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

-------