Environmental epidemiology = L'éco-épidémiologie [full issue]

WORLD HEALTH - l QUAR I ERLY · RAPPORT TRIMES I RIEL DE SANITAIRES MONDIALES tesponse/ Reponse ENVIRONMENTAL EPIDEMIOLOGY L'ECO-EPIDEMIOLOGIE Vol. 43, No. 3, 1990 World Health Organization Organisation mondiale de la Sante Gen eve Dose The World Health Organization i~ a spccialiLed agency of the United Nation\ with primary re\pOnsibility for international health matters and public health. Through thi~ or- ganiLation, which was created in 1948, the health profe-.,:,,ionals of -..<.une 160 countrie:,, exchange their knowledge and experience with the aim of making pO!-.:,,iblc the attain- ment by all citizen<s of the world by the year 2000 of a level of health that will pem1it them to lead a socially and e<.:onomically productive life. By means of direct technical ,ooperation with its Member States, and by :,,timulating "uch cooperation among them, WHO promotes the dcvclopmem of compn::hen:,,i\-e health services, the prevention and control of <lisea!i.es, the improvement of environmental con- dition!<., the development of health manpower. the coordination and development of bio- medical and healtl\ services research, and the planning and implementation of health pro- gramme ..... These broad fields of endeavour cncompa..,..., a wide variety of activitie~. Mtch as devel- oping systcm<s of primary health care that reach the whole porulation of Member coun- trie'I; promoting the health of mother:,, and children: combating malnutrition: controlling malaria and other- communicable diseases including tuherculrn,is and leprosy; having achieved the eradication of smallpox, promoting mass immunization again'il a number of other preventable disea\es; improving mental health: providing safe water :,,upplics; and training health per.,onnel of all categories. Progre\s towards better health throughout the world abo demands international coop- eration in ~uch matters as establishing international 'Standard:,, for biological suh~tances. resticides and pharmaceutieal5; fomlulating environmental health criteria; recommending international nlmproprietary names for <lrug!i; administering the International Heallh Re- gulations; revbing the International Cla:,,siftcation of Di<sea!-.es, Injuries, and Cause:-. of Death; and collecting and disseminating health 'ltatistical infonnation. Further information on many aspects of WHO"s work 1s presented in the Organin- tion 's publications. The World· health statistics quarterly replaces (s.ince 1978) the monthly 'world health statistics report (published s.ince 1967) and its forerunner the Epid~mio/ogical and 1·ital statistic·j· report ·(puhlished !i.ince 1947). It deals with the detailed analysis of selected .health topics of current interest. Starting with Vol. 41 (1988), the Quarterly contains articles in either French or English with a summary in both languages. Annual sub:,,cription .................................. . Price per copy .................................... . Sw. fr. 90.- Sw. fr. 24.- Material from the Quarterly may he reproduced providing due acknowledgement is made. L'Organbation mondialc de la Sante (OMS), crCC·c en 1948. e'it une in!ititution ~pCCia- li<..C!c:: de:-. Nations Unies a qui incombe, sur le plan international. la re\pOn~ahilitC princi- palc en matitre de question~ sanitaire'I et <le same puhlique. Au ... ein de:: I 'OMS, le:,, pro- fe'isionnel'i de la santC de quclque 160 pay .... C!changent des connai<ssam.:e'i et de!<. donnC-c:,, <l'exp<:rience en vue de faire acc<:der <l'ici l'an 2000 tous lcs habitanl'S Ju rnondc :'L un niveau de santC' qui leur pem1ette de mcner unc vie ..,ocialement et economiqucment pro- ductive. Griicc a la cooperation technique qu'cllc praLique avec !<.e\ Etat!<. Mcmbre\ ou qu"elle :,,timule entrc eux, l'OMS s'emploie i1 promom·oir la mi:,,e sur pied de scrvicc::s <le 'Santi complel'i. la prevention et l'endiguc::ment de..., maladies, l'amClioration de l'environne- 111ent, le devcloppemem de .... per:,,onnels de santC, la coordination et le progrCs de la re- cherche hiomCdicale et <le la re<.:herche !i.Ur lcs services <le sante, ain!-.i 4ue la planifica- tion et rexecution des programme'i de:: ~ante. Le vaste domainc oll 'I 'exerce I 'action de I 'OMS compOL1c:: des activite!<.. trCs diverse..,: developpemcnt des soins de !i.ante primaircs pour que toute .... le:,, populations pui'isent y avoir accC:s; promotion de la sante matemelle et infantile: la luttc contre la malnutrition: lutte contre le paludisme et d'autre:,, maladies transmis~ibles. donL la tuherculosc et la ICpre; I 'eradication de la variolc· Ctant re!ali'l<!e. promotion de la vaccination de ma .... -..e contre un certain nombrc <l'autres maladies evitahle!<.; amC-lioration de la -..ante mentalc: approvbionnemcnt en c::au ~aine; formation de personncb de sant<: de Loutes categories. II e!<.t d'autre'I secteurs encore oll. unc cooperation intemationale s'impose pour assurer un meilleur <!tat de same a travers le mondc:: et l"OMS collabore notamment aux tfi.che:,, \uivantes: C!tahlissement d ·etalons intemationaux pour le~ produits biologique!-.. les pc:,,ti- cides et les prCparations phannaceutiques: fonnulation de critC!res de .... aluhritC de I 'envi- ronncmcnt, recommandations relatives aux denominations communes intcrnationales pour les <sub\lances phannaccutique~; application du Reglcmcnt sanitaire international; rCvi- 'iion de la Cla:,,sification intcrnationale des maladies, traumatismes et <..:au<..e:-. de decCs: rassemblcmcnt et diffusion d'infom1ations statistlques sur la !<.ante. On trouvera dan .... le!i. publications de I 'OMS de plu!i. amplcs rcnscignemenl-" sur de nombreux a<spects de:-. travaux de I 'Organbation. Le Rapporr trim<'.\·tril'I dt.' statistiqties sanitaires mondiah'.\ remplace (<lepui<s 1978) le Ramwrt d<' .\lati.\liques sanitaires mondiah'.\ (publiC depuis 1967) et :,,on precurseur le Rupport i11idimiologique et demograpltiqu£' (publiC <lepuis 1947). n prCscnte de'i analy:,,e!<. detaillCcs sur des !-.Ujet:,, !i.p6cifiqucs d'interCt courant. A compeer du Vol. 41 ( 1988 ), le Trimt.•strh•I presente des article:-. originaux en fran<;ais ou en anglais, accompa- gne:-. d"un resume <lans le!) deux langues. Prix de I 'abonnement annuel ............................. . Fr.,. 90.- Le nurnero ....................................... . Fr. s. 24.- La reproduction d'extraits du Trimestriel est autorisee, sous re- serve d' indication de la source. IX ISSN 0043 - 8510 PRINTED IN SWITZERLAND 90/8603 - Atar S.A., Geneva 5000 Cover design: Gilbert Auberson * T M F N Symbols used in tables Preliminary, approximate or estimated data. Data not available. Nil or magnitude negligible. Category not applicable. Total. Male. Female. Absolute numbers. © World Health Organization 1990 The designations employed and the presentation of mate- rial in this publication do not imply the expression of any opinion whatsoever on the part of the Secretariat of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. The statistics which relate to the Federal Republic of Germany and the German Democratic Republic include the relevant statistics regarding Berlin for which separate data have not been supplied. This is without prejudice to any question of status which may be involved. Where the designation "country or area" appears in the headings of tables, it covers countries, territories, cities or areas. Signed articles express the opinions of the authors and do not necessarily represent the findings or policy of the World Health Organization. Couverture: Gilbert Auberson * T M F N Explication des signes Donnee preliminaire, approximative ou estimative. Donnee non disponible. Zero OU quantile ncgligeable. Categorie non applicable. Total. Masculin. Feminin. Nombres absolus. © Organisation mondiale de la Sante 1990 Les appellations employees dans cette publication et la presentation des donnees qui y figurent n'impliquent de la part du Secretariat de !'Organisation mondiale de la Sante aucune prise de position quant au statut juridique des pays, territoires, villes ou zones, ou de leurs autorites, ni quant au trace de leurs frontieres ou limites. Les statistiques se rapportant ii la Republique democra- tique allemande et ii la Republique federate d' Allemagne comprennent les statistiques pertinentes concernant Berlin pour lequel des donnees separees n'ont pas ete fournies. Toute question de statut demeure reservee. Lorsque l'appellation «pays ou zone)) apparait dans le titre des tableaux, elle couvre les pays, territoires, villes ou zones. Les articles signes expriment les vues de leurs auteurs et ne correspondent pas necessairement aux conclusions ou ii la politique adoptee par l'Organisation mondiale de la Sante. WORLD HEALTH STATISTICS QUARTERLY RAPPORT TRIMESTRIEL DE STATISTIQUES SANITAIRES MONDIALES Vol. 43, No. 3, 1990 ENVIRONMENTAL EPIDEMIOLOGY CONTENTS The impact of inadequate sanitary conditions on health Page in developing countries. Sharon R. A. Huttly . . . . . . . . . . 118 Indoor air pollution in developing countries. B. H. Chen, C. J. Hong, M. R. Pandey & K. R. Smith . . . . . . . . . . . . . 127 Acute pesticide poisoning: a major global health pro- blem. J. Jeyaratnam............................. 139 Assessing the health impact of urbanization. Brian T. Williams..................................... 145 Urban air pollution in Latin America and the Caribbean: health perspectives. Isabelle Romieu, Henyk Weitzenfeld & Jacobo Finkelman............................. 153 Lead in petrol - the mistake of the XXth century. Carl M. Shy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 Health risks associated with pollution of coastal bathing waters. Louis J. Saliba & Richard Helmer . . . . . . . . . . . . 177 The role of environmental and occupational hazards in the adult health transition. Tord Kjellstri:im & Linda Rosenstock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 Smoking-related deaths in developed countries - an update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 Corrigenda Vol. 43, No. 2, 1990 P. 62, Fig. 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 P. 106, right-hand column, second paragraph, the sec- ond sentence should read: Among males, the highest and lowest death rates occur in Hungary (1 566.0 per 1000 OOO) and Japan (837.5 per 100 OOO) respectively. L'ECO-EPIDEMIOLOGIE SOM MAIRE lnsuffisances de l'assainissement et sante dans les pays Pages en developpement [resume). Sharon R. A. Huttly . . . . . . 124 Pollution de l'air a l'interieur des habitations dans les pays en developpement [resume). B. H. Chen, C. J. Hong, M. R. Pandey & K. R. Smith . . . . . . . . . . . . . . . . . . 136 Intoxications aigues par les pesticides: probleme de sante majeur dans le monde [resume]. J. Jeyaratnam . . 143 Appreciation de l'impact de !'urbanisation sur la sante [resume). Brian T. Williams . . . . . . . . . . . . . . . . . . . . . . . 151 Pollution de l'air dans les villes d'Amerique latine et des Cara'ibes: perspectives sanitaires [resume]. Isabelle Romieu, Henyk Weitzenfeld & Jacobo Finkelman . . . . . . 165 Essence contenant du plomb - l'erreur du xx• siecle [resume]. Carl M. Shy. . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 Risques pour la sante lies a la pollution des eaux cotieres utilisees pour la baignade [resume]. Louis J. Saliba & Richard Helmer. . . . . . . . . . . . . . . . . . . . . . . . . 184 Le role des risques environnementaux et professionnels dans la transition sanitaire chez les adultes [resume]. Tord Kjellstri:im & Linda Rosenstock . . . . . . . . . . . . . . . . 195 Deces lies au tabagisme dans les pays developpes - une mise a jour. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 Rectificatifs Vol. 43, N° 2, 1990 P. 62, Fig. 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 P. 106, colonne de droite, deuxieme paragraphe, la seconde phrase devrait se lire comme suit: Among males, the highest and lowest death rates occur in Hungary (1 566.0 per 100 OOO) and Japan (837.5 per 100 OOO) respectively. - 118 - THE IMPACT OF INADEQUATE SANITARY CONDITIONS ON HEALTH IN DEVELOPING COUNTRIES Sharon R. A. HuttlY' The recognition of a link between poor sanitary conditions and cholera infection formed the basis of John Snow's pioneering epidemiological work about 150 years ago. Since then, considerable improve- ments in access to sufficient clean water and hygienic excreta-disposal facilities in the in- dustrialized countries have drastically reduced the incidence of water- and excreta-related diseases. In the developing world, however, these diseases ac- count for a large proportion of morbidity and mor- tality, especially among young children, and this is largely attributed to the lack of basic sanitary fa- cilities. In the face of this, the International Drinking Water Supply and Sanitation Decade (1981-1990) was launched amid high hopes of providing global coverage of services.b Concerted efforts were under- taken by governments and aid agencies and a va- riety of schemes were pursued. Review of progress throughout the Decade showed that the original coverage targets would not be realized and that large numbers of people would remain without adequate sanitary facilities by 1990.c Aside from advances in technically-related matters, the Decade has brought about an improved understanding of the relationship between poor sanitary conditions and health. This article reviews current knowledge of the major water- and sanitation-related diseases and the potential impact of improved sanitary facilities on health. Water-supply and sanitation facilities Coverage Fig. 1 shows the proportion of developing-country populations with access to adequate water-supply and sanitation facilities according to WHO regions and urban/rural areas as of 1985. Further details are available elsewhere.d Most striking are the dif- ferences in coverage between urban and rural areas, in each region. In addition, sanitation has received less attention than water supply, especially in rural areas. The proportion of countries reporting that more than half of their population were without access to safe, adequate water supplies was 38% and without adequate sanitation was 46% (compared to 45% and 47%, respectively, in 1980). The global coverage targets for 1990 are 87% and 61% in urban and rural areas respectively for water coverage, with • Research Fellow, Maternal and Child Epidemiology Unit, London School of Hygiene and Tropical Medicine, London, United Kingdom. b United Nations. Repon of the United Nations water conference, Mar de/ Plata, March 14-25. New York, 1977. (Document E/CONF 70/29). 'World Health Organization. Review of progress of the International Drinking Water Supply and Sanitation Decade, 1981-1990: eight years of implementation. Report by the Director-General. (Doc- ument EB83/3, 1988). • World Health Organization. The International Drinking Water Supply and Sanitation Decade: review of mid-decade progress (as at December 1985). (Document CWS/87.5, 1987). corresponding figures of 78% and 31% for sanita- tion. It should be noted that these are coverage figures only and do not necessarily reflect the proper functioning or utilization rates of the facilities. Relationship with poor health The mechanisms by which water- and excreta- related diseases are transmitted must be understood in order to evaluate the effect of poor sanitary conditions on health and the likely impact of im- proved facilities. A framework first proposed by Bradley in 1968 (1) and modified by Feachem in 1975 (2) provided an important advance in this re- spect. Water-related diseases were categorized as: (a) faecal-oral (water-borne or water-washed), for example diarrhoeal diseases, infectious hepatitis; (b) water-washed only, for example scabies, con- junctivitis; (c) water-based, for example dracunculiasis (guinea- worm disease), trachoma; (d) water-related insect-vector transmitted, for ex- ample malaria, onchocerciasis. Poor excreta-disposal facilities and hygienic prac- tices are mostly associated with those diseases in categories (a) and (b). Attention has been primarily focused on those dis- eases transmitted through the faecal-oral route, since they are the most widespread and constitute a large proportion of the morbidity/mortality burden of the population, especially among young children. The faecal-oral route is not confined to one type, however. Transmission through contaminated water, food, hands and eating utensils or via the ground is possible. Hence improvements in water quality and quantity, in excreta-disposal facilities and in hygienic practices, are all important for the interruption of pathogenic transmission. Potential for health impact The Decade has brought about a much greater un- derstanding of the mechanisms and magnitude of the impact of water-supply and sanitation inter- ventions on health. Disappointing and conflicting results from a number of health-impact evaluations had led many to believe that the often huge invest- ments necessary were not cost-effective. Thus, es- pecially where resources were limited, water supply and sanitation were assigned a relatively low priority by aid agencies and governments alike. However, many researchers have argued that apart from the basic right of people to a clean and adequate water supply and good sanitation, the health impact of these facilities is complex and therefore not easily seen or understood, but can be substantial neverthe- less. Rapp. trimest. statist. sanit. mand., 43 (1990) 119 - FIG.1 WATER SUPPLY AND SANITATION COVERAGE BY WHO REGION, 1985 COUVERTURE DES SERVICES D'APPROVISIONNEMENT EN EAU ET D'ASSAINISSEMENT PAR REGION OMS, 1985 A. WATER SUPPLY - APPROVISIONNEMENT EN EAU 100 l 80 CD 2 CD :ll CD "O c: 0 N 60 :, a. 0 0.. I ~ "O 40 "' 2: "' "' c: 0 ~ :, a. 0 20 0.. 0 "'"' ;E_g <C~ "' .~cii C:CD ~ "' ffi?g :, <f>W g <C ' c:C: '" .a; cii -g "'"' :Q E <OC/) a;·g u ~-6 u <C =a: 0 "' 5.~ <l>c: :, O<I> :::;;!!! er Cl)<( c: - ~ -"' i~ "' 0.. w:::. WHO Region - Region OMS - ~~ .Q O <!l:::;; Urban areas Zones urbaines Firstly, health-impact evaluations of water-supply and sanitation projects are themselves complex to conduct. Although much progress has been made in this area, many studies suffered from methodologi- cal flaws which rendered their results difficult to interpret (3). Secondly, extensive reviews (4, 5) have concluded several important facts: • The level of intervention, the functioning and the utilization of services all influence the impact on health. • Improvements in water quantity and excreta dis- posal appear more effective than those in water quality alone. It is often the case, however, that several sanitary improvements are implemented together in an integrated programme. Thus, assessing the relative health impacts of individual components is rendered more difficult. • An integrated approach is often desirable whereby attention is paid to upgrading water- supply and sanitation facilities alongside appro- priate education to maximize the use of facilities and improve hygienic practices. • Most attention has focused on supply rather than demand for services and more anthropological inputs are necessary for understanding non- technical problems, especially in the area of hygienic practices. Wld hlth statist. quart., 43 ( 1990) 8. SANITATION - ASSAINISSEMENT ~ "' 2 CD "' "' "' "O c: 0 -~ "3 a. 0 0.. I t "O "' ~ "' c: 0 -~ "3 a. 0 0.. ~ Rural areas 1.£.£.d Zones rurales 100 80 60 40 20 0 "''" u:, ·.:a <( :E < "'"' 1;; C:(1> .~ea "''" ffi?g u:, w 13 ~ ·;:: C" -6 c:c: Q,)•;:: :, "''" t:.J;1 E·<I> Cl) ffi § ~8 <C E <C :, :§.~ "O l;;CD "' <1>c: CD:l 'ijj ::iif!! 3:~ <C c: - -'" ·u ~:g "' 5i~ 0.. WHO Region - Region OMS ~~ .Q O <!l::; • The presence of certain risk factors (for example, low socioeconomic status), and the level of pathogen exposure, also influence health impact. • Improvements in water-supply and sanitation fa- cilities may be a necessary but not a sufficient condition for improving health (6). That is, aside from any direct impact or not, provision of these facilities may greatly enhance the impact of other interventions on health. Estimation of the public health impact of adequate sanitary conditions is thus confounded by the above facts. ,It cannot be assumed, for example, that in populations served by improved facilities, water- and excreta-related diseases will be greatly reduced. A multifactorial approach is needed to bring about the desired benefits. Water- and excreta-related diseases As mentioned above, there is a wide range of water- and excreta-related diseases. A review of all these diseases, and the impact of improved sanitary fa- cilities on them, is not feasible here. Instead, those which constitute the major health burden and/or are often the focus of health-impact assessments of water and sanitation interventions are reviewed. - 120 - Data are presented on diarrhoeal diseases, soil- transmitted nematode infections and dracunculiasis (guinea-worm disease). Nutritional status is also considered. Diarrhoeal diseases Diarrhoeal diseases are common throughout the de- veloping world and are a leading cause of morbidity and mortality in young children. Older children and adults are also affected although the consequences are usually less severe. Children. Snyder & Merson (7) estimated from longi- tudinal community-based studies that the median incidence rate of diarrhoea in children under 5 was 2.2 episodes per child per year, being higher (3.0) if only studies with frequent surveillance were con- sidered. The annual number of diarrhoea-associated deaths in 1980 was estimated as 4.6 million (13.6 per 1 OOO population under 5). Morbidity rates were gen- erally higher in older infants (6-11 months) and children 12-23 months, while mortality peaked in children under 2. Similar trends have been observed in many other epidemiological studies. More recently, incidence estimates based on 7 350 cross-sectional surveys in 70 countries have yielded a global median incidence rate of 3.4 episodes per child per year.• Diarrhoea was estimated to account for 35.8% of all deaths in the under-5s in the period 1981-1986. Based on these and demographic data for 1989/ young children in developing countries (ex- cluding China) experience an estimated 1 500 million episodes of diarrhoea per year and 4 million diar- rhoea-associated deaths (12.0 per 1 OOO population under 5) (Table 1). The morbidity/mortality burden is high in all regions although there is considerable • World Health Organization. Seventh programme report, 1988-1989. Programme for Control of Diarrhoeal Diseases. (Document WHO/ CDD/90.34). 'World Health Organization. Global estimates for health situation assessment and projections, 1990. (Document WHO/HST/90.2). variation, with children in Africa generally suffering the most. It should be noted, however, that there is also substantial variation within regions. Adults. Community-based data on adult diarrhoeal morbidity are scarce but estimates suggest that inci- dence rates are 3-5 times lower than those found in children 6-11 months (8). Diarrhoea may also be associated with a sizeable proportion of adult deaths but there are few available data. Three studies from sub-Saharan Africa reported diarrhoea to be associ- ated with 4-21% of adult deaths (8). In women of reproductive age in rural Bangladesh, 10.1% of deaths were due to diarrhoea (9). Impact of water-supply and sanitation improve- ments. Most studies have reported the impact on young children (under 5) and so the following dis- cussion is restricted to this age group. In their review of data from 67 studies in 28 countries, Esrey et al. (4) found that improved water supplies and sanitation facilities demonstrated a median reduction of 22% in diarrhoeal morbidity rates of children under 5. If only the studies of better quality were considered, the median reduction was 27%. Impact was found to vary according to a number of factors, as noted above. Data from several more recent studies, in various geographical settings and involving different levels of intervention, have also produced quite varied results. In rural areas of Nigeria and Bangladesh, longitudinal evaluation studies were conducted of integrated water, sanitation and hygiene education programmes (10, 11 ). In Nigeria, an impact of the project on diarrhoea morbidity was found only in limited subgroups of the population. In Bangladesh, however, diarrhoeal morbidity rates in under-5s, similar prior to the intervention project, became 25% lower in the intervention area than in the control area. Several results are now available from the "new breed" of health-impact evaluation studies, adopting a health-facility-based case-control design. In Malawi and the Philippines, children from families using good-quality water supplies and sanitation TABLE 1. ESTIMATES OF DIARRHOEAL MORBIDITY AND MORTALITY AMONG CHILDREN <5, BY WHO REGION (EXCLUDING CHINA), 1989 TABLEAU 1. ESTIMATION DE LA MORBIDITE ET DE LA MORTALITE DUES AUX MALADIES DIARRHEIQUES CHEZ LES ENFANTS DE <5 ANS, PAR REGION OMS (NON COMPRIS LA CHINE), 1989 South-East Eastern Western Africa Americas Asia Mediterranean Pacific• Global Afrique Ameriques Asiedu Mediterranee Pacifique Mon de Sud-Est orientale occidenta(a Number of countries - Nombre de pays ......... 43 32 11 23 14 123 Number of children <5 (millions) - Nombre d'enfants <5 ans (millions) ..................... 87.3 55.0 167.0 64.5 22.8 396.6 Number of episodes of diarrhoea per child per year - Nombre annuel d'episodes de diarrhee par enfant .. - 4.4 4.6 2.9 3.0 2.5 3.4 Total number of diarrhoea episodes (millions) - Nombre total d'episodes de diarrhee (millions) . . . . . . . . . . 383.7 252.6 476.1 193.9 57.3 1 362.6 Percentage deaths associated with diarrhoeab - Pour- centage de deces lies a la diarrheeb . . . . . . . . . . . 37.7 35.2 28.3 39.1 29.0 35.8 Number of diarrhoea deaths (thousands) - Nombre de deces dus a la diarrhee (milliers) . . . - - - . . . . . . . 1 498 339 1 630 916 129 4 827 Diarrhoea death rate per 1 OOO - Taux de mortalite due a la diarrhee (pour 1 OOO) ................ 17.4 6.0 9.5 14.0 5.5 12.0 Case fatality rate per 100 episodes - Taux de letalite (pour 1 oo episodes) ...................... 0.39 0.12 0.35 0.52 0.23 0.36 • Excluding China - La Chine non comprise. • Source: World Health Organization. Sixth programme report, 1986-1987. Programme for Control of Diarrhoeal Diseases. (Document WHO/CDD/88.28) (anglais seulement). Rapp. trimest. statist. sanit. mond., 43 (1990) - 121 - facilities experienced 20% less diarrhoea (12, 13). The evaluation of a rural sanitation programme in Lesotho (which included a hygiene-education com- ponent) showed that in households with a latrine, children suffered 24% less episodes of diar- rhoea (14). In Sri Lanka, water quality was believed to play an important role in the association found between improved water supplies and diar- rhoea (15). Compared to children from families using unprotected supplies, those from families using handpumps or protected traditional wells experi- enced 46% and 35% fewer episodes of diarrhoea respectively. In their review (4), Esrey et al. showed a median reduction of 21% in mortality due to all causes associated with sanitary improvements (30% in the better-quality studies). Only two studies (in urban areas) reported on diarrhoea-associated mortal)ty, both showing a positive impact with a reduction of about 40%. More recently, data from rural Bangladesh, adjusted for confounding variables, showed post-neonatal mortality rates to be 68% lower among families with a latrine (16). In an urban area of southern Brazil, the availability of piped water was an important factor-diarrhoea mortality, after controlling for confounding variables, was 79% lower among those children from households with in-house supplies compared to those of households with no easy access to piped water (17). The infectious dose (ID) of a diarrhoea-causing pathogen (the number of ingested organisms re- quired to cause diarrhoea) varies according to the particular pathogen and various host factors. Esrey et al. (4) suggested that water and sanitation im- provements would be more likely to have an impact on diarrhoeas caused by high-ID pathogens (for example, cholera) than on those caused by low-ID pathogens (for example, Entamoeba histolytica, rotavirus). Their review of etiology-specific impact studies supported this hypothesis, with reductions of over 40% of median and high ID pathogenic diar- rhoeas and 0-2% of protozoa! diarrhoeas. As noted previously, there are several routes for the faecal-oral transmission of diarrhoea-causing patho- gens. Thus the mechanisms whereby the provision of improved water-supply and sanitation facilities reduce this transmission are recognized as complex and their measurement has not proved easy. For example, while high levels of faecal contamination in food and water are generally assumed to be associated with diarrhoeal diseases, a direct relation- ship has rarely been quantified (18, 19)--perhaps largely because of the difficulties inherent in d?ing so. Most evidence is indirect, based on observations of diarrhoea and water-pollution seasonality or on peak diarrhoea rates coinciding with the introduction of (contaminated) supplementary foods. Several studies have reported high levels of food and water contamination in the domestic environment (18, 20- 22). Even where clean water is supplied, high de- grees of contamination occur between the source and use in the home (14, 23-25). This may partly account for the growing evidence that improvements in water quantity or excreta disposal are of greater importance than water quality in the reduction of diarrhoeal diseases (5, 26). Studies of food-handling practices have reported contamination of ~and_s and utensils, as well as of raw foods, as contributing to food contamination.9 Food preparation and storage • Esrey, S. A. & Feachem, R. G. Interventions for the control of diarrhoeal diseases among young children: promotion of food hygiene. (Document WHO/CDD/89.30). Wld hlth statist. quart., 43 (1990) practices also play a role. Undercooked or inad- equately reheated foods allow the survival and pro- liferation of bacterial pathogens, as does storage at ambient temperatures. However, little is known about the link between food-hygiene practices and either food contamination with enteric pathogens or the risk of diarrhoea. This uncertainty about the role of food as a diarrhoea-transmission vehicle is acknow- ledged by Esrey & Feachemg who estimated that it may contribute to 15-70% of all diarrhoea episodes. The above findings also indicate a role for appro- priate hygiene education. The introduction of im- proved water supplies (and in some_ cases, access _to greater quantities of water) and suitable faecal d_1s- posal facilities, provides an ideal situation for 1~- tegrated hygiene education programmes and this has been strongly promoted. This broadens the nec- essary range of skills required in implementing s~ni- tary interventions to include those of anthropologists and communications personnel. Although evidence is scarce on the eff~ctive_nes~ of hygiene-education programmes, reductions in diar- rhoea incidence rates of 14-48% have been reported in a review of three studies (27). In addition, an educational intervention in urban Bangladesh was based on three hygienic practices which had been recorded as associated with high rates of diarrhoea. During the six months following the interv~ntion, diarrhoea incidence rates were 26% lower in the intervention than in the control communities (28). An integrated water, sanitation and hygiene intervention programme in rural Bangladesh examined diarrhoea incidence in relation to household hygienic prac- tices (29). In both the intervention and control areas, diarrhoea incidence rates were inversely related to the number of improved hygienic practices (related to water-source use, disposal of faeces and maternal handwashing) that were reported by the household. Intestinal helminth infections Intestinal helminth infections constitute some of the most common in human beings throughout the world, the most prevalent being the soil-transmitted nematode infections. Inadequate sanitation facilities and poor hygienic practices provide for easy trans- mission via the faecal-oral route. Despite their high prevalence, these infections are generally a_ccorded relatively low priority in health-care planning. Im- provements in water-supply and sanitation facilities and appropriate health education, alongside chemo- therapeutic treatment, seem to offer the best hopes for control. The most widespread soil-transmitted nematodes are Ascaris lumbricoides, hookworms and Trichuris trichiura. The global number of people afflicted with these three infections has been estimated as 1 OOO million, 900 million and 500 million respectively (30). Data from a number of geographical areas, however, have shown that infection with T. trichiura occurs as frequently as that of A. lumbricoides (31 ). Simulta- neous infection is known to occur often. The number of worms per person (intensity of infection), how- ever, is not uniform, with most people ~arbouring a few worms and a small number suffering a heavy burden. It is the latter group who are most suscep- tible to complications, such as intestinal obstruction. Other consequences include poor nutritional status, anaemia and, in the case of T. trichiura, bloody and sometimes persistent diarrhoea (32). - 122 - Although case-fatality rates are low, the sheer mag- nitude of the prevalence rates and worldwide dis- tribution have led to estimates of A. lumbricoides contributing to 100 OOO deaths per year (33). Actual prevalence and intensity rates of these nema- tode infections differ widely both between and within countries, partly depending on environmental and socioeconomic conditions. Epidemiological studies from many different geographical areas, however, have shown common trends (34). Preva- lence of infection rises rapidly in the first years of life, and then levels out in adulthood. Intensity, on the other hand, tends to peak in school-age children for A. lumbricoides and T. trichiura and in adults for hookworm infections. Impact of water-supply and sanitation improve- ments. Despite the acknowledged role of poor sani- tary conditions in the transmission of these nema- tode infections, the impact of improved water-supply and sanitation interventions on them has not been well studied. Data from Saint Lucia showed that prevalences of infections with A. lumbricoides and T. trichiura in young children were significantly lower in areas with improved sanitary conditions (35), as was reinfection (6 months after chemotherapy) with A. lumbricoides (36). Crowding and type of excreta- disposal facility were the only significant predictors of reinfection found in a multivariate analysis. A study of A. lumbricoides and T. trichiura, conducted in two urbans slums in Bangladesh, showed lower rates of prevalence, intensity and reinfection (9 months after chemotherapy) among young children in the area with improved sanitation (F. J. Henry et al., submitted for publication). In a rural area of the United Republic of Tanzcinia, however, a latrine cam- paign plus treatment against hookworm only result- ed in a decrease in the intensity of hookworm infections for up to 6 months after the inter- ventions (37). While in Mozambique, an evaluation of a programme for improved pit-latrine construction showed no association between the type of latrine used and Ascaris infection or presence of Ascaris eggs in the soil (38). The authors concluded, how- ever, that behavioural factors were largely respon- sible for the apparent lack of impact of the latrines. Dracunculiasis Dracunculiasis (guinea-worm disease) has been identified as the only major water-related infection which can be eradicated through water-supply im- provements alone. As such, its elimination by 1990 was adopted as a subgoal of the International Drink- ing Water Supply and Sanitation Decade. Although this target will not be achieved, great progress has been made in this direction and the goal is still seen as attainable, probably within the next decade. Dracunculiasis is a painful, debilitating disease which confers no immunity and therefore frequently recurs each year through the continued use of infect- ed water supplies. It is endemic in 19 African coun- tries, India, Pakistan, Saudi Arabia and Yemen (39), occurs mostly in remote rural areas and is greatly underreported. It has been estimated that ap- proximately 10 million people are infected each h Esrey, S. A. & Habicht, J. P. Nutritional anthropometric indica- tors for evaluating water and sanitation projects. Paper presented at the International workshop on measuring the health impact of water and sanitation programmes, Cox's Bazaar, Bangladesh, 21-25 November 1983. year (39) but accurate data are limited. Recently, however, the Decade has focused attention on the disease in endemic areas and many affected coun- tries have undertaken eradication campaigns and greater surveillance. Details of the parasite's life cycle and epidemiology of the disease are given elsewhere (39). Adult female worms, ingested through contaminated water, emerge through painful blisters on the victim's skin about 12 months after infection. Although not in itself a fatal disease, complications such as tetanus resulting from secondary infection of the blisters are not uncommon. However, its social and economic consequences are most striking. Incidence rates vary widely in different areas, with some villages report- ing up to 70% of their population affected in a year. Cases are most common among the economically active, however, and are usually seasonal, often occurring at critical times in the agricultural cal- endar. School absenteeism rates have also been noted as markedly higher during "the guinea-worm season". In a prospective study conducted in south- eastern Nigeria, the mean duration of symptoms was 12.7 weeks (40). Severe disability (unable to leave the household compound) was found in 58% of episodes, these severe symptoms lasting 2-12 weeks. While few studies have sufficient data to quantify the economic impact of this disease, losses would appear to be substantial. Rao (41) estimated that in India, among the economically active, about 70 days of wages are lost for each episode. Possible indirect impacts on nutritional status have also been suggested. Impact of water-supply improvements. Rapid de- clines in the incidence of guinea worm have been noted in countries where eradication campaigns (usually involving some form of water-supply im- provement) have been undertaken (42, 43). An eval- uation study conducted in mid-western Nigeria showed that within three years of the implementa- tion of handpumps, village point-prevalence rates of dracunculiasis could fall from over 50% to 0% or near 0% (44). Prevalence rates in the control villages remained virtually unchanged. Moreover, the greatest impact was seen in villages with well-sited and correctly functioning boreholes. Socioeconomic benefits, including increased agricultural production, were also perceived by the villagers. In the evalu- ation of a similar programme in south-eastern Nigeria, the exclusive use of borehole water was inversely related, and the prevalence of dracunculia- sis positively related, to the household borehole distance (10). The problem of people consuming contaminated water away from the home, for example while farming, has been noted in these and other studies. This emphasizes the importance of a health-education component. Nutritional status Malnutrition is widely prevalent in the developing world due to various interrelated factors such as morbidity, diet, environmental conditions and socio- economic levels (45). In addition it is a common cause or associated cause of child death. Improve- ments in water-supply and sanitation facilities may have an impact on nutritional status not only via the well-documented malnutrition/morbidity relation- ship, but also through more time available for child care (24, 46, 47) and greater agricultural produc- tion (44). Indeed, Esrey & Habicht" have suggested that anthropometric indicators are as sensitive as Rapp. trimest. statist. sanit. mond., 43 ( 1990) - 123 - diarrhoeal indicators to sanitary improvements and can be measured more precisely. Impact of water-supply and sanitation improve- ments. Of the six studies reviewed by Esrey et al. (4) which investigated the impact of improved sanitary facilities on nutritional status, all showed a positive association (although not always significant). More recently, a water and sanitation intervention in south-eastern Nigeria reported a significant impact on acute but not chronic malnutrition in young children (10). The evaluation of a latrine programme in Lesotho showed that children from households with a latrine had better height-for-age (D. L. Daniels et al., submitted for publication). In Bangladesh, where nutritional status is particularly poor, no such impact by improved water and sanitation facilities was detected despite a significant reduction in diar- rhoeal diseases (48). A hygiene-education interven- tion in urban Bangladesh, aimed at sanitary be- haviours such as excreta disposal, similarly showed an impact on diarrhoea but not on nutritional status (49). Other studies have shown associations between water- and sanitation-related factors and nutritional status (47, 50-52) but this was not the case in Sri Lanka (53). It appears likely that the potential impact of improved sanitary conditions on nutritional status depends on many other factors, as is the case for diarrhoeal diseases. Conclusions The magnitude of the health problems associated with poor sanitary conditions is evident from the above. It is also clear, however, that simply provid- ing a clean, adequate water supply and good sanita- tion facilities does not automatically result in any significant improvement in health. There are many factors to consider, both in programme design and implementation, and in health-impact evaluation. One such issue in assessing health impact is the multifactorial nature of most of the major health indicators considered. Indeed, in the case of diar- rhoeal diseases a range of appropriate interventions has been promoted (54). Similarly, nutritional im- provements are also attempted in a variety of ways. It has been suggested that improvements in water- supply and sanitation facilities would realize rel- atively little health benefit in either low or high socioeconomic populations, an appreciable impact only being seen in middle-level communities (55). Others have argued for a minimum level of sanitary improvement before any health benefit can be ex- pected (56). Briscoe (6) has pointed out several flaws in the analytic methods used to assign low priority to water-supply and sanitation improvements. These flaws include misleading cost-effectiveness analyses that fail to take account of the multiple health ben- efits, direct or indirect, of sanitary improvements. Also, it was concluded that long-term effects on child survival are probably substantial, in addition to the more immediate impacts on morbidity. Thus while it is clear that there are many factors to consider, a major role for improving water supplies and sanitation facilities in development activities seems justified (6, 57). As the International Drinking Water Supply and Sani- tation Decade draws to a close, continued efforts are necessary to reach the many people who still lack coverage by adequate sanitary facilities. The insights gained in recent years into the relationship between sanitary conditions and health are invaluable for future programmes which should capitalize on these experiences. Improvements ranging from more tech- nological aspects (for example, the development of low-cost facilities, appropriate siting of improved water supplies) to more behaviour-related factors (for example, improving the disposal of faeces of young children, who rarely use latrines) can all help combat the appalling morbidity and mortality bur- den suffered in the developing world. Acknowledgements Dr F. J. Henry's contributions to earlier drafts of this article are gratefully acknowledged. SUMMARY One of the achievements of the International Drink- ing Water Supply and Sanitation Decade (1981-1990) is a better understanding of the relationship between poor sanitary conditions and health. This article reviews current knowledge of the major water- and excreta-related diseases and the potential impact of improved sanitary facilities on health. Four health indicators are considered-diarrhoeal diseases, soil-transmitted nematode infections, dracunculiasis (guinea-worm disease) and nutritional status. These constitute some of the major health problems in the developing world. Although there are large variations in morbidity and mortality rates both between and within geographical areas, some global estimates are presented. Children under 5 experience an average of 3.4 episodes of diarrhoea per year and a diarrhoeal mortality rate estimated Wld hlth statist. quart .. 43 (19901 at 12 per 1 OOO, leading to 4 million diarrhoea- associated deaths per year. Diarrhoea may also be associated with a sizeable proportion of adult deaths. Approximately 1 OOO million people are be- lieved to be infected with Ascaris lumbricoides, 900 million with hookworms and 500 million with Trichuris trichiura. Dracunculiasis, primarily occur- ring in remote rural areas, is known to be greatly underreported, but estimates suggest that 10 million people are infected each year. Malnutrition is widely prevalent and a common cause or associated cause of child death. The complex issues involved in measuring the health impact of improved sanitary facilities and the mechanisms by which these impacts may occur are discussed. These complexities, plus the differences in sanitary improvements and environmental set- - 124 - tings, and the diverse nature of the health indicators considered, all contribute substantial variations in the health impact observed. There is good evidence, however, that sanitary improvements have multiple health benefits, direct and indirect, both in the short and the long term. The insights gained during recent years into the relationship between poor sanitary conditions and health are invaluable for improving water-supply and sanitation programmes. During this time, many recommendations have been made; these include emphasis on technological issues and, increasingly, on behaviour-related factors. An integrated approach has been advocated, involving hardware implemen- tation alongside appropriate education in order to maximize facility usage and hygienic practices and to bring about the desired health benefits. RESUME lnsuffisances de l'assainissement et sante dans les pays en developpement Parmi les succes de la Decennie internationale de l'eau potable et de l'assainissement (1981-1990) figure une meilleure comprehension des relations entre de mauvaises conditions d'hygiene et la sante. Cet article contient un bilan des connaissances ac- tuelles sur les principales maladies liees a l'eau et aux excreta et sur l'impact que pourrait avoir sur la situation sanitaire !'amelioration des services d'assainissement. Ouatre indicateurs sanitaires sont pris en considera- tion: les maladies diarrheiques, les infestations par les nematodes, la dracunculose (ou maladie due au ver de Guinee) et l'etat nutritionnel. lls representent certains des principaux problemes de sante dans le monde en developpement. Malgre d'importantes variations des taux de morbidite et de mortalite entre regions geographiques et a l'interieur de memes regions, quelques estimations mondiales sont presentees. Les enfants de mains de 5 ans presentent en moyenne 3,4 episodes de diarrhee par an et un taux de mortalite par diarrhee estime a 12 pour 1 OOO, ce qui correspond a 4 millions de deces lies a la diarrhee par an. La diarrhee peut egalement etre associee a une proportion non negligeable de deces d'adultes. On estime a environ 1 OOO millions le nombre des personnes infestees par Ascaris lum- bricoides, 900 millions par les ankylostomes et 500 millions par Trichuris trichiura. Les cas notifies de dracunculose, qui sevit surtout dans les zones rurales isolees, sont, on le sait, tres inferieurs au nombre reel des cas mais, d'apres les estimations, 10 millions de personnes seraient infectees chaque annee. La malnutrition est tres repandue et une cause courante ou associee de mortalite infanto- juvenile. Les problemes complexes poses par la mesure de l'impact sur la sante de !'amelioration des services d'assainissement et les mecanismes ainsi mis en ceuvre sont passes en revue. Ces complexites, ajoutees aux differences dans les ameliorations apportees aux services d'assainissement et les con- textes environnementaux ainsi qu'a la diversite des indicateurs sanitaires pris en consideration, font que les impacts observes varient sensiblement. Cela etant, ii est amplement prouve que !'amelioration des conditions d'hygiene est a maints egards favo- rable pour la sante, directement et indirectement, a court et a long terme. Les etudes faites ces dernieres annees sur les re- lations entre une mauvaise hygiene et la sante ont donne des resultats precieux pour !'amelioration des programmes d'approvisionnement en eau et d'assainissement. Parallelement, de nombreuses re- commandations ant ete formulees. Elles privilegient entre autres les aspects technologiques et, de plus en plus, les facteurs lies au comportement. On a preconise une approche integree associant des reali- sations concretes et des activites educatives ap- propriees pour optimaliser !'utilisation des in- stallations et les pratiques d'hygiene et obtenir ainsi les resultats attendus sur le plan sanitaire. REFERENCES- REFERENCES 1. BRADLEY, D. J. & EMURWON, P. Predicting the epi- demiological effect of changing water sources. East African medical journal, 45: 284-291 (1968). 2. FEACHEM, R. G. Water supplies for low-income communities in developing countries. Journal of the Environmental Engineering Division, Pro- ceedings of the American Society of Civil En- gineers, 101 (EE5): 687-702 (1975). 3. BLUM, D. & FEACHEM, R. G. Measuring the impact of water-supply and sanitation investments on diarrhoeal diseases: problems of methodology. International journal of epidemiology, 12 (3): 357-365 (1983). 4. EsREY, S. A. ET AL. Interventions for the control of diarrhoeal diseases among young children: im- proving water supplies and excreta-disposal fa- cilities. Bulletin of the World Health Organiza- tion, 63 (4): 757-772 (1985). ESREY, S. A. ET AL. La lutte contre les maladies diarrheiques du jeune enfant: interventions visant a ameliorer l'approvisionnement en eau et !'elimination des excreta [resume]. Bulletin de /'Organisation mondiale de la Sante, 63 (4): 769 (1985). 5. ESREY, s. A. & HABICHT, J.-P. Epidemiologic evi- dence for health benefits from improved water Rapp. trimest. statist. sanit. mond., 43 ( 1990) - 125 - and sanitation in developing countries. Epi- demiologic reviews, 8: 117-128 (1986). 6. BRISCOE, J. A role for water supply and sanitation in the child survival revolution. Bulletin of the Pan American Health Organization, 21 (2): 93- 105 (1987). 7. SNYDER, J. D. & MERSON, M. H. The magnitude of the global problem of acute diarrhoeal disease: a review of active surveillance data. Bulletin of the World Health Organization, 60 (4): 605-613 (1982). SNYDER, J. D. & MERSON, M. H. Ampleur du pro- bleme des diarrhees aigues: analyse des don- nees de surveillance active [resume]. Bulletin de /'Organisation mondiale de la Sante, 60 (4): 612 (1982). 8. KIRKWOOD, B. R. Diarrhoea. In: Feachem, R. G. & Jamieson, D. T. (eds), Disease and mortality in sub-Saharan Africa. New York, Oxford University Press, 1990. 9. FAUVEAU, V. ET AL Epidemiology and cause of deaths among women in rural Bangladesh. Inter- national journal of epidemiology, 18 ( 1 ): 139-145 (1989). 10. HurrLY, S. R. A. ET AL. The Imo State (Nigeria) Drinking Water Supply and Sanitation Project -2. Impact on dracunculiasis, diarrhoea and nutritional status. Transactions of the Royal Society of Tropical Medicine and Hygiene, 84: 316-321 (1990). 11. Az.12, K. M. A. ET AL. Reduction in diarrhoeal dis- eases in children in rural Bangladesh by environ- mental and behavioural modifications. Trans- actions of the Royal Society of Tropical Medicine and Hygiene: in press (1990). 12. YOUNG, B. & BRISCOE, J. A case-control study of the effect of environmental sanitation on diar- rhoea morbidity in Malawi. Journal of epidemio- logy and community health, 42 ( 1 ): 83-88 ( 1988). 13. BALTAZAR, J. ET AL Can the case-control method be used to assess the impact of water supply and sanitation on diarrhoea? A study in the Philippines. Bulletin of the World Health Organ- ization, 66 (5): 627-635 (1988). BALTAZAR, J. ET AL. La methode cas-temoins peut- elle etre utilisee pour evaluer l'impact de l'approvisionnement en eau et des conditions d'hygiene sur la diarrhee? Exemple d'application aux Philippines [resume]. Bulletin de /'Organi- sation mondia/e de la Sante, 66 (5): 635 (1988). 14. DANIELS, D. L. ET AL A case-control study of the impact on diarrhoea morbidity of improved sani- tation in Lesotho. Bulletin of the World Health Organization, 68 (4): 455-463 (1990). DANIELS, D. L. ET AL. Effets de !'amelioration de l'assainissement sur la morbidite liee a la diar- rhee au Lesotho: etudes cas-temoins [resume]. Bulletin de /'Organisation mondiale de la Sante, 68 (4): 463 (1990). 15. MERTENS, T. E. ET AL Childhood diarrhoea in Sri Lanka: a case-control study of the impact of improved water sources. Tropical medicine and parasitology, 41: 98-104 (1990). 16. RAHMAN, M. ET AL Impact of environmental sani- tation and crowding on infant mortality in rural Bangladesh. Lancet, ii: 28-31 (1985). 17. V1CTORA, C. G. ET AL Water supply, sanitation and housing in relation to the risk of infant mortality from diarrhoea. International journal of epi- demiology, 17 (3): 651-654 (1988). Wld hlth statist. quart., 43 (1990) 18. BLACK, R. E. ET AL Contamination of weaning foods and transmission of enterotoxigenic Es- cherichia coli diarrhoea in children in rural Bangladesh. Transactions of the Royal Society of Tropical Medicine and Hygiene, 76: 259-264 (1982). 19. HENRY, F. J. ET AL Environmental sanitation, food and water contamination and diarrhoea in rural Bangladesh. Epidemiology and infection, 104 (1): 79-85 (1990). 20. BLACK, R. E. ET AL Incidence and etiology of infantile diarrhoea and major routes of transmis- sion in Huascar, Peru. American journal of epi- demiology, 129 (4): 785-799 (1989). 21. ROWLAND, M. G. M. ET AL Bacterial contamination in traditional Gambian weaning foods. Lancet, i: 136-138 (1978). 22. HENRY, F. J. ET AL Bacterial contamination of weaning foods and drinking water in rural Bangladesh. Epidemiology and infection, 104 (2): 253-259 (1990). 23. BLUM, D. ET AL The Imo State (Nigeria) Drinking Water Supply and Sanitation Project. -1. De- scription of the project, evaluation methods and impact of intervening variables. Transactions of the Royal Society of Tropical Medicine and Hygiene, 84: 309-315 (1990). 24. FEACHEM, R. G. ET AL. Pollution, hygiene and health. In: Water, health and development-An interdisciplinary evaluation. London, Tri-Med Books Ltd., 1978. (pp. 112-138). 25. HENRY, F. J. & RAHIM, Z. Transmission of diar- rhoea in two crowded areas with different sani- tary facilities in Dhaka, Bangladesh. Journal of tropical medicine and hygiene, 93 (2): 121-126 (1990). 26. CAIRNCROSS, S. The benefits of water supply. In: Pickford, J. (ed.), Developing world water II. London, Grosvenor Press, 1987. (pp. 30-34). 27. FEACHEM, R. G. Interventions for the control of diarrhoeal diseases among young children: pro- motion of personal and domestic hygiene. Bulletin of the World Health Organization, 62 (3): 467-476 (1984). FEACHEM, R. G. Lutte contre les maladies diar- rheiques chez les jeunes enfants: promotion de !'hygiene personnelle et domestique [resume]. Bulletin de /'Organisation mondiale de la Sante, 62 (3): 475 (1984). 28. STANTON, B. F. & CLEMENS, J. D. An educational intervention for altering water-sanitation be- haviors to reduce childhood diarrhoea in urban Bangladesh - II. A randomized trial to assess the impact of the intervention on hygienic be- haviors and rates of diarrhoea. American journal of epidemiology, 125 (2): 292-301 (1987). 29. ALAM, N. ET AL Mothers' personal and domestic hygiene and diarrhoea incidence in young children in rural Bangladesh. International journal of epidemiology, 18 ( 1 ): 242-247 ( 1989). 30. WHO Expert Committee. Public health signifi- cance of intestinal parasitic infections. Bulletin of the World Health Organization, 65 (5): 575-588 (1987). COMITE OMS D'EXPERTS. Importance des para- sitoses intestinales en sante publique. Bulletin de /'Organisation mondiale de la Sante, 66 (1): 23-34 (1988). 31. BUNDY, D. P. Epidemiological aspects of Trichuris and trichuriasis in Caribbean communities. - 126 - Transactions of the Royal Society of Tropical Medicine and Hygiene, 80: 706-718 (1986). 32. HARLAND, P. S. E. G. Trichuriasis in Africa and the Caribbean. In: Walker-Smith, J. A. & McNeish, A. S. (eds), Diarrhoea and malnutrition in child- hood. London, Butterworths & Co. Ltd, 1986. 33. Ascariasis. Lancet, i: 997-998 (1989). 34. ANDERSON, R. M. The population dynamics and epidemiology of intestinal nematode infections. Transactions of the Royal Society of Tropical Medicine and Hygiene, 80: 686-696 (1986). 35. HENRY, F. J. Environmental sanitation, infection and nutritional status of infants in rural St. Lucia, West Indies. Transactions of the Royal Society of Tropical Medicine and Hygiene, 75: 507-513 (1981). 36. HENRY, F. J. Reinfection with Ascaris lumbricoi- des after chemotherapy: a comparative study in three villages with varying sanitation. Trans- actions of the Royal Society of Tropical Medicine and Hygiene, 82: 460-464 (1988). 37. TANNER, M. ET AL Longitudinal study on the health status of children in a rural Tanzanian community: parasitoses and nutrition following control measures against intestinal parasites. Acta tropica, 44: 229-244 (1986). 38. MULLER, M. ET AL Evaluation of a sanitation pro- gramme using eggs of Ascaris lumbricoides in household yard soils as indicators. Journal of tropical medicine and hygiene, 92 (1): 10-16 (1989). 39. HOPKINS, D. R. Dracunculiasis: an eradicable scourge. Epidemiological reviews, 5: 208-219 (1983). 40. SMITH, G. S. ET AL Disability from dracunculiasis: effect on mobility. Annals of tropical medicine and parasitology, 83 (2): 151-158 (1989). 41. RAO, C. K. Epidemiology of dracunculiasis in India. In: Workshop on opportunities for the control of dracunculiasis; contributed papers, Washington, D.C., June 16-19, 1982. Washington, D.C., National Academy Press, 1982. 42. HOPKINS, D. R. Dracunculiasis eradication: a mid- Decade status report. American journal of tropi- cal medicine and hygiene, 37 (1): 115-118 (1987). 43. Dracunculiasis-Global surveillance summary, 1986. Weekly epidemiological record, 62 (45): 337-339 ( 1987). Dracunculose - Surveillance mondiale en 1986 (resume). Re/eve epidemiologique hebdoma- daire, 62 (45): 337-339 (1987). 44. EDUNGBOLA, L. D. ET AL. The impact of a UNICEF- assisted rural water project on the prevalence of guinea-worm disease in Asa, Kwara State, Nigeria. American journal of tropical medicine and hygiene, 39 (1): 79-85 (1988). 45. HAAGA, J. ET AL An estimate of the prevalence of child malnutrition in developing countries. World health statistics quarterly, 38 (3): 331-346 (1985). HAAGA, J. ET AL. Estimation de la prevalence de la malnutrition infantile dans les pays en develop- pement. Rapport trimestriel de statistiques sani- taires mondiales, 38 (3): 331-346 (1985). 46. POPKIN, B. M. & SOLON, F. S. Income, time, the working mother and child nutrition. Environ- mental child health, 8: 156-166 (1976). 47. TOMKINS, A. M. Water supply and nutritional status in rural northern Nigeria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 72 (3): 239-243 (1978). 48. HASAN, K. Z. ET AL Lack of impact of a water and sanitation intervention on the nutritional status of children in rural Bangladesh. European journal of clinical nutrition, 43: 837-843 (1989). 49. STANTON, B. F. ET AL. Educational intervention for altering water-sanitation behavior to reduce childhood diarrhoea in urban Bangladesh: im- pact on nutritional status. American journal of clinical nutrition, 48 (5): 1166-1172 ( 1988). 50. HEBERT, J. R. Effects of water quality and water quantity on nutritional status: findings from a south India community. Bulletin of the World Health Organization,63 (1): 143-155 (1985). HEBERT, J. R. Influence de la quantite et de la qualite de l'eau sur l'etat nutritionnel: obser- vations faites dans une collectivite du sud de l'lnde [resume]. Bulletin de /'Organisation mon- diale de la Sante, 63 (1): 153-154 (1985). 51. EsREY, S. A. ET AL. Drinking water source, diar- rhoeal morbidity and child growth in villages with both traditional and improved water sup- plies in rural Lesotho, Southern Africa. American journal of public health, 78 ( 11 ): 1451-1455 (1988). 52. BERTRAND, W. E. ET AL Differential correlates of nutritional status in Kinshasa, Zaire. International journal of epidemiology, 17 (3): 556-567 (1988). 53. COUSENS, S. N. ET AL. The anthropometric status of children in Kurunegala district in Sri Lanka: its relationship to water supply, sanitation and hygiene practices. Tropical medicine and para- sitology, 41: 105-114 (1990). 54. FEACHEM, R. G. Preventing diarrhoea: what are the policy options? Health policy and planning, 1 (2): 109-117 (1986). 55. SHUVAL, H. I. ET AL. Effect of investments in water supply and sanitation on health status: a threshold saturation theory. Bulletin of the World Health Organization, 59 (2): 243-248 (1981). SHUVAL, H. I. ET AL. De l'effet des mesures d'hygiene collectives sur la sante de la popula- tion: theorie seuil-saturation [resume]. Bulletin de /'Organisation mondiale de la Sante, 59 (2): 248 (1981). 56. KAWATA, K. Water and other environmental inter- ventions-the minimum investment concept. American journal of clinical nutrition, 31: 2114- 2123 (1978). 57. OKUN, D. A. The value of water supply and sani- tation in development: an assessment. American journal of public health, 78 ( 11 ): 1463-1467 (1988). Rapp. trimest. statist. sanit. mand., 43 (1990) - 127 - INDOOR AIR POLLUTION IN DEVELOPING COUNTRIES B. H. Chen.a C. J. Hong,b M. R. Pandey<' & K. R. Smithd Air pollution in general and indoor air pollution in particular have been associated in many people's mind with industrialization and urbanization and, thus, with the cities of developed countries where most of the measurements have been made. Re- cently, however, the UNEP/WHO Global Environment Monitoring System (GEMS)• has demonstrated quite convincingly that the worst ambient conditions re- ported today obtain in the cities of developing coun- tries. In a similar fashion, although most studies of indoor air quality have been carried out in devel- oped-country buildings, for many important pol- lutants the largest indoor-pollutant concentrations and exposures are also found in the developing countries in both rural and urban households. Worldwide, there are four principal classes of indoor pollutants (1), those derived from: (i) combustion, e.g. carbon monoxide, particulates, and sulfur and nitrogen oxides from cooking stoves, space heaters and cigarettes (environ- mental tobacco smoke--ETS); (ii) building materials, furnishings and chemical products, e.g. pesticides, volatile organic com- pounds, formaldehyde; (iii) the ground under the building, i.e. radon; (iv) biological processes, e.g. maid, mildew, mites. As with ambient emissions, the impact of indoor emissions on concentrations depends directly on ventilation (airflow and mixing volume). In general, most housing in developed countries lies in tem- perate latitudes and thus has relatively low air ex- change rates, i.e. typically one air change per hour or less. Under these conditions, even rather small emission rates can result in indoor concentrations that exceed those outdoors, considering also that indoor air quality is affected by outdoor levels. Since people spend a great proportion of time indoors, typically 90% or more, indoor pollution is frequently the predominant factor in total exposure. In the developing world, however, much housing lies in tropical and subtropical regions and typically has much higher ventilation rates. In these circum- stances, it takes fairly high emission rates to pro- duce significant pollutant levels. Unfortunately, there seem to be many situations in which such strong indoor sources exist. Moreover, in spite of being a minority, several hundred million people in develop- ing countries require space heating for a significant part of the year, and thus share with developed • International Programme on Chemical Safety, World Health Organization, Geneva. b School of Public Health, Shanghai Medical University, Shanghai, China. ' Bir Hospital, Kathmandu, Nepal. d Environment and Policy Institute, East-West Center, Honolulu, Hawaii, United States of America. • Global Environment Monitoring System. Assessment of urban air quality. United Nations Environment Programme/World Health Organization, 1988. (Unpublished document). Wld hlth statist. quart., 43 (1990) countries the difficult trade-off between space- heating fuel efficiency, ventilation and indoor air quality. This review focuses on indoor combustion-derived pollutants, principally because comparatively little work has been published on the other three cat- egories in developing countries. Although the main purpose is to review the available epidemiological studies, the relevant sources and pollutants will be briefly discussed first. Indoor sources Due to the generally higher ventilation rates, ETS has not been singled out in developing-country studies to date. With the alarming increase in smok- ing rates and the changing housing stock in some developing countries, however, such research may well be warranted. Although such minor combustion sources as incense and mosquito coils have occa- sionally been examined, most efforts have con- centrated on cooking and heating stoves. After tobacco, the most common indoor pollution source of concern in developed-country studies has been gas cookstoves. In the global context, however, gas stoves are near the upper end of an evolution in the quality of household fuels, sometimes called the energy ladder (2). Although dependent on local con- ditions, at the lowest rung might lie those house- holds that rely on dried animal dung and scavenged twigs and grass for cooking. The next step might be crop residues, followed by wood, and then charcoal. The first non-biomass fuel might be kerosene, as in India, or coal, as in China. Highest on the ladder lie bottled and piped gases and, for some rich com- munities, electricity. In general, each step involves an increase in the cooking system's levels of technology, cleanliness, efficiency and cost. On a global scale, it is estimated that more than half of the world's households cook daily with un- processed solid fuels, i.e. biofuels or coal. An un- known, but significant, proportion of this activity takes place in conditions where much of the air- borne effluent is released into the living area. Although ventilation rates are often relatively high, the emission factors for such fuels are so great that indoor concentrations and exposures can be quite significant. Compared with gas stoves, for example, even stoves using one of the cleaner biofuels, wood, typically release 50 times more pollution in cooking an equivalent meal (2). · Although less prevalent, space heating with un- processed solid fuels can also lead to high indoor air-pollution levels. For part of the year in the high- altitude areas of developing countries in all four major regions (Asia, Africa, Latin America and Oceania), unvented space heating with biomass fuels is common. In addition, in much of temperate China, - 128 - significant indoor concentrations result because of the high emission rates of coal fuels in simple stoves, although the pollution is less serious with venting, as is the case in most Chinese cities. It is impossible at this point to derive a reliable estimate of the total developing-country population exposed to what might be termed "excessive" in- door concentrations, partly because there are few internationally-recognized standards for household concentrations. Assuming that they would be at least as stringent as outdoor standards, the total number of people exposed clearly rivals or exceeds that exposed to excessive ambient concentrations in all the world's cities,• i.e. several hundred million. Given the greater time spent indoors and the known magnitude of concentrations in many situations however, the total global dose-equivalent (amount actually inhaled) is probably higher for some impor- tant pollutants, e.g. particulates (3). Pinpointing the actual sources of exposure is further complicated in many urban and even some rural areas, because indoor concentrations are influenced by outside as well as inside sources. Pollutants In spite of the long history of fascination with fire and hundreds of years of scientific effort, the ability to predict the detailed behaviour of combustion in any but the simplest configurations of fuel and air has eluded us. Nevertheless, empirical evidence has shown that unprocessed solid fuels produce hun- dreds of chemical compounds because of the incom- plete combustion that occurs under the operating conditions of simple cookstoves, which are often little more than three-rocks or a small open-ended box of clay. Both coal and biomass act this way, although the mix of compounds from each is signifi- cantly different. Unlike coal, biomass generally con- tains few intrinsic contaminants, such as sulfur, trace metals and ash, and thus can, under proper con- ditions, be burned with no releases other than the products of complete combustion, i.e. carbon dioxide and water. Unfortunately, it has turned out to be difficult to reliably create these conditions in modestly-priced household devices. As with tobacco smoke, which is also the result of open biomass combustion, such cookstove smoke is a complicated mixture of aerosol (droplets and solid particles) and gases. Biomass smoke contains sig- nificant amounts of several of the important pol- lutants, for which most countries have set standards: carbon monoxide, particulates, hydrocarbons and, to a lesser extent, nitrogen oxides. Perhaps more im- portant however is that the aerosol contains many organic compounds that are thought to be toxic, carcinogenic, mutagenic or otherwise worrisome (4). Coal smoke contains all this plus additional pol- lutants, e.g. sulfur oxides, inorganic ash particles and heavy metals such as lead. It is difficult to make generalizations about the relative amounts of each kind of pollutant however, because emissions can change with relatively small changes in fuel quality, configuration and combustion, and comparatively little detailed research has focused on such con- ditions. I For a review of exposure studies related to biomass, see reference (4); for those examining coal, see reference (6). Epidemiology Although, by international comparison, the ex- posures received in households in many developing countries are large, epidemiological studies are still difficult. Expected short-term (acute) effects, such as acute respiratory infections (ARI), have many other important risk factors, some of which correlate with pollution exposures. Long-latency effects, such as cancer and chronic obstructive lung disease (COLD), also involve other risk factors. Moreover, it is dif- ficult to associate these effects with previous ex- posures because of the long interval between these exposures and the onset of illness (the appearance of symptoms) (5). Physiological and biochemical markers, such as lung function and immunological status, are only indicators of the particular path- ology. For the sake of conciseness this review focuses on developing-country studies published since 1980 and which include a health-related endpoint. Most of the early studies are cited, but not discussed in detail, and purely environmental studies are only men- tioned occasionally, i.e. those that looked only at exposures.I The studies have been divided according to general category of source, biomass or coal, and the health-related endpoint being examined. Most have been done either in South Asia with rural populations using biomass, or in China with urban groups using coal. Biomass- South Asia, Africa, Guatemala and Papua New Guinea The vast majority of rural inhabitants in these parts of the world use unprocessed biomass fuels. Although significant in some communities, the utili- zation of these fuels in cities has been dropping as households move up the energy ladder, i.e. to ker- osene and bottled gas. Proportionately, household coal use is minor, but it involves tens of millions of people and seems to be on the increase as the availability of biomass fuels dwindles and petroleum fuels remain relatively expensive. Carboxyhaemoglobin (HbCO). In a study of carbon monoxide (CO) concentrations in 180 biomass- burning kitchens of two Guatemalan villages at dif- ferent elevations (250 m and 1 350 m), Dary et al. (7) also determined HbCO blood concentrations in women. They found lower HbCO levels in women cooking in well-ventilated kitchens, but no difference in concentrations between houses having similar ventilation but situated at different altitudes. In both well-ventilated and poorly-ventilated kit- chens, however, there were higher HbCO levels in the higher-altitude village, with blood reaching 3.5% HbCO in some women as a result of concentrations of 30-50 ppm during the cooking period. In other words, at the same CO concentration, exposures at higher elevations produced higher HbCO levels, a finding consistent with known physiological mech- anisms and a significant consideration in those many high-altitude areas of the developing world where solid fuels are burned for space heating without ventilation. Lung function. To test for the effect of domestic smoke, Pandey et al. (8) used a dry portable spiro- meter to measure lung function in a random sample of 150 women aged 30-44 years living in a rural area on the outskirts of Kathmandu valley at an elevation Rapp. trimest. statist. sanit. mond., 43 ( 1990) - 129 - of 1 600 m. The area is totally free from industrial and atmospheric pollution and women spend con- siderable time near the hearth, which is used for both cooking and heating. Smoke concentrations are high because dwellings are ill-ventilated and without chimneys. The selected sample comprised 6 groups of 25 subjects in each of 3 self-reported durations of daily smoke exposure for both smokers and non- smokers. All the spirometric tests (FVC, FEVI and FMEF 25-75) were performed in the standard way recommended by the American Thoracic Society. Variations in age, height, arm-span and weight be- tween the three levels of exposure to domestic smoke in both the smokers and non-smokers were compared but no significant differences were found. There was a fall of mean FVC, FEVI and FMEF 25-75, however, as duration of exposure increased. This decline was found to be statistically significant among the smokers but not among the non- smokers. Similar results were found after adjusting for age and height. Respiratory symptoms. In Ahmedabad, western India, studies of the incidence of cough, cough with expectoration, dyspnoea and lung abnormalities found a statistically significant (P < 0.05) higher inci- dence among women cooking with smoky fuels (9-11). The investigators also noted complaints by the women about the irritating effect of the smoke, particularly on the eyes. In a three-year prospective study of clinical morbid- ity and air pollution in three areas of Bombay city and in one rural area nearby, Kamat et al. (12) found that the rural subjects were smaller in size, had lower lung function for their body stature and a higher incidence of cough and dyspnoea. Many of the respiratory symptoms exhibited by the rural population were similar to those of the moderately or severely polluted urban areas rather than the non-polluted areas. Use of wood for kitchen fuel is one of the explanations given by the authors for this finding. In a study conducted in Ladakh, (Keith Ball, personal communication) a high prevalence of chronic respir- atory symptoms was noted, particularly among women, which rises steeply with age to affect the majority of older people. About one-quarter of men and one-third of women over 50 have lung functions suggestive of obstructive airways disease, and this becomes worse in winter. Tobacco smoking has so far developed only in about 20% of the men but is virtually unseen in women. It has been suggested that domestic pollution, which is significant through- out much of the year in this high-altitude area, appears to play a part in creating this pattern. Chronic obstructive lung disease. In a survey carried out in a housing complex in Chandigarh (India) and in a nearby village, a study was made of the records of 2 180 women aged over 20, working with different domestic fuels (13). Sixty-six women (3%) had symp- toms of chronic bronchitis, the highest number being in those who used the chulla (traditional un- vented cookstove) for cooking. The cooking fuels were categorized in the following groups: I-liquefied petroleum gas (LPG); II-kerosene; Ill-coal; and IV-chulla (wood and cow dung). Five per cent of the subjects in group IV had chronic bronchitis compar- ed to 1.5% and 1.3% in groups I and II, respectively. This difference was statistically significant (P < 0.01 ). Analysis of peak expiratory flow rate showed it to be low in subjects in group IV, but after adjusting for Wld hlth statist. quart., 43 (19901 age and height, no difference was found between the four groups. Hospital statistics in Nepal showed an unusually high proportion (46%) of cor pulmonale in hospital cardiac admissions (14), indicating the need for field studies to determine the distribution and magnitude of chronic bronchitis and chronic cor pulmonale, and to identify the factors responsible (15). Four sites were selected for this purpose: urban Kathmandu, nearby villages from Kathmandu Valley in the hills region, a village district in the plains region (terai) near India, and a district in the mountains. A survey of the prevalence of chronic bronchitis in the area of the hill villages was used to determine the required sample size in other areas. In the plains village, a simple random sampling method was used to select the sample. Two-stage random sampling was done in the urban areas and in the mountain district. In all the study areas, houses were ill-ventilated and without chimneys. Most of the houses had only two or three rooms, but many had two floors. Cooking was done on traditional stoves in a corner of one of the ground-floor rooms in the morning and evening (the villagers take only two main meals a day). Most houses in the hills and mountains were customarily heated by means of an open fire in a fireplace known as an ageno dug into the floor of a ground- floor room. Members of the family sit around the ageno in the morning and evening to keep them- selves warm. The ageno is also used for preparing cattle fodder in the morning and snacks for family members. Except in urban Kathmandu, there was no general industrial air pollution. A questionnaire, modified slightly for different sites, recorded demographic information and exposure- related information such as smoking habits, location of kitchen, type of fuel used for cooking and heating and average time spent daily near the fireplace, along with the information required to diagnose chronic bronchitis according to the British Medical Research Council criteria. Emphysema and chronic cor pulmonale were diagnosed according to the WHO Expert Committee criteria. The individuals were classified as smokers, past smokers, or non- smokers and into various strata according to self- reported number of hours spent daily near the stove where, presumably, the highest smoke con- centrations are found. This study (15) found a high prevalence (crude and age-adjusted) of crude chronic bronchitis and cor pulmonale (10-30%). A striking feature was the sim- ilarity of rates in men and women. The difference in prevalence rate between the sexes was not statisti- cally significant in any study area (P > 0.10 - > 0.98). This conflicts with most other studies, which have shown a lower prevalence rate among women. Tobacco smoking is common in both men and women in all but the urban area, where the smoking rate among women was only 14%. In all areas, however, women were light smokers (less than 10 cigarettes or equivalent per day). Heavy smoking among men was much more prevalent and the difference was statistically significant in all three rural areas (P < 0.05 - < 0.001) although not in urban Kathmandu. This suggests that the high pre- valence of chronic bronchitis in women is primarily occupational since women engaged in cooking meals for the family are exposed for proportionately longer hours at higher concentrations of domestic smoke pollution than men. The difference in smoke - 130 - exposure hours was significantly higher (P < 0.001) in women in all the study areas. In the hill villages, a statistically significant (P < 0.001) positive correlation was found between prevalence of chronic bronchitis and exposure to domestic smoke pollution in both smokers and non- smokers (Fig. 1). This supports earlier studies in India and Papua New Guinea (16-19) which suggest- ed the possibility of a relationship between chronic bronchitis and domestic smoke pojlution. The trend of increasing prevalence rates with increasing smoke exposure persisted even after elimination of the age effect (20). A similar statistically significant correla - tion was also found in other study sites. In the mountain village, it was not possible to compare the prevalence between exposed and non-exposed groups, because everybody was exposed to do- mestic smoke from space heating. Unlike in other areas, however, many were exposed more than 8 hours per day. This provided an opportunity to test for an effect at longer hours of exposure. A signifi- cant correlation between bronchitis and smoke ex- posure was noted also beyond 8 hours of exposure among the smokers of both sexes. Hence, there seems to be a dose-response relationship between domestic smoke pollution and chronic bronchitis. All the cases of chronic car pulmonale identified in this study were complications of chronic bronchitis. Although most of the women smokers were light o World Health Organization. Acute respiratory infections in children. Dossier ARl -5 of the Respiratory Infections Programme and th e Programme Support Servi ce. Geneva, 1987. smokers, the study recorded similarity of prevalence of chronic car pulmonale in the two sexes. This suggests that domestic smoke pollution is an impor- tant factor in producing car pulmonale. Childhood acute respiratory in fections (AR/). Be- tween 4 and 5 million children < 5 die in the world each year from ARI, mostly in developing countries. Most ARI studies have focused on the important issues of microbial causative agents, case- management with antibiotics, and effectiveness of vaccination . Research on risk factors has been judged to be of lesser importance in spite of the dominance of risk-factor reduction in the history of ARI control in the developed countries. Domestic smoke pollution has been assumed to be an impor- tant risk factor for ARI in many parts of the develop- ing world ,9 but there have been few studies, par- ticulary of the relationship with- pneumonia, which is nearly always involved in fatal cases. In a study carried out before 1980 in Papua New Guinea, Anderson (2 1) failed to find an effect of environmental differences, including household smoke, on acute respi ratory symptoms among schoolchildren. In a study of younger children in South Africa, Kossove (22) found that 70% of Zulu infants with respi ratory symptoms lived in households with cookfire smoke, compared to 33% of a partially matched group who did not have such symptoms. This was significant at the P < 0.005 level, although there was no significant difference in the maternally- reported exposure times for the groups. No con- founding factors were investigated . FIG. I PREVALENCE OF CHRONIC BRONCHITIS IN NEPALESE HILL VILLAGES ACCORDING TO SELF-REPORTED DAILY TIME SPENT NEAR COOKING/HEATING STOVE, AROUND 1984 a PREVALENCE DE LA BRONCHITE CHRONIQUE DANS LES VILLAGES NEPALAIS DE MONTAGNE EN FONCTION DU TEMPS PASSE PAR JOUR A PROXIMITE DU POELE (CUISINE/CHAUFFAGE), VERS 1984 a 35 30 l 25 " " c: " ~ 20 ~ 0.. I l " " c: " <ii > " 0:: 15 10 o I I I I I I I I I I I I I I I I I I 0-0.9 1-1.9 2-3.9 4 + Hours near stove per day - Heures passees pres du poele par jour Non-smokers - Non fumeurs • Men - Hammes • Women - Femmes Smokers - Fumeurs - - - - Men - Hammes __ Women - Femmes a The strong associations are statistically significant (p<0.001) for all smokers and for women non-smokers (20) - Une association nette est statistiquement sig nif icative (p<0,001) pour !'ensemble des fumeurs et pour Jes femmes qui ne fument pas (20). Rapp. trimest. statist. sanit. mond., 43 ( 19901 - 131 - As part of WHO-sponsored indoor air pollution studies in Kenya and the Gambia (23,h), ARI inci- dence in Kenyan children < 5 was monitored bi- weekly in 36 houses for 42 weeks. During this period, stationary indoor 24-hour monitoring was conducted twice in each house, for nitrogen dioxide and suspended particulates. Although respirable par- ticulate concentrations were quite high (mean of 1.4 mg/m3), no correlation was found between ARI incidence and indoor concentrations of either pol- lutant, or with exposures of the children, as deter- mined from the concentrations and time spans con- structed by interviewing the mothers. Moreover, the measured concentrations were quite homogeneously distributed among the households. Indeed, the intra-household variation between the two measurement days was greater than the inter- household variation, although households were randomly chosen from four groups stratified on the basis of house characteristics, roof type and kitchen arrangement, thought to influence ventilation, and thus indoor air quality. As the authors point out, given the apparent lack of natural variation in ex- posure among households, it would be quite difficult to establish smoke as a risk factor with this type of study design. In the Gambia, on the other hand, where indoor woodsmoke levels were also high, but homo- geneously distributed, there seemed to be significant differences in exposures because some young children are carried on their mother's back during cooking. Based on a multiple logistic regression, Campbell et al. (24) report an adjusted odds ratio of 2.8 for episodes of breathing difficulty among 280 children carried by their mothers. The only other significant risk factor was paternal smoking. A semi-quantitative epidemiological study was con- ducted in Nepal by Pandey et al. (25), to assess the relation between maternally-reported hours spent per day near the hearth by infants and children < 2, and episodes of life-threatening, moderate and severe ARI. In the study area, a hill region of Nepal, traditional stoves burning biomass fuels, mainly wood and crop residues, are used for cooking and heating in unventilated houses with no chimneys. The area is totally free from industrial pollution. The first substudy was conducted during a six- month period in 1984. A statistically significant (P < 0.01 l association was found when the exposed group was compared with the non-exposed in the two age strata i.e. < 1 year (Fig. 2a) and 1-2 years. In order to confirm and validate the findings, an addi- tional three-month substudy was conducted in the same population. This substudy employed separate groups of observers to determine ARI episodes and smoke exposure times, thereby assuring a tireater degree of independence in observation of the two variables. A statistically significant (P < 0.01) associ- ation was found in this second substudy as well (Fig. 2b). There were no obvious confounding factors that might account for these findings. Socioeconomic condition is not an important factor as more than 98% of the population live within a subsistence 'World Health Organization. Indoor air quality in the Basse area, the Gambia. (Document GEMS/HEALS WHO/PEP/88.3, 1988). ; Wang, J. & Chen, B. H. [The influence of air pollution an immunological parameters of children]. Shanghai Medical Uni- versity, 1987. (Graduate thesis). (In Chinese). Wld hlth statist. quart., 43 (1990) economy, and nearly all the houses are similar in construction. This preliminary study seems to suggest that domestic smoke pollution is a risk factor of ARI. As with all available studies, however, there were insufficient resources to fully monitor exposure, measure and correct for confounding vari- ables, and to conduct the many expensive and time- consuming quality-control procedures needed for full-scale modern environmental epidemiology in- vestigations. Coal - East Asia In China, coal is and will remain for decades the main energy source. Particulates and sulfur oxides emitted from coal combustion are two of the main urban atmospheric pollutants. In Shanghai, some 5% of total coal consumption is used for domestic cook- ing, making it the most important urban cooking fuel. For example, until 1987 Tianjin was the only city where all the families used gas stoves and in Beijing, only 70% of the families used gas or lique- fied petroleum gas, while in Shanghai about half of the population use coal-cake stoves. Little indoor air quality research has been done in rural areas where most of the population uses crop residues and wood. Research on indoor pollution and its health effects was initiated in China towards the end of the 1970s and over 100 papers have been published (6) on the subject. A series of epidemiological surveys have been carried out to evaluate the effect of indoor pollution on human health. In contrast to studies in developed countries, the gas-user group is usually considered as the unexposed control group, and the coal-user group as the exposure group. Immunological function. The influence of indoor air pollution on immunological function was investigat- ed in groups of primary-school students aged 10-13 exposed (or not) to coal smoke at home and at school (26). Compared to the control group, the amount of SlgA in saliva of the exposure group decreased by 32.5%, and the activity of lysoenzyme by 17.3%. There were no differences in distribution of age, sex and ETS exposure between the groups (P values, all > 0.05). Wang & Chen monitored the immunological para- meters of primary-school pupils from families using different fuels for cooking.; The results were statistic- ally significant (< 0.05) for white blood-cell count (mean reduction: 13%), saliva lysoenzyme (13%), transfer of lymphocyte (50%), lgG (29%), lgM (28%) and lgA (54%), but not in the PHA skin test. Carboxyhaemoglobin (HbCOJ. The blood HbCO con- centration in women using coal stoves was 2.1 :±:: 0.9%, compared to those using gas stoves at 1.4 :±:: 1.0% (P < 0.01). Women using coal stoves without chimneys had 4.3%, while those using chim- neys haQ 3.0% (27). A study by Hu & Liu of 450 suburban schoolchildren < 13 near Beijing (28) found significantly higher HbCO levels in those living in traditional houS'ing where pit stoves were used for heating and cooking, compared to those residing in modern buildings with stoves and chimneys. At 4.1% and 2.8%, how- ever, both groups were high. The authors compare - 132 - FIG.2 ACUTE RESPIRATORY INFECTIONS IN INFANTS (0-1) IN THE HILL REGION OF NEPAL LISTED BY SEVERITY AND MATERNALLY REPORTED DAILY TIME NEAR STOVE a INFECTIONS RESPIRATOIRES AIGUE~ CHEZ LES NOURRIS~ONS (0-1 AN) DANS LA REGIQN DE MONTAGNE DU NEPAL, ~LASSEES SELON LA GRAVITE ET SELON LE TJ:MPS PASSE PRES DU POl:LE CHAQUE JOUR AU DIRE DE LA MERE a A Based on six months' data oolleded in 1984 (233 infants) - Sur la base des donnees rerueillies en 1984 sur une periode de six rnois (233 nourrissons) 2.00 1.75 E "' c: 1.50 "' t;; a. "' ~ 1.25 51 ·a ~ 1.00 E "' £; 0.75 "' ~ 0.50 0 "' ·a w 0.25 0-0.9 1-1.9 2-3.9 4+ Reported hours per day near stove - Heures passees par jour a proximite du poele B. Based on three months' data collected in 1989, with the same population (247 infants) but with separate teams diagnosing ARI and determining exposure - Sur la base des donnees recueillies en 1989 sur une periode de trois mois dans la meme population (247 nourrissons), des equipes distinctes etant chargees de diagnostiquer les IRA et de determiner !'exposition 2.00 1.75 E "' ~ 1.50 t;; a. ~ 1.25 ~ ·a w 1.00 I E "' :S 0.75 ~ ~ 0.50 0 "' ·a w 0.25 0-0.9 1-1.9 2-3.9 4+ Reported hours per day near stove - Heures passees par jour a proximite du poele Mild cases (grade I) - Cas benins (degre I) D Moderate cases (grade II) - Cas moderes (degre II) • Severe cases (grades Ill and IV) - Cas graves (degre Ill et IV) a The trends are significant in both cases (25) - Les tendances sont significatives dans les deux cas (25). these results with an earlier urban study,i where HbCO levels of 4.3%, 3.0% and 1.1% were found for residents in homes heated by stoves without flues, by stoves with flues, and by circulating hot water respectively. This study also found other effects. There were significant differences in the incidence of lower respiratory diseases among girls (RR= 3, p. < 0.05) but not for boys nor for upper respiratory diseases. Small or non-significant differences were found for immune and eye responses and lung function, after correcting for height, age, weight, family income and parental smoking (28). i Chang X. M. {An epidemiological study on indoor air pollution). Second National Conference on Environmental Health, October 1984, Nanjing. (In Chinese). Although small-scale combustion of all coal and biomass fuel is often associated with high CO emissions, poor combustion of all unprocessed bio- mass and high-volatile coals also emits large amounts of hydrocarbon aerosol and gas, e.g. aldehydes. Although these may have long-term health effects because of the intense irritation they cause, they act in the short term to warn when CO concentrations approach acutely toxic levels. In other words, it is extremely rare for someone to die from acute CO poisoning from such fuels. Even though measured CO concentrations can sometimes approach dangerous levels, the irritation due to hydrocarbons will awaken householders and prompt them to escape or open a window (4). Rapp. trimest. statist. sanit. mond., 43 (1990) - 133 - Unfortunately, this is not the case with the three principal low-volatile solid fuels: charcoal, naturally low-volatile coal such as anthracite, and processed (devolatilized) coal. These may release no more CO than the high-volatile solid fuels, but they do so with little of the irritating hydrocarbons that provide warning. As a result apparently, hundreds if not thousands of people die every year in northern China, Korea, the Islamic Republic of Iran and other developing countries where overnight space heating is provided by such fuels (mainly coal), along with excessive leakage of combustion gases into living areas, compounded by poor household ventilation. A study in Korea, for example, has shown a strong seasonal variation in hospital admissions for CO poisoning, peaking during the cold season where heating is used (29). Mortality is not the only effect- high, but sublethal, exposures can cause serious poisoning. Lung function. Li & Hongk monitored the lung func- tions (FVC, FEV1, FEV1%) of 213 women aged 50-70 years, including 167 women using coal and 46 using gas for cooking. Even after adjustment for potential confounding factors, i.e. age, body weight, height and ETS, there was a statistically significant differ- ence between the coal group and the gas group in all lung functions tested. In another study, the FEV1/FVC of persons using coal stoves was lower than those using gas stoves, while the reverse relationship was found for MEF25, MEF50 and MEF75. This was possibly due to higher NOx concentrations in homes using gas. In Xuanwei County, Yunnan Province (southern China), the airway resistance of farmers and school- children in homes where poor-grade bituminous coal was used for cooking was higher than in homes where wood and straw were used for cooking (P < 0.001) (30). Respiratory symptoms and non-cancer respiratory diseases. There seems to be an increase in respira- tory symptoms and prevalence of respiratory dis- • Li, L. & Hong, C. J. [Kitchen air pollution and its health effects]. Shanghai Medical University, 1988. (Graduate thesis). (In Chinese). ' Lin, H. S. et al. [Relation of air pollution and human health in a higher density residential area of Shanghai]. Symposium on Environmental Health, Taiyuan, 1983. (In Chinese). m Hong, C. J. et al. [Studies on air pollution and its impact on health in Shanghai]. Shanghai Medical University, 1989. (In Chinese). eases in coal-using families. A study' was conducted in two areas of Shanghai with populations of about 77 OOO and 17 OOO respectively. Two samples were selected, using cluster sampling: 1 316 using coal- cake stoves and 721 using gas stoves. The respira- tory symptoms and prevalence of respiratory dis- eases were found to be 79% and 44%, respectively, for people using coal stoves and gas stoves (P < 0.05). The prevalence of respiratory diseases among housewives aged 40-65 are 50% and 40%, re- spectively, for coal-using homes and gas-using homes, where P < 0.05, and relative risk for coal- using housewives is 1.9 (28). The prevalence of common cold among persons using stoves with or without chimneys were 22% and 52% respectively; RR= 2.0 (without chimney), attributable risk - 25% (31). The prevalence of chronic bronchitis, em- physema, cough, productive cough and breathless- ness in women using coal stoves were significantly higher than among those using gas stoves. There were no significant differences in the prevalence of coronary heart disease, hypertension and a tight feeling in the chest between the coal- and gas-using groups.1 Another study,m carried out in Shanghai on 12 037 people, using the American Thoracic Society DLD-78 questionnaire, found that in areas where atmos- pheric pollution was light, the prevalence of COLD was 9.6% in the gas-user group and 15.6% in the coal-user group. In heavily-polluted areas, these prev- alences were 12.8% and 17.3%, respectively. Multi- variate analysis indicated that indoor air pollution had a stronger impact than outdoor air pollution. Another survey of respiratory symptoms and dis- eases was carried out among retired women and nonemployed women aged 45 years and over (32). Of the 393 women who responded to the question- naire, 186 used gas for cooking and 207 coal. Indoor pollutant concentrations typically found in each type of home were determined, based on small represen- tative samples. The results are listed in Table 1. In order to determine whether factors other than indoor pollution (such as age, living standard, edu- cation, ETS, kitchen style and living space per per- son) were important influences, a multifactor logistic regression analysis was conducted. The result showed that domestic smoke was an important con- tributing factor to chronic bronchitis, emphysema, TABLE 1. PREVALENCE OF RESPIRATORY SYMPTOMS AND DISEASES IN WOMEN USING DIFFERENT FUELS FOR COOKING, 1985 TABLEAU 1. PREVALENCE DES SVMPT0MES ET MALADIES RESPIRATOIRES CHEZ LES FEMMES UTILISANT DIFFERENTS TYPES DE COMBUSTIBLES POUR LA CUISINE, 1985 Disease or symptom Maladieou symptome Asthma - Asthme . . . . . . . . . . . . . Chronic bronchitis - Bronchite chronique . Bronchodilatation . . . . . . . . . . . . . . Emphysema - Emphyseme . . . . . . . . Coronary heart disease - Cardiopathie coronarienne . . . . . . . . . . . . Hypertension . . . . . . . . . . . . . . . Cough - Toux ............. . Expectoration . . . . . . . . . . . . . . . Chest tightness - Gene respiratoire . . . Shortness of breath - Essoufflement Source: Reference (32) - Reference (32). Wld hlth statist. quart .. 43 (1990) Gas user group (186casesl Groupe utilisant le gaz (186cas) Number - Nombre % 6 3.3 22 11.8 3 1.64 4 2.15 27 14.5 60 32.3 33 17.7 24 12.9 55 29.6 18 9.68 Coal user group (207 cases! Groupe utilisant le charbon (207 cas) Number - Nombre % 15 7.2 51 24.6 13 6.23 21 10.1 25 12.1 66 31.3 83 40.1 53 25.6 77 37.2 53 25.6 X'(M-H) Number - Nombre 2.2 7.5 4.6 7.3 0.69 0.03 19.1 7.5 2.1 13.4 p % >0.05 <0.01 <0.05 <0.01 >0.05 >0.05 <0.01 <0.01 >0.05 <0.001 - 134 - bronchodilatation, cough, expectoration and short- ness of breath. These conditions were 1.8-4.5 times more common among the group exposed to coal smoke. No significant difference was found for coronary heart disease. In Hong Kong, Koo et al. (33) found small (10-20%) but statistically significant differences in nitrogen dioxide exposure levels as well as reported respira- tory symptoms (allergic rhinitis and bronchitis) among non-smoking mothers in homes using piped gas (lower exposure) for cooking, compared to those using kerosene or bottled gas. A similar effect was found for incense-burning and kitchen ventilation, but not for mosquito coils or ETS. No effect was found in children. Cancer. Although the lung-cancer risks of occu- pational exposures as well as active smoking, and to a lesser extent ETS, have been extensively document- ed, the relation between lung cancer and ambient air pollution remains uncertain. Indoor pollution, however, is now starting to be linked to lung cancer, especially among women cooking over coal-fired stoves, although their level of cigarette smoking and occupational exposure is much lower than men's. Xu et al. (34) analysed female lung-cancer preva- lence in 16 Chinese metropolitan areas, each with a population over 500 OOO and located at various lati- tudes. They found that the correlation coefficients between female lung-cancer prevalence and latitude were statistically significant. The higher the latitude, the higher the lung cancer prevalence. Yang et al." compared several factors with the high and low lung-cancer prevalence regions in Wuhan (standardized rates: 26 vs. 12 per 100 OOO) to find that there were significant differences in stove types and density, sulfur dioxide and nitrogen oxides, but not for carbon monoxide. In Guangzhou, there were 349 housewives among 1 243 female lung-cancer deaths (35). These women (28%) had no previous occupational exposure, 40% of them were non-smokers, and none seemed to be exposed to significant outdoor air pollution. The epidemiological survey showed that female lung- cancer occurrence was related to exposure to coal combustion and cooking (36). The investigators monitored kitchens for suspended particulates, de- posited dust, and benzo(a)pyrene in deposited dust. They found (as expected) that concentrations were much higher in kitchens using coal compared to those using liquefied petroleum gas for cooking. Based on a case-control analysis of 662 deaths (446 males and 216 females), the lung-cancer odds ratio (OR) of smoking and cooking was analysed and is presented in Table 2 (36). Exposure to cooking was found to be an important contributing factor to female lung cancer, but not to male lung cancer. The combination of smoking and cooking exposure has a synergistic effect on the prevalence of lung cancer in both males and females. TABLE 2. LUNG-CANCER ODDS RATIO (ORI OF SMOKING AND COOKING - CASE-CONTROL ANAL VSIS, 662 PAIRS, 1985 TABLEAU 2. ODDS RATIO (ORI DE CANCER DU POUMON PAR EXPOSITION AU TABAC ET AUX FUM~ES DE CUISINE - ANALYSE CAS-T~MOINS, 662 PAIRES, 1985 ·M p F p OR OR (1) S+NS+E+NE• . . . . . . 1.0 1.0 (2) NS+NE ... 0.31 >0.05 0.31 >0.05 (3) S+NE 4.4 <0.01 0.6 >0.05 (4) NS+E .......... 1.4 >0.05 3.1 <0.05 (5) S+E ........... 4.2 <0.01 5.8 <0.01 a S = smokers - fumeurs. NS = non-smokers - non-fumeurs. E = exposed (coal stove) - exposes (cuisiniere ii charbon). NE = non-exposed (gas stove) - non exposes (cuisiniere ii gaz). Source: Reference (36) - Reference (36). showed that a high coal-consumption index (HCCI, OR= 10.6), indoor smog pollution in winter (OR= 15.2) and low ceiling height of the living room (OR= 12.5) were the main risk factors for female lung adenocarcinoma. It was further confirmed by indoor air monitoring that the average air con- centrations of total suspended particulates (TSP) and benzo(a)pyrene in bedrooms in winter were 4.4 and 27 times higher than the respective maximum allow- able air concentrations in China. No significant cor- relation was found between female lung adeno- carcinoma, and cigarette smoking and ETS. In another Xuanwei study, the standardized lung- cancer mortality in 1973-1975 was 26 per 100 OOO, higher than the average urban lung-cancer mortality in China. In Xuanwei, lung-cancer prevalence was different in various regions, the highest being over 150 per 100 OOO, while the lowest was under 1 per 100 OOO, a difference of about a factor of 200. Although rare today, in the recent past high-tar bituminous coal burned in open pits was the main domestic fuel used in high lung-cancer prevalence regions, while anthracite coal and wood were used in low-prevalence regions. The prevalence of lung cancer in farmers was higher than among people in other occupations and there was little difference between sexes. This evidence suggests the presence of a strong carcinogenic factor linked to lung cancer in nonemployed women (38). Research involving indoor air monitoring, epidemiological studies and animal exposure tests in situ have been conducted by the Chinese Academy of Preventive Medical Sciences and local public health and anti-epidemic stations with assistance from the United States En- vironmental Protection Agency (39, 40). In Hong Kong, however, no signlficant association was found between lung cancer and cooking fuels (41) or ETS (42). The indoor-pollutant concentrations were much less than those found in the above- mentioned lung-cancer studies in China. Although there have been reports linking domestic smoke with the high rates of nasopharyngeal cancer among southern Chinese populations (43, 44) and in other groups, e.g. Kenya (45, 46), most reviews have con- Another matched case-control study and measure- .. eluded that the risk must be small or non-existent ment of indoor air pollution was carried out in a compared to risk factors such as diet (47). northern Chinese city (Harbin) to analyse risk factors for female lung adenocarcinoma (37). The results "Yang, X. et al. {The relation of indoor air pollution and human health - a primary study]. Workshop on Chinese environmental health, 1983. (In Chinese). Fluorosis. Increased fluorosis induced by high con- centrations of fluoride in indoor air and food orig- inating from coal combustion has been reported recently (48-51). In the suburbs of Xiangtan (Hunan Province) the fluoride concentration in drinking- water was less than 1 mg/litre, but in coal it was Rapp. trimest. statist. sanit. mond., 43 (1990) - 135 - 190-330 mg/kg. Fluoride concentration originating from coal combustion was 0.018-0.066 mg/m3 in indoor air and 1.9-3.3 mg/kg in vegetables. The prevalence of dental fluorosis there was as high as 35% in residents of all ages. In a neighbouring area, the prevalence of dental fluorosis in a control group cooking with wood was only 4.9% (48). In the suburb of Luoyang (Henan Province) the fluoride concentration of coal was 370-640 mg/kg, and the indoor-air fluoride concentration was meas- ured at 0.016 mg/m3• The vegetable and grain fluoride concentration was 0.95-3.4 mg/kg. This in- creased to 2.7-14 mg/kg after a half-year storage period indoors. There were 1 486 persons suffering from fluorosis among 2 246 persons aged over 6. The prevalence rate was 66% (49). Increased preva- lence of fluorosis due to coal combustion in residents has also been found in Xianwen (Sichuan Province), Ankang (Shanxi Province), Benxi (Liaoning Province) and Huanggang (Hubei Province). Stroke. A group of 957 men were followed for 12 years in a study of risk factors for heart disease and stroke carried out in Shanghai (52). During the study period, 24 subjects died. In addition to high blood pressure, age and smoking, the study found ex- posure to household coal smoke to be a significant risk factor. A dose-response relationship was found with the relative risk of heavy exposure (only coal use) of nearly 11, after correction for the other major risk factors. The slight-exposure groups (using both gas and coal) had a relative risk of nearly 4. Conclusion Available evidence in the developing world would seem to argue strongly that indoor air pollution from cook-fire smoke is a risk factor for chronic lung disease in adults, particularly women, and acute respiratory disease in young children. In addition, coal smoke may be an important risk factor for lung cancer in women. The total number of people ex- posed to high concentrations of smoke is likely to be several hundred million. However, too few resources have been allocated to such studies for it to be possible to quantify these effects. Such efforts are needed, however, in order to determine the effectiveness of smoke-exposure reduction measures in comparison with other ap- proaches for preventing or mitigating these dis- eases. In particular, there is a need to link detailed epi- demiological study designs using internationally ac- cepted diagnostic procedures with detailed and sys- tematic exposure determinations. This is not easy to do with cancer and chronic obstructive lung disease, since exposure histories going back several decades must be established. One of the highest priorities for indoor air pollution epidemiology is to initiate a study of the risk factors for ARI, because of its importance as a cause of early childhood morbidity and mortality in many developing countries, and in view of the high indoor exposures that seem to prevail in many areas. A longitudinal intervention study lasting at least two years and including all seasons before and after intervention would most likely lead to results that are scientifically sound and useful for designing large-scale intervention programmes (53). A growing trend among those concerned with en- vironmental epidemiology is the realization that, in order to understand the impact of pollution on health, it is necessary to monitor dwelling places rather than just the ambient environment. Total ex- posure assessment is the term for such efforts, which are beginning to reveal important sources and control measures different from those on which past efforts have focused (4). Indeed, from a global stand- point, the present balance of research, monitoring and control efforts in air pollution should perhaps be shifted towards households using unprocessed solid fuels, which is where the largest number of suscept- ible persons can be found. SUMMARY Of the four principal categories of indoor pollution (combustion products, chemicals, radon and bio- logicals), research in developing countries has focused on combustion-generated pollutants, and principally those from solid-fuel-fired cooking and heating stoves. Such stoves are used in more than half the world's households and have been shown in many locations to produce high indoor con- centrations of particulates, carbon monoxide and other combustion-related pollutants. Although the proportion of all such household stoves that are used in poorly ventilated situations is uncertain, the total population exposed to excessive concentrations is potentially high, probably several hundred million. A number of studies were carried out in the 1980s to discover the health effects of such stove exposures. The majority of such studies were done in South Asia in homes burning biomass fuels or in China with coal-burning homes, although a sprinkling of studies examining biomass-burning have been done in Oceania, Latin America and Africa. W/d hlth statist. quart., 43 (1990) Of the health effects that might be expected from such exposures, little, if any, work seems to have been done on low birthweight and eye problems, although there are anecdotal accounts making the connection. Decreased lung function has been noted in Nepali women reporting more time spent near the stove as it has for Chinese women using coal stoves as compared to those using gas stoves. Respiratory distress symptoms have been associated with use of smoky fuels in West India, Ladakh and in several Chinese studies among different age groups, some with large population samples. Acute respiratory infection in children, one of the chief causes of infant and childhood mortality, has been associated with Nepali household-smoke exposures. Studies of chronic disease endpoints are difficult because of the need to construct exposure histories over long periods. Nevertheless, chronic obstructive lung disease has been associated with the daily time spent near the stove for Nepali women and found to be elevated among coal-stove users compared to - 136 - gas-stove users in Shanghai. In contrast to early reports, there seems to be little or no risk of naso- pharyngeal cancer from cookstove smoke. Several studies in China, however, have found smoke to be a strong risk factor for lung cancer among non- smoking women. In addition, severe fluorosis has been observed in several parts of China where coal fluoride levels are high. The high indoor concentrations and the potentially large (also poor and thus vulnerable) populations at risk, argue for more resources to be devoted to . epidemiological studies of large numbers of exposed populations, using appropriate quality-control meas- ures in order to generate reliable quantitative dose- response information. In the meantime, however, enough is known to warrant household, community and government efforts to reduce exposures through education and introduction of improved stoves, cleaner fuels and enhanced ventilation. RESUME Pollution de l'air a l'interieur des habitations dans les pays en developpement Des quatre principaux types de pollution a l'interieur des habitations (produits de combustion, substances chimiques, radon et substances biologiques), la re- cherche dans les pays en developpement a privilegie les produits de combustion et notamment les pol- luants resultant de !'utilisation de combustibles soli- des pour la cuisson et le chauffage. Les poeles incrimines sont en service dans plus de la moitie des habitations du monde et ii a ete prouve en de nombreux endroits qu'ils etaient a l'origine de concentrations elevees de matieres particulaires, d'oxyde de carbone et d'autres polluants lies a la combustion. Bien que l'on ne connaisse pas avec precision la proportion de poeles ainsi utilises dans des locaux mal aeres, le nombre total de personnes exposees a des concentrations excessives de pol- luants est eleve, probablement plusieurs centaines de millions. Dans le courant des annees 80, plusieurs etudes ont ete consacrees aux effets sur la sante de !'exposition a ces produits de combustion. La plupart ont ete conduites en Asie meridionale dans des habitations ou sont utilises des combustibles tires de la bio- masse ou en Chine la ou l'on utilise du charbon, bien que certaines etudes sur !'utilisation de la bio- masse aient ete faites en Oceania, en Amerique latine et en Afrique. Parmi les effets de ce type de pollution qui peuvent etre redoutes, ii semble que l'on n'ait pratiquement pas ou pas du tout etudie l'insuffisance ponderale a la naissance et les problemes ophtalmologiques, bien que certains comptes rendus anecdotiques fas- sent la relation entre ces problemes et ce type de pollution. Une diminution de la fonction pulmonaire a ete relevee chez des femmes nepalaises ayant signale passer beaucoup de temps aupres du poele ainsi que chez des femmes chinoises qui utilisent des cuisinieres a charbon par rapport a celles qui utilisent des cuisinieres a gaz. Des sympt6mes de detresse respiratoire ont ete associes a !'utilisation de combustibles produisant de la fumee dans l'ouest de l'lnde, au Ladakh et a l'issue de plusieurs etudes chinoises portant sur differents groupes d'age, cer- taines aupres d'echantillons importants de popula- tion. Des cas d'infections respiratoires aigues chez les enfants, l'une des principales causes de mortalite infanto-juvenile, ont ete associes a !'exposition a la fumee dans des habitations nepalaises, d'apres le temps passe par les enfants aupres des poeles. Les etudes sur les maladies chroniques sont diffici- les car ii est necessaire d'etablir des antecedents d'exposition sur de longues periodes. Neanmoins, la maladie pulmonaire obstructive chronique a ete as- sociee au temps passe chaque jour aupres du poele pour des femmes nepalaises et son taux d'incidence a ete juge eleve parmi les utilisatrices de poeles a charbon compare aux utilisatrices de poeles a gaz a Shanghai. Contrairement a ce qui avait ete prece- demment indique, la fumee degagee par les cuisi- nieres ne presente, semble-t-il, que peu de risque de cancer du rhino-pharynx. Toutefois, plusieurs etudes menees en Chine ont montre que la fumee etait un facteur de risque important pour le cancer du pou- mon chez les femmes non-fumeuses. En outre, des cas graves de fluorose ont ete observes dans plu- sieurs regions de Chine ou les taux de fluorure dans le charbon sont eleves. Les concentrations elevees de polluants a l'interieur des habitations et !'importance des populations exposees (egalement pauvres, done vulnerables) jus- tifient que des ressources accrues soient consacrees a des etudes epidemiologiques d'echantillons impor- tants de population en prenant les mesures de con- tr61e de la qualite necessaires a l'obtention de don- nees quantitatives fiables sur les rapports dose- reponse. Cela etant, ce que l'on sait devrait deja suffire a inciter les menages, les communautes et les gouvernements a redoubler d'efforts pour reduire !'exposition a ces types de pollution, par !'education et !'introduction de poeles ameliores, de combusti- bles plus propres et de meilleurs systemes d'aera- tion. REFERENCES- REFERENCES 1. SAMET, J. M. ET AL. Health effects and sources of indoor air pollution. American review of respira- tory disease, 136: 1486-1508 (1987). 2. SMITH, K. R. Indoor air pollution and the pollution transition. In: Kasuga, M. (ed.), Indoor air quality. Berlin, Springer Verlag, 1990. (Supplement to the International archives of environmental health). 3. SMITH, K. R. Total exposure assessment: impli- cations for developing countries. Environment, 30 (10): 16-20; 28-35 (1988). Rapp. trimest. statist. sanit. mond., 43 (1990) - 137 - 4. SMITH, K. R. Biofuels, air pollution and health-A global review. New York, Plenum, 1977. 5. DE KONING, H. W. ET AL. Biomass fuel combustion and health. Bulletin of the World Health Organi- zation, 63 (1 ): 11-26 (1985). DE KONING, H. W. ET AL. Combustion de biomasse et sante, 63 (2): 215-232 (1985). 6. CHEN, B. H ET AL. [Indoor air pollution due to coal combustion in China. Modern medicine in China]. 1: 285-289 (1989). (In Chinese). 7. DARY, 0. ET AL Carbon monoxide contamination in dwellings in poor rural areas of Guatemala. Bulletin of environmental contamination and toxicology, 26 (1 ): 24-30 (1981 ). 8. PANDEY, M. R. ET AL. Domestic smoke pollution and respiratory function in rural Nepal. Tokai journal of experimental and clinical medicine, 10 (4): 471-481 (1985). 9. INDIA NATIONAL INSTITUTE OF OCCUPATIONAL HEALTH. Comparative epidemiological studies of the ef- fects of air pollutants. In: Annual report, 1979. Ahmedabad, Indian Council of Medical Research, 1979. 10. AGGARWAL, A. L. ET AL. Assessment of exposure to benzo(a)pyrene in air for various population groups in Admedabad. Atmospheric environ- ment, 16 (4): 867-870 (1982). 11. PATEL, Y. S. ET AL. Indoor air pollution problems: traditional versus modern fuels. In: Berglund, B. et al. (eds), Proceedings of the Third Inter- national Conference on Indoor Air Quality and Climate. (Vol. 4). Stockholm, Swedish Council for Building Research, 1984. 12. KAMAT, S. R. ET AL (EDS). Bombay air pollution - Health study. Bombay, Municipal Corporation of Greater Bombay, 1984. 13. MALIK, S. K. Preliminary report on exposure to domestic cooking fuels and chronic bronchitis. Indian journal of chest diseases and allied sciences, 27 (3): 171-174 (1985). 14. PANDEY, M. R. & GHIMIRE, M. Prevalence of var- ious types of heart disease in Kathmandu. Journal of the Nepalese Medical Association, 13: 33 (1975). 15. PANDEY, M. R. ET AL. Chronic bronchitis and cor pulmonale in Nepal. Kathmandu, Mrigendra Med- ical Trust, 1988. 16. WOOLCOCK, A. J. ET AL. Studies of chronic lung diseases in New Guinea populations. American review of respiratory disease, 102: 575-590 (1970). 17. BATTACHARYA, S. N. ET AL. Chronic bronchitis in rural populations. Indian journal of chest dis- eases and allied sciences, 17 (1): 1-7 (1975). 18. PADMAVATI, S. & ARORA, S. S. Sex differences in chronic cor pulmona/e in Delhi. British journal of diseases of the chest, 70: 251-259 ( 1976). 19. ANDERSON, H. R. Chronic lung disease in the Papua New Guinea highlands. Thorax, 34: 647- 653 (1979). 20. PANDEY, M. R. Prevalence of chronic bronchitis in a rural community of the hill region of Nepal; Domestic smoke pollution and chronic bronchitis in a rural community of the hill region of Nepal. Thorax, 39: 331-336; 337-339 (1984). 21. ANDERSON, H. R. Respiratory abnormalities in Papua New Guinea children: the effects of loca- lity and domestic wood-smoke pollution. Inter- national journal of epidemiology, 7: 63-72 (1978). Wld hlth statist. quart., 43 ( 19901 22. KossovE, D. Smoke-filled rooms and respiratory diseases in infants. South African medical journal, 61: 622-624 (1982). 23. BOLEIJ, J. S. ET AL Domestic air pollution from biomass burning in Kenya. Atmospheric environ- ment, 23 (8): 1677-1681 (1989). 24. CAMPBELL, H. ET AL Indoor air pollution in de- veloping countries and acute respiratory infec- tions in children. Lancet, i: 1012 (1989). 25. PANDEY, M. R. ET AL Domestic smoke pollution and acute respiratory infection in a rural com- munity of the hill region of Nepal. Environment international, 15: 337-340 (1989). 26. CHEN, B. H. & WONG, J. [Effects of indoor and outdoor pollution on the immune function of schoolchildren. Journal of environment and health]. 6 (4): 1-10 (1989). (In Chinese). 27. CHEN, B. H. ET AL [Health stress from domestic coal combustion. Journal of environment and health], 4 (6): 1 (1987). (In Chinese). 28. Hu, H. S. & Liu, Y. C. Evaluation of indoor air pollution and its effect on human health in Beijing's rural areas. Environment international, 15: 321-335 (1989). 29. K1M, Y. S. & KREISEL, W. Effects of household ex- posures to carbon monoxide poisoning in Korea. In: Berglund, B. et al. (eds). Proceedings of the Third International Conference on Indoor Air Quality and Climate. (Vol. 4). Stockholm, Swedish Council for Building Research, 1984. 30. Lu, B. Y. ET AL [Research on human airway re- sistance in high and low lung-cancer prevalence regions in Xuanwei County. Journal of environ- ment and health]. 1 (4): 12 (1984). (In Chinese). 31. ZHOU, M. H. ET AL. [An epidemiological survey on indoor air pollution. Journal of environment and health]. 2 (5): 20 (1985). (In Chinese). 32. ZHOU, G. D. ET AL [An epidemiological survey on indoor pollution from various fuel combustion. In: Chen, B. H. et al. (eds), Indoor air pollution and its health effects). Shanghai, Shanghai Med- ical University, 1987. (In Chinese). 33. Koo, L. C. ET AL. Personal exposure to nitrogen dioxide and its association with respiratory ill- ness in Hong Kong. American review of respira- tory disease, forthcoming. 34. 35. 36. Xu, Z. Y. ET AL [Air pollution and lung cancer in Liaoning Province. In: Proceedings of air pollu- tion and mortality). Beijing, Ministry of Public Health, 1984. (In Chinese). Du, Y. S. ET AL. [A preliminary analysis of trends of lung cancer mortality in Guangzhou, China. Acta academiae medicinae Guangzhou], 12 (1): 1-22 (1984). (In Chinese). Ou, X. L. ET AL. [Relation between housewife lung cancer and domestic coal use. Journal of en- vironment and health], 6 (1): 4 (1989). (In Chinese). 37. WANG, F. L. ET AL. [Analysis of risk factors for female lung adenocarcinoma in Herbin - Indoor air pollution. Chinese journal of preventive medi- cine], 23 (5): 270 (1989). (In Chinese). 38. YUNNAN PROVINCIAL PUBLIC HEALTH AND ANTI-EPIDEMIC STATION. [Indoor air pollution monitoring in high and low lung-cancer prevalence regions in Xuanwei County. Journal of environment and health), 1 (5): 14 (1984). (In Chinese). 39. HE, X. ET AL. [Lung cancer and indoor air pollu- tion in China. In: Huang, H. (ed.), Medicine in - 138 - China: cancer, early diagnosis and contron. Hong Kong, University of Hong Kong, 1986. (In Chinese). 40. MUMFORD, J. L. ET AL. Lung cancer and indoor air pollution in Xuanwei, China. Science, 235: 217- 235 (1987). 41. Koo, L. C. ET AL. Do cooking fuels pose a risk for lung cancer? A case control study of women in Hong Kong. Ecology of disease, 2 (4): 255-265 (1983). 42. Koo, L. C. Environmental tobacco smoke and lung cancer - Is it the smoke or the diet? In: Bieva, C. J. et al. (eds), Present and future of in- door air quality. Amsterdam, Elsevier Science Publications, 1989. 43. WEN, C. P. Nasopharyngeal cancer. In: Quinn, J. E. (ed.), China medicine as we saw it. Washington, D. C., Department of Health, Educa- tion and Welfare, 1974. (NIH75-684). 44. ARMSTRONG, R. w. ET AL. Self-specific environ- ments associated with nasopharyngeal car- cinoma in Selangor, Malaysia. Social science and medicine, 120: 149-156 (1978). 45. CLIFFORD, P. Carcinogens in the nose and throat. Proceedings of the Royal Society of Medicine, 65: 682-686 (1972). 46. HOFFMANN, D. & WYNDER, E. L. Respiratory car- cinogens. In: International symposium on identi- fication and measurement of environmental po/- lutants. Ottawa, Canadian National Research Council, 1972. 47. Yu, M. C. ET AL. Epidemiology of nasopharyngeal carcinoma in Malaysia and Hong Kong. National Cancer Institute monograph, 69: 203-207 (1985). 48. XIAO, Y. Y. ET AL. [Observations on dental fluo- rosis induced by indoor air pollution from do- mestic fuel combustion. Journal of environment and health], 4 (1): 36 (1987). (In Chinese). 49. WANG, C. L. ET AL. [An endemic fluorosis induced by domestic fuel combustion. Journal of en- vironment and health], 4 (3): 12 (1986). (In Chinese). 50. YANG, D. J. [Suggestions on the diagnosis and identification of fluorosis induced by food pollut- ed from coal combustion, using x-rays. Chinese journal of preventive medicine], 21 (3): 136 (1987). (In Chinese). 51. DAI, G. J. ET AL. [A survey on fluorosis indu- ced by coal combustion. Chinese journal of preventive medicine], 20 (4): 217 (1986). (In Chinese). 52. ZHANG, Z. F. ET AL. Indoor air pollution of coal fumes as a risk factor of stroke, Shanghai. American journal of public health, 78: 975-977 (1988). 53. PANDEY, M. R. ET AL. Indoor air pollution in devel- oping countries and acute respiratory infection in children. Lancet, i: 427-429 (1989). Rapp. trimest. statist. sanit. mond., 43 (19901 - 139 - ACUTE PESTICIDE POISONING: A MAJOR GLOBAL HEAL TH PROBLEM J. Jeyaratnam• Pesticides are a group of chemicals used predomi- nantly in agriculture and against vectors in vector- borne diseases such as malaria, filariasis, etc. There are several definitions of a pesticide; the Food and Agriculture Organization of the United Nations (FAQ) (1) defines a pesticide as any substance or mixture of substances intended for preventing, ·des- troying or controlling any pest, including vectors of human or animal disease, unwanted species of plants or animals causing harm during or otherwise interfering with the production, processing, storage or marketing of food, agricultural commodities, wood and wood products or animal feedstuffs or which may be administered to animals for the con- trol of insects, arachnids or other pests in or on their bodies. The term includes substances intended for use as a plant-growth regulator, defoliant, dessicant or fruit-thinning agent or agent for preventing the premature fall of fruit and substances applied to crops either before or after harvest to protect the commodity from deterioration during storage and transport. • Associate Professor, Department of Community, Occupational and Family Medicine, National University of Singapore, Singapore. It is evident that a pesticide, so defined, is used for the variety of benefits it provides to mankind. But in so doing there are certain undesirable and unwanted effects of pesticide usage which cannot be ignored. This article focuses attention on the unnecessary acute health effects that arise from the use of pesti- cides. Man may be exposed to pesticides in a variety of ways; at different dose levels and for varying periods of time. A schematic representation of the manner of human exposure to pesticides is shown in Fig. 1. In most countries all of these modes of pesticide ex- posure prevail, but what is most important is that each country or region identify the mode of ex- posure and resultant hazard which is most important to its own circumstances. For instance, in the in- dustrialized world the problem of acute pesticide poisoning has largely been controlled and the main focus of attention is on the possible health effects arising from exposure to low levels of pesticides over a long period of time. Such exposures usually arise from environmental contamination as well as from pesticide residues in food, whereas the situ- ation is quite the reverse in the countries of the developing world. In these countries the main health FIG.1 POPULATION GROUPS AT RISK GROUPES DE POPULATION A RISQUE Suicides and mass poisoning Pesticide formulators, mixers, applicators and pickers - Suicides et intoxications collectives Preposes a la formulation, au melange et a !'application des pesticides et cueilleurs Pesticide manufacturers, formulators, mixers, applicators and pickers - Fabricants de pesticides, preposes a la formulation, au melange et a !'application des pesticides et cueilleurs All population groups - Taus les groupes de population Source: Reference (?)-Reference (7). Wld hlth statist. quart., 43 ( 1990) exposure Exposition unique et breve de tres forte intensite Long-term high-level exposure Exposition prolongee de forte intensite Long-term low-level exposure Exposition prolongee de faible intensite - 140 - problem arising from pesticides is that due to acute poisoning. No doubt the concerns of long-term ex- posure to low levels of pesticides also obtain in these countries, but they should not be considered a priority health issue as they are in the countries of the industrialized world. The need to set priorities in combating the health problems arising from pesti- cide usage is illustrated (2) by the fact that only 1-2% of scientific papers published on pesticides and health has addressed the issue of acute pesticide poisoning. This article will mainly focus on acute pesticide poisoning, as it is a major health concern of pesticide usage. Current status of knowledge Any figures concerning the extent of acute pesticide poisoning on a global scale are largely based, by necessity, on estimates. The first such estimate was made in 1973 by the World Health Organization (3) which suggested that 500 OOO cases of acute serious pesticide poisoning occurred annually. This estimate included only hospitalized cases of unintentional poisoning (excluding suicide attempts). At that time it was considered to be an unacceptably large prob- lem, requiring efforts to substantiate this estimate as well as to control the problem. In 1982 (4) a national study of hospital cases of acute pesticide poisoning demonstrated that Sri Lanka, a country with a popu- lation of 12 million, had approximately 10 OOO per- sons admitted to hospitals for acute pesticide poisoning annually, resulting in almost 1 OOO deaths. The public health importance of this figure was highlighted by the fact that the deaths due to acute pesticide poisoning for that particular year were almost twice the total number of deaths due to malaria, poliomyelitis, whooping cough, diphtheria and tetanus, the traditional public health problems of developing countries. These figures included sui- cide attempts and suicides, which comprised about two-thirds of the hospitalized poisonings (4-6). The equivalent annual figures at a global level in 1985 (5) were estimated at approximately 3 million cases hospitalized and approximately 220 OOO deaths. Recently, a WHO task group reviewed the available estimates and other pesticide poisoning data and summarized the overall public health impact of pesticides (7), as shown in Fig. 2. WHO states that "the estimated 3 million cases of acute severe poisonings may be matched by a greater number of unreported, but mild, intoxications and acute con- ditions such as dermatitis" (this figure includes sui- cide attempts). It should be noted that the data shown in Fig. 2 are the inverse of the pyramid shown in Fig. 1. This is because the associated morbidity is small although a large number of per- sons are potentially at risk from long-term low-level exposure to pesticides. On the other hand, although the numbers exposed to high levels of pesticides for a short period of time are small, their morbidity and mortality are high. On the basis of this data, WHO states that "there is no segment of the general population that is sheltered from exposure to pesti- cides and potentially serious health effects, although a disproportionate burden is shouldered by the developing world and high-risk groups in each country". Suicides have been incriminated as a major factor in the causation of acute pesticide poisoning. They contribute to approximately two-thirds of all causes of acute pesticide poisoning. The herbicide paraquat is extensively used as an agent for suicides. Para- quat poisoning is a major problem in Malaysia particularly, with 73.4% of such poisonings due to suicides, 13.8% to accidents and only 1.07% to occu- pational accidents (8). Suicide in any society is a social problem which requires attention from many disciplines. The reason for the extensive use of pesticides as an agent for suicide in developing countries is the ready avail- ability of extremely toxic pesticides. In the developed countries, pesticides are responsible for only a small percentage of all poisonings, whereas in the de- veloping countries they are a major contributor to poisoning (Table 1) (9). The ready availability to the general public of these toxic pesticides should be controlled to limit the current epidemic of acute pesticide poisoning in the countries of the develop- ing world. FIG.2 ESTIMATED OVERALL ANNUAL PUBLIC HEALTH IMPACT EFFETS ANNUELS GLOBAUX ESTIMATIFS SUR LA SANTE PUBLIQUE Approximate number of poisonings in each group Nombre approximatif d'intoxications par groupe Single and short-term exposure (including suicides) - Exposition unique et breve (y compris les suicides) Long-term exposure, specific chronic effects - Exposition prolongee, effets chroniques specifiques Long-term exposure, unspecific chronic effects (cancer) - Exposition prolongee, effets chroniques non specifiques (cancer) Source: Reference: (7)- Reference (7). 3 OOO OOO (220 OOO deaths/deces) Rapp. trimest. statist. sanit. mond., 43 ( 1990) - 141 - TABLE 1. PROPORTION OF ACUTE POISONINGS DUE TO PESTICIDES, SELECTED COUNTRIES, 1980s TABLEAU 1. POURCENTAGE DES INTOXICATIONS AIGU~S DUES AUX PESTICIDES, CERTAINS PAYS, ANNiEs 80 Indonesia - lndonesie . . . . . . . . Brazil - Brasil . . . . . . . . . . . . United Kingdom - Royaume-Uni . . Australia - Australie . . . . . . . . . Country-Pays Canada . . . . . . . . . . . . . . . . . . United States of America - Etats-Unis d'Amerique Only relatively recently have data on acute pesti- cide poisoning in Africa become available. Choudhury (10) estimated that 11 million cases of pesticide intoxications occur annually in Africa. This figure includes minor non-hospitalized poisoning cases and is therefore not directly comparable with the figures given above. Table 2 indicates the extent of the problem in some of the African countries. In making global estimates, there are a variety of pitfalls which are likely to give rise to inaccurate estimates of the extent of the problem. The main issues which distort the picture are: (i) misdiagnosis of acute pesticide poisoning; (ii) some studies con- fined to hospital cases, others confined to minor poisoning or mere pesticide exposure; (iii) most studies confined to a limited region and not nation- ally representative; and (iv) incomplete compilation of data. In Indonesia, although the officially-collected records do not indicate a problem, local studies estimate that there are 30 OOO cases of pesticide poisoning annually, of whom approximately 2 400 require hos- pitalization.b In Thailand (11) the epidemiological sur- veillance report records 2 094 cases of pesticide poisoning with no deaths for the year 1985, while the data collected by the National Environmental Board record a total of 4 046 cases resulting in 289 deaths, indicating the great variation even in official records. The epidemiological surveillance data in Thailand are routinely obtained on the basis that O Suma'mur, P. K. Current situation of chemicals and accidents/ episodes involving toxic chemicals in Indonesia. Paper presented at a WHO intercountry workshop on chemical safety in countries of the South-East Asia Region, 29 October-2 November 1989. c Jeyaratnam, J. Pesticide projet undertaken by the Asian Associa- tion of Occupational Health Research Committee. Percentage of total acute poisonings Pourcentage des intoxications aigues (toutes causes reunies) 28.0 16.0 5.0 3.0 2.4 0.8 pesticide poisoning is one of the 54 notifiable dis- eases in that country, whereas the National Environ- mental Board collects data from a variety of sources. Given this situation, there is an urgent need to collect accurate data on the different aspects of acute pesticide poisoning, particularly in the coun- tries of the developing world. Such data should not merely be looked upon as data to establish the extent of the problem, but rather as the starting point for programmes for the control of acute pesti- cide poisoning. Further, such data are necessary to monitor and evaluate the efficacy of different inter- vention programmes that may be implemented for the control of acute pesticide poisoning. The situation among agricultural workers The global estimates of acute pesticide poisoning are largely based on hospital data. Very few surveys have been undertaken to study the problem of acute pesticide poisoning among agricultural workers. A survey undertaken in Asian countries (6) examined the problem, and the situation in two countries with the most reliable data (Table 3) indicates the extent of pesticide-poisoning episodes as perceived by the workers themselves. On this basis, if it is taken that maybe on average 3% of agricultural workers in developing countries suffer an episode of pesticide poisoning a year (6), it would mean that for the 830 million agricultural workers in the developing world (12), there are about 25 million cases of occupational pesticide poisoning. The bulk of these episodes of poisoning do not get recorded, as they are considered minor and often self-limiting, and most of the patients do not seek medical attention. A survey of the prefer- red choice of medical attention for episodes of TABLE 2. NUMBER OF POISONINGS IN SOME AFRICAN COUNTRIES, 1980s TABLEAU 2. NOMBRE D'INTOXICATIONS DANS CERTAINS PAYS D'AFRIQUE, ANNiEs 80 Country - Pays Sudan - Soudan . . . . . . . . . . . . . . . . . . . . . . . . United Republic of Tanzania - Republique-Unie de Tanzanie . Kenya .............................. . Uganda - Ouganda . . . . . . . . . . . . . . . . . . . . . . . . . Mozambique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cameroon - Cameroun . . . . . . . . . . . . . . . . . . . . Zimbabwe. . . . . . . . . . . . . . . . .......... . Cote d'Ivoire . . . . . . . . . . . . . . . . . . . . . . . . . . . Malawi ............................. . Senegal - Senegal . . . . . . . . . . . . . . . . . . . . . . . Mauritius - Maurice . . . . . . . . . . . . . . . . . . . . . . . From reference ( 11) - D'apres la reference (17). Wld hlth statist. quart., 43 (1990) Population (millions) 24 23 22 17 15 11 10 10 8 7 2 Percentage agricultural labour force Pourcentage de la main-d'ceuvre agricole 80 85 80 80 70 80 80 80 85 80 75 Annual number of cases of pesticide poisoning Nombre annuel de cas d'intoxication par les pesticides 384 OOO 368 OOO 350 OOO 272 OOO 240 OOO 175 OOO 160 OOO 160 OOO 128 OOO 112 OOO 3200 - 142 - TABLE 3. PERCENTAGE AGRICULTURAL WORKERS AND PESTICIDE USERS WITH PESTICIDE POISONING, SELECTED COUNTRIES, 1980s TABLEAU 3. POURCENTAGE D'OUVRIERS AGRICOLES ET D'UTILISATEURS DE PESTICIDES VICTIMES D'INTOXICATIONS PAR DES PESTICIDES, CERTAINS PAYS, ANN~ES 80 Agricultural workers Country - Pays ever poisoned Ouvriers agricoles deja intoxiques (%) Malaysia - Malaisie . . . . . 13.3 Sri Lanka .......... . 4.6 Source: Reference (6) - Reference (6). poisoning among agricultural workers indicated vari- ation. In Indonesia 69.9% of patients sought treat- ment in hospitals, in Malaysia 67.8% and in Sri Lanka 83.5%, but in Thailand only 8.4% of patients chose to seek treatment in hospitals. Control of acute pesticide poisoning The essential starting point in any control pro- gramme is to establish the extent of the problem. In the case of acute pesticide poisoning it has become abundantly clear over the last few years that the problem on a global scale is immense. It has also become evident that this particular aspect of pesti- cide poisoning is almost exclusively a concern of the developing world. The heartening feature in this gloomy picture is that the industrialized countries have been able to successfully control the problem. This must mean that there are valuable lessons to be learnt and it is possible to do likewise in the coun- tries of the developing world. The available data suggest that acute pesticide poisoning is a major health concern in the developing world today. Yet very little has been done. The situation in Sri Lanka is a case in point: in 1982, scientists in Sri Lanka published data (4) to indicate that acute pesticide poisoning was an issue of even greater significance than the traditional public health problems of com- municable diseases seen in developing countries, yet there has been no significant progress since (Table 4). There is an urgent need to take action. The problem can be solved if all interested parties collaborate. It is wasteful and unnecessary to em- bark on activities which purely seek to blame the agrochemical industry. It is recognized that pesti- cides are primarily used for their beneficial effects, but responsible action must be taken to eliminate or minimize the associated hazards. Thus the people, national governments, agrochemical industries, scientists and international agencies all have a role to play in any control programme. The role of responsible agents The role of governments The ultimate responsibility to control the use of pesticides so as to minimize health hazards devolves to national governments. They must continue, and whenever necessary strengthen, health education O World Health Organization. Recommended classification of pesti- cides by hazard and guidelines to classification 1988-1989. (Document WHONBC/88.953, 1988). Agricultural workers Pesticide users Pesticide users poisoned/year ever poisoned poisoned/year Ouvriers agricoles Utilisateursde pesticides Utilisateurs de pesticides intoxiques/an intoxiques intoxiqu0s/an (%) (%) (%) 6.7 14.5 7.3 2.7 11.9 7.1 programmes among pesticide users, particularly to ensure safe practices. Though many countries have enacted legislation, en- forcement remains insufficient. As an immediate cor- rective measure it may be appropriate to consider selective enforcement or selective legislation to con- trol those pesticides considered to be most hazard- ous. For this purpose the WHO document Recom- mended classification of pesticides by hazard and guidelines to classificationd would be most use- ful-pesticides classified extremely hazardous and highly hazardous should be identified for stricter controls. Agricultural activities are undertaken in remote rural areas, which often most lack health-care facilities. The primary health care approach can be regarded as most suitable for such situations. Further, such an approach involves the consumer (in this case the worker) in the process of health-care delivery, thereby making it more effective. The primary health care approach also incorporates the multisectoral approach so essential in the control of pesticide poisoning. TABLE 4. HOSPITAL ADMISSIONS FOR PESTICIDE POISONING, SRI LANKA, 1984-1988 TABLEAU 4. ADMISSIONS DANS LES HOPITAUX POUR INTOXICATION PAR DES PESTICIDES, SRILANKA,1984-1988 Year-Annee Total number Deaths Nombre total Deces 1984 16 085 1 459 1985 14 423 1 439 1986 14 413 1 452 1987 .......... 12 841 1 435 1988 .......... 12 997 1 524 Source: Ministry of Health, Sri Lanka - Ministere de la Sante, Sn Lanka. The role of the agrochemical industry The agrochemical industries are often not included in control programmes. This is a great drawback which needs to be rectified, as these organizations can contribute significantly to the control of poison- ing, particularly in the following areas: • research into developing appropriate personal protective equipment for tropical countries; • prevention of marketing of pesticide mixtures; • maintenance and repair of spray equipment; • research to develop hazard-free spray equipment; • use of safe pesticide containers which are un- likely to be accident prone. Rapp. trimest. statist. sanit. mond., 43 (1990) 143 - The role of international agencies The international agencies, particularly WHO and the International Labour Organisation (ILO), have contri- buted a great deal in their attempts to control pesti- cide poisoning. They should continue their efforts, with particular emphasis on education and training on safety in the use of pesticides (13, 14) and applied research activities, and should play the role of inter- mediary for the involvement of agrochemical in- dustries in safety activities. SUMMARY The global problem of acute pesticide poisoning has been confirmed as extensive by a variety of independent estimates. Further, it is also recognized to be a problem confined to the developing coun- tries. Most estimates concerning the extent of acute pesticide poisoning have been based on data from hospital admissions which would include only the more serious cases. The latest estimate by a WHO task group indicates that there may be 1 million seri- ous unintentional poisonings each year and in addi- tion 2 million people hospitalized for suicide attempts with. pesticides. This necessarily reflects only a fraction of the real problem. On the basis of a survey of self-reported minor poisoning carried out in the Asian region, it is estimated that there could be as many as 25 million agricultural workers in the developing world suffering an episode of poisoning each year. This article emphasizes the need to con- trol the problem on a collaborative basis by all con- cerned, including national governments, agrochemi- cal industries, international agencies, scientists and victims. RESUME Intoxications aigues par les pesticides: probleme de sante majeur dans le monde Diverses evaluations independantes ant confirme l'etendue du probleme mondial des intoxications aigues dues aux pesticides. II a en outre ete reconnu que ce probleme se limitait aux pays en developpe- ment. Les evaluations de l'ampleur des intoxications aigues par les pesticides s'appuient pour la plupart sur les donnees concernant les admissions dans les hopitaux qui ne concernent en principe que les cas les plus graves. Selan !'estimation la plus recente d'un groupe de travail de l'OMS, le nombre annuel des intoxications accidentelles s'eleverait a 1 million, outre les 2 millions de personnes hospitalisees pour tentative de suicide au moyen de pesticides. Cela _ne reflete necessairement qu'une fraction du probleme reel. Sur la base d'une enquete effectuee en Asie concernant les intoxications mineures signalees par les victimes elles-memes, on estime a 25 millions le nombre de travailleurs agricoles qui sont victimes chaque annee d'un episode d'intoxication dans les pays en developpement. Cet article souligne la necessite de resoudre ce probleme avec la collabora- tion de tous les interesses, a savoir les autorites nationales, l'industrie agrochimique, les organismes internationaux, les specialistes scientifiques et les victimes. REFERENCES- REFERENCES 1. FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NATIONS. International code of conduct on the distribution and use of pesticides. Rome, FAQ, 1986. 2. JEYARATNAM, J. 1984 and occupational health in developing countries. Scandinavian journal of work environment and health, 11: 229-234 (1985). 3. WHO Technical Report Series No. 513, 1973 (Safe use of pesticides: twentieth report of the WHO Expert Committee on Insecticides). OMS Serie de Rapports techniques N° 513, 1973 (Securite d'emploi des pesticides: vingtieme rap- port du Comite OMS d'experts). 4. JEYARATNAM, J. ET AL. Survey of pesticide poison- ing in Sri Lanka. Bulletin of the World Health Organization,60 (4): 615-619 (1982). Wld hlth statist. quart., 43 (19901 JEYARATNAM, J. ET AL. Enquete sur les in- toxications par les pesticides a Sri Lanka [resume]. Bulletin de /'Organisation mondiale de la Sante, 60 (4): 618-619 (1982). 5. JEYARATNAM, J. Health problems of pesticide usage in the Third World. British journal of industrial medicine, 42: 505-506 (1985). 6. JEYARATNAM, J. ET AL. Survey of acute pesticide poisoning among agricultural workers in four Asian countries. Bulletin of the World Health Organization,65 (4): 521-527 (1987). JEYARATNAM, J. ET AL Enquete sur !'intoxication aigue par les pesticides chez les travailleurs agricoles dans quatre pays d' Asie [resume]. Bulletin de /'Organisation mondiale de la Sante, 65 (4): 526-527 (1987). - 144 - 7. WORLD HEALTH ORGANIZATION/UNITED NATIONS EN- VIRONMENT PROGRAMME. Public health impact of pesticides used in agriculture. Geneva, WHO, in preparation. 8. WONG, K. T. & NG, T. S. Alleged paraquat poison- ing in Perak. Medical journal of Malaysia, 39 (1): 52-55 (1984). 9. JEYARATNAM, J. Acute poisonings caused by chemicals. In: Extended abstracts of the Inter- national Symposium on Health and Environment in Developing Countries, Haikko, Finland, August. 29-30, 1986. Helsinki, Institute of Occupational Health, 1986. 10. CHOUDHURY, A. W. Health hazards of pesticide use in Africa. In: Lehtinen, S. et al. (eds), Pro- ceedings of the East Africa Regional Symposium on Chemical Accidents and Occupational Health. Helsinki, Institute of Occupational Health, 1989. 11. KRITALUGSANA, S. Pesticide poisoning studies and data collection in Thailand. In: Tend, P. S. & Heang, K. L. (eds), Pesticide management and integrated pest management in South-East Asia. Manila, Island Publishing House, 1988. 12. INTERNATIONAL LABOUR ORGANISATION. Medium-term plan, 1982-87. Geneva, ILO, 1980. 13. INTERNATIONAL LABOUR ORGANISATION. Guide to safety and health in the use of agrochemicals. Geneva, ILO, in preparation. 14. WHO Technical Report Series No. 720, 1985 (Safe use of pesticides: ninth report of the WHO Expert Committee on Vector Biology and Con- trol). OMS Serie de Rapports techniques N° 720, 1985 (Securite d'emploi des pesticides: neuvieme rap- port du Comite OMS d'experts). Rapp. trimest. statist. sanit. mond., 43 (1990) - 145 - ASSESSING THE HEAL TH IMPACT OF URBANIZATION Brian T. Williams" Urbanization involves the processes of social aggregation, migration, modernization, industrializa- tion and urban living. All these make individual impacts upon population health status, but the changes in health status attributable to urbanization are not the simple total of the effects of each of these features; and the overall impact depends upon such factors as the pace of urbanization, whether migration is unidirectional or not, and the char- acteristics of migrants compared with those left be- hind; the climate, terrain, housing, transport sys- tems, type of industrial base of the urbanized area; the volume, nature and coverage of its health-care facilities; and, more fundamentally, how an urban population is defined. This article examines the complexity of attempting to assess the extent to which urbanization promotes health, or otherwise. In rapidly urbanizing populations, because of the great variability in the coverage and accuracy of important life events, occurrence of disease and encounters with health-care providers (numerator data), as well as imprecise estimates of the size and other characteristics of the populations concerned (denominator data), the exact impact on health of these social changes is not generally ascertainable statistically. One cannot, for example, readily pro- duce reliable death rates from all or any causes, nor accurately compute how many years of potential life are lost to these causes. Still less can we learn from published statistics what proportion of the surviving population is handicapped in various respects, nor for what proportion of their lives the sufferers are affected. Health benefit of urbanization Overall impact The crudest way of estimating the health effect of urbanization is to relate the degree to which popula- tions are urbanized to a proxy measure of popula- tion health status. Child mortality rates and average life expectancy at birth are such indicators. UNICEF's 1989 report (1) provides data which permit these comparisons (Table 1), albeit from the converse standpoint of what might determine the levels of child mortality. High child-mortality rates, hence low- ered life expectancy, hence higher proportions of potential life lost are associated with lower levels of urbanization. The net effect of urbanization is, in these respects, beneficial therefore. Nevertheless, urbanized living per se does not imply mortality reduction. Among low-income developing countries in 1983, for example, Uganda's population, • Department of Public Health Medicine, University of Sheffield Medical School, Sheffield, United Kingdom. Wld hlth statist. quart., 43 ( 1990) TABLE 1. UNDER-5 MORTALITY RATE (<5 MR) PER 1 OOO LIVE BIRTHS AND PROPORTION (%) OF POPULATION URBANIZED, 1987 TABLEAU 1. TAUX DE MORTALln CHEZ LES MOINS DE 5 ANS (TM <5) POUR 1 OOO NAISSANCES VIVANTES ET PROPORTION(%) DE LA POPULATION URBANIS!aE, 1987 <5MRper1 OOO live births TM <5 pour 1 OOO naissances vivantes >170 95- 170 31 - 94 "" 30 Median percentage population urbanized Number of countries Mediane des pourcentages de Nombre de pays population urbanisee 23 33 41 30 53 29 73 39 Source: Reference ( I) - Reference (1). 13.4% of which was urbanized, experienced an infant mortality of 94 per 1 OOO live births, roughly similar to that of its continental neighbour Zambia (101 per 1 OOO), 46.8% of whose population were urbanized; and, among middle-income developing countries, Brazil (70.7% urbanized) had an infant mortality rate (71 per 1 OOO) little lower than that of Kenya (15.7% urbanized) at 82 per 1 OOO. Unfortunately, doubts about the completeness of ascertainment com- promise the value of such transnational com- parisons, even when health indices are concerned with the urban segments of the population. Attributing changes in health status As stated above, urbanization has several com- ponents which may affect individual health status, alone or in combination. The impact of some of these components has been described in various types of comparative study. Statistical correlations of survey data. Studies of this type compare the health status of populations of areas which differ in one or more features of urban- ization. The comparisons may be made using report- ed statistical measurements or survey data. One of the most thorough of these recently reported the impact of urbanization and industrialization on the populations of the various states of Nigeria (2). The populations of urban zones (localities of 20 OOO or more) held distinct advantages socioeconomically over those in rural areas. For example, in the 1983- 1984 national household survey, proportionately more urban adults were literate (51% against 24%); urban workers' wage levels were more than double those of rural workers; 58% of the urban population had access to piped, treated water but only 6% of those in rural areas; relatively fewer urban residents had no toilet facilities (17% against 54%); and although over 80% of the population lived in villages, only 50% of health clinics were situated in rural areas. The better health facilities and socio- economic conditions in urban centres were associ- - 146 - ated with lower infant and child mortality rates throughout the previous 15-year period. In 1975- 1979, the infant mortality rates in urban and rural zones were 63.9 and 88.5 per 1 OOO live births re- spectively. The contribution made by a number of socioeco- nomic and ecological factors to variation in two health outcome measures - life-table infant mortality rates and the expectation of life at birth - were examined in a regression analysis using data from the 19 Nigerian states. For each state the social indicators "urbanization index" (percentage of the population living in localities of 20 OOO or more), overcrowding (person/room density) and per capita generated revenue, the ecological indicators depict- ing solid waste generated and atmospheric pollu- tion, an industrialization indicator (percentage of per- sons engaged in industrial establishments). and a health services indicator (beds per 1 OOO population), were correlated with the respective health-status in- dicators in the analysis. The variables explained only 40% and 43% of the total variation in the distribution of infant mortality rate and life expectancy respectively. Overcrowding and pollution accounted for most of explained vari- ation in infant mortality rates, overcrowding being the weightier factor. These factors were also the most important determinants of life expectancy, but this time pollution was the more critical explanatory factor. Thus, while urbanization and industrialization overall had positive effects on health status in Nigeria, there were negative effects of two elements, overcrowding and pollution. This, then, represents the imbalance which characterizes the positive and negative effects of urbanization. Statistical correlations (the case of tuberculosis and leprosy). A review of leprosy prevalence rates and tuberculosis incidence rates among 29 African coun- tries, 13 of which were classed as rural and 16 as urban (more than 12% of the population living in urban areas) illustrated the effect of partial cross- interference between tuberculosis and the milder, paucibacillary form of leprosy (3). It showed a nega- tive correlation between leprosy and urbanization in the 16 urban countries. In a partial correlation analy- sis involving all the countries, there was a strong negative urban influence on the ratio of the total number of leprosy cases per new case of tuberculo- sis per 100 OOO population (R = -0.598; P = 0.007). It was argued that the rise of the city in Africa, acting through a combination of influences including tuber- culosis, has led to a decline of leprosy. As well as the cross-interference effect, the re- viewers listed other plausible reasons for the lower levels of leprosy observed in towns, such as dimin- ished concealment of infection in towns; less toler- ance of the condition among families; better access to health care leading to earlier immunization, dia- gnosis and treatment; better compliance with treat- ment regimes; reduced infection rates as a result of there being proportionately fewer effective carriers in towns, and less favourable circumstances for transmission. Urbanization may be said, therefore, to have made a substantial contribution towards the eradication of leprosy. Prevalence surveys. One of the earliest cross- sectional studies to compare health characteristics of urban and rural populations was made in the United Kingdom (4), where the prevalence of pulmonary disease in mail-delivery workers in urban and rural areas was compared. After controlling for smoking habits, it was found that the pulmonary functions of the urban outdoor workers were impaired to a great- er extent and in a higher proportion than those of rural workers, and this was attributed to the levels of atmospheric pollution in urban areas. More recently, a study of 13 636 registered, mentally handicapped people in the Galician provinces of north-west Spain reported a prevalence of mental handicap of 5 per 1 OOO total population (3.4 per 1 OOO severely handicapped) (5). The districts with the highest rates (up to 15 per 1 OOO) were the most isolated, mountainous, depressed and disadvan- taged, while the lowest rates were in districts which included, or were adjacent to, cities. As Table 2 shows, severe subnormality was signifi- cantly more prevalent in rural than in urban areas, whereas milder handicap was more common in urban than in rural areas. Characteristic of the rural areas was low socioeconomic, cultural and health- care levels and a high degree of endogamy. Severe subnormality was favoured by the poor health care available in rural areas, particularly with regard to obstetrics and paediatric services, by endemic men- ingitis and goitre, and by selective emigration of the healthier individuals. Less severe mental deficiency however tended to go undetected in rural areas, where integration of the mentally handicapped is facilitated by socio-occupational requirements which are less demanding than in an urban environment. This study exemplifies an aspect of the health bur- den, not of the urban but of the rural environment. TABLE 2. PERCENTAGE DISTRIBUTION OF THE MENTALL V HANDICAPPED BY IQ FOR URBAN AND RURAL GALICIA AND GALICIA AS A WHOLE, 1983 TABLEAU 2. DISTRIBUTION EN POURCENTAGE DES HANDICAP~S ME.NTAUX SELON LE QI POUR LES ZONES URBAINE ET RURALE DE GALICE ET LA GALICE DANS SON ENSEMBLE, 1983 Area -Zone Urban - Urbaine . . . . . . . . . . . Rural - Rurale . . . . . . . . Total ................. . • p< 0.001. Profound Profonde (0-19) 6.51 11.39 10.24 Deficiency level (10) - Niveau de deficience (01) Severe - Severe Slight - Legere Grave Moderate Mild Borderline Moderee Faible Li mite (20-34) (35-49) (50-70) (70-85) 16.59 38.23 24.09 14.58• 25.87• 37.66 18.52 6.56 23.68 37.68 19.83 8.45 Total 100.0 100.0 100.0 Rapp. trimest. statist. sanit. mond., 43 (1990) - 147 - The effects of migration, and particularly the ten- dency for immigrants to exhibit health char- acteristics more like those of the host population, are illustrated in a study of the prevalence of dia- betes mellitus in Tunisia involving health interviews and blood glucose level estimations in random sam- ples of 3 826 adults from an urban area of Tunis and 1 787 from a rural district of Siliana (6). Overall, the age-standardized prevalence rate was much higher in urban adults, especially women (3.5% vs. 0.6% in women, 4.6% vs. 2.3% in men). Within the urban sample, the prevalence rate was similar in subjects born in urban or rural areas, but a family history of diabetes was more often reported by the urban sample. The findings illustrate how "westernization" of urban areas in developing countries leads to the adoption of Western morbidity patterns. The "patch" approach. Rapid urbanization may bring some rapid health benefits, but the attendant psy- chosocial changes may threaten mental health. This is thought to be due to the depressing effects of poverty and overcrowding, and sociocultural disin- tegration, detaching individuals from their support- ive communities or old modes of life. Children might find the adjustment difficult to make, especially in the early stages. The picture of child behaviour in the population of an area (or patch) of Khartoum (Sudan) had been studied in 1965. By 1980, when the study was re- peated, the area had been urbanized. Compared to 1965, the 1980 study showed an increase of be- haviour problems among boys aged 7-15 years (Table 3) (7). Physical health status had improved overall however, but there was a strong connection between poor somatic health and high rates of behaviour deviance. Among those studied in 1980, the prevalence of any behavioural problems was not found to be signifi- cantly different among newcomers (28.1%) from that in the indigenous population (31.8%). But unlike the indigenous children, who showed a uniform dis- TABLE 3. COMPARISON OF HEALTH STATUS, KHARTOUM CHILDREN AGED 3-15, 1965 AND 1980• TABLEAU 3. COMPARAISON DE L'ETAT DE SANTE CHEZ LES ENFANTS DE 3 A 15 ANS, KHARTOUM, 1965 et 1980• 1965 1980 (%) (%) Behavioural symptoms - Symptomes comportementaux 3-6 years/ans None - Aucun . . . . . . . . . . . . Mild - Lagers . . . . . . . . . . . . Moderate - Moderes . . Severe - Graves 7-15 years/ ans None - Aucun . . Mild - Lagers . . . Moderate - Moderes . . . . . . . Severe - Graves . . . . . . . . . 64 6 25 5 61 10 17 10 Physical health - Etat physique Healthy - En bonne sante . . . . . . 19 Mild illness - Maladie benigne . . . . 42 Moderat; illness - Maladie de moyenne grav1te . . . . . . . . . . . . 33 Severe illness - Maladie grave 6 • p <0.001. W/d hlth statist. quart., 43 (19901 45 30 16 9 48 19 17 17 32 31 35 2 tribution of behaviour problems between preschool and school-age groups, the children of newcomers showed a significantly higher rate of such problems among older children aged 7-15 years (34.7%) than among younger ones (19.0%) (p < 0.05), indicating that adjustment becomes more difficult as childhood proceeds. While urbanization may enhance the economic prop- erties and health of able-bodied adults, it is liable at first to prove psychologically disadvantageous to children. The longitudinal study. The literature on the pro- gressive effects on the health of cohorts of indi- viduals who migrate from rural to urban areas is sparse (in contrast to transnational migrants). Rel- atively more is known about the way health status changes through life stages from studies such as those of the Framingham and North Karelia popula- tions. These however do not address the issue of urbanization. One study of changes in perceived health status and morbidity reported by 426 urban black American youths in a series of interviews when they were aged 12-17, 18-23, then 26-31 showed no continuous nor progressive health decline over the three life stages (8). Morbidity reports increased in number between adolescence and post-adolescence, but this was succeeded by stabilization (Table 4). The authors of this report suggested that the findings reflected the stressfulness of adolescent life for urban blacks. The usefulness of self-reported morbidity and disability as a means of ascertaining health status is discussed further later in this article. The case-control study. Intra-urban differences in incidence and prevalence of disease are well rec- ognized. What explains these differences? We know that health disadvantage is associated with low socioeconomic status, but of what is this a marker? In a population-based case-control study in metro- politan areas of southern Brazil, children dying in infancy from diarrhoea were compared to neigh- bourhood controls in terms of several social and environmental variables (9). Increased risk of death from diarrhoea was significantly associated with non-availability of piped water, the absence of a flush toilet, residence in a poorly-built house and household overcrowding. These factors interact however and, after adjustment for the confounding effects, the only association to remain statistically significant was with the availability of piped water. Those without easy access to piped water were 4.8 times more likely to die from diarrhoea (95% confi- dence interval 1.7-13.8) and those with water piped to their plot but not to their house 1.5 times more likely (95% Cl 0.8-3.0). This highlighted the pre-emptive importance of a piped water supply. Excess or avoidable mortality The concept Deaths resulting from certain causes may be classed as potentially avoidable altogether, such as those resulting from accident, or at least potentially of being delayed, such as those from infantile diar- rhoea, bronchial asthma or cirrhosis of the liver, so that the individual dies later and of some other disease condition. The limit of what is theoretically achievable may be set according to age-specific or - 148 - TABLE 4. MEAN NUMBER OF CHRONIC HEALTH PROBLEMS AND GLOBAL APPRAISAL OF HEALTH STATUS AT THREE LIFE STAGES, URBAN BLACKS, LATE 1960s TO 1983-84 TABLEAU 4. NOMBRE MOYEN DE PROBLf:MES DE SANT~ CHRONIQUES ET APPR~CIATION GLOBALE DE L'~TAT DE SANT~ A. TROIS AGES DIFF~RENTS, POPULATION URBAINE NOIRE, FIN DES ANN~ES 60 A. 1983-84 Adolescence Late 1960s Fin des annees60 Post-adolescence 1975-76 Young adult Jeunesadultes 1983-84 Number of chronic health problems - Nombre de problemes de sante chroniques Males - Hammes (n = 199) Mean - Moyenne . . . . . . . . . . . Standard deviation - Ecart type . . . Females - Femmes (n = 212) Mean - Moyenne . . . . . . . . . . . Standard deviation - Ecart type 3.20 (2.97) 5.90 (4.14) 4.62 4.08 (3.51) (3.33) 6.08 5.29 (4.34) (3.75) Global appraisal - Appreciation globale Males - Hammes Very/pretty good - Tres bon/assez bon Fair/poor - Moyen/mauvais . . . . . . . . . . Females - Femmes Very/pretty good - Tres bon/ assez bon Fair/poor - Moyen/mauvais (%) 88 13 88 12 (%) (%) 84 87 16 12 82 80 17 20 TABLE 5. COMPARISON OF STANDARDIZED MORTALITY RATIOS (SMRI <65 YEARS, AND RATIO OF YEARS OF POTENTIAL LIFE LOST (RYPLL), MEXICO AND ARGENTINA WITH REFERENCE TO THE UNITED STATES OF AMERICA, 1982 TABLEAU 5. RAPPORT COMPARATIF DE MORTALIT~ (SMR) <65 ANS ET RAPPORT COMPARATIF D'ANN~ES DE VIE POTENTIELLE PERDUES (RYPLLI, MEXIQUE ET ARGENTINE PAR RAPPORT AUX ~TATS·UNIS D'AM~RIOUE, 1982 SMR <65 ......... . RYPLL .......... . M 184.8 235.0 Mexico - Mexique F 204.5 263.9 cause-specific mortality rates which are current in a developed country, or to the lowest reliable rates ever recorded in any country (10). A country or area's performance may then be assessed using the particular indicator of performance against the standard target, and the quantity of excess mortality, measured as average potential years of life lost per death or per unit of population, is then calculated for each or all causes. As stated earlier, this demands at least reasonably complete recording of the event of death and a knowledge of the size and structure of the population at risk. Accurate knowledge of the underlying medical cause of death will be possible only in the most developed societies. Using this approach for two middle-income coun- tries, it was calculated that in 1982, with reference to existing age- and sex-specific mortality experience in the United States of America, Mexico experienced considerably higher premature mortality and higher loss of potential life years (Table 5) (11). Argentina's experience was similar but not quite so markedly different from that of the United States. The same approach could be utilized to compare the impact on health of progressive urbanization of countries or smaller localities, stratifying for the percentage of the population living in urban situ- ations. In a similar way, particular aspects of urban- ization may be explored; for example, the impact of Total 191.9 245.8 M 132.2 157.4 Argentina -Argentine F 138.0 187.7 Total 134.3 168.4 urban development on the conservation of potential years of life from sentinel categories of conditions like infections and parasitic conditions, or a sub- group of these, e.g. intestinal infectious diseases, which reflect an area's standards of sanitation and nutrition. In this particular respect, Mexico's male population lost 2 702 times more potential years of life in 1982 than would have been expected by reference to the United States, and Argentina's males 418 times more. Few if any studies have looked specifically at the health impact of urbaniza- tion in this way. Where data sources are deficient Most developing countries do not have the capacity to provide population and mortality data in such detail. There have been some imaginative attempts to fill the gaps, however, such as the efforts of Matlab, the rural field station of the Inter- national Centre for Diarrhoeal Disease Research (Bangladesh), which documents vital events in its region with a considerable degree of accuracy and understands the limiting factors associated with its methods of data collection. The tools can be devel- oped only within the context of small-area studies however, but if the small area acts as a typical sentinel area for the wider region or country, it is possible to use its health status experience to mon- itor the impact of developments associated with Rapp. trimest. statist. sanit. mond., 43 (1990) - 149 - urbanization. Using data from Matlab, D'Souza re- cently demonstrated the use of a Preventable Death Index (PDI) (12), the principle of which is to take a measure of mortality experience, such as the infant mortality rate, and to place it in the context both of how much worse the mortality experience on that indicator might have been (for example, an infant mortality rate of 300 per 1 OOO live births), and how much further improvement might be anticipated be- fore the "hard rock" level was reached at which further socioeconomic, environmental and elemen- tary preventive health measures produce diminish- ing returns. The index takes into account the fact that changes in the infant mortality rate (IMR) are limited in their usefulness as an overall mortality indicator for guiding health policy, since falls in the IMR are not evenly distributed in terms of the dif- ficulty of controlling mortality. D'Souza points out that a drop of 10 points from an initial IMR level of 180 per 1 OOO is much more easily achieved through health intervention than an equivalent drop from an initial level of, say, 55 per 1 000. He transforms the IMR logistically to an Infant Mortality Index (IMI) which lies on a scale of O (equivalent to an IMR of 30) to 100 (IMR = 300) on which the intervals better portray the difficulty involved in disease control. His Preventable Disease Index for a study population is then IMI for the study population minus the IMI for a standard population, the latter reflecting the "hard rock" level of IMI. This index has clear advantages for comparing the experience of an area on two occasions spanning a period of urbanizing development, or two areas sub- jected to different rates or levels of development. For example, as D'Souza shows, the IMR for China fell from 252 to 45 per 1 OOO between 1950 and 1982, but measured by the IMI, the fall was from 94.6 to 50.3, giving a better perspective of the achievement. Sim- ilarly, as Table 6 shows, the achievement of further reduction of the 1950-level IMR in Sri Lanka com- pared with that of Bangladesh was likely to be more difficult, and viewing the achievement of the lower rates in each country by 1980 respectively through the medium of the IMI is a better indicator of the dimension of Sri Lanka's achievement. Urbanization and disability Measuring the consequences of disease Reference was made earlier in this article to a study from Spain which reported a higher prevalence of severe mental subnormality among a rural com- pared with an urban population. The mentally sub- normal were integrated into rural society more easily however. Thus, the impairment of mental deficiency led to a limitation in the powers of those affected to perform tasks which required complex reasoning, judgement or initiative; that is, they were disabled in that respect. Nevertheless, the rural indi- viduals integrated more easily, i.e. they were less handicapped in their social relationships. The sequence of conditions - impairment, disability, handicap - is the basis of the International Classifi- cation of Impairments, Disabilties and Handicaps (ICIDH) (13). This classification provides a framework for dealing with the consequences of ill-health and as such, and because the classification does not depend upon the accurate, or even any ascertain- ment of the underlying clinical cause, the impact on the health of the population of an area, in which rapid change such as that resulting from urbaniza- tion is taking place, may be described and compared at intervals in terms of what the inhabitants can or cannot do for themselves. There are no international census or survey recom- mendations however which would harmonize the classification of disabled people. Moreover, there are a variety of ways of identifying disabled or potenti- ally disabled persons in surveys. Some of these are intended to screen out persons with impairments, others those with disabilities or handicaps. In the framework of the United Nations World Pro- gramme of Action concerning Disabled Persons, the International Disability Statistics Data Base (DISTAT) was developed from a review of surveys and regis- tration systems in various countries (14). This is leading to the development of international stand- ards for the use in surveys of ICIDH and for classify- ing causes of impairment, technical aids used by disabled people and services received (15). Disability-free life expectancy Representative population surveys of disability, usu- ally involving person-interviewing, will produce sex- age-specific prevalence levels of disability. The risk of being disabled, or more disabled, at subsequent quinquennial or decennial points may be computed therefore and, as for the calculation of life expect- ancy by applying the reciprocal approach, the likeli- hood of remaining disability-free through the subse- quent periods may also be computed. If the average life expectancy at birth, or average further life ex- pectancy at a particular age, is ascertainable from accurate age-specific mortality data, the two meas- ures may be combined to produce an index of disability-free life expectancy. Thus in this way it may be calculated that males in England & Wales in 1985 had a life expectancy at birth of 71.8 years and that, on average, 58.7 of these years would be free of disability (16). TABLE 6. COMPARATIVE INFANT MORTALITY RATE (IMR) AND INFANT MORTALITY INDEX (IMI), BANGLADESH AND SRI LANKA, 1950 AND 1980 TABLEAU 6. TAUX DE MORTALITE INFANTILE (/MR) ET INDICE DE MORTALITE INFANTILE (IMI), BANGLADESH ET SRI LANKA, 1950 ET 1980 Bangladesh . . . . . . . . . . Sri Lanka .......... . Wld hlth statist. quart., 43 (1990) IMR per 1 OOO live births-pour 1 OOO naissances vivantes 1950 179.5 90.7 1980 132.6 40.6 Difference Difference 46.9 50.1 1950 84.9 67.2 IMI 1980 76.8 47.9 Difference Difference 8.1 19.3 - 150 - Data from disability surveys now exist for more than 50 countries (with the above caveats however), and they are, for the most part, countries with reliable mortality statistics and some estimates of the pro- portion of the population who are urbanized. There is as yet however no systematic report of inter- national comparisons of urban/rural differences in disability, nor any longitudinal study which has re- ported any changes in levels of disability according to changes in the degree of urbanization. Does urbanization create a health burden? The concept of urbanization is, in its way, an ap- propriate predictor or correlate of health status, but only in the same way as socioeconomic status. Higher socioeconomic categories are usually asso- ciated with better health status than low categories, but within the higher categories some groups fare less well, such as women aged over 50 years, who are more likely to develop breast cancer than their less-advantaged counterparts. In general, urbanization raises the overall status of public health. Clearly, there are groups who fare less well than they might have done in a non-urbanized environment, such as casualties of road-traffic ac- cidents, or those who find themselves as the urban homeless, even in developed countries (17). The real challenge is how to tease out from the mosaic of different sickness entities what is actually attribut- able to becoming "urbanized". The countries in which urbanization will continue rapidly well into the next century are, in the main, those whose national information systems are less well developed and whose outputs are incomplete. There is an opportunity, therefore, to exploit those systems which already exist or to construct a limited number of sentinal community health status measur- ing stations at vantage points in urbanizing areas and in corresponding places in the same region or country which are not already directly affected, but may be affected indirectly by migration to the city. The examples from Bangladesh and Spain quoted above indicate the approach that might be adopted. Enumeration and health monitoring of these sentinel populations is likely to be practicable, affordable and cost-effective. The measuring and analytical tools are being forged, as we have seen. Opportunities can be created to produce information which will assist in appraising how urbanization may be organized so that the maximum health bonus may in future be sought and the worst of the health problems avoided. SUMMARY Several components of urbanization influence health status, but it is difficult to attribute changes in health status to any particular component. The overall impact may be estimated by relating the degree of urbanization of populations to some proxy measure, like the under-5 mortality rates. In this respect the net effect of urbanization is shown to be beneficial. A variety of survey and computational methods have been used to clarify the relationship. Some illustrate the effects of urbanization upon particular clinical conditions, such as promoting the eradication of leprosy, others of particular components, such as overcrowding and pollution, on infant mortality. To help set goals, excess or avoidable mortality may be computed for a country or region by relating its experience to current mortality levels in a developed country; and changes in the levels of avoidable mortality from sentinel conditions such as infectious diseases may be related to changes in particular aspects of urbanization, e.g. improvements in levels of sanitation. Migration to the urban environment imparts the tendency to acquire the health characteristics of the host population. Rapid urbanization causes problems of psychosocial adjustment for older children. Urbanization may impact upon the incidence and prevalence of disability. Where sex-age disability- survey data exist, they may be combined with age- specific mortality data to construct an index of dis- ability-free life expectancy, a more subtle measure for assessing the progressive impact of urbanization. Up to now, however, there have been no inter- national studies of how levels of disability change according to the progress of urbanization, and there are no international census or survey recom- mendations for harmonizing the classification of dis- abled people. Only approximately one-third of the world's countries have the capacity to mount such studies at present. There has been sufficient progress in establishing typical small-area sentinel populations to suggest that a series of such monitoring stations could be used to measure the impact on health of different degrees of urbanization or other socioeconomic changes. Rapp. trimest. statist. sanit. mond., 43 ( 19901 - 151 - RESUME Appreciation de l'impact de l'urbanisation sur la sante Plusieurs elements de !'urbanisation influent sur l'etat de sante, mais ii est difficile d'attribuer les changements observes a tel ou tel d'entre eux en particulier. II est possible d'evaluer l'impact global en rappor- tant le degre d'urbanisation des populations a une mesure de substitution, par exemple les taux de mortalite chez les mains de 5 ans. Dans ce domaine, !'influence nette de !'urbanisation apparait comme benefique. Differentes methodes d'enquete et de calcul ant ete utilisees pour preciser cette relation. Certaines illus- trent les effets de !'urbanisation sur certains etats cliniques (eradication de la lepre par exemple), d'autres d'elements particuliers, tels le surpeuple- ment et la pollution, sur la mortalite infantile. Pour aider a fixer des objectifs, on peut calculer la surmortalite ou mortalite evitable pour un pays ou une region en rapportant les effets observes dans le pays OU la region consideres aux taux de mortalite dans un pays developpe; !'evolution des taux de deces evitables dus a certaines affections sentinelles, telles les maladies infectieuses, peut etre rapportee aux changements concernant certains aspects de !'urbanisation, par exemple a !'amelioration des con- ditions d'hygiene. Les migrations vers les villes sont associees a une tendance a acquerir les caracteristiques sanitaires de la population d'accueil. L'urbanisation rapide est cause de problemes d'adaptation psychosociale chez les enfants plus ages. L'urbanisation peut avoir un effet sur !'incidence et la prevalence des incapacites. Lorsqu'on dispose de donnees d'enquetes sur les incapacites selon le sexe et l'age, on peut les associer aux donnees de mor- talite en fonction de l'age pour construire un indice de l'esperance de vie sans incapacites, mesure plus subtile d'evaluation de l'impact progressif de !'urbanisation. Jusqu'ici cependant, aucune etude internationale n'a ete consacree a la fac;:on dont les taux d'incapacite evoluent selon le rythme de !'urbanisation, et ii n'existe ni recensement inter- national ni recommandations pour harmoniser la classification des handicapes. Environ un tiers seule- ment des pays du monde aurait les moyens d'orga- niser de telles etudes a l'heure actuelle. La constitution de populations sentinelles carac- teristiques sur des petites superficies a marque suf- fisamment de progres pour qu'il soit possible d'envisager d'utiliser de telles stations de sur- veillance pour mesurer l'impact sur la sante de differents degres d'urbanisation ou d'autres modi- fications socio-economiques. REFERENCES- REFERENCES 1. UNITED NATIONS CHILDREN'S FUND. State of the world's children 1989. Oxford, Oxford University Press, 1989. FONDS DES NATIONS UNIES POUR L'ENFANCE. Situa- tion des enfants dans le monde 1989. New York, UNICEF, 1989. 2. ADEGBOLA, 0. The impact of urbanization and in- dustrialization on health conditions: the case of Nigeria. World health statistics quarterly, 40 (1 ): 74-83 (1987). ADEGBOLA, 0. L'impact de !'urbanisation et de !'industrialisation sur la situation sanitaire: l'exemple du Nigeria. Rapport trimestriel de statistiques sanitaires mondiales, 40 ( 1 ): 7 4-83 (1987). 3. HUNTER, J. M. & THOMAS, M. 0. Hypothesis of leprosy, tuberculosis and urbanization in Africa. Social science and medicine, 19 (1 ): 27-57 (1984). 4. HOLLAND, W. W. & REID, D. D. The urban factor in chronic bronchitis. Lancet, i: 445-448 (1965). 5. DIAZ-FERNANDEZ, F. & GESTAL-OTERO, J. J. The influ- ence of habitat on the prevalence of mental handicap. International journal of epidemiology, 16 (1): 52-56 (1987). 6. PAPOZ, L. ET AL. Diabetes mellitus in Tunisia: description in urban and rural populations. Inter- national journal of epidemiology, 17 (2): 419-422 (1987). Wld hlth statist. quart., 43 (1990) 7. RAHIM, S. I. A. & CEDERBLAND, M. Effects of rapid urbanization on child behaviour and health in a part of Khartoum, Sudan. Journal of child psy- chology and psychiatry and allied disciplines, 25: 629-641 (1984). 8. BRUNSWICK, A. F. & MERZEL, C. R. Health through three life stages: a longitudinal study of urban• black adolescents. Social science and· medicine, 27 (11): 1207-1214 (1988). 9. V1cTERA, C. G. ET AL. Water supply, sanitation and housing in relation to the risk of infant mortality from diarrhoea. International journal of epi- demiology, 17 (3): 651-654 (1988). 10. UEMURA, K. Excess mortality ratio with reference to the lowest age-sex-specific death rates among countries. World health statistics quarterly, 42 (1): 26-41 (1989). UEMURA, K. Rapport de surmortalite fonde sur les plus bas taux de mortalite des pays par sexe et par groupe d'age [resume]. Rapport de statis- tiques sanitaires mondiales, 42 (1): 40 (1989). 11. PLAUT, R. & ROBERTS, E. Preventable mortality: in- dicator or target? Applications in developing countries. World health statistics quart0 rly, 42 (1): 4-15 (1989). PLAUT, R. & ROBERTS, E. La mortalite evitable: indi- cateur ou objectif? Applications dans les pays en developpement [resume]. Rapport trimestriel de statistiques sanitaires ~ondiales, 42 ( 1 ): 15 (1989). - 152 - 12. D'SouZA, S. The assessment of preventable in- fant and child deaths in developing countries: some applications of a new index. World health statistics quarterly, 42 (1 ): 16-25 (1989). D'SouZA, S. L'evaluation des deces evitables de nourrissons et d'enfants dans les pays en deve- loppement: quelques applications d'un nouvel indice [resume]. Rapport trimestriel de statis- tiques sanitaires mondiales, 42 (1): 25 (1989). 13. WORLD HEALTH ORGANIZATION. International Classi- fication of Impairments. Disabilities and Handi- caps - a manual of classification relating to the consequences of disease. Geneva, WHO, 1980. ORGANISATION MONDIALE DE LA SANTE. Classification internationale des handicaps: deficiences, inca- pacites et desavantages - un manuel de classi- fication des consequences des maladies. Paris, CTNERHI-INSERM, 1988. 14. UNITED NATIONS. United Nations Disability Statis- tics Data Base, 1975-1986 - technical manual. New York, Department of International Economic and Social Affairs, Statistical Office, 1988. 15. CHAMIE, M. Survey design strategies for the study of disability. World health statistics quarterly, 42 (3): 122-140 (1989). CHAMIE, M. Methodes d'enquete sur l'incapacite [resume]. Rapport trimestriel de statistiques sanitaires mondiales, 42 (3): 138-139 (1989). 16. BEBBINGTON, A. C. The expectation of life without disability in England and Wales. Social science and medicine, 27 (4): 321-326 (1988). 17. ROPERS, R. H. & BOYER, R. Perceived health status among the new urban homeless. Social science and medicine, 24 (8): 669-678 (1987). Rapp. trimest. statist. sanit. mond., 43 ( 1990) - 153 - URBAN AIR POLLUTION IN LATIN AMERICA AND THE CARIBBEAN: HEAL TH PERSPECTIVES Isabelle Romieu," Henyk Weitzenfeld," & Jacobo Finkelman• Over the last few years, urban development and growing industrialization have resulted in air pollu- tion becoming a major issue in some countries of Latin America and the Caribbean. One of the causes is the rapid increase in the size of the urban popula- tion due to a continued high fertility rate and to internal migration from rural areas. Both increasing urbanization and the continued growth of megacities have generated wastes that pollute air and water and degrade renewable resources. In addition to industrial processes often concentrated in the cities, vehicle emission and stationary-source fuel combus- tion have become the primary sources of air pollu- tion in Latin American cities. Although air quality standards have been established in some Latin American countries, these are frequently exceeded. This article reviews the problem of air pollution in Latin America and the Caribbean and the potential for adverse health effects in their population. Health effects of air pollution Adverse health effects have been associated with three major types of air pollution related to three major source categories: (i) the sulfur oxide and particulate complex arising from the combustion of sulfur-containing fossil fuels, particularly in fossil- fueled power plants; (ii) photochemical oxidants formed in the atmosphere from a complex chemical reacti.on between precursor hydrocarbon compounds and nitrogen oxides which are largely related to motor-vehicle emission, as well as carbon mon- oxide; and (iii) a miscellaneous class of air pol- lutants such as hydrogen sulfide, lead and cadmium, which are mostly emitted by localized point sources, such as smelters, refineries and manufacturing plants (1), although fossil fuel is still a major source of lead in most Latin American countries. Su/fur oxide and particulate matter Short-term exposure effects. Variation in the 24-hour average of sulfur dioxide (S02) and total suspended particulates (TSP) have been associated with in- creased mortality and morbidity, and reductions in pulmonary functions (2). During the first half of this century, episodes of marked air stagnation have resulted in well- documented acute excesses of mortality in areas where fossil-fuel combustion produced very high levels of S02 and TSP. In one notable episode in London, concentrations of S02 and smoke rose above 500 µg/m3• The people primarily affected were those with preexisting heart or lung disease or both, and the elderly, although with a closer look at the • Pan American Health Organization, Pan American Center for Human Ecology and Health, Mexico. ~ FEV1 = forced respiratory volume after 1 second. Wld hlth statist. quart., 43 (1990) data, it seems that children < 5 were also severely affected (3). Following this major episode, attention was turned to studies on more moderate day-to-day variations in mortality within large cities in relation to pol- lutants (4). These acute mortality studies of S02 and particulates suggested a dose-response relationship between 24-hour levels of these pollutants and ex- cess mortality, particularly at values over 500 µg/m3 (5). However new analysis of the London winter data 1958-1959 to 1971-1972, controlling for important confounding variables such as temperature and humidity, indicates the absence of a threshold level for the adverse effect of British smoke, and a statis- tically significant pollutant effect on mortality below 150 µg/m3 was observed (6). Similar results were reported by Schwartz & Marcus who noted a signifi- cant association between mortality and British smoke to below 100 µg/m3 (7). Fig. 1 shows excess mortality in relation to particulate matter. Major episodes are shown on the right-hand side of the graph (with excess mortality ranges, where avail- able). The effect of day-to-day variation analysed with a time series is displayed as lines indicating the slope of excess mortality (8). From these figures it can be derived that an increment of total suspended particulates (TSP) from 100 µg/m3 to 300 µg/m3 would lead to an excess mortality ranging from 3% to 6%. Short-term peak concentrations of sulfur dioxide and particulate matter may also increase morbidity, es- pecially in individuals with higher sensitivity than the general population, such as those with asthma and bronchitis. Based on epidemiological studies among these populations, a minimum level of black smoke and S02 of 250 µg/m3 is estimated to be needed to produce effects, including exacerbation of symptoms and asthma attacks (2), although effects may be observed at lower levels among highly- sensitive bronchitis patients (9). Results from studies of hospitalization (10, 11) and emergency-room visits also tend to support this association (12). Ostro (13) reported a 3.2% increase in acute respiratory dis- eases in relation with a change of 1 µg/m3 in fine particulate matter (< = 2.5 µg, reference mean = 23.2 µg/m3) in adults aged 18-65. In some studies, observed deviations in pulmonary function levels in children have been associated with short-term fluctuations in particulate concentration (2). From data collected by Dockery et al. (14) during air-pollution episodes, it can be calculated that in the most sensitive children (approximately 25% of his population), there was a deficit in pulmonary func- tion at least 4 times greater compared to those of average sensitivity, corresponding to a decrease in FEV1h of 0.39 ml per each µg/m3 of exposure (TSP). The minimum level for effect was judged to be 180 µg/m3 TSP. Although the significance of such phys- iological change on future lung function is not yet clarified, there is concern that such decrease could lead to chronic lung disease in adulthood. - 154 - FIG.1 RELATIONSHIP OF EXCESS MORTALITY TO TOTAL SUSPENDED PARTICULATES (TSP) RELATION ENTRE LA SURMORTALITE ET LE TOTAL DES PARTICULES EN SUSPENSION (TPS) t :~ <ii t: 0 E Ji I t .E"' <ii t: 0 E "' ~ w 100 ,- 0 10 - 10 100 200 0 0 0 Cl oa •• I I I I 700 1000 Cl Cl Cl Cl Cl 10 OOO Total suspended particulates (µgim') - Total des particules en suspension (µg/m') nme-series analyses - Analyses de series chronologiques Episodes * = Mean values - Valeurs moyennes • • • • • • • = Philadelphia (TSP) - Philadelphie (TPS) - - - - = London - Landres = New York City - Ville de New York Modified from reference (8) -Adapte de la reference (8). Long-term exposure effects. Correlation studies be- tween S02 and particulate levels, and mortality from cardiorespiratory diseases, have usually indicated that this complex accounts for a proportion of ap- proximately 4% of the variation in death rates among cities (2). Many factors, such as differences in smoking habits, occupation or social conditions may contribute to the disparities in death rates attributed to S02 and particulates, but the results of studies carried out in different parts of the world in various years imply a relatively consistent associa- tion between long-term residence in more polluted communities and increased mortality rates (1). How- ever, given the presence of different pollutants, it has been fairly difficult to identify the main contribu- tor to excess mortality. Some authors (8, 15) found that TSP played a major role, while others (16) found that particle exposure measures related to the respir- able or toxic fraction of the aerosol (sulfate fraction of particulates (TS04 )) were those most consistently associated with annual cross-sectional mortality rates. Community-based health studies conducted among adults and children have indicated a detectable in- crease in the frequency of respiratory symptoms and the frequency of respiratory illness in communities where annual mean concentrations of both black smoke and sulfur dioxide exceed 100 µg/m 3 (4). Children represent a high-risk population because of the potential effect of pollutants on lung growth and lung-function development. Several studies investi- gated the relation between respiratory illness and symptom rates, and air pollution levels. In a study conducted among a preadolescent population aged • = Pittsburgh, 1975 Cl = London - Landres O = NewYork 6-9 in six cities of the United States of America (17), frequency of chronic cough was significantly associ- ated with the annual average concentrations of three air pollutants (TSP, TS04 and S02) during the year preceding the examination (p < 0.01 ). The ranges were: 39.3-114.1 µg/m 3 for TSP; 3.4-68.3 µg/m3 for S02; and 5.4-18.8 µg/m3 for TS04• Rates of bronchitis and a composite measure of lower respiratory ill- ness were significantly associated with average par- ticulate concentration (p < 0.05). Similar results have been confirmed in a second cross-sectional survey of the same population (18), and a subset of this population (10%) with asthma or persistent wheez- ing, experienced a higher rate of pulmonary symp- toms in relation to increased pollutants. There was no evidence of impaired pulmonary function asso- ciated with pollutant levels. However such chronic effects may be masked by exposure-associated acute changes in pulmonary function days or hours prior to testing (19). Although results from the Six Cities Study referred to children over 5 years, younger children should be equally or more sensitive to pollutants. This is important because severe lower respiratory infection in childhood has been associ- ated with both persistently lower lung function (20), and with subsequent development of chronic ob- structive pulmonary disease (COPD) (21). The causal role of air pollution for chronic obstruc- tive pulmonary disease (COPD) has never been dem- onstrated conclusively, mainly because of meth- odological difficulties in epidemiological studies. Nonetheless, there is some evidence that chronic exposure to smoke may play an important role (22). Lambert et al. (23) studied 10 OOO British residents Rapp. trimest. statist. sanit. mond .• 43 ( 1990) - 155 - aged 35-69 in England, Wales and Scotland. These authors reported that age-standardized prevalence rates for symptoms increased with heavier air pollu- tion, independent of cigarette smoking. They also noted a synergic effect between cigarette smoking and air pollution. Other correlation studies have been undertaken in adult populations in different countries. Although the interpretation of such studies is limited by the crude exposure measure- ment, results are consistent and suggest that areas with higher pollution levels are associated with higher prevalence rates of chronic bronchitis (4). In developing countries, prevalence rates of chronic bronchitis often appear to be much higher than in industrialized countries and to have sex ratios tend- ing to 1, which cannot be explained solely by cigar- ette smoking (22). Although exposure to multiple risk in developing countries may be much higher than in developed countries, these results suggest that· atmospheric and indoor smoke pollution may largely account for the differences (an exposure much more common among women and children). Effect of photochemical oxidant complex, nitrogen dioxide and carbon monoxide Ozone and other photochemical oxidants. The ob- served health effects of photochemical-oxidants ex- posure cannot be attributed only to oxidants be- cause photochemical smog typically consists of ozone (03), N02, sulfate and other agents. These pollutants may have additive or synergistic effects on human health, but ozone appears to be the most biologically active (24). Most of the studies on the health effects of 0 3 have focused on the effect of short term (1-2 hour) exposure. However, recent research has shown that effects can be produced by exposures as short as 5 minutes, and that various effects become pro- gressively larger as exposures at a given concentra- tion are extended in time up to 6.6 hours. The inhalation of 0 3 causes concentration-dependent de- creases in the average lung volumes and flow rates during forced expiratory manoeuvres, and mean de- crements increase with increasing depth of breath- ing (25). Changes in pulmonary functions associated with 1-3 hours of 0 3 exposure in normal subjects during exercise have been reported for different parameters (26). The severity of respiratory and other symptoms parallels the impairment of pul- monary function. 0 3 can also induce increased non- specific airway sensitivity to inhalation challenge testing with bronchoconstrictive agents (2). Field and epidemiological studies have indicated a number of acute effects of 0 3 and other photochemi- cal oxidants. Effects which have been associated with hourly average oxidant levels beginning at about 200 µg/m3 (0.10 ppm) include: eye, nose and throat irritation, cough, throat dryness, thoracic pain, increased mucous production, rales, chest tightness, substernal pain, lassitude, malaise and nausea. Although smokers and subjects with preexisting pul- monary diseases do not appear to be more sensitive than others to 0 3, within the apparently normal population there is a range of responsiveness to 0 3 that is reproducible (26). Several authors (10, 11, 27) studied hospital admission rates for respiratory symptoms in relation to 0 3 levels and reported an increase in the relative risk of asthmatic attack (27) c FVC = forced vital capacity. d PEFR = peak expiratory flow rate. Wld hlth statist. quart., 43 ( 1990) and an increase in hospital-admission rates for re- spiratory diseases (10, 11) in relation to high 0 3 levels. Decrease in pulmonary functions in children and young adults have been reported at hourly average 03 concentrations in the range of 160- 300 µg/m3• Based on estimates from Spector et al. (28), moderate physical activity for a range of 03 exposure from 0.019 to 0.113 ppm for 1 hour could lead to a decrement of 1.0 ml/ppb for FVCC and 1.4 ml/ppb for FEV1b; this would result in a decre- ment of 180 ml for FVC and 250 ml for FEV1 for a concentration of 0.2 ppm 03. From these data, aver- age decrements in FVC, FEV1 and PEFRd of 4.9%, 7.7% and 17% respectively were predicted for the current EPA standard for 03 (0.12 ppm). The effect on pulmonary function appears transient and some authors suggest "adaptation" to chronic exposure (29). The potential adverse effect of such adaptation is not known. Different experimental studies in ani- mals and humans have shown that 0 3 increases airway permeability and particle clearance, causes airway inflammation and decreases in bactericidal capacity, as well as morphological alteration in the lung (25). The long-term health effect of exposure to high levels of 03 is still unclear, but there is good reason for concern that repeated insults of this type could lead to chronic impairment of lung develop- ment and function. The long-term exposure effects of 0 3 have been investigated in epidemiological studies of popula- tions living in southern California, which suggest that chronic oxidant exposures of subjects aged 19-59, who never smoked, do affect baseline respira- tory function. In a study conducted in two com- munities from this region with different levels of 03 exposure, baseline pulmonary functions were lower and there was a greater rate of decline in lung function over 5 years in the high oxidant community (30). There is some evidence that 0 3 may act syn- ergistically with other pollutants, such as sulfates or N02 (31). Euler et al. (32) evaluated the risk of chronic obstructive pulmonary disease due to long- term exposure to ambient levels of total oxidants and N02 in a cohort of 7 445 Seventh-day Adventists who had both resided in California for at least 11 years, and were aged at least 25. The results suggest a significant association between chronic symptoms and total oxidants above 0.10 ppm. However when cumulative exposure to TSP was entered in the model, only TSP exposure above 200 µg/m3 showed statistical significance. Data from Kryzanowski et al. (33) suggested the absence of a threshold value for 0 3 effects on diurnal variability of peak expiratory flow rate. In a study conducted in a representative community population sample of adults and children, the variability of ventilatory flow in children significantly increased with increasing 0 3, and the effects were 5 times greater in asthmatic children than in the others. In adults, effects of high 03 days were seen only in non-asthmatics. Nitrogen dioxide (N02). The health effect of N02 has been mainly related to a potential increase in respir- atory infection among children. The major studies were done in the neighbourhoods of chemical plants (34) and reported an increased risk of lower respira- tory illness in children who resided in high N02 exposure areas. Other outdoor sources of N02 are exhaust from petrol and diesel engines; N02 effects from these sources cannot be readily separated from those of other potentially toxic agents in the exhaust nor from those of substances secondarily generated by photochemical reactions and therefore difficult to quantify. - 156 - Epidemiological studies have investigated several outcomes related to N02 exposure including respira- tory illness and symptoms in children, airway reac- tivity in asthmatics and changes in lung function, but results are inconsistent (35). Speizer et al. (36) in- vestigated the effect of indoor N02 concentration due to gas-stove use and reported an average rate difference between exposed and non-exposed of 32.5 per 1 OOO children for respiratory infection be- fore age 2. These results were not confirmed by other studies which did not find a significant effect of the use of gas-combustion appliances (4). Carbon monoxide. Health effects related to carbon monoxide are due to the capacity of this pollutant to bind to haemoglobin and form carboxyhaemoglobin (HbCO). At a low HbCO level (< 10%), effects are mainly cardiovascular among subjects with preexis- tent chronic angina (seen at levels of HbCO of 2.9-4.5%), and neurobehavioural with impairment of vigilance (at 5% HbCO), exhaustion in exercising young adults (at 5% HbCO). and dizziness (at 10-15% HbCO) (37). Although health effects have been re- ported according to lowest observed levels, it seems that decreasing exposure can lead to decreasing health effects on a continuous mode below 2.9% HbCO. Although there is no adequate evidence that carbon monoxide per se is directly involved in the path- ogenesis of cardiovascular and chronic obstructive pulmonary disease, studies carried out among cigar- ette smokers suggested that HbCO plays a role in the genesis of these diseases. Exposure of pregnant women and their fetus has been related to reduced birthweight and retarded postnatal development (38). This is of primary interest because in many developing countries pregnant women are exposed to high levels of carbon monoxide, mainly due to indoor biofuel combustion. Lead. Blood-lead levels in both children and adults increase in the presence of this multimedia contami- • United Nations Environment Programme. [First preliminary report on the state of the environment in Latin America and the Carribean - Air quality]. Mexico City, UNEP, 1987. (In Spanish). nant in the air. Automobile exhaust is an important source of lead exposure in the urban environment and results in contamination of road dust, clothes, hands and inhalable particulates (39). The decline observed in blood-lead levels in urban children dur- ing the period of substitution of leaded by unleaded petrol suggests that reducing the amount of lead in petrol is an important strategy for reducing overall lead exposure. The populations at highest risk of adverse effects are women in reproductive years and small children, because even relatively low blood- lead levels (< = 10 µg/dl) may lead to impaired neuro- logical development (40). Average level of outdoor air pollutants In some countries of Latin America and the Caribbean, air-pollution problems are becoming an environmental priority (Table 1). The urban areas more affected by anthropogenic pollutant emissions are: the area of Sao Paulo in Brazil, the city of Santiago in Chile and the metropolitan area of Mexico City. It is important to mention that similar situations could well occur in other cities of Latin America and the Caribbean.• A summary of annual averages for total suspended particulates (TSP) and sulphur dioxide (S02) for specific cities of Latin America are presented in Figs2 & 3 (41-48). The geometric mean levels of TSP at individual sites ranged from 39 µg/m 3 to 520 µg/ m3• For most cities the occurring average values are between 100 µg/m 3 and 300 µg/m 3• These values should be incremented by approximately 10% to estimate arithmetic mean values on which the WHO guidelines for TSP are based. The figures from the majority of the countries with available data are well over the guideline values proposed by a WHO task force (60-90 µg/m 3 ) (4). There is no evidence of a consistent trend among the different cities. The only notable decrease occurred in Cubatao (Brazil) (change in yearly average of Pm 10 from 169 µg/m 3 in 1984 to 104 µg/m 3 in 1988). During the period 1983- 1988, Santiago (Chile) experienced a slight increase in the level of particulates. TABLE 1. HUMAN SETTLEMENTS WITH AIR POLLUTION PROBLEMS AS PRIORITY TABLEAU 1. ETABLISSEMENTS HUMAINS OU LES PROBL~MES DE POLLUTION ATMOSPHERIQUE SONT CONSIDERES COMME PRIORITAIRES Country-Pays Argentina - Argentine Brazil - Bresil . . . . Colombia - Colombie Chile - Chili . . . . . Cuba ........ . Ecuador - Equateur . Guatemala ..... Mexico - Mexique . . Peru - Perou . . . . . . . . . . . . . . . . . . . Dominican Republic - Republique dominicaine Uruguay .. Venezuela .................... . City or region -Villeou region High priority - Priorite elevee Low priority - Faible priorite Sao Paulo Rio de Janeiro Belo Horizonte Bogota Santiago Mexico City (metropolitan area) - Mexico (zone metropolitaine) Guadalajara Monterrey Lima-Callao Caracas (metropolitan area) - Caracas (zone metropolitaine) Gran Buenos Aires Salvador (Bahia) Brasilia Medellin Cali Concepcion Havana - La Havane Guayaquil Guatemala City Puebla San Luis Potosi Aguascalientes Santo Domingo - Saint-Domingue Montevideo Maracaibo Valencia Source: United Nations Environment Programme. [First preliminary report on the state of the environmentin Latin America and the Caribbean - Air quality). Mexico City, UNEP, 1987. (In Spanish) - Programme des Nations Unies pour !'Environnement. [Premier rapport preliminaire sur l'etat de l'environnement en Amerique /atine et dans /es Caraibes - Qualite de l'airj. Mexico, PNUE, 1987. (En espagnol.) Rapp. trimest. statist. sanit. mond., 43 ( 1990) - 157 - FIG.2 SUMMARY OF THE ANNUAL GEOMETRIC AVERAGE OF TOTAL SUSPENDED PARTICULATES (TSP) IN SOME LATIN AMERICAN CITIES TABLEAU RECAPITULATIF DES MOYENNES GEOMETRIQUES ANNUELLES DE TPS (TOTAL DES PARTICULES EN SUSPENSION) DANS CERTAINES VILLES D'AMERIQUE LATINE 10 2 3 4 5 6 7 8 9 100 2 3 4 6 7 1000 San Jose 1983-1986 Caracas 1983-1986 Bogota 1983-1986 Sao Paulo ' 1983-1988 Cubatao' 1983-1988 Santiago 1983-1988 Mexico 1985-1987 Karen olsen Policia metropolitana Min de Salud San Rafael San Jose C.B.S. Sena Artes Nossa Senhora de 6 Congonhas Mooca Guarulhos Sao Gaetano del sul Vila Nova Vila Parisi La Granja Min de Salud Cia de Bomberos Padahuel CFE Nezahualcoyotl Rio de Janeiro Xalostoc 1984-1987 Copa cabana Santa Teresa Sao Joao de Meriti Bonsucesso Range of individual site annual averages - Etendue des moyennes annuelles sur chaque site i Combined site average over the period - Moyenne pour les sites combines sur la p0riode consid0r0e II] I I 60-90 µglm3 a • Guideline values proposed by a WHO task force (arithmetic mean) - Valeurs recommandees par un groupe de travail de l'OMS (moyenne arithmetique). ' Calculated from Pm1 o levels (assuming Pm1 o = 50% TSP) - Calcule a partir des niveaux de Mp1 o (en prenant pour hypothese Mp1 o = 50% de TPS). Sulfur dioxide levels ranged from 16 µg/m3 to 160 µg/m 3• There is an overall decreasing trend in Mexico City, Sao Paulo and Santiago with a de- · crease of 24%, 30% and 28% respectively between I 1985 and 1988. Most of the available data show mean averages below the guideline values proposed by a WHO task force (40-60 µg/m3) with the excep- tion of Mexico City where particularly high levels were noted in the north-east of the city (annual average of 129 µg/m3 in 1987, Xalostoc monitoring station). In relation to intercountry comparison, it should be noted that the data presented here are based on different monitoring systems which were not stand- ardized by an intercomparison study and therefore can only provide an approximation of relative rank- ing. For N02, available data show levels generally below the standard, except for Mexico City where all re- ported monitoring stations show figures well over W/d hlth statist. quart., 43 11990) the United States Environmental Protection Agency (USEPA) standard for a yearly average (100 µg/m3 ). This is cited for comparison as WHO has not suggested an annual guideline value. Data on recent levels of air pollutants are presented according to specific cities and types of measure- ment (Table 2). Results show that in all cities one or more areas probably greatly exceed the proposed guideline values for TSP mentioned above. As pre- viously noted, geometric means are the most fre- quently available data . For Brazil, results refer to the arithmetic mean of particulate matter less than 10 µm (Pm 10) for which the only available standard on a yearly basis is equal to 50 µg/m3 (USEPA). 0 3 was only measured in Mexico and Sao Paulo; the maximum one-hour values exceeded the WHO guidelines for Europe (100-200 µg/m3, 0.05-0.10 ppm) (2) in all reported stations in Mexico City, with 88% of the measurements above 0.11 ppm in the south- west area (Pedregal), and in some areas of Sao Paulo and Cubatao (Brazil). - 158 - FIG. 3 SUMMARY OF THE ANNUAL S02 AVERAGES IN SOME LATIN AMERICAN CITIES TABLEAU RECAPITULATIF DES MOYENN!:S ANNUELLES DE S02 DANS CERTAINES VILLES D'AMERIQUE LATINE 10 2 3 4 5 6 7 8 9 100 2 4 5 6 7 1000 Sao Paulo Nossa Senhora de 6 1983-1988 Congonhas Guarulhos Sao Gaetano del sul Cubatao Vila Nova 1983-1988 Vila Parisi Santiago La Granja 1983-1988 Padahuel Min de Salud Cia de Bomberos Mexico CFE 1985-1987 Nezahualcoyotl Xalostoc Range of indi vid ual site an nu al averages - Etendue des moyennes annuelles su r chaque site 40-60 µg/m3 a i Combined site ave rage ove r the period - Moyenne pour les si tes combines sur la p8riode consid8r8e • Guideline values proposed by a WHO task force (arithmetic mean) - Valeurs recommandees par un groupe de travail de l'OMS (moyenne arith metique). Health implications for Latin America and the Caribbean Population exposed Based on estimates made by UNEP,e high priority must be given to the control of air-pollution prob- lems in several cities of Latin America (Table 1). These cities are: Sao Paulo, Rio de Janeiro and Belo Horizonte (Brazil), Bogota (Colombia), Santiago (Chile), Mexico City, Guadalajara and Monterrey (Mexico) , Lima-Callao (Peru) and Caracas (Venezuela). The populations of these cities repre- sent approximately 81 million people or 26.5% of the total urban population of Latin America and 19% of its total population (49, 50). If the population of cities with less severe air-pollution problems is con- sidered, the at-risk population will reach a total of 119 million people, or 39% of the total urban popula- tion of Latin America. Based on the age distribution of the urban popula- tion of the specific countries previously cited (51), and considering cities with high-priority air-pollution problems, the following information was derived: approximately 30 million children (0-14 years), 47 million adults (15-59 years) and 4 million elderly people (60 years or more) are exposed to air- pollutant levels that exceed WHO guidelines for adequate health protection . Epidemiological studies To date, very few epidemiological studies have been carried out to determine the potential health effects of air pollutants in Latin America. ' FEV25_75 = forced expiratory flow from 25% to 75% of the forced vital capacity. In Mexico, awareness of air-pollution problems and the potential for adverse health effects increased in the early 1970s. Between 1970 and 1973 a study was conducted to determin-e chronic health effects of ai r pollution in four areas of Mexico City with different levels of pollution. Respiratory infections and symp- tomatology were monitored in children , elderly people and subjects with chronic respiratory dis- eases. Comparison between different areas sug- gested an increased respiratory morbidity in areas of higher pollution (52,53). During ensuing years, several studies were con- ducted among schoolchildren in areas of Mexico City with contrasting levels of pollutants. General results suggest that children living in the most polluted areas have a slight decrease in lung func- tion. However, data are cross-sectional and informa- tion on pollutant levels is often missing (54-56) or inaccurate (57). In 1980, a study was conducted in two public elementary schools in Mexico City, one located in the industrial area of Xalostoc (north of Mexico City) and the other in the suburban area of San Lorenzo characterized by less industrialization and large open spaces (south of Mexico City) . Four hundred and sixty-eight children from the first to the sixth grade were screened using questionnaires and spirometry to determine their pulmonary functions; regression slopes for forced vital capacity (FVC) and forced expiratory flow (FEV25_75)f were compared. There was no difference between residents of the two communities in terms of acute or chronic respir- atory symptoms and illness. However, results suggested that young children in Xalostoc (aged < 10) had a lower FVC than their counterparts in San Lorenzo, which could be explained by higher pol- lutant exposure (55). Another study was conducted in schoolchi ldren 6-13 years in three areas of Mexico City-Xalostoc (north), Morazan (centre) and Rapp. trimest. statist. sanit. mond., 43 (1990) - 159 - TABLE 2. SUMMARY OF AIR QUALITY MEASUREMENTS IN DIFFERENT LATIN AMERICAN CITIES TABLEAU 2. TABLEAU RECAPITULATIF DES MESURES DE LA OUALITE DE L'AIR City/monitor Ville/ lieu d 'observation WHO guidelines Recommandations de l'OMS Mexico City- Mexico (1987) Xalostoc Museo .. . . Nezahualcoyotl CFE ... . Pedregal . Bogota (1986) Sena Artes Andes ... Caracas (1986) El Silencio California . . Rio de Janeiro (1987) Sao Joao de Meriti Santa Terea Copacabana Bonsucesso San Jose (1986) Policia metropolitana . Karen olsen . . . . . Santiago (1988) Ministerio de Salud . Providencia . . Padahuel . .. Sao Paulo (1988) Cambuci . .. Santo Andre centro Santo Amaro . . . . Santo Bernardo do Campo Cubatao (1988) Vila Nova Vila Parisi . . City/monitor Ville/lieu d'observation WHO guidelines Recommandations de l'OMS Mexico City- Mexico (1987) Merced . ... Ciudad Estrella Pedregal .. Xalostoc . . Caracas (1985) El Silencio Trinidad . .. Santiago (1988) Cia. de Bomberos La Granja .. . Sao Paulo (1988) Mooca .... Congonhas .. Cerqueira Cesar Lapa .. . . Cubatao (1988) Vila Nova Vila Parisi . . DANS DIFFERENTES VILLES D'AMERIOUE LATINE TSP (µg/ m')' - TPS (µg/m3) • Annual geometric mean Moyenne geometrique Max.24h annuelle 60-90 100-150 490 1 204 250 1 494 250 990 150 355 143 550 180 620 60 250 98 247 43 71 123 448 45 106 66 108 151 268 91 3339 49 1279 242 665 195 1 142 308 975 669 3269 1869 5369 1409 1 0689 1589 1 1609 589 1469 2089 8189 N02 (µglm ' )' Annual arithmetic mean Moyenne arithmetique Max.24h annuelle 100 150 226 620 156 526 151 564 56 125 22 100 88 284 26 85 49 105 62 S02 (µg!m ' )' Annual arithmetic mean Moyenne arithmetique annuelle 40-60 129 77 62 69 126 91 1 1641 38 9 16 47 29 13 18 11 14 O, (ppm)' Max.1 h 0.05-0.10 0.36 0.22 0.34 0.14 0.16 0.10 0.29 0.17 0.14 Max.24h 100-150 369 225 326 122 252 143 29 68 204 84 84 82 74 90 • Total suspended particulates : air quality standards for the annual TSP average are : 80 µg!m 3 in Brazil , 75 µglm' in Chile and 76.8 µg!m' in Colombia ; for 24 h average, the values are : 240 µg/m3 in Brazil, 260 µglm' in Chile, 400 µg/m3 in Colombia, 275 µglm' in Mexico, 300 µglm' in Peru and 260 µg/m3 in Venezuela. - Total des particules en suspension : les normes de qualite de l'air pour la moyenne annuelle du TPS sont les suivantes: 80 µg!m3 au Bresil , 75 µglm' au Chili et 76,8 µglm' en Colombie ; pour les moyennes sur 24 heures, les valeurs sont les suivantes: 240 µglm' au Bresil , 260 µg/m3 au Chili, 400 µglm' en Colombie, 275 µglm' au Mexique, 300 µglm' au Perou et 260 µg/m3 au Venezuela. ' Air quality standards for the annual S02 average are: 80 µglm' in Brazil, Chile and Venezuela; the 24 h average standard value is 375 µglm' in Mexico. - Les normes de qualite de l'air pourles moyennes annuelles de S02 sont les suivantes: 80 µg/m3 au Bresil, au Chili etau Venezuela; la valeur moyennetype sur24 heures est de 375 µg/m3 au Mexique. ' Air quality standards for the annual N02 average are : 100 µglm' in Brazil and Chile, and 80 µg!m3 in Venezuela ; the 24 h average standard is 300 µglm' in Chile. - Les normes de qualite de l 'air pour la moyenne annuelle de N02 sont les suivantes: 100 µg!m' au Bresil et au Chili, et 80 µg/m3 au Venezuela; la valeur moyenne type sur 24 heures est de 300 µg!m' au Chili. ' Air quality standards for 1 h maximum 0 3 are: 0.08 ppm in Brazil and 0.11 ppm in Mexico. - Les normes de qualite de l'air pour la valeur maximum de O, sur 1 heure sont les suivantes: 0,08 ppm au Bresil et O, 11 ppm au Mexique. ' Environmental Protection Agency Standards (no WHO guidelines available) - Normes de I' Environmental ProtectionAgency(pas de recommandations de l'OMS disponibles). ' Values for 1984 - Valeurs pour 1984. • Calculated from Pm10 levels (assuming Pm 10 = 50% TSP) - Calcule a partir des niveaux de Mp10 (en prenant pour hypothese Mp 10 = 50% de TPS). Wld hlth statist. quart., 43 (1990) - 160 - FIG.4 REPORTED PREVALENCE OF RESPIRATORY ILLNESS AMONG CHILDREN AGED 6-13 IN THREE AREAS OF MEXICO CITY, NOVEMBER 1985 - JUNE 1988 PREVALENCE NOTIFIEE DES MALADIES RESPIRATOIRES CHEZ LES ENFANTS DE 6-13 ANS DANS TROIS QUARTIERS DE MEXICO, NOVEMBRE 1985 - JUIN 1988 t Q) u c Q) <ii > •Q) 0:: I l Q) u c Q) <ii > ~ CL 70 60 50 40 30 20 10 * MORAZAN 03 = 0.005 ppm TSP = 100 µglm' n = 68 Lower respiratory illness ~-· ··: :_: · ····. Maladies des voies respiratoires inf€rieures l;l;WlD Upper respiratory illness ~ Maladies des voies respiratoires sup0rieures ililiib * !!:::!'.!:;:; :; : ;:; :; : ;:~: !ilii@illl ~::~!iii! fliiU'.iili!~i :·'. ~'.: '. :~ :~:: illtl:lt! ··iiiifoi PED REGAL 03 = 0.15 ppm TS P = 25 µglm' n = 69 · XALOSTOC 03 = 0.07 ppm TSP = 280 µglm' n = 63 o3 = Ozone (arithmetic mean over the period) - Ozone (moyenne arithmetique sur la periods consideree). TSP = Total suspended particulates (arithmetic mean over the period) - Total des particules en suspension (moyenne arithmetique sur la periods consideree). * = Significant difference (p<0.05) - Difference significative (p<0,05). Pedregal (south)-during the period November 1985- June 1986. During the follow-up period, respiratory illnesses were higher in the southern part of the city (Pedregal) and the northern part (Xalostoc) than in the centre, where the level of contaminants was substantially lower but still exceeded air-quality standards for TSP. An increased number of episodes of infectious conjunctivis were also recorded in Xalostoc (57). More recently, a pilot study (58) was conducted in the south-western part of Mexico City to study the acute effect of 0 3 among 147 children aged 6-9. Spirometry was performed at two-week intervals and the mean and distribution of average lung function over the six-month observation period were compared to those of Mexican-Americans in Texas. Results showed no consistent effect of acute 0 3 exposure on forced expiratory volumes. These findings are not consistent with results from studies in the United States (25); the inconsistency could be explained by differences in activity levels or by differential indoor and outdoor 0 3 levels, resulting in an error in exposure measurement. Alternatively, this result could be due to tolerance to 0 3 acquired from high levels of chronic exposure. Other studies were conducted among populations living close to a cement plant. In one study (56) respiratory symptoms were reported on question- naires by individuals aged 16-65 years living close to the plant and by a control group. Some respiratory symptoms were more frequent in the exposed indi- viduals (living close to the plant) than in the control Rapp. trimest. statist. sanit. mond., 43 (1990) - 161 - group. Pulmonary-function measures in individuals reporting a high prevalence of respiratory symptoms were lower than predicted values. However, this study did not control for smoking status. Another investigation conducted in 1983-19849 studied the prevalence of respiratory symptoms among 295 resi- dents in three different communities in the south- western part of Mexico City. Two of these com- munities were located in the vicinity of a cement factory, the third one was not close to any source of emission. A significant relation was observed be- tween residency length and the incidence of chronic, non-productive cough in the two areas close to the cement plant. Among subjects living far from any major source, acute respiratory symptoms were more frequent and could possibly be related to 0 3 exposure. In Cubatao (Brazil), the main pollution problem is related to particulate matter due to industrial emis- sion. Between 1982 and 1984, the arithmetic annual mean of the respirable fraction of total particulates (Pm10) was well over 100 µ/m3 in one part of Cubatao (annual arithmetic mean of Pm10: 186 µg/m3 in Vila Parisi for 1984). During the same period all emer- gency visits were recorded in the community of Cubatao and the relative share of respiratory emer- gency visits was determined. Following control inter- vention, the level of particulates decreased (annual arithmetic mean 173 µ/m3 in 1985; 165 µg/m3 in 1986; 151 µg/m3 in 1987; and 116 µg/m3 in 1988 in Vila Parisi) and the proportion of respiratory emergency visits in the area of Vila Parisi decreased from 30.7% to 23% in 1987, with a decrease of visits for bronch- itis and asthma from 15% to 11% (59). Another study was conducted in Cubatao among schoolchildren to explore the relation between pul- monary functions and air pollution (60). It included two cross-sectional evaluations of pulmonary func- tions in 1 110 children aged 6, selected by stratified random sampling in 16 schools of Cubatao. The first study was conducted in June 1983 and enrolled 480 children. Among them, 55.3% experienced a reduc- tion of at least one of the ventilatory parameters (FVC,c FEV1,b FEV2s-1s/ PEFR,d MVVh). Another evalu- ation of pulmonary functions was conducted in November 1985 among 630 children of the same schools and showed that only 36.5% of the children presented some altered pulmonary volumes. These results have been attributed to the overall decrease in the level of particulates observed between 1983 and 1985 in Cubatao. However, such a decrease was not observed in all areas of Cubatao, and global data are therefore difficult to interpret. Furthermore, spirometric measurements were realized in two dif- ferent groups of children and at different seasons. Still, these findings are interesting and warrant further investigation. A study on the health effects of air pollutants was conducted in Chile between November 1987 and March 1988. Four schools located close to monitor- ing stations were selected in Santiago, and two others in Los Andes (a city considered as non- polluted). The total sample size was 300 children, aged 9-13. Prevalence of respiratory symptoms, peak-flow measurements and school absenteeism were recorded on 46 randomly-selected weekdays and on 10 high-pollution days. Results were corn- • Vega, S. et al. Respiratory symptoms and air pollution in the southwest region of Mexico City, 1983-84. Secretariat of Environ- mental Health, Mexico. h MW= maximum voluntary ventilation. Wld hlth statist. quart., 43 ( 1990) pared between Santiago and Los Andes. All respira- tory symptoms (cough, hoarseness, wheezing and nocturnal respiratory symptoms) were significantly higher in Santiago than in Los Andes, after con- trolling for confounding variables (bronchial hyper- reactivity, smoking of the mother, pollution within the house). Spirometric measurements (especially FEV1) were also significantly higher in children of Los Andes. Although data on air pollutants were not validated, results suggested an association between cough, nocturnal respiratory symptoms and hoarse- ness, and Pm10 levels (61). · During the same period, a surveillance system for respiratory illness was implemented through private medical practitioners in Santiago and Los Andes. The age-adjusted prevalence for respiratory illnesses (upper respiratory infections-URI; and lower respir- atory infections-LAI) was significantly higher in Santiago than in Los Andes, particularly during the winter season where pollutant levels are the highest in Santiago. URI seemed to be related to particulate, 502 and 03 levels (61). However the analysis, which included pollutant levels, assumed a linear relation between respiratory illnesses and pollutants and did not account for autocorrelation of observations and lag time. Overall the data suggest an association between air pollution and respiratory illnesses, although no causal relation with any specific pol- lutant was clearly established. Hypothetical scenario The objectives of this section are to obtain a rough estimate of the potential health effects of air pollu- tion in Latin America, and to describe a quantitative approach that could be used to estimate the health impact of pollutants in specific situations. Epi- demiological data from Latin America are sparse, but data on population distribution and levels of pollutants are available. Therefore, to determine the potential extent of health effects due to air pollutants in Latin America, a scenario was hypothesized in which subjects living in cities would be exposed to a given level of particulate matter, and data from the literature cited above were used to extrapolate the expected number of events in different strata of the population (see Table 3 & Annex). This pollutant was chosen because it is of main concern in Latin America and data from the literature are the most consistent. The assumptions were as follows: • Population: 81 million with the following age distribution: 0-5 years, 15%; 6-9 years, 13%; 10-14 years, 11%; 15-59 years, 56%; and 60+ years, 5%, with 50% smokers among the elderly (60+). This age distribution is very similar to that observed in the urban population of Latin American coun- tries (51). • Pollutant levels: three different levels were hypo- thesized for total suspended particulates (TSP): (i) a very high level with an annual geometric mean of 250 µg/m3; (ii) a high level with an annual geometric mean of 150 µg/m3; (iii) a mod- erate level with an annual geometric mean of 100 µg/m3• The groups exposed to these different levels of TSP were: (i) 14.5 million exposed to a very high level; (ii) 23.5 million to a high level; (iii) 43 million to a moderate level. The reference guideline value for TSP used to calculate the health impact was 75 µg/m3• - 162 - TABLE 3. HEALTH EFFECTS IN A HYPOTHETICAL POPULATION OF 81 MILLION EXPOSED TO DIFFERENT LEVELS OF TOTAL SUSPENDED PARTICULATES (TSP) TABLEAU 3. EFFETS SUR LA SANTE D'UNE POPULATION HVPOTHETIQUE DE 81 MILLIONS DE PERSONNES EXPOSEES A. DIFFERENTES QUANTITES DE TPS (TOTAL DES PARTICULES EN SUSPENSION) TSP annual mean - Moyenne annuelle de TPS Items - Aubriques 250µglm' 150/Lglm' 100/Lg/m' Total Populations exposed (millions) - Nombre de personnes exposees (millions) Children 0-14 - Enfants 0-14 ans . . . . . . . . . . . . . . . . . . . . . Adults 15-59 - Adultes 15-59 ans . . . . . . . . . . . . . . . . . . . . Elderly 60+ - Personnes a.gees 60 ans et plus . . . . . . . . . . 5.7 8.1 0.7 9.2 13.2 1.1 16.8 31.7 24.0 45.3 2.2 4.0 Total .................................... . 14.5 23.5 43.0 81.0 Excess number - Surmortalite/surmorbidite Mortality (thousands/year) - Mortalite (milliers par an) . . . . . . . . . . Chronic cough in children (million/year) - Toux chronique chez l'enfant 11.5 7.9 4.9 24.3 (millions par an) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 0.76 0.47 2.3 RAAD• in adults (million days/year) - LCTR• (millions de jours par an) Chronic bronchitis in elderly (thousands) - Bronchite chronique chez les personnes a.gees (milliers) . . . . . . . . . . . . . . . . . . . . . . . . 32.0 50.0 21.0 33.0 12.0 65.0 22.0 105.0 • Respiratory-related restriction in workers's activity - Limitation de la capacite de travail due a une affection respiratoire. Total mortality To assess the potential effect of TSP on life span, data derived from Evans et al. (15) were used. These authors summarized cross-sectional mortality studies in a quantitative approach while controlling for major confounding variables. Cross-sectional analysis is likely to capture geographical global dif- ferences of both long- and short-term effects of TSP on mortality, since annual mortality rates by defini- tion include the cumulative effects of short-term episodes (15). Using the estimates provided by the authors that relate death rate to particulate matter, it was derived that annual means of 100 µg/m3, 150 µg/m3 and 250 µg/m3 would lead in our hypo- thetical population to an excess number of deaths of 4 900, 7 900, and 11 500 respectively or a total of 24 300 subjects. If an annual total mortality rate of 5 per 1 000 is assumed in our population (51), ex- posure to TSP could account for as much as 6% of the mortality. However, this estimation is ap- proximate since population structures differ between Western and Latin American countries. Morbidity among children To determine the increment in pulmonary symptoms and diseases among children exposed to outdoor air pollutants, the estimates used were derived by Ware et al. (17), who studied children aged 6-9 living in six cities of the United States. From the coefficient provided by the authors, it was calculated that an- nual means of 100, 150 and 250 µg/m3 would lead to an increases of 2.8%, 8.3% and 19.3% in the fre- quency of chronic cough. If these estimates are applied to the population of children aged 0-14, who are likely to be affected in the same way or even more (62), the excess number of children affected would be approximately 2.3 million. Similar es- timates could be derived for a composite variable including either cough, bronchitis or respiratory ill- ness in the previous year. The excess number of children affected would then reach 4 million. Those children affected chronically are likely to have lower respiratory functions and could then remain impair- ed as adults, and be at increased risk of developing chronic obstructive lung disease. Morbidity among adults To determine the impact of air pollution in adults, estimates provided by Ostro (63) on the association between days of restricted activity (RAD) and air pollutants were used. From his analysis the follow- ing relation can be derived (63): Fine particulate matter-related RAD per person = 0.0048 x average RAD x fine particulate change Assuming that 25% of TSP is composed of fine particulates, the level of such particulates in our scenario would be 62.5 µg/m3 , 37.5 µg/m3 and 25 µg/m3• Given a mean of 19 RAD per person as reference (63), it was estimated that in our hypotheti- cal population, there would be close to 65 million days per year of respiratory-related restriction in workers' activity. Morbidity among the elderly Among the elderly, respiratory morbidity due to air pollutants can be derived from results provided by Lambert et al. (23). These authors compared the prevalence of chronic bronchitis among British sub- jects living in areas with low (100 µg/m3 ) and high (200+ µg/m3 ) annual means of TSP. They observed a prevalence difference of 2% between residents of areas with low and high TSP among non-smokers, and a prevalence difference of 6% among smokers. Although these authors do not provide estimates of the prevalence of chronic bronchitis in areas with TSP annual means exceeding 200 µg/m3, and the shape of the relation between the rate of chronic bronchitis and TSP levels is not known, it seems realistic to assume an increase in the rate of chronic bronchitis with increasing levels of TSP. Assuming that 50% of the elderly are smokers, the following increase in the prevalence of chronic bronchitis was assigned: 1% among subjects moderately exposed, 3% among subjects highly exposed and 7% among subjects very highly exposed. The excess number of chronic bronchitis cases in our population would then be 105 800. Conclusion Some limitations of this approach have to be men- tioned. Firstly, populations differ and the effects observed in a caucasian American population may not be applicable to other populations in which co-factors may act to modify health effects, although Rapp. trimest. statist. sanit. mond., 43 (1990) - 163 - effects are more likely to be minimized in this ap- proach. In this regard, the often considerable levels of ir'ldoor air pollutants encountered in many coun- tries of the developing world (64) may act cumu- latively or synergistically. Secondly, the composition of particulates is likely to be different in countries in which coal burning is not the major source of com- bustion. Thirdly, the inhalable part of these total particulates will also differ, implying a modification of the health effects. Finally, results presented should be interpreted with caution because some of the estimates are based on a single study; in addi- tion, extrapolations have been made outside the range of TSP observed in these studies. However, this scenario is conservatively based on existing data, and pollution levels are those observed in some developing countries. In this context, the health effect as estimated here is considerable and warrants priority control interventions. Further epi- demiological studies are urgently needed in Latin America to evaluate the health impact of specific pollutant compounds as well as their interactions in populations exposed to high levels of pollution, and to assess the effectiveness of interventions. ANNEX Total mortality The dose-response function derived by Evans et al. (15) is as follows: TSP-related annual deaths per 100 OOO population = 0.45 x TSP change. Taking 75 µg/m3 as guideline value, it was derived that the excess number of deaths at 100 µg/m3, 150 µg/m3 and 250 µg/m3 would be: 0.45 x 25 = 11.3 per 100 OOO or 4 859 in a population of 43 million; 0.45 x 75 = 33.8 per 100 OOO or 7 943 in a population of 23.5 million; 0.45 x 175 = 79 per 100 OOO or 11 455 in a popula- tion of 14.5 million. The total excess number of deaths would then be: 4 859 + 7 943 + 11 455 = 24 257. Assuming a mor- tality rate of 5 per 1 OOO (which is a reasonable es- timate in Latin American countries (51)) this number corresponds to: 24 257/(0.005 x 81 OOO OOO]= 0.06 or 6% of the annual mortality. Chronic cough among children Based on estimates from Ware et al. (17), the follow- ing relation can be derived: TSP-related change in chronic cough (%) = bx PR x (1-PR) x TSP change where b = 0.01; PR = average probability of chronic cough. In this study (17), the average frequency of chronic cough was 12.5% for an average TSP annual mean of 114 µg/m 3• Using the relation previously pre- sented, it was then derived that the difference in the frequency of chronic cough between an area with an average TSP annual mean of 114 µg/m3 and of 250 µg/m3 would be: 0.01 x 0.125 x (0.875) x (250 - 114) = 0.15 or 15%. And similarly for annual TSP means of 150 µg/m 3 , 100 µg/m3 and 75 µg/m 3 (guideline value), the dif- ferences in the frequency of chronic cough would be + 4%, - 1.5% and - 4.3%, corresponding to a preva- lence of chronic cough of: 27.5% (12.5% + 15%) for an annual TSP mean of 250 µg/m3 and similarly for annual TSP means of 150 µg/m 3, 100 µg/m3 and 75 µg/m3 to a prevalence of 16.5%, 11% and 8.2%. Taking the prevalence of chronic cough calculated for the guideline value (75 µg/m3), it was then derived that the excess frequency of cough for Wld hlth statist. quart., 43 (1990) 100 µg/m3, 150 µg/m3 and 250 µg/m3 would be 2.8% (11% - 8.2%), 8.3% and 19.3%, respectively. In the hypothetical population studied here, children aged 0-14 represent 39% of each group of popula- tion exposed to different TSP levels. Applying the rates calculated for each TSP level, the excess number of children with chronic cough would be 2 334 100 (in very high level areas: 1 100 100; in high level areas: 763 600; in moderate level areas: 470 400). Respiratory-related restriction in workers' activity (RRAD) Based on estimates from Ostro (63), the following relation can be derived: FPM-related RAD/person/year= 0.0048 x average RAD x FPM change in annual average where: FPM = fine particulate matter; RAD = restricted activity day; average RAD/person/year= 19 days (estimate for the United States of America). Assuming that 25% of the TSP is composed of fine particulates (< 2.5 µm), it was estimated that the guideline value for FPM should be approximately 20 µg/m3 (75 µg/m3 x 0.25) and that the levels of fine particulates in this scenario would be: 62.5 µg/m3, 37.5 µg/m3 and 25 µg/m3, respectively. The difference between the level of fine particulates in the reference situation and our scenario is there- fore: 42.5 µg/m3 (62.5 - 20), 17.5 µg/m3 and 5 µg/m3• Given an average RRAD of 19 days/person/year, the excess number of days/person/year of respiratory- related restricted activity would then be: 3.9 (0.0048 x 19 x 42.5), 1.6 and 0.5, which correspond in the total adult population to 64 710 OOO days/year ((3.9 x 8 100 OOO) + (1.6 x 13 200 OOO) + (0,5 x 24 OOO OOO)]. Chronic bronchitis among the elderly Estimates derived from Lambert et al. (19) were used. These authors contrast two areas with dif- ferent levels of pollution: - low pollution level= 100 yearly average TSP; - high pollution level = 200+ yearly average TSP. - 164 - The prevalence of chronic bronchitis in subjects aged 65-69 is: 4% in the high-pollution area and 2% in the low-pollution area, corresponding to a preva- lence difference of 2% among non-smokers; and 10% in smokers in the low-pollution area and 16% in the high-pollution area, corresponding to a prevalence difference of 6% (19). If 50% of the subjects in this population are smokers, the prevalence difference of chronic bronchitis between high- and low-pollution areas is then 4%, corresponding to a change of an- nual TSP average of 100 µg/m3• From an assumed linear relation between TSP and chronic bronchitis, it was derived that the prevalence of chronic bronchitis would be 5%, 6%, 8% and 12% in areas with 75µg/m3, 100µg/m3, 150µg/m3 and 250 µg/m3 TSP annual means, respectively. In the scenario presented here, the population aged 60+ includes 725 OOO subjects in areas with a very high level of TSP (250+ µg/m3); 1 million subjects in areas with a high level of TSP (150 µg/m3); and 2.2 million subjects in areas with a moderate level of TSP (100 µg/m3). The excess proportion of chronic bronchitis (compared to the prevalence in the area with 75 µg/m3 ) would be: 7% in areas with very high levels of TSP; 3% in areas with high levels of TSP; and 1% in areas with moderate levels of TSP. The excess number of chronic bronchitis cases would be: 105 750 subjects ((725 OOO x 0.07) + (1 100 OOO x 0.03) + (2 200 OOO x 0.01 )]. Calculation of population exposed ( 42-44) • Total urban population of Latin America (1987): 303.69 million. • Total population of high-priority cities (1987): 81.27 million (26.5% of total urban population of Latin America). • Total population of all cities with air pollution priorities: 119 million (39% of total urban popula- tion of Latin America). • Age distribution of urban population in different countries (44). SUMMARY In the last few years, air pollution has become a major issue in some countries of Latin America and the Caribbean because of urban development and growing industrialization. In addition to industrial processes often concentrated in the cities, vehicle emission and stationary-source fuel combustion are the primary sources of air pollution. Although air- quality standards have been established in some Latin American countries, these are frequently exceeded. Adverse health effects of air pollution have been mainly associated with the following pollutants: sulfur dioxide and particulate matter, photochemical oxidants, nitrogen dioxide and carbon monoxide, and lead. Short-term as well as long-term effects can be expected at levels exceeding WHO guidelines. The Latin American urban areas most affected by anthropogenic pollutant emissions are: the area of Sao Paulo (Brazil), the city of Santiago (Chile) and the metropolitan area of Mexico City. However, situ- ations similar to those prevailing in these cities could well occur in other cities of Latin America and the Caribbean. The population exposed to air-pollutant levels ex- ceeding WHO guidelines can be estimated to 81 million or 26.5% of the total urban population of Latin America and 19% of its total population. These estimates correspond to 30 million children (0-14), 47 million adults (15-59) and 4 million elderly people (60+). To date a very limited number of epidemiological studies have been carried out to determine the potential health effects of air pollutants in Latin America. To obtain a rough estimate, a scenario was hypothesized in which subjects living in cities would be exposed to a given level of air pollutant, using data from the international literature to extrapolate the expected number of events in different strata of the hypothetical population. The estimated health effects are considerable and warrant priority control intervention. This is true although epidemiological studies are needed to eval- uate the health impact of specific pollutant com- pounds as well as their interactions in Latin American populations exposed to high levels of pollution. RESUME Pollution de l'air dans les villes d' Amerique latine et des Cara"1bes: perspectives sanitaires L'acceleration de !'urbanisation et de !'industrialisa- tion ces dernieres annees a fait que la pollution de l'air est maintenant un probleme majeur dans cer- tains pays d'Amerique latine et des Cara"ibes. En plus des industries souvent concentrees dans les villes, les gaz d'echappement des vehicules a moteur et les produits de combustion constituent les princi- pales sources de pollution atmospherique. Bien que des normes de qualite de l'air aient ete etablies dans certains pays d' Amerique latine, elles sont souvent depassees. Les effets nefastes de la pollution de l'air sur la sante sont surtout associes aux polluants suivants: anhydride sulfureux et matieres particulaires, oxydants photochimiques, dioxyde d'azote et oxyde de carbone, et plomb. Lorsque les concentrations depassent les normes recommandees par l'OMS, Rapp. trimest. statist. sanit. mond., 43 (1990) - 165 - des effets a court terme et a long terme peuvent etre redoutes. Les zones urbaines d'Amerique latine les plus tou- chees par les emissions de polluants anthropogenes sont: la region de Sao Paulo (Brasil), la ville de Santiago (Chili) et la zone metropolitaine de Mexico. Toutefois, des situations analogues a celles obser- vees dans ces villes peuvent tres bien se rencontrer dans d'autres villes d'Amerique latine et des Cara'ibes. Le nombre de personnes exposees a des concentra- tions de polluants atmospheriques superieures aux normes de l'OMS peut etre estime a 81 millions, soit 26,5% du total de la population urbaine d'Amerique latine et 19% de sa population totale. Ces esti- mations correspondent a 30 millions d'enfants (0-14 ans), 47 millions d'adultes (15-59 ans) et 4 millions de personnes agees (60 ans au plus). Tres peu d'etudes epidemiologiques ant ete organi- sees jusqu'ici pour determiner les effets potentiels sur la sante de la pollution atmospherique en Ameri- que latine. Pour obtenir une estimation approxima- tive de ces effets, on a imagine un scenario selon lequel les habitants des villes seraient exposes a une concentration donnee de polluants atmospheriques, en utilisant les donnees de la litterature internationa- le pour extrapoler les manifestations attendues a differentes couches de la population hypothetique choisie. Les effets ainsi evalues sont considerables et justi- fient des mesures de lutte prioritaires. Cela etant, des etudes epidemiologiques s'imposent pour eva- luer les effets sur la sante de polluants particuliers ainsi que leurs interactions au sein des populations d' Amerique latine exposees a des concentrations elevees de polluants. REFERENCES- REFERENCES 1. AMERICAN THORACIC SOCIETY. Health effects of air pollution. New York, American Lung Association, 1978. 2. WORLD HEALTH ORGANIZATION. Air quality guidelines for Europe. Copenhagen, WHO Regional Office for Europe, 1987. (WHO Regional Publications, European Series No. 23). 3. MINISTRY OF HEALTH. Mortality and morbidity dur- ing the London fog of December 1952. London, Ministry of Health, 1954. (Reports on Public Health and Medical Subjects No. 95). 4. WORLD HEALTH ORGANIZATION. Su/fur oxides and suspended particulate matter. Geneva, WHO, 1979. (Environmental health criteria No. 8). ORGANISATION MONDIALE DE LA SANTE. Oxydes de soufre et particules en suspension. Geneve, OMS, 1979. (Criteres d'hygiene de l'environne- ment N° 8). 5. WARE, J. H. ET AL Assessment of health effects of atmospheric sulfur oxides and particulate mat- ter: evidence from observational studies. En- vironmental health perspectives, 41: 255-276 (1981). 6. 0STRO, B. A search for a threshold in the rel- ationship of air pollution to mortality: a reanaly- sis of data on London mortality. Environmental health perspectives, 58: 397-399 (1984). 7. SCHWARTZ, J. & MARCUS, J. Mortality and air pol- lution in London: a time series analysis. American journal of epidemiology, 131: 185-194 (1990). 8. LIPFERT, F. W. Mortality and air pollution: is there meaningful connection? Environmental science and technology, 19: 764-770 (1985). 9. WALLER, R. E. Air pollution and community health. Journal of the Royal College of Phys- icians, 5: 362-368 (1971). 10. BATES, D. V. & S1zro, R. Relationship between air pollutants levels and hospital admissions in Southern Ontario. Canadian journal of public health, 74: 117-122 (1983). Wld hlth statist. quart., 43 (1990) 11. BATES, D. V. & S1zro, R. The Ontario air pollution study: identification of the causative agent. En- vironmental health perspectives, 79: 69-72 (1989). 12. SAMET, J. M. ET AL The relationship between air pollution and emergency room visits in an in- dustrial community. Journal of the Air Pollution Control Association, 31: 236-240 ( 1981 ). 13. Osrno, B. D. Estimating the risk of smoking, air pollution and passive smoke in acute respiratory conditions. Risk analysis, 9: 189-196 (1989). 14. DOCKERY, D. W. ET AL Change in pulmonary func- tion in children associated with air pollution epi- sodes. Journal of the Air Pollution Association, 23: 937-942 (1982). 15. EVANS, J. S. ET AL Cross-sectional mortality studies and air pollution risk assessment. En- vironment international, 10: 55-83 (1984). 16. OzKAYNAK. H. ET AL Assessment of population health risk resulting from exposures to airborne particles. In: Lee, S. D. et al. (eds), Aerosols: re- search, risk assessment and control strategies. Proceedings of the second US-Dutch inter- national symposium, Williamsburg, 1985. Chelsea, Mi., Lewis Publishers Inc., 1985. 17. WARE, J. H. ET AL Effects of ambient sulfur oxides and suspended particles on respiratory health of preadolescent children. American review of re- spiratory disease, 133: 834-842 (1986). 18. DOCKERY, w. D. ET AL. Effects of inhalable particles on respiratory health of children. American re- view of respiratory disease, 139: 587-594 (1989). 19. L1PPMANN, M. & L10Y, P. J. Critical issues in air pollution epidemiology. Environmental health perspectives, 62: 243-258 (1985). 20. GOLD, D. R. ET AL Acute lower respiratory illness in childhood as a predictor of lung function and chronic respiratory symptoms. American review of respiratory disease, 140: 877-884 (1989). 21. SAMET, J. M. ET AL The relationship between re- spiratory illness in childhood and chronic air- - 166 - flow obstruction in adulthood. American review of respiratory disease, 127: 508-523 (1983). 22. BUMGARTNER, J. R. & SPEIZER, F. Chronic obstruc- tive pulmonary disease. In: Jamison, D. T. & Mosley, W. H. (eds), Evolving priorities in de- veloping countries. Washington, D.C., World Bank, Population, Health and Nutrition Division, in preparation. 23. LAMBERT, P. M. & REID, D. D. Smoking, air pollu- tion and bronchitis in Britain. Lancet, i: 853-857 (1970). 24. WORLD HEALTH 0RGANIZATON. Photochemical oxi- dants. Geneva, WHO, 1978. (Environmental health criteria No. 7). ORGANISATION MONDIALE DE LA SANTE. Oxydants photochimiques. Geneve, OMS, 1979. (Criteres d'hygiene de l'environnement N° 7). 25. L1PPMANN, M. Effects of ozone respiratory function and structure. Annual review of public health, 10: 49-67 (1989). 26. WORLD HEALTH ORGANIZATION. Ozone and other photochemical oxidants. In: Air quality guide- lines for Europe. Copenhagen, WHO Regional Office for Europe, 1987. (WHO Regional Pub- lications, European Series No. 23). 27. WHITTEMORE, A. & KORN, E. L. Asthma and air pol- lution in the Los Angeles area. American journal of public health, 70: 687-696 (1980). 28. SPEKTOR, D. M. ET AL Effects of ambient ozone on respiratory function in active normal children. American review of respiratory disease, 137: 313-320 (1988). 29. HACKNEY, J. D. ET AL Adaptation to short-term re- spiratory effects of ozone in men exposed re- peatedly. Journal of applied physiology, 43: 82- 85 (1977). 30. DETELS, R. ET AL The UCLA population studies of chronic obstructive disease-9. Lung function changes associated with chronic exposure to photochemical oxidants: a cohort study among never smokers. Chest, 92: 594-603 (1987). 31. KLEINMAN, M. T. ET AL. Health effect of acid aerosols formed by atmospheric mixtures. En- vironmental health perspectives, in press. 32. EULER, G. L. ET AL Chronic obstructive pulmonary disease symptom effects of long-term cumula- tive exposure to ambient levels of total oxidants and nitrogen dioxide in California Seventh-Day Adventists. Archives of environmental health, 43: 279-285 (1988). 33. KRYZANOWSKI, M. ET AL. Subchronic respiratory ef- fects of long-term exposure to ozone in Tucson. In: 82nd annual meeting of the Air and Waste Management Association. Anaheim, Ca., Air and Waste Management Association, 1989. 34. WORLD HEALTH ORGANIZATION. Oxides of nitrogen. Geneva, WHO, 1977. (Environmental health criteria No. 4). ORGANISATION MONDIALE DE LA SANTE. Oxydes d'azote. Geneve, OMS, 1978. (Criteres d'hygiene de l'environnement N° 4). 35. SAMET, J. M. ET AL. Health effects and sources of indoor air pollution (part 1 ). American review of respiratory disease, 136: 1486-1508 ( 1987). 36. SPEIZER, F. E. ET AL. Respiratory disease rates and pulmonary function in children associated with N02 exposure. American review of respiratory disease, 121: 3-10 (1980). 37. WORLD HEALTH ORGANIZATION. Carbon monoxide. In: Air quality guidelines for Europe. Copen- hagen, WHO Regional Office for Europe, 1987. (WHO Regional Publications, European Series No. 23). 38. WORLD HEALTH ORGANIZATION. Carbon monoxide. Geneva, WHO, 1979. (Environmental health criteria No. 13). ORGANISATION MONDIALE DE LA SANTE. Monoxyde de carbone. Geneve, OMS, 1980. (Criteres d'hygiene de l'environnement N° 13). 39. SNEE, R. D. Silver Valley lead study: further ana- lysis of relationship between blood lead and air lead. Journal of the Air Pollution Control Associ- ation, 32: 170-175 (1982). 40. SMITH, M. A. ET AL (EDS). Lead exposure and child development - An international assessment. Norwell, Ma., Kluwer Academic Publishers, 1989. 41. ENVIRONMENTAL PROTECTION UNIT. [Air pollution in Bogota, 1983-1986]. Bogota, Division of Environ- mental Health, 1987. (In Spanish). 42. DEPARTMENT OF ATMOSPHERIC POLLUTION CONTROL. [Report on concentrations in the atmosphere of total suspended particulates in seven cities in, Venezuela during 1986]. Maracay, General Dir- ectorate of Malariology and Environmental Health, 1987. (In Spanish). 43. NATIONAL ECOLOGY COMMISSION. [Atmospheric pol- lution]. Mexico City, National Ecology Commis- sion, 1988. (In Spanish). 44. ENVIRONMENTAL HEALTH SERVICE OF THE SANTIAGO METROPOLITAN REGION. [Atmospheric pollution in 1988]. Santiago, Environmental Health Service, 1988. (In Spanish). 45. COMPANY OF ENVIRONMENTAL SANITATION TECHNOLOGY. [Air quality in the metropolitan region of Sao Paulo and in Cubatao]. Sao Paulo, Company of Environmental Sanitation Technology, 1987. (In Portuguese). 46. COMPANY OF ENVIRONMENTAL SANITATION TECHNOLOGY. [Report on air quality in the metropolitan region of Sao Paulo and in Cubatao]. Sao Paulo, Com- pany of Environmental Sanitation Technology, 1988. (Report Series) (In Portuguese). 47. MINISTRY OF HEALTH. [Concentration of suspended particulates in the metropolitan area of San Jose, 1985-1986]. San Jose, Ministry of Health of Costa Rica, 1987. 48. FUNDA~AO ESTADUAL DE ENGENHARIA DO MEIO AMBI- ENTE. [Air quality in the metropolitan region of Rio de Janeiro, 1984-198'7J. Rio de Janeiro, FEEMA, 1989. (In Portuguese). 49. UNITED NATIONS. Patterns of urban and rural population growth. New York, United Nations, 1980. (Population studies No. 68). 50. UNITED NATIONS. The prospects of world urbaniza- tion. New York, United Nations, 1987. (Popula- tion studies No. 101 ). 51. UNITED NATIONS. Demographic yearbook. New York, United Nations, 1988. 52. VIZCAINO MURRAY, c. P. F. ET AL. [Environmental pollution and children's health]. Salud publica de Mexico, 15: 91-100 (1973). (In Spanish). 53. PACHECO, C. R. [Atmospheric pollution and its ef- fects on lung function]. Gaceta medica de Mexico, 114: 467-470 (1978). (In Spanish). 54. PEREZ-NERIA, J. ET AL. [Atmospheric pollution and spirometry in children in the federal district]. Rapp. trimest. statist. sanit. mand., 43 (1990) - 167 - Boletin medico de/ hospital infantil de Mexico, 35: 415-419 (1978). (In Spanish). 55. NAMIHIRA, D. ET AL. A study of spirometry in children from Mexico City. Pediatric pulmono- logy, 2: 337-343 (1986). 56. DIAZ MEJIA, G. S. ET AL. [Evaluation of respiratory symptoms in a cement-plant district in the Mexican Republic]. Salud publica de Mexico, 26: 438-446 (1984). (In Spanish). 57. CASTILLEROS, M. [Effects of environmental pollu- tion on the health of schoolchildren between zones of the metropolitan area of Mexico City. In: Punte, S. & Legorreta, J. (eds), Environment and quality of life]. Mexico City, Plaza lbaldes, 1988. (Collection Desarrollo Urbano). (In Span- ish). 58. GOLD, D. ET AL Ozone level as a predictor of lung function in a population of Mexican school- children [abstract]. American review of respiratory disease, 138: 428 (1989). 59. SECRETARIAT OF STATE FOR HEALTH. [Analysis of gen- eral attendance and visits related to respiratory Wld hlth statist. quart., 43 ( 1990) and cardiovascular problems at the emergency centre of the community of Cubatao-S.P. (1980- 1987)]. Cubatao, Secretariat of State for Health, 1988. (In Portuguese). 60. FISCHER, F. M. & HOFMEISTER, v. A. The effects of air pollution at different seasons on lung func- tion in normal schoolchildren. Sao Paulo, Uni- versity of Sao Paulo, Department of Environ- mental Health, 1987. 61. REGIONAL METROPOLITAN SERVICES (SERPLAC). [Epi- demiological study of the effects of atmospheric pollution]. Santiago, SERPLAC, 1989. (In Span- ish). 62. MELIA, R. ET AL. Children as a sentinel for respira- tory diseases in the United Kingdom. In: Leaver- ton, P. (ed.), Environmental epidemiology. New York, Praeger, 1982. 63. DE KONING, H. W. ET AL Biomass fuel combustion and health, Bulletin of the World Health Organi- zation, 63 (1): 11-26 (1985). DE KONING, H. W. ET AL Combustion de biomasse et sante. Bulletin de /'Organisation mondia/e de la Sante, 63 (2): 215-232 (1985). - 168 - LEAD IN PETROL: THE MISTAKE OF THE XXTH CENTURY Carl M. Shy• This article aims to provide some perspective of the public health consequence of using lead as an addi- tive in petrol. The emphasis will be on leaded- petrol use in the United States of America, although the conclusions are applicable to all countries where leaded motor-vehicle fuel is still in use. Through its use in paint and petrol, lead became one of the most widely dispersed toxins of the XXth century. The subtle effects of lead accumulation in human tissue, by multiple routes of exposure, have become known only in recent years (1). Although lead has long been recognized as a neurotoxin and haematological poison at high doses, little was known of its adverse effects at concentrations now considered to be "normal" in human blood. This lack of anticipation of the low-dose effects of lead resulted in a major mistake by the United States Public Health Service in 1925, when a Blue Ribbon Committee appointed by the Surgeon General con- cluded that there were "no good grounds for pro- hibiting use of ethyl [that is, tetraethyl lead] gas- oline" (2). The arguments marshalled by proponents and opponents of leaded petrol are familiar because they apply to many other economically useful but potentially harmful substances such as pesticides, acid air pollutants and synthetic organic chemicals. Proponents stated that leaded petrol was essential to the industrial progress of America; that it would lower the cost of motor-vehicle travel and allow for the manufacture of more powerful engines; that people would be exposed to low and insignificant doses; that illness among exposed workers was due to their own carelessness; that such exposures would not occur in the community; and finally that any innovation entails some risk. Opponents argued that lead is a slow, cumulative poison and should not be dispersed into the environment, that workers were not responsible for lead-induced illness and death, and that the burden of proof on the safety of motor-vehicle combustion of leaded petrol should be on industry. In a paper on tetraethyl lead published in 1925 in the Journal of the American Medical Association, Alice Hamilton, one of the foremost industrial physicians of this century, and her col- leagues from the Workers' Health Bureau (3) said: "The evidence so far available seems to show a real danger of chronic lead poisoning connected with garage work when ethyl gasoline is used and a possible danger to the public from lead dust in the streets of large cities ... The discharge of... [lead] particles ... which fall to the ground ... on crowded streets of cities, might constitute a far from negligible danger". More than 50 years passed before Alice Hamilton's fears about population exposure to lead in com- busted petrol were widely appreciated (4, 5). • Professor, Department of Epidemiology, School of Public Health, University of North Carolina, Chapel Hill, United States of America. Population exposure to lead from leaded petrol Combustion of leaded petrol since 1925 accounts for about 90% of total atmospheric lead (5). The history of global lead emissions can be constructed from chronological records of lead deposition in polar snow strata, marine and freshwater sediments and annual rings of trees. Murozumi et al. (6) provide a revealing time-profile in their plot of lead con- centrations in snow strata of northern Greenland. Lead in snow and ice increased linearly with time from the beginning of the industrial revolution, from 1750 until 1950, a 200-year interval during which lead concentrations increased threefold. Between 1950 and 1965, however, lead concentrations in snow and ice strata increased more than threefold again, owing to the delayed transport to northern Greenland of lead emissions from combusted leaded petrol, which was introduced in the United States in 1923 (Fig. 1). In the peak year 1972, 250 OOO metric tons of lead were utilized in the United States for leaded petrol, an average of 2.4 pounds of lead per person per year (5). Between 1972 and 1984, consumption of lead for petrol additives correlated by year with a decline in ambient lead con- centrations. The effect of these temporal reductions in the consumption of leaded petrol, and in air-lead levels, on blood-lead levels of the United States population was remarkable, as shown in Fig. 2 (7). Blood-lead levels of a probability sample of the United States population, as observed in the Second National Health and Nutrition Examination Survey (NHANES II) (BJ, declined by 37% between 1976- and 1980. During the same period, the estimated lead intake in the diet of teenage males showed no change (4). The correlation between changes in blood lead, adjusted for demographic variables, and leaded petrol use, was particularly strong for white children aged 6 months-5 years, with a correlation coefficient of 0.95. Thus, while inhalation of lead in air contributes only 1-2% of the total lead intake of humans, indirect exposure to atmospheric lead via ingestion and inhalation of lead in dust, soil, food and water can contribute up to 50% of the total lead intake. In support of the NHANES II results are temporal declines in blood-lead levels of black children aged 2-3 years examined in lead-screening programmes in New York City and Chicago between 1970 and 1980 (9, 10), among 5-6 year-old school- children studied in Newark, N. J., from 11 847 births between 1979 and 1981 in Boston (12). Based on NHANES II results, it is evident that blood- lead levels vary by age and race (Fig. 3); the highest levels occur in the first few years of life, and blacks have higher lead levels than whites at all ages. Superimposed on these demographic factors is an effect of degree of urbanization; high blood-lead levels are found among black children of central cities of the largest metropolitan areas, and the lowest values among white rural children (Table 1). Depending upon the boundary for defining excess blood lead, from 200 OOO to 2 380 OOO children aged Rapp. trimest. statist. sanit. mond., 43 (1990) - 169 - FIG.1 LEAD CONCENTRATION PROFILE IN SNOW STRATA OF NORTHERN GREENLAND PROFIL DE LA CONCENTRATION DE PLOMB DANS LES COUCHES DE NEIGE DU GROENLAND SEPTENTRIONAL 0.20 0.18 ., Cl ai 0.16 c • Q) " Cl 0.14 ,!!: • Cl -= .0 0.12 E 0 0:: I 3 0.10 0 c 0.08 "' Cl • • 'a, -= 0.06 " "' CD ...J 0.04 0.02 o 800 1750 1800 1850 1900 1950 YearO-AnO .... ...... Age of samples (years) -Age des echantillons (annees) Source: Reference (6) - Reference (6). FIG.2 PARALLEL DECREASE IN BLOOD-LEAD VALUES OBSERVED IN THE SECOND NATIONAL HEALTH AND NUTRITION EXAMINATION SURVEY AND AMOUNTS OF LEAD USED IN PETROL DURING 1976-1980, UNITED STATES OF AMERICA BAISSE ':uR~'tkEsLiN~~ 1~UL1 i~~~?r~t~~}E a't'lSN\~Jt 8t~EJ-~.f~V1~rs~~~~~~~~:::c~~~~7~~;9~t,TIONALE ETATS-UNIS D'AMERIQUE 110 Lead used in petrol Plomb utilise dans !'essence 16 / ~ -= 15 CD c c CD >, 0 E 14 .!!l E ·CD .0 E g 13 Q. I ~ I Average blood lead Taux de plombemie moyen -= 12 "' -a; ~ " 11 "' ~ .,,, 0 ~ 10 CD Cl !!! CD > < 9 ~ 1976 1977 1978 1979 1980 Year-Annee Source: Reference (7) - Reference (7). Wld hlth statist. quart., 43 (1990) - 170 - FIG.3 BLOOD-LEAD LEVELS BY RACE AND AGE IN THE UNITED STATES OF AMERICA ACCORDING TO THE SECOND NATIONAL HEALTH AND NUTRITION EXAMINATION SURVEY, 1976-1980 TAUX DE PLOMBEMIE SELON N~~lit<i'iJ1b~ri :~~T~T:is.:~~~,.O~fT~J~.1?~7~-~~RES LA DEUXIEME ENQUETE 22 21 20 19 ~ 18 :i.- 17 ~~ 16 O Cl i2: 15 ~ Cl> 1: c 14 Cl>C UCI> c >, 13 00 "E i·~ 12 . .., 11 ".a 8E 10 :C..Q cO.. 9 "' Cl> 8 ::;; 7 6 5 0 10 20 30 40 50 60 70 Age (years/ans) Source: Reference (8) - Reference (8). TABLE 1. RANKING OF BLOOD LEAD IN CHILDREN AGED 6 MONTHS-5 YEARS BY DEGREE OF URBANIZATION, 1976-1980 TABLEAU 1. PLOMBEMIE CHEZ LES ENFANTS DE 6 MOISA 5 ANS PAR DEGRE D'URBANISATION, 1976-1980 Geometric mean Rank- Rang blood lead Plombemie geometrique moyenne 1. (Highest - Le plus eleve) 20.8 2. 19.2 3. 16.5 4. 15.6 5. 14.4 6. (Lowest- Le plus bas) 12.7 Source: Reference (8) - Reference (8). Race Black - Noire Black - Noire Black - Noire White - Blanche White - Blanche White - Blanche 6 months-5 years living in metropolitan areas of the United States in 1984 were estimated to have blood- lead levels associated with adverse health effects (Table 2). There are several major environmental sources of these excess blood-lead levels. As shown in Table 3, lead in paint, petrol dusts and soil is responsible for the largest number of more highly exposed children. Unfortunately, there is considerable overlap of children exposed to these environmental sources, in that children living in central cities are often simulc taneously exposed to the highest air and soil/dust lead levels and to the highest density of older homes painted with lead-based paint. An important source of lead exposure to the fetus is lead in the blood of Degreeofurbanization -Degred'urbanisation Central cities of largest metropolitan areas - Villes centrales des zones metropolitaines les plus grandes Non-central cities of largest metropolitan areas; metropolitan areas <1 million population - Villes peripheriques des zones metropoli- taines les plus grandes; zones metropolitaines <1 million d'habitants Rural - Zones rurales Central cities of largest metropolitan areas - Villas centrales des zones metropolitaines les plus grandes Non-central cities of largest and all other metropolitan areas - Villes peripheriques des zones metropolitaines les plus grandes et toutes les autres zones metropolitaines Rural areas - Zones rurales pregnant women. Not only does lead transfer across the placenta to the fetus, but women appear to mobilize lead as well as calcium from their own body stores during pregnancy and transfer these substances to the fetus (13). The Agency for Toxic Substances and Disease Registry (14) estimated that there were 400 OOO pregnant women living in met- ropolitan areas in 1984 whose blood lead exceeded 10 mg/100 ml, a level associated with potentially adverse development of fetuses and newborn in- fants. That atmospheric lead, and therefore leaded petrol, contributes substantially to the total lead intake of human populations has been demonstrated in two studies. Firstly, the NHANES II survey showed a 37% Rapp. trimest. statist. sanit. mond., 43 ( 19901 - 171 - TABLE 2. ESTIMATED NUMBER OF CHILDREN AGED 6 MONTHS-5 YEARS WITH EXCESS BLOOD LEAD IN METROPOLITAN AREAS, UNITED STATES OF AMERICA, 1984 TABLEAU 2. NOMBRE ESTIMATIF D'ENFANTS DE 6 MOISA 5 ANS AVEC UNE PLOMBEMIE EXCESSIVE, ZONES METROPOLITAINES DES ETATS-UNIS D'AMERIQUE, 1984 Metropolitan area - Zone metropolitaine Large - Grande (;;.1 million) . . . . . . . . . . . . (Percentage children - Pourcentage d'enfants) Small - Petite (<1 million) ............ . (Percentage children - Pourcentage d'enfants) Total metropolitan areas - Total pour les zones metropolitaines. (Percentage children - Pourcentage d'enfants) . . . . . . . . Source: Reference (14) - Reference (14). >15 1493400 (21%) 483 OOO (13%) 2 380 600 (17%) Blood lead - Plombemie (l'g/100cc) >20 >25 459 500 128 200 (6%) (2%) 142 OOO 40 300 (4%) (1%) 715 500 200 700 (5%) (1%) TABLE 3. SOURCES OF CHILDHOOD LEAD EXPOSURE, UNITED STATES OF AMERICA, 1984 TABLEAU 3. SOURCES D'EXPOSITION AU PLOMB CHEZ LES ENFANTS, ETATS-UNIS D'AMERIQUE, 1984 Source of lead - Source de plomb 1. Maternal blood - Sang maternal . 2. Paint - Peinture . . . . . . . . . . 3. Petrol - Essence . . . . . . . . . . 4. Stationary sources - Sources fixes 5. Dusts and soils - Poussieres et sol . 6. Drinking-water - Eau potable . . . . 7. Food - Aliments .......... . Estimated number of children highly exposed Nombre estimatif d'enfants fortement exposes (millions) 0.4 a 5.9 b 5.6 c 0.013 d 5.9-11.7 e 3.8 1 <1.09 • Number of pregnant women in 1984 with blood lead > 1 O l'g/100 cc in standard metropolitan statistical areas - Nombre de femmes enceintes en 1984 avec une plombemie > 10 l'g/100 cc dans les zones statistiques standard des grandes agglomerations. • Number of children <7 living in oldest, highest-point residential units - Nombre d'enfants <7 ans vivant dans les unites residentielles les plus anciennes et les plus elevees. ' Number of children potentially exposed to lead from combusted petrol and residing in the 100 largest cities - Nombre d'enfants potentiellement exposes au plomb provenant de !'essence en combustion et residant dans les 100 plus grandes villes. • Number of children exposed to primary and secondary lead smelter, and who have blood-lead levels >20 l'g/100 cc - Nombre d'enfants exposes a une fonderie primaire ou secondaire de plomb, avec une plombemie > 20 l'Q/100 cc. • Number of children exposed to lead paint, leaded petrol combustion, or stationary sources of lead - Nombre d'enfants exposes a la peinture au plomb, a la combustion d'essence contenant du plomb ou a des sources de plomb fixes. I Number of children exposed to residential drinking-water containing the Environmental Protection Agency proposed level of 20 l'g/M - Nombre d'enfants exposes chez eux a de l'eau de boisson contenant le niveau de plomb propose par I' Environmental Protection Agency de 20 l'g/m. ' Number of children exposed to sources of food exceeding recommended lead concentrations - Nombre d'enfants exposes a des aliments de provenance depassant les concentrations en plomb recommandees. Source: Reference (14) - Reference (14). decline of blood lead in probability samples of the United States population between 1976 and 1980, as previously discussed. Secondly, an ingenious iso- tope lead experiment was conducted in Turin (Italy) during the 1970s and early 1980s (15). In this com- munity study, the geological source of lead for leaded petrol was changed to alter the isotopic ratio of lead-206 to lead-207. This altered isotopic ratio was then measured in air and in the blood of local inhabitants, and the change in blood-lead isotopic ratios was used to estimate the proportion of lead in the blood of an exposed population attributable to emissions from combustion of leaded petrol. The mean fraction of blood lead directly attributable to inhalation of emission of leaded petrol with the altered isotopic ratios was 21 ::!:: 10% in Turin, and 10-11% in the surrounding countryside. Incorpo- rating into the calculations the non-inhalation contri- bution of leaded petrol, via ingestion of lead in dust, soil, food and drinking-water, an estimated 60% of blood-lead levels in Turin residents was attributed to leaded petrol emissions. The point is that leaded petrol emissions contribute substantially to blood- lead levels; that this contribution comes not only from direct inhalation of air lead but from ingestion of dust, food and water impacted by fallout of atmospheric lead; and that the contribution is larger when other sources of excess exposure to lead are less important. Wld hlth statist. quart., 43 ( 1990) Health effects of ambient air lead Table 4 summarizes the most current estimates of blood-lead levels at which adverse effects may occur in children (5, 14). Prior to 1975, a blood lead of 40 mg/100 ml was considered to be the lowest observable effect level. However, in the past 15 years, a number of well-conducted longitudinal studies of children conclude that adverse develop- mental effects occur in early life in association with blood lead in the range of 15-25 mg/100 ml (16). The developmental effects of chronic low-level lead ex- posure in early life include: • reduced birthweight; • impaired mental development in the first two years of life; • IQ deficits in school-age children; and • disturbances in sensory pathways within the cen- tral nervous system persisting for five or more years. The cognitive and neurosensory effects of slightly increased blood-lead levels are particularly difficult to study for several reasons: (i) there are many different tests of cognition and neurosensory func- tion; (ii) these tests are not equally standardized; (iii) social class, home environment, birth order, inheritance and other factors are important deter- - 172 - TABLE 4. LOWEST OBSERVABLE EFFECT LEVEL (Pb-B) IN CHILDREN• TABLEAU 4. EFFET MINIMUM OBSERVABLE (Pb-sang) CHEZ LES ENFANTS • Lowest effect Pb-B Effet minimum plombemie (l'Q/dl) 10-15 (prenatal and postnatal - periodes prenatale et postnatale) 15-20 <25 30 Neurological effects Effets neurologiques Deficits in neurobehavioural development (Bayley and McCarthy Scales); electrophysiological changes - Retard du developpement neurocomportemental (Echelle de Bayley et McCarthy); modifications electrophysiologiques Lower IQ, slower reaction time (studied cross-sectionally) - Baisse du QI, prolongement du temps de reaction (elude transversale) Slowed nerve conduction velocity - Baisse de la vitesse de conduction nerveuse Heme synthesis effects Effets sur la synthese del'heme ALA-D inhibition - Inhibition de ALA-D EP elevation - Elevation de la PE Other effects Autres effets Reduced gestational age and weight at birth; reduced size up to age 7-8 - Age gestationnel et poids reduits a la naissance; taille reduite jusqu'a l'age de 7-8ans Impaired vitamin D metabolism; Py-5-N inhibition - lnsuffisance du metabolisme de la vitamine D; inhi- bition de Py-5-N 40 Reduced haemoglobin, elevated CP and ALA-U - Baisse du taux d'hemo- globine; elevation de la PC et de ALA-U 70 Peripheral neuropathies - Neuro- Anaemia - Anemie pathies peripheriques 80-100 Encephalopathy - Encephalopathie • Adapted from reference (5) - D'apres la reference (5). minants of cognitive abilities, and some of these factors are difficult to measure; finally, (iv) an oper- ational model for understanding the effect of lead on growth in cognitive and neurosensory function is lacking. Thus it is not surprising that there are inconsistencies in the existing body of studies re- lating relatively low blood lead to intellectual and sensory development. However there is an im- pressive convergence of evidence from experimental animal studies and from human epidemiological ob- servations on the effect of lead on learning and neurological function. Even those epidemiological studies that fail to find a statistically significant effect of slightly high blood-lead levels consistently detect effects in the postulated direction of adverse out- comes. Two epidemiological studies illustrate the nature and complexity of the findings. Bellinger et al. (17, 18) studied the relationship between umbilical cord blood lead and early cognitive development between 6 and 24 months of age. Cord-blood samples were obtained for 9 489 births in Boston at the Brigham and Women's Hospitals, and those in the lowest 10th, middle 10th and highest 10th percentiles of lead (n=1 207) were eligible for study; the final sample consisted of 249 of these infants. The en- rolled infants were healthy at birth and their parents were in the top two social strata according to Hollingshead's index; 87% of the families were white. Cord-blood lead levels in the highest 10th percentile groups ranged only from 10-25 mg/ 100 ml, a level considered to be within normal variation. Infants were tested at 6, 12, 18 and 24 months, using the Bayley Scales of Infant De- velopment. Fig. 4 shows results for the three cord- blood lead groups in terms of the mental develop- ment index score at six-month intervals. The mental development index is a composite age-corrected scale that assesses sensory-perceptual acuities, Colic, otherGI effects; kidney effects - Colique, autres effets gastro-intestinaux, affections renales memory, learning and problem-solving ability, verbal communication, and early ability to form gen- eralizations, as well as other cognitive functions; 15-20% of infants did not complete all four test runs, and losses were greater among infants born to nonwhite or unmarried mothers and to mothers of lower age, education level and socioeconomic status. Mental Development Index scores were re- gressed against cord-blood lead and 12 factors associated with infant development, including mother's age, race, IQ, education, smoking and alcohol habits, social class, care-giving environment and infant's sex, birthweight and birth order, and gestational age. At all ages, infants in the high prenatal exposure group scored lower than infants in the other two groups, and the difference between the high and each of the other two exposure groups was statistically significant. Test scores were not associated with concurrent postnatal blood-lead levels or with cumulative postnatal lead exposure up to the time of assessment. Blood-lead differences between the three groups were much larger for cord-blood than for postnatal blood-lead levels. The striking feature of this study is that the mean blood-lead level of the study sample as a whole was about one-half the mean level of United States pre- school-age children, based on measurements made in the NHANES II survey. Likewise, study children came from families in the higher socioeconomic strata. These factors reduced the likelihood of find- ing an effect of prenatal lead exposure during the first two years of life, especially if a favourable home and family environment can offset the potentially small effects of prenatal lead on cognitive develop- ment. In their later publication (18), the authors pre- sented evidence that the performance of children in the lower socioeconomic stratum (of this relatively high socioeconomic sample) was adversely affected at lower levels of prenatal exposure, i.e. cord-blood Rapp. trimest. statist. sanit. mond., 43 (1990) - 173 - FIG.4 MEAN MENTAL DEVELOPMENT INDEX SCORES AT FOUR AGES IN INFANTS ACCORDING TO THE LEAD LEVEL IN UMBILICAL-CORD BLOOD • INDICES MOVENS DU DEVELOPPEMENT MENTAL A Ql,IATRE STADES CHEZ LES NOURRISSONS SELON LE TAUX DE PLOMBEMIE CORDALE • 120 Cord-blood lead group Groupe de plombemie cordale .......... Low - Faible Medium - Moyenne High - Elevee ~ a, E E 116 a, E a, Q. Q. ~ > .., "t) ::, "t) a, .!./ "t) .5 I 112 ~ 8 ., 1il "t) .5 E Q) E Q. ~ > a, Cl 108 <ii E a, ::; 104 6 12 18 24 Age at testing (months) - Age au moment de l'examen (mois) a Scores are least-squares means obtained by regressing Mental Development Index scores on the cord-blood lead group and 12 variables considered to be potential confounders. Error bars represent 1 SO. For clarity, bars extend only in one direction - Les indices moyens sont obtenus selon la methode des moindres carres par regression par rapport au groupe de plombemie cordale et par rapport a 12 variables considerees comme des facteurs confondants potentiels. Les barres d'erreur representent 1 ET. Pour plus de clarte, les barres sont toutes dirigees dans le meme sens. Source: Reference (17) - Reference (17). lead levels of 6-7 mg/100 ml, than children in the higher socioeconomic stratum. However these re- sults are based on small numbers of children in each lead and social class stratum. The second study, conducted in Port Pirie, South Australia, investigated a cohort of 539 children born between 1979 and 1982 to women living in a com- munity near a lead smelter (19-20). This study pro- vided a considerably larger sample in a community with presumably higher environmental lead ex- posures and probably a distribution of families of lower socioeconomic strata than in the Boston study. The mean blood-lead concentration in the Port Pirie children at age 15 months was 21 mg/ 100 cc, and one-third of the children had levels above 25 mg/100 cc on one or more occasions. Maternal intelligence, home environment and the children's mental development were assessed, the latter with use of the McCarthy Scale of Children's Wld hlth statist. quart., 43 ( 1990) Abilities, from which a general cogn1t1ve index was derived to reflect reasoning, concept formation and memory. Blood-lead concentrations at each age and the in- tegrated postnatal average blood-lead level were inversely related with the general cognitive index score at age 4 years, as shown in Fig. 5. General cognitive index score is adjusted for 16 factors iden- tified a priori as potential determinants of mental development. The regression line of the figure shows a drop of 7.2 units in the general cognitive index score (where 100 is the expected average score value for a population) associated with an increase in blood lead from 10 to 30 mg/100 ml. There was no evidence that cognitive function at age 4 was more influenced by recent than by earlier blood-lead levels, or that children whose blood-lead levels increased with age were more affected than those with constant levels. - 174 - FIG.5 LINEAR RELATION, AS ESTIMATED FROM MULTIPLE REGRESSION ANALYSIS, BETWEEN THE INTEGRATED POSTNATAL AVERAGE BLOOD-LEAD CONCENTRATION (SHOWN ON A LOGARITHMIC SCALE) AND THE GENERAL COGNITIVE INDEX (GCI) CORRl~~l 1 ~~5~N~~~~~i¥!t~~ftHAERLr:~~~l~~~:i~~TJr~i ~~J~1~~'6~~14i~~~iN°Jr~r:RcA~~ULEE 1201 GCI Score - lndice de maturation cerebrale 115 Change Line of best fit (regression coetticient = 15.0) ./ Droite d'ajustement / ,~···~"""'~"" -"'" = 7.2 unit Modification = 7,2 unites [ 110 Mean score / lndice moyen = 107, 1 105 100 95 Source: Reference ( 19) - Reference ( 1!Jj. These two studies suggest that there is no clear threshold for the effects of lead on cognitive de- velopment in early life. While the failure to find a threshold is not surprising, the existence or lack of a threshold cannot be readily demonstrated in obser- vational studies because enormous sample sizes are needed to compare effects at low levels of exposure. The practical conclusion of these and other studies is fairly clear with respect to lead-there should be as little human exposure as possible, and all evidence points to a greater risk of a variety of adverse effects, particularly effects on cognition and haematological function, at what were formerly considered normal blood-lead levels. Lessons learned retrospectively Today, 65 years after the United States Surgeon General's Blue Ribbon Committee concluded that there was no good evidence for prohibiting use of leaded petrol, it can reasonably be claimed that lead is the environmental health disaster of the XXth century. Even discounting the thousands of children who were acutely and severely poisoned by inges- tion of lead-based paint residues, there are still millions of children whose cognitive and neurosen- sory development was probably compromised by excess lead intake attributable to leaded petrol com- bustion. To this can be added the known lead- induced effects on fetal development, haematologi- cal function and blood pressure, and the yet-to-be- discovered effects of mobilization of high body stores of lead among the elderly during periods of bone resorption (21). 0.25 0.5 1.0 Change in PbB - Modifi· I cation de la plombemie Blood-lead concentration (µmol/1) - Plombemie 1.5 2.5 How could this environmental health calamity have been avoided? With hindsight, were there major clues missed by the 1925 Blue Ribbon Committee that are applicable to future technological develop- ment? Several can be listed. • Lead is a cumulative toxin. Lead combusted in petrol adds to the body burden imposed by other environmental sources. Cumulative toxins from mul- tiple sources magnify the risk of excessive popula- tion exposures. Other heavy metals and many chlori- nated hydrocarbons also pose these risks. • Leaded petrol was initially associated with energy consumption, which increased dramatically over the next 50 years. Even slight increases in hazardous byproducts of energy consumption have the po- tential for a widespread adverse public health im- pact. • Combustion of leaded petrol in millions of auto- mobiles resulted in extraordinarily dispersive con- tamination of residential environments. Therefore, any other additive to petrol should raise similar concerns, and any hazardous pollutant emitted by motor-vehicle emissions will similarly expose a large segment of the population to potential adverse health effects. Ozone and carbon monoxide are known concerns. • Although the adverse effects of low levels of blood lead were unknown in 1925, there was ample reason for concern. Lead was a known poison to enzyme systems, and as such had the potential to interfere with essential biochemical pathways at low doses, with virtually no threshold. It would have been reasonable to extrapolate from high-dose acute human effects, for which there was abundant evi- Rapp. trimest. statist. sanit. mond., 43 ( 1990) - 175 - dence, to more subtle low-dose effects. In the face of profound and clear-cut adverse effects and the ab- sence of studies at low doses, linear extrapolation of effects to low doses is a judicious public health risk assessment. Instead, the assumption of zero effect at low doses was made by default. This assumption continues to be made for population exposure to combinations of carcinogenic substances discharged into public drinking-water supplies. The same holds for environmental chemicals which are reproductive toxins and neurotoxins. • Government, industry and scientific advisers to these institutions are obsessed with accumulating virtually incontrovertible proof of causality before making regulations to protect public health. This was true for environmental lead. Prior to 1955, there were no regulations in the United States on the amount of lead allowed in paint. Lead-free petrol was introduced only in the 1970s, and this was done in order to protect catalytic converters rather than human beings. The burden of proof of causality is excessive, particularly in relation to an agent such as lead in combusted petrol which affects human popu- lations by a combination of direct and indirect ex- posures, through inhalation and ingestion. In the face of reasonable evidence of human toxicity, the burden of proof should be shifted to the polluter, to demonstrate absence of effects at low doses. • If a new technology having the potential for widespread population exposure to environmental contaminants emerges, one cannot rely on high- dose toxicological studies to assess human risk, particularly for a complex technology. There is ample time to evaluate human exposures, some- times even by controlled experiments, certainly by obtaining good-quality baseline exposure and health data in pilot communities, and by performing ap- propriate epidemiological follow-up of the com- munities. Protection of public health must be an absolute goal, after which the economic benefit of technologies can be considered. Our XXth century experience with environmental lead should drive this lesson firmly home. In 1786, Benjamin Franklin wrote: "The Opinion of the mischievious Effect of Lead is at least Sixty Years old; and you will observe with Concern how long a useful Truth may be known and exist, before it is generally received and practiced on" (22). Following Franklin's advice, let us not delay another 60 years, as we did before removing lead from petrol, before acting on the evidence of the "mischie- vous effects" of other pervasive environmental hazards. SUMMARY The health aspects of the use of lead in petrol were evaluated in the 1920s in the United States of America and, in spite of warnings from certain lead- toxicity experts, lead addition to petrol became standard international practice. Available data now show that lead in petrol at the scale of use in the 1970s produced significant en- vironmental lead contamination and increased aver- age blood-lead levels in the general population. National sample surveys of blood-lead levels in the United States carried out annually from 1976 show a decreasing trend closely correlated with the use of lead in petrol. Recent longitudinal epidemiological studies have concluded that the exposure levels associated with lead in petrol can cause a reduced average mental ability in children. These studies accounted for the potential confounding from socioeconomic and other factors. The practical conclusion from the studies reviewed is that there should be as little human lead exposure as possible, because there may be no threshold for the effects occurring and many thousand children have already been affected in the United States and other countries. The environmental health calamity caused by lead in petrol could have been avoided if the initial warn- ings had been heeded and better preliminary re- search of the health issues had been carried out. Nevertheless, incontrovertible proof of causality should not be required before regulations are made to protect public health. RESUME Essence contenant du plomb- l'erreur du xxe siecle Les incidences pour la sante de la presence de plomb dans !'essence ant ete evaluees dans les annees 20 aux Etats-Unis d'Amerique et, malgre les mises en garde de certains experts en matiere de toxicite du plomb, la pratique consistant a mettre du plomb dans !'essence s'est repandue dans le monde. On sait maintenant que le plomb contenu dans !'essence, a l'echelle au ii etait utilise dans les annees 70, a serieusement contamine l'environne- ment et entraine une augmentation du taux moyen de plombemie dans !'ensemble de la population. Les enquetes nationales par sondage sur les taux de Wld hlth statist. quart., 43 (1990) plombemie aux Etats-Unis faites chaque annee depuis 1976 font apparaitre une tendance decrois- sante, etroitement liee a la quantite de plomb utilisee dans !'essence. Des etudes epidemiologiques longitudinales recen- tes ant montre que !'exposition au plomb contenu dans !'essence, selon le niveau, pouvait se traduire par une baisse de la capacite intellectuelle moyenne des enfants. Ces etudes tiennent compte des dis- positions possibles dues aux facteurs socio- economiques et autres. La conclusion pratique qui se degage de ces etudes est que l'homme doit eviter - 176 - au maximum toute exposition au plomb car ii se peut qu'il n'y ait pas d'exposition sans effets et des milliers d'enfants ant deja ete affectes aux Etats-Unis et dans d'autre pays. Le fleau que constitue la presence de plomb dans !'essence pour !'hygiene de l'environnement aurait pu etre evite si l'on avait tenu compte des premiers avertissements donnes et que les incidences sani- taires avaient d'abord ete etudiees plus en pro- fondeur. II ne devrait cependant pas etre necessaire que la preuve irrefutable soit faite qu'il existe une relation de cause a effet pour proteger la sante publique au moyen d'une reglementation appro- priee. REFERENCES - REFERENCES 1. WORLD HEALTH ORGANIZATION. Air quality guidelines for Europe. Copenhagen, WHO Regional Office for Europe, 1987. (WHO Regional Publications, European Series No. 23). 2. ROSNER, D. & MARKOWITZ, D. A "gift of God"? -The public health controversy of leaded gaso- line during the 1920s. American journal of public health, 75: 344-352 (1985). 3. HAMILTON, A. ET AL. Tetra-ethyl lead. Journal of the American Medical Association, 84: 1481- 1486 (1925). 4. lANSDOWN, R. & YULE, w. (EDS). Lead toxi- city-History and environmental impact. Baltimore, The John Hopkins University Press, 1986. 5. us ENVIRONMENTAL PROTECTION AGENCY. Air quality criteria for lead. Research Triangle Park, N.C., Office of Health and Environmental Assessment, 1986. (EPA Report No. EPA-600/8-83028aF-dF). 6. MuROZUMI, M. ET AL. Chemical concentrations of pollutant lead aerosols, terrestrial dusts and sea salts in Greenland and Antarctic snow strata. Geochimica cosmochimica acta, 33: 1247-1294 (1969). 7. ANNEST, J. L. ET AL. Chronological trend in blood- lead levels between 1976 and 1980. New England journal of medicine, 208: 1373-1377 (1983). 8. MAHAFFEY, K. R. ET AL. National estimates of blood-lead levels, United States, 1976-1980. New England journal of medicine, 307: 573-579 (1982). 9. BILLICK, I. H. ET AL. Analysis of pediatric blood- lead levels in New York City for 1970-1976. En- vironmental health perspectives, 31: 183-190 (1979). 10. BILLICK, I. H. ET AL. Relation of pediatric blood-lead levels to lead in gasoline. Environmental health perspectives, 34: 213-217 (1980). 11. GAUSE, D. ET AL. Reduction in lead levels among children in Newark. Journal of the Medical Society of New Jersey, 74: 958-960 (1977). 12. RABINOWITZ, M. B. & NEEDLEMAN, H. L. Temporal trends in the lead concentrations of umbilical- cord blood. Science, 216: 1429-1432 (1982). 13. ALEXANDER, F. W. & DELVES, H. T. Blood-lead levels during pregnancy. International archives of occu- pational and environmental health, 48: 35-39 (1981). 14. AGENCY FOR TOXIC SUBSTANCES AND DISEASE REGISTRY. The nature and extent of lead poisoning in children in the United States-A report to Con- gress. Bethesda, Maryland, Department of Health and Human Services, 1988. 15. FACCHETTI, s. & GEISS, F. Isotopic lead experi- ment-Status report. Luxembourg, Commission of the European Communities, 1982. (Publication No. EUR 8352 EN). 16. us ENVIRONMENTAL PROTECTION AGENCY. Supple- ment to the 1986 EPA air quality criteria for lead. (Vol. I, addendum - pp. A1-A67). Washington, D.C., Office of Health and Environmental Assess- ment, 1989. (EPN600-8-89/049A). 17. BELLINGER, D. ET AL. Longitudinal analyses of pre- natal and postnatal lead exposure and early cognitive development. New England journal of medicine, 316: 1037-1043 (1987). 18. BELLINGER, D. ET AL. Low-level lead exposure, social class, and infant development. Neuro- toxicology and teratology, 10: 497-503 (1989). 19. MCMICHAEL, A. J. ET AL. Port Pirie cohort study-Environmental exposure to lead and children's abilities at the age of four years. New England journal of medicine, 319: 468-475 (1988). 20. WIGG, N. R. ET AL. Port Pirie cohort study -Childhood blood-lead and neuropsychological development at age two years. Journal of epi- demiology and community health, 42: 213-219 (1988). 21. S1LBERGELD, E. K. ET AL Lead and osteoporosis -Mobilization of lead from bone in post- menopausal women. Environmental research, 47: 79-94 (1988). 22. McCoRD, C. P. Lead and lead poisoning in early America-Benjamin Franklin and lead poisoning. Industrial medicine and surgery, 22: 393-399 (1953). Rapp. trimest. statist. sanit. mond., 43 ( 1990) - 177 - HEAL TH RISKS ASSOCIATED WITH POLLUTION OF COASTAL BATHING WATERS Louis J. Saliba• & Richard Helmer" The main channel of human exposure to pollutants discharged into coastal waters is through contact with polluted water or beach sand while swimming or bathing, including ingestion or inhalation of path- ogens. Potentially, all the diseases which are spread by the faecal-oral route and whose etiological agents are shed in the faeces of all individuals or carriers could be contracted by swimming in sewage- polluted waters. Cabelli (1) has reported such dis- eases to include: (i) bacterial diseases such as salmonellosis (including typhoid and paratyphoid fevers), shigellosis (bacillary dysentery), cholera and gastroenteritis caused by enteropathogenic E. Co/i, Yersinia enterocolitica, etc.; (ii) viral diseases such as infectious hepatitis, illnesses caused by enteroviruses (polioviruses, coxsackieviruses A and 8, echoviruses, reoviruses and adenoviruses), and "non-specific" gastroenteritis caused by the human rotavirus and parvo-like viruses; and (iii) diseases caused by a variety of protozoan and metazoan parasites, such as amoebic dysentery, giardiasis, ascariasis, etc. In so far as the actual ingestion of water during swimming or bathing is necessarily limited, with the exception of pathogens with a relatively low infec- tive dose, the diseases above can be contracted more easily through the consumption of raw or partially-cooked fish or shellfish. Other pathogens reported (2) as causing human infections through this route include Vibrio parahaemolyticus and Clos- tridium botulinum (type E), whose native habitat is the sea. There is, in fact, extensive evidence of the spread of diseases to man following the consump- tion of polluted shellfish. Although a wide range of diseases have been described (3), the main ones are typhoid and paratyphoid fever, salmonellosis, Vibrio parahaemolyticus infections, viral hepatitis type A (infectious hepatitis), paralytic shellfish poisoning and cholera. Although the shellfish industry is large and continuously expanding, the consumption of raw or only lightly cooked or preserved shellfish taken from polluted areas (mainly coastal) or follow- ing incorrect storage or treatment has resulted in disease occurrence on a scale possibly greater than would be expected from the total amounts of shell- fish consumed. Apart from diseases affecting the gastrointestinal tract, a number of diseases and disorders affecting the eye, ear, skin and upper respiratory tract have been associated with bathing. This particular cat- egory of infective conditions may be caused by microorganisms such as Staphylococcus aureus, Pseudomonas aeruginosa, Clostridium welchii and Candida albicans, and may cause infection as a result of being forced into breaks or tears in the skin, or into ruptures in delicate membranes in the ear or • Senior Scientist, WHO/EURO Project Office, Mediterranean Action Plan, Athens, Greece. O Scientist, Division of Environmental Health, World Health Organi· zation, Geneva. Wld hlth statist. quart., 43 ( 1990) nose resulting from the trauma associated with div- ing into water (3). These microorganisms have been described by Mood & Moor.e (4) as often present in man, but giving rise to disease only when, for one reason or another, the resistance of the individual who harbours them is lowered. Although all four species mentioned may be found also in polluted water, the suggestion that an individual suffering from infection has acquired it from polluted water is considered as a matter to be treated with reserve (4), as the bather is very likely to have been carrying the organism beforehand, and the disease may be largely determined by the individual's susceptibility rather than by exposure to the organism in the environment. Infective diseases associated with bathing (as well as with seafood consumption) are largely caused through pollution of coastal waters by municipal sewage. Another type of health problem associated with pollution of the marine environment arises from the discharge of industrial and related effluents into the immediate coastal zone, coupled with chem- ical pollutants reaching the sea through other routes, including the atmosphere. Among the major pol- lutants in this category are mercury and cadmium compounds, and a number of organohalogen com- pounds, particularly chlorinated hydrocarbons. All of these are highly persistent in the marine environ- ment, and accumulate along the food chain, reaching their highest concentrations in filter- feeders, such as bivalves, and in predatory fish. The regular consumption by man of these types of seafood may give rise to a number of diseases or disorders. While such chemicals constitute an overall rather than a coastal marine pollution problem, the coastal zone is generally the most affected, par- ticularly in the case of bivalves and inshore-dwelling fish species exposed to major local point-sources of chemical pollution. Health problems arising from coastal pollution but not associated with exposure through bathing are, however, beyond the scope of this article. Coastal recreational water quality The increasing use of the sea as a recreational amenity, coupled with the spread of international tourism, has led to major concern among health authorities regarding actual and potential hazards to bathers and other users of the sea for sport and recreation in those areas affected by sewage pollu- tion. The main response to the problem has been twofold: reduction of pollution at source wherever feasible through the establishment of sewage treat- ment plants and/or construction of submarine outfall structures taking the effluent out to sea, and the development and enforcement of recreational water quality criteria and standards aimed at ensuring, as far as possible, that recreational marine areas pre- sent no health hazard caused by sewage pollution. Both approaches are necessarily intimately inter- - 178 - linked, and the design and location of treatment plants and outfall structures depend on specific water-quality requirements in affected amenity areas. In their turn, water-quality criteria and stand- ards must be based on firm epidemiological evi- dence if the aim is to protect the health of exposed population groups. In this context, causality and quantitative description of the correlation between levels of contamination and morbidity in such ex- posed groups is a subject that has occupied gener- ations of epidemiologists, microbiologists, public health specialists and physicians. Ideally, any criteria and standards applied should be based on well-conducted epidemiological investi- gations which establish a clear dose-response re- lationship and thus allow the definition by the com- petent authorities of a level of acceptable risk in terms of water quality. In this regard, it has been stated c that the large number of epidemiological studies which have attempted to quantify the pos- sible links between bathing in sewage-polluted waters and health risks as evidenced by observed effects have produced dissimilar results, and a con- sistent relationship has so far not been proved, possibly because, apart from the different way each study has been conducted, a number of imper- fections and confounding factors influencing the in- terpretation of results obtained have been practically impossible to surmount. This situation has led to a host of different coastal recreational water-quality criteria and standards being applied round the world, and to considerable controversy surrounding their implementation. The Mediterranean situation The Mediterranean basin can be taken as a typical example of the problems of coastal pollution in so far as they affect recreational areas, and of the difficulties both in achieving adequate safety meas- ures and in obtaining cohesion and harmonization between 18 countries which, though varying to a degree in their state of socioeconomic development, do have practically the same ecological conditions, overall if not in detail. Apart from the 130 million inhabitants estimated to live permanently along the Mediterranean coastline, over 100 million tourists visit the area annually. Tourism, both domestic and foreign, is heavily con- centrated during the summer months, and the sea constitutes the major recreational amenity for both local and visiting populations. As a result, during this period most beaches, particularly those located in the vicinity of cities and tourist resorts, are heavily overcrowded. The prevailing warm climatic con- ditions, including a fairly high water temperature, result in a comparatively long time spent on the beach, especially in the sea itself. The bather is therefore subjected to a longer exposure period both to seawater and (where appropriate) to beach sand, than is the case in more temperate regions. Apart from pollution of the water itself, there is also the heterogeneous nature of beach populations to con- 'Jones, F. & Kay, D. Recreational water quality: the relationship between epidemiological studies and recreational activities in water. Report of a biological standards seminar, Middlesex Poly- technic, 9 February 1990. d World Health Organization. Microbiological quality control in coastal recreational and shellfish areas in the Mediterranean. (WHO/EURO document ICP/CEH 083/6, 1989). sider. Brisou (5), in describing the health situation around the Mediterranean nearly two decades ago, has stated that "the milling throngs of holiday makers that stretch from the southern tip of Spain to Turkey, taking in the Italian and Yugoslav shore of the Adriatic, Greece and the Greek Islands, for two or three months each year, and which add up to over a hundred million human beings in transit, constitute an aggregation particularly receptive to the spread of infections, especially considering that 90% of all the dirt generated by this mass of human beings is directly discharged into the receptacle in which they take their dips". During the same period, the Mediterranean was described by Ritchie Calder (6) as a dying sea, a statement associated with one particular ecological school of thought. Irrespective of the exactness of this description as applicable to the Mediterranean sea as a whole, the situation in coastal areas (where the bulk of observations had been carried out) was certainly serious. From the viewpoint of human health hazards related to bathing (as well as to shellfish consumption), this situation was largely a result of the fact that over 90% of municipal sewage was discharged in the raw untreated state from outfall structures opening onto the immediate shoreline, in many cases in the immediate vicinity of heavily-frequented bathing bj:laches. Although the situation has improved considerably since that time, problems still exist in a number of areas and tour operators make detailed enquiries as to the state of beaches in localities affecting their potential clients. Mediterranean quality criteria and standards In line with global practice, recreational water quality standards in the Mediterranean are based on accept- able concentrations of bacterial indicator organisms (mainly faecal coliforms, supplemented to a lesser extent by faecal streptococci) and, in some in- stances, pathogens such as Salmonella and entero- viruses. The four countries (France, Greece, Italy and Spain) which are member states of the European Economic Community are bound by the relevant EEC Directive (7), which also includes an assortment of physicochemical parameters, apart from the micro- biological. Standards exist in practically all countries in the region, but differ to a large extent both as to the particular microorganism(s) monitored, and the "acceptable" levels of each.d Apart from this, even were the same standards to be prevalent, comparison would still be difficult, owing to differences in sampling techniques, analytical methodologies and interpretation of results. Considerable progress towards the achievement of at least partial standardization has been effected by the World Health Organization (WHO) and the United Nations Environment Programme (UNEP) within the framework of the Long-term Programme of Pollution Monitoring and Research in the Mediterranean Sea (MEO POL Phase II), operational since 1981, but there are a number of practical and technical limi- tations. One such problem is that of methodology. In an effort to achieve harmonization, a series of rec- ommended methods for the determination of con- centrations of the major bacterial parameters were prepared by WHO, issued by UNEP within the frame- work of its series of reference methods for marine pollution studies, and distributed to Mediterranean laboratories. At the political level, the division within the region of preference between two major tech- Rapp. trimest. statist. sanit. mond., 43 (1990) - 179 - niques (membrane filtration and multiple test tube) led to the eventual recommendation of either, and to the issue of both, even though the two are not directly comparable. Apart from the technique itself, culture media rec- ommended can produce unexpected results. Methods used for seawater analysis have been developed from those used for potable water which, in the unpolluted state, is bacteriologically pure. The natural marine bacterial flora often interfere with bacterial indicator counts, producing false positive or false negative readings, and as natural flora vary with the area, it has been found difficult to produce standard methodology guaranteed to produce the same results wherever applied. This problem is not confined to the Mediterranean. Comparison of the state of microbiological pollution in coastal rec- reational waters in different areas can therefore be quite meaningless if restricted solely to counts ob- tained without taking sampling and analytical tech- niques, not to mention quality control, into account. A WHO working group on guides and criteria for recreational quality of beaches and coastal waters met in Bilthoven in 1974,e with the aims of making a critical review of available information on beach pollution and existing preventive measures in var- ious part of Europe; evaluating the health risks to swimmers and bathers resulting from pollution of beaches and coastal waters; and considering criteria and guidelines for seawater and beach pollution of microbiological, chemical and physical origin. In dis- cussing microbiological criteria and guidelines, the group agreed that, because of the high variability of bacterial counts in the marine environment, recom- mended upper limits for indicator organisms should be expressed in broad terms of orders of magnitude, rather than as rigidly stated specific numbers. It was also agreed, however, that highly satisfactory bathing areas should show E. Coli counts of consis- tently < 100 per 100 ml and, to be considered ac- ceptable, bathing waters should not give counts • World Health Organization. Guides and criteria for recreational quality of beaches and coastal waters. Report on a working group, Bilthoven, 28 October-1 November 1974. Copenhagen, WHO Regi- onal Office for Europe, 1975. (Document EURO 3125(1)). ·' United Nations Environment Programme/World Health Organiza- tion. Assessment of the state of microbial pollution of the Mediterranean sea and proposed measures. (Document UNEP/WG. 91/6, 1983). g United Nations Environment Programme. Report of the fourth ordinary meeting of the contracting parties to the convention for the protection of the Mediterranean sea against pollution and its related protocols, Genoa, 9-13 September 1985. (Document UNEP/ IG.56/5, 1985). consistently > 1 OOO E. Coli per 100 ml at the time, backed by direct epidemiological data; but, neverthe- less, the group believed that on general public health grounds, the level of contamination of bathing waters with faecal microorganisms should be kept at as low a level as was reasonably feasible. The principal investigators from 27 Mediterranean laboratories participating in a WHO-organized pilot project on coastal water quality control within the framework of the UNEP-sponsored MEO POL pro- ~ gramme (carried out between 1976 and 1981), took the 1974 Bilthoven conclusions as a baseline for extrapolation to Mediterranean conditions, and made recommendations on the basis of which WHO and UNEP proposed, in 1983, interim quality criteria for bathing waters in the region/ utilizing both faecal coliforms and faecal streptococci as mandatory indi- cators, and specifying the membrane filtration cul- ture technique (Table 1). These criteria were de- signed to alleviate the situation pending the perfor- mance and results of microbiological/epidemio- logical studies correlating microbiological water quality with observed health effects. The part of the proposal concerning faecal coliforms, with the addi- tion of the multiple test tube technique as an altern- ative method of analysis, was adopted by Medi- terranean states on a joint basis as an interim measure in 1985.9 This interim measure did not affect the several countries in the region which already had stricter limits in force. Quality criteria and standards in other countries As is the case with their Mediterranean counterparts, the rest of EEC Member States have their standards based on the already-mentioned 1976 Directive (7). These standards are reproduced in Table 2. It is interesting to note that they apply equally to all the Commmunity's bathing areas (freshwater and ma- rine) from the North Sea to the Mediterranean, irre- spective of the difference in prevailing conditions, and it has been suggested at seminars reviewing the status of implementation of the Directive that the standards themselves should be reviewed with a view to taking into account relatively stricter meas- ures for the Mediterranean. Standards in other European countries vary in detail, but are mainly based on the 1974 Bilthoven recommendations. Elsewhere, the principal standards in force are in Northern America. A selection are reproduced in Table 3. Some of them apply solely to freshwater as opposed to marine coastal areas, but afford a good idea of quality requirements utilized. TABLE 1. INTERIM QUALITY CRITERIA FOR MEDITERRANEAN COASTAL RECREATIONAL WATERS PROPOSED BY UNEP/WHO, 1983 Concentrations per 100ml Minimum Parameter not to be eKceeded number of Analytical Interpretation 50% 90% samples method method of the time Faecal coliforms 100 1 OOO 10 Membrane filtration, } m-FC broth or agar incubated at44.5 ± o.2°Cfor24h Graphical or analytical Faecal adjustment to a lognormal streptococci 100 1 OOO 10 Membrane filtration, probability distribution KF-streptococcus agar incubated at 36 ± 0.5 °c for 48 h Wld h/th statist. quart., 43 ( 1990) - 180 - TABLE 2. EEC MICROBIOLOGICAL QUALITY REQUIREMENTS FOR BATHING WATER COUNCIL DIRECTIVE OF 8 DECEMBER 1975 (76/160/EEC) Minimum Parameters G I sampling Method of analysis and inspection frequency Fermentation in multiple tubes. Subculturing of the 1 Total coliforms /100 ml 500 10 OOO Fortnightly positive tubes on a confim1ation medium (1) Count according to MPN (most probable number) 2 Faecal 100 2 OOO Fortnightly colifom1s /100 ml (1) or membrane filtration and culture on an appropriate medium such as Tergitol lactose agar, endo agar, 0.4% Teepol broth, sub-culturing and identification of the suspect colonies In the case of 1 and 2, the incubation temperature is variable according to whether total or faecal coli- forms are being investigated 3 Faecal 100 - (2) Litsky method streptococci /100 ml Count according to MPN (most probable number) or filtration on membrane. Culture on an appropriate medium 4 Salmonella /1 I - 0 (2) Concentration by membrane filtration. Inoculation on a standard medium. Enrichment-subculture on isolating agar-identification 5 Enteroviruses PFU/10 I - 0 (2) Concentration by filtration, flocculation or centrifuging and confirmation G =Guide. I - Mandatory. (0) Provision exists for exceeding the limits in the event of exceptional geographical or meteorological condttions. (1) When a sampling taken in previous years produced results which are appreciably better than those in this Table and when no new factor likely to lower the quality of the water has appeared, the competent authorities may reduce the sampling frequency by a factor of 2. (2) Concentrations to be checked by the competent authorities when an inspection in the bathing area shows that the substance may be present or that the qualtty of the water has deteriorated. Source: Reference (7). Epidemiological studies on health effects The basic aim of investigations into the health ef- fects of swimming or bathing have always been as originally defined by Stevenson in 1953 (8), i.e. to determine what difference in illness incidence might be expected from swimming in waters containing various degrees of bacterial pollution. Since that time, a number of studies have been carried out in an attempt to define the levels of risk following exposure to different concentrations of bacteria in bathing waters. The relevant literature, including details of the design and results of the studies performed, have been reviewed by Shuval (3) and by Jones & Kay (9). Outline of main results The Stevenson United States Public Health Service study of 1953 (8) has been described (3) as the first systematic prospective epidemiological study deal- ing with recreational waterborne disease in which microbial indicator organism concentration was used as a key intervening variable. Conducted in a number of sites in the form of three separate studies, the investigation showed higher morbidity rates among bathers, particularly children under 10 years, with ear, eye, nose and throat ailments repre- senting over 50% of the reported symptoms, with gastrointestinal illness accounting for less than 20%. A relationship between morbidity patterns and water quality was established at two sites, the latter based · on total coliform density. The results of this study were dismissed by Moore (10) as statistically dubious, on the grounds that: (a) the selection of groups for evaluation from among the whole popu- lations surveyed was biased; (b) all illnesses report- ed were lumped together; (c) the comparability of the two swimming groups in other material respects was not discussed; and (d) the inclusion of all illnesses reported during the week after bathing made the association of such ailments with the bathing episode more tenuous. Between 1953 and 1959, the British Public Health Laboratory Service carried out an extensive study of bathing in sewage-contaminated coastal waters (11, 12). The study showed no differentiation between bathers and non-bathers, even on beaches showing high coliform counts, and it was concluded that bathing in sewage-polluted water carried only a negligible risk to health, even on beaches that were aesthetically unsatisfactory. Rapp. trimest. statist. sanit. mond., 43 (1990) - 181 - TABLE 3. RECREATIONAL WATER QUALITY STANDARDS IN SELECTED COUNTRIES/AREAS, NORTHERN AMERICA Agency Sampling regime Total coliforms Faecal coliforms Other Reference 1. United States National Not less than 5 samples ,.;200/100 ml, nor shall more National Technical Technical Advisory taken over not more than than 10% of the samples Advisory Committee a 30-day period exceed 400/100 ml Committee, 1968 2. Province of Not less than 5 samples Running geometric mean Province of British Columbia taken over not more than ,.;200/100 ml; nor shall more British Columbia, a 30-day period than 10% of the samples 1968 exceed 400/100 ml 3. Inland Waters ,o;;500/100ml ,o;;200/100ml Environment Directorate, Canada, 1972 Department of the Environment 4. Alberta Not less than 5 samples Geometric mean Geometric mean Alberta Environment taken over not more than ,.;1 000/100ml ,o;;200/100ml Environment, 1977 a 30-day period 5. Ontario Ministries At least 1 0/30 days Geometric mean Geometric mean Faecal Ontario Ministry of Health, Environment ,,;;; 1 000/100 ml ,o;;100/100ml streptococci, of Health, 1975; Pseudomonas, Ontario Ministry Staphylococcus of the Environment, 1978 6. Committee of the Great Geometric mean Faecal Committee of the Lakes, Upper Mississippi ,o;;200/100ml streptococci Great Lakes, Upper River Board of State No samples to exceed Pseudomonas Mississippi River Sanitary Engineers 1 000/100ml Board of State Sanitary Engineers, 1975 7. United States Not less than 5 samples Geometric mean United States Environmental taken over not more than ,.;200/100 ml, nor shall more Environmental Protection Agency a 30-day period than 10% of the samples Protection Agency, exceed 400/100 ml 1976 8. Saskatchewan Not less than 5 samples Geometric mean Geometric mean Saskatchewan Environment taken over not more than ,.;1 000/100ml ,o;;200/100ml Environment, 1977 a 30-day period 9. Manitoba Clean Not less than 5 samples Median ,.;SOO MPN/100 ml Median ,.;200 MPN/100 ml Province of Environment per month Commission The Cabelli Environmental Protection Agency (EPA) study (1), carried out in the United States of America between 1972 and 1978, was a large-scale prospec- tive epidemiological-microbiological study, including both marine and freshwater beaches and involving 26 OOO subjects in three locations in the United States. The design characteristic involved subject recruitment on the beach itself, classification as swimmers only of persons immersing their head in the water, and follow-up interviews 7-10 days after the initial beach questionnaire. A system was also devised for validation of gastrointestinal symp- tomatology, though information regarding other ailments was also solicited. A range of micro- organisms, including both indicators and pathogens, was analysed in the water samples collected. The results showed that "the risks of gastroenteritis associated with swimming in marine waters impact- ed with municipal wastewaters was related to the mean enterococcus density in the water" (13-15). Prospective epidemiological studies of the same type (generally referred to as "Cabelli-style studies") were carried out in a number of countries between 1982 and 1989, including Israel (16), Canada (17), Egypt (18) and Hong Kong (19). Most of these in- volved marine beaches, and microorganisms em- ployed to assess water quality varied from one to a combination of total coliforms, faecal coliforms, fae- h Marino et al. Unpublished report, 1983. ; Borrego et al. Unpublished report, 1989. Wld h/th statist. quart., 43 (1990) Manitoba, 1979 cal streptococci, enterococci, E. coli and staphylo- cocci. Practically all of these obtained results in the form of higher morbidity among bathers as com- pared to non-bathers, the best correlation with water quality being with one or other microorganism where different beaches were compared. Two studies were conducted in Spain. The former (20) involved a large-scale cross-sectional study utilizing over 20 OOO respondents on various beaches in Malaga and Tarragona, where interviews were carried out. The highest morbidity rates ob- served concerned skin infections (2%), followed by ear and eye infections (nearly 1.5%), the latter group being considered by the authors as significantly associated with immersion of the head in seawater. Faecal streptococci concentrations gave better cor- relation with morbidity rates than faecal coliforms. The latter study, carried out on a number of Medi- terranean and Atlantic beaches in 1982,h was roughly similar in design, and again gave higher morbidity rates for ear, eye and skin infections. A positive correlation was found between water quality and morbidity, but in this case the best correlation with such morbidity was obtained with faecal coli- forms, threshold concentrations for increases in morbidity being calculated as in the range of 400 FC per 100 ml. A more recent study in progress in Spain; is being carried out on two beaches in Malaga, determining water quality in terms of total coliforms, faecal coli- forms, faecal streptococci, Pseudomonas aeruginosa, - 182 - Staphylococcus aureus, Aeromonas hydrophila, Vibrio species and Candida albicans. During the first year of the study (1988), surveys of test population groups were performed by a modified Cabelli-type questionnaire. Preliminary conclusions so far have been that: (a) enteric symptomatology was higher among swimmers than non-swimmers on the more polluted beach; (b) there was a significant excess in dermatological and respiratory symptoms among swimmers as compared to non-swimmers, irrespec- tive of recorded bacterial densities in seawaters; (c) no statistical significance was evident between swimming-associated symptom rates and faecal in- dicators; and (d) fungal density in beach sand was the same for both the polluted and non-polluted beach, and no relation was found between sand- associated dermatitis symptoms and dermal patho- gens. One other problem associated with this study was the fact that, out of 1 447 individuals surveyed, only 4.6% could be defined as non-swimmers, and the control group was therefore very small. The study is continuing with a larger population group. Studies carried out in France included one by Foulon in 1983 (21) on five beaches, the water quality of which was tested for concentrations of total coli- forms, faecal coliforms and faecal streptococci. Nearly 5 OOO individuals were interviewed on the beach and just over 1 500 followed up on the basis of a completed answer-card questionnaire. The re- sults indicated differences between the incidence of conjunctivitis and skin infections between bathers and non-bathers, as well as in the incidence of abdominal discomfort, nausea and pruritis between those immersing their head in the water and those refraining from doing so. There was, however, no correlation between morbidity and water quality. A later study (22), carried out in the Bassin de I' Ardeche on freshwater beaches, concerned over 5 700 subjects in the summers of 1985 and 1986. The water-quality indicators selected, apart from total coliforms, faecal coliforms and faecal streptococci, were Aeromonas species and Pseudomonas aerugi- nosa. Subjects were kept under constant observation for periods of between 3 and 8 days and symptoms recorded. Higher morbidity rates were found among bathers, as opposed to non-bathers, for each type of disease observed (gastrointestinal, skin, eye and ORL). Total coliform and faecal coliform con- centrations gave the best correlation with overall morbidity. When taking each type of symptom sep- arately, correlation varied with the different indica- tors, faecal streptococci providing the best correla- tion with gastrointestinal symptoms. Analysis of re- sults produced the conclusion that a concentration of 100 faecal coliforms per 100 ml (the EEC "guide" standard) corresponded to an incidence of 15.3 cases of overall gastrointestinal morbidity per 1 OOO person-days, and 2 OOO per 100 ml (the EEC "manda- tory" standard) to an incidence of 20.4 cases per 1 OOO person-days. The guide standard for faecal streptococci (100 per 100ml) was calculated as cor- responding to 23 cases per 1 OOO person-days. The calculation on expected incidence of gastrointestinal morbidity in non-bathers was not made but, on the basis of the authors' results, this would come to 8.5 cases per 1 OOO person-days. The authors' con- clusions include the statement that the results cannot be safely extrapolated to seawater. In their review of the literature, Jones & Kay" tabu- lated the summary results of the main Cabelli-style prospective epidemiological studies (Table 4). They reached the conclusion that there are "no good epidemiological data on which to base the im- plementation of scientifically justified quality stand- ards for recreational waters which would control for a defined level of risk". In an unpublished paper on epidemiological data presented at the WHO working group on health impact from human exposure to recreational waters, held in Rimini from 27 February to 2 March 1990, Colley, in quoting such con- clusions, points to some general problems in the conducting of epidemiological studies carried out so far, including: (i) difficulties in defining exposed and non-exposed population groups; (ii) uncertainty in the degree and duration of bathing exposure; (iii) availability of only limited information on pol- lutants present in the bathing water; (iv) inadequate identification of illnesses and their relationship to bathing; and (v) difficulties in the interpretation of any associations between illnesses and exposure to polluted water, and in ascribing a cause-effect re- lationship. He also points out that a number of issues remain to be addressed in future epi- demiological studies, including subgroups that may need to be identified both in terms of estimating risk TABLE 4. SUMMARY RESULTS OF THE MAIN CABELLI-STYLE PROSPECTIVE EPIDEMIOLOGICAL STUDIES• Author Date Country/area Fresh/sea Indicator R>• Symptoms' Stevenson 1953 United States of America Both Total coliforms NR ENT/GI/R Cabe Iii 1982 United States of America Both Enterococci 0.56 GI Seyfried 1985 Canada Fresh Total staphylococci 0.19 R/GI Faecal coliforms 0.08 Faeca I streptococci 0.03 Cheung 1988 Hong Kong Sea E.coli 0.53 S/GI Staphylococci EISharkawi 1983 Egypt Sea Enterococci 0.79 GI E. co/i 0.77 Fattal 1986 Israel Sea Enterococci NR GI E.coli NR GI Mujeriego 1982 Spain Sea Faecal streptococci NR S/E/ENT/GI Foul on 1983 France Sea Faecal streptococci NR E/S/GI Total coliforms Faecal coliforms • After Jones, F. & Kay, D. Recreational water quality: the relationship between epidemiological studies and recreational activities in water. Report of a biological standards seminar, Middlesex Polytechnic, 9 February 1990. •R2 • Coefficient of determination; NR R not reported. c E • eye infections; S • skin complaints; GI • gastrointestinal symptoms; ENT• ear nose and throat infections; R • respiratory illness. Rapp. trimest. statist. sanit. mond., 43 (1990) - 183 - and in addressing of preventive measures, the adequate follow-up of exposed and non-exposed persons, the obtaining of some measure of the "dose" bathers receive in estimating risk of illness, and the differentiation between marine and non- marine bathing exposures. Similarly important are non-bathing exposures, particularly consumption of contaminated food, as illness which has no relation to bathing may be present in a population, and must be recognized in order to prevent confusion with bathing-acquired illness, as well as sufficient evi- dence that the pathogen or agent causing a disease is also present in the water. According to Jones & Kay,C an unpublished thesis (the most recent and detailed attempt to replicate the Cabelli protocol) presented by Lightfoot in 1989 has highlighted a problem which may invalidate all previous prospective investigations, the design flaw inherent in most of which is that they have sought to measure disease perception, rather than incidence. Her comments include the possibility that utilization of medical and laboratory confirmation might have altered the results based on illness reporting, that such illness reporting was performed by contact persons and not confirmed by clinicians or labora- tory testing, and that future investigators would be well advised to attempt recording the duration of exposure for individuals, and to carry out more frequent water sampling each day. WHO has issued a protocol for epidemiological- microbiological studies in the Mediterranean to cor- relate coastal recreational water quality with health effects) Originally produced in 1985, this protocol i World Health Organization. Correlation between coastal water quality and health effects: report of a joint WHO!UNEP meeting, Folfonica, 21-25 October 1985. Copenhagen, WHO Regional Office for Europe, 1986. (WHO/EURO document ICP/CEH 001 m06). was rediscussed during WHO/UNEP expert con- sultations in 1987 and in 1989, mainly owing to the need for simplification to cater for limitations of Mediterranean institutions. As this protocol is based practically entirely on Cabelli-type prospective studie&, the basic design will have to be looked into again in the light of the recent evidence quoted above. Conclusions On the basis of evidence available so far, it would be difficult to attempt to quantify the actual health risks arising from bathing in sewage-polluted coastal bathing waters, as well as to correlate such risks to specific levels of water pollution as expressed in terms of the more routinely-measured bacterial indi- cator organisms. From the qualitative viewpoint, however, the evidence clearly indicates that health risks do exist, and are most pronounced in areas directly exposed to pollution by untreated sewage. Perhaps the best attempt at quantification was pro- duced by the 1983 Cabelli study (1), which developed data for regression lines for both total and "highly credible" gastrointestinal symptoms (i.e. nausea, vomiting, diarrhoea, and stomach-ache). The latter data were used by the EPA in developing a regres- sion line in a recommended health effects criterion for marine recreational waters (reproduced in Fig. 1), as it was considered that there was greater credibil- ity for those types of severe symptoms, and that the consequences of illness were of greater public health importance and thus more amenable to eco- nomic analysis (14). On the basis of such a criterion, one would expect to find 25-40 cases of gastrointestinal symptoms per FIG.1 RECOMMENDED HEALTH EFFECTS CRITERION FOR MARINE RECREATIONAL WATERS• CRITERES RECOMMANDES POUR LES EFFETS SANITAIRES DES EAUX MARINES A USAGE RECREATIF • 8g 00 ~ iii- a.5 "'a. E"' ~~ E o, ~-~ _.a ~.!!! ~-«I "'"' -~~ ;, "' "'"' """ O>x -:, ~-§ ~i -c E 11>"7 <a e ~ gj "'O> '""' . "'g'E e'£l §~ ;,: >, (/JC/) 70 60 50 40 30 20 10 0 10 0 Regression line (X on Y) - Droite de regression (XIV) 95% confidence limit - Limites de confiance ii 95% r= 0.75 Mean enterococcus density per 100 ml - Densite moyenne d'enterocoques pour 100 ml • After reference ( 74)- D'apres la reference ( 14). Wld hlth statist. quart., 43 { 1990) - 184 - 1 OOO persons exposed to seawater containing a concentration of 100 enterococci per 100 ml. In a hypothetical population of 1 million, this would mean 25 000-40 OOO cases, and further extrapolation to normal tourist and local populations in various regions of the world, particularly in tropical and subtropical climates, would accentuate the concern one has to accord to such a situation. Much higher estimates have been made by European tourist authorities, according to which some 40% of tourists on vacation at Mediterranean coastal resorts become ill at some time during or immediately after their visit (Shuval, personal com- munication). Undoubtedly, however, a portion of such tourist illness is associated with the consump- tion of contaminated food or unsafe drinking-water. A considerable amount of work is still required before a satisfactory dose-response relationship is obtained, on the basis of which quality criteria and standards calculated to present no appreciable hazards as a result of bathing can be developed. Apart from the confounding factors which have to be overcome in future microbiological/epidemio- logical studies correlating bathing water quality and health effects, including the major problems of (a) separating pollution-caused and bathing-asso- ciated effects, and (b) ensuring the effectiveness of indicator organisms selected, local circumstances, both environmental and socioeconomic, will have to be fully taken into account in the development of future protocols for such studies. SUMMARY A number of bacterial, viral and other diseases can be contracted by man through exposure to sewage- polluted bathing-water or beach sand. The increas- ing use of the sea for recreation has led to major concern regarding health hazards to both local and tourist populations. Epidemiological studies attempt- ing to correlate microbiological water quality with health effects have produced different results, lead- ing to a wide variation in recreational water quality criteria and standards applied, and to considerable controversy regarding their implementation. The Mediterranean provides a good example of a problem region where health risks are accentuated as a result of high utilization of bathing areas coupled with long exposure periods. Coastal pollu- tion by sewage is still a major concern and control measures vary considerably. A number of microbiological/epidemiological studies have been carried out since 1953 in an attempt to define the levels of risk following exposure to dif- ferent concentrations of bacteria in bathing waters. Most of these have been prospective studies whose design involved subject recruitment on the beach itself, classification of swimmers on the basis of immersion of the head in the water, and follow-up interviews after 7-10 days, together with a system for validation of gastrointestinal symptomatology. Practically all studies showed higher morbidity among bathers as compared to non-bathers, but correlation between specific symptoms and bacterial indicator concentrations varied considerably. On the basis of the 1972-1978 Cabelli Environmental Protec- tion Agency (EPA) study in the United States of America, later developed by EPA into a recommend- ed health-effects criterion for marine recreational waters, one would expect 25-40 gastrointestinal cases per 1 OOO persons exposed to seawater con- taining 100 enterococci per 100 ml. Extrapolation of these figures to annual bathing populations indicates the potential magnitude of the problem. Much more work is needed however before a sat- isfactory dose-response relationship is obtained, principally because of confounding factors inherent in all studies carried out so far, which still require a solution. RESUME Risques pour la sante lies a la pollution des eaux cotieres utilisees pour la baignade L'exposition a des eaux de baignade ou a du sable pollues par des eaux usees peut favoriser des mala- dies bacteriennes, virales et autres. L'engouement toujours plus marque pour les distractions bal- neaires accroit les preoccupations a l'egard des risques qu'elles presentent pour la sante des popula- tions locales et des touristes. Des etudes epidemio- logiques executees pour tenter d'etablir une correla- tion entre la qualite microbiologique de l'eau et ses effets sur la sante ont donne differents resultats conduisant a !'adoption de criteres et de normes tres varies pour la qualite des eaux de baignade et a des controverses considerables sur leur application. La Mediterranee est un bon exemple de region a problemes ou les risques pour la sante sont accen- tues par un taux d'utilisation eleve des zones de baignade associe a de longues periodes d'exposi- tion. La pollution des eaux cotieres par les eaux usees y est encore un probleme majeur et les mesu- res de controle varient considerablement. Plusieurs etudes microbiologiques/epidemiologiques ont ete entreprises depuis 1953 pour tenter de defi- nir les niveaux de risque apres exposition a differen- tes concentrations de bacteries dans les eaux de baignade. La plupart etaient des etudes prospectives Wld hlth statist. quart., 43 (1990) - 185 - consistant a recruter les sujets de l'etude sur la plage, a classer les nageurs selon qu'ils mettaient plus au mains la tete sous l'eau et a organiser des entretiens de suivi au bout de 7-10 jours, le tout associe a un systeme de validation de la symptoma- tologie gastro-intestinale. La quasi-totalite de ces etudes a fait apparaitre un taux plus eleve de morbi- dite chez les baigneurs par rapport aux non- baigneurs mais la correlation entre les sympt6mes particuliers et les concentrations de bacteries variait sensiblement. A l'issue de l'etude Cabelli de !'Envi- ronmental Protection Agency, conduite de 1972 a 1978 aux Etats-Unis d'Amerique, les criteres recom- mandes pour les eaux de mer utilisees pour la baignade etablissent que l'on peut attendre 25-44 cas de gastro-enterite pour 1 OOO personnes expo- sees a de l'eau de mer contenant 100 enterocoques pour 100 ml. L'extrapolation de ces chiffres au nom- bre annuel de baigneurs illustre l'ampleur du pro- bleme. Un travail enorme s'impose dans ces conditions pour obtenir une relation dose-reponse satisfaisante, principalement en raison des facteurs confondants inherents a toutes les etudes executees jusqu'ici et qui n'ont toujours pas ete resolus. TABLEAU 1. CRITERES PROVISOIRES DE QUALITE POUR LES EAUX C0TIERES MEDITERRANEENNES A USAGE RECREATIF PROPOSES PAR LE PNUE/OMS, 1983 Concentrations par100ml Nombre Para metre a ne pas depasser minimal Methode Methode 50% 90% d'echantillons d'analyse d'interpr0tation du temps Coliformes fecaux 100 1 OOO 10 Filtration sur membrane, } bouillon m-FC ou gelose incubee a.44,5 ± 0,2 °c pendant 24 heures Ajustement graphique Streptocoques analytique a une distribution fecaux 100 1 OOO 10 Filtration sur membrane, lognormale gelose KF-streptocoques incubee a 36 ± 0,5 °c pendant 48 heures TABLEAU 2. NORMES CEE DE QUALITE MICROBIOLOGIQUE DES EAUX DE BAIGNADE DIRECTIVE DU CONSEIL DU 8 DECEMBRE 1975 (76/160/CEE) Frequence Parametres G I d'echantillon- Methode d'analyse ou d'inspection nage minimal a Fermentation en tubes multiples. Repiquage des 1 Coliformes totaux /100 ml 500 10 OOO Bimensuelle tubes positifs sur milieu de confirmation (1) Denombrement selon NPP (nombre le plus probable) 2 Coliformes 100 2 OOO Bimensuelle fecaux /100 ml (1) ou filtration sur membrane et culture sur milieu approprie tel que gelose lactose au tergitol, gelose d'endo, bouillon au teepol 0,4%, repiquage et identification des colonies suspectes Pour les points 1 et 2, temperature d'incubation variable, selon que l'on recherche les coliformes totaux ou les coliformes fecaux 3 Streptocoques 100 - (2) Methode de Litsky fecaux /100 ml Denombrement selon NPP (nombre le plus probable) ou filtration sur membrane. Culture sur un milieu approprie 4 Salmonelles /1 I - 0 (2) Concentration par filtration sur membrane. lnocula- tion sur milieu type. Enrichissement, repiquage sur gelose d'isolement, identification 5 Enterovirus PFU/10 I - 0 (2) Concentration par filtration, par floculation ou par centrifugation et confirmation G = guide. I - Imperative. (0) Depassement des limltes prevues en cas de conditions geographiques ou meteorologiques exceptionnelles. (1) Lorsqu'un echantillonnage effectue au cours des annees precedentes a donne des resultats sensiblement plus favorables que ceux prevus au present tableau et lorsqu'aucune condition susceptible d'avoir diminue la qualite des eaux n'est intervenue, la frequence d'echantillonnage peut etre reduite d'un facteur 2 par les autorites competentes. (2) Teneur a verifier par les autorites competentes lorsqu'une enquete effectuee dans la zone de baignade en revele la presence ou une deterioration de la qualite des eaux. Source: Reference ( 7). Wld hlth statist. quart., 43 (1990) - 186 - TABLEAU 3. NORMES DE QUALITE POUR LES EAUX A. USAGE RECREATIF DANS CERTAINS PAYS/ZONES, AMERIQUE SEPTENTRIONALE Organisme Normes de prelevement Coliformes totaux Coliformes fecaux Autres Reference 1. National Technical Au minimum 5 echantillons ,;;200/1 OD ml; pas plus National Technical Advisory Committee preleves sur une periode de 10% des echantillons Advisory Etats-Unis maxi male de 30 jours depassant 400/100 ml Committee, 1968 2. Province de Au minimum 5 echantillons Moyenne geometrique Province de Colombie preleves sur une periode courante ,;;200/100 ml; Colombie britannique maxi male de 30 jours pas plus de 10% britannique, 1968 des echantillons depassant 400/1 DO ml 3. Direction generale des ,s;5Q0/100ml ,s;2Q0/100ml Environnement eaux interieures, Canada, 1972 Departement de l'environnement 4. Environnement Au minimum 5 echantillons Moyenne geometrique Moyenne geometrique Environnement Alberta preleves sur une periode ,;;1 000/100ml ,s;200/100ml Alberta, 1977 maxi male de 30 jours 5. Ministere de la sante Au mains 10 sur 30 jours Moyenne geometrique Moyenne geometrique Streptocoques Ministere de la sante et de l'environnement ,;;1 000/100ml ,s;100/100ml fecaux, del'Ontario, 1975; de !'Ontario Pseumodonas Ministerede Staphylococcus l'environnement de !'Ontario, 1978 6. Committee of the Great Moyenne geometrique Streptocoques Committee of the Lakes, Upper Mississippi ,;;20011oom1; fecaux, Great Lakes, Upper River Board of State aucun echantillon ne doit Pseudomonas Mississippi River Sanitary Engineers depasser 1 000/1 OD ml Board of State Sanitary Engineers, 1975 7. Environmental Au minimum 5 echantillons Moyenne geometrique Environmental Protection Agency preleves sur une periode ,;;200/100 ml; pas plus Protection Agency, Etats-Unis maxi male de 30 jours de 10% des echantillons Etats-Unis, 1976 depassant 400/100 ml 8. Environnement Au minimum 5 echantillons Moyenne geometrique Moyenne geometrique Environnement Saskatchewan preleves sur une periode ,;;1 000/lOOml ,;;20011oom1 Saskatchewan, 1977 maxi male de 30 jours 9. Commission Au minimum 5 echantillons Mediane ,;;SOO MPN/100 ml Mediane ,;;200 MPN/100 ml Province du Environnement parmois Manitoba, 1979 Manitoba TABLEAU 4. RESULTATS RECAPITULATIFS DES PRINCIPALES ETUDES EPIDEMIOLOGIQUES PROSPECTIVES DE TYPE CABELL!• Eaudouce/ Auteur Date Pays/zone eaudemer lndicateur R>• Sympt6mes 0 Stevenson 1953 Etats-Unis d' Amerique Douce+mer Coliformes totaux NR ENT/GI/R Cabe Iii 1982 Etats-Unis d'Amerique Douce+mer Enterocoques 0,56 GI Seyfried 1985 Canada Douce Staphylocoques totaux 0,19 R/GI Coliformes fecaux 0,08 Streptocoques fecaux 0,03 Cheung 1988 Hong Kong Mer E.coli 0,53 S/GI Staphylocoques EISharkawi 1983 Egypte Mer Enterocoques 0,79 GI E.coli 0,77 Fattal 1986 Israel Mer Enterocoques NR GI E.coli NR GI Mujeriego 1982 Espagne Mer Streptocoques fecaux NR S/E/ENT/GI Foul on 1983 France Mer Streptocoques fecaux NR E/S/GI Coliformes totaux Coliformes fecaux • D'apres Jones, F. & Kay, D. Recreational water quality: the relationship between epidemiological studies and recreational activities in water. Rapport d'un seminaire sur les normes biologiques, Middlesex Polytechnic, 9 fevrier 1990. (Anglais seulement). b R2 = Coefficient de determination: NA• non indique. c E • infections oculaires; S • troubles cutanes; GI = troubles intestinaux; ENT• infections nez-gorge-oreilles; R • affections respiratoires. Rapp. trimest. statist. sanit. mond., 43 (1990) - 187 - REFERENCES - REFERENCES 1. CASELLI, V. J. Health-effects criteria for marine recreational waters. Research Triangle Park, N.C., US Environmental Protection Agency, 1983. (R & D report No. EPA-600/1-80-031). 2. WHO Technical Report Series No. 550, 1974 (Fish and shellfish hygiene: report of a WHO Expert Committee convened in cooperation with FAO). OMS Serie de Rapports techniques N° 550, 1974 (Hygiene du poisson et des fruits de mer: rap- port d'un Comite d'experts de l'OMS reuni en cooperation avec la FAO). 3. SHUVAL, H. I. Thalassogenic diseases. Geneva, United Nations Environment Programme, 1986. (UNEP regional seas reports and studies, No. 79). 4. MOOD, E. W. & MOORE, B. Health criteria for the quality of coastal and bathing waters. (Periodical publication). 5. BR1sou, J. An environmental sanitation plan for the Mediterranean seaboard. Geneva, World Health Organization, 1976. (Public Health Papers, No. 62). BR1sou, J. Mesures a prendre en vue d'assurer la salubrite du littoral mediterraneen. Geneve, Organisation mondiale de la Sante, 1975. (Cahiers de sante publique n° 62). 6. RITCHIE-CALDER, LORD. Pollution of the Medi- terranean. In: Pollution of the Mediterranean Sea-Pacem in Maribus II. Proceedings of a Con- ference in Malta, 29 June-5 July 1971. Berne, Herbert Lang, 1972. 7. EUROPEAN ECONOMIC COMMUNITY. Council Directive concerning the quality of bathing water. Official journal of the European Communities, L 31: 1-7 (1976). (76/160/EEC). COMMUNAUTE ECONOMIOUE EUROPEENNE. Directive du Conseil concernant la qualite des eaux de baignade. Journal officiel des Communautes europeennes, L 31: 1-7 (1976). (76/160/CEE). 8. STEVENSON, A. H. Studies of bathing water quality and health. American journal of public health, 43: 529 (1953). 9. JONES, F. & KAY, D. Bathing waters and health studies. Water services, 93 (1117): 87-89 (1989). 10. MOORE, B. The case against microbial standards for bathing beaches. In: Gameson, A. L. H. (ed), The discharge of sewage from sea outfalls. Oxford, Pergamon Press, 1974. 11. PUBLIC HEALTH LABORATORY SERVICE. Sewage con- tamination of coastal bathing waters in England and Wales: a bacteriological and epidemiological study. Journal of hygiene, 57 (4): 435-472 (1959). Wld hlth statist. quart., 43 (1990) 12. MEDICAL RESEARCH COUNCIL. Sewage contamination of bathing beaches in England and Wales. Med- ical Research Council memorandum, 37: 1-32 (1959). 13. CASELLI, V. J. ET AL. Swimming-associated gas- troenteritis and water quality. American journal of epidemiology, 115 (4): 606-616 (1982). 14. CASELLI, V. J. ET AL A marine recreational water quality criterion consistent with indicator con- cepts and risk analysis. Journal of the Water Pollution Control Federation, 55: 1306-1314 (1984). 15. CASELLI, V. J. Swimming-associated illness and recreational water quality. Water science and technology, 21 (2): 13-21 (1989). 16. FATTAL, B. ET AL. The association between sea- water pollution as measured by bacterial indica- tors and morbidity of bathers at Mediterranean beaches in Israel. Chemosphere, 16 (2/3): 565- 570 (1987). 17. SEYFRIED, P." L. ET AL. A prospective study of swimming-related illness. American journal of public health, 75 (a): 1068-1075 (1985). 18. EL SHARKAWI, F. M. & HASSAN, M. N. E. R. The relation between the state of pollution in Alexan- dria swimming beaches and the occurrence of typhoid among bathers. Bulletin of the High Institute of Public Health, Alexandria, 12: 337- 351 (1982). 19. CHEUNG, w. H. s. ET AL. Health effects of beach- water pollution in Hong Kong. In: Proceedings of the Institute of Water and Environmental Man- agement-Proceedings of the annual conference. New Haven, Connecticut, Yale University School of Medicine, 1976. 20. MuJERIEGO, R. ET AL. Recreation in coastal waters: public health implications. In: vr• journees d'etudes sur les pollutions marines en Mediter- ranee, Cannes, 2-4 decembre 1982. Monaco, Commission internationale pour !'exploration scientifique de la mer Mediterranee, 1983. 21. FouLON, G. ET AL. Etude de la morbidite humaine en relation avec la pollution bacteriologique des eaux de baignade en mer. Revue franr;aise des sciences de l'eau, 2 (2): 127-143 (1983). 22. CENTRE ALPIN DE RECHERCHE EPIDEMIOLOGIOUE ET DE PREVENTION SANITAIRE. Etude epidemiologique des effets sur la sante de la contamination bacterio- logique des eaux de baignade: cas des eaux de riviere du Bassin de l'Ardeche - Rapport de syQthese. Grenoble, CAREPS, 1987. - 188 - THE ROLE OF ENVIRONMENTAL AND OCCUPATIONAL HAZARDS IN THE ADULT HEAL TH TRANSITION Tard Kjellstrom" & Linda Rosenstock" The "health transition" is defined as the phase of major change in the nature and extent of ill-health associated with changes in economic and social development. This extends the concept beyond the more narrow health transition introduced by Caldwell< to describe the behavioural changes re- lated to maintaining and improving health. In broad terms, the health transition involves a shift towards improved levels of health and wellbeing, with a greater proportion of diseases being noncommunic- able (non-infectious chronic and degenerative dis- eases, such as cancer and cardiovascular disease), even though the actual mortality rates from these diseases are concomitantly declining. Among the numerous implications of this transition is the changing impact of occupational and environmental hazards. The occurrence of these hazards changes with economic and social development, and their impact on health varies in different age groups in such a way that changes in the population structure influence the overall effect of these hazards. In addi- tion, since the environmental and occupational hazards are largely preventable, their health impact will to a great extent be determined by policies and actions for prevention. In order to consider more • Division of Environmental Health, World Health Organization, Geneva. O Occupational Medicine Programme, University of Washington, Seattle, United States of America. 'Caldwell, J.C. Introductory thoughts on health transition. Paper presented at the Rockefeller Foundation workshop on cultural, social and behavioural determinants of health, Canberra, Australia, 1989. fully the role of these hazards on health, it is neces sary first to review the historical background and the important components of the health transition. The demographic transition The demographic transition, a concept first de- scribed about 50 years ago, refers to the effects of evolving patterns of vital status-birth rates and death rates-on population size and distribution. This transition takes place in three stages: (i) popula- tion equilibrium, where birth rates (fertility) and death rates (particularly infant mortality) are both high; (ii) population explosion, where birth rates remain high in the face of decreasing mortality rates; and (iii) return to population equilibrium, where in response to a complex set of factors arising from decreasing mortality rates, fertility rates also decline. These stages are depicted in Fig. 1. This pattern has been observed in many countries (1-4), although in a number of developing countries and subpopulations within developed ones, a slowing down has been noted in the rate of decline in fertility-an important factor in the transition from stage (ii) to stage (iii) (1, 5, 6). The epidemiological transition The "epidemiological transition", the next term to be coined in transition theory, was grounded in the demographic transition and first described by Omran FIG.1 40 30 20 10 STAGES OF DEMOGRAPHIC TRANSITION ETAPES DE LA TRANSITION DEMOGRAPHIQUE Stage 1 - 1 • etape ...... ......................... ... •• •• •• •• Stage 2 - 2' etape •• • .· •• •• •• •• 50-100 nme (years) - Temps (annees) Stage 3 - 3' etape ...................... ....... Population ...... t--+-- (absolute numbers/ nombres absolus) Rapp. trimest. statist. sanit. mond., 43 (1990) - 189 - in 1971 (7). The epidemiological transition refers to the secular changes in patterns of health and disease in relation to social, economic and demographic factors. According to this theory, as countries evolve through higher levels of development, improved so- cial, economic and health conditions cause a trans- ition from short life expectancy, with high rates of particularly early mortality from infectious diseases, to increased survival into older ages and the re- sultant relatively greater proportion of deaths from noncommunicable disease, or from what Omran re- ferred to as "degenerative or man-made diseases" (Fig. 2). It should be noted that the original analysis of the epidemiological transition (7) was based on the proportional mortality rates for different dis- eases. As shall be discussed later in this article, this type of analysis may lead to the mistaken impres- sion that noncommunicable disease mortality rates (per 1 OOO population) are increasing, which is gen- erally not the case. Infectious disease mortality is falling more rapidly than noncommunicable disease mortality, which leads to the relative increase of the latter as a part of total mortality. A transition through three stages of changing pat- terns of disease (1. pestilence and famine; 2. reced- ing pandemics; 3. degenerative and man-made dis- eases) was described by Omran (7) as occurring at three different rates among different populations. Accordingly, he described three models: (i) the classic or Western model, as typified by England and the United States of America, where changes evolved over centuries; (ii) the accelerated transition model, as seen in Japan, where stages 2 and 3 were compressed into decades; and (iii) the contemporary delayed model, typified by many developing coun- tries. In contrast to the first two models which were largely driven by changes in social and economic conditions and not medical and health care (3, 7), in the delayed model there has been a massive influx of medical and health-care technology influencing the disease pattern, but a slowing of the rate of the decline in mortality and fertility. Omran (8) subse- quently described a fourth model of this transition, the transitional model: a variant of the delayed model typified by some developing countries, such as Sri Lanka and Singapore, where the mortality decline continues at a regular rate and is followed within decades by declining fertility rates which are due at least in part to organized family planning, education and social development efforts (8). More recently other authors have proffered refine- ments of the theory; two different fourth stages have been suggested. The first is the age of delayed degenerative diseases. Because of unanticipated rapid improvements in survival among the older ages, there is a shift in the age distribution of causes of death to older ages. The result is increased morbidity, because those affected are more suscept- ible and more likely to become disabled, although incidence rates still decline in this oldest age stratum (2). The second suggestion for a fourth stage in the epidemiological transition, the hybristic stage (from the term hybris-or hubris-meaning excessive self-confidence), is derived from the obser- vation that many societies, as they become more developed, have a larger proportion of deaths at- tributable to detrimental individual behaviours and lifestyles (9). In this stage, so-called social pathol- ogies (such as cigarette smoking, violence, diseases and accidents attributable to alcohol, drug abuse, or FIG.2 THE EPIDEMIOLOGICAL TRANSITION IN A TYPICAL DEVELOPED COUNTRY TRANSITION EPIDEMIOLOGIQUE DANS UN PAYS DEVELOPPE TYPE 1900 Years - Annees 1980 Based on reference ( 7) - O'apres la reference ( 7). Wld hlth statist. quart., 43 (1990) - 190 - sexual behaviours) assume a more dominant influ- ence and rise more rapidly than the proportionate increase in degenerative disease deaths (4, 9-11 ). This unanticipated development in the epidemiologi- cal transition, including most notably the AIDS epi- demic, can be associated with significant losses in life expectancy, particularly among young adults. Other transition concepts With the increase in life expectancy that follows the avoidance of premature deaths during infancy, child- rate of evolution (the classic transition model); for many countries now undergoing development, a relatively high level of both risks (traditional and modern) exists simultaneously. An unexplored po- tential of this risk overlap is a synergistic or inter- active effect of both types of risk causing worse sequelae than if the risks were merely additive. For example, inadequate water supply combined with the introduction of pesticides is likely to heighten the risk of pesticide poisoning-as occupational and environmental exposures are increased from pro- longed contact of pesticides on skin and clothing. hood and female reproductive years, comes not the The health transition concept has five fundamental eradication of premature and preventable deaths but components: (i) the health situation itself; (ii) the a shift in their causes and affected populations (10). treatment given to ill people (the curative inter- The term risk transition was recently coined (12) to ventions); (iii) the specific causal factors for ill-health refer to the changing pattern of disease deter- (e.g. tobacco smoking); (iv) the preventive inter- minants that accompanies economic development. ventions against such causal factors; and (v) the There is a change from traditional environmental socioeconomic, behavioural or political framework exposures (such as bacterial water contamination within which curative and preventive interventions and diarrhoeal diseases) and their associated health are carried out. An important consideration in the risks, to those associated with agricultural moderniz- health transition is that different levels of preventive ation and industrialization (such as pesticide use and action may occur at the same economic develop- resultant acute and chronic morbidity). As depicted ment level (13). Many specific preventive inter- in Fig. 3, an important implication of the risk transi- ventions (such as seat belts in cars, strict food tion for the developing world is that because of the hygiene in commercial premises or urban air pollu- direct and indirect influence of the developed world, tion control) are not an inevitable part of economic modern risks and technologies are introduced during development itself. These interventions require well- earlier phases of economic development than in the informed and conscious decisions by community past. This gives rise to a situation different from that leaders and other community members. As Fig. 4 of developed countries which underwent a slower shows, there may be a situation of a steadily in- Low-Faible FIG. 3 THE HEALTH RISK TRANSITION• TRANSITION DES RISQUES POUR LA SANTE • Stages of development - Stades de developpement High -Eleve a In countries undergoing the classic epidemiological transition (A), modern risks increase at a time when traditional risks have virtually been removed. But, increasingly, devel· oping countries are facing the influx of modern risks earlier in their development (B). The shilt of the curve to the lelt causes the risk overlap. - Dans les pays soumis ii la transition epidemiologique classique (A), les risques modernes augmentent lorsque les risques traditionnels ont ate quasiment elimines. Mais les pays en developpement sont de plus en plus exposes a un courant de risques modernes a un stade plus precoce de leur developpement (B). Le deplacement de la courbe vers la gauche entrains le recouvrement des risques. Adapted from reference ( 12) - D'apres la reference (12). Rapp. trimest. statist. sanit. mond .• 43 (1990) - 191 - FIG.4 INTERVENTION TRANSITION • TRANSITION DES INTERVENTIONS • ~ c: c: 0 ·;;; "' "'"' 'C-me l;j~ .c"' -::i "'C' C:u, o·;:: ~ 8W o- ~ai .,E -"' "'" -c: c: 0 "'~ e·;: c:C: ef ">-m !~ o- !j:ll aig I::·;:: a lfi 8-c "' u c: "' 'C ·;:; .!: '"' iii w I .c "' :f <I> :0 ... u. I !!: 0 ..J A B Hazard occurrence Incidence des risques (Low prevention - Prevention faible) (High prevention - Prevention elevee) Time or economic development - Temps ou developpement economique • A: In low prevention situations, both hazard occurrences and health risks are increasing - Lorsque les mesures de prevention sont insuttisantes, !'incidence des risques et les risques pour la sante augmentent. B: In high prevention situations, there is a separation between rising hazard occurrences and decreased associated risk rates - Lorsque les mesures de prevention sont importantes, l'ecart se creuse entre !'incidence croissante des risques et la baisse du taux des risques apparentes. creasing occurrence of a particular hazard, but a peaking and then decreasing health risk (at com- munity level) of this hazard due to the introduction of preventive interventions. This transition from low prevention to high prevention may be labelled the "intervention transition". Good examples of this transition are traffic-accident mortality in most European countries (peaked in the 1970s) and urban air pollution morbidity/mortality in London (peaked in the 1950s). The importance of adult health in the health transition One of the important postulates of the epidemiologi- cal transition was that the shift in health and disease patterns was not equally distributed by age or gen- der, with the greatest changes occurring among women and children (7). Accordingly, in stage 2 (receding pandemics), the improvements in survivor- ship occur disproportionately in children and women in the adolescent and reproductive years, probably because of the increased susceptibility of those groups to infections and diseases of nutritional defi- ciency. At the same time, public health efforts have also been targeted to these subpopulations, par- ticularly children. These and other forces have con- tributed to the overall improvements in life expect- ancy, with a significantly higher proportion of Wld hlth statist. quart., 43 ( 1990) children in the world now surviving until age 15. In addition, as countries evolve through economic de- velopment, significant improvements in survivorship have also been occurring among adults of both sexes; in many developed countries the rate of decline in mortality for adults aged 15-59 has par- alleled that seen for children < 5. These and other features of the health transition among adults are described in a recent report from the World Bank (14). A convenient way of summar121ng the mortality trends for different broad age groups is the use of "cumulative mortality rates" (CUMR) or "mortality risks", which express the probability that a person of a certain age (e.g. 15 years) will die before another age (e.g. 60 years). With the terminology used in this article it may be denoted CUMR (15, 59). The CUMR (15, 59) for most developed countries has decreased from about 40% to less than 15% during this cen- tury, but many developing countries still have a CUMR (15, 59) at the higher level (14). Based on data from the World health statistics an- nual (15) and the life tables by Preston et al. (16), CUMR (15, 59) was calculated for a selection of countries representing four different economic de- velopment levels. These results support the findings in the World Bank report (14). Male adults in the developing world, however, face significant risks of - 192 - dying before age 60, with estimates ranging from 20-50% (Fig. 5). Interestingly, in some developing countries the risk of dying in adulthood for males has remained relatively high while the risk for adult females has approached that of the most developed countries. As also shown in Fig. 5, the widely dis- parate mortality risk among countries at differing levels of development reflects the remarkable gains in adult survivorship during the demographic tran- sition-and underlines the need for identifying and preventing causes of adult mortality. A number of factors contribute to the increasing importance of changing patterns of health and dis- ease among adults (particularly those aged 15-59) in the health transition . These include: (i) an increasing number of children are now surviving into adulthood (approximately two-thirds of the population in the developing world is over age 15); (ii) adults have important societal roles (in reproduction, support of others and production), and face a unique pattern of hazards (including occupational, environmental and psychosocial); (iii) there is significant potential to improve survivorship within these age groups, as shown by overall declining mortality rates with in- creasing development; (iv) the widening gender gap in some countries, with males facing a slowing in the decline of adult mortality rates, suggests that resources should be directed at hazards likely to contribute to observed gender differences; and (v) even within developed countries, some subpopu- lations face very high risks of adult mortality and should be targeted for interventions (14). Cause-specific adult mortality in the health transi- tion: discerning the role of occupational and environ- mental hazards One of the main tenets of the health transition is that as overall mortality rates decline so too do non- communicable-disease mortality rates, even though these diseases make up a greater proportion of the causes of death. This seem ing paradox (which might be termed the "proportion fallacy") of rising propor- tionate mortality in the face of decreasing absolute risk is best demonstrated for cardiovascular dis- eases. As illustrated in Fig. 2, these diseases are now the leading cause of death in countries which have undergone the epidemiological transition, even though there has been an unanticipated and ac- celerated decline in age-specific mortality rates for these conditions in some countries (2, 9, 17). It should be pointed out that the category "other causes" in Fig. 2 includes all the deaths without specific diagnosis-the undefined deaths. These de- crease rapidly after the 1950s, and the cause-specific mortality trends need to be adjusted for the changes in the proportion of undefined deaths. Most of these deaths appear to be due to heart disease or can- cer (16). The changes in mortality rates concomitant with development are illustrated in Table 1, which shows the calculations made of the decline in mortality risk for adult men (aged 15-59), based on some of the limited data available (15, 16). The figures are ad- justed for the undefined deaths, distributing these FIG. 5 50 ~ <ii t:: 0 E Q) " ·~ :, E 40 ::, " ~ g ir I "' c 30 ·;;_ " 0 "' ·~ Q) > ~ ~ 20 0 RISK OF DYING BETWEEN THE AGES OF 15-59 FOR MEN AND WOMEN RISQUE DE MORTALITE CHEZ LES HOMMES ET LES FEMMES DE 15 A 59 ANS - Males Hommes ~ Females Femmes Japan - Japan Sweden - Suede Chile - Chi Ii Egypt - Egypte Rural India - lnde rurale Country - Pays Rapp. trimest. statist. sanit. mond. , 43 ( 1990) - 193 - TABLE 1. MORTALITY RISK FOR ADULT MEN(%) AGED 15-60, BY STAGE OF DEVELOPMENT• TABLEAU 1. RISQUE DE MORTALITia CHEZ LES ADULTES DU SEXE MASCULIN (%) DE 15-60 ANS, SELON LE NIVEAU DE DiaVELOPPEMENT• Very low Low Intermediate High Tresfaible Faible Mayen Eleve Heart disease - Cardiopathie 10 10 5 4 Cancer ............ 4 4 4 4 Injuries - Traumatismes . . . 10 6 5 3 Infection ............ 16 1.5 0.9 0.2 Chronic lung disease - Affection pulmonaire chronique . 4 0.8 0.3 0.2 All causes of death - Toutes les causes de deces . 50 25 20 13 Country examples - Exemples nationaux . . . . . . Rural India Egypt - Egypte Chile Sweden - Suede lnderurale Sri Lanka Chili Japan - Japan • Calculated from mortality data published in the World health statistics annual 1987 & 1988 - Calcule a partir des donnees de mortalite publiees dans l'Annuaire de statistiques sanitaires mondia/es 1987 & 1988. equally into "heart disease" and "cancer". Also shown is the trend towards a decline in overall fatal injury rates. Cancer death rates are relatively stable, in large part due to the ongoing lung-cancer epi- demic which has occurred simultaneously with markedly declining gastric carcinoma rates. Chronic lung disease mortality risk decreases almost as dramatically as mfectious diseases (Table 1). These evolving cause-specific mortality risks are guides to the importance of specific environmental and occu- pational hazards in the health transition. In order to explore further these hazards in relation to the health transition, a discussion of some specific exposure-effect relationships is necessary. The first factor to consider is cigarette smoking, a hazard that has actually slowed the rate of improve- ment expected in the health transition. The role of cigarette smoking as the greatest single contributor to the lung-cancer epidemic is well-acknowledged: the reduction of cigarette use would have the greatest impact of any single intervention on the number of cancer deaths (18). Tobacco smoke ex- posure has increased over the course of the health transition, only now declining in some of the most developed countries. But tobacco smoking alone is not responsible for the absolute and proportionate increased mortality risks from lung cancer. Other environmental hazards are known to independently cause lung cancer (such as radon, asbestos, arsenic and constituents of indoor and outdoor air pollu- tion). In the United States, for example, up to 15% of lung-cancer deaths for men and 5% for women have been attributed to occupational hazards (18). Chronic obstructive lung diseases are also strongly related to smoking in the developed countries with high smok- ing prevalence (19), whereas in non-smoking popula- tion groups these diseases are related to infections, indoor air pollution from biomass smoke, outdoor air pollution and occupational dust exposures (20). Thus, during the health transition communities are likely to experience a complex transition of the hazards associated with lung diseases, all of these hazards being preventable. One of the greatest threats to the future health of developing-country populations is increasing tobacco-smoking habits. Unless the promotion of tobacco smoking is cur- tailed or eliminated, the health transition will be markedly slowed in these countries. Another important aspect of the role of environ- mental hazards is their potential synergism, where the combined effect of two hazards multiplies the risk of each individual hazard alone. Synergism, also known as interaction or effect modification, is well Wld hlth statist. quart., 43 (1990) understood for the contribution of cigarette smoking and asbestos exposure to risk for lung cancer. As shown in Table 2, the combined effect of smoking and asbestos in an occupational cohort (21) with heavy asbestos exposure is a 50-fold increase in the risk for lung cancer; asbestos exposure has quin- tupled the lung-cancer risk both for the smoker and the non-smoker. One important aspect of synergism between two carcinogens is that the elimination of TABLE 2. RELATIVE RISKS - RELATION BETWEEN ASBESTOS, SMOKING AND RISK FOR LUNG CANCER IN ASBESTOS WORKERS TABLEAU 2. RISQUES RELATIFS - RELATION ENTRE L'AMIANTE, LE TABAGISME ET LE RISQUE DE CANCER DU POUMON CHEZ LES TRAVAILLEURS DE L'AMIANTE Smoking• - Tabagisme• No - Non Yes - Oui Asbestosexposureb - Exposition a l'amianteb No - Non Yes - Oui 1 10 5 50 • 11 smoking is eliminated, eliminate 90% lung cancers in asbestos workers - En eliminant le tabagisme, on elimine 90% des cancers du poumon chez les travailleurs de l'amiante. " If asbestos is eliminated, eliminate 80% lung cancers in exposed asbestos workers - En eliminant l'amiante, on elimine 80% des cancers du poumon chez tes travailleurs de l'amiante exposes. Based on data in reference (19) - D'apres les donnees de la reference (19). either one has a significant effect on overall reduc- tion for risk of disease (22). In the population de- scribed in Table 2, elimination of the smoking ex- posure alone would result in 90% fewer lung-cancer deaths. But even if the smoking behaviour were unchanged in the group of workers who were ex- posed to asbestos, elimination of the asbestos ex- posure would result in 80% fewer lung-cancer deaths in this group. The implications for prevention are obvious, but so are the concerns raised by this example of what might be considered risk overlap in the risk transition; much of the developing world is now facing an exponential increase in exposure to cigarette smoke at the same time as occupational and environmental exposures to asbestos may be increasing. In contrast with the situation 40-50 years ago in now developed countries, these simultaneous exposures are evolving with full knowledge of their dire consequences. The contribution of deaths from motor-vehicle col- lisions during the health transition is also illustrative of avoidable causes of adult mortality on the path to - 194 - development. This environmental hazard and its consequences also illustrate some of the complexity of the difference between rates and proportions. Here, the rate of motor-vehicle fatalities per 100 OOO population typically increases with increasing de- velopment, but with development comes other evi- dence of decreasing risk in the face of increased exposure to a hazard; the mortality per vehicle or mile travelled typically decreases significantly. In contrast, in the developing world the rapid increase in use of motor vehicles is associated with rapidly increasing vehicle-related mortality, particularly for the passengers in motor-vehicle/pedestrian and motor-vehicle/bicycle collisions (23). Thus, although there may be stable or increasing rates of motor- vehicle accident deaths per population with develop- ment, the finding in some developing countries (e.g. Nigeria and Kenya) of a mortality rate per vehicle which is 30-50 times the rate found in the United States or Sweden suggests great potential for pre- venting these avoidable causes of death with their attendant high economic and social costs. In some subpopulations, motor-vehicle accidents have a con- siderable effect on mortality; among the Navajo tribe in the United States, for example, motor-vehicle accidents are the leading cause of death, with an age-adjusted increase 5.5 times that for the United States population (6). It has been estimated that for Navajo males, the elimination of such accidents alone would result in a gain of life expectancy of over 5 years. Another factor of importance in the health transi- tion-the changing pattern in case-fatality rates-is well demonstrated by the accumulating data on pesticide production, use and health effects. Recent estimates suggest that each year worldwide there are 3 million acute severe pesticide poisonings, of which about two-thirds may be suicide attempts, with 220 OOO deaths (24, 25). Although there has been a marked increase in pesticide production and use, much of it in the most highly industrialized countries, most of the burden of pesticide poisoning is borne by developing countries, where it has been estimated that 99% of fatal pesticide poisonings occur (26). The reasons for this are complex, but a number of factors in the developed world have contributed to reduce the number and severity of poisonings, including technological advances, re- sources for protective clothing and equipment, sub- stitution with less toxic chemicals, and medical ser- vices. Although reliable data on the full impact of environ- mental and occupational factors on the improved survivorship in the adult health transition are not available, a very crude analysis suggests they play a substantial role. As shown in Table 1, there has been an improvement of about 12% in mortality risk in the male adult age group (15-59 years) in the transition from low to high development (because the data are less reliable for countries at a very low stage of development, considered here is the change from 25% to 13% from low to high development). About 3% of the reduction of mortality risk is from avoid- ance of injuries, including agricultural and industrial accidents, and violence (homicide). There has been a 1.3% decrease of adult mortality risk from infections, which are largely influenced by environmental fac- tors (tuberculosis, septicaemia and tetanus (15)). In addition, a 0.6% decrease of chronic lung-disease mortality risk could also be linked to environmental exposures to a considerable degree (indoor air pollu- tion, organic dusts in agriculture, inorganic dusts in mining and industry, etc.). Perhaps as much as one-third of the gains in adult health in the health transition are due to changes in disease patterns largely or entirely explained by occupational and environmental factors. Further data collection and analysis are needed to identify cause-specific changes, particularly in relation to specific hazards, in order to better target resources and design pre- vention strategies for adults in countries with varying stages of development. Conclusions In conclusion, this discussion has reviewed several aspects of transition theory in relation to the poten- tial role of occupational and environmental factors as populations undergo social and economic devel- opment. The analysis of changing patterns of cause- specific mortality and hazard occurrence allows a framework for developing specific strategies to pro- mote and accelerate the decline in avoidable causes of death. As countries evolve through economic development, there is a progressive change in the type of environmental and occupational hazards ex- posures from the traditional hazards of poor sanita- tion, indoor air pollution and agricultural accidents to the modern hazards of urban air pollution, toxic chemicals and traffic accidents. The effect on mor- tality of the traditional hazards appears to be higher than the modern hazards, in that mortality risk re- lated to the environment appears to decrease con- siderably during the health transition. The sooner appropriate prevention interventions are undertaken, the sooner will the resultant benefits in better health be realized. SUMMARY In the course of economic development from the pre-industrial to the post-industrial stages, human communities go through major social and economic transitions. The demographic transition, a long es- tablished concept, refers to the change from a stage of high birth rates and mortality rates to one of low birth rates and mortality rates. The mortality rates have always decreased before the birth rates, giving rise to a period of rapid population growth in each society. The recently-coined term "health transition" provides a more detailed view of the different aspects of the mortality decline and of the changes in morbidity and causes of morbidity which also accompany economic development. The health transition is associated with major changes in exposure to environmental and occu- pational health hazards. The traditional hazards of the pre-industrial stage include unsafe drinking- water, poor sanitation, infected food, tropical disease vectors, indoor air pollution from biomass smoke, and accidents in agriculture and fishing. Gradually these are replaced by or overlap with the modern hazards, such as urban air pollution, tobacco smok- ing, pesticides, occupational hazards in new in- dustries and traffic accidents. The health impact of these hazards depends very much on the preventive measures taken. For some hazards there is an in- creasing health risk, in parallel with the increasing Rapp. trimest. statist. sanit. mond., 43 ( 19901 - 195 - occurrence of the hazard, which is followed by a decreasing health risk as preventive actions become effective. This intervention transition does not occur automatically with economic development, but re- quires specific policies and plans, as well as com- munity involvement. The adult age group (15-59 years) is likely to be at particular risk for environmental and occupational hazards which emerge during the health transition. The cumulative mortality rate (CUMR (15,59)) or risk of dying between the 15th and 60th birthday, is used as a measure of the health impact. In developed countries this CUMR is about 13% for men and, about 6% for women, a decrease from a CUMR of about 40% at the turn of the century. In developing countries there are geographical areas with a CUMR (15, 59) as high as 50% both for men and women, but at the national level the highest CUMR (15, 59) is about 25% in countries with systematic mortality data recording systems. An analysis of the influence of environmental and occupational hazards on this decrease of mortality shows that a considerable proportion, maybe one- third of this change, is due to decreasing mortality from diseases associated with such hazards. Acci- dent-injury mortality, infectious-disease mortality, and lung-disease mortality decrease considerably during the health transition. Although some of this decrease can be attributed to improved medical treatment, the changing pattern of occurrence of different hazards and the application of effective preventive interventions no doubt play a major role in the health transition. RESUME Le role des risques environnementaux et professionnels dans la transition sanitaire chez les adultes Au fur et a mesure du developpement economique, de l'ere preindustrielle a l'ere postindustrielle, les communautes humaines passent par des transitions sociales et economiques majeures. La transition de- mographique, notion bien connue, correspond a une evolution au cours de laquelle les taux de natalite et de mortalite baissent considerablement. Les taux de mortalite ant toujours diminue avant les taux de natalite, d'ou une periode de croissance demogra- phique rapide dans toutes les societes. L'expression «transition sanitaire1>, forgee recemment, donne une idee plus fine des differents aspects du declin de la mortalite ainsi que de !'evolution de la morbidite et des causes de morbidite qui accompagnent le deve- loppement economique. La transition sanitaire est liee a des changements majeurs de !'exposition aux risques que comportent pour la sante l'environnement et le travail. Les ris- ques traditionnels de l'ere preindustrielle sont no- tamment l'eau de boisson polluee, l'insalubrite du milieu, la contamination des aliments, les vecteurs de maladies tropicales, la pollution a l'interieur des habitations due a la fumee de la biomasse et enfin les accidents dont sont victimes agriculteurs et pe- cheurs. Ces risques sont peu a peu remplaces - quand ils ne viennent pas s'y ajouter - aux risques du monde moderne, comme la pollution de l'air des villes, !'usage du tabac, les pesticides, les risques professionnels dans les industries nouvelles et les accidents de la circulation. L'impact de ces risques sur la sante depend pour beaucoup des mesures preventives prises. Dans certains cas, le risque pour la sante s'accroit parallelement a sa frequence crois- sante, laquelle est suivie d'une diminution du risque pour la sante a mesure que les actions preventives font la preuve de leur efficacite. Cette transition Wld hlth statist. quart., 43 ( 1990) d'intervention ne survient pas automatiquement avec le developpement economique; elle exige des politiques et plans precis ainsi qu'un engagement communautaire. Le groupe d'age des adultes (15-59 ans) est selon toute vraisemblance particulierement expose a des risques environnementaux et professionnels qui ap- paraissent au cours de la transition sanitaire. Le taux de mortalite cumule (cumulative mortality rate - CUMRJ (CUMR (15, 59)) au la probabilite de deces entre le 15° et le 60• anniversaire, sert a mesurer !'impact sanitaire. Dans les pays developpes, le CUMR est d'environ 13% pour les hommes et 6% pour les femmes, une diminution d'un CUMR d'envi- ron 40% au debut de ce siecle. Dans les pays en developpement, ii y a des regions geographiques au le CUMR atteint 50% pour les hommes comme pour les femmes mais, au niveau national, le CUMR (15, 59) maximal dans les pays au les donnees de mortalite sont systematiquement enregistrees est d'environ 25%. Lorsqu'on analyse l'effet des risques environ- nementaux et professionnels sur cette diminution de la mortalite, on constate qu'une part tres importante de cette baisse - peut-etre un tiers - est due a la diminution de la mortalite par maladies associees a ces risques. La mortalite par traumatisme du a des accidents, par maladie infectieuse et par maladie pulmonaire, diminue considerablement au cours de la transition sanitaire. Quoiqu'il soit vraisemblable qu'une partie de cette diminution est due a !'amelio- ration du traitement medical, !'evolution des modali- tes d'apparition de differents risques et !'application d'interventions preventives efficaces jouent certaine- ment un role majeur dans la transition sanitaire. - 196 - REFERENCES - REFERENCES 1. TEITELBAUM, M. S. Relevance of demographic transition theory for developing countries. Sci- ence, 188: 420-425 (1975). 2. OLSHANSKY, s. J. & AULT, A. B. The fourth stage of the epidemiologic transition: the age of delayed degenerative diseases. The Mi/bank quarterly, 64: 355-391 (1986). 3. MCKINLAY, J. B. & MCKINLAY, S. M. The question- able contribution of medical measures to the decline of mortality in the United States in the twentieth century. Mi/bank Memorial Fund quar- terly, 55 (3): 405-428 (1977). 4. OMRAN, A. R. Forum: population planning. A cen- tury of epidemiologic transition in the United States. Preventive medicine, 6: 30-51 (1977). 5. JOSE, M. v. & BORGARO, R. Demographic and epi- demiologic transition: problems for research [in Spanish with English abstract]. Salud publica mexicana, 31: 96-205 (1989). 6. BROUDY, D. N. & MAY, P.A. Demographic and epidemiologic transition among the Navajo Indians. Social biology, 30: 1-16 (1983). 7. OMRAN, A. R. The epidemiologic transition. A theory of the epidemiology of population change. Mi/bank Memorial Fund quarterly, 49: 509-538 (1971). 8. OMRAN, A. R. The epidemiologic transition theory. A preliminary update. Journal of tropical pediatrics, 29: 305-316 (1983). 9. ROGERS, R. G. & HACKENBERG, R. Extending epi- demiologic transition theory: a new stage. Social biology, 34: 234-243 (1987). 10. LITVAK, J. ET AL. The growing noncommunicable disease burden, a challenge for the countries of the Americas. Bulletin of the Pan American Health Organization,21 (2): 156-171 (1987). 11. YOUNG, T. K. Are subarctic Indians undergoing the epidemiologic transition? Social science and medicine, 26: 659-671 (1988). 12. SMITH, K. R. The risk transition. Honolulu, En- vironment and Policy Institute, 1988. (Working paper No. 10). 13. The epidemiological transition. Lancet, ii: 670 (1977). 14. FEACHEM, R. ET AL. (EDS). The health of adults in the developing world. Washington D.C., World Bank, 1990. (In press). 15. WORLD HEAL TH 0RGANIZA TION. World health statis- tics annual. Geneva, WHO, 1989 (and earlier volumes, annually). 16. PRESTON, S. H. ET AL. Causes of death-Life tables for national populations. New York, Seminar Press, 1972. 17. PISA, Z. & UEMURA, K. Trends in mortality fr~m ischaemic heart disease and other cardio- vascular diseases in 27 countries from 1968 to 1977. World health statistics quarterly, 35 (1): 11-47 (1982). P1sA, Z. & UEMURA, K. Tendances de la mortalite par cardiopathies ischemiques et autres mala- dies cardio-vasculaires. Rapport trimestriel de statistiques sanitaires mondia/es, 35 ( 1 ): 11-47 (1982). 18. DOLL, R. & PETO, R. The causes of cancer: quanti- tative estimates of avoidable risks of cancer in the United States today. Journal of the National Cancer Institute, 66: 1194-1309 (1981). 19. WORLD HEALTH ORGANIZATION. Tobacco or health. Copenhagen, WHO Regional Office for Europe, 1988. (Smoke-free Europe: 4). ORGANISATION MONDIALE DE LA SANTE. Le tabac OU la sante. Copenhague, Bureau regional OMS de l'Europe, 1988. (Europe sans tabac: 4). 20. SEATON, A. ET AL. Crofton and Douglas' respira- tory disease. Oxford, Blackwell Scientific Pub- lications, 1989. (Fourth edition). 21. HAMMOND, E. C. ET AL. Asbestos exposure, cigar- ette smoking and death rates. Annals of the New York Academy of Sciences, 303: 473-490 (1979). 22. SARACCI, R. Interaction and synergism. American journal of epidemiology, 12: 465-466 (1980). 23. STANSFIELD, S. K. ET AL. Injury and poisoning. In: Jamison, D. T. & Mosley, W. H. (eds), Evolving health sector priorities in developing countries. Washington, D.C., World Bank, Population, Health and Nutrition Division, in preparation. 24. JEYARATNAM, J. 1984 and occupational health in developing countries. Scandinavian journal of work environment and health, 11: 229-234 (1985). 25. WORLD HEALTH ORGANIZATION/UNITED NATIONS EN- VIRONMENT PROGRAMME. Public health impact of pesticides used in agriculture. Geneva, WHO, in preparation. 26. WHO Technical Report Series No. 513, 1973 (Safe use of pesticides: twentieth report of the WHO Expert Committee on insecticides). OMS Serie de Rapports techniques N° 513, 1973 (Securite d'emploi des pesticides: vingtieme rap- port du Comite OMS d'experts). Rapp. trimest. statist. sanit. mond., 43 (1990) - 197 - SMOKING-RELATED DEATHS IN DEVELOPED COUNTRIES-AN UPDATE In the World health statistics quarterly Vol. 43 (2), page 113, estimates of smoking-attributable mor- tality in the developed countries were presented in the article by Alan D. Lopez, Who dies of what? A comparative analysis of mortality conditions in developed countries around 1987. These estimates were hypothetical mortality estimates which sug- gested what the annual death toll from smoking in developed countries would be if the smoking- attributable fractions for the United States of America were to apply in each country (the United States experience was used since the relative-risk estimates used in the calculation of smoking- attributable fractions are readily available from sev- eral large-scale epidemiological studies carried out in that country). Subsequently, the World Health Organization has developed a new methodology for calculating smok- ing-related deaths which scales the experience of the United States according to the estimated impact of previous smoking patterns in each country. The scaling factor is specific to each country and is calculated from the ratio of excess lung-cancer mor- tality compared with a non-smoking population. Based on an application of this new methodology, the annual number of smoking-attributable deaths in the developed countries in the early 1990s is esti- mated at around 1.8 million, with 790 OOO of these occurring before age 65. Women account for about one-fifth of the estimated total number of smoking- attributable deaths. Wld hlth statist. quart., 43 (1990) DECES LIES AU TABAGISME DANS LES PAYS DEVELOPPES - UNE MISE A. JOUR Dans le volume 43 (2), p. 113, du Rapport trimestriel de statistiques sanitaires mondiales, des chiffres es- timatifs de la mortalite imputable au tabagisme dans les pays developpes sont presentes dans !'article d' Alan D. Lopez, Qui meurt de quoi? Analyse com- paree de la mortalite dans les pays industrialises vers 1987. II s'agit de chiffres hypothetiques indi- quant ce que serait la mortalite annuelle due au tabagisme dans les pays developpes si les fractions de la mortalite imputables au tabagisme aux Etats- Unis d'Amerique s'appliquaient a tous les pays (on s'est fonde sur !'experience des Etats-Unis parce que les estimations du risque relatif utilisees pour le calcul des fractions de la mortalite imputables au tabagisme sont aisement disponibles dans ce pays grace a plusieurs enquetes epidemiologiques de grande envergure). Ulterieurement, !'Organisation mondiale de la Santa a mis au point une nouvelle methode pour le calcul des deces lies au tabagisme qui adapte les chiffres des Etats-Unis a l'impact estimatif de !'evolution du tabagisme dans chaque pays. Le facteur d'adaptation est specifique a chaque pays et fonction du taux de surmortalite par cancer du poumon par rapport a une population de non-fumeurs. En appliquant cette nouvelle methodologie, on es- time le nombre annuel des deces imputables au tabagisme dans les pays developpes au debut des annees 90 a environ 1,8 million, dont 790 OOO avant l'age de 65 ans. Les femmes representent environ un cinquieme du nombre estimatif total des deces im- putables au tabagisme. CORRIGENDUM 1. Natural focus Foyer natu rel - 198 - FIG. 1 YELLOW FEVER - EPIDEMIOLOGICAL CYCLE IN WEST AFRICA FIEVRE JAUNE- CYCLE EPIDEMIOLOGIQUE EN AFRIQUE OCCIDENTALE 4. Sylvatic outbreak Epidemie selvatique 7. Urban outbreak (village) Epidemie urbaine («rurale») RECT/FICAT/F 2. Isolated endemic emergence Emergences endemiques isolees 5. Intermediate outbreak Epidemie intermediaire 8. Transovarial transmission of the virus Transmission transovarienne du virus 3. Multiple endemic emergence Emergences endemiques multiples 6. Urban outbreak (town) Epidemie urbaine (en ville) 9. Post-transovarial transmission circulation Circulation post transmission transovarienne Northern limit of the yellow-fever endemic zone - Limite nerd de la zone d'endemie amarile (CA. 15' Lat. N. ) 0 ... ....-;;-··;·~·;··· ... :•-·;)' . ,., ,•: ·W :• .... ,, •• : f- •• '• ·-• ... •, • : 0 ... ..:):::.-·~ : 0 :j: \ .... •' ~.t .!: .. Relative importance of the man/wild vector contact Importance relative du contact homme/vecteurs sauvages Wild vectors which may be responsible for monkey-to- monkey transmission of the yellow-fever virus Vecteurs sauvages pouvant assurer fa transmission intersimienne du virus amaril -~ ... •" ., .·· /.:·~--·-······ . ....... __ ~ "- i---------········-./ -------------: ·-·' •. • ·.'!"'" ... ----------... -----------1 ·W : ;•,\I.~ \ " .r:: ~ !E. " V) 'iii c "' 'O :, 51 " c "' > "' (/) I "' c c ~ V) ~ j 'is ill .; '" :,a j '" c ~ U) "' c c g? ill :,a j E ill ~ :,a § I .!l! :c a. e .0 E 0 . .; 3l ~ -~ ~-§ >:, ffi ~ c .0 "'" 'O V) :,a, (/) c '"' .0 > :,n, (/)(/) ~ ,Q) §i NO ~! 0" ~§ "-N ;;; ~'" 0:, -'O ~:~ g"9 ii E'~ "0 b : ·-·-' •• 11,•. ".. w : • .~ .. ,~ • ,,~.:-~ : ·~·-· ,,, •... l .. •;!i .; ..... ·· ~ ·····-.. ~~··· <( 0 :j: w 0 a: w .. , , ... ·-- ,"""' ...... ..... .... t .. .. ---- ··, ' ······ . . . ..... _ ... __ ·W f- 0 :j: ·w 0 z w b w a: <( <( w a: <( 0 :j: w 0 z w w :::, a :j: -tu 0 z w w 0 z w ('.) a: w ::;; ····~····· © ......... © .......... :~t· -.-. ...... 4 ;.~,--· f.\f.i·;····· . ... ~~ ..... • f•~. rt~· • \ ----f"/.··..;....' f •'t C::..· \--,..----1'. .... · .......... ill~·-· ·/·-· . .,.-........ · • ®l~I \ _. \!I .... -~ I ... ···•······ t,~:© : •,, ... . 0 w z 0 N w z 0 N w 0 z w ('.) a: w ::;; w 0 :j: w 0 z w @ ... -,,,, ... • f • ••• , .... \ : '.,, .. . :·-·-· ...... , ?--• . •' ·"· : i--------- \. :-'., ;..:--!-___ _,. _________ _., .... •"" : . ······!. .. ~.: t A. luteocephalus A. furcifer A. metallicus A. opok A. africanus A. neoafricanus A. vittatus A. aegypti A. taylori A. luteocephalus A. furcifer A. taylori A. opok A. africanus A. neoafricanus A. vittatus A. meta/ficus A. aegypti A. africanus A. luteocephalus A. furcifer A. opok A. cordellieri A. flavicolfis A. taylori A. vittatus A. aegypti A. africanus A. cordellieri A. flavicollis A. vittatus A. aegypti 0 U)U- u_ ' ·······•. -!• ... • ···-~~ .. -----------'t-----------+-----------+----------1 : ....... , ...... : " ;;; c " tii~ .E ~-2' c .E ~ 'iii CE a: ~~ °'-~.,, "~ >O wu_ 1 • .,r'I~"':: i .... •. n , ... : ·--?~1r.r~~) @ ·-·, ·""' . ! © \ \ I • • '·-·"' __., Wild vectors Vecteurs sauvages Domestic vectors Vecteurs domestiques • Monkeys Singes • Man Homme A. africanus A. cordellieri A. flavicolfis R.C. Emergence front line Front d'emergence Rapp. trimest. statist. sanit. mond., 43 ( 1990) - 199 - From the World Health Organization The WHO Environmental Health Criteria Series The Environmental Health Criteria series was launched in 1976 in response to concern over the risks to human health and the environment posed by the growing number of chemicals on the market and in the environment. In planning the series, the aim was to give national authorities all the information needed to understand the specific hazards posed by a chemical and then devise appropriate protective measures, whether for the health of workers, the safety of the general public, or the survival of the environment. In keeping with this aim, volumes in the series issue authoritative conclu- sions about human and environmental risks based on a study of virtually everything ever written about a selected chemical, physical factor, or environ- mental pollutant. Substances to be evaluated are selected according to an internationally agreed list of priorites. Drafts are first prepared by individual experts or national institutions, circu- lated worldwide for comment, and then reviewed and revised by one or more international groups of experts in order to make the information as complete and accurate as possible. Virtually since the first volume a p- peared, the series has earned high praise from the scientific community for the precision and accuracy of information, the completeness of coverage, and the relevance of conclu- sions and recommendations. In addition to volumes concerned with individual chemicals, the series also publishes guides addressing key methodological problems of data interpretation, risk assessment, and carcinogenicity and toxicity testing. The majority of volumes dealing with chemical substances are the result of work performed by the International Programme on Chemical Safety (IPCS), a joint venture of the United Nations Environment Programme, the Interna- tional Labour Organisation, and the World Health Organization. Volumes assessing sources of non- ionizing radiation are prepared jointly by the International Radiation Protec- tion Association and the World Health Organization, with funding from the United Nations Environment Pro- gramme. The series also benefits from funding by a number of governmental agencies, who appreciate both the critical impor- tance of a healthy environment and the capacity of objective, validated, and complete information to contribute to that goal. World Health Organization, Distribution and Sales, 1211 Geneva 27, Switzerland - 200 - The WHO Environmental Health and Chemical Safety Subscription Package A yearly service offering automatic and immediate delivery of all WHO publications dealing with environmental health and chemical safety Introduced as a convenience for customers, the WHO Environmental Health and Chemical Safety Sub- scription Package includes all 1991 issues of the Environmental Health Criteria series and the related Health and Safety Guides plus all Order form D Please enter my 1991 subscription to the WHO Environmental Health and Chemical Safety Subscription Package at Sw.fr. 300.-/US $240.00 Name Address other serial and non-serial books focused on a topic important to environmental health, pollution, chemical toxicology, and chemical safety. The WHO Environmental Health and Chemical Safety Subscription Package 1991 subscription rate: Sw.fr. 300.-/US $240.00 (includes surface postage and handling charges) D Payment enclosed D Please charge to my credit card D Visa D American Express D Eurocard/Access/Mastercard Card no. Expiry date Date of order Signature WHO· Distribution and Sales· 1211 Geneva 27 · Switzerland PUBLICATIONS OF THE WORLD HEALTH ORGANIZATION 1990 WORLD HEAL TH FORUM An international journal of health development (Separate editions in English. French, Spani'ih. Arabic, Chinese and Rus!"-.ian) World health forum is a quarterly juumal for policy-maker:-., health planners, a<lminis- tratuf'i. health educators. and public health workers of all kind!-,. It provide!-, a medium for the presentation and discus~ion of new concepts in public health an<l new approaches to health problems, and it is devoted to the improvement of health through the promo- tion of health ~ervice:-. covering the entire population and the undertaking of a wide va- riety of public health mea~ures, whether or not they are ~upportcd by WHO. The Forum is the main organ of WHO's Member Stares for the international exchange of health in- fonnation and an active in .... trumenl in technical cooperation among developing countries.. Suhscription (4 numbers) ................................. . Price per copy ...................................... . BULLETIN OF THE WORLD HEALTH ORGANIZATION Sw. fr. 60.- Sw. fr. 18.- The Bulletin is the principal scientific organ of WHO; its role is to review progress in medical and related sciences ("update'' articles) and to bring to light new knowledge by publi~hing original papers on scientific re .... earch in the laboratory and field. The 81ttletin contains original articles in either English or French. with a !-.Ummary in the other language. One volume is published annually, con,;;isting of six numb<!rs. Sub'icription (6 numbers) ................................. . Price per copy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . WEEKLY EPIDEMIOLOGICAL RECORD (Bilingual: English and French) Sw. fr. 130.- Sw. fr. 22.- Prepared for the guidance of health administrationri. and health authoritieri., the Weekly t•t,idemioloxical n•cord contains notifications made under the International Health Regu- lation .... and infonnation concerning their application. The Record also contains epidemio- logical information on communicable disease!-. of international importance. Annual subscription ................................... . Sw. fr. 150.- WORLD HEALTH STATISTICS QUARTERLY The World health .<.tati.tticr quarterly replaces (since 1978) the World health statistics report (published since 1967) and its forerunner the Epidt•minlngh·al and \'ital .\'latistics report (published since 1947). It deals with the detailed analysis of ,elected health topics of cun-ent interest. The Quarterly contains articles in either French or English with a summary in both languages. Annual subscription ................................... . Price per copy ...................................... . WORLD HEAL TH STATISTICS ANNUAL (Bilingual: English and French) Sw. fr. 90.- Sw. fr. 24.- The forerunner of thiri. series was the Annual epideminloxical re11ort nf the League of Nations. It was followed by the Annual e11idemiological and \'ital statistics issued by the World Health Organization. Latest puhl ication .... : 198). Vit.tl .... tatistics and life tables, environmental health, causes of death (single volume), 531 page,. . . . . . . . . . . . . . . . . . . . . . . . Sw. fr. 80.- 1986. Vital statistics and life tables, evaluation of the Global Strategy for Health for All, causes of death (single volume), 692 pages .... 1987, Vital .... tatbtics and lite tables, oral health and care, causes of death (single volume), 455 pages .•...................... 1988, Vital statistic, and life tables, the health professions in the 1980s, causes of death (single volume), 513 pages ................ . 1989, Vital statistics and life tables. the WHO MONICA Project, causes of death (single volume), 443 pages .................... . Sw. fr. 90.- Sw. fr. 90.- Sw. fr. 90.- Sw. fr. 90.- PUBLICATIONS DE L'ORGANISATION MONDIALE DE LA SANTE 1990 FORUM MONDIAL DE LA SANTI<~ Revue internationale de developpement sanitaire (Editions <iC!parCcs en fran~ai.'., anglais, e~pagnol, arabe. chinois et rus~e) Forum mondial de la sum( est une revue trimestriellc destinee aux re"!lpon'iables des politiques .... anitaires, aux planificatcur<i, administrateurs et educateurs \anitaires, cnfin aux travailleurs de la \ante publiquc de toutcs caregorie ..... Trihune pour la presentation et la discus!-.ion de nouveaux concepts en santC publique et de nouvelle"i approches des pro- blemes de 'iantC. Fnrum sc cons.acre a !'amelioration de la ... ancC par la promotion de serviceri. de santC couvrant la population tout entiCre et c.l'une vastc gamme de mesures de .... ante publique, qu'cllcs soicnt ou non sourenue.'. par !'OMS. II est le principal organe a la disposition des Etars Mcmbres de l'OMS pour l'Cchange international d'infom1a- tion~ sanitaires en mC:mc temps qu'un insLrument de:: cooperation technique entre pays en developpemem. Abonnement (4 numerm,) ............................... . Le numero ....................................... . BULLETIN Fr. s. 60.- Fr. s. 18.- DE L'ORGANISATION MONDIALE DE LA SANTE Le Bulletin est le principal organe scientifiquc de I 'OMS; ii a pour rOle <le pa .... ser en revue les progrC'i de~ sciences mCdicales et apparcntCcs (article5. de la rubrique (<Le point») et de mcttrc en lumierc les connaissance~ nouvelle" en prCsentant dans des arti- cles originaux lcs rCsultats de recherches scientifique...,, au laboratoire et sur le terrain. Le Bulletin conticnt des articles originaux en fran<;ais ou en anglais, accompagnCs d'un r<!sum.e dans l'autre languc. Un volume est publie annuellement, Ctant forme de .'.ix numCro!). Abonnement (6 numeros) ............................... . Le numero ....................................... . RELEVE EPIDEMIOLOGIQUE HEBDOMADAIRE (Bilingue: fran,ais et anglais) Fr.,. 130.- Fr. s. 22.- Destine aux administrations sanitaires et aux services de santC, le Rele\•f !(JidtmioloRi- que hehdomadairl' contient le5. notilications exigCes par le RCglement sanitaire interna- tional. ainsi quc d'autrcs renseignements concemant !'application de cc rCglement. Le Relel'i contient Cgalcment des informations epidCmiologique'i concernant les maladies transmissibles d 'unc imponance internationale. Prix de I 'abonnement annuel .............................. . Fr. s. 150.- RAPPORT TRIMESTRIEL DE STATISTIQUES SANITAIRES MONDIALES Le Rapport trimt•striel de sratistiqul's sanitain•.\ mondiales, remplace (depuis 1978) le Rapport di' statistiques .wnitaires mm1dialt'.\' (publiC depuis l 96 7) et son pr6eurseur le Rapport epidemiolo!lique ,•t demoKraphiquc (publie depuis 1947). II presente de, analyr.,.es dCtaillCes sur des sujets !-.J)eCifique .... d'intCrCt courant. Le Trimeslril'I pre .... ente de!) articles originaux en fran1rais ou en anglais, accompagnCs d 'un re .... umC dans les deux langues. Prix de l'abonnement annucl Le numero ....................................... . ANNUAIRE DE ST A TISTIQUES SANIT AIRES MONDIALES (Bilingue: fram;ais et anglai"ll) Fr.,. 90.- Fr. s. 24.- Cet annuaire rcmplace lcs Statisriqm•s tpidemio/ogique.\· et d(!mographiques annuel/e.\· publiCes par I 'Organisation mondiale de la SantC et qui avaient, elles-mC:mes, remplace le Rapport t'pidimiologique annuel public par la SociCtC des Nations. Publications reccntes: 1985, Mouvement de la population er tables de survie, salubrltl! de l'en- vironnement, cau .... es de dCcCri. ( I \eul volum.e). 531 pages. . . . . . . Fr. s. 80.- 1986, Mouvement de la population et table'i de survie, evaluation de la Srrat6gie mondiale de la .... ante pour tau!), cause"i de dCces ( I seul volume). 692 pages . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fr. s. 90.- (987. Mouvemcnt de la population er tables de survie, .... ante et soins bucco-dentaires, causes de deces ( 1 .... eul volume), 455 pages . . . . . Fr. s. 90.- 1988. Mouvemcnt de la population et table5. de survie, la situation de" professions de sance dans le5. annees 80, cause~ de dCcCs ( I seul volume), 513 pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fr. s. 90.- 198(}, Mouvemcnt de la population et tables de survie, le Projet MONICA de rOMS. causes de deces ( I seul volume). 443 pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fr. s. 90. - OMS ENVIRONMENT AL EPIDEMIOLOGY The growing concern ahout environmental pollution at the local, regional and global levels has incrca,ed the demand for information about the consequences of such pollution both from government agencies and the community at large. The consequences for the natural environment, agriculture and other economic activities need to be quantified in order to assess the direct economic impact of pollution. But the effects on human health and wellbeing may be even more important, as the ultimate goal of economic activities should be to improve life on earth. In their ultimate extension, some of the global pollution problems threaten the very survival of living species. including the human race. In a more immediate perspective, every year millions of people get ill, and may even die, from exposure to environmental health hazards. Environmental epidemiology is the science that helps us measure, interpret and forecast the effects of environmental hazards on health. It analyses the occu1Tence of good health, ill health and death in population groups in relation to the various environmental factors, be they of biological, chemical or physical origin. This science not only describes the situation in different communities, countries or continents, but it is also essential for identifying the ways in which people can be protected from environmental adverse health effects. The World Health Organization is committed to guide, encourage and facilitate environmental epidemiology activities in Member States and uses this science to analyse the public health impact of environmental hazards, at global and regional levels. This issue of the World health statistics quarterlr reviews the public health impact of several major environmental health problems. Examples from specific countries are given and some of the information is analysed from a regional perspective, but the overall theme is global. the objective being to promote a healthy planet. The topics cover different types of air pollution, water pollution, hazardous chemicals and the development process itself. It is seen that with economic development there is generally an improvement of health in the long term. a, reflected by the health transition. Much of this health improvement is due to the elimination nr reduction of environmental hazards exposures, but development may also lead to new hazards, such as pesticides and lead in petrol, which environmental epidemiology can identify and quantify as a step towards health protection. Millions of people are e.xpose<l to some of the hazards reviewed: for indoor air pollution it may be several hundred million, and even more for outdoor air pollution. Severe acute pesticide poisoning may occur in 3 million people each year and attacks of poisoning symptoms may affect as many as 25 million each year. The review, highlight the severity of the situation, both for the traditional hazards due to poor sanitation and lack of safe drinking-water and food, and for the modern hazards such as industrial pollution and use of toxic chemicals. A general conclusion from these reviews is the need for accurate and timely environrhental epidemiology data at the local level to assist public health authorities in developing environmental health policies and to guide effective preventive actions. There is also a need to improve our epidemiological knowledge at the global level, and it is hoped that this issue will encourage further activities in this field. L'ECO-EPIDEMIOLOGIE La pollution de l 'environnement suscitant une inquietude croissante aux niveaux local, regional et mondial, les organismes publics et la population dans son ensemble demandent a etre de mieux en mieux informes sur ses consequences. II faut les quantifier, ces consequences, pour le milieu nature!, !'agriculture et <l'autres activites economiques, afin d'en evaluer I 'impact economique direct. Cepen<lant, les effets de la pollution sur la sante et le bien-etre des etres humains sont peut-etre plus importants encore, puisque les activites economiques cloivent avoir pour finalite ]'amelioration de la vie sur la Terre. A leur stade ultime, certains des problemes de pollution clans le moncle menacent la survie meme des especes vivantes, sans epargner la race humaine. Dans I' immediat, chaque annee des millions de personnes tom bent malades ou meme meurent a la suite de I' exposition a des dangers d 'orclre environnemental. L 'epidemiologie environnementalc est la science qui nous aide a mesurer, a interpreter et a prevoir les effets des risques lies a l'environnement sur la sante. Elle vise a analyser la situation sanitaire - bonne sante, mauvaise sante, mortalite - de groupes de population en fonction de divers facteurs environnementaux, qu'ils soient d'origine biologiquc, chimique ou physique. Mais eette science ne decrit pas seulement la situation dans clifferents lieux, pays ou continents, elle est egalement essentielle pour identifier les moyens de proteger les gens contre des conditions ecologiques nefastes pour leur sante. L'Organisation moncliale de la Sante a pour mission de guider, <.!'encourager et de faciliter les activites d'epiclemiologie environnementale <lans les Etats Membres et fait appel a cette science pour analyser !'impact sur la sante publique des dangers lies a I 'environnement, aux niveaux regional et mondial. On trouvera dans le present numero du Rapport trimestriel de statistiques sanitaire.1· mondiales des etudes de !'impact sur la sante publique de plusieurs grands problemes environnementaux. Des exemples concemant <lifferents pays sont cites et ccrtaines clonnees soot analysees dans une perspective regionalc, mais le theme general est universe!, l 'objectif etant la promotion de la salubrite de la planete tout entiere. Les sujets traites portent sur clifferents types de pollution: pollution de I 'air, pollution de l'eau, substances chimiques dangereuses, processus de developpement par lui-meme. On constate que le developpement economique s'accompitgne generalement d'une amelioration de la sante a long terme, comme le montre la «transition sanitaire». Cette amelioration est due en grande partie a !'elimination ou a la reduction de !'exposition a des clangers ecologiques, mais le developpement peut aussi crecr de nouveaux clangers, comme les pesticides ou le plomb dans I 'essence, quc l 'epiclemiologie environnementale peut identifier et quantifier afin de mieux proteger la sante. Des millions de gens sont exposes a certains des dangers evoques: c'est ainsi que le chiffre cloit etre de plusieurs centaincs de millions pour la pollution de l'air a l'interieur, et encore plus pour la pollution atmospherique exterieure. Les intoxications aigues clues aux pesticides toucheraient 3 millions de personnes chaque annee tandis que 25 millions - nombre considerable - presenteraient a un moment ou a un autre des symptomes d 'intoxication. Les articles mettent en evidence la gravite de la situation tant pour les dangers traclitionnels dus a de mauvaises conditions cl'hygiene et au manque d'eau potable et de nourriture que pour les clangers modernes tels que la pollution industrielle et I 'utilisation de substances chimiqucs toxiqucs. La conclusion generate qui en est tiree est qu 'on a besoin de clonnees d'epiclemiologie environnementale exactes et recentes au niveau local pour aider les responsables de la sante publique a elaborcr une politique de salubrite de I 'environncmcnt et orienter des actions preventives efficaces. Les connaissanccs epi<lemiologiques doivent etre ameliorees au niveau mon<lial et ii faut csperer que ce numero favoriscra la poursuite des activites dans ce domaine.