-si-CTiLnr^o
csj^-Ln^-CTiLOi — o^CT*cvjr^ooo*d- ro o o> p — ro-^i — I^COCSJ ^-OOOOi — CDtDkD
i — i — i — i — OJLT>> — IT)
OOO^-C\Jt£)i — O^OOr- i — t£>)^ OCMlOOfOLDO OVOCMCDrO^-kO
\£jooococ>«* CD o oo o in .— *d- o o o o t— o in o in tn o co •— «*
^•oco^or^cvjo irJ^oCT>^j-«*r— ^- , — oomexjid-CTtLn 1^*1 — • — ro^a-o^o
r-. •— co r^ r^ en co ojor^ooocn i^-cMco'^^-coo locsjrorooorocsj
O -— O CD O O O r- •— CD r— ,— •— O ^-i— ^-i— i— i— r— r—^^-r—,— ,— ,_
ID^-<*OCMOOO .— ^oorococor^ mcnr— cnr-^-*£> o 01 •— co oo co in
cMmi^^-o^tiDO ocTioj^ocoom cO<^om^ocDc\j om^-i — inrocn
co*3-cO'£>ooc\j ^i-oocvjcnornr-- romcocnr^r-.r^ o co in CM .— co r--
inr^i^^-c\jco*d- r^coincocoLnro Oi — coo^' — ^ CT»COIOV^O" — in
r- ,-
r-^ro^— voini — CM ^-^or^r^nroLn rooic\Jioinm<- o>tD^<*fO^r^.
co^j-ini — minm r^mr-^unLnmr^ t^)tOf^CT>ir>m^D • — cvicocninco^-
cMCMi — roi — CM rocoi — mcom roin^-
OiOOO*±OO OOOOCOOCD OOOCDOC3O OOOCD^OOCD
CMOOO' — Oi — OOOOOCDO OOOCDOOO OOOOOOO
roi — r^.ror»-cOLr) ^-COCOCMOCO^J- r^miooi^cocri r-^iocncocoinco
•— co co oo co ro co rotiDCOtoi^.coco ror--cooococsjoo cor-^r--oococMr^
ro CM ro ro
0^^^.-,.^ 0 -- -- r- ^ ^ CM 0_____™ 0 r. ^ ^ r- ,- CM
CNjrO^OCTlCMCMCn CMOOVOCnCMCMCO CNirOiXJCDCMCMCO CMrOvOCT^CMCMrO
iiiiiii iiiiiii TiiiTTi iiiiiii
.-^^r^o^^ .-^^^o,-^ ,-^^^0^^ ^^^^To^^r
t, c
•D •— «) CD
CJi >>
> +J >1
(0 'ra^
<— rj re
1 t0"
> •*<
-Q >,
^- >>
•O (fl i.
ai E re
*• -^ E
J «tz
t ^°-
£ HP
"O c Id
CO ,— r-
^- c c
0^ o O
* — +J +J
m nj
m z -z.
O i i
aConcentra
Standard
Standard
Salt Lake Basin Studies
2-29
-------
CO
CO
CO
Q
U
LLJ
Z> CM
d £
[ii CT>
CC r~
"- x
Q O
2 5
< <
9 g
I- CT)
< <-
CC CC
I- m
•Z. oa
LU S
O £
cIS
O Z
t; w
< LU
^ E
O H
r: D
o. o
Q O
SB
Sx
I I
c/l <
i— ^
2 o
es
_OJ
XI
(0
l/l
o
•M -
>
S-
OJ
_Q
o
e
CT,
P.
c
4-J
c:
OJ
^
O
CJ
o-5
cnm
c -
F- "O
o s-
0)
o
:3
_Q
£
4->
[/)
-O
ft
C
OJ
CT
OJ
S-
u
T"S
4-> tC
OJ XI
O fC
OJ 4-
C£J v
u
<1) 4-
o a
-C C-
SI
a
5
L
a.
(/
.c
c
c
-c
c
a
d
CM
•a
4->
IS)
-
-o
4->
l/l
a>
CT>
en
r-^
IT)
fO
c:
O
4->
fO
>
CD
-a
c
i
c
OJ
E
E
13
X
£
E
t=
c
£
Ul
c
o
4-»
)
C
'•^
J ni
> >
- s_
} V}
o
Community
orooo«*CTiro oooorocor^- ooooooo C5txic3CDc\jLn«±
OOOOOi— OCO OOOOCOCDLD OOOOOOO O^-OOC\Jr-^
CDr^.t^O'-orocn ocr>e\jr^toroco comCDco^J-ocJi 1-ocn<£) oco^d-OLncMCM ^J-CNJI — m^j-i — CD Oi — Lnro^ovDr^
Ln^CMro^ror- LnLnooroun^J-co < — • — • — • — ' — CM m«d-i — i — COCMLO
CJ^r-cMr^-CTii — cooLT>cvjV^-«3-oi r-t^joocoro' — o oor^cvji — OLTJ
tnp^CTip^tDi — cxj coor^r^mLnco CNJ^J-I — CTioaim r^.^DOcricni«a-ro
vo^rocvjuoino o^J-i^-CMcTiLf)' — ^j-o^l"COCJiror^ C0-«s-i — ajrococo
«d-Lnr^^j-otr>c\j CDI — i — ro^oi — LD coLn**coc\j^Dco CT»cxJi-r)pv. r^i — u3P^c\JM3ii3 cnr^"X)OOro^^- cxjvoooooroLnco
^d-r-^cno^r^-f^' — mcocncof — nou r^i^-r^Lncor^i — uncococomi — o
r^.«3-cour>"=3-Ok£> cocni — OLDLDt-. cncnLDOroOLn c\joor^«d-r^Lncr>
r^Lnror-.cvjLnc\i tntni — incovocx; P-.I — r^criooicj oc\jr^c\JC\io>CT»
°.^-,-,-^^1-^ ^
inr^-ocor-^c^i — c\jrooc\jcxji.nCD o^Cvjoucnr^csj i^-i^>Lno*>r*-*a-i —
mcrico^ocoixico csjt^DcouDcncr^i — «^-<^3r-«d-rocnu3 r^-ir>'^F— kocoa^
CO^ISLDI^D^DVOI — r-.cOLr>ikor-^c\j cxjroco^d-rorO'd- rou3^fLnLr>LOCO
^)- "d" r--cocor-.ChCft c\jmLT>rotDdLncn i — c\jcvjroc\jc\joo CM^-ro^-ron^
, — i — *d-i — ix>OLn -ci-CMLnocoooLr) moj- — i — cncDcn o^3-cOv£)p*-c\JO
c\jncnr^r^rocT> r"io*3-^toii£>ur> co • — CM^OCMI — lo1^
romcM^d-cxjm^j- rO"!d-ro^j-c\jroi^5 i — i — csjcxji — i — CM i — CMOomojcxi^t
• — mcnoo^^DiX) rocoococoi — i — i — ^j-i — roor-^r^ Ln^d-criCMixjcMOJ
i£)^^-ro|^oi-r)cr) usLn^-csjujinin r^k£>Lnmr^LOiD cni^^±»ii3U3
OOOCDCDOO OCDOOOOO OOOOOOO OOOOOOO
k£3O^r^o^^J-i-r>' — cocDCDCMCvjCTir- r^cniDCOr— csj^o loair^mcocnCM
C\J^J-Lnr--CDUDO CD^C7>incOO^J- O^fCTiCMkOCOCM OOJI-^i — ^r-^r-*c\i nnro^t-nrom ^d-tD*d-LnLnLnO^
1-^.CMroO' — cnro O^i — r^.i^.c\jro cocni^CMcnuncTi ro^j-ti3COCOCMO
•^rcxJCMi — cor--io tnco«a-r--.*3-i — • — cni — «d-ur>Ln^fi — i — c\j*d-m^j-voi —
cOCTi^or-.cor^.cM coCT^vot^o^cO'd- n^-«d-«d-«*^ri£) MDCQLnmr-.tiDi —
, — «^-iomcMc\jkO rorokocoooro o ur> i — • — mrocTi uj r^- oo r^. ^- «^ ^*
r-.*iic\jkococoii3 ocor-.oc\jc\jp-* r-nr-^LnocDr- i — orocMCMCMOO
i^>r— cxjior^r^i — roino^mooroo koroLnoocMCMCM CTI^O<*COCOCO^'
p — CNJCNJi — CMCMfO CMCNJi — • — COro^J- . — i— r— < — CMCMCM i — rocncjicocnLoco rococncocn*^-co mr — OOCOCD^J-CT>
ro n ro ro
0 ^r_OJOl_^.^r-(_(XJOl_ ,_,_,_CJOl_1_1_1_._cM
CMro^ocncMCMro CMroi£)CT>cvjcMro cMfO~^.cncMC\jro CMrokDcncMCM<^>
i r i i i i i i i i i i i i r i i i i i i i i i i i i i
\ "-^ --^ \ -*-, \ \ -^^-v-^-^-^-^-^ --v.-^.-^-^-^.-^"^ •^.•^-^."^--^"^-^
QJ
^>£.
ro
—1 VI
C C 4->
CO
I s.0"
0 5i
>> ,
T3 rd i-
Qj E *d
u l^ -i-
JD Q- S-
•r- G-
+-> "ro >^
E ro
-------
CO
00
>- CM
O 15
o o
LU OC
cc <
"-1
o .
^ °
to ?
2
o £5
P 00
< s
O
S o
CO -i-
2<
2 tt
O u_
LT3
es
I/) fe-s
O CT.Q
•r- C -
-<-> -i- -D
> QJ fO
s-
-o
>,
(J
OJ
u
£
(J
•r- -a
s- s-
4-J re
QJ -£
O ffi
OJ •*-
CD t/1
(J
0 S-
QJ -t-J
CD
S-
<1J
(/I
.0
O
o
-o
o
QJ
Q_
OJ
"O
•M
C/l
_
-a
-t~>
LO
CT>
en
O>
r^
in
c
o
4->
fO
>
<1J
13
fO
QJ
E
c
E
X
(V
E:
E
13
F
c
•E
i/>
c
o
-Jj
c:
4->
fd
>
_Q
O
Community
CO O O O ' — CT* CO ^" ' — O O ID CTt to ' — r--. O O i — OJ r^ "-O CO i — LO i — O> ^O
^rooO'^oro cOi — oo^rii — ^ r-.rocDOcnfOLn o o • — **vocvjr^
, , — i — (T)CMC\Ji — r—
r*«.«a-C3O' — P--UD oJcoo> — LO'j-o^ r^c\JOLnr--rOi — ro«— o^ffoor^
Lonoococ\ju3 ^r-*CDi — <^><*o^r mi — CDCOC\JCT>^O cot£>o^i — > — ^o
(v^ ro i — i — c\i i — c\j i — co n i— ^i~ csj ^f
i — mcricOiCLrii — UDCOI — r*. m r*. CT» r^cOi — r^tni-ncM vo^J-CTi«d-oo^oo
coaip^cni— •— o eg-— r-nocor- r— -x> «d- co in o oj r—u^ro-— OIT^I—
c\ji — nojco kocxii — i — «3-c\jir) »£> CM i — i — Lnrorn coLnexjoo^T^'Lr)
^001 — «d-rnmi-o r--Lncnroo'X)i — cxjin^-Lniooo «3-^-CMr^r-^i — r^
ix>^-tx>r-.cy^. — i£) cQ"X>r-vO^i_or--ro rocT^o^i — r-cr>r^ COi — r-O^oro
CMi— i— i— »— CO i— CM •— C\J CO •— i— C\J^-CXIl£.^-r-CMCOCOCO
r--coi!S-cTi'3-cor^ r-cMncokooco • — uD*^-'X)coco«cj- cr>r-^i — o^oor^^o
r*-OLouDf\Jco^)- o^r-^r^-i — oar- coiDcocrioj^-r-* *c-oLnr^o^ir)^o
C\j,_ i — , — rO. — OJi — • — OJi — OJi — • — LT)COi — i — OdOOCM
coLfirocncoaicvj coco^i^-ojr-.Lr) ^D LD cri LT> <^3 r^ «* mocnco^foco
r-,Ln^LncOLOCO r- r*. ID LO « — r — o> ^J-cOLOOOoo«d-o
, i — , — i — , — OJi — i — f — OJi — Od
, — ^-^o- — r^-cocn ocTicoO' — «=j-^^r^^vo uD-^-^-^-^-^Lr) tor^yp^oou^O
cnuDr--cocjicocn ojcooj^r^-^ro cncni — •d-coovj-es- i^ojrMLor^oooo
^i- CM ' — rocoojco ^-oocoro«!3- o-jojcoroi=d-ooco co^^j-rotrjooi-o
LT)I — cDr-OLTiro i — roc7tyDi--«^l-co roix)or--r— CNJ^J- oja^^l-coooMi —
I^D ko i^ CM un LO LO co i-T) ro ro LD uo 10 co us **• ^ r*^ to t^ > — r~^ ID i£> r--- r*-~ 10
OOOOOOO CDOOOOCDO OOOCDOOO i — OOOOOO
o ^i~ ro to r-- ^r i^j i — i^> ^r LT) cri to co cr> ^o LO LO cri ii> r-- CM i — co r*- co 01 co
• — ^OkDCOCMiOCTi OOJO^roroOxJ COO«*OOCOCM OiDOdCytLD*^-i —
roLncoiDCOLnr-. Lnr-.LOLni — r-.o cocOmu3CMr--cn c\jLnoicoi£>o\Jvo
LO vc LO cO i — i — i — o 01 oj LO ko ^o ^o oj Ln CD ID CTI CTI CO «^- crt CM ID CTI 01 O
cOi — cocM*3-«3-co rrj^i-nco^^-ch • — > — LncMLnuncM uSLncO' — counco
CMOJ • — coroco i^joji — i — unm^o r-coi — CMU^^OCO co^iJCMroiDcoio
"^-OOCMCOCDOJ OCMOi — CMOCM OJi — OOOOCM OOJOJi — CMi — O
*rcQi — (£> ^ CTI i — cooi — coi — 01 — Lnocnioco«^-r^ ir)CMcncocM> — i —
i — cocncOr-.cocr> rocncncocriuDCO rococococo^J-co nr--cococrt'd-o
ro co co ro
0_ ^^-^cg o^-r-.-r-^CM O ^- ^ ^- ^- r- CXJ O^r-r~^-^CM
CXICO--,CT>CMCMrO CMCOVCCTlOJCXjrO CMCO^CnCMCMCO CMCOLDCTtCMC\JOO
Tiii*iTi TtiiTTi iiiiiii iiiitif
i — i — «^- i-- o i — • — i — i — «^J-r~,oi — • — i — i — *d- r--. o i — ' — < — i — — 03 Cn
en ra OJ tj
0 <-0 ^ K
s- -o
rO S_
TJ (O
C T3
I SI
co •— •—
E: (O fD
\ C C
Cn O O
C C
ro +J
LO LO
Salt Lake Basin Studies
2-31
-------
I-
OQ
a <
LU S
OC .
U- '
a>
oc
O ">
Z H
O t
o z
si
x §
28
0 to
II
II
^ 3
=! tf
< 2
Q "-
CO
.0
10
V)
o
•1-
>
S-
(U
8
ro
E
^
O
S-
c
QJ
U
C
o
CJ
«
M2_
S1
w tfl
ro
0
_Q
S-
•M
C/)
-O
C
(U
QJ
S-
u_
a
•r- "C
S- S-
(T
CD i/
L,
0) +J
CD QJ
JC L,
£1
r
3
S.
a
t/
_c
c
c
1
L
a
CM
TD
+J
t/1
-
C/)
en
en
in
CO
o
ro
>
QJ
-a
c
c
1
E
X
E
.
C
O
•1-
.
ervatio
o
-M
OOOOOOO OOOOOOO OOOOOOO OOOOOOO
OOOOOOO OOOOOOO OOOOOOO OOOOOOO
o r*» o o co i— oo co r— o o co CM i — o o o o 10 co o oo o CM o co en co
o CM o o m CM in oo in o o 10 in u? oooo^p — o coo< — OCMOCO
in co vo CM vn en in ^ h^. o CM ^o CM en ^~ r**- CM co • — CM co CM oo in >—~ co ^- CM
£^3£££§ 5SSSKSS SESS^JSK SSS"^S2
cvji^oor^coo^o (vioooo-o^vo ait^cvjcncvioo 03^-CMUiin^-c\i
co 0\ co 10 CD r^ o o CM in in CM o co ^o ^ CM co cy> tn in co m CM co co uo r^«
r- r^. \& i— co en i— co ^ co en co r^ cr> co »^ 01 m cr» ^o vo ' — o c\i ^ ^o oo o^
52*SStS5 Si^StS^So S™"*^?^ S5S3«Sw£
^. r— o CM cri r*« i — 10 co co i — CD co co co r-- co o o UD ^" co i— co o o CM in
SSSSSSS g£8§SS£ £S£S£S;g] SSS^SSS
^WC-luDr-inO mi^CVJr-CSJOCO CVJCOCO(NJCT>(M^ CV,CT>CVjmU3(OW
S^C^S§?iS SSSS^SS S^^^S^S S^^^S^S
CSJOCMOOr-lD^f f-IOMCMmi^O CSJr-^fOOVOCT.*J-
co r^ r^ co co co in o en o co o ^~ ^~ en o co co co i — o O r— vo co co en cr>
coi^inujmr-CT, couiin^-un^.3- ir>Lnoo«-r^ai maiiocoio^l-un
§22^S"S Ki£££;e£g ^^^^^^^2 om^cM^^o
^SSSSSE SKSSS5S SCSS2SS S5SSSSS
OOOOOOO OOOOOOO OOOOOOO OOOOOOO
m-^^uDO^ oo.c^comoro ^cooor^^o^ m^-j.tco^r-
LnCOCMOO^-CO«d- in«d-CMCMCOCOun CMi — i — r— CM •— CM CMi — i — CMi — CM
r-^^n^in^cM COCMCOOOO^O vo^^^mroco iD^cncococo^
SMCMCO^CoS SScMCO^^S CO^^r^COOJCM COCM^r^COCMCO
^ooor^ooo- — in co i — cococom uiCMcni — oo co co coin^cooococrt
i — i — i — CMCMi — i — i — CM CMCM i — r— i — • — i —
co co ^ co r— co o CM in CT) ^ CD en r^ ".o CD o • — en o oo in o o o c\j o co
ocMCor^mcMiD o ^- o co oo o en r*--cDCMc\jiocDin CMOOOinocM
CO •— i— •—
^ocotn^voooio ^j-coio^- +J !- C
T3 r— ro CT>
O t/1 ^ S
^^
-M
c
OJ
J-
(U
o.
QJ
o
c
3 ~^- E
o a en
° 0 *
(*- O O
0 ^"ft
& -a" .
= h?
QJ ro *-
3 "^ *O
% ro c
4_ (/)+->
to
*> ?* ^
* 2&
% «^
,— 3 03
1 ^
>> ^ "*
"° ^ x
-o « i-
s •- 1
1 °-i
•*-> "* >)
42 c^
"° 0 0
>3 '^T:
ro ro
V) z ^
c
0 I '
+j i — CM
dConcentra
Standard
Standard
2-32
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
CO
O
h-
00
p
CO
O
j_ tM
is
35
UJ DC
CC <
CO
Z or
O LLJ
o
? c/T
O uj
O =
o
o
_Q
(0
l/l
o
+J -
>
s-
3.
c
o
-t->
fO
ncentr
<_)
ss
CTU3
C *
i- T3
OJ to
0) T3
O C
(U -•->
l/l
(H
o
3
-Q
S-
4->
~O
>>
O
OJ
3
OJ
S_
<_»
••- T3
S_ S-
-!-> fO
cy-o
O n
O V
U
1 -i-
0 S-
O Q
1
.C (J
+J •!-
1- (U
< e
rt3
>
S-
QJ
l/l
-Q
O
O
-a
0
£
Q_
OJ
-o
4J
00
-
*D
•*->
t/0
O>
CTi
O>
r^
tn
c
o
+-)
(TJ
>
ai
-o
c
fO
E
c
fO
E
E
3
X
£
E
3
E
c
iE
01
c:
o
+0
C
ervatic
-Q
O
Community
10 in o o vo o ^" LnCTioo(^'J"ix. dooocrif— ^- OOOJOOO^LD
ooV£) OO^^CTiCOCTlCNJO ^DOJ^ti — CO^-CO i — ^±LT)C\JOC\jr^
«j-oror^ooooCTi ^f-^t-i — uDLni^^c co> — ^j-cno-jcnio i — cn*d-cDCOooo^
^Dror^-LDLDoofo roooooLn<£>^i'«d- c\ii — cvjLOCMro i — i — i — ni — oj
r^cor-^oocooo^r CMI — Ln*±co^J-^- ^-ro<^3c\ii£Ji — o^ ovovocjiroor*-
"^D 00 ro OCOT^I — <*«a-n mi — OLOUD*^OJ
td-*d-cMroo«cj-i — r-*^o*fr*3-oj|*ovo comc\jrocTi^-Ln CNJ-td-oocxj^f^n
^-vocr>c\jocr>Ln uii — ko^J-oiuncM r-^coco*d--*^-i — Lni^mooo^J-
r^c\ji — roincxjn rorocMroinro^ i — ' — > — c\j«^c\jro r— i — i — c\jnc\jcsj
LnLnmuDCvJCTicTi CTt^-cDCOLr)^"^ CTic\jo^ «3-ror^i — r^ioir) roc\jLncn^-^-c\j
i— O O •— i— O O OOOi— •— OO O O O •— O O O OOOOOOO
*j-*3-cnrr)Ln«3-csj ini^ocr>OCTio> COCT>LT>OO^U">^I-
cnco«d-«*cor^ro cT»oo^or^ — O^ID^-OCO?>J o> — ^oicocn^-
OOOOOOi— OOOOOCD' — OOOOi — CDi — i — i — CDOi — Oi —
rOH3Ln*^-C\JCDi — OCTiCTi^£)i^3«a-CT> CTii — «J-CJiLnr^.i — COOIOLnCPtfO'JS
ojroiovor^r^uD or^cvjoiO3Ooc> «d-i£j<^>r-ro«d-oo «d-r^.ojr-.roooj
\£> OJ i — OJ^-OJ^J- OOOJOJOJinrOlD r— i — i — (M^FCSJ^- . — i — i — i — COOJ^-
00 un co c\j »— r- ro ypomcnoocxj miOOOmp-^-iD cjicnmorom^o
^o i— ro Lf) Ln LO en oJCMUDu-)<3-^i-oj r^LD^]-<*r-vr^r^ ooo*3--*oo(\j
! — i — oj oj *a- i— i — coroLo ^<*n i — ojcxjco
i — oufOLD^oojo rooj^j-*d-Loc\jn OCDCO^J-OJOO o o « — oooo
r-OOOOOO OOOOOOO OOOCDOOO OOOOOOO
^rcO' — ti3«a-CT»i — cooi — oOi — or^ Lnoo^oco^r^ looocrtcooJi — • —
• — coCTicor-roo-t c^o^CTicocnvooo rooO(jiooco«^-co nr».cococr>^J'CT»
ro fo ro ro
"-^ *-., "--^ --V. --^ --~ \ --~.*-^~--,--^---,--^.--x. -»>,--*. *^« -^. *-^ •*-^. ^-^ ^^--^--^-^'-~^^^
tNjroi£>o^CMOjro ojro^ocnoJOjn ojroi£>o^oJOjro ojrovocTtCMCsjro
i i i i i i i i i i i i i i i i i i i i i i i i i i i i
--\\\-^\\ -^\\\-^\\ \--,\\^-^^ ^^.\^^-^\
QJ
^£
(D
_l W
C c nj
CO (O QJ fO
o co v 2:
0> » £
^3 c c
OJ (U
••->
s- s-
ro ro
T3 T>
C C
«3 ra
+J 4->
-------
Table 2.1.A.8. MONTHLY TOTAL DUSTFALL
RATES (g/m2/mo) FOR UTAH CHESS
COMMUNITIES,
NOVEMBER 1970 - FEBRUARY 1972
Month
Nov. 70
Dec.
Jan. 71
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 72
Feb.
Ogden
3.57
1.16
5.29
8.10
7.39
7.58
4.51
3.19
2.53
3.40
-
3.62
1.45
0.73
1.68
4.02
Salt Lake
City
2.34
0.77
4.84
4.13
7.00
-
6.50
0.60
—
6.89
1.24
0.59
2.00
1.81
—
3.52
Kearns
1.11
0.09
1.19
2.72
4.76
4.29
4.29
2.69
1.99
2.07
0.74
0.96
1.11
—
1.38
2.98
Magna
2.86
1.65
2.45
2.72
19.00
2.64
—
3.89
2.60
2.50
4.82
3.14
3.29
3.70
3.33
3.44
Table 2.1.A.10. MONTHLY LEAD
CONCENTRATIONS (mg/m2/mo)
IN DUSTFALL FOR UTAH
CHESS COMMUNITIES,
NOVEMBER 1970 - FEBRUARY 1972
Month
Nov. 70
Dec.
Jan. 71
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 72
Feb.
Ogden
3.45
3.58
12.86
8.74
10.61
8.95
6.98
4.60
3.90
4.77
—
7.43
2.41
2.48
4.02
7.23
Salt Lake
City
5.71
2.23
12.02
11.71
13.52
7.54
17.94
0.57
—
4.61
5.12
1.77
7.29
—
10.31
8.49
Kearns
1.19
0.46
3.21
4.36
3.97
4.25
6.54
6.98
1.38
4.43
0.60
7.20
2.18
1.56
1.61
2.07
vlagna
22.98
1.84
3.21
1.95
7.75
6.08
—
4.96
1.26
5.46
4.02
4.82
2.41
2.02
2.30
2.53
Table 2.1.A.9. MONTHLY CADMIUM
CONCENTRATIONS (mg/m2/mo)
IN DUSTFALL FOR UTAH
CHESS COMMUNITIES,
NOVEMBER 1970 - FEBRUARY 1972
Month
Nov. 70
Dec.
Jan: 71
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 72
Feb.
Ogden
0.06
0.05
0.14
0.08
0.07
0.15
0.18
0.03
0.07
0.04
—
0.09
0.02
0.01
0.03
0.02
Salt Lake
City
0.12
0.00
0.07
0.07
0.10
0.12
0.19
0.03
—
0,15
0.08
0.28
0.04
—
0.08
0.03
Kearns
0.00
0.00
0.05
0.15
0.09
0.39
0.19
0.63
0.05
0.07
0.02
0.11
0.14
0.02
0.05
-
Magna
0.17
0.00
0.05
0.00
0.19
0.20
-
0.14
0.05
0.07
0.16
0.23
0.07
0.04
0.05
0.02
Table 2.1.A.11. MONTHLY ZINC
CONCENTRATION (mg/m2/mo)
IN DUSTFALL FOR UTAH
CHESS COMMUNITIES,
NOVEMBER 1970 - FEBRUARY 1972
Month
Nov. 70
Dec.
Jan. 71
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan. 72
Feb.
Ogden
1.43
0.92
4.07
3.86
4.82
7.81
5.99
2.61
5.28
4.21
—
5.76
1.26
1.19
1.95
3.10
Salt Lake
City
2.74
1.70
5.48
3.79
6.43
4.88
6.32
1.04
—
0.62
6.13
1.11
3.00
—
6.32
4.71
Kearns
0.24
0.64
1.72
3.10
3.88
5.28
3.99
7.62
2.64
3.99
0.90
3.55
3.79
0.73
5.05
5.51
Magna
3.21
0.37
0.92
0.46
6.52
5.05
-
4.68
2.98
3.54
28.33
7.69
3.33
2.76
2.76
3.10
2-34
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
CO
LLJ
O
O
CO
CO
LU
I
O
O
LLJ
CC
CC
O
o
1
CM.—
1
CM. —
00
oo . —
1
1
O- r—
1
CM.—
Z
1
CM.—
CO
1
CO J~
c z-
•f—
•1- 1
M- D- ,—
QJ I—
o
0
0 1
•r- CM.—
re z
rcu i
S_ CM.—
o co
1
CO i —
1
z •—
1
Q- . —
CM. —
1
CMi —
CO
1
CO . —
1
Q- i—
1
C
ro
13 4->
"o
oo cn
sT ID
0 0
CO VD
CO CO
0 0
sr VO
sr if)
0 0
O CO
si- VO
CD O
sr cn
sT CO
0 0
r- CO
S3- Is-
CD CD
CM ID
CM CO
O O
sr ID
CO VD
0 0
in cn
co vo
o o
cn cn
o o
r~- cn
•a- sr
O CD
Is- CO
CM CO
O CD
i— cn
si- Is-
O O
CO Is-
sl- CO
O O
. — Is-
OO VO
o o
oo cn
sr ID
CD O
cn VD
CO sT
O CD
CO CO
si- CO
0 0
O CD
VD CD
O T—
CD VO
o in
•— o
i . —
co z
1— 00
VO Is- O
in CM vo
O CD CD
cn i — vo
CM si- CO
000
CO OO •—
sj- CM ID
o o o
co cn sr
ID CM ST
o o o
O CM OO
CO i— CM
O 0 0
r- r- CO
i-- co in
O O CD
o cn oo
CM . — CM
o o o
i-- cn Is-
1D i— ST
o o o
OO CD sr
LD CM sT
O O O
VO si- CO
CO i— CM
O O O
in o ID
si- CM ID
O 0 0
VO 00 ID
CM CM CO
O O O
cn ,— i—
r-. co in
O O O
. — CO CTl
Is- CO sr
o o o
ID si- O
ID CO ID
000
si- CO 0
VO si- O
00.—
VO O CO
sr O sr
0 <— 0
O VO sj-
O si- VO
i— O CD
co Is- cn
OO si- ID
O O O
oo cn oo
O O O
1 CM CM
CO O O
oo oo z
VO
LD
CD
ID
CO
o
ID
O
£
O
in
o
ID
ID
o
co
o
ID
o
CO
in
CD
CM
VD
o
0
VO
o
cn
CM
CD
1 —
in
o
sT
VO
o
0
o
"-
o
ID
CD
CO
CD
ID
ID
CD
ID
O
CO
0
1
D-
OO
ID in
O CD
r— VD
co co
0 0
CM VD
in in
o o
O CO
VD in
0 0
cn vo
CO CO
O O'
VO CM
Is- Is-
0 0
O VO
sr CO
0 0
cn CM
o o
1 — VD
o o
si- 0
in sr
0 0
VO r—
o o
co co
CD O
o o
CO 0
O i—
O CD
O CO
•— 0
VD ID
o o
cn ,—
sj- in
o o
CO i —
co co
o o
i— cn
O CD
co cn
OO Is-
o o
OO r-
0 0
CM CM
1 1
Z CO
OO CO
CD CO
CO ID
0 0
. — Is-
CM CM
o o
CO CM
CM in
0 0
O VO
CO sT
CD 0
CO O
i — CO
0 0
CO sT
CO ID
CD O
CM O
CM CO
0 0
CM ID
CO sr
0 0
si- CO
CM CO
0 0
cn o
. — CO
0 0
o o
CO O
o •—
0 0
o co
i— 0
CO sT
o o
i— VO
CO sr
0 0
cn o
CM VO
o o
in in
co in
CD CD
OO CD
CM CM
o o
ID ID
CM ST
o o
co cn
CO sT
o o
CM sT
o o
1 1
CM CM
0 O
CO Z
cn
CO
o
cn
0
Is-
co
o
0
o
VO
o
in
0
VO
co
0
co
ID
o
cn
o
o
o
•—
CD
co
CD
cn
o
CD
o
ID
o
CM
VO
0
CO
CM
CD
*
CD
VO
CO
o
cn
o
cn
0
Q_
CO
1—
r- CO
si- ID
0 0
OO sT
0 0
ID CO
CO VO
0 0
o sr
in sr
0 0
O r—
co sr
0 0
.— o
0 0
s}- OO
CM in
0 0
cn CD
VO O
O i—
O cn
O VD
i— O
cn co
si- in
O CD
CO ID
CO ST
0 0
s}- CM
CM CO
o o
Is- CM
VD Is-
o o
VD cn
0 0
CO .—
ID ID
0 0
si- Is-
O CD
o cn
CM I—
0 0
VO Is-
in in
0 0
cn ID
VD VO
O 0
ID sT
co co
0 0
CO CO
1 1
z co
OO CO
CO CD
si- Is-
0 0
ID 1 —
VO CM
O CD
cn cn
in sr
0 0
si- cn
sT ID
0 0
O Is-
sl- CO
0 0
cn CD
CO CD
o •—
o cn
O CO
i— 0
CO CD
ID Is-
O CD
s}- .—
CM Is-
0 0
VD . —
CO ID
o o
0 «*•
CO ID
O CD
CM CO
CM CO
CD O
VO CM
co Is-
0 O
o vo
st Is-
o o
. — ID
CO ID
0 0
CO CO
CM ID
CD CD
cn . —
. — co
CD CD
CD •—
CM 1 —
CD O
ID CO
CO Is-
o o
CM . —
CM si-
0 0
1 1
CM CM
o o
co z
r-- co co
si- ID VO
O CD O
CM CO Is-
CO CO VD
O CD CD
O CD O
ID ID CD
O O i—
cn o o
CO O ID
0 .— 0
O cn o
O CO ID
.— 00
Is- cn cn
CO ID si-
CD O CD
o si- cn
sr si- in
000
i— sr co
sT sT VO
o o o
O O ID
CO ID CO
o o o
VO O Is-
VO si- CO
o o o
O ID CM
co si- in
o o o
co o co
i— CO CM
o o o
vo co vo
CO ID ID
o o o
CT> O CM
CO VO ID
o o o
. — Is- sr
in sr in
000
co sr i—
CM sT ID
o o o
CM cn oo
. — CM CM
O O CD
O CO CO
co in sr
o o o
sr co vo
CO VO ID
0 O O
sr o sr
o o o
Q. 1 1
CO Z CO
1— co oo
ID CD
sr o
O i—
CD ID
CD sj-
r- CD
Is- co
VD VD
o o
CO CO
CO ID
0 0
CM Is-
CO sj-
O CD
Is- O
CM Is-
0 0
ID CO
VO si-
CD CD
sr co
sr in
o o
00 •—
o o
cn cn
i— CO
CD CD
Is- CO
CM VO
0 0
. — o
CM CO
o o
VD «*•
co in
0 0
CO ID
0 0
sr vo
co in
O CD
vo o
co vo
0 0
. — Is-
sr CM
0 0
cn vo
CM in
CD O
vo cn
CO ID
o o
CO OO
CO sj-
o o
1 1
CM CM
o o
co z
•• E
O 1
fO O
£-0
4J <0
c
CO •
tJ O)
•r- -O
4J X
ro •!-
3
.— i.
0 4-
^
(/I
03 1
CT; CM
ro O
vt
r—
c ro
•r- 3
(O •!-
0 -P
4- ro
•r- 0.
+->
cu cu •
•o -a cu
•r- C T3
CU •!-
4-> -a
(0
Salt Lake Basin Studies
2-35
-------
DC on
II
oc
O
u.
LLI
OC
DC
O
O
O
CMr—
O ^
Z
CMr—
00
00 r—
1
Z r—
1
O_ i—
00^-
1
CMr —
Z
1
CMr—
00
1
00 r—
+J 1
E Z r—
OJ r-
O
O
c
O 1
•r- CMr—
to ~z.
o oo
t ^
i
00 r—
1
Z r—
oLr-
*~
1
CMr—
Z
1
CMr—
00
CO r—
1
Z r—
1
O- i —
*"
1
c
t— CO
O
cn
CM
o
0
CO
o
co
0
CO
o
co
o
CM
CM
0
CO
CM
o
CO
CO
o
cn
CM
0
co
LO
o
LO
o
o
£
0
co
CM
0
oo
CM
0
CM
o
LO
CM
o
CO
CD
LO
oo
o
oo
o
0
0
1 —
1
O-
oo
cn «*•
co co
O CD
«fr CO
CM CO
o o
CO LO
0 0
CM CO
LO LO
O 0
«* LO
o o
CM OJ
LO LO
o o
CO l~~
O 0
*r o
LO r-
O 0
LO LO
LO r~»
0 0
5 S
0 0
CO Cn
CO CM
o o
CO r—
LO «*
0 0
cn i—
<*!- CO
CD O
co co
LO CO
o o
CM CO
^r LO
O CD
co «=f
*f CO
0 0
oo ro
o o
r> o
p***, ^^
O r—
^D pX"
f~} r*N,
r— O
i— LO
CO CO
0 0
1 1
z oo
oo oo
CO LO
CO CM
o o
LO OO
0 r—
o o
CD CM
CM CM
0 0
LO 1^
CO CM
0 0
co p.*.
0 0
0 0
Is- LO
CM 00
o o
1^ CO
O r—
0 0
0 r—
OJ OJ
0 0
r— LO
OJ OJ
0 0
LO OJ
0 0
cn r—
CD O
i^ cn
0 0
«*• l~x
OJ OJ
0 0
LO I—
oj oo
0 0
LO 1^
oj oo
0 0
r— O
00 O
O r—
O r—
o oo
r— O
co «d-
«3- co
0 0
oo oo
o o
oo oj
0 0
1 1
OJ OJ
0 0
oo z
oj oo oj
•* LO LO
o o o
oo oo oo
CO ^- **•
o o o
LO CM 01
LO r-~ LO
o o o
r-~ LO oo
LO r>* LO
0 O O
r^ *st~ if"*1*
CM CM CM
000
LO i — OO
^- r^. LO
o o o
cn CM r —
^- LO LO
O 0 O
LO 00 CO
LO r^ ^^
O CD O
Or — CO
LO cn co
o o o
LO LO «*
O CD O
cn tj- cn
co «a- co
O O CD
r— CO LO
LO LO ^T
000
CM «S" O
LO cn o
O O r—
LO CD ^r
LO o cn
0 r— 0
O LO CM
O LO LO
r— 0 0
CO OO CM
O CD CD
LO LO "«*•
CM OJ CM
o o o
CO CO r—
LO 00 CO
O 0 0
oj co cn
^- LO "^
o o o
CM CO CO
r- oj oj
000
1 CM C\J
Q_ 1 1
co z oo
r— 00 00
CM CM
oo oo
CD O
O LO
CO O
0 0
CO r —
*r oo
o o
cn r—
o o
CM *j-
CO CO
0 0
CO t^
LO CO
o o
i— cn
LO OJ
0 0
cn r-*.
LO CO
o o
CM LO
LO CO
0 0
00 O
Kt OJ
0 0
CO O
*i- o
O r—
O CO
o «*•
r— 0
LO cn
«3- oo
o o
CO <*•
LO «*•
o o
r— cn
LO CO
0 0
cn r—
0 0
r^ cn
O CD
r— CTI
*r CM
0 0
CO CT>
LO CO
o o
r-^ LO
r— O
0 0
1 1
CM CM
o o
CO Z
i — cn r>~
CO «* LO
CD O CD
^1- LO OJ
OJ CO «3"
o o o
LO LO O
000
r— CO LO
^- r^ LO
o o o
CO O CO
^r oj oj
O O CD
LO LO VO
CO 1-^ LO
0 O 0
r*** ^o O"^
CO *± LO
o o o
LO r— O
•*!• CO 0
0 0 r—
CO O r—
•4- 0 CO
O r— CD
0 00 LO
o «r **-
r— CD O
O LO 1 —
CM CO OO
o o o
CO CM CT>
«* LD LO
000
t-^ co oo
^ 00 1^
o o o
r— r— CO
LO cn t>-
000
r^ o LO
LO LO LO
CD O CD
CM LO r —
r— CM CM
000
LO r— O
i— CM CM
000
t~~ LO O
000
r— LO ^J-
** LO LO
o o o
co cn co
LO OJ CO
000
i co co
0. 1 1
co z oo
1— oo oo
O f^
o o
*3- cn
LO CM
0 0
cn r—
<3~ LO
0 0
LO LO
"3- LO
0 0
r — OJ
0 0
oo o
LO O
O r—
o oo
O LO
r— O
cn LO
LO LO
0 0
LO LO
o o
1 — LO
co oo
CD CD
cn r-
OJ CO
O CD
r— CO
LO LO
0 0
r— CO
LO LO
0 0
OJ r—
LO r-~
0 0
cn LO
•<*• *3"
CD O
OO LO
r— CO
0 0
r^ r*^
O CM
0 0
r~~ oj
^f LO
O 0
00 CM
^- LO
o o
00 CM
CM OJ
0 0
CO CO
1 1
CM OJ
0 0
co z
oo
CM
CD
CO
o
OJ
00
0
00
o
0
o
1—
OJ
0
,d
o
CO
CM
0
o
CM
o
co
o
CO
CD
CM
CO
0
CM
0
CM
CD
CM
o
0
CD
CO
CD
o
LO
o
*
o
co
o
^
Q-
oo
oo i —
«* 1 —
O CD
O CO
OJ LO
o o
0 0
LO O
0 r—
o o
O LO
r— 0
CO OJ
r— 00
0 0
LO r—
LO LO
0 0
LO cn
o o
LO O
LO r^
0 0
OO LO
1 — LO
o o
r — LO
o o
5- oo
o o
cy> co
o o
co cn
LO LO
0 0
LO CM
0 0
r^ LO
LO LO
O CD
r^ CM
CM OJ
O CD
LO O
OO CM
o o
CO LO
LO LO
0 0
CM 00
LO «*
o o
CO 00
0 0
•d- *i-
1 1
z oo
oo oo
OJ O
*J- 0
O r—
O OJ
0 «*
r— O
OO CM
LO r^
0 0
o ^~
OJ **•
0 0
co co
r— CM
0 0
cn r*^
CM LO
0 0
LO *J-
O 0
CM CO
«* LO
0 O
LO O
CO LO
O CD
oj oo
o o
LO OO
O 00
o o
O 00
oo oo
0 0
00 CM
^- LO
0 0
CO ^~
*J- LO
0 0
co oj
oo ^~
o o
OO LO
r— OJ
0 0
LO OJ
CD OJ
0 0
00 LO
CO LO
0 0
C
o cn
•(J •!-
c c
-O CM
z
ns 0)
r— ><
O •!-
i-
10 t-
C r—
cn 3
1 CM
r— O
CO CO
c: cu
ro ro
4 3
co
•r- Q)
to
CD 3
ro ,
-"co
co
"ro •
CO I/)
r— 4J
•+5
•p i.
•t- ro
00 O-
ro
2-36
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.1.A.14. EMISSIONS FROM SMELTER AND ESTIMATED SULFUR DIOXIDE
AND SUSPENDED SULFATE (SS) CONCENTRATIONS FOR MAGNA AND KEARIMS
1940 - 1971
Year
1971
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
1955
1954
1953
1952
1951
1950
1949
1948
1947
1946
1945
1944
1943
1942
1941
1940
S0£ emissions
from smelter,
tons/day
193
261
322
281
276
301
302
311
260
349
331
247
224
202
250
448
445
489
557
581
600
649
554
614
616
475
575
734
888
892
976
970
Concentrations ,a jjg/m
Diffusion
model
Magna S02
41
56
69
60
59
64
65
67
56
75
71
53
48
43
54
96
95
105
119
124
128
139
119
131
132
102
123
157
190
191
209
208
CHESSb
Magna
S02
61.8
(84)
(103)
(90)
(88)
(96)
(97)
(100)
(83)
(112)
(106)
(79)
(72)
(65)
(80)
(143)
(142)
(157)
(178)
(186)
(192)
(208)
(177)
(197)
(197)
(152)
(184)
(235)
(284)
(286)
(313)
(310)
SS
12.4
(16.7)
(21.0)
(18.1)
(17.7)
(19.3)
19.3
20.0
16.7
22.4
21.3
15.9
14.4
13.0
16.1
28.8
28.6
31.4
35.8
37.3
38.6
41.7
35.6
39.4
39.6
30.5
37.0
47.1
57.0
57.3
62.7
62.4
Kearnsc
so2
21.7
(29)
(36)
(32)
(31)
(34)
(34)
(35)
(29)
(39)
(37)
(28)
(25)
(23)
(28)
(50)
(50)
(55)
(63)
SS
7.8
(10.5)
(13.2)
11.4
11.2
12.2
12.2
12.6
10.5
14.1
13.5
10.0
9.0
8.2
10.1
18.1
18.0
19.7
22.5
Utah State Division of
Health, S02d
Magna
107.4
120.5
(162)
(141)
(139)
(150)
(153)
(157)
(132)
(176)
(167)
(125)
(113)
(101)
(127)
(226)
(223)
(247)
(280)
(291)
(301)
(327)
(280)
(308)
(310)
(240)
(289)
(369)
(446)
(449)
(491)
(489)
Kearns
34.1
(38)
(51)
(45)
(44)
(48)
(49)
(50)
(42)
(56)
(53)
(40)
(36)
.(32)
(40)
(72)
(71)
(78)
(89)
Estimated values in parentheses.
bWest-Gaeke method.
cKearns was not established until 1953.
Conductometric method; subject to interference.
Salt Lake Basin Studies
2-37
-------
Table 2.1.A.15. ANNUAL PRODUCTION (1Q3 tons)
OF COPPER BY THE SMELTER
ATGARFIELD,UTAH,
1940-197116
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
212
269
283
271
279
197
227
267
114
226
302
326
318
271
249
2-38
HEALTH rONSEQUENCES OF SULFUR OXIDES
-------
Table 2.1.A.16. POLLUTANT SUMMARY FOR UTAH CHESS COMMUNITIES, 1940 - 1971a
Year
1971
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
1955
1954
1953
1952
1951
1950
1949
1948
1947
1946
1945
1944
1943
1942
1941
1940
Oqden
so2
8.0b
TSP
78&
78C
95C
91c
103C
94C
121C
SS
5.6b
4.4d
—
2.9d
—
4.8d
—
6.8d
Salt Lake City
so2
15. 3b
8.0°
28.0°
17. Od
17. Od
20. Od
12.0°
17. Od
(22)
26. Od
(27)
(22)
(22)
(19)
(24)
(30)
(36)
(36)
(28)
TSP
81. 2b
88. Od
90. Od
76. Od
77. Od
93. Od
114. Od
127. Od
122. Od
151.0d
145. Od
121. Od
117. Od
101. Od
132. Od
163. Od
195. Od
195. Od
153. Od
SS
~^3T
5.2d
(5.4)
5.2d
3.6d
6.9d
8.0d
7.3d
8.5d
(11.6)
(11.01
7'7^
6.5d
(6.6)
(9.7)
(12.8)
06.1)
(16.2)
(11.8)
Kearns
so2
21.7
(29)
(36)
(32)
(31)
(34)
(34)
(35)
(29)
(39)
(37)
(28)
(25)
(23)
(28)
(50)
(50)
(55)
(63)
TSP
44b
(50)
(50)
(39)
(28)
(45)
(44)
(34)
(34)
(37)
(36)
(37)
(24)
(32)
(40)
(42)
(39)
(36)
(46)
SS
7.8b
(8.3)
(9.0)
(8.6)
(8.5)
(8.8)
(8.8)
(8.9)
(8'.3)
(9.3)
(9.1)
(8.2)
(7.9)
(7.7)
(8.2)
(10.4)
(10.4)
(10.9)
(11.7)
Magna
so2
61. 8b
(84)
(103)
(90)
(88)
(96)
(97)
(100)
(83)
(112)
(106)
(79)
(72)
(65)
(80)
(143)
(142)
(157)
(178)
(186)
(192)
(208)
(177)
(197)
(197)
(152)
(184)
(235)
(284)
(286)
(313)
(310)
TSP
66b
(75)
(75)
(58)
(43)
(67)
(66)
(51)
(51)
(55)
(54)
(55)
(37)
(48)
(60)
(64)
(59)
(54)
(68)
(72)
(69)
(71)
(50)
(57)
(68)
(29)
(57)
(77)
(83)
(81)
(69)
(63)
SS
12. 4b
(14.2)
(15.9)
(14.8)
(14.6)
(15.3)
(15.4)
(15.7)
(14.1)
(16.7)
(16.2)
(13.8)
(13.1)
(12.5)
(13.9)
(19.5)
(19.4)
(20.8)
(22.7)
(23.4)
(23.9)
(25.4)
(22.6)
(24.4)
(24.4)
(20.3)
(23.2)
(27.8)
(32.2)
(32.4)
(34.8)
(34.6)
Pol.lutants are identified as follows: S02 - sulfur dioxide, TSP - total suspended particulate,
SS - suspended sulfate. All values are annual arithmetic means. Estimated values are in
parentheses.
For Kearns and Magna, estimates of annual SO? exposures were derived by multiplying the yearly
smelter emission of SOj by the ratio of the 1971 measured annual average S02 concentration to
the 1971 S02 emission rate (193 tons/day).
For Kearns and Magna, estimates of suspended sulfates were derived from estimates of S02, using
the following regression equations for 1971:
Kearns - SS = 0.10 (S02) + 5.41
Magna - SS = 0.09 (S02) + 6.66
For Kearns and Magna, estimates of annual TSP exposures were derived by multiplying the yearly
smelter production of copper by the ratio of the 1971 measured annual arithmetic mean TSP con-
centration to the 1971 copper production rate (260,000 tons/year). The assumption is made
that the smelter is the main source of TSP in Kearns and Magna. For Salt Lake City, estimates
of suspended sulfates were derived from annual average concentrations of TSP using the following
regression equation:
Salt Lake City - SS = 0.101 (TSP) - 3.65
This equation was derived using all annual data from Salt Lake City in which both SS and TSP
levels were known. The correlation coefficient is 0.74. S02 concentrations for Salt Lake City
were derived using an average ratio of S02 to TSP for all available Salt Lake City data.
bCHESS data.
cUtah State Division of Health data.
National Air Surveillance Network data.
Salt Lake Basin Studies
2-39
-------
2.2 PREVALENCE OF CHRONIC RESPIRATORY
DISEASE SYMPTOMS IN ADULTS: 1970 SURVEY
OF SALT LAKE BASIN COMMUNITIES
Dennis E. House, M.S., John F. Finklea, M.D.,Dr. P.H.,
Carl M. Shy, M.D., Dr.P.H., Dorothy C. Calafiore, Dr. P.H.,
Wilson B. Riggan, Ph. D., J. Wanless Southwick, Ph. D.,
and Lyman J. Olsen, M.D.
2-41
-------
INTRODUCTION
Air pollution episodes characterized by extremely
high levels of sulfur dioxide and total suspended
participates are associated with increased respiratory
disease symptoms and increased mortality.1"4 Studies
of the chronic effects of relatively low levels of sulfur
dioxide have been less conclusive. Bell, Petrilli et al.,
and Tsunetoshi et al. found positive associations
between low-level sulfur dioxide concentrations and
chronic bronchitis.5'7 Holland et aL.and Holland and
Reid likewise found an increased number of episodes
of cough, phlegm, and chest illnesses in men who
lived in a polluted area of London. British males age
50 to 59 living in London were compared to men of
the same ages in three towns where the sulfur dioxide
concentrations were lower and found to manifest
more frequent and more severe chronic respiratory
disease symptoms.8'9 In contrast, Zeidberg et al.
could find no relationship between respiratory
diseases and very low-level sulfur dioxide concentra-
tions.10 Furthermore, Anderson, Ferris, and
Zichmantel found no association between the
prevalence of chronic respiratory disease and low
levels of sulfur dioxide.11'13 Most of these studies
were hampered by an inability to separate the effects
of sulfur dioxide from those of suspended particu-
lates. Inadequate air monitoring and intervening
variables such as age, socioeconomic status, and
smoking habit limit the scope of most earlier studies.
cigarettes would be additive; and fourth, that relative-
ly brief peak exposures in a community somewhat
more distant from the point exposure source would
be accompanied by increases in one or more indices
of chronic respiratory disease morbidity.
MATERIALS AND METHODS
Community Selection
On the basis of existing aerometric, meteorologic,
and topographic data, four Salt Lake Basin com-
munities were chosen to represent a gradient of sulfur
dioxide exposure. These four communities were: a
section of Ogden (the Low exposure area), a section
of Salt Lake City (the Intermediate I exposure area),
Kearns (the Intermediate II exposure area), and
Magna (the High exposure area). The primary source
of sulfur dioxide emissions was a large copper smelter
located approximately 5 miles northwest of the High
exposure community. The Intermediate II exposure
community is located approximately 8 miles south-
east of the smelter, the Intermediate I exposure
neighborhood approximately 13 miles east of the
smelter, and the Low exposure community approxi-
mately 38 miles northeast of the smelter.
The purpose of this paper is to describe a survey
designed to estimate the effect of sulfur dioxide and
suspended sulfate exposure on the prevalence of
chronic respiratory disease symptoms. The study was
conducted in communities that have similar total
suspended particulate levels but represent an
exposure gradient for sulfur dioxide and suspended
sulfates, which largely emanate from a single indus-
trial source, a copper smelter. The corporation
involved began reducing the emissions of air pol-
lutants several years ago and is making further
substantial efforts to control pollutant discharges.
Four specific hypotheses were tested: first, that
exposure to elevated community air pollution levels
in the immediate vicinity of the industrial source
would be accompanied by significantly increased
indices of chronic respiratory disease frequency or
severity; second, that the prevalence of chronic
respiratory disease symptoms in the more polluted
community would increase with increasing length of
residence when compared to populations living in
cleaner communities; third, that the effects of com-
munity air pollution and self-pollution by smoking
Collection of Health and Demographic Data
During autumn 1970, chronic respiratory disease
information was ascertained from parents of
elementary, junior high, and senior high school
students. In each of the four communities, three
elementary schools, one junior high school, and one
senior high school were selected on the basis of
similar socioeconomic status. Each community was a
middle-class, predominantly white area. Elementary
school children carried home the questionnaire, while
parents of junior and senior high school students were
contacted by mail.
The questionnaire, which was adapted for self-
administration from that version standardized by the
British Medical Research Council, inquired about the
duration and presence of cough, phlegm (excluding
phlegm from the nose), and shortness of breath. The
reliability of this type of self-administered question-
naire has been tested and found to be very satisfac-
tory.7 An example of the form used is presented
elsewhere.14
2-42
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
The mother or female guardian was instructed to
answer the questions for parents living in the home.
Other information ascertained included length of
residence in the community, educational attainment
of head of household, cigarette smoking habit, age,
and whether the parent had occupational exposure to
respiratory irritants. It should be made clear that
covariate information was applicable to the time at
which the survey was conducted. For example,
exsmokers were individuals who at the time of the
survey were not smoking, but who had smoked in the
past.
Assessing Air Pollution Exposure
At the time of this study, air monitoring stations
were located within each community. Each station
was on a building rooftop at a height of about 25 feet
above the ground. At each station, 24-hour integrated
samples of sulfur dioxide (modified West-Gaeke
method), total suspended particulates (high-volume
samplers), suspended sulfates, and suspended nitrates
were monitored daily. Daily measurements of nitro-
gen dioxide (Jacobs-Hochheiser method) were
also made; however, the validity of the method has
subsequently been questioned by the Environmental
Protection Agency. Dustfall was determined from
monthly samples. A full description of the monitor-
ing station locations and aerometric procedures has
been presented elsewhere.15
While these stations allow estimates of pollutant
exposure for prospective studies, it was also necessary
to estimate exposure for past time periods. Sulfur
dioxide had been monitored in the Low and Inter-
mediate I communities since 1965 and in the High
community since 1970 by the Utah State Division of
Health. In addition, measurements of total suspended
particulates and suspended sulfates dating back to
1953 were available for the Intermediate I com-
munity. Estimates of sulfur dioxide, total suspended
particulates, and suspended sulfate concentrations in
the High exposure community for 1940-1970 and the
Intermediate II exposure community for 1950-1970
were obtained by a mathematical dispersion model,
which utilized emissions from the industrial source
and extensive local meteorological data, and by
observed relationships among pollutants. Observed
suspended particulate, suspended sulfate, and sulfur
dioxide concentrations for 1970-1971 were used to
calibrate the models used to estimate exposure levels
for previous years.
Data Analysis
For the purposes of statistical analysis, a scale of
symptom severity was devised by a panel of physi-
cians based upon the number, clinical significance,
and duration or persistence of symptoms. Although
the number of symptom categories was somewhat
arbitrary and their severity ranking was subjective and
no more than ordinal in nature, the following ranking
of chronic respiratory symptoms categories was de-
veloped:
1. No symptoms.
2. Cough alone for less than 3 months each year.
3. Phlegm with or without cough for less than 3
months each year.
4. Cough without phlegm for 3 months or more
each year.
5. Phlegm without cough for 3 months or more
- each year.
6. Cough and phlegm for 3 months or more each
year.
7. Cough and phlegm for 3 months or more each
year and shortness of breath.
Reported respiratory symptoms were used to
devise three indices of morbidity; first, chronic
bronchitis was defined in accordance with the British
Medical Research Council (cough and phlegm on most
days for at least 3 months each year), and prevalence
rates were compared across areas. The second variable
was a mean respiratory symptom score defined as the
arithmetic average of the respiratory symptom score
of each member of a particular population group. The
third variable was the arithmetic average of the
respiratory symptom score for those members of the
population who reported any chronic respiratory
symptoms, that is, for those members who had
symptom scores 2 through 7 only. The first two
variables reflect frequency of chronic respiratory
disease symptoms, with the mean symptom score
having the advantage of considering a gradient of
symptoms but the disadvantage of lacking the ac-
ceptance and the assured clinical relevance of clas-
sically defined chronic bronchitis. The mean severity
score, which is the third variable, was utilized as a
measure of severity of reported symptoms.
Hypotheses were tested using a general linear
model for categorical data.16 This technique uses
weighted regression on categorical data and allows
estimation of each effect in the model after adjusting
for all other effects in the model. Hypotheses were
tested by Chi Square procedures. Covariates used in
Salt Lake Basin Studies
243
-------
the analysis model were age and cigarette smoking
habit. Ir.terr fion terms were not included.
RESULTS
Environmental Exposure
Pollutant exposures already projected from aero-
metric and dispersion model estimates were averaged
for each community by decade from 1940 through
1959 and by 4-year intervals thereafter (Table 2.2.1).
More extensive aerometric data for the year 1971 are
isolated in the table for comparative purposes. Proce-
dures utilized in these projections are detailed else-
where.15
Estimated arithmetic mean sulfur dioxide levels
for the period 1950-1970 show that the community
exposure gradient very likely existed for two decades
prior to this study. During this two-decade period,
estimated annual average sulfur dioxide levels in all
communities, except the town nearest the industrial
source, have been quite low (17 to 50 Mg/m3) and
well within the National Annual Primary Ambient Air
Quality Standard for sulfur oxides. In the High
exposure community, estimated annual levels of
sulfur dioxide have been falling steadily from a high
of 234 jug/m3 three decades ago to 92 to 95 jug/m3
for 1960-1970. For 1971, the levels were somewhat
lower,
Two factors accounted for lower sulfur dioxide
levels during 1971 in the High exposure community.
New emission controls for the industrial source were
implemented in 1971. In addition, a near cessation of
industrial activity in July 1971 caused an extreme
drop in ambient levels of sulfur dioxide.
In the 3-year period before the survey was
conducted, 1968-1970, total suspended participate
levels were minimally elevated (84 to 88 Mg/m3) in
the Low and Intermediate I exposure areas. Earlier
annual exposures (1950-1967) in the Intermediate I
community were probably moderately high (103 to
151 jUg/m3). In the other two communities, sus-
Table 2.2.1. AIR POLLUTANT EXPOSURE ESTIMATES (jug/m3) IN
FOUR UTAH COMMUNITIES, 1940-1971
Pollutant and community
Sulfur dioxide
Low
1 ntermediate 1
Intermediate II
High
Total suspended particulates
Low
Intermediate 1
Intermediate II
High
Suspended sulfate
Low
Intermediate 1
Intermediate II
High
Suspended nitrates0
Low
Intermediate 1
Intermediate II
High
1940-49
NAa
NA
_b
234
NA
NA
_b
63
NA
NA
_b
28
—
—
—
—
1950-59
NA
28
50
142
NA
151
<50
60
NA
11
10
20
—
—
—
—
1960-63
NA
27
33
95
NA
134
<50
53
NA
8
8.7
15.2
—
—
—
—
1964-67
NA
17
33
95
108
103
<50
57
5.8
7.7
8.7
15.3
—
—
—
—
1968-70
NA
18
32
92
88
84
<50
70
3.7
4.7
8.6
15.0
—
—
—
—
1971
8
15
22
62
78
81
45
66
5.6
7.3
7.8
12.4
2.7
3.3
2.4
2.0
aNA—not available.
kjhe Intermediate II city did not exist until 1953.
°Only fragmentary data available for suspended nitrates from 1940-1970.
2-44
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
pended particulate levels (<50 to 70 jug/m3) were
thought to have been well below the National Annual
Primary Ambient Air Quality Standard for the last
three decades. Estimated and measured suspended
particulate sulfates evidenced a community gradient
in pollution exposure similar to that for sulfur
dioxide, being exceedingly low (<5.8 jug/m3) in the
Low exposure community, moderate (5 to 11 jug/m3)
in the Intermediate I and II communities, and high
(12 to 28 jug/m3) near the industrial emissions source.
Suspended particulate nitrate levels (2.0 to 3.3
Mg/m3) were similar for all areas.
Three air quality trends should be highlighted:
First, there appeared to be a general decrease in the
levels of sulfur dioxide and particulate air pollution
over the last decade, which was probably due, at least
in part, to the pollution control efforts of local
industry and governmental units. Second, the Inter-
mediate I and II communities, as well as the High
community, appeared to experience elevated annual
suspended sulfate levels attributable to the point
source of industrial emissions. Annual average sulfur
dioxide levels in the Intermediate I and II com-
munities were not greatly elevated. Third, there
appeared to be similar community exposure gradients
for sulfur dioxide and suspended sulfates, which
makes it difficult to separate the effects of these
pollutants.
Community Characteristics
Four community characteristics of special impor-
tance in the assessment of chronic respiratory disease
are respondent rates, exclusion rates, socioeconomic
status, and cigarette smoking habits. Parents of
elementary school children completed and returned
87 percent of the questionnaires, while parents of
junior and senior high children, to whom question-
naires were mailed, completed and returned 35
percent of the questionnaires (Table 2.2.2). For both
groups of parents, intercommunity differences were
observed. Parents of elementary school children in
the two Intermediate exposure areas returned a
somewhat smaller proportion of questionnaires than
the High and Low exposure communities (p <
0.001). Parents of junior and senior high school
children in the Low and Intermediate I exposure
areas returned smaller proportions of questionnaires
than in the Intermediate II and High exposure
communities (p < 0.001).
Because 2 years seemed a minimum time in which
chronic respiratory disease might develop, re-
spondents were excluded from the analyses when
they had not lived in their community for at least 2
years prior to the survey. Parents were also excluded
if they reported occupational exposures to irritating
fumes, dusts, or aerosols. Lastly, parents were
excluded if information needed for the analyses was
missing from the questionnaire. The percent of
mothers excluded was approximately the same in all
four communities; more fathers were excluded in the
High exposure community than in other communities
because many fathers from the High exposure
community worked in a nearby industry where they
reported occupational exposure to irritants, fumes,
dusts, or gases, including sulfur dioxide (Table 2.2.2).
A sample of 27 to 36 nonrespondent families from
each community were interviewed by telephone and
found to be similar to respondents in education,
smoking habits, family size, and length of residence.
Table 2.2.2. COMMUNITY CHARACTERISTICS: QUESTIONNAIRE RESPONSE,
EXCLUSIONS, AND EDUCATIONAL ATTAINMENT OF FATHERS
Community
Low
Intermediate 1
Intermediate II
High
Total
Percent of families responding
Distributed
through
elementary
school
91
83
86
88
87
Mailed to
families of
junior and senior
high school students
34
32
40
37
35
of respondents
excluded3
Mothers
16
21
19
22
19
Fathers
27
34
35
50
37
Percent of
fathers who
completed
high school
78
66
68
73
71
aBecause of incomplete response, less than 2 years residence, or occupational exposure.
Salt Lake Basin Studies
2-45
-------
Among the families responding to the question-
naire, there were small intercommunity differences in
educational achievement (Table 2.2.2). The Low
exposure area had an educational ranking only
slightly higher than the High exposure community,
which, in turn, ranked above the two Intermediate
exposure areas. This intercommunity socioeconomic
pattern could hardly diminish intercommunity dif-
ferences in chronic respiratory disease that might be
attributable to residence in the High exposure corn-
munity because comparison communities ranked
socioeconomically higher and lower than the High
exposure community.
Sex and cigarette smoking are such strong deter-
minants of chronic respiratory disease that the study
population of each community was categorized by
sex and history of cigarette smoking before chronic
respiratory disease morbidity indices were evaluated
(Table 2.2.3). There were no significant intercom-
munity differences in the intensity of smoking among
respondents who currently smoked cigarettes. Like-
wise, cigarette smokers of both sexes living in the
High exposure community were no more likely to be
heavy smokers (> 1-1/2 packs per day) than their
counterparts in the other three communities. When
the smoking intensity of either mothers or fathers
was considered, the High exposure community
ranked midway among the four study areas. The low
percentage (28 percent) of parents who reported
being current smokers can be explained by the fact
that the study communities were composed largely of
individuals of a religious group whose members are
typically nonsmokers.
Chronic Bronchitis Prevalence
The suspected determinants of chronic bronchitis
other than ambient air pollution were first evaluated
by calculation of sex-specific chronic bronchitis
prevalance rates (Table 2.2.4). As expected, cigarette
smoking proved the strongest determinant, followed
by sex and educational attainment. However, educa-
tional attainment was linked to cigarette smoking,
with those who were less educated more likely to be
cigarette smokers. Within sex- and smoking-specific
categories, no consistent trend that was attributable
to educational attainment could be found. For all but
one of the examined categories—smoking, education,
and age—fathers were found to have higher rates than
mothers. The only exception to this sex effect was a
somewhat lower chronic bronchitis prevalence among
fathers who had stopped smoking cigarettes. Little or
no differences could be attributed to aging in this
population, probably because of the narrow age span.
Seemingly elevated rates in the 29-year-or-younger
group were based upon a relatively small sample size.
To determine the effects of ambient air pollution
on the prevalence of chronic bronchitis, smoking- and
sex-specific rates for each community were cal-
culated, and these revealed a clear trend towards
excess illness in the High exposure community (Table
2.2.5). In four of six possible smoking- and sex-
specific comparisons, chronic bronchitis rates in the
Intermediate II exposure community fell below the
markedly elevated rates of the High exposure area
and above the relatively low rates of the Low and
Table 2.2.3. STUDY POPULATION DISTRIBUTED BY COMMUNITY, SEX, AND SMOKING STATUS
Community
Low
Intermediate I
Intermediate II
High
Total
Nonsmokers
Mothers
755
755
772
667
2949
Fathers
396
367
350
265
1378
Exsmokers
Mothers
75
101
114
84
374
Fathers
230
177
241
133
781
Smokers
Mothers
214
286
295
212
1007
Fathers
272
311
354
209
1146
Total
1942
1997
2126
1570
7635
2-46
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.2.4. SEX-SPECIFIC CHRONIC BRONCHITIS PREVALENCE
DISTRIBUTED BY SMOKING HISTORY,
EDUCATIONAL ATTAINMENT, AND AGE
Determinant
Smoking history:
Nonsmokers
Exsmokers
Smokers
Education:
High school
Age:
<29
30 to 39
40 to 49
>50
Chronic bronchitis prevalence, percent
Mothers
3.5
5.9
17.1
7.7
7.2
6.1
9.4
6.2
7.1
4.2
Fathers
3.7
4.2
20.9
13.8
9.1
8.1
13.4
9.6
8.8
10.5
Both sexes
3.6
4.8
18.3
10.3
8.0
7.0
10.5
7.6
7.9
8.1
Table 2.2.5. SMOKING- AND SEX-SPECIFIC CHRONIC BRONCHITIS PREVALANCE
RATES (percent) DISTRIBUTED BY COMMUNITY
Community
Low
Intermediate I
Intermediate II
High
Nonsmokers
Mothers
2.3
2.0
4.7
5.2
Fathers
3.0
3.6
2.3
6:8
Exsmokers
Mothers
5.3
4.0
7.0
7.1
Fathers
2.6
3.4
5.4
6.0
Smokers
Mothers
17.8
14.7
15.3
22.2
Fathers
19.9
18.6
20.1
26.8
Intermediate I exposure communities. Rates were
highest for current cigarette smokers of either sex
who lived in the most polluted community.
In five of the eight possible intracommunity
comparisons between exsmokers and nonsmokers in
Table 2.2.5, chronic bronchitis rates for exsmokers
were somewhat higher than those of lifetime non-
smokers. In every intracommunity comparison, rates
for exsmokers were distinctly lower than those for
current cigarette smokers. Differences between the
sexes were inconsistent across smoking categories.
Among current cigarette smokers, males consistently
exhibited a higher chronic bronchitis prevalence,
while the converse was true for exsmokers. No real
sex pattern could be found among nonsmokers.
Tests for statistical significance of these effects, as
well as the effect of age, are presented separately for
mothers and fathers in Table 2.2.6. Significant
increases in the prevalence of chronic bronchitis in
both sexes were associated with cigarette smoking
and residence in the more polluted community.
Significant increases did not occur in the Inter-
mediate II exposure area. The nonsignificant Chi
Squares for the fit of the models indicate that linear
models containing only smoking, age, and community
pollution exposure effects adequately fit the data.
An additional analysis was performed with the
data on the parents of junior and senior high school
students omitted. In this analysis, the excess in the
prevalence of chronic bronchitis in the High exposure
Salt Lake Basin Studies
2-47
-------
Table 2.2.6. ANALYSIS OF VARIANCE FOR HEALTH OBSERVATIONS,
CHRONIC BRONCHITIS RATES
Determinant
Pollutant
exposure
Smoking3
Age
Fit of
model
Degrees
nf
freedom
3
1
1
10
Fathers
X2
11.15
178.23
3.12
8.54
Probability
p<0.001
p<0.001
0.05
-------
Table 2.2.8. ANALYSIS OF VARIANCE FOR HEALTH OBSERVATIONS,
MEAN RESPIRATORY SYMPTOM SCORES
Determinant
Pollutant
exposure
Smoking8
Age
Fit of
model
Degrees
of
freedom
3
1
1
10
Fathers
X2
25.62
344.58
1.75
10.67
Probability
p<0.001
p<0.001
0.10.9
Mothers
X2
11.87
8.03
7.11
11.30
Probability
p<0.01
p<0.01
p<0.01
0.3
-------
Occupational Exposure
Length of Residence
Fathers in the High exposure community with
confirmed occupational exposure to irritant dusts,
fumes, gases, or aerosols were compared to the
fathers who were not occupational^ exposed to
pollutants. There were 174 nonsmoking and 105
smoking occupationally exposed fathers, and 385
nonsmoking and 205 smoking fathers who were not
occupationally exposed. Preliminary analyses
indicated that age was interacting with exposure,
hence separate analyses were done on the age groups
over and under 40. For the age group under 40, there
was no significant difference between occupationally
exposed and unexposed fathers in either chronic
bronchitis prevalence, symptom scores, or severity
scores (Table 2.2.11). Significantly higher chronic
bronchitis prevalence rates and mean symptom scores
were seen in a similar comparison involving older
occupationally exposed fathers. Conditional severity
scores were also increased, but not significantly, in
older exposed workers.
The net effect of excluding occupationally
exposed parents from the main analyses was to
diminish any effects attributable to pollution. The
incremental effect of occupational exposure in the
older workers was roughly twice the effect at-
tributable to ambient air pollution and over half that
of cigarette smoking. Occupational exposures to air
pollutants, self-pollution by cigarette smoking, and
ambient air pollution appear to be additive hazards.
Safety standards set to protect workers from inhaled
irritants will be imprecise, if not ineffective, unless
self-pollution by smoking and ambient air pollution
exposures are duely considered in the standard-setting
process.
Two questions prompted analysis of duration of
residence in the study communities: first, did com-
munity air pollution influence migration patterns by
deterring more vulnerable citizens from moving into a
polluted community or by increasing out-migration
of citizens whose ill health may be aggravated by
pollution? Second, what length of exposure to the
pollutant levels found in the present study was
associated with significant increases in chronic
bronchitis prevalence?
Relative prevalence ratios and expected chronic
bronchitis prevalence rates were computed for five
residence categories in the High exposure community
based upon the pooled experience of the three Low
and Intermediate exposure communities. Smoking-
specific relative prevalence plots suggested that a
deficit of persons with chronic bronchitis moved into
the High exposure area and a probable excess moved
out after a lapse of 10 to 12 years (Figure 2.2.1). This
effect was most pronounced in the current non-
smokers group, which was composed of exsmokers
and lifetime nonsmokers. Possibly, people subject to
chronic bronchitis avoided self-pollution by not
smoking and minimized ambient air pollution by
their choice of occupation and residence. Comparison
of the observed and expected chronic bronchitis rates
at successively greater lengths of residence established
that significant increases in illness occurred both in
smokers and current nonsmokers after 4 to 7 years.
Relative Effect of Community Air Pollution
and Cigarette Smoking
Cigarette smoking and ambient air pollution are
both significant determinants of chronic bronchitis
Table 2.2.11. SMOKING - ADJUSTED CHRONIC BRONCHITIS RATES, MEAN RESPIRATORY
SYMPTOM SCORES, AND MEAN RESPIRATORY SEVERITY SCORES BY
OCCUPATIONAL EXPOSURE IN THE HIGH EXPOSURE COMMUNITY
Occupational exposure
to smoke, dust.
or fumes
Yes
No
Chronic bronchitis
rate, percent
<40
yr olds
18.4
16.0
>40
yr olds
25.9a
17.9
Overall mean
severity scores
<40
yr olds
2.56
2.60
>40
yr olds
3.26a
2.58
Mean severity
scores of fathers
with symptoms
<40
yr olds
4.63
4.54
>40
yr olds
5.07
4.67
Significantly increased.
2-50
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
6 8 10
RESIDENCE DURATION years
Figure 2.2.1. Relative risk of chronic
bronchitis in High exposure community
compared to pooled risk for all other
communities.
morbidity, but what is the relative importance of
these insults? Are their effects additive? To answer
these questions, the experience of lifetime non-
smokers living in the three relatively clean com-
munities was used to compute a base prevalence rate
for each sex separately. For each smoking and
community pollution category, the excess prevalence
was calculated by subtracting the appropriate base
rate from the smoking and community specific
prevalence rate (Table 2.2.12).
Should the deleterious effects of cigarette smoking
and ambient air pollution be additive, then the sum
of the excess prevalence for smokers living in clean
communities plus the excess prevalence for non-
smokers from the polluted community should equal
the excess prevalence for smokers living in the
polluted community. When these excess prevalences
were summed, both sexes had somewhat higher
excess prevalences (117 and 128 percent) than
predicted by the additive model. When the same
additive model was applied to exsmokers, air pol-
lution and cigarette smoking again appeared to be
additive (64 and 79 percent) of the predicted value.
To assess the relative importance of air pollution
where the effects of air pollution and cigarette
smoking are additive, the excess prevalence of chronic
bronchitis for nonsmokers in the most polluted
community may be divided by the excess prevalence
for current cigarette smokers living in cleaner com-
munities. When this was done for males, air pollution
Table 2.2.12. RELATIVE IMPORTANCE OF CIGARETTE SMOKING AND
AMBIENT AIR QUALITY AS DETERMINED BY COMPARISON OF
EXCESS PREVALENCE FOR CHRONIC BRONCHITIS
Current
smoking
status
Females
Lifetime nonsmokers
Exsmokers
Smokers
Males
Lifetime nonsmokers
Exsmokers
Smokers
Community
air
quality
Clean (pooled)5
Dirty0
Clean (pooled)
Dirty
Clean (pooled)
Dirty
Clean (pooled)
Dirty
Clean (pooled)
Dirty
Clean (pooled)
Dirty
Excess
prevalence
of chronic
bronchitis3
0.00 (3.0)
2.2
3.0
4.1
12.8
19.2
0.00 (3.0)
3.8
0.9
3.0
16.6
23.8
Simple additive
model combining
adverse effects of
smoking and pollution
Expected
—
—
—
5.2
—
15.0
—
—
—
4.7
—
20.4
Observed/
expected
—
—
—
0.79
—
1.28
—
—
—
0.64
—
1.17
aBase rates in parentheses. Excess prevalence = smoking- and community-specific prevalence rate minus
base rate.
"Based on the Low, Intermediate I, and Intermediate II exposure communities combined.
°Based on the High exposure community.
Salt Lake Basin Studies
2-51
-------
was found to have about one-fourth (0.23) the effect
of cigarette smoking; for females, air pollution was
found to have about one-tenth (0.09) the effect of
cigarette smoking. In summary, excess prevalence
comparisons indicate that a large portion of excess
chronic bronchitis morbidity can clearly be attributed
to ambient air pollution and that the combined
effects of pollution and cigarette smoking appear to
be additive.
DISCUSSION
Three of the four hypotheses advanced in this
study were validated by the analyses. The frequency
of chronic bronchitis and overall mean respiratory
symptom scores were significantly increased among
current nonsmokers and cigarette smokers of both
sexes who lived in the High exposure community.
There was, however, no evidence that the severity of
chronic respiratory disease symptoms differed be-
tween communities. Chronic bronchitis prevalence
was found to increase significantly after 4 to 7 years
exposure to the elevated ambient air pollution levels
found in the most polluted community. Excess
prevalence comparisons indicated that excess chronic
bronchitis morbidity attributed to community air
pollution and cigarette smoking are additive. This
study could not differentiate the effects of elevated
annual average sulfur dioxide concentrations from
those of peak exposures to sulfur dioxide. Further-
more, where sulfur dioxide levels were elevated,
suspended sulfates were also quite high. However, in
the Intermediate I and II communities, exposures to
elevated levels of suspended sulfates occurred in the
absence of elevated annual average exposures of
sulfur dioxide. These communities did not differ
significantly from the cleaner community, even
though there was a trend towards more chronic
bronchitis and higher mean symptom scores for
respiratory symptoms. One might interpret this find-
ing as suggesting that annual average exposures to
suspended sulfates of 11 to 12 Mg/m3 could be near
the threshold for induction or magnification of
chronic respiratory disease symptoms. Since past
suspended particulate exposures in the Low and
Intermediate I exposure areas had been moderately
elevated, it is possible that these less polluted
communities may still harbor a small residual
morbidity excess, making the present estimates of air
pollution effects conservative.
There was also some evidence that people with
chronic respiratory disease symptoms hesitated to
migrate to the most polluted community and a
suggestion that residents who developed such
symptoms preferentially migrated from the High
exposure community to a more salubrious climate.
Ambient air pollution was found to have roughly
one-tenth to one-fourth the deleterious effect of
cigarette smoking and one-half that of occupational
exposure to irritant gases, dusts, fumes, or aerosols.
As expected, cigarette smoking history, sex, and
occupational exposures were significant determinants
of chronic respiratory disease frequency. Educational
attainment and age were not strong determinants in
the present study. Moreover, intercommunity dif-
ferences in educational attainment were small. Each
of these variables was duely considered in the
analysis.
One intervening variable, the influence of non-
respondents, must be considered further. To equalize
the prevalence of chronic bronchitis, major intercom-
munity differences in morbidity would have to affect
nonrespondents in a direction opposite to the dif-
ferences observed in respondents. To equalize chronic
bronchitis rates, nonrespondent smokers in the
cleaner areas would have to double illness rates of
respondents, and nonrespondent cigarette smokers
from the cleaner areas would have to report one-and-
one-half more illness than respondents. Such differ-
ences are conceivable, but hardly likely. One other
exclusion, that of occupationally exposed workers in
the High exposure community, clearly diminished
intracommunity differences attributable to air pollu-
tion.
After the effects of all intervening variables were
duly considered, we concluded that it is only prudent
from a public health standpoint to assume that
increases in the prevalence of chronic bronchitis can
be related to 4- to 7- year exposures to ambient air
pollution. Such exposures were characterized by
elevated levels of sulfur dioxide (92 to 95 Mg/m3)
accompanied by low suspended particulate levels (53
to 70 Mg/m3) and elevated suspended sulfate levels
(15 Mi/m3) in the absence of elevated levels of
suspended nitrates. Clearly, additional clinical and
toxicological laboratory research will be necessary to
define the relative importance of sulfur dioxide and
suspended sulfates. Additional epidemiologic studies
will also help piece together the puzzle since urban
levels of sulfur dioxide are falling much more rapidly
than levels of suspended sulfates. Physical and
chemical characterization of suspended sulfates must
be improved, and a more complete understanding of
the relevant atmospheric chemistry is clearly needed.
2-52
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
SUMMARY
The prevalence of chronic bronchitis was found to
increase significantly among smokers and nonsmokers
of both sexes after 4 to 7 years exposure to ambient
air pollution characterized by elevated annual levels
of sulfur dioxide (92 to 95 jug/m3) and suspended
sulfates (15 jug/m3) accompanied by low levels of
total suspended particulates (53 to 70 Mg/m3).
Among the 7635 adults who comprised the study
population, chronic bronchitis prevalence was highest
in smokers, intermediate in exsmokers, and lowest in
lifetime nonsmokers. The effects of air pollution and
self-pollution by smoking cigarettes appeared to be
additive. Occupational exposure to irritating dusts,
fumes, gases, or aerosols was a strong determinant of
illness after age 40, having twice the effect of air
pollution and over half the effect of cigarette
smoking. Community pollution seemed to influence
migration patterns in that there was a relative deficit
of persons with chronic respiratory disease symptoms
among new arrivals in the High exposure community;
furthermore, there was evidence that symptomatic
persons preferentially migrated away from the High
exposure community. All segments of society should
gain substantial health benefits with the control of
sulfur dioxide air pollution.
REFERENCES FOR SECTION 2.2
1. Firket, J. Fog along the Meuse Valley. Trans.
Faraday Soc. 32:1192-1197,1936.
2. Firket, J. The Cause of the Symptoms Found in
the Meuse Valley during the Fog of December
1930. Bull. Roy. Acad. Med. Belgium.
11:683-739,1931.
Yamagata. Epidemiological Study of Chronic
Bronchitis with Special Reference to Effect of
Air Pollution. Int. Arch. Arbeitsmed. 29:1-27,
1971.
Holland, W. W., D. D. Reid, R. Seltser, and R. W.
Stone. Respiratory Disease in England and the
United States; Studies of Comparative Preva-
lence. Arch. Environ. Health. 70:338-345,1965.
3. Schrenk, H. H., H. Hermann, G.O. Clayton,
W. M. Gafafer, and H. Wexler. Air Pollution in
Donora, Pennsylvania; Epidemiology of the
Unusual Smog Episode of October 1948. Federal
Security Agency, Division of Industrial Hygiene,
Public Health Service, U.S. Department of
Health, Education and Welfare. Washington, D.C.
Public Health Bulletin 306. 1949.
4. Wilkens, E. T. Air Pollution Aspects of the
London Fog of December 1952. Roy. Meteorol.
Soc. 1.50:267-271,1954.
5. Bell, A. The Effects on the Health of Residents
of East Port Kembla, Part II. In: Air Pollution by
Metallurgical Industries (Vol. 2). Sydney,
Australia, Division of Occupational Health, New
South Wales Department of Public Health, 1962.
p. 1-144.
6. Petrilli, R. L., G. Agnese, and S. Kanitz.
Epidemiology Studies of Air Pollution Effects in
Genoa, Italy. Arch. Environ. Health. 72:733-740,
1966.
7. Tsunetoshi, Y., T. Shimizu, H. Takahashi, A.
Ichinosawa, M. Ueda, N. Nakayama, and Y.
9. Holland, W.W. and D. D. Reid. The Urban
Factor in Chronic Bronchitis. Lancet. 1:445-448,
February 27, 1965.
10. Zeidberg, L.D., R.A. Prindle, and E. Landau. The
Nashville Air Pollution Study; III. Morbidity in
Relation to Air Pollution. Amer. J. Public
Health. 54:85-97,1964.
11. Anderson, D. O., B.C. Ferris, and R.
Zickmantel. Levels of Air Pollution and Respira-
tory Disease in Berlin, New Hampshire. Amer.
Rev. Respiratory Dis. 90:877-887, 1964.
12. Anderson, D. 0. and B. G. Ferris. Air Pollution
Levels and Chronic Respiratory Disease. Arch.
Environ. Health. 10:307-311,1965.
13. Ferris, B. G. and D. 0. Anderson. The Prevalence
of Chronic Respiratory Disease in a New
Hampshire Town. Amer. Rev. Respiratory Dis.
86:165-177, August 1962.
14. Questionnaires Used in the CHESS Studies. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Salt Lake Basin Studies
2-53
-------
Protection Agency. Research Triangle Park, N. C. Health Consequences of Sulfur Oxides: A Report
Publication No. EPA-650/1-74-004. 1974. from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N. C.
Publication No. EPA-650/1-74-004. 1974.
15. Hertz, M. B., L. A. Truppi. T. D. English, G. W.
Sovocool, R.M. Burton, L. T. Heiderscheit, and 16. Grizzle, I.E., C. F. Starmer, and G.G. Koch.
D. 0. Hinton. Human Exposure to Air Pollutants , Analysis of Categorical Data by Linear Models.
in Salt Lake Basin Communities, 1940-1971. In: Biometrics. 25(3):489-504, September 1969.
2-54 HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
2.3 FREQUENCY OF ACUTE LOWER RESPIRATORY
DISEASE IN CHILDREN: RETROSPECTIVE SURVEY
OF SALT LAKE BASIN COMMUNITIES, 1967-1970
William C. Nelson, Ph.D., John F. Finklea, M.D., Dr. P.H.,
Dennis E. House, M.S., Dorothy C. Calafiore, Dr.P.H.,
Marvin B. Hertz, Ph.D., and Donald H. Swanson, B.S.
2-55
-------
INTRODUCTION
Children provide a unique opportunity to assess
the effect of air pollutants on the respiratory tract.
The young are vulnerable to acute respiratory disease.
Unlike adults, children have not usually been
subjected to self-pollution by cigarette smoking, to
occupational exposures to respiratory irritants, or to
a bewildering series of residential changes. Any of
these factors can profoundly confound needed
estimates of past exposures to ambient air pollutants.
When ambient air pollutant levels can be largely
attributed to a single industrial point source or when
long-term community aerometric data are available,
one can determine both recent and past exposures.
Furthermore, since air pollution control technology
must be directed towards single pollutants, it is
especially important to disentangle the effects of
multiple pollutant exposures. This has rarely been
possible in the past.
British studies of children exposed to particulate
and sulfur dioxide pollution indicated a direct
relationship between level of exposure and incidence
of respiratory disease. Frequency and severity of
acute respiratory tract infections increased with the
level of exposure to sulfur dioxide and particu-
lates.1'3 Significant findings have been reported from
other countries. School children aged 10 to 11 in a
polluted area of Japan had higher frequency of
nonproductive cough, irritation of the upper respira-
tory tract, and increased mucus secretion than
matched children in a less polluted area.4 Similar
findings were reported among school children in
Irkutsk, U.S.S.R.5 While these previous studies
yielded remarkably consistent results, each suffers
from the inability to separate the effects of sulfur
dioxide from those of suspended particulates. More-
over, none could relate the length of sulfur dioxide
and particulate exposure to excesses in acute respira-
tory morbidity. However, a single report has linked
nitrogen dioxide exposures lasting 3 years to an
increase in childhood respiratory illness.6
The retrospective survey described here attempted
to isolate the effects of sulfur dioxide and suspended
sulfate air pollution from the influence of elevated
levels of total suspended particulates and oxides of
nitrogen. Three specific questions were asked: First,
would a greater proportion of children from the
polluted communities experience acute lower respira-
tory illness? Second, would a greater proportion of
children from the polluted communities experience
repeated episodes of acute lower respiratory illness?
Third, would pollutant exposures lasting longer than
2 years alter the susceptibility of children to acute
lower respiratory illness?
MATERIALS AND METHODS
The study was carried out in four communities in
the Salt Lake Basin that exhibited similar particulate
and nitrogen dioxide pollution levejs but represented
an exposure gradient for sulfur dioxide and sus-
pended sulfates. These four communities were: an
•Ogden neighborhood (the Low exposure area), a Salt
Lake City neighborhood (the Intermediate I exposure
area), Kearns (the Intermediate II exposure area), and
Magna (the High exposure area). The primary source
of sulfur dioxide emissions was a large smelter located
approximately 5 miles northwest of the High
exposure community. The Intermediate II exposure
community is located approximately 8 miles south-
east of the smelter, the Intermediate I exposure
neighborhood approximately 13 miles east of the
smelter, and the Low exposure community approxi-
mately 38 miles northeast of the smelter.
Collection of Health and Demographic Data
During autumn 1970, lower respiratory illness
(LRI) information was ascertained retrospectively by
parental reporting for each of their children 12 years
old or younger. In each of the four communities,
three elementary schools, one junior high school, and
one senior high school were selected for study
inclusion on the basis of similar socioeconomic status
(SES). These areas were each middle-class white
communities. Each elementary school child carried
home a School and Family Health Questionnaire.
Parents of junior and senior high school children were
contacted by mail and asked to complete the same
questionnaire. (An example of the questionnaire used
is presented elsewhere.)7 Response was encouraged
by reminder notes to nonrespondents, and a sample
of nonrespondents was interviewed by telephone.
The questionnaire inquired about the frequency of
treatment by a physician for pneumonia, croup, or
bronchitis (including bronchiolitis or deep chest
infections other than pneumonia or croup) during the
period beginning in September 1967. Hence all rates
used in the analyses (except for the two figures) are
3-year rates. Other information ascertained included
hospitalizations for lower respiratory illnesses, name
of children's physician, history of asthma, length of
residence in the community, socioeconomic status
measured by education of head of household,
parents' smoking status, and family census. Physician
2-56
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
records were reviewed to validate parental reporting
for approximately a 15 percent sample of question-
naires. The sample included an equal number of
children reported sick and well. Additionally, a
sample of 53 physicians in the four areas were
queried about their definition of the bronchitis
syndrome.
Assessing Air Pollution Exposure
At the time of this study, air monitoring stations
were located within each community. Each station
was on a building rooftop at a height of about 25 feet
above the ground. At each station, 24-hour integrated
samples of sulfur dioxide (modified/West-Gaeke
method), total suspended participates (high-volume
samplers), suspended sulfates, suspended nitrates, and
nitrogen dioxide (Jacobs-Hochheiser method) were
monitored daily. Dustfall was determined from
monthly samples. A full description of the location of
monitoring stations and methodology has been
presented elsewhere.8
These stations allow estimates of pollutant
exposure for prospective studies, but it was also
necessary to characterize past exposures. Sulfur
dioxide had been monitored in the Low and Inter-
mediate I communities since 1965 and in the High
community since 1970 by the Utah State Division of
Health. In addition, measurements of total suspended
particulates and suspended sulfates dating back to
1953 were available for the Intermediate I com-
munity. Estimates of sulfur dioxide, total suspended
particulate, and suspended sulfate concentrations in
the High exposure community for 1940-1970 and the
Intermediate II exposure community for 1953-1970
were obtained by a mathematical dispersion model
procedure utilizing emissions output of the industrial
source and extensive local meteorological data.
Observed suspended particulate, suspended sulfate,
and sulfur dioxide concentrations for 1970-1971
were utilized to calibrate the model used to predict
exposure levels for previous years.8
Hypotheses Testing
The hypotheses tested were: first, that the
frequency of physician treatment for any lower
respiratory illness, croup, bronchitis, and pneumonia,
as well as the frequency of hospitalization for these
illnesses, adjusted for appropriate covariates, would
correspond to pollutant exposures; second, that there
would be no significant intercommunity differences
in the physician definition of lower respiratory illness
syndromes for the four communities; and third, that
the accuracy of illness reporting by parents would be
similar for all communities. For each morbidity
condition, two analyses were done. In the first, the
dependent variable was the percent of children
reporting one or more episodes. For the second, the
dependent variable was the percent of children
reporting repeated (two or more) episodes.
Hypotheses were tested by a general linear model
for categorical data.9 This technique utilizes weighted
regression on categorical data and allows estimation
of each effect in the model adjusted for all other
effects in the model. Significance was ascertained by
Chi Square procedures. Covariates used in the model
were age (three categories: 1 to 4, 5 to 8, 9 to 12),
sex, and educational attainment (two categories:
< completed high school, > completed high school),
which was an index of socioeconomic status.
RESULTS
Environmental Exposure
Pollutant exposures were projected for the illness
reporting period covered by the survey (1967-1970)
and for the earlier growth and development years of
the childhood population (Table 2.3.1). The latter
period was partitioned into two 4-year intervals,
1963-1966 and 1959-1962, which reflected changes
in air quality and coincided with the earliest
exposures of children age 5 through 8 and 9 through
12. Younger children, age 1 through 4, were primarily
exposed to pollution levels monitored during the
1967-1970 study period.
The postulated intercommunity exposure
gradients were confirmed for sulfur dioxide and
suspended sulfates, and such gradients apparently had
existed throughout the lifetime of the children
surveyed. However, from 1959 to 1970, estimated
annual average sulfur dioxide levels were quite low
(16 to 33 jug/m3) in all but the High exposure
community, which apparently had distinct elevations
(91 to 94 //g/m3). Elevations in estimated suspended
sulfates (9 to 15 /ug/m3) occurred in the two
communities near the industrial emissions source. The
highest estimated annual sulfate levels were found in
the most proximal community.
Annual arithmetic mean suspended particulate
levels (82 to 133 jug/m3) were moderately elevated in
the Low and Intermediate I communities, but not in
Salt Lake Basin Studies
2-57
-------
Table 2.3.1. PROJECTED AIR POLLUTANT EXPOSURES IN FOUR
UTAH COMMUNITIES, 1959-1971
Pollutant and community
Sulfur dioxide
Low
Intermediate I
Intermediate II
High
Total suspended
particulate
Low
Intermediate I
Intermediate II
High
Suspended sulfates3
Low
Intermediate I
Intermediate II
High
Suspended nitrates'3
Low
Intermediate I
Intermediate II
High
Concentration, /Lig/m^
1959-1962
_c
27
32
94
_c
133
<50
51
_c
8
9
t5
—
—
—
—
1963-1966
_c
16
33
94
109
114
<50
59
6
8
9
15
—
—
—
—
1967-1970
_c
18
33
91
92
82
<50
62
4
5
9
15
—
—
—
—
1971
8
15
22
62
78
81
45
66
6
7
8
12
2.7
3.3
2.4
2.0
aData given for 1959 to 1970 are estimated.
bQnly fragmentary data are available for suspended nitrates from 1959 to 1970.
cNot available.
the Intermediate II and High exposure areas ( <50 to
62 /Ltg/m3). No data for nitrogen dioxide are given
because the validity of the Jacobs-Hochheiser method
of analysis has been seriously challenged. However,
there was no evidence that the National Primary
Ambient Air Quality Standard (100 Mg/m3) was
exceeded in any of the four communities.
Two factors accounted for relatively lower 1971
annual average of sulfur dioxide in the High exposure
community compared with 1959-1970 values. New
emission controls for the industrial source were
implemented in 1971 and, more importantly,
cessation of industrial activity in July 1971 caused a
precipitous drop in sulfur dioxide levels.
Community Characteristics
Intercommunity differences in response rates, non-
respondent characteristics, respondent exclusions,
socioeconomic status, family size, and cigarette
smoking might alter the interpretation of inter-
community differences in acute respiratory illness
reported in the present study. Overall, almost two-
thirds of the distributed questionnaires were
completed and returned (Table 2.3.2). As expected,
distribution through elementary schools was more
effective than the mail-out to families of junior and
senior high school students. Intercommunity differ-
ences in respondent rates were observed (p < 0.001).
The High exposure community ranked second in the
proportion of questionnaires returned. A small
sample of nonrespondents (27 to 36) in each com-
munity was interviewed by telephone. This sample
was adequate to ascertain a demographic profile but
too small to yield useful morbidity information.
Nonrespondents in the High and Low exposure
communities faithfully mirrored the demographic
characteristics of respondents. More variation, but no
large deviations, were found between respondents and
nonrespondents in the two Intermediate exposure
areas.
2-58
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.3.2. COMMUNITY CHARACTERISTICS: QUESTIONNAIRE
RESPONSE AND RESPONDENT EXCLUSIONS
Community
Low
Intermediate 1
Intermediate II
High
Total
Percent of families responding
Distributed through
elementary school
91
83
86
88
87
Mailed to families
of junior and senior
high school students
34
32
40
37
35
Percent of nonasthmatic
children excluded
Because of missing
data on education.
sex, or illness
6
7
7
6
7
'Because of missing
data on age
g
9
10
8
9
One-sixth of the nonasthmatic children in respond-
ing families were excluded because of missing data on
sex, educational attainment of the head of household,
illness reporting, or age. No children less than 1 year
of age were included in the principal analysis, and
these accounted for most of the age exclusions.
However, such children were included in a special
analysis that attempted to relate duration of air
pollution exposure to increased lower respiratory
illness. Almost identical percentages of respondents
were excluded from each community, with the High
exposure community having one of the two lowest
overall exclusion rates.
Small intercommunity differences were also found
in educational attainment (p < 0.001). Small, but
nonsignificant, community differences were found in
mean family size and proportion of parents who were
current cigarette smokers (Table 2.3.3). In each case,
the High exposure community ranked intermediate
when the four study communities were compared. In
summary, there was no evidence of intercommunity
differences that would bias the results of the survey.
Asthmatics
As planned in the protocol, asthmatics were
excluded from the final analysis of illness. In each
area, about 5 percent of the children were reported as
having a history of asthma. The number of asthmatics
proved too small to permit adequate adjustments for
other determinants of acute lower respiratory illness.
Except for the High exposure area children, who were
reported having significantly higher illness rates for
croup, no significant area differences in lower
respiratory illnesses or hospitalizations were observed
(Table 2.3.4). However, asthmatics exposed to higher
levels (9 to 15 Mg/m3) of suspended sulfates, whether
living in the High or Intermediate II exposure areas,
tended to report total respiratory illness, pneumonia,
bronchitis, and hospitalizations more frequently than
asthmatics living in communities with lower sulfate
Table 2.3.3. COMMUNITY CHARACTERISTICS: EDUCATIONAL
ATTAINMENT OF FATHERS, MEAN FAMILY SIZE, AND CURRENT
CIGARETTE SMOKING HABITS OF PARENTS
Community
Low
Intermediate 1
Intermediate II
High
Total
Percent of fathers
who completed high school
78
66
68
73
71
Mean family
size
6.0
5.7
6.4
6.0
6.0
Percent of parents who
currently smoke cigarettes
Mothers
20
25
25
22
23
Fathers
31
37
38
35
35
Salt Lake Basin Studies
2-59
-------
Table 2.3.4. ASTHMATIC CHILDREN REPORTING AT LEAST ONE LOWER RESPIRATORY
ILLNESS DISTRIBUTED BY COMMUNITY AND CLINICAL DIAGNOSIS
Community
Low
Intermediate I
Intermediate II
High
Number
of
subjects
96
no
156
113
Percent reporting each respiratory syndrome
Any LRI
50
56
62
60
Croup
29
24
30
43
Bronchitis
43
54
56
58
Pneumonia
10
12
18
15
Hospitalization
10
12
17
14
and sulfur dioxide exposures. Thus, the exclusion of
asthmatic children appears to diminish any overall
excesses in lower respiratory illness in the High and
Intermediate II exposure areas that might later be
attributed to sulfur dioxide or suspended sulfate
pollution.
Total Acute Lower Respiratory Illness
Nonasthmatic respondents who comprised the
study population of 8991 were then distributed by
length of residence within their respective com-
munities (Table 2.3.5). Sample sizes for 2 years of
residence were quite small, and were combined with
durations of < 1 year prior to analysis. Younger
children experienced more lower respiratory illnesses,
with attack rates highest for 2 year olds and steadily
decreasing thereafter to a plateau beginning at age 10
(Figure 2.3.1). Average attack rates over the study
period showed similar patterns for both sexes.
Separate analyses indicated that socioeconomic status
was related to illness reporting in a complex fashion.
Thus, any analysis relating intercommunity dif-
ferences in illness to ambient air pollution should
isolate or adjust for the effects of age, duration of
residence, and socioeconomic status. Data in the
following tables reflect such adjustment; unadjusted
data are presented in the Appendix, Tables 2.3.A.1 to
2.3.A.5.
The effect of each of the postulated major personal
determinants of acute lower respiratory disease (and
of each specific lower respiratory disease component
considered) was tested. In only 4 of 30 tests was
there evidence of deviation from the model as-
sumption, and 3 of these occurred with tests involv-
ing croup. Results of the analyses for total acute
respiratory illness are given in Table 2.3.6. As
expected, age was a powerful determinant of illness
frequency, but no consistent sex or socioeconomic
effects were found.
Children living in the High exposure area were
found to report single and repeated episodes of acute
Table 2.3.5. NONASTHMATIC CHILDREN USED IN ALL ILLNESS
ANALYSES DISTRIBUTED BY
COMMUNITY AND RESIDENCE DURATION
Number of subjects
Community
Low
Intermediate 1
Intermediate II
High
Total
<1 yr
residence
153
236
253
222
864
2yr
residence
81
99
90
94
364
>3yr
residence
1719
1968
2161
1915
7763
Total
1953
2304
2504
2231
8991
2-60
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
AGE AT LAST BIRTHDAY, years
Figure 2.3.1. Age-specific annual mean attack
rates for acute lower respiratory illness.
lower respiratory illness significantly more frequently
than children living in the other three areas (Table
2.3.7). Such excesses were entirely due to significant-^
remarked increases in reported illness among children
residing in the High pollution area for 3 or more
years. Although not significant, children living in the
High exposure community less than 3 years reported
less illness, suggesting that families with exquisitely
susceptible children may avoid the High exposure
community and that exposures of up to 3 years may
be necessary before excess morbidity occurs. There
are at least two possible explanations of the low rates
in children who have lived in the High exposure
community less than 3 years: First, since the presence
of chronic respiratory disease symptoms in adults
accurately predicted the occurrence of excess lower
respiratory disease in their children,10 perhaps
parents of the more susceptible children were
prevented from being employed and then moving into
the more polluted community by occupational health
preplacement screening. Secondly, the High pollution
community is small and hence may have fewer
endogenous strains of infectious organisms that
typically affect migrants than the larger, more hetero-
genous Low and Intermediate I pollution com-
munities.
Lesser increases also occurred in the Intermediate
II exposure community, suggesting that either in-
frequent peak exposures to sulfur dioxide or long-
term exposures to suspended particulate sulfates
might also be accompanied by excess acute lower
respiratory morbidity.
Table 2.3.6. ANALYSIS OF VARIANCE FOR CATEGORICAL DATA:
SUMMARY OF PROBABILITIES BY NUMBER OF EPISODES
AND RESIDENCE DURATION,
ANY LOWER RESPIRATORY DISEASE
Factor
Pollution
exposure
Age
Sex
Socioeconomic
status
Pollution
exposure
Age
Sex
Socioeconomic
status
Number
of
episodes
^1
>2
Probability of effect for indicated
residence duration8
<3yr
NS
<0.001
NS
<0.05
NS
<0.001
NS
NS
>3yr
<0.001
<0.001
NS
NS
<0.001
<0.001
NS
NS
Any length
<0.001
<0.001
NS
<0.10
<0.001
<0.001
NS
NS
^S—not significant, p >0.10.
Salt Lake Basin Studies
2-61
-------
Croup
Laryngotracheobronchitis, or croup, was analyzed
in a similar fashion (Table 2.3.8). Croup mimicked
the findings already reported for overall lower respira-
tory illness. Significant increments in single or
repeated attacks of croup were especially striking
among children residing in the High exposure area for
as long as 3 years. As before, new arrivals in the High
exposure area actually had less illness than their
counterparts in the less polluted communities. No
trend towards a morbidity increment was noted in
the Intermediate II community, which was exposed
to relatively high suspended sulfate levels. Croup was
reported significantly more often by the better
educated families and was more frequent in younger
children (Table 2.3.9).
Bronchitis
Reporting patterns for acute bronchitis were
similar to those of croup and overall acute lower
respiratory illness (Table 2.3.10). Significant
morbidity excesses in single or repeated illness
episodes could be attributed to residence in the High
exposure community. Again, these excesses were
entirely due to significant increases occurring after 2
years of residence. As before, younger children were
more likely to experience single or repeated episodes
diagnosed as acute bronchitis, and advantaged
families were more likely to report repeated episodes
among their children (Table 2.3.11).
Table 2.3.7. AGE-SEX-SES ADJUSTED 3-YEAR
ATTACK RATES FOR ANY LOWER RESPIRATORY ILLNESS
BY RESIDENCE DURATION AND NUMBER OF EPISODES
Community
Low
Intermediate 1
Intermediate II
High
Low
Intermediate 1
Intermediate II
High
Number of
illness episodes
>1
>2
Attack rate, percent
<3yr
residence
25.6
27.8
27.5
22.9
19.2
14.6
14.2
12.3
>3yr
residence
27.3
26.5
29.0
38.2
15.2
14.6
17.2
23.4
Any length
residence
27.1
27.0
28.7
36.1
15.7
14.6
16.9
21.9
Table 2.3.8. AGE-SEX-SES ADJUSTED 3-YEAR ATTACK RATES FOR CROUP
BY RESIDENCE DURATION AND NUMBER OF EPISODES
Community
Low
Intermediate I
Intermediate II
High
Low
Intermediate I
Intermediate II
High
Number of
illness episodes
>1
>2
Attack rate, percent
<3yr
residence
15.7
11.7
11.7
9.4
10.3
5.0
5.0
4.7
>3yr
residence
16.9
14.5
16.5
26.4
8.0
5.7
7.8
13.8
Any length
residence
16.9
14.3
16.3
24.3
8.4
5.7
7.5
12.5
2-62
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.3.9. ANALYSIS OF VARIANCE FOR CATEGORICAL
DATA: SUMMARY OF PROBABILITIES BY NUMBER
OF EPISODES AND RESIDENCE DURATION, CROUP
Factor
Pollution
exposure
Age
Sex
Socioeconomic
status
Pollution
exposure
Age
Sex
Socioeconomic
status
Number
of
episodes
>1
>2
Probability of effect for indicated
residence duration3
<3 yr
NS
<0.001
NS
<010
NS
<0.05
NS
<0.10
>3yr
<0.001
<0.001
NS
<0.10
<0.001
<0.001
NS
<0.10
Any length
<0.001
<0.001
NS
<0.05
< 0.001
<0.001
NS
<0.05
aNS-not significant, p >0 10.
Table 2.3.10. AGE-SEX SES-ADJUSTED 3-YEAR ATTACK RATES
FOR BRONCHITIS BY RESIDENCE DURATION
AND NUMBER OF EPISODES
Community
Low
Intermediate I
Intermediate II
High
Low
Intermediate I
Intermediate II
High
Number of
illness episodes
>1
>2
Attack rate, percent
<3yr
residence
169
18.8
17.6
16.4
11.3
6.9
8.6
47
>3yr
residence
165
16.7
17.1
23.6
6.5
65
7.8
10.8
Any length
residence
16.6
17.3
17.3
22.8
7.0
6.7
8.2
10.1
Table 2.3.11. ANALYSIS OF VARIANCE FOR CATEGORICAL
DATA: SUMMARY OF PROBABILITIES BY NUMBER
OF EPISODES AND RESIDENCE DURATION, BRONCHITIS
Factor
Pollution
exposure
Age
Sex
Socioeconomic
status
Pollution
exposure
Age
Sex
Socioeconomic
status
Number
of
episodes
>1
>2
Probability of effect for indicated
residence duration8
<3yr
NS
<0.001
NS
NS
NS
<0.01
NS
<0.05
>3yr
<0.001
<0.001
NS
NS
<0.001
<0.001
NS
<0.01
Any length
<0.001
<0.001
NS
NS
<0.01
<0.001
NS
<0.05
aNS—not significant, p > 0 10
Salt Lake Basin Studies
2-63
-------
Pneumonia
Single or repeated episodes of physician-diagnosed
pneumonitis were infrequent in all study com-
munities (Table 2.3.12). No trends in intercom-
munity differences attributable to pollution exposure
were seen. As expected, pneumonia reporting
decreased with age. However, increases in pneumonia
bordering on statistical significance were related to
lower socioeconomic status (Table 2.3.13).
Hospitalizations Resulting from Acute Lower
Respiratory Illness
Hospitalizations, which were even less frequent
than pneumonia, seemed influenced by the same
factors as pneumonia (Tables 2.3.14 and 2.3.15).
Table 2.3.12. AGE-SEX-SES ADJUSTED 3-YEAR ATTACK RATES FOR PNEUMONIA
BY RESIDENCE DURATION AND NUMBER OF EPISODES
Community
Low
Intermediate 1
1 ntermediate 1 1
High
Low
Intermediate 1
Intermediate II
High
Number of
illness episodes
>1
>2
'
Attack rate, percent
<3yr
residence
4.9
7.7
7.7
6.0
3.0
2.8
2.6
2.0
>3yr
residence
4.4
6.5
5.0
4.8
1.2
1.3
1.1
1.2
Any length
residence
4.4
6.8
5.5
4.9
1.0
1.7
1.9
1.1
Table 2.3.13. ANALYSIS OF VARIANCE FOR CATEGORICAL DATA:
SUMMARY OF PROBABILITIES BY NUMBER OF EPISODES
AND RESIDENCE DURATION, PNEUMONIA
Factor
Pollution
exposure
Age
Sex
Socioeconomic
status
Pollution
exposure
Age
Sex
Socioeconomic
status
Number
of
episodes
>1
>2
Probability of effect for indicated
residence duration8
<3yr
NS
<0.001
NS
NS
NS
NS
NS
NS
>3yr
NS
<0.001
NS
NS
NS
<0.001
NS
NS
Any length
NS
<0.001
NS
<0.10
NS
<0.001
NS
<0.10
aNS-not significant, p >0.10.
2-64
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.3.14. AGE-SEX-SES ADJUSTED 3-YEAR HOSPITALIZATION RATES
BY RESIDENCE DURATION AND NUMBER OF ADMISSIONS
Community
Low
Intermediate 1
Intermediate II
High
Low
Intermediate 1
Intermediate II
High
Number of
admissions
> 1
>2
Attack rate, percent
<3yr
residence
3.6
4.4
4.9
4.2
2.9
1.8
2.7
1.8
>3yr
residence
2.5
2.8
2.1
2.1
0.4
0.6
0.5
0.8
Any length
residence
2.4
3.0
2.6
2.7
0.5
0.5
0.7
0.7
Table 2.3.15. ANALYSIS OF VARIANCE FOR CATEGORICAL DATA:
SUMMARY OF PROBABILITIES BY NUMBER OF ADMISSIONS
AND RESIDENCE DURATION, HOSPITALIZATION
Factor
Pollution
exposure
Age
Sex
Socioeconomic
status
Pollution
exposure
Age
Sex
Socioeconomic
status
Number
of
admissions
>1
>2
Probability of effect for indicated
residence duration8
<3yr
NS
<0.01
NS
NS
NS
NS
NS
NS
>3yr
NS
<0.001
NS
<0.05
NS
<0.05
NS
NS
Any length
NS
<0.001
<0.10
NS
NS
<0.01
NS
NS
aNS-not significant, p >0.10.
Recent hospitalizations were more often reported for
younger children from less advantaged families. Thus,
the increases in acute respiratory illness morbidity
attributed to community air pollution, while serious
enough to merit physician attention, were unlikely to
have been immediately life-threatening. Otherwise,
hospitalizations would have paralleled overall
morbidity and increased significantly in the more
polluted community. Adjusted reporting rates for
hospitalizations and pneumonia both tended to be
higher in newly arrived children in every study
community.
Age-specific Attack Rates
Selected age-specific lower respiratory illness
attack rates adjusted to remove the potentially
Salt Lake Basin Studies
2-65
-------
confounding effects of sex and socioeconomic status
were calculated for three age groups (1 to 4, 5 to 8,
and 9 to 12) in each study community (Table
2.3.16). Since children of every residence duration
were considered, this analysis provided a conservative
estimate of intercommunity differences attributable
to ambient air pollution exposures. The effects of
sex, socioeconomic status, and ambient air pollution
were then tested by the linear model previously
described, and a probability summary was compiled
(Table 2.3.17). Intercommunity differences
attributable to the experience of children living in the
High exposure community, which were thought to be
a manifestation of combined sulfur dioxide-
suspended sulfate exposures, could be separated in
the analyses from effects that accompanied the
overall exposure gradient for suspended particulate
sulfates. The expectation was that effects of the lower
sulfate (9 jug/m3) exposures experienced by children in
the Intermediate II community might not be manifest
until later in childhood than would effects of the
higher sulfur dioxide (91 to 94 jug/m3) and
Table 2.3.16. SEX-SES ADJUSTED 3-YEAR AGE-SPECIFIC ATTACK RATES FOR ANY LOWER
RESPIRATORY ILLNESS AND BRONCHITIS DISTRIBUTED BY COMMUNITY
Illness
Any lower
respiratory
illness
Bronchitis
Community
Low
Intermediate 1
Intermediate II
High
Low
Intermediate 1
Intermediate II
High
Attack rate, percent
1 to 4 yr
olds
38.6
36.9
37.3
49.0
24.6
24.0
21.4
30.9
5 to 8 yr
olds
26.6
29.1
30.0
32.2
15.9
17.9
18.6
19.3
9 to 12 yr
olds
15.7
14.7
19.5
25.3
9.0
9.2
11.9
17.2
Table 2.3.17. PROBABILITY SUMMARIES FOR THE EFFECTS OF POLLUTION EXPOSURE,
SEX, AND SOCIOECONOMIC STATUS ON AGE-SPECIFIC ACUTE LOWER RESPIRATORY
ILLNESS RATES
Illness
Any lower
respiratory
illness
Bronchitis
Age,
years
1 to 4
5 to 8
9 to 12
1 to 4
5 to 8
9 to 12
Source of variation3
Pollution exposure
Sulfur dioxide
<0.001
NS
<0.001
<0.05
NS
<0.001
Suspended
sulfate
<0.01
<0.05
<0.001
NS
<0.10
<0.001
Sex
NS
NS
NS
NS
NS
NS
Socio-
economic
status
NS
NS
NS
NS
NS
NS
aNS—not significant, p > 0.10.
2-66
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
sulfate levels (15 Mg/m3) found in the High exposure
community. This was indeed the case. The High
exposure community reported more illness at every
age when either total acute respiratory illness or acute
bronchitis was considered. Intercommunity dif-
ferences were large in the youngest and oldest
children and attenuated in those who were just
entering school or kindergarten. Perceptible illness
gradients corresponding to the exposure gradient for
suspended sulfates were found in the youngest school
children and were unmistakable and statistically
significant in children over 9 years of age.
There was evidence, then, that 9- to 13-year
exposures to annual average suspended sulfate levels
of 9 jug/m3 may be accompanied by increased
respiratory illness. Even though this illness gradient
could best be explained by suspended sulfate
exposures, there remains a distinct possibility that
short-term peak exposures to sulfur dioxide might be
responsible. This is unlikely since the most recent
monitoring data revealed that 24-hour peak sulfur
dioxide exposures during the last quarter of 1970 and
the year 1971 were similar in the two Intermediate
exposure areas. Similar analyses for croup,
pneumonia, and hospitalization were not performed
because of restricted sample sizes. In the age-specific
rate analysis, the previously recorded weak, yet
significant, effects of socioeconomic status were not
apparent.
Clinical diagnosis of croup, the lower acute respira-
tory illness syndrome most likely to show the earliest
initial effect on acute lower respiratory illness, was
selected to determine the shortest sulfur dioxide-
suspended particulate exposure that will induce a
significant increment in acute lower respiratory
illness. Age-specific attack rates for croup among
children whose families lived in the Low and Inter-
mediate I communities for 3 or more years were
compared to attack rates for similar children residing
in the High exposure area (Figure 2.3.2). In this
analysis, the experience of boys and girls was
combined to give attack rates greater stability. With-
out doubt, excess croup was present in the 1 to 2
year old age group. Finer analyses showed an effect in
the first year of life. Was the excess in croup caused
by short-term peak pollution exposures or longer
term annual exposures? This remains a key question
that could not be answered by the present retrospec-
tive report but may be answered by prospective
studies currently in progress. There can be no doubt,
however, that exposures to elevated sulfur dioxide
(91 Mg/m3) and suspended sulfate (15 Mg/m3) levels
4 S
AGE AT LAST BIRTHDAY, years
Figure 2.3.2. Age-specific annual mean
attack rates for croup distributed by
community pollution level.
lasting only a year can be accompanied by significant
increases in the frequency of childhood croup.
Age-specific illness rates were also utilized in tests
of the significance of indoor air pollution emanating
from cigarette smoking by parents. When this was
done, the cigarette smoking habits of parents were
found to be a significant determinant of acute lower
respiratory conditions. The model may have failed to
completely separate smoking effects from socio-
economic influences since persons of lower socio-
economic status were more likely to be cigarette
smokers. In any case, cigarette smoking within the
home was associated with significant increases in
acute bronchitis and pneumonia in preschool children
age 1 to 4, who are most heavily exposed to home
pollution from side-stream cigarette smoke. Un-
explained paradoxical effects upon the reporting of
croup were found: both families without any smokers
and families with heavy smokers reported croup more
frequently than families with light smokers. The
effects of cigarette smoking, however, could not
explain the previously discussed intercommunity
differences in childhood acute lower respiratory
illness because both the prevalence of smoking and
the intensity of cigarette smoking were similar for all
Salt Lake Basin Studies
2-67
-------
communities. Moreover, significant statistical
relationships persisted between ambient air pollution
and increased reporting of acute bronchitis and croup
in analytical models that considered the cigarette
smoking habits of parents.
Diagnostic Patterns and Validation
of Questionnaire-reported Illness
Striking intercommunity differences in physician
diagnoses for the same set of respiratory symptoms in
a young patient could be reflected by significant
intercommunity differences in illness reporting by
parents who responded to the survey questionnaire.
This did not occur. Fifty-three physicians who cared
for children comprising the study population were
asked to diagnose six respiratory syndromes as either
"bronchitis" or "not bronchitis." The symptom
complexes in each syndrome were graded ranging
from pharyngitis with minimal, but unmistakable,
lower respiratory symptoms to classical pneumonia.
Bronchitis was purposefully chosen since bronchitis
in children is a somewhat imprecise diagnosis. This
allowed a better chance to discover any latent
intercommunity differences. Physicians and their
diagnostic decisions were tallied for each community
(Table 2.3.18). Since it was common for physicians
to have clinical practices that included children from
three of the four study communities, a pool of their
answers was tallied. The fact that children from three
geographically clustered study communities consulted
a common pool of physicians was in itself an
unexpected safeguard against diagnostic bias. Physi-
cians from different communities generally had the
same diagnostic pattern when considering pharyn-
gitis, classical bronchitis, early pneumonia, or classical
pneumonia. Differences were observed when either
the mild bronchitis-deep chest cold syndrome or the
croup syndrome was considered. Although no
significant differences were found, physicians in the
Low exposure community were somewhat more
likely to call croup bronchitis. Thus, any small degree
bias traced to variations in diagnostic customs would
be most likely to involve the Low exposure com-
munity but not the other three communities.
The questionnaire used in this study classified deep
chest colds as bronchitis, which is lower respiratory
illness, and this made validation of questionnaire
responses a difficult task since physicians will often
note only acute respiratory infection or upper respira-
tory illness when, in fact, they diagnose a chest cold.
At times, physicians only record symptoms that are
of value in helping recall the patient's medical
history. Other possible errors include imperfect
memory retention, imprecise doctor-patient com-
munication, and imprecise enumeration. For the
validation estimate, either the specific diagnosis of
bronchitis or any acute respiratory illness confirmed
by medical record was accepted as validation of lower
respiratory illness reported by the questionnaire
respondents. On the other hand, a questionnaire
reporting no illness was not considered in error unless
a specific lower respiratory illness syndrome was
diagnosed. For each area, a 15 percent sample of
questionnaire responses was validated by the
described review of medical records. For each physi-
cian included in the validation sample, approximately
equal numbers of children reported well and sick with
bronchitis were selected. (For this reason, the use of
sensitivity and specificity measures to arrive at
corrected rates is not valid.)11 Within the sample,
attempts were made to validate all lower respiratory
illness episodes. The validation effort revealed no
Table 2.3.18. CLINICAL DIAGNOSIS QF_ACUTEBRONCHITIS: NUMBER OF
PHYSICIANS CHARACTERIZING RESPIRATORY ILLNESS SYNDROMES
AS "BRONCHITIS" OR "NOT BRONCHITIS"
Low
Syndrome I
description Bronchitis
Pharyngitis
0
Deep chest
cold or
mi Id bron-
chitis
13
Croup or
laryngo-
tracheo-
bronchitis 1 4
Classical
bronchitis
Severe bron-
chitis or
early
pneumonia
Pneumonia
13
Not
bronchi ti s
17
4
13
• 4
8
3
9
14
Pool of physicians
from Int. I, Int. II,
and High
exposure areas
1 Not
Bronchi ti s bronchi ti s
4
18
8
28
29
5
32
18
28
8
17
31
Intermediate I
Bronchitis
2
Not
bronchitis
13
7
8
3
11
8
2
12
4
7
13
Intermediate II
| Not
Bronchitis
1
8
2
bronchitis
11
4
10
10 ! 2
7
1
5
11
High
Not
Bronchitis bronchitis
1 8
3 6
3 6
2
4 , 5
2 7
.
2-68
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.3.19. VALIDATION OF QUESTIONNAIRE RESPONSES REPORTING CHILDREN
EITHER SICK OR WELL AGAINST PHYSICIAN RECORDSa
Any lower respiratory il
Sick i Well
Iness] Bronchitis
sick '
I validated validated validated
Community
percent
1
Low
Intermediate I
Intermediate II
High
Total
n percent n
70 91 92
69 108 94
78 132: 79
78
72
103
434
84
87
85
percent
69
87 69
90
82
61 73
323 74
n
Well
validated
percent
77 96
85 96
109 82
86 86
357
90
n
100
114
114
79
407
Allowances were made to adjust for multiplicity of diagnostic terms used
to describe mixed upper and lower respiratory infections. Otherwise, the
agreements would be substantially reduced.
Dn is the number of records examined.
significant intercommunity differences (Table
2.3.19). Since the High exposure community ex-
•hibited high to intermediate values for any of the
validating procedures, there was no evidence of bias
that could explain intercommunity differences at-
tributed to pollution. Given the diagnostic impreci-
sion and other inherent errors involved, the question-
naire proved to be an adequate, though not perfect,
reporting instrument.
DISCUSSION
Children living in communities characterized by
elevated ambient levels of sulfur dioxide and/or
suspended particulate sulfate air pollution were
reported to have significantly higher attack rates for
single or repeated episodes of illnesses grouped as
acute lower respiratory disease. Consistent excesses
were observed for any acute lower respiratory illness
and for croup and bronchitis, two of the most
common clinical syndromes within the grouping. No
intercommunity differences in episodes of pneumonia
or hospitalizations were reported. Morbidity excesses
were observed after only 1 year of exposure to
elevated levels of both sulfur dioxide and suspended
particulate sulfates. More subtle morbidity excesses
attributed to suspended sulfate exposures were first
apparent after exposures lasting 5 to 8 years and
unmistakable after 9 to 12 years.
Asthmatic history, age, socioeconomic status,
cigarette smoking in the home, and recent family
migration were intervening variables that had a
significant effect on the frequency of acute lower
respiratory illness. In the present report, the two
most powerful determinants of illness, history of
asthma and age, were either isolated or adjusted.
Weaker covariates, including duration of residence
and socioeconomic status, were duly considered in
the analysis, as was the effect of parental cigarette
smoking. Two other factors that might seriously bias
the study were intercommunity differences in the
customary diagnostic patterns of physicians serving
the study areas and overreporting by respondents in
the more polluted communities. Special surveys
assured that neither of these had occurred.
Other community characteristics were also closely
scrutinized, but none would seem likely to invalidate
the study. Smoking patterns, family size, and educa-
tional attainment of the head of household were
similar for all study communities. Exclusions affected
each of the communities equally, and the illness
experience of asthmatics who were excluded from the
main analysis showed the same morbidity patterns as
the nonasthmatics who comprised the principal study
population. Samples of nonrespondents from each
community were interviewed by telephone and found
to exhibit demographic patterns almost identical to
those of respondents. Still, there is a remote pos-
sibility that nonrespondent illness patterns could
differ enough from those of respondents to alter the
findings of the study. However, nonrespondents in
different communities would have to behave
anomalously. Doubling the reporting rate of non-
respondents in Low and Intermediate exposure com-
munities could equalize overall acute lower
respiratory illness reporting if nonrespondents in the
Salt Lake Basin Studies
2-69
-------
High exposure area reported like initial respondents.
In the High exposure area, nonrespondents might
have the same impact only by reducing their illness
reporting to a level one-fifth that of respondents.
Such wide variations are unlikely in view of the
consistent paucity of other major intercommunity
differences involving illness covariates.
One might also choose to question the exposure
estimates projected from emissions profiles and
dispersion models after calibration with aerometric
measurements.8 Exposure projections, while not
precise enough to quantitate short-term peak levels,
are most useful tools for the assessment of annual
average exposures.
After carefully considering each of the com-
plexities involved, we concluded that significant
increases in acute lower respiratory morbidity can be
attributed to exposures of 1 or more years charac-
terized by elevated annual average levels of sulfur
dioxide (91 Mg/m3) accompanied by elevated levels of
suspended sulfates (15 Mg/m3) in the presence of low
levels of total suspended particulates (62 /zg/m3), and
suspended nitrates (2.0 jug/m3). We also found that
morbidity excesses accompanied longer (5 to 9 year)
exposures to elevated annual average levels of sus-
pended sulfates (9 /xg/m3) in the absence of eleva-
tions of other ambient air pollutants.
The magnitude of the observed morbidity excesses
attributed to combined sulfur dioxide and suspended
sulfate exposure is worthy of comment (Table
2.3.20). Among nonasthmatic children living as long
as 3 years in the High exposure area, overall increases
were 42 percent in the proportion of children
reporting a single illness and 57 percent in the
proportion of children reporting repeated episodes.
Morbidity excesses in repeated respiratory illness that
were attributable to suspended sulfate pollution of
the Intermediate II community were a more modest
15 percent in the absence of high levels of sulfur
dioxide. Among asthmatics, the increases in total
acute lower respiratory illness (25 percent) and in
croup (13 percent) were substantial, but very small
increases occurred with every clinical syndrome. In
the High exposure community, puzzling, un-
explained, statistically insignificant deficits were
observed in pneumonia attack rates and hospitaliza-
tions.
Reductions in ambient sulfur dioxide concentra-
tions to levels specified by Primary Ambient Air
Table 2.3.20. EXCESS ACUTE LOWER RESPIRATORY ILLNESS AMONG CHILDREN EXPOSED
3 OR MORE YEARS TO ELEVATED LEVELS OF SULFUR DIOXIDE
Illness
Any lower
respiratory
illness
Croup
Bronchitis
Pneumonia
Hospital izations
Number
of episodes
reported
>1
>2
>1
>2
>1
>2
>1
>2
>1
>2
Percent excess3
In asthmatic
children living
in High exposure
community
25
_b
13
_b
5
_b
3
_b
2
_b
Nonasthmatic children living in
Intermediate II
community
8
15
6
15
3
20
_c
_c
_c
1
High exposure
community
42
57
69
103
42
66
_c
c
_c
60
Combined experience of Low and Intermediate I communities used as a reference point. Excess risk plus 1.00 equals relative
risk.
Not available.
°Deficits noted for these categories are based upon relatively rare events.
2-70
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Quality Standards accompanied by a reduction in
suspended sulfates could well significantly reduce
acute lower respiratory illness in large numbers of
children currently or recently exposed to high pollu-
tion levels. However, little is now known about the
persistence of induced susceptibility to acute respira-
tory infection. Repeated surveys, controlled clinical
studies, and well designed toxicological studies will be
necessary to document and understand the expected
improvements in human health.
SUMMARY
In a survey involving 9000 preschool and elemen-
tary school children, the frequency of single or
repeated episodes of acute lower respiratory illness
significantly increased (40 to 50 percent) after
exposures to elevated levels of sulfur dioxide (91
Aig/m3) and suspended particulate sulfates (15 Mg/m3)
lasting over 2 years. Longer exposures, roughly 9
years, to lower annual average levels of suspended
sulfates (9 jug/m3) resulted in smaller increments (22
percent) in acute lower respiratory morbidity. As
expected, a personal history of asthma and age were
the most powerful determinants of acute lower
respiratory ilhiess. When parents polluted the home
environment by smoking cigarettes, their preschool
children more frequently became ill with bronchitis
or pneumonia. Socioeconomic factors behaved in a
complex fashion. Families with higher social status
reported increased croup, increased bronchitis, and
decreased pneumonia. Physician diagnostic customs
were examined and found similar in all communities.
A 15 percent sample of questionnaire responses was
validated from physician records, and no intercom-
munity differences in confirmation of questionnaire
responses were noted.
REFERENCES FOR SECTION 2.3
1. Douglas, J. W. B. and R. E. Waller. Air Pollution
and Respiratory Infection in Children. Brit. J.
Prevent. Soc. Med. 20:1-8, 1966.
2. Lunn, J. E., J. Knowelden, and A. J. Handyside.
Patterns of Respiratory Illness in Sheffield
Infant School Children. Brit. J. Prevent. Soc.
Med. 27:7-16, 1967.
3. Holland, W. W., T. Haul, A. E. Bennett, and A.
Elliott. Factors Influencing the Onset of Chronic
Respiratory Disease. Brit. Med. J. 2:205-208,
April 1969.
4. Toyama, T. Air Pollution and Its Health Effects
in Japan. Arch. Environ. Health. 5:153-173,
1964.
5. Manzhenko, E. G. The Effect of Atmospheric
Pollution on the Health of Children. Hygiene and
Sanitation (Moscow). 57:126-128, 1966.
6. Pearlman, M. E., J. F. Finklea, J. P. Creason,
C. M. Shy, M. M. Young, and R. J. M. Horton.
Nitrogen Dioxide and Lower Respiratory Illness.
Pediatrics. 47:391-395, 1971.
7. Questionnaires Used in the CHESS Studies. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
8. Hertz, M. B., L. A. Truppi, T. D. English, G. W.
Sovocool, R. M. Burton, L. T. Heiderscheit, and
D. 0. Hinton. Human Exposure to Air Pollutants
in Salt Lake Basin Communities, 1940-1971. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N. C.
Publication No. EPA-650/1-74-004. 1974.
9. Grizzle, J. E., C. F. Starmer, and G. G. Koch.
Analysis of Categorical Data by Linear Models.
Biometrics. 25(3):489-504, September 1969.
10. House, D.E., J. F. Finklea, C.M. Shy, D. C.
Calafiore, W. B. Riggan, J. W. Southwick, and
L. J. Olsen. Prevalence of Chronic Respiratory
Disease Symptoms in Adults: 1970 Survey of
Salt Lake Basin Communities. In: Health Conse-
quences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N. C.
Publication No. EPA-650/1-74-004. 1974.
11. Vecchio, T. Predictive Value of a Single
Diagnostic Test in Unselected Populations. New
Eng. J. Med. 274:1171-1173, May 26, 1966.
Salt Lake Basin Studies
2-71
-------
APPENDIX
Table 2.3.A.1. UNADJUSTED 3-YEAR ATTACK RATES FOR ANY LOWER RESPIRATORY
ILLNESS BY RESIDENCE DURATION AND NUMBER OF EPISODES
Community
Low
Intermediate I
Intermediate II
High
Low
Intermediate I
Intermediate II
High
Number of
illness episodes
> 1
> 2
Attack rate, percent
3yr
residence
24.3
24.8
26.7
33.8
13.4
13.2
15.3
21.1
Any length
residence
24.2
25.4
26.4
32.4
13.9
13.6
15.2
20.0
Table 2.3.A.2. UNADJUSTED 3-YEAR ATTACK RATES FOR CROUP
BY RESIDENCE DURATION AND NUMBER OF EPISODES
Community
Low
Intermediate I
Intermediate II
Number of
illness episodes
>1
High j
Low
Intermediate I
Intermediate II
High
>2
Attack rate, percent
< 1 yr
residence
19.0
15.7
12.6
12.6
11.1
5.1
4.7
4.5
2yr
residence
8.6
9.1
17.8
9.6
8.6
6.1
6.7
6.4
>3yr
residence
15.1
13.6
15.1
23.3
7.2
5.2
6.9
12.7
Any length
residence
15.1
13.6
14.9
21.6
7.6
5.2
6.7
11.7
Table 2.3.A.3. UNADJUSTED 3-YEAR ATTACK RATES FOR BRONCHITIS
BY RESIDENCE DURATION AND NUMBER OF EPISODES
Community
Low
Intermediate I
Intermediate II
High
Low
Intermediate 1
Intermediate II
High
Number of
illness episodes
>1
>2
Attack rate, percent
<1 yr
residence
20.9
24.6
17.0
18.5
13.1
9.3
9.9
6.3
2yr
residence
4.9
11.1
16.7
10.6
3.7
5.1
5.6
3.2
>3yr
residence
14.5
15.2
16.2
21.5
6.2
5.8
7.4
10.2
Any length
residence
14.6
16.0
16.3
20.7
6.6
6.2
7.5
9.5
2-72
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.3.A.4. UNADJUSTED 3-YEAR ATTACK RATES FOR PNEUMONIA
BY RESIDENCE DURATION AND NUMBER OF EPISODES
Community
Low
Intermediate 1
Intermediate II
High
Low
Intermediate 1
Intermediate II
High
Number of
illness episodes
>1
>2
Attack rate, percent
<1 yr
residence
6.6
8.9
7.9
7.7
3.9
3.8
2.8
0.9
2yr
residence
1.2
5.1
4.4
1.1
0.0
1.0
1.1
0.0
>3yr
residence
4.2
5.9
4.5
3.8
1.3
1.8
1.7
1.1
Any length
residence
4.2
6.2
4.9
4.1
1.5
2.0
1.8
1.0
Table 2.3.A.5. UNADJUSTED 3-YEAR HOSPITALIZATION RATES
BY RESIDENCE DURATION AND NUMBER OF ADMISSIONS
Community
Low
Intermediate 1
Intermediate II
High
Low
Intermediate 1
Intermediate II
High
Number of
admissions
>1
>2
Attack rate, percent
<1 yr
residence
3.3
4.7
6.3
3.6
1.3
0.4
2.4
0.5
2yr
residence
0.0
4.0
4.4
4.3
0.0
1.0
1.1
1.1
>3yr
residence
2.2
2.5
1.7
2.2
0.2
0.5
0.2
0.6
Any length
residence
2.2
2.8
2.2
2.4
0.3
0.5
0.5
0.6
Salt Lake Basin Studies
2-73
-------
2.4 AGGRAVATION OF ASTHMA BY AIR POLLUTANTS:
1971 SALT LAKE BASIN STUDIES
John F. Finklea, M.D,, Dr. P.H., Dorothy C. Calafiore, Dr. P.H.,
Cornelius J. Nelson, M.S., Wilson B. Riggan, Ph. D.,
and Carl G. Hayes, Ph. D.
2-75
-------
INTRODUCTION
The purpose of this study was to determine what
levels of ambient air pollutants are linked with excess
illness in asthmatics who comprise 3 to 5 percent of
our total population.1-2 Previous studies have
indicated that complex urban air pollutant mixtures
aggravate asthma as determined from diaries of
patients, emergency room visits, or physician and
outpatient records.3'8 Air pollutants generated by
agricultural enterprises and specific types of industrial
activity were also found to aggravate asthma.9"12
Other studies revealed that the effects of meteoro-
logical variables, particularly temperature, must be
considered in attempts to quantify the role of air
pollutants in the aggravation of asthma.13"18 Un-
fortunately, there remains a relative deficiency in the
dose-response information on the adverse effects of
air pollutants upon human health, and such informa-
tion must provide the basis for National Primary Air
Quality Standards. Some of the complexities involved
were illustrated by a study that linked air pollutants
from a coal-fueled power plant to asthmatic attacks
in a dose-response fashion but was unable to dis-
entangle the effects of single pollutants.19 There
remains, then, a real need to define the dose-response
functions relating air pollutants to an excessive
number of asthmatic attacks.
Western smelter communities are generally located
in a fashion that places residential neighborhoods
away from the usual path of stack emissions contain-
ing high concentrations of sulfur dioxide. On oc-
casion, however, meteorological shifts expose com-
munities to short-term peak pollutant elevations
involving stack emissions or particulates that have
been resuspended from waste disposal sites. In ad-
dition, lower but significant elevations in annual
average pollution levels are observed in smelter
communities in the absence of high concentrations of
suspended particulates, suspended nitrates, nitrogen
dioxide, gaseous hydrocarbons, and carbon
mon ixide. We expected, therefore, that western
smelter communities might provide an opportunity to
quantify adverse health effects attributable to com-
munity air pollution exposures characterized by
elevated levels of sulfur dioxide and suspended
sulfates.
The diary study reported here comprises the first
26 weeks of surveillance of asthmatics living in Utah.
Surveillance began in March 1971 and continued for a
total of 67 weeks. The study sought to test the
following two hypotheses: first, that excessive asthma
attacks among panels of known asthmatic cases
would be related to ambient air pollution exposures
in a dose-response fashion after removal of the effects
of variations in ambient temperature; and second,
that any effects attributable to sulfur dioxide and
suspended sulfates could be disentangled from the
effects of suspended particulates and suspended
nitrates.
MATERIALS AND METHODS
Selection of the Study Population
Four Salt Lake Basin communities where previous-
ly instigated CHESS studies were in progress were
used for the asthma study. Located immediately west
of the Rocky Mountain Wasatch Range, they are:
sections of Ogden and Salt Lake City, Kearns, and
Magna. For the other studies, they were selected to
represent a pollutant gradient and are thus referred to
as Low exposure, Intermediate I, Intermediate II, and
High exposure, respectively. The principal point
source of sulfur dioxide emissions, a copper smelter,
was located approximately 38 miles southwest of
Ogden, 13 miles west of Salt Lake City, 8 miles
northwest of Kearns, and 5 miles north of Magna.
Thus, there would be assurance of some level of
sulfur dioxide exposure for all four panels. Unlike
other CHESS studies, which compare intercom-
munity differences, the asthma study will focus on
the experience of each community's panel over time.
Asthmatics living in these communities were
identified through a School and Family Health
Questionnaire administered in December 1970. Ad-
ditional asthmatics were found by searching the
records of community agencies and through the help
of local practicing physicians. Each asthmatic was
personally interviewed. The interviewer explained the
study, elicited consent to participate, administered an
asthma panel questionnaire, and taught the asthmatic
the diary method for recording daily symptoms.
(Examples of the questionnaires and diary form used
are given elsewhere.)20 Asthmatics, diagnosed by a
physician, were asked to become panelists after they
were interviewed if they gave a history of wheezing
and shortness of breath with each asthma attack and
if they had two or more attacks during the previous
year. Those who agreed to participate lived within a
2-mile radius of an air monitoring station. Highest
priority for selection was given to nonsmokers over
16 years of age. The study began with 46 panelists
each in the Low exposure and Intermediate I com-
munities, 55 in the Intermediate II community, and
48 in the High exposure community. Replacements
2-76
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
for panelists dropped from study were added during
the 26 weeks of diary reporting, which began on
March 7, 1971, and ended on September 4, 1971.
Diary Coverage
Each panelist received a diary weekly by mail. He
recorded attacks as they occurred each day and
returned the completed diary by mail at the end of
each week. When instructions for filling out the diary
were given by the summer interviewers, each panelist
was asked to indicate on the diary that he had an
attack whenever he became aware of shortness of
breath combined with wheezing. Nonresponse and
diaries requiring clarification were investigated by
telephone. Diaries received more than 12 days after
the last day covered by the diary and diaries that
could not be clarified were not accepted for data
processing. A log book of diary entries was main-
tained to permit constant evaluation of panelists'
performance. The log book was reviewed weekly to
identify over- and underreporting of asthma attacks,
which might indicate misunderstanding of what was
to be reported, to look for repeated clarification of
diaries usually attributable to a misunderstanding of
the diary method, and to evaluate frequency of
failure to return diaries. Appropriate contact was
made with the panelists to improve performance, and
continued nonresponders were dropped from the
study.
Monitoring Air Pollution Exposures
Air monitoring stations were installed in each
community at a second-story building level approxi-
mately 25 feet above the ground. Measurement
devices at each station included high-volume samplers
for measuring total suspended particulates, dustfall
buckets, and sulfur dioxide and nitrogen dioxide
bubblers. Continuous 24-hour monitoring was main-
tained throughout the study period. Sampler filters
and bubblers were replaced daily and sent for
laboratory analysis. Dustfall buckets were replaced
monthly. Minimum and maximum temperatures were
recorded at the International Airport that serves the
four communities. In the laboratory, sulfur dioxide
and nitrogen dioxide gas bubblers were analyzed by
methods that are standard for the Environmental
Protection Agency, that is, the West-Gaeke and
Jacobs-Hochheiser methods, respectively. After de-
hydration, contents of the dustfall buckets were
weighed. High-volume sampler filters were measured
gravimetrically and analyzed for sulfate and nitrate
fractions. The latter were determined by an
automated analysis reduction-diazo coupling method.
For the sulfate fraction, a turbidimetric method was
used, and turbidity was measured by spectro-
photometer. More detailed descriptions of exposure
monitoring are piesented elsewhere.21
Statistical Analysis
Daily asthma attack rates were computed by
dividing the number of panelists who reported one or
more asthma attacks on a given day by the number
who returned their diaries for that day. This rate
computation method was used because some panelists
were dropped from the study during the 26 weeks
and others were added. Also, some panelists were
occasionally negligent in returning their diaries or did
not complete diaries when they were temporarily out
of the area. Four Intermediate I, four Intermediate II,
and three High exposure panelists were not included
in the analysis because they failed to return any of
the diaries mailed to them.
Data for a single day of the study, then, consisted
of the following items: a daily asthma attack rate, a
24-hour average level for each pollutant measured,
and a minimum and maximum temperature. These
data items for each community were available for
each day of the study with the exception of some
missing pollutant data. In all analyses, the asthma
attack rate was considered the dependent variable.
The independent variables were temperature and
24-hour average pollutant levels.
The statistical analysis involved a sequence of five
steps: first, plotting weekly averages of asthma attack
rates, pollutant levels, and minimal temperature;
second, calculating a simple correlation matrix for
daily asthma attack rates, pollutant levels, and
minimum temperature; third, performing multiple
regression analyses to look for pollutant effects after
removal of temperature effects; fourth, constructing
temperature-specific relative risk and excess risk
models for various pollutant levels; and fifth,
computing temperature-specific dose estimates to
determine, if possible, threshold levels for pollutant
effects.22-24
To calculate relative risk functions, the total ex-
perience of all four panels was combined into a
series of temperature-specific relative risk models for
appropriate concentrations of each of the pollutants
of interest. Since the four panels differed to some
extent from each other in the intrinsic frequency of
asthmatic attacks, a separate temperature-specific
Salt Lake Basin Studies
2-77
-------
relative risk for each was computed based upon the
experience of the panel during the days falling in the
lowest concentration range for the particular pol-
lutant of interest. Then, for each panel, relative risks
for each subsequent pollutant level in the tempera-
ture-specific series were computed using the rate for
each successively higher pollutant level as the numera-
tor and the base rate from the lowest pollutant level
as the denominator. Relative risks were then weighted
by the number of person-days of risk that each
represented and combined across communities to
achieve a single pooled rate for each temperature-
specific pollutant level. The person-days of risk
associated with any pooled rate were tallied so that
an idea of rate stability might be conveyed. Because
the analyses included days on which one or another
pollutant might be missing in one or another com-
munity, temperature-specific groups of person-days
will not balance.
Percentage of excess risk was computed from a
base rate. For each of the pollutants of interest, the
base rate consisted of the asthma attack rate for
warm (Tmm > 50 °F), low-pollutant days. Percentage
of excess risk was calculated by dividing the ap-
propriate pooled asthma attack rate by the base rate
and then subtracting one.
Threshold functions were computed by the
method of Hasselblad et al.23 In setting air quality
standards, the existence of a threshold concentration
for a health effect is implicitly assumed. It is
important, therefore, to ascertain if the dose-response
relationship can be reasonably estimated by a
function indicating no response until some nonzero
threshold concentration is exceeded. Most statistical
procedures assume a strictly monotonic functional
relationship between two variables, and therefore a
significant association exists at all pollutant levels, no
matter how low.
As a simple alternative, we hypothesized, for a
threshold function, a segmented line having zero
slope below exposure level x and with positive slope
at levels above x. The point x is estimated by the
least squares method. This function has been
designated a "hockey stick" function.23 This least
squares technique had been earlier applied to other
more general problems.24
The estimated threshold, estimated slope,
estimated intercept, and the assumption of a linear
response above the threshold permitted calculation of
an excess risk at the current short-term air quality
standard. Multiplying the difference between the
standard and the estimated threshold by the
estimated slope of the response function provided an
estimated asthma attack rate at these allowable
once-per-year levels. Percentage of excess risk was
calculated by dividing this estimated attack rate by
the estimated attack rate intercept and then subtract-
ing one.
RESULTS
Environmental Exposures
Pollutant exposure data for 1971 are given in
Table 2.4.1. During the spring and summer months
covered by the study, aerometric stations in the High
and Intermediate II communities reported short-term
sulfur dioxide and suspended sulfate exposures that
were much higher than the short-term exposures in
the Low and Intermediate I communities. This same
pattern was not seen for total suspended particulates
and suspended nitrates. During the summer quarter,
the effect of a strike on sulfur dioxide emissions was
reflected by the relatively low median, 90th per-
centile, and maximum exposures in the High and
Intermediate II exposure communities. The experi-
ence of panelists during other seasons when peak
pollutant exposures are much higher are not covered
in this publication, but will be covered in later
reports.
One other observation should be emphasized: no
single 24-hour exposure during this spring and
summer study period exceeded the maximum daily
total suspended particulate level (260 Mg/m3) al-
lowable once yearly under the National Primary
Ambient Air Quality Standard, and only 2 days
exceeded the similar once-per-year allowable sulfur
dioxide standard (365 Atg/m3). Therefore, adverse
effects in this study attributable to either sulfur
dioxide or total suspended particulates were within
currently allowable air quality limits for a single
24-hour period. Furthermore, none of the study
communities exceeded the current National Primary
Annual Average Air Quality Standard for total
suspended particulates or sulfur dioxide during the
study.
Characteristics of the Study Population
Although the study design did not call for inter-
community comparisons, every attempt was made to
2-78
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
LU
CO
i
1
Q
(A
ID
O
I
<
D
ce
CO
oc
LU
i
LU
ce
§
CL
X
<
Q
CM
E
Oi
3.
1/1
C
•£>
10
c
QJ
O
C
0
u
4_)
c
(0
1 \
3
^~
r—
O
CL
f^
Ol
£
X
to
2ET
!
OJ
•r—
^
OJ
S-
O)
0.
.c
4J
O
Ol
ro
T3
£
t-
QJ
^ j
£
•r-
-
ro
U.
i.
0)
pr
£5
3
00
01
c
•r-
J^.
a.
00
S-
QJ
•M
c
3
_
ro
ii
S^.
QJ
|5
3
00
O!
E
i_
CL
00
S_
ttJ
C
•r—
5
QJ
g
3
00
CD
C
ii
Q.
00
•4—) ^»j
C -4->
ro •!-
•!-> -0 C
3 C 3
i— £
0 0
CL 0
^•iniocM r~-mcotD CMCMPOIO • — r— i—
i — CO CM CM i — «*•
t— 01 co 01 COCMOCM <• in vo <• mi — r-o
ininmin r--oocMoo romioio CMCO«*CM
CM OO CO CM CO CM «f
co 01 m r— r*— r— m co co 01 co CM tn to ^- «5j-
i— in Ol Ol VOCOCMOO r— i— CM CO
^— r— ^~ ^~
oil — r-*i — cocor — ^ oiinin'oo ^iDLn^
inr-.r--.oo CMoiin^l- i— > — CM
CM LO CM i— f— CM
oo ^~ CM ^- r*— ^— oo vo i — ^t* r-. r— in r-. ^° ^*
r-coinin r--cor--oo r— i— ^- co
CMOIVOOI OOCMCOsi- CMCMOOO OCMOlO
CM"Sj-r--CO t>.| — OOCO i — CMCMCO i — i— i —
01 *iO CM co Oi r*— • 01 o^ r-— oo *~~* oo ^^ ^J* oo co
•— CM «*• oicomoo i— i—
^— co ^^ in 01 oi r— . oo i—
^~
tnoovo o^ovoio ^~io vo i — CM i — i — i —
r— ^™ *^ r— . CO CO IO f—
•— •— i—r-I VOCOCOLO i— C>J
*3~r-.^i-oi i — vo^j-i — in in in i*-. COCOCMCM
us vo «a- in
U3coo«* oooor~.r-s «d-minco CM CM i— i—
i — i — — < >— " at i— i QJ i— i
T3 ro ro
Q) QJQJ O> 0)11 <4- QJO) S_ Q,Q)
^^ 4^ 4-^ ~^3 4-^ <4-^ ^— [ * [ * i % i % » <
•r— roro CQJ rOrO 3 roro -. ro ro
X •!-•!- QJ 4-> •!-•!- 10 •!- -r- C T- •!-
OT3TD Q.rOT3-a -DT3 T3T3
•i— Q)Q) (/)i — QJCO ~O QJQJ 13 OJOJ
"O EE 33 EE Q) EE QJ EE
l-i-
M-OCE-i- rOS-OCG-r- Q. O C C •!— Q. O C C -r-
i — — f i— i >— i a: -Mro — it— ii— ire 10 — i i— i i— i a: 10 i i— i i— i a:
3 00. 3 3
oo H- oo oo
Salt Lake Basin Studies
2-79
-------
select panels as alike as possible. The four panels were
similar in size, age, sex, educational attainment,
smoking habits, and frequency of asthmatic attacks
during the previous year (Table 2.4.2). The Inter-
mediate I community was somewhat different in that
fewer panelists were current cigarette smokers. About
half of the panelists in each community reported a
history of one or more nonasthmatic allergies,
including atopic dermatitis, allergic rhinitis, and
urticaria. There were, however, no intercommunity
differences in the history of such allergic conditions
or in the duration of asthma itself.
A difference was noted in turnover among the
panels during the study period. In the Intermediate II
community, one-third of the original panel withdrew
from study by the 26th week, while attrition was
19.5, 23.9, and 16.6 percent for Low, Intermediate I,
and High exposure panels, respectively. None of the
Low exposure panelists lost were replaced; however,
two-thirds of panelists lost from the other panels
were replaced with new panelists. These differences
are important because new panelists are usually
characterized by a tendency to overreport during
their first month of participation. Three clusters of
newly recruited panelists were identified: four
panelists were added during week 15 to the Inter-
mediate I panel, seven during week 15 to Interme-
diate II panel, and five during week 16 to the High
exposure panel. Examination of diary recordings of
panelists added to Intermediate I and II panels
showed they had reported very few asthma attacks.
The five new High exposure panelists, however,
reported a significant number of attacks.
The overall response rate for the four panels was
good. Five-sixths of the almost 5000 diaries mailed
were useable in the analysis. Weekly response rates
varied somewhat for any particular week in all four
communities.
Temporal Patterns
An overall impression of the relationship between
the frequency of a disorder and factors that might
influence that disorder may be gained by scrutiny of
temporal patterns. Though we expected that short-
term 24-hour pollutant exposures, not weekly
exposures, would be most likely to aggravate asthma,
weekly averages of attack rates, minimum daily
temperature, and selected pollutants were never-
theless computed and plotted to gain a knowledge of
seasonal trends. Experiences of all four communities
were evaluated, but only plots for the Low and High
exposure communities are presented for illustrative
purposes.
When average weekly asthma attack rates and
average daily minimum temperature for the same
week were plotted together, asthma rates in both the
High and Low exposure communities were seen to
decrease as temperature increased. Weekly asthma
attack rates were, however, more erratic than weekly
temperature averages (Figure 2.4.1). Asthma rates
were lowest in both communities during late summer,
weeks 18 to 22 of the study. Attack rates for the
Intermediate II and High exposure communities
Table 2.4.2. CHARACTERISTICS OF ASTHMA DIARY STUDY POPULATION
Community
Low
Intermediate 1
Intermediate II
High
Total
Population characteristics
Total
number of
subjects
(N)
(46)
(49)
(66)
(50)
(211)
Subjects
>16 years
of age,
percent
61
49
42
56
51
Female
subjects,
percent
52
51
46
46
48
>high school
completed by
household head,
percent
80
67
85
74
77
Adults
currently
smoking,
percent
25
17
39
36
30
^10 asthma
attacks during
past year,
percent
37
45
41
38
40
2-80
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
i i i i i i i i i i i i i i i i
LOW EXPOSURE COMMUNITY
| I I I I I I I I I I I i I
HIGH EXPOSURE COMMUNITY
Figure 2.4.1. Weekly mean asthma attack
rates compared with weekly average minimum
temperatures.
LOW EXPOSURE COMMUNITY -
I II I M II II I II II I-
HIGH EXPOSURE COMMUNITY
Figure 2.4.2. Weekly mean asthma attack
rates compared with weekly mean sulfer di-
oxide levels.
dropped to their summer lows 2 weeks earlier than
the other two communities.
Weekly average sulfur dioxide, total suspended
particulate, and suspended sulfate air pollutants were
each in turn plotted in concert with weekly average
asthma attack rates. Sulfur dioxide levels in the Low
exposure community were quite low, except for one
peculiar week during summer, and generally had no
discernible relationship to either asthma attacks or
the previously described temperature trend (Figure
2.4.2). The High exposure community was quite
different. During March and April (weeks 1 to 9), the
coldest weeks of the study, there was an apparent
inverse relationship between sulfur dioxide levels and
asthma attack rates, similar to the pattern seen in
Figure 2.4.1 for temperature and asthma attack rates.
Thus, the effect of temperature could be masking the
effect of sulfur dioxide during these cold months.
Levels of sulfur dioxide in the High exposure com-
munity reached zero levels during weeks 18 to 20 due
to a smelter strike in July.
i i i M M i i : ii i i i i M
LOW EXPOSURE COMMUNITY^
2 160
8
i- !«
§
I
I I I I I 1 II I I II I I I I I
HIGH EXPOSURE COMMUNITYJ»
24 6 I 10 12 14 16 U 20 22 24 X
TIKE, neks
Total suspended particulates were then considered
in the same fashion (Figure 2.4.3). No clear relation-
ship between particulates and asthma was seen for
Figure 2.4.3. Weekly mean asthma attack
rates compared with weekly mean total sus-
pended particulate levels.
Salt Lake Basin Studies
2-81
-------
either community, with the possible exception of the
last 10 weeks in the High exposure community.
Next, a similar plot for suspended sulfates was
evaluated (Figure 2.4.4). In the High exposure com-
munity, suspended sulfate levels followed the same
general trend noted for asthma attack rates; weekly
average suspended sulfate levels were consistently low
in the Low exposure community, and unrelated to.
fluctuations in asthma.
The temporal pattern of suspended nitrates (not
illustrated) seemed related to the seasonal tempera-
ture trend but not to asthma.
The nitrogen dioxide concentrations were
calculated and were included in the statistical
analyses. These analyses indicated that nitrogen
dioxide levels in the four areas were relatively low.
Their temporal pattern was related to that of sus-
pended nitrates and temperature, but as was the case
for suspended nitrates, was not strongly related to
asthma fluctuations. However, recent reevaluation by
the Environmental Protection Agency indicates that
the measurement method for nitrogen dioxide is
subject to interferences and variable collection ef-
ficiencies. Therefore, the nitrogen dioxide data, while
still summarized elsewhere,21 are not included in this
report.
Based on the temporal patterns analyzed, we
could conclude only that temperature and perhaps
one or more pollutant might be influencing asthma
attack rates and that suspended sulfates might play an
important role in the High exposure community.
Correlation and Stepwise Multiple Regression
Analysis
To gain a better view of the complex tangle of
interrelated factors, a simple correlation matrix was
constructed (Table 2.4.3). Temperature was unques-
tionably the strongest and most consistent apparent
determinant of asthma attack rates, with asthma
frequency decreasing as temperature increased.
M I I 1 I I I I I ! I I 1 M till I I I I 1
-\ LOW EXPOSURE COMMUNITY
3 40
I I I I I I I I I I I I I I I M
HIGH EXPOSURE COMMUNITY
2 I t 8 10 12 14 16 18 20 2! 24 26
TIME, weks
Figure 2.4.4. Weekly mean asthma attack
rates compared with weekly mean sulfate
fraction levels.
Suspended nitrates seemed at first glance to exert
the greatest effects upon asthma. One might falsely
conclude that decreasing nitrate air pollution actually
increased asthma. Since nitrates were present in low
levels and were strongly and positively correlated
with temperature, it seemed probable that the
negative correlations only reflected a temperature
effect.
The negative correlation of sulfur dioxide with
temperature was statistically significant in the High
exposure community. The suspended sulfate correla-
tion with temperature, although also negative, was
not statistically significant. It is probable that any
deleterious effects exerted by these pollutants were
masked by the temperature effect. Furthermore,
sulfur dioxide and suspended sulfates seemed too
closely correlated to allow later separation.
There were no consistent correlations between
total suspended particulates and any of the other
pollutants. In relation to temperature, total sus-
pended particulates were either negatively or not at
all correlated.
Hopefully, one could separate the effects of total
suspended particulates and suspended nitrates from
each other and from those of sulfur dioxide or
suspended sulfates. Stepwise multiple regression tech-
niques were employed to separate definite pollutant
2-82
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.4.3. SIMPLE CORRELATION COEFFICIENTS OF ASTHMA DAILY ATTACK RATES,
SELECTED POLLUTANTS AND MINIMUM TEMPERATURES3
Pollutant and
community
Total suspended
particulate
Low
Intermediate 1
Intermediate II
High
Suspended nitrates
Low
Intermediate 1
Intermediate II
High
Suspended sulfates
Low
Intermediate 1
Intermediate II
High
Sulfur dioxide
Low
Intermediate 1
Intermediate II
High
Minimum daily
temperature
Low
Intermediate 1
Intermediate II
High
Attack
rate
0.036
0.103
0.140
0.181b
-0.1 60C
-0.281 b
-0.268b
-0.279b
-0.1 50C
-0.106
-0.025
0.269b
0.080
0.000
0.002
0.065
-0.378b
-0.427b
-0.44 1b
-0.469b
Total
suspended
particulate
0.274b
0.1 87C
0.115
-0.105
0.221b
0.008
0.056
0.242
0.001
-0.059
0.021
0.253b
0.018
-0.119
0.007
-0.248b
Suspended
nitrates
0.553b
0.374b
0.191°
-0.019
-0.006
-0.144
0.002
-0.228C
0.646b
0.649b
0.625b
0.689b
Suspended
sulfates
0.004
0.2 19C
0.508b
0.265b
0.469b
0.223b
0.086
-0.112
Sulfur
dioxide
0.012
0.1 62C
-0.052
-0.2836
Based on 100 to 182 observations.
Significant at p < 0.001.
Significant at p < 0.05.
effects from those of temperature. Since the effect of
low temperature may itself be enhanced by con-
comitant pollutants, the multiple regression proce-
dure is a conservative estimate of any pollutant
effects.
Multiple regression analysis showed a significant
effect of temperature in every community (Table
2.4.4); however, the linear model probably did not
completely remove all effects of temperature. Total
suspended particulates were noted to have a signifi-
cant adverse effect in the Intermediate II and High
exposure communities, and suspended sulfates had a
strong effect in the High exposure community. No
residual adverse effect was observed for sulfur dioxide
and suspended nitrates. Clearly, further analyses
Salt Lake Basin Studies
2-83
-------
Table 2.4.4. SUMMARY OF MULTIPLE REGRESSION ANALYSES CONSIDERING
THE EFFECT OF MINIMUM TEMPERATURE ON ASTHMA ATTACK RATE,
FOLLOWED BY THE EFFECTS OF AIR POLLUTANT LEVELS
Community
Low
Intermediate 1
Intermediate II
High
Source of variation3
Temperature
alone
<0.001
<0.001
<0.001
<0.001
Temperature plus:
Sulfur
dioxide
<0.10
NS
NS
NS
Total suspended
particulates
NS
NS
<0.05
<0.02
Suspended
sulfates
NS
NS
NS
<0.005
Suspended
nitrates
<0.10
NS
NS
NS
NS—not significant, p > 0.10.
should focus primarily on two pollutants (particulates
and suspended sulfates) and also on sulfur dioxide to
define the relative and excess risks attributable to
each and to establish, if possible, an appropriate
mathematical dose-response model, whether linear,
threshold, or curvilinear.
Temperature-specific Risk Models
By computations described in the "Materials and
Methods" section, relative risks were calculated for
the pollutants of interest (Table 2.4.5). The strongest
effect by far was noted for suspended sulfates at
warmer temperatures (T^ > 50 °F); 24-hour levels
as low as 6.1 Mg/m3 were linked to an increased
relative risk for asthma. Similar findings were
obtained when the temperature divide was set at
TJJJJJJ = 40 °F. The inverse relationship observed for
temperature and asthma attack rates in the High
exposure community could be masking the true
extent of the sulfate effect. One can be confident
that the sulfate effect was not influenced spuriously
by sulfur dioxide; the sulfate effect could be seen in
the Low and Intermediate I communities where the
influence of modest levels of sulfur dioxide could not
be discerned in the asthma attack rates.
The effect of total suspended particulates seemed
independent of temperature. Roughly equal in-
crements in attack rates occurred at the same total
suspended particulate levels in either temperature
range.
The effect of low daily minimum temperature
(Tjuju = 30 to 50 °F) obviated or obscured any
aggravation of asthma caused by 24-hour sulfur
dioxide levels of <365 /zg/m3. Since sulfur dioxide
levels in the High exposure community decreased as
temperature increased and since asthma attack rates
follow the opposite temperature trend, one cannot be
sure that temperature is still not hiding a true sulfur
dioxide effect.
We concluded that suspended particulates and
suspended sulfates at relatively low levels exerted an
independent aggravating effect on the frequency of
asthma attacks. Any effect of low levels of sulfur
dioxide that might have occurred at higher ambient
temperatures could not be separated with certainty
from the effect of suspended sulfates.
Excess risk due to the effect of temperature or air
pollution fluctuations on the frequency of asthma
attack rates were determined from appropriate excess
risk functions as described in the "Materials and
Methods" section (Table 2.4.6). The base rate for
suspended sulfates computed when Tj^ > 50 °F at
a level of < 6 Mg/m3 of the pollutant was 10.8. There
was an increasing excess risk of asthma attacks with
increasing levels of suspended sulfate when tempera-
tures were > 50 °F. At the higher levels of suspended
sulfates (> 10.1 Mg/m3) the excess risk computed was
50 percent. An approximately equal excess risk (53
percent) was found in the 30 to 50 °F temperature
range at low levels of the pollutant (< 6 Mg/m3).
This pattern was not seen for either total sus-
pended particulates or for sulfur dioxide. Instead
2-84
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.4.5. POLLUTANT AGGRAVATION OF ASTHMA AS QUANTIFIED
BY TEMPERATURE-SPECIFIC RELATIVE RISK MODELS
Pollutant
Sulfur
dioxide
Total
suspended
particulates
Suspended
su (fates
Minimum
temperature,
°F
30 to 50
>50
30 to 50
>50
30 to 50
>50
Integrated
24-hour
concentration,
Mg/m3
<60
61 to 80
81 to 365
>365
<60
61 to 80
81 to 365
<60
61 to 75
76 to 244
<60
61 to 75
76 to 244
<6
6.1 to 8
8.1 to 10
10.1 to 28
<6
6.1 to 8
8.1 to 10
10.1 to 32
Relative risk8
1.00(16.9)
0.70
0.93
0.81
1.00(12.4)
1.16
1.12
1.00(16.1)
1.12
1.15
1.00(12.1)
1.08
1.17
1.00(16.6)
0.95
0.85
0.89
1.00(10.8)
1.17
1.35
1.50
Person-
days of
observation
13,871
942
1,576
100
12,847
672
216
11,789
1,940
3,663
9,734
3,930
3,207
13,212
1,946
894
1,346
10,816
3,094
965
1,957
Base rate for temperature-specific relative risk is in parentheses.
Person-days for pollutants may not balance across temperature-days because one or more pollutants may be missing on one
or more days.
both showed highest excess risk on cooler days (30 to
50 °F) at the lower pollutant levels and approximate-
ly one-half to one-third less excess risk than this on
warm days (T rr^n > 50°F) at pollutant levels of > 61
Mg/m3.
Threshold and Linear Dose-response
Functions
Threshold functions were fitted for sulfur dioxide,
total suspended particulates, and suspended sulfates
to determine if there is a safe exposure below which
pollutant-induced aggravation in asthma will not
occur or if a linear dose-response function can be
established so that one can balance socially
acceptable control measures with a socially ac-
ceptable increase in asthma attacks.
Threshold functions for sulfur dioxide, considering
two warm temperature ranges (T j^ > 40 °F and
T/min > 50 °F), are plotted in Figure 2.4.5. In either
case, the best estimate was that an effect began at a
concentration near zero, but there was a reasonable
probability (0.05) the threshold might be as high as
23 to 54 jug/m3 (Figure 2.4.5). The steeper slope
belongs to the set of warmer days.
At the present 24-hour once-a-year allowable
sulfur dioxide level of 365 Mg/m3, the threshold
functions predict that asthma attack rates would
Salt Lake Basin Studies
2-85
-------
Table 2.4.6. RELATIVE EFFECT OF TEMPERATURE AND AIR POLLUTANTS ON ASTHMA
ATTACK RATES AS QUANTIFIED BY EXCESS RISK FUNCTIONS
Pollutant
Sulfur dioxide
Total suspended
particulates
Suspended
sulfates
Minimum
temperature
>50°F
30 to 50 °F
>50°F
>50°F
>50°F
30 to 50 °F
Either range
Either range
>50°F
30 to 50 °F
>50°F
>50°F
>50°F
Pollutant concentration,
M9/m3
<60
<60
61 to 80
81 to 365
<60
<60
61 to 75
76 to 260
< 6
< 6
6.1 to 8.0
8.1 to 10
>10.1
Excess
risk of asthma,3
percent
00(12.4)
36
16
12
00(12.1)
33
10
16
00(10.8)
53
17
35
50
Base rate for excess risk is in parentheses.
100 150 TOO
SULFUR DIOXIDE CONCENTRATION, f
Figure 2.4.5. Effect of minimum daily tem-
perature and sulfur dioxide on daily asthma
attack rates.
increase by 51 percent based on cooler days (Tjujjj >
40 °F) and by 254 percent based on warmer days
(Tmin > 50 °F) (Table 2-4-7)- If t116 allowable
once-per-year level were lowered, for example, to 335
Mg/m3, the concomitant reduction in expected excess
illness would be modest, at best a one-tenth reduction
in the asthma excess. On colder days (Tju^ < 40 °F)
no sulfur dioxide effect can be ascertained.
Threshold functions could also be established for
excess asthma attributable to total suspended particu-
lates for two temperature ranges (Tjjjjjj = 30 to 50 °F
and Tjnju > 50 °F) in the Low and Intermediate I
communities (Figure 2.4.6 and Table 2.4.7). As
expected from the temperature-specific relative risk
analyses, the theoretical threshold for the lower
temperature range (107 Mg/m3) is higher than the
threshold for the higher range (71 /ig/m3). In both
cases, 24-hour suspended particulate levels now al-
lowable once per year (260 Mg/m3) might be
expected to more than double asthma attack rates.
Figure 2.4.6 also shows that the threshold levels
for total suspended particulates containing an
elevated suspended sulfate component were identical
(26 jug/m3) for the two temperature ranges con-
2-86
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 2.4.7. TEMPERATURE-SPECIFIC THRESHOLD ESTIMATES FOR THE EFFECT OF
SELECTED POLLUTANTS ON ASTHMA ATTACK RATES
Pollutant
Sulfur dioxide
Total suspended
particulates
without high
sulfates
Total suspended
particulates
with high
sulfates
Suspended
sulfates
Minimum
daily
temperature,
°F
>40
>50
30 to 50
>50
30 to 50
>50
30 to 50
>50
Intercept
(attack
rate,
percent)
12.0
80.0
14.8
9.8
17.1
11.0
16.0
9.8
Estimated
effects
threshold,
M9/m3
54
23
107
71
26
26
17.4
1.4
Slope
0.000198
0.000595
0.001141
0.000882
0.000440
0.000686
0.005500
0.003274
Estimated percent
excess risk at
presently allowed
once-per-year level
51
254
118
170
60
146
9a
62a
aSince no standard exists for suspended sulfates, the level of 20 jug/m , which is now found on 5 to 10 percent of days in urban
areas, was arbitrarily chosen to illustrate the excess risk.
sidered. For the currently allowable once-a-year
24-hour particulate standard of 260p.g/m3, an excess
in asthma attacks of 60 percent for the lower
temperature range and 146 percent for the higher
range might be expected. Clearly, the chemical
composition of suspended particulates is a critically
important determinant of their biological effects.
Substantial new efforts are required to define and
control particulate pollution.
A last threshold function was calculated for
suspended sulfates (Figure 2.4.7 and Table 2.4.7),
and the threshold level for the higher temperature
range (Tjjjyj > 50 °F) was very low (1.4 jug/m3) while
the level for cooler days (T^ = 30 to 50 °F) was
higher (17.4 Mg/m3). At levels of 20 Mg/m3, which are
found on 5 to 10 percent of the days in our large
cities, asthma rates could be expected to increase by
one-tenth for the lower temperature range and by
three-fifths for the higher range.
DISCUSSION
The data reported here suggest association of
excess asthma attacks with several complex environ-
mental influences. Because of their preliminary
nature, they must be interpreted with caution. These
first 26 weeks of study covered the period March 7
through September 4,1971, the greater part of which
is the summer season and the known period of lowest
pollution in the Salt Lake Basin. Thus, during this
period, effects of pollution would be less amenable to
identification and measurement. In addition,
problems of interferring factors, such as airborne
allergens, were probably present during the summer
months of this initial part of the study.
The possible effect of seasonal peaks in plant
allergens presents a problem that is broad in scope
and difficult to define. Various methods have been
used to identify and quantitate pollens, none of
which have been altogether successful. Efforts are
presently in progress to develop a method for
measuring protein content of ambient air. Until
feasible and reliable measurement technology is
available, the problems of assessing the effects of
airborne allergens are prohibitive. In the presently
reported study, the probable net effect of such agents
would have been to make associations between
pollutants and asthma more difficult to establish.
Any allergen peak that happened to coincide with a
high pollutant exposure peak in any of the four
communities would probably also influence asthma
Salt Lake Basin Studies
2-87
-------
mm
£500 F
Tm.n=30°to50°F
7
(107)
100 150 200
PARTICULATE CONCENTRATION, |l
Figure 2.4.6. Effect of total suspended
particulates with and without a high sulfate
content on asthma attack rate.
£' •
attack^rates in the other nearby study communities,
increasing asthma attack rates for all panels and
possibly obscuring true differences attributable to
ambient air pollution.
8 12 IS 20
SUSPENDED SULFATE CONCENTRATION, ft/a?
Figure 2.4.7. Effect of minimum daily tem-
perature and suspended sulfates on daily
asthma attack rates.
We also considered possible sources of bias
inherent in the population studied. Exclusion of some
panelists from the study was of first importance.
Every effort was made to locate all possible
asthmatics living within the monitoring range. Only
those diagnosed as asthmatics by a physician at some
time in their lives were accepted. This exclusion of
medically undiagnosed persons was felt justified
because of the nature of asthma; the frightening
pulmonary syndrome makes it unlikely that persons
subject to acute asthma attacks would fail to seek
medical care. Another small group excluded were
those who would not agree to participate because
they planned to move or claimed no recent asthma
attacks.
Initially, nonsmokers and adults were to receive
selection priority, but the number of asthmatics in
the communities was so limited that none were
excluded for smoking habits or age. A second source
of bias that must be considered, then, is the influence
of smoking cigarettes. An unexpected result of a finer
analysis was an inconsistently high asthma attack rate
found for nonsmokers compared with the rates of
smokers. However, this observation is tenuous as the
smoker rate was based on very small numbers. We
speculate, however, that the severity of asthma in the
2-88
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
nonsmokers was greater, causing them to stop
smoking or to be lifetime nonsmokers. The habit of
smoking could not bias the study since each panel
had a roughly equal proportion of smokers.
Finally, the addition of panelists in clusters during
the study must be considered as a possible source of
bias. Clustering of new panelists occurred in three
communities. In Intermediate I and II this clustering
had no effect on attack rates because of the few
attacks reported by the new panelists. Attacks
reported by a cluster of five panelists added in the
High exposure community during week 16 had the
effect of reducing pollutant impact on attack rates
because they were added at a point in time when the
attack rate of this community was at its lowest level.
An unusual event that occurred during July was a
strike involving the major source of sulfur dioxide
emissions. The shut-down of operations by the strike
was accompanied by a pronounced improvement in
air quality and a reduction in asthma attack rates that
occurred sooner and were larger than seasonal reduc-
tions observed in the more distant study communities
some 2 weeks later. Thus, a reduction in pollution
similar to that which might result from stringent
pollution control action measurably benefited
asthmatics, although it undoubtedly had a major
adverse economic impact.
Analysis of environmental measurements linked
specific air pollutants with excess asthma attacks.
Clearly, low ambient temperature was found to be
the strongest determinant. This finding is consistent
with previous studies conducted by other investiga-
tors.1'19'25 The effect of low temperature was
apparent during the first 8 weeks of surveillance when
weekly means ranged between 24 and 39 °F. One
might expect a further increase in mean attack rates
attributable to the low ambient temperatures of
Utah's winter during the remainder of the 67-week
study.
Temporal patterns and simple correlation analysis
were clouded by the powerful temperature effect but
provided clues pointing to specific pollutants, most
importantly the effects of suspended sulfates. The
intervening strong temperature variable could be
isolated and adjusted for in further analyses. Multiple
regression results showed a temperature effect in
every community and revealed a significantly high
association of suspended sulfates after the tempera-
ture effect was removed. Significant residual effects
were also seen in the two most polluted communities
for total suspended particulates.
The temperature effect was further delineated by
the relative and excess risk models. Suspended
sulfates and sulfur dioxide effects for low levels of
the pollutants were reflected in increased relative
risks for asthma on warm days and an increase in
relative risk on both warm and cold days for exposure
to higher levels of total suspended particulates.
Excess risks could be measured for all three of these
pollutants.
After considering all facets of the study, we
concluded that increased asthma attacks could be
related to 24-hour exposures involving modestly
elevated levels of suspended particulate matter (71
Mg/m3) during warm weather (Tmin > 50 °F) and
higher elevations (107 Mg/m3) in colder weather
(Tmin = 30 to 50 °F). Secondly, we felt it prudent to
consider the threshold level for sulfur dioxide on
warm days (T^ > 40 °F) to be 23 to 54 jug/m3.
Finally, we concluded that the threshold for aggrava-
tion of asthma by suspended sulfates was 1.4 ju§/m3
on warm days (Tj^ > 50 °F) and 17.4 /ug/m3 on
cooler days (Tmin = 30 to 50 °F).
These results suggest that achieving the present
primary air quality standards will not necessarily
protect asthmatics from excess illness attributable to
ambient air pollution. Excess asthma attributable to
sulfur dioxide might be expected on 5 to 10 percent
of summer days. Excesses attributable to total sus-
pended particulates could occur on up to 5 percent of
summer days and 30 percent of fall and winter days.
Most importantly, excesses due to suspended sulfates
are likely to occur on 10 percent of fall and winter
days and 90 percent of summer days. Suspended
sulfates tend to penetrate through suburban residen-
tial rings into more rural areas, thus exposing a much
larger population. Demonstrated pollutant-induced
excesses in asthma are an important public health
problem since asthma ranks as one of the leading
causes of chronic illness. Relatively small increases in
asthma attack rates could easily involve several
million extra days of disability.
Analysis of the complete 67 weeks of this study,
now in progress, will permit comparison of seasonal
variation in factors of biological origin, varying
meteorologic factors, and differing levels and
combinations of the air pollutants. Data from the
1973 study conducted by the same methods and in
these same communities will offer further oppor-
tunity to evaluate the effects of pollutants over a
greatly extended period of time.
Salt Lake Basin Studies
2-89
-------
SUMMARY
Four panels, totaling 211 asthmatics, reported
asthma attacks using weekly diaries for 6 months
during 1971. Excess asthma attack rates could be
attributed to colder days and to ambient air pol-
lutants. Total suspended particulates and suspended
sulfates had significant effects at minimum tempera-
ture ranges greater than 30 °F. The strongest effects
were attributed to suspended sulfates; the most
consistent effects were caused by suspended particu-
lates; and the least impressive effects could be
attributed to sulfur dioxide exposures for the higher
temperature range considered (Tmm > 50 °F). The
threshold estimates for the lower temperature range
(Tmin = 30 to 50°F) were 107 Mg/m3 for total
suspended particulates and 17.4 jug/m3 for suspended
sulfates. For the higher temperature range, the
threshold levels were 23 to 54 jug/m3 for sulfur
dioxide, 71 jug/m3 for total suspended particulates,
and 1.4 Mg/m3 for suspended sulfates. The impor-
tance of chemical composition in determining effects
thresholds was seen in the comparison of the thres-
hold level for suspended particulates containing high
sulfate with that for particulates with relatively low
sulfate content (26 and 71 Mg/m3, respectively).
REFERENCES FOR SECTION 2.4
1. Broder, J., P.P. Barlow, and R.J.M. Horton. The
Epidemiology of Asthma and Hay Fever in a
Total Community, Tecumseh, Michigan. J.
Allergy. 55:513-523, 1969.
2. Chronic Conditions and Limitations of Activity
and Mobility: United States, July 1965-June
1967. In: Vital and Health Statistics from the
National Health Survey. Health Services and
Mental Health Administration, Public Health
Service, U.S. Department of Health, Education,
and Welfare. Rockville, Md. Series 10, No. 61.
January 1971.
3. Gersh, L.S., E. Shubin, C. Dick, and F.A.
Schulaner. A Study of the Epidemiology of
Asthma in Children in Philadelphia. J. Allergy.
39: 347-357, 1967.
4. Schrenk, H.H., H. Heimann, G.O. Clayton, W.M.
Gafafer, and H. Wexler. Air Pollution in Donora,
Pennsylvania: Epidemiology of the Unusual
Smog Episode of October 1948. Federal Security
Agency, Division of Industrial Hygiene, Public
Health Service, U.S. Department of Health,
Education, and Welfare. Washington, D.C. Public
Health Bulletin 306. 1949.
5. Schoettlin, C.E. and E. Landau. Air Pollution
and Asthmatic Attacks in the Los Angeles Area.
Public Health Reports. 76: 545-548,1961.
6. Zeidberg, L.D., R.A. Prindle, and E. Landau. The
Nashville Air Pollution Study; I. Sulfur Dioxide
and Bronchial Asthma: A Preliminary Report.
Amer. Rev. Respiratory Dis. 84: 489-503,1961.
7. Yoshida, K., H. Oshima, and M. Swai. Air
Pollution and Asthma in Yokkaichi. Arch.
Environ. Health. 13: 763-768, 1964.
8. Yoshida, K., H. Oshima, and M. Swai. Air
Pollution in the Yokkaichi Area with Special
Regard to the Problem of "Yokkaichi Asthma."
Ind. Health. 2: 87-94,1964.
9. Figley, K.D. and R.H. Elrod. Endemic Asthma
Due to Castor Bean Dust. J. Amer. Med. Assoc.
90: 79-82, 1928.
10. Mendes, E. and A.B.V..Centra. Collective Asthma
Simulating an Epidemic Provoked by Castor
Bean Dust. J. Allergy. 25:253-259, 1954.
11. Cowan, D.W., H.J. Thompson, H.J. Paulus, and
P.W. Mielke. Bronchial Asthma Associated with
Air Pollutants from the Grain Industry. J. Air
Pollut. Contr. Assoc. 75:546-552,1963.
12. Cowan, D.W. and H.J. Paulus. Relationship of
Air Pollution to Allergic Diseases. Division of Air
Pollution, Public Health Service, U.S. Depart-
ment of Health, Education, and Welfare.
Washington, D.C. Final report under Research
Grant AP 00090. 1969. 136 p.
13. Lewis, R., M.M. Gilkeson, and R.O. McCalden.
Air Pollution and New Orleans Asthma. Public
Health Reports. 77: 947-954,1962.
14. Weill, H., M.M. Ziskind, R. Dickerson, and V.J.
Derbes. Epidemic Asthma in New Orleans. J.'
Amer. Med. Assoc. 790:811-814, 1964.
2-90
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
15. Weill, H., M.M. Ziskind, V. Derbes, R. Lewis,
R.J.M. Horton, and R.O. McCalden. Further
Observations on New Orleans Asthma. Arch.
Environ. Health. 8: 184-187, 1964.
16. Lewis, R., M.M. Gilkeson, and R.W. Robison. Air
Pollution and New Orleans Asthma (2 volumes).
New Orleans, Tulane University, 1962.
17. Carroll, R.E. Epidemiology of New Orleans
Epidemic Asthma. Amer. J. Public Health. 58:
1677-1683, 1968.
18. Chrobok, H., 0. Wojcik, and J. Zajiczek.
Influence of Meteorological and Weather Condi-
tions on Bronchial Asthma, Cardiac Infarction
and Mortality. Prezeglad Lekarski (Warsaw). 22:
419-421, 1966.
19. Cohen, A.A., S. Bromberg, R.M. Buechley, L.I.
Heiderscheit, and C.M. Shy. Asthma and Air
Pollution from a Coal Fueled Power Plant. Amer.
J. Public Health. 62:1181-1188,1972.
20. Questionnaires Used in the CHESS Studies. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
21. Hertz, M.B., L.A. Truppi, T.D. English, G.W.
Sovocool, RJVI. Burton, L.T. Heiderscheit, and
D.O. Hinton. Human Exposure to Air Pollutants
in Salt Lake Basin Communities, 1940-1971. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004.1974.
22. Draper, N.R. and H. Smith. Multiple Regression
Mathematical Model Building (Chapter 8) and
Multiple Regression Applied to Analyses of
Variance Problems (Chapter 9). In: Applied
Regression Analyses. New York, John Wiley and
Sons, 1966. p. 234-262.
23. Hasselblad, V., G. Lowrimore, and C.J. Nelson.
Regression Using "Hockey Stick" Function. U.S.
Environmental Protection Agency. Research
Triangle Park, N.C. Unnumbered intramural
report. 1971.
24. Quandt, R.E. The Estimation of the Parameters
of a Linear Regression System Obeying Two
Separate Regimes. J. Amer. Statistical Assoc.
55:873-880,1958.
25. Tromp, S.W. Influence of Weather and Climate
on Asthma and Bronchitis. Review of Allergy.
Volume 22, November 1968.
Salt Lake Basin Studies
2-91
-------
CHAPTER 3
ROCKY MOUNTAIN STUDIES
3-1
-------
3.1 HUMAN EXPOSURE TO AIR POLLUTANTS
IN FIVE ROCKY MOUNTAIN COMMUNITIES,
1940-1970
Thomas D. English, Ph.D., Jose M. Sune, M.A.,
Douglas I. Hammer, M.D., M.P.H., Lawrence A. Truppi, M.S.,
Wave E. Culver, M.S., Richard C. Dickerson, B.S.,
and Wilson B. Riggan, Ph.D.
3-3
-------
INTRODUCTION
Western smelting communities usually have little
pollution from automobiles or from nonsmelting
industrial sources. Ambient air pollution in these
communities is characterized by high levels of both
sulfur dioxide and total suspended particulates. These
levels are considered high since they are in excess of
National Primary Air Quality Standards. Since
smelters represent an approximate point source of
pollutants, estimates of previous community
exposures to both sulfur dioxide and particulates can
be made through the use of production data.
f
MONTAMA
-X^HELENA EAST HELENA (3 MILES!
100 MILES
C"l03 BILES \
^ ~~0 BOZEMAN
IDAHO
Health studies of respiratory morbidity in these
communities should help to distinguish the effects of
simple mixtures and of air pollutants like sulfur
dioxide, total suspended particulates, and suspended
sulfates from the effects of more complex urban
mixtures. Such scientific studies can be used to
support National Primary Air Quality Standards that
will adequately protect the public health and yet will
not unnecessarily penalize the economy. In
November 1970, the frequencies of lower respiratory
disease in children through age 12 and of chronic
respiratory disease symptoms in their parents were
surveyed retrospectively in five Rocky Mountain
communities.1 >2 This report provides environmental
exposure information for these health studies and
reports the recent history of ambient exposure to
sulfur dioxide, total suspended particulates, and
suspended sulfates in the five communities. The
present survey was a sequel to the Helena Valley Area
Environmental Report.3
COMMUNITY DESCRIPTION
Five communities were selected on the basis of air
quality data published by the States of Idaho and
Montana and advice from health department officials
in both states.4'8 Kellogg is in the Idaho panhandle,
and Anaconda, East Helena, Helena, and Bozeman are
in Montana (Figure 3.1.1). Kellogg is about 250 miles
northwest of Helena and East Helena; Helena,
Anaconda, and Bozeman form an isosceles triangle,
with all three sides being approximately 100 miles in
length.
Bozeman, Montana
Bozeman, with a population of about 18,500, is
essentially an agricultural community, the site of
Montana State University, and the county seat of
Figure 3.1.1. Location of the five Rocky
Mountain study communities.
Gallatin County. Bozeman is about 4800 feet above
sea level and is located in a large valley bounded on
the east by the Bridger Mountains and on the south
by the Madison Range. The Gallatin River flows
about 7 miles east of the community. There are no
large manufacturing industries in the area. A number
of inversions have been noted in the Bozeman area,
but the air pollution is quite low.7
Helena-East Helena, Montana
Helena and East Helena are about 3 miles apart in
the Helena Valley and have populations of about
23,000 and 1700, respectively, according to the 1970
U. S. Census. The valley is about 25 miles north to
south and 35 miles east to west. Helena, with an
average elevation of 4100 feet, is located on the south
side of this intermountain valley, which is bounded
on the north and east by the Big Belt Mountains and
the west and south by the main chain of the
Continental Divide. The southern parts of the city
have an elevation of about 3800 feet. The average
height of the mountains above the valley floor is
about 3000 feet. There are no large manufacturing
concerns located in the city. Helena is the state
capital and the county seat of Lewis and Clark
County. Since Helena residents live upwind of the
smelter in East Helena, they are only intermittently
exposed to ambient air pollutants emitted by the
smelter (Figure 3.1.2).
East Helena, which has been the site of a lead
smelter since 1888, also has a zinc fuming plant, a
zinc and talc processing plant, and offices and retail
business establishments. Virtually all East Helena
3-4
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
13-18 > 1
WIND SPEED mph
5 10
z
PERCENT OCCURRENCE
Figure 3.1.2. Surface climatologic wind rose,
Helena Municipal Airport, November 1949-
October 1954.
residents do their major shopping in Helena, and
some of them work in Helena. The elevation of East
Helena is about 3900 feet, and the ground slope is
much less than that of Helena. The ground south of
East Helena, where the smelting operations are
located, is 30 to 50 feet higher than the city.6 East
Helena residents also live upwind of the smelter and
are thus only occasionally exposed to the emitted air
pollutants.
Anaconda, Montana
Anaconda, with a population of about 9800, is
situated just north and west of the Continental Divide
at an elevation of about 5300 feet. The mountains
rise abruptly at the south side of the city. Anaconda
is at the southern end of the Deer Lodge Valley,
which is approximately 35 miles long and 8 to 10
miles wide, sloping in a northerly direction from
Anaconda to Garrison (4500 feet). Anaconda is the
county seat of Deer Lodge County and has been the
site of a large copper smelter since the turn of the
century.4'5 The smelter processes copper and zinc
ores from Anaconda mines and does no custom
smelting. The present 585-foot stack was built in
1902 and is almost 1500 feet above the valley floor
because it rests upon a tall slag heap. Electrostatic
precipitators were installed in 1917 and have been
upgraded technologically about every 2 years since
then. Since 1955, the company has used ores from
open pit mines, which are lower in sulfur content
than the underground ores that were used previously.
Since 1965, some sulfur has been metallurgically
eliminated. Hoods were also installed on the con-
verters in 1966 to reduce sulfur dioxide emissions. An
acid plant, which will be completed in the near
future, is expected to reduce sulfur dioxide emissions
to about 300 tons/day.
Kellogg, Idaho
Kellogg has a population of about 3800 and is in
the famous Couer d'Alene mining district of the
Idaho panhandle. Kellogg sits at the base of a narrow
canyon on the west slopes of the Bitterroot Mountain
Range at an elevation of about 2300 feet. The Bunker
Hill Company complex in Kellogg consists of a mine,
an ore concentrator, a smelter, a zinc plant, a
fertilizer plant, and a research and analytical labora-
tory. The Bunker Hill smelter has been a custom lead
and zinc smelter since the turn of the century.
Approximately two-thirds of its ores are from a wide
variety of domestic and foreign mines of other
companies. The zinc plant stack is 250 feet high and
is about 810 feet above the valley floor. The smelter
stack is 200 feet high and is about 337 feet above the
valley floor.
Climatology
The climate for all five communities may be
described basically as modified continental. Winter
temperatures are affected by Artie cold waves and
occasionally may drop well below zero. Summertime
temperatures are moderate, with maximum readings
generally under 90 °F and very seldom reaching
100 °F. Total precipitation varies from approximate-
ly 10 inches in the semiarid northern portion to 30
inches in the humid areas along the Continental
Divide. Most of the precipitation occurs from May to
July, the rest of the year being relatively dry. Snow
may fall from September through May. Cold air
drains into the valleys from surrounding mountain
slopes at night. Hence, strong and persistent tempera-
ture inversions are common throughout this area of
Idaho and Montana.
ESTIMATING HUMAN EXPOSURE TO AIR
POLLUTANTS
In the two nonsmelter communities, Bozeman and
Helena, long-term exposure to sulfur dioxide, total
suspended particulates, and suspended sulfates was
Rocky Mountain Studies
3-5
-------
based on recently measured values.3'6'9 Prior
exposure was assumed to have been similar to recent
exposure since there were no known large changes in
either community emissions or meteorologic patterns.
Estimates of long-term exposure to sulfur dioxide,
total suspended particulates, and suspended sulfates
in the three smelter communities, East Helena,
Kellogg, and Bozeman, were based upon annual metal
production data, estimates of stack emissions of both
particulates and sulfur dioxide, and observed air
quality measurements.
Metal Production and Emission Controls
Annual metal production estimates for each
smelter10"13 were supplemented with actual produc-
tion data and sulfur dioxide and particulate emissions
data provided to the U. S. Environmental Protection
Agency by each corporation.14"16 A summary of the
annual copper, lead, and zinc production data for
East Helena, Anaconda, and Kellogg for the years
1940-1970 is given in Table 3.1.1. Production figures
Table 3.1.1. ANNUAL PRODUCTION OF COPPER, LEAD, AND ZINC IN THREE
ROCKY MOUNTAIN SMELTER COMMUNITIES, 1940-197010'13'16
Year
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
1955
1954
1953
1952
1951
1950
1949
1948
1947
1946
1945
1944
1943
1942
1941
1940
Production, tons/year
East Helena
lead
57,721
45,343
42,824
50,160
67,133
65,019
69,926
63,693
62,433
58,650
74,412
37,882
69,310
67,517
62,133
58,774
72,694
72,144
60,265
61,033
79,394
64,817
55,217
57,737
28,097
39,995
49,463
59,232
67,010
70,197
—
Anaconda
copper
120,412
103,314
69,480
65,483
128,061
115,489
103,806
79,762
94,021
104,000
91,972
65,91 1
90,683
91,512
96,426
81,542
59,349
77,617
61,948
57,406
54,478
56,61 1
58,252
57,900
58,481
88,506
117,856
138,295
140,465
128,712
129,071
Kellogg
Lead
123,106
123,986
124,134
122,247
113,194
93,753
86,282
93,430
93,961
107,857
31,658
91,940
91,769
103,420
99,816
88,126
82,520
69,635
57,220
52,501
73,294
43,414
58,150
51,038
33,707
37,974
52,638
53,265
67,389
65,739
54,642
Zinc
95,637
105,700
102,946
92,134
90,000
90,990
91,760
81,296
76,756
74,735
26,449
61,191
55,454
68,832
57,799
56,625
47,404
54,037
54,340
54,468
53,922
41,853
42,067
41,801
34,833
33,110
36,563
41,129
39,916
39,285
37,471
3-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
for the American Smelting and Refining Company
(Asarco) at East Helena and for the Bunker Hill
Corporation at Kellogg were taken from the year-
books of the American Bureau of Metal
Statistics.10"13 In the case of the Anaconda Corpora-
tion, the yearbooks give total values for smelting
performed at Anaconda, Montana; Miami, Arizona;
and Hayden, Arizona. In order to estimate the
amount of smelting at the Anaconda plant at
Anaconda, Montana, figures for the total copper
produced in the state of Montana from domestic ores
were used. These production figures may be slightly
low because they do not include the ore that
Anaconda obtains from Canada. For example, in
1971, 7339 tons of ore, matte, and regulous copper
were imported into the United States from Canada. If
the assumption were made that all of these
imports were smelted in Anaconda, Montana, this
would only increase the estimated production by
approximately 7 percent.
Sulfur dioxide and particulate emission factors for
copper smelting, lead smelting, and zinc smelting17
were applied to the annual metal production figures
in order to determine average daily stack emissions of
both sulfur dioxide and particulates. Estimates of
stack emissions for both particulate and sulfur
dioxide for East Helena for the years 1941-1970 were
provided by Asarco. Estimated sulfur emissions for
the Bunker Hill plant in Kellogg were provided for
the years 1940,1945, 1950,1955,1959,1965,1970,
and 1971 by the Bunker Hill Corporation. Estimates
of particulate emissions for the years 1970 and 1971
were also provided by the same company. Estimates
of particulate emissions for the years 1965 and 1971
were provided by the Anaconda Corporation. Sulfur
dioxide emission measurements were provided by
Anaconda for the year 1971. These emissions were
compared with those determined by application of
emission factors. On the basis of these comparisons,
estimates of pollution control efficiencies were made
for the three areas.
At the Anaconda smelter, electrostatic pre-
cipitate rs were installed in 1917 and upgraded tech-
nologically about every 2 years. Anaconda's precipita-
tors are presently estimated to be 90 percent
efficient. It is quite reasonable to assume this
particulate emission control efficiency for the period
1940-1970. In the case of sulfur dioxide, the data
provided indicate a 23 percent control of sulfur
dioxide emissions for the years 1965-1971. Since
1955, pit mined ores, which have less sulfur concen-
tration than vein mined ores, have been used. It is
difficult to estimate the actual controls during the
period 1940-1964 because of the change in both the
type of ore used and the metallurgical process. An
approximate average control of 11 percent for sulfur
dioxide emissions was assumed for this period.
Since sulfur emissions data are available for 8 years
in Kellogg, better estimates of sulfur dioxide emission
controls can be made. The evidence indicates approxi-
mately 50 percent control during the period
1959-1971, 43.5 percent control during the period
1955-1958, and no control during the period
1940-1954. In the case of particulate emissions, the
estimated factors are 97.5 percent control from
1968-1971, 95 percent control from 1965-1967, and
90 percent control from 1940-1964. Summaries of
estimated sulfur dioxide and particulate emissions for
East Helena, Anaconda, and Kellogg during the
period 1940-1970 are shown in Tables 3.1.2 and
3.1.3.
Exposure Estimates for East Helena
By comparing observed air quality data to average
daily pollutant emissions from a given isolated point
source, one can estimate the ratio of annual average
pollutant concentration to the average rate of pol-
lutant emission from the point source. In the case of
East Helena, annual average total suspended par-
ticulate concentrations are available for the years
1966, 1967, 1968, and 1969. The ratios of observed
annual average total suspended particulate concentra-
tion to average daily particulate emitted are shown
for these years in Table 3.1.4. The average ratio is
1.05 (jug/m3)/(ton/day). The range of the ratios is
0.50. Using the techniques of inefficient statistics
described by Dixon and Massey,18 the standard
deviation of the data was estimated to be 0.24. Using
the scientific convention of expressing variation at
the 50 percent confidence level, the ratio of total
suspended particulate concentration to tons of par-
ticulate emitted per day was estimated to be 1.05 ±
0.16(jug/m3)/(ton/day).
Similar date for sulfur dioxide are also given in
Table 3.1.4. Using the same analysis, the ratio of the
concentration of sulfur dioxide to tons of sulfur
dioxide emitted per day was estimated to be 0.33 ±
0.10 (Mg/m3)/( ton/day). These ratios were multiplied
by the appropriate daily emissions of sulfur dioxide
and particulates (Tables 3.1.2 and 3.1.3) to obtain
estimated annual average concentrations of total
suspended particulate and sulfur dioxide for the time
period 1941-1971. The results of these calculations
for East Helena are given in Table 3.1.5. These total
suspended particulate and sulfur dioxide estimates are
Rocky Mountain Studies
3-7
-------
Table 3.1.2. ESTIMATED AVERAGE DAILY
SULFUR DIOXIDE EMISSIONS IN THREE
ROCKY MOUNTAIN SMELTER COMMUNITIES,
1940-19701°-17
(tons/day)
Table 3.1.3. ESTIMATED AVERAGE DAILY
PARTICULATE EMISSIONS IN THREE ROCKY
MOUNTAIN SMELTER COMMUNITIES,
1940-19701017
(tons/day)
Year
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
1955
1954
1953
1952
1951
1950
1949
1948
1947
1946
1945
1944
1943
1942
1941
1940
SOz emissions
East
Helena
239
221
153
104
128
118
116
113
108
120
143
71
115
120
112
113
146
148
116
116
139
102
108
108
61
95
108
111
119
125
-
Anaconda
635
545
367
346
675
609
629
483
570
630
558
399
550
555
584
495
360
471
375
348
330
334
353
352
355
537
714
838
852
781
783
Kellogg
262
266
262
245
233
219
212
203
200
140
67
164
220
260
234
219
291
289
267
259
285
205
332
218
166
181
205
220
242
237
219
Year
1970
1969
1%8
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
1955
1954
1953
1952
1951
1950
1949
1948
1947
1946
1945
1944
1943
1942
1941
1940
Particulate emissions
East
Helena
0.323
0.304
0.298
0.219
0.298
0.304
0.323
0.328
0.271
0.246
0.287
0.095
0.263
0.367
0.304
0.520
0.441
0.583
0.353
0.413
0.564
0.479
0.583
0.994
0.846
1.139
1.084
0.646
0.452
0.301
-
Anaconda
8.9
7.6
5.1
4.8
9.5
8.5
7.8
5.9
7.0
7.7
6.8
4.9
6.7
6.8
7.1
6.0
4.4
5.7
4.6
4.2
4.0
4.2
4.3
4.3
4.3
6.5
8.7
10.2
10.4
9.5
9.5
Kellogg
3.21
3.41
3.36
6.24
5.97
5.66
5.52
5.25
5.05
3.84
1.70
4.33
4.23
4.89
4.32
4.05
7.20
7.22
6.75
6.57
7.35
5.23
5.77
5.52
4.26
4.22
2.17
2.20
2.78
2.72
2.77
presented graphically in Figures 3.1.3 and 3.1.4 in
order to indicate the relative precision of the
estimates.
Suspended sulfate data for East Helena are avail-
able. During the period 1965-1966,66 measurements
of suspended sulfate indicated an annual average
concentration of 7.9 Mg/m3. The observed total
suspended particulate during this period of time was
96 Mg/m3. Hence the ratio of suspended sulfate
observed to particulate observed during this period
was 0.082. A series of 25 measurements made during
the time period April to August 1968 indicated a
suspended sulfate to total suspended particulate ratio
of 0.057. In order to estimate the range of this ratio,
measurements that were made during the time period
June through October 1969 at two stations were
compared. These measurements indicated a
suspended sulfate to total suspended particulate ratio
range of 0.037. Using the statistical technique
described previously, the ratio of suspended sulfate to
total suspended particulate for East Helena was
estimated to be 0.063 ± 0.022 (jug/m3)/ ( /ig/m3 ).
By multiplying this factor by the total suspended
3-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.1.4. RATIO OF POLLUTANT CONCEN-
TRATIONS TO POLLUTANT EMISSIONS FOR
EAST HELENA, MONTANA
Pollutant
and
year
Total sus-
pended
particulate
1969
1968
1967
1966
Sulfur
dioxide
1971
1970
1969
Emissions,
tons/day
111
109
80
109
264
239
221
Observed annual
average
concentration.
Mg/m3
136
97
104
87
78
52
104
Ratio
1.23
0.89
1.30
0.80
0.296
0.217
0.470
particulate concentrations in Table 3.1.5, estimates
were made of the suspended sulfate concentrations
for East Helena during the period 1940-1970. The
results of these calculations are also given in Table
3.1.5. A graphical summary of the estimates of the
suspended sulfates concentrations for East Helena is
shown in Figure 3.1.5.
Examination of Table 3.1.5 shows that the
estimates of annual average total suspended par-
ticulate for East Helena were consistently above the
National Primary Air Quality Standard of 75 Mg/m3
for an annual geometric mean.19 For example, during
the time period 1968-1970, average total suspended
particulate concentration was estimated to be 115
Mg/m3. Similarly, the estimated average total
suspended particulate concentration was 146 Mg/m3
during the decade 1950-1959, and 270 Mg/m3 during
the 9-year period 1941-1949.
In contrast to these high total suspended particu-
late concentrations, the sulfur dioxide levels in East
Helena (Table 3.1.5) have been lower than the
National Primary Air Quality Standard (80 Mg/m3
annual arithmetic mean).19 For example, during the
time period 1968-1970, the average sulfur dioxide
concentration was estimated to be 67 Mg/m3. During
the period 1964-1967, an even lower level of sulfur
dioxide concentration, 38 Mg/m3, was estimated. The
estimated annual average had also been low during
the decades of the forties and fifties. Suspended
sulfate estimates (Table 3.1.5) were 7.3 Mg/m3 during
the period 1968-1970, and 9.2 Mg/m3 during the
Table 3.1.5. ESTIMATED ANNUAL AVERAGE
CONCENTRATIONS OF SULFUR DIOXIDE,
TOTAL SUSPENDED PARTICULATE, AND
SUSPENDED SULFATE FOR EAST HELENA,
1941-1970
(M9/m3)
Year
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
1955
1954
1953
1952
•1951
1950
1949
1948
1947
1946
1945
1944
1943
1942
1941
Pollutant
S02
78.9
72.9
50.5
34.3
42.2
38.9
38.3
37.3
35.6
39.6
47.2
23.4
38.0
39.6
37.0
37.3
48.2
48.8
38.3
38.3
45.9
33.7
35.6
35.6
20.1
31.4
35.6
36.6
39.3
41.3
TSP
123.8
116.5
114.2
83.9
114.2
116.5
123.8
125.7
103.9
94.3
110.0
36.4
100.8
140.7
116.5
199.3
169.1
223.4
135.3
158.3
216.2
183.6
223.4
381.0
324.2
436.5
415.4
247.6
173.2
115.4
SS
7.8
7.3
7.2
5.3
7.2
7.3
7.8
7.9
6.5
5.9
6.9
2.3
6.4
8.9
7.3
12.6
10.7
14.1
8.5
10.0
13.6
11.6
14.1
24.0
20.4
27.5
26.2
15.6
10.9
7.3
decade of the fifties. At present there is no National
Primary Air Quality Standard for suspended sulfates.
Exposure Estimates for Kellogg
Two years of data for total suspended particulate
were available for Kellogg, Idaho.20-21 During 1970,
the total suspended particulate concentration
averaged 107 Mg/m3. During 1971, the corresponding
value was 92 Mg/m3. By comparing these values with
the particulate emissions for Kellogg in Table 3.1.3,
an estimate can be made of the ratio of total
suspended particulate concentration to particulate
Rocky Mountain Studies
3-9
-------
o
I—
cc
LU
O
1
o:
-------
ro
{*
<£
DC
LU
O
c
-------
1968-1970, the average suspended sulfate level was
estimated to be 11.3 Mg/m3. A value of 19.9 Mg/m3
was estimated for the period 1964-1967. During the
decade of the fifties, an average of 19.3 Mg/m3 was
estimated.
Exposure Estimates for Anaconda
During the period 1961-1962, the annual total
suspended particulate concentration was found to be
84.5 /ng/m3. In 1971, the average suspended particu-
late level was 52 Mg/m3. By comparing the observed
total suspended particulate concentration with the
particulate emitted from the Anaconda plant, a ratio
of 9.1 ± 2.3 (Mg/m3)/(ton/day) was determined. This
ratio was multiplied by the particulate emission for
Anaconda shown in Table 3.1.3 to estimate the total
suspended particulate concentrations for the years
1940-1970. The results of these calculations are
shown in Table 3.1.7. Estimated total suspended
particulate concentrations averaged 65 Mg/m3 during
1968-1970, which represented a moderate decrease
Table 3.1.6. ESTIMATED ANNUAL
AVERAGE CONCENTRATIONS OF
SULFUR DIOXIDE, TOTAL
SUSPENDED PARTICULATE, AND
SUSPENDED SULFATE FOR
KELLOGG, 1940-1971 (M9/m3)
Year
1971
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
19t5
19b-»
1953
1952
1951
1950
1949
1948
1947
1946
1945
1944
1943
1942
1941
1940
Pollutant
SO 2
411 8
372.0
377.7
3720
3479
3309
311 0
301 0
288.3
2840
198.8
95 1
232.9
3124
369.2
3323
311 0
4132
4104
3791
3678
404.7
291 1
471 4
309.6
2357
2570
291 1
3124
3436
3365
311 0
TSP
995
995
105.7
1042
1934
185.1
1755
171 1
162.8
156.6
1190
527
1342
131 1
151 6
1339
1256
b3
109
109
11 6
11.4
21.2
20.3
197
18.2
178
172
131
5.7
14.7
14.4
16.6
14.7
13.8
223 2 ! 24 4
223 8 24 5
209.3
203.7
227.9
162 1
178.9
171.1
132.1
130.8
673
68.2
862
843
859
22.9
223
24.9
177
196
18.7
145
14.3
74
75
9.4
9.2
94
from an estimated average of 69 Mg/m3 during
1964-1967.
For 1965, the Montana State Department of
Health reports the average annual concentration of
sulfur dioxide in Anaconda was 80Mg/m3. For 1971,
the corresponding value was 286 Mg/m3. These values
are based on the results of analysis of sulfation plates.
Comparing these observed concentrations with the
sulfur dioxide emissions in Table 3.1.2, a ratio of
0.343 ± 0.253 (Mg/m3)/(ton/day) was obtained for
the Anaconda area. In order to estimate the average
sulfur dioxide concentrations during the period
1940-1970, the sulfur dioxide emissions shown in
Table 3.1.2 were multiplied by this factor. The
annual values of sulfur dioxide in Anaconda are
estimated to be high (Table 3.1.7). For example, in
the period 1968-1970, an average concentration of
177 Mg/m3 was estimated. The corresponding
estimated values were 193 Mg/m3 for 1964-1967,192
Mg/m3 for 1960-1963,153 Mg/m3 for 1950-1959, and
203Mg/m3 for 1940-1949.
Since no data were available for suspended sulfate
concentrations in the Anaconda area, an approach
Table 3.1.7. ESTIMATED ANNUAL
AVERAGE CONCENTRATIONS OF
SULFUR DIOXIDE, TOTAL
SUSPENDED PARTICULATE, AND
SUSPENDED SULFATE FOR
ANACONDA, 1940 - 1971 (M9/m3)
Year
1971
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
1955
1954
1953
1952
1951
1950
1949
1948
1947
1946
1945
1944
1943
1942
1941
1940
Pollutant
S02
1770
217.8
1869
125.9
118.7
231 5
208.9
2157
1657
195.5
216.1
191.4
136.9
1887
1904
200.3
169.8
123.5
161.6
128.6
119.4
1132
1146
121.1
120.7
121 8
1842
2449
287.4
2922
267.9
268.6
TSP
65.2
806
689
462
43.5
86.1
77.0
707
53.5
634
698
61.6
44.4
607
61.6
64.3
54.4
399
51.6
41.7
381
362
381
39.0
39.0
39.0
589
78.8
92.4
94.2
86 1
86 1
SS
7.2
8.9
76
5.1
4.8
9.6
8.5
7.8
5.9
7.0
7.7
6.8
49
6.7
6.8
7 1
6.0
4.4
5.7
4.6
4.2
40
42
4.3
4.3
4.3
6.5
8.7
10.3
10.5
9.6
9.6
3-12
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
identical to that used in Kellogg was used. The ratio
of suspended sulfate concentration to the total
suspended particulate concentration was estimated to
be 0.11 ± 0.06 on the basis of the results found in
East Helena, Helena, and Magna. Estimates of
historical sii«tlorHe:i sulf«, con 'ntrations for
Kellogg v >aade by mui ' lying H'J°. factor by the
total suspended particulate values snown In Table
3.1.7. The values of suspended sulfate for the time
period 1940-1970 are also given in Table 3.1.7. For
example, for 1968-1970, an average suspended sulfate
level of 7.2 jug/m3 was estimated. During the period
1964-1967, this level was estimated to be 7.7 Mg/m3.
HUMAN EXPOSURE SUMMARY
A graphical summary of the estimated annual
concentrations of total suspended particulate for
Anaconda, East Helena, and Kellogg for the time
period 1940-1970 is shown in Figure 3.1.6. In
general, Anaconda is characterized by total suspended
particulate levels that were below 75 jiig/m3, the
National Primary Air Quality Standard for total
suspended particulate, whereas the levels in both East
Helena and Kellogg tended to be considerably above
this value. In the decade of the forties, total
suspended particulate in East Helena was estimated to
have averaged 278 Mg/m3. The weighted average of
340 total suspended particulate samples taken at four
stations in Bozeman during 1967-19687 was 50
Mg/m3. This value was used as an estimate of total
suspended particulate concentration during the
period 1940-1970 for Bozeman. For Helena, 13 years
of data were available for the estimation of historical
trends for total suspended particulate. These data
indicate geometric means ranging from 41 to 75
Mg/m3.
The summary graph of the estimated sulfur
dioxide concentrations for Anaconda, East Helena,
and Kellogg is shown in Figure 3.1.7. Both Anaconda
and Kellogg experienced estimated sulfur dioxide
concentrations in excess of the present National
Primary Air Quality Standard of 80 Mg/m3, whereas
East Helena experienced concentrations consistently
below this level. For example, during the decade of
the fifties, East Helena was estimated to have had an
average sulfur dioxide concentration of 39 Mg/m3 in
contrast to an estimate of 353 Mg/rn3 for Kellogg.
Measured ambient sulfur dioxide levels in Bozeman
during 1967 were 10 Mg/m3; in Helena, measured
sulfur dioxide levels were 26 jug/m3 in 1967 and
1969.7'9 These respective values were used as an
estimate of sulfur dioxide exposure for 1940-1970
for the latter two communities.
A similar summary of the estimated annual sus-
pended sulfate concentrations for Anaconda, East
Helena, and Kellogg is shown in Figure 3.1.8. In this
case, during the decade of the sixties, Anaconda and
East Helena have similar exposures, whereas those in
Kellogg are considerably higher. The weighted average
of 340 samples of suspended sulfates taken in
Bozeman during 1967-19687 was 3.3 Mg/m3. A
similar weighted average for 123 samples in Helena
during 1966-19676 was 4.9 jug/m3. These values were
used to represent exposure for 1940-1970 in
Bozeman and Helena.
A summary of the overall exposure to sulfur
dioxide, total suspended particulate, and suspended
sulfate for the five communities is given in Table
3.1.8. Since the estimated sulfur dioxide levels in
Anaconda and Kellogg are so much higher than in the
other communities, these communities are referred to
as High I and High II. Bozeman, Helena, and East
Helena are referred to, respectively, as Low I, Low II,
and Low III.
Rocky Mountain Studies
3-13
-------
O
1—
I
LU
O
O
O
cc
Q-
0
1940 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70
TIME, years
Figure 3.1.6, Estimated annual average total suspended particulate concentrations
versus time for Anaconda, East Helena, and Kellogg.
500 —
-------
0
1940 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70
TIME, years
Figure 3.1.8. Estimated annual average suspended sulfate concentrations versus
time for Anaconda, East Helena, and Kellogg.
Table 3.1.8. ESTIMATED ANNUAL AVERAGE AMBIENT SULFUR DIOXIDE, TOTAL SUSPENDED
PARTICULATE, AND SUSPENDED SULFATE EXPOSURES IN FIVE ROCKY
MOUNTAIN COMMUNITIES, 1940-1970
Pollutant and
community3
Sulfur dioxide
Low I
Low 1 1
Low III
High I
High II
Total suspended
particulate
Low I
Low 1 1
Low III
High!
High 1 1
Suspended sulfate
Low I
Low 1 1
Low III
High I
High 1 1
Estimated annual average concentration, jug/m
1940-
1949
10
26
34
203
316
50
60
270
65
115
3.3
4.9
17.1
6.8
12.8
1950-
1959
10
26
39
153
353
50
60
146
49
174
3.3
4.9
9.2
5.4
19.3
1960-
1963
10
26
40
192
217
50
48
106
62
121
3.3
4.9
6.7
6.9
13.5
1964-
1967
10
26
38
193
323
50
55
107
69
179
3.3
4.9
6.8
7.7
19.9
1968-
1970
10
26
67
177
374
50
45
115
65
102
3.3
4.9
7.3
7.2
11.3
Low I is Bozeman; Low II, Helena; Low III, E. Helena; High I, Anaconda; and High II, Kellogg.
Rocky Mountain Studies
3-15
-------
ACKNOWLEDGMENTS
The authors are grateful to the Idaho and Montana
State Departments of Health and to the American
Smelting and Refining Company, the Anaconda
Corporation, and the Bunker Hill Company (a sub-
sidiary of Gulf Resources and Chemical Corporation)
for supplying the U. S. Environmental Protection
Agency with air pollutant measurements, available
production and emissions data, and other helpful
information regarding the history of these smelter
operations.
REFERENCES FOR SECTION 3.1
1. Finklea, J. F., D. I. Hammer, D. E. House, C. R.
Sharp, W. C. Nelson, and G. R. Lowrimore.
Frequency of Acute Lower Respiratory Disease
in Children: Retrospective Survey of Five Rocky
Mountain Communities, 1967-1970. In: Health
Consequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
2. Hayes, C.G., D.I. Hammer, C.M. Shy, V.
Hasselblad, C. R. Sharp, J. P. Creason, and K. E.
McClain. Prevalence of Chronic Respiratory
Disease Symptoms in Adults: 1970 Survey of
Five Rocky Mountain Communities. In: Health
Consequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N. C.
Publication No. EPA-650/1-74-004. 1974.
7. A Study of Air Pollution in Townsend-Three
Forks, the Gallatin Valley, and West Yellow-
stone, November 1967-November 1968. Montana
State Department of Health. Helena, Montana.
43 p.
8. Idaho Air Quality—Methods of Measurement and
Analysis of Recent Data. Idaho State Depart-
ment of Health. Boise, Idaho. August 19, 1970.
23 p.
9. Maughan, D. Personal communication to J.M.
Sune, U. S. Environmental Protection Agency,
Research Triangle Park, N.C. Montana State
Department of Health and Environmental
Sciences. Helena, Montana. June 12,1972.
10. Yearbook of the American Bureau of Metal
Statistics, 51st Annual Issue for 1971. New
York, American Bureau of Metal Statistics, 1972.
p. 34, 57, 84.
3. Helena Valley, Montana, Area Environmental
Pollution Study. U. S. Environmental Protection
Agency. Research Triangle Park, N. C. Office of
Air Programs Publication No. AP-91. January
1972.193 p.
4. A Study of Air Pollution in Montana, July
1961-July 1962. Montana State Board of Health.
Helena, Montana. 107 p.
5. A Study of Air Pollution in the Deer Lodge
Valley-August 1965-June 1966. Montana State
Department of Health. Helena, Montana. 32 p.
6. A Study of Air Pollution in the Helena-East
Helena Area-October 1965-October 1968.
Montana State Department of Health. Helena,
Montana. 35 p.
11. Yearbook of the American Bureau of Metal
Statistics, 44th Annual Issue for 1964. New
York, American Bureau of Metal Statistics, 1965.
p. 34, 56,81.
12. Yearbook of the American Bureau of Metal
Statistics, 35th Annual Issue for 1955. New
York, American Bureau of Metal Statistics, 1956.
p. 29,49,74.
13. Yearbook of the American Bureau of Metal
Statistics, 29th Annual Issue for 1949. New
York, American Bureau of Metal Statistics, 1950.
p. 24,38,60.
14. Baker, G. M. Personal communication to D. I.
Hammer, U.S. Environmental Protection
Agency, Research Triangle Park, N. C. The
Bunker Hill Company, Kellogg, Idaho. June 22,
1972.
3-16
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
15. Laird, F. Personal communication to D. I. Ham-
mer, U. S. Environmental Protection Agency,
Research Triangle Park, N. C. Anaconda Cor-
poration, Tucson, Arizona. July 6,1972.
16. Nelson, K.W. Personal communication to D.I.
Hammer, U.S. Environmental Protection
Agency, Research Triangle Park, N. C. American
Smelting and Refining Company, Salt Lake City,
Utah. June 22,1972.
17. Compilation of Air Pollutant Emission Factors.
U. S. Environmental Protection Agency, Re-
search Triangle Park, N. C. Office of Air Pro-
grams Publication No. AP42. February 1972.
166 p.
18. Dixon, W. J. and F. J. Massey. Introduction to
Statistical Analysis (2nd ed.). New York,
McGraw-Hill Book Company, Inc., 1957.
19. U. S. Environmental Protection Agency. National
Primary and Secondary Air Quality Standards.
Federal Register. Vol. 36, No. 84, April 30,
1971.
20. Eiguren, A. J. Personal communication to D. I.
Hammer, U.S. Environmental Protection
Agency, Research Triangle Park, N. C. Idaho
State Department of Health, Air Pollution Con-
trol Commission, Boise, Idaho. October 5, 1971.
21. Michael, M. Personal communication to D.I.
Hammer, U.S. Environmental Protection
Agency, Research Triangle Park, N. C. Idaho
State Department of Health, Environmental
Improvement Division, Boise, Idaho, April 3,
1972.
22. Spirtas, R. and H. J. Levin. Characteristics of
Particulate Patterns-1957-1966. National Air
Pollution Control Administration, Public Health
Service, U. S. Department of Health, Education,
and Welfare. Raleigh, N. C. Publication No.
AP-6 I.March 1970. 101 p.
Rocky Mountain Studies
3-17
-------
3.2 PREVALENCE OF CHRONIC RESPIRATORY
DISEASE SYMPTOMS IN ADULTS: 1970 SURVEY
OF FIVE ROCKY MOUNTAIN COMMUNITIES
Carl G. Hayes, Ph.D., Douglas I. Hammer, M.D., M.P.H.,
Carl M. Shy, M.D., Dr. P.H., Victor Hasselblad, Ph.D.,
Charles R. Sharp, M.D.^M.P.H., John P. Creason, M.S.,
and Kathryn E. McClain
3-19
-------
INTRODUCTION
Studies in several countries have linked episodic
community air pollution exposure with increased
morbidity and mortality among people with chronic
bronchitis.1'12 The role of long-term exposure to
lower ambient levels of sulfur dioxide in the induc-
tion and exacerbation of chronic respiratory disease is
somewhat less conclusive. United States, Australian,
British, Italian, and Japanese studies have related
ambient sulfur dioxide and particulate exposures with
chronic bronchitis prevalence and an increase in
chronic respiratory symptoms in adults.13"17 British
and U.S. prospective studies of patients with chronic
bronchitis have shown a positive correlation between
acute symptom reporting and daily air pollution
levels.18'20
In* addition to ambient air pollution, chronic
respiratory disease can result from pollution from
other sources, such as occupational exposure to
irritating fumes and dusts and exposure to cigarette
smoke. Demographic characteristics and residential
mobility also are related to risk of chronic bronchitis
and less severe symptoms of chronic respiratory
disease. It has not been possible to consider all of
these important covariates in most previous studies.
Furthermore, most studies have not been designed to
distinguish between the effects of sulfur dioxide and
other pollutants such as oxides of nitrogen, sus-
pended sulfates, suspended nitrates, and total
suspended particulates. This problem was summarized
succinctly in the following recent comments sub-
mitted by an American industrial spokesman "...in no
study was SO2 the sole ah- pollutant.... Accordingly,
none of the studies reliably isolates SOj as the
causative factor which accounted for the observed
differences in health effects."21
Western smelting communities provide an un-
usual opportunity to observe populations exposed to
high sulfur dioxide levels without the elevated
particulate levels with which they are usually ac-
companied. The same corporation spokesman
incisively points out that, "In the comparatively open
rural areas of the West where smelters are typically
located, particulate levels are normally very low....
Moreover, the composition of particulate matter in
western areas, where population density is low .and
natural gas is the man source of space heating, can be
expected to be quite different from that found in the
urban areas in England or the eastern United States
that are dealt with in epidemiological studies."21 The
need for distinguishing the health effects of sulfur
dioxide from those of other air pollutants is a
compelling reason to study western smelting com-
munities. Point source emissions from smelters usual-
ly constitute the only major source of pollution, and
production and emissions data are often available for
the past 20 to 30 years. Thus, one can reconstruct
estimates of average ambient air concentrations of
sulfur dioxide, suspended sulfates, and total sus-
pended particulates in smelting communities for
several decades.
Three smelting and two nonsmelting communities
were selected for this survey of chronic respiratory
disease. Available information indicated that these
cities would provide a marked contrast in previous
exposure to sulfur dioxide. Three specific hypotheses
were proposed: first, that frequency and severity of
chronic respiratory symptoms vary across areas with
sulfur dioxide exposure; second, that rates of chronic
bronchitis are related to length of exposure in
polluted communities; third, that the combined
deleterious effects of air pollution and cigarette
smoking are greater than those of either factor alone.
METHODS
Community Selection
Five communities in the Montana-Idaho area were
selected for study. In three of these communities,
large smelters have been operating for over 50 years,
while in the other two, there has been no heavy
industry and best available information indicated no
major pollution sources. In addition to the obvious
pollution potential of the communities, their selec-
tion was based on routine aerometric monitoring
data, collected by Montana and Idaho State health
departments in conjunction with Federal
agencies.22'29 In two of the communities, daily
samples from continuous monitors were available; but
in others, less sophisticated monitoring techniques
had been utilized, and conversions were necessary to
estimate sulfur dioxide levels. Subsequently, in order
to characterize previous long-term exposure, ad-
ditional monitoring has been instituted and historic
smelter production data calibrated to these measure-
ments have been utilized to reconstruct earlier pollu-
tion levels.3 °
Bozeman, Montana, the southernmost city, is a
farming and college community. Its 1970 population
of 18,500 represented an increase of 40 percent
during the previous decade. Approximately 100 miles
north are Helena and East Helena. Both towns are
located on a sloping valley 25 to 35 miles wide and
3-20
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
tend to be swept by winds. Helena, the state capital,
had a 1970 population of 23,000, representing a 12
percent growth since 1960. East Helena has a
population of less than 1700 and has experienced an
11 percent growth since 1960. Although a lead
smelter has been in operation for over 80 years in
East Helena, its residents are protected to some
extent by the location of the town with respect to
the prevailing wind patterns. Anaconda, located 90
miles southwest of Helena in a deep valley partially
surrounded by high mountains, is the site of a large
copper smelter. Its 1970 population of 9800 has
decreased 'by 19 percent since 1960. Kellogg, Idaho,
250 miles northwest of Helena, has both lead and
zinc smelting operations, and is located in a narrow
valley that limits ventilation. Its 1970 population of
3800 had decreased by 25 percent during the
previous 10 years.
In summary, Helena and Bozeman were relatively
clean nonindustrial cities, while East Helena,
Anaconda, and Kellogg each had smelting operations.
When compared to Anaconda and Kellogg, East
Helena city residents were exposed to less of the
impact of the smelter emissions.
Data Collection
In each of the selected communities, families for
study were selected through elementary school
records. During November of 1970, a School and
Family Health Questionnaire with explanatory letter
was sent home by children to be completed by their
parents and returned to the school. An example of
the form used is presented elsewhere.31 In Kellogg,
Anaconda, and East Helena, all existing elementary
schools were included. Only two schools chosen for
their comparability were selected in each of the larger
cities of Helena and Bozeman. Three of the eight
schools in Anaconda were Catholic parochial. The
remainder in all cities were from public school
systems.
The questionnaire included a set of seven ques-
tions dealing with chronic respiratory disease
symptoms (adapted for self-administration from
those used by the British Medical Research
Counsel).32 Another set of questions referred to
pollution exposure history, including smoking
patterns, occupational exposure to irritating smoke,
dust, or fumes, and residential history. In addition,
age, sex, race, and education were ascertained.
Morbidity Indices
Responses to questions regarding respiratory
symptoms of shortness of breath and frequency and
duration of phlegm production (from chest) and
cough were used to develop a scale of severity as
follows:
1. No symptoms.
2. Cough alone for less than 3 months each year.
3. Phlegm production with or without cough for
less than 3 months each year.
4. Cough without phlegm production for 3
months or more each year.
5. Phlegm without cough for 3 months or more
each year.
6. Cough and phlegm for 3 months or more each
year.
7. Cough and phlegm for 3 months or more each
year with shortness of breath.
Severity categories 6 and 7 are considered equivalent
to chronic bronchitis for the purpose of this study.
Three indices were derived to summarize
morbidity of various study segments: first, chronic
bronchitis prevalence rates; second, mean symptom
scores computed by averaging the score, or rank, on
the scale of severity; and third, mean severity scores
computed by averaging the symptom score only for
those who reported symptoms. Through the use of
these indices, one is able to gain insight into observed
differences in terms of minor and severe symptoms.
Data Analysis
Distributions of important covariates were
compared across areas, as were the effects of those
covariates on chronic respiratory disease. Appropriate
controls or adjustments were utilized for display of
the data, and a general linear model for categorical
data was employed for the statistical analyses.3 3 Chi
Square partitioning procedures were used to test
hypotheses.
RESULTS
Environmental Exposure
Air pollution exposure since 1940 has been re-
constructed from existing aerometric measurements
and available smelter production information (Table
Rocky Mountain Studies
3-21
-------
3.2.1).30 Sulfur dioxide exposure clearly distin-
guishes two of the smelter communities, Anaconda
(High I) and Kellogg (High II), from the other smelter
community, East Helena (Low III), and the two
nonsmelter communities, Bozeman (Low I) and
Helena (Low II).
This distinction is not seen in the total suspended
particulate or the suspended sulfate averages. East
Helena has the highest total suspended particulate
average and is intermediate, more closely resembling
the High communities, in suspended sulfate levels.
The size of the East Helena population did not permit
its analysis as a separate Intermediate pollution area;
therefore, it was considered in the analyses as a Low
sulfur dioxide community. Levels of sulfur dioxide,
as well as total suspended particulates and suspended
sulfates, have been higher on the average in the High
II than in the High I community.
Response Rates
Questionnaire response rates were similar across
the five cities, ranging from 83.3 to 85.0 percent
(Table 3.2.2). The small differences that were
observed did not indicate a likely source of bias
between the High and Low cities since the pooled
response for the Low areas and for the High areas
were almost equivalent. Unfortunately, non-
respondents could not be personally interviewed as
has been done in more recent Community Health and
Environmental Surveillance System (CHESS) studies.
Characterization of the Study Population
Over 97 percent of the respondents were white
and approximately 1 percent were either black,
Indian, Mexican-American, or Oriental. The
remainder did not indicate their race. Socioeconomic
Table 3.2.1. ESTIMATED ANNUAL AVERAGE AMBIENT SULFUR DIOXIDE, TOTAL SUSPENDED
PARTICULATE, AND SUSPENDED SULFATE EXPOSURES IN FIVE ROCKY
MOUNTAIN COMMUNITIES, 1940-1970
Pollutant and
community3
Sulfur dioxide
Low I
Low II
Low 1 1 1
High I
High 1 1
Total suspended
particulate
Low I
Low 1 1
Low III
High I
High 1 1
Suspended sulfate
Low I
Low II
Low III
High I
High II
Estimated annual average concentration, jug/m3
1940-
1949
10
26
34
203
316
50
60
270
65
115
3.3
4.9
17.1
6.8
12.8
1950-
1959
10
26
39
153
353
50
60
146
49
174
3.3
4.9
9.2
5.4
19.3
1960-
1963
10
26
40
192
217
50
48
106
62
121
3.3
4.9
6.7
6.9
13.5
1964-
1967
10
26
38
193
323
50
55
107
69
179
3.3
4.9
6.8
7.7
19.9
1968-
1970
10
26
67
177
374
50
45
115
65
102
3.3
4.9
7.3
7.2
11.3
Low I is Bozeman; Low II, Helena; Low III, E. Helena; High I, Anaconda; and High II, Kellogg.
3-22
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
status as indicated by education (proportion of heads
of household completing high school) and crowding
(percent of households with more than one room per
person) reflected the industrial or nonindustrial
nature of the cities (Table 3.2.2). As expected, a
more educated population living in less crowded
housing was found in the Low pollution com-
munities.
Respondents were asked about exposure to coal
dust, cutting oils, asbestos, mine dust, smelter fumes,
or raw cotton dust through their occupation.
Exclusion of those who reported prolonged exposure
to these irritating dusts or fumes resulted in a loss of
approximately one-third of the males from most
subsequent analyses for the High pollution areas
(Table 3.2.2). Parents included in all analyses are
distributed in Table 3.2.3 by smoking status, sex, and
age group for individual communities and for the
pooled High and pooled Low pollution areas.
Chronic Respiratory Disease Evaluation
Chronic Bronchitis
Prevalence rates for chronic bronchitis are dis-
played by age and by educational attainment for
parents in each smoking category in Table 3.2.4. A
stronger age effect was seen in fathers than mothers,
Table 3.2.2. COMMUNITY CHARACTERISTICS: RESPONSE RATES, EDUCATIONAL STATUS
OF FATHERS, CROWDING INDEX, AND OCCUPATIONAL EXPOSURE TO IRRITATING SMOKE,
DUST, OR FUMES
Community
Pooled
Low
Low 1
Low II
Low III
Pooled
High
High 1
High It
Families
completing
questionnaire
1488
803
483
202
1626
846
780
Percent
response
84.9
85.0
84.7
84.9
83.5
83.7
83.3
Education
(percent of
heads of house-
hold completing
high school)
84.9
90.7
80.3
73.6
60.5
63.2
57.5
Crowding index
(percent of
households with
<1.0 person/
room)
81.4
87.3
76.5
70.9
67.9
67.7
71.9
Percent of
fathers excluded
because of
occupational
exposure3
2£
1.6
1.4
11.9
35.6
35.1
36.1
Three mothers from the Low exposure communities and 17 mothers from the High exposure communities were also excluded.
Table 3.2.3. NUMBER OF NONINDUSTRIALLY EXPOSED MOTHERS AND FATHERS
BY SMOKING STATUS, AGE, AND COMMUNITY
Communi ty
Pooled Low
Low I
Nonsmokers
Mothers
< 40 ! ^40
448 ; 293
270 170
Low II i 115 92
Low III
63 31
Pooled High ' 434 236
High I 1 193 144
High II I 241 . 92
Total
882 529
,
Fathers
< 40
197
145
36
16
124
59
65
321
140
204
128
51
25
78
43
35
282
Exsmokers
Mothers
< 40
137
82
37
18
143
70
73
280
140
87
49
29
9
71
42
29
158
Fathers
< 40
188
118
37
33
109
52
57
297
140
Smokers
Mothers 1 Fathers
40
227 342
126 147
76 122
25 73
119 470
75 222
140 < 40
176 253
71 118
83 87
22 48
221 304
154 132
44 248 67 172
346
812 I 397 i 557
140
228
91
100
37
206
118
88
434
Rocky Mountain Studies
3-23
-------
while the educational effect was inconsistent. A
maiked contrast between smokers and nonsmokers
was apparent, with exsmokers more closely
resembling nonsmokers. Smoking fathers had more
bronchitis than smoking mothers in all age groups.
For nonsmokers and exsmokers, this male excess was
only seen in the older age groups.
Smoking- and sex-specific rates for each com-
munity, given in Table 3.2.5, are quite variable due to
the relatively small numbers of cases in most cate-
gories. When pooled High and pooled Low com-
munities were compared, bronchitis prevalence was
higher within the High exposure community in five of
the six smoking- and sex-specific categories. Likewise,
sex- and age-adjusted rates within the pooled High
exposure population showed excess bronchitis in each
smoking category. Differences were statistically
significant only in the combined nonsmoking and
exsmoking population (Table 3.2.6); however, there
were approximately equal differences between High
and Low exposure areas in smokers of both sexes.
Sex-specific smoking intensity comparisons between
the High and Low exposure communities are shown
in Table 3.2.7. The similarity between High and Low
exposure communities and the absence of a pattern in
the slight differences that occur indicates that the
observed area differences among smokers cannot be
explained by smoking intensity.
Table 3.2.4. SMOKING- AND SEX-SPECIFIC PREVALENCE RATES (percent) FOR
CHRONIC BRONCHITIS BY EDUCATION AND AGE3
Category
Education:
High school
Age:
<29
30-39
40-49
>50
Nonsmokers
Mothers
2.06
2.20
1.36
1.09
1.31
2.63
2.61
Fathers
3.23
3.66
1.95
0.00
0.68
4.10
6.25
Exsmokers
Mothers
5.83
1.43
2.51
2.63
3.86
2.38
0.00
Fathers
4.81
3.49
2.13
0.00
2.72
3.24
5.06
Smokers
Mothers
14.50
11.75
10.85
13.07
11.17
14.95
7.69
Fathers
21.18
15.70
19.46
14.55
15.59
21.00
28.41
Chronic bronchitis rates are equivalent to crude rates for symptom severities 6 and 7.
Table 3.2.5. PREVALENCE OF CHRONIC BRONCHITIS IN IMONINDUSTRIALLY EXPOSED PARENTS:
INDIVIDUAL AND POOLED COMMUNITY RATES (percent) BY SEX AND SMOKING STATUS
Community
Pooled Low
Low I
Low II
Low III
Pooled High
High I
High II
Nonsmokers
Mothers
1.08
1.36
0.48
1.06
2.54
3.56
1.50
Fathers
1.25
1.10
0.00
4.88
3.47
4.90
2.00
Exsmokers
Mothers
3.12
3.05
4.55
0.00
2.80
1.79
3.92
Fathers
1.45
0.00
5.31
0.00
4.82
4.72
4.95
Smokers
Mothers 1 Fathers
11.78 , 17.05
8.72 12.44
14.15 ! 20.86
13.68 ' 20.00
12.88 18.63
13.83 18.40
11.75 i 18.85
I
Sex- and age-adjusted rates
Non-
smokers
1.08
3.12
' Ex-
smokers
2.46
3.56
Smokers
14.00
15.72
3-24
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.2.6. ANALYSIS OF VARIANCE FOR HEALTH OBSERVATIONS IN SMOKERS
AND NONSMOKERS, CHRONIC BRONCHITIS
Factor
Sex
Education
Age
Exposure
Fit of
model
Degrees
of
freedom
1
1
1
1
11
Smokers
X2
14.70
3.15
8.61
0.66
14.19
Probability (p)
<0.0003
<0.07
<0.004
<0.5
<0.2
Nonsmokers3
X2
1.12
0.48
10.10
8.73
13.85
Probability (p)
<0.3
<0.6
<0.002
<0.005
<0.2
Exsmokers and lifetime nonsmokers were combined for this analysis to obtain a larger sample size.
Table 3.2.7. SMOKING INTENSITY AMONG
SMOKERS IN POOLED HIGH AND LOW
EXPOSURE COMMUNITIES
Cigarettes
smoked
per day
1 to 5
6 to 15
16 to 25
>25
Percent smoking
Fathers
Low
9.3
15.0
48.5
27.3
High
8.8
16.7
52.9
21.6
Mothers
Low
12.8
18.4
51.4
17.4
High
9.1
19.0
52.7
19.2
Respiratory Symptom and Severity Scores
In each of the six smoking- and sex-specific
comparisons (Table 3.2.8), High exposure com-
munities had higher symptom scores. As indicated in
Table 3.2.9, area differences among nonsmokers were
highly significant (p < 0.002), and area differences
among smokers approached significance (p < 0.07).
Again, the pattern of differences was very consistent.
Symptom scores were also higher for males and for
smokers within each pooled exposure group.
Table 3.2.8. MEAN SYMPTOM SCORES IN IMONINDUSTRIALLY EXPOSED PARENTS: INDIVIDUAL
AND POOLED COMMUNITY AVERAGES BY SEX AND SMOKING STATUS
Nonsmokers
Community Mothers
Pooled Low 1.23
Low I 1.21
Low II 1.20
Low III 1.34
Pooled High 1.29
High I 1.31
High II ; 1.28
Fathers
1.33
1.27
1.32
1.71
1.38
1.46
1.30
Exsmokers
Mothers
1.30
1.20
1.49
1.23
1.45
1.44
1.46
Fathers
1.41
1.32
1.68
1.27
1.58
1.58
1.58
Smokers
Mothers
2.20
1.98
2.31
2.47
2.34
2.42
2.23
Fathers
2.65
2.53
2.72
2.75
2.72
2.87
2.58
Sex- and age-adjusted rates
Non-
smokers
1.27
1.34
Ex-
smokers Smokers
1.35 2.39
1.49 2.52
Table 3.2.9. ANALYSIS OF VARIANCE FOR HEALTH OBSERVATIONS IN SMOKERS
AND NONSMOKERS, MEAN RESPIRATORY SYMPTOM SCORES
Factor
Sex
Education
Age
Exposure
Fit of model
Degrees
of
freedom
1
1
1
1
11
Smokers
X2
32.39
3.00
3.02
3.32
17.53
Probability (p)
<0.0001
<0.08
<0.08
<0.07
<0.09
Nonsmokers3
X2
20.60
0.54
6.84
10.32
11.58
Probability (p)
<0.0001
<0.5
<0.009
<0.002
<0.04
Exsmokers and lifetime nonsmokers were combined for this analysis to obtain a larger sample size.
Rocky Mountain Studies
3-25
-------
Mean severity scores among parents who reported
some chronic respiratory symptoms did not differ
between High and Low exposure communities or
between sexes (Tables 3.2.10 and 3.2.11). Thus, there
was no evidence for overreporting or exaggeration of
minor symptoms among symptom-positive re-
spondents from High exposure communities. Lower
severity scores in High exposure communities would
be anticipated if overreporting of minor symptoms
occurred. Exsmokers and cigarette smokers had
higher severity scores than nonsmokers. These results
suggest that the severity of chronic respiratory
symptoms, once present, did not differ between
exposure groups but did differ between smokers and
nonsmokers.
Length of Residence
In order to assess the importance of duration of
exposure to community air pollution, chronic
bronchitis rates were computed according to the
length of residence in High or Low pollution areas.
Rates for smokers and current nonsmokers were
computed. Within each residence duration category,
the ratio of bronchitis in High exposure communities
to bronchitis in Low exposure communities was
calculated (Figure 3.2.1). Absolute bronchitis rates in
Low exposure communities are given in Table 3.2.12.
Recent immigrants (duration of residence of less than
2 years) who were nonsmokers or exsmokers had
more disease in all areas than residents of 2 or more
Table 3.2.10. MEAN SYMPTOM SEVERITY SCORES FOR IMOIMINDUSTRIALLY EXPOSED PARENTS WHO
REPORTED SYMPTOMS: INDIVIDUAL AND POOLED COMMUNITY AVERAGES
BY SEX AND SMOKING STATUS
I
! Nonsmokers
Community
Pooled Low
Low I
Low II
Low III
Pooled High
High I
High II
Mothers
4.27
4.24
4.38
4.20
3.98
4.39
3.62
Fathers
4.05
4.08
4.11
3.90
4.50
4.61
4.33
I
I
Exsmokers
Mothers [
4.73
5.33
5.00
3.00
4.31
4.07
4.62 !
i i
Fathers
4.11
3.89
4.50
3.67
4.38
4.36
4.41
Smokers
Mothers j Fathers
4.23 4.52
4.29
4.20
4.18
4.33
4.40
4.24
4.44
Sex- anc
Non-
smokers
4.13
4.58
4.55
4.48 4.24
4.52
4.43
age-adjust
~Ex- ' "
smokers
4.47
3d rates
Smokers
4.38
4.29 j 4.41
i
i
Table 3.2.11. ANALYSIS OF VARIANCE FOR HEALTH OBSERVATIONS IN SMOKERS
AND NONSMOKERS, SEVERITY SCORES FOR THOSE WITH SYMPTOMS
Factor
Sex
Education
Age
Exposure
Fit of model
Degrees
of
freedom
1
1
1
1
11
Smokers
X2
3.53
0.20
9.72
0.06
6.65
Probability (p)
<0.05
<0.8
<0.002
<0.09
<0.82
Nonsmokers3
X2
0.11
2.11
3.59
1.51
20.35
Probability (p)
<0.8
<0.14
<0.055
<0.2
<0.04
Exsmokers and lifetime nonsmokers were combined for this analysis to obtain a larger sample size.
3-26 HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
"0 1 ! 3 4 5 6 7 8 9 10 II 1! > 12
RESIDENCE DURATION, years
Figure 3.2.1. Risk of chronic bronchitis in
High exposure communities relative to base-
line experience of Low exposure communities.
years duration. In High exposure communities,
smoking and nonsmoking long-time residents (more
than 4 years) had more disease than residents of
intermediate duration. Residence duration-specific
bronchitis ratios were greater than unity among
nonsmoking and' smoking residents of intermediate
and long duration in High exposure communities. Of
particular importance was the excess bronchitis in
nonsmoking adults of 2 to 3 years residence duration
in High exposure communities. These data suggested
a relatively recent effect of exposure on chronic
disease prevalence. A trend towards increasing excess
bronchitis with longer residence in High exposure
communities was also apparent among nonsmokers
and smokers. Excess bronchitis among smokers in
High exposure areas was relatively less than among
nonsmokers, reflecting the larger effect of cigarette
smoking, compared with pollutant exposure, on
bronchitis prevalence.
Table 3.2.12. CHRONIC BRONCHITIS ATTACK
RATES (percent) FOR SMOKERS AND CURRENT
NONSMOKERS IN LOW EXPOSURE COMMUNITIES
BYJDURATJOJ\K)F RESIDENCE3
Residence duration, years
Smoking
status
Current
nonsmokers
Smokers
<1
2.4
161
>1 to<3
11 0
170
>3 to < 7
07
135
>7to<12
1.5
9.9
>12
1 6
14.4
Baseline rates for Figure 321.
Relative Importance of Community Air Pollution and
Cigarette Smoking
In order to distinguish the relative effects of
cigarette smoking and air pollution, chronic
bronchitis rates for those who had been exposed to
each of these factors alone were compared with rates
for those presumably exposed to neither factor
(Table 3.2.13). Prevalence rates were computed for
nonsmokers, exsmokers, and smokers residing in
polluted and in clean areas. Bronchitis rates of
nonsmokers from clean areas were used as a baseline.
Table 3.2.13. RELATIVE IMPORTANCE OF CIGARETTE SMOKING AND
AMBIENT AIR QUALITY AS DETERMINED BY COMPARISON OF
EXCESS PREVALENCE OF CHRONIC BRONCHITIS
1
Additive model for
Community
Sex and ' air quality
smoking status [ (pooled)
Female
Lifetime nonsmoker Clean
Exsmoker
Smoker
Male
Lifetime nonsmoker
Exsmoker
Smoker
Dirty
Clean
Dirty
Clean
Dirty
Clean
Dirty
Clean
Dirty
Clean
Dirty
Excess prevalence3
0.00
11081
1 46
2.04,
1 72
1070
11 80
000
(1 25)
2.22
020
357
15.80
17.38
smoking and pollution
Expected
-
-
_
350
_
12.16
-
-
_
242
_
1802
Observed/
expected ,
-
—
-
0.49
-
097
-
-
_
1 48
_
0.96
3Base rates in parentheses Excess prevalence = smpking- and community-specrfic prevalence rate minus base rate
Rocky Mountain Studies
3-27
-------
The isolated effect of pollution was computed by
the excess prevalence among nonsmokers of High
versus Low exposure communities. This difference
was 1.46 for females and 2.22 for males. The isolated
effect of smoking was computed as the difference in
prevalence between smokers and nonsmokers in Low
exposure communities. These differences were 10.70
for female smokers and 15.80 for male smokers.
Thus, among both females and males, the effect of air
pollution on bronchitis prevalence was 14 percent of
the effect of smoking. The effect of current air
pollution approached that of past cigarette smoking
(in exsmokers) on bronchitis prevalence among
females. In males, current air pollution had a stronger
effect than past cigarette smoking.
If the effects of air pollution and smoking were
additive, excess bronchitis among smokers in Low
exposure communities added to excess bronchitis
among nonsmokers in High exposure communities
should equal the excess bronchitis among smokers in
High exposure communities. In each case, excess
bronchitis must be derived with baseline rates of
nonsmokers in Low exposure communities. When
these calculations were made, 97 percent of the
excess bronchitis expected in High exposure areas
among smoking females and 96 percent among
smoking males could be accounted for by the simple
additive model. However, the additive model under-
estimated the combined effect of exsmoking and
pollution among females, .and overestimated this
combined effect among males.
Occupational Exposure
Smoking- and age-specific chronic bronchitis
prevalence rates, mean symptom scores, and severity
scores for males with and without occupational
exposure in High exposure communities were
computed (Table 3.2.14). For this analysis, occupa-
tional exposure was defined as: exposure on the job
for at least 5 years to either coal dust, cutting oils,
asbestos, mine dust, smelter fumes, or raw cotton
dust. There were 478 males with occupational
exposure and 940 without occupational exposure.
Consistently higher prevalence rates and symptom
scores were observed in occupationally exposed males
of each smoking-age category. Severity scores were
higher in occupationally exposed smokers but lower
in nonsmokers with occupational exposure. Results
indicate that bronchitis morbidity is increased in
smokers and nonsmokers by community and occupa-
tional exposures, and that exclusion of oc-
cupationally exposed males from previous analyses
yielded a conservative estimate of the bronchitis
morbidity effect of community air pollution.
Utilizing the model previously used to test for
additive effects of cigarette smoking and community
exposure, the effects of occupational exposure were
evaluated. Since occupational exposure in the Low
pollution communities was quite rare, its effect was
estimated as the difference between the prevalence in
occupationally exposed and nonoccupationally
exposed nonsmokers in the High pollution com-
munities. This difference was added to the previously
calculated excess prevalence for smoking status and
community exposure to derive an expected excess
risk for those individuals who had experienced
exposure from all three sources (Table 3.2.15). The
additive model predicted 82 and 88 percent of the
excess risks, again indicating that the effects of
occupational exposure, superimposed on community
exposure and cigarette smoking, appear to be ad-
ditive.
Table 3.2.14. BRONCHITIS PREVALENCE, SYMPTOM SCORES, AND SEVERITY SCORES
BY AGE IN MALES WITH AND WITHOUT OCCUPATIONAL EXPOSURES IN POOLED
HIGH EXPOSURE COMMUNITIES
Smoking
category
Nonsmokers
and
exsmokers
Smokers
Occupational
dust exposure
Yes
No
Yes
No
Prevalence, percent
<40
2.2
2.1
17.4
13.8
>40
7.0
6.6
27.3
25.7
Symptom
score
<40
1.62
1.33
2.89
2.50
>40
1.90
1.66
3.40
3.05
Severity
score
<40
3.80
4.39
4.49
4.29
>40
4.68
4.79
4.87
4.71
3-28
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.2.15. RELATIVE IMPORTANCE OF OCCUPATIONAL EXPOSURE, CIGARETTE SMOKING,
AND AMBIENT AIR QUALITY AS DETERMINED BY COMPARISON OF
EXCESS RISK OF CHRONIC BRONCHITIS
Smoking status
Lifetime nonsmoker
Exsmoker
Smoker
Community
air quality
(pooled)
Clean
Dirty
Dirty
Clean
Dirty
Dirty
Clean
Dirty
Dirty
Occupational
exposure
Unexposed
Unexposed
Exposed
Unexposed
Unexposed
Exposed
Unexposed
Unexposed
Exposed
Excess
prevalence
of chronic
bronchitis
0.00
(1.25)a
2.22
4.43
0.20
3.57
3.78
15.80
17.11
17.78
Additive model for
smoking, occupational
exposure, and pollution
Expected
4.63
20.23
Observed/
expected
0.82
0.88
Base rate in parentheses.
DISCUSSION
In two Rocky Mountain smelter communities
exposed to elevated levels of atmospheric sulfur
dioxide and suspended sulfates, but not to total
suspended particulates, the prevalence of chronic
bronchitis and overall mean respiratory symptom
scores were significantly increased in nonsmokers and
exsmokers by comparison with three Low exposure
communities in the same region. Excess bronchitis
morbidity occurred among nonsmokers, exsmokers,
and current smokers of both sexes in the High
exposure smelter communities.
Three hypotheses were proposed and supported by
this study: First, chronic bronchitis morbidity was
related to sulfur dioxide exposure. Second, the excess
of chronic bronchitis tended to increase with longer
duration of residence in High exposure communities.
Third, the combined deleterious effects of air pol-
lution and cigarette smoking were greater than those
of either factor alone.
The distribution of certain important covariates of
bronchitis prevalence differed between study com-
munities due to differences in principal occupations
among areas. A substantial portion of fathers in High
exposure communities were excluded from most
analyses because of occupational dust and fume
exposure, the effects of which could not be isolated
from those of community exposure. Further, in High
exposure communities, educational attainment was
below that of Low exposure areas. However, no
consistent or statistically significant association be-
tween educational attainment and bronchitis
prevalence was found in the five study communities.
Differences in age distribution occurred but did not
correspond to exposure contrasts between com-
munities. Bronchitis prevalence increased with age
among males in all smoking categories, but not among
females. All bronchitis morbidity indices were there-
fore age adjusted. This adjustment compensated in
part for area differences in educational achievement,
since the more educated communities were generally
younger. As anticipated, a strong and statistically
significant male excess in bronchitis morbidity was
present with each smoking category. Analyses of
age-adjusted bronchitis rates were performed for sex-
and smoking-specific groups or were adjusted for
both age and sex within each smoking category.
Smoking categories in High and Low communities
were very similar with respect to the distribution of
smoking intensity. Thus, differences between smokers
in High and Low communities could not be at-
tributed to area differences in smoking intensity.
Rocky Mountain Studies
3-29
-------
The mean respiratory symptom score, based on a
seven-point ordinal scale of chronic respiratory
symptom reporting, proved to be a useful index. The
symptom score avoided making a simple dichotomy
of symptom responses into disease or no disease.
Early or less prolonged chronic respiratory symptoms
were given greater weight than complete absence of
symptoms. The symptom score snowed a more
consistent response gradient between communities
within each smoking-sex category than was true for
bronchitis prevalence rates (compare Tables 3.2.5 and
3.2.8). The consistency may be accounted for by the
fact that the spectrum of morbidity is broader with
the symptom score and that when sample sizes were
small,- the symptom score was less variable in the
presence of a low morbidity prevalence. The
symptom score strongly confirmed results indicated
by prevalence rates. Symptom scores have the dis-
advantage that relative and excess rates cannot be
computed and comparison with past studies cannot
be made.
Severity scores in the population reporting any
respiratory symptoms showed that disease, though
present in excess, was not more severe in the High
exposure community. In addition, the severity score
was useful in eliminating the possibility that over-
reporting bias could explain observed community
differences in morbidity. The severity score gave no
evidence that less severe or minor respiratory
symptoms were exaggerated or overreported in any of
the sex- and smoking-specific categories within High
exposure communities. If less severe symptoms had
been overreported, lower severity scores would have
been anticipated from High exposure areas. This
clearly was not so.
Analysis of bronchitis morbidity by duration of
residence proved to be very important. The associa-
tion of excess disease with recent change of
residence34"39 was confirmed in this study and
demonstrated the need to analyze separately those
who moved within the past year. In High exposure
areas, a U-shaped disease distribution was observed
over a residence duration scale. High morbidity
occurred in most recent immigrants and in long-time
residents. When most recent immigrants" were
excluded, the risk of bronchitis in High relative to
Low exposure communities increased with longer
residence time. This pattern was more consistent for
smokers than nonsmokers. However, nonsmokers
present in High exposure areas for only 2 to 3 years
had 2.6-fold more bronchitis than similar residents of
Low exposure areas. Smokers of 2 to 3 years
residence duration did not as yet have bronchitis in
excess by comparison with smokers in Low exposure
areas. These observations suggested that nonsmokers
may respond more rapidly to a new but polluted
environment. Such effects might be anticipated, given
the relatively greater chronic insult of smoking
compared with air pollution. An additional contribu-
tion to the relative risk below unity among smokers
who resided in High exposure areas less than 4 years
was the possibility that smokers with respiratory
symptoms would be disinclined to move into a
polluted environment.
A further important conclusion for air quality
standards could be derived from the residence dura-
tion analysis. To relate chronic disease to environ-
mental exposure, environmental standards require
intelligence concerning length of exposure as well as
concentration. Oftentimes, only current air quality
data are available. Assumptions are made that excess
chronic disease, requiring cumulative exposure, was
caused by long-term exposures similar to current
levels. This assumption cannot usually be supported,
especially in recent years when vigorous efforts at air
pollution abatement have been underway. Attributing
chronic disease to recent exposure could result in
unduly restrictive standards if the disease resulted
from past high exposures. This challenges the environ-
mental health scientist to define the time-weighted
concentration likely to cause excess chronic disease, a
task that is formidable and usually beyond current
knowledge. However, documentation of excess
bronchitis in 2- to 3-year residents of High exposure
communities provides evidence that current levels,
which can be readily quantified, can be associated
with excess disease. This association is valid if the 2-
to 3-year residents moved from a low exposure area,
an assumption that unfortunately could not be
validated because data on location of prior residence
were not obtained. Furthermore, the pattern of
increasing excess illness with longer length of
residence in High exposure areas is consistent with
the above assumptions. Therefore, the finding of
excess bronchitis in 2- to 3-year residents who were
nonsmokers provides evidence that exposure to sulfur
dioxide in the range of 177 to 374 Mg/m3 and
suspended sulfates of 7.2 to 19.9 jug/m3 for 2 to 3
years can result in excess chronic bronchitis; and for
smokers, the corresponding period is 4 to 7 years.
These exposures occurred in the presence of only
moderately high total suspended particulates, 65 to
179 Mg/m3. Since High exposure areas were non-
ferrous smelter communities, the possibility that
metallic particulates potentiated observed effects is
distinctly present, especially in view of experimental
findings that metallic sulfates have greater adverse
effects on the respiratory tract than does sulfur
dioxide.40'41
3-30
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Comparisons of prevalence rates for bronchitis
among nonsmokers and smokers in Low and High
exposure communities revealed a smoking effect
about 7 times greater than air pollution in this
population. Simple addition of the effects of these
two variables explained remarkably well the observed
excess bronchitis among smokers living in High
exposure communities. That is, excess bronchitis due
to community air pollution was of the same magni-
tude in smokers and nonsmokers.
Occupationally exposed males had excess bron-
chitis over and above that attributable to community
air pollution. Again, an additive model using com-
munity air pollution, occupational exposure, and
cigarette smoking as factors was sufficient to reason-
ably explain the excess bronchitis among smoking,
Occupationally exposed males in High exposure com-
munities. To protect the health of workers living in
High exposure communities, occupational health
standards should be coupled with ambient air quality
standards. High community exposure forces industry
to hire more vulnerable populations who can ill
afford any added pollutant exposure. Thus, when
starting with a higher baseline of illness, protection of
employee health requires simultaneous reduction of
community and occupational hazards.
One of the Low exposure communities (Low III)
was a smelter town. Residential areas of this town
were situated upwind of the smelter in a valley where
prevailing winds minimized the entire impact of the
smelter emissions. However, total suspended particu-
lates (106 to 270 Mg/m3) and suspended sulfate
concentrations (6.7 to 17.1 Mg/m3) were higher than
levels in the other two Low exposure communities
(45 to 60 Mg/m3 for total suspended particulate; 2.9
to 3.8 Mg/m3 for suspended sulfates). Sulfur dioxide
was low in all Low exposure communities (10 to 67
jug/m3). Increased exposure to particulates in the
Low exposure smelter community (Low III) appeared
to exert an adverse effect on bronchitis morbidity, as
evidenced by higher mean respiratory symptom
scores (Table 3.2.8) among nonsmokers and smokers
of both sexes when compared with comparable
groups in the other two Low exposure areas. The
total number of respondents from this smelter com-
munity was too small to draw firm conclusions from
these findings. Small sample sizes .were reflected in
the instability of bronchitis prevalence rates within
smoking- and sex-specific groups for each community
(Table 3.2.5). Inclusion of this smelter community in
the pooled Low exposure group tended to raise
baseline morbidity rates and thus to diminish the
magnitude of excess bronchitis found in High
exposure communities.
SUMMARY
Chronic bronchitis prevalence and symptom scores
in nonoccupationally exposed adult residents of two
High exposure Rocky Mountain smelter communities
significantly exceeded morbidity rates found in
residents of three Low exposure communities. 'Age-
adjusted excess bronchitis was observed within sex
and smoking categories. Nonsmokers in High
exposure communities had excess bronchitis
prevalence rates 2.4 to 2.8 times those of nonsmokers
in Low exposure areas. The effect of community
exposure was additive to that of cigarette smoking,
which exerted an effect on chronic bronchitis ap-
proximately 7 times stronger than that of air pollu-
tion. Symptom scores confirmed area differences
shown from bronchitis prevalence rates. Though more
prevalent, bronchitis was not more severe in High
exposure communities. Comparison of residence
duration-specific rates demonstrated that prolonged
residence in High exposure areas magnified the excess
bronchitis attributable to pollutant exposure. Occu-
pational exposure further increased observed bron-
chitis morbidity. Excess bronchitis occurred with 2-
to 3-year exposure to sulfur dioxide concentrations
of 177 to 374 £tg/m3 and suspended sulfate concen-
trations of 7.2 to 19.9 /ug/m3 in the presence of low
total suspended particulates. Metallic sulfates may
well have accounted for the findings of excess
bronchitis.
REFERENCES FOR SECTION 3.2
1. Firket. J. Fog along the Meuse Valley. Trans.
Faraday Soc. 32:1192-1197,1936.
2. Firket, J. The Cause of the Symptoms Found in
the Meuse Valley during the Fog of December
1930. Bull. Roy. Acad. Med. Belgium.
77:683-739,1931.
3. Schrenk, H. H., H. Heimann, G. 0. Clayton,
W. M. Gafafer, and H. Wexler. Air Pollution in
Donora, Pennsylvania; Epidemiology of the Un-
usual Smog Episode of October 1948. Federal
Security Agency, Division of Industrial Hygiene,
Public Health Service, U. S. Department of
Health Education, and Welfare. Washington,
D.C. Pubh'c Health Bulletin 306. 1949. p. 173.
Rocky Mountain Studies
3-31
-------
4. Gore, A. T. and C.W. Shaddick. Atmospheric
PollutI .,•! and Mortality in the County of
London. Brit. J. Prev. Soc. Med. 72:104-113,
1958.
15. Holland, W. W., D. D. Reid, R. Seltser, and R. W.
Stone. Respiratory Disease in England and the
United States; Studies of Comparative Preva-
lence. Arch. Environ. Health_70:338-345,1965.
5. Burgess, S. G. and C. W. Shaddick. Brochitis and
Air Pollution. Roy. Soc. Health J. 79:10-24,
1959.
16. Petrilli, R. L., G. Agnese, and S. Kanitz. Epidemi-
ology Studies of Air Pollution Effects in Genoa,
Italy. Arch. Environ. Health. 72:733-740, 1966.
6. Scott, J. A. The London Fog of December 1962.
Med. Officer. 709:250-252,1963.
7. Greenburg, L., M. B. Jacobs, B. M. Drolette, F.
Field, and M. M. Braverman. Report of an Air
Pollution Incident in New York City, November
1953. Public Health Reports. 77:7-16,1962.
8. McCarroll, J. andW. Bradley. Excess Mortality as
an Indicator of Health Effects of Air Pollution.
Amer. J. Public Health. 55:1933-1942,1966.
9. Glasser, M., L. Greenburg, and F. Field. Mortal-
ity and Morbidity during a Period of High Levels
of Air Pollution, New York, November 23-25,
1966. Arch. Environment. Health. 75:684-694,
1967.
10. Logan, W. P. D. Mortality in the London Fog
Incident 1952. Lancet. 264:336-338, 1953.
11. Greenburg, L., F. Field, J.I. Reed, and C. L.
Erhardt. Air Pollution and Morbidity in New
York City. J. Amer. Med. Assoc. 752:161-164,
1962.
12. McCarroll, J. R., E. J. Cassell, E. W. Walter, J. D.
Mountain, J. R. Diamond, and I. R. Mountain.
Health and the Urban Environment; V. Air
Pollution and Illness in a Normal Urban Popula-
tion. Arch. Environ. Health. 74:178-184, 1967.
13. Bell, A. The Effects on the Health of Residents
of East Port Kembla, Part II. In: Air Pollution by
Metallurgical Industries (Vol. 2). Sidney,
Australia, Division of Occupational Health, New
South Wales Department of Public Health, 1962.
p. 1-144.
14. Holland, W. W. and R. W. Stone. Respiratory
Disorders in United States East Coast Telephone
Men. Amer. J. Epidemiol. 52:92-101, 1965.
17. Toyama, T. Air Pollution and Its Health Effects
in Japan. Arch. Environ. Health- 5:153-173,
1964.
18. Lawther, P. J. Climate, Air Pollution, and
Chronic Bronchitis. Proc. Roy. Soc. Med.
57:262-264,1958.
19. Carnow, B. W., M. H. Lepper, R. B. Shekelle, and
J. Stamler. The Chicago Air Pollution Study:
Acute Illness and S02 Levels in Patients with
Chronic Bronchopulmonary Disease. Arch.
Environ. Health. 75:768-776. 1969.
20. Lawther, P. J., R. E. Waller, and M. Henderson.
Air Pollution and Exacerbations of Bronchitis.
Thorax. 25:525-539,1970.
21. American Smelting and Refining Company. Salt
Lake City, Utah. Written Comments to Proposed
National Ambient Air Quality Standards for
Sulfur Oxides, Federal Register, 36:1502,
January 30, 1971. Submitted to National Air
Pollution Control Administration, Washington
D.C., March 15,1971.
22. Helena Valley, Montana, Area Environmental
Pollution Study. U. S. Environmental Protection
Agency. Research Triangle Park, N. C. Office of
Air Programs Publication No. AP-91. January
1972. 193 p.
23. Idaho Air Quality—Methods of Measurement and
Analysis of Recent Data. Idaho State Depart-
ment of Health. Boise, Idaho. August 19, 1970.
23 p.
24. Eiguren, A. J. Personal communication to D. I.
Hammer, U.S. Environmental Protection
Agency, Research Triangle Park, N. C. Idaho
State Department of Health, Air Pollution Con-
trol Commission. Boise, Idaho. October 5, 1971.
3-32
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
25. Michael, M. Personal communication to D. I.
Hammer. U.S. Environmental Protection Agency,
Research Triangle Park, N.C. Idaho State
Department of Health, Environmental Improve-
ment Division. Boise, Idaho. April 3, 1972.
26. A Study of Air Pollution in the Deer Lodge
Valley-August 1965-June 1966. Montana State
Department of Health. Helena, Montana. 32 p.
27. A Study of Air Pollution in the Helena-East
Helena-Area-October 1965-October 1968.
Montana State Department of Health. Helena,
Montana. 35 p.
28. A Study of Air Pollution in Montana, July
1961-July 1962. Montana State Board of Health.
Helena, Montana. 107 p.
29. A Study of Air Pollution in Townsend-Three
Forks, the Gallatin Valley and West Yellowstone,
November 1967-November 1968. Montana State
Board of Health. Helena, Montana. 43 p.
30. English, T. D., J. M. Sune, D. I. Hammer, L. A.
Truppi, W. E. Culver, R. C. Dickerson, and W. B.
Riggan. Human Exposure to Air Pollutants in
Five Rocky Mountain Communities, 1940-1970.
In: Health Consequences of Sulfur Oxides: A
Report from CHESS, 1970-1971. U.S. Environ-
mental Protection Agency. Research Triangle
Park, N.C. Publication No. EPA-650/1-74-004.
1974.
33. Grizzle, J. E., C. F. Starmer, and G. G. Koch.
Analysis of Categorical Data by Linear Models.
Biometrics. 25(3):489-504, September 1969.
34. Anderson, D. O. and B. G. Ferris, Jr. Community
Studies of the Health Effects of Air Pollution—a
Critique. J. AirPollut. Contr. Assoc. 75:587-593,
December 1965.
35. Eascott, D. F. The Epidemiology of Lung Cancer
in New Zealand. Lancet. 1:37-39, 1956.
36. Dean, G. Lung Cancer in South Africans and
British Immigrants. Proc. Roy. Soc. Med.
57:984-987,1964.
37. Haenszel, W. Cancer Mortality among the
Foreign-Born in the United States. J. Nat. Cancer
Inst. 26:37-132, 1961.
38. Mancuso, T. F. and E. J. Coulter. Cancer
Mortality among Native White, Foreign-Born
White, and Non-White Male Residents of Ohio:
Cancer of the Lung, Larynx, Bladder, and
Central Nervous System. J. Nat. Cancer Inst.
20:79-105, 1958.
39. Buechley, R., J. E. Dunn, Jr., G. Linden, and L.
Breslow. Excess Lung Cancer Mortality Rates
among Mexican Women in California. Cancer.
10:63-66, 1957.
31. Questionnaires Used in the CHESS Studies. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
32. Standards for Epidemiologic Surveys in Chronic
Respiratory Disease. National Tuberculosis and
Respiratory Disease Association. New York,
N. Y. 1969. 36 p.
40. Ambdur, M. 0. and M. Corn. The Irritant
Potency of Zinc Ammonium Sulfate of Different
Particule Sizes. Amer. Ind. Hyg. Assoc. J.
24:326-333, July-August 1963.
41. Ambdur, M. O. and D. Underbill. The Effect of
Various Aerosols on the Response of Guinea Pigs
to Sulfur Dioxide. Arch. Environ. Health.
76:460-468, 1968.
Rocky Mountain Studies
3-33
-------
3.3 FREQUENCY OF ACUTE LOWER RESPIRATORY
DISEASE IN CHILDREN: RETROSPECTIVE SURVEY
OF FIVE ROCKY MOUNTAIN COMMUNITIES, 1967-1970
John F. Finklea, M.D., Dr. P.H.,
Douglas I. Hammer, M.D., M.P.H., Dennis E. House, M.S.,
Charles R. Sharp, M.D., M.P.H., William C. Nelson, Ph.D.,
and Gene R. Lowrimore, Ph.D.
3-35
-------
INTRODUCTION
The purpose of this study was to determine if
long-term or short-term exposures to air pollutants
will aggravate or induce acute respiratory illnesses in
children. Acute respiratory diseases are the most
common cause of illness in children, and the lower
respiratory component may very well be a portent of
chronic respiratory disease in adults.1'3 Unlike
adults, elementary school children are generally not
subjected to occupational pollution exposures or
self-pollution by cigarette smoking. Children are also
less likely to have experienced a variety of com-
plicated long-term ambient air pollution exposures
related to residential mobility. Hence, they are an
excellent group in which to isolate the adverse health
effects induced by community air pollution.
in studies that quantify the relationship between
ambient air pollution and acute respiratory illness.
High levels of sulfur dioxide often occur in
western smelting communities where the total sus-
pended particulate levels are often quite low. Many of
these towns are small and have relatively little other
pollution from automobiles or fossil fuels. Production
and emissions data for the past years are usually
available. Since smelters are point sources of pol-
lutants, one can model previous community expo-
sures to sulfur dioxide and particulates for several
decades. We describe here a retrospective survey of
the frequency of acute lower respiratory illnesses
among elementary school children in five Rocky
Mountain communities.
Several foreign studies have implicated sulfur
dioxide and particulates in ambient air as a cause of
increased respiratory symptoms and illnesses in
children. In England, different studies of children
have shown an increased frequency and severity of
acute upper and lower respiratory morbidity and a
decrease in pulmonary function among exposed
children.4"6 In Japan, Toyama found that 10- and
11-year old school children living in a highly polluted
community had a higher frequency of nonproductive
cough, upper respiratory tract irritation, and in-
creased mucous secretion than comparable children in
a less polluted area.7 Manzhenko reported similar
findings in exposed Russian children as well as a high
prevalence of abnormal pulmonary X-ray findings.8
None of these studies were designed to distinguish the
effects of individual air pollutants such as sulfur
dioxide, total suspended particulates, and suspended
sulfates from the more complex urban mixtures. In
addition, the lifetime pollution exposure doses of the
children could not be calculated and the duration of
exposure necessary to induce effects was not usually
described. However, two studies of families living
near a point source of nitrogen dioxide were able to
link significant increases in respiratory illness to
nitrogen dioxide exposures lasting 3 or more
years
9,10
Respiratory morbidity in children is influenced by
age, sex, and socioeconomic status. Duration of
residence may be an important determinant of disease
because of community pollution exposure and be-
cause individuals moving into a new community may
initially suffer higher rates of illness.11 Children with
a history of asthma do have a higher frequency of
lower respiratory illness than do nonasthmatics.6
These personal covariates must be isolated or adjusted
MATERIALS AND METHODS
Community Selection
Five communities were selected on the basis of air
quality data published by the States of Idaho and
Montana and advice from health department officials
in both states.12"16 This study was a sequel to the
Helena Valley Area Environmental Report.17 Com-
munities were ranked Low or High on the basis of
their estimated exposure to sulfur dioxide during the
period covered by the study and for the prior decade.
Bozeman, the Low I community, is a prosperous,
growing, farming and college community of 18,500
located about 100 miles southeast of Helena. Helena,
the Low II community, is the state capitol of
Montana, has a population of 23,000 and is located in
a broad, usually well ventilated valley. East Helena,
the Low III community, is a small town of 1700 just
3 miles from Helena and has been the site of a lead
smelter since 1888, a zinc plant since 1927, and a
paint pigment plant since 1955. Virtually all East
Helena residents do their major shopping in Helena,
and some of them work in Helena. Anaconda, the
High I community, with a population of about 9800,
is located about 90 miles southwest of Helena and has
been the site of a large copper smelter since the end
of the last century. Kellogg, the High II community,
with a population of 3800, has been the site of a large
lead smelter for over 75 years and is in Shoshone
County, Idaho, about 252 miles northwest of Helena.
It is located at the bottom of a narrow valley in the
world famous Couer d'Alene mining district in the
Idaho panhandle. Both Anaconda and Kellogg
reported population losses in the 1970 census.
3-36
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Collection of Health and Demographic Data
Elementary school children (grades 1 to 6) in each
city were asked in November 1970 to take home an
explanatory letter and questionnaire to their parents.
All elementary schools in Kellogg, East Helena, and
Anaconda were included in the survey. Socio-
economically similar schools in the larger cities,
Helena and Bozeman, were selected after consultation
with the Superintendent of Schools and school
principals. The frequency of lower respiratory disease
in children was ascertained by means of a School and
Family Health questionnaire,18 which was completed
by the mother in each family and returned to the
school. The mother was instructed to give respiratory
disease information on all children in the family 12
years of age or younger.
The questionnaire inquired about the frequency of
treatment by a physician for pneumonia, croup, or
bronchitis (including bronchiolitis or deep chest
infections other than pneumonia or croup) during the
3-year period beginning in September 1967. Hence,
all rates used in the analyses (except in the three
figures) are 3-year rates. Other information as-
certained included hospitalizations for lower respira-
tory illnesses, name of children's physician, history of
asthma, length of residence in the community,
socioeconomic status measured by education of head
of household, parents' smoking status, and family
census. Physician records were reviewed to validate
parental reporting for approximately a 15 percent
sample of questionnaires. The sample included an
equal number of children reported sick and well.
Additionally, a sample of 18 physicians in the study
areas were queried about their definition of the
bronchitis syndrome.
Assessing Air Pollution Exposures
Annual estimated averages of sulfur dioxide, total
suspended particulates, and suspended sulfates were
derived from production figures and emissions
estimates. A detailed report of these procedures is
given elsewhere.19
Hypotheses Testing
The hypotheses tested were: (1) that the fre-
quency of physician treatment for any lower respira-
tory illness, croup, bronchitis, and pneumonia, as well
as the frequency of hospitalization for these illnesses,
adjusted for appropriate covariates, would correspond
to pollutant exposures, (2) that appropriately
adjusted acute lower respiratory illness rates would be
more elevated in children exposed to pollution for
several years than in newly arrived residents, and (3)
that the accuracy of illness reporting by parents
would be similar for all communities.
For each morbidity condition, two analyses were
done. In the first, the dependent variable was the
percentage of children reporting one or more
episodes. For the second, the dependent variable was
the percentage of children reporting two or more
episodes.
Hypotheses were tested by a general linear model
for categorical data.20 This technique utilizes
weighted regression on categorical data and allows
estimation of each effect in the model adjusted for all
other effects in the model. Significance was as-
certained by Chi Square procedures. Covariates used
in the model were age (three categories: 1 to 4, 5 to
8, 9 to 12), sex, and educational attainment (two
categories: < completed high school, > completed
high school), which was an index of socioeconomic
status (SES).
RESULTS
Exposure to Air Pollutants
The estimated annual pollutant levels for each
community for the study period and the decade
preceding showed a distinct gradient for sulfur
dioxide exposure (Table 3.3.1). Estimated concentra-
tions in the Low I and Low II communities (10 to 26
Atg/m3) were well within the relevant national
ambient air quality standards. Estimated total sus-
pended particulate levels (45 to 55 Mg/m3) and
suspended sulfates (2.9 to 3.3 jug/m3) in these two
communities were also low. Based on the estimated
annual pollutant levels, it appears that air quality in
these two communities had varied little during the
lifetime of the study population.
The Low III community differed from the other
Low exposure communities in that estimated annual
average suspended particulates (99 to 115 Mg/m3)
were elevated above the National Primary Air Quality
Standard, while estimated suspended sulfates, for
which there is no air quality standard, were also
somewhat elevated (6.2 to 7.3 Mg/m3). Moreover,
because of industrial emissions, the Low III com-
Rocky Mountain Studies
3-37
-------
Table 3.3.1. ESTIMATED ANNUAL SULFUR DIOXIDE, TOTAL SUSPENDED PARTICULATE,
AND SUSPENDED SULFATE LEVELS IN FIVE ROCKY MOUNTAIN COMMUNITIES
Community
Low 1
Low II
Low III
Highl
High II
Annual pollutant concentrations, p.g/m
Sulfur dioxide
During
study
(1968-70)
10
26
67
177
374
Decade
prior to
study
(1958-67)3
10
26
37
187
270
Total suspended
particulate
During
study
(1968-70)
50
45
115
65
102
Decade
prior to
study
(1 958-67) a
50
55
99
63
146
Suspended sulfate
During
study
(1968-70)
3.3
4.9
7.3
7.2
11.3
Decade
prior to
study
(1958-67)8
3.3
4.9
6.2
7.0
16.2
aBased on production figures and emissions estimates.
munity probably experienced peak short-term
exposures of sulfur dioxide that were considerably
higher and more frequent than the two other Low
exposure communities. The somewhat higher
estimated annual average sulfur levels in the Low III
community (37 to 67 Mg/m3) probably reflected
these short-term peak exposures. Ideally, the Low III
community would be approached separately in an
attempt to isolate the effects of short-term peak
exposures to sulfur dioxide and moderately elevated
suspended particulates. Unfortunately, the small size
of the Low III community prevented such an
analysis. On the basis of estimated pollutant
exposures, it seems that pooling the Low III
community with the two cleaner Low exposure
communities is a conservative procedure that would
tend to diminish any differences attributable to air
pollutants.
It is estimated that the two High exposure
communities were both exposed to annual average
sulfur dioxide levels (177 to 374 Mg/m3) well above
the relevant air quality standards. The estimated
annual average sulfur dioxide levels In the High I
community were consistently elevated (177 to 187
jug/m3) during the entire lifetimes of the children
surveyed. Estimated sulfur dioxide levels in the High
II community were higher (270 Mg/m3) during the
infancy and early childhood of the study population,
and appeared to increase even more (374 jug/m3)
during the study. On the other hand, estimated
annual average suspended sulfate levels (11.3 to 16.2
Mg/m3) and total suspended particulate levels (102 to
146 Mg/m3) in the High II community were distinctly
higher than levels in the High I community where
total suspended particulates (63 to 65 MgM3) were
well within the National Ambient Air Quality Stand-
ard and suspended sulfates were less drastically
elevated (7.0 to 7.2 fjg/m3). In none of the com-
munities would one expect to find significant eleva-
tions of nitrogen dioxide, suspended nitrates, carbon
monoxide, or gaseous hydrocarbons.
Community Characteristics
Intercommunity differences in reported illness
might be influenced by large intercommunity dif-
ferences in questionnaire completion rates, socio-
economic status, crowding, indoor pollution by
cigarette smoking, and exclusions of respondents
prior to analysis (Table 3.3.2).
Families from the pooled High and pooled Low
exposure communities completed and returned the
study questionnaires equally well, and no large
3-38
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.3.2. COMMUNITY CHARACTERISTICS: QUESTIONNAIRE COMPLETION RATES,
EDUCATIONAL ATTAINMENT OF HEAD OF HOUSEHOLD, CROWDING, AND SMOKING HABITS
Community
Pooled Low
Low 1
Low II
Low III
Pooled High
Highl
High II
Percent of
families
completing
questionnaire
84.9
85.0
84.7
84.9
83.5
83.7
83.3
Percent of
household heads who
completed high school
85
91
80
74
61
63
58
Percent of
households with
— 1 person
per room
81
87
77
71
70
68
72
Percent of
parents who
are current
cigarette smokers
Mothers
33
26
39
41
42
44
40
Fathers
37
29
48
46
54
52
56
community differences were noted within the two
pollution strata. Significant differences were noted,
however, in socioeconomic status as indexed by
educational attainment of the head of household. The
two High pollution communities were definitely less
well educated than the Low exposure communities (p
< 0.001). Household crowding, an index of family
size and economic level, was also significantly more
frequent in the High exposure communities (p <
0.001). Both mothers and fathers from the High
exposure communities were also more likely to be
cigarette smokers than mothers and fathers from the
Low exposure communities (p < 0.001). Fortunately,
there were no intercommunity differences in the
intensity of cigarette smoking among either mothers
(p > 0.20) or fathers (p > 0.10) who had the habit.
What might be the effect of the differences
between the characteristics of High exposure and
Low exposure communities? Other studies show that
higher socioeconomic status is linked with increased
frequency of croup and bronchitis coupled with a
decreased frequency of pneumonia. These studies
inconsistently associated heavy cigarette smoking by
parents with increased bronchitis in their children.
Paradoxically, children of light smokers were found
to have less croup than children of nonsmokers or
heavy smokers who reported similar rates for croup.
Clearly, social status and cigarette smoking might bias
the present study in a complex fashion. Since
cigarette smoking and lower educational attainment
are inextricably linked, inclusion of both these
variables in an adjustment procedure might cause
over-adjustment of illness rates. Furthermore, the
sample size and the statistical analysis limited the
possible number of adjustments, and a number of
other disease determinants, including age, sex, and
residential mobility, clearly required it. Preliminary
analysis indicated that consistent overall illness dif-
ferences within pollution strata were attributable to
socioeconomic differences. Thus, socioeconomic ad-
justment was clearly necessary. On the other hand,
intercommunity differences in smoking habits were
less striking, and the expected effects of smoking on
the clinical syndromes that comprise the lower
respiratory illness grouping were to some extent
counterbalancing. Croup might tend to be modestly
increased in the Low exposure communities because
of the modest excess of current nonsmokers there.
On the other hand, the same factor would tend to
slightly accentuate bronchitis rates in the High
exposure communities. Therefore, no adjustments
were made for differences that might be attributable
to cigarette smoking.
As specified in the study protocol, children of the
families who completed the health questionnaire were
excluded from the study because of missing question-
naire data or presence of active asthma requiring
treatment by physician during the study period.
Rocky Mountain Studies
3-39
-------
One-tenth of the respondents were excluded because
of incomplete or uninterp re table questionnaire
responses, and there were no remarkable differences
between the High and the Low exposure communities
(Table 3.3.3). Almost half of the exclusions were
caused by missing data on age and by exclusion of all
children under 1 year of age.
Acute Lower Respiratory Disease Evaluation
Asthmatics
Approximately 5 percent of the respondent
population were asthmatics, and these were excluded
from the principal analysis. The number of asthmatics
proved too small to allow adequate adjustments for
other strong determinants of acute lower respiratory
disease (Table 3.3.4). Nevertheless, parents living in
the High exposure communities consistently reported
significantly more total acute lower respiratory
illness, croup, and bronchitis among their asthmatic
children than did parents from the Low exposure
communities. Clinically, the greatest proportionate
morbidity excess was among children having the
croup syndrome. The frequency of croup and that of
hospitalizations exactly paralleled the long-term
exposure gradients for sulfur dioxide and suspended
particulate sulfates.
Increased croup and hospitalization rates were
apparent in the Low HI exposure community, but
with the small sample size, this excess was not
statistically significant. Since asthmatics from both
High exposure communities and the Low III com-
munity also more frequently reported hospitaliza-
tions, they were combined for testing, and excess
from these three communities approached statistical
significance. The described excesses for croup and
hospitalizations could be attributed to repeated short-
term peak sulfur dioxide exposures, to peak sus-
pended sulfate exposures, or to longer, lower level
exposures of these pollutants. It seems reasonable
that exclusion of asthmatics from the principal
analysis would lead to a conservative estimate of
intercommunity differences attributable to ambient
air pollution.
Total Acute Respiratory Illness
After all exclusions, the study population of 5773
was distributed by length of residence so that one
could distinguish the exposure duration necessary for
the emergence of any respiratory morbidity ac-
companying high levels of ambient air pollutants
(Table 3.3.5). This procedure also isolated any
increases in illnesses that may have been associated
Table 3.3.3. COMMUNITY CHARACTERISTICS: REASONS FOR
EXCLUSION OF NONASTHMATIC CHILDREN
Community
Pooled Low
Low 1
Low II
Low III
Pooled High
High 1
High II
Total
Percent of nonasthmatic
children excluded
Because of missing
data on education,
sex, or illness
5^
4
6
4
4.
3
6
5
Because of missing
data on age
6
8
4
5
5
4
6
5
Total
11
12
10
9
9
7
12
10
3-40
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.3.4. PERCENT OF ASTHMATIC CHILDREN REPORTING AT LEAST ONE LOWER
RESPIRATORY ILLNESS DISTRIBUTED BY COMMUNITY AND CLINICAL DIAGNOSIS
Community
Pooled Low
Low 1
Low II
Low III
Pooled High
Highl
High II
Number
of
children
149
75
43
31
141
54
87
Percent reporting lower respiratory illness8
Any lower
respiratory
disease
52
52
51
52
65
65
64
Croup
14
11
14
23
23
26
21
Bronchitis
44
48
47
42
60
61
59
Pneumonia
17
12
23
23
22
24
21
Hospitalization
li
15
21
26
25
28
23
aThe cluster of High exposure communities reported significantly more total lower respiratory illness (p < 0.05), croup
(p < 0.05), and bronchitis (p < 0.01). If the Low III community were included in the High exposure cluster, hospitalization
increases in that cluster would be nearly significant (0.10 < p < 0.05).
Table 3.3.5. NUMBER OF NONASTHMATIC CHILDREN USED IN ALL ILLNESS ANALYSES
DISTRIBUTED BY COMMUNITY AND RESIDENCE DURATION
Community
Pooled Low
Low I
Low II
Low III
Pooled High
High I
High 1 1
Total
Number of children
<1 yr
residence
512
290
148
74
492
180
312
1004
> 1 to < 3 yr
residence
213
145
42
26
251
68
183
464
>3yr
residence
1772
786
640
346
2533
1505
1028
4305
Total
2497
1221
830
446
3276
1753
1523
5773
Rocky Mountain Studies
341
-------
with the fact of family migration since increases in
many types of illness are reported after major changes
in life situations.11 An overwhelmingly important
determinant of acute respiratory illness is age, with
the highest rates in preschool children (Figure 3.3.1).
If the experience of children less than 1 year of age
were included, the origin point in Figure 3.3.1 would
fall at a point substantially lower than that plotted
for 1-year olds. Infants in the first year are in part
protected by passive transfer of maternal antibodies
and relative isolation within their homes. However,
the tremendous excess in lower respiratory illness
observed in older preschool children probably did not
result from the fading of passive immunity but rather
from the fact that the youngest children were not yet
born, as were their older siblings, when pandemic
influenza swept across the United States in late 1968
and early 1969. Higher attack rates were generally
found among boys, but females inched into the lead
position at age 12. This observed sex-age pattern is
classically reported for many infectious diseases. The
important fact is that any attempt to assess pollution
effects must adjust for differences in age. Sex
adjustment, while desirable, is not as necessary
because there was no evidence of intercommunity
differences in the sex ratio.
Both unadjusted illness attack rates and rates
adjusted- for the effects of age, sex, and socio-
economic status were computed for the acute respira-
t 6 >
AGE AT LAST 8RITHDAY, (eats
Figure 3.3.1. Age-specific annual mean attack
rates for acute lower respiratory illness.
tory illness grouping as a whole and for each of its
constituent clinical syndromes. Unadjusted data are
given in the Appendix, Tables 3.3.A.1 to 3.3.A.5. For
adjusted rates, sample size restrictions required
frequent pooling. Therefore, residence durations of
less than 3 years were combined, as were those of 3
years or more. Three other points deserve emphasis.
First, unadjusted illness rates for nonasthmatic
children differed from illness rates in asthmatics in
that no intercommunity differences were immediate-
ly apparent. The second observation was that length
of residence was a powerful determinant of illness
that fully justified separate consideration in the
analysis. The third puzzling point was that the illness
behavior of children from families who had recently
migrated seemed paradoxical, with children from
families moving into the cluster of less polluted
communities having more illness during the first years
of residence than similar children moving into the
more polluted cluster. Thereafter, children from the
less polluted communities generally reported progres-
sively lower attack rates, while those from the more
polluted cluster reported progressively higher attack
rates. It seems possible that families with vulnerable
children somehow preferentially avoided moving into
the more polluted cluster. Since the presence of
chronic respiratory disease symptoms in adults ac-
curately predicted the occurrence of excess lower
respiratory disease in their children,21 perhaps
parents of the more susceptible children were
prevented from being employed in and then moving
into the more polluted cluster by occupational health
preplacement screening. Another possible explanation
is that the communities in the High pollution cluster
are smaller and hence have fewer endogenous strains
of infectious organisms that typically affect migrants
than the larger, more heterogeneous communities in
the Low pollution cluster.
After the rates were appropriately adjusted,
excesses in single and repeated attacks of any acute
lower respiratory illness were apparent among
children living in the High exposure communities for
3 or more years (Table 3.3.6). The significance of
each of the major determinants of acute respiratory
disease was tested and the results are summarized
(Table 3.3.7). Age, sex, and pollution exposure all
had significant impacts on attack rates for repeated
episodes of lower respiratory illness among children
living in the High pollution cluster for 3 or more
years. The observed two failures of model fit, which
were not themselves independent of one another,
could be explained by chance alone if one considered
the total number of model fits tested.
3-42
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.3.6. AGE-, SEX-, SES-ADJUSTED 3-YEAR ATTACK RATES FOR ANY LOWER
RESPIRATORY ILLNESS BY RESIDENCE DURATION AND NUMBER OF EPISODES
Pooled
communities3
Low
High
Low
High
Number of
illness episodes
>1
^2
Attack rate, percent
<3yr
residence
25.0
22.0
9.2
9.1
>3yr
residence
21.3
24.2
9.2
12.1
Any length
residence
22.9
23.8
9.6
12.4
Low—Low I, Low II,and Low III; High — High I and High II.
Table 3.3.7. AIMLYSJS^OF^AmMQE^P^^AJEGC)RICAL. DATA: SUMMARY^OF PROBABILITIES
BY NUMBER OF EPISODES AND RESIDENCE DURATION, ANY LOWER
RESPIRATORY DISEASE
Factor
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Number
of
episodes
^1
^2
Probability of effect for indicated
residence duration3
<3yr
NS
<0.001
NS
NS
NS
NS
<0.001
NS
<0.01
<0.02
>3yr
<0.10
<0.001
<0.001
NS
NS
<0.001
<0.001
<0.001
NS
NS
Any length
NS
<0.001
<0.001
<0.05
NS
<0.01
<0.001
<0.001
<0.01
<0.05
NS—not significant, p > 0.10.
Croup
Excess laryngotracheobronchitis, or croup, mor-
bidity in the High pollution cluster accounted for
a large portion of the previously described excess in
total acute lower respiratory illness (Table 3.3.8).
Pollution effects were significant in explaining higher
attack rates for single or repeated episodes of illness
(Table 3.3.9). New arrivals in the more polluted
communities reported less illness than those in the
cleaner communities. Younger children, especially
those under 4, and girls had higher attack rates.
Socially advantaged families also reported more croup
than less advantaged families.
Bronchitis
Appropriately adjusted bronchitis rates trended in
the same direction as those for total lower respiratory
illness (Table 3.3.10). However, for bronchitis the
differences were not significant (Table 3.3.11). The
Rocky Mountain Studies
3-43
-------
Table 3.3.8. AGE-, SEX-, SES-ADJUSTED 3-YEAR ATTACK RATES FOR CROUP
BY RESIDENCE DURATION AND NUMBER OF EPISODES
Pooled
communities3
Low
High
Low
High
Number of
illness
episodes
>1
>2
Attack rate, percent
<3yr
residerice
8.1
7.7
2.4
2.2
>3yr
residence
6.2
11.2
3.1
4.8
Any length
residence
7.2
10.5
3.2
4.3
Low—Low I, Low II, Low III; High—High I and High II.
Table 3.3.9. ANALYSIS OF VARIANCE FOR CATEGORICAL DATA: SUMMARY OF PROBABILITIES
BY NUMBER OF EPISODES AND RESIDENCE DURATION, CROUP
Factor
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Number
of
episodes
>1
>2
Probability of effect for indicated
residence duration3
<3yr
NS
<0.001
<0.02
NS
NS
NS
<0.05
NS
<0.10
NS
>3yr
< 0.001
< 0.001
<0.01
NS
NS
<0.05
< 0.001
<0.01
NS
NS
Any length
< 0.001
< 0.001
<0.01
<0.02
NS
<0.10
< 0.001
<0.02
<0.10
NS
NS—not significant, p > 0.10.
Table 3.3.10. AGE-, SEX-, SES-ADJUSTED 3-YEAR ATTACK RATES FOR BRONCHITIS
BY RESIDENCE DURATION AND NUMBER OF EPISODES
Pooled
communities3
Low
High
Low
High
Number of
illness
episodes
>1
>2
Attack rate, percent
<3yr
residence
16.2
14.5
5.3
4.7
>3yr
residence
14.9
15.8
5.1
6.7
Any length
residence
15.6
15.5
5.2
6.3
aLow— Low I, Low II, and Low III; High—High I and High II.
3-44
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.3.11. ANALYSIS OF VARIANCE FOR CATEGORICAL DATA: SUMMARY OF
PROBABILITIES BY NUMBER OF EPISODES AND RESIDENCE DURATION, BRONCHITIS
Factor
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Number
of
episodes
>1
^2
Probability of effect for indicated
residence duration8
<3yr
NS
< 0.001
NS
NS
NS
NS
<0.01
NS
NS
NS
>3yr
NS
< 0.001
< 0.001
NS
NS
<0.10
<0.01
<0.01
NS
NS
Any length
NS
< 0.001
<0.01
NS
NS
NS
< 0.001
< 0.001
NS
NS
NS—not significant, p > 0.10.
direction of the differences was hardly surprising
since acute bronchitis is the largest single component
of the total lower respiratory illness grouping. Again,
children in the cleaner communities tend to have
more illness initially but less after 3 years than
children from polluted communities. Significant
differences attributable to age and sex again occurred.
As with croup, more advantaged families tended.
although not significantly, to report higher attack
rates.
Pneumonia
Attack rates for single episodes of pneumonia
mimicked the community pollution, age, and sex
trends previously described (Table 3.3.12). Excesses
in the more polluted communities, though not
statistically significant (Table 3.3.13), are never-
theless worrisome since pneumonia in children is a
serious, potentially life-threatening illness. Fortunate-
ly, practically no excess in repeated episodes of
pneumonia was observed. Socioeconomic effects,
while not significant, did trend opposite to those
observed for croup and bronchitis in that children
from less advantaged families had higher pneumonia
attack rates.
Hospitalizations
The nonsignificant trend toward more frequent
hospitalizations in the polluted communities,
previously found among asthmatics, was replicated in
the larger nonasthmatic population (Tables 3.3.14 and
3.3.15). Moreover, a nonsignificant tendency toward
higher rates for repeated hospitalization was apparent
in the more polluted communities, and this finding
gives warning that the excess in illness induced in the
more polluted communities may be more severe. The
effects of age, sex, and Socioeconomic status paral-
leled those described for penumonia, as might be
expected since many childhood pneumonia cases
frequently require hospital care.
Age-specific Attack Rates
Because croup appeared to be the clinical
syndrome most likely induced, at least in part, by
sulfur dioxide and suspended sulfate exposures, an
attempt was made to link exposure duration with the
onset of higher attack rates. It was necessary to attain
adequate sample sizes and to isolate, balance, or
adjust for the duration of residence, for age, and for
Rocky Mountain Studies
3-45
-------
Table 3.3.12. AGE-, SEX-, SES-ADJUSTED 3-YEAR ATTACK RATES FOR PNEUMONIA
BY RESIDENCE DURATION AND NUMBER OF EPISODES
Pooled
communities3
Low
High
Low
High
Number of
illness episodes
>1
^2
Attack rate, percent
<3yr
residence
5.0
6.7
1.1
1.4
residence
5.2
5.7
1.4
1.4
Any length
residence
5.4
6.0
1.2
1.3
Tow—Low I, Low II, and Low III; High—High I and High II.
Table 3.3.13. ANALYSIS OF VARIANCE FOR CATEGORICAL DATA: SUMMARY OF
PROBABILITIES BY NUMBER OF EPISODES AND RESIDENCE DURATION, PNEUMONIA
Factor
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Number
of
episodes
^1
^2
Probability of effect for indicated
residence duration3
<3yr
NS
<0.001
NS
NS
NS
NS
<0.01
NS
NS
NS
>3yr
NS
<0.001
<0.01
NS
NS
NS
<0.01
<0.02
NS
NS
Any length
NS
<0.001
<0.05
NS
NS
NS
<0.01
<0.02
NS
NS
aNS—not significant, p > 0.10.
sex. The problem of preventing Socioeconomic bias
further complicated the issue. A simple analytical
scheme was devised, and the influence of intervening
variables was assessed prior to testing for significance.
The analysis was limited to children of families who
had lived for 3 or more years in the same study
community. The population was partitioned accord-
ing to community air pollution, the sexes were
combined, and adjacent ages were pooled. No sex or
age biases were apparent in the two resulting distribu-
tions. Any bias attributable to overreporting among
the more advantaged families would hamper efforts
to attribute intercluster differences to ambient air
pollution since the Low pollution cluster contained
proportionately more of the advantaged families.
Age-specific attack rates for each cluster were then
plotted (Figure 3.3.2). Community clusters had
practically identical attack rates at ages 1 and 2, but
never thereafter. Attack rates for children from the
more polluted cluster decreased much more slowly
than those for children from the less polluted
community cluster. The gap between the two clusters
3-46
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.3.14. AGE-, SEX-, SES-ADJUSTED 3-YEAR HOSPITALIZATION RATES
BY RESIDENCE DURATION AND NUMBER OF ADMISSIONS
Pooled
communities8
Low
High
Low
High
Number of
illness episodes
^1
>2
Residence duration
<3yr
residence
4.3
6.3
1.2
1.6
^3yr
residence
4.3
5.4
1.3
1.4
Any length
residence
4.8
5.7
1.0
1.2
3Low—Low I, Low II,and Low III; High—High I and High II.
Table 3.3.15. ANALYSIS OF VARIANCE FOR CATEGORICAL DATA: SUMMARY OF PROBABILITIES
BY NUMBER OF EPISODES AND RESIDENCE DURATION, HOSPITALIZATION
Factor
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Exposure
Age
Sex
Socioeconomic
status
Fit of model
Number
of
episodes
^1
^2
Probability of effect for indicated
residence duration3
<3yr
NS
<0.001
NS
NS
NS
NS
NS
NS
NS
NS
>3yr
NS
<0.001
<0.01
NS
NS
NS
<0.001
<0.02
NS
NS
Any length
NS
<0.001
<0.05
NS
NS
NS
<0.001
<0.01
NS
NS
NS—not significant, p > 0.10.
appeared to be closing by age 11 to 12. Statistically
significant increases were demonstrable in the High
pollution areas by age 3 to 4. There can be little
doubt, then, that an exposure duration of 3 years
was strongly coupled to increased croup morbidity.
Attack rates within 4-year age groupings (1 to 4, 5
to 8, and 9 to 12) for families having different
educational levels were computed to look for any
peculiar interactions involving age, social status, and
the two most common lower respiratory illness
syndromes: croup and bronchitis. A relative risk
related to pollution exposure was then constructed
using the appropriate attack rate for the more
polluted communities as the numerator and the
attack rate for the cleaner communities as the
denominator (Figure 3.3.3). The full impact of
intercommunity differences in social class became
clear. Pollution had a greater relative effect on less
advantaged families than on advantaged families. This
excess was most marked when croup was considered.
One factor contributing to such an excess may well
be the poorer housing of less advantaged families.
Such housing would more likely be located closer to
Rocky Mountain Studies
3-47
-------
468
AGE AT LAST BRITHDAY.jeats
Figure 3.3.2. Age-specific annual mean attack
rates for croup distributed by community air
pollution.
the point sources of pollution and might well be less
well constructed, allowing freer penetration by pol-
lutants. Poorer families are also less likely to have
modern heating and air conditioning systems, which
may in part dampen the penetration of air pollutants
into the home.
Physician Diagnostic Patterns and Validating
Questionnaire-reported Illness Using
Physician Records
Possible effects of differences in diagnostic habits
of practicing physicians or overreporting of acute
lower respiratory illness on the significant morbidity
excesses thus far attributed to ambient air pollutants
were considered. There is no evidence, however, for
the occurrence of such a bias, and much evidence
against it.
Eighteen physicians who cared for children who
comprised the study population were asked to
diagnose six respiratory syndromes as "bronchitis" or
1 T
BRONCHITIS
6
ACE.yeais
Figure 3.3.3. Difference in _th_e_ relative risk of
croup and bronchitis which can be attributed
to residence in polluted communities.
"not bronchitis." The syndromes were graded ranging
from pharyngitis with minimal, but unmistakable,
lower respiratory symptoms to classical pneumonia.
Bronchitis was purposely chosen since bronchitis in
children is a somewhat imprecise diagnosis, thus
allowing a better chance to discover any latent
intercommunity differences. Physicians' diagnostic
decisions were tallied for pooled communities (Table
3.3.16). No remarkable intercommunity differences
could be attributed to differences in diagnostic
custom. Physicians, as expected, agreed with each
other better at the extremes of the syndrome and
tended to disagree when the clinical picture was less
distinctive.
The questionnaire used in this study classified
deep chest colds as bronchitis, which is a lower
respiratory illness. This made validation of question-
naire responses a difficult task since physicians will
often note only acute respiratory infection or upper
respiratory illness when, in fact, they diagnose a
"chest cold." At times, physicians only record
symptoms which are of value in helping recall the
3-48
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.3.16. CLINICAL DEFINITION OF ACUTE BRONCHITIS: NUMBER OF PHYSICIANS
CHARACTERIZING SIX RESPIRATORY ILLNESS SYNDROMES AS
"BRONCHITIS" OR "NOT BRONCHITIS"
Syndrome
description
Pharyngitis
Deep chest cold or
mild bronchitis
Croup or
laryngotracheobronchitis
Classical bronchitis
Severe bronchitis or
early pneumonia
Pneumonia
Physicians from
Low exposure areas
Bronchitis
0
3
2
3
1
0
Not bronchitis
3
0
1
0
2
3
Physicians from
High exposure areas
Bronchitis
3
13
4
11
6
0
Not bronchitis
12
2
11
4
g
15
patient's medical history. Other possible errors
include imperfections in memory, diagnosis, doctor-
patient communication, and enumeration. For the
validation estimate, either the specific diagnosis of
bronchitis or any acute respiratory illness confirmed
by a medical record was accepted as validation of
lower respiratory illness reported by the question-
naire respondents. On the other hand, a questionnaire
reporting no illness was not considered in error unless
a specific lower respiratory illness syndrome was
diagnosed. For each area, a 15 percent sample of
questionnaire responses was validated by the
described review of medical records. For each physi-
cian included in the validation sample, children sick
with bronchitis and an appropriate number of
children denying lower respiratory illness were
selected. (For this reason, the use of sensitivity and
specificity measures to arrive at corrected rates is not
valid.)22 Within the "sick sample" attempts were
made to validate all lower respiratory illness episodes.
There was a trend toward poorer validation of
questionnaire-reported illness in the High pollution
communities (Table 3.3.17). However, within the
"well sample," there was closer agreement with
physician records in the High pollution communities.
While probably not of sufficient magnitude to bias
the study, these trends were disturbing and difficult
to interpret.
DISCUSSION
For nonasthmatic children, elevated annual
average exposures, and perhaps repeated short-term
peak exposures, to sulfur dioxide and suspended
sulfates were significantly associated with excesses in
repeated episodes of any lower respiratory disease.
The same significant association was found for single
and repeated episodes of croup. Nonsignificant
morbidity excesses in bronchitis, pneumonia, and
hospitalizations because of acute lower respiratory
illness were also found in the High exposure cluster
(Table 3.3.18). The analysis also revealed that
suspected determinants of acute respiratory disease,
including age, sex, social class, and change in family
situation, were behaving in predictable
fashion.1 >2>23-26 Once this was assured, construction
of an appropriate age-specific model for croup, which
was the disorder most influenced by pollution, was
easily accomplished. The model clearly established
that 3- to 4-year exposures to elevated annual average
concentrations of sulfur dioxide (177 to 374jug/m3)
accompanied by elevated annual average levels of
suspended sulfates (7 to 11 Mg/m3) would enhance
host susceptibility for croup. It cannot be categorical-
ly stated that this effect was not attributable to
repeated, short-term peak exposures that were much
higher, though this seems unlikely because of the
Rocky Mountain Studies
3-49
-------
Table 3.3.17. VALIDATION OF QUESTIONNAIRE RESPONSES REPORTING CHILDREN
EITHER SICK OR WELL AGAINST PHYSICIAN RECORDS3
Community
Pooled Low
Low 1
Low II
Low III
Pooled High
High 1
High II
Overall
Any lower respiratory illness
Sick
validated
Percent
88
90
92
78
75
70
83
81
nb
86
42
26
18
108
61
47
194
Well
validated
Percent
75
78
74
74
85
84
86
80
n
191
77
68
46
162
85
77
353
Bronchitis
Sick
validated
Percent
80
84
87
71
71
65
78
76
n
77
37
23
17
91
51
40
168
Well
validated
Percent
76
77
78
81
93
94
92
85
n
200
82
71
47
180
96
84
380
Allowances were made to adjust for multiplicity of diagnostic terms used to describe mixed upper and lower respiratory
infections. Otherwise, the agreements would be substantially reduced.
n is the number of records examined.
relatively long latent period before the effect became
manifest. Quantifying the adverse impact of pol-
lutants in children who were not asthmatics was
much more difficult than in asthmatic children, as the
statistical analysis was forced to cope with bother-
some intercommunity differences in social class as
well as with migration, which was probably itself
influenced by pollution. The investigators had care-
fully attempted to avoid any intercommunity social
class differences, but were constrained by the paucity
of suitable geographically and climatologically similar
communities. Moreover, grouping of the Low III
smelter community with the Low pollution cluster,
while perfectly justified by its relatively low annual
average sulfur dioxide level, was a course selected
from a narrow array of poor options. We believe this
led to a conservative estimate of pollution effects.
Appropriate adjustments were able to cope with the
confusing effects of intervening variables, and a
consistent morbidity pattern became apparent.
For asthmatic children, elevated sulfur dioxide and
suspended sulfate air pollution was significantly
linked to excess morbidity for the acute respiratory
grouping as a whole and for croup and bronchitis
separately. As with nonasthmatic children, non-
significant morbidity excesses in pneumonia and
hospitalizations because of acute lower respiratory
illness were found in the High exposure cluster (Table
3.3.18).
What limitations still might alter the conclusions
of the present study? Overreporting and non-
respondent bias are two. The worst projectable
overreporting possible, ignoring compensating socio-
economic effects, might reduce the magnitude of the
observed excesses but could not begin to equalize the
large excess morbidity observed.
It is also quite unlikely that the study results
would have changed if the children with missing in-
formation (10 percent) and the nonrespondents (15
percent) had been included in the analysis. One can
calculate how the inclusion of these children might
have affected the present study outcome. To equal
the significantly higher rates for any lower respiratory
disease in the High communities, a greater than
two-fold increase in the rate among Low community
nonrespondents coupled with a rate of zero among
High community nonrespondents would have been
necessary. Alternately, a 50 percent increase among
the Low community nonrespondents along with a
similar decrease for those in the High community
3-50
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.3.18. EXCESS ACUTE LOWER RESPIRATORY ILLNESS AMONG CHILDREN EXPOSED
3 OR MORE YEARS TO ELEVATED LEVELS OF SULFUR DIOXIDE
Illness
Any acute
respiratory
disease
Croup
Bronchitis
Pneumonia
Hospitalization
Number of
illness episodes
>1
>2
>1
>2
>1
>2
>1
>2
>1
>2
Excess in high
exposure communities, percent
Among
asthmatics
25
NAb
60
NA
36
NA
27
NA
31
NA
Among
nonasthmatics3
14
42
81
55
6
31
10
0
26
8
Adjusted for age, sex, and socioeconomic status.
NA—not available.
would have made the overall rates equal. To equal the
significantly higher rates for croup, the Low com-
munity nonrespondent rate would have to be in-
creased 2.9 times at the same time that the rate for
the High community nonrespondents was zero. Rates
would also have to be more than 75 percent higher
for Low community nonrespondents and more than
75 percent lower among those of the High com-
munity for the reported rates of croup to be equal.
Such outcomes would be extremely uncommon.
by elevations of annual average levels of total
suspended particulates (99 to 115 jug/m3) and inter-
mediate suspended sulfates (6 to 7 jUg/m3). Stringent
control of sulfur dioxide, suspended particulates, and
suspended sulfates should be accompanied by
tangible reductions in lower respiratory illness in all
children.
SUMMARY
Based on the above analyses and considerations,
we conclude that excessive acute lower respiratory
illnesses can be expected among asthmatic and
nonasthmatic children who are exposed for longer
than 2 years to elevated annual average sulfur dioxide
levels (177 /ug/m3) accompanied by elevated annual
average suspended sulfate levels (7.2 Mg/m3) in the
presence of low levels of suspended particulates (65
jug/m3). We also find strongly suggestive evidence that
excess lower respiratory illness in asthmatics can be
linked to even lower pollution levels: annual average
sulfur dioxide levels as low as 67 Mg/m3 accompanied
Acute respiratory morbidity in asthmatic children
began to increase with exposures to low annual
average levels of sulfur dioxide (67 Mg/m3) when
moderately elevated levels of total suspended particu-
lates (99 to 115 Mg/m3) were present. Significant
excesses in acute lower respiratory illnesses, especially
croup, appeared among asthmatic and nonasthmatic
children exposed to higher pollution levels for 3 or
more years. Such levels involved estimated annual
average sulfur dioxide levels of 177 ng/m3 ac-
companied by low estimated annual average levels of
particulates (65 jug/m3) and minimally elevated
Rocky Mountain Studies
3-51
-------
estimated average levels of suspended sulfates (7.2
jug/m3). There was also suggestive evidence that more
serious infections, including pneumonia, that might
require hospitalization tended to be more frequent in
children exposed to these elevated pollution levels.
Other determinants of respiratory illness included
age, sex, social status, and recent intercommunity
migration. Future studies will be required to validate
the morbidity reductions that should follow pollution
control.
REFERENCES FOR SECTION 3.3
1. Acute Conditions (Incidence and Associated Dis-
ability, United States-June 1965-June 1966).
National Center for Health Statistics, Public
Health Service, U. S. Department of Health,
Education, and Welfare. Washington, D. C. PHS
Publication No. 1000, Series 10, No. 38. June
1967. 61 p.
2. Dingle, J. H., G. F. Badger, and W. S. Jordan.
Patterns of Illness. In: Illness in the Home.
Cleveland, The Press of Western Reserve Univer-
sity, 1964. p. 33-37.
3. Reid, D. D. -The Beginnings of Bronchitis. Proc.
Roy. Soc.Med. 62:311-316, 1969.
4. Douglas, J. W. B. and R. E. Waller. Air Pollution
and Respiratory Infection in Children. Brit. J.
Prevent. Soc.Med.20:1-8, 1966.
5. Lunn, J. E., J. Knowelden, and A. J. Handyside.
Patterns of Respiratory Illness in Sheffield Infant
School Children. Brit. J. Prevent. Soc. Med.
27:7-16, 1967.
6. Holland, W.W., T. Halil, A. E. Bennett, and
A. Elliott. Factors Influencing the Onset of
Chronic Respiratory Disease. Brit. Med. J.
2:205-208, April 1969.
7. Toyama, T. Air Pollution and Its Health Effects
in Japan. Arch. Environ. Health. 5:153-173
1964.
8. Manzhenko, E. G. The Effect of Atmospheric
Pollution on the Health of Children. Hygiene and
Sanitation (Moscow). 57:126-128,1966.
9. Shy, C. M., J. P. Creason, M. E. Pearlman, K. E.
McClain, F. B. Benson, and M. M. Young. The
Chattanooga School Children Study: Effects of
Community Exposure to Nitrogen Dioxide; II.
Incidence of Acute Respiratory Illness. J. Air
Pollut. Contr. Assoc. 20:582-588, September
1970.
10. Pearlman, M. E., J. F. Finklea, J. P. Creason,
C. M. Shy, M. M. Young, and R. J. M. Horton.
Nitrogen Dioxide and Lower Respiratory Illness.
Pediatrics. 47:391-395, 1971.
11. Rahe, R. H., J. D. McKean, and R. J. Arthur. A
Longitudinal Study of Life-change and Illness
Patterns. J. Psychosom. Res. 70:355-366, 1967.
12. A Study of Air Pollution in Montana, July
1961-June 1962. Montana State Board of Health.
Helena, Montana. 107 p.
13. A Study of Air Pollution in the Deer Lodge
Valley-August 1965-June 1966. Montana State
Department of Health. Helena, Montana. 32 p.
14. A Study of Air Pollution in the Helena-East
Helena Area-October 1965-October 1968.
Montana State Department of Health. Helena,
Montana. 35 p.
15. A Study of Air Pollution in Townsend-Three
Forks, the Gallatin Valley, and West Yellow-
stone, November 1967-November 1968. Montana
State Department of Health. Helena, Montana.
43 p.
16. Idaho Air Quality—Methods of Measurement and
Analysis of Recent Data. Idaho Department of
Health. Boise, Idaho. August 19, 1970. 23 p.
17. Helena Valley, Montana, Area Environmental
Pollution Study. U. S. Environmental Protection
Agency. Research Triangle Park, N. C. Office of
Air Programs Publication No. AP-91. January
1972. 193 p.
18. Questionnaires Used in the CHESS Studies. In:
Health Consequences of Sulfur Oxides: A Report
3-52
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
22. Vecchio, T. Predictive Value of a Single Diag-
nostic Test in Unselected Populations. New Eng.
J. Med. 274:1171-1173, May 26, 1966.
19. English, T. D., J. M. Sune, D. I. Hammer, L. A.
Truppi, W. E. Culver, R. C. Dickerson, and W. B.
Riggan. Human Exposures to Air Pollutants in
Five Rocky Mountain Communities, 1940-1970.
In: Health Consequences of Sulfur Oxides: A
Report from CHESS, 1970-1971. U.S. Environ-
mental Protection Agency. Research Triangle
Park, N.C. Publication No. EPA-650/1-74-004.
1974.
23. Tucher, D., J. E. Coulter, and J. Downes.
Incidence of Acute Respiratory Illness among
Males and Females; Study No. 5. The Milbank
Memorial Fund Quarterly. XXX(1):42-60,
January 1952.
24. Tucher, D. Incidence of Pneumonia in Two
Communities in New York State. The Milbank
Memorial Fund Quarterly. XXX:224-23S, 1952.
20. Grizzle, J. E., C. F. Starmer, and G. G. Koch.
Analysis of Categorical Data by Linear Models.
Biometrics. 25(3):489-504, September 1969.
21. Hayes, C. G., D.I. Hammer, C. M. Shy, V.
Hasselblad, C. R. Sharp, J. P. Creason, and K. E.
McClain. Prevalence of Chronic Respiratory
Disease Symptoms in Adults: 1970 Survey of
Five Rocky Mountain Communities. In: Health
Consequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
25. Miller, F. J. W., S. D. M. Court, W. S. Walton,
and E. G. Know. The General Pattern of Respira-
tory Disease (Chapter XXX); and Colds,
Bronchitis, and Pneumonia (Chapter XXXI). In:
Growing Up in Newcastle upon Tyne. London,
Oxford University Press, 1960. p. 194-217.
26. Family Use of Health Services (United States-
July 1963-June 1964). National Center for
Health Statistics, Public Health Service, U. S.
Department of Health Education, and Welfare.
Washington, D. C. PHS Publication No. 1000,
Series 10, No. 55. July 1969. p. 18.
Rocky Mountain Studies
3-53
-------
APPENDIX
Table 3.3.A.I. UNADJUSTED 3-YEAR ATTACK RATES FOR ANY LOWER RESPIRATORY
DISEASE BY NUMBER OF EPISODES AND DURATION OF RESIDENCE
Community
Pooled Low
Low 1
Low II
Low III
Pooled High
Highl
High II
Pooled Low
Low 1
Low II
Low III
Pooled High
High 1
High II
Number of
illness episodes
>1
>1
>2
>2
••
Attack rate, percent
<1 yr
residence
27.3
27.9
29.7
20.3
19.3
20.6
18.6
11.5
13.4
11.5
4.1
9.1
8.3
9.6
> 1 to < 3
yr<
residence
17.4
17.9
19.0
13.0
24.7
14.7
28.4
9.9
11.0
7.1
7.7
9.2
5.9
10.4
>3yr
residence
19.2
20.7
16.7
20.5
21.3
21.7
20.7
8.5
9.5
7.8
7.2
11.6
12.9
9.8
Any length
residence
20.7
22.1
19.2
20.0
21.2
21.3
21.2
9.2
10.6
8.4
6.7
11.1
1Z2
9.8
Table 3.3.A.2. UNADJUSTED 3-YEAR ATTACK RATES FOR CROUP BY NUMBER OF EPISODES
AND DURATION OF RESIDENCE
Community
Pooled Low
Low I
Low II
Low III
Pooled High
Highl
High II
Pooled Low
Low I
Low II
Low III
Pooled High
High I
High II
Number of
illness episodes
^1
^1
>2
>2
Attack rate, percent
<1 yr
residence
9.2
9.3
10.8
5.4
5.7
3.3
7.1
3.1
3.8
2.7
1.4
1.8
1.7
1.9
> 1 to < 3
yr
residence
6.1
7.6
2.4
3.8
10.0
7.4
10.9
1.4
2.1
0.0
0.0
4.0
2.9
4.4
>3yr
residence
5.4
6.1
4.2
6.1
9.5
10.3
8.4
2.8
2.9
2.3
3.2
4.3
5.7
2.3
Any length
residence
6.2
7.0
5.3
5.8
9.0
9.5
8.4
2.7
3.0
2.3
2.7
4.0
5.2
2.5
3-54
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 3.».A.3. UNADJUSTED 3-YEAR ATTACK RATES FOR BRONCHITIS BY NUMBER
OF EPISODES AND DURATION OF RESIDENCE
Community
Pooled Low
Low 1
Low II
Low III
Pooled High
High!
High II
Pooled Low
Low 1
Low II
Low III
Pooled High
High)
High II
Number of
illness episodes
>1
>1
>2
>2
Attack rate, percent
<1 yr
residence
17.4
18.3
16.9
14.9
14.0
15.6
,13.1
6.4
8.6
4.1
2.7
5.1
5.0
5.T
> 1 to < 3
yr
residence
12.2
11.0
16.7
11.5
14.7
7.4
17.5
5.6
5.5
4.8
7.7
4.4
1.5
5.5
>3yr
residence
13.9
14.8
12.7
14.2
14.6
14.9
14.2
4.8
5.2
4.5
4.3
6.3
6.6
5.7
Any length
residence
14.5
15.2
13.6
14.1
14.7
14.7
14.4
6.0
6.1
4.5
4.3
6.0
6.3
5.6
Table 3.3.A.4. UNADJUSTED 3-YEAR ATTACK RATES FOR PNEUMONIA BY NUMBER OF
EPISODES AND DURATION OF RESIDENCE
Community
Pooled Low
Low I
Low 1 1
Low III
Pooled High
High I
High 1 1
Pooled Low
Low I
Low 1 1
Low II I
Pooled High
High I
High 1 1
Number of
illness episodes
>1
>1
>2
>2
Attack rate, percent
<1 yr
residence
6.4
6.6
8.1
2.7
9.6
8.9
10.4
1.2
1.0
2.0
0.0
1.4
1.7
1.3
> 1 to < 3
yr
residence
4.7
6.2
2.4
0.0
5.2
2.9
6.0
0.5
0.7
0.0
0.0
1.6
1.5
1.6
^3yr
residence
4.2
3.9
4.5
4.3
4.6
4.9
4.2
1.0
1.5
0.6
0.6
1.1
1.1
1.0
Any length
residence
4.7
4.8
5.1
3.8
5.2
5.2
5.2
1.0
1.3
0.8
0.4
1.2
1.2
1.1
Rocky Mountain Studies
3-55
-------
Table 3.3.A.5. UNADJUSTED 3-YEAR RATES FOR HOSPITALIZATION BY NUMBER
OF ADMISSIONS AND DURATION OF RESIDENCE
Community
Pooled Low
Low 1
Low II
Low III
Pooled High
Highl
High II
Pooled Low
Low 1
Low II
Low III
Pooled JHk[h_
High 1
High II
Number of
illness admissions
>1
>1
^2
>2
Attack rate, percent
<1 yr
residence
4.5
4.5
4.7
4.1
5.5
6.7
4.8
0.6
1.0
0.0
0.0
1.4
1.7
1.3
>1 to<3
yr
residence
3.3
2.8
7.1
0.0
6.0
2.9
7.1
0.0
0.0
0.0
0.0
1.2
1.5
1.1
>3yr
residence
3.6
2.5
4.7
4.0
4.3
5.0
3.2
0.8
0.6
1.1
0.9
0.9
1.2
0.4
Any length
residence
3.8
3.0
4.8
3.8
4.6
5.1
4.0
0.7
0.7
0.8
0.7
1.0
1.3
0.7
3-56
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
CHAPTER 4
CHICAGO-NORTHWEST INDIANA STUDIES
4-1
-------
4.1 HUMAN EXPOSURE TO AIR POLLUTANTS IN THE
CHICAGO-NORTHWEST INDIANA METROPOLITAN REGION,
1950-1971
David O. Hinton, B.S., Thomas D. English, Ph.D.,
Blaine F. Parr, Victor Hasselblad, Ph.D.,
Richard C. Dickerson, B.S.,
and Jean G. French, Ph.D.
4-3
-------
INTRODUCTION
In order to carry out epidemiologjc studies on the
health effects of ambient pollution, it is important to
have measurements of previous air pollution expo-
sures for the community under study as well as
continuing air monitoring surveillance. In the Chi-
cago-Northwest Indiana area, such data were available
from the City of Chicago's 20-station air monitoring
network and from the Environmental Protection
Agency's (EPA) National Air Surveillance Network
(NASN).
Two studies, concerned with the effects of sulfur
dioxide and total suspended particulates on the
health of selected groups, were carried out in the
Chicago-Northwest Indiana area during 1969-1970.
The Chicago-Northwest Indiana area was chosen as
the setting for these studies on the basis of the
following criteria: (1) pollution originated from
numerous sources, (2) pollutant levels were relatively
high for both total suspended particulates (TSP) and
sulfur dioxide (802), (3) recorded total suspended
particulates for 1969 showed an exposure gradient
within Metropolitan Chicago, and (4) air monitoring
networks were available for both Chicago and out-
lying areas.
One health effects study was concerned with the
acute respiratory disease experience of nursery school
children and their families within the City of Chi-
cago.1 Participating families living within 1.5 miles of
one of the city's 20 monitoring stations were grouped
into three exposure categories (Highest, High, Inter-
mediate) based on 1969 levels of total suspended
particulates. The distribution of the monitoring sta-
tions in the City of Chicago is shown in Figure 4.1.1.
In the other study, differences in the prevalence of
chronic respiratory disease symptoms were examined
in military recruits from the Chicago-Northwest
Indiana Metropolitan area.2 Using reported residen-
tial histories and known or estimated pollution levels,
subjects were assigned to one of three community
exposure strata: relatively clean outlying areas (Low);
Chicago suburbs (High II); or the urban core of the
Chicago-Northwest Indiana area (High I). The relative
location of these areas is shown in Figure 4.1.2.
These two studies were preliminary to the full
development of the national Community Health and
Environmental Surveillance System as described else-
where.3'4 The main objective of the two Chicago
studies was to determine health effects of combined
exposure to particulates and sulfur oxides. The
Figure 4.1.1.
tions (City).
City of Chicago monitoring sta-
purpose of this paper is to document the air quality
data from which exposure estimates were made.
TOPOGRAPHY
Chicago is situated on a flat, or at most gently
rolling, plain with an average elevation of 620 feet
above mean sea level. While the terrain itself is not
rough enough to significantly affect airflow, discon-
tinuities created by local heat sources and sinks do
affect low-level mixing. This discontinuity is caused
by the large water body mass of Lake Michigan and
results in lake breezes in the spring and less strong
land breezes in the fall. The central Loop area and
most of the lake shore have an additional disconti-
nuity in airflow caused by high-rise apartment build-
ings. Just how much this affects mixing and disper-
sion of pollution is not fully known at this time.
4-4
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
O PALATINE
QNILES
WILMETTE
FRANKLIN PARK
HILLSIDE O
CICERO
BLUE ISLAND
HARVEY O N—H OHAMMOND
CALUMET CITYHyASTCH|CAGO
I
FLOSSMOOR O j
CHICAGO HEIGHTSQ !
CHICAGO SUBURBS (HIGH II)
OUTLYING AREAS (LOW)
Figure 4.1.2. Chicago-Northwest Indiana
Metropolitan area monitoring stations (NASN).
CLIMATOLOGY
The climate of Chicago is predominately con-
tinental, although it is modified to some extent by the
proximity of Lake Michigan. The weather, influenced
by warm, humid air masses originating from the Gulf
of Mexico, by cold, relatively dry air flowing out of
the Polar and Arctic regions, and by more moderate
masses of air with origins in the North Pacific Ocean,
is frequently changeable. Daily temperature ranges
are rather large. Annually, the temperature averages
58.7°F. Seasonally, temperatures of 99°F or higher
are recorded at Midway Airport in about one-half of
the summers and temperatures of -10°F are recorded
during one-half of the winters. July is normally the
warmest month and January the coldest. The normal
heating season is from mid-September to early June,
while the air conditioning season is usually from
mid-June to early September.
Precipitation averages 33.23 inches a year, with
summer the wettest season and winter the driest.
Sunshine amounts are generally moderate in summer
but quite low in winter due to considerable cloudi-
ness produced locally by lake effects from Lake
Michigan. Visibility reductions are frequent because
of the availability of adequate water vapor combined
with hygroscopic aerosols.
Wind rose summaries for both Midway Airport and
O'Hare Field are shown in Figures 4.1.3 and 4.1.4.
Normally, the winds generated by the warm season
lake breeze reach only a few miles inland. A well
developed breeze during the warm season may reduce
temperatures near the shore by more than 10°F
relative to those further inland.
PERCENT OCCURRENCE
5.0
6.0
9.0
PERCENT OCCURRENCE
12 16
1
13-17 18-23 >23
WIND SPEED, mph
Figure 4.1.3. Surface climatologic wind rose,
Chicago Midway Airport, 1951 -1960.
Chicago-Northwest Indiana Studies
4-5
-------
PERCEHT OCCURRENCE
7.6
6.6
8.1
PERCENT OCCURRENCE
12
16
1
4-7
1
8"12 13717 18-23 ' >23
WIND SPEED, rnpti
Figure 4.1.4. Surface climatologic wind
rose, Chicago O'Hare Airport, 1951-1960.
The climate of suburban and outlying areas of the
Chicago-Northwest Indiana Metropolitan area is not
greatly affected by Lake Michigan but is generally
typical of the Northeastern Great Plains area of the
United States.
POLLUTION SOURCES
Pollution sources in general are most prevalent
near the Loop area of Chicago and decrease out-
wardly away from Lake Michigan. Distribution and
type of air pollution sources are shown in Figure
4.1.5. These sources are composed primarily of power
generating plants, residential and commercial sources,
automobiles, and a variety of industrial sources that
are not dominated by any one specific type of
product. Pollution effluents from these multiple
sources are continually influenced by changing clima-
tological conditions, creating both complex and
varying exposures to area residents. Because of the
natural attraction for people to live in close proxim-
O MULTIPLE RESIDENCES
O COMMERCIAL ESTABLISHMENTS
O INDUSTRIAL PLANTS
A PUBLIC UTILITY POWER PLANTS
Figure 4.1.5. Major air pollution sources in
the City of Chicago.
ity to their place of employment, the higher pollution
areas are also the most densely populated.
Figure 4.1.6 shows total sulfur dioxide source
distribution during 1967. Nonutility industries con-
tributed approximately 10 percent of the total sulfur
dioxide emission. This is in sharp contrast to the
power generation industry, which contributed 67
percent. Residential areas contributed approximately
15 percent of the total emissions during the same
year. Sulfur dioxide emission from residential space
heating is given in Figure 4.1.7.
Great progress has been made by the City of
Chicago in both reducing emissions and obtaining air
quality data throughout the city. Industry and power
generating companies have assisted in emissions re-
duction by both utilizing controls and converting to
dual fuel capabilities. The overall effectiveness of
these measures is readily apparent when current
annual pollutant concentrations are compared with
prior concentrations (Figure 4.1.8).
Although significant progress has been made in
improving the general air quality in the Chicago area,
it should be noted that total suspended particulate
levels even in the Intermediate area exceeded the
National Primary Air Quality Standard of 75 jug/m3
in 1971. Further improvements in air quality will be
4-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
RESIDENTIAL COAL,
HAND FIRED
RESIDENTIAL COAL
STOKER FIRED
RESIDENTIAL
LIGHT OIL
RESIDENTAL
HEAVY OIL
T5] " '
53,551 1
14.974J
~n 5|362
RESIDENTIAL |17
GAS "1
COMMERCIAL COAL
HAND FIRED
COMMERCIAL COAL
STOKER FIRED
COMMERCIAL
LIGHT OIL
COMMERCIAL
HEAVY OIL
COMMERCIAL
GAS
INDUSTRIAL COAL
HAND FIRED
INDUSTRIAL COAL,
STOKER FIRED
INDUSTRIAL
LIGHT OIL
INDUSTRIAL
HEAVY OIL
INDUSTRIAL
GAS
UTILITY COAL
UTILITY GAS
~~j 5,832
14,903 1
^] 4,238
J 22,239 [
'
275
44,287 |
826
20,009 1
25
373,120
22 i
. ,. J K—L .-
370 380
SULFUR DIOXIDE EMISSIONS 10 tons
EMISSIONS, tons/year
I I 0 - 200
200 • 500
1 SOD • 10M
j 1000 • 1500
I 1500 • 2500
Figure 4.1.7. Sulfur dioxide emission from
residential space heating, City of Chicago,
1968.
Figure 4.1.6. Annual sulfur dioxide source
distribution for City of Chicago, 1967.
expensive; hence, it is essential to conduct studies
that objectively determine the health effects of air
pollutants to ensure that the cost of achieving the
primary standards are not higher than necessary to
adequately protect human health and welfare.
MONITORING HUMAN EXPOSURE
TO AIR POLLUTANTS
Air quality data from two monitoring networks
were available in the study area:
1. City of Chicago 20-station network (City).
2. Environmental Protection Agency National Air
Surveillance Network (NASN), including one
Continuous Air Monitoring Program (CAMP)
station.
City
Within the metropolitan area, the City of Chicago
Department of Environmental Control maintains a
network of 20 air monitoring stations (Figure 4.1.1).
Twenty-four-hour-average sulfur dioxide samples are
obtained twice weekly and analyzed by the West-
Gaeke method.5 Total suspended particulates are
obtained three times per week using the high-volume
sampler.6 The weight or mass of particulate matter
per unit volume of air is determined by drawing a
measured volume of air through a previously weighed
clean filter and then reweighing the soiled filter.
NASN
The NASN sampling frequency differs from that
of Chicago. NASN uses 26 randomly selected days
per year for both total suspended particulates and
Chicago-Northwest Indiana Studies
4-7
-------
TOTAL SUSPENDED PARTICULATE
I»
SULFUR DIOXIDE
6E 67
TIME, ,6313
Figure 4.1.8. Pollution exposure estimates,
City of Chicago, 1964-1971.
sulfur dioxide. Each sampler location, identified by
station name in Figure 4.1.2, represents the central
business-commercial district of that particular area. In
such a location, measured concentrations are among
the higher concentrations found in the immediate
area. These values may, therefore, slightly overesti-
mate area-wide exposure. In the absence of area data,
however, these measurements were used as the basis
for historical exposure estimates.
NASN collection procedures are similar to those of
the City network for total particulates. The high-
volume sampler is used to collect 24-hour samples.
The suspended sulfate fraction is determined by the
methylthymol blue method. Twenty-four-hour sulfur
dioxide samples are analyzed using the West-Gaeke
method.5 7 Continuous measurements of sulfur diox-
ide concentration are made at the CAMP station in
Chicago using a conductometric method.8 The princi-
ple of this method is the oxidation of sulfur dioxide
to sulfuric acid by aqueous hydrogen peroxide and
the subsequent measurement of the electrical con-
ductivity of the solution.
RESULTS AND DISCUSSION
Frequency Distribution
Continuous sulfur dioxide measurements for Chi-
cago are summarized in Figure 4.1.9 in terms of
cumulative frequency. These data, shown for several
different years, indicate that general air quality is
improving, although the National Daily Primary
Standard for sulfur dioxide (365 Mg/m3) was ex-
ceeded 8 percent of the days in 1970. This should be
compared with a corresponding 42 percent in the
1962-1964 period.
Exposure Estimates
Air pollution meteorological models were devel-
oped for the Chicago area by both Argonne National
Laboratories and the Mitre Corporation.9'10 These
models were useful for determining the distribution
of air pollutants in this area. Although both models
were available for use in the study and were in general
agreement for central Chicago, the Mitre model
predicted a lower concentration of sulfur dioxide in
the southwestern area than did the Argonne model.
Measured data from the City network, from which
the exposure estimates were made, were best sup-
ported by the Mitre model. An example isopleth for
sulfur dioxide during 1968 based on the Mitre model
is shown in Figure 4.1.10.
Exposure data are presented in two parts: first,
exposures within the City of Chicago for 1964-1971,
and second, exposures in the Chicago-Northwest
Indiana area for 1950-1971. Data on which these
estimates are based are given in the Appendix, Tables
4.I.A.I to 4.I.A.7. Calculations performed on the
data, described in the following paragraphs, are
illustrated in Figures 4.I.A.I to 4.I.A.9.
City of Chicago
In order to estimate prior exposures, data for the
period 1950 through 1970 were used. Except for
annual particulate dustfall in Chicago (Figure
4.1.A.I), many data points were missing. Relation-
ships between pollutants, such as the ratio of total
suspended particulates to monthly dustfall in Table
4.1.A.3, were computed to provide reasonable esti-
mates for missing data points.
To estimate missing total suspended particulate
values, an average ratio was determined from a graph
of the ratio of total suspended particulates to average
4-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
1000
900
800
700
600
500
400
300
200
CO
I,
B 100
I 90
70
60
50
30
20
10
NATIONAL DAILY AMBIENT
AIR QUALITY STANDARD
1
99
10 30 50 70 90
PERCENT LESS THAN INDICATED CONCENTRATION
Figure 4.1.9. Cumulative frequency distribution of sulfur dioxide concentrations, City of Chicago.
Chicago-Northwest Indiana Studies
4-9
-------
Figure 4.1.10. Sulfur dioxide isopleths for
Chicago and suburbs based on 1968 emissions.
particulate dustfall versus time (Figure 4.1.A.2).
Multiplication of the particulate-to-dustfall ratio by
corresponding yearly dustfall values gave estimated
particulate values for 1950-1953 (Figure 4.1.A.3).
The correlation of total suspended particulates to
dustfall is 0.75, and the regression line is plotted in
Figure 4. l.A.4. Although the limitation of extrapolat-
ing to earlier years is recognized, this still provides a
crude estimate for previous exposures.
Using the same approach, the ratio of suspended
sulfates to dustfall was plotted versus time (Figure
4.1.A.5) and projected values were used for missing
data for sulfates (Figure 4.1.A.6).
A similar comparison between the sulfur dioxide
data and dustfall data showed no consistent relation-
ship. Sulfur dioxide exposure estimates for the City
of Chicago were based instead on CAMP data.
Comparison of 1967 average sulfur dioxide values
from NASN Site 1 with those from the CAMP station
indicated that the CAMP value was higher by 29
percent. Further comparison of CAMP values with
the nearby City Station F for 1964-1969 (Table
4.I.A.I) indicated that CAMP values were approxi-
mately 15 percent higher than City values. CAMP
data reduced by 25 percent are considered compara-
ble to results from NASN Site 1. This assumption is
reasonable since CAMP sampling utilizes a conducto-
metric method instead of the West-Gaeke method
and thus tends to report higher readings due to
positive interference caused by other acidic gases.
Data for the City of Chicago in Table 4.l.A.4 are,
therefore, CAMP data reduced by 25 percent. Expos-
ure estimates for total suspended particulates and
sulfur dioxide in the City of Chicago are summarized
in Table 4.1.1 for the three study areas used in the
acute respiratory disease study.1 Sulfur dioxide data
for 1968 compare favorably with values predicted by
the model.
Chicago-Northwest Indiana
Exposure estimates for study areas used in the
chronic respiratory disease study2 were projected
from recent measurements (Figures 4.1.A.7 to
4.1.A.9). Sulfur dioxide, total suspended particulate,
suspended sulfate, and particulate dustfall exposures
are categorized by area into Chicago-Northwest Indi-
ana urban core (High I), suburban (High II), and
relatively clean outstate (Low) in Table 4.1.2.
SUMMARY
Sulfur dioxide, total suspended particulate, and
dustfall measurements for the City of Chicago and for
the Chicago-Northwest Indiana Metropolitan area
have been obtained by the City of Chicago and the
Environmental Protection Agency (NASN and
CAMP) in recent years. Based on these measured
values, it was possible to project pollutant exposure
in these areas for the period 1950-1970.
The exposure estimates indicated a definite pollu-
tion exposure gradient among Chicago areas con-
sidered in an acute respiratory disease study (Inter-
mediate, High, Highest) and between Chicago-
4-10
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.1.1. ESTIMATES OF EXPOSURE
FOR CITY OF CHICAGO,
1950-1970
Pollutant and
community
Total suspended
particulate
Intermediate
High
Highest
Sulfur dioxide
Intermediate
High
Highest
Concentration, jug/m3
1950- 58a
—
-
243
—
-
1959-633
123
140
165
130
130
250
1964-68
121
137
163
109
107
170
1969-70
111
126
151
57
51
106
Projected values.
Table 4.1.2. ESTIMATES OF EXPOSURE FOR CHICAGO-
NORTHWEST INDIANA METROPOLITAN REGION, 1950-1970
Pollutant and
community
Sulfur dioxide, jug/m3
Urban core (High l)a
Suburban (High II)
Relatively clean
outstate (Low)
Total suspended parti-
culates, jug/m3
Urban core (High I)
Suburban (High II)
Relatively clean
outstate (Low)
Suspended sulfates, fig/m3
Urban core (High I)
Suburban (High II)
Relatively clean
outstate (Low)
Dustfall, g/m2/mo
Urban core (High I)
Suburban (High II)
Relatively clean
outstate (Low)
Pollutant levels over time
1950-59
—
—
-70
244
174b
~80
20.6b
—
<8
19.4
—
—
1960-65
282
—
~70
177
141b
-80
17.3
—
<8
15.2
—
—
1966-68
157
100
43
149
118
76
14.1
—
7.7
14.2
—
—
1969-70
96
217C
19
155
103
71
14.5
—
9.3
12.6b
—
—
aValues for Chicago proper utilized.
Projected values.
C1969 data only.
Chicago-Northwest Indiana Studies
4-11
-------
Northwest Indiana areas considered in a chronic
respiratory disease study (High I, High II, Low). In
some instances, sulfur dioxide exposures in the High
II area equaled or exceeded those in the High I area,
possibly as a result of proximity to point sources.
Low exposures for the outlying Low area, attributed
to fewer sources per unit area, were verified.
REFERENCES FOR SECTION 4.1
1. Finklea, J.F., J.G. French, G.R. Lowrimore, J.
Goldberg, C.M. Shy, and W.C. Nelson. Pros-
pective Surveys of Acute Respiratory Disease in
Volunteer Families: Chicago Nursery School
Study, 1969-1970. In: Health Consequences of
Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Publica-
tion No. EPA-650/1-74-004. 1974.
2. Finklea, J.F., J. Goldberg, V. Hasselblad, C.M.
Shy, and C.G. Hayes. Prevalence of Chronic
Respiratory Disease Symptoms in Military
Recruits: Chicago Induction Center,
1969-1970. In: Health Consequences of Sulfur
Oxides: A Report from CHESS, 1970-1971.
U.S. Environmental Protection Agency. Re-
search Triangle Park, N.C. Publication No.
EPA-650/1-74-004. 1974.
3. Shy, C.M., W.B. Riggan, J.G. French, W.C.
Nelson, R.C. Dickerson, F.B. Benson, J.F.
Finklea, A.V. Colucci, D.I. Hammer, and V.A.
Newill. An Overview of CHESS. In: Health
Consequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Pro-
tection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
4. Riggan, W.B., D.I. Hammer, J.F. Finklea, V.
Hasselblad, C.R. Sharp, R.M. Burton, and C.M.
Shy. CHESS, a Community Health and En-
vironmental Surveillance System. In: Proceed-
ings of the Sixth Berkeley Symposium on
Mathematical Statistics and Probability (Vol.
6). Berkeley, University of California Press,
1972.
5. West, P.W. and G.C. Gaeke. Fixation of Sulfur
Dioxide as Sulfitomercurate III and Subsequent
Colorimetric Determination. Anal. Chem.
25:1816-1819, 1956.
6. U.S. Environmental Protection Agency. Na-
tional Primary and Secondary Air Quality
Standards; Reference Method for Determina-
tion of Suspended Particulates in the Atmos-
phere (High Volume Method). Federal Register.
J6(84):8191-8194, April 30, 1971.
7. National Air Sampling Network (NASN) Labor-
atory Methodology. In: Health Consequences
of Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Publica-
tion No. EPA-650/1-74-004. 1974.
8. Air Quality Criteria for Sulfur Oxides. National
Air Pollution Control Administration, Public
Health Service, U.S. Department of Health,
Education, and Welfare. Durham, N.C. NAPCA
Publication No. AP-50. January 1969. p. 21-22.
9. Wosko, T.D., M.T. Matthies, and R.F. King. A
Methodology for Controlling Air Pollution
Emergency. Argonne National Laboratories. Ar-
gonne, 111. ANL/ES-14. In preparation.
10. Golden, J. The Air Quality Display Model
Applied to Cook County, Illinois. The Mitre
Corporation. Washington, D.C. MRT-4148. July
1970. p. 27-53. (Document has not been
approved for public distribution.)
4-12
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
APPENDIX
Table 4 1.A.1. ESTIMATES OF SULFUR DIOXIDE EXPOSURE FOR CITY OF
CHICAGO (CITY NETWORK)
Concentration, ug/m^
Communi tya
Station
Highest Cooley (D)
GSA (F)
Crane (G)
Carver (Q)
Hyde Park (L)
High Austin (H)
Kelly (J)
Intermediate
Chicago Voc. (0)
Logan Sq. (V)
Clay (R)
Taft (A)
lakeview (B)
Steinmetz (C)
Lindbolm (K)
Stevenson- (M)
Calumet (N)
Fenger (P)
F?rr (I)
Sullivan (T)
Hale (W)
1964
210
445
183
105
314
262
105
105
-
52
131
210
131
131
52
52
105
-
341
52
1965
199
267
170
86
1966
144
191
1C8
100
241 236
204 217
113 110
52 86
-
55
81
206
94
121
66
84
52
139
186
100
97
89
173
73
131
115
141
102
136
186
149
1967
60
212
147
73
178
186
92
66
Data
63
68
139
68
105
110
118
63
155
115
a Grouped according to 1969 total suspended particulate
Table4.1.A.2. ESTIMATES
1968
89
147
94
52
141
115
66
39
1969
76
136
68
71
107
76
76
13
1970
152
131
1 105
! 71
134
18
81
18
1 1971
71
58
39
58
79
39
58
16
Average
125
198
122
77
179
140
88
49
hot available - - -
42
42
89
37
79
71
58
13
98
73
68
26
79
34
60
39
42
34
110
110
60
55
97
60
58
18
16
60
34
94
45
31
66
42
34
29
26
31
39
34
60
65
132
68
90
63
137
58
138
144
91
levels.
OF TOTAL SUSPENDED PARTICULATE EXPOSURE
FOR CITY OF CHICAGO (CITY
NETWORK)
Concentration, uq/rtr
Community3 Station
Highest Cooley (D)
GSA (F)
Crane (G)
i Carver (Q)
Hyde Park (L)
High ; Austin (H)
Kelly (J)
: Chicago Voc. (0)
Intermediate
Logan (V)
Clay (R)
Taft (A)
Lakeview (B)
1 Steinmetz (C)
lindblom (K)
Stevenson (M)
Calumet (N)
Fenger (P)
Farr (I)
Sullivan (T)
Hale (W)
*
1964
186
178
163
128
173
139
146
132
142
123
94
134
101
125
114
119
124
157
124
134
1965 1966
166 175
173 174
155 165
147 156
164 177
135 143
132 , 147
125 145
147 162
130 135
98 ' 100
128 132
100 ; 111
127 136
112
117
118
146
118
130
125
127
131
176
144
138
1967
167
172
142
142
147
132
132
133
139
128
82
117
93
118
113
115
117
-
125
120
1968
173
161
143
146
194
132
137
138
152
129
100
138
101
127
119
118
119
129
108
I
144
1969
166
160
156
156
172
133
133
133
141
132
96
124
113
119
109
121
122
126
105
123
1
1970
150
134
141
134
140
112
115
117
124
118
88
116
86
101
106
111
105
126
101
120
1971
145
125
133
116
118
107
106
111
112
113
85
104
80
91
97
100
103
119
90
101
Average
166
160
150
141
161
129
131
129
140
126
93
124
98
118
112
116
117
140
114
126
aGrouped according to 1969 total suspended particulate levels.
Chicago-Northwest Indiana Studies
4-13
-------
Table 4.1.A.3. AIR QUALITY DATA AND RELATIONSHIPS,
CITY OF CHICAGO (NASN)
Year
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
1958
1957
1956
1955
1954
1953
1952
1951
1950
Dustfall
parti cul ate,
g/m2/mo
12. 4a
12.83
15.2
14.6
12.9
14.4
14.1
15.8
15.1
17.1
14.4
17.1
18.3
19.1
20.1
18.5
19.6
18.8
19.7
i 22.3
20.5
Sulfate/
dustfall
ratio
0.881
0.849
0.891
1.10
1.35
1.07
1.03
1.46
1.043
1.053
0.907
1.063
1.21
1.03
1.05
l,08a
1.093
1.10a
1.11
Suspended
sul fates,
ug/m3
14.8
14.13
13.4
12.4
11.5
15.9
19.0
16.9
15.5
24.9
15.0
18.0
16.6
20.2
24.4
19.0
20.6
20.33
21.53
24.53
22.0a
Total
suspended
parti cul ates,
ug/m3
124
172
151
143
159
173
176
144
123
190
182
240
190
194
290
231
223
243a
256a
294a
273a
TSP/
dustfall
ratio
10. 3a
10. la
9.93
9.79
12.32
12.01
12.48
9.11
8.15
11.11
12.64
14.04
10.38
10.16
14.43
12.49
11.38
12. 9a
13. Oa
13.23
13.3a
Sulfur
dioxide,
wg/m3
72
184
174
232
221
341
459
393
296
350a
' 390a
Estimated values.
Table 4.1.A.4. ESTIMATES OF SULFUR DIOXIDE EXPOSURE FOR CHICAGO-NORTHWEST INDIANA
METROPOLITAN AREA (NASN)
Commimi ty
High I
High II
Low
Area
Chicago3
E. Chicago
Hammond
Harvey
Wilmette
Chicago Heights
Hillside
Cicero
Parke Co.
Monroe Co.
Evansville
Concentration, ug/m^
1960
1961
292b 280b
1962
222
1963
295
105
1964
344
100
j
I
i
i
i
1965 I 1966
I
256
128
166
107
1967
174
117
; i
1968
130
75
194
160
58 60 141
44
66
172
248
47
10
24
1969
138
101
85
94
1970
54
57
58
174
292
291
232
14
32
8
9
25
aCAMP measurements times 0.75.
^Estimated values.
4-14
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
H-
CO
LU
^
o
6
o
<
u
0
s
u.
LU
ee
Q. ~
uSw
LU<
||
ui o
V) UI
£5
o~
t/J
LU
UI
in
»—
®
.0
(Q
o
eyi
10
t— i
CO
VO
1 ~*
s
, a
i 8
; 2
vo
i
CO
g
1 ""*
r
I CM
IO
t-4
*o
Ot
«— t
co u5
E cn
ov
n
cn
* in
c cn
O ^H
10 CO
t. in
c 2
g -
§ 5
o cn
t— <
vo
in
cn
in
in
m
in
cn
CM
in
cn
• in
cn
1°
la
to
Community
CM
CM
in
CO
cn
in
CO
i-H
r-H
CO
CM
i-H
O
en
2
§
CM
§
cn
o
cn
CM
CO
CM
CO
CM
CM
CO
CM
in
CM
cn
CM
CO
CM
0
cn
m
u
?
cn oo
CO CM
f"^ cn
CO CM
£ o
co *r
\O CM
IO CO
CO «*
O CM
CM i-H
CO
CM
»-l VO
cn co
ft I-H
cn
CO CM
in co
<-H .— 1
g
i-H
*& r^«
O in
CM r-l
CM
CM
CO
CO
CM
cn
CM
in
CM
E. Chicago
Hanmond
*3" CO ^" C7t ^ CTt VO CM
o p^cocoin rococo
i-H r-l
COO^CDCMCnOCMlO^-CO «*CM COO
cM«-tcoc5r».ftinO'*corv. <*^- coco
*^ co ^- o co o vo cn o in co ^ c\i co in
t— * i— i tNi co r^* co in o% co ot cn *f *t co CM
oorN.^-cr»inoO'-*CMvo ^•»-' ^~
CM r** o r** (O ^~ co cn co ^ cn ^ in r**>
f-Hcoocn'— *I-«CMCOCOOO . co co
I-H
O
r* in 10
in en > cn i_
•M -0 -r- IQ
o ro ^ • o d
i- 0> OJ •— 01 C OOt_>"O
•M gjc+Jvio-o >r- o i-
»E'r-4-»'-«Cnf-O»— (US-O-*-*
EO>i/)4->a) O W>«OE
U* H
'•o ^*
o! i
4-* fft
U
01 r—
*^> i—
O
-------
o
C3
<
O
I
O
CL<
XLU
LUCC
COQ
So
CO I—
•"g
** r-- m
in o UD
r-H r*»- VjO
IO C\J O
O
00
8.
4-16
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.1.A.7. DUSTFALL PARTICULATE MEASUREMENTS
FOR CHICAGO NASN STATION
Year
1950
1951
1952
1953
1S54
1955
1956
Dustfall,
g/m2/mo
20.5
22.3
19.7
18.8
19.6
18.5
20.1
Year
1957
1958
1959
1960
1961
1962
1963
Dustfall,
g/m^/mo
19.1
18.3
17.1
14.4
17.1
15.1
15.8
Year
1964
1965
1966
1967
1968
1969
1970
Dustfall ,
g/m2/mo
14.1
14.4
12.9
14.6
15.2
12.89
12. 4a
Estimated values,
1960
TIME, years
Figure 4.1 .A.1. Dustfall particulate versus time, City of Chicago, 1950-1970.
1970
Chicago-Northwest Indiana Studies
4-17
-------
u.
15
12
o
£
S 6
a
Q_
to
• ACTUAL
A ESTIMATED
1950
1955
1960
1965
1970
TIME, years
Figure 4.1^A.2 . Total suspended particulate/dustfall particulate versus time, City of Chicago,
1950-19707
1960
TIME, years
Figure 4.1.A.3. Total suspended particulate versus time, City of Chicago, 1950-1970.
4-18
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
325
300
275
<£ 250
jf 225
3 200
o
£ 175
Q 150
UJ
g 125
^ 100
50
25
y =-49.5+14.4x
1 = 0.747
n =15
10 15
DUSTFALL, g/m2/mo
20
25
Figure 4.1 .A.4. Least squares fit, total suspended particulate versus dustfalI City of Chicago
1954-1968. ' '
L6
K 1.2
Q <
UJ
£
00
a
0.8
0-4
0
• ACTUAL
Q ESTIMATED
1950
1955
1965
1970
1960
TIME, years
Figure 4.1.A.5. Suspended sulfate/dustfall particulate versus time, City of Chicago, 1950-1970.
Chicago-Northwest Indiana Studies
4-19
-------
LU
O
o
o
10
a
LU c
a 3
a.
=D
in
0
1950
ACTUAL
ESTIMATED
1955
1965
1960
TIME, years
Figure 4.1.A.6. Suspended sulfates versus time, City of Chicago, 1950-1970.
1970
400
300
a:
o
o
x
o
a
tr
200
100
1950
HIGH I ^
HiGHi- -^ACTUAL
O ESTIMATED
1955
1960
TIME, years
1965
Figure 4.1.A.7. Sulfur dioxide exposure estimates, Chicago-Northwest Indiana Metropolitan
area, 1950-1970.
4-20
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
• HIGH k
A HIGH II-^ACTUAL
• LOW-'''
9 ESTIMATED
TIME, years
Figure 4.1.A.8. Total suspended particulate exposure extimates, Chicago-Northwest Indiana
Metropolitan area, 1950-1970.
o
o
C>O
a
LLl
a
LLl
D_
15
10
II HIGH |..^
• LOW—-*ACTUAL
O ESTIMATED
1950
1955
I960
1965
1970
TIME, years
Figure 4.1 .A.9.Suspended sulfate exposure estimates, Chicago-Northwest Indiana Metropolitan
area, 1950-1970.
Chicago-Northwest Indiana Studies
4-21
-------
4.2 PREVALENCE OF CHRONIC RESPIRATORY DISEASE
SYMPTOMS IN MILITARY RECRUITS: CHICAGO
INDUCTION CENTER, 1969-1970
John F. Finklea, M.D., Dr. P.H., Julius Goldberg, Ph.D.,
Victor Hasselblad, Ph.D., Carl M. Shy, M.D., Dr. P.H.,
and Carl G. Hayes, Ph.D.
4-23
-------
INTRODUCTION
This study represents an attempt to link
chronic bronchitis to long-term air pollutant exposure
in a population of adolescents and young adults.
Self-pollution by cigarette smoke is consistently
accompanied by a significant decrease in ventilatory
function and a clear increase in the risk of chronic
bronchitis among adults of both sexes.1"10 In older
adults, exposures to ambient air pollution are likewise
accompanied by reductions in ventilatory function
and increases in the prevalence of chronic bron-
chitis.1 1-1 s In late adolescence, exposures to elevated
levels of air pollution perceptibly reduce ventilatory
function, but the significance of this finding in
contributing to the development of chronic respira-
tory disease symptoms in later life is still unknown.8
Studies involving adolescents and young adults
have the advantage of allowing the assessment of
relatively long-term air pollution exposures while
weakening the effects of several intervening variables
that hamper such studies in older adults. Selection of
adolescents minimizes complex effects introduced by
residential mobility, long-term self-pollution by ciga-
rette smoking and occupational exposures involving
pulmonary irritants. Unfortunately, studies of
chronic respiratory disease symptoms and ambient air
pollution exposures in young adults are hampered by
two major difficulties: first, large populations are
required to quantify the relatively small differences in
symptoms that would be expected so early in life;
and second, securing the cooperation of free-living
young adults without introducing one or more serious
study biases is most difficult. Another major disad-
vantage is that very few geographic areas have
sufficient historical air monitoring data to allow good
estimates of previous air pollution exposures.
By and large, these disadvantages were alleviated
for a brief period during 1967-1969. Armed conflict
in Southeast Asia resulted in a rapid acceleration of
the number of military recruits being processed at
induction centers. There, recruits completed medical
history questionnaires and underwent physical exami-
nations. With the cooperation of the Department of
Defense, it was a relatively easy task to supplement
the standard military medical history with a question-
naire dealing specifically with chronic respiratory
disease symptoms. The Chicago Induction Center
processes recruits from the more polluted Chicago-
Northwest Indiana metropolitan urban core, from the
suburbs that surround that core, and from relatively
clean outstate areas of Illinois and Indiana. Of equal
importance, the City of Chicago and other govern-
mental units in the metropolitan area have main-
tained one or more air monitoring stations with
varying degrees of sophistication since 1926.
Thus, it was possible to obtain a large sample of
black and white male adults who were selected in the
same fashion from more polluted and relatively clean
communities. In no case did the recruits realize they
were participating in a study of impact of air
pollution. There was also ample opportunity to
collect blood samples that may later relate immuno-
logic and genetic risk factors to chronic respiratory
disease symptom prevalence. Furthermore, linking
the chronic respiratory disease survey to military
records provides an opportunity for lifetime follow-
up to assess the impact of the early onset of chronic
respiratory disease symptoms on changes in smoking
habits, on occupational choices, and on residential
preferences. Currently, our understanding of the
natural history of chronic respiratory disease is
clouded by failure to quantify the effects of these
variables.
Three specific hypotheses were tested in the study
now reported: first, that residence in high pollution
areas is accompanied by increases in the prevalence
and/or severity of chronic respiratory disease symp-
toms; second, that the prevalence of chronic bron-
chitis is related to the length of residence in polluted
areas; and third, that air pollution and cigarette
smoking together have a greater adverse effect than
either alone. Although other factors could have
significant effects, it is difficult to perceive how they
could affect the different areas in different manners.
Stringent environmental control measures began to
improve air quality during the late sixties. In the
future, we hope to document that the prevalence
and/or severity of chronic respiratory disease is
lessened by continued improvements in ambient air
quality.
METHODS
Health Questionnaire
During the period from June 24, 1969, to Febru-
ary 20, 1970, all military recruits undergoing physical
examination at the Chicago Induction Center received
a brief, self-administered questionnaire16 appended
to the routine military medical history form. In
addition to standard questions regarding age and race,
six questions that dealt specifically with chronic
4-24
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
respiratory disease symptomatology were asked. The
remainder of the questionnaire pertained to smoking
patterns, residential history, and educational attain-
ment. Participants were not informed that responses
would be used in a study of air pollution.
Morbidity Indices
Frequency and duration of cough and phlegm
production were used to classify each respondent on
the following scale of symptom severity:
1. No symptoms.
2. Cough alone for less than 3 months each year.
3. Phlegm with or without cough for less than 3
months each year.
4. Cough alone for 3 months or more each year.
5. Phlegm alone for 3 months or more each year.
6. Cough and phlegm for 3 months or more each
year.
With modification for self-administration, respiratory
symptom questions were similar in scope and wording
to those developed by the British Medical Research
Council, and the sixth ranking on the symptom
severity scales was defined as chronic bronchitis for
the purpose of the study. Three summary morbidity
indices were developed from these severity rankings:
first, chronic bronchitis prevalence rates; second,mean
symptom scores, which reflect the frequency of all
reported symptoms weighted by their severity; and
third, mean severity scores for those persons who
reported one or more chronic respiratory disease
symptoms.
Community Pollution Exposures
Air quality for areas in and around Chicago was
projected from monitoring data from the National
Air Surveillance Networks, the City of Chicago, and
other local and state air pollution agencies. The great
majority of those recruits from outside Chicago and
its suburbs came from small towns or rural areas
where they were exposed to low, almost background,
levels of pollution. A small proportion resided in
somewhat polluted outstate industrial areas, but the
paucity of aerometric measurements in these areas
made it difficult to quantify pollution exposures for
this population segment.
Using reported residential histories and known or
estimated pollution levels, subjects were assigned to
one of three exposure communities: relatively clean
outstate areas, Chicago suburbs, or the urban core of
the Chicago-Northwest Indiana conurbation. A small
group (<3 percent) of recruits were from states other
than Illinois and Indiana, and these were excluded
from the analyses.
Statistical Analyses
Comparisons were made across the three commu-
nities for the distribution of known or suspected
covariates and adjustments or controls were em-
ployed where appropriate. Hypotheses were tested in
a general linear model for categorical data.1 7 This
procedure is very similar to the analysis of variance
procedures used for continuous data, except that
sums of squares are replaced by Chi Squares. All rates
were adjusted for educational differences, and the
rates for smokers were adjusted for differences in the
amount smoked.
Environmental Exposures
Air pollution exposures were averaged over four
time periods, which reflected either growth and
development milestones of the study population or
major shifts in air quality (Table 4.2.1).18 The first
period, 1950-1959, coincided with infancy and child-
hood phases of growth and development. The second
period, 1960-1965, covered puberty and adolescence.
The third period related to late adolescence, which
also marked the onset of improvements in air quality
for the City of Chicago, and the induction study
years of 1969-1970 constituted the final period.
During the 1950-1968 period, residents of the urban
core area were exposed to extremely high levels of
sulfur dioxide, suspended particulates, and suspended
sulfates. Based on the limited available data, the
annual average sulfur dioxide levels were estimated to
be 157 to 282 /ig/m3, and the total suspended
particulate levels to be 149 to 244 /Kg/in3, which is 2
to 3 times the relevant National Primary Air Quality
Standards.
Residents of suburban areas were perhaps more
fortunate, but they were also exposed to excessive
pollutant levels: estimated annual concentrations of
sulfur dioxide (approximately 100 Mg/m3) and sus-
pended particulates (118-174 ^g/m3) both exceeded
the present relevant National Primary Air Quality
Standards. On the other hand, residents of outstate
areas enjoyed better, but probably quite variable, air
quality: estimates for sulfur dioxide exposures were
generally below the National Primary Air Quality
Chicago-Northwest Indiana Studies
4-25
-------
Table 4.2.1. ESTIMATES OF EXPOSURE FOR CHICAGO-NORTHWEST INDIANA
METROPOLITAN REGION, 1950-1970
Pollutant and
community
Sulfur dioxide, jLig/m3
Urban core3
Suburban
Relatively clean
outstate
Total suspended parti-
culates, jug/m3
Urban core
Suburban
Relatively clean
outstate
Suspended sulfates, jug/m3
Urban core
Suburban
Relatively clean
outstate
Dustfall, g/m2/mo
Urban core
Suburban
Relatively clean
outstate
Pollutant levels over time
1950-59
—
-70
244
174b
~80
20.6b
—
< 8
19.4
—
-
1960-65
282
—
~70
177
141b
-80
17.3
—
<8
15.2
—
-
1966-68
157
100
43
149
118
76
14.1
—
7.7
14.2
—
-
1969-70
96
217C
19
155
103
71
14.5
—
9.3
12.6b
—
-
Values for Chicago proper utilized.
Projected values.
C1969data only.
Standard, while particulate exposures, including a
substantial agricultural background component, were
somewhat below the relevant national standard.
Since 1965, sulfur dioxide and particulate levels in
the urban core have been substantially reduced. By
1971, City of Chicago monitoring data indicated
sulfur dioxide levels were at or below the primary
standard. In 1971, suspended particulate levels (97
Mg/m3) were still somewhat above the National
Primary Air Quality Standard, even though they were
only half the levels monitored 10 years earlier. Air
monitoring data from suburban areas, however, did
not indicate the same dramatic improvement trend.
Characterization of the Study Population
Two-thirds of the 52,000 recruits were less than
21 years of age and only 5 percent were older than
24. Tests for intercommunity differences were lim-
ited to the 45,000 recruits who completed all
portions of the questionnaire. To quantify any
intercommunity bias caused by these exclusions, the
stepwise selection process was summarized for whites
and blacks (Tables 4.2.2 and 4.2.3). The largest
complete exclusion, roughly 15 percent of the popu-
lation, was caused by reluctance of the respondents
to record ethnic information. Only roughly 5 percent
4-26
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.2.2. WHITE RECRUITS INCLUDED IN AND EXCLUDED FROM PRINCIPAL ANALYSIS
DISTRIBUTED BY COMMUNITY AND REASON FOR EXCLUSION3
Community
Relatively clean
outstate
Suburbs
Urban core
Overall
Overall
sample
size
18,864
12,773
13,769
45,406
Percent
excluded because
of missing
ethnic data
13
12
19
15
Sample
size
16,436b
11,298
11,057
38,791
Percentage of white recruits
With complete data
Included
62
66
63
63
Partially
excluded because
of short residence
32
29
31
31
Excluded because of
missing data
Smoking
1
1
1
1
Chronic respiratory
disease symptoms
5
4
6
5
aNot included in the analyses are 19 recruits whose address was not listed and 126 members of miscellaneous racial groups.
Not included in the analyses are 2387 recruits from areas other than Illinois and Indiana.
Table 4.2.3. BLACK RECRUITS INCLUDED IN AND EXCLUDED FROM PRINCIPAL ANALYSIS
DISTRIBUTED BY COMMUNITY EXPOSURE AND REASON FOR EXCLUSION
Community
Relatively clean
outstate
Suburbs
Urban core
Overall
Sample
size
871
441
5422
6734
Percentage of black recruits
With complete data
Included
46
50
48
48
Partially
excluded because
of short residence
38
35
34
35
Excluded because of
missing data
Smoking
3
3
3
3
Chronic respiratory
disease symptoms
13
13
14
14
Chicago-Northwest Indiana Studies
4-27
-------
of the remaining white recruits and 15 percent of
black recruits were excluded because of failure to
answer questions on smoking habits or chronic
respiratory disease symptoms. The remaining recruits
were then divided into two groups: those who
reported no change of address during the preceding 3
years and those who reported a recent move. Anal-
yses of overall intercommunity differences in indices
of chronic respiratory morbidity were restricted to
the more stable group, while the analysis that sought
an exposure duration threshold utilized both stable
residents and newcomers. Black recruits appeared
somewhat more mobile than white recruits. Overall,
intercommunity exclusions were remarkably similar.
Intercommunity patterns in smoking history and
educational attainment were then considered for each
of the three exposure communities (Tables 4.2.4 and
4.2.5). There was a trend towards increased smoking
in recruits of both races who lived in urban core
areas. The excess in smokers was particularly striking
among urban blacks (Table 4.2.4). Urban residents,
whether smokers or nonsmokers, were also less likely
to have completed high school than suburban or
outstate recruits. Interestingly enough^ this trend was
closely related to smoking habits. Smokers in all
communities were less likely to have completed high
school, and this effect was most marked in the urban
core community.
Table 4.2.4. STUDY POPULATION FOR PRINCIPAL ANALYSIS DISTRIBUTED BY COMMUNITY,
RACE, AND SMOKING STATUS3
Community
Relatively clean
outstate
Suburbs
Urban core
Total
Blacks
Percent
nonsmokers
34.5
28.5
10.4
15.3
Percent
smokers
65.5
71.5
89.6
84.7
Total
789
411
4,169
5,369
Whites
Percent
nonsmokers
38.6
36.9
36.1
37.4
Percent
smokers
61.4
63.1
63.9
62.6
Total
10,713
7,676
7,390
26,779
aLimited to those living at same address for >3 years.
Table 4.2.5. EDUCATIONAL ATTAINMENT OF STUDY POPULATION
DISTRIBUTED BY COMMUNITY AND SMOKING STATUS
Percent who completed high school
Community
Relatively clean
outstate
Suburbs
Urban core
Blacks
Nonsmokers
80.9
77.8
75.0
Smokers
59.2
52.4
47.6
Whites
Nonsmokers
93.4
94.7
89.2
Smokers
82.3
83.0
68.5
4-28
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
HEALTH EVALUATION
Chronic Bronchitis
Prevalence rates for chronic bronchitis were dis-
tributed by age, and when no age trends were found
in the narrow range surveyed, prevalence rates for the
entire population were distributed by educational
attainment and smoking status (Table 4.2.6). Recruits
who only completed high school were found to have
somewhat lower chronic bronchitis rates than those
who either did not finish high school or went beyond
high school. As expected, cigarette smoking was
related to chronic bronchitis prevalence in a dose-
response fashion. Heavier smokers (>1 pack daily)
reported rates 5 times those found in nonsmokers.
Exsmokers, pipe smokers, and cigar smokers had rates
lower than cigarette smokers, but higher than lifetime
nonsmokers.
Residents of the more polluted urban core and
suburban communities were found to have somewhat
higher prevalence rates than residents of the rural
areas (Table 4.2.7). These differences were significant
in the smoking blacks and the nonsmoking whites
(Table 4.2.8). The difference in smoking whites was
nearly significant, the difference being due to a higher
rate in the suburbs. The pattern among the nonsmok-
ing blacks was mixed and not significant
Table 42.6. CHRONIC BRONCHITIS PREVALENCE DISTRIBUTED
BY SMOKING STATUS AND EDUCATIONAL LEVEL
Population
characteristic
Chronic bronchitis prevalence,
percent
Entire study population
Education3
Did not complete high school
Completed high school
Education beyond high school
Smoking
Never smoked
Exsmokers, or current pipe
or cigar smokers
Current cigarette smokers:
«1/2 pack/day)
(1/2 to 1 pack/day)
(>1 pack/day)
12.9
12.7
12.2
13.6
5.5
7.8
18.2
(8.0)
(10.5)
(28.1)
aThe chronic bronchitis prevalence for nonrespondents to this question was
14.6 percent.
The chronic bronchitis prevalence for nonrespondents to this question was
11.6 percent.
Chicago-Northwest Indiana Studies
4-29
-------
Table 4.2.7. CHRONIC BRONCHITIS PREVALENCE DISTRIBUTED BY COMMUNITY,
RACE, AND SMOKING STATUS
Community
Relatively clean
outstate
Suburbs
Urban core
Chronic bronchitis prevalence, percent
Blacks
Nonsmokers
9.0 (8.8)
9.4 (7.8)
9.4 (9.5)
Smokers
9.3 (8.8)
12.6(12.7)
12.9 (13.0)
Whites
Nonsmokers
4.3 (4.2)
5.5 (5.4)
5.2 (5.4)
Smokers
16.7 (17.6)
19.8 (18.8)
18.3(17.8)
Prevalence rates adjusted for education are given in parentheses.
Table 4.2.8. ANALYSIS OF VARIANCE FOR RESPIRATORY MORBIDITY MEASUREMENTS
DISTRIBUTED BY RACE AND SMOKING STATUS, CHRONIC BRONCHITIS PREVALENCE
Factor
Area
Smoking
level
Education
Fit of model
Degrees
of
freedom
2
2
2
4(20)a
Blacks
Nonsmokers
X2
0.50
-
0.45
2.83
P
0.7788
-
0.7985
0.5866
Smokers
X2
28.79
96.61
27.04
28.79
P
<0.0001
<0.0001
<0.0001
<0.0001
Whites
Nonsmokers
X2
10.23
-
1.43
1.10
P
0.0060
-
0.4892
0.8943
Smokers
X2
4.85
1298.65
109.28
35.24
P
0.0885
<0.0001
<0.0001
0.0189
First number for nonsmokers, second for smokers.
Mean Respiratory Symptom and Severity
Scores
As previously described, mean respiratory symp-
tom scores for each ethnic and smoking category were
compiled for each exposure community. These scores
were first averaged over the total population, and
then a severity score was calculated for those persons
who reported symptoms.
For the total study population, mean symptom
scores (Table 4.2.9) were generally lowest in the
relatively clean outstate communities, intermediate in
the suburbs, and highest in the urban core, which also
reported the highest pollution levels during most of
the study period. The only exception was among
white smokers, who showed almost no differences
after adjusting for educational differences. Inter-
community differences in symptom scores were
4-30
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.2.9. MEAN RESPIRATORY SYMPTOM SCORE DISTRIBUTED BY~COMMUNITY,
RACE, AND SMOKING STATUS
Community
Relatively clean
outstate
Suburbs
Urban core
Mean symptom score3
Blacks
Nonsmokers
2.30(2.10)
2.11 (2.24)
2.30 (2.34)
Smokers
2.50 (2.47)
2.62 (2.61)
2.72 (2.71)
Whites
Nonsmokers
1.78(1.76)
1.81 (1.82)
1.88(1.84)
Smokers
2.98 (2.90)
2.98 (2.93)
2.90(2.91)
Scores adjusted for education are given in parentheses.
significant for smoking blacks, and nearly significant
for nonsmoking whites (Table 4.2.10). Symptom
scores for the other groups were not significantly
different.
Mean severity scores were computed for those who
reported any chronic respiratory disease symptom
(Table 4.2.11). These scores had a mixed pattern, and
although there were significant differences by area
(Table 4.2.12), these differences were not always
consistent with the estimated pollution gradient. In
addition, the test for the fit of the model was
significant in two of the analyses, indicating some
significant interactions.
Table 4.2.10. ANALYSIS OF VARIANCE FOR RESPIRATORY MORBIDITY
MEASUREMENTS DISTRIBUTED BY RACE AND SMOKING
STATUS, MEAN SYMPTOM SCORES
Factor
Area
Smoking
level
Education
Fit of
model
Degrees
of
freedom
2
2
2
4(20)a
Blacks
Nonsmokers
X2
0.12
-
2.10
0.79
P
0.9418
-
0.3499
0.9398
Smokers
X2
8.72
320.91
55.70
24.61
P
0.0128
<0.0001
<0.0001
0.2168
Whites
Nonsmokers
X2
4.82
-
20.27
0.30
P
0.0898
-
0.0001
0.9898
Smokers
X2
0.92
2325.52
75.83
20.80
P
0.6313
<0.0001
<0.0001
0.4090
aFirst number for nonsmokers, second for smokers.
Chicago-Northwest Indiana Studies
4-31
-------
Table 4.2.11. MEAN SEVERITY SCORES FOR RECRUITS WITH RESPIRATORY SYMPTOMS
DISTRIBUTED BY COMMUNITY, RACE, AND SMOKING STATUS
Community
Relatively clean
outstate
Suburbs
Urban core
Mean severity scores3
Blacks
Nonsmokers
4.00 (4.02)
3.67 (3.63)
3.94 (3.93)
Smokers
3.66 (3.62)
3.94 (3.98)
3.91 (3.91)
Whites
Nonsmokers
3.97 (3.99)
4.20(4.18)
4.04 (4.05)
Smokers
4.18(4.21)
4.33 (4.32)
4.24 (4.19)
aScores adjusted for education are given in parentheses.
Table 4.2.12. ANALYSIS OF VARIANCE FOR RESPIRATORY MORBIDITY
MEASUREMENTS DISTRIBUTED BY RACE AND SMOKING
STATUS, MEAN SEVERITY SCORES
Factor
Area
Smoking
level
Education
Fit of model
Degrees
of
freedom
2
2
2
4(20)a
Blacks
Nonsmokers
X2
2.99
—
8.12
1.11
P
0.2242
-
0.0173
0.8927
Smokers
X2
11.27
72.54
23.58
44.55
P
0.0036
<0.0001
<0.0001
0.0013
Whites
Nonsmokers
X2
12.17
. -
33.23
2.24
P
0.0023
-
<0.0001
0.6917
Smokers
X2
18.04
326.24
265.64
55.47
P
<0.0001
<0.0001
<0.0001
<0.0001
aFirst number for nonsmokers, second for smokers.
Length of Residence
The effect of length of exposure to community air
pollution was also assessed by computing chronic
bronchitis prevalence rates for selected durations of
residence at the same address. Only white recruits
were used in these analyses because of the large
sample sizes needed for the additional categorization.
Figure 4.2.1 shows an increase in the <1 -year
residence category. Smokers from all areas showed a
decrease in rates for those with an intermediate
length of residence, such as 4 to 7 years. All areas
4-32
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
i I i r
NONSMOKERS
O —O URBAN CORE A
Q —Q SUBURBS
A^—A OUTSTATE
2-3 4-7 8-11
LENGTH OF RESIDENCE, years
Figure 4.2.1. Bronchitis prevalence by area
and length of residence. Prevalence rates for
smokers adjusted for level of smoking.
showed an increase in the > 12-year residence cate-
gory, and the urban core and suburbs showed the
greatest increase in this category. The pattern for
nonsmokers was mixed, although the increase in the
< 1-year residence still appeared. Both the urban core
and suburbs showed a sharp increase in the > 12-year
residence category, whereas the outstate areas did
not.
Table 4.2.13 gives the relative effects of smoking
and ambient air quality. These differences in preva-
lences were computed only for whites because of
small sample sizes for blacks in some of the areas.
Smokers appear to have an excess prevalence of about
13 percent, or about four times the rate for nonsmok-
ers. Residents of the urban core and suburbs each had
an excess prevalence of 1.2 percent, or approximately
1.3 times the rate for outstate areas. The combined
effects of increased ambient air pollution and smok-
ing appeared to be additive.
DISCUSSION
Military recruits from more polluted metropolitan
areas were found to have increases in two measures of
chronic respiratory disease frequency. These increases
were noted for smokers and nonsmokers of both
major ethnic groups, although not all of the differ-
ences were statistically significant. Residence for 11
or more years in the urban core community appeared
to be associated with an increase in chronic bronchitis
prevalence. As expected, cigarette smoking was asso-
ciated in a dose-response fashion with the prevalence
of chronic bronchitis. Air pollution and cigarette
smoking seemed to exert additive effects upon
chronic bronchitis prevalence, although the pollution
effects were not strong enough to determine this
accurately.
Newcomers to an area clearly showed increases in
bronchitis prevalence. This effect usually dropped off
in the 2- to 3-year residence category. The effect of
pollution possibly can be detected as early as 8 to 11
years, but the largest differences occur in the
> 12-year category. This peculiar "U" shaped
distribution in the polluted areas may be partially
explained by short-lived irritant symptoms induced in
newcomers. The human body may then develop a
tolerance to irritant effects that is not overwhelmed
until exposures persist for at least 12 years. Excess
morbidity among recent arrivals might also be related
to a general increase in morbidity, which can in turn
be related to changes in life style.19
The possible effects of six other intervening
variables should be considered. These variables are the
reliability of pollution estimates, possible effects of
indoor pollutants, apparent deficit of symptoms
among residents of the urban core community,
socioeconomic status of families, ethnic differences
within the white racial grouping, and the effect of
exclusions. The procedures and rationale utilized in
estimating pollutant exposures are detailed else-
where.18 Estimates are more reliable for the urban
core than for suburban and outstate areas. Actual
exposures to many residents of the urban core areas
may have been somewhat lower than estimated from
the available monitoring data. Existing area-wide
models were available for a recent year, but they were
in substantial disagreement with each other and with
monitoring data from suburban stations. In general,
models seemed to underestimate suburban exposures.
The exposure estimates in this paper are intended as
useful approximations from which to estimate health
hazards; they are not precise exposure doses. In
Chicago-Northwest Indiana Studies
4-33
-------
Table 42.13. RELATIVE IMPORTANCE OF CIGARETTE SMOKING
AND AMBIENT AIR QUALITY AS DETERMINED BY COMPARISON
OF EXCESS PREVALENCE FOR CHRONIC BRONCHITIS3
Current
smoking
status
Nonsmokers
Smokers
Community
Outstate
Suburban
Urban core
Outstate
Suburban
Urban core
Excess
prevalence
of chronic
bronchitis
0.0 (4.2)
1.2
1.2
13.4
14.6
13.6
Simple additive
model combining
adverse effects of
smoking and pollution
Expected
-
-
t
-
14.6
14.6
Observed/
expected
-
-
-
-
1.00
0.93
aOnly whites were considered since sample sizes for blacks were insufficient.
Base rates are given in parentheses. Excess prevalence equals smoking and community-specific
prevalence rate minus base rate.
addition, indoor levels of pollutants were not meas-
ured in the present study. Side-stream cigarette
smoke and nitrogen dioxide from the use of gas for
cooking or space heating may be significant indoor
hazards. Domestic use of gas and cigarette smoking
would be expected to be more frequent in the urban
communities, but their combined effect would hardly
be expected to account for the large differences
attributed to ambient air pollution.
If one accepts that ambient air pollution induces
or makes manifest a constitutional predisposition
towards chronic respiratory disease, then higher
symptom rates should have been observed in the
urban core. A possible hypothesis is that ambient
pollution levels in the suburbs exceeded the threshold
necessary for the appearance of respiratory symptoms
and that further increments in pollution would have
their effects by increasing the prevalence and severity
of symptoms in later life. To some extent, this
speculation was supported by increases in the overall
mean symptom scores for urban core residents.
Another factor that may influence symptom re-
porting is the socioeconomic status of the family of
the recruit. Lower social status might be accompanied
by real increases in disease coupled with a greater
reluctance to report symptoms. The overall effect of
such variation was in part controlled by consideration
of racial, smoking, and educational variables. In any
case, the net effect of social differences would be
opposite for the poorer urban core and the more
prosperous suburban communities. With this in mind,
one can hardly postulate that pollution-related area
differences would disappear. The effects of differ-
ences between ethnic subgroups in the study popula-
tion cannot be quantitated, but they are unlikely to
approach the magnitude of the reported intercom-
munity differences.
A more difficult problem is that of exclusions.
Failure to complete the study questionnaire probably
resulted in elimination of large numbers of less well
educated recruits. Some estimates can be made by
examining the responses of recruits who completed a
4-34
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
portion of the questionnaire. For example, chronic
bronchitis rates of those who did not classify them-
selves by race were identical to rates for whites and
rates for those who refused to answer the smoking
question were like light smokers. In other words,
examination of incomplete responses revealed nothing
that would diminish intercommunity differences.
After the effects of intervening variables were
considered, we concluded that it would be prudent to
protect the public health by assuming that increases
in chronic bronchitis prevalence can be related to
ambient air pollution levels characterized by annual
average sulfur dioxide levels of 96 to 217 /ug/m3
accompanied by annual average suspended particulate
levels of 103 to 155 ;ug/m3 and annual average
suspended sulfate levels of about
SUMMARY
Chronic respiratory disease prevalence rates among
military recruits who lived in more polluted urban
core and suburban areas of Chicago were significantly
higher than chronic bronchitis rates in recruits from
rural and smaller urban areas in the outstate Illinois-
Indiana area. Excesses in morbidity were noted
among blacks, whites, smokers, and nonsmokers,
although not all were statistically significant. The
hazards of ambient air pollution and self-pollution by
cigarette smoking seemed additive. Available evidence
indicates that exposures lasting 12 years or more to
ambient air pollution characterized by elevated an-
nual average levels of sulfur dioxide (96 to 217
jug/m3), suspended particulates (103 to 155 jug/m3),
and suspended sulfates (14 jug/m3) were accompanied
by significant increases in the frequency of chronic
respiratory disease symptoms.
REFERENCES FOR SECTION 4.2
1. Higgins, I.T.T. Tobacco Smoking, Respiratory
Symptoms, and Ventilatory Capacity; Studies in
Random Samples of the Population. Brit. Med. J.
5118:325-329, 1959.
2. Ashford, J.R., S. Brown, D.P. Duffield, C.S.
Smith, and J.W.J. Fay. The Relationship between
Smoking Habits and Physique, Respiratory
Symptoms, Ventilatory Function, and Radiologi-
cal Pneumoconiosis amongst Coal Workers at
Three Scottish Collieries. Brit. J. Prevent. Soc.
Med. 75:106-117, 1961.
3. Ferris, B.C. and D.O. Anderson. The Prevalence
of Chronic Respiratory Disease in a New Hamp-
shire Town. Amer. Rev. Respiratory Dis.
56:165-177, August 1962.
4. Hyatt, R.E., A.D. Kistin, and T.K. Mahan.
Respiratory Disease in Southern West Virginia
Coal Miners. Amer. Rev. Respiratory Dis.
89:387-40 I.March 1964.
5. Van der Lende, R., R. Ter Brugge, J.P.M.
DeKroon, H.J. Sluiter, G.J. Tammeling, K. De-
Vries, and N.G.M. Orie. The Organization of an
Epidemiologic Investigation. T. Soc. Geneesk.
44:148-157, 1966.
6. Huhti, E. Prevalence of Respiratory Symptoms,
Chronic Bronchitis, and Pulmonary Emphysema
in a Finnish Rural Population; Field Survey of
Age Group 40-64 in the Harjavalta Area. Acta
Tuberc. Scand. Suppl. 61:1-111, 1965.
7. Payne, M. and M. Kelsberg. Respiratory Symp-
toms, Lung Function, and Smoking Habits in an
Adult Population. Amer. J. Public Health.
54:261-277, February 1964.
8. Holland, W.W., T. Halil, A.E. Bennett, and A.
Elliott. Factors Influencing the Onset of Chronic
Respiratory Disease. Brit. Med. J. 2:205-208,
1969.
9. Addington, W.W., R.L. Carpenter, J.F. McCoy,
K.A. Duncan, and K. Mogg. The Association of
Cigarette Smoking with Respiratory Symptoms
and Pulmonary Function in a Group of High
School Students. J. Oklahoma State Med. Assoc.
63:525-529, November 1970.
10. Seely, I.E., E. Zuskin, and A. Bouhuys. Cigarette
Smoking: Objective Evidence for Lung Damage
in Teenagers. Science. 772(3984): 741-743, May
14, 1971.
11. Holland, W.W. and R.W. Stone. Respiratory
Disorders in United States East Coast Telephone
Men. Amer. J. Epidemiol. 52:92-101, 1965.
12. Deane, M., J.R. Goldsmith, and D. Tuma. Res-
piratory Conditions in Outside Workers. Arch.
Environ. Health. 70:323-331, 1965.
13. Holland, W.W. and D.D. Reid. The Urban Factor
in Chronic Bronchitis. Lancet. 7:445-448, Febru-
ary 27, 1965.
14. Reid, D.D. Environmental Factors in Respiratory
Disease. Lancet. 7:1237-1242, June 14, 1958.
15. Reid, D.D. Environmental Factors in Respiratory
Disease. Lancet. 7:1289-1294, June 21, 1958.
16. Questionnaires Used in the CHESS Studies. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Chicago-Northwest Indiana Studies
4-35
-------
Protection Agency. Research Triangle Park, N.C. 1950-1971. In: Health Consequences of Sulfur
Publication No. EPA-650/1-74-004. 1974. Oxides: A Report from CHESS, 1970-1971. U.S.
17. Grizzle, I.E., C.F. Starmer, and G.G. Koch. Environmental Protection Agency. Research
Analysis of Categorical Data by Linear Models. Triangle Park, N.C. Publication No. EPA-
Biometrics. 2J(3):489-504, September 1969. 650/1-74-004. 1974.
18. Hinton, D.O., T.D. English, B.F. Parr, V. Hassel- 19. Rahe, R.H., J.D. McKean, and R.J. Arthur. A
blad, R.C. Dickerson, and J.G. French. Human Longitudinal Study of Life-change and Illness
Exposure to Air Pollutants in the Chicago- Patterns. J. Psychosom. Res. 70:355-366, 1967.
Northwest Indiana Metropolitan Region,
4-36 HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
4.3 PROSPECTIVE SURVEYS OF ACUTE RESPIRATORY
DISEASE IN VOLUNTEER FAMILIES:
CHICAGO NURSERY SCHOOL STUDY, 1969-1970
John F. Finklea, M.D., Dr. P.H., Jean G. French, Dr. P.H.,
Gene R. Lowrimore, Ph.D., Julius Goldberg, Ph.D.,
Carl M. Shy, M.D., Dr. P.H., and William C. Nelson, Ph.D.
4-37
-------
INTRODUCTION
A question of current interest is whether control
of sulfur dioxide air pollution is likely to be
accompanied by reductions in acute respiratory
morbidity. Sulfur dioxide levels in our major urban
areas are being significantly reduced by stringent,
costly air pollution control measures. Acute res-
piratory illness is an important, multibillion dollar
public health problem that is the leading cause of
acute morbidity in our nation.1'2 Substantial health
benefits might result if modest decreases in the
frequency or severity of acute respiratory disease
were to accompany improvements in air quality.
Another, possibly more important benefit is that
reducing the frequency or severity of acute respira-
tory disease might lessen the chance of developing
chronic bronchitis in later life.3 Stringent control of
ambient air pollution, cessation of self-pollution with
tobacco smoke, and continued improvements in
industrial hygiene offer the greatest promise for the
primary prevention of chronic bronchitis.
A number of previous studies attributed excesses
in acute respiratory illness to ambient air pollution
characterized by elevated levels of sulfur dioxide and
suspended particulates.3"6 Increases in illnesses in-
volving the lower respiratory tract were more consist-
ently observed than increases in illnesses involving the
upper respiratory tract. One study also linked urban
pollution exposures to increased morbidity during an
influenza epidemic.7 Previous investigations have
been hampered by the inability to measure pollutant
exposures. All have failed to ascertain the duration of
exposure needed to induce susceptibility or the
persistence of susceptibility once it has been induced.
This dose-response information is needed to achieve a
level of environmental quality that will protect
human health but not be needlessly costly.
Acute respiratory illness can best be assessed in
nuclear families. Within a nuclear family, children
attending nursery school manifest the highest inci-
dence rates for acute respiratory illness.8 Quantifying
the effects of air pollution on acute respiratory
disease in families and disentangling the effects of
pollutants from those of age, socioeconomic status,
and self-pollution by cigarette smoking is a complex
task A succession of investigations utilizing standard-
ized techniques but separated in time and space
would be needed to accumulate the mosaic of
necessary information. The City of Chicago was
chosen for one of the earliest studies because rela-
tively complete air monitoring data allowed estima-
tion of air pollution exposures and because air
pollution control efforts began to achieve recog-
nizable reductions in sulfur dioxide and total sus-
pended particulates.
The present study sought to detect evidence of
decreases in acute respiratory illness frequency or
severity that might have accompanied recent reduc-
tions in ambient air pollution levels. It also attempted
to ascertain which family segment was most suscepti-
ble to the deleterious effects of air pollutants as
indexed by increased vulnerability to acute res-
piratory disease. The importance of socioeconomic
status, residential mobility, and self-pollution by
cigarette smoking as determinants of acute respira-
tory illness was assessed to assure the validity of any
morbidity excess attributed to ambient air pollution.
METHOD
Setting and Study Population
The study population consisted of families of
nursery school children aged 2 to 5 years who were
attending day care centers throughout the City of
Chicago. The sample was not intended to be repre-
sentative of the general population but rather to
provide sufficient numbers of families with both
young and older children so that effects of pollution
control measures could be better assessed. Middle-
class residential neighborhoods such as those con-
sidered in this study do represent a large proportion
of the population and are apt to be more homoge-
neous with respect to social class distribution. Also,
neighborhood characteristics change more slowly
than in central city neighborhoods.
Those day care centers selected for study were
those inside Chicago city limits and situated within
1.25 miles of an air monitoring station. Letters were
sent to the parent or guardian of children currently
enrolled in each school requesting that the total
family enroll in the study. For those who agreed to
participate, a family census was obtained along with
information about age of family members, crowding,
education, occupation, smoking, and use of gas in the
home for cooking or for heating.
Illness Monitoring
Volunteer families were called once every 2 weeks
and questioned about the presence of illness, fever,
respiratory symptoms, restricted activity, middle ear
4-38
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
infection (otitis media) diagnosed by a physician, and
other physician consultation visits. A standardized
questionnaire9 was used to obtain the information,
and telephone interviews were conducted by trained
interviewers. After the telephone interviews were
completed, all the questionnaires were reviewed for
incorrect or missing information and necessary cor-
rections were made. Whenever possible, information
on symptoms was obtained from the mother; the
presence or absence of illness was left to her
discretion. The mother was asked whether or not any
new illness had occurred in the past 2-week period. If
there was an affirmative response, additional informa-
tion was elicited concerning the presence of fever,
restriction of activity, whether or not lower or upper
respiratory symptoms were present, and whether a
physician was consulted. If a member of the family
had visited a physician, the informant was asked
whether or not the physician had diagnosed otitis
media. The upper respiratory illness classification was
given to those reporting any or all of the following:
cough (dry, nonproductive), head cold, sore throat,
sinus or postnasal drip, or runny nose. Lower
respiratory illnesses were limited to chest colds with a
persistent, productive cough, bronchitis, or pneu-
monia. In the present study, the definition of acute
lower respiratory illness was much more restrictive
than that used in a later study in New York where
any illness with both upper and lower tract symptoms
was classed as an acute lower respiratory infection.
To evaluate the validity of parental diagnosis, a small
sample of parents' replies were compared with physi-
cians' records. A total of 100 physicians were
contacted concerning information on 133 individual
visits.
Air Monitoring
The air monitoring information for this study is
based on measurements from 20 sampling stations
located throughout the City of Chicago.10 Total
suspended particulates (TSP) were measured at each
station three times a week, and sulfur dioxide was
measured twice a week. Total suspended particulate
levels were initially assumed to be representative of
those census tracts having 20 percent or more of their
area inside a circle with a radius of 1.25 miles
originating at the sampling station. Monthly arith-
metic mean values as well as yearly arithmetic mean
values were computed for each of the areas for the
year 1969. Each station was assigned to one of four
quartiles (Low, Intermediate, High, Highest) based on
the monitored suspended particulate levels for 1969,
and analyses were made on the basis of this partition-
ing. Air monitoring data obtained during the 1970
period of study raises questions about the appropri-
ateness of the groups selected, but the decision was
made to proceed with the evaluation of the data on
the basis of these groups. This was considered to be
reasonable because during 1971 only one station
from the Intermediate group and one from the High
group seemed to be misassigned on the basis of sulfur
dioxide levels, and these were properly assigned on
the basis of total suspended particulates. Since total
pollution is the factor that affects human health, any
other grouping of the communities also would be
questionably appropriate. As each family entered the
study, it was assigned to a census tract determined by
the family's residential address. If the census tract fell
within the area of a monitoring station, the family
was then assigned to the appropriate quartile. If the
census tract did not fall within the area of a
monitoring station, it was then plotted on a map of
the City of Chicago to determine its proximity to the
various stations. Those families in the immediate area
of a station were included in the study and assigned
an appropriate number. The two lower exposure
quartiles were very similar with respect to 1969
pollution concentrations of total suspended particu-
lates and sulfur dioxide. These populations were
combined for the morbidity analyses to gain a larger
population in the cleaner areas.
The three exposure categories grouped according
to the 1969 levels of total suspended particulates
were:
1. Intermediate exposure (combined Low and Inter-
mediate quartiles), where sulfur dioxide levels were
generally within the limits set by the National
Secondary Air Quality Standard for sulfur dioxide
and suspended particulate levels were below the
median annual level for Chicago stations.
2. High exposure, where sulfur dioxide levels were
within the National Secondary Air Quality Stand-
ard but suspended particulate levels were above
the median level for Chicago stations and averaged
1-1/2 times the relevant National Primary Ambient
Air Quality Standard.
3. Highest exposure, where sulfur dioxide levels
exceeded the National Primary Ambient Air Qual-
ity Standard and suspended particulates were
approximately twice the relevant National Primary
Air Quality Standard.
Chicago-Northwest Indiana Studies
4-39
-------
Statistical Analysis
The statistical analysis involved five steps. First,
selected characteristics of the family groupings that
comprised each exposure category were compared to
detect gross socioeconomic and demographic differ-
ences that might bias the study. Second, temporal
trends in mortality reporting were evaluated to isolate
any epidemic outbreaks for special analysis. Third,
intercommunity differences in illness frequency were
assessed by relative risk models. Fourth, differences
in the severity of illness were quantified and indices
of excess morbidity computed. Fifth, the impact of
selected intervening variables, including social status,
mobility, and cigarette smoking, was evaluated.
The analytical protocol allowed partition of the
entire study period into several seasonal segments so
that any epidemic outbreak might be investigated
separately. This procedure also minimized the ad-
verse effects of family withdrawals or temporary
nonparticipation since the experience of a family was
not considered in the analysis if contact could not be
made during any portion of a particular seasonal
segment. Three segments were identified for the
present study. The first season, which lasted from
.December 14, 1969, to March 21, 1970, encompassed
an epidemic caused by A2/Hong Kong influenza. A
second, noninfluenza respiratory disease season in-
cluded two periods, the early spring of 1970 and the
fall of 1970, and ended just after Thanksgiving. The
third (summer) segment lasted from May 31, 1970,
through September 19, 1970.
When assessing illness frequency, attack rates for
upper, lower, and total acute respiratory illness were
computed separately for five family segments: fath-
ers, mothers, older siblings, nursery school children,
and younger siblings. The experience of other adults
living in the homes was excluded from the study. To
ascertain the population reporting repeated bouts of
acute respiratory illness, the illness experience of each
volunteer was considered. Repeated excessive illness
was defined as more than three illnesses involving the
upper respiratory tract, more than one illness involv-
ing the lower respiratory tract, or more than four
total acute respiratory illnesses. Intercommunity dif-
ferences in the frequency of acute respiratory illness
were ascertained by comparing two measures: attack
rates and proportion of the study population report-
ing excess respiratory illness.
Intercommunity differences in illness severity were
ascertained by comparison of severity scores that
ranged from Level I through Level V. Briefly, Level I
was illness without restricted activity, Level II in-
cluded restricted activity but not fever, Level III
included restricted activity and fever, and Level IV
included all of the preceeding plus a physician visit.
Level V applied only to children and was limited to
illnesses characterized by restricted activity, fever,
and middle ear infection diagnosed by a physician.
Level V was chosen to reflect a common, presumably
bacterial, complication of acute upper respiratory
illness. Levels II through V were further characterized
'by the mean number of restricted activity days
incurred.
The primary statistical tools used in this report
were the techniques developed by Grizzle et al.11 for
the analysis of categorical data. These techniques
permit one to treat categorical data in a manner
similar to the way continuous data are treated by the
theory of least squares analysis. Since we wanted a
conservative Type II error while maintaining a reason-
able Type I error, a significance level of p < 0.10 was
chosen.
RESULTS
Air Pollution Exposures
Health impairments attributed to air pollution
might be related to either current or prior exposures
and each of these could be related to either imme-
diate or persistent impairments. Effects of recent or
remote exposures might in turn be related to repeated
short-term elevations of pollutants or to longer term
annual average levels. Annual average air pollution
exposure estimates during and for the decade prior to
this study were available. However, it was not
possible to evaluate recent or remote short-term peak
exposures to air pollutants because constructing an
aggregate frequency distribution for the multiple
stations comprising each exposure stratum proved
very difficult. The network of stations was designed
to assess long-term exposures, and sampling was
restricted to 2 to 3 days each week. There was also
some unavoidable lack of comparability among moni-
toring data from stations within particular exposure
strata because of differences in measurement tech-
niques used through the preceeding decade. However,
appropriate scale factors have been utilized to provide
a common base for comparison.
Within the confines stated above, reasonable esti-
mates of annual average air pollutant exposures could
nevertheless be made for sulfur dioxide and total
suspended particulates (Table 4.3.1). No matter what
4-40
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.3.1. ESTIMATED ANNUAL AVERAGE POLLUTION LEVELS
IN CHICAGO STUDY COMMUNITIES
Pollutant
Sulfur
dioxide
Total
suspended
particulate
Community
Intermediate
High
Highest
Intermediate
High
Highest
Years prior
6-10
(1959-63)
130
130
250
123
140
165
Concentration,
to study
1-5
(1964-68)
109
107
170
121
137
163
ug/m3
During study
(1969-70)
57
51
106
111
126
151
exposure stratum was considered, those children who
were in elementary school during the study had been
exposed to excessive levels of sulfur dioxide (130 to
250 jug/m3) and suspended particulates (123 to 165
/ug/m3) during their early years of life. Children less
than six who lived in the Intermediate and High
exposure communities were exposed to levels of
sulfur dioxide of 107 to 109 /ug/m3 during their early
childhood years. Children from the Highest exposure
community were exposed to sulfur dioxide levels of
170 jug/m3. During the study years (1969-1970),
children living in the Intermediate and High exposure
communities were exposed to sulfur dioxide levels
(51 to 57 jug/m3) that had fallen below the relevant
National Secondary Ambient Air Quality Standard.
Their counterparts in the Highest exposure com-
munity continued to be exposed to elevated levels
(106 jUg/m3) of sulfur dioxide, even though these
levels were appreciably reduced from concentrations
of the earlier decade. Suspended particulate levels in
every exposure community were highest during the
earlier periods considered, and their subsequent de-
creases were more modest. Levels for each of the
exposure communities exceeded the relevant National
Primary Air Quality Standard throughout the 12-year
period. There were important intercommunity differ-
ences in total suspended particulate levels that made
further pooling of study populations inadvisable.
During the study, suspended particulate levels in the
Intermediate community (111 /ug/m3) were modestly
elevated, while those in the two more polluted
communities exhibited stepwise increases (126 and
151/ug/m3).
Quantitatively less information is available for
other measured ambient air pollutants. Annual aver-
age suspended sulfate levels were elevated (estimated
as 18 jug/m3 for 1959-1963, 14.4 jug/m3 for
1964-1968, and 14.5 /ug/m3 for 1969-1970), but
probably varied little between community groupings.
Suspended sulfate levels appeared to decrease during
the decade prior to the study. Percentagewise, this
decrease in suspended sulfates resembled the fall in
sulfur dioxide levels more than the gradual decline in
suspended particulates. Suspended nitrate levels in
the area were relatively low (1 to 2 /ug/m3 during
1960-1967).
Characteristics of the Study Population
Intercommunity differences involving intervening
variables that might require isolation or special
adjustments during the principal analyses were evalu-
ated (Table 4.3.2). Restriction of recruitment to
families with children in nursery school was itself a
factor that ensured a certain amount of socioeco-
nomic homogeneity. The study subjects in each
community were well educated, but a greater propor-
tion of fathers in the Highest exposure community
had attended graduate school. With the exception of
a higher number of professionals in the highest
pollution sector, there were no major intercom-
munity differences in the occupation of the fathers.
In every community, the great majority of families
cooked with gas, a practice that may cause elevated
indoor levels of nitrogen dioxide. Cigarette smoking
Chicago-Northwest Indiana Studies
4-41
-------
Table 4.3.2. PANEL CHARACTERISTICS
Characteristic
Educational attainment
of fathers
High school
> 4 yr college
Families using domestic
gas
Current cigarette
smokers
Mothers
Fathers
Families moving during
past 3 yr
Percent of indicated segment
by community
Intermediate
92
40
83
40
33
46
High
90
42
87
38
29
44
Highest
94
64
84
42
36
56
seemed somewhat less frequent among parents living
in the High exposure community, and families living
in the Highest exposure communities reported greater
residential mobility. There were essentially no inter-
community differences in smoking intensity of those
with that habit. When one compared only the
Intermediate and Highest communities, the only
differences were in residential mobility and educa-
tional attainment. These covariates were given special
attention in the analysis.
A total of 627 families representing 2705 indi-
viduals participated in the study, with the number of
families participating in each community differing
substantially (Table 4.3.3). The Intermediate expo-
sure community had the most families (335) and the
High area had the fewest (72). The number of
families considered in the analysis is somewhat less
than this due to attrition during the study.
Physicians' offices were surveyed to validate physi-
cian visits reported to the telephone interviewer.
Four-fifths, or 106, of the reported visits could be
identified in physician records and mother and
physician concordance could be examined. For all
respiratory symptoms, a 95.3 percent concordance
was obtained; for upper respiratory infection, 80.2
percent; for presence of fever, 75.5 percent; and for
earache, 85.8 percent. There was no reason to believe
there was any difference among communities in
illness reporting.
Seasonal Variation and Epidemic Influenza
The study began November 1969 and terminated
November 1970. A summer component was included
to allow calculations of estimates that would be
useful in evaluating similar studies that are restricted
to school months, i.e., the season of highest incidence
in the United States. Acute respiratory disease attack
rates for each of the study communities during each
reporting period were plotted to define epidemic
outbreaks or serious reporting anomalies (Figure
4.3.1). Excessively high illness reporting during the
initial calling period, not uncommon in panel studies,
probably represented overreporting during a period
when the telephone interview is viewed as a special
event rather than another routine task. For this rea-
son, data for the first three reporting periods were
not included in the analysis.
One month after the study began, an epidemic
wave of illness struck every community. The out-
break, subsequently associated with A2/Hong Kong
influenza, burned slowly through the city before
subsiding in early March to baseline attack rates for
the respiratory season. Thereafter, illness rates fell to
seasonal lows in July and rose sharply when school
resumed in September. The less polluted Intermediate
community generally manifested the lowest attack
rates during the respiratory seasons (influenzal and
noninfluenzal), but no noteworthy intercommunity
4-42
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
The mother or female guardian was instructed to
answer the questions for parents living in the home.
Other information ascertained included length of
residence in the community, educational attainment
of head of household, cigarette smoking habit, age,
and whether the parent had occupational exposure to
respiratory irritants. It should be made clear that
covariate information was applicable to the time at
which the survey was conducted. For example,
exsmokers were individuals who at the time of the
survey were not smoking, but who had smoked in the
past.
Assessing Air Pollution Exposure
At the time of this study, air monitoring stations
were located within each community. Each station
was on a building rooftop at a height of about 25 feet
above the ground. At each station, 24-hour integrated
samples of sulfur dioxide (modified West-Gaeke
method), total suspended particulates (high-volume
samplers), suspended sulfates, and suspended nitrates
were monitored daily. Daily measurements of nitro-
gen dioxide (Jacobs-Hochheiser method) were
also made; however, the validity of the method has
subsequently been questioned by the Environmental
Protection Agency. Dustfall was determined from
monthly samples. A full description of the monitor-
ing station locations and aerometric procedures has
been presented elsewhere.1 s
4
While these stations allow estimates of pollutant
exposure for prospective studies, it was also necessary
to estimate exposure for past time periods. Sulfur
dioxide had been monitored in the Low and Inter-
mediate I communities since 1965 and in the High
community since 1970 by the Utah State Division of
Health. In addition, measurements of total suspended
particulates and suspended sulfates dating back to
1953 were available for the Intermediate I com-
munity. Estimates of sulfur dioxide, total suspended
particulates, and suspended sulfate concentrations in
the High exposure community for 1940-1970 and the
Intermediate II exposure community for 1950-1970
were obtained by a mathematical dispersion model,
which utilized emissions from the industrial source
and extensive local meteorological data, and by
observed relationships among pollutants. Observed
suspended particulate, suspended sulfate, and sulfur
dioxide concentrations for 1970-1971 were used to
calibrate the models used to estimate exposure levels
for previous years.
Data Analysis
For the purposes of statistical analysis, a scale of
symptom severity was devised by a panel of physi-
cians based upon the number, clinical significance,
and duration or persistence of symptoms. Although
the number of symptom categories was somewhat
arbitrary and their severity ranking was subjective and
no more than ordinal in nature, the following ranking
of chronic respiratory symptoms categories was de-
veloped:
1. No symptoms.
2. Cough alone for less than 3 months each year.
3. Phlegm with or without cough for less than 3
months each year.
4. Cough without phlegm for 3 months or more
each year.
5. Phlegm without cough for 3 months or more
each year.
6. Cough and phlegm for 3 months or more each
year.
7. Cough and phlegm for 3 months or more each
year and shortness of breath.
Reported respiratory symptoms were used to
devise three indices of morbidity; first, chronic
bronchitis was defined in accordance with the British
Medical Research Council (cough and phlegm on most
days for at least 3 months each year), and prevalence
rates were compared across areas. The second variable
was a mean respiratory symptom score defined as the
arithmetic average of the respiratory symptom score
of each member of a particular population group. The
third variable was the arithmetic average of the
respiratory symptom score for those members of the
population who reported any chronic respiratory
symptoms, that is, for those members who had
symptom scores 2 through 7 only. The first two
variables reflect frequency of chronic respiratory
disease symptoms, with the mean symptom score
having the advantage of considering a gradient of
symptoms but the disadvantage of lacking the ac-
ceptance and the assured clinical relevance of clas-
sically defined chronic bronchitis. The mean severity
score, which is the third variable, was utilized as a
measure of severity of reported symptoms.
Hypotheses were tested using a general linear
model for categorical data.16 This technique uses
weighted regression on categorical data and allows
estimation of each effect in the model after adjusting
for all other effects in the model. Hypotheses were
tested by Chi Square procedures. Covariates used in
Salt Lake Basin Studies
243
-------
INFLUENZA
SEASON
12/14/69-3/21/70
NONINFLUENZAL
RESPIRATORY SEASON
(FIRST PERIOD)
3/22/70-5/30/70
SUMMER-SEASON
5/31/70-9/19/70
NONINFLUENZAL
RESPIRATORY SEASON
(SECOND PERIOD)
9/20/70-11/28/70
• HIGHEST
• HIGH
A INTERMEDIATE
8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58
TIME, week of study
Figure 4.3.1. Acute respiratory illness attack rates versus time, December 1969
to November 1970.
reported in the Highest community. On the other
hand, attack rates for upper tract illness and total
acute respiratory illness were not elevated in the
Highest community at that time. Since the analyses
indicated a consistent pattern of variance in respira-
tory illness rates during both epidemic and nonepi-
demic periods, subsequent analyses included data
from the entire study period.
Overall Illness Frequency
For each community, overall acute respiratory
disease attack rates per 100 person-weeks at risk were
computed separately for each of the five family
segments. Sets of relative risks were then calculated
for each family segment using the Intermediate
community as a base. With a longer time span, it was
possible to assess the experience of the relatively
limited population that comprised the High com-
munity. Even then, illness rates for this group were
somewhat unstable since these rates were based upon
only 850 to 1600 person-weeks of risk. Families from
the Highest community consistently reported exces-
ses in total respiratory illness and even larger excesses
in lower respiratory illness, the only exception being
infants, who reported no excess lower tract disease
(Table 4.3.5). Rates for the High community were
much more variable but tended to be higher than
those in the Intermediate community. One might
conclude that the more polluted groupings tended to
have more total illness, and an excess in the poten-
tially more dangerous illnesses involving the lower
respiratory tract.
Illness Severity
Each illness was assigned a severity score based
upon the previously defined five-level severity scale.
To facilitate intercommunity comparisons., sets of
relative severity scores were calculated for each
family segment. The procedure was identical to that
utilized in the relative risk model except that the base
severity score not an attack rate (Table 4.3.6).
ative severity scores were quite variable, and
4-44
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.3.4. RELATIVE RISK OF ACUTE RESPIRATORY ILLNESS DURING
A2/HONG KONG INFLUENZA OUTBREAK
Family
segment
Fathers
Mothers
Older
siblings
Nursery
school
children
Younger
siblings
Community
Intermediate
Highest
Intermediate
Highest
1 ntermediate
Highest
Intermediate
Highest
Intermediate
Highest
Relative risk of
acute respiratory illness3
One
episode
1.00 (28)
1.46
1.00 (46)
1.24
1.00 (45)
1.20
1.00 (69)
1.07
1.00 (74)
1.08
More than
one episode
1.00(6)
0.44
1.00(12)
0.74
1.00(14)
1.45
1.00 (28)
0.97
1.00 (34)
1.51
Base rate, that is percentage ill during outbreak, is in parentheses.
pollutant-attributable excesses in severity were found
only for illnesses involving the lower respiratory tract
in fathers and school children.
Illnesses occurring in the more polluted groupings
were as severe as illnesses among families living in the
less polluted Intermediate exposure area. As prev-
iously discussed, strict definition of acute lower
respiratory illness was reflected in higher severity
scores when contrasted to severity scores for upper
tract illness. The average number of restricted activity
days was also calculated for upper and lower tract
illnesses of a given severity level (Table 4.3.7). Since
the patterns indicated by the data were more con-
sistent when residentially stable families were con-
sidered, the analysis in Table 4.3.7 is restricted to
families that had not moved within the past 3 years.
Effects of residential mobility are discussed in more
detail in a later section. Increasing severity scores
from Level II to Level III did not increase the number
of restricted activity days in upper tract illnesses
among children of any age or in lower tract illness
among school children. A partial explanation might
be that febrile upper tract illnesses are more often
bacterial and thus more readily respond to antimicro-
bial therapy than do some of the more lingering
infections produced by other agents.
The excess number and excess percentage of
restricted activity days attributable to living in the
Highest exposure community as compared with living
in the Intermediate community were calculated from
community-specific attack rates and severity scores
for each family segment (Table 4.3.8). Since the
excess days are based on 100 person-weeks of risk, a
good estimate of excess restricted activity in a single
year can be obtained by dividing the excess days
given in Table 4.3.8 by two. (The percentages, of
course, are applicable to any time frame chosen.) The
annual excess is impressive and would involve about 1
extra day of restricted activity every year for each
member of the family. A nuclear family with three
children living, in the Highest exposure community
might experience as much as 12 percent more
restricted activity days than a family living in the
Intermediate exposure community. The greatest per-
centage increase in restricted activity was occasioned
by acute lower respiratory illness in fathers.
There were no intercommunity differences in the
percentage of illnesses requiring a physician's visit.
Chicago-Northwest Indiana Studies
4-45
-------
Table 4.3.5. RELATIVE RISK OF ACUTE RESPIRATORY ILLNESS
FOR OVERALL STUDY PERIOD
Family
segment
Fathers
Mothers
Older
siblings
Nursery
school
children
Younger
siblings
Community
Intermediate
High
Highest
Intermediate
High
Highest
Intermediate
High
Highest
Intermediate
High
Highest
Intermediate
High
Highest
<\
Relative risk of acute respiratory illness3
Upper
tract
1.00 (2.66)
0.96
1.14
1.00 (4.28)
1.21
1.14
1.00 (4.03)
1.36
1.02
1.00 (7.70)
1.09
1.02
1.00 (7.87)
1.12
1.21
Lower
tract
1.00(0.44)
0.89
1.48
1.00 (0.72)
1.08
1.46
1.00 (0.53)
1.64
1.32
1.00(1.67)
0.92
1.38
1.00 (2.87)
0.37
0.94
Total
1.00(3.09)
0.95
1.19
1.00(5.00)
1.19
1.19
1.00 (4.56)
1.39
1.06
1.00 (9.37)
1.06
1.09
1.00(10.31)
0.94
1.16
aBase rate per 100 person-weeks of risk is in parentheses.
However, physician visits for otitis media complicat-
ing an acute respiratory illness were elevated among
children younger than age six who lived in the most
polluted community (Table 4.3.9).
Vulnerability to Repeated Acute
Respiratory Illness
During the study, family members who reported
either four total respiratory illnesses or more than
three upper or one lower illness were judged to have
repeated excessive illness. When the illness experience
of residentially stable families in the study communi-
ties was reviewed, excessive repeated total acute
respiratory illness was consistently more common in
families from the Highest community (Table 4.3.10).
Validity of Analyses
Several analyses were performed to determine
whether intervening variables rather than differences
in air pollution levels could account for the observed
intercommunity difference in morbidity. Intervening
variables considered included socioeconomic status
(as indexed by educational attainment), residential
mobility, and cigarette smoking.
Since residential mobility and educational attain-
ment appeared to be related to each other, the effects
of educational attainment were first evaluated for the
entire study population and later evaluated for
residentially stable families only. To evaluate educa-
tional differences, attack rates for each type of acute
respiratory illness were adjusted for the educational
attainment of fathers using a five-level educational
achievement scale, and appropriate sets of relative
risks were calculated. The adjustment changed the
relative risk pattern very little.
Residential mobility determines pollution expo-
sures and may itself influence the incidence of acute
illness, since changes in life situations seem to
increase morbidity from diverse causes.12 Mobility
and social status are themselves interrelated in that
socioeconomically advantaged upper-middle-class
4-46
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.3.6. RELATIVE SEVERITY SCORES FOR ACUTE RESPIRATORY ILLNESS
Family
segment
Fathers
Mothers
Older
siblings
Nursery
school
children
Younger
siblings
Community
Intermediate
High
Highest
Intermediate
High
Highest
Intermediate
High
Highest
Intermediate
High
Highest
Intermediate
High
Highest
Relative severity score3
Upper
tract
1.00(1.63)
0.92
0.91
1.00(1.59)
0.97
0.91
1.00(2.13)
1.08
1.03
1.00 (2.26)
0.87
0.92
1.00(1.99)
0.89
1.08
Lower
tract
1.00(2.57)
1.47
1.38
1.00(3.23)
0.75
0.74
1.00(3.82)
1.66
1.07
1.00(4.27)
0.77
0.91
1.00 (3.79)
0.77
0.97
Total
1.00(1.76)
1.21
1.12
1.00(1.82)
0.88
0.86
1.00(2.33)
1.31
0.99
1.00(2.62)
0.93
1.04
1.00(2.62)
1.12
0.93
aBase severity score in parentheses.
Table 4.3.7. RESTRICTED ACTIVITY ACCOMPANYING SEVERITY SCORES FOR ACUTE RESPIRATORY
ILLNESS AMONG RESIDENTIALLY STABLE FAMILIES
Family
segment
Parents
Older
siblings
Nursery school
children and
younger siblings
Severity
score
I
II
III
IV
I
II
III
IV
V
I
II
III
IV
V
Mean number of restricted activity
days at each level of severity
Upper
tract
0
2.11
2.32
4.34
0
2.11
1.95
3.53
5.60
0
2.38
2.14
3.77
4.73
Lower
tract
0
3.05
3.33
5.00
0
2.62
2.20
4.24
3.00a
0
3.78
4.00
5.24
4.50
Total
0
2.27
2.54
4.68
0
2.21
2.00
3.67
5.17
0
2.75
2.58
4.23
4.69
aUnusual mean based on one illness.
Chicago-Northwest Indiana Studies
4-47
-------
Table 4.3.8. EXCESS RESTRICTED ACTIVITY NECESSITATED BY MORE FREQUENT OR MORE
SEVERE ACUTE RESPIRATORY ILLNESS
Family
segment
Fathers
Mothers
Older
siblings
Nursery
school
children
Younger
siblings
Total
for nuclear
family of
five
Excess restricted activity days per 100 person-weeks of risk
Upper tract
Percent
excess
-13
-14
2
4
20
9
Excess
days
0.47
0.74
0.13
0.68
3.71
4.79
Lower tract
Percent
excess
97
40
23
40
9
17
Excess
days
1.36
1.06
0.52
3.33
-1.24
5.03
Total
Percent
excess
18
23
6
15
8
12
Excess
days
0.89
1.80
0.65
4.01
2.47
9.82
Table 4.3.9. SEVERITY OF ACUTE RESPIRATORY ILLNESS IN
RESIDENTIALLY STABLE FAMILIES
Family
segment
Fathers
Mothers
Older
siblings
Nursery
school
children
Younger
siblings
Community
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Percent of acute respiratory illnesses
Requiring
restricted
activity
38
49
34
32
58
62
60
56
60
57
Requiring
physician
visit
10
18
7
7
24
30
27
26
25
26
Accompanied by
physician-diagnosed
otitis media
-
—
1.9
1.1
3.5
4.9
1.5
8.3
4-48
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.3.10. REPEATED ACUTE RESPIRATORY ILLNESSES AMONG MEMBERS OF
RESIDENTIALLY STABLE FAMILIES
Family
segment
Fathers
Mothers
Older
siblings
Nursery school
children and
younger
siblings
Community
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Percent of members with repeated acute
respiratory illnesses
More than 3
upper tract
illnesses
6
8
20
35
24
33
49
56
More than 1
lower tract
illness
5
26
4
6
5
13
21
20
More than 4
total
illnesses
6
14
15
21
17
33
49
56
families seem more mobile than less advantaged
working-class families. Residentially stable families
were defined as those who had not changed addres
during the preceeding 3 years. Acute respiratory
disease attack rates for each family segment were
compiled for mobile and residentially stable families,
and sets of relative risks were calculated as previously
described (Table 4.3.11). The most striking result of
the length-of-residence partition was an increase in
the magnitude of the intercommunity differences in
acute respiratory illness found among the residen-
tially stable families. These excesses, most of which
were statistically significant (Table 4.3.12), could be
attributed to long-term exposures to air pollution. On
the other hand, attack rates among the more mobile
population were quite variable, and overall there
seemed to be little difference except for a question-
able increase in total acute respiratory illness among
children exposed to the highest level of ambient air
pollution. The analysis suggested that the excess
morbidity might either be due to exposures lasting 3
years or exposures of indeterminate length involving
ambient levels occurring 1 or 2 years before the study
in the most polluted community.
To determine whether educational attainment and
mobility were related in some complex fashion that
would bias the study, attack rates were compiled for
residentially stable families and partitioned into two
educational attainment groups on the basis of grad-
uate school attendance (Table 4.3.13). Attack rates
were relatively unstable because of small sample sizes,
but a clear pattern was apparent: excesses in illness
attributable to pollution were greater among the
economically advantaged. While this finding might
tend to diminish the differences attributed to pollu-
tion by the data in Table 4.3.12, the differences were
hardly of a magnitude to account for the entire
pollution effect. Educational differences exerted a
statistically significant influence only in the case of
upper tract illness (Table 4.3.12).
Cigarette smoking constitutes an intense self-pollu-
tion hazard and contributes to indoor air pollution
for members of the family who do not smoke. Since
the Intermediate and Highest communities reported
similar smoking habits, it was unlikely that this factor
could bias the study. Nevertheless, smoking-specific
attack rates were calculated for acute respiratory
illness, and relative risk models were computed for
each family segment (Table 4.3.14). Cigar and pipe
smokers were not included in the smoking category.
Among mothers, current cigarette smokers seem to
experience somewhat more illness than lifetime
Chicago-Northwest Indiana Studies
4-49
-------
Table 43.11. RELATIVE RISK OF ACUTE RESPIRATORY ILLNESS
AMONG STABLE AND MOBILE FAMILIES3
Family segment
and
community
Fathers
Intermediate
Highest
Mothers
Intermediate
Highest
Older siblings
Intermediate
Highest
Nursery school
children
Intermediate
Highest
Younger
siblings
Intermediate
Highest
Relative risk of acute respiratory illness
Upper tract
Mobile
1.27
1.20
1.14
1.24
1.14
1.17
1.05
1.21
1.16
1.65
Stable
1.00 (2.59)
1.21
1.00 (3.96)
1.46
1.00 (4.09)
1.37
1.00(7.57)
1.12
1.00 (7.65)
1.27
Lower tract
Mobile
1.10
0.90
1.53
1.22
0.81
0.82
1.25
1.53
0.90
1.00
Stable
1.00 (0.42)
2.29
1.00(0.64)
1.45
1.00 (0.63)
1.44
1.00(1.63)
1.30
1.00 (2.84)
0.93
Total
Mobile
1.25
1.14
1.19
1.28
1.09
1.25
1.09
1.24
1.09
1.43
Stable
1.00 (3.00)
1.36
1.00 (4.60)
1.46
1.00 (4.73)
1.80
1.00 (9.20)
1.15
1.00(10.49)
1.18
aMobile families had moved within the past 3 years; stable families had not.
Base rate per 100 person-weeks in parentheses.
nonsmokers, but this did not prove true for fathers.
Mothers who were exsmokers also appeared to
have a somewhat greater relative risk when com-
pared to lifetime nonsmokers. Assessing the ef-
fect of smoking in the home on children was more
difficult. Attempts to construct dose-response grad-
ients for smoking in the home were partially frus-
trated by small sample sizes. Families could, however,
easily be divided into those with no smoking in the
home and those with smokers. The latter families
were divided into homes where the mother smoked
and an indeterminant category where the mother was
a nonsmoker and either the father was a current light
smoker or someone else in the home smoked. When
the appropriate relative risk models were constructed,
there was a definite increase in the risk of acute lower
respiratory illness in the youngest children that could
be associated with cigarette smoking in the home.
This difference was significant at the p < 0.05 level,
as shown in Table 4.3.12. These children may be in
repeated, intimate contact with indoor air pollution
caused by side-stream cigarette smoke from cigarettes
left burning by their busy mothers. Evaluation of the
smoking-specific rates left some confusion as to the
* impact upon overall acute lower respiratory attack
rates. To answer the question, smoking-adjusted acute
lower respiratory attack rates were calculated (Table
4.3.15). The adjustment further strengthened the
relative risks attributable to pollution in children and
in adults.
4-50
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.3.12. PROBABILITY SUMMARY FOR EFFECTS OF SELECTED DETERMINANTS
ON ACUTE RESPIRATORY DISEASE FREQUENCY AMONG
RESIDENTIALLY STABLE FAMILIES
Determinant
Pollution
exposure
Socioeconomic
status
Personal
cigarette
smoking
Cigarette smoking
in home
Study
period
Overall
Influenza
Noninfluenzal
respiratory
Summer
Overall
Overall
Overall
Family
segment
All members
Parents
Children
All members
All members
All members
All members
Parents
Children
Significance3 i
Upper
tract
<0.05
< 0.025
NS
<0.10
< 0.025
NS
< 0.001
NS
<0.10
Lower
tract
<0.05
NS
<0.05
NS
NS
<0.05
NS
<0.10
< 0.001
Total
< 0.005
<0.01
<0.05
< 0.025
< 0.005
<0.10
NS
NS
<0.05
NS — not significant, p > 0.10. All significant associations involve an increase in illness frequency.
Age-specific Attack Rates
Age-specific acute respiratory attack rates among
children' from residentially stable families (Figure
4.3.2) can be used to estimate the benefits of recent
pollution control or the length of exposure necessary
to increase the vulnerability of a child to acute lower
respiratory illness.
Children from the Highest exposure community
were found to have higher attack rates during the first
3 years of life. This finding suggests that infants living
in the Intermediate exposure community benefit
substantially from control of sulfur dioxide. There
was a decline in attack rates after 3 years of age, and
intercommunity differences were quite small or non-
existent. A possible explanation for this observation
may be that the child who first enters nursery school
or elementary school is already maximally challenged
by a succession of infectious agents and that little or
no further increase in illness can be attributed to the
more subtle environmental insults.
DISCUSSION
This study corroborates the findings of Douglas
and Waller4 and Lunn et al.s Acute respiratory
morbidity was significantly lower among families
living in neighborhoods where sulfur dioxide levels
had been substantially decreased. Furthermore, the
most striking improvements involved reductions in
total acute respiratory illness among the youngest
children and acute lower respiratory illness among
adults.
The gradient of pollution exposure developed for
this study was based on total suspended particulate
concentrations. As a result, some families were
included in the Intermediate community although
Chicago-Northwest Indiana Studies
4-51
-------
Table 4.3.13. EFFECT OF EDUCATIONAL ATTAINMENT UPON RELATIVE
RISK OF ACUTE RESPIRATORY ILLNESS IN
RESIDENTIALLY STABLE FAMILIES
Family segment
and
community
Fathers
Intermediate
Highest
Mothers
Intermediate
Highest
Older siblings
Intermediate
Highest
Nursery school
children
Intermediate
Highest
Younger siblings
Intermediate
Highest
Relative risk of acute respiratory illness3
Upper tract
High
socioeconomic
status
1.00 (2.35)
1.49
1.00(4.30)
1.37
1.00 (5.02)
1.27
1.00 (8.76)
1.03
1.00(6.18)
1.89
Low
socioeconomic
status
1.03
0.85
0.92
0.93
0.80
0.64
0.80
0.65
1.15
0.97
Lower tract
High
socioeconomic
status
1.00 (0.43)
2.09
1.00 (0.43)
2.02
1.00 (0.56)
3.57
1.00(1.73)
1.26
1.00 (3.09)
0.65
Low
socioeconomic
status
0.91
1.40
2.16
1.65
1.30
0.55
0.80
1.39
0.79
1.17
Total
High
socioeconomic
status
1.00 (2.78)
1.58
1.00 (4.73)
1.43
1.00 (5.58)
1.58
1.00(10.50)
1.06
1.00 (9.27)
1.48
Low
socioeconomic
status
1.01
0.94
1.04
1.00
0.85
0.63
0.80
0.77
1.03
1.04
aBase rate for 100 person-weeks of risk in parentheses.
they were exposed to sulfur dioxide levels higher than
those of some families in the High and Highest
communities. However, a separate analysis of those
families with higher sulfur dioxide exposures showed
that each population segment, with the exception of
the youngest children, had respiratory disease rates
that were higher than the mean rate for the Inter-
mediate community. Thus, inclusion of these families
in the Intermediate community would tend to ob-
scure differences in disease rates between the com-
munities that could be attributed to pollutant levels.
No lessening of illness severity could be docu-
mented, except for a decrease in otitis media among
the youngest children living in the Intermediate
community. Lessened illness frequency in the com-
munity with improved air quality combined with
unchanged severity resulted in substantial reductions
in the restricted activity burden associated with acute
respiratory illness. Another important finding was
that reductions in ambient air pollution might reduce
the vulnerability of families to an epidemic outbreak
of influenza.
Residential mobility, socioeconomic status,
and cigarette smoking in the home were shown
to be determinants of acute lower respiratory disease
in children, but none of these factors accounted for
the excesses attributed to ambient air pollution.
Another potential indoor air pollution problem,
nitrogen dioxide emissions from cooking gas, could
hardly bias the study since there were no significant
intercommunity differences in the use of domestic
gas for cooking or space heating.
It would be a most serious error to assert that the
present study demonstrated the full benefits of air
4-52
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 4.3.14. EFFECT OF CIGARETTE SMOKING UPON RELATIVE RISK
OF ACUTE RESPIRATORY ILLNESS IN RESIDENTIALLY STABLE FAMILIES
Family
segment
Fathers
Mothers
Older
siblings
Nursery
school
children
Younger
siblings
Community
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Current
personal
or family
cigarette
smoking
status
Lifetime non-
smoker
Lifetime non-
smoker
Exsmoker
Exsmoker
Current cigarette
smoker
Current cigarette
smoker
Lifetime non-
smoker
Lifetime non-
smoker
Exsmoker
Exsmoker
Current cigarette
smoker
Current cigarette
smoker
No current
smoking
No current
smoking
Indeterminant
Indeterminant
Mother smokes
Mother smokes
No current
smoking
No current
smoking
Indeterminant
Indeterminant
Mother smokes
Mother smokes
No current
smoking
No current
smoking
Indeterminant
Indeterminant
Mother smokes
Mother smokes
Relative risk of acute respiratory
illness3
Upper
tract
1.00(2.71)
1.28
0.89
0.89
0.73
0.96
1.00 (3.78)
1.46
2.51
2.64
1.09
1.06
1.00 (4.84)
0.83
0.57
2.13
0.99
0.41
1.00 (8.56)
0.80
0.86
1.14
0.84
0.83
1.00 (7.82)
0.93
0.55
2.05
0.90
1.41
Lower
tract
1.00(0.27)
4.44
1.65
1.59
2.14
1.48
1.00(0.52)
1.38
1.92
21.15b
1.98
2.10
1.00 (0.32)
5.18
1.78
4.47
3.06
0.00
1.00(1.72)
1.78
1.03
1.16
0.67
0.77
1.00(2.18)
1.15
0.77
0.46
1.68
2.29
Total
1.00(2.98)
1.57
0.97
0.94
0.87
1.00
1.00(4.30)
1.45
2.44
2.32
1.20
1.18
1.00(5.16)
1.14
0.65
2.27
1.12
0.39
1.00(10.28)
0.96
0.89
1.14
0.81
0.81
1.00(10.00)
0.98
0.60
1.70
1.07
1.60
aBase rate per 100 person-weeks of risk is in parentheses.
Unstable risk based on a very small sample.
Chicago-Northwest Indiana Studies
4-53
-------
4.3.15. SMOKING-ADJUSTED RELATIVE RISK OF ACUTE RESPIRATORY
DISEASE IN RESIDENTIARY STABLE FAMILIES
Family
segment
Fathers
Mothers
Older
siblings
Nursery
school
students
Younger
siblings
Community
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Intermediate
Highest
Relative risk of acute respiratory illness3
Upper
tract
1.00 (2.39)
1.25
1.00 (4.06)
1.27
1.00 (4.40)
1.01
1.00 (7.82)
0.99
1.00 (6.87)
1.48
Lower
tract
1.00(0.41)
1.90
1.00 (0.70)
1.64
1.00 (2.64)
1.46
1.00(1.53)
1.16
1.00(2.54)
1.09
Total
1.00(2.80)
1.33
1.00(4.76)
1.25
1.00(7.04)
1.18
1.00(0.35)
1.02
1.00 (9.41)
1.37
aBase rate per 100 person-weeks at risk is in parentheses.
pollution control on acute respiratory morbidity. The
limited nature of air monitoring data precluded
quantifying possible short-term peak exposure ef-
fects. The age-specific attack rate pattern presented is
consistent with the emergence of increased suscepti-
bility to acute respiratory illness after less than 3
years of exposures to elevated pollutant levels.
Perhaps less than 1 year of exposure is required. Such
exposures cannot be fully evaluated without a com-
prehensive monitoring of short-term pollutant levels.
Another problem relates to the confidence with which
a pollutant or pollutants might be incriminated as
causal agents for the observed impairments. The
evidence indicates the Intermediate community most
likely benefitted from reductions in sulfur dioxide
and perhaps suspended sulfates. Suspended particu-
late reduction, although more modest, could well
have played a role in the observed improvements.
After weighing the strengths and limitations of the
present study, we concluded that there was strong
evidence that reducing sulfur dioxide levels from an
annual average somewhat above the relevant National
Primary Air Quality Standard to a level below the
standard (107 to 109 jug/m3 compared with 51 to 57
/ug/m3) would reduce acute respiratory disease mor-
bidity. Reductions in total suspended particulate
and/or suspended sulfate levels could well result in
additional reductions in acute respiratory illness, but
the reductions in illness associated with reductions in
sulfur dioxide are felt to be distinct frorn, those that
might result from reductions in other pollutants.
SUMMARY
A prospective survey of over 2500 members of
volunteer families indicated that those 'individuals
living more than 3 years in high pollution areas
(sulfur dioxide, 107 to 250 pg/m3; total suspended
particulates, 137 to 165 A/g/m3) had significantly
increased rates of acute respiratory illness, restricted
activity, and otitis media when compared to indi-
viduals living in less polluted areas (sulfur dioxide,
109 to 130 jug/m3; total suspended particulates, 121
to 123 fig/m3). It is also possible that more recent
lower air pollution levels contributed to increased
respiratory illness. During the study, sulfur dioxide
levels in the lower pollution areas were around 57
Mg/m3 as compared with 51 to 106 ng/m3 in the
higher pollution areas; for total suspended particu-
4-54
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Figure 4.3.2. Age-specific attack rates for
acute respiratory illness among children
from residentially stable families.
lates, the levels are around 111 /ug/m3 as compared
with 126 to 151 jug/m3.
Children and adults living in neighborhoods with
cleaner air were measurably less susceptible to etiolo-
gjcally undefined illness and to epidemic influenza as
well. Families with higher socioeconomic status gen-
erally reported more illness. Residential mobility
proved a significant determinant of illness, and
cigarette smoking by mothers seemed related to
increased susceptibility among the youngest pre-
school children. None of these intervening variables
could account for the effects attributed to ambient
air pollutants. It was concluded that improvements in
air quality carr significantly reduce the incidence of
respiratory illnesses.
REFERENCES FOR SECTION 4.3
1. Acute Conditions, Incidence and Associated Dis-
ability. In: Health Statistics, Series B. National
Center for Health Statistics, Public Health Serv-
ice, U. S. Department of Health, Education, and
Welfare. Washington, D.C. PHS Publication No.
584. 1969.
2. Clark, D. and B. MacMahon. Preventive Medicine
(Chapter 25). Boston, Little Brown and Com-
pany, 1967.
3. Reid, D.D. The Beginnings of Bronchitis. Proc.
Roy. Soc. Med. 62:311-316,1969.
4. Douglas, J. W. B. and R. E. Waller. Air Pollution
and Respiratory Infection in Children. Brit. J.
_ Prevent. Soc. Med. 20:1-8, 1966.
5. Lunn, J. E., J. Knowelden, and A. J. Handyside.
Patterns of Respiratory Illness in Sheffield Infant
School Children. Brit. J. Prevent. Soc. Med.
27:7-16,1967.
6. Toyama, T. Air Pollution and Its Health Effects
in Japan. Arch. Environ. Health. 5:153-173,
1964.
7. Dohan, F. C. Air Pollutants and Incidence of
Respiratory Disease. Arch. Environ. Health.
3:887-895,1961.
8. Dingle, J. H., G. F. Badger, and W. S. Jordan.
Patterns of Illness. In: Illness in the Home.
Cleveland, The Press of Western Reserve Univer-
sity, 1964. p. 33-37.
9. Questionnaires Used in the CHESS Studies. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA-6 50/1-74-004. 1974.
10. Hinton, D. O., T. D. English, B. F. Parr, V.
Hasselblad, R. C. Dickerson, and J. G. French.
Human Exposure to Air Pollutants in the Chica-
go-Northwest Indiana Metropolitan Region,
1950-1971. In: Health Consequences of Sulfur
Oxides: A Report from CHESS, 1970-1971. U.
S. Environmental Protection Agency. Research
Triangle Park, N. C. Publication No. EPA-
650/1-74-004. 1974.
11. Grizzle, J. E., C. F. Starmer, and G. G. Koch.
Analysis of Categorical Data by linear Models.
Biometrics. 25(3):489-504, September 1969.
12. Rahe, R. H., J. D. McKean, and R. J. Arthur. A
Longitudinal Study of life-change and Illness
Patterns. J. Psychosom. Res. 70:355-366, 1967.
Chicago-Northwest Indiana Studies
4-55
-------
CHAPTER 5
NEW YORK STUDIES
5-1
-------
5.1 HUMAN EXPOSURE TO AIR POLLUTION
IN SELECTED NEW YORK
METROPOLITAN COMMUNITIES, 1944-1971
Thomas D. English, Ph.D., Walter B. Steen, B.S.,
Robert G. Ireson, M.S., Peggy B. Ramsey, B.S.,
Robert M. Burton, B.S., and L. Thomas Heiderscheit, M.S.
5-3
-------
INTRODUCTION
The New York metropolitan area, with over ten
million people exposed to a complex mixture of air
pollutants, provides a unique setting for studying the
influence of pollutants on human health. Although
the patterns of pollution in a metropolitan area obvi-
ously are much more complex than those of less
highly urbanized situations, the size of potential
study groups and the size of the population possibly
experiencing adverse effects make it imperative that
the studies be undertaken. In the metropolitan area,
measurement of major pollutants—sulfur dioxide
(SO^), total suspended particulates (TSP), suspended
nitrates (SN), suspended sulfates (SS), and dustfall—
provides useful indices of total pollution that can be
used as a basis for comparison with other areas. The
major metropolitan areas are also the locations in
which improvements in air quality are most likely to
be effected at an early date. It follows then that these
areas will be the first in which data related to the
Benefits of control can be obtained. For these
reasons, the Community Health and Environmental
Surveillance (CHESS) program of the Environmental
Protection Agency (EPA) was initiated in the New
York area in April 1970.
Community Description
Within the New York City metropolitan area,
three communities were selected on the basis of
several years of prior data to comprise a gradient
of exposure to sulfur oxides and total suspended
particulates. These are the Westchester sector of
the Bronx, a community expected to have high
exposures to both sulfur oxides and total
suspended particulate; the Howard Beach Section of
the Queens, an intermediate exposure com-
munity; and Riverhead, Long Island, a low
exposure community.
The New York CHESS area is characterized
by a flat terrain with elevations of 100 feet or
less above sea level. Bronx, Queens, and River-
head, are all located near rivers or bays (Figure
5.1.1). The climate of the area can best be de-
fined as modified continental. Weather situations
that make up the climate pattern move from a
generally westerly direction. The maritime in-
fluence is shown by the uniform occurrence of
precipitation throughout the year in contrast to
the summertime concentration of rainfall that is
typical of an inland regime. The maritime in-
fluence also moderates temperatures in a manner
that produces differences between "in-city" and
suburban locations that are often of considerable
magnitude.
Other factors influencing the climate of the
area are coastal storms that produce strong winds
and heavy precipitation and low-pressure systems,
also associated with unsettled weather, that fre-
quently pass through the area. The prevailing
westerly winds experienced are illustrated by
wind roses from John F. Kennedy and LaGuardia
airports (Figure 5.1.2 and 5.1.3).
Annual mean temperature in New York City
is somewhat higher than that of most places at
the same latitude in the United States, except
Pacific coastal localities. In summer, New York
City may be hot, but extended heat waves are
rare. Autumn is milder than spring.
Pollution Sources
In the metropolitan New York area, numerous
and varied pollution sources produce a complex
mixture of pollutants, the composition of which
has not been defined completely. The measure-
ment of major pollutants is acknowledged to be
one means of indexing this total pollution.
Sulfur dioxide and suspended particulates are
produced primarily by the combustion of fossil
fules and thus by the industrial and domestic
units that burn these fuels. Electric power plants
have been shown to be the major producers, but
Figure 5.1.1. New York CHESS communities.
5-4
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
7.0
8.0
6.0
5.0
8.0
8.0
9.0
12
OCCURRENCE, %
4-7 8-12
WIND SPEED, mph
13-17 19-23
Figure 5.1.2. Surface climatologic wind
rose for John F. Kennedy Airport, New
York. (Percentage of time wind is in
direction shown.)
'6.U
9.0
12 16
OCCURRENCE,
1
J-7
7
17
WIND SPEED, mph
Figure 5.1.3. Surface climatologic wind
rose for La Guardia Airport, New York.
(Percentage of time wind is in direction
shown.)
a variety of other industrial facilities contribute
to the total sulfur dioxide emissions. Many of
these sources, as well as incinerators and auto-
mobiles, contribute significantly to total partic-
ulate emissions.
Monitoring Human Exposure to Air
Pollutants
The CHESS strategy employed for monitoring
human exposure to environmental pollutants has
been described in other CHESS reports.1'2 The
New York populations under surveillance were
located within 1.5 miles of the CHESS air
monitoring sites.
Location and Description of Monitoring
Sites
All New York monitoring sites were located
on tops of buildings 30 to 45 feet above ground
level. CHESS stations in New York City were
located on the same sites as those of the New
York City Department of Air Resources
(NYC-DAR). Since the NYC-DAR was also
measuring current air quality and had several
years of past data available, the analyses in this
monograph utilized information from both
sources to the fullest extent possible.
Two aspects of the location of the monitoring
sites must be considered in relating the data
New York Studies
5-5
-------
obtained to the exposure of the populations
being studied. First, the air monitoring station in
the Bronx community was situated on top of a
three-story court house in the center of a busy
commercial area, while those in the Queens and
Riverhead communities were located on two-story
buildings surrounded by residential dwellings.
Thus, at the Bronx station, proximity to heavy
traffic would tend to increase measured pollutant
levels, while greater height above ground would
tend to decrease these values. The stations in the
Queens and Riverhead communities probably
more accurately reflected the ambient exposure
experienced by nearby residents. Secondly, the
Queens community lies about 1 mile west of the
John F. Kennedy International Airport; odors
and noise from the airport have been the object
of many complaints by the residents of that
community. Because of budget limitations, hydro-
carbons, one of the important classes of pollut-
ants emitted by aircraft engines, were not
monitored at any of the CHESS stations.
Pollutant Measurement Methods
Similar techniques were used by CHESS and
NYC-DAR for the measurement of total sus-
pended particulates3 and particulate dustfall.4
Total suspended particulate values recorded by
NYC-DAR were consistently about 20 percent
higher than values recorded by CHESS. To
measure concentrations of sulfur dioxide, CHESS
used a modification of the West-Gaeke method5
and the NYC-DAR used a hydrogen peroxide
technique.6 NYC-DAR estimated that 15 percent
higher concentrations of sulfur dioxide should be
expected with the hydrogen peroxide technique.7
A "detailed comparison of NYC-DAR and CHESS
sulfur dioxide measurements is presented later in
this paper. The same techniques for determining
coefficient of haze (COH) were used by CHESS
and NYC-DAR.8 NYC-DAR also made determina-
tions of total suspended particulate by utilizing a
dust count instrument that operated by the
principal of light scattering.9
Nitrogen dioxide concentrations were measured
by CHESS and NYC-DAR using a modification
of the Jacobs-Hochheiser method, which at the
time of data collection was the procedure
accepted by the Federal government. The tech-
nique has recently come under much criticism. It
is widely held that the Jocobs-Hochheiser method
does not give accurate results because of several
factors, among which are a fluctuating absorption
efficiency and the probability of interference
from other environmental pollutants. Nitrogen
dioxide data reported in this paper, therefore,
should only be used in light of the recent
findings concerning the accuracy of the measure-
ment method that was employed.
To estimate the fraction of the total sus-
pended particulate that is respirable, a separate
set of measurements was made by CHESS using
an Atomic Energy Commission cyclone as a pre-
filter. This device collects all particles larger than
10 nm, filters out 50 percent of the 3.5-^m
particles, and has no effect on 2-fjm or smaller
particles. The respirable particulates penetrating
the cyclone are collected on the backup filter.
Precision of Measurements
The precision of daily CHESS sulfur dioxide
collection and analytical techniques is estimated
to be ±27 percent for the arithmetic mean based
on duplicate measurements made over a period
of 19 months in Birmingham, Alabama.2 Com-
parison of monthly CHESS and NYC-DAR sulfur
dioxide data over a period of 13 months resulted
in a correlation coefficient of 0.57 and a mean
relative error of ±25 percent based on monthly
arithmetic means (Figure 5.1.4). Part of the rela-
tive error is due to differences in measurement
method; as mentioned earlier, the hydrogen
peroxide method used by NYC-DAR has been
estimated to give results about 15 percent higher
than actual values.
Precision of CHESS total suspended particulate
measurements is estimated to be ±6 percent for
the arithmetic mean based on duplicate measure-
ments made over a period of 28 months in
Birmingham, Alabama, Charlotte, North Carolina,
and Greensboro, North Carolina.2 A comparison
of CHESS and NYC-DAR total suspended partic-
ulate data over a period of 1 year gave a corre-
lation coefficient of 0.80 and a mean error ±10
percent (Figure 5.1.5). Recent evidence indicates
that CHESS measurements may be approximately
20 percent low due to loss of material from the
filter during shipment.
Precision of CHESS total dustfall measure-
ments is estimated to be ±25 percent for the
arithmetic mean based on duplicate measurements
made over a period of approximately 2-1/2 years
in Birmingham, Charlotte, and Greensboro.2 It is
believed that the precision of the total dustfall
5-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
O
O
175
150
125
100
75
50
t/o
25
Y = 27.2+0.34X
R=57
N=40
MEAN ERROR =±25.3
25 50 75 100 125 150 175
NYC-DAR S02 CONCENTRATION,MS/m3
200
225
250
Figure_5/K4. Comparison of monthly average CHESS and New York City Department of Air
Resources sulfur dioxide data for Queens and Bronx.
CO
00
•a:
a:
o
o
D_
CO
130
120
110
100
90
80
70
60
50
40
YM.5+0.78X
R = 0.80
N=40
ME AN ERROR = ±10.1%
40 50 60 70 80 90 100 110
NYC-DAR TSP CONCENTRATION, jig/m
120
130
140 150
Figure 5.1.5. Comparison of monthly average CHESS and New York City Department of Air
Resources total suspended particulate data for Queens and Bronx.
New York Studies
5-7
-------
measurement obtained by NYC-DAR is compara-
ble.
Sampling Frequency
Both CHESS and NYC-DAR data for sulfur
dioxide, nitrogen dioxide, and total suspended
particulates are based on 24-hour integrated
samples. The CHESS respirable suspended partic-
ulates measurements are also based on 24-hour
samples. Both CHESS and NYC-DAR data for
COH are based on 2-hour samples. Dustfall
sampling periods are 1 month. A summary of
the methods, sampling frequencies, and sampling
times for both CHESS and NYC-DAR is given in
Table 5.1.1.
given in the Appendix, Tables 5.I.A.I to 5.I.A.6.
Frequency distributions of daily sulfur dioxide
concentrations in the three New York commu-
nities are illustrated in Figure 5.1.6. In general,
the data show a log-normal frequency distribu-
tion; however, data for Queens deviate from the
straight line for cumulative frequencies greater
than 85 percent. Bronx concentrations exceed
those in Queens about 15 percent of the time.
As a point of reference, the National Annual
Primary Ambient Air Quality Standard of 80
fj.g/m^ for sulfur dioxide was exceeded on 5
percent of the days in Riverhead, 20 percent of
the days in the Bronx, and 25 percent of the
days in Queens. Daily mean values never ex-
ceeded 365 ng/m^ (the National Daily Primary
Ambient Air Quality Standard).
Quality Control
Because data outputs of an air sampling pro-
gram are subject to many sources of error, an
effective quality control program is an essential
element of an environmental surveillance system.
Effective real-time controls are essential to
minimized field errors, systematic drift, chemical
laboratory errors, data transfer errors, computer
punch card errors, analysis errors, etc. A sys-
tematic approach has been used to provide a
quality control system for CHESS.^
RESULTS AND DISCUSSION
Frequency distributions of daily total suspended
particulate concentrations for the New York commu-
nities are illustrated in Figure 5.1.7. The data are
log-normally distributed, the slopes of the lines for
the three areas are approximately the same, and the
relative positions of the lines indicate a significant
concentration gradient among the three communities.
The 24-hour mean total suspended particulate con-
centrations did not exceed the National Daily Pri-
mary Air Quality Standard (260 jitg/m-*) in any of the
communities. However, the geometric annual mean
for total suspended particulate in the Bronx is slightly
above the National Annual Primary Air Quality
Standard.
Daily Frequency Distribution
Daily frequency distributions for the pollutants
measured in the New York communities are
Daily frequency distributions for suspended
sulfates in the New York communities (Figure
5.1.8) show that Riverhead generally had lower
concentrations and that levels in Bronx and
Table 5.1.1. METHODS, SAMPLING FREQUENCIES, AND SAMPLING TIMES USED TO MEASURE
POLLUTANTS IN NEW YORK AREA
Pollutant
Sulfur dioxide
Nitrogen dioxide
Dustfall
Total particulate
Smoke shade
Dust count
Respirable suspended
particulate
Method
NYC-DAR
Hydrogen peroxide
Jacobs-Hochheiser
Jar
High-volume samplers
AISI
Light scattering
CHESS
West-Gaeke
Sampling frequency,
samples/yr
NYC-DAR
365
Jacobs-Hochheiser 365
Bucket
High-volume samplers
AISI
—
AEC cyclone
12
365
4380
Grab
samples
--
CHESS
365
365
12
365
4380
—
365
Sampling time
NYC-DAR
24 hours
24 hours
Monthly
24 hours
2 hours
—
CHESS
24 hours
24 hours
Monthly
24 hours
2 hours
--
24 hours
5-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
400
300|
200
100
60
40
cc.
o
o 20
CXI
10
DAILY AIR QUALITY STANDARD
ANNUAL AIR QUALITY STANDARD
10
30
50
70
90
98
PERCENT LESS THAN OR EQUAL TO INDICATED CONCENTRATION
Figure 5.1.6. Cumulative frequency of sulfur dioxide concentrations, New York CHESS, 1971
New York Studies
5-9
-------
300
200
o
t
Cd
o
O
100
60
fe 40
30
20
10
_J. 1 [ 1 U-1 4
•DAILY AIR QUALITY STANDARD
ANNUAL AIR QUALITY STANDARD
5 10 30 50 70 90 95 99
PERCENT LESS THAN OR EQUAL TO INDICATED CONCENTRATION
Figure 5.1.7. Cumulative frequency of total suspended particulate concentrations, New York
CHESS, 19717
Queens were similar. Frequency distributions for
the nitrate fraction of total suspended partic-
ulates, for respirable particulates, and for nitrogen
dioxide are not plotted but are given in Tables
5.1.A.4,5.1.A.5,and5.1.A.6.
Seasonal Trends
Quarterly arithmetic averages for sulfur
dioxide, nitrogen dioxide, total suspended partic-
ulates, suspended sulfates, and suspended nitrates
are presented for the three New York CHESS
communities in Table 5.I.A.7. These data, from
the fourth quarter of 1970 through the first
quarter of 1972, represent seasonal averages.
Monthly arithmetic mean sulfur dioxide con-
centrations for the three New York communities
over a 19-month period (Figure 5.1.9) demon-
strate an annual cyclic pattern, with peak con-
centrations occurring in January and minimum
values during the summer months. During most of the
year, sulfur dioxide levels were highest in Bronx and
lowest in Riverhead. Although it is not applicable to
monthly means, the National Annual Primary Air
Quality Standard is shown in Figure 5.1.9 as a basis
5-10
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
for comparison. Six of the 19 monthly mean levels of
sulfur dioxide for the Bronx exceeded the Annual
Standard, as did 2 of the 19 values for Queens.
Monthly arithmetic mean values for total sus-
pended particulates over a 19-month period do
not exhibit an annual cyclic pattern (Figure
5.1.10). The data show a significant gradient,
with the highest concentrations in the Bronx and
the lowest in Riverhead. Fifteen of 19 mean
monthly values from the Bronx exceed the
National Annual Primary Ambient Air Quality
Standard, as did 2 of the 19 mean values from
Queens.
Monthly arithmetic mean values for suspended
sulfates over a period of 18 months for New
York indicated no seasonal trends (Figure
5.1.11). In general, sulfate levels conform to the
gradient observed for total suspended particulate
and sulfur dioxide; however, the gradient between
the Bronx and Queens is not as well defined.
Monthly arithmetic mean values for suspended
nitrates in the New York communities are shown
in Figure 5.1.12. The gradient is similar to that
for suspended sulfates.
Monthly arithmetic mean nitrogen dioxide
concentrations exhibited a reasonably consistent
40
30
20
10
o
H
cc
UJ
o
o
o
10 30 50 70 90 98 99
PERCENT LESS THAN OR EQUAL TO INDICATED CONCENTRATION
99.9
99.99
Figure 5.1.8. Cumulative frequency of suspended sulfate concentrations, New York
CRESS, 1971.
New York Studies
5-11
-------
ANNUAL AIR QUALITY STANDARD
OCT
1970
DEC
APR
OCT
DEC
FEB
JUN AUG
TIME, months
Figure 5.1.9. Sulfur dioxide concentration versus time in months, New
York CHESS.
APR
1972
ANNUAL AIR QUALITY STANDARD
BRONX
APR JUN AUG
TIME, months
Figure 5.1.10. Total suspended particulate concentration versus time in
months, New York CHESS.
5-12
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
o
6
O
o
OCT DEC FEE APR JUN AUG OCT DEC FEB APR
TIME, months
Figure 5.1.11. Suspended sulfate concentration versus time in months,
New York CHESS.
TIME, months
Figure 5.1.12. Suspended nitrate concentration versus time in months,
New York CHESS.
New York Studies
5-13
-------
gradient for the New York communities (Figure
5.1.13). Since October 1970, none of the
monthly mean values have exceeded the National
Annual Primary Ambient Air Quality Standard of
100jug/m3.
Monthly arithmetic mean values for respirable
suspended paniculate (i.e., that portion of total
suspended particulates with maximum diameters of
10 /zm) for Bronx, Queens, and Riverhead over a
9-month period (Figure 5.1.14) demonstrated a gra-
dient similar to that of total suspended particulate.
These communities are characterized by typical
respirable particulate concentrations of about 40
/-ig/m . The data are insufficient to provide indica-
tions of annual cyclic patterns.
Monthly arithmetic mean COH values for the
New York communities over a 6-month period
(Figure 5.1.15), like the pollutants previously
discussed, demonstrate exposure levels that de-
crease in order from Bronx, to Queens, to River-
head. The respective average COH values for
these communities were 1.25, 0.65, and 0.27.
During most of the year, monthly total dust-
fall (Figure 5.1.16) also demonstrates a gradient
of exposure that decreases in order from Bronx,
to Queens, to Riverhead. Cadmium in monthly
dustfall samples is much more inconsistent in
ANNUAL AIR QUALITY STANDARD
APR JUN AUG
TIME, months
Figure 5.1.13. Nitrogen dioxide concentration versus time in months, New York CHESS.
5-14 HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
AUG
1971
SEP OCT NOV DEC JAN
TIME, months
FEE
MAR
APR
1972
Figure 5.1.14. Respirable suspended particulate concentration versus time in months, New York
CHESS.
SEP
1971
NOV DEC JAN
TIME, months
Figure 5.1.15. Coefficient of haze versus time in months, New York CHESS.
New York Studies
FEB
1972
5-15
-------
Bronx than in Queens or Riverhead (Figure
5.1.17). Bot1 Queens and Bronx are characterized
by monthly average cadmium values of about 0.1
mg/m2/mo, while the levels in Riverhead are
about 0.05 mg/m^/mo. Zinc in monthly dustfall
samples shows a gradient during most of the
year, decreasing from Bronx, to Queens, to
Riverhead (Figure 5.1.18). The respective average
monthly zinc values in dustfall are 16, 8, and 4
mg/m^/mo. A similar gradient is demonstrated by
lead in monthly dustfall (Figure 5.1.19); mean
monthly values over 17 months from Bronx,
Queens, and Riverhead were 16, 10, and 2
mg/m^/mo, respectively.
SEP
1970
NOV
MAR MAY JUL
TIME, months
SEP
NOV
JAN
1972
Figure 5.1.16, Dustfall concentration versus time in months,
New York CHESS.
E
O
TIME, months
Figure 5.1 .17. Cadmium concentration in dustfall versus time in months,
New York CHESS.
5-16
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
o
§
o
1970
TIME, months
Figure 5.1.18. Zinc concentration in dustfall versus time in months,
New York CHESS.
O
O
o
NOV
JAN
MAR
SEP
1970
TIME, months
Figure 5.1.19. Lead concentration in dustfall versus time in months,
New York CHESS.
New York Studies
5-17
-------
Interrelationships among Pollutants
Total Suspended Particulates, Suspended Nitrates,
and Nitrogen Dioxide
Comparisons of 24-hour integrated daily
measurements for 1971 (Table 5.1.2) gave evi-
dence that the weights of the nitrate fractions of
total suspended particulate samples were consist-
ently related to the total weights of the samples
in the community with highest pollution levels
(Bronx), but poorly correlated in the lowest
pollution community (Riverhead). Nitrogen
dioxide also correlated strongly with total sus-
pended particulates in the high pollution com-
munity, but more weakly in the communities
with less pollution. Nitrate levels showed weak
correlation with nitrogen dioxide levels.
Table 5.1.2. CORRELATION COEFFICIENTS FOR
TOTAL SUSPENDED PARTICULATES, SUSPEND-
ED NITRATES, AND NITROGEN DIOXIDE, NEW
YORK CHESS, 1971
Community
Bronx
Queens
Riverhead
Pollutant
TSP
SN
N02
TSP
SN
NO2
TSP
SN
NO2
Correlation coefficient
TSP
1.000
0.594
0.545
1.000
0.430
0.374
1.000
0.198
0.368
SN
0.594
1.000
0.361
0.430
1.000
0.190
0.198
1.000
0.323
N02
0.545
0.361
1.000
0.374
0.190
1.000
0.368
0.323
1.000
Total and Respirable Suspended Particulates
Comparisons of nine monthly arithmetic means
for total and respirable suspended particulates
indicated a rather consistent relationship. Ratios
of respirable to total particulates decreased from the
areas of low pollution to the area of high pollution
(Figure 5.1.20).
Total Suspended Particulates, Suspended Sulfates,
and Sulfur Dioxide
Comparisons of 24-hour integrated daily
measurements for 1971 (Table 5.1.3) indicated
BRONX
— QUEENS
RIVERHEAD
JULY OCT
TIME, monlb
APR
197!
Figure 5.1.20. Total suspended particulate
(high volume sampler) and respirable sus-
pended particulate (cyclone) concentrations
versus time in months, New York CHESS.
Table 5.1.3. CORRELATION COEFFICIENTS FOR
TOTAL SUSPENDED PARTICULATES, SUSPEND-
ED SULFATES, AND SULFUR DIOXIDE, NEW
YORK CHESS, 1971
Community
Bronx
3ueens
-iiverhead
Pollutant
TSP
SS
S02
TSP
SS
SO2
TSP
SS
S02
Correlation coefficient
TSP
1.000
0.615
0.150
1.000
0.637
0.399
1.000
0.772
0.315
SS
0.615
1.000
0.130
0.637
1.000
0.282
0.772
1.000
0.282
SO2
0.150
0.130
1.000
0.399
0.282
1.000
0.315
0.282
1.000
5-18
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
that the weights of the sulfate fractions of total
suspended particulate samples were consistently
related to the total weights of the samples. This
correlation was strongest (0.772) in the
community with lowest total suspended
particulate levels (Riverhead) and weakest (0.615)
in the community with highest levels (Bronx).
Similar comparisons of atmospheric sulfur dioxide
and concurrent suspended sulfates showed less
significant correlations.
Long-term Exposure Trends
Estimates of previous exposures to airborne
pollutants, extending back to 1944 (Table
5.1.A.8), were derived from information provided
by NYC-DAR, supplemented by recent CHESS
measurements and limited data from the National
Air Sampling Network (NASN). In general, the
city wide estimates were based on 12 years of
recorded data (1958-1970) for atmospheric sulfur
dioxide concentrations and 14 years (1956-1970)
of data for monthly dustfall. Because fuel con-
sumption and emission control technology re-
mained reasonably constant for several years prior
to 1958, it was assumed that the ratio between
dustfall and sulfur dioxide concentration also
remained constant during those years. Therefore,
sulfur dioxide concentrations were estimated on
the basis of available dustfall measurements for
1956-1957. Because the fluctuations in concen-
trations of total suspended particulate and sulfur
dioxide follow the same general pattern, total
suspended particulate concentrations back to
1956 also were estimated. Measured values for
suspended sulfates for 1956-1970 were available
from the Manhattan 121st Street station, and
these values were used as citywide values. Pollu-
tion estimates for Queens and the Bronx were
based on measured values or on relationships
with citywide data, as described in the following
paragraphs.
Annual arithmetic mean dustfall values for
New York City are shown in Figure 5.1.21. In
addition to citywide averages, data are presented
from Bronx, Queens, and Manhattan. The values
shown for Bronx and Queens prior to 1960 are
borough averages. Starting with 1969, dustfall for
Queens and Bronx was collected at the present
CHESS sites. For 1960-1969, values for Queens
and Bronx in Table 5.1.A.8 were obtained by
interpolation. Major decreases in dustfall levels
occurred during the periods 1944-1952 and
1962-1969. For example, the citywide value in
1969 was only 14 percent of the 1944 value.
Long-term citywide trends for sulfur dioxide
concentrations were determined by annual
summarization of monthly data for 1958-1970.
The citywide sulfur dioxide data for 1956-1957
were projected by graphically estimating the ratio
of sulfur dioxide concentration to dustfall and
multiplying the annual dustfall by the appropriate
ratio. Citywide data for 1958-1970 were available
from NYC-DAR. From 1965 to 1969, NYC-DAR
measured sulfur dioxide in Queens and Bronx,
providing an indication of average values for the
boroughs. Prior to 1965, all values were esti-
mated in the same manner as the 1956-1957
citywide values. Sulfur dioxide data for these
areas were collected at the present CHESS sites
after 1969. Most of the NYC-DAR measurements
utilized the hydrogen perioxide technique; how-
ever, at some locations an electroconductivity
method was used.
As indicated in Figure 5.1.22, dramatic reduc-
tions in average sulfur dioxide concentrations in
New York City have occurred since 1964. How-
ever, the 1970 mean ambient levels still exceeded
the National Annual Primary Air Quality
Standard, even though they were approximately
66 percent lower than the 1964 values.
Annual arithmetic mean concentrations of total
suspended particulate from 1956 to 1970 are
shown in Figure 5.1.23. Citywide data for
1956-1957 were estimated by using the observed
time-dependent ratio of total suspended partic-
ulate to dustfall. Citywide data indicate that
suspended particulate concentrations increased
from 1958 to 1963. Since 1963, suspended
particulate concentrations have decreased by 63
percent; however, the 1970 mean annual ambient
level of 105 jug/m^ was still in excess of the
National Annual Primary Ambient Air Quality
Standard (75 Aig/m-'). In both Bronx and Queens,
annual values were estimated by multiplying the
appropriate annual dustfall values by the citywide
ratio of total suspended particulate to dustfall.
Suspended sulfate data obtained from the
121st Street sampling station in Manhattan are
shown as citywide values in Figure 5.1.24. City-
wide suspended sulfate data for 1959-1961 were
obtained by interpolation. The observed annual
ratios of suspended sulfate to dustfall for New
York City were used to estimate the suspended
sulfate levels in Queens and Bronx. These esti-
mates indicate that suspended sulfate levels,
which show a gradual decrease since 1960, are
quite similar in Queens and Bronx.
New York Studies
5-19
-------
• QUEENS
• BRONX
» CITYWIDE
• MANHATTAN
o QUEENS
a BRONX
v CITYWIDE
O MANHATTAN
1944
1950
1955
TIME, years
Figure 5.1.21. Dustfall concentration versus time in years
for New York City.
100
1956
I960
1965
TIME, years
Figure 5.1.22. Sulfur dioxide concentration versus time in years
for New York City.
5-20
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
50
QUEENS o QUEENS
BRONX o BRONX
SOUTHHAMPTON (RIVERHEAD) A SOUTHAMPTON (RIVERHEAD)
CITYWIDE v CITYWIDE
1956
Figure 5.1
York City.
1960
1965
1970
TIME, years
.23. Total suspended particulate concentration versus time in years for New
O
O
£3 10
• QUEENS
• BRONX
CITYWIDE
o QUEENS
BRONX
v CITYWIDE
1956 1960 1965 1970
TIME, years
Figure 5.1.24. Suspended sulfate concentration versus time in years for New York City.
New York Studies
5-21
-------
Annual NYC-DAR Manhattan dust count data
from 1958 to 1971 are shown in Figure 5.1.25.
The curve indicates that the dust count, related
primarily to small particles, decreased by a factor
of 5.6 from 1959 to 1966, then started an
upward trend that continued through 1971.
These recent increases indicate that, although
total suspended particulate concentrations are
decreasing, the quantities of respirable particulates
could be increasing.
A summary of COH data for New York City
is shown in Figure 5.1.26. City wide annual COH
values decreased from 2.5 in 1960 to 1.0 in
1971.
3.
u>
s
o
CJ
1958
TIME, years
Figure 5.1.25. Dust count versus time in years for Manhattan.
o QUEENS
n BRONX
A CITYWIDE
1960 1965
TIME, years
Figure 5.1.26. Coefficient of haze versus time in years for New York City.
5-22
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
SUMMARY
REFERENCES FOR SECTION 5.1
CHESS data and available data from the New
York City Department of Air Resources have
been utilized to provide a comprehensive data
base that is representative of historical air quality
in the New York City metropolitan area. Pollu-
tant levels for selected time periods are summa-
rized in Table 5.1.4. Collection methods were
compared for equivalency and correlation. Where
data were not available, estimates were made
based on the best available information. These
data show a consistent improvement in air quali-
ty for all .pollutants considered. Monthly sulfur
dioxide concentrations follow a cyclic pattern,
with peak concentrations occurring in January.
No such pattern was exhibited for total sus-
pended particulates, respirable suspended partic-
ulates, or suspended sulfates.
2.
Riggan, W.B., D.I. Hammer, J.F. Finklea, V.
Hasselblad, C.R. Sharp, R.M. Burton, and C.M.
Shy. CHESS, a Community Health and Environ-
mental Surveillance System. In: Proceedings of
the Sixth Berkeley Symposium on Mathematical
Statistics and Probability. Berkeley, University of
California Press, 1972.
CHESS Measurement Methods, Precision of
Measurements, and Quality Control. In:
Health Consequences of Sulfur Oxides: A
Report from CHESS, 1970-1971. U.S. Envi-
ronmental Protection Agency. Research Triangle
Park, N.C. Publication EPA-650/1-74-004. 1974.
Table 5.1.4. ARITHMETIC MEAN POLLUTION CONCENTRA-
TIONS. NEW YORK CITY, 1956-1971
Pollutant and
location
Sulfur dioxide.
Citywide
Queens
Bronx
Total suspended
particulates.
/ig/m3
Citywide
Queens
Bronx
Suffolk Co.
Suspended
sulfates.
Mg/m3
Citywide
Queens
Bronx
Dustfall,
g/nrWmo
Citywide
Queens
Bronx
Suffolk Co.
Pollutant levels
1956-
1959
423
402
359
172
164
147
-
25.5
24.8
22.3
24.1
22.6
20.3
—
1960-
1963
478
424
362
257
228
195
-
26.4
23.3
19.9
23.7
21.2
18.1
-
1964-
1967
452
423
445
182
139
152
-
28.2
14.6
15.3
16.8
13.2
13.9
—
1968-
1970
227
174
247
99
84
108
41
20.5
8.6
14.8
8.7
6.7
9.7
2.3
1971
105
94
107
105
98
105
40
-
10.1
14.4
8.4
8.0
7.9
-
New York Studies
5-23
-------
3. Air Quality Criteria for Particulate Matter.
National Air Pollution Control Administra-
tion, Public Health Service, U.S. Department
of Health, Education, and Welfare. Durham,
N.C. NAPCA Publication No. AP-49. January
1969.
4. Collection and Analysis of Dustfall (Settleable
Particulate). In: Annual Book of ASTM
Standards. The American Society for Testing
and Materials. Philadelphia, Pa. ASTM Test
Method D1739-70. 1970.
5. U.S. Environmental Protection Agency.
National Primary and Secondary Air Quality
Standards; Reference Method for the Determina-
tion of Sulfur Dioxide in the Atmosphere (Para-
rosaniline Method). Federal Register. 36
(84):8187-8190, April 30, 1971.
6. Air Quality Criteria for Sulfur Oxides.
National Air Pollution Control Administra-
tion, Public Health Service, U.S. Department
of Health, Education, and Welfare. Durham,
N.C. NAPCA Publication No. AP-50. January
1969.
7. Review of Air Quality in New York City.
New York City Department of Air Re-
sources. New York, N.Y. Internal memo-
randum. January 1972.
8. Particulate Matter in the Atmosphere — Opti-
cal Density of Filtered Deposit. In: Annual
Book of ASTM Standards. The American
Society for Testing and Materials. Phila-
delphia, Pa. ASTM Test Method D1704-61.
1970. p. 498-505.
9. Ferrand, E. Personal communication to T.D.
English, U.S. Environmental Protection
Agency, Research Triangle Park, N.C. New
York City Department of Air Resources,
New York, N.Y. November 1971.
5-24
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
APPENDIX
CO
g
00
£E
§ g
I i
uJ O
o 1
UJ S
Q O
Z "~
CO CO
11
CO
UJ
o
o
o
x U
2 CO
Q [2
rr X
D O
co j;
5 S
< z
Q ^
. O
'"I ""
<
^"
LO
Observations
exceeding
standard^, %
(O
c
o
•r-
4-4
a
_Q
•c—
T3
U
C
Ol
3
CT
U-
Concentration, vg/m^
CM
•o
to
"O
4->
CO
S
in
o
o
0
LO
o
CO
o
r-H
Geometric
standard
deviation
U
1 "£ E
O 4-> ro
CD 0) OJ
13 E E
1
JZ (J
4-> -i- E
•r- +-> CO
S- OI 0)
<£ E E
Maximum
Minimum
i
ro
> in
!- C
(U O
in -I-
S^
c.
1- 0
O -r-
-)->
"O
'E 01
""• ^3 CO ^" C3 CO CTl ^O tO CO <~~t CT^ CJi *~l CO
I— Ir-t r-H r-< i— t CM CM t— 1 i— 1 CM *— 1 i-H CMCMCM f-HCMCJ
CMCMLOUDOO^O LOLOCOOCOOCO «— I «— I CM O ^f P^. CT)
rHi — 1 i— 1 i— 1 «— 1 i— 1 i-H CM *— 1 i-H i— t T— t
LOCOOJlOrAJOCO COIT>CMCT>CNJ-HVO I^I^CM-.^--HCO
^f VO ^f *~^ r~' C*) 00 CTi 00 CO CO CO ^" ^O !**•• CM "^" CO CO ^f P"^
?
L0-.r^00000^ M-CMVO^CXJO^co CMVOrtOVO-HCO
i— t »— 4
LO P^ r-n ^0 oo ^^ ""^ c\i uo r*» co CTt LO CM to co CM oo P*^ CT> co
COP^-LOP^.COLOCM P^.rOP^-CTiCOCO'O COLOCTtOOOUO
rHr-Hi — 1 CM'—'CO LOCO'S'CMLOCOLO VoO CO «— * CM VO CO P^-
O CM P** COLOOCM LO "^J" LO ^O CO P** t— I ^ LO CTl P^ IjO t— I ^-
^ococo^coj-, ooco^oooocvj^ r-vqin^^voo
^o^cMLnco.*^ CM^o^in^ocn ^I^^^^-OCNJ
»-H CO CM CO ^O CM {Q CO »~H O^ LO «— * CO »*H CT> LO < — ' «~H CO CO P"^
t— 1 r- 1 i— 1 i— 1 i — 1 CM *-*•
i— i O O CM CT) •— i O «-HO^OOO^OO COCMLOCO^DCOO
i-.^-cvJococoo. ^^ocoi-r-o, ^,^1000-.^^
T— 1 i— 1 i — 1 CM'~|CM «.3-COCOCM*±CO^f VO CO t— 1 i— 1 IO CM Ijp
co^cMcn^.«>r- ^.OI--OVCM-.V£> o^min^mw
i — i co *~^ co co CM cr> LO r^x CM •"• cy> co cji oo LO co r*1- co CTI w>
^^-vo^-o.a.Ln t-,^00^-0,^00 ^ |»— (CO
1 1 1 1 •— III I I 1 1 I— III I 1 1 1 i-H 1 1
Ol
•E S
Ol Ol C
> 01 0
n^ O" CD
E
0)
t
01
Q.
U
O
O
8"
= °°E
§
•O •>
i- -a
10 i-
-D 10
c -a
10 E
+•» ro
ro i— -i-
r^ rO i—
CD" 5
S~
E
•o
O)
+J
ro
CD
3.
(/)
(U
O
C
O
O
rO
|b
E ro
C .^
rx s-
D_
ro ro
E E
O O
«~H CM
H
fO fO
x> -o
c: c
fO (O
New York Studies
5-25
-------
o
LU ;
oe
"- X
II
2 O
o s;
UJ LU
O fc
Z LU
O <"
°
LU LU
11
Q
LU LU
Q I
H O
S *
I O
So s-
_j S
S s
O re
< z
Q g
CNJ D
rf 00
< ~
10
« X en
> 0» c
i. -r-
(A C
J3 O (U
4J
£
"•x.
3.
O
£
S-
-u
OJ
c
o
O
&«
«t
-o
s-
T3
C
+->
i/l
fO
o
-i->
-Q
-*->
t/l
S>
01
cr
O)
V
U-
O
•i- T
i- I
+J n
<1J T:
O a
at
c
5!
i
_C (
•i- 4
i- C
-
-o
+j
oo
Ol
CTl
in
CT*
01
r-^
LD
ro
•~
c:
3 O
5 •)->
3 IO
0 >
J Ol
n T3
J
- c
J 0)
) E
J
J 10
i i
E
ZI
X
z;
E
3
E
c
dE
3 c/1
> C
- O
1) «-
n •>->
D
3
C
= 5
E>
^ (U
i; f)
I. -Q
O
Community
OOOOOOO C3OOOOOO OOOO«d-"X>O
OOOOOOO OOOOOOO OOOOCMOO
o^ ix. i — OOIX.CMCSJ <^5rx.r*vcn,Lnrx.i — i — tprof'orxcof**'
coroi — oo«^-ro> — ^d-LO^cMco^-m oof^^cxJLO0^
CXJi — •CMf>"'OvJC\IC\J COLDLnLOtDLOLn
ixsrocMr^cTir^oo orootnr--c\]^o LOr--.OOr-.tQ'=3-
•d-roCTiLOLoroLn c\jp-.OcnofriLr> • — co*=t-cr». — oo
oo^ocxirxcoLoco ro
LOuiLnLOLnLn^- com
i — rorocvjai^^ « — LOJ-o^-rOr— tr> oojcoi — ( — o>u>
cnkof^-ooo^rx, oror— ^oooio u> m o f— to vo co
oororo«d--«^-roro r^.\or*xrxrx.(X)r^ px^ocnocncrt
Loco^ooicrto i — roi-^i — i — i — o nooo^rxooLO
i — cocr><3-cno^ro tOLorxorx,rN.i^> rxLO^orx^co^o
ro.r— <± o o c\j co r— CM
cNic\jcoi^><^-<^-co CSJ^OVOOLO^OI in co rx. r— o O co
cr»'xfco<*-p^roix. <*-^coocsjc\jLr) coco^o«3-a^rx.c\j
r--Lr>i-O(7icorx,c\j CNJI^LOOLO^LO cOk.ocj>i — a^coo
r— i — i — i — r-i — CNJ rorom«3-rop'}ro co^fCO^'d-^j-Ln
cOLDUDLOLncorn ^LD O^VDfOi — CMCOCSJ UDh-NPxO^^^O
Cftmcnocoroo cOLncnLorxcni— r- co •— i— TX> rx, en
rx.oo^oOLnoorx. counLOLno^Looj ir> co *st •— o r— o
O^J-crir-'O^c\j roc\jooc\jLO LoroocxJinoco cnrx.or^rorx.cM
pxcooocooo<^-co ^£)0ocr»cx)co^rx rxr-xcncocof^co
ro no ro
CNjOO^OCnCMCMOO CMOO^DCTiCMCNJCO C\JCO'*DCT>OJC\jrO
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
\--v\-^-^*X,-^ ^^•^,-^-^X^-^-^ •Xv-^'^-X.-^-X 10
oo +->
i r-5
U -i- S-
<•--
T3
Ol
S- >>
10 S-
E 10
Q- S.
a.
E 10
< Q
IO IO
o o
IO IO
I I
c
01
o
s- s_
10 IO
•a -a
c c
IO to
4J +J
OO OO
5-26
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
O
I—
CO
c/)
D CM
>• r*.
O 05
1 5
O ce
LJ <
°
en
CC
W w
O LU
!I
o
a ty
LU T.
o o
z ^
LU Q;
i I
>- ty
CO
E
CD
3
C
o
ro
0) TD
O IT
0) 4-
O tr
c
0 4-
1
•r- 4-
< s
1
It
>
S.
a
i/
_c
c
M-
C
T:
c
s.
a
a
CJl
Ol
Ol
LT)
o
o
c
o
4->
CO
>
CD
-o
rO
E
c
CD
E
3
X
E
E
c:
•E
o
) '!-
4->
C
O
•r-
>
S-
(U
) V)
XI
o
Community
ioi^r-~ooi^u3 o 01 01 .-i *a- •* *3- •*
ro ^~ CM ^f" CM co ro CM ro co co ro co ro co co ro LO co ro CM
i — CT* CM CO CM OV CM (71 CM r-- COCOr^-CMCO*^3CO
CM CM r— CM CM CM CM CM CM CM CO CM CM CM CM CM CM CO CM CO CM
ooa>o^oo^-u3 ^I^CMCM^-C^CO LO^CMVOOI^CO
CMr*-.toi — cncOr— CM CT> r^. r-- CM •— • •— < •— < CM CM *-» LO ^f o
CMi — i — CM. — i — CM CM •— I i— ( CM CM CM CM CMCMCMCOCMCMCM
CMLOl — i — COLOCO COCOOO^OO^On CO ^" "— i CJt CM O^ 00
LO^-COCMCMr-CM I^^J-'^r^LO^-^ LOtOCOCOtOLOLO
COP^OCOf— LOi — i— I CO «— I <± CM CO «* U><— iCOCMCOLnr- 1
r-CO^OO^OCOO ^-r-IOlCMCM-HCM CO^-C^^-CMCMCO
r '
r^^OCS.^CM^, ^OCM^CMOICO r-Ol^CM^^-Cv,
i-H I i— 1 r— 1 t— 1 f— 1
LO 00 f~~t ^" CT) CM CM C3 ^~ CO CO i— I LO 00 CO «~H O"l CO "sD P*^ <^
rAj^t-ro^i-Ln^r^ coootouir^^co LOCO^-ID^UJOO
00,0^,-^ocvj S%5K^gS SSSS3S8
ooooooo ooooooo ooooooo
U3COLOCVJCVJL001 COr-^^rH^CM OlLn^^^U,^,
OlOO<£)dOOOO CMCMO*tCMCM«-HCO r-HCOOCOCOCMCO
^-^-O^CMCOCMO LOr^COr-LOCM-. LO(-sOCMP^COO
CM CO r~-. r~~ r~ ^3 CM ^" CO t~> ^" CO CO ^~ CO ^" CM ^Q ^" ^f *^"
oo^-tDOiro^-^- OOO^-LOCOCOCM r-,^.kocn-«oir-.
r*v r-» i — i — Oi — LO OOOl^OlDCOCTtCO LO LO •— < CM ** CNJ CM
^3" LO CO *^~ CO LO CO CM CO CO CO CO CO CO CO CO ^" LO C^O LO CO
o^-^oo^-cn oo^-a.co^>o.c^ ooror^r-ro^
OCMi — COCO. — «d- O LO •— I CM CM r-t CO OLOCMCOLOCM^D
CO CO CO
O r— r— r— i— r— CM
•~~ I~" r~* CM CO tO O"* CM CM CO CM CO VO CH CM CM CO
1 1 1 1 1 1 1 ,-4 ^Ht— 1 rH .—It— 1
° ^~ *~~ r~ ^~ ^ S^ iiitiii itiiiii
i— ) r-H
"O
OJ
-C t/>
S- C X
OJ O
oc cr CQ
4
V,
<
(
(
t
t
M
(
4
t-
J.
T
4
J
4
-
V.
•
4
-
C.
ni
c
ai
o
01
a.
u
o
cr
a>
S-
3
E
-a
-------
Q
I s
LU O>
2 =
S o
LL 6C
Q i
co
— LU
i- oa
< ^
CC LU
O LU
O r;
< i
oc S
LU CO
i I
a o
CO ^
2 cc
CO O
Q ^
cc
* O
. u.
<
in
^
Ol
«l
G
o
4->
fO
S-
4->
0
C
O
ro
O
J!
•M
•5
5>
Ol
OJ
t-
u
•1- 1
i.
+J
O> 7
O
01 •>
CO
1
O ^
0
i
_c
•i-> •
s-
CC
.
C
c
01
01
LO
O
O
o
o
o
c
3 0
0 •*->
3 10
a >
-> 0)
/> -a
j
= E
J
- c
1) (U
E E
E
a
s
E
E
£
a
1- c
ai o
l/l •!—
Q +*
D
C
1- O
O -r-
•)->
O ro
O >
S- Lnror^io^-
oir-cooOLnocM o^-ioicoo.^ ^-.^^cntom^
t— 4 i— * «— * i— t i— t
ooooooocoo co CM «— * o vo oo **• cno^ovo^"r~Lr,cv,^oi*vo r.cor-.oOLOoi^ COOOOOO^-OLO
OOOOOOO O O O •— 1 «-H O T-I CDCDOCMi— lr-tO
co ^~ ' — i co r*"-- CTI ^~ co c^ co r*^ r*^> CM co **"" LO co co CM LO o^
OOOOOOO OOOOOOO OOOOQO--<
^OCO^LOOOI Ln^-^oio.ocM «t «. o, - oo w *
^^o^^a,^ ^r^^^cvjL^o ^cnooco^^f-
i — Ir-Hi — 1 CM CO i— ' CM ^H i— 1 CM ^~ LO CO ^" rH i-H ^J- ^- LO ^~ CO
„,__
f-H i— l CM CM t— 1 CM •— I •— • CM i— t
CMCMOO-.0- COCM^O^-OCM -O^LnCMOO
OOOOOO'O OOOOOOO OOOOOOO
CO CO CO
C\J CO CC.J CT> CM CM CO CM CO lO CTv CM CM CO CM CO ^.Q O^ CM CM CO
1 1 1 t 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
i—1 i— ( i-H
0)
•E £ x
QJ CD C
> QJ O
•^ z» i-
CC O- CQ
o
i.
01
o.
>
-Q
-o
(U
-o
ro
o
+J
ta
-------
tn
O
i-
oo
oc
O
UJ
O
UJ
U. CM
< I
w O
*
UJ >-
(A
UJ
D
U
2 °
CO
Q tf
a.
3
I?
a
m
Q-
£2
oc
D
in
<
r-
LO
co
E
CT
,}
entration ,
0
c
0
o
1
en
CTi
C LD
O CTt
+J
;3
r g
4->
l/l
-o
>, s
u
c
QJ
3
CT 0
C1J LO
LJ-
0
co
0
Geometric
standard
deviation
o
1 £ C
O -t-> t3
CL> CL1 CU
C3 E E
I
-C O
+J -i- C
M- 4-> (0
S- 0) CU
< E E
Maximum
Minimum
Observa-
tions
c
0 •!-
-t->
~O (U
o >
-r- i-
O) i/i
0- -0
o
Community
00
Ol
LO
r~-
LO
LO
r-~
LO
ro
LO
co
LO
in
CM
< — i
o
C\]
LO
LO
LO
LO
O
•3-
r^
CM
CTI
i— t
ro
LO
C\l
01
«s-
CO
ro
10
,_,
01
i
i — i
r-~
-o
(C
01
xi
S-
O)
>
cr:
Ol
LO
CO
CO
CO
r-~
r-^
r^
o
01
CO
CO
i — i
ro
cr.
O
C\J
OJ
n
LO
LD
o
CO
CT1
CO
CD
LO
n
»— t
ro
01
o
0
! —
^H
CM
i — 1
O
O
LO
CM
CTi
CO
CM
LO
CO
Ch
LD
CM
•-H
t— t
CM
LO
O
LD
CO
0
LO
r\
t^
CM
r\
LO
i^
o
o
ro
rv.
C\J
CO
i
CM
^
CM
CO
LD
O
LO
LO
r--.
co
LO
LO
CO
«3-
r^
cn
sa-
LO
^f
ro
r-^
CM
CM
CM
LO
O
LO
r-~
CO
^~
i — i
*!•
CO
CO
LO
CO
LO
*d-
ro
^H
O!
i — 1
r-~
t/i
c
QJ
O)
13
Cr
LO
CM
LT>
LO
- i — 1
CO
LO
CO
1 — ,
CO
LO
•=1-
LO
LO
CO
CO
01
.— )
^
LO
CO
o
co
CM
CM
r^
CO
CO
LO
01
>3-
CM
1 — 1
CO
^-,
CM
O
1 — 1
0
CM
CO
"=f
CM
LO
LO
r-^
«3-
LO
CO
o
00
CM
CM
O
CM
«»
CM
CO
r^
O
•51-
CM
^1-
O-i
CM
LO
C\J
CTt
Ol
CM
LO
CO
CM
n
1
CM
\
. — 1
r^.
LD
00
O
CM
CO
CO
ro
r^
CM
LO
LD
^3-
CO
LO
O
•3-
LO
•*
CM
LO
*a-
o
LO
CO
^•
OT
CM
LO
r^
00
CO
o
r-i
>a-
LO
i — i
CT>
!~~
X
c=
o
i.
CO
o
CO
cn
Ol
CO
o
o
•a-
co
o
o
LO
*3-
O
*3-
CO
CO
CM
CO
Ol
I— t
Ol
LO
O
1^
CO
CO
LO
CO
*±
CO
LO
Ol
1 — 1
1-
01
LO
i— <
CM
1
O
t— 1
CO
CO
^J-
co
r-^
cn
o
r*^
o
0
LO
ro
LO
CO
LO
CM
O
01
>a-
o
i — i
CM
CO
CO
O
LD
r^
c-g
CO
1
CM
---.
1 — 1
+->
c
,
U
E
3
cr
*0
3
u
>)
&
stributed
-o
ro
E
01
3.
in
c.
o
Concentrat
New York Studies
5-29
-------
QC
CM
> ft
O ?
LU X
D O
O «:
LU <
& S
u. i
CO
LU LU
O P
o z
0 ^
LU §
9 1
% o
2 co
Z X
LU Q
0 -
o
< cc
0 O
U-
co
in
Si
CD
Observations
exceeding
standards, b %
.. -
Frequency distribution3
OJ
T3
CO
-o
cn
LO
cn
o
cn
CD
p^
O
in
o
CO
0
l-H
O E
1- i- t~
+-* ra -t-3
(U -O ro
E E -r-
ai -M ai
t n (/) ^3
O
0 '£ E
ai 4-> ra
C3 QJ CU
1
CO .E U
E 4-> -i- E
<; -r- 4-> ro
cr f- a> (u
3. -
+j ra
ai
8 1
E
E
3E
i
i "°
> in
t- E
ai o
U) -f-
J3 -U
o
Period of
observation
z
c;
^
o
c_>
O
O
o
o
LO
LO
O
LO
LO
CO
cn
o
o
oo
o
o
OJ
o
LO
OJ
CO
LO
0
OJ
OJ
rv
OJ
OJ
cn
LO
O
CD
rv
oo
1
-
Riverhead
o o
0 0
o o
o o
in o
OO LO
CO «sP
cn co
CO OO
OJ 00
LO r —
in co
OJ OJ
CO CO
LO i—
r— OJ
00 00
OJ LO
rv in
co o
i — CO
in LO
*d- co
LO LO
0 0
rv LO
^ ?
CO OJ
00 p~
rv in
rv \o
CD PS-
CD CO
in ps.
CO CO
LO cn
i i
^ P^
0 0
0 O
O CD
0 0
OJ LO
i— CD
co rv
rv LO
rv co
LO LO
OJ 00
LO cn
in co
OJ i —
LO ^}-
00 CM
in rv
in LO
CM i—
r- LO
in i —
cn LO
co cn
0 0
•a- rv
CO LO
OJ I —
o rv
cn o
OJ OJ
rv rv
LO LO
CO CO
LO O
LO O
1- 0
rv oj
CO
•—
OJ CM
1 1
O i—
o
0
^.
"
OJ
oo
OJ
in
LO
IV
LO
sf- in
00 OJ
OJ
• cn
- -a-
cn co
•—
cn CM
i i
Pv O
CD O
O 0
•— IV.
CO LO
CO LO
OO i —
co PV
co co
o cn
CO CO
cr> co
CO CO
cn. rv
LO LO
0 co
>*• o
in LO
i— cn
5- >*
LO CM
CO *,T
OJ CO
LO O
LO «*•
0 0
cn ^~
cn in
«* in
OJ LO
rv cn
m in
•a- rv
CO CO
cn
cn
•el- ,-
OO ps-
OO CO
oo
r— OJ
OJ CO
1 1
i— CM
- -
CD o
o o
pv cn
oo rv
•sf cn
OJ Iv
OO OJ
PV ,-
pv in
i— 0
CO CO
OJ LO
r— cn
i — OJ
OJ i—
CO Pv
CO OO
pv cn
LO in
i — OO
r- LO
LO ^3-
CM O
cn oj
i — CO
O cn
O 0
cn co
LO LO
LO in
r-s rv
LO i—
\& LO
oo cn
LO cn
CO LO
O oo
O CO
oo cn
r- CO
r— rv
CO LO
X
c.
o
i-
ca
0 0
0 O
cn o
Sst- LO
^j- in
• — in
rv co
oj cn
o co
CO CO
CO CO
i — in
I — OJ
CO CO
i— co
in co
LO LO
00 Pv
OO OJ
in in
pv i —
OO r—
CO *3-
CO LO
in
-------
Table 5.1.A.7. QUARTERLY ARITHMETIC MEAN POLLUTANT CONCENTRATIONS
IN NEW YORK CITY CHESS COMMUNITIES
(yg/m3)
Pollutant and
location
Sulfur dioxide
Ri verhead
Queens
Bronx
Nitrogen dioxide
Ri verhead
Queens
Bronx
Total suspended
parti culates
Ri verhead
Queens
Bronx
Suspended sul fates
Ri verhead
Queens
Bronx
Suspended nitrates
Ri verhead
Queens
Bronx
Time, quarter of year
4(1970)
20.7
63.1
80.7
45.5
102.2
118.9
35.1
61.2
69.5
12.0
14.3
13.9
1.4
1.7
1.6
1(1971)
32.2
60.9
74.5
40.6
63.1
70.9
33.5
59.6
81.6
10.4
13.8
14.9
1.3
1.6
1.9
2(1971)
21.5
53.8
41.1
13.8
47.8
61.8
32.6
63.1
84.9
7.7
10.9
12.0
1.0
2.7
4.3
3(1971)
10.1
31.7
26.7
17.2
55.9
70.2
36.7
66.8
82.4
11.1
14.8
16.1
2.1
4.6
4.8
4(1971)
32.0
60.3
70.7
30.7
64.9
74.2
34.8
63.7
97.9
11.3
13.5
14.6
3.1
5.3
5.0
1(1972)
40.4
63.0
88.7
27.7
59.5
73.8
33.8
65.6
86.2
11.5
14.0
14.0
2.5
4.0
3.7
New York Studies
5-31
-------
CO
o
cn
i
«*
S
o
^
S
LL.
CO
C *•""•»
O CO
o eu
.c
-i£ S-
i— eu
o >
i- -i-
f- CZ.
V}
O>
oo
GO cn
1— 31
t— O
i— E
IO \
E
3 cn
O
oo
uo ^~
t/0 cn
" E
CL. \
1— 3.
" E
CM\
O cn
GO p.
r-"o
10 --,
-P E
Q
CO
oo cn
* E
Q- "-s.
OO cn
H- P-
" E
CM*--.
O cn
i— O
m •-•*.
3 cn
O
CO
to Cn
00 3.
CO
- E
a. -*•*
oo cn
h- 3.
•CO
*
51
3 Cn
Q
t-
rO
cu
CO CO
CO ,—
CM CO
*± in
in r—
in in
co r--
r-. i—
in co
0 0
to to
CM CM
CM O
i— r--
r— cn
00 CO
i— CM
O O
00 O
CM CM
CM CO
CM •—
«3- co
O Cn
0 0
oo cn
o cn
CM LO
in co
CO
1 —
co r-* i —
co o «*
CM r*- co
cn o CM
CO i— O
ro «tf- CM
o r- co
co cn CM
r- CM *d-
CM co in
r-v co <3-
! — co m
O **" CM
CO CO ^J"
o o o
•— CM •«-
CM CM r-.
r— CM CO
CM CO r-
o^ co r-.
CM r— CO
to ^ cn
o o o
cn CM to
CO P*. to
cn cn CA
to co co
in i— • • cn
in r-* co
to ^- in
'd- CO CO
CM CM CD
* 10 CO
«!••-,-
ID r^ co
o »}• •*
ir> vo i —
4O IO CO
* •§ CO
O IT) IT)
^S- IT) IO
00 i — CO
CO m i —
oj n co
O CO CvJ
cn eu co
i — CM CNJ
I/} CJ> ^f
^ CM CNJ
O o r-.
eo i— i—
• — CM CM
in ^ ro
IO I>O ^O
cn
CM
CM
•-
co
r-~
oo
cn
a>
o
CM
o
oo
CM
CM
cn
ro
CM
O
VO
en
,-
VO
CM
cn
IO
CO
CO
oo
CM
in
in
f>
in
o
CM
CO
o
oo
CM
o
CO
in
^
CM
cn
in
cn
CM
to
CM
in
in
5
o
CM
r-
CM
a>
CM
CO
%
.—
0
CO
CM
CM
in
to
CM
CM
in
CM
CO
in
cn
in co
**• CM
CM CM
CO CM
r*. in
r- in
o <*
co co r-^ to CM o CM
CM o cn cn co cn m
CM CM r— r— CM r— CM
in cn
CM CO
CM •—
tO CM
cn CM
to cn
CO CM
in CM in CM CM i-* «*
o r^ un to to *3- r-.
r- CM tO CO
CM ^ LTi CM
CO CM CM CM
co" in
co m
CO CM
CM m
^ —
,>*. to co 10 to to o
CO CM CM r— CO O in
CM CM CM, CM CM CM CM
i-x to in ^- co CM r-
in m m in to to m
cn cji cn c^ en cn cn
CM r- CM co co PN. rs.
co co cn r- cn o to
CM CM CO "^ CO LO in
cn CT^ cy* cn cn cn c^
w
=>
>
cu
t-
3
1/1
•8
C
0)
^
0)
•D
**~
2
s.
cu
=
o
• IO
c/> 4J
cu •(->
4-> CU
C i-
H. ti
IO
a. 4->
•r- CM
(/) +J
i — C
IO IO
> s
"O E
CD O
4-> S-
i *-
V) (Q
LU O
rd -Q
5-32
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
5.2 PREVALENCE OF CHRONIC RESPIRATORY
DISEASE SYMPTOMS IN ADULTS:
1970 SURVEY OF NEW YORK COMMUNITIES
Harvey E. Goldberg, M.D., John F. Finklea, M.D., Dr. P.H.,
Cornelius J. Nelson, M.S., Walter B. Steen, B.S.,
Robert S. Chapman, M.D., Donald H. Swanson, B.S.,
and Arlan A. Cohen, M.D.
5-33
-------
INTRODUCTION
A number of studies have documented the
adverse health effects of air pollution on people
with established chronic lung disease. Mortality
from bronchitis has been observed to increase
disproportionately during air pollution epi-
sodes.1"10 Morbidity of people with chronic lung
disease was also increased during air pollution
episodes, as reflected by the number of hospital
emergency clinic visits that occurred during two
episodes in New York City.11'1^ Furthermore,
prospective studies of large groups of bronchitic
subjects have shown a positive correlation be-
tween the day-to-day variation in acute symptoms
and air pollution levels.13"15
The place of long-term air pollution exposure
in the etiology of chronic lung disease is harder
to define. In studies of acute symptoms, one can
usually record factors influencing the incidence of
illness, such as smoking habits, occupational
exposure to toxic substances, age, place of resi-
dence, air pollution exposure, and socioeconomic
status. Such factors usually are constant for the
duration of the study. However, chronic disease
investigations must deal with fluctuation of these
important covariates over time. Prospective studies
that obtain detailed accurate covariate informa-
tion have been prohibitively expensive, while
retrospective studies require an accuracy of recall
over many years that is difficult to attain. There-
fore, most chronic disease studies have been
cross-sectional comparisons of symptom rates in
different areas at the same time.16"21 These
studies have usually assumed that air pollution
exposures, smoking habits, socioeconomic status,
and place of residence of subjects at the time of
study are fair approximations of these charac-
teristics over a lifetime. It is also not possible to
find communities alike in every respect except
exposure to air pollution. Thus, the evidence
adduced from any single cross-sectional study is
relatively weak. If the apparent associations
shown between air pollution exposure and
chronic lung disease are to be considered credi-
ble, they must be found at different times by
several investigators in different areas.
In the present study, three middle-class com-
munities were chosen to represent a gradient of
air pollution exposure. These communities are
part of the Environmental Protection Agency's
Community Health and Environmental Surveil-
lance System (CHESS) program. Parents of
elementary school children in these communities
were asked to provide information about their
own respiratory symptoms. The study had three
purposes: (1) to determine whether chronic
respiratory disease symptom rates in adult popu-
lations, age 20 to 50, corresponded to a gradient
for past or present exposure to air pollution; (2)
to initiate a prospective study to determine the
effect of pollution control on chronic respiratory
disease; and (3) to relate chronic respiratory
symptom rates among migrants to their past and
recent air pollution exposures.
METHODS
Community Selection
Available air quality data for the 25 years
preceding the study (1944-1970) were obtained
from the New York City Department of Air
Resources (NYC-DAR) and the Suffolk County
(New York) Department of Health. These data,
in conjunction with the most recent available
census data and current information on house
and property values, were used to select com-
munities alike socioeconomically but different in air
pollution exposure.
•
After considering long-term air pollution
exposure trends, Riverhead, Long Island, was
chosen as a Low exposure community, the
Howard Beach section of Queens as an Inter-
mediate exposure community, and the West-
chester section of the Bronx as a High exposure
community. As a result of recent improvements
in air quality, the Queens and Bronx commu-
nities were found to have similar current pollut-
ant concentrations, often below National Primary
Air Quality Standards. Hence, on the basis of
recent measurements, the latter two communities
were redesignated Intermediate I and Intermediate
II, respectively. On a historical basis, however,
the Intermediate communities have experienced
pollution exposures well above those experienced
by the Low community.
Collection of Health and Demographic
Data
Parents of all children attending elementary
schools located within 1.5 miles of air moni-
toring stations in the three communities were
5-34
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
asked to participate in the study. In May 1970,
each child was given a questionnaire at school to
be filled out by his parents and returned to
school. The questionnaire, based on the standard
British Medical Research Council questionnaire on
chronic bronchitis, elicited information concerning
cough, production of phlegm from the chest,
shortness of breath, as well as age, sex, smoking
habits, occupational exposure to toxic dust or
fumes, and residential mobility. (An example of
the form used has been presented elsewhere.)22
Data on race were obtained from school records.
A random subsample of families, stratified by
race, was interviewed at home to obtain informa-
tion on education and income. This subsample
contained only whites from the Intermediate I
and II communities; both blacks and whites were
interviewed in the Low exposure community.
Assessing Air Pollution
NYC-DAR has maintained monitoring stations
within the Intermediate communities for many
years. Levels of sulfur dioxide (peroxide titration
method), total suspended particulate (24-hour
samples), soiling index (coefficient of haze,
2-hour spots), and monthly dustfall were
recorded. This information was used to recon-
struct past exposures to air pollutants.23 More
recent air quality data collected by the Suffolk
County Department of Health confirmed that
exposures to air pollutants in the Low exposure
community were below the National Primary
Ambient Air Quality Standards for suspended
particulates and oxides of sulfur. However, very
few aerometric data and no emissions inventories
were available in Suffolk County, so no detailed
estimates of past exposures could be made for
the Low exposure community.
Supplementary air monitoring was carried out
by the Environmental Protection Agency at sites
that were also utilized by agencies of local
governments. At each station, 24-hour integrated
samples of sulfur dioxide (modified West-Gaeke
method), total suspended particulates (high-volume
samplers), suspended sulfates (high-volume
samplers), and suspended nitrates (high-volume
samplers) were collected daily. Dustfall was de-
termined from monthly samples. Twenty-four-hour
nitrogen dioxide measurements were also made;
however, the validity of the method used (Jacobs-
Hochheiser) has subsequently been questioned due to
variable collection efficiencies and interferences.
Analysis of Data
Reported respiratory symptoms were used to
devise three indices of morbidity, which were
grouped by smoking status and sex within each
community. First, chronic bronchitis was defined
in accordance with the British Medical Research
Council (cough and phlegm for at least 3 con-
secutive months of the year), and prevalence
rates were compared across areas. Second, all
respiratory symptoms were used to construct a
severity gradient in the following manner:
1. No symptoms.
2. Cough alone for less than 3 months each
year.
3. Phlegm with or without cough for less
than 3 months each year.
4. Cough without phlegm for 3 months or
more each year.
5. Phlegm without cough for 3 months or
more each year.
6. Cough and phlegm for 3 months or
more each year.
7. Cough and phlegm for 3 months or
more each year and shortness of breath.
Respondents were classified according to their
reported symptoms and assigned a score corre-
sponding to their rank on this gradient. These
scores were then averaged for all respondents to
permit area comparisons. The mean respiratory
symptom score thus utilized a greater range of
symptom information furnished by questionnaire
respondents than did the chronic bronchitis
prevalence rate. Finally, since the mean respira-
tory symptom scores are markedly influenced by
the number of persons reporting no symptoms,
mean severity scores were obtained by averaging
symptom scores only for those persons reporting
respiratory symptoms. Severity scores are an
index of severity, given the presence of
symptoms. Bronchitis prevalence rates, symptom
scores, and severity scores were adjusted for the
covariables of smoking, age, and sex using multi-
variate techniques. Statistical significance of
differences attributable to community exposure
was then tested by a linear Chi Square model.24
RESULTS
Environmental Exposure
New York City pollutant exposures were
averaged over periods that roughly reflect
New York Studies
5-35
-------
major changes in fossil fuel usage (Table 5.2.1).
Prior to 1949, the major fuel employed was
coal. Unfortunately, reliable aerometric data from
this period are too sparse to permit meaningful
estimation of sulfur oxide exposures. During the
decades 1949-1958 and 1959-1968, petroleum
fuels were predominant. These fuels often had a
high sulfur content but caused less particulate
pollution than coal. In the most recent period,
1969-1971, much emphasis had been placed on
control of particulate pollutants and the use of
low-sulfur fuels. Estimates of total suspended
particulates and sulfur dioxide for the years
1949-1958 were projected from citywide and
borough-specific dustfall measurements. Citywide
total suspended particulate and sulfur dioxide
measurements were available from 1958 and these
were used to project borough-wide estimates for
Queens and Bronx. These estimates were cali-
brated using borough-specific monitoring data
available for both suspended particulates and
sulfur dioxide since 1968. More recent (1971)
monitoring data for all study communities are
presented in Table 5.2.2.
Two air quality trends are unmistakable.
First, the Intermediate communities were exposed
to markedly elevated levels of air pollutants for
at least 20 years prior to 1970. Second, striking
improvements in air quality, as indexed by
particulate and sulfur dioxide measurements, were
made prior to 1969. From 1949 to 1968, annual
average sulfur dioxide levels in the Intermediate
communities (estimated at 343 to 404 Mg/m^)
were- apparently 4 to 5 times the present
National Annual Primary Ambient Air Quality
Standard of 80 Mg/m^, which is based upon the
need to protect human health. Moreover, annual
suspended particulate concentrations in the Inter-
mediate communities during the same period (esti-
mated at 141 to 173 jUg/m^) were up to two and one
Table 5.2.1. ARITHMETIC MEAN POLLUTION CONCEN-
TRATIONS IN NEW YORK CITY, 1949-19703
Pollutant and
community''
Sulfur dioxide, jug/m^
Citywide
Queens
Bronx
Total suspended par-
ticulate, jug/in^
Citywide
Queens
Bronx
Suspended sulfate.
Mg/m3
Citywide
Queens
Bronx
Dustfall, g/m^/mo
Citywide
Queens
Bronx
Pollutant levels
1949-58°
416
397
343
169
162
141
26
24
22
24
22
17
1959-68
443
404
395
203
173
166
26
19
18
19
17
16
1969-70
210
144
210
101
80
104
20
9
16
9
6
9
aAII measurements provided by New York City Department of Air Re-
sources.
Queens and Bronx are the boroughs in which the respective Intermedi-
ate I and II communities are located.
°Sulfur dioxide and total suspended particulate levels based upon pro-
jection from dustfall measurement.
5-36
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.2.2. MEAN ANNUAL POLLUTION CON-
CENTRATIONS, 1971, NEW YORK CHESS COM-
MUNITIES
Table 5.2.3. NUMBER AND PERCENT OF FAMI-
LIES RETURNING QUESTIONNAIRE BY COM-
MUNITY
Pollutant
Sulfur dioxide, Mg/m^
CHESS
NYC-DAR
Total suspended
particulate, jug/m^
CHESS
NYC-DAR
Suspended sulfates, ng/m^
Suspended nitrates, jug/m^
Dustfall, g/m^/mo
Community
Low
23
—
34
-
10.2
1.9
1.22
nter me-
diate 1
51
63
63
84
13.2
3.5
3.80
Interme-
diate II
51
58
86
104
14.3
4.1
5.13
half times the National Annual Primary Ambient Air
Quality Standard (75 jug/m3 geometric mean).
High annual sulfate levels (estimated at 18 to 24
Atg/m3) evidently prevailed as well. Because of
stringent control measures, levels of these pollutants
have fallen during the last few years. By 1971,
suspended particulate levels, though still somewhat
above national standards, were only half those en-
countered in earlier decades. Sulfur dioxide levels,
which had been reduced by a factor of 4, were within
the recommended standards, and sulfates were
reduced to about 14 jug/m3. Unfortunately, it was
not possible to construct comparable long-term
exposure trends for other pollutants such as oxides of
nitrogen, carbon monoxide, and various particulate
fractions. The available data indicated that differ-
ences in the air pollution levels for the two Inter-
mediate communities were not marked even though
borough-wide estimates for Queens were somewhat
higher than those of the Bronx before 1968 and lower
thereafter. Air monitoring data also indicated that
Riverhead was indeed a clean community with all
pollution levels substantially below those specified by
National Ambient Air Quality Standards. Suspended
particulate sulfate levels in Riverhead were somewhat
elevated, perhaps reflecting the intrusion of fine
particulate from more central areas of the New
York-New Jersey conurbation.
Community
Low
Intermediate I
Intermediate II
Total
Families returning
questionnaire
1367
1529
1868
4764
Percent
response
73
78
85
78.4
Response Rates
Questionnaires were completed and returned
to the schools by 78 percent of the 4764
families to whom they were sent (Table 5.2.3).
Response rates were somewhat better in the
Intermediate areas (78 and 85 percent, respec-
tively, in the Intermediate I and II communities)
than in the Low pollution area (73 percent). If
white respondents alone were considered, area
differences in response rates would be lessened,
with the Low and Intermediate I exposure areas
becoming almost identical. It was not possible to
interview nonrespondents personally.
Characterization of the Study Population
The study population was predominantly white,
although a sizeable black segment resided in the Low
pollution community and a small black segment
resided in the Intermediate I community. Since the
overall size of the black populations proved too small
for meaningful analysis of blacks alone, only whites
were considered in the analyses. The age distribution
of parents (Table 5.2.4) reflects the method of
population selection. Overall, about 50 percent of the
parents were in the 31- to 40-year-old age group.
There were no significant intercommunity differences
in age distribution within smoking- and sex-specific
categories. Smoking histories were obtained for
parents both to determine community smoking
patterns and to adjust for this variable in the
evaluation of respiratory symptoms. Cigarette smok-
ing patterns were quite similar for the three study
populations: between 40 and 45 percent of each
population were smokers, and 35 to 40 percent of
New York Studies
5-37
-------
these smoked at least 15 cigarette's daily. Although
the proportion of smokers was approximately equal
in women and men, men more often smoked one or
more packs of cigarettes per day. In this study, about
20 percent of the total population were exsmokers,
and these were classified separately for most analyses;
however, sample size considerations necessitated
pooling current smokers and exsmokers for the
analysis of migration effects. Since the Low pollution
community had a somewhat greater proportion of
exsmokers (25 percent compared with 20 and 16
percent), such pooling would probably tend to reduce
symptom rates and symptom severity for that com-
munity.
Table 5.2.4. DISTRIBUTION OF STUDY POPULA-
TION BY COMMUNITY, SMOKING STATUS, SEX,
AND AGE3
Category
Nonsmokers
Mothers
<40
>40
Fathers
<40
>40
Exsmokers
Mothers
<40
>40
Fathers
<40
>40
Smokers
Mothers
<40
>40
Fathers
^40
>40
Total
Community
Low
148
49
104
70
123
21
99
45
207
60
139
77
1142
Interme-
diate 1
,298
113
229
155
190
43
142
80
449
86
327
165
2277
Interme-
diate II
380
149
288
211
177
49
123
75
481
126
338
188
2585
Total
826
311
621
436
490
113
364
200
1137
272
804
430
6004
Comparison of socioeconomic patterns (Table
5.2.5) indicated that socioeconomic differences
did exist, with the Intermediate I community
ranking above the Low and Intermediate II
communities. The Intermediate I community was
the best educated, had the highest family
income, and was the least crowded in living
accommodations. The Intermediate II community
had the poorest educational background, showed
the lowest family income, and was the most
crowded in housing. The Low pollution com-
munity ranked intermediate for all three
variables.
Less than 1 percent of the population in
any of the study communities reported occupa-
tional exposure to irritating smoke, dust, or
fumes; therefore, no separate analysis of these
individuals was performed.
Chronic Respiratory Disease Evaluation
Chronic Bronchitis Prevalence
Prevalence rates for chronic bronchitis were
computed for the combined Intermediate and
Low exposure populations by age, sex, and
Table 5.2.5. COMMUNITY CHARACTERISTICS:
FAMILY INCOME, RACIAL COMPOSITION, AND
EDUCATIONAL STATUS OF FATHERS
(percent)
Characteristic
White3
High school
graduate'3
College graduate^
Family income
^$8000^
Community
Low
64.6
63.0
(246)
13.4
(246)
23.1
(238)
Interme-
diate 1
93.4
49.1
(281)
24.2
(281)
8.7
(276)
Interme-
diate II
88.0
55.3
(235)
7.2
(235)
33.9
(224)
aOnly whites were utilized because small black population
samples precluded meaningful analyses.
aBased on entire population returning questionnaires.
''Based on subsample of whites interviewed in their homes.
Numbers in parentheses represent size of subsample.
5-38
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
smoking habit to determine the overall effects of
these variables (Table 5.2.6). As expected, rates
in all age-sex categories were higher in smokers
than nonsmokers. Chronic bronchitis rates for
mothers who were exsmokers closely approxi-
mated rates for nonsmokers; fathers who were
exsmokers reported rates intermediate between
those reported for lifetime nonsmokers and those
found in current cigarette smokers. Bronchitis
rates were also higher in fathers than in mothers
in all of the age- and smoking*specific compari-
sons. No consistent upward trend in bronchitis
prevalence was noted with aging in this relatively
young population.
On the other hand, comparison of smoking- and
sex-specific rates between the Intermediate and Low
communities showed marked area differences for
smokers, exsmokers, and nonsmokers of both sexes
(Table 5.2.7). For nonsmokers, bronchitis rates in the
combined Intermediate communities were 2 to 4
times those in the Low pollution community. Al-
though this relative difference was less striking in
exsmokers and smokers, the absolute differences in
rates were generally similar. Much smaller, incon-
sistent differences were observed between the two
Intermediate communities. In each community,
mothers-whether smokers, exsmokers, or
nonsmokers—had lower rates than fathers. Tests for
statistical significance of these effects as well as the
Table5.2.6. AGE-, SMOKING-, AND SEX-SPECIFIC
CHRONIC BRONCHITIS PREVALENCE RATES*
(percent)
Category
Nonsmokers
Mothers
Fathers
Exsmokers
Mothers
Fathers
Smokers
Mothers
Fathers
Overall rate
Sample size
Age
<40
5.7
12.9
6.9
17.3
17.0
20.1
13.6
4242
>40
4.5
15.6
4.4
17.0
18.8
20.7
14.8
1762
effect of age are presented separately for mothers and
fathers in Table 5.2.8. Differences in rates between
the three communities and between the Low and
combined Intermediate communities were significant
for both mothers and fathers (p < 0.01). Differences
between the two Intermediate communities were not
significant. No significant age effect was seen in either
mothers or fathers; but as expected, smoking was
consistently highly significant. Among mothers, there
were no significant differences between lifetime
nonsmokers and exsmokers; however, significant
differences were found among fathers.
Mean Respiratory Symptom and Severity Scores
Respiratory symptom and severity scores were
also summarized for the three communities by
smoking status foi mothers and for fathers. The mean
symptom score is the average symptom rating in the
total subgroup designated; the mean severity score
represents the average symptom rating in only those
who reported symptoms.
Mean respiratory symptom scores (Table 5.2.9)
for residents of the Intermediate communities, like
chronic bronchitis prevalence rates, were consistently
higher than scores for the Low exposure community.
Table 5.2.7. SMOKING- AND SEX-SPECIFIC
CHRONIC BRONCHITIS PREVALENCE RATES
DISTRIBUTED BY COMMUNITY
(percent)
aCrude rates for symptom severity 6-7.
Category
Nonsmokers
Mothers
Fathers
Exsmokers
Mothers
Fathers
Smokers
Mothers
Fathers
Smok ing-adjusted
rates
Mothers
Fathers
Community
Low
2.0
4.6
3.8
13.9
13.9
13.9
4.7
11.5
Interme-
diate 1
7.5
18.0
9.0
18.0
19.8
21.3
11.6
18.4
Interme-
diate II
4.9
14.2
4.5
18,7
16.6
22.1
10.6
17.4
New York Studies
5-39
-------
Table 5.2.8. ANALYSES OF VARIANCE FOR HEALTH OBSERVATIONS,
CHRONIC BRONCHITIS PREVALENCE RATES
Factor
Community prevalence
rates
Low vs pooled
Intermediate
Intermediate 1 vs
Intermediate II
Smoking
Exsmokers vs
nonsmokers
Age
Model fit
Degrees
of
freedom
2
1
1
2
1
1
4
Mothers
X2
9.48
7.84
2.14
73.72
0.43
0.11
1.08
Probability(p)
0.0091
0.0054
0.14
0.0001
>0.5
>0.5
0.90
Fathers
X2
28.07
28.07
0.00
20.18
5.37
0.00
1.85
Probability(p)
0.0001
0.0001
>0.5
0.0002
0.02
>0.5
0.77
Moreover, smoking-adjusted differences between the
two Intermediate communities faithfully mirrored
the previously described differences in chronic bron-
Table 5.2.9. SMOKING-AND SEX-SPECIFIC MEAN
RESPIRATORY SYMPTOM SCORES DISTRIBUTED
BY COMMUNITY
(percent)
Community
Category
Nonsmokers
Mothers
Fathers
Exsmokers
Mothers
Fathers
Smokers
Mothers
Fathers
Smoking-adjusted
rates
Mothers
Fathers
Low
1.29
1.81
1.37
2.05
2.30
2.48
1.61
2.11
Interme-
diate 1
1.76
2.41
2.00
2.56
2.68
2.76
2.07
2.57
Interme-
diate II
1.61
2.35
1.72
2.51
2.53
2.76
2.02
2.52
chitis prevalence. In both sexes, scores for exsmokers
were intermediate between the higher scores of
current smokers and the lower scores of lifetime
nonsmokers. An analysis of variance was performed,
testing the effects of community pollution exposure,
of smoking, and of age upon mean symptom score
(Table 5.2.10). The results mimicked those of the
analysis testing the effects of these variables on
chronic bronchitis prevalence rates. Thus, mean
symptom scores proved a useful confirmatory analysis.
Sex- and smoking-specific mean severity scores
were then calculated (Table 5.2.11) and statistically
tested by the analysis of variance model previously
described (Table 5.2.12). These conditional severity
scores were consistently and significantly higher for
males, both smokers and nonsmokers, living in the
Intermediate communities. Severity scores were gen-
erally higher for male exsmokers than for males who
still smoked cigarettes. This result suggested that the
presence of respiratory symptoms may have been
related to cessation of smoking, with the more
symptomatic males being more likely to stop smok-
ing. Severity scores for female exsmokers were
intermediate between scores for lifetime nonsmokers
and for current cigarette smokers. This result suggest-
ed that factors other than the threat of current
respiratory symptoms induced mothers of young
children to stop smoking cigarettes.
5-40
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.2.10. ANALYSIS OF VARIANCE FOR HEALTH OBSERVATIONS,
MEAN RESPIRATORY SYMPTOM SCORES
Factor
Community
symptom scores
Low vs pooled
Intermediate
Intermediate I
vs Intermediate II
Smoking
Exsmokers vs
nonsmokers
Age
Model fit
Degrees
of
freedom
2
1
1
2
1
1
4
Mothers
x*
25.32
23.95
2.11
144.37
2.36
0.23
1.79
Probability(p)
0.0001
0.0001
0.14
0.0001
0.12
>0.5
0.78
Fathers
X2
25.27
25.73
0.02
29.14
2.52
0.06
5.33
Probability(p)
0.0001
0.0001
>0.5
0.0001
0.11
>0.5
0.255
Effects of Migration on Prevalence of Chronic
Bronchitis
Based upon questionnaire responses, the study
population was grouped into four categories: those
who had always lived in New York City; those who
moved to New York City from relatively clean,
smaller cities of less than 200,000 population; those
who moved into the Low exposure community from
a city larger than 200,000; and those who had always
lived in relatively clean areas. Respondents who
reported more than one move across the clean and
polluted areas were excluded from the analyses.
Because the sample of exsmokers was small, this
group was combined with current smokers in the data
analysis. The proportions of males and females were
similar in each residence-smoking category. Chronic
bronchitis prevalence rates for each residence-
smoking category (Table 5.2.13) were calculated and
found to be 1.6 to 5 times higher in smokers than in
lifetime nonsmokers. A novel and potentially impor-
tant finding was that migrants to polluted areas had
developed elevated bronchitis rates similar to those of
lifelong residents of the polluted communities. Mi-
grants to the Low exposure community exhibited
chronic bronchitis prevalence rates that were similar
to those of lifelong residents of relatively clean areas.
Unfortunately, sample size limitations precluded esti-
mating a time interval of exposure to which one
could attribute either beneficial or harmful effects.
Effect of Air Pollution Relative to Smoking
To assess the relative importance of ambient
air pollution and cigarette smoking, smoking- and
area-specific prevalence rates were compared
Table 5.2.11. MEAN SEVERITY SCORES IN PRES-
ENCE OF ANY CHRONIC RESPIRATORY DISEASE
SYMPTOM
Community
Category
Nonsmokers
Mothers
Fathers
Exsmokers
Mothers
Fathers
Smokers
Mothers
Fathers
Smoking-adjusted
rates
Mothers
Fathers
Low
4.40
4.19
4.58
4.77
4.67
4.56
4.44
4.57
1 nterme-
diate 1
4.45
4.88
4.67
4.79
4.86
4.73
4.72
4.85
Interme-
diate II
4.37
4.68
4.62
4.99
4.69
4.82
4.88
5.01
New York Studies
5-41
-------
Table 5.2.12. ANALYSIS OF VARIANCE FOR HEALTH OBSERVATIONS,
SEVERITY SCORES FOR THOSE WITH SYMPTOMS
Factor
Community
severity scores
Low vs pooled
Intermediate
Intermediate 1
vs Intermediate II
Smoking
Exsmokers vs
nonsmokers
Age
Model fit
Degrees
of
freedom
2
1
1
2
1
1
4
Mothers
X2
0.81
0.00
0.81
2.94
2.33
0.45
5.65
Probability(p)
>0.5
>0.5
>0.5
0.23
0.12
>0.5
0.23
Fathers
X2
6.62
6.62
0.00
4.64
4.64
0.40
0.26
Probability(p)
0.04
0.01
>0.5
0.10
0.03
>0.5
>0.5
Table 5.2.13. ASSOCIATION OF CHANGES IN CHRONIC BRONCHITIS PRE-
VALENCE WITH CHANGES IN POLLUTION EXPOSURE3
Chronic bronchitis prevalence rate, %
Exposure history
Always lived in New York City
Moved to New York City from
less populous area
Moved to Low exposure
area from large city
Always lived in relatively
clean area
Nonsmokers
Rate
9.2
10.8
2.0
2.6
Population
at risk
1000
139
100
192
Exsmokers and current
cigarette smokers
Rate
19.3
15.8
10.1
12.3
Population
at risk
2743
152
298
527
aBoth sexes combined.
5-42
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
(Table 5.2.14). Rates for mothers and fathers
were compared separately. In each case, the base
rate was the prevalence of chronic bronchitis
among lifetime nonsmokers living in the Low
community. The experience of the two Inter-
mediate communities was pooled. The excess
prevalence attributable to smoking was computed
by subtracting the chronic bronchitis prevalence
among nonsmoking fathers (or mothers) in the
Low community from the prevalence among
smoking fathers (or mothers) in the Low com-
munity. The excess prevalence attributable to
pollution was computed by subtracting the
chronic bronchitis prevalence among nonsmokers
in the Low community from the prevalence
among nonsmokers in the Intermediate com-
munities. The excess prevalences attributable to
smoking and to pollution among exsmokers were
computed in the same manner, using prevalence
rates of exsmokers instead of current smokers.
One measure of the importance of ambient
air pollution relative to smoking may be ex-
pressed by a ratio whose numerator is the excess
prevalence attributable to pollution and whose
denominator is the excess prevalence attributable
to present or past smoking. When such ratios
were computed for females, air pollution was
found to have 34 percent the effect of current
cigarette smoking and 222 percent the effect of
former cigarette smoking. For males, ambient air
pollution was computed to have 122 percent the
effect of either current or former cigarette smok-
ing. Overall, air pollution seemed almost (78
percent) as important a factor as current cigarette
smoking and more important (172 percent) than
previous cigarette smoking.
If the hazards of self-pollution by cigarette
smoking and ambient air pollution were additive,
then the sum of the excess prevalence for
smokers living in the Low community and the
excess prevalence for nonsmokers from the Inter-
mediate communities should equal the excess
prevalence for smokers living in Intermediate
communities. The same procedure may be applied
for exsmokers. When these procedures were
applied, the combination of current cigarette
smoking and air pollution was found to explain
83 to 101 percent of the expected excess
bronchitis under a simple additive model. The
comparable values when the additive model
involved exsmokers were 67 to 83 percent.
Together, these models suggested that the effects
Table 5.2.14. RELATIVE IMPORTANCE OF CIGARETTE SMOKING AND AMBIENT
AIR QUALITY AS DETERMINED BY COMPARISON OF EXCESS PREVALENCE OF
CHRONIC BRONCHITIS3
Current cigarette
smoking status
Female
Lifetime nonsmoker
Exsmoker
Smoker
Male
Lifetime nonsmoker
Exsmoker
Smoker
Community
Low
Intermediate
Low
Intermediate
Low
Intermediate
Low
Intermediate
Low
Intermediate
Low
Intermediate
Excess prevalence
of chronic
bronchitis
0.00(2.0)
4.0
1.8
4.8
11.9
16.1
0.00(4.6)
11.3
9.3
13.7
9.3
17.1
Simple additive
model combining adverse
effects of smoking
and pollution
Expected
—
—
—
5.8
—
15.9
—
—
—
20.6
—
20.6
Observed/
expected
—
—
—
0.83
—
1.01
—
—
—
0.67
—
0.83
aBase rates in parentheses. Excess prevalence = smoking-and area-specific prevalence rate minus base
rate.
New York Studies
5-43
-------
of ambient air pollution and smoking were
roughly additive.
DISCUSSION
In this study, chronic respiratory disease
frequency, whether defined by symptom score or
as a symptom profile compatible with chronic
bronchitis, was significantly higher in the two
Intermediate communities. People living in the
Low pollution community had a markedly lower
prevalence of respiratory symptoms than people
living in the Intermediate communities. This
difference was observed for people over as well
as under 40 years of age; for both sexes; and for
smokers, exsmokers, and nonsmokers.
Two important additional observations strength-
en the association of ambient air pollution with
chronic respiratory disease symptoms: First, migrants
moving to the Low community from polluted com-
munities had lower rates than lifetime residents of the
Intermediate communities, while migrants to the
Intermediate communities from clean communities
were found to assume the higher rates of the
Intermediate communities. Second, excess prevalence
models indicated that the hazards of air pollution
exposure and continued or past self-pollution by
cigarette smoking were nearly additive. These models
also indicated that under the circumstances of this
particular study, the contribution of air pollution to
observed prevalence rates, relative to smoking, was
substantial. The figures describing this contribution
are not meant to be exact measurements; rather, they
are intended to point out that pollution may exert a
stronger effect than had been previously suggested.
The observed effects of covariates in the present
study were consistent with previous findings on the
determinants of chronic bronchitis in that smokers
and males were found to have substantially higher
rates than nonsmokers and females. Failure to define
a linear age gradient may be explained in part by the
relatively young age of the study population.
Before assessing the full implications of the
present report, critical consideration must be
given to the impact of four factors: socio-
economic differences, ethnic exclusion, nonre-
spondent morbidity, and differences between
recent and remote air pollution exposures.
First, if one considers that lower socio-
economic status might be associated with an
increase in chronic bronchitis prevalence rates,
then the documented intercommunity socioeco-
nomic differences would dampen the observed
morbidity excess found when the more prosper-
ous Intermediate I community was compared to
the less advantaged Low pollution area. Con-
versely, the same socioeconomic effect would
accentuate observed morbidity differences between
the less prosperous Intermediate II community
and the relatively more prosperous Low pollution
area. For this particular study, the opposite
argument, that higher socioeconomic status might
be associated with an increase in reported pre-
valence rates, might be applied in view of the
fact that the more prosperous Intermediate I
community reported the highest rates. In either
case, socioeconomic bias in this study would be
minimal because of the intermediate socioeco-
nomic status of the Low pollution community. It
is extremely unlikely that any residual soc;o-
economic bias could explain the substanua.
intercommunity prevalence differences that were
observed.
Second, ethnic exclusions facilitated the
analyses, but inclusion of blacks would not have
altered the findings of the study. Blacks and
whites exhibited the same symptom gradient
between Low and Intermediate communities, and
there were no significant differences attributable
to racial differences alone. The only anomalous
finding was that black male nonsmokers had
somewhat higher chronic bronchitis prevalence
and mean symptom scores than black male smokers.
This anomaly could well be explained by the rela-
tively small sample size of blacks.
Third, morbidity in nonrespondents might
dilute the observed differences in morbidity. For
this to occur, area differences in morbidity re-
porting among nonrespondents would have to be
opposite to the demonstrated trend in morbidity.
To equalize the observed area differences, chronic
bronchitis symptoms among the nonrespondents
in the Low area would have to average over 4
times the rates reported for respondents. In
addition, the assumption depends upon none of
the nonrespondents in the Intermediate commu-
nities having chronic bronchitis. Such a set of
events would be indeed extraordinary.
Fourth, estimating past and present ambient and
indoor air pollution exposures is a difficult problem
that has not been satisfactorily resolved. Nevertheless,
particulate and sulfur dioxide pollution in past
5-44
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
decades was estimated. More precise measurements
were made for current exposures, and the apparent
beneficial and detrimental influences of migration
and nonmigration were assessed. It was not possible
to differentiate or quantify the importance of re-
peated short-term peak exposures relative to longer
lasting low-level pollutant exposures. Personal pollu-
tion by cigarette smoking has been duly considered in
this report. The importance of indoor air pollutants,
such as nitrogen oxides from gas appliances and
smoke from other people's cigarettes, has yet to be
determined.
There is reasonable evidence that residents of the
Intermediate communities were exposed to very high
pollution levels during their childhood and/or young
adult years. Such early exposures may have con-
tributed to the prevalence of chronic respiratory
disease in later life. Symptomatic improvement,
however, might eventually be expected if observa-
tions based upon the experience of migrants and
exsmokers are valid.
Even if childhood exposure to air pollution was
not a factor in the development of chronic bronchitis,
present pollution levels may still cause such symp-
toms to persist and perhaps even to worsen. More
optimistically, the tide of excess morbidity may
already have begun to ebb, and this decrease could
accelerate as air quality is improved. On the other
hand, one could interpret the present study as
suggesting that as yet unspecified pollutants capable
of inducing or aggravating chronic respiratory disease
are not well controlled. In any event, a single
cross-sectional study cannot provide the definitive
explanation. A mosaic of information involving multi-
ple epidemiologic and laboratory studies is clearly
indicated.
Based upon the available data, we believe it is
prudent to consider that long-term exposures to
annual average sulfur dioxide levels of about 144 to
404 /ig/m^ coupled with annual average suspended
particulate levels of about 80 to 173 jug/m^ and
suspended sulfate levels of about 9 to 24 /xg/m^, in
the presence of other fossil fuel combustion products,
substantially increases the risk of chronic bronchitis.
It is also possible that lower pollution levels may well
aggravate or support the persistence of chronic
respiratory disease symptoms. Annual sulfur dioxide
levels of 50 to 60 jug/m-^ accompanied by annual
average suspended sulfate levels of about 14 jug/m^
and annual arithmetic mean total suspended par-
ticulate levels of about 60 to 105 jug/m^ could be
associated with such effects. It is thus quite possible
that the present National Primary Ambient Air
Quality Standards for sulfur dioxide and suspended
particulates contain either a small safety factor or no
safety factor at all. Finally, increasing attention
should be devoted to elucidating the health implica-
tions of suspended particulate sulfates, which are
evidently decreasing more slowly than sulfur dioxide
or suspended particulates.
SUMMARY
Chronic respiratory disease morbidity was
assessed in 6004 parents of elementary school
children residing in either a clean Long Island
community or one of two more polluted com-
munities in New York City. Long-term exposures
(up to 20 years) to sulfur dioxide levels of an
estimated 144 to 404 jUg/m-' and suspended sul-
fate levels of an estimated 9 to 24 ng/m-* were
associated with marked increases in chronic
bronchitis prevalence among smokers, exsmokers,
and nonsmokers of both sexes age 20 to 50.
Shorter term, more recent exposures to lower
pollutant levels may also induce, aggravate, or
sustain chronic respiratory disease symptoms.
Such short-term exposures might involve annual
average sulfur dioxide levels of 50 to 60 jug/m-'
accompanied by annual average suspended sulfate
levels of about 14 jug/m^ and total suspended
particulate levels of about 60 to 105 jug/rrP. In the
present study, the observed differences between
communities could not be accounted for by age,
occupational exposures, smoking habits, or ethnic
factors. It is unlikely that nonrespondent bias or
intercommunity socioeconomic differences could
have explained the observed intercommunity differ-
ences in prevalence rates. An analysis of migration
patterns indicated that improvement of air quality
was associated with striking decreases in chronic
bronchitis prevalence and that moving into a polluted
city from a relatively clean area was associated with
an increased risk of chronic bronchitis. Air pollution
and self-pollution by cigarette smoking appeared to
be additive hazards. Improvements in air quality may
be expected to benefit the population by reducing
chronic respiratory disease symptom frequency and
severity.
REFERENCES FOR SECTION 5.2
1. Firket, J. Fog along the Meuse Valley. Trans.
Faraday Soc. 52:1192-1197, 1936.
New York Studies
5-45
-------
2. Firket, J. The Cause of the Symptoms
Found in the Meuse Valley during the Fog
of December 1930. Bull. Roy. Acad. Med.
Belgium. 11:683-739, 1931.
3. Shrenk, H.H., H. Heimann, G.O. Clayton,
W.M. Gafafer, and H. Wexler. Air Pollution
in Donora, Pennsylvania; Epidemiology of the
Unusual Smog Episode of October 1948.
Federal Security Agency, Division of In-
dustrial Hygiene, Public Health Service, U.S.
Department of Health, Education, and Wel-
fare. Washington, D.C. Public Health Bulletin
306. 1949.
4. Gore, A.T. and C.W. Shaddick. Atmospheric
Pollution and Mortality in the County of
London. Brit. J. Prev. Soc. Med. 72:104-113,
1958.
York City. J. Amer. Med. Assoc.
782:161-164, 1962.
12. McCarroll, J.R., E.G. Cassell, E.W. Walter,
J.D. Mountain, J.R. Diamond, and I.R.
Mountain. Health and the Urban Environ-
ment; V. Air Pollution and Illness in a
Normal Urban Population. Arch. Environ.
Health. 74:178-184, 1967.
13. Lawther, P.J. Climate, Air Pollution, and
Chronic Bronchitis. Proc. Roy. Soc. Med.
57:262-264, 1958.
14. Carnow, B.W., M.H. Lepper, R.B. Shekelle,
and J. Stamler. The Chicago Air Pollution
Study: Acute Illness and SC>2 Levels in
Patients with Chronic Bronchopulmonary
Disease. Arch. Environ. Health. 75:768-776,
1969.
5. Burgess, S.G. and C.W. Shaddick. Bronchitis
and Air Pollution. Roy. Soc. Health J.
79:10-24, 1959.
15. Lawther, P.J., R.E. Waller, and M. Hender-
son. Air Pollution and Exacerbations of
Bronchitis. Thorax. 25:525-539, 1970.
6. Scott, J.A. The London Fog of December
1962. Med. Officer. 709:250-252, 1963.
7. Greenburg, L., M.B. Jacobs, BJM. Drolette, F.
Field, and M.M. Braverman. Report of an
Air Pollution Incident in New York City,
November 1953. Public Health Reports.
77:7-16, 1962.
8. McCarroll, J. and W. Bradley. Excess Mor-
tality as an Indicator of Health Effects of
Air Pollution. Amer. J. Public Health.
56:1933-1942, 1966.
9. Glasser, M., L. Greenburg, and F. Field.
Mortality and Morbidity during a Period of
High Levels of Air Pollution, New York,
November 23-25, 1966. Arch. Environ.
Health. 75:684-694, 1967.
10. Logan, W.P.D. Mortality in the London Fog
Incident 1952. Lancet. 264: 336-338, 1953.
11. Greenberg, L., F. Field, J.I. Reed, and C.L.
Erhardt. Air Pollution and Morbidity in New
16. Holland, W.W. and R.W. Stone. Respiratory
Disorders in United States East Coast Tele-
phone Men. Amer. J. Epidemiol. 52:92-101,
1965.
17. Holland, W.W., D.D. Reid, R. Seltser, and
R.W. Stone. Respiratory Disease in England
and the United States; Studies of Compara-
tive Prevalence. Arch. Environ. Health.
70:338-345, 1965.
18. Petrilli, R.L., G. Agnese,. and S. Kanitz.
Epidemiology Studies of Air Pollution Effects
in Genoa, Italy. Arch. Environ. Health.
72:733-740, 1966.
19. Toyama, T. Air Pollution and Its Health
Effects in Japan. Arch. Environ. Health.
5:153-173, 1964.
20. Zeidberg, L.D., R.A. Prindle, and E. Landau.
The Nashville Air Pollution Study; HI. Mor-
bidity in Relation to Air Pollution. Amer. J.
Public Health. 54:85-97, 1964.
5-46
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
21. Bell, A. The Effects on the Health of
Residents of East Port Kembla, Part II. In:
Air Pollution by Metallurgical Industries (Vol.
2). Sydney, Australia, Division of Occupa-
tional Health, New South Wales Department
of Public Health, 1962. p. 1-144.
22. Questionnaires Used in the CHESS Studies.
In: Health Consequences of Sulfur Oxides: A
Report from CHESS, 1970-1971. U.S. Envi-
ronmental Protection Agency. Research Triangle
Park, N.C. Publication No. EPA-650/1-74-004.
1974.
23. English, T.D., W.B. Steen, R.G. Ireson, P.B.
Ramsey, R.M. Burton, and L.T. Heiderscheit.
Human Exposure to Air Pollutants in
Selected New York Metropolitan Commu-
nities, 1944-1971. In: Health Consequences
of Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Publi-
cation No. EPA-650/1-74-004. 1974.
24. Grizzle, J.E., C.F. Starmer, and G.G. Koch.
Analysis of Categorical Data by Linear
Models. Biometrics. 25(3):489-504, September
1969.
New York Studies
547
-------
5.3 PROSPECTIVE SURVEYS OF ACUTE RESPIRATORY
DISEASE IN VOLUNTEER FAMILIES:
1970-1971 NEW YORK STUDIES
Gory J. Love, Sc.D., Arlan A. Cohen, M.D.,
John F. Finklea, M.D., Dr. P.H., Jean G. French,
Dr. P.H., Gene R. Lowrimore, Ph.D., William C. Nelson, Ph.D.,
and Peggy B. Ramsey, B.S.
5-49
-------
INTRODUCTION
Acute respiratory infections consistently rank
as the leading cause of acute illness in the
United States. During the influenza year July
1968 to June 1969, the average American
endured 1.2 acute respiratory illnesses, which was
about twice the expected rate for a noninfluenza
year.1 The direct economic costs of morbidity
from acute respiratory infections are difficult to
determine precisely but are estimated to be well
over $5 billion per year.2 Children suffer the
highest incidence of acute respiratory disease,
about twice that of adults, and there is increas-
ing evidence that a history of repeated acute
lower respiratory illnesses in childhood may
contribute to the development of chronic
bronchitis in later life.3
Accordingly, one way to reduce the frequency of
chronic respiratory disease in adults may be to iden-
tify and control factors that might induce or aggra-
vate acute respiratory illnesses in children. Exposure
to high levels of atmospheric pollution, as indexed by
elevated levels of sulfur dioxide and total sus-
pended particulates, may be an important
determinant of acute respiratory illness. British
investigators evaluated the effects of long-term
exposures among more than 3000 children living
in areas representing different levels of air pollu-
tion.4 They found an excess number of lower
respiratory tract illnesses in children from the
most polluted areas and a significant trend
towards more hospitalization for lower respiratory
tract infections in the same group of children. A
second English study confirmed the association of
lower respiratory tract illnesses with elevated
pollution levels and implicated air pollution as an
important determinant of upper respiratory
disorders.5 In another study, Japanese students
living in a highly polluted area reported a higher
frequency of respiratory symptoms than did children
in a less polluted area.^
The laboratory evidence accumulated thus far on
the toxicological effect of sulfur dioxide indicates
that the pure gas is a mild respiratory irritant that is
largely absorbed in the upper respiratory tract. The
basic physiologic response appears to be a mild degree
of broncho-constriction that causes a measurable
increase in flow resistance. However, sulfur dioxide in
combination with suspended particulates was found
to be more irritating than sulfur dioxide alone.?>8
Few clinical or toxicological studies involving sulfur
dioxide, suspended particulates, or suspended sulfates
in combination with various metals have been re-
ported.9
The prospective epidemiologic study reported
here considered the acute respiratory illness
experience of families living in three areas of
metropolitan New York. For the period late
September 1970 to early May 1971, panels of
volunteer families residing in two more urban
communities were compared with families living
in a less polluted suburban fringe community. An
important advantage of the New York study
setting was the availability of historic air moni-
toring data from which estimated past exposures
could be reconstructed. The recent rapid decline
in air pollutant levels in the area also permitted
the investigators to search for early health bene-
fits from air pollution control.
The study tested two hypotheses: first, that
the frequency and/or severity of acute respiratory
illness would be higher in the more polluted
communities; and second, if increases were ob-
served, that at least a portion of the increased
vulnerability could be attributed to air pollution
exposures of a defined duration. In addition, an
attempt was made to quantify the effects of
indoor air pollutants and to measure the
importance of residential mobility, socioeconomic
status, and the history of chronic respiratory
disease in parents as determinants of acute
respiratory illness in children.
METHODS
Setting and Study Population
Three areas that were similar with regard to
socioeconomic factors, ethnic composition, and
meteorological conditions but differed from each
other with respect to atmospheric pollution were
selected for study. After considering long-term air
pollution exposure trends, Riverhead, Long Island,
was chosen as a Low exposure community, the
Howard Beach section of Queens as an Inter-
mediate exposure community, and the West-
chester section of the Bronx as a High exposure
community. As a result of recent improvements
in air quality, the Queens and Bronx were found
to have similar current pollutant concentrations,
often below National Primary Air Quality
Standards. Hence, the latter two communities
were redesignated Intermediate I and Intermediate
II, respectively.
5-50
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Initial contact with the population residing in
a community was made with the cooperation of
local school boards, who permitted question-
naires10 to be distributed to parents by their
children who attended specific schools. Informa-
tion obtained from the questionnaire identified
the demographic characteristics of the study
population. Comparison of the population return-
ing the completed questionnaire with a sample of
the approximately 10 percent who did not indi-
cated that the two groups did not differ
demographically.
Panels of subjects selected for study were
restricted to families who resided within 1 to 1.5
miles of the air monitoring station, had a func-
tioning telephone, had at least one school child
12 years of age or younger, and had lived at
their present address for at least 1 year. When-
ever possible, selections were based on the
presence of preschool siblings. The study was
limited to white families because the paucity of
black families living in the more polluted com-
munities precluded ethnic matching. A home visit
was made to each prospective study family to
gain their willingness to participate and to collect
additional information regarding age, occupational
pollution, smoking habits, use of gas in the
home, and symptoms of chronic respiratory
disease. These covariates, with the exception of
occupational exposure, were all considered in
subsequent analyses. Since the survey indicated
that fewer than 1 percent of the fathers worked
in a polluted environment, this factor was con-
sidered to be insignificant.
Illness Monitoring
Families were called once every 2 weeks and
questioned about the presence of illness, fever,
respiratory symptoms, restricted activity, including
absences from school or from work or confine-
ment to bed, and whether contact had been
made with a physician either by phone or office
visit. If a family member had visited a physician,
information was elicited concerning the physi-
cian's diagnosis. A standardized questionnaire10
was used to obtain the information and the tele-
phone interviews were conducted by trained
interviewers. Whenever possible, information on
symptoms was obtained from the mother or
female guardian, and determination of the
presence or absence of illness was left to her
discretion. If the mother was not available, the
rnformation was obtained from whomever was
considered to be the most responsible resident in
the home.
The upper respiratory illness classification was
given to those reporting any one of the following:
cough (dry, nonproductive), head cold, sore throat,
sinus or postnasal drip, or runny nose. Lower
respiratory illnesses were characterized by a chest
cold, a deep productive chest cough, wheezing, croup,
bronchitis, or pneumonia. Whenever both upper and
lower tract symptoms were present, the diagnosis of
acute lower respiratory illness was arbitrarily
assigned. As a check of reliability of reporting,
randomly selected households were reinterviewed
several days after the routine calls. This reinterview
was assigned to experienced interviewers who had not
previously contacted the families that were being
checked.
Monitoring Air Pollution Exposures
Measurements of pollutant levels were ob-
tained through the cooperation of the New York
City Department of Air Resources (NYC-DAR)
and the Suffolk County (New York) Department
of Health and from CHESS air monitoring sta-
tions located within each community studied.
Stations were located on roof tops 30 to 45 feet
above ground level. At each station, continuous
measurements were obtained for sulfur dioxide
(modified West-Gaeke method), total suspended
particulates, suspended sulfates, and suspended
nitrates. Nitrogen dioxide measurements were
made by the Jacobs-Hochheiser method, but
because it is now known that this method does
not provide accurate data in some instances, the
information was not used in the analysis. Newer
knowledge related to the inaccuracy of the
Jacobs-Hochheiser method indicates that the
method overestimates actual levels lower than 90
/ig/m . Since levels measured during the period
of this study were consistently well below 90
/ig/m , it is believed that nitrogen dioxide did
not contribute significantly to impairments in
health associated with exposure to air pollution.
Analyses of air pollution measurements were
made on the basis of 24-hour samples. In addi-
tion, monthly dustfali samples were obtained.
The air monitoring procedures were detailed else-
where.11 Historical data were obtained from
aerometric stations maintained by NYC-DAR in
the Intermediate communities. These stations
monitored sulfur dioxide (peroxide titration
method), total suspended particulates, monthly
New York Studies
5-51
-------
dustfall, and soiling index. From these data,
estimates of the annual average air pollutant
exposures for the decade prioi to the study were
reconstructed.11
Statistical Analysis
Provisions were made to partition the total
study period into four segments so that separate
analyses might be focused on any period of
epidemic illness. This procedure also minimized the
adverse effects of family withdrawals or tempo-
rary nonparticipation, since the experience of a
family was not considered in the analysis if
contact could not be made during any portion
of a particular 8-week study segment. Attack
rates for upper, lower, and total acute respiratory
illness were computed separately for four family
segments: fathers, mothers, elementary school
children, and preschool children. The experiences
of older children or other adults living in the
home were excluded from the analyses. The
population reporting repeated bouts of acute
respiratory illness was defined by considering the
illness experience of each volunteer. Intercommu-
nity differences in the frequency of acute respira-
tory illness were ascertained by comparing two
measures: (1) attack rates for acute respiratory
illness and (2) the proportion of the study
population reporting excess respiratory illness.
Repeated excess acute respiratory illness during
the study was defined as more than two episodes
of either upper or lower tract illness.
Intercommunity differences in illness severity
were ascertained by comparison of severity
scores, which ranged from Level I through Level
V. Briefly, Level I denoted" illness without re-
stricted activity, Level II included restricted
activity but not fever, Level .III included re-
stricted activity and fever, and Level IV included
all of the preceding plus a physician visit. Level
V applied only to children and .was defined as
restricted activity and fever accompanied by otitis
media (middle ear infection) diagnosed by a physi-
cian. Level V was chosen to reflect a common,
presumably bacterial, complication of acute upper
respiratory illness. Levels II through V were further
characterized by the mean number of restricted
activity days incurred.
The effects of probable determinants of
illness were assessed by comparison of specified
or adjusted attack rates and by statistical tests
using a general linear model for categorical
data.12
RESULTS
Air Pollution Exposures
Respiratory defense mechanisms might be
impaired by either short-term pollutant exposures,
long-term exposures, or both. The present study
considered both types of exposure. Long-term
exposures were indexed by annual average pollut-
ant levels during the study and by estimates for
the prior decade, which coincided with early
growth and development years of the childhood
population. Short-term air pollutant exposures
were indexed by the median, ninetieth percentile,
and maximum levels during the 32-week study
period.
As expected, the Low exposure community
had lower pollutant levels than the two Inter-
mediate exposure areas during the study, and
evidence indicated that these conditions had
existed during the previous decade as well (Table
5.3.1). Suspended sulfate levels, however, were
higher than expected in the Low exposure area
even though no large point emission sources of
sulfur oxides were nearby. The two Intermediate
exposure areas generally had higher pollutant levels
than the Low exposure area and were similar to each
other. During the study, the Intermediate commu-
nities had annual average sulfur dioxide levels that
approximated the annual average specified by the
National Secondary Air Quality Standard (60 jug/m-'
or 0.02 ppm). Total suspended particulate levels
approximated or just exceeded the annual geometric
average level specified by the National Primary Air
Quality Standard (75 jUg/m^). Suspended particulate
sulfafe- levels (13 to 14 ng/m^) were elevated in both
of the • Intermediate communities, compared to the
Low exposure community. Suspended particulate
nitrates were similar in the two Intermediate com-
munities (^4 £(g/m->) and were also higher than in the
Low community. The Intermediate I community was
adjacent to a large internationalairport and may have
been polluted with higher levels-of hydrocarbons or
other materials associated directly with this facility.
Because the full potential impact of these pollutants
was not recognized at the time the study was
initialed, the siting of the study community was
considered to be acceptable, and aerometric measure-
ments of hydrocarbons were not included in the
study design.
5-52
HEALTH CONSEQUENCES OF SULFUR
-------
Air pollution levels during the decade prior to
the study were indicated to be much higher than
present values in both of the Intermediate exposure
communities. Sulfur dioxide concentrations were
probably 3 to 5 times the current Primary Air
Quality Standard, and suspended particulate con-
centrations were approximately twice the current
standard. Estimated suspended sulfate levels (10
to 20 p.g/m^) were about 1.5 times higher in the
preceding decade than in 1971. Thus, most
families participating in the study had, by and
large, probably been exposed to much higher
pollution levels during the preceding decade than
during the study period. This improvement in air
quality resulted from stringent control efforts
carried out by private enterprise and govern-
mental agencies.
Median, ninetieth percentile, and maximum
levels, indicative of short-term pollutant exposure
Table 5.3.1. ESTIMATED ANNUAL AVERAGE
LEVELS OF SELECTED AIR POLLUTANTS IN NEW
YORK METROPOLITAN STUDY COMMUNITIES
Pollutant and
community
Sulfur dioxide
Low
Intermediate 1
Intermediate II
Total suspended
particulate
Low
Intermediate 1
Intermediate II
Suspended sulfate
Low
Intermediate 1
Intermediate II
Suspended nitrate
Low
Intermediate 1
Intermediate II
Years prior to study3
6 to 10
(1961-65)
<30
431
396
<40
201
182
MO
20
18
NAC
NA
NA
1 to 5
(1966-70)
<30
256
321
<40
97
123
MO
10
15
NA
NA
NA
During
1971
23
51 to 63b
51 to 58b
34
63 to 84b
86to104b
10.2
13.2
14.3
1.9
3.5
4.1
aStudy began i n September 1970 and ended i n May 1971.
The upper end of the range represents analysis of samples made
by the New York City Department of Air Resources. The low
figure represents analysis of samples collected during the same
period and at the same site then mailed to the Research Triangle
Park laboratory in North Carolina.
cNA-not available.
during the 32-week study period, are given in
Table 5.3.2. Similar values for the summer
following the survey are included for comparison.
The maximum allowable short-term (24-hour) air
quality limits for sulfur dioxide and suspended
particulates were not exceeded during the study.
Peak exposures to sulfur dioxide (143 to 349
jug/lip) and suspended nitrates (5 to 20 jug/m^)
were higher in the Intermediate exposure areas
than in the Low area. Surprisingly, peak sus-
pended sulfate exposures in the Low exposure
area (32 to 51 jug/m^) were similar to those in
the Intermediate communities (28 to 42 jug/irr).
Intercommunity patterns for median 24-hour
exposure levels paralleled annual average expo-
sures, and ninetieth percentile values exhibited a
community pattern similar to that of peak
exposures. Seasonal factors were operative, with
overall higher pollutant exposures occurring
during fall and winter. Total suspended partic-
ulates showed less of a seasonal trend than the
other measured pollutants.
Characteristics of the Study Population
Intercommunity differences in socioeconomic
status, indoor air pollution, and mobility were
assessed (Table 5.3.3). The Low exposure com-
munity ranked between the two Intermediate
communities when either of two indices of socio-
economic status was considered. Educational
attainment of fathers was lower and relative
crowding was more frequent in the Intermediate
II community and least frequent in the Inter-
mediate I community. Diminished social status
was accompanied by a moderate increase in the
proportion of parents who smoked cigarettes and
a decrease in residential mobility. Almost all
families in the two Intermediate communities
utilized gas for cooking as compared to only
three-fifths of families in the Low exposure
community. Since the two Intermediate exposure
areas differed only slightly in their pollution
exposures, the respiratory disease experience of
families from these two communities could
reasonably be pooled. The pooled Intermediate
grouping more closely matched the Low exposure
area in socioeconomic status, smoking habits, and
mobility. Smoking intensity, among those who
smoked differed little between the communities.
Over 3000 subjects participated in the study;
these subjects are distributed by family segment
and study period in Table 5.3.4. The number of
New York Studies
5-53
-------
Table 5.3.2. DAILY AIR POLLUTANT EXPOSURE PATTERNS
IN THREE NEW YORK COMMUNITIES DISTRIBUTED BY SEASON
Pollutant and
community
Sulfur dioxide
Low
Intermediate I
Intermediate II
Total suspended parti culates
Low
Daily pollutant levels, ug/mj
Median
Fall i Winter
1970 1971
13 18
56 34
64 35
32 29
Intermediate I 56 55
Intermediate II
57
75
Suspended sul fates i •
Low 11
9
Intermediate I 14 12
Intermediate II
Suspended nitrates
Low
Intermediate I
Intermediate II
14
1
2
1
14
1
1
2
Spring Summer
1971 1971
15 8
48
29
19 21
29 34
57
60
77 | 78
90th percent! le Maximum
Fall
1970
Winter
1971
Spring
1971
45 76 '45
120 i 119 108
178 162
59 56
93 90
118 121
104
53
102
128
i
6 9 23 18 16
9
10
1
2
3
12 22 20
14 j 22
2
4
4
3
3
3
23
17
22
1
2
3
4
2
5
9
Summer
1971
Fall
1970
21 131
71
58
58
231
268
97
101 170
118
216
22 ' 48
27 28
32
4
8
9
36
3
6
5
Winter
1971
125
287
349
134
156
188
51
39
35
6
6
6
Spring
1971
114
143
219
91
Summer
1971
52
107
123
104
193 152
217 211
32
36
42
42
37
53
4
11
20
6
15
14
Table 5.3.3. PANEL CHARACTERISTICS
Characteristic
Fathers with high
school diploma, %
Families with ^1.0
person per room, %
Families using
domestic gas, %
Current cigarette
smokers, %
Mothers
Fathers
Families moved
during previous
5 years, %
Community
Low
73
85
60
37
45
36
Intermediate
I
76
92
95
37
47
50
II
60
72
98
47
53
38
Pooled
69
83
96
41
50
45
families varied in each area, with the Inter-
mediate I community having the largest number
(241) and the Intermediate II community having
the smallest number (187). The number of
families considered in the analysis is somewhat
lower due to attrition during the study and the
restriction that excluded individuals unless they
reported for every period within any study
segment. Attrition was somewhat higher in the
Low exposure community than in the two
Intermediate communities.
Acute Respiratory Disease Evaluation
Seasonal Variation
The study, which began in late September
1970 and terminated in early May 1971, encom-
passed an entire respiratory illness season. The
overall acute respiratory illness attack rates for
each of the study communities were plotted to
detect any epidemic wavelets, but none occurred
5-54
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.3.4. STUDY PARTICIPANTS INCLUDED IN ANALYSES DISTRIBUTED
BY SEASON AND COMMUNITY
Family segment
and community
Fathers
Low
Intermediate I
Intermediate II
Pooled Intermediate
Mothers
Low
Intermediate I
Intermediate II
Pooled Intermediate
School children
Low
Intermediate I
Intermediate II
Pooled Intermediate
Preschool children
Low
Intermediate I
Intermediate II
Pooled Intermediate
Total
Low
Intermediate I
Intermediate II
Pooled Intermediate
Period of study
Weeks
0-8
191
231
140
371
211
238
147
385
564
536
331
867
170
148 •
95
243
1136
1153
713
1866
Weeks
9-16
185
232
181
413
210
241
187
428
561
525
401
926
135
149
121
270
1091
1147
890
2037
Weeks
17-24
184
227
183
410
206
235
190
425
551
516
423
939
133
144
125
269
1074
1122
921
2043
Weeks
25-32
163
Participated
during all
periods
132
218 179
175
393
186
224
184
408
492
501
403
904
122
145
122
267
963
1088
884
1972
157
336
146
183
161
344
388
406
349
755
84
119
97
216
750
887
764
1651
New York Studies
5-55
-------
(Figure 5.3.1). Rates were high throughout fall
and winter before declining in the spring. The
attack rates for the two Intermediate areas were
generally above that of the Low pollution area.
Higher illness rates in the Intermediate I area
were reported, despite the fact that current
pollution levels were generally somewhat lower
than those reported in the Intermediate II area.
The discrepancy possibly was related to the
additional pollution associated with the nearby
airport, but this could not be substantiated with
the data collected.
To assess the effect of pollution on inter-
seasonal variations in respiratory disease, the time
span of study was divided into roughly equal
seasonal segments:
1. Season I, September 27 to November 21,
1970.
2. Season II, November 22, 1970, to
January 16, 1971.
3. Season III, January 17 to March 13,
1971.
4. Season IV, March 14 to May 8, 1971.
Significant intercommunity differences in illness fre-
quency were found in every season (Table 5.3.5).
Illnesses involving, the lower respiratory tract were
significantly increased when residentially stable
families (i.e., families who had not moved within the
past 5 years) living in the two Intermediate commu-
nities were compared to families living in the Low
Table 5.3.5. STATISTICAL ANALYSIS OF
EFFECTS OF AIR POLLUTION ON
FREQUENCY OF ILLNESS BY
SEASON3
Season
1
II
III
IV
Probability (p)b
Upper
tract
illness
0.01
NS
NS
NS
Lower
tract
illness
0.005
0.001
0.01
0.025
All
respiratory
illness
0.001
0.001
0.10
NS
aAnalysis included all members of residen-
tially stable families, i.e., those who had
not moved within the past 5 years.
bNS-not significant, p > 0.10.
£2
cc
o
o±
UJ
o.
o
-------
exposure community. Acute upper respiratory
illnesses were significantly increased in the Inter-
mediate communities only during the initial
seasonal segment, and the greatest increases in
overall acute respiratory illness occurred during
the first two seasonal segments.
Illness Frequency
For each community, overall acute respira-
tory disease attack rates were computed sepa-
rately for fathers, mothers, elementary school
children, and preschool children. To facilitate
intercommunity comparisons, relative risk ratios
were calculated, utilizing the experience of the
Low exposure community as a base. The relative
risk for any respiratory illness category involving
family members from the Low exposure commu-
nity was set at 1.00, and the relative risks for
other communities were calculated by using the
attack rate for the Low exposure community as
the denominator. Sets of relative risks were
compiled for each family segment and for each
class of respiratory illness (Table 5.3.6). Because
the classification of data obtained always cate-
gorized disease as lower tract illness if a single
lower tract symptom was reported, the overall
rates for lower tract illnesses were somewhat
higher than those for upper tract illnesses—a con-
dition recognized as being contrary to that ob-
served in other studies. This factor should not
affect the significance of the study results, how-
ever, since data were collected in the same
manner in each community. Relative risks were
also computed for the pooled Intermediate ex-
posure group, which closely mimicked the socio-
economic characteristics of the Low exposure
area. The tabulated base rates must be inter-
preted with caution since they consider only the
32-week season of greatest respiratory illness. The
base rate expressed in illnesses per 100 person-weeks
of exposure, in fact, roughly approximates the acute
respiratory illness experience of the 32-week-long
seasons, not the experience of a single year.
Table 5.3.6. RELATIVE RISK OF ACUTE
RESPIRATORY ILLNESS
Family segment
and community
Fathers
Low
Intermediate i
Intermediate II
Pooled Intermediate
Mothers
Low
Intermediate 1
Intermediate II
Pooled Intermediate
School children
Low
Intermediate 1
Intermediate II
Pooled Intermediate
Preschool children
Low
Intermediate 1
Intermediate II
Pooled Intermediate
Relative risk of acute
respiratory illness3
Involving
upper
tract
1.00(1.64)
0.94
0.65
0.82
1.00(2.17)
0.97
0.94
0.96
1.00(2.70)
1.04
0.94
1.00
1.00(3.30)
1.04
0.87
0.97
Involving
lower
tract
1.00(1.80)
1.38
1.08
1.25
1.00(2.28)
1.30
1.16
1.23
1.00(3.51)
1.31
1.10
1.22
1.00(4.59)
1.63
1.15
1.42
All acute
respiratory
illness
1.00(3.44)
1.17
0.88
1.05
1.00(4.45)
1.14
1.05
1.10
1.00(6.22)
1.20
1.03
1.13
1.00(7.88)
1.40
1.03
1.20
aBase rates in parentheses refer to illnesses per 100 person-weeks at risk.
New York Studies
5-57
-------
Intercommunity differences in relative risks
were apparent, with the Intermediate communities
consistently reporting significant increases in
lower respiratory tract illness and total acute
respiratory illness when compared with the Low
exposure community (Table 5.3.7). The most
striking finding was a marked (22 to 43 percent)
excess risk for acute lower respiratory illness
among families in the pooled Intermediate com-
munities. In general, the Intermediate II com-
munity reported lower rates than did the Inter-
mediate I community. Although upper respiratory
illness was somewhat reduced among families
living in the Intermediate areas, the general trend
for total acute respiratory disease seemed to be
more illness in these more polluted areas. The
reduction in upper tract illness was not statis-
tically significant.
Illness Severity
Each illness was assigned a severity score
based upon the previously defined five-level
Table 5.3.7. STATISTICAL ANALYSIS OF EF-
FECTS OF AIR POLLUTION ON FREQUENCY OF
ILLNESS BY FAMILY SEGMENT3
Family
segment
All members
Adults
Children
Probability (p)b
Upper
tract
illness
NS
NS
NS
Lower
tract
illness
0.005
0.001
0.07
All
respiratory
illness
0.025
0.005
NS
Statistical analysis was restricted to residentially stable fami-
lies, i.e., those who had not moved within the past 5 years.
bNS-not significant, p > 0.10.
severity scale. To facilitate intercommunity com-
parisons, sets of relative severity scores were
calculated for each acute respiratory illness class
and for each family segment (Table 5.3.8). The
Table 5.3.8. RELATIVE SEVERITY SCORES FOR ACUTE RES-
PIRATORY ILLNESS
Relative severity score3
Family segment
and community
Fathers
Low
Intermediate I
Intermediate II
Pooled Intermediate
Mothers
Low
Intermediate I
Intermediate II
Pooled Intermediate
School children
Low
Intermediate I
Intermediate II
Pooled Intermediate
Preschool children
Low
Intermediate I
Intermediate II
Pooled Intermediate
Upper
respiratory
illnesses
1.00(1.38)
1.19
0.98
1.10
1.00(1.55)
1.19
1.04
1.14
1.00(2.11)
1.17
1.09
1.14
1.00(2.21)
1.19
1.09
1.15
Lower
respiratory
illnesses
1.00(1.76)
1.07
1.13
1.10
1.00(1.92)
1.14
1.13
1.14
1.00(2.28)
1.22
1.13
1.18
1 .00 (2.39)
1.12
1.16
1.14
All acute
respiratory
illnesses
1.00(1.58)
1.13
1.11
1.12
1.00(1.74)
1.21
1.09
1.16
1.00(2.16)
1.21
1.12
1.17
1.00(2.31)
1.29
1.21
1.26
5-58
aBase severity score in parentheses.
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
procedure was identical to that utilized in the
relative risk model, except that the base was a
severity score not an attack rate. Severity scores
for all categories of respiratory illness were con-
sistently elevated in the Intermediate commu-
nities, with the single exception of upper respira-
tory tract illnesses among fathers in the Inter
mediate II community. The greatest increases in
relative severity scores (7 to 22 percent) occurred
with acute lower respiratory illnesses. Statistical
analysis for all family members, restricted to
residentially stable families, indicated that in-
creases in- upper, lower, and all respiratory illness
were statistically significant (p < 0.01, 0.001, and
0.001, respectively).
The average number of restricted activity
days was then calculated for upper and lower
tract illnesses falling within a given severity score
(Table 5.3.9). There were some anomalous
findings: for parents, school children, and pre-
school children increasing severity scores for
upper tract illnesses from Level II to Level III
reduced rather than increased the number of
restricted activity days, and for preschool
children a similar decrease was found for lower
tract illness. One explanation might be that
febrile upper tract illnesses are more often
bacterial and thus respond more readily to
antimicrobial therapy than do the more lingering
infections produced by other agents. Despite
these anomalies, the severity score proved a
useful concept.
By applying mean severity scores to com-
munity-specific attack rates for each family
segment, one could calculate the excess per-
centage of restricted activity days that might be
attributable to living in one of the two Inter-
mediate communities (Table 5.3.10). Direct
estimates of yearly morbidity excesses cannot be
made from Table 5.3.10 because results are expressed
Table 5.3.9. AVERAGE NUMBER OF RESTRICTED ACTIVITY
DAYS ACCOMPANYING EACH SEVERITY SCORE FOR ACUTE
RESPIRATORY ILLNESS3
Family segment and
severity score*3
Parents
1
II
III
IV
School children
1
II
III
IV
V
Preschool children
1
II
III
IV
V
Mean number of restricted activity
days at each level of severity
Upper tract
illness
0
3.17
2.44
4.36
0
2.72
2.44
4.82
4.50
0
3.63
2.91
4.29
4.89
Lower tract
illness
0
3.59
3.92
5.34
0
3.22
3.91
5.73
5.96
0
3.77
3.15
5.63
6.28
All
illness
0
3.41
3.42
5.12
0
3.01
3.34
5.39
5.45
0
3.73
3.10
5.31
5.82
aAnalysis restricted to residentially stable families, i.e., those who had not
moved within the past 5 years.
"Severity score: I - illness without restricted activity, II - restricted activity but
not fever, III - restricted activity and fever, IV - all of the preceding plus a
physician visit, and V - restricted activity, fever, and otitis media diagnosed
by a physician (limited to children).
New York Studies
5-59
-------
Table 5.3.10. EXCESS RESTRICTED ACTIVITY DAYS NECESSITATED BY MORE FREQUENT
OR MORE SEVERE ACUTE RESPIRATORY ILLNESS
Family
segment
Fathers
Mothers
School
children
Preschool
children
Nuclear
family
Excess restricted activity days per 100 person-weeks of risk
Upper tract illness
Percent
excess
12
34
-3
-8
2
Excess
days
0.24
1.30
-0.21
-0.88
0.45
. _
Lower tract illness
Percent
excess
55
58
33
33
40
Excess
days
2.70
4.35
3.97
5.50
16.52
All illness
Percent
excess
42
50
20
17
26
Excess
days
2.94
5.65
3.76
4.62
16.97
in excess restricted activity days per 100 person-
weeks of risk. These rates were themselves based on
the 32-week-long season of greatest respiratory ill-
ness. A better estimate of the excess number of days
of restricted activity due to respiratory disease per
100 person-weeks of risk for a single year might be
obtained by dividing the tabulated number of excess
restricted activity days by three. The tabulated
percentage excesses, however, should be relatively
consistent over the shorter or longer period. The
yearly excess morbidity is impressive and would
involve 1 to 2 extra days of restricted activity every
year for each member of the family. A nuclear family
with two children residing in the Intermediate com-
munities might experience 5.67 extra days of
restricted activity, or 26 percent more restricted
activity than an identical family living in the Low
exposure community. The greatest percentage
increase in restricted activity was occasioned by acute
lower respiratory illness in adults; the greatest in-
crease in actual days of illness occurred among
preschool children.
Families living in the Intermediate commu-
nities also indicated that their illnesses were more
severe by reporting that a higher percentage of
their acute respiratory illnesses required restricted
activity or a physician visit (Table 5.3.11), as
might be reasoned from the relative severity
scores. The prevalence of otitis media, which was
another index of severity, did not differ between
communities. By applying community-specific physi-
cian utilization rates to community-specific attack
rates, estimates of excess physician visits attributable
to residence in the Intermediate communities can be
derived (Table 5.3.11). Among the volunteer families,
the combination of more illness and more severe
illness almost doubled physician visits resulting from
acute respiratory illness. Clearly, the estimated in-
creases in physician utilization and restricted activity
constitute a potentially costly public health problem.
Vulnerability to Repeated Acute Respiratory
Illness
During the study, family members who re-
ported more than two upper or two lower
respiratory illnesses were judged to have repeated
excessive illness. When the illness experience of
residentially stable families in the study commu-
nities was reviewed, excessive repeated total acute
respiratory illness was somewhat more common
in children from more polluted communities
(Table 5.3.12). However, no trend towards inter-
community differences was noted for acute
lower respiratory disease, which might more
likely be accompanied by an increased long-term
risk for chronic lower respiratory disease.
Validity of the Analyses
Additional analyses were undertaken to de-
termine if the previously discussed relationships
5-60
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.3.11. SEVERITY OF ACUTE RESPIRATORY ILLIMESSa
Family segment
and community
Fathers
Low
Pooled
Intermediate
Mothers
Low
Pooled
Intermediate
School children
Low
Pooled
Intermediate
Preschool
children
Low
Pooled
Intermediate
Acute respiratory illness
Percent requiring
restricted activity
Upper
tract
30
29
Lower
tract
32
49
Total
31
41
Percent requiring
physician visit
Upper
tract
4
4
28 \ 46 29 3
45 61
54
i
5
51 60 56 ; 14
68 76
73 22
39
63
59 53
74
70
15
23
Lower
tract
5
12
13
16
14
30
21
30
Total
5
9
8
n
Percent accompanied
by otitis media
__
--
--
14 3.7
26
19
28
3.5
7.2
5.6
aAnalysis restricted to residentially stable families, i.e., those who had not moved
within the past 5 years.
New York Studies
5-61
-------
Table 5.3.12. REPEATED ACUTE RESPIRATORY ILLNESSES3
Family segment
and community
Fathers
Low
Pooled Intermediate
Mothers
Low
Pooled Intermediate
School children
Low
Pooled Intermediate
Preschool children
Low
Pooled Intermediate
Percent of members with repeated acute
respiratory illness
> two upper
tract illnesses
9
2
4
5
10
11
2
14
> two lower
tract illnesses
5
5
2
7
16
15
23
28
> four total
illnesses
4
2
1
4
10
12
16
25
aAnalysis restricted to residentially stable families, i.e., those who had not
moved within the past 5 years.
between air pollutants and acute respiratory
illness could be resulting spuriously from the
effects of differences caused by reporting errors,
socioeconomic factors, cigarette smoking, or
familial susceptibility. The first factor investigated
was the quality of respondent information.
Greater than 90 percent reproducibility was
found for the fact of acute respiratory illness,
and there was greater than 98 percent repro-
ducibility for the fact of no new illness. No
intercommunity differences in the quality of
respondent information were found.
Residential mobility determines pollution
exposures and may itself influence the incidence
of acute illness since changes in life situations
seem to increase morbidity from diverse causes.13
In the present study, residentially stable families
were defined as those who had not changed
address during the preceding 5 years. Some of
the mobility was more apparent than real as some
movement likely took place within a community
rather than across communities. This was espe-
cially true in the Intermediate I community.
Mobility-specific acute respiratory illness attack
rates and appropriate sets of relative risks were
calculated by the procedures previously described
(Table 5.3.13). Mobility-specific total acute respi-
ratory illness rates formed a consistent pattern:
adults who recently moved to the Low exposure
community reported acute lower respiratory
illness rates much like those reported in the
Intermediate communities. Since most of the
families moving into the suburban fringe area
were migrating from urban core areas, this
residual susceptibility might be attributed to
previous long-term air pollution exposures. In
contrast, school children moving into the Low
community reported lower respiratory illness rates
about half way between those of the
Intermediate and Low communities. Furthermore,
preschool children who were recent migrants to
the Low pollution area reported no excess lower
respiratory disease, but did report higher rates of
upper respiratory illness than any other group,
migrant or stable. These findings could be inter-
preted as indicating that air pollution control
should most quickly benefit children. Mobility-
specific rates reinforced the observation that large
increases in acute respiratory illness, principally
lower tract illness, could be attributed to ambi-
ent air pollution exposure. The overall effect of
including more mobile families in the initial
5-62
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.3.13. ACUTE RESPIRATORY ILLNESS AMONG STABLE AND MOBILE FAMILIESa
Family segment
and community
Fathers
Low
Pooled
Intermediate
Mothers
Low
Pooled
Intermediate
School children
Low
Pooled
Intermediate
Preschool
children
Low
Pooled
Intermediate
Relative risk of acute respiratory illness
Upper tract
illness
Mobile
1.00(1.77)
0.95
1.00(2.51)
0.91
1.00(2.80)
1.09
1.00(2.71)
1.26
Stable
0.85
0.65
1.04
0.83
0.93
0.94
1.48
1.19
Lower tract
illness
Mobile
1.00(1.66)
1.39
1.00(1.80)
1.56
1.00(3.25)
1.23
1.00(5.47)
1.25
Stable
1.27
1.35
1.67
1.12
1.08
1.21
0.73
1.21
All respiratory
illness
Mobile
1.00(3.43)
1.16
1.00(4.31)
1.18
1.00(6.06)
1.16
1.00(8.18)
1.25
Stable
0.95
0.89
1.30
1.00
1.00
1.18
0.98
1.15
Stable families had lived at the same address during the 5 years preceding the
study; mobile families had moved during the 5-year period.
DBase rate per 100 person-weeks of risk in parentheses.
New York Studies
5-63
-------
Table 5.3.14. EFFECT OF SOCIOECOIMOMIC STATUS UPON ACUTE RESPIRATORY ILLNESS3
Family segment
and community
Fathers
Low
Pooled
Intermediate
Mothers
Low
Pooled
Intermediate
School children
Low
Pooled
Intermediate
Preschool
children
Low
Pooled
Intermediate
Relative risk of acute respiratory illness
Upper tract
illness
Higher
statusc
1.00(1.96)
0.68
1.00(2.29)
1.72
1.00(3.20)
1.11
1.00(3.21)
1.04
Lower
status
0.42
0.91
0.56
0.96
0.66
0.83
0.41
1.72
Lower tract
illness
Higher
status0
1.00(1.69)
1.23
1.00(1.52)
1.62
1.00(4.03)
1.15
1.00(6.90)
1.14
Lower
status
0.66
1.59
1.86
2.19
0.96
0.90
0.64
0.48
All illness
Higher
status0
1.00(3.65)
0.94
1.00(3.81)
1.24
1.00(7.23)
1.14
1.00(10.12)
1.10
Lower
status
0.53
1.22
1.08
1.38
0.82
0.87
0.57
0.87
aAnalysis restricted to residentially stable families, i.e., those who had not moved
in the past 5 years.
Base rate per 100 person-weeks of risk in parentheses.
Higher signifies homes'with <_1.00 person per room.
5-64
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
analysis was to partially obscure the excess in
illness that is most logically linked to air pollu-
tion exposures.
Differences in socioeconomic status were
found to have a statistically significant (p <
0.005) effect on frequency of upper respiratory
disease and total respiratory disease. Pooling the
Intermediate exposure communities, however,
should have minimized, if not eliminated, any
bias due to social class difference. Socioeconomic-
specific rates based on crowding in the home
were computed for family segments in the Low
exposure and pooled Intermediate exposure
communities. Sets of relative risks were then
calculated (Table 5.3.14). Families in the lower
social category reported significantly less upper
respiratory illness than did families in the higher
social category. The differences in illness that
were thought attributable to pollution exposure
were not substantially diminished by comparison
of socioeconomic-specific rates.
Cigarette smoking constitutes an intense self-
pollution hazard, as well as contributing to
indoor air pollution for other members of the
family who do not themselves smoke. Pooling
the two Intermediate communities left a small
residual excess of smokers in the more polluted
communities. Statistical analysis (Table 5.3.15)
indicated that both personal smoking and smoking in
the home significantly affected frequency of upper
and lower respiratory illness, though not total respira-
tory illness. The indicated decrease in upper tract
illness may be associated with the study design, which
arbitrarily assigned the lower tract classification when
both upper and lower tract symptoms were present.
To evaluate these effects, smoking-specific attack
rates were calculated for acute lower respiratory
illness, and relative risk models were computed for
each family segment (Table 5.3.16). Among adults,
current cigarette smokers seemed to experience more
illness than lifetime nonsmokers. Exsmokers also
appeared to have a greater relative risk when com-
pared to lifetime nonsmokers. Assessing the effect of
smoking in the home on children was more difficult.
Attempts to construct dose-response gradients for
smoking in the home were partially frustrated by
small sample sizes. Families could easily be divided
into those with no smoking in the home, those with
smoking in the home, and an indeterminant group.
Families were assigned to the indeterminant group if
the mother was an exsmoker, if the father was a
current smoker, or if someone else in the home
smoked. Exsmoking mothers were placed in the
indeterminant category because it was not known
when mothers ceased to smoke or if they resumed the
habit during the study. When the appropriate relative
risk models for children were constructed, a signifi-
cant inciease in the risk of acute lower respiratory
attack rates was noted (Table 5.3.17). The adjust-
ment somewhat diminished the relative risk attrib-
utable to pollution in children but increased the
comparable risk among adults. Clearly, there were
pollution effects quite independent of self-pollution
by cigarette smoking.
Table 5.3.15. STATISTICAL ANALYSIS OF EFFECT
OF SMOKING ON FREQUENCY OF ILLNESS3
Determinant
Personal cigarette
smoking
Cigarette smoking
in home
Family
segment
Adults
Children
Probability (p)b
Upper
tract
illness
0.005C
0.005C
Lower
tract
illness
0.005
0.01
All
respiratory
illness
NS
NS
aAnalysis restricted to residentially stable families, i.e., those who
had not moved within the past 5 years.
bNS - not significant, p > 0.10.
cAssociated with a decrease in illness frequency.
New York Studies
5-65
-------
Table 5.3.16. EFFECT OF CIGARETTE SMOKING
ON ACUTE LOWER RESPIRATORY ILLNESS3
Current smoking status
Fathers
Lifetime nonsmoker
Exsmoker
Smoker
Mothers
Lifetime nonsmoker
Exsmoker
Smoker
School children
No smoking in home
Indeterminant
Smoking in home
Preschool children
No smoking in home
Indeterminant
Smoking in home
Relative risk of acute
lower respiratory
illness by community'3
Low
1.00(1.67)
1.19
0.78
1.00(0.98)
2.32
2.08
1.00 (3.49)
0.82
1.39
1 .00 (6.04)
1.10
0.98
Pooled
Intermediate
0.96
1.51
1.47
2.09
2.66
3.35
0.80
1.49
1.30
0.82
1.16
1.27
Table 5.3.17. SMOKING-ADJUSTED ACUTE
LOWER RESPIRATORY ILLNESS
ATTACK RATES3
aAnalysis restricted to residentially stable families, i.e., those
who had not moved within the oast 5 years.
"Base rate per 100 person-weeks of risk is in parentheses.
The value of chronic bronchitis in parents as
a predictor of childhood respiratory illness was
then evaluated (Table 5.3.18). When study areas
were pooled, the few children of parents with
chronic respiratory disease were found to have
excessive acute lower resphatory illness. This
excess was most consistent in families who did
not smoke cigarettes and was most frequent in
the Intermediate communities, suggesting that air
pollution may be an additive hazard to a familial
predisposition involving vulnerability of the lower
respiratory tract.
Age-specific Attack Rates
Age-specific acute respiratory attack rates
among children from residentially stable families
were examined to determine if reasonable esti-
mates might be made of the length of exposure
Family segment
and community
Fathers
Low
Pooled Intermediate
Mothers
Low
Pooled Intermediate
School children
Low
Pooled Intermediate
Preschool children
Low
Pooled Intermediate
Relative rate of acute lower
respiratory illness15
1.00(1.58}
1.41
1.00(1.72)
1.55
1.00(3.97)
1.09
1.00(6.12)
1.10
aAnalysis restricted to residentially stable families, i.e., those
who had not moved within the past 5 years.
Base rate per 100 person-weeks of risk in parentheses.
necessary to increase the vulnerability of a child
to acute lower respiratory illness (Figure 5.3.2).
Children from the Low exposure community
were found to have somewhat higher attack rates
during the first year of life than those from the
Intermediate communities, suggesting the possi-
bility either that longer exposures were necessary
to induce a state of increased vulnerability or
that older cohorts were carrying a residual
impairment imprinted during earlier, much higher
exposures to air pollution. Significant excess
illness occured in the children from the Inter-
mediate communities at 2 to 4 years of age,
suggesting that 1 to 3 years might be the ex-
posure duration necessary to increase suscepti-
bility. One other finding seems to be important.
Intercommunity differences were obliterated when
children first attended school, suggesting that
intense exposure to respiratory agents can over-
whelm any differences caused by more subtle
environmental factors like community air pollu-
tion. Smaller pollution-related increases in attack
rates reappeared during the last 4 years of
elementary school.
5-66
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.3.18. RELATIONSHIP BETWEEN CHRONIC
BRONCHITIS IN PARENTS AND ACUTE RESPIRA-
TORY ILLNESS IN THEIR CHILDREN
Age
of child
School
Preschool
Bronchitis
in either
parent
No
Yes
No
Yes
Relative risk of acute lower
respiratory illness8
Homes
without
.smokers
1.00(3.60)
2.43
1.00(5.55)
1.56
Homes
with
smokers
1.00(4.55)
0.97
1.00(5.78)
1.54
All homes
1.00(4.14)
1.19
1.00(5.67)
1.55
aBase rate per 100 person-weeks of risk in parentheses.
Figure 5.3.2. Age-specific attack rates for
acute lower respiratory illness among children
living in communities with differing air pollu-
tion exposure.
DISCUSSION
The study demonstrated that acute lower
respiratory tract illnesses are more frequent and
more severe among families exposed to elevated
levels of ambient air pollution. Evidence for
increases in severity included additional days of
restricted activity and additional physician visits.
Residential mobility, socioeconomic status, ciga-
rette smoking in the home, and family history of
chronic bronchitis were all shown to be deter-
minants of acute lower respiratory disease in
children, but none of these factors accounted for
the excesses attributed to ambient air pollution.
Two other intervening variables might alter
the interpretation of the study: (1) indoor air
pollution by the domestic use of gas for cooking
or unvented space heating and (2) the presence
of other not easily identified or unappreciated air
pollutants. An attempt was made to disentangle
the possible effects of cooking with gas from
other determinants, but this was not possible
because almost everyone in the Intermediate areas
used gas and because sample sizes were very
small when the experience of families in the
Low exposure community was evaluated. The prob-
lem may be important since cooking with gas is
alleged to generate substantial quantities of nitrogen
dioxide, a pollutant that alters susceptibility to acute
respiratory illness.14'15 Unfortunately, no definitive
answer can be given from the present study. Prelim-
inary analysis did attribute a modest excess in the
frequency of acute lower respiratory illness to indoor
air pollution associated with the use of gas and
indicated that this effect would diminish the excess in
illness attributable to pollution by about 10 percent.
On the other hand, adjustment for the use of gas
would increase the frequency of acute upper respira-
tory disease attributable to pollution and leave
unchanged the pollution-linked excesses found for
total acute respiratory illness.
New York Studies
5-67
-------
Other pollutants were monitored in the
communities during or after the study now
reported. Metal concentrations measured in dust-
fall and high-volume samples showed an increase
in the Intermediate areas. Respirable particulate
levels were found to parallel and to constitute
roughly two-thirds of the total suspended partic-
ulate level. The Intermediate I exposure area,
which manifested the highest illness rates, is
located only a short distance from a large inter-
national airport and may be exposed to as yet
undefined pollutants arising from aircraft
emissions
REFERENCES FOR SECTION 5.3
1. Acute Conditions, Incidence and Associated
Disability. In: Health Statistics Series B.
National Center for Health Statistics, Public
Health Service, U.S. Department of Health,
Education, and Welfare. Washington, D.C.
PHS Publication No. 584. 1969.
2. Clark, D. and B. MacMahon. Preventive Medi-
cine. Boston, Little Brown and Company,
1967.
Despite these uncertainties, we concluded
that increases in acute lower respiratory disease
morbidity can be attributed to exposures of 2 to
3 years involving annual average sulfur dioxide
levels of 256 to 321 Mg/m accompanied by
elevated annual average levels of total suspended
particulate of 97 to 123 Mg/m3 and annual aver-
age suspended sulfate levels of 10 to 15 Mg/nA
There is a distinct possibility that increased
susceptibility to acute lower respiratory illness is
maintained or even induced by substantially
lower ambient air pollution exposures involving
annual average sulfur dioxide levels of 51 to 63
/ug/m^ accompanied by annual average total
suspended particulate levels of 63 to 404 Aig/irH
and annual average suspended sulfate levels of 13
to 14 jug/m3.
SUMMARY
A prospective survey of over 3000 volunteers
from families living in three communities with
differing air pollution exposures attributed signifi-
cant increases in the frequency and severity of
acute lower respiratory illnesses to ambient air
pollutants. Exposures of 2 to 3 years involving 256 to
321 jug/m3 of sulfur dioxide accompanied by 97 to
123 /ug/m3 of total suspended particulates and 10 to
15 ng/m °f suspended sulfates were linked to
excessive illnesses. There was suggestive evidence that
control of air pollution substantially decreases sus-
ceptibility to acute lower respiratory illness. The
average family of four might expect to experience 5
extra days of restricted activity and one extra
physician visit each year as a result of the excess
acute lower respiratory disease morbidity. Illness
frequency was found to increase with increasing
socioeconomic status, recent family migration, ciga-
rette smoking in the home, and parental history of
chronic bronchitis.
3. Reid, D.D. The Beginnings of Bronchitis.
Proc. Roy. Soc. Med. 62:311-316, 1969.
4. Douglas, J.W.B. and R.E. Waller. Air Pollu-
tion and Respiratory Infection in Children.
Brit. J. Prevent. Soc. Med. 20:1-8, 1966.
5. Lunn, I.E., J. Knowelden, and A.J. Handy-
side. Patterns of Respiratory Illness in Shef-
field Infant School Children. Brit. J. Prevent.
Soc. Med. 27:7-16, 1967.
6. Toyama, T. Air Pollution and Its Health
Effects in Japan. Arch. Environ. Health.
5:153-173, 1964.
7. Frank, N.R. et al. Effects of Acute Controlled
Exposure to S02 on Respiratory Mechanics in
Healthy Male Adults. J. Appl. Physiol. 77:252,
1972.
8. Frank, N.R., M.O. Ambdur, and J.L.
Whittenburg. A Comparison of the Acute
Effects of SC>2 Administered Alone or in
Combination with NaCl Particles on the
Respiratory Mechanics of Healthy Adults.
Int. J. Air Water Pollut. 5:125, 1964.
9. Ambdur, M.O. and M. Corn. The Irritant
Potency of Zinc Ammonium Sulfate of
Different Particle Sizes. Amer. Ind. Hyg.
Assoc. J. 24:326-333, July-August 1963.
10. Questionnaires Used in the CHESS Studies.
In: Health Consequences of Sulfur Oxides: A
Report from CHESS, 1970-1971. U.S. Envi-
ronmental Protection Agency. Research Triangle
Park, N.C. Publication No. EPA-650/1-74-004.
1974.
5-68
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
11. English, T.D., W.B. Steen, R.G. Ireson, P.B.
Ramsey, R.M. Burton, and L.T. Heiderscheit.
Human Exposure to Air Pollutants in Se-
lected New York Metropolitan Communities,
1944-1971. In: Health Consequences of Sul-
fur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Publication
No. EPA-650/1-74-004. 1974.
13. Rahe, R.H., J.D. McKean, and R.J. Arthur.
A Longitudinal Study of Lifechange and
Illness Patterns. J. Psychosom. Res.
70:355-366, 1967.
14. Ehrlich, R. and M.C. Henry. Chronic Tox-
icity to Nitrogen Dioxide: Effect on Resist-
ance to Bacterial Pneumonia. Arch. Environ.
Health. 77:860, 1968.
12. Grizzle, I.E., C.F. Starmer, and G.G. Koch.
Analysis of Categorical Data by Linear
Models. Biometrics. 25(3):489-504, September
1969.
15. Valand, S.B., J.D. Action, and Q. Myrvils.
Nitrogen Dioxide Exhibition of Viral Induced
Resistance in Alveolar Monocytes. Arch.
Environ. Health. 20:303, 1970.
New York Studies
5-69
-------
5.4 AGGRAVATION OF ASTHMA BY AIR POLLUTANTS:
1970-1971 NEW YORK STUDIES
John F. Finklea, M.D., Dr. P.H., John H. Farmer, Ph.D.,
Gory J. Love, Sc.D., Dorothy C. Calafiore, Dr. P.H.,
and G. Wayne Sovocool, Ph.D.
5-71
-------
INTRODUCTION
The air pollutant levels that aggravate asthma and
the complex relationships that determine the re-
sponse of asthmatics are subjects of major public
health significance because 3 to 5 percent of the U.S.
population are asthmatics. ^ National Primary
Ambient Air Quality Standards are based upon the
need to minimize or prevent deleterious effects to the
health of especially vulnerable population groups,
including asthmatics.
Identification of the precipitating factors in
most acute diseases is relatively easy. Asthmatic
attacks, though capricious, are acute, and evi-
dence of airborne substances being capable of
precipitating an attack is well documented.3"15 If
short-term air pollutant exposures precipitate
asthmatic attacks, then it should be possible to
establish a dose-response relationship between
asthma attack rates and pollutant levels. However,
the effects of air pollutants must be separated
from concurrent meteorological influences, from
effluents of agricultural activity, and from pollu-
tants peculiar to certain industries. Two additional
demanding tasks are the need to separate the
effects of one pollutant from those of another
and the need to assure that human exposures are
adequately monitored. Any single study may be
unable to accomplish these goals, but a series of
well-designed standardized studies separated in
time and space can provide the needed health
intelligence.
The study reported here had three specific
goals: first, to ascertain, in a large urban area,
which components of the ambient air pollution
mixture might cause an excess in the asthma
attack rates for panels composed of known
asthmatics; second, to establish threshold levels
for these pollutants; and third, to establish the
relative importance of air pollutants and mete-
orologic factors. The study was conducted from
October 1970 to May 1971 by the U.S.
Environmental Protection Agency (EPA) with the
cooperation of the New York City Department
of Air Resources (NYC-DAR) and the Suffolk
County (New York) Department of Health.
METHODS
The Study Population
Panelists were asthmatics who lived in three
New York communities selected for EPA
Community Health and Environmental Surveil-
lance System (CHESS) studies because they had
similar socioeconomic, racial, and ethnic makeup
but differed in pollution exposure. NYC-DAR
and the Suffolk County (New York) Department
of Health guided the selection of the CHESS
communities by furnishing historical air quality
data and by facilitating the necessary air moni-
toring. After considering long-term air pollution
exposure trends, Riverhead, Long Island, was
chosen as a Low exposure community, the
Howard Beach section of Queens as an Inter-
mediate exposure community, and the Westchester
section of the Bronx as a High exposure commu-
nity. As a result of recent improvements in air
quality, the Queens and Bronx were found to
have similar current pollutant concentrations,
often below National Primary Air Quality
Standards. Hence, the latter two communities
were redesignated Intermediate I and Intermediate
II, respectively. The three communities chosen
represent air pollution exposures involving sulfur
dioxide, suspended sulfates, and total suspended
particulates. Other pollutants are also present
including nitrates, nitrogen dioxide, organics,
metals, carbon monoxide, and gaseous hydro-
carbons. The most recent census data, along with
more current information on house and property
values, supplied the socioeconomic profiles.
Rosters of possible asthmatic panelists were
compiled from hospital clinic records and records
of practicing physicians. Each prospective subject
was interviewed by trained interviewers to obtain
information regarding the nature, frequency, and
severity of asthma. Smoking history, history of
occupational exposure to respiratory irritants, and
socioeconomic status were also determined.1**
Those subjects who had been diagnosed by a
physician as being asthmatic and whose symp-
toms consisted of wheezing with an accompany-
ing dyspnea were accepted if they had experi-
enced three such episodes during the previous
year. Selection of subjects with these criteria plus
general agreement on the part of the physicians
as to a definition of asthma enabled a uniform
selection of panelists. Only panelists who lived
within a 1.5-mile radius of monitoring stations
were enrolled. The three panels initially consisted
of 50 subjects in the Intermediate I community,
50 in the Intermediate II community, and 52 in
the Low community.
Diary Coverage
Each panelist received a diary weekly by
mail,16 recorded attacks as they occurred each
5-72
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
day, and returned the completed diary by mail
at the end of the week. When instructions for
completing the diary were given by the inter-
viewers, each panelist was asked to indicate on
the diary that he had an attack whenever he
became aware of shortness of breath combined
with wheezing. Nonresponse and diaries requiring
clarification were investigated by telephone.
Diaries received more than 12 days after the last
day covered by the diary and diaries that could
not be corrected were not accepted for data
processing. A log book was maintained to permit
constant evaluation of each panelist's perform-
ance. Diary entries were transferred to the
panelist's individual log page each week. The log
book was reviewed weekly to identify over- and
underreporting of asthma attacks, which might
indicate misunderstanding of what was to be
reported, to look for repeated corrections of
diaries, which were usually due to a misunder-
standing of the diary method, and to evaluate
frequency of failure to return diaries. Appropriate
contact was made with the panelists to improve
performance, and continued nonresponders were
dropped from the study.
Monitoring Air Pollution Exposures
Air monitoring stations were located in each
community on the roof of a two- or three-story
building 30 to 45 feet above the ground.
Measurement devices at each station included
high-volume samplers for suspended particulates,
dustfall buckets, and bubblers for nitrogen
dioxide and sulfur dioxide. Continuous 24-hour
monitoring was maintained throughout the study
period. Sampler filters and bubblers were replaced
daily, while dustfall buckets were replaced
monthly. In the laboratory, sulfur dioxide and
nitrogen dioxide were analyzed by standard EPA
methods: West-Gaeke and Jacobs-Hochheiser,
respectively. After dehydration, contents of the
dustfall buckets were weighed. High-volume
sampler filters were measured gravimetrically and
analyzed for sulfate and nitrate fractions. The
latter was determined by a reduction-diazo
coupling method with automated analysis. For
the sulfate fraction, a turbidimetric method was
used, and turbidity was measured by spectro-
photometer. A more detailed description of air
monitoring methods is presented elsewhere.^
Continuous measurements of temperature were
obtained for the duration of the study from
airports near the study communities.
Recent reevaluation by EPA indicates that
the measurement method used for nitrogen
dioxide is subject to interferences and variable
collection efficiency. Therefore, the nitrogen
dioxide data, while still included elsewhere,17 are
not included in this paper. The measurements
made, however, indicated that nitrogen dioxide
levels in the three areas were relatively low.
Statistical Analysis
Daily asthma attack rates were computed by
dividing the number of panelists who reported
one or more asthma attacks on a given day by
the number who returned their diaries for that
day. This rate computation method was used
because some panelists dropped from the study
during the study period and others were added
and because some diaries were not returned by
panelists who traveled away from the area or
who were occasionally forgetful.
The data for a single day of the study,
then, consisted of the following items: a daily
asthma attack rate, 24-hour averages of the
measurements of selected pollutants, and a
minimum and a maximum temperature. These
data items for each community were available for
each day of the study with the exception of
some missing data for the pollutants. In all
analyses, the asthma attack rate was considered
the response variable and the others the inde-
pendent variables.
The statistical analysis involved five sequential
steps: first, plotting weekly averages of asthma
attack rates, pollutant levels, and minimal
temperature; second, calculating a simple correla-
tion matrix for daily asthma attack rates, pollut-
ant levels, and minimum temperature; third,
performing multiple regression analyses to look
for pollutant effects after removal of temperature
effects; fourth, constructing temperature-specific
relative risk and excess risk models for various
pollutant levels; and fifth, computing temperature-
specific dose-response estimates to locate thresh-
old levels for pollutant effects. 18-20
To calculate relative risk functions, the total
experience of all four panels was combined into
a series of temperature-specific relative risk
models for appropriate concentrations of each of
the pollutants of interest. Since the panels
New York Studies
5-73
-------
differed to some extent from each other in the
intrinsic frequency of asthmatic attacks, a
separate temperature-specific relative risk for each
was computed based upon the experience of the
panel during the days falling in the lowest
concentration range for the particular pollutant
of interest. Then, for each panel, relative risks
for each subsequent pollutant level in the
temperature-specific series were computed using
the rate from the lowest pollutant level as the
denominator. Relative risks were then weighted
by the number of person-days of risk that each
represented and combined across communities to
achieve a single pooled rate for each temperature-
specific pollutant level. The person-days of risk
associated with any pooled rate were tallied so
that an idea of rate stability might be conveyed.
Because the analyses included days on which one
or another pollutant might be missing in one or
another community, temperature-specific groups
of person-days will not balance.
Threshold functions were computed by the
method of Hasselblad et al.1^ In setting air
quality standards, the existence of a threshold
concentration for a health effect is implicitly
assumed. It is important, therefore, to ascertain
if the dose-response relationship can be reason-
ably estimated by a function indicating no
response until some nonzero threshold concen-
tration is exceeded. Most statistical procedures
assume a strictly monotonic functional relation-
ship between two variables, and therefore, a
significant association exists at all pollutant levels,
no matter how low.
As a simple alternative, we hypothesized, for
a threshold function, a segmented line having
zero slope below exposure level x and positive
slope at levels above x. The point x is estimated
by the least squares method. This function has
been designated a "hockey stick" function.^
This least squares technique had been earlier
applied to other more general problems.-^0
The estimated threshold, estimated slope, esti-
mated intercepts, and the assumption of a linear
response above the threshold permitted calcula-
tion of an excess risk at the current short-term
air quality standard. Multiplying the difference
between the standard and the threshold by the
estimated slope of the response function provided
the estimated asthma attack rate at these levels
presently allowable once per year. Percentage of
excess risk was calculated by dividing this esti-
mated attack rate by the estimated attack rate
intercept and then subtracting one.
RESULTS
Environmental Exposures
Fluctuations in short-term peak pollutant
levels would be more likely to cause fluctuations
in asthma attack rates than would exposures over
longer averaging times. The median, 90th per-
centile, and maximum 24-hour pollutant levels
were tallied for each study community and for
each season (Table 5.4.1). As expected, the Inter-
mediate exposure areas were generally found to have
higher values than the Low exposure community at
all three points on the cumulative frequency distribu-
tions for each pollutant. Communities differed most
markedly in exposures involving sulfur dioxide, total
suspended particulates, and suspended nitrates. Rela-
tively minor intercommunity differences were found
for suspended sulfates because of unexpectedly high
levels in the Low exposure community that may have
resulted from windborne transport of the pollutant
into the suburban fringe community from more
urbanized areas. Dustfall metals also differed, with
higher levels of cadmium, lead, zinc, and other trace
metals being found in the Intermediate exposure
communities. Despite substantial differences in short-
term air pollution exposure, no community reported
sulfur dioxide concentrations as high as the 24-hour
level (365 /ug/irP) that is allowable once per year
under the relevant National Primary Air Quality
Standard. Similarly, all communities met the short-
term National Air Quality Standard for suspended
particulates.
On the basis of annual average levels, the
Low exposure community was well within all
National Primary Ambient Air Quality Standards.
The Intermediate exposure communities, while
achieving substantial recent improvements in air
quality, nevertheless exceeded the sulfur dioxide
levels permitted by the National Secondary Air
Quality Standard and the suspended particulate
levels permitted by the National Primary Air
Quality Standard. In summary, though air quality
was substantially better on the suburban fringe
than in the city itself, all three study commu-
nities were within or nearly within the National
Ambient Air Quality Standards. Thus, any ad-
verse health effect attributed to air pollutants
5-74
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.4.1. DAILY AIR POLLUTANT EXPOSURE PATTERNS IN THREE NEW YORK COMMUNITIES
DISTRIBUTED BY SEASON
I Daily pollutant
levels, ug/m
j Median j 90th percentile j Maximum
Pollutant and Fall
community
Sulfur dioxide
Low
Intermediate I
1970
13
56
Intermediate II 64
Total suspended parti culates
Low 32
Intermediate I 56
Intermediate II
57
Suspended sul fates
Low 11
Intermediate I .14
Intermediate II
Suspended nitrates
Low
Intermediate I
Intermediate II
14
1
2
1
Winter
1971
18
34
35
29
55
75
9
12
14
1
1
2
Spring I Summer Fall
1971 1971
1970
" " ~|
15 8
48 , 29
19
45
120
21 178
29 34 59
57 60
77 78
6 9
9
10
1
2
3
12
93
118
23
22
14 22
1
2 3
4
4
3
3
Winter
1971
76
119
162
56
90
121
18
20
23
2
3
4
Spring | Summer
1971 1971
45
108
104
21
71
58
53 58
102
128
16
101
118
22
17 27
22 32
2 4
5
9
8
9
rFall
1970
131
231
268
97
170
216
Winter
1971
125
287
349
134
156
188
48
28
36
3
6
•
51
39
35
6
6
6
Spring
1971
114
Summer
1971
52
143 107
219 123
91 104
193 152
217 211
32 42
36
37
42 53
4 1 6
11
20
15
14
would probably be occurring at levels below
those currently thought to protect the public
health.
Characteristics of the Study Population
Although the analysis focused on the asthma
attack experience of each panel over time rather
than on intercommunity differences, every effort
was made to select panels as alike as possible.
The three panels were similar in number of
participants, age, sex, smoking habit, and fre-
quency of asthmatic attacks during the previous
year (Table 5.4.2). The Intermediate II
community was somewhat different in that fewer
adult panelists were high school graduates. The
significant difference in educational attainment of
Intermediate II panelists might well affect report-
ing, but one would expect that the lower educa-
tional level of this panel might lead to under-
reporting for the self-administered questionnaires
used and thus make associations with pollutants
more difficult. About half of the panelists in
each community gave a history of one or more
nonasthmatic allergies including atopic dermatitis,
allergic rhinitis, and urticaria. There were, how-
ever, no differences in the history of such
allergic conditions or the duration of asthma
itself. The overall response rate for the three
panels was good; 75 percent of the more than
4700 diaries mailed were useable in the analysis.
Weekly response rates varied somewhat for any
particular week, but no notable intercommunity
differences occurred.
Temporal Patterns
Seasonal trends in asthma attack rates,
pollutant levels, and minimum temperatures can
help identify major determinants of illness that
must be isolated or adjusted for in more detailed
analyses. Seasonal trends were constructed
separately for each community by plotting
weekly averages for the variables. Asthma attack
rates were plotted together with temperature and
then in sequence with each of the air pollutants
considered. All three communities were evaluated,
but for illustrative purposes, only plots for the
Low and Intermediate I communities are pre-
sented.
The temperature patterns were typical for
the late fall, winter, and early spring months of
New York Studies
5-75
-------
Table 5.4.2. CHARACTERISTICS OF ASTHMA DIARY
STUDY POPULATION
Population
characteristics
Number of subjects
Subjects >16 years
old, %
Female subjects, %
>High school com-
pleted by house-
hold head, %
Adults currently
smoking, %
> 10 asthma attacks
during past year,
%
Community
Low
52
50
48
42
19
56
Intermediate
1
50
50
38
60
16
60
II
46
46
43
11
38
70
Total
148
48
43
39
12
61
the study (Figure 5.4.1). The Low exposure com-
munity's asthma attack curve was highest in the
early weeks of study, with a general downward
trend thereafter. The attack rate curves of the
Intermediate communities showed no overt
seasonal trends. No clear association between
temperature and asthma attack rates could be
discerned in the plots for these variables. It
should be noted, however, that asthma data were
collected during a period characterized by a rela-
tively narrow temperature band, when the
majority of mean temperatures were near or
below the freezing point.
No simple relationship linked sulfur dioxide
levels with asthma attack rates (Figure 5.4.2). A
possible temperature effect on sulfur dioxide
levels can be seen: concentrations were highest in
late fall and early winter when temperatures were
at their lowest levels, low during the blustery
weeks of late winter and early spring, and ele-
vated again during April and May.
No seasonal trend was found for total sus-
pended particulates (Figure 5.4.3). Weekly fluctu-
ations in particulate levels were wider in the
Intermediate community. The plots suggested a
tenuous relationship between asthma attack rates
and suspended particulates in the Intermediate
community but not in the Low community.
Suspended sulfate levels proved quite variable
in both communities, with the highest concen-
trations observed during the coldest weather
(Figure 5.4.4). The plots suggested no overt re-
lationship between sulfates and asthma attack
rates in any community. When the seasonal plots
were repeated for suspended nitrates (not illus-
trated), levels tended to be higher in the spring
but not related to fluctuations in asthma attack
rates.
Correlation and Stepwise Multiple Regression
Analyses
A complex web of constantly changing,
interacting factors makes the separation of the
underlying determinants of asthma attack rates
difficult. One step in reducing the problem was
to compute a simple correlation matrix to
discern the relationships between temperature,
asthma attack rate, and pollutant levels. When
this was done, pollutants were generally found to
be correlated with each other (Table 5.4.3); the
single exception was the nonsignificant correlation
between sulfur dioxide and suspended nitrates.
Pollutants varied in their relationship to
temperature: suspended nitrates rose as tempera-
ture rose; total suspended particulates were little
5-76
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
TEMPERATURE
M
Figure 5.4.1. Weekly mean attack rates
compared with weekly average minimum
temperatures.
Figure 5.4.3. Weekly mean attack rates
compared with weekly total suspended
particulate levels.
TIME, week ol study
Figure 5.4.2. Weekly mean attack rates
compared with weekly mean sulfur
dioxide levels.
Figure 5.4.4. Weekly mean attack rates
compared with weekly mean suspended
sulfate levels.
New York Studies
5-77
-------
Table 5.4.3. SIMPLE CORRELATION COEFFICIENTS OF DAILY ATTACK
RATES, POLLUTANTS, AND TEMPERATURE
Pollutant and
community
Total suspended
particulate
Low
Intermediate 1
Intermediate II
Suspended nitrates
Low
Intermediate 1
Intermediate II
Suspended sulfates
Low
Intermediate 1
Intermediate II
Sulfur dioxide
Low
Intermediate 1
Intermediate II
Minimum
temperature
Low
Intermediate 1
Intermediate II
Attack
rate
0.058
0.058
0.064
0.127
0.120
0.086
0.1 45a
0.106
-0.109
-0.022
0.061
-0.211b
-0.002
0.056
0.1 77a
TSP
0.483b
0.402b
0.604b
0.795b
0.506b
0.625°
0.364b
0.381 b
0.230b
0.006
0.029
0.075
SN
0.356°
0.112
0.289°
0.118
0.076
-0.026
-0.078
0.222°
0.258°
ss
0.369°
0.339°
0.338°
-0.1533
-0.255b
- 0.241 b
SO2
-0.226°
-0.195°
-0.365b
Significant at p < 0.05.
Significant at p < 0.001.
altered by temperature fluctuations; and sus-
pended sulfates and sulfur dioxide rose as
temperatures fell, probably reflecting the use of
fossil fuels for heating. It seemed possible to
separate the effects of most single pollutants
from the effects of other pollutants, with the
exception of sulfur dioxide and suspended sul-
fates.
No consistent significant correlations were
found between attack rates and any of the
pollutants or between attack rates and tempera-
ture. A significant positive correlation of asthma
attacks with ambient temperature was found for
the Intermediate II community. It is difficult to
interpret this relationship. It is also hard to
explain the seemingly beneficial effect of elevated
sulfur dioxide observed in the Intermediate II
community.
Multiple regression analysis was used to
clarify these relationships and to separate the
effects of pollutants from those related to fluctu-
ations in temperature (Table 5.4.4). First,
temperature alone was considered; then the
effect of pollutants after removal of any effects
that could be related to temperature was tested.
This analysis may be a conservative estimate of
pollutant effects because some of the temperature
effect may really be due to pollutants, such as
sulfur dioxide and suspended sulfates, that are
inversely correlated with temperature. In contrast
to findings of other studies, no strong tempera-
ture effect was apparent. Suspended nitrates were
linked weakly to excessive rates in the Low
exposure area, perhaps reflecting their own corre-
lation with sulfates. More importantly, multiple
regression analyses unmistakably linked elevated
suspended sulfates with excessive asthma and
5-78
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.4.4. SUMMARY OF MULTIPLE REGRES-
SION ANALYSIS CONSIDERING EFFECT OF MIN-
IMUM TEMPERATURE FOLLOWED BY EFFECTS
OF AIR POLLUTANTS3
Source of
variation
Temperature
alone
Temperature
adjusted
S02
TSP
SS
SN
Community
Low
NS
NS
NS
<0.05
<0.10
Intermediate
1
NS
NS
NS
<0.10
NS
II
NS
<0.02
NS
NS
NS
aNS-not significant, p > 0.10.
suggested that sulfur dioxide also contributed to
increases in attack rates. Since sulfur dioxide,
suspended particulates, and suspended sulfates are
closely correlated with each other, and since two
of the three had significant effects, it was
decided to limit further analyses to these three
pollutants.
Temperature-specific Relative Risk Models
Linear functions predicted that temperature
alone had little effect on asthma attack rates.
Wide ranges in temperature were monitored
during the study period, but there were fewer
than 20 warm days (Tmin > 50 °F). Perhaps,
then, the present study was centered in a colder
temperature range where there could be little or
no further increases in asthma attributable to
temperature alone. Temperature-specific relative
risk models would avoid assumptions of a linear
temperature effect and might allow separate
estimates for the effects of pollutants on cold,
freezing days (Tmjn < 30 °Ff and on somewhat
warmer days (T^ = 30 to 50 °F).
In the relative risk models, a consistent
temperature-related pattern appeared (Table
5.4.5). None of the pollutants seemed to have
any measurable effect on asthma attack rates on
very cold, freezing days. However, asthma attack
rates for cleaner days at each temperature level
were quite similar. When temperatures increased
50 °F) dose-related increments in asthma attack
rates occurred for total suspended particulates
and for suspended sulfates, but not for sulfur
dioxide. The same pattern was apparent on the
few warm days (T^j, > 50 °F) included in the
study. For temperatures above 30 °F, elevated
asthma attack rates were noted for suspended
particulate exposures of 76 to 260 M8/m and
for suspended sulfate exposures of 10 Mg/m^.
Analysis of temperature-specific attack rates thus
proved extremely helpful in identifying
temperature ranges within which pollutant effects
occurred, the general pollutant level accompany-
ing the adverse health effect, and the magnitude
of the effect. Increases as high as 13 percent in
asthma attack rates can occur with particulate
levels below the 24-hour level currently allowable
once per year. A 10 percent increase in the risk
of asthma occurred on days when suspended
sulfate concentrations exceeded 10 /ug/m3.
However, control of air pollution requires more
precise estimates of response thresholds and
dose-response relationships than are possible from
the relative risk analyses employing temperature-
specific attack rates.
Threshold and Linear Dose-response
Relationships
To determine the highest pollutant level that
would fail to have an adverse effect on asthma
attacks, the concurrent effects of ambient
temperature and the limitations of the study
were considered. The present study was restricted
to a rather narrow range of sulfur dioxide and
suspended particulate concentrations and focused
upon cooler seasons, making threshold estimates
for warmer days difficult. Sulfur dioxide levels
exceeded 200 jug/m3 and suspended particulate
levels exceeded 150 Mg/m^ on less than 5
percent of the study days. Suspended sulfate
levels ranged widely, with peak values over 50
Mg/m3. One might therefore expect that good
threshold estimates would be limited to the
lower temperature ranges and to suspended sul-
fates, with less of an opportunity to detect
thresholds for suspended particulates or sulfur
dioxide.
Threshold estimates were calculated for total
suspended particulates and for suspended sulfates,
but no linear or threshold effects were demon-
strable for sulfur dioxide (Table 5.4.6). When
to the next level above freezing (Tmin = 30 to days were cool, yet not frigid (Tmjn = 30 to 50
New York Studies
5-79
-------
Table 5.4.5. AGGRAVATION OF ASTHMA QUANTIFIED
BY TEMPERATURE-SPECIFIC RELATIVE RISK MODELS
Pollutant and
minimum
temperature
Sulfur dioxide
<30°F
30 to 50 °F
>50°F
Total suspended
particulates
<30°F
30 to 50 °F
>50°F
Suspended sulfates
<30°F
30 to 50 °F
>50°F
Integrated
24-hour level,
jug/m3
<60
61 to 80
81 to 365
>365
<60
61 to 80
81 to 365
>365
<60
61 to 80
81 to 365
>365
<60
61 to 75
76 to 260
>260
<60
61 to 75
76 to 260
>260
<60
61 to 75
76 to 260
>260
<6.0
6.1 to 8.0
8.1 to 10
>10
<6
6.1 to 8.0
8.1 to 10
>10
<8.0
8.1 to 10
>10
Relative risk3
1.00(21.0)
0.99
0.91
—
1.00(22.2)
1.06
0.98
—
1.00(21.5)
0.87
0.97
—
1.00(21.1)
0.97
1.01
—
1.00(21.3)
1.02
1.13
—
1.00(21.2)
0.92
1.05
—
1.00(19.1)
1.04
1.02
0.97
1.00(19.0)
1.04
1.09
1.08
1.00(22.1)
1.03
1.10
Person-
days of
observation"
2832
515
2228
—
10477
1481
2835
—
1902
203
287
—
4063
898
1351
—
10265
1766
2959
—
1545
396
451
—
433
702
1110
4007
2549
2300
2708
7274
915
515
1046
5-80
aBase rate for temperature-specific relative risk is in parentheses.
bPerson-days for pollutants may not balance across temperature-days
because one or more pollutants may be missing on one or more days.
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.4.6. TEMPERATURE-SPECIFIC THRESHOLD ESTIMATES FOR THE
EFFECT OF SELECTED POLLUTANTS ON ASTHMA ATTACK RATES3
Pollutant
Total suspended
particulates
Suspended
sul fates
Minimum
daily
temperature.
°F
30 to 50
30 to 50
>50
Intercept
(attack
rate per
person)
0.237
0.240
0.245
Estimated
effects
threshold.
;ug/m3
56
11.9
7.3
Slope
0.000254
0.000371
0.005009
Estimated percent
excess risk at
specified 24-hour
exposures
22% at 260 M9/m3
4% at 35jug/m3
57% at 35 /zg/m3
aThreshold functions could not be calculated for sulfur dioxide or for other temperature ranges involv-
ing the tabulated pollutants.
°F), excess asthma attacks were first noted at
total suspended particulate levels of 56 jug/m^
(Figure 5.4.5). The dose-response function further
estimates that elevations of total suspended
particulates to the once-yearly allowable level of
260 jug/m^ might be expected to induce a 22
percent increase in asthma attack rates. No
threshold fit for total suspended particulates
could be made for the large number of freezing
days or the smaller number of warmer days.
Threshold functions could be fitted for suspended
particulate sulfates (Figure 5.4.6). Levels of 12
jug/m^ induced modest excesses in asthma on
cooler days (Tmin = 30 to 50 °F). More striking
excesses at a lower threshold (7.3 jig/m^) were
found for warmer days (Tjj^ > 50 F). The
dose-response function predicts that a suspended
sulfate exposure of 35 jug/n-r would be associ-
ated with a 4 percent increase in asthma attacks
on cooler days and a 57 percent increase on
warmer days.
100 150
TSP CONCENTRATION,
15 20
SS CONCENTRATION, ii
Figure 5.4.5. Effect of total suspended par-
ticulates with minimum daily temperature be-
tween 30 and 5QOF on asthma attack rate.
Figure 5.4.6. Effect of minimum daily tempera-
ture and suspended sulfates on daily asthma
attack rate.
New York Studies
5-81
-------
DISCUSSION
Excessive asthma attack rates attributable to
air pollutants occurred only on days when
minimum temperatures were above freezing. Both
elevated levels of total suspended particulates and
suspended sulfates were linked to increases in
asthma attacks, and threshold levels were cal-
culated for both. The threshold level for sus-
pended sulfates at cool temperatures (Tmjn = 30
to 50 °F) was higher (12 Mg/m3) than that
found for warmer temperatures (7.3 /Jg/m3 for
Tmin > 50 °F). Only a single suspended partic-
ipate threshold could be calculated: 56 ;ug/m3
for cooler days. The implications of these thresh-
olds are substantial. Actual monitoring data
(1971-1972) showed that suspended particulates
in New York City will exceed the calculated
threshold in residential areas 50 to 70 percent of
all days. The picture for suspended sulfates is
equally disturbing: summer exposures in all
communities, including the Low exposure area,
exceeded the warm weather threshold on 50 to
70 percent of all days, and the threshold for
lower temperatures was exceeded on half of fall
and winter days.
The findings reported here should be in-
terpreted in light of several possible qualifying
factors. Although the study was conducted during
the low-pollen season, the possibility that plant
allergens may have been present during the few
warm weeks at the beginning and at the end of
the surveillance period must be considered. Efforts by
numerous investigators to isolate and quantify pollen
effects have met with varying levels of success. A
promising method, measurement of protein content
of ambient air, may offer a reliable means in the
future for evaluating this variable. In the present
study, the net effect of plant allergens would have
been to obscure the pollutant associations upon all
three asthma panels. Although the study design may
have eliminated the pollen season problem for a large
part of the surveillance period, it also limited surveil-
lance largely to a period when ambient temperatures
were near or below the freezing level. This limitation
could account for finding little or no temperature
effect, although temperature has been observed to be
the strongest determinant of asthma attacks in other
studies.1.9,21,22
Another problem is the possible role of
other unmeasured pollutants, which might have
interacted with the pollutants that were measured
to produce the observed deleterious health
effects. There is, after all, reasonable evidence to
speculate that suspended sulfates in the more
polluted communities might be combined with
fine particulate metals to form more potent irritants
than suspended sulfates in the less polluted area,
which may be partially of oceanic origin. Higher
metal contents were recorded in the more polluted
communities for both dustfall and suspended par-
ticulate samples.
Other intervening variables of importance
may be associated with intrinsic characteristics of
the study population. For instance, self-pollution
by cigarette smoking might obscure the effects
caused by ambient air pollution. Paradoxically,
asthmatic smokers seemed to be as responsive to
air pollutants as asthmatic nonsmokers. Perhaps
the sicker asthmatics, who are most prone to
severe repeated attacks, are less likely to smoke
and may even be so stressed that they respond
somewhat less to air pollutants. Validity of
reported asthma attacks is another possible inter-
vening variable. It was not possible in the
present study to validate asthmatic episodes by
physician visits as was done in previous in-
vestigations.9 However, past validations showed
that panelists recognized asthmatic attacks and
accurately report them. Thus, there is little
reason to believe that clinical findings in the
present study would have revealed a study bias.
After weighing the available evidence, we
concluded that 24-hour suspended sulfate levels
of 12 jug/m3 on cooler days (Tj^ = 30 to 50
°F) and 7.3 Aig/m3 on warmer days (Tmjn > 50
°F) were thresholds for the induction of exces-
sive asthma attacks. Total suspended particulate
levels of 56 jug/m3 on cooler days (Tmm = 30
to 50 °F) constituted a similar threshold. No
firm evidence could be found to associate eleva-
tions in sulfur dioxide (100 to 180 jug/m3 on 10
percent of days) with excessive asthma attack
rates on either cold or warmer days.
SUMMARY
Increased asthma attack rates among 148
panelists, who completed weekly diary reports,
were linked to suspended particulate and sus-
pended sulfate air pollution. Pollutants exerted an
effect only on days in which the minimum
temperature was above freezing (Tmjn = 30 to
5-82
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
50 °F). The estimated threshold level for total
suspended participates was very low, 56 /Jg/m^,
and that for suspended sulfates was 12 Aig/m-*.
On warmer days (Tmjn > 50 °F), the threshold
estimate for suspended sulfates was lower, 7.3
jUg/m3, and no good estimate for the total
suspended particulate threshold could be estab-
lished. Suspended particulate and suspended sul-
fate air pollution, like ambient temperature, may-
very well eventually be identified as important
determinants of asthma attack rates in urban
America.
REFERENCES FOR SECTION 5.4
1. Broder, J., P.P. Barlow, and R.J.M. Horton.
The Epidemiology of Asthma and Hay Fever
in a Total Community, Tecumseh, Michigan.
J. Allergy. 53:513-523, 1969.
2. Chronic Conditions and Limitations of Activi-
ty and Mobility: United States, July 1965-
June 1967. In: Vital and Health Statistics
from the National Health Survey. Health
Services and Mental Health Administration,
Public Health Service, U.S. Department of
Health, Education, and Welfare. Rockville,
Md. Series 10, No. 61. January 1971.
3. Gersh, L.S., E. Shubin, C. Dick, and F.A.
Schulaner. A Study on the Epidemiology of
Asthma in Children in Philadelphia. J. Aller-
gy. 59:347-357, 1967.
4. Zeidberg, L.D., R.A. Prindle, and E. Landau.
The Nashville Air Pollution Study; I. Sulfur
Dioxide and Bronchial Asthma: A Pieliminary
Report. Amer. Rev. Respiratory Dis.
54:489-503, 1961.
5. Schoettlin, C.E. and E. Landau. Air Pollution
and Asthmatic Attacks in the Los Angeles
Area. Public Health Reports. 76:545-548,
1961.
6. Yoshida, K., H. Oshima, and M. Swai. Air
Pollution and Asthma in Yokkaichi. Arch.
Environ. Health. 73:763-768, 1964.
7. Yoshida, K., H. Oshima, and M. Swai. Air
Pollution in the Yokkaichi Area with Special
Regard to the Problem of "Yokkaichi
Asthma." Ind. Health. 2:87-94, 1964.
8. Schrenk, H.H.. H. Heimann, G.O. Clayton,
W.M. Gafafer, and H. Wexler. Air Pollution
in Donora, Pennsylvania; Epidemiology of the
Unusual Smog Episode of October 1948.
Federal Security Agency, Division of Indus-
trial Hygiene, Public Health Service, U.S.
Department of Health, Education, and Wel-
fare. Washington, D.C. Public Health Bulletin
306. 1949.
9. Cohen, A.A., S. Bromberg, R.M. Buechley,
L.T. Heiderscheit, and C.M. Shy. Asthma and
Air Pollution from a Coal Fueled Power
Plant. Amer. J. Public Health. 62:1181-1188,
1972.
10. Carnow, B. The Relationship of S02 Levels
to Morbidity and Mortality in "High Risk"
Population. University of Illinois. (Presented
at American Medical Association Air Pollu-
tion Medical Research Conference. New
Orleans. October 5, 1970.)
11. Lewis, R., M.M. Gilkeson, and R.O.
McCalden. Air Pollution and New Orleans
Asthma. Public Health Reports. 77:947-954,
1962.
12. Weill, H., M.M. Ziskind, R. Dickerson, and
VJ. Derbes. Epidemic Asthma in New
Orleans. J. Amer. Med. Assoc. 790:811-814,
1964.
13. Weill, H., M.M. Ziskind, V. Derbes, R. Lewis,
R.J.M. Horton, and R.O. McCalden. Further
Observations on New Orleans Asthma. Arch.
Environ. Health. 8:184-187, 1964.
14. Lewis, R., M.M. Gilkeson, and R.W. Robison.
Air Pollution and New Orleans Asthma (2
volumes). New Orleans, Tulane University,
1962.
15. Carroll, R.E. Epidemiology of New Orleans
Epidemic Asthma. Amer. J. Public Health.
55:1677-1683, 1968.
16. Questionnaires Used in the CHESS Studies. In:
Health Consequences of Sulfur Oxides: A
Report from CHESS, 1970-1971. U.S. Envi-
ronmental Protection Agency. Research Triangle
Park, N.C. Publication No. EPA-650/1-74-004.
1974.
New York Studies
5-83
-------
17. English, T.D., W.B. Steen, R.G. Ireson, P.B.
Ramsey, R.M. Burton, and L.T. Heiderscheit.
Human Exposure to Air Pollution in Selected
New York Metropolitan Communities,
1944-1971. In: Health Consequences of Sulfur
Oxides: A Report from CHESS 1970-1971. U.S.
Environmental Protection Agency. Research
Triangle Park, N.C. Publication No. EPA-
650/1-74-004. 1974.
18. Draper, N.R. and H. Smith. Multiple Regres-
sion Mathematical Model Building (Chapter 8)
and Multiple Regression Applied to Analyses
of Variance Problems (Chapter 9). In:
Applied Regression Analyses. New York,
John Wiley and Sons, 1966. p. 234-262.
19. Hasselblad, V., G. Lowrimore, and C.J.
Nelson. Regression Using "Hockey Stick"
Function. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Unnum-
bered intramural report. 1971.
20. Quandt, R.E. The Estimation of the Param-
eters of a Linear Regression System Obeying
Two Separate Regimes. J. Amer. Statistical
Assoc. 5J:873-880, 1958.
21. Tromp, S.W. Influence of Weather and Climate
on Asthma and Bronchitis. Review of Allergy.
Vol. 22, November 1968.
22. Finklea, J.F., B.C. Calafiore, C.J. Nelson,
W.B. Riggan, and C.G. Hayes. Aggravation of
Asthma by Air Pollutants: 1971 Salt Lake
Basin Studies. In: Health Consequences of
Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Publication
No. EPA-650/1-74-004. 1974.
5-84
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
5.5 FREQUENCY AND SEVERITY OF
CARDIOPULMONARY SYMPTOMS IN ADULT PANELS:
1970-1971 NEW YORK STUDIES
Harvey E. Goldberg, M.D., Arlan A. Cohen, M.D.,
John F. Finklea, M.D., Dr. P.H., John H. Farmer, Ph.D.,
Ferris B. Benson,B.A., and Gory J. Love, Sc.D.
5-85
-------
INTRODUCTION
Elderly people, particularly those with heart
or lung disease, seem vulnerable to very high
levels of air pollutants as shown by their in-
creased mortality and morbidity during air pollu-
tion episodes.1'1 ^ However, the influence of
smaller day-to-day pollutant fluctuations on the
health of people with chronic heart or lung
disease is less well defined. In Los Angeles,
hospital admission rates for people with bron-
chitis were shown to increase with increasing
daily concentrations of ambient sulfur dioxide.1-^
During several winter seasons in London, panels
of bronchitics were asked to maintain daily
diaries indicating their overall state of health.11*
They reported increased morbidity when ambient
concentrations of sulfur dioxide exceeded 600
;ug/m3. Patients with chronic bronchitis in
Chicago were asked to maintain a record of
dates of onset and duration of acute respiratory
illness, emergency visits to physicians or hospitals,
and hospitalizations.15 Increased acute morbidity
among severe bronchitics over age 55 was associ-
ated with sulfur dioxide concentrations of 150 to
300 Mg/m or greater.
The present study investigated the influence
of both air pollution and meteorologic condi-
tions, as indexed by temperature, upon selected
daily symptom rates in elderly subjects. We
hypothesized that air pollutants exert a signifi-
cant influence on cardiopulmonary symptom rates
independent of the influence of temperature. Our
purposes were to document pollutant thresholds
for adverse health effects, to define dose-response
relationships between air pollutants and the aggra-
vation of cardiac or respiratory symptoms, and
to estimate the relative importance of fluctua-
tions in temperature and fluctuations in pollutant
levels.
METHODS
Community Selection
Three socioeconomically similar communities
were selected on the basis of air quality measure-
ments supplied by the New York City Depart-
ment of Air Resources (NYC-DAR) and the
Suffolk County (New York) Department of
Health, current information on house and prop-
erty values, and the most recently available
census data. Riverhead, Long Island, was chosen
as a Low exposure community, the Howard
Beach section of Queens as an Intermediate
exposure community, and the Westchester section
of the Bronx as a High exposure community.
Due to recent changes in air quality, the Queens
and Bronx communities were found to have
similar current pollutant concentrations, often
below National Primary Air Quality Standards.
Hence, the latter two communities were redesig-
nated Intermediate I and Intermediate II, re-
spectively.
Population Selection
Participants were selected from lists provided
by the New York City Housing Authority, local
Golden Age clubs, and religious organizations.
Each prospective participant was personally inter-
viewed concerning the presence of cough, phlegm
production, angina pectoris, enlarged heart, short-
ness of breath, leg swelling, heart attack history,
occupational exposure to pulmonary irritants,
personal smoking habits, residential mobility, age,
sex, race, and general socioeconomic status. Based
upon this interview, participants were classified
(according to criteria given elsewhere)16 as "well"
when no significant history of cardiac or respira-
tory disease was elicited, as having "heart
disease" when a history of cardiac symptoms was
reported, as having "lung disease" when a history
of respiratory symptoms was obtained, and as
having "heart and lung disease" when both
cardiac and respiratory symptoms were reported.
Within each community, between 190 and 340
subjects at least 60 years old were enrolled.
About half of the subjects were classified as
healthy, with the remainder being fairly evenly
distributed over the three disease categories. All
subjects lived within 1.5 miles of the air pollu-
tion monitoring stations.
Symptom Reporting
Different diaries were distributed to each of
the four panels, as detailed elsewhere.16 Where
common symptoms could be expected, similar
questions were asked. On a daily basis, "well"
panelists were asked about the presence of chest
pains, wheezing, cough, and phlegm; "heart"
panelists were asked about shortness of breath,
angina or dull chest pain, and leg swelling;
"lung" panelists were asked about cough or
phlegm, wheezing and shortness of breath, angina
or dull chest pain, leg swelling, and cough or
5-86
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
phlegm. Participants were asked to indicate for
each day whether their symptoms were never
present, better, the same, worse, or much worse
than usual. On a weekly basis, all panelists were
asked about the presence of a cough, cold, or
sore throat, a visit to a doctor due to illness, or
an admission to the hospital. All panelists were
asked to maintain their symptom diaries daily for
the 32-week period October 8, 1970, through
May 22, 1971.
Pollution and Weather Monitoring
One pollution monitoring station was estab-
lished in each community to provide a repre-
sentative sampling of population exposure. At
each station, sulfur dioxide (SO2), total sus-
pended particulates (TSP), suspended sulfates
(SS), suspended nitrates (SN), and nitrogen
dioxide (NO2) were monitored continuously and
analyzed on a 24-hour sample basis.17 Dustfall
was also monitored continuously, but analyzed
on a monthly sample basis. Continuous measure-
ments of temperature, humidity, barometric
pressure, and wind speed and direction were
obtained from nearby airports. A more detailed
description of the monitoring program is pre-
sented elsewhere.17
Statistical Analyses
Specific symptom rates were computed for
each of the 224 days. The numerator contained
a count of all people who responded "worse" or
"much worse." This pooling was necessary since
there were very few "much worse" responses.
The denominator contained a count of all people
who answered "much worse," "worse," "same,"
"better," or "never present." The rate is, in
effect, a measure of the percent of people at
risk who felt worse on a given day. A more
general index was based upon whether or not
panelists reported aggravation of any existing
cardiopulmonary symptom or induction of any
new symptom. This index was designated the
"combined symptom" variable.
The statistical analyses involved six steps.
First, intrinsic differences between panels and
communities were evaluated. All additional steps
considered temporal rather than spatial variations.
Second, seasonal patterns were sought by plotting
intracommunity average weekly symptom attack
rates and weekly averages for each pollutant.
Third, simple correlations between symptom rates
and pollutant levels were calculated for each
community and possible interrelationships were
defined. Fourth, stepwise multiple regression
analysis sought to identify pollutant effects after
removal of any linear temperature effects.18
Fifth, temperature-specific relative risk models for
the induction or aggravation of symptoms were
constructed for selected pollutant levels within
each community, and a combined estimate of
pollutant effects on the total elderly population
was made. Sixth, temperature-specific estimates of
linear or threshold functions for pollutants were
made, and the impact was estimated for existing
air quality standards and current air pollution
levels.19'20
RESULTS
Pollutant Exposure
Short-term peak exposures to air pollutants,
rather than long-term exposures, would most
likely be significant determinants of day-to-day
variations in the severity of cardiopulmonary
symptoms. Consequently, median, 90th percentile,
and maximum pollutant levels were tabulated for
each study community and for each season
(Table 5.5.1).
Aerometric data collected during the study
period substantiated the community difference
indicated by long-term pollution levels17 between
the Low pollution community and the two Inter-
mediate communities. During the study (fall to
spring in Table 5.5.1), median concentrations of
sulfur dioxide (13 to 18 /ig/m^) and total sus-
pended particulates (29 to 32 jug/m-') in the
Low exposure community were about half the
levels of the Intermediate communities. The
differences for suspended nitrates and suspended
sulfates were smaller, with nitrate levels of 1 to 3
jug/m-' and sulfate levels of 6 to 14 /ig/m^ being
recorded for the three communities. The difference in
long-term annual average exposures between the two
Intermediate communities was slight for all of the
pollutants previously mentioned.
As expected from study of longer term
averages, the Intermediate exposure communities
almost always exhibited higher concentrations at
all three points on the cumulative frequency
distribution. Intermediate communities had much
New York Studies
5-87
-------
Table 5.5.1. DAILY AIR POLLUTANT EXPOSURE PATTERNS
IN THREE NEW YORK COMMUNITIES DISTRIBUTED BY SEASON
Pollutant and
community
Sulfur dioxide
Low
Intermediate I
Intermediate II
Total suspended particulates
Daily pollutant levels, ug/m
Med
Fall 1 Winter
1970 1 1971
13 18
56 34
64 I 35
Low 1 32 | 29
Intermediate I i 56 j 55
Intermediate II j 57 i 75
Suspended sulfates
Low 11 9
Intermediate I
Intermediate II
Suspended nitrates
Low
Intermediate I
Intermediate II
14 i 12
14 , 14
1 1 '
2 1
1 | 2
ian
Spring
1971
" ' ~T "
Summer
1971
15 8
48 29
19 21
29 34
57 60
77 78
6 9
9
10
1
2
12
14
2
4
4
Fall
1970
45
120
178
59
93
118
23
22
22
3
3
3
90th percentile
Winter Spring
1971 1971
- - - j
76 | 45
119 108
Summer
1971
21
71
162 104 58
56 53
90
102
121 128
58
101
118
i [
18 16
20
23
17
22
2 2
3
4
5
9
22
27
32
4
8
9
Maximum
Fall
1970
131
231
268
97
170
Winter
1971
Spring
1971
I
125 114
287
349
143
Summer
1971
52
107
219 , 123
134 91 104
156
216 188
i
48
28
36
3
6
5
51
39
35
6
6
6
193
152
217 ; 211
32
42
36 37
42 53
4
11
20
6
15
14
higher peak levels of sulfur dioxide, total sus-
pended particulates, and suspended nitrates.
Somewhat surprisingly, intercommunity differences
in suspended sulfates were quite small. Levels in
the Low exposure community were higher than
expected, possibly reflecting penetration by sul-
fates from the more urbanized communities, or
even sulfates of oceanic origin. More predictably,
dustfall metals and respirable particulate levels were
higher in the Intermediate communities. None of
the study communities reported sulfur dioxide con-
centrations that exceeded the National Daily Primary
Air Quality Standard (365 ^g/m3). Short-term total
suspended particulate measurements were also within
the allowable standard (260 jug/m3).
In summary, all three communities achieved
both short- and long-term air quality standards
for sulfur dioxide. The two Intermediate commu-
nities just exceeded the allowable annual average
level of total suspended particulates but did not
violate the short-term standard. The Low ex-
posure community had total suspended particulate
levels well below allowable levels for either short-
er long-term exposure. Clearly, any excess mor-
bidity attributable to short-term exposures involv-
ing total suspended particulates or sulfur dioxide
occurred despite the fact that these communities
were within currently allowable exposure limits.
Characteristics of the Study Population
Nitrogen dioxide concentrations were calcu-
lated and were included in the statistical
analyses. These analyses indicated that nitrogen
dioxide levels in the three communities were
relatively low. However, recent reevaluation by
the Environmental Protection Agency indicates
that the measurement method for nitrogen di-
oxide is subject to interferences and variable
collection efficiencies. Therefore, these data, while
summarized elsewhere,17 are not included in this
report.
Characteristics of the study population are
given in the Appendix, Table 5.5.A.I. Panelists in
the Low pollution community were younger than
those in the Intermediate communities; i.e., a
greater proportion of panelists were less than 70
years old. Based upon background symptom
histories obtained by personal interview, there
were no major differences among the "well"
panelists. "Heart" as well as "lung" panelists in
the Low pollution community were somewhat
less symptomatic than those in the Intermediate
communities, with the only major difference
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
occurring for the symptom shortness of breath, for
which panelists with both "heart and lung" disease
reported the highest background. Otherwise, panelists
with both "heart and lung" disease reported the
highest background symptom rates of all groups, but
no large differences were observed between the three
communities.
A marked decrease in panelist participation
occurred over the course of the year in spite of
frequent phone calls to remind those people who
did not return their diaries on time (Table
5.5.A.2). Over the course of the study, three of
four panels in each of the Intermediate pollution
communities had a dropout rate exceeding 50
percent of those initially enrolled. The lowest
attrition rates were observed in the Low pollu-
tion community, where all panels retained at
least 70 percent of enrollees. At the end of the
study, there were 142, 117, and 144 active
panelists in the Low, Intermediate I, and Inter-
mediate II communities, respectively.
To gain a better idea of gross intercommu-
nity differences in symptom reporting, mean
daily symptom rates were determined for each
panel within each community (Table 5.5.2). As
expected, intra- and intercommunity differences
were apparent. "Well" panelists usually had the
lowest symptom rates. "Heart and lung" panelists
had the highest reported symptom rates, and the
rates of the other two panels were intermediate.
For all panels and all symptom reporting, rates
were lowest in the Low pollution community. As
previously indicated, the baseline demographic
characteristics and symptom status of various
panels were not the same in each community
(Table 5.5.A.I). Differences in daily symptom
rates between communities, shown in Table 5.5.2,
should be attributed to the fact that panels in
Intermediate neighborhoods were older, had more
smokers, and were otherwise inherently different
from the Low exposure panel at the start of the
study. In the Intermediate communities, panelists
generally reported symptom rates 2 to 5 times those
of the Low pollution community. The symptom
ratios for shortness of breath and angina among
"heart" panelists in the Intermediate communities
were even greater, being 8 to 13 times those
reported in the Low pollution community. No
major differences in symptom rates were observed
between the Intermediate communities. Fortu-
nately, differences between panelists in the Inter-
mediate communities were small, thus allowing
Table 5.5.2. MEAN DAILY SYMPTOM RATES OF PANELISTS WITH
RESPECT TO THE LOW POLLUTION COMMUNITY
Panel
Well
Heart
Lung
Heart
and
lung
Symptom
Cough and phlegm
Chest pain
Wheezing
Shortness of breath
Angina or chest pain
Leg swelling
Shortness of breath
Cough and phlegm
Wheezing
Shortness of breath
Cough and phlegm
Angina or chest pain
Leg swelling
Low pollution
community
symptom rate3
7.0
0.8
1.1
1.6
4.2
2.2
5.2
10.3
4.9
6.7
11.9
9.1
2.2
Intercommunity symptom
ratios (Intermediate/Low)
Intermediate I
2.2
4.1
2.6
12.8
7.9
5.4
4.6
2.2
3.9
3.7
1.9
2.9
6.1
Intermediate II
2.3
4.5
4.5
11.6
8.6
2.4
3.6
2.0
3.0
4.4
2.9
4.2
8.2
aMean daily rate is based upon the number of respondents marking worse or much worse
x 100 divided by the total number of panelists responding to that day's diary question.
New York Studies
5-89
-------
the pooling of panels if this became necessary.
We concluded that there were gross intercom-
munity differences that reflected a combination
of intrinsic intercommunity differences in the
symptom severity and possibly an additional
component attributable to fluctuations in ambient
temperature and pollutant levels. Since the study
was not designed to compare intercommunity
response rates, further analyses were restricted to
temporal variations in symptom rates within each
community, thus essentially eliminating the prob-
lem of intrinsic intercommunity difference.
Temporal Patterns
Insight into the major determinants of the
aggravation of cardiopulmonary symptoms might
be gained by evaluating seasonal trends in
symptom reporting, ambient temperature, and
ambient pollutant level. Once identified, these
determinants could be isolated or adjusted for in
later analyses. For this reason, weekly averages
were compiled for the "combined symptom" rate
in each panel and plotted over the weeks of
study. Similar plots were prepared for tempera-
ture and for each pollutant. Worsening of any
one of the symptoms that comprised the com-
bined cardiopulmonary index proved the most
stable indicator. For illustrative purposes, the
experience of panelists with combined "heart and
lung" disease living in the Low and Intermediate
I exposure areas is presented in Figure 5.5.1.
Similar plots were evaluated for all panels in
every community. Few consistent seasonal trends
were noted among panels or within communities.
There was a slight tendency towards increased
symptom reporting during late fall and winter
among three of the four panels: the "well,"
"lung," and "heart and lung" categories. Since
fluctuations in temperature are accompanied by
fluctuations in mortality caused by heart and
lung disease in the elderly population,21 it
seemed reasonable that ambient temperature
might be an important determinant of symptom
severity in elderly patients. Moreover, low
temperatures have been shown to aggravate
asthma22 and might therefore aggravate symp-
toms caused by other chronic pulmonary dis-
orders. In the two Intermediate communities,
"heart" panelists reported higher rates during the
earlier fall weeks of the study, after which the
rates gradually subsided.
Exceptions to these trends were not infre-
quent, as illustrated in the Intermediate I
community (Figure 5.5.1) by the percent of
panelists with "heart and lung" disease who re-
ported a worsening of symptoms that manifested
absolutely no seasonal trend. One other possibly
important observation should be stressed: there
was no obvious overreporting in the early weeks
of the study.
OCT
1970
1Z 16 20
TIME, week of study
MAY
1971
Figure 5.5.1. Weekly symptom aggravation
rates, "heart and lung" panel.
As expected, pollutant levels followed sea-
sonal trends that were actually summations of
the processes that govern pollutant emissions, i.e.,
atmospheric transformation and atmospheric
dispersion. Seasonal trends in environmental
factors were sequentially reviewed. Ambient
temperature followed a gently U-shaped curve,
with temperatures falling from the beginning of
the study until late February (week 17) (Figure
5.5.2). Sulfur dioxide levels rose as ambient
temperature fell, with the Intermediate commu-
nities attaining daily levels twice those of the
Low exposure community (Figure 5.5.3). Simul-
taneous plots of symptom aggravation rates and
exposure suggested a temporal correlation be-
tween symptom ratio and sulfur dioxide in the
Low exposure community. Sulfur dioxide levels
were low during March and early April, perhaps
because dispersing processes were more than
adequate even though emissions related to com-
bustion of fossil fuels for heating might still have
been substantial. On the other hand, total sus-
pended particulate levels did not exhibit any
5-90
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
TIME «eek of study
Figure 5.5.2. Weekly average maximum
and minimum temperatures during study
(1970-1971).
Figure 5.5.4. Weekly average total sus-
pended particulate concentrations and
symptom aggravation rates for "heart and
lung" panel.
LOW l ','• I I I I
"~ / \ ''\ AGGRAVATION
' \ ' * RATE
/•\ •' ^'' V A, „ A
r i
. INTERMEDIATE I
Figure 5.5.3. Weekly average sulfur
dioxide concentrations and symptom
aggravation rates for "heart and lung"
panel.
clear seasonal trends (Figure 5.5.4). Levels were
somewhat lower in the first month of the study
and seemed related to symptom rates in the Low
but not the Intermediate communities. Like sul-
Figure 5.5.5. Weekly average suspended
sulfate concentrations and symptom
aggravation rates for "heart and lung"
panel.
fur dioxide, the sulfate fraction of total sus-
pended particulates showed a clear seasonal trend
in the more polluted communities (Figure 5.5.5);
New York Studies
5-91
-------
even the Low exposure community was found to
have large variations in wintertime suspended sul-
t'ate levels, with some peak weekly averages
higher than those found in more polluted com-
munities. The simultaneous symptom rate plot
suggested that sulfates were an important de-
terminant. The nitrate fraction of total suspended
partkulates followed somewhat different temporal
patterns in the Low and Intermediate exposure
communities, but no consistent seasonal trend
was discerned (Figure 5.5.6). The temporal
pattern of nitrates could not be easily related to
simultaneous plots of symptom rates.
An overview of seasonal trends leads to the
following observations: (1) The effects of
temperature and pollutants were likely to be
entangled in a complex fashion. (2) The effect
of temperature could well mask effects attribut-
able to pollutants, especially any sulfur dioxide
effects. (3) Suspended sulfates might well con-
strain or entirely preclude any statistical analyses
that utilized the Low exposure community as a
control population. On the other hand, intrusion
of relatively high suspended sulfate concentrations
into the Low exposure community provided a
rare double-blind study of air pollution effects
since neither panelists nor survey teams could
have any idea that the Low exposure community
was in fact quite polluted.
Figure 5.5.6. Weekly average suspended ni-
trate concentrations and symptom aggravation
rates for "heart and lung" panel.
Table 5.5.3. NUMBER OF STATISTICALLY SIGNIFICANT CORRELATIONS BETWEEN
VARIOUS SYMPTOMS AND DAILY POLLUTANT OR TEMPERATURE LEVELS
Symptom
Shortness of
breath
Cough and
phlegm
Angina or
chest pain
Wheezing
Leg
swelling
Direction
of
correlative
Positive
Negative
Positive
Negative
Positive
Negative
Positive
Negative
Positive
Negative
Number of correlations
TSP
1
0
2
0
1
0
1
1
1
1
SN
0
1
1
3
0
0
1
1
0
0
ss
6
0
6
0
5
0
3
0
1
0
S02
1
2
4
2
5
2
2
1
0
2
Max.
temp
1
4
1
Q
2
3
0
4
1
0
Win
temp
2
3
1
7
3
2
0
3
4
0
5-92
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Simple Correlation Analyses
In attempting to define effects thresholds or
dose-response functions for pollutants, a hierar-
chical approach seemed advisable. Consequently,
we first evaluated community-specific simple
correlations for individual symptoms and for the
combined symptom index within each panel
before proceeding to multiple regression analyses
(Tables 5.5.3, 5.5.4, and 5.5.A.3). The indi-
vidual symptom rates and combined symptom
index were often strongly and negatively corre-
lated with temperature. Among the "heart"
panelists, however, symptoms were positively
correlated with temperature. Suspended sulfate
and sulfur dioxide were the only pollution
variables that significantly correlated with the
combined symptom rate in a consistently positive
manner. Fewer significant correlations with
symptom rates were observed for total suspended
particulates and suspended nitrates.
When the symptoms shortness of breath,
cough and phlegm, angina or chest pain, wheez-
ing, and leg swelling were considered in relation
to daily variations in air quality, the strongest
and most consistent positive correlations with
suspended sulfate pollution were observed for
shortness of breath, cough and phlegm, and
angina or chest pain. Additionally, the symptoms
Table 5.5.4. SIMPLE CORRELATION OF SYMPTOM AGGRAVATION WITH POLLUTANTS
AND AMBIENT TEMPERATURE3
Panel
Heart
Lung
Heart and lung
Well
Community
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Correlation coefficient
S02
-0.097
-0.005
0.091
0.010
0.143b
-0.291C
0.230C
-0.088
0.078
0.140b
0.008
0.206C
TSP
-0.131
0.032
-0.107
0.005
0.053
0.054
0.160b
-0.053
0.088
0.009
-0.055
0.106
SS
-0.070
0.172b
0.086
0.015
-0.356C
0.047
0.282C
0.264C
0.178b
0.021
0.199C
0.204b
SN
-0.042
0.018
-0.133
0.146b
-0.155b
0.103
0.158b
-0.087
-0.019
-0.003
-0.173b
0.049
Max.
temp
0.039
0.047
-0.120
-0.205C
-0.511C
0.097
-0.331C
-0.145b
-0.191C
-0.247C
-0.399C
-0.411C
Symptom aggravation for lung panelist involved worsening of dyspnea, cough, or
phlegm production; for heart panelists, increased frequency or severity of angina
or dyspnea; for patients with both disorders, worsening of any symptom. For
well panels, appearance of cough, phlegm, or angina was regarded as aggravation
of a chronic disorder.
3p < 0.05.
:p < 0.01.
New York Studies
5-93
-------
angina and cough and phlegm were frequently
positively correlated with sulfur dioxide, although
some inconsistencies existed. Wheezing and leg
swelling showed no consistent pattern relating to
air pollution.
When the four panels were considered as
health indicators of daily variations in air pollu-
tion, the "heart and lung" panelists proved the
most reliable index (Table 5.5.4). Among the
"heart and lung" panelists, significant positive
correlations with suspended sulfates were found
in all communities for the symptoms shortness of
breath, cough and phlegm, and angina. Among
the other three panels, significant associations of
symptom rates with pollution were fewer, and
the Intermediate I community was seemingly
responsible for most of the significant associa-
tions between symptom rates and suspended
sulfates. Based upon the principal relationships
suggested by the simple correlations, further anal-
ysis focused primarily on the combined symptom
index and on the three symptom categories cough
and phlegm, dyspnea, and angina.
It was also necessary to review the relation-
ships linking pollutants to each other and to
ambient temperature and this was done by study
of a simple correlation matrix (Table 5.5.5).
Every pollutant in every community was highly
correlated with total suspended particulates. Like-
wise, every pollutant except total suspended
particulates was correlated significantly with ambient
temperature as indexed by either maximum (Tmax)
or minimum (Tmm) daily temperatures. As expected
from their seasonal trends, sulfur dioxide and
suspended sulfates increased as ambient temperature
fell, and suspended nitrates decreased as temperature
fell. Another important relationship was the strong
positive association between sulfur dioxide and
suspended sulfates. In summary, study of the simple
correlation matrix served three purposes. First, atten-
tion was focused upon three specific symptom
categories and a combined symptom index; second,
strong associations between symptom rates and
temperature demonstrated the need to isolate this
variable in subsequent analyses; and third, complex
relationships between pollutants and temperature
were demonstrated.
Table 5.5.5. SIMPLE CORRELATIONS OF POLLUTION AND
TEMPERATURE
Pollutant
TSP
S02
SS
SN
Community
Low
Intermediate 1
Intermediate II
Low
Intermediate 1
Intermediate II
Low
Intermediate 1
Intermediate II
Low
Intermediate 1
Intermediate II
Correlation coefficient
SO2
0.364a
0.381 a
0.230a
SS
0.7953
0.506a
0.625a
0.3693
0.3393
0.3883
SN
0.4833
0.402a
0.604a
0.118
0.076
-0.026
0.3563
0.112
0.289a
Max.
temp
0.006
0.029
0.075
-0.2663
-0.1 95a
-0.3653
-0.1 53b
-0.2553
-0.2413
-0.078
0.2223
0.2583
bp < 0.05.
5-94
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Stepwise Multiple Regression Analyses
Since symptoms in most panelists seemed
strongly influenced by ambient temperature levels,
multiple regression techniques were employed to
remove all symptom variation that would be
attributed to temperature before considering the
effect of any pollutant. As previously discussed,
several important pollutants, including sulfur
dioxide and suspended sulfates, were inversely
associated with ambient temperature. After
removal of temperature effects, the effect of
each single pollutant was considered. Next,
pollutants were evaluated in combination until
each possible permutation was considered. Any
effects on symptoms due to these pollutants
could easily be obscured by the planned multiple
regression. Thus, multiple regression estimates of
the effects of these pollutants were quite con-
servative. A second qualification should be made
before interpreting the multiple regression anal-
ysis: the effects of temperature upon symptom
rates are not really simple or linear. Symptoms
tend to be aggrated by extremely low and by
excessively high ambient temperatures. Thus,
multiple regression might cause spurious associa-
tions between pollutants such as sulfur dioxide
and suspended sulfates that decrease with eleva-
tions in temperature. Nevertheless, multiple regres-
sion could prove useful in selecting pollutants for
temperature-specific analyses and in interpreting
such analyses.
A probability matrix was employed to sum-
marize the multiple regression analyses performed
for the combined symptom variable and for the
three specific symptoms thought most likely to
be responsive to fluctuations in air pollutants
(Table 5.5.6). The different panels did not re-
spond in a homogeneous fashion. Because of the
infrequent reporting of any single symptom, only
the combined symptom variable could be ana-
lyzed for the "well" panels. The "well" and
"heart disease" panels also showed less frequent
significant effects than the other panels. The
sickest panel, those with "heart and lung"
disease, reported significant symptom aggravations
attributable to environmental factors most fre-
quently. The "lung disease" panel followed a
pattern quite similar to panelists with "heart and
lung" disease.
As indicated by the preliminary analyses,
ambient temperature was the environmental factor
that most frequently exerted a significant influ-
ence on the cardiopulmonary symptom status of
every panel. Decreasing ambient temperatures
were usually accompanied by symptom aggrava-
tion and increased symptom reporting. In the
"heart" panel, temperature was related to an
opposite effect. The Low exposure community
more consistently reported significant temperature
effects than the two Intermediate communities.
All of the air pollutants were linked to
significant fluctuations in cardiopulmonary symp-
toms in one or more panels and in one or more
communities. The strongest and most consistent
pollutant effects were found for suspended sul-
fates, which were linked to a worsening of
symptoms. Inclusion of the sulfate variable just
after temperature left no significant increases in
symptom aggravation that were attributable to
other pollutants. On the other hand, the worsen-
ing of symptoms attributable to suspended sul-
fates always persisted after the removal of effects
attributable to temperature and all other pollut-
ants. In no case was an increase in suspended
sulfates coupled with an amelioration of symp-
toms. As previously discussed, the significant
aggravating effect of sulfates is especially note-
worthy because multiple regression analyses could
tend to wash out any sulfate effect. The other
two pollutants that seemed most likely to be
affecting panelists adversely were suspended par-
ticulates and sulfur dioxide. Suspended particulate
levels, which were largely independent of
temperature, seemed to predict symptom aggrava-
tion concurrent with, but not nearly as well as,
suspended sulfates. Sulfur dioxide effects, which
like suspended sulfate effects were clouded by
the initial removal of all effects associated with
temperature, were most frequently related to
symptom aggravation in the "heart and lung"
panel. In the "heart" and "lung" panels, elevations in
sulfur dioxide were at times accompanied by a
beneficial effect, probably spurious, that could result
either from chance alone or from the previously
recounted, probably nonlinear, temperature effect.
Suspended nitrate concentrations exerted
scattered significant effects that appeared incon-
sistent in direction. Although the Low exposure
community was involved in more than half of
New York Studies
5-95
-------
Table 5.5.6. SUMMARY OF MULTIPLE REGRESSION ANALYSES FOR EFFECTS OF TEMPERATURE
AND POLLUTANTS ON SYMPTOM AGGRAVATION
Panel
Well
Heart
Lung
Heart and
lung
Symptom
Combined cardio-
pulmonary symptoms
Combined cardio-
respiratory symptoms
Shortness of breath
Angina
Combined cardio-
respiratory symptoms
Shortness of breath
Cough and phlegm
Combined cardio-
respiratory symptoms
Shortness of breath
Cough and phlegm
Angina
Community
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Low
Intermediate I
Intermediate II
Source of variation3
Temperature
alone
«0.001b
<0.001b
<0.001b
NS
NS
NS
NS
NS
NS
<0.005
NS
<0.10
0.10.
Negative relationship.
5-96
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
the "significant" associations, it is unlikely that
the low levels of suspended nitrates found in this
community actually aggravated illness. Nitrates
were curiously correlated with both increasing
temperature and increasing levels of sulfates and
sulfur dioxide, even though these latter pollutants
were inversely correlated with temperature.
Perhaps these two trends resulted in spurious
associations for suspended nitrates. Other explana-
tions are also possible.
Multiple regression analyses gave unequivocal
evidence that suspended sulfates were closely
linked to aggravation of cardiopulmonary symp-
toms in the sickest elderly population. Weaker,
but probably adverse, effects might be attributed
to total suspended particulates and sulfur dioxide.
Thus, temperature-specific symptom rate analyses
could focus on these three pollutants, although it
was unlikely that their effects could be com-
pletely separated from one another.
Temperature-specific Relative Risk and
Excess Risk Models
exposure community be separately considered.
The study was partitioned into colder days, when
the minimum daily temperature (Tmm) was 20
to 40 °F, and a second category of warmer
days, when minimum temperatures were >40 °F.
Relative risks were calculated for each temperature-
specific category using as a base the symptom
aggravation rates for low pollution days. Relative
risks were then computed by using the symptom rate
for each successively higher pollutant concentration
as the numerator and the base rate from the lowest
pollution level as the denominator. In the Inter-
mediate communities where pooling was allowable,
the relative risks-for each community were weighted
by the number of person-days at risk that each
represented and combined across the two commu-
nities to achieve a single temperature-specific relative
risk for a given pollutant level. Relative risks based
upon fewer than 500 person-days of observation have
been identified in the tables since they were quite
unstable. Relative risks for each panel were combined
into a model, derived from Health Interview Survey
estimates of chronic, activity-limiting disorders in the
population 65 or older,23 to .show the experience of
each panel. The formula used was:
Multiple regression analyses probably gave
conservative estimates for the effects of sulfur
dioxide and suspended sulfates since part of the
effects that might be related to these pollutants
was removed by the stronger simultaneously
associated effect of temperature. Temperature-
specific relative risk models avoid the assumption
of linear temperature effects and instead consider
days within specific temperature ranges. The
effects of ambient temperature are not entirely
removed since any specific temperature range
may be quite broad. Thus, effects attributed to
either sulfur dioxide or suspended sulfates might
still contain a residual temperature-related aggra-
vating effect. Therefore, temperature-specific
analyses that attribute increased relative risks to
any one pollutant are liable to somewhat over-
state the effect of that pollutant because of
enhancement caused by other pollutants and by
concurrent small fluctuations in ambient
temperature.
Preliminary analyses indicated that the experi-
ence of specific panels might be pooled for the
two Intermediate areas. Conversely, panels from
the Low exposure area were either intrinsically
less symptomatic or reported fewer symptoms
because of lower pollution levels. In either case,
.prudence dictated that panels from the Low
total = 0.683 well +0.219 heart
+ 0.080 lung +0.018 heart and lung
The temperature-specific relative risk model
for sulfur dioxide showed some evidence for
increased symptoms at 24-hour exposures greater
than 81 jug/m , but many relative risks were
unstable because of the relatively few person-days
at risk upon which the estimates were based
(Table 5.5.7). In general, relative risks showed a
tendency to increase on colder days (Tmjn = 20
to 40 °F) and with elevated levels of sulfur
dioxide. Only the "heart" panel failed to show
some evidence of a sulfur dioxide effect at lower
temperatures. Estimates for the elderly population
as a whole reflect the heavy weight attached to
the "well" panel in the preceding formula. On
warmer days (Tmm > 40 °F) there was less
evidence of a sulfur dioxide effect. The more
stable pooled rates for the Intermediate commu-
nities snowed no increase from any panel.
When temperature-specific relative risk rates
for total suspended particulate exposures were
evaluated (Table 5.5.8), the observed effect was
quite small and seemed to involve "lung" and
"heart and lung" panelists on the colder (Tmjn =
20 to 40 °F) days. There was also a suggestion
New York Studies
5-97
-------
Table 5.5.7. RELATIVE RISK OF CARDIOPULMONARY SYMTPOM AGGRAVATION
IN ELDERLY PANELISTS EXPOSED TO VARYING LEVELS OF SULFUR DIOXIDE3
Panel
Well
Heart
Lung
Heart and lung
Total elderly
population
SO2 concentration
(24 hr), iUg/m3
<60
61 to 80
81 to 365
<60
61 to 80
81 to 365
<60
61 to 80
81 to 365
<60
61 to 80
81 to 365
<60
61 to 80
81 to 365
Tmjn = 20to40°F
Low
1.00(1.8)
1.01
1.45
1.00(5.7)-
0.63b
0.81b
1.00(17.0)
1.18b
1.01b
1.00(21.4)
1.61b
1.29b
1.00(7.9)
0.90
1.28
Pooled
Intermediate
1.00(17.6)
0.93
1.09
1.00(38.2)
1.05
1.04
1.00(29.4)
0.90b
0.98
1.00(52.6)
1.05
1.01
1.00(23.7)
0.96
1.07
Tmin>40°F
Low
1.00(5.9)
1.02b
0.85b
1.00(4.9)
0.65b
2.04b
1.00(12.9)
1.12b
0.78b
1.00(15.4)
1.19b
0.94b
1.00(6.4)
0.95
1.11
Pooled
Intermediate
1.00(14.7)
0.89
0.99
1.00(39.4)
0.98
0.90
1 .00 (24.2)
0.95
0.84
1.00(49.4)
0.86
0.93
1.00(21.5)
0.91
0.96
aBase rates are gtven in parentheses
Relative nsk is based on less than 500 person-days and is considered unstable
Table 5.5.8. RELATIVE RISK OF CARDIOPULMONARY SYMPTOM AGGRAVATION
IN ELDERLY PANELISTS EXPOSED TO VARYING LEVELS OF TOTAL
SUSPENDED PARTICULATESa
Panel
Well
Heart
Lung
Heart and lung
Total elderly
population
TSP concentration
(24hr),A 4°°F
Low
1.00 (6.1)
1.15b
0.66b
1.00 (5.0)
0.60b
1.20b
1.00 (16.5)
1.03b
0.85b
1.00(16.0)
0.42b
1 25
1.00 (6.9)
1.01
0.80
Pooled
Intermediate
1.00(14.3)
0.94
1.10
1.00(33.6)
1.04
0.99
1.00 (22.6)
1.18b
1.01
1.00(47.5)
1.04
0.96
1.00 (20.9)
0.98
1.07
aBase rates are given in parentheses
^Relative risk is based on less than 500 person-days and is considered unstable
5-98
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
of increase in "heart" and "well" panels on the
warmer days. It was difficult, however, to relate
unequivocal adverse health effects to the levels of
suspended particulate observed during this study.
Suspended particulate sulfate exposures were
significantly correlated with adverse health effects
(Table 5.5.9). In the Intermediate communities,
elevated sulfate levels were consistently related to
increments in symptom rates in all panels except the
"heart" panel on colder days. In general, the relative
risks for the Low exposure community reflected the
same trend, even though they were themselves
somewhat erratic. There was good evidence that a
threshold effect existed somewhere between 6 and 10
Hgjm^, with the sickest panel being the most
susceptible. The best estimates for the total elderly
population predicted that annual average suspended
sulfate levels of 10 to 12 /ug/m-' would be accom-
panied by a 6 percent morbidity excess on colder
days and a 32 percent morbidity excess on warmer
days.
In summary, temperature-specific relative risk
analyses confirmed that the sickest ("heart and
lung") panelists were more likely to report ex-
acerbations of their symptoms and that sus-
pended sulfates were the strongest contributory
pollutant determinant. The relationships for the
three pollutants considered are illustrated in
Figure 5.5.7 for "heart and lung" panelists.
Table 5.5.9. RELATIVE RISK OF CARDIOPULMONARY SYMPTOM AGGRAVATION IN
ELDERLY PANELISTS EXPOSED TO VARYING LEVELS OF SUSPENDED SULFATE3
Panel
Well
Heart
Lung
Heart and lung
Total elderly
population
SS concentration
(24hr),jug/m3
<6.0
6.1 to 8
8.1 to 10
10.1 to 12
>12
<6.0
6.1 to 8
8.1 to 10
10.1 to 12
>12
<6.0
6.1 to 8
8.1 to 10
10.1 to 12
>12
<6.0
6.1 to 8
8.1 to 10
10.1 to 12
>12
<6.0
6.1 to 8
8.1 to 10
10.1 to 12
>12
Tmin = 20to40°F
Low
1.00(8.1)
1.07
0.90
1.10
0.99
1.00(5.7}
1.27
0.81b
0.93
0.86
1.00(11.2)
1.12
1.07b
1.24
1.08
1.00b(18.7)
1.00b
1.02b
1.35b
1.53b
1.00(8.0)
1.12
0.90
1.08
0.98
Pooled
Intermediate
1.00 (17.2)
0.86
1.08
1.08
1.12
1.00 (41.7)
0.90
0.97
0.93
1.00
1.00b(27.3)
0.92
1.06
1.20b
1.10
1.00b (51.3)
1.08
1.16
1.16
1.15
1.00(24.0)
0.88
1.06
1.06
1.09
r Tmin>40°F
Low
1.00(6.3)
0.83
0.97
0.73b
1.08
1.00(4.7)
1.09b
0.98
1.02b
1.21b
1.00 (12.6)
0.97b
0.79b
1.41b
1.15b
1.00b(12.9)
1.40b
1.56b
1.16b
1.22b
1.00(6.6)
0.91
0.97
0.86
1.12
Pooled
Intermediate
1.00 (10.9)
1.24
1.25
1.44
1.37
1.00 (36.0)
0.99
1.04
1.09
1.19
1.00 (23.2)
1.07
1.00
0.93
1.20
1 .00 (46.8)
1.01
1.02
1.17
1.11
1.00(18.0)
1.17
1.18
1.32
1.31
aBase rates are given in parentheses.
"Relative risk is based on less than 500 person-days and is considered unstable.
New York Studies
5-99
-------
Overall, morbidity excesses in the elderly popula-
tion could be predicted when increased exposures
to suspended sulfates occurred.
Excess risk models were used to convey the
importance of ambient temperature alone and of
suspended sulfates alone in the exacerbation of
symptoms and to relate pollutant-induced risks to
a more familiar effect like seasonal fluctuation in
temperature (Table 5.5.10). Excess morbidity
attributable to normal seasonal temperature fluc-
tuations usually somewhat exceeded that attribut-
able to suspended sulfates.
Threshold and Linear Dose-response
Functions
In setting air quality standards, the existence
of a threshold concentration for a health effect
is implicitly assumed. It is important, therefore,
to ascertain if the dose-response relationship can
be reasonably estimated by a function indicating
no response until some nonzero threshold concen-
tration is exceeded. Most statistical procedures
assume a strictly increasing functional relationship
between two variables, and therefore that a signifi-
cant association exists at all pollutant levels, no
matter how low.
Tmin = 20 to
S02
105
rrr |-
n
40 °F
1 00
102
Iri -
Si
10B
1 16
115
—
«60 6180 »80 «60 6175 » 75 «60 6 1« 81100 »100
Tmin >40°F
ss
S02
TSP
104
AUA
» iSO 6175 75 «60 6180 81-100 »100
POLLUTANT CONCENTRATION,^"3
Figure 5.5.7. Relative risk of symptom aggt a-
vation versus sulfur dioxide, total suspended
particulate, and suspended sulfate concentra-
tions: two minimum temperature ranges,
"heart1 and lung" panel.
As a simple alternative, we hypothesized, for
a threshold function, a segmented line having a
zero slope below exposure level x and a positive
slope at levels above x. The point x is estimated
by the least squares method. This function has
been designated a "hockey stick" function.19
This least squares technique had been earlier
applied to other more general problems.20 The
purpose is to estimate the threshold, not the
entire dose-response function.
Temperature-specific threshold functions were
computed using the experience of the pooled popula-
tions from the Intermediate communities (Table
5.5.11). Illustrative plots of threshold functions for
the "heart and lung" panel' for each of three
pollutants are presented in Figure 5.5.8. Only one
threshold for sulfur dioxide could be estimated, 181
/ug/m3 for the sickest panel on the colder days (Tmm
= 20 to 40 °F). For total suspended particulates,
threshold estimates were fitted for symptom induc-
tion in the "well" panel (68 Atg/m3) on warmer days
C^min > 40 °F) and in the "heart and lung" panel
(47 jug/™3) on cooler days (Tmin = 20 to 40 °F). For
suspended sulfates, one or more threshold estimates
could be made for every panel. The estimate for the
"heart and lung" panel on the colder days was 9.2
Mg/m3. On warmer days, the elderly in general and
the "lung" panel in particular tolerated higher sulfate
thresholds (10.9 ^g/m3) than on cold days when the
threshold was zero. For the other two panels,
computed threshold estimates ranged from 0 to 9.8.
Confidence intervals overlapped for most of the point
estimates of suspended sulfate effects. Overall, sus-
pended sulfate levels of 2 to 6 jug/™3 could be
Assumed to represent a reasonable threshold on colder
days while 9 to 10 /ug/m3 would be a reasonable
overall effects threshold estimate for warmer days.
5-100
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.5.10. EXCESS CARDIOPULMONARY SYMPTOMS IN ELDERLY
PANELISTS ATTRIBUTABLE TO TEMPERATURE AND TO SUSPENDED
SULFATE EXPOSURES
Panel
Well
Heart
Lung
Heart and
lung
Total elderly
population
Community
Low
Intermediate
Low
Intermediate
Low
Intermediate
Low
Intermediate
Low
Intermediate
Percent excess attributable to
Colder
temperatures3
29
58
21
16
-12
29
36
16
17
16
Suspended sulfates"
Tmin = 20 to
40° F
10
8
-7
-7
24
20 c
35 c
16
8
6
Tmin>40°F
-27C
44
2C
9
41 c
— 7
16C
17
-14
32
aPercent excess on cool days (Tmj = 20 to 40° F) with low pollution levels as compared
with warmer days (Tmjn > 40 F) with low pollution levels.
^Percent excess on days with suspended su If ate concentrations of 10.1 to 12;u9/m3 as com-
pared with days of 6.0 M9/m^ or less.
cRelative risk is based on less than 500 person-days and is considered unstable.
Table 5.5.11. TEMPERATURE'SPECIFIC THRESHOLD ESTIMATES FOR THE EFFECT OF SELECTED
POLLUTANTS ON ELDERLY PANELISTS REPORTING CARDIOPULMONARY SYMPTOMS
Pollutant
and
panel
Sulfur dioxide
Heart and lung
Total suspended
parti culates
wen
Heart and lung
Suspended
sul fates
Well
Heart
Lung
Heart and lung
Minimum daily
temperature
(Tmin). °F
20 to 40
>40
20 to 40
20 to 40
>40
>40
20 to 40
>40
20 to 40
Intercept
(aggravation
rate per person)
0.53
0.14
0.51
0.17
0.13
0.37
0.25
0.22
0.51
Estimated
effects
threshold,
pg/m3
181
68
47
0
1.9
9.8
0
10.9
9.2
Slope
0.001352
0.000456
0.000880
0.001132
0.001769
0.007817
0.003479
0.007557
0.003277
Percent of
days threshold
exceeded in
M.Y. in 1971
<5
50 to 60
70 to 80
100
100
70
100
50
70
Estimated percent
excess risk at
presently allowable
once-a-year levels6
47
63
37
33
65
85
70
134
26
Suspended sulfate level of 50 ug/m3, which occurs infrequently in New York, was chosen arbitrarily since no standard
exists.
New York Studies
5-101
-------
s
s
»h
SO,
Implications of Threshold Estimates
Threshold estimates and their health and air
quality implications are summarized in Table
5.5.11. If one accepts 181 nglm as a sulfur
dioxide threshold on colder days, New York
study areas now exceed this level on less than 5
percent of the days during each year. However,
if sulfur > dioxide levels climbed as high as the
currently permissible once yearly 24-hour level
(365 jug/m^), a substantial morbidity excess, 47
percent over the base, might be expected. For
total suspended par-ticulates, the relevant thresh-
old levels are exceeded on 50 to 80 percent of
the days during each year and morbidity excesses
of 37 to 63 percent might be expected among
elderly groups. For suspended sulfates, the low-
temperature thresholds are exceeded every day;
the thresholds for warmer days are exceeded on
50 to 70 percent of all days. Morbidity excesses
at peak 24-hour exposures currently monitored
(A/50 jug/m5) would range from 26 to 134 per-
cent. These findings suggest that short-term
exposures involving sulfur dioxide, total sus-
pended particulates, and, more importantly, sus-
pended sulfates will adversely affect the health of
our senior citizens even after the current urban
air quality goals for sulfur dioxide and partic-
ulates are achieved.
10 15
POLLUTMIT CO«CENTRftTIOH,jj/n3
Figure 5.5.8. Temperature-specific threshold
estimates for symptom aggravation by sulfur
dioxide, total suspended particulates, and
suspended sulfates: minimum temperature,
20 to 40 °F; "heart and lung" panel.
DISCUSSION
All three goals of the study were at least
partially attained. Three air pollutants were found
to be associated with aggravation of chronic
cardiopulmonary disease symptoms in panels of
elderly patients. These pollutants were sulfur
dioxide, total suspended particulates, and, most
strongly, suspended particulate sulfates. Conserva-
tive estimates utilizing patterns of significance in
multiple regression analyses suggested that most,
if not all, of the adverse effects could be
attributed to suspended sulfates since no signifi-
cant residual effects were noted after the effects
of ambient temperature and suspended sulfates
were removed. Temperature-specific relative risk
estimates and threshold calculations are more
5-102
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
likely to attribute a significant role to any single
pollutant than the more conservative multiple
regression analyses. Such was the case, with all
pollutants causing aggravation of symptoms in the
most vulnerable, sickest panelists. Again, tempera-
ture-specific analyses assigned larger and more
consistent effects and lower thresholds to sus-
pended sulfates than to the other two pollutants.
Unexpectedly high levels of suspended sul-
fates in the Low exposure community provided a
double-blind study of the effects of this pollut-
ant. Despite the relatively small panel sizes and
fewer polluted days, elderly panelists from this
community generally reported higher symptom
rates on days when sulfate levels exceeded 10.0
/zg/m . Furthermore, the total elderly population
of panelists from the Low exposure community
had estimated higher symptom rates on days
when sulfur dioxide exceeded 80 (J.g/m than on
less polluted days. The deleterious effects of
suspended sulfates in the Low exposure commu-
nity were significant in both the multiple regres-
sion and temperature-specific analyses.
The effects of short-term peak exposures in
today's urban air should be placed in a broader
perspective. Air quality has improved, suggesting
that peak exposures and deleterious health effects
on patients with chronic heart and lung disorders
might have been much greater in the past. More-
over the pollutant effects now observed in one
of our largest metropolitan areas are generally of
the same magnitude as deleterious effects of
normally accepted seasonal and short-term
changes in ambient temperature. The estimated
threshold effects for suspended particulates and
suspended sulfates are close to the observed
background levels monitored in remote locations.
Analyses reported here, however, imply that
major, potentially life-threatening health impair-
ments in our large elderly population can be
related to commonly monitored levels of urban
pollutants. Furthermore, modern heating and air
conditioning, to some extent, are ameliorating the
very large mortality and morbidity effects of
wide swings in ambient temperature. Therefore, if
urban air pollution is not well controlled and the
use of central heating and air conditioning con-
tinues to increase, pollutant effects could easily
surpass those now attributed to ambient tempera-
ture.
Several problems that arose in the present
study must be considered, including intercom-
munity differences, difficulties in panelist recruit-
ment, panelist participation, and the role of
indoor air pollutants. Panelists from the Inter-
mediate communities differed intrinsically from
panelists in the Low exposure community; pollut-
ant effects could be identified only by time
series analyses within each exposure community.
Panelists from the Intermediate communities were
also somewhat less faithful participants. The utili-
zation of elderly panelists for the study of day-
to-day changes in disease symptoms involved
many difficulties. Initial recruitment posed the
problem of finding elderly people with heart
and/or lung disease who would agree to
participate in the study. Recruitment was very
time-consuming as only about 5 percent of those
interviewed both met our criteria and agreed to
participate. Once recruited, motivation became a
major problem. Since the potential benefits of
pollution control may not be realized by these
people and we provided no special incentive for
them to assure their participation, it proved diffi-
cult to maintain their interest. Many of the most
elderly found questionnaires difficult to under-
stand; thus, frequent explanations were necessary
to assure that the panelists filled in their diaries
daily rather than at the end of the week. Many
panelists found the study too bothersome in
requiring maintenance of a daily diary for 32
weeks, and there was a marked decrease in active
participants over the course of the study. The
questions asked on a daily basis also contributed to
the difficulty of this study. Recruitment of panelists
with lung disease was contingent solely upon a
history of cough, phlegm, or both for more than 3
months out of a year. Thus, all "lung" and "heart and
lung" panelists could potentially vary in this symp-
tom, and therefore each diary response was valid for
evaluation of daily symptom change. However, only
38 percent of all "lung" panelists had a history of
shortness of breath, so that daily variation in this
symptom was based upon a limited subsample of this
panel. Similarly, the background history varied in the
number of panelists who could be expected to exhibit
day-to-day changes in other diary symptoms.
Despite these frustrating difficulties, the use of
elderly panelists proved feasible and provided
health intelligence needed for air quality criteria
and indirectly for evaluation of National Primary
Ambient Air Quality Standards.
Indoor air quality could not be evaluated in
this study. However, suspended sulfates, which
proved to be the pollutant that manifests the
strongest relationships to cardiopulmonary
symptoms, are primarily fine particulates that
should penetrate indoors.^
New York Studies
5-103
-------
After careful consideration of all the evi-
dence, we concluded that there was preponderant
evidence that fluctuations in suspended sulfate
levels can be linked to significant worsening in
the symptoms of patients with chronic heart and
lung disease. Suspended sulfate levels commonly
found in the major cities of our country ( ~10
jug/trr*) have a deleterious impact even when 24-hour
sulfur dioxide and suspended particulate levels are
well within the limits set by National Primary Air
Quality Standards. We also believe it is possible, but
not likely, that the relatively low 24-hour levels of
sulfur dioxide (180 /zg/m^) and total suspended
particulates (50 to 70 /ng/m^) may also have an
adverse effect on the symptom status of the sickest
panelists. Such levels are well below the currently
acceptable upper limits for 24-hour peak exposures
involving these pollutants. We could not be certain
that each of these pollutants has an independent
effect, but believe it prudent to assume that any of
the aforementioned pollutants may result in adverse
health effects in elderly populations. Repeated
studies should be conducted to strengthen confidence
in the threshold values presented in this report.
REFERENCES FOR SECTION 5.5
1. Firket, J. Fog Along the Meuse Valley. Trans.
Faraday Soc. 32:1192-1197, 1936.
2. Firket, J. The Cause of the Symptoms
Found in the Meuse Valley during the Fog
of December 1930. Bull. Roy. Acad. Med.
Belgium. 77:683-739, 1931.
3. Schrenk, H.H., H. Heimann, G.O. Clayton,
W.M. Gafafer, and H. Wexler. Air Pollution
in Donora, Pennsylvania; Epidemiology of the
Unusual Smog Episode of October 1948.
Federal Security Agency, Division of Indus-
trial Hygiene, Public Health Service, U.S.
Department of Health, Education, and Wel-
fare. Washington, D.C. Public Health Bulletin
306. 1949.
4. Gore, A.T. and C.W. Shaddick. Atmospheric
Pollution and Mortality in the County of
London. Brit. J. Prev. Soc. Med. 72:104-113,
1958.
SUMMARY
Over 500 elderly panelists who were well or
who reported cardiopulmonary symptoms par-
ticipated in a diary study linking aggravation or
induction of symptoms to fluctuations in ambient
air pollutants and minimum daily temperature.
Panels were previously diagnosed by a physician
as being well (without cardiopulmonary symp-
toms), having heart disease, having lung disease,
or having combined heart and lung disease. Both
temperature and pollutants were associated with
changes in symptom status. Decreasing tempera-
tures induced or aggravated symptoms in all but
the heart disease panelists, who reported worsen-
ing of symptoms with elevations in ambient
temperature. Elevated levels of sulfur dioxide,
total suspended particulates, and especially
suspended sulfates were linked to symptom
exacerbations. However, within temperature-
specific ranges, suspended sulfates exhibited the
strongest and most consistent association with
symptom aggravation. The best estimate of thresh-
old levels was <10 Mg/m3 for suspended sul-
fates. The implications of these thresholds in light of
current air pollution levels and the relative impor-
tance of seasonal swings in temperatures were dis-
cussed, and the problem was judged to be of
considerable public health significance.
5. Burgess, S.G. and C.W. Shaddick. Bronchitis
and Air Pollution. Roy. Soc. Health J.
79:10-24, 1959.
6. Scott, J.A. The London Fog of December
1962. Med. Office. 7(99:250-252, 1963.
7. Greenburg, L., M.B. Jacobs, B.M. Drolette, F.
Field, and M.M. Braverman. Report of an
Air Pollution Incident in New York City,
November 1953. Public Health Reports.
77:7-16, 1962.
8. McCarroll, J. and W. Bradley. Excess Mor-
tality as an Indicator of Health Effects of
Air Pollution. Amer. J. Public Health.
56:1933-1942, 1966.
9. Glasser, M., L. Greenburg, and F. Field.
Mortality and Morbidity during a Period of
High Levels of Air Pollution, New York,
November 23-25, 1966. Arch. Environ.
Health. 75:684-694, 1967.
10. Logan, W.P.D. Mortality in the London Fog
Incident, 1952. Lancet. 264:336-338, 1953.
5-104
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
11. Greenburg, L., F. Field, J.I. Reed, and C.L.
Erhardt. Air Pollution and Morbidity in New
York City. J. Amer. Med. Assoc.
752:161-164, 1962.
12. McCarroll, J.R., E.J. Cassell, E. W. Walter,
J.D. Mountain, J.R. Diamond, and l.R.
Mountain. Health and the Urban Environ-
ment; V. Air Pollution and Illness in a
Normal Urban Population. Arch. Environ.
Health. 74:178-184, 1967.
13. Sterling, T.D., J.J. Phair, S.V. Pollack, D.A.
Schumsky, and I. deGrott. Urban Morbidity
and Air Pollution. Arch. Environ. Health.
73:158-170, 1966.
14. Lawther, P.J., R.E. Waller, and M. Hender-
son. Air Pollution and Exacerbations of
Bronchitis. Thorax. 25:525-539, 1970.
15. Carnow, B.W., M.H. Lepper, R.B. Shekelle,
and J. Stamler. The Chicago Air Pollution
Study: Acute Illness and SO2 Levels in
Patients with Chronic Bronchopulmonary
Disease. Arch. Environ. Health. 75:768-776,
1969.
16. Questionnaires Used in the CHESS Studies.
In: Health Consequences of Sulfur Oxides: A
Report from CHESS, 1970-1971. U.S. Envi-
ronmental Protection Agency. Research Tri-
angle Park, N.C. Publication No. EPA-650/1-74-
004. 1974.
17. English, T.D., W.B. Steen, R.G. Ireson, P.B.
Ramsey, R.M. Burton, and L.T. Heiderscheit.
Human Exposure to Air Pollution in Selected
New York Metropolitan Communities,
1944-1971. In: Health Consequences of Sul-
fur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Publication
No. EPA-650/1-74-004. 1974.
18. Draper, N.R. and H. Smith. Multiple Regres-
sion Mathematical Model Building (Chapter 8)
and Multiple Regression Applied to Analyses
of Variance Problems (Chapter 9). In:
Applied Regression Analyses. New York,
John Wiley and Sons, 1966. p. 234-262.
19. Hasselblad, V., G. Lowrimore, and C.J. Nel-
son. Regression Using "Hockey Stick" Func-
tion. U.S. Environmental Protection Agency.
Research Triangle Park, N.C. Unnumbered
intramural report. 1971.
20. Quandt, R.E. The Estimation of the Param-
eters of a Linear Regression System Obeying
Two Separate Regimes. J. Amer. Statistical
Assoc. 53:873-880, 1958.
21. Buechley, R.W., J. VanBruggen, and L.E.
Truppi. Heat Island = Death Island? Environ.
Res. 5(l):85-92, March 1972.
22. Cohen, A.A., S. Bromberg, R.M. Buechley,
L.I. Heiderscheit, and C.M. Shy. Asthma and
Air Pollution from a Coal Fueled Power
Plant. Amer. J. Public Health. 62:1181-1188,
1972.
23. Chronic Conditions and Limitations of
Activity and Mobility: United States, July
1965-June 1967. In: Vital and Health Statis-
tics from the National Health Survey. Health
Services and Mental Health Administration,
Public Health Service, U.S. Department of
Health, Education, and Welfare. Rockville,
Md. Series 10, No. 61. January 1971.
24. Wagman, J., R.E. Lee, Jr., and C.J. Axt.
Influence of Some Atmospheric Variables on
the Concentration and Particle Size Distribu-
tion of Sulfate in Urban Air. Atmospheric
Environ. 7:479-489, 1967.
New York Studies
5-105
-------
APPENDIX
Table 5.5.A.1. BACKGROUND INTERVIEW INFORMATION
Panel and characteristic
Well
Over 70 years old
Smokers
Males
Education-high school
or less
Shortness of breath
Leg swelling
Medication, heart or water
Heart
Over 70 years old
Smokers
Males
Education-high school
or less
Shortness of breath
Leg swelling
Medication, heart or water
Angina
Chest pain with exertion
or upset relieved by rest
Heart attack
Large heart
Lung
Over 70 years old
Smokers
Males
Percent of enrolled panelists
Low
31
48
38
92
7
15
7
25
50
42
88
34
27
34
27
36
43
34
27
31
65
I nterme-
diate I
53
71
37
97
16
25
5
38
51
42
94
57
52
62
21
74
45
21
48
55
41
Interme-
diate II
51
68
40
97
15
16
4
35
66
36
97
48
31
56
21
71
36
20
59
47
47
Panel and characteristic
Education-high school
or less
Shortness of breath
Leg swelling
Cough and phlegm > 3
months
Cough and phlegm > 3
months and shortness
of breath
Heart and lung
Over 70 years old
Smokers
Males
Education-high school
or less
Shortness of breath
Leg swelling
Medication, heart or water
Cough and phlegm > 3
months
Cough and phlegm > 3
months and shortness
of breath
Angina
Chest pain with exertion or
upset relieved by rest
Heart attack
Large heart
Percent of enrolled panelists
Low
85
17
14
35
3
31
56
63
88
63
38
69
19
31
25
38
38
38
I nterme-
diate 1
97
41
28
13
22
46
56
42
98
83
63
63
6
46
13
27
27
21
Interme-
diate II
97
53
24
15
27
55
49
47
98
71
54
60
15
29
13
31
34
21
5-106
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.5.A.2. CHANGE IN PANEL MEMBERSHIP WITH TIME
Panel
Well
Heart
Lung
Heart and lung
All panels
Week of study
1
30
1
30
1
30
1
30
1
30
Number of active panel members
Low
103
74
41
33
32
23
16
12
192
142
Intermediate 1
82
38a
90
33a
33
13a
57
33
262
117a
Intermediate II
144
57a
88
32a
41
22
70
33a
343
144a
aGreater than 50 percent reduction of panel size over the course of the 30 weeks.
Table 5.5.A.3. CORRELATION COEFFICIENTS FOR SYMPTOMS, POLLUTANTS, AND TEMPERATUREa
Symptom
and panel
Shortness of breath
Lung
Heart
Heart and lunq
Cough and phlegm
Well
Lung
Community
Low
Int. I
Int. II
Low
Int. I
Int. II
Low
Int. I
Int. II
Low
Int. I
Int. II
Low
Int. I
Int. II
TSP
-0.032
0.035
0.008
-0.088
-0.055
0.140
0.091
0.050
0.013
0.036
-0.044
0.068
-0.039
-0.002
0.014
SN
0.096
-0.130
0.072
-0.109
0.072
0.098
0.098
-0.027
0.054
-0.009
-0.139
-0.019
0.067
-0.118
0.155
SS
-0.043
0.362
-0.007
-0.074
0.130
0.328
0.284
0.322
0,115
0.053
0.233
0.174
-0.008
0.277
0.034
S02^
-0.085
0.175
-0.305
0.034
-0.183
0.073
0.036
0.049
0.031
0.125
0.031
0.202
-0.042
0.110
-0.288
Max.
temp
-0.153
-0.456
0.086
0.122
0.109
0.018
-0.143
-0.249
-0.075
-0.237
-0.443
-0.406
-0.168
-0.486
0.207
Min.
temp
-0.105
-0.383
0.164
0.074
0.128
0.036
-0.132
-0.143
-0.006
-0.229
-0.433
-0.388
-0.125
-0.444
0.249
New York Studies
5-107
-------
Table 5.5.A.3 (continued). CORRELATION COEFFICIENTS FOR SYMPTOMS, POLLUTANTS,
AND TEMPERATURES
Symptom
and panel
Heart and lung
Angina or chest pain
Well
Heart
Heart and lung
Wheezing
Well
Lung
Leg swelling
Heart
Heart and lung
-
Community
Low
Int. I
Int. II
Low
Int. I
Int. II
Low
Int. I
Int. II
Low
Int. I
Int. II
Low
Int. I
Int. II
Low
Int. I
Int. II
Low
Int. I
Int. II
Low
Int. I
Int. II
r • "
TSP
0.175
-0.067
0.188
0.040
-0.012
-0.016
-0.046
0.044
-0.085
0.067
-0.032
0.110
0.066
-0.072
-0.013
-0.116
0.113
0.130
-0.096
-0.038
0.027
-0.018
-0.137
0.184
SN
0.081
-0.113
0.072
0.106
0.044
-0.084
-0.090
0.056
-0.109
0.048
-0.077
0.030
0.077
-0.008
-0.017
-0.045
-0.162
0.176
0.038
0.069
0.071
0.046
-0.004
0.097
SS
0.252
0.119
0.270
-0.023
0.145
0.078
-0.038
0.136
-0.001
0.166
0.205
0.200
0.230
0.165
-0.040
-0.092
0.295
-0.050
-0.064
-0.026
-0.080
0.158
-0.040
0.020
so2
0.171
-0.234
0.122
-0.054
-0.181
0.150
-0.169
0.119
0.127
0.143
-0.004
0.215
0.048
0.035
0.088
0.136
0.247
-0.163
0.004
0.024
-0.257
-0.118
-0.081
-0.081
Max.
temp
-0.128
-0.124
-0.293
-0.038
-0.008
-0.197
0.198
0.065
0.130
-0.224
-0.079
-0.184
-0.125
-0.248
-0.205
-0.094
-0.593
0.001
0.089
0.296
0.105
0.088
-0.045
0.001
Min.
temp
-0.132
-0.076
-0.256
-0.107
-0.050
-0.106
0.206
0.116
0.145
-0.244
-0.033
-0.120
-0.109
-0.150
-0.135
-0.089
-0.580
-0.860
0.125
0.275
0.121
0.159
-0.021
-0.022
Underlined are significant at the 5 percent level.
5-108
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
5.6 VENTILATORY FUNCTION IN SCHOOL CHILDREN:
1970-1971 NEW YORK STUDIES
Carl M. Shy, M.D., Dr. P.H., Victor Hasselblad, Ph.D.,
John F. Finklea, M.D., Dr. P.H., Robert M. Burton, B.S.,
Mimi Pravda, B.S., Robert S. Chapman, M.D.,
and Arlan A. Cohen, M.D.
5-109
-------
INTRODUCTION
Previous studies in Japan, Great Britain,
Italy, and North America have shown decreased
lung function in children living in areas of high
air pollution.1"6 Based on these reports, lung
function in school children, indexed by
0.75-second timed forced expiratory volume
(FEVQ 75), was chosen as one of the health
indicators in the Environmental Protection
Agency's Community Health and Environmental
Surveillance System (CHESS). Highly significant
decreases in ventilatory function are seen with
aging, chronic obstructive lung disease, and self-
pollution by cigarette smoking.7 Thus, significant
decrements in ventilatory function early in life
may reasonably be viewed as a risk factor for
later respiratory disorders. Furthermore, cessation
of self-pollution with cigarette smoke is accom-
panied by an amelioration of chronic respiratory
disease symptoms.8 Therefore, stringent control
of ambient air pollution might be associated,
over time, with improvements in defective venti-
latory function and with decreased risk of
chronic respiratory disease that can be attributed
to earlier elevated air pollution exposures. This
study was designed to test three hypotheses: (1)
early childhood exposure to ambient air pollution
reduces ventilatory function in later childhood,
(2) improvements in air quality tend to reverse
deficits in ventilatory function, and (3) monthly
variations in ambient air pollution exposures are
paralleled by variations in ventilatory function.
METHODS
Three New York neighborhoods were selected
to represent an air pollution gradient among
communities reasonably similar in general socio-
economic characteristics. After considering long-
term air pollution exposure trends, Riverhead,
Long Island, was chosen as a Low exposure
community, the Howard Beach section of Queens
as an Intermediate exposure community, and the
Westchester section of the Bronx as a High
exposure community. As a result of recent im-
provements in air quality, the Queens and Bronx
neighborhoods were found to have similar current
pollution levels, often below the National Primary
Air Quality Standards. Hence, the latter two com-
munities were redesignated Intermediate I and
Intermediate II, respectively. During the study
months, air quality averages for both suspended
particulates and sulfur dioxide were between the
National Primary and National Secondary Air
Quality Standards. However, during the previous
15 years, the Intermediate exposure communities
had experienced annual average pollution levels
that were 2 to 5 times those recorded during the
present study.9
Within each community, children in three
elementary schools, including kindergarten through
sixth grades (ages 5 to 13 years) and located
within 1.5 miles of a single air monitoring sta-
tion, were enrolled in the study. Few children
were bused into any of the participating schools.
Each air monitoring station sampled air from the
immediate home neighborhood of the children.
Four rounds of pulmonary function testing were
conducted in the schools in the late fall
(November-December 1970), early winter (January
1971), late winter (February-March 1971), and
spring (April 1971). The National Cylinder Gas
(NCG) Pulmonary Function Indicator was used to
measure 0.75-second forced expiratory volume
(FEVo_75). This instrument is small, light, easily
portable, and capable of giving immediate digital
read-outs of FEVQ 75, FEVj Q, and vital capac-
ity. The instrument measures air flow by means
of a change in temperature across a heat-sensitive
platinum element within a transducer. Air flow is
electronically integrated over time to produce a
volume measurement. The pulmonary function
indicators were calibrated against a Stead-Wells
volume spirometer. They were tested for repro-
ducibility by obtaining six or seven successive
FEVg 75 measurements for each indicator with
trained subjects and comparing these measure-
ments with those obtained with a Stead-Wells spi-
rometer connected in series. Percent differences
between the pulmonary function indicator and the
spirometer for 25 trials ranged from -7.0 to +6.6 and
averaged -0.7 (Table 5.6.1).
Four teams, trained to obtain FEV measure-
ments from school children, collected data simul-
taneously; each team was rotated across all areas.
Each pulmonary function indicator was used a
similar number of times in each community. Age,
sex, and race data were obtained from school
records and verified at the time of testing. On
the day of each test, standing height was
measured and children were queried about the
presence of cold, cough, or sore throat. Each
child was tested until two readings agreed within
10 percent of the child's best performance. The
maximum of the two best results was used in
the analysis.
5-110
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Measurements of total suspended particulates,
suspended nitrates, and suspended sulfates were
obtained daily using a high-volume sampler. Gas
bubblers were used to monitor nitrogen dioxide
by the Jacobs-Hochheiser method and sulfur
dioxide by the West-Gaeke method. Soiling index
was recorded at 2-hour intervals on tape
samplers.^
Two aspects of the location of the moni-
toring sites must be considered in relating these
air quality data to the exposure of these commu-
nities. First, the air monitoring station in the
Intermediate II community was situated on top
of a three-story court house in the center of a
busy commercial area, while those in the Inter-
mediate I and Low communities were located on
two-story buildings surrounded by residential
dwellings. Thus, at the Intermediate II station,
proximity to heavy traffic would tend to increase
measured pollutant levels, while greater height
above ground would tend to decrease these
values. The stations in the Intermediate I and
Low communities probably more accurately re-
flected the ambient exposure experienced by
nearby residents. Secondly, the Intermediate I
community lies about 1 mile west of the John
F. Kennedy International Airport; odors and
noise from the airport have been the object of
many complaints by the residents of that
community. Because of budget limitations, hydro-
carbons, one of the important pollutant classes
Table 5.6.1. CALIBRATION OF PULMONARY FUNCTION
INDICATOR AGAINST STEAD-WELLS SPIROMETER
Pulmonary
function
indicator
1
2
3
4
Overall
average
Trial
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
7
FEV0 75, liters
PFI
3.01
1.40
2.82
2.86
1.74
2.68
3.33
3.41
3.38
3.50
3.52
3.64
3.00
3.00
3.03
3.27
3.24
3.16
3.02
2.83
3.06
2.81
2.51
2.09
2.02
Stead-Wells
2.99
1.44
2.85
2.85
1.82
2.65
3.58
3.53
3.54
3.46
3.63
3.67
3.08
3.04
2.95
3.21
3.11
3.10
3.14
2.93
3.08
2.84
2.44
1.96
2.02
Percent
difference
+0.7
-2.8
-1.1
+0.4
-4.4
+1.1
-7.0
-3.4
-4.5
+1.2
-3.0
-0.8
-2.6
-1.3
+2.7
+1.9
+4.2
+1.9
-3.8
-3.4
-0.6
-1.1
+2.9
+6.6
+0.0
Average
percent
difference
-1.0
-2.9
+1.1
+0.1
-0.7
New York Studies
5-111
-------
emitted by aircraft engines, were not monitored at
any of the CHESS stations during the 1970-1971
study period. Therefore, the influence of aircraft
emissions on the exposure of residents could not be
determined from the pollutant data available.
RESULTS
Population Characteristics
An analysis of socioeconomic characteristics
of the families of children participating in the
study, documented through personal interviews of
a sample of families, showed only slight differ-
ences between study areas:
1. Age - same distribution in all areas.
2. Income - slightly higher in the Interme-
diate I community and lower in the
Intermediate II community when com-
pared with the Low exposure commu-
nity.
3. Education - slightly higher in the Inter-
mediate I community and lower in the
Intermediate II community when com-
pared with the Low exposure com-
munity.
4. Duration of residence - slightly shorter in
the Intermediate I community, but
mobility occurred principally within a
single area.
5. Size of family - somewhat larger in the
Low exposure community.
6. Smoking by parents - no appreciable
difference among communities.
Pollutant Exposures
Estimates of air pollution exposure during
the lifetime of the study population, that is, the
12 years preceding the study, were constructed
from monitoring records of the New York City
Department of Air Resources (Table 5.6.2). These
estimates indicated that New York City children
aged 9 to 13 years at the time of study had
been exposed to high levels of sulfur dioxide
(about 364 to 435 /ig/m3) for the first 5 to 10
years of life. Children 5 to 8 years old were exposed
to these higher levels for a shorter period of time, 5
years or less. Annual suspended sulfate levels were
also high during the decade prior to the study
(estimated at 9 to 25 jug/m3). Both age groups were
exposed to rapidly decreasing pollution levels for 3
years prior to the study, with sulfur dioxide levels
being reduced by about three-fifths and suspended
particulates by one-third. The decrease for suspended
Table 5.6.2. ESTIMATED AIR POLLUTION EXPOSURES OF STUDY COMMUNITIES DURING
YEARS PRIOR TO VENTILATORY FUNCTION TESTING
Pollutant
Sulfur
dioxide.
/ig/m3
Total
suspended
particulates.
Suspended
sulfates.
M9/m3
Dustfall,
g/m2/mo
Community
Low (Riverhead)
Intermediate 1 (Queens)
Intermediate II (Bronx)
Low
Intermediate 1
Intermediate II
Low
Intermediate 1
Intermediate II
Low
Intermediate 1
Intermediate II
Estimated annual average exposure3
1958-1962
NA
424
364
-V35
200
172
NA
25
22
NA
23
20
1963-1967
NA
425
435
^35
158
164
NA
16
16
NA
14
14
1968
NA
233
320
52
92
117
NA
9
12
NA
8
11
1969
NA
131
236
31
75
97
NA
11
10
3
6
10
1970
NA
157
184
39
85
111
NA
5
22
2
6
8
aNA-not available.
5-112
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
sulfates, which averaged an estimated 5 to 22^g/m^,
was not as consistent.
During the period of ventilatory testing, the
Intermediate II area was exposed to marginally
higher levels of suspended particulates and sulfur
dioxide than was the Intermediate I area (Table
5.6.3). As expected, both of these communities
were found to have pollution levels significantly
higher than the Low exposure area, where pollut-
ant levels were substantially below the National
Primary and Secondary Air Quality Standards for
sulfur dioxide and suspended particulates. Concen-
trations in the Low exposure area tended to be
Table 5.6.3. ARITHMETIC AVERAGE POLLU-
TANT CONCENTRATIONS FOR STUDY AREAS,
NOVEMBER 1970 TO MAY 1971
Pollutant
Total
suspended
particulate.
Mg/m3
Suspended
nitrates.
jug/m3
Suspended
sulfates.
/zg/m3
Sulfur
dioxide.
Mg/m3
Soiling,
COHs
Period
Nov-Dec
Jan
Feb-Mar
Apr-May
Avg
Nov-Dec
Jan
Feb-Mar
Apr-May
Avg
Nov-Dec
Jan
Feb-Mar
Apr-May
Avg
Nov-Dec
Jan
Feb-Mar
Apr-May
Avg
Nov-Dec
Jan
Feb-Mar
Apr-May
Avg
Exposure
Low
35
43
29
30
33
1.6
1.8
1.1
1.1
1.3
13
15
8
7
10
27
45
26
25
29
0.38
0.51
0.36
0.199
0.34
Interme-
diate 1
65
69
55
59
61
1.9
1.8
1.6
2.2
1.9
16
17
12
9
13
75
105
39
62
66
0.70
1.05
0.62
0.51
0.68
Interme-
diate II
62
83
81
72
74
1.5
1.9
2.0
2.6
2.0
14
18
13
9
13
95
147
39
59
76
1.02
1.57
1.01
0.91
1.06
about one-half those measured in the Interme-
diate exposure areas. The data suggest that the
Intermediate exposure areas may be considered in
the same pollutant category and that the Low
exposure community represented a much cleaner
area. Concentrations of sulfur dioxide and partic-
ulates were highest in all areas during January.
On the basis of the Jacobs-Hochheiser nitro-
gen dioxide measurements, there was reason to
conclude that exposures to nitrogen dioxide were
similar in the Intermediate communities, and
substantially lower in the Low exposure com-
munity. However, due to the unreliability of the
Jacobs-Hochheiser method, quantitative descrip-
tions of nitrogen dioxide exposure were not
warranted.
Ventilatory Function
Only white children present for all four test-
ing periods were included in data analysis. Black
children were excluded since their sample sizes in
the Intermediate communities were too small to
allow significance testing. Mean height- and age-
adjusted FEVrj 75 values for each of the four
test periods are presented in Figure 5.6.1 for the
2364 white children who participated in all four
tests. These means were obtained from 587
children in the Low exposure area, 835 in the
Intermediate I area, and 942 in the Intermediate
MALES'_
T~
FEMALES
INTERMEDIATE II
Figure 5.6.1. Height- and age-adjusted mean
FEVg.75 for males and females in each com-
munity.
New York Studies
5-113
-------
II area. Month-to-month variations in FEVg 75
values generally followed the same pattern for
both sexes in the three study areas. Pulmonary
function was consistently lowest in all areas in
the February-March test period. The temporal
pattern of FEVg 75 results was unrelated to
variations in monthly average pollutant concen-
trations (Table 5.6.3). However, variations in
monthly means of daily minimum temperatures
(Table 5.6.4) may have influenced the overall
response. In general, ventilatory function was best
in November, December, and April when
temperatures were somewhat higher than the
winter months of January through March. How-
ever, ventilatory function during January, the
coldest month, was better than in February and
March, which were warmer.
Except for the first period of testing, chil-
dren in the Intermediate I area consistently had
the poorest ventilatory performance. During the
first period, lowest pulmonary function was
recorded in Intermediate II area children; per-
formance of children from the Intermediate I
area was significantly higher, and performance of
children from the Low exposure area was highest
but relatively close to that of those from the
Intermediate I area. For the remaining three test
periods, males from the Intermediate II commu-
nity had higher FEVQ 75 readings than males
from the Low exposure area, while the opposite
was true for females.
An analysis of covariance was employed to
test the significance of identified important de-
terminants of FEVQ 75, including sex, age,
height, and area of residence. Significance tests
were computed for each test period and for all
periods combined. These analyses showed strong
effects of sex, age, and height on FEVg 75 of
school children. A consistent FEVQ 75 difference
between the Low exposure community and the
combined Intermediate 1 and Intermediate II
communities appeared in each of the four test
periods. This difference was explained mostly by
the low performance of children of both sexes
living in the Intermediate I exposure area. Except
for the first of the four test periods, the per-
formance of children from the Intermediate II
and Low exposure communities did not signifi-
cantly differ when FEVQ 75 values were averaged
across both sexes.
Since the cumulative dose of air pollution
exposure of younger children (5 to 8 years old)
was less than that of older children (9 to 13
years old), separate analyses were performed,
using the previously described procedure, to test
the hypothesis that older children from the
polluted areas would demonstrate a greater decre-
ment in ventilatory function than younger chil-
dren from those areas. Sex- and age group-
specific height- and age-adjusted FEVQ 7 5 for
each exposure area and each time period were
calculated (Table 5.6.5). No consistent inter-
community differences in ventilatory function
were found for younger children. There was a
modest but steady improvement over time in
ventilatory function among young males living in
the Intermediate II exposure area. A similar
trend, considerably dampened, was suggested for
young females from the same community.
On the other hand, when all testing periods
were grouped together, older children of both
sexes from each of the Intermediate communities
had lower ventilatory function than children in
the Low exposure community. Decrements were
modest, ranging from 2 to 3.4 percent, and were
significant only for males (Table 5.6.6). When
each testing period was considered separately,
significant decrements in ventilatory function
were often found for older children residing in
the Intermediate communities. On two occasions,
a single sex group of older children from one of
Table 5.6.4. MONTHLY AVERAGE DAILY MINIMUM TEMPERA-
TURE (°F) FOR EACH STUDY AREA DURING PULMONARY
FUNCTION TESTING
Community
Low
Intermediate 1
Intermediate II
Nov
40.4
42.4
44.1
Dec
29.1
29.5
31.2
Jan
20.4
21.7
23.1
Feb
26.5
28.5
30.9
Mar
31.7
32.8
34.0
Apr
36.5
39.7
41.6
5-114
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 5.6.5. SEX- AND AGE GROUP-SPECIFIC AGE- AND
HEIGHT-ADJUSTED MEAN FORCED EXPIRATORY VOLUME
(FEV0 75) DISTRIBUTED BY TEST PERIOD AND COMMUNITY
Test period
and community
Nov-Dec
Low
Intermediate 1
Intermediate II
Jan
Low
Intermediate 1
Intermediate II
Feb-Mar
Low
Intermediate 1
Intermediate II
Apr
Low
Intermediate 1
Intermediate II
Average
Low
Intermediate 1
Intermediate II
(Sample size)
Height-adjusted FEVg 75, liters
Males
5 to 8
yr old
1.204
1.171
1.193
1.178
1.196
1.203
1.141
1.134
1.203
1.210
1.191
1.243
1.183
1.173
1.211
(645)
9 to 13
yr old
1.821
1.792
1.726
1.776
1.708
1.793
1.703
1.649
1.682
1.747
1.660
1.718
1.762
1.702
1.730
(582)
Females
5 to 8
yr old
1.222
1.082
1.065
1.087
1.091
1.091
1.049
1.031
1.066
1.093
1.086
1.106
1.088
1.072
1.082
(596)
9 to 13
yr old
1.698
1.738
1.647
1.701
1.602
1.690
1.643
1.604
1.593
1.689
1.621
1.656
1.683
1.641
1.647
(541)
the Intermediate communities had ventilatory
performance that was higher than in the Low
exposure community. On one other occasion, the
decrements for older children observed in the
Intermediate communities were not significant.
As expected, height and age differences were
both strong determinants of ventilatory function
for younger and older children. The regression
coefficients for age and height by sex and age
group are given in Table 5.6.7 since they may
be of value to other investigators.
DISCUSSION
White children aged 9 to 13 years living in
fairly polluted (Intermediate) communities were
found to have decreased ventilatory function
when compared to similar children from a rela-
tively clean (Low) community. For males, these
decrements were statistically significant; for fe-
males they were not. These decrements were
replicated for both sexes over time, with only
two ' of ten possible community comparisons
yielding anomalous trends. Younger children (5
to 8 years old) from polluted communities, who
were relatively less exposed than their older (9
to 13) classmates, did not exhibit any decrement
in ventilatory function that was attributable to
air pollution. In fact, in young males from the
area with the highest recent pollution exposures,
a trend towards improving ventilatory function
was seen during the school year. A less marked
but perceptible improvement trend was noted for
females from the same community.
New York Studies
5-115
-------
Table 5.6.6. ANALYSIS OF VARIANCE FOR EACH TEST PERIOD BY SEX AND BY AGE
GROUP (F VALUES)
Factor
Males 5 to 8 years
Age
Height
Area
Ethnic differences
Females 5 to 8 years
Age
Height
Area
Ethnic differences
Males 9 to 13 years
Age
Height
Area
Ethnic differences
Females 9 to 13 years
Age
Height
Area
Ethnic differences
Degrees
of
freedom
1
1
2
1
1
1
2
1
1
1
2
1
1
1
2
1
Nov-Dec
36.1 1a
200. 19a
1.81
2.37
37.64a
168.363
4.58b
0.01
7.23a
324.73a
5.89a
0.04
1.44
348.42a
6.60a
0.71
Jan
9.82a
213.483
0.80
2.00
1 1 .76a
118.563
0.03
0.21
3.38
308.023
14.73a
0.78
4.65b
314.123
8.47a
0.04
Feb-Mar
5.46b
256.59a
1 1 .04a
1.83
17.35a
167.47a
2.27
0.01
1.55
289. 17a
4.62a
0.10
0.01
307. 17a
2.80
1.38
Apr
47.033
180.233
7.00a
0.43
28.313
156.563
0.50
1.85
3.26
277.913
5.24a
0.43
0.57
403.883
3.55b
0.40
Average
34.88a
339.313
5.23a
1.22
37.77a
240.833
0.48
0.10
4.87b
412.263
3.71b
0.24
1.40
465. 10a
2.41
0.70
Significant at 0.01 level.
Significant at 0.05 level.
Table 5.6.7. REGRESSION COEFFICIENTS FOR EFFECTS OF AGE (in liters/yr)
AND HEIGHT (in liters/in.) ON FEVQ 75 BY SEX AND BY AGE GROUP
Factor
Males 5 to 8 years
Age
Height
Females 5 to 8 years
Age
Height
Males 9 to 13 years
Age
Height
Females 9 to 13 years
Age
Height
Nov-Dec
0.0549
0.0433
0.0540
0.0399
0.0347
0.0624
0.0159
0.0667
Jan
0.0306
0.0479
0.0356
0.0395
0.0233
0.0598
0.0286
0.0633
Feb-Mar
0.0211
0.0486
0.0383
0.0416
0.0159
0.0582
0.0010
0.0647
Apr
0.0725
0.0476
0.0578
0.0475
0.0240
0.0595
0.0098
0.0699
Average
0.0448
0.0469
0.0465
0.0422
0.0242
0.0600
0.0135
0.0661
5-116
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
For the elementary school population sampled
in this study, early childhood exposure to ele-
vated air pollution levels lasting 5 to 10 years
appears to have been accompanied by a decre-
ment in ventilatory function. The study also
suggests that improvement in ambient air quality
may be followed by improved ventilatory func-
tion. The study could identify no effect of
month-to-month changes in air quality on month-
to-month changes in ventilatory function. These
changes in ventilatory function may well be
attributable in part to a physiologic seasonal vari-
ation. The effect of short-term pollution expo-
sures in the present study may have been ob-
scured by a temperature effect, with lower
temperatures adversely affecting ventilatory func-
tion.
The influence of nine factors must be con-
sidered before assessing the implications of the
present study for air quality standards: (1) socio-
economic differences, (2) ethnic factors, (3)
migration, (4) urban heat island, (5) indoor air
pollutants, (6) ambient air pollutants other than
those measured, (7) learning, (8) current or
recent acute respiratory disease, and (9) the
machinery used in testing.
Socioeconomic differences tend to cancel
themselves in the present study, and thus could
hardly account for the observed differences in
lung function. One of the Intermediate commu-
nities ranked socioeconomically higher and the
other lower than the Low community. Social
differences in the study areas were also quite
modest when compared to those encountered in
mortality studies of the effects of air pollution
on health.
Ethnic differences could have biased the
study had not the experience of blacks been
excluded from the analyses. Blacks have been
shown to have lower measured ventilatory per-
formance than whites of comparable age, sex,
and height.10"13 Too few black children from
the Intermediate areas were available to achieve
the sample size needed for significance testing.
Intercommunity differences in ventilatory per-
formance attributable to differences in the pro-
portions of Jewish or Italian surnames were also
sought, but none were found (Table 5.6.6).
Migration effects, while not specifically con-
sidered, would probably tend to dampen rather
than accentuate the observed differences between
the Intermediate and Low communities. Children
moving to more polluted areas from clean areas
are more likely to have better ventilatory func-
tion than life-long residents of polluted areas.
Similarly, children moving from polluted areas
would tend to bring their impaired ventilatory
function to the relatively clean communities.
Data from other CHESS studies indicated that up
to 25 percent of families in the study commu-
nities had made such moves across pollution
categories.
Lower ambient temperatures seemed generally
to be associated with decrements in ventilatory
function. However, area temperature differences
would also probably tend to dampen rather than
accentuate the observed differences attributable to
air pollution. The Intermediate communities were
integral parts of a compact urban heat island,
and winter temperatures were somewhat warmer
in New York City than on Long Island. This
fact might tend to improve ventilatory perform-
ance in the Intermediate communities relative to
the Low community.
The effects of indoor air pollutants from
cigarette smoking or from domestic use of gas
for cooking or space heating could not be quan-
titated. Though the smoking habits of elementary
school children were not documented, there was
no particular reason to suspect that their smok-
ing habits would be very different in the Inter-
mediate and Low pollution areas. The prevalence
of cigarette smoking among children is probably
most strongly influenced by parental smoking
habits and by socioeconomic status. In this
study, parental smoking habits were documented
to be very similar in all three communities. Also,
because the Low pollution community was socio-
economically intermediate between the Intermediate
I and II communities, differences in smoking habits
between Intermediate and Low exposure areas
probably would have tended to cancel in this study.
Environmental pollutants other than those
measured might also contribute to the observed
decrements in ventilatory function. One environ-
mental factor that stands out, distinguishing the
Intermediate I exposure community from the
other study areas, is its proximity to one of the
country's busiest international airports. We were
unable to monitor air pollutants peculiar to air-
craft emissions. Even with the best existing
instrumentation, it would be difficult to monitor
the hydrocarbons emitted by aircraft but not by
automobile traffic and therefore to specify which
pollutants are unique to the area.
New York Studies
5-117
-------
Ventilatory performance, like other perform-
ance testing, is likely to be enhanced by learn-
ing. One would have to postulate significant area
differences in the effects of learning to account
for the present results, and this would not seem
justified. In pulmonary function studies in all
CHESS areas, investigators have been consistently
impressed by the ease with which even the
youngest children learn the FEV procedure.
However, there is little doubt that older children
grasp the procedure more thoroughly than
younger ones do. Thus the results obtained from
older children are probably somewhat more reliable
than those from younger children.
Concurrent or recent acute respiratory disease
might adversely affect ventilatory performance.
Excesses in such disorders did in fact occur in the
Intermediate communities.^4 However, the study
design minimized the effects of acute respiratory
illnesS by excluding symptomatic children from the
analysis. One point should be emphasized here: air
pollution seems to have played an important role in
the excess respiratory disease experience of children
in the more polluted communities. Thus, though
impaired ventilatory function may be partly attrib-
utable to respiratory illness history, this history may
in turn be partly attributable to air pollution expo-
sure.
In CHESS studies subsequent to the one
reported here, the investigators became aware of
a tendency of the pulmonary function indicator
to drift a maximum of about 350 ml in the
course of a day. The drift was usually downward
but occasionally upward. In the present study,
the presence and magnitude of drift were never
determined. There is no reason to suspect, how-
ever, that instrument drift, if present, would have
altered one community's results by a different
magnitude or in a different direction from
another's. (The National Cylinder Gas Pulmonary
Function Indicator was replaced in CHESS
pulmonary function studies by a waterless dry-
seal volume spirometer in 1972.)
In summary, combined covariate effects are
likely to have dampened rather than accentuated
true differences between the Intermediate and
Low communities, or not to have affected these
differences at all. We concluded that 9 or more
years exposure to annual sulfur dioxide levels of
an estimated 131 to 435 jUg/m-^ accompanied by
suspended participate levels of about 75 to 200
/zg/m-' and suspended sulfate levels of about 5 to
25 jiig/m-' can be associated with a small but
significant impairment in ventilatory function that
persists even as Primary Ambient Air Quality
Standards for these pollutants are achieved.
Furthermore, though it is unlikely, one cannot
be certain that lower annual levels of sulfur
dioxide (66 to 76 ug/m^) accompanied by lower
annual levels of suspended sulfates (13 ^ig/m^)
and suspended particulates (61 to 74 jug/m^) do
not facilitate the persistence of previously in-
duced impairments in ventilatory function. Con-
firmation ' of the adequacy of present air quality
standards must await the passage of time and the
•necessary retesting of ventilatory function.
Finally, time may well prove that more attention
should be devoted to the effects of suspended
sulfates.
SUMMARY
Ventilatory function (FEVg 75) tests were
performed by 2364 New York metropolitan
elementary school children four times during the
1970-1971 school year. Older white children,
aged 9 to 13 years, from a relatively unpolluted
suburban fringe community on Long Island were
found to have significantly better ventilatory
function than children from two more polluted
urban communities. This difference existed in the
face of recent improvements in air quality. Older
children exposed for more than 8 years to annual
average sulfur dioxide levels estimated at 131 to 435
jug/m3, accompanied by levels of suspended par-
ticulates estimated at 75 to 200 ng/m^ and suspended
sulfates estimated at 5 to 25 Mg/m^ showed signifi-
cant decreases in FEVg 75, but younger children, age
5 to 8 years, did not. The effects of age, height,
socioeconomic differences, ethnic factors, migration,
ambient temperature, exposure to indoor air pollu-
tants and to pollutants not monitored, and con-
current acute respiratory disease were all considered.
REFERENCES FOR SECTION 5.6
1. Watanabe, H., F. Kaneko, H. Murayama, S.
Yamaoka, and T. Kawaraya. Effects of Air
Pollution on Health; Report No. 1: Peak
Flow Rate and Vital Capacity of Primary
School Children. Reports of the Osaka City
Institute of Hygiene. 26:32-37, 1964.
2. Toyama, T. Air Pollution and Its Health
Effects in Japan. Arch. Environ. Health.
5:153-173, 1964.
5-118
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
3. Holland, W.W., T. Halil, A.E. Bennett, and
A. Elliott. Factors Influencing the Onset of
Chronic Respiratory Disease. Brit. Med. J.
2:205-208, April 1969.
4. Lunn, J.E., J. Knowelden, and A.J. Handy-
side. Patterns of Respiratory Illness in Shef-
field Infant School Children. Brit. J. Prevent.
Soc. Med. 27:7-16, 1967.
5. Petrilli, R.L., G. Agnese, and S. Kanitz.
Epidemiology Studies of Air Pollution Effects
in Genoa, Italy. Arch. Environ. Health.
72:733-740, 1966.
6. Anderson, D.O. and C. Kinnis. An' Epi-
demiologic Assessment of a Pediatric Peak
Flowmeter. Amer. Rev. Resp. Dis. 95:73-80,
1967.
New York Metropolitan Communities,
1944-1971. In: Health Consequences of
Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. Publication
No. EPA-650/1-74-004. 1974.
10. Wilson, M.G. and D.J.
Value of Determining
Lungs of Children. J.
75:1107-1110, 1922.
Edwards. Diagnostic
Vital Capacity of
Amer. Med. Assoc.
11. Smillie, W.G. and D.L. Augustine. Vital
Capacity of Negro Race. J. Amer. Med.
Assoc. 57:2055-2058, 1926.
12. Roberts, F.L. and J.A. Crabtree. The Vital
Capacity of the Negro Child. J. Amer. Med.
Assoc. 55:1950-1954, 1927.
7. Higgins, I.T.T., P.O. Oldham, A.L. Cochrane,
and J.C. Gilson. Ventilatory Function in
Miners: A Five Year Follow-up Study. Brit.
J. Ind. Me.d. 79:65-76, 1962.
13. Damon, A. Negro-white Differences in Pul-
monary , Function (Vital Capacity, Timed
Vital Capacity and Expiratory- Flow Rate).
Human Biol. 35:380-393, 1966.
8. Higgins, I.T.T., M.W. Higgins, J.C. Bilson, H.
Campbell, W.E. Walters, and E.G. Ferris.
Smoking and Chronic Respiratory Disease:
Findings in Surveys Carried Out in 1957 and
1966 in Stavely, in Derbyshire, England.
Chest. 59(5, Suppl.):345-355, 1971.
9. English, T.D., W.B. Steen, R.G. Ireson, P.B.
Ramsey, R.M. Burton, and L.T. Heiderscheit.
Human Exposure to Air Pollution in Selected
14. Love, G.J., A.A. Cohen, J.F. Finklea, J.G.
French, G.R. Lowrimore, W.C. Nelson, and
P.B. Ramsey. Prospective Surveys, of Acute
Respiratory Disease in Volunteer Families,
1970-1971 New York Studies. In: Health Con-
sequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C. Publi-
cation No. EPA-650/1-74-004. 1974.
New York Studies
5-119
-------
CHAPTER 6
CINCINNATI STUDY
6-1
-------
6.1 VENTILATORY FUNCTION IN SCHOOL CHILDREN:
1967-1968 TESTING IN CINCINNATI NEIGHBORHOODS
Carl M. Shy, M.D., Dr. P.H., Cornelius J. Nelson, M.S.,
Ferris B. Benson,B. A., Wilson B. Riggan, Ph.D.,
Vaun A. Newill, M.D., S.M.Hyg.,
and Robert S. Chapman, M.D.
6-3
-------
INTRODUCTION
Prolonged exposure to air pollution in the early
years of life can produce adverse changes in pulmo-
nary function of children. Holland et al.1 detected
reduced function in urban children attending schools
in the more polluted areas of Kent, England. Children
residing in a heavily polluted metropolitan area of
Japan had significantly lower forced-expiratory-
volume measurements than children in a rural area;
these differences were greatest during winter months
when urban pollutant levels were highest.2 Similar
results have not been reported in United States cities
having only moderately elevated concentrations of
particulate and sulfur oxide pollution. The purpose of
this study was to investigate the influence of
temporal and geographic variation in urban air pol-
lution on ventilatory performance of school children
living in a moderately large metropolitan area of the
United States. Possible effects of both acute and
chronic exposures were sought.
MATERIALS AND METHODS
The study was designed to include a pair of public
elementary schools in each of six neighborhoods
differing in socioeconomic level, race, or pollution
exposure, as shown in Table 6.1.1. The six schools in
Table 6.1.1. COMPOSITION OF STUDY SECTORS
ACCORDING TO SOCIOECONOMIC-RACIAL
AND EXPOSURE CATEGORIES
Socioeconomic
and racial
category
Upper-middle
white
Lower-middle
white
Lower-middle
black
Exposure category
Clean
School 1
School 2
School 5
School 6
School 9
School 10
Polluted
School 3
School 4a
School 7
School 8b
School 1 1
School 12b
This school was reassigned to the clean exposure category on
the basis of pollutant data obtained during the 8-month
study.
Schools 8 and 12 are, respectively, the white and black por-
tions of the same school.
polluted neighborhoods were located in the main
industrial valley of central metropolitan Cincinnati
and were matched for socioeconomic and racial
characteristics with six schools in a nonindustrial river
valley on the east side of the metropolitan area.
Visual estimates of home market values were used to
gauge the general socioeconomic level of the
neighborhood. These estimates were substantiated by
questionnaires administered to households of par-
ticipating children at the end of the study. One
school had approximately equal numbers of black
and white children and is listed twice in Table 6.1.1
as Schools 8 and 12. All children in one or two
classrooms of the second grade of the elementary
schools were asked to participate in the study to
achieve sample sizes of 60 to 75 children in each of
the six study sectors.
Ventilatory performance, as measured by a
0.75-second forced-expiratory volume (FEV0 7S) on
a Stead-Wells spirometer, was obtained 12 times from
each child: once weekly in the months of November
1967 and February and May 1968. Tests were
administered on Tuesday and Wednesday mornings
by six teams, which were systematically rotated
among schools. The best of two satisfactory forced-
expiratory maneuvers obtained on each test day was
used for all computations. Criteria for a "satis-
factory" maneuver were: (1) full inspiration before
exhaling, (2) maximum, uninterrupted exhalation,
and (3) no leakage of air from around the mouth-
piece. Each child was asked whether he currently had
a cold, cough, or sore throat on the test day. Standing
heights were measured to the lower half inch once
monthly, and a linear regression of FEV0 75 on
height was computed to adjust for area differences in
height of children.
In the final month of study, experienced Census
Bureau interviewers visited homes and administered a
health questionnaire to the mother of each par-
ticipating child. Questions were asked about the
child's past history of illness, parental income, educa-
tion, occupation, and duration of residence at current
address.
Air monitoring stations were placed in locations
within three blocks of each school to provide samples
representative of the air quality in the neighborhood
served by the school. Twenty-four-hour integrated
samples of total suspended particulates, suspended
sulfates and nitrates, and sulfur dioxide were col-
lected daily during the 8 months of the study at the
monitoring station adjacent to each school. At one
school in each of the six study sectors, both outdoor
6-4
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
and indoor soiling index and sulfur dioxide were
monitored. During the 12 hours preceding pulmonary
function testing, 4-hour sequential samples of soiling
index and sulfur dioxide were obtained at these
schools. Daily measurements of 24-hour total sus-
pended particulates, suspended sulfates, and sus-
pended nitrates were collected with high-volume air
samplers designed to capture particles less than 90
microns in diameter.3 Soiling-index samples were
collected on a sequential tape sampler with a 4-hour
sampling cycle at a flow rate of 15 cubic feet per
hour.4 Sulfur dioxide was measured by bubbling air
through a solution of tetrachloromercurate and
analyzed by the colorimetric method of West and
Gaeke.s
RESULTS
Pollutant concentrations at School 4 (polluted,
upper-middle class, white sector) were consistently
lower than expected and were similar to schools in
the clean areas of the study. School 4 was therefore
reassigned to the clean, upper-middle class, white
sector, leaving only one school in the polluted
category of this socioeconomic level.
Concentrations of the various pollutants averaged
over the 7 months of study are presented for each
study sector in Figure 6.1.1 and for each school in
Table 6.1.2. Seasonal averages for total suspended
120
100
80
60
40
20
0
3.5
3.0
2.5
2.0
1.5
10
0
_ 77
-114-
96
82
96 _
78
0.0(- 9.4
8.3
8.0-
6.0-
4.0
2.0-
9.5
8.3
8.8 8.9
_ 2.9
h- 2.4
CP CP CP " CP CP CP
UMW LMW LMB UMW LMW LMB
TOTAL SUSPENDED PARTICULATES SUSPENDED SULFATES
60i
1.2
2.4
2.6
2.7
C P C P C P
UMW LMW LMB
SUSPENDED NITRATES
_ J139.2
50J
40.3
44.5«4
C P C P C P
UMW LMW LMB
$02 OUTDOORS
UMW - UPPER-MIDDLE WHITE
C - CLEAN LMW - LOWER-MIDDLE WHITE
P - POLLUTED LMB - LOWER-MIDDLE BLACK
Figure 6.1.1. Pollutant concentration, arithmetic averages over entire 7
months of study withfn~each study sector, Cincinnati, 1967-1968.
Cincinnati Study
6-5
-------
Table 6.1.2. ARITHMETIC AVERAGES OF POLLUTANT CONCENTRATIONS MEASURED
OVER ENTIRE 7 MONTHS OF STUDY FOR EACH SCHOOL,
OCTOBER 24, 1967 to MAY 22, 1968
(n = 211 sampling days)
Stratum
Clean UMW
Polluted UMW
Clean LMW
Polluted LMW
Clean LMB
Polluted LMB
School
1
2
4
3
5
6
7
8
9
10
11
12
Suspended particulates,
Mg/m3
Total
61
85
85
96
88
76
133
96
78
78
97
96
Sulfate
8.2
7.7
9.1
9.4
8.2
8.4
10.1
8.9
8.9
8.7
8.9
8.9
Nitrate
2.2
2.4
2.7
2.9
2.4
2.3
3.6
2.8
2.6
2.6
2.6
2.8
Outdoor
Soiling,
COH
0.87
1.05
0.90
1.35
0.98
1.04
S02,
jug/m3
37
39
47
39
44
37
57
44
50
39
47
44
Indoor
Soiling,
COH
0.72
0.77
0.92
1.25
0.85
0.91
S02,
Aig/m3
10
5
16
15
21
16
6
particulates ranged from 61 to 92 /zg/m3 in clean
neighborhoods, and from 76 to 131 jug/m3 in
polluted neighborhoods, a relative difference of 33
percent. During the winter months (December-
February) suspended particulate concentrations
increased by 25 to 30 percent in all areas (Table
6.1.3), but relative differences between clean and
polluted neighborhoods remained the same.6
A total of 394 second-grade children, representing
93 percent of enrollment in selected classrooms,
participated in the study during all 3 months of
pulmonary function testing. As shown in Table 6.1.4,
sample sizes ranged from 44 to 99 children in the
various study sectors.
Within the same socioeconomic category, residents
of clean and polluted neighborhoods were similarly
distributed with respect to educational achievement
of the head of household (Figure 6.1.2). Differences
between socioeconomic categories were large. Income
and occupation of father, market value of home,
monthly rent and number of persons per bedroom
showed the same area distribution patterns as educa-
tion.
Analysis of FEV0 7S results was performed only
on students who participated in each of the 3 months
of pulmonary function testing. From 5 to 13 percent
of children in the various study sectors had a history
Table 6.1.3. SEASONAL VARIATION OF
TOTAL SUSPENDED PARTICULATE AND
OUTDOOR 24-HOUR AVERAGE SULFUR
DIOXIDE WITHIN EXPOSURE SECTORS
(/ig/m3)
Exposure category
aeasun ana
socioeconomic-
racial category
Fall (Oct.-Nov.)
UMW
LMW
LMB
Winter (Dec.-Feb.)
UMW
LMW
LMB
Spring (Mar.-May)
UMW
LMW
LMB
Clean
TSP
62
66
60
86
92
86
75
82
80
S02
29
29
29
58
60
66'
26
26
31
Polluted
TSP
79
88
76
107
131
106
92
111
98
S02
24
39
37
58
73
68
29
34
29
of asthma or bronchitis; these children, numbering 46
in all, were eliminated from the analysis, reducing the
total sample size to 348.
Weekly FEV0 75 values for each sector were
adjusted to a standing height of 50 inches by means
of the equation:
6-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
adjusted FEVQ 75 = average FEV0 7S +
b(50 - average height)
A sex-specific regression coefficient, b, was calculated
by fitting a least-squares-regression equation to
observed FEV0 75 values plotted against standing
height for each sex. Sex-specific, height-adjusted
FEV0 7S values were thus computed for each week.
A 'combined value for both sexes was derived by
giving equal weight to values for males and females.
These combined values are presented in Figure 6.1.3
for each of the 12 weeks of testing. Relatively low
FEV0 7S values were independently associated with
three factors: (1) black children, (2) residence in
polluted area, and (3) the month of February. Among
white children, the contrast between FEV0 7 s values
of children in polluted and clean neighborhoods was
greatest during February. FEV0 7S values of black
children in polluted areas were not consistently
different than values of black children in clean areas.
Table 6.1.4. NUMBER OF CHILDREN TESTED
DURING ALL 3 MONTHS OF STUDY
WITHIN EACH SECTOR
Socioeconomic-
racial category
Upper-middle
white
Lower-middle
white
Lower-middle
black
Total
Exposure category
Clean
99
67
62
228
Polluted
44
53
69
166
Total
143
120
131
394
A multivariate analysis of variance model was used
to test the significance of race, socioeconomic
contrast between white children, residence in pol-
luted or clean area, standing height, and sex on
FEV0 75 for the 3 test months (Table 6.1.5). This
analysis confirmed the highly significant (p < 0.01)
effects of height, sex, and race on FEV0 7 s. Among
white children, higher socioeconomic level was as-
sociated with higher FEV0 75 performance. High
suspended particulate levels were consistently as-
sociated with significantly lower FEVQ 75 results.
Among pollutants, however, suspended sulfates
exerted the strongest influence on lung function of
the children. Suspended nitrates also exerted a
significant effect on FEV0 7S. The relative effects of
race, sex, and suspended sulfate and nitrate exposure
on FEV0 7S performance are illustrated in Figure
6.1.4. FEV0 75 values have been adjusted for dif-
ferences in all factors other than those contrasted in
that set. Of the four factors, suspended sulfate
exposure had the largest independent effect on lung
function. On the average, adjusted FEV0 7S results
for children in polluted areas were 17.4 percent
below those for children in clean areas. Lung function
of black children was 9.3 percent below that of white
children, while results for girls were 6.7 percent
below those for boys. The pollution effect was
somewhat diminished by the fact that area dif-
ferences in pollution had no effect on the perform-
ance of black children.
Indoor and outdoor sulfur dioxide, soiling index,
and suspended particulate levels measured over the
24-hour or 4-hour period directly preceding pulmo-
nary function tests did not consistently correlate with
FEVo.75 values. Ventilatory performance of children
thus did not appear to be acutely affected by
variations in pollutant levels on the day of the test.
100
80
i-60
LU
o
£40
o_
20
0
HIGH SCHOOL^
NOT
COMPLETED
.5
•1M13.7
COMPLETED HIGH SCHOOL
19.220.0
COLLEGE EDUCATION
2.5
C P
UNIW
C P
LHIW
C P
LK1B
C P
C P
LMW
C P
LMB
C P
UMW
C P
LMW
C P
1MB
Figure 6.1.2.
1967-1968.
Educational achievement of fathers within each study sector, Cincinnati,
Cincinnati Study
6-7
-------
OCLEANUMW
D POLLUTED UB1W
A CLEAN LMW
• POLLUTED LMW
• CLEAN LMB
APOLLUTED LMB
NOVEMBER
1967
FEBRUARY
1968
TESTING PERIODS, weeks
Figure 6.1.3. Height-adjusted FEVo.75 during each of 12 weeks of testing for each study
sector, Cincinnati, 1967-1968.
FEV0 75 results obtained from students admitting
to a cold, cough, or sore throat were compared with
results from the same children when they denied
these symptoms. No consistent difference in
FEV0 7S values, as tested by paired t-tests, was
found.
Average concentrations of total suspended particu-
lates in Cincinnati are available from the National Air
Surveillance Network station over the period 1957 to
1968.7'13 These years cover more than the life span
of the 7- to 9-year-old children participating in this
study. At the National Air Surveillance Network
6-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 6.1.5. EFFECTS OF AIR POLLUTION, SOCIOECONOMIC-RACIAL CATEGORY,
SEX, STANDING HEIGHT, AND MONTH OF TEST ON FEV0 75:
TESTS OF SIGNIFICANCE BY MULTIVARIATE ANALYSIS OF VARIANCE
Factor
Suspended particulates
Sulfates
Nitrates
Total particulates
UMW versus LMW
Race
Sex
Standing height
" Overall mean FEV0.7s
Regression coefficients8
Nov.
-0.0145
-0.0045
-0.0000
0.0265
-0.0702
0.0300
0.0553
1.1592
Feb.
-0.0621
-0.0626
-0.0008
0.0314
-0.0687
0.0399
0.0509
1.1500
May
-0.0511
-0.0420
-0.0005
0.0548
-0.0741
0.0388
0.0528
1.1906
Significance test
F
8.79
3.56
1.08
6.07
11.84
6.84
66.56
P
<0.01
<0.05
NSb
<0.01
<0.01
<0.01
<0.01
aRegression coefficients for socioeconomic level, race, sex, standing height, and mean FEVg 75 values are
given from the multivariate analysis using sulfates alone as the pollution variable. Regression coefficients
changed very little when the analysis was reported for the other pollution variables.
bNS—not significant.
SOCIOECONOMIC- UIWW
RACIAL LMW
CATEGORY LMB
SEX
MALE
FEMALE
SUSPENDED C
SULFATE p
SUSPENDED
NITRATE
258
7//////////////////////AUVI
1.00 1.05 1.10 1.15 1.20 1.25 1.30
HEIGHT-ADJUSTED MEAN FEVQ 75, liters
Figure 6.1.4. Relative influence of socioeconomic-racial difference, sex,
and air pollution on average FEVrj.75 performance of school children,
Cincinnati, 1967-1968.
Cincinnati Study
6-9
-------
station, total suspended particulate concentrations
averaged 125 Mg/m3 (geometric mean) over the 1957
to 1961 period, representing the infant years of
participating children. During early childhood years,
from 1962 to 1966, these concentrations averaged
129 Mg/m3 (geometric mean). During the years of this
study (1967-1968), however, total suspended particu-
late concentrations had fallen to an average of 105
jug/m3 (geometric mean). Thus, suspended particulate
concentrations at the National Air Surveil-
lance Network station were approximately 20 percent
higher during the early childhood years than during
the years of this study. If diminution in pulmonary
function were a response to early childhood
exposures rather than to recent pollution exposures,
the observed effect on ventilatory performance would
be related to total suspended particulate levels
estimated to be approximately 20 percent higher than
those reported during the study in high exposure
neighborhoods. That is, total suspended particulate
concentrations are estimated to have been in the
range of 115 to 137 jug/m3 (arithmetic mean) during
early childhood years of the participants from high
exposure neighborhoods. A similar time trend for
suspended sulfates in high exposure neighborhoods
would raise the estimate of exposure associated with
diminished lung function in neighborhoods of white
children from observed values between 8.9 and 10.1
/jg/m3 to values between 10.7 and 12.1 |itg/m3. For
suspended nitrates, estimates of potentially detri-
mental exposure would be raised from 2.8 and 3.6
/zg/m3 to 3.4 and 4.3 jUg/m3. These worst-case and
least-case estimates are given for pollutant exposures
most strongly associated with diminished lung func-
tion of the children participating in this study.
DISCUSSION '
This study showed significant area differences in
ventilatory performance of second-grade school
children. Area differences were associated with both
race and chronic exposure to moderate levels of
suspended sulfates and nitrates in the air.
Densen and his coworkers14'15 also noted lower
pulmonary function and lower smoking-specific
chronic respiratory symptom rates among adult black
New York City postal workers compared with whites
in the same occupation. The strong racial difference
in ventilatory function among school children has
also been previously reported.16"19 The ventilatory
performance of black children was variously reported
as 7 to 15 percent18'19 below that of white children,
and differences persisted when similar socioeconomic
groups of the two races were compared17 or when
adjustments were made for differences in stem height
and chest expansion.18 One possible explanation for
the relatively low FEV0 7S values of the black
children in our study was their response, conditioned
by racial attitudes prevalent in their early years, to
the all-white teams administering the test.
Of the three measured components of suspended
particulate matter (total, sulfates, and nitrates), sus-
pended sulfate levels had the strongest quantitative
association with area differences in FEV0 75 (Figure
6.1.4). Area differences in suspended sulfates be-
tween clean and polluted neighborhoods were largest
among UMW and LMW sectors, but almost non-
existent among LMB sectors (Figure 6.1.1). The same
gradients of exposure existed for suspended nitrates.
Decrements in FEVo.vs performance paralleled the
gradients of exposure to suspended sulfates and
nitrates. Clean and polluted black sectors, however,
showed relatively large differences in total suspended
particulates—a difference not reflected in FEV0 7S
performance. Thus, for establishing appropriate dose-
response relationships, quantitative differences in
pollutant exposure of UMW and LMW study sectors
are the only useful considerations. That is, LMB areas
showed no area difference in lung function that was
consistent with differences in pollution exposure.
Mean pollutant exposures in white sectors with low
and high exposure, respectively, were as follows
(values given are arithmetic means in clean versus
polluted sectors): 80 vs 105 £ig/m3 for total sus-
pended particulates, 8.3 vs 9.5 Aig/m3 for suspended
sulfates, and 2.4 vs 3.1 fxg/m3 for suspended nitrates
(Figure 6.1.1).
These levels were probably about 20 percent
higher during 7 to 9 years of exposure preceding this
study for children who were lifetime residents of
Cincinnati. The fact that lung function of children in
high-exposure neighborhoods improved in spring
when air quality also improved suggests that current
pollutant exposures, in addition to chronic exposures,
may exert an effect on ventilatory performance.
Unfortunately, little information on lung function
values is as yet available to test this hypothesis. The
springtime improvement in FEV0 75 may also, in
part, reflect a physiologic seasonal variation in
ventilatory function. The results of this study support
the hypothesis that chronic exposure to these
relatively low levels of suspended particulates was
associated with significantly lower ventilatory
performance of second-grade school children. These
effects occurred in the presence of low concentra-
tions of sulfur dioxide (less than 52 Mg/m3) in all
areas and constitute one of the few cases where
health effects can be attributed to particulate pollu-
tion independently of atmospheric levels of gaseous
sulfur dioxide.
6-10
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
As shown in Figure 6.1.3, the magnitude of
differences in ventilatory performance of children in
clean and polluted schools in the two white socio-
economic categories was greatest during February,
when both low outdoor temperatures and high
concentrations of air pollutants were experienced.
Cold temperatures alone did not significantly affect
the FEV0 75 results of children in the clean areas,
and ventilatory performance of children was not
affected by acute exposure to pollutants on the day
of the test. An interaction between cold weather and
elevated pollutant concentrations in polluted
neighborhoods alone is a possible explanation for
these results. The data in Figure 6.1.3 suggest that the
impairment of ventilatory performance among
children of this age group was not irreversible, for
during May, FEVQ 75 differences between white
children in clean and polluted neighborhoods were
diminished. In retrospect, it would have been highly
desirable to have repeated ventilatory performance
tests on these same children in November of 1968, to
determine whether differences in FEV0 7S values
again became very slight or disappeared altogether.
Other investigators have reported an association
between ventilatory performance of school children
and air pollution exposure. On comparing pulmonary
function of 10- to 11-year-old school children in a
heavily polluted area and a clean area of Osaka,
Japan, during each of 12 successive months,
Watanabeetal.20 found peak flow rates to be associated
with geographic and temporal differences in exposure
to sulfur dioxide and particulate matter. Area differ-
ences in peak flow rates were greater during winter
months when pollutant concentrations were highest
in the polluted area. Total vital-capacity measure-
ments did not show this relationship with area;
however, no corrections were made for observed
differences in height between the children of the two
areas. In a similar comparison of 10- and 11-year-old
children in an industrialized area and a rural area of
Kawasaki, Japan, Toyoma2 reported diminished peak
flow rates and one-half-second forced-expiratory-
volume measurements in children of the highly
polluted area. Area differences were again greatest
during the season of highest pollution; however, in
Kawasaki highest concentrations of pollution oc-
curred during the warmer months of April through
August. Thus pollution exposure exerted its effect
independently of cold weather. Sulfur dioxide and
particulate matter levels were simultaneously
elevated, preventing discrimination of the effects due
to each pollutant alone.
Petrilli et al.21 reported that peak flow and mid-
maximal expiratory flow rates of adolescents in a
relatively polluted industrial area of Genoa, Italy,
were lower than tests obtained from adolescents in a
residential sector of the same city. Slight area
differences in age and height were observed, but
pulmonary function results were not adjusted for
these differences, nor were socioeconomic charac-
teristics of the two areas described.
Anderson and Kinnis,22 comparing first-grade
school children in a British Columbia kraft pulp mill
town with an adjacent clean community, found age-
and height-adjusted peak flow rates of male children
in the mill town to be significantly lower than male
rates in the control town; these differences were not
found among female children. No air pollution
measurements were made in this study.
Five- and six-year-old children living in the most
polluted of four sectors of Sheffield, England, had
significantly lower height-adjusted FEV0 7S values
than children in the other sectors.23 the high-
exposure sector had high concentrations of both
sulfur dioxide and suspended particulates. Pulmonary
function measurements were made during summer
months when pollution levels were relatively low and
acute respiratory infections were infrequent.
Holland et al.1 demonstrated significantly lower
peak expiratory flow rates, after adjusting for the
effect of social class, family size, and past history of
respiratory illness, in urban school children of Kent,
England, living in a moderately polluted residential
area.
McMillan et al.24 failed to find an effect of
photochemical oxidant pollution on peak flow rates
of third-grade school children living in two suburbs of
Los Angeles. Substantial area differences in ethnic
composition and family stability were present, how-
ever, and may have influenced peak flow rates in a
direction opposite to that of air pollution.
These studies point to a consistent relationship
between impaired ventilatory performance of
children 5 to 10 years of age and exposure to sulfur
dioxide in combination with particulate matter.
Impairment was not associated with overt clinical
disease and could be detected only by comparison
with performance of children in control communities
matched for socioeconomic characteristics. Important
covariates of ventilatory performance in children,
including age, height, socioeconomic level, and
history of respiratory disease, were taken into ac-
count in some of these studies. The influences of
cigarette smoking and of occupational exposures to
respiratory irritants were minimized by selection of
elementary school children for study. Performance of
children in the polluted communities generally
improved during seasons of low pollution, but did not
attain the level of their counterparts in clean areas,
suggesting a chronic but still reversible state of
functional impairment.
The present study confirms the findings of those
from other countries. Significantly lower ventilatory
performance, adjusted for height, socioeconomic
Cincinnati Study
6-11
-------
level, and race, was present in elementary school
children exposed to relatively moderate levels of
particulate matter, ranging from 76 to 131 jug/m3
compared to children exposed to 61 to 92 jug/m3.
Lower ventilatory performance was most highly
associated with exposures to suspended sulfate con-
centrations of 8.9 to 10.1 Mg/m3 monitored during
the 7 months of the study, and with suspended
sulfate concentrations of 10.7 to 12.1 Mg/m3
estimated for the lifetime exposure of children who
always lived in the same high exposure neighborhoods
in Cincinnati. This effect appears to have occurred
independently of atmospheric sulfur dioxide, levels of
which were low (less than 52 Aig/m3) in all areas.
Absence of an air pollution effect among blacks is
difficult to evaluate in view of the very low perform-
ance of black children relative to the white and the
possibility that the low performance may have been a
culturally determined response of these children to
the white interviewers.
SUMMARY
Ventilatory performance, measured as FEV0 7S,
of white children residing in a moderately polluted
industrial valley of Cincinnati was significantly lower
than the performance of white children living in a
clean area of the metropolitan region. Areas were
matched for socioeconomic and racial characteristics.
Tests were performed once weekly for a month
during three different seasons of the school year.
Area differences in test results were greatest during
the cold winter month when pollution levels were
highest in all sectors. Socioeconomic level had an
influence on the ventilatory performance of white
children, but black children of both exposure areas
had the lowest test results of all groups. Pollution
exposure had no consistent area effect on the
performance of the black children. In none of the
study sectors were test results significantly associated
with pollution concentrations during the 24 or 4
hours immediately preceding the tests.
Concentrations of total suspended particulates,
suspended sulfates and nitrates, soiling index, and
sulfur dioxide were measured daily at monitoring
stations established adjacent to all schools in the six
study sectors. Indoor and outdoor concentrations
were monitored at one school in each study sector.
Sulfur dioxide levels were low-the 7-month
arithmetic mean was less than 52 jug/m3-in all
sectors. The largest differences in area exposure were
in the concentrations of suspended particulates.
Levels of total suspended particulates were 76 to 131
jug/m3 in polluted sectors and 61 to 92 jug/rn3 in
clean sectors, a relative difference of 32 percent. In
the polluted sectors, the average suspended sulfate
level was 9.5 Mg/m3, compared to 8.3 /ug/m3 in the
clean white sectors, a relative difference of 13
percent. In the polluted white sectors, the average
suspended nitrate level was 3.1 ngjm3, compared to
2.4 ;ug/m3 in the clean white sectors, a relative
difference of 23 percent. Since sulfur dioxide levels
were low in all areas, observed area differences in
FEV0 7S were probably attributable to differences
either in total suspended particulate, suspended
sulfate, or nitrate exposure. Black children were
exposed to substantially different levels of total
suspended particulate, yet showed no appreciable
differences in ventilatory performance. Because
ventilatory performance varied most strongly with
exposure to suspended sulfates in all racial and
socioeconomic groups, the observed area effect on
FEV0 75 was probably attributable mostly to sus-
pended sulfates. However, the socioeconomic-racial
exposure gradients of suspended nitrates were
identical to the gradients of suspended sulfates.
Clearly, the relative effects of total suspended par-
ticulates, suspended sulfates, and suspended nitrates
need further elucidation. The possibility that
suspended sulfates and nitrates exert a combined
detrimental effect on ventilatory function is of
particular concern.
This study substantiated the evidence from several
other countries that pulmonary function of young
elementary school children is adversely affected by
chronic moderate levels of air pollution exposure.
ACKNOWLEDGMENT
The authors gratefully acknowledge the services of
Miss Kathryn McClain who carefully measured all
spirograms and helped to prepare tabulations for this
paper.
REFERENCES FOR SECTION 6.1
1. Holland, W.W., T. Halil, A.E. Bennett, and A.
Elliott. Factors Influencing the Onset of Chronic
Respiratory Disease. Brit. Med. J. 2:205-208,
April 1969.
2. Toyama, T. Air Pollution and Its Health Effects
in Japan. Arch. Environ. Health. 5:153-173,
1964.
6-12
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
3. U.S. Environmental Protection Agency. Refer-
ence Method for the Determination of
Suspended Particulates in the Atmosphere (High
Volume Method). Federal Register. 56(84):
8186-8201, April 30, 1971.
4. Hemeon, W.C.L., G.F. Haines, and H.M. Ide.
Determination of Haze and Smoke Concentra-
tion by Filter Paper Samplers. Air Repair.
5:22-28, 1953.
5. West, P.W. and G.C. Gaeke. Fixation of Sulfur
Dioxide as Sulfitomercurate III and Subsequent
Colorimetric Determination. Anal. Chem.
25:1816-1819,1956.
6. Bromberg, S.M. Aerometric Aspects of the
Cincinnati School Study. Internal report. Divi-
sion of Health Effects Research, National
Environmental Research Center, U.S. Environ-
mental Protection Agency, Research Triangle
Park, N.C. January 1970. 71 p.
7. Spirtas, R. and H.J. Levin. Characteristics of
Particulate Patterns-1957-1966. National Air
Pollution Control Administration, Public Health
Service, U.S. Department of Health, Education,
and Welfare. Raleigh, N.C. Publication No.
AP-6 I.March 1970. 101 p.
8. Air Quality Data (1962), National Air Sampling
Network. Public Health Service, U.S. Department
of Health, Education, and Welfare. Cincinnati,
Ohio. 1964.
9. Air Pollution Measurements of the National Air
Sampling Network, Analyses of Suspended Partic-
ulates, 1963. Public Health Service, U.S. Depart-
ment of Health, Education, and Welfare.
Cincinnati, Ohio. 1965.
10. Air Quality Data from the National Air Sampling
Networks and Contributing State and Local
Networks, 1964-1965. Public Health Service,
U.S. Department of Health, Education, and
Welfare. Cincinnati, Ohio. 1966.
11. Air Quality Data from the National Air Sampling
Networks and Contributing State and Local
Networks, 1966. National Air Pollution Control
Administration, Public Health Service, U.S.
Department of Health, Education, and Welfare.
Durham, N.C. NAPCA Publication No.
APTD-68-9. 1968.
12. Air Quality Data for 1967 from the National Air
Surveillance Networks and Contributing State
and Local Networks, Revised 1971. Office of Air
Programs. U.S. Environmental Protection
Agency. Research Triangle Park, N.C. OAP
Publication No. APTD-0741. August 1971.
13. National Air Surveillance Network 1968
Geometric Means for Urban Stations, Suspended
Particulate Matter. Internal report. Office of Air
Programs, U.S. Environmental Protection
Agency, Research Triangle Park, N.C. 1972.
14. Densen, P.M., E.W. Jones, H.E. Hyman, and J.
Breuer. A Survey of Respiratory Disease among
New York City Postal and Transit Workers; 1.
Prevalence of Symptoms. Environ. Res.
1:265-286, 1967.
15. Densen, P.M., E.W. Jones, H.E. Hyman, J.
Breuer, and E. Reed. A Survey of Respiratory
Disease among New York City Postal and Transit
Workers; 2. Ventilatory Function Test Results.
Environ. Res. 2:277-296, 1969.
16. Wilson, M.G. and D.J. Edwards. Diagnostic Value
of Determining Vital Capacity of Lungs of
Children. J. Amer. Med. Assoc. 75:1107-1110,
1922.
17. Smillie, W.G. and D.L. Augustine. Vital Capacity
of Negro Race. J. Amer. Med. Assoc.
57:2055-2058, 1926.
18. Roberts, F.L. and J.A. Crabtree. The Vital
Capacity of the Negro Child. J. Amer. Med.
Assoc. 55:1950-1954,1927.
19. Damon, A. Negro-white Differences in
Pulmonary Function (Vital Capacity, Timed
Vital Capacity and Expiratory Flow Rate).
Human Biol. 55:380-393, 1966.
20. Watanabe, H., F. Kaneko, H. Murayama, S.
Yamaoka, and T. Kawaraya. Effects of Air
Pollution on Health; Report No. 1: Peak Flow
Rate and Vital Capacity of Primary School
Children. Reports of the Osaka City Institute of
Hygiene. 26:32-37, 1964.
Cincinnati Study
-------
21. Petrilli, R.L., G. Agnese, and S. Kanitz. 23. Lunn, J.E., J. Knowelden, and A.J. Handyside.
Epidemiology Studies of Air Pollution Effects Patterns of Respiratory Illness in Sheffield Infant
Genoa, Italy. Arch. Environ. Health. 72:733-740, School Children. Brit. J. Prevent. Soc. Med.
1966. 27:7-16,1967.
24. McMillan, R.S., D.H. Wiseman, B. Hanes, and
P.P. Wehrle. Effects of Oxidant Air Pollution on
22. Anderson, D.O. and C. Kinnis. An Epidemiologic peak Expiratory Flow Rates in Los Angeles
Assessment of a Pediatric Peak Flowmeter. School Children. Arch. Environ. Health.
Amer. Rev. Resp. Dis. 95:73-80, 1967. 75:941-949,1969.
6-14 HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
CHAPTER 7
SUMMARY AND CONCLUSIONS
7-1
-------
7.1 HEALTH CONSEQUENCES OF SULFUR OXIDES:
SUMMARY AND CONCLUSIONS BASED UPON
CHESS STUDIES OF 1970-1971
John F. Finklea, M.D., Dr. P.H., Carl M. Shy, M.D., Dr. P.H.,
Gory J. Love, Sc.D., Carl G. Hayes, Ph.D.,
William C. Nelson, Ph.D., Robert S. Chapman, M.D.,
and Dennis E. House, M.S.
7-3
-------
INTRODUCTION
Epidemiologic studies of the U. S. Environmental
Protection Agency's Community Health and Environ-
mental Surveillance System (CHESS) program are de-
signed to develop dose-response information relating
short-term and long-term pollutant exposures to ad-
verse health effects.1 This paper summarizes effects
associated with sulfur oxide exposures in CHESS com-
munities in New York and the Salt Lake Basin during
studies conducted in 1970-1971. In addition, studies
conducted in Idaho-Montana (Rocky Mountain
studies), Chicago, and Cincinnati, in which health indi-
cators similar to those used in CHESS were employed,
are included. The individual research reports on which
this study is based often suggested pollutant-disease
associations but left a number of problems unan-
swered. These problems include (1) the relative con-
tribution of various air pollutants, especially sulfur
dioxide (S02), total suspended particulates (TSP),
and suspended sulfates (SS), to observed disease
frequencies; (2) the importance of intervening in-
fluences, or covariates, such as occupational expo-
sures, socioeconomic status, residential mobility, and
cigarette smoking; (3) the association between
chronic disease prevalence and current versus past
pollutant exposures; and (4) the precise pollutant
threshold for excess disease in exposed communities.
Obviously, epidemiologic studies alone cannot
resolve any one of the above problems. The find-
ings summarized in this paper must be substan-
tiated by replicated observations in different years
and under different circumstances. Well controlled
human and animal studies are required to isolate
several of the important intervening variables that are
inherent to studies of free living populations, and to
elucidate the precise nature of the pollutant-disease
relationship. Hence, the conclusions put forth at this
time cannot be definitive, but are offered in the sense
of developing more refined quantitative and scientific
hypotheses concerning pollutant-health effect associ-
ations in a real-life environment. In the CHESS
program, we are repeating our observations, using
essentially the same health indicators in the same
(and more) communities.1 These results will provide
one form of data verification required for scien-
tifically defensible air quality standards.
In relating observed health effects to possible
pollutant thresholds, wherever practical and possible,
three threshold estimates were provided: a worst case
estimate, which attributes an observed adverse health
effect to the lowest pollution exposure suggested by
the epidemiologic studies after considering only the
strongest and most established covariates; a least case
estimate, which attributes an observed adverse health
effect to the highest pollution exposure level sug-
gested by the epidemiologic studies after considering
effects of all covariates; and a best judgment estimate
based upon a synthesis of several studies. The best
judgment estimate duly considers interactions be-
tween pollutants and is at times based upon special
analyses that were necessary when individual studies
raised questions regarding interactions involving pol-
lutants or intervening variables.
RESPONSES TO LONG-TERM POLLUTANT
EXPOSURES
Summary of Chronic Respiratory
Disease Studies
Chronic bronchitis prevalence rates observed in
four CHESS areas are given in Table 7.1.1.2"5 (For
the studies summarized, chronic bronchitis was
characterized by the presence of cough and phlegm
on most days for at least 3 months each year.) In
each of the four studies, a very consistent pattern of
excess chronic bronchitis was found among residents
of more polluted communities. In each case, these
differences were statistically significant. Mean respira-
tory symptom scores, which take into account less
severe as well as more classical chronic respiratory
symptoms, are tabulated for each CHESS area in
Table 7.1.2. Symptom scores substantiate the con-
sistent pattern of excess respiratory symptoms among
participants from more polluted neighborhoods. (As
indicated by the difference in community designa-
tions, pollutant gradients in the several CHESS areas
do not represent equal class intervals. Pollutant ranges
associated with adverse health effects are presented
later.)
In the Salt Lake, Rocky Mountain, and New York
surveys2'3'5 where parents of school children were
studied, several consistent findings were observed.
For both smokers and nonsmokers, there was a male
excess of chronic respiratory disease, whether defined
in terms of bronchitis prevalence or of symptom
scores. Chronic bronchitis rates and symptom scores
were higher among male and female smokers. Finally,
male and female nonsmokers, exsmokers, and current
smokers from more polluted areas had higher chronic
7-4
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 7.1.1. SMOKING- AND SEX-SPECIFIC CHRONIC BRONCHITIS PREVALENCE RATES
BY COMMUNITY IN FOUR CHESS AREAS
CHESS area and
community
Salt Lake Basin
Low
Intermediate 1
Intermediate II
High
Rocky Mountain
Pooled Low
Pooled High
Chicago
Black
Outstate (Low)
Suburban (High II)
Urban (High 1)
White
Outstate (Low)
Suburban (High II)
Urban (High 1)
New York
Low
Intermediate 1
Intermediate II
Percent chronic bronchitis
Nonsmoker
Male
3.0
3.6
2.3
6.8
1.25
3.47
8.8
7.8
9.5
4.2
5.4
5.4
4.6
18.0
14.2
Female
2.3
2.0
4.7
5.2
1.08
2.54
2.0
7.5
4.9
Exsmoker
Male
2.6
3.4
5.4
6.0
1.45
4.82
13.9
18.0
18.7
Female
5.3
4.0
7.0
7.1
3.12
2.80
3.8
9.0
4.5
Smoker
Male
19.9
18.6
20.1
26.8
17.05
18.63
8.8
12.7
13.0
17.6
18.8
17.8
13.9
21.3
22.1
Female
17.8
14.7
15.3
22.2
11.78
12.88
13.9
19.8
16.6
bronchitis rates and symptom scores than those from
less polluted areas.
Despite considerable variation in the population
characteristics and pollutant exposures of the above
three CHESS areas, the relative contribution of
cigarette smoking alone was greater than the effect of
the air pollution gradient (Table 7.1.3), with the
exception of males in New York. Among males in the
Salt Lake and Rocky Mountain CHESS areas, and
among all females, air pollution alone was associated
with an excess bronchitis rate (when compared with
nonsmokers of less polluted neighborhoods) ranging
from 1.5 to 3.8 percent. Among New York City
males, this excess was 11.3 percent — an unusually
high figure requiring verification in subsequent study
years. Cigarette smoking alone accounted for an
excess bronchitis rate of 9.3 to 16.6 percent in the
three CHESS areas. Thus, the relative contribution of
air pollution alone ranged from one-third to one-
seventh as strong as that of cigarette smoking as a
determinant of chronic bronchitis prevalence in com-
munities (with the exception of males in New York,
where air pollution appeared to make a slightly larger
contribution than smoking - a finding difficult to
accept in the light of other evidence). The range of
observed differences in the relative contributions of
smoking and pollution is not surprising in view of the
quantitative and qualitative differences in pollution
profiles of the communities studied as well as the
community differences in smoking patterns. The sum
of the evidence suggests that, while personal cigarette
smoking is the largest determinant of bronchitis pre-
valence among parents of school children, air pollu-
tion itself is a significant and consistent contributing
Summary and Conclusions
7-5
-------
Table 7.1.2. MEAN RESPIRATORY SYMPTOM SCORES FOR ALL STAGES
OF CHRONIC RESPIRATORY SYMPTOMS IN FOUR CHESS AREAS
CHESS area
and community
Salt Lake Basin
Low
Intermediate I
Intermediate II
High
Rocky Mountain
Pooled Low
Pooled High
Chicago
Black
Outstate (Low)
Suburban (High II)
Urban (High I)
White
Outstate (Low)
Suburban (High II)
Urban (High I)
New York
Low
Intermediate I
Intermediate II
Mean symptom score
Nonsmoker
Male
1.54
1.56
1.48
1.94
1.33
1.38
2.10
2.24
2.34
1.76
1.82
1.84
1.81
2.41
2.35
Female
1.36
1.37
1.55
1.73
1.23
1.29
1.29
1.76
1.61
Exsmoker
Male
1.48
1.54
1.82
1.87
1.41
1.58
2.05
2.56
2.51
Female
1.65
1.79
1.88
2.02
1.30
1.45
1.37
2.00
1.72
Smoker
Male
2.73
2.88
2.99
3.32
2.65
2.72
2.47
2.61
2.71
2.90
2.93
2.91
2.48
2.76
2.76
Female
2.57
2.53
2.44
2.98
2.20
2.34
2.30
2.68
2.53
factor, leading to increased bronchitis rates in non-
smokers as well as smokers from polluted com-
munities.
Among young white military recruits studied in
the Chicago area,4 air pollution was associated with a
considerably smaller excess in bronchitis rates (Table
7.1.3) than was found in the other CHESS areas, and
the contribution of air pollution was relatively much
less than that of cigarette smoking. However, there
was evidence that even among these young (18- to
24-year-old) inductees, respiratory symptoms were
more prevalent among persons from more polluted
communities. Black and white inductees showed
similar effects of pollution on bronchitis rates (Table
7.1.1). In the case of blacks, these effects were
superimposed on higher base rates among persons
residing in relatively clean outstate areas. Whether
these high rates are attributable to sources of indoor
pollution or other environmental factors, and
whether these baseline rates are indeed verifiable,
remains to be studied. Strangely, cigarette smoking
alone was associated with no excess bronchitis in
blacks, while cigarette smoking and air pollution
combined accounted for more bronchitis than the
additive effect of both pollution and smoking. Sam-
ple sizes among blacks in these categories were small.
Among whites, air pollution and cigarette smoking
were generally additive in their effect on bronchitis
rates among smokers within polluted communities of
each of the four CHESS areas.
Attempts were made to assess the length of
residence in polluted areas required for development
of excess bronchitis rates.2'4 These findings should
be accepted in a very preliminary vein because
7-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
7.1.3. EXCESS CHRONIC BRONCHITIS ATTRIBUTABLE TO AIR POLLUTION
AND SMOKING, AND RELATIVE CONTRIBUTION OF EACH FACTOR
Factor
Males
Air pollution alone
Smoking alone
Air pollution + smoking
Air pollution alone:
smoking alone
Females
Air pollution alone
Smoking alone
Air pollution + smoking
Air pollution alone:
smoking alone
Excess chronic bronchitis prevalence, percent3
New
York
11.3
9.3
17.1
1:0.82
4.0
11.9
16.1
1:3.0
Salt
Lake
3.8
16.6
23.8
1:4.4
2.2
12.8
19.2
1:5.8
Rocky
Mountain
2.2
15.8
17.4
1:7.2
1.5
10.7
11.8
1:7.1
Chicagob
1.2
13.4
<13.6
1:11.2
aExcess prevalence: absolute excess above rate experienced by nonsmok-
ers in the less polluted community of the same CHESS area.
Analysis restricted to whites because of small black sample size.
relatively small sample sizes were available for anal-
ysis after populations were subdivided into smoking-
and residence-duration-specific groups. The overall
evidence suggests that immigrants into polluted areas
reported excess chronic bronchitis after 2 to 7 years
of exposure. Further evidence from the New York
study5 indicated that movement from polluted to
clean communities could effect a substantial decline
in bronchitis rates, while migration into a polluted
community seems to result in high bronchitis rates
like those of the long-time residents of more polluted
neighborhoods. These conclusions should be taken as
hypotheses for further testing, but they justify some
optimism about current efforts to improve air qual-
ity. An important feature of the CHESS program is
the plan to resurvey residents of the more polluted
neighborhoods during and after achievement of de-
sirable air quality.1 These studies can provide con-
siderably more firmness to the conclusions stated in
this monograph.
Other covariates such as age, race, sex, socio-
economic status, and occupational exposure were
controlled, insofar as possible, by the selection of
study areas and by appropriate adjustments in statisti-
cal analyses. Covariates other than occupational
exposure played a relatively minor role as determi-
nants of bronchitis prevalence. Participants with
known occupational exposure were analyzed sepa-
rately, and in none of the above quantitative assess-
ments concerning air pollution and cigarette smoking
was the occupationally exposed group included.
Occupational exposure to irritating dusts, fumes, and
aerosols added to the effects of ambient air pollution
and cigarette smoking in producing a higher preva-
lence of chronic bronchitis among exposed workers.2 •3
Summary and Conclusions
7-7
-------
In general, occupational exposures made a quan-
titative contribution somewhat larger than that of air
pollution and one-half as large as cigarette smoking.
Table 7.1.4 lists current (i.e. during the year of the
survey) and past exposures (within 10 years)6"9
estimated for those communities in which excess
bronchitis was observed in the four CHESS studies.
The precise exposure or dose that should be associ-
ated with excess respiratory symptoms could not be
determined because accurate measures of past expo-
sures were not made and the duration of exposure
required to produce excess respiratory disease is not
known. Current exposures may be taken as a worst
case estimate for the chronic bronchitis effect, and
past exposures as a least case estimate. In the best
judgment of the investigators, excess chronic bron-
chitis in the Salt Lake Basin could be reasonably
attributed to sulfur dioxide levels of 92 to 95 jUg/m3
and/or suspended sulfate levels of 15 jug/m3. This was
the only CHESS area in which low concentrations of
total suspended particulates occurred in the presence
of elevated sulfur oxide pollution. Pollutant concen-
trations measured in 1971 were unlikely determinants
of the excess bronchitis rates in the High exposure
Salt Lake Basin community. In the other CHESS
areas, combinations of particulate matter and sulfur
oxide exposures occurred, and the investigators
judged that the lowest pollutant concentrations that
could reasonably be associated with excess chronic
bronchitis were past exposures to 100 to 177 /ig/m3
sulfur dioxide, 80 to 118 jig/m3 total suspended
particulates, and 9 to 14 ;ug/m3 suspended sulfates.
The individual contribution of each pollutant could
not be identified.
From these data, it appeared that excess bronchitis
may be reasonably associated with community expo-
sures to sulfur oxides alone, in the form of annual
levels of 92 to 95 ng/m3 sulfur dioxide and 15 /zg/m3
suspended sulfates. When higher levels of particulate
matter are present, annual exposures to 100 Mg/m3
sulfur dioxide, 120 /ug/m3 total suspended particu-
late, and 14jug/m3 suspended sulfate are reasonably
associated with excess bronchitis. None of the CHESS
areas experienced elevated exposures to total sus-
pended particulates without concomitant increases in
sulfur oxide levels. Overall, these data support the
existing National Primary Ambient Air Quality Stand-
ards of 80 Mg/m3 annual mean (arithmetic) for sulfur
dioxide and 75 jug/m3 annual mean (geometric) for
total suspended particulates. A National Standard for
suspended sulfates has not been established.
Summary of Lower Respiratory Disease
Studies of Children
In the lower respiratory disease studies conducted
in the Salt Lake Basin and Rocky Mountain CHESS
areas,1 °>11 three findings were consistently observed.
First, for all combinations of disease and numbers of
Table 7.1.4. RANGE OF POLLUTANT EXPOSURES ASSOCIATED WITH
EXCESS CHRONIC BRONCHITIS
CHESS
area
Salt Lake
Rocky Mountain
Chicago
New York
Current exposures
(annual average), jug/m3
SO,
(80)b
62
177-374
96-217
49.9-62.9
TSP
(75)b
66
65-102
103-155
63.1-104.0
SS
(no standard)
12.4
7.2-11.3
14.5
13.2-14.3
Exposures within past 10
years (annual average)3, pg/m3
S02
(80)b
92-95
177-374
100-282
144-404
TSP
(25)b
53-70
62-179
118-177
80-173
SS
(no standard)
15.0-15.3
6.9-19.9
14.1-17.3
9-24
Estimated from emissions data and pollutant trends.
National Primary Air Quality Standard. The particulate standard is a geometric mean; the equivalent arithmetic
mean would be about 85 jug/m3.
7-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
illness episodes, no significant association between
total lower respiratory disease and pollution was
found for children whose parents had been residents
of their communities for less than 3 years. Second,
for single and repeated episodes of croup and
repeated episodes of any lower respiratory disease,
families of children who had lived 3 or more years in
the High exposure communities reported more illness
across all ages of children from 0 to 12 years than did
their counterparts in the less polluted communities.
Third, for single and repeated illness episodes and for
all residence durations, there were no associations of
pollution exposure with pneumonia or number of
hospitalizations for total lower respiratory diseases.
The only inconsistencies noted were that for children
who had lived 3 or more years in their community,
both single and repeated episodes of bronchitis and
single episodes of any lower respiratory diseases were
significantly associated with pollution exposure in the
Salt Lake study, whereas these associations were not
found in the Rocky Mountain study.
The effects of the age and socioeconomic covari-
ates were very consistent in the two lower respiratory
disease studies. In almost every instance, significantly
higher illness rates occurred at younger ages, while
there were very few significant associations with
socioeconomic levels of the household. In the Salt
Lake study, no significant differences in illness rates
of males and females were observed. In the Rocky
Mountain study, male and female children who had
lived in their communities for less than 3 years had
similar illness rates, with the exception of single
episodes of croup. But, in every instance, male
children who had been residents of their community
for 3 or more years had higher illness rates than
females. The reason for this inconsistent sex effect is
unknown.
The increase in the rates of single or repeated
episodes of lower respiratory disease, croup, and
bronchitis attributable to high air pollution exposure
can be determined from the data in Table 7.1.5. This
table gives the illness rates over a 3-year reporting
period for children who had been residents of their
communities for 3 or more years. During the 3-year
periods covered by the two studies, the mean annual
sulfur dioxide concentrations in the High exposure
communities were 92 Mg/m3 in the Salt Lake Basin
study and as high as 177 jug/m3 in the Rocky
Mountain study. Hence, a worst case estimate of the
annual sulfur dioxide concentration associated with
increased lower respiratory disease is 92 /xg/m3, while
the least case estimate is 177 jug/m3- (The National
Primary Standard for sulfur dioxide is 80 jug/m3
annual arithmetic mean.) During the same periods,
mean annual suspended sulfate concentrations in the
High exposure communities were 15 jug/m3 in the
Salt Lake Basin study and as low as 7.2 Aig/m3 in the
Rocky Mountain study. For suspended sulfates, a
worst case estimate of the annual concentration
associated with increased lower respiratory disease is
7.2 /ig/m3 while the least case estimate is 15 jug/m3.
(A National Standard for this pollutant has not been
established.) Total suspended particulate levels in the
Rocky Mountain communities ranged from 65 to 102
jug/m3, representing the worst case and least case
estimates for this pollutant, respectively. (The Na-
tional Primary Standard for total suspended particu-
lates is 75 /ug/m3 annual geometric mean.)
It is interesting to note that larger increases in
total lower respiratory disease and two of its com-
ponents were observed in the High pollution com-
munity of the Salt Lake Basin study than in the
corresponding communities in the Rocky Mountain
study. Also, the mean annual suspended sulfate
concentration was higher in the High pollution
community in the Salt Lake Basin study than in the
Rocky Mountain study; the opposite was true for
sulfur dioxide. This suggests that increases in lower
respiratory disease frequency are probably associated
with suspended sulfates rather than sulfur dioxide.
Several factors should be remembered when inter-
preting the results of the lower respiratory disease
studies. First, the data were collected by asking the
children's parents about illness frequency over a
3-year period. Hence, the recall ability of the parents
could affect the validity of the data, as could the
degree of cooperation of the parents. However, it
does not seem likely that this source of error affected
the communities differently and thereby affected the
community comparisons. Second, there could be
differences in diagnostic criteria among the communi-
ties in a study. In both lower respiratory disease
studies, a sample of physicians were asked to diagnose
six respiratory syndromes so that a determination of
differences in diagnostic criteria could be made. No
differences were found in either study. Third, the
communities observed in the studies were mainly
white and middle class. Therefore, the results of these
studies may not apply to other ethnic or socio-
economic groups. Fourth, a majority of the pollution
exposure data in both studies were estimated from
emissions data. The degree to which these give
reasonable estimates of individual exposures may be
questionable.
Summary and Conclusions
7-9
-------
Table 7.1.5. AGE-, SEX-, AND SOCIOECONOMIC-STATUS-ADJUSTED 3-YEAR
LOWER RESPIRATORY DISEASE ATTACK RATES PER 100 CHILDREN BY COMMUNITY,
NUMBER OF EPISODES, AND STUDY AREA
Disease
category
All lower
respiratory
disease
Croup
Bronchitis
Community
Low
High
Low
High
Low
High
Low
High
Low
High
Low
High
Number
of
episodes
> 1
>2
> 1
>2
> 1
>2
Study area3
Rocky
Mountain
242 (NS)
ft* (p< 0.001)
^'2 (p< O.OOD
3-g (p<0.05)
]« INS,
67 (NS>
Salt Lake
Basin
322 (p< 0.001)
23^ (p< 0.001)
^'4 (p< 0.001)
111 0.05.
Rates given for low pollution exposure are weighted averages of the age-, sex-, and socioeconomic-status-adjusted
rates for the Low, Intermediate I, and Intermediate II communities.
On the basis of the two lower respiratory disease
studies summarized, and in the best judgment of the
investigators, it seems reasonable to conclude that
there is a positive association between lower respira-
tory disease frequency in children and pollution
exposure, and that excess respiratory disease may
reasonably be associated with community exposures
to approximately 95 jug/m3 sulfur dioxide and 15
jUg/m3 suspended sulfates. From these studies, there
is no evidence that elevated levels of total particulate
matter are required to produce the adverse effect.
Summary of Acute Respiratory Disease
Studies of Families
Table 7.1.6 summarizes findings for total acute
respiratory disease (combined upper and lower tract
disease) among family members in the Chicago and
New York studies.12'13 With the exception of
fathers, who often have greater occupational expo-
sures and daily changes of exposure due to place of
work, a consistent excess acute respiratory disease
rate was reported among family members living in
more polluted neighborhoods. The relative excess in
acute respiratory illness rates within more polluted
neighborhoods varied from 3 to 40 percent. A range
estimate of 5 to 20 percent relative excess includes all
but the most extreme values. Unfortunately, the
Intermediate community for the Chicago study also
had elevated pollutant concentrations of sulfur oxides
and particulates; this community, therefore, does not
afford a satisfactory baseline illness rate. In New
York, the Intermediate I community consistently
reported considerably higher illness rates in all family
7-10
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 7.1.6. RELATIVE RISK OF TOTAL ACUTE RESPIRATORY ILLNESS
IN FAMILIES LIVING IN NEW YORK AND CHICAGO AREAS
Family
segment
Preschool
children3
School
children
Mothers
Fathers
Community
Chicago
Intermediate
High
Highest
Intermediate
High
Highest
Intermediate
High
Highest
Intermediate
High
Highest
New York
Low
Intermediate 1
Intermediate II
Low
Intermediate 1
Intermediate II
Low
Intermediate 1
Intermediate II
Low
Intermediate 1
Intermediate II
Relative risk of total
acute respiratory illness
Chicago
1.00 (9.37)
1.06
1.09
1.00 (4.56)
1.39
1.06
1.00 (5.00)
1.19
1.19
1.00(3.09)
0.95
1.19
New York
.00 (7.88)
.40
.03
.00 (6.22)
.20
.03
1.00(4.45)
1.14
1.05
1.00(3.44)
1.17
0.88
aFor Chicago, data are for nursery school children only.
segments than did the Intermediate II community,
even though measured pollutant concentrations were
somewhat lower in the Intermediate I neighborhood.
Other environmental factors, including the proximity
of a large international airport, may have influenced
the illness reporting of residents in the Intermediate I
New York community. For these reasons, it was
difficult to determine the magnitude of excess illness
associated with specific pollutant levels in these two
studies. A conservative estimate (i.e., closer to the
least case than the worst case estimate) would be that
exposures (Table 7.1.7) to 210 Mg/m3 sulfur dioxide
with 104 /ug/m3 total suspended particulates and
approximately 16 /Jg/m3 suspended sulfates were
associated with a 5 to 20 percent excess of acute
respiratory illness in various family members. This
estimate largely discounts the high illness experience
of the Intermediate I New York community and the
relatively low current sulfur dioxide levels in New
York and Chicago. Further observations of acute
respiratory illness in these and other CHESS areas are
being made and should considerably refine the
quantitative estimates given above.
The Chicago study12 also provided evidence of
increased susceptibility to epidemic A2/Hong Kong
influenza among otherwise healthy families exposed
during the previous 3 years to atmospheric levels of
106 to 119 jug/m3 sulfur dioxide, 151 to 159jug/m3
total suspended particulates, and 14 jug/m3 suspended
sulfates.
The effects of other factors on the incidence of
acute respiratory disease (Table 7.1.8) were most
interesting. In both Chicago and New York City
areas, socioeconomic status was a significant factor in
the incidence of upper respiratory disease; more
illness was reported by respondents of the upper-
middle socioeconomic level. No effect of socio-
economic level on reporting of lower respiratory
disease was observed.
Although personal cigarette smoking was associ-
ated with a decrease in illness frequency in New
York, it apparently had little effect on the initial
contracting of respiratory infections, but smoking
was a significant determinant in the development of
lower tract illness as a result of the initial infection.
Parental smoking also was a significant factor in the
Summary and Conclusions
7-11
-------
Table 7.1.7. RANGE OF POLLUTANT EXPOSURES IN NEW YORK
AND CHICAGO COMMUNITIES HAVING EXCESS ACUTE RESPIRATORY
DISEASE RATES IN FAMILIES
Period and
pollutant
Current
exposures
S02 (80)a
TSP (75)a
SS (no
standard)3
Previous 2
years
S02
TSP
SS
Previous 5
years
S02
TSP
SS
Annual average concentration, M9/m3
Chicago
High
51
126
14.5b
83
135
14.1b
107
137 h
14.4b
Highest
106
151 h
14.5b
119
159
14. 1b
170
163
14.4b
New York
Intermediate I
50 to 63
63 to 84
13.2
144
80
13.4
256
94
23
Intermediate II
50 to 58
87 to 104
14.3
210
104
16.2
321
117
23
aNational Primary Air Quality Standard. The particulate standard is a geometric mean; the equivalent arithmetic
mean would be about 85 M9/m3.
Estimate from the Chicago stations of the National Air Sampling Network.
development of lower tract illness among nonsmoking
young members of their households. This is an
important association, providing additional evidence
that smoking is more than a means of self-pollution
and affects other individuals in the immediate en-
vironment as well.
Data collected during acute respiratory disease
studies are difficult to interpret because, in addition
to the effects of all other environmental factors, the
incidence of illness depends first of all on exposure to
infectious agents. The fact that many of these agents
are more virulent than others in itself may affect the
ease with which infections are recognized and re-
ported. Also a mild attenuated agent may protect
against infection with a more virulent one, thus the
recognizable illness rate may depend upon the se-
quence in which successive exposures to infectious
agents occur. These difficulties must be recognized,
but controlling for them is impossible without a
prohibitively large and expensive laboratory activity.
Consequently, the credibility of the results depends
not only on the careful collection and analysis of
data, but also on the reproducibility of associations
between increased illness and higher pollution expo-
sure. This latter factor provides the greatest strength
to the data reported in this monograph. The consist-
ency with which increased illness rates were observed
to be associated with higher pollution exposure levels
in different parts of the country and in the various
segments of the population add greatly to the
credibility of the results.
Differences observed between metropolitan areas,
e.g., between New York and Chicago, were antici-
pated to be greater than those that would be found
between neighborhoods within a single area. These
differences can be accounted for by the fact that data
7-12
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 7.1.8. SIGNIFICANCE OF EFFECTS OF SELECTED FACTORS
ON ACUTE RESPIRATORY DISEASE
Factor
Air pollution
Socioeconomic
status
Personal cigarette
smoking
Parental smoking
effect on
children
Statistical significance of factor3
Upper tract
disease
Chicago
<0.05
< 0.001
NS
<0.10
New York
NS
< 0.005
<0.005b
< 0.005b
Lower tract
disease
Chicago
<0.05
NS
<0.10
< 0.001
New York
< 0.005
NS
< 0.005
< 0.001
aNS - not significant, p>0.10.
Associated with decrease in illness frequency.
were collected in each metropolitan area by different
survey groups; thus techniques were consistent within
the same area but may have varied somewhat from
area to area as a result of differences in execution of
the same study protocol. Furthermore, a more
conservative definition of lower respiratory symp-
toms was employed in the Chicago than in the New
York surveys. As a result, lower respiratory diseases
were reported at lower frequencies in Chicago than in
New York. These definitions have since been stand-
ardized for all CHESS areas.
Summary of Pulmonary Function Studies
Ventilatory function of elementary school child-
ren, measured by the 0.75-second forced expiratory
volume (FEV0.75), was diminished in areas of
elevated exposure to sulfur oxides. In all cases,
observed decrements were subtle. In the New York
study14 only the older children (age 9 to 13 years)
who had been exposed to substantially elevated
pollutant concentrations for the first 5 to 10 years of
life showed evidence of reduced ventilatory function.
The best available estimates of these remote annual
average exposures were as follows: sulfur dioxide, 131
to 435 jug/m3; total suspended particulates, 75 to 200
Aig/m3; suspended sulfates, 18 to 28 jug/m3.
From the New York study, the authors could not
determine the relative importance of specific pollut-
ants in reducing ventilatory function. From the
Cincinnati study,15 however, suspended sulfates
emerged as a pollutant of particular concern. In all
Cincinnati neighborhoods, sulfur dioxide concentra-
tions were at or below the moderate level of 57
A
-------
and 96 ng/m3 respectively), might alone account for
reduced ventilatory function. The New York study
strongly indicated a more moderate interpretation,
however. It was the authors' best judgment that 8 to
9 years of exposure to about 10 to 13 jug/m3 of
suspended sulfates might reduce ventilatory function.
If these suspended sulfate exposures were accom-
panied by exposures to about 200 to 250 Mg/m3 of
sulfur dioxide and about 100 to 150 jug/m3 of total
suspended particulates, further reductions in FEV0.7S
might be expected.
Clearly, these best judgment estimates are based
on suggestive, nonconclusive evidence. In Cincinnati,
for example, the socioeconomic-racial patterns of
exposure to suspended nitrates were very similar to
suspended sulfate exposure patterns. Though absolute
levels of suspended nitrates were much lower than
suspended sulfates, possible effects of suspended
nitrates could not be ruled out. Also, the ventilatory
performance of black children in Cincinnati remains
somewhat confusing. At present, it is impossible to
disentangle the effects of objective environmental
factors from these children's possible subjective re-
sponses to the all-white testing teams.
The contribution of covariates to pulmonary
function results in school children is summarized in
Table 7.1.9. Height, age, sex, and race are well
recognized in the literature as significant determi-
nants of pulmonary function in children, and these
variables were taken into account in analyzing the
CHESS data. Table 7.1.10 summarizes the CHESS
pulmonary function findings to date. These studies
have been repeated in New York and other CHESS
areas for 2 successive years and will be reported in
subsequent papers.
Summary of Threshold Estimates for Long-
term Pollution Effects
Table 7.1.11 summarizes worst case, least case,
and best judgment estimates of the pollution levels
Table 7.1.9. SUMMARY OF EFFECTS OF COVARIATES OBSERVED IN
CHESS PULMONARY FUNCTION STUDIES
Covariate
Effect
Height
Age
Sex
Race
In all studies, height was the most significant determinant of FEV0.7s. being somewhat
less important in children aged 5 through 8 years (0.045 liter/inch) than in children
aged 9 through 13 years (0.063 liter/inch).
In New York, children aged 9 through 13 years demonstrated area differences in FEV0.75
while children aged 5 through 8 years did not.
In all children tested, age was a significant determinant of FEV0.7s, being somewhat
more important in children aged 5 through 8 years (0.045 liter/year) than in children
aged 9 through 13 years (0.019 liter/year).
In New York, area differences in FEV0.7s were statistically significant for boys aged
9 through 13 years, but not for girls of the same age.
The FEVo.75 of boys was consistently higher than that of girls of the same age and
height.
In Cincinnati, the FEV0.7s of black children was consistently lower than that of white
children.
In Cincinnati, white children demonstrated area differences in FEV0.7s, while black
children did not.
7-14
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 7.1.10. SUMMARY OF FINDINGS IN CHESS PULMONARY FUNCTION STUDIES
Location
Time
Age group tested
Findings
Cincinnati
1967-68
Second grade
New York
1970-71
All elementary grades
• White children exposed to average suspended sulfate levels
of 9.5/ig/m3 had lower FEV0 75 than white children ex-
posed to average suspended sulfate levels of 8.3 ;ug/m3.
• The FEVo.75 of black and white children was lowest in
winter.
• The FEVo.75 of black children did not vary with air pollution
exposure.
• The FEV0.7s of black children was consistently lower than
that of whites.
• The FEVo.75 of white children aged 9 through 13
years, who had been exposed to high levels of
sulfur oxides and particulates during the first
decade of life, was lower than that of children who
had not been so exposed.
• This finding was statistically significant in
males, but not in females.
• The FEV0.75 of children aged 5 through 8 years
did not vary consistently with pollution exposure.
• The FEVo.75 of children in all grades was lowest
in winter.
that can be associated with adverse effects on the
health indicators discussed in the previous sections.
RESPONSES TO SHORT-TERM
POLLUTANT EXPOSURES
In contrast to the health indicators previously
summarized in this paper, studies on panels of
asthmatic and cardiopulmonary subjects gave the
investigators the opportunity to relate daily changes
in symptom status to daily changes in pollutant
levels.
Summary of Asthma Panel Studies
One pattern immediately emerged from the asthma
studies conducted in the Salt Lake Basin and New
York.16-17 As shown in Table 7.1.12, daily asthma
attack rates in the Salt Lake Basin were more
consistently correlated with colder outdoor tempera-
ture than with any measured pollutant. Therefore, an
analysis of asthma attack rates against daily pollutant
concentrations was carried out within two ranges of
minimum temperature, 30 to 50°F and greater than
50°F. These data are summarized for sulfur dioxide,
total suspended particulates, and suspended sulfates
in Figures 7.1.1 and 7.1.2 for Salt Lake and New
York, respectively. Inspection of these figures reveals
one quite consistent finding: asthma attack rates were
most closely related to stepwise increases in the levels
of suspended sulfates. Virtually no relationship be-
tween sulfur dioxide and attack rates appeared. Total
suspended particulates (with the exception of Tmin >
50°F for New York) and suspended sulfates (with the
exception of Tmjn = 30 to 50°F for Salt Lake) were
positively and stepwise correlated with daily asthma
attack rates. In the Salt Lake Basin, where the effects
of total particulates and suspended sulfates were
partitioned, a higher frequency of asthma attacks was
Summary and Conclusions
7-15
-------
Table 7.1.11. SUMMARY OF CHESS STUDIES RELATING LONG-TERM POLLUTANT
EXPOSURES TO ADVERSE EFFECTS ON HUMAN HEALTH
Adverse effect
Increase in prevalence of
chronic bronchitis in adults
Increases in acute lower
respiratory tract
infections in children
Increase in frequency or
severity of acute
respiratory illness in
families
Subtle decreases in childhood
ventilatory function
Type of
estimate
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Duration
of
exposure.
years
3
10
6
3
3
3
1
3
3
1
9
8-9
Annual average levels linked to adverse
health effects, jug/m3
Sulfur
dioxide
(80)a
62
374
95
92
177
95
50
210
106
57
435
200
Total
suspended
particulates
(75)a
65
179
100
65
102
102
104
159
151
96
200
100
Suspended
sulfates
(no standard)3
12
20
15
7.2
15
15
14
16
15
9
28
13
aNational Primary Ambient Air Quality Standard. The particulate standard is a geometric mean; the equivalent
arithmetic mean would be about 85 jiig/m3.
Table 7.1.12. SIGNIFICANCE OF CORRELATION OF ASTHMA
ATTACKS WITH ENVIRONMENTAL FACTORS
Factor
Minimum
temperature
Total suspended
particulates
Sulfur dioxide
Suspended
sulfates
Significance of correlation3
Salt Lake
Low
<0.01b
NS
NS
< 0.05b
Intermediate
<0.01b
NS
NS
NS
High
<0.01b
<0.01
NS
<0.01
New York
Low
NS
NS
NS
<0.05
Intermediate I
NS
NS
NS
NS
Intermediate II
<0.05
NS
<0.01b
NS
aNS-not significant, p > 0.05.
I nverse correlation.
7-16
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Tmjn = 30 TO 50 °F
SO2
Tmin -500F
S02
1 16
TSP
053 A
TSP
SS
095
SS
sa 6180 '80 160 61 K >75 <60 618081100 'WO
POLLUTANT CONCENTRATION yg m3
Figure 7.1.1. Relative risk of asthma attack
versus sulfur dioxide, total suspended parti-
culate, and suspended sulfate concentrations:
Salt Lake area, two minimum temperature
ranges.
Tmin=30T050°F
'mm
so.
SS
- so.
Tmin>500F
SS
TSP
^60 6180 '80 ^60 6175 * 75 ^80 81100'IQO
POLLUTANT CONCENTRATION tig m*
Figure 7.1.2. Relative risk of asthma attack
versus sulfur dioxide, total suspended partic-
ulate, and suspended sulfate concentrations:
New York area, two minimum temperature
ranges.
observed at the same daily total suspended particulate
concentration when a high sulfate fraction was
present in the atmosphere (Figure 7.1.3). Thus, it
appeared that sulfate levels were a stronger determi-
nant of asthma attack rates than total suspended
particulates were in the Salt Lake Basin. However, in
the cold dry climate of the Basin, the effect of cold
temperatures was considerably stronger than that of
sulfates, and the pollutant threshold for the asthma
response was much higher in colder than in more
moderate temperatures (Figure 7.1.4).
In New York, asthma attack rates were more
consistently associated with daily suspended sulfate
levels than with either sulfur dioxide or total sus-
pended particulates (Figure 7.1.2). As in Salt Lake,
the pollutant threshold for the asthma response was
higher in colder than in more moderate temperatures
(Figure 7.1.5), but unlike Salt Lake, attack rates were
generally higher on days with more moderate than
with colder temperatures.
Thus, the effect of temperature was somewhat
inconsistent between the two study areas; colder
temperatures were associated with higher attack rates
in the Salt Lake Basin but not in New York. It is
difficult to compare the temperature effect in the
two studies because they extended over different
seasons of the year. In each case, the sulfate threshold
was higher on colder days; and, in each case, elevated
daily sulfate levels were quite consistently associated
with increased asthma attack rates.
Summary of Cardiopulmonary Panel Study
The pattern of daily aggravation of symptoms in
cardiopulmonary subjects in New York18 was very
similar to that of asthma with respect to temperature
and pollutants. In each of the three New York
neighborhoods, cold temperatures were directly re-
lated to increased symptom rates in subjects with
combined heart and lung disease (Table 7.1.13).
Summary and Conclusions
7-17
-------
=50° F
(107)
100 150 ' 200 2SO
PARTICULATE CONCENTRATION, ft, m3
Figure 7.1.3. Effect of total suspended
particulates with and without a high sulfate
content on asthma attack^rate: Sajt Lake
area, two minimum temperature ranges.
Elevated suspended sulfates were the only pollutant
consistently associated with symptom aggravation, as
shown in Table 7.1.13 and Figure 7.1.6. Daily sulfur
dioxide and total suspended particulate concentra-
tions could not be associated with symptom aggrava-
tion in the "heart and lung" panel, which was the
most sensitive to variations in daily pollutant concen-
trations.
Summary of Threshold Estimates for
Short-term Pollution Effects
The pollutant thresholds for sulfur dioxide, total
suspended particulate, and suspended sulfates among
the several cardiopulmonary asthmatic panels at
different temperature ranges are summarized in Table
7.1.14. Although this table presents pollutant thres-
holds for sulfur dioxide and total suspended particu-
lates, the above discussion should make it clear that
suspended sulfate levels demonstrated the only con-
sistent relationship with daily aggravation of symp-
toms in these diseased panelists. Thus, while adverse
effects were occuring at daily concentrations below
the National Daily Primary Air Quality Standard for
sulfur dioxide and total suspended particulates, the
investigators would attribute these effects to sus-
pended sulfate concentrations on those days rather
than to those pollutants. It was the best judgment of
the investigators that significant aggravation of car-
diopulmonary symptoms could be attributed to
24-hour suspended sulfate levels as low as 8 to 10
Mg/m3 on both cooler days (Tmjn = 20 to 40°F) and
warmer days (Tmjn > 40°F). The investigators
intuitively felt that the chemical composition and
particle size involved in sulfate exposures were critical
determinants of the threshold for the adverse re-
sponse. Since these sulfate-symptom relationships
were manifested even in the less polluted communi-
ties of the Salt Lake and New York CHESS areas,
there was evidence that suspended sulfates emanating
from point or urban sources penetrated well beyond
the suburban ring and adversely affected persons
living in more distant communities. Such penetration
might involve smaller respirable particles, acid mist,
or other atmospheric transformation products of
sulfur oxide emissions. The magnitude of such remote
exposures and their overall health importance cannot
yet be quantified.
CONCLUSIONS
In this paper, we have summarized an original
series of studies describing a variety of pollutant
relationships with several health indicators. The pol-
lutants of main concern were sulfur dioxide, total
7-18
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
8 12 16 20
SUSPENDED SULFATE CONCENTRATION, ft'a?
24 28
Figure 7.1.4. Effect of minimum daily
temperature and suspended sulfates on
daily asthma attack rates: Salt Lake area.
ID 15 20 25
SUSPENDED SULFATE CONCENTRATION, |lf/«3
Figure 7.1.5. Effect of minimum daily
temperature and suspended sulfates on
.daily asthma attack rate: New York area.
Table 7.1.13. SIGNIFICANCE OF CORRELATION OF CARDIOPULMONARY
SYMPTOM AGGRAVATION WITH ENVIRONMENTAL FACTORS: NEW YORK
AREA, "HEART AND LUNG" PANEL
Significance of correlation3
Factor
Maximum
temperature
Total suspended
particulates
Sulfur dioxide
Suspended
sulfates
Low
<0.01b
<0.05
<0.01
<0.01
Intermediate 1
< 0.05b
IMS
NS
<0.01
Intermediate II
<0.01b
NS
NS
<0.05
aNS-not significant, p > 0.05.
Inverse correlation.
Summary and Conclusions
7-19
-------
ss
TSP
104
S02
560 6180 =-«0 «60 6175 75 «60 11-1,0 81100 »100
POLLUTANT CONCENTRATION,^™3
Figure 7.1.6. Relative risk of symptom aggra-
vation versus sulfur dioxide, total suspended
particulate, and suspended sulfate concentra-
tions: New York area, two minimum tempera-
ture ranges, "heart and lung" panel.
suspended particulates, and suspended sulfates. We
have examined the impact of community differences
in exposure to these pollutants on chronic respiratory
disease in adults, acute lower respiratory disease in
children, acute respiratory illness in families, aggrava-
tion of symptoms in subjects with preexisting asthma
and cardiopulmonary disease, and lung function of
school children. These health indicators were selected
because past studies by many investigators indicated
that the frequency of these responses in a community
was affected by sulfur oxides and particulates. Our
studies more than substantiate these findings.
Our results can be divided into two groups: (1)
health indicators responsive to cumulative long-term
pollutant exposures and (2) health indicators sensitive
to daily or shorter-term variations in pollutant expo-
sure. Least case, worst case, and best judgment
estimates concerning the pollutant thresholds for
long-term exposures were given in Table 7.1.11. Best
judgment estimates are recapitulated for long-term
exposures in Table 7.1.15. Our findings support the
existing National Primary Standards for long-term, or
annual average, exposures, insofar as we have meas-
ured the desirability of these standards in terms of
the health indicators mentioned.
With regard to short-term exposures, least case,
worst case, and best judgment estimates were given in
Table 7.1.14, and best judgment estimates are recapit-
ulated in Table 7.1.16. Our data indicate that adverse
effects on elderly subjects with heart and lung disease
and on panels of asthmatics are being experienced
even on days below the National Primary Standard
for 24-hour levels of sulfur dioxide and total sus-
pended particulates. However, as is evident from the
presentation, these adverse health effects should be
attributed to suspended sulfate levels rather than to
the observed concentrations of those pollutants. The
consistency of the relationship between symptom
aggravation and sulfate levels, and the lack of
consistency for this relationship with other pollut-
ants, leads us to this conclusion.
Having identified atmospheric suspended sulfates
as an environmental pollutant of present concern to
health, we by no means have acquired sufficient
intelligence to establish a National Standard for this
pollutant. We know little about the environmental
determinants of atmospheric suspended sulfates, and
less about the means to control sulfate levels or to
bring about significant reductions in sulfate con-
centrations in urban, suburban, and rural areas
(particularly of the northeastern United States). In
identifying the need for control of sulfates, we have
raised a series of unstated questions and issues. Are all
sulfates equally biologically reactive? Are sulfates
reactive because of the chemical properties associated
with specific chemical compounds, or because of
physical properties such as particle size or pH? Are
sulfates equally reactive in humid and dry air, and at
warm and cold temperatures? These biological issues
must be addressed and satisfactorily resolved, because
our strategies to control sulfate levels may be
critically dependent on the nature of the sulfate-
biologic response relationship. If acid mist is the
problem, we may be able to neutralize the sulfur
oxides emitted at the source, without more stringent
reductions in sulfur oxide emissions than are pres-
ently required to achieve National Primary Standards.
On the other hand, if atmospheric transformation
products of sulfur dioxide are implicated, we may be
forced to restrict even more severely the sulfui
7-20
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table 7.1.14. SUMMARY OF THRESHOLD ESTIMATES FOR ADVERSE EFFECTS
OF SHORT-TERM EXPOSURES
Adverse effect
Aggravation of cardio-
pulmonary symptoms
in elderly
"Well" panel
•
"Heart" panel
"Lung" panel
"Heart and lung"
panel
Aggravation of
asthma
Type of
estimate
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Worst case
Least case
Best judgment
Minimum
temper-
ature, °F
20-40
>40
>40
20-40
>40
20-40
>40
30-50
>50
Daily average levels linked to
adverse health effects,3 jug/m3
Sulfur
dioxide
(365) b
81-365
NE
NPE
81-365
NE
NPE
NPE
NE
NPE
NPE
NE
NPE
NPE
NE
NPE
181
NE
NPE
NPE
NE
NPE
NPE
NE
NPE
23
NE
180-250C
Total
suspended
particulate
(260)b
NPE
NE
NPE
68
NE
80-100
76-260
NE
NPE
76-260
NE
NPE
76-260
NE
NPE
47
NE
80-100
76
NE
NPE
61-75
NE
105
61-75
NE
70
Suspended
sulfate
(no standard)
< 1
NE
8-10
2
10
8-10
10
10-20
10
6
NE
10
11
12
12
9
10
10
6
17
10
8
NE
9-10
< 1
10
8
aNE—no effect below Primary Standard, or simply no effect for suspended sulfates, for which no Primary
Standard has been established. NPE-no proven effect below Primary Standard, or simply no proven effect for
suspended sulfates.
National Primary Air Quality Standard.
cThis judgment is based on presently summarized studies and on a previously reported CHESS study of asthma in
New Cumberland, West Virginia.19
Summary and Conclusions
7-21
-------
Table 7.1.15. BEST JUDGMENT ESTIMATES OF POLLUTANT THRESHOLDS FOR ADVERSE
EFFECTS OF LONG-TERM EXPOSURES
Effect
Increased prevalence of chronic
bronchitis in adults
Increased acute lower respiratory
disease in children
Increased frequency of acute
respiratory disease in families
Decreased lung function of
children
Threshold (annual average), /ug/m3
Sulfur
dioxide
(80) a
95
95
106
200
Total suspended
particulates (75)a
100
102
151
100
Suspended
su If ates
(no standard)3
15
15
15
13
aNational Primary Air Quality Standard. The particulate standard is a geometric mean; the equivalent arithmetic
mean would be about 85 jug/m3.
Table 7.1.16. BEST JUDGMENT ESTIMATES OF POLLUTANT THRESHOLDS
FOR ADVERSE EFFECTS OF SHORT-TERM EXPOSURES
Threshold, jug/m3
Effect
Aggravation of cardiopulmonary
symptoms in elderly
Aggravation of asthma
Sulfur
dioxide
(365)a
>365
180-250
Total
suspended
particulate
(260)a
80-100
70
Suspended
su If ates
(no standard)3
8-10
8-10
3National Primary Air Quality Standard.
7-22
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
content of fossil fuels. Until more definition of these
issues is achieved, however, our findings strongly
argue against any measures that would allow more
sulfur loading of the atmosphere.
REFERENCES FOR SECTION 7.1
1. Shy, C.M., W.B. Riggan, J.G. French, W.C.
Nelson, R.C. Dickerson, F.B. Benson, J.F. Fink-
lea, A.V. Collucci, D.I. Hammer, and V.A.
Newill. An Overview of CHESS. In: Health
Consequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C. Publi-
cation No. EPA-650/1-74-004. 1974.
2. House, D.E., J.F. Finklea, C.M. Shy, D.C. Cala-
fiore, W.B. Riggan, J.W. Southwick, and L.J.
Olsen. Prevalence of Chronic Respiratory Disease
Symptoms in Adults: 1970 Survey of Salt Lake
Basin Communities. In: Health Consequences of
Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection Agen-
cy. Research Triangle Park, N.C. Publication No.
EPA-650/1-74-004. 1974.
3. Hayes, C.G., D.I. Hammer, C.M. Shy, V. Hassel-
blad, C.R. Sharp, J.P. Creason, and K.E. McClain.
Prevalence of Chronic Respiratory Disease
Symptoms in Adults: 1970 Survey of Five
Rocky Mountain Communities. In: Health Con-
sequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C. Publi-
cation No. EPA-65 0/1-74-004. 1974.
4. Finklea, J.F., J. Goldberg, V. Hasselblad, C.M.
Shy, and C.G. Hayes. Prevalence of Chronic
Respiratory Disease Symptoms in Military Re-
cruits: Chicago Induction Center, 1969-1970. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
5. Goldberg, H.E., J.F. Finklea, C.J. Nelson, W.B.
Steen, R.S. Chapman, D.H. Swanson, and A.A.
Cohen. Prevalence of Chronic Respiratory Dis-
ease Symptoms in Adults: 1970 Survey of New
York Communities. In: Health Consequences of
Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection Agen-
cy. Research Triangle Park, N.C. Publication No.
EPA-650/1-74-004. 1974.
6. Hertz, M.B., L.A. Truppi, T.D. English, G.W.
Sovocool, R.M. Burton, L.T. Heiderscheit, and
D.O. Hinton. Human Exposure to Air Pollutants
in Salt Lake Basin Communities, 1940-1971. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
7. English, T.D., J.M. Sune, D.I. Hammer, L.A.
Truppi, W.E. Culver, R.C. Dickerson, and W.B.
Riggan. Human Exposure to Air Pollutants in
Five Rocky Mountain Communities, 1940-1970.
In: Health Consequences of Sulfur Oxides: A
Report from CHESS, 1970-1971. U.S. Environ-
mental Protection Agency. Research Triangle
Park, N.C. Publication No. EPA-650/1-74-004.
1974.
8. Hinton, D.O., T.D. English, B.F. Parr, V. Hassel-
blad, R.C. Dickerson, and J.G. French. Human
Exposure to Air Pollutants in the Chicago-
Northwest Indiana Metropolitan Region,
1950-1971. In: Health Consequences of Sulfur
Oxides: A Report from CHESS, 1970-1971. U.S.
Environmental Protection Agency. Research Tri-
angle Park, N.C. Publication No. EPA-
650/1-74-004. 1974.
9. English, T.D., W.B. Steen, R.G. Ireson, P.B.
Ramsey, R.M. Burton, and L.T. Heiderscheit.
Human Exposure to Air Pollution in Selected
New York Metropolitan Communities,
1944-1971. In: Health Consequences of Sulfur
Oxides: A Report from CHESS, 1970-1971. U.S.
Environmental Protection Agency. Research Tri-
angle Park, N.C. Publication No.
EPA-650/1-74-004. 1974.
10. Nelson, W.C., J.F. Finklea, D.E. House, D.C.
Calafiore, M.B. Hertz, and D.H. Swanson. Fre-
quency of Acute Lower Respiratory Disease in
Children: Retrospective Survey of Salt Lake
Basin Communities, 1967-1970. In: Health Con-
sequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Pro-
tection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
11. Finklea, J.F., D.I. Hammer, D.E. House, C.R.
Sharp, W.C. Nelson, and G.R. Lowrimore. Fre-
quency of Acute Lower Respiratory Disease in
Children: Retrospective Survey of Five Rocky
Mountain Communities, 1967-1970. In: Health
Consequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C. Publi-
Summary and Conclusions
7-23
-------
cation No. EPA-650/1-74-004. 1974.
12. Finklea, J.F., J.G. French, G.R. Lowrimore, J.
Goldberg, C.M. Shy, and W.C. Nelson. Prospec-
tive Surveys of Acute Respiratory Disease in
Volunteer Families: Chicago Nursery School
Study, 1969-1970. In: Health Consequences of
Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection Agen-
cy. Research Triangle Park, N.C. Publication No.
EPA-650/1-74-004. 1974.
13. Love, G.J., A.A. Cohen, J.F. Finklea, J.G.
French, G.R. Lowrimore, W.C. Nelson, and P.B.
Ramsey. Prospective Surveys of Acute Respira-
tory Disease in Volunteer Families: 1970-1971
New York Studies. In: Health Consequences of
Sulfur Oxides: A Report from CHESS,
1970-1971. U.S. Environmental Protection Agen-
cy. Research Triangle Park, N.C. Publication No.
EPA-650/1-74-004. 1974.
14. Shy, C.M., V. Hasselblad, J.F. Finklea, R.M.
Burton, M. Pravda, R.S. Chapman, and A.A.
Cohen. Ventilatory Function in School Children:
1970-1971 New York Studies. In: Health Conse-
quences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C. Publi-
cation No. EPA-650/1-74-004. 1974.
15. Shy, C.M., C.J. Nelson, F.B. Benson, W.B.
Riggan, V.A. Newill, and R.S. Chapman. Ventila-
tory Function in School Children: 1967-1968
Testing in Cincinnati Neighborhoods. In: Health
Consequences of Sulfur Oxides: A Report from
CHESS, 1970-1971. U.S. Environmental Protec-
tion Agency. Research Triangle Park, N.C. Publi-
cation No. EPA-650/1-74-004. 1974.
16. Finklea, J.F., D.C. Calafiore, C.J. Nelson, W.B.
Riggan, and C.G. Hayes. Aggravation of Asthma
by Air Pollutants: 1971 Salt Lake Basin Studies.
In: Health Consequences of Sulfur Oxides: A
Report from CHESS, 1970-1971. U.S. Environ-
mental Protection Agency. Research Triangle
Park, N.C. Publication No. EPA-650/1-74-004.
1974.
17. Finklea, J.F., J.H. Farmer, G.J. Love, D.C.
Calafiore, and G.W. Sovocool. Aggravation of
Asthma by Air Pollutants: 1970-1971 New York
Studies. In: Health Consequences of Sulfur Ox-
ides: A Report from CHESS, 1970-1971. U.S.
Environmental Protection Agency. Research Tri-
angle Park, N.C. Publication No. EPA-
650/1-74-004. 1974.
18. Goldberg, H.E., A.A. Cohen, J.F. Finklea, J.H.
Farmer, F.B. Benson, and G.J. Love. Frequency
and Severity of Cardiopulmonary Symptoms in
Adult Panels: 1970-1971 New York Studies. In:
Health Consequences of Sulfur Oxides: A Report
from CHESS, 1970-1971. U.S. Environmental
Protection Agency. Research Triangle Park, N.C.
Publication No. EPA-650/1-74-004. 1974.
19. Cohen, A.A., S. Bromberg, R.M. Buechley, L.T.
Heiderscheit, and C.M. Shy. Asthma and Air
Pollution from a Coal Fueled Power Plant. Amer.
J. Public Health. 62:1181-1188, 1972.
7-24
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
APPENDICES
A-l
-------
APPENDIX A:
CHESS MEASUREMENT METHODS,
PRECISION OF MEASUREMENTS, AND
QUALITY CONTROL
This appendix presents the pollutant measurement
methods employed by the Environmental Protection
Agency's (EPA) Community Health and Environ-
mental Surveillance System (CHESS), describes the
quality control procedures employed to ensure
optimum operation of the system, and presents
results of duplicate testing to determine the precision
of the measurements.
In recent years, additional types of pollutant
measurement devices and automated or improved
analysis methods have been incorporated in the
CHESS system. In order to provide a complete and
current description, these new methods and pro-
cedures, as well as those employed during the studies
described in this report, are included.
in a second polypropylene tube. The air next passes
through a 23-gauge hypodermic needle, which be-
haves as a critical orifice controlling the flow rate at
approximately 500 ml/min. A second moisture trap is
placed between the needle and the vacuum source.
Moisture traps are used tn reduce corrosion for both
the vacuum pump and needle. They also function as
containers for overflow of TCM solution should the
train be assembled incorrectly. A calibrated roto-
meter attached at the ambient air inlet is used to
measure the flow rate at the beginning and end of a
sampling period. A change in flow rate of ±50 ml/min
voids a sample. If flow decreases below 400 ml/min,
the system is adjusted.
Principle and Applicability
MEASUREMENT METHODS
Sulfur Dioxide Determination (West-Gaeke
Method)1
The method is the reference method as specified in
the Federal Register.2 CHESS sampling and analysis
procedure deviated from the reference method only
by using a flow rate of 500 ml/min in lieu of 200
ml/min, and 35 ml of absorbing solution in lieu of 50
ml.
Hardware
The 24-hour sulfur dioxide bubbler system is
composed of relatively simple parts (Figure A.I).
From an ambient air inlet, air is drawn through a
15-cm-long bubbler stem of 6-mm glass tubing. One
end of the tubing is drawn to an inner tip diameter of
0.06 cm to control bubble size. The air then bubbles
through 35 ml of a 0.04 M sodium tetrachloro-
mercurate (TCM) solution contained in a 164- by
32-mm polypropylene tube. The airstream then
passes through a glass wool moisture trap contained
Sulfur dioxide is collected by bubbling air through
a sodium tetrachloromercurate solution to form a
solution of nonvolatile dichlorosulfitomercurate. The
ion produced during sampling is determined colori-
metrically by reacting the exposed absorbing reagent
with formaldehyde and acid-bleached pararosaniline
hydrochloride to form the intensely colored pararos-
aniline methyl sulfonic acid.
The method is applicable to collection of 24-hour
samples in the field and subsequent analysis in the
laboratory by manual or automated methods.
Range, Sensitivity, Precision, Accuracy, and Stability
The system obeys Beer's Law up to 1.0 absorbance
units.
The error for the combined sampling and analyti-
cal techniques is ±10 percent in the concentration
range below 26 /ug/m^, with increasing accuracy with
concentration in the range of 26 to 2600 ;Ug/rrP.
After sampling, the concentration of the dichloro-
sulfitomercurate ion decreases at an average rate of
A-3
-------
Vh TO AMBIENT AIR SOURCE
NEEDLE
COPPER IVIANIFOLD
GLASS IVIANIFOLD
BUBBLER STEM
REAGENT
Figure A.1. Sulfur dioxide bubbler train.
1.5 percent per day. It is necessary, therefore, to
analyze these samples with a minimum of delay.
Analysis Apparatus
The following equipment is required for the
analysis:
1. Volumetric flasks - 100, 500, and 1000 ml.
2. Graduated cylinders - 100 ml.
3. Pipets — 1, 2, 3, 4, 5, 8, and 10 ml, volumetric
transfer; 1 and 2 ml graduated in 0.1-ml
intervals; 1 and 10 nil pipets are desirable but
not necessary.
4. Test tubes.
5. Spectrophotometer or colorimeter capable of
measuring absorbance or percent transmittance
at 560 nm.
Reagents
Sampling — The absorbing reagent, 0.04 M sodium
tetrachloromercurate (TCM), is prepared by dissolv-
ing 10.88 grams (0.04 mole) mercuric chloride and
4.68 grams (0.08 mole) sodium chloride in 1 liter of
distilled water. (The solution is highly poisonous,
and, if spilled, should be flushed from the skin with
water immediately.)
Analysis — The following reagents are required for
the analysis:
1. Sulfamic acid — 0.84 gram of sulfamic acid
diluted to 100 ml with distilled water.
2. Formaldehyde - 0.5 ml of 38 percent
formaldehyde diluted to 100 ml with dis-
tilled water. Solution is prepared weekly.
A-4
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
3. Stock pararosaniline hydrochloride (PRA)
solution — 0.20 gram of PRA is dissolved in
100 ml of distilled water and filtered after
48 hours. The PRA used must have an assay
of better than 95 percent and an absorbance
maximum at 543 or 544 nm. The solution is
stable for at least 3 months if stored in a
dark, cool place.
4. Acid bleached PRA — The solution is pre-
pared by pipetting 20 ml of stock PRA into
a 100-ml volumetric flask, adding 6 ml of
concentrated hydrochloric acid, and, after 5
minutes, diluting to 100 ml with distilled
water. The color of the solution, which
should be pale yellow with a greenish tint, is
checked to assure that it has been mixed
properly. In an amber bottle, the solution
can be stored for a week at room tempera-
ture or for about 2 weeks under refrigera-
tion.
5. Stock sodium thiosulfate solution (0.IN) —
The stock solution is prepared by dissolving
25 grams of sodium thiosulfate (Na2S203-
5H20) in 1000 ml of freshly boiled, cooled,
distilled water and adding 0.1 gram of
sodium carbonate to the solution. The solu-
tion is allowed to stand for 1 day before
standardization.
6. Sodium thiosulfate titrant (0.01 AO - 100
ml of stock sodium thiosulfate solution is
diluted to 1000 ml with freshly boiled,
cooled, distilled water. (Normality of result-
ing solution is equal to normality of stock
solution times 0.100.)
7. Stock iodine solution (0.1 N) - The stock
solution is prepared by mixing 12.7 grams of
iodine, 40 grams of potassium iodide, and 25
ml of water in a 250 ml beaker. The solution
is then transferred quantitatively to a volu-
metric flask and diluted to 100 ml with
distilled water.
8. Iodine solution (0.01 A/) - 50 ml of the
stock iodine solution is diluted to 500 ml
with distilled water.
9. Starch indicator solution — This solution is
prepared by triturating 0.4 gram soluble
starch and 0.002 gram mercuric iodide
(preservative) with a little water and adding
the paste slowly to 200 ml of boiling water.
The solution is boiled until it is clear,
cooled, and transferred to a glass-stoppered
bottle.
10. Sulfite solution for standardization - 0.8
gram sodium metabisulfite (Na2S2C>5) is
dissolved in 500 ml of recently boiled,
cooled, distilled water. (Concentration is
roughly 1000 jug SO2/ml.)
11. Stock sulfite solution — By titration ol the
sulfite solution for standardization, as de-
scribed in the following paragraphs on
standardization, the amount of sodium
metabisulfite needed to yield a solution with
an accurate concentration of 1000 yug
S02/ml is determined. A fresh solution of
this amount of sodium metabisulfite in
distilled water is prepared weekly.
12. Working sulfite standards — A solution of 10
/ug SO2/ml is formed by diluting 2.0 ml of
the stock sulfite solution to 200 ml with
0.04 M TCM. By appropriate further dilu-
tions with TCM, working standards in the
range 0.1 to 1.0/ug/ml are obtained.
Standardization
Standard Sodium Thiosulfate - This solution is
prepared by standardization of the approximately 0.1
N solution of sodium thiosulfate (reagent 5 of the
analysis section) prepared previously. First, 1.5 grams
of primary standard potassium iodate (previously
dried at 180 °C) is weighed to the nearest 0.1 mg and
dissolved in water, and the solution is diluted to 500
ml in an iodine flask. Then 50 ml of this iodate
solution, 2.00 grams of potassium iodide, and 10 ml
of 1 TV hydrochloric acid are mixed in a 500-ml
volumetric flask. The flask is stoppered and allowed
to stand for 5 minutes. The stock thiosulfate solution
is then used to titrate the solution in the flask. The
titration is continued until the solution becomes pale
yellow. At this point, 5 ml of starch indicator
(solution 9) is added, and the titration is continued
until the blue color disappears. The normality of the
sodium thiosulfate solution is calculated from the
following equation:
N = 2.80 W/Vt
(A.I)
where
N = normality of stock thiosulfate solution
W = weight of potassium iodate, grams
V( = volume of thiosulfate required, ml
The factor 2.80 results from multiplying the
conversion from grams to milligrams (1000) by the
fraction of iodate used (0.1) and dividing by the
equivalent weight of potassium iodate (35.67).
Stock Sulfite Solution - In determining the weight of
sodium metabisulfite needed to produce a concen-
tration of 1000 jug S02/ml for the stock sulfite
solution (solution 10), two 500-ml iodine flasks, each
containing 50 ml of 0.01 N iodine (solution 8) are
Appendix A
A-5
-------
used. To flask A (blank), 25 ml of distilled water is
added; to flask B (sample), 25 ml of solution 10. The
flasks are stoppered, allowed to react for 5 minutes,
then titrated with 0.01 TV sodium thiosulfate (solu-
tion 6). Titration is continued until each flask turns a
pale yellow. Then 5 ml of starch solution 9 is added
and the titration is continued until the blue color
disappears. The concentration of sulfur dioxide in
sulfite solution 10 is then calculated from the
following equation:
C}= (32,000/25) (Va-Vb)(N) (A.2)
where
Cj = concentration of sulfur dioxide in
solution 10, Mg/ml
Va = volume of thiosulfate for blank, ml
Vjj = volume of thiosulfate for sample, ml
N = normality of thiosulfate (from Equa-
tion A.I)
32,000 = milliequivalent weight of sulfur di-
oxide, jUg
25 = volume of sulfite solution, ml
The weight of sodium metabisulfite that will give
the desired concentration of sulfur dioxide for
the stock sulfite solution (solution 11) when dis-
solved in 500 ml of distilled water can then be
calculated from
or
where
= W1(C2/C1)
(A.3)
W2 =
weight of sodium metabisulfite needed
for solution 1 1
weight of sodium metabisulfite used for
solution 10 (known)
concentration desired for solution 11
(1000 jug/ml)
Cj = concentration of solution 10 (from
Equation A.2)
€2 =
Analysis Procedure
Samples are removed from the sampling train,
properly sealed, and returned to the laboratory for
analysis. At the laboratory, the sample is brought
back to its original volume by the addition of distilled
water to compensate for water loss during sampling.
Then, 0.1 ml of sulfamic acid solution (solution 1) is
added and mixed with the sample. In a test tube, 5 ml
of the sample and 5 ml of TCM are mixed. Blanks
consisting of 10 ml unexposed TCM are also prepared
and treated in the same manner as the samples. Each
sample is mixed with 1 ml of formaldehyde, then
with 1 ml of bleached PRA. After 20 minutes to
allow complete color development, the transmittance
of the sample is determined at 560 nm on a
colorimeter previously zeroed on the blank. At least
three freshly prepared standards are run with each set
of samples.
A manual method of analysis was used by CHESS
until October 1972, after which analysis was accom-
plished on the Technicon Autoanalyzer. At the time
of changeover, the TCM concentration was changed
from 0.1 M to 0.14 M; the additional 0.04 M TCM
was needed to prevent clogging in the Autoanalyzer.
The change in TCM concentration was tested by EPA
for accuracy and was also incorporated into the
Federal Register method. Results of the TCM change
caused no loss of precision in the analysis.
The sulfur dioxide concentration of the samples is
determined from the following equation:
Cs = log(%T)(100) (0.02tSs + 85) (A.4)
2.30258
where
Cs = sulfur dioxide concentration, /ig/ml
%T = percent transmittance of sample
from colorimeter
t = time between sampling and anal-
ysis, days
Ss = slope of Beer's law curve deter-
mined from standards
2.30258 = conversion from natural to com-
mon logarithm
0.02 = average rate of decay of samples
A spectrophotometer, measuring absorbance
rather than transmittance, can also be used. For
absorbance, concentration is determined from the
following equation:
= As(0.02tSs
(A.5)
where
Ac =
absorbance
slope of Beer's law curve determined
from standards
time between sampling and analysis,
days
0.02 = average rate of decay of samples
t =
A-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Nitrogen Dioxide Determination (Jacobs-
Hochheiser Method)3
The method is the reference method as specified in
the Federal Register.4 CHESS sampling and analysis
procedure deviated from the reference method only
by using a flow rate of 500 ml/min in lieu of 200
ml/min and 35 ml of absorbing solution in lieu of 50
ml. In addition, two bubbler tubes in series with a
restricted stem were used in lieu of one bubbler tube
with a fitted stem.
Recent Evaluation of the Method
Testing by EPA personnel in early 1972 revealed
and identified apparent deficiencies with the Jacobs-
Hochheiser method. Further reevaluation showed
that the collection efficiency of the reference method
varies nonlinearly with nitrogen dioxide concentra-
tion from 15 percent at 740 Mg/m^ to 50 to 70
percent at 20 to 50 (j.g/m^. A second deficiency
found is a positive interference caused by the presence
of nitric oxide (NO) in the ambient atmosphere.
Results of the evaluation, shown in Figure A.2 and
Table A.I, have been summarized in the Federal
Register5 and an Environmental Science and Technol-
ogy article.6 The findings affect the range, sensitivity,
interferences, and accuracy as specified in the refer-
ence method and noted in the following paragraphs.
The concentration of nitrogen dioxide in samples
collected from the ambient air, such as by CHESS
monitoring stations, is in the range where it is
questionable to assume a constant collection effi-
ciency of 35 percent.
Hardware
The 24-hour nitrogen dioxide gas bubbler system
(Figure A. 3) is very similar to the sulfur dioxide
bubbler described previously. The air is bubbled
through a sampling train of two 164- by 32-mm
polypropylene tubes containing 0.1 TV sodium hy-
droxide. Otherwise, components and operation are as
described for the sulfur dioxide bubbler.
70
O>
60
o 50
LjJ
g 40
u_
UJ
H 30
20
10
o 1ST PERMEATION TUBE
D 2ND PERMEATION TUBE
A 3RD PERMEATION TUBE
• 4TH PERMEATION TUBE
•0-
" 30 90 150 210 270 330 390 450 510 570 630 690 750
N02 CONCENTRATION, jig/m3
Figure A.2. Response to the nitrogen dioxide reference method
during EPA evaluation.
Appendix A
A-7
-------
Table A.1. EFFECT OF NITROGEN OXIDE ON THE
REFERENCE METHOD FOR NITROGEN DIOXIDE
Concentration, jug/m3
N02
100
102
105
122
180
244
248
215
311
316
318
356
NO
0
63
127
627
0
1205
1279
1242
0
111
332
1060
Ratio
NO/NO2
0.0
0.6
1.2
5.1
0.0
4.9
5.2
5.8
0.1
0.4
1.1
3.0
Expected N02
recovered,
percent
39
39
38
36
29
24
23
26
20
20
20
18
Apparent N02
recovered, percent
38
38
52
57
29
45
55
50
17
30
33
44
NEEDLE CPPPER MANIFOLD
BUBBLER STEM
TO AMBINET
AIR SOURCE
Figure A.3. Nitrogen dioxide bubbler train.
A-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Principle and Applicability
Nitrogen dioxide is collected by bubbling air
through a sodium hydroxide solution to form a stable
solution of sodium nitrite. The nitrite ion produced
during sampling is determined colorimetrically by
reacting the exposed absorbing reagent with phos-
phoric acid, sulfanilamide, and N-1-naphthyl-
ethylenediamine dihydrochloride (NEDA).
The method is applicable to collection of 24-hour
samples in the field and subsequent analysis in the
laboratory.
Range, Sensitivity, and Interferences
The range of the analysis is 0.04 to 1.5 jug
N02/ml. A concentration of 0.04 jug N02/ml will
produce an absorbance of 0.02 using 1-cm cells. The
interference of sulfur dioxide is eliminated by con-
verting it to sulfuric acid with hydrogen peroxide
before analysis.
Precision, Accuracy, and Stability
The relative standard deviations are 14.4 and 21.5
percent, respectively, at nitrogen dioxide concentra-
tions of 140 and 200 Mg/m3 based on an automated
analysis of samples collected from a standard test
atmosphere. The precision for manual analysis would
probably be different, and no accuracy data are
available for manual analysis. Samples are stable for
at least 6 weeks.
Analysis Apparatus
The following equipment is required for analysis:
1. Volumetric flasks - 50, 100, 200, 250, 500,
and 1000 ml.
2. Graduated cylinder — 1000 ml.
3. Pipets - 1, 2, 5, 10, and 15 ml, volumetric; 2
ml graduated in 0.1-ml intervals.
4. Test tube.
5. Spectrophotometer or colorimeter — Capable
of measuring absorbance at 540 nm; band
width is not critical.
Reagents
The absorbing reagent, 0.1 TV sodium hydroxide, is
prepared by dissolving 4.0 grams of sodium hydrox-
ide in distilled water and diluting to 100 ml.
For the analysis, the following reagents are re-
quired:
1. Sulfanilamide - 20 grams of sulfanilamide is
dissolved in 700 ml of distilled water, mixed
with 50 ml of concentrated phosphoric acid (85
percent), and diluted to 1000 ml. This solution
is stable for a month if refrigerated.
2. NEDA - 0.5 gram N-1-naphthylethylene-
diamine dihydrochloride is dissolved in 500 ml
distilled water. This solution is stable for a
month if refrigerated and protected from light.
3. Hydrogen peroxide — 0.2 ml of 30 percent
hydrogen peroxide is diluted to 250 ml with
distilled water. This solution can be used for a
month if protected from light.
4. Standard nitrite solution - This solution is
prepared by dissolving a calculated amount of
desiccated sodium nitrite (NaNC>2, assay of 97
percent or greater) and diluting to 1000 ml
with distilled water. The amount of sddium
nitrite used to yield a solution with a nitrogen
dioxide concentration of 1000 jug/ml is calcu-
lated by the following equation
G- 100(1.500/A)
(A.6)
where
G = amount of sodium nitrite required,
grams
A = assay, percent (thus requiring the 100
factor)
1.500 = gravimetric factor in converting NO 2
to NaN02
5. Working nitrite standards — A solution of 10 jug
N02/ml is formed by dilution of 1 ml of the
standard nitrite solution to 100 ml with 0.1 N
sodium hydroxide. By appropriate further dilu-
tions with sodium hydroxide, working stand-
ards of 0.05, 0.1, 0.2, 0.3, 0.4, and 0.5 fig/ml
are obtained.
Analysis Procedure
Sample tubes are removed from the sample train,
properly sealed, and returned to the laboratory for
analysis. At the laboratory, the sample is brought
back to its original volume by the addition of distilled
water to compensate for water loss during sampling.
To 10 ml of the collected sample in a test tube, 1.0
ml of hydrogen peroxide solution, 10.0 ml of
sulfanilamide solution, and 1.4 ml of NEDA solution
are added, with thorough mixing after the addition of
each reagent. A blank, using 10 ml of absorbing
Appendix A
A-9
-------
reagent, is prepared in the same manner. After a
10-minute color-development period, the absorbance
of the sample is measured against the blank at 540
mil. At least three standards are run with each set of
samples.
A manual method of analysis and the Technicon
Autoanalyzer method were both used by CHESS
until June 1, 1972, after which only the automated
method was continued.
Calculation
The nitrogen dioxide concentration of the samples
is determined from the following equation:
(A-7)
where
Cn = nitrogen dioxide concentration, jug/ml
An = absorbance of sample
Sn = slope of Beer's law curve determined
from standards
Total Suspended Particulate Determination
(High-volume Sampler)7
SHELTER
FACEPLATE AND
GASKET
FLOWMETER
FILTER
ADAPTER
ASSEMBLY
MOTOR
UNIT
The method is the reference method as specified in
the Federal Register.8
Principle, Applicability, and Hardware
The high-volume air sampler (Figure A.4) is used
to collect relatively large quantities of particulate
matter on filter papers. The airflow rate is sufficiently
high to allow collection of a sample adequate for
gravimetric and chemical analysis. The high-volume
sampler has three basic parts: (1) the faceplate and
gasket, (2) the filter adapter assembly, and (3) the air
pump and motor unit.
Figure A. 4. High-volume
particulate sampler.
rate greater than 40 liters/min voids the sample. Flow
rates falling below 1130 liters/min are investigated
and, depending on the cause, may void the sample.
Each unit is inspected and repaired every 25
calendar days in a program of preventive mainte-
nance. A shelter of suitable design protects the
sampler from adverse weather and vandalism.
The sampler pulls ambient air through a Gelman
type A 20- by 25-cm (8- by 10-inch) glass fiber filter
at the rate of 1130 to 1700 liters/min (40 to 60
ft3/min). The motor operates on approximately
90-volt, 60-hertz alternating current utilizing a "buck
or boost" transformer between the sampler and the
power outlet.
The sampler and rotameter are calibrated as a unit
for accurate flow rate determinations. Start and stop
flows are recorded for each sample. A change in flow
Range and Sensitivity
When the sampler is operated at an average flow
rate of 1700 liters/min for 24 hours, an adequate
sample will be obtained even in an atmosphere having
concentrations of suspended particulates as low as 1
jug/m3. If particulate levels are unusually high, a
satisfactory sample may be obtained in 6 to 8 hours
or less. For determination of average concentrations
of suspended particulates in ambient air, a standard
sampling period of 24 hours is recommended.
A-10
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Weights are determined to the nearest milligram,
air flow rates are determined to the nearest 28
liters/min (1 ft3/min), times are determined to the
nearest 2 minutes, and mass concentrations are
determined to the nearest microgram per cubic meter.
Interferences
Particulate matter that is oily, such as photochem-
ical smog or wood smoke, may block the filter and
cause a rapid drop in airflow at a nonuniform rate.
Dense fog or high humidity can cause the filter to
become-wet and severely reduce the airflow through
the filter. Glass-fiber filters are comparatively insensi-
tive to changes in relative humidity, but the collected
particles can be hygroscopic.^ If the water can be
reversibly desorbed, there is no error in weighing
because the filters are equilibrated at low humidity
prior to weighing. Glass fiber filter media having a
high pH can generate sulfate on the surface of the
media if sulfur dioxide gas is present in the sampling
atmosphere.
Precision, Accuracy, and Stability
Based upon collaborative testing, the relative
standard deviation (coefficient of variation) for a
single analysis variation (repeatability of the method)
is 3.0 percent. The corresponding value for multi-
laboratory variation (reproducibility of the method)
is 3.7 percent.^
The accuracy with which the sampler measures the
true average concentration depends on the amount of
reduction of airflow rate and on the variation of the
mass concentration of dust with time during the
24-hour sampling period.8
Procedure
Prior to sampling, each filter is exposed to the
light source and inspected for pinholes, particles, or
other imperfections. Particles are removed with a
small brush, and filters with visible imperfections are
rejected. The filters are equilibrated in the filter
conditioning environment for 24 hours, and then
weighed. Tare weight and filter identification number
are recorded.
After sampling, the filter is again equilibrated for
24 hours, then reweighed. The weight of particulate
matter collected is the weight of the filter after
sampling minus its tare weight. After weighing, the
filters may be subjected to detailed chemical analysis.
Calibration
The instrument is routinely calibrated for accurate
measurement of the volume of air sampled. A
primary standard positive-displacement rotary meter
is used in conjunction with a calibrating orifice-water
manometer assembly. Details of calibration procedure
are specified in the Federal Register.^
Suspended Sulfate Determination
In the CHESS sampling program, the manual
turbidimetric (Sulfaver) method of analysis for sul-
fate was used until September 1971. At that time, the
methylthymol blue method replaced the Sulfaver
method in order to automate the procedure, thus
making possible a higher rate of analysis and reducing
the chance for operator error.
Analysis Apparatus
The following equipment is required for analysis:
1. A balance room or desiccator maintained at 20
to 25 °C and less than 50 percent relative
humidity for filter conditioning.
2. An analytical balance equipped with a weighing
chamber designed to handle unfolded 20- by
25-cm filters and having a sensitivity of 0.1 mg.
3. A light source. Frequently a table of the type
used to view X-ray film is used.
4. A numbering device capable of printing identi-
fication numbers on the filters.
5. Rootsmeter for airflow rate calibration.
Sample Collection and Extraction
Suspended sulfate concentration is determined
from ambient particulate matter collected by the
standard high-volume sampler with glass fiber or
other filter media. A sufficient quantity of particulate
matter is collected in a typical 24-hour sampling
period to allow chemical analysis for the sulfate ion.
After the exposed filter is equilibrated and weighed, a
2- by 20-cm (3/4- by 8-inch) strip of it is placed in an
erlenmeyer flask, 25 ml of distilled water is added,
and the mixture is refluxed for 30 minutes. After the
solution has cooled, it is filtered through acid-washed
Whatman No. 42 (or equivalent) filter paper and then
diluted to 50 ml with distilled water.
Appendix A
A-ll
-------
Turbidimetric Method^ °
Principle and Applicability—An aqueous extract of
the sample is treated with barium chloride in the
presence of Sulfaver, a proprietary stabilizing agent,
forming barium sulfate crystals of uniform size. The
absorbance of the barium sulfate suspension is
measured by a spectrophotometer or filter photom-
eter. The sulfate ion concentration is determined by
comparison to a previously determined standard
calibration curve.
Precision and Accuracy -Based upon the analysis of
128 samples in triplicate, sulfate ion concentration
determinations are reproducible within ±5.8 percent,
producing a 95 percent confidence level, provided
there is uniformly good technique and interferences
are absent. The accuracy based upon recovery stand-
ards is estimated to be ±11.2 percent.
Sensitivity—Sensitivity may be increased by concen-
trating the sample or decreased by diluting it. The
method is adequately sensitive for levels found in
ambient environments.
Interferences—The accuracy of the method is changed
when the concentration of the sample is changed
since the size and shape of the barium sulfate crystals
are affected by the strength of other ions present in
the solution. In addition, if the concentration of
sulfate is greater than 40 jug/ml of sample, the
accuracy of the method decreases as the barium
sulfate suspension loses stability. The standards
recommended under standardization were selected to
give maximum operating range. Reproducibility and
accuracy are also affected by ion strength, pH,
concentration of reagents, and temperature. For the
indicated accuracy, the following conditions should
be specified with respect to the conditions existing
for preparation of the standard curve: (1) pH = ±1
unit, (2) temperature = ±5 °C, (3) Sulfaver added =
0.50 ± 0.01 gram, and (4) the specific conductance of
aqueous extracts should not vary more than 300
/umhos/cm.
Reagents—A\l chemicals must be American Chemical
Society analytical-reagent grade. The following re-
agents are required:
1. Sulfaver powder: A proprietary item (Cat. No.
396, Hach Chemical Co., Ames, Iowa), which
contains barium chloride stabilizing and condi-
tioning reagents.
2. Stock standard sulfate solution-A 500 jug
SO4/ml solution prepared by dissolving 0.74
gram anhydrous sodium sulfate (Na2SO4) in
distilled water and diluting to 1000 ml in a
volumetric flask.
3. Working standard sulfate solution—A 100 jug
S04/ml solution prepared by diluting 10.0 ml
of the stock standard to 100 ml in a volumetric
flask with distilled water. This solution may not
be stable for more than 24 hours; therefore,
fresh solutions should be prepared immediately
before use.
Analysis Apparatus—The following equipment is re-
quired for the analysis:
1. Refluxing apparatus—125-ml erlenmeyer
flasks fitted with standard taper joints and
matching water-cooled 200-ml-long con-
densers.
2. Heating apparatus—Electric hotplate with
thermostatic control.
3. Filter paper—Whatman filter paper No. 42.
4. Photometer—Spectrophotometer or filter
photometer for use at 500 nm with a light
path of 1 to 5 cm for preparing standard
sulfate solution.
5. Measuring spoon-0.5-gram capacity (Hach
Chemical Co., Ames, Iowa).
6. Stirrer-Adams Cyclo-Mixer (or equivalent)
with timer.
7. Graduated cylinders.
8. Volumetric flasks.
9. Pipettes.
10. Test tubes.
11. Cuvettes.
12. Spectrophotometer or colorimeter—Capable
of reading absorbance in the 400-nm range
for analyzing sample.
Analytical Procedure-Fmm the filter extract, pre-
pared as described earlier in this section, pipet 20 ml,
or a suitable aliquot made up to 20 ml, into a 25- by
150-mm test tube. Add 1 level spoonful (0.5 gram) of
Sulfaver and mix with the Cyclo-Mixer for exactly 60
seconds, making certain that the samples and stand-
ards are mixed the same length of time and at the
same temperature. After at least 5 but less than 20
minutes, transfer the sample to a cuvette and read theT
transmittance. To correct for the filter and the
reagents, subtract the absorbance of a distilled water
blank carried through the same procedure.
Estimate the sulfate concentration in the 20-ml
sample by referring the absorbance to the standard
calibration curve, preparation of which is described in
the next section.
A-12
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Standardization—Prepare a series of standards from
the working standard solution according to Table
A.2.
Table A.2. PROPORTIONS FOR MIXING STAND-
ARDS FOR SUSPENDED SULFATE ANALYSIS
Working standard
solution, ml
0
1
2
4
6
8
Distilled
water, ml
20
19
18
16
14
12
Sulfate concentration,
A
-------
6. Sodium hydroxide—Prepared by dissolving 7.2
grams of sodium hydroxide in 800 ml of
distilled water, allowing solution to cool, and
diluting it to 1000 ml with distilled water.
Analytical Procedure—Uncomplexed methylthymol
blue is measured using a Technicon Autoanalyzer.1-'
Since the measurements obtained do not conform to
Beer's law, a Technicon Linearizer is also required to
obtain readings that are directly proportional ito
concentration. Concentrations from the linearizer,
in micrograms per milliliter, are converted to micro-
grams per cubic meter of air by Equation A.8.
Since cations, such as calcium, aluminum, and
iron, would interfere with the measurement by
complexing the methylthymol blue, these ions are
removed by passage through an ion-exchange column.
The column consists of a length of glass tubing 19 cm
long, 2.0 mm inner diameter, and 3.6 mm outer
diameter filled with the commercial ion-exchange
resin Bio-Rex 70 (Bio-Rad Laboratories, New York,
N.Y., 20-50 mesh, sodium form).
ally 100 percent of the nitrate is reduced to nitrite
after 10 to 15 minutes at 52 °C. The nitrite is
converted to nitrous acid, which diazotizes sulfanil-
amide. The resulting complex is coupled with
N- 1-napthylethylenediamine dihydrochloride
(NEDA) to produce an azo dye. The intensity of the
dye color is measured at 540 nm on a colorimeter and
the absorbance is related to concentration of sus-
pended nitrate.
Range and Sensitivity—The range is 0.1 to 50.0 jug
N03/ml or approximately 0.03 to 15 /ug N03/m3.
The sensitivity is 0.1 /ug N03/ml or approximately
0.03 /ug N03/m3 of air.
Interferences—The method is free from interferences
normally encountered in procedures based on the
nitration of aromatic compounds. Turbidity in the
final solution will cause an erroneous increase in the
apparent nitrate concentration. This effect may be
eliminated by use of a spectrophotometer with a
monochromatic light source.
Suspended Nitrate Determination
In the CHESS sampling program, the hydrazine
sulfate-copper sulfate method of analysis for nitrate
was used until December 1971. At that time, it was
replaced with the copper-cadmium reduction method
in order to refine the baseline stability of the
instrumentation used. A Technicon Autoanalyzer was
used for both methods.
Sample Collection and Extraction
The suspended nitrate concentration is determined
from ambient particulate matter collected on glass
fiber filter media by the standard high-volume
sampler. A sufficient quantity is collected in a typical
24-hour sampling period to allow chemical analysis
for the nitrate ion. After the exposed filter is
equilibrated and weighed, a 2- by 20-cm (3/4- by
8-inch) strip is cut and placed in an erlenmeyer flask,
25 ml of distilled water is added, and the mixture is
refluxed for 30 minutes. After the solution has
cooled, it is filtered through acid-washed Whatman
No. 42 (or equivalent) f" Tjer, and then diluted
to 50 ml with distilled wa...
Hydrazine Sulfate-Copper Sulfate Method14
Principle and Applicability—The nitrate (NC<3) is
reduced to nitrite (NC>2) by hydrazine sulfate. Virtu-
Nitrites, if present, will react quantitatively and
give false high nitrate values. Correction may be made
by a parallel analysis for nitrite by substituting
distilled water for the reducing agents, copper sulfate
and hydrazine. Subtraction of the nitrite value from
the total value gives nitrate concentration. Nitrite is
usually a negligible interference.
Precision, Accuracy, and Stability—The standard devi-
ation for the analytical portions of this method is
±0.02 /ug N03/ml for samples containing 1.0 /ug
N03/ml. The color developed must be measured
within 30 minutes after the addition of the diazo-
tizing-coupling reagent.
Analysis Apparatus—The following equipment is
required for the analysis:
1. Refluxing apparatus consisting of a 125-ml
erlenmeyer flask fitted with standard taper
joints and a matching water-cooled 200-mm-
long condenser.
2. An electric hot plate with thermostatic control.
3. Graduated cylinders - 50 and 2000 ml.
4. Volumetric flasks - 50, 100, 500, and 1000 ml.
5. Pipettes - 1, 2, 5, 10, 20, 50, and 100 ml.
6. A spectrophotometer or colorimeter capable of
reading absorbance at 460 nm. This may be
part of an automated system.
A-14
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Reagents—The following reagents are required for the
procedure:
1. Sodium hydroxide (10.0 TV) — Prepared by
dissolving 400 grams sodium hydroxide in 500
ml freshly boiled and cooled (carbon dioxide-
free) distilled water, allowing the mixture to
cool, and diluting it to 1000 ml with carbon
dioxide-free distilled water.
2. Sodium hydroxide (0.3 TV) - Prepared by
diluting 30 ml of the 10.07V sodium hydroxide
solution to 1000 ml with carbon dioxide-free
distilled water.
3. Acetone solution (10%) — Prepared by diluting
100 ml acetone to 1000 ml with distilled water.
4. Diazotizing-coupling reagent - Prepared by
completely dissolving 200 ml of 85 percent
0-phosphoric acid and 80 grams sulfanilamide
in approximately 500 ml distilled water, adding
4.0 grams NEDA and diluting to 2000 ml.
Solution is stable for 4 to 5 weeks when stored
in a dark bottle and kept in a cool place.
5. Stock copper sulfate solution — Prepared by
dissolving 2.5 grams copper sulfate (CuS04-
5H20) in 500 ml distilled water and diluting to
1000ml.
6. Working copper sulfate solution - Prepared by
diluting 10 ml of the stock solution to 1000 ml
with distilled water.
7. Stock hydrazine sulfate solution - Prepared by
dissolving 13.73 grams hydrazine sulfate in
water and diluting to 1000 ml. The solution is
stable for 6 months or longer.
8. Working hydrazine sulfate solution - Prepared
by diluting 50.0 ml of stock solution to 1000
ml with distilled water. The solution is stable
for 1 month.
9. Standard nitrate solution — Prepared by dis-
solving 0.8153 gram anhydrous potassium
nitrate in distilled water and diluting to 100 ml.
This solution contains 50 /ug N03/ml. After
dilution, 1 ml chloroform is added for preserva-
tion. A fresh solution is prepared every week.
Analysis Procedure-Far the manual method of anal-
ysis, 25 ml of the filter extract, prepared as described
earlier in this section, is pipetted into a 50 ml
volumetric flask. To this is added successively (since
order of addition affects final color development),
with complete mixing after each addition, 1 ml of
working copper sulfate reagent, 1 ml of 0.3 N sodium
hydroxide, and 1 ml of working hydrazine reagent.
The flask is placed in a 52 °C water bath for 30
minutes (timed). The flask is removed and 2 ml of
acetone solution and 6.0 ml of the diazotizing-
coupling reagent are added with complete mixing
after each addition. The solution is diluted to 50 ml
with distilled water and mixed well. Absorbance is
then measured against a true blank after at least 10
but not more than 30 minutes has elapsed.
This analysis may be automated. The sampling
pattern for analysis is: sample, reagent-washout,
sample; and the sampling rate is 45 samples per hour.
Nitrate concentration in micrograms nitrate per milli-
liter is read from a standard curve prepared using
standard solutions, described in the next paragraph,
in the automatic analyzer.
Calibration Standards-By diluting 50.0 ml of the
500-jug NOs/ml solution to 500 ml with distilled
water, a solution containing 50 ]Ug N03/ml is ob-
tained. To obtain concentrations of 0.5, 1.0, 5.0,
10.0, 25.0, and 40.0 jug/ml, respectively, 1-, 2-, 10-,
20-, 50-, and 80-ml portions of the above solution are
pipetted into a series of 100-ml volumetric flasks and
diluted to the mark with distilled water. These
standards are analyzed by either the manual or
automated method. A plot of absorbance versus
nitrate concentration gives a standard curve, which
relates absorbance to concentration for ambient air
samples.
Calculation—The concentration of suspended nitrate
is expressed as micrograms of nitrate per cubic meter
of air at standard conditions (25 °C and 760 mm Hg).
The following equation is used:
Csn=ArVsC,F/V
where
Csn = nitrate concentration,
Ar = ratio of sample area to total exposed
filter area (12)
Vs = volume of filter extract (50 ml)
GI = concentration of nitrate in solution,
jug/ml
F = dilution factor
V = volume of air sampled, m^
or
= 600CiF/V
Copper-Cadmium Reduction Method^ $
(A.9)
Principle and Applicability-This automated pro-
cedure, performed with a Technicon Autoanalyzer,
for the determination of nitrate and nitrite utilizes
the procedure whereby nitrate is reduced to nitrite by
Appendix A
A-15
-------
a copper-cadmium reductor column.15 The nitrite
ion then reacts with sulfanilamide under acidic
conditions to form a diazo compound. This com-
pound then couples with N-1-napthylethylenediamine
dihydrochloride (NEDA) to form a reddish-purple
azo dye.
Range, Sensitivity, and Interferences—The concen-
trations of oxidizing or reducing agents and poten-
tially interferring metal ions are well below the limits
causing interferences. When present in sufficient
concentration, metal ions may produce a positive
error, i.e., divalent mercury and divalent copper may
form colored complex ions having absorotion bands
in the region of color measurement. The range of
measurement is 0 to 50 jug/ml.
The performance at 40 samples per hour using
aqueous standards is as follows:
Sensitivity (0.72 absorbance units) 2.0 ppm
Detection limit 0.04 ppm
Coefficient of variation 0.02%
(95% confidence level at 1.0 ppm)
Reagents—Distilled water is not required for the
preparation of all reagents for this analysis. However,
if distilled water is not readily available, care is taken
to ensure that the water used is completely free of
contamination. Reagents are stored in glass bottles,
and contact with air is avoided. The following
reagents are required:
1. Ammonium chloride - First alkaline water is
prepared by adding just enough ammonium
hydroxide to distilled water to attain a pH of
8.5. Then 10 grams of ammonium chloride is
dissolved in the alkaline water and diluted to
1000 ml. Finally, 0.5 ml of Brij-35 is added.
2. Color reagent - Prepared by completely dis-
solving (with heat if needed) 200 ml concen-
trated phosphoric acid and 20 grams of sulfanil-
amide in approximately 1500 ml distilled
water. Then 1 gram of NEDA is dissolved in the
solution and it is diluted to 2000 ml. Finally,
1.0 ml Brij-35 is added. Stored in a cold, dark
place, the solution is stable for 1 month.
3. Stock standards - Prepared by dissolving 0.72
gram of potassium nitrate in distilled water and
diluting to 1000 ml. The solution is stored in a
glass bottle with a few drops of chloroform as a
preservative. Working standards, ranging from 5
to 50 /Lig/ml are prepared from the stock
standard daily by serial dilution.
4. Cadmium powder — Coarse cadmium powder,
99 percent pure. To remove grease and dirt, the
fillings are rinsed with a little clean diethyi
ether or 1 N hydrogen chloride followed by
distilled water. The metal is then air-dried and
stored in a well stoppered bottle.
Reductor Column—The reductor column is a U-
shaped 35-cm (14-inch) length of 2.0-mm inside
diameter glass tubing filled with cadmium powder
that has been carefully washed with copper sulfate
solution and distilled water to remove colloidal
copper.
Analysis—Nitrate concentrations in micrograms
nitrate per milliliter are obtained from a standard
curve prepared using standard solutions in the auto-
matic analyzer. Concentrations in micrograms per
cubic meter of air are calculated by Equation A. 9.
Coefficient of Haze Determination
(Optical Density of Filtered Deposit
Method)16
Principle and Applicability
The tape sampler (Figure A.5) is used to measure
suspended particulate concentration of ambient air
by the soiling index method. Air is drawn through a
section of white filter paper, and the optical density
of the resulting soiled spot is measured by compari-
son of the transmittance of the spot with' that of a
clean section of the same filter paper.'
Sensitivity, Precision, and Accuracy
The maximum amount of deposited matter that
can be evaluated accurately by this method is that
corresponding to 50 percent attenuation of incident
light or a maximum optical density of about 0.30.
For heavy particle loading, which will give rise to
higher optical densities than the maximum range, the
sampling cycle must be shortened accordingly. For
lightly contaminated atmospheres, the sampling
period may be lengthened.
Calibration of the gas flow meter, variations of the
aspiration rate through the paper during sampling, the
filtration efficiency of the collection media, and the
accuracy of the optical density measurement must all
be considered for accurate and precise measurements.
A-16
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
SAMPLE AIR
INLET
TIMING
MECHANISM
SUPPLY SPOOL
VACUUM
PUMP
AND
FLOW METER
PUMP MOTOR
Figure A.5. Tape sampler.
Hardware
The tape sample contains the following functional
subassemblies:
1. Sampling probe with filter screen.
2. Head with clamp to hold inlet tube to paper.
3. Paper holder with roll of filter paper.
4. Interval timer.
5. Calibrated flow meter.
6. Vacuum pump.
7. Spectrophotometer or photoelectric colorim-
eter for measuring optical density.
Operation
automatically advanced, exposing a clean area to the
air stream for the next sample. A laboratory optical
instrument such as the photoelectric colorimeter is
used for determining the degree of soiling of the
samples.
The flow meter of the tape sampler is periodically
calibrated in the laboratory with a standard wet test
gas meter. The calibration curve established is sent to
the field with the instrument.
Analysis
The concentration of the sample is measured by
optical density, expressed as log 0o/I), where Io is
the intensity of light transmitted through a clean part
of the filter paper next to the sample and I is the
intensity of light transmitted through the sample. The
computed optical density is converted to the coeffi-
cient of haze (COH) unit, which is 100 times the
optical density. To account for the variables of flow
rate and sample time, the size of the sample is
expressed in linear units of air; specifically, the results
are calculated as COHs per 1000 linear feet as
follows:
COHs/1000 lin ft = 100 [log (IO/I)] H-
where
rt
1000A,
(A. 10)
I0 = intensity of light transmitted through
clean filter paper
I = intensity of light transmitted through
sample
r = flow rate, ft3/min
t = sample time, minutes
As = area of filter spot, ft2
Conversion:
1 lin ft = 30.5 cm
1 ft 2 = 930.25 cm2
1 ft3 = 2.84xlo4Cm3
Fine Particulate Determination (Cyclone
Separator Method)17 >18
Hardware
The air stream is passed through a 2.5-cm-(l-inch-)
diameter area of the filter for a sampling period of 2
hours. The flow rate, nominally 7 liters/min (0.25
ft3/min) is observed and recorded daily. At the end
of the 2-hour sampling period, the filter paper is
For separating the respirable fraction from the
total suspended particle mass, a 1.25-cm-(0.5-inch-)
diameter stainless steel cyclone collector is used as a
prefilter to allow only the smaller respirable size
particles to be collected on a filter.18 The cyclone
Appendix A
A-17
-------
separator has been calibrated by several laboratories
using aeros. ^ of known size, density, electrostatic
charge, size distribution, state of agglomeration, etc.
The two most recent calibrations have been docu-
mented by the American Industrial Hygiene Associ-
ation.19'20
For field sampling, the cyclone is connected in
series with a downstream preweighed 27-mm-diam-
eter filter and filter pad enclosed in an airtight plastic
cassette (Figure A.6). Downstream of the cassette is a
limiting orifice to control the flow rate at 9 liters/min
(0.3 ft-'/min). A vacuum pump moves ambient air
through the train.
In parallel with the cyclone sampler, an open-face
37-mm cassette simultaneously captures a total sus-
pended particulate (TSP) sample on a similar 37-mm
RAIN HAT
MAST SUPPORT
AND AIR LINE
CRITICAL ORIFICE,
9.0 LITER/MIN.
CYCLONE SEPARATOR
Figure A.6. Cyclone sampler and shelter assembly.
A-18
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
preweighed filter. The open-face cassette also has a
limiting orifice of the same flow rate as the in-line
cassette and cyclone so that comparative mass num-
bers are available for correlations. The total sus-
pended particulate airflow is moved by the same
vacuum pump that operates the cyclone sampler.
Principle and Applicability
Particles moving in an air stream tend to follow
their original straight line direction of motion when
the streamlines of airflow are deflected by an
obstacle. Making use of this principle, a cyclone
separator is utilized to separate the fine respirable
suspended particulate (RSP) fraction of suspended
airborne particles from the larger particles found in
ambient air.18 The air sample is drawn first through
the cyclone where the larger particles are removed
and discarded by impaction and settlement. The small
particles follow the air vortex, pass through the top
of the cyclone, and are captured on a filter for
weighing and further analysis.
Mass concentration of both the fine fraction and
total particulate matter found in ambient air is
determined from the weight of samples collected and
the volume of air passed through the train during
each sampling period. Mass concentration is expressed
as micrograms per cubic meter.
The fraction not passing through the cyclone
separator represents the size particles that are filtered
or trapped before reaching the uncilliated portions of
the lungs. The portion of the sample collected on the
cassette filter after the cyclone represents the
particles that reach the lower regions of the respira-
tory tract. Shown in Figure A.7, along with collected
particle sizes, is the fraction of each size particle not
reaching the deep regions of the lungs. The curve was
developed by the Atomic Energy Commission and has
been accepted by other researchers, including the
American Conference of Government Industrial
Hygienist.18
Although the cyclone was designed primarily for
use in industrial hygiene surveys, it has been adapted
and proven to be successful for sampling mass
concentrations of suspended particulates in ambient
air. The critical orifice stabilizes air flow; therefore, a
fairly representative sample is obtained with good
repeatability.
Range, Sensitivity, Precision, Accuracy, and Stability
If the sampler is operated continuously for 24
hours at a flow rate of 9 liters/min, a meaningful
sample will be obtained, provided the ambient air
o 100
| 90
u^
80
CJ3
u-
o -
«c H-
a: o
70
60
50
40
30
20
10
0
FRACTION OF PARTICLES
PASSED TO FILTER
• CASSETTE
FRACTION OF PARTICLES
SEPARATED OUT BY CYCLONE
'01 2 3 4 5 7 6 8 9 10
AERODYNAMIC PARTICLE SIZE AT UNIT DENSITY, pm
Figure A.7. Fractionation of fine particulates
by cyclone separator.
particulate concentration is at least 30 jug/m3. Due to
the small rate of flow, low-concentration sampling
range is sensitive to weighing procedures.
Filter and sample weights are determined by using
a precision balance. Weights are measured to the
nearest 0.01 mg. Flow rates are measured to the
nearest 0.1 liter/min. Real sampling times are re-
corded to the nearest minute.
High humidity or rainfall pulled into the filter may
dissolve the portion of the sample that is water
soluble.
The filter does not have water vapor gathering
properties; however, the particles themselves can be
hygroscopic and can introduce errors in the weight
determinations if care is not exercised in equilibrating
the samples to a fixed environmentally controlled
humidity and temperature level before weighing.
Results of duplicate sampling at the Durham,
North Carolina, Ambient Air Station has given a
correlation coefficient of 0.97 for cyclone samplers
operated simultaneously side by side, and a correla-
tion coefficient of 0.98 for the open face TSP
cassettes operated simultaneously side by side. At an
average mass concentration of 115 jug/m^ of par-
ticulate matter in ambient air, the standard deviation
Appendix A
A-19
-------
is 4.9
-------
Calculations
The true airflow rate is calculated from the
equation:
Q = V/T
(A. 11)
where
Q = airflow rate, m-^/min
V = volume of air sampled, m^
T = sampling time,.min
The volume of air sampled is determined from the
equation:
Qi + Qf
V= * xT
where
Qi = initial airflow rate, m^/min
Qf = final airflow rate, m3/min
The mass concentration of suspended particulates
(TSP or RSP) is determined by the following equa-
tion:
(Wf-Wj)x 106
(A. 13)
where
Cp = mass concentration of suspended par-
ticulate, jug/m^
Wj = initial weight of filter, grams
Wf = final weight of filter, grams
1Q6 = conversion of grams to micrograms
Size Distribution Determination for Fine
Particulates (High-volume Cascade
Impactor Method)17
Principle and Applicability
An inertia! impactor operates on the principle
that, with enough momentum, particles moving in an
air stream tend to follow their original straight line
direction of motion when the air stream is deflected
or curved in direction, and thus will be impacted on a
surface in their path. Discrimination of particles by
their momentum results in particle size classification
since momentum is directly related to the particle
size, shape, and density.
In the cascade impactor, a series of parallel plates
containing a set number of sized holes are stacked
with selected spaces between each plate. Also, be-
tween each pair of plates is a replaceable collection
medium for collecting the impinged particles. The
holes in each plate are offset from the holes in the
succeeding plate in order to abruptly bend the air
stream. In addition, the holes get smaller with each
consecutive plate in order to increase the velocity
step-wise with each plate, thus creating increasing
levels of momentum for the particles left in the air
stream.
(A. 12) Hardware
The impactor sampler (Figure A.8) is a multistage,
multijet unit consisting of five multiorifice aluminum
plates separated by sized neoprene rubber gaskets. At
the end of the train of five plates is a standard 20- by
25-cm (8- by 10-inch) high-volume sampler filter to
collect submicron particles. Each plate is approxi-
mately 30 cm (12 inches) in diameter and contains
300 uniform-sized jets. The alignment of the 30-cm
plates and resulting airflow patterns direct the par-
ticles onto the surface of the next plate in series
below. The collection surface is covered with a cut
and perforated collection medium, such as glass fiber
or aluminum foil, which has been preconditioned and
weighed. The size and location of the jets are held to
very close tolerances to give proper particle size
fractionation. The collection medium contains punch-
ed holes so that it provides a collection surface for
the plate above, but does not obstruct the jet
openings in the plates on which it is mounted. The
head interface plate contains four dowel pins and a
center post for properly aligning the plates with
respect to each other and with the collection medi-
um. The assembled head is positioned on a 23- by
28-cm (9- by 11-inch) interface gasket and attached
to a conventional high-volume filter holder inside the
high-volume shelter. A standard high-volume air
mover is used to provide airflow through the system.
Regulation of the airflow through the size fraction-
ator is achieved by adjusting a variable voltage
transformer and noting the corresponding pressure
drop through the assembled unit as indicated by a
manometer. To collect particles accurately in prede-
termined size fractions, the fractionator should
operate at 565 liters/min (20 ft-Vmin). Since most of
the particulate sample is collected on the upper stages
of the sampler where airflow does not penetrate the
collection medium, decreases in flow rate during
sampling are minimized.
Appendix A
A-21
-------
SAMPLING SHELTER
HIGH-VOLUME
CASCADE IMPACTOR
HEAD
MANOMETER
HIGH-VOLUME
BLOWER
BACKUP FILTER
_T ADAPTOR HOUSING
VARIABLE VOLTAGE
TRANSFORMER
Figure A.8. Cascade impactor.
Range, Sensitivity, and Interferences
For 24-hour sampling, the instrument can accu-
rately measure ambient concentrations down to 10
/Jg/m3. Weights on each filter average between 5 and
50 mg and are determined on a balance with a
sensitivity of 0.01 mg. Overloading and subsequent
flow-rate drop are eliminated for ambient air sampling
since only the particles <0.93 jum impinge on
collection media through which the airflow must
pass. Sulfate formation on certain types of filter
media may cause erroneous elevated mass numbers.^
A-22 '
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Correlations between duplicate instruments and with
the standard high-volume sampler show a high degree
of correlation.1^'22
Particle Size Fractionation
The impactor fractionates particles into five aero-
dynamic size ranges:
Stage 1 5,5 urn and greater
Stage 2 2.40 to 5.5 urn
Stage 3 1.75 to 2.40 jum
Stage 4 0.93 to 1.75/urn
Stage 5 0 to 0.93 urn
The instrument has been calibrated for particle size
cutoff by the Environmental Protection Agency.21
Airflow Calibration
Each high-volume sampler head should be initially
calibrated and a pressure drop versus altitude curve
drawn for each serial numbered head. This pressure
drop is determined with the high-volume sampler
head installed and a backup filter (glass fiber type A)
placed on the 20- by 25-cm filter holder. Any other
type backup filter would require a new calibration in
the field. A small orifice of 0.914 mm (0.036 inch) is
drilled in the brass fitting (screwed in the interface
plate) where a flowmeter (rotometer) may be
attached in place of the manometer. Since the orifice
only measures partial flow, the flowmeter must be
originally calibrated with the complete high-volume
air sampler using a standard high-volume rootsmeter
or equivalent. The method of calibration is very
similar to that specified in the Federal Register for
calibrating the standard high-volume samplers.8
Care and Grooming
The high-volume sampler head is covered when not
in use. Before use, stages are examined for foreign
material in the holes by holding them up to a light.
With proper operation, handling, and care the holes
should not become plugged. Normal skin oils can
cause partial plugging of the smaller holes if hands or
fingers are brushed across the plates. When necessary,
the jet stages are cleaned in an ultrasonic bath with a
suitable solvent or carefully washed in a detergent
water solution and rinsed and dried thoroughly.
The neoprene gaskets are also cleaned periodically,
as required, in plain water and dried thoroughly
before use. If necessary, the gaskets may be dusted
with talc or baby powder after cleaning to speed up
drying time. Any dampness will cause the collection
paper to adhere to the gaskets.
Component Part Description and Function
The major component parts of the high-volume air
sampler head may be divided into nine sections:
1. Jet plates — There are five jet plates ranging in
thickness from 6.35 to 1.25 mm (0.25 to 0.050
inch).
2. Interface adapter plate - This plate interfaces
between the four circular plates and the standard
high-volume sampler.
3. Gasket-spacer — There are five 6.35-mm-(0.25-
inch-) thick neoprene gasket-spacers to separate
the plates. In damp weather or high humidity, the
gasket-spacer may show a tendency to adhere to
the collection paper. A slight amount of dusting
talc or baby powder on the top and bottom side of
the gasket and careful handling of the perforated
collection papers will minimize this problem.
4. Speedball handle — This handle is designed to
convenience tightening down the head by hand
and also carrying the head from the field to the
lab. Two to three complete turns should be
sufficient to seal the plates and gaskets. While air is
being pulled through the sampler, it can be
determined if the unit is sealed by pressing both
sides of the plates down by hand and observing
any change in pressure drop on the manometer.
5. Perforated collection paper — There are two
configurations of specially designed, cut and
perforated collection paper. Configuration I holes
correspond to the jets of plates 2 and 4; Configura-
tion II holes correspond to jets of plates 3 and 5.
6. Interface gasket — This part seals the interface
plate to the high-volume filter holder.
7. Backup filter assembly - The 20- by 25-cm
backup filter is located below the interface plate
(standard high-volume filter holder).
8. Manometer — A manometer is provided to accu-
rately set the unit for an equivalent pressure drop
corresponding to 565 liters/min. The scale is
calibrated with red gage oil to correspond with
centimeters of water.
9. Variable voltage transformer — Some type of
variable voltage transformer is required to properly
adjust the flow rate (measured by the pressure
drop across the head with a manometer and
referenced to a calibration curve furnished with
each serial numbered head).
Appendix A
A-23
-------
Sample Preparation and Analysis
Sheets of glass fiber filter medium are used for
collecting the particles. They are equilibrated in an
environmentally controlled chamber for 24 hours,
weighed, and loaded into the cascade impactor. The
impactor is shipped to the sampling site, operated for
24 hours, and returned to the laboratory. Upon
arrival, the exposed collection medium is equilibrated
again for 24 hours and then reweighed. Particle mass
for each size fraction is determined and data are
expressed in micrograms of particulate matter per
cubic meter of air sampled.
Dustfall Determination (Settleable Particulate
Method)23
Principle and Applicability
The dustfall bucket is an open-topped cylindrical
container used to collect the larger and more dense
fractions of particulate atmospheric pollutants that
fall out of suspension under the normal force of
gravity. Made of durable, nonreactive polyethylene
and supplied with a press-on lid, the entire unit is
suitable for shipment. The dustfall bucket is 20 cm
(8-1/4 inches) high and 30 cm (7-1/2 inches) in
diameter at the top, with a slight taper to the bottom.
It has a capacity of approximately 5 liters (5 quarts).
For sampling, a clean, dry bucket is suspended on an
aluminum mast at a height of at least 1.8 meters (6
feet) above ground and is allowed to remain for 1
month. After sampling, the bucket is returned to the
laboratory, where the contents of the bucket are
emptied into a glass beaker, any water is evaporated,
and the weight of the remaining matter is determined.
Residue is analyzed chemically for trace metals.24
Settleable particles are defined as any particles, liquid
or solid, small enough to pass through a 1-mm screen
and large enough to settle in the collector.
Interferences
Care must be taken to avoid matter from trees,
bird droppings, and other such deposits. Loss of
material from the collector by action of the wind
must be prevented.
Analysis Apparatus
The following equipment is required for the anal-
ysis:
1. Sieve—20 mesh, brass.
2. Analytical balance-160-gram capacity, 1-mg
sensitivity.
3. Drying oven—capable of maintaining oven
temperature of 105 ± 2 °C.
4. Desiccator.
5. Steam bath (or water bath).
6. Instrumentation for trace metal analysis.
Procedure
v
Dustfall—The sample collected in the dustfall bucket
(or other suitable container) is transferred through a
20-mesh sieve into a 2-liter beaker. Sufficient distilled
water is added to the bucket to ensure complete
transfer of the contents. The sample is evaporated to
low volume on a steam bath, then quantitatively
transferred to a previously dried (at 150 °C) and
tared 100-ml beaker. Volume is further reduced by
evaporation on the steam bath, then the sample is
dried in an oven held at 105 °C. After being cooled in
a desiccator, the sample is weighed to the nearest
milligram.
Dustfall is expressed as grams per square meter per
month, determined as follows:
D=^xf (A.14)
where
D = dustfall, g/m^/mo
Wd = weight of total dustfall, grams
AJ = area of top of collecting jar (0.028 m^
for standard bucket), m2
t = sampling period, days
The term 30/t is used to standardize the measurement
to a 30-day month. The sample is normally collected
over a calendar month (28 to 31 days).
Trace Metals—A suitable volume of nitric acid
(HNOs) diluted with an equal volume of water, is
added to the dustfall sample after dustfall has been
determined. The beaker containing the sample is then
heated (below the boiling point) for about 1 hour to
solubilize the trace metals. The sample is then
concentrated to a convenient volume to remove
excess acid and facilitate transfer.
The sample is decanted into a centrifuge tube. The
beaker is rinsed with distilled water to assure com-
plete transfer, and additional water is added to bring
A-24
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
the sample to a known, convenient volume. Then the
sample is centrifuged to remove solids. The remaining
solution is analyzed on an atomic absorption instru-
ment or other instrumentation using a reagent blank
prepared in the same manner as the sample. Standard
curves for each metal are prepared by plotting
absorbance versus metal concentration in parts per
million by weight for each metal at the appropriate
wave length. (Development of x-ray fluorescence
methods offers much promise for future methods of
analysis.)
The weight of each trace metal in the dustfall
sample can then be calculated as follows:
VSF 30
X T
106A;
(A. 15)
where
Wm = weight of metal, mg/m2/ml
Cm = concentration of metal in sample, ppm
Vs = volume of sample, ml
F = dilution factor, mg/ml
Aj = area of top of collecting jar, m2
t = sampling period, days
QUALITY CONTROL
In order to understand the CHESS quality control
program, it is essential to understand how the quality
control operation fits into the overall operation of
the CHESS air monitoring program. A schematic
diagram of the overall CHESS operational plan is
shown in Figure A.9. The rectangles indicate normal
routine functional operations. The circles indicate
operations especially designed to provide quality
control. Each of the quality control points has been
assigned a number to facilitate the system descrip-
tion.
Materials and Equipment Control
All incoming materials, such as chemicals, filter
papers, etc., are routinely inspected to ensure that
they meet specifications. Before these materials are
used in the CHESS operational areas, scheduled
checks are made at quality control point 1. Chemical
reagents are prepared for gas bubbler tubes in the
chemistry laboratory. These reagents are periodically
checked at quality control point 2 to ensure uni-
formity and absence of any systematic drifts. After
this operation, the filters and reagent tubes are mailed.
to the CHESS operational areas. ,
All incoming instrumentation and equipment
undergo an incoming inspection to ensure that
specifications are met. Before this equipment is used
in the CHESS operational areas, periodic checks are
performed at quality control point 3. Some equip-
ment, such as a high-volume sampler motor, must be
attached to other assemblies before the complete unit
is functional. In this case, the entire unit may require
calibration. At quality control point 4, periodic
checks of the fitness of the assembly and the
accuracy of the calibrations are performed. The
equipment is then ready to be mailed to CHESS
operational areas.
The CHESS area station operators are trained by
Environmental Protection Agency personnel to
ensure uniformity of field procedures. The operators
change bubbler tubes and filter paper, adjust timing
of samplers, record both start and stop airflows,
record both start and stop clock times, fill out site
identification information, note special comments,
perform routine simple maintenance, and phone
CHESS area engineers if an emergency arises. The
CHESS area engineer ensures that emergency situa-
tions are solved in a timely manner.
Another quality control feature of the CHESS air
monitoring program is the use of systematic duplica-
tion of instrumentation. Replicate samples are ob-
tained and analyzed in order to determine the
reproducibility of individual instruments and an
estimate of the range associated with individual
measurement techniques. The above controls are
represented by quality control point 5.
Routine preventive maintenance is used to mini-
mize field equipment failures. Field equipment is
returned in accordance with a preventive maintenance
schedule from the CHESS area to the CHESS central
maintenance area. This equipment, along with equip-
ment that has failed in the field, is centrally repaired,
refurbished, and calibrated. Scheduled checks of this
equipment are made at quality control point 6 before
this equipment is returned to a CHESS operational
area.
Equipment Early Warning System
Upon receipt of field samples, CHESS central
personnel immediately examine the field data cards
to determine any abnormality in the performance of
the CHESS equipment or,the CHESS area operators.
Appendix A
A-25
-------
MATERIALS
INSTRUMENTATION
ISSUE
MONTHLY
DATA
SUMMARY
INCOMING
CHEMICAL
AND FILTER
INSPECTION
PREPARATION
AND SUPPLIES
FIELD SAMPLES
AND DATA CARDS
CALIBRATION
PROBLEM
DESCRIPTION
TO AREA
ENGINEER
DATA CARD
WARNING
SYSTEM
, AREA ^
ENGINEERS
•CHANGES
Figure A.9. CHESS quality control.
This operation is represented by quality control point
7. The area engineer is quickly notified of discrep-
ancies by means of the data card warning system,
which is represented as quality control point 8.
Typical problems that can be identified by this
system include:
1. Deterioration of routeman's performance.
2. Leaks in bubbler system.
A-26
3. Need for replacement of parts.
4. Need for periodic site inspection.
5. Invalid samples.
The area engineer resolves the particular problem
with the CHESS area operator and sends out appro-
priate replacement items. This function is represented
as quality control point 9.
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Chemical Analysis
The CHESS area specimens are analyzed in the
chemistry laboratory, following documented quality
control procedures.25 The analyses follow Federal
Register procedures where possible. Systematic
checks of procedures are made by analyzing standard
samples that contain known amounts of a specific air
pollutant. Data are calculated in the lab before release
in order to ensure their validity. A visual inspection
of each data card is made in order to ensure
completeness. These operations are represented by
quality control point 10. Data, such as COH levels,
not requiring measurement at the laboratory are
checked at control point 11, and data cards are
prepared.
Computer Analysis
The field data cards are sent to the computer
group for computer card punching. These cards are
inspected for card punch errors before they are
analyzed by the computer. This inspection is repre-
sented by quality control point 12.
The data are analyzed by computer according to
an analysis protocol. Examples of monthly averages
for given CHESS monitoring stations are shown in
Figure A. 10. These averages provide a basis of
20
15
CO
00
=1,
u$
LJ
in
O
Q-
oo
10
A OGDEN
O SALT LAKE CITY
O KEARNS
D MAGNA
DEC JAN 1971
APR
JULY OCT JAN 1972
TIME, months
Figure A. 10. Drawing based on graphical output of monthly data.
APR
Appendix A
A-27
-------
r
o
CM T-i »N *-«
t-l «H (Nl »-H .-» *
oo >o \o m
• • • • o
O O\ to O «O
IH r*-
n
o
o
z
A-28
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
comparison for individual data points. If a point
appears to be in error, both the chemistry laboratory
and the computer group are notified as required. The
above functions are represented by quality control
point 13. After the aerometry field data have been
successfully analyzed, they are compiled by the area
engineer and issued as a monthly data summary. A
typical monthly summary is shown in Figure A. 11.
PRECISION OF MEASUREMENT
In order to determine the precision of the air
monitoring techniques used in the CHESS program, a
series of special studies were initiated. The studies
were performed in four cities in the southeastern
United States—Birmingham, Alabama, Charlotte and
Greensboro, North Carolina, and Chattanooga,
Tennessee. Duplicate air monitoring equipment was
installed in selected sites in order to determine the
daily range and daily percentage error for a given air
measurement.
The range is determined by taking the absolute
value of the difference between the pollutant con-
centration measured with the original equipment and
that measured with the duplicate equipment. An
estimate of the daily error between original and
duplicate is made by dividing one-half the range by
the average value.
(C0 - Cd)/2
L-e — a^-
(C0 + Cd)/2
C0 - Cd
- — ^
C0 + Cd
(A. 16)
where
%E = daily percentage error
Co = daily concentration from original equip-
ment
Cd = daily concentration from duplicate
equipment
The daily percentage errors over the period of time
for particular studies are tabulated in the order of
increasing error, for example, 2, 5, 9 percent, etc. The
rank of these percentage errors is listed next to a
given percentage. An estimate of cumulative fre-
quency is made by dividing the rank of a given
percentage error (n) by the total number of duplicate
samples (m), plus one.
CF =
m+ 1
(A. 17)
The estimate of the error at the 50 percent
cumulative frequency is obtained by determining the
percentage error value that corresponds to a cumula-
tive frequency of 50 percent. The 50 percent fre-
quency is chosen because 50 percent of the data have
errors that are smaller than the value, and likewise 50
percent are greater.
Sulfur Dioxide
A study of the precision of sulfur dioxide measure-
ments was made in Birmingham over a 20-month
period. The results of this study are shown in Figure
A. 12. The 50 percent cumulative frequency corre-
sponds to an error of ±18 percent.
10 20 30 40 50 60 70 80
CUMULATIVE FREQUENCY, percent
90 100
Figure A.12 Precision of CHESS sulfur dioxide
measurements (Birmingham, October 1969-
May 1971).
Appendix A A-29
-------
50
40
cc
o
£
LU
O
30
20
10
0
Total Suspended Particulate
A similar 12-month study was performed in three
southeastern CHESS cities (Birmingham, Charlotte,
and Greensboro) to determine the precision of total
suspended particulate measurements. A graph of the
data from this study is shown in Figure A. 13. The 50
percent cumulative frequency corresponds to an error
of 4 percent.
0 10 20 30 40 50 60 70 80 90 100
CUMULATIVE FREQUENCY, percent
Figure A.13. Precision of CHESS particulate
measurements (Birmingham, Charlotte, and
Greensboro, May 1971-April 1972).
0 10 20 30 40 50 60 70 80 90 100
CUMULATIVE FREQUENCY, percent
Figure A.14. Precision of CHESS nitrogen
dioxide measurements (Chattanooga, September
1971-March 1972).
10 20 30 40 50 60 70 80 90 TOO
CUMULATIVE FREQUENCY, percent
Figure A.15. Precision of CHESS dustfall
measurements (Birmingham, Charlotte, and
Greensboro, October 1969-March 1972)
A-30
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Nitrogen Dioxide
A 7-month study of the precision of nitrogen
dioxide measurements was performed in Chatta-
nooga. A graphical presentation of the percent error
frequency distribution is shown in Figure A. 14. The
50 percent cumulative frequency corresponds to an
error of ±13 percent.
Total Dustfall
An 18-month study of the precision of CHESS
monthly dustfall measurements was performed in
three southeastern CHESS cities (Birmingham,
Charlotte, and Greensboro). The data obtained in this
study are summarized in Figure A, 15. The 50 percent
cumulative frequency corresponds to an error of 16
percent.
REFERENCES FOR APPENDIX A
1. West, P. W. and G. C. Gaeke. Fixation of Sulfur
Dioxide as Sulfitomercurate III and Subsequent
Colo rime trie Determination. Anal. Chem.
25:1816-1819, 1956.
2. U. S. Environmental Protection Agency. National
Primary and Secondary Air Quality Standards;
Reference Method for the Determination of
Sulfur Dioxide in the Atmosphere (Pararosaniline
Method). Federal Register. J6(84):8187-8190,
April 30, 1971.
6. Hauser, T. R. and C. M. Shy. Position Paper:
NOX Measurement. Environ. Sci. Technol.
6:890-894, October 1972.
7. Silverman, L. and F. G. Viles. A High-Volume
Air Sampling and Filter Weighing Method for
Certain Aerosols. J. Ind. Hygiene Toxicol.
30:124, 1948.
8. U. S. Environmental Protection Agency. National
Primary and Secondary Air Quality Standards;
Reference Method for the Determination of
Suspended Particulates in the Atmosphere (High
Volume Method). Federal Register. 36(84)'
8191-8194, April 30, 1971.
9. McKee, H, C., R. E. Childers, and 0. Saenz.
Collaborative Study of Reference Method for
Determination of Suspended Particulates in the
Atmosphere (High Volume Method). Southwest
Research Institute. San Antonio, Texas. Contract
CAP 70-40, SwRI Project 21-2811. June 1971.
10. Parr, S. W. and W. D. Staley. Determination of
Sulfur by Means of the Turbidimeter. Ind. Eng.
Chem. (Annual Ed.). 5:66-67, 1931.
11. Sulfate in H20 and Waste H2O. In: Technicon
Autoanalyzer II Methodology. Technicon Cor-
poration. Ardsley, N.Y. Industrial Method
118-71W. January 1971.
12. Lazrus, A. L., K. C. Hill, and J. P. Lodge. A New
Colorimetric Micro-determination of Sulfate Ion.
In: Automation in Analytical Chemistry (Vol. 1).
New York, Mediad Inc., 1965. p. 291-293.
3. Jacobs, M. B. and S. Hochheiser. Continuous
Sampling and Ultramicro Determination of
Nitrogen Dioxide in Air. Anal. Chem.
50:426-428, 1958.
13. Operating Instructions for the Technicon Air
Pollution Detection Systems. Technicon Cor-
poration. Ardsley, N.Y. Instruction Manual
T-67-105. 1968.
4. U. S. Environmental Protection Agency. National
Primary and Secondary Air Quality Standards;
Reference Method for Determination of Nitrogen
Dioxide in the Atmosphere (24-hour Sampling
Method). Federal Register. 38(84): 8200-8201,
April 30, 1971.
5. U. S. Environmental Protection Ager cy. National
Primary and Secondary Air Qualit •' Standards;
Reference Method for Determimituti, L "'trogen
Dioxide. Federal Register. .?#(! H : (5174-
15191, June8, 1973.
14. Morgan, G. B., E. C. Tabor, C. Golden, and H.
Clement. U. S. Environmental Protection
Agency, Research Triangle Park, N.C. Automated
Laboratory Procedures for Analysis of Air Pollut-
ants. (Presented at the 59th Annual Air Pollution
Control Association Meeting. San Francisco.
June 20-24, 1966.)
15. Nitrates. In: Technicon Autoanalyzer II Method-
ology. Technicon Corporation. Ardsley, N.Y.
Industrial Method 100-70W. January 1971.
Appendix A
A-31
-------
16. Particulate Matter in the Atmosphere — Optical
Density of Filtered Deposit. In: Annual Book of
ASTM Standards. The American Society for
Testing and Materials. Philadelphia, Pa. ASTM
Test Me thodD 1704-61. 1970. p. 498-505.
17. Burton, R. M,, W. M. Kozel, and F. B. Benson.
Development of Fine Particulate Sampling
Methods in Support of CHESS Health Studies. U.
S. Environmental Protection Agency. Research
Triangle Park, N.C. (In preparation).
18. Guide for Respirable Mass Sampling. Amer. Ind.
Hygiene Assoc. J. 31:133, 1970.
19. Ettinger, H. J., J. E. Partridge, and G. W. Royer.
Calibration of Two-Stage Air Samplers. Amer.
Ind. Hygiene Assoc. J. 31:537, 1970.
20. Knuth, R. H. Recalibration of Size-Selective
Samplers. Amer. Ind. Hygiene Assoc. J. 30:379,
1969.
21. Eaton, W. C., R. L. Penley, and R. M. Burton.
Calibration of the Andersen 2000 High-Volume
Particle Sizing Collection Head. U. S. Environ-
mental Protection Agency. Research Triangle
Park, N.C. (In preparation).
22. Burton, R. M., J. N. Howard, R. L. Penley, P. B.
Ramsey, and T. A. Clark. Field Evaluation of the
High-Volume Particle Fractionating Cascade
Impactor. J. Air Pollut. Contr. Assoc.
25:277-281, April 1973.
23. Collection and Analysis of Dustfall (Settleable
Particulate). In: Annual Book of ASTM Stand-
ards. The American Society for Testing and
Materials. Philadelphia, Pa. ASTM Test Method
D1739-70. 1970.
24. Colucci, A. V., T. Hinners, and J. Kent. Analysis
Methods for Trace Metals. U. S. Environmental
Protection Agency. Research Triangle Park, N.C.
Unnumbered intramural report.
25. Sovocool, G. W. The Origin, Detection, and
Elimination of Errors in Chemical Analysis: A
Scheme for Quality Control. U. S. Environmental
Protection Agency. Research Triangle Park, N.C.
Unnumbered intramural report. March 20, 1972.
A-32
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
APPENDIX B:
NASN LABORATORY METHODOLOGY
In estimating human exposure to air pollutants for
Community Health and Environmental Surveillance
System (CHESS) studies, data from CHESS sites are
supplemented by available data from other Federal,
state, and local sampling sites. One of the most
frequently available sources of such supplemental
data is the Environmental Protection Agency's
National Air Sampling Network. NASN laboratory
methodology is therefore described in this appendix.
SUSPENDED PARTICULATE MATTER
The weight or mass of particulate matter per
volume of air is usually determined by drawing air
through a weighed filter with a high-volume air
sampler, and then reweighing the soiled filter. Par-
ticular laboratory procedures, however, may differ
from one laboratory to another. The following
procedure is used by the Division of Air Surveillance
for all sampling conducted by the NASN.
Each city's high-volume sampler is located in the
central business-commercial district at a site that is, as
nearly as practical, comparable with the correspond-
ing sites in other cities. In such a location, measured
concentrations are among the higher concentrations
found in the city, and therefore cannot be interpreted
as the citywide average.
The high-volume samplers, operating at from 1130
to 1700 liters/min (40 to 60 ft3/min), collect
particles from about 2200 m-* of air during the
24-hour sampling period. The NASN stations are
operated during 25 randomly selected days per year.
The filters used are 20- by 25-cm (8- by 10-inch) fiber
filters selected for low and uniform background
concentrations of those substances to be measured.
To eliminate any filters having pinholes or other
flaws that could affect airflow, the filters are screened
on a light table. Prior to weighing, they are equili-
brated for 24 hours at 24 °C and not more than 50
percent relative humidity. Because any crease in a
filter can seriously affect the airflow, filters are
weighed with a balance that permits weighing without
bending.
The filters are then distributed unfolded to the
cooperating local agencies. After sampling, the filters
are folded in half with the collected particulate
matter inside and returned to the laboratory. After a
filter with the collected particulate matter is again
equilibrated for at least 24 hours at 24 °C and not
more than 50 percent relative humidity, it is weighed
to determine the amount of particulate matter
collected. The equilibration at low relative humidity
prior to this weighing is crucial, because many of the
compounds collected (ammonium sulfate in partic-
ular) are quite hygroscopic. Figure B.I shows the
effect of humidity on the measured weights of soiled
filters from three widely differing sampling locations.
The increase in weight at the higher humidities
represents the amount of atmospheric moisture
collected by the particulate matter on the filters.
ORGANIC PARTICULATE MATTER
The numerous specific organic compounds that are
removed by benzene include many of the manmade
organic pollutants. Natural organic airborne particles,
such as pollens and molds, generally are not soluble in
benzene.
Benzene-soluble organics are determined by ex-
traction of aliquots of the samples with redistilled
benzene in a Soxhlet extractor for from 6 to 8 hours.
This extraction removes more than 90 percent of the
benzene-soluble organic materials from the sample.
After extraction, .the solution is concentrated and is
filtered into a tared test tube. The benzene solvent is
then evaporated at 60 °C, and the organic matter
remaining is weighed.
B-l
-------
CINCINNATI, OHIO
ALTOONA, PA.
0 10 20 30 40 50 60 70 80 90 100
RELATIVE HUMIDITY, percent
Figure B.1. Effects of relative humidity
on weight of atmospheric participates at 24°C.
The benzo(a)pyrene (BaP) component of the
benzene-soluble organic residue is determined after
first separating the different components by thin-
layer chromatography, then removing the BaP from
the thin-layer plate and thin-layer adsorbent. The BaP
is then dissolved in sulfuric acid and its concentration
measured by fluorescence spectroscopy.
NONMETALLIC INORGANIC
CONSTITUENTS
Ammonium, nitrate, and sulfate ions are auto-
matically analyzed using an aqueous extract of an 8.3
percent aliquot of the particulate sample. The sample
aliquot is refluxed with 50 ml of distilled water in a
125-ml flask, cooled and filtered, and then reex-
tracted with 10 to 15 ml of water. In filtration of this
extract, the filter and flasks are washed to a total
volume of 50 ml of filtrate, which is then mixed and
used for the three analyses.
subsequently combined with N-1-napthylethylene-
diamine. The resulting compound is measured spec-
trophotometrically at 535 nm. This method offers
distinct advantages in avoiding interferences from
other water-soluble compounds found in atmospheric
particulate matter.
Sulfate ion in the filtrate is determined by the
methylthymol blue method, which is appropriate
only with automatic analysis techniques because the
methylthymol blue dye is oxidized by atmospheric
oxygen. The filtrate is reacted with a reagent consist-
ing of equal parts of methylthymol blue dye and
barium chloride, kept at a pH of 2.8 to prevent the
formation of a chelate complex from the dye and the
barium. Any sulfate ion in the sample reacts with the
barium, leaving an excess of methylthymol blue dye
that is proportional to the amount of sulfate present.
The pH is then raised to 12.4, at which point the
barium not removed by the sulfate forms a chelate
complex with the methylthymol blue dye and the
excess dye turns yellow. The intensity of the yellow
color is then determined colorimetrically at 480 ^m.
Water-soluble fluorides are also measured using a
fluoride selective electrode technique.1 The analysis
is performed on an aliquot of the same water extract
prepared for the ammonium, nitrate, and sulfate
analyses. A sodium citrate-carbonate buffer solution
is added to adjust the pH ionic strength and control
interferences from metal ions. The fluoride-selective
electrode generates a potential in proportion to the
fluoride activity in the sample that is measured
against the potential of a standard reference elec-
trode.
METALS
In the metals analysis, an aliquot of the sample is
ashed and then extracted with a mixture of distilled
hydrochloric acid and nitric acid. The manner of
ashing the sample is critical in the determination of
the more volatile metals such as lead, zinc, and
cadmium. At one time, samples were ashed in a
muffle furnace at 500 °C, but they are now ashed in a
low-temperature asher (at about 150 °C). After
ashing and extraction, metals contents of urban
samples are determined with an emission spectograph.
GASES
Ammonium ion in the filtrate is reduced to nitrite
by alkaline hydrazine (pH about 11). Sulfanilamide is
then added to form a diazo compound, which is
The gas samples are gathered with a bubbler-train
sampler. Air is drawn successively through a mem-
brane filter, a manifold, a bubbler, another filter, and
B-2
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
finally a critical orifice. The membrane prefilter
removes particulate matter; the glass manifold per-
mits the connection of up to five bubblers in parallel.
The bubbler itself is filled with the appropriate
absorbing reagent and is followed by a postfilter,
which serves to protect the hypodermic needle
critical orifice, which controls the airflow. Following
sample collection, the bubbler tubes are shipped to
the laboratories for analysis.
The analytical method used for sulfur dioxide is
the West-Gaeke method2 with the addition of sul-
famic acid to the method modified for the Auto-
Analyzer by Welch and Terry.3 Sulfur dioxide is
collected from the air sample by the complexing
action of sodium tetrachloromercurate and is deter-
mined colorimetrically by measuring the red-violet
color produced when a bleached rosaniline hydro-
chloride solution and dilute aqueous formaldehyde
react with the sulfur dioxide in the collecting
solution. The sulfamic acid prevents interference
from nitrogen dioxide.
Nitrogen dioxide measurements are based on the
Jacobs-Hochheiser and Saltzman techniques.4-5 When
the sample air containing nitrogen dioxide is drawn
through a sodium hydroxide solution, a dilute solu-
tion of sodium nitrite and sodium nitrate is produced.
In the laboratory, this collecting solution is reacted
with solutions containing phosphoric acid, sulfanil-
amide, and N-1-naphthylethylenediamine dihydro-
chloride. The transmittance of the resulting azo dye is
measured colorimetrically.
The analytical method for aldehydes is based on
the fact that MBTH (3-methyl-2-benzothiazolone
hydrozone hydrochloride) reacts with aliphatic alde-
hydes in the presence of acid ferric chloride. Alde-
hydes are collected by drawing air through a 0.05
percent solution of MBTH contained in a special
polypropylene collection tube. The special tube is
required because both the MBTH and the aldehydes
are unstable when allowed to come in contact with
rubber, polyethylene, or Tygon, materials that are
otherwise used in the gas sampler. The sampling tubes
contain 50 ml of distilled water when sent into the
field. After the sample is returned to the laboratory,
color is developed by the addition of ferric chloride-
sulfamic acid solution, and measured colorimetrically.
Results are expressed as formaldehyde.
REFERENCES FOR APPENDIX B
1. Frant, M.S., and J. W. Ross, Jr. Electrode for
Sensing Fluoride Ion Activity in Solution. Sci-
ence. 145:1553-1554, 1966.
2. West, P. W. and G. C. Gaeke. Fixation of Sulfur
Dioxide as Sulfitomercurate III and Subsequent
Colorimetric Determination. Anal. Chem.
25:1816-1819,1956.
3. Welch, A. F. and J. P. Terry. Developments in
the Measurement of Atmospheric Sulfur Dioxide.
J. Amer. Ind. Hygiene Assoc. 21:316, 1960.
4. Jacobs, M. B. and S. Hochheiser. Continuous
Sampling and Ultramicro Determination of
Nitrogen Dioxide in Air. Anal. Chem.
30:426-428, 1958.
5. Saltzman, B. E. Colorimetric Microdetermination
of Nitrogen Dioxide in the Atmosphere. Anal.
Chem. 2(5:1949-1955, 1954.
Appendix B
B-3
-------
APPENDIX C:
QUESTIONNAIRES USED IN THE CHESS STUDIES
The following questionnaires were used for the
CHESS studies covered in this report:
1. School and Family Health Questionnaire
(Table C.I).
2. Acute Respiratory Disease Questionnaire
(Table C.2).
3. Asthma Panel Questionnaire (Table C.3).
4. Asthma Diary (Table C.4).
5. Elderly Panel Questionnaire (Table C.5).
6. Elderly Panel Health Questionnaire (Table
C.6).
7. "Well" Panel Diary (Table C.7).
8. "Heart" Panel Diary (Table C.8).
9. "Lung" Panel Diary (Table C.9).
10. "Heart and Lung" Panel Diary (Table C. 10).
The School and Family Health Questionnaire
(Table C.I), modified for self-administration for these
studies from a standardized British Medical Research
Council questionnaire for chronic respiratory disease,
was distributed through schools in the Salt Lake and
Rocky Mountain study communities, and a similar
form was distributed through schools in the New
York communities. A modification of the form was
used in Chicago to obtain health and demographic
information for military recruits. Information ob-
tained from the questionnaire was used to assess the
prevalence of chronic respiratory disease in adults in
the Salt Lake, Rocky Mountain, and New York areas,
prevalence of chronic respiratory disease among
military recruits in the Chicago area, and prevalence
of acute lower respiratory illness in children in the
Salt Lake and Rocky Mountain areas.
The Acute Respiratory Disease Questionnaire
(Table C.2) was used to obtain information on acute
respiratory disease in volunteer families in the
Chicago and New York communities. The forms were
completed by trained interviewers on the basis of
telephone calls that covered 2-week periods.
The Asthma Panel Questionnaire (Table C.3) was
used to enroll and classify panelists for asthma studies
in the Salt Lake and New York communities. The
Asthma Diary (Table C.4) was completed by the
panelists on a weekly basis.
The Elderly Panel Questionnaire and Elderly Panel
Health Questionnaire (Tables C.5 and C.6) were used
to enroll and classify panelists for a study of cardio-
pulmonary symptoms in the New York communities.
On the basis of answers to questions on the health
questionnaire, participants in this study were par-
titioned into four panels:
1. "Well" - no to questions 3, 6, 8, 9, 10, 12, 13,
and 14.
2. "Heart" - yes to questions 8, 9, 10, 12, 13, or
14; no to questions 3 and 6.
3. "Lung" — yes to questions 3 or 6; no to
questions 8, 9, 10, 12, 13, and 14.
4. "Heart and lung" — yes to questions 8, 9, 10,
12, 13, or 14 and yes to questions 3 or 6.
The panelist then completed weekly diaries appro-
priate to their respective panels (Tables C.7 to C.10).
C-l
-------
Table C.1. SCHOOL AND FAMILY HEALTH QUESTIONNAIRE
FORM APPROVED
Budget Bureau
(CARD,, No.85-R0,33
SCHOOL AND FAMILY HEALTH QUESTIONNAIRE
FAMILY SURNAME:.
(COL. 9-28)
ADDRESS:
(COL. 29-59)
TELEPHONE:
(COL 60-66)
Please write on the lines above
your family's surname (last name),
address, and telephone number.
The information requested in this
questionnaire will be held in strict
confidence. It will be averaged for
groups of people only.
(COL 79-80)
I 0 I 1 I
C-2 HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
LU
CC
O
s
I
<
LU
X
I
O
(/}
•a
V
O
_a>
^i
n>
i
^
O
or
UJ
l—
0
0
^
o:
UJ
3:
j—
u_
CJ
yr
0
Z
o
o
3UESTIONS
i—
_j _
UJ 0
2z rL
u
c 6 „ 2 ™ n Z ' ^
LU ra - ^
LL "5 £ - U £ .. 'J
5I
CC to ,5 r- ~ 01
-^ °* _1 -1 -1
s j - « - » - -»
lt=li l i s.
S"c3°Q ° tr ™
3 5 "« *> 3 Z to
- c S-5 * s. o S t
iS - « •§ " £ p E S;
sf " ® = - ^ ® ^ - ^
^ * i °f 2* °z "^
«~ a- : ~1
Hill! It rl If
^ — "f E -- 2 "° o 2 Q. "
S ~ § S " ii 1 "° u* c" J=
^£^°-ss = •= "'z ig
s»t^i- gi ?S s »
ogl^gg o | 8< 05
Q>-(JTO>,O QTJ >LU O —
I.I
s;|
s* ?
!*»
•O -5 ff
2|S
= -5*
SJg
OjT =
« o S
lp.
w 0>S£
||;1
flfi1
LU
O
Z
s.'
I
^
^ .
1- fl
on
S:
uZ
Q)
I
«
0
"I -
"5
s
c
Is
S 0.
£ £.
— CO
~ -5
8.5
11
O o>
5o | , ~ Z
i . j
"8 „ 8 „
R " 41 i - >
2I_ " - 1
5
•8 8 .
"Ul >
i- O)
lili
_ W « 0 —
!§!! =?
" ^ f ° ro"
||sl |
jijf i
mi i
•^ § s * -?
!tlii =
>. £ c » £ >,
O - ^ ® " O
O S 3 £ B Q
CM
I
1
u '
0 '
III!
,
j ^
ff S
w __ r
i CO ^ ^ CT
CT) 0) — —
' I ! M
f 1 1 f
1 re Q. - "
, a ^ » JJ -
LT «
— O^
3 Z C
S Q o
CD £ £ e-
1 £ S S
? ° S £
o o ^ e y
0 *- z S
c - 0 c
« CO f ° £
3 >- UJ ^ O
8 : D ^ E
D O ^ 0)
0 UJ [T - S
IT LU x: £
1 ^ ||
5 j. LLJ D nj
to! 0
Appendix C
C-3
-------
oc
<
LU
O
Q
o
O
1
c
O
OJ
-li
!M
HI
« in
-~°
i bj£ 1 s 5
_ ZD §»=
O°?D° sJ^i
-, $ s 1 S = * 1 8
s°
-£ E 5
3 - .*
S|6
C-4
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
LU
DC
\
LU
O
X
LU
I
O
o
o
•a
> > ~
1
5 o> g »
" o _ X
c * O
o ° to
1. , I ,
c
- CD "
" ill
8 ^ „- „; g
L- -'
I
U ,_, LJ ,_
in CU o O
" ill
o . >
- -
TO
R g O £
-- - " -
I
J I I §„
Z > > > ~
- - - —
I III
1 Has your child been treated by
a doctor since September 1967
for pneumonia7
1
5 O> Q d>
1 * l ^«
u w « w ^
1
" III
8 „- „- „- 8
C
I S 1 !„
c
j I j jg
^ " — -J
c
S S " S
g J 1
0 . g
L
c
c
-J § 1 iw
z >>•>--
c
8 ^- " ^g
_
2 Has your child been treated by
a doctor since September 1967
for an attack of croup''
1
S o S £
0 ° " Q>
u w" w « <•>
- o « w ai 5
c
^:
° ill
o . i1
' 1
c
(0
" ill
O - °>
C
I 1 I i a
"" 0 01 1? IU j
L-
S a) J 0)
0 " * "
_
1
8i g o J
8 § !- "s
^Z 1 1 1 *
1
c
8 l l ^
z > > >
— •- -
1
o § ^ 1.
L
3 Has your child been treated by
a doctor since September 1967 for
an attack of bronchitis, bronchi-
olitis, or other deep chest in-
fection''
5 01 S fi>
* ill
o o
Z > >
1 j J
c
1 § 1 i ,„
z > >- > -
c
! ! ! i«
>
c
S (u g
^
1
s S S »
?0 !•!• ^
-
c
J g 5 E
Z > 1 > *
.
-I = * C
8 °- " «
_
4 Has your child been in^he
h_o_sp_tta_l since September 1967
for one of the illnesses mention-
ed in questions 1. 2, or 3 above''
™R
J
»
o
i
o
S
S
J
O
3
sa
a
o
.
-1
o
5 What is the full name and address
of the doctor who takes care of
chitd to a clinic for medical care
3
n
n
i
i
3
i
S
i
'. f
n E
i I
Appendix C
C-5
-------
DC
<
a
i
LU
I E
§ °
n in
D Q g 5 6
S S
O CO
IX CL Q Q Q. O O
8 S S
c 8 ?
JT ° "
<1) 3 ™
I 1 »
8
O
V)
0)
I
o
— I
.« i
z o s o
C-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
LU
GC
O
UJ
O
UJ
I
d
i
o
•§
o
o>
8£
to — f.
1:1 -s
ii i o
— ra w o
-1«
X £ "6 ^ >-
!@ i
D
[S]
§
0
0
Q
n
u
n
D
U
D
D
Appendix C
C-7
-------
cc
<
CO
LLI
O
LLJ
S3
LU
LU
a:
a.
CO
O
.a
(B
LU
Z5
O
a.
LU
_J
LLI
o
cr
j
^
~
;^,©c^
(o) ^1 (TM) (r-7l '>*.~\ -.^V ' " ' Jl'
' i ' , i , <•-, fS)
(,,-, (0 (^ © ») © © (r~; © ©
t/5
a
5
cc
cc
CO
z
z
I
a.
LU
2
z
O
0
cc
LLI
a.
w
a
D
D
8
"^
u
a
c
"o
a
» - S
" 0 C
•a 2 2 -^
O o > £ o
•a 2 3 £
_
E
o
0)
5 »
> y
"D ^
C ""
.g >
O £
if sel
-) S "* * C
•= 32
Ul ™
"• S a£ ° E
o SboN
^.
LL.
^ 3
£ = * g
ra -O c =
CO
CJ
G
G
O
©
0
©
O
©
O
0
0
0
0
©
©
e
0
©
©
0
0
©
©
©
0
0
©
©
©
©
0
©
O
O
0
O
©
O
O
o
0
©
©
0
©
©
0
©
0
©
0
O
0
O
o
0
©
G
©
©
©
0
o
0
0
©
c
3
G
G
0
O
G
G
0
©
©
©
©
0
0
©
©
©
0
0
0
©
©
0
©
©
©
©
O
(3
O
0
O
©
O
0
O
0
©
0
0
©
©
0
©
©
©
0
o
©
©
G
Q
0
Q
©
0
©
©
©
©
©
0
©
©
O
o
o
o
0
0
©
0
©
©
©
©
0
0
©
o
©
Q
Q
0
O
©
0
©
0
0
O
o
©
©
0
o
0
o
©
o
©
G
©
©
©
0
0
©
©
0
0
©
0
O
©
O
©
©
©
0
©
(3
O
G
©
O
O
O
O
o
0
0
0
©
©
O
0
©
©
©
©
O
©
0
©
©
O
o
o
©
©
o
o
o
o
G
O
G
0
G
0
G
(3
G
0
0
O
©
G
©
0
O
0
Q
O
0
©
O
©
O
©
0
©
G
(3
G
G
G
G
G
O
G
c;
G
(3
G
G
©
0
0
0
O
O
O
O
G
G
©
O
0
0
0
©
,, ' . , i, . i r-, • I . r
>«, - , ., ,/.,tv • i, ,; ,o; ?,
G
'G
G
O
0
G
0
G
O
0
G
G
O
G
G
0
0
G
0
O
O
0
o
o
o
G
O
0
O
0
0
G
G
O
G
0
G
G
(3
O
G
©
G
G
G
G
©
©
G
Q
0
O
O
O
o
O
©
0
O
O
O
O
;3
G
G
3
G
G
Q
(3
G
(3
^
G
'3
G
G
O
0
0
0
G
G
0
O
0
O
G
O
O
O
0
O
(3
o
o
0
O
*_J
G
G
G
G
O
(3
(3
G
G
O
G
G
0
O
0
O
O
0
o
©
0
O
O
0
0
G
G
O
G
O
G
G
G
G
G
'}
G
G
G
G
G
O
0
0
G
O
O
©
©
0
0
O
©
O
©
O
vG
©
O
G
©
©
(3
G
0
'3
'-'
©
0
O
G
O
©
©
©
©
O
O
O
0
0
0
0
0
0
0
o
"/
' )
o
G
O
G
O
o
©
G
^
G
O
0
f~*\
O
0
G
0
0
O
0
O
©
O
0
O
©
O
©
G
C-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table C.3- ASTHMA PANEL QUESTIONNAIRE
ASTHMA PANEL QUESTIONNAIRE
OMB No. 85-570028
(CARD 1)
I.D. #_
Asthma Subject's Name
(COL 13-441 (LAST)
Respondent's Name
(ENTER ONLY IF RESPONDENT IS
OTHER THAN ASTHMA SUBJECTI
COL 179 -801 01
(CARD 2
Subject
(COL
Subject
Subject
Subject
Reason
Intervie
Date of
's Address
13-56) IHOUSE NUMBER) (STREET)
ICITYl (STATE)
's Telephone Number
(COL 57-631
plans to move before next March I I Yes 1 ) No
1 COL 164) 2
agrees to participate in panel Q Yes l~] No
' COL (651 2
for refusal
ICOL 66-671
wer's Name
ICOL 68-69)
interview j | |
MO DAY YR
(COL 70-75)
IZIPI
rn
COL 179-80] | 0 | 2
Appendix C
C-9
-------
Table C.3. (continued). ASTHMA PANEL QUESTIONNAIRE
(CARO 3}
1. IDENTIFYING INFORMATION
1 Current aae
COL 113
2 Year of birth
•
COL 120]
4. Race Q Indian
MI t
r~] Mexican -American or
2 Spanish-American
[~~1 Negro
COL (15 18) 3
COL 19)
3. Sex- Male
' D
Female
'°
II. ASTHMA HISTORY
1 . How old were you when
(ENTER AGE
2 How often have you had
the past year?
| [Oriental
I (White
5
Q] Other
6 (PLEASE SPECIFY
your asthma first began'
N YEARSI
asthma attacks within
3. When you have an asthma attack which of the
following symptoms do you usually have'
a. Shortness of
b Wheezing in
c. Fever
breath
the chest
d. Increased sputum or phlegm production
4. Do your asthma attacks
keep you from school or
your usual activities'
ever get severe enough to
work or from carrying on
COL .21 22
COL 123
QNone
D1-4
2
D5-9
3
| 1 1 0 or more
4
QYes Q No COL (241
1 2
niYes Q No COL 125)
1 2
QYes dN° COL ii
1 2
Q Yes Q] No COL i »
I 2
QJYes Q] No COL 128)
1 2
C-10
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table C.3. (continued). ASTHMA PANEL QUESTIONNAIRE
5. Da you have asthma attacks in"
Winter (Jan. -Feb. -Mar.)
Spnno (Aor • May - June)
Summer (Jul - Aua -Sept.)
Fall (Oct • Nov. -Dec.)
6. Do many of your attacks occur in
Winter (Jan. - Feb. -Mar )
Snrmq (Apr. - May - June)
Summei ijul • Aua. -Sept.)
Fall (Oct Nov. Dec >
No seasonal predominance
7. How many cigarettes do you usually smoke now?
Never smoked
Ex -smoker
Less than '; pack per day (1-fi cigarettes)
About Vi pack oer day (6-14 ciqarertes)
About 1 nack oer day (15—25 cigarettes)
About 1V, packs per day l2fi-34 cigarettes)
About 2 or more packs per day i35 or more cigarettes)
8. At your job are theie fumes or dusts that tend to bring on
asthma attacks'
9. Did you ever have a skin rash or hives after taking drugs
or after eating certain foods?
CDYes
jJYes
[>»
DYes
O»
d>-
Qres
DY-S
QYes
DN°
DN°
DNO
DNO
2
DN°
2
DN°
2
Q|No
2
QNO
2
DN°
2
COL 1291
COL (30)
COL 1311
COL 132)
COL (331
COL 1341
COL 13SI
COL 1361
COL 137]
CO 18
D
D
D
4
D
5
6
D
7
DYes
1
QYes
i
QNO
2
QNO
2
COL 1391
COL 140]
Appendix C
C-ll
-------
fable C.3. (continued). ASTHMA PANEL QUESTIONNAIRE
10. Do you tend to get asthma attacks:
a. When it aets unusual Iv cold outside?
b. When you get respiratory infections, like colds?.
c. When you get upset?
QJYes
1
QYes
QNO
[ |No
COL (41]
COL i«i
COL (43)
11. Do you have hay fever?
£]NO
GENERAL HISTORY
1. How many rooms are there in your living quarters? Do
not count bathrooms, porches, balconies, foyers, halls,
or ha If rooms.)
COL 145- 46)
QOne nsix
1 6
[~JTwo r~|Seven
2 7
Q Three QjEight
3 8
QFour QjNme
[^Eleven
11
FHTwelve or
»2 more
2. How many people live in your household?
COL 147-48)
QjOne | [Seven
01 07
QJTwo | [Eight
02 08
QJThree [ [Nine
03 09
| [Ten
10
[ [Eleven
11
QjTwelve
Five
05
Q
06
3. What educational level did the head of your
household complete?
COL (49)
[^[Elementary school
1
Q]Part of high school
2
Q High school graduate
3
QjPart of college
4
QCol lege graduate
5
Graduate school
Other
(PLEASE SPECIFY)
COL (79-801 |0 [ 3_]
C-12
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
ro r—
T3 ro
o z
s: o
IO
00 i
oo oo <
LU LU :
?- >- I
a a a
ro
-o >,
LU ce LU
z o z
o E: o
CO CO et
O LU LU I
Z >- >- r
nan
3 O1
X3 CM
CO
LU
UJ CC LU
z o :
o E; o
co co ->->—
n D a
rOCO
-OCM
00
£>,
-D 3
CU T3 I
11 i ry
Z O Z
o 2: o
(/) CO cC
nan
§
6
00
EC
z
LU
a.
LU
Q
<:
LU
n:
rOCM
-O
oo >,
CO^-
3 3
LU
LU o: i
z o z
o 2: o
CO OO - >- I—
ana
CO
cu
ro
-o
S-
co
ro
CU
VD
>i<
ra
-o >,
Ci—
O 3
UTS
LU
111 ry 11 i
Z O Z
o s: o
CO CO eC
O LU LU 3:
Z >- >- I—
a a a
LO
>>CM
-O >,
00 "73
LU CC I
Z O ^
O 2! O
OO OO =t
O LU LU T
Z >- >- ^
ana
cu
ra
c
ro (
CU ro
-O
O
3 CU
o -o
>> o
U)
-a •!-
••- a.
Q CU
0
LU
CO
OO
z
o
\~
~3
o-
LU
I—
o:
LU
OO
z
LU
CO
«t
LU
O.
g
OO
1 — 1
Q.
uu
a:
E:
i—
••j^
at.
Z
"'•
LU
a:
Q
1— H
Q
— )
o
•—
z • 2:
s: CM E: CM
=c T Q. T
ID 2: 10 E:
i i -
2: =C Z O-
CM CM
r— ID i— IO
n n n n
z • s:
E: CM E: CM
e£ T 0- T
vo 2: ID E:
1 1 •
s: =c z o.
CM CM
i — (D i — *JD
n D D n
z • 2:
E: CM 2: oj
et 1 O- 1
CO
ro -O
JC O
2 oo
1 < e-i
•=C CU
CO
O LU
Z >-
n a
CO
CD LU
Z >-
a a
CO
O LU
Z >-
a a
CO
O LU
Z >-
n a
00
O LU
Z >-
a n
O LU
•z. >-
a a
oo
O LU
z ^
a n
OO LL.
0 LU 1-1
Z >- 0
1 1 1
a a ft
j
CO
O UJ
z >-
n n
CO
O UJ
z >-
n n
CO
O LU
Z >-
n a
CO
O LU
z :>-
a n
CO
O LU
Z >-
n a
i
C/l
O UJ
^ >-
a a
CU CO
E -c
O 4-J
^:
c
s- co
3 -C
O S P-
>, c
c o en
•r- O CO
-c ja
3 S-
o o co
>^Xi "O
S^ 0
cu en oo
1- T- -r-
O) QJ CL
3 C CU
oo
O LU
Z >-
a a
i— en
ro c
E E-r-
oj o s- a.
s- s- o cu
cu v«- c cu
> i—
CU 3 S- 00
00 O 3
>> 0 E
CO >, O
-D CL S-
O CO "M-
00 CO i —
•^ ^ O S-
O- O O
CO O _c r>.
•!-> O >,-(->
CO 00 4-> J=
-c ^: -i- en
•4-J en « > T-
3 J^ -i- C
00 O S- 4->
rO C O O 4->
3 CO 5 ro ra
OO LL.
O LU i—i
z >- c_>
LU
a D ft
CO U-
O LU 1—1
Z >- O
LU
a n ft
OO U-
CD LU 1— 1
Z >- r_>
LU
D a ft
OO LL.
O LU »— i
Z >- CJ
n n ft
>-
CO LL.
C> LU I-H
z: >- o
UJ
a n ft
>.'
CO U_
O LU i—.
•^ >- <__)
aa ft
ro >,
°if- **"
15 O O
O CO
>, CO Q.
oo oo
4-> 3 r>.
ro ro CO CU
-C U -O >, >, CU
>, C ro 10
CO "O <*- 3
O T3 O 1-1 ro
3
cr.
ra
-D
c:
3
OO
ro
CO
c:
ro
o
a.
Appendix C
C-13
-------
Table C.5. ELDERLY PANEL QUESTIONNAIRE
ELDERLY PANEL QUESTIONNAIRE
OMB No 85-570028
I. D.
Subject's Name
(COL 13-44] (LAST)
Respondent's Name
(ENTER ONLY IF RESPONDENT IS
OTHER THAN SUBJECT]
Subject's Address
COL (13-561 (HOUSE NUMBER]
Subject's telephone number
(COL 57-631
Subject plans to move before next March'
Subject agrees to participate in panel-
1 COL 164) 2
No
I COL 1651 2
Reason for refusal:
Interviewer's Name'
COL (79 -801 0 1
Date of Interview
MO DAY YR
COL [70-73!
COL I79-80I 02
C-14
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table C.5. (continued). ELDERLY PANEL QUESTIONNAIRE
(CARD 31
IDENTIFYING INFORMATION
(CARD 3'
IDENTIFYING INFORMATION
1. Current age'
2. Year of birth'
COL '1 'i
3. Sex Q] Male
i
I i Female
4. Race
| I Indian
i
£jj Mexican-American or
2 Spanish -American
n Negro
3
| [Oriental
4
Q White
5
[^Other
6 IPLEASF SPECIPYI
5. How many cigarettes do you usually smoke now?
Never smoked
Ex -smoker
Les? than "4 pack per day (1 -5 cigarettes).
About '/2 pack per day 16-14 cigarettes)
About! pack per day (15 25 cigarettes).
About 1'/2 packs per day (26- 34 ogarettes)_
About 2 or more packs per day (35 or more cigarettes)^
D
i
D
2
3
D
4
D
3. At your job, are you now or have you been frequently
exposed to irritating smoke, dust, or fumes? (Do not
include neighborhood or home exposures.)
IF YOU ANSWERED "NO" TO QUESTION 6, SKIP THE
NEXT THREE QUESTIONS BELOW.
ia. If the answer to question 6 is "yes", what kind of
irritant were you exposed to? (For example' coal dust,
cutting oils, asbestos, mine dust, smelter fumes, raw
cotton dust.)
I COL 25|
Appendix C
C-15
-------
Table C.5. (continued). ELDERLY PANEL QUESTIONNAIRE
6b. If the answer to question 6 is "yes", what kind of
work did you perform in this job? (For example-
miner, maintenance, assembly line, supervisor.)
6c. If the answer to question 6 is "yes", how long were
you exposed'
Less than 1 year
1 to 5 years
6 to 10 years
More than 10 years
III. GENERAL HISTORY
1. How many rooms are there in your living quarters? Do
not count bathrooms, porches, balconies, foyers, halls,
or halfrooms.)
8. How many people live in your household?
9. What educational level did the head of your household
complete?
n
'COL 21)
n
t
n
2
D
3
n
4
1COI 22 23) 7
,—1, Seven
QjOne "-^
D tight
QThree LJKwe
r-A | [Ten
| [Four ' — '
D^ | [Eleven
Five ' — '
rA I (Twelve or
Six 1 — 1
I — 1 more
'COL 24 25
LH <~*1e | | Seven
01 07
[^] Two [ |Eighl
O 2 08
PJ Three 1 |Nme
03 09
r~| Four | [Ten
04 to
j^JFive [ [Eleven
05 11
r~jSix r~jTwelve or
oe 12 more
COL (26)
[^Elementary school
1
QjPart of high school
2
j^JHigh School graduate
3
QPart of college
4
[^College graduate
5
[^Graduate school
6
| [Other
7 (PLEASE SPEC IFYI
COL (79 80) |0 | 3|
C-16
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Table C.6. ELDERLY PANEL HEALTH QUESTIONNAIRE
HEALTH QUESTIONNAIRE
Study
I. Name
ICOL 13 331 IFIRSTI
Address
ICOL 34- 771
FORM APPROVED
Budget Bureau
No 85-R0133
I COL 79-80) |o | 1 |
1. Do you usually cough first thing in the morning in winter?
2. Do you usually cough during the day or night in winter?
If you answered "Yes" to question 1 or 2, please answer question 3.
Yes
Yes
D'
3. Do you cough like this on most days or nights for as much as three months Yes
each year? | |,
No ICOL t3l
o
NO ICOL Ml
n*
NO COL15I
|z
4. Do you usually bring up phlegm from your chest first thing on getting up in Yes No
the morn+ng? [~~|i |~~|;
5. Do you usually bring up phlegm from your chest during the day or at night Yes No
in winter? |~~|T [ \2
If you answered "Yes" to question 4 or 5. please answer question 6.
6. Do you bring up phlegm like this on most days for as much as three months Yes No
each year? [ ]i [ [2
7. Have you ever had heart trouble?
Yes
8. Have you ever had pain or discomfort in your chest brought on by exertion Yes
9. Have you ever been told by your doctor that you had angina, or angina Yes
pectons? r~]i
0. Have you ever been told by your doctor that your heart was large?
11. Do you get short of breath walking with other people at an ordinary
pace on the level?
12. Have your legs ever become swollen?
Yes
Yes
// so, what was the cause of the swelling?
13. Have you ever had a heart attack?
Heart trouble
Lung trouble
Kidney trouble
Varicose veins
Other
4. Do you take any medicine for your heart or to help you lose water?
Yes
Yes
No
o
No
No
No
Yes No !COL 23)
D' D'
No ICOL 241
No ICOL 261
D. Q
No ICOL 271
Appendix C
C-17
-------
Table C.6. (continued). ELDERLY PANEL HEALTH QUESTIONNAIRE
15. How old are you?
16. What is your sex?
17. How many years have you lived at your present address?
ICOL 28- 29)
Years
Male Female ICOL 3<»
D< D>
.Years
18. How many years have you lived in your present borough or county?
19. What is your home telephone number
ICOL 33-34!
. Years
Thank you very much for your cooperation.
C-18
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
o
=3
O
z
Q.
CC
O
Q.
ce
<
Q
LLJ
O-
tu
a
UJ
>-
D
o
z
D
o
z
D D
D
D
r>
6
_OJ
.0
<0
a:
o
o
o
o
a;
UJ
UJ
^
LU
X
1-
0
z
K
3
Q
t
^
O
a:
x
•
UJ
Ef.
s
o:
o
a
d
u
X*
o
o
o
(J
3:
UJ
z
UJ
>
<
X
o
o
>-
a
a
^
o
z
_i
UJ
LU
U.
UJ
o;
LU
^
3
O
>-
LU
CO
13
<
U
UJ
to
n
111
I/)
X
h~
D:
o
0
<
a
in
^~
CO
>
O
>-
UJ
>
ri
CO
LU
}_
D
0
z
D
t*-
^
LU
UJ
3:
CO
X
1—
-J
<
H~
a!
CO
o
X
<
£
o
LU
1=
i!
<
g
>•
LLJ
0£
UJ
*
c>
o
o
U-
o
>-
a
<
x
I
a.
u
x
UJ
5
X
t-
3
O
UJ
_i
X
a.
LU
X
<
a.
UJ
I
a.
I
UJ
UJ
UJ
£ O
o >
g *
LU 1U
UJ X
X l-
u
Ul
ui
>•
H
<
CO
5
u_
CK
3
X
t-
O
LU
*
CO
UJ
3
1-
1
1
CO
Z
a.
to
UJ
5
O
N
LU
UJ
I
5
UJ
_l
x
a.
g
Appendix C
C-19
-------
cc
•f,
a.
<
Q
<
a.
cc
<
00
6
3C
K-
U-
CD
LU
cc
o
5
a
h-
vi
>-
o
a:
u.
•^
o
a:
X
o
*/>
o
^f
o
LU
•~
5
O
d
2
5
o
3
V)
Z
LU
Ml |/)
5> uj
Ql QC
2 a °-
* LU LU a
I VI m 1- LU
o a: 3 i- >
ID O < LU UJ
2 S vi 05 Z
a D a n D
z
LU
£ 13
a; Q;
2 o: °-
* ui uj a:
5 vl ui H UJ
o: s i- >
o o 5 LU LU
5 S
(0 UJ
of a:
O Q, a.
ui uj o;
U 2 ^ »- >
r) O 5 uj uj
2 ^ (/> CO Z
a n a D a
,_
z
Ul
UJ to
VI LU
a: o:
9 K °-
LU ui o:
X UJ 1— LU
U o; 5 i- >
D O < UJ UJ
2 ^ CQ z
D D D D G
\-
z
UJ
LU
to Ul
a: o;
O -v a.
fe Q.
^ uj uj o;
U D; ^ H >
ID O < UJ UJ
2 ^ v> ca Z
D a a n a
i-
z
UJ
UJ to
10 UJ
o; o:
o 0,0,
* ui LJJ or
X tO UJ 1— UJ
U Or 5 H- >
D O
Z) O -< LU UJ
5 S i/> DO Z
D D D D D
-si
o I- , to ^
o ^ :£ ' n
>-» _ i/i
"~ fQ V) CZ
. H o 2
,-« ja * -.5
LU uj
UJ ^7
O §
. .z
O m LU I
z >• > i-
ana
LU w
UJ EV y
,z
VI l/l <
Q UJ UJ X
z >- »-
a D a
UJ uj
ui i^ z
O Z?
*/r 1/1" •<
O UJ ui x
Z >- >- h-
D D a
Ul ui
uj ce z
O 2 ^
VI (/>" -^
O UJ LU X
Z >- >- )-
D p n
UJ ui
UJ &. "7,
O 2
, .z
O LU LU X
Z >- >- I-
n n n
UJ uj
uj o; -r
o 1°
. . z
O UJ UJ X
Z >- >- h-
nan
UJ Ml
UJ &. -r
§ i°
- ,z
O LU LU X
Z > >- 1-
a p D
C t3 iS
•S"S.E
J^ Ci3 "3 f1"'
-, I- TJ £
0 -**- cr o
> »/i ex »-r
ro -ic _t_, CM
31 *-* (/i *-.
_£J &j to
CM PJ U CJL
h-
5
LU v)
w> Ul
Q£ Qi
O Q; 0.
* uj LU a
I tO UJ H UJ
U CC 2 1- >
13 O ^ Ul UJ
2 ^
n o 5 ui uj
2 S «rt CO Z
p a p p p
Z
UJ tO
t/) UJ
cc a:
g a: °-
* uj uj o;
X vi uj >— UJ
U a: ^ »- >
13 O < UJ UJ
2 * t/i CQ Z
p a p p p
t-
z
UJ
LU O
LU
o; a:
O |V £*•
* uj LU o;
X VJ UJ 1- Ul
U oc ^ i- >
23 O < UJ UJ
2 3: £ CQ Z
p n p p P
H
Z
UJ
Ul to
v) UJ
a: n:
O « 0-
5? Cc
* uj ui a:
X vi ui 1- UJ
U Oi. 2 h- >
ra o < LU LU
2 ^ vi ca z
P P P P P
t-
z
LU
UJ tO
to UJ
fy Q;
* UJ LU Of
X i/> UJ H- UJ
U CK ^ h- >
Z5 O < LU UJ
2 ^ t-o 03 Z
P P P P P
1-
Z
UJ
LU vi
to LU
a: o;
g a °-
^ m LU a;
X to LU H UJ
H) O < UJ UJ
2 * «/> cfl Z
P P P P P
G E3
.5 j= cu r^.
M =j "^ ">
c: |-: a> n
^ >»^ =
"*** O) .£
— *- ^ S
n •- •"
• O ° k.
00 >^ == o
Ul
cc
n:
I—
ce
o
a
z
=3
UJ
_J
a.
O
a.
>-
D
>-
a
* a
>-
D
<-- n
^e LJ
O
z
a
M ~
H ^
55
"
=" * ^
5 <« 2
o " -a
ce £ «
II I c= T3
r^
-------
>-
0
tt.
LU
O
31
<
Q.
Oi
6
_03
J2
(D
LU
3:
o
>-
a;
o
re
o
•f
LU
CC
O
LL.
? a.
<
z _•
>-
<
o
cc
13
t-
<
V)
>-
•^
Q
a
u_
>-
<
Q
i/)
or
o
1C
K
>_
^j
Q
LU
O
S
>-
^
O
v>
LU
3
H
>-
^
Q
~
*
"~
Q
•%.
ID
ij—
Z
UJ
LU vi
v> LU
O DC OL
LU UJ O;
X »O UJ *~ UJ
(j a 5 H >
i g 1 S g
G G G a a
t
£-
LU
w 12
VI UJ
a oc
O a a.
^ MI UJ OH
6 £ S E >
D O 3 uj uj
2 * Co ttJ Z
G G G a a
H
z
LU
(/> Ul
tt a
° a "•
LU 111 K
X v> uj t- UJ
U a -y t- >
3 O 5 UJ LU
2 f. v> CO Z
G a G a a
*-
z
UJ
UJ VI
V) UJ
ce a
O — CL
i
uj uj Q:
X i/) UJ h— LU
u o: 5 v- >
13 o < LU uj
S ^ vi CQ Z
G G G G G
t-
LU v;
v) LU
LT CK
g K °-
* uj uj a
u a; ^ (— >
ID O 5 LU UJ
2 * vi to z
G a G a G
»-
-£
UJ V)
o: a
0 or CL
f LU UJ Q;
*T* vi in (~ m
u a: s i- >
^ o 3 uj uj
^ =z 3 ca %
C G G D G
»-
z
UJ
UJ lO
VI LU
D; a:
g a Q-
UJ UJ CK
X t/> iii 1 — UJ
u a ^ i- >
IJ O 5 UJ UJ
3 s 5 m z
a a a D a
E
bo
QJ
— *"
°-"S fe ^
»- ^ to —
o c "" ro
._ ~ "J
u a: s H- >
3 O < UJ UJ
5 ^ i^ CO Z
a G a G a
t-
z
Ul
Uj to
So uj
S a
0 a: °-
* m uj o:
I to UJ 1- Ul
U C£ 5 »- >
13 O < UJ UJ
S S tn m z
G G a G a
i~
z
UJ
UJ vi
(/> LU
a a
0 a Q-
* UJ UJ Q;
X "0 uj H- UJ
U a: 5 H >
3 O < UJ uj
S * v> m z
G G G a a
»-
z
LU
III VJ
V) LU
a a
O o: °-
* uj uj o;
5 S ^ t >
rs o < LJJ uj
5 * ^ 03 Z
G G G G G
t-
z
LU
UJ *"
a o;
g K °-
* uj uj a:
X v> uj t— LU
u or 5 H- >
Z) O 5 LU LU
2 * 03 2
c G G a a
t-
z
UJ
UJ i/>
V) UJ
a: o:
0 K a.
* LU LU a
X LO uj t— LU
u a 5 K- >
ID O < UJ UJ
S * vi CO Z
G G G G G
»-
z
UJ
LU V)
V> UJ
Ct QL
o a °-
* uj LJJ a
X V> LU I" UJ
U t¥ ^ h- >
13 O < LU LU
5 *: VI CO Z
a a a G a
3 m
o - *;
>-, OJ &
C £ n
.E 5 -a
w> 3^ o
c _n:
"rsi o a>
_
^ >-. r~-
rj p> "
QJ -o -c: ^
-^ 5 *^ w
^ ^ " a
— ^ - ^
• J^ ° Jc
(XJ O =- *-
l~
Z
LU
io LU
K Ct
0 a Q.
LJJ LU o;
X V> LU *~ UJ
U a: 5 t- >
13 O < UJ UJ
2 * v> CO Z
a G a G G
t-
z
UJ
UJ" t/1
t/1 UJ
a a
g a *•
* uj m o:
X V) m h- UJ
u o; 3 >- >
3 O 3 "J uj
Z * vi m z
G G G G D
i-
z
UJ
Ul to
t/> Ul
a: Q;
0 a o.
* uj uj a
X *^> HI ^~ UJ
U K s h- >
2 o 3 uj uj
S J: t^ OJ z
a a G a a
H
z
UJ
UJ to
v,- uj
a: a:
o a a.
* UJ LU CC
5oO uj ^~ UJ
D± S >- >
O O rf UJ LU
2 * Co cQ Z
G G G G G
K-
z
LU
UJ f
a a:
Q a: a
LU uj o;
X vo m t — UJ
u (x. -s *- >
r> O 5 tu uj
2 * Co m z
G G G a a
H-
z
LU
LU vl
v) UJ
o: o:
0 a °-
* uj uj a:
x vi uj *"• uj
U a: ^ i- >
;D o «f LU LU
s * v> m z
a G G G a
i-
z
UJ
UJ V>
v> UJ
o; cc
0 a: ^
^ LU LU Q;
X vi uj 1— LU
U o: s i- >
H> O < UJ LU
2 * v) aj z
a a G G a
aj~
"5 ^
"> §
v» •*• i_
°> JZ 0 "^
o E 6 2
"1^ >-> V> ^
»- -S o> *=
n ^ r- "J
o o i^ -^
>-i _ -*-1
— ^ s s
. I" 0 'S
^o ja i: ja
o
.J
D
s n
I—
z
LU
LU
O
a 2
n
o
a i
D
S 15 -5.
(T3 *-• .—
z:
-------
DC
<
Q
<
a.
Q
o=
i
^.
O
v>
z
o
UJ
;w
:£
<
a
V*
UJ
13
h-
>-
•<
O
•z
o
£
>•
2
£
v>
V-
z
UJ
UJ (/>
»/J UJ
Of OE:
g Of a.
* iu uj a;
x \n uj t- uj
U a: ^ »- >
O O < UJ LU
2 * £ 03 Z
DDDDD
t-
z
UJ
LU t/>
»/) Ul
o: of
g a; &•
* UJ UJ Of
X UJ t~ UJ
U of 5r »- >
§O < LU UJ
* oo CO Z
DDDDD
i-
•z
UJ
ui y>
ft UJ
K o:
g a °-
* ui ui o:
5S) uj H "J
Q: 5j H >
D O 5 UJ Ul
s S 3 to z
D DO DD
i—
~p_
HI
UJ i/>
i/) in
CC Of
2 a: o-
UJ UJ Q;
5 S S t >
3 O 5 Ul ui
S S « to z
D D D DD
l-
z
UJ
LU
r3 o 5 LU LU
S * w ca z
DDDDD
H
Z
LU
LU O
t/> Ul
(V y
g & o-
* uj uj o:
X ui uj t- IU
U o: ^ j- >
i i 1 S S
DDDDD
H
z
Ul
UJ t/>
t/> LU
Of Q£
g Of 0-
* 111 ui o:
X vo UJ 1- UJ
u a: s t- >
i i 1 S S
DDDDD
W JS
V> 0 <=
S = ., 2
f^r
j5 ra - J>
s-ss«
*3 *^ 5 rQ
OJ= * *-
3^% «j O ^^
J2|Sg?
.^§13
•-H o S vi :i
UJ Ul
Ul Of 2
§ i°
.2
to" to •<
O Ul UJ X
2 >->-!-
D DD
Ul Ul
uj u: z
§ i°
. s-z
to to •<
O LU LU X
Z >->-»-
D DD
UJ uj
LU tKv-
o i°
?
wi" v> ^
O LU LLJ X
Z >- >-J-
D DD
Ul uj
LU o; 2
Z O o
O S
, -2
*/> - >-l-
D DD
LULU
LU Of ^
^r o n
0 §°
- -3
«/) to - >-H-
D DD
LU m
iu a: -2
Z O o
0 5
0 ,2
t/>" */l -<
O UJ LU X
Z >- >-H-
D DD
Ul ui
UJ tt-r
Z OQ
O 5
. -Z
WJ U) <
O UJ UJX
Z >- >-l-
D DD
>»
^.i^?^
•^ y™ «>
„ c ,-t ro
o •°S'Q-
J^° S^t;
" •••fi*' ?!
•S-3°E?o
3= n = ~ •«=
^-^2-gS
t—
z
Ul
LU
i>0 UJ
o: oz
Q Q, Q.
^ UJ LU CC
X VI m I" UJ
U a: 5 i- >
o o 5 uj uj
2 * 3 CD z
DDDDD
t-
z
UJ
Ul O
t/1 UJ
D; &.
g a: o-
* Ul Ul CC
X 40 Ul H- Ul
U QC 5 1- >
3 O 5 UJ UJ
S * S 00 2
DDDDD
K
Z
IU
LU V>
t/1 UJ
o; of
g o- Q.
* Ul Ul Of
5V) ui )- Ul
a: S >- >
=1 O < Ul Ul
S S S CO 2
DDDDD
t-
z
Ul
III I/I
CO UJ
o: CK
g a °-
^ ui ui a;
5V5 UJ ^ UJ
o: 5 H >
3 o 5 UJ uj
5 * 10 CO 2
DDDDD
l-
z
UJ
Ul */>
VI LU
o: S
g a: °-
* uj LU a;
X (/l m *~ UJ
U K 5 H >
r> o 5 iu LLJ
5 * v> ca z
DDDDD
i-
z
LU
UJ t/>
«X> Ul
a o;
g a °-
ui LU o;
5w ui *- "JJ
Of 5 I- >
^D O 5 UJ UJ
S 3= 3 03 Z
DDDDD
t-
2
UJ
Uj
V) UJ
a cc
0 a- o.
* in ui Q;
I Ul UJ 1- Ul
o o; 5 i- >
i i S S S
DDDDD
M «? S
.E&gS
~ -o ° U
*o> 2
*« S 5E:
M £ g « ™
£ .. 2 F S
t, 3 5 3
- S.-5 "• ts
. = |5S
ro — t s3 +-t
tr-
z
UJ
Ul t/I
v) LU
o: oi
g a °-
* ui ui o;
5W> Uj h- Ul
OC 2 H >
ID O 5 UJ UJ
2 * t/i m z
D D D DD
t-
z
Ul
Ul u*>
Ul Ul
cc a:
g o: o-
f ui • ui o:
X ui ui t- UJ
O CK S 1- >
13 O < Ul Ul
Z * Ul CO 2
DDDDD
t-
z
UJ
LU ">
vi UJ
Of Of
g a "-
* ui ui o:
510 UJ >~ Ul
a: 5 h- >
ID O < Ul UJ
S * ui CO 2
DDDDD
»-
•z.
Ul
UJ Ul
ul Ul
a. a.
g o: "•
* Ul Ul CK
X uj IU t- UJ
U o: 2 P >•
ID O < Ul Ul
S * ui eo 2
DDDDD
H
2
UJ
ui to
ul UJ
a: cc
g or °-
* uj ui o:
X Ul Ul 1- Ul
U a: s P >
3 O 5 ui ui
S S to CD 2
D D.D DD
l-
z
UJ
UJ ft
to LU
of a:
i K t
^ tu ui cc
5 to ul >- Ul
£ s >- >
3 O 5 ui ui
S * ul CO 2
DDDDD
i-
2
UJ
UJ ft
to UJ
a: cc
g a °-
* UJ Ul CC
5 S S t= >
i i I s s
D D D DD
St_
C m °<^.
•B.S - » g
g
s^ * a 3
•- ° -fi 01 c
0 S. 3 £'2
>•• e E *
^&t||
-•g.sl£
o
2:
o
Q-
n 12
>- *»
I—I ^
^ o U >
-g D
r — i
D
UJ Z
- ^ n
p S; "-1
^» ._ J< «
•rf °. a 5
a-
z "5
o 8
£ •&
= 3 s «
o" " -S .3
K £ »
UJ C
ta tu
s = * i
S 5 S
a 1 s
C-22
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
APPENDIX D:
ABBREVIATIONS AND CONVERSION FACTORS
ABBREVIATIONS
°C degrees Celsius (Centigrade)
CAMP Continuous Air Monitoring Program
CHESS Community Health and Environmental
Surveillance System
COH coefficient of haze
cm centimeters
EDTA ethylenediamine tetraacetic acid
EPA U.S. Environmental Protection Agency
°F degrees Fahrenheit
FEVQ.75 0.75-second forced expiratory volume
g grams
Hg mercury
LMB lower-middle class black
LMW lower-middle class white
LRI lower respiratory illness
m meters
M molar
max. maximum
MBTH 3-methyl-2-benzothiazolone hydrozone
hydrochloride
mg milligrams
min minutes
min. minimum
ml milliliters
mm millimeters
mo months
mph miles per hour
N normal
NASN National Air Sampling Network
NDEA N-1-naphthylethylenediamine dihydro-
chloride
nm nanometers
NO nitric oxide
NC>2 nitrogen dioxide
NYC-DAR New York City Department of Air
Resources
p probability
ppm parts per million
PRA pararosanijine hydrochloride
RSP respirable suspended particulates
sec seconds
SES socioeconomic status
SN suspended nitrates
S02 sulfur dioxide
SS suspended sulfates
TCM sodium tetrachloromercurate
Tmin minimum temperature
TSP total suspended particulate
UMW upper-middle class white
X2 Chi Square
yr years
fj.g micrograms
nm micrometers
CONVERSION FACTORS
Primary data, such as pollutant concentrations, in
this report are given in metric units. For convenience
some intermediate data, such as emission rates, used
in the calculations are given in nonmetric units as
originally reported. If desired, metric equivalents for
those data can be obtained by application of the
conversion factors given below.
To obtain
Multiply
cubic feet
degrees Fahrenheit*
feet
inches
miles
tons
*To obtain Celsius (C) temperature readings from Fahrenheit
(F) readings, use the following formula: C = (5/9)(F - 32).
2.83x ID'2
5/9
3.05 x 10'1
2.54 x lO'2
1.61 x 103
1.02 x 103
cubic meters
degrees Celsius
(Centigrade)
meters
meters
meters
kilograms
D-l
-------
APPENDIX E:
BIBLIOGRAPHY
HEALTH EFFECTS
General
Acute Conditions, Incidence and Associated Disa-
bility. In: Health Statistics Series B. National Center
for Health Statistics, Public Health Service, U. S.
Department of Health, Education, and Welfare. Wash-
ington, D. C. PHS Publication No. 584. 1969.
Acute Conditions (Incidence and Associated Disa-
bility, United States-June 1965-June 1966). Na-
tional Center for Health Statistics, Public Health
Service, U. S. Department of Health, Education, and
Welfare. Washington, D. C. PHS Publication No.
1000, Series 10, No. 38. June 1967. 61 p.
Ambdur, M. 0. and M. Corn. The Irritant Potency of
Zinc Ammonium Sulfate of Different Particle Sizes.
Amer. Ind. Hygiene Assoc. J. 24:326-333, July-
August 1963.
Ambdur, M. 0. and D. Underbill. The Effect of
Various Aerosols on the Response of Guinea Pigs to
Sulfur Dioxide. Arch. Environ. Health. 7(5:460-468,
1968.
Anderson, D. O. and B. G. Ferris, Jr. Community
Studies of the Health Effects of Air Pollution-a
Critique. J. Air Pollut. Contr. Assoc. 75:587-593,
December 1965.
Ashford, J. R., S. Brown, D, P. Duffield, C. S. Smith,
and J. W. J. Fay. The Relationship between Smoking
Habits and Physique, Respiratory Symptoms, Ventila-
tory Function, and Radiological Pneumoconiosis
amongst Coal Workers at Three Scottish Collieries.
Brit. J. Prevent. Soc. Med. 75:106-117, 1961.
Metallurgical Industries (Vol. 2). Sydney, Australia,
Division of Occupational Health, New South Wales
Department of Public Health, 1962. p. 1-144.
Buechley, R., J. E. Dunn, Jr., G. Linden, and L.
Breslow. Excess Lung Cancer Mortality Rates among
Mexican Women in California. Cancer. 70:63-66,
1957.
Buechley, R. W., J. VanBruggen, and L. E. Truppi.
Heat Island = Death Island? Environ. Res. 5(l):85-92,
March 1972.
Carnow, B. The Relationship of SO2 Levels to
Morbidity and Mortality in "High Risk" Population.
University of Illinois. (Presented at American Medical
Association Air Pollution Medical Research Confer-
ence. New Orleans. October 5, 1970.)
Chronic Conditions and Limitations of Activity and
Mobility: United States, July 1965-June 1967. In:
Vital and Health Statistics from the National Health
Survey. Health Services and Mental Health Adminis-
tration, Public Health Service, U. S. Department of
Health, Education, and Welfare. Rockville,Md. Series
10, No. 61. January 1971.
Clark, D. and B. MacMahon. Preventive Medicine
(Chapter 25). Boston, Little Brown and Company,
1967.
Dean, G. Lung Cancer in South Africans and British
Immigrants. Proc. Roy. Soc. Med. 57:984-987, 1964.
Dingle, J. H., G. F. Badger, and W. S. Jordan. Patterns
of Illness. In: Illness in the Home. Cleveland, The
Press of Western Reserve University, 1964. p. 33-37.
Bell, A. The Effects on the Health of Residents of
East Port Kembla, Part II. In: Air Pollution by
Eascott, D. F. The Epidemiology of Lung Cancer in
New Zealand. Lancet. 7:37-39, 1956.
E-l
-------
Family Use of Health Services (United States—July
1963-June 1964). National Center for Health Sta-
tistics, Public Health Service, U. S. Department of
Health, Education, and Welfare. Washington, D. C.
PHS Publication No. 1000, Series 10, No. 55. July
1969.
Firket, J. The Cause of the Symptoms Found in the
Meuse Valley during the Fog of December 1930. Bull.
Roy. Acad. Med. Belgium. 77:683-739, 1931.
Firket, J. Fog along the Meuse Valley. Trans. Faraday
Soc. 52:1192-1197, 1936.
Glasser, M., L. Greenburg, and F. Field. Mortality and
Morbidity during a Period of High Levels of Air
Pollution, New York, November 23-25, 1966. Arch.
Environ. Health. 75:684-694, 1967.
Gore, A. T. and C. W. Shaddick. Atmospheric
Pollution and Mortality in the County of London.
Brit. J. Prev. Soc. Med. 72:104-113, 1958.
Manzhenko, E. G. The Effect of Atmospheric Pollu-
tion on the Health of Children. Hygiene and Sani-
tation (Moscow). 57:126-128, 1966.
Petrilli, R. L., G. Agnese, and S. Kanitz. Epidemi-
ology Studies of Air Pollution Effects in Genoa, Italy.
Arch. Environ. Health. 72:733-740, 1966.
Rahe, R. H., J. D. McKean, and R. J. Arthur. A
Longitudinal Study of Life-change and Illness Pat-
terns. J. Psychosom. Res. 70:355-366, 1967.
Schrenk, H. H., H. Hermann, G. 0. Clayton, W. M.
Gafafer, and H. Wexler. Air Pollution in Donora,
Pennsylvania; Epidemiology of the Unusual Smog
Episode of October 1948. Federal Security Agency,
Division of Industrial Hygiene, Public Health Service,
U. S. Department of Health, Education, and Welfare.
Washington, D. C. Public Health Bulletin 306. 1949.
Scott, J. A. The London Fog of December 1962.
Med. Officer. 709:250-252, 1963.
Greenburg, L., F. Field, J. I. Reed, and C. L. Erhardt.
Air Pollution and Morbidity in New York City. J.
Amer. Med. Assoc. 752:161-164, 1962.
Greenburg, L., M. B. Jacobs, B. M. Drolette, F. Field,
and M. M. Braverman. Report of an Air Pollution
Incident in New York City, November 1953. Public
Health Reports. 77:7-16, 1962.
Seely, J. E., E. Zuskin, and A. Bouhuys. Cigarette
Smoking: Objective Evidence for Lung Damage in
Teenagers. Science. 7 72(3984):741-743, May 14,
1971.
Sterling, T. D., J. J. Phair, S. V. Pollack, D. A.
Schumsky, and I. de Grott. Urban Morbidity and Air
Pollution. Arch. Environ. Health. 75:158-170, 1966.
Haenszel, W. Cancer Mortality among the Foreign-
Born in the United States. J. Nat. Cancer Inst.
26:37-132, 1961.
Logan, W. P. D. Mortality in the London Fog
Incident, 1952. Lancet. 264:336-338, 1953.
Toyama, T. Air Pollution and Its Health Effects in
Japan. Arch. Environ. Health. 5:153-173, 1964.
Wilkens, E. T. Air Pollution Aspects of the London
Fog of December 1952. Roy. Meteorol. Soc. J.
50:267-271, 1954.
McCarroll, J. and W. Bradley. Excess Mortality as an
Indicator of Health Effects of Air Pollution. Amer. J.
Public Health. 56:1933-1942, 1966.
McCarroll, J. R., E. J. Cassell, E. W. Walter, J. D.
Mountain, J. R. Diamond, and I. R. Mountain. Health
and the Urban Environment; V. Air Pollution and
Illness in a Normal Urban Population. Arch. Environ.
Health. 74:178-184, 1967.
Mancuso, T. F. and E. J. Coulter. Cancer Mortality
among Native White, Foreign-Born White, and Non-
White Male Residents of Ohio: Cancer of the Lung,
Larynx, Bladder, and Central Nervous System. J. Nat.
Cancer Inst. 20:79-105, 1958.
Zeidberg, L. D., R. A. Prindle, and E. Landau. The
Nashville Air Pollution Study; III. Morbidity in
Relation to Air Pollution. Amer. J. Public Health.
54:85-97, 1964.
Acute or Chronic Respiratory Disease
Addington, W. W., R. L. Carpenter, J. F. McCoy, K.
A. Duncan, and K. Mogg. The Association of Ciga-
rette Smoking with Respiratory Symptoms and Pul-
monary Function in a Group of High School Stu-
dents. J. Oklahoma State Med. Assoc. 65:525-529,
November 1970.
E-2
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Anderson, D. O. and B. G. Ferris. Air Pollution Levels
and Chronic Respiratory Disease. Arch. Environ.
Health. 70:307-311, 1965.
Anderson, D. 0., B. G. Ferris, and R. Zickmantel.
Levels of Air Pollution and Respiratory Disease in
Berlin, New Hampshire. Amer. Rev. Respiratory Dis.
90:877-887, 1964.
Burgess, S. G. and C. W. Shaddick. Bronchitis and Air
Pollution. Roy. Soc. Health J. 79:10-24, 1959.
Carnow, B. W., M. H. Lepper, R. B. Shekelle, and J.
Stamler. The Chicago Air Pollution Study: Acute
Illness and SC»2 Levels in Patients with Chronic
Bronchopulmonory Disease. Arch. Environ. Health.
75:768-776, 1969.
Deane, M., J. R. Goldsmith, and D. Tuma. Respira-
tory Conditions in Outside Workers. Arch. Environ.
Health. 70:323-331, 1965.
Densen, P. M., E. W. Jones, H. E. Hyman, and J.
Breuer. A Survey of Respiratory Disease among New
York City Postal and Transit Workers; 1. Prevalence
of Symptoms. Environ. Res. 7:265-286, 1967.
Dohan, F. C. Air Pollutants and Incidence of Respira-
tory Disease. Arch. Environ. Health. 5:887-895,
1961.
Douglas, J. W. B. and R. E. Waller. Air Pollution and
Respiratory Infection in Children. Brit. J. Prevent.
Soc. Med. 20:1-8, 1966.
Ehrlich, R. and M. C. Henry. Chronic Toxicity to
Nitrogen Dioxide: Effect on Resistance to Bacterial
Pneumonia. Arch. Environ. Health. 77:860, 1968.
Ferris, B. G. and D. O. Anderson. The Prevalence of
Chronic Respiratory Disease in a New Hampshire
Town. Amer. Rev. Respiratory Dis. 5(5:165-177,
August 1962.
Frank, N. R. et al. Effects of Acute Controlled
Exposure to S02 on Respiratory Mechanics in
Healthy Male Adults. J. Appl. Physiol. 77:252, 1972.
Frank, N. R., M. 0. Amdur, and J. L. Whittenburg. A
Comparison of the Acute Effects of S02 Admin-
istered Alone or in Combination with NaCl Particles
on the Respiratory Mechanics of Healthy Adults. Int.
J. Air Water Pollut. 8:125, 1964.
Higgins, I.T.T. Tobacco Smoking, Respiratory
Symptoms, and Ventilatory Capacity; Studies in
Random Samples of the Population. Brit. Med. J.
5775:325-329, 1959.
Higgins, I. T. T., M. W. Higgins, J. C. Bilson, H.
Campbell, W. E. Walters, and B. G. Ferris. Smoking
and Chronic Respiratory Disease: Findings in Surveys
Carried out in 1957 and 1966 in Staveley, in
Derbyshire, England. Chest. 59 (5, Suppl.):345-355,
1971.
Holland, W. W., T. Halil, A. E. Bennett, and A.
Elliott. Factors Influencing the Onset of Chronic
Respiratory Disease. Brit. Med. J. 2:205-208, April
1969.
Holland, W. W. and D. D. Reid. The Urban Factor in
Chronic Bronchitis. Lancet. 7:445-448, February 27,
1965.
Holland, W. W., D. D. Reid, R. Seltser, and R. W.
Stone. Respiratory Disease in England and the United
States; Studies of Comparative Prevalence. Arch.
Environ. Health. 70:338-345, 1965.
Holland, W. W. and R. W. Stone. Respiratory Dis-
orders in United States East Coast Telephone Men.
Amer. J. Epidemiol. 52:92-101, 1965.
Huhti, E. Prevalence of Respiratory Symptoms,
Chronic Bronchitis, and Pulmonary Emphysema in a
Finnish Rural Population; Field Survey of Age Group
40-64 in the Harjavalta Area. Acta Tuberc. Scand.
Suppl. 67:1-111, 1965.
Hyatt, R. E., A. D. Kistin, and T. K. Mahan.
Respiratory Disease in Southern West Virginia Coal
Miners. Amer. Rev. Respiratory Dis. 59:387-401,
March 1964.
Lawther, P. J. Climate, Air Pollution, and Chronic
Bronchitis. Proc. Roy. Soc. Med. 57:262-264, 1958.
Lawther, P. J., R. E. Waller, and M. Henderson. Air
Pollution and Exacerbations of Bronchitis. Thorax.
25:525-539, 1970.
Lunn, J. E., J. Knowelden, and A. J. Handyside.
Patterns of Respiratory Illness in Sheffield Infant
School Children. Brit. J. Prevent. Soc. Med. 27:7-16,
1967.
Appendix E
E-3
-------
Miller, F. J. W., S. D. M. Court, W. S. Walton, and E.
G. Know. The General Pattern of Respiratory Disease
(Chapter XXX) and Colds, Bronchitis, and Pneu-
monia (Chapter XXXI). In: Growing Up in Newcastle
Upon Tyne. London, Oxford University Press, 1960.
p. 194-217.
Payne, M. and M. Kjelsberg. Respiratory Symptoms,
Lung Function, and Smoking Habits in an Adult
Population. Amer. J. Public Health. 54:261-277,
February 1964.
Asthma
Broder, J., P. P. Barlow, and R. J. M. Horton. The
Epidemiology of Asthma and Hay Fever in a Total
Community, Tecumseh, Michigan. J. Allergy.
55:513-523, 1969.
Carroll, R. E. Epidemiology of New Orleans Epidemic
Asthma. Amer. J. Public Health. 55:1677-1683,
1968.
Pearlman, M. E., J. F. Finklea, J. P. Creason, C. M.
Shy, M. M. Young, and R. J. M. Horton. Nitrogen
Dioxide and Lower Respiratory Illness. Pediatrics.
47:391-395, 1971.
Reid, D. D. The Beginnings of Bronchitis. Proc. Roy.
Soc.Med. 62:311-316, 1969.
Reid, D. D. Environmental Factors in Respiratory
Disease. Lancet. 7:1237-1242, June 14, 1958.
Reid, D. D. Environmental Factors in Respiratory
Disease. Lancet. 7:1289-1294, June 21, 1958.
Shy, C. M., J. P. Creason, M. E. Pearlman, K. E.
McClain, F. B. Benson, and M. M. Young. The
Chattanooga School Children Study: Effects of
Community Exposure to Nitrogen Dioxide; II. Inci-
dence of Acute Respiratory Illness. J. Air Pollut.
Contr. Assoc. 20:582-588, September 1970.
Tsunetoshi, Y., T. Shimizu, H. Takahashi, A. Ichin-
osawa, M. Ueda, N. Nakayama, and Y. Yamagata.
Epidemiological Study of Chronic Bronchitis with
Special Reference to Effect of Air Pollution. Int.
Arch. Arbeitsmed. 29:1-27, 1971.
Tucher, D. Incidence of Pneumonia in Two Commu-
nities in New York State. The Milbank Memorial
Fund Quarterly. XXX:224-238, 1952.
Tucher, D., J. E. Coulter, and J. Downes. Incidence
of Acute Respiratory Illness among Males and Fe-
males; Study No. 5. The Milbank Memorial Fund
Quarterly. XXX(1):42-6Q, January 1952.
Valand, S. B., J. D. Action, and Q. Myrvils. Nitrogen
Dioxide Exhibition of Viral Induced Resistance in
Aveolar Monocytes. Arch. Environ. Health. 20:303,
1970.
Chrobok, H., 0. Wojcik, and J. Zajiczek. Influence of
Meteorological and Weather Conditions on Bronchial
Asthma, Cardiac Infarction and Mortality. Prezeglad
Lekarski (Warsaw). 22:419-421, 1966.
Cohen, A. A., S. Bromberg, R. M. Buechley, L. I.
Hiederscheit, and C. M. Shy. Asthma and Air Pollu-
tion from a Coal Fueled Power Plant. Amer. J. Public
Health. 62:1181-1188, 1972.
Cowan, D. W. and H. J. Paulus. Relationship of Air
Pollution to Allergic Diseases. Division of Air Pollu-
tion, Public Health Service, U. S. Department of
Health, Education, and Welfare. Washington, D. C.
Final report under Research Grant AP 00090. 1969.
136 p.
Cowan, D. W., H. J. Thompson, H. J. Paulus, and P.
W. Mielke. Bronchial Asthma Associated with Air
Pollutants from the Grain Industry. J. Air. Pollut.
Contr. Assoc. 75:546-552, 1963.
Figley, K. D. and R. H. Elrod. Endemic Asthma Due
to Castor Bean Dust. J. Amer. Med. Assoc. 90:79-82,
1928.
Gersh, L. S., E. Shubin, C. Dick, and F. A. Schulaner.
A Study on the Epidemiology of Asthma in Children
in Philadelphia. J. Allergy. 59:347-357, 1967.
Lewis, R., M. M. Gilkeson, and R. O. McCalden. Air
Pollution and New Orleans Asthma. Public Health
Reports. 77:947-954, 1962.
Lewis, R., M. M. Gilkeson, and R. W. Robison. Air
Pollution and New Orleans Asthma (2 volumes). New
Orleans, Tulane University, 1962.
Mendes, E. and A. B. V. Centra. Collective Asthma
Simulating an Epidemic Provoked by Castor Bean
Dust. J. Allergy. 25:253-259, 1954.
E-4
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Schoettlin, C. E. and E. Landau. Air Pollution and
Asthmatic Attacks in the Los Angeles Area. Public
Health Reports. 7(5:545-548, 1961.
Roberts, F. L. and J. A. Crabtree. The Vital Capacity
of the Negro Child. J. Amer. Med. Assoc.
55:1950-1954, 1927.
Tromp, S. W. Influence of Weather and Climate on
Asthma and Bronchitis. Review of Allergy. Vol. 22,
November 1968.
Weill, H., M. M. Ziskind, R. Dickerson, and V. J..
Derbes. Epidemic Asthma in New Orleans. J. Amer.
Med. Assoc. 790:811-814, 1964.
Weill, H., M. M. Ziskind, V. Derbes, R. Lewis, R. J.
M. Morton, and R. 0. McCalden. Further Observa-
tions on New Orleans Asthma. Arch. Environ. Health.
5:184-187, 1964.
Yoshida, K., H. Oshima, and M. Swai. Air Pollution
and Asthma in Yokkaichi. Arch. Environ. Health.
13:763-768, 1964.
Yoshida, K., H. Oshima, and M. Swai. Air Pollution in
the Yokkaichi Area with Special Regard to the
Problem of "Yokkaichi Asthma." Ind. Health.
2:87-94, 1964.
Shy, C. M., J. P. Creason, M. E. Pearlman, K. E.
McClain, F. B. Benson, and M. M. Young. The
Chattanoga School Children Study: Effects of Com-
munity Exposure to Nitrogen Dioxide; 1. Methods,
Description of Pollutant Exposure and Results of
Ventilatory Function Testing. J. Air Pollut. Contr.
Assoc. 20:539-545, August 1970.
Smillie, W. G. and D. L. Augustine. Vital Capacity of
Negro Race. J. Amer. Med. Assoc. 57:2055-2058,
1926.
Watanabe, H., F. Kaneko, H. Murayama, S. Yamaoka,
and T. Kawaraya. Effects of Air Pollution on Health;
Report No. 1: Peak Flow Rate and Vital Capacity of
Primary School Children. Reports of the Osaka City
Institute of Hygiene. 26:32-37, 1964.
Wilson, M. G. and D. J. Edwards. Diagnostic Value of
Determining Vital Capacity of Lungs of Children. J.
Amer. Med. Assoc. 75:1107-1110, 1922.
Zeidberg, L. D., R. A. Prindle, and E. Landau. The
Nashville Air Pollution Study, I. Sulfur Dioxide and
Bronchial Asthma: A Preliminary Report. Amer. Rev.
Respiratory Dis. 54:489-503, 1961.
Ventilatory Function
Damon, A. Negro-white Differences in Pulmonary
Function (Vital Capacity, Timed Vital Capacity and
Expiratory Flow Rate). Human Biol. 55:380-393,
1966.
Densen, P. M., E. W. Jones, H. E. Hyman, J. Brewer,
and E. Reed. A Survey of Respiratory Disease among
New York City Postal and Transit Workers; 2.
Ventilatory Function Test Results. Environ. Res.
2:277-296, 1969.
Higgins, I. T. T., P. D. Oldham, A. L. Cochrane, and
J. C. Gilson. Ventilatory Function in Miners: A Five
Year Follow-up Study. Brit. J. Ind. Med. 79:65-76,
1962.
McMillan, R. S., D. H. Wiseman, B. Hanes, and P. F.
Wehrle. Effects of Oxidant Air Pollution on Peak
Expiratory Flow Rates in Los Angeles School Chil-
dren. Arch. Environ. Health. 75:941-949, 1969.
MEASUREMENT AND ANALYTICAL
METHODS
Anderson, D. 0. and C. Kinnis. An Epidemiologic
Assessment of a Pediatric Peak Flowmeter. Amer.
Rev. Resp. Dis. 95:73-80, 1967.
Burton, R. M., J. N. Howard, R. L. Penley, P. B.
Ramsey, and T. A. Clark. Field Evaluation of the
High-volume Particle Fractionating Cascade Impactor.
J. Air Pollut. Contr. Assoc. 25:277-281, April J973.
Burton, R. M., W. M. Kozel, and F. B. Benson.
Development of Fine Particulate Sampling Methods
in Support of CHESS Health Studies. U. S. Environ-
mental Protection Agency. Research Triangle Park,
N. C. (In preparation.)
Collection and Analysis of Dustfall (Settleable Par-
ticulate). In: Annual Book of ASTM Standards. The
American Society for Testing and Materials. Phila-
delphia, Pa. ASTM Test Method D1739-70. 1970.
Dixon, W. J. and F. J. Massey. Introduction to
Statistical Analysis (2nd ed.). New York, McGraw-
Hill Book Company, Inc., 1957.
Appendix E
E-5
-------
Draper, N. R. and H. Smith. Multiple Regression
Mathematical Model Building (Chapter 8) and Multi-
ple Regression Applied to Analyses of Variance
Problems (Chapter 9). In: Applied Regression Anal-
yses. New York, John Wiley and Sons, 1966.
p. 234-262.
McKee, H. C., R. E. Childers, and 0. Saenz. Collabora-
tive Study of Reference Method for Determination of
Suspended Particulates in the Atmosphere (High
Volume Method). Southwest Research Institute. San
Antonio, Texas. Contract CAP 70-40, SwRI Project
21-2811. June 1971.
Eaton, W. C., R. L. Penley, and R. M. Burton.
Calibration of the Andersen 2000 High-volume Par-
ticle Sizing Collection Head. U. S. Environmental
Protection Agency. Research Triangle Park, N. C. (In
preparation.)
Ettinger, H. J., J. E. Partridge, and G. W. Royer.
Calibration of Two-stage Air Samplers. Amer. Ind.
Hygiene Assoc. J. 31:537, 1970.
Frant, M. S. and J. W. Ross, Jr. Electrode for Sensing
Fluoride Ion Activity in Solution. Science.
745:1553-1554,1966.
Graybill, F. An Introduction to Linear Statistical
Models, Vol. 1. New York, McGraw-Hill Book
Company, Inc., 1961.
Grizzle, J. E., C. F. Starmer, and G. G. Koch.
Analysis of Categorical Data by Linear Models.
Biometrics. 25(3):489-504, September 1969.
Guide for Respirable Mass Sampling. Amer. Ind.
Hygiene Assoc. J. 57:133, 1970.
Hauser, T. R. and C. M. Shy. Position Paper: NOX
Measurement. Environ. Sci. Technol. 6:890-894,
October 1972.
Hemeon, W. C. L., G. F. Haines, and H. M. Ide.
Determination of Haze and Smoke Concentration by
Filter Paper Samplers. Air Repair. J:22-28, 1953.
Jacobs, M. B. and S. Hochheiser. Continuous Sam-
pling and Ultramicro Determination of Nitrogen
Dioxide in Air. Anal. Chem. 50:426-428, 1958.
Knuth, R. H. Recalibration of Size-selective Samplers.
Amer. Ind. Hygiene Assoc. J. 30:379, 1969.
Lazrus, A. L., K. C. Hill, and J. P. Lodge. A New
Colorimetric Microdetermination of Sulfate Ion. In:
Automation in Analytical Chemistry (Vol. 1). New
York, Mediadlnc., 1965. p. 291-293.
Martin, D. 0. An Urban Diffusion Model for Estimat-
ing Long Term Average Values of Air Quality. J. Air
Pollut. Contr. Assoc. 27(1):16-19, 1971.
Morgan, G. B., E. C. Tabor, C. Golden, and H.
Clement. U. S. Environmental Protection Agency,
Research Triangle Park, N. C. Automated Laboratory
Procedures for Analysis of Air Pollutants. (Presented
at the 59th Annual Air Pollution Control Association
Meeting. San Francisco. June 20-24, 1966.)
Parr, S. W. and W. D. Staley. Determination of Sulfur
by Means of the Turbidimeter. Ind. Eng. Chem.
(Annual Ed.). 5:66-67, 1931.
Particulate Matter in the Atmosphere-Optical Den-
sity of Filtered Deposit. In: Annual Book of ASTM
Standards. The American Society for Testing and
Materials. Philadelphia, Pa. ASTM Test Method
D1704-61. 1970. p. 498-505.
Quandt, R. E. The Estimation of the Parameters of a
Linear Regression System Obeying Two Separate
Regimes. J. Amer. Statistical Assoc. 55:873-880,
1958.
Riggan, W. B., D. I. Hammer, J. F. Finklea, V.
Hasselblad, C. R. Sharp, R. M. Burton, and C. M. Shy.
CHESS, a Community Health and Environmental
Surveillance System. In: Proceedings of the Sixth
Berkeley Symposium on Mathematical Statistics and
Probability (Vol. 6). Berkeley, University of Cali-
fornia Press, 1972.
Saltzman, B. E. Colorimetric Microdetermination of
Nitrogen Dioxide in the Atmosphere. Anal. Chem.
26:1949-1955, 1954.
Scaringelli, F. P., B. E. Saltzman, and S. A. Frey.
Spectrophotometric Determination of Atmospheric
Sulfur Dioxide. Anal. Chem. 59:1709-1719, Decem-
ber 1967.
Selected Methods for the Measurement of Air Pollut-
ants. Division of Air Pollution, Public Health Service,
U. S. Department of Health, Education, and Welfare.
Cincinnati, Ohio. PHS Publication No. 999-AP-ll.
1965.
E-6
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
Silverman, L. and F. G. Viles. A High-volume Air
Sampling and Filter Weighing Method for Certain
Aerosols. J. Ind. Hygiene Toxicol. 30:124, 1948.
Standards for Epidemologic Surveys in Chronic
Respiratory Disease. National Tuberculosis and
Respiratory Disease Association. New York, N. Y.
1969. 36 p.
U. S. Environmental Protection Agency. National
Primary and Secondary Air Quality Standards;
Reference Method for Determination of Nitrogen
Dioxide. Federal Register. 38(110): 15174-15191,
June 8, 1973.
U. S. Environmental Protection Agency. National
Primary and Secondary Air Quality Standards;
Reference Method for Determination of Nitrogen
Dioxide in the Atmosphere (24-hour Sampling
Method). Federal Register. 3S(84):8200-8201, April
30, 1971.
U. S. Environmental Protection Agency. National
Primary and Secondary Air Quality Standards; Refer-
ence Method for the Determination of Sulfur Dioxide
in the Atmosphere (Pararosaniline Method). Federal
Register. 36(84):8187-8190, April 30, 1971.
U. S. Environmental Protection Agency. National
Primary and Secondary Air Quality Standards; Refer-
ence Method for the Determination of Suspended
Particulates in the Atmosphere (High Volume
Method). Federal Register. J6(84):8191-8194, April
30, 1971.
Van der Lende, R., R. Ter Brugge, J. P. M. DeKroon,
H. J. Sluiter, G. J. Tammeling, K. DeVries, and N. G.
M. Orie. The Organization of an Epidemiologic
Investigation. T. Soc. Geneesk. 44:148-157, 1966.
Vecchio, T. Predictive Value of a Single Diagnostic
Test in Unselected Populations. New Eng. J. Med.
274:1171-1173, May 26, 1966.
Welch, A. F. and J. P. Terry. Developments in the
Measurement of Atmospheric Sulfur Dioxide. J.
Amer. Ind. Hygiene Assoc. 27:316, 1960.
POLLUTANT CONCENTRATIONS
Air Pollution Measurements of the National Air
Sampling Network, Analysis of Suspended Particu-
lates, 1963. Public Health Service, U. S. Department
of Health, Education, and Welfare. Cincinnati, Ohio.
1965.
Air Quality Criteria for Particulate Matter. National
Air Pollution Control Administration, Public Health
Service, U. S. Department of Health, Education, and
Welfare. Durham, N. C. NAPCA Publication No.
AP-49. January 1969.
Air Quality Criteria for Sulfur Oxides. National Air
Pollution Control Administration, Public Health
Service, U. S. Department of Health, Education, and
Welfare. Durham, N. C. NAPCA Publication No.
AP-50. January 1969.
Air Quality Data (1962), National Air Sampling
Network. Public Health Service, U. S. Department of
Health, Education, and Welfare. Cincinnati, Ohio.
1965.
Air Quality Data from the National Air Sampling
Networks and Contributing State and Local Net-
works, 1964-1965. Public Health Service, U. S.
Department of Health, Education, and Welfare.
Cincinnati, Ohio. 1966.
Air Quality Data from the National Air Sampling
Networks and Contributing State and Local Net-
works, 1966. National Air Pollution Control Admin-
istration, Public Health Service, U. S. Department of
Health, Education, and Welfare. Durham, N. C.
NAPCA Publication No. APTD-68-9. 1968.
Air Quality Data for 1967 from the National Air
Surveillance Networks and Contributing State and
Local Networks, Revised 1971. Office of Air Pro-
grams, U. S. Environmental Protection Agency.
Research Triangle Park, N. C. GAP Publication No.
APTD-0741. August 1971.
Compilation of Air Pollutant Emission Factors.
Office of Air Programs, U. S. Environmental Protec-
tion Agency. Research Triangle Park, N. C. OAP
Publication No. AP-42. February 1972. 166 p.
West, P. W. and G. C. Gaeke. Fixation of Sulfur
Dioxide as Sulfitomercurate III and Subsequent
Colorimetric Determination. Anal. Chem.
25:1816-1819, 1956.
Helena Valley, Montana, Area Environmental Pollu-
tion Study. Office of Air Programs, U. S. Environ-
mental Protection Agency. Research Triangle Park, N.
C. OAP Publication No. AP-91. January 1972. 193 p.
Appendix E
E-7
-------
Idaho Air Quality — Methods of Measurement and
Analysis of Recent Data. Idaho Department of
Health. Boise, Idaho. August 19, 1970. 23 p.
A Study of Air Pollution in Townsend-Three Forks,
the Gallatin Valley, and West Yellowstone, November
1967-November 1968. Montana State Department of
Health. Helena, Montana. 43 p.
Spirtas, R. and H. J. Levin. Characteristics of Particu-
late Patterns - 1957-1966. National Air Pollution
Control Administration, Public Health Service, U. S.
Department of Health, Education, and Welfare.
Raleigh, N. C. NAPCA Publication No. AP-61. March
1970. 101 p.
A Study of Air Pollition in the Deer Lodge Valley-
August 1965-June 1966. Montana State Department
of Health. Helena, Montana. 32 p.
A Study of Air Pollution in the Helena-East Helena
Area - October 1965-October 1968. Montana State
Department of Health. Helena, Montana. 35 p.
Wagman, J., R. E. Lee, Jr., and C. J. Axt. Influence of
Some Atmospheric Variables on the Concentration
and Particle Size Distribution of Sulfate in Urban Air.
Atmospheric Environ. 7:479-489, 1967.
Yearbook of the American Bureau of Metal Statistics,
29th Annual Issue for 1949. New York, American
Bureau of Metal Statistics, 1950.
Yearbook of the American Bureau of Metal Statistics,
35th Annual Issue for 1955. New York, American
Bureau of Metal Statistics, 1956.
Yearbook of the American Bureau of Metal Statistics,
44th Annual Issue for 1964. New York, American
Bureau of Metal Statistics, 1965.
A Study of Air Pollution in Montana, July 1961-July
1962. Montana State Board of Health. Helena,
Montana. 107 p.
Yearbook of the American Bureau of Metal Statistics,
51st Annual Issue for 1971. New York, American
Bureau of Metal Statistics, 1972.
E-8
HEALTH CONSEQUENCES OF SULFUR OXIDES
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-650/1-74-004
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
HEALTH CONSEQUENCES OF SULFUR OXIDES:
A Report from CHESS, 1970-1971
5. REPORT DATE
May 1974
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
National Environmental Research Center
Office of Research and Development
U.S . Environmental Protection Agency
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
1A1005
11. CONTRACT/GRANT NO
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Final, 1970-1971
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Community Health and Environmental Surveillance Systems (CHESS) studies provide
dose-response information relating short- and long-term air pollution exposures to
adverse health effects. This report presents results of studies in CHESS communities
in New York and the Salt Lake Basin during 1970-1971. Studies in Idaho-Montana,
Chicago, and Cincinnati, employing health indicators similar to those used in CHESS,
are also included. Attention is focused on effects of sulfur oxides, but the relative
contribution of various pollutants, especially sulfur dioxide, particulates, and
suspended sulfates, is considered. Health indicators of long-term pollution effects
included acute and chronic respiratory illness and ventilatory function. Indicators
of short-term effects were cardiopulmonary symptoms and asthma. Threshold
estimates for the pollutants considered support existing National Primary Air
Quality Standards for long-term exposures. For short-term exposures, the studies
indicated adverse effects even on days below the Standards for 24-hour levels of
sulfur dioxide and particulates. These effects, however, appear to be associated
with suspended sulfates.
KEY WORDS AND DOCUMENT ANALYSIS
Acute respiratory disease
Air pollution (health effects)
Air quality standards
Asthma
Cardiopulmonary illness
CHESS (Community Health
and Environmental Sur-
veillance System)
Chronic respiratory disease
Epidemiology
Health effects (air pollution)
Particulates
Respiratory disease
Sulfur dioxide
Sulfur oxides
Suspended sulfates
Total suspended
particulates
Ventilatory function
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
21. NO. OF PAGES
454
EPA Form 2220-1 (9-73)
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
4. TITLE AND SUBTITLE
Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
type or otherwise subordinate it to mam title. When a report is prepared in more than one volume, repeat the primary title, add volume
number and include subtitle for the specific title.
5. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e g., date of issue, date of
approval, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc./. List author's affiliation if it differs from the performing organi-
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Leave blank.
15. SUPPLEMENTARY NOTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with, Translation of, Presented at conference of,
To be published in, Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 words or less) factual summary of the most significant information contained in the report. If the report contains a
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI FIELD GROUP - Field and group assignments are to be taken from the 1965 COSATI Subject Category List. Since the ma-
jority of documents are multidisciplmary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary Field/Group assignments that will follow
the primary posting(s).
18. DISTRIBUTION STATEMENT
Denote releasabihty to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
the public, with address and price.
19. &20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1 (9-73) (Reverse)
------- |