COMMISSION OF THE
EUROPEAN
COMMUNITIES
UNITED STATES
ENVIRONMENTAL
PROTECTION AGENCY
WORLD HEALTH
ORGANIZATION
INTERNATIONAL SYMPOSIUM
PROCEEDINGS
Recent Advances
in the Assessment of the Health Effects
of Environmental Pollution
Volume I
^
._ .
Paris. 24 to 28 June 1974
-------
LEGAL NOTICE
The Commission of the European Communities
and its departments decline all
responsibility with regard to the use of
the information contained herein.
-------
TAGUNGSBERICHTE
INTERNATIONALES SYMPOSIUM
NEUESTE ERKENNTNISSE IN DER BEURTEILUNG
DER GESUNDHEITLICHEN FOLGEN DER UMWELTVERSCHMUTZUNG
PROCEEDINGS
INTERNATIONAL SYMPOSIUM
RECENT ADVANCES IN THE ASSESSMENT
OF THE HEALTH EFFECTS OF ENVIRONMENTAL POLLUTION
ACTES
SYMPOSIUM INTERNATIONAL
PROGRES REGENTS DANS L*EVALUATION DES
EFFETS DE LA POLLUTION DE L'ENVIRONNEMENT SUR LA SANTE
ATT I
SIMPOSIO INTERNAZIONALE
RECENTI PROGRESSI NELLA VALUTAZIONE
DEGLI EFFETTI DELL'INQUINAMENTO DELL' AMBIENTE SULLA SALUTE
VORSLAG
INTERNATIONAAL SYMPOSIUM
RECENTE VORDERINGENBIJ DE VASTSTELLING VAN
DE GEVOLGEN VAN MILIEUVERONTREINIGING VOOR DE GEZONDHEID
Paris (France), 'June 24-28, 1974
organized jointly by
CEC - Commission of the European Communities
Directorate General for Social Affairs - Health Protection Directorate
EPA - United States Environmental Protection Agency
and
WHO - World Health Organization
Published by the Commission of the European Communities
Directorate General Scientific and Technical Information and Information Management
Luxembourg, 1975
-------
[LI
ORGAMISATIONSAUSSCHUSS / ORGANIZING COMMITTEE / COMIT£ ORGANISATEUR
COMITATO ORGANIZZATIVO / ORGANISEREND COMITE^
Vorsitzend / Chairmen / Presidents / Presidenti / Voorzitters:
B.H. DIETERICH (W.H.O.)
S.GREENFIELD (E .P A.)
P. RECHT (C.E.C.)
Stellvertretende Vorsizende / Vice Chairmen / Vice-presidents /
Vicepresidenti / Vice-Voorzitters:
D. EARTH (E.P.A.)
M. CARPENTIER (C.E.C.)
J. KUMPF (W.H.O.)
Generalsekretar / Secretary General / Secretaire general /
Segretario generate / Algemeen Secretarie:
J. SMEETS (C.E.C.)
Wissensahaftliahe Sekretariat / Scientific Secretariat /
Secretariat scientifique / Segretariato aeientifiao /
Wetengahappelijk Secretariaat:
A. BERLIN (C.E.C.)
R. ENGEL (E.P.A.)
V.B. VOUK (W.H.O.)
Assistentin / Assistant / Assistante / Assistente / Aseietent:
G. TREU-RICCO' (C.E.C.)
Wissenschaftliche Berater / Scientific advisors /
Conseillers scientifiques / Consiglieri saientifici /
Wetenschappelijke Advieeure:
Ph. BOURDEAU (C.E.C.), G.C. BUTLER (Canada)
M. COLOMBINI (Italy), G. DEAN (Republic of Ireland),
J.F. FINKLEA (U.S.A.), L. FRIBERG (Sweden),
H.E. GRIFFIN (U.S.A.), W.W. HOLLAND (U.K.),
M. KEY (U.S.A.), A. LAFONTAINE (Belgium),
D. RALL (U.S.A.), H.-W. SCHLIPKOTER (F.R. Germany),
R. SENAULT (France), J. SPAANDER (Netherlands),
L. TEPPER (U.S.A.), N.K. WEAVER (U.S.A.)
-------
TAGUNGSBERICHTE
INTERNATIONALES SYMPOSIUM
NEUESTE ERKENNTNISSE IN DER BEURTEILUNG
DER GESUNDHEITLICHEN FOLGEN DER UMWELTVERSCHMUTZUNG
Band I
ERdFFNUNGSSITZUNG
POPULATIONSUNTERSUCHUNGEN
UNTERSUCHUNG DER WIRKUNGEN AUF DEN MENSCHEN (panel)
TOXIKOLOGISCHE UNTERSUCHUNGEN
EXPOSITIONSMESSUNGEN (Panel)
Band II
UNTERSUCHUNG DER WIRKUNGEN AUF DEN MENSCHEN
TIERUNTERSUCHUNGEN
STOFFWECHSEL
WECHSELBEZIEHUNGEN
EXPOSITIONS INDIKATOREN
Band III
GEWEBSMESSUNGEN
ERFORDERNISSE IM HINBLICK AUF MESSUNGEN
UMWELTMESSUNGEN
EXPOSITIONSMESSUNGEN
MODELLE
ZUSATZLICHE BERICHTE
Band IV
DIE WISSENSCHAFTLICHEN DATENGRUNDLAGEN, DIE FUR DIE
ENTSCHEIDUNG UBER DEN GESUNDHEITSSCHUTZ BENOTIGT WERDEN
(Aussprache im rahmen des gesaratauaschusses)
EXPERIMENTELLE FORSCHUNG
ERFORDERNISSE IM HINBLICH AUH MESSUNGEN (Panel)
GEWEBSMESSUNGEN (Panel)
UNTERSUCHUNG DER GESUNDHEITLICHEN WIRKUNGEN
SCHLUSSITZUNG
TEILNEHMERLISTE
-------
VL
PROCEEDINGS
INTERNATIONAL SYMPOSIUM
RECENT ADVANCES IN THE ASSESSMENT OF THE HEALTH
EFFECTS OF ENVIRONMENTAL POLLUTION
Volume I
OPENING SESSION
POPULATION STUDIES
HUMAN EFFECTS STUDIES (Panel)
TOXICOLOGICAL STUDIES
EXPOSURE MONITORING (Panel)
Volume II
HUMAN EFFECTS STUDIES
ANIMAL STUDIES
METABOLISM
INTERACTIONS
INDICATORS OF EXPOSURE
Volume III
TISSUE MEASUREMENTS
MONITORING NEEDS
ENVIRONMENTAL MEASUREMENTS
EXPOSURE MONITORING
MODELS
SUPPLEMENTARY PAPERS
Volume IV
THE SCIENTIFIC DATA BASE REQUIRED FOR DECISIONS TO
PROTECT HUMAN HEALTH (Plenary discussion group)
EXPERIMENTAL INVESTIGATIONS
MONITORING NEEDS (Panel)
TISSUE MEASUREMENTS (Panel)
HEALTH EFFECTS STUDIES
CLOSING SESSION
LIST OF PARTICIPANTS
-------
VII
ACTES
SYMPOSIUM INTERNATIONAL
PROGRES REGENTS DANS DEVALUATION DES
EFFETS DE LA POLLUTION DE L'ENVIRONNEMENT SUR LA SANTE
Volume I
SEANCE D'OUVERTURE
ETUDES SUR LA POPULATION
ETUDES DES EFFETS SUR L*HOMME (Panelj
ETUDES TOXICOLOGIQUES
MESURE DE L'EXPOSITION (Panel)
Volume II
ETUDES DES EFFETS SUR L'HOMME
ETUDES SUR LES ANIMAUX
METABOLISMS
INTERACTIONS
INDICATEURS D'EXPOSITION
Volume III
MESURES RELATIVES AUX TISSUS BIOLOGIQUES
BESOINS EN MATIERE DE MESURE DE L'EXPOSITION
MESURES DANS L'ENVIRONNEMENT
MESURE DE L'EXPOSITION
MODELES
RAPPORTS SUPPLEMENTAIRES
Volume IV
LES DONNEES SCIENTIFIQUES REOUISES A
LA PRISE DE DECISION POUR LA PROTECTION DE LA SANTE HUMAINE
(Groupe de discussion en assemblee pi§niere)
RECHERCHES EXPERIMENTALES
BESOINS EN MATIERE DE MESURE DE L'EXPOSITION (Panel)
MESURES RELATIVES AUX TISSUS BIOLOHIQUES (Panel)
ETUDES DES EFFETS SUR LA SANTE
SEANCE DE CLOTURE
LISTE DES PARTICIPANTS
-------
VJ II
ATT1
SIMPOSIO INTERNAZIONALE
RECENTI PROGRESSI NELLA VALUTAZIONE
DEGLI EFFETTI DELL;INQUINAMENTO DELL'AMBIENTE SULLA SALUTE
Volume I
SESSIONS INAUGURALE
STUDI SULLA POPOLAZIONE
STUDI DEGLI EFFETTI SULL'UOMO (Panel)
STUDI TOSSICOLOGICI
MISURA DELL'ESPOSIZIONE (Panel;
Volume II
STUDI DEGLI EFFETTI SULL'UOMO
STUDI SUGLI ANIMALl
METABOLISMO
INTERAZIONE
INDICATORI DI ESPOSIZIONE
Volume III
MISURE NEI TESSUTI BIOLOGICI
NECESSITY RELATIVE ALLA MISURA DELL'ESPOSIZIONE
MISURE AMBIENTALI
MISURA DELL'ESPOSIZIONE
MODELLI
RAPPORTI SUPPLEMENTARI
Volume IV
I DATI SCIENTIFIC! DI BASE RICHIESTI PER
PRENDERE DEC IS ION I AL FINE DI PROTEGGERE LA SALUTE UMANA
(Gruppo di discussione plenaria)
INDAGINE SPERIMENTALE
NECESSITA' RELATIVE ALLA MISURA DELL'ESPOSIZIONE (Panel)
MISURE NEI TESSUTI BIOLOGICI (Panel)
STUDI DEGLI EFFETTI SULLA SALUTE
SESSIONS FINALE
ELENCO DEI PARTECIPANTI
-------
IX
VORSLAG
INTERNATIONAL SYMPOSIUM
RECENTE VORDERINGEN BIJ DE VASTSTELLING
VAN DE GEVOLGEN VAN MILIEUVERONTREINIGING VOOR DE GEZONDHEID
Band I
OPENINGSZITTING
BEVOLKINGSSTUDIES
ONDERZOEKINGEN NAAR EFFECTEN BIJ DE MENS (Panel)
TOXICOLOGISCHE ONDERZOEKINGEN
METING VAN EXPOS IE (Panel)
Band II
ONDERZOEKINGEN NAAR EFFECTEN BIJ DE MENS
ONDERZOEKINGEN BIJ DIEREN
METABOLISMS
INTERACTIES
INDICATOREN VAN EXPOSIE
Band III
METINGEN VAN BIOLOGISCH WEEFSEL
EISEN VOOR MET TOT STAND BRENGEN VAN TOEZICHT OP EXPOSIE
METINGEN IN HET MILIEU
METINGEN VAN EXPOSIE
MODELLEN
SUPPLEMENTAIRE DOCUMENTEN
Band IV
WETENSCHAPPELIJKE GEGEVENS NOODZAKELIJK VOOR HET NEMEN VAN
BESLISSINGEN TER BESCHERMING VAN DE GEZONDHEID VAN DE MENS
(Pienaire discussiegroep)
EXPERIMENTEEL ONDERZOEK
EISEN VOOR HET TOT STAND BRENGEN VAN TOEZICHT OP EXPOSIE (Panel)
METINGEN VAN BIOLOGISCH WEEFSEL (Panel)
ONDERZOEKINGEN NAAR GEVOLGEN VOOR DE GEZONDHEID
SLOTZITTING
DEELNEMERSL1JST
-------
XI
PREFACE
Many countries are recognizing the importance of the availability of well
established exposure-effect relationships to develop various approaches to
protect the public health. This is leading to the use of existing health and
monitoring data which in our experience, gained to date, could have been more
complete and comparable. To improve this situation the Commission of the
European Communities, the United States Environmental Protection Agency and
World Health Organization organized in Paris, from 2k to 28 June 197*4, an
International Symposium on the "Recent Advances in the Assessment of the
Health Effects of Environmental Pollution".
The aim of this scientific meeting was to evaluate the methodologies and
techniques currently available for establishing exposure-effect relationships
for environmental pollutants. The subject matter covered by the Symposium
emphasized three main areas which could offer approachs to a more common
denominator in methodology. Hopefully, wide acceptance of such a denominator
will help to ensure that future data will be more directly applicable to an
assessment of the nature, role, and significance of health effects.
Hopefully this contribution will be useful for establishing appropriate
environmental control plans to prevent or reduce adverse health effects, from
exposure to environmental pollutants.
on behalf of the Organizing Committee,
D.H. DIETERICH (WHO) S.M. GREENFIELD (EPA) P. RECHT (CEC)
-------
XIII
EDITORIAL NOTE
This presentation contains the full proceedings of the Symposium,
Volumes I and IV include the broad coverage papers presented at plenary
sessions as well as all panel discussions; and the opening and closing ses-
sions which set forth essentially the intent and conclusions of the Symposium
respectively. Volumes IT and III contain all papers presented at specialized
sessions as well as the supplementary reports.
The reports and communications are published in their original language
Dutch, English, French, German, and Italian, with an English translation of
the abstracts. The opening and closing sessions are published in the original
language and English. We wish to point out that the views expressed in the
various communications are the responsibility of the authors and may or may
not represent the views of their sponsors. Through our extensive editing, and
rearrangement of the order of the presentations we have produced a shorter
and hopefully more readable document.
We anticipate that these proceedings will be a valuable contribution to
the necessary exchange of scientific knowledge, to the enhancement of human
health, and to the protection of the environment.
BERLIN, TUGEL, SMEETS
-------
XV
INHALTSVERZEICHNIS
CONTENTS
TABLE DES MATIERES
INDICE
INHOUDSOPGAVE
-------
XVII
ERbFFNUNGSSITZUNG
OPENING SESSION
SEANCE D OUVERTURE
SESSIONE INAUGURALE
OPENINGSZITTING
RUSSELL W.PETERSON 3
Chairman, President's Council on Environmental Quality
P.J. HILLERY 7
Vice-president of the Commission of the European Communities
T. ADEOYE LAMBO 11
Deputy Director-General, World Health Organization
RENE MAHEU 17
Directeur General, UNESCO
SIMONE VEIL 25
Ministre de la Sante, France
POPULATIONSUNTERSUCHUNGEN
POPULATION STUDIES
ETUDES SUR LA POPULATION
STUDI SULLA POPOLAZIONE
BEVOLKINGSSTUDIES
The effect of environmental factors on health - methods for
population studies 39
W.W. HOLLAND
Household survey of the incidence of resporatory disease in
relation to environmental pollutants 47
R.I. MITCHELL, R. WILLIAMS, R.W. COTE, R.R. LANESE,
M.D. KELLER
Relationship of air pollution to prevalance of lower respiratory
illness and lung function in Australian school children . . . 63
S.R. LEEDER, A.J. WOOLCOCK, J.K. PEAT,
C.R.B. BLACKBURN
Respiratory symptoms and ventilatory capacity in a cohort of
Londoners born in 1952 - 53 77
R.E. WALLER, A.G.F. BROOKS, M.W. ADLER
A study on health effect indices concerning population in
cadmium-polluted area 91
H. WATANABE, H. MURAYAMA
Clinical investigation in Northwest Quebec, Canada, of
environmental organic mercury effects 105
A.D. BERNSTEIN
-------
xv in
Scaling of annoyance in epideim'ologieal studies 119
B. BERGLUND, U. BERGLUND, T. LINDVALL
Tfie environmental and the protection of human health 145
E.J. BURGER, Jr.
The strengths and weaknesses of population studies in assessing
environmental health effects 161
D.L. JACKSON, V.A. NEWILL
Designing the exposure/response matrix in environmental
health studies 181
J.H. KNELSON
Dose-response relationships linking short-term air pollution
exposures to aggravation of cardio-respiratory illness • • • 193
R. CHAPMAN, J. FRENCH, J. FINKLEA, H. GOLDBERG
Effects of some approximations in analyses of radiological
response to coalmine dust exposure 211
M. JACOBSEN
Valutazione dell1 inquinamento da rumore prodotto da una
.centrale termoelettrica su di un quartiere . 231
F. MERLUZZI, A. GRIECO, M. BERNA, A. DONZELLI,
N. MORRESI
Mercury-nelenium association in persons exposed to inorganic
mercury .245
L. KOSTA, A.R. BYRNE, V. ZELENKO
UNTERSUCHUNG DER WIRKUNGEN AUF DEN MENSCHEN (Panel)
HUMAN EFFECTS STUDIES (Panel)
ETUDES DES EFFETS SUR L HOMME (Panel)
STUDI DEGLI EFFETTI SULL UOMO (Panel)
ONDERZOEKINGEN NAAR EFFECTEN BIJ DE MENS (Panel)
Methods for a european study on possible effects of air
pollution in children 263
Ph. BOURDEAU
Presentation et discussion du protocole d'une enquete
epidemiologique francaiee BUT lea relations entre pollution
atmospherique et affections respiratorr.es chroniques .... 269
D. BRILLE, P. BOURBON, JT LELLOUCH, P. ORIOL
Cber Veranderungen einiger biologischer Parameter bei Schulkindern
in Gebieten rait unterschiedlich starker Luftverunreinigung . . 279
P. SCHMIDT, L. PELECH, R. DOLGNER
Air pollution: Methods to study its relationship to respiratory
disease in British schoolchildren 289
L. IRWIG, D.G. ALTMAN, R.J.W. GIBSON, C. FLOREY
-------
XIX
Epidemiologic studies of adverse health effects associated
with exposure to air pollution 301
G.J. LOVE
Predictive models for estimating the health impact of future
energy sources 313
B.W. CARNOW
L1evaluation du risque pour la sante de la population exposed
aux poliuants atmospheriques irritants. Aspects me"thodologiques 333
B. BARHAD, M. CUCU
Study on long-terra effects on health of air pollution. . . . 339
J. KUMPF, M. ARHIRII, B. GRAB, M.J. SUESS
Panel discussion 351
TOXIKOLOGISCHE UNTERSUCHUNGEN
TOXICOLOGICAL STUDIES
ETUDES TOXICOLOGIQUES
STUDI TOSSICOLOGICI
TOXICOLOGISCHE ONDERZOEKINGEN
Toxicologic assessment of the health effects of sulfur dioxide
and sulfate particulates 371
J.W. CLAYTON jr.
Biological effects of the polychlorinated tjiphenyls in
nonhuman primates 385
J.R. ALLEN, L.A. CARSTENS, D.H. NORBACK
EXPOSITIONSMESSUNGEN (Panel)
EXPOSURE MONITORING (Panel)
MESURE DE L EXPOSITION (Panel)
MISURA DELL'ESPOSIZIONE (panel)
METING VAN EXPOSIE (Panel)
Recapitulation of environmental exposure ..." 401
T. D. ENGLISH
Sampling, analysis and composition of airborn particulate
material in Belgium
R. DAMS, M. DEMUYNCK
422
The champ air quality monitoring program **"
G. LAUER, F.B. BENSON
Studies of trace substances in an urban atmosphere 431
T.J. KNEIP, M. EISENBUD, M. KLEINMAN, D. BERNSTEIN
Air pollution problems in Latin America 451
R. HADDAD
Panel diaousaion 459
-------
EROFFNUNGSSITZUNG
OPENING SESSION
SEANCE D'OUVERTURE
SESSIONS INAUGURALE
OPENINGSZITTING
Vorsitzender - Chairman - President - Preeidente - Voorzitter
Simone VEIL - Ministre de la Sante (France)
-------
OPENING ADDRESS
OR, RUSSELL W, PETERSON
Chairman President's Council on Environmental Quality
United States of America
I wish to express, on behalf of the United States'delegation,
our appreciation for being able to work with all of you in this
Symposium, which promises to be helpful to all of us.
During the next five days, we shall be dealing with a wide
range of research areas related to man's health and the environ-
ment to which his health is so intimately bound. In arranging
for the present series of meetings, the Commission of the
European Communities, the World Health Organization, and the
U.S. Environmental Protection Agency have brought together an
eminent group of scientists from the fields of medicine, chem-
istry, toxicology, epidemiology, physics, engineering, and
mathematics. This spectrum of disciplines in itself gives some
indication of the vast resources and many factors we must bring
to bear on developing truly effective programs to protect health
and enhance our environmental quality.
Let us for a moment place some of these factors in perspec-
tive, because they will surely impinge upon our discussions this
week. Certainly any viable environmental protection program
must be predicated on several components. One of the most basic
of these relates to the promulgation of a firm body of environ-
mentally protective guides, laws, or codes. Sound legislation
or protective guidelines, based on thorough knowledge, should
enable us to assure that the protection of the health and welfare
of our citizens is given full scope within the entire range of
our activities. Science can contribute much to the development
of our environmental codes by expanding our knowledge of the
impact man exerts on ecological systems and the effects which
this impact in turn exerts upon man.
Authority to direct environmental protection efforts or
programs must be vested in a responsible and responsive agent.
Much has been said and written about the global nature of
environmental pollution, and I believe we are all realizing
that any one institution charged with environmental concerns
cannot limit itself to the problems within its own geographical
boundaries. Its decisions should be based upon balanced consi-
derations, insofar as possible, of a myriad of national and
world-wide, direct and indirect, consequences of any one action.
-------
Our experience in the U.S. continues to reveal that the choices
are not easy. We are becoming progressively aware, for example,
of the need to evaluate the total health, social, and economic
effects of available options for controlling pollution prior to
their selection as the measure of choice. Similarly, in this
time of dwindling energy supplies, we all face the task of
meeting our world-wide energy requirements while insuring that
public health is protected from potential adverse effects of
energy production; this requires that as we seek expanded and
alternative energy technologies, we exercise prudence lest we
trade an energy crisis for a health crisis.
A third component we must consider is the attitudes of
society toward environmentally protective measures. In any
decision-making process, the public should be informed of the
need for and rationale behind a given action. Thus broadly
speaking, we must ultimately include in our deliberations
the extent to which society is willing to pit the risks of
pollution against the benefits of an ever-growing world-wide
technology. The benefits are usually widely proclaimed. This
is not normally the case with risks. Modern man is barraged with
a variety of health insults - the risks of some are evident, the
risks of others can be estimated, but the risks remain unknown to
In the case of air pollutants, for example, early evidence
suggested that health risks were associated with exposure to
sulfur dioxide. We now have sufficient information to postulate
that these risks stem more directly from exposure to sulfate
compounds produced by atmospheric conversion of sulfur dioxide
than to the sulfur dioxide itself, and we can even make some
predictions about potential exposure levels from various station-
ary and mobile air pollution sources.
On the other hand, the precise, reproducible methods
necessary for monitoring exposure to and quantifying the risks
from certain air pollutants, notably nitrogen oxides, remain
to be established. Further, there exists a variety of other
substances, not nearly as widely studied as the major air poll-
utants, which are suspected as toxicants but for which we lack
adequate information to assess health risks. A case in point
relates to the carcinogenic properties of vinyl chloride
recently observed under conditions of occupational exposure in
the U.S. We have only meagre data on the health effects of
this Substance under a very particular set of circumstances and
virtually little data on its non-occupational implications.
We are thus confronted with some fundamental questions about
whether to regulate or control a number of materials, to what
levels we need to reduce their presence, and how to best control
those materials which have been identified as posing a known
risk to public health. Setting standards, establishing regul-
ations, issuing protection guides, or other similar forms of
-------
national and world-wide decisions require that we possess tech-
nically adequate and defensible data. So that rational environ-
mental and technological choices can be made, then, we must
obtain knowledge about the known risks associated with each
alternative open to us, the risks we can reasonably estimate, the
benefits of our actions, and the areas where our data are in-
sufficient for appropriate action. Because only a concerted
research effort can provide this information, the scientific
aspect of any environmental protection program should be con-
sidered an underlying and most essential component.
This of course places a tremendous burden of responsibility
on the scientific community. It is imperative, therefore, that
we as scientists develop means of obtaining technical information
which are comparable, that we develop experimental approaches and
methodologies on which we agree. But what are some of the areas
where we most critically need to formulate common approaches?
In assembling data on any given class of pollutants, we need
to measure the levels of pollution emanating froia the source or
sources we have identified. We need to monitor human exposure
and be able to estimate incremental exposures with time if the
source were to remain uncontrolled. To do this, we need sampling
and measurement techniques which are valid and reproducible not
only within the same laboratory but between the laboratories
using these techniques. We need comparable systems and techniques
for monitoring exposures as well as predictive models which are
equivalent. By the same token, we need analagous methods to
determine the environmental transport and possible transformations
of the pollutants in question, their effects on ecosystems, and,
most importantly, their impact on human health.
Perhaps at this point we might do well to ask ourselves,
"What should health intelligence assess?" We should certainly
consider how best to define susceptible populations, the validity
of employing animal studies to predict human health effects,
and what kinds of indicators are most sensitive for determining
human health effects. We should also consider how best to study
effects associated with an extended period of latency, with an
extended low level of exposure, or with subclinical manifesta-
tions. Inherent in our assessment must be an awareness of the
need to determine an adequate, but not excessive, margin of
safety for the establishment of prudent health protection guides.
Over the next few days, we shall be discussing such topics
as the health implications of exposure to noise, to pollutants
resulting from fossil fuel combustion, to pesticides, and to
other materials pervading the environment. Our fundamental
objective in studies of this nature is to determine exposure
and to discern exposure-effect relationships. Herein lies the
essence of the contribution we /is scientists must provide for
successful environmental protection programs and where we must
concentrate our attempts at achieving comparable results. We
will inevitably come to grips with the concept of threshold
-------
limits. We shall consider factors shaping the dose-response
curve, such as age and nutritional status of the subject, in-
ter current disease, duration, level, route, and method of ex-
posure, as well as the proper application of statistics. We will
also come to grips with the validity of experimental design for
studying the possibly non-threshold effects of pollutants such
as those which are carcinogenic, mutagenic, or teratogenic, and
whether the experimental design reflects the actual conditions
of use or exposure.
I am certain that specific points of disagreement on many
of the subjects I have mentioned will surface, but constructive
controversy is desirable. Only by communicating with one another
on our areas of scientific discord can we resolve them, both on
a national and an international scale. I would like to add
that I am optimistic about the attempts to achieve standardiz-
ation of methods and inter-laboratory quality assurance which
are already underway within and among a number of nations. The
air monitoring program of WHO is certainly a substantive effort
in this direction. In addition, the development of the WHO
Environmental Health Criteria Documents should prove to be of
real value in fostering the collaborative accumulation of
research information, the interpretation of data, and the iden-
tification of gap areas in our knowledge. Whether we are aff-
iliated with the academic, industrial, or governmental sectors,
we can all benefit from these efforts. I would urge that such
programs be strongly supported and extended to include as many
laboratories from as many nations as possible.
I believe that the present conference, with its emphasis
on comparable approaches for discerning exposure-effect relation-
ships, will prove to be another major step in promoting the mutua
scientific understanding we critically require for effective
health protection programs. I look forward to our collective
deliberations.
-------
OPENING ADDRESS
DR, P, J, HILLERY
Vice-President of the Commission of the European Communities
It is a great pleasure for me, on behalf of the Commission
of the European Communities - one of the co-organizers - to wel-
come you to this international symposium
The fact that so many prominent personalities in this field
have decided to attend is not only a great honour for the organ-
izing Committeee and the Commission, but it also underlines the
considerable interest there is in the problems involved in the
theme of this symposium which is devoted to the protection of
public health against environmental pollution.
We are also extremely pleased that so many other inter-
national organizations, as well as experts from nearly 5O
countries, are present. This demonstrates the world-wide inter-
est there is in studying and exchanging views on the scope of
the problems involved and in gathering together useful information
and scientific data.
I would like to extend a special welcome to the represent-
atives of the other European Community Institutions. The fact
that the European Parliament and Economic and Social Committee
are represented at this symposium is a cause of considerable
encouragement to the Commission.
Indeed the interest of the European Community in the prob-
lems of the environment was underlined by the Council of Minis-
ters' decision on 19 July 1973 to establish an environmental
action programme.
Various activities in this field have since been developed
and a number of practical measures have already been undertaken
with the aim of reducing environmental pollution, increasing
health protection and improving the quality of people's lives.
Since its inception in 1958, the European Community has
achieved a substantial increase in the rate of economic growth,
which has literally transformed the life and face of Western
Europe and brought to the great majority of its people higher
living standards and wider horizons.
However, this greater prosperity has not resolved all the
problems. On the contrary, it has exacerbated some and created
-------
8
new ones also. There are now problems caused by growth itself:
problems of industrial pollution, of a deteriorating environment,
of a conflict of values - in some cases between industry and
society - disruptions to the pattern of life, and many others.
Unless these problems can be resolved, economic growth
will fail to provide the quality of life which people rightly
expect. The aim of the European Communities' environmental pol-
icy, and of the Communities' Social Action Programme is to improve
the conditions in which people of our countries live and work.
To reach this desirable goal, it is essential that, as a
first step, scientific criteria be drawn up based on adequate
information, so that we have common reference data for the
elaboration of programmes and standards for the protection of
man and his environment. The knowledge of such factors as the
observed effects on human health of exposure to various pollu-
tants is essential to an objective evaluation of the environmenta
risks involved in any particular undertaking or work process.
It goes without saying that the approach to such studies
should be scientific and that conclusions should be based on
proven facts. Only in this way will it be possible to make the
precise assessments of the nature, and significance of the
health effects of pollutants.
Without going too deeply into the various interests concerned
in programmes to protect man's environment, I would say that
considering the risks to public health involved in pollution
and considering the benefits to be derived from the application
of scientific preventive measures, there is an indispensable
need for close collaboration between scientists, economists and
political leaders.
To tackle effectively problems of such a large scale at
regional or national level is nowadays virtually impossible.
To attempt it would be a waste of effort, energy and finance.
International collaboration is needed. For this reason we wel-
come the multi-national nature of this conference organized by
three co-sponsors of very different background.
This congress, where more than 25O papers will be presented
in more than 1O plenary and parallel sessions is structured in
such a way that a very broad scientific field is covered. We
realise that the meeting by itself is not a panacea. The prob-
lems are too complicated and the many existing local and regional
situations differ greatly whether considered from an industrial,
ecological, social, demographic or economic point of view.
But we all agree that there is sufficient common ground
for many of the problems and many useful lessons that can be
drawn from the pooling of information and experience.
-------
Much has changed since the first international congress
on public health was held in 1851. Since that time there has
been an exponential growth in scientific information and fun-
damental methods of research have been developed to include not
only deductive analysis but also inductive experimental tech-
niques, especially in the new discipline of toxicology.
Exchanges of views between scientists and the conclusions
drawn by them in their publications and meetings can now have
a very great impact on political discussions. It is therefore
essential that the work of scientists be concerned with critical
analysis so that the competent authorities can take decisions on
the basis of solid research and advice.
A special effort will, I trust, be made during this meeting
to achieve a better comprehension of the mechanisms and para-
meters of pollution and its effects on public health. If this
target is attained, then this Symposium can be considered a
success. In this context I might refer to the very successful
colloquium "Environmental Health Aspects of Lead" organized by
the European Communities in collaboration with the United States
Environmental Protection Agency, which was held in Amsterdam
nearly two years ago. In recent scientific literature reference
is very often made to both the proceedings and the papers of
that Symposium. This should encourage new efforts.
Before ending, I would also like to thank on behalf of the
Commission, the other co-sponsors of this Symposium - the United
States Environmental Protection Agency, and the World Health
Organization with its headquarters in Geneva and its regional
office in Copenhagen - for their co-operation during the pre-
paration of this congress. I have been informed that despite
the great geographical separation of the three organizations
and all the administrative complications arising from this, the
preparatory work has been done in an atmosphere of friendly and
close co-operation. We are very grateful for this.
I would also like to express my thanks to the Director
General of the Unesco and his collaborators for allowing the
use of this splendid congress building and also for helping us
with the necessary preparations.
In conclusion, I should like to express the hope that this
Symposium will not only be a scientific but also a human success,
and that all who have to take decisions concerning the protection
of man and his environment will draw useful lessons and infor-
mation from these discussions which will help them to carry out
their responsibilities in the best interests of mankind.
-------
OPENIMP ADDRESS
DR, T, ADEOYE LAMBO
Deputy Director-General, World Health Organization
It is with great pleasure that, on behalf of Dr. Mahler,
Director-General of the World Health Organization, I address
this distinguished gathering of scientists and health workers
who will discuss during this week one of the most important
aspects of the relationship of the environment to man's health.
The mere fact that more than 90O participants from about
50 countries are attending this Symposium, organized jointly by
the Commission of the European Communities, the United States
Environmental Protection Agency and the World Health Organization,
shows the increasing awareness of the medical and health profession,
the scientific community and the public at large of the need to
improve the quality of the environment in which we live and work.
The promotion of environmental health is one of the con-
stitutional obligations of the World Health Organization. The
WHO Constitution, adopted by the International Health Conference
held in New York in 1946, stipulates that the function of the
organization is, inter alia, "to promote, in cooperation with
other specialized agencies where necessary, the improvement ot
nutrition, housing, sanitation, recreation, economic or working
conditions and other aspects of environmental hygiene and to
promote cooperation among scientific and professional groups
which contribute to the advancement of health".
The subject of the Symposium belongs to the most important
aspects of environmental health because the assessment of health-
effects of environmental pollution and other hazards is one of
the primary criteria on which governments should base their
policies and decisions related to environmental control. These
decisions must be based on the best available scientific know-
ledge and not on the prevailing political atmosphere or emotional
reaction of the public, because they involve expenditure of
public funds of enormous magnitude.
I should like to give you only one example to illustrate
this point. A WHO study presented to the Twenty-fifth World
Health Assembly in 1972 revealed that in 197O, 1258 million
people in 9O selected developing countries did not have an
adequate community water supply, and that by 198O this will
increase to 1781 million, unless adequate measures are taken.
-------
12
To reach the target set by the World Health Assembly for the
Second Development Decade, i.e. that 60% of urban dwellers should
be served by house connexions and 4O% by public stand points,
and that 25% of the rural population should have reasonable access
to safe water, the total average annual investment required will
be of the order of US $ 1.3 billion.
Safe community water supply is, of course, only one element
of the total environmental quality improvement; other elements
such as waste management, air pollution control, food safety,
improvement of housing and recreational conditions, require
annual investments which are in some instances even higher than
the one mentioned in this example. The decision on how much
should be done to improve environmental conditions, obviously
requires as sound a scientific basis as it is possible to have
at a given time, and the major component of this basis is the
quantitative relationship between the level of exposure to
environmental hazards and the resulting effects on health and
ecosystems.
Although this Symposium will deal only with one specialized
aspect of environmetal health, that is with health effects of
environmental pollutants - and quite rightly so in order to enable
in-depth and critical examination of this problem - I should like
to point out that the World Health Organization looks at environr-
mental quality problems in a much wider context in agreement
with the definition on which the WHO constitution is based, that
"health is a state of complete physical, mental and social well-
being and not merely the absence of disease or infirmity."
In view of this, I feel it necessary to stress the need for
a holistic and ecological approach in social and health action
to improve the human environment and underline the importance
for human health and man's well-being of all aspects of the
human environment, physical, chemical and biological as well as
psychosocial and socio-economic factors? and the increasing
awareness that psychosocial factors, as well as the bio-physical
ones, can precipitate or counteract physical and mental ill healtt
profoundly modify the outcome of health action and influence the
general quality of life.
Since bio-physical factors, such as environmental pollutants,
will be extensively discussed during this Symposium, I should not
like to take your time in describing the direct influence they
may have on human health. it seems to me more appropriate to
point out that socio-economic and psychosocial influences, such
as poverty, affluence, population growth, mobility of populations,
urbanization, psychosocial stress and the cultural environment
can profoundly modify the effects of environmental pollution and
other bio-physical factors. In other words, there is a syner-
gistic relationship between these two sets of factors. Failure
-------
13
to consider various interactions of environmental factors, such
as the effects of other pollutants, smoking, alcohol, various
drugs and industrial or agricultural chemical exposures has been
common. Moreover, environmental pollutants should be considered
to be not only primary toxicants but also modifiers or co-factors
of natural diseases. In view of this complex relationship, it is
imperative to have the full participation of many disciplines
in many of these programmes.
There is evidence, for example, that the effects of air
pollution can be promoted by concomitant climatological conditions,
socio-economic factors/ stress situations and smoking habits.
Besides well-defined effects, such as the deterioration of res-
piratory function or incidental peaks in mortality and morbidity
during disastrous accumulations of pollutants in a prolonged
inversion period, there is evidence that chronic exposure to
low levels of atmospheric pollutants may be a factor in non-
specific respiratory disease, and may cause interference with
bodily functions, detriment of performance and, as suggested by
animal experimentation, premature ageing. Because of this, the
results of environmental pollution studies should be carefully
evaluated as to their possible interaction with biassing influ-
ences. There is a great need for a multi-dimensional prediction
model of the status of health and physiological functions in
relation to race, sex, age, climate, geography, nutrition, smok-
ing and other habits and living conditions.
Smoking cigarettes, i.e. the inhalation of cigarette smoke,
is a most important factor producing disease per se and in pre-
paring the respiratory tract, and probably other systems in the
body to react unfavourably to noxious gases, vapours and partic-
ulates. There is evidence that in the case of certain types of
exposure to asbestos, miner's bronchitis and the respiratory
function of iron and steel-foundry workers, the moderate smoker
already has an increased risk. He is less able to cope with the
superimposed impact of atmospheric pollutants upon the respir-
atory organs. Smoking habits should be carefully taken into
account not only when planning and performing air pollution
research, but also in the evaluation of the epidemiological and
experimental results.
To a considerable extent geography, topography and local
climate determine the fate of pollutants. This is essentially
true for changes in the temperature gradient of the atmosphere,
which can lead to the accumulation of air pollutants over a number
of hours or days. Persistent pollutants such as mercury, cadmium
lead and some organochlorine compounds may appear in several
environmental media (air, water, soil, food) and may reach man
through a variety of environmental pathways; they may accumulate
in the environment or may change their physical state and chemical
form; they may also be metabolized in micro-organisms and higher
plants or animals, producing sometimes more toxic compounds.
-------
14
Water that is drained from polluted land may become harmful if
used as drinking water for animals or man. This is increasingly
so in urbanized and industrialized areas with a high water demand.
Some air pollutants may contaminate surface waters and, in some
instances, increase the background concentration of certain toxic
substances causing, as a result of accumulation, an adverse eco-
logical condition; fluorine and arsenic are well-known examples.
Pollution of the marine environment is another concern
because of its implications for health (use for recreational
purposes, marine food products), but there are many other aspects
of this problem (ecological and economic) which are of great
importance.
In order to determine the effects on human health, human
beings have to be studied. Extrapolation of results on animals
to human beings has limited applicability. Studies on human
volunteers are limited by the duration and magnitude of exposures
that can be ethically used. Because of this, advantage must be
taken of those situations where human populations are exposed
to various levels and combinations of pollutants. Epidemiological
studies are used to evaluate such exposures.
Criteria for environmental quality for the protection of
human health must depend on information that defines the limits
of concentration of environmental pollution which permit main-
tenance of normal growth, structure an<^ function. Such criteria
are often derived from experimental studies on animals and, when
possible, on human volunteers. Another source is from exper-
iences in industry. Whatever the source, the information must
be unequivocal as to its reliability and relevance to exposures
of human beings to pollutants. This has not characterized much
of the past experimental data.
The insiduous adverse effects of environmental factors, such
as carcinogenic, teratogenic and mutagenic effects, that may
result from long-term environmental exposure are at present of
greater social concern than the acute or subacute effects be-
cause of a far greater uncertainty about the chronic adverse
effects. There is a strong pressure therefore, for a preventive
approach through predictions and evaluations of such effects
before human data on such effects have been obtained.
Health effects also to be considered are those that have a
bearing on the general welJ-being of man but without clear evident
of disease, at least as evaluated by methods now available. The
response may often consist of only a feeling of discomfort and
annoyance. It is not known at present what is meant by "annoy-
ance" and what bearing it has on human health in the long run.
A number of persons who report that they are annoyed also report
other effects such as disturbances of sleep, headache of nausea.
This complex and ill-defined emotional and behavioural reaction
-------
15
may cause sensitive or vulnerable persons to develop a psychoso-
matic disorder or may aggravate existing diseases. There is an
urgent need to study this further.
To facilitate the international comparisons of data neces-
sary to the solution of many of the problems of environmental
health, it is essential that investigators use standardized
techniques and scientific methodology of proven validity and that
the findings should be presented uniformly.
Economic and ecological effects of pollution must be taken
into consideration to assess the total effect on man. They
include effects on weather, vegetation, domestic and wild animals,
damage to aquatic and other food resources, destruction of mater-
ials, as well as soiling and impairment of aesthetic values and
recreational amenities resulting in an economic loss to man;
their appraisal is an essential prerequisite for setting up of
environmental quality standards and for regulatory control of
pollution.
In addition, the cost to society of controlling pollution
must be determined and then compared to the benefits resulting
from its control, and risks to health arising from inadequate
control. The degree of health protection desired above the
acceptable minimum level {which at the very least should assure
protection from death and illness directly resulting from ex-
posure to pollution) is a matter of political decision and is
generally purchased at increasing increments in control costs;
these costs are also directly related to the deadline by which
the control is to be achieved; the shorter the time deadline,
the higher are the costs. The level of protection desired
depends, of course, on the type and severity of health effects
and the size of population involved, but other factors (social,
cultural and economic) and other health priorities must be con-
sidered as well.
As the Director-General of WHO very recently reported to
the World Health Assembly, health oriented environmental poll-
ution control programmes differ from the traditional medical and
public health programmes, and the relevant policy formulation
is a new challenge because it involves not only science, tech-
nology, legislation and administrative enforcement, but also
economic policy as related to energy production and use, indus-
trial structure, land use, national and international trade,
consumption, and conservation of natural resources; and finally
problems of damages and compensation, allocation of costs and
fiscal problems of taxes and charges.
These are only a few of the problems encountered in the
assessment of health effects of environmental pollutants and its
administrative uses. I have no doubt that this week's delibera-
tions will make most valuable contributions to the elucidation
-------
16
of some of these questions, and thus provide a better scientific
basis for those who have to make decisions on measures to be
taken to prevent or control environmental hazards.
May I conclude by wishing you great success in your meeting
this week.
-------
17
DISCOURS INAUGURAL
DR, RENE MAHEU
Directeur General Unesco
(pr$8ent£ par M, le Dr. M. Batisse)
Le Directeur General de 1*Unesco, M. Rene Maheu, appele"
dans une autre salle par les travaux de notre Conseil Executif
qul se reunit en ce moment, regrette de n*avoir pu assister a
cette seance d'ouverture. II m'a pri6 de 1'excuser aupres de
vous et de vous accueillir ici en son nom.
II m'est particuliereraent agre"able de souhaiter la bien-
venue dans cette maison aux eminents sp£cialistes qui, sur
I1 invitation de trois organisations avec lesquelles 1"Unesco
entretient des relations etroites s'appretent a examiner les
progres recents dans 1*evaluation des effets de la pollution de
1'environnement sur la sante.
Ce sujet particulier n'est pas de ceux auxquels 1*Unesco
apporte par son action une contribution directe puisque, dans
le systeme des Nations Unies, les problemes d'ordre sanitaire
sont par definition du ressort de I1Organisation Mondiale de la
Sante. Hais ce n'est pas pour autant un sujet auquel I1 Unesco
peut rester indifferente, car tout se tient lorsqu'il s'agit
d'environnement, de developpement economique et social, du
devenir de 1'homme sur notre planete devenue trop etroite.
Aussi ne me paralt-il pas inutile de rappeler ici brievement
les grandes lignes ainsi que certains traits de 1'action de
1*Unesco en matiere d'environnement, et de souligner combien vos
preoccupations et les notres peuvent etre complementaires dans
ce vaste domaine.
Pour I1Unesco, 1'environnement de 1'homme est constituS
non seulement par 1*ensemble des facteurs et des situations
physiques et biologiques qui influencent sa vie, mais aussi par
les conditions sociales et culturelles diverses dans lesquelles
il se trouve. C'est done un concept tres large qui comporte a
la fois des elements objectifs et subjectifs, et qui s'etend du
cadre de vie journalier dans la maison ou 1' agglomeration jus-
gut au plan regional et global des interactions entre 1'homme
et la nature.
-------
18
Face a un sujet aussi vaste, 1'Unesco a entrepris, des sa
fondation, diffgrentes activitfis qui, au cours des vingt-cinq
dernieres annees, se sont dfiployees surtout dans le domaine des
sciences naturelles. Cependant 11 est apparu peu S peu neces-
salre de rSserver, dans les entreprlses de recherche et de for-
mation scientifiques relatives aux ressources naturelles et 3
1*environnement, une part de plus en plus grande aux facteurs
proprement humains; en d'autres termes, il s'agit dSsormais
d'accroitre la contribution des sciences sociales, des sciences
du comportement et de l'§ducation.
Cette Evolution est deja sensible dans le cas du programme
intergouvernemental et interdisciplinaire sur 1' nomme et la
biosphere (MAS). Ce programme de cooperation scientifique, qui
offre aux Etats membres la possibility d*examiner en commun des
probleroes concrets, constitue a 1'heure actuelle 1'element pr€-
pond£rant des activitfis de 1'Onesco relatives SL 1' environnement.
II est dirigS par un Conseil international de coordination au-
quel siegent les reprSsentants de 25 pays et ceux des principales
organisations internationales compgtentes, en particulier I1QMS.
II s'appuie sur des comitfis nationaux - il en existe 7O 3 1'heure
actuelle - charges d*assurer la participation des pays aux projet
internationaux du MAB qui les int£ressent.
Ces projets internationaux visent, d'une maniere g£n£rale,
a determiner les interactions entre 1'homme et les grands eco-
systemes qui constituent la biosphere, de facon a assurer a la
fois I1 utilisation la plus judicleuse possible des ressources
et la protection des int6rets a long terme de I'humanitg. Us
sont consacres notamment a 1*etude des effets gcologiques des
activit£s humaines sur les ecosysternes des forets tropicales
sur les terres a paturage, ou sur les systemes insulaires ou
montagneux, et ils examinent en meme temps certains processus
SL 1'intfirieur de ces systemes tels que les cons6quences 6colo-
giques de 1*utilisation des pesticides et des engrais/ ou les
interactions entre la transformation de 1'environnement et
1*evolution demographique et genetique. Parmi les projets du
MAB, 11 en est un qui vise a creer un r£seau mondial de zones
protegees, les "reserves de la biosphere", qui pourraient etre
utilisees non seulement pour la protection des esp^ces qu'elles
contiennent, mais aussi conme zones temoins et conme zones d'ex-
perimentation sur les effets de la pollution gfinerale des eco-
systemes et sur la surveillance continue de cette pollution.
Parallelement a ce programme interdisciplinaire sur I1 homme
et la biosphere qui est centre sur une mfithode ecologique,
1'Unesco poursuit deux autres grands programmes scientifiques
qu'il convient de mentionner ici consacres respectlvement a
1'eau et a la mer. Pour ce qui est de 1'eau, la decennie hydro-
logique Internationale, lancee en 1965 et axee principalement
sur la comprehension, 1*evaluation quantitative et 1*utilisation
rationnelle des ressources en eau, va prendre fin cette annee.
-------
19
Elle sera suivie, des 1975, par un programme hydrologique inter-
national oil I1 accent sera mis davantage sur les aspects qualita-
tifs, notamment sur la pollution des eaux et sur I1influence de
I'homme sur les ressources hydriques. Dans le domaine des
sciences de la mer, 1'Unesco assure le Secretariat de la Com-
mission oceanographique intergouvernementale et s
-------
2Q
En mime temps nous atteignons ici une autre dimension de
1*action/ tout aussi essentielle: celle de I1Education.
ration des relations entre 1'homme et la nature, qui est in-
dispensable au developpement harmonieux de 1'individu et des
soci6t€s, exige, en effet, une meilleure comprehension des
interactions qui sont a la base des systemes dont dependent la
vie et 1'gpanouissement de 1'horame. Pour sa part, 1'Unesco
s'emploie a favoriser cette Education a tous les niveaux, qu*il
s'agisse d'enseignement scolaire ou d*Education des adultes, de
la formation avanc£e d'Scologistes ou de la refonte de 1'ensei-
gnement donnfi aux architectes, aux urbanistes et, surtout, aux
ing£nieurs, qui sans doute apparaissent comme de grands pertur-
bateurs de 1'environnement, mais qui peuvent aussi, grace aux
conclusions de travaux comme les votres et les notres, apporter
des rem&des a la situation pr£sente.
Ces quelque indications sur les activit€s de 1'Unesco en
matiere d'environnement auront suffi, je I'esp&re, a vous montrei
1'intSret que nous attachons aux r£sultats de votre reunion.
De la rfialitS ou de la gravitS des dangers que les diverses
pollutions constituent pour la sant£ de 1'homme et que vous
allez devoir eValuer depend, en effet, pour une large part, la
nature et l'€tendue des mesures qui devront etre prises, ce qui
ne manquera pas d'influer sur le programme de I1 Organisation.
Nous vous demandons done de consid€rer cette maison comme la
v5tre et vous souhaitons le plus grand succds dans vos travaux.
-------
21
OPENING ADDRESS (translation)
DR, RENE MAHEU
Director-General, Unesco
(presented by Dr. M. Batisse)
The Director-General of Unesco, Mr. Ren6 Maheu, has been
called to attend a meeting of our Executive Board. He is sorry
he is unable to be present at this opening session and has
asked me to convey his apologies and welcome you on his behalf.
It is an especial pleasure for me to welcome here today
the eminent specialists who, at the invitation of three organ-
izations with which Unesco maintains close contact, are to ex-
amine recent progress in evaluating the health effects of
environmental pollution.
Strictly speaking, Unesco does not act directly in this
area since, by definition, within the United Nations system
health problems are the responsibility of the World Health
Organization. Yet Unesco cannot remain indifferent, for where
the environment is concerned, economic and social development
and the very future of man on this increasingly crowded planet
are at stake.
I therefore feel that in this context it would be worth-
while for me to run over the basic principles and some specific
aspects of Unesco action in the environmental field, and to show
how much, in this vast area, your concerns and ours are often
complementary.
Unesco considers man's environment to consist not only of
all the physical and biological factors and situations which
affect our lives, but also of the various social and cultural
conditions which surround us. The concept is thus a very broad
one with both subjective and objective elements extending from
everyday life in the home or town to interaction between man
and nature at the regional and global level.
Confronted with such an extensive field, Unesco has since
its inception,initiated various activities which, over the last
25 years, have been mainly concerned with the natural sciences.
However, it has become increasingly clear that research and
scientific training establishments concerned with natural
-------
22
resources and the environment should concentrate more and more
on specifically human factors - in other words/ from now on we
should see that the social and behavioural sciences and education
play a greater part.
This trend can already be seen in the intergovernmental
interdisciplinary programme on"Man and the Biosphere"(MAB).
This scientific cooperation project enabling Member States to
examine concrete problems jointly is the most important of
Unesco's environmental activities at the present time. It is
directed by an international coordinating council consisting of
representatives from 25 countries as well as from the WHO and
the other major international organizations concerned. It is
backed up by national committees - of which there are now 7O -
responsible for ensuring that countries take part in the inter-
national MAB projects which are of interest to them.
Generally speaking, the aim of these international projects
is to determine the interaction between man and the major eco-
systems which constitute the biosphere, with a view to ensuring
the most judicious use of resources together with the protection
of humanity's long-term interests. They concern the study of
the ecological effects of human activities on the ecosystems
in tropical forests, grazing lands and mountainous or insular
systems, as well as the examination of certain processes within
these systems such as the ecological consequences of the use of
pesticides and fertilizers or the interaction between changes
in the environment and demographic and genetic trends. One of
the MAB projects is aimed at creating a world network of pro-
tected areas or "biosphere reserves" which could be used both
for the protection of the species within them and as control and
experimental zones for studying and continuously monitoring the
overall effects of the pollution of ecosystems.
Side by side with this interdisciplinary programme which
examines man and the biosphere from the ecological viewpoint,
Unesco has two other major scientific programmes worthy of men-
tion here. One is concerned with fresh water, the other with
the sea. The International Hydrological Decade which started in
1965 and is mainly concerned with the understanding, quantitative
evaluation and rational use of fresh water resources, is to end
this year. In 1975 it will be followed by an International
Hydrological Programme which will focus on the qualitative as-
pects, particularly water pollution and man's influence on water
resources. In the area of the marine sciences Unesco provides
the Secretariat for the Intergovernmental Oceanographic Commissic
and seeks to promote a combination of observation, research and
training activities for studying and measuring marine pollution
and for improving knowledge of oceanic systems and problems in
coastal areas.
-------
23
I felt it important on this occasion to mention these inter-
national research programmes on the interaction between human
activities and the workings of and changes in the major natural
systems whether on land, in fresh water or in the sea, their
subject is the overall natural framework which contains the
specific effects of pollution on human health in which you are
interested.
However, it seems to me that apart from this, we have at
least two other points of contact - the quality of life and
education.
It may well be that "the quality of life" has not yet been
properly defined - it certainly covers many aspects giving rise
to varying subjective interpretations and evaluations. Yet
there are also some extremely important objective aspects in-
cluding, at the top of the list, the health considerations on
which your work is to be based. In any event, there is no doubt
that public opinion is showing increasing concern for the var-
ious aspects of the problems in question. Indeed, together
with underdevelopment, this is probably the subject which is
now the greatest burden on man's conscience. What is known as
the environmental crisis is, in fact, basically one aspect of
a civilization crisis which has arisen, on the one hand from
man's fears about the survival of the species as a result of his
aberrant behaviour towards nature and natural resources, and,
on the other hand,from his spiritual dissatisfaction derived from
the way of life he has created for himself. The crisis is part-
icularly apparent in the ever more monstrous urban areas where
man really seems to be "denaturing" himself and losing his true
identity.
This is why Unesco has undertaken this effort to stimulate
and coordinate thought, research and discussion with a view to
gaining a better understanding of man's immediate environment
and helping design a better setting and habitat for our lives.
In this field, appreciation of the quality of the environment,
psychological, sociological and cultural factors, and the motives
of individual and collective behaviour are becoming essential
to the guidance of public action, and, I believe, link up with
the issues which concern you today on the subject of human
health.
This brings us to another, equally vital aspect of our
action - education. Improved relations between man and nature
so essential to the harmonious development of the individual and
of societies - are in fact dependent on a better understanding
of the interactions underlying the systems on which man's life
and development depend. For its part, Unesco strives to promote
such education at all levels, whether it be through schools,
adult education, the advanced training of Geologists, or through
a new conception of the training of architects, urbanists and,
-------
24
above all, engineers who doubtless appear as major perturbers of
the environment,but who could also remedy the present state of
affairs by taking note of the conclusions drawn from work such
as yours and ours.
These few indications about Unesco activities in the environ-
mental field will, I think, convince you of our interest in the
results of your meeting. The nature and extent of the measures
adopted will depend to a large extent on the existence or gravity
of the various pollution hazards to human health which you are
about to evaluate. This analysis will no doubt influence our
Organization's programme.
We hope you will feel at home here and we wish you every
success in your work.
-------
25
DISCOURS INAUGURAL
SIMONE VEIL
Ministre de la Santg - France
Les degradations subies par 1'environnement du fait des
transformations du monde moderne et les risques que cette situa-
tion fait encourir a la santS des populations, constituent 1'un
des principaux sujets de preoccupation de notre epoque.
Depuis le debut du siecle, les pouvoirs publics avaient
orients de maniere priorxtaire leur action vers les aspects
bacteriologiques de la pollution du milieu naturel, et poursuivi
notamment la lutte contre les grandes endemies, telles que le
cholera ou la fievre typholde. A cet £gard, du fait du develop-
pement de la vaccination et de I1hygiene, ainsi que des mesures
d'assainissement prises contre les vecteurs des maladies, ron-
geurs ou insectes, des progres considerables ont pu etre rea-
lises, meme si, depuis peu, de nouvelles inquietudes se font
jour, en raison de facteurs recents tels que la resistance ac-
crue de certains microbes aux antibiotiques ou le developpement
des voyages intercontinentaux.
Or, depuis a peu pres une trentaine d'annees, la pollution
de 1'environneraent ne cesse de s'accroitjre dans les soci§tes
industrialists et cet Stat de fait constitue pour la sante des
etres humains une menace de plus en plus redoutable. II est
des £ present certain, en effet, que nombreux et graves sont les
troubles qui sont lies a la degradation du milieu, qu'il s'agisse
de la pollution de I1air, de 1'eau, des produits alimentaires ou
encore de 1'effet du bruit, ou -de 1'action des micro-polluants,
recontrSs dans ses diffbrents milieux.
Les raisons de cette situation sont a present bien connues
et font dejck 1'objet de larges discussions dans I1 opinion pub-
lique. Elles tiennent essentiellement a cinq grandes causes:
le rejet d'Snormes quantites de substances toxiques par certaines
industries, 1'utilisation massive en agriculture de multiples
produits chimiques, et notamment des engrais et des pesticides,
le surpeuplement de quelques grands centres urbains dans lesquels
se trouvent concentrees des centaines de milliers de personnes,
la circulation intensive des vShicules individuels dans les
villes et enfin 1*innovation technologique qui fait apparaitre
sur le marche, chaque annee, 4OO a 50O substances chimiques
nouvelles dont il est rarement possible de prSvoir toutes les
consequences eventuelles sur la sant6 de la population.
-------
26
Deux tentations, des lors, se font jour. Celle de 1'im-
mobilisme qui tendrait S entraver systematiquement tout nouveau
progres technique de peur de nouvelles nuisances/ et celle, au
contraire, du fatalisme, qui inciterait a s'en remettre a la
providence, ou aux decouvertes futures de la m£decine, du soin
de coropenser toutes les consequences negatives pour la sante
publique des transformations de notre societ§.
Ces deux attitudes sont £galement excessives. La premiere
aurait conduitr au milieu du siecle dernier/ a ne jamais creuser
un tunnel de chetnin de fer pour 6viter aux voyageurs le risque
de pfirir €touff€s. La seconde, conjugu£e avec une confiance
aveugle dans la technique, peut mener a certains desastres tels
que ceux de Minamata ou de la Thalidomide.
Or, l*€tat d1 esprit qui doit aniroer les milieux scientifi-
ques et les responsables politiques des divers pays est tout
autre. II s'agit en effet, a present, de considerer avec objecti-
vitfi et rSalisme, et en substituant la rigueur et la precision
scientifique a 1'irrationnalitS des impressions et des angoisses,
ce probleme majeur des risques pour la sante du fait de 1'envi-
ronnement. La reunion de votre symposium manifeste avec tout
1'eclat desirable notre volonte commune d'aboutir a ce rfisultat.
Incontestablement, la situation en matiere de maladies
et de troubles dus a la degradation du milieu naturel est preoc-
cupante.
La pollution de 1'air, provoquee par I1action conjuguge de
1*Industrie» du chauffage en milieu urbain, des rejets de subs-
tances toxiques par les v^hicules automobiles et de I'incin6ra-
tion des de'chets, entralne dans certaines dres g£ographiques une
agression pulmonaire dont les consequences peuvent §tre parti-
culierement redoutables. II semble en effet qu'il y ait une
correlation Stroite entre la pollution atmospherique et certaines
maladies des bronches, et en particulier 1'emphyseme pulmonaire
et le cancer du poumon. Chacun se souvient, par exemple, du
smog de Londres de 1952 qui a provoqu€ pres de 5.OOO deces en
une semaine du fait del1aggravation subite de maladies cardia-
ques et respiratoires.
La pollution de 1'eau est a peine moins redoutable. Provo-
quee aussi bien par 1'excessive concentration urbaine que par
le rejet de de'chets industrials, elle peut entralner des intoxi-
cations aigues et des maladies chroniques du fait de certains
616ments micro-polluants, tels que les hydrocarbures, les pesti-
cides ou les detergents, dont les effets peuvent ne s'averer to-
xiques qu'aprds plusieurs ann^es d'absorption d'eau polluee.
-------
27
Cependant la degradation de la qualitS de I1air et de 1'eau
ne sont pas les seuls sujets de preoccupation en matidre de
sant§ publique. Certaines substances, rfipandues dans tout notre
environnement - 1'eau, I1air et les aliments - constituent 6ga-
lement un p6ril redoutable, qu'il s'agisse de I'amiante, du
plomb, du mercure, du cadmium, du dioxyde de soufre, ou des pes-
ticides organophosphorSs dont la toxicite peut se trouver multi-
plied par 1'effet de phenomenes de synergie encore mal connus.
A ces divers elements de pollution d'autres facteurs viennent
s'ajouter, tel que notamment le bruit, qui dans certaines cir-
constances peut diminuer 1'acuitS auditive et atteindre 1'Squi-
libre nerveux de 1'homme, c'est-a-dire le diminuer et le dS-
grader.
C'est pourquoi, apres une longue p£riode de temps od il
semblait que les dficouvertes de la medecine n'allaient plus ces-
ser de faire progresser l'esp£rance de vie des homines, 1'in-
qui£tude a nouveau grandit.
La tache a laquelle seront consacres les travaux de votre
symposium est done d'une importance capitale. Elle est aussi,
nous le savons tous, d'une extreme difficulte.
L1incidence nocive de la pollution de 1'environnement sur
la santS humaine est a present incontestable. Particulierement
significatif est a cet egard ce passage d'une communication de
savants ame'ricains de l'universit§ de Cleveland qui d£clarent
"il y a maintenant peu de doute que beaucoup de maladies chroni-
ques, jusqu'a present consid€r€es comme spontan€es, et en parti-
culier le cancer, sont provoquees par des polluants de 1'envi-
ronnement ". Mais si, globalement, la toxicltS de la pollution
est indeniable il s'agit d'un phSnomSne tellement impalpable
dans la multiplicity de ses manifestations, aussi bien en ce
qui concerne les agents polluants que les personnes atteintes
dans leur sante, que les probldmes pos§s aux autorite's scienti-
fiques sont d'une complexity a peine imaginable. Pour ne prendre
que le seul exemple de 1'eau, combien de difficultes a risoudre,
qu'il s'agisse de la detection rapide de pollutions accidentel-
les, de la contamination des eaux souterraines, du conditionnement
des eaux de table ou du developpement de certaines maladies
comme 1'hepatite virale ...
Cette tache gigantesque, vous 1'accomplissez, dans vos
sp£cialit€s respectives, par les mSthodes scientifiques les
plus varie"es, de la recherche fondamentale a la recherche ap-
pliquSe, de 1'envergure des enquetes en mati§re d'^pidemiologie
a la poursuite de 1'infiniment petit pour la detection des
substances les plus fines, du questionnaire a 1'examen somatique,
et des experiences realis^es sur les animaux aux Etudes consacrees
notamment en milieu hospitaller, a 1'etiologie probable des
divers troubles affectant les malades. Ce symposium doit per-
mettre a present d'op^rer une confrontation g^nSrale du rSsultat
de toutes ces recherches.
-------
28
Je crois, a cet egard, important de souligner tout 1'interet
qui s1attache a une harmonisation plus complete des donnees
scientifiques en matiere de pollution et de repercussion sur
la sante. Qu'il s'agisse notamment des protocoles operatoires
des enquetes epidemiologiques, des methodes de mesure des para-
metres de la pollution ou du choix des types de population
retenus pour I1etude d'un phenomene, il est en effet essentiel
que tous les milieux scientifiques interesses s1accordant a
utiliser les memes definitions et les memes instrumets de mesure
afin que leurs travaux puissent aboutir a des resultats parfaite-
ment clairs et non contestes.
L'objectif supreme de toutes ces etudes si diversifies
consiste evidemment dans la possibility de determiner en toute
connaissance de cause des normes precises pour reglementer
1'utilisation des substances polluantes, deteeter les accidents
ou les abus possibles, et les sanctionner s'il y a lieu.
II y a bien longtemps que les pouvoirs publics se preoc-
cupent de ces problemes puisqu'en matiere de pollution de 1'eau
et de I1air, les premieres mesures importantes sur le plan juri-
dique remontent en France au tout dfibut du siecle et que nom-
breux sont les textes legislatifs et regleroentaires qui sont
intervenus surtout depuis une quinzaine d'annees dans le domaine
de la protection de 1'environnement. Mais il faut aller encore
beaucoup plus loin/ et notamment ameliorer 1'etendue, la fiabili-
te et 1'automatisation des reseaux de controle pour aboutir aux
reseaux de surveillance globale de 1'environnement recommandes
par la conference des Nations-Unies tenue en 1972 a Stockholm.
C'est en ce sens que s'orientent les pouvoirs publics fran^ais
qui ont mis en oeuvre a Rouen un reseau continu de mesure de la
pollution atmospherique qui va etre egalement install^ a Fos-
sur-Mer, en vue de declancher une eventuelle alerte.
On conceit aisement que 1'ensemble de cette reglementation
pose un nombre considerable de problemes techniques notamment en
ce qui concerne la constatation du caractSre polluant de certains
produits et leur detection dans le milieu ambiant. II se pose
egalement des problemes juridiques, dans la mesure ou il faut
fixer des normes et les sanctionner, et aussi de problemes
economiques, puisqu1aussi bien la mise en place de dispositifs
antipolluants peut dans bien des cas grever plus ou moins lour-
dement le cout de certains produits. Toutes ces difficultes ne
peuvent 6tre resolues qu'au prix d'une collaboration etroit^
entre les milieux de la recherche scientifique, les administra-
tions et professions touchant au domaine de la sante, plusieurs
dfipartements ministeriels, et en particulier 1'environnement,
1*Industrie et 1*agriculture, et aussi, et plus encore peut-
etre, la population dans son ensemble.
-------
29
Dans ce domaine capital du choix entre d'une part, 1'acce-
leration sans limite du progres technologique - avec ses conse-
quences bSnefiques sur la progression du niveau de vie de 1'ensem-
ble des citoyens et ses "retombees" nefastes pour leur sante -
et d'autre part, le controle ainsi que, le cas echeant, la
limitation de cette evolution dans un souci de prudence,il est
essentiel que 1'opinion soit completement informee de tous les
elements du probleme et decider du genre de solution qui lui
paralt le plus opportun. L1education, 1'information, la discus-
sion, la persuasion me paraissent en effet des arguments beau-
coup plus decisifs que la contrainte. Si I1hygiene corporelle
par example, est un facteur essentiel de la sante d'une popula-
tion, mieux vaut evidemment en persuader 1'ensemble des citoyens
que de vouloir en ce domaine controler tout un chacun et saction-
ner la carence des plus negligents ...
Les autorites de sante publique ont, pour leur part, un
role primordial a reroplir dans 1'Elaboration et la mise en
oeuvre d'un programme rationnel de lutte contre les nouvelles
agressions du monde moderne. Les donnees medicales, dans la
raesure ou elles sont susceptibles de prot£ger la sante de 1'etre
humain contre les agressions multiples de 1'environnement, doi-
vent en effet prendre le pas sur toutes les autres considerations.
Dans la definition d'une politique de sauvegarde de 1'envi-
ronnement, les autorites sanitaires constituent certainement
"1'interlocuteur privilegie" des services mis en place pour as-
surer la coordination des actions entreprises en ce domaine.
C'est dans cette optique que le Ministere francais de la sant£
aentrepris, confornament aux indications de 1'Organisation
Mondiale de la Sante, de restructurer ses services d1hygiene
publique. En outre 1'Institut National de la Sante et de la
Recherche M6dicale consacre une part importante de ses activites
a des Etudes sur les consequences de la degradation de I1envi-
ronnement sur la sante.
Telles sont quelques unes des observations que je tenais
a presenter 3 1'ouverture de ce symposium, en ma qualite de
Ministre Fran?ais de la Sante. Je ne doute pas qu'il ne con-
tribue puissamment a faire plus de lumiere sur des problemes
que nous avons encore beaucoup de peine a apprehender dans toute
1'etendue de leur complexite. Plus tard, sans doute, sera-t-il
possible d'Svaluer de maniere plus precise le cout economique de
la lutte a mener sur tous les fronts contre la pollution de
milieu naturel et aussi de determiner au terme de dSbats publics
particuli&rement approfondis - une notion de risque acceptable en
matiSre d1environnement comparable a celle qui est deja retenue,
au moins de maniere implicite, en matiere de vaccination obli-
gatoire.
-------
30
Aussi les preoccupations qui sont les notres ne doivent-
elles en aucun cas nous inciter au pessimisme. A partir du
moment, en effet, oft I1ensemble de la population prend conscien-
ce du peril que represente pour elle la degradation du milieu
ambiant, cette Evolution des esprits autorise les plus grands
espoirs. Ainsi, d£ja, dans notre pays/ des rivieres totalement
pollutes ont-elles pu permettre a nouveau le retour de la vie
animale et vSgStale, les produits non bio-degradables ont-ils
6t6 presque totalement interdits dans les detergents, et des
r eg lenient a tions plus protectrices sont en cours d1 Elaboration
dans plusieurs domaines importants relevant de la Santfi Publique
et de 1'Environnement.
Dans cette gigantesque bataille engagee pour la protection
de 1'etre humain contre les menaces multiples de son cadre de
vie, vos travaux vont apporter une contribution de tout premier
plan. Car beaucoup, nous le savons tous, reste S faire, pour
que la planete terre qui constitue - avec sa mince couche
d'humus et son fragile vernis de civilisation - notre seule
chance de survie, ne soit plus considSre'e comme une immense
carriere ouverte 5 toutes les convoitises, mais qu'elle devienne
enfin, pleinement et d€finitivement, la terre des homines.
-------
31
OPENING ADDRESS
SIMONE VEIL (translation)
Minister of Health - France
One of the major causes of concern to the present generation
is the deterioration of the environment as a result of modern
industrial development and the health hazards to which the pop-
ulation is consequently exposed.
Since the beginning of the century, public authorities have
given priority to the bacteriological aspects of pollution of
the natural environment, and have emphasized control of the main
endemic diseases such as cholera of typhoid. Significant progress
has been made in this field through the development of vaccination,
improvements in hygiene and campaigns against disease carriers
such as rodents or insects, though of late further problems have
arisen as a result of new factors such as the increased resistance
of certain microbes to antibiotics or the growth of intercontin-
ental travel.
For some thirty years, environmental pollution has been
growing steadily in industrialized societies, and this situation
poses an increasingly grave threat to the health of human beings.
Indeed it is already established that many serious effects are
associated with the deterioration of the environment in its
various forms: air and water pollution, contamination of food-
stuffs, the effects of noise or the action of micropdllutants.
The reasons for this situation are now well known and are
already the subject of a wide-ranging public debate. There are
five main causes: disposal of huge quantities of toxic substances
by certain industries, massive use in agriculture of .numerous
chemical products, especially fertilizers and pesticides, over-
population of a number of large urban centres in which hundreds
of thousands of people are concentrated, heavy private traffic
in towns, and finally technological progress, which introduces
each year on the market four to five hundred new chemical pro-
ducts, whose possible effects on the health of the population
can rarely be foreseen.
Two temptations now arise. The first is the temptation to
freeze all action, consistently blocking all technical progress
for fear of creating new sources of pollution. The second is
the temptation to adopt a fatalistic approach, to rely on prov-
idence or on future medical discoveries to counteract all the
negative effects on public health of the development of our
society.
-------
32
Both attitudes are extreme. Had the first been current in
the middle of the last century, no railway tunnels would have
been driven lest passengers died of suffocation. The second, in
combination with blind trust in technology, can lead to disasters
such as Minamata or thalidomide.
Scientific circles, however, and politicians exercising
responsibility in the different countries should tackle the
matter in a very different frame of mind. An objective and rea-
listic approach replacing irrational impressions and fears by
the discipline and precision of science must now be adopted to
the major problem of environmental health hazards. This sym-
posium is a dramatic assertion that-, it is our common purpose to
achieve this end.
The situation with regard to illnesses and disorders caused
by deterioration of the natural environment undoubtedly gives
cause for concern.
In certain geographic areas, air pollution resulting from
the combined effects of industry, urban heating, toxic emissions
by motor vehicles and waste incineration leads to pulmonary
effects with particularly serious consequences. There appears
to be a close correlation between atmospheric pollution and
certain bronchial diseases, especially emphysema and lung cancer.
Who, for example, does not remember the London smog of 1952 which
caused nearly 5,OOO deaths in one week as a result of the aggra-
vation of heart and respiratory diseases?
Water pollution is scarcely less serious. It is caused
both by excessive urbanization and by the discharge of industrial
waste, and can lead to acute poisoning and chronic illnesses
through the action of certain micropollutants, such as hydro-
carbons, pesticides or detergents, the toxic effects of which
may not become apparent until polluted water has been absorbed
over a period of years.
Deterioration of air and water quality, however, are not
the only public health problems. Certain substances dispersed
throughout our environment - in water, air and foodstuffs -
also pose a grave threat. They include asbestos, lead, mercury
cadmium, sulphur dioxide and organo-phosphate pesticides, the
toxicity of which may be multiplied by as yet unknown synergic
effects. In addition to these various pollutants, there are
other factors, such as noise, which in certain circumstances
may lead to impaired hearing and affect the nervous balance by
upsetting and damaging the human nervous system.
-------
33
For all these reasons anxiety is again increasing after a
long period in which it seemed that medical discoveries would
continue indefinitely to improve life expectation.
The task to which this symposium is devoted is therefore of
vital importance, but as we are all well aware, it is also very
difficult.
It cannot now be denied that environmental pollution has a
harmful effect on human health. In this connection the following
passage from a statement by American scientists from the Univer-
sity of Cleveland is of particular significance: "There is now
little doubt that many chronic illnesses, especially cancer,
which until now were thought to develop spontaneously, are
caused by environmental pollutants." But while in general terms
the toxic effects of pollution are undeniable, the phenomenon is
so elusive and appears in so many forms both as regards the pol-
lutants and the persons affected that the scientific authorities
are faced with problems whose complexity can scarcely be imagined.
To take water alone as an example, how many difficulties must be
overcome, whether these lie in the rapid detection of accidental
pollution, contamination of subterranean water, treatment of
drinking water or the development of certain diseases such as
virulent hepatitis.
This is an enormous task, which you accomplish in your re-
spective fields by the most varied scientific methods ranging
from pure to applied research, from large scale investigations
in epidemiology to highly detailed work on the detection of
trace substances from the questionnaire to somatic examination,
and from experiments on animals to studies devoted, in clinical
surroundings, to the determination of the probable etiology of
the various symptoms affecting victims of disease. The purpose
of the present symposium is the general confrontation of all
these research activities.
I should like at this point to emphasize the importance
of fuller harmonization of scientific data regarding pollution
and its effects on health. Whether we consider case reports
in epidemiological inquiries, methods of monitoring pollution
parameters or the choice of population groups for the study of
a particular phenomenon, it is essential that all scientific
bodies concerned agree on common definitions and common measuring
instruments so that the results of their studies may be absolutely
clear and beyond dispute.
The ultimate aim of this whole range of studies is obviously
to have a basis of knowledge on which to establish precise stand-
ards allowing the use of pollutants to be controlled, accidents
and any violations of regulations to be detected and sanctions
imposed where appropriate.
-------
34
The public authorities have long been concerned with these
problems, and the first major legal measures in France with
regard to air and water pollution date from the very beginning of
of the century, while numerous laws and regulations concerning
environmental protection have been promulgated, especially in the
last fifteen years. Much more remains to be done, however,
especially improving the comprehensiveness, reliability and
automation of monitoring systems with a view to establishing the
type of global environmental monitoring system recommended by
the 1972 United Nations conference in Stockholm. The French
authorities are moving in this direction by establishing at Rouen
a continuous monitoring network for atmospheric pollution. The
same system will also be set up at Fos-sur-Mer to give the alert
if necessary.
It will be readily appreciated that this corpus of legis-
lation raises a large number of technical problems especially
determining that certain products are pollutants and detecting
them in the environment.
It also poses legal problems, as standards have to be set
and enforced, and economic problems, as the installation of
pollution control plant frequently affects the cost of certain
products to varying degrees. All these difficulties can be
solved only if there is close cooperation between the scientific
research community, the administrative services and professions
concerned with health, several government departments, especially
the Departments of the Environment, Industry and Agriculture,
and most important of all perhaps, the population as a whole.
Now that we are faced with this crucial choice between
unchecked acceleration of technological progress - with its
beneficial effect in improving the standard of living and its
harmful "fall-out" affecting health - and surveillance and if
necessary restriction of this development as a precautionary
measure, it is essential for the public to be fully aware of all
aspects of the problem and to choose the form of solution it
finds most suitable. Education, information, discussion and
persuasion seem to me more effective than compulsion. If, for
example, personal hygiene is a major factor affecting the health
of a population, it is obviously better to convince all the
citizens of its necessity than to attempt to check each indi-
> vidual and punish the worst offenders.
The public health authorities have a central rolf* to play
in devising and implementing a rational programme to control
the new hazards arising in the modern world. Insofar as they
protect health against the numerous environmental insults,
medical requirements must take precedence over all other con-
siderations.
-------
35
In drawing up an environment policy, the health authorities
must certainly have a "special relationship" with the departments
set up to coordinate all action in this field. This is the ap-
proach which the French Ministry of Health has adopted in reorgani-
zing the public health departments in accordance with the recom-
mendations of the World Health Organization. The National Insti-
tute of Health and Medical Research moreover, devotes much of
its energies to investigating the effects on health of environ-
mental pollution.
These, are a few of the points which, as French Minister of
Health, I felt I ought to raise by way of opening this symposium.
I am sure this assembly will make a major contribution to eluci-
dating problems which are still difficult to grasp in all their
complexity. At a later date it will no doubt be possible to
arrive at a more precise estimate of the economic cost of the
war to be waged on all fronts against pollution of the natural
environment and also to define after the fullest public debate
the concept of an acceptable degree of risk in environmental
matters similar to that already applied, at least implicitly, in
the field of compulsory vaccination.
The problems with which we are concerned thus give no
grounds for pessimism. Indeed, if the population as a whole
becomes aware of the danger to which it is exposed through pol-
lution of the environment, this evolution of attitudes is justi-
fication for the very highest hopes. In France for example,
animal and vegetable life has already returned to rivers which
were totally polluted, virtually all use of non-biodegradable
products in detergents has been prohibited, and in several areas
under the responsability of Health and Environment regulations
designed to afford a higher degree of protection are being drafted.
Your work will play a vitally important role in this vast
battle to protect the human bieng against the multifarious threats
posed by the world around him. As we are all well aware, much
is still to be done before the planet Earth, which - with its
fine layer of humus and its thin veneer of civilization - is
our only hope of survival, is no longer regarded as a huge quarry
to be exploited for the satisfaction of every form of greed but
becomes, in the fullest sense, the home-land of mankind.
-------
POPULATIONSUNTERSUCHUNGEN
POPULATION STUDIES
ETUDES SUR LA POPULATION
STUDI SULLA POPOLAZIONE
BEVOLKINGSSTUDIES
Voreitzender - Chairman - President - Presidente - Voorzitter
D. RALL (U.S.A.) '
Stellvertretender Vorsitzender - Vice Chairman - V-iee-prSs-ident
Vicepresidente - Vice-Voorzitter
W. GOERKE
(Bundesrepublik Deutschland)
-------
39
THE EFFECT OF ENVIRONMENTAL FACTORS ON HEALTH -
METHODS FOR POPULATION STUDIES
W, W, HOLLAND
Department of Clinic?! Epidemiology and Social Medicine,
St. Thomas's Hospital Medical School, London, United Kingdom
ABSTRACT
The paper deals with the epidemiological methods and
sources of information which are useful for the investigation
of the influence of environmental factors on health. Sugges-
tions are made for developments in identification of new health
hazards and of monitoring the health of population groups.
-------
As we all know, there are various techniques that can be
used to study the effect of environmental factors on the health of the
population. Certain information, such as the nature of the condition
under investigation, how frequently it occurs and who it is most likely
to affect, oust be considered when a study is being designed so that
the most appropriate technique can be chosen.
When a specific environmental factor can be isolated, its
effect on health can be detected relatively simply. In the studies of
the development of cancer of the lung in ashestos and nickel workers
(Doll [l] ; Uewhouse [2] ), for example, the investigators knew that
they were dealing with a specific industrial process which had been
used in the manufacture of these substances for many years. Cancer of
the lung is a relatively rare condition and in order to establish its
incidence in the population at risk it is necessary to study either very
large population groups or small groups of individuals exposed for a
very long period of time. Furthermore, since the condition is
uniformly fatal, it is sufficient to identify individuals who die rather
than those who develop symptoms. In these particular studies,
therefore, the number of years a given industrial population was at risk
from either asbestos or nickel and how many cases of cancer of the lung
had occurred were calculated and compared with the incidence of cancer
of the lung in the general population over the same time period. Results
showed that the risk of developing lung cancer was very much greater in
both asbestos and nickel workers than in the population as a whole.
One assumption made in this study was that the smoking habits of these
workers were similar to those of the population at large. However, as
the excess risks were high this was unlikely to have changed the outcome
subs tantially.
In most investigations of environmental health hazards,
however, the picture is obscured by the variety of factors involved.
The general techniques that are currently used by epidemiologists in
such investigations follow classical lines.
A clue to the need for investigation is often first found in
published statistical information, perhaps by examining mortality ratios
for people in different industries, living in different towns or regions,
or from different social class or economic groups. This may be
followed by scrutiny of morbidity or illness statistics and although the
-------
41
quality of this type of data varies greatly, it may help to form
precise hypotheses. It creates the possibility, for example, of
examining hospital discharge rates for different areas, or sickness
absence rates in a particular industry or for those living in a
particular region or town. Disability or early retirement rates are
usually available for the working population and can provide crude
evidence of a health hazard related to the working environment.
Let us now , however, consider the far more common way in
which one undertakes a study, namely the investigation of the prevalence
or incidence of a condition in a defined population. In any
population the frequency of disease at any moment in time depends on the
rate of inception, recovery and recurrence and on mortality among those
afflicted by it. As it is inception that is the most relevant in
epidetriological studies of cause of disease, records are considered of
prime value because they reflect the onset of the disease. Too often,
however, such data are either absent or inadequate and the only
alternative is a survey of the current prevalence of disease within the
community.
Prevalence studies are designed to estimate the number of
people with a particular disorder of defined severity in a population at
one point in time. In practice this means the number counted not at
any one point in time but over a certain short period. The first step
then is to define the population to be studied. A possible difficulty
here is the choice of area and this will, of course, depend on the
particular hazard to be investigated. Some years ago, for example,
we were interested in determining the influence of air pollution on the
aetiology of chronic bronchitis. The best way to answer this question
would obviously have been to investigate random samples of the
population from areas with different levels of pollution. However,
since we knew that social class and occupational factors influenced the
onset of this condition, re would have had to look at very large
numbers of individuals in each area to be able to exclude their possible
effects on differences in prevalence. To avoid this difficulty, we
chose individuals in the same occupation, vhich was not liable to be at
risk from chronic bronchitis, and examined all those following this
occupation in two areas. Ue thus standardised for exposure to
occupational hazards and social class, and were able to identify the
-------
42
effect of two independent factors, namely cigarette smoking and level
of air pollution (Holland and Reid [3]).
A further difficulty in the choice of a population is that it
is necessary to examine individuals who have been exposed to the
environmental factor under investigation for a known period of time.
And here migration may interfere. If an area is prone to
contamination from a particular health hazard, i.t is likely that people
will move away - and it may be that the most susceptible will move.
Studies by Bradford Hill [4] among others have shown that migrants from
an area differ in susceptibility and resistance to disease from those
who remain. It is thus important to have some idea of the differential
rate of migration of individuals exposed or not exposed to a particular
hazard. In industrial studies this is relatively easy since the
records of a particular industry in terms of premature death and
retirement of workers can be obtained.
It is also possible, in looking at environmental hazards,
to use a population of children. They tend to remain in the same area
at least for a certain known period of time and, since they usually go
to school near their homes, it is relatively easy to determine the
effect of an environmental factor.
In comparisons of different groups of the population exposed
to environmental hazards, it is also important to choose appropriate
control groups. In considering an environmental hazard in an
industry, for example, it is possible to examine illness or morbidity
amongst husbands and wives, using the wives as controls. It is
unlikely that they will be exposed to the same occupational hazards as
their husbands whereas both are likely to be exposed to the same
environmental hazard.
Whatever population is chosen, it is important to investigate
that population as completely as possible. Information on those who
do not participate is essential - they may, for example, have refused
because they are ill or because they fear the consequence of having the
disease discovered.
Having defined the population and decided what variables to
measure, the next step is to collect the necessary information. One
method of doing this is by questionnaire. This is usually better than
a clinical history which may not identify and record in a repeatable
-------
way what the individual actually says. The importance of using a
standardised questionnaire, such as that developed by the EEC and the
MRC on chronic bronchitis, has been amply demonstrated. But whatever
questionnaire is used, it is vital that the sensitivity, specificity
and precision of the individual questions are validated (Cochrane and
Holland 5 ). And interviewing techniques must also be standardised.
Physiological tests are another method of obtaining the
necessary information. Such tests must be simple and if possible cheap
to administer, acceptable to those who receive them, and must produce
accurate and repeatable results. They must also be 'sensitive1
(give a positive finding in those who have the condition under
investigation)and 'specific (give a negative finding in those who do
not have the condition under investigation). Those who administrate
the tests must be trained, equipment standardised, and factors such as
the time of day and season of the year recorded as these may affect the
results considerably.
These then are the basic criteria for the prevalence or
retrospective investigations which are commonly used to estimate the
relative risk of a particular health hazard. Incidence or longitudinal
investigations, which are obviously much more expensive to carry out,
are appropriate when the absolute risk must be ascertained.
So far, 1 have discussed the standard epidemiological
techniques used in population studies throughout the world. Let us
now, for a moment, look towards the future and consider some of the
possible techniques.
The first such technique which, in my view, merits serious
consideration is tracer index case or rare event analysis. Various
situations in the recent past have emphasised the importance of
identifying a rare event - perhaps the best known example is the
thalidomide tragedy. We must equip oufselves to recognise harmful
new agents in the environment so that these can be removed or
iowbilised before too much damage is caused. It is difficult to
visualise how such an epidemiological method could be put into
effective practice. One possibility would be for various international
agencies to collect information on specific conditions and identify
suspicious clustering of cases. Examples of this type of monitoring
exist already in the World Health Organisation and in the Office of
-------
44
Population Censuses and Surveys in the area of congenital malformation.
But this is not sufficient and further possibilities to extend this
technique oust be explored.
The monitoring of events in society,is another tecnnique
which requires development. Sophisticated methods exist for the
surveillance of infectious diseases such as malaria and smallpox and
these have proved their worth. However, we have not yet gone far
enough in trying to monitor changes in chronic disease brought about
by alterations in the normal environment. The difficulty with this,
of course, is the time aspect - the diseases which are common
nowadays take many years to develop. For the development of such a
technique, therefore, it is necessary to begin to think of tracer
elements or changes in function or symptoms rather than the full-blown
developments of the disease itself. One example of this type of
monitoring is our study of the surveillance of growth in schoolchildren
to see whether entry into the EEC and changes in social welfare policy,
such as the provision of free school milk, have any effect on their
nutrition (6). Again 1 would suggest that much research needs to be
done in this area in future years.
Finally, may I make a plea for more attention to be paid to
analysis of disease and symptom complexes. We are now adding vast
numbers of chemicals to our water supply and yet we know little if
anything about the possible effects of health. Since it is unlikely
that these will necessarily produce changes in the pattern of known
diseases, we must be able to assess the occurrence of new diseases
and symptoms. Thus a recent printout of the analysis of the water
supply in London revealed the presence of considerable numbers of
chemicals, admittedly in very minute quantities. In future we may
be able to control our population with greater ease than at present
but perhaps they will all be sterile!
In this talk I have tried to give a brief outline of current
standard methods of studying the effects of environmental factors on
health and to suggest some possible areas for future development.
This is a difficult and complex field where issues are clouded by
economic and political considerations which cannot be ignored. But
if we are to cope successfully with this increasingly important area
of health, we must equip ourselves with more sensitive, sophisticated
and flexible methods of collecting the evidence.
-------
45
References
1 DOLL, R. , "Cancer of the lung and nose in nickel workers"
Brit. J. industr. Med.t 15, 217-223 (1958).
2 NEWHOUSE, H.L., "A study of the mortality of workers in an
asbestos factory", Brit. J. industr. Meg., 26, 294-301 (1969)
3 HOLLAND, \/.W. and REID, D.D., "The urban factor in chronic
bronchitis", Lancet, ^, 445-448 (1965)
4 HILL, A.B., "internal migration and its effects upon the
death-rates: with special reference to the County of Essex.
Medical Research Council Special Report Series, No. 95 (1925)
5 COCHRANE, A.L. and HOLLAND, W.U., "Validation of screening
procedures", Brit, ned. Bull. , 27_ (1), 3-8 (1971)
6 The National Study of Health and Growth - progress report.
pp 44-47. Annual Report of the Social Medicine and Health
Services Resea-ch Unit, St. Thomas's Hospital and Medical
School, 1973-74.
DISCUSSION
CARNOW (U.S.A.)
The rates of disease were less in the American cohort than
the urban or rural U.K. group. What U.S. group was studied and
how do you account for the difference?
HOLLAND (U.K.)
We studied male A.T.& T. van drivers aged 40 - 59 in three
U.S. cities. We examined 96 - 98% of all men classified as dri-
vers in the three cities. The reason for the difference was
probably that levels of air pollution were lower in the U.S.
cities than in the three English country towns.
-------
47
HOUSEHOLD SURVEY OF THE INCIDENCE OF RESPIRATORY
DISEASE IN RELATION TO ENVIRONMENTAL POLLUTANTS
RALPH I. MITCHELL/"*" RONALD WILLIAMS/ ROGER W, COTE*
RICHARD R, LANESE"*"1" AND MARTIN D, KELLER++
+ Battelle, Columbus Laboratories, Columbus, Ohio, USA
++ The Ohio State University, College of Medicine, Columbus,
Ohio, USA
ABSTRACT
A study was undertaken to determine the incidence of res-
piratory disease in households in a midwestern suburban middle-
class community, in a relatively unpolluted area. The sample
included 441 families, divided into two groups; those utiliz-
ing gas cookery and those utilizing electric cookery. Family
health and demographic data were obtained from the participants.
The period of the study was one year. Reports of acute res-
piratory illness were obtained through bi-weekly telephone
calls to each of the households. The respondents were asked
to report respiratory illness in any member of the household,
and to indicate the presence or absence of a set of signs and
symptoms. Ambient air was analyzed, indoors and outdoors, in
a sample of the households, and pulmonary function tests were
conducted on a 42 percent sample of the participants, represen-
ting both types of household.
The results of the study are presented with regard to re-
ported indicence of respiratory illness in the two types of
household. AID analysis and multiple regression were carried
out to determine the best set of independent variables as pre-
dictors of respiratory disease incidence. This is the first
-------
48
in a series of studies, Subsequent studies will follow-up the
reported respiratory illness to ascertain objective signs of
illness and obtain bacterial cultures, Similar studies are
also contemplated in areas in which the ambient air is more
heavily polluted,
-------
49
1. Introduction
In the combustion of fossil fuels, many substances are produced that
are toxic at certain concentrations. A substance of particular interest
is N02. Although it is known that NC>2 may be fatal to persons exposed to
high concentrations, such as 500 ppm that may be encountered in silos,
very little is known about long-term exposure to low levels of concentra-
tion. Recent studies in Chattanooga [l,2,5| indicated that persons living
in areas with elevated levels of ambient N(>2, reported a high incidence
of respiratory illness. This caused some concern regarding the possible
effects of N02 on the incidence of respiratory illness. It is known that
the open flame in gas ranges can produce short-term elevation of NO
2
concentration in the cooking area. However, the possible effects on the
respiratory health of members of such households have not been assessed
adequately. Accordingly, the present study was initiated to acquire data
for making such assessments.
In the Chattanooga area, it was found that in addition to the NO
there were other pollutants such as nitric acid, suspended nitrates,
and particulates at unusually high levels of concentration. In order to
eliminate such factors, it was felt to be of value to study areas with
low ambient air pollution levels.
2. Design
The area chosen for the present study is the City of Upper Arlington,
a middle class suburb on the periphery of Columbus, Ohio. In June, 1972,
questionnaires were distributed through the elementary schools in
Upper Arlington. The questionnaires were developed by the Environmental
Protection Agency for the Community Health and Environmental Surveillance
System (CHESS) [_4j and included questions regarding demographic variables,
smoking habits, and a variety of health problems. The completed question-
naires were separated into two groups; households with gas ranges; and
households with electric ranges. Families that expected to move within
12 months were excluded from the sample. Among over 1,000 completed
questionnaires, 17.5 percent of the households had gas ranges. In order
to augment the number of gas-cooking households, additional households
were visited and a second appeal was made through the schools. The final
sample included 441 households, 232 of which used electricity for cooking,
and 209 of which used gas. The final sample included only the families
-------
50
expressing willingness to participate in the program. Table I presents
the population sample whose respiratory disease experience is reported
in the present paper.
TABLE I. POPULATION SAMPLE PARTICIPATING IN STUDY
Upper Arlington Study Population
Electric Gas
Households 232 209
Mothers 230 207
Fathers 222 197
Children, 0-12 years 407 356
Table II presents the overall age and sex distributions of the
entire study sample.
TABLE II. AGE AND SEX DISTRIBUTION OF STUDY SAMPLE
Age
0-5
6-10
11-15
16-20
21-40
41-60
61-Up
Upper Arlington Study Population
Sex
Male
69
176
213
77
215
212
6
Total 968
Female
59
173
208
90
289
158
7
984
The program was initiated in November, 1972. Participants were
contacted by telephone every two weeks to obtain reports on respiratory
illness among members of the household during that period. A special
calendar was supplied to each household for record keeping. The informa-
tion obtained was recorded on two sets of records. A master card was
maintained for each family in case the working mark-sense card was lost
or unreadable. The master card contained all the information obtained
-------
51
by the telephone interviewer. The survey continued for 26 periods to
yield one full year of data.
The study population was invited to have pulmonary function
screening tests for the detection of chronic respiratory impairment.
Also, in order to determine the Levels of NO and NO, a sample of the
gas and electric households were monitored for these substances.
2.1 Pulmonary Function TeajEinp
Measures of Forced Vital Capacity and 0.75 Second Forced Expiratory
Volume were made using a 10-liter dry rolling-seal spirometer. Prior to
testing, each subject was instructed regarding the proper method of
performing the forced vital capacity maneuver. Care was taken to insure
that each subject inspired to maximum capacity and kept the back reason-
ably straight during forced expiration. The test was repeated a number
of tines until the results obtained were reproducible (± 57,). A minimum
of three trials were obtained with each subject, and the best value for
each of the variables was recorded.
2.2 Monitoring for NO
^_^_^^^MMV^^^M^^^^.^B^^_^BI2£
Since the study was concerned with the possible relationship between
the incidence of respiratory disease and exposure to oxides of nitrogen,
it was necessary to measure this substance in a sample of the households.
A total of 83 homes with gas cooking modes and 50 homes with electric
cooking modes were monitored for NO and NO over periods of 24 hours using
the modified Jacobs-Hochheiser (J-H)method [5] . In addition, 53 outdoor
samples were taken with the J-H units in the areas of the indoor monitor-
ing. Continuous chemiluminescence measurements were also made for three-
day periods in 46 homes to evaluate the instantaneous variation in indoor
NO. and NO levels.
Peak NO. levels in gas cooking mode households, during cooking
periods, were generally eight times higher than the 24-hour average.
In several households, these peak values exceeded 100 pphm. No peak
values of this type were noticed in electric cooking mode households,
other than a rise in NO levels during periods of high traffic, in the
homes near thoroughfares. Measurements were made in the period of
January through April, 1973. Table III presents a summary of the NO2 and
HO data.
-------
52
TABLE III. SUMMARY OF THE N02 AND NO DATA
NO Monitoring
X
Type
Gas homes
Electric homes
Outdoors
Range ,
N02
0.5-11
0-6
1.5-5
pphmn
NO
1-41
0-34
0.5-22
Average .
N02
5
2
3
pphm
NO
11
7
4
2.3 Statistical Method
The present study seeks to examine the relationship of a large number
of variables to the incidence of respiratory illness jj\ the study popula-
tion. In this type of study, the combined effects and interactions of
these variables are as important as the individual or main effects of each
independent variable. Traditional analytical methods are typically bound
by rather stringent assumptions concerning distributions and linearity,
and frequently neglect interaction. In nearly all studies like the
present one, these assumptions are known to be unwarranted, or they are
incapable of being validated by examination of the data. The Automatic
Interaction Detector method (AID) of Sonquist and Morgan [6] , utilized
in the present study, provides a multivariate approach to analysis
relatively free of these restrictions. The AID program is based on a
sequential analysis of variance. It selects the best set of predictive
or explanatory variables with respect to the dependent variable, in this
case, reported incidence of respiratory disease. It also simultaneously
seeks out relationships and interactions among these variables and
indicates the significance of the relationship of independent to dependent
variables, in terms of the portion of the variance (differential occur-
rence of respiratory illness) that is explained.
AID was used for analysis of the incidence of respiratory disease in
the samples of mothers and of fathers. Multiple regression was used for
the analysis of the incidence in children, as will be explained in the
section on Results. Table IV presents the predictive or explanatory
variables used in the AID and multiple regression methods.
-------
TABLE IV. VARIABLES USED IN THE AID AND MULTIPLE REGRESSION METHODS
Mothers and Fathers - Columbus Children - Columbus Mothers - Long Island
Explanatory Variables
Age (year of birth)
Group (Gas/Electric)
Cough - 3 months a year
Cough - day or night
Cough - day or night - 3 months a year
Phlegm in morning
Phlegm 3 months a year
Phlegm day or night
Cough or phlegm 3 weeks
Shortness of breath
Current cigarette smoker
Smoked 5 packs or more in life
Number of cigarettes per day
Age started smoking
Age quit smoking
Number of years smoked
Number of years since quit smoking
Exposed to irritant on job
Dependent Variables
1. All respiratory illness
(mothers and fathers)
2. Lower respiratory illness
(mothers)
Explanatory Variables
Age (year of birth)
Group (Gas/Electric)
Sex
Size of house (number of rooms)
Numbers of years family lived
in community
Father's educational level
Chronic heart disease
Chronic lung disease
Chronic asthma
Pneumonia
Croup
Bronchitis
Other chest infections
Hospitalized
Dependent Variable
All respiratory illness
Explanatory Variables
Group (Gas/Electric)
Cough or phlegm - 3 months
a year
Current cigarette smoker
Cough in morning
Cough day or night
Cough day or night
3 months a year
Phlegm in morning
Phlegm day or night
Phlegm 3 months a year
Shortness of breath
Ever smoked
Number of cigarettes per day
Exposed to irritants on job
Duration of exposure to
irritants
Kind of irritants
Where lived as a child
(up to age 20)
Where lived from age 20-30
Where lived after age 30
Family size
Dependent Variables
1. All respiratory illness
2. Lower respiratory
illness
-------
54
3. Results
Figure 1 presents a comparison of the annual incidence of all reported
respiratory illness, per hundred persons, in the subject categories indi-
cated. In all cases, the rates were higher in households(utilizing elec-
tricity for cooking. In general, rates were highest in children under 12,
and higher in mothers than in fathers. An additional comparison is
presented for reported lower respiratory disease (chest colds) among the
mothers. This will be related in the Discussion section to data from
another study.
4OO
w 30O
J
1
£ 200
I
100
E - Electric cookery
G - Gas cookery
ARD -All respiratory
disease
LRD - Lower respiratory
disease
243
E
222
G
172
E
159
G
389
E
377
G
159
143
E G
Mothers Pothers Children Mothers
ARD ARD under 12 LRD
ARD
Fig. 1. Comparison of annual incidence of respiratory illness.
Figure 2 presents the AID analysis for all reported respiratory ill-
nesses among the mothers in the Arlington sample. The overall rate for
the entire population of mothers was 234 illnesses per hundred mothers
per year; mothers cooking with electricity reporting an incidence of 243;
while those cooking with gas reported 222. These are the figures
presented in the bar graph in Figure 1. The AID analysis further indi-
cated a higher incidence among younger mothers, who indicated in the
household questionnaire that they usually bring up phlegm. Younger
mothers, in general, reported higher incidence than older mothers, even
when they did not initially report bringing up phlegm. Beyond this, none
of the explanatory variables contributed significantly to explaining the
differences in incidence. In fact, the AID analysis indicated that
-------
55
<34 years
"Yes"
n = 5
I =572
: 28 years
Bring Up Phlegm
Day or Night in Winter
"No"
n = 6
I = 494
I = Incidence of re-
ported acute respira-
tory disease per 100
persons per year
Fig. 2. AID analysis for respiratory illnesses among
mothers - Arlington sample.
virtually none of the variance was explained whether or not the mothers
cooked with gas or electricity. Figure 2 is the only AID diagram
included in the present report to illustrate the method. None of the AID
results proved effective in explaining differences in incidence.
AID analysis was also carried out on reports of lower respiratory
illness among the mothers. The overall population of mothers had an
incidence of 151 per hundred, per year. As shown in the bar graph of
Figure 1, the mothers cooking with electricity had an incidence of 159 as
compared with an incidence of 143 among mothers cooking with gas. However
this difference was of such low significance that it could explain only
1/50 of 1 percent of the variance. In fact, in this case, all of the
explanatory variables together could predict only one percent of the
variance, leaving 99 percent unexplained.
AID analysis was carried out for all reported respiratory illness
among the fathers. The overall incidence in the population of fathers
was 166 per hundred, per year. The fathers in households with electric
-------
56
cookery reported an incidence of 172, as compared with 159, in the gas
household sample. However, once again the differences between the gas
and the electric samples explained only 1/100 of one percent of the
variance, and all of the explanatory variables together explained less
than one percent. Trends were noted indicating that younger fathers
reported more respiratory illness. The only other variable that may be
related is smoking; those who smoked more for more years, or quit more
recently, reported more illness than nonsmokers or those who quit long
ago. However, it must be emphasized that none of these variables signi-
ficantly explain the incidence of respiratory disease. The AID analysis
does not even select the variable of gas or electric cookery for inclu-
sion, since this has so little relation to differences in the occurrence
of illness in the two groups.
The incidence of illness among children from birth to 12 years was
analyzed by multiple regression. Since, in earlier analyses, all of the
explanatory variables, and, in particular, the variable of gas or elec-
tric cookery contributed little to explaining incidence of illness, the
expense of AID analysis was not warranted for this group, and multiple
regression was deemed sufficient to answer the questions posed. The
overall incidence for this population of children was 384 per hundred
per year. The rate was higher among children than among the mothers or
fathers. Children in the electric cooking homes had an incidence of 389
as compared with 377 in the gas cooking sample. However, once again the
method indicated no predictive power.
3.1 Lung^unction Tests
The entire population of the sample was invited to have lung function
examinations. The tests were conducted in the manner indicated in the
section on Design. Table V presents the age, sex, and group (gas or elec-
tricity) of the sample tested. Overall, approximately 42 percent of the
2
study population was tested. Chi square (X ) analysis indicated no signi-
ficant differences between the gas and electric household participants,
in terms of age and sex representation. Table VIpresents the analysis of
variance of the results of Forced Vital Capacity and the 0.75 Second Forced
Expiratory Volume. There were no significant differences between the
groups with respect to the Total Forced Vital Capacity, and a small but
significant difference appeared with regard to the 0.75 Second Forced
-------
57
Volume. The participants from the gas cooking households had slightly
better lung function measurements, on the average. However, the differ-
ences explained only an extremely small portion of the variance in
pulmonary function, approximately one half of one percent. These tests
were carried out in order to detect the possible presence of a large
number of individuals with pulmonary impairment in one or the other group,
since this might bias the results with regard to the reported incidence
of respiratory illness. These data do not indicate such bias.
3.2 Discussion
The overall outcome of the analyses employed in the present study
indicates no significant difference in reported respiratory illness
between the members of households cooking with gas and those cooking with
electricity. This was evident in the comparison of incidence rates among
mothers, fathers, and children from birth to 12 years of age.
Additional analyses were carried out with regard to the reported
lower respiratory disease among mothers in order to allow comparison of
the present study results with those obtained in an epidemiologic study
conducted by the Environmental Protection Agency in a suburban community
in Long Island, N. Y. The latter study involved 146 households, 59 of
which cooked with electricity, and 87 with gas. The study was similar to
the present one in Upper Arlington, and was conducted over a period of 30
weeks. Inquiries were made regarding the incidence of respiratory illness
every two weeks. Through the courtesy of the Environmental Protection
Agency, computer tapes containing the data of the Long Island Study were
made available to the authors so that they could be subjected to AID
analysis in the same way as the Upper Arlington data. Table Vllpresents a
comparison of the two studies with regard to all reported respiratory
illness, and lower respiratory illness (chest colds) only. The data of the
Long Island Study were extrapolated to a period of one year so that Inci-
dence rates could be directly compared with those of the Upper Arlington
study. It is realized that this may introduce some error, since seasonal
variation may be associated with changes in respiratory disease incidence.
However, the basic hypothesis being tested involves comparison of incidence
rates among the gas cooking and the electric cooking households in each
study area. In Long Island, as in "pper Arlington, the incidence of all
respiratory disease was higher among mothers cooking with electricity.
-------
TABLE V. SUMMARY OF PARTICIPANTS IN LUNG FUNCTION TESTS
Age
Group
0-5
6-10
11-15
16-20
21-40
41-60
61-Up
Totals
0-5
6-10
11-15
16-20
21-40
41-60
61-Up
Totals
Total
31
96
112
42
104
120
2
507
30
83
111
49
146
85
&
508
Electric
Number
Tested
5
54
44
10
48
40
2
203
4
51
44
11
65
38
I
214
Families
Tested, 7.
16
56
40
24
46
33
100
40
13
61
40
22
46
45
_25
42
Gas Families
Number
Total Tested
Male
38
80
101
35
111
92
4
461
Female
29
90
97
41
143
73
3
476
Participants,
2
59
42
8
45
37
2
195
Participants
6
42
43
13
75
30
1
210
Tested, 7,
5
74
42
23
41
40
_50
42
21
47
44
32
52
41
33
44
Total
69
176
213
77
215
212
&
968
59
173
208
90
289
158
7
984
Number
Tested
7
113
86
18
93
77
_4
398
10
93
87
24
140
68
2
424
Tested, 7,
10
64
40
23
43
36
66
41
17
54
42
27
48
43
_29
43
Ln
00
-------
59
TABLE VI. ANALYSIS OF VARIANCE IN PULMONARY
FUNCTION TESTS
Characteristics
Number of participants
Mean age
Mean height
Mean weight
Mean FVC
Mean FEV-75
F test for FVC (1
F test for FEV-75
Group 1
Electric
416
24.2
61.5
118
3.02
2.32
,816 df) = 0
(1,816 df) - 4
Group 2
Gas
405
23.8
61.6
m 116 m
3 O4
^i^ ^i^
\ s i *3 n ^ '
2.39
.647, ,
.455UJ
(1) Adjusted values on basis of age, sex, and height by
co-variate analysis.
(2) Significant at Alpha 0.05.
In the Long Island study, the overall incidence was 302 respiratory ill-
nesses per hundred mothers per year; those with electric cookery reporting
322, as compared with 289 in the gas cooking households. However, as in
the Upper Arlington sample, these differences were insignificant. All of
the explanatory variables together were only able to account for 4 percent
of the variance in the Long Island Study. While this was somewhat higher
than Upper Arlington, it still left 96 percent of the variance in illness
unexpla ined.
TABLE VII. RESPIRATORY DISEASE REPORTED BY
SAMPLE OF MOTHERS
Rates of Illness
All Respiratory Lower Respiratory
N Disease Disease
Long Island Electric 59 322 99
Gas 87 289 133
Total 146 302 120
Upper Arlington Electric 230 243 159
Gas 207 222 143
Total 437 234 151
(1) Reported illnesses per 100 persons per year.
(2) "Chest colds".
-------
60
The Long Island sample mothers, in general, presented less lower
respiratory disease (chest colds). The population incidence was 120 per
hundred per year, with the mothers in the electric group reporting 99,
and the mothers in the gas group reporting 133. However, again the differ-
ences were not significant, with only a minute fraction of the occurrence
of illness explained. All of the predictor variables in the Long Island
Study together explained only 5 percent of the incidence.
The data presented do not demonstrate sufficient differences to reject
the null hypothesis, namely, no difference between the incidence of respi-
ratory illnesses among mothers in the electric and gas cooking households.
In both the Upper Arlington and the Long Island studies, insofar as the
latter has been analyzed, it would appear that populations living in
relatively unpolluted areas do not differ significantly with respect to
reported respiratory illness in relation to mode of cookery. The under-
lying assumption of the present study was that exposure to NO. generated
by gas cookery might have some effect upon the incidence of respiratory
disease. If this were so, it would be reasonable to conclude that either
the concentration or the length of the exposure were not sufficient to
cause differences in the type of acute illness reported1. It may be
possible that certain other noxious agents in the general environment,
as mentioned by Warner and Stevens [?] , might be required to bring out
the incidence described in other studies.
There may be an inherent weakness in all of the studies that have
been considered due to the inability of the household respondent to convey
accurate information regarding the incidence of respiratory infection in
all members of the household. There may be sufficient differences between
respondents to introduce error that may seriously affect the outcome of
studies with relatively small samples. Certainly, it would be difficult
to feel secure about the differentiation of upper and lower respiratory
disease by the respondents. Such distinctions are difficult to make, and
may vary greatly with the experience and perceptiveness of the respondent.
Further analysis of the specific symptoms that were reported in the
present study will be presented in another paper. A second study is now
in progress in which a nurse-epidemiologist will examine individuals
reporting respiratory illnesses and take nasopharyngeal cultures. At the
same time, examination will be made of "well" members of the same house-
-------
61
hold and of members of control households in the study population. By
this means, it will be possible to obtain more rigorous information
regarding the meaning of a reported illness, in terms of objective
physical findings and bacterial culture.
References
1 SHY, C.M. , CREASON, J. P., PEARLMAN, M. E. , McCLAIN, K. E. , BENSON,
F. B., and YOUNG, M. M. , "The Chattanooga School Children Study:
Effects of Community Exposure to Nitrogen Dioxide. I. Methods,
Description of Pollutant Exposure, and Results of Ventilatory Function
Testing", J. Air Pollution Control Assoc.. 20(8): 539-545 (1970).
2 SHY, C. M., CREASON, J. P., PEARLMAN, M.E., McCLAIN, K. E., BENSON,
F. B., and YOUNG, M. M., "The Chattanooga School Children Study:
Effects of Community Exposure to Nitrogen Dioxide. II. Incidence
of Acute Respiratory Illness", J. Air Pollution Control Assoc., 20(9):
582-588 (1970).
3 SHY, C. M., HASSELBLAD, V. , BURTON, R. M., NELSON, C. J. and COHEN,
A. A., "Air Pollution Effects on Ventilatory Function of U. S. School
Children", Arch. Environ. Health, 27:124-128 (1973).
4 School and Family Health Questionnaire Developed by the Division of
Health Effects Research, Environmental Protection Agency, OMB #158-R
0019.
5 MERRYMAN, E, L., "Evaluation of the Arsenite-Modified Jacobs-
Hochheiser Procedure", Environ. Sci. Tech., 7:1056 (1973).
6 SONQUIST, J. A. and MORGAN, J. N., "The Detection of Interaction
Effects." Monograph # 35. Survey Research Center, Institute for
Social Research, Univ. of Michigan (1964).
7 WARNER, P. 0. and STEVENS, L., "Revaluation of the 'Chattanooga
School Children Study' in the Light of Other Contemporary Govern-
mental Studies: The Possible Impact of These Findings on the
Present NO Air Quality Standards", J. Air Pollution Control Assoc.
23(9) 769-772 (1973).
-------
62
DISCUSSION
KJELLSTROM (Sweden)
I understand that you only found a cause-effect relation-
ship between smoking and respiratory disease. Why do you then
spend so much time on showing prevalences among electric cooking
homes and gas cooking homes? Why did you make the study and who
funded it?
MITCHELL (U.S.A.)
The objective of our study was to determine if cooking with
gas would increase the incidence of respiratory disease as com-
pared to an identical population which cooked with electricity.
The study was funded by the American Gas Association with coop-
eration from the United States Environmental Protection Agency.
GOLDSMITH (U.S.A.)
You informed us that you distributed about five thousand
questionnaires and received about a thousand replies. Does
this not introduce a great opportunity for bias? The bias can
be aggravated by how the purpose of the study was presented to
the possible respondents. Even if the purpose were disguised,
how can one expect to generalize from a sample which is so poorly
defined?
MITCHELL (U.S.A.)
We distributed 5,000 questionnaires to the entire element-
ary school population. There were several children from the
same family in different grades. We received replies from a
thousand families, which was more than one-half returns; this
represents a fairly good sample.
-------
63
RELATIONSHIP OF AIR POLLUTION TO PREVALENCE OF LOWER
RESPIRATORY ILLNESS AND LUNG FUNCTION IN
AUSTRALIAN SCHOOL CHILDREN
++ + +
S, R, LEEDER, A, J, WOOLCOCK, J, K, PEAT/ C, R, B, BLACKBURN
•••Department of Medicine, University of Sydney, Australia
•H-Medical School, St. Thomas's Hospital/ London
ABSTRACT
The association of different levels of air pollution with
prevalence of past loner respiratory illness and lung function
(assessed by the maximum expiratory flow-volume curve) was stu-
died in two populations of Australian school children - 12,000
in Sydney and nearly 300 in a steel-industry coastal city 60
miles south of Sydney (Port Kembla).
Levels of sulphur dioxide and smoke pollutants were low by
international standards in Sydney, but high peak concentrations
of sulphur dioxide were recorded in Port Kembla.
Children attending school in the more polluted areas of
Sydney and in Port Kembla had the same prevalence of past, trea-
ted lower respiratory illness as children in the less polluted
regions, A small, statistically significant correlation between
lower forced vital capacity and higher pollution areas was
found in older Sydney boys. This correlation was independent
of the effects of other factors such as height, social class
and smoking habits.
Ho significant decrease in maximum expiratory flow rates
was found in children in the more polluted areas when the effects
of their smaller vital capacities were allowed for. Interpre-
tation of reduced expiratory flow rates attributed to pollution
is difficult unless change* in vital capacity arc takvn into
account.
-------
64
1. INTRODUCTION
Chronic obstructive lung disease (chronic bronchitis and
emphysema) may begin in the small peripheral airways of the lung
many years before symptoms occur or conventional tests of lung
function reveal abnormalities (Macklem 1972). There is epidemio-
logical evidence that factors operating during childhood may be
associated with the consequent development of chronic obstructive
lung disease in adult life (Holland et al 1969, Reid 1969), and
these factors may affect the small airways.
Extensive obstruction of the small airways has been found at
post mortem in patients with only mild symptoms of chronic obstruc-
tive lung disease dying of unrelated causes (Hogg et al 1968).
Such obstruction may remain asymptomatic for years, undetectable
with conventional tests of lung function because of the low
resistance of these airways to airflow, afforded by their shortness
and large cross sectional area (Weibel I960, Macklem and Mead, 1968).
Impaired lung function detected with special tests, consistent
with damage to the small airways, has been found in asymptomatic
smokers who have nortnal spirometry (Ingram and O'Cain 1971);
following 'upper' respiratory infection (Picken et al 1972); and
after exposure to high levels of air pollutants including ozone
especially when the subject exercises as v/ell (Bates et al 1972).
While the antecedent relationship of disease in the small airways
to chronic obstructive lung disease remains to be established,
assessment of the effects of such factors as air pollution on the
function of the peripheral airways in populations is justified
(Bates 1972). Such assessments are not easy in epidemiological
-------
65
studies because of the lack of a totally satisfactory field test
for this purpose (Macklem 1972).
The present study of Australian school children commenced in
1971 and is designed to assess the effect of various factors on
the development of normal lung function between the ages of 7 and
17 years. Factors including acute respiratory illness, poor
social environment, cigarette smoking and air pollution are beinp
studied. Children are examined annually. Ventilatory function
is measured from the maximum expiratory flow volume curve.
This paper outlines the methods used in the study of the
relationship of low concentrations of air pollution with sulphur
dioxide and smoke with the ventilatory function and prevalence of
respiratory illness of Australian children, and sorae early results.
2. METHODS
2.1 Sampling
Two samples of children were drawn, the larger one from
Sydney, and the smaller from Port Kembla. The base population
from which the larger sample was drawn was of children attending
all school within 20 miles of central Sydney, in third grade of
primary school (7-8 years) or first form of secondary school
(12-13 years) in 1971. The sample contained approximately 1 in 6
third grade pupils and 1 in 7 first formers of the base population.
Almost 9396 of the sample (12,000) were studied, the remaining
children either being absent or lacking parental consent. Slightly
more children not examined were of parents of low social class
and of non-Australian birthplace than were children in the study.
-------
66
To observe the effects of occasional high levels of pollution
(in comparison with the effects of relatively constant low levels
in Sydney), 283 children in third to sixth grade (aged 7-13 years)
were studied in Port Kembla. Port Kembla is a coastal steel-
works town 60 miles south of Sydney. Ninety percent of the school
population were examined with a similar non-participant bias as in
Sydney.
2.2 Measurement of Lunp; Function
The maximum expiratory flow volume (MEFV) curve was derived
by computer analysis of the largest of two or three forced vital
capacity manoeuvres which were recorded for each child on magnetic
tape using portable equipment taken to the school (Leeder et al
1974). The MEFV curve was chosen because it provided maximum
expiratory flow rates during the latter half of forced maximal
expiration as well as conventional spiroraetric values. These
later flow rates are considered less effort-dependent and more
reflective of resistance within the smaller airways than are
earlier (e.g. peak) flov/ rates (Macklem and Mead 1968).
From the MEFV curve were obtained values for peak expiratory
flow rate (P3FR), forced vital capacity (FVC), timed expiratory
volumes (FWQ>5 se(.( FEV0.75 gec) and the maximum expiratory flow
rates (V ) at 50. 25 and 10°S FVC (Figure 1). Measurements were
max
converted to body temperature equivalents.
2.3 Collection of Demographic. Social and Medical Data
Demographic details were obtained by written questionnaire
from the children's parents as were parental occupations and
-------
67
PEFR
Maximum
Expiratory
Flow
Rate
TOTAL
LUNG
CAPACITY
(TLC)
Volume
RESIDUAL
VOLUME
(RV)
Figure 1 Schematic maximum expiratory flow volume curve, from
which may be obtained forced vital capacity (TLC-RV)
and peak expiratory flow rate (PEFR) as well as
maximum expiratory flow rates at any point during the
FVC manoeuvre.
The interval A-RV commencing at 50% FVC is less effort
dependent and more sensitive to minimal airways damage
than the flow rates along the curve from TLC to A.
-------
68
countries of birth. Details of past respiratory illness were
sought using the saaie questionnaire, and information about the
smoking habits of children was obtained by confidential interview
(Leeder and Woolcock 1973).
2.4 Air Pollution Measurements
Annual average concentrations of sulphur dioxide (S0_) and
smoke density were obtained for selected regions of Sydney and
Port Kembla (as well as peak hourly concentrations for the latter)
from the statutory pollution control body in New South Wales
(Pollution Report 1972). Schools were then ranked by our own
estimates, based upon this information, according to their
probable pollution exposure.
2.5 Analysis
Analyses of variance in the measurements of lung function of
1942 boys aged 12-13 years in 1971 (non-smokers) were computed
using a regression technique. To exclude the confounding effects
of social class, height, weight and age, these several factors were
included as covariates in all analyses. Similarly, to remove the
effects of forced vital capacity upon measurements of expiratory
flow rate (as they were strongly correlated), it was included as an
additional covariate in these analyses.
The regressions v/ith height and mean values of lung function
measurements of children living in Port Kembla were compared with
those of children living in Sydney.
The frequency of a history of past respiratory illness was
compared for boys and girls of all ages living in the more and
-------
69
less polluted areas of Sydney, and for children living in Port
Kembla with those of Sydney children, and examined for significant
differences using the X test.
3. RESULTS
3.1 Air Pollution Levels
The annual average concentrations of sulphur dioxide varied
from approximately 120 to 1*0 micrograms/cubic metre (m } in the
most to least polluted areas of Sydney in 1971. Concentrations
of this pollutant has been declining slowly for several years in
Sydney (Pollution Report 1972) but the geographical distribution
has not changed substantially.
The annual average SO- concentration recorded in Port Kembla
in 1972 was 95 micrograms/ar with the highest 2k hour average of
over 1000 micrograms/m and the highest hourly reading in excess of
3000 micrograms/m .
Smoke density varied in Sydney from 22 to 60 raicrograms/nr
and was distributed in a similar geographical pattern to the SO^.
Dust deposit in Port Kembla ranged from 10 to 30 grams/m / month
and smoke, measured as coefficient of haze (CCH), had an annual
average of 1.2 COH units/1000 linear feet.
3»2 Prevalence of Respiratory Illness
The prevalence of a history of past, treated respiratory
illness in boys and girls in Sydney is given in Table 1. Althougl
the prevalence rates varied considerably from school to school,
this was not dependent on pollution rating. Similar prevalence
-------
70
rates were obtained in Port Kembla to those in Sydney children of
the same age.
TABLE I
Percentage prevalence of past treated bronchitis* asthma,
and other chest or lung problems in a sample of Sydney
(Australia) school children in 1971.
7-8 years old
n - 1,246
Disease Boys
Asthma alone 2.2
Bronchitis alone 24.0
Asthma with
bronchitis 7.2
Other chest or
lung problems
alone or with 6.1
asthma and bronchitis
Girls
1.0
22.0
4.4
5.5
12-13 years old
n = 2,517
Boys
2.1
20.5
5.9
7.2
Girls
1.2
16.2
4.3
6.2
3.3 Lung Function Measurements
In Sydney boys aged 12-13 years in 1971, significant variance
(F = 15.9) in forced vital capacity was attributable to the pollu-
tion factor with a significantly lower mean FVC {tf.8%, p < .001)
-------
71
among those boys attending schools in more polluted areas (Table 2).
Results are not yet available for older girls and the younger
children in our study.
TABLE II
Analysis of variance in lung function measurements
attributable to an air pollution factor in 1924
boys aged 12-23 years in Sydney in 1971.
MEASUREMENT MEAN SUM OF
Source (Pollution)
Peak expiratory
_— fc * & / J
flow rate
j-FVC 9.190
1j"FEV 0.243
U.V 50% FVC 4.399
ff max
SQUARES F
Error
0.929 2.45
0.577 15.93***
0.098 2.48
0.723 6.08**
*** p<.001
** p<.01
I Height,weight,age and social class as covariates
•ff As for -J-, with FVC as additional covariate.
A small a*-iount of variance in V 50?'. FVC was still attributable
max
to the pollution rating even after the inclusion of FVC as a
covariate in the analysis but the reduction of mean value v.ras not
significant (2.1%, p> .05). Vfhile other measurements of lung
function were negatively correlated to a small degree v;ith pollution
rating, the mean differences were all small and r-ntiroly accounted
for by the reduction in mean vital capacity.
Regression coefficients and nenn values for prediction of
forced vital capacity and Vmax 50^ FVC according to heifht in the
-------
72
Port Kembla children did not differ significantly from those found
in Sydney children.
4- DISCUSSION
Sydney enjoys clean air by international standards. A
spreading suburbia affords the opportunity to study the effects of
relatively sraall gradients of pollution without the interference
of marked social stratification or urban rural gradient. Photo-
chemical pollution has increased in recent years and aore informa-
tion on its magnitude is now becoming available and will be
considered in future studies.
In marked contrast to past treated lower respiratory illness,
particularly asthma, the pollution factor in our study was not
related to any large reductions in V , particularly during the
ZTlGl A.
latter half of forced expiration. That no variation in the
prevalence of respiratory illness could be attributed to pollution
is consistent with this observation.
Several explanations may be offered for our findings.
Firstly, the concentrations of pollution we have been studying may
be entirely safe, and cause no trouble to the lungs of children.
The small reduction in FVC would then have to be attributed to some
other factor not included in our analyses of covariance, or some
non-linear effect of one of the factors already included in the
analysis.
Secondly, the measurements of air pollution, lung function
and respiratory illness may all be too insensitive to detect subtle
changes. However the geographical stability of the pollution
patterns, and the known association in our study of impaired lung
-------
73
function with positive answers to our questions about past res-
piratory illness make these explanations unlikely. It may how-
ever be profitable to investigate the association of certain
respiratory symptoms both with specific impairments lung function
and pollution gradients in the future in determining the dose-
response relationship between impaired respiratory health and
pollution levels.
Thirdly, the lower FVC observed in older boys in the more
polluted areas of Sydney may simply reflect a variation in their
/
lung size, perhaps due to altered growth characteristics of the
lung and airways when exposed to low levels of pollution over a
long time. Lung volume measurements would be required to determine
if this were the case.
Changes in lung function attributable to low levels of
pollution may be different from those due to high concentrations.
Other studies have revealed a greater frequency of respiratory ill-
ness and reduced peak expiratory flow rates in children exposed to
high concentrations of pollution (Holland et al 1969, Toyama 196^)
sometimes in the presence of a normal FVC. Low concentrations of
pollution may alter the growth pattern of the lung, whereas higher
levels alter the function of the large airways and hence reduce the
PEFR (Bates 1972). Zapletal et al (1973) found some evidence of
impaired airway function, using the MEFV curve, in 6 of 111
children, 10 to 11 years old who lived in Most, Czechoslovakia,
but the pollution levels there were substantially higher than in
our study.
The Port Kembla study was of a small number of children,
and small differences in lung function may not be apparent for this
-------
reason. However, it was clear that no marked long-term effects
on lung function could be attributed to occasional very high con-
centration peaks.
5. CONCLUSIONS
Thus far, the study of the association of air pollution by
low levels of sulphur dioxide and sicoke with the prevalence of
lower respiratory illness and lung function in Sydney children has
not revealed any long tera effects apart from a marginally lower
forced vital capacity in older boys. The study points up the
need to allow for changes in lung growth characteristics in children
living in different areas before concluding that isolated measure-
ments of reduced expiratory flow raten represent lung or airway
damage.
Acknowledgements:
He thank Dr. A. Eell of the division of Health and Pollution
Control of the Health Conaiission of NS7/ for providing measurements
of pollution, and the staff and children of the survey schools for
their participation. The study was supported by the Australian
Tobacco Research Foundation and S.R. Leeder was in receipt of an
Australian National Health and Medical Research Council Scholarship.
-------
75
BIBLIOGRAPHY
BATES, D.V., "Air pollutants and the human lung", Amer. Rev. Resjx Disease.
105, 1 (1972).
BATES, D.V., BELL, G., BURNHAM, C., HAZUCHA, M., MATHA, T. , FENGELLY, L.D. ,
and SILVERMAN, F, , "Short-term effects of ozone on the lung" J. AppjL Physio 1
32(2); 176 (1972).
HOGCjJ.C., MACKLEM, P.T., and THURLBECK, W.M., "Site and nature of airway
obstruction in chronic obstructive lung disease," New Eng J. Med. 278;1355 (l96b).
HOLLAND, W.W., HALIL, T. , BENNETT, A.E. , and ELLIOTT, A., "Factors affecting
the onset of chronic respiratory disease," Brit. J. Med. 2; 205 (19^9).
INGRAM, R.H., Jnr. and O'CAIN, C.F., "Frequency dependence of compliance in
apparently healthy smokers versus non-smokers," Bull Physio-Path Resp. 7;195
(1971).
LEEDER, S.R., and WOOLCOCK, A.J., "Cigarette smoking in 12-13 year old
Sydney school children," Med. J. Aust. 2; 674 (1973).
LEEDER, S.R., WOOLCOCK, A.J., and BLACKBURN, C.R.B., "Prevalence and natural
history of respiratory disease in New South Wales school children," Int, J.
•id. 3(1); 15 (1974).
MACKL3M, P.T., "Obstruction in small airways - a ,challenge to medicine
Amer. J. Hed. 52(6); 721 (1972).
MACKLEM, P.T., and MEAD, J., "Factors determining maximum expiratory flow
rate in dogs," J. Appl. "Physiol 25(2); 159 (1968).
PICKEN, J.J., NIEWOEHNER, D.E., and CHESTER, E.H., "Prolonged effects of
viral infections of the upper respiratory tract," Amar. J. Med. 52(6); 638
(1972).
POLLUTION REPORT, "Division of Occupational Health and Pollution Control
Laboratories, New South Wales Health Commission," (1972).
REID, D.D., "The beginnings of bronchitis," Proc. Roy. Spc._Med. 62; 311 (1969).
TOYAMA, T., "Air pollution and its health effects in Japan," Arch. Environ.
Health 8; 153 (1964).
WEIBEL, E.R., "Morphometre of the human lung," Springer_Verla^t Berlin (1963).
ZAPLETAL, A., JECH, J., PAUL, T., and SAMANEK, M., "Pulmonary function
studies in children living in an air polluted area," Amer. Rev_._Resp_._J).is.
107; 400 (1973).
-------
76
DISCUSSION
JOOSTING (Netherlands)
A small, statistically significant correlation between low-
er forced vital capacity and higher pollution areas was found in
older Sydney boys. You mentioned that this correlation was in-
dependent of the effects of other factors such as height, weight
and social class. You did not mention the influence of smoking
explicitly. Could smoking habit be a relevant explanatory var-
iable if we recognize the fact that nowadays smoking starts al-
ready at the age of ten?
LEEDER (Australia)
Regular cigarette smoking is certainly an important factor
to consider when measuring ventilatory capacity, even in child-
ren as you suggest. We included a variable which reflected the
age of onset of smoking and the amount smoked in our calculation
of effects of factors affecting lung function, but the factor
was of no significance in explaining reduced values. This may
well be due to the relatively young sample we have studied, and
we will continue to investigate the effect of smoking in our
ongoing studies.
-------
77
RESPIRATORY SYMPTOMS AND VENTILATORY CAPACITY
IN A COHORT OF LONDONERS BORN IN 1952-3
R, E, WALLER/"*" A, G, F, BROOKS+ AND M, W, ADLER++
+ MRC Air Pollution Unit, St. Bartholomew's Hospital Medical
College, London, United Kingdom
++ Department of Clinical Epidemiology and Social Medicine,
St. Thomas's Hospital Medical School, London, United Kingdom
ABSTRACT
In the first stage of this study 800 Londoners who reached
age 18 in 1970, and who were exposed in the neo-natal period
to very high concentrations of pollutants during the London
fog of December 1952 were examined, using a modified form of
the MRC Questionnaire on Respiratory Symptoms. Ventilatory
capacity was also measured* using a portable spirometer and a
peak flow meter. In the second stage, a further 800 subjects
who reached age 18 in 1971, and were not exposed to the 1952
episode3 were examined in the same way.
Results from these two samples have now been compared,
and no differences have been found in the prevalence of respir-
atory symptoms, nor in the mean values of FEV^, FVC and PEF,
between subjects born before the fog and those born after it.
Respiratory symptoms were much commoner among the smokers
than among the non-smokers and there were small differences in
ventilatory capacity between these categories. The main fac-
tor associated with lowered ventilatory capacity was a history
of respiratory illness, and in view of earlier findings, this
may indicate an indirect long-term effect of exposure to pollu-
-------
78
1. Int roduct ion
In an earlier study, Douglas and Waller f 1; found a
relationship between the incidence of lower respiratory
tract infections in children and exposure to air pollution,
as estimated from domestic coal consumption in the areas
where they lived. This relationship was already evident
when the children were nine months old, and it persisted
at least until age 15. Several other studies in the
United Kingdom have shown a higher incidence of respiratory
diseases or reduced ventilatory capacity among children
living in relatively polluted areas (Lunn et al. [2! , fsl ,
Holland e_t aJL. [4] , Colley and Reid I 5"]) and the possibility
that this has some bearing on the gradual development of
chronic bronchitis has been discussed by these and other
authors. In each of the above-mentioned studies exposures
to pollution were assessed in terms of mean values over
long periods, and there was no way of determining whether
occasional exposures to very high concentrations were of
special importance. The most outstanding episode of high
pollution on record, in terms of adverse effects on health.
was the London fog of December, 1952. It was estimated
(Ministry of Health Tel) that some 4,OOO people died
prematurely as a result of exposure to the very high
concentrations of smoke, sulphur dioxide and associated
pollutants that persisted then for four days. These
"excess" deaths were mainly among the elderly, but a small
proportion were of infants less than one year old. There
was evidence of an increased incidence of acute respiratory
disease at the time, but there was little quantitative
information on its impact on infants in particular.
The opportunity has now been taken to study respiratory
symptoms and ventilatory capacity among two samples of
Londoners, one born during 1952, before the fog, and the
-------
79
other during 1953, just after it.
2. Method
Three small areas of Greater London, chosen to contain
about 1% of the total population and a cross-section of
all social classes, were selected, and within these the
names of all residents reaching age 18 during 1970 were
listed from the electoral registers. After sending out
introductory letters, each subject was visited in his (or
her) own home in the evenings, phasing the interviews over
a 12 month period from October 197O. The same procedure
was adopted for the second sample, except that 1971
registers were used and the interviews started in October
1971. The average age at interview was then just the same
in the two samples, and although there may have been some
biases in the registers themselves, the two samples were
very well matched in respect of every factor other than
the designed one-year difference in date of birth.
The aim in this study was to compare the results for
pre-fog and post-fog births, using the day on which the
fog finally cleared - 9th December, 1952 - as the dividing .
line, and restricting these contrasts to London-born
subjects. Local contrasts between the three areas were
not sought, since it was known that there had been much
movement of families within London since 1952.
A modified MRC Questionnaire on Respiratory Symptoms
(Medical Research council [7]) was used. All the standard
questions were included, with additional ones on childhood
diseases, place of birth and home conditions when the
subjects were babies. (Most of the subjects were still
living with their parents, who helped with this latter
information when necessary.) Five spirometric measurements
were made on a portable instrument (McDermott et al \_Q\} -
-------
80
recording FEV1 and FVC each time, followed by five measure-
ments of peak expiratory flow (Wright and McKerrow T^l).
Calibration checks were made throughout the field-work and
all instruments were returned to the laboratory at regular
intervals to maintain comparability.
3. Results
The total number of subjects seen in the two years
was 1586, representing 88% of the available sample (people
who had already left the district permanently were discounted
but students who were away only in term-time were included).
The proportion of males (52.5%) was a little higher than
that of females, and 86% of the total were born in Greater
London.
Over 4O questions on respiratory symptoms, previous
illnesses, smoking habits and environmental conditions
were asked of each subject, and the results were analysed
initially in terms of the proportion of positive answers
to individual questions. Not all the subjects who reported
persistent phlegm also said that they had a persistent
cough, and despite efforts to ensure that the answer
related to phlegm from the chest, there appeared to be
some confusion with phlegm from the nose in this young
age-group (nasal catarrh was the most commonly reported
symptom, with nearly 4O% reporting catarrh in the winter).
The proportion giving positive answers to both the cough
and phlegm questions (Q5 and Q1O) was therefore determined
as an index of productive cough, indicative of simple
chronic bronchitis (Medical Research Council [lOj). One
other general feature was that the proportion of girls
reporting slight breathlessness (shortness of breath when
hurrying on level ground or walking up a slight hill) was
more than twice that of the boys. This marked difference
-------
81
was found in all areas throughout the survey, and it may
be characteristic of adolescent girls, but this mild degree
of breathlessness was not in general related to reduced
ventilatory capacity.
Results from only the principal questions are shown
in the tables below. The wheezing question included (Q16)
is that relating to attacks of shortness of breath with
wheezing, but normal between attacks (indicative of asthma),
and any chest illness that had interrupted usual activities
for more than a week during the past three years is included
under Q21. The group of questions Q22 to 31 relates to
previous illnesses, mainly bronchitis, asthma and pneumonia,
and the proportion of subjects reporting one or more of
these is quoted. This information depends on recall
(usually by the parents) over an 18 year period, and it is
therefore not very reliable.
Preliminary analyses were done on the ventilatory
capacity measurements to determine the relationship with
height and weight. As in other studies a linear
relationship with height was found, and the results quoted
in the tables have been standardized to heights of 174 cm
for men and 162 cm for women. After standardizing for
height there was no consistent variation with weight, and
since all subjects were virtually the same age (the mean
age at interview was 18.7 yr) there was no need to
standardize for age. Results in this paper are based on
the highest of the five measurements of FEV^, FVC and PEF
made on each subject, but the mean of the last three has
also been used.
-------
82
TABLE I
PREVALENCE OF RESPIRATORY SYMPTOMS (% POSITIVE) AND VENTIL-
ATORY CAPACITY AMONG SUBJECTS BORN BEFORE AND AFTER THE
1952 FOG. LONDON-BORN SUBJECTS ONLY.
Question
Persistent cough
Persistent phlegm
Cough and phlegm
Breathlessness (Gr.2+)
Attacks wheezing
Chest illnesses
Previous illnesses
FEVj_ litres, mean
FVC litres , mean
PEF I/rain, mean
No of subjects
Males
Pre-
fog
8.9
10.3
4.5
11.7
Post-
fog
7.9
8.4
4.7
13.1
12.7* 7.2
12.7
30.2
4.34
5.16
549
291
10.0
24.3
4.39
5.12
546
428
Females
Pre-
fog
8.7
6.0
2.7
26.8
4.7
12.8
23.2
3.20
3.71
431
298
Post-
fog
5.9
6.2
2.9
27.1
7. 1
14.5
24.5
3.26
3.76
425
339
Significant difference (P < O.O5, X test)
Comparisons between pre-fog and post-fog groups are
shown in Table I. For most of the symptoms and measurements
considered there were no significant differences between
the results for subjects born before the fog and those born
-------
83
after it. Among males the proportion reporting attacks
of shortness of breath with wheezing (Q16) was higher for
the pre-fog than for the post-fog subjects, but this was
not consistent between the two sexes.
TABLE II
PREVALENCE OF RESPIRATORY SYMPTOMS {% POSITIVE) AND
VENTILATORY CAPACITY AMONG NON-SMOKERS AND SMOKERS
Question
Persistent cough
Persistent phlegm
Cough and phlegm
Breathlessness
Attacks wheezing
Chest illnesses
Previous illnesses
FEV, litres, mean
FVC litres , mean
PEF 1/min, mean
No of subjects
Males
Non-
smokers
6.2*
7.3*
3.8
7.9*
9.1
10.2
25.8
4.42*
5.15
550
453
Smokers
12.9
12.4
6.6
16.3
9.7
12.9
28.4
4.32
5.12
545
380
Females
Non-
smokers
4.2*
2.6*
0.6*
18.7*
5.4
8.7*
22.7
3.26*
3.74
432*
497
Smokers
12.5
12.5
7.5
45.1
7.5
21.6
27.6
3. 18
3.74
422
255
2
* Significant difference (P < O.05, X test for
questions and t test for ventilatory capacity)
-------
84
The main factor affecting the prevalence of symptoms
was smoking, even though smoking histories in these young
subjects were short, and Table II shows results for smokers
compared with those for non-smokers (the latter group
includes ex-smokers). The prevalence of the main resp-
iratory symptoms (cough, phlegm and breathlessness) was
TABLE III
PREVALENCE OF RESPIRATORY SYMPTOMS (% POSITIVE) AND
VENTILATORY CAPACITY IN RELATION TO HISTORY OF RESPIRATORY
ILLNESSES
Question
Persistent cough
Persistent phlegm
cough and phlegm
Breathlessness
Attacks wheezing
Chest illnesses
FEVi litres, mean
FVC litres, mean
PEF 1/min, mean
No. of subjects
Males
Pneu-
monia
11.8
10.3
7.4
14.7
13.2*
23.5*
4.21*
5.05
525*
68
No ill-
nesses
7.7
7.9
3.6
9.5
3.0
5.4
4.42
5.14
553
608
Females
Pneu-
monia
15.1
5.7
5.7
37.7*
11.3
24.5*
3.17
3.74
417
53
No ill-
nesses
5.8
5.1
2.3
23.0
3.2
7.7
3.25
3.74
430
570
2
* Significant difference (P < O.O5, X test for
questions and t test for ventilatory capacity)
-------
85
significantly higher among smokers than among the others,
and the contrast was particularly marked for females.
There were no differences in FVC between the groups, but
the mean values of FEV^ were a little lower among smokers
than among the others (in females there was a significant
difference in PEF also). There was no reason to believe
that differences in smoking habits could have biased
the results in Table I, and further analyses showed that
the findings were still the same when smokers and non-
smokers were considered separately.
Subjects who said that they had suffered from
bronchitis or asthma during their childhood were seen to
have a high prevalence of current respiratory symptoms and
reduced ventilatory capacity. Those who reported at least
one illness with pneumonia (or pleurisy) also tended to
have a higher prevalence of symptoms and lower ventilatory
capacity than others (Table III). There was also a
relationship between current respiratory symptoms and
ventilatory capacity, the greatest contrast being between
subjects reporting persistent cough with phlegm and those
free from respiratory symptoms.
The results were considered in relation to many other
factors that had been studied, including the social class
of the family, the size of family, the type of house in
which they lived and the season of birth, but none of
these had any substantial or consistent effect on the
prevalence of symptoms, nor on ventilatory capacity.
Questions on parental smoking had been asked, and there
was some evidence that this was related to an increased
prevalence of symptoms, but in view of the general
correlation between the subjects' own smoking habits and
those of the parents, it was difficult to determine whether
this was an independent effect.
-------
86
4. Discussion
The results of this study have shown quite clearly
that exposure in infancy to the London fog of December 1952
has had no dominant effect on the development of respiratory
symptoms up to age 18. The maximum age of subjects in
this study at the time of the fog was 9 months, and it is
possible that such young children may have been kept
indoors, slightly (but by no means wholly) protected from
the pollution. A more extensive study would be required
to follow up people who were a little older at the time
of the fog. The most important point may however be that
each of the two groups studied (pre-fog and post-fog) was
exposed in the remaining childhood years to the high levels
of pollution that prevailed in London through the 1950's
and into the 196O;s. Although the 1952 fog has not been
matched in severity, there have been many episodes of high
pollution in London, for example in January 1956, December
1957 and December 1962 (Waller and Commins fill ). There
are few data for comparable groups of 18-year-olds brought
up in relatively clean conditions, but the prevalence of
respiratory symptoms in the whole of the sample was higher
than in a group of medical students studied at the same
age (Lawther. Brooks and Waller fl2l ) This latter group
must however be regarded as selected, and further work is
planned on general population samples in London and in a
relatively "clean" area.
The importance of respiratory illnesses in childhood
in relation to the subsequent development of respira-
tory symptoms, which was apparent in this study, has also
been seen in some other investigations. In the longitudinal
study of a sample of children born in 1946, the prevalence
of symptoms at age 2O, although affected mainly by smoking,
has been found to be related to the past history of
-------
87
respiratory illnesses (Colley, Douglas and Reid [l3J ) .
It had already been shown (Douglas and Waller ^ij) that
the incidence of lower respiratory tract infections in
this same sample was related to the estimated exposure to
pollution and there may therefore be a link between
childhood exposures to pollution, repeated respiratory
infections and the subsequent development of chronic
bronchitis. The retrospective information on childhood
illnesses in the present study of 1952-53 born subjects
was not adequate to substantiate this point any further,
but the dramatic changes in pollution in London during
the last 2O years may provide an opportunity for further
studies, in which the prevalence of symptoms in the present
1952-53 cohort can eventually be compared with that in a
1962-63 cohort, born around the time of the last of the
major London fogs, and finally with that in a 1972-73
cohort, born into the "clean air" conditions of today.
References
1 DOUGLAS, J.W.B., WALLER, R.E., "Air pollution and
respiratory infection in children", Brit. J. prev. soc.
Med.,20, 1 (1966).
2 LUNN, J.E., KNOWELDEN, J., HANDYSI.OE, A.J., "Patterns of
respiratory illness in Sheffield infant schoolchildren",
Brit. J. prev. soc. Med., 21, 7, (1967).
3 LUNN, J.E., KNOWELDEN, J., ROE, J.W., "Patterns of
respiratory illness in Sheffield junior schoolchildren",
Brit. J. prev. soc. Med., 24. 223 (1970)
4 HOLLAND, W.W., HALIL, T., BENNETT, A.E., ELLIOTT, A.,
"Factors influencing the onset of chronic respiratory
disease", Brit, med. J., 2, 2O5 (1969).
5 COLLEY, J.R.T., REID, D.D., "Urban and social origins
of childhood bronchitis in England and Wales", Brit.
med. J. , 2, 213 (197O) .
-------
88
6 MINISTRY OF HEALTH, Mortality and morbidity during the
London fog of December 1952. H.M. Stationary Office,
London (1954).
7 MEDICAL RESEARCH COUNCIL, Questionnaire on respirat-or-y
svmbtoms. M.R.C.. London (1966)
8 MCDERMOTT, M,, McDERMOTT, T.J., COLLINS, M.M., "A
portable bellows spirometer and timing unit for the
measurement of respiratory function", Med. biol.Engng..
6, 291 (1968)
9 WRIGHT, B.M., McKERROW, C.B., "Maximum forced expiratory
flow rate as a measure of ventilatory capacity",
Brit, med. J.. 2, 1O41 (1959).
10 MEDICAL RESEARCH COUNCIL, "Definition and classific-
ation of chronic bronchitis for clinical and
epidemiological purposes". Lancet. 1, 775 (1965).
11 WALLER, R.E., COMMINS, B.T.. "Episodes of high poll-
ution in London, 1952-1966", Proc. Int. Clean Air
Conf.. London, Part 1, 228, N.S.C.A., London (1966).
12 LAWTHER, P.J., BROOKS, A.G.F., WALLER, R.E.,
"Respiratory function measurements in a cohort of
medical students", Thorax. 25. 172 (197O).
13 COLLEY, J.R.T., DOUGLAS, J.W.B., REID, D.D.,
"Respiratory disease in young adults: influence of
early childhood lower respiratory tract illness,
social class, air pollution and smoking", Brit, med. J..
3, 195, (1973)
DISCUSSION
JACOBSEN (U.K.)
The author suggested that the observed higher prevalence of
respiratory symptoms among those who reported respiratory disease
in early life may indicate that exposure to the 1952 fog may have
had an indirect effect (via early childhood disease). Is this
not a hypothesis that can be verified from the data? Would it
not imply that children exposed to the smog should have shown
higher symptom-prevalence?
-------
89
WALLER (U.K.)
It is true that if exposure to the 1952 fog had had an in-
direct effect on the subsequent development of respiratory symp-
toms, by enhancing the incidence of respiratory illnesses among
babies at the time, then this should have shown up in the con-
trasts between the pre- and post-fog samples. The problem is
that we do not have an adequate index of the exposure of indivi-
dual members of the samples to pollution at that time. Whether
they were kept indoors or taken outside during the fog would
have made a material difference to their exposure to at least
some components of the pollution and it would have been inter-
esting to have been in a position to compare results in relation
to an index of exposure within the pre-fog sample.
STUPFEL (France)
You show that there are differences between males and fe-
males concerning phlegm but not concerning FEV. We have been
able to demonstrate experimental sex related differences in
acute mortality between males and females in mice and rats ex-
posed to high concentrations of carbon monoxide and automotive
exhaust gas. Do you have data for mortality differences rela-
ted to sex and to acute air pollution episodes independent of
smoking habits and occupational exposure?
WALLER (U.K.)
There are sex-related differences in the prevalence of some
of the symptoms considered in our study and in the absolute val-
ues of the ventilatory capacity measurements, but in general
factors such as smoking have the same kind of effect on these
quantities within each sex.
We do not have detailed information on mortality in the 1952
fog by sex, but full details of day-to-day changes in mortality
by age, sex and cause have been collected for more recent periods
and these will be reported soon.
KEVANY (Ireland)
Would age at first school attendance have been a more sen-
sitive time to judge pre- and post-smog effects as school going
children would be subject to more uniform exposure through tra-
velling to school.
-------
90
WALLER (U.K.)
Ages under 1 year were chosen for the present study because
of evidence from other work that babies in that age range appeared
to be particularly susceptible to environmental influences. The
age-range 5-6 could provide an alternative: the children may
then be affected more by cross-infections on first starting school
but at least it is known that they have to go out in all weathers.
We have some information on children who were aged 6 in 1952,
from our earlier study of the "1964 cohort" and this will be ex-
amined further.
-------
91
A STUDY ON HEALTH EFFECT INDICES CONCERNING POPULATION
IN CADMIUM-POLLUTED AREA
HIROMU WATANABE AND HISAKO MURAYAMA
Public Health Institute of Hyogo Prefecture, Japan
ABSTRACT
1. Stages of Cd metabolism, classified by urinary Cd excretion
and tubular function.
Accumulation stage: the stage when no abnormal increase of
urinary Cd excretion is found, (school age in low and high pol-
luted area).
Saturation stage: the stage when increase of urinary ex-
cretion of low molecular weight protein is not clear, and there
is found an increase of urinary Cd excretion,(adult age in low
polluted area).
Excretion stage (the former term): the stage when a sharp
increase of urinary excretion of Cd, Cu and low molecular weight
protein is observed, (adult age in high polluted area).
Excretion stage (the latter term): the stage when excre-
tion of low molecular protein in urine is still clear, but ex-
cretion of Cd in urine has tendency to decrease, (adult age in
high polluted area).
2. Indices of tubular disorder by Cd intake.
Lysozyme, B9~M G and Proline in urine are reasonable indices
-------
92
for tubular dysfunction. The pattern of proline excretion in
urine is different between male and female.
Increase of 69-W G in serum is reasonable index of tubular
&
dysfunction, though it is often seen in aged persons having no
finding of tubular dysfunction.
It seems that female has other accelerating factor concer-
ning development of tubular dysfunction by Cd intake. Tubular
disorder by intake of Cd polluted rise is accelerated by aging.
It seems that Cd level in urine is reflected to Cd level
in blood and, on the other hand, Cu level in urine is not rela-
ted to that in blood.
The meaning of Cu intake for the cause of tubular dysfunc-
tion is not possible to be ignored.
-------
93
Introduction
Cadmium (Cd, in abbreviation) pollution of rice was noticed in the
areas around Ikunc Mine (located in the central part of Hyogo Prefecture)
in the early spring of 1971. Then, Cd pollution health effect surveys were
made on inhabitants of JO years or more (the number, 13,052), in the areas,
where rice contained Cd of over 0.*+ ppm, and also in the areas adjacent to
them. The total number of those areas are 9, which are subdivided into 54
districts.
Based on the results of the surveys, the following considerations were
made on Cd health effect indices.
All subjects belonged to farmers' family and, as a staple food, ate
rice produced in the areas where they lived.
High-polluted area; an area where 0.6-1.0 ppm of Cd was detected in
unpolished rice. (Copper ( Cu, in abbreviation) concentration in the same
rice was about 6 ppm.) ( H in abbreviation)
Low-polluted area; an area where around 0.3 ppm of Cd was detected in
unpolished rice. (Cu in the same rice was about 3 ppm) (L in abbreviation)
Non-polluted area; an area where no Cd environmental pollution was
found and Cd below 0.1 ppm was detected in unpolished rice. (Cu in the same
rice was about 2 ppm) (N in abbreviation)
Time of survey;
NO.l. April, 1971. At that time, the subjects were recommended not to
eat rice containing Cd over 1.0 ppm.
NO.2, October, 1971. 6 months after the recommendation.
N0.3« June, 1972. About 1 year after the recommendation.
NO.4. June-October,1973- About 2 yearjafter the recommendation.
Results
I. Relationship between in take of Cd-polluted rice and urinary excretion.
1. Mean Cd concentrations in urine arj hair of adults and schoolchildren,
by Cd concentration in unpolished rice harvested in the areas where
they lived. (Time NO. 1)(Table 1)
(1) Kean Cd concentration in urine of adults in H area is higher than
that of L area. (2) Cd concentration in urine of schoolchildren is not
.related to Cd pollution. (3) Cd concentrationin urine of adults is higher
than that of schoolchildren. (4) Cd concentration in hair of adults is
higher than that of schoolchildren. (5) Cd concentration in hair is not
-------
94
Table 1. Mean Cd concentrations in urine and hair of adults and school-
children, by Cd concentrations in unpolished rice harvested
in the areas where they lived.
(Examined in April,1971.)
Gd
concentration
in unpolished
rice
ppm
0.33
0.35
0.61
0.88
1.10
Number
of area
Ik
11
3
5
6
Adult
Urinary
protein
positive rate
%
V.k
10.2
6.9
12.1
13.5
Cd in
urine
;ug/i
7.1
7.7
11.1
13.0
14.2
Cd in
hair
PS/K
0.67
O.82
0.78
0.73
1.06
School-children
Cd in
urine
jug/1
2.3
1.8
3.4
3.5
2.0
Cd in
hair
PS/S
0.49
0.41
0.37
0.67
0.63
Table 2\ G/A ratio and concentration of protein, Cd and Cu in urine by
Cd pollution. (Protein possitive persons)
Cd
concentration
in unpolished
rice
0.33PP°
0.35
o.6l
0.88
1.10
n
35
3*
49
50
55
G/A ratio
0.60
0.62
1.47
1-53
1.80
protein
in urine
ll^*1
7.4
*.9
5.1
4.7
Cd in urine
5.8Mg/1
5-5
7.8
11.7
10.5
Cu in urine
3^.7Mg/1
36.6
44.7
50.1
49.0
The method of the quantitative analysis of urinary protein is Kings-bury
-Clark's .
-------
Fig. 1 Comparison of cumulative frequency distributions of
logarithmic values of urinary protein concentration
(by biuret method), about those subjects who showed
positiveness for urinary protein test (by sulfosalicylic
Cumulativ
99.99'
99.9.
99-
00
80
60
C.I
acid method) ,between H and N areas.
40 100 *tOO 800
Urinary protein concentration r.g/dl
-------
96
2. G/A ratio and geometrical mean of concentrations of protein, Cd and
Cu in urine of subjects who showed possitive for urinary protein test (by
TCA method), by Cd concentrations in unpolished rice harvested in the areas
where they lived. (Time NO.l) (Table 2)
(1) G/A ratio and Cd or Cu concentrations in urine in H area are
higher than those of L area. (2) Concentrations of protein in urine
is not related to Cd pollution.
3. Prevalence of glucosuria.
(l) prevalence of glucosuria is higher in H area than in L area. (By
age and sex matched data) H area: 4.1 %', L area: 2.7 % (combistix test
paper) (Time NO.2)
k. Comparison of cumulative frequency distributions of logarithmic values
of urinary protein concentrations between Hi and N area. (Time NO.3) (Fig..l)
(l) Subjects in non-polluted area were 82, Who were the whole number
having showed urinary protein concentration of 1*4 mg/dl or more.
Subjects in H area were 99« who were sampled, matched with them in N
area concerning compositions of age and sex, and also showed urinary protein
concentration of lU mg/dl or more.
This means this comparison is the complate comparative study on such
subjects who showed relatively high urinary protein excretion, between H
and N areas. Both distribution diagrams are quite similar except one
(marked with an arrow) in H area. Therefore, the existence of some hetero-
geneous group, i.e., a group of high urinary protein excretion, seems to
be very unlikely.
5. Comparison of urinary protein concentrations between H and N areas by
biuret method with sex and age matched data. (Time NO.3)
(1) Geometrical mean of H area : 2.19 mg/dl (n=3?6), L area : ?.. ^2
mg/dl (n=250). It was not shown the significant difference between H and
L areas.
6. Relation between Cd and protein or between protein ard G/A ratio in
urine. (Time NO. 2)
(l) Relation between Cd and protein in urine, by Cd concentrations of
unpolished rice was not significant. (2) Relation between G/A ratio and
urinary protein concentration was significant in H area.
II. Relationship among aging, Cd intake and urinary Cd excretion, protein
or others. (Time NO.2)
1. Relation between aging and urinary Cd concentration was unclear.
(Table 3)
-------
97
Table 3. Relation among G/A ratio, Cd and Cu in urine by aging.
(Protein possitive persons)
§
0
o
H- 0
3 P-
C 0
•1 O
H- 3
3 0
rt> O
3
ci-
ts P
<5 £T.
I-1 0
v_x 3
H- 0
3 £
C O
*i o
H- 3
3 O
CD n
3
rt-
"t P
I-1 O
— 3
^^^^^^^^^^^^^_a
Age
30-39
4O-49
50-59
60-69
70-79
£0-89
30-39
40-49
50-59
60-69
70-79
80-89
30-39
40-49
50-59
60-69
70-79
n
8
11
8
18
17
6
8
11
8
18
17
6
^^—^^-^^—
8
9
8
18
17
80-89 1 •*
Low-
Polluted area
0..47 ± 0.17
0.62 ± 0.46
0.55± 0.26
0.70± 0..48
0.90 + 0.60
1.43 ± 0.47
7.75 ± 5.00
8.45± 5.14
6.00 ± 4.27
6.22± 3.88
ID. 29 ± 6.70
8.33± 3-93
__— — — — - — —
53-8 ± 37.£
39.4+16.7
32. 5 ± 27-1
33-9 ±19.1
53-2 ± 29.2
4?. 5± 20. 6
n
16
19
23
47
62
13
16
19
23
47
62
13
18
21
22
46
61
12
Highly-
Polluted area
0.93± 0.17
1.24i 0.38
1..56 ± 1.07
1.75± 0.72
2.21+ 0.6?
2.O2± 0.74
11,50± 9.64
13.10 ±7.70
14.04 +9.49
12.44± 5.92
11.61 ± 6.58
11.51 ± 6.29
43. 3 ± 26.0
44.0 ± 21.0
56. 8 ± 37.1
53-7+ 27.8
62.1 ± 30.4
105.8 ± 77.7
•
-------
98
2* Increase of urinary Cu concentration or G/A ratio was related to age.
( Table J) Urinary protein and glucose were related to age too. And these
increases were more accerelated in H area than in L or N area.
III. Relationship between G/A ratio and Cd or Cu concentration in urine or
hair. (Time NO.2) (Table *0
Table k. Variation of Cd or Cu concentration in urine and hair by grade
of G/A ratio. (Protein possitive persons)
G/A ratio
n
Age
Urine
Hair
Protein mg/dl
concentration
G/A ratio
Cd /ig/1
concentration
Cu jug/1
concentrat ion
Cd jug/g
concentration
Cu jug/g
concentration
Pb /ig/g
concentration
0.71
and below
in
54
5-5
0.4
4.7
23
0.5
13
10
0. 72 ~ 1.54
11
61
4.6
1.0
9.6
31
0.9
13
9
1.55
and over
12
74
8.5
2.3
9.4
53
0.4
11
5
Remark; the method of quantitative analysis of urinary protein is
Kingsbury-Clarke.
1. Increase of Cu concentration in urine is related to G/A ratio.
2. Cd concentration in urine has tendency to decrese in higher grade of
Q/A ratio.
3. Cd, Cu and Pb concentrations in hair have tendency to decrease in
higher grade of G/A ratio.
IV. Comparison of urinary excretive components influenced by tubular
-------
99
5» Prevalence of Bj.-M G and others in urine in H and N area.
( Atrandom sample)
Glucose
Protein
(10mg/dl<)
Ba-M G
Lysozyme
T-amino-N
(200mg/l<)
Age m
Highly-polluted area
n=39
3 0. 8
5 3. 9
k 1. 0
2 0.5
3 0. 8
7 6. k
Non-polluted area
n=56
2 3. 2
*t 1. 1
3. 6
3. 6
3 3. 9
7 ±. 6
Table 6. Prevalence of Lysozyme positive by aging.
(Urinary glucose positive persons)
Area
H
L
N. Age
SexV
male
female
Total
male
female
Totale
50-59
% (n)
0 (9)
0 (7)
0 (16)
0 (47)
9 (2*0
3 (71)
60~69
% (n)
0 (7)
29 (13)
19 (20)
0 (*K3)
9 (33)
k (73)
70-79
% (n)
69 (I1*)
76 (25)
7^ (39)
6 (3D
AO (32)
23 (63)
80 --
% (n)
87 (8)
62 (6)
75 (1A)
Mf (9)
61 (28)
57 (37)
Total
* (n)
M (38)
52 (51)
^9 (89)
5 (127)
29 (2MO
16 (107)
-------
100
1. Prevalence of Bj.-M G and others positive in urine for at random
subjects. (Table 5)
(1) Positive of urine component examined with the way as follows :
Glucosuria-testape {+), Protein (10 mg/dl over), Bt-M G (lOfig/ml over),
Lysozyme (1 fig/al over), T-amino-N (200 mg/1 over). (2) Subjects were
selected at random in population aged 70 years ovrer in H and N areas.
(3) The differences of prevalenceof Lysozyme and fti-M G between H and N
area, were clear.
2« Prevalence of Lysozyme positive by aging. (Table 6)
(1) Excretion of Lysozyme in urine waa incresed by aging and was
accerelated by Cd pollution.
Table ?• Concentration of B^-M G and others in urine.
(Urinary glucose, protein and Globulin positive persons)
Area
Sex
Urine
Serum
Glucose
(mg/dl)
Protein
(mg/dl)
H E p , , ^
(jug/ml)
Lysozyme
(jig/ml)
B*-M G
(Mg/ml)
T-amino-N ^i}
Proline(ffig/i)
Cd
Cu (MS/I)
20 (Mg/D
B*-M G , , .
(jug/ml)
IP (mg/dl)
Hiehly-polluted area
male
71.6
38.9
8.0
0.7
27.3
249.2
0.3
13.7
79.8
322.9
3.7
3.0
female
77.7
36.1
9.0
2.7
37.0
250.4
19.2
9.*
62.6
183.6
4.0
3.1
Low-polluted area
male
70.0
15.3
1.5
0.1
17.0
144.4
0.2
6.7
32.6
322.1
3.0
3.0
female
68.2
39.5
6.O
1.3
32.0
206.7
3.5
6.7
49.2
222.2
3-9
3.5
-------
101
Table 8. Correlation coefficient between Cd and Iii-M G in urine.
Cd
Uu
Sex
male
female
male
female
Area
H
L
H
L
H
L
H
L
n
15
15
18
28
15
15
18
28
Iii-M G
- 0.01
**
0.63
- 0.31
0.32
**
0.74
***
0.83
O.IK)
• * »
0.60
Proline
- 0.18
- 0.07
- 0.54*
0.20
0.84**
*
0.51
* *
0.6?
0.25
Lysozyme
- 0.07
0.26
- 0.26
* *
0.55
0.93
* *
0.7^
0.37
* *
0.52
* P<0.05 , ** P
-------
102
Table 10. Cd or Cu concentration in blood, hair and urine,
Blood
(jig/ml)
Hfcir
-------
103
3. Comparison of mean value of glucose, fi*-M G, Cd in urine and Bj-M G,
inorganic P in serum. (Time NO.'t) (Table 7)
(l) Subjject was persons, who had finding of tubular dysfunction such
as glucosuria, proteinuria and G fraction positive by Disc electrophoretic
examination simultaneously. (2) It was clear the difference of concentra>*-
tions of Lysozyme and proline in urine between male and female. (3) It was
not clear the difference of concentration of each component between H and
L area except proline.
U. Relation of Cd or Cu to B2-M G, Proline and Lysozyme in urine.(Table 8)
(1) Cd concentration in urine was not related to those of Iii-M G,
Lysozyme or proline in H area, but it was related to thos in L area.
(2) Cu concentration in urine was related to those of &-V( G, Lysozyme or
Proline in each area.
5. Relation of fJ2-M G in urine or serum to protein, R B P and others.
(Table 9)
(1) It is seemed that fi>2-M G in serum is related to tubular dysfunction.
V. Body burden Index.
1. Comparison of Cd or Cu concentration in blood, hair and urine between
H and L areas. (Time NO.4) (Table 10)
(1) Cd concentration in blood in H area was higher than that of L area.
2. Relationship among Cd or Cu concentration in hair, urine and blood.
(Table 11)
(l) Correlation coefficient between Cd concentrations of blood and
urine in H and N area was significant.
DISCUSSION
WASSERMANN (Israel)
Table 3 in your paper indicates a positive age correlation
with Cadmium concentration in urine both in low and highly pol-
luted areas. The same appears for Copper in urine in both
areas. Since this correlation is positive, only up to age
groups of approximately 40 - 49, perhaps these groups of 30 - 39,
40 - 49 years would be most significant for characterizing the
urinary concentration of a population in an area.
-------
104
WATANABE (Japan)
At first it was expected that some effects would be found
on the group of 30 - 50, but there were no findings other than
an increased Cadmium excretion in urine.
CHAMBERS (Ireland)
Do the differences in excretion showing tubular dysfunction
between females and males occur at all times of the menstrual
cycle or is there some change during the cycle?
WATANABE (Japan)
We found the difference of tubular dysfunction between male
and female on the group of over 70 years old. In view of this,
any relation with the menstrual cycle should not be considered.
-------
105
CLINICAL INVESTIGATION IN NORTHWEST QUEBEC, CANADA
OF ENVIRONMENTAL ORGANIC MERCURY EFFECTS
AARON D, BERNSTEIN
Medical Services, Health and Welfare, Ottawa, Canada
ABSTRACT
In 1971, residents of three Cree Indian communities in
Northwest Quebec, Canada, were investigated for possible clini-
cal effects of organic mercury. This was done after finding
that mercury levels of fish caught in the area ranged to greater
than three parts per million.
In the first phase of the investigation, over three hundred
persona had blood samples drawn for mercury determination. Many
of these individuals had hair specimens for mercury determination
collected as well. Approximately half this number had field
clinical testing done. Clinical evaluation included full or
screening types of history and physical examinations, electro-
myograms, and maze performance testing.
Twenty-two individuals were found to have whole blood mer-
cury levels greater than one hundred parts per billion, the high-
est reading being three hundred and six parts per billion. There
was good correlation between hair and blood mercury levels.
There appeared to be some relationship between the individual's
fish consumption and his blood mercury level. Ho significant
clinical findings suggestive or organic mercury excess were
detected.
In the second phase of the investigation, five individuals
from one of these communities, four of whom had previously shown
-------
106
blood mercury levels greater than one hundred parts per billion*
underwent extensive in-patient evaluations at the Montreal Neu-
rological Hospital in Montreal, Canada. In addition to routine
hospital tests, detailed neurological examinations, as well as
electroencephalograms* audiograms, visual field determinations,
blood cytogenetic studies, electromyogram and nerve conduction
studies, and blood and hair mercury level determinations were
carried out.
Here again* no significant clinical findings suggestive of
organic mercury excess were detected.
-------
107
1. Introduction
In the spring of 1971, samples of whole blood from twenty-nine Cree
Indians of the Waswanipi band in Northwest Quebec, Canada were tested for
mercury by the Federal Department of Fisheries and Forestry (now integrated
into the Department of the Environment). This was done after finding that
mercury levels of fish caught in the Waswanipi area ranged to over 3 parts
per million (ppm). The average mercury level found in the Waswanipi blood
samples was 55 parts per billion (ppb) with a high of 135 ppb. It was
thus felt necessary to carry out more extensive testing.
In addition, because it was reported that mercury content of fish in
Mistassini Lake northeast of the Waswanipi area was quite high, some
testing was also carried out amongst the Mistassini band of Indians.
2. Phase I
2.1 Method^
Complete history and physical examinations including specific inquiry
about fish consumption, as well as collection of blood for determination
of mercury levels, were first carried out in June 1971 on Waswanipi
Indians, who are of Cree origin, living in Matagami, Quebec.
These initial history and physical examinations were negative for
symptoms and signs of organic mercury poisoning. Therefore a screening
battery of tests based on recommendations made at the Symposium on Human
Mercury Exposure (held at the Center for Disease Control, Atlanta,
Georgia, June 25-26, 1971) was devised to be used for the Waswanipi
Indians in the area of Miquelon, Quebec, who were evaluated in July, 1971.
The procedure at Miquelon was to enquire how the individual was
feeling. He was then specifically asked about parestheslae. If there
were any significant positive answers to these questions, a complete
history and physical examination was carried out. If not, a battery of
tests were carried out. This battery included:
(1) Detection of any hand tremor.
(2) Testing for dysdiadochokinesia.
(3) Finger-nose testing with eyes both open and closed.
(A) Testing of vibration sensation at the ankles with a tuning fork.
(5) Testing of two-point discrimination of the index finger tips
using a paper clip.
(6) Having the person draw a circle.
-------
108
(7) Testing of auditory acuity - first with a watch and If this
was positive (not heard), using a tuning fork.
All the above subjects were also questioned about fish consumption.
These subjects plus others who did not undergo examinations (particularly
school children who spent much of their recent time away from the
community) had blood collected for mercury determinations.
The mercury determinations on the blood samples were carried out
by the Fisheries and Forestry Department In Montreal using atomic
absorption. Portions of six of these samples were sent to the laboratory
of Dr. T. W. Clarkson at University of Rochester for independent
verification.
These areas were visited shortly thereafter by a technician from the
Institute of Environmental and Industrial Health of the University of
Michigan who (1) performed electromyograms using a portable apparatus to
obtain power spectra plots; (2) administered a maze performance test; and
(3) collected hair for mercury determination to be done by atomic
absorption at the University of Michigan.
In late July, 1971* a medical student collected blood specimens for
mercury determinations from Cree Indians living in Mistasslnl, Quebec, and
in August, 1971, he repeated blood collections in Matagaml and Mlquelon,
Quebec.
It had been hoped that visual field examinations would be performed
during the summer of 1971 as well. However, because of problems entailed
in obtaining persons competent to carry out such examinations, these
were not done until October 1971, at which time they were carried out by
two different optometrists - one for Matagaml and one for Mlquelon.
Advance notice was generally given the community involved that
examinations were to be carried out, and virtually anyone who presented
himself for testing had it carried out.
2.2 Results
The results are summarized in Table I. Seventy-nine persons at
Matagaml had blood tested for mercury, ten of these had one or more
readings of greater than 100 ppb, the mean was 41.04 ppb and the highest
reading obtained was 306 ppb. This latter reading was obtained on a
fifty-one year old man in August 1971 (he had had no blood taken in
June, 1971). One hundred and forty-one persons from Miquelon were tested,
three had readings greater than 100 ppb, the mean was 21.63 ppb and the
-------
TABLE I
Summary of Phase I Evaluation
T«*C Sit*
Hatagami, Quebec
(Waawanlpi Band)
Miqutlon, Quebec
(Haswantpi Bind)
Miataealnl, Quebec
(Miataeaini Band)
BLOOD MERCURY
Number of
P«non»
Tested
79
141
181
NumUr>.
100
ppb
10
3
9
Highest
Value
Recorded
306 ppb
148 ppb
155 ppb
Mean
41.04
ppb
21.63
ppb
36.76
ppb
HAIR MERCURY
Number of
Paraona
Teated
11
tf
r = 0.82
vlth blood
traluea
£
Hlghe»c
Value
Recorded
44.85 ppa
Physical
Examina-
tiona
45 full
NO
Electro-
fflyograiQ* t
Mai* Per-
formance
11
SIGNIFICANT
69 acreenlng
29 full
NO
Not done
56
SIGMIFICAMT
Not done
Viiual rialda
8
FEIDINGS
Hot done
FINDINGS
19
NO SIGNIFICANT
FIOTISCS
-------
110
highest reading here was 148 ppb. In Mistasslni one hundred and eighty-
one persons were tested, nine had readings greater than 100 ppb, the
mean Was 36.76 ppb and the highest reading was 155 ppb.
Six of the blood specimens collected from Matagaml in June, 1971
were centrifuged shortly after collection, so that red blood cell mercury
levels could be determined. Portions of these blood specimens were
analyzed at University of Rochester while an equal portion was sent to be
analyzed by Dr. Magos of the Medical Research Council. There was a
reasonably good relationship between the Fisheries and Forestry results
and those of Magos.
Eleven persons from Hatagaoi and fifty-six persons from Miquelon
had hair specimens analyzed at the University of Michigan. The highest
reading, 44.85 ppm was on a specimen from a thirty-eight year old man
living at Matagaml, who had had a blood mercury level of 172 ppb. A
correlation of .82 was found between blood and hair mercury levels.
Full physical examinations were carried out on forty-nine individuals
at Matagaml and on twenty-nine individuals at Miquelon. Screening exam-
inations were done on sixty-nine individuals from Miquelon. Electro-
myograms and maze performance tests were carried out on eleven individuals
from Matagaml and forty-six individuals from Miquelon. Finally, visual
field examinations were carried out on eight of the previously tested
Individuals from Matagaml and on nineteen of the previously tested
individuals from Mistasslni. In all these clinical examinations no
significant findings were detected.
However, there did appear to be a relationship between admitted heavy
fish consumption and raised blood mercury levels in that where asked,
those individuals with levels greater than 100 ppb admitted to fish
consumption more than once per week whereas the majority of individuals
stated their fish consumption was once a week or less. Also, with one
exception, all those children who had blood specimens collected in early
summer 1971 and had attended school in towns away from home for the
previous two years or more, had mercury levels of 0 (the one exception had
a blood mercury level of 21 ppb).
One final point of interest is brought out in comparing the cumu-
lative blood mercury levels of the first twenty persons from each com-
munity who came forward to have blood specimens collected, with the final
-------
Ill
twenty. The results suggest that those within a community who came
forward first to be tested are less likely to have high mercury levels
than those who came forward towards the end of testing.
TABLE II
Total Blood Mercury (ppb) for Groups of 20
Matagami Miquelon Mistassini
(June 1971) (July 1971) (August 1971)
First twenty 253 498 496
Final twenty 1115 784 896
2.3 Discussion
It must be pointed out that due to the nature of this field study,
valid statistical analysis, particularly of the physical examinations
(including visual fields), is virtually impossible to perform. Nonetheless
there was no suggestion of any clinical sign or symptom appearing more
often than expected within such populations.
The electromyogram power spectra plots were subjected to more rigid
statistical analysis, but here too no significant correlation was found
with blood or hair mercury levels.
The blood mercury levels done by the Fisheries and Forestry
Department appear reasonably valid, both statistically and clinically.
They were found to correlate highly with hair mercury levels. Too, a
good relationship was found between Fisheries and Forestry red blood cell
mercury levels and those done by Magos of the Medical Research Council.
And, for further support, there appeared to be a relationship between
fish consumption and blood mercury levels.
The reasons why those who first come forward to have their testing
done have lower mercury levels than those who come at the end are not
clear-cut. However, in general, it seems that women and children tend to
come forward first, while the men, particularly those who have been
-------
112
working or who spend much time "in the bush" tend to come forward last.
Possibly, too, those who are more "urbanized" I.e. those who tend to
spend more time in the community, utilizing the supplies of the local
general store and relying less on natural sources for their food supply
thus consuming less fish than the man "out in the bush", are less wary of
the outsider.
This point is brought out as a caution for future studies, to ensure
that appropriate selection is made when choosing a sample population from
within a community. Special efforts might be considered to test indi-
viduals out hunting, trapping and fishing in the bush.
More detailed studies of fish mercury levels, carried out in
September and October, 1971 revealed that the levels ranged to a
of 4.44 ppm in the Waswanipi area and to 0.84 ppm in the Mistassini areaPO
3. Phase II
3.1 Introduction
In that no significant clinical findings were detected in the field
studies of Phase I, it was decided to offer some of the individuals with
higher mercury levels the opportunity to be investigated more thoroughly.
3.2 Methods
Originally It was hoped to select ten individuals, some with known
high mercury levels and others with more "normal" levels, to undergo
complete investigation at the Montreal Neurological Hospital in Montreal,
Quebec. However, due to uncontrollable circumstances, only five persons
were investigated, four of whom had previously shown blood mercury levels
of greater than 100 ppb and one who had not previously been tested. At
the time of the investigation, these individuals were aware that high
mercury levels had been found in the area, but were not yet aware whether
they themselves were amongst those who had the higher levels.
All testing was carried out on an outpatient basis within an Interval
of five days in February, 1972. Use was made of two Cree Indian inter-
preters. The following procedures were carried out:
1. Complete medical and neurological examinations - performed by
a neurology resident of the hospital.
2. Psychological testing - performed by hospital psychologists.
It must be stated that this aspect of the evaluation created
fl) Unpublished data, Environment Canada.
-------
113
the most difficulties, as there were inadequate norms from
which to work and the language barriers were virtually
insuperable despite the presence of the interpreters.
3. Electromyogram and nerve conduction studies.
4. Electroencephalograms.
5. Audiograms.
6. Visual field determinations.
7. Cytogenetic studies - in which blood was drawn at the hospital
and sent to the Health Protection Branch of National Health &
Welfare in Ottawa for analysis.
8. Blood and hair mercury levels - specimens being collected at
the hospital and sent to the same department as in No. 7.
Blood mercury levels were determined using atomic absorption -
while hair mercury levels were determined utilizing neutron
activation.
Routine hospital tests done included:
9. Chest X-rays.
10. Hemograms.
11. Complete Urinalyses.
12. SMA - 12
3.3 Results and Discussion
Details of the persons Investigated and their findings are available
upon request to the author. (See appendix)
Although non-specific abnormalities of the type expected within such
a population were, of course, found, no significant clinical findings
suggestive of a chronic effect of mercury were seen. Electromyography,
electroencephalography, visual fields, and blood cytogenetics are all
normal in those persons with the elevated mercury levels.
4. Conclusions
1. Some moderately elevated blood mercury levels have been found
amongst Indians of the Waswanipi and Mistassini Bands.
2. There is a satisfactory correlation of blood and hair mercury
levels in the studies done.
3. There appears to be some relationship between an individual's
fish consumption and his mercury level.
4. No significant clinical findings suggestive of organic mercury
excess have been detected.
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
116
Recommendations
1. Blood and/or hair should be screened for mercury periodically
in areas where high environmental mercury levels are found.
2. In screening individuals, special effort should be made to
obtain specimens from individuals who have spent much time
outside of their settlement, fishing, trapping and hunting.
3. Detailed clinical evaluations should be offered to selected
individuals with higher mercury levels.
A. Open and frank communication with individuals and groups being
evaluated should be encouraged.
DISCUSSION
DANIELSON (Sweden)
1. Was mercury in fish measured as organic mercury?
*
2. What was the average mercury level in fish?
3. Was fish consumption measured in more specific terms, e.g.
g/day?
BERNSTEIN (Canada)
1. No, not in the specific studies referred to in the paper.
It is generally accepted however, on the basis of investi-
gations conducted elsewhere that the majority of mercury present
in fish muscle is methylmercury.
As mentioned in the paper, six blood specimens were centri-
fuged to separate the cellular fraction. In these specimens
the majority of mercury was found in the red cell fraction,
lending support to the concept that one was dealing predominantly
with organic mercury.
2. In the Waswanipi area, average whole mercury fish levels
ranged from 0.16 ppm to 1.98 ppm depending on the species of
fish and the body of water tested.
In the Mistassini area the average levels ranged from 0.25
to 0.60 ppm.
3. No, not in this study. In a few instances where this has
been attempted, under comparable circumstances, the information
derived has not been very satisfactory.
-------
117
BRAMAN (U.S.A.)
Did you determine if the source of environmental mercury
pollution (prior to absorption by fish) was actually inorganic
or organic mercury?
BERNSTEIN (Canada)
No, not in this study. Sources in comparable situations
elsewhere in Canada would appear to be predominantly inorganic
mercury.
It must be pointed out that natural sources of mercury (as
contrasted to man-made sources) seem to play a significant role
in the total environmental mercury level found in regions such
as those we studied.
-------
119
SCALING OF ANNOYANCE IN EPIDEOLOGICAL STUDIES
B, BERGLUND+, U, BERGLUND^ AND T, LINDVALL
+ Psychological Laboratories, University of Stockholm, Sweden
++ Acoustical Laboratory, Department of Environmental Design,
Royal Institute of Technology, Stockholm, Sweden
+++ Department of Environmental Hygiene, Karolinska Institute
& Swedish Environment Protection Board, Stockholm, Sweden
ABSTRACT
The measurement of annoyance evoked by environmental agents
often involves survey techniques. Specific problemsr such as
effects of response criteria and comparability between popula-
tion responses, are riot well solved by using the assumptions
behind the traditional survey methods. In criteria work the
meaningfulness and effectiveness of such data are critical,
strongly underlining the need for improvements of the methods.
The introduction of modern scaling techniques into surveys can
give the necessary improvements. Investigations on general
annoyance with regard to pulp mill odors and aircraft noise
demonstrate that it is possible to measure the population res-
ponse criteria as well as the population response as such.
Guidelines are given for annoyance measurements that result in
calibrated measures permitting comparisons between differently
exposed areas, that fulfill the assumptions for commonly used
statistical procedures, and thus enable the determination of
meaningful dose-response relationships.
-------
120
1. Introduction
A recurrent task in environmental research and community
planning is to measure annoyance evoked by environmental agents in
the community. The typical tool has been the sociological survey
from which dose-response relationships have been constructed for
e.g., community noise (e.g., Traffic Noise in Residential Areas (1);
Rylander, Sb*rensen & Kajland (2); Patterson & Connor (3)).
In an annoyance survey, usually a representative group of
persons is presented questions on annoyance with several well
defined response categories. Two principle approaches are common
for estimating the annoyance with regard to an environmental agent;
either (1) a battery of questions constituting a Guttman-like
attitude scale to obtain annoyance scores or (2) self-rating
questions to obtain direct annoyance estimates. Both approaches
were employed for example in the Heathrow studies (McKennell (4)).
This paper will not deal with attitude scales but the measurement
of degree (intensity) of annoyance as expressed in responses to
self-rating questions.
Traditionally, the results of the self-rating questions are
treated in two different ways depending on whether the aim is to
measure the extensity of annoyance (cf. Galloway & Jones (5)) or
the mean degree of annoyance in the population. The extensity
of annoyance is determined by simply calculating the proportions of
persons responding in certain categories (e.g. Rylander, SBrensen
& Kajland (2); Patterson & Connor (3)). Such proportions are
sometimes regarded as a sufficient basis for dose-response descrip-
tions. If instead the mean degree of annoyance is to be quantified,
category scaling has been utilized (e.g. Grandjean et al. (6);
-------
121
McKennell (7)).
Annoyance surveys are mostly conducted in populations chronica-
lly exposed to the agent in question. This circumstance always
limits the generalization of results for criteria work. The
responses from exposed individuals may be biased due to the exposure
itself, differences in attitudes, and other so called extra-
expositional factors, i.e. their response criteria* may differ from
the criteria of the non-exposed individuals. Similarly, the
response criteria for populations exposed but to a varying degree
may differ (see e.g., Thibant & Kelly (8)). This problem has been
noticed in the laboratory (Borsky & Leonard (9)) as well as in the
field (McKennell (10)). To obtain comparability between differently
exposed populations the variation of the response criteria must be
known and, if necessary, corrected for.
Psychophysical scaling research has developed quantitative
methods suitable for measuring the degree of annoyance with regard
to environmental agents in such a way as to permit a satisfactory
control of the response criteria variation (for a review of these
methods see Ekman & Sjb'berg (11)). The present team has demonstra-
ted the applicability of psychophysical scaling methods in the
annoyance survey. The emphasis is put on the measurement of mean
degree of annoyance and the need for calibration procedures; these
seem to be two critical factors arising when annoyance responses
are compared or a prognosis is made of the annoyance in the
community.
Response criteria are here defined as the respondents- locations
of the response-category boundaries on the continuum of annoyance.
-------
122
2. Theoretical Considerations
When self-rating questions are chosen for annoyance surveys
many investigators utilize only a small amount of the information
contained in the survey data. The proportions of persons in the
response categories above an arbitrary cut-off point are lumped
together, disregarding the distribution of the respondents over
the categories. This way of utilizing the data is illustrated in
Fig. 1 for a 7-category scale where the cut-off point is set at the
4th category boundary. The cut-off point could for example
dichotomize the respondents into "little annoyed", and "much
annoyed" persons. The main question is to what extent the pro-
portion of "much annoyed" persons represents not only extensit^y but
also the intensity of annoyance. In the following, three theoreti-
cal models of measurement are discussed with regard to their
appropriateness to estimate extensity as well as intensity of
annoyance.
Trallic
-
(?) Decree of
annoyance
Fig. 1. Theoretical frequency distributions of annoyance evoked by
aircraft noise and traffic noise. The shaded area shows
the proportion of persons more annoyed than is represented
by the 4th category boundary.
-------
123
2.1 Model 1
The assumptions behind the traditional treatments and uses of
survey data are rarely discussed in the literature from a measure-
ment point of view. The unspoken but basic assumption can be
stated as follows: All persons, not only within an area but also
between differently exposed areas, use fixed response criteria.
This assumption Is Illustrated in the left diagram of Fig. 2
(Model I) where hypothetical distributions of responses to, e.g.,
traffic noise, are shown for two differently exposed areas, A and
B. The shaded part of the distributions represents the proportion
of persons that respond in annoyance categories above a certain
boundary (here the Ifth) on a verbal category scale. The critical
question is whether the cut-off point is fixed and equal in Area A
and B. This is a necessary condition even if the extensity of
annoyance is to be measured. Moreover, the choice of the cut-off
point, i.e., its location on the annoyance scale, is more or less
critical depending on the form and location of the response
distribution.
Although Model I is frequently used for determining the
extensity as well as the intensity of annoyance, several factors
critical for the validity of the basic assumptions are reported in
the literature. For example, acclimatization towards greater
tolerance to the exposure conditions (McKennell (7)) can be one such
factor. Also, the findings of Hazard (12) and Grandjean et al.
(6) indicate a dependence between annoyance response to the target
agent and the exposure conditions in general. Demonstrated
differences in the interpretation of the phrasing of questions and
category labels constitute another factor (cf. Jones & Thurstone
-------
124
(13); Jonsson (14). Further, practical experience has shown that
induced changes in attitudes may be accompanied by drastic changes
in proportions of annoyed persons (Sorensen (15)). Such results
may reflect not only a change of degree of annoyance response to an
environmental agent but also a dislocation of the response criterion.
2.2 Model II
The center diagram of Fig. 2 illustrates Model II where the
4th category boundary is assumed to be fixed within an area but
differs between areas in terms of degree of annoyance. In this
hypothetical example, the 4th boundary represents a higher degree of
annoyance in Area B than in Area A. If the 4th category boundary
is chosen as the cut-off point xhe proportion of annoyed persons
will be the same in both areas. When the comparison is made on
equal terms, i.e. at the same degree of annoyance, the proportion
of annoyed persons is higher in Area B. Thus, if there is a
difference in response criteria between areas they should be
compared with respect to the same cut-off point on the annoyance
scale, regardless of the verbal labels of the response alternatives.
To do this, a calibration of the underlying annoyance scales must
be made.
2.3 Model III
If the meaning of the 4th category boundary varies between and
within areas, the measurement problem becomes even more complicated.
An illustration of Model III is presented in the right hand diagram
of Fig. 2. The distribution to the left represents the annoyance
felt by a population in response to a disturbance and the distribu-
tion to the right represents the annoyance associated with the 4th
-------
125
"OOEl I
MODEL 01
i I
Fig. 2. Models for three different views on response criteria in
differently exposed populations.
Aircraft no/ir Traffic rtoisr
Area A
Dvgree of *nnoy«nc«*
Traffic ne/jr
Arra B
Fig. 3-
-------
126
category boundary. Both distributions contain the same individuals.
It follows that when Areas A and 6 are compared, simple proportions
cannot be used. Accordingly, if Model III describes reality it is
impossible to obtain a measure of the extensity of annoyance.
Instead comparison of areas must be made in terms of calibrated
measures of mean degree of annoyance.
Fig. 3 shows Model III with hypothetical annoyance agents vair-
craft and traffic noise) together with the set of response alterna-
tives to a self-rating question. The two agents as well as each
category boundary give rise to response distributions. The
distances between the category boundaries are allowed to vary within
as well as between areas. It should be stressed that according to
the model the response given by each individual depends not only on
the degree of annoyance evoked by the environmental agent but also
on his response criteria, i.e. where he locates the category
boundaries. By Thurstonian scaling of responses to self-rating
questions all the information collected is utilized.
3. Bnpirical Studies
In the following the three theoretical models (Fig. 2, Models
I-III) are applied to data from self-rating questions on annoyance
collected In surveys on pulp mill odors and aircraft noise. The
stress is put on the applicability and meaningfulness of the models
: rather than on any specific dose-response relationship. A detailed
presentation of the two surveys is given elsewhere (Berglund, Berg-
lund, Jonsson & Lindvall (16), Berglund, Berglund & Lindvall (17)).
Both surveys covered differently exposed subareas. The questionalras
included items on "annoyance in general" with regard to several
environmental agents. The respondents were to rate their annoyance
-------
127
on verbal category scales with 6 or 7 categories (cf. Fig. 3).
3.1 The Traditional Approach
Model I is implicit in the traditional treatment of data for
estimating the extensity of annoyance. The model requires that
the response criteria corresponding to the cut-off points are
fixed both within and between areas. A cut-off point is thus
assumed to represent the same degree of annoyance in all the areas
compared. In Fig. k results are shown for the two self-rating
questions on annoyance treated according to Model I. Two cut-off
PULP HILL ODOH
AIRCRAFT NOISE
1.00
.ts
,*S
.00
—
I
—
TT,
;X
//
/
f s ,-
i
—
\
—
9 T.UU
.7S
I
.SO
.25
^—
\
\
I
—
I
1
I
Fig. ^. Proportion of annoyed persons responding to a question on
general annoyance in two surveys. The open and shaded bars
represent the results when two different category boundaries
are used as cut-off points. (Sample sizes ?6-175 persons).
points have been chosen for the illustration: One representing all
responses including and above "fairly little annoyed" (open bars)
and the other including and above "fairly much annoyed" (shaded
bare).
The results for both the odor and the noise questions show
that the rank orders of the areas are not identical for the two cut-
-------
128
off points. Even though the differences in extensity of annoyance
are email, this is the expected outcome if the response distributions
differ between areas. That response distributions can vary with
the exposure to community noise has been demonstrated nicely by
Borsky fr Leonard (9). Mainly for the theoretical reasons presented
we are inclined to abandon Model I and to strive for models of
measurement that require less rigid assumptions.
3.2 Alternative Approaches
In Model II and III the locations of the verbal categories
along the annoyance continuum are allowed to be different among
areas. In Model III, this location may also vary within the areas.
These conditions are illustrated in Fig. 3. The measurement approach
to satisfy Model II is limited to a calibration of the underlying
annoyance scales so that the verbal response categories are
comparable. It should be stressed that a calibration cannot be
made on the proportions directly. Therefore a solution according
to Model III is imperative. Model III requires a mathematical
treatment of the data according to Thurstone's Law of Category
Judgment (see Torgerson (18)). The specific assumptions behind
the solutions applied are that the variances of each response
category boundary are constant and independent of the specific
environment agents.
In the treatment of the collected data on "annoyance in
general" with regard to pulp mill odors and aircraft noise, the
least rigid model, i.e. Model III, has been used. It should be
noted that Models I and II are special cases of Model III so that a
solution according to Model III will also be a solution of Model I
and II.
Thurstone's Law of Category Judgment demands that the frequency
-------
129
of response be tabulated for each environmental agent for each
response category. This frequency is then transformed to a pro-
portion of the total number of responses. The proportion is
further transformed into a normal deviate. It is from the normal
deviates that the continuum of annoyance can be derived (Gulliksen's
method of least squares; cf. Torgerson (18)). Once the annoyance
scale is established, locations of the boundaries for the verbal
categories and the environmental agents are easily determined.
Single annoyance values are obtained for each agent and category
on the same interval scale of annoyance. The interval property
of the scale means that the annoyance scales for different areas
can be calibrated to a common reference point and a common unit of
measurement.
3.2.1 Scales of Response Category Boundaries
The Thurstonian scaling of annoyance in the two surveys gives
the locations for the category boundaries that are presented in the
lower part of Fig. 5. The upper parts of the diagrams present the
conventionally used response model for self-rating questions (e.g.
Borsky & Leonard (9); GrandJean et al. (6)). It simply states that
the interdistances between the category boundaries are of constant
size. Obviously the empirically obtained category interdistances
are unequal within as well as between areas for both the surveys.
The fact that the category sizes vary between areas invalidates
the use of proportions of persons above any response category (cf.
Model I) as a measure of extensity of annoyance. The result that
the category sizes are unequal is by no means unexpected. For
example, in well controlled laboratory experiments adverbs of
degree (e.g. slightly, quite, very) have been shown to produce
-------
PULP MILL ODOR
AIRCRAFT NOISE
Conventionally assumed category ilzt
Arta A
Arm B
Arra C
Art* 0
Empirical category tlzr
Arra
Arra JF
Conventionally assumed category size
Art* I t-
(D
A fee I t-
Afea Y I-
Emplrical category size
Fig. 5« Thurstonian scales of the category boundaries from two
surveys. For comparison the conventionally assumed
locations of the category boundaries are shown in the
upper part of the diagrams.
OJ
o
-------
131
unequal steps when used as modifiers of an adjective (e.g. "annoyed")
(Cliff (19); for survey data see McKennell (10)).
3.2.2 Calibrated Scales of Annoyance to Environmental Agents
The Thurstonian scaling method provides annoyance measures of
both response category boundaries and environmental agents on inter-
val scales. However, direct comparisons of annoyance responses
between different areas require a calibration of the scales. This
means, as stated earlier, that the annoyance scales must be trans-
formed to the same unit of measurement and to a common reference
point. There are several possible ways to do this calibration.
Common to all of them is that at least tuo reference anchors have
to be included in the survey. In the present investigations the
following anchors were tried: (a) annoyance evoked by an experi-
mentally introduced reference stimulus (standard odor or standard
noise) and (b) annoyance evoked by the traffic noise in a nearby
city. The units of measurement were defined as the distances
between the two anchors on the scale, the anchors also serving as
the common reference point. The choice of anchors is merely an
illustration of the calibration principle and should not be regarded
as a recommendation of specific anchors.
The usefulness of Model II can be illustrated by the calibrat-
ed annoyance scales. In Fig 6, left hand diagram, the results of
the aircraft noise question treated according to Model II are shown
for areas comparable from a noise exposure point of view (take-offs
only). In the figure frequency distributions of the annoyance
responses with regard to aircraft noise are drawn. The curves are
obtained from normal ogives fitted to the cumulative proportions
and the annoyance scale values. For illustration a cut-off point is
set at the same degree of annoyance value for all areas, N.B. not
-------
132
t 50
-
50
f 50
tree 1
Arm
J7V.
© ©
Arro IS
2fl.
® © ® ©
Degree of annoyance
Ar?0 I
A TO HI
Cy S> C*3
© Q) ® (5)
Degree ofannoyance
Fig. 6. Thurstonian scales of "annoyance in general" with regard to
aircraft noise for three differently exposed areas. The
left hand diagram shows obtained frequency distributions of
annoyance. The shaded parts of the distributions show the
proportion of the populations that are more annoyed than is
represented by the cut-off point. The right hand diagram
shows the mean degree of annoyance with regard to aircraft
noise in the three populations. All scales are calibrated,
allowing comparisons between areas.
at a certain verbal category boundary (cf. Fig 2, Model I and II).
The results in Fig. 6 show the proportion of persons that are
more annoyed than the degree of annoyance represented by the cut-
off point (shaded part of the curve). Being obtained from
calibrated scales, the proportions in the different areas are compar-
able on equal terms. Calibrated scales are a minimum requirement
-------
133
if administrative decisions are to be based on proportions of
annoyed persons.
Another tactic is to measure the mean degree of annoyance
evoked in the population by the environmental agent. Such annoyance
studies have been performed earlier but we have not found any case
where Model III has been applied. Instead, Model I has been used
that requires equal category sizes on the response scale; however,
in doing so Borsky and Leonard (9) have clearly pointed out the
problem with incomparable scale units. Direct measurement of degree
of annoyance is possible to perform on the traditional survey data
provided the sample sizes and number of response categories are
large enough. The results from such an analysis (Model III) are
presented in Fig. 6, right hand diagram. The data are from the
aircraft survey. The annoyance scales are calibrated in the same
way as was done in the left hand diagram. As is evident from the
figure, the "aircraft noise in general" is most annoying in Area I,
less annoying in Area III and least annoying in Area IV just as
would be expected from the physical exposure point of view.
Moreover, by the same technique (Model III) it is possible to
compare the degree of annoyance for several environmental agents
both within and between areas provided the proper calibration
procedures have been applied. This possibility is illustrated by
the empirical results from one of the areas in the aircraft noise
investigation, in Fig. ?• The curves are fitted by the same
procedure as in Fig. 6. In this specific area "aircraft noise in
general" was shown to be less annoying than the reference anchors.
The degree of annoyance for each agent is expressed by the mean of
the distribution. These means are located on an interval scale
of annoyance and therefore quantitative comparisons are possible.
-------
134
so
25
Area I
O Aircraft
A Traffic n«/»r
D Standard niitr
© © (D © (5) (D Degree of annoyance
Fig. 7- Empirically obtained frequency distributions of annoyance
with regard to three different noises. The means of the
distributions represent the degree of annoyance.
3.2.3 Concluding Reaarks
It has been demonstrated that the extensity approach of treat-
ing annoyance data in terms of proportions gives uncertain informa-
tion about the mean degree of annoyance. Even if a supplementary
calibration of annoyance scales for the different areas is made, the
proportions may still be useless for prognosis purposes. This stems
from the fact that if we admit the existence of response criteria
variation it is impossible to say precisely how much a certain
change in physical exposure will change the proportion of annoyed
people. However, if instead the mean degree of annoyance to the
disturbance factor is actually measured on a calibrated interval
scale, then a meaningful dose-response relationship can be establish-
ed. Only then can precise recommendations be given e.g., as to how
much the exposure level has to be reduced in order to meet
established celling limit values with respect to population annoyance.
Before extensity of annoyance according to Model I could be
considered for decisions in community planning and investment proof
would have to be presented that the basic assumptions are realistic.
But it should also be noted that by increasing the sophistication
in the data treatment the requirements put upon the underlying
assumptions become more lenient. From this point of view the
-------
135
application of Model III is safest in practical work. Further,
the appropriateness of the basic assumptions for all the three
models can be checked only by solutions based on Model III.
4. Guidelines for Improvements of the Annoyance Survey
The following recommendations are given for improvement of
annoyance scaling based on self-rating questions in epidemiological
studies. The recommendations are given primarily from a psycho-
metric point of view.
First, whenever economically possible Thurstonian scaling
should be used to determine the mean degree of population annoyance.
Second, to permit such a scaling the questionaire must include a
large number of response categories (at least 6), preferably using
only adverbs of degree as modifiers of the annoyance adjective.
Third, the survey should include questions on the annoyance evoked
by several environmental agents, not only by the target factor.
Fourth, to allow for a satisfactory data treatment subsamples of
respondents must be comparatively large; ideally at least 200
persons are needed in each subsample to get a reliable population
measure of annoyance. Fifth, if areas or populations are to be
compared, the annoyance scales must be calibrated. Such a
calibration is particularly necessary when annoyance responses to
different environmental agents are to be determined. To achieve
this at least two reference anchors have to be introduced into the
survey. There is a strong need for development of suitable
reference anchors for practical work.
ACKNOWLEDGMENTS
The present work was supported by grants from the Swedish
Environment Protection Board, the Swedish Council for Social
Science Research, and Swedish National Council for Building
Research.
-------
136
References
1. Traffic Noise in Residential Areas, Report from the National
Swedish Institute for Building Research, No. 36 E (1968).
2. Rylander, R., Sorensen, S., Kajland, A., "Annoyance reac-
tions from aircraft exposure", Journal of sound and Vibration,
24, 419-444 (1972).
3. Patterson, H.P., Connor, W.K., "Community responses to air-
craft noise in large and small cities in the USA". In pro-
ceedings of the International Congress on Noise as a Public
Health Problem, Dubrovnik, May 13-18, 1973. U.S. Environ-
mental Protection Agency, Washington, D.C. (1973)
4. McKennell, A., "Aircraft noise annoyance around London
(Heathrow) airport". Central Office of Information, London
(1963).
5. Galloway, W.J., Jones, G., "Motor vehicle noise: Identifi-
cation and analysis of situations contributing to annoyance".
In Proceedings of the International Congress on Noise as a
Public Health Problem, Dubrovnik, May 13-18, 1973. U.S.
Environmental Protection Agency, Washington, D.C. (1973).
6. Grandjean, E., Graf, P., Lauber, A., Meier, H.P., Miiller, R.,
"A survey of aircraft noise in Switzerland", In Proceedings
of the International Congress on Noise as a Public Health
Problem, Dubrovnik, May 13-18, 1973. U.S. Environmental
Protection Agency, Washington, D.C. (1973).
7 McKennell. A., "Second survey of aircraft noise annoyance
around London (Heathrow) airport", Her Majesty's Stationery
Office, London (1971) . *
8. Thibant, J.W., Kelley, H.H., The Social Psychology of Groupes,
Wiley, New York (1959) .
9. Borsky, P.N., Leonard, H.S., "A new field survey-laboratory
methodology for studying human response to noise". In
Proceedings of the International Congress on Noise as a
Public Health Problem, Dubrovnik, May 13-18, 1973. U.S
Environmental Protection Agency, Washington, D.C. (1973)
10. McKennell, A., "Psycho-social factors in aircraft noise
annoyance". In Proceedings of the International Congress
on Noise as a Public Health Problems, Dubrovnik, May 13-18,
1973. U.S. Environmental Protection Agency, Washington D.C
(1973).
11. Ekman, G., Sjoberg, L., "Scaling", Annual Review of Psychology
16, 451-474 (1965).
12. Hazard, W.R., "Predictions of noise disturbance near large
airports". Journal of Sound and Vibration, 15, 425-445
(1971).
-------
137
13. Jones, L.V., Thurstone, L.L., "The psychophysics of
semantics: An experimental investigation", Journal of
Applied Psychology, 39, 31-36 (1955)..
14. Jonsson, E., "On the formulation of questions in medico-
hygienic interview investigations", Acta Socioloqica, 7,
193-202 (1964) .
15. Sorensen, S. , "On the possibility of changing the annoyance
reactions to noise by changing the attitudes to the source of
annoyance", Nordisk Hygienidk Tidskrift, Suppl. 1. 1-76
(1970).
16. Berglund, B., Berglund, U., Jonsson, E., Lindvall, T.,
"On the scaling of annoyance to environmental factors",
Unpublished manuscript, University of Stockholm and
Karolinska Institute, Sweden (1974).
17. Berglund, B., Berglund, U., Lindvall, T., "A study of re-
sponse criteria in populations exposed to aircraft noise",
Unpublished manuscript, University of Stockholm and
Karolinska Institute, Sweden (1974).
18. Torgerson, W.S., Theory and Methods of Scaling, Wiley, New
York (1958).
19 Cliff, N., "Adverbs as multipliers", Psychological Review,
66, 27-34 (1959).
DISCUSSION
(Editorial Comment: the following discussion represents a con-
siderable expansion of the exchange of views which took place
orally).
RYLANDER and SORENSEN (Switzerland)
The paper "Scaling of annoyance in epidemiological studies"
presented by Berglund et al concerns methods to assess the
presence of annoyance reactions in populations exposed to envi-
ronmental agents. The authors criticize the conventional
methods, mainly because of a postulated variation in response
criteria and propose the use of a method involving Thurstonian
scaling to evaluate annoyance.
The following presents a critical analysis of the paper.
Initially the traditional social survey method for the deter-
mination of annoyance will be reviewed, whereafter the paper by
Berglund et al will be discussed.
Measurement of annoyance
Annoyance due to environmental agents may be generally de-
fined as "a feeling of displeasure associated with any agent or
condition realized or believed by an individual or a group to
be adversely affecting them" (Borsky 1972).
-------
138
In exposure to single events the momentary reaction can be
defined as acute annoyance. When a collection of individuals
is exposed to several events over a longer time period, certain
individuals might experience chronic annoyance.
The chronic annoyance reaction is related to activity dis-
turbances caused by the exposure. It is also dependent on
extra-expositional factors such as the susceptibility of the
individual, socio-economic conditions and attitudes. Due to
the influence of these factors the relation between individual
annoyance and the exposure level is never very high (e.g. Hazard
1971) . The. relation between the exposure and the mean reaction
in a sufficiently large group of the population is however pre-
cise and the reproducibility is good (Alexandre 1974).
To study the extension of annoyance in a population, social
survey methods are used. They must be performed according to
several criteria which have been well defined. For instance,
in order to prevent the respondent from consciously exaggerating
or underestimating his annoyance, the reason for the investi-
gation must be masked. Questions on the specific agent studied
have to be included in a battery of questions on other agents
in the vicinity. The choice of the verbal expression for an-
noyance is also important (Jonsson 1964).
The annoyance is usually expressed in various degrees -
little annoyed, rather annoyed, very annoyed or as various steps
on an annoyance scale, or a multiple question index. The dis-
tribution of the various degrees of annoyance in a population is
'not normal and the intervals between the various degrees are un-
equal (Jonsson and Fog, 1968). It has been shown that the
highest degree of annoyance is best related to the level of the
exposure agent. Lower degrees of annoyance are more influenced
by extra-expositional factors and thus less suitable for the
construction of dose-response relationships (Rylander et al
1972a, Sorensen et al 1974).
Discussion of Berglund paper
In their presentation Berglund et al discuss the possibi-
lity that the response criteria for persons expressing annoyance
may be different between areas with different exposure levels
or in newly exposed areas. In principle the points raised are
important and should be evaluated against information available
in the literature.
Concerning the variation in response criteria between areas
with different exposure levels available data seem to contradict
this hypothesis. Alexandre (1974) and Rylander et al (19723)
demonstrated close dose-response relationship for a large num-
ber of areas exposed to varying levels of aircraft noise. The
same consistency has largely been found for traffic noise in
English and French investigations.
As a support for the hypothesis that response criteria will
vary Berglund et al discuss acclimatization to the agent, a dep-
endence to the agent, phrasing of the questions and attitudes.
-------
139
All this represents well known phenomena (Borsky 1972) and the
traditional social surveys are designed so that the influence
or these factors is minimized.
The problem concerning reactions in exposed and non-exposed
areas is of particular importance for prognostic purposes. Ex-
perimental data are available also on this aspect. Jonsson et
al (1974) reported data from annoyance due to oil refinery odors
and noise from traffic, shooting ranges and aircraft. They
compared the prognostic values on annoyance based upon field
studies around similar locations with values found when the real
exposure situation has been established. The agreement was
found to be very good. These results do not support the theory
that the variation between non-exposed and exposed areas would
be of importance for the accuracy of planning.
To test the possibility of varying response criteria in
areas with different exposure levels, Berglund et al use data
from field surveys on odors and aircraft noise. They group
the annoyance data from different areas in these studies into
two categories: "fairly little annoyed and more" and "fairly
much annoyed and more".
They found in figure 4 in their paper that the ranking
order of the areas is different for the two categories of an-
noyance. From this they conclude that the response distribu-
tion differs and that the response criteria vary between areas.
In their report they are not, however, using the relevant an-
noyance data from the studies they cite. The annoyance has
been assessed using a question on annoyance constructed by the
authors themselves. The usual question "how much annoyed are
you?" reflects the individual's experience in his actual expo-
sure situation. The question "what do you think about aircraft
noise in general?" constructed by Berglund et al, has no refer-
ence to the particular exposure situation of the individual.
The criteria for annoyance studies as referred to in the previ-
ous section have thus not been adhered to.
If instead the relevant annoyance data from the studies
are used, it is seen that the ranking order between areas is
not different for the various groups of annoyance (Table 1) .
Rankings for different categories of annoyance has also been
evaluated in all areas in the investigation from which Berglund
et al have chosen 5 areas. The results are shown in Table 2.
It is seen that by and large the ranking order for the two
groups of annoyance are similar. Differences in ranking order
are present between two groups of areas. The differences for
the percentages are however very small and within the accepted
variation.
The data analysis presented by Berglund et al to support
their hypothesis must therefore be rejected.
-------
uo
Epilogue
Even if it is not possible to prove the hypothesis presen-
ted by Berglund et al concerning the shortcomings of the tradi-
tional methods for the study of annoyance, it is still of in-
terest to discuss their proposals for improved social surveys.
They recommend the utilization of all degrees of the an-
noyance response to construct an average annoyance level. As
only the highest degree of annoyance is well correlated to the
exposure level (see section on measurement of annoyance), this
approach seems to be less valuable.
The second proposal is the correction of the annoyance
using "reference anchors". This would be an interesting ap-
proach if verbal differences were present in the expression of
annoyance reaction. This has not been demonstrated.
The correction procedure requires that the distance between
the anchors is identically evaluated by all individuals in the
study. Concerning the examples for anchors "how to you consi-
der traffic noise in the city during rush hours", this reference
point - in the absence of any data on its validity - seems to be
even more influenced by experience and extra-expositional factors
than the traditional questions on annoyance criticized by the
authors for this reason.
The second anchor involves an evaluation of an acute expo-
sure to noise or odors. It has earlier been shown that the
exposure to acute stimuli correlates better to the physical
noise level than to chronic annoyance (Borsky 1974). Thus
this anchor does not appear to be useful.
In view of the above, it seems difficult to postulate that
the distance between the two anchors will be constant, until
experimental data have been produced which demonstrate this.
Other anchors might be more useful but in the absence of
any suggestions and data to demonstrate their validity this
will remain speculative.
Conclusion
Social survey methods are incomplete in many ways. It
would be especially desirable to increase the accuracy for the
expression of the annoyance reaction in the community, as well
as make it possible to use the reaction of the individual rather
than a group mean. The paper presented by Berglund et al
contains many interesting ideas. In view of the absence of
data which prove their validity they will remain speculative
and do not present a support for the conclusions found in the
paper.
-------
141
Table 1
Extent of annoyance (%) and ranking order (R.O.) between differ-
ent areas in odor/aircraft noise studies using relevant data for
the different groups of annoyance.
AREA
Aircraft noise
MMA 3
MMA 2
MMA 5
MMA 1
MMA 4
Odor
BBY 1
BBY 2
BBY 3
LITTLE ANNOYED
OR MORE
% R.O.
41 1
18 2
12 3
9 4
6 5
93 1
84 2
49 3
RATHER ANNOYED
OR MORE
% R.O.
13 1
12 2
6 3
5 4
3 5
72 1
65 2
34 3
Table 2
Extent of annoyance (%) and ranking order (R.O.) between areas
with different exposure levels to aircraft noise (data from
Rylander et al 1972).
AREA
OSL 6
OSL 4
OSL 5
CPH 5
CPH 4
OSL 2
OSL 3
CPH 6
CPH 3
CPH 2
CPH 1
LITTLE ANNOYED
OR MORE
%
78
73
67
62
55
48
47
40
22
13
12
R.O.
1
2
3
4
5
6
7
8
9
10
11
RATHER ANNOYED
OR MORE
%
70
71
66
60
55
44
45
40
18
11
12
R.O.
2
1
3
4
5
6
7
8
9
10
11
-------
142
References
1. Alexandra, A., (1974) Aircraft noise annoyance - temporal
and spatial comparisons in Noise as a Public Health Problem,
Dubrovnik 1973.
2. Borsky, P.H., (1972) Sonic boom exposure effects-annoyance
reactions, J.Sound Vibr. 20:527-530.
3. Hazard, W.R., (1971) Predictions of noise disturbance near
large airports, J. Sound Vibr. 15:425-445.
4. Jonsson, E., (1964), On the formulation of questions in
medico-hygienic interview investigations, Acta Sociol.,
7:193-202.
5. Jonsson, E,, Sorensen, S. and Arvidsson, O., (1974), Reli-
ability of forecast of annoyance reactions of exposure to
noise and air pollution. Arch. Envir. Health, in press.
6. Jonsson, E. and Fog, H. (1968), Traffic noise in residen-
tial areas, Natl. Swedish Inst. Building Research, 36 E :
1968.
7. Rylander, R., Sorensen, S., Berglund, K. and Brodin C.
(1972a) Experiments on the effect of sonic boom exposure
on humans. J. Acoust. Soc. Amer. 51 : 790-798.
8. Rylander, R., Sorensen, S. and Kajland, A. (1972b). An-
noyance reactions from aircraft noise exposure. J. Sound
Vibr. 24 : 419-444.
9. Sorensen, S., Rylander, R. and Berglund, K. (1974).
Reaction patterns in annoyance response to aircraft noise
in Noise as a Public Health Problem, Dubrovnik 1973.
BERGLUND (Sweden)
The comments of Dr. Rylander made during the session have
later been changed and considerably expanded. Accordingly, I
am forced to give a more elaborate answer than X gave in the
session.
i. One of the main objectives of the present symposium has
been the concern for improvements of methodologies. Our paper
presents a new psychophysical approach applicable in the annoy-
ance survey. Therefore, I am surprised by Dr. Rylander's
negative reaction to our paper. I cannot understand how our
suggested improvements of methodology could be harmful to the
annoyance surveys.
-------
143
2. To avoid misunderstandings I will repeat the main points
of our paper. These are:
- to openly display the assumptions behind different methods
of annoyance assessment,
- to clarify the requirements of the annoyance measurement
methods before they can be used in the construction of
precise dose-response functions, and
- to recommend the use of calibration procedures and to sug-
gest how such calibrations can be done.
Although the methodological problems brought up by us may
have implications for specific investigations, I will restrict
myself to respond on materials presented in this session.
3. Dr. Rylander argues that we criticize the conventional
measurements because of a postulated response criteria variation.
We do not postulate such a variation but base our view on empi-
rical evidence presented by us and other researchers. The full
references are given in our paper. Dr. Rylander refers to the
same findings and underlines the problem as such. He argues
that the influences of the response criteria are minimized by
the design of the survey. That has, however, not been shown.
We do criticize the traditional methods used for scaling
annoyance because it is implicitly postulated that the response
criteria are invariant when proportion of persons is used as a
measure of annoyance. We recommend methods that permit us to
check the influence of the response criteria and, if needed,
make the necessary corrections.
4. Dr. Rylander advocates that the reproducibility of his
results is a measure of the value of the traditional survey and
concludes that the problems raised by us are irrelevant in
practice. We have to keep in mind that reproducibility is an
elusive indicator because it is inversely related to the preci-
sion of the measurements. Thus a crude measurement technique
will always result in a high probability of good reproducibility.
It should be noted that the annoyance measures obtained in the
traditional surveys that Dr. Rylander refers to are at their
best rank orders of annoyance. It would be surprising if the
rank orders could not be reproduced.
5. In the paper we show that the choice of the cut-off point
is more or less critical depending on the distribution of res-
ponses over categories. The problem is illustrated by Dr.
Rylander's own results that are presented in Tables 1 and 2 in
his written comments. The good outcome of the rank orders is
not a strong argument for the traditional survey, especially
when Table 2 is consulted. There all persons, except for a
few, are located above both the cut-off points he compares.
In my opinion this result can have implications for Dr. Rylan-
der 's notion of a frequency maximum of aircraft flyovers as
presented in his paper at this conference.
-------
144
6. Dr. Rylander argues that the question of annoyance used by
us is too general to have any reference to the noise exposure
situation of the individual. Of course, there may be differ-
ences in the interpretation of slightly different questions,
but I can hardly believe that the responses to the question on
"annoyance to aircraft noise in general" are not at all related
to the experiences of the aircraft noise exposures.
Regardless of this strange dispute on the phrasing of ques-
tions, Dr. Rylander's comments are irrelevant for the principles
of measurement raised by us. It is up to him to show that
"his" question is not influenced by response criteria variation.
The absence of calibrated measures has many times led to invalid
conclusions from results, e.g., in environmental toxicology.
The need for calibration is the same in behavioral epidemiology.
7. Dr. Rylander's argument that only the highest degree of
annoyance is well correlated to the exposure level does not go
along with his earlier statement that the rank order of annoy-
ance in Tables 1 and 2 are the same, independent of cut-off
points. I also dislike the practice of selecting parts of
the collected data in order to verify hypotheses about dose-re-
sponse relationships. The aim of annoyance measurements must be
to establish a scale of the annoyance felt by a population and
then all the relevant data must be considered.
8. Dr. Rylander's comments on the calibration procedure are
confusing. He says that the reference anchors must be
experienced in the same way by all persons. It that were the
case, there would be no need for calibration. In fact most
comments made by Dr. Rylander strengthen our arguments for cali-
bration procedures and improvements of the methods. Why we
need calibration and what calibration does to the data are ex-
plained in the paper. For further information I refer to text-
books on theory of measurement.
9. Dr. Rylander has doubts about the validity of the improved
method and concludes that our ideas are speculative. By definiti*
the improved measurement procedure must be at least as valid as
the traditional method, because the same basic data are used.
What is added by the improvements is precision of measurement,
provided the less restricting assumptions accompanied by the
»improvements are correct. The assumptions can and must be
checked in every case, regardless of method used, whenever soft
data are the basis for dose-response function.
10. I now realize that it will take time to introduce Thurstoniai
scaling methods as well as calibration procedures into applied
research on annoyance. It is interesting to note that Thurstone
presented his scaling methods as long ago as 1927. Since the
late 50's, when the computer made the methods practical, the
methods have been a valuable tool for quantifiying psychological
variables in the laboratory. It is my hope that also the in-
vestigators in the field will find the methods useful and also
realize the need for calibration procedures and a well founded
measurement theory. This is particularly important for inves-
tigators concerned with soft data such as are involved when
gauging population annoyance.
-------
145
THE ENVIRONMENT AND THE PROTECTION OF HUMAN HEALTH
EDWARD J, BURGER/ JR,
Office of the President's Science Adviser, Science and Technology
Policy Office, National Science Foundation, Washington DC., USA
ABSTRACT
The study of the effects on human health of environmental
agents derives its importance because of the opportunity it
affords for prevention of disease and disability. Environmen-
tal health research provides the knowledge base for governmental
decisions aimed at avoidance of undesirable or hazardous expos-
ures. The list of chemical and physical substances treated in
this manner is a very large one extending from air and water
pollutants to therapeutic drugs.
The case for this activity is a compelling one when consid-
ering the opportunity afforded for betterment of health and the
magnitude and far-reaching nature of the decisions which are
based on this body of science, Xet, the body of knowledge
available for decision-making is poor - both in quality and
quantity and does not correspond to the capacity of science.
This paper discusses some of the characteristics of this
area of scientific investigation and which sets it apart from
others. It examines the consequences of lack of adequate in-
formation for decision-making. Finally, it makes a number of
recommendations for the improvement of the scientific fabric
for regulation.
-------
146
1. Introduction
The fundamental motivation behind the study of the effects on human
health of environmental agents is the potential opportunity for prevention
of disease and disability. Environmental health research provides the
base of knowledge and understanding fundamental to decisions (generally,
government decisions) and standards aimed at avoidance of undesirable
or hazardous exposures. Very large numbers of chemical and physical
environmental substances are treated in this manner. Government
decisions aimed at protection of health are made for both products and
by-products of man's endeavors, including pesticides, ionizing radiation,
air and water pollutants, food additives, etc. Although not always
thought of as "environmental" agents, therapeutic drugs do, in fact,
comprise a portion of man's chemical environment, present some of the
same challenges of scientific understanding, and are treated similarly
as far as patterns of research and methods of control are concerned.
It is interesting to note that, in the United States (and not uniquely
so), most of the environmental legislation and environmental activity
is performed in order to protect human health. The importance of an
adequate base of scientific knowledge of the health effects of environ-
mental agents is deserving of great attention. The extent to which the
opportunity for prevention of environmentally-caused diseases is taken
rests squarely upon the scientific fabric available for government
decisions and upon the ways in which this base of scientific information
is used in decision-making. Yet, to an unhappy extent, from a govern-
mental point of view, the quality and quantity of the knowledge base for
decisions is less than this subject deserves and less than science is
capable of offering.
To a great extent our ability to understand the biological effects of
chemical and physical substances such as pesticides, food additives and
air pollutants has not kept pace with our technological ability to develop
and use new substances. While some environmental agents seem to be
clear causes of ill health, we must readily admit that detailed and
systematic knowledge in this area lags far behind the levels of quantifi-
cation and reliability accessible to contemporary science.
• In 1970, the President's Science Adviser commissioned a broad
review of chemical substances and their relationship to human health.
The report of this review was published and released by the Science
Adviser earlier this year. _!/ The review was intended to be a broad,
yet extensive one covering several aspects of this complex area --
economics, legal questions and administrative procedures. The report
reflects many of these studies and analyses. A large amount of attention
was given to the subject of the base of scientific knowledge available for
decisions.
-------
147
The case for assuring an adequate knowledge base is a compelling
one and the growing breadth of governmental decision processes attests
to the Importance of the research activity. The opportunity for preven-
tion is particularly large and attractive.
In the United States, the crude death rate in 1900 was 17.2 per 1000
population. This rate fell fairly consistently through the first half of
the twentieth century and by I960 was 9. 5 per 1000. (Figure 1). Between
1900 and I960, large relative declines in mortality took place during
infancy and childhood. In absolute terms, however, the declines in
mortality at older ages were also substantial. Various deviations from
this trend can be traced to epidemics of infectious disease which oc-
curred from time to time before the 1940's.
Deaths per 1.000
population
20r
15
10
Age-adjusted rate
1900 1920 1940
SOCKZ: Various reports of the National Office of Vital Statistics.
1950
Figure 1 Crude Death Rates in 1900-1960 per 1000 Population
in the United States.
It is important to examine the changes that have taken place in the
rank ordering of causes of death by disease category. Notable has been
the relative decline in importance of infectious diseases as compared to
chronic degenerative diseases (cancer, diseases of the heart, arterio-
sclerosis, cirrhosis). (Figure 2). In 1900, diseases of the heart caused
eight percent of the deaths in the United States and cancer caused less
than four percent. By 1960, influenza and pneumonia were the only
infectious diseases ranking in the top ten causes of death. (Together,
these amounted to less than four percent in 1969. )• At the same time
-------
148
by I960 diseases of the heart were responsible for 38 percent of all
deaths and cancer for over 15 percent. Most important, these increases
were substantially greater than could be accounted for by the decreases
in deaths from infectious diseases and the increased portion of the pop-
ulation reaching advanced age. Thus, while cancer, for example, has
"taken the place" of other causes of death, the increased age and size
of the population does not entirely explain the increase in cancer deaths
since 1900. J2/ The environmental contribution to human cancer in
general is believed to be possibly quite large. ZJ
It is interesting to note that in the United States, the long-term fall
in overall death rates which characterized the first half of the twentieth
century was interrupted in the decade of the 1950's by a flattening
off. 4/ For males, the shape of the curve for mortality actually changed
sign during the 1960's and passed from negative through zero to positive.
The "excess" mortality (excess over what would have been expected
from previously prevailing trends) was found to have been contributed
principally by eight causal categories of which bronchopulmonary
diseases accounted for over 50 percent. Figure 3 illustrates the major
causes of death which were responsible for "excess" mortality in 1967
compared wit hi 960. The apparent influence of environmental factors
(in a broad sense) on these extra deaths -- notably cigarette smoking
along with other forms of air pollution and alcohol --is very promi-
nent. II, ±1
Finally, a striking series of observations of the past few years
implicating environmental factors in the causation of disease has come
from the series of migratory studies. 5J In these studies, persons who
migrate from one area of the world to take up residence in another
appear to adopt the patterns of incidence of disease (notably chronic
degenerative diseases) and causes of mortality of their new hosts in
place of those characteristic of their former residence. Again, the
implication that it is "something" in the environment which has led to
the observed altered disease and mortality patterns is a very strong one
indeed.
In spite of this strong, compelling case, in spite of the evident op-
portunity for a preventive endeavor, the storehouse of scientific infor-
mation available for judgments and decisions concerning environmental
health remains remarkably poorly stocked. The challenge of this task
of research has not been taken up to any extent proportional to the
importance of the issue nor commensurate with the capacity for scientific
investigation. This marked disparity between the potential for scientific
understanding and the dedication which this subject has so far attracted
prompted one of the speakers at a forum on the Health Effects of Air
Pollutants at the U.S. National Academy of Sciences to remark recently:
"We [scientists] owe someone an apology. How is
it we could arrive at this point with so little understanding?" _6/
-------
149
Figure 2: Leading Causes of Death by Disease Category in the
United States.
Rank
Cause of death
1900
(All Causes)
1 Pneumonia and influenza
2 Tuberculosis (all forms)
3 Gastritis, etc.
4 Diseases of the heart
5 Vascular lesions affecting the CNS
6 Chronic nephritis
7 All accidents*
8 Malignant neoplasms (cancer)
9 Certain diseases of early infancy
10 Diphtheria
(TOTAL)
Deaths per 1 Percent of 1
100,000 pop. | all deaths |
(1,719) (100)
202.2 11.8
194.4 11.3
142.7 8.3
137.4 8.0
106.9 6.2
81.0 4.7
72.3 4.2
64.0 3-7
62.6 3.6
40.3 2.3
(64*)
*Violence would add 1.4#; horse, .vehicle and railroad
accidents provide 0.8%.
I960
(All causes)
1 Diseases of the heart
2 Malignant neoplasms (cancer)
3 Vascular lesions affecting the CNS
4 All accidents*
5 Certain diseases of early infancy
6 Pneumonia and influenza
7 General arteriosclerosis
8 Diabetes mellitus
9 Congenital malformations
10 Cirrhosis of the liver
(946) (100)
366.4 38.7
147.4 15.6
107.3 11.3
51.9 5.5
37.0 3.9
36.0 3.5
20.3 2.1
17.1 1.8
12.0 1.3
11.2 1.2
(TOTAL) (8536)
"Violence would add 1.5$; motor vehicle accidents
provide 2.3$; railroad accidents provide less than 0.1$.
-------
150
Figure 3 Major Causes of Death in 196? Compared to I960 in
the United States.
Principal contributors to rise in mortality in 196? compared to 1960
for white males. Diseases and conditions shown are those for which
the number of extra deaths comprise more than 10 per cent of the
total deaths from that cause.
Percentage Increase in
Cause of Death Deaths 196? versus I960
Carcinoma of the lung 22.9
Carcinoma, other, unspecified 1^.2
Circulatory, other 1?.0
Bronchitis 39.3
Bronchopneumonia, other 39***
Cirrhosis 15.3
Motor vehicle accidents ,^.6
Homicide 30.1
2. What Characterizes Present Knowledge ior Environmental Health
Decision-Making?
It seems useful to review some of the important characteristics
which mark this area of science and scientific investigation. We have
already noted that the totality of the investment in this field has been
less than its importance would dictate and that its fruits have been
accordingly scarce. As a reflection, several environmental standards
(such as air quality and emission standards) have been based on rela-
tively little scientific information.
The science of environmental health has at times been described as
a frontier science. This title has been used to point up the frequency
with which the results of preliminary scientific investigations have been
taken directly from the laboratory and reflected in judgments and de-
cisions regarding the environment. This process necessarily avoids
(or evades) the avenues of interpretation of experimental results by the
peers of the investigator, publication in scientific journals, confirmation
through further experimentation and, therefore, the several means for
maturation of investigative findings traditional to most of science. The
consequences of preliminary, frontier, scientific evidence indecision-
making for environmental health may, in the long run, be very serious
and may not necessarily guarantee betterment of health.
A measurable amount of the scientific investigation in behalf of
-------
151
environmental health has been of a low order of sophistication. The
term, testing, has been used at times to characterize some of the in-
vestigation -- implying a process of accumulation of relatively gross
observations reflective not of a particular hypothesis or matter of
scientific insight but done more or less at random according to the
spirit, the more data the better. TJ This phenomenon has come about
in part because of certain government policies which have encouraged
such investigations. They may be laid in part, also, to the fact that
this area of science has not always engaged the best professional and
scientific minds.
Much of the scientific investigation for environmental health is
performed only with great difficulty (still further limiting the number
of willing and able investigators). Studies involving human subjects,
for example, clearly essential for certain kinds of insight, are extra-
ordinarily time consuming, are difficult logistically, and generally
produce statistical associations rather than establish cause and effect
relationships. Similarly, some investigations are aimed at an under-
standing of the low-dose and of dose-response relationship. These
investigations, by definition, deal with low-probability events and
therefore require very large experimental samples in order to assure
statistical validity.
Finally, some mention should be made of the difficulties in matching
the products of scientific research to the public expectations and to the
requirements reflected in various legal instruments. For example,
the establishment of environmental standards implies in many cases
the identification of effective thresholds of effects. Major problems
arise when multiple biological effects occur, when several organ
systems are involved, and when effective thresholds cannot be estab-
lished (or, in fact, do not exist). Further, the establishment of
relationships between hazardous exposures and human subjects implies
detailed or, at least, statistical knowledge of biological variability and
knowledge as to how that variability is altered by diseases to give a
pattern of known susceptibilities. This complex question of identification
of susceptible members of the population and appreciation of suscepti-
bility in reasonably detailed biological terms is one to which some dis-
cussion has been devoted recently although there remain large numbers
of unanswered questions. 8/ Again, the degree of biological understand-
ing presently available is far less than is implied by the present-day
political and legal instruments. The science and its interpretation are
generally found to be much more complex than is implied by most legal
instruments and standard-setting procedures.
3. Consequences of an Inadequate Knowledge Case
The first and foremost consequence of an inadequate fund of
scientific information concerning environmental health is missed oppor-
tunities for prevention of ill health. This, of course, is a serious con-
-------
152
sequence since protection of human health is the rationale for environ-
mental control and regulation. Thus, exposures to methyl mercury
might have been avoided if the process of microbiological methylation
of mercury had been recognized and if the unusual biological properties
of alkyl mercury had been appreciated. Similarly, several years'
exposures of occupational groups to radon-filled atmospheres in uranium
mines and to coal dust occurred before the biological implications of
these exposures were appreciated.
A second consequence is that of inappropriate decisions which are
occasionally offered in the place of incomplete information. This is
often an unfortunate consequence regardless of how well-intentioned
the decisions may have been. Environmental decisions characteris-
tically imply large expenditures of money and have far-reaching impacts
(as, of course, they are designed to do). Because they provoke large
rearrangements in our lives, however, they require and deserve the
best possible foundation of information.
Decisions made in behalf of environmental integrity or improvement
are typically definitive choices, or at least they are often perceived as
such by lawmakers and by members of the public. Accordingly, once a
decision has been taken, further research is often discouraged or sus-
pended. There are, for example, few continuing efforts to determine
the human health effects of DDT. Clearly, therefore, if decisions and
rational actions are to be founded in scientific truth, we are at some
peril if actions are taken on half-way knowledge and there is no promise
of better information and possibly later confirmation or revision.
A striking (yet not well recognized) consequence of inadequate
knowledge in recent years has been the result of trading inadvertently
one environmental hazard for another. The removal of tripolyphosphate
from among the ingredients of washing detergents produced the unex-
pected and unfortunate consequence of compromising the flame-proofing
properties of chemicals added to clothing materials. Polychlorinated
biphenyls, chosen as fluids for heat exchangers specifically because of
their high level of stability at high temperatures, underwent certain
restrictions. These restrictions may lead to an increase in the number
of accidental industrial fires. More recently, the election of catalytic
* exhaust devices on automobiles in order to reduce certain automotive
emissions may lead to an unexpected increase in the outpouring of other
substances (sulfates and sulfuric acid mists) whose implications for
human health are expected to be important.
Finally, and perhaps most important, the lack of sound scientific
data for decision-making in the environment typically leads the political
process to be unusually conservative (in the direction of protection of
health). The consequent balancing of considerations and judgments
becomes almost more difficult or impossible in the face of an implied
threat than in the face of a recognized and documented hazard.
-------
153
It is with these considerations in mind that the President's Science
Advisory Committee urged that:
"Where knowledge is so inadequate as to make the
reality of a possible threat quite tenuous, the proper
response is to seek more knowledge, not either to take
drastic action or to do nothing. " !_/
4. What is Needed
The task of gaining new knowledge for this area of public decisions
is an enormous one to the extent that it has appeared overwhelming to
some who are charged with establishing its priorities. Yet, however
acutely that message is felt within the scientific community, it is not
well appreciated by others. There is need for biological insight across
a broad front for environmental substances. At least as important,
however, is a strong need for an increase in the level of sophistication
of the research performed. Much decision-making reflects the results
of gross "testing" -- the fulfillment of "standard" or "established"
procedures. By contrast, there is an impelling need to insert a new
but clearly attainable degree of sophistication into the research per-
formed for the environment and health. Investigations should reflect
specific hypotheses and scientific insight and should aim at uncovering
an understanding of biological mechanisms of environmentally-related
diseases.
In the United States we have come to recognize the need for more
new knowledge in the areas intermediate between basic biology on the
one hand and research closely related to immediate regulatory problems
on the other. This has dictated a policy of encouraging a variety of
various institutions serving this area.
Human studies represent a separate challenge. Clearly, some
types of information can only be gathered from human experience.
Commitments to epidemiological research and other types of human
investigations imply and require long-term dedication and continuity.
There is a strong need to engage in forecasting --to look around
the corner in order to anticipate problem areas and to set priorities
for research. This subject has attracted much attention in recent
years. However, there are as yet relatively few tangible results.
Attention has recently been afforded a methodology to forecast new and
forthcoming technological events, new materials, and new uses of
materials, specifically in order to point directions for biological re-
search. In addition, a project in its early stages in the National
Institutes of Health aims at ranking chemical substances as potential
hazards according to the amounts produced, patterns of use, and dis-
tribution and modes of disposal.
Finally, there is a clear need for a reconciliation between the
-------
154
traditional scientific conservativeness and deliberateness in judgment
and the need for interpretation of scientific information for public and
social understanding. The public is bound to pose broad questions on
what the scientific data mean for the public's health. Since there is a
technical or scientific base for their questions, they deserve a sound
scientific judgment. Clearly, a regulatory agency of government needs
to react appropriately and soon in the face of new scientific findings,
clearly implicating a hazard to human health. Few would argue with
the judgments taken in behalf of thalidomide. Yet, there is perhaps
an equally strong need to preserve the deliberate quality of review and
interpretation in order to assure the quality of information. Because
of the scientific and technical nature of these decisions, scientists must
be engaged actively in the process of interpretation.
References
1 Chemicals and Health, a Report of the Panel on Chemicals and
Health of the President's Science Advisory Committee, Science
and Technology Policy Office, National Science Foundation,
(September 1973).
2 Cancer Rates and Risks, U.S. Public Health Science Publication
No. 1148, Department of Health, Education, and Welfare,
Washington, D. C.
3 Man's Health and the Environment. Some Research Needs,
National Institutes of Health, Department of Health. Education,
and Welfare, Washington, D. C. , (1970).
4 Leading Components of Upturn in Mortality for Man, United States
1952-1967, Washington, D. C. , Department of Health, Education,
and Welfare, National Center for Health Statistics, Publication
No. (HSM), 72-1008, Series 20, Number 11, (1971).
5 BUELL, P., DUNN, J. E. , Cancer. 18, 565, (1965).
6 NELSON, N. , Summary Remarks, Academy Forum on the Health
Effects of Air Pollutants. National Academy of Sciences.
Washington, D. C. . (February 7, 1971).
7 GOLBERG, L., Safety of Environmental Chemicals, The Need and
the Challenge, Presented before the Panel on Chemicals and Health
of the President's Science Advisory Committee, Washington, D. C. ,
(February 7, 1971).
8 FINKLEA, J. , Conceptual Basis for Establishing Standards,
Remarks presented to the Academy Forum on Air Pollutants,
National Academy of Sciences. Washington, D. C. , (October 1973).
-------
155
DISCUSSION
GOLDSMITH (U.S.A.)
h« e °r; Bu£ger h*s raised a set of questions about the store
But L°Li~°^atl0Veedf f°r health Flection decisions.
But to press the analogy further we must obtain a better inven-
tory of what the store house contains and what of valuJ is there
Prof. Holland for example noted the value of vital statistics
But now we have permitted the already available health statis-
tical resources to be of less value than it should be by not
using modern data handling methods. The result is that we
lack the ability to detect and respond to trends and gradients
in morbidity and mortality which may be affected by environmen-
I note comments in your paper on the poor quality and inad-
equacy of environmental health research. I feel obliged to
respond that not all the faults are those of environmental health
scientists whose work has had such an unstable level and pattern
of support. For example work of the quality and scope of that
for example, of Dr. Waller, cannot be undertaken. In 1967 the'
U.S. Department of Commerce's Technical Advisory Committee Panel
on the Electrically Powered Automobile noted the difficulty of
sustaining an adequate programme of research on health effects
of motor vehicle exhaust with the pattern of research support
suitable for short-term problems. Yet the situation today is
as discouraging as it was then. The faults lie to some extent
with the general scientific community which has been satisfied
with short-term research support, with an inadequately responsive
research policy, and to a small extent only with the environ-
mental health scientists. We need and deserve a programme of
stable, adequately supported, well planned research among coun-
tries and including intergovernmental agencies as well.
BURGER (U.S.A.)
The purpose of my paper was mainly to offer some observa-
tions on the state of affairs relating to Governmental decisions
for the environment and on the knowledge available for those
decisions. Your comment about the importance of continuity of
research in this area is, of course, absolutely correct. It is
one to which I can only subscribe enthusiastically and, in fact,
I alluded to it myself in my presentation. I applaud also your
own efforts to learn how to glean further clues relating to en-
vironment and health by better exploiting data which already
exist.
However, I must insist that what I observe as a mismatch of
the capacity of science and decisions for the environment is
symptomatic of additional factors and will persist unless some
fundamental questions are settled. A major problem seems to be
-------
156
one of having permitted health to act as a surrogate for other
desirable features in the environment - thus forcing political
decision-makers at times to search desperately among the health
sciences for arguments for action. A second is the short period
of time considered by many as permitted for coming to decisions.
This factor alone discourages long-term investments in background
research. I would like to disabuse this audience of the idea
that decisions must be made in every case rather than waiting for
adequate information. Here, I quoted from the report of the
PSAC Panel on Chemicals and Health on this very point.
A major problem which has inhibited adequate long-term in-
vestments in environmental health research and, in some cases,
appropriate sophistication of research, has been a relative lack
of understanding or concern among members of the environmental
community for the importance of good scientific endeavor and
information. This has been a serious factor from which many
other problems seem to have been derived.
BERLIN (Sweden)
I appreciate your pointing out that there is a lack of
knowledge about important areas in connection with environmental
control. What is your advice to the administration? Should we
not allow further change of our environment until enough knowledge
is obtained or should we avoid interfering in the change of our
environment until enough knowledge is present?
BURGER (U.S.A.)
I cannot, of course, sustain an argument which says that
no action towards betterment of the environment should be taken
until we are fully equipped with all of the information. How-
ever, I do join many others in being very uncomfortable at times
over very large decisions taken on the basis of almost no infor-
mation. By the same token, it is not true, I believe, that a
quest for a decision must inevitably and immediately be met by
a decision to act. A decision at times to temporize while
awaiting additional information may or will usually serve the
public's best interest.
WASSERMANN (Israel)
Today's sessions are encompassed between the doubts of
Dr. Peterson and Dr. Burger with regard to the concern for the
validity of research data before taking a decision.
It is rather necessary to consider some valuable epidemio-
logical resources which are not enough used at present. I
refer pro domo as an occupational health professional to the
need for giving more attention to the amount of scientific in-
vestigation which could be carried out and obtained from the
investigation and data in the industrial plants.
-------
157
BURGER (U.S.A.)
I can only concur with Dr. Wassermann's recommendation in
favor of exploitation of experience from the occupational en-
vironment in considering the general community environment.
As some other participants in this symposium have indicated, in-
formation concerning occupational and industrial health experi-
ence is not always available in the published literature.
Neither is this experience always systematically perused when
available when considering prospective human epidemiological
studies.
RECHT (C.E.C.)
The problems raised by Dr. Burger are extremely important
and point clearly to the relationships between the scientific
world, public opinion and the authorities responsible for deci-
sion making.
Certainly, with the use of the mass media and the interest
shown by public opinion in environmental problems means that
public opinion has easier access to scientific information, but
does not mean that these data are easier to interpret as a result,
Quite the opposite, in fact; in the United States a few years
ago we had the example of the nuclear controversy, when the
simple adoption by a few scientists of a position different from
that of the scientific majority meant that doubts were raised as
to the whole subject of nuclear expansion and that biased and
even false interpretations of scientific data were taken as a
basis for emotional attitudes.
A closer relationship must thus be established between sci-
ence and politics, and scientists must assess the possible psy-
chological effects of information which they impart sometimes
incautiously and often in all good faith. It is also certain
that this Symposium will fulfil the need for arriving at the
widest possible concensus of scientific opinion on a number of
methods, concepts and doctrines on which any real initiatives
to reduce environmental pollution and annoyances will be based.
BURGER (U.S.A.)
I spoke to Dr. Recht's point concerning the relationship
between scientists and public information. In my experience,
it is an exceedingly important consideration in practice and
has, at times, gone far toward influencing individual decisions.
Responsibilities for active yet appropriate participation appear
to lie with several parties to this question - the scientist,
the press, and government agencies. (For a fuller discussion,
I would refer the reader to the report of the Panel on Chemicals
& Health of the President's Science Advisory Committee referred
to in my talk).
-------
POPULATIONSUNTERSUCHUNGEN
POPULATION STUDIES
ETUDES SUR LA POPULATION
STUDI SULLA POPOLAZIONE
BEVOLKINGSSTUDIES
(Continued)
Vorsitzender - Chairman - President - President - Voorzitter
M. COLOMBINI (Italia)
-------
161
THE STRENGTHS AND WEAKNESSES OF POPULATION
STUDIES IN ASSESSING ENVIRONMENTAL HEALTH EFFECTS
DAVID L, JACKSON AND VAUN A, NEWILL
Special Assistants to the Administrator, Environmental Protec-
tion Agency, Washington DC, USA
ABSTRACT
The basic purpose of the environmental regulatory process
and the impact of the political and social framework on the
design of health effects studies is briefly reviewed. The
strengths and weaknesses of clinical and toxicological studies
are outlined, and a more detailed exposition for population
studies is presented. The problems of response indicators,
exposure quantification, causality of association, combined
pollutant effects, and the threshold concept are discussed.
Four areas for further development in population studies are
•presented.
-------
162
It is often both helpful and revealing to step back and attempt to
achieve a fresh perspective by identifying the primary purposes of work
in the environmental health sciences. This is particularly true for those
whose work is related in the last analysis to environmental regulatory
decision-making.. The primary purpose of the complex network of legisla-
tion, as well as federal and state regulations on environmental pro-
tection, in the United States and in many other nations has been and con-
tinues to be the protection of public health. Initial programs to improve
public sanitation, when coupled with the advances in acute infectious
disease control over the past 75 years, have significantly lengthened the
life expectancy of the average individual in many nations. As an in-
creasing percentage of the population has reached the seventh decade and
beyond, the possibilities of developing chronic degenerative diseases
has increased. These chronic diseases, most frequently seen in the older
age groups, may be caused or exacerbated by chronic exposures to relative-
ly low levels of many different environmental contaminants. In the past
decade, regulatory d^cisiop^rak* "g has l^T"** much more «Trny»l*»x and the
legislative framework much more extensive. The primary objective, how-
ever, remains the protection of public health. Recent legislative in-
itiatives have introduced the broader concepts of protection of general
welfare, consideration of ecological impact, and improvement in the
quality of life as secondary goals for environmental regulations. The
role of economic and social factors in regulatory decision^raiH *yj raises
heated controversy amongst both scientists and administrative/political
decision makers. In sane activities in the U.S.A., e.g., in determination
of primary ambient air quality standards, regulatory decision-making is
constrained by law to consider only the protection of public health. In
other areas, not only is public health a major factor, but ecological,
economic and social factors play a significant role in the ultimate
decision-making as well.
The central issue addressed in this paper is how health science
data, par-t-iniiar'iy population studies, can be most effectively \isgcl in
regulatory decision-making. In this context, all the steps in decision
making must be considered, frrm pnnfcigiq i Aan-H f•jfcvH^in and assessment
to alternative control strategies, implementation and enforcement of a
control/action program, and ultimately to an assessment of the efficacy
of the action program. How this can be done depends, to a large extent,
en what societal goals are chosen by political decision makers, as they
construct the framework in which the issues will be evaluated. That is,
specifically, the level of environmental protection to be achieved by
any regulatory action must be predicated on a clearly stated choice be-
tween two different approaches. These designated levels of protection
reflect two philosophical viewpoints. The first is the fia^ra] zero
tolerance or no risk approach, which states basically that when dealing
with any toxic substance in the environment, no health risk is acceptable.
Therefore, when zero tolerance is the basic premise, it leads to zero
exposure programs. This approach has been adopted by the U.S. Congress
for dealing with carcinogenic materials in foods and drugs under the
Delaney Clause. A slightly less stringent corollary is the threshold or
approach. This concept is the basis
-------
163
for the Clean Air Act provisions establishing primary ambient air qua-
lity standards. In this case, exposure limits are set at levels where no
adverse health effect can be measured.
More recently another conceptual framework has been advanced—that
of socially-acceptable risk, This approach has been used to determine
sane industrial exposure standards and most recently was used as the
basis for the U.S. Environmental Protection Agency's (EPA's) proposed
benzidine efficient guideline. There are many difficult questions in-
herent in applying this approach, both from statistical/mathematical
point of view and from moral/ethical considerations. This concept re-
quires two decisions. First, one must make certain assumptions as he
extrapolates from known dose/response data to the low dose/low risk
portion of the dose/response curve. There are many models that have been
proposed for this extrapolation. Ttoo of the more connonly employed are
the Mantel-Bryant probit model (1) and the linear "one hit" model (2).
Thus these abstract models have dollar and cents reality when used in
the decision-making process of a regulatory agency. The plenary
discussion will develop in more detail some of the trade-offs
involved in balancing health effects with economic, social and
energy related factors as they relate to the assessment of a socially-
acceptable risk and to a determination of regulatory policy.
What tools and mechanisms are available to generate the health-
related data necessary to formulate rational, efficient and effective
environmental regulatory control strategies?
It is helpful to view environmental regulations as a four-step
process (Fig. I). [3]
SIMPLIFIED SCHEMATIC OF THE
ENVIRONMENTAL CONTROL PROCESS
ENVIRONMENTAL PROBLEM IDENTIFICATION AND
ASSESSMENT - SET PRIORITIES
For each problem in priority
order
EVALUATE ALTERNATIVE CONTROL STRATEGIES-CHOOSE
STRATEGY AND TIME FRAME FOR ACTION
I
IMPLEMENT ACTION PROGRAM
I
ASSESS ACTION PROGRAM-FEEDBACK OF PROBLEMS
Fig. 1. Simplified schematic of the environmental control
process.
-------
164
1. Problem identification and assessment.
2. Developing alternative control strategies.
3. Inplenentation and enforcement of an action program.
4. Action program assessment.
In all steps except implementation and enforcement, it is crucial to have
an adequate scientific data base as the foundation for the politioal-
-egulatory decision process. Problem identification often requires much
less data than that required to develop alternative strategies and select
the most appropriate regulatory action. One cannot assesTthe inpact of
any actLon program unless he has adequate base line data to assess the
situation prior to the initiation of that program, plus wen-designed
Llow-up studies to assess the "benefit" of the progran.
Health scientists have three major weapons in their arsenal
clinical studies, toxicology and epidemiology or population studies
2).
Epidemiology
Clinical
Studies
Toxicology
Overlap: identical biological endpoints
Fig. 2. Methods to demonstrate biological response to
pollutants.
Clinical research studies can be used to gather human data on
either normal or diseased individuals regarding the absorption, metabo-
lism, and excretion of pollutants. ihis is particularly useful in the
evaluation and in-depth analysis of humans accidently exposed to high
levels of pollutants where new parameters of exposure and Indicators of
response can be identified. The advantages of clinical studies are: (a)
that there is a controlled or at least relatively accurately-quantifiable
exposure to a pollutant, ilrus in experimental protocol studies one
can obtain improved dose measurements; (b) often one can use each person
as his own internal control (before, during and after exposure), thus
effectively controlling many covariates; (c) vulnerable subjects may be
-------
165
included by design to increase the sensitivity for low-dose responses;
(d) cause-effect relationships are more easily deciphered; (e) cross-
species extrapolation of the results is not necessary. This maximizes
the usefulness of these data in setting health-related pollutant
standards. There are three major limitations to clinical studies:
1. The exposure profile is artificial.
2. No long-term exposures are possible, so only acute effects can
be determined.
3. There is a small but real hazard to the exposed subject.
Toxicological studies are a second tool, which use many response
systems, including whole animal, tissue culture and biochemical systems.
The advantages of toxicology studies are: (a) maximal dose-response
data can be obtained, though this is often incomplete at the low dose
portion of the curve; (b) data acquisition is relatively rapid; (c)
cause-effect relationships are more easily established; (d) studies on
the mechanism of response, such as pollutant absorption, distribution,
metabolism and excretion, can be performed. However, toxicological
studies often lack appropriate human disease models in laboratory animals,
thus the effects of the controlled and well-characterized-pollutant expo-
sure are evaluated in most cases only in healthy animals. This limits
the sensitivity of this tool in evaluating subtle but possibly important
changes in certain relatively cannon human disease states. Only recent-
ly has this tool been used to identify and evaluate delayed responses
or chronic cumulative effects of exposure. It is also very difficult
to duplicate in the laboratory the exact exposure profile for a given
pollutant in a ccmnunity. The generic problem of extrapolating data
from laboratory animals to the human subject remains particularly
difficult, especially when one is attempting to use this data to esti-
mate a "threshold" level for adverse human response.
Epidemiology or population studies, the main focus of our consid-
erations here today, offers certain unique advantages. It provides
studies on the effects of pollution exposure on populations in real-
life settings. Thus the exposure profile is natural, reflecting the
real-world situation. There is no problem with extrapolation of the
results from one species to another. This tool enables the environ-
mental health scientist to study the most vulnerable population groups,
not just the healthy population. It allows evaluation of the effects
of current exposure levels, crucial for meaningful and timely regula-
tory decision-making, and of the effects of exposure to chronic low
levels of various environmental pollutants.
There are certain key limitations to epidemiolcgical studies,
which must be kept in mind when assessing how most effectively to
use population studies in regulatory decision-making. The precise
quantification of the exposure profile for each individual is very
difficult, if not impossible, to determine. A corollary of this problem
is that often one cannot obtain any dose-response curves. IJpidemiologi-
cal studies by their very nature must deal with a multiplicity of co-
variates, often making isolation of the single variable under study
virtually impossible*
-------
166
Unless the pollutant under investigation in an epidemiological
study has already been introduced into the environment for a long
enough period of time to cause a measurable effect, such studies are
often unrewarding and nay lead to an incorrect epidemiological "clean
bill of health" for a recently-introduced new chemicaVpollutant.
Thus in the study of new chemicals, epidemiology has limited useful-
ness and certainly is of little predictive value. Finally epidemiologi-
cal studies are much less powerful tools than toxicological studies or
clinical studies for establishing causality in the association of a
given exposure and a measured effect.
How then can epidemiological studies be designed to maximize the
strengths, minimize the weaknesses and thus make the study of maximal
utility in regulatory decision-making?
First a more precise understanding of the spectrum of responses
for exposures to any given pollutant must be developed. There are
five haslc levels of response to exposure of any given pollutant (Fig.
3}. [4] At the lowest levels, affecting the largest population groups,
exposure to a pollutant may increase the total body burden of the pol-
Adverse
'Health
Effects
Pathophysiologic
Changes
Physiologic Changes of
Uncertain Significance
Pollutant Burdens
Proportion of Population Affected-
Fig. J. Spectrum of biological response to pollutant
exposure.
-------
167
lutant, without any measurable effect on the health, metabolism, or
physiology of the organism. Exposure to lead with increased body burden
but blood lead levels of less than 25 ugm/100 ml. blood is an example
of this "response1' level. The next level would be exposures which lead
to metabolic or physiological changes ot unknown sigroricance, such as
the increased urinary excretion of delta-aminolevulinic acid associated
with blood lead levels of 40 ugn/100 nd. blood. This blood lead level
is not associated with significant organ dysfunction or any increase in
acute or delayed morbidity. The third level is that exposure which
causes a physiological or metabolic change that is clearly prenorbid,
such as exposures leading to decreased mucociliary clearance in the
respiratory system or changes in pulmonary function tests. Next are
exposures that elicit frank disease, such as initiating an acute asthmatic
attack or causing an increase in acute lower respiratory tract infection.
The final level is that which leads to death. With the multitude of
covariates often associated with increased mortality, it is difficult to
establish exposure to a given pollutant as the sole "cause" of the excess
deaths, the role played by diet, aging, accidental injury, and individual
variability in susceptibility must be considered.
In the vast majority of cases, one deals with a response continuum
and rarely with a discontinuous, all or nothing event. The information
gap between the level of pollutant body burden and that of overt disease
state must be bridged before the role of pollutant exposure in total
community morbidity can be precisely defined. The population studies
are particularly effective in studying those groups most susceptible to
the adverse effects of environmental factors, such as the very young
and the aged, or to evaluate the effect of pollutant exposure on popu-
lations with specific diseases, such as asthma or chronic respiratory
disease.
It is also of great value in assessing the effect of pollutant
exposure during periods when an individual's susceptibility is temporar-
ily impaired, as during pregnancy, the adolescent growth spurt or acute
lower respiratory tract infection. [5] Populations in whom suscepti-
bility is permanently altered, such as alpha-one antitrypsin deficiency
states or abnormal hemoglobins, can also be studied.
Disease states are the end product of an intricate and complex
causal web* [6] Each individual strand must be isolated and evaluated.
But to understand fully the "gestalt" of environmental pollutant
exposure, nutritional status, genetic composition and personal habits,
the strands must be woven together again. Man's natural intuitive con-
cepts of the nature of response may not be entirely accurate. For
example, there is no reason why one could not postulate that exposure to
an environmental pollutant could cause an increased severity of preexist-
ing disease without increasing the overall incidence of that <3i »=*»-«*»- in
this case, classical incidence studies would miss the significant "health
effect."
The horizons of the responses measured must be broadened. In a
large number of the classical environmental pollution population
studies, either death rate, pulmonary disease symptoms or cardiopulmonazy
symptoms have been chosen as the dominant or sole cirteria for judging
adverse health effects. These areas of response may well be the major
patterns for exposure to the cuimoii airborne pollutants, but it is
-------
168
essential that otter parameters of adverse responses must be developed.
Both more precise and sensitive biochemical and physiological indicators
in maixy organ systems are reguired. Recent work with enzyme systems as
response indicators or system dosimeters shows much early promise.
The interface of response and exposure presents its own difficulties
in designing and conducting optimal epidemiological studies. One must
be able to assess accurately the duration of exposure, the concentration
of the pollutant (s) and the rates of change in levels of exposure. He
should know the route of exposure, which is not as simple as it may
appear, e.g., the possibility of swallowing inhaled fibers after they
are cleared by nuoociliary action in the respiratory tract. He must
generate data on the uptake, metabolism, and excretion of the pollutant.
It is important to be ^ble to assess individual variability in response
related to physiological variability, f*i •imatg, occupational history,
and psychosocial factors. This responsiveness is also t-gT'**f^ by
tolerance and adaptation. The physical and chemical properties of a
given pollutant also affect their potential to elicit an adverse effect.
Thus the size, density, viscosity, volatility, solubility, chemical re-
activity and electric charge of the pollutant must be quantitated as
precisely as pog«i.M«*- The more data available about the photochemical
and +h**TTBnrfrtanri evt i transformations which a pollutant undergoes as it moves
through the environment from source to receptor, the more precisely one
can assess the toxic potential of any given pollutant.
The measurement of exposure is fraught with difficulties. [7] The
population under study in the U.S.A. is highly mobile and often transient.
Twenty percent of the U.S. population changes place of residence every
year. On. the average, 1/3 of the population of a U.S. oonnunity has liv-
ed there less than three years. The pollutant mix is also variable,
changing with both time of day and season. For air pollutant exposures,
the activity pattern of an individual during a "normal day" carries
him through areas of potentially markedly differing pollutant mixes
and concentrations. The duration of exposure in each area is different
for different individuals. There are not enough different air quality
monitoring sites in any one area to construct an exact exposure profile,
even if »?"! could agree on a reasonable activity pattern or patterns. In
the U.S. Ccnraunity Health and Qwironmental Surveillance Studies (CHESS),
each area selected for study must meet certain criteria, with no large
point source of pollution, relatively simple flat topography and rela-
tively complete census tract data available. Then a single monitoring
station is constructed near the center of the area. However, for reasons
of accessibility of building sites, often the sampling site is far to one
side of the area. Recently, mobile samplers have been used to check
the uniformity of pollution levels in the area. It is unclear what con-
stitutes the most effective technique for dealing with data from an area
in which the pollution levels are not uniform. The question of exposure
profiles, cannot be resolved with the present data base, as many will
leave the ^r**^ for significant p"aT"?"f^s of -t-jpip each day. This often causes
exposure to very different pollutant levels in these other areas.
One way to avoid seme of the problems with quantifying exposures
is to develop personal exposure meters, analogous to the X-R film badge.
EPA is supporting the development of a personal exposure meter for
sulfates, and it would appear that the technology could be developed
-------
169
for some other pollutants as wall. Such devices would have certain
very important benefits. They would permit validation testing of
activity/exposure models for much larger populations than would wear
the monitors. They would help determine the appropriate placement of
stationary monitoring sites for larger scale studies of similar popu-
lations. They can help assess the health inpact of differing pollutant
exposure levels. However, they suffer from several drawbacks. The
Hisenberg principle of uncertainity introduced by observation may play
a role if behavior is modified by knowledge the individual is being
monitored. Any study using such personal monitors would be expensive,
limiting the sample size. It would require excellent participant
cooperation. Reliability would have to be firmly established, and it
would be difficult to use the data to establish a threshold or to con-
struct an appropriate control strategy.
The alternative to expanded development of personal exposure
meters is expanded use of stationary monitors. This generates data
which can be applied to large numbers of individuals, if one assumes
certain activity patterns. These data are necessary for accurate
assessment of the effects of short-term exposures. Continuous mon-
itoring equipment is also expensive and the complex technology is prone
to breakdowns at the most inopportune times. Too often the monitoring
data from these stations is disassociated in space and time from the
measured health effects. Frequently disease incidence studies in
metropolitan areas are correlated with exposure estimates based on
data from one or at most a few stationary monitoring units. These data
may not be representative of community-wide exposure. The derived
exposure estimate may imply exposures higher than actually occurred.
Cft the other hand, ascribing excess chronic disease to currently-
measured levels of pollutants may underestimate the real long-term
exposure, since air pollution levels in the late 1940's and 1950's
were higher than in 1970-74. Perhaps the weakest link in our attempts
to establish the exposure-response relationship is in obtaining an
accurate quantitative estimate of exposure. To base regulatory decisions
on very precise and extensive health response data, but relatively poor
exposure data, can lead to gross errors in regulatory judgments.
Cnce statisticians have shown an association between an exposure and
a response, how can one best evaluate whether or not there is a causal
relationship as the basis for that association? Hill £8] has developed
a set of criteria to judge whether or not an observed exposure-response
relationship is causal. Although these criteria were developed as guide-
lines for occupational health studies, they can be adapted for application
to general population studies. The major criteria are: (a) consistency
of observation; (b) coherence of data; (c) plausibility of association,
and; (d) strength of association.
Other factors which can aid in the evaluation of the significance
of exposure-response associations include: (a) the presence or absence
of an exposure-response gradient. When there is a stepwise increase in
the frequency or severity of the adverse health effect, this supports a
causal relationship. Linear relationships over an exposure (dose) grad-
ient are difficult to explain by the intervention of other variables.
Such gradients are most easily created in experimental settings, but can
be approached in population studies by investigating the relationship of
response to exposure gradients across geographic areas, differences in
-------
170
length of residence in high exposure areas, and migraticn gradients
constructed from various combinations of childhood and adult exposures
of the same individuals; (b) Intervention: The effect of societal
programs designed to protect the public health from observed adverse
health effects associated with exposure to a given pollutant can be
helpful in establishing causal relationships. If, as desirable environ-
mental ambient levels of the pollutant are achieved, the frequency/
severity of adverse effects diminishes, there is a strong suggestion of
causality. The high cost, both in the public and private sector, of
many of the environmental regulatory programs warrants a vigorous
national program of community health and environmental surveillance to
evaluate with more precision the benefits of these abatement efforts; (c)
Control of covariates; The causal nature of an exposure-response associ-
ation is convincingly identified when, after the effects of known covari-
ates are first displayed, increased disease risk within covariate classes
can be clearly demonstrated in high exposure populations. For example,
in any studies of chronic bronchitis prevalence, smokers and males
show more disease than nonsnokers and females respectively. An air
pollution-chronic bronchitis study should reveal the above smoking-sex
differences in prevalence rates, thereby assuring readers that the study
has internal consistency. If smoking and sex specific groups in high
exposure neighborhoods have excess chronic bronchitis, the hypothesis
that air pollution exposure causes excess chronic bronchitis is con-
siderably more convincing than if the smoking-sex covariates were not
analyzed. Most epidemiologic studies of air pollution require similar
analysis of excess disease risk within covariate categories, with partic-
ular attention given to age, sex, stoking, socioeconomic level and
duration of residence at current location. When covariates are system-
atically analyzed for relationships to the health indicator under study,
residual excesses in disease frequency can be attributed to differences
in pollution exposure with a reasonable degree of confidence.
Such population studies have been classically designed to hold
constant all variables but the one under study, and this has been a
powerful tool. Yet as one looks at more complex exposure profiles,
with pollutants which could act in a synergistic, additive, or ameliora-
tive manner, he must develop and employ more sophisticated multivariate
analyses. Be will require methodology to identify possible interactions
of various pollutants. The relatively limited rfat-a base available in
most cases makes it difficult to complete such an analysis. The data
base must be expanded to include increased data on the combined effects
of mil tip! ff pollutants. The physio-chemical interactions occurring among
different pollutants in the environment often alter the biological
activity of the pollutants, as well as modify the response in the exposed
individual. A recent example of a study of combined synergistic effects
was presented by Bates and Hazucha [9] in 1973. They stated "our most
recent experiments have been concerned with an qdditv*vO. problem,
namely a possible interaction between ozone and S02- The episode of
pollution in Botterdan that occurred two years ago had several puz-
zling features. Although the ozone levels did not reach much beyond
0.2 ppm, and the S02 levels at the same time were about 0.2 ppm also,
there was considerable morbidity, particularly amongst people bicycling
in such an atmosphere. These levels of the individual constituents
seemed to be too low to have caused the considerable symptoms which were
reported. There is evidence that ozone and 902 together affect plants
at lower concentrations than each would individually, and since nowa-
-------
171
days ozone exposure and SO2 exposures may be expected to occur together,
it seemed most important to us to try and study the interaction of the
two." This Figure (Pig. 4} [9] shows the effect on four measures of
pulmonary function of 0.37 ppra of S02 and 0.37 ppm of ozone exposure
independently, and then the enhancement of effects when both gases are
present simultaneously. It is interesting to note that the effect of
i
u
THIS FIGURE SHOWS THE EFFECT ON FOUR MEASURES OF PULMONARY
FUNCTION OF 0.37 PARTS PER MILLION OF S02 AND 0.37 PARTS PER
MILLION OF OZONE INDEPENDENTLY. AND THEN THE ENHANCEMENT OF
EFFECT WHEN BOTH GASES ARE PRESENT SIMULTANEOUSLY1
• 0.37 ppm S02 (N=4)
0.37 ppm 03 |N-3]
0.37 ppm S02+03 (N=4|
100
90
BO
70
60
50
100
90
SO
70
60
,50
1-5
2-0 2-5
2-5 0 0-5
TIME fhrsj
\. BASED ON FIGURE 14, PAGE 534 OF U.S. CONGRESS DOCUMENT NO. 93-15
Fig. 4-. Effect of S0_ and ozone on pulmonary function.
the two gases administered simultaneously is greater than the sum of the
effect of each individually, thus demonstrating a synergistic effect.
Similar synergistic reactions may well explain many other episodes of
pollution-related disease in Europe, Japan, and the U.S.A. Until popu-
lation studies which can evaluate this possibility are conducted, it
will continue to be difficult to select the most appropriate control
level for each individual pollutant to ensure adequate protection of
public health, while avoiding unnecessarily restrictive standards.
A final area of methodological concern is the use of the concept
of "threshold" in epidemiological studies. A biological effect may
not be observed until the exposure (dose) reaches a certain level. This
"threshold" dose is defined as the minimum dose required to produce a
-------
172
detectable effect. Ohe nervous system has been cited as an example or
an "all or none" or "threshold" biological system. However, it is now
veil established that, on a membrane level, much of the response in the
nervous system is a continuum and is not "all or none." The concept
of a threshold dose for a potentially harmful pollutant is important
from both a practical and a theoretical viewpoint. It implies that
there is a "safe" limit or standard which can be specified. However,
as the indicators of response become more sensitive r the "threshold"
may appear to df-CTTfvine. Par many pollutants, there may be no real "safe
dose," [5] while for others existence of a "threshold" may reflect the
development of tolerance to a particular pollutant. This adaptation to
long-term, low-level exposure to a given adverse environment varies within
a population and is determined by an individual's anatomic, physiologic,
and biochemical characteristics [5]. Even if we set aside the consider-
ation of the biological reality of the threshold concept, the statistical
methods employed to approximate this level have their own limitations.
The EPA National Environmental Research Center group has adapted from
Quandt 110] an approach to estimate "threshold" by fitting the data to
two best-fit lines, one with zero slope and one with a positive slope
(Fig. 5) [HI. This "hockey stick" analysis then uses an iterative
STATISTICAL ASPECTS OK SURVEILLANCE
100
20
1 I T
W
V*
tf
CO
tt 40
e
o
°o o -
I I* I I I I
04 8 12 16 ZO 24 28 32
OXIDANT, ppha
Per cent decreased performance versus oxidant level, with fitted "hockey stick"
function.
Fig. 5. Statistical aspects of surveillance.
-------
173
process to reach the least squares combination of the two lines and the
point of inflection is then taken as the threshold. This has been used
in setting the U.S. primary ambient air quality standards, and is particu-
larly helpful if a series of studies generate similar "hockey stick"
thresholds. Hcwever, the inflection point determination may not be
unique and the analysis assumes that the positive slope response curve
is linear. Also, the method is univariate and the only method for
evaluation of multiple covariates is to assume their presence and add a
statistical factor to account for their influence. Finally, the 95 per-
cent confidence limit around the inflection point is often very broad
and occasionally vail encompass "zero" dose. Both the scientific com-
munity and the political decision-makers must continually reassess the
validity of the assumption of applicability of a threshold concept to a
given pollutant-response association. It nay well turn out that some of
these cases will require, in the future, assessment of the socially-
acceptable risk rather than a set "no adverse effect" threshold.
As one looks to the future, what directions will be taken in
population studies to maximize their usefulness in the area of environ-
mental health regulation? It is absolutely necessary that more emphasis
be placed on developing baseline data prior to the implementation of
any environmental regulatory program. If this data is not obtained,
neither the regulatory agencies nor the health scientist community will
be able to assess adequately the effectiveness of regulatory programs
or to convince the people and the political power structure that
further programs are justified and worthwhile. A noteworthy example
of this is the need to design and conduct such population studies
before and after implementation of early projects under the $18 billion
waste water treatment plant construction grant program in the U.S.A.
A second area where increased future emphasis is indicated is in
cooperative programs between population study designers and large health
services delivery systems. There is a large potential data bank that
can be gathered by organizations such as National Health Services in
Europe and Canada and private Health Maintenance Organizations (HMO's)
in the U.S. Seme Union programs, including Rubber Worker's Union in
the U.S., have made health data recording and long-term follow-up
part of their contractual agreements with their employers. Environ-
mental health scientists should develop similar data banks. This source
has been much more effectively utilized by our colleagues in Europe
and Canada, but EPA is presently evaluating a feasibility study on
cooperative programs with several of the larger HMO's. The major
challenge in this area is to develop an infornational frame-work so
that these long-range data sources can efficiently collect those data
that will be most useful in the future.
Future epidemiological studies must also maximize the temporal
and spatial concordance of exposure monitoring systems and health data.
The coordination between the Regional Air Pollution System (RAPS) and
the Ocmnunity Health and Envirorroental Surveillance SystemfCHESS) in
St. Louis is a step in this direction. But on a broader scale, measure-
ment methodology and the health information should be standardized on a
regional, national, and eventually international basis. This is parti-
cularly true as the distant effects, both regional and global, of air
pollution are more fully understood.
-------
174
I would like to place special gnfhag-icg on tte need to generate
health infarmntion fran the rT*Tlatiqn groups most affected by environ-
mental pollution. The central-city Inhabitants of many cities in the
world are exposed to the highest levels of pollution, but these popu-
lations have been only infrequently included in epidaniological studies.
It is obvious that studies in these areas are more expensive and far
more difficult than studies dealing with suburban igi«Mie class, relative-
ly well-educated population groups. Most of the classic subjective
tools for collecting ep"« A=mlologtral data (diaries; long, frrplex data
forms) are not effective in the inner city setting. Until the methodo-
logy to gather reliable epidemiological data from inner city popula-
tions is dgWlcgS, a aignifirsmf void in the data base required for
optimal regulatory decision-making will exist. Cne avenue for improved
data quality might be to increase reliance on objective tests, such as
pulmonary function tests. The ccnnunity experience of many university
hospitals in center city areas is a potential source of expertise for
optimizing the quality of subjective dgt^ from the diverse
groups in these areas. The goals of such studies must be translated
into clearly-perceived positive motivational factors, relevant for
the individuals included in the study.
In sunnation, this paper has attempted to focus on some of the
problems which face environmental health scientists and how population
studies can be designed to maximize their usefulness and impact in the
onmental regulatory decision-making process. It also suggests
a few of the future paths which epidemiologists must take the le*cl in,
as they strive to solve these methodological and informational problems.
It will have served its purpose if it generates an exchange of ideas and
perceptions on the issues raised or on the many issues not developed due
to time constraints.
BIBLIOGRAPHY
1 MANTEL, N., and HORN, W.R., "Safety testing of carcinogenic agents,"
J. Natl. Cancer Inst. 27, 455 (1961).
2 GROSS, M.A., KLTZHUUH, O.G., and MANTEL, N., "Evaluation of safety
for food additives: an illustration involving the influence of
methyl salicylate on rat reproduction," Biometrics 26, 181 (1970).
3 NBHILL, V.A., "The administrative need for environmental health
research," presented at the Annual Conference of the National Qivircn-
mental Health Association, New York City (1972).
-------
175
4 NEWTLL, V.A., "Regulatory decision making: ate scientist's role,"
presented at the National Academy of Science Symposiun, "Statistic**
and Man in the Environment, " Washington, D.C. (1974) .
5 Message from the President of the United States, "The report of the
Department of Health, Education and Welfare and the Environmental
Protection Agency on the health effects of environmental pollution,"
House of Representatives Document ND. 92-241 (1972) .
6 MacMftBON, B., PUGH, T.F., and IPSEN, J., Epidemiologic Methods, pp.
18-21, Little, Brown and Company, Boston, Massachusetts (1960).
7 SHY, C.M., "Health intelligence for air quality standards," presented
at the Meeting of the President's Air Quality Advisory Board. St. Louis,
Missouri (1973).
8 HILL, A.B. , "The environment and disease: Association or causation,
Proc. Royal Soc. Med. , 58:295 (1965).
9 BOXES, D.V. and HAZUCHA, M. , "The short-term effects of ozone on
the hunan lung," proceedings of the Conference on Health Effects of
Air Pollutants, HAS, October, 1973, pp. 513-540. U.S. Congress
Docunent Serial No. 93-15 (1973) .
10 GUAMJT, R.E. , "The estimation of the parameters of a i-ipear regression
system obeying two separate regimes," J. Amer. Statis. Soc., 53, 873
(1958) . -- -
11 NELSON, W.C., BASSELBLRD, V., and G.R. LOWKM3RE, "Statistical aspects
of a connunity health and environmental surveillance system," Proc.
6th Berkeley Symp. on Mathenpt^^al Statistics and Pir*»abJ'Lity, pp.
125-133, Berkeley, California (1970).
DISCUSSION
ZIELHUIS (Netherlands)
Dr. Jackson distinguished between "public health" and
"quality of life". However, health is a unified concept,
although with different aspects: - medical
- behavioural
- ecological
each of these studied by different disciplines.
There is no distinction between human health and quality
of life, but only a sequence of priorities in regulatory mea-
sures of prevention.
The priority of prevention should not be determined by
scientists (or physicians), but by the population itself. Very
probably, the population has different priorities than, for
-------
176
example, medical science. Priority setting is a political
decision, and not the outcome of scientific discussion. Res-
tricting health to medical aspects, already sets a priority,
maybe not in concordance with priorities as set by the popula-
tion.
JACKSON (U.S.A.)
Dr. Zielhuis raises some very important points. In my
full text, the separation of "Classical" public health and the
"quality of life" reflected the chronological sequence of the
introduction of these concepts into the environmental protection
legislative framework. In my own view, the concept of the
quality of life is an integration of many factors, including the
absence of disease, behavioral, socio-economic and ecological
considerations. For a society to maximize this integral re-
quires prioritization of these factors. These societal deci-
sions are clearly beyond the sole province of the technique
experts. I have tried to distinguish between the role of the
scientific community and the political decision makers. The
scientific community must identify and assess problem areas,
generate the data, evaluate the data, and present an objective
interpretation of the data to the political decision makers and
the general public. This does not mean a scientist should not
participate in public policy debate, but he must do some reali-
zing he represents only a small segment of the population. For
his position to be accepted by the general public, it must be
made understandable and cogent to the decision makers and the
public. It is clear the U.S. Congress reflected an inherent
priority decision on the major role for protection of the popu-
lation from disease, when it passed the Clean Air Act in 1970.
This policy decision was taken by the elected representatives
of the people, not by an elite corps of scientists. In the
past four years, the areas of concern have been broadened, as
detailed in the paper, reflecting a growing awareness that all
factors must be considered in most decision making processes.
However, the protection of the population's "medical health"
remains a cornerstone of the environmental regulatory process,
though it should no longer be considered the entire edifice,
eliminating consideration of other important factors.
MAGE (Denmark)
Is not figure 5 faultily constructed? since the fitted
functions must go to zero% decreased performance at zero pphm
oxidant by definition. Is not the correct form, as dictated
by this physical principle a horizontal line from the origin
(0,0) followed by a linear section with a positive slope?
-------
177
Correct
100
100
.-' II
Incorrect because it
must go to zero
100
100
JACKSON (U.S.A.)
Dr. Mage has difficulty with the "hockey stick" analysis.
In this, he has many companions. In the full text of my paper,
many of the limitations of the threshold concept in general and
the "hockey stick" analysis in particular are noted. However,
when multiple studies generate many similar "hockey stick"
thresholds, the value so determined can be useful in regulatory
decision making. Figure 5 was chosen from a paper by Nelson
et al to represent a graphic example of this technique. The
"per cent decreased performance" refers; to the performance of
a high school cross-country running team, which ran the same
course each week, timing each individual runner. Each team
member's time was compared to his time from the previous week.
The percentage of the team that ran slower on a given week was
plotted against ambient oxidant levle measured at the race
course. Approximately 16% of the team ran slower than the
previous week at all measured oxidant levels between 4 and 12
pphm. Only when oxidant level rose above 12 pphm did team
performance show any additional decrement. Two points should
be made. Although it is true there is no data on team perfor-
mance in oxidant levels between zero and 4 pphm, the constancy
of the 16% decreased performance, at decreasing oxidant levels
from 12 to 4 pphm, would make a fall of the percentage decreased
performance to zero at zero oxidant concentration quite unlikely.
The myriad of other random causes for a team member running
slower than the previous week other than oxidant levels makes
this data intuitively plausible. If one subtracts the 16% as
a baseline and plots change in decreased performance percentage
above this baseline vs. oxidant level, one generates a "zero
effect at zero level" curve that Dr. Mage prefers.
-------
178
MOKEMATKENGUEMBA (Central African Republic)
Pollution is often said to be a problem peculiar to the
industrialized countries. I should like to know whether:
1. Research or studies on pollution have been done in the
developing countries in general and in Africa in parti-
cular? What were the findings obtained?
2. What forms of pollution are found in the developing
countries?
JACKSON (U.S.A.)
I am particularly pleased that Dr. Mokematkenguemba raises
the question of environmental issued in the developing nations.
These nations have many significant environmental problems,
though most are different than those facing the heavily indus-
trialized nations. The major environmental problems in the
developing nations have, in the past, been related to classical
public health areas, particularly infectious diseases. My
knowledge of the scientific literature from many of these areas,
particularly Africa, is somewhat limited. There have been many
important studies on the relation of water resource development
projects and the wildfire spread of schistosomiasis, particular-
ly in the Aswan Valley, West Africa and Tanzania. It is abso-
lutely mandatory that new schistosomiasis control methods such
as the slow-release organotin molluscicides, be thoroughly eva-
luated and vigorous control efforts must be couples with any
new water resource project in these areas. The ecological and
human health disaster precipated by the drought in the Sahelian
area of West Africa has only recently been brought effectively
to general public awareness in many of the industrialized
nations. The U.S. Center for Disease Control has conducted
epidemiological surveys in the Sahel, in conjunction with mobile
health care delivery teams. They noted very high incidence
figures for measles and tuberculosis especially in the children
of those most severely affected - the nomadic herdsman. These
illnesses are clearly related to nutritional deprivation and
overcrowded, hastily constructed refugee camps. To meet the
immediate human needs in this area, continued and expanded ef-
forts from all concerned nations are required. I am not per-
sonally aware of studies in these areas on the more common
industrial related pollutants.
In reply to the second part of your question, the problems
of cross-frontier pollution or the distant effects of pollutants
must receive more attention. The pollution of major water
supplies, including the oceans, affects many nations, not only
the industrialized nations. The evidence of particulate air
pollution and DDT in Antarctica and the finding of fine parti-
culate matter from the Sahara in the upper atmosphere over the
Caribbean both serve to emphasize the "shrinking" nature of the
-------
179
world. Bryson has hypothesized on the global effects of parti-
culate air pollution. He implicates the increase in industrial-
related air pollution over the past 25 years in the slight dec-
rease in temperature noted in the polar regions over the same
time span. This change in polar temperature, according to the
hypothesis/ caused a shift in the high pressure anticyclone and
hence in the edge of the monsoon rain belts. This is advanced
as at least a partial explanation for the drought in the Sahel.
With the very real possibility of increased coal utilization in
the near future it is vital to evaluate this hypothesis and test
its validity, so as to either substantiate or disprove this pos-
sibility.
A particularly difficult area for the developing nations
will be balancing the benefits of rapidly increasing industrial
production essential for improving the material standard of
living for the people of these nations, and the real "costs" of
the pollution generated by such increased industrial activity.
Perhaps some of the experience and mistakes of the industrialized
nations can serve as useful data as each developing nation con-
structs its own priority system - in effect, its own optimizing
of the "quality of life".
-------
181
DESIGNING THE EXPOSURE/RESPONSE MATRIX
IN ENVIRONMENTAL HEALTH STUDIES
JOHN H, KNELSON
Human Studies Laboratory, National Environmental Research Center,
Environmental Protection Agency, Research Triangle Park, NC, USA
ABSTRACT
Assessing the -influence of an environmental agent on pub-
lic health is a difficult multifactorial problem. Bovever,
some unifying concepts in research design are not) being used to
provide a better data base for emission control strategies. The
Exposure/Response Matrix is essentially a three-dimensional ap-
proach to determining the relationship between environmental
stress or insult, and the resulting health effect. One dimen-
sion represents pollutant concentration gradient, the second is
range of susceptibility within a population and the third is the
spectrum of response within population subsets of different sus-
ceptibility. This method of approaching study design and inter-
preted results is applicable to classic toxicology, epidemic-
logy and clinical research.
The pollutant concentration gradient, ranging from back-
ground through levels actually occurring in varying geographic
areas to those levels recognized as producing overt toxicity must
be measured unambiguously and accurately for population studies
and reproduced in controlled environmental laboratories for
toxicologic and clinical studies. The pollutant gradient or a
body pollutant burden is usually treated as the independent
variable.
-------
182
Humans are not uniformly susceptible to various environmen-
tal influences on their health. This normal range of suscepti-
bility must be recognized when characterizing the population in
an epidemiologic study, when establishing criteria for subject
selection in a clinical study, or when preparing animal models
of human disease.
Just as there is a range of susceptibility within the popu-
lation* there is a spectrum of response within each more or less
susceptible subset of the population. The range of response is
usually a function of pollutant concentration, duration of expo-
sure, and the presence of co-stressors such as exercise, thermal
stress, or other co-existing pollutants.
The final goal of environmental health effects research is
the construction of this three-dimensional dose-response relation-
ship to calculate risk factors which reflect the degree of hazard,
as well as the size of the population at risk for any set of
environmental circumstances.
-------
183
INTRODUCTION
Modem man Is well aware of many stresses and health hazards
associated with his rapidly evolving industrial society. These
concerns have been expressed in a variety of laws and regulations
designed to assure occupational safety and consumer protection
over the past several decades. Only relatively recently, however,
has a serious attempt been made to protect the general population
from harmful effects of existing and emerging industrial and
technologic processes.
Complex interactions of environmental factors have focused
attention on the necessity for a systematic approach to the evaluation
of environmental health hazards. Association of changes in health
status with changes in environmental factors is made even more
difficult by the complexity of environmental stress in general.
It is clear some simplifying concepts must be used to organize and
interpret environmental health effects data. These simplifying
concepts are also useful in establishing a systematic approach to
acquisition of new health effects data.
RANGE OF SUSCEPTIBILITY
Humans are not uniformly susceptible to environmental stress. In
epidemiologic studies, greater correlation between pollutant gradient
and health effects has been seen in the incidence of respiratory
infection in children, in the excess number of asthma attacks, and in
excess symptoms in people with cardiopulmonary disease. In clinical
research it is evident that persons with arteriosclerotic heart disease
are much more susceptible to effects of carbon monoxide than are normal
subjects.
In order to make meaningful statements concerning Range of
Susceptibility, we must know (1) the nature of the health effect in
nore or less susceptible individuals, (2) the size of susceptible
populations, and (3) their geographic distribution with respect to
concentration of pollutants in the environment. Even when health
effects for susceptible segments of the population have been determined,
-------
184
the size and distribution of that segment is usually inadequately
known. Not only improved health effects research, but improved
national health survey statistics are required to better estimate
the influence of environmental factors on public health.
SPECTRUM OF RESPONSE
Whereas there is a range of susceptibility within a population,
"spectrum of response" refers to the various ways an Individual
can manifest the effects of environmental stress. The spectrum of
response is usually a function of pollutant concentration, duration
and exposure, and the presence of co-stressors such as exercise,
thermal stress, or other pollutants. The spectrum has been
classified into categories ranging from the mildest to the most
severe. They are: (1) pollutant body burden, (2) changes of
uncertain significance, (3) pathological changes, (4) morbidity,
and (5) mortality. Because the proportion of any population falling
into any category is greatest with (1) and the least with (5), it is
evident that the concept of "range of susceptibility" is closely
related to that of response spectrum. Consideration of the two
concepts allows the construction of a three-dimensional dose-response
relationship which can be used to calculate risk factors that reflect
the degree of hazard as well as size of the population at risk. This
kind of prediction is the final goal of environmental health effects
research. We are trying to answer the question, "How many
individuals in each category of our population are subjected to what
levels of risk for a given range of pollutant concentration?"
THE EXPOSURE-RESPONSE MATRIX
The exposure-response matrix uses dose, established by
environmental monitoring and modeling or by pollutant burden
analysis, as the independent covariate. The two inter-related
dependent covariates, as described, are Spectrum of Response and
Range of Susceptibility. Without the third covariate, such a
plot is simply the classic Dose/Response relationship. With the "z*
axis representing the Range of Susceptibility there is additional
-------
185
information concerning total assessment of risk relating the degree
of response at a particular dose to specifically susceptible
population subsets.
An example of the Exposure-Response matrix is presented in
Table I. Here data concerning the effects of acute carbon monoxide
exposure on cardiac function have been used. Range of Susceptibility
has been divided into four categories: (A) young normal men, (B)
middle-aged normal men, (C) men with pre-clinical or asymptomatic
arteriosclerotic heart disease, and (D) men with overt arteriosclerotic
heart disease. Spectrum of Response is presented according to the
described classification scheme;
(!) Pollutant burden - carboxyhemoglobfn levels above background
(2) Physiologic changes of uncertain significance - decreased
exercise tolerance in young normals, changes in systolic time
intervals in middle-aged normals
(3) Pathologic changes - increased electrocardiographic
abnormalities in middle-aged normals and those with pre-clinical
as well as overt arteriosclerotic heart disease
(4) Morbidity - restriction of activity in patients with angina
pectoris
(5) Mortality
Reasonable estimates for the size of the population at risk
can be made for each of the five categories:
(1) Pollutant burden - A recent survey in the United States has
indicated that 45% of the non-smoking population had carboxy-
hemoglobin concentrations greater than 1.5%
(2) Physiologic changes of uncertain significance - those changes
seen in young normals apply to the entire population. Those
seen in middle-aged normals apply to that fraction 40 years
old or greater (35% of the US population)
(3) Pathologic changes - those seen in middle aged normals
includes that part of the population with either pre-
clinical or overt arteriosclerotic heart disease
-------
186
TABLE I
i*
*^~ "^^
0X1
DlCL
C O)
(O O
oe to
t/i
A
B
C
D
o>
Ic
Jin
0>O»
oca
CO
<*-
o
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
COHb%
0-5
+
+
-
-
—
+
+
-
_
-
•f
+
+
-
-
+
+
*
+
-
5-10
+
+
-
-
—
+
+
-.-
-
-
+
+
+
-
-
+
0
0
0
0
10-15
•§•
+
0
0
™"
+
+
0
0
0
+
+
0
0
0
+
0
0
0
0
15-20
+
+
0
0
'^
+
+
0
0
0
+
+
0
0
0
+
0
0
0
0
+ = effect seen
- = no effect
0 = Inadequate data
-------
187
(4) Morbidity - Patients with angina pectoris comprise about
2.1% of the US population. About 5% have definite or suspect
coronary artery disease.
(5) Mortality - Although some studies show a relationship between
ambient carbon monoxide levels and excess mortality in
hospitalized patients, other studies have not shown such a
relationship. Relationships expressed in Table I are
appropriate only for acute exposures and do not consider the
problem of chronic carbon monoxide intoxication.
No systematic scheme to derive a numerical score for a matrix such
as that presented in Table I has yet been developed. Such a scheme would
be very useful in assessing the adequacy of the data base for any
environmental standard. A more sophisticated and powerful approach is
the use of these data to calculate a damage function. Such a function
will be most useful in performing Cost/Benefit analysis to optimize
environmental control. This function could take the following form:
(1) D = f (XL X2, . . ., Xr . . ., Xk)
(2) X = g (S, R, £)
(3) R = h (rlt r2, r3, n,, rs)
(4) £ = J (Ei, E2, . . . E., . . . E^
(5) E = q (e, p)
(6) e = m (de'/dt
o,
(7) p = n (rlf r2, r3, n,, r5)
Where:
0 = Overall population damage function
X = Population subset damage function
S = Categories of susceptibility
-------
188
R = Response spectrum for each S.
£ = Exposure density function for each population subset
Vu
E = Dose frequency distribution for each geographic area under
consideration
e = Exposure integral (dose) for each population risk subgroup, p.
e"= Air quality data bits
p = Population risk subgroup function based on size of
population in each response spectrum category with
a weighting factor for degree of injury in each
category
The elements of these equations identify many components of a
useful analysis. Currently available environmental monitoring
data as wellas health effects data are still inadequate for such
a multivariate analysis to be meaningful. Derivation of such
an equation is not an empty academic task, however, if it serves to
focus attention on the problems of comprehensive environmental
analysis.
CONCLUSIONS
Evaluation of various environmental control strategies
ideally incorporates an analysis of costs of control compared to
health benefits of improved environmental quality. Cost estimates,
especially when projected over a ten or twenty year period, may
suffer from inadequate technical information. Estimation of health
benefits, however, suffers from conceptual as well as technical
inadequacies. A three-dimensional Exposure-Response Matrix is
presented as a basis for improving synthesis of a human health
damage function as well as providing a better framework for
controlled environmental laboratory studies.
-------
189
DISCUSSION
GOLDSMITH (U.S.A.)
In the presentation o"f the model is it not too elaborate to
assume that for each population subset there are a large number
of response variables (rn) when in most population subsets only
one or rarely two are likely to be observed.
Our colleagues, especially those in Sweden, have demonstra-
ted that annoyance reactions are common, readily measured, and
in our experience a very common health complaint. I include
eye irritation, respiratory irritation, sense of difficulty in
breathing, and possibly tangentially, noise annoyance. I be-
lieve that these effects are considered by the public to be health
effects; why do you and Drs. Jackson and Newill exclude annoy-
ance reactions from population health responses? In doing so,
do you not imply that these reactions have no long-term disease
or disability implications? On what basis can we assume that
these annoyance reactions are not associated significantly with
risk of increased body burden, physiological impairment, or
pathological changes?
The assumption that costs of medically measurable effects
are more readily measured than annoyance reactions is very shaky.
Costs of change of residence, loss of property value, and demands
for relief and abatement are readily measured costs, compared to
these, costs of increased body burden, abnormality of ECG, or
even impairment of exercise capacity in angina patients, are dif-
ficult to assign a dollar value to.
KNELSON (U.S.A.)
I share Dr. Goldsmith's concern with the complexity of the
general damage function presented. However, it is the advantage
of using symbolic notation, that one can render very complex re-
lationships relatively comprehensible. Compared to the real
world, the list of formulae I have proposed are simplistic.
Including dose response variables for population subsets with
respect to range of susceptibility as well as spectrum of res-
ponse is necessary to show the interaction between these two
interdependent vectors in defining the size and nature of the
population at risk.
With respect to Dr. Goldsmith's second statement, I do not
assume that medical effects are more easily measurable than nui-
sances. I simply have no experience with the latter problem
and am therefore not competent to discuss it. I am sure, how-
ever, that others could provide the necessary data for inclusion
in a more general statement of damage.
-------
190
DEAN (Ireland)
We are taking part in a week's symposium on the Health Ef-
fects of Environmental Pollution involving over 200 papers. No
paper considers the greatest threat of environmental pollution
which is radiation, perhaps from a limited nuclear war.
Should we not include in our discussions how the human race
can survive the environmental pollution which may result from a
sudden increase in radiation, - and as a corollary express our
concern at the failure to restrict preparations for a nuclear
missile war or the inadvertant release of a massive amount of
radiation?
KNELSON (U.S.A.)
I am sure we all share your concern. However, in the con-
text of this symposium it seems to me impossible to even begin
to deal with the important issue you have raised.
(EDITORIAL SOTE: It was a decision of the Organizing Committee
not to include the topic of risks due to ionizing radiation in
this Symposium in view of the many meetings devoted to this sub-
ject. Furthermore^ several specialized international organi-
zations are now reviewing these risks.)
SCHLIPKOTER (Federal Republic of Germany)
When setting threshold values two points are difficult to
judge:
1. What importance must be attached to odour annoyance and
irritation effects? Irritation effects should be regarded
as having negative effects on health. Do you agree with
this standpoint?
2. How should subjects with previous damage (illnesses) be con-
sidered? Which illnesses and degree of illness should still
be taken into consideration? If persons with severe emphy-
sema or bronchitis were included the threshold values would
have to be very low.
KNELSON (U.S.A.)
1. Dr. Schlipkoter has raised the question of quantitating the
nuisance aspects of environmental stress. This problem
has been approached by several groups of investigators,
usually employing a subjective ranking score. I agree
with the importance both you and Dr. Goldsmith attach to
annoyances, irritation syndromes and what might be referred
to as interference with quality of life in general. I
-------
191
believe these aspects of environmental stress can be inclu-
ded in the overall damage function I have defined.
I agree also with Dr. Schlipkoter's concern with exacerbation
of chronic illness related to environmental factors. It is
precisely interactions of this sort that should be recognized
in the overall assessment of a general damage function. As
you know, in the United States we do recognize the need to
protect particularly susceptible segments of the population.
This principle is most evident with respect to CO and coro-
nary artery disease. We have described the exacerbating
effect of respirable sulfates on asthmatics and have work in
progress concerning the effects of nitrogen dioxide and ozone
on persons with chronic lung disease.
-------
193
DOSE-RESPONSE RELATIONSHIPS LINKING SHORT-TERM
AIR POLLUTION EXPOSURES TO AGGRAVATION
OF CARDIO-RESP1RATORY ILLNESS
R, CHAPMAN/ J, FRENCH, J, FINKLEA AND H, GOLDBERG
Environmental Protection Agency, Research Triangle Park, NC, USA
ABSTRACT
In 2970 and 1972, the US Environmental Protection Agency
conducted several epidemiologic studies which assessed the ef-
fects on human health of day-to-day fluctuations in air pollution
levels. Three of these studies are summarised in this report.
Two studies, one in the New lork City metropolitan area and one
in Utah's Salt Lake Basin, investigated pollution effects in
groups of asthmatic children and adults. The third, in the
New York area, investigated pollution effects in groups of eld-
erly patients with combined cardio-pulmonary disease. In all
studies, subjects maintained symptom diaries for about eight
months. A total of 359 asthmatics and about 200 elderly patients
participated.
In Utah, outdoor temperature showed a strong negative cor-
relation with asthma attack rates. On subfreezing days, the
temperature effect was strong enough to obscure any pollutant
effects that may have been present. On days of minimum temper-
ature between 30° and 50° Farenheit, attack rates increased con-
aistently as total suspended particulate (TSP) levels increased.
On days of minimum temperature above BO , attack, rates increased
consistently as both TSP and suspended sulfate (SS) levels in-
creased. Levels of SS, above which excess attacks could be
-------
194
expected, were estimated at 17.4,ug/m for days of minimum tem-
perature between 30° and 50°, and at 1.4,ug/m for warmer days.
In Sew York, as in Utah, the estimated SS threshold for
excess asthma attacks was lower on days of minimum temperature
7 7
above 50° than on cooler days (7.5,ug/m and 12 ,ug/m respect-
ively). Effects of temperature and pollution on asthma attack
rates were generally not as strong in New lork as in Utah.
In cardio-pulmonary patients, decreasing temperatures pro-
moted increased rates of aggravation of symptoms. The effects
of pollution on cardio-pulmonary symptoms were slightly stronger
on cold days than on warmer days. The level of SS above which
excess symptom rates could be expected was estimated at 9.2 ,ug/
m for days_of minimum temperature between 20° and 40°. For
warmer days, a reliable point estimate of the SS threshold
could not be made.
-------
195
1. Introduction
This report will summarize the results of three epidemiologic
studies, ' ' conducted by the U.S. Environmental Protection
Agency (EPA). Two studies tested the hypothesis that daily changes
in ambient air pollution directly affect the frequency of asthma
attacks. The third study tested the hypothesis that daily changes
in ambient air pollution directly affect the severity of symptoms
of chronic cardie-pulmonary disease. Each of the three studies
was the first of a series in its area. One study of asthma was
performed in the Salt Lake Basin in Utah, in the vjestern United
States, where the major pollution source is a large copper smelter.
The major pollutants in this area are oxides of sulfur, specifically
suspended sulfates and sulfur dioxide. Two studies, one of asthma
and one of chronic cardie-pulmonary disease, v;ere performed in the
New York City Metropolitan Area, where the major pollutants are
oxides of sulfur and suspended particulate matter.
2, Methods
In New York and Utah, asthmatics and cardio-pulmonary patients
were identified and enrolled through clinical records, health
questionnaire information, and house-to-house interviews. Asthmatics
were eligible for study if they had had at least one asthmatic attack
in the past year, and if their attacks involved, at a siinimum,
wheezing and shortness of breath. Both children and adults were
enrolled in the anth:na studies. Patients v/ith chronic cardio-
pul.-nonary disease were all nt least 60 years old, and had had their
illnesr, diagnosed by a doctor.
The Utah asthma study was conducted in four communities of
different sulfur oxide exposures. The New York asthma and cardio-
-------
196
pulmonary studies were each conducted in three communities of
different sulfur oxide and particulate exposures. Two communities
were within the New York City limits; the third was on suburban
Long Island. Air monitoring stations were installed in resident-
ial areas in each community. Daily measurements of sulfur dioxide,
suspended sulfates, and total suspended particulates were obtained
from each station. Each study participant lived within two miles
of his community's air monitoring station. In all three studies,
study participants used self-administered symptom diaries to report
asthma attack frequency or cardio-pulmonary symptom severity. In
Utah, data were collected during the 26-week period from March 7,
1971 through September k, 1971. In New York, data were collected
during the 32-week period from October 8, 1970 through May 22, 1971.
In each community in Utah and New York, about 50 asthmatics were
enrolled. In the New York cardio-pulmonary study, a total of
about 200 participants were enrolled.
3. Results
3.1 Utah Asthma Study
In the Utah asthma study, the frequency of attack rates was
more strongly and consistently correlated to minimum daily tempera-
ture than to any other environmental variable. As temperature
decreased, attack frequency increased. Indeed, on subfreeaing
days (minimum temperature below 30°F), no effect of air pollution
on attack rates could be detected.
On days of minimum temperature between 30 and 50°F, increases
in asthma attack rates were associated consistently with increases
in total suspended particulates (TSP). For all days in this
temperature range on which the daily average TSP level was at or
-------
197
below 60^/ig/m , the asthma attack rate was 16.1 attacks per hundred
person-days of diary coverage. (In other words, for each 100
diaries from days having these temperature and pollution conditions,
16.1 diaries would report one or more asthma attacks.) When TSP
levels rose to 61-75 ji£/m , the attack rate rose by 12%, to 18.0.
On days having TSP levels of at least 76 ^f/m3, the attack rate rose
by an additional 3%, to 18.5 (Figure 1). On days of minimum
temperature between 30 and 50°F, increases in sulfur dioxide (S0_)
and suspended sulfates (SS) were not consistently associated with
increased attack rates.
UTAH, 1971: RELATIVE RISK OF ASTHMA ATTACK ON DAYS OF LOW AND HIGH EXPOSURES
TO SO?, TSP, AND SUSPENDED SULFATES, (MINIMUM TEMPERATURE 30 TO 50 CF)
1.50
1.00 —
*z
o
0.50 —
<:
u_
o
^"~
1.00
(16.9>*
0.70
0.93
1.00
(16.1)
1.12
1.15
•"'
£60 61-80 £81
S02, ug/m
£60 61-75 £76
TSP,
—
1.00
ilC C\
ilo.b)
095
Oac
.83
083
0.50 —
<6 6.1-8 8.MO £10.1
SS, pg/m3
* Baseline attack rates in parentheses
FIGURE 1
-------
198
On warmer days, of minimum temperature aboTe 50 F, increases
in both TSP and SS were consistently associated with increased
attack rates (Figure 2). When TSP levels were below 6O ug/m on
such days, the attack rate was 12.1 attacks per 100 person-days.
•z
When TSP increased to 61-75 Jig/m t the attack rate rose by &%+ to
13.1. When TSP levels were above 75 ^ig/m , the attack rate rose
by 1756 above the baseline level, to 1^.2. On warmer days, only
UTAH, 1971: RELATIVE RISK OF ASTHMA ATTACK ON DAYS OF LOW
AND HIGH EXPOSURES TO S02, TSP, AND SUSPENDED SULFATES
(MINIMUM TEMPERATURE £51°F)
J..TV
f AA
1.00
O
£ 0.50
•X
1.50
H-
UJ
ce
100
0.50
0
1.50
1.35 I
1.17 I
1.00 |
—
(10.8)
—
3
SS, jig/ra
FIGURE, 2
* Baseline attack rates in parentheses
-------
199
small increases in attack rates v;ere observed with increasing
SOo levels.
Attack rates rose most sharply and consistently with increas-
ing SS levels on warm days. When SS were below 6 ug/nr on such
days, the attack rate was 10.8 attacks per 100 person-days. When
SS increased to 6-8 ug/m , the attack rate rose by 17% to 12.6.
When SS increased to 8-10 ug/m , the attack rate increased by an
additional 18%, to 1/^.6. Finally, on warmer days when SS levels
exceeded 10 ug/m , the attack rate was 16.£. This figure
represents an increase in attack rate of 50% above days v;hen sulfate
levels were below 6 ug/m .
Least squares estimation techniques v/ere applied to the Utah
data to determine threshold levels of SS above whish excess asthma
attacks could be expected (Figure 3). Estimates indicated that on
days of minimum temperature between 30 and 50°F, excess attacks
might not be expected until SS levels reached 17.^ yug/m-5. On
warmer days, excess attacks might be expected when SS exceeded the
very low level of l.tf /ig/m
20
'0 4 8 12 16 20 24 28
FIGURE 3 SUSPENDED SULFATES, |it/m3
Utah, 1971: temperature - specific estimates of suspended sulfate
levels above which excess asthma attacks occur.
-------
20O
^. Discussion
Each of the three studies described in this report was the
first of a series. Clearly, the findings of each study require
verification by the subsequent studies in its series. However,
the studies described here stand alone in identifying SS as a
pollutant of major concern. On the basis of the findings
described, the investigators felt that adverse effects on both
asthma and combined cardio-pulmonary disease might be expected when
daily average SS levels reach 8-10 /ug/nr or above. In American
cities, daily SS levels exceed 8-10 M6/m a substantial proportion
of the time. Indeed, during the fall and winter of 1971, SS levels
exceeded 8-iO yug/nr over 509» of the time in one of Utah's study
communities. During the period of study in New York City, daily
SS levels exceeded 8-10 yug/nr 70$ of the tine.
The threshold SS estimate of 8-10 wg/ar may in fact be higher
than the true threshold in some geographic areas. Theoretical and
empirical considerations support this point. Theoretically, it is
thoroughly conceivable that, with increased knowledge and investi-
gative skill, scientists will detect adverse health effects in the
future at near-zero concentrations of SS and other pollutants.
Bnpirically, the low threshold SS level of 1.1* ug/nr was estimated
from the data collected from Utah asthmatics on warm days.
There are at least three possible explanations for the
observed discrepancy between the New York and Utah threshold
estimates. First, the reporting habits of asthmatics may have
been different in the two areas. Second, the least squares
estimation techniques yielded broad confidence limits around the
point estimates of thresholds. We cannot stete unequivocally
-------
201
3.2 New York Asthma Study
In New York, asthma attack rates were generally hi-gher than
in Utah. This difference probably reflects differences in the
original composition of study samples rather than more severe
pollutant effects in New York than in Utah. In fact, neither
temperature nor pollution exerted effects on attack rates in New
York as strongly or consistently as they did in Utah. In New York,
on days of minimum temperature between 30 and 50 F, attack rates
generally increased with increasing TSP and SS levels (Figure If) .
NEW YORK, 1970-71: RELATIVE RISK OF ASTHMA ATTACK ON DAYS OF
LOW AND HIGH EXPOSURE TO S02, TSP, AND SUSPENDED SULFATES
(MINUMIM TEMPERATURE 30 TO 50 °F)
1.J.O
.1Z
1.08
1.04
v-- t nn
o
£ 0.96
* 0
u.
o
*£ 1 1C
a I.ID
ee
SS 1.12
i-
uJ 1.08
o±
1 AA
1.1M
.00
0.96
0
1.13
"~ 1.06 ~~
~~ 1.02 ~~
1.00 1.00
(22.2)' 0.98 !2Ui
— —
III 1 1 1 1
£60 61-80 281 £60 61-75 £76
$02, |ig/ra3 TSP, iig/m3
1-09 1M
1.04
1.00
(19.0)
1 1 1 1 1 1
6.1-8 8.1-10 £10.1
SS, |i(/m3
*Base attack rates in parentheses
Figure k.
-------
202
On such days, the attack rate rose from 21.3 attacks per 100
person-days when TSP levels were below 61 yug/m , to 2^.1, an
increase of 1396» when TSP levels exceeded ?6 yug/m . On days in
the same temperature range, the attack rate rose from 19*0 when SS
levels were below 6.1 ug/m , to 20.7. an increase of 9%, when SS
levels were between 8 and 10 iig/m , When SS levels exceeded
10 yug/m, a very slight fall in the attack rate (1%) was observed.
On warmer days in New York, when the minimum temperature
exceeded 50°F, the attack rate continued to increase with increas-
ing SS levels (Figure 5). On such days the attack rate rose by
105^, from 22.1 attacks per 100 person-days when SS levels were
below 8 ug/m , to 2*t.3 when SS levels exceeded 10 yug/m . On these
NEW YORK. 1970-71: RELATIVE RISK OF ASTHMA ATTACK ON DAYS OF
LOW AND HIGH EXPOSURE TO S02, TSP, AND SUSPENDED SULFATES
(MINIMUM TEMPERATURE £51 °F)
1.08
1.04
1 Oft
0.96
AA4
0.88
0.84
—
_^
^^»
—
1.00
(21.51*
0.87
0.97
1.00
(212)
0.92
1.05
—
^^^m
^^^B
J i
I 1 I L
I i
£ £60 61-80 281 £60 61-75 276
* S02.K/-3 T$P,pe/rf
ui *•"
| 1.M
£ 1.04
tc
1 AA
1.00
0.96
0.92
0.88
0.84
^ p
— 1.03
1.00
_
—
—
—
(22.11
^^^
—
^—
—
—
—
1 1 1 1 1 1
£8 8.1-10 210.1
Figure 5 . * »««*11»» »tt«ck r«tt« i» »ar««tb»«««
-------
203
warmer days, the attack rate did not increase consistently with
TSP levels.
No consistent effect of SO- on attack rates was observed in
Hew York in either temperature range.
Estimates of threshold 5S effects on attack rates were develop-
ed from the "New York data (Figure 6). As in Utah, the SS threshold
was higher in New York on warm days than cooler ones (11.9 Pg/nr and
7»3 Mg/» » respectively). The difference in temperature-specific
threshold estimates was not as- great, nor was the warm-temperature
threshold estimate as low, in New York as in Utah.
NEW YORK, 1970-71: TEMPERATURE-SPECIFIC ESTIMATES OF
SUSPENDED SULFATE LEVELS ABOVE WHICH EXCESS ASTHMA
ATTACKS OCCUR
0=
IU
o.
Ill €/»
£9
O P
vt
30 TO 50 °F ^5
SUSPENDED SULFATES, |ig/m3
Figure 6.
3.3 New York Study of Cardio-Pulmonary Disease
For cardie-pulmonary patients in New York, decreasing tempera-
Wires were strongly associated with increased frequency of aggra-
vation Of symptoms. On days of minimum temperature between 20°
and 40°F, the frequency of aggravation rose quite steadily as SS
levels increased (Figure 7). On such days, the rate of symptom
-------
204
NEW YORK, 1970-71: RELATIVE RISK OF CARDIO-PULMONARY SYMPTOM
AGGRAVATION ON DAYS OF LOW AND HIGH EXPOSURES TO S02, TSP,
AND SUSPENDED SULFATES (MINIMUM TEMPERATURE 20 TO 40 °F)
i.lt
1.08
1.04
= 1.00
e
—
1.00
| (52.61*
5 ffl 1
- £60
1.05
^n 5,2
1.01
III III
61-80 £81 £60 61-75 >76
° S02, CB/ui3 JSP, jjg/m3
£ 1.16
:>
5 LU
UJ
ft*
i ntt
l.UO
1.04
1.00
1.08
__ 1.00
nl
A.»D
i.tU
1.15
^_
—
—
1 i 1 1 1 1 1
6.1-8 8.1-1010.1-12 £12.1
SS, pg/n)3
attack rates in parentheses
Figure ?.
aggravation was 51.796 when SS levels were 6 ug/ra-5 or below. When
SS levels were between 8 and 12 ug/m , the aggravation rate
increased by 16% to 59.5#. When SS levels were above 12 ug/m5,
the aggravation rate declined very slightly (1%) to 59.0%. On
these cooler days, elevations in symptom severity associated with
rising TSP and S02 were small, 5°$ at most, and not very consistent.
On wanner days, of minimum temperature above ^0°F, there v/as
also a distinct tendency toward increased frequency of symptom
aggravation as SS levels increased (Figure 8). The aggravation rate
was 46.8% when SS levels were 6.0 ^g/nr or less. When SS levels
rose to 10-12 ug/m , the aggravation rate increased 1?^, to 5if. S%.
-------
205
NEW YORK, 1970-71: RELATIVE RISK OF CARDIO-PULMONARY SYMPTOM
AGGRAVATION ON DAYS OF LOW AND HIGH EXPOSURES TO S02, TSP, AND
SUSPENDED SULFATES (MINIMUM TEMPERATURE
i.li
1.08
i M
1 CM
0 96
•E 0.92
? w»
^ 0.84
—
_
—
1.00
(49.41*
0.86
0.93
1.00
WA)
1.04
0.96
—
—
— 1
I
I
I
I
I
fcO
«r
£60 61-80
S02,
£60 61-75 276
TSP,
S l.UI
\—
5 i.«
£ 112
1.08
1.04
1.00
0.96
0.92
0.88
084
A
Uf
^^"
_
— ,M 1.01
—
—
—
(46.8)
1
1.02
1.11 _
—
—
-^
—
—
— J
M SS,
Base attack rates in parentheses
Figure 8.
When SS levels were above 12 yUg/m^, there was a slight (6#) decline
in the aggravation rate. On warmer days, there were no consistent
elevations in symptom severity associated with increasing TSP and
S02 levels.
The anbient level of SS at which elevated rates of symptom
aggravation could be expected were estimated by least squares
estimation techniques. For cardio-pulmonary patients on days of
minimum temperature between 20 and i*0°F, this estimated SS level
was 9.2 ug/m^. For warmer days, a reliable point estimate of a
t
threshold could not be made.
-------
206
NEW YORK, 1970-71: ESTIMATE OF SUSPENDED SULFATE LEVEL
ABOVE WHICH CARDIO-PULMONARY SYMPTOM AGGRAVATION OCCURS
o
<
0=
C3
o
OL
or
o
a.
o
oc
5 10 15 20
DAILY SUSPENDED SULFATES, jig/m3
25
Figure 9.
that the thresholds estimated for Utah and New York were signific-
antly different from each other. Third, the lower threshold in
Utah may indicate that the physical and chemical composition of SS
in Utah's atmosphere was different from that in New York's. These
possibilities are currently being tested.
REFERENCES
1. Finklea, J.F., et al. Aggravation of Asthma by Air Pollutants:
1971 Salt Lake Basin Studies. Health Consequences of Sulfur
Oxides: A Report From CHESS, 1970-71. In press at U.S.
Government Printing Office.
2. Finklea, J.F., et al. Aggravation of Asthma by Air Pollutants:
1970-71 Hew York Studies. Health Consequences of Sulfur Oxides:
A Report From CHESS, 1970-71. In press at U.S. Government
Printing Office.
3. Goldberg, H.E., et al. Frequency and Severity of Cardiopulmonary
Symptoms in Adult Panels: 1970-71 New York Studies. Health
Consequences of Sulfur Oxides: A Report From CHESS, 1970-71.
In press at U.S. Government Printing Office.
-------
207
DISCUSSION
MAGE (Denmark)
Is it not correct that daily suspended sulphates at o mg/m
must give zero excess attacks? If as you state that the "noise"
in the data prevent you from discerning the lower portion of
the curve, what justification is there for drawing the curve as
horizontal? Why did not the study filter this "noise" out front
the data? You define excess asthma attacks in a nonspecific
manner, since your basis is not defined for what constitutes
zero excess attacks.
CHAPMAN (U.S.A.)
The hockey stick function is an expression of simple attack
rates, not "excess attack" rates. The zero line for "excess
attacks" would be the same line as the horizontal portion of the
hockey stick function. In that sense, the hockey stick function
does indeed intersect "excess attacks" at zero.
Clearly, in any population of active asthmatics, there will
be attacks even in the complete absence of sulfates. This is
unavoidable. Thus, the hockey stick function does not, indeed
cannot, intersect the zero line for simple attack rate.
HINE (U.S.A.)
1. Is the concept of the continuum of effects rather than the
use of the classic dose-response curve on a political, socio-
logical or biological basis?
2. Were the differences shown statistically significant for
increased responses by the asthmatics in Salt Lake and New
York?
3. Is there a plateau effect with increased levels of pollutant
as regards responses of the selected population?
4. Is there a lower arm to your sigmoid curve?
-------
208
CHAPMAN (U.S.A.)
1. We have certainly not ruled out the possibility that classic
dose-response curves exist for many pollutants. It does seem
reasonable, though, on a biologic basis, that as health indi-
cators become more and more sensitive, effects will be observed
at lower and lower pollution levels.
2. In addition to the temperature-specific relative risk models
presented, simple correlation and stepwise multiple regression
analyses of these data were performed. Correlation and regres-
sion coefficients were generally not statistically significant
for SO, and TSP, but were quite consistently significant for
sulfates.
3. The plots of relative risk for the New York asthma and
cardio-pulmonary studies (Figures 4, 7 and 8) suggest that there
may be a "plateau effect" for sulfates. This finding definitely
requires verification in replicate studies.
4. To date, we have not mode detailed analyses of the lower
end of the exposure spectrum. However, the presence of a
horizontal portion at the lower end of most of our threshold
estimation functions to date suggests that, for the parameters
studied, there may be a lower horizontal portion in the dose-
response curves.
PHAM (France)
1. Your paper raises the problem of the use of health indi-
cators. You used the number of attacks registered by the sub-
jects as an indicator, what effect would the treatment that
these subjects could receive have on subsequent attacks?
2. If you use a bronchial susceptibility test (for example,
measurement of change in air way resistance) would it not
lower the observed thresholds?
CHAPMAN (U.S.A.)
1. You are raising a problem that is most pertinent, and most
difficult to control for. I believe that, in most cases, the
attack rate in a treated asthmatic would be lower than the rate
in the same patient if he were untreated. Many of the asthmatics
in our studies routinely take some sort of preventive medication,
so that the attack rates that we publish might best be considered
minimum attack rates.
It seen»s unlikelv to me that medication taken to relieve
an attack in progress would have much effect on the incidence of
future attacks, since the duration of activity of routine asthma
medication is usually quite short. However, I also believe one
-------
209
could argue that present medication might affect future attack
rates positively of negatively. To my knowledge, few data are
available on this point.
In our studies of asthma, we ask all patients daily whether
they have taken medication for asthma, and whether they have
taken more or less medication than usual. In general, the pro-
portion of patients taking medication differs little from
community to community. Whatever effect medication might have
on attack rates probably is distributed evenly in all comm-
unities. Thus,though our published attack rates may be low, our
conclusions about attack rates in one community relative to
another are probably valid.
2. It is very difficult to say whether pulmonary tests would
have been more sensitive indicators than number of attacks,
since few data are as yet available on this point. In the next
year, EPA will be testing the sensitivity of just such indica-
tors as you have suggested.
BULCRAIG (U.K.)
As a chemist I feel it is essential to distinguish between
neutral sulphats and acid sulphates in studies of the relation-
ship of diseases to SO- and suspended sulphates. It is well
known that SO2 is oxidised to sulphate in perhaps 2 or 3 days
but there ari ample data in the United Kingdom to show that
the endproduct is mainly ammonium sulphate or sodium sulphate.
Sulphuric acid has been found in foggy conditions, and the
Threshold Limit Values recommended by the U.S.A. Association
of Industrial Hygienists and the UK Department of Health show
that SO3 or sulphuric acid are rated as 12 times more toxic
than SO- while neutral sulphates are not even included in the
list of substances for which a TLV is recommended, I there-
fore regard as essential to mesure the acidity as well as the
total of suspended sulphates. This might explain the signifi-
cant difference between Utah and New York since Dr. Chapman
mentioned the existence of a large smelter in Utah and furnaces
often emit a higher proportion of SCU to SO- whereas in New York
the sulphates may be mainly neutral as in most cities in the
U.K.
Chemically it would not be difficult to measure both
factors and until this is done one cannot be sure of the inter-
pretation of the apparent relationship of the diseases to total
suspended sulphates .
-------
210
CHAPMAN (U.S.A.)
I agree wholeheartedly with your general point, that com-
plete chemical and physical characterization of local sulfates
is an extremely important goal. One should bear however, in
mind that certain metal sulfates such as zinc ammonium sulfate
can cause increased airway resistance in animals. Thus, even
metal sulfates may exert toxic effects in human populations.
-------
211
EFFECTS OF SOME APPROXIMATIONS IN ANALYSES
OF RADIOLOGICAL RESPONSE TO COALMINE DUST EXPOSURE
M, JACOBSEN
Institute of Occupational Medicine, Edinburgh, United Kingdom
ABSTRACT
Problems of analysing incomplete epidemiological data to
provide guidance on coalmine dust standards are considered. An
interim analysis of results from a longitudinal study in British
nines had provided estimates of probabilities of developing var-
ious degrees of coalworkers ' pneumoconiosis following exposure
to different concentrations of respirable dust. These dose
response relationships were used to guide decisions on dust
standards. Data which became available subsequently have per-
mitted quantitative assessments of how certain approximations
used for the interim analysis affected results.
Averages of dust concentrations at coalfaces sampled in a
colliery during a ten year period proved to be good estimates
of average cumulative dust exposures of face workers in the col-
lieries. However, re-grouping data according to men's indivi-
dual dust exposures showed that attribution of a single (mean)
dust-dose to men working in the same colliery introduced a bias
into the analysis. Risks associated with high dust concentra-
tions had been overestimated; at lower concentrations they had
been underestimated, to a lesser extent. A simplified stochastic
model of radiological progression was used in the interim analysis
to permit temporal extrapolation of results. This took no for-
mal account of variations in dust exposure prior to the research.
-------
212
Later* calculations give more variable results, but provide no
evidence of serious bias arising from the simplifying assumption.
An earlier finding concerning higher risks of radiological res-
ponse to dust among men with pneumoconiosis is confirmed. The
practical importance of this observation is discussed.
-------
213
1. Introduction
Use of epidemiological data as a guide to decisions on pexmissible
levels for pollutants presupposes relevance of study designs to the
problem of interest and the use of valid methodology. Even when these
conditions are met, difficulties arise if specific action is required
before an epidemiological study is complete, and guidance is sought
from the incomplete data. The plausibility of assumptions used to
interpret th* data say rest partly en informed value judgements rather
than on the results available} but inherent in this situation is the
opportunity to study the effects of approximations as more data become
available.
Such was the case in the period leading to the introduction of
gravimetric standards for airborne dust in British coalmines in 1970*
An interim analysis of results from a longitudinal study in the sines
provided estimates of probabilities of developing various degrees of
coalworkers* pneumoconiosis following exposure to different
concentrations of resplrable dust (Jacobsen et al jjl]). Application
of these results to conditions In British coalmines, to assess likely
effects of different possible dust standards, required a number of
operational assumptions. The data justifying them have been described
by Chamberlain et al [2], Consequences of possible deviations from
conditions assumed have been discussed quantitatively by Jacobsen
The present paper is concerned with some other approximations which
were necessary to derive the probability relationships themselves.
The Fneumoooniosis Field Research began in 1952 with two sain aims:
to establish what kinds of dust cause pneumoconiosis and to determine
what environmental conditions should be maintained if mineworkers are not
to be disabled by the dust they breathe* The research design has been
described fully by Fay and Rae [4] • In essence the plan was to take
full-sice chest X-rays of all working miners in 25 collieries at five-
year Intervals* About JO 000 men were involved at any one time and the
response rate was consistently higher than 90 percent. The collieries
were selected from all the British coalfields to provide a representative
range of environmental conditions that night prove to be important in the
development of pneumoconiosis. Research investigators stationed
-------
214
permanently at the collieries carried oat a statistically designed
programs* of dust sampling close to the workmen in defined occupational
groups. Records were accumulated of the tine spent by each Ban in these
groups and occupational histories prior to the research were compiled by
interview for all men surveyed. In this way it was hoped to establish a
quantitative relationship between radiological appearances and
individuals* dost exposures.
3. The inter** st***^
3.1 Ob.1eotiTe
Che first British coalmine duet standards had been expressed in
terms of the number of particles in the sice range 1 to 5 jta per ca? of
sampled air and war* Beasured using the Standard Thermal Precipitator.
Hew dost sampling instruments had been developed by the late 1960s which
measured the mass concentration of respirable dust. This was the
preferred measure of the medical hasard (Bedford and Warner [5];
Hamilton and Valton £6]) whose use for routine monitoring of dust levels
had bean postponed only by the earlier non-availability of suitable
sampling instruments* The decision by the Rational Coal Board to
introduce gravimetric sampling underground by 1970 determined the
objective of the Interim study which began in 1967: a relationship
between radiological pneumooozdosis and mass concentrations of airborne
dust was required urgently to guide the necessary decisions on
appropriate new dust standards.
These are described and discussed in detail by Jacobsen et al [?J.
The essentials are summarised below.
3.2.1 Radiological
Pairs of X-rays of 4 122 coalface workers employed in 20 of the
research collieries were classified for pneumoconiOBia by eight doctors
experienced in the radiology of pneumoconioais. Each pair of X-rays
represented the start and end of a ten-year period of dust sampling in
the collieries concerned. Every doctor examined all pairs* Both films
from a pair were classified at the same time using the 12-pdnt scale of
abnormality now incorporated in the International labour Organisation's
classification of paeumoconiosis (ILO (&]). The eight classifications
of each film wars averaged, as described by Jacobsen et al [7]*
-------
215
7.2.2 Environmental
The mean dust concentration over all coalface occupational groups
sampled over the ten-year period was available for each of the 20
collieries included In the special study. The concentrations were
expressed in the units in which they had been measured, that is
particles/cm3. These 20 colliery mean concentrations were converted
into mass concentration units (ag/fe3) using separate factors for each
colliery; the factors had been determined experimentally from the
results of simultaneous sampling instrument trials in each colliery.
The statistical analyses described by Jacobsen et sjl [?] showed
that radiological changes over ten years were correlated with the 20
colliery mean mass concentrations (r • 0*87) and the results were used to
estimate risks of developing simple pneumoconlosis over periods longer
than ten years*
4, Effects of some assumptions
The mean concentration of all coalfaces sampled in a colliery was
taken to be an estimate of the mean exposure for the sample of coalface
workers in the colliery. Soon after the original analysis had been
reported, calculations of men's individual exposures over the ten years
were completed. It transpired that the colliery mean coalface
concentrations used in the interim analysis correlated well (r •* 0.94)
with the mean dust expo sore a of the men studied, although the latter were
calculated from detailed records of individual attendance in occupational
groups, including some groups away from the coalface* Thus it could be
concluded that the positive correlation between radiological results and
measures of airborne dust found earlier was not an artefact arising from
reliance on results from all coalfaces sampled, as opposed to the
individually time-weighted concentrations in groups where the men had been
working* Notwithstanding this reassurance, a deviation from linearity
ia the relationship between the two sets of mean data suggested a
possible Mas, fig. 1 shows the relationship between colliery mean
results and the mean of concentrations experienced by individuals in
those collieries* This figure demonstrates what was not obvious from
an earlier analysis by Jacobsen et aj. (vj. A bias in variability about
the theoretical straight line which would represent a 1 : 1 relationship
-------
216
ia evident in fig* 1 at lover as veil aa at hig»r concentrations* Thua
the data naed in the interim analyais appears to hare not only
undeieetLnated dust lerela experienced by individuals at relatively higi
oonoentrationa bat alao to hare overestimated then at lower
concentrationa«
Mean Concentrations Experienced
By Individuals (mg/m3)
Colliery Mean Coalface Dust Concentrations
(mg/m3)
Variability of mean duat concentrations experienced by
individuals in * colliery over ten years in relation to
col-He*y mean coalface dust concentrations.
f-SD
RANGE
-------
217
4.2 Pse of
The within-colliery variability in M.g. 1 is seen to be an
increasing function of the colliery Bean concentrations; the highest
range of individuals* Bean concentrations in one colliery was from 2.1 to
16.1 mg>fe3. s*h« effect of this variation on the estinated ten-year lose-
response relationship has been described by Jacobsen QjJ. It is
consistent with the bias evident in Fig. 1 but the over-esttaation of
lever concentrations was seen to relate mainly to a minority of the nan
studied who began the ten-year period with some radiological evidence of
pneimwaoaiesis. Hot discussed previously is the question of hov the
revised ten-year dose—response relationships affect temporal
extrapolations of results which were made in the original analysis,
Jaoobsen et al |l ,?] had produced estimates of dose-related risks of
developing category 2 or higher siaple pneuaoceniosis over a 35-year
working life at the coalface. These estimates have been used to
deteraine tie likely long-ten effects of dust standards both in the UK
(Jacobsen |~j]) and in the USA (Doyle [lo]). Table I compares estimates
of risks over 35 years. Those relating to concentrations experienced
by individuals have been calculated in the save way as that used in the
earlier analysis of colliery aean coalface concentrations, but results
frea the same 4 122 men concerned have been re-grouped, according to
convenient ranges of individually experienced mean concentrations.
In the aiddle region of the concentrations experienced the estimates are
similar. the bias arising from calculations based on colliery aean dust
concentrations affects mainly the estimated risks at higher concentrations.
This is seen more clearly from Fig. 2 where smooth curves have been fitted
to the results (using the angular transformation of estimated
probabilities). At 4 mgA5 ths two curves give the same probability
estimates. At 8 n*/*5 the difference amounts to 0.08 probability units.
4.3 Assumption of a Barker model for temporal egtrarelation Of
4.3.1 The model
A third possible source of bias in the previously published results
arises from some assumptions used to produce the 35-year-risk predictions
from data relevant to lO-year periods. The model depicts changes in the
-------
218
Table I. Estimates of probabilities of developing category 2
or higher simple pneumoconiosis during a 35-year
working life under various dust conditions
a. using colliery mean dust concentrations
b. using mean concentrations experienced
by individual men in the collieries.
Men
Colliery
S
D
0
P
C
G
J
I
B
M
£
X
T
W
Y
V
L
Q
A
F
Total
a.
grouped according
to collieries
Number
of men
129
113
319
234
215
268
254
161
211
261
76
160
281
99
293
335
196
291
80
146
4122
Kean dust
cone.
(mg/nr )
1.2
1.6
1.6
1.6
2.5
2.6
3.4
3.6
4.2
4.4
4.4
4.5
4.8
5.0
5.0
5.1
5.9
5.9
7.2
8.2
b.
Men grouped according
to individual exposures
Probability
estimates
a. b.
0
0
0
0
0.0036
0.0069
0.0074
0.0077
0.0212
0.0420
0.0185
0.0484
0.0509
0.0448
0.0454
0.0456
0.1107
0.1000
0.1428
0.2095
0
0
0.0080
0.0140
0.0270
0.0354
0.0744
0.0486
0.0387
0.0373
0.0394
0.1608
Dust
Mean
1.10
1.72
2.21
2.68
3.20
3.70
4.18
4.71
5.19
5.91
6.95
9.47
concentration
(mg/m5)
Range
0.5-1.5
-2.0
-2.5
-3.0
-3.5
-4.0
-4.5
-5.0
-5.5
-6.5
-7.5
->/1.5
Number
of men
348
427
510
430
301
350
355
291
243
335
227
305
4122
-------
219
234567
Mean Dust Concentration (mg/m3)
8
9
10
Fig. 2. Bias arising from use of colliery mean data as estimates
of individual exposures. Pitted curves to estimates of
probabilities of developing category 2 or higher simple
pneumoconiosis during 35 years at various mean dust
concentrations. (Data from Table I.)
colliery mean data
mean concentrations experienced
by individuals
-------
220
12 ordered X-ray classifications as a random process in discrete tine
units* Table II shows, for example, the frequencies of observed changes
over 10 years among the 355 »en listed in Table I as having experienced
mean dust concentrations between 4*0 and 4*5 mg/a5 during that period.
Table II. Matrix of transitions over 10 years for 355
coalface workers exposed to mean dust
concentrations from 4.0 but less than
4.5
INITIAL
o/-
0/0
0/1
1/0
1/1
1/2
2/1
2/2
2/3
3/2
3/3
3/4
Total
RADIOLOGICAL CLASSIFICATION
FINAL
O/-. 0/0 0/1 1/0 1/1 1/2 2/1
1 2
214 25 6 1 1
19 11 4 2 1
10 5 4 2
1813
5 4
5
1 216 44 28 18 12 16
2/2 2/3 3/2 3/3 3/4
1
1 1
3 1
2 1
3 1 1
2
3
10 5 5
Total
3
247
38
23
13
13
8
5
2
3
355
If the numbers in each cell are divided by the corresponding row totals
then the resulting array may be regarded as an estimate of a matrix of
transition probabilities for the relevant time period. By hypothesis,
particular transition probabilities are functions of dust exposure
during the relevant transition period. The validity of the hypothesis
-------
221
can be tested using observed data; strong evidence consistent with it,
for ten-year periods, was produced by Jacobsen et al Qt7j and again by
Jaeobsen [9], Assume now that for a given dust exposure during a period,
transition probabilities are independent of earlier dust exposure. With
this assumption estimates of transition probabilities for integral
multiples of the observed tine periods are possible by using a simple
Markov-chain procedure. For a 20-year period the 1&-year matrix is
squared; for a 30-year period it is cubed, and so on. Again, if the
assumption is valid, then transition probabilities relevant to half an
observed time period can be estimated by finding a suitably structured
matrix which when multiplied by itself gives a product equal to the
original matrix.
4*3*2 Effect of thf* ae8"yption
Statistical independence of radiological transition probabilities
during consecutive periods with ^mUM1 dust exposures is, at first
sight, an unreasonable assumption, for at least two reasons* On the one
hand it can be argued that because radiological changes are associated
with accumulation of dust in the lungs the probability of no transition
during a second period is likely to be lower than in the preceding
period; the dust accumulated, but not detected radiologieally, during
the first period will pre-dispose individuals to showing radiological
changes on the claealficatoiy scale during the second period. On the
other hand it may be that some individuals are inherently less
susceptible to the development of radiological changes than others*
Probabilities of no change during the second period will then be higher
than in the preceding period, since the hypothesised less susceptible
individuals are more likely to constitute the group at risk during the
second period. Notwithstanding these two objections, the assumption
was made for the interim analysis. It was rationalised by appealing
firstly to the substantial observed differences in risks of radiological
change depending on the initial category. This factor presumably
reflects the combined effects of the two opposite tendencies and possibly
others. Given then that the data available covered a wide range of
previous exposures it was thought likely that a simple Markov model would
yield reasonable estimates of the probabilities required*
Subsequent examination of the results has confirmed that
probabilities of no radiological change over a ten-year period for a
-------
222
given dust exposure tend to decrease with increasing previous dust
exposure* Table III for example shows results for sub-groups totalling
308 of the 355 next considered in Table II* It will be seen for instance
that the percentages of sen who showed no radiological change fron the
two lowest categories conhined decreases fron 68/77 «= 88 percent for men
with less than ten years' previous exposure, to 42/53 = 79 percent for
nsn with previous exposures ranging from 20 to 30 years* The observed
proportions progressing radiologically show the expected trend with
increasing previous exposures, as do the prevalences in the three sets
of starting categories* «»<•»«!• trends were evident for other groups
exposed to different concentrations of dust during the ten years*
In principle, extrapolation of results beyond ten years should be
•ore reliable if differences depending on previous exposure are taken
into account* Table IT shows the matrix product [A] obtained by
successive multiplication of transition matrices corresponding to the
three sub-groups defined in Table III. Also shown for comparison is
[B], obtained by cubing the matrix corresponding to the pooled results
from all 355 men (Table II). The latter calculation, which requires
the approximation used for the interim analysis, is based on results
from, all men exposed to 4 to 4.5 mg/s5 over ten years, including 47 men
whose previous exposure to airborne dust exceeded 30 years* Using the
simplified model [fij k of the 5 probabilities of no radiological change
are lover than when previous exposures are taken into account. For this
range of dust concentrations ten of the 15 possible probabilities of
change were overestimated using the simplifying assumption. Analogous
comparisons of results from men who experienced other ranges of dust
concentrations during the ten years gave variable results (Table 7).
The example in Table IV concerning risks for men exposed to dust
concentrations between 4 and 4.5 mg/fe? is of particular interest because
this concentration range corresponds to the maximum consistent with
application of the current British coalmine dust standards.
5. Discussion
Serious problems of public health nowadays engender a continuous
research effort, even if this weak is conducted by different research
groups rather than one. Opportunities for reviews of available
scientific evidence concerning health hacards present themselves. Such
opportunities should be taken.
-------
223
Table III.
INITIAL
STATE
Matrices of transitions over 10 years for
308 coalface workers exposed to mean dust
concentrations from 4.0 tut less than
4.5 mg/m^; "by years of previous exposure
to dust.
FINAL
STATE
0/-AO/0 0/1 1/0 1/1 1/2 2/1+.
68
)
0
h
92 6 3
7321
3 5
2
2 2
1 2_
. — •
42 8 2 1
-,52 2
4411
5 1
2 3
10 _.
101
13
6
2
4
2.
129
53
16
10
6
5
10,
~~1 100
Years of previous
exposure to dust:
Total fo
77 98
less than 10;
100
78
10
5 10, but less
2 than 20;
3
2_
100
53
16
10 20, but less
6 than 30.
5
10_
100
-------
224
Table IV. Estimated transition matrices for a 30-year
exposure to dust concentrations from 4oO
•2
but less than 4.5 rag/m
B
: obtained from the product of
matrices based on data in Table III
: obtained by cubing the matrix based on
data in Table II.
W-
.6375
.6540
M-B-
.1940 .0866
.2356 .2572
.2000
.0210
.0962
.2000
.4167
.0116
.0337
.0500
.0833
.2000
.0493
.3774
.5500
.5000
.8000
1
.1477 .0839
.1250 .1900
.0822
0457
2156
2107
3318
.0231
.1151
.1130
.0687
.0569
.0496
.3543
.5941
.5995
.9431
1
-.0165
.0503
.1106
.0027
.0672
.1178
-.0247
-.1194
-.0107
.0849
-.0115
-.0814
-.0630
.0146
.1431
-.0003
.0231
-.0441
-.0995
-.1431
0
-------
225
A major lesson to be learnt from the results reported is that
inferences baaed on grouped summaries of environmental data nay give
general guidance on the form of a dose-response relationship, but
quantitative estimates may be biased. Careful examination nay indicate
the direction in which the bias is likely to operate* In the present
ease the indications were that the effect of the bias would be to
overestimate rather than underestimate risks to health. The error
incurred due to use of the simplifying assumption to extrapolate results
is relatively small. Cubing a matrix which represents a weighted average
of three sub-matrices (and one other) gave, as expected, a fairly good
approximation to the product of the three sub-ma trices. Moreover, the
approximate transition matrices were based on larger numbers of
observations, and therefore gave estimates of transition probabilities
with relatively lower variances. This is reflected in the better
correlation of these probabilities with increasing dust concentrations
(Table V).
Diagonals of both [VJ and (~BJ in Table IV demonstrate higher
probabilities of further radiological change for men who begin a period
with X-rays classified in categories other than the two lowest. The
phenomenon was noted in earlier analyses (Jacobsen et al fljj
Jacobsen J9]) and is now confirmed. It has been suggested by Oldham
jllj and Bosslter £12] that the higher risks for men with pneumoconiosis
initially may be due entirely to the nature of the scale of radiological
abnormality; no more sophisticated hypothesis is required to explain
the observed results. This argument in no way diminishes the practical
importance of the observation. So long as the 12-point scale is used to
record radiological abnormality, so long can classification into
categories other than the lowest be regarded as an index of increased
risk. A dust standard designed to restrict long-tent risks for new
entrants to the industry with no pneumoconiosis will demonstrably be
associated with much higher risks for older men who already show some
early signs of the disease. The Rational Coal Board's regular X-ray
surveys of all British miners £13) thus automatically constitute a
medical screening programme; this supplements routine dust sampling
designed to ensure compliance with, the standard.
-------
226
Table V. Estimates of probabilities of developing category 2
or higher simple pneumoconiosis during 30 years1
exposure to various dust concentrations
Range of dust
concentration
(mg/m )
0.5-1.5
-2.0
-2.5
-3.0
-3.5
-4.0
-4.5
-5.0
-5.5
-6.5
-7.5
^7.5
Total
a
b
by simple Markov chain calculation
using information on previous
exposure to dust.
a. b.
Simple Markov Initial state augmented
chain by years' previous exposures
N
548
427
510
430
301
350
355
291
243
335
227
305
4122
to dust
Mean dust
concentration
(mg/m5)
1.10
1.72
2.21
2.68
3.20
3.70
4.18
4.71
5.19
5.91
6.95
9.47
Probability
Estimates
a. b.
.0000 .0000
.0000 .0000
.0050 .0000
.0087 .0000
.0195 .0354
.0236 .0078
.0496 .0493
.0331 .0243
.0254 .0091
.0220 .0000
.0256 .0675
.1141 .1552
Mean dust
concentration
(mg/m3)
1.12
1.72
2.21
2.68
3.21
3.71
4.17
4.71
5.19
5.91
6.95
9.49
N
266
336
417
351
254
307
308
257
213
291
198
281
3479
N = number of men whose results over 10 years
contributed to probability estimates
-------
227
I am grateful to the many researchers who have conducted this
project during the last 22 years and to the miners who volunteered their
co-operation. Special thanks are due to ay colleagues at the Institute
of Occupational Medicine, particularly Mr. M.D. Attfield, and to the
Chief Medical Officer of the National Coal Board, Dr. J.S. Mclintoek,
for pennies!on to publish. Br, J.M. Rogan, Director of the research
from its inception till 1977, saw the potentialities and difficulties
associated with analysis of incomplete epidemiologies! data. He
initiated the interim study and guided it to its conclusion* Opinions
expressed in this paper are mine,
References
1. JACOBSEN, M«, BAB, S., WALTON, W.H. and ROGAN, J.M.
•New dust standards for British Coal Mines", Nature (Lend.)
227, 445 (1970).
2. CHAMBERLAIN, E.A.C., MAKOVER, A.D» and WALTON, W.E.
"Hew gravimetric dust standards and sanpling procedures for British
coal Bines", T*>i«i«»d Particles III. Edt W.H. Walton, pp. 1015-1050,
Uhvin Bros., Old WokLng. England (1971).
3. JACOBSSN, M. "The basis for the new coal dust standards'4,
Mining Engineer, 131, 269 (1972).
4. FAT, J.W.J.. and HAS, S. "The Pneumoooniosis Held Research of
the Hational Coal Board", V"*t QCCUP. HVK.. 1, 149 (1959).
5. BEDFORD, T. and WARNER, C.G. "Chronic pulmonary disease in South
Wales coalainers, II* Environmental Studies", Spec. Rep. Ser...
Med. Res. Coun. No. 244 (1943).
6. HAMMON, R.J. and WALTON, W.H. "The selectire sampling of
reanlrable dust" Tpfrtalad Particles *»d Vapours. Bd: C.N. Davies,
pp. 465-475, Pergamon, Oxford, England (1961).
7. JACOBSBN, M., RAB, S.. WALTON, W.H. and ROGAN, J.M. "The relation
between pneumoconiosis and dust exposure in British ooal mines",
Inhaled Particles III. Edr W.H. Walton, pp. 903-919, Onwin Bros.,
Old Woking, England (1971).
8. INTERNATIONAL LABOUR OPFICB. "ILO U/C International classification
of radiographs of pneumoconioses", Qccmoational Safety and HeaTtt
Series No. 22 (rev.), I.L.O., Geneva, Switserland (1972).
-------
228
9. JACOBSEH, H. "Progression of coalvorkars' pneuaoconiosia in
Britain in relation to environmental conditions underground",
Proceedings of the Conference on Technical Jteaarorea of Drwt
Prevention
-------
229
2. We have not studied our data systematically for a possible
"peak concentration" effect. Reisner (1971) has reported
such an effect among coalminers in the Ruhr, although this
was relatively small compared with the effect of cumulative
dust exposure. Rogan et al (1973) have reported an inter-
action between age and dust exposure on pulmonary function
(FEV10) among the same cohort of miners considered in my
paper, and this interaction is consistent with Reisner's
observation.
3. We are currently analysing our data to determine whether
and to what extent the quartz content of the dust affected
results.
References
REISNER, M.T.R. (1971). In Inhaled Particles III. Ed:
W. H. Walton, pp 921 - 931. Unwin, Old Woking, Surrey, England.
ROGAN, J.M., ATTFIELD, M.D., JACOBSEN, M., RAE, S., WALKER, D.D.,
and WALTON, W.H. (1973) Brit. J. Industr. Med.. 30, 217 - 226.
STOPPS (Canada)
Were particle size distribution measurements of the coal
dust in the various mines made? What was the relationship
between the particle size and the gravimetric measurements?
JACOBSEN (U.K.)
Particle size measurements were made on dust samples collec-
ted during the first 10-year periods of observation. The con-
centrations of airborne dust from these samples had been expressed
as the number of particles in the size-range 1 - 5 ,ug/m of sampled
air. Unfortunately, as reported by Rogan et al (1961), attempts
to use estimated parameters from the observed particle size dis-
tributions in order to derive corresponding mass concentrations
were not successful. This is why a series of field trials were
conducted to determine experimentally the relationship between
particle-count concentrations and mass concentrations in each of
the 24 collieries involved in the study (reference (2) in the
paper).
Reference
ROGAN, J.M., RAE, S. and WALTON, W.H. (1967) in Inhaled Particles
and Vapours II, Ed: C.N. Davies, pp. 493 - 508. Pergamon,
Oxford, England.
-------
231
VALUTAZIONE DELl'INQUINAMTO DA RUMORE PRODOTTO DA
UNA CENTRALE TERMOELETTRICA SU DI UN QUARTIERE
FRANCA MERLUZZI+, ANTONIO GRIECO , MARCO BERNA ,
ALBERTO DONZELLI+, NELLO MORRESI
+ Clinica del Lavoro, Universita di Milano, Italia
-H- Sezione di Acustica Applicata dell1Istituto di Pisica
Tecnica, Politecnico di Milano, Italia
RIASSUNTO
Lo seopo dell'indagine e state quello di valutare il distur-
J>o arrecato dalla trasmissione del rumore, e delle vibrazioni
meccaniche di una centrale termoelettrica sugli abitanti di un
quartiere delta Riviera Ligure. E' stata oondotta un'indagine
oggettiva attraverso la misur-azione diretta del rumore e delle
Vibrasioni a centrals teTmoelettrica ferma e in funzione, nonche
un'indagine "soggettiva" mediante un questionario delle opinioni
della popolazione del quartiere. Per garantire un elevate numero
di risposte e stato utilizzato un questionario "mirato" e "parte-
cipato"-
Sono stati recapitati a domicilio e consegnati a ciassun
abitante del quartiere di eta pari o superiore a 18 anni 358
questionari ohe sono stati totalmente compilati e quindi utiliz-
zati ai fini statistici, I questionari, opportunamente codifi-
cati, sono stati elaborati mediante il calcolatore elettronico.
La rumorositd nisurata di notte all'interno di 6 abitazioni
diverse, con le finestre aperte oscillava tra 43 e 32 dBA con la
centrale termoelettrica ferma e tra 4835 e 52 dBA con la centrale
in funzione. Le vibrazioni misurate sulla zona centrale di
un vetro di una finestra chiusa hanno presentato un massimo
di ampiezza per le frequenze comprese fra i 12 e 30 Hz. I
risultati della elaborazione statistic delle risposte del
questionario hanno permesso di concludere one il rumore e le
-------
232
vibrazioni della centrale termoelettrica rappresentano pep gli
abitanti del quartiere una fonte di disturbo talmente grave da
poter essere definite* "intollerabile".
Gli autori sottolineano I'importanza del questionario
quale strumento per la raccolta della soggettivita e per la
valutazione del grado di disturbo provocato da rumors.
ABSTRACT
The aim of the survey was to assess the effects of noise
and mechanical vibrations produced by a thermal generating
station on the inhabitants of a district in the Riviera Ligure
region. Apart from conducting an objective survey consisting
of direct measurements of noise and vibrations at the thermal
generating station under operational and non-operational con-
ditions , a "subjective" survey was carried out on the basis of
a questionnaire on the opinions of the surrounding inhabitants.
In order to ensure a large number of responses, an "active" and
"passive" questionnaire was used.
Three hundred and fifty-eight questionnaires were completed
by means of a door to door survey of all inhabitants over 18
years of age in the neighbourhood and were then used for stati-
stical purposes. After proper coding, these questionnaires were
processed by computer.
The noise measured at night inside six different types of
home with the windows open varied between 43 and 32dBA when the
generating station was not operating and between 48,5 and 52 dBA
when it was. The vibrations measured at the centre of closed
window panes were found to be most intense for frequencies from
12 to 30 Ez. Statistical -processing, of the replies contained
in the completed questionnaires led to the conclusion that the
noise and vibrations from the generating station were so distur-
bing to the inhabitants of the district as to be considered
"intolerable".
The authors consider the survey emphasizes the significance
of questionnaires as a means of collecting subjective opinions
so as to assess the degree of disturbance caused by environmental
pollutants, among which noise' is certainly to be included.
-------
233
1. Introduzione
Nella vastissima letteratura che tratta della tollerabilita del rumo
re c'e ormai accordo quasi unanime sul fatto che la valutazione del distur
bo da rumore non debba prescindere dal giudizio della popolazione interes-
sata (Cederloff e coll., 1963; Noise final Report, 1963; Robinson e coll.,
1963; Bolt e coll., 1964; Bishop, 1966; Kryter e coll., 1969). Nonostante
questo accordo unanime, ]e norme e le leggi attualmente esistenti che stabi-
liscono i livelli massimi di rumorosita tollerabili senza fastidio non so
no univoche. In Italia, in particolare, la normativa in questo campo e
particolarmente carente, ed estremamente esigue sono le indagini condotte
sulla popolazione per stabilire il quadro di disturbo corrispondente ai
vari livelli di rumorosita in diverse situazioni. Per questo motivo, e sta
ta condotta un'indagine sulla popolacione di un quartiere disturbato dal
rumore e dalle vibrazioni prodotte da una centrale termoelettrica (C.T.E.)
costruita nelle immediate vicinanze delle abitazioni.
2. Metodi
II quartiere presso il quale e stata svolta la nostra indagine e com
posto da circa 20 caseggiati di 3-4 piani ciascuno, situate alia perife -
ria di un piccolo comune della Riviera Ldgure e conf inante, da un lato,
con la restante parte del paese e, dall'altro, con una centrale termoelet
trica ivi insediata da 4 anni (vedi Fig. l).
Una strada e un muro di circa 4 metri dividono il quartiere dalla centra-
le termoelettrica che occupa una zona in precedenza costituita per la mag
gior parte da prato. H quartiere e abitato da circa 400 persone, rappre-
sentate da impiegati, operai, piccoli commercianti, marittimi. Una percen
tuale irrilevante degli abitanti lavora presso la centrale termoelettrica.
Gli appartamenti sono parte di proprieta, parte in affitto e parte a ri-
scatto. La costruz.ione, 1'acquisto, o comunque 1'occupazione di queste abi
tazioni ha preceduto di alcuni anni il sorgere della centrale nella zona
conf inante.
Altri insediamenti industrial!, costituiti da 3 fabbriche, si trovano a
circa un chilometro di distanza in linea d'aria: pertanto, prima dell'in-
sediamento della centrale, questa area poteva essere considerata come una
zona "residenziale", mentre ora e stata trasformata in zona "semi-indu
striale" o "mista".
Al fine di stabilire il grado di disturbo provocato dal rumore della cen-
trale termoelettrica, e stata svolta una duplice indagine: soggettiva,
raccogliendo mediante un questionario le opinioni di tutta la popolazione
del quartiere, e oggettiva, misurando direttamente il rumore e le vibra-
zioni nelle abitazioni.
-------
234
RICA
FIGURA 1 - Planimetria del quartiere dove e stata svolta
1'indagine. Le lettere dell'alfabeto indicano
i punti dove sono state effettuate le misure
fonometriche. I numeri indicano le zone in
cui e state suddiviso il quartiere.
-------
235
2.1. II Questionarig
E' stato utilizzato un questionario "mirato" e "partecipato" secondo
una metodica originale alia cui preparazione abbiamo collaborate nel corse
di precedent! indagini (Barlassina, 1972; Picotti, 1972; Grieco e coll.,
1973). Secondo questa metodica e stata preparata una bozza di questiona-
rio, successivamente presentata e discussa nel corso di due riunioni con
UU gruppo di rappresentanti del quartiere e quindi in una ulteriore riu-
nione collettiva con tutta la popolazione della zona interessata. La for-
na definitiva del questionario arricchita e verificata seguendo il metodo
sopra esposto e recante una serie di istruzioni rivolte ai destinatari, e
risultata composta da 44 domande, raggruppate in 4 sezioni.
la sezione I e dedicata a raccogliere notizie relative ai dati anagrafici
del soggetto (nome e cognome esclusi) e alia collocazione della sua abita
zione in rapporto con il quartiere e con la C.T.E. Le domande della II e
IH sezione sono rivolte ad evidenziare eventual! segni riferibili a di-
Sturbo da rumore o da vibrazioni, ad individuare la provenienza di tale
rumore e/o vibraaioni nonche il grado di intensita. A questo scope sono
stati utilizzati riferimenti a condizioni reali analitiche che consentiva-
no anche di stabilire vari livelli del disturbo. Per esempio, viene richie
sto se il rumore sia tale da interferire con la conversazione, con la con-
centrazione e con il riposo notturno, Infine con la IV sezione viene indci
gata la presenza o meno di una sintomatologia patologica soggettiva connes_
sa con alterazioni neurc—vegetative rapportabili verosimilmente al rumore
e/o alle vibrazioni.
Una oopia del questionario e stata consegnata ad ogni abitante del quarti£
re di eta pari o superiore a 18 anni, e successivamente raccolta dopo la
sua compilazione. In questo modo sono stati recapitati a domicilio comples_
sivamente 358 questionari che sono stati totalmente utilizzati ai f ini sta
tistici.
Per la elaborazione dei dati del questionario^ al fine di valutare la ril£
ranza della distanza della C.T.E. dal quartiere sul disturbo, e stato rite
nuto opportune suddividere il quartiere in 5 zone relativamente omogenee
per numero di abitanti e different! per distanza ed orientamento nei ri-
guardi della C.T.E. Quindi si e proceduto a valutare la eventuale esisten
za di diff erenze tra le zone e a determinare in quale delle 5 zone il di-
sturbo da rumore e/o da vibrazioni e la morbilita soggettiva fossero mag-
giori rispetto alle altre mediante il calcolo dei parametri che seguono.
A) indice di disturbo da rumore (I.D.R.): questo indice, inteso come enti-
ta dell'interferenza del rumore con le normali attivita fisiologiche e di
relazione (conversazione, concentrazione, sonno) e state calcolato somman-
do le risposte del questionario che inequivocabilmente confermano 1'ipote-
si di esistenza del disturbo. B) indice di disturbo da vibrazioni (I.D.V.);
questo indice, inteso come valutazione soggettiva della presenza e della
oitita degli effetti delle vibrazioni sulle cose, e stato calcolato come
il precedente sommando le risposte che confermano 1'ipotesi di esistenza
del disturbo da vibrazioni. C) indice di morbilita sogpettiva (I.M.S.):
-------
236
questo indice inteso come frequenza di sintomi patologici soggettivi, e
state calcolato sotnraando le risposte positive,
La elaborazione statistica del dati e stata effettuata mediante i seguenti
procedimenti. A) E1 stata calcolata,ed espressa in valori percentuali la
distribuzione delle frequenze e delle risposte per ciascuna delle domande
del questionario. B) E' stata effettuata una serie di tabulazioni crocia
te delle variabili eta, sesso e scolarita con le risposte alle domande re-
lative ai tre indici di disturbo considerati. C) E1 stato verificato se
rispettivamente I.D.R., I,D,V., I.M.S. subiscano variazioni in rapporto
con le 5 zone del quartiere in relazione con i piani di abitazione.
2,2. Misure fonometriche e vibrometriche
Le misure fonometriche e vibrometriche sono state effettuate dalla Se_
zione di Acustica Applicata dell'Istituto di Fisica Tecnica del Politecni-
co di Milano.
I rilievi di n"""re, eseguiti di notte in alcune abitazioni del quartiere
indicate nella piantina topografica (vedi Fig, l), sono stati ripetuti tre
volte: con la centrale termoelettrica ferma e le finestre aperte, con la
centrale termoelettrica in funzione e le finestre rispettivamente aperte
e chiuse. Sono stati usati: due misuratori di livello sonoro Bruel & Kjaer
tipo 2203 e 2204 dotati di capsule microfoniche Brflel & Kjaer tipo 4131,
registratore magnetico Nagra IV e uno Nagra III, un amplificatore Brflel &
Kjaer tipo 2606, un registratore di livello Brflel & Kjaer tipo 2305 e un
analizzatore di distribui-ione statistica Brflel & Kjaer tipo 4420, Le misu-
re sono state eseguite tra la mezzanotte e le 3 di mattina, nell'aprile e
giugno 1972, all'interne delle abitazioni prescelte, con il fonometro po-
sto a circa 2 metri dalla finestra aperta e chiusa. In ciascuna postazione
e stata fatta una registrazione magnetica dei segnali derivati dal fonome-
tro, adattato in scala "lineare" previa incisione del segnale di livello
sonoro campione, Successivamente, in laboratorio le registrazioni sono sta
te filtrate in curva A e trascritte su carta, ricavandone allo stesso tem-
po la distribuzione statistica dei livelli sonori e lo spettro sonoro.
Allo scopo di accertare la presenza di vibrazioni aeree immesse nel quartie
re dalla centrale in funzione, sono stati eseguiti dei rilievi vibrometrici
applicando un captatore accelerometrico sulla zona centrale della finestra
chiusa. IL segnale elettrico derivato e stato inciso su un registratore a
nastro magnetico a modulazione di frequenza. In fase di trascrisione, il
nastro e stato riprodotto a velocita quadrupla di quella usata nella regi-
strazione, e quindi analizzato con filtro selettivo e continue, Sono stati
usati un accelerometro Brflel & Kjear tipo 4330, un preamplificatore Brflel
& Kjaer tipo 2625, un registratore magnetico Brflel & Kjaer tipo 7001, un
amplificatore-analizzatore Brtiel 4 Kjaer tipo 2107, un registratore di li
vello Brflel & Kjaer tipo 2305.
-------
237
ZONA
1
2
3
4
5
I.D.R.
234 (6256)
185 (4950
151 (6350
271 (65$)
227 (7656)
I.D.V.
124 (58*)
105 (48*)
64 (SO*)
125 (S3*)
100 (61$)
I.M.S.
126 (5556)
87 (39*)
42 (37*)
88 (46*)
75 (5056)
TABELLA N° 1 -
Valori degli indici di disturbo da rumore (I.D.R.), di
disturbo da vibrazioni (I.D.V.) e di morbilita sogget-
tiva (I.M.S.) corrispondenti a ciascuna delle 5 Zone
del quartiere. Tra parentesi sono riportati i valori
percentuali rispetto al mimero totale delle risposte.
PIANI
Terrene
1°
11°
III0
ryo
I.D.R.
83 (49*)
237 (63*)
258 (57*)
288 (69*)
189 (68*)
I.D.V.
51 (47*)
115 (55*)
123 (53*)
136 (59*)
84 (55*)
I.M.S.
24 (23*)
96 (47*)
105 (43*)
114 (55*)
78 (54*)
TABELLA N° 2 -
Valori degli indici di disturbo da rumore (I.D.R.), di
disturbo da vibrazioni (I.D.V.) e di morbilita sogget—
tiva (I.M.S.) corrispondenti a ciascuno dei 5 piani
del quartiere. Tra parentesi sono riportati i valori
percentuali rispetto al numero totale delle risposte.
-------
238
3. Risultati
Per brevita non vengono riportati i dati analitici relativi a ciascuna
delle 44 domande del questionario. Nelle tabelle 1, 2, 3, 4, 5 e 6 vengono
riferiti i risultati piu significativi ottenuti dalla elaborazione stati-
stica del questionario*
Nella tabella numero 7 estrapolando e riassumendo dai grafici forniti dal_
la sezione di acustica applicata sono indicati i risultati delle misure fo
nometriche• n rumore della centrale e di tipo continue, costituito per la
maggior parte da basse frequenze con un massimo di intensita fra i 50-200
Hz; e possibile distinguere un rumore di fondo che oscilla di + 2,5 dBA
circa sul quale si inscrivono rari picchi di breve durata.
Per quanto concerne i rilievi vibrometrici, dagli spettrogrammi che sono
stati ottenuti si puo concludere per la presenza nella zona centrale di
una finestra chiusa di vibrazioni che hanno il massimo di intensita compre_
so tra le frequenze di 12 e 30 Hz.
I valori dell'accelerazione non sono stati rilevati.
4* Discussione
L'esame dei risultati emersi dal questionario attraverso la valutazione
della distribuzione delle frequenze delle risposte consente di stabilire
se il rumore e le vibrazioni rappresentano un reale fastidio per la popola
zione del quartiere. Tra le varie condizioni del quartiere ipotizzate come
causa di disturbo (la carenza di mezzi di trasporto, il tipo di vicini,
il rumore del traffico, il fumo e la polvere delle fabbriche, il rumore
della C.T.E., il chiasso dei bambini, le vibrazioni della C.T.E., etc.),
la rumorosita della centrale termoelettrica e indicata come la principale
dal 6S!t dell a popolazione, mentre il 6l# indica come tale le vibrazioni.
H 44J& delle persone indica nel rumore e il 1\% nelle vibrazioni provenien
ti dalla C.T.E. il fattore di disagio del quartiere sul quale vorrebbe in-
tervenire prioritariamente. la somma di questi due valori supera largamen-
te quella dei valori che si riferiscono a tutte le altre "cose da cambia-
re" considerate globalmente. II rumore costringe il 28# della popolazio-
ne ad alzare la voce nella conversazione in casa, disturba la concentra-
zione del 46j6 dei soggetti, costringe 1'86# a chiudere le finestre anche
d«estate. n 5&% della popolazione stenta a prendere sonno e nel 39#
dei casi la C.T.E. e indicata come la causa di ci&. Al 29£ delle per-
sone accade di ess ere svegliata improwisamente di notte, nell'80^ dei
casi a causa del rumore della C.T.E., il 69# delle persone svegliate non
riprende facilmente sonno. L*86£ degli abitanti dichiara di percepire la
esistenza nel proprio appartamento di vibrazioni e queste sono presenti nel
25JC dei casi "sempre" e nel 34)6 "spesso". L1 intensita delle vibrazioni e
cosi elevata che il 15# dei soggetti rif erisce di sentire vibrare il pavi
oento, il 1656 nota spostamenti di oggetti, il 28^ rif erisce di danni sia
pure modesti alle pareti e il 27% ai pavimenti. H (>1% della popolazione
soffre di insonnia, il 695f di cefalea, 1'8^6 di astenia mattutina, il 66^
di inappetenza, il 65J6 e giudicato dai f amiliari o amici "una persona ner-
-------
ZONE
1-2-3-4-5
I.D.R.
X*
55,50
g.d.l.
4
P
<0,05
I.D.V.
xz
8,71
g.d.l.
4
P
>0,05
I. M.S.
r
17,96
g.d.l.
4
P
<0,05
n
TABELLA N° 3 - Valori di X , del gradi di liberta (g.d.l.) e del livelli di probabilita (P)
per le 5 zone in rapporto a ciascuno del 3 indici.
Is)
U!
PIANI
T- 1-2- 3-4
I.D.R.
X*
29,52
g.d.l.
4
P
<0,05
I.D.V.
Xz
4,51
g.d.l.
4
P
>0,05
I. M.S.
r
35,49
g.d.l.
4
P
<0,05
TABELLA N° 4 - Valori di X *, dei gradi di liberta (g.d.l.) e dei livelli di probabilita (P)
per i 5 piani in rapporto a ciascuno dei 3 indici.
-------
240
vosa". Nella grande maggioranza del casi tali disturb! sono insorti negli
ultimi 2—3 anni. E1 risultato, inoltre, che le donne sono piu disturbate
degli uomini soprattutto per quanto riguarda il sonno, lo stesso vale per
la classe di eta intermedia (36-55 anni) e per coloro che dichiarano la
piu bassa scolarita. Dall'esame delle tabelle numero 1 e 3, in base ai va
lori degli indici I.D.R. e I.M.S. risulta che le zone piu disturbate sono
la 5, la 1 e la 4 rispetto alia 3 e alia 2. Per quanto riguarda 1'I.D.V.
non esistono differenze significative tra le varie zonej tuttavia le tabel
le 5 e 6 indicano che le zone 1, 5 e 4 sono le piu interessate dalle vibra^
zioni. L'esame della figura 1 dimostra che le zone piu disturbate sono le
piu vicine alia C.T.E. Sono state messe in evidenza differenze significa-
tive anche tra i 5 piani di abitazione per gli indici I.D.R. e I.M.S. (ve-
di tab. 2 e 4) per cui risulta che gli abitanti del 3° e 4° piano sono piu
disturbati di quelli del 1°, 2° e del pianterreno. Questo fenomeno e verc—
similmente in rapporto con 1'azione protettiva verso il rumore esplicata
dalle raura di cinta della C.T.E., che si frappongono tra questa e i piani
di abitazione piu bassi. Non sono state evidenziate differenze significati
ve tra i piani di abitazione per quanto riguarda 1'I.D.V.
I risultati'e le considerazioni esposti permettono di affermare che il di-
sturbo da rumore esiste inequivocabilmente, che la C.T.E* ne e di gran lun
ga la principale fonte responsabile, che le vibrazioni sono present! in un
gran numero di abitazioni, che esse costituiscono dopo il rumore la secon-
da causa di disturbo e che anch'esse provengono dalla C.T.E. I risultati
ottenuti dal questionario concordano con quanto emerso dalle misure fononre
triche. Infatti, anche con queste si dimostra che le misurazioni effettua-
te nel punto piu lontano dalla C.T.E. hanno fornito i valori piu bassi in
termini di intensita del rumore, sebbene le differenze tra le zone non sia
no significative (vedi tab. 7). Cio che invece risulta molto evidente e
1'incremento di rumorosita riscontrato con 1'entrata in funzione della
C.T.E., che varia da un minimo di 8 dBA ad un massimo di 18 dBA. Per quan
to riguarda la rumorosita nei diversi piani di abitazione, non e possibile
stabilire un confronto in quanto le misurazioni non sono state effettoate
in tutti i piani.
Le misure vibrometriche eseguite valgono soltanto a confermare 1'esistenza
delle vibrazioni denunciate dalla popolazione, e dimostrano che la loro
nassima intensita corrisponde a vibrazioni aventi una frequenza compresa
fra 12 e 30 Hz. A questo proposito, giova ricordare che le vibrazioni a
bassa frequenza (fino a circa 100 Hz) sono quelle capaci di provocare, a
parita di accelerazione e spostamento, i maggiori effetti biologici, in
quanto la frequenza di risonanza di numerosi organi ed apparati e compresa
entro questi valori (Collana C.E.C.A. di Igiene e Medicina del Lavoro,1967;
Scherrer e coll., 196?; Odescalchi, 1972).
Non e possibile giudicare se le vibrazioni present! nel quartiere siano do
tate di una accelerazione tale da mettere in risonanza organi e apparati.
Tuttavia, il fatto stesso che buona parte della popolazione ne percepisce
la presenza induce a considerarle una fonte non trascurabile di disturbo.
-------
241
ZONA
1
2
3
4
5
SI
62
25
29
46
55
NO
29
55
44
49
40
N.R.
9
20
27
5
5
TOT ALE
(79)
(85)
(56)
(81)
(57)
N = (358)
22,55
g.d.l. = 4
P < 0,05
TABELLA N° 5 - Domanda n° 30: Le vibrazioni hanno provocato secondo
te> danni alle cose?
Percentuali delle risposte e delle non risposte (N.R.) date dai sog-
getti divisi in base alia zona di appartenenza. Tra parentesi il to-
tale su cui e stata calcolata la percentuale.
ZONA
1
2
3
4
5
SEMPRE
20
15
13
30
51
SPESSO
41
25
27
40
33
QUALCHE
VOLTA
28
47
37
21
7
N.R.
11
13
23
9
9
TOT ALE
(79)
(85)
(56)
(81)
(57)
N- (358)
=49,41 g.d.l. P < 0,05
TABELLA N° 6 - Domanda n° 27: In case affermativo, le vibrazioni
sono present!:
Percentuali delle risposte e delle non risposte (N.R.) date dai sog-
getti divisi in base alia zona di appartenenza. Tra parentesi il to-
tale su cui e stata calcolata la percentuale.
-------
242
LUOGO BELLA
MISURAZIONE
A
B
C
D
B
F
CENTRALE TERMOELETTRICA
FERMA
FINESmE APBRTE
39
40
35
40
27
33
CENTRALE
IN
FINESTRE
APERTE
49,5
48,5
TERMOELETTRICA
FUNZIONE
FINESTRE
CKEUSE
38,5
39,0
52,0
51,0
48,0
50,0
35,0
32,0
33,0
TABELLA N° 7 -
Rumorosita misurata nei 6 appartamenti prescelti,
indicata come livello sonoro globale espresso in
dBA, con centrale termoelettrica ferma e in fun-
zione, con finestre delle abitazioni aperte «
chiuse.
-------
243
5. Conclusion!
Considerando globalmente le risposte alle domande del questionario e
possibile affennare che il rumore e le vibrazioni della C.T.E. arrecano di
sagio e disturbo a circa il 60-70$ della popolazione del quartiere.
Questa percentuale cosi elevata Indica chiaramente che il livello di rumo-
rosita presente e superiore alia tollerabilita della popolazione e che per
tanto deve essere considerate inaccettabile.
Questa conclusione alia quale siamo giunti dimostra I'importanaa fondamen-
tale dell'uso del questionario ogni qualvolta si debba giudicare del distur
bo arrecato da un inquinante ambientale che, come il rumore, alteri lo sta
to di benessere soggettivo senza determinare modificazioni oggettive fa-
cilmente obiettivabili. Eseguendo unicamente delle misure fonometriche e
confrontandone i risultati con una delle norme o raccomandazioni che rego-
lano questa materia, avremmo ottenuto dei risultati piuttosto incerti o
imprecisi.
Per esempio, con 1'applicazione della Norma ISO 1996 che e la piu usata in
Italia, e possibile giungere ad indicare come livello di ruraorosita ottima
le per il quartiere un valore compreso tra 25 e 50 dBA, a causa delle di-
verse possibilita di scelta fornite da questa norma. Avremmo potuto, pertai
to, dare dei giudizi estremamente contrastanti, non potendo basare la no-
stra scelta su dei dati concreti. D'altro lato, non riteniamo che la stes-
sa Norma possa essere usata indifferentemente per valutare nunori con ca-
ratteristiche cosi diverse quali, per esempio, il rumore del traffico ur-
bano e quello di una centrale termoelettrica. Infatti, il primo e estre
mamente variabile nel tempo, mentre il secondo e persistente e duraturo
per anni interi. Non e escluso che questo carattere di estreraa persistenza,
senza possibilita di soste, possa aumentare il grado di disturbo arrecato
alia popolazione.
Sulla base di queste considerazioni riteniamo che I1 uso del questionario
sia di fondamentale importanza nella valutazione della tollerabilita del-
1'inquinamento da rumore e che costituisca 1'unica garanzia affinche non
vengano accettati arbitrariamente dai tecnici livelli di rumorosita che
molto spesso rappresentano il risultato di un compromesso con una realta
che non si vuole modificare, piuttosto che la preoccupazione di salvaguar-
dare la salute e il benessere della popolazione.
-------
244
Bibliografia
Barlassina C. - Messa a punto di un metodo per la preparazione e la valuta
zione del questionario come contribute allo studio e alia
prevenzione della nocivita negli ambienti di lavoro.
Tesi di laurea - Facolta di Medicina -- Universita di Mila-
no, 1972.
Bishop D.E. - Judgments of the relative and absolute acceptability of
aircraft noise.
J. Acoust. Soc. Am. 40, 108-122, 1966.
Bolt H.H,, Beranek L.L., Newman R.B. - Judgment of the relative and absolu
te acceptability of actual and recorded aircraft noise*
Kept. n° 107, 1964.
C.E.C.A. - Effetti e debellamento dei rurnori.
Collana d'Igiene e di Medicina del Lavoro, n° 7, Lussembur
go, 1967.
Cederloff R., Jonsson E*, Kajland A. - Annoyance reactions to noise from
motor vehicles. An experimental study.
Acustica 13, 270-279, 1963.
Grieco A., Molteni G., Berna M.E., Piccoli B., Cesana G., Beggi P., Fannuz
zi A«, Basagni M. - Comunicazione al 36° Congresso Nazionale di Medicina
del Lavoro, Pugnochiuso (Bari) 8-10 novembre 1973*
Kryter K.D., Johnson P.J., Young J.R. - Judgment Tests of Flyover Noise
from Various Aircraft.
NASA Kept. CR-1635. Stanford Research Institute. National
Aeronautics and Space Administration, Washington, D.C.,
1969.
Noise Final Report. HMSO London, 1963.
Odescalchi C.P. - Ergonomia: elementi di igiene del lavoro.
5* Etas Kompass, Milano, 1972.
Picotti G. - Inchiesta operaia alia San Remo.
Rassegna di Medicina dei Lavoratori-supplemento al n° 4/
1972 dell'Assistenza Sociale - pp. 324-344, 1972.
Robinson D.W., Bovsher J.M., Copeland W.D. - On Judging the Noise from
Aircraft in Flight.
Acustica 13, 324-336, 1963.
Scherrer J. - Physiologic du Travail (Ergonomie)
Ed. Hasson, Parigi, 1967*
-------
245
MERCURY-SELENIUM ASSOCIATION IN PERSONS EXPOSED TO
INORGANIC MERCURY
L, KOSTA+, A, R, BYRNE++ AND V, ZELENKO+++
+ Faculty of Natural Sciences and Technology, University of
Ljubljana, Yugoslavia
++ J. Stefan Institute, University of Ljubljana, Yugoslavia
+++ Medical Faculty, University of Ljubljana, Yugoslavia
ABSTRACT
For the last feu years we have been measuring the distri-
bution of mercury in organs obtained from autopsies of persons
working in or living around the mercury mine of Idrija, Sloven-
ija. These investigations revealed an unexpected distribution
pattern with very high uptake by thyroid and pituitary glands
(by a factor of a thousand) accompanied by a very long biologi-
cal half life. Kidney, normally considered to be the prime
accumulator> ranked only third behind these glands. So far we
had eight cases of professionally exposed miners and there was
no exception to this pattern. In eleven subjects of the general
population of Idrija, there was a greater variability but thyroid.
pituitary and kidney still showed a considerable increase in
concentration as compared to nonexposed persons, with values
near background to factors of about fifty. Tables presenting
mercury concentrations for three population groups (professional,
population, controls) for many additional organs are given.
negligible methylmercury concentrations have also been found.
-------
246
1. Introduction
In spite of numerous animal studies, information
on inorganic mercury accumulation in humans is very
limited, as noted by Nordberg and Skerfving(l) in the
recent compilation of Vostal and Eriberg, and by
Berlin(2). Over the last few years, we have been ana-
lysing mercury in human organs from persons working in,
or living around the mercury mine at Idrija, Slovenija,
end comparing them with controls of no known exposure.
This is part of a general survey of the uptake and
distribution of mercury in the biosphere around the
mine which has included plants(3,4) and fungi(5)» ani-
mals (6) and the aquatic life of the contaminated river
Idrijca(4,6) and man(20). Since Parizek and co-workers
have shown(7»8) that selenium compounds exert a de-
toxifying effect (or in certain circumstances, a po-
tentiating effect) on the toxicity of mercury compounds,
and recent analytical results by Ganther et al(9) on
tuna and especially on marine mammals by Koeman et al(10)
indicate a close interrelation between the natural con-
centrations of the two elements in these species, we
have extended our studies to the occurrence of both
elements in the organs of exposed persons. Additionally
we have also determined the methylmercury contents of
some organs from mercury miners and the Idrija popula-
tion.
2. Experimental
2.1. Analysis
Total mercury was determined by neutron activation
analysis using a quantitative volatilization separation
of the ^Hg activity(ll). The volatilization techni-
que as applied to total mercury analysis has been
further developed(16) so that additionally selenium,
as 120 day '^Se, may also be quantitatively volatilized
-------
247
simultaneously from the same sample ana measured sepa-
rately from mercury. Methylmercury was analysed by a
new method(12) using conventional gas chromatograpnic
detection but with an isothermal distillation of vola-
tile methylmercury cyanide in a Conway cell, instead
of the usual Westoo extraction procedure(13-15)•
2.2. Sampling
With the exception of some thyroid tissue
excised by surgery, all samples were autopsy material
and thus, apart from the relatively small number of
subjects we have been able to examine, there was the
further disadvantage that little control could be
exercised over their distribution in terms of age,
exposure etc. Of the eight mine workers so far ana-
lysed, only one died while still employed at the
mine, the others having been in retirement for pe-
riods of from two to sixteen years.
Samples,stored at -25°C, were analysed in most
cases without any pretreatment to avoid contamination
problems or volatilization losses, the latter being
as especial hazard with both mercury and selenium
analyses. While for the larger organs (liver, brain)
this procedure can be criticised on the grounds of
inhomogenity, replicate analyses showed a surprisingly
small variation, and also such errors do not affect
the mercury to selenium ratio measured simultaneously
from the same aliquot.
3. Results and Discussion
3.1. Mercury Distribution
The average distribution of mercury in the three
groups, the highly exposed mine workers, the general
population of Idrija and non-exposed controls, is shown
in Table I.
-------
Table I. Average Hg content with standard deviation of human
organs in ppm fresh weight
G
r o u p :
Thyroid
Pituitary
Kidney
Liver
Lung
Brain
_l_ "V
55-2*28.5 27.1*14.9 8.4*4.9 0.26*0.25 1.11*0.89 0.70*0.64'
Idrida Population 0.70*0.45* 0.46*0.54 0.66*1.13 0.107*0.059 0.127*0.100 0.058*0.045*
Non-Exposed 0.030*0.0570.040*0.026 0.16*0.18 0.050*0.017 - 0.0058*0.0055 g
Controls (16) (6) (5) (6) ^'
Figures in parentheses refer to the number of subjects analysed,
+excluding P.M., Excluding T.A.
10
-p'
-------
249
Striking is the very high accumulation of mercury in
the thyroid and pituitary glands in exposed persons,
being a factor of about 100 greater that the popula-
tion group and about 1000 over the control group.
Previously, kidney has been considered to be the
prime accumulating organ, but in all eight mine
workers, it ranked only third behind thyroid and
pituitary. As noted above, seven of them had been
in retirement for periods ranging up to 16 years,
but the pattern of distribution of mercury in the
organs found in all seven and the other subject
(see Table II) was quite similar, with only a some-
what higher value in the liver of the latter, as
might be expected. Thus on the basis of the distri-
bution being virtually invariant with time since
cessation of exposure, it is clear that retention
of mercury in the organs with highest levels, namely
thyroid, pituitary, kidney and brain, is very strong
indeed.
As regards accumulation in the thyroid, though
there was some previously published evidence suggesting
its accumulating ability, its magnitude in man does
not seem to have been observed. Artagavayieta{19)
197
used Hg as a scanning agent for thyroid tumours
in man. Also Suzuki et al(17) found significant uptake
of mercury in the thyroid of rats and rabbits follo-
wing a single injection of mercury compounds, parti-
cularly inorganic ones, though only short time periods
were followed. We did not observe(Is) very striking
accumulation (50-100 ng/g fresh weight) in the thyroid
of rabbits exposed for long periods to natural mercury
contamination in Idrija. Investigation of mercury-rich
thyroid(20) showed that little mercury was contained
in extractable proteins, and there only in the highest
molecular weight fraction. A small scale survey(21)
of some other trace elements in thyroid did not reveal
any other correlations. Obviously in order to see the
effects of mercury accumulation in the thyroid, detailed
clinical investigations would be required involving
-------
250
Table II. Mercury content of human organs in ng/g,
fresh weight-basis
Subject
Group I
XB.A. ,
O.A.,
P. A.,
L.F.,
F.M;,
J.V.,
B.P.,
T.A. ,
B.A.
O.A.
P.A.
L.F.
F.M.
J.V.
B.P.
T.A.
Thyroid Pituitary Kidney Liver Brain Lung
- Hg mine workers
56y (5)
62y (10)
42y (0)
52y (2)
60y (5),l7y
59y (5),21y
7ly (16)33y
71y(16),29y
Blood Muscle
1.6 8-5
4.8 14.4
10.0
17.4 106
7.0
78
72
20
41300
21300
34500
19250
7800
29700
26500
101000
Spleen
25
23
188
64300X
53660
18000
14500
13800
23900
4110O
5500
9430
18530
6660
7510
5970
2320
11400
Heart Pancreas
18
98
36
42
352
108
120
791
184
360
370
37
65
Testis
295
37
526
181 127
1500 1580
791 614
1430
160
830
—
2400
Adrenal G.
223
238
age, years in retirement, years exposed to Hg, respectively.
-------
251
Table II continued
Subject
Thyroid Pitudtery Kidney
Group II
G.V.
K.F.
B.M.
S.Z.
B.S.
P.M.
D.I.
*
»
>
»
i
»
»
Liver
Brain
Lung
- Population of Idrija
62y, m
40y, m
75y, f
43y, f
I9y, m
7ly, f
0.3 y
730
958
676
122
69
14400
26
315
163
975
776
590
20
202
79
316
526
4000
760
610
124
63
131
48
184
67
85
39
15
24
76
21
110
44
244
168
122
42
68
66
Blood Muscle Spleen Heart Pancreas Testis Adrenal G,
G.V.
K.F.
B.M.
S.z.
B.S.
P.M.
3.7
3.6
8.5
4.2
17
12.2
4.9 11.8
17
10 20
24
15.1 8.7
14.8
13
17
36.6
20
21
-------
252
thyroid function, measurements related to size and
number of follicles and the statistical incidence of
malfunctions in exposed subjects.
The rather high values found in brain confirm
the data of Takahata(22) on two mercury mine workers.
Great variation between different sections and again,
of course, the extreme retention are noticeable;
this is also illustrated in Table 5 below. Also two
out of three miners showed rather high levels (0.3
and 0.5 ppm fresh weight) in the testes, in agreement
with the slow rate of excretion found in animal expe-
riments by Berlin(23).
3.2. Methylmercury
The main source of contamination for both miners
and the Idrija population are mercury vapour and dusts,
In addition to the high levels in soil (typically
50-500 ppm), the exhaust stack of the distillation
plant discharges 20-JO kg of mercury daily to the
Idrija atmosphere. We have examined some subjects from
both groups to determine the proportion of mercury
which is accumulated in the methyl form, in the case
of kidney, liver and brain.
As shown in the Table III, the absolute amounts found
from 3 to 30 ng/g fresh weight, in both groups are
low and very similar. Expressed as a fraction of the
total mercury, in the population group it nevertheless
amounts to 20 % on average in the liver and 26 % in
the brain. It can probably be concluded therefore,
that in vivo methylation does not occur, but that
the small amounts represent ingested foodstuffs con-
taining methylmercury (fish, meat, eggs).
-------
253
Table III. Total and methylmercury contents in
human organs in ppm fresh weight
Subject Kidney Liver Brain
Hg MeHg Hg MeHg Hg MeHg
Exposed Group
P. A.
L.F.
18.5
6.66
0.026
0.013
0.79
0.18
0.022
0.010
0.79
1.43
0.009
0.004
Idrija Population
G.V.
K.F.
B.M.
v V
s.z.
B.S.
0.20
0.079
0.52
0.53
4.0
0.006
O.OC5
0.019
0.010
0.011
0.12
0.063
0.15
0.048
O.ls
0.014
0.015
0.028
0.014
0.023
0.039
0.015
0.024
0.076
0.021
0.009
0.003
0.010
0.026
0.002
3.3- Mercury and Selenium
The results of the simultaneous analyses of
mercury and selenium from the same organ sample are
shown in Table IV.
A remarkable correlation is noticeable between the
contents for the organs with the highest mercury
values, namely thyroid, pituitary and kidney. The
weight ratios of mercury to selenium approximate to
the molar ratio (2.54) in these organs. The same
is true for brain as shown in Table V.
The correlation in subject T.A. is found in different
brain sections. Since selenium, as an essential element,
seems always to be present at levels at or above 0.1 ppm,
whereas mercury in non-exposed persons can fall to
almost zero, a near molar ratio will only be observed
with rather elevated mercury levels. However, even
slightly elevated mercury levels seem to result in
above normal selenium values.
Since we have not yet been able to analyse both
-------
Tat>le IV. Hg and Se in human organs in ppm fresh weight
Subject:
Group I -
Group II
Group III
X4 month
Thyroid
Hg Se Hg/Se
Pituitary Kidney
Hg Se Hg/Se Hg Se Hg/Se
Liver
Hg Se
Mine Workers
P.M. (mj
J.V. (m)
B.P. (m)
T.A. (m)
- Idrija
P.M. (f)
D.I.X
7-8 3.2 2.4
29.7 12.6 2.4
26.5 12.5 2.2
101 41.1 2.5
Population
14.4 5.72 2.5
0.026 0.16 - 0
14.5 6.1
13.8 6.4
27.7 13.3
47.9
0.59 0.59
.020 0.54
2.4 7.5
2.1 5-97
2.1 2.32
11.4
0.76
0.61
2.5 3.0
2.1 2.8
1.6 1.5
5.1 2.2
0.70
0.47
0.36
0.37
0.057
0.065
0.067
0.085
0.38
0.40
O.J6
0.38
0.21
0.12
- Controls
D.J. (m)
K.H. (f)
infant
0.039 0.79 - 0
.057 0.45
Or0023 0.46 - 0.044 0.34
F.M. exposed 17 years
J.V. exposed 21 years
B.P. exposed 33 years
T.A. exposed 29 years
, 5 years in
, 5 years in
, 16 years in
t 16 years in
0.57
0.011
retirement
retirement
retirement
retirement
0.82 -
0.54 -
0.059
0.010
0.19
0.17
in
P.M. aged 82, life-long Idrija resident (son a miner)
-------
255
selenium and mercury in a miner still exposed to mer-
cury, we should strictly speak of the effect as a
molar mercury-selenium retention ratio, though it
would seem likely that it will also "be observed in
man as an accumulative effect. Koeman et al(10)
observed the same molar ratio over more than two
decades of mercury concentration in the livers of
marine mammals. Here also the mercury was present
predominantly in the inorganic form. In our case, it
seems that both mercury and selenium are excreted from
liver rather more rapidly than the other mercury
accumulating organs, where very high levels are
retained after up to 16 years. (It is interesting to
note that, according to Iwata et al(23), Uzioka(24)
in I960 found high levels of selenium, as well as
inorganic mercury, in fish, cats and humans polluted
with methylmercury).
Table V. Eg and Se contents of human brain sections
in ppm fresh weight
Sample
Cerebellum
Striatum &.
cortex
Thalamus
Hypothalamus
Substans . nigra
Occip. cortex
Mod . oblongata
H
2.
2.
13
2.
8.
2.
6
42
34
.2
96
16
66
T
1
0
6
1
3
1
.A.
Se
.38
.93
.05
.43
.33
.15
Hg/Se H
1
2
2
2
2
2
.8
.5
.2
.1
.5
.3
0
0
0
0
B.P.
g Se
.15
.10 0.17
.073 0.17
.Ifl 0.23
P.K.
Hg
0.071
0.020
0.24
Se
0.17
0.15
0.25
-------
256
In attempting to explain this effect, the strong
mercury-selenium linkage should be considered, and
the protective effects demonstrated against mercury
compounds by selenium administration(7,8»23). The
correlation between mercury and selenium definitely
is causal, as is confirmed in the Idrija case by our
findings of normal levels of selenium in dusts and
soils (ca. 1 pptn). One can suggest that the effect is
a natural protective effect operating in conditions
of higher exposure. A selenium supplemented diet might
be indicated for workers exposed to high mercury le-
vels(8)« The detailed metabolic explanation must await
extensive further investigations, and we hope that
these analytical results will stimulate further research
into this phenomenon, which is clearly of considerable
importance in the toxicity of mercury.
References
1 Nordberg, G.F., Skerfving, S., Mercury in the
Environment, Priberg, L., Vostal, J., Editors, Ch.4,
CEC Press, Cleveland, Ohio (1972).
2 Berlin, M., Mercury Contamination in Man and His
Environment, pp 169-177, IAEA Technical Reports
Series No 137, IAEA, Vienna 1972.
3 Byrne, A.H., Kosta, L., Vestnik Slov.Kem.Drustva
17, 5 (1970).
4 Kosta, L., Byrne, A.H., Stegnar, P., Zelenko, V.,
Hadiotracer Studies of Chemical Residues in Pood
and Agriculture, pp 47-59, Vienna 1972.
5 Stegnsr, P., Kosts, L., Byrne, A.R., Ravnik, V.,
Chemosphere 2, 57 (1973).
-------
257
6 Byrne, A.R., Dermelj, M., Koste, L., Nuclear
Techniques in Environmental Pollution, pp 415-427?
IAEA, Vienna 1971 (N.B. Table VII should be jig/g,
not ng/g as printed).
7 Parizek, J., Oskadalova, I., Kalouskova, J.,
Babicky, A., Benes, J., Newer Trace Elements in
Nutrition, Mertz, W., Cornatzer, W.E., Editors,
pp 85-H9, Marcel Dekker, New York (1971).
8 Parizek, J., Nuclear Activation Techniques in the
Life Sciences, pp 177-194, IAEA, Vienna 1972, and
references therein.
9 Ganther, H.E., Goudie, C., Sunde, M.L., Kopecky, M.J.,
Wagner, P., Sang-Hwan Oh, Hoekstra, W.G., Science
175, H22 (1972).
1C Koeman, J.H., Peeters, W.H.M., Koudstaal-Hol, C.H.M.,
Tjioe, P.S., De Goeij, J.J.M., Nature 243, 385 (1973).
11 Kosta, L., Byrne, A.R., Talanta 16, 1297 (1969).
12 Zelenko, V., Kosta, L., Talanta 20, 115 (1973).
15 Westoo, G., Acta Chem.Scand. 20, 213 (1966).
14 idem, ibid, 21, 1790 (1967).
15 idem, ibid, 22, 2277 (1968).
16 Byrne, A.R., Kosta, L., Talanta, 1974, in press.
17 Suzuki, T., Miyarna, T., Katsunuma, H., Ind.Health
4^ 69 (1966).
1« Stegnar, P., Byrne, A.R., Kosta, L., Bioloski
Vestnik 21, 29 (1973).
19 Artagavayieta, D., Degrossi, O.J., Pecorini, V.,
Kucl.Med. 9, 350 (1970)
-------
258
20 Kosta, L., Zelenko, V., Stegnar, P., Havnik, V.,
Dermelj, A., Byrne, A.B., Isotope Tracer Studies
of Chemical Residues in Pood and the Agricultural
Environment, pp 87-102, IAEA, Vienna 1974.
21 Kosta, L., Zelenko, V., Eavnik, V., Levstek, M.,
Lermelj, M., Byrne, A.H., Comparative Studies of
Food and Environmental Contamination, pp 541-550,
IAEA, Vienna 1974.
22 Takahata, N., Hayashi, H., Watanabe, B., Anso, T.,
Folia Phychiatr.Neurol.Jap. 24, 59 (1970).
23 Iwata, fl., Okamoto, H., Ohsawa, Y., Hes.Commun.
in Chen.Pathology and Pharmacology £, 673 (1973).
24 Uzioka, T., Kumamoto Igaku Z. 34, 383 (in Japanese)
DISCUSSION
BRAMAN (U.S.A.)
What analytical method was used for methylmercury analysis
of tissues?
BYRNE (Yugoslavia)
Full details are described in reference (12). In brief,
an isothermal distillation of methylmercury cyanide, generated
in situ in the tissue, was employed. The volatile product was
trapped on a ring of cysteine impregnated paper in the outer
compartment of a Conway dish. Following acidification and ex-
traction into benzene, the usual gas chromatographic detection
was used.
PFANNHAUSER (Austria)
Have you any indications which kind of proteins bind mer-
cury? Is it conceivable that detoxification by selenium is
based on a greater binding of Se to these proteins.
-------
259
BYRNE (Yugoslavia)
It is generally considered that mercury is bound on the
sulphur containing ligands of proteins. The molar Hg/Se ratio
suggests that mercury is bound to selenium, but the nature of
the attachment to the protein is uncertain.
CHAMBERS (Ireland)
I am not quite sure of the source of the material for the
analysis of the metals in the organs. Were the organs taken
biopsy specimens or were they post-mortem specimens? If the
latter case is the source, have any comparisons been made with
biopsy material?
BYRNE (Yugoslavia)
They were all post-mortem specimens with the exception of
most of the thyroid samples in the control group. In the latter
group of samples no differences were noted between biopsy and
post-mortem samples.
de BRUIN (Netherlands)
Three points arise:
1. Was there any exogenous source of Se contamination present
in your investigation, or did the interaction concern only an
influence of naturally occurring Se upon the distribution of
exogenous Hg in the human organs?
2. Is there any evidence of thyroid and or generalized endo-
crinal disturbance in people working in or living around the Hg
mine?
3. Concerning the type of interaction Hg/Sej has Se a protec-
tive or rather synergistic effect upon Hg toxicity?
BYRNE (Yugoslavia)
1. As stated in the paper, selenium levels were found to be in
the normal range for the natural environment.
2. Not to our knowledge, though this should be investigated.
3. The interaction of Se and Hg is complex and can be of
either type, see for example, the work of Parizek and co-workers
referred to in the text.
-------
260
WEBER (Yugoslavia)
1. What is, in your opinion, the point of having the results
of mercury analyses given with apparently high precision when
the notion of fresh or wet weight of organs in general is a
rather undefined concept?
2. Have you any idea why mercury is accumulated in the thyroid
gland?
BYRNE (Yugoslavia)
1. If selenium is also simultaneously analysed in the same
sample, precise values are necessary to give a precise ratio*
In general high precision is unnecessary, though still meaning-
ful if the water content is also measured.
2. We have postulated an association with iodine and its meta-
bolic cycle in the body. The scheme, involving the thyroid and
pituitary glands, is described in more detail in reference (20)
of the paper.
HIKE (U.S.A.)
1. Why did you not analyse Hg and Se in tissues with higher
cysteine content such as the skin and hair?
2. Were the values expressed in ppm by weight or milliequiva-
lents per mass?
BYRNE (Yugoslavia)
1. Because of the strong likelihood of external mercury contam-
ination, we consider analyses of such samples to be unreliable.
Also selenium containing shampoos are in common use.
2. All results are expressed as ppm or ng/g on a fresh weight
basis.
CLEMENTS (Italy)
Was the blood concentration of Se measured? In fact the
ratio of Hg/Se in blood can indicate whether the Hg/Se ratio of
2.5 is due to an accumulation or to particular retention of Se
in man in the presence of Hg concentrations.
BYRNE (Yugoslavia)
Yes such measurements were made. For mercury the blood
values were in the normal range because of the interval between
cessation of exposure and measurement. We hope to measure the
ratio for miners still exposed.
-------
UNTERSUCHUNG DER WIRKUNGEN AUF DEN MENSCHEN
HUMAN EFFECTS STUDIES
ETUDES DES EFFETS SUR L'HOMME
STUDI DEGLI EFFETTI SULL'UOMO
ONDERZOEKINGEN NAAR EFFECTEN BIJ DE MENS
Panel
Vorsitsender -Chairman - President - Ppesidente - Voorzitter
F.G. HUETER (U.S.A.)
-------
263
METHOD FOR A EUROPEAN STUDY ON
POSSIBLE EFFECTS OF AIR POLLUTION IN CHILDREN
PH, BOURDEAU
on behalf of the Commission's Working Party of experts on
epidemiology.
Commission of the European Communities, Directorate General for
Research, Science and Education, Brussels, Belgium.
ABSTRACT
A coordinated epidemiological survey on the relationship
between air pollution and incidence and prevalence of respira-
tory affections in children has "been planned with the partici-
pation of about ten institutes in the Member States of the European Communities.
This broad participation will provide results from a very
large number of observations obtained in a significant number
of geographical areas categorised according to their air pollu-
tion characteristics.
This survey will be carried out by means of a questionnaire
(on health and socio-economic status of children) administered
by fieldworkers, of physical measurements (height, weight, ven-
tilatory function) performed on children in the schools, and
air pollution measurements, in particular SOg and suspended
particulate matter.
The protocol for the survey has been developed by a Working
Party of experts on epidemiology set up in 197S to advise the
C.E.C. on its Environmental Research Programme in this area.
A pilot study will be first undertaken on a limited number
of children followed by the main study on 20,000-30,000 children.
-------
264
1. Introduction
Epidemiological surveys are essential for the understanding of
the health effects of air pollution. Investigations on the relative
influence of individual pollutants are particularly useful.
Although it has been shown that air pollution influences incidence
and prevalence of respiratory symptoms and disease, as well as
levels of ventilatory function in children and adults, very little
is known about the independent effects of individual pollutants such
as for instance smoke and sulphur dioxide. Areas with very con-
trasting degree of air pollution will be chosen in the Member States
of the European Communities, in which appropriate investigations may
be carried out.
2. Purpose of investigation
To determine the influence of air pollutants on respiratory
symptoms, respiratory disease and ventilatory function in school
children.
3- Methods
3.1 This survey will require a questionnaire designed to
determine the incidence of past respiratory disease and present res-
piratory symptoms. The questionnaire will also need to elicit
details of social and occupational factors of the family that could
influence the incidence and prevalence of respiratory symptoms and
disease and the levels of ventilatory function, as well as to aid
conparability. It should, therefore, contain questions concerning
the child's health, (e.g. occurrence of cough and/or breathing
difficulties) and the socio-economic status of the family (e.g. type
of housing and type of work of parents).
The questionnaire will be completed by the fieldworkers inter-
viewing the child's parents at school and/or at home. By prefer-
ence the mother of the child should provide the answers.
3.2 In addition the following physical measurements on the
individual child will have to be made: height, weight and ventilat-
ory function* Instructions for such measurements as well as for the
administration of the questionnaire will be given in a Manual to be
used by the fieldworkers.
-------
265
4. Areas
The sites for investigation (a total of 20-25 areas) should be
chosen according to the type of pollution. They should be character-
ized by very contrasting levels of air pollutants, mainly SO and
suspended particulate matter. Rural and low-polluted urban areas
should also be considered.
It would also be of interest to locate areas polluted by other
substances (e.g. NOxl NH3, photochemical oxidants, etc...).
The characterization of atmospheric pollution in the areas to be
selected should be based on regular measurements performed over a
period of at least three years. The concentrations of atmospheric
pollutants should be determined through comparable methods in the
frame of air mgnitoring programmes of regional networks.
The measurements could be completed during the implementation of
the survey by means of monitoring instruments installed in the schools.
Because of the importance of climatic factors, meteorological
observations, should also be made during the investigations (temper-
ature, wind rose, relative humididty etc...)
It is essential that the selection of the areas where to conduct
the epidemiological survey be made in cooperation with experts on air
monitoring.
Potential contractors will have to provide information on air
pollution measurements currently made in the proposed areas by means
of an appropriate form. Instructions regarding the air pollution
measurements to be carried out in the proposed areas and the results
to be reported are contained in a separate protocol.
5- Population to be studied
In each area where the study is to take place, 1.000 - 1.500
school pupils aged six years or more, and less than eleven years at
the time of the beginning of the survey should be the object of the
investigations. Both sexes are to be included in about equal pro-
portion.
-------
266
6. Recruitment and training of personnel
The persons (fieldworkers) who will be in charge of the adminis-
tration of the questionnaires and the execution of the physical
measurement* during the survey should be, by preference, recruited
among nurses, or student nurses, and/or social workers.
Uniform training procedures should be established within each
country for the recruited personnel.
In order to be quite certain that the measurements performed in
different areas are comparable, only one team of fieldworkers should
be responsible for the measurements in all the individual projects
in a given country. It would be advisable to set up a joint train-
ing session for such responsible personnel from each country.
The administration of the questionnaire will be carried out by
the two fieldworkers assigned to single areas.
The work of the teams in the various countries during the
period of survey is to be supervised by at least two of the above
mentioned responsible fieldworkers* Thqrwill re-measure, with the
local team, a sample of the children in order to make certain that
the results which have been obtained are comparable.
The work schedule to be established for the fieldworkers should
include periods of: - training and familiarization
- interviews (administration of questionnaire)
- physical measurements
- refresher courses
- additional interviews and measurements (cases
of dubious answers and incorrect results)
The number of weeks required for each period may vary from
country to country according to qualifications af recruited personnel,
working conditions and families and children responding promptness.
Since variations in ventilatory function, symptoms, etc... occur
according to season, as short a time as possible should be spent on
field observations (physical measurements and interviews to fill in
the questionnaires).
-------
267
?• Data handling programme
7.1 Data collection
For each area : - gathering of the results of physical measure-
ments (recorded on special forms) and filled
questionnaires for each child
- report on air pollutants measurements per-
formed by local air monitoring stations
(including data for at least three years pre-
ceding the beginning of the survey) and/or
by means of instruments installed in the
schools
- report on meteorological observations made
during the investigations (temperature, wind
rose, relative humidity..*)
7.2» Data analysis
The data from each area of each country, transferred locally on
punch cards| are processed in a central place :
- analysis of questionnaire answers by age-sex-area categories
- analysis of results of lung function measurements by age-sex-area
categories. Means, frequency distribution, s.d., s-e. etc...
- analysis of air pollution and meteorological data
- establishment of the relationship between air pollution character-
istics and respiratory symptoms and disease in children.
-------
269
PRESENTATION ET DISCUSSION DU PROTOCOLE D'UNE
ENQUETE EPIDEMOLOGIPUE FRA.NCAISE SUR LES RELATIONS
ENTRE POLLUTION ATfOSPHERIOUE ET AFFECTIONS
RESPIRATOIRES CHRONI6UES
D, BRILLE , P, BOURBON , J, LELLOUCH , P, ORIOL
+ Unite de Recherches de Physio-pathologie Respiratoire (U.68)
INSERM, HSpital Saint-Antoine, Paris, France
+4- Unit£ de Recherches de Pollution atmospherique (U.57) INSERM
Vigoulet Auzil pres Toulouse, France
444 Unite de Recherches Statistiques (U.21) INSERM, Villejuif,
France
Etude pr6par6e avec les responsables de I'enque'te dans les 8
centres;
- Bordeaux; P. FREOUR, J. F. TESSIER, M. BERNADOU, J.G. FAUGERE,
J. DUFOIR
- Lille; M. GERVOIS, G. DUBOIS, C. VOISIN, HOUDRET
- Lyon; S. PERDRIZET, J. BOURDEIX, J. VIOLET, N. GALLY
- Mantes; D. BRILLE, P. ORIOL, F. KAUFFMANN, Cl. BOUDENE,
J. GODIN
- Marseille P. LAVAL, J. CHARPIN, A. ARNAUD, J. P. KLEISBAUER,
et Fos ; V. VIALA, M. J. SOMMER
- Rouen; J. P. LEMERCIER, A. LECADET, FLAUGNATI
- Toulouset R. BOLLINELLI, Y. ROUGH, P. BOURBON, J. ALARY
RESUME
A la demands du Ministere de la -Protection de la nature et
de I'Environnement et en collaboration avec le Minist&re de la
Sante Publique et de la Securite Sooiale, une etude epidemiolo-
gique sur les relations entre pollution atmospherique moyenne
et affections respiratoires ohroniques ou d. repetition, a ete
-------
270
entreprise en France. Le protocole et le$ methodes de travail
ont et& 6laboves en commun par 7 equipes travaillant dans 8
centres.
28 zones de 1000 sujets environ ont ete constitutes (2 d 4
zones par centre) en tenant compte de plusieurs facteurs:
- importance de la pollution anterieurement connue en aci-
dite forte (exprimee en SO-) et en fumees no-ires, de facon a
obtenir des zones aussi contrastees que possible. A ces mesures
ont ete ajoutees celles du dioxyde de soufref d'oxydes d1'azote3
des poussi&res par gravimStrie.
- existence d'une population suffisante en fonction de cri-
teres choisis: categories 8ocio~professionnellesf age, duree
de re&idencet nationality.
Des renseignements sur lf£tat respiratoire des sujets, leurs
antecedents, leur niveau de vie ... sont recueillis d I'aide d'un
questionnaire et d'un examen spirographique par des enquSteurs,
d domicile.
QueIques points du protocole sont discutes et notamment la
difficult^ de trouver des populations socialement comparables
habitant des zones peu etendues et fortement contrastees au
point de vue pollution.
ABSTRACT
An epidemiological survey on the connection between mean
atmospheric pollution levels and chronic or recurring respira-
tory ailments has been undertaken in France at the request of
-------
271
the Ministry for the Protection of Nature and the Environment
and the Ministry of Public Health and Social Security. The
report and working methods were drawn up jointly by seven teams
working in eight centres.
Several factors were considered in the choice of 28 zones
of approximately 1000 test subjects (2 to 4 zones per centre):
- the extent of previously known pollution from strong aci-
dity (expressed in SO ) and black smoke so as to have zones which
contrasted as far as possible. In addition, measurements for
sulphur dioxide and nitrogen oxides were included, as well as
gravimetrical dust counts.
- presence of sufficient numbers of population corresponding
to selected criteria: social and professional groups, age,
length of residence, nationality.
Data on the test subjects ' respiratory condition, past his-
tory, standard of living, etc. are collected by means of a
questionnaire and a spirographic test carried out at the subject's
home by research workers.
Certain aspects of the report are discussed, notably the
difficulty in finding socially comparable population groups
living in zones which are not too widespread but at the same
time are strongly contrasted with regard to pollution levels.
-------
272
I. INTRODUCTION
En 1973, a la demande du Ministere de la Protection de la Nature et de
1'Environnement, en collaboration avec le Ministere de la Sante Publique et
de la Securite Sociale, il a ete decide d'entreprendre une etude sur les
relations entre la pollution atmospherique moyenne et les affections respi-
ratoires chroniques ou a repetition par la methode epidemiclogique.
Le principe de 1'enquete est le suivant :
- choisir des zones dont la pollution est de type et de niveau moyen
different ;
- examiner sur le plan respiratoire (les voies respiratoires etant
considerees comme un reacteur privilegie) des populations residant dans
ces zones.
Cette etude est conduite dans 8 centres (villes de Bordeaux, Lille,
Lyon, Mantes, Marseille, Rouen, Toulouse et region de Fos).
Une telle etude se heurte a 3 types de difficulty's :
- complexite du facteur pollution ;
- multiplicite des facteurs etiologiques des affections respiratoires ;
- mise au point d'un protocole comraun et harmonisation des methodes
de travail de 7 equtpes differentes.
Le protocole de cette enquete est presente ici et certains points
discutes.
2. IA POLLUTION
2.1 Choix des postes
Toutes les villes disposaient avant le debut de 1'enquete d'un
reseau de mesure de 1'acidite forte (exprimee en SO^) et des funices noires
(postes "SF"). Les mesures faites en 1971 ont permis de determiner des
points aussi contrastes que possible en qualite et en quantite, permettant
des comparaisons entre les differents centres et quelquefois dans un centre
entre differentes zones (Bordeaux, Lyon, Marseille).
Les moyennes mensuelles se situaient entre 20 et !70i; g/m^ pour le
S02 et entre 30 et 200 pour les fumees noires.
2.2 Choix des polluants
Pour essayer de preciser le facteur "pollution", il a ete decide
-------
273
d'adjoindre aux mesures de 1'acidite forte et des fumees noires considerees
comme des "indices de pollution" des mesures du dioxyde de soufre par la
methode "Bourbon-Malboscq", des oxydes d'azote (NO, N02) par la methode
"Griess-Saltzman", des poussieres par gravimetric.
Les problemes propres aux mesures, particulierement ceux concer-
nant leur hannonisation, feront 1'objet d'une prochaine communication.
2.3 Representativite des points de prelevement
II est difficile de dire dans quelle mesure la pollution determi-
nee aux points d'emplacement des postes est representative de la pollution
d'une zone. Tout au plus, peut-on affirmer que 1'etendue des zones que
1'on veut ainsi caracteriser, doit etre aussi limitee que possible. II
reste que les personnes habitant dans ces zones sont inegalement soumises
a cette pollution : migrations quotidiennes, temps passe a 1'interieur des
maisons ...
3. LA POPULATION
La population a ete choisie en fonction de deux exigences :
- habiter a proximite de postes de mesure interessants par leurs
resultats ;
- avoir une certaine homogeneite sociale.
28 zones de 1000 sujets environ, presentant les critires d'admission a
1'enquete, ont ainsi ete constitutes {234 zones par centre) a partir des
donnees demographiques du recensement general de 1968 (INSEE).
La selection des menages, et a 1'interieur des menages des sujets,
necessite un recensement exhaustif de la population pour lequel un
questionnaire a ete etabli.
3.1 Criteres de selection des menages
Les menages sont selectionnes en fonction du chef de menage0 ;
celui-ci doit etre age de 25 a 59 ans, de nationalite franchise (pour evi-
ter des difficultes linguistiques dans une enquete par questionnaire),
resider dans la zone depuis 3 ans au moins (duree minimum pour pouvoir
juger des effets d'une pollution moyenne). Le chef de menage doit apparte-
nir a 1'une des categories socio-professionnelles0 suivantes : "patrons
"Definitions de 1'Institut National de la Statistique et des Etudes
Economiques (INSEE)
-------
ENQUETE EPIDEMIOLOGIQUE COORDONNEE EN FRANCE
a) 'Nombre de sujets admls pour la pre'-enquete
Menaces,
vijsit 6s :
6 175
" Admit
Manages
681 °
» a I'enquete"
Sujets
1 452 C ' 2" Adult"
( 227 Enfant s I
!
0 N'Staient admis que les manages dont le chef Stait ou cadre
moyen ou employ6 (C.S.P. 4^ ou 5^ de 1 'INSEE) .
Is)
-•J
b) Folds dee difference critSres dans I'elimination :
- dea menages :
- dee sujets
. Nationalite du chef de manage -
. Anciennet€ de residence 3 ans •
. C.S.P. du chef de manage . . -
• Age
5 Z
20 %
79 2
32 Z
TABLEAU 1 : RESULTATS du RECENSEMENT de la POPULATION lors
de la PRE-ENQUETE dans CINQ CENTRES.
-------
275
de 1'Industrie et du commerce", "professions liberales - cadres
superieurs", "cadres moyens - employes", "personnel de service".
3*2 Criteres de selection des sujets
L1etude porte sur les adultes des deux sexes, ages de 25 a
59 ans et leurs enfant de 6 a 10 ans. Ces deux tranches d'age
constituent, separement, des groupes relativement homogenes. De plus,
les enfants de 6 a 10 ans, en grande majorite non fumeurs et scolar-
ises pres du domicile, sont essentiellement soumis a la pollution
atmospherique de la zone etudiee.
3«3 Discussion du choix de la population
Le choix de la population est le seul point de protocule que
les enseignements de la pre-enquete ont fortement marque et qui
merite discussion.
Cette population n'est pas aussi homogene que souhaite*
Initialement, il avait ete prevu de ne retenir que d«ux categories
socio-professionnelles voisines : "cadres moyens" et "employes".
One zone de pre-enqu€te comportant 1000 menages environ a ete etudiee
dans chaque centre. La ou 1'INSEE avait recense 1000 menages dont
:200 "cadres moyens - employes", comprenant 600 sujets, les criteres
de selection propres a 1'enquSte (duree de residence, age, nationalite)
n'ont permis de retenir que 100 menages et 200 sujets en moyenne.
Le tableau 1 montre les resultats du recensement des menages obtenus
lors de la pre-enque"te conduite dans 5 centres. Le faible rendement
eat couteux en temps et en argent et entraine une extension de la
surface de la zone.
Ces constatations ont conduit a elargir les categories socio-
professionnelles retenues; ceci introduit un facteur d'heterogeneite
qu'il faudra prendre en compte au moment de 1'analyse.
Cependant, certaines categories, comme les "ouviers", n'ont
pas e6e retenues car la pollution professionnelle est parfois bien
superieure a la pollution atmospherique generale et, par sa diversite,
aurait introduit un biais difficilement contrdlable sans une etude
precise du lieu de travail (entreprise, poste de travail ...).
D'autres categories ont ete eliminees car inexistantes en ville
("agriculteurs") ou tres heterogenes ("personnes non active", "autres
categories").
-------
ENQUETE EPIDEMIOLOGIQUE COORDONNEE EN FRANCE
four
6 175 menage B
visitis
dans 5 centres
C. S. P.
I.N.S.E.E.
4-5 '
2-3-4-5-7 °°
Menage a
admis a
1 'enquete
681
11 %
1 204
19 %
Sujets
Adultes
1 225
2 209
admis H 1 '«
Enfant 8
227
412
mquete
Total
1 452
2 621
* Cadrei noyens et Employes
*• " " " + Patrons de 1'Industrie et du Commerce,
Cadres superieurs et Professions liberales, Personnels de service.
- sont exclus : les Ouvriers, Agriculteurs, divers.
TABLEAU 2 : NOMBRE de SUJETS qui AURAIENT ETE ADMIS par
ELARGISSEMENT des CATEGORIES SOCIO-PROFESSIONNELLES
-------
277
Avec les nouvelles categories socio-professionnelles admises a
1'enquSte, les renseigneoents collectes lors de la pre-enquSte
montrent que pour 6 175 menages visites, 1 20k auraient ete admis,
soit 2 621 sujets (tableau 2).
Get elargissement permet de doubler, pour une meme surface,
le nombre de sujets ou pour un me me nombre de sujets, de diminuer le
rayon de la zone de 30$.
4. RECUEIL des DONNEES CONCERNAHT les SUJETS ADMIS a 1'ENQUETE
Les donnees concernant les sujets admis a 1'enquete sont recueillies a.
domicile a. I1aide d'un questionnaire et d'un examen spirometrique simple
qui permettent d'obtenir des renseignements sur 1'etat respiratoire et sur
les facteurs etiologiques a prendre en compte au moment de 1'analyse.
4.1 Etat respiratoire du sujet
4.1.1 Un_qu_ej^ti^oimaire_dj;rJAr£ du questionnaire CECA pour 1'etude
de la bronchite chronique et de 1'emphyseme pulmonaire est
utilise. En fonetion des buts propres a 1'etude, que1ques questions ont ete
supprimees (ainsi une seule question est posee sur la dyspnee d'effort),
d'autres ont ete ajoutees, notatmnent sur les voies aeriennes superieures.
Un questionnaire du meme type a ete elabore pour les
enfants et doit etre pose aux parents.
4.1.2 £n_examen_ £pjrom£tr_i^ue_ ^impLe est fait a chaque sujet a
1'aide d'un spirographe a sec portatif (Vitalograph) qui
permet de calculer sur un trace d'expiration forcee, CV, VEMS, VEMS/CV,
VEM 25-75 Z.
4.2 Facteurs etiologiques a prendre en compte
Une partie du questionnaire permet de noter pour chaque sujet,
1'age, le sexe, les antecedents pathologiques, le tabagisme, 1'histoire
professionnelle, 1'exposition professionnelle eventuelle, le niveau
d1 etudes, la duree de residence dans la zone etudiee. Les questions sur
i
1'histoire professionnelle ont ete simplifiees a la suite de la pre—enquete.
Pour chaque menage, quelques renseignements sur les conditions de
logement sont releves a 1'aide d'un questionnaire permettant d'avoir une
idee de la pollution interieure eventuelle par le chauffage et de preciser
le niveau social.
-------
278
5. CONCLUSION
L'etude du protocole de 1'enquete epidemiclogique francaise sur les
relations entre pollution atmospherique moyenne et affections respiratoires
chroniques est une illustration des difficultes de toute etude de ce type.
Le role probablement mineur de la pollution atmospherique et, particu-
lierement de son taux moyen, dans la pathologic respiratoire exige pour
e*tre mis en evidence que I1 etude se diroule dans des zones fortement
contrastees. La multiplicity des facteurs etiologiques connus et inconnus
entraine des exigences parfois contradictoires. Le choix, pour cette
enquete, a etc difficile entre : etudier une population heterogene mais
groupee autour des points de mesure de la pollution, ou une population
homogene, au point de vue social, mais alors dispersed.
Une prg-enquete a ete menee pour tester le protocole et les methodes.
Elle a montre la necessite d'elargir les categories socio-professionnelles
et d'intensifier 1'harmonisation des methodes par une etude de comparabili-
te a tous les niveaux : mesures chimiques, questionnaires, examens spiro-
metriques.
Etude subventionnee par le Ministere de la Protection de la Nature et
de 1'Environnement et par le Ministere de la Sante Publique et de la
Securite Sociale.
-------
279
UBER VERANDERUNGEN EINIGER BIOLOGISCHER PARAMETER
BEI SCHULKINDERN IN GEBIETEN MIT
UNTERSCHIEDLICH STARKER LUFTVERUNREINIGUNG
P, SCHMIDT/ L, PELECH,++ R, DOLGNER"1"*"1"
+ Institut fur Fortbildung der Arzte und Pharmazeuten, Praha,
CSSR
++ Institut fur Hygiene und Epidemiologie, Praha, CSSR
+++ Medizinisches Institut fur Lufthygiene, Dvisseldorf, BRD
KORZFASSUNG
Vie vergleiohende Untersuohung von Bevdlkerungskollektiven
in stark luftverunreinigten Gebieten und in Gegenden mit relativ
sauberer Aussenluft ist eine der MSgliahkeiten, den Einfluss der
Luftverschmutzung auf den Gesundheitszustand des Mensohen zu er-
forsQhen. Im Ranm&n einer solohen epidemiologisonen Studie
vurden von uns in den Jahren 1968-74 Kindergruppen in der CSSRt
der BRD und der DDR untersuoht. GegenUber den in den Kontroll-
regionen lebenden Kindern wurde hierbei in den exponierten Kol-
lektiven wiederholt ein hSufigeres Vorkommen vergrSsserter und
zerklttfteter Gaumenmandeln sowie vergrdsserter region&rer Lymph-
knoten beobaohtet. Die Zahl pathogener Keime auf den Tonsillen
war hier hdher, der Muramidasespiegel im Speichel niedriger als
bei den Kontrollen. Die Kinder in den luftverunreinigten
Arealen zeigten regelmdesig eine relative Verspatung der Knochen-
reifung und eine relativ erhVhte Erythrozytenkonzentration;
ferner fanden wir den mittleren Hdmoglobingehalt und die Cnemo-
resistenz der roten BlutkSrperohen verandert. Der von uns positiv
gefUhrte internationale Vergleioh beweist die Allgemeingultigkeit
dieser Ergebniese.
-------
280
ABSTRACT
The comparative study of population groups in areas of high
air pollution and in districts with relatively clean ambient air
is one method of investigating the influence of air pollution on
human health. As part of an epidemiologioal study of this kind
we examined groups of children in the CSSR, the BED and the DDR
from 1968 to 1974. It was repeatedly noted that enlarged and
fissured tonsils as well as enlarged regional lymph nodes oc-
curred more frequently among the exposed groups than among the
children living in the control areas. The number of pathogenic
germs on the tonsils was higher and the muramidase level in the
saliva was lower than among the control groups. The children
in the air polluted areas were regularly relatively late in
reaching bone maturity and had a relatively high erythrocyte
count; further, we found changes in the average haemoglobin
value and the chemoresi stance of the erythrocytes. The inter-
national comparison conducted by us proves the general validity
of these results.
-------
281
Einfuhrung
Eine MSglichkeit, die Auswirkungen der atmospharischen Ver-
unreinigung auf die menschliche Gesundheit zu erfassen, ist
die vergleichende Untersuchung von Bevolkerungskollektiven,
die in luftverschmutzten bzw. relativ luftreinen Gebieten
unter im iibrigen moglichst Shnlichen Bedingungen leben. Als
besonders geeignet hierfiir betrachten wir in Ubereinstimmung
mit Kapalfn (1) und Symon (2) die 7-12jahrigen. Denn in die-
sem Alter sind die uns interessierenden Entwicklungsparameter
von der PubertMt noch unbeeinfluBt, und wir kSnnen in dieser
Altersstufe eine schnellere und intensivere Reaktion auf Um-
weltreize erwarten als beim Erwachsenen, ohne fiirchten zu
miissen, daB die Abwehrmechanismen noch nicht voll entwickelt
sind Oder die Expositionsdauer unzureichend ist.
Studienpj.an
Fiir unsere Untersuchungen des biologischen Status von Schul-
kindern haben wir die Regionen mit unterschiedlich starker
Luftverunreinigung aufgrund der Angaben und nach den
Empfehlungen der in der &SSR, der BRD und der DDR fur die
Immissionsuberwachung zustandigen Gremien ausgewahlt. In den
einzelnen Untersuchungsraumen dieser drei Staaten wurden Kol-
lektive mit einem mittleren Alter von rund lo Jahren aufge-
stellt, die beide Geschlechter in etwa gleichem Anteil um-
faBten; insgesamt haben wir von 1968 bis heute ca 45oo Kinder
untersucht. Bei den meisten unserer Feldstudien wurden vor
der eigentlichen Untersuchung Fragebogenaktionen durchgefuhrt,
teilweise auch die Eltern direkt interviewt, wobei wir der
Expositionsdauer des Kindes, den von ihm durchgemachten Er-
krankungen, seiner Ernahrungsweise und den sozio-Skonomischen
VerhSltnissen der Familie besondere Beachtung schenkten. Die-
se anamnestischen Daten wurden, soweit moglich, zum Zeitpunkt
der klinischen, somatometrischen und Labor-Untersuchung jedes
Kindes iiberpriift und ggf. erganzt. In die Endauswertung nahmen
wir nur die Daten jener Kinder hinein, die nach Anamnese und
klinischer Beurteilung als gesund zu bezeichnen waren.
-------
282
I 0 I S I L I £
« Uupll* in » l*t*«k TO.n,
QTJJ
Ti
Tl
Abb. 1. Tonsillenbefund "maBig bis stark verandert"
(2/3) der Knaben; Haufigkeitsprosentwerte von k Jahren
aus Gelsenkirchen (expon.) sowie den Raumen Hunsriick
und Freiburg (Kontrolle)
(mtittj ••
DIL1»I *
2/J • l«t.«» T0.71.TJ.TJ)
100
H
inn
72
TJ
tolwvkli
Abb. 2. Lyophknotenbefund "ma'Big bis stark verandert"
(2/3) der MMdcnen; HSufigkeitaprOeentwert* von k Jahren
aus Qelsenkirchen (expon.) sowie den Raumen Hunsriick
und Freiburg (Kontrolle)
-------
283
Summotionskurvtn dtr Differenz zwischcn
Kolender- und Knochtnolter in Monottn
Bittirftld
Btrtin
Seifftn
Arnftld
-32 -26 -20 -U -8 -2
*10
22 Mon.
Abb. 3. Wert "Knochenalter minus Kalenderalter"
(Monate) der Knaben; Summenhaufigkeiten fur Bitter-
feld und Selffen (expon.) sowie Arnfeld und Berlin
(Kontrolle)
- •
»T ^^"
_m— • • •
"
1.5* -
IB* -
TT
.JOl
Abb. *+. Knockenreifungszustand der deutschen
Population, bezogen auf die tschechischen Standard
werte; Summenhaufigkeit
-------
284
Studienergebnisse
Die Befunde der Kinder, die stadtisch-industriellen Luftver-
unreinigungen unterschiedlicher Starke exponiert waren, las-
sen beim Vergleich rait den Kontrollbefunden fur die Mehrzahl
der diagnostischen Parameter Unterschiede erkennen (Dolgner,
3; Schmidt, 4). Die beobachteten VerSnderungen bewegen sich
meist in der physiologischen Schwankungsbreite; die Unter-
schiede zwischen den belasteten Gruppen einerseits und den
Kontrollkollektiven andererseits sind dennoch fast immer
statistisch signifikant. Einige dieser Veranderungen konnte
Schmidt (5) im Tierversuch reproduzieren und dadurch ihren
Kausalzusammenhang mit Komponenten der Luftverunreinigung
wahrscheinlich raachen.
Im einzelnen fand man bei den exponierten Kindern stets rela-
tiv haufiger als bei den Kindern in den Reinluftgebieten ver-
grofierte und zerkliiftete Gaumenmandeln (Abb. 1) und vergroBer-
te regionale Lymphknoten (Abb. 2). Diese Beobachtung haben
wir bei alien unseren Aktionen in der &5SR, der BRD und der
DDR, und zwar wiederholt, gemacht. Die Kinder in den verun-
reinigten Gebieten erwiesen sich ferner regelmaBig in ihrer
Knochenreifung gegeniiber den Kontrollkindern als verspatet,
was wir als ein Zeichen ihres relativ geringeren biologischen
»•
Alters verstehen miissen (Abb. 3) . Der Ossifikationsgrad wurde
von uns auf der Basis des tschechischen Standards bestimmt,
dessen Anwendbarkeit nach vergleichenden Studien von Pelech
(6) gegeben ist (Abb. 4). Weiterhin zeigten die Kinder in
praktisch alien von uns untersuchten Belastungsgebieten eine
h6here Erythrozytenkonzentration als die Kinder in den Rein-
luftarealen (Abb. 5). Entsprechend war bei generell einheit-
lichen Hamoglobinwerten der mittlere Hb-Gehalt des Erythro-
zyten (HbE) unter Expositionsbedingungen, wenn auch nicht
immer eindeutig, relativ vermindert. Die Chemoresistenz der
roten BlutkQrperchen schlieBlich, bestimmt nach der Erythro-
gramm-Methode, fanden wir in den belasteten Kollektiven gegen-
uber den Kontrollgruppen sowohl erhSht als auch erniedrigt
(Abb. 6). Die genannten Unterschiede bestehen bei Jungen und
-------
285
Erythrozyten
i.6
*
>f
1
I
*
I'll!!
1'
IT
111
III
III
III
III
III
III!
kill
A
I
1
III
III
-------
286
HMtt am /man • Mtan m
• m
.*
cttu U
uttltc
afun
. H
• *
.M
Abb. ?• Haufigkeit (5O des Befundes "normale Mundflora" (1) und
von pathologisehen Keimbefunden (2) in den Tonsillenkrypten; Herbst
1971* Teplice (expon.) sowie Ceska Lipa und Liberec (Kontrolle)
MURAMIDAZA -1972
201-
• V
£10
^.
V
0
—
—
;I-
;'::
;•"
I-'.
'1
*
l~
••
1
j?
{•
jj
B%
*
»•
A
£
• *
>
!•
•
!*I
•
*•
»*•
•
,\
.*.
r
1
1
r
::\
'• '
:••
;*;
•-•
^:
!•:
•. .
•*.
".;
>.•
e
m
E
I
gj
A
,*.
^:
A
^
I 4
A
X
A
A
£
A
»
•*•
*%
•V
^
:L
1
1
p?€
po
^H
4
KJ
^
pq
§
•••
A
r*
.*.
K
h •
>•.
.N
»••
^
^k(
j5
r
•4
j.
1;
t%
;;".
|:;
;.;
i'
•*«
v
•*•
S
',•1
,*.
,\
.*.
V
X
•*.
•"
^
S
•*•
*^
^.
«•
.*.
§
u
Abb. 8. Lysozymkonzentration ii& Speichel (mg/ml); Hittelwerte vor
(pred) und nach (po) der klinischen Untersuchung des Jahres 1972
aus Teplice (expon.) und Ceska Lipa (Kontrolle)
-------
287
Madchen gleichsinnig; wir haben jedoch den Eindruck gewonnen,
daB die Knaben auf eine Xnderung der Immissionssituation
schneller und starker reagieren als die Madchen.
In Nordbohmen bezog sich ein spezieller Teil der Unter-
suchung auf die Frage, ob sich die Kinder auch bzgl. der
Keimbesiedlung der Tonsillenoberflache und Tonsillenkrypten
sowie hinsichtlich des Lysozymgehaltes des Speichels unter-
scheiden. Pfeifer (7) fand in tlbereinstimmung mit dem klini-
schen Bild bei den Exponierten haufiger als bei den Nicht-
exponierten pathogene Keime (Abb. 7); Schmidt (8) sah bei den
belasteten Kindern eine relative Erniedrigung des Muramidase-
spiegels, der als Parameter der unspezifischen Abwehrlage
gilt (Abb. 8).
Unsere Studie hat daruberhinaus ergeben, aaB sich die von
tschechischen Autoren fur tschechische Bevolkerungskollektive
ausgearbeitete Methode der Gruppendiagnostik vernunftig und
erfolgreich auch auf andere Populationen anwenden lafit. Unter
der Voraussetzung, daB man die einzelnen Populationen mit Er-
folg zumindest orientierend standardisiert hat, sind die mit
dieser Methode gewonnenen Daten im Prinzip international
vergleichbar.
Literatur
1. KAPALfN, V., Vyvoj de"tf za ruznych 2ivotnfch podmfnek, Dis-
sertationsarbeit, Hygiene-Institut, Praha 1964.
2. SYMON, K., et al., >*s- Hyg. , 5, 88 (I960).
3. DOLGNER, R., et al., £s. Hyg., 16, 62 (1971).
4. SCHMIDT, P., PELECH, L. , Acta hyg., epidem. et microbiol.
(Praga), 3, XX (1974).
5. SCHMIDT, P., Experimental study on the effects of sulphur
oxides upon organism, Collection of Scientific Reports
of the Institute of Hygiene, Prague 197O.
6. PELECH, L., Cs. Hyg., 14, 195 (1969).
7. PFEIFER, I., et al., Cs. Hyg., 17, 347 (1972).
8. SCHMIDT, P., RICHTER, J., PFEIFER, I., Cs. Hyg., 18, 299
(1973).
-------
289
AIR POLLUTION: METHODS TO STUDY ITS
RELATIONSHIP TO RESPIRATORY DISEASE IN
BRITISH SCHOOLCHILDREN
L, IRWIG, D, G, ALTMAN, R, J. W, GIBSON AND C. FLOREY
Department of Clinical Epidemiology and Social Medicine,
St. Thomas's Hospital Medical School, London, United Kingdom
ABSTRACT
This paper describes the design and some of the methods
used in a study of the relationship between air pollution and
childhood respiratory disease.
The data were collected in the first year of a four year
mixed longitudinal study intended to detect effects of changes
in air pollution levels on respiratory disease in children.
Two aspects of methodology are discussed:- first the sel-
ection of survey questions concerning respiratory illness on
the basis of the association of these reported illnesses with
impaired peak expiratory flow rates; secondly, the use of quan-
tal response regression to estimate the effects of atmospheric
smoke and sulphur dioxide on respiratory disease after allowing
for the effects of interfering factors (e.g. age3 sex and social
class).
Analysis by this method on a subsample of the study child-
ren showed a positive association between some of the selected
reported illnesses and both smoke and sulphur dioxide.
-------
290
1. Introduction
There is substantial evidence that children in areas \vith high
concentrations of atmospheric smoke and sulphur dioxide have more
respiratory disease than those in areas of low pollution (1,2).
The objective of our study is to determine whether changes in
respiratory disease prevalence in children over successive years
are related to changes in measured levels of air pollution. The
opportunity to test this hypothesis has arisen in Britain since the
introduction in 1956 of the Clean Air Act which prohibited the
emission of dark smoke by industry. The domestic provisions of
the Act gave Local Authorities power to set up Smoke Control Areas
in which the burning of fuels other than 'smokeless fuels' v/as
forbidden. Smoke Control Areas are still being designated and
smoke emissions and concentrations are decreasing over time. (3)-
This paper discusses some of the methodology of the study.
2. Method
2.1 Study Design
The study is being carried out in association with a national
study of the health and growth of approximately 11,000 school-
children aged 5 to 11 years. All children attending selected
primary schools in 28 areas of England and Scotland are being
examined annually. The areas represent a random sample stratified
by an index designed to indicate their relative wealth. The sample
was weighted to include a higher proportion of poorer areas.
-------
291
TABLE I
Study Design
Age
(years)
11
10
9
6
7
6
5
A
B
C
D
E
F
G
B
C
D
E
F
G
H
C
D
E
F
G
H
I
D
E
F
G
H
I
J
1973 1974 1975
year of study
1976
The design of the study is shown in Table I in which cohorts
are identified by different letters. This mixed longitudinal
design lends itself to several methods of analysis. For example,
we can compare respiratory disease prevalence in equivalent age
groups over successive years. We can also trace the changes in
respiratory disease prevalence in any cohort and compare these with
the changes experienced by another cohort which may have been
exposed to different levels of pollution.
2.2 Measurement of Respiratory Disease
2.21 Questionnaires
Self-administered questionnaires are completed by parents of
children in the study. Questions are asked about each child's
-------
292
respiratory disease episodes and symptoms as shov/n in Table II
and about each family's social and economic status.
TABLE II
List of questions about child's health
1. Does he or she usually cough first thing In the morning?
(Exclude clearing throat or single cough).
2. Does he or she usually cough during the day or at night?
(Exclude clearing throat or single cough).
3. Does his or her chest ever sound wheezy or whistling?
4. Do colds usually go to his or her chest?
5. Has he or she had earache in the last 12 months?
6. Has he or she suffered from any of these Illnesses in the last
12 months? - Asthma
- Bronchitis
- Pneumonia
7. Has this child stayed in hospital overnight or longer in the
last 12 monthsV
Please give the reason for each separate admission and the
length of stay.
B. Which of the following words do you think best describes this
child's health? - Excellent
- Good
- Fair
- Poor
9. Has this child been seen by a doctor within the last two weeks?
2.22 Peak expiratory flow rates
Peak expiratory flow rates are measured using Wright Peak Flo?/
Meters. Five readings are taken from each child and the mean of
the last three is used in analysis. Height and weight are
measured at the same time. All measurements are taken by trained
nurses in each area in the presence of one of the fieldworkers
-------
293
from our department who remeasures 12% of the children. The
comparability of the measurements taken by the three fieldworkers
is also assessed periodically. These data from the nurses and the
fieldworkers are used to estimate the size of measurement error due
to differences in machines and techniques.
2.3 Measurement of Air Pollution
Smoke and sulphur dioxide are sampled using a standardized
hydrogen peroxide sampler as recommended by the Organisation for
Economic and Cooperative Development. The same type of meter is
used in the United Kingdom National Study of Air Pollution conduct-
ed by Warren Spring Laboratory.
A meter is sited at or within half a mile of each school in
the study and 2k hour readings are obtained. Reflectometry on
smoke stains and titration of sulphur dioxide are performed under
the supervision of people trained by Warren Spring Laboratory.
3. Methods of Analysis
3.1 Choice of Respiratory Disease Measurement
The quality control system described above allowed us to
compare the peak expiratory flow measurements obtained by one of
our fieldworkers using a standard meter with those obtained by a
nurse usins a different meter on the same sample of children. Our
analyses indicated that most of the measurement errors were due to
differences between meters rather than between observers. The
meter differences were present even after recent calibration.
Table III shows the magnitude of the difference between peak
expiratory flow rates obtained on one meter by one fieldworker and
-------
294
TABLE III
Mean differences in peak expiratory flow rate (PEFR) between
five meters and a standard meter
Nurses'-
Meters
A
B
C
0
E
No. of
Children
81
97
56
32
29
Mean
Difference
Cl/min)
17.0
14.9
30.4
-9.4
16.6
Standard
Error
2.0
1.9
4.1
4.2
4.1
P
(t-test)
< .001
< .001
< .001
< .05
< .001
Each child's PEFR was measured on a nurse's meter and a
standard meter. Five readings were taken with each meter.
The meters were used in random order.
those obtained by nurses using some of the other meters. In this
table each of the meters A to E have been used by more than one
nurse. In addition to the differences shown, readings tended to
vary over the time that a meter was in service. It seems likely
that the differences in peak flow rates expected between areas of
high and low pollution would be small in comparison with the
magnitude of inter-meter variation shown in the table.
The biases arising from differences between meters can be
eliminated in small surveys, by randomly allocating the meters over
the whole study population. However, in studies where populations
-------
295
are being measured in many widely separated areas it is not admin-
istratively feasible to do this.
One of the methods of overcoming this difficulty is to use
reported illnesses which are associated with reduced peak flow rates,
This method was tested on a sample of the children examined in 1973.
For each question the population was divided into those replying
yes and those replying no. The peak flow rates for these groups
were adjusted for differences in age, sex, height, wei-ght and which
meter had been used (analysis of covariance) and then compared.
The adjusted :nean flow rates for the two groups for each question
are given in Table IV. The difference between the rates and the
probability of the difference arising by chance are also given.
The data suggest that the first six questions in the table may
be used as indicators of lower respiratory tract disease.
3-2 Statistical Techniques for Handling Reported Respiratory
Disease
A quantal response (i.e. yes or no, in this case) was available
for each question. We wanted to know first whether age, sex and
social class were associated with the response, second whether,
after adjustment for these variables, smoke and SO- were also
related to the response and, third, how the proportion of people
giving a positive response changed with change in smoke and SO-.
Simple tabulation of the data showed that age, sex, social
class and area (or pollution level) were each separately associated
with positive responses. A. model was required to predict the
response using age, sex, social class, smoke and S02 as predictors.
The coamonly used multiple regression analysis requires that the
predicted variable is continuously and normally distributed at each
-------
296
13
TABLE IV
PeaK expiratory flow rates, (PEFR), adjusted for
age, height, weight and meter, by reported disease
Question
1. Asthma
in last
year
2. Wheeze
3. Fair/
Poor
Health
4. Bron-
chitis in
last year
5. Cough
6. Colds
to chest
7. Doctor
visit in
last 2
weeks
6. Ear-
ache in
last 12
months
9. Hospi-
tal for
upper
respira-
tory dis-
ease in
last 12
months
% reporting PEFR[1/min)
symptom if no symp-
(Total=3115) torn
1.8 242.2
10.2 243.2
5.4 242.3
5.3 242.5
8.4 242.6
25.5 244.0
9.4 242.1
27.8 242.0
2.0 241.6
PEFR (1/minl Difference in Probability
if has symp- PEFR (1/min) of differ-
tom ence arising
by chance
212. B 29.4 < 0.0005
226.2 17.0 < 0.0005
227.0 15.3 < 0.0005
228.0 14.4 < 0.0005
231.2 11.5 < 0.0005
233.7 10.2 < 0.0005
237.7 4.4 < 0.10
241.1 0.9 <0.25
240.8 0.8 <0.90
-------
297
valie of the independant variables if the statistical tents are to
be valid. Since the response was not continuously distributed, a
regression method specially designed to handle quantal data was
used.(^) This provides a regression equation which predicts the
probability that a croup of individuals with characteristics
described by the independent variables would give a positive res-
ponse to any particular question. Each regression coefficient
could be tested for statistical significance and the significance
Of combinations of variables can be calculated.
This method has been used on some of the information collected
during 1973. An equation was computed for the 1,816 children for
whom sex, are, social class and ambient air pollution levels were
known. The air -pollution levels used were the tiean smoke and SO
for November 1973. These variables were used to predict the
3j
logit of the proportion of subjects reporting colds to the chest.
None of the independent variables v/ere correlated with any other,
except smoke v,rith SO-,. These tv,ro showed such high correlation
that they were entered into separate equations. Both pollutants
were found to be significantly associated with a history of colds
going to the cheat {p < .05). The equation for predicting colds
to the chest showed increased risk for this reported illness for
males, younger children, thooe in lov/er social classes and those
living in areas vith high smoke or S02 pollution.
By appropriate substitution in the equation, the proportion of
children who would be expected to have colds to the chest was
* The logit used in the analysis was 0.5 (Ln (P/l-p) ) where
P is the proportion of children v/ith a characteristic.
-------
298
estimated for the group at greatest risk (males, aged 5, social
class V).
Over the range of smoke levels in this analysis (10-130 /ug/nr),
it was predicted by the equation that for each increment of 10
of smoke (ignoring S02)» 0.77$ more of this population would have
colds to the chest. In the group at least risk (females, aged 11
in social class I) approximately 0.^0/6 of the population would he
affected with each 10 ug/nr increment in smoke. Similar results
were obtained for S0_, ignoring smoke.
Using the same technique, no significant relationship was
found between air pollution (smoke and S02 separately or combined)
and the other respiratory diseases listed in Table IV. The relation-
ships however were always in the direction of more disease in more
polluted areas.
Both smoke and S02 were significantly associated with respira-
tory disease defined as a positive answer to at least one of the
six questions indicative of lower respiratory tract disease
(p < .005). This relationship is still under study.
4. Comment
Reported illness provides a useful measure of respiratory
disease in children. Questions can be selected on the basis of
their relationship to impaired peak expiratory flow rates. Binary
data of this type can then be analysed using a quantal response
regression which predicts the probability of any defined group
having a reported illness.
The model we have used assumes a linear relationship between
a set of predicting or independent variables and the logit of the
-------
299
proportion of subjects reporting illness. This model appeared to
fit the data well (P=0.5, X2 for fit).
We have used the predictive quality of the equation to
estimate the changes to be expected in the prevalence of colds to
the chest in the groups at highest and lowest risk (as judged by
the equation). These predictions suggest that diminution of
smoke or S02 levels from 130 yug/nr to 10 yug/nr would result in a
decline in prevalence from 49% to k&% in the highest risk group and
from 1796 to 1256 in the lowest risk group.
These predictions have been made for the extreme groups and
may carry rather large standard errors. A method for calculating
the standard error is not yet available however, so we may only
suggest the change in the prevalence of 'colds to the chest' over
the given range of pollution levels lies somewhere between 5°4 and
9*.
Quantal regression has thus permitted us to make predictions
from prevalence data about the changes we .-night expect to find in
longitudinal data. We must now await the end of the study to see
whether the changes occurring in respiratory disease prevalence as
air pollution levels change correspond to our predictions.
5. Acknowledgements
This study is supported by a grant from the Department of
Health and Social Security.
We thank all the doctors, nurses, and teachers involved in the
study areas and the staff of Warren Spring Laboratory for their
continued help.
We also thank Miles. A. Tavender, J. Palmer, A. Edwards and
J. Cox for their administration and fieldwork.
-------
300
Professor V/.W. Holland has been instrumental in initiating
this study and formulating the design. We thank him for his advice
and help throughout the study.
6. References
1. Colley, J., Reid, D., "Urban and Social Origins of Childhood
Bronchitis in England and Wales" British Medical Journal
2, 212-217 (1970)
2. Holland, \V., Halil, T., Bennett, A., Elliott, A., "Factors
influencing the Onset of Chronic Respiratory Disease"
British Medical Journal 2, 205-208 (1969)
J>. Warren Spring Laboratory, "National Survey of Air Pollution
1961-71", Volume 1 Her Majesty's Stationery Office, London
(1972)
J*. Naylor, A.F., "Comparisons of regression constants fitted by
maximum likelihood to four common transformations of binomial
data" Ann. Hum. Genet., London 27, 2tfl (19&4)
-------
301
EPIDEMIOLOGIC STUDIES OF ADVERSE HEALTH EFFECTS
ASSOCIATED WITH EXPOSURE TO AIR POLLUTION
G, J, LOVE
Human Studies Laboratory, National Environmental Research Center
Environmental Protection Agency, Research Triangle Park, NC, USA
ABSTRACT
Epidemiologic studies of the effects of air pollution have
'been conducted by the Human Studies Laboratory of the U.S. Envi-
ronmental Protection Agency since 1969. These studies have
been carried out in groups of communities selected to be similar
with respect to climate and demographic characteristics but to
differ with respect to air pollution levels. Measurements of
health status within each community were made in a manner which
permitted differences to be associated with differences in expo-
sure to pollution. Particular effort was made to select the
groups of communities to provide gradients for particular pollu-
tants or combinations of pollutants in the hope that the total
data obtained might provide evidence concerning the relative
significance of specific pollutants, especially sulfur dioxide
(SO ) and the total suspended particulates (TSP).
&
The results of these studies have shown consistent associ-
ations of impaired health status with exposure to higher levels
of pollution. They have, however, left a number of problems
unanswered including the one concerning the relative significance
of SO or TSP. On the other hand the studies indicated that
the atmospheric levels of suspended sulfates may be more directly
related to impaired health than is either SO ^ or TSP.
-------
302
As the results of chronic exposure, studies were analyzed and
the data interpreted, ranges of current or prior exposure values were
developed to represent the "worst case" and the "least case" air
pollution level which might have been incriminated in a causal
manner with the adverse health effects measured (Table 1). Then
within these ranges "best judgement" estimates were made concerning
the levels that In the opinion of the investigators most probably
were related to the increased adverse effects.
For the short-term studies comparisons of the variation in
health status were made with 24 hour daily mean levels of pollution
(Table 2).
The studies of chronic respiratory disease ~ indicated that
the relative contribution of air pollution alone ranged from one-
third to one-seventh that of cigarette smoking as a determinant of
chronic bronchitis prevalence in communities. The sum of the
evidence suggested that, while personal cigarette smoking was the
largest determinant of bronchitis prevalence among parents of school
children, air pollution itself was a significant and consistent
contributing factor, leading to increased bronchitis rates in
nonsnokers as well as in smokers from polluted communities.
Attempts to assess the length of residence in polluted areas
required for development of excess bronchitis rates suggested
that immigrants into polluted areas reported excess chronic
t
bronchitis after two to seven years of exposure.
-------
TABLE 1
SUMMARY OK CHESS STUDIES RELATING LONG-TERM POLLUTANT EXPOSURES INVOLVING SULFUR DIOXIDE, TOTAL
SUSPENDED PARTICULATES AND SUSPENDED SULFATES TO ADVERSE EFFECTS ON HUMAN HEALTH
Adverse Effect
on Human Health
Increase in Prevalence of
Chronic Bronchitis in Adults
Increases in Acute Lower
Respiratory Tract
Infections 1n Children
Increase in Frequency or
Severity of Acute
Respiratory Illness 1n
Otherwise Healthy 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
i
Duration
m
Exposure
(Years)
3
10
6
3
3
3
1
3
3
1
9
8-9
Annual Average Levels Linked to Adverse
Health Effects (ug/m3)
Sulfur
Dioxide
62
374
95
92
177
95
50
210
106
57
435
200
Total
Suspended
Particulates
65
179
100
65
102
102
104
159
151
96
200
100
Suspended
Sul fates
12
20
15
7.2
15
15
14
16
15
9
28
13
-------
TABLE 2.
SHORT-TERM EXPOSURES:
POLLUTANT THRESHOLDS FOR ADVERSE EFFECTS
(BEST JUDGMENT)
EFFECT
THRESHOLD, 24-hour fig/ni3
S02
TSP
SS
w
o
AGGRAVATION OF
SYMPTOMS IN ELDERLY
AGGRAVATION OF
ASTHMA
>365
80-100
8-10
180-250
70
840
-------
305
In the best judgement of the investigators, excess chronic
bronchitis in the Salt Lake Basin of Utah could reasonably be
attributed to annual mean sulfur dioxide levels of 92 to 95 yg/m3
and/or suspended sulfate levels of 15 yg/m3. In the other areas,
the investigators judged that the lowest annual mean pollutant concen-
trations which could reasonably be associated with excess chronic
bronchitis were past exposures to 100-177 yg/m sulfur dioxide,
3 3
80-118 yg/m total suspended particulates and 9-14 yg/m suspended
sulfates.
From these data, it appeared that excess bronchitis may be
associated with community exposures to sulfur oxides alone, in
3 "3
the form of annual levels of 92 to 95 yg/m S(L and 15 yg/m
suspended particulate sulfates. When higher levels of particulate
3 3
matter were present, annual exposures to 100 yg/m S02> 120 yg/m
3
total suspended partlculate and 14 yg/m suspended sulfate were
associated with excess bronchitis.
Lower respiratory disease (LRD) studies of children
indicated that for single and repeated episodes of croup and
repeated episodes of any LRD, families of children who had lived
for three 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.
It is interesting that larger increases in total LRD were
observed in the high pollution community of the Salt Lake Basin
-------
306
study than In the high pollution communities in another Rocky
Mountain study area. Also, the mean annual suspended sulfate
concentration was higher in the high pollution community in the
Salt Lake Basin study than in the other Rocky Mountain study but
the opposite was true for SCL. This suggested that the increases
in LRD frequency were more likely associated with suspended sulfates
than with S02.
In the best judgement of the investigators, it seemed
reasonable to conclude that there is a positive association between
lower respiratory disease frequency in children and pollution
exposure, and that excess respiratory disease may be associated
with community exposures to approximately 95 yg/m SO. and 15
vg/m suspended sulfates.
Acute respiratory disease studies of families in Chicago
o
and New York studies showed that, with the exception of fathers,
a consistent excess acute respiratory disease rate was reported
among family members living in more polluted neighborhoods. The
relative excess varied from 3 to 40 percent. A conservative
3 3
estimate would be that exposures to 210 vg/m SO. with 104 vg/m
3
total suspended particulates and approximately 16 yg/m suspended
sulfates was associated with a 5 to 20 percent excess of acute
respiratory illness in various family members.
-------
307
The Chicago study also provided evidence of increased
susceptibility to epidemic A2/Hong Kong influenza among other-
wise healthy families exposed during the previous three years to
atmospheric levels of 106 to 119 Pg/m3 S02, 151 to 159 yg/m3
total suspended participates, and 14 yg/m suspended sulfates.
9-10
Pulmonary function studies showed that ventilatory
function of elementary school children, measured by the three-
quarter second forced expiratory volume (FEVQ 75), was diminished
in areas of elevated exposure to sulfur oxides. In New York only
the older children (age nine to 13 years) who had been exposed
to substantially elevated pollutant concentrations for the first
five to ten years of life suffered reduced ventilatory function.
The best available estimates of these remote annual average
3
exposures were as follows: sulfur dioxide, 131-435 yg/m , total
suspended particulates, 75-200 vg/m , suspended sulfates, 18-28
pg/m .
Ventilatory function in white children exposed to suspended
sul fate concentrations of about 9.5 yg/m was lower than that
of white children exposed to concentrations of about 8.3
o
vg/m . Black children studied were all exposed to suspended
sulfate concentrations of about 8.9 yg/m , and these children
demonstrated no differences in ventilatory function.
-------
308
From a Cincinnati study, it was conceivable that one year's
exposure to 9 yg/m of suspended sulfates, in the presence of
moderate levels of sulfur dioxide and total suspended particulates
3 3
(57 ug/m and 96 yg/m , respectively) might alone account for
reduced ventilatory function.
Among asthmatics and study subjects with cardiopulmonary
illness in the Salt Lake Basin daily attack rates were more
consistently correlated with lower outdoor temperature than with
-any measured pollutant. Asthma attack rates were most closely
related to stepwise increases in the levels of suspended sulfates.
Virtually no relationship between S0_ and attack rates appeared.
Total suspended particulates and suspended sulfates were
positively and stepwise correlated with daily asthma attack rates.
In New York, asthma attack rates were more consistently
associated with daily suspended sulfate levels than with either
12
SO- or TSP . The pattern of daily aggravation of symptoms in
4
cardiopulmonary subjects in New York was very similar to that
of asthma. Low temperatures were directly related to increased
symptom rates in subjects with combined heart and lung disease.
Elevated suspended sulfates were the only pollutant consistently
associated with symptom aggravation.
It was the best judgement of the investigators that
significant aggravation of cardiopulmonary symptoms could be
-------
309
attributed to 24-hour suspended sulfate levels as low as 8-10
yg/m on cooler days (20-40°F) or wanner days (41+°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 response.
With regard to short-term exposures then it was evident that
adverse health effects were attributable to suspended sulfate
levels rather than to the observed concentrations of SO- and
for TSP.
The identification of atmospheric suspended sulfates as an
environmental pollutant of concern to health, it is now necessary
to obtain sufficient information to recommend the extent to
which emissions should be controlled to protect against them.
Little is known about the formation of atmospheric suspended
sulfates and even less is known about the means to control them.
To develop adequate information about this need it is also
necessary to determine if all sulfates are equally reactive;
if sulfates are reactive because of the chemical properties
associated with specific chemical compounds; or because of
physical properties such as particle size or pH; and if sulfates
are equally reactive in humid and dry air and at high and low
temperatures. These biological issues must be addressed and
satisfactorily resolved, because our strategies to protect
human health may be critically dependent on the nature of the
-------
310
sulfate-biologic response relationship. Until more definition
of these issues is achieved, however, our findings indicate
strongly that we are just beginning to learn about the real
impact of atmospheric pollution on human health.
These issues are being addressed in the current research
program of the Human Studies Laboratory.
SUMMARY
Epidemiologic studies of the health effects of air pollution
in the United States have indicated that both acute and chronic
health effects are associated with exposure to ambient air containing
3 3
approximately 100 pg/m S0_, 100 yg TSP and 15 ug/m suspended
sulfates. The levels of suspended sulfates were associated more
closely and more consistently than were the levels of either
TSP or S02.
-------
311
REFERENCES
1. D. E. House, J. F. Finklea, C. M. Shy, D. C. Calafiore,
W. B. Riggan, J. W. Southwick and L. J. 01 sen. 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, 1972, pp.
2. C. G. Hayes, D. I. Hammer, C. M. Shy, V. Hasselblad, C.
R. Sharp, J. P. Creason and Kathryn 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, 1972, pp.
3. J. F. Finklea, J. Goldberg, V. Hasselblad, C. M. Shy,
C. G. Hayes. Prevalence of Chronic Respiratory Disease Symptoms
in Military Recruits, 1969-1970. In: Health Consequences
of Sulfur Oxides: A Report from CHESS, 1972, pp.
4. H. E. Goldberg, J. F. Finklea, C. J. Nelson, Walter Steen,
R. S. Chapman, D. H. Swanson and A. A. Cohen. Prevalence of
Chronic Respiratory Disease Symptoms in Adults: 1970 Survey
of New York Communities. In: Health Consequences of Sulfur
Oxides: A Report from CHESS, 1972, pp.
5. W. C. Nelson, J. F. Finklea, D. E. House, D. C. Calafiore,
M. B. Hertz, and D. H. Swanson. Frequency of Acute Lower Respiratory
Disease in Children: Retrospective Survey of Salt Lake Basin
in Communities, 1967-1970. In: Health Consequences of Sulfur
Oxides: A Report from CHESS, 1972, pp.
6. J. F. Finklea, 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, 1972, pp.
7. J. F. Finklea, J. G. French, G. R. Lowrimore, J. Goldberg,
C. M. Shy and W. C. Nelson. Prospective Surveys of Acute Respiratory
Disease in Volunteer Families 1969-1970 Chicago Nursery School
Study. In: Health Consequences of Sulfur Oxides: A Report
from CHESS, 1972, pp.
8. G. J. Love, A. A. Cohen, J. F. Finklea, J. G. French, G.
R. Lowrimore, W. C. Nelson, P. B. Ramsey. Prospective Surveys
of Acute Respiratory Disease in Volunteer Families 1970-1971
New York Studies. In: Health Consequences of Sulfur Oxides:
A Report from CHESS, 1972, pp.
-------
312
9. C. M. Shy, V. Hasselblad, J. F. Finklea, R. M. Burton,
M. Pravda, R. S. Chapman and A. A. Cohen. Ventilatory Function
in School Children: 1970-1971 Testing in New York Communities.
In: Health Consequences of Sulfur Oxides: A Report from CHESS,
1972, pp.
10. C. M. Shy, C. J. Nelson, Ferris Benson, W. B. Riggan, V.
A. Newill and R. S. Chapman. Ventilatory Function in School
Children: 1967-1968 Testing in Cincinnati Neighborhoods. In:
Health Consequences of Sulfur Oxides: A Report from CHESS,
1972, pp.
11. J. F. Finklea, 0. C. Calafiore, C. J. Nelson, W. B. Riggan,
C. 6. Hayes. Aggravation of Asthma by Air Pollutants: 1971
Salt Lake Basin Studies. In: Health Consequences of Sulfur
Oxides: A Report from CHESS, 1972, pp.
12. J. F. Finklea, J. H. Fanner, A. A. Cohen, 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 Oxides: A Report from CHESS, 1972, pp.
-------
313
PREDICTIVE MODELS FOR ESTIMATING THE HEALTH
IMPACT OF FUTURE ENERGY SOURCES
BERTRAM W, CARNOW
Occupational and Environmental Medicine, University of Illinois
School of Public Health, USA
ABSTRACT
Decisions regarding future energy usage will depend on a
number of major parameters, including those most essential to
life, and must be based on consideration of the pollutants pro-
duced by the source, control methods for the pollutants, and,
most important, the health effects produced by the energy source
used.
Models for quantitating health effects on varying populations
and the predictive capacity for assessing these effects have been
developed and will be discussed; and, the types of studies that
are needed to increase the required data base for this predictive
capacity will be explored.
Results of a study examining visits to the emergency rooms
in eighteen major hospitals in the City of Chicago during twenty-
four hours on the same day of the week that was carried out for a
period of eighteen months will be presented. Documentation of
all admissions to the emergency rooms for cardiac, respiratory,
and other diseases was maintained and levels of various air pol-
lutants at the hospital and in the area of the residence from
which the patients came to the hospital were evaluated. The
study examines these relationships and their effect on the cardio-
pulmonary system and how these results can be used in the predictive
-------
314
model being developed. The model is based on the concept of a
continual dose-response relationship between air pollutants and
health and requires definition of various categories in the pop-
ulation at high risk. By quantitative definition of differences
in disease at low and high pollution levels in those at high risk
and determination of the total numbers of these individuals in
the population, the numbers that are affected at varying levels
of pollution can be ascertained.
-------
315
The extraordinary increase in the use of fuels for the production of
energy for industrial use, transportation, and the production of
electricity, has created serious problems of environmental contamination.
Ihe dilemma faced by all industrial societies is one of achieving and
maintaining a healthy environment in the face of this apparent huge energy
need and limited or inaccessible clean energy sources.
Seme pollution accompanies any energy production and use. Decisions
regarding future uses of alternative energy sources must, therefore, be
considered with great care. While these decisions must examine many
social, economic, and political factors, an essential parameter is the
health effects; and, not only current, but future impact must be estimated
for each energy source with increasing use, as well as for new and
untried sources.
Pollutants frequently occur together, thus decisions regarding
alternatives must consider two preliminary sets of questions in order to
determine priorities for development of technology for the selective
removal of individual pollutants. These questions are:
1) Which pollutants are the major health stressors?
2) What is the extent of the contribution, qualitatively and quanti-
tatively, of each to etiology or pathogenesis of various diseases?
3) What meteorologic and other environmental factors contribute as
environmental stressors?
A study examining this series of questions was recently carried out
in the City of Chicago and will be discussed.
Critical questions which then must be asked concerning the health
cost of a particular energy source include:
1) How many people are being adversely affected by pollutants from
various energy sources, such as coal, oil, natural gas, and nuclear
energy?
2) Who are the affected people and what diseases are caused or
exacerbated by these substances?
3) How many more would be affected by the increased pollutants
produced from a particular source for a given increase in energy
production?
A model developed to begin to answer these critical questions in
order to predict the intact on populations of changes in pollution levels
engendered by qualitative or quantitative changes in fuel sources for
-------
316
-------
317
for energy use, will be presented.
HOSPITAL EMERGENCY ROOM STUDY
METHODS i
Emergency room admissions for all major respiratory and cardiac
causes were monitored from midnight on the Monday of each week to midnight
on Tuesday for eighteen months in fourteen hospitals scattered throughout
the City of Chicago. (Figure I) A form listing each major cardiac and
respiratory disease was used by the nurse in charge of the emergency room
to enter a diagnosis for each individual seeking care. In addition, the
age, sex, race, and place of residence were noted. These data were
collected each week and computerized.
Pollution and Meteorologic Measurements:
The City of Chicago maintains a continuous monitoring network of
eight stations which measure SO , coefficient of haze (COH) , CO, and for a
lesser period of time, NO . (Figure II) In addition, a twenty station
Jv
network measured twenty-four hour levels of SO and particulate, three
days each week. Using these data and with the knowledge of the residence
location of each individual who was admitted to the emergency rooms, two
indices of air pollution exposure were used in analyzing the data. The
first was a citywide average, or arithmetic average of air pollution
levels observed at all functioning monitoring stations. The second was
a weighted average of levels at these stations where, for a given hospital,
if n. = the number of subjects admitted for cardiac or respiratory causes
who reside in areas closest to monitoring station i for the
entire period of study.
N = ^ n. = the total number of subjects examined for cardiac or
respiratory causes,
P. » The air pollution level observed at monitoring station i. for a
id
given day, d,
The weighted pollution average, w , on a day d is defined by the
formula :
v n.
P
Pid
In justifying the use of the weighted average as a valid indication
-------
318
of air pollution exposure, two assumptions were made:
1) Air pollution levels observed at a given station should generally
reflect the exposure in areas closest to that station, and
2) The population from which a given hospital drew its patients
was reasonably stable, at least for the duration of the present study, so
ni
that the observed proportion (-rr-) of patients residing in the neighbor-
it
hood of Station i would be a reasonable estimate of the actual proportion
of the number of individuals residing near Station i most likely to seek
treatment at that hospital.
A consideration of the formula for the weighted average shows that
this quantity will tend to reflect exposure in areas where there is a high
proportion of individuals seeking treatment, whereas, the citywide average
gives all areas equal weight. Since the distribution of place of residence
of individuals who seek emergency treatment at a particular hospital is
far from uniform (Figure III) but instead, reflects socioeconomic factors
and accessibility to the hospital, the use of a weighted average appears
justifiable.
Each hospital then compared admissions to its emergency week from
week to week, throughout the eighteen months of study in relation to
variations in these pollution levels.
As a preliminary step in determining whether significant associations
held between the number of emergency room admissions and air pollution,
multiple linear regression techniques were applied to the data. To
briefly summarize this method, a linear relationship is assumed to hold
between a so-called "dependent" variable (the number of admissions, in
this case) and a set of "independent variables, or predictors, whose
behavior may influence the behavior of the dependent variable. Because
of measurement error in the dependent variable, observed values of the
dependent variable will tend to fluctuate about the hypothetical line or
plane defined by the linear relationship. Multiple regression analysis
provides estimates of the values of the dependent variable in terms of the
independent variables, such that the deviation of the observed values of
the dependent variable from the estimated values is minimized. The
percentage of total variation in the dependent variable explained by the
independent variables (R ) can also be calculated and tested for statis-
tical significance. When variables are observed in time, as in this case,
rather than under controlled laboratory conditions, it is possible that an
-------
319
tr vum • SIT*
Figure III : Distribution by place of residence
of patients making cardiac/respiratory
emergency room visits to Cook County
Hospital (Chicago)
iii—t—e—s—s—t—s—*—*—i—t-
5S5 S3 25 52:3
Figure IV : Chicago cardiac/respiratory emergency
room admission study. Seasonal var-
iation in SO and respiratory admis-
sions from October, 1971 through
March, 1973. Cook County.
-------
320
apparent association between them may be due to the action of unobserved
variables correlated with some of the observed variables. Most of the
variables in the present study are correlated in tine; for example,
Figure IV shows pronounced seasonal variation in the number of respiratory
admissions and in the weighted SO. average. Since many variables exhibit
a similar relationship in time, but are clearly not factors which
precipitate episodes of chest illness, all variables in the present study
were seasonally adjusted before applying any statistical technique to the
analysis of the data. Each variable was seasonally adjusted by subtracting
from its observed value on a given day, a "moving" average of the four
values, one and two weeks before and following the day in question.
The moving average serves as a rough extimate of seasonal trend with the
irregular or short-term component of variation removed; hence, the
deviations of the values for individual days from the moving average
should be independent of season of year and reflect short-term variability
only. As an example, the heavy black curves in Figure IV represent the
moving average about which individual daily values fluctuate, and in
Figure V, it can be seen that the seasonally adjusted values plotted
against time, no longer exhibit seasonal variation.
The number of emergency room admissions on a given day can be
affected by conditions which actually have no bearing on actual illness;
e.g., some individuals tend to stay at home, even when requiring immediate
medical attention, because of inclement weather. Since air pollution is
inversely associated with factors such as heavy precipitation, which
tends to disperse air pollution particles, a direct association of
admissions with air pollution may be observed, merely because individuals
stayed at home during heavy precipitation which resulted in low air
pollution levels. Various types of climatological conditions were there-
fore added to the model in order that their effects could be separated
from the effects of the air pollutants.
The addition of the climatological variables to the model resulted in
a total of twelve independent variables as compared to thirty days of
observations for Cook County Hospital in winter. Mathematically, the
addition of more variables to the model will always result in a larger
value of R , which is the percentage of total variation in the dependent
variable explained by the model; however, when the number of variables is
large compared to the number of observations, the estimate of the
-------
321
ssT
i M H H i i ^ 5 s |
Figure V { Chicago cardiac/respiratory emergency
room admissions study. Seasonally
adjusted respiratory admissions ani
S02 levels from October, 1971 through
March 1972 for Cook County Hospital
Figure VI
athv
i
fount emu
TO K 11
»ay A s^
yf
TIM » MHMM1C MU
/
1 SIUCT CKMCI lm
I
\ CSTIMTl HULTIM
[ C'*"iU IN tmiS MM
\
CSTIMTf KtULllUt DU-KCf
IN A.'*ICHT AtR fiUALtTT
\
IAMIC WEI
laio
"\_^ Pathwa
£li|p»rf mmuaci or
unw» I i« TOTH.
rOFMATIO H (tMUMIC
MU
1
ASSOS VMIATIOM V
» uvui or ttmon
I *t HIM «K io»
nuntoii Lt«e.»
V
Est|»ftTC WMMTITATTVI
VWIATrM ID HMTACITT
AND NMKtniTY IN CATEMm
X AT VMrriNF nuuTiOH
LEVEL*
il!?l:ufF Hf«LFI. ffFCCTS 111 |
CATEKHR I: P,1 OU>D
Figure VI : Model for Quantitating and Predicting
Health Impact of Energy Consumption
-------
322
Chicago Cardlac/IMpiratoir tmnymgr •«» AdalMiona Study
Distribution of Hater of EB>. MOB VUlts by DlaanoiU
Hospital la Cook Otmty
•••bar el ToMday* 1" sturty IB 80
(9/14/71-1/27/73, no Intonation for 2/29/72)
Par Cut Avcrao;
of par
Diaanoais Total
Acota •rondiltl* 38.1 1974 24.7
Piwumli 7.1 96* 4.6
Sinitfltll J.5 182 2.1
Acute Traciuobroncbltii 0.0 2 0.0
*cut« MMpintozy 1.1 1« 2.0
HMplratoiy 22.0 1141 14.1
Infarction 0.7 3& O.4
Con9*«tlv* Bcazt rallu* 1.2 62 O.B
Gthar CwdiK 7.1 3M 4.6
Total •Mpiratoxy 91.O 4711 58. »
Total Cardiac 9.0 465 5.8
Total Cardiae/DMpiratorr 100.O S178 64.7
II
Oilcaoo Cardlac/MMplratory •*8ij8«L| MOCBI AdBlscloaa Stody
DtBtribatloa of Nwrtwr of tmrynrr fotm vi>it> by Diapie«l«
13 WaptuOa («*cl»d1m Ox*
Par Otat
Uaanoaia of Total*
Aattaa
Acnta •roachitla
Paaweaia
Siftoaitla
Acuta Trachaobronchitia
Ac&ta •aaplratory
Otter Kaaplntoir
Myocardlal laf aretioa
Ceeoaativa Baart rallvjm
Otter Cardiac
11.1
41.4
>.«
0.4
2.0
0.1
S.4
1S.C
•.7
*.J
773
2400
set
25
1M
It
34»
•22
SM
304
10.1
12. t
7.J
0.2
1.4
0.1
4.«
10.7
7.0
5.0
Total •Mplntezy t*.4 4317 W.7
Total Cudiac M.t 1742 22.t
Total C«rdl«c/*Mpiratorr 100.0 COS* 7».l
• TIM Matter of cfeMrvatioM varla* fto» OM aoapltal to awKlktr. Tkarafoi*.
tte parc«tt of total atia»« im tta mmtmft ptttmmt«a« atn all tka iadtvidaal
koapltala a»d tte awraaa par TaaaJay la tW mm of tte T*M<»y «*»r«9aa
«*ar tte Udlvldval tnapitala.
-------
323
variation of the observations about the estimated regression line or
plane (residual mean square) can increase, particularly if the additional
variables do not account for a large portion of the variation. To find the
"best" predictive model among the twelve independent variables necessitates
12
making 2 , or 4,096, separate regression analyses for each choice of the
dependent variable. Therefore, a stepwise regression procedure was
applied, which ordinarily provides a good, if not optimal, solution.
Briefly, on the first step of this procedure, the independent variable
having the highest correlation with the dependent variable is selected
first, and on successive steps, other independent variables are selected
which have the highest correlation with the independent variable when the
effects of the variables selected on preceding steps are eliminated. The
procedure continues in this fashion until either all variables are
selected, or until the contribution of additional factors is negligible,
according to some criterion.
RESULTS;
The hospitals selected were generally the larger facilities scattered
throughout the city. Admissions to the emergency room for most of them
were relatively small for each diagnostic category day and frequently
reflected the inaccessibility of the private physician because of the time
of occurrence of illness. Cook County Hospital, the largest hospital in
the city, received almost 50% of all admissions, reflecting its use as a
primary care facility by the lower socioeconomic population, mostly inner
city blacks, who generally had no physician or ongoing medical care.
Table I reveals the frequency distribution of admissions by diagnostic
categories, Table II, those of all the other hospitals, except Cook County.
While the pattern generally is similar, myocardial infarction appears to
be more frequent in the other hospitals. This is because at Cook County
these patients are admitted directly to the hospital and do not go through
the emergency room.
The analyses carried out to date for hospitals other than Cook County
reveal some evidence of associations between some pollutants and some
diseases, particularly hospitals like Englewood, which also services a
majority of black patients. The results, however, are not consistent and,
while this may reflect the manner in which the emergency rooms are used,
no conclusions can be drawn. Further, analyses are being carried out, and
-------
324
Chi raw Cardi AC/RHP* r«tanr Aofciioioi.
t-rccoouro E«tl.t«iM Vuiltini in tho
SMMully U]»ud." mpiul l> O
ttoam Etuay. BMt« Jteoraoolc* Mad*l Fbund by
•vtor* 01 Moiuion tor 5.1-Tt.J Condition!. Ml Viria
x* Coiaty. Mora U Hlnttr. •»».
CcnUUMi
ftotlM
ftc*t*
•nncti-
iti*
Pim-
•OBI*
*»«-
CUOiol
lafuct
COM-
R.
roil
T»U1
RUC.
»U1
CM41K
Co*ff?f «td. orrar* reto toot eatff. olff. fm
.»oj
S3
i.»
l.M
1.11
S.OT*
1.01
O.JJ
7.IJ*
• .14
l.«
lt.fl*
.KG.'
0.01
O.OJ
>.u.
ua>'
0.13
0.01
«.»<
-o.u
O.tl
l.K
f«^>
<.3»
0.10
3.V
4.04
0.01
4. SI'
•«t«r
praclp
-«.5I
1.40
>.T>
1.41
l.W
l.U
1
•l~t
-1.41
i.n
l.M
imnm'
-4.M
1.71
i.»
O.Ti
l.OS
O.5I
O.S1
«.»
>.41
-14. It
4. IT
».«•
mn for tuihf iiiilint VuiiblM
•ninic'
O.S7
o.u
i.n
-0.44
0.14
i.»
Mi*
•|Ml
-O.M
O.U
2.t7
K.1.
h«id.
-O.14
0.0*
1.40
0.10
O.OS
l.M
Sin-
-O.Oi
0,0}
l.U'
0.01
0.01
1.4*
Sty
-S^
0.71
O.M
1.10
O.U
O.JO
1.4*
0.20
0.0*
l.OJ*
Cat.
• Ubt
O.U
•i.n
O.M
O.U
0.11
-l.li
0.01
l«.
Dt
v«w
4
4
5
1
«
I
1
•>
0.48
O.JI
0.41
0.04
O.S4
0.40
O.X>
r
5.1J-
3.47
l.»7«
1.17
4.54*
S.»T
».»•
Hodtl wit* *Mll«*t n
cnvt PHOTO »4 mmo.
IMn>0 !• m
1 :.tvt»B 11 « !»*•« of «kMk»r tho •!•• »!«-• fn» tko Honk
Mrly wlnom.
pailtio nlw*
lii mltl« nlon Itialnu nontai
tatta, n*t)*tiv* valui
F viadt, Mtj«tiw
jid this analysis will be limited to Cook County Hospital for the winter
onths (November, 71 - April, 72 and November, 72 - March, 73).
The best models found by the stepwise procedure (smallest residual
ean square) for various choices of the dependent variable are presented
n Table III and summarized in Table IV. They reveal that emergency
ntry for respiratory causes, particularly total respiratory, acute bron-
of ltotmMia.1 Ce*ffiei.Mt* Mtl stMlatlc..!
. HiBtor.
twmuff of T.*.M novlDi]
•a,
of
ftc«M ftrwctilti*
l-l '
1*1 •
orofo-, o.f. . ^MBO2* *
1> op U4.I of Mutlar tko «i»d Mom (n
l^leoto -oourlr »l«o«.
1 ««W ll m U4u of «hoco«kli»ki.
-------
325
chitis and pneumonia, appear to be strongly associated with increases in
SO , but not with CO. On the other hand, congestive heart failure
correlates positively with carbon monoxide and negatively with temperature.
Other factors which appear to add significantly to the same relationships
include absence of sunshine and reduced windspeed for asthma, and reduced
windspeed for congestive heart failure.
The trend of these analyses, in addition to those achieving 5% levels
of significance, is generally in the appropriate direction. This study
suggests that in this population of low socioeconomic blacks, using an
emergency room as primary care physician, there appears to be a relation-
ship between certain air pollutants, particularly SO_, and respiratory
disease, and to a lesser degree between the incidence of congestive heart
failure and CO levels.
These are very preliminary conclusions and additional analyses of the
large amount of data gathered are being carried out.
PREDICTIVE QUANTITATION OF HEALTH EFFECTS OF AIR POLLUTANTS:
The emergency room study sought to further examine relationships of
various categories of diseases of the respiratory and cardiac systems to
different pollutants. Air pollution effects on chronic bronchitics, a
disease known to be affected by irritant gases, have been studied by this
and other groups in the United States and other countries. Additionally,
we have examined the relationship between cardiac and respiratory deaths
during an air pollution episode in the city. Many of these studies found
increased morbidity and mortality at increasing levels of some of these
pollutants, particularly SO and particulate.
The model is proposed as a method, however, crude, to begin to answer
the critical questions posed initially. It is based on a fundamental
assumption (i.e. that there is no population threshold). The environment
is essentially hostile and humans, as biological organisms, with varying
degrees of resistance and adaptive capacities, are in continuous struggle
with it. Any factor in the environment which increases environmental
hostility or anything lowering the resistance of humans, decreases adapt-
ability. Marginal people, that is those with poor adaptive capacity (i.e.
because of serious lung disease, heart disease, asthma, or other advanced
chronic diseases), were the populations examined in most studies of the
effects of air pollutants on humans, since it was postulated that the
-------
326
effects of irritant gases would be strongest on those with inadequate
cardiac or pulmonary reserves.
123
A number of studies, ' ' including the Chicago Registry Study of
4
Bronchitics suggest that there are increased illness and death rates
in some members of the population at relatively low levels of pollution,
with the increases in these rates parallelling the rise in pollution
levels. It, thus, would appear that at every level of pollution, and not
at some defined "threshold" depending upon the adaptive reserve of the
individual, someone becomes sick and someone's life is shortened. While
there may be a biological threshold for tissues, the dose-response
relationship, even for individuals, will vary with ageing, acute illness,
stress, fatigue, increased physical activity, changes in nutrition, and
many other factors. The emergency room and other studies which we also
carried out, suggest the possibility that socioeconomics may very well be
a very significant factor in reducing adaptive capacity. Certainly in a
heterogeneous population there is marked dose-response variability and no
single population threshold can be defined. It, therefore, follows that
definition of the major population groups at high risk and examination of
the variation in risk at different pollutant levels represents a reason-
able method for examining and predicting health impact on a population.
A major weakness in the predictions is the absence of quantifiable data on
low level, long term effects of pollutants on health.
Figure VI represents two pathways followed to estimate the health
effects of various energy systems. Two examples will be given for
quantitating and predicting the health impact of energy consumption.
Pathway "A" represents the steps necessary to define variations in
air quality per unit use of energy, which can then be correlated with the
biological endpoints. This is summarized briefly in Table V, which shows
the fuel consumption in the geographic areas under study. For the
example of respiratory disease, pre-school children in cities over 10,000
population with SO2 levels above 100 mg/m were examined. The estimated
total energy consumption in this area approximates 4.1 quadrillion BTU's.
For the example estimating excess asthmatic attacks, the City of Chicago
data were used. Here the total estimated energy consumption equals 1.2
quads per year.
The human health assessment track is shown in Pathway B, Figure VI.
The geographic area to be studied consists of a total population of 21
million people. From a number of studies it has been shown that approxi-
-------
327
KTBHW HJEL OMStffTIW
IN tEOMMC ME*
IN PIC-SCHOOL anua
ESTMnto aaet oMSurriON (IN GUVS) w FUB. JM> use
USE ooEemr. 1369, CITIES oat 30.000 PVUXTICN MTH
us/f?
WE
MB
coteciN.
»««««.
lUETUC
CBEMTION
TCTM4
CO*.
0.1
0.3
O.g
1,1
on.
0.6
0.3
0.1
1.1
as
0.7
0.9
0.1
2.0
TOW.
L3
L5
1.3
1.1
rwra /61HTIT1C ATKB
ESHWIB) TOTM. B«»er ODNSUrTlOH IN CH1CMO • 1.2 ONE/VEM
mately 1.07 million of these are pre-school children. Three studies
summarized in Figure Vllfharacterize the change in incidence of acute
respiratory disease in children under five correlated with variations in
annual SO. levels. The results are used to assess the variation in
Figure VII
IKIMKE of ACUTE tesf IHAIOKT BtitAst IN
CMUJBEH (hue* 5 COHIELATES WITH AMIML
S02 LIVEU
B 36Z
- 32J
g 2S
S ?*I
a 201
ii
| V
IW
A
X
X x
x "
X
X
75 105135 1S5 195 ZZ5255
IS
SOUKC: U) cxct* ETA, 197J.
U) J.M.I. MUCUS.* liEA HALLM. M. J-
(C)
-------
328
Morbidity at high and low S(>2 levels. This was found to be 19% (5-24%) .
Thus, it appears that an excess of 255,000 respiratory infections might
occur annually in this population of 1.07 million children under the age of
CamiMTiM OF SOj TO TOTAL AMICT Ai*
POUJITIM LEVEL* IT FSEL MB USE CATESOKIES, 1969,
CITIES OVE* 10,000 POPULATION WITH SOj LEVELS AMVE
UK
REIIDEWTIAL
, MB
GMBKIAL
ImrmiAL
'ELECTRIC
NHQt
CCHEJUtTIM
TOTAL*
COM.
3.2Z
55.81
M.OZ
53.01
OIL GA*
13.«
1.5X
l.«
16.71
TOTAL
17.41
37.3X
IS.tt
7P.OX"
HBTI: 30Z or AMIEMT AIM LEVELS or POLLUTION AM
ESTIMTED TO RESVLT MM WM-OCKY SOURCES
five at the higher pollution levels. The number of excess illnesses per
quad BTO's of energy can then be estimated when one considers the percent-
age contribution of SO2 to total ambient air pollution levels by fuel and
use categories. (Table VI) This is summarized in Table VXI. An excess
of 00,000 respiratory infections results from coal use in this population
for residential and commercial, 305,000 from industrial use,
45,000 from electric power generation, with a weighted average for coal of
120,000 excess acute respiratory infections per quadrillion BTU's.
flm. hmoni 21 ItujaA
kuTm comtn*TiM SJBIB Pf _am_E
or MO. tOUBITIOl W ElTCCT COM. OlL GM
oaiorr »2 tna» tP
•9,000 37,000 -
37.3 M*EME • 335,900 13.000 - 65,000
t 15.*
•isaot • *5,000 3S,000 - 30,000
».0 SsiAsR""*1** 120,000 39,300 - «,000
•Bl MNBI S TEARS tf ABE
I MBATIVC COTINWttM OT EACN SSMCE
-------
329
Figure VIII
Another example examines excess asthma attacks in relation to sulfates.
The area defined is the City of Chicago with a population of 3.4 million.
In this population, the asthma prevalence rate is conservatively estimated
to be 3% for a total of 100,000 asthmatics in the city. The graph in
Figure VIII assesses variation of morbidity in samples of asthmatics at
varying pollution levels taken from the CHESS and Cohen Studies which
examined the relationship between sulfate concentrations in the air and
excess asthmatic attacks. No excess was found below 7.5 ug/m . An
increase in the rate of excess asthmatic attacks per person day, accom-
panied the increase in sulfate concentration. Figure IX shows the time
distribution of airborne sulfate concentration (number of days per year
Figure IX
-------
330
at various concentrations of sulfates for Chicago). An estimate of the
quantitative variation in asthmatic attacks at varying levels of sulfate
can then be made, using the following formula.
oo
TOTAL EXCESS f
ASTHMATIC ATTACKS * J (c) (c)
O
n
X Ei • Di A ci
£ 763,000 EXCESS ATTACKS/YEAR
A - TOTAL NUMBER OF ASTHMATICS IN A COMMUNITY
E, = EXCESS ASTHMATIC ATTACKS/PERSON - DAY
D = DISTRIBUTION OF TIME WITH RESPECT TO CONCENTRATION
OF AIRBORNE SULFATE
f DAYS/YEAR ~l
A SULFATE CONCENTRATION J
CONCENTRATION OF AIRBORNE SULFATE
From this calculation it would appear in this example that 763,000
excess asthmatic attacks occur per year due to total sulfate exposure.
Using the following formula it is suggested that 445,000 excess asthmatic
attacks might be expected to occur per quadrillion BTU's of energy from
sulfate producing sources.
i EXCESS ASTHMATIC ATTACKS X% CONTRIBUTION OF SULFATES
ESTIMATED ENERGY CONSUMPTION
= # EXCESS ASTHMATIC ATTACKS
QUAD
This model can be used to examine other populations at high risk and
the variation in risk accompanying a change in the environmental quality.
A better definition of those segments of the population at high risk and
more sophisticated studies measuring variations in health effects of
multiple pollutants in cohorts of these people will increase the accuracy
of the model. Better measurements of more pollutants to better quantify
environmental stressors also increase its accuracy. While reliability of
-------
331
the model is severely limited because of the data, more adequate measure-
ments of both environmental and biological endpoints will enable studies
to be carried out which will more accurately quantitatively assess health
impact.
In Conclusion:
1) A Model is suggested for making predictive quantitations of the
health impact of changes in energy use with accompanying changes in levels
of pollution;
2) The excess numbers of asthmatic attacks, 455,000/quad, and acute
respiratory illness in pre-school children, 120,000/quad, which might be
expected to occur suggest that increases in energy use may very signifi-
cantly affect the health of large segments of the population.
3) Studies similar to the one discussed and additional studies of
variation in morbidity and mortality to define the effects on high risk
groups at low and high levels of pollution permit the introduction of
health impact considerations into decision making regarding energy
alternatives.
Finally, it must be recognized that infinite growth in energy use in
a finite system is not possible. Industrial societies soon will be
required to accept a state of energy equilibrium, i.e. the new use of a
quantum of energy to serve an important societal need will require the
phasing out of an old use, considered less important and less necessary.
REFERENCES
1. Carnow, B. W. and Carnow, V.: "Mr Pollution and the Concept of No
Threshold" chapter in Advances in Environmental Science and Technology,
Vol. 3, published by John Wiley and Sons, N.Y., 1973.
2. Glasser, M. L., Greenburg, and Field,: Mortality and Morbidity During
a Period of High Level Air Pollution. Arch, of Environ. Health
15;364, December, 1967.
3. Health Consequences of Sulfur Oxides, A Report from CHESS, 1970-71,
EPA 650/1 74-004, May, 1974.
4. Carnow, B. W., Lepper, M. H., Shekelle, R. B., and Stamler, J.:
The Chicago Air Pollution Study: SO Levels and Acute Illness in
Patients with Chronic Bronchopulmonary Disease, Arch, of Environ.
Health, Vol. 18, May, 1969.
Work Su-.'-'Oj-to^ by : St-.tc- of IMinoi - In--tit'.:te for iiivii-nr: e •: t "1
V Tity ~r.r; the Chic-, -o L'ir.-- A:-.-oci~- 'lion
-------
333
L'EVALUATION DU RISQUE POUR LA SANTE DE LA
POPULATION EXPOSEE AUX POLLUANTS ATMOSPHERIQUES IRRITANTS,
ASPECTS METHODOLOGIQUES
B, BARHAD ET M, CUCU
Institut d'HygiSne et de Sante Publique, Bucarest, Roumanie
RESUME
Les recherches epidemiologiques effectuees dans des loaa-
lites & atmosphere polluee par des agents irritants (suspensions
et dioxyde de soufre) ont permis 1'estimation de la valeur de
la sensibilit£ de certainesmethodes appliquees dans les Etudes
sur la population^ dans les conditions ou le niveau de pollution
de I'air n'est pas trap Sieve.
Des methodes utilisees3 la correlation statistique de la
morbidite et de la mortalite de la population s'est averse moins
sensible que la comparaison statistique des territoires pollues
et non pollues. De meme les etudes effectuees sur des groupes
de population a sensibilite accrue envers I'aotion des polluants
irritants se sont averees d'une plus grande valeur que I'analyse
de la mortality et de la morbidity de la population enti&re.
Nous avons obtenu ainsi des resultats oonoluants par I'etude
de la prevalence de la bronchite chronique et des symptomes res-
piratoires chez la population adulte, a I'age de risque (40-60
cms), Dans le cas de I'investigation de I'action de la pollu-
tion atmospnerique sur la sante des enfants, les resultats les
meilleurs nous les avons obtenie akez le groupe d'age de 8-11 ans.
Les tests epid&mioloaiques utilises chez ce groupe d'age ont ete:
I'etude des causes de I'absenteisme seolairet la prSvalenoe des
-------
334
maladies et des symptomes respiratoires3 ainsi que I'investiga-
tion des modifications de certaines aonstantes hematologiques,
ABSTRACT
Epidemiological research carried out in localities where
the atmosphere is polluted by irritants (suspended matter and
sulphur dioxide) has enabled us to estimate the sensitivity of
certain methods used for studying the population under conditions
in which the level of air pollution is not too high.
Of the methods used, statistical correlation of the morbi-
dity and mortality of the population proved less sensitive than
statistical comparison of polluted and non-polluted areas.
Similarlyf studies carried out on population groups highly sen-
sitive to the effects of irritant pollutants proved of greater
value than analysis of the mortality and morbidity of the whole
population.
Thus we obtained conclusive results by studying the preva-
lence of chronic bronchitis and respiratory symptoms in the
adult population most at risk (age 40-60). With regard to the
study of the effects of atmospheric pollution on child health,
our beat results were obtained with the 8-11 age group. The
following epid&miological tests were used for this group: the
study of reasons for absence from school* the prevalence of res-
piratory illnesses and symptoms and the study of variations in
certain haematological constants.
-------
335
Des amples recherch.es epidemiclogiques effectives dans
des locality's a atmosphere pollute par des agents irritants,
represented surtout par des suspensions non toxiques comma
polluants majeurs et par le bioxyde de soufre et le bioxyde
d*azote comma polluants secondaires, nous ont permis lfesti-
mation de la sensibilit6 de certalnes m^thodes appliqu^es
dans les Etudes sur la population. Les zones 6tudi4es u'e-
taieuat pas somnises a une pollution excessive, fait demontr6
par 1'enregistrement continuel des principaux polluants.
Dans ces conditions, les etudes de correlation statis-
tique se sont av6r6es moins sensibles en ce qui concerne
1'effet de la pollution de I1air sur la aante de la popula-
tion, vis-a-vis des Etudes de comparaison statistique, par
rapport a une population temoin d'une zone faiblement pollute
en respectant eVidemment avec rigueur toutes les conditions
de selection correcte des echantillons. Aussi bien, pour les
conditions respectives, I1 analyse de la morbid.it6 et de la
mortality de la population en general n*a pas r^ussi a, mettre
en evidence des effets notables sur la sent6, le ph^nomene
e*tant reflate par 1'investigation de I'Stat de sant6 de cer-
tains groupes de la population a sensibility accrue envers
I1 action des polluants atmosph^riques.
Dans le cas de I1exposition aux polluants irritants
aentionne's on a obtenu la mise en Evidence de I1 influence sur
la sante tant de la population adulte que de la population
infantile.
La population adulte la plus sensible s'est averse le
groupe entre 4o-6o ana, ou on a pu mettre en Evidence une
frequence accrue des broncnopathies chroniques non sp6cifi-
qoes et des symptomes respiratoires. Afin de mettre en Evi-
dence les malades et les symptomatiques, on a utills6 le
questionnaire propose par le Medical Research Council, avec
quelques modifications, et 1*investigation de la fonction
respiratoire a I1 aide &• 1'expiregrapiie Soddard. L« inves ti-
gation de la fonction respiratoire apres 1' application des
bronchodilatateurs augmente la sensibilite" de la m^thode. La
mfithode s'est av^rde adequate a l^tude, mais elle porte
-------
336
1'empreinte des difficulty's techniques (1* interview Indivi-
dual, le rendement bas de 1* examination spirographique) ffait
qui rend difficile la poursuite de grands groupes. Compte
tenu des difficult6s techniques et de la dimension relative-
ment reduite des groupes, 1*Influence de la pollution sur la
prevalence de la bxonchite chronique et des symptomes respi-
ratoires eat parne plus concluante dans le cas de la pour-
suite prospective des groupes studies.
En ce qui concerne les 6tudes sur la population infan-
tile, le groupe le plus sensible s'est av6r6 celui compris
entre 8-11 ans, d'oh on obtient une bonne cooperation et on
evite le risque de 1'habitude de ftimer comme facteur inter-
far ant. Dea methodes utilises pour la mise en Evidence de
1*existence de 1*influence de la pollution de I1 air sur la
saute* des enfants, 1*6tude de 1'absente'isme scolaire et celui
de la frequence des maladies et des symptomes respiratoires
et - en partie - des modifications hematologiques sont cel-
les qui ont donne1 des resultats concluants.
Llabsente'isme scolaire et ses causes ont 6t£ e'tudie'es
a I1 aide d'une fiche individuelle des e*leves, qui a Ste1 rem-
plie par les cadres didactiques, avec le support de 1'infir-
miere de iMcole. Poursuivie assez longtemps, au moins pour
une ann£e scolaire, cette fiche a riussi a, mettre en Eviden-
ce certaines particularites de la pathologic des Ecoliersj
les maladies pr^sentant des differences significatives du
point de vue statistique etant surtout les affections aigues
et chroniques des voles respiratoires sup^rieures et du pou-
mon. Les autres causes de maladie n'ont pas present6 de dif-
ferences entre les zones pollutes et les zones temoin, sans
pollution* fitant donn6 qu'il s'agit d'une methode simple et
facile a appliquer, nous la oonsid^rons - malgre' son manque
de sp^cificiti - utile pour apprecier 1'effet de la pollution
atmospherique snr la sante*
La frequence des maladies et des symptomes respiratoi-
res a e*t6 e'todle'e a 1'aide d'un questionnaire rempli par les
parents et a 1'aide de l*examen de la fonotion respiratoire
chez les enfants des groupes e'tudie's.
-------
337
Le questionnaire, comprenant des questions sur les
symptomes respiratoires chez les enfants, sur les ante'ce'-
dents pathologiques pulmonalres personnels et he're'docolate'-
raux, ainsi que sur certain«s conditions sodales (le degre
d'instruction des parents et les conditions d1habitat, y
compris le syateme de chauffage) ont 6t6 remplis en utili-
sant la mdthode de 1*interview en groupe (2o-3o parents),
que nous trouvons sup^rieure a la mdthode de 1'interview in-
divlduel ou de I'autoadministration. Nous avons pr^f^rS
cette mdthode puisque, en la testant parallelement aux deux
autres, nous avons obtenu un rendement suplrieur par compa-
raison a I1 interview individuel, en evitant les erreurs les
plus fre"quentes observ^es dans le cas de I'autoadministra-
tion. Cette mdthode nous a permis aussi 1'examination durant
le mdme jour de la fonction respiratoire, en nous permettant
ainsi d'obtenir un taux des r^ponses sup^rieur (presque 9o#),
Dans 1*investigation de la fonction respiratoire nous avons
utilise1 la de*termination du d£bit respiratoire maximal, en
utilisant le peack-flow-meter Wright, ainsi que la capacite*
vitale force'e et le debit expiratoire maximal. Pour ces
deux mesures nous avons utilise le vitalographe que nous
trouvons satisfaisant en ce qui concerne le rendement et la
sensibilite1, Afin de simplifier les enregis trements et d*aug-
menter le rendement, nous sommes en train de tester 1'appa-
reii Monaghan aussi.
L*interpretation de ces donates comporte certaines
difficulty's, sur tout en ce qui concerne les valeurs normales
des fonctions respiratoires, lvStablissement des valeurs
normales nationales represent ant une de nos preoccupations
actuelles.
Les examens h^matologiques n'ont pas pr6sente, par
leur manque de specificit^, une sensibility particuliere
dans 1*appreciation des efrets de la pollution de I1air sur
la sant4 des enfants. Nous signalons pourtant plusieurs
tests h^matologiques qui ont indiqu6 - avec une certaine
Constance - des differences chez les groupes d'enfants des
zones polluees et non polluees, a savoir la quantite d*hemo-
-------
338
globine des hematics, la frequence dee lymphocytes grands par rapport
aux petits, ainsi que la presence des corps toxiques dans lee hematics
Nous considerons que 1'etude epidemiologique des groupes de
population a sensibilite accrue permet 1'evaluation des effets de la
pollution de I1air sur la sante, meme dans des conditions ou le
niveau de la pollution n'atteint pas de valeurs excessives.
References
1. COLLEY, J.R.T., "respiratory Disease in Children", Brit, med.
Bull., 27. 9 (1971).
2. FREOUR, P., COUDHAY, P. BARRIERE, P., "Symptomatologie broncho-
respiratoire dans une population urbaine", Rev. Tuberc.
Pneuaol., 31, 1139 -(19&7).
3. HOLLAND, V., HALIL, T., BENNETT, A.E., ELLIOT, A., "Factors
Influencing the Onset of Chronic Respiratory Disease", Brit, med.
J.. 205 (1969).
4. LAWTHER, P.J., MARTIN, A.E., WILKINS, E.T., "Epidemiology of Air
Pollution", Proc. Soc. Roy. Med.. 57. 969 (196*+).
5. MARTIN, A.E., "Mortality and Morbidity Statistics and Air Pollu-
tion", Proc. Soc. Roy. Hed., 57. 969 (196M.
6. RACOVEANU CARMEN, "Cercetari epidemiologice aeupra bronsitei
cronice In doua zone fara poluarea atmosferei", Stud. Cercet.
Hed. int., 12. 35^ (1971).
7. REID, D.D., FLETCHER, C.M., "International Studies in Chronic
Respiratory Diseases", Brit. Med. Bull.. 2?, 59 (197D-
8. SCHMIDT, P., PELECH, L., "Changes of some Diagnostic Indicators
in Children from Regions with Varied Degree of Atmospheric Pollu-
tion", Acta hygiena, epidemiol, microbiol., 15, 52 (1973).
9. x X x "Long Term Programme in Environmental Pollution Control
in Europe", WHO Regional Office for Europe, Copenhagen (1971).
x X x "The Health Effects of J
for Europe. Copenhagen (1968).
1C. x X x "The Health Effects of Air Pollution", WHO Regional Office
-------
339
STUDY ON LONG-TERM EFFECTS ON HEALTH
OF AIR POLLUTION
J, KUMPF+, M. ARHIRII + , B, GRAB"*"*" AND M, J, SUESS*
+ World Health Organization, Regional Office for Europe,
Copenhagen, Denmark
++ World Health Organization, Division of Health Statistics,
Headquarters, Geneva, Switzerland
ABSTRACT
Epidemiological studies are required -in order to ascertain
the long-term effects on health of exposure to relatively moder-
ate levels of pollution of ambient air.
To avoid the influence of other factors, such as occupa-
tional exposure and smoking, studies on children have been re-
commended.
School children, 8-10 years old, living in areas with
different levels of air pollution, divided into "highly polluted"
(annual median concentration of SO more than 100.ug/mS and an-
nual median concentration of standard smoke over 40 ,ug/m2) and
"low polluted" (S02 below 50/ug/m3; standard smoke below 30 .ug/
m ), are investigated by means of a questionnaire filled in by
parents or guardians, and by means of lung function tests (Peak
Flow Rate, FEV 0.75 and FVC).
Using a common working protocol and reporting forms, studies
ave currently carried out in five European countries. The re-
sults obtained in 1972 and 1973 studies will be analysed by a
Working Group to meet in Dusseldorf in April 1974.
-------
340
1. Introduction
Within the scope of its long-term programme in environ-
mental pollution control, approved by the Regional Committee
for Europe of the World Health Organization in 1969 (1/2), the
WHO Regional Office for Europe has organized a study on the
long-term effects on health of air pollution (3).
The objectives of the study are:
(1) to investigate the relationship between long-term exposure
to air pollution and indices of chronic respiratory
disease in children living under different social and
domestic circumstances in a number of European countries;
(2) to evaluate the relationship between air pollution and
respiratory disease with a view to providing additional
criteria for air quality standards.
2. Method used
While more reliable data were expected to result from a
prospective (longitudinal) study, it was decided that, in view
of the time and expenditure required for such a study, the
prevalence should first be investigated by a cross-sectional
study, organized so as to permit a follow-up to be made in a
prospective or longitudinal study. The parameters were limited
to those thought to be easily quantifiable and comparable.
2.1. Air pollution
While several pollutants have been found to exert adverse
influences on health, two pollutants were selected for the
purposes of this study, viz. S02 and "standard smoke". These
have been found to be typical of the reducing type of air pol-
lution in urban areas in Europe, causing or aggravating respir-
atory symptoms. Tnese are also the pollutants measured most
frequently and for the longest period of time. If other pol-
lutants were measured, they could also be recorded and reported.
-------
341
To guide local teams in the selection of study areas, two
levels of pollution were defined:
(1) High-level pollution: median annual concentration
SO2> 1OO mg/m
and median annual concentration standard
smoke > 4O mg/m
(2) Low-level pollution: median annual concentration
S02<50 mg/m3
and median annual concentration standard
smoke <30 mg/m
In each country participating in the study, at least two
areas, one with high and the other with low pollution, were
selected. Since chronic health effects were to be investigated,
the air quality records, for at least the previous five years,
were thought to be necessary, especially for the areas with high
pollution levels.
The methods recommended for SO_ measurement were the acidi-
metric hydrogen peroxide method, with automatic sampling and
potentiometric titration - OECD (4) - and WHO (5,6,7,) or the
West and Gaek colometric method, the former being used as re-
ferences method if other procedures for SO measurement were
employed.
For the determination of standard smoke, the reflectance
method (5) was suggested.
The density of the measuring network was to be increased
where necessary through the installation of new stations, each
area to be covered by at least one station. The frequency of
sampling was already determined by the methods of analysis
selected, i.e. 24-hourly values for SO_ and standard smoke. Of
great importance was the location of the sampling site, which
had to be close to the level of inhalation by the population.
-------
342
The description of the network, the instruments and methods
employed, the tables of all available air quality data, the
summary of the data (monthly extreme and mean values) the 9O
and 98 percentiles and the median were to be stated.
2.2. Medical and socio-economic data
Schoolchildren 8 to 1O years old were selected as being a
fairly susceptible group, generally uniformly exposed to ambient
air pollution, not yet affected by smoking and puberty, and no
longer under the direct influence of infectious respiratory
diseases of early childhood. Children of this age group could
also be expected to collaborate well in the pulmonary function
tests and their relatively short life span would permit their
anamnesis with respect to respiratory afflictions to be rather
well established.
The sample proposed for each pollution area was 2OOO child-
ren, i.e. 4OOO for each country (two areas in each country) .
To standardize for socio-economic conditions, the educational
attainment of the parents and the details of the households were
to be recorded (i.e. number of rooms per household, method of
heating, number of persons per household.)
Within the study areas, random samples of schools were to
be chosen and, within the school classes selected, all children
would be investigated. Special schools attended by children,
the majority of whom did not live within the study area, were
to be excluded.
The studies were conducted in spring, starting almost at
the same time and taking about two months.
The personal and medical data were obtained through a
self-administered questionnaire and a medical examination of
the child.
-------
343
The questionnaire, accompanied by an explanatory letter,
was sent to the parent or guardian who was to provide the per-
sonal, socio-economic and health data on the child. These
data included information on the presence or otherwise of res-
priatory symptoms during the previous winter season, the pre-
sence of such symptoms in previous years and the child's history
with regard to certain respiratory and infectious diseases.
The medical examination yielded, in addition to personal
data, information regarding height and weight, the peak res-
piratory flow rate as measured Wright peak flow meter - (five
tests; the flow meters were standardized at the London School
of Hygiene and Tropical Medicine), the forced vital capacity
(FVC) - (three times, preferably by wet spirograph) and the
forced expiratory volume in O.75 seconds (FEV O.75).
3. The pilot studies
To test the working protocol for the study, pilot studies
were conducted in two countries in 1972, taking one school in
an urban district in one country (Denmark), 331 children in the
7-11 year age-group 295 of which completed the investigation
and, in another country (Poland), 991 children aged 8-1O years
from two schools in a smaller low-level pollution area (monthly
means September 1972 - January 1973 at two stations: 5-15O .ug/m3
SO 2 and 4O-25O ,ug/m smoke ) and from three schools in a larger
high-level pollution area (monthly means September 1972 - Jan-
uary 1973 at two stations : 4O_ 19O ,ug/m3 SO2 - peak >4OO, and
80-24O ,ug/m smoke - peak 120O) .
Difficulties were encountered with the FVC and FEV measure-
ments , leading to the adoption of the peak flow measurement as
the compulsory measurement and the FEV O.75 and FVC as optional
tests.
The study protocol was adjusted accordingly by a working
group which discussed the results of the two pilot studies (8).
-------
344
Contact was established with the Commission of the European
Communities in an effort to achieve a common approach to the
studies conducted by both organizations and to obtain comparable
results.
4. The main study
In 1973, two countries conducted the main studies (Denmark
and the Netherlands) and another two started pilot studies
(Czechoslovakia and Poland) to test the questionnaire in their
particular circumstances. However, since these studies were
undertaken using the corrected working protocol and were similar
in size and approach to the other main studies, their results
are thought to be comparable to the studies conducted in
Denmark and the Netherlands and were evaluated accordingly. A
fifth country is to conduct the main study in 1974 (Poland).
The Danish study was carried out in spring 1973 on 85O
children in a high-class district of the capital (annual means
3 3
1971/72: 79/72 Aig/nT S(>2 and 37/32 ,ug/m smoke) , 1300 child-
ren living in a working-class district of the capital (annual
means 1971/72: 87/66 ,ug/m SO, and 31/26 ,ug/m smoke) / both
at the same levels of air pollution, on 16OO children in a low-
pollution urban area (annual means 1972: 28 /ug/m smoke) and
finally on 8OO children in a rural area with very low pollution
(annual means 1972: 9 ,ug/m SO. and 7 ,ug/m smoke). The data
which are not yet fully evaluated, do not seem to indicate a
clear- correlation between prevalence of respiratory diseases
and air pollution when results are adjusted for social class
differences.
In the Netherlands, 21O4 children aged 9-11 years, were
investigated in a high-level pollution area (annual means at
seven stations: 1OO-19O .ug/m SO. and 25-45 ,ug/m smoke) and
276 in a low-level pollution area (annual mean at one station:
SO /ug/m SO- and 1O /ug/m smoke). A general trend towards
higher prevalence of respiratory symptoms was found in the high
*
pollution area, although the findings did not reveal clear and
significant differences.
-------
345
The study in Czechoslovakia was conducted on 4276 children
^.ged 8-11 years in two highly polluted areas (annual mean SO
concentration 1969/70/71: 170/130/150 ,ug/m3 and 12O/140/170,ug/in
respectively; estimated mean air-borne dust aerosol in 1970:
109 and 170 /ug/ra respectively^, and in four less polluted
/ 3
areas (annual mean SO_ concentration for 1972: 4O/8O/106/75,ug/m
respectively? estimated air-borne dust aerosol in 1970 - not
stated/76/105/99 ,ug/m respectively). Preliminary analysis
shows significant differences in the prevalence of respiratory
symptoms as between high-level and low-level pollution areas.
The study conducted in Romania in 1973 yielded preliminary
indications of a higher prevalence of respiratory symptoms in
a high-level pollution area (1974 children) as compared with a
low-level pollution area (1918) children).
5. Evaluation
A numerical evaluation of the study results is at present
underway, and a preliminary assessment (9) of the results has
shown that the trends observed will be established more clearly
by a longitudinal (prospective) study which is being set up for
the next three to five years.
The cross-sectional study showed that methods and equipment
have to be closely standardized to exclude experimental bias.
While this appears to be easier for the medical part of the
study, if the peak flow rate is taken as the significant meas-
urement, no comparison is possible between countries in respect
of air quality measurements obtained. Although air pollution
levels in the various study areas within a country can be com-
pared, standardized equipment will have to be provided to make
valid inter-country comparison possible. This should go hand
in hand with the training of equipment operators.
During the cross-sectional study all Wright peak flow
meters were calibrated at one place (London School of Hygiene
and Tropical Medici-.e) , whereas no air quality measurement
-------
346
equipment was calibrated. However, even with the calibrated
flow meter, the questionnaire replies to questions on respira-
tory symptoms seem to yield more conclusive answers than the
measured expiration values in relation to levels of air pollution.
In international study special attention had to be paid
to language problems. Since the standard questionnaire had
been prepared in English and was to be used in the local language,
the correctness of the translation into the local language was
checked by having it re-translated into English by a translator
not connected with the study. In spite of this precaution, 4
questions were consistently replied to in an unexpected way in
one of the study areas.
Consultants visited the various pilot areas during or after
the studies. It is planned to organize such visits before and
during the longitudinal study.
References
1. Long-term programme in environmental pollution control.
Report submitted to the 19th session of the WHO Regional
Committee for Europe (Doc. EUR/RC19/1O).
2. Long-term planning in the field of environmental pollution.
Resolution of the 19th session of the WHO Regional
Committee for Europe (Doc. EUR/RC19/R5).
3. World Health Organization, Regional Office for Europe (1973)
Long-term programme in environmental pollution control in
Europe. The long-term effects on health of air pollution.
Report on a Working Group, Copenhagen, 14-18 February 1972.
4. Organization for Economic Co-operation and Development
(1964) Methods of measuring air pollution. Report of the
Working Party on methods of measuring air pollution and
survey techniques, Paris.
5. Katz M., (1969) Measurement of air pollutants: guide to
the selection of methods, Geneva, World Health Organization.
6. World Health Orhanization, Selected methods for sampling
and measuring sulfur dioxide (Doc. WHO/EP/7O.2).
7. World Health Organization, Selected methods for sampling
and measuring particulate matter in air (Doc. WHO/EP/71.5) .
-------
347
8. World Health Organization, Regional Office for Europe (1973)
Working Group on the Study on Chronic Respiratory Diseases
in Children in relation to Air Pollution, Rotterdam,
26-28 February 1973. Summary report.
9. World Health Organization, Regional Office for Europe
(1974) Working Group on the Study on Chronic Respiratory
Diseases in Children in relation to Air Pollution,
Diisseldorf, 17-19 April 1974. Summary report.
-------
PANEL DISCUSSION
-------
351
SUMMARY OF THE DISCUSSION
This panel, chaired by F.G. Hueter, addressed itself to
the subject of evaluation of health risks of populations exposed
to irritant atmospheric pollutants. The emphasis in this regard
was on the methodologic aspects of these human population
studies.
The chairman initiated the discussion period by identifying
two major considerations which had come to light through the
formal presentations:
- one consistent thread through the formal presentations was
the use of health questionnaires and respiratory function
studies to elicit human health effects data. The chairman
questioned the relative importance of developing and applying
other non-respiratory and more sensitive biologic parameters
for the assessment of adverse air pollution health effects;
- another consideration which was highlighted during the formal
presentations was the importance of the sub-groups used in
population studies. Selecting appropriate, susceptible,
populations bears directly on the validity and applicability
of the resultant data.
There was general agreement among panel members that the
epidemiologic study, as a methodologic approach to the assessment
of air pollution effects, has been perfected to the point that
the variables employed can be very carefully controlled. In a
discussion of populations to be considered, for instance, the
group agreed that special populations, such as survival popula-
tions, can be identified and avoided at the outset for a given
study.
The group recognized that there is still an insufficient
data base for making decisions on air pollution control, and, in
-------
352
comparability in human population studies. If standardized
methodologies can be developed and applied, then data from
many different countries can be correlated to provide a
firmer data base. The group also recognized that for certain
types of studies, world-wide uniformity of methodology is not
now feasible. Because of such factors as socio-cultural
orientation, differences in nutrition, differences in genetics,
and the like, certain populations cannot be compared.
A second general concensus was that the effort towards
comparability can be enhanced considerably by international
cooperative studies. It was agreed that such a pooling of
expertise could result in great strides in the control of air
pollution.
DISCUSSION
HUETER (U.S.A.)
A number of the presentations have contained a thread of
continuity, but, on the other hand, there have been obvious
differences in experimental approaches used by various inves-
tigators. Time does not allow a discussion of all these
differences, but I feel that there are two major areas which
the panel could address. One of them is that most reports con-
cern the health questionnaires and respiratory function measure-
ments. Several reports also included hematological and
enzymatic, bacteriological type of data. I would like the panel
to consider what is the importance of developing and/or applying
non-respiratory or other more sensitive biological parameters
for assessing both qualitatively and*quantitatively the adverse
-------
353
health effects from exposure to air pollutants. The second
point is that various investigators have studied health effects
using different subgroups of the population. The question is,
"what are the most appropriate susceptible populations to study
for discerning and quantifying adverse health effects of air
pollution". I would like to request Dr. Carnow to begin the
discussion on the aspect of questionnaires and respiratory
function.
CARNOW (U.S.A.)
Everybody is very defensive about questionnaires and I
would like to raise the question concerning more sensitive
parameters. We used questionnaires to define the levels of
inadequacy and disease of 56O bronchitics in the City of Chicago.
We followed these people for some five years and our data have
been published. In relation to the integrity of the question-
naires, we carried out some validation studies and found a
number of interesting things. There were very high positive
correlations between the dyspnea index (developed by the British),
the cough-phlegm index, frequency of past history of illness
index, and spirometry. We also found a very high correlation
with subsequent mortality, in that 8O% of all deaths were
cardio-respiratory and most occurred in those we had classified
as Group 4 - severe bronchitics. Classification was based solely
on information from the questionnaires. Thus, questionnaires
appear to be at least an important prognostic and, perhaps, a
diagnostic tool.
There is more regarding the validity of questionnaires.
We analyzed 36 months of data and found that those people who,
in the initial questioning, said that they were sensitive to
high humidity, indeed proved to be sensitive. There was a
significant relationship between their illness over the 36
months and days of the highest humidity. Those who said they
were sensitive to heat appeared to be more sensitive and had more
-------
354
illness episodes related to heat. Those who reported that they
suffered from frequent respiratory illness, were found to have
a much higher frequency of respiratory illnesses.
In many ways, I think people know about their environmental
sensitivities. Responses to questionnaires can be extremely
valuable if they are standardized and if personal questions
are not included by the questioners; who must be trained for a
short period of time to ask standard primary and secondary
questions. In summary, I think that questionnaires should not
be thought of as crude or of little value. They are a very
good way to obtain data.
SCHMIDT (CSSRJ
Wir haben heute zum Thema Asthma und ins besondere zum
Problem der auslosenden Faktoren zahlreiche Beitrage gehort.
Eine weitere MSglichkeit in diesem Zusammenhang
bletet die Ueberwachuhg des Arzneimittelverbrauchs. fiekannt
lich best eh t in unserein Land ein zentralisierter Gesundheits-
dienst, was dem Kollegen Hruby aus Briinn die M5glichkeit gab,
den Verbrauch von Antiasthmatika in Regionen mit starken und
relativ geringer Umweltverschmutzung zu vergleichen. Dabei
zeigte sich, dass der Verbrauch in den verschmutzten Gebieten
htther war.
In der Frage, welche Kriterien bei Untersuchungen von
Kindern oder anderen Bevblkerungsgruppen am besten zugrunde zu
legen sind, sowie in der Frage der Bedeutung der sich aus
LungenfunktionsprUfungen, epidemiologischen Fragebogenaktionen
und spezifischen Kinderuntersuchungen ergeben den Abweichungen
bei anderen biologischen Parametern vertreten wir die Ansicht,
dass die meisten oder alle unsere Bef uncle physiologischer Art
sind. Es handelt sich also nicht urn pathologische Verander-
ungen. Die festges tell ten Veranderungen fiihren jedoch zu der
Eukenntnis, dass der Organismus des Erwachsesen bzw. des Kindes
-------
355
einen Langzeitstress ausgesetzt 1st. Die Frage, welche
Altersgruppe sich am besten fur Untersuchungen eignet,
1st, wenn es sich urn Untersuchungen auf der Grundlage von
Anamnesen oder Fragebogen handelt, nicht von groBer Bedeutung;
werden jedoch gezielte klinische und Laboruntersuchungen durch-
gefvihrt, dann halten wir aus folgenden Griinden die Altersgruppe
zwischen 7 und 12 Jahren fur die geeignetste:
1. Kinder haben zwangslSufig bereits gewisse Zeit unter
bestimmten Bedingungen gelebt, d.h. entweder in einer Region
mit oder ohne Umweltverschmutzung.
2. Kinder arbeiten bei den Untersuchungen willig mit.
3. Kinder rauchen nicht, trinken keinen Alkohol und halten
sich nicht am einem Arbeitsplatz auf, so dass einige der
zahlreichen Fremdfaktoren ausscheiden. Auch spielen in diesem
Alter die Pubertatsakzelerationsfaktoren noch keine Rolle.
With reference to the contributing factors, a lot has
been said about asthma at this Symposium. There is yet
another possibility and this is to observe the use of drugs.
The centralized health service network in our country, enabled
Eruby, from Brno, to compare the use of asthma drugs in pol~
luted and relatively unpolluted areas. The investigation
showed that they were used more in exposed areas.
Turning to the question of the best criteria to use for
surveys of children or of the general population, and to the
significance of the variations in other biological parameters
uhich have emerged from lung tests, epidemiological question-
naires and specific child surveys, we find that the majority
of, or all our findings, are physiological in nature; we are
not dealing with pathological changes. However, the changes
found indicate that the adult's or child's organism has to^
fight against long-term stress and this is important. Which
age group do we consider most suitable for the survey? This
is not so important for surveys based on case histories or
questionnaires but when the survey has specific objectives
and is carried out clinically and in the laboratory, we
consider the age group from 7 to 12 years to be the best for the
following reasons:
-------
356
1. children have already been obliged to live in a given
environment for a certain period of time, i.e. in a polluted
or unpolluted area;
2. children are cooperative.
3. children do not smoke or drink and do not spend their time
at a place of work, which means that we can eliminate some of
the many extraneous factors; children of this age are not
yet affected by factors arising at puberty.
BRILLE Mme (France)
Pour repondre a cette question, 11 faut Her deux des points
que vous avez indiques, le choix des differentes populations
d'une part et d'autre part s'il faut etudier les effets de la
pollution a long terme, ou s'il faut au contraire etudier les
effets courts, aigus, subaigus, car le choix de la population
depend du type d1etude que 1'on veut faire. D'apres ce qui est
deja connu par des etudes epidemiologiques et a la suite de
constatations cliniques, les effets aigus sont particulierement
bien etudies sur certaines populations susceptibles faites de
vieillards, de malades respiratoires ou cardio-respiratoires
deja connus et d'enfants tres jeunes, de nourrissons. La, je
crois qu'il n'y a pas de doute que pour les effets aigus, il
faut etudier des populations susceptibles. Ce choix repose
aussi sur le resultat d'etudes experimentales faites chez
I1 nomine, comme par example, celles qui ont ete faites a I1Unite
de Recherche "Pollution Atmospherique" de Londres, par Waller,
qui en parlerait mieux que moi. Ces etudes ont montre que si
1'on expose des sujets normaux a des concentrations de polluants
atmospheriques bien superieures a celles que I1on trouve dans
1'atmosphere, en particulier pour le SO-, on n1arrive pas a
provoquer de symptomes ni meme de modifications plus fines,
comme celles de 1'augmentation des resistances de voies aerien-
nes, sauf dans quelques cas et en utilisant des taux enormes
qui n'ont strictement rien a voir avec les conditions recontrees
-------
357
dans 1'air ambiant. Dans ce type d1experience on n'obtient de
reactions que si I1on etudie des sujets asthmatiques ou ayant
une reponse positive a un test a 1'acethylcholine, c'est a
dire une hyperexcitabilite broncho-motrice: on peut alors
obtenir des augmentations de resistance en les exposant a des
taux voisins de la pollution atmospherique generale.
En ce qui concerne I1etude des effets chroniques a long
terme, le choix de la population n'est pas aussi clair et il
est actuellement 1'objet de discussions. On peut dire qu'il y
a actuellement deux attitudes: I1une, de n'etudier que des
enfants et je pense que ce choix sera defendu par Irwig, et
1'autre qui consiste a ne pas se limiter aux enfants, mais a
etudier des adultes. C'est I1attitude que nous avons choisie
apre"s une discussion de tous les responsables de notre etude
en France, appuyee sur les donnees de la litterature. Je
rappelle les avantages d'une etude d'enfants d'age scolaire:
ils ne fument pas, ils restent toute la journee dans la zone
oti ils habitent (les mesures de pollution atmospherique de
cette zone sont done representatives de ce a quoi ils sont
soumis) et ils n'ont pas d1exposition professionnelle; en outre
a cet age un test peut etre fait pour mesurer la fonction
pulmonaire. Je pense cependant qu'il y a deux inconvenients
a n'etudier que des enfants: 1'un est que nous ne pouvons pas
affirmer que la susceptibilite des voies aeriennes d'un enfant
sont les temoins exacts de la susceptibilite des voies aeriennes
d'un adulte. L'autre inconvenient decoule de 1'effet additif
possible, peut-Stre me"me plus qu'additif, de plusieurs irritants;
a ce point de vue il est possible que 1'on mette plus facilement
en evidence 1'effet de la pollution atmospherique chez les
furneurs que chez les sujets non fumeurs; ceci apparalt dans
des etudes comme celle de Lambert et Reid en Grande Bretagne.
Dans le meme ordre d'idees, on peut penser que les ouvriers
soumis a une exposition professionnelle ont des voies aeriennes
plus susceptibles de reagir a la pollution atmospherique.
One etude des effets de la pollution atmospherique dans la
-------
358
population ouvriere pourrait done etre fructueuse, mais pour
etre valable, elle devrait absolument comprendre une etude
detaillee des postes de travail, avec entre autres, des mesures
de I1atmosphere du lieu de travail. Ainsi que je 1'ai dit
precedemment, I1exposition professionnelle est tres diverse et
est beaucoup plus difficile a prendre en corapte dans 1'analyse
que la consommation de tabac. C'est pourquoi il nous a paru
raisonnable, tout en le regrettant, de limiter notre etude aux
autres categories socio-professionnelles.
L'age des adultes a etudier merite aussi discussion et
1'on doit ajouter les raisons qui, dans notre etude, ont conduit
a retenir les sujets ages de 25 a 59 ans: a cote de probiernes
de croissance, les jeunes adultes sont rarement stables quant
a leur domicile (et a ce point de vue les sujets de plus de
6O ans, du fait de la cessation d'activite leur ressemblent) il
a done paru deraisonnable de les etudier dans une etude d'effets
a long terme ou 1'on souhaite une residence d'au mo ins trois
ans dans la zone etudiee. En ce qui concerne les sujets Sges
de 6O ans et plus, la pathologic respiratoire est chez eux
trop frequente par 1*accumulation de nombreux facteurs etiolo-
giques et trop souvent intriquee a d1autres maladies pouvant
retentir sur la fonction pulmonaire, pour pouvoir etre consideree
comme un "reacteur" a la pollution atmospherique.
Pour I1etude cooperative francaise, il a done ete decide
d'etudier a la fois les enfants de 6 a 1O ans, et les adultes
de 25 a 59 ans pour les raisons que je viens de dormer; mais je
serais tres interessee de savoir s'il y a d1autres arguments en
plus de ceux qui ont deja ete developpes, en faveur du choix de
belle ou telle population pour 1*etude des effets a long terme.
In order to answer that question, the two points to which
you have referred should be taken together; on the one hand the
populations to be selected and on the other the question whether
the long-term effects or the short-term, acute and sub-acute
effects of pollution should be studied, for the choice of
-------
359
population depends on the type of survey one has in mind.
According to the information already obtained from epidemiological
studies and clinical observation, the acute effects have been
subjected to thorough examination on certain sensitive populations
consisting of old persons, persons in whom respiratory or cardio-
respiratory disease has already been diagnosed and very young
children, infante. I think there can be no doubt that, in
dealing with acute effects, sensitive populations should be
studied. This choice is also based on the result of experimental
studies made in man such as, for example, those carried out in
the "Atmposheric Pollution" research unit in London by Mr. Waller,
who is better able than I to discuss them. These studies show
that, if normal subjects are exposed to levels of atmospheric
pollution greatly in excess of those found in the atmosphere, in
particular of SO„, it is not possible to produce symptoms, nor
even minor changes such as increased resistance of the air pas-
sages, except in a few cases and by using very high concentrations
which have nothing in common with the conditions in the ambient
atmosphere. In this type of experiment, reactions are obtained
only if the subjects studied are asthmatic or have a positive
response to an acetylcholine test, i.e., if they suffer from
broncho-motor hyperexcitability: it is then possible to obtain
increased resistance by exposing them to concentrations close
to those found in general atmospheric pollution.
With reference to the study of chronic long-term effects,
the choice of population is not so obvious and it is at present
the subject of discussion. It is fair to say that, at present,
there are two points of view: one, that only children should be
included in the survey - a choice which I think Dr. Irwig would
support - and the other, which consists in not limiting the
survey to children but in including adults as well. This is
the policy which we have adopted after a discussion, based on
the information in the literature, with all those in charge
of our survey in France. I would again point out the advantages
of a survey of children of school age: they do not smoke, they
spend the entire day in the zone where they live (the atmospheric
pollution readings for that zone are therefore representative of
the pollution to which they are subjected) and in their case
there is no occupational exposure; besides, at that age a test
can be carried out to measure pulmonary function. I think
however that there are two disadvantages in restricting the
survey to children: the first is that we cannot be sure that
the sensitivity of a child's air passages coincide exactly
with that found in an adult. The other disadvantage stems from
the possible additive effect, perhaps even more than additive,
of several irritants; from this point of view it is possible
that the effects of atmospheric pollution can be demonstrated
more easily in smokers than in non-smokers; this is clear in
studies such as those by Lambert and Reid in the United Kingdom.
-------
360
By the same token, one might think that the air passages
of workers subjected to occupational exposure are more sensitive
to atmospheric pollution* A study of the effects of atmospheric
pollution in the working population might therefore be useful^
but in order to be worthwhile, it would have to include a
detailed study of workplaces with, inter alia, measurements of
the local atmosphere. As I have already said, occupational
exposure is very varied and much more difficult to take account
of in the analysis than tobacco consumption. For that reason^
we reluctantly concluded that it was better to limit our
survey to the other eocio-professional categories.
The age of the adults to be included in the survey should
also be considered and we must therefore give the reasons why
we selected subjects aged between 25 and 59: apart from problems
of growth, young adults seldom live for long in the same place
(and in this respect, persons over 60 years of age, when they
retire, resemble them), so it seemed inadvisable to include
them in a survey of long-term effects for which residence of
at least three years in the zone surveyed is desirable. In
persons over 60, the frequent instances of respiratory disorders
are too often due to the accumulation of numerous etiological
factors, and too often linked with other diseases liable to
affect pulmonary function, for them to be seen as a "reaction"
to atmospheric pollution.
For the combined French survey, it was therefore decided
to use children from 6 to 10 years old and also adults from
25 to 59, for the reasons I have just explained: but I should
be very interested to know if there are other arguments, in
addition to those which have been put forward, in favour of
selecting any given population for a survey of long-term
effects.
IRWIG (U.K.)
Basically I do not disagree with Dr. Brille. As in all
epidemiological studies, the population that one should be
looking at, should be chosen in terms of the question that is
being asked. If one wishes to use populations as an indicator
of the biological effects of pollution, for example if one
wanted to compare the relative effects of smoke and S02, it would
seem reasonable to choose that population which was sensitive
to the pollutants under study, which was easy to obtain, and in
which there were minimal other interfering factors. It seems
t
that there is a fair amount of consensus that the childhood
-------
361
population forms such a group. if on the other hand, one wishes
to take on the enormous task of estimating the potential benefits
of improvement of environmental pollution levels, then one has
to study all the population groups and all changes in health
status within those populations. So if one wanted to compare,
the cost of improving air pollution levels with the benefit
obtained, then it would be necessary to have a look at all
population subgroups within that community.
KNELSON (U.S.A.)
I believe the two questions our colleagues have been
addressing in the last few minutes are inextricably intertwined.
The choice of health parameters as a means of exploring the
relationships between environmental stress and the health status
Of population, as well as choosing the subsets within that
population cannot be discussed separately. There are some
interrelationships between the two. I agree and certainly I
think it has been well demonstrated by virtue of our experience
with the rather clear cut results of major catastrophic episodes
of the last couple of decades that the very young, the very
old and infirm, as we always say in our manuscripts, are
those that are most susceptible; but, in what way?
Of course, in these catastrophic episodes, we were observing
largely excess mortality, as well as excess overt morbidity.
As we attempt to delineate in somewhat more detail and more
refined fashion, the relationships that we are exploring here,
I think we need to look carefully at other tools and to what
subsets of the population these tools may be amenable. We have
concentrated, as has been the experience of all the people
on the panel, primarily on pulmonary disease. Most of us,
historically, have been associated with air pollution health
effects studies and obviously, the lung and the respiratory
tract in general is the organ of insult. The lung constitutes
the greatest interface between man and his environment so it
-------
362
is always a logical organ to explore. But I think we must
not be narrow-minded in continuing to slice thinner and thinner
and explore in more and more detail the relationships between
air pollution and lung function at the expense of other health
effects and at .the expense of susceptible population sub-groups
other that the young and the old and the infirm. We have in
our laboratories in the U.S. Environmental Protection Agency
begun to use in population studies the tools that are developed
in clinical laboratories. As we have begun to design our
population studies with this in mind, it has given us an oppor-
tunity to enlarge the scope of our investigation where we are
no longer limiting ourselves to the classic cardio-pulmonary
kind of examinations. I agree wholeheartedly with Dr. Carnow
that the continued use of questionnaires is a very valid tool
and I am reassured by his experience that the large investment
we are making in this kind of health indicator is likely to
continue to pay off as it has in the past. However, there are
specific areas that deserve more attention, and we are using our
experience in the clinical laboratory now to design experiments
with populations to explore, for example, cytogenetic effects -
we only recently have the tools to do that - and to explore the
immune status of populations. We know from questionnaires that
one of the most reliable indicators of health effects of environ-
mental stress is an excess of respiratory tract infection of a
variety of kinds. However, what are the effects on the more
subtle parameters of immune status in the populations that do
not overtly express the environmental insult with lung disease
or respiratory disease and how is their immune status altered so
that other infectious diseases are allowed to manifest them-
selves? We have recently begun to learn about the whole field
of carcinoma-associated antigens, and we already have some
preliminary evidence that environmental insult is somehow
involved in the whole story of there carcinoma-associated
antigens. These are offered simply as examples of how I believe
the investigators who are designing tthe clinical as well as the
-------
363
epidemiologic studies should broaden their scope both with
respect to the tools that they use and the populations in which
they use them.
KUMPF (W.H.O.)
I would only like to mention for international studies,
perhaps some practical points, that is the accessibility of the
study population, the feasibility of the study including the cost
and if it is decided to study respiratory affliction, it may be
that school children are not the most susceptible group, but
certainly they are accessible and the cost of the study can be
kept rather low using school health examination. This has
been a major factor in determining the age group in our study.
In addition, school children of the age groups selected,
i.e. from 8 to 1O years, have as a study group all the advantages
mentioned already by Dr. Schmidt, i.e. they are not yet affected
by smoking and puberty, are fairly uniformly exposed to air
pollution, especially if the school is located in the neighbour-
hood where the children live, as would be expected with
elementary schools for that age group; they are expected to
collaborate well with the pulmonary function tests and can be
followed up later.
HUETER (U.S.A.)
How much stress should be placed on the air pollution
measurements?
BOURDEAU (C.E.C.)
I think this is a point which has been brought out in-
directly, at least by the report we heard on the WHO study.
-------
364
The measurement of independent variables is of course of
capital importance in drawing conclusions from epidemiological
surveys and we ought to try to avoid the pitfalls that have
affected previous studies by trying to make those air pollution
measurements as comparable as possible. Special emphasis has
been placed on this point in organizing the European Community
survey. The SO- measurements may be considered to be fairly
comparable. The problem of particulates is more complex since
in some countries they are measured as smoke, in others by
gravimetric methods. Moreover, it has been shown by the very
interesting studies from the CHESS programme, that quite possibly
sulphates will show closer relationship with respiratory
effects then any other parameters of air pollution. Considerable
thought will be given to that in the implementation of the
European Communities' study.
HUETER (U.S.A.)
Closing the discussion between panel members, the panel is
now opened to questions from the floor.
ZIELHUIS (Netherlands)
1. International studies conducted by EEC and by WHO - Europe
have been discussed. It is not a waste of manpower and money
if within Europe both studies are conducted at the same time?
2. Drs. Bourdeau, Brille, Kumpf and Knelson spoke about annual
average levels for S02 etc. However, 24 hrs. levels usually are
distributed according to log-probability scale. Should exposure
data be studied according to the specific annual distribution?
3. International comparison assumes similar dose-response
relationships. However differences between countries exist in
-------
365
- genetics, e.g. alpha-antrypsine deficiency
- socio cultural factors, determining the subjective
interpretation of e.g. irritation
- nutrition status (quantitative, qualitative,
e.g. milk, protein intake) may determine sus-
ceptibility.
So, international comparability is only possible if all the
covariables are taken into account. Is this done in EEC and WHO
studies?
4. Dr. Carnow proposed a predictive model. He stressed the
effect of socio-economic status on susceptibility. This pre-
diction of excess bronchitis/asthma per increase of energy con-
sumption assumes no change in socio-economic status. This
prediction strongly supports a change in socio-economic
politics.
KUMPF (W.H.O.)
Mr. Chairman, I could perhaps try to reply to two of the
questions:
There is a difference in the area covered by the 2 studies
as the European region of WHO extends much beyond the nine member
states of the European Community. However, the method of
approach is similar with the main difference that WHO uses a
mailed questionnaire whereas CEC uses one administered by
interview.
CEC has participated in all or most of the meetings of WHO
on this suudy, and of course CEC, in a way is in a position to
profit from the results of the WHO study. The WHO study will
have to be tightened up as far as the air quality measurements
are concerned, hopefully with the collaboration of CEC.
-------
366
Differences in nutritional status between the various
national groups are not being considered in the study, however,
differences in socio-economic conditions are summarized in the
information of the living conditions of the family of the child
and the educational level of the parents. These seemed to be
the only parameters which could easily be determined in the
various countries under the different socio-economic conditions,
BOURDEAU (C.E.C.)
I can say very little in addition to what Dr. Kumpf has
said. The two studies are coordinated and should yield more
information than the sum of two isolated studies; concerning
the non-environmental link or non-air pollution link differences
between populations, an attempt was made to get at that by
having in each country zones of low pollution and zones of
fairly high pollution either in SO. or particulates, wherever
available.
Regarding the use of annual concentrations whereas we
have a log-probability scale, his, of course is a point but
we have not only the average annual concentrations. These are
used actually to select the areas, but afterwards, winter
concentrations, monthly concentrations, and even daily average
concentrations will be available which might be used in the
analysis of the data, if needed.
CABNOW (U.S.A.)
Concerning Dr. Zielhuis* comment that the prediction
strongly supports a change in socio-economic politics. This
is true, and is, of course, one of the values of prediction.
No matter how crude it is, it gives us some opportunity to vary
alternatives so that in looking at energy we can consider low,
moderate, or high scenarios and, as I said, try to structure
-------
367
a more rational energy program based on that. The way we are
doing it now is to react to emergency situations and the
solutions are frequently worse than the problems that were origi-
nally present. In addition, these models are completely flexible
and can be modified as time goes on. As variables change, new
data can be added to the model (i.e., new epidemiclogical data
on guantitation of biological responses or the finding of new
risk populations).
As we become more sophisticated in our environmental
modeling (i.e., better diffusion models and monitoring equip-
ment) one can add these also. By the way, for SO- and the
pollutants we measured, 15-minute values were used, from which
we derived 24-hour averages at each station. They were not
annual averages nor were they citywide. I think citywide or
annual averages are generally inadequate as estimators of
population exposure.
We not only used individual 15-minute values at each
station, but we developed a linear model using all the stations
and an infinite number of data points to estimate square mile
values for each 15-minute of each day for 36 months in order
to estimate exposures of the 56O bronchitics monitored for
illness in one of our studies.
KARHAUSEN (C.E.C.)
If lower respiratory disorders like emphysema, and asthma
have a strong genetic component and if the geographic distri-
bution of the genetically susceptible individual is not homo-
geneous (as indicated by the results of the WHO study in
Roumania), the results of those studies might not be about the
sole relationship between environmental factors and respiratory
illness but actually between genetic factors, environmental
factors and respiratory disorders.
-------
368
CARNOW (U.S.A.)
You are talking about national group characteristics or
genetic or ethnic factors which may be important in disease.
I think there is something that is very important in what you
are saying. I do not know if you meant this also, but there
is frequently self-selection of populations and I think that
considerable caution should be exercised because of this.
In the United States greater numbers of people die from
emphysema in Arizona and Idaho, states that generally have
little pollution, than in other states. I think that those
families that can afford to relocate may leave an area of high
pollution because they have children who have frequent respiratory
illnesses. Those remaining in the highly polluted areas may be
non-pollution sensitive (i.e. no asthma, allergies, alpha anti-
trypsin deficiency, etc.) so you may be dealing with a survival
population. Just briefly, we looked at a group of octagenerians
in Chicago and found that they were very resistant to pollution.
They had lived in Chicago for 65 years, were generally non-
smokers, had no history of respiratory disease, and had survived
everything Chicago had to offer. They were a survival population
who are very resistant and draw general conclusions regarding the
effects of pollutants on people in general.
-------
TOXIKOLOGISCHE UNTERSUCHUNGEN
TOXICOLOGICAL STUDIES
ETUDES TOXICOLOGIQUES
STUDI TOSSICOLOGICI
TOXICOLOGISCHE ONDERZOEKINGEN
Vorsitzender - Chairman - President - Preeidente - Voorzitter
H.W. SCHLIPKDTTER
(Bundesrepublik Deutschland)
Stellvertretender Vorsitzender - Vice-chairman - Vice-president
Vicepresidente - Vice-Voorzitter
N.K. WEAVER (U.S.A.)
-------
371
TOXICOLORIC ASSESSMENT OF THE HEALTH EFFECT OF SULPHUR
DIOXIDE AND SULFATE PARTICULATES
J. WESLEY CLAYTON JR.
Health Effects Division
Environmental Protection Agency, Washington DC, U.S.A.
ABSTRACT (reduced by the Editorial Board)
The health effects of particulates in the inhalation of
sulfur dioxide and other gases has not yet been satisfactorily
defined. It is known that several factors contribute to the
"biological response elicited when particles and gases are inhaled.
Among these are the size of the particle, its shape, the con-
centration of particles in the gas medium, and the physio-chemical
properties of the particulate, especially its reactivity with
components of the gaseous phase and its solvent action related
to the gaseous phase.
While sulfur dioxide in sufficient concentration is a re-
spiratory irritant, it rarely, if ever, occurs at irritation
concentrations in the urban environment even in "episodes." It
has become, therefore, a so-called "index of pollution" and not
a cause per se of health consequences.
On these grounds it then becomes important to examine the
health effects of the particulate matter dispersed in the urban
air, notably, particulate sulfate aerosols which result from
catalytic oxidation of the sulfur dioxide effluent from stationary
power sources, or from catalytic converters employed to reduce
carbon monoxide and hydrocarbons emissions from mobile sources
of pollutants.
The study described in this paper on the particulate, ob-
-------
372
the Eastern United States disclosed no injurious effects when
particulate o/<5 microns in diameter was inhaled by monkeys
for 25 months* 22 hours a day* at concentrations of 100 or 500
ug/m .
In spite of the lack of apparent injurious effects from
sulfur dioxide in the long-term studies cited above, it has been
shown by Amdur and Underhill (Arch. Environ. Health 16:460 (1968
3
that sodium chloride aerosol (10 mg/m ) exacerbates the irritancy
properties of sulfur dioxide (2 ppm) as determined by increased
pulmonary flow resistance following single, one-hour exposures
of guinea pigs. The authors infer from these data that the
resistance to airflow through the respiratory tract is a function
of the solubility -of sulfur dioxide in the particulate aerosol.
Further information has been developed which suggests that
the irritant properties of sulfur dioxide might not only be en-
hanced by its solubility in a particulate phase but also by its
subsequent oxidation to sulfuric acid. Preliminary studies by
Amdur and Underhill (vide infra) have shown that insoluble* solid
aerosols, such as carbon, manganese dioxide, fly ash, triphenyl-
phosphate, and iron oxide, were inactive in potentiating the
irritant effect of sulfur dioxide in guinea pigs exposed for one
hour. However, soluble, solid aerosols, such as manganese chlo-
ride, sodium orthovanadate, and ferrous sulfate, all significantly
potentiated the irritant property of sulfur dioxide.
It is evident from these toxicologic data that not all
airborne contaminants potentiate the irritant action of sulfur
dioxide in single exposures, but that potentiation of toxic
effects is dependent on the ambient catalysis of sulfur dioxide
and S04 - and the latter on the oxidizing capacity of the catalytic
aerosol. Many sulfates in aerosol form possess far greater tox-
icologic activities than sulfur dioxide. The health consequences
of the conversion of sulfur dioxide to sulfate either in the
atmosphere or as a result of the oxidation of sulfur in fuels £>y
catalytic converters demand increased research before an adequate
margin of safety care be defined for sulfate emissions.
-------
373
In the assessment of the effects on human health resulting
from exposure to environmental contaminants, the data base
derived from human experience is seldom adequate to make sound
judgments as to the deleterious or beneficial effects on humans
of any substance or group of substances. Therefore, the stan-
dard setting process employed by a regulatory agency to control
exposures to pollutants is hampered by the limitations of the
health intelligence requisite to the establishing of a scien-
tifically defensible standard.
On the other hand, even wnen substantial human data are
available on a particular pollutant, there is seldom a consensus
on the meaning or significance of this information especially
with respect to an environmental standard.
While it may be held that the best way to study man is to
study men themselves,significant limitations inhere in the use
of people for the assessment of health effects.
Animal studies, too, are limited. The central issue herein
is whether the pollutant, in effective quantities, gains access
to the site of action. Species or strain variations, differences
from man in the absorption, distribution, metabolism, and
excretion inveigh against the easy acceptance that any animal
provides valid clues for examination of effects of the same
agent on human. Therefore, biochemical and toxicologic simil-
arities between animals and man should be elucidated before the
animal model can be satisfactorily employed to quantify the human
response to an environmental pollutant. If the ideal were
possible, animal studies to define relative toxicity should be
followed by low-dose metabolic studies in healthy humans, and
then a representative animal model should be sought so that
definitive biochemical, pharmacologic, and chronic toxicologic
studies can be performed.
Considering the toxicology of sulfur dioxide and sulfate
compounds or the so-called sulfates, animal experiments yield
data applicable to man because of the similarity of response
to the single agents. However, the health effects of particulates
-------
374
combined with sulfur dioxide or sulfates has not yet been
satisfactorily defined.
It has been shown by Amdur and Underbill (Arch. Environ.
Health 16-46O 1968) that sodium chloride aerosol (10 mg/m3)
exacerbates the irritancy properties of 2ppm sulfur dioxide
as determined by increased pulmonary flow resistance in guinea
pigs following single, one-hour exposure. Furthermore, aerosols
of other salts, for example, potassium chloride and ammonium
thiocyanate, in which sulfur dioxide is more readily soluble,
produced a greater enhancement of the irritant effect than
equivalent air concentrations of sodium chloride.
Further information has been developed which suggests that
the irritant properties of sulfur dioxide might not only be
enhanced by its solubility in a particulate phase but also by
its subsequent oxidation to sulfuric acid. Preliminary studies
by Amdur and Underbill (vide infra) have shown that insoluble,
solid aerosols, such as manganese chloride, sodium orthovanadata,
and ferrous sulfate, all significantly potentiated the irritant
capacity of sulfur dioxide.
Long-term inhalation studies employing guinea pigs and
monkeys exposed to sulfur dioxide, particulate matter, and sul-
furic acid aerosols elucidate their biologic action. Exposure
times were 22-23 hours a day for one year in guinea pigs and 1.5
years in monkeys. Guinea pigs inhaling 25O, 2,500, or 13/OOO
ug/m3 exhibited no deleterious effects as judged by survival,
body weight, hematology, clinical biochemistry, respiratory
function, or pathology.
As illustrated in Table I, an exposure level of 13,OOO
ug/m3 was associated with increased lung diffusion, reduced
lung disease, increased survival, and slight liver change.
Another group of guinea pigs were exposed to a mixture of
13,OO ug/m3 S0_ plus 5OO ug/m3 of respirable particulate. No
deleterious effects were observed. This finding pertains to the
functional, as well as the morphologic parameters recorded.
In this group of guinea pigs, the liver cells disclosed no
-------
375
vacuolation as was observed in the animals not inhaling partic-
ulate material. We do not understand the significance of this
difference. Vacuolation in rodent liver cells may signify
reversible adaptive changes of little health consequence. That
this minimal response was not recorded in the animals breathing
SO2 plus particulate can be taken as evidence that the combin-
ation exposure imposed no undue stress on the animals. A pair
of experiments was conducted with guinea pigs exposed to sulfuric
acid mist alone or combined with particulate.
TABLE I — Effects on Guinea Pigs of Chronic Exposure to
S02 and Particulate
Nominal Concentrations Results
S02 13,000 ug/m3 Increased Lung Diffusion
Reduced Lung Disease
Increased Survival Rate
Liver Cell Vacuolatlcr,
S02 13,000 ug/m3 Normal Lung Diffusion
- Normal Lung Disease
with 500 ug/nr> Particulate Normal Survival Rate
No Liver Cell Vacuolation
Tables II and III summarize the results. Significant harm
to these animals occurred when the guinea pigs receiving 610
ug/m3 (O.61O mg/m3) were accidentally exposed to 2,300 mg/m3
for about 2O minutes. As consequences of this event, one-third
of the animals died, and respiratory dysfunction and lung injury
occurred. Recovery in lung function was noted in three weeks
after the over-exposure when the guinea pigs were exposed to
filtered air only, not H-SO^ aerosol. Because the high level
exposure in this experiment was aborted by the accidental over-
exposure, an additional group was exposed to 9OO ug/m3. Only
a reduction in growth rate was noted in this experiment, and
this was not attributable to H_SO4 exposure because of the
-------
376
absense of any respiratory or pulmonary changes usually associated
with inhalation of an irritating aerosol.
TABLE II - Exposure of Guinea Pigs to H2SO4 Aerosol
(1)
Concentration
ug/m3
Particle 31 20
microns
Results
-0-
80
100
610<2>
No deleterious effects
< 1 Reduced growth rate only
> 1 Reduced growth rate only
<1 30* mortality
900
<1
Reduced growth rate
Increased respiratory
resistance
Lung Injury (acute)
Recovered in three weeks
Reduced growth rate only
(1) Exposed 12 months, 22 hours dally.
(2) Accidental overexposure at week 23 to estimated concentration
of 2,300 ng/m3, 20 minutes.
In the combination experiment (Table III) in which guinea
pigs inhaled H2SO aerosol mixed with particulate at concentra-
tions of 8O ug/m3 H2SO4 with 450 or 460 ug/m3 particulate, no
deleterious effects were demonstrated functionally or anatomicallj
Not even was a reduction in growth rate observed which had been
noted (Table IV) in guinea pigs exposed to 8O or 1OO ug/m3 of
H2SO aerosol. Reduction in growth rate would seem not to be
critical in assessing effects of these inhaled toxicants whose
primary point of attack is on the respiratory system at least
without other evidence of systemic toxicity.
-------
377
TABLE III - Exposure of Guinea Pigs to Particulate and
H2S04 Aerosol (1)
Combined Concentration Particle Size
Agents ug/ra3 microns Results
H2S04 +
Particulate
H2S04 +
Particulate
80
460
80
450
-------
378
As shown in Table IV, Phase II, a continuous exposure at
13,OOO ug/m3 for 78 weeks, without the intervention of an over-
exposure, effects on health attributable to S02 were not observed.
Tables V and VI disclose no harmful effects on monkeys
exposed for 18 months from these concentrations of respirable
particulate at 5OO ug/m3 alone or in combination with SO2 at
levels of 250, 2,500 or 13,OOO ug/m3.
TABLE v — Exposure of Monkeys to Particulate
Concentration
160 ug/m3
460 ug/m3
Particle Size
Average
5u
5u
Exposure
Duration
Observations
(1)
18 months No harmful effects
18 months No harmful effects
(1) Included: Body weight, survival, clinical chemistry, behavior,
respiratory functions, pathology
TABLE VI " Ejects of Exposure of Monkeys to S02 and Particulate
Combinations (18 Months)
Nominal
S02
ug/m3
250
2,500
13,000
Concentrations
Particulate
ug/m3
500
500
500
Particle
Diameter
u
<5
Results
Increased growth
No deleterious effects
on lung function, blood
and biochemical parameters
or histology
-------
379
When monkeys were exposed continuously for 18 months, as
shown in Table VII to sulfuric acid aerosol at levels of 38O
ug/m3 to 4,79O ug/m3 employing two particle size ranges, two
deaths occurred from pulmonary hemorrhage, and significant
functional changes occurred.
TABLE VII — Effects on Monkeys of Inhaling H-SO. Aerosol
(18 months)
Concentration Particle
ug/m3 Diameter Effects*
2,430 l-5u 1, 2, 3, 4
380 l-5u 1, 4 (slight)
4,790 lu 1, 2, 3, 4
480 lu 2 (slight)
* 1. Increased Respiratory Rate
2. Impaired Lung Ventilation
3. Lowered Arterial Oxygen
4. Lung-Epithelial Hyperplasia, Thickening of Bronchiolar Wall
Evidence of functional changes included increased res-
piratory rate, deterioration in the distribution of inspired
air, and lowered arrerial oxygen content. The two lower con-
centrations were associated with lesser effects. Diffusional
capacity of the lung was not affected at any concentration.
Alteration of the lung structure included epithelial hyperplasia.
and thickening of the bronchiolar wall.
When combined with particulate at 50O ug/m3 (<5 u diam), a
sulfuric acid aerosol concentration of 99O ug/m3 produced
structural changes in the lungs of monkeys that were similar
to those observed when particulate was not present in the
atmosphere. No such alterations were observed at the lower
sulfuric acid aerosol concentration of 110 ug/m3 combined with
a particulate concentration of 500 ug/m3.
-------
380
It is evident that the particulate present in these chambers
exerted little if any influence on the response to inhaled
sulfur dioxide or sulfuric acid aerosol. The latter proved to
be the greater irritant to the respiratory system.
Exposures of monkeys 22 hours a day for 2 years to 1O,OOO
ug/m3 calcium sulfate produced no finctional or anatomic abnor-
mality. Observations or measurements included: body weight,
hematology, blood gases, respiratory functions, clinical bio-
chemistry, organ weight, and pathology. Even when 25,OOO ug/m3
of sulfur dioxide were mixed with 10,000 ug/m3 of calcium sulfate,
no detrimental changes were observed in the same parameters
cited above.
These toxicologic investigations bear on potential human
health effects in several ways.
1. Except at concentrations exceeding 13,OOO ug/m3, i.e.,
overtly irritating sulfur dioxide per se does not appear to be
an air pollutant of concern to a healthy population. Even when
mixed with respirable particulate collected from fossil fuel-
burning electric generating plants, sulfur dioxide was not
harmful to guinea pigs or monkeys. This particulate was relatively
insoluble, and little conversion to sulfate is suspected.
2. The oxidation products of sulfur dioxide are of concern,
even to a healthy population particularly sulfuric acid aerosol,
as judged by these experiments using guinea pigs and monkeys.
3. We know that differential susceptibility to air pollutants
in the human population inveighs against uniformity of response.
By using different animal species, and approximation to the
human situation is feasible and dose-response data can be
obtained - a feature not readily obtainable from epidemiology.
4. This toxicologic study has clearly revealed that all sulfate
compounds do not act alike as pulmonary irritants. For example,
we h-ve shown that calcium sulfate is several orders of magnitude
less a toxic irritant than sulfuric acid aerosol. This finding
aptly demonstrates that animal experiments can sort out the more
-------
381
hazardous substances from those which are less so. On the basis
of relative toxicity as derived from animal investigations,
priorities can be defined for controlling certain pollutants and
criteria can be developed for assessment of health effects.
Of course toxicological data are inferential in nature.
Animal experiments yield inferential data in that animal species
may differ markedly among themselves and also may be quite
different from man in reacting to foreign substances. Man may
be more or less susceptible than any one animal species, and
it is not possible to decide a priori which animal species is
closest to man. Further, the homogeneity of animal groups does
not at all copy the variegated human population at risk. And,
finally, the stresses of human society are usually not part of
a well controlled animal study. The stresses experienced by
man may make him more or less susceptible to attack from pol-
lutahts.
Even data derived from human experiments are inferential
because of the small numbers of experimental subjects relative
to the vast number of their brothers, the relative brevity of
the experimental exposure relative to the duration of most
hazards, and the variety of physiological conditions and psy-
chological stresses resident in modern human society that
cannot be adequately incorporated into a human experiment.
The decision to establish an environmental standard with
a finite probability of human exposure involves balancing the
toxicological risks of use against the benefits of use in
society. Many factors bear on this decision. Animal studies
provide a preliminary or initial basis for the judgment. Acute
and chronic toxicity studies are fundamental in establishing
safety guidelines for the hazards involved in the use of a
substance. Functional and mechanism studies should assume a
greater role in the decision. There is a need for experimental
designs to be closely related to use patterns to enhance the
validity of the inferences from toxicological data. Ultimately,
the decision leading to an environmental standard is a judgement.
-------
382
reflecting some personal or social system of values, and it is
based on experience in actual use and documented by continuing
epidemiologic studies of the populations at risk.
DISCUSSION
GOLDSMITH (U.S.A.)
Your data appear to show that there ^s impairment of growth
rate by levels of H2SO. as low as 80,ug/m in guinea pigs, with
particle size ^ 1/ug. How should such effects be interpreted
with respect to human health?
CLAYTON (U.S.A.)
Such effects should be interpreted with caution. Unless
body weight changes are accompanied bv other, related dvsfunc-
tions they should not be anticipated as significant to human
health effects.
LAFONTAINE (Belgium)
My question is perhaps more of a comment than an actual
question: experimental exposure to SO_, H_SO. aerosols, and
certain combinations of SO. and particulates should be accom-
panied by exposures to bacteria or viruses with varying degrees
of pathogenic!ty. This would permit closer comparison with
the effects in humans and would perhaps explain the scatter
found in the results. We must, of course, consider measurements
made in the atmosphere (SO- and particulates) as nothing more
than indices.
CLAYTON (U.S.A.)
I agree with Or. Lafontaines' comment. Studies on this
issue show that intercurrent infection with microorganisms is
exacerbated by inhaled irritants.
-------
383
NEWHOUSE (Canada)
I question the validity of the animal model and the axiom
that toxicologic data can be extrapolated in large part to man.
I base my comments on the alterations noted by Proctor in nasal
mucociliary transport in normal man exposed, to 5 or even 1 ppm
SO2 alone during short-term acute exposures, as well as data
from our laboratory (and others) showing significant changes in
the maximum midexprivatory flow rate (an index of probably small
airway obstruction) during similar exposures. While this may
not have important implications in chronic exposure conditions
it does, I believe, cast some doubt on the validity of the
animal model.
CLAYTON (U.S.A.)
As I made clear in my statements in this paper, extension
to man should be made with conservatism. Your data appear int-
eresting but a dose-response curve should be obtained at levels
<1 ppm in order to increase the relevance for ambient conditions
-------
385
BIOLOGICAL EFFECTS OF THE POLYCHLORINATED
BIPHENYLS IN NONHUMAN PRIMATES
J, R, ALLEN, L, A. CARSTENS AND D, H, NORBACK
Department of Pathology and Regional Primate Research Center,
University of Wisconsin, Madison, Wisconsin, USA
ABSTRACT
Daily ingestion of polychlorinated biphenyls (PCBs) at doses
between 25 ppm and 300 ppm over a period of 2 to 3 months caused
morbidity and mortality in nonhuman primates (Macaco, mulatto).
Changes included loss of body weight, alopecia, subcutaneous
edema, acne, conjunctivitis, ascites, hydrothorax, gastric mu-
cosal hyperplasia, liver hypertrophy and bone marrow hypoplasia.
Terminally, these animals had decreased hemoglobins, hematocrits
and serum protein, shifts in the serum albumin to globulin ratios
and a neutrophilia. A dose of 2.5 ppm PCS in the diet was suf-
ficient to cause alopecia, focal edema and conjunctivitis in
monkeys within 2 months. Animals that have survived the inges-
tion of 250 to ZOO mg of PCB over 2 months maintained levels of
PCB in their adipose tissue for periods in excess of 1 year.
In addition, offspring of these animals had detectable levels
of PCBs in their tissues.
In contrast to the injurious effects experiences by adult
nonhuman primates exposed to PCBs, one-month-old infant monkeys
that were exposed to Ig of PCB for 30 days were essentially free
of any of the clinical, gross and microscopic alterations of PCB
intoxication. These observations as well as metabolic studies
suggest that the metabolism of the PCBs is important in deter-
mining their toxicity.
-------
386
1. Introduction
Although the polychlorinated biphenyls (PCBs) have been used exten-
sively for industrial purposes for over 40 years, it has only been during
the last decade that they have been regarded as environmental contamin-
ants [Jensen [1]). These compounds are extremely stable, not hydro-
lyzed by water, acids or alkali, and are able to withstand temperatures
up to 650°F without disintegrating. They have been used in elastomers
adhesives, paints, varnishes, printing inks, putty, and as general
fillers. Since they do not conduct electricity, they have also found
widespread use in electrical equipment such as transformers. Their
stability and low vapor pressures make them well suited as lubricants,
hydraulic fluids, liquid seals, cutting oils, vacuum diffusion pump oils,
and as vapor suppressants and insecticide formulations. The same proper-
ties that make them ideal for commercial use also enhance their resis-
tance to degradation in the environment. The magnitude of PCB contamina-
tion is further exemplified by their presence in coho salmon, milk fat,
poultry, eggs, and fish (Kolbye [2]). Primarily as a result of food
contamination, detectable levels of PCBs were found to be present in over
30% of randomly sampled inhabitants of the United States (Yobs [3]).
Early outbreaks of PCB intoxication were limited to industrial acci-
dents where workers reported incidents of chloracne (Jones and Alden,
[4], Good and Pensky [5]). It was not until 1968 that the greater
health significance of PCB exposure became known. Following the consum-
ption of PCB contaminated rice oil by over 1,000 Japanese, diverse symp-
toms and lesions developed. These persons became nauseated and lethargic,
and developed chloracne and subcutaneous edema. Infants born to exposed
mothers were small and exhibited discolored skin and had an abnormal
eyeidischarge. Many of the symptoms and lesions that developed during
the acute phase of the illness have persisted for years (Kuratsune [6],
Kuratsune et^ al^ [7]).
2. Experimental Studies
2.1 High Level PCB Exposure. In order to further clarify the bio-
logical effects of various levels of PCB exposure in primates, a series
of experiments have been conducted. In these experiments, adult Maeaca
mulatta monkeys were given diets containing 25, 100. and 300 ppm of a
PCB (Aroclor 1248, Monsanto Co., Inc., St. Louis, Missouri, U.S.A.) for
periods of 2 to 3 months (Allen and Norback [8], Allen et_ al_. [9],
-------
387
[10]). Although each of these dosages produced signs and lesions com-
patible with PCB intoxication within one month, the total PCB intake
varied from 250 to 400 rag for the 25 ppm group, 0.8 to 1.0 g for the 100
ppm group, and 3.6 to 5.4 g for the 300 ppm group. Mortality occurred in
the groups receiving the two larger doses within 2 months and within 4
months on the lower regime. Hematological changes developed gradually
during the period of experimentation; however, they were most obvious in
the 300 ppm group. These animals showed a decrease in hemoglobin (grams
per 100 ml) from 13.0 ± 1.0 to 10.6 ± 0.3 within 12 weeks, and a corres-
ponding decrease in hematocrit from 39.0 ± 2.0% to 32.0 ± 1.0%. Although
the total white cell count was not appreciably altered, there was a
decrease in the number of lymphocytes and a concomitant increase in the
neutrophil population. Decreases in the level of serum protein were
accompanied by a reduction in the percentage of albumin and increases in
the globulin fraction. Reduced levels of serum lipids, cholesterol, and
triglycerides accompanied the alterations in serum protein (Allen et al.
[10]).
A gradual weight loss was recorded throughout the experimental
period for all PCB-fed animals. The 25% decrease in body weight of the
300 ppm group was particularly striking. Following one month of expo-
sure, there was moderate to marked loss of hair from the head, neck,
and back, and edema of the mouth and eyelids. Loss of eyelashes, exces-
sive lacrimation, and conjunctivial congestion were also apparent.
Small pustules involving hair follicles were particularly obvious around
the mouth and on the cheeks and neck.
Liver biopsies taken from the animals after a short period of expo-
sure to PCBs showed a decided increase in the smooth endoplasmic reticu-
lum of the hepatic cells (Figure 1). Biochemically, the liver homogen-
ates contained decreased levels of DMA (mg/mg liver) and UNA (mg/mg DNA)
and increased levels of protein (mg/mg DNA) and increased activity of
microsomal mixed function oxidases. In those animals that were sacri-
ficed,, the enlarged livers continued to show proliferation of the endo-
plasmic reticulun. However, instead of the endoplasmic reticulum being
distributed throughout the cytoplasm it was arranged in distinct packets
of closely associated membranes (Figure 2). Microsomes prepared from
those livers showed a decided decrease in activity of mixed function oxi-
dases (Allen e_t al_. [9]).
-------
388
RER
SER
FIGURE 1. Hepatocytes obtained from a liver biopsy of a monkey that had
received 300 ppm PCB in the diet for 2 weeks. Note the abundant smooth
endoplasmic reticulum (SER) and short segments of rough endoplasmic
reticulum (RER). Normal appearing mitochondria (M) are randomly dispersed
among the ER. Small lipid droplets (L) appear as round lucent areas in
the cytoplasm. X 8,420.
FIGURE 2. Large dense packets of smooth endoplasmic reticulum (SER) pre-
vail in the cytoplasm of hepatocytes obtained from monkey exposed to
300 ppm PCB in the diet for 2 months. Microsomes prepared from liver
cells having similar morphological features were hypoactive enzymatically.
Abundant lipid droplets prevail in the cytoplasm not occupied by ER.
The mitochondria (M) appear normal. X 6,920.
-------
389
The most severe pathological changes recorded at necropsy were loca-
ted in the stomach where marked hypertrophy of the mucosal lining and
edema of the stomach wall, particularly of the fundic and pyloric regions,
were apparent. Microscopically, the hypertrophic, gastric mucosa was
several times thicker than the control mucosa and was comprised of
greatly elongated hyperplastic glands containing mucous secreting cells
(Figures 3 and 4) . Prevention of discharges of the secretions owing to
the depth and apposition of the glands predisposed to the development of
large mucous cysts. There was also widespread penetration of the muscu-
laris mucosaewith invasion of the submucosa by the mucosal epithelium.
There were microscopic lesions of the skin, particularly of the face, of
PCB-exposed animals. Numerous intrafollicular cysts .surrounded by edema-
tous leukocyte-filled connective tissue were apparent.
At necropsy a portion of the tissues from the animals was obtained
for the .determination of PCB levels by the previously reported gas chroma -
tographic procedures (Allen et_ al_. , {9]}. Sites of major storage, in
addition to fat, were the adrenals, liver, and pancreas.
The persistence of the PCBs in the tissues of exposed animals was
demonstrated in the group receiving 25 ppm PCBs in the diet for 2 months
(total intake approximately 300 mg) . Immediately after the PCB-supple-
mented diet was discontinued, the subcutaneous adipose tissue contained
an average of 127 yg per g fat (Allen et^ al_. [11]). Even after one year,
concentrations of PCBs in the fatty tissue averaged 30 pg per g fat.
Additional distribution data were obtained from a PCB-fed female that
delivered an infant eight months after being taken off the PCB diet.
Levels of PCB (pg/g tissue) in the adipose tissue of this infant were
approximately one-half that present in the tissues of the mother, thus
indicating transplacental movement of these compounds (Allen et al. ,
2.2 Low Level PCB Exposure. Recently, experiments have been initi-
ated to determine if levels of PCBs permitted in foods destined for
human consumption would be toxic when consumed on a continuous basis.
Adult rhesus monkeys are being given diets containing 2.5 and 5.0 ppm PCBs
(Aroclor 1248). These levels are one-half and equal to the levels per-
mitted in certain foods. Interestingly enough, these animals developed
loss of hair and eyelashes within 2 to 3 months. They also have swollen
upper eyelids and congested conjunctiva.
-------
390
SM
FIGURE 3. The gastric mucosa of a normal adult rhesus monkey is depicted.
Note the narrow band of muscularis mucosae(M) and underlying submucosa
(SM) and smooth muscle (S) components of the stomach wall; hematoxylin
and eosin stain. X 25.
FIGURE 4. Stomach of a monkey that had received 300 ppm PCB in the diet
for 2 months. Note the hyperplastic gastric mucosa and presence of
dilated glands. The majority of the submucosa has been replaced by
glandular elements of the mucosal epithelium. Note the large cystic
spaces that have resulted from the stasis of mucous in these glands
(C); hematoxylin and eosin stain. X 25.
-------
391
2.3 jnfant Studies. In order to determine if adult and infant pri-
mates would respond similarly to PCB exposure, one-month-old rhesus mon-
keys were given 1 g (35 mg/kg body weight) of PCB over a 30 day period.
The total consumption of PCBs by these infants was essentially the same
as that received by the adults given 100 ppm in the diet as mentioned
previously. These animals were able to survive without any obvious ill
effects at doses that had been fatal to adults (Abrahamson and Allen,
[12]). When the tissues from the infant and adult monkeys were evaluated
chromatographically it was established that the concentration of PCB was
lower and fewer isomerswere present in the adult tissues. Those isomers
that were present in infants and absent in the adult tissues consisted
primarily of those having'lower chlorine levels (Table 1).
TABLE 1. CHROMATOGRAPHIC DATA ON ADIPOSE TISSUE OBTAINED FROM
RHESUS MONKEYS EXPOSED TO AROCLOR 1248a
Retention
Time
(Sec)
57
73
82
102
130
146
156
201
236
306
392
472
507
Standard
0.12
2.85
4.07
18.45
6.91
1.52
1.30
40.12
2.86
10.61
7.65
3.46
0.03
, Newborn ,
Mother Infantc Adult
7.70 1.56
2.80
10.49 19.33
7.58
0.30 11.34 63.52
50.96 4.14 15.57
20.14 76.79
7.17
Infant6
0.17
14.04
14.27
47.56
7.93
11.63
3.92
0.47
8 Monsanto Co, Inc., St. Louis, Missouri U.S.A.
Tissue taken from female monkey at time of cesarian section, 8 months
after discontinuation of PCB diet (total PCB intake - 300 g).
C Infant tissue taken immediately following cesarian delivery.
d Tissues of adult monkey taken after 1 month on PCB diet (total PCB
intake - 1 g).
6 Tissue from infant monkey taken after 1 month on PCB diet (total PCB
intake - 1 g).
-------
392
2.4 Metabolic Studies. Metabolic studies were conducted in adult
rhesus monkeys that had been given a single intragastric dose of a PCB
containing multiple isoners (Aroclor 1248) and a single isomer (2,5,2',
S'-tetrachlorobiphenyl). It was demonstrated that approximately 90% of
the compound was absorbed from the gastrointestinal tract. Ten percent
of the original dose was detected chromatographically in the urine and
feces within 14 days with the greatest percentage being excreted between
the second and eighth day. When similar studies were conducted with
tritiated 2,5,21,S'-tetrachlorobiphenyl, approximately 85% of the radio-
activity was recovered, primarily in the feces within 2 weeks, and major-
ity of it was eliminated by 72 hours. An additional 10% was detected
in the animal tissues. Of the PCB that remained in the animal, the
largest percentage was present in the adipose tissue and secondly in
organs having high fat content.
3. Discussion
When nonhuman primates were exposed to levels of PCBs similar to
that experienced by the Japanese in the 1968 PCB outbreak, they deve-
loped signs and lesions similar to those recorded in human cases. Sub-
cutaneous edema, particularly around the eyes and lips, and hyperemic
conjunctivitis with excessive secretion of the Meibomian glands were
consistently present. There were also follicular accentuations of the
skin characterized by dilatation of the sebaceous ducts and development
of keratin cysts. The severely affected patients and primates had a
decrease in erythrocytes, reduced hemoglobin, and a neutrophilia-
Although not documented in the exposed humans, the nausea and anorexia
they experienced was suggestive of a gastritis as was the case in the
experimental animals. Liver hypertrophy, proliferation of the hepatic
endoplasnic reticulum, and increased microsomal enzyme activity were
apparent in both species. In addition, the persistence of these lesions
for extended periods following the discontinuation of PCB exposure was
recorded in the Japanese outbreak and has also been observed in experi-
mentally exposed nonhuman primates.
The data and observations cited in this report clearly indicate that
PCBs are toxic to primates over a wide dose range. Particularly signi-
ficant is the appearance of lesions within less than 2 months in monkeys
that were fed diets containing 2.5 and 5.0 ppm. Although the intake of
PCBs by these animals was on a continuoui basis and the likelihood of the
-------
393
general population consuming a diet containing this quantity of PCBs is
small, it does point out that only small amounts of these compounds are
required to produce toxicity in primates. In addition, since PCBs accu-
mulate in tissues of exposed mammals, continuous exposure to even very
minute amounts may eventually reach a level sufficient to cause toxic
manifestations.
It is of interest that the toxicity of PCBs in infant and adult
monkeys is markedly different. Levels of exposure that were fatal to
adults produced no gross effects in one-month-old infants. It has
been shown that infant monkeys have a poorly developed mixed function
oxidase system (Allen and Chesney [13]) .These observations are further
substantiated by the chromatographic data obtained on the tissues of the
two age groups of animals. Adult tissues contained lower levels and
fewer isomers of PCBs than did the infant tissues, thus indicating more
rapid metabolism or excretion of these compounds by older animals.
These data suggest that PCB metabolites are responsible for the dele-
terious effects produced by PCBs in monkeys.
Unlike primates, infants of lower animals suffered much more severely
from PCB exposure than did adults (Mclaughlin e£ al^. [14], Keplinger et^
al^. [15], Dahlstrom [16], Hays and Risebrough [17], Ringer e£ al_. [18],
DeLong et_ al_. [19]). Clarification of these findings was recently
obtained in rats by the use of liver enzyme inhibitors (SKF 525-A --
p-dimethylaminoethyldiphenylpropylactate and chloramphenicol) and indu-
cers (phenobarbital). In these experiments, rats pretreated with enzyme
inhibitors, which may parallel the enzymatic status of the infant liver,
experienced heavy mortality while animals pretreated with enzyme inducers
suffered no deaths. Thus it would appear that the metabolism of PCBs may
be necessary for toxicity in monkeys while the reverse may be true in rats.
Many of the lesions that develop in PCB intoxicated primates may be
directly related to tissue exposure to the PCBs while other lesions may
be secondary. The loss of weight, decrease in serum protein and moderate
anemia could be related to the reduced food intake that occurs in the
animals with severe hyperplastic gastritis which develops as a result
of the irritating effects of the PCBs on the gastric mucosa. The hyper-
trophic hyperactive livers that develop in these animals are a result
of the stimulatory effects of these compounds on the hepatic endoplasmic
reticulum. However, when the liver exposure to the PCBs is sufficiently
-------
394
great, degenerative changes develop particularly in the organelle con-
taining the majority of PCBs, the endoplasmic reticulum. The skin lesions
may also be directly related to localization of high levels of PCBs in
their tissue. Keratinization of the sebaceous gland ducts and hair
follicles following the irritating exposure to PCBs leads to formation
of keratin cysts. In addition, edema and inflammation that develops
around the effected glands and hair follicles may be a direct result
of injurious effects on these tissues by the PCBs.
Many of the lesions that develop following PCB exposure may result
secondarily from the increased metabolism of endogenous substances by
the hyperactive hepatic endoplasmic reticulum. Following PCB exposure
there is enhanced metabolism of steroid and steroid-like compounds.
Vitamin A which is steroid-like compositionally is deficient in quail,
rats (Bitman et al. [20]) and monkeys (Allen, unpublished observations)
following PCB exposure. Since vitamin A deficiencies resulting from an
increased rate of metabolic decomposition have been reported to produce
gastric hyperplasia (Cramer [21]), gastric carcinoma (Fujimaki [22]), an
follicular dermatitis (Mason [23]), it is not unlikely that a deficiency
of this vitamin may be responsible at least in part for many of the
lesions resulting from PCB intoxication.
Acknowledgements
This research was supported by the University of Wisconsin
Sea Grant, the Food Research Institute of the University of Wisconsin,
and United States Public Health Service grants ES-00
-------
395
References
1. JENSEN, S., "Report of a new chemical hazard", New Sclent.. 32, 612
(1966). —
2. KOLBYE, A.C., "Food exposures to polychlorinated biphenyls" Environ
Health Persp.. 1, 85 (1972). '
3. YOBS, A.R., "Levels of polychlorinated biphenyls in adipose tissue
of the general population of the nation", Environ. Health Persp
1, 79 (1972). L*'
4. JONES, J.W., ALDEN, H.S., "An acneform dermatergosis", Arch. Dermatol
Syphilol., 33, 1022 (1936).
5. GOOD, C.K., PENSKY, N.} "Halowax acne (cable rash); cutaneous erup-
tions in marine electricians due to certain chlorinated naphthalenes
and diphenyls", Arch. Dermatol. Syphilol.. 48, 251 (1943).
6. KURATSUNE, M., "An epidemiologic study on "Yusho" or chlorobiphenyl
poisoning", Fukuoka Acta Medica, 60, 513 (1969).
7. KURATSUNE, M., YOSHIMURA, T., MATSUZAKA, J., YUMAGUCHI, A.,
"Epidemiologic study of Yusho, a poisoning caused by rice oil con-
taminated with a commercial brand of polychlorinated biphenyls"
Environ. Health Persp.. 1, 119 (1972).
8. ALLEN, J.R., NORBACK, D.H., "Polychlorinated biphenyl and triphenyl-
induced gastric mucosal hyperplasia in primates", Science, 179, 498
9. ALLEN, J.R., ABRAHAMSON, L.J., NORBACK, D.H., "Biological effects of
polychlorinated biphenyls and triphenyls on the subhuman primate.
Environ. Res.. 6, 344 (1973).
10. ALLEN, J.R., CARSTENS, L.A., BARSOTTI, D.A., "Pathological changes in
nonhuman primates exposed to low levels of polychlorinated biphenyls",
Toxicol. Appl. Pharmacol., accepted (1974) .
11. ALLEN, J.R., NORBACK, D.H., HSU, I.e., "Tissue modifications in mon-
keys as related to absorption, distribution and excretion of poly-
chlorinated biphenyls", Arch. Environ. Contam. Toxicol., in press
(1974). —
12. ABRAHAMSON, L.J., ALLEN, J.R., "The biological response of infant
nonhuman primates to a polychlorinated biphenyl", Environ. Health
Persp., 2, 81 (1973).
13. ALLEN, J.R., CHESNEY, C.F., "Effect of age on the development of
cor pulmonale in nonhuman primates following pyrrolizidine alkaloid
intoxication", Exp. Molec. Path., 17, 220 (1972).
-------
396
14. MCLAUGHLIN, J., MARLIAC, J.P., VERRETT, M.J., MUTCHLER, M.K.,
FITZHUGH, O.H., "The injection of chemicals into the yolk sac of
fertile eggs prior to incubations as a toxicity test", Toxicol. Appl.
Pharmacol.. 5, 760 (1963).
IS. KEPLINGER, M.L., FANCHER, O.E., CALANDRA, J.C., WHEELER, E.P., "Toxi-
cological studies with polychlorinated biphenyls", Proceedings of
NIEHS Polychlorinated Biphenyl Meeting, December 20-21 (1971).
16. DAHLSTROM, N., as quoted by TINKER, J., "The PCB study: seagulls
aren't funny anymore", New Scient.Science J., 16 (1971).
17. HAYS, H., RISEBROUGH, R.W., In Natural History, as reported in
"Controversy continues over PCBs", Chem. Eng. News 49, 32 (1971).
18. RINGER, R.K., AULERICH, R.J., ZABIK, M,,"Effect of dietary poly-
chlorinated biphenyls on growth and reproduction in mink", Proc.
Amer. Chem. Soc., 12, 149 (1972).
19. DELONG, R.L., GILMARTIN, W.G., SIMPSON, J.G., "Premature births in
California sea lions: association with high organochlorine pollutant
levels", Science. 181, 1168 (1973).
20. BITMAN, J., CECIL, H.S., HARRIS, S.J., "Biological effects of poly-
chlorinated biphenyls in rats and quail", Environ. Health Persp., 1,
145 (1972).
21. CRAMER, W., "Papillomatosis in the forestomach of the rat and its
bearing on the work of fibiger," Amer. J. Cancer, 31, 537 (1937).
22. FUJIMAKI, Y., "Formation of gastric carcinoma in albino rats fed
on deficient diets", J. Cancer Res., 10, 469 (1926).
23. MASON, K.E., "Effects of vitamin A deficiency in human beings",
In The Vitamins, W.H. Sebrel, Jr., R.H. Harris (eds.), Academic
Press, New York, 1, 137 (1954).
; DISCUSSION
WASSERMAN (Israel)
It would be of high epidemiological interest to learn about
the following biological effects observed in the Japanese pop-
ulation poisoned by PCB's.
Among the biological effects produced by PCB's, the decrease
of FBI and the hyperactivity in the hypophyso-thyroidadrenal axis
was observed by us. Tribute should be paid to Jensen for^d
the PCB's in man, a finding which opens a new view in man's
exposure to environmental hazards. •.
-------
397
ALLEN (U.S.A.)
The residual effects of polycholrinated biphenyl intox-
ication in man seem to be long lasting. According to Dr.
Kuratsune of the Department of Public Health, Kyushu Univer-
sity in Japan, many of those persons who were exposed to the
PCB contaminated rice oil continue to show ill effects after
3 years. We would suggest that Dr. Kuratsune be contacted
about the levels of PBI and hyperactivity in these patients as
related to the hypophyso-thyroid adrenal axis. Furthermore, it
goes without saying that all of us are indebted to Dr. Jensen
and his astute observations which exposed the PCB problem.
CLAUDE (France)
In your text you mention the reduced blood level of chol-
esterol and trlglycerides after intake of PCB. One could make
the hypothesis that there was a blockage of hepatic secretion
of the lipoproteins. Have you made chemical or histological
observations showing a certain frequency of hepatic steatoses
which would confirm the above hypothesis?
ALLEN (U.S.A.)
These animals do have a moderately fatty liver. However,
we have no data that would indicate one way or the other if this
fatty liver is related to the blockage of lipoprotein secretion
by the liver. In addition, we are not aware of any other data
along these lines from other laboratories.
CHAMBERS (Ireland)
I note that Dr. Allen did not make any reference to the
inununo-suppressive effect of the PCBs which has been described
in other animals, rabbits I believe, by workers in Utrecht. Is
this phenomenon not considered to occur in monkeys?
ALLEN (U.S.A.)
Experiments along these lines on monkeys have not been
conducted. It can be said, however, that these animals do
become leukopenic and are much more susceptible to opportune
pathogens. Whether this is related solely to their debilitated
condition or to an immunosuppressive effect of the PCBs is not
clear.
-------
UNTERSUCHUNG DER WIRKUNGEN AUF DEN MENSCHEN
EXPOSURE MONITORING
MESURE DE L'EXPOSITION
MISURA DELL'ESPOSIZIONE
METING VAN EXPOSIE
Panel
Vorsitzender - Chairman - President - Presidente - Voovzittev
M.J. SYROTA (France)
-------
401
RECAPITULATION OF ENVIRONMENTAL EXPOSURES
THOMAS D, ENGLISH
Environmental Protection Agency, National Environmental Research
Center, Research Triangle Park, NC, USA
ABSTRACT
In order to quantitate the effects of air pollutant exposure
on diseases such as chronic respiratory disease, annual estimates
of air pollutant concentrations for approximately twenty years
are required. This paper describes general techniques for the
recapitulation of environmental exposure under a wide variety
of topographic, climatic and source emission distribution condi-
tions. Annual wind rose and atmospheric stability conditions
are analytically coupled with source emission data to estimate
past spatial distributions of various pollutants.
Example recapitulations of sulfur dioxide concentrations
and total suspended particulate concentrations are presented for
both densely populated areas such as Metropolitan New York, and
sparsely populated areas such as Helena, Montana. The use of
dustfall data as an indicator of other pollutants such as sus-
pended sulfates is examined statistically. Methods are des-
cribed for calibrating computer based air pollution meteorologi-
cal models to local topographic and climatic conditions. Tech-
niques for determining the limits of precision of air pollutant
exposure estimates are described.
Measurements of air pollutant exposure for communities in
the vicinity of LOB Angeles constitute a data base upon which
estimates of air pollution concentrations for a 10,OCO square
mile area (the California South Coast Air Basin) are made.
-------
402
Annual average isoplethes for Oxidant, ff02 and total suspended
particulate are constructed for the time period from 19S5 to
1972. Annual average isoplethes of the daily maximum hourly
average concentrations which provide an estimate of the magni-
tude of peak air pollutant exposures are discussed and spatial
distribution of Oxidant is presented. Long-term trends for 4
communities are described.
-------
403
1. INTRODUCTION
A1r pollution monitoring is a young science. Consequently, the
methodologies used to measure pollutants change frequently. Further-
more, the number of sites at which the air is measured is rapidly
increasing. This state of flux makes it difficult to recapitulate
air pollutant historical exposures. In some locations which are very
Interesting from an epidemiological point of view, past exposure data
may be very sparse. In other cases, the available air monitoring data
may not be for a community near the city under investigation. In
large cities such as New York there are gaps in the historical air
quality records.
This paper presents techniques that have been used to estimate
the values of past air quality under a wide variety of conditions.
Various techniques are described for interpolation and extrapolation
of available air pollutant measurements. These techniques include use
of air quality models, construction of isoplethes from air monitoring
station data and estimation of the value of one pollutant from measured
values of another pollutant.
2. AIR QUALITY MODELS
Air quality models vary in sophistication from the simple "box
model"[l] to very complex models which numerically estimate the space
and time distribution of a pollutant through a set of coupled non-
linear differential equations^?] The class of air quality simulation
models^3] is intermediate in complexity. These models primarily use
wind-stability roses coupled with gaussian diffusion of pollutants
from sources of pollution to estimate ground level exposure. The
relative merits of models is currently the subject of considerable
debate.L4J The general precision, accuracy and applicability of a
variety of models to air quality criteria pollutants will be carefully
established in a Regional Air Pollution Study which is being conducted
by the Environmental Protection Agency.
-------
404
The goal of most all air quality models Is to relate emissions to
ground level exposure. In this paper various approaches are used to
estimate exposures for a wide variety of locales ranging from smelter
communities in the Rocky Mountains to large metropolitan cities such
as Los Angeles and New York.
3. RECAPITULATIONS BASED ON PRODUCTION
Western smelting communities usually have little pollution from
automobiles or from non-smelting industrial sources. Ambient air
pollution in these communities is characterized by levels of both
sulfur dioxide and total suspended particulates 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.
Estimates of long-term exposure to sulfur dioxide, total
suspended particulates, and suspended sulfates in the three smelter
communities were based upon annual metal production data, estimates
of stack emissions of both particulates and sulfur dioxide, and
observed air quality measurements. Annual metal production
estimates for each smeltert5! were supplemented with actual
production data and sulfur dioxide and particulate emissions data
provided to the U. S. Environmental Protection Agency by each
corporation.
Sulfur dioxide and particulate emission factors for copper
smelting, lead smelting, and zinc smelting'- •" were applied to the
annual metal production figures in order to determine average
daily stack emissions of both sulfur dioxide and particulates.
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 pollutant emission from the point source. In the case of East
Helena, annual average total suspended particulate concentrations
-------
405
are available for the years 1966, 1967, 1968, and 1969. The average
ratio of observed annual average total suspended particulate con-
centration to average daily particulate emitted is 1.05 T^^—
ton/day
The range of the data is 0.50. Using the techniques of inefficient
statistics described in Dixon and Massey.C?] the standard deviation
of the data is estimated to be 0.24. Using the common scientific
convention of expressing variation at the 50% confidence level, the
ratio of TSP concentration to tons of particulate emitted per day is
estimated to be 1.05 + .16
ton/day
4. RECAPITULATIONS BASED ON ISOPLETHES
Measurements of air pollutant exposure for communities in the
vicinity of Los Angeles is available from over forty separate
locations extending back in time as far as 1955. Unfortunately, all
stations are not available for every year. For example, in 1955
data from only seven stations is available; in contrast, data from
thirty-four stations is available in 1972. Continuous monitoring
instruments were used to record Oxidant, Nitrogen Dioxide and Sulfur
Dioxide. Estimates of environmental exposures for communities where
the air was not monitored can be made by constructing a set of isoplethes
for each pollutant for each year. The pollutant values for the
communities of interest can then be read directly from the isopleth
maps. The isoplethes are constructed by linearly interpolating
the values observed at pairs of adjacent stations on a line connecting
the stations. This process is repeated for all adjacent site
locations except those where topographic considerations, such as a
mountain, would invalidate the interpolation. Pollutant concentrations
at nearby locations external to the network of stations can be estimated
through linear extrapolation. Smooth lines are drawn between locations
of equal pollutant concentration. In cases where doubt or ambiguity
exists, concerning the detailed shape of a segment of the isopleth,
more detailed interpolations can be made to resolve the problem.
Historical trends are obtained by estimating the annual exposures
at the communities of interest from the isoplethes and graphically
showing these values as a function of year. Examples of oxidant
-------
Annual Average of Daily Maximum Hour Oxldant Concentration Vs Year
for Los Angeles Vicinity Communities
400- •
*Anahe1m
+Glendora
BJSanta Monica
•Vista
300-•
200"
100-
£
4J
I
o
o
ro
•o
*!"
X
o
1955
1960
Year
1965
1970
FXdORE 1.
-------
407
5. RECAPITULATIONS BASED ON INDICATORS
Indicators of pollution such as dustfall readings can be used
to estimate values of other pollutants such as sulfur dioxide or
suspended sulfates. Examinations of New York City data over the
time period 1954 through 1970 indicated that the ratio of the
annual average of suspended sulfate to the annual average of dustfall
did not appear to change substantially with year. The data indicated
that this ratio was 11.5 + 1.5 jjjjjjjjfo. This ratio was used to
estimate the values of suspended sulfate for years where dustfall data
existed and suspended sulfate data did not. The above approach
could not be used for sulfur dioxide estimates, since the available
data indicated that the ratio of annual average of S0» to the annual
average of dustfall did show a significant time trend.
6. CONCLUSIONS
Methods for recapitulation of past environmental exposures based
on production data, isoplethes and pollution indicators have been
described. Estimates of the accuracy of these techniques have been
provided. These approaches provide a useful method of estimating
past exposures to air pollutants.
-------
408
REFERENCES
1. Gifford, F. A., "Urban Air Pollution Modeling", NOAA Research
Laboratories, Atmospheric Turbulence and Diffusion Laboratory
Oak Ridge, Tenn., Dec. 1970.
2. Fox, D. G. and Pooler, F., "The Regional Air Pollution Study -
Updated". Proceedings of the Fourth Meeting of the Expert Panel
on A1r Pollution Modeling. N-30, NATO/CCMS, Oberursel, Germany,
May 1973.
3. Stern, A. Proceedings pf^Syjnposium on Multiple-Source Urban
Diffusion Models, Environmental Protection Agency AP-86, 1970.
4. Gifford, F.A., "The Simple ATDC Urban Air Pollution Model"
Proceedings of the Fourth Meeting of the Expert Panel on
A1r Pollution Modeling, N-30 NATO/CCMS Oberursel, Germany,
May 1973.
5. Yearbook of the American Bureau of Metal Statistics. 29th, 35th,
44th, and 51st Annual Issues, New York, American Bureau of Metal
Statistics.
6. Compilation of Air Pollutant Emission Factors. U. S. Environmental
Protection Agency, Research Triangle Park, N. C. Office of Air
Programs Publication No. AP-42. February 1972. 166p.
7. Dixon, W. J. and F. J. Massey, "Introduction to Statistical Analysis,"
2nd Ed., McGraw Hill, 1957.
-------
409
SAMPLING, ANALYSIS AND COMPOSITION OF AIRBORNE
PARTICULATE MATERIAL IN BELGIUM
R, DAMS AND M, DEMUYNCK
with the technical assistance of M. Nagels and R. Steelandt.
Institute of Nuclear Sciences, Rijksuniversiteit Gent, Belgium
ABSTRACT
During one year, aerosols have been sampled at 14 industrial,
urban and rural locations in Belgium. At each station daily 24
hour samples were taken on cellulose filters. All samples were
analysed for total suspended particulate and a large number for
about 40 inorganic components including most metals and halogens.
When properly summed, the measured constituents account for 50
to 75% of the total aerosol mass. The analytical techniques
applied include instrumental neutron activation analysis, flame-
less atomic absorption and gravimetry.
Mean, maximum and most frequent values at all 14 stations
for total suspended particulate and typical chemical compositions
were calculated. The variation of the levels over the country
and under the influence of the meteorological parameters are
discussed. The generally much higher ambient air levels in
industrial and urban areas, than in rural areas, give first
indications for the pollution origin of the aerosols sampled.
Inspection of the aerosol composition and its geographical dis-
tribution as a function of wind direction clearly indicate the
existence of important localised emission sources for elements
such as iron, zinc, cadmium, selenium, antimony, indium, silver,
mercury, etc.
-------
410
Intensive area-wide simultaneous samplings in industrial
and urban areas reveal the existence of strong concentration
gradients in the neighbourhood of steel plants and non-ferrous
industries. Measurements of the mass versus aerodynamic par-
ticle size for all elements, by means of 8 stage cascade im-
pactors, give information on the penetration and retention in
the lungs.
-------
411
1. Introduction
A 14 station aerosol sampling network was set up in order to evaluate
the particulate levels in the ambient air over Belgium. During the period
of October 1972 to October 1973 daily 24 hour samples were taken at all
14 stations. The sampling locations included the 5 major residential
centra of Belgium namely Gent, Antwerpen, Brussel, Liege, Charleroi
(GR, AR, BR, LR, CH) and the adjoining industrial areas (GI, AI, LI).
Samples were also taken in the medium-sized city Mechelen (ME),in the
rural areas of Ploegsteert and Mol (PL, ML), in the background stations
Houffalize and Dourbes (HF, DB) and in a sea-coast station Zeebrugge (ZB).
The samplers used in this study have been described extensively by
Dams and Heindryckx CO . They consist of a high vacuum rotary pump
equiped with a flow- and vacuum-meter and a 10 cm diameter P. V.C.
filterholder placed in a protective P. V. C. shelter. On these samplers
Whatman 41 cellulose filters are being used at a flow rate of approxi-
mately 400 m per 24 hours which results in the collection of 10 to 100
mg particulate material.
A number of days with relatively constant wind directions was selec-
ted for chemical analysis of 35 elements in the particulate material.
These selected samples were analysed by instrumental neutron activa-
tion analysis using a simplified irradiation-counting scheme (Dams et
al. L2] , Heindryckx and Dams C3] ). A short neutron irradiation was
followed by a 8 min gammaspectrometric count on a Ge(Li) detector
after a decay period of 6 min. After a longer irradiation, cooling times
of 1 to 3 days and 15 to 20 days were allowed for two additional gamma-
spectrometric measurements. The gammaspectra were recorded on
magnetic tape and the data were automatically reduced with a PDP 9
computer. A limited number of the selected samples was analysed for
some additional components. Sulphate, nitrate and ammonium were
determined spec trophotome trie ally. Silicon was measured after activa-
tion with 14 MeV neutrons and fluorine by reactor neutron activation
using a very short irradiation and counting scheme, owing to the small
half-life (H.2 s) of the isotope formed (Van Grieken and Dams C4D ,
Dams et al. E53 ). Cadmium and lead were determined by flameless
atomic absorption (Janssens and Dams [6] , [7]).
-------
412
2. Total Suspended Particulate Matter (TSP)
Total suspended particulate matter is one of the basic components to
be measured in air pollution studies and can be regarded as a good mea-
sure for general air quality. For all 14 Belgian stations mean levels,
coefficients of variation and integrated frequency distributions have been
calculated, the 50% and 95% levels are shown on a map in figure 1.
>ctober 1972 - octotwf'973
BELGIUM
Fig. 1. Total suspended particulate matter at 14 stations in Belgium.
Yearly means (50% and 95% percentiles) in ^g. m
The results suggest a generally high pollution level all over Belgium.
Only 6 out of 14 meet the EPA primary ambient air quality standard
for the U.S.A. (75 ^ig. m as annual geometric mean). Atmospheric
conditions such as ventilation, have of course an important influence
-------
on the daily levels of the particulate material. At some sites however,
local industrial emissions seem to dominate the general trend caused
by meteorological parameters. In all stations, TSP levels strongly de-
pend on wind direction. This is due to the fact that the ventilation varies
generally with wind direction, low ventilation coefficients being nearly
always associated with easterly winds. The TSP levels do not form a
single log-normal distribution but they can be separated into two log-
normal populations, one for wind directions from NNW to SSW and the
other for the sector NNE - SSE. This is illustrated for the station PL in
figure 2. Similar trends are found for the other stations.
STATION PL
Fig. 2. Total suspended particulate matter as a function of wind direc-
tion, normalised to the yearly mean, at Ploegsteert (PL).
-------
414
Important industrial contributions are recognised most easily in
areas with intensive steel industry, such as the Meuse valley and the
cities Liege and Charleroi. In figure 3, the mean TSP values for the
different wind directions in the city of Charleroi are plotted. While the
N
STATION CH
Fig. 3. Total suspended participate matter as a function of wind direc-
tion, normalised to the yearly mean, at Charleroi (CH).
relatively high values for the eastern sector can be explained as a re-
sult of the low ventilation, the high concentrations during westerly
winds are due to contributions from steel mills west of the city. Other
anthropogenic sources are more homogeneously distributed over the
country and as a consequence their influence is not so obvious. If however
the mean TSP values for the different days of the week are calculated a
smooth pattern is obtained (figure 4). The results are roughly in agree-
ment with the pattern of the daily electricity consumption which may be
regarded as a measure of general human Activity.
-------
415
120
110
100
90
80
SUN MON TUE WEN THU FRI SAT SUN
WEEKDAY
Fig. 4. Normalised weekly cycle ot total suspended particulate matter
averaged over one year at 14 stations.
3. Chemical Composition of Particulate Matter
The concentrations of most elements show larger local and daily
variations than the TSP levels. Elemental ambient levels do not only
depend on the particulate concentration, but specific emissions diffe-
rentiate the composition of the dust. It was therefore preferred to com-
pare particulate compositions instead of ambient air levels. In Table I
the mean composition of the suspended particulate matter in Belgium
is summarised. These arithmetic means are based on the analysis of
246 samples for most elements and of 14 samples for the components
Si, F, Hf, Pb, SO", NO ~ and NH . As a measure for the constancy
of these concentrations, the relative standard deviations are given.
The concentrations of the majority of elements are relatively constant
and similar to the concentrations in particulates collected at remote
European sites { Rahn C8] ). For all elements enrichment factors have
been calculated as the ratio of the concentration in the aerosol to the
concentration in crustal rock, normalised to the Si ratio (Mason [93 ).
In spite of this similarity,inspection of the composition and enrich-
ment factors still allows the elements to be grouped into classes with
-------
416
TABLE I
Mean composition of aerosols collected at 14 stations in Belgium.
Relative standard deviation (S. D. %) and soil-enrichment factors (E. F. )
for all components determined
Ele-
ments
F*
Na
Mg
Al
^
Si*
Cl
K
Ca
Sc
Ti
V
Cr
Mn
Fe
Co
Cu
Mean
(ppm)
2,970
31,200
5,440
15,040
41,900
56,750
14, 700
29, 700
4.0
1,180
480
240
1,830
31,900
39
750
Zn 11,700
Ga 43
As I . 270
1
Se 60
Br
2,370
S.D.
98
100
85
49
30
88
54
58
48
95
54
126
102
95
185
101
143
79
129
115
74
E. F.
31
7.3***
1.7
Ele-
ment
Ag
Cd
:„
1.2 1 Sb
++
1.0**
*•*•*
2,900***
3.7
I
Ce
Ba
5.4 || La
1.2 Ce
1.8 Sm
If
23 Eu
16 Lu
13 Hf*
4.2 W
10 Au
90 Hg
1.100 Pb*
20 Th
1,000 S0i" *
8,000 NO" *
6, 300*** NH4+ *
Mean
(ppm)
10
340
12
280
125
9.5
1,220
18
30
1.75
0.52
0. 76
1.80
18
0.62
8.5
10,500
2.7
132,000
60,500
36,400
S.D.
131
231
428
224
50
98
143
66
64
92
81
186
55
248
118
127
71
60
27
34
43
i
E.F.
1,000
11,000
800
9,300
1,600
20
20
4
3
2
3
10
4
80.
1,000
700
5,300
2.5
-
-
-
Based on a limited number of data
Reference
*** Enriched by natural sea spray
-------
417
similar behaviour. The first group contains the elements with small
coefficients of variation and low enrichment factors namely Al, Si, K
Ca, Sc, V, I, La, Ce, Hf, Th and the components SO~~, NO~, NH4+.
The nearly constant concentration of these elements in the aerosols
suggests emissions by various diffuse sources whereas the enrichment
factors in the vicinity of unity point towards emissions in soil-like
ratios. Nearly all these elements belong to the geochemical class of
the lithophile or oxyphile elements. In total they account for about 10%
of the aerosol mass. The ions, often called secondary aero sols, amount
to approximately 23%. The second group is composed of elements
having slightly larger coefficients of variation and still low enrichments
factors. Most of these elements have important industrial sources
such as steel emissions (Fe, Mn, Cr, etc. ), other anthropogenic sources
(Br, F, etc. ) or significant natural contributions (Na, Cl, Br, Mg).
Although some of those elements have very intensive industrial and
other anthropogenic sources, their enrichment factors remain low on
the average. This is explained by the influence of steel mills and fly ash.
These elements account for almost 15% of the total aerosol-mass and
most of them are considered as lithophilic or siderophilic. The elements
of the third group (Cd, In, Sb, Zn, As, Se, Ag, Au, Hg) have a larger
enrichment factor and a relatively large coefficient of variation. Their
major source is the non-ferrous metallurgy and some of them are of
toxicological importance. All these elements are geochemically classi-
fied as chalcophilic. Although their enrichment factor in air particulates
can be as high as 10, 000, their total mass accounts only for less than
2% of the particulate matter. A summation of all the components de-
termined in the aerosol reveals that only about 50% of the aerosol-mass
is identified. The organic fraction, soluble in benzene, is 7.5% on the
average and the particulate contains also about 5% water. The remaining
38% is probably organic material insoluble in benzene or carbon black.
In Pasadena-aerosols, carbon concentrations of 20 to 45% have been
found (Mueller et al. ClOD ).
By comparing the elemental compositions of the particulate matter
collected at one or two stations but under different wind directions re-
levant informations concerning the location of major emission sources
-------
418
can be obtained. An example of such a study is given in figure 5. The
D OO 500 XBO XXX
PI A:.'
"ig. 5. Selenium concentration in the aerosol as a function of wind
direction, at two locations in Antwerp. The concentrations
are proportional to the areas of the sectors.
selenium concentrations vary strongly at stations AR and AI as a function
of wind direction and point towards a large facility for non-ferrous
metallurgy.
-------
419
4. Mass-size Functions of the Elements
The distribution of the elements over the aerosol size range can give
important indications about the nature of the source, the residence time
in the air and the penetration and deposition in the human respiratory
system (Hatch and Gross £113 ). By means of Andersen cascade im-
pactors the aerosols are separated in 8 fractions, with 50% cutoff
diameters ranging from 11 to 0. 43 urn and the even smaller particles
are collected on a final filter. Figures 6 and 7 illustrate some results of
such a measurement in an urban-industrial area (LR) and in a background
area (HF). The elements can roughly be divided in three classes namely
those with maximum concentrations on the giant particles (^> 3 >un),
Br
fl-O2l O43 0.65 1-1
-i—r r i 1*1 n-r
21 33 4.7 7-0
1C
Fig. 6.
0.1 1
50% CUTOFF DIAM.^m)
Mass-size functions, as measured with Andersen cascade
impactor at background station, Houffalize (HF).
-------
A 20
Fig. 7.
1000-
100-
o
z
z 10T|
o
a:
UJ
o
z
o
u
Br
A;
Br
M'O2 O43 Q65 11 2.1 33 47 70 11
_1 I I I I III I
0.1
50°,o CUTOFF DIAM.(|jm'
Mass-size functions, as measured with Andersen cascade
impactor at urban-industrial station, Liege (LR).
those with maximum concentration on the medium-sized particles {0. 5-2
um) and those mainly associated with the smaller particles « i um).
The latter group shows generally a maximum concentration on the filter
and consists of relatively volatile elements such as Se, As, Sb, V, Hg,
Pb, Cl, Br and I. Condensation sources favour this kind of mass-size
distribution. The elements on medium-sized particles such as Zn, Cd,
In, Ag, Au are inhaled deeply into the lungs and have a relatively high
-------
421
probability for deposition. The elements Al, Si, Ti, Fe, Ca, Mg, Th,
Sc, etc. are generally produced by dispersion sources, which results
in their association with the very large particles.
5. Conclusion
This short summary of a relatively large study illustrates that in
Belgium the general pollution level of the aerosols is elevated in most
areas of the country. The chemical composition of the aerosols varies
as a function of location and as a function of wind direction. Especially
the chalcophile elements show large concentration gradients, indicative
of important emission sources. Their general association with the small
particles makes them potentially hazardous. Since the elements can he
divided in groups with common sources and similar atmospheric be-
haviour, a selected number of indicator elements can be found for each
group.
ACKNOWLEDGEMENT
The N. C. S. L. V. (Nationaal Centrum voor de studie van Luchtveront-
reiniging door Verbranding) is greatfully acknowledged for financial
support.
References
1. DAMS, R. , HEINDRYCKX, R. , "A High Volume Air Sampling System
for Use with Cellulose Filters", Atmosph. Environ. , J7, 319 (1973)
2. DAMS, R. , ROBBINS, J. A. , RAHN, K. A. , WINCHESTER. J. W. ,
"Nondestructive Neutron Activation Analysis of Air Pollution Particu-
lates". Anal. Chem. . 42, 861 (1970)
3. HEINDRYCKX, R. , DAMS, R. , "Evaluation of Three Procedures for
Neutron Activation Analysis of Elements in Atmospheric Aerosols
using Short-lived Isotopes", Radiochem. Radioanal. Letters. 16,
209 (1974)
4. VAN GRIEKEN, R. , DAMS, R. , "Determination of Silicon in Natural
and Pollution Aerosols by 14 MeV Neutron Activation Analysis",
Anal. Chim. Acta. 63, 369 (1973)
-------
422
5. DAMS, R. , BILLIET, J. , HOSTE, J. , "Neutron Activation Analysis
of F, Sc, Se, Ag and Hf in Aerosols using Short-lived Isotopes"
J. Environ. Analyt. Chern., in press
6. JANSSENS M. J. , DAMS, R. , "Determination of Lead in Atmospheric
Particulates with a Graphite Tube", Anal. Chim. Acta. 65, 41 (1973)
7. JANSSENS, M. J. , DAMS, R. , "Determination of Cadmium in Atmos-
pheric Particulates with a Graphite Tube", Anal. Chim. Acta.70.25 (197*0,
8. RAHN, K. A. , Inst. Nucl. Sci. , Ghent University, unpublished data
9. MASON, B. , Principles of Geochemistry. 3rt* Ed. , Wiley & Sons,
New York (1966)
10. MUELLER, P.K. , MOSLEY, R. M., PIERCE, L. B. , "Chemical
Composition of Pasadena Aerosol by Particle Size and Time of Day",
J. Colloid Interface Sci. . 39, 235 (1972)
11. HATCH, T. F., GROSS, P., Pulmonary Deposition and Retention of
Inhaled Aerosols, Acad. Press, New York (1964)
-------
423
THE CHAMP AIR QUALITY MONITORING PROGRAM
GEORGE LAUER+ AND FERRIS B, BENSON +
Rockwell International Science Center, Thousand Oaks, Ca, USA
Environmental Protection Agency, Research Triangle Park, NC,
USA
ABSTRACT
The CHAMP air monitoring system has been designed and in-
stalled to obtain short term, as well as long term* air monitor-
ing data to assess the influence of pollutants on the health of
five communities in the United States. The philosophy of design
uill be discussed with an emphasis on the techniques and method-
ology utilized to insure that the data obtained is valid. Pro-
gram requirements, both present and anticipated, will be presen-
ted and the system as it now exists will be described.
-------
424
The Community Health Air Monitoring Program (CHAMP) system has been
built to provide the Human Stuoies Laboratory of the Environmental Protec-
tion Agency (EPA) the physical data for its Community Health Environmental
Surveillance System (CHESS). The objective of the CHESS program is to deter-
mine the health effects of ambient airborne pollutants on the population in
order to establish realistic community air quality standards.
The CHAMP system is the product of an evolutionary growth of air qual-
ity monitoring which had its inception years ago. It is a product of the
experience gained by EPA personnel in the air quality monitoring field cou-
pled to the physical measurement skills developed within the aerospace com-
munity by the Rockwell International Science Center.
The system, at present, consists of 23 remote air quality monitoring
stations located in five geographical areas within the United States, cou-
pled to the central controller located at the National Environmental Research
Center (NERC) at Research Triangle Park, North Carolina. This paper will
describe the general features of the CHAMP system, stressing those features
which we feel have made new contributions to the state-of-the-art.
A block diagram of the remote monitoring station is shown in Figure 1.
Figure l
DATA MONITORING SYSTEM
AMBIENT SAMPLING SYSTEM
I
AEROMETRIC
SENSORS
METED ROUOGICAL
SENSORS
SOURCE
GASES
P
CONTROL
SYSTEM
VACUUM &
:ERO AIR SYSTEM
^^
CALIBRATION
SYSTEM
DATA ACQUISITION
SYSTEM
DATA TRANSMISSION
SYSTEM
Block Diagram of Remote Monitoring Station
-------
425
The sensors include aerometric pollutant analyzers, meteorological sensors,
and a particulate sampling system. A major technical requirement of the
program is to insure that all data obtained meets specific validity criteria.
In order to meet this objective, each critical point in the measurement
system is monitored. Each of the aerometric sensors employed must be cali-
brated at regular intervals; therefore, a dynamic calibration system has
been included as part of the station; as the calibration directly affects
the data validity, the critical elements of this subsystem are monitored.
Aerometric Sensors
At present the following sensors and measurement methods are used in
the CHAMP stations:
OXIDES OF NITROGEN (NO-NOX) - Determined by measurement of the light
given off in the reaction of 03 and NO. N02 is converted to NO prior to
reaction and is determined by subtraction.
OZONE - Determined by measurement of the amount of light given off by
the reaction of ozone and ethylene.
SULFUR DIOXIDE - Determined by measurement of the light emitted by
sulfur atoms excited in a hydrogen-rich flame. The sample is pretreated to
remove H2$ using a hot silver wire.
HYDROCARBONS - Methane is separated from other hydrocarbons in a gas
chromatographic column. The hydrocarbon determination is made using a
flame ionization detector.
PEROXYACETYLNITRATE. PAN - Separated utilizing a gas chromatograph.
Determined utilizing an electron capture detector.
TOTAL SUSPENDED PARTICULATE - Determined using a High Volume sampler
equipped with a flow controller.
SIZE SEPARATED PARTICULATE - Determined by filter impaction after size
separation utilizing a 20 micron impactor following by a 3.5 micron impac-
tor.
In addition, the following parameters constitute the primary data:
- Wind Speed
- Wind Direction
- Barometric Pressure
- Relative Humidity
- Ambient Temperature
These are measured using standard meteorological sensors with the
exception of the relative humidity sensor which utilizes a solid state semi-
conductor surface.
-------
426
Each aerometrlc sensor has been modified to permit electrical readout
of the complete operational status of the unit. These status indications
Include power supply on/off, range, etc. In addition, each critical flow
within the analyzer is monitored; examples include hydrogen flow in the SO-
analyzer, sample and ethylene flow in the ozone analyzer, sample and oxygen
flow 1n the oxides of nitrogen analyzer, etc. All other points which affect
the validity of the data are monitored. These include sample inlet flow,
valve status. Internal temperature of the shelter, etc.
Aerometric Sensor Calibration Subsystem
As indicated above, a complete dynamic calibration subsystem is pro-
vided in each station. This system has been designed to provide multipoint
calibration capabilities of each aerometric analyzer.
Maintenance of calibration sources at the low concentrations required
Is extremely difficult, If not impossible. This is due to reaction and ir-
reversible adsorption of the pollutant gas on the walls of the cylinders.
We, therefore, have utilized sources at reasonably high concentrations and
have provided a dynamic dilution system as shown in Figure 2. The design
objectives for the calibration system Include precision which is superior by
at least a factor of five to the instrument being calibrated and the ability
to Implement, at a later date, a fully automated calibration capability.
It was, therefore, decided at the beginning to utilize mass flow sensors
throughout; these provide significantly Improved flow measurement over vis-
ually read rotometers, as well as the ability to premanently record the cali-
bration process Itself.
Figure 2
CALIBRATION SUBSYSTEM
CAUMATHN
HUME OAS
ZEHO
AMI
Mi
M
!••••• ^ ,.
!••••• ^^
• *•••• ^^
_, TO
MANIFOLD
Dynamic Dilution System'
-------
427
Data Acquisition System (DAS)
The fundamental improvement in the CHAMP monitoring system in this area
is the utilization of a minicomputer as the primary logical element of the
system. A block diagram of the CHAMP DAS is shown in Figure 3. We have
employed a Digital Equipment Corporation PDP-8/M minicomputer with 16K of
core, hardware multiply/divide, power fail/auto restart, a 60 Hz real-time
clock, and a modem interface. The minicomputer is interfaced to a Pertec
9-track, 800 bpi magnetic tape, a Xincom multiplexor/ADC, a digital display
unit, 96 bits of digital input, and 96 bits of digital output.
Figure 3
DAIA ACQUISITION SET
TELETYPE
OPTICAL
ISOLATORS
^^^^
•fcaurt
^^Mi
96 BITS
DIGITAL
INPUT
OPTICAL
ISOLATORS
M^H
^^^M
M^M
^MM
96 BITS
DIGITAL
OUTPUT
MODEM
CONTROLLER
POP - 8/M MINICOMPUTER
MUX/ADC
CONTROLLER
MAG TAPE
CONTROL
MAG TAPE
6 DIGIT
DISPLAY
43 CHANNEL
MUX & ANALOG
TO DIGITAL
CONVERTER
CHAMP Data Acquisition System
The utilization of the minicomputer in the station permits a highly
flexible mode of operation. One of the prime objectives of the program is
to acquire data which can be validated. The term "valid" implies that all
aspects of the remote station operation are functional and within tolerance.
The implementation of this concept is shown in Figure 4 for the oxides of
nitrogen analyzer. This instrument operates on the principle of measuring
the amount of light emitted by the chemical reaction NO + 03 -* N02 + QJnv.
-------
428
T
SAMPLING
SYSTEM
AMBIENT
INLET
®—
-------
429
has been properly performed. The operator can also control the operating
parameters for each instrument; these include the averaging time and the
calibration constants.
At present, the system consists of 23 remote stations located in
Southern California, Utah, New York, Birmingham, Alabama, Charlotte, and
Research Triangle Park, North Carolina. Each can be individually inter-
rograted by a central controller via dial up voice grade lines. The remote
station is equipped with a full duplex, two wire autoanswer modem. Full
duplex operation is achieved by utilizing 1200 baud forward (remote to
central) transmission and 150 baud reverse (central to remote) transmission.
The data is transmitted in 512 byte blocks with byte parity and record
checksums. Retransmission is utilized if a communication error is found.
The central controller consists of a dual processor system, configured
of a PDP-11/05 connected to a PDP-11/40 via a "bus window". The configur-
ation was chosen to provide for rapid communications processing simultan-
eously with batch processing of the data. A block diagram of the central
controller is shown in Figure 5. The central controller has been provided,
Figure 5
CENTRAL CONTROLLER HARDWARE
PDP-11/40
3ZK CORE
!«ORY MANAGEMENT
FLOATING POINT
CLOCK
Block Diagram of Central Controller.
principally, as a means to aid in validation of the physical data as it is
recieved. The validation procedure consists of checking that the various
critical instrument parameters are within tolerance, that the various valves
are correctly set, and that the calibrations of the instruments have been
-------
430
made at the proper interval. In addition, the controller is utilized to
store recent data for comparison checks, property location control, etc.
Overall validity of the data of a long term requires a systematic
quality assurance program. Such a program has been implemented for the
CHAMP system. A full five point calibration is given each aerometric
analyzer every two weeks. Other sensors are calibrated at periods consis-
tent with the known drift and failure rate of the device. An archive of
each instrument's performance is kept in order to track degradation and
systematic problems.
Each calibration source is certified by two independent means in the
laboratory. Source gas tanks are rechecked every three months to insure
that degradation has not taken place. Each station is visited by quality
assurance personnel at irregular intervals to establish that the local
operator is performing, the normal routine maintenance tasks required.
In conclusion, the CHAMP system provides the EPA the tool by which
aerometric data for the CHESS program can be obtained. By utilization of
state-of-the-art technology, much of it derived from aerospace experience,
the validity of the data is enhanced, thus establishing a high confidence
factor for any air quality standards derived from the program.
The work upon which this publication is based was performed pursuant
to Contract No. 68-02-0758 with the Environmental Protection Agency.
-------
431
STUDIES OF TRACE SUBSTANCES IN AN URBAN ATMOSPHERE
THEO J, KNEIP/ MERRIL EISENBUD, MICHAEL KLEINMAN
AND DAVID BERNSTEIN
New York University Medical Center, New York, NY, USA
ABSTRACT
An investigation has been undertaken of certain trace sub-
stances in the atmosphere of New York City in order to: 1) im-
prove our understanding of the seasonal and year-to-year vari-
ations in the concentrations of these substances; 2) ascertain
their particle size distribution; 3) determine the extent to
which human tissue burdens are due to atmospheric exposure; and
4) enable one to apportion the total suspended particulates
among some of the major sources, such as space heating, power
generation, ocean spray, and automobile traffic. In addition,
certain characteristics of settled dust have also been studied
in order to understand the relationship of settled to airborne
dust and to develop methods by which the source of settled dust
can be ascertained through analytical procedures. This paper
summarizes the findings to date, with particular emphasis on the
new methods that have been developed in the course of our study.
-------
432
1. Introduction
A study of the composition of the urban aerosols of
New York city has been undertaken in the hopes of achieving
better understanding of the sources, distribution, and fate of
aerosols in the urban environment. The overall objectives of
this program are 1) To use the physical and chemical properties
of suspended and settled dusts to characterize their sources.
2) To ascertain the extent to which trace elements present in
suspended dust are absorbed by New York City residents.
In order to achieve these objectives we have undertaken to:
a) Determine temporal patterns in the variation of trace metal
and ion concentration in total suspended particulate (TSP) and
settled dust.
b) Establish the effects of meteorological parameters on the
TSP and trace element concentrations.
c) Define the sources of the trace metals or ions.
d) Relate the aerosol trace element compositions to human tissue
burdens„
A system for sampling airborne particulates is in use that is
capable of continuous operation for 7 days at a constant flow
rate of Q57 m /minute, prom 1967 through 1970, and 1972 through
1973, samples have been collected at 4 locations for analysis
by atomic absorption spectrophotometry as reported by
Kneip, et. al. [1]. By use of weekly rather than daily samples,
short-term fluctuations were successfully averaged out, with
a substantial reduction in sampling load.
In addition, sampling of dustfall was initiated in 1972.
Weekly samples from each of the four stations are analyzed
for 9 metals, and monthly composites are analyzed for 4 to 6
additional elements or ions. (Table I) Monthly dustfall
samples are analyzed for all of these materials. Methods to
quantitatively analyze trace element concentrations in sus-
-------
433
TABLE I
ELEMENTS ANALYZED IN AIR FILTER SAMPLES
Element Standard Deviation (%)
WEEKLY
Cadmium 10
Chromium 18
Copper 11
Iron 22
Lead 2
Manganese 7
Nickel 8
Vanadium 14
Zinc 3
MONTHLY
Calcium
Magnesium
Potassium
Sodium
Chloride
Nitrate
pended particulates, settled dust and human tissue specimens
were restudied to obtain increased accuracy and sensitivity
and improved precision. These methods were tested using
available National Bureau of Standards reference materials.
2. Experimental
Weekly samples of suspended particulates are collected at
four locations in New York City and one location in a rural
area. The locations and characteristics of these stations are
given in Table II. Station No. 5, which has been installed in
Sterling Forest, about 40 miles northwest of New York City, is
on a hill about 1 mile from the nearest road and should serve
-------
434
TABLE II
LOCATIONS OF SAMPLING SITES
Station
1
Address
376 Hudson St.
(Houston St.)
500 First Ave.
(30th St)
Sedgewick Ave.
(near Fordham Rd)
Kissena Blvd.
(near 69th Rd)
Region
S.W. Manhattan
East Side of
Manhattan
Western Bronx
North Queens
Long Meadow Rd Tuxedo, N.Y.
Height Above
Street
13 stories
14 stories
3 stories
4 stories
700' above
road level
as an indicator of the background level of the aerosols enter-
ing the New York City area on northwest winds.
Samples are obtained by drawing air through a 10.5 cm x
25.4 cm glass fiber filter with a Roots-Connorsville pump.
The flow is controlled at 0,57 m^/rain by use of a by-pass valve
actuated by a feedback loop to maintain a constant pressure
drop across an orifice which is in line with the sampling head.
A section of about 45 cm2 is cut from the filter which
has a total exposed area of about 406 cm . The fraction taken
as the sub-sample is determined by a weight fraction technique,
The metals are dissolved by wet ashing in a nitric-perchloric
acid mixture (4:1) to the first appearance of perchloric acid
fumes.
3. Results and Discussion
3.1 TSP and Composition Data
The samples of the New York City aerosol have been ana-
lyzed in this way for eleven trace metals previously listed in
Table I. The precisions given in the table were obtained by
analysis of replicate portions of the 10.5 cm x 25.4 cm glass
fiber filters and include the variapce due to sampling, sub-
sampling, sample preparation and analysis.
-------
435
The general trends in the metals concentrations are shown
in Table III for Stations 1, 2 and 3. Decreases have been ob-
served for copper, chromium, lead and manganese as well as
vanadium in the 1969-1973 period.
TABLE III
TgACE^METALS IN SUSPENDEp PARTICULATES
AT STATION 1, STATION 2, AND, STATION 3
ANNUAL GEOMETRIC MEANS (ng/m )
STATION 1
Element
Pb
V
Cu
Cr
Fe
Mn
Na
Ni
Cd
STATION 2
Pb
V
Cu
Cr
Fe
Mn
Zn
Ni
Cd
STATION 3
Pb
V
Cu
Cr
Fe
Mn
Zn
Ni
Cd
1968
1969
1972
1973
2990
1000
212
63
_
71
2380
_
23
1900
564
187
42
—
61
490
-
11
1390
81
152
28
1301
45
498
29
16
1143
169
95
22
1723
55
634
58
9
2100
874
526
33
_
89
670
_
10
1361
59
45
11
1332
26
350
28
5
1193
68
62
7
1281
31
416
67
5
3820
1230
133
49
_
54
730
150
14
2760
795
45
23
_
40
1120
122
9
1880
48
48
5
1590
27
232
127
4
1520
67
43
4
1310
31
273
221
3
-------
436
The TSP levels in the city have shown a general decline
as noted in the data of Table IV. Current annual means have
declined to less than 90 u.g/m and one site (station 3) reached
the primary quality goal of 75 ug/m3 in 1973.
TABLE IV
TOTAL SUSPENDED PARTICUIATE 1968-1973
ANNUAL GEOMETRIC MEANS |j.g/m
1968 1969 1972 1973
STATION 1 125 104
STATION 2 - 134 79 81
STATION 3 113 108 79 75
The concentration of vanadium has dropped by an order of
magnitude during this same period as shown in Figure 1. From
1968 through 1969 the airborne v concentration decreased by
about one-third to one-half as compared to a reduction of 50%
in the sulfur content in fuel oil burned for space heating.
During this period, Consolidated Edison was required to burn
residual fuel oil with no more than 1% sulfur, while oil used
for space heating was reduced gradually from 2.2% S to 1.0% S.
From 1969 to 1973 a further drop occurred with airborne vana-
dium falling in 1972-73 to one-tenth the 1968 levels. Fuel
sulfur contents in 1972-73 were restricted to 0.3% for heating
oils and 0.5% for power station usage.
3.2 Meteorological Factors
The ambient concentration of a pollutant in the atmos-
phere depends on the nature of its source, its dilution or
dispersion by atmospheric motions, and its rate of removal
from the atmosphere by washout in rainfall and by direct im-
pingement onto surfaces. If one is to use observed trends in
airborne trace metal concentrations as indications of changes
in source emission characteristics, it is necessary to compen-
sate in some way for variations in 'the meteorological phenomena
-------
437
1200
•g 800
s,
0>
400
0
Annual Geometric Mean V Concentration
A Station No. 1
• Station No. 2
• Station No. 3
160
80
0
Annual Geometric Mean TSP Concentration
3
2(
1
n
Percent Sulfur in Fuel Oil
"^-^^^-Home Heating Oil Consolidated Edison
t . i i 1 1 — IJ • 1 1 1
1968 1969 1970
Figure 1 t Variation in the atmospheric concentration of vanadium (V) and
total suspended particulates (TSP) as well as in the concentration
of sulphur in fuel oil in New York city for the period 1968-1973
-------
438
as reported by Rubin [2] . The meteorological factors which
seem most important for this purpose are precipitation, which
relates to atmospheric scavanging processes as reported by
Engleman [3fc wind direction, which is important in evaluating
emissions from major point sources; and wind speed and mixing
depth as reported by Holzworth 14 1, which relate to atmospher-
ic dispersion processes.
We have used seasons related to heating needs in our
studies rather than calendar seasons, so that Winter is
December, January and February; Spring is March, April and
May; Summer is June, July and August; and Fall is September,
October and November. No consistent rainfall pattern has been
found as a function of season. For this reason we have, for
the moment, postponed detailed examination of the effects of
precipitation on observed concentrations of airborne metals
and particulates. The degree to which pollutants diffuse
vertically depends to a great extent on the vertical temper-
ature gradient of the atmosphere. In New York City, this
profile is determined daily, shortly after sunrise, using
balloon-borne temperature and altitude sensors which relay
information to ground based stations. Using the method
described by Holzworth [4], these profiles can be used to
calculate the height to which mixing will occur.
Daily dispersion factors were computed for this study as
•»
the product of the mixing depth and the wind speed. Average
weekly values for mixing depth, wind speed and dispersion
factor were examined and both wind speed and mixing depth were
found to exhibit seasonal variations, being lower during the
summer and higher in the winter. The week-to-week variations
are large and the indulging pattern is difficult to visualize
in unsmoothed data. Four-week moving averages of the disper-
sion factor show a very distinct pattern, with a definite
minimum during the summer and maximum during the winter months.
-------
439
This pattern has significance in that during periods of high
dispersion, the concentration of airborne pollutants from re-
latively constant sources woul^ be re^uce^ because of dilution
effects. During periods of low dispersion, as during the
summer months, pollutant concentrations would appear to be in-
creased even though source emission rates were unchanged.
A means of compensating for this variable dispersion is
to normalize particulate concentrations to a constant disper-
sion factor. The TSP and trace metal data for the period
February 1972 to March 1973 have been normalized to a constant
2
dispersion Eactor of 3100 m /sec, the average factor for 1972,
by multiplying the concentration by the monthly dispersion
factor and dividing by 3100.
The effect of this transformation on the data is clearly
shown by several seasonal shifts. For instance, peak values
for Total Suspended Particulate shift from the summer to
winter. This indicates that an important source of particulate
emission during the winter months is obscured by the effect of
the dispersion factor. It is reasonable to assume that space
heating is such a source. Support for this hypothesis is
found in the record of degree days. TSP values as a function
of degree days (7 day weekly averages) indicates that these
two variables are strongly correlated, 10.001. The winter
peak in the normalized TSP concentration evidently reflects
the increased emissions from space heating.
Based on these results and the assumption that other
sources remain constant over the year, a baseline average was
estimated from the summer results and this average subtracted
from weekly winter TSP values to estimate the magnitude of the
heating systems as sources. These calculations show that
heating emissions contribute 20% of the annual average TSP
with the highest weekly fraction estimated at 60%.
The data for lead, vanadium and copper have also been
examined in "as is" and in dispersion normalized forms. The
-------
440
results for copper show no significant seasonal relationships
both before and after normalization. The corrected values for
vanadium show effects similar to those for TSP with more dis-
tinct winter peaks after normalizing. The lead data show no
seasonal relation before normalizing, however, the normalized
values are distinctly lower during the summer, a period when
schools are closed and many people take vacations. No quanti-
tative detailed traffic analyses are available, but qualitative
observations from many sources agree that the summer is a
period of significantly lower traffic.
Dispersion normalization of the TSP and metals data
appears to offer a means of obtaining correlations between
such data and source emissions free of variations caused by
the variable meteorological conditions from season to season.
3.3 Source Related Trace Elements
One of our major interests in the trace element pollu-
tants in urban aerosols stems mainly from the need to achieve
a better understanding of the sources of urban particulate
pollutants. Larsen [5] estimated the total contribution of
the automobile to the Boston aerosol through a study of lead
and other aerosol components from an automotive tunnel. More
recently, Kneip et. al. [6] and Friedlander [7] have examined
the relationship between aerosol composition data and source
contributions.
The tracers expected to be of use are sodium and chlorine
for sea salt; calcium, aluminum, silicon and potassium for soil?
lead for automotive sources; vanadium and possibly nickel for
fuel oil (principally residual); and copper, zinc, chromium
and others as industrial tracers. Based upon the emission
inventory and source sampling data in the literature we have
selected the following five elements as probably associated
with the major sources of the urban aerosol.
-------
441
The literature values for source terms for each of these
elements were evaluated to determine the similarity of the
sources to those in the system under study, and therefore the
likelihood that these terms would be representative for our
study and values selected as preliminary source term factors
as given in Table V. As sampling and analysis of sources was
prohibitively expensive, a mathematical technique has been
developed to confirm the factors selected. The method and
results were discussed previously by Kneip [6].
TABLE V
TRACER ELEMENTS CHARACTERISTIC
OF SELECTED EMISSION AEROSOL SOURCES
Element Associated Source Element Fraction in
Source Aerosol
Pb Automobile 0.09
V Fuel Oil Burning 0.047
Cu Incineration 0.01
jja Sea Salt Aerosols 0.33
K Soil Particulates 0.024
Using the literature values for the concentration of each
element in the aerosols emitted by the sources, a matrix is
solved to establish the fraction of each of the elements attri-
butable to the individual sources. These fractions are then
used in a correlation-regression analysis to recalculate the
source terms of the elements for each aerosol source as co-
efficients in the equation obtained by combining and rearrang-
ing equations as follows:
-------
442
TSP
£)EJ
F; - = f. . X E
3-D ID
C.
E . = F. . X C.
D ID i
TSP = y. .1 XF.. XC.
^i.D T 13 l
•L • •
j = an individual source
i = a selected element
E. = ambient concentration of aerosol associated
3 with the source j
F.. = Fraction of airborne concentration of element
i associated with the source j
f j • = Fraction of element i in aerosol from source j
C. = Concentration of element i in ug/m in the
ambient aerosol.
The sources chosen must account for the total amount of each
of the elements selected, and the TSP total must be a summa-
tion of the sources under study.
What we are attempting at this point can be expressed as
follows. We have assumed that there are six contributors to
the TSP. These are automobiles, fuel burning {for both space
heating and power production), incinerators (both municipal
and apartment house), suspended particles of soil, salt spray
and particles from undefined industrial and commercial sources
(labelled for convenience "other"). The particles associated
with the first five of these sources can be traced by certain
elements: Pb, V, Cu, K, and Na, respectively.
The multiple regression solution to the equations shown
above essentially tries to account for changes in the TSP from
week to week in terms of changes in the selected metal concen-
trations. That portion of the TSP variation which is unrelated
-------
443
to changes in the metal concentrations is attributed to changes
in the undefined sources (others). From this relation using
the fractions of each element due to the various sources (F..},
we can obtain recalculated values for the element fractions
(f. .) -
JO
These fractions for element composition in emitted
aerosols, as generated from actual ambient aerosol compositions,
can then be used to calculate the aerosol mass associated with
each source and the apparent percentage that each source
contributes to the total aerosol. No significant change in
the calculated element fractions (f. .) is found upon successive
recalculations indicating a satisfactory fit of the data to the
equations in use. The results for these fractions are given
in Table VI. The results for the source contributions to the
ambient TSP are given in Table VII.
Element
Pb
Cu
V
TABLE VI
ELEMENT FRACTIONS CALCULATED
FOR ASSOCIATION TO SELECTED SOURCES
Element Fraction (fi
Related
Source Station 2 Station 3
Auto 0.055 0.076
Incineration 0.013 0.051
Fuel Oil 0.008 0.006
^pooled
Data
0.097
0.012
0.005
-------
444
TABLE VII
PERCENT OF TSP ASSOCIATED WITH ELEMENT
February - February 1972 - January 1973
August 1972
Element
Pb
V
Cu
Na & K
Other
Medical
Center
32
11
4
17
37
Medical
Center
31
10
5
(18)
54 (36)
Carpenter
Hall
32
11
1
-
55
Combined
Data
22
14
4
-
59
( ) Values estimated on the assumption that the February-
August contribution was unchanged.
These percentages represent the fractions of the TSP
which can be mathematically correlated to the variations in
composition as discussed. This method associates the TSP with
the ambient aerosol, not the actual aerosol emitted by the
source. Thus, agglomeration, attachment, condensation, ira-
paction, or other causes of growth or dispersion of particles
have already occurred prior to the measurements used in these
calculations. Moreover, the association between Pb and 20-30%
of TSP may result from the fact that Pb is present in auto-
mobile exhaust and that automobiles generate dusts in a variety
of ways, including abrasion of brake linings, tires and road
surfaces, or comminution and resuspension of settled dust on
street surfaces.
The data do provide a means of judging the probable
importance of further examination of each source. For instance
the estimates for vanadium related aerosols show a total of
10 to 14% of the TSP as related to oil burning sources. We
have estimated that space heating contributes about 20% of
5
-------
445
the TSP at Station No. 2 during 1972 based on the "excess"
normalized TSP observed during the winter heating season.
This result compares well with the estimates of 10 to 14%.
It appears that further reductions in these emissions are not
likely to be productive if concentrated on the large power
stations burning residual fuels, but must include efforts to
reduce the emissions related to the heating systems in the
city. As a definite winter peak occurs in the TSP, such
efforts may be crucial in reaching a 75 n.g/ra3 annual mean.
3.4 Particle size distributions^
To obtain information about the particle size spectrum
of the urban aerosol, our colleague, Morton Lippmann, has
designed a size selective particle classifier which will
enable long-term sampling of large volumes of the urban air.
Analyses of these samples will permit correlation of trace
metal concentrations as a function of particle size and permit
comparisons to emission source characteristics and meteorolog-
ical data.
The device uses an array of size selective centrifugal
separators (cyclones) as precollectors in four-parallel, two-
stage systems using glass fiber filters for the second stage.
This type of cyclone has been in use for short-term "respirable"
dust sampling in industry, but until now has not been incor-
porated in a long-term urban aerosol particle size collector.
The air is moved by an integral motor-blower units (Radeco
# 809V). The cyclones collect particles larger than the
selected particle size cut while the filter papers collect
particles smaller than the selected particle size cut. One
filter is used with no precollector to provide a total value
independent of constant flow sampler. By carefully adjusting
the inlet geometries to get equal inlet velocities, it has
been possible to get reasonable agreement of TSP and the sums
of the weights for the two stages in each classifier. The
-------
446
TSP values obtained by our constant flow sampler and the array
of particle size classifying samplers are given in Table VIII.
The 3.5 \w stage has given high total flows due to a flow
control problem from the low speed setting used to obtain the
3.5 iim cut. A new feedback circuit will be used which will
afford improved control and lower the total flow to the de-
sired value.
TABLE VIII
TOTAL SUSPENDED PARTICULATE
FROM SIZE SELECTIVE STAGES
Mass collected (ug/m )
Run 1 Run 2
Sampler
Constant Flow
O.3 m /rain
Total Stage
50* 3.5 pro
50% 2.5 urn
50% 1.5 urn
50% 0.5 urn
114
115
142
118
121
101
Lost
110
130
No Sample
100
103
3.5 Dustfall
Monthly samples of dustfall collected at these stations
in New York City have been analyzed for Cd, Cu, Cr, Fe, Mn,
Hi. Pb, Vf and Zn. Two years of data do not disclose any
discernable trends in the data as a function of time. A pre-
liminary attempt to correlate dustfall levels with rainfall
has not been fruitful. For example, although November 1972
had twenty times the precipitation as August 1972, there was
no significant difference between the masses of the trace
elements deposited (P
-------
447
tional data will be needed if seasonal trends are to be
delineated.
3.6 Deposition Velocities
The deposition velocity is defined as the particle depo-
sition flux to a surface divided by the particle concentration
2
above that surface. Thus, ug of particles/cm /month, deposited
in a dustfall collector, divided by the average concentration
of particulates in air measured during that month, in units of
ug/m , multiplied by a dimensional constant of 0.38 yields
deposition velocities in units of cm/sec. Average deposition
velocities calculated for each metal are shown in Table IX
along with particle size estimates based on a Stoke's Law
approximation and an assumed particle size of 0.5 um for lead.
TABLE IX
AVERAGE DEPOSITION VELOCITIES
AND ESTIMATED PARTICLE SIZES
Element
Lead
Nickel
Cadmium
Vanadium
Zinc
Iron
Manganese
Copper
Average
Deposition
Velocity
(cm/sec)
0.65±0.08
1.34*0.26
2.59+0.58
2.41±0.17
3.85+0.90
4.70+1.06
5.71+0.95
7.04±0.76
95% Confidence Range of
Apparent Particle Size,
MMD
(U)
0.46-0.5Z
0.64-0.78
0.87-1.08
0.92-0.99
1.06-1.35
1.18-1.48
1.34-1.60
1.76-1.83
-------
448
3.7 Trace Elements in Human Tissues
Analysis of human tissues has been undertaken to improve
the knowledge of the relationships between trace metal-tissue
distributions and the various routes of exposure, with parti-
cular emphasis on the urban aerosol source. The tissue distri-
bution of the elements in various tissues generally agree with
the values reported by Tipton [8]. Lead, nickel and chromium
have been found to show elevated levels in the combined Hilar
and pulmonary lymph nodes as compared to the lung. This dis-
tribution which indicates a phagocytic clearance mechanism,
and a clearance half-time of 39 days for lead from our data are
in qualitative agreement with the fact that we find lead in
the New York aerosol to be less than 20% soluble in water.
4. Conclusions
These studies have provided data on the concentrations
of trace elements in the aerosol to which residents are ex-
posed in New York City. Correlations of the composition and
TSP variations have shown associations to exist between lead,
vanadium, copper, sodium and potassium and some 40% to 65%
of the total aerosol. The aerosol associated with lead (and
possibly the automobile) accounts for 20 to 30% of the total,
while that associated with vanadium (and probably oil burning
sources) accounts for 1O% to 14%. A further effort must be
made to determine whether other major source correlations can
be found in the remaining aerosol currently attributed to
"other" sources.
A means has been evaluated for correcting for meteorolog-
ical variations and TSP and metal concentrations have been
normalized to constant mixing conditions. The results have
revealed distinct source-related seasonal variations which had
been partially or wholly obscurred by the variations and inter-
actions in the meteorological factors. Vanadium, lead and TSP
show maxima and minima in the normalized data which now can be
-------
449
related to source terms independent of the variable meteoro-
logical effects. The normalized TSP and vanadium results
clearly reflect space heating and closely correlate mathe-
matically to degree day data. The annual space heating con-
tribution has been calculated to be about 20% on an annual
basis from these results and is in fair agreement with the 10
to 14% estimated through the correlation method.
Dustfall deposition velocities have been determined for
several elements and the size of the particles associated with
these elements has been estimated using an empirically derived
Stokes Law approximation based on the fact that Pb particulates
have a mass median diameter (nund) of about 0.5 pjn. Copper
appears to be on relatively large particles, 1.8 |xm mmd while
nickel and vanadium which are believed to originate as fumes
in oil burning sources, have mass median diameters of less than
1 um.
The studies are being extended using a new size selective
sampler to determine particle size-composition relationships,
and possible interrelationships with human tissue burdens.
Further efforts are being made to refine the data and the
computational techniques in order to better assess the source-
aerosol and tissue burden-aerosol relationships.
-------
450
SBFSKENCES
1 KNEIP, T.J., EISENBUD, M., STREHLOW, C.D., and
FREUDENTHAL, P.C.,"Airborne particulates in New York
City", A.P.C.A.J. 20. 144-149 (1971).
2 RUBIN, E.S., "The influence of annual meteorological
variations on regional air pollution modeling: A case
study of Allegany County, Pennsylvania", J.A.P.C.A. 24,
349 (1974).
3 ENGLEMENN, R., "Scavanging prediction using ratios of
concentrations in air and precipitation". Presented
at the 1970 Precipitation Scavanging Conference,
Richland, Washington (1970).
4 HOLZWORTH, G.C., "Mixing depths, wind speeds and air
pollution potential for selected locations in the
United States", Appl. Meteorol. 6, 1039 (1967).
5 LARSEN, R.I., "Air pollution from motor vehicles",
Annals of the New York Academy of Sciences, 136,
275 (1966).
6 KNEIP, T.J., KLEINMAN, M.T., and EISENBUD. M., "Relative
contribution of emission sources to the total airborne
particulates in New York City", Proceedings of the
Third International Clean Air congress, Dusseldorf,
Federal Republic of Germany (1973).
7 FRIEDLANDER, S.K., " Relating particula te pollution to
sources: Case of the LOS Angeles aerosol". Proceedings
of the Third International Clean Air Congress, Dusseldorf,
Federal Republic of Germany (1973).
8 TIPTON, I.H. and COOK, M.J., "Trace elements in human
tissue, part II Adult subjects from the United States",
Health Physics 9. 103 (1963).
This research is supported in part by the Edison Electric
Institute and American Petroleum Institute, and is part of a
center grant from the National Institute of Environmental Health
Sciences, Grant No. ES00260.
-------
451
AIR POLLUTION PROBLEMS IN LATIN AMERICA
R, HADDAD
Pan American Center for Sanitary Engineering and Environmental
Sciences, Pan American Health Organization, Lima, Peru
ABSTRACT
Latin America includes 20 countries of different size,
socio-economic condition and type of government. The total
population is reaching 300,000,000, with 50% rural, but inclu-
ding two of the ten largest cities in the world, with nearly
10,000,000 inhabitants in each.
In the late 50's some measurements of air pollutants were
made. In 1967, the Pan American Health Organization started a
continental network of sampling stations. The Pan American
Network of Air Pollution Monitoring Stations has now 92 stations
in 25 cities of IS countries. By December 1973 it had collec-
ted over 300,000 data on particles and sulphur dioxide, which
show the magnitude of the problems affecting the main cities.
Some of them present monthly averages three or more times higher
than the reference levels.
The main air pollution problems seem to be industrial com-
bustions. Motor vehicles, with a total around 10,000,000, are
concentrated in the main cities. These, combined with plenty
of sunshine cause a problem in some of them. Home heating is
seldom used, and only a small proportion of garbage is inciner-
ated.
-------
452
An Under-Secretariat for Environmental Protection Das esta-
blished by Mexico -in 1972, and by Argentina and Brazil in 1973,
A few control programs are at present in operation, and some
institutions have been especially created to make air pollution
studies. The main deterrent for good air pollution programs is
the scarcity of well-trained personnel. Academic and short
courses, as well as scholarships and travel grants, are trying
to solve this problem.
Research on air pollution is just starting in Latin America.
Most of it deals with measurement of air pollutants, but some has
been done on health effects, mainly in Chile and Mexico. These
studies are still too few in number, but they are showing the
right path to the Latin American scientists and those from other
developing countries.
-------
453
Latin America includes 20 countries of very different size, popula-
tion, socio-economic condition and even type of government. It is diffi-
cult to compare the giant Brazil with the tiny El Salvador; a fairly well-
developed, industrialized country, like Argentina, with a mainly agricul-
tural one, like Haiti; a socialist country as Cuba, with one with a market
economy, as Venezuela; etc. Per capita incomes differ also widely, going
from less than US$ 100 to nearly US$ 1,200 per year. The environmental
problems are also very different. The total population is reaching
300,000,000 including two of the ten largest cities in the world, with
nearly 10,000,000 inhabitants each, and 50% of the total as rural popula-
tion, very disseminated. Air pollution is still incipient, but some of
the most important cities, like Mexico City, Buenos Aires, Sao Paulo and
others, are quite polluted.
The Latin American governments began to worry about air pollution
during the late 50Ts when some sporadic measurements were made, mainly by
universities and occupational health departments. In 1965 the Directing
Council of the Pan American Health Organization (PAHO), to which the 20
countries belong, recommended to its Director the initiation of air and
water pollution programs, to help the Member Governments in the develop-
ment of adequate policies. At the end of 1966 a permanent air pollution
regional consultant was hired.
Since any good program should start with a thorough knowledge of
the magnitude of the problems, PAHO, through its Pan American Center for
Sanitary Engineering and Environmental Sciences (CEPIS), decided to ini-
tiate a continental network of air pollution sampling stations. It was
decided to collect monthly samples of settled dust, by means of a dust
jar, and daily samples of suspended dust and sulphur dioxide, measured by
the loss of reflectivity of a white filter paper and by the hydrogen per-
oxide method respectively. A very detailed Manual of Operations was pre-
pared, in Spanish. The Pan American Network of Air Pollution Monitoring
Stations (REDPANAIRE) started its operation in June 1967. By the end of
that year it had eight stations, and 92 by the end of 1973, covering 25
cities in 13 countries. New others are being added continuously, and it
is expected that by 1975 the Network will include most of the Latin Ameri-
can countries and their main cities.
-------
454
In 1971 a First Report of the Network Results was published analyz-
ing around 40,000 data collected until December 1970. A preliminary ver-
sion of a Second Report, including 300,000 data collected until December
1973, was published this year, with the final version expected to be ready
by July. The analysis of the data shows that several cities in the Net-
work had high results, doubling and tripling in some cases the recommended
reference levels. This is especially true for settled dust, with the
highest values found in some of the stations of Mexico City, and in
Cordoba and Buenos Aires, both in Argentina.
Sao Paulo, Buenos Aires, Mexico, Havana and Rio de Janeiro have the
highest values for suspended dust, with figures over three times the ref-
erence level. Some cities, like Buenos Aires, show well-defined seasonal
variations, and a clear upward trend. Sulphur dioxide concentrations seem
to be the worst in Mexico City, Caracas, Santiago and Sao Paulo, with
•ontbly averages three and even four times higher than the reference level.
A comparison between the 1970 and 1973 reports shows important in-
crements for several cities and the clear need to initiate control pro-
grams designed to eliminate the damages already occurring and to prevent
the appearance of new problems.
Besides the REDPANAIRE measurements, other activities are being
developed by several countries. The most important can be found in
Mexico City, where a Philips automatic and computerized network of 20
stations, to measure five of the main pollutants, is under installation.
This will be an important experience and when under full operation should
give a good picture of the situation, guide the Mexican control program,
and serve as a training ground for Latin American personnel.
\
Mention can be also made of the measuring programs in Rio de
Janeiro and Sao Paulo. In Caracas, Venezuela, besides eight stations of
the REDPANAIRE there is a carbon monoxide detector in operation, and sev-
eral other instruments are being considered. Colombia, with the most ex-
tensive national network of REDPANAIRE type stations, a total of 25 dis-
tributed among six cities, intends to expand it and is expecting the
arrival of a carbon monoxide detector. Several other cities are planning
also to expand their measurement programs.
The main air pollution problems in the Latin American countries
«
to be industrial combustion, and in some cases combustion of solid
-------
455
wastes. Motor vehicles, even if at a low number on the average, with a
total of 9,108,400 in 1970, are fairly well concentrated in the main ci-
ties, which combined with plenty of sunshine causes a problem in some of
them, mainly Mexico City, Sao Paulo, Buenos Aires and Santiago. Due to
the hot climate, home heating is seldom used, with the exception of
Argentina and Chile, the southernmost countries. Burning of garbage in
apartment buildings is usually a problem, due to poor incinerators design,
but fortunately there are not too many of them, for the time being, even
in the largest cities. Municipal incineration is also a rarity, with
Buenos Aires as one of the exceptions.
Better understanding of the magnitude of the problem has moved some
of the Latin American Governments to act. An Onder-Secretariat for Envi-
ronmental Protection was established by Mexico in 1972, followed in 1973
by similar institutions in Argentina and Brazil. The Mexican Dnder-Secre-
tariat has a staff of over 450, more than 50% of them professionals. A
basic environmental protection law and a regulation on particles emission
have been promulgated, and some control is trying to be applied to the
150,000 factories in the country. However important these efforts, the
air pollution problem in Mexico City is perhaps one of the most serious
in Latin America.
Control programs are in operation in other countries. Mention can
be made of those in Argentina; Rio de Janeiro and Sao Paulo, Brazil;
Chile; Colombia; Peru, and Venezuela. Peru, the biggest fisher, and ex-
porter of fishmeal, in the world, with a catch of 12 million tons of an-
chovy in 1971, its record year, has been fairly successful in the control
of odors emitted by the factories surrounding its capital, Lima. Chile
and Venezuela have established institutions for air pollution studies, the
Institute of Occupational Health and Air Pollution Research in Chile, and
the Center for Research on Environmental Pollution in Venezuela. The last
one, with the technical support of PAHO and with an important financial
aid of the UNDP, is expected to expand air pollution programs in this
country.
The most important deterrent to the development of good air pollu-
tion programs in the Latin American countries is the scarcity of well-
trained personnel. PAHO is helping to correct this by giving technical
assistance in the preparation of programs, legislation, regulations, etc.,
-------
456
and for the solution of specific problems. Seventeen countries have bene-
fited from this activity. Assistance has been also given to the universi-
ties for the development of air pollution courses, like those in Argenti-
na, Brazil, Chile, Colombia, Mexico and Venezuela. Seven short courses on
air pollution have been presented in five countries, as well as lectures
at other 24 short courses and seminars. In 1968, a Latin American Seminar
on Air Pollution was held in Rio de Janeiro, with delegates from eight
countries. The information collected at this opportunity, and later pub-
lished, is still the best and more complete available. PAHO, through its
Fellowship Program, has also assigned scholarships and travel grants.
Research on air pollution is just starting in Latin America, and
all the work being done is of the applied type. Most of it deals with
measurements aimed at the establishment of the real magnitude of the prob-
lems. Besides the activities of REDPANAIRE and other measurement pro-
grams, work on concentration of specific pollutants has been done in at
least eight countries. Mexico and Peru have dealt with the efficiency of
the measuring methods.
Some research has been done on the health effects of air pollutants.
The Institute of Occupational Health and Air Pollution Research of Chile
has published at least three papers. One dealt with concentration of car-
boxihenoglobin in the blood of people exposed to carbon monoxide. The
second one makes a comparison of several clinical and physiological cha-
racteristics of a group of people from Santiago, the capital, with another
from a rural town, supposedly non-exposed. The last one made a study of
the effect on children of airborne pesticides. Mexico prepared one of
the most extensive studies, looking for a correlation between the results
of the 8EDPANAIRE stations and clinical and physiological observations
made on inhabitants of Mexico City. Even if these studies are still in
their preliminar stage, and are too few in number, they are showing the
right path for the Latin American scientists, and those from other devel-
oping countries, to get a real knowledge of their own problems.
-------
PANEL DISCUSSION
-------
459
SUMMARY OF THE DISCUSSION
SYROTA (France)
La table ronde debute par cinq exposes traitant tous de
techniques de mesure ou d'analyse de polluants atmospheriques.
Pour estimer sur un site donne quelle etait la repartition
des concentrations de polluants dans le passe, M. English
propose diverses techniques qui interpolent ou extrapolent les
rares mesures disponibles :
- 1'utilisation de modeles de qualite de 1'air;
- 1'utilisation des donnees de production;
- la construction de courbes d'egale concentration, par
interpolation lineaire des mesures disponibles;
- 1'emploi d'indicateurs de pollution, pour estimer a
partir des retombees connues d'un polluant celles d'un
autre.
M. Lauer presente le programme CHAMP de 1'EPA qui vise
a acquerir des donnees sur la qualite de 1'air, afin d'evaluer
1'influence des polluants sur la sante de cinq communautes
americaines (programme CHESS). II comprend 23 stations de mesure
situees dans cinq zones geographiques. Tout est continuellement
controle grace a un mini-ordinateur: calibrage, analyse et
validite des donnees, ainsi que fiabilite des stations et des
appareils de mesure.
M. Dams a evalue les niveaux d'empoussierage en Belgique.
Toutes les particules en suspension ont ete recueilles sur des
filtres de cellulose pendant un an dans 14 sites industriels,
urbains et ruraux, puis analyses par activation electronique,
absorption atomique et gravimetrie. Les sources d'emission
importantes peuvent en particulier etre localisees en comparant
les compositions chimiques elementaires des particules collectees
sur un ou plusieurs sites, mais sous differentes directions de
vent.
-------
460
M. Kneip montre que, grace a I1analyse et aux compositions
des aerosols recueillis a New-York, les elements les plus
caracteristiques a 1'etat de traces qui peuvent etre associes
a des sources urbaines sont: le plomb, le vanadium, le cuivre,le
sodium et le potassium. L1importance de leur absorption par
les new-yorkais est etudiee a partir des retombees de particules,
de leur dimension, de leur Vitesse de deposition et de leurs
traces dans les tissus humains.
M. Vouk fait le point sur la pollution de 1'air en Amerique
Latine. Depuis 1967, la Pan American Health Organisation {PAHO)
a etabli un reseau de mesure qui comprend actuellament 92
stations dans 25 villes de 13 pays et qui mesure le SO. (par
la methode a 1'eau oxygenee) et les particules en suspension
(par reflectometrie). Les principales sources de pollution sont
les combustions industrielles et le probierne le plus preoc-
cupant reste le manque de personnel qualifie. Des etudes
epidemiclogiques sont en cours.
La discussion qui s'ouvre alors porte principalement sur
les analyses et les resultats obtenus par 1'equipe de M. Kneip,
sur le coGt d'un reseau de mesure sur la validite des donnees
collectees par la PAHO et sur les mesures de particules de
sulfates qui semblent mieux correlees avec les effets sur la
sante que le SO2.
Mais le veritable objet de la table ronde - qui n'etait pas
de donner des details sur les systemes de mesure existants, mais
de discuter sur la maniere dont ces systemes peuvent etre utilises
pour mesurer 1*exposition des personnes aux polluants de I1air -
n'est aborde au cours du debat que par MM. Vouk et Sherwood.
Puisque les mesures effectuees a partir d'installations fixes
semblent ne pas fournir de donnees appropriees, parce que
d'une part la population surveillee se deplace souvent et
d'autre part les appareils ne sont pas forcement etablis dans les
endroits oil la population est exposee aux plus fortes pollutions,
le port de petits compteurs individuels de genre de ceux qui
-------
461
sont utilises pour evaluer la contamination radioactive des
personnes, est suggere. Mais cette methode assez pertinente
semble avoir rencontre certaines difficultes a I1EPA qui a
annule ses essais rfecenunent.
The discussion began with five statements dealing with
techniques for measuring OP analysing atmospheric pollutants.
As a means of estimating what the distribution of pollutant
concentrations have been in the past on a given site, Mr. English
suggested various techniques for interpolating or extrapolating
the few measurements actually available;
- use of models of air quality;
- use of production data;
construction of equal-concentration curves by linear
interpolation of the available measurements;
- use of pollution indicators to predict the immission
of a given pollutant on the basis of the known
immission of another.
Mr.Lauer introduced the EPA's CHAMP programme, the aim of which
was to obtain data on the quality of air so as to determine the
effects that pollutants had on the health of five American
communities (CHESS programme). There were 23 measuring stations
in five geographical areas and everything was monitored con-
tinuously by means of a mini-computer; calibration, analysis
and validity of data, as well as the reliability of the
measuring stations and equipment.
Mr. Dams said he had studied dust levels in Belgium. All
the particules in suspension were collected on cellulose
filters for a year at 14 industrial, urban and rural sites,
then analysed by electronic activation, atomic absorption and
gravimetry. The chief emission sources in particular could be
located by comparing the elementary chemical composition of
particles collected at one or more sites and correlated with
wind direction, which should be variable,
Mr. Kneip showed that, from the analysis and the composition
of the aerosols collected at New York, the most typical
elements in the form of traces which could be associated with
urban sources were lead, vanadium, copper, sodium and potassium.
The extent to which they were absorbed by New Yorkers was _
studied on the basis of the immission of the particles, their
size, rate of deposition and the trace quantities of them present
in human tissue.
-------
462
Mr. Vouk spoke of air pollution in Latin America. Since
196? the Pan American Health Organization (PAHO) had set up a
measuring network which currently comprised 92 stations in
25 towns in 13 countries. Sulphur dioxide was measured by the
hydrogen peroxide method, and the particles in suspension by
reflectometry. The main sources of pollution were -industrial
combustion processes, and the problem giving rise to most
concern was still the lack of qualified staff. Epidemiological
studies were currently in progress.
The main topics of the discussion which followed were the
analyses carried out by Mr. Kneip's team and the results obtained
on the coot of a measuring network, the validity of the data
gathered by the PAEO, and measurements of sulphate particles
which seemed better correlated with the effects on health than
was sulphur dioxide.
However, the true purpose of the panel discussion was not
to go into the details of existing measuring systems, but to
examine the ways in which such systems might be used for
measuring human exposure to air pollutants, and that had been
dealt with during the debate only by Mr. Vouk and Mr. Sherwood.
It appeared that measurements taken at fixed measuring stations
did not provide the appropriate information. That was due
partly to the fact that the measuring equipment was not neces-
sarily located at the points where the population was exposed to
the greatest pollution. The wearing of small personal counters
of the type used to assess the radioactive contamination of
human beings was suggested, but although the method was certainly
appropriate^ the EPA had apparently encountered certain problems
with it, and had recently suspended its testing.
DISCUSSION
SYROTA (France)
Cette table ronde a pour titret la mesure de 1*exposition
a la pollution atmospherique. II s'agit IS d'un probleme
particulierement important et difficile.
La mesure est a la base de toute etude scientifigue. II est
pourtant malaise de faire admettre cela, dans le domaine de la
pollution atmospherique, puisqu'il n'existe pas, a ce jour, de
-------
463
methods de reference, d'etalon si vous voulez, mime pour les
polluants les plus repandus, si bien que 1'utilisation des
connaissances acquises dans le monde sur les effets de differents
polluants sur la sante ne peut donner que des resultats contes-
tables. La mesure porte en effet sur des traces, des concen-
trations de 1'ordre de milliardiemes, et tout resultat est
influence par 1'effet d'autres polluants, qui, du reste, ne sont
pas toujours les memes.
II existe neanmoins, dupuis quelques annees, de nombreux
reseaux de mesure, permanents ou non, dans le roonde, probablement
plusieurs milliers d'appareils et les questions que 1'on peut se
poser consistent a savoir s'ils existent depuis suffisamment de
temps pour que I1on puisse en deduire d1importance des doses
absorbees par la population et s'ils donnent effectivement des
indications en rapport avec 1"exposition de la population.
Autrement dit, connatt-on la structure, la composition et I1 im-
plantation ideales d'un reseau de mesure destine a evaluer
1'exposition de la population?
II faut se souvenir que 1'on a affaire a une population
variee, qui comprend des enfants, des adultes, des vieillards,
chacun pouvant etre malade ou en bonne sante, travailler ou non,
se deplacer peu ou continuellement pendant une partie de la
journee. L'essentiel des mesures concerne actuellement I1air
ambiant et les atmospheres de travail.
Dans ce dernier cas, les problemes sont relativement simples,
les concentrations maximales admissibles sont souvent definies
et, dans 1'ensemble, peu contestees. Dans I1air ambiant, on
s'interesse a des concentrations qui sont souvent cent fois
plus faibles et on ne peut definir de teneurs limites puisque
le but affiche de tous les gouvernements est de proteger I1 ensemble
de la population sans exception. Rien n'empe'che en effet de
penser que la concentration en polluants susceptible de troubler
davantage la sante du citoyen marginal dont 1'etat sanitaire est
le plus degrade, soit particulierement faible.
-------
464
Dans 1'air ambiant, souvent, les reseaux de me sure mis en
place servant a indiquer surtout 1'evolution de la pollution dans
le temps, la oil elle est deja relativement elevee et a controler
les installations les plus polluantes. Les polluants mesures
sont alors les meilleurs traceurs de la pollution surveillee
comme par exemple le dioxyde de soufre pour la combustion ou le
monoxyde de carbone pour les vehicules, sans qu'ils soient
forcement ceux qui entrainent des effets nocifs sur la population.
D'ailleurs la lumiere n'est pas faite sur la nature, la
forme, 1'etat et les dimensions, des polluants qui, directement
ou indirectement entrainent des effets nocifs sur la sante.
Les phenomenes de synergic qui interviennent a coup sur et les
correlations entre la duree d*exposition et les risques encourus
sont encore mal connus. Ceci incite tout natur el lenient :a effectuer
des mesures tres nombreuses.
Les recherches deviennent alors complexes et mettent en jeu
des compilations enormes de donnees: il suffit de peu de capteurs
et de peu de semaines pour se retrouver a la tete de millions de
resultats de mesure.
Lorsgue I1on s'interesse i 1'exposition, il faut disposer
de capteurs dont on puisse theoriquement integrer les resultats
en fonction des deplacements de la population. C'est une entreprise,
ardue: comment, en effet, connaltre les effets de la pollution
due aux automobiles? Au milieu des rues ou la circulation est
intense, on trouve des dizaines de parties par million de pol-
luants, sur le trottoir on n'en trouve plus que la moitie, et
dans les rues plus calroes on n'en trouve plus que des traces.
Combien de temps passe-t-on dans les rues les plus polluees?
De m&ne, comment tenir compte de la pollution a 1'interieur des
locaux d1habitation? Certes, une partie de cette pollution provienfe
de 1'exterieur, mais s'interesse-t-on assez aux polluants emis
dans le logement, par exemple par les cuisinieres a gaz, par les
fours autonettoyants ou par les installations de chauffage
individualles?
-------
465
Voila, brievement trace, le decor, pas tres rose, dans
lequel se place cette table ronde. Ayant entendu les exposes
introductifs la discussion est maintenant ouverte.
The title of this round table conference is 'The measurement
of exposure to atmospheric pollution'. This is a particularly
important and difficult problem.
Measurement is at the basis of all scientific study. In
the field of atmospheric pollution^ however, it is not easy
to persuade people that this is so, since there is at present no
method of reference, no standard, even for the most common
pollutants, which means that the use of the knowledge acquired
throughout the world on the effects of various pollutants on
health can produce only questionable results. Measurements
are being made on mere traces, concentrations of the order of
one thousand millionth partt and any result is influenced by
the effect of other pollutants, which, moreover, are not
always the same.
Nevertheless, for some years now numerous measurement
networks, permanent or temporary, have been in operation in all
parts of the world, probably comprising several thousand
instruments, and the question may be asked whether they have been
in existence long enough to enable the volume of intakes
absorbed by the population to be deduced and whether they in
fact provide consistent information regarding the exposure of
the population. In other words, do we know the ideal structure,
make-up and layout in space of a measurement network intended to
evaluate the exposure of the population?
We must remember that we are dealing with a varied pop-
ulation, which includes children, adults and old persons, each
of whom may or may not be in good health, may or may not work,
may travel little or extensively during part of the day. At
the moment most measurements are carried out on the outside
ambient air and on the atmosphere of work places.
In the case of the latter, the problems are relatively
simple, since maximum allowable concentrations are generally
well-defined and on the whole give rise to little controversy.
As regards the outside ambient atmosphere, interest centres on
concentrations which are often a hundred times lower, and it
is not possible to establish threshold limits, since the avowed
aim of all governments is to protect the population as a whole,
without exception. One cannot help believing that the con-
centration of pollutants which is likely to affect the health of
citizens living on the fringe of society, whose physical con-
dition ia particularly poor, is especially low.
-------
466
In the case of the ambient air, the measurement networks
installed are often intended to indicate, above all, trends of
pollution over a certain period of time, in oases where it is
already relatively severe, and to monitor the most highly
polluting installations. The pollutants measured then are those
providing the best indication of the pollution monitored, such
as sulphur dioxide in the case of combustion or carbon monoxide
in that of motor vehicles, but not necessarily those which
have harmful effects on the population.
Moreover, much light remains to be shed on the nature, form,
state and dimensions of the pollutants which, directly or
indirectly, have harmful effects on health. The synergistic
processes which inevitably occur and the correlations between
the length of exposure and resulting hazards are as yet little
"known. The natural consequence is that large numbers of
measurements are taken.
Research then becomes complex and involves the compilation
of enormous amounts of data; in the space of a few weeks, a few
monitoring devices can produce millions of measurements.
In order to measure exposure, it is necessary to have
monitoring devices, the results of which may theoretically be
integrated according to population movements. This is a most
difficult task, for how can one determine the effects of pol-
lution from motor cars? In streets where traffic is heavy, pol-
lutants are present in dozens of parts per million, but at the
pavement this amount is halved and in quieter streets only
traces can be found. How much time do people spend in the most
polluted streets? Similarly, how can one allow for pollution
in dwellings? True, some of this pollution comes from outside,
but is enough attention paid to pollutants emitted inside the
dwelling, by gas cookers for example, self-cleaning ovens or
individual heating installations?
This, then, is the setting, and not a very rosy one, for
this round table conference. Sow that the introductory talks
are over, the discussion can be opened.
KNEIP (U.S.A.)
I think I can immediately see one distinguishing feature
between out studies and the studies that we have just heard
about from the panel members. All three that we have heard are
very large government supported programmes, ours is a study
being performed at a private university supported by the American
Petroleum Institute and the Electric Power Research Institute.
-------
467
Our approach has been to attempt to understand some of the com-
position relationships in the urban aerosol through a program which
could be handled by a small group in terms of number of samples and
analyses to be performed. Despite the fact that our group numbers
only five people, we have been able, by use of the weekly sampling
technique noted in our paper, to effectively define a number of
the relationships affecting both the concentration and composition
of the total suspended particulate matter in the New York City air.
While the larger programs may be highly productive, carefully
planned efforts with well-defined objectives may be carried out
quite well by such a small group as ours.
In addition to the relations reported in our paper, we have
recently accumulated enough data on tissue analysis to begin
analyzing and correlating these results. We found that along
with the reduction of lead in air, we have a reduction of lead
in lung tissue of the people in the city as represented by our
sample, and in fact we have a very strong linear correlation
between the two.
VOUK (W.H.O.)
As has already been pointed out, the purpose of this panel is
not to discuss various details of the existing air pollution
monitoring systems but rather to exchange views with the audience,
as to what extent monitoring systems, designed for other
purposes such as checking the iroplemen*ation of regulations or
standards, or for evaluating the effectiveness of control measures,
can be used for assessing human exposure. Obviously, levels of
air pollutants should not be identified with the true exposure
of an individual or population group. Since exposure has been
defined as the amount of a pollutant reaching the target (i.e.
man or, to be more precise, his respiratory system) under specified
conditions {i.e. concentration, frequency of exposure and its
duration), I hoped that questions from the audience would help
us to clarify some of the aspects of the design of air quality
monitoring systems for assessing human exposure. For instance,
how should such a system be designed, are fixed sampling stations
-------
468
adequate for measuring human exposure, or should we perhaps aim
at personal monitoring equipment, similar to that used sometimes
in industrial hygiene or, more frequently, in radiological
protection?
Reference was made to Mr. Haddad's paper, not because it
describes an air pollution monitoring system which has been
designed to measure human exposure, but because it was one of the
first attempts to collect, on an international basis, information
on basic air pollution data in a number of Latin American countries
that will allow a comparative study of the situation in the
cities included in the network and help to determine the trend of
the problem; also, because it will awaken greater interest in
the participating countries in air pollution problems and help
them to set up a well-organized administrative structure with
trained personnel for air pollution control. The Pan-American
Air Pollution Monitoring Network, described in Mr. Haddad's
paper, was initiated in 1967 and by December 1973 there were
93 monitoring stations located in 26 cities in 14 countries.
The air quality variables generally measured include settled dust
(monthly average), suspended dust (daily average) and sulfur
dioxide (daily average). This activity, initiated by the
Regional Office of WHO for the American and the Pan-American
Health Organization is a part of the WHO air pollution monitoring
programme which, in addition to Latin America, includes 13
Member States with three monitoring stations in each of them.
The Pan-American Air pollution Monitoring Network has not yet
been completely integrated with WHO activities in other countries
built is planned to do so in the near future.
BRAVO (Mexico)
I would like to comment as to the validity of the data that
has been collected by the PAHO network in Latin America.
Mr. Lauer has stressed the importance of valid measurement techniqxf
yet the PAHO network measured suspended particles using the
-------
469
reflectance method to report yg/m . This method is based on
standard smoke curve - developed in England and since we have no
heating by coal it does not apply to our estimation of total
suspended particles. The data obtained is not useful for control
strategies design.
VOUK (W.H.O.)
I would like to answer this question very briefly. Of course,
the measurement of suspended particles using the reflectance method,
developed originally in England and adopted by the OECD, is not
the ideal method to be applied in Latin American countries.
However, the purpose of the Pan-American Air Pollution Monitoring
Network has not been to obtain absolute values for suspended
particulate matter in air but to follow the trends, and for this
purpose I think the method is good enough. It may be mentioned
in passing that none of the existing routine methods for measuring
suspended particulate matter are satisfactory. High-volume sampler
results are also not comparable if obtained in different conditions
because the particle size distribution will greatly influence
the results. In addition, if we want to use the index of suspended
particulate matter for evaluating exposure/effect relationships for
man, a detailed chemical analysis of the sample is required in
addition to the particle size distribution. This is in fact the
major problem in all epidemiolgical studies performed so far.
When one of the WHO Expert Committees which met in 1972 in Geneva
was asked to propose criteria for particulated and sulfur dioxide
it found that a large proportion, almost 9O per cent, of the
available epidemiological information had tc be discarded because
the exposure measurements were not comparable. The criteria for
suspended particulates and sulfur dioxide which were published
in WHO Technical Reports Series No. 5O6 refer actually to the
smoke shade method, not because this method is recommended, but
because the only information that seemed to be valid was
obtained by measuring exposure to suspended particulate matter
using this method.
-------
470
Of course, we need also better methods for the design of
control strategies, but here again the choice has to be left
to the countries. The WHO has prepared and will publish soon
a Manual on Selected Methods for Sampling and Measuring Urban
Air Pollutants. This manual proposes for each pollutant several
methods which are applicable under certain conditions and also
for each pollutant a single "comparison" method, by which the
methods routinely used in different countries can be calibrated
and compared, at least approximately.
PHAM (France)
A combien peut-on es timer le coQt pecunier et le nombre de
personnel pour une organisation de roesures, coirane celle
presentee par M. English?
What io the east and the personnel requirements for a
measuring network as that presented by Dr. English?
ENGLISH (U.S.A.)
To give you a rough idea, consider the regional air pollution
study in St. Louis. That study is a five year program involving
25 very complete air monitoring stations which measure essentially
all pollutants and meteorological parameters. The study includes
gathering data not only from ground stations but also from air-
borne platforms such as 3 helicopters and 11 fixed-wing aircraft.
It includes a meteorological sounding network to determine the
three dimensional temperature and wind velocity distribution.
It also includes a modeling effort to compare the results which
are obtained through the measurements with those predicted
through a variety of air pollution meteorological models. The
price of this programme is roughly 25 million dollars, 5 million
dollars per year for five years.
-------
471
WESOLOWSKI (U.S.A.)
How close together must the monitoring stations be in an
area such as the Los Angeles Basin, where physical and Chemical
kinetics are an extremely important component of the pollution
problem, in order for the isopleth model you described to be
truly useful?
ENGLISH (U.S.A.)
Isopleths estimate concentrations of a pollutant at
locations where the pollutant has not been measured. These
estimates can be performed in many different ways which produce
identical results at the air monitoring stations. For example,
you can weigh the values obtained at each air monitoring station
using the inverse square of the distance from the location in
question to each station. This is a popular approach. Others
prefer to use a direct inverse relationship. Neither of these
approaches has any sound physical justification for its use.
The relationships which I have presented are based on the criteria
of simplicity. The simplest approach that reproduces the
multiple station data set at the station locations, and estimates
the values between stations is linear interpolation. In order
to estimate the error in this approach, we made some concentration
estimates for the Los Angeles area in which individual stations
were left out of the analysis. This produced changes that were
quite small, for example approximately 2O%.
SHERWOOD (U.K'.)
Occupational hygiene has demonstrated that there is seldom
any relationship between fixed position air samplers and exposure
of workers. Why therefore should fixed position samplers in
cities represent exposure of any of its inhabitants? An example
is quoted of the use of a personal S02 sampler in London that
-------
472
showed very wide variation of concentration depending on location
of the individual, and raised questions of the biological
significance of short-term peak concentrations. There is need
for better estimation of exposure of individuals in cities and
elsewhere.
ENGLISH (U.S.A.)
We agree quite strongly with the need for personal environ-
mental monitors. We have had a task force in the Environmental
Protection Agency looking into the feasibilities of such monitors
and they do appear to be feasible. One other comment, the
usefulness of past exposure estimates must be looked at in the
frame-work of the CHESS (Community Health Environmental Sur-
veillance System) programme. The CHESS programme selects
communities within which the epidemiological studies take place
to cover a wide range of pollutant concentrations. For example
if three socio-economicaly comparable communities are studied to
determine the health effects of a predominate pollutant, one
community will have high concentrations, one moderate and one
low concentrations. Within this kind of frame-work one can
make useful estimates of past exposures.
STEENSBERG (Denmark)
We should be cautious not to be caught by "measurement
euphoria".: Refined monitoring systems are needed in some parts
of the world also for research purposes. But our knowledge of
the health implications cannot follow the technical refinements
in measurement technique. We have at this symposium had two
contributions dealing with particulate sulfates which seem
to correlate better than SO2 and total suspended particulates
with health effects. Is WHO undertaking studies to see whether
this parameter should be included in more routine monitoring
alongside SO. and particulates?
-------
473
VOUK (W.H.O.)
Yes indeed. Among the methods which the World Health
organization is now examining for routine air pollution measure-
ments and for comparison purposes is also a method for
measuring sulfuric acid aerosols and sulphate particles. It
seems, however, that we shall not be able to recommend a
satisfactory method very soon and that more research will be
needed to develop an adequate method. One of our collaborating
institutions is at the moment studying this problem.
KNEIP (U.S.A.)
I would like to comment on this sulphate problem. There has
been a long-term study of the relationship of SO- and mortality
in New York City. It was originally published about 1972 by
Goldberg and Schimmel, and Schimmel is in the process of
republishing it with an additional set of data. I believe he
is now covering 5 to 7 years of mortality data still showing
a very strong correlation between variations in S02 and variations
in mortality from health records in New York City. However,
during the period of time involved, the S02 level in the city
has been reduced by a substantial amount, a factor of 4 or more,
but the mortality rate has not decreased. Thus, in all of these
cases when we are talking about monitoring and comparing to
exposures of the public, and we find a correlation, it should be
made absolutely clear that such a correlation does not constitute
a demonstration of a cause and effect relationship. If the SO-
correlation that Schimmel finds represented a cause and effect
relationship, the mortality should have decreased when the SO2
decreased. It did not.
The sulphate has to be looked at the same way. This is a
mathematical correlation, it does not prove that sulphate is
doing anything having to do with human health although it may
be an index or an indicator. Someone has to get behind it and
find out why the relationship exists.
-------
VERZEICHNIS
INDEX
INDEX
INDICE
INDEX
-------
XXIII
AASETH, J., 913
ADLER, M.W., 72
ALBERT, R.f 1167, 2O61, 2O68,
2069
ALESSIO, L., 1123, 1129, 1130,
1167
ALLART-DEMUL, C., 13O3
ALLEGRINI, M., 1697
ALLEN, J.R., 385, 397
ALTMAN, D.G., 289
ALTSHULER, B., 2061
ALTSHULLER, A.P., 21O9
ANDERSON, J., 1449, 1461, 1468,
1469
ANGERER, J., 1317, 1327, 1328
ARHIRII, M., 339
ARONOW, R., 1177
ARSAC, F., 603
AUBERT, M., 1613
BABCOCK, L.R., 2083
BACKHAUS, F., 2231
BAKER, F.D., 879
BARHAD, B., 333
BARQUET, A., 695
BARRATT, R.S., 1397, 1779
BARSAN, E.T., 1073
BARTH, D.S., 1875, 1877, 1921,
1922, 1925/1926, 1928, 193O,
1931, 1933, 1934, 1935, 1939,
194O, 1950, 1959, I960, 1962,
1963, 2073,
BASTENIER, H., 1303
BATES, D.V., 1007, 1967, 1978,
1979, 20O1, 2003,~2T6"2
BATTI, R., 1531
BEACONSFIELD, P., 2397
BECK, E.G., 1031, 1O4O, 1O41
BEITZ, L., 1417, 1430
BELCHER, R., 1779
BELL, A., 1101
BENARIO, M., 2169
BENINSON, D., 845, 1878, 1924,
1927, 1934, 1935, 1941, 1949,
BENSON, F.B., 423
BERGLUND, B., 119
BERGLUND, U., 119, 142
BERLIN, A., 552, 611, 629, 693,
859, 1087, 1100, 1238, 1959,
2003, 2185, 2254, 2257
BERLIN, M., 156, 491, 895, 1259
BERNA, M., 231
BERNER, A., 1729
BERNSTEIN, A.D., 1O5, 116, 117,
1O29, 2298
BERNSTEIN, D., 431
BIANCO, A., 1039
BIERSTEKER, K., 1881, 1924, 1926,
1933, 1934, 1943, 1957, 1958,
1959
BIGNON, J., 1189, 1196, 1197
BINDER, R.E., 669
BITTEL, R., 714, 1441, 1449,
1469
BLACKBURN, C.R.B., 63_
De BOECK, R., 1131
Den BOER, M.C., 1247
BOGDANOVIC, E., 2271
BONNAUD, G., 1169
BONNEFOUS, M., 51O
Underlined numerals denote authors
Arabic numerals denote participants in discussions
-------
XXIV
BORDAS, A., 1145
del BORGHI, M., 18O7
BORLAUG, N., 2397
de BORTOLI, M. , 1287
BOTZENHART, K. , 1757
BOUDENE, C., 6O3, 612
le BOUFFANT, L., 1645, 1651
BOUHUYS, A., 669, 675
BOUQOIAUX, J. , 1239, 1298
BOURBON, P., 269
BOURDEAU, Ph., 263 363, 366,
510
BOUVILLE, A., 1531
BRADEN, H., 485
BRAETTER, P., 2255, 23O1
BRAMAN, R.S., 117, 258, 1328,
1363, 1370, 1397, 1954
BRAVO A, H., 468, 2091, 216O,
2161
BREIDENBACH, A.H., 751
BRILLE, D., 269, 356, 667
BROCKHAOS, A., 781
BROMBERG, P.A., 1989
BROOKS, A.G.F., 77_
BRUAUX, P., 1131, 1143
CACCURI, S., 1823
CAGNETTI, P., 1451
CALANDRA, J.C., 772
CANTON, J.H., 1479, 1489
CAPURRO, P.U., 1579
CARNOW, B.W., 45, 313, 353, 266,
368, 588, 1129, TI67
CARPENTER, L., 1729
CARPI di RISMINI, A., 2397
CARSTENS, L.A., 385
CARSTENSEN, J., 969
CARTER, M.H., 1399
del CASTILHO, P., 2185
CASULA, D., 1693
CERNIK, A.A., 1207, 1221, 1237,
2254
CERQUIGLINI-MONTERIOLO, S., 13O1
1383
CHAMBERS, P., 1O4, 259, 397, 51O,
7O2, 2O45
CHANTEUR, J.f 2386
CHAPMAN, R., 193, 2O7, 2O8, 21O,
645, 658
CHARLTON, J., 112O
CHATTOPADHYAY, A., 1685
CIALELLA, N., 2225
BRUCH, J., 781, 791, 1O4O, 12O6 CICOLELLA/ A>, 1661
de BRUIN, A., 259, 911, 2O59
BDCHET, J.P., 631, 887, 2185
BULCRAIG, W.R. , 2O9
BURGER, E.J.jr., 145, 155
156, 157
CIGNA ROSSI, L., 1451
CLAEYS-THOREAU, F., 1131
CLAUDE, J.R., 397
CLAYTON, J.W.Jr., 371, 383, 383
CLEMENTS, G.F., 260, 1O28, 1451,
2298
COFFIELD, T.H., 858
BUSH, B., 879, 885, 911, 1339
BUSTUEVA, K., 1OO9
BUTLER, G.C., 1431, 1884, 193O
1937, 1948, 1955, 1956, 1978
BUXTON, R. St.J., 1113
BYRNE, A.R., 245, 258, 259, 26O,COLOMBINI' M' ' 159
1370
DOLMAN, R., 485
-------
XXV
COLUCCI, A.V., 1043
COOPER, W.C., 555, 568, 569.
1196
COTE, R.W., 47_
CROCKER, K., 2068
CROSSMANN, G., 14O7
CRUZ, R., 1685
CUCU, M., 333
DAHL, R., 2231
DAMS, E., 513, 1430
DAMS, R., 409
DANIEL, H., 1645, 1651
DANIEL, J.W., 877, 9O2, 1O41
DANIELSON, L., 116, 885
DAVID, O., 588, 1549
DAVIDOW, B., 545
DAVIES, J.E., 695
DAVOUST, P., 2384
DEAN, G., 190, 643
DEHNEN, W., 781
DELEANU,H., 1583
DELCARTE, E., 1675
DELVES, H.T., 2215. 2258
DEMUYNCK, M., 4O9
DENNIS, C.A.R., 1O29, 1543,
1953
DEPAUS, R., 1341
von DEPKA, J., 5O6
DERWENT, R.G., 1669
DESBORDES, J., 1199
DEVOTO, G., 1693
DIEHL, J.F., 958, 1121
DIETERICH, B.H., 2451
DIETL, F., 1853
DI FERRANTE, E., 1956
DJURIC, D., 1829
DOBIN, D.D., 1223
DOBRYSZYCKA, W., 685
DOLGNER, R., 279
DONALDSON, W.T., 1399, 14O5,
1406
DONNIER, B., 1613
DONZELLI, A., 231
DUBOIS, L., 1331
DUGANDZIC, M. 2285
DUMONT, M., 2231
DUNCAN, K.P., 2175
DUNCAN, L.J., 1241
DUPUIS, P.J., 491, 1196, 1314
EDWARDS. H.W., 1277, 1285, 1286
EFTHYMIOU, M.L., 1789
EGELS, W., 1407
EISENBUD, M., 431
ENGLISH, T.D., 4O1, 47O, 471,
472
EPSTEIN, S.S., 552, 569, 749,
814, 112O, 2367, 2382, 2383,
2384, 2385, 2386
van ESCH, G.J., 1O17
ESPINOSA, M.E., 2O91
EVENDIJK, J.E., 1351
FACCHETTI, S., 1287, 1298, 1299
FAGNIART, E., 1675
FAIRWEATHER, F.A., 1113
FALK, H.L., 2331, 2348, 2349,
2350
FAVRETTO, L., 1511
FAVRETTO GABRIELLI, L. , 1511
FERRAIOLO, E.G., 18O7
-------
XXVI
FERRARI-BRAVO, P., 1797
FINE, P.R., 1223
FINKLEA, J., 193, 645
FISCHER, A.B., 1O31
FISHBEIN, L,, 725, 749
FLOREY, C., 289
FONDIMARE, A., 1199, 1206
FOURNIER, E., 1433, 1439, 1789
FREDERIKSON, M., 1959
FREEMAN, G., 685, 833, 844,
1007
FRENCH, J. , 193, 645
FRIBERG, L., 23O7, 2315, 2316
FRIEDMANN, J., 5O7, 51O, 511
FRIEDRICHS, K.H., 715, 723,
724
FOMAROLA, G., 18O7
FURIOSI, N.J., 833
GADDO, P.P., 1287
GAFFEY, W.R., 555
GAGE, J.C., 895
GAGLIONE, P., 1287
CARDI, R., 18O7
GARDNER, D.E., 7O5, 713, 714,
GARIBALDI, P., 1287
GARNIER, A., 1441
GENT, M., 1263
GHELBERG, N.W., 1145, 1583
GHETTI, P.P., 1957
GIANANI, G.( 1697
GIBBS, G.W., 1197, 12O6, 2271,
2296, 2297, 23O1, 23O2
GIBSON, R.J.W., 289
von GIERKE, H., 1249
GIOVANNINIt I., 1797
CLAUDE, P., 1341
GLOBUS, G.f 507
GODIN, J., 6O3
GOERKE, W., 37
GOLDBERG, A.M., 793, 803
GOLDBERG, H., 193
GOLDSMITH, J.R., 62, 155, 189,
382, 585, 675, 1889, 1923,
1935, 1936, 1942, 1945, 1950,
1956, 2156, 2165, 2179,
GOLDSTEIN, I., 1275
GONO, E., 591
GOOTJIS, P., 1316, 1951
GRAB, B., 339
GRADISKI, D., 1631, 1661
de GRAEVE, J., 523
GRAOVAC-LEPOSAVICr L., 1829
GRASSO, C., 861, 878
GREENLAND, R.D., 805
GREVE, P.A., 1479
GRIECO, A., 231
GRIFFIN, H.E., 23O5
GROLL-KNAPP,E., 989
GROS, R., 2225
GRUENER, N., 1O67
GRUNSPAN, M.J., 773, 20O3
GUINEE, V.F., 545, 551, 552,
553, 1166
HAAG, A.t 1317
HAASE, J., 1417
HADDAD, R., 451
HAGEDORN-GOETZ, H., 2231
HAIDER, M., 475, 989, 999, 2393
HARDWICK, D.F., 961
HARKE, H.P., 1327, 1773
HARRISON, P.R., 1741
HARRISON, R.M., 1111, 1285, 1349,
1370, 1405, 1783
-------
XXVH
HAYES, C., 645
HAZUCHA, M., 1979, 20O1
HEBBELINCK, D., 13O3
HEM, B., 1189
HENDERSON, P.Th., 2O47
HENIN, J.P., 1651
HERNBERG, S., 568, 692, 1129,
1142, 2395
HERTZ, M.B., 1763
HICKEY, N., 658
HILLERY, P.J., T.
HILPERT, K., 2231
HINE, C.H., 207, 26O, 1O4O,
1130, 1937, 2029, 2298
HINTON, D.O., 1769
HISLOP, J., 959
HO, M.T., 1519
HOFMAN, B., 1Q17
HOGGER, D., 1247, 2157
HOLL, K., 612
HOLLAND, W.W. , 39_, 45
HOLM, S., 895, 902
HOLMQVIST, I, 613, 629
HORIE, Y., 2143
HOSEIN, H.R.f 669
HOUCK, C.L., 879
HOWER, J., 591, 6O1
HUETER, F.G., 261, 352, 363,
364
HUTCHINSON, T.C., 1685
HUTH, F., 715
HUUNAN-SEPPALA, A., 2263
IMPENS, R., 1675
IOVENITTI, L., 1797
IRWIG, L., 289, 36O
IWANKIEWICZ, S., 685
IZMEROV, N.F., 24O9
JACKSON, D.L., 161, 176, 177.
178, 1956
JACOBSEN, M., 89, 211, 228, 229,
1478, 1502, 2365
JACYSZYN, K., 685
JAMIN, P., 523
JANSEN, G., 513, 999, 2044
JEANMAIRE, L., 2225, 2252, 2256
JERVIS, R.E., 1685
JOHNSON, D.L., 1363
JOOSTING, P.E., 76, 2OO5, 2O29
JOST, D., 2115
JUHOS, L., 833
KAMINSKI, E.J., 551, 8O3, 1168
KARCHER, W., 1341, 1349
KARHAUSEN, L., 367
O'KEEFFE, A.E., 2109, 2164, 2166
2168, 2169, 2172
KEITZ, E.L., 1241, 1247, 1248
KELLER, M.D. , 4_7_
KEVANY, J., 89
KILPIO, J.O., 2263
KIRCHMANN, R., 1675
KJELLSTROM, T., 62, 1221, 1939,
2197, 2252, 2254, 2257, 2299,
2303, 2328, 2365, 2383
KLAHRE, P- / 2231
KLEINMAN, M., 431
KNAUTH, P., 2O31, 2043, 2O44
KNEIP, T.J., 431, 466, 473
KNELSON, J.H., 181, 189, 190,
361, 973
KOPPLE, J.D., 847
KORICANAC, Z., 2285
KOSTA, L., 245
KOTLAREK-HAUS, S., 685f 692, 693
-------
xxvm
KRACKE, W., 1853 LLOYD, W.J., 1O85
KREBS, H., 239J7 L0/ Fa-Chun, 879
V3?n?r ??f?K' F*W" 959' I028' van LOON' J" 1685r 1849
J.UD:>, ^Jo2
KEEOZER, W.. 601. 1853, 2045, l°RKE' D"^i' 83°' 831
2299 LOSER, E., 817
KUMMER, J., 1303, 1314, 1315, LOVE, G.J., 3O1, 645
1316, 1339 LOWE, A.C., 2O91
KOMPFM.J., 339, 363, 365, 1957 LVNAM, D.R., 543, 791
LACOURLY, G., 1441
LAFONTAINE, A., 382, 1131, 1942
2437
LAMBO, T.A., 11
LANESE, R.R. , 4J7
LANG, R., 659
LANGER, H., 1757
LANZOLA, E., 1697
LAO, R.C., 1331, 1339, 134O,
1468
LAOER, G., 423, 2156, 2163
LAUWERYS, R., 542, 568, 631,
831, 887, 1238, 2185
LEEDER, S.R., 63, 76, 1111
LEFEVRE, M.J., 988, 2O29
LEGRAND, M., 1131
LEHNERT, G., 1317
LEHTO, V.P., 1007, 2315
LELLOUCH, J., 269
LENGHEL, I., 1583
LEUNG, S., 1248
LEVERE, T.E., 493, 506
LEVY, D., 1263, 1275, 1276
LEVY, E.A., 1797
LIND, B., 2197
LINDVALL, T., 119, 1954
LINKMAN, L., 2197
MAGADUR, J.L., 1631, 1661
»MAGE, D., 176, 2O7, 1275, 2068
2156, 2158, 2162, 2164, 2165,
2168, 2170, 2174, 2176, 2178,
2179, 2181, 2315, 2364
MAGE, D.T., 2O97
MAGI, F., 857
MAGNAVAL, R., 1531
MAHEU, R, , r?
MALONE, D,W., 1569
MANOJLOVIC, N., 1O31
MARCUS, A.H., 15O5
MARQUARDT, H., 16O7
MARTIN, A.E., 1113, 112O, 1121,
1122, 2305
MARTIN, J., 973
MARTIN, J.C., 1645, 1651
MATERNE, D., 631
McCABE, E.B., 1168
MCDONALD, G.C., 1491
McGUIRE, J.M., 1399
MCNEIL, J.L., 571, 584, 586,
588, 589
McNESBY, J.R., 1371, 1383
MEININGER, J., 6O3
van MEIRHAEGE, A., 2044
MENENTI, M., 1797
MERIAN, E., 929, 1396, 2175
-------
XXIX
MERLUZZI, F., 231
MICHAELSON, I.A., 805
MILIC, S., 1829, 2285
MILOVANOVIC, Lf,, 2285
MIRE, B.f 973
MITCHELL, C.A., 669
MITCHELL, R.I., 47, 62
MOKEMATKENGUEMBA, G., 178,
1941, 1951
MOLDOVAN, N., 1583
MOLLARET, P., 1938
MONCELON, B., 1661
MONCHAUX, G., 1189
MONKMAN, J.L., 1O15, 1331
MOONEY, T.F., 1637
MOORE, M.R., 535, 1171, 2252
MOORE, W., 751
MOREAU, M., 1315
MORGADE, C., 695
MORGAN, G.B., 2O73
MORRESI, N., 231
MORRIS, S.C., 677, 683
MOSE, J.R., 1617
MOULE, Y., 967
MRAK, E.M., 1965
MURAYAMA, H., 91
MUSSENDEN, R., 833
NAGDA, N.L., 2O83
NANGNIOT, P., 1675
NEEDLEMAN, H.L., 584, 1155,
1167, 1168, 1169, 1936, 2433
NEUBERGER, M., 989
NEWHOUSE, M., 383, 1263, 1953
NEWILL, V.A., 161
NIEUWSTRATEN, N., 1351
NORBACK, D.H., 385
NORSETH, T., 913
NOTTEN, W.R.F., 2O47, 2059, 2O6O
NURNBERG, H.W., 2231
OLEKSYK, E,, 685
OLOFFS, P.C., 702, 723, 75O,
961, 1065, 1953, 235O
OMENN, G.S., 1563, 1952
OREL, J.V., 2177
ORIOL, P. , 269
OTT, W.R., 2097
OUW, K.H., 11O1
PACIGA, J., 1685
PACKHAM, R.F., 1468
PATTI, F., 2225
PEAT, J.K., 62
PECORA, L., 1823
PELECH, L., 279
PERROTEY, J., 1199
PERRY, R., 1285, 1328, 1385,
1405, 1783
PETERS, R., 2397
PETERSON, R.W. , 3_
PFANNHAUSER, W., 258, 858, 1328,
1719, 2253
PHAM, Q.T., 2O8, 470, 973, 988
PHILP, J,M., 235O
PICHE, L., 1314, 1468
PITTWELL, L.R., 23O3
PIVA, C., 1789
POTT, F., 715
POWELL, W. , 833_
-------
XXX
PRINZ, B., 591, 1471, 1489
PTASNIK, J.A., 571
QUAGLIARDI, A., 1287
QUINOT, E.f 1519
RABINOWITZ, M., 847
RAFFONELLI, A. , 695
RAINSBURY, R., 2397
RALL, D., 37
RAMACIOTTI, D., 659, 2316
REAY, J.S.S., 1669
RECHT, P., 157, 1894, 1928,
1938, 1946, 1961, 2641
REEVES, A., 724, 1637, 2385
RETHFELD, H., 14O7, 1416
REUSMANN, G., 591
REUTER, L., 1O66
REY, P., 659, 667
ROBERTS, T.M., 1685
ROELS, H., 631, 887, 1237,
2185
ROGGI, C., 1697
RONDIA, D., 523, 535, 1339
ROOSKEN, A.A., 1351
ROOTS, L.M., 1113
van ROOYEN, G.I., 2382
ROSCA, Gti., 1073
ROSCA, S., 1073
ROSIVAL, L., 877, 1187, 1237
ROSSANO, A.T., 2083
ROSSI, A», 1823
RUBIN, R.J., 903, 911, 912
RUDEN, H., 1757
RUDOLF, W., 2115
RUTENFRANZ, J., 2O31
RUTZEL, H., 2231
RYLANDER, R., 137, 477, 484,
667
SABATINI, G.C., 1697
SADOUL, P., 973
SALAZAR, S., 2O91
SANOTSKY, I.V., 1OO9, 2349,
2409, 2433
SANSONI, B. , 1853
SANTARONI, G.P.f 1O30, 1451
SANTOLUCITO, J.A. , 7O2, 1O51,
2387, 2394
SAUERHOFF, M.W., 8O5
SAVING, A., 931, 1749
SAYERS, M.H.P., 12O7
SCASSELLATI-SFORZOLINI G., 931,
1749
SCHALLER, K.H., 1O87
SCHNEIDER, T., 1954, 2O71, 2155,
2158, 2162, 2166, 2168, 2172,
2174, 2175, 2178, 2182
SCHILLING, R.S.F., 669
SCHLATTER, C., 723
SCHLIPKOTER, H.W., 19O, 228,
369, 521, 667, 723, 771,
2328
SCHMIDT, P., 279, 354
SCHOENBERG, J.B., 669
SCHUCK, E.A., 2073, 2159, 2161,
2164, 2171, 2173, 2176, 2177,
2179, 2180, 2181, 2182
SCHULLER, P.L., 1017
SCHULZ, C.O., 9O3
SCHWING, R.C., 1491, 15O2, 1503,
2178, 2181
SCOPPA, M., 958
SEBASTIEN, P., 1189
SECCHI, G.C., 1123
-------
XXXI
SEIFERT, D., 1407
SERWER, D., 2383
EL-SEWEFY, A.Z., 1589
SHAHEEN, H., 1589
SHAMS El-DEEN, A., 1589
SHAPIRO, I.M., 1155
SHAPIRO, M.A., 677
SHERWOOD, R.J., 471, 134O, 1954
SHUVAL, H.I., 1067
SILBERGELD, E.K., 793, 814, 1O51
SILVERMAN, A.P., 16O1
SKJAERASEN, J., 1233
SLATER, D.H., 1783
SHEETS, J., 1087, 2185
SMIDT, U., 1557
SMITH, R.G., 1637
SMITH, G., 1729
SORENSEN, S., 137, 477
STANKOVIC, B., 2285
STANKOVIC, M., 2285
STANKOWSKA, K., 685
STANLEY, R., 1729
STARA, J.F., 714, 751, 771, 772,
813
STEELE, T.D., 1954, 1961, 217O
STEENSBERG, J., 472, 1006, I960
2432
STEPANEK, V., 1473, 1478, 15O3
STEPHAN, W.I., 1779
STERN, A.C., 2143
STEVENS, L., 1O01
STEWART, H.N.M., 1669
STIDL, H.G., 989
STOEPPLER, M., 2231, 2259
STOOFF, W., 1479
STOPPS, G.J., 229
STUIK, E.J., 537
STUPFEL, M., 89, 770, 844,
1275, 1625, 1963, 2O43
SUESS, M.J., 339
SULAIMAN, A.B., 657, 683, 1952
SUTHERLAND, L.C., 484, 485, 491
SWYNGEDOUW, I., 1131
SYROTA, M.J., 399, 457, 459,
462
SZADKOWSKI, D., 1841, 2183, 2249,
2256, 2259
SZPERLINSKI, Z., 1711
TAKABATAKE, E., 2197
TATI, M., 2251
TAYOT, J., 1199
TER HAAR, G., 8O3, 1177
THOMAS, R.S., 1331
THOMAS, T.J., 1569
THOMPSON, J.M., 1779
TOMATIS, L., 1053, 2317, 2328,
2329
TOMPKINS, E., 2197
den TONKELAAR, E.M., 1O17, 1O28,
1029, 1030
TRAKOWSKI, A.C., 2455
TREMOLIERES, J., 19O2, 1921,
1927, 1944, 1950
TRUFFERT, L., 2255, 23OO
TSUCHIYA, K., 2197, 2351, 2364,
2365
TULLIEZ, J., 921, 929
TUMASONIS, C.E., 879
TWIBELL, J.D., 1385, 1397, 1398
TYTUN, A., 545
VALENTA, P., 1416, 2231
VALJAREVIC, V., 1829
-------
XXXII
VANINI, G., 1289 YORDANOV, D., 1815
VEIL, S., 1, 25
VERBERK, M.M., 20O5 ZAPHIROPOULOS, M., 1961
VOINIER, B., 659 ZEEUW M 1947
VOUK, V.B., 467, 469, 473, 7!Tpv,Vn 'v "?47
1912, 1925, 1930, 1931, 1947 ZELENKO, V., 247
VOORI, E., 2263, 2296 ZIELHUIS, R.L., 175,^364,^.537,
1168, 2O02, 2060, 2348, 2384
WALDBOTT, G.L., 1575 ZUNIC, R., 2386
WALLER, R.E., 77_, 89, 90 ZWIERS, J.H.L. , 1017_
WANNAG, A., 1233, 1237, 1238
WARNER, P.O., 1O01, 10O6, 1OO7,
1008
WASSERMANN, D., 1O53
WASSERMANN, M., 1O3, 156, 396,
683, 7O2, 713, 877, 1O53,
1065, 1066, 1439, 1955
WATANABE, H., 9^, 1O4
WEAVER, N.R., 369
WEBER, O., 260, 2261, 2295, 2297,
2299, 23O1, 2302, 2303
WEICHERT, N., 1417
WEIR, F.W., 1989, 2O01, 20O2,
20O3, 2004
WEISS, B., 2415, 2432, 2433
WESOLOWSKI, J.J., 471, 1729
WETHERILL, G.W., 847, 857, 858,
859, 1298
WILCOX, S.L., 1241
WILLIAMS, H., 16O1
WILLIAMS, M.K., 551
WILLIAMS, R., 47_
WISSMATH, P., 1853
WOIDICH, H., 1719
WOOLCOCK, A.J., 63_
WORTH, G., 1557
YANIV, S.L., 1249, 126O
-------
SALES OFFICES
Belgique — BelgiS
Ireland
United States of America
Monireur beige - Belgisch Staatsblad Stationery Office
Rue de Louvain 40-42 - Leuvenseweg 40-42
IDOOBruxelles - lOOOBrussel Beggar s Bush
Tel. 5120026 Dublin 4
CCP 000-2005502-27 - Tel 688433
Postrekening 000-2005502-27
European Community Information Service
2100 M Street, N W
Suite 707
Washington, D.C. 20037
Tel 29651 31
Sous-depot — Agentschap
Librairie europeenne — Europese Boekhandei
Rue de la Loi 244 - Wetstraat 244
1040 Bru xelles - 1040 Bfussel
Danmark
Italia
Libreria dello Stato
Piazza G Verdi 10
00198 Roma - Tel. (6) 8508
CCP 1/2640
Telex 62 008
Aganzie
00187 Roma - Via XX Settembre
(Palazzo Ministero
del Tesoro)
20121 Milano - Galleria
Vittorio Emanuele
3 - Tel. 8064 06
Schweiz — Suisse — Svizzera
Librairie Pa/of
6. rueGrenus
1211 Geneve
Tel. 31 89 50
CCP 12-236 Geneve
J H. Schultz — Bogttanael
Montergadeig
1ll6KobenfiavnK
QiroHonto 1195
Tel 1-1195
Deutschland (BR)
Ver.'ag Bundesanzeiger
5Kom1 - BreiteStrafle - Postfach 108006
Tei (0221)2103-18
(Fernschreicer Anzeiger Bonn
08882595)
PostscnecKkonto83400 Koln
Grand-Duche
de Luxembourg
Office des publications officieiles des Com-
munautes europ6ennes
5, rue du Commerce
Bofte postale 1003 - Luxembourg
Tel 490081 -CCP 191-90
Compte courant bancaire:
BIL 8-109/6003/300
Nederland
Staatsctrukkerii- en uitgeveriioedritf
Chnstollel Plantijnstraal.
S-Gravenhage
Tel. (070)814511
Postgiro425300
Sverige
Librairie C E. Fritze
2. Fredsgatan
Stockholm 16
Post G iro 193. Ban k G i ro 73/4015
Espana
Librena MunOi-Prensa
Castell637
Madrid 1
Tel. 2754655
Other countries
France
Service de vente en France des publications
ties Commjnaut6s europevnnes
Journal otliciel
26. rue Desaix
75732Pans- CedexIS
Tel. (1) 5786139 - CCP Paris 23-96
United Kingdom
H.M. Stationery Office
P.O. Box 569
London SE1 9NH
Tel. 01-9286977. ext. 365
Office lor Official Publications of the
European Communities
5, rue du Commerce
Boite postale 1003 - Luxembourg
Tel. 490081 -CCP 191-90
Compte courant bancaire:
BIL8-109/6003/300
------- |