Docstoc

Mise en page 1

Document Sample
Mise en page 1 Powered By Docstoc
					ABSTRACT BOOK - Table of Contents

Guidelines to read the abstracts numbers
The code indicates the day and time of the day a session is scheduled.
• The first letter corresponds to the day of the session (Sunday, Monday, Tuesday or Wednesday).
• The second and/or third letter: S is for Symposium, M is for morning, AA is for the beginning of the afternoon and AB is for
  the end of the afternoon.
• Finally, the number corresponds to the order of the session.
After the dash, the letter O indicates the oral communications, PD indicates poster discussion. The number gives the order of the
abstracts.
Example: SS2-43 means Sunday Symposium 2 – old reference 43
            TM6-O means Tuesday Morning session number 6 – oral communications
            TAB2-PD means Tuesday end of the afternoon session number 2 – Poster Discussion

1) Sunday September 3rd
    Plenary Session
       SPL - Plenary :Early ages exposures to environmental hazard........................................................................................xx
    Symposia session
       SS1       - Regional, Multi-City Time-Series Studies of Air Pollution and Health: Progress and Prospects..................xx
       SS2       - Measuring in-Utero Exposures in Longitudinal Children Studies: What, Why, How and When ..................xx
       SS3       - Biomarkers in human studies and relevance to disease outcomes............................................................xx
       SS4       - Intake fraction: Past, Present, and Future ..................................................................................................xx
       SS5       - Biological Monitoring to Assess Exposure to Biological and Chemical Warfare Agents ............................xx
       SS6       - Modelling Children's Exposures to PM, Ozone and Air Toxics ...................................................................xx
       SS7       - Human Tracking Technologies for Exposure Analysis .................................................................................xx
    Oral Communication Session
       SM3-O - Children's health and environmental chemicals ........................................................................................xx
       SAA1-O - Cancer and the Environment .....................................................................................................................xx
       SAA2-O - Ethical issues in environmental health: violence and precautions .............................................................xx
       SAA5-O - Air pollution and pregnancy outcomes......................................................................................................xx
       SAB1-O - Ecosystem and Human Health ...................................................................................................................xx
       SAB2-O - Biomolecular methods (I)...........................................................................................................................xx
       SAB5-O - Children's health, air pollution and noise ..................................................................................................xx
    Poster Discussion Session
       SM1-PD - Pesticides and Persistant Organic Pollutants .............................................................................................xx
       SM2-PD - Environmental equity (I) ............................................................................................................................xx
       SM4-PD - Air pollution and reproduction, health of children and other vulnerable groups.......................................xx
       SM5-PD - Spatial analysis in environmental health ...................................................................................................xx
       SAA3-PD - Heavy metals .............................................................................................................................................xx
       SAA4-PD - Health and meteorology ............................................................................................................................xx
       SAB3-PD - Dietary exposures ......................................................................................................................................xx
       SAB4-PD - Air pollution: Mechanisms and health .......................................................................................................xx
    Poster Session
       Run 1: Sunday 3rd and Monday 4th ...............................................................................................................................xx

2) Monday September 4th
    Plenary Session
       MPL - AWARDS PLENARY ...............................................................................................................................................xx
    Symposia session
       MS1     - Childhood exposures to bioavailable metals in soil and household dust in residential environments ......xx
       MS2     - Ethical Issues in Reanalysis of Data from Policy-Relevant Studies.............................................................xx
       MS3     - RESPIRE - The Guatemala Randomized Intervention Trial..........................................................................xx
       MS4     - Uncertainty, complexity and precaution in environmental health..............................................................xx
       MS5     - Characterizing the Risks of Climate change ..............................................................................................xx
       MS6     - Assessing exposures to Chemicals in consumer Products: Global Policy and Emerging Sciences Issues ...xx
       MS7     - The Current Advances in Exposure and Health Effects Assessment of Traffic Exhaust
                  particulate air pollution .............................................................................................................................xx
       MS8     - Population Exposure Modelling and Urban Air Quality Management in Europe .......................................xx
    Oral Communication Session
       MM1-O - Gene-environment Interactions .................................................................................................................xx
       MM2-O - Risk assessment methodologies ................................................................................................................xx
       MM3-O - Electromagnetic Fields ...............................................................................................................................xx
       MAA1-O - Temperature-related short term health effects ..........................................................................................xx
       MAA3-O - Occupational health (I) ..............................................................................................................................xx
       MAB4-O - Exposure assessment.................................................................................................................................xx
ABSTRACT BOOK - Table of Contents

  Poster Discussion Session
    MM4-PD - Air pollution sources, exposure and health................................................................................................xx
    MAA2-PD - Children's health and the environment .....................................................................................................xx
    MAA4-PD - Biomolecular methods (II)..........................................................................................................................xx
    MAB1-PD - Environmental health indicators ................................................................................................................xx
    MAB2-PD - Chemicals in consumer products ...............................................................................................................xx
    MAB3-PD - Occupational health (II) .............................................................................................................................xx

3) Tuesday September 5th
  Plenary Session
     TPL      - Plenary :Environment and Genetics...........................................................................................................xx
  Symposia session
     TS1      - Exposure scenario development and exposure assessment in the context of the REACH legislation........xx
     TS2      - Contribution of epidemiology for risk assessment after environmental Ionising radiation exposure ........xx
     TS3      - Longitudinal Studies of Particulate Matter-Related Mortality and Morbidity ............................................xx
     TS4      - Health effects from exposure to radio frequency electromagnetic fields from mobile telephony..............xx
     TS5      - Human Biomonitoring (HBM) as a key tool in environment and health....................................................xx
     TS6      - Dietary carcinogens: epidemiology and exposure assessment...................................................................xx
     TS7      - Risk assessment tools and decision making process for a cleaner air: the example of .............................xx
     TS8      - From Avian to Human Influenza ................................................................................................................xx
  Oral Communication Session
     TM1-O - Hormonally active compounds and reproductive health............................................................................xx
     TM2-O - Society and the environment .....................................................................................................................xx
     TM3-O - Air pollution: PM exposure : sources and health effects............................................................................xx
     TM4-O - Climate Change and Human Health ..........................................................................................................xx
     TM5-O - Environmental Equity (I) ............................................................................................................................xx
     TM6-O - Air pollution: Short term cardio-vascular effects........................................................................................xx
     TAA1-O - Indoor air (I) ..............................................................................................................................................xx
     TAA2-O - Water (I).....................................................................................................................................................xx
     TAA3-O - Air pollution: Health effects of PM and ozone exposure ...........................................................................xx
     TAB1-O - Novel study designs in environmental epidemiology ................................................................................xx
  Poster Discussion Session
     TAB2-PD - Indoor air (II) .............................................................................................................................................xx
     TAB3-PD - Asthma and allergies (II) ...........................................................................................................................xx
     TAB4-PD - Water (II)....................................................................................................................................................xx
  Poster Session
     Run 2: Tuesday 5th and Wednesday 6th .........................................................................................................................xx

4) Wednesday September 6th
  Plenary Session
     WPL - PLENARY / PANEL: Meeting the challenges of environmental Health ..................................................................xx
  Symposia session
     WS1      - The Dioxin exposure and human health 30 years after the Seveso, Italy accident ....................................xx
     WS2      - Short term effects of weather....................................................................................................................xx
     WS3      - Housing and Health: Emerging concerns associated with residential exposure to chemicals....................xx
  Oral Communication Session
     WM1-O - Asthma & allergies (I)................................................................................................................................xx
     WM2-O - Air pollution long term effects...................................................................................................................xx
  Poster Discussion Session
     Wednesday September 6th .............................................................................................................................................xx
     WM3-PD - Exposure modeling ....................................................................................................................................xx
SUNDAY SEPTEMBER 3
                                                                                                                                                                                                      SUNDAY SEPTEMBER 3
        Plenary Session
        Symposia Session:

SPL1           ON THE CONTINUUM OF EXPOSURES AFFECTING CHILDREN’S HEALTH
               – DEFINING A CONCEPTUAL PARADIGM FOR THE EARLIEST CRITICAL WINDOWS
Germaine M. Buck Louis, PhD, MS
Childhood is marked by a period of rapid development and underscores the need for researchers to be cognizant of the highly interrelated and time-dependent processes underlying successful
human development and homeostasis. Without doubt, children, broadly defined, are often more adversely affected by environmental exposures than adults, indicative or their inherent biologic
susceptibility and vulnerability arising from exposure pathways that are atypical of adult behavior. For example, children’s physiology differs from adults in that they have faster inhalation
rates, higher consumption of food and liquids per body weight, a larger body surface to weight ratio as well as a host of other differences. The World Health Organization estimates that
children represent approximately one-third of the world’s population and that approximately 30% or more of the global burden of disease can be attributed to environmental factors. Despite
children’s at risk status for many environmental exposures, risk assessment paradigms typically do not capture the entire life span, underscoring the need for efforts to ensure the safety of
all.
To aid in understanding the relation between environmental agents and children’s health status, several authors have developed conceptual paradigms for grounding research and to aid in
the interpretation of results. Critical windows have been defined in relation to varying operational definitions of childhood such as chronologic time denoting age (inclusive of gestational age),
in relation to school placement or other legal statues and, more recently, with regard to developmental stages. As yet, no universally accepted critical windows paradigm exists, though most
epidemiologists recognize the presence of three broadly defined domains – conception, pregnancy and following birth. The latter two domains often are referred to as the prenatal or postnatal
periods. Each of these three critical windows has been operationalized somewhat differently by authors. While there is a large amount of epidemiologic evidence regarding critical windows
for in utero exposures and structural birth defects, timed exposures in relation to functional endpoints at birth, during childhood or beyond have received limited systematic study, underscoring
one of many important critical data gaps for population-based epidemiologic research.
Growing interest in the so-called fetal origins of disease hypothesis (or more generically, the developmental origin of health and disease) during the past few decades has facilitated recognition
and appreciation of the importance of critical windows in grounding epidemiologic research to derive answers to questions about the impact of environmental factors on children’s health.
Reproductive epidemiologic research has continued to generate a body of evidence, albeit limited in scope, supporting the importance of the peri-conception interval as the earliest critical
window for human development, one that is often overlooked. For some time, this oversight was a reflection of limited tools for capturing exposures in the absence of clinically recognized
pregnancies coupled with limited statistical models and software for the widely known clustering of pregnancy outcomes and the longitudinal measurement of exposures. Fortunately, such
methodologic limitations no longer preclude researchers’ ability to capture the peri-conception critical window. The absence of a biomarker for human conception prior to implantation has
challenged epidemiologists in the design and conduct of research. However, this limitation is being overcome by the commercially available technologies for timing data collection to the
menstrual cycle and, hence, to the time of ovulation and estimated conception. Thus, epidemiologists are not impeded by the absence of methodologies for population based research capturing
the peri-conception window. In fact, this is an era of extreme excitement for designing research responsive to critical data gaps so that lingering questions may be addressed aided by the
continual development and refinement of research methodologies including the use of the home for collection of data and biospecimens.
So, how might epidemiology proceed in answering questions about the impact of the environment on child health? Research aimed at estimating health risks in relation to environmental
exposures requires prospective epidemiologic designs with longitudinal capture of data and biospecimens for classification of exposure, disease and even susceptibility. This approach allows
a spectrum of reproductive and developmental outcomes to be assessed minimizing the risk of missing an adverse effect on one or more endpoints including those that might reflect subtle
changes in function. The ability of such research to obtain valid and reliable data on exposure and a spectrum of health outcomes is dependent upon a conceptual framework that recognizes
and accommodates the many complexities and eccentricities of human development. In assessing the spectrum of human developmental outcomes, epidemiologists are increasingly able to
address competing risks for outcomes conditional on birth, such as embryonic or fetal mortality. This strategy is consistent with the timed dependent nature of human development and
provides a more complete understanding of study findings conditional on a pregnancy resulting in a birth.
This talk presents an overview of a critical windows conceptual paradigm that offers a framework for delineating parental exposures and time-varying exposures, including the use of novel
technologies suitable for population based epidemiologic research. This talk will emphasize the peri-conception critical window using the Longitudinal Investigation of Fertility and the
Environment Study for illustration.




SPL2            EARLY LIFE EXPOSURES AND CHRONIC DISEASE RISK IN LATER LIFE
Davey-Smith G, University of Bristol Department of Social Medicine, Bristol, United Kingdom
Exposures acting during the pre-natal period and during infancy and childhood are increasingly being identified as potential causes of chronic disease in late adulthood. This presentation will
review evidence from a range of study designs that implicate nutritional factors, infections and other environmental exposures acting during early-life sensitive periods in generating increased
risk of metabolic and cardiovascular disorders in later life. The problems of establishing causality when outcomes are related to exposures acting across many years, with the possibilities of
confounding, bias and reverse causation being manifold, will be discussed, as will novel study designs for establishing causality. Examples will be drawn from studies of early-life origins of
cardiovascular disease, such as those relating exposures during the intrauterine period or during the early post-natal period with blood pressure, insulin resistance and cardiovascular disease
in later life. Situations in which similar exposures acting at different stages of development have different long term consequences will be addressed. The notion that predictive adaptive
responses may underlie some of the associations between early-life exposures and later-life health outcomes will be discussed.




                                                                                                5
                                                                                                                                                                                                         SUNDAY SEPTEMBER 3
         Symposia Session: SS1

SS1            REGIONAL, MULTI CITY TIME SERIES STUDIES OF AIR POLLUTION AND HEALTH:
               PROGRESS AND PROSPECTS
Symposium Organisers: A Cohen (Health Effects Institute), HR Anderson (St Georges Hospital Medical School/U London)

SS1-01 WORLD          WIDE SPATIAL VARIATION IN DAILY MORTALITY RELATIVE RATES DUE TO
               SHORT TERM EXPOSURE TO AIR POLLUTION: A META ANALYSIS
R Atkinson (St Georges, University of London), AJ Cohen (Health Effects Institute, Boston), JC Carrington (St Georges, University of London), HR
Anderson (St Georges, University of London)
Hundreds of time-series studies of the health effects of short-term exposure to air pollution have now been conducted worldwide. Collectively, these studies provide compelling evidence of
the adverse effects of short-term exposure, but also pose problems of interpretation due to variation in analytic methods and reporting and the possibility of publication and analytic bias.
Heterogeneity in effect estimates may also be due to spatial variations arising from differences in meteorological and other environmental conditions, health care systems and other spatially
related factors. A number of recent meta-analyses of published results have summarized this literature, but few have focused on global patterns in the relative rates of mortality and morbidity
and evidence as to the causes of their spatial variation. The Air Pollution Epidemiology Database (APED) from St. George’s, University of London, is a comprehensive database of epidemiologic
studies of the effects of short-term exposure to air pollution published in the peer-reviewed literature. It contains effect estimates from 374 time-series studies published worldwide (141
Europe, 148 North America, 32 Latin America, 53 Asia and the Western Pacific) and descriptive data concerning the locales in which these studies have been conducted. Data from APED will
be used to investigate some of the sources of heterogeneity in the effect estimates including spatially varying factors, analytic methods used, etc. The data will also be used to illustrate some
of the issues arising from such meta-analyses of the published literature, including publication bias and other study-specific limitations. Finally the advantages and disadvantages of individual
studies versus multi-city studies will be discussed.




SS1-02 APHEA                    PROJECT. AIR POLLUTION AND HEALTH: A EUROPEAN APPROACH
K. Katsouyanni (Coordinator Dept. of Hygiene and Epidemiology, University of Athens Medical School, Greece) on behalf of the APHEA group
The APHEA project started in 1993. Its main objective was to investigate the short-term effects of air pollution on health establishing a European network of scientists, based on a large data
base analyzed in a uniform, updated and standardized way. Eleven research groups participated in the first phase of the project, which resulted in clear estimates for the effects of particles
and other pollutants. The results were used for regulatory purposes and received the attention of the press. A new contract with the E.C. led to an enlargement of the network to 22 research
groups from 20 countries and the data base to 32 cities for which daily data on several health outcomes, pollutants concentrations and information on confounders and potential effect
modifiers were recorded. The APHEA project was also supported by 3 accompanying measures contracts. Through these additional contracts, aspects related to methodology, policy and
international collaborations were addressed.
The results of all stages of the APHEA project have been published in several papers in peer-reviewed journals (and cited for more than 1500 times). They concern many aspects of the short-
term effects of ambient particles and gaseous pollutants (ozone, nitrogen dioxide and sulphur dioxide): the effects of different lags, dose-response curves, the mortality displacement issue,
the independent effects of each pollutant, effect modification patterns and methodological issues.
For example, it was found that when PM10 concentrations increase by 10µg/m3 the daily number of deaths from all causes increases by 0.52%; and that of deaths from cardiovascular and
respiratory causes by 0.76% and 0.71% respectively (all figures are statistically significant). The effects are larger on the first and second day for total and cardiovascular mortality, but persist
for longer, whilst for respiratory mortality there are more pronounced and prolonged lagged effects. The dose-response curve does not deviate from linearity within the range of observed
concentrations. A pattern of effect modification has been observed: the particles effects are larger in cities with higher NO2 concentrations, in warmer climates and in cities with larger
proportion of elderly. Effects of ozone on daily mortality were observed only during the warm season. It was found that when 8-hour ozone concentrations increase by 10µg/m3, the total daily
number of deaths increases by 0.22% and the number of respiratory deaths increases by 0.70%.
The APHEA project results have been used for HIA by other European projects and investigators and for regulatory purposes by the U.S. E.P.A. and the E.C.




SS1-03 RECENT          DEVELOPMENTS OF THE NATIONAL MORBIDITY MORTALITY AIR POLLUTION
               STUDY: 1987-2000
Dominici F, Peng RD, Zeger SL, Samet JM, Johns Hopkins University Bloomberg School of Public Health
Multi-site time series studies of particulate matter (PM) and health have provided evidence that daily variation in air pollution levels is associated with daily variation in mortality and morbidity
counts. Key areas of research concern: 1) assessing uncertainties in the adjustment for potential confounders and in the selection of the statistical model; 2) quantifying spatial and seasonal
variation of PM10 health effects to understand toxicity of various PM components; and 3) developing a computational framework for reproducible research.
To address the above questions we: 1) explore statistical properties of semiparametric regressions to adjust for confounding bias; 2) extend the NMMAPS hierarchical semiparametric Poisson
regression models to allow for seasonally and spatially varying PM effects; and 3) introducing a R software package (NMMAPS-R add web-site) which implements the statistical methods
described above and serves as a platform for conducting systematic and reproducible research.
We apply methods to the National Morbidity, Mortality, and Air Pollution Study (NMMAPS), a national study covering 100 U.S. cities for the years 1987-2000 for PM10 and for the years 1999-
2000 for PM2.5.
We found strong evidence that lag 1 exposure to PM10 continues to be associated with all-cause and cardiorespiratory mortality, with the greatest effect in the Eastern U.S. Under the NMMAPS
basic model, a 10 increase in PM10 at lag 1 is associated with a 0.19 and 0.24 percent increase in all-cause mortality and cardiorespiratory mortality (95% PI 0.10,0.28 and 0.13,0.36,
respectively). In addition a 10 increase in PM2.5 at lag 1 is associated with 0.29 and 0.38 percent increases in all-cause mortality and cardiorespiratory mortality (95% PI 0.01,0.57 and -
0.01,0.82, respectively).
These findings provided key epidemiological evidence for the EPA’s review of the U.S. National Ambient Air Quality Standards for particulate matter.




                                                                                                  6
                                                                                                                                                                                                      SUNDAY SEPTEMBER 3
        Symposia Session: SS1

SS1-04 AIR        POLLUTION AND HEALTH: A COMBINED EUROPEAN AND NORTH AMERICAN
               APPROACH (APHENA)
Samet JM (Johns Hopkins Bloomberg SPH), Katsouyanni K (U Athens Medical School), Principal Investigators for the APHENA Team
The project brings together the investigators who have carried out the Air Pollution and Health: A European Approach (APHEA 1 and 2) studies, the National Morbidity and Mortality Study
(NMMAPS) in the United States, and national studies in Canada. The findings of these studies have substantial implications for public health policies directed at controlling health effects of
air pollution; each has shown adverse effects of air pollution on morbidity and mortality at current concentrations of air pollution in Europe and North America. A principal objective of APHENA
is to characterize heterogeneity in the effects of particulate matter on mortality and morbidity across the included cities. The principal goals include 1) development of core models for the
first-stage (within-city) analyses of time-series data on mortality and hospitalization for the APHENA cities; 2) development of harmonized data bases on potential modifying factors across
the cities; and 3) parallel and combined analyses of data on air pollution and mortality and morbidity, with exploration of heterogeneity in the effect of particulate matter and its determinants.
There have now been multiple meetings and exchanges to develop a protocol for first-stage (within-city) analyses, to review findings of these analyses, and to carry out second-stage analyses.
The approach to the first-stage analyses is based on initial simulation studies and extensive exploratory and sensitivity analyses. The decisions taken specified the outcome categories,
pollutants, core models, and sensitivity analyses. Additionally, a set of potential effect modifiers for the second-stage analysis was identified. The first-stage analyses have been completed for
mortality and hospitalization and second-stage analyses are being completed for mortality. The results provide estimates for exposure to PM10 and ozone and outcomes including the daily
total, CVD and respiratory numbers of deaths by age categories. Several models have been applied with varying degrees of freedom and smoothers and results compared. For PM10, estimates
are generally similar in the United States and Europe; for ozone, estimates tend to be higher in the U.S. compared with Europe and the highest in Canada. The second-stage analyses tests
whether factors representing pollution and population characteristics explain heterogeneity of first-stage estimates.




SS1-05 EFFECTS OF AMBIENT PARTICULATE MATTER ON MORTALITY IN EUROPEAN AND
NORTH AMERICAN CITIES: AN ANALYSIS WITHIN THE APHENA PROJECT
E Samoli (Dept. of Hygiene and Epidemiology, University of Athens Medical School, Greece) on behalf of the APHENA group
Introduction: APHENA is a collaborative effort joining the multi-city European APHEA and the U.S. NMMAPS projects, as well as independent Canadian investigators. The effects of ambient
particulate matter (PM10) were studied in 90 U.S. cities (16 with daily measurements and 74 with one out of six days measurements), with population from 250,000 to 9 million; 22 European
cities (with daily measurements), population from 200,000 to 7 million; and 12 Canadian cities (with one out of six days measurements), population from 200,000 to 20 million. The daily
number of total, respiratory and cardiovascular deaths was studied for all ages and among those above and under 75 years old.
Methods: The analysis was done in two stages. The first stage concerns city-specific data sets and involves extensive exploratory and sensitivity analyses using Poisson regression models. In
order to control for several time-varying confounders we used both natural and penalized splines as smoothers. We also applied models using varying degrees of freedom for seasonality
control. The second stage involves combining the effect estimates obtained from first stage and attempting to explain heterogeneity by predefined potential effect modifiers.
Results: Results for all studied outcomes and from models using varying methods for seasonality control will be presented. Thus for example, the daily increase of cardiovascular deaths
associated with 10µg/m3 increase in PM10 concentrations using 8 degrees of freedom per year and natural splines, in cities with daily measurements, was 0.31% (95% CI: 0.04, 0.59) for
European cities and 0.30 % (-0.18 , 0.78) for U.S. cities. The effects estimated for Canadian cities were generally higher.
Conclusions: In summary, we found an increase in mortality (in most cases statistically significant) with increasing PM10 concentrations. Although the estimates in Europe and the U.S. were
similar, in Canada they were considerably higher. The effects were higher in the elderly compared to those under 75 years old. Higher effects were observed in cardiovascular mortality. Effect
modification patterns between the two continents showed differences with environmental and climatic variables but consistency with socio-economic status indicators.




SS1-06 EFFECTS OF AMBIENT OZONE ON MORTALITY IN EUROPEAN AND NORTH
AMERICAN CITIES: AN ANALYSIS WITHIN THE APHENA PROJECT
Roger Peng (Johns Hopkins Bloomberg School of Public Health) on behalf of the APHENA Investigators
The effects of ambient ozone were studied in 90 U.S. cities (50 with year-round daily measurements and 88 with at least summer-only measurements), with population from 250,000 to 9
million; 22 European cities (with daily measurements, population from 200,000 to 7 million); and 12 Canadian cities (with daily measurements), population from 200,000 to 20 million. The
daily numbers of total, respiratory and cardiovascular deaths was studied as the outcome for all ages and among those > 75 years old. The analysis was done in 2 stages. The first stage
concerns city-specific data sets and involves extensive exploratory and sensitivity analyses. The second stage involves combining the effect estimates and attempting to explain heterogeneity
by predefined potential effect modifiers. Results from core models and sensitivity analysis will be presented.
The sensitivity analysis includes various methods for smoothing seasonality and different choices of degrees of freedom. Thus for example, the daily increase of all non-accidental deaths
associated with 10 µg/m3 increase in ozone concentration using 8 degrees of freedom per year and natural splines, in cities with year-round measurements, was 0,16% (95% CI: 0,02, 0,30)
for European cities and 0,20 % (0,11 , 0,28) for US cities. The effects estimated for Canadian cities were generally higher. Effect modification patterns between the two continents showed
differences with environmental and climatic variables, but consistency with SES indicators.




                                                                                                7
        Symposia Session: SS1

SS1-07 PUBLIC HEALTH AND AIR POLLUTION IN ASIA (PAPA): A MULTI-CITY STUDY FOR
SHORT TERM EFFECTS OF AIR POLLUTION ON MORTALITY
CM Wong, on behalf of PAPA groups (members):
Bangkok (N Vichit-Vadakan1, B Ostro4, N Vajanapoom1, W Aekplakorn2, and S Wangwongwatana3)
  1
    Thammasat University; 2Mahidol University; 3Department of Pollution Control of Thailand, Bangkok, Thailand; 4 California Environmental Protection Agency, Oakland,
  California, USA
Hong Kong (CM Wong1, JSM Peiris2, TQ Thach1, PYK Chau1, KP Chan1, RY Chung1, GN Thomas1, TH Lam1, TW Wong3, and AJ Hedley1)
  1
  Department of Community Medicine, The University of Hong Kong; 2Department of Microbiology, The University of Hong Kong; 3 Department of Family and Community
  Medicine, The Chinese University of Hong Kong, Hong Kong, China
Shanghai (BH Chen1, HD Kan1, NQ Zhao1, GX Song2, GH Chen3, ZC Shan3, and CY Guo2)
  1
  School of Public Health, Fudan University; 2Shanghai Municipal Center of Disease Control and Prevention; 3Shanghai Environmental Monitoring Center, Shanghai,
  China
Wuhan (ZM Qian1, QC He2, HM Lin1, LL Kong2, NN Yang3, WS Liu4, DJ Zhou3, DP Liao1, and JJ Dan3)
  1
    Pennsylvania State University College of Medicine, Pennsylvania, USA; 2Wuhan Academy of Environmental Science; 3Wuhan Centres for Disease Prevention and
  Control; 4Wuhan Center of Environmental Monitoring, Wuhan, China
Introduction: To meet its goal of assessing the health effects of air pollution across key and often understudied areas, in a collaborative study for Public Health and Air Pollution in Asia (PAPA),
a common protocol was developed among 3 cities in China and a city in Thailand in the first wave, and another 3 cities in India in the second wave. Individual teams followed the guidelines
for development of core models and assessment of the effects, taking into account city-specific conditions. The objectives are to (i) establish a common protocol and estimate the short-term
effects of air pollution in Asian cities, and to (ii) assess the effects of air pollutants on mortality within each city and the pooled effects in all the participating cities.
Methods: Daily counts of mortality for non-accidental causes (all ages, 0-4, 5-44, 45-64 and 65+ age groups) and cardio-respiratory diseases (all ages) were modeled by Poisson regression.
Trend and seasonality using smoothers of natural spline with 4-6 degrees of freedom (df), temperature and relative humidity using smoothing spline each with 3 or 4 df, days of the week and
other city-specific conditions using dummy variables, were adjusted for in a core model for each health outcome. Adequacy of the core models was evaluated by partial autocorrelation of the
residuals. The effects of daily air pollutant concentrations of PM10, NO2, SO2 (24 hour average) and O3 (8 hour average) were estimated.
Results: Preliminary findings were that ambient concentrations for PM10 varied across the 4 first wave cities with mean levels of 52.1 µg/m3 (Bangkok), 51.6 µg/m3 (Hong Kong), 102.0 µg/m3
(Shanghai), and 141.8 µg/m3 (Wuhan). Statistically significant excess risks were observed across the 4 cities, ranging from 0.2 to 1.1% per 10 µg/m3 increase of PM10 for non-accidental causes
in all ages, with highest excess risk found in Bangkok followed by Hong Kong, Wuhan and Shanghai. Similar studies with a modified common protocol was developed for the 3 Indian cities,
Ludhiana, Chennai and Delhi, in the second wave.
Discussion and conclusions: Asian cities, like many cities in other parts of the world, suffer similar effects of excess mortality associated with increased air pollution. Despite regulations put
in place from local agencies to control air quality, there is clear evidence of the need for legislation to support a comprehensive program of air quality controls to reduce adverse health effects
associated with pollution.




SS1-08 MULTI-CITY                        STUDY OF AIR POLLUTION AND HEALTH EFFECTS IN LATIN AMERICA
Romieu I (Instituto Nacional de Salud Publica, Mexico), Gouveia N (University of Sao Paolo, Brazil), Cifuentes L (Pontifica Universidad Catolica de
Chile)
The overall goal of this project is to use a common analytic framework to examine the association between exposure to outdoor air pollution and health effects in several Latin American and
Caribbean (LAC) cities, in order to obtain comparable and updated information on effects of air pollution on several age/cause groups. This will be achieved through developing a common
protocol for the design and analysis of time series data to be followed by all centers involved in the project. An important focus will be the detailed examination of the impact of such exposures
in specific subgroups of the population considered most vulnerable such as infants and young children and the analysis by socioeconomic position.
In the first steps we will collect data and perform quality assurance for air pollution and mortality data from the participating cities from Brazil, Mexico and Chile. Candidate cities so far are:
Mexico City, Guadalajara, Toluca, Monterey and Juarez City (Mexico); Sao Paulo, Rio de Janeiro, Porto Alegre and Vitoria (Brazil); Santiago, Antofogasta, Calama, Copiapo and Temuco (Chile).
Assessment of the availability and completeness of data in other large Latin American cities will also be pursued and those who fit the inclusion criteria will be included.
Descriptive data of air pollutants and frequency distribution of mortality by cause and age groups will be carried out. First single-city analysis will be conducted using a standardized analytical
protocol (first stage analysis). Heterogeneity will be explored and data combined using meta-regression models (second stage analysis). Overall and subgroup analyses will be conducted using
Generalized Additive Models (GAM) in Poisson regression.
We will also explore the potential effect modification of socio-economic position (SEP) and other variables on the association of air pollution and mortality using multi-level modeling (second
stage analysis). A common SEP indicator will be identified that can equally represent the SEP of populations in the different cities.
This project aims to contribute to the international scientific discussion on the conduct and interpretation of time-series studies of the effects of short-term exposure, and to provide relevant
information to decision makers in the LAC Region to support prevention and control strategies of air pollution.




                                                                                                 8
                                                                                                                                                                                                       SUNDAY SEPTEMBER 3
          Symposia Session: SS2

SS2       MEASURING IN UTERO EXPOSURES IN LONGITUDINAL CHILDREN STUDIES: WHAT,
          WHY, HOW AND WHEN
Dr Georges SALINES, Dr Stéphanie VANDENTORREN, Institut de Veille Sanitaire, Saint-Maurice, France

SS2-01          IN UTERO ENVIRONMENTAL EXPOSURES IN THE NATIONAL CHILDREN’S STUDY –
                MEASUREMENT STRATEGIES
J. Quackenboss1, P. Mendola2, W. Galke3 For the National Children’s Study, Interagency Coordinating Committee and Program Office
1
    US EPA, NERL; 2 US EPA, NHEERL; 3 NICHD NCS PO
The National Children’s Study (NCS) is a longitudinal cohort study which is designed to follow a nationally representative sample of American children from prior to conception or early
pregnancy until age 21 years. The Children’s Health Act of 2000 directs a consortium of United States Federal agencies to: (1) evaluate the effects of both chronic and intermittent exposures
on child health and human development; and (2) investigate mechanisms of developmental disorders and environmental factors that influence health and developmental processes to identify
factors that are helpful, harmless and harmful. Exposures very early in pregnancy are difficult to measure directly because women may be unaware of their pregnancies and the logistics of
timing and collecting samples during these early windows are challenging. The NCS is focused on enrolling women who are likely to become pregnant as well as those early in pregnancy in
order to assess the impact on child health and development due to in utero exposures that occur early in pregnancy. Couples planning pregnancy will be visited up to four times in the 6
months following enrollment (or moving into this group) to assess pregnancy status, exposures, and to collect biological specimens. Once pregnant, study visits are scheduled for each trimester:
a home and clinic visit in the first trimester; a clinic visit in the second and third trimester. Environmental factors to be studied are prioritized based on the core hypotheses for the NCS which
have been developed over the past five years with input from a variety of scientists and stakeholders (www.nationalchildrensstudy.gov). The NCS defines environment broadly including
chemical, physical, biological, behavioral, and social factors, as well as genetic factors and their interactions with environmental exposures. In order to evaluate the effects of both chronic and
intermittent exposures, the Study needs to develop approaches to capture the frequency, magnitude, and duration of peak or intermittent exposures and carefully consider the impact of these
measures on both participant burden and Study costs. Approaches for exposure assessment in the NCS will likely involve a combination of environmental and biological samples, together with
questionnaire and diary reporting. A strategy for their use is based on two related concepts: 1) core measures obtained for the entire cohort and the use of validation sub-samples for more
detailed exposure measurements (i.e., for measurement error adjustment); and 2) a hierarchical approach that relates measures obtained at different levels of aggregation (e.g. individual,
residential, community, and region). A comprehensive pilot study will assess the feasibility of enrolling women using a household-based screening approach and will demonstrate the
acceptability of biologic and environmental sample collection during pregnancy. This is an abstract of a proposed presentation and does not necessarily reflect EPA policy.




SS2-02          CHARACTERIZING EXPOSURES TO NONPERSISTENT PESTICIDES DURING PREGNANCY
                AND EARLY CHILDHOOD IN THE NATIONAL CHILDREN’S STUDY
Whyatt RM1 and Bradman A2
1
 Columbia Center for Children’s Environmental Health, Mailman School of Public Health, Columbia University, New York, New York ; 1Center for Children’s
Environmental Health Research, School of Public Health, University of California, Berkeley, California. The authors were members of the Chemical Exposures Work Group
of the National Children’s Study
The National Children’s Study (NCS) is a longitudinal birth cohort study initiated in the United States to evaluate the relationships between children’s health and the environment. Plans are
to follow 100,000 children from preconception or early pregnancy until adulthood. Among multiple health outcomes, the study is proposing to investigate whether pre- and/or postnatal
exposures to nonpersistent pesticides increase the risk of poor performance on neurobehavioral and cognitive exams during infancy and early childhood. We have reviewed measurements for
exposure assessment and propose a sampling scheme that can be used to characterize exposures during pregnancy and early childhood. Characterization of exposures will be challenging.
Nonpersistent pesticides include many chemicals with biologic half-lives on the order of hours or days. Exposures can occur through multiple pathways (e.g., food and residential or agriculture
pesticide use) and by multiple routes (inhalation, ingestion, and dermal). Effects may depend on the developmental stage when exposure occurs. Sequential sampling will be required and
should involve a combination of environmental and biologic monitoring, as well as collection of questionnaire data. Exposure dosimeters will likely include biologic markers, personal and
indoor air sampling techniques, collection of dust, surface and dermal wipe samples, and dietary assessment tools. Criteria for sample selection include evaluation of the time-frame of exposure
captured by the measurement in relationship to critical windows of susceptibility, the cost and validity of the measurements, participant burden and variability in exposure routes across
populations and at different age periods. As part of the NIEHS/EPA Children’s Environmental Health Center, we have been conducting longitudinal birth cohort studies to evaluate effects of
prenatal exposure to nonpersistent pesticides on fetal growth and infant neurocognitive development. We have utilized a number of strategies to assess exposure, including measurements of
pesticides in personal air and house dust samples collected during pregnancy and in maternal and child urine, maternal and umbilical cord blood, and meconium. Our research has shown
significant associations between certain insecticides and adverse fetal growth and neurocognitive development. Lessons learned from this research will be discussed.




SS2-03          BIRTH COHORTS IN THE FAROE ISLANDS: ASSESSMENT OF PRENATAL TOXICANT
                EXPOSURE AND 14 YEARS OF FOLLOW-UP
Philippe Grandjean, Pal Weihe, Frodi Debes, and Esben Budtz-Jørgensen
(University of Southern Denmark; Harvard School of Public Health; Faroese Hospital System; and University of Copenhagen)
The Faroe Islands are located in the North Atlantic between Shetland and Iceland. The population of 48,000 mainly relies on seafood, and traditional diets also include meat and blubber from
the pilot whale. The meat is contaminated with methylmercury and the blubber with lipophilic pollutants, mainly PCB and DDE. From 1985, prospective studies have been carried out to
characterize the adverse effects of seafood contaminants.
In the design of these studies decisions were made in regard to the types of biological samples to collect for contaminant analyses (as exposure biomarkers). Beneficial parameters, such as
essential nutrients and of breast-feeding were also recorded. Follow-up for developmental patterns, especially in regard to neurobehavioral and immunological outcomes, was carried out at
carefully selected ages that were deemed optimal for the main effect parameters.
Cohort 1: A cohort of 1022 singleton births was assembled in the Faroes during a 21-month period of 1986-1987. The methylmercury exposure was assessed by analysis of cord blood, cord
tissue, and maternal hair. The mercury concentrations covered a span of about 1000-fold. The first detailed examination (neuropsychological, neurophysiological and neuropediatric) took place
at age 7 years, i.e., just before school entry. A total of 917 of the eligible children (90.3%) completed the examinations. The children were re-examined at age 14, with a similar participation
rate. Postnatal exposures were assessed only in connection with these examinations and averaged about one-fifth of the prenatal levels. Clear dose-response relationships were observed for
deficits in attention, language, and memory, and in neurophysiological parameters. Postnatal exposure was not associated with these outcomes.
Cohort 2: The next cohort was established during a 12-month period in 1994-1995 and included 182 singleton term births from consecutive births with emphasis of high-exposure
communities. Maternal serum from the last antenatal examination and milk were also collected in this case, with the purpose of analyzing PCB and related pollutants. These children were
examined by the Neurological Optimality Score at age two weeks and again by a variety of tests at approximately one-year intervals. Examinations at age 10 years have just been completed.
Mercury-associated deficits were again found, most clearly at age two weeks and at school age.
Cohort 3: This cohort consists of 650 children born during 1998-2000. Following a two-week examination, a subgrouo was seen at age 18 months to ascertain antibody responses to childhood
immunizations. A comprehensive examination was then carried out just before and one month after the 5-year vaccination. These children are now being re-examined at age 7 years with
emphasis also on neurobehavioral function. Mercury exposures have declined in this community, and emphasis is now on lower-level mercury neurotoxicity and the independent effects of
PCBs and related substances.




                                                                                                 9
        Symposia Session: SS2

SS2-04          THE TECHNICAL WORKSHOP ON OPTIMIZING THE DESIGN AND INTERPRETATION OF
                EPIDEMIOLOGIC STUDIES FOR ASSESSING NEURODEVELOPMENTAL EFFECTS FROM
                IN UTERO CHEMICAL EXPOSURE
Christopher Cox, Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health
While many epidemiologic studies of children’s environmental health have been completed, and more are being planned, a comprehensive examination of the methodologies commonly used
in past studies has not been conducted. Such an examination could: assess the strengths and limitations of the methodological approaches used to date; consider scientific and technical
advances in relevant methodologies; and help to elucidate optimal approaches to estimating exposure, identifying and measuring neurodevelopmental effects, and interpreting epidemiologic
data. In short, such an examination could serve to identify the key methodological factors that ultimately determine the value and strength of a study.
The Technical Workshop on Optimizing the Design and Interpretation of Epidemiologic Studies for Assessing Neurodevelopmental Effects from In Utero Chemical Exposure provided a forum
for examining these methodologies. An expert panel of researchers was selected by the steering committee for the workshop, and asked to provide written responses to a series of questions
in advance of the workshop itself. The workshop was held on September 14, 2005, in conjunction with the Twenty-second International Neurotoxicology Conference in Raleigh-Durham, North
Carolina. The objective of the Workshop was to begin to develop a scientific consensus on the key principles and considerations for optimizing the design and interpretation of epidemiologic
studies of in utero exposure to chemicals and subsequent neurodevelopmental effects. The goal of the workshop was to provide not only useful input for the design of future investigations,
but also metrics that can be used to judge the adequacy of reported studies. Workshop deliberations included only scientific methodological issues (i.e., the development of ‘best practices’
for future study design, conduct, reporting and interpretation); the workshop did not include an evaluation of the findings or conclusions from previous epidemiological studies of
environmental health. A summary of the daylong discussions will be published in a special issue of NeuroToxicology.
This presentation will review the recommendations of the workshop, with emphasis on the measurement of exposure. An illustration will be provided by a study of the developmental effects
of prenatal exposure to methylmercury from fish consumption, The Seychelles Child Development Study, with which the presenter has been affiliated for many years.




SS2-05          THE COMPLEX ENTERPRISE OF MODELLING PRENATAL EXPOSURE TO CIGARETTES:
                WHAT IS “ENOUGH?
Kate E. Pickett1, Vanja Dukic2, Paul J. Rathouz2 Kristen Kasza2, Kathryn Weaver3, Marina Niessner2, Lauren S. Wakschlag3
1
 Dept. of Health Sciences, University of York; 2Dept. of Health Studies, University of Chicago; 3Dept. of Psychiatry, University of Illinois at Chicago
Our research programme focuses on modelling pathways from maternal smoking during pregnancy to problem behaviours in exposed offspring, with an emphasis on establishing whether
exposure plays a causal role or whether it is a marker for other risk factors. As the timing, intensity and duration of foetal exposure may be critical in such aetiological studies, enhancing the
quality of exposure measurement is essential. Yet developing an integrated approach to this question is difficult because of intra-individual variations in smoking topography and metabolism,
and methodological constraints inherent in even the most meticulous exposure measurement. Our team brings several different disciplinary approaches, (epidemiology, biostatistics,
developmental psychopathology, pharmacology, and health psychology) and datasets (data from national surveys, small samples with intensive multi-method assessment) to bear on this issue.
Our research addresses three questions:
• Do prospective vs. retrospective, repeated vs. single measures and biomarker vs. self-report methods provide useful information in characterising patterns of smoking during pregnancy?
• Can we combine different methods to enhance the precision of exposure measurement without compromising feasibility in epidemiological studies?
• What psychosocial and contextual maternal factors are associated with different patterns of smoking in pregnancy, and how might these confound aetiological studies of behavioural
  outcomes in offspring?
We have demonstrated substantial within-person fluctuations in pregnancy smoking. We have also developed new statistical approaches including: (1) characterizing patterns of smoking
during pregnancy based on repeated, prospective self-report and biological measures using random-effects models; (2) using Bayesian methods to create a latent exposure measure via
calibration of self-reported smoking, based on variability in cotinine levels at various levels of self-reported smoking. In an on-going prospective study we are examining associations between
prospective and retrospective approaches. We test the utility of these new measures by assessing their predictive value in modelling foetal growth parameters and behavioural outcomes during
the first years of life.
We have also found that women who quit smoking during pregnancy have a lower prevalence of interpersonal problems, problems with adaptive functioning, and problematic health
behaviours than either those who smoked in pregnancy or women who never smoked. We have demonstrated the relevance of this finding in a study of foetal exposure and infant
temperament, in which pregnancy quitting was protective against behavioural risk.
Smoking during pregnancy is a complex health and maternal behaviour. Precision of exposure measurement is critical for disentangling effects related to the “type of women who smoke”
from teratological effects of smoking itself.




SS2-06          COMPARISON BETWEEN SELF REPORT AND BIOCHEMICAL ASSESSMENT OF
                SMOKING IN PREGNANCY.
Triche EW, Belanger K, Hellenbrand K, Leaderer BP, Bracken MB. Department of Epidemiology and Public Health, Yale School of Medicine, New Haven,
CT, USA.
Background: Having reliable data on important in-utero exposures such as smoking has important implications for short term (e.g., perinatal) and longer term (e.g., childhood and adulthood)
outcomes. One recently completed cohort study allows us to examine the reliability of self-report data of in-utero exposure to tobacco. If biochemical verification is found necessary, then
exposures must be assessed during pregnancy.
Methods: The Nutrition in Pregnancy study (NIP) screened over 11,000 women early in pregnancy and enrolled 2,478 women with various levels of self-reported tobacco consumption. Women
gave informed consent prior to enrollment. At enrollment interview (<24 weeks gestation), women were asked detailed questions about smoking exposures in early pregnancy. At the same
visit, we collected urine samples, which were analyzed for cotinine, nicotine, and creatinine concentrations. Self-report and biochemical measures of tobacco exposure were compared.
Results: 193 women had levels of urinary cotinine >80 ng/ml, a level which indicates active smoking (Smoke-Free Families project). According to interview, only 169 women reported active
smoking in the prior day; 153 women reported smoking in the 3rd month of pregnancy, and 273 reported smoking at some point since the pregnancy began. Of the 193 smokers identified by
cotinine, 29 (15%) reported not smoking “yesterday”. Of the 29 women, 13 reported no passive smoke exposure on the previous day. While biochemical verification may identify women who
misrepresent their smoking status, it is only valid for a 7-day window (SRNT Workgroup); thus this measure may not reflect exposures at different times in pregnancy. In our data, of 273
women who reported smoking at some time before interview, only 191 (70.0%) had urinary cotinine levels >80 ng/ml at interview. 100 of 273 women (37%) reporting smoking in 1st trimester
had cotinine levels < 80 ng/ml. Analysis comparing self-report to biochemical measurement of other exposures assessed at different points during the pregnancy and immediately postpartum
is underway.
Conclusions: Self-reported exposure during the perinatal period may be inaccurate, particularly for exposures such as smoking during pregnancy when there is strong social and personal
pressure to quit. For primary assessment of these exposures, biochemical verification may be important. However, biochemical measures are subject to different problems, including being valid
for only a short window of time.




                                                                                               10
                                                                                                                                                                                                    SUNDAY SEPTEMBER 3
        Symposia Session: SS2

SS2-07          DEVELOPMENT OF TWO PROSPECTIVE BIRTH COHORTS IN SOUTHERN CALIFORNIA
M. Wilhelm, and B. Ritz, Department of Epidemiology, University of California, Los Angeles
Data concerning the impact of pre- and postnatal air pollution exposures on development of reactive airway disease are sparse, even though the first six years of life comprise a period of high
susceptibility due to the immaturity of the human lung. We are currently planning to conduct two prospective birth cohort studies to examine the influence of perinatal air pollution exposures
on subsequent development of asthma and other adverse health outcomes in Los Angeles, California, one of the smoggiest cities in the U.S. In a nested case-control survey of approximately
2,500 women who delivered infants in 2003, we obtained detailed information on residential mobility, exposure to indoor air pollution sources, time-activity patterns and risk behaviors during
pregnancy and collected home dust wipe samples. We are planning to re-contact participants when the children are 3 and 6 years of age to assess the incidence and prevalence of respiratory
and allergic symptoms and to perform lung function testing using portable spirometers and skin prick testing in a nested subgroup of children (at age 6). We will derive prenatal and childhood
air pollution exposures using data from ambient air monitors, traffic density metrics and pollution surfaces estimated through land use regression (LUR) and geostatistical modeling. In a second
prospective birth cohort study of 600-1,000 Los Angeles families enrolled at delivery and followed for five years to include second pregnancies, we plan to create a biological sample repository
by collecting maternal blood and urine and cord blood at delivery and during pregnancy. We will examine the impact of prenatal exposures to ubiquitous environmental toxins on child health.
We will administer a questionnaire at birth of the first pregnancy designed to assess prenatal traffic-related and consumer product exposures, and will also create pollution exposure measures
using existing data sources. Based on these preliminary measures, we will select a subsample of approximately 60 women for extensive personal air sampling of polycyclic aromatic
hydrocarbons (PAHs), ultrafine particles, organic and elemental carbon, and hopanes approximately six months post-delivery and analysis of blood and urine for the PAH biomarkers1-
hydroxypyrene (1-OHP) and PAH DNA-adducts. Future work includes similar monitoring once during each trimester of the second pregnancy. Infants enrolled in these unique cohorts comprise
especially vulnerable populations since by design many of them will be low birth weight or preterm and live in disadvantaged neighborhoods in LA County.




SS2-08          ENVIRONMENT AND CHILD’S HEALTH: THE INMA SPANISH STUDY
Núria Ribas-Fitó1, Rosa Ramón2, Ferran Ballester3, Joan Grimalt4, Alfredo Marco5, Nicolás Olea6,Manuel Posada7, Marisa Rebagliato2, Adonina Tardón8,
Maties Torrent9, Jordi Sunyer1,10 on behalf of the INMA Study Group.
1
  Institut Municipal d’Investigació Mèdica. Barcelona, Spain ;2 Departamento de Salud Pública. Universidad de Alicante. Alicante, Spain ;3 Escola Valenciana d’Estudis
en Salut-CS-GV. Valencia, Spain ;4 Institut d’Investigacions Químiques i Ambientals. Barcelona, Spain ;5 Hospital Universitario La Fe- CS-GV. Valencia, Spain ;6 Hospital
Universitario San Cecilio – SAS-UGR. Granada, Spain ;7 Instituto de Investigación de Enfermedades Raras- ISCIII. Madrid, Spain ;8 Universidad de Oviedo, Oviedo,
Spain ;9 Àrea de Salut de Menorca – IB-Salut. Maó, Spain ;10 Universitat Pompeu Fabra. Barcelona, Spain
The INMA (INfancia y Medio Ambiente (=Spanish for Environment and Childhood)) is a population based cohort study in different Spanish cities, that focuses on prenatal environmental
exposures and growth, development and health from early foetal life until childhood. The study focuses on five primary areas of research: (1) growth and physical development; (2) behavioural
and cognitive development; (3) asthma and allergies; (4) sexual and reproductive development; and (5) environmental exposure levels and pathways. The general aims of the project are: 1.
To describe the degree of individual prenatal exposure to environmental pollutants, and the internal dose of chemicals during pregnancy, at birth and childhood in Spain; 2. To evaluate the
impact of the exposure to different contaminants on foetal and infants’ growth, health and development; 3. To evaluate the role of diet on foetal and infants’ growth, health and development;
and 4. To evaluate the interaction between persistent pollutants, nutrients and genetic determinants on foetal and infant growth, health and development.
Extensive assessments are going to be carried out on 3300 pregnant women and children. Data are going to be collected by physical examinations, questionnaires, interviews, ultrasound and
biological samples. Pregnant women are being assessed at 10-13, 20 and 28-32 weeks of gestation to collect information about environmental exposures and foetal growth. The children will
be followed until the age of 4.




                                                                                              11
        Symposia Session: SS3

SS3 BIOMARKERS IN                            HUMAN STUDIES AND RELEVANCE TO DISEASE OUTCOMES
SS3-01 BIOMARKERS          IN NEWBORNS – THE IMPACT OF ENVIRONMENTAL EXPOSURE TO
               CARCINOGENIC PAHS
Radim J. Sram, Blanka Binkova, Jan Dejmek, Irena Chvatalova, Ivo Solansky, Jan Topinka, Institute of Experimental Medicine AS CR and Health
Institute of Central Bohemia, Prague, Czech Republic
Introduction. The molecular epidemiology methods were used to analyze the impact of air pollution in pregnancy outcome studies in the Czech Republic.
Methods. PM10, PM2.5, and carcinogenic polycyclic aromatic hydrocarbons (c-PAHs) were daily determined using stationary monitoring. The placental bulky DNA adducts were determined
by 32P-postlabeling assay. The genetic polymorphisms of metabolic genotypes (GSTM1, GSTP1, GSTT1, CYP1A1*2A, CYP1A1*2C, EPHX1, NAT2) were determined by PCR. Cotinine and vitamins
A, C, E, were analyzed in plasma. Data on pregnancies were extracted from medical records and questionnaires.
Results. Organic compounds adsorbed to air particles (PM10) induced DNA adducts and embryotoxicity in vitro studies. C-PAHs were mostly responsible for the genotoxic activity, contributing
to 45-50% of all DNA adducts induced by these complex mixtures. DNA adducts in placentas were affected by air pollution, smoking, genotypes, vitamin C levels. Higher DNA adducts were
observed in nonsmoking mothers delivering children with IUGR (intrauterine growth retardation). In the Pregnancy Outcome Project, an increased risk of IUGR was established for mothers
who were exposed to c-PAHs > 15 ng/m3 during the first month of gestation. Birth weight was significantly decreased by smoking, ETS and genetic polymorphisms of CYP1A1*2A, CYP1A1*2C,
GSTM1, and EPHX1.
Discussion and Conclusions. c-PAHs seem to be an important source of genotoxic and embryotoxic activities of organic mixtures associated with urban air particles. These are new results,
indicating the significant impact of c-PAHs to the early stages of fetus development.
Supported by the Czech Ministry of Environment VaV/740/5/03 and by the EC QLK4-CT-2002-02198.




SS3-02         XPD GENE POLYMORPHISMS, AIR POLLUTION, AND BRONCHITIS IN EARLY
               CHILDHOOD
Irva Hertz-Picciotto1 Jan Topinka2 Miroslav Dostal2 Jesse P. Joad3 Poh-Sin Yap1 Teri Greenfield1 Radim Sram 2
1
 Department of Public Health Sciences, University of California; Davis; 2 Laboratory of Genetic Ecotoxicology, Institute of Experimental Medicine, AS CR and Health
Institute of Central Bohemia, Prague, Czech Republic; 3 Department of Pediatrics, University of California; Davis, CA
Introduction: Bronchitis in early childhood is a common occurrence. It can reflect both viral and bacterial infection, as well as hypersensitivity involving wheeze. We examined genetic
polymorphisms and gene-environment interactions in relation to early childhood bronchitis.
Methods: A birth cohort comprising a probability sample of 20% of deliveries between May 1994 and March 1999 in two districts in the Czech Republic was followed up at 3 to 4.5 years of
age. Pediatric records were abstracted for all illnesses. Bronchitis was defined as a physician diagnosis of J20 (acute bronchitis). Parents completed questionnaires at birth and at follow-up
regarding sociodemographics, lifestyle, indoor air pollution sources, breastfeeding, day care, and other factors potentially related to respiratory illnesses. PCR-RFLP techniques were used for
genotype analysis of xenobiotic metabolizing (GSTM1, GSTT1, etc.) and DNA repair genes (XPD-6, XPD-23) in samples of fetal tissue from placentas of 800 births. Poisson regression was used
to model the number of episodes of bronchitis during the first two years of life in relation to child’s genotype, with adjustment for ethnicity, maternal smoking, child’s sex, mother’s age and
education, season of birth and use of coal for home heating or cooking. Additionally, XPD-6 was examined for modification of the effects of acute air pollution in a repeated measures logistic
regression analysis.
Results: More than 60% of the children experienced 1 or more episodes of bronchitis in the first two years of life. Xenobiotic metabolizing genes did not predict the number of bronchitis
events, but XPD-6 polymorphisms did. Those with the XPD-6 AA genotype (40% of children) experienced bronchitis more frequently than the homozygous CC and heterozygous children
combined (RR=1.45, 95% confidence interval (CI)=1.36, 1.55). Those with the CC genotype had a similar risk to the heterozygotes. Genotype did not modify the effects of ambient air pollution.
Discussion & Conclusion: XPD genes encode for nucleotide excision repair and basal transcription, and their transcription may be regulated by insulin. In our data, the C-allele of XPD-6
appears to confer some protection against bronchitis in early childhood, after adjusting for ethnicity and other factors. Neither XPD-6 nor XPD-23 renders children more susceptible to short-
term respiratory effects of ambient particles or PAHs. Although studied mainly for their involvement in cancer risk, XPD genes may play a role in immune competence.




SS3-03         MOLECULAR EPIDEMIOLOGICAL STUDIES IN 1,3-BUTADIENE EXPOSED CZECH
               WORKERS
Richard J. Albertini1,3, Radim J. Sram2, Pamela M. Vacek3, Janice A. Nicklas3, Jake McDonald4, James A. Swenberg5
1
  BioMosaics, Inc. Burlington, VT (USA); 2 Laboratory of Genetic Ecotoxicology, Prague, CZ; 3 University of Vermont, Burlington, VT (USA); 4 Lovelace Respiratory Research
Institute, Albuquerque, MN (USA); 5 University of North Carolina, Chapel Hill, NC (USA)
Introduction: 1,3-butadiene’s (BD) mode of action as a rodent carcinogen is genotoxicity, an important precursor for cancer.
Methods: A continuum of biomarkers was measured in BD exposed workers in the Czech Republic following extensive exposure assessments over several months. .
Results: Mean 8-hour TWA BD exposure levels for the 25 controls, 24 monomer production workers and 34 polymerization workers (all male) in an initial study were 0.010, 0.290 and 0.810
ppm. Biomarkers measured included urinary metabolites, haemoglobin adducts, SCEs, HPRT gene mutations, chromosome aberrations and metabolic genotypes at several loci. Urinary
metabolite and haemoglobin adduct concentrations were significantly elevated, correlating with BD exposure levels. Detoxification metabolism as reflected by urinary metabolite ratios showed
the dominance of hydrolysis. No significant relationships were seen between BD exposures and gene mutations or cytogenetic endpoints. Some polymorphisms did modulate detoxification
metabolism but did not otherwise influence biomarker responses. A second study compared females and males. In addition to the haemoglobin adducts measured in the first study, analysis
of a diepoxide specific adduct was added. BD levels had fallen since the first study, TWAs are now 0.180 and 0.0035 ppm for 23 female exposed and 26 female control workers, respectively
and 0.370 and 0.007 ppm for 30 male exposed and 25 male control workers, respectively. Urinary metabolite concentrations again reflected BD exposures in males but not in females who
actually excreted less urine metabolite per unit BD exposure than did males. Metabolite ratios again reflected the dominance of hydrolytic detoxification and were not different between the
sexes. Although haemoglobin adduct determinations are incomplete, diepoxide specific adducts were non-detectible in any subjects. Again, neither HPRT mutations nor cytogenetic changes
were related to BD exposures in either sex.
Discussion and Conclusions: Biomarkers are useful for mechanistic cancer assessments allowing comparisons between animal experiments and human molecular epidemiological studies.
Metabolism, metabolite accumulations and genotoxic effects can now be quantified by molecular epidemiology. The BD studies are illustrative. BD’s genotoxicity and therefore its
carcinogenicity depend on electrophilic metabolites, particularily the diepoxide, that result from oxidative metabolism. Humans form much less of this metabolite per unit dose of BD than do
rodents. This is reflected in its diminished genotoxicity in humans, with the weight of evidence (including these Czech studies) indicating that it has yet to be demonstrated under modern
exposure conditions. This has implications for assessing its cancer risk.




                                                                                              12
                                                                                                                                                                                                    SUNDAY SEPTEMBER 3
        Symposia Session: SS3

SS3-04          THE USE OF BIOMARKERS IN CANCER RISK ASSESSMENT
R. Julian Preston, National Health and Environmental Effects Research Laboratory, US EPA, Research Triangle Park, NC, USA
The recent guidelines recommended by different regulatory organizations and agencies for conducting cancer risk assessments for exposures to environmental chemicals emphasize the use
of mechanistic data and non-tumor endpoints in support of dose-response characterization. These guidelines are based on the use of a mode of action approach for describing the formation
of chemically-induced tumors. A mode of action is defined by the U.S. EPA as “a sequence of key events and processes starting with interaction of an agent with a cell, proceeding through
operational and anatomical changes and resulting in cancer formation”. A general approach for establishing potential risk to humans is to develop a mode of action for a rodent model by
which it is proposed that a particular chemical can induce tumors. Examples of such a mode are DNA-reactivity, receptor-mediated and cytotoxicity with regenerative proliferation, appreciating
that any chemical can work via one or more modes. A defined set of key events is used for each mode of action. These key events are empirically observable precursor steps that are each
necessary but by themselves not sufficient to produce a tumor. Key events are thus non-tumor endpoints that can predict tumor outcomes, qualitatively and perhaps quantitatively. As such
they are biomarkers of response. Having established a mode of action in a rodent (since for most environmental chemicals tumor data are not available from epidemiological studies), it is
necessary to establish if the same process is operating in humans. This can be achieved by use of a Human Relevance Framework that considers if the key events in the rodent could occur in
humans. Again biomarkers of effect in humans could serve as predictors of key events. Once it is established that a particular mode of action could occur in humans, a quantitative risk
assessment is recommended. Biomarkers of response can be used to help describe the shape of the tumor does-response at low doses and for establishing lowest observable effect levels
(LOELs) for extrapolation purposes. Thus, the development of informative biomarkers of response remains a critical need for reducing uncertainties in cancer risk assessments [This abstract
does not necessarily reflect U.S. EPA policy]




SS3-05          MOLECULAR EPIDEMIOLOGY OF CARDIOVASCULAR DISEASES AND GLAUCOMA
Silvio De Flora and Alberto Izzotti, Department of Health Sciences, University of Genoa, Italy
Introduction. Several chronic degenerative diseases share common pathogenic mechanisms, such as DNA damage, oxidative stress, and chronic inflammation.
Methods. Bulky DNA adducts and 8-oxo-dG were measured by 32P postlabeling in human and rat tissues. Metabolic polymorphisms were evaluated by RFLP-PCR.
Results and Discussion. DNA adducts in abdominal aorta smooth muscle cells (SMC) from 85 atherosclerotic patients were significantly correlated with atherogenic risk factors, including
age, number of cigarettes, hypertension, blood triglycerides and total/HDL cholesterol, fluorescent DNA adducts, and 8-oxo-dG. 8-Oxo-dG levels were 3-fold higher in the aorta inner layer than
in the corresponding medium layer. Ethenodeoxyadenosine was significantly higher in SMC from current smokers than in SMC from nonsmokers or former smokers. The 4977 bp mtDNA
deletion was detected in aorta SMC. Plasma malondialdehyde was correlated with the number of cigarettes. Plasma homocysteine was higher in 84 atherosclerotic patients than in 74 controls.
In atherosclerotic patients, plasma homocysteine was positively correlated with blood triglycerides and inversely correlated with total glutathione and fruit consumption. Moreover, it was
increased by homozygosity for slow MTHFR. DNA adducts were increased in patients having a null GSTM1 genotype, while GSTT1, NAT1 and NAT2 did neither affect DNA adducts nor 8-oxo-
dG. Trombophilic polymorphisms affected the 4-5% of the atherosclerotic patients. Formation of DNA adducts in the thoracic aorta of smoke-exposed rats was modulated by the oral
administration of chemopreventive agents.
DNA adducts in rodent heart accumulate with age and with exposure to cigarette smoke. The heart was quite sensitive to smoke-induced DNA alterations, and was not effective in their
removal. Since cardiac myocytes are perennial cells, they are not prone to develop cancer in adults, but occurrence of DNA damage is presumably associated with degenerative conditions.
Interestingly, smoke-induced DNA adducts in rat heart were increased by the simultaneous ingestion of alcohol, while they were attenuated by chemopreventive drugs and dietary agents.
Glaucoma is the most frequent cause of irreversible blindness worldwide. We demonstrated that a significant increase of 8-oxo-dG, correlated with the increase of intraocular pressure and
with visual field defects, occurs in trabecular meshwork cells of open-angle glaucoma patients. Both within controls and glaucoma patients, 8-oxo-dG levels were significantly higher in subjects
having a GSTM1-null polymorphism.
Conclusions. The results obtained in both humans and animal models suggest that DNA alterations occur not only in cancer but also in cardiovascular diseases and glaucoma.




                                                                                              13
         Symposia Session: SS4

SS4            INTAKE FRACTION: PAST, PRESENT, AND FUTURE
Susan Greco and Matti Jantunen

SS4-01 INTAKE                    FRACTION: WHERE DO WE COME FROM?
Debbie Bennett, University of California, Davis, USA
Activities such as comparative risk analysis, life-cycle assessment, emissions trading and sustainable development are creating a growing demand for information about the potential exposure
and adverse effects of the thousands of chemicals released to the environment. This demand has fostered measurement and modeling efforts that link emissions to effects for the release of
substances from a wide range of human products and activities, such as consumer goods, cooking, smoking, energy conversion, industrial production, and agricultural activities. A source-to-
intake metric is useful, allowing researchers to compare the intake resulting from various sources in a clear, straightforward manner. As a result, researchers see a need for, and have proposed,
various terms to express source-to-intake relationships. The first known articulation of the source-to-intake relationship was in the field of radiation protection by the ICRP in 1982. Some who
worked on radiation dosimetry proposed that this approach could be extended to other materials. Over the last two decades, there have been many examples of calculations relating the
inhaled intake of a substance to the amount emitted to air, particularly to contrast releases to the outdoor and indoor environments. There were also efforts to estimate multimedia intake by
a population relative to a specified release (to air, water, or soil), in these cases contrasting the difference between various chemicals. These examples point out two of the critical attributes
in a soure-to-intake relationship, specifically, differences in the emission scenario and differences between compounds. Historical references, by lead authors including Cohen, Harrison, Smith,
Evans, Guinee, Hertwich, Jolliet, and Bennett, and the motivation for these efforts, will be presented. The diversity of fields of inquiry between the investigators, often with a lack of intersection,
led to a lack of consistency, particularly with regard to units and calculation boundaries. To promote consistency, a working group formed with the goal of developing a consistent definition
and terminology for this idea. The group proposed the term intake fraction (iF) as the primary label for quantifying the source-to-intake relationship. The intake fraction is defined as the
integrated incremental intake of a pollutant, summed over all exposed individuals, and occurring at any time, released from a specified source or source class, per unit of pollutant emitted.




SS4-02 INTAKE                    FRACTION: WHERE ARE WE?
Olivier Jolliet, Manuele Margni, University of Michigan, Center for Risk Science and Communication, Dept. of Environmental Health Sciences, School
of Public Health, Ecole Polytechnique de Montréal, CIRAIG, Canada
The intake fraction – the fraction of a pollutant’s emissions that is taken in by a population - is an interesting and useful concept that has been defined within a multidisciplinary collaborative
effort, bringing together scientists from multimedia modelling, risk assessment, outdoor and indoor air pollution, and Life Cycle Impact Assessment. As mentioned by Nazaroff, “intake fraction
is a lens through which we can view the problem and bring important aspects into focus that have never been seen before”. So far, most of the research and evaluation emphasis has been
set on the fate component of the evaluation, with e.g. the OECD comparison exercise by Fenner et al. for multi-media models. This comparison has shown that improving the quality of input
data is of the highest priority, with models providing rather comparable results using the same quality input data. Significant progress has also been achieved in the modeling of intake fraction
for primary and secondary pollutants (iF from 10-5 to 10-7) and in the estimation of indoor intake fractions that lead to higher values of 10-2 to 10-4. Recently, a stronger emphasis is being placed
on estimating exposure, which has improved the assessment of the food chain where intake fractions of 10-2 to 10-3 are reported. In the case of bioaccumulation, this is occurring with a)
improved correlation or better modeling of the bioconcentration factors in vegetation, meat, milk and eggs, b) the modeling of the full foodweb within the aquatic ecosystem, and c) the better
description of coastal marine environments where 80% of the fish catch takes place. Further developments are emerging to link intake with absorption, uptake and metabolism in the body,
coupling multi-media fate and pharmacokinetic models to cover the whole cause-effect chain from source to body burden.




SS4-03 APPROACHES FOR IF MODELLING: IF DISTRIBUTIONS ACROSS POPULATION,
GEOGRAPHIC AREA AND TIME
Asikainen A.1; Briggs, D.2; Kukkonen J.3; Jantunen M.1
1
  National Public Health Institute / Dept. Environmental Health, FINLAND; 2 Imperial College / Dept. Epidemiology and Public Health, UK; 3 Finnish Meteorological Institute
/ Dept. Air Quality Research, FINLAND
Introduction. Intake fraction (iF) is the integrated intake of a pollutant released from a source and summed over all exposed individuals (individual iFs) during a given time, per unit emitted.
This concept directly relates exposure to a source. Being dimensionless, iF values from different sources are particularly helpful for risk comparison. The published iFs have been mostly
generated by dispersion and exposure modelling. Population iF values lack any information about the high end of the individual iFs caused by short-term events in environmental dispersion
and by individual behaviour. The objective of the present study is to develop and validate modelling techniques to predict the full distributions of individual iFs of the emissions from a source
to a target population.
                                                                 Methods. A probabilistic, GIS-based individual iF modelling approach is applied. The case studies, selected for model
                                                                 application, all have independently collected exposure data for model validation. First applications employ benzene exposure
                                                                 measurements in EXPOLIS database for modelling (i) distribution of the individual inhalation iFs for traffic benzene in Helsinki
                                                                 city, (ii) spatial distribution of the population iF over the Metropolitan area, and (iii) temporal distribution of the population iF
                                                                 over the days of a year. Other applications look at ingestion iF for dioxins from a chemical plant, inhalation, dermal and ingestion
                                                                 iFs for chloroform formed by drinking water chlorination, and inhalation iF for primary PM from a co-generating station.
                                                                 Results. First model run results are expected in the summer of 2006. The graph presents our case (i) model validation data
                                                                 calculated from EXPOLIS database for individual iF distribution. The distribution is close to lognormal, with a relative standard
                                                                 deviation 104% and respective integrated population iF of 1.47*10-4.
                                                                 Conclusions. iF modelling normalises all exposures relative to source emissions. Therefore risk managers can observe the
                                                                 relative risks of unit emissions from different sources as well as the relative benefits of different risk management options
                                                                 independently of the variations and uncertainties caused by specific local source emission data.
                                                                 Acknowledgements. The intake fraction project is funded by Cefic-LRI Contract LRI-B3.3-KTLE-0511.




                                                                                                  14
                                                                                                                                                                                                       SUNDAY SEPTEMBER 3
        Symposia Session: SS4

SS4-04 INTAKE                   FRACTIONS FOR COMPOUNDS RELEASED TO THE INDOOR ENVIRONMENT
Debbie Bennett, Robert Canales, University of California, Davis, Harvard School of Public Health, USA
A better understanding of contaminant fate and transport in indoor environments is needed to make predictions about concentrations in the home, to interpret measured data on contaminant
concentrations in residential air and on surfaces, and to aid in quantifying exposure to products used indoors. A general framework for modeling fate and transport indoors is presented and
coupled to an exposure model to simulate source-to-exposure processes. The resulting output is a time series of residues and concentrations in the modeled compartments. Exposure is
calculated through the inhalation, dermal, and non-dietary exposure routes. From the time series of indoor exposure, the intake fraction is calculated. We calculate and compare intake fraction
values and determine primary routes of exposure for a suite of organic compounds, including pesticides, formaldehyde, phthalates, and benzo[a]pyrene. A sensitivity analysis is completed to
look at which input values influence the output variance. Intake fractions did not vary significantly from 2.03E-4 between compounds. Although the values are quite similar, both the time
profile and the routes of exposure are very different between compounds. The differences in the exposure profile result from differences in the chemical properties between the compounds.
Understanding the differences in exposure profiles is critical for understanding impacts of the use of various compounds in the indoor environment.




SS4-05 PERSONAL         EXPOSURES TO TRAFFIC RELATED AIR POLLUTION WHILST WALKING
               AND IN CAR: IMPLICATIONS FOR THE ASSESSMENT OF INTAKE FRACTIONS
David Briggs, John Gulliver, Kees de Hoogh, Department of Epidemiology and Public Health, Imperial College London, UK
Introduction. Journey-time exposures to traffic-related air pollution make up a large proportion of people’s exposures to air pollution, and for many people accounts for periods of peak
exposure. They thus provide important components of intake fraction calculations for traffic-related air pollution. Choice between walking and travel by car is especially important in this
respect, though relatively few data are available on the relative exposures under these different modes. This study was aimed to compare monitored exposures whilst following the same
routes in-car and walking, in order to assess the implications for computation of intake fractions.
Methods. Measurements of personal exposure to particulates were made in a car, and whilst simultaneously walking the same routes, using portable light-scattering devices for total
suspended particulates (TSP) and PM10, PM2.5 and PM1 fractions and a pTrac monitor for ultrafines. Forty-seven routes across London were studied, representing different street types, traffic
conditions and urban land cover types. Comparisons between in-vehicle and walking exposures were made both in terms of average and total journey-time exposures, using regression analysis
and Student’s t-test for paired samples. Effects of traffic composition and flow on the levels of exposure under each mode were also examined.
Results. Strong correlations were seen between PM10 and TSP exposures, and between PM2.5 and PM1 exposures, both in-car and whilst walking (r > 0.8). Correlations between other fractions
were weak or moderate. Close correlations between average in-vehicle and walking exposures were also seen for the ultrafine, PM1 and PM2.5 fractions (r = 0.62-0.85); correlations for the
coarser components (PM10 and TSP) were much weaker (r < 0.15). Ratios of average exposures whilst walking to those in-car were approximately 3.5:1 for TSP and PM10, 2:1 for PM2.5 and
PM1, and 1.3:1 for ultrafines. Because walking generally took three-to-four times longer than driving each route, and because of higher respiration rates, total intake whilst walking would be
substantially greater still when walking compared to travelling by car.
Discussion and Conclusions. Marked differences were found between in-vehicle and walking exposures. Walking exposures are significantly higher than those in-vehicle for all particulate
fractions, and markedly so when the travel time is taken into account. Results suggest that pedestrians are likely to account for a disproportionately high percentage of the intake fraction of
traffic-related particulates, and as such are a potentially especially at-risk group. Better information on time activity patterns, including travel mode and duration, are required properly to
assess the intake fractions of traffic-related particulates, and to identify particularly vulnerable groups.




SS4-06 INTAKE                   FRACTION: AN APPLICATION IN THE TURIN PROVINCE
Benedetto Carella, Politecnico of Turin, Pierpaolo Mudu, WHO-ECEH Rome
In this study we present an application of the intake fraction based on data gathered in the southwest area of the city of Turin in Italy. The intake fraction (iF) is the time-integrated mass of
pollutant inhaled by the population divided by the mass of the pollutant released into the environment. In our case, the iF was calculated for motor vehicle generated air pollutants, CO and
NOx, using a box model as well as a one-compartment model. By using these two methods to calculate the mass of the pollutant inhaled by the population, we can investigate the influence
of using different models on the iF. In the one-compartment model, the iF is calculated using a matrix algebra solution to a first-order steady-state mass balance differential equation. Solving
the differential equation, the mass vector, M (in g), is the emissions vector, E (in g/day) divided by the rate coefficient matrix, k (in units of inverse time). The concentration vector, C, which
describes the concentration distribution of a chemical in an air compartment system, was obtained by combining the mass balance results, M, with the volume vector V. The quantity of mass
calculated from the mass-balance equation is three times bigger than the mass obtained from Gaussian dispersion model IMMPROG2000 that was employed in this application. Using the
box-model, we obtained higher iFs that were 2 to 3 orders of magnitude larger than using the one-compartment model. Much of this difference can be attributed to the difference in
concentration outputs from the two models. For example, for CO, the output of the box model was 38-132 ppm, while for the one-compartment model it was 1-3.89 ppm. The differences in
concentration reflect the different hypotheses at the basis of the models. This study will also consider a comparison between the results obtained and other results published in the literature.




                                                                                                15
        Symposia Session: SS4

SS4-07 OPTIMIZATION        OF URBAN TRANSPORTATION FLEET PM2.5 EMISSION REDUCTION
               STRATEGIES USING AN INTAKE FRACTION FRAMEWORK
Susan Greco, Harvard School of Public Health, USA
Transportation agencies often have the opportunity to retrofit current bus fleets or replace existing vehicles, with resulting reductions in emissions of fine particulate matter (PM2.5) and other
pollutants. As not all vehicles are upgraded at once, given limited resources, the agencies must choose a subset of vehicles to upgrade. At present, there is no systematic way for transportation
agencies to choose among buses in such a way that maximizes public health benefits, as applying dispersion models or measuring concentrations at numerous sites is implausible. An intake
fraction framework, in which the overall population exposure per unit emissions from a bus route is characterized, can inform questions of health efficiency. However, environmental justice
concerns are often raised regarding diesel buses in urban areas, and strictly optimizing on efficiency may have negative impacts on equity.
In this study, we evaluate a hypothetical vehicle procurement decision by the Massachusetts Bay Transportation Authority (MBTA) in the Boston, Massachusetts area. We estimate fine
particulate matter intake fractions (emissions-to-exposure relationships) for the 13 highest-ridership bus routes of the MBTA, which impact populations from differing socioeconomic strata.
To estimate the intake fractions, we applied the CAL3QHCR line source model and integrated the outputs with populations within 1000 m (using U.S. Census and ArcGIS) for each road segment
in the model region. Previous studies indicated that this approach yielded significant heterogeneity in intake fractions within an urban area, implying that differentials may exist across bus
routes. For the 13 bus routes in question, the intake fraction for residential populations varied from 6.5 to 19 per million. The effect of vehicle emissions on passengers on the bus was explored
as well. Equity considerations were incorporated by evaluating demographic patterns among the impacted populations and by developing an efficiency-equity frontier, considering which bus
routes were dominated across both dimensions (i.e., lower intake fraction and worse equity outcome) and which sat on the optimal frontier.
The intake fraction concept in this application allowed for a first-order evaluation of the health benefits of emission reductions (to be compared with the costs of such efforts), the optimal
emission reduction strategy for the MBTA to pursue, and the areas in which more refined modeling might be warranted in the future, with explicit consideration of equity as well as efficiency.




SS4-08 CALIFORNIA         POWER SECTOR EMISSIONS: STATEWIDE INHALATION INTAKE AND
               MORTALITY RISK FROM PRIMARY PM2.5 EMISSIONS
GA Heath (Energy and Resources Group, University of California Berkeley) and WW Nazaroff (Department of Civil and Environmental Engineering,
University of California Berkeley)
Introduction. The California power sector is one of the cleanest in the USA. Nevertheless, air pollutant emissions from California electricity generation units are still an environmental health
concern both locally and statewide. Heath et al., (Atmos. Env., 2006, submitted) estimated intake fractions (iF) for primary PM2.5 emissions from a sample of existing California central stations
(CS). This research uses those iF values to estimate statewide exposure burden and premature mortality impacts resulting from primary PM2.5 emissions from existing California CS.
Methods. Based on a combination of Gaussian plume modeling and GIS-based inhalation exposure assessment, Heath et al. (2006) estimated iFs for primary PM2.5 emissions from 25
California CS. These plants are representative of the full set of 45 combustion-based California power plants larger than 50 MW that do not produce useable heat (i.e., non-cogeneration). iFs
are similar to emission factors (EF) in that total impact (e.g., emissions) can be estimated using data on total activity (e.g., electricity generated). We estimate statewide inhalation intake of
primary PM2.5 from non-cogeneration, combustion-based California CS in 1999 using the central estimate of iF from our sample and data on total 1999 emissions. This estimate of statewide
intake is then combined with a concentration-response (C-R) function provided in the HEI reanalysis of the ACS PM health impacts study (Krewski et al., 2000) and modified for use with
inhaled mass to estimate premature mortality in the California population over 30 y old.
Results. We estimate statewide primary PM2.5 intake from 1999 California CS emissions as 2.4 kg and 1.4 kg for those over 30 y old (assuming homogeneous population spatial distribution
by age). Assuming a population-average breathing rate of 12 m3/d for the exposed, the Krewski et al. C-R function of 0.5% increase in annual death rate per ug/m3 increase in annual mean
concentration of PM2.5 implies an intake-response function of 0.11% increase in annual death rate per mg inhaled. This leads to our estimate of 21 premature deaths in the California
population over 30 y old from exposure to California CS primary PM2.5 emissions in 1999.
Conclusions. Intake fractions can be used to estimate health impacts (or benefits) from human exposure to emissions from not only individual sources but also whole source classes. Levy et
al. (2002) estimated 42 premature deaths from PM2.5 emissions from nine Illinois coal-burning power plants, indicating the relatively small impact of California’s natural gas-fired central
station stock.




SS4-09 INTAKE        FRACTIONS FOR EUROPEAN ECB EXISTING CHEMICALS PRIORITY
               SUBSTANCES LIST
Matti Jantunen, Aleksi Jetsu, Arja Asikainen, KTL-Environmental Health, Kuopio, FINLAND
Introduction: European Chemical Bureau’s priority substances list contains 141 existing chemicals, for which risk assessment tasks have been assigned to member countries. Some of these
risk assessment reports (RAR) have been completed, others are still under work. Intake fraction, iF, is the integrated intake of a pollutant released from a source summed over all exposed
individuals, per unit of emitted pollutant. Because iF is dimensionless, it enables straightforward comparisons between the different sources and environmental fates of pollutants as they
relate to human exposure and also risk. RARs do not contain any iF information, and therefore Cefic-LRI is funding a new study to prepare an iF-database for ECB listed priority substances,
and to develop iF modeling tools for different chemicals and sources.
Methods: As the first task, iF-values for ECB listed priority substances are extracted from literature. Further in the project, these data are expanded to represent other similar chemicals, iF
modelling techniques are developed, evaluated and applied to exiting datasets to create new iF information, and to predict individual iF distributions within populations.
Results: Very early in the study iF data has been found for 34 of the 141 ECB priority substances (20/42 for the 1st priority). For 33 substances we have iF for both atmospheric and aquatic
dispersion. Both range from below 10-7 to above 10-5., and there is no correlation between the atmospheric and aquatic iF values. For seven compound we found iF for both ambient and indoor
air dispersion, and also individual iF distributions within different urban populations. Compared to ambient air, indoor air dispersion iF values are 2 - 4 orders of magnitude higher. For urban
populations, iF from indoor sources range from 1.6 - 5*10-5, and the individual iF values from below 10-5 to above 10-4. For one priority substance, benzene, we have results from 4 different
studies. Most of these data are strictly based on modelling, but the ABS-study of the indoor iF distribution for 7 European cities data was computed from dilution measurements and monitored
time activities.
Conclusions and Discussion: iF is not a new concept, but with its present name and definition it was defined as late as 2002. Since then a surprising amount of iF methods data have already
been published, and the field is developing rapidly.




                                                                                               16
                                                                                                                                                                                                       SUNDAY SEPTEMBER 3
        Symposia Session: SS4

SS4-10 MULTI-CONTINENTAL                                    INTAKE FRACTIONS FOR HUMAN TOXICITY ASSESSMENT
Olivier Jolliet, University of Michigan, Center for Risk Science and Communication, Dept. of Environmental Health Sciences, School of Public Health;
Manuele Margni and Louise Deschenes, Montréal, CIRAIG, Canada; François Gigante and David Rochat, Ecole Polytechnique Fédérale de Lausanne,
Switzerland
Background: It is not surprising that colleagues from developing countries are reluctant to use exposure assessment methods and data developed for Europe or North America in cases where
such information is lacking for their own country. To address this need, this paper aims to develop multi-continental intake fractions for toxic air emissions and to analyze under which
conditions this spatial distinction makes a significant difference compared to generic iF values. For this purpose, a simplified but accurate correlation is determined to extrapolate continent-
specific intake fractions and characterization factors of a wide range of substances for Oceania, Africa, South America, North America and Asia, starting from the results of Europe as a base
continent.
Methods: The multimedia fate and exposure model, IMPACT 2002, is used for this purpose with global databases providing reliable and consistent data across all continents.
Results and Discussion: For a given chemical, intake fractions can vary typically by up to a factor-of-10 between continents (maximum factor = 100), mainly as a function of population
density for inhalation, and as a function of the total agriculture production per km2 for ingestion. This is significant but still limited compared to the variation between substances, of 6 orders
of magnitude in intake fraction and of 12 orders of magnitude in cumulative risks. Moreover, the ranking between continents remains almost the same for all chemicals. Therefore generic
intake fractions calculated at continental level, such as the one proposed by the common life cycle assessment method, are in most cases suitable for comparative purposes in any other
continent. However, continent-specific intake fractions are required if one is interested in evaluating absolute values or in comparing impact between scenarios with emissions in very different
continents.




SS4-11 EFFICIENCY-EQUITY        TRADEOFFS ASSOCIATED WITH POWER PLANT POLLUTION
               CONTROL STRATEGIES
Jonathan Levy, Harvard School of Public Health
In many pollution control decisions, there are tensions between efficiency and equity in deciding on optimal control strategies. While the intake fraction concept provides an understanding
about the “exposure efficiency” of emissions reductions from different sources, it provides limited insight about the distribution of exposure benefits, which is a key component of distributional
equity considerations. Development of a quantitative measure that could appropriately address distributional equity would allow for efficiency-equity tradeoffs to be formally incorporated in
the decision-making process and could allow for broader applicability of intake fraction findings. In previous studies, we determined that the Atkinson index provides a quantitative indicator
of inequality that is interpretable from an exposure and risk perspective, provided that it is applied appropriately (i.e., without value judgments about the relative importance of transfers at
different percentiles of the risk distribution, including baseline distributions of risk as well as marginal benefits of controls, and considering multiple competing policy alternatives). In this
study, we develop a number of theoretical control scenarios for power plants in the United States, focusing on the various ways by which emissions reductions could be distributed to meet
hypothetical national emissions caps for nitrogen oxides, sulfur dioxide, and primary fine particulate matter. For each control scenario, we quantify the resulting public health benefits, applying
power plant intake fractions developed from a source-receptor matrix previously validated in this context and using particulate matter concentration-response functions from the
epidemiological literature. In addition, we use the Atkinson index to determine the change in spatial equity of exposures or risks for each control scenario, based on background fine particulate
matter concentrations and baseline mortality rates at county-level resolution. The public health benefits and changes in the Atkinson index allow us to determine the optimal frontier of policies
that are not dominated from both efficiency and equity perspectives. We consider the sensitivity of our conclusions to the pollutants evaluated, the inequality indicator selected, and other
critical dimensions of the analysis, and we compare our policy options with those that would be selected without the benefit of this information. Our work demonstrates an approach for
combining risk assessment outputs based on intake fraction estimates with concepts of equity, in a format that is meaningful for public policy.




SS4-12 U.S.        URBAN SCALE INTAKE FRACTION OF MOTOR VEHICLE EMISSIONS: TRENDS
               DURING 1950 – 2000.
Julian D. Marshall, University of British Columbia, USA
Introduction: The intake fraction depends on factors such as the size of the exposed population, proximity between people and emissions, and pollutants’ environmental persistence. Previous
work (Marshall et al., Atmos. Env. 39, 1363, 2005) suggests that for nonreactive motor vehicle emissions, urban-scale intake fraction is proportional to linear population density (LPD). LPD is
the population of an urban area divided by the square root of urban land area (LPD = P/A0.5). An LPD value of 20 people per meter, for example, would mean that a 1-meter wide strip of land
fully traversing an urban area would contain 20 people. LPD accounts for urban-scale population and proximity, but not pollutants’ persistence. To a first approximation, long-term changes in
urban-scale intake fraction of nonreactive vehicle emissions are expected to be proportional to changes in LPD.
Methods: This investigation analyzed a panel dataset of LPD values by decade (1950 – 2000) for US Census-designated Urban Areas. A Census block (or block group) is defined as “urban”
if it has a population density greater than 1000 mile-2 and is surrounded by census blocks with density greater than 500 mile-2. An “urban area” is a contiguous group of urban Census blocks
with a combined population of at least 50,000 people.
Results: For US Urban Areas during 1950 – 2000, on average, population increased by a factor of ~2, and land area increased by a factor of ~7, yielding a factor of ~3 decline in urban
population density (“sprawl”). In specific urban areas, LPD may increase or decrease during the time considered. Surprisingly, however, the distribution of LPD values – including mean, median,
and shape of the distribution – remained approximately constant. (During 1950 – 2000, mean LPD declined only 9%, from 24 m-1 to 22 m-1, and population-weighted LPD declined 26%, from
41m-1 to 30 m-1.) This finding indicates that intake fraction of urban vehicle emissions may go up or down over time in specific urban areas, but to a first approximation the distribution of
values is expected to be approximately constant during 1950-2000.
Discussion: These findings have important implications for exposure assessors and for urban planners. They suggest that the average exposure impact of each kilogram of urban vehicle
emissions has remained roughly constant over time. They also suggest a scaling rule for how urban land area expanded (1950 – 2000) in response to population growth: on average, each
new urban resident occupied roughly twice the land area of existing residents.




                                                                                                17
        Symposia Session: SS4

SS4-13 THE         RELATIONSHIP OF POPULATION INTAKE FRACTION (IF) TO ENVIRONMENTAL
               PERSISTENCE
Thomas E. McKone, University of California, Berkeley and Lawrence Berkeley National Laboratory.
There is international interest in understanding and managing persistent organic pollutants. This has resulted in efforts to characterize global persistence for a large number of chemical
substances. In previous work we demonstrated for a large set of organic chemicals that intake fraction is proportional to the overall multi-media residence time, Tov, of a chemical. The
proportionality constant of this relationship has dimensions of time, represents the characteristic time for population intake (CTI) of the chemical from the environment, and is on the order of
1 million days for the continental US. In this poster, we examine the reliability and consistency with which screening-level assessments of persistence provide reliable estimates of intake
fraction for population-level exposure. We focus on both inhalation and food pathways. With a combination of models, emissions data, ambient air measurements, food residue data, and
biomonitoring data, we evaluate results for three case studies that give insight about the how iF depends on Tov. First we consider the case of benzene emissions in the South Coast air basin
of California. Here the relative uniformity in space and time of the emissions and the abundance of reliable emissions and monitoring data make possible the harmonized use of models and
measurements to show that CTI is simply the ratio of basin air flushing rate divided by the ensemble population breathing rate. The second case study makes use of US EPA data on emissions
and body burdens of dioxin-like compounds.These compounds are persistent in the environment and within biological organisms. Human exposures are predominantly through food. The US
EPA data make possible the use of biomonitorinng data in the form of human lipid levels and food lipid levels to estimate iF for dioxin-like compounds in North America. We use two alternative
methods to evaluate iF and its associated variability and uncertainty-an empirical approach based on measurements and a probabilistic approach using the CalTOX multimedia fate and
exposure model. Here evaluation of the two results provides a process for characterizing uncertainty in the CTI concept. Finally, we use data from the Canadian Environmental Protection Act
Priority Substance List (PSL1) assessments and the CalTOX multimedia model to evaluate how well empirical observation are consistent with interpretations from the model for a broad range
of chemicals. For the 14 substances available in the PSL1, the trend in the empirical and model results confirms our postulated relationship between iF and POV.




SS4-14 DEVELOPING        INTAKE FRACTION ESTIMATES WITH LIMITED DATA: HOW DO
               METHODS COMPARE?
Gretchen Stevens, Harvard School of Public Health, USA
In order to estimate the health benefits of reducing vehicular emissions, analysts must estimate the change in population exposure that results from a given change in emissions. However,
the detailed atmospheric modeling needed to accurately estimate changes in population exposures is often difficult to implement in settings where data are limited or policy decisions are
needed in the short term. Intake fraction, defined as the fraction of a pollutant or its precursor that is inhaled by the population, is a summary metric that has been promoted as a way to
allow exposures to be estimated when detailed modeling is not possible. However, most intake fraction estimates to date have been derived in settings with adequate data, and economic
and technical resources. It remains unclear how to most appropriately estimate intake fractions in settings where those resources may not be available. In this paper, we calculate particulate
matter intake fractions for Mexico City using five methods, some more resource-intensive than others. First, we create a simple box model to describe dispersion of primary particulate matter
in the Mexico City basin. Second, because many variables used by the simple box model show diurnal patterns, we develop a dynamic box model. Third, we extrapolate intake fractions for
primary PM, ammonium sulfate, and ammonium nitrate from U.S. values using a regression model. Fourth, we calculate intake fractions by assuming a linear relationship between emissions
and population-weighted concentrations of primary PM, ammonium nitrate, and ammonium sulfate (a particle composition method). Finally, we estimate particulate matter intake fractions
from detailed atmospheric dispersion and chemistry models run for only a short period of time. Results vary by up to an order of magnitude, from 26 to 260 per million for primary PM. The
particle composition method results in the highest intake fraction values, while the extrapolation from U.S. values yields the smallest values. The former result may be due to underestimates
in the emission inventory or important external sources of emissions, while the latter could be explained by extrapolating outside of the range of observed population values. Using several
methods to calculate intake fraction provides a starting point for evaluating the quality of data inputs used in each method and provides first-order approximations of intake fractions and
their uncertainties. Our results emphasize that formal uncertainty analysis, with special attention to model uncertainty, would be important when using any single rapid-assessment method
in a health benefits analysis.




SS4-15 EVALUATION         OF THE SEASONAL VARIATION OF INTAKE FRACTIONS (IF) FOR THE
               PRIMARY FINE PARTICLE (PM2.5 ) EMISSIONS IN FINLAND FOR VARIOUS SOURCE
               SECTORS
Marko Tainio1, Mika Hujo1, Mikhail Sofiev2, Jaakko Kukkonen2, Ari Karppinen2, Niko Karvosenoja3, Jouni T Tuomisto1.
1. Centre for Environmental Health Risk Analysis, National Public Health Institute (KTL), Finland; 2. Finnish Meteorological Institute (FMI); 3. Finnish Environment
Institute (SYKE).
The evaluation of exposure is a prerequisite for the estimation of the environmental risks. The concept of intake fraction (iF) has been developed in order to create common metrics for different
exposure assessments. Especially in air pollution studies, the intake fraction method has been used in numerous risk studies. In this study, we have assessed the intake fractions for Finnish
primary fine particles. The monthly variation of the iF’s are assessed for different emission sectors. The spatial allocation of emission rates has been estimated on a resolution of 1 km x 1 km
over Finland by the Finnish Environment Institute (SYKE). The atmospheric dispersion of primary particles originated in Finland for the year 2000 has been modeled at the Finnish Meteorological
Institute (FMI) for the whole of Europe. The intake fractions were estimated by combining emission rates, predicted atmospheric concentration data, the data regarding the location of the
European population, and average breathing rate data. The intake fractions were estimated separately for all the months, corresponding to each of the six main emission sectors. All the data
was incorporated to a numerical grid of a resolution of 10 x 10 km over the whole of Europe. Calculations were done with the program R (version 2.1.1.). The iFs exhibited a substantial
seasonal variation. Intake fractions were higher during the winter months (January-March), compared with those during the summer months (May-July). The iFs of all the various emission
sectors showed a similar seasonal variation. The comparison of standard deviations showed that the variation of iFs between emission sectors was more moderate than the seasonal variation.
The highest intake fractions were identified for traffic (mean 7.9E-07), and the lowest for agricultural dust (mean 5.1E-07). The seasonal variation of intake fractions was more pronounced
than that between various emission sectors. The strong seasonal variation is due to (i) the substantial seasonal variation of meteorological conditions, especially in Northern Europe and (ii)
the notable differences of population densities in areas that surround Finland (e.g., pollutant transport to Lapland vs. to the St. Petersburg region). The mean intake fraction was highest in
case of vehicular traffic; this fine particle pollution source category is therefore particularly important for public health.




                                                                                               18
                                                                                                                                                                                                       SUNDAY SEPTEMBER 3
        Symposia Session: SS4

SS4-16 USING        INTAKE FRACTION ESTIMATES FOR ENERGY POLICY: APPLICATION TO
               EXTERNALITY ESTIMATION AND EMISSIONS TRADING
Andrew M. Wilson, Harvard School of Public Health, USA
Intake fraction (iF) research is generally justified by either or both of two arguments: (i) quantitative iF estimates enable useful comparisons of the exposure impact (per unit emissions) within
and across pollution source types; and (ii) use of iF values allows us to leverage information from existing exposure model studies to make streamlined estimates for policy applications that
would otherwise require expensive, time-consuming, and largely repetitive new model studies. To date, most iF work has concentrated on methods and estimates for the first justification. This
study addresses the second justification, as it explores the use of power-plant iFs for two general energy policy applications: externalities and emissions trading programs. This poster will focus
on important insights enabled by iF estimates on the magnitude and variability of externalities from power plant emissions and on the relative efficiency of emissions versus exposure trading
programs. Using a database of iF values for over 500 US power plants, we assess the importance of national mortality impacts of fine particulate matter air pollution to a set of externality
and emissions trading scenarios and discuss how these results relate to those found in the current literature on these energy applications. In doing so, we use iF to connect findings from the
public health and environmental modeling community to issues of current importance in the realm of energy policy.




SS4-17 THE         INFLUENCE OF GEOGRAPHIC LOCATION ON POPULATION EXPOSURE TO
               EMISSIONS FROM POWER PLANTS THROUGHOUT CHINA
Ying Zhou, Harvard School of Public Health, USA
This analysis seeks to evaluate the influence of emission source location on population exposure in China to fine particles and sulfur dioxide. We use the concept of intake fraction, defined as
the fraction of material or its precursor released from a source that is inhaled or ingested by a population. We select 29 power-plant sites throughout China and estimate annual average intake
fractions at each site, using identical source characteristics to isolate the influence of geographic location. In addition, we develop regression models to interpret the intake fraction values
and allow for extrapolation to other sites. To model the concentration increase due to emissions from selected power plants, we used a detailed long-range atmospheric dispersion model,
CALPUFF. Primary fine particles have the highest average intake fraction (1x10-5), followed by sulfur dioxide (5x10-6), sulfate from sulfur dioxide (4x10-6), and nitrate from nitrogen oxides
(4x10-6). For all pollutants, the intake fractions span approximately an order of magnitude across sites. In the regression analysis, the independent variables are meteorological proxies (such
as climate region and precipitation) and population at various distances from the source. We find that population terms can explain a substantial percentage of variability in the intake fraction
for all pollutants (R2 between 0.86 and 0.95 across pollutants), with a significant modifying influence of meteorological regime. Near-source population is more important for primary coarse
particles while population at medium to long distance is more important for primary fine particles and secondary particles. A significant portion of intake fraction (especially for secondary
particles and primary fine particles) occurs beyond 500 km of the source, emphasizing the need for detailed long-range dispersion modeling. These findings demonstrate that intake fractions
for power plants in China can be estimated with reasonable precision and summarized using simple regression models. The results should be useful for informing future decisions about power-
plant locations and controls.




                                                                                                19
        Symposia Session: SS5

SS5            BIOLOGICAL MONITORING TO ASSESS EXPOSURE TO BIOLOGICAL AND CHEMICAL
               WARFARE AGENTS
John Barr

SS5-01 DETECTION        AND DIFFERENTIATION OF PROTEIN TOXINS AND HUMAN PATHOGENS BY
               MASS SPECTROMETRY
John R. Barr, Centers for Disease Control and Prevention
Biological warfare has been used since antiquity and the threat remains today. There are many possible biological warfare and bioterrorism agents. Botulinum neurotoxins, ricin, and Bacillus
anthracis (the bacterium that causes anthrax) are often considered among the top bioterrorism threat agents. We have recently developed mass spectrometry based methods that can detect
and differentiate these important biological warfare agents.
Botulinum neurotoxins (BoNTs) are proteases that cleave specific cellular proteins essential for neurotransmitter release. There are seven BoNT serotypes (A-G), four of which usually cause
human botulism (A, B, E, and F). We have developed a rapid, mass spectrometry-based method (Endopep-MS) to detect, differentiate, and quantify active BoNTs. This method uses the highly
specific protease activity of the toxins with target peptides that are specific for each toxin serotype. The product peptides derived from the endopeptidase activities of BoNTs are detected by
matrix-assisted laser-desorption ionization time-of-flight mass spectrometry. We can also quantify the amount of toxin and determine the subtype of the toxin by digesting the toxin with
trypsin and then sequencing the resulting tryptic peptides by MS/MS.
Bacillus anthracis is a spore forming bacterium that produces two toxins that are composed of 3 proteins. These three proteins are protective antigen (PA), edema factor (EF), and lethal factor
(LF). PA combines with LF to form the binary lethal toxin (LTx). Anthrax LTx is known to cleave members of the mitogen-activated protein kinase kinase (MAPKK) family and is lethal. We have
developed a method that can detect anthrax lethal factor and anthrax lethal toxin by employing a specific antibody to PA or LF and using the specific enzymatic activity of LF coupled with
mass spectrometry to detect the toxin. We have applied this to spiked serum samples, serum from rhesus macaques that had inhalation anthrax.
Ricin is found in the bean of the castor plant, Ricinis communis, and is one of the most toxic and easily produced plant toxins. Ricin enzymatically inactivates protein synthesis via cleavage
of an adenine residue (A4324) of the 28S ribosomal subunit which results in the blockage of elongation factor-2 binding. We have developed methods that can detect ricin in a sample by
employing antibody capture followed by tryptic digestion and mass spectrometry detection of the tryptic fragments. We also have developed a method that can detect an alkaloid, ricinine,
which is specific to the castor plant and is often co-purified with ricin.




SS5-02 DIAGNOSIS        OF OP EXPOSURE BY MASS SPECTROMETRIC ANALYSIS OF HUMAN
               BUTYRYLCHOLINESTERASE ADDUCTS
Daan Noort, TNO Defense, Security and Safety
Introduction. Methods to diagnose exposure to organophosphate anticholinesterases, e.g., nerve agents, can play a pivotal role in case of a terrorist attack with nerve agents. In the same
context, confirmation of non-exposure of worried citizens is of utmost importance. We here present an agent-specific diagnostic method for detection of OP exposure that is based on the
mass spectrometric analysis of phosphylated peptides after pepsin digestion of modified butyrylcholinesterase (BuChE). Furthermore, we present a generic mass spectrometric method, based
on the chemical conversion of the phosphylated peptides obtained after pepsin digestion of BuChE.
Methods. Isolation of human butyrylcholinesterase from plasma is based on procainamide affinity chromatography. Subsequent pepsin digestion under acidic conditions can be performed
without pretreatment of the protein. The digests do not need further purification prior to LC-tandem MS analysis. Phosphylated nonapeptide standards were prepared by solid phase peptide
synthesis, followed by reaction with a suitable phosphorylating reagent.
Results. After exposure of plasma to various organophosphates (nerve agents, pesticides) the expected nonapeptides could be analyzed in the pepsin digests. It was found dat d7-sarin-
exposed plasma could conveniently be used as an internal standard. After treatment of the various digest with Ba(OH)2 in the presence of a suitable nucleophile (a thiol or amine), the
phosphylated nonapeptides were transformed into a common nonapeptide, that could be analyzed sensitively by means of LC tandem MS.
Discussion. The advantages of the presented methods are obvious. It can be envisaged that the generic method can be used as a rough screening method in case of large numbers of samples,
as is to be expected after a terrorist incident. Subsequently, after finding a positive sample, the digest can be analyzed in a more specific way, by reanalyzing the original digest for all kinds
of phosphylated nonapeptides in a more thorough and laborious way. Also, the analysis of phosphylated peptides in enzymatic digests of BuChE surpasses the limitations of the fluoride
reactivation method, since it can also deal with dealkylated (“aged”) phosphylated BuChE. In conclusion, two promising methods for diagnosis or biomonitoring of OP exposure have been
developed, based on LC tandem MS analysis of modified nonapeptides derived from BuChE.




SS5-03 LABORATORY        ANALYSIS OF BIOMEDICAL SAMPLES FOR THE VERIFICATION OF
               CHEMICAL WARFARE AGENT EXPOSURE
J. Richard Smith and Benedict R. Capacio, United States Army Medical Research Institute of Chemical Defense, Analytical Toxicology Division,
Aberdeen Proving Ground, MD 21010-5400
Over the past 15 years, our laboratory has been actively engaged in the development of analytical methods for verifying chemical warfare agent (CWA) exposure in clinical samples such as
blood, urine, and tissue. Most CWAs are rapidly hydrolyzed in aqueous solutions, consequently our earlier methods measured CWA breakdown products resulting from hydrolysis or enzymatic
reactions. The breakdown compounds are relatively non-volatile and required derivatization prior to analysis using gas chromatography/mass spectrometry (GC/MS). We were successful in
developing methods capable of quantitatively measuring the breakdown compounds of several CWA nerve agents and vesicants at the low ng/ml concentration range. The methods were
validated using animal models which demonstrated that the breakdown products could be detected over a period of several days. The primary limitation of this approach was the rapid
excretion of the breakdown compounds in the urine or blood. An alternate verification approach is the analysis of CWA adducts to biomolecules such as proteins. Adducts of biomolecules are
generally stable and provide an opportunity to verify a CWA exposure much longer after an actual event has occurred. Recently, our laboratory has developed a mass spectrometric method
to measure the products formed from the interaction of sulfur mustard with blood proteins. Adducts are formed from the reaction of sulfur mustard with free carboxylic acid sites of plasma
proteins. The adducts, hydroxyethylthioethyl (HETE) esters, could be separated from the protein by the addition of base to release thiodiglycol. The thiodiglycol was derivatized and analyzed
using negative ion chemical ionization GC/MS to produce a very sensitive, quantitative method. Detectable levels in blood were still observed 45 days after exposure in a nonhuman primate
receiving a 1 mg/kg, IV dose of sulfur mustard. The method was also applied to blood specimens obtained from an accidental human exposure to sulfur mustard. We have also analyzed
phosphylated plasma proteins resulting from nerve agent exposure using a fluoride reactivation method which regenerates the parent nerve agent. It has been evaluated as a quantitative
method for low dose exposure to the nerve agents sarin and soman. In a guinea pig model, both plasma and tissue samples generated abundant levels of reactivated nerve agent following
a series of low dose, subcutaneous exposures. Reactivation levels were found to decrease rapidly after time of exposure, although they were still measurable after three days. In postmortem
samples, the rate of decrease was found to be much slower. Taken together or in part, the analytical methods presented here provide a mechanism for verifying a chemical warfare agent
exposure in biomedical samples such as blood or urine.




                                                                                               20
                                                                                                                                                                                               SUNDAY SEPTEMBER 3
        Symposia Session: SS5

SS5-04 URINARY                    METABOLITES AS INDICATORS OF EXPOSURE TO CHEMICAL WARFARE
AGENTS
Robin Black, Robert Read and James Riches, Defence Science & Technology Laboratory, Porton Down, Salisbury, UK
The detection of urinary metabolites may provide the simplest means of confirming an exposure to a chemical warfare agent. Urinary excretion accounts for the major part of an absorbed
dose. Most metabolites are relatively small molecules, they are readily analysed by GC-MS, usually after derivatisation, or by LC-MS, and most analytical standards are easily obtained. A
disadvantage is that most of the absorbed dose is excreted within a few days of exposure. Urinary metabolites have been well characterised for sulphur mustard and nerve agents, and
hydrolysis products have been identified in urine for Lewisite and nitrogen mustards. Metabolites of sulphur mustard and the nerve agent sarin have been detected in exposed human subjects.
In a few cases, urinary metabolites may not be unequivocal indicators because of the presence of low background levels in non-exposed individuals. A general overview of urinary metabolites
as indicators of CW agent poisoning will be presented, illustrated by examples from our laboratory. Recent additions to our analytical methodology have included the use of LC-MS-MS for
the analysis of beta–lyase metabolites of sulphur mustard, and the use of a simple ion trap mass spectrometer for the GC-MS-MS analysis of thiodiglycol and the hydrolysis products of nerve
agents.




                                                                                            21
        Symposia Session: SS6

SS6            MODELING CHILDREN’S EXPOSURES TO PM, OZONE AND AIR TOXICS
Halûk Özkaynak and Amanda Wheeler

SS6-01 LONG TERM EFFECTS OF AIR POLLUTION ON CHILDREN’S HEALTH: STUDY DESIGN
CHALLENGES TO DISENTANGLE THE GEMISCH
Nino Künzli, MD PhD; ICREA & Institut Municipal de Investigacio Medica (IMIM) Barcelona
Long-term chronic consequences of air pollution are of major public health interest. With increasing evidence for adverse effects of the complex mixture – or Gemisch – occurring already in
childhood, there is growing interest in understanding the contribution of specific constituents and sources to these chronic health effects. Disentangling the Gemisch requires rethinking of
exposure assessment and study designs. A major challenge originates from the substantial differences in spatial characteristics of air pollution constituents and/or sources. As a consequence,
spatial gradients of ambient concentrations and of personal exposure, in particular, largely differ across constituents. For example, the concentration ratio of annual means of PM2.5
characteristics across two locations within the city of Antwerp was 1.06 for sulfur; 1.14 for total mass; 1.22 for redox activity; 1.75 for blackness; 2.14 for Fe, and 2.18 for NO2 (European
Community Respiratory Health Survey). Therefore, residential location and personal time-activity patterns affect long-term exposure to these constituents to very different degrees. This is
confirmed, e.g., in an HEI study using personal exposure measurements among children in Amsterdam and Helsinki. Whereas outdoor and personal concentrations agreed well for total mass
or S, this was not at all the case for other constituents (e.g. Si).
To disentangle effects of the Gemisch, source specific pollutants and/or constituents need to be characterized on the appropriate spatial scale. The increase in the number of pollutants or
constituents monitored at a single location in a community (e.g. “speciation networks”) will not serve the purpose of disentangling long-term effects of sources and constituents on children’s
health in multi-community studies. Instead, chronic effect studies need spatially dense measurements of source specific constituents (or markers thereof), with a particular focus on those of
pathophysiological relevance (e.g. ultrafine particles, metals, redox activity of particles etc.). These measurements are also needed for the development and validation of exposure models for
these constituents.
Investigations of effects of traffic related pollutants on artery wall thickness – a preclinical marker of atherosclerosis – among the South Californian Children’s Health Study will be used as an
example to discuss future exposure measurement strategies.




SS6-02 ISSUES        AND CHALLENGES IN MODELING CHILDREN’S LONGITUDINAL EXPOSURES:
               AN OZONE CASE STUDY
Jianping Xue,Valerie Zartarian, Halûk Özkaynak, USEPA National Exposure Research Laboratory, RTP, NC, USA
Modeling children’s exposures is a complicated, data-intensive process. Modeling longitudinal exposures, which are important for regulatory decision making, especially for most air toxics,
adds another level of complexity and data requirements. Because it is difficult to model inter- and intra- personal variability for exposure model inputs, there is potential for inaccurate
estimation of upper percentiles of longitudinal exposure distributions. In order to develop a scientifically sound exposure prediction model, we need to resolve how to do the following: (1)
obtain longitudinal data needs for time activity and pollutant measurements; (2) separate intra- and inter- person variability of model inputs; (3) link inputs from different data sources and
fit those data as inputs for the model, which will preserve its variance-covariance structure; and (4) use cross-sectional data to simulate longitudinal data, which is difficult and expensive to
collect. In this presentation, we address these issues by applying both an existing and a new technique for estimating personal ozone exposures of school-age children living in two California
communities. The first modeling analysis employed commonly used methods for estimating exposures using a microenvironmental exposure model that used independent distributions in
simulations, which were fit to observed inputs for the model. In the new modeling methodology, various variance components derived from the underlying data were put back into the model,
so that proper variance-covariance relationships in model inputs were maintained. Contribution of intra-personal, inter-personal, seasonal, and area’s variances predicted by the new model,
were: 38%, 19%, 39% and 4%, which were quite close to those derived from the original data while intra-personal variance is 91% for the old method without decomposition of the variance.
The standard deviation and 99th percentile of overall personal ozone exposures from the model with the new method and the old method without decomposition of variance of intra, inter
personal and other factors, were: 12 ppb, 53 ppb and 21 ppb, 89 ppb, respectively, and in comparison to 12 ppb, 52 ppb for the observed personal ozone data. Results show that this method
not only can keep variance-covariance structures of inputs and output in a simulation, but also can accurately predict high percentiles of longitudinal exposures that are important for
regulatory applications. Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy.




SS6-03 CHILDREN’S        EXPOSURE TO AMBIENT AIR POLLUTION: THE APPLICATION OF
               DIFFERENT EXPOSURE ASSESSMENT METHODOLOGIES.
Wheeler AJ1, Villeneuve P1, Smith-Doiron M1, Mahmud M1, Dales R1 and Brook JR2.
1
 Health Canada, Ottawa, Canada, 2Environment Canada, Toronto, Canada
The Border Air Quality Strategy (BAQS), an international agreement between Canada and the US, has led to a prioritization of research in Windsor, Ontario. Much of the health and exposure
assessment research has focussed on the potential impact of exposure to ambient air pollution on the health of pre-adolescent children.
In September 2004, Health Canada issued a cross-sectional survey to all children, aged 7-14, enrolled in Windsor schools (n = 20,159) to determine their respiratory health status. Other
respiratory risk factor data collected in the survey included: parental health, housing conditions, and self-reported measures of exposure to traffic. A response rate of 63.8% (n = 12,694) was
obtained. Chronic exposure to a variety of air pollutants is being estimated for these participants using different methods.
On-going air pollution measurements include intra-urban monitoring at approximately 50 locations measuring particulate matter (fine and coarse mode), nitrogen dioxide, sulfur dioxide,
polycyclic aromatic hydrocarbons, volatile organic compounds and acid vapour. These pollutants are being monitored for a two-week period in four different seasons between 2004 and 2007.
Air pollution exposures are also being assessed using personal, indoor and outdoor microenvironmental monitoring at 48 residences during cold and warm seasons throughout 2005 - 2007.
This exposure study has a total of 1440 sample days (480 annually) which will be used to assess infiltration factors and activity patterns which influence personal exposure.
Both of these air pollution monitoring studies are also co-located at both of the Environment Canada National Air Pollution Surveillance (NAPS) sites in Windsor, which are the typical central
monitoring sites used when assigning exposure in health effects research. Chronic exposure will be estimated using historical data obtained from NAPS and compared with results obtained
from the BAQS exposure assessment methods.
While the use of personal monitoring is the preferred method for characterizing an individual’s exposure to air pollution, this method is not always viable in epidemiological studies. The
different methods of assessing exposure for the same population will allow a comprehensive investigation of the sources of exposure measurement error. This will be assessed by conducting
an inter-method reliability analysis. Correlational analyses using the Pearson and Spearman Rank correlation coefficients will identify any association between exposure metrics generated
using these data sources, and moreover, explore whether agreement is differential by health status, age, and socioeconomic status. Findings from this inter-method reliability study will assist
in the application of methods to adjust associations between air pollution and health outcome data for the effects of exposure measurement error.




                                                                                               22
                                                                                                                                                                                                   SUNDAY SEPTEMBER 3
        Symposia Session: SS6

SS6-04 EXPOSURE                       TO POLYCYCLIC AROMATIC HYDROCARBONS IN CHILDREN IN THE UNITED
               STATES
Richard Y. Wang, Samuel P. Caudill, Courtney D. Sandau, Andreas Sjödin, Zheng Li, Lovisa C. Romanoff, Larry L. Needham, Donald G. Patterson.
National Center for Environmental Health, Centers for Disease Control and Prevention.
Introduction: People are exposed to polycyclic aromatic hydrocarbons (PAHs) from various sources, including food, smoking, and air pollution. The purpose of this presentation is to
characterize the exposure to PAHs in children in the United States. Method: A cross sectional survey consisting of persons aged 6 years and older from the US general population participating
in the National Health and Nutrition Examination Survey (2000-2001). Demographic data for children (6 to 11 years), adolescents (12 to 18 years), and adults (> 19 years) were collected.
Urinary levels for metabolites of fluoranthene (3-hydroxyfluoranthene), fluorene (2-, 3-, 9-hydroxyfluorenes), naphthalene (1-, 2- hydroxynaphthalenes), phenanthrene (1-, 2-, 3-, 4-, 9-
hydroxyphenanthrenes), and pyrene (1-hydroxypyrene) were measured by GC/HRMS. These levels were used as an index of exposure to PAHs. Covariates included age, gender, urinary
creatinine, serum cotinine for exposure to tobacco smoke, and race/ethnicity. Statistical comparisons were performed by ANCOVA and alpha was set at 0.025.
Results: The levels for 1-hydroxypyrene at the geometric mean and the 95th percentile for children were as follows: 59.7 ng/L (95CI, 52.8-67.5) and 272 ng/L (95CI, 224-453), respectively.
Differences in levels for PAH metabolites were observed among age groups. In children, the levels for 1-hydroxypyrene, 1- and 3-hydroxyphenanthrenes, and 2- and 3-hydroxyfluorenes were
higher than those for adolescents or adults. Adolescents had lower levels for 9-hydroxyfluorene and 1- and 2- hydroxynaphthalenes compared to other age groups. There were no differences
among age groups for the other PAH metabolites.
Conclusion: Children are exposed to several polycyclic aromatic hydrocarbons. Levels for the metabolites of these chemicals in children can differ in comparison with older age groups. The
extent to which these differences are attributed to variations in exposure or pharmacokinetics will require additional research to define.




SS6-05 ESTIMATING        CHILDREN’S INTAKE OF AIR TOXICS AT REGIONAL SCALES: THE
               CALTOX MODELING AND EVALUATION PROCESS
Thomas E. McKone, University of California, Berkeley and Lawrence Berkeley National Laboratory, Berkeley, CA, USA
Polycyclic Aromatic Hydrocarbons (PAHs) are chemicals derived from incomplete combustion; ubiquitous in both rural and urban environments; and for which human exposure and intake
accumulates from multiple sources, multiple environmental media, and multiple exposure pathways. In previous work, we used the CalTOX regional multimedia mass-balance model to evaluate
and update emissions estimates in rural and urban regions of the state of Minnesota, USA. By comparing model estimates of outdoor PAH airborne concentrations with those reported by the
Minnesota Children’s Pesticide Exposure Study (MNCPES), we probabilistically evaluated the reliability of our emissions estimates for specific PAHs. In this follow-on study we use the CalTOX
model to assess children’s (ages 6 to 15) cumulative intake of PAHs from air, food, and water. We develop a regional model that is applied to archetypal urban and rural regions of the US to
assess intake fraction—the cumulative individual intake (mol/d) within a region divided by estimated emissions to that region (mol/d). As part of its National Health and Nutrition Evaluation
Survey (NHANES), the US Center for Disease Control and Prevention (CDC) has compiled data on levels of mono-hydroxy metabolites in urine for a number of PAHs. We use NHANES 1999-
2000 data to evaluate our model estimates of PAH individual intake fraction in both urban and rural environments. Based on our previous work for urban and rural regions of Minnesota, we
develop emissions factors for PAHs that are empirically linked to population. Due to model and parameter uncertainty the CalTOX results have a large uncertainty. But the CalTOX results
indicate that the intake fraction depends on the persistence of the pollutant and the degree of bioaccumulation in vegetation and animals. Individual intake fractions are in the range of 10-
14 to 10-12, with variations due to both the chemical properties of the PAH and the extent to which food pathways account for cumulative intake. We use the NHANES biomarker data to
confirm or refute the intake fraction estimates obtained from the model. The NHANES results provide a second but still highly uncertain estimate of individual intake fraction. However, the
intake fractions implied by the NHANES results are consistent with the model results both in terms of the magnitude of specific PAHs and with regard to an apparent trend of increasing intake
fraction with increasing PAH persistence and level of bioaccumulation.




SS6-06 MODELING          DAILY EXPOSURE OF CHILDREN’S EXPOSURE TO PM, OZONE, NO2, AND
               BIOLOGICAL AGENTS IN THE FRESNO ASTHMATIC CHILDREN’S ENVIRONMENT STUDY
               (FACES)
Frederick W. Lurmann, Paul T. Roberts and David L. Vaughn, Sonoma Technology, Inc., Petaluma, California, USA
S. Katharine Hammond, Charles Perrino, Mervi Hjelmroos Koski, Ira Tager, University of California, Berkeley, California, USA
A multi-pollutant personal exposure model has been developed for assigning daily exposures to children participating in the Fresno Asthmatic Children’s Environment Study (FACES). This
presentation describes the methods and results for estimating children’s personal exposure to ozone, NO2, PM2.5 mass, elemental carbon (EC), organic carbon, PM10 mass, endotoxin, agricultural
fungi, Cladosporium, Alternaria, total fungal spores, and total pollens grains. The modeling approach relies on a central air monitoring station measurements to characterize the day-to-day
temporal variations in outdoor concentrations, spatial mapping of pollutant concentrations measured (at various times) at 85 locations in the community to characterize intra-urban gradients
in ambient concentrations, and indoor-outdoor measurements made in 80 homes to guide selection of parameters for the mass balance equation used to estimate indoor concentrations.
Subject-specific home survey and daily human activity questionnaire data are used in the microenvironmental exposure model. The principal findings are:
1) On most days, personal exposure estimates vary between subjects by a factor of two for PM2.5 mass and by a factor of three or more for other pollutants considered here. The large variations
   between subjects in estimated exposure to pollutants of ambient origin suggests that the use of central site ambient concentrations alone for individual exposure assignments in FACES
   may result in considerable exposure misclassification.
2) The between-subject variations in personal exposure estimates are generally greater for biological agents than conventional pollutants, and greater for primary pollutants (such as EC) than
   secondary (ozone) or combined primary/secondaryd pollutants (PM2.5). For example, the personal exposure estimates may range from 100 to 800 total pollen grains/m3 and from 10 to 250
   Alternaria spores/m3 on one day, which are much larger ranges than are estimated for conventional pollutants.
3) The mean estimated personal exposure concentrations of pollutants of ambient origin are consistently lower than the ambient concentrations measured at the FACES central monitoring
   station. On average, the mean personal exposure concentrations range from 15% of central site ambient concentrations for total pollen grains to 59% of central site ambient concentrations
   for PM2.5 mass. The pollutant ranking (from highest to lowest) for mean ratio of personal exposure to central site ambient concentrations is PM2.5 mass, endotoxin, EC, agricultural fungi,
   NO2, PM coarse, Alternaria, ozone, Cladosporium, and total pollen grains. The systematic variations in personal exposure levels relative to ambient concentrations are primarily a result of
   lower indoor than outdoor concentrations and secondarily a result of spatial differences in ambient concentrations within the community.




                                                                                              23
        Symposia Session: SS6

SS6-07 PERSONAL,         RESIDENTIAL, AND CENTRAL SITE PM2.5, CO, NO2, AND VOCs FOR A
               COHORT OF PREGNANT WOMEN IN TRUJILLO, PERU
Luke P. Naeher, Ph.D.1 Brandon Cassidy, B.S.1, P. Barry Ryan, Ph.D.2 Charlene W. Bayer, Ph.D.3 Manuel Aguilar-Villalobos, M.S.4
1
  The University of Georgia, College of Public Health, Athens, GA, USA; 2Emory University, Rollins School of Public Health, Atlanta, GA, USA; 3Georgia Technology
Research Institute, Atlanta, GA USA; 4Asociacion del Aire Ambiental, Lima, Peru
Objectives: In this study, we measured personal, residential, and central site (urban and rural) measures of particulate matter ≤ 2.5 um (PM2.5), carbon monoxide (CO), nitrogen dioxide (NO2),
and select volatile organic compounds (VOCs), in a cohort of 99 pregnant women from Trujillo, Peru during their first trimester of pregnancy.
Methods: With the help of Trujillo City Hall, pregnant women were recruited voluntarily for this study in their first trimester in April-July 2004. Separate 48-hr personal, kitchen, and main
living area air sampling for PM2.5, CO, NO2 and VOCs was done in the first or early second trimester. A detailed personal time activity diary for the subject was done in parallel with this 48-hr
air sampling, along with a detailed background and descriptive questionnaire for each subject and their residence. Central site measures of PM2.5, CO, NO2 and VOCs were collected for the full
study period at an urban (downtown Trujillo) and rural (local airport) location. Air sampling equipment included SKC pumps and BGI Triplex cyclones with 2.0µm pore size Teflon filters (PM2.5),
Dräger Pac IIIs (CO), Palmes Tubes (NO2) and stainless steel tenax tubes (VOCs). Real-time PM2.5 measures (DusTrak) were also collected at the central sites. Subjects were grouped based on
the type and number of cooking fuels they used: natural gas, wood, kerosene, carbon briquette, >1 fuel not including natural gas, and >1 fuel including natural gas.
Results: Natural gas-only kitchens had the lowest or near lowest 48-hr area PM2.5 [57.2±85.1µg/m3 (average ± standard deviation); n=32], NO2 (13.2±13.8ppb; n=31) and CO (1.44±2.3ppm;
n=32). Wood-only kitchens had the highest 48-hr personal PM2.5 (141.9±102.8µg/m3; n=15] and NO2 (13.49±11.25ppb; n=17), and carbon briquette-only kitchens had the highest 48-hr
personal CO (1.80±2.49ppm; n=13). Kerosene-only kitchens contained the highest 48-hr area toluene (209.6±235.0µg/m3; n=2), ethylbenzene (12.88 ±13.65µg/m3±; n=2), m-/p-xylene
(39.82±38.66µg/m3; n=2) and o-xylene (15.60±16.76µg/m3±; n=2), while carbon briquette-only kitchens had the highest 48-hr area benzene (26.17±43.07µg/m3; n=7). Daily (PM2.5) and
weekly (NO2 and VOCs) air pollutant levels at the downtown monitoring station (PM2.5: 27.6±10.6µg/m3; n=86) (NO2: 9.0±8.6ppb; n=13) (benzene: 16.3±9.0µg/m3; n=12) (toluene
176.1±242.7µg/m3; n=12) (ethylbenzene 16.0±8.1µg/m3; n=12) (p-/m-xylene: 77.0±37.9µg/m3; n=12) (o-xylene: 24.8±12.2µg/m3; n=12) were consistently higher than those at the airport. The
relationship between personal, kitchen, main living area, downtown Trujillo, and airport measures of PM2.5, CO, NO2 and VOCs will be explored.
Conclusions: The use of highly polluting-fossil fuels resulted in elevated kitchen area concentrations within residences and personal exposures for those living within the homes.




SS6-08 BIOMASS         SMOKE EXPOSURE AMONG GUATEMALAN INFANTS PARTICIPATING IN A
               RANDOMIZED TRIAL OF CHIMNEY STOVES
John McCracken1, Anaite Diaz2, Eduardo Castro2, Rufus Edwards3, Louise Ryan1, Joel Schwartz1, Zohir Chowdhury,4 Kirk R. Smith4
1. Harvard University School of Public Health, 2 Universidad del Valle de Guatemala. 3. University of California, Irvine, 4. University of California, Berkeley.
Objectives: As part of a randomized trial to study the effect of indoor air pollution on acute lower respiratory infections (ALRI) among Guatemalan infants, we aimed to assess the effectiveness
of the plancha improved wood-burning stove as an exposure-reduction intervention and estimate child-specific longitudinal exposures during the first 18 months of life.
Methods: 534 households with young infants participated in the trial, with half assigned to the intervention arm and half continuing to use their traditional open fire until the end of the trial.
Personal carbon monoxide (CO) exposures were used as a proxy for biomass smoke exposures as there were no other important local sources of CO. Gastec passive diffusion tubes were worn
by the infants for 48 hours roughly every three months until they were 18 months old. 2190 measures were obtained among 524 of the children. Mixed models with random intercepts for
child were used to account for correlation among repeated measures and to estimate child-specific exposures based on the “best linear unbiased predictor” (BLUP). Since the data were
lognormally distributed, log transfomed exposures were used in regression models and the geometric mean (GM) and geometric standard deviation (GSD) were used to summarize the
distributions.
Results: At baseline before intervention, the control and intervention arms had similar child CO distributions, 2.4±2.4 ppm and 2.5±1.8 ppm, respectively. In the control arm, there was a small
reduction to 2.0±2.2 ppm during the trial, which may be attributed to changing time-location patterns associated with age. However, a much more dramatic reduction in exposure to 1.0±2.3
ppm was observed in the intervention arm. Adjusting for gender, age, season, and weekday, the mixed model estimated a 44% reduction (95%CI: -48 to -39%) in child CO associated with
the improved stove intervention. Younger age, female sex, and the rainy season were associated with slightly higher exposures. The intraclass correlation coefficient (ICC) was 0.28, indicating
the limited reliability of a single measure as a surrogate for long-term exposure.
Discussion: The plancha stove provided large reductions in child CO exposure. The low ICC indicates that most of the variability among the data is within-subject rather than between-subject.
We will examine this issue in light of the repeated measures on each child and the observed predictors of exposure and consider the implications for child-specific predictions using the BLUP
and attenuation of the hazard ratio for the effect of biomass smoke exposure on infant ALRI.




SS6-09 CHANGE          OF MODELLED EXPOSURE TO TRAFFIC RELATED POLLUTANTS DURING THE
               FIRST 6 YEARS OF LIFE IN THE MUNICH BIRTH COHORTS (GINI UND LISA)
Verena Morgenstern and Joachim Heinrich, GSF-Institute of Epidemiology, Neuherberg, Germany
There is a strong need for individual based exposure assessment to air pollutants. In particular for long-term exposure studies on health effects of traffic related air pollutants (TAP), data from
geographic information systems (GIS) are useful. GIS data could be used to calculate distances to major roads and to model exposure to traffic related air pollutants by complex modelling
techniques. In the framework of the EU funded TRAPCA project such regression models were developed and successfully applied to the residential addresses of the members of our birth
cohorts. Since families are moving, in particular when families are growing, the effects of moving on life long exposure estimates to TAP or changed exposure levels over time might improve
total exposure assessment of the cohort members. Within an ongoing birth cohort study in the metropolitan area of Munich we aim to analyse the change of the GIS-based modelled exposure
to TAP during the first 6 years of life. The children of the two ongoing birth cohorts GINI and LISA in the metropolitan area of Munich, which were recruited between the years 1995-98, will
be used to study effects of long term exposure health effects to TAP. A total of 3586 newborns were recruited in this area and 86% could be followed-up until the age of 6 years. During the
follow-up period of 6 years 1390 children had moved at least once (39%) within the Munich metropolitan area and of 710 children residential addresses were available at the age of 6 years.
Residential addresses of the newborns were geocoded and GIS data on street type (as a proxy for traffic volume), land coverage, population, and household density were available in order
to calculate distances between residential addresses and major roads as well as estimates of concentrations of TAP at the residential addresses on the bases of the complex model.
Within the group of movers between birth and 6 years the distance to major road at birth and at 6 years was moderately correlated (r=0.38). Also the modelled exposure to PM2.5 and black
smoke was moderately correlated with r=0.42 and r=0.73. For NO2 the correlation was very low (r=0.10).
However, there was as well a tendency to wider distances to major roads and to lower exposures to TAP, especially for NO2, at the age of 6 years compared to the results from the residential
addresses at birth. These data indicate that young families lower their exposure to TAP by moving into the suburbs of the metropolitan area.




                                                                                                24
                                                                                                                                                                                                           SUNDAY SEPTEMBER 3
         Symposia Session: SS6

SS6-10 MODELING          CHILDREN’S EXPOSURES TO DIESEL EXHAUST FROM COMMUTING BY
               DIESEL SCHOOL BUSES BEFORE AND AFTER ENGINE
L.-J. Sally Liu. University of Washington, Seattle, USA. University of Basel, Institute of Social & Preventive Medicine, Switzerland.
Each weekday in the United States, more than 24 million children ride a diesel school bus to and from school. Diesel exhaust (DE) has been associated with worsening asthma and allergies,
among other important health effects. Reducing DE exposures among children has become a major regulatory initiative, with federal, state, and local jurisdictions investing billions of dollars
in retrofitting diesel engines in school buses. The retrofitting programs typically involve installing diesel oxidation catalysts (DOC) or diesel particle filters (DPF), with the crankcase ventilation
case (CVC) recently being validated for application. An important question remains unanswered: what magnitude of impact will this major environmental policy have on the exposure of
children who commute by diesel bus? This study assesses children’s exposure and health effects from DE before and after the retrofit in school bus fleets. This 5-year study includes 400
asthmatic commuters (ages 9-11) in the Seattle area. Children are recruited from the Seattle and Tahoma School Districts to represent 3 exposure scenarios: 1) 150 children riding newer diesel
buses that will be retrofitted with DOC, CVC, and DPF sequentially; 2) 200 children riding older diesel buses, which will be retrofitted with CVC and DPF after the first and second year of the
study, respectively; and 3) 50 children riding cars throughout the study as the control group. Exposure monitoring is performed on a subset of 100 subjects and the school buses they ride each
year. These personal and on-bus exposure measurements include ultrafine PM (using real-time P-TRAK), PM2.5 (real-time active pDR), black carbon (through light absorbance), PAHs (real-time
EcoChem), and NO2 (active TEA tube). In addition, we used a dual-tracers method (iridium for tailpipe exhaust and deuterated alkane for crankcase emission) and an on-road dilution tunnel
approach for source sampling to quantify and characterize the self-pollution from the buses in a pilot study. Source apportionment analyses and modeling are performed subsequently to
distinguish children’s DE exposure to the bus’s self-pollution from other on-road sources. Exposure models are developed to predict individual children commuters’ monthly and annual
exposures to total DE and bus’s self-pollution while commuting, respectively. This paper compares accuracy and issues of various modeling approaches for the children population.




SS6-11 LONG         TERM PERSONAL EXPOSURE TO TRAFFIC RELATED AIR POLLUTION AMONG
               SCHOOL CHILDREN, A VALIDATION STUDY.
Sofie Van Roosbroeck*, José Jacobs*, Nicole A.H. Janssen**, Gerard Hoek* and Bert Brunekreef*.
* Environmental and Occupational Health Group, IRAS, Utrecht University, Utrecht, The Netherlands.
** Center for Environmental Health Research, National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands.
Introduction: Several recent studies suggest an association between long-term exposure to traffic-related air pollution andhealth. Most studies use indicators of exposure, such as traffic
density on the street of residence or outdoor air pollution as a measure of exposure. Very little information is available about the validity of these measures as an estimate of long-term personal
exposure to traffic-related air pollution. In this study, we therefore assessed outdoor and personal exposure to traffic-related air pollution in locations with different degree of traffic intensity.
Methods: Personal exposure to traffic-related air pollution was monitored in children (10-12 year), attending four different schools with varying proximity to traffic in Utrecht. Subjects wore
personal PM2.5 monitors for four 48-hour periods spaced over a 9-month period.
Personal monitors are carried in small backpacks and simultaneously, Ogawa passive samplers are attached as a badge to the children’s clothing to provide personal NOx measurements. In
addition to PM2.5 mass, absorption of PM2.5 filters is measured as a marker for diesel soot (EC).
Results: Fifty-four children enrolled into the study. Of these, 15 attend a school near a highway (97,800 vehicles/24h), 11 attend a school near a busy local road (45,200 vehicles/24h), and
28 attend an urban background school. Due to pump failure and sickness of some participants, 34 measurements were lost, resulting in 182 complete personal measurements.
Personal and outdoor traffic-related long-term exposures were higher in schools near a busy road. Children attending the school with the highest traffic intensities were also exposed to the
highest personal concentrations. Personal exposure to ‘soot’, PM2.5, NO, NOx and NO2 was respectively 30%, 12%, 52%, 31% and 7% higher for children attending schools near busy roads
compared to children attending schools at an urban background location.
Results of the study show that personal exposure to traffic-related air pollution was significantly higher for children attending schools near busy roads.




SS6-12 COMPARISON         OF ULTRAFINE TRAFFIC PARTICLE EXPOSURES OF PUPILS
               COMMUTING TO SCHOOL BY CAR OR WALKING
Otto Hänninen*, Tarja Yli-Tuomi, Nick Hodges, Janet Robinson, John Gulliver, David Briggs, and Matti Jantunen, * Kansanterveyslaitos (KTL), Finland
Introduction. Traffic congestion occurs frequently in school areas especially during the morning hours when many parents drive their children to school by car, causing problems in the traffic
flow, and increasing injury risks and air pollution exposures of the children, drivers, and residents. Ultrafine particles are a good indicator of local traffic and are suspected to cause health risks
due to their combustion origin and small size. To support development of science based environmental policies and to protect the public from unnecessary environmental risks, the current
work evaluates the effects of commuting to school by walking versus by car to the exposures of school children to these particles.
Material and methods. Ultrafine particle measurements and a database of 1222 journeys of 374 pupils during 422 diary days was used in the modelling. 651 trips were done to or from
school. Commuting by car and walking represent together 94 % of school journeys, rest being divided rather evenly between school buses, taxis, buses, and cycling. School distances of 0.5
to 2 km representing 45% of pupils were selected for the current work to compare situations where the parents have a realistic choice for taking their children to school. Daily exposures of
the children living within this range from the school were modelled for the two commuting options.
Results. Average home-school distance in the 10 schools in Leicester for all pupils living 0.5 – 2 km from their school was 0.9 km, amounting to daily commuting distance of 1.8 km.
Commuting this distance by car (average speed 30 km h-1) results in 3.6 minutes and by walking (speed 3 km h-1) 36 minutes daily (0.25 and 2.5 % of daily time, respectively). Daily average
exposures to ultrafines are increased from 2900 to almost 3500 particles per cubic centimetre, representing a 15% raise.
Discussion and conclusions. For short distances the use of a car does not provide a benefit over walking, and any policy supporting not using cars for overly short distances can be highly
recommended to avoid unnecessary traffic in the congested school areas. For long distances the attractiveness of walking as a commuting option reduces quickly.
Time spent in traffic increases ultrafine particle exposures rapidly. Children living in distances from school suitable for both walking and being transported by car would experience 15% higher
exposures if walking. Moreover, inhalation rates are higher during walking, leading to increased intake of particles. Filtration of particles by the car ventilation system was not accounted for.




                                                                                                  25
        Symposia Session: SS6

SS6-13 ISSUES       AND CHALLENGES IN MODELING CHILDREN’S LONGITUDINAL EXPOSURES:
              AN OZONE CASE STUDY
Jianping Xue1, Halûk Özkaynak1, Valerie Zartarian1, John Spengler2
1
 USEPA National Exposure Research Laboratory, RTP, NC, USA; 2Harvard University School of Public Health, Boston, MA, USA
Development and Evaluation of a Model for Estimating Long-term Average Ozone Exposures of Children
Long-term average exposures of school-age children can be modelled using longitudinal measurements collected during the Harvard Southern California Chronic Ozone Exposure Study over
a 12-month period: June, 1995-May, 1996. The data base contains over 200 young children with personal, indoor, and outdoor data, as well as time activity and other questionnaire information
collected over a whole year period. This large data base was analyzed and used to fit a set of hierarchal long-term ozone models, ranging from simple to more complicated. Other relevant
data such as central-site ozone measurements were also used in fitting the exposure models. The long-term average ozone exposure models were then evaluated against a subset of the data
that were set aside for model evaluation. Results from these comparisons show that the model with both indoor and outdoor ozone variables had the best predictive power, with slope 0.95
and R2 0.91 for Upland, and 1.07 and 0.93, respectively for the Mountain area. The model with using only the indoor ozone variable also had high predictive power. In contrast, the slope and
R2 based on comparison of predictions from the ozone model just using the outdoor ozone data collected by at the central stations, with measured exposures, were 0.98 and 0.78 for Upland
and 1.11 and 0.78 for the Mountain area, respectively. Using time activities to weight indoor and/or outdoor ozone reduces model prediction for the Upland area but has no impact on the
predictions for the Mountain area. The slope and R2 based on prediction of the ozone model just using time activities were 0.51 and 0.17 for Upland and 1.2 and 0.50 for Mountain area. We
conclude that the model based on using the more readily available central site ozone measurements can not reliably predict the weekly profile of annual exposures for each individual, but
may predict well long-term average population exposures. Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy.




                                                                                            26
                                                                                                                                                                                                       SUNDAY SEPTEMBER 3
        Symposia Session: SS7

SS7            HUMAN TRACKING TECHNOLOGIES FOR EXPOSURE ANALYSIS
Kai Elgethun, Texas A&M University, USA; Steen Solvang Jensen, National Environmental Research Institute, Denmark

SS7-01 INTEGRATED       GPS/POLLUTION SENSOR DATALOGGER FOR PERSONAL EXPOSURE
               MEASUREMENTS OF CHILDREN
Kai Elgethun PhD MPH, *Michael Yost PhD MS
Department of Geography & School of Rural Public Health, Texas A&M University, *Department of Environmental & Occupational Health Sciences, University of
Washington
Personalized exposure assessment methodology is evolving rapidly with the miniaturization of sensors and the widespread availability of location based services such as global positioning
systems (GPS). GPS receivers can be used as tracking devices to record people’s time-location and linear velocity. Velocity (when not travelling in a vehicle) in turn can be used as a metric for
exertion level and thus as a weighting factor for breathing rate. Velocity above 10 m/s can be attributed to vehicle travel and therefore excluded from exertion estimates. Using differentially-
corrected GPS (dGPS), high resolution (±3m RMS) near-continuous human time-location data can be collected to replace traditional self-report categorical data for characterization of
exposures within and between microenvironments. It may be argued that dGPS data even permit time-location and exertion to be attributed to specific ‘nanoenvironments’ in areas where
GPS reception is consistent. The aim of this project is to generate personal inhalation time-location-dose profiles that capture dynamic air pollution exposure patterns of children living in urban
areas. Toward the further refinement of exposure profiles, particularly for transient peak exposures, it is logical to synchronize dGPS time-location measurements with those from a real-time
air contaminant microsensor (using time steps as short as 10 seconds). The GPS Personal Acquisition Logger (GPS-PAL) dGPS instrument (Elgethun et al. 2003, 2006) is being modified to
integrate interchangeable air pollution microsensors. The logger has pre-existing input slots and available memory for this purpose. The modified device, dubbed the GPS Personal Exposure
Tracker (GPS-PET), will simultaneously record time, location, and contaminant concentration in the breathing space of children, with the sensor mounted to the lapel of a lightweight vest. The
unit is lightweight (approx. 250 g) and tamper-resistant and was designed for children, though it has obvious utility for adults as well. The advantage of an integrated GPS datalogger over
multiple instruments is twofold: 1) it is less heavy and less cumbersome, thus reducing burden on subjects; and 2) it records a single time-stamp for both location and concentration, thus
reducing effort and potential error introduced by post hoc synchronization.




SS7-02 A         GIS-GPS MODELLING SYSTEM FOR PERSONAL EXPOSURE TO TRAFFIC AIR
               POLLUTION
Steen Solvang Jensen, Senior Scientist, PhD, Department of Atmospheric Environment, National Environmental Research Institute, Denmark
The presentation will focus on a human exposure modelling system for traffic air pollution developed by the National Environmental Research Institute (NERI) for use in Danish air pollution
epidemiological studies and human exposure studies.
The model system has a high spatial resolution (address) and high time resolution (hour).
The system is named AirGIS and integrates dispersion models, available digital maps (buildings, roads, address points), Danish national administrative databases on buildings and populations,
traffic data, and a Geographic Information System (GIS). The Danish Operational Street Pollution Model (OSPM) is used to calculate exposures. One of the unique features of the AIRGIS system
is that it is able to estimate required street geometry parameters for the OSPM model in an automatic way.
Apart from modelling exposure at address level the system is also able to model exposure under transport provided tracking data is available. Therefore, a system has been developed for
tracking people using Global Positioning System (GPS).
A commercial mobile phone (Benefon Track Pro) is used that has a built-in GPS. SMS sent from the mobile phone is used for transferring positions of a study subject e.g. every 20s to a computer.
A tracking program developed by NERI collects the GPS data.
A developed dynamic area-tracking feature ensures that SMS is only sent when the subject is moving (to save costs on sending SMS).
The movement of a study subject can be tracked on digital maps in real time for on-line monitoring in field studies.
The application of the AIRGIS system will be demonstrated based on a study that included tracking of study subjects in Copenhagen, measurements of personal exposure, and modelling of
personal exposure using AIRGIS. Comparison of modelled and measured data will be presented for nitrogendioxid (NO2).




SS7-03 TRACKING                       HUMAN EXPOSURE TO ULTRA FINE PARTICLES IN COPENHAGEN USING
               GPS
Martin Hvidberg, Researcher, Department of Atmospheric Environment, National Environmental Research Institute, Denmark
Presentation of a research project involving simultaneous tracking of study subjects via GPS and air quality measurements of particle numbers.
Research indicates that the ultra-fine fraction of particles poses the largest risk to human health. A large number of these particles are found in an outdoor urban environment and originates
mainly from traffic. Vehicle exhaust contains ultra-fine particles. The emission depends on traffic density, the mix of vehicle classes and fuel types together with the vehicle age distribution
and driving pattern.
The project is specifically looking at particle number concentration measured with hand-held CPC equipment.
The presentation will be focusing on map-matching of GPS data as input to a personal exposure model, and on comparison of measured and modelled exposure to particle number
concentration. The project includes personal exposure measurements of persons riding on bicycle along a predefined route in the city centre of Copenhagen as well as data from a passenger
car driving along a predefined route between suburban areas and downtown Copenhagen.
Due to the moderate quality of the GPS measurements presumable due to the low sky view factor in the narrow streets in Copenhagen, the spatial accuracy of the tracking points are not
always suitable to generate the exact route by use of simple proximity assumptions. Individual points may be off by 100m or more from the street that was actually followed.
This can cause some map matching routines to create complex and unrealistic routes.
It is the intention to try a number of map matching methods to evaluate their capability of predicting the correct route from data with high spatial uncertainty and low auto correlation. The
correct route is known and can, by a human operator, be deduced from the flow of points from the GPS.
An analysis of the GPS points and their location compared with the correct route forms the basis of an analysis of which parameters that influence the GPS tracking when it goes wrong.
Potential errors are: GPS equipment quality, low Sky View Factor (SVF), high travel speed, sharp turns around street corners etc.




                                                                                                27
         Symposia Session: SS7

SS7-04 USING        GPS AND GIS TO ASSESS THE EXPOSURE OF TEENAGERS TO TOBACCO,
               ALCOHOL AND FAST FOOD OUTLETS IN CHENGDU, CHINA
Michael Jerrett, PhD; Myles Cockburn, PhD; Guanjun Feng, PhD; Ping Sun PhD; Chih Ping Chou, PhD; Ning Chen MSc; Paula Palmer PhD; James Pike,
PhD; Andy Johnson, PhD Department Preventive Medicine, University of Southern California
This paper details a pilot study undertaken in Chengdu, China, in the summer of 2005. The study focuses on the spatial aspects of assessing teenagers’ exposure to tobacco, alcohol and fast
food outlets by using GIS. 59 middle school students (age 12-14) were selected from a large population cohort known as the China Seven City Study (CSCS). We trained 30 data collectors
from a local medical school to use GPS to collect the home addresses of these students. We also developed a protocol for marking a representative sample of the tobacco, alcohol and fast
food outlets. Data collectors filled in an information sheet about each outlet, detailing the store type, the visibility of the store, and the signs the store used for tobacco, alcohol and fast food
advertisements. Once the data were imported into ArcGIS (version 9.0), a buffer was done to each student’s home address to include the area within distances of 250, 500, 750 and 1000
meters. A spatial join between the buffered areas and the stores gives us the statistics of how many stores are located within each student’s address within the corresponding distance, as
well as the distance between each student’s home and the nearest store. The analysis was augmented with a GeoStats Data Logger to measure likely exposures in routes frequented by the
teenagers. A professional videographer followed the routes to document the landscape time-synchronized to the GPS position.
The GPS recorded locations with generally more than 35 m accuracy. In total, over 1100 store locations were marked with the GPS. With the results entered into the GIS, we quantified objective
and more precise assessment of the teenagers’ exposure to tobacco, alcohol and fast food outlets. The density of tobacco stores for these teenagers is high in that in the 1000 meters’ range,
the mean number of tobacco stores is 33. The GeoStats data logger functioned well, and the results suggest high exposure to unhealthy products in routes throughout the neighborhood.
Video of the landscape offers promise as technique to cross-validate assignment of exposure and to track temporal changes in the neighborhood structure of this rapidly evolving urban system.
The largest challenge was obtaining high quality geospatial data to represent the urban landscape digitally. This pilot study is a first step towards developing a complex model for the CSCS
that examines the relationship between teenagers’ exposure to these outlets and their addictive behaviors and obesity.




SS7-05 INDOOR                    LOCATION TRACKING USING BLUETOOTH PROXIMITY BEACONS
Paul N. Kizakevich, Michael McCartney*, Ann Zhang, Robert Furberg, Steve Duncan, Roy Whitmore
RTI International, Research Triangle Park, NC 27709, *MLM Technical Services, Durham, NC
Introduction: The purpose of this work is to develop methods for collecting longitudinal data on human exposure- related activities. Methods and instruments are being developed for a broad
range of data types: activity, location, energy expenditure, environmental conditions, diet, and use of three types of consumer products – pesticides, cleaning products, and personal care
products. Integration of data streams through a common ambulatory platform is integral to our strategy for collecting information all day, each day for a week. Our goal is to achieve low
enough participant burden that people will sustain participation in longitudinal studies for 1 week in each quarter of the year.
Methods: To relieve the participant burden of noting presence in residential microenvironments, we developed an automated monitoring system for indoor location tracking using Bluetooth
wireless technology. Autonomous beacons are placed in key residential microenvironments (e.g., kitchen, bedroom, bathroom, garage, vehicle) and range-adjusted to the perimeter of each
room. The range is determined according to room size, beacon placement within the room, and potential interference with adjacent rooms. Throughout the 7-day monitoring session, each
beacon periodically tests (e.g., 30 second intervals) whether the participant’s Pocket PC is within range. When the participant’s Pocket PC is detected, the beacon sends a time-stamped
microenvironment location code to the Pocket PC for data logging. The resulting data log records a week-long history of the time spent in each residential microenvironment. Controlled bench
tests are being conducted to establish reliability, longevity, sensitivity and specificity of the tracking methodology. Observational field tests are being conducted which by placing beacons in
investigator’s homes and logging automated and manually-affirmed presence in various rooms throughout the day. A pilot field test of our overall system, including location tracking, will be
conducted during the Spring of 2006. Five pilot test subjects will be randomly assigned to each of eight experimental treatments. Each subject will participate for 7 consecutive days.
Results: Results will be available after the controlled bench, observational field, and pilot field tests have been conducted in the Spring of 2006.
Discussion and Conclusions: Preliminary results have shown that automated detection is closely aligned with manual data. Technical elements of the system design will be presented, along
with results of the various test methodologies. Results from this project will determine a set of collection methods that will produce accurate estimates of human exposure-related activities
in future longitudinal studies.




                                                                                                 28
                                                                                                                                                                                                           SUNDAY SEPTEMBER 3
         Oral Communication Session

SM3-O                       CHILDREN’S HEALTH AND ENVIRONMENTAL CHEMICALS
SM3-O-01                    INCORPORATING EARLY LIFE SUSCEPTIBILITY INTO RISK ASSESSMENT: THE
                            EXAMPLE OF EPA’S GUIDANCE ON EARLY LIFE EXPOSURE TO CARCINOGENS
T WOODRUFF1
(1) US Environmental Protection Agency, San Francisco.
Early-life is a time of toxicokinetic and toxicodynamic changes, which can increase susceptibility from exposure to carcinogens. Ability to ascertain susceptibility is limited by available data,
which comes primarily from animal studies, where dosing typically begins 6-8 weeks after birth or from occupational adults studies. To assess whether exposures during childhood could result
in increased cancer risk later in life, a review of the existing scientific literature was undertaken and studies providing sufficient information to quantitatively assess early-life susceptibility to
carcinogens were analyzed. Animal studies were selected for analysis with similar laboratory conditions. Data were available from 27 rodent studies for 18 different chemicals (12 with a
mutagenic mode of action). Susceptibility was calculated as the estimated ratio of juvenile to adult cancer potencies for three study types: acute dosing; repeated dosing; and dosing over the
animal’s lifetime. For mutagenic chemicals, a geometric mean ratio of early life to adult cancer potency of 10 for more chronic exposures was found. For acute dosing, the geometric mean
ratio was 1.5. However, there was considerable variability among tissue types for the acute dosing studies, with geometric mean ratios >1 for kidney (1.6), leukemia (5.9), liver (8.1), lymph
(1.8), reticular tissue (6.5), thymic (2.8), lymphoma (1.8), and uterus/vagina (1.6). Ratios were < 1 for forestomach (0.076), harderian gland (0.48), ovaries (0.033), and thyroid (0.05). Examples
of potentially unique windows of susceptibility include: mammary tumors, exposure during rat’s puberty gave a ratio of 7.1 compared to 0.071 from earlier in life exposure; and nerve tumors,
exposure right after birth gave a ratio of 10 compared to 2.3 for later exposure. Chemicals causing cancer through nonmutagenic modes of action had a geometric mean ratio of 3.4 for
lifetime, 2.2 for repeat exposures. Based on these observations and physiological considerations, EPA adopted new guidelines accounting for potential susceptibilities from early-life exposure
to mutagenic carcinogens. In the absence of adequate chemical-specific data, a factor of 10 is applied to the adult based potency estimate for exposures occurring the first 2 years of life. For
ages 2-16, a factor of 3 is applied. No adjustment is recommended for nonmutagenic chemicals and these are evaluated on an individual basis. EPA’s new guidance provides an example of
how specific data on early-life susceptibility can be incorporated into risk assessment. * View of author and not USEPA. (http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=116283).




SM3-O-02                    LONG TERM ORGANOCHLORINE EXPOSURE AND PRECOCIOUS PUBERTY
S KOIFMAN1, P SARCINELLI1, AC ROSA1, IE MATTOS1, IA CARVALHO2, RJ KOIFMAN1
(1) Oswaldo Cruz Foundation, Rio de Janeiro. (2) National Cancer Institute, Rio de Janeiro.
Introduction. In the early sixties, a pesticides factory producing HCH, DDT and other organochlorine compounds located in a rural settlement in a developing country has interrupted its
activity. Thousand tons of pesticides remained in the facilities without major control procedures and were used by the population living around the factory. Several endocrine effects associated
to organochlorine exposure, including precocious puberty, have been point out.
Objectives. To ascertain the distribution of the age of menarche among local women according to organochlorine blood levels.
Methods. A survey was carried out in the contaminated area, blood samples were collected from the universe (1367 residents) and a questionnaire filled with data on personal and lifestyle
(diet, reproductive antecedents, breastfeeding, other) exposures. Blood levels of selected organochlorine pesticides were ascertained by electron capture chromatography. This survey was
approved by the an Ethical Committee.
Results. Overall population median DDE levels in the area was 26.8 ng/mL and median beta-HCH level was 19.1 ng/mL. Women with menarche 9 yr. old or younger showed median DDE levels
of 36.7 ng/ml compared to 16.7 among those 15 yr.or older (p<0.000001). According to beta-HCH, median levels were, respectively, 40 and 18.5 ng/mL (p< 0.000001).
Discussion. An important estrogenic effect associated to DDT has been reported in the literature. Immigrating children from developing countries to Belgium have also been reported as
presenting high DDE levels, which was associated to precocious puberty. Our study seems to support these results, considering that the observed median organochlorine levels were twice
higher among women with menarche 9yr. or younger compared to those above 14 yr. old.
Conclusion. Very high organochlorine median blood levels were observed among women with menarche at an early age in an area contaminated by such pesticides.




SM3-O-03                    PHYSICAL MATURATION AND EXPOSURE TO ENVIRONMENTAL CHEMICALS: A
                            COHORT ANALYSIS
C RUBIN1, S KIESZAK1, A HOLMES1, M MARCUS1, J HERON2, D BARR1, A CALAFAT1, A SJODIN1, M MCGEEHIN1, R JONES2, J GOLDING2
(1) Centers for Disease Control and Prevention, Atlanta. (2) University of Bristol, Bristol.
Introduction: In 1999 the U.S. Centers for Disease Control and Prevention and the Avon Longitudinal Study of Parents and Children (ALSPAC) began collaborating to evaluate maturation
indicators within the ALSPAC cohort of British children. Beginning prenatally in 1990, the families in this cohort have contributed questionnaire information at least annually, as well as periodic
physical examination and collection of biologic samples. To date, Tanner stage information has been collected each year from approximately 6000 boys and girls aged eight to 11 years.
Methods: Questionnaire data were analyzed to determine the proportion of girls at each Tanner stage of breast development and pubic hair development, and the proportion of boys at each
Tanner stage of genital development and pubic hair development. All statistical analyses were conducted using SAS 9.1. We also analyzed pooled biologic samples to determine the presence
of environmental chemicals that have been implicated as endocrine disrupters (e.g., DDE, PCBs, PBDEs, bisphenol A, phthalates); both maternal and child samples were analyzed.
Results: At age 11 years, 87% of girls were at least Tanner Stage 2 for breast development; entrance into Tanner Stage 2 defines the beginning of maturation. Two percent of 11 year old girls
reported having reached adult maturity level (Tanner Stage 5) for breast development. Approximately 90% of 11 year old boys were at least Stage 2 of genital development; 1% were at Stage
5. In comparison to reference populations, ALSPAC boys appeared to be manifesting earlier genital development than boys in other European and U.S. reference populations; these differences
were not observed for girls. All pooled biologic samples had detectable levels of the various environmental chemicals or their metabolites.
Discussion and Conclusions: Earlier genital development may be related to exposure to endocrine disrupting chemicals in the environment. The next phase of this research includes
laboratory analysis of banked samples from those children who have now been identified as exhibiting either accelerated or delayed physical development.




                                                                                                  29
         Oral Communication Session

SM3-O-04                    LONGITUDINAL ANALYSIS OF FACTORS PREDICTING ORGANOPHOSPHATE
                            PESTICIDE EXPOSURE TO CHILDREN LIVING IN AN AGRICULTURAL AREA AT
                            AGES 6, 12 AND 24 MONTHS.
A BRADMAN1, M HARNLY2, J SCHWARTZ1, D BARR3, T MCKONE1, B ESKENAZI1
(1) CENTER FOR CHILDREN’S ENVIRONMENTAL HEALTH RESEARCH, UC BERKELEY, BERKELEY, CA. (2) CA DEPT. OF HEALTH SERVICES, DIV. OF ENVIRONMENTAL
AND OCCUPATIONAL DISEASE CONTROL, RICHMOND, CA. (3) CENTERS FOR DISEASE CONTROL AND PREVENTION, DIVISION OF LABORATORY SCIENCES, ATLANTA,
GA.
Introduction: Farmworker children are vulnerable to higher pesticide exposures because they live and play in nearby agricultural fields, are exposed to pesticide residues on their parents’
work clothes and may ingest pesticides in the breast milk of a farmworker mother. We investigated pesticide exposures to children living in an agricultural area with intensive pesticide use.
Methods: We measured six organophosphate (dialkyl phosphate (DAPs)) urinary metabolites in samples from children when they were 6 (n=407), 12 (n=401) and 24 (n=383) months old..
DAPs were summed to obtain a measure of total exposure to approximately 80% of pesticides applied in the study area. Dimethyl (DMP, DMTP, DMDTP) and Diethyl (DEP, DETP, DEDTP)
metabolites were also totaled and evaluated individually. We investigated potential exposure risk factors, including age, sex, season, occupation of parents and household members
(farmworker/not farmworker), wearing of work clothes into the home, distance of home to the fields, home pesticide use and diet.
Results: Levels increased with age, with total median DAP levels of 37, 54 and 76 nmol/L, respectively (p < 0.01). This increase was driven by the total dimethyl metabolites, which rose from
a median of 16 nmol/L to 50 nmol/L between 6 and 24 months. Total diethyl metabolites did not significantly increase over these ages (median = 11 nmol/L to 13 nmol/L, respectively). At 6
months of age, summer season, having farmworker household members, the wearing of agricultural work clothes into the home and consumption of canned fruits or vegetables were
significantly associated with higher levels of total DAP urinary metabolites (p < 0.05). At 12 months, living less than 200 feet from an agricultural field was significantly associated with higher
total DAP levels (p < 0.05). At 24 months, none of the exposure risk factors were statistically associated with total DAP metabolite levels. The findings for total DAP metabolites were driven
by the dimethyl metabolites. For diethyls alone, higher levels were associated with canned fruit or vegetable consumption at 6 months and living less than 200 feet from an agricultural field
at 12 months.
Discussion and Conclusion: Overall, these finding suggest that diet and occupational risk factors are associated with pesticide exposures to very young children. Results for older children
were not consistent, potentially due to higher variability in diets, behavior and locations where children spent time. Findings presented will include the results of multivariate statistical models.




SM3-O-05                    NO INFLUENCE OF PRENATAL EXPOSURE TO PCDD/FS AND DIOXIN LIKE PCBS
                            ON THYROID HORMONES OF MOTHERS AND THEIR NEWBORNS
M WILHELM1, J WITTSIEPE1, F LEMM1, U RANFT2, U KRAEMER2, G EBERWEIN3, M GRESHAKE1, M KRAFT4, K RAUCHFUSS3, G WINNEKE2
(1) Ruhr-University Bochum, Dept. of Hygiene, Social and Environmental Medicine, Bochum. (2) Institut für umweltmedizinische Forschung (IUF) an der Heinrich-
Heine-Universität Düsseldorf, Düsseldorf. (3) Landesumweltamt Nordrhein-Westfalen, Essen. (4) Ministerium für Umwelt und Naturschutz, Landwirtschaft und
Verbraucherschutz des Landes Nordrhein-Westfalen, Düsseldorf.
Introduction: Prenatal exposure to PCB and other persistent organochlorine compounds (POPs) may affect neurobehavioral development of infants and children. This effect may be mediated
through disruption of thyroid hormone homeostasis. However, epidemiological studies reveal no consistent influence of POPs on thyroid status. This birth cohort study evaluates the influence
of the prenatal exposure to PCDD/Fs and dioxin-like PCBs on the neurobehavioral development. Here the influence on the thyroid hormones is reported.
Methods: Blood samples of 182 pregnant women at the age of 19 to 42 years at time of birth living in an industrialized area were taken between 2000 and 2003. All samples were analyzed
for their content of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) as well as dioxin-like PCBs following extraction and sample clean-up by HRGC/HRMS. Thyroid hormones
(TSH, T3, T4, FT3, FT4) were measured in serum samples of the pregnant women and in the cord serum by immunometric assay. Associations between POP exposure and thyroid hormone
status were studied by multiple linear regression analysis. The list of confounders/covariates introduced into the model comprised of alcohol/smoking during pregnancy, duration of pregnancy,
number of pregnancies, nationality, lead and cadmium in blood, selenium in serum of the pregnant women, medication during pregnancy (especially iodide and thyroid hormones) and disease
of the thyroid gland.
Results: Blood levels of WHO-TEQ were in the range of 4.34 to 75.04 pg/g lipid base (median: 25.86). Median values for thyroid hormones in serum of pregnant women were: FT3 4.0 pmol/l, FT4
8.8 ng/l, T3 2.0 µg/l, T4 102.2 µg/l, TSH 1.3 mlU/l and for cord serum: FT3 2.3 pmol/l, FT4 10.9 ng/l, T3 0.7 µg/l, T4 113.7 µg/l, TSH 7.7 mIU/l. Multiple regression analysis showed no influence
of the PCDD/F and PCB exposure on thyroid hormones of mothers and their newborns.
Discussion: This study supports the view, that at the lower dose levels exposure to PCDD/Fs and dioxine-like PCBs does not decrease serum thyroid hormone concentrations. Analysis of
neurological as well as of neuropsychological parameters of this ongoing birth cohort study will show, if at the current exposure levels PCDD/Fs and dioxine-like PCBs affect the neurobehavioral
development of the infants.




SM3-O-06                    LEVELS OF PCB118 ARE ASSOCIATED WITH THYROID HORMONE
                            CONCENTRATIONS IN CHILDREN FROM GENERAL POPULATION
M ALVAREZ-PEDREROL1, N RIBAS-FITÓ1, M TORRENT2, D CARRIZO3, J GRIMALT3, J SUNYER1
(1) Respiratory and Environmental Health Research Unit, Institut Municipal Investigació Mèdica (IMIM), Barcelona. (2) Àrea de Salut de Menorca, IB-SALUT,
Menorca. (3) Department of environmental Chemistry; IIQAB-CSIC, Barcelona.
INTRODUCTION: Thyroid hormones (TH) are essential for normal development of the human fetal brain. Recent reports have suggested that some polychlorinated biphenyls (PCB’s) or other
persistent organohalogen compounds (OC’s) may alter TH levels in both animals and humans. The mechanisms involved in the alteration of TH homeostasis have been investigated in several
studies but are still not fully understood. Because of the structure similarity of some OC’s and TH, they are suspected either to decrease or mimic their biological action. The possible mechanisms
include (1) direct interaction with the thyroid gland; (2) increased biliary clearance of thyroxine (T4) through the induction of thyroid metabolizing enzymes; and (3) competitive binding to
thyroid hormone transport proteins like transthyretin (TTR) resulting in a decreased plasma T4 level. This study evaluates the effects of background exposure to some OC’s on the levels of free
thyroxine (FT4), thyroid stimulating hormone (TSH) and triiodothyronine (T3) in 4-year children from general population.
METHODS: Children (n=255) from a Spanish general population birth cohort in Menorca Island (a popular tourist destination with no industry) were assessed at four years of age. Thyroid
hormones (TSH, T3, FT4) and OC’s (PCB’s 28, 52, 101, 118, 153, 138, 180, DDE, DDT, HCB, HCH) were analyzed. OC’s were categorized in quartiles.
RESULTS: PCB118 (range: 0.018-0.42ng/ml) showed a linear dose-response relationship with levels of T3 (range: 95-204ng/dl) and FT4 (range: 0.76-1,38ng/dl) observing a decrease of 3,2ng/dl
(p=0.402), 4.8ng/dl (p=0.209) and 11,5ng/dl (p=0.003) of T3 concentrations in children from the second, third and fourth quartile of PCB118, respectively, and a decrease of 0.03ng/dl
(p=0.266), 0.06ng/dl (p=0.019) and 0.07ng/dl (p=0.003) of FT4 concentrations in children from the second, third and fourth quartile of PCB118, respectively, which was not confounded for
body-mass index, gestational age, breastfeeding, mother smoking, sex, location, and fish consumption. No significant relations were observed between thyroid hormones and other OC’s.
DISCUSSION AND CONCLUSIONS: This study shows an association between levels of PCB118 and thyroid hormones (T3 and FT4) in children from general population (children not
particularly exposed to OC’s). Further research is needed to study if this exposition to background levels can affect neurodevelopment in children, as a consequence of a decrease of thyroid
hormones concentrations in brain.




                                                                                                 30
                                                                                                                                                                                                 SUNDAY SEPTEMBER 3
        Oral Communication Session

SM3-O-07                   IMPACT OF PRENATAL CHLORPYRIFOS EXPOSURE ON NEURODEVELOPMENT IN
                           THE FIRST THREE YEARS OF LIFE AMONG INNER CITY CHILDREN
VA RAUH1, R GARFINKEL1, R PERERA1, H ANDREWS1, D BARR2, D WHITEHEAD2, D TANG1, RM WHYATT1
(1) Columbia Center for Children’s Environmental Health, Mailman School of Public Health, Columbia University, New York. (2) The Centers for Disease Control and
Promotion, New York.
Introduction:
Chlorpyrifos (CPF) is a known neurodevelopmental toxicant in animals, but effects on human development may be undetected because of the absence of overt signs of toxicity. Despite a
recent ban on residential use, agricultural applications continue in the U.S. and abroad. As part of a prospective cohort study in an urban minority population, neurodevelopmental effects of
prenatal exposure to CPF were evaluated in 254 children through 36 months. Previous reports from this study showed widespread prenatal CPF exposure and significant adverse impacts on
birth weight and birth length.
Method:
Eligible women were non-smoking, aged 18-35, registered for prenatal care by 20 weeks, free of diabetes, hypertension, HIV, and drug abuse, and resident in the area for at least one year.
All children had reached 36 months with complete maternal interview data, biomarkers of CPF exposure (from umbilical cord blood), postnatal assessment of the caretaking environment; and
neurobehavioral outcomes. Outcomes included cognitive and motor development at 12, 24, and 36 months on the Bayley-II Scales of Infant Development and behavior problems on the Child
Behavior Checklist (CBCL). Age-specific regression models were adjusted for race/ethnicity, sex, gestational length, maternal education, maternal IQ, prenatal ETS, and quality of the home
environment. To estimate the CPF effect on neurodevelopment over time, general linear modeling for repeated measures was used.
Results:
Children prenatally exposed to high CPF levels, greater than 6.17 picograms/gram (pg/g) plasma, were significantly more likely than low-exposed children to experience delay in psycho-motor
(p=0.002) and cognitive (p=0.024) development at 36 months, and this effect was not mediated by birthweight or birth length. Highly exposed children were significantly more likely than
less exposed children to manifest symptoms of attention disorder (p=0.009), attention deficit hyperactivity disorder (ADHD) (p=0.046), and pervasive developmental disorder (PDD) (p=0.025),
based on the CBCL. Children exposed to high levels of CPF were 11, 6 and 5.6 times as likely to score in the clinical range on attention problems, ADHD, and PDD, respectively.
Discussion and Conclusions:
This report is one of the first to show that prenatal organophosphate exposure generates a pattern of neurodevelopmental delay similar to the effects of other known neurotoxicants (lead,
tobacco, cocaine), marked by a decline in the preschool period and lower scores for exposed children. The association with behavior problems, known to be co-morbid with subsequent learning
disabilities and academic problems, is a new finding and deserves further study.




SM3-O-08                   ORGANOPHOSPHATE PESTICIDES AND NEURODEVELOPMENT IN YOUNG
                           MEXICAN AMERICAN CHILDREN
B ESKENAZI1, AR MARKS1, K HARLEY1, A BRADMAN1, C JOHNSON1, DB BARR2, N MORGA3, NP JEWELL1
(1) Center for Children’s Environmental Health Research, School of Public Health, University of California, Berkeley, Berkeley, CA. (2) National Center for
Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA. (3) Center for the Health Assessment of Mothers and Children of Salinas, Clinica de
Salud del Valle de Salinas, Salinas, CA.
Introduction: Despite animal evidence suggesting neurotoxic effects of organophosphate (OP) pesticides, few published studies investigate the neurodevelopmental toxicity of OP pesticides
in humans. This study investigates exposure to OP pesticides and neurodevelopment of infants from low-income Mexican farmworker families in California.
Methods: Participants from this birth cohort study include 445 singletons with psychomotor (PDI) and mental (MDI) development assessed using the Bayley Scales of Infant Development at
6 (n=405), 12 (n=405), and/or 24 (n=379) months of age. Six non-specific dialkylphosphate (DAP) metabolites of OP pesticides were measured in urine collected from mothers twice during
pregnancy (mean=13 and 26 weeks gestation) and from children at the time of the neurodevelopmental assessments. The six DAPs were summed to obtain total DAPs at each time point,
representing exposure to approximately 80% of OPs used in the Salinas Valley {California EPA, 2002 #27}. Total DAPs were transformed to the log10 scale. Pregnancy DAPs were averaged
before log10 transformation.
Results: Median total DAPs levels were 103.2 and 111.8 nmol/gram in pregnant mothers and 37.0, 54.4, and 75.6 nmol/gram in children at ages 6, 12, and 24 months, respectively. Bayley
MDI scores (mean ± standard deviation) were 95.6 ± 7.0, 100.8 ± 8.9, 85.8 ± 11.7 at 6, 12, and 24 months, respectively; corresponding PDI scores were 96.4 ± 10.6, 106.0 ± 12.7, and 97.6
± 10.6, respectively.
• We found a 3.51-point decrease in 24-month MDI (p=0.02) with each 10-fold increase in average pregnancy DAPs, but no association at 6 and 12 months. We found a 1.49-point increase
  (p=0.04) in 12-month MDI and a 1.97-point increase (p=0.04) in 24-month MDI associated with each 10-fold increase in children’s DAPs concurrent to the assessments. We observed no
  interaction between prenatal and concurrent levels, and no association between DAPs metabolites and PDI.
• Analyses using diethyl and dimethyl phosphate metabolites, which sum to make total DAPs, will also be presented. Analyses of DAPs and maternal reports of behavior from the Child Behavior
  Checklist results are ongoing and will also be discussed.
Discussion and Conclusions: These results suggest that prenatal exposure, but not postnatal exposure, to OPs may be negatively associated with mental neurodevelopment during early
childhood. We have no ready explanation for the positive associations we observe between children’s DAPs metabolites and MDI, although we note that the outcomes may precede these
exposures.




SM3-O-09                   IN UTERO PESTICIDE EXPOSURE AND NEURODEVELOPMENT IN THREE
                           NIEHS/EPA CHILDREN’S CENTER BIRTH COHORTS
B ESKENAZI1, K HARLEY1, A BRADMAN1, L FENSTER1, M WOLFF2, S ENGEL2, V RAUH3, R WYATT3, F PERERA3
(1) Center for Children’s Environmental Health Research, UC Berkeley, Berkeley. (2) Center for Children’s Environmental Health and Disease Prevention, Mt. Sinai
Medical Center, New York. (3) Center for Children’s Environmental Health, Columbia University, New York.
Introduction: Despite evidence in animal studies and known human exposure, the effect of in utero exposure to organophosphate (OP) and organochlorine (OC) pesticides on human
neurodevelopment has rarely been studied in humans. In 1999-2000, three of the NIEHS/EPA Children’s Centers established birth cohorts, two in New York City and one in rural California,
to examine the association of OPs and OCs with neurodevelopment.
Methods: All three Centers collected maternal urine and blood samples, two administered the Brazelton Neonatal Behavioral Assessment Scale (BNBAS) shortly after delivery (Ns = 381 and
311), and all administered the Bayley Scales of Infant Development when the children were 1 (Ns = 416, 200, and 523) and 2 years of age (Ns = 389, 276, and 419). One center also
administered the Bayley at 6 months and another at 3 years. Metabolites of OP pesticides, including three diethylphosphates (DEP), three dimethylphosphates (DMP), and malathion
dicarboxylic acid (MDA), were measured in maternal urine in two centers and OP parent compounds were measured in maternal and cord blood in the other center. DDT/DDE and other OC
compounds were measured in maternal and/or cord blood at all three centers. Multivariate analyses of the association of OPs and OCs with BNBAS and/or Bayley performance are being
conducted separately in each population and in pooled analyses.
Results: Median levels of DEP were comparable in the New York and California populations, but median levels of DMP and DDE were considerably higher in California mothers. Higher levels
of total OP metabolites were associated with an increase in the number of abnormal reflexes on the BNBAS in two different cohorts, although results differed for subgroups of metabolites.
Higher chlorpyrifos levels in cord blood were associated with significantly poorer performance on the Bayley MDI and PDI at age 3 years, as well as with ADHD and attentional problems, at
one Center. Analyses of OP metabolites and Bayley score are currently underway at two Centers and will be presented. DDT/DDE was not associated with any of the BNBAS domains, but
higher levels of DDT/DEE were associated with poorer performance on the Bayley MDI at 1 and 2 years of age.
Discussion: The racial/ethnic make-up of the populations and the magnitude and types of exposures differed across Centers; yet there is some consistency in results. Data will be presented
from the three Centers, results will be compared and contrasted, and findings from pooled analyses will be presented.




                                                                                             31
        Oral Communication Session

SM3-O-10                    IN UTERO EXPOSURE TO BACKGROUND CONCENTRATIONS OF DDT AND
                            COGNITIVE FUNCTIONING AMONG PRESCHOOLERS
N RIBAS-FITÓ1, M TORRENT2, D CARRIZO3, L MUÑOZ-ORTIZ1, J JÚLVEZ1, JO GRIMALT3, J SUNYER1
(1) Respiratory and Environmental Health Research Unit, Institut Municipal Investigació Mèdica (IMIM), Barcelona. (2) Àrea de Salut de Menorca, IB-SALUT,
Menorca. (3) Department of Environmental Chemistry, IIQAB-CSIC, Barcelona.
Background: p,p’-DDT (bis[p-chlorophenyl]-1,1,1-trichloroethane) is a persistent organochlorine compound that was used worldwide as an insecticide. Exposure to DDT and DDT metabolites
(such as p,p’-DDE) occurs both in utero and through breastfeeding. Given that brain growth continues after birth, both pre- and postnatal periods are critical windows of vulnerability for the
nervous system. DDT has been reported to be neurotoxic in animals, affecting the sodium channels in nerve cell membranes, but little is known about its effects in children. The aim of the
study is to evaluate the association of cord serum levels of DDE and DDT with neurodevelopment at age 4.
Methods: Two birth cohorts in Ribera d’Ebre and Menorca (Spain) were recruited between 1997-1999 (n=475). Infants were assessed at age 4 using the McCarthy Scales for Infant
Development. Organochlorine compounds were measured in cord serum. Children’s diet and paternal sociodemographic information was obtained through questionnaire.
Results: DDT cord serum concentration at birth was inversely associated with the Verbal, Memory, Quantitative and Perceptual-Performance skills at age 4. Children with DDT concentrations
in cord serum above 0.20 ng/ml had a mean decrease of 7.86 (SE=3.21) points in the verbal scale and a mean decrease of 10.86 (SE=4.33) points in the memory scale when compared to
those with concentrations below 0.05 ng/ml. These associations were stronger among girls.
Conclusions: Prenatal exposure to background low-level concentrations of DDT was associated with a decrease in the cognitive skills among preschoolers. These results should be considered
when evaluating the risk and benefits of spraying DDT in antimalarial and other disease-vector campaigns.




SAA1-O                      CANCER AND THE ENVIRONMENT
SAA1-O-01                   POPULATION MIXING AT THE PLACE OF RESIDENCE AT THE TIME OF BIRTH
                            AND INCIDENCE OF CHILDHOOD LEUKAEMIA IN FRANCE
J RUDANT1, B BACCAINI2, M RIPERT1, A GOUBIN1, S BELLEC1, D HEMON1, J CLAVEL1
(1) INSERM, U754, IFR69, Université Paris-Sud XI, Villejuif. (2) INSEE, Provence-Alpes-Côte d’Azur, Marseille.
Introduction
Childhood leukaemia is suggested to be a rare response to a common viral infection, which could develop under particular conditions of population mixing in isolated areas. The aim of the
present study was to investigate the association between the risk of childhood leukaemia before age 7 years and population mixing at the place of residence at the time of birth in France.
Methods
We considered retrospectively all French births between the 1st January 1990 and the 31st December 1998. All leukaemia cases less than 7 years old that occurred during that period were
identified from the French National Registry of Childhood Leukaemia and Lymphomas (NRCL) and their places of residence at birth were sought. Isolated status, population density and
demographic data on the Commune (smallest French administrative unit) were obtained from the 1990 and 1999 population censuses. Population mixing was estimated by the number of
incomers from another commune, another département or another région.
Results
An increased risk of acute lymphoblastic leukaemia was found with higher levels of migration for children residing at birth in isolated communes with a population density ≥ 50 people per
km2, strongest association being observed for migrations defined at the higher levels of aggregation (région: standardized incidence rates ratio (SIRR) = 2.6, 95%CI: 1.5-4.5; département:
SIRR = 2.4, 95%CI: 1.1-4.7; commune: SIRR = 2.0, 95%CI: 0.8-4.6). No association was observed with lower population densities. For children residing in non-isolated communes at birth, the
results were similar but less marked. The risk tended to increase only for population densities ≥ 5000 people per km2 (région: SIRR = 1.6, 95%CI: 1.0-2.5; département: SIRR = 1.8, 95%CI:
1.0-3.2; commune: SIRR = 1.2, 95%CI: 0.8-1.8).
Discussion and conclusions
The findings were then consistent with epidemic models and thus support the hypothesis of an infectious aetiology related to population mixing. Population density may be seen as an indicator
of the opportunity of contacts between inhabitants and should therefore be taken into account when investigating an infectious hypothesis. This is the first systematic study of population
mixing at the place of residence at the time of birth to be conducted on a national scale.




SAA1-O-02                   PCBS AND NON HODGKIN’S LYMPHOMA: RESULTS FROM THREE COHORTS
L ENGEL1, F LADEN2, A ANDERSEN3, P STRICKLAND4, A BLAIR5, L NEEDHAM6, D BARR6, M WOLFF7, K HELZLSOUER8, D HUNTER2, Q LAN5, K
CANTOR5, G COMSTOCK8, J BROCK6, D BUSH8, N ROTHMAN5
(1) DEPT. OF EPIDEMIOLOGY AND BIOSTATISTICS, MEMORIAL SLOAN-KETTERING CANCER CENTER, NEW YORK, NY. (2) CHANNING LABORATORY, DEPT. OF
MEDICINE, BRIGHAM AND WOMEN’S HOSPITAL AND HARVARD MEDICAL SCHOOL, BOSTON, MA. (3) THE NORWEGIAN CANCER REGISTRY, OSLO. (4) DEPT. OF
ENVIRONMENTAL HEALTH SCIENCES, JOHNS HOPKINS BLOOMBERG SCHOOL OF PUBLIC HEALTH, BALTIMORE, MD. (5) DIV. OF CANCER EPIDEMIOLOGY AND
GENETICS, NATIONAL CANCER INSTITUTE, NATIONAL INSTITUTES OF HEALTH, DEPT. OF HEALTH AND HUMAN SERVICES, BETHESDA, MD. (6) NATIONAL CENTER
FOR ENVIRONMENTAL HEALTH, CENTERS FOR DISEASE CONTROL AND PREVENTION, ATLANTA, GA. (7) DEPT. OF COMMUNITY AND PREVENTIVE MEDICINE,
MOUNT SINAI SCHOOL OF MEDICINE, NEW YORK, NY. (8) DEPT. OF EPIDEMIOLOGY, JOHNS HOPKINS BLOOMBERG SCHOOL OF PUBLIC HEALTH, BALTIMORE, MD.
Introduction: The incidence of non-Hodgkin’s lymphoma (NHL) unrelated to HIV infection has steadily risen over the past several decades and remains substantially unexplained. There is
some evidence that polychlorinated biphenyls (PCB) measured in blood or fat tissue may be associated with increased risk of NHL. Although PCB congeners vary in their biological activity, the
relation between individual congeners and NHL risk has not been previously examined using prospectively collected biospecimens. We examined congener-specific associations in three
prospective cohorts.
Methods: Prediagnostic serum or plasma concentrations of selected PCB congeners were measured among NHL cases and controls from 3 prospective cohorts: Janus (190 cases,190 controls)
in Norway, and Clue I (74 cases, 147 controls) and the Nurses’ Health Study [NHS] (30 cases, 78 controls) in the U.S. All blood samples were collected in the 1970s or 1980s. We used logistic
regression to calculate odds ratios (OR) and 95% confidence intervals (CI) for the relations between risk of NHL and lipid-corrected plasma or serum concentrations.
Results: Several congeners (i.e., 118, 138, and 153) that were present at higher levels and were moderately to highly correlated with each other showed exposure-response trends with risk
of NHL in all three cohorts. The strongest trends were seen for congener #118, with ORs and 95% CIs for increasing thirds/fourths of concentration relative to the lowest third/fourth of: 1.0
(0.5-2.0), 1.2 (0.6-2.3), and 1.7 (0.9-3.5) in Janus; 4.9 (1.6-15.3), 3.5 (1.0-11.8), and 5.4 (1.7-17.1) in Clue I; and 2.5 (0.8-8.5) and 2.5 (0.7-9.0) in NHS (p for trend < 0.05 in all 3 cohorts).
These associations were observed primarily among subjects diagnosed closer to the date of blood collection in the two cohorts with sufficient cases to permit stratification by time. Among
cases diagnosed within the median years of follow-up (16 years in Janus, 12 years in Clue I), ORs and 95% CIs for increasing fourths of concentration of congener #118 were: 2.4 (0.9-6.5),
4.9 (1.6-15.3), and 5.3 (1.5-18.8) (p trend < 0.005) in Janus and 8.1 (1.0-68.9), 6.6 (0.7-59.0), and 13.0 (1.6-106.8) (p trend < 0.05) in Clue I. Similar patterns were seen for congeners #138
and #153 and for total PCBs. Limited evidence of exposure-response trends was also observed for several other congeners. The primary DDT metabolite, p,p’-DDE, was not significantly
associated with NHL in most analyses, but slightly to moderately confounded the PCB associations.
Conclusions: The results from these three cohorts suggest that PCB levels in blood are associated with increased risk of NHL.



                                                                                                32
                                                                                                                                                                                                   SUNDAY SEPTEMBER 3
        Oral Communication Session

SAA1-O-03                  PESTICIDE USE AND PROSTATE CANCER INCIDENCE IN A PROSPECTIVE
                           COHORT STUDY
MCR ALAVANJA1, R MAHAJAN1, LE BEANE-FREEMAN1, JH LUBIN2, CJ HINES3, K THOMAS4, J COBLE1, DP SANDLER5, JA HOPPIN5, A BLAIR1
(1) Occupational and Environmental Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health,
Department of Health and Human Services, Rockville, MD. (2) Biostatistics Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, National
Institutes of Health, Department of Health and Human Services, Rockville, MD. (3) National Institute for Occupational Health and Safety, Centers for Disease Control,
Department of Health and Human Services, Cincinnati, OH. (4) U.S. Environmental Protection Agency, Research Triangle Park, NC. (5) Epidemiology Branch, National
Institute of Environmental Health Sciences, National Institutes of Health, Department of Health and Human Services, Research Triangle Park, NC.
Introduction: Prostate cancer has been observed to occur in excess among farmers and other agricultural workers in a number of countries around the world.
Methods: A prospective cohort of 89,658 pesticide applicators and their spouses were followed for cancer incidence for an average of 8.5 years. Members of the cohort completed
questionnaires about pesticide use, occupations, lifestyle and medical history during enrollment (1993-1997) and at two subsequent periods of time. A total of 1289 incident prostate cancers
cases were identified among the 57,311 male pesticide applicators through Dec 2003 through linkage with population-based cancer registries. Two exposure metrics were used to estimate
exposure. One was an estimate of total days of a specific pesticide use in a life-time; the second, total days of exposure weighted by a published estimate of exposure intensity.
Results: Male cohort members experienced a small, but significant, elevated risk of prostate cancer compared to the general populations of the study states (SIR = 1.24, 95% CI 1.18-1.33).
A family history of prostate cancer among first-degree relatives of male cohort members conferred a two-fold excess risk of prostate cancer. Significant linear exposure-response trends
(p < 0.05) were observed for six specific pesticides (fonofos, phorate, coumaphos and chlorpyrifos which are organothiophosphates; butylate, an herbicide; and permethrin, an insecticide), but
only among those with a family history of prostate cancer, confirming a preliminary published report from this study.
Discusion/Conclusions: The consistency of our findings with two different exposure metrics, in both study sites and with preliminary published results suggests that genetic susceptibility
among pesticide applicators may play an important role in prostate cancer etiology. Shared environmental risk factors for prostate cancer among family members, however, may also explain
these results. Planned molecular studies nested within the same cohort to assess genetic and epigenetic modes of action might clarify and explain these associations.




SAA1-O-04                  BLADDER CANCER AND OCCUPATIONAL EXPOSURES: RESULTS OF A
                           MULTICENTRIC CASE CONTROL STUDY
P ANDUJAR1, Y IWATSUBO1, D VORDOS2, F CONSO3, B GATTEGNO4, V RAVERY5, E FONTAINE6, T BILLEBAUD7, Y ALLORY1, M SIBONY1, C BERTHAUT3,
S CHAMMING’S3, D CHOPIN1, C ABBOU2, JC PAIRON1
(1) INSERM E03.37 Université Paris XII, Créteil. (2) Service d’Urologie Hôpital Henri Mondor, Créteil. (3) Institut Interuniversitaire de Médecine du Travail de Paris-Ile
de France, Paris. (4) Service d’Urologie Hôpital Tenon, Paris. (5) Service d’Urologie Hôpital Bichat, Paris. (6) Service d’Urologie Hôpital Ambroise Paré, Boulogne. (7)
Service d’Urologie Centre Hospitalier Intercommunal de Créteil, Créteil.
The aim of this multicentric case-control study was to evaluate the frequency of occupational causes of bladder cancer and the distribution of exposure situations. Between 1998 and 2004,
316 male incident cases of bladder cancer and 316 male controls matched on age, were recruited in five hospitals. Each subject had a face to face interview to collect information on tobacco
consumption and on the complete job history occupational exposure to polycyclic aromatic hydrocarbon (PAH), aromatic amines (AA) and nitrosamines (NA) evaluated by two occupational
experts, blindly according to the case/control status. For each subject, a cumulative exposure index (CEI) to PAH, AA and NA was calculated, taking into account probability, frequency and
exposure level for all job periods during the working life. Logistic regression analyses were performed to calculate odds ratios (OR) with corresponding 95% confidence intervals (95% CI) for
bladder cancer occurrence and PAH, AA and NA exposures, adjusted for age and tobacco smoking.
As expected, this study shows a dose-effect relationship between tobacco consumption levels and risk of bladder cancer (OR = 1.97 (95% CI: 1.14-3.41) for low tobacco consumption level
(1-20 pack-years (PY), 5.36 (95% CI: 3.14-9.16) for high tobacco consumption level (21-40 PY) and 11.66 (95% CI: 6.63-20.48) for very high tobacco consumption level (> 40 PY). A dose-
response relationship was observed between probability of exposure to PAH or cumulative exposure index to PAH and risk of bladder cancer after adjustment on age and tobacco consumption
(OR = 2.31 (95% CI: 1.39-3.85) for definite exposure and OR = 1.93 (95% CI: 1.08-3.43) for higher CEI). By contrast no relationship was identified between AA or NA and bladder cancer in
this series. Excess bladder cancer risks were observed in various occupations, such as motor vehicle drivers (p<0,0002), earthmoving plant or material handling drivers (p<0,002) and sheet-
metal workers (p<0,0006) and in various industries, such as wood industry (p<0,0002), ship building and repairing and manufacture of railroad equipment, motor vehicles and aircraft
(p<0,0005), freight transport by road (p<0,006) and retail trade (p<0,004).
The lack of dose-response relationship between bladder cancer AA or NA exposures and bladder cancer could reflect a decrease of exposures to AA and NA since a few decades. By contrast,
the relationship between PAH exposure and bladder cancer justifies to go deeply into exposure situations identified of this study in order to propose specific prevention programs.




SAA1-O-05                  ASSOCIATION OF NITRATES, TRIHALOMETHANES, AND ARSENIC IN DRINKING
                           WATER AND TESTICULAR CANCER RISK
C WOLFF1, P ENGLISH2
(1) Impact Assessment, Inc., La Jolla. (2) California Department of Health Services, Richmond.
Introduction
Testicular cancer is the most common cancer among young men in the U.S. The disease has few known risk factors. Chemical contaminants in drinking water are suspected disrupters of the
endocrine system, and may affect estrogen levels. In this study, we used a large number of cases from the California Cancer Registry to determine if an association exists between chemical
contaminants in drinking water and risk of testicular cancer.
Methods
We used the California Water Quality Monitoring Database to obtain levels of sampled water contaminants. We frequency-matched controls to cases based on the SES characteristics of the
cases’ 1990 or 2000 U.S. Census block groups. Of 7,119 geocoded cases, we frequency-matched 7,095 controls. We defined the exposure period as two years before the diagnosis date of the
cases. We developed a probability weighting methodology that took into account several data sources to link a subject to a water system. Water systems were surveyed to characterize relative
concentration contributions of upstream sampling stations for each study subject during the corresponding exposure period. Logistic regression was used for multivariate analysis.
Results
A strong dose-response pattern was found for nitrates in surface water only, with significant odds ratios in every category. The AOR for nitrates in surface water above the 95th percentile
(compared to the first quartile) was 1.99 (95% C.I.=1.28, 3.10). For arsenic, significant adjusted odds ratios emerged in surface water above the 3rd quartile and above the 95th percentile,
compared to the first quartile (AOR=1.39 [95% C.I.=1.03, 1.88] and AOR=2.52 [95% C.I.=1.54, 4.10], respectively). In surface water, we found an apparent interaction between nitrates and
trihalomethanes (THMs), with a strong dose response when THMs were present. For surface water sources only with THMs present, AOR’s for nitrate ranged from 1.76 for the second quintile
(95% C.I.=1.25, 2.48) to 2.91 for the fifth quintile (95% C.I.=1.69, 5.01), compared to the first quartile.
Discussion
The risk of testicular cancer associated with nitrate levels increased in a dose-response fashion in households receiving surface sources of water only. We found that in the presence of THMs
in surface water, the dose-response pattern became even steeper, with the risk associated with the highest level of nitrate concentrations increasing to threefold. These data are suggestive of
an effect of nitrates and THMs on testicular cancer risk, but the lack of an appropriate lag time, and more exact consumption data limits the interpretability of the findings.




                                                                                              33
        Oral Communication Session

SAA1-O-06                  URINARY CADMIUM LEVELS, BREAST CANCER RISK, AND ENVIRONMENTAL
                           TOBACCO SMOKE EXPOSURE
J MCELROY1, M SHAFER1, A TRENTHAM-DIETZ1, J HAMPTON1, P NEWCOMB2
(1) University of Wisconsin, Comprehensive Cancer Center, Madison. (2) Fred Hutchinson Cancer Research Center, Seattle.
Cadmium, a highly persistent heavy metal, is categorized as a probable human carcinogen by the U.S. Environmental Protection Agency. In our population-based case-control study of Wisconsin
women (USA), we interviewed by telephone 246 case women aged 20-69 years and 254 randomly selected similarly aged controls from September 2004 to February 2005. The urine collection
kits were carefully designed to minimize trace element contamination during specimen collection and handling in each participant’s home. After returning the samples by postal mail, cadmium
was quantified using inductively-coupled plasma mass spectrometry. The participation proportion for those completing the interview was 75% for eligible cases and 71% for controls in the
parent study. We performed two cadmium exposure analyses—to evaluate breast cancer risk in the case-control population and to evaluate the possible association of environmental tobacco
smoke exposure (ETS) in the control population. Creatinine-adjusted cadmium levels ranged from 0.02 µg/g to 4.55 µg/g for cases (excluding one observation at 30.95 µg/g) and for controls
at 0.08 µg/g to 2.64 µg/g. Approximately 60% of the controls reported exposure to parental smoking in the home during childhood. For smoking status 12% of the control population were
current smokers and 26% were former smokers. Recent exposure (approximately 2 years prior to the interview) to smoke at work or in the home was rare (10% exposed at work and 6%
exposed at home). In contrast, participants were more likely to be exposed to ETS in social settings with only 16% reporting no recent ETS exposure in social settings. For our analysis of breast
cancer risk and urinary cadmium exposure, we observed a two-fold increased breast cancer risk (odds ratio 2.29; 95% confidence interval 1.25-4.20) for the highest creatinine-adjusted
cadmium category (•0.53 µg/g) compared to the lowest creatinine-adjusted cadmium category (<0.26 µg/g) after adjustment for established breast cancer risk factors (p- trend = 0.01). In
the control population, comparison of least squares means creatinine-adjusted cadmium levels for smoking status of never, former, or current was significantly different between each group
(p-values < 0.003) with current smokers having the highest cadmium level and never smokers having the lowest level. In comparison to control participants who reported no ETS exposures,
the least squares means creatinine-adjusted cadmium levels in women exposed as a child to ETS in the home or as an adult in recent home, work or social settings were not different.




SAA1-O-07                  SIMIAN VIRUS 40 AND MESOTHELIOMA: AN EPIDEMIOLOGICAL STUDY IN
                           GREAT BRITAIN
M PRICE1, A DARNTON2, D MCELVENNY2, J HODGSON2
(1) UNIVERSITY OF BRISTOL, BRISTOL. (2) HEALTH AND SAFETY EXECUTIVE, BOOTLE.
Introduction: Simian Virus 40 (SV40) is a DNA virus that has been shown capable of infecting and transforming cells in various species. Laboratory studies have suggested that inoculation
with SV40 is associated with various types of cancer – including mesothelioma. However, a recent study concluded that SV40 is at most only rarely present in human mesotheliomas, and that
its positive detection in many Polymerase Chain Reaction (PCR) based studies may result from the wide use of assay designs susceptible to false-positive results. The aim of this study was to
test the hypothesis, via an ecological analysis, that exposure to SV40 via contaminated polio vaccines is a risk factor for mesothelioma in humans in Great Britain.
Methods: Information from Ministry of Health reports about what proportion of children were adminstered the potentially contaminated polio vaccines, when and at what ages was used to
construct two birth cohorts most likely to be exposed to SV40 (those born in 1951-55 and 1956-1960) and a further cohort likely to be unexposed (those born in 1962-66). Two SV40-exposed
cohorts were constructed as a means of controlling indirectly for confounding by asbestos exposure. Age adjusted mesothelioma mortality rates for each exposed cohort were then compared
separately for males and females with those for the unexposed cohort.
Results: In males, neither comparison showed a significant difference in mesothelioma mortality between the exposed and unexposed cohorts. However, in both comparisons the
mesothelioma mortality rate for females was statistically significantly higher in the
exposed cohort (Table):
Conclusion: Though based on small numbers, these results are somewhat intriguing. The
reduction in female mesothelioma mortality in the unexposed cohort could be consistent
with an SV40 effect. However, the different results for males and the lack of any a priori
evidence of a sex-specific SV40 effect makes this and other interpretations problematic.
Other explanations include SV40 being a risk factor for mesothelioma in both sexes with
the effect in males being masked by an increase in asbestos exposure at younger ages,
or simply that over time there was a decrease in the levels of asbestos exposure among
young women but not among young men. The fact that the mesothelioma rate in males
does not reduce between 1956-60 and 1962-66 is somewhat surprising in itself, given
that in general asbestos exposures of men born in this period are likely to have fallen.




SAA2-O                     ETHICAL ISSUES IN ENVIRONEMENTAL HEALTH: VIOLENCE AND PRECAUTION
SAA2-O-01                  EXPOSURE TO VIOLENCE – HEALTH IMPACT ON REFUGEES
J LINDERT1, H TRAUE2, F LAMOTT3
(1) Queen Mary, University of London, London. (2) University of Ulm, Ulm. (3) University of Ulm, Ulm.
Introduction: Worldwide there are up to 30 millions of refugees. The impact of violence on mental health of refugees is an expanding area of research. There is ample evidence that the
prevalence rates for mental disorders among refugees can be very high, but there are little data refugees from Kosovo as regards prevalence of post-traumatic stress disorder, anxiety and
depression in relation to quantity and quality of traumatic events. We aimed to assess the prevalence of post-traumatic stress disorder, anxiety and depression of these refugees and to analyse
symptom levels and potential dose-response relationships.
Methods: Cross-sectional cluster sample survey was conducted among 178 refugees. Exposure to traumatic events was measured using the “Traumatic experiences checklist” of the
“Harvard–Trauma-Questionnaire” (HTQ). Exposure assessment was completed by qualitative interviews. Main outcome measures were posttraumatic stress disorder (PTSD) symptoms,
depression and anxiety. All refugees were interviewed face-to-face using the “Hopkins-Symptom-Checklist 25” (HSCL-25). Main biographical data was collected with a standardised
demographic measure.
Results All refugees were exposed to high levels of violence during war (almost 90%), most commonly reported events were forced expulsion and deprivation of water, food and shelter. The
rates of PTSD (44%), depression (42%) and anxiety (40%) were extremely high. Differences in symptom levels between age groups with a peak for the age group between 40-60 were found.
Conclusion: High levels of psychological problems are likely to occur in populations affected by mass violence. Further longitudinal research is urgently needed to acquire more knowledge
of the impact of man-made disaster on mental health.




                                                                                               34
                                                                                                                                                                                                   SUNDAY SEPTEMBER 3
        Oral Communication Session

SAA2-O-02                  LONG TERM HEALTH IMPACT OF GENOCIDE AND ORGANISED VIOLENCE
J LINDERT1, S PRIEBE1
(1) Queen Mary, University of London, London.
Background: 250 million have died from genocide in the 20th century; millions more are at risk for the effects of exposure to mass violence and genocide. Data and knowledge about
longterm mental health impact of exposure to mass atrocities remains scattered. We aimed to review available knowledge on mental health effects of exposure to genocide and mass and to
compare the results to knowledge about mental health impact of environmental disasters.
Methods: We systematically reviewed all published papers located by keywords *genocide, *disaster, *mass violence, *transgenerational effects in Pubmed, Science Citation Index and
Psychinfo published 1945-2005. Additionally articles were hand searched in English and German journals. We developed quality criteria i.e. use of standardized measures, number of people
assessed, training of the interviewers. Case reports and studies with less than 100 people were excluded from the review, remaining 167 studies to be analyzed.
Results: The systematic review of studies on the long-term health impact of genocide suggests that exposure to organized violence has an impact on mental health far in excess of other
health emergencies. Results varied and level of mental disorders like depression, anxiety and posttraumatic stress disorder varied in relation to the decade when the study had been carried
out (2%-88%). The earlier the study had been carried out less impact has been found. A relationship between languages in which the study had been conducted was found, with studies
published in English language showing higher impact on mental health than studies conducted in German speaking countries. Some studies in German speaking countries showed no effect.
The younger studies assessing mental health impact with standardised measures showed higher levels of morbidity (up to 80%) in the affected populations. Weaknesses of studies are that
mainly cross-sectional studies have been carried out. Studies on Mental health impact of environmental disasters showed less mental health impact.
Conclusions: The studies have limitations, but still suggest, that man made disasters have long lasting mental health effects, which are more frequent and more sustained than those from
environmental disasters.




SAA2-O-03                  MALTHUSIAN PRESSURES: DO THEY EXPLAIN GENOCIDE?
E RICHTER1, R BLUM1, T BERMAN1, G STANTON2
(1) Hebrew University-Hadassah, Jerusalem. (2) University of Mary Washington, Fredericksburg.
Introduction and Purposes: To prevent genocide, we have carried out preliminary explorations of the reciprocal relationships between Malthusian pressures, resulting from high population-
low resource ratios and risks for genocide. Our purpose is to determine whether (1) Malthusian pressures and zero-sum rivalries for power over depleted, diminishing or contaminated resources
increase risks for genocide; (2) genocide increases risks for environmental degradation, including wanton environmental abuse— and vice versa; (3) the explanatory power of so-called
“upstream” environmental indicators of carrying capacity depletion is sufficient to account for genocide without attention to the known political determinants of genocide; (4) unsustainable
population/resource ratios increase risks for violence; (5) intervention to prevent or stop genocide can be sustainable, without attention to preventing wanton environmental and ecologic
abuse.
Methods: Descriptive time lines for recent genocides and their precursors, population resource ratios for settings from recent genocides and settings without genocides, case studies of effects
of wanton ecologic abuse on vulnerable populations.
Results: The histories of recent genocides (former Yugoslavia, Rwanda and Darfur) confirm that (a) genocide cannot occur without the ideology and decisions of its authoritarian perpetrators,
a lopsided intergroup power equation and indifference or indecision of outside bystanders; (b) simple population-land ratios do not appear to have explanatory power for vulnerability to
genocide and mass atrocities; (c) Malthusian pressures, combined with choices to exacerbate zero-sum rivalries over water, arable land, or natural resources may have exacerbated the risks
from known political and socio-economic predictors in Rwanda and Darfur, but not in former Yugoslavia (Bosnia and Kosovo), which by themselves were not sufficient to result in genocide;
(d) collapse of socio-economic and governmental infrastructures following genocide produces massive sustained damage to carrying capacity and sustainability; (e) victims of mass atrocities,
if they return to their environments and continue to be persecuted, have more severe problems than those of environmental refugees; (f) wanton ecological abuse can induce flight, death,
injury, and multigenerational impacts, including genotoxic risks, without organized mass expulsions and violence.
Conclusions: Simple descriptive comparisons of recent genocides do not appear to support the hypothesis that Malthusian pressures based on simple or more complex resource ratios account
for genocide and mass atrocities. International early warning and effective response systems are needed to deter or prevent political decisions to carry out genocide. But these systems have
to include long-term measures to resolve zero-sum conflicts over environmental resources and prevent toxic risks to vulnerable populations from wanton ecologic abuse.




SAA2-O-04                  METHODS TO ANALYSE INDICATORS OF INTENTIONAL MASS VIOLENCE
J LINDERT1, O VON EHRENSTEIN2, S PRIEBE1
(1) Queen Mary, University of London, London. (2) University of Berkeley, Berkeley.
BACKGROUND: Genocide and war were the main causes of preventable death in the 20th century. Research has shown that organised violence develops in stages from the first stage (loss
of opportunities for the persecuted group) to the last stages (loss of existence, loss of memory) in the aftermath. Several concepts of stages exist, all agree on a continuum of violence.
Probability and intensity of subsequent stages increases aggression. Each stage can be measured with indicators i.e. changes in language, devaluation of others, polarisation, preparation and
denial. Further known indicators are the readiness to accept violence as a means for solving conflicts, inequality in distribution of resources and opportunities, level of discrimination and
change in the legal context. Until now data are scattered and remain often un - catalogued within countries. Accurate empirical estimates of the burden of organised violence are necessary
for anticipatory planning of prevention. We intent to provide a comprehensive set of early indicators and to introduce epidemiological tools for describing and understanding the historical
and spatial organisation of organised violence.
METHODS: Because each stage of organised violence has its own indicators, stages can be assessed with standardised measures by measuring specific indicators like i.e. percentage of
weapons in a country and number of riots. We present geographical information systems (GIS) combined with sociological and psychological methods as a new interdisciplinary violence
surveillance system to map and analyse rates and distributions of genocide indicators. Furthermore we introduce a comprehensive set of standardised indicators based on interdisciplinary
research for surveillance of violence. By applying risk categories using GIS to population data areas of differing risk can be defined. In an effort to integrate existing mapping activities we
propose a global partnership of violence mapping activities.
RESULTS: Historical examples i.e. from Germany, from the Soviet Union, from Kampuchea show the utility of this method. Other actual examples will be given in the presentation and countries
will be analysed with the standardised set of indicators.
CONCLUSION: Case studies show that using GIS can be useful for scoring profiles of risk, based on a priory defined quantitative criteria. But there is a need for safeguards against abuse,
stereotyping, using models for hating up situations. Genocide and violence control programmes could operate within this framework for mapping organised violence at early stages collecting
in a standardised way data to prevent violence and to identify settings at risk.




                                                                                              35
         Oral Communication Session

SAA2-O-05                   WHAT DID JOHN SNOW KNOW? A CAUTIONARY APPROACH TO THE
                            PRECAUTIONARY PRINCIPLE
B GOLDSTEIN1
(1) University of Pittsburgh Graduate School of Public Health, Pittsburgh.
John Snow is justly revered for his careful analysis of data that led to a successful preventive intervention and established the scientific basis for epidemiology, particularly environmental
epidemiology. Advocates of the Precautionary Principle often cite Snow’s work as an example of effective precautionary action – which it certainly is. Unfortunately, some of the same advocates
go one step further and claim that since Snow did his work before the establishment of the germ theory of disease, precautionary actions need not be based on mechanistic understanding
of disease processes. Further, mere observation of an association is sufficient for action, and research into understanding of cause and effect relationships has at best a low priority. Yet Snow’s
action was firmly based upon his knowledge of scientific mechanisms involving the respiratory system that allowed him to discard the prevalent notion that cholera, and other contagious
diseases, were necessarily spread through the air. Most of Snow’s work would now be called pulmonary physiology and toxicology. His thorough study of the basic workings of the lung included
evaluation in laboratory animals of respiratory mechanics and gas exchange. Snow’s studies of inhalation toxicology included relating together the inhaled dose of chloroform, blood
chloroform levels and the degree of central nervous system narcotism. These interests led to his being one of the foremost experts in the then new field of anesthesiology. A milestone in the
acceptance of anesthesia by the general public was the submission of Queen Victoria to obstetrical anesthesia in 1853, one year before the Broad Street Pump episode for which Snow is now
so well known. Snow was chosen to be Queen Victoria’s anesthesiologist. The episode of the Broad Street Pump does not provide ammunition for those who in the name of the precautionary
principle would downplay the importance of mechanistic science in public health and prevention. Rather, it illustrates the adage that science consists of going up alleys to see if they are blind.
Snow had learned enough about basic pulmonary physiology and toxicology to realize that that inhalation of a miasma was a blind alley for understanding the cause of cholera, something
that was not obvious to anyone else at the time. Public health interventions are most likely to be effective when they occur in the context of basic scientific understanding.




SAA2-O-06                   DEALING REFLEXIVELY WITH UNCERTAINTY IN ENVIRONMENTAL HEALTH
                            ISSUES.
M CRAYE1
(1) European Commission Joint Research Centre, Ispra.
Institutions increasingly recognize that uncertainty in relation to environmental health issues can no longer be suppressed or denied, but has to be dealt with in an appropriate way.
The experience of intractable controversies, such as those on the effects of dioxins, endocrine disruptors,….inspired such institutional action.
In depth analyses of these policy conflicts clearly show they are not merely the consequence of the absence of conclusive factual information. In these exemplary cases, uncertainty in the
knowledge base goes hand in hand with plural risk framings and diverging interests. A strict ‘science based decisionmaking’ logic rather intensifies than appeases these conflicts and is
inadequate to cope with strategic behaviour towards uncertainty (as shown by the flourishing consultancy business of ‘manufacturing uncertainty’ in relation to workplace hazards and
environmental health impacts in the US).
Following these in depth analyses, there is a need to develop reflexive alternatives to the dominant approach to uncertainty. This dominant approach limits the assessment of scientific
uncertainty to technical analyses downstream in the scientific research process, i.e. quantification - in terms of ranges and error-bars – of what is seen as ‘unsolvable’ uncertainty.
Reflexive alternatives put the issue of uncertainty within a framework of deliberative governance (involvement of stakeholders and civil society). They aim at increasing insight in the nature
and sources of expert disagreement and its relation with the socio-political context.
We developed a discussion protocol and a discussion scheme to interactively explore uncertainty in relation to different scientific framings, societal perspectives and policy options. The
discussion scheme is based on the concept of ‘pedigree of knowledge’. The discussion protocol has to to guarantee conditions for a reasoned debate.
The approach has been deployed to assess scientific studies, that had been produced in the context of a socio-political debate on possible health effects from waste incineration. The results
obtained show the approach has potential to trigger a profound social debate and a negotiated management of risk. Pro-active use of the approach could enhance the quality and robustness
of the knowledge input in policy making.
In the near future, the approach will be applied to assess uncertainty and expert disagreement in relation to the health effects of low dosis ionizing radiation.




SAA2-O-07                   THE PERILS OF ONE-SIDED RISK ASSESSMENT AND PRECAUTION: THE CASE
                            OF POPS IN FISH
T ASSMUTH1
(1) Finnish Environment Institute, Helsinki.
In environmental and health policy and analysis a common belief is that precaution, as outlined in the precautionary principle (PP), automatically will lead to more safety. This fallacy essentially
arises from a vague and one-sided definition of precaution. POPs are often hailed as a prime case justifying PP. However, as the risks from many POPs have been much reduced, through mixed
schemes of precaution, a major shift has occurred. Unintended effects or counter-veiling risks from actions meant to be precautionary have become prominent. In the case of fish, they
especially include loss of its benefits to health, possibly far exceeding the health risks from its POPs (Assmuth and Jalonen 2005). A more comprehensive and varied consideration of risks
(including loss of opportunity), impacts and consequences, as well as agents or factors causing risks is thus called for.
The EU approach to fish POPs has been market regulation based on dioxin levels in foods, instead of the traditional dietary advice and instead of preventing dioxin formation. Only recently
have health benefits been addressed (SPCFC 2005); fisheries management has not been integrated. This may in part reflect the inability of the regulatory system to broadly consider complex
risks and impacts, and pressures from the public and others driven by fears for dioxins in food (more than for loss of benefits from foods), i.e. asymmetric and misguided precaution. Those
stressing the health benefit argument have on the other hand often simplified it too far, downplaying e.g. qualities, distributions and uncertainties of risks and benefits, and sensitivity to
options like healthy dietary alternatives. Instructive parallels are seen with other cases of comparative risk assessment and debates of ‘rational risk reduction’ (e.g., Finkel 1995). On both sides
in the debate, illusionary belief reigns in knowability, and scientific authority is invoked. Uncertainties are on the other hand also exaggerated. TDIs for dioxins and R/BA in fish consumption
are discussed as examples. POPs and their effects are lagged, as are perceptions of their risks and the societal responses. Some over-reaction takes place to risks already (being) phased out,
due in part due to stigmatization of classic POPs as symbolic boundary objects, and can result in ‘overdoing’ PP.
The case analysis is generalized to other areas of environmental and health risk management, emphasizing the following science-policy links: a) communication and deliberation between
sectors and interest groups; b) navigation between panic action and ‘paralysis by analysis’: c) intermingling of values and facts. Broad consideration of risks, impacts and uncertainties is viewed
as a ‘reality check’ for both those inflating (some) risks and those deflating (some) risks, and a key challenge for development of PP-science interactions. Some pointers for such a development
are given.
References: Assmuth T, Jalonen P. Risks and management of dioxin-like compounds in Baltic Sea fish: An integrated assessment. Copenhagen, Nordic Council of Ministers. TemaNord 2005:568.
376 p. Finkel AM. 1995. Toward less misleading comparisons of uncertain risks: the example of aflatoxin and alar. Environ. Health Perspect. 103;4:376. SPCFC. 2005. Opinion of the Scientific
Committee on Contaminants in the Food Chain on a request from the Eur. Parliament related to the safety assessment of wild and farmed fish. Question N EFSA-Q-2004-23. EFSA J. 236:1.




                                                                                                36
                                                                                                                                                                                                       SUNDAY SEPTEMBER 3
        Oral Communication Session

SAA2-O-08                   TOWARDS A COMMON UNDERSTANDING OF THE PRECAUTIONARY PRINCIPLE
JP VAN DER SLUIJS1, M KAISER2
(1) Copernicus Institute for Sustainable Development and Innovation, Utrecht University, Utrecht. (2) National Committee for Research Ethics in Science and Technology
(NENT), Oslo.
Over the past decades, the Precautionary Principle (PP) has become an underlying rationale for a large and increasing number of international treaties and declarations in the fields of
sustainable development, environmental protection, health, trade, and food safety. In its most basic form, the PP is a strategy to cope with scientific uncertainties in the assessment and
management of risks. It is about the wisdom of action under uncertainty. Precaution means taking action to protect human health and the environment against possible danger of severe
damage. However, in the international arena, different views exist of what precaution is and the PP has different interpretations.
The work to be presented has been carried out by an international expert group and aims to reduce the gaps in the understanding of the principle and to clarify the PP for decision-makers
and scientists in order to achieve a more informed debate of the principle and to serve as reference for possible further implementations of the PP. It sketches the history of the PP, reviews
concepts and definitions of the PP and identifies common elements in the various definitions. On that basis a new working definition of the PP is presented, formulating the PP on the basis
of positive criteria, including the demand that the possible harm referred to, in spite of being uncertain, needs to have some scientific backing. Furthermore, it allows for a wide range of
precautionary actions, provided they appear effective in order to either avoid or diminish the possible harm. This answers the criticism that the PP is too narrow a tool for innovation policy as
long as it only provides the go or no-go options.
The ethical basis of the PP and the questions of responsibility, inter- and intra-generational equity and deliberative democracy are explored and legal issues discussed. Special focus is put on
the characteristics of complex systems and the concepts of robustness and resilience as well as the multiple dimensions of uncertainty in scientific assessment. These uncertainties are at the
heart of the PP. The concept of risk is dealt with and associated decision-making problems for which the PP can be helpful.
The paper further addresses a range of application issues of the PP: implications of the PP for science, implications for policy and governance, implications for industry and trade and social
and cultural implications of the PP.




SAA5-O                      AIR POLLUTION: PREGNANCY OUTCOMES AND CHILDREN HEALTH
SAA5-O-01                   A COHORT STUDY OF AIR POLLUTION IMPACTS ON BIRTH OUTCOMES
M BRAUER1, C LENCAR1, L TAMBURIC2, M KOEHOORN3, E NETHERY1, P DEMERS1, C KARR4
(1) School of Occupational and Environmental Hygiene, The University of British Columbia, Vancouver. (2) Centre for Health Services and Policy Research, The University
of British Columbia, Vancouver. (3) Department of Health Care and Epidemiology, The University of British Columbia, Vancouver. (4) Department of Environmental and
Occupational Health Sciences, University of Washington, Seattle.
Introduction: Evidence is accumulating that gestational exposure to air pollution adversely affects pregnancy outcomes including reduced birth weight and gestational age. However, data
are mixed regarding the importance of individual pollutants and the timing of exposure. Few studies have focused on individual-based spatial contrasts in exposure which we hypothesize to
be more informative than exposure based on regional monitoring data. We evaluated the impacts of air pollution during discrete exposure windows on low birth weight for gestational age
and premature birth.
Methods: Using the Border Air Quality Study’s administrative data cohort we identified 75,078 singleton births (from a total of 84,242 during 1999-2002) with complete covariate (gender,
Infant/Mother ethnicity [First Nations Status], parity, Month-yr of birth, income, maternal education) and maternal residential history data. Exposures were estimated for the home address of
the mother by month of pregnancy. Using logistic regression, we estimated the risk of mean (entire pregnancy, first and last month / first and last 3 months of pregnancy) air pollution
concentrations on intrauterine growth retardation (IUGR: <10th percentile of birth weight for the full cohort at a given gestational age) and pre-term birth (<37 weeks) using inverse-distance
weighting of all monitors in the study area for CO, NO2/NOx, O3, PM10 and PM2.5, as well as temporally-adjusted land use regression (LUR) models for NOx and Black Carbon (filter absorbance).
Results: PM2.5 and Black Carbon concentrations were associated with IUGR (e.g. During the first 3 months of pregnancy, PM2.5 OR=1.14, 1.04-1.24 per 1 µg/m3 increase, Black Carbon OR=1.05,
1.01-1.09 per 0.5x10-5 m-1 increase in filter absorbance). ORs were slightly elevated for ambient monitoring estimates of PM10, NO2, NOx and to a lesser extent for LUR estimates of NOx.
Analyses of preterm births did not reveal statistically significant associations, with the exceptions of PM2.5 during early pregnancy (e.g. OR =1.20, 1.08-1.35, per 1 µg/m3 increase in first three
months), and LUR estimates of NOx for all exposure windows.
Discussion: Overall, risk estimates derived from LUR models were not consistently higher than inverse-distance weighted monitor based estimates. Analyses indicated significant associations
between PM2.5 in early pregnancy with both IUGR and preterm birth. For other pollutants, no consistent patterns suggested specific exposure windows of greater relevance. As these results
indicate associations between air pollution and birth outcomes in a large population-based cohort, future efforts should be directed to defining exposure sources and potential mechanisms
of toxicity, with emphasis on early pregnancy.




SAA5-O-02                   EXPOSURE TO TRAFFIC RELATED ATMOSPHERIC POLLUTANTS DURING
                            PREGNANCY AND TERM BIRTH WEIGHT: A COHORT STUDY RELYING ON A
                            STOCHASTIC EXPOSURE MODEL
R SLAMA1, V MORGENSTERN2, J BOUYER1, J HEINRICH2
(1) Inserm and INED (U569), “Epidemiology, Demography and Social Sciences”, 94276 Le Kremlin-Bicêtre. (2) GSF-National Research Center for Environment and
Health, Institute of Epidemiology, 85764-Munich.
Introduction: Our aim was to characterize the influence of maternal exposure to traffic-related air pollutants during pregnancy on term birth weight.
Methods: Women from the Munich LISA cohort who delivered singleton non-low birth weight babies after 37 gestational weeks in 1998 and 1999 were questioned after delivery. Women
who had moved out during pregnancy (19% of the cohort) were excluded. A stochastic (regression) model predicted mean atmospheric pollutant levels. This exposure model was defined using
measurements at 40 sites across the city; the home addresses were geocoded and the length of each type of road, household density, and land coverage in the surrounding of the home were
extracted from a Geographic Information System and used to predict annual atmospheric pollution levels at each home. The probability of giving birth to a child weighing less than 3,000 g
was analyzed by a Poisson regression model with bootstrap-estimated confidence intervals (CI), adjusted for gestational duration, sex, maternal smoking, maternal size and pre-pregnancy
weight. As the exposure model did not incorporate seasonal variations, we repeated the analyses after introduction of interaction terms between the pollutant and the season of birth.
Results: Air pollutant levels were estimated for 1,016 live births, out of which 142 weighed less than 3000 g (14.0%). The adjusted prevalence-ratios (PR) of birth weight below 3000 g
comparing the lowest and highest quartile of the pollutants levels were 1.46 for PM2.5 concentration (95% CI, 0.98; 2.30), 1.44 for PM2.5 absorbance (95% CI, 0.93; 2.39) and 1.07 for NO2
concentration (95% CI, 0.73, 1.62). The estimates were higher among the 278 births that occurred from January to March. In this group, the adjusted PRs of birth weight below 3000 g
corresponding to the highest exposure quartile were 2.81 for PM2.5 (95% CI, 1.38; 6.11), 1.57 for PM2.5 absorbance (95% CI, 0.72; 3.57) and 2.01 for NO2 (95% CI, 0.96; 4.28).
Discussion: This study is to our knowledge the first to use a stochastic model to assess individual exposure to traffic-related atmospheric pollutant levels in relation to birth weight. A
deleterious effect of pollutant levels on birth weight was clearer for births occurring during the first trimester of the year, which corresponds to the end of the period with the highest air
pollution levels. This illustrates the need to incorporate seasonal effects in stochastic models estimating pollutant levels when the health outcome is likely to be influenced by short-term
variations in pollution.




                                                                                                37
        Oral Communication Session

SAA5-O-03                  RELATION BETWEEN ATMOSPHERIC POLLUTANTS AND HEAD CIRCUMFERENCE
                           IN UTERO AND AT BIRTH: A COHORT STUDY RELYING ON ULTRASOUND
                           IMAGING DURING PREGNANCY
R SLAMA1, S SINNO-TELLIER1, O THIÉBAUGEORGES2, V GOUA4, A FORHAN3, B DUCOT1, I ANNESI-MAESANO5, J HEINRICH6, M SCHWEITZER2, G
MAGNIN4, J BOUYER1, M KAMINSKI7, MA CHARLES3, & EDEN STUDY GROUP8
 (1) Inserm and INED (U569), Le Kremlin-Bicêtre. (2) Maternité Régionale de Nancy, Clinique Universitaire de Gynécologie Obstétrique, Nancy. (3) Inserm U780, Villejuif.
(4) CHU de Poitiers, Service de Gynécologie Obstétrique, Poitiers. (5) Inserm U707, Paris. (6) GSF-National Research Center for Environment and Health, Institute of
Epidemiology, Munich. (7) Inserm U149, Villejuif. (8) Maternité Régionale de Nancy, CHU de Poitiers, Inserm IFR69, Villejuif.
Introduction: Our aim was to study the influence of exposure of pregnant women to atmospheric pollutants on fetal growth assessed during pregnancy and at birth.
Methods: The population is part of the ongoing Eden cohort, conducted in two French maternity hospitals. Pregnant women were included before 24 gestational weeks. Head circumference
was assessed with standardized ultrasound measurements on two occasions during the second and third trimesters, and at birth by a midwife. Atmospheric pollutant levels (NO2) in the vicinity
of the home were estimated from the measurements of the nearest background monitoring station and averaged between the dates of fertilization and of each examination. Women who
moved out during pregnancy were excluded. Birth weight and head circumference were analyzed with a linear regression model adjusted for gestational age at the examination, sex, maternal
tobacco use, maternal size and pre-pregnancy weight, parity, gestational diabetes and centre. Head circumference was also dichotomized (above versus below first quartile of its distribution)
and studied by logistic regression.
Results: Exposure to NO2 was assessed for 366 women living less than 2 km away from an air monitoring station, out of 1641 live births. Adjusted mean birth weight decreased by 130 g
(95% confidence interval, CI, 18; 242) in the highest tertile of exposure ([NO2]>31.4 µg/m3), compared to the lowest tertile. Adjusted mean head circumference at birth decreased by 3.1 mm
(95% CI, -1.0; 7.3) in the highest tertile of exposure to NO2, compared to the lowest tertile. For head circumference assessed by ultrasound imaging at 30-34 gestational weeks, the mean
adjusted decrease was 3.8 mm between the first and third tertiles of exposure (95% CI, -0.9; 8.5). The odds of a head circumference at birth below 34 cm was multiplied by 2.6 (95% CI, 1.2;
5.2) in the highest exposure tertile; the corresponding odds-ratio (OR) was 2.0 (95% CI, 0.8; 4.9) when birth weights below 3 kg were excluded. The adjusted ORs of a head circumference
below 29 cm at 30-34 gestational weeks and below 19 cm at 20-24 gestational weeks associated with the highest exposure tertile were respectively 2.9 (95% CI, 1.2; 6.9) and 2.1 (0.8; 5.6).
Discussion: To our knowledge, this study is the first to describe the relation between atmospheric pollutants and intra-uterine growth assessed by ultrasound during pregnancy. High NO2
concentration was associated with a decreased head circumference at birth and in utero, possibly as early as at 20-24 gestational weeks.




SAA5-O-04                  ADVERSE BIRTH OUTCOMES AND EXPOSURES TO AIR POLLUTANTS ASSESSED
                           BY SPATIAL AND TEMPORAL MODELING IN INCHEON, REPUBLIC OF KOREA
JH LEEM1, B KAPLAN2, Y SHIM2, C TYLENDA3
(1) Department of OEM, Inha University, Incheon. (2) Division of Human Studies, ATSDR, Atlanta. (3) Division of Toxicology and Environmental Medicine,ATSDR, Atlanta.
Introduction: The association between adverse birth outcome and exposure to air pollutants has recently become a major concern. We investigated this relationship in Incheon, Republic of
Korea, using spatial and temporal modeling to better infer individual exposures
Methods: The birth cohort data consisted of 52,113 singleton births in 2001-2002 and included residential address, gestational age, gender, birth date and order, parental age and education.
A geographic information system and kriging methods were used to construct spatial and temporal exposure models. Associations between exposure and adverse birth outcome were
evaluated using univariate and multivariate log binomial regressions. Based on the gestational age, birth date, and the mother’s residential address, each mother’s potential exposure to air
pollutants during critical periods of the pregnancy was estimated. The effect sizes of air pollutants assessed by spatial and temporal exposure models were compared with those by temporal
models.
Results: The adjusted risk ratios for preterm delivery in the highest quartiles of the first trimester exposure were a) 1.26 (95% CI: 1.11-1.44) for CO, b) 1.27 (95% CI: 1.04-1.56) for PM10, c)
1.24 (95% CI: 1.09-1.41) for NO2, and d) 1.21 (95% CI: 1.04-1.42) for SO2. The adjusted risk ratios for low birth weight in the highest quartiles of the first trimester exposure were a) 0.86
(95% CI: 0.73-1.02) for CO, b) 1.13 (95% CI: 0.97-1.33) for PM10, c) 1.16 (95% CI: 0.99-1.35) for NO2. d) 1.05 (95% CI: 0.90-1.23) for SO2.
Discussion and conclusions: Our study showed that relatively low concentrations of air pollution under current air quality standards during the pregnancy may contribute to an increased
risk of adverse birth outcomes, such as preterm delivery and low birth weight. Further studies about biological mechanisms of these kinds of adverse birth outcomes are needed.




SAA5-O-05                  AIR POLLUTION AND ADVERSE PREGNANCY OUTCOMES: WHAT IS THE ROLE
                           OF GENDER? A SYSTEMATIC REVIEW
R GHOSH1, T PLESS-MULLOLI1, J RANKIN1, S GLINIANAIA1
(1) UNIVERSITY OF NEWCASTLE UPON TYNE, NEWCASTLE UPON TYNE.
Introduction: Gender differences in pregnancy outcomes e.g. low birth weight (LBW), preterm birth (PTB) and fetal growth retardation are well established. Recent studies on air pollution
and adverse pregnancy outcomes suggest an increase in risk, albeit small. We undertook a systematic review to study the interaction between gender and air pollution on pregnancy outcomes.
Methods: English language articles were retrieved from Medline, Embase, Science Citation Index and other environmental databases using a comprehensive list of keywords relating to gender,
air pollution and pregnancy outcomes (1966 to 2005).
Studies were included if:
exposure measurements were well defined - personal monitoring, approximation from ambient levels, modelled estimates and proximity to pollution sources
pollutants were specified: • effects on pregnancy outcomes were reported • effect estimates were reported by gender
Studies were excluded if they included: • accidental or occupational exposures multiple births
A tailored checklist based on Bracken’s guideline for observational studies, was used to assess methodological quality. Quality was assessed based on the applicability of the findings and
validity of the individual study results. We had concern about exposure measurements in two studies while some others did not adjust for socio-economic status and maternal smoking. Overall
the quality of the included studies was considered satisfactory.
Results: A total of 11 studies, conducted between 1986 and 2003, were included. Of these, four evaluated LBW; one each evaluated very low birth weight (VLBW) and fetal growth, and six
examined PTB. PTB studies were more homogeneous in terms of study design and analysis than the LBW studies. Female babies were at higher risk of LBW (AORs ranged from 1.07-1.62) and
male babies were found to be at higher risk for PTB (AORs ranged from 1.11-1.20) from the adjusted logistic models. In addition, there was some evidence of an interaction between gender
and air pollution for LBW when the highest and lowest exposure categories were compared separately for male and female babies. The risk of LBW, which is commonly high for female babies,
was higher in males after accounting for interaction; however the results were not statistically significant.
Discussion and conclusions: This review provides some limited evidence of an interaction between air pollution and gender on pregnancy outcomes. Any interactive effect should be
interpreted with caution in light of study limitations such as exposure misclassification, multiple comparisons and residual confounding. Further studies using high quality datasets are needed
to ascertain the effect.




                                                                                              38
                                                                                                                                                                                                                         SUNDAY SEPTEMBER 3
         Oral Communication Session

SAA5-O-06                     AMBIENT AIR POLLUTION, MATERNAL SMOKING AND RESPIRATORY RELATED
                              INFANT MORTALITY
T WOODRUFF1, L DARROW2
(1) US Environmental Protection Agency, San Francisco. (2) Emory University, Atlanta.
Introduction: Ambient levels of air pollution have previously been observed in association with postneonatal infant mortality due to respiratory causes. We further investigated this hypothesis
in 96 counties throughout the U.S., including an assessment of confounding by maternal smoking status.
Methods: Birth and infant death data for all U.S. births over the study period 1999-2002 were obtained and linked to particulate and gaseous air pollution monitoring data. Pollution exposure
was calculated as the average county concentration during the first two months of the infant’s life. The occurrence of respiratory postneonatal infant mortality was evaluated in relation to air
pollution levels using generalized estimating equations, adjusting for maternal, temporal, and regional factors. Using the subset of birth records which included maternal smoking status, we
ran models with and without smoking to assess potential confounding by maternal smoking. In addition, we assessed the relationship between indices of air pollution and respiratory death
after stratification by birthweight status: low birth weight (<2500 grams) and normal birth weight (?2500 grams).
Results: After linkage and exclusions, the study population consisted of 3.6 million births with 404 respiratory related deaths in 96 counties. In single pollutant models, postneonatal respiratory
death was associated with PM10 (per interquartile range increase OR= 1.16; 95% CI, 1.04-1.29), PM2.5 (OR=1.05; 95% CI, 0.87-1.27) and CO (OR=1.14; 95%CI, 0.93-1.39). The subset of birth
records containing maternal smoking data included 2.5 million births, with 333 postneonatal respiratory deaths. Controlling for smoking in the model had no effect on the parameter estimates
for any of the pollutants: PM2.5, CO or PM10. When the study population was stratified by birth weight status, effect estimates for PM2.5, CO and PM10 were elevated in both strata, however the
association was stronger for all pollutants in the normal birth weight stratum. In multipollutant models, only PM remained elevated and significant.
Discussion: These findings support a relationship between particles and postneonatal infant respiratory death. Results suggest that effects may be stronger in normal birth weight infants;
possible explanations for effect modification by birth weight will be discussed. This study provides no evidence that the relationship between county level air pollution and postneonatal
respiratory death is confounded by maternal smoking status. *Views of the authors and not of their institutions.




SAA5-O-07                     HEALTH IMPACT ASSESSMENT OF PM10 ON MORTALITY AND MORBIDITY IN
                              CHILDREN IN CENTRAL EASTERN EUROPEAN CITIES
A PALDY1, J BOBVOS2, M LUSTIGOVA3, H MOSHAMMER4, EM NICIU5, P OTOREPEC6, V PUKLOVA3, K SZAFRANIEC7, T ZAGAR GALE8, M NEUBERGER4,
R KUBINOVA3, F PASCAL10, N STILIANAKIS9, S MEDINA10, D DALBOKOVA10, M KRZYZANOWSKI10
(1) “Fodor Jozsef” National Center for Public Health, National Institute of Environmental Health, Budapest. (2) Capital Institute of the National Public Health Officer’s Service, Budapest. (3) National Institute
of Public Health, Prague. (4) Institute of Environmental Health, Medical University of Vienna, Vienna. (5) National Institute of Public Health, Bucharest. (6) Institute of Public Health, Ljubljana. (7) Jagiellonian
University, Cracow. (8) National Institute of Oncology, Ljubljana. (9) Joint Research Centre, Institute for Environment and Sustainability, Ispra. (10) Institute for Public Health Surveillance, Saint Maurice Cedex.
Background and objective: The 4th Ministerial Conference on Environment and Health adopted the Children’s Environment Health Action Plan in 2004. The aim of regional priority goal III
is to prevent and reduce respiratory diseases due to outdoor air pollution, and the frequency of asthmatic attacks. The objective of this study was to estimate the potential benefits in terms
of childhood deaths and morbidity that could be prevented by improving ambient air quality in Central-Eastern European cities of the Apheis network participating in the ENHIS programme.
Methods: Exposure-Response Functions for HIA in children were selected for long-term effects of PM10: total and respiratory postneonatal mortality (Lacasaña et al 2005), postneonatal
Sudden Infant Death Syndrome (SIDS) (Woodruff et al. 1997), and for short-term effects we looked at hospital respiratory admissions in 0-14yr children (Anderson et al 2004).
Six Central-Eastern European cities summing 8 millions inhabitants provided data on PM10, mortality and morbidity of children for the year 2001. Attributable cases were estimated using
software created by the French Air Pollution and Health Surveillance Programme (PSAS-9). Two scenarios were used for long-term effects of PM10 and postneonatal mortality (total, respiratory
and SID syndrome): reduction of the annual mean value of PM10 to a level of 40 µg/m3, to 20 µg/m3 (Limit of 1999/30/CE Directive for 2010), and by 5 µg/m3. For HIA on short-term effects of
PM10 and hospital respiratory admissions in children under 15 year: reduction of PM10 levels to a 24-hour value of 50 µg/m3 in all days exceeding this value (Limit of 1999/30/CE Directive), to
20 µg/m3 and reducing by 5 µg/m3 of all the 24-hour values.
Results: The yearly mean values of PM10 ranged from 22.0 to 62.0 µg/m3. All other things being equal in the participating cities, reduction of the annual mean value of PM10 to a level of 40
µg/m3 would prevent yearly around seven cases of total postneonatal mortality in Cracow and Bucharest. A reduction to a level of 20 µg/m3 would prevent 21 total postneonatal deaths in the
six cities, one respiratory and two SID cases in three cities with available data. A short-term reduction of PM10 concentration to a 24-hour value of 20 µg/m3 would prevent 186 hospital
admissions for respiratory diseases in three cities.
Discussion: The results indicate that reductions in PM10 concentrations can beneficially protect children’s health in Central-Eastern European cities.
ACKNOWLEDGEMENTS: The ENHIS programme is supported by the European Commission DG SANCO (Grant: SPC 2003112). The Apheis programme was supported by the European
Commission DG SANCO programme of Community action on pollution-related diseases (contract Nos. SI2.131174 [99CVF2-604]; SI2.297300 [2000CVG2-607]; SI2.326507 [2001CVG2-602]
and the participating institutions in 12 European countries.




SAB1-O                        ECOSYSTEM AND HUMAN HEALTH
SAB1-O-01                     THE IMPACT OF EXPOSURE AND DOSE METRIC SELECTION ON THE CONDUCT
                              OF HUMAN AND ECOLOGICAL EXPOSURE ASSESSMENTS
T MCCURDY1, V SANDIFORD1, H RICHMOND1, S GRAHAM1
(1) US EPA, RESEARCH TRIANGLE PARK.
Introduction. There are numerous differences that have to be accounted for when undertaking human versus ecological exposure assessments. Some of the most important are addressing
(1) the different spatial scales involved in where the two “populations” of interest are located (e.g., regional v. local; urban v. non-urban); (2) the level of “population” aggregation involved
(e.g., species v. sub-group cohorts); (3) the issue of mobility (people move in time and space while vegetative ecosystems generally are fixed in space), and (4) the time scale of exposures
(e.g., episodic v. long-term, acute v. cumulative). Many of these issues affect what exposure and dose intake metrics are used in an exposure assessment, and these metrics are the topic of
this presentation.
Methods. A comparative analysis of EPA’s 1997 and 2006 evaluations of the primary and secondary national ambient air quality standards (NAAQS) for ozone is used to highlight important
differences and similarities in exposure and intake dose metrics appropriate for evaluating human health and ecological system impacts.
Results. While both assessments explicitly addressed all four “dimensions” of exposure and intake dose metrics (intensity, duration, frequency, and pattern), there were significant differences
in the actual values used for each dimension. The different metrics used for ecosystem and human health resulted in quite different ozone standards being proposed in 1997 to protect human
health and ecosystem damage. No standards have yet been proposed for 2006.
Discussion. Metrics greatly affected the content and scale of the assessments undertaken and how they were reviewed by the scientific community. In addition, selection of the metrics used
in the two types of assessments had significant and far reaching impacts on EPA’s procedures and the focus of subsequent new research.




                                                                                                         39
        Oral Communication Session

SAB1-O-02                   INTEGRATED MULTIMEDIA DECISION MAKING FOR HUMAN AND ECOLOGICAL
                            RISK ASSESSMENT: A NATIONAL SCALE STUDY OF LAND APPLICATION OF
                            ARSENIC BEARING WASTES
J BABENDREIER1
(1) U.S. Environmental Protection Agency, National Exposure Research Laboratory, Athens.
Introduction
Requisite to the development and application of sound waste management policies, decision-makers must discern the impact of a given waste management approach upon both human and
ecological receptors. To configure such policies solely upon the assessment of impact to humans, for example, may result in unacceptable risk to ecological receptors. Historically, intuition
reflecting the most likely affected community and exposure pathway of concern has guided the focus of risk assessments upon one or the other class of receptors. Presumably, this has been
well founded for single-medium assessments in light of relative toxicity for a given contaminant and location of receptors at a given site. For many contaminants though, it is actually the case
that significant risks are presented to both receptor classes, expressed through a variety of media/pathway exposures, where significance depends upon a host of decision variables the
decision-maker must consider and ultimately set for a given policy (e.g., distance from source, acceptable risk level, and % population protected).
Methods
The U.S. EPA’s Multimedia, Multipathway, and Multireceptor Risk Assessment (3MRA) model is capable of providing assistance in the identification of the determinant receptor class of concern
for a variety of exposure profiles in land-based disposal scenarios. Intrinsic to its integrated design, 3MRA facilitates the evaluation of multiple permutations of decision variables, and provides
additional abilities to evaluate uncertainty in risk within and across subpopulations defined by receptor class subgroups, habitats, and cohorts (e.g., residents, children; terrestrial habitats,
birds). In this work, using 3MRA and associated national, regional, and site-based data sets, an assessment of risks to human and ecological receptors is considered for land application of
Arsenic-bearing wastes across the conterminous United States.
Results
Acceptable concentration levels in waste disposed, and associated identification of the dominant receptor class of concern, are sensitive to, among other variables, the selection of acceptable
risk levels typically assigned for human and ecological receptors, and the maximum receptor distance from land application units considered.
Discussion and Conclusions
The primary issue exposed - lacking an integrated approach for assessment of human and ecological receptors in a multimedia world, a decision-maker cannot readily discriminate nor
document risks to a less rigorously considered receptor class. Multimedia, multipathway, and multireceptor assessments provide decision-makers with a more complete view of the landscape,
allowing for explicit consideration of the full suite of decision options, and associated levels of risks avoided and risks incurred by both human and ecological receptors.



SAB1-O-03                   APPROACHES TO HUMAN AND ECOSYSTEM EXPOSURE OF VULNERABLE AND
                            SENSITIVE POPULATIONS TO MERCURY
D MANGIS1, J JOHNSTON2, E SUNDERLAND3
(1) US EPA, RESEARCH TRIANGLE PARK. (2) US EPA, ATHENS. (3) US EPA, BOSTON.
Introduction: Both human and ecosystem exposure studies evaluate exposure of sensitive and vulnerable populations. We will discuss how ecosystem exposure modeling studies completed
for input into the US Clean Air Mercury Rule (CAMR) to evaluate the response of aquatic ecosystems to changes in mercury deposition will be used to improve human exposure modeling for
methylmercury.
Methods: The ecosystem study completed for CAMR utilizes the mercury models WASP, WCS, BASS, and SERAFM to evaluate depositional change dynamics in five freshwater case studies,
forecasting the impacts of mercury reductions to ecosystems and fish tissue concentrations.
An upcoming Lake Ontario study will evaluate the impacts of local US and Canadian mercury sources and long range mercury transport to fish of Lake Ontario and their impact on human
exposure. For the human exposure pathway we will present an example that utilizes an integrated modeling framework: HYSPLIT for the source receptor model, WASP for aquatic fate and
transport, BASS for trophic dynamics and bioaccumulation and MENTOR for human exposure to perform an integrated analysis. This study will allow predictions of how regional emissions
reductions can impact local human mercury exposures in the Lake Ontario basin.
Results: Results from the five freshwater case studies showed that most freshwater systems will achieve 90% of the benefits of the mercury emissions reductions as the result of CAMR in
2-3 decades. Some systems may respond faster (5-10 years), and watershed dominated systems will likely take 50 years or more to respond. Attenuation of methylmercury after load reductions
are calculated for northern pike and yellow perch by size class to illustrate body burdens across species and size classes that are cost prohibitive to sample effectively.
Discussion: The time lag in ecosystem response has a major effect on the benefits of regulations and how quickly these benefits are translated into human health benefits.
By coupling ecosystem process models developed to evaluate the impacts of mercury reductions on sensitive ecosystems with atmospheric source models, and human consumption models,
we can improve our human exposure risk analyses of mercury control scenarios, and better evaluate the impacts of local mercury hotspots to ecosystems and local fish consumptive human
populations. The next step is to take this to the coastal and oceanic systems to determine how much reduction of mercury is needed to protect coastal and ocean ecosystems, and humans,
since most mercury exposure comes from ocean fish consumption.




SAB1-O-04                   THE IMPACT OF DRYLAND SALINITY AND WATERLOGGING ON
                            MOSQUITO BORNE DISEASE IN WESTERN AUSTRALIA
A JARDINE1, P WEINSTEIN1, MDA LINDSAY2, CA JOHANSEN3, A COOK1
(1) School of Population Health, The University of Western Australia, Perth. (2) Mosquito-Borne Disease Control Branch, Western Australian Department of Health, Perth.
(3) Discipline of Microbiology, School of Biomedical and Chemical Sciences, The University of Western Australia, Perth.
Introduction
Anthropogenic changes in land use and land cover are primary drivers of ecological disruption, and have the potential to strongly influence human vulnerability to vector-borne diseases,
particularly those carried by mosquitoes. One million hectares in the south-west of Western Australia (5.5% of the region) are currently affected by secondary soil salinisation (dryland salinity)
and waterlogging caused by extensive clearing of native vegetation, and this is predicted to increase to 5.4 million ha (29%) by 2050. This has led to a steady rise in salt concentration in
inland aquatic systems, which progressively compromises ecosystem integrity and reduces aquatic biodiversity. Thus mosquito species with a high salinity tolerance may be favoured over
endemic freshwater-breeding species as available freshwater habitat and predator populations decrease. Extensive waterlogging may also increase the overall area of available mosquito
breeding habitat. One such salinity tolerant species is Ochlerotatus camptorhynchus, a major vector of Ross River virus (RRV) disease in temperate Australia. The relationship between salinity,
mosquitoes and disease distribution are currently being investigated in the heavily salinised Wheatbelt region of Western Australia.
Methods
Mosquito collections were undertaken in regions with varying salinity levels between October 2004 and December 2005 to investigate the seasonal dynamics of the mosquito populations
present. Associations were tested using correspondence analysis and non-metric multi-dimensional scaling (NMDS) methods. All blood-fed mosquitoes collected were tested by direct enzyme-
linked immunosorbent assay (ELISA) for blood-meal source identification. Serum neutralisation assay testing is currently being carried out to determine the prevalence of IgG antibodies to
RRV of residents within these regions. These data will be combined using GIS to analyse the overall impact of dryland salinity on mosquito-borne disease risk.
Results
Oc. camptorhynchus breeding was significantly associated with salinity affected sites. The primary bloodmeal source was sheep and marsupials, the latter of which are natural vertebrate hosts
for RRV. Results of the serosurvey will also be presented.
Discussion and Conclusions
Dryland salinity is significantly associated with breeding of the main mosquito vector for RRV in south-western Australia. Changes in disease risk in rural communities are currently being
evaluated by a concurrent seroprevalence survey. This study indicates the potential impact of environmental degradation on arboviral disease ecology in semi-arid environments. The risk of
vector-borne disease in this region of Australia is expected to escalate given the predicted increase in dryland salinity over the next 50 years.



                                                                                                40
                                                                                                                                                                                                                                                                                      SUNDAY SEPTEMBER 3
            Oral Communication Session

SAB1-O-05                              HUMAN AND ECOLOGICAL EXPOSURE SCIENCE: DIVERGENCE AND
                                       RAPPROCHEMENT
ARAUJO1, W SUTER2
(1) U.S. Environmental Protection Agency, National Exposure Research Laboratory, Research Triangle Park, NC. (2) U.S. Environmental Protection Agency, National Center
for Environmental Assessment, Washington, DC.
Introduction: To the extent that they had their common origins in concerns with adverse impacts of contaminants on the individual, human and ecological exposure sciences have elements and approaches in common; source, exposure, dose, effect is a joint paradigm.
Divergences between the fields occurred as ecosystems exposures began to address the impacts of non-contaminant stressors, the interactions among stressor and nonstressor components of environments, and stressor effects that manifest themselves at the levels of ecological
populations and communities, including the consideration of cascading exposures. Those requirements drove the ecological exposure field to develop spatially explicit methods for exposure assessment whereby both stressors and the organisms of interest are distributed on
and impacted by landscapes, as well as population monitoring and modeling tools to assess the implications of multiple stressors. In the meantime, the field of human exposure assessment focused on improved understanding of relationships between stressor sources, individual
behaviors, and exposure; biochemical and genetic markers of exposure, and the metabolic and physiologic pathways that determine internal dose. A number of models are available for exposure assessment; however, few are used as tools for both human and ecosystem risks.
This discussion will consider two modeling frameworks that have recently been used to support human and ecological decision making.
Methods: Recent years, however, have seen each field expand its treatment of scale and context for exposures and effects. Ecologists increasingly look to molecular indicators of exposures that correspond to population-level effects; human exposure researchers increasingly
examine how people and their behaviors are distributed unevenly over the landscape, creating disparities in both exposure intensities and in susceptibilities. The study will compare and contrast a Bayesian approach for pooling pre-implementation model forecasts with post-
implementation measurements to assess compliance with the relevant water quality standard with an integrated multimedia national-scale study of land application of arsenic-bearing wastes (the latter is presented separately at this conference). A specific application of each
model will provide a case study. The first case will address the U.S. EPA’s Multimedia, Multipathway, and Multireceptor Risk Assessment (3MRA) model from the standpoint of determining a variety of exposure profiles for receptor classes of concern for land-based arsenic
disposal scenarios. This approach will be compared to a present a Bayesian process for updating model output using subsequent monitoring data. The updated results are a combination of our understanding of the system (reflected in model results) and the information in the
data. The Bayesian SPARROW model was used to predict mean and standard deviation of the nitrogen loading as the basis for constructing the prior distribution of the nitrogen concentrations in the Neuse River Basin, North Carolina. The comparison will consist of a judgment
by the presenters of the advantages and disadvantages of these model applications for ecological and exposure assessments.
Results: This presentation takes stock of the histories, present states and futures of the fields of human and ecological exposure assessment and begins to frame a comparison of the fields.
Results for specific spatial and temporal cases are available for both approaches. The lessons learned for decision makers will consist of another level of results gathered from decision makers who have recently used these models in ecological and/or human exposure analyses.
Discussion and Conclusions: Subsequent presentations in this session will expand and substantiate these analyses using case studies to address similarities, differences, and the potential for cross-field synergies at scales ranging from the sub-cellular to the geographic. We
look forward to an interactive session that will improve exposure science by bringing together the human health and ecosystem exposure scientists to learn from each other; improve understanding how methods developed in one area can be translated to another; and improve
our understanding how to combine human and ecosystem models to achieve beneficial results. This discussion will address real-world applications of these models to ecological and human decision making, and how to improve data collection (e.g., sensors) and methods of
data analysis (e.g., pattern recognition). The discussion will also address advances in scientific knowledge, as well as the problems with using highly uncertain models results in decisions with unanticipated outcomes. For example, adaptive management, using a Bayesian
approach, offers an alternative that is immediately available.




SAB1-O-06                              A ROLE FOR ECOHEALTH APPROACHES IN PANDEMIC PREVENTION
DF CHARRON1
(1) International Development Research Centre, Ottawa.
The failure to control infectious diseases is a major international health issue. Long-standing problems such as infectious gastroenteritis, malaria, cholera and dengue fever exact an enormous
toll on the world’s poorest populations. Eradication earlier this century of smallpox and the development of increasingly potent antimicrobials contributed to the belief that most major
infectious diseases could eventually be controlled. Yet infectious diseases continue to be a scourge world-wide. Rich industrialized countries again fear pandemic disease like Severe Acute
Respiratory Syndrome (SARS) and Influenza. Many other pandemic diseases are a daily burden for developing countries. Pandemic spread of infectious diseases is characterized by the large
number of people infected, contagion and rapid spread, difficult treatment and containment, and occurrence in many geographic loci. Pandemic Influenza and SARS are perhaps best known
because of their high mortality and very rapid spread around the world from one geographic source. Cholera, dengue and HIV/AIDS easily qualify as pandemics, as do childhood infectious
gastroenteritis, malaria, and Hepatitis B virus. Pandemics are distinguishable by our inability or failure to contain disease spread.
Pandemics appear to emerge and persist in developing countries more so than in richer countries, this for several reasons. First, many diseases capable of pandemic spread are zoonoses
(transmissible between animals and people) that emerge in populations in intimate contact with livestock or wildlife. Secondly, densely populated human settlements with poor sanitation and
health care fuel pandemics. Thirdly, pandemics spread best if neglected for a time, this being more likely to happen outside of rich developed countries. Finally, new diseases appear to emerge
most readily in areas undergoing rapid environmental change, such change also preventing communities from being resilient to the population-level impacts of disease.
Much health research into pandemics is based on biomedical control strategies: vaccination, containment, quarantine, and eradication. These measures are proven effective in certain
circumstances (smallpox) but they do not favour upstream investigations into the emergence of disease. Ecohealth approaches tackle all drivers of pandemic spread – social, economic and
environmental. Ecohealth approaches could be used to understand how social conditioning surrounding animal rearing may favour the evolution of influenza. Participatory methods typical of
ecohealth elucidate new approaches that will be culturally appropriate yet effective in stopping the transmission of disease at some critical juncture (say, interrupting spread of influenza from
chickens to pigs and then to people). These and other strengths of the approach are discussed.




SAB1-O-07                              COMPARING THE UTILITY OF MULTIMEDIA MODELS FOR HUMAN AND
                                       ECOLOGICAL EXPOSURE ANALYSIS: TWO CASES
K RECKHOW1, R DIGUILIO1, J BABENDREIER2, D VALLERO2
(1) Duke University, Nicholas School of the Environment, Durham, NC. (2) U.S. Environmental Protection Agency, Research Triangle Park, NC.
Introduction
A number of models are available for exposure assessment; however, few are used as tools for both human and ecosystem risks. This discussion will consider two modeling frameworks that
have recently been used to support human and ecological decision making.
Methods
The study will compare and contrast a Bayesian approach for pooling pre-implementation model forecasts with post-implementation measurements to assess compliance with the relevant
water quality standard with an integrated multimedia national-scale study of land application of arsenic-bearing wastes (the latter is presented separately at this conference). A specific
application of each model will provide a case study. The first case will address the U.S. EPA’s Multimedia, Multipathway, and Multireceptor Risk Assessment (3MRA) model from the standpoint
of determining a variety of exposure profiles for receptor classes of concern for land-based arsenic disposal scenarios. This approach will be compared to a Bayesian process for updating model
output using subsequent monitoring data. The updated results are a combination of our understanding of the system (reflected in model results) and the information in the data. The Bayesian
SPARROW model was used to predict mean and standard deviation of the nitrogen loading as the basis for constructing the prior distribution of the nitrogen concentrations in the Neuse River
Basin, North Carolina. The comparison will consist of a judgment by the presenters of the advantages and disadvantages of these model applications for ecological and exposure assessments.
Results
Results for specific spatial and temporal cases are available for both approaches. The lessons learned for decision makers will consist of another level of results gathered from decision makers
who have recently used these models in ecological and/or human exposure analyses.
Discussion and Conclusions
This discussion will address real-world applications of these models to ecological and human decision making, and how to improve data collection (e.g., sensors) and methods of data analysis
(e.g., pattern recognition). The discussion will also address advances in scientific knowledge, as well as the problems with using highly uncertain models results in decisions with unanticipated
outcomes. For example, adaptive management, using a Bayesian approach, offers an alternative that is immediately available. The discussion will include interim findings and work in progress
(including sharing relevant data and plans).




                                                                                                                                      41
        Oral Communication Session

SAB1-O-08                  COMMON ISSUES IN HUMAN AND ECOSYSTEM EXPOSURE ASSESSMENT: THE
                           SIGNIFICANCE OF PARTITIONING, KINETICS, AND UPTAKE AT BIOLOGICAL
                           EXCHANGE SURFACES
T MCKONE1, W RILEY1, R MADDALENA1, R ROSENBAUM2, D VALLERO3
(1) Lawrence Berkeley National Laboratory, Berkeley, CA. (2) Swiss Federal Institute of Technology, Lausanne. (3) U.S. Enviromental Protection Agency, National Exposure
Research Laboratory, Research Triangle Park, NC.
Introduction
Exogenous chemicals enter organisms through critical surfaces in the lung, gills, gut, and skin. Transfer across these boundaries is the first step in characterizing the ratio of tissue dose to
external exposure. Surface processes and fugacity are important elements of both human and ecosystem exposure assessments.
Methods
Three cases will be analyzed to describe the chemical properties that determine transfer across biological surfaces and transport to target tissues. First, the dermal uptake for humans will be
compared to uptake on vegetation surfaces and how it relates to surface chemistry. The next step is to consider uptake in the oral and respiratory tracts of humans and gills of fish. Here both
metabolism and chemical properties can determine the blood levels in controlled exposure studies. Finally, a mass balance will be conducted on how uptake at the gut surface of cattle controls
the carry over of chemicals to meat and milk. Here again chemical properties, in particular lipid partition ratios, set maximum limits on the amount of a chemical that can be transferred from
ingestion to specific tissues.
Results
The case analyses will demonstrate the role of fugacity in both ecosystem exposure-response relationships and in human dose-response and exposure assessment approaches.
Discussion and Conclusions
The mechanisms and processes that occur in humans with those of other biota will be discussed, principally to improve the understanding of the types of fugacity used and that should be
used in exposure assessment paradigms in ecosystems versus those in humans. This will include discussion on how these processes change with scale and complexity of systems, as well to
compare the sensitivity of exposure in ecosystems and humans to various kinetic and equilibrium (partitioning) processes. Finally, discussions and recommendations will be dedicated to
designing algorithms and selecting data for both ecosystem and human exposure models; this will include any interim findings.




SAB2-O                     BIOMOLECULAR METHODS (I)
SAB2-O-01                  ESTIMATES OF AGE SPECIFIC URINARY EXCRETION RATES FOR CREATININE
                           AMONG CHILDREN
PP EGEGHY1, KW THOMAS1
(1) USEPA NATIONAL EXPOSURE RESEARCH LABORATORY, RESEARCH TRIANGLE PARK, NC.
INTRODUCTION: Poor compliance and high burden often make the collection of 24-hour urine samples impractical in large-scale studies involving young children. Instead, untimed spot or
grab samples are typically collected, and metabolite concentrations are adjusted for variations in water content of urine using creatinine concentration. Creatinine excretion, however, is
dependent upon muscle mass. Children’s muscle mass increases with age. Failure to appropriately account for creatinine excretion may result in “over-correction” of young children’s urinary
metabolite concentrations when compared against creatinine-adjusted samples from adults, or even from children who are just a few years older.
METHODS: Measurements in urine collected from children 3-12 years old in the Minnesota Children’s Pesticide Exposure Study (MNCPES), together with available information on void volumes
and void times, were used to estimate age-specific urinary excretion rates (UER) for creatinine with the following equation: UER = CuVu/(tc - t1), where Cu is the creatinine concentration, Vu is
the total void volume, tc is the collection time and t1 is the time of the last urination.
RESULTS: Estimated urinary excretion rates increased with each two-year age category. Creatinine excretion rates (mean ± SD) increased from 10 ± 4 mg/hr for 3-4 year olds up to 24 ± 12
mg/hr for 11-12 year olds. Similarly, urine output rates increased from 13 ± 6 ml/hr for 3-4 year olds to 19 ± 7 ml/hr for 11-12 year olds. Differences based on sex within the age categories
were neither substantial nor consistent, suggesting that sex is not an important predictor of creatinine excretion for pre-pubescent children.
DISCUSSION AND CONCLUSIONS: These results suggest that naïve adjustment by creatinine concentration, without consideration of the age-dependence of the physiological mechanisms
controlling its excretion, may introduce sizeable error and is inappropriate when comparing metabolite concentrations among different age groups.
Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy.




SAB2-O-02                  DERMAL EXPOSURE AND URINARY 1-HYDROXYPYRENE AMONG ASPHALT
                           ROOFING WORKERS
M MCCLEAN1, R RINEHART2, R HERRICK3
(1) BOSTON UNIVERSITY SCHOOL OF PUBLIC HEALTH, BOSTON. (2) OCCUPATIONAL SAFETY AND HEALTH ADMINISTRATION, WASHINGTON DC. (3) HARVARD
UNIVERSITY SCHOOL OF PUBLIC HEALTH, BOSTON.
Introduction: Historically, the assessment of occupational exposure to polycyclic aromatic compounds (PACs) has relied primarily on air monitoring; however, there has been increasing
evidence that dermal contact is a significant route of PAC exposure. The primary objective of this study was to identify determinants of dermal PAC exposure among asphalt roofing workers
and use urinary 1-hydroxyprene (1-OHP) measurements to evaluate the effect of dermal exposure on total absorbed dose.
Methods: The study population included 26 asphalt roofing workers who performed three primary tasks: tearing off old roofs (tear-off), putting down new roofs (put-down), and operating
the kettle at ground level (kettlemen). During multiple consecutive work shifts, dermal patch samples were collected from the underside of each worker’s wrists and were analyzed for PACs
and pyrene. During the same work week, urine samples were collected at pre-shift, post-shift, and bedtime each day and were analyzed for 1-OHP (219 urine samples). Linear mixed effects
models were used to evaluate the dermal PAC and pyrene measurements for the purpose of identifying important determinants of exposure, and to evaluate urinary 1-OHP measurements for
the purpose of identifying important determinants of total absorbed dose.
Results: Dermal exposures to PAC and pyrene were found to vary by roofing task (tear-off > put-down > kettlemen) and by the presence of coal tar pitch (presence > absence). The adjusted
mean PAC exposures associated with tear-off (no pitch: 159 ng/cm2/hr, pitch: 1,436 ng/cm2/hr) were 8 times higher than among kettlemen (21 ng/cm2/hr, 189 ng/cm2/hr) whereas the presence
of coal tar pitch was associated with an 9-fold increase in dermal PAC exposure. Similarly, rooftop workers (who perform both tear-off and put-down atop buildings) had post-shift urinary 1-
OHP concentrations (no pitch: 0.70 mmol/mol creatinine, pitch: 3.6 mmol/mol creatinine) that were significantly higher than kettlemen (0.47 mmol/mol creatinine, 2.44 mmol/mol creatinine)
while the presence of coal tar pitch was associated with an 5-fold increase in post-shift measurements of total absorbed dose. Additionally, the dermal measurements of both PAC and pyrene
were found to be significant determinants of urinary 1-OHP.
Conclusions: The presence of coal tar pitch was the primary determinant of dermal PAC exposure, though the observed task-based differences suggest that exposure to asphalt also
contributes to dermal PAC exposure. We found that dermal exposure was a significant determinant of total absorbed dose, suggesting that control strategies aimed at reducing occupational
exposure to PACs should include an effort to minimize dermal exposure.




                                                                                               42
                                                                                                                                                                                                     SUNDAY SEPTEMBER 3
        Oral Communication Session

SAB2-O-03                  URINARY EXCRETION OF MALONDIALDEHYDE AND 1-HYDROXYPYRENE IN
                           RESTAURANT WORKERS
CH PAN1, CJ CHEN2, CC CHAN3
(1) Institute of Occupational Safety and Health, Council of Labor Affairs, Executive Yuan, Taipei. (2) Institute of Occupational Safety and Health, Council of Labor Affairs,
Executive Yuan, Taipei. (3) Institute of Occupational Medicine and Industrial Hygiene, College of Public Health, National Taiwan University, Taipei.
Introduction: Restaurant workers may have the risk for occupational exposure to polycyclic aromatic hydrocarbons (PAHs) in cooking oil fume. Urinary malondialdehyde (MDA) and 1-
hydroxypyrene (1-OHP) are biological markers of oxidative damage and PAHs metabolism, respectively. We attempted to evaluate the relationship between MDA and 1-OHP in urine for
restaurant workers in this study.
Methods: We studied 314 non-smoker workers of Chinese type restaurants who had been employed for at least 1 year in northern Taiwan. All participants were asked to answer a
questionnaire which inquired their type of job, life style, health condition. After subjects completed the questionnaires during their annual physical examinations, a single urine sample was
collected. The annual physical examinations were arranged during the work shifts. All workers were asked to wash their hands prior to urine collection to avoid environmental contamination.
All urine samples were analyzed for 1-OHP and MDA by high performance liquid chromatograph (HPLC).
Result: For male kitchen workers, median MDA and 1-OHP concentrations (mmol/mol creatinine) were 287.4 and 2.0, respectively. These levels were significantly higher than those for male
banquet workers: 216.4 mmol/mol creatinine (p=0.024) and 0.4 mmol/mol creatinine (p=0.01), respectively. There was a significantly positive correlation between individual urinary MDA
concentrations and 1-OHP concentrations, after adjusting several potential confounders.
Discussion and Conclusions: Kitchen workers may have the high risk for occupational exposure to polycyclic aromatic hydrocarbons in cooking oil fume. We need further studies to confirm
this finding.




SAB2-O-04                  USING SALIVARY BIOMARKER FOR PESTICIDE EXPOSURE AND RISK
                           ASSESSMENT: POSSIBILITIES AND PITFALLS
C LU1
(1) Dept. Env. Occ. Health, Rollins School of Public Health, Emory U., Atlanta.
PLEASE DO NOT MODIFY THE TEMPLATE FRAME
otherwise the abstract will be rejected by the Programme Committee

Introduction: The use of salivary biomarker as a possible tool for measuring pesticide exposure has recently been investigated in a series of animal and human studies. Salivary biomarker, in
comparison to other specimen sample collections, is convenient and non-invasive. The measurements of pesticide concentrations in saliva have the potential of reliability to reflect levels in
plasma, and therefore render the best estimate of exposure. This advantage will allow salivary biomarker to be used for the purpose of absorbed pesticide dose or health risk estimation.
Method: Animal dosing studies using male Sprague Dawley rats were conducted to determine the feasibility of saliva biomonitoring for 3 different pesticides, atrazine, diazinon, and
chlorpyrifos. Pesticides were all intravenously injected, and a series of time-matched saliva and arterial blood samples were collected immediately after and up to 6 hours post pesticide
administration. Salivary biomarker was later collected from a total of 69 human participants, pesticide applicators and their children, in three separated studies conducted in the US, Nicaragua,
and Thailand. The purpose of these field studies was to validate the results obtained from the animal studies.
Results: Salivary concentrations of atrazine and diazinon were parallel to plasma concentrations in rats after intravenous injection. The correspondence of such concentration relationship was
independent from different doses and salivary flow rates. The results from the pharmacokinetic analysis indicated that salivary concentration of atrazine and diazinon could be used to predict
plasma concentration of both pesticides. Salivary diazinon concentrations were significantly correlated with the time-matched plasma samples collected from the same applicators, which is
in agreement with the results from animal models. However, different results suggested that salivary biomarker could not be used to estimate chlorpyrifos concentrations in plasma. Salivary
concentrations of chlorpyrifos could not be quantified 35 minutes post intravenous chlorpyrifos injection, whereas plasma concentrations could be measured as long as 350 minutes post
chlorpyrifos administration in rats. Saliva samples collected from pesticide applicators in Nicaragua clearly validated the results from the laboratory animal studies.
Discussion/Conclusion: Although, in theory, salivary biomarker can be used to assess exposure and estimate absorbed dose for those three pesticides that were studied. However, the
pharmacokinetics of chlorpyrifos prohibits the excretion of chlorpyrifos into saliva. The results from those animal and human studies not only has demonstrated that salivary biomarker can be
used to assess exposures to pesticides, but also provide the insight of its application to future exposure and health risk assessment.




SAB2-O-05                  ASSESSMENT OF MATERNAL AND INFANT EXPOSURES TO PESTICIDES AND
                           PERSISTENT POLLUTANTS
D BARR1, J KHOURY2, Y XU2, S BAEZ2, C BEARER3, K DIETRICH2, R HORNUNG2, M DAVIS1, L NEEDHAM1, B LANPHEAR2
(1) Centers for Disease Control and Prevention, Atlanta, GA. (2) Cincinnati Children’s Hospital, Cincinnati, OH. (3) Rainbow Children’s Hospital, Cleveland, OH.
Background: In utero exposures to environmental toxicants have been linked to a variety of adverse health outcomes including low birth weight and adverse neurodevelopmental effects.
Methods: The Health Outcome and Measures in the Environment (HOME) study, a prospective birth cohort involving 400 families in the greater Cincinnati area, was designed to test a variety
of health outcomes associated with environmental exposures. Women were enrolled at 16 weeks (+ 3 weeks) gestation. Maternal blood and urine samples were collected at 16 and 26 weeks
gestation and at delivery. We collected cord blood and meconium samples from the infant after birth. The biological samples were analyzed for a variety of environmental toxicants including
non-persistent pesticides, organochlorine pesticides and polychlorinated biphenyls (PCBs). Toxicants were measured using mass spectrometry-based methods. All analytes were quantified
using isotope dilution calibration. Typically, relative standard deviations of repeat analyses were less than 10%. Positive and negative control samples were included in all analytic runs to
ensure proper operation of the system.
Results: We detected metabolites of commonly used organophosphrous (OP) and pyrethroid insecticides in the majority of the maternal urine samples. Diethylphosphate (DEP), a common
metabolite of the OP pesticides chlorpyrifos and diazinon, was the most commonly detected OP metabolite in maternal urine. The geometric mean values of DEP ranged from 3.8 µg/L to 4.6
µg/L during pregnancy, which were ~ 4 times higher than concentrations found in a national (U.S.) reference population (NHANES). 3-Phenoxybenzoic acid (3-PBA) a common pyrethroid
metabolite, was detected in 98% of maternal urine samples. The range of 3-PBA, 0.76 µg/L to 0.93 µg/L, were 3 times higher than levels found in a national (U.S.) reference population. The
persistent organochlorines PCBs and p,p’-DDE (a metabolite of DDT) were detected in over 95% of maternal blood, cord blood and meconium samples. In a subsample of mother-infant pairs
with lipid corrected values (n=50), we found that mean maternal serum PCB concentrations were highly correlated with cord serum concentrations (r=0.67) and moderately correlated with
meconium PCB concentrations (r=0.38). Mean maternal serum p,p’-DDE concentrations were highly correlated with both cord serum (r=0.68) and meconium (r=0.62) p,p’-DDE concentrations.
Conclusion: Collectively, these data demonstrate pervasive environmental exposures in a representative sample of pregnant women. The high degree of correlation among the biological
matrices for these chemicals suggest that maternal samples collected throughout pregnancy may serve as good surrogate measures of fetal exposure. The ultimate validation of these
biomarkers would enhance our ability to predict reproductive and neurobehavioral toxicity.




                                                                                               43
        Oral Communication Session

SAB2-O-06                  DETERMINATION OF PERFLUOROOCTANESULFONATE AND PERFLUOROOCTANOIC
                           ACID IN MATERNAL AND CORD BLOOD SAMPLES
O MIDASCH1, J ANGERER1
(1) Institute and Outpatient Clinic of Occupational, Social and Environmental Medicine, Erlangen.
Introduction
Perfluorooctanesulfonate (PFOS) and Perfluorooctanoic acid (PFOA) are raw materials or used as manufacturing aids for some perfluorinated chemicals and can be released of those by biotic
and/or metabolic decomposition. Due to their ubiquitous occurrence, persistence and bioaccumulative properties they can be found in plasma and serum samples of the general population
all over the world. In animal studies PFOS and PFOA provoked various types of cancer and showed developmental toxic potential besides other adverse health effects. On the basis of the
comparison of plasma sample pairs from mothers and their infants’ cord blood we wanted to show if neonates are exposed to PFOS and PFOA in earliest state of development.
Methods
We determined PFOS and PFOA in 11 plasma samples of mothers and in 11 plasma samples from the cord blood of their infants. An analytical method based on serum protein precipitation
followed by HPLC with MS/MS-detection was employed. As internal standards we used isotopically labelled 1,2,3,4-13C-PFOS and 1,2-13C-PFOA.
Results
We found PFOS and PFOA in every plasma sample analysed. In maternal plasma samples PFOS concentrations were consistently higher compared to those of the related cord plasma samples
(median: 12.7 µg/L vs. 7.2 µg/L). In the case of PFOA we observed only minor differences between PFOA concentrations within the analysed sample pairs (median: 2.4 µg/L vs. 3.1 µg/L for
maternal and cord plasma samples, respectively).
Discussion and Conclusions
For both substances a crossing of the placental barrier could be shown. For PFOS we observed a decrease from maternal to cord plasma concentrations by a factor of 0.41 to 0.80. However,
PFOA obviously crosses the placental barrier unhindered. These findings show that neonates are exposed to PFOS and PFOA via their mothers’ blood. Given the current situation that only
little is known about the consequences of PFOS and PFOA exposure in the early state of development of humans and the fact that in animal studies both substances showed developmental
toxic effects future research regarding human health effects is indispensable.




SAB2-O-07                  LUNG RETAINED ASBESTOS FIBRE ANALYSIS: TRENDS OVER 25 YEARS IN
                           ONE INSTITUTION AND THEIR IMPLICATIONS.
BW CASE1, A DUFRESNE1, P SEBASTIEN1
(1) McGill University, Montreal.
INTRODUCTION: Lung-retained asbestos, as light-microscopic asbestos “bodies” (AB) and as electron microscope ascertained asbestos fibre (AF) analysis, was the first exposure biomarker.
It began in the 1950’s with counts of AB and reached maximum use in the 1990’s with AF used in cohort and case-control studies.
METHODS: A 25 year experience in one university setting, using analyses of AB and AF representing general populations; environmental, domestic, and occupational exposures; “sentinel”
animal data; and exposure assessment in epidemiological cohorts and case-referent studies of mesothelioma, was assessed. Asbestos lung content was by 2006 evaluated in several thousand
individuals on four continents. Methods were identical over time and results were available for all categories of AF length, width, aspect ratio, and chemistry as well as AB quantities. Time
trends were assessed to determine important uses for asbestos-related exposure and disease.
RESULTS: The following trends were identified: Lung AF, and to a lesser extent AB, are declining over time. We find this to be true not only in the general population and to populations of
individuals with the three common asbestos-related diseases (lung cancer, mesothelioma and asbestosis). The latter finding may make the procedure less useful in future determinations of
risk attribution and compensation hearings. The findings also suggest increased clearance of all fiber types. This may explain recent reports of past overestimation of future mesothelioma risk,
especially in the Netherlands, Norway, Sweden, and Australia. AF and AB have consistently outperformed other indices of exposure as indicators of risk and radiological change. This was
especially true in application of our data to workers at a vermiculite mine in Libby, Montana, where we find that even asbestos bodies in sputum predict radiological extent of asbestosis more
accurately than do historical estimates of cumulative exposure derived from work histories. AF and AB analysis have identified environmental risk where it was missed by conventional air
measurements. One example is our identification of high concentrations of tremolite asbestos in animals near natural asbestos deposits in the Western Slope of the Sierra Mountains in
California. In North America, there has been a shift over time from crocidolite to amosite as the principal fiber associated with mesothelioma risk. Tremolite from natural sources has become
increasingly important over all continents.
CONCLUSIONS AND DISCUSSION: The results illustrate both the past and the future of biological markers of exposure. Trends suggest that AF and AB will be less useful for occupational
assessment, due to declining concentrations. On the other hand, the observations suggest greater clearance (including amphibole clearance) than was hitherto appreciated, and fit well the
recent observations of Berry et al. in Wittenoom (Australia) crocidolite miners. The already observed decrease in predicted mesothelioma cases is likely to accelerate as a result. Our findings
are useful for contemporary environmental risk assessment, in that unlike past exposures, these exposures are ongoing and indeed current. The result has been that we can detect these
exposures in at-risk populations either through sentinel animals (as in our observations in California) or directly in autopsy samples of general populations.




SAB5-O                     CHILDREN’S HEALTH, AIR POLLUTION AND NOISE
SAB5-O-01                  NOISE EXPOSURE IN TAIWANESE SCHOOL CHILDREN AND DISTURBANCE ON
                           SPEECH COMPREHENSION
YC LIN1, CM CHIANG2, JL WU3, HJ SU1, YL GUO4
(1) Graduate Institute of Environmental and Occupational Health, College of Medicine, National Cheng Kung University, Tainan. (2) Department of Architecture, National
Cheng Kung University,, Tainan. (3) Department of Otolaryngology, National Cheng Kung University Hospital, Tainan. (4) Department of Occupational and Environmental
Medicine, College of Medicine, National Taiwan University, Taipei.
Data is lacking on the quantity of interference in school children’s comprehension of class material. The objective of the study is to quantify the effect of environmental noise on school
children’s speech comprehension. Speech comprehension of sixth-grade children are compared among 12 schools with noise monitoring station in Taiwan. Standardized phonetically balanced
word lists with sound pressure of 75, 80, 85, 90, and 95 dBA are used in the field to test the speech recognition ability of children under the interference of background noise. To create a
signal contour map of classroom, random noise generator SF-06 is used to generate sound level of 75, 80, 85, 90, 95dBA and sound level meter NL-11 is used to measure the sound level at
different positions in an empty classroom. Received sound level by test children are derived from combination of coordinate of seat and noise contour map of classroom. Signal to noise ratio
(S/N ratio) are derived by subtracting classroom noise level from received signal level by test children. Relationship of S/N ratio and score of word recognition test are assessed to evaluate
the interference of comprehension in students. Total of 393 children of 24 classes among 12 test schools were included in the analysis. Linear regression analysis indicates that S/N ratio are
associated with speech comprehension and also provides an equation for S/N ratio to infer the score of speech comprehension-in-noise test (Total score (%)=37.5+2.84 (S/N ratio)-0.043 (S/N
ratio)2). According to the noise monitoring station data of daytime 8 hours average in the past 4 years, over 80% schools had monitored school noise of higher than 60 dBA. By the equation,
when the teaching sound levels are 90, 85dBA and the school noise is higher than 60dBA, >41.4%, and >50.6% test words cannot be comprehended correctly. The study shows that present
school noise causes significant interference with speech comprehension, which should not be ignored. However, the study cannot specify the sources of noise. Further study is therefore
warranted in order to prevent interference on teaching.




                                                                                              44
                                                                                                                                                                                                  SUNDAY SEPTEMBER 3
        Oral Communication Session

SAB5-O-02                  EFFECTS OF SULFUR DIOXIDE ON LUNG GROWTH IN YOUNG CHILDREN
                           RESIDING NEAR LIGNITE COAL POWER PLANT
N VICHIT-VADAKAN1, W AEKPLAKORN2
(1) Thammasat University, Bangkok. (2) Mahidol University, Bangkok.
Effects Of Sulfur Dioxide On Lung Growth in Young Children Residing Near Lignite-Coal Power Plant
Nuntavarn Vichit-Vadakan, Thammasat University; Wichai Aekplakorn, Mahidol University

Introduction: October of 1992 marked the first episode of high ambient concentration of sulfur dioxide (2000 ug/m3 during one hour) in Mae Moh District, Lampang Province, Thailand where
a 2,625 megawat lignite-coal power plant is located. In response to this ill-fated event, the government commissioned, among other mitigation measures, an epidemiological study to
determine the health effects of the Mae Moh residents. One of the objectives of the study was to investigate the effects on SO2 on lung growth on young children.
Methodology: This cohort study, extending from 1995 to 1999, recruited 279 eight year old children in the exposed area and 226 children in the non-exposed area at the start of the study.
Efforts to ensure data quality included intensive training of children to enhance their spirometry performance, daily calibration of spirometers, and review of spirometry measurements by
pulmonologist. Lung functions were measured once annually at the same time period between May and June. Mixed models were used in the data analysis.
Results: The ambient SO2 level decreased consistently from 1995 to 2000 as a result of several control measures and the installation of Flue Gas Desulfurization as shown in Figure 1. The
annual average declined from 28 ug/m3 in 1995 to 3 ug/m3 in 2000.
                                                                                                          Figure 1:
Lung functions measured in FVC, FEV1, and PEFR increased statistically with age in both the exposed and non-
exposed groups and female were higher after controlling for several covariates. In comparing the lung function
growth, FVC and PEFR grew at the same rate, however, FEV1 increased more significantly in the exposed group.
Although the children in all 4 subdistricts in the exposed area showed a comparable lung growth, children in Sob
pad subdistrict located adjacent to the power plant started off with a lower lung function but grew at a faster rate
than the other 3 districts and eventually caught up with the other children.
Discussion: Although further studies are needed, this study showed that the effects of SO2 on lung function
growth may be reversible. Once the ambient level of SO2 started to decline as a result of mitigation measures
instituted by EGAT (Electricity Generating Authority of Thailand), lung functions of children living in the most
affected area showed a faster growth rate and eventually reached the same levels as other children.




SAB5-O-03                  GERES IV: TIME TRENDS IN EXPOSURE OF GERMAN CHILDREN TO PAHS,
                                        ,
                           COTININE, PCP AND HEAVY METALS
K BECKER1, A CONRAD1, N KIRSCH1, M KOLOSSA-GEHRING1, C SCHULZ1, M SEIWERT1, B SEIFERT1
(1) Federal Environmental Agency (Umweltbundesamt), Berlin.
The German Environmental Surveys (GerESs) are representative population studies carried out repeatedly since the mid-1980s. The objectives of GerESs are, inter alia, a) to evaluate the body
burden of pollutants of the general population, b) to investigate the contribution of different compartments (air, water, food) to the body burden and c) to develop models to assess exposure.
Although children had been included to some extent in earlier GerESs, GerES IV (2003-2006) is the first survey focussing solely on children. The study participants were selected at random
from registration offices.
In GerES IV polycyclic aromatic hydrocarbons (PAH), pentachlorophenol (PCP), cotinine, and heavy metals (Cd, Pb, Hg) have been measured in blood and urine. When comparing GerES IV
results to those obtained in GerES II (1990/92) for children aged 6 to 14 years living in the households of adult participants, the following trends can be observed:
• PAH metabolite concentrations in urine decreased remarkably since 1990/92, especially for East-German children. The gap between the levels observed earlier in East-Germans and West-
  Germans has widely been closed. This finding is most likely related to the successful reduction of air pollution in East-Germany after the German reunification.
• The median level of PCP in urine decreased from 4.5 µg/l to < 0.6 µg/l. This is due to the fact that production and application of PCP and PCP containing products was banned in 1989.
• In both surveys around 50 % of the children were living in households with at least one smoker and thus exposed to environmental tobacco smoke. The cotinine levels confirm the ongoing
  relevance of this exposure pathway.
• The concentration of lead in blood decreased continuously since the ban of leaded gasoline. The current lead level of 18 µg/l (32 µg/l in GerES II) is the lowest mean concentration reported
  from German studies on children so far.
• In 1990/92 the cadmium levels in blood and urine were close to the limit of detection (GM: 0.1 µg/l) and remained unchanged.
• Mercury in blood and urine decreased due to a reduced exposure through fish consumption and amalgam fillings. In Germany amalgam fillings are no longer recommended for children.
These results from GerES IV demonstrate the success of measures to reduce human exposure for selected chemicals.
The financial support of the Federal Ministries for the Environment, Nature Conservation and Nuclear Safety and of Education and Research is gratefully acknowledged. Field work during
GerES IV was carried out by the Robert Koch Institute, Berlin.




SAB5-O-04                  CHILDREN’S EXPOSURE TO DIESEL EXHAUST AT A RURAL ELEMENTARY
                           SCHOOL
C SCHREINER1, E GONYO1, K SHANDILYA1, J BUONOCORE1, X ZHOU1, G SALAMERO1, H MOORE1, J BROWN1, A ROSSNER1, A FERRO1
(1) CLARKSON UNIVERSITY, POTSDAM.
Diesel exhaust has been associated with health effects in humans, including respiratory illness and cancer. Children are of particular concern regarding diesel exposure from schools buses
because their lungs are still developing and they can be more susceptible to health effects than adults. In this study a direct exposure assessment was conducted at the elementary school in
Potsdam, NY in late October 2005 for five days and early February 2006 for four days. Sampling was performed for two hours in the morning and two hours in the afternoon corresponding
with the school bus drop off and pick up times. Matched indoor/outdoor semi-continuous instruments were used to measure ultrafine particle (UFP) counts, particle bound PAHs, and particle
size distribution. One indoor and several outdoor integrated filter samplers collected PM2.5 mass on both quartz and Teflon® filters. The filter samples were analyzed for mass and organic and
elemental carbon (OC/EC).
The highest PM2.5 concentrations were found indoors and the outdoor concentrations in the bus pick up and drop off area were found to be slightly higher than those from the upwind site.
All outdoor PM2.5 concentrations were below 12 mg/m3. The elemental carbon measurements were found to be below the detection limit during the fall field intensive, which is likely due to
the fact that the buses shut off immediately upon arrival at the school, reducing the source strength, and the sampling period was short. The sampling period was increased from 4 hours to
8 hours for the winter field intensive. Time series data from the semi-continuous instruments showed clear increases in UFP number concentration to more than 100,000 #/cc during the times
that the buses arrived and departed from the school compared to less than 10,000 particles per cubic centimeter (#/cc) before the buses arrived. A strong correlation between the indoor air
and the outdoor particle concentrations was found, and the penetration of the outdoor particles indoors was found to be almost 100 percent. The results of this study illustrate that the bus
pick up and drop off next to the elementary school results in a short-term, high concentration exposures to UFP both outside and inside the school.




                                                                                             45
         Oral Communication Session

SAB5-O-05                   THE IMPACT OF AIR POLLUTION ON BRONCHIOLITIS
P DEMERS1, C KARR2, M KOEHOORN3, C LENCAR1, L TAMBURIC4, M BRAUER1
(1) School of Occupational and Environmental Hygiene, The University of British Columbia, Vancouver. (2) Department of Environmental and Occupational Health
Sciences and Department of Pediatrics, University of Washington, Seattle. (3) Department of Health Care and Epidemiology, The University of British Columbia,
Vancouver. (4) Centre for Health Services and Policy Research, The University of British Columbia, Vancouver.
Introduction: Bronchiolitis, a major cause of infant morbidity, has been associated with environmental tobacco smoke. Its association with air pollution has been evaluated in several stud-
ies, but the results are mixed. We evaluated the impacts of several air pollutants on bronchiolitis in a nested case-control study using improved exposure methods.
Methods: The study population were singleton children born in a metropolitan area in 1999 and 2000. Birth records were linked to health records to obtain data on all health care system
encounters, as well as residential history and census tract information as part of the Border Air Quality Study. After excluding 723 missing census data and 22 with <25 weeks of gestation
40,288 were available for analysis. Infants were followed from the 2nd to 12th month and 5,678 outpatient and hospitalization cases were identified. Five controls born on the same date were
incidence density matched to each case. Mean exposures to NOX, CO, SO2, PM10, PM2.5 and O3 were estimated using inverse distance-weighted average of all monitors in the study region. NOx
and Black Carbon (filter absorbance) were estimated using temporally-adjusted land use regression (LUR) models. Exposures were estimated for the home address of the children for lifetime,
one month, two months prior to diagnosis/reference date. Conditional logistic regression was used to adjust for infant sex, ethnicity [First Nations Status], and prematurity, as well as neigh-
borhood income and maternal education.
Results: Based on lifetime exposure, the strongest association was observed with O3 (OR=1.38, 95% CI=1.19-1.60 per 10 µg/m3). Smaller, suggestive observations were observed for PM10
(OR=1.02, CI=1.00-1.05, 1 µg/m3), PM2.5 (OR=1.09, CI=0.98-1.22, 1 µg/m3), and black carbon (OR=1.05, CI=0.95-1.16, 0.5x10-5 m-1 increase in filter absorbance). No association was observed
for NOX based on either ambient monitors or LUR. Protective associations were observed for SO2 (OR=0.95, CI=0.92-0.97, 1 µg/m3) and CO (OR=0.93, CI=0.90-0.97, 100 µg/m3). These results
were relatively unchanged when exposure in specific time windows was considered, with the exception of O3 where the effects were somewhat reduced (OR=1.26, CI=1.10-1.44, 1 month
prior and OR=1.28, CI=1.11-1.48, 2 months prior). Seasonal stratification indicated that O3 and PM10 associations were highest in fall (ORs=1.71 and 1.07, respectively) and winter (ORs=1.69
and 1.05, respectively).
Conclusions: This is the first study to observe an association between physician diagnosis and hospital admissions for bronchiolitis and ozone exposure. These findings add some support for
previously reported associations between bronchiolitis and exposure to particulate matter.




SAB5-O-06                   RESIDENTIAL AMBIENT AIR POLLUTION AND LUNG FUNCTION
B OFTEDAL1, B BRUNEKREEF2, W NYSTAD1, SE WALKER3, P NAFSTAD4
(1) NORWEGIAN INSTITUTE OF PUBLIC HEALTH, OSLO. (2) INSTITUTE FOR RISK ASSESSMENT SCIENCES, UTRECHT. (3) NORWEGIAN INSTITUTE FOR AIR RESEARCH,
KJELLER. (4) UNIVERSITY OF OSLO, OSLO.
Introduction
Several studies have reported associations between ambient air pollution and lung function. However, most of these studies used exposure on aggregated level, and not on individual level.
The aim of this study is to study the relation between residential long-term, both in early and late years of life, and short-term air pollution exposures and lung function in 9-10 year old children
living in one city.
Methods
Air pollution exposure was assessed by a dispersion model based on emissions, meteorology, topography and background air pollution concentrations. The model calculates hourly
concentrations of NO2, PM10 and PM2.5 for each km2 and at thousands of receptor points. Early and late long-term exposure was calculated as the average concentration in first year of life and
in the calendar year preceding the lung function examination, respectively. Short-term exposure was the average concentration of lag1-lag3 relative to the examination date. Pulmonary
function outcomes (PEF, FEF25%, FEF50%, FEV1 and FVC) were measured in 2679 children, and only children with valid outcomes and valid address in both early and late years were included in
the analyses (n=1914). We used multiple linear regression stratified by sex, to study the relation between residential air pollution and lung function, and adjusted for height, body mass index,
age, birth weight, any asthma confirmed medically including symptoms last 12 months, average temperature of lag1-lag3, parents’ education, smoking habits and ethnicity.
Results
The long-term residential concentration in first year ranged from 1.5 to 84.0 •g/m3, from 5.1 to 53.9 •g/m3, and from 5.1 to 26.9 •g/m3, for NO2, PM10, and PM2.5 respectively. The late
concentrations were lower. The pollutants were highly correlated (correlation coefficients r=0.77-0.95). Results showed that all the pollutants were associated with PEF, especially in girls;
10 •g/m3 increase of early NO2, PM10 and PM2.5 were associated with 38.3 ml/s (-65.1, -11.5) (95% CI), 83.0 ml/s (-153.3, -12.7) and 125.9 ml/s (-232.9, -18.9) decline in PEF respectively. We
found similar associations with FEF25% and FEF50%, but no statistically significant long-term effects on FEV1 and FVC. For late long-term exposure the associations were generally higher, especially
for particulate matter. The short-term effects were lower than the long-term effects, and higher in girls than boys.
Discussion and Conclusions
This study investigated both long- and short-term effects of ambient air pollution at residential addresses, and moderate air pollution levels may be associated with lung function in 9-10 year
old children, particularly girls.




SAB5-O-07                   INDICATORS OF CHILDREN’S HEALTH AND THE ENVIRONMENT IN NORTH
                            AMERICA
E PHIPPS1, A BÉRUBÉ2, T WOODRUFF3, M RAMIREZ4
(1) North American Commission for Environmental Cooperation, Montreal. (2) Health Canada, Ottawa. (3) U.S. Environmental Protection Agency, San Fransisco. (4)
Cofepris, Secretaria de Salud, Mexico.
Introduction
The Commission for Environmental Cooperation (CEC), in partnership with public health organizations and the governments of Canada, Mexico and the United States, released in January
2006 the first-ever.
Materials and Methods
The CEC, the International Joint Commission, the Pan American Health Organization, the World Health Organization (WHO), and the governments of Canada, Mexico and the United States
collaborated in the development and selection of the children’s environmental health indicators and the release of this report. The presentation will describe the collaborative process followed
to select appropriate indicators of children’s environmental health in North America, data gathering and harmonization as well as the main findings of the report.
Results
The report presents 13 indicators under three thematic areas: asthma and respiratory disease, effects of exposure to lead and other toxic substances, and waterborne diseases. It finds that
North American children remain at risk from environmental exposures and that children’s health reporting must be improved to address the data gaps identified in the report. Only one of the
indicators, addressing asthma in children, was fully reported by all three countries. These data show a rising number of childhood asthma cases across North America. One possible contributor
is outdoor air pollution such as ground-level ozone and particulate matter, which remains a problem for all three countries. In Mexico, exposure to smoke from indoor burning of wood or
charcoal is also a problem, as 18 percent of the country’s population continued to burn biomass for cooking and heating in 2000. For lead exposure, case studies from all three countries
demonstrate improvements in children’s blood lead levels due to interventions such as the removal of lead from gasoline. However, there is little biomonitoring data available in Canada since
there has been no national blood level survey in the country since 1978. Other exposure pathways for lead remain a concern, such as older homes with lead-based paint. Recently collected
data in the United States show that 25 percent of homes have a “significant lead-based paint hazard, which could be from deteriorating paint, contaminated dust or contaminated soil outside
the house.” Finally, information is also presented on indicators of water quality and sanitation.
Conclusion
Based on the findings of the report, recommendations for health surveillance, environmental monitoring and research on the health effects of children’s exposure to environmental hazards in
North America are discussed.



                                                                                                 46
                                                                                                                                                                                                         SUNDAY SEPTEMBER 3
         Oral Communication Session

SAB5-O-08                   EXPOSURE TO TRAFFIC RELATED AIR POLLUTION IN EARLY LIFE AFFECTS
                            LUNG FUNCTION, AIRWAY DISEASE AND SENSITIZATION IN 4 YEAR OLD
                            CHILDREN
E NORDLING1, N BERGLIND1, E MELÈN1, G EMENIUS1, J HALLBERG2, F NYBERG3, G PERSHAGEN1, M SVARTENGREN2, M WICKMAN1, T BELLANDER1
(1) Dept of Occup and Environ Health, Stockholm County Council & Institute of Environ Medicine, Karolinska Institutet, Stockholm. (2) Dept of Occup and Environ
Health, Stockholm County Council & Department of Public Health Sciences, Karolinska Institutet, Stockholm. (3) Institute of Environ Medicine, Karolinska Institutet &
AstraZeneca R&D Mölndal, Stockholm & Mölndal.
Introduction: Urban air pollution can trigger asthma symptoms but there is conflicting evidence on effects of long-term exposure on lung function, as well as on the onset of airway disease
and allergy in preschool children. The aim of this study was to assess the impact of exposure to air pollution from road traffic and from house heating early in life, on lung function, wheezing
and sensitization in children up to the age of four.
Methods: The spatial distribution of air pollution from traffic was assessed using source-specific nitrogen oxides (traffic-NOx) and inhalable particulate matter (traffic-PM10) as indicators, based
on emission databases and dispersion modelling in a 25 to 500 meter gridnet. Air pollution from residential heating was assessed as sulphur dioxide (heating-SO2). Estimated levels were used
to assign monthly residential outdoor levels for the first year of life of 4,081 children in a prospective birth cohort, by linking to geocoded home addresses. The average source-specific
contribution from road traffic to residential NOx and PM10 levels were 23.2 and 3.9 µg/m?, respectively, during the children’s first year of life. Parents provided questionnaire data on symptoms
and exposures when the children were approximately 2 months, 1, 2 and 4 years old. At 4 years, 73% of the children participated in clinical examination including specific IgE in blood and
peak expiratory flow (PEF).
Results: Higher residential levels of air pollution from traffic during the first year of life were associated with lower lung function at four years of age: 5.4 l/min lower PEF (95% CI: 0.05, 10.7)
for a 6 µg/m? (5th-95th percentile) difference in residential traffic-PM10.These levels tended to be associated with persistent wheezing, odds ratio (OR) 1.64 (95% CI: 0.90-3.00), and with
sensitisation to inhalant allergens, especially to pollen, OR 2.30 (95% CI: 1.23-4.29). Results were similar using traffic-NOx as indicator. No associations to heating-SO2 (average source-specific
contribution 2.5 µg/m?) were seen.
Conclusions: In a Swedish birth cohort we found associations between residential outdoor levels of air pollution from traffic during the first year of life, and three different indicators of airway
disease in 4-year old children: lower peak expiratory flow, persistent wheezing and sensitisation to pollen. This indicates that modest exposure to air pollution from traffic early in life may
have a negative effect on the respiratory health of pre-school children.




                                                                                                 47
        Symposia Session:

SM1-PD                      PESTICIDES AND PERSISTENT ORGANIC POLLUTANTS
SM1-PD-01                   MULTIPLE PESTICIDE EXPOSURE AND CHROMOSOME ABERRATIONS AMONG
                            ORANGE FARMERS FROM NORTHERN THAILAND
T PRAPAMONTOL1, U MEVATEE2, A MANGKLABRUKS2, N UDOMVONG3, T KAEWTHUMMANUKUL3, S SRIBOONRUANG4, S NIMSAKUL1, C KINGKEOW1,
S HONGSIBSONG1
(1) RESEARCH INSTITUTE FOR HEALTH SCIENCES, CHIANG MAI UNIVERSITY, CHIANG MAI. (2) FACULTY OF MEDICINE, CHIANG MAI UNIVERSITY, CHIANG MAI. (3)
FACULTY OF NURSING, CHIANG MAI UNIVERSITY, CHIANG MAI. (4) SCHOOL OF ECONOMICS, CHIANG MAI UNIVERSITY, CHIANG MAI.
Introduction: Farmers are prone to multiple pesticide exposure especially in an intensive agriculture. In northern Thailand, growing of mandarin oranges has been increasing rapidly may be
due to suitable geographical condition. We conducted a prospective study in orange farmers on pesticide exposure and its effect from Chiang Mai Province in 2004-5.
Methods: Fifty family farmers (F1 and F2) and two consumer groups, as referent groups, were enrolled. One consumer group was from the same districts of farmers (C1, n= 68) and the other
was from Chiang Mai City (C2, n= 54). Farmers and consumers were followed for 3 and 2 visits (V1, V2 and V3) in approx. 6 mo. interval, respectively. Demographic data were collected using
structured questionnaires. Venous blood samples (~10 ml) were collected for acetylcholinesterase (AChE) in red cells, organochlorine pesticides (OCPs) in plasma, and chromosome aberrations
(CAs) in lymphocytes. AChE activity was assayed using Ellman-based method, OCPs using a Prapamontol and Stevenson method and CAs in lymphocytes using a WHO standard method.
Results: Farmers (F1 males and F2 females) aged from 27-60 years old and consumers (C1, and C2, both sexes) aged 24-62 years old. Inhibitions of F1 and F2 AChE activities in V2 versus V1
were 37.1 and 32.9% (both, P<0.001) and in V3 versus V1 were 38.6 and 36.2 % (both, P< 0.001), respectively. Inhibitions of AChE activities of C1 and C2 in V2 versus V1 were 26.3 % (P<
0.001) and 7.13% (NS), respectively. Two major OCP components detected were p,p’-DDE (100%) and p,p’-DDT(98-100%) except in C2 was only 1.85%. Geometric means of p,p’-DDE in F1,
F2, C1, and C2 were 21.0, 21.9, 13.7, and 7.78 ng/ml and of p,p’-DDT were 1.82, 1.95, 0.84, and 0.26 ng/ml, respectively. CAs was performed in 2 visits. Percentage of CAs in F1, V1 versus
V3 were 1.42±1.26 versus 2.00±1.26 (P=0.03) and in F2, V1 versus V3 were 1.22±1.43 versus 1.98±1.87 (P=0.03). Percentage of CAs in C1, V1 versus V2 were 1.50±1.32 versus 2.41±1.85
(P=0.001) and in C2, V1 versus V2 were 1.35±1.25 versus 1.30±1.28 (NS).
Discussion and conclusions: AChE activities of F1, F2 and C1 showed significantly and consistently inhibition pattern. Levels of p,p’-DDE and p,p’-DDT in farmers were significantly higher
than in consumers. Farmers and C1 therefore could be exposed to cholinesterase-inhibiting pesticides and OCPs. Percentage of CAs in farmers and C1 were significantly increased but not in
C2. This may be the cytogenetic effect consequences from multiple pesticide exposure though confounding factors such as age, sex, education and chronic ill health should not be ignored.




SM1-PD-02                   ASSOCIATION BETWEEN ENVIRONMENTAL EXPOSURE TO PCBS AND HEARING
                            IMPAIRMENT IN 8-9 YEARS OLD CHILDREN
T TRNOVEC1, E SOVCIKOVA1, M HUSTAK2, S WIMMEROVA1, A KOCAN1, J PETRIK1
(1) SLOVAK MEDICAL UNIVERSITY, BRATISLAVA. (2) MILITARY AVIATION HOSPITAL, KOSICE.
Introduction: A chemical plant produced PCBs between 1959 and 1984. Disposal of effluent into a river resulted in contamination of a site sized about 70 by 30 km. It accommodates
approximately 250000 inhabitants and is recognized as one of the most heavily polluted areas with PCBs in the world. The data for sum of PCB serum concentrations for 433 8-9 years-old-
children living in this area in 2002 were: mean 528.2 ng/g of serum lipids, median 321 ng/g of serum lipids and limits of the quartiles-18, 175, 321, 618, 6477 all in ng/g of serum lipids. It
was reported that auditory thresholds for 1-kHz tones were elevated by approximately 25 dB in animals exposed to PCB (Goldey ES et al. Toxicol Appl Pharmacol 1995 135:77). They were
linked to a loss of outer hair cells (Crofton KM et al. Hear Res 2000 144:196). Therefore hearing examination was included into our study program.
Methods: The concentrations of selected PCB congeners in serum of 8-9 years-old-children were determined using GC/ECD. Hearing examination involved pure tone audiometry (Madsen
Orbiter 922), transient evoked otoacoustic emissions (Otodynamics ILO 88), auditory brainstem response (Otometrics BERA System) and tympanometry (Siemens SD-30 tympanometer).
Results: The main findings:
• An association between negative middle ear pressure and PCB serum concentrations.
• Increase of the hearing thresholds at low frequencies associated with higher PCB serum concentrations.
• No association between the serum PCB concentrations and any of the auditory brainstem response parameters.
• Higher serum PCB concentrations associated with decreased emission levels of the transient evoked otoacoustic emissions.
• The sound pressure level of otoacoustic emissions was associated with results of several neurobehavioral tests applied in the same cohort of children and with the proportion of the
   permanent teeth with demarcated and diffuse opacities or hypoplasia.
Discussion and Conclusions: The data of Longnecker MP et al. (Neurotoxicol Teratol. 2004, 26:629) did not support adverse effects of PCBs during early-life on sensorineural hearing loss.
The exposure levels were twofold higher than recent background levels in the United States. In contrast to the previous data at our much higher exposure levels, subclinical but significant
increase of hearing thresholds and a decrease of sound pressure level of transient evoked otoacoustic emissions were associated wth PCBs serum concentrations. A follow-up study on the
same cohort will bring additional information on related time trends.




SM1-PD-03                   SERUM CONCENTRATIONS OF ORGANOCHLORINE COMPOUNDS AND
                            MAMMOGRAPHIC DENSITY IN A HIGHLY EXPOSED POPULATION IN TRIANA,
                            ALABAMA
J RUSIECKI1, J CASH2, C RAINES2, L BRINTON3, S ZAHM3, T MASON4, L NEEDHAM5, A BLAIR3, S SIEBER3, R HOOVER3
(1) Uniformed Services University of the Health Sciences, Dept of Preventive Medicine, Bethesda. (2) University of Alabama, College of Nursing, Huntsville. (3) National
Cancer Institute, Division of Cancer Epidemiology and Genetics, Bethesda. (4) University of Southern Florida, College of Public Health, Tampa. (5) Centers for Disease
Control, National Center for Environmental Health, Atlanta.
Introduction: Between the late 1940’s and early ‘70’s, a now defunct chemical factory discharged approximately 400 tons of DDT and associated waste into the Tennessee River upstream
from Triana, Alabama, a rural, predominantly African-American community of about 600 people. Residents were exposed to DDT and PCBs primarily via ingestion of river fish. A 1979 CDC
investigation found average DDT levels nearly 10 times higher than the U.S. population. The NCI and University of Alabama Huntsville College of Nursing, with the collaboration of the Triana
Area Medical Fund (TAMF), have investigated the association between mammographic density (a major predictor of breast cancer risk) and serum concentrations of various organochlorines,
many of which have estrogenic properties.
Methods: Women who were part of the TAMF were offered mammograms and a standard blood chemistry analysis. Breast density was determined for 187 women, along with analytic
measurements (ng/g lipid-adjusted) of the organochlorine pesticides DDT, DDE, •-HCCH, heptachlor epoxide, oxychlordane, trans-nonachlor, mirex, and PCB congeners 74, 99, 118, 153, 170,
180, 183, 187, 194, and 206. For both the left and right breasts, we measured percent of total and average absolute breast density. Linear regression analyses for each chemical and breast
density, using both continuous actual and log transformed values, were conducted, adjusting for age, body mass, parity, smoking and use of hormone replacement therapy. Analyses were also
stratified by three age groups (19-40, 41-54, 55+).
Results: The average age of study women was 49 (range: 19-91), and 80% were African-American. Most organochlorine values were not significantly related with mammographic densities,
although there were significant inverse associations between log transformed average percent density and several organochlorine measures. Among all subjects, we observed significant
negative correlations for •-HCCH (•-coefficient=-0.31, p=0.01) and pp’DDT (-0.31, p=0.05); •-HCCH was also significantly correlated among the 41-54 year old (•-coefficient=-0.54, p=0.02) subjects.
Significant relationships were also observed for PCB-183 among the 41-54 year olds (-1.03, p=0.01) and DDE among 55+ year olds (-0.29, p=0.02). The only significant finding for absolute
breast density was for oxychlordane among the 55+ year old subjects -0.54, p=0.05).
Conclusions: This is the first study to assess relationships between organochlorines and mammographic patterns. We saw no evidence for increased risk of mammographic densities with
organochlorine concentrations. In fact there was fairly consistent evidence that some of these estrogenic organochlorines were inversely associated with percent density, a finding contradictory
to the notion of estrogenic activity increasing mammographic density.


                                                                                                48
                                                                                                                                                                                                      SUNDAY SEPTEMBER 3
        Symposia Session:

SM1-PD-04                  EVIDENCE OF TCDD EXPOSURE 17 YEARS AFTER CESSATION OF PRODUCTION
                           OF THE HERBICIDE 2,4,5-TRICHLOROPHENOXYACETIC ACID (2,4,5-T) IN A
                           NEW ZEALAND COMMUNITY
J FOWLES1, C STEVENSON2, M NOONAN2, L GALLAGHER1, V BAKER1, D READ3, D PHILLIPS1
(1) Institute of Environmental Science and Research, Wellington. (2) Air and Environmental Sciences, Auckland. (3) Ministry of Health, Wellington.
Introduction. The exposure of residents of Paritutu, New Zealand, to 2,3,7,8-TCDD emissions from a local source of 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) manufacture, from 1962-1987,
was investigated using serum analyses and exposure modelling.
Methods. Plume dispersion, multipathway exposure, and toxicokinetic models were used to predict individuals with the highest TCDD serum levels in 2004 from a self-nominated population
of 379 former and current residents. A spatial analysis based on 74 soil samples predicted TCDD soil concentrations in the area, from which historical air concentrations and exposures were
estimated. Twenty-two men and 30 women were selected for serum testing based on soil TCDD concentrations, age, sex, years of residence, and contributions to exposure from home grown
produce. Participants resided for at least one year at an address within approximately 2 km of the 2,4,5-T plant during the production period, and indicated they had not been in an occupation
involving agrichemical use or exposure. Serum was tested for chlorinated dibenzo-p-dioxins, chlorinated dibenzofurans, and dioxin-like PCBs.
Results. There was significant elevation of TCDD in the serum of the subjects, when compared to national background concentrations. No other dioxin congeners were elevated. Increases in
TCDD levels were 3.8-fold (arithmetic mean), or 2.7-fold (geometric mean). The highest individual value was 33.3 ppt (elevated 15-fold). Those residing in the area for at least 15 years showed
the highest values, as a group, with an increase in TCDD of 14.6-fold (arithmetic mean), or 12.4-fold (geometric mean). Inhalation was the main route of exposure. Home-grown produce was
weakly correlated with an increase in exposure. The TEQ elevation in the study overall was 1.2-fold compared to 1997 levels, and 2.0-fold compared to estimated 2004 levels.
Conclusions. The residents living in Paritutu, between 1962 and 1987, were exposed to aerial emissions of TCDD. The signature of this molecule in the serum was evident in 2004, 17 years
after aerial exposures would have ceased with the end of 2,4,5-T production in 1987. It could be inferred from the lower serum values in the individuals living in Paritutu only during more
recent times (i.e. post 1980) that there is no significant ongoing contribution to TCDD exposure in the community. The elevations of serum TCDD seen in 2004, would have corresponded to
larger increases in 1987 and earlier with estimated peak serum levels of TCDD in long term residents between 40 and 80 ppt (geometric mean) above background (assuming a 10-year half-
life).




SM1-PD-05                  IMPACT OF FOLPET ON RESPIRATORY SYSTEM : IN VITRO STUDIES
M CANAL-RAFFIN1, B L’AZOU1, B MARTINEZ2, E SELLIER3, F FAWAZ4, C OHAYON1, I BALDI1, J CAMBAR1, M MOLIMARD2, N MOORE2, P BROCHARD1
(1) Santé Travail Environnement EA 3672 Université Bordeaux2, Bordeaux. (2) Departement of Pharmacology, INSERM U657, Université Bordeaux2, Bordeaux. (3) Centre
de Ressources en Microscopie Electronique et Microanalyse Centre (CREMEM) Université Bordeaux1, Bordeaux. (4) Bases thérapeutiques des inflammations et
infections EA 3677, Université Bordeaux2, Bordeaux.
Introduction: Folpet is a fungicide of contact, widely used since 50 years in vineyards, slightly soluble in water and mostly present as particles in environment. Since few years, pesticides
have been found in rural and urban air in France. Folpet was the main of these pollutants, measured at concentrations between 4 and 30 ng/m3, which might be particularly damaging to
vulnerable persons (pregnant women, children, the elderly). However, no study on its penetration by inhalation and on its respiratory toxicity is currently available. The objective of this work
was to determine if folpet particles can be inhaled by humans and affect cells of the respiratory tract.
Methods: Three products were tested: folpet and two of its commercial forms (Folpan 80WG® and Myco 500®). Preliminary studies of granulometry and morphometry characteristics of folpet
particles were conducted with scanning electron microscopy and laser coulter. Their stability were studied by chromatography and cytotoxicity tests were conducted on human bronchial
epithelial cells (16-HBE). Cells were exposed to folpet, Folpan 80WG®, Myco 500® and degradation products (phthalimide and phthalamique acid) at different concentrations (0.1-100 µM) and
for different times.
Results: Granulometric study showed that majority of the particles (75%) had a size under 5mm; so they could be easily inhaled by humans. These particles were relatively stable in time :
75% of folpet remained in the particle suspension after 19 days, under the usual conditions of use. In the cells toxicity study, the inhibitory concentration (IC50) was between 10 and 20 µM
for folpet and commercial products after 24h of exposure. The cytotoxic effect occured rapidly after 1h of exposure. Degradation products of folpet, did not show any toxicity at tested
concentrations (0.1-100 µM).
Discussion-Conclusion: These results show that the folpet particles present in the air could be easily inhaled by humans and can damage respiratory cells in vitro. These data also constitute
the first step towards the determination of a dose-effect correlation and a risk assessment for human health. This work confirms the need for further studies on the effect of environmental
pesticides on respiratory system.




SM1-PD-06                  MULTIVARIATE ANALYSIS OF CORRELATES OF PESTICIDE LEVELS IN HOUSE
                           DUST FROM AN AREA WITH INTENSE AGRICULTURE
M HARNLY1, A BRADMAN2, M NISHIOKA3, R MCLAUGHLIN1, E SILVER4, D SMITH1, B ESKENAZI2
(1) California Dept Health, Richmond, CA. (2) UC Berkeley, School of Public Health, Berkeley, CA. (3) Battelle Memorial Institute, Cincinatti, Ohio. (4) Impact Assessment
Inc, Richmond, CA.
INTRODUCTION: House dust can serve as a reservoir for pesticides and is a potential contributor to children’s pesticide exposures. Greater clarification of the sources of pesticides in dust is
needed.
METHODS: In an area with year-round production of high-value fruit and vegetable crops, we collected, as part of a health study of mothers and their children, house dust samples using a
high volume surface sampler, longitude and latitude coordinates (GPS), and questionnaire responses from 241 homes. In addition, growers in the area are legally required to report agricultural
pesticide use and we ascertained nearby use (within 2.5 km) of participant’s homes. Dust samples were analyzed for 18 pesticides by gas chromatography /mass spectroscopy. Quantile
regression was used to determine associations between dust concentrations and independent variables including nearby agricultural pesticide use, weather parameters, and household-specific
variables, i.e., farmworkers living and wearing work clothes/shoes in the house, pesticide use in the homes, potential soil track-in measures (e.g., housing density, children living in the home,
pets/animals in the home), and other covariates, (e.g., living in poverty). For multivariate models, variables were entered stepwise and eliminated from further models if the proportional
reduction in absolute deviations (i.e., the pseudo-R-square) did not improve.
RESULTS: Diazinon and chlorpyrifos (organophosphate insecticides), chlorthal-dimethyl (a chlorinated phthalate herbicide) and permethrin were most frequently detected (86% or more of
samples, detection limits 1-5 ng/g ) and were selected for further statistical analysis. Initial univariate regression analysis demonstrated significant (p<0.01) associations of chorthal-dimethyl
concentrations with independent variables. Specifically, farmworkers wearing work shoes/clothes in the home, dirt entrance to the home, and living in poverty, were significantly associated
with chlorthal-dimethyl dust concentrations. Further, agricultural pesticide use on the days 1-14, 15-30, and 31-60 prior to dust collection were significantly associated with chlorthal-dimethyl
concentrations; pesticide use was associated with a 114, 62, and 44% increase in median dust concentrations for every 100 pounds used near the homes in these respective periods. In a
multivariate model of chlorthal-dimethyl concentrations, with agricultural pesticide use and corresponding weather parameters (mean daily temperature and rainfall for the respective period)
in the model, the pseudo R-square was substantially improved over the univariate models with pesticide use alone (i.e., 14% compared to 9%). For each pesticide, final multivariate models
on the median, and the 90th percentile, of dust concentrations will be presented.
CONCLUSIONS: Our results suggest that nearby agricultural pesticide use and household-specific variables are impacting children’s exposure to pesticides.




                                                                                               49
        Symposia Session:

SM1-PD-08                  PILOT STUDY OF THE POTENTIAL FOR HUMAN EXPOSURES TO PET BORNE
                           DIAZINON RESIDUES FOLLOWING LAWN APPLICATIONS
M MORGAN1, D STOUT1, D BARR2
(1) U.S. ENVIRONMENTAL PROTECTION AGENCY, RESEARCH TRIANGLE PARK. (2) NATIONAL CENTER FOR ENVIRONMENTAL HEALTH, CENTERS FOR DISEASE CONTROL
AND PREVENTION, ATLANTA.
This study examined the potential for indoor/outdoor pet dogs to be an important pathway for transporting diazinon residues into homes and onto occupants following residential lawn
applications. The primary objective was to investigate the potential exposures of children and their pet dogs to diazinon after a lawn application at their homes. Six families and their pet dogs
were recruited into the study. Monitoring was conducted at pre-, 1, 2, 4, and 8 days post-application of a commercial formulation of diazinon to the lawn by the homeowner. Environmental
samples consisted of indoor air, soil, dust, and transferable residues from lawns and floors. Urine samples were collected from both a child (< 13 years) and his/her parent. Samples taken from
dogs included paw wipes, fur clippings, and transferable residues using cotton gloves. Transferable residues were collected from dogs by a technician wearing a cotton glove and stroking the
dog five times (head-to-rump) each sampling day. Diazinon was analyzed in all environmental samples by GC/MS, and the metabolite, 2-isopropyl-6-methyl-pyrimidin-4-ol (IMPy), was analyzed
in urine by HPLC/MS-MS. We report here the levels of diazinon measured in the paw wipes, fur clipping, and transferable residue samples and urinary IMPy concentrations for children. The
mean diazinon loadings in paw wipe samples were the lowest at pre-application (0.06 ± 0.04 ng/cm2) and highest at Day 1 post-application (88.1± 100.1 ng/cm2). The mean loadings of
diazinon in fur clippings were at least 15 times higher at each sampling day post-application than at pre-application (0.8 ± 0.4 ng/cm2). Post-application diazinon loadings in glove samples
(technician) ranged from 0.01 to 59.2 ng/cm2 during the four sampling days. The children’s mean urinary IMPy concentrations were similar at pre- (1.5 ± 2.0 ng/mL) and post-application (1.3
± 1.5 ng/mL) of diazinon, indicating they did not have substantial increases in absorbed doses to this insecticide after these events. The results showed that children and their pet dogs were
exposed to diazinon residues after residential lawn applications. In addition, the dogs were a good vehicle for the uptake of diazinon residues onto the fur and paws and likely tracked the
residue into the homes and onto occupants through intimate contacts (i.e., petting).
Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy.




SM1-PD-09                  THE MOVEMENT OF DIAZINON RESIDUES INTO HOMES FOLLOWING
                           APPLICATIONS OF A GRANULAR FORMULATION TO RESIDENTIAL LAWNS.
D STOUT1, M MORGAN1, P EGEGHY1
(1) US EPA, RTP, NC.
A pilot study was conducted to examine the movement of diazinon, an organophosphate insecticide, following applications of a granular formulation to residential lawns. The objectives
included evaluating methods for collecting environmental samples and examining the transport and fate of diazinon from a source to the indoor living areas of homes. Although not discussed
here, the study also emphasized the role of pet dogs as vehicles for the translocation of diazinon residues and potential human exposures. Measures included the formulation, soil, particles
from doormats, transferable residues from residential turf and indoor flooring, indoor air from living rooms and children’s bedrooms, and vacuum dislodgeable dust. Samples were collected
from six single family homes located in the mid-Atlantic region of the USA in the summer of 2001. Environmental samples were collected prior to the application by the homeowner and at
days 1, 2, 4, and 8 following the application. Soil concentrations, an indicator of source strength, were highest immediately following the application and declined by an average of 51% by
day 8. Polyurethane foam (PUF) roller samples collected from turf ranged from 0.1 to 970 ng/cm2 over the study period. Particle-associated residues collected from doormats located at
entryways into the home declined from day 1 to day 8 by an average of 75% demonstrating the movement of residues from the source towards the indoors of the homes. Indoor air
concentrations reached maximal levels from 1 to 2 days following the application and declined over the remainder of the study. Indoor PUF roller measures showed little to no increase above
background. Concentrations in vacuum dislodgeable dust were variable over time, but consistently above background concentrations. Results demonstrate that both the physical translocation
of particle-bound residues and the intrusion of volatilized diazinon contribute to indoor levels. Physical processes such as pet activity and foot traffic moved particle-bound residues to the
entryway of the homes. Increased airborne concentrations demonstrate the intrusion of diazinon from the outdoor source. Elevated diazinon concentrations measured from vacuum
dislodgeable dust suggests the movement and deposition of volatilized and/or particle-bound residues. In summary, applications to residential lawns resulted in an increase of diazinon above
background concentrations in all homes. Lawn applications are a source for potential occupant exposures both on treated lawns and inside homes.
Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy.




SM1-PD-10                  BIOMARKERS OF PESTICIDE EXPOSURE: ADULT AND ADOLESCENT
                           FARMWORKER COMPARISONS
L MCCAULEY1, M LASAREV2, J MUNIZ1, YR LIN1, K ANGER2
(1) University of Pennsylvania, Philadelphia. (2) Oregon Health & Science University, Portland.
Agricultural workers may be exposed to pesticide spray, drift, and residues on plants ; however little evidence is available on acceptable levels of pesticide exposure in this population. It has
been suggest that developmental factors- physical, cognitive, and psychological- may place youth workers at increased risk for occupational illness and injury, including the potential for
pesticide exposure. We conducted an investigation of 133 hispanic farmworkers and 56 appropriate controls and assessed whether short-term biomarkers of exposure to pesticides (including
insecticides and fungicides) differed in adolescents employed in agriculture compared to adult farmworkers, controlling for type of agricultural work, hours worked, and reported hygiene
practices. All agricultural workers were employed in crop harvesting and were recruited from agricultural temporary housing and from evening and school year ESL programs. The mean age
of the adult workers was 29 years with a mean of 9.3 years of agricultural work and average age of the adolescent workers was 15.3 years with an mean of 3.0 years of agricultural work.
Surveys were administered on work histories and work practices and included questions on type of work activity, pesticide application, type of crop(s), work hours per week, use of protective
clothing, bathing, and laundry practices. A spot urine sample was collected for the measurement of pesticide urinary biomarkers. Urinary samples were analyzed for alkylphosphate metabolites
from organophosphates and for the metabolite of Captan (tetrahydrophthalimide (THPI), a fungicide frequently used on berry crops. All pesticide metabolite levels were adjusted for urinary
creatinine concentration. Levels of THPI were were shifted significantly higher in the total agricultural when compared to controls. We found that median levels of the major OP metabolite
(DMTP) among teen farmworkers was slightly higher compared to adult workers, though not statistically higher (3.2 ng/mL vs. 1.1 ng/mL). A similar, but non-significant trend was observed
for THPI (0.073 µg/mL vs 0.025 µg/mL). Our previous studies have found that adolescent farmworkers are less knowledgeable about pesticide safety, but protective practices did not differ
significantly from those of adults. These results indicate that exposure levels among these two populations do not differ and while these results are reassuring, they can not be generalized to
all crops in which adolescents work. Future analyses will examine differences in markers of potential health effects in adolescent farmworkers compared to adults with comparable levels of
pesticide exposure.




                                                                                               50
                                                                                                                                                                                                     SUNDAY SEPTEMBER 3
        Symposia Session:

SM1-PD-11                  PESTICIDE DOSE ESTIMATES FOR CHILDREN OF IOWA FARMERS AND
                           NON FARMERS
B CURWIN1, H MISTY1, W SANDERSON2, C STRILEY1, D HEEDERIK3, H KROMHOUT3, S REYNOLDS4, M ALAVANJA5
(1) National Institute for Occupational Safety and Health, Cincinnati. (2) University of Iowa, Iowa City. (3) Utrecht University, Utrecht. (4) Colorado State University, Fort
Collins. (5) National Cancer Institute, Rockville.
Introduction
Farm children have the potential to be exposed to pesticides. Biological monitoring is often employed to assess this exposure; however, the significance of the exposure is uncertain unless
doses are estimated.
Methods
In the spring and summer of 2001, 117 children (66 farm, 51 non-farm) of Iowa farm and non-farm households were recruited to participate in a study investigating potential take-home
pesticide exposure. Each child provided an evening and morning urine sample at two visits spaced approximately one month apart. Estimated doses were calculated for atrazine, metolachlor,
chlorpyrifos and glyphosate from urinary concentrations derived from the spot urine samples and compared to EPA reference doses.
Results
For all pesticides except glyphosate, the doses from farm children were higher than doses from the non-farm children. The difference was statistically significant for atrazine (p-value = 0.0001)
but only marginally significant for chlorpyrifos and metolachlor (p-value = 0.07 and p-value = 0.1 respectively). The highest doses for atrazine, chlorpyrifos, metolachlor and glyphosate were
0.085, 1.96, 3.16, and 0.34 µg/kg/day respectively. None of the doses exceeded the EPA chronic reference values for atrazine, metolachlor and glyphosate; however, all of the doses for
chlorpyrifos exceeded the EPA chronic population adjusted reference value. Doses were similar for male and female children. A trend of decreasing dose with increasing age was observed for
chlorpyrifos.
Discussion and Conclusion
There are several limitations in our pesticide dose estimates that are common to the estimation of doses from spot urinary concentrations. First it was assumed that the spot urine samples
were representative of average daily pesticide excretion and that the doses estimate average daily doses. Second, the merits of creatinine adjustment for spot urine samples are being debated,
especially in children. Lastly, it was assumed that the amount of pesticide metabolite excreted in urine, after adjusting for the fraction of metabolite excreted, was equivalent to an absorbed
pesticide dose. Estimation of pesticide dose from farm children’s urine samples allows comparison to EPA reference doses and therefore provides an indication of the significance of pesticide
exposure. Additional longitudinal studies which better estimate daily pesticide doses over the course of a year are needed to truly determine the health significance of pesticide exposures.




SM1-PD-12                  POTENTIAL OCCUPATIONAL AND NON-OCCUPATIONAL EXPOSURES TO
                           PESTICIDES AMONGST PREGNANT WOMEN IN NORTHERN KWAZULU NATAL
                           PROVINCE OF SOUTH AFRICA
S NAIDOO1, L LONDON2, R CLAPP3, PA JANULEWICZ3, S MNYAIZA1, V VIEIRA3, R WHITE3
(1) Centre for Occupational and Environmental Health, University of Kwazulu Natal, Durban. (2) Occupational and Environmental Health Research Unit, University of
Cape Town, Cape Town. (3) Department of Environmental Health, Boston University School of Public Health, Boston.
Despite their importance in the public health control of vector-borne disease, and their role in promoting agricultural productivity, pesticides may present significant public health problems
for exposed populations, particularly in developing countries. In particular, prenatal exposure to pesticides is increasingly thought to be responsible for a wide variety of health impacts on
pregnant women and their unborn fetuses, potentially manifesting in increased rates of birth defects, growth stunting and impaired childhood development.
Pilot research on 50 mother and child pairs in the northern Kwazulu Natal province of South Africa was conducted in 2005 to develop methods for a larger cohort study to assess the impact
of prenatal pesticide exposure. This paper reports on pilot data outlining potential occupational and non-occupational exposures to pesticides in this rural population.
Results. The mean age was 23 (range 16 to 44 years), and most women had completed 7 or more years of schooling.
Major potential sources of non-occupational exposure included reuse of washed pesticide containers (66%) for water storage or washing of clothes, indoor spraying by the malaria control
programme with DDT (94%), lack of any secure locked storage facility for pesticides (98%) and use of commercial domestic pesticide applications (cans) for domestic pest control (86%).
Although less common, other important potential sources of exposure included use of agricultural pesticides for domestic pest control (6%), mixing of pesticides with floor polish to control
domestic pests (4%) at home and 8% of the women were found to have empty containers from the Malaria control programme on their farm. Given that only two respondents had access to
running water, the potential for domestic pesticide contamination is exacerbated.
About 10% of respondents were involved in agricultural activities at the time of interview, including weeding, planting, and spraying of pesticides. Application was usually by backpack
although one women reported applying using a tractor spray device. None of the women could recall the names of the chemicals with which they worked, nor any hazards associated with
these chemicals. Use of protective equipment, and awareness of the label information was reported by only one respondent each.
These data suggest substantial opportunities exist for exposure to pesticides in rural farming populations in the region. For most women, exposures appear ubiquitous through a variety of
routes, while for a smaller group of women, high exposures typical of occupational exposures are present as well, with little awareness or protective measures.




SM1-PD-13                  INCREASES IN SERUM PCBS DURING EARLY PREGNANCY
M BLOOM1, G BUCK-LOUIS1, A LIU1, E SCHISTERMAN1, P KOSTYNIAK2
(1) Division of Epidemiology, Statistics, & Prevention Research, NICHD/NIH/DHHS, Bethesda, MD. (2) Department of Biotechnical & Clinical Laboratory Sciences,
University at Buffalo, SUNY, Buffalo, NY.
Introduction: Critical data gaps exist regarding the toxicokinetics of polychlorinated biphenyls during the peri-conception period, thereby, limiting our ability to assess potential reproductive
and/or developmental toxicants during this window.
Methods: To address this gap, 99 women planning pregnancies were recruited upon discontinuation of contraception and prospectively followed until hCG-confirmed pregnancy or 12 at risk
menstrual cycles. Blood specimens were obtained from 73 women at baseline, 63 specimens upon hCG-confirmed pregnancy, 52 following delivery, 13 after pregnancy loss, and 9 after 12
months without achieving pregnancy. Women completed daily diaries for the collection of lifestyle factors. Sera were analyzed for 88 PCB congeners using gas chromatography with electron
capture detection run in batches of 10 with 4 quality control samples. Serum lipids were quantified using enzymatic methods. Congeners were summed into three PCB groupings: total,
estrogenic (PCB IUPAC #s 4+10, 8+5, 15+17, 18, 31, 44, 47, 48, 52, 70, 77+110, 99, 101, 126, 136, 153, and 188), and anti-estrogenic (PCB IUPAC #s 77+110, 105, 114, 126, 156+171, and
169). Mean daily change in PCB concentration expressed as ng/g serum between baseline and each of three critical windows - early pregnancy, pregnancy loss and infertility – was estimated
using linear regression adjusting for baseline PCB values and average daily change in serum lipids as covariates.
Results: Among 50 women with live births, serum PCB concentrations increased for all groups. Mean daily increase in total PCBs from baseline to delivery was 0.05 ng/g (95% CI -0.01, 0.11)
and 0.04 (95% CI, 0.01, 0.07) for estrogenic congeners. Greater mean increases in PCB concentrations were observed between baseline and after an early loss. Concentrations remained
constant among 9 women who failed to become pregnant. While baseline PCB concentrations positively predicted the magnitude of the change during early pregnancy, serum lipids reflected
no effect across pregnancy or its spontaneous termination.
Discussion and Conclusions: Increases in serum PCB levels were observed among women after becoming pregnant irrespective of whether pregnancy ended in live birth or loss. Evidence
that higher early increases in serum PCBs was associated with pregnancy loss suggests the need to consider embryonic mortality in relation to PCB concentrations. The findings support early
mobilization of PCBs in the serum of newly pregnant women underscoring the ability of contaminants to impact embryonic and fetal development. The absence of changes among women
failing to become pregnant emphasizes corroborates the role of pregnancy in affecting serum concentrations in women.




                                                                                               51
        Symposia Session:

SM1-PD-14                  EVIDENCE FOR SEX SPECIFIC EFFECTS OF TCDD ON MALE AND FEMALE
                           PROGENY EXPOSED IN UTERO
B LE MAGUERESS-BATTISTONI1, F ODET1, C GUIGON2, A VÉROT1, R GUYOT1, R EL RAMY1, P DESSEN3, V LAZAR3, N COUDOUEL2, C BLEUX2, M FOREST1,
P LEDUQUE1, S MAGRE2
(1) INSERM U418 UMR INRA 1245 UCBL1, Lyon. (2) UMR CNRS 7079, Paris. (3) Institut Gustave Roussy, Villejuif.
Reproductive and developmental disorders are the most sensitive toxic effects caused by 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). The present study aimed at examining specific end-points
in males and females exposed in utero. To that purpose, pregnant rats were given one oral dose of TCDD (0.2 microg/kg) on gestational day 15. Preliminary experiments established that the
dose 1 microg/kg induced maternal and fetal toxicity. TCDD 0.2 microg/kg caused no changes in growth curve, testis weight, or testicular histology. Furthermore, intratesticular levels of
testosterone, 17-beta estradiol or DHT measured at 3, 20, 32, 45 and 90 days postnatal (dpn) were normal. However, we noticed an increased germ cell apoptosis during the juvenile period,
and a 30% decrease in sperm numbers at 67 and 90 dpn. We next used a microarray gene expression profiling (Agilent technology) to identify genes with altered transcript levels at specific
time-points during testicular development (3 and 32 dpn). Although data have yet to be validated using a quantitative RT-PCR procedure, it is worth to note that most genes identified encode
proteins involved in cell cycle and apoptosis. In female progeny, TCDD caused no changes in the onset of puberty or in fertility at the beginning of the reproductive life. However, we noticed
an increase in ovarian weight during the prepubertal period. We also observed that the growth of follicles in the prepubertal ovaries was accelerated. This was demonstrated using in situ
hybridization and specific markers of follicular maturation. Furthermore, preliminary results obtained from microarray gene expression profiling and quantitative RT-PCR allowed the
identification of endocrine-related genes induced by TCDD. Our data are in line with the hypothesis of an alteration of the endocrine function in the prepubertal ovary. Collectively, our data
suggest that an embryonic in utero exposure to an endocrine disruptor (TCDD) may alter the endocrine environment of the germinal lineage during the prepubertal period, in the female
progeny. In the male progeny, the TCDD treatment would result in an abnormal germ cell differentiation in the juvenile period with a subsequent alteration of the adult spermatogenic capacity.
PLEASE DO NOT MODIFY THE TEMPLATE FRAME
otherwise the abstract will be rejected by the Programme Committee




SM1-PD-15                  DIOXIN EXPOSURE AND QUALITY OF OVARIAN FUNCTION
M WARNER1, B ESKENAZI1, D OLIVE2, S SAMUELS3, L NEEDHAM4, D PATTERSON4, S MILES1, P VERCELLINI5, P GERTHOUX6, P MOCARELLI6
(1) University of California, Berkeley, CA. (2) University of Wisconsin Medical School, Madison, WI. (3) State University of New York, Albany, NY. (4) Centers for Disease
Control and Prevention, Atlanta, GA. (5) Mangiagalli Hospital, University of Milan, Milan. (6) Hospital of Desio, University of Milano-Bicocca, Desio-Milano.
Introduction: 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD), a widespread environmental contaminant and known endocrine disruptor, has been associated with alterations in ovarian function
and steroid levels in animal studies. No studies have examined the potential effects of TCDD exposure on human ovarian function.
On July 10, 1976, as a result of a chemical explosion, residents of Seveso, Italy experienced the highest levels of TCDD in a human population. Twenty years later, the Seveso Women’s Health
Study (SWHS), a retrospective cohort study of the reproductive health of the women, was initiated. We aimed to examine the relationship of TCDD levels measured in sera collected near the
explosion and quality of ovarian function (ovarian follicles, ovulation rate, serum hormone levels) 20 years later.
Methods: The SWHS cohort comprises 981 women who were 0 to 40 years in 1976, resided in the most contaminated areas at the time of the explosion, and had archived sera collected
soon after the explosion. Individual TCDD exposure was measured in archived sera by high-resolution mass spectrometry. Eligible for the ovarian function analysis were 363 women who were
20 to 40 years old and did not currently use oral contraceptives.
We examined the association of archived serum TCDD levels with serum hormone levels (estradiol, progesterone) at follow-up among 129 women who were in the luteal phase of their
menstrual cycle. Women were classified as having ovulated if their serum progesterone level was >3 ng/ml. We examined the relationship of archived serum TCDD levels with ovarian follicles
(number, size) among 96 women who underwent transvaginal ultrasound during the follicular phase of their menstrual cycle.
Results: The median lipid-adjusted serum TCDD level for the 363 women was 77.3 ppt (range: 2.8-17,300). Of 129 women in the luteal phase, 87 (67%) were classified as having ovulated.
Serum TCDD (logTCDD) was not associated with odds of ovulation (adjusted odds ratio (adj-OR) = 0.99; 95% Confidence Interval (CI): 0.5, 1.9). Among those who had ovulated, logTCDD was
associated with a non-significant decrease in progesterone (adj-beta = -0.70; 95%CI: -2.42, 1.01) and estradiol (adj-beta = -1.81; 95%CI: -10.4, 6.8). Among women in the follicular phase,
serum TCDD was not associated with number or size of ovarian follicles.
Discussion and Conclusions: These results suggest TCDD may not have significant effects on human ovarian function 20 years after exposure in this highly exposed population. Detailed
results will be presented and interpreted in light of study advantages and limitations.




SM1-PD-16                  DOES POLYCHLORINATED BIPHENYLS AND DIBENZOFURANS INDUCE SPERM
                           ANEUPLOIDY IN MEN
YL GUO1, PC HSU2, W HUANG3, PL KUO3
(1) National Taiwan University, Taipei. (2) National Kaohsiung First University of Science and Technology, Kaohsiung. (3) National Cheng Kung University, Tainan.
Introduction. Among men heavily exposed to polychlorinated biphenyls (PCBs) and polychlorinated dibenzofurans (PCDFs), increased percent of oligospermia, reduced sperm capability of
penetrating oocytes, and reduced percent of male offspring were reported by our study group. The mechanism of such effects is unknown. We test the hypothesis that effects on gender ratio
and sperm functioning is related to sperm chromosomal aneuploidy.
Methods. During 1978-79, over 2,000 Taiwanese ingested rice oil contaminated with PCBs and its pyrolytic product, mainly PCDFs. This episode was referred to as “Yucheng” (“oil-disease”
in Chinese) exposure. In 1999-2000, exposed men and their controls aged 37-50 years were recruited for physical examination followed by semen analysis. Semen samples were studied for
chromosomal aneuploidy by fluorescent in situ hybridization (FISH) with an established procedure in our laboratory. Hybridization procedures were done using centromere-specific probes
directly labeled by fluorescence dyes (Vysis, Downers Grove, IL), including CEP X (SpectrumAqua), CEP Y (SpectrumOrange) and CEP 8 (SpectrumGreen). Each cell was scored for the number
of fluorescent spots.
Results. A total of 50 men exposed to PCBs/PCDFs and 34 unexposed matched controls volunteered to participate in the study. The age, body height and weight, percent of smokers, and
period of abstinence from ejaculation were not different between exposed and control men. Percents of sperm with normal numbers of chromosomes X,Y, and 8 were not different between
the two groups. Percents of sperm with aneuploidy of sex chromosomes or chromosome 8, and those with diploidy were not different between Yucheng and control men. Ratios of normal
X/normal Y sperm were not different between the exposed and control group. However, among Yucheng men, 4 (8%) had normal X/normal Y sperm ratio of >1.4, and among controls, none
had such elevated X/Y ratio (not statistically significant by Fisher’s exact test).
Discussion and Conclusions. Chromosomal aneuploidy, diploidy, and X/Y sperm ratio were not increased in men highly exposed to PCBs/PCDFs despite reduced sperm capability of oocyte
penetration, mildly increased abnormal morphology, and changed gender ratio in exposed men. The mechanisms of the latter remained to be determined.




                                                                                             52
                                                                                                                                                                                                        SUNDAY SEPTEMBER 3
         Symposia Session:

SM1-PD-17                   DEVELOPMENT OF A PBPK MODEL FOR 2,2’,4,4’ TETRABROMODIPHENYL
                            ETHER IN RATS AND ITS EXTRAPOLATION TO HUMANS FOR MEASUREMENT
                            OF IN UTERO EXPOSURES.
J RAYMER1, W STUDABAKER1, L MICHAEL1, CE GARNER1
(1) RTI International, RTP.
Introduction. Exposures to environmental contaminants in-utero and during childhood create the potential for a variety of adverse health effects including abnormal or disturbed development
of organ systems. Work at RTI is focused on the development of a PBPK model for 2,2’4,4’-tetrabromodiphenyl ether (BDE-47) to permit the prediction of exposure both in-utero and during
the neonatal period. The model was developed for rats, extrapolated to humans and evaluated using maternal blood, cord blood, meconium, and breast milk. In addition, PBDE content in
house dust collected from the residences of the mothers was analyzed for PBDEs.
Methods. Sprague-Dawley rats were dosed either with a single dose or 10-day repeated (iv or PO) at a total dose of 0, 1 or 10 mg/kg. Samples of blood, urine, and feces were collected at
various times post-dose. Liver, kidney, adipose, muscle, skin, and brain were collected 10 days post-dosing in pregnant and non-pregnant animals. Tissues and fluids from near-term fetuses
were also acquired including amniotic fluid, placentae, fetal brain, liver, and the remainder of the fetus. The extraction method involved the addition of a surrogate standard followed by
homogenization/extraction with hexane, SPE cleanup using Florisil, and analysis by GC/ECD. Berkeley Madonna (v 8.0) software was used for modeling.
A total of 25 women were recruited into the study. Maternal blood, cord blood, and meconium were collected. Baseline data were acquired by hospital staff. At least four weeks following
delivery, a visit was made to the participants’ homes and samples of breast milk and floor dust were collected.
Results. Recoveries of BDE-47 from blood and tissues ranged from 90-110%. Concentrations of BDE-47 were measured in all human samples and ranged from 0.9 to 2.9 pg/mL and from 0.3
to 2.6 pg/mL in maternal and cord blood, respectively, not adjusted for lipid. Other PBDEs also appeared to be in the samples.
Discussion and Conclusions. BDE-47 concentrations in cord blood were similar to those in maternal blood and show in-utero exposure. Additional data to be available include breast milk
concentrations; lipd adjustments will allow for more consistent evaluations. Analysis of the meconium data will provide insight into this medium for cumulative exposure estimations.
This research has been supported by grant R830756010 from the U.S. Environmental Protections Agency’s STAR program.




SM2-PD                      ENVIRONMENTAL EQUITY (II)
SM2-PD-01                   PARTICIPATION IS THE KEY TO PARTICIPATORY COMMUNITY ENVIRONMENTAL
                            HEALTH STUDIES.
DF GOLDSMITH1, PR DAVIDSON2
(1) Dept Env/Occup Health, George Washington Univ, Washington. (2) Dept Sociology, George Washington Univ, Washington.
Background: Epidemiologists are key to developing effective responses to environmental hazards when they impact communities. However, we know that absent strong community
investment in the importance of health research, there will be poor response rates, leading to biased risk estimates. Thus, in democratic societies how to encourage high enough participation
in health surveys so that findings will have substance for both the community and for science and medicine. We use as an example a neighborhood exposure to leaking chemical storage and
a participatory community environmental health (PCEH) research effort, focusing on both acute and chronic health effects. Health would be assessed by both symptom surveys and clinical
exams for both adults and children residing in the neighborhood and parallel assessments in a control community.
Processes for increasing community participation: In the literature there are several methods for increasing and encouraging study participation—1) link provision of health care with
survey follow-up; 2) develop joint investigator-community efforts to enhance outreach; 3) provide modest financial or gift incentives for each participation event where data are collected or
health education offered; 4) offer prizes or lottery tickets for participants (depending on underlying cultural acceptance in the community); 5) direct appeal to community leaders and members
for participation as the only basis for assessing valid, scientific findings of the link between exposure and illnesses.
Discussion: Each of these methods or combination of methods must be placed in context of noncoercive study participation, consistent with ethical informed consent. However PCEH research
must also acknowledge that it occurs in a fishbowl including local and national media, language and cultural barriers, politics, lawsuits seeking redress of perceived or real grievances, laws
limiting the sharing of health data (HIPPA in U.S.) or access to health or exposure registry information. These all combine to erect barriers between investigators and community members, and
directly impact willingness to participate. Thus, when there is confusion, a logical community response is to choose not to participate. This may then lead to overrepresentation in the study
population of residents with other agendas, and bias rendering the PCEH research of uncertain scientific or epidemiologic value.
Conclusion: Attention needs to be paid to increasing community participation in serious environmental health and contamination studies. Epidemiologists conducting PCEH must focus on
the study impacts of these issues, or else our findings will not achieve the goals we have: community insight (co-learning) to what has happened to their health nor to a scientific understanding
of the health effects related to human exposure to toxins.




SM2-PD-02                   USE OF SPATIAL AUTOCORRELATION TO INVESTIGATE CLUSTERING OF
                            HEALTH DEPRIVATION
E SOFIANOPOULOU1, T PLESS-MULLOLI1, S RUSHTON2
(1) School of Population and Health Sciences, University of Newcastle upon Tyne, Newcastle upon Tyne. (2) Institute for Research on Environment and Sustainability,
University of Newcastle upon Tyne, Newcastle upon Tyne.
Introduction
The susceptibility of people to develop ill health can be influenced by characteristics of their neighbourhood. Moreover, populations in socio-economic deprived areas are considered more
susceptible to poor health compared to those in affluent areas. Here, we aimed to use methods of spatial statistics and apply them in the context of environmental epidemiology. Specifically,
we aimed to test for spatial autocorrelation to investigate clustering of indicators of health, in the Northeast of England.
Methods
We abstracted data on the English Health Deprivation and Disability Index (HDDI), which comprises of four measures of ill-health. These were downloaded for the smallest possible spatial
unit comprising of a mean population of 1500, called Lower Super Output Areas (LSOA).
We used Geographical Information Systems to map HDDI for the Northeast of England, as well as for each county and unitary authority (UA) in the region.
We used the global Moran’s I statistic to investigate spatial autocorrelation. Global autocorrelation statistics provide a single measure for clustering (positive spatial autocorrelation) or
dispersion (negative spatial autocorrelation) of an attribute, in a spatial unit. A Moran’s Index value near +1.0 indicates clustering, while an index value near -1.0 indicates dispersion.
Results
We detected positive spatial autocorrelation in all spatial units assessed. A high positive z-score for all tests, indicated that the observed clustering was statistically significant (p<0.01). The
strongest positive spatial autocorrelation was detected in counties Durham (Moran’s I=0.68, z=10.4) and Northumberland (Moran’s I=0.68, z=8.5). Clustering of health deprived areas was
apparent in the eastern part of both counties. Weak positive spatial autocorrelation was found in Darlington UA (Moran’s I=0.18, z=3.8) and Hartlepool UA (Moran’s I=0.24, z=3.2). Moderate
positive spatial autocorrelation was detected for the Northeast of England (Moran’s I=0.34, z=22.4) and the other spatial units.
Discussion and Conclusions
The global Moran’s I test was successful in identifying spatial autocorrelation of health indicators, separating spatial units with high, moderate and weak clustering. In those areas where
county and UA level clustering was moderate or weak the local Moran’s I test may be able to detect more localised clustering in smaller spatial units. We conclude that susceptibility to develop
ill health is strongly clustered at county level in some counties but not in others.




                                                                                                53
        Symposia Session:

SM2-PD-03                   ENVIRONMENTAL EQUITY IN THE NETHERLANDS
H KRUIZE1, PPJ DRIESSEN2, K VAN EGMOND3, P GLASBERGEN2
(1) NATIONAL INSTITUTE OF PUBLIC HEALTH AND THE ENVIRONMENT, BILTHOVEN. (2) UNIVERSITY OF UTRECHT, UTRECHT. (3) ENVIRONMENTAL ASSESSMENT
AGENCY, BILTHOVEN.
Introduction
Environmental equity has received a huge amount of attention in the USA, and lately in Canada, South-Africa and United Kingdom as well. Recently, we performed an environmental equity
study in the Netherlands, in which we 1) analysed socio-economic differences in exposure to environmental risks; 2) investigated the mechanism leading to these differences; 3) investigated
differences between actual environmental quality and the way citizens perceive it; and 4) assessed the effects of different perspectives on equity on these issues. We will present the results
of our analyses on socio-economic differences in exposure to environmental risks.
Methods
We used secondary data and selected exposure to traffic noise, NO2, safety risks of aircraft, transportation and industry, and access to public green areas as environmental indicators, and
analysed them separately and accumulated. Exposure was defined to be ‘bad’ if levels were higher than the (legal) standard or ‘good’ if levels were lower than the target value. Household
income was selected as socio-economic indicator. Analyses at the household level were performed for the Netherlands nationwide, and for two regions with a high spatial pressure and large
economic activities, the Rijnmond region and Amsterdam Airport region
Results
Generally, exposure to environmental ‘bads’ is limited for traffic noise and NO2, but occurs more often in the two selected regions, together with a lesser access to public green areas.
Simultaneous exposure to ‘bads’ of traffic noise and NO2 does not occur often, except for the Rijnmond region, where the average income level is lower as well.
Differences between income categories are most pronounced for railroad noise and access to public green areas. For the other indicators no consistent trend was found. Higher incomes more
often have access to (accumulation of) environmental ‘goods’.
Discussion and Conclusions
Exposure to levels above a standard is generally limited in the Netherlands for traffic-related pollutants, which can be attributed to environmental regulation. It occurs more often in the
Rijnmond region, due to a high spatial pressure and heavy traffic. Differences between income categories in exposure to environmental ‘bads’ exist, but are mostly small.
Higher income categories have more access to -accumulation of- environmental ‘goods’ than lower income categories, probably because these higher income categories generally have more
choice to select their residential location. In addition, they may have more access to information and may be better in defending themselves against unwanted environmental situations.




SM2-PD-04                   THE PAISA PROJECT: STUDYING THE RELATIONSHIPS BETWEEN AIR
                            POLLUTION, SOCIO-ECONOMIC STATUS AND ASTHMA ATTACKS
O LAURENT1, C SEGALA2, E RIVIERE3, C SCHILLINGER3, L ROUIL4, S DEGUEN1, S HAVARD1, L FILLEUL5, D BARD1
(1) LERES (ENSP), RENNES. (2) SEPIA Sante, MELRAND. (3) ASPA, SCHILTIGHEIM. (4) MECO (INERIS), VERNEUIL EN HALATTE. (5) CIRE Aquitaine (INVS), BORDEAUX.
Socioeconomic gradients in health are well documented in developed countries, but incompletely explained. Besides, many environmental exposures are suspected to be health risk factors,
while relatively few are well-established. However, research in these different domains is essentially carried out independently. Since a part of health socioeconomic gradients might be
explained by environmental exposures, our objective is to explore relationships between socio-economic status (SES), environmental exposures and health. We chose to study a short term
effect of an environmental exposure, that is, asthma attacks and atmospheric pollution, in relation to SES. Such a work on asthma was never conducted in Europe. The design of our studies
is ecological and will test the associations between levels of atmospheric pollution and i) emergency visits of physicians for asthma attacks and ii) figures of respiratory drug sales in relation
to SES. The setting is an eastern French conurbation of 450.000 inhabitants. The statistical unit is the smallest area (IRIS) for which SES data are available through the French National Institute
of Statistics (INSEE). Data on emergency visits have been obtained from the two emergency and healthcare networks operating in the conurbation for years 2000-2002. Data on respiratory
drug sales for 2004 have been obtained from the 3 French Medical Insurance Systems. Levels of PM10, O3, NO2, and SO2 in each IRIS will be modelled on an hourly basis from emission
inventories, meteorological data and pollution measurements of the Local Air Monitoring Association using the deterministic model ADMS-Urban. An uncertainty and sensitivity analysis of the
air pollutants modelling will be carried out, allowing analyzing the effects of modelling uncertainty on observed associations. We’ll study i) the distribution of both health indicators across
IRISes exhibiting contrasted SES; ii) the distribution of ambient air pollutants concentrations across IRISes exhibiting contrasted SES; iii) the modulating role of IRISes SES on the relationship
between air pollution levels and asthma attacks using both case crossover and time-series approaches. Potential confounding factors (influenza epidemics, pollen release bursts and
meteorology) will be taken into account in the analyses. Age and sex specific analyses will be carried out as far as possible, depending on the number of health events registered. As health
data have been collected during first 2006 quarter, first results on the distribution of health indicators across IRISes exhibiting contrasted SES will be presented. This work is funded by the
French National Research Agency.




SM2-PD-05                   THE MODIFYING EFFECTS OF INDIVIDUAL SOCIO-ECONOMIC STATUS ON
                            MORTALITY RISK ASSOCIATED WITH AIR POLLUTION
CQ OU1, RY CHUNG1, TQ THACH1, KP CHAN1, YK CHAU1, L YANG1, TH LAM1, AJ HEDLEY1, CM WONG1
(1) The University of Hong Kong, Hong Kong.
Department of Community Medicine, School of Public Health, The University of Hong Kong, Hong Kong SAR, China
Introduction: While socio-economic inequity is a continuous concern in public health, its role in modifying air pollution effects is now a major research interest. We will explore whether
individuals with low socio-economic status (SES) are more susceptible to the acute effects of ambient air pollution than those with high SES.
Methods: The study included 24,656 Hong Kong Chinese who died of natural causes at the age of 30 or above in 1998. Information on individual socio-economic status as well as lifestyle
and physical status was obtained by interviewing the informants via a standardized questionnaire. Based on physical demands in jobs, occupational classes were categorized into three groups:
blue-collar, jobless and white-collar. Levels of education were classified into no formal education, primary education and secondary or higher education. Two types of housing were considered:
self-owner and public estate. Poisson regression was performed to assess excess risks per 10 mg/m3 increase of air pollutant (ER) for each SES subgroup. The differences in ER between SES
subgroups were estimated by inclusion of interaction terms (i.e. indicator of SES times concentration of each air pollutant) in the model for data with all subgroups put together. Case-only
logistic regression was used as sensitivity analysis with adjustment for potential confounding by lifestyle and physical status.
Results: We found that the effects of air pollution were significantly associated with types of housing, occupational classes, and levels of education. The differences in ER between people
living in public estate and those owning private house were 1.63% (95% CI: 0.02%, 3.61%) and 1.65% (0.45%, 2.84%) for NO2 and PM10, respectively. The highest effects of air pollution
were observed in those with primary education among three education subgroups, and the differences in ER were significant for NO2, O3, and PM10 (p<0.05) but not for SO2. ER for all pollutants
were significantly greater in blue-collar workers compared to the other two occupational classes (p<0.05). The results changed slightly but in general remained significant and were consistent
after adjustment for potential confounders.
Discussion and conclusions: The result provides insight to the role of individuals’ socio-economic status in modifying the effects of air pollution on mortality.




                                                                                                54
                                                                                                                                                                                                      SUNDAY SEPTEMBER 3
        Symposia Session:

SM2-PD-06                  HEALTH IMPACT OF INDOOR AIR POLLUTION: A STUDY OF GROWING URBAN
                           CENTRE ALIGARH (INDIA)
A RAHMAN1, A SARKAR2
(1) JAMIA MILLIA ISLAMIA UNIVERSITY, DEPARTMENT OF GEOGRAPHY, NEW DELHI. (2) DEPARTMENT OF POLICY STUDIES, TERI SCHOOL OF ADVANCE STUDY, NEW
DELHI.
The environmental concern for air pollution has largely been focused on the question of pollution outdoors. However the attention of scientists has shifted on the quality of air indoor i.e.
within the buildings. The indoor air pollutants are nearly the same as that of out door one. But in some instances the concentration of indoor pollutants exceeds the standard set for the
outdoor. These pollutants reach such a high level although they are emitted in small volume because pollutants can not escape out easily from the buildings due to lack of proper ventilation
in most of the houses. Indoor air pollution is more harmful and poses greater health hazards because on an average a person spends nearly 16 to 18 hours indoors i.e. inside the house or in
the offices. So it was thought worth studying the indoor air pollution and its impact on the health of a growing urban centre of India, Aligarh.
This study is based on purposive sampling method and about 3 per cent i.e. (2,185) households were sampled. These households were divided into five income groups i.e. from very low to
very high-income groups and an attempt was made to see the income-wise indoor air pollution conditions and its health impact on the sampled residents of Aligarh city. The paper addresses
the use of various fuels for cooking (LPG, kerosene, wood, coal, saw dust, dung cake, dry leaves), place of cooking, outside smoke coming in, ventilation and floor space in the sleeping rooms.
Relationship between the indoor air pollution and health impact were drawn using the Chi-Square technique by adjusting other confounding variables such as tobacco smoking, chewing
tobacco, allergic bronchial asthma, previous history of chronic respiratory diseases, TB etc.
The study unfolded that indoor air pollution, crowding and humidity contributes to the respiratory problems in the more than 50% of the samples residents. It is seen that even upper-middle
class people who live along the main roads are more at the risk of respiratory infections because the use smoky fuels and also emission form heavy traffic for most of the day time. Further
low quality housing and leaky roofs also facilitate the spread of respiratory infections to a large number of poorer residents. The disease tends to be associated with income and poverty as
well as the living and working environment. A large section of the poorer households (74% in low and 89% in the very low income) are suffering from respiratory diseases mainly because
they are labourers working in various lock, polishing and other household industries. But only 10 to 15 % of the high and very high-income households are also suffering from the respiratory
problems. They are the parents and grand parents of the rich people and due to the old age they are more at the risk of respiratory diseases especially when they walk in the smoky
neighborhood.
There is an urgent need to improve the quality of life of vulnerable section of the society. People of Aligarh can only get some respite from the respiratory diseases only with the use of cleaner
fuel for cooking by giving financial subsidy to purchase it, use of improved stoves in the poorer income households, providing good working environment and provision of better low class
housing facilities by the Government.




SM2-PD-07                  THE HEALTH IMPACT OF COLD HOMES ON OLDER PEOPLE:
                           A POPULATION BASED STUDY
J RUDGE1, R GILCHRIST1
(1) London Metropolitan University, London.
Introduction
This research aims to contribute to the evidence base for links between cold homes and health among older people. Fuel poverty is defined as ‘the inability to afford adequate warmth’ and
its health effects are conventionally measured by annual numbers of excess winter deaths among the older population. Because of consequent implications for winter pressures on hospital
beds, this work attempts to measure excess winter morbidity rather than mortality. Cross-disciplinary approaches are required to identify and address fuel poverty, since housing, socio-
economic and health issues are involved. The reported methodology (and its further development) could help to encourage interagency partnership towards dealing with health consequences
of cold housing.
Methods
A population approach was used to examine the relationship between excess winter hospital admissions and the risk of older people living in cold homes in one London borough. We counted
emergency respiratory hospital episodes for people aged 65 years and over, for the 4 years 1993-1996 at enumeration district (ED) level, where EDs include approximately 220 households,
on average. We calculated a Fuel Poverty Risk Index (FPR) for all EDs, based on combined factors of low income, population age, poor energy efficiency of dwellings and under-occupation of
dwellings. Statistical analysis was made of links between excess winter counts of episodes, FPR and other explanatory variables.
Results
Annual winter and summer morbidity counts for individual EDs were frequently zero. In previous work, therefore, data was aggregated for groups of EDs, and showed FPR to be a predictor
of excess winter morbidity for older people. Subsequently, all 450 EDs have been considered, and various distributions used to allow for over-dispersed data. This confirmed an interaction
between FPR and excess winter counts. The ecological fallacy is recognised as associated with small area studies, but as many explanatory variables as possible are under examination to
characterise the population at risk.
Discussion and Conclusions
FPR is a predictor for excess winter respiratory morbidity at small area level – and could start to provide information about costs to health services of cold, energy inefficient housing. There
may be public health implications for increased health-driven energy-efficiency investment in addition to that demanded by energy-saving targets, which tend to skew energy efficiency
investment in favour of fuel-rich households. Ongoing work, funded by the ESRC, aims to improve and refine the methodology by further detailed statistical exploration of potential
confounding variables, including weather, in the project data.




SM2-PD-08                  EXPOSURE TO NOISE AND WELL BEING IN SOUTH GERMANY
N MEYER1, H SPEGEL1, I KREUZMAIER3, A SCHULZE1, C HEUMANN3, R VON KRIES2, D NOWAK1, K RADON1
(1) Institute for Occupational and Environmental Medicine, Ludwig-Maximilians-University, Munich. (2) Institute for Social Pediatrics and Adolescent
Medicine, Ludwig-Maximilians-University, Munich. (3) Department of Statistics, Ludwig-Maximilians-University, Munich.
Introduction: So far, most noise studies took only one source of noise exposure into account (e.g., aircraft or traffic noise). The aim of this study was to assess the individual noise exposure
of different age groups using personal monitoring in an inner-city general population sample.
Methods: Within this population-based cross-sectional survey a random sample of 268 children, 310 adolescents and 430 adults living in a big city were invited to take part in the study.
Personal noise exposure was assessed using Larson Davies dosimeters (lower sensitivity range: 40 dB (A), dynamic range: 75 dB (A)). For each subject individual exposure levels were assessed
every 5 seconds over a 24 hour period. Mean exposure levels with 95% confidence intervals (95%CI) were analysed stratified for age group and by time of the day. In addition, personal
exposure levels during school holidays were compared to those during the school year.
Results: The overall response rate was 63%. During morning and afternoon hours mean levels of noise exposure were significantly higher among children as compared to adolescents and
adults. In contrast, children had significantly lower levels of noise exposure during the night (table). Morning exposure levels were significantly lower for measurements done during school
holidays as compared to those obtained during the school year.
                                                                                                                  Table: Mean (95% CI) level of noise exposure by time of the day and age
                                                                                                                  group
                                                                                                                  Conclusions: During daytime higher levels of noise exposure might be seen in
                                                                                                                  children. Part of this exposure might be self-induced by e.g. activities at the
                                                                                                                  schoolyard. In order to assess differences in noise exposure between urban
                                                                                                                  and rural areas, two rural towns will be included in the study. In addition, the
                                                                                                                  potential association between exposure and well-being will be analysed.




                                                                                               55
        Symposia Session:

SM2-PD-09                  SOCIAL DETERMINANTS AND DISTRIBUTION OF CHRONIC LEAD POISONING
                           AMONG THE URBAN CHILDREN: A TALE OF TWO CITIES IN INDIA
S GHOSH1, A SARKAR2, M ANAND3
(1) COUNCIL FOR SOCIAL DEVELOPMENT, NEW DELHI. (2) DEPARTMENT OF POLICY STUDIES, TERI SCHOOL OF ADVANCED STUDIES, NEW DELHI. (3) SCHOOL OF
PUBLIC HEALTH AND TROPICAL MEDICINE, TULANE UNIVERSITY, NEW ORLEANS.
Introduction:
Chronic lead poisoning among children is a major public health problem in India and is believed to be associated with the ambient air pollution. Although previous research endeavors have
dealt with the health impact or emphasized on the exposure pathways, there are very few studies that explore and detail the social dimensions of the issue or identify lead levels among
children in developing countries. The objective of the present study is to find out plausible linkage between the various social factors and lead level among children in two Indian cities.
Methods:
Data for the present study is essentially drawn from India’s National Family Health Survey (NFHS-2) and also from systematic review of relevant literature to get a more qualitative picture.
The survey primarily collected information regarding lead level in blood among 1,078 children born during three years preceding the survey from the two metro cities (Delhi and Mumbai),
where the level of ambient air pollution was very high. Since the present study focuses on the social determinants and distribution of the lead exposure, a number of demographic and social
variables like age, birth order, nutritional status, living standard, educational attainment of mothers, religion and caste, use of main cooking fuel and utensils are controlled in the analysis.
Altogether nine sets of multivariate logistic regression models, three each for two cities and combination of these three, are employed in the analyses to see how demographic and social
factors are responsible for elevated lead level at different stages of analyses.
Results:
About 50% and 45% of the children under age three years living in the vulnerable areas of the two cities have elevated lead levels in their blood (•10•g/dL). Significant association has been
observed between elevated lead level and age of the child and also with childhood anemia, maternal education and household’s standard of living in the multivariate analysis even after
controlling all other demographic and socio-economic variables.
Discussions & conclusions:
The study shows that lead exposure particularly among the children is a serious public health problem in the metros in India. People from deprived sections require more attention as they are
the most vulnerable. Urgent action from the Government is needed to provide clean environment by reducing the use of leaded gasoline on the one hand and overall socio-economic
development of the urban poor through special schemes and programme on the other.




SM2-PD-10                  SCHOOL ENVIRONMENT AND CHILDREN HEALTH IMPACT IN DEVELOPING
                           COUNTRY
SD JOSHI1, N PANDIT2, SK BK2, B SHAHI2
(1) NEPAL MEDICAL COLLEGE ANDTEACHING HOSPITAL NEPAL, KATHMANDU. (2) COMMUNITY HEALTH AND ENVIRONMENTAL SOCIETY NEPAL, KTM.
Background:
Developing country like Nepal has poor health status in school students
Objective: To know and evaluate the indoor environmental condition of government and private-schools and the health impacts on students.
Methods:
A cross sectional studied of representative samples of 35 schools of selected region of Nepal including government and private schools (N=3572) from 2003-2005. Onsite observation and
health check up & interview with students and teachers were done. Specific scores was given in each criteria. The data were analysed and edited in EPI info program.
Results:
The results shows that 89% of government school and 45%of private-schools have poor environmental condition and 69% of government schools students are suffering from environmental-
health problem while only 22% of private school student are suffering from some kind of diseases. Government-school don’t have the standard classroom, adequate sports facility, safe drinking
water, light and ventilation in comparison with private schools.
Conclusions:
We conclude that the main causes are poor socio-economic status, illiteracy of parents, negligence, hard housework for children, diseases, malnutrition, incomplete immunization and lack of
health education. The poor environment condition includes crowed students in a classroom, poor ventilation, shortage of clean drinking water, untidy clothes of students, poor nutrition and
lack of greenery in the school area, school near by road, air pollution and lack of environmental awareness among teachers and parents. The government schools have limited budget, resources
with compared to private schools and most of the lower and lower middle class family children are studying in government schools which covers nation 82% of total students.
Recommendation:
The government should allocate the special budget to the government schools and it should be utilized from the available resources such as good ventilation, limited student in a class,
awareness among teachers and parents. Last but not the least, this type of programs are helpful to prevent from environmental health hazards also.




SM2-PD-11                  ASSOCIATIONS BETWEEN SOCIO-ECONOMIC STATUS AND ENVIRONMENTAL
                           EXPOSURES IN FINLAND
M JANIS1, K PASANEN1, M KOLEHMAINEN2, N KARVOSENOJA3, P VERKASALO1
(1) National Public Health Institute, Kuopio. (2) University of Kuopio, Kuopio. (3) Finnish Environment Institute, Helsinki.
Background. Substantial differences in health have been described between different socioeconomic groups. Some remote parts of Finland are known to be disadvantaged both in terms of
health and socioeconomic circumstances. Until recently little attention has been paid to the possibility that environmental exposures may partly determine socioeconomic or health status.
Objectives. This study investigates associations between socioeconomic and environmental variables in Finland. We also explore the applicability of geoinformatics and computationally
intelligent methods in environmental health.
Materials and methods. We used a geographic information system which stores for year 2002 registry data (in 250 m x 250 m grid squares) from the Statistics Finland, the National Land
Survey of Finland, and the Finnish Environment Institute. Source proximity was assessed for each grid square. Associations between socioeconomic and environmental variables were assessed
using Pearson correlation coefficients, K-means clustering and neural networks. GIS was used to integrate different data sources, to visualize spatial variation of original data and results from
K-means clustering, and to explore residuals of the MLP-models.
Results. The strongest positive associations (Pearson correlation coefficients from 0.49 to 0.63) were between household income, education and social class, and the strongest negative
associations between household income, education, social class and unemployment (? 0.50). In greater Helsinki area, household income showed a clear positive association with living area
(in m2 per person) (0.66) and proximity to shoreline (0.24), and a clear negative association with population density (-0.32), proximity to railroads (-0.19) and main roads (-0.18), and outdoor
PM10 levels (-0.22). Elsewhere the respective correlations were 0.23, 0.04, -0.22, -0.10, -0.20 and 0.04. In Helsinki area, the association between household income, education, social class
and proximity to industrial sites was relatively weak (-0.08 to -0.07), and the association was still weaker in other parts of the country (-0.04 to -0.03). MLP-models confirmed the relationship
between environmental exposures and household income (IA ? 0.80). Large scale maps showed clear variations in study variables. K-means clustering resulted in interesting spatially-clustered
combinations of socioeconomic and exposure variables.
Discussion. Our preliminary analyses suggest weak associations between socioeconomic and environmental factors in Finland. These associations were stronger in greater Helsinki area.
Interpretation of results is complicated by small number of exposed grid cells, selection of study area, and non-linearity of effects. Additional work is needed to properly characterize the
associations and their effects on health.




                                                                                               56
                                                                                                                                                                                                     SUNDAY SEPTEMBER 3
        Symposia Session:

SM2-PD-12                  THE EFFECT OF ENVIRONMENTAL STRESSORS ON BLOOD PRESSURE AND
                           HYPERTENSION IS GREATER IN THE DUTCH ETHNIC MINORITY GROUPS
C AGYEMANG1, C VAN HOOIJDONK1, W WENDEL-VOS1, J UJCIC-VOORTMAN2, E LINDEMAN3, M DROOMERS1
(1) Centre for Prevention and Health Services Research (pb 101); RIVM, Bilthoven. (2) Dept. of Epidemiology Documentation and Health Promotion, GGD Amsterdam,
Amsterdam. (3) O&S, Gemeente Amsterdam, Amsterdam.
Background: High blood pressure (BP) is one of the important causes of cardiovascular diseases and its role is set to continue. In western societies, BP levels and prevalence of hypertension
differ by ethnic group with most studies showing higher levels and higher rates in the minority ethnic groups than their European counterparts. The explanations for the higher BP levels and
the higher prevalence of hypertension in ethnic minority populations still remain unclear. Evidence strongly suggests that the neighbourhood in which people live influences their health.
Despite this the environmental effect on BP and hypertension in different ethnic groups has hardly ever been examined. Because many ethnic minority populations live in disadvantaged
neighbourhoods with high levels of stress, neighbourhood factors may provide important clues for explaining the higher BP levels and hypertension rates in ethnic minority populations in
western countries.
Objective: To investigate associations between environmental stressors (crime, housing density, nuisance from alcohol and drug misuse, quality of green space and social participation) and
systolic and diastolic blood pressure (BP) and hypertension in different ethnic groups in Amsterdam.
Design: Multilevel regression analyses were performed separately for each ethnic group
Participants: The study sample consisted of 517 native-Dutch, 404 Turkish and 365 Moroccan ethnic groups living in 15 neighbourhoods.
Main outcome measures:Systolic BP, diastolic BP and hypertension
Results: Amongst Moroccan, high density housing and nuisance from drug misuse were associated with a higher systolic BP while high quality of green space and social participation were
associated with a lower systolic BP. High level of nuisance from drug misuse was associated with a higher diastolic BP. Amongst Turkish, high level of crime and nuisance from motor traffic
were associated with a higher diastolic BP. High quality of green space was associated with lower odds of hypertension in the Moroccan ethnic group. Similar associations were also observe
amongst the Dutch but none of the differences were statistically significant.
Conclusion: The study findings indicate that the adverse effects of neighbourhood stresses on BP and hypertension are greater in ethnic minority groups. These findings may indicate that the
higher BP levels and prevalence of hypertension found in some ethnic minority groups may be partly due to their greater susceptibility to the adverse neighbourhood environment. Primary
prevention measures targeting these neighbourhood attributes may have a major impact in reducing high BP related morbidity and mortality in the Netherlands.




SM2-PD-13                  REGIONAL VARIATIONS IN EXPOSURE TO PCBS, ORGANOCHLORINE PESTICIDES,
                           AND METALS IN THE UNITED STATES
L PHILLIPS1, K MAHAFFEY1, R CLICKNER2, R JEFFRIES2, Y AOKI3
(1) U.S. Environmental Protection Agency, Washington, DC. (2) Westat, Rockville, MD. (3) Association of Schools of Public Health, Washington, DC.
INTRODUCTION
Environmental contamination may vary geographically in the United States as a result of historic regional differences in the production and use of chemical substances. These differences,
coupled with differences in the fate and transport of chemicals, may result in regional variations in the levels of human exposure. This study used data on serum concentrations of environmental
contaminants collected as part of the National Health and Nutrition Examination Survey (NHANES) to evaluate regional variations in exposure.
METHODS
Serum concentrations of polychlorinated biphenyls (PCBs) and organochlorine pesticides from approximately 4,200 participants of NHANES during the years 1999-2002, and metals from
approximately 16,900 participants for lead and 5,200 for mercury, were used in this analysis. These data were categorized according to the geographic regions reported for the participants
(four inland regions: the northeast, south, midwest, and west; and three coastal regions: the Atlantic, Pacific, and the Gulf of Mexico. The data were further categorized according to age,
gender, and race/ethnicity. Descriptive statistics (e.g., mean, standard error, percentile distribution) were established for each category, and rank order analyses were performed to evaluate
exposure to multiple chemicals within each class (i.e., PCBs, pesticides, and metals), and to all chemical classes combined.
RESULTS
The results of this analysis indicate that NHANES participants living in the inland south had the highest exposures to organochlorine pesticides while those living in the inland northeast had
the highest exposures to PCBs. For example, the inland south had the highest geometric mean concentration of the pesticide transnonachlor (23.74 ng/g; 95% CI = 21.426.4 ng/g), while the
inland northeast had the highest geometric mean concentration of PCB 118 (10.77 ng/g; 95% CI 9.512.2). Participants in the inland northeast also had the highest exposures to lead, while
those living in the Atlantic coast region had the highest exposure to mercury.
DISCUSSION
This analysis indicates geographic differences in human exposure to certain classes of contaminants. Differences in exposure according to age and race/ethnicity were also observed. It should
be noted that the rank order method used in this study provides an index of relative human exposure and does not provide for differences in the toxicity of these compounds and thus does
not provide an indication of relative risk to human health. The findings and conclusions in this presentation have not been formally disseminated by the U.S. Environmental Protection Agency
and should not be construed to represent any Agency determination of policy.




SM2-PD-14                  SOCIOECONOMIC DISPARITIES IN ENVIRONMENTAL EXPOSURES AND
                           CHILDREN’S HEALTH: RESULTS OF A LITERATURE REVIEW IN THE
                           FRAMEWORK OF THE EU-FUNDED NETWORK PINCHE
M KOHLHUBER1, J HEINRICH2, G BOLTE1
(1) Department of Environmental Health, Bavarian Health and Food Safety Authority, Oberschleissheim. (2) Institute of Epidemiology, GSF National Research Center for
Environment and Health, Neuherberg.
Introduction: The aims of workpackage 5 within the Policy Interpretation Network on Children’s Health and Environment (PINCHE) were to review and evaluate the current evidence of the
impact of socioeconomic factors on children’s health and environment in Europe in the four selected themes air pollution, carcinogens, neurotoxicants and noise. The aim of the network was
to develop policy recommendations on local, regional, national and EU level.
Methods: The workpackage conducted a strategic literature search and review on socioeconomic disparities in children’s health and environment and evaluated the literature identified by
the PINCHE network. Also reports, books and grey literature were considered. Additionally a survey of studies on children’s health and environment funded by the EU framework programmes
4 and 5 was conducted.
Results: Overall there was a lack of information and data in reviews, original studies and other sources regarding the impact of socioeconomic factors on children’s health and environmental
exposure in Europe. Only a small percentage of the literature found in the PINCHE network considered socioeconomic factors. Whereas in the theme air pollution there is an increasing number
of studies which take social disparities into account, in studies of childhood cancer hardly any reviews or original studies were found. Especially in Eastern European countries there is still a
lack of information. The scarce data showed that there is an inverse social gradient: the burden of environmental exposures and environmentally related diseases is increased in children with
lower social status. Possible exceptions are atopic diseases which are more prevalent in children from families with higher social status and PCB exposure because of breastfeeding which is
more common in mothers with higher social status. The survey of EU-funded studies showed that the most common indicator to assess children’s social status is parental education which is
considered most often as confounder in data analysis. However it became clear that awareness increased over the last years to consider social status also as independent variable or effect
modifier in the analyses.
Discussion and conclusions: There is a need to improve the research of social disparities by inclusion of several indicators of socioeconomic position in data collection and by analysing
socioeconomic factors as influencing variable as well as effect modifier. Without a more detailed knowledge of socioeconomic disparities in children’s health and environment, it will not be
possible to develop sound specific and targeted measures for environmental health promotion.




                                                                                               57
          Symposia Session:

SMA-PD                          AIR POLLUTION AND REPRODUCTION, HEALTH OF CHILDREN AND OTHER
                                VULNERABLE GROUPS
SM4-PD-01                       PERIOD OF HEIGHTENED VULNERABILITY TO AIRBORNE POLYCYCLIC
                                AROMATIC HYDROCARBONS DURING GESTATION
H CHOI1, J SPENGLER3, WY TSAI5, P KINNEY6, W JEDRYCHOWSKI2, YH TU5, A PENAR2, D CAMANN4, E MROZ2, E MASTERS1, F PERERA1
(1) Columbia Center for Children’s Environmental Health, Mailman School of Public Health, Columbia University, New York. (2) Epidemiology and Preventive Medicine, Jagiellonian University, Krakow. (3)
Department of Environmental Health, Harvard School of Public Health, Boston. (4) Department of Analytical and Environmental Chemistry, Southwest Research Institute, San Antonio. (5) Department of of
Biostatistics, Mailman School of Public Health, Columbia Univeristy, New York. (6) Department of of Environmental Health Sciences, Mailman School of Public Health, Columbia Univeristy, New York.
Introduction: Conflicting epidemiologic evidence exists regarding vulnerable gestational period to air pollutants. Although little is known about period of greater vulnerability to polycyclic aromatic hydrocarbon (PAH)
exposure on birth weight, animals studies suggest first trimester, or earlier half of the gestation as the most vulnerable period to ETS and/or benzo[a]pyrene exposure. Therefore, hypothesis that high exposure to PAHs during
first trimester is associated with a larger decrement in fetal growth was tested.
Methods: Healthy, non-smoking pregnant women between 18 – 35 years of age, who have no current occupational exposure to PAHs were recruited during 2nd trimester from prenatal healthcare clinics. Personal cumulative
inhalation exposure to nine PAHs was measured by placing a portable air monitor in a backpack, which was worn for a consecutive 48-hour period by the women. To assess other source of PAHs exposure, the women were
also given a health, lifestyle, and environmental exposure interview. Upon delivery, the mother’s and infant’s medical record data were collected. Newborn’s total gestational exposure was estimated with generalized estimating
equation based on a single PAH measurement of the entire cohort and three measurements among a subset. Predicted gestational exposure was defined as either a linear function of mean exposure levels from each gestational
month, or a mean over each trimester. Risk of high exposure in given gestational month or trimester was determined by developing a linear regression models of BW, BL and BHC, respectively.
Results: Fitness of total gestational exposure model to the current data was examined with mean deviance score. The mean deviance of the chosen model was much closer to 1 than the initial model, suggesting a fairly good
fit. Furthermore, compared to the birth outcome models based on the actual PAH measurements, fits of the models based on estimated exposure were very similar. For birth weight and birth length, the predicted PAH exposure
during first trimester had the most adverse effect, controlling for pre-pregnancy weight, maternal height, gestational age, gender and parity. For birth head circumference, the predicted PAH exposure during first and second
trimester yielded comparable sizes of adverse effect, controlling for the same set of confounders. Considering high correlation between ambient PAH level and season, prenatal PAH exposure at given gestational month was
alternatively defined using season during each gestational month as a surrogate for PAH exposure level. Enduring winter during the first three gestational months had most adverse effect on birth weight and birth length.
Specifically, effect size was largest for those whose first gestational month coincided with December to February.
Conclusions: Current results lend support to the hypothesis that fetus is most susceptible to airborne PAHs during first trimester.




SM4-PD-02                       ADVERSE BIRTH OUTCOMES AND AMBIENT AIR POLLUTION
L DARROW1, M KLEIN1, M MARCUS1, D FLANDERS1, A CORREA2, E KAHN3, M STRICKLAND1, J MULHOLLAND4, P TOLBERT1
(1) Rollins School of Public Health, Emory University, Atlanta. (2) National Center on Birth Defects and Developmental Disabilities, U.S. Centers for Disease Control and
Prevention, Atlanta. (3) Division of Public Health, Georgia Department of Human Resources, Atlanta. (4) Georgia Institute of Technology, Atlanta.
Introduction: A growing body of evidence suggests that the developing fetus may be particularly susceptible to adverse effects of urban air pollution. To further investigate this hypothesis,
we are examining the occurrence of premature birth, intrauterine growth restriction (IUGR) and low birth weight in relation to ambient air pollution levels using time-series and spatio-temporal
analyses.
Methods: Data from all births and fetal deaths occurring between 1994 and 2004 in a major U.S. metropolitan area were obtained from vital records. Preterm birth was defined as spontaneous
delivery before 37 weeks of gestation; we estimated gestational age using the birth date and the date of the mother’s last menstrual period. Infants whose birth weight fell into the bottom
10 percent for their sex and gestational age were classified as intrauterine growth-restricted. Crude birth weight, which is determined by both intrauterine growth rate and length of gestation,
was also assessed. In ongoing analyses, the occurrence of preterm birth is being examined in relation to air pollution levels during two windows of a priori interest based on previous reports:
the final six weeks of pregnancy and the final week of pregnancy. IUGR and birth weight are being examined in relation to pollution levels during the third trimester. In secondary analyses,
we are examining alternative exposure windows during gestation for each outcome. Pollution exposures are being characterized using speciated particulate matter monitoring data as well
as gaseous measurements. Outcomes are being analyzed as aggregated daily counts in generalized estimating equations, in individual-level logistic regression models and, in the case of crude
birth weight, in linear regression models. Depending on the model, results are adjusted for individual-level covariates, temporal trends, and meteorology.
Results: The assembled cohort consists of approximately 770,000 births. Prevalence of preterm delivery over the study period is 11.6%. Approximately 8.5% percent of these infants have a
birth weight below 2500 grams. Results from the ongoing time-series modeling will be presented.
Discussion: Strengths of this study include the large study size and the availability of speciated PM monitoring data. Results will address current uncertainties regarding the relationship of
air pollution and adverse birth outcomes, including the gestational windows of susceptibility and the specific pollutants responsible for observed associations.




SM4-PD-03                       A PROSPECTIVE STUDY OF PERICONCEPTIONAL AMBIENT AIR POLLUTANT
                                EXPOSURES AND PREECLAMPSIA RISK
C RUDRA1, M WILLIAMS1
(1) CENTER FOR PERINATAL STUDIES, SWEDISH MEDICAL CENTER, SEATTLE.
Introduction: Preeclampsia, the onset of hypertension and proteinuria during pregnancy, is a common disorder that is epidemiologically and physiologically similar to atherosclerotic
cardiovascular disease. No prior studies have directly examined air pollutant exposures in relation to preeclampsia, to our knowledge. We measured the associations between periconceptional
fine particulate matter (PM2.5) and carbon monoxide (CO) exposures and subsequent preeclampsia risk.
Methods: We developed linear regression models to estimate monthly average ambient PM2.5 and CO concentrations using traffic, weather, and population characteristics as predictor
variables. We fit these models using data measured at regional air monitoring stations. We applied the model coefficients to estimate residential concentrations during the seven months
surrounding conception among 1,699 participants of a prospective cohort study who enrolled between 1996 and 2002. We defined preeclampsia according to criteria established by the
American College of Obstetricians and Gynecologists. We used multivariable logistic regression models to estimate associations between periconceptional air pollutant concentrations and
preeclampsia risk.
Results: The peak monthly CO concentration within the seven months surrounding conception was non-significantly associated with preeclampsia risk. Second and 3rd tertile adjusted risk
ratios (RR) and 95% confidence intervals (CI) were 1.26 (0.67-2.37) and 1.73 (0.91-3.27), respectively, versus 1st tertile peak CO concentration. The average CO concentration during the month
of conception and the following 3 months (3rd tertile RR 1.49, CI 0.76-2.90) was more strongly related than the average concentration during the 3 months before conception (3rd tertile RR
1.02,CI 0.54-1.93). PM2.5 exposures were not strongly related to preeclampsia risk (3rd tertile RRs 0.88-1.11).
Discussion and Conclusions: Though not conclusive, this study is the first to suggest that ambient CO exposure during the months surrounding conception may increase preeclampsia risk.
These results should be confirmed in future studies.




                                                                                                              58
                                                                                                                                                                                                   SUNDAY SEPTEMBER 3
        Symposia Session:

SM4-PD-04                  ASSOCIATION BETWEEN AMBIENT AIR POLLUTION AND BIRTH WEIGHT IN
                           NOVOCHEBOKSARSK, RUSSIA.
D DIMITRIEV1, A DIMITRIEV1, Y KONSTANTINOVA1
(1) Chuvash state pedagogic university of name I.Ja.Jakovlev, Cheboksary.
Introduction. We have evaluated the relationship between ambient air pollution and the occurrence of low birth weight using routinely collected data in Novocheboksarsk, Chuvash Republic,
Russia .
Methods. We used an epidemiologic case-control study design to examine whether exposure of ambient air pollutants to influenced the occurrence of low birth weight (LBW). The study
comprised 120 term LBW infants and 240 term infants weighing 2,500 g or more; date of mother’s last menstrual period was used to define gestational age. Matching variables were parity,
mother’s height, age, education and ethnic, infant’s gender. Birth data and information on maternal characteristics were obtained from the medical documentations (“History of delivery”,
“History of newborn”). To estimate trimester exposures, we used measurements of ambient nitrogen dioxide (NO2), phenol, sulfur dioxide (SO2), ammonia (NH3), chlorine ( Cl ), which were
collected at 3 monitoring posts (the mean distance between study subject’ home and monitoring post was approximately 2,5 km, and although 90% of subjects lived within 3 km, the
maximum distance was 4,5 km). The exposure variables were categorized into percentiles of the exposure distribution: < 25th, 25 to < 75th, >= 75th. For the categorical analysis the group
of infants with exposures under the 25th was used as the reference category. We calculated OR across trimester exposure categories for each pollutant.
Results. Our results suggest that exposures to ambient phenol, chlorine and ammonia increase the risk for term LBW. Infants with phenol second trimester exposure falling within the 25th
(0.005 mg/m?) and < 75th (0.008 mg/m? ) (OR 1.85; CI 1.16, 2.98), >= 75th (OR 3.26; CI 1.59, 6.60) percentiles were at increased risk for term LBW when compared to those in the reference
category (< 25th percentile). Exposures to atmospheric chlorine above the 75th percentile (0.06 mg/m?; 25th – 0.03 mg/m?) at second trimester yielded an odds ratio of 1.98 (CI: 0.99, 3.8).
Third trimester exposures to NH3 were associated with an increased risk for LBW among infants with exposures greater than or equal to the 75th percentile of the exposure distribution (0.042
mg/m?; 25th – 0.021 mg/m?) (OR 2.36; CI 1.1, 4.06).
Discussion and conclusion. Although the effects of unmeasured risk factors (smoking of mother, occupational exposure), could not be excluded with certainty, our findings suggest that in
Novocheboksarsk there may be a relationship between maternal exposure to ambient phenol, chlorine and ammonia and the risk of LBW.




SM4-PD-05                  TRAFFIC RELATED AND SOCIOECONOMIC INDICATORS IN ASSOCIATION WITH
                           LOW BIRTH WEIGHT AND PRETERM BIRTHS IN EASTERN MASSACHUSETTS
                           BETWEEN 1996 AND 2002.
A ZEKA1, J SCHWARTZ1, SJ MELLY1
(1) Exposure, Epidemiology and Risk Program, Department of Environmental Health, Harvard School of Public Health, Boston.
Introduction: Previous evidence has suggested that physical and social characteristics of places are important factors in determining human health. These characteristics may define gradients
of exposure, and as result, they may define gradients of population health outcomes. Exposures to air pollution have been shown to affect indicators of health, for example increasing risk of
adverse birth outcomes. The present work was carried out with the main goal to study the association between indicators of city-specific gradients of traffic and socioeconomic measures and
low birth weight and preterm births in Eastern Massachusetts.
Methods: We obtained data on birth outcomes for seven counties of Eastern Massachusetts from the Massachusetts Department of Public Health (MDPH) for years 1996-2002. Data on
socioeconomic measures were available at the census-, block group- (received from the US Bureau of Census), and individual-level (obtained from MDPH). Traffic count, distance to major,
secondary and tertiary roadways, percent of land use in each tract as recreational and conservation area were also available and were assigned to each subject based on their tract and address
location. Effects of socioeconomic indicators, traffic, and land use on low birth weight, preterm births, and “small for gestational age” were estimated through mixed effects models. We
controlled for conditions of the mother during pregnancy, mother age, race, gender of infant, amount smoked during pregnancy, previous preterm births or previous infants weighting more
than 4 000gr, prenatal care index measured by Kotelchuck Index, and level of education of mother – as a measure of socioeconomic level.
Results: The study included 359 875 births for the seven year period. Analyses for birth weight were restricted only among infants born between 37 and 45 weeks of gestation. Analyses for
Gaussian distribution of this outcome showed an important positive association with distance to major roadways and percent land use for recreation and conservation, and an inverse
association with cumulative traffic density to the nearby roads. Lower median income and greater proportion below poverty at the census tract level were associated with lower birth weight.
Preterm births were defined as those of less than 37 weeks of gestation. “Small for gestational age” was defined as the lowest decile by gestational week, mother race, and infant gender.
Models for these binary outcomes showed similar findings with those for birth weight.
Conclusions: Differences in physical and social characteristics of the places that may infer differences in exposures to air pollutants, may predict gradients of birth outcomes.




SM4-PD-06                  CORD BLOOD LYMPHOCYTES AND SHORT TERM EXPOSURE TO NITROGEN
                           OXIDES (NO2, NO, NOX), FINE PARTICLES (PM2.5), AND POLYCYCLIC
                           AROMATIC HYDROCARBONS (PAH) BEFORE DELIVERY
CEW HERR1, I HERTZ-PICCIOTTO1, M DOSTAL2, P ASHWOOD3, M LIPSETT4, PS YAP1, JP JOAD5, RH SHUMWAY6, I BENES2, R SRAM2, CEW HERR7
(1) DEPARTMENT OF PUBLIC HEALTH SCIENCES, UNIVERSITY OF CALIFORNIA, DAVIS. (2) LABORATORY OF GENETIC ECOTOXICOLOGY, INSTITUTE OF EXPERIMENTAL MEDICINE,
AS CR AND HEALTH INSTITUTE OF CENTRAL BOHEMIA, PRAGUE. (3) DEPARTMENT OF RHEUMATOLOGY, ALLERGY AND CLINICAL IMMUNOLOGY UNIVERSITY OF CALIFORNIA,
DAVIS. (4) DEPARTMENT OF EPIDEMIOLOGY AND BIOSTATISTICS, UNIVERSITY OF CALIFORNIA, SAN FRANCISCO. (5) DEPARTMENT OF PEDIATRICS, UNIVERSITY OF CALIFORNIA,
DAVIS. (6) DEPARTMENT OF STATISTICS, UNIVERSITY OF CALIFORNIA, DAVIS. (7) INSTITUTE FOR HYGIENE AND ENVIRONMENTAL MEDICINE UNIVERSITY OF GIESSEN, GIESSEN.
Background: The mechanisms by which ambient air pollutants influence health may include immunomodulation. We previously demonstrated differences in the distribution of cord
lymphocyte immunophenotypes with chronic and short-term exposure to PM2.5 and PAH in mothers infant pairs in the Czech Republic. Others have observed acute effects, i.e. a decrease in
absolute numbers of T-lymphocytes in peripheral blood, following nitrogen dioxide exposure in chamber studies.
Objective: To determine whether maternal exposure to outdoor nitrogen oxide pollution in the two weeks prior to delivery is related to systemic immunological alterations in the fetus, as
measured in cord blood.
Method: Cord blood samples were analyzed by flow cytometer (FACSort) to determine immunophenotypes (T-lymphocytes: CD3+, CD4+ subsets, CD8+, B-lymphocytes: CD19+, and natural
killer cells: CD16+CD56+) for 1,397 births in the Czech Republic between 1994 and 1998. Daily 24-hr average concentrations of nitrogen oxides (nitrogen oxide (NO), nitrogen dioxide (NO2)
and total nitrogen oxides (NOx)), fine particulate matter (PM2.5) and polycyclic aromatic hydrocarbons (PAHs) were measured at two centrally located monitors. Missing data were imputed
using standard algorithms. For each pollutant, we evaluated the 14-day average ambient concentration preceding the date of birth in relation to cord blood lymphocyte proportions, using
multiple linear regression models adjusted for potential confounders. Information on covariates was collected by interview with the mother.
Results: All pairwise Spearman correlation coefficients ranged from 0.51 to 0.66. For a 30 µg/m3 (corresponding to two standard deviations) increase in average NO2 exposure in the 14 days
before birth the fraction of NK cells decreased by -2.70% [CI (Confidence Interval) -5.29,-0.06] when adjusting for 3 day-average temperature before delivery. Models adjusting for 45 day-
average temperature yielded greater estimates: -4.38% [CI -7.39, -1.38]. In models adjusting for season of birth as well as 3 day-average temperature the decrease in NK cell fractions was
slightly smaller -2.30% [CI -5.12, -0.52]. Other lymphocyte subsets changed only negligibly. Models addressing NO and NOx exposure showed similar results
Conclusion:These results suggest that exposure to nitrogen oxide pollution shortly before birth may affect fetal production of NK cells relative to other lymphocytes. The relative contribution
of nitrogen oxides vs. other seasonal factors requires further investigation. In other work, we found PAHs and PM2.5 were associated with decreases in T-lymphocytes and increases in B-
lymphocytes, but not with changes in NK cell fractions, suggesting that the present findings are not due to confounding by co-pollutants. Possible interactions cannot be ruled out.


                                                                                              59
         Symposia Session:

SM4-PD-07                   HEALTH IMPACT ASSESSMENT OF PM10 AIR POLLUTION ON POSTNEONATAL
                            MORTALITY IN 31 EUROPEAN CITIES
K CAMBRA1, E ALONSO1, S MEDINA2, A LE TERTRE2, S CASSADOU2, L PASCAL2, F FRANK2, E DIAZ DE QUIJANO3, N VALERO3, L KNUDSEN4, L LOPEZ-
CARRASCO5, M SCHUEMANN6, R ATKINSON7, F GARCIA8, H MOSHAMMER9, S TORO10
(1) Department of Health, Basque Government, Vitoria-Gasteiz. (2) InVS- Institute for Public Health Surveillance, Saint Maurice-Cedex. (3) Public Health Agency of
Barcelona, Barcelona. (4) University of Copenhagen, Copenhagen. (5) Regional Ministry of Health, Madrid. (6) Department of Health and Science, Hamburg. (7) St
George’s Hospital Medical School, London. (8) Valencian School of Public Health Studies, Valencia. (9) Institute of Environmental Health, University of Vienna, Vienna.
(10) Andalusian School of Public Health, Granada.
Introduction: Health consequences of air pollution (AP) exposure are not equally spread among the population, and some groups, like children, may be at particular risk. Here we present the
results of the health impact assessment (HIA) of suspended particles of less than 10 µm (PM10) on postneonatal mortality, carried out within the Environment and Health Information System
(ENHIS) programme in 31 cities of the Apheis network.
Methods: Exposure-Response Functions for HIA were selected according to the following criterion: 1) Summary estimates from meta-analysis or original studies involving large populations;
and 2) outcomes for which the overall evidence of a causal contribution of air pollution is high. The selected RR for 10 µg/m3 increase of PM10 were 1.048 for total- and 1.216 for respiratory
postneonatal mortality (Lacasaña, 2005) and 1.12 for Postneonatal Sudden Infant Death Syndrome (PSIDS) (Woodruff, 1997)
We used PM10 data from 2001 or 2002. Three scenarios based on the annual mean of PM10 were defined: a reduction of 5 µg/m3 and the reduction to 40 and 20 µg/m3, both limits of 1999/30/EC
Directive. Each city used its own mortality rates. Attributable cases were estimated using the software created by the French Air Pollution and Health Surveillance Programme (PSAS-9).
Results: PM10 annual mean was between 20 and 40 µg/m3 in most of the participating cities. All other things being equal, in these cities (totalling almost 45 million inhabitants), a reduction
of the annual mean value of PM10 by 5 µg/m3 would be associated with a decrease of 23 postneonatal deaths per year (4.7/100000 children), 5 from respiratory causes (1.4/100000 children)
and 7 from PSIDS (1.8/100000 children). The accomplishment of the limit for 2010 of 1999/30/EC Directive (20 µg/m3 ) would prevent 56 postneonatal deaths, 13 from respiratory causes and
9 from PSIDS.
Discussion and Conclusions: These results add more evidence that AP continues to pose a significant threat to public health in urban areas in Europe, particularly for children. Nevertheless,
postneonatal mortality does not seem to be a sensitive indicator to assess the impact of PM on children’s health in Western cities, where mortality rates are very low. The main obstacle to a
more comprehensive HIA of outdoor AP on children is the lack of routine morbidity data.
ACKNOWLEDGEMENTS: ENHIS is supported by the European Commission DG SANCO (Grant: SPC 2003112). APHEIS was supported by the European Commission DG SANCO (contracts:
SI2.131174 [99CVF2-604]; SI2.297300 [2000CVG2-607]; SI2.326507 [2001CVG2-602] and the participating local institutions.



SM4-PD-08                   PM2.5 EXPOSURE AND RISK OF HOSPITALIZATION FOR INFANT BRONCHIOLITIS
C KARR1, KA MILLER1, T LARSON1, JQ KOENIG1
(1) University of Washington, Seattle.
Introduction
We recently reported a positive association between lifetime PM 2.5 exposure and risk of infant bronchiolitis in a high pollution region of the U.S. We sought to further examine this question,
as well as differing exposure windows, in a U.S. population with moderate particulate matter pollution as part of the Border Air Quality Study.
Methods
Subjects were derived from linked birth record-hospital discharge records of infants born 1997-2002. Cases (a hospital discharge record for bronchiolitis in infancy) were randomly matched
to ten controls without this diagnosis based on date of birth, length of birth hospitalization, and gestational-age. Short term (7, 14 days), subchronic (30, 60 days) and lifetime mean exposures
referenced prior to the case diagnosis date were assessed based on regulatory network monitoring data (proximal monitor within 20 km). Risk estimates for an increase in PM2.5 of 10 mg/m3
and quartiles of exposure were estimated using conditional logistic regression.
Results
2,770 case infants and 24,009 controls were assigned exposure. Their mean (SD) distance to a PM2.5 monitor was 8.1 km (4.7). The 7,14, 30, and 60 day and lifetime mean (SD) concentrations
of PM2.5 were 11.9 (5.3), 12.1 (4.5), 12.3 (4.0), 12.8 (3.8) and 12.3 (3.3) µg/m3, respectively. We found no statistically significant increased risk for infant bronchiolitis with any of the exposure
windows investigated, after adjustment for confounders (sex, mother smoking during pregnancy, mother’s education, parity, insurance and benefits type, infant race/ethnicity). Risk estimates
were highest for lifetime exposure (OR; 95% CI per 10 •g/m3 increase, 1.07; 0.83-1.36). This estimate increased with restriction to subjects with monitors within 5 km (1.25; 0.72-2.17) or to
subjects with a diagnosis of bronchiolitis due to respiratory syncytial virus (RSV), specifically (1.17; 0.88-1.56).
Discussion and conclusions
For North American infants, bronchiolitis is the leading cause of hospitalization and RSV is the most important etiologic agent. In contrast to our findings in a high pollution region, this study
did not find a statistically significant association with chronic PM2.5 exposure. Discrepancies may reflect differences in concentrations and composition of particulate matter and monitoring
network density in the two regions. Sensitivity analyses based on refined exposure assessment (closer monitors) and case definition (RSV) were consistent with previous findings of higher risk
estimates for chronic exposure windows compared to shorter term exposures. Our next steps include using spatial modeling of traffic and woodsmoke exposure to improve the assessment of
chronic particulate air pollution in this study population.




SM4-PD-09                   THE RELATIONSHIP BETWEEN OUTDOOR AIR QUALITY AND RESPIRATORY
                            SYMPTOMS IN YOUNG CHILDREN
C RODRIGUEZ1, A HINWOOD1, R TONKIN1, J HEYWORTH2, M KUSEL3, N DE KLERK3, P SLY3, P FRANKLIN2, T RUNNION4, A BLOCKLEY4, L LANDAU5
(1) Centre for Ecosystem Management, Edith Cowan University, Perth. (2) School of Population Health, The University of Western Australia, Perth. (3) Telethon Institute
for Child Health Research and Centre for Child Health Research, The University of Western Australia, Perth. (4) Department of Environment, Perth. (5) Faculty of Medicine,
Dentistry and Health Sciences, Perth.
Introduction: Few studies have addressed the relationship between outdoor air pollution and risk of acute respiratory illness in young children and those that have are in areas with high
levels of outdoor air pollutants. As air pollution is decreasing in many regions of the world it is important to understand whether low levels of pollution can still have adverse effects. In this
study we investigated the relationship between exposure to outdoor air pollution and symptoms associated with respiratory illness (SARI) in a cohort of children at higher risk of the adverse
effects of pollutants due to their young age (0-5years) and susceptibility to atopic disease.
Methods: Two hundred and sixty three children at high risk of developing asthma or atopy were recruited antenatally and followed until 5 years of age. During this period all respiratory
symptoms experienced by the children were recorded by their parents and daily concentrations of ozone (O3), carbon monoxide (CO), nitrogen dioxide (NO2), particle matter (PM2.5), visibility
(BSP) and meteorological data (ambient temperature and humidity) were collected from network monitoring sites. Logistic regression models were used to investigate associations between
individual air pollutants and respiratory symptoms with lags of 0 and 5 days as well as additive exposure over 5 days.
Results: The relationships between O3 (1hr and 8hr) concentrations and raised body temperature during acute respiratory illnesses were significant for a lag of 0 days. CO (8hr) was significantly
associated with wheeze/rattle and runny/blocked nose for both a lag of 5 days and additive exposure over 5 days. A significant association was also observed between NO2 (24hr)
concentrations and cough for both lag 0 and additive exposure over 5 days. Both PM2.5 and BSP (1hr) were also significantly associated with cough (lag 0 days). However, air pollutant
concentrations were generally below national standards throughout the study period.
Conclusions: This study suggests that even with air pollution concentrations below Australian standards there is a relationship between air pollutants and symptoms associated with
respiratory illnesses in children in Perth.




                                                                                                 60
                                                                                                                                                                                                        SUNDAY SEPTEMBER 3
         Symposia Session:

SM4-PD-10                   PARENTAL STRESS, TRAFFIC, AND NEW ONSET ASTHMA IN A COHORT OF
                            CHILDREN
R MCCONNELL1, K BERHANE1, J MILAM1, L YAO1, F LURMANN2, J RICHARDSON1, M JERRETT1
(1) University of Southern California, Los Angeles. (2) Sonoma Technology, Inc, Sonoma.
Recent literature suggests that both traffic related pollutant exposure and psychosocial stress are risk factors for childhood asthma. However, there has been little study of synergistic effects
or of associations of these risk factors with new onset asthma in prospective studies. We examined the relationship of local traffic-related exposure and of parental stress to the risk of incident
asthma in the southern California Children’s Health Study. A cohort of 5 – 7 year old children was recruited in 2003, and a parent of each child completed a questionnaire on the child’s history
and symptoms of asthma and relevant risk factors. Parental stress at study entry was assessed using a well-validated 4-question perceived stress scale. The child’s residential exposure was
assessed by modeling exposure to local traffic-related pollutants. A yearly parent-completed questionnaire identified children with new onset asthma through 2005. The relative risk of new
onset childhood asthma associated with parental stress was 1.19 over the inter-quartile range of stress in the study population (95% confidence interval 1.00, 1.41). There was no association
of incident asthma with traffic modeled pollution by itself, but there was a significant interaction of parental stress with the child’s residential traffic exposure. The relative risk of asthma
associated with the inter-quartile range of traffic-modeled pollutant exposure was 2.48 (1.37, 4.50) among children with parents in the upper quartile of perceived stress. The pattern of effects
was similar among children with no lifetime history of wheeze at study entry, indicating that the associations with incident asthma did not merely reflect exacerbation of pre-existing (but
undiagnosed) asthma. These results suggest that parental stress may make children more susceptible to traffic-related asthma.




SM4-PD-11                   SOCIOECONOMIC STATUS, PARTICULATE AIR POLLUTION AND DAILY
                            MORTALITY: DIFFERENTIAL EXPOSURE OR DIFFERENTIAL SUSCEPTIBILITY
F FORASTIERE1, M STAFOGGIA1, C TASCO2, S PICCIOTTO1, N AGABITI1, G CESARONI1, CA PERUCCI1
(1) Dept. Epidemiology, Local Health Authority, Rome. (2) Environmental Authority, Lazio Region, Rome.
Introduction: Short-term increases in particulate air pollution are linked with increased daily mortality and morbidity. Since socioeconomic status (SES) is a strong determinant of overall heath
status, we investigated whether social class is an effect modifier of the PM10-daily mortality association, and possible mechanisms for this effect modification.
Methods: Area-based traffic emissions, income and socioeconomic status (SES) were available for each resident in an Italian city. We selected all non-accidental deaths (83,253 subjects)
occurring among city residents (aged 35+ years) during the period 1998-2001. For each deceased individual, all the previous hospitalizations within 2 years before death were available via
a record linkage procedure. PM10 daily data were available from two urban monitoring sites. We adopted a case-crossover analysis where control days were selected according to the time
stratified approach (same day of the week during the same month). Conditional logistic regression was used.
Results: Due to the social class distribution in the city, exposure to traffic emissions was higher among those with higher area-based income and SES. Meanwhile, people of lower social class
had suffered to a larger extent from chronic diseases before death than more affluent residents, especially diabetes, hypertension, heart failure and COPD. Overall, PM10 (lag 0-1) was strongly
associated with non-accidental mortality (1.2% increase, 95%CI= 0.7-1.7%, per 10 µg/m3). The effect was more pronounced among persons with lower income and SES (1.9% and 1.6% per
10 µg/m3, respectively) compared to those in the upper income and SES levels (0.3% 0.2%, respectively).
Discussion and conclusions: The results confirm previous suggestions of a stronger effect of particulate air pollution among people in low social class. Given the uneven geographical
distributions of social deprivation and traffic emissions in this city, the most likely explanation is a differential burden of chronic health conditions conferring a greater susceptibility to less
advantaged people.




SM4-PD-12                   PARTICULATE AIR POLLUTION AND SHORT TERM MORTALITY FOLLOWING
                            ACUTE MYOCARDIAL INFARCTION
G WELLENIUS1, J SCHWARTZ2, M MITTLEMAN1
(1) Cardiovascular Epidemiology Research Unit, Beth Israel Deaconess Medical Center, Boston, MA. (2) Department of Environmental Health, Harvard School of Public
Health, Boston, MA.
Introduction: Short-term increases in respirable particulate matter (diameter <= 10µm, PM10) have been linked to increased risk of acute myocardial infarction (MI). PM10 levels preceding an
MI may also be associated with short-term mortality following the event. For example, PM10 may plausibly increase the risk of medical complications (e.g.: life-threatening cardiac arrhythmias)
and thereby increase the risk of short-term mortality. Thus, we evaluated the association between daily PM10 and the risk of short-term mortality (in- hospital death, death within 48 hours of
admission, death within 30 days of admission) among Medicare recipients (age >= 65 years) hospitalized from the Emergency Department for acute MI (ICD-9: 410) in 7 US cities between
1986 and 1999.
Methods: We performed logistic regression to separately estimate the effect of a 10 µg/m? increase in PM10 in Chicago, IL, Cleveland, OH, Detroit, MI, Minneapolis/St. Paul, MN, New Haven,
CT, Pittsburgh, PA and Seattle, WA. We considered PM10 on the admission day (lag 0) or the day prior to admission (lag 1) in separate models. In all models, we controlled for time trends and
apparent temperature (lags 0 and 1) using generalized additive models with penalized splines. Subsequently, we obtained a combined random-effects estimate from the city-specific effect
estimates.
Results: There were 188,914 hospital admissions from the Emergency Department with a primary discharge diagnosis of acute MI during the observation period. Of these, 16.4% died in-
hospital, 5.6% died within 48 hours of hospitalization, and 20.2% died within 30 days of hospitalization. A positive association between PM10 on the day before admission and each outcome
was observed in 5 of the 7 cities. Overall, a 10 µg/m? increase in PM10 at lag 1 was associated with a 1.2% (95% CI: 0.4, 2.0; p=0.0039) increase in risk of in-hospital mortality, a 0.9% (95%
CI: -0.4, 2.2; p=0.16) increase in the risk of death within 48 hours, and a 0.6% (95% CI:-0.5, 1.7; p=0.26) increase in the risk of death within 30 days. No effect was seen when PM10 on the
admission day was considered.
Discussion and Conclusions: These results suggest that ambient PM10 may be associated with a small increased risk of short-term mortality following acute MI in this population of elderly
Americans. Additional studies with more accurate data on the timing of MI onset and more detailed information on patient characteristics and medical course are needed to confirm or refute
these findings.




                                                                                                61
        Symposia Session:

SM4-PD-13                  HEALTH EFFECTS OF DIESEL EXHAUST IN ASTHMATIC PATIENTS: THE ROLE
                           OF SPECIFIC EXPOSURES IN A REAL WORLD STUDY
J MCCREANOR1, J ZHANG2, P CULLINAN1, F CHUNG1, J STEWART EVANS1, E MALLIAROU1, M SVARTENGREN3, P OHMAN-STRICKLAND2, L JARUP1, M
NIEUWENHUIJSEN1
(1) Imperial College, London. (2) University of Medicine and Dentistry of New Jersey, Piscataway. (3) Karolinska Institute, Stockholm.
Back ground and aims: Epidemiological evidence suggests a link between respiratory morbidity and urban fine particles, many of which are produced by diesel-powered vehicles. We studied
the health effects of diesel exhaust in asthmatic patients exposed within a ‘natural exposure chamber’.
Methods: On separate occasions non-smoking adult asthmatics were exposed, for 2 hours, in Oxford Street, London (where only pedestrians, diesel powered buses and taxicabs are permitted)
or Hyde Park, a large, nearby open park mostly free of vehicles. We measured PM2.5, elemental carbon, NO2 and ultrafine particles during each session. Lung function, oxidative stress and
inflammatory markers, and symptoms were monitored during exposures and for 24 hours after.
Results: Levels of air pollution was higher at Oxford Street compared to Hyde Park. Analyses showed that lung function (FEV1 and FVC) decreased from baseline at both exposure sites; this
decrement was largest and more sustained in Oxford Street. The decline in lung function showed the strongest relationship with levels of elemental carbon and ultrafine particles and, to a
lesser extent, NO2 and PM2.5. Some markers of inflammatory and oxidative stress showed similar trends, providing a possible explanation of the mechanism for the observed lung function
reductions.
This study suggests that asthmatic subjects exposed to ambient urban levels of diesel exhaust, in a natural environment, incur worsening respiratory function and that this is related to specific
combustion pollutants in the environment.




SM4-PD-14                  RELATIONSHIPS BETWEEN PERSONAL, INDOOR AND OUTDOOR LEVELS OF
                           SULFATE, ELEMENTAL CARBON AND PM2.5 FOR A COHORT OF INDIVIDUALS
                           WITH CARDIOPULMONARY DISEASE
KW BROWN1, JA SARNAT2, BA COULL1, HH SUH1, JD SPENGLER1, P KOUTRAKIS1
(1) HARVARD SCHOOL OF PUBLIC HEALTH, BOSTON. (2) EMORY UNIVERSITY, ATLANTA.
Background: Previous exposure studies have validated the use of ambient fine particulate matter (PM2.5) measurements as surrogates of exposure in PM time-series epidemiologic studies. It
is uncertain, however, whether ambient measurements of specific PM components are equally good surrogates. Panel exposure assessment studies provide a direct means of exploring the
strength of the relationships between personal, home indoor, home outdoor and ambient particulate levels. For this study, we conducted a panel study measuring personal, home indoor, home
outdoor and ambient site concentrations of PM2.5 as well as its components, sulfate and elemental carbon (EC), for individuals with cardiopulmonary disease. We examined particulate pollutant
associations among the microenvironments by season and subject.
Methods: Simultaneous 24-hour integrated personal, home indoor, and home outdoor PM2.5, SO42- and EC concentrations were measured in 25 single-family homes during the summer and
winter. All homes and subjects were measured for seven consecutive days. Integrated 24-hour PM2.5, sulfate and EC samples were also measured at a central monitoring site.
Results: A total of 210 person-days of exposure data were collected. We found ambient sulfate to be strongly correlated with personal and home indoor sulfate for all individuals without an
indoor source of sulfate in the home. Outdoor or ambient sulfate accounted for approximately 80% of the variability in personal and indoor sulfate concentrations, despite the fact that absolute
exposure to sulfate was at times considerably lower than outdoor levels. In contrast to the results for sulfate, personal/indoor and outdoor/ambient associations for EC and PM2.5 were weaker,
likely due to indoor and/or local sources of EC and PM2.5. There were relatively weak EC outdoor-ambient associations, especially during the summer (mean home-specific Spearman’s correlation
coefficient = 0.26), indicating a greater degree of spatial heterogeneity for this pollutant.
Discussion and Conclusion: These findings provide further evidence that ambient sulfate monitors may serve as good indicators of correlations, but not absolute exposures, to particles of
ambient origin for particle health effects studies. The spatial variability of EC was notable in this study, although our results did not show stronger personal-outdoor associations compared to
personal-ambient associations. These findings suggest that, for this metropolitan area, locating monitors closer to participants’ homes may not be effective when trying to assess personal EC
exposure in future studies. Our results also indicate that housing characteristics are likely the central driving factor determining the strength of the personal-outdoor or personal-ambient
relationships for PM2.5 and these two components.




SM4-PD-15                  ARRHYTHMOGENESIS INDUCED BY PARTICULATE AIR POLLUTANTS: ARE
                           OBESE PEOPLE MORE SENSITIVE?
JC CHEN1, PH STONE2, DC CHRISTIANI3
(1) Department of Epidemiology, University of North Carolina School of Public Health, Chapel Hill, NC. (2) Cardiovascular Division, Brigham and Women’s Hospital,
Harvard Medical School, Boston, MA. (3) Occupational Health Program, Department of Environmental Health, Harvard School of Public Health, Boston, MA.
Introduction:
Both epidemiologic and experimental studies have demonstrated the arrhythmogenesis induced by exposure to particulate air pollutants, especially PM2.5 (particulate matter with aerodynamic
diameter •2.5•m) and combustion-source particles. Emerging evidence has suggested that obesity may impart both toxicokinetic and toxicodynamic susceptibilities to the adverse health
effects of air pollution. Whether obese subjects are more sensitive to PM-associated arrhythmogenesis is unclear.
Methods:
We examined the PM2.5–mediated acute effects on supraventricular extrasystole using ten 24-h and thirteen 48-h ambulatory ECG recordings collected over 2 years from 18 male boilermakers
(aged 39.5±8.5) who were exposed to high-levels of metal particulates (4-h moving average PM2.5:1.00±0.17 mg/m3 during working hours). The presence of supraventricular extrasystole within
each 5-min epoch was identified through standardized algorithm. Personal PM2.5 exposures were monitored continuously in the workplace, community, and home environments. Subjects with
measured body mass index >30 kg/m3 were classified as obese. We employed GEE (generalized estimation equation) to model the odds ratio (OR) of having any supraventricular extrasystole
associated with PM, adjusting for temporal correlation (1st-order autoregressive), individual fixed effects, age, body mass index, baseline mean arterial blood pressure, workdays, and activities
that changed over the course of measurement (workdays, smoking, drinking alcohol, drinking coffee, eating, sleeping, exercising, and sleeping).
Results:
Half (50%) of the study subjects were obese. Subjects with obesity were more likely to have supraventricular extrasystoles (11% vs. 3% of all recordings). Among non-obese subjects, there
was no discernible effect of PM2.5 on supraventricular extrasystoles. Among obese subjects, each 1-mg/m3 increase in preceding 4h-moving average PM2.5 was associated with the presence of
supraventricular extrasystole (OR:1.98; 95% CI: 1.10, 3.56), which was statistically significantly (p=0.01) greater than the corresponding effects among non-obese subjects (OR:1.02; 95% CI:
0.62, 1.69).
Discussion and Conclusions:
We found that the induction of supraventricular extrasystole was associated with high levels of metal particulate exposure in obese men but not in those without obesity. Our study results
support the hypothesis that people with obesity are more sensitive to arrhythmogenic effects of particulate air pollutants.




                                                                                               62
                                                                                                                                                                                                          SUNDAY SEPTEMBER 3
         Symposia Session:

SM4-PD-16                   EFFECT MODIFIERS OF THE ASSOCIATION BETWEEN AIR POLLUTION AND
                            CARDIAC READMISSIONS AMONG SURVIVORS OF FIRST MYOCARDIAL
                            INFARCTION
S PICCIOTTO1, S VON KLOT2, F FORASTIERE1, A PETERS2, M STAFOGGIA1, CA PERUCCI1
(1) DEPARTMENT OF EPIDEMIOLOGY, LOCAL HEALTH AUTHORITY, ROME. (2) INSTITUTE OF EPIDEMIOLOGY, GSF (NATIONAL RESEARCH CENTER FOR ENVIRONMENT
AND HEALTH), NEUHERBERG.
Introduction: A recent study found that among survivors of a first myocardial infarction, elevated levels of ambient air pollution are associated with increased hospital admissions for cardiac
causes. Whether this association occurs uniformly in this group or is modified by other factors is unknown. We analyzed effect modification by gender, area-based socioeconomic status, and
chronic comorbidities in a cohort of myocardial infarction survivors with index event in the period 1998-2000 and follow-up through the end of 2004.
Methods: Hospitalisation data were obtained from the regional Hospital Discharge Registry. Persons (aged 35+ years) admitted to the hospital for a first acute myocardial infarction were
included in the cohort if they survived for at least 28 days (N=7382). They were followed up from the 29th day until death, first readmission for a cardiac cause (new myocardial infarction,
angina, dysrhythmia, or heart failure), or end of study. Vital status was ascertained and hospital readmissions identified via record linkage. Daily air pollution exposures were calculated as the
means from several monitors measuring PM10 and from single monitors for ozone (urban background level, daily maximum 8-hour average, only during summer months) and particle number
counts (representing ultrafine particles, estimated for 1998-March 2001 from other pollutants and meteorological measurements). Poisson models were applied to stratified datasets, adjusting
for trend, temperature, day of week, and holidays. Relative risks were calculated for an increase of one interquartile range of the respective pollutant (PM10 22.03ug/m3, ozone 39.11ug/m3,
ultrafines 28350/cm3) for various lags. Results reported are for the mean of lags 0 and 1.
Results: Among 5160 men and 2222 women who survived a first myocardial infarction, 1675 and 719 (respectively) were subsequently hospitalised for a cardiac cause. The overall relative
risk (RR) for ultrafine particles was 1.090 (95% confidence interval (CI)=1.005-1.181), for PM10: 1.039 (95%CI=0.987-1.094), and for ozone: 1.089 (95%CI=0.994-1.193). No evidence was
found to indicate that area-based socioeconomic status modified these effects. However, gender was an effect modifier (e.g. ultrafines, men’s RR=1.045, 95%CI=0.952-1.148, women’s
RR=1.196, 95%CI=1.058-1.352). There was convincing evidence that diabetes modified the effects of particles (e.g. ultrafines, RR=1.260 (95%CI=1.107-1.434) vs. RR=1.037 (95%CI=0.947-
1.136)) and that COPD modified the effect of ozone.
Discussion and Conclusions: Various characteristics of post-infarction patients may make them more susceptible to another hospital admission for cardiac causes. These characteristics
include gender and co-morbidity. In particular, diabetics experience a greater effect of particulate air pollution than other myocardial infarction survivors.




SM5-PD                      SPATIAL ANALYSIS IN ENVIRONMENTAL HEALTH
SM5-PD-01                   IMPROVING THE SPATIAL RESOLUTION IN EXPOSURE ASSESSMENT OF AIR
                            POLLUTION AND ROAD TRAFFIC NOISE FOR HEALTH EFFECT STUDIES
Y DE KLUIZENAAR1, CGM OUDSHOORN2, HME MIEDEMA1
(1) TNO, Delft. (2) Roessingh Research and Development, Enschede,.
Introduction
Living near a busy road may induce cardiovascular health effects. Studies into potential mechanisms and epidemiological studies indicate that cardiovascular health effects may be related to
both air pollution and environmental noise. However, uncertainty regarding exposure-effect relationships exists and the relative contribution of both factors is unknown. Accurate assessment
of long term air pollution and road traffic noise exposure, improving spatial resolution, is needed for epidemiological exposure-effect studies investigating these relationships.
Methods
We present an approach for refined long term exposure assessment of subjects in epidemiological studies to air pollution and noise, taking local spatial variations into account. This approach
is applied in a case study in Leiden, the Netherlands, to make a detailed assessment of spatial variation of air pollution and road traffic noise in a city. The correlation between air pollution
and road traffic noise is investigated. The impact of factors responsible for differences in the spatial variation of air pollution and noise is studied. Errors introduced at lower spatial resolution
in exposure assessment, and possible consequences for reliability and validity of estimated relative risks are investigated for this case.
Results
Spatial variation in air pollution and noise exposure was substantial. The difference between the lowest and highest 5% of dwellings exceeded 15 •g m-3 (NO2) and 25 dB(A) (Lnight). Although
air pollution (NO2) and road traffic noise (Lnight) are correlated, substantial variation in air pollution exposure at a given noise level was found. With lower spatial resolution, an error was
introduced in exposure estimates that appeared partly systematic and partly random. Lower levels seemed to be overestimated, while high exposure levels were underestimated. Systematic
and random error appeared to distort the relative risk estimate in a different direction: systematic error only seemed to lead to overestimation, while the random error only appeared to lead
to underestimation of the relative risk. Combined the two types of errors could distort the relative risk estimate in both directions, depending on the spatial resolution. Relative risk was
underestimated for both air pollution and road traffic noise at the lowest spatial resolution investigated here.
Discussion and Conclusions
This paper illustrates that improving spatial detail in exposure assessment, taking local spatial variation into account, may help improve reliability and validity of exposure-effect estimates
and may help elucidate relationships between long term air pollution and road traffic noise exposure, and (cardiovascular) health effects.



SM5-PD-02                   EVALUATION OF A LAND USE REGRESSION MODEL TO ASSESS EXPOSURE TO
                            AIR POLLUTION DURING PREGNANCY: USE OF GPS TRACKING AND PERSONAL
                            MONITORING
E NETHERY1, SE LECKIE1, JD MARSHALL1, M BRAUER1
(1) School of Occupational and Environmental Hygiene, University of British Columbia, Vancouver.
INTRODUCTION: The Border Air Quality Study (BAQS) includes a cohort study of the relationship between exposure to traffic-based air pollutants during pregnancy and adverse birth
outcomes. This cohort study (BAQS) estimates exposures using a land-use regression model and geo-coded home locations. To evaluate this approach, we compared measured and modeled
exposures for a sample of pregnant women.
METHODS: We measured and modeled 48-hour exposures (NO, NO2, NOX, and black carbon) for 31 pregnant, non-smoking women. Measured personal exposures (Ogawa passive samplers
and Personal Exposure Monitors [PEMs]) were compared with two modeled exposure estimates both based on locations (home and work) and a land-use regression model of ambient pollution
concentrations. The first estimate simply assumes subjects stayed at home all day; the second accounts for time at home and at work using time-varying location data from a datalogging GPS
receiver. In addition, since 6-digit postal codes, but not actual address locations, were available for the BAQS cohort study we also modeled exposures based on postal codes for each of the
subjects’ home and work locations.
RESULTS: Modeled outdoor exposures at home address locations were highly correlated with those based on postal code centroids for all pollutants [Spearman’s rho: 0.95 – 0.99]. Therefore,
we used postal code estimates for all subsequent analyses. For NO2, geometric mean measured exposures for each of 4 groups defined by quartiles of modeled home exposure were: 10, 19,
20, 22 ppb (significant differences between 1st quartile and 2nd, 3rd or 4th: Kruskal-Wallis p-value<.05). Similar group differences were seen for NOX and NO (K-W p<.05) but not for black carbon.
For NO2, there was a modest correlation (Spearman’s rho=0.43) between measured and modeled home exposures. Incorporating modeled work exposures slightly improved this correlation
(rho= 0.45). No such associations were observed for other pollutants (range rho: 0.03 to 0.08).
DISCUSSION: Modeling personal exposure using home postal codes was comparable to using exact home addresses. This is important for large cohort studies, since privacy regulations
typically restrict geocoding to the postal code level. For the pregnant women monitored in this study, modeled exposure quartiles were moderately predictive of high or low measured
exposures for NO, NO2 and NOX. These results indicate that exposure classification of this population based on postal code geocoding and land use regression models is appropriate for NO,
NO2 and NOX but not black carbon.




                                                                                                 63
        Symposia Session:

SM5-PD-03                  ZIP CODES OR STREET ADDRESSES? COMPARING AMBIENT OZONE
                           EXPOSURES FOR ALTERNATIVE SPATIAL RESOLUTIONS OF COHORT SUBJECTS
S SORET1, M GHAMSARY1, D SHAVLIK1, WL BEESON1, LH CHEN1, S WIAFE1, J PRESS2, S KNUTSEN1
(1) School of Public Health, Loma Linda University, Loma Linda. (2) University of California, Riverside.
Introduction: Accurate estimation of exposures is of the outmost importance for drawing valid inferences about the spatial relationship of risk factors with health outcomes of concern. In
air pollution epidemiology, estimates are often derived from monitoring data. With GIS technology, we can map individuals to their street address rather than to a central location by ZIP code.
Assuming the validity of the spatial interpolation method chosen and the representativeness of exposure data, an independent question remains: Does enhanced positional accuracy of subjects
result in reduced exposure misclassification? We compared the estimated ozone exposures assigned to our cohort subjects when their locations are resolved alternatively by ZIP-code centroids
and by street addresses.
Methods: Monitors with representative data for at least 9 months in three years or more from 1996 to 1999 were selected and the average annual concentration calculated. We then identified
cohort subjects with valid, geocoded residence information during 1996-1999 (n =584). Differences in exposure with respect to positional accuracy were assessed by implementing two spatial
interpolation methods: Inverse Distance Weighting (IDW) and Ordinary Kriging (OK). OK- and IDW-derived ozone concentrations were then assigned to each subject’s mapped ZIP centroid
and street address.
Results: Ozone estimates for both spatial resolutions were highly correlated regardless of the interpolation method used (RIDW = 0.978 and ROK = 0.987). The estimated kriging standard
errors were also highly correlated for both sets of predictions (R = 0.94). The average kriging standard error was virtually the same for predictions resolved at the street-level (0.0705 ppm)
and those for ZIPs (0.0700). The average difference between exposures assigned to ZIP and street addresses was positive but of small magnitude. For OK predictions, this value reached 0.0001
ppm (SD = 0.00096 ppm), while that for IDW estimates was even smaller, 0.00006 ppm (SD = 0.0012 ppm).
Discussion and conclusions: Refining the locational resolution of subjects did not result in a large change in estimated exposures. Both interpolation methods predicted greater exposures
for ZIP centroids. Kriging generated greater differences between predictions for ZIPs and street addresses. Thus, greater spatial accuracy is of less concern when we use IDW for estimating
ozone exposures. Using ZIP centroids seem to be a reasonable approach for modeling ozone exposures. This notion is yet to be corroborated for other types of small-area units. Our results
lend validity to previous findings from our study in which subject locations were mapped to ZIP centroids.

PLEASE DO NOT MODIFY THE TEMPLATE FRAME
otherwise the abstract will be rejected by the Programme Committee




SM5-PD-04                  ESTIMATION OF URBAN AIR POLLUTION EXPOSURE DURING PREGNANCY
                           USING A GIS-BASED METHODOLOGY
I AGUILERA1, M ALVAREZ1, C FERRER1, B JACQUEMIN1, J JULVEZ1, N RIBAS-FITÓ1, C VILLANUEVA1, J SUNYER1, G HOEK2, B BRUNEKREEF2
(1) Environmental and Respiratory Research Unit, Municipal Institut of Medical Research, Barcelona. (2) Institute for Risk Assessment Sciences, Universiteit Utrecht,
Utrecht.
Introduction: INMA (Environment and Childhood) Study is a collaborative research network that includes several birth cohorts in Spain to study the impact of environmental pollutants and
diet on children’s development and health. We aim to estimate individual exposure to air pollution during pregnancy and first years of life in the cohort of Sabadell (n=800), a city in the
metropolitan area of Barcelona.
Methods: NO2 was measured in two periods of 48-hours and 7-days during 2005. 57 measurement sites were selected to represent the gradient of exposure to air pollution in the study
population. NO2 was also measured simultaneously in 40 homes of pregnant women with outdoor passive samplers. All the women were on their third trimester of pregnancy.
Geographic data (e.g., street and buildings map, households and population density) and traffic data (e.g., street type, traffic intensity and bus routes) were collected from the city council
using GIS. The measuring sites and the cohort addresses were geocoded.
A regression model was conducted to identify the best geographic predictors of air pollution in the city. The model was achieved by linking the NO2 data to GIS variables (such as population
density, traffic intensity and length of roads) within 200, 500 and 1000 meter buffers around each measuring site. From this model we will predict outdoor NO2 concentrations at the home
addresses of all the cohort members. Outdoor NO2 data obtained in the 40 dwellings will be used to validate the model.
Results: NO2 mean levels for the two periods were 51.7 ?g/m3 for traffic sites and 32.3 ?g/m3 for background sites. The following table shows the preliminary linear regression model for NO2
(R2 = 0.70)
This preliminary model is being applied to predict NO2 levels at all the cohort
addresses in order to study the association between prenatal air pollution
exposure and reproductive effects.
Discussion and conclusions: Preliminary results suggest that a substantial
fraction of the spatial variability in NO2 is explained by traffic-related
variables. A GIS-based methodology is useful for individual exposure
assessment in epidemiologic studies, in comparison to traditional approaches
that use air monitoring data alone.




SM5-PD-05                  GIS-BASED RETROSPECTIVE ASSESSMENT OF TRAFFIC RELATED AIR
                           POLLUTION EXPOSURES, AND SOCIAL ENVIRONMENTAL INFLUENCES ON
                           CHILDHOOD ASTHMA
J CLOUGHERTY1, J LEVY1, PB RYAN2, R WRIGHT3
(1) Harvard School of Public Health, Boston, MA. (2) Emory University - Rollins School of Public Health, Atlanta, GA. (3) Channing Laboratory — Harvard Medical School,
Boston, MA.
Introduction: Asthma disproportionately impacts children in lower-income urban communities, where environmental exposures such as traffic-related air pollution and social stressors
including exposure to violence (ETV) are elevated. Stress may disrupt biological systems related to inflammation through mechanisms overlapping those altered by physical pollutants, making
those individuals dually exposed most susceptible to asthma. However, few studies have been able to incorporate both elements, in part because data and methods have not been available
for retrospective exposure assessment. For this study, we used GIS-based land use regression (LUR) techniques to retrospectively assign estimates of residential exposures to traffic-related air
pollutants, taking account of temporal variability in exposures, in order to examine the effect of multiple exposures on asthma outcomes, including interactions between social and physical
exposures.
Methods: Intra-community spatial and temporal variability in pollution concentrations was examined using monthly nitrogen dioxide (NO2) measurements for 13 sites over 19 years in East
Boston and surrounding neighborhoods. Temporally-specified LUR models were then developed to create residence-based and time-specific estimates of long-term exposure trajectories for a
community-based pregnancy cohort recruited between 1986 and 1992, taking account of residential history and the time course of disease development. Residence-based pollution estimates
were merged with individual-level questionnaire data on age, sex, race, socioeconomic position, tobacco smoke exposure, exposure to violence (ETV), and asthma/wheeze phenotypes
ascertained at cohort enrollment and follow-up. Using multivariate logistic regression, we examined the relative and multiplicative contributions of ETV and NO2 indicators on childhood
asthma/wheeze outcomes.
Results: LUR models indicated that NO2 concentrations were associated with distance to major road, kernel-weighted traffic density within 500 meters, and population density, with strong
temporal stability in spatial patterns. Preliminary analyses indicate elevated univariate odds ratios for diagnosis of bronchitis and wheeze only among children in the highest tertiles at follow-
up for both ETV and NO2 exposure (OR= 1.33 (1.01, 1.75) and 1.22 (1.004, 1.49), respectively). Likewise, using backward elimination, only male sex and the interaction term between ETV and
NO2 exposure were significant (p < 0.10) in predicting bronchitis [OR=2.11 (0.86, 5.15) and OR= 3.26 (1.06, 10.0), respectively]. Only the interaction term between ETV and NO2 was significant
in predicting wheeze [OR = 2.02 (0.84, 4.83), comparing highest to lowest quartile].
Discussion: Our findings demonstrate the viability of GIS-based land use regression for retrospective long-term exposure estimation, and suggest a possible multiplicative effect between ETV
and traffic-related air pollution in the expression of asthma/wheeze phenotypes.


                                                                                               64
                                                                                                                                                                                                     SUNDAY SEPTEMBER 3
        Symposia Session:

SM5-PD-06                  INTRA URBAN RESIDENTIAL VARIABILITY IN PM2.5, EC, NO2, AND GIS BASED
                           EXPLORATION OF METEOROLOGICAL AND SITE-SPECIFIC MODIFICATION IN
                           THE TRAFFIC CONCENTRATION RELATIONSHIP
J CLOUGHERTY1, L BAXTER1, R WRIGHT2, J LEVY1
(1) Harvard School of Public Health, Boston, MA. (2) Channing Laboratory, Harvard Medical School, Boston, MA.
Introduction: The Asthma Coalition on Community, Environment, and Social Stress (ACCESS) is a birth cohort study aimed at understanding environmental, social, and genetic influences in
the etiology of childhood asthma. While several prior GIS-based land use regression models have explained variability in traffic-related pollution, applying these techniques in a residential
setting, it may be important to evaluate factors that may modify the traffic-pollution relationship, including local meteorology and site characteristics. We considered a large suite of traffic
and land use indicators, selecting optimal indictors for different pollutants, and also explored meteorological and site characteristics that may modify the traffic-concentration relationship.
Finally, we employed multiple methods for candidate selection across very highly collinear traffic indicators.
Methods: We measured fine particles (PM2.5), nitrogen dioxide (NO2), and elemental carbon (EC), indoors and outdoors of 44 homes representing a range of traffic densities and neighborhoods
across Boston. Samples were collected for one week per home, winter and summer, 2003 to 2005. Multiple traffic indicators were derived using Massachusetts Highway Department data and
traffic counts collected outside homes. We performed multivariate regression analyses using central site measurements to correct for temporal heterogeneity, as well as traffic and land use
indicators, and interaction terms for traffic by meteorology and site characteristics.
Results: PM2.5 was strongly associated with the central site monitor, indicating little intra-urban spatial heterogeneity. Additional variability was explained by total roadway length within 100
meters, smoking or grilling, and population density (R2 = 0.76). EC showed greater spatial variability, which was predicted by roadway length within 200m, significantly modified by wind
parameters (hours of “still” winds, below 2 m/s), and was lower during heating season (R2 = 0.43). NO2 also showed significant spatial variability, and higher NO2 was associated with
population density, cooling season, and roadway length within 50m, modified by site characteristics (presence of an obstruction between the monitor and roadway) (R2 = 0.56).
Discussion: In our final models, we found significantly more spatial variability in EC and NO2 than in PM2.5, roadway lengths provided stable predictors in residential neighborhoods where
traffic counts were sparse, and meteorology and site characteristics significantly modified the traffic-exposure relationship. Further, we found that regression tree clustering analysis provided
appropriate multivariate suites of predictors from across collinear candidates, and, by dichotomizing concentration outcomes, may be more robust to other residential sources and measurement
error.




SM5-PD-07                  INFORMATION FUSION FOR COMPUTATIONAL ASSESSMENT OF AIR QUALITY
                           AND HEALTH EFFECTS
DA SARIGIANNIS1, A GOTTI1, N SOULAKELLIS2
(1) EUROPEAN COMMISSION, JOINT RESEARCH CENTRE, IHCP, PCE, ISPRA. (2) UNIVERSITY OF THE AEGEAN, DEPARTMENT OF GEOGRAPHY, MYTILENE, LESBOS.
This work presents the functional layout, the operational features and characteristic results of the information fusion method ICAROS NET. This is an innovative technique for the assessment
of air quality and the related potential health effects at the urban and regional scales. It is based on multi-layer fusion of environmental and epidemiological data and models aiming at
reducing the error inherent to environmental measurements and their statistical interpretation. ICAROS NET exploits to the fullest the information potential of Earth observation data,
atmospheric chemical and transport models, and ground-based measurements. The assimilation of information from all three data sources into an optimised computational model allows the
estimation of tropospheric particulate loading at very high precision and very high spatial resolution.
The computational system and the corresponding method for human exposure and health effect assessment due to airborne toxicants is based on:
• Integration and assimilation of different environmental data types including data coming from satellite-based remote sensing, ground-based air quality measurements, and advanced
  atmospheric modeling,
• Emergence of meta-knowledge through the use of inference techniques for systemic data validation and forecasting,
• Sharing of the knowledge acquired through environmental data fusion among urban policy-makers and concerned stakeholders using current and future telematics infrastructure, and
• Decision support for air quality management at the urban and trans-boundary scales based on multi-criteria analysis and interactive optimization.
ICAROS NET technology shows great potential for integrated air quality and health management. The information fusion methodology implemented herein functionally integrates all three
information sources (Earth observation, modelling and analytical measurements) and reduces the overall error of the method to circa 10% with regard to estimation of human exposure to
airborne toxicants. Results from the implementation of this method for advanced spatially-resolved assessment of human exposure and the associated health effects at the urban (e.g. Athens,
Greece) and regional (e.g. Lombardy, Italy) scales are given. This overall system performance is a significant improvement over the best atmospheric models and the pollutant concentration
maps produced by spatial interpolation of measurements from the ground. Based on the increased accuracy of the result of information fusion in terms of air quality characterization and
taking advantage of its specially designed geo-database, ICAROS NET provides an equally powerful tool for calculating the relative risk on human health posed by airborne chemicals in fine
particle form. On the basis of average values of PM loading in Athens, acute mortality due to airborne particles was estimated to have increased by 0.1-1% in 2002 and by 1-6% (in the city
centre) in 2003. Chronic mortality values were significantly higher ranging from 10% increase in 2002 up to 30% (in some hot spots) in 2003. Morbidity indicators such as increase in hospital
admissions with respiratory problems showed a similarly increasing trend. Analogous results at the regional scale in Lombardy, northern Italy, showed a definitively increasing trend in both
morbidity and mortality indicators, albeit more contained than in the case of Athens for the period 2001-2003.




SM5-PD-08                  A GIS SPATIO-TEMPORAL MODEL OF AMBIENT AIR POLLUTION EXPOSURE
M BUZZELLI1, J SU1, B AINSLIE1, D STEYN1, M BRAUER1, T LARSON2
(1) The University of British Columbia, Vancouver. (2) The University of Washington, Seattle.
Introduction: Until the mid-1990s time series analysis set the standard for air pollution epidemiology. The advent of spatial analysis and geographic information systems (GIS) has led to a
rise in studies of chronic exposure and health effects based on the rationale that intra-urban variations in ambient air pollution concentrations are as great as inter-urban differences. Such
studies typically rely on local spatial covariates derived from circular buffers to predict concentrations/exposures at receptor sites, as a means of averaging the annual net effect of
meteorological influences. To date spatial studies of chronic exposures and temporal studies of acute exposures have not been adequately integrated.
Methods: This paper presents a GIS-based methodology for air pollution exposures that vary both through time and across space and consider the role of topography and meteorology. As a
first calibration of this approach, a source area analysis integrating wind speed, wind direction and cloud cover was used to predict daily CO, NOx and NO2 concentrations over 19 days at 19
monitoring stations. Hourly buffer sections (“wedges”) identifying source areas that impact receptor sites were created for each station and estimated hourly concentrations were aggregated
back to daily totals and regressed against daily monitoring stations measurements.
Results: In total 8,644 hourly wedge buffers were created and aggregated to daily totals. Estimated daily concentrations were used to predict monitored pollution in the area (wedge) of
interest. Predictiveness varies by pollutant and meteorological condition (stability class) with predicted-observed correlations ranging from r = 0.30-0.60. The final analysis masks (clips) land
uses within each wedge buffer revealing the land uses contributing to exposures at different points in time and space. Land use composition varies significantly from traditional buffering
(circular, grid) approaches, in turn impacting how other approaches such as land use regression may be implemented.
Conclusion and discussions: Source area analysis is used to develop time-based exposures that are also sensitive to location within urban regions by incorporating basic meteorological and
topographic inputs. Beyond this calibration stage the model may then be used to estimate and validate spatio-temporal exposures at points of interest for epidemiological studies. The model
may also be used to refine land use regression approaches by refining the quantities and types of land uses and activities used as covariates in those models.




                                                                                               65
         Symposia Session:

SM5-PD-09                   PREDICTING AMBIENT PM2.5 INFILTRATION FOR INDIVIDUAL RESIDENCES AT
                            A REGIONAL SCALE
P HYSTAD1, E SETTON1, P KELLER1, R ALLEN2, D CLOUTIER-FISHER1, L FOSTER1, M BRAUER3
(1) University of Victoria, Victoria. (2) University of Washington, Seattle. (3) University of British Columbia, Vancouver.
Introduction: Infiltration is an important determinant of exposures to ambient PM2.5. Since people spend more than 60% of their time at home it is important to incorporate residential
infiltration into exposure assessments. The feasibility of a regional infiltration model based on residential building characteristics, simple meteorological and topographic variables was explored
as part of the Border Air Quality Study (BAQS) for the Georgia Basin Puget Sound airshed (GBPS).
Methods: The potential of using spatial property assessment data (SPAD) to acquire residential building attributes was first examined. It was found that these data are suitable for use with
geographic information systems (GIS) and contain detailed information for individual residences (~2.4 million in GBPS) that can influence infiltration, such as building age and improvements,
square footage, condition, value, and number of stories. Infiltration rates calculated from ten-day continuous indoor and outdoor Nephelometer measurements at 62 residences were then
compiled from a previous study in Seattle, Washington. These data were linked to SPAD using residential addresses, and meteorological and topographic data compiled for each residential
location. Univariate and multivariate regression were then conducted.
Future work will involve incorporating Seattle infiltration data with current infiltration monitoring of 80 residences in Victoria BC, Canada. The two samples will be combined to evaluate a
final infiltration model and predict infiltration of ambient PM2.5 for all 2.4 million homes in the region.
Results: Preliminary findings indicate that meteorological data and housing characteristics can predict a large percentage of the variance in residential infiltration efficiencies. Temperature
and relative humidity explained 53% of the variance in infiltration. Predicting infiltration by a dichotomous heating season (October-February) and non-heating season (March-September)
                                                                                                                                                                            _
explained 49% of the variance in PM2.5 infiltration. Significant differences in infiltration were found between residences classified as apartments/condominiums (x= 0.70), group homes
 _                                 _
(x= 0.51), and detached homes (x= 0.60). Age of residence and infiltration showed no significant relationship, but this is likely due to a lack of newer homes in the sample; current monitoring
aims to address this issue. Infiltration of PM2.5 into detached homes was also influenced by number of stories (r2 = 0.14, p = 0.13) and value of residence (r2 = 0.11, p = 0.15).
Conclusions: An infiltration model based on property assessment data, meteorology, and topography predicted a large portion of the variability in ambient PM2.5 infiltration for individual
residences. Property assessment data are readily available for North America and in the absence of indoor ambient PM2.5 measurements provide valuable information to predict indoor ambient
PM2.5 concentrations throughout large regions.




SM5-PD-10                   SPATIAL DISTRIBUTION OF URBAN WINTER WOODSMOKE CONCENTRATIONS IN
                            AN URBAN AREA
T LARSON1, J SU2, AM BARIBEAU2, M BUZZELI2, E SETTON3, M BRAUER2
(1) University of Washington, Seattle. (2) University of British Columbia, Vancouver. (3) University of Victoria, Victoria.
Introduction: In many areas, including major cities, residential wood burning is a significant wintertime source of PM2.5. As part of a larger study on health effects of air pollution (the Border
Air Quality Study), our objective was to develop a spatial model of seasonal average ambient woodsmoke levels.
Methods: Two-week average PM2.5 and levoglucosan (a marker for wood smoke) concentrations were concurrently measured at up to six sites in the study region from October, 2004 to April,
2005. In addition, pre-selected routes spanning the major population areas in and around Vancouver, B.C. were traversed during 19 cold, clear winter evenings (9PM to 1AM) from November,
2004 to March, 2005 by a vehicle equipped with GPS receiver and a nephelometer. Fifteen second average values of light scattering coefficient (bsp) and spatial position were recorded during
each mobile sampling period. Over 12,000 individual bsp measurements were adjusted for variations between evenings and then combined into a single, highly resolved map of nighttime
winter bsp levels. A GIS-based spatial regression model was developed using selected spatial co-variates to predict these temporally adjusted bsp values.
Results: Several nighttime ‘hotspots’ were identified in surrounding residential communities, distinguished by their elevated bsp values relative to bsp values measured at the locations of the
regulatory network PM2.5 monitors. In addition, the PM2.5 values measured by our own fixed-site network were highest in these ‘hotspot’ areas. At our fixed site locations, the temporally
adjusted average bsp values were also highly correlated with the average levoglucosan levels (R2=0.85 ). A relatively simple but robust (R2=0.64) model using hydrological catchment areas to
predict the nighttime bsp levels throughout the study area was developed that included census-based housing and economic covariates. The bsp values predicted by this model were correlated
with the average levoglucosan concentrations at our fixed site locations (R2= 0.66).
Discussion and Conclusions: To our knowledge, this is the first measurement-based implementation of land use regression applied to urban woodsmoke levels. This approach resulted in
our ability to identify the location of elevated, persistent wintertime levels of woodsmoke that were not captured by a relatively dense regulatory ambient monitoring network. Given that
woodsmoke sources are dispersed in numerous residential locations throughout the metropolitan area, this result was anticipated, but supports a need for approaches such as the one
described here to assess exposure to woodsmoke in urban areas.




SM5-PD-11                   A NOVEL GEOSTATISTICAL APPROACH FOR MODELING, VISUALIZING AND
                            PROPAGATING SPATIAL UNCERTAINTY IN CANCER MORTALITY MAPS
P GOOVAERTS1
(1) BioMedware Inc, Ann Arbor.
Cancer mortality maps are used by public health officials to identify areas of excess and to guide surveillance and control activities. Quality of decision-making thus relies on an accurate
quantification of risks from observed rates which can be very unreliable when computed from sparsely populated geographical entities or for diseases with a low frequency of occurrence.
Smoothing methods have been developed to improve the reliability of these estimates by borrowing information from neighboring entities. Filtering local details of the spatial variation of the
risk leads, however, to the detection of larger clusters of low or high cancer risk while most spatial outliers are filtered out. Static maps of risk estimates and the associated prediction variance
also fail to depict the uncertainty attached to the spatial distribution of risk values and does not allow its propagation through local cluster analysis.
This paper presents a geostatistical methodology to generate multiple realizations of the spatial distribution of risk values. These maps are then fed into spatial operators, such as in local
cluster analysis, allowing one to assess how risk spatial uncertainty translates into uncertainty about the location of spatial clusters and outliers. This novel approach is applied to age-adjusted
breast and pancreatic cancer mortality rates recorded for white females in 295 US counties of the Northeast (1970-1994).
Geostatistical simulation generates risk maps that are more variable than the smooth risk map estimated by Poisson kriging and reproduce better the spatial pattern captured by the risk
semivariogram model. Local cluster analysis of the set of simulated risk maps leads to a clear visualization of the lower reliability of the classification obtained for pancreatic cancer versus
breast cancer: only a few counties in the large cluster of low risk detected in West Virginia and Southern Pennsylvania are significant over 90% of all simulations. On the other hand, the cluster
of high breast cancer mortality in Niagara county, detected after application of Poisson kriging, appears on 60% of simulated risk maps.
The approach presented in this paper enables researchers to generate a set of simulated risk maps that are more realistic than a single map of smoothed mortality rates and allow the
propagation of cancer risk uncertainty through local cluster analysis. Coupled with visualization and querying capabilities of geographical information systems, animated display of realizations
can highlight areas that depart consistently from the general behavior observed across the region, guiding further investigation and control activities.




                                                                                                 66
                                                                                                                                                                                                      SUNDAY SEPTEMBER 3
        Symposia Session:

SM5-PD-12                  GEOSTATISTICAL ANALYSIS AND RISK ASSESSMENT ON SOIL ARSENIC IN
                           XINYI DONGHAI AGRICULTURAL AREA, JIANGSU PROVINCE, CHINA
Y ZHAO1, Z BAI1, R SHI2, X FU1, H GAO2, Y HE2
(1) College of Environmental Science and Engineering, Nankai University, Tianjin. (2) Agro-Environmental Monitoring Center, Ministry of Agriculture of China, Tianjin.
The soil arsenic (As) concentrations of Xinyi-Donghai agricultural area, Jiangsu province, China, was investigated in 2003. Total 651 soil samples were collected according to the stratified
systematic sampling method. The Geostatistical analysis was used to calculate variogram, fit model and estimate spatial distribution of soil As. The analysis results showed that the soil As
concentrations in the area were in a range of 1.6 - 19.0 mg/kg with normal distribution, and the background level was 7.0 mg/kg. The variogram presented the anisotropic feature and had
the best correlations in the SE-NW direction. A linear model with sill was applied to fit the variogram and sequential Gaussian simulation method was applied to estimate the spatial
distribution of soil As. The interpolation results showed that the concentrations of soil As in the suburbs were higher than those in other agriculture area, and those in the West monitoring
area were higher than those in the East. Meanwhile, the risk assessment information system (RAIS) was adopted to estimate the carcinogenic risk values of four exposure routes to
contaminants in soil and homegrown produce. The default values of some parameters of the risk equations were modified in conformity to the actual conditions of the target area. The results
showed that carcinogenic risk of incidental ingestion, inhalation and dermal contact pathways were acceptable to the human health. The ingestion of fruit and vegetables is the most important
pathway that would affects human health, and the risk value is in a range of 3.34 x 10-04 - 3.97 x 10-03. Because the calculation method of risk value in agricultural land use is different from
that of residential land use, there is not an acceptable risk value that is available in the hazard assessment for the soil As pollution in the agricultural land. Considering the background level
of the monitoring area, the cancer risk of ingestion of fruit and vegetables is not acceptable if it is larger than 2.51 x 10-03. Therefore, human activities have caused soil As contamination in
local agricultural soil of Xinyi-Donghai area.
Key words: geostatistics, soil arsenic, sequential Gaussian simulation, risk assessment, agricultural land use




SM5-PD-13                  RISK OF CONGENITAL ANOMALIES IN RELATION TO GEOGRAPHIC DENSITY OF
                           LANDFILL SITES IN ENGLAND
P ELLIOT1, S RICHARDSON1, JJ ABELLAN1, A THOMSON2, C DE HOOGH1, L JARUP1, D BRIGGS1
(1) Imperial College London, London. (2) London School of Hygiene and Tropical Medicine, London.
Introduction: Previous studies have suggested an increased risk of congenital anomalies near landfill sites, but their design did not permit risks from proximity to one or a large number of
sites to be differentiated.
Methods: Using powerful statistical and GIS techniques, we carried out a small-area study of the variation in the incidence of congenital anomalies (non-chromosomal specific and all
combined, for the period 1983 -1998) by 5x5 km grid square in England, and its relation to geographic density of 8804 landfill sites, adjusting for confounders. A landfill exposure index for
the births in each grid square was computed. This measure was calculated separately for sites handling ‘special’ (i.e. hazardous) wastes (607 sites) and non-special or unknown waste types.
Results: For special waste sites, adjusted risks (per additional special landfill site within 2 km of residence) were 1.11 (95% credibility interval 1.01 to 1.22) for hypospadias and epispadias,
1.08 (0.93 to 1.26) for cardiovascular defects, 1.10 (0.99 to 1.22) for neural tube defects, and 1.02 (0.87 to 1.20) for abdominal wall defects. For all anomalies combined, the risk was 1.03
(0.96 to 1.10). No excess risks were found for sites handling non-special or unknown waste types.
Discussion and conclusions: Though mechanisms and pathways for environmental exposures to landfill emissions are poorly understood, the results of this and other studies indicate a
geographic association between risk of congenital anomalies and proximity to special (hazardous) waste sites. Further work is needed to evaluate possible risks and routes of exposure of
landfill emissions to human populations




SM5-PD-14                  INCINERATOR AND SPATIAL EXPOSURE DISTRIBUTION: AN EXAMPLE OF
                           SMALL AREA STUDY IN ITALY
A RANZI1, M CANTARELLI2, L ERSPAMER1, V FANO3, R BACCHI2, P LAURIOLA1, F FORASTIERE3, CA PERUCCI3
(1) Environmental Epidemiology Department – Regional Agency for Environmental Prevention of Emilia Romagna, Modena. (2) Public Health Department - Local Health
Authority, Forlì. (3) Department of Epidemiology, Local Health Authority, Rome.
Introduction: The industrial area of Coriano (Forlì, Italy) is characterized by the presence of two incinerators. Increasing concern has arisen among local population and policy makers about
the possible adverse effects on health. An European project has been carried out, aimed at the implementation of an environmental-health surveillance system in the urban area near the
incinerators. A Geographical Information System (GIS) database containing environmental, population and health data has been developed, in order to better define population exposure in
epidemiological analyses of the area.
Methods: The study area is defined as the circle of 3.5 km radius centered on the two incinerators. Residents in the study area between 1/1/1990 and 31/12/2003 have been included in the
study. Residence history and the geographical coordinates of residencies have been traced back by the General Registry Office. Individual residential data were linked with the Mortality
database (from 1990 to 2003), the Cancer Registry database (from 1990 to 2003) and the Hospital Admission database (from 1999 to 2003). Estimated average emissions and concentration
of environmental pollutants (NOx, SOx, COV, CO, TSP, C6H6, Ni, Pb, HCl) were classified on a 500m2 grid determined by geographical coordinates. Different source of pollution were considered,
such as road traffic, industrial premises, incinerators and heating.
Results: A total of 39,586 individuals have been resident in the study area during the study period and have been followed-up. In the integrated database, each record includes the
demographical information as well as the information of the health episodes (hospital admission, cancer incidence, death) and the average concentration of pollutants in the square of
residence. Two approaches of data analysis of the cohort are currently under way using data from this database: distance from the incinerators and disease mapping (using Bayesian estimators)
where a 500 meter square constitutes the unit of analysis. The preliminary results indicate a significantly increased risk of mortality among women living in the vicinity of the incinerators
compared to those living far away, for all causes, colon and breast cancer, diabetes, and cardiovascular diseases.
Discussion and conclusions: Retrospective studies may use a combination of dispersion modeling, life-course data and indicators to reduce exposure misclassification in both short and long-
term studies. The use of environmental data in the analyses has opened a discussion about the best indicator for each pollution source, in particular incinerators. Since heavy metals seem to
be more directly associated with incinerator pollution than other pollutants, i.e. they are good markers of the pollution sources, they are to be employed in the epidemiological analyses to
confirm initial findings.




                                                                                               67
         Symposia Session:

SM5-PD-15                   DISPERSION MODELLING AS A DIOXIN EXPOSURE INDICATOR IN THE VICINITY
                            OF A MUNICIPAL SOLID WASTE INCINERATOR: A VALIDATION STUDY
JF VIEL1, N FLORET1, E LUCOT1, JY CAHN2, PM BADOT1, F MAUNY1
(1) University of Franche-Comté, Besançon. (2) Inserm, University Hospital, Grenoble.
Introduction. Whether low environmental doses of dioxin affect the general population, particularly in the vicinity of a municipal solid waste incinerator, is debated. In a previous study, we
found a 2.3-fold risk for non-Hodgkin lymphoma associated with residence in areas classified as highly exposed to dioxin emitted from a municipal solid waste incinerator. The main limitation
of this study lay within the use of a first-generation Gaussian-type dispersion model as a proxy for dioxin exposure, since its validity or appropriateness had not been assessed before. The aim
of this study was therefore to validate this geographic-based exposure through PCDD/F measurements from soil samples.
Methods. Isopleths of predicted ground-level concentrations were digitalized, contoured onto the surface of the map and classified in four categories of increasing exposure. Seventy-five
sampling points were determined in relation to homogeneous geological and topographical conditions. PCDD/F concentration, pH, organic carbon concentration, cation exchange capacity,
and geomorphology and ecology features were assessed for each soil sample.
Results. PCDD/F soil concentrations ranged from 0.25 to 28.06 pg World Health Organization toxic equivalent (WHO-TEQ)/g dry matter. An interaction between measured dioxin concentrations
and topography complexity (simple terrain on the northeast side with gentle hills of moderate slope, and complex terrain on the southwest side with more pronounced hills and valleys) was
found.
                                                                                                                Table: Means (standard deviations) of dioxin soil concentrations (WHO-TEQ/g dry
                                                                                                                matter) per geographic-based exposure and topography complexity categories.
                                                                                                                These results were confirmed by multivariate models, adjusting for organic
                                                                                                                carbon concentration and altitude. In simple terrain, a strong gradient across
exposure categories was highlighted. Conversely, in complex terrain, the diffusion model overpredicted ground-level concentrations, particularly in the high exposure zone.
Discussion and conclusions. First-generation modelling provided a powerful tool for dioxin relative classification of exposure in simple terrain, reinforcing the results of our case-control
study (since 89.5% of cases and 90.7 of controls lived on the northeast side). However, a more advanced atmospheric diffusion model should have been used for refined assessment in complex
terrain. This study confirms that use of first-generation models should not be transferred to another geographic region without validation with measurement data from the new area.




SM5-PD-16                   PESTICIDE EXPOSURE ASSESSMENT FOR A POPULATION BASED
                            CASE CONTROL STUDY OF CHILDHOOD CANCERS
K ELGETHUN1, S HOREL2, S CAROZZA2
(1) Dept. of Geography, Texas A&M University, College Station. (2) Dept. of Epidemiology & Biostatistics, Texas A&M University, College Station.
Introduction. The overall goal of this epidemiological case-control study is to estimate risk of developing specific childhood cancers associated with prenatal exposure to pesticides used in
agricultural settings in Texas. During the study period, all but one of the five top crops by acreage in Texas were treated with chemicals that show evidence of possible human carcinogenicity
and many of these crops are sprayed from crop dusting airplanes, increasing potential for off-target drift. Cotton was chosen for evaluation of potential pesticide drift exposure because of
the high volume, repeated aerial spraying of this crop.
Methods. The study area comprised 12 Texas counties with intense cotton farming. To identify specific cotton fields during the study period, Landsat 5 TM data was multitemporally merged
then classified through supervised classification. Training data for the classification were obtained through land cover identification and location with GPS of over 200 cotton fields. Once
training and image classification was completed, an accuracy assessment was conducted to determine the effectiveness of the classification.
Meteorological data were obtained for morning hours (likely spray time) for locations and time windows (months) of interest. Windows corresponded to times when potential in utero exposure
could have occurred. Meteorological data and other parameters were used as inputs for drift modeling (AgDrift v.2.03). Wind direction-specific model output was then integrated in a GIS for
each cotton field and each time window, and potential exposures were assigned to individuals based on model-predicted pesticide deposition and air concentration at geocoded birth
residence.
Results. Landsat 5 TM data effectively covered the study area and time period and frequent over-flights allowed for the inclusion of multiple dates per year. Mutiltemporal data merging
combined images obtained on different dates and proved effective for agricultural crop identification. Estimated exposure was primarily driven by wind velocity and direction. For some
individuals, multiple fields contributed to exposure.
Discussion and conclusions. These GIS-based methods allow for the assessment of potential historical exposure to cotton pesticides of children living in farming communities and will allow
for expansion of the study area. Information resulting from these exposure assessment tools is being used for epidemiologic analysis of risk of specific childhood cancers. Because causes of
cancer in children continue to largely elude identification, leaving few avenues for developing prevention strategies, the possible role of agricultural pesticide exposure in the development of
childhood cancers offers a tantalizing potential for preventing at least some of these cases.




SAA3-PD                     HEAVY METALS
SAA3-PD-01                  LEAD EXPOSURE OF YOUNG CHILDREN IN A FRENCH URBAN AND INDUSTRIAL
                            AREA: BLOOD LEAD LEVELS AND FACTORS OF VARIATION
C NISSE1, A LEROYER1, M TONNEAU2, I BATKIN3, F DOUAY4, L LABAT5, B DEHON5, P JEHANNIN6, G AUQUE2, H FOURRIER4, JM HAGUENOER1
(1) EA 2690-UNIVERSITE de LILLE 2, LILLE. (2) INSTITUT DE SANTE AU TRAVAIL DU NORD DE LA FRANCE, LILLE. (3) CENTRE ANTI POISON, LILLE. (4) INSTITUT
SUPERIEUR D’AGRICULTURE, LILLE. (5) LABORATOIRE DE TOXICOLOGIE ET GENOPATHIES-CHRU, LILLE. (6) DDASS DU NORD, LILLE.
Introduction: Lead poisoning prevention is a priority of public health in France. To evaluate the prevalence of lead poisoning in young children and study the main factors still influencing
blood lead levels today, we proposed the determination of blood lead concentration (BLLs) in children living in a urban sector with important traffic, many old houses and next to a battery
manufacturing plant in activity since 1890. In this area, concentrations of lead in soils of schools, parks and private gardens ranged from 39 to 2818 mg/kg; in schools and houses, outside
lead dust concentrations ranged from 71 to 20249 g/m2, and inside lead dust concentrations from 12 to 9653 g/m2.
Methods: Blood tests were performed with parental agreement among children under 6 years, in 12 schools and 2 day nurseries situated at less than 1 km of the battery plant. The parents
completed a questionnaire about risk factors for lead exposure.
Results: 1213 children participated (participation rate = 83 %). Geometric mean of BLLs was 24 µg/l [23.4 – 24.7]. Nine children had BLLs over 100 µg/l (0.75 %).
The relations between blood lead level and exposure risk factors such as behavior, environment, hygiene, alimentary habits were studied in a multivariate model. The parameters significantly
related to BLLs were: the age (higher BLLs between 3 and 4) and the sex (higher in boys), the date of construction of the residence (higher in houses built before 1948), the clothing hygiene
of the child (higher when the clothes cleanliness was bad), the location of the residence with regard to the factory (higher when located at the North-Est, under the wind of the battery plant),
the period of the test (higher in september-october), the fact of having a relative working in the factory, the inactivity of the mother. The presence of scales of painting, of water lead pipes, of
softener of water, and recent works in the house of the children did not influence the BLLs.
Conclusions: Surprisingly the mean blood level and the prevalence of lead poisoning were quite low in this population attended to be at high risk for exposure. Even at such low levels, usual
risk factors of lead poisoning were related to BLLs. This is of great interest to determine, who, today is still at risk for lead exposure and which are the priority axes of prevention to lower BLLs
as much as possible.




                                                                                                 68
                                                                                                                                                                                                       SUNDAY SEPTEMBER 3
        Symposia Session:

SAA3-PD-02                  CHILDHOOD LEAD POISONING SCREENING ACTIVITY IN FRANCE BETWEEN
                            1995 AND 2002
F CANOUI-POITRINE1, P HARRY2, D POISOT3, M MATHIEU-NOLF4, C CEZARD4, S SABOURAUD5, J ARDITTI6, J MANEL7, R GARNIER8, O GUILBERT9, C
LECOFFRE1, P BRETIN1
(1) InVS, SAINT-MAURICE. (2) Angers Poison Center, ANGERS. (3) Bordeaux Poison Center, BORDEAUX. (4) Lille Poison Center, LILLE. (5) Lyon Poison Center, LYON. (6)
Marseille Poison Center, MARSEILLE. (7) Nancy Poison Center, NANCY. (8) Paris Poison Center, PARIS. (9) Toulouse Poison Center, TOULOUSE.
Introduction: Lead poisoning can be responsible for neurodevelopmental and hematopoietic damage especially in young children. Deteriorated lead-based paint in old buildings is one of the
main risk factors. In 1995, a Childhood Lead Poisoning National Monitoring System began recording blood lead level and characteristics of all children who had lead blood tests; the screening
was targeted to children with a probability of high exposure to lead. The objectives of the present study were 1) to enumerate children screened for lead poisoning and new cases, and 2) to
describe their distribution over time as well as geographically, their characteristics and their risk factors to lead poisoning.
Methods: The study’s population was children who had one or more lead blood test in France between 1995 and 2002. In order to take a precise census of the number of children screened
for lead poisoning and new cases, two sources of data were crosschecked: data recorded in the Monitoring System and questionnaires filled by the local authorities of public health in 2003.
Results: A total of 36,151 children (35/100 000 children per year) were screened at least once. Most of them (95%) were under 7. The probability, for a child under 7, to have a lead blood
test once was 0.6 %. The annual incidence of detected lead poisoning cases was 17/100 000 children under 7 (N=5974). During the study period, the percentage of screened children who
had a lead blood level above the threshold of 100 •g/L decreased regularly (from 24.5% in 1995 to 8.5% in 2002). The screening activity was concentrated in three of the 22 French regions
(Ile-de-France = 61%, Rhônes-Alpes =13% and the North area= 9%) and 66% of the new cases lived in Paris or the suburbs even though the population of children was less than 20% of
the total of French children. Moreover, 64% of the screened children lived in an old and deteriorated housing. The percentage of old and uncomfortable housing in each region wasn’t correlated
to the percentage of lead poisoning cases.
Discussion and conclusion: The incidence of childhood lead poisoning decreased regularly since 1995. Nevertheless, the French childhood lead poisoning screening activity stayed too much
heterogeneous. Even if risk factors of lead poisoning aren’t equally distributed in the French territory (old housing, industrial sites), numerous children exposed to lead are not screened in
many regions and all the new cases are not identified.




SAA3-PD-03                  LEAD CONCENTRATION IN THE BLOOD OF CHILDREN AND ITS ASSOCIATION
                            WITH LEAD IN SOIL AND AMBIENT AIR-TRENDS BETWEEN 1983 AND 2000
                            IN A GERMAN INDUSTRIAL CITY
U RANFT1, T DELSCHEN2, M MACHTOLF3, D SUGIRI1, M WILHELM4
(1) Institut für umweltmedizinische Forschung (IUF) an der Heinrich-Heine-Universität Düsseldorf, Düsseldorf. (2) Landesumweltamt Nordrhein-Westfalen, Essen.
(3) Institut für Umweltanalysen (IFUA), Bielefeld. (4) Abteilung für Hygiene, Sozial- und Umweltmedizin, Ruhr-Universität Bochum, Bochum.
Introduction: For risk assessment and to derive trigger values of lead (Pb) in soil, results of biomonitoring studies on the association between Pb concentration in soil and Pb concentration
in the blood of exposed persons are of high importance, especially for young children who are at higher risk of exposure. This study evaluates the importance of soil as intake pathway of
ambient Pb. As Pb in ambient air is also associated with Pb in blood, it has to be considered too.
Methods: In a series of 5 cross-sectional studies between 1983 and 2000, 843 children (6-11-y-old), living in the industrial city of Duisburg, provided blood samples which were analysed for
Pb. Potential confounders (gender, parents’ educational level, ETS, unfavourable heating) were obtained by questionnaire. Based on routine data of Pb in soil and deposition, spatial distribution
of Pb in soil (0–10 cm depth) was estimated. Pb exposure in ambient air was calculated using routine air quality data and Lagrange dispersion modelling. Therefore, exposure data of Pb in
soil and ambient air at the child’s place of residence could be individually assigned. Multiple regression analysis was applied to estimate adjusted association measures.
Results: Median (95th percentile) exposure of Pb in soil was about 200 (950) mg/kg and nearly constant over study time whereas Pb in ambient air decreased from 0.47 (0.47) to 0.03 (0.19)
µg/m? and Pb in children’s blood from 85 (162) to 31 (67) µg/l. 59% of the variance of Pb in blood could be explained by regression whereas Pb in ambient air contributed 51% and Pb in
soil 3%, respectively. If one hypothesises an increase of Pb in ambient air by 0.45 µg/m?, Pb in blood would increase by 161% whereas a hypothetical increase of Pb in soil by 800 mg/kg
would result in an increase of Pb in blood by 30%.
Discussion and conclusions: Besides ambient air, soil is a relevant pathway for intake of ambient Pb. Since Pb in ambient air considerably decreased within the last 20 years, but Pb
concentrations in soil remained nearly constant, the Pb blood concentrations decreased markedly in the same period. Indications exist that Pb in blood below 100 µg/l may be of health concern.
Toddlers (1-3 year) are the most exposed children. Their Pb concentration in blood is about one third higher than that of 6-y-old children. Therefore, soil can relevant contribute to the Pb intake
of young children.




SAA3-PD-04                  MATERNAL SELF ESTEEM MODIFIES THE NEUROTOXICITY OF LEAD
P SURKAN1, L SCHNAAS2, RJ WRIGHT3, MM TÉLLEZ-ROJO2, L HÉCTOR2, H HU1, M HERNÁNDEZ-AVILA2, DC BELLINGER3, RO WRIGHT3
(1) Harvard School of Public Health, Boston. (2) Instituto Nacional de Salud Pública, Cuernavaca. (3) Harvard Medical School, Boston.
Background: Social stressors and lead poisoning are environmental risk factors for poor neurodevelopment. Self-esteem is negatively correlated with perceived stress and may act as a buffer
for social stress. Previous studies of child development have considered social factors as potential confounders of lead poisoning. There are biological reasons to believe that stress modifies
the toxicity of lead poisoning instead.
Methods: This study was conducted among mother-infant pairs living in Mexico City. Blood lead, maternal self-esteem (Coopersmith Score), and Bayley Scale Mental Development Index (MDI)
were collected at 24 months of age. Bivariate associations were determined by using a chi-square or Student t-test. A multiple linear regression model was constructed using the MDI as the
dependent variable. Self-esteem, sex, blood lead, maternal IQ, and maternal years of education were entered as covariates. In the interaction analysis, self-esteem was divided into quartiles
and dummy coded so that the lowest 3 quartiles were used as the referent group.
Results: Complete data on blood lead, maternal self-esteem and MDI scores were available from 270 mother-child pairs. A one unit increase in Coopersmith score predicted a 0.4 unit increase
in MDI (p=0.01) and a one unit increase in blood lead was associated with a 0.3 unit decrease in MDI (p=0.2). There was a significant interaction between self-esteem score and blood lead
in predicting the 24 month MDI. The highest quartile of self-esteem was associated with a 12 point increase in MDI among the children with lead poisoning (blood Pb>10 •g/dL) and only a
2 point increase among children with blood lead levels <10 •g/dL(interaction p=0.03). Similarly, blood lead predicted poorer MDI scores within the lowest 3 quartiles of self-esteem (lead
beta=-0.4; p=0.04) but not within the highest quartile of maternal self-esteem (lead beta=0.8 p=0.11).
Discussion: Maternal self-esteem predicted performance on Bayley’s Mental Developent Index in children exposed to lead. Furthermore, higher maternal self-esteem scores modified lead’s
toxicity. The direction of the effect suggests that maternal self-esteem mitigates the effects of lead. These results suggest that social factors may modify lead toxicity. Given this, social
interventions may be effective treatments for lead poisoning. Further research is needed within other domains of social stressors to confirm and validate these findings.




                                                                                                69
        Symposia Session:

SAA3-PD-05                 DIETARY CALCIUM AS A MODIFIER OF THE RELATIONSHIP BETWEEN LEAD
                           BURDEN AND BLOOD PRESSURE AMONG POSTPARTUM WOMEN
A ROY1, A ETTINGER1, H HU1, M HERNANDEZ-AVILA2
(1) Harvard School of Public Health, Boston. (2) National Institute of Public Health, Cuernavaca.
Introduction: Cumulative lead burden and low dietary calcium have been independently associated with increased risk of hypertension. Research has shown that mobilization of cumulative
bone lead stores peak during lactation when calcium demand is also high. This may lead to increases in maternal blood pressure. We tested the hypothesis that dietary calcium intake modifies
the relationship between lead burden and blood pressure in postpartum, lactating women.
Methods: We examined the effects of blood and bone lead on systolic and diastolic blood pressure among N=629 lactating women at one-month postpartum in Mexico City. Subjects were
comprised of a cohort of women recruited for later participation in a randomized placebo-controlled trial of calcium supplementation during lactation and, thus, women with existing
hypertension or pregnancy-related hypertension disorders were excluded at baseline. Blood lead analysis was performed using graphite furnace atomic absorption spectrophotometry. Bone
lead was measured using a K-X-ray fluorescence instrument. Calcium intake was assessed by a self-administered, semi-quantitative, food frequency questionnaire designed to estimate usual
dietary intake over an extended period of time (one year) prior to completion of the questionnaire.
Results: Over fifty percent of women reported receiving less than the recommended dietary intake for pregnant and lactating women age 19-50 of calcium (1000 mg/day). Mean levels of
lead were: blood 9.3 mcg/dL (SD=4.7, range 1.8-60.1); patella 14.7 mcg/g (SD=16.3, range <1-85.9); tibia 9.9 mcg/g (SD=10.3, range <1-76.5). After adjustment for age, arm circumference,
breastfeeding status (exclusive or partial), smoking status, and alcohol intake, increasing intake of calcium was associated with lower systolic blood pressure (p=0.006) and increasing patella
lead with higher systolic blood pressure, though not statistically significant, at one-month postpartum. An interaction term between bone lead and calcium was also not significant. However,
among women receiving less than 1000 mg calcium/day, the bone lead effect on systolic and diastolic blood pressure was three and 12 times, respectively, that of women receiving at least
1000 mg calcium/day. Effects were similar for the association between blood lead, calcium and blood pressure.
Discussion/Conclusions: Our results suggest that dietary calcium may be a modifier of the relationship between lead burden and blood pressure among postpartum, normotensive women.
Since our cohort did not include women with frank hypertension, our results may be an underestimate of the true association. Dietary calcium may be helpful in prevention of blood pressure
elevations induced by chronic exposure to lead particularly among women with low dietary calcium intake.




SAA3-PD-06                 A RANDOMIZED CONTROLLED TRIAL OF CALCIUM SUPPLEMENTATION TO
                           REDUCE BLOOD LEAD LEVELS (AND FETAL LEAD EXPOSURE) IN PREGNANT
                           WOMEN
M TÉLLEZ-ROJO1, H LAMADRID-FIGUEROA1, A MERCADO-GARCÍA1, K PETERSON2, D BELLINGER2, A ETTINGER2, M SOLANO-GONZÁLEZ1, M
HERNÁNDEZ-ÁVILA1, H HU2
(1) Instituto Nacional de Salud Pública, Cuernavaca. (2) Harvard School of Public Health, Boston.
Introduction: Prenatal lead exposure has been associated with delays in neurobehavioral development and reduced intelligence quotient later in life. Calcium supplementation during
pregnancy may attenuate fetal exposure by inhibiting maternal bone lead mobilization and intestinal absorption of lead. This study evaluated the effect of a 1.2 g/d calcium supplementation
during pregnancy on maternal blood lead concentrations, and the extent to which calcium’s effect is mediated by inhibition of bone resorption.
Methods: During 2001 to 2003, we recruited 670 pregnant women who were at less than 14 weeks of gestation. We conducted a double-blind randomized controlled trial. Women were
randomly assigned to receive calcium (n=334) or placebo (n=336). The calcium supplement consisted of 600 mg of elementary calcium in the form of calcium carbonate twice a day; the
placebo was identical in appearance to the calcium supplement and consisted of starch tablets. Blood lead concentrations were measured at baseline, 24 and 34 weeks of gestation by means
of atomic absorption spectrophotometry. Urinary N-telopetides of type I collagen (NTx) concentrations as a marker of bone resorption were measured by ELISA in a subsample of 296 women.
The effect of the supplement was evaluated by a multivariate random-intercept regression model, to account for the longitudinal design.
Results: After adjusting for age and dietary calcium intake at baseline, calcium supplementation was associated with an overall average reduction of 11% in maternal blood lead
concentrations relative to placebo (p<0.01). This reduction was more evident at 24 weeks of gestation (14% reduction, p<0.01) and marginal at 34 weeks (7% reduction, p=0.11). Urinary
NTx concentrations were reduced by an average of 15% in the supplement group relative to placebo (p<0.01); reduction in NTx was larger at 34 weeks of gestation (16%, p<0.01) than at
24 weeks (10%, p=0.08). After adjusting for NTx, average overall reduction in blood lead concentrations relative to placebo was 9% (p=0.09).
Discussion and conclusions: Results of this study suggest that a 1.2 g/d calcium supplementation during pregnancy reduces maternal blood lead concentrations. The effect appears to be
at least partially mediated by inhibition of bone lead mobilization; much of it also appears to be attributable to inhibition of intestinal absorption of the metal.




SAA3-PD-07                 SPATIAL VARIATION IN METAL BIOMARKERS OF PERIPARTUM WOMEN NEAR
                           A MINING RELATED SUPERFUND SITE
A ZOTA1, CJ PACIOREK2, AS ETTINGER1, C AMARASIRIWARDENA3, JD SPENGLER1, DC BELLINGER4, M OSBORN5, H HU1, RO WRIGHT3
(1) Department of Environmental Health, Harvard School of Public Health, Boston, MA. (2) Department of Biostatistics, Harvard School of Public Health, Boston, MA.
(3) Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA. (4) The Children’s Hospital Boston, Boston,
MA. (5) Integris Regional Medical Center, Miami, OK.
Introduction: Limited evidence has shown that geographic proximity to hazardous waste may have consequential health effects particularly for susceptible subpopulations such as pregnant
women and infants. Geostatistical techniques provide a powerful way to examine these potential relationships. The purpose of this analysis is to examine if women who live closer to mining
waste (“chat”) have higher biomarker levels of lead and arsenic relative to those who live farther away.
Methods: We collected and analyzed biomarker data on lead and arsenic levels in peripartum women living proximate to the Tar Creek Superfund Site, a region contaminated with mining
waste. Maternal biological samples were obtained at delivery (venous blood lead, n=243; hair arsenic, n=88) and analyzed by inductively-coupled plasma mass spectrometry. Residential
address information was geo-coded in ArcGIS 8.3 using the Census TIGER Files as a reference road network. Student T-tests were used to examine differences in log-transformed maternal
biomarker levels by two different classifications representing varying proximity to chat piles (inside vs. outside superfund boundaries, towns of Picher/Cardin vs. other). Generalized additive
models and kriging methods were used to model the spatial covariance of blood lead levels.
Results: Mean (SD) values for maternal blood were: lead (0.8 ± 0.5 µg/dL, n=243), and for hair: arsenic (24.9 ± 37.6 ppb, n=88). Women who lived within the Superfund site had significantly
higher lead (0.97 vs. 0.78 µg/dL, p=0.003) and arsenic (33.7 vs. 20.6 ppb, p=0.02) levels. In addition, women who lived in towns closest to the chat piles had the highest levels of lead (1.12
vs. 0.80 µg/dL, p=0.001) and arsenic (43.5 vs. 22.0 ppb, p=0.009). Of 243 participants, 85% of residences were successfully geo-coded. Results from preliminary kriging analyses suggest that
the non-spatial variance was greater than spatial variance. However, the spatial term in the generalized additive model was significant (g(si), df= 4.3, p=0.01), and predictive spatial surfaces
support the results from the T-tests suggesting differences in maternal blood lead levels by location.
Discussion/Conclusions: These data suggest differences in human exposure by proximity to hazardous waste confirming the hypothesis that contact to toxins may be greater for populations
residing near a Superfund site. Failing to detect a significant spatial correlation with kriging may be due to the fact that participants are not evenly distributed throughout this rural region
rather they are clustered in dispersed towns. Future studies will evaluate our cohort for personal exposures to mining-related metal contaminants in the home environment.




                                                                                              70
                                                                                                                                                                                                       SUNDAY SEPTEMBER 3
        Symposia Session:

SAA3-PD-08                  RISK ASSESSMENT AND REMEDIATION PROPOSALS IN AN ABANDONED LEAD
                            SMELTER COMPLEX IN CHIHUAHUA, MEXICO
LH SANIN1, MT ALARCON-HERRERA2, M ORNELAS1, F DÍAZ-BARRIGA3, S REZA1, I ROMIEU4
(1) FEN. Universidad Autónoma de Chihuahua, Chihuahua. (2) Centro de Investigación en Materiales Avanzados (CIMAV), Chihuahua. (3) Facultad de Medicina.
Universidad Autónoma de San Luis Potosí, San Luis Potosí. (4) Instituto Nacional de Salud Pública de México, Cuernavaca, Morelos.
INTRODUCTION: Lead has well known deletereous effects, especially affecting women and children under 5 years old.
Risk assessment of the residual lead intoxication risk is presented, 10 years after the closure of the one of the world’s largest smelter complex, to present the viable remediation alternatives
for the encountered contamination.
METHODS: Seven years earlier, a risk assessment of the area was performed, in accordance with the EPA (Environmental Protection Agency) methodology, including stratification of the zone
in function of the distance from the smelter. In each of the three defined strata, a random sample of mothers and their children between ages 1 and 5 was selected; capillary blood samples
were taken, and these samples were analyzed through voltammetry (lead care) with strict quality control. In each of the sampled households, soil, paint, dust and water samples were taken
for later analysis through anodic voltammetry. Pre-schools from each stratum were selected as well, and in each of them capillary blood samples from children under age 6, as well as
environmental samples corresponding to each school, were taken. Recently, an evaluation of stratum zero (the area of the facilities) and strategic points of the other strata was performed.
The two groups of data were compared, and possible remediation alternatives were evaluated for the presentation of a proposal.

a percentage of 40% ≥ 10 µg/dL and 20% > 20 µg/dL. Lead concentrations showed an inverse relation to the distance from the smelter. The global lead mean in mothers was 5.22 µg/dL
RESULTS: In the initial evaluation, a total of 324 children were analyzed whose blood lead levels ranged from 3.0 to 59.0 µg/dL; they had a geometric mean of 11.45 µg/dL (∑ ± 8.72 ) and

(∑ ± 3.36). The analyzed superficial soil samples show a range of 100 > 9000 mg/Kg. The predictors of lead in children, through Multiple Linear Regression, are related to exposure to soil.
The recent evaluation shows concentrations from 509 mg/Kg in strata I and II to over 15 000 mg/Kg in stratum zero. The remediation of the site through phytostabilization was found to be
an appropriate measure. This, together with a permanent educational program that includes hygienic and dietary recommendations, would result in the minimization of the associated risk,
limiting the damage from exposure in children above all.
CONCLUSIONS: It is estimated that soil is contributing to over 90% of the lead exposure, especially in small children, with values higher than those established by the EPA as requiring
remediation measures. It is clear that there is residual contamination in the zone adjacent to the smelter, which must be remediated. To reduce the risk of exposure, it is recommended to
perform a phytostabilization of the inhabited areas, together with educational activities and epidemiological vigilance.




SAA3-PD-09                  RISK OF CANCER AND ENVIRONMENTAL EXPOSURE TO CADMIUM IN A
                            PROSPECTIVE POPULATION STUDY
T NAWROT1, M PLUSQUIN1, J HOGERVORST2, H ROELS3, H CELIS1, L THIJS1, J VANGRONSVELD2, E VAN HECKE1, JA STAESSEN1
(1) University of Leuven (KULeuven), Leuven. (2) University of Hasselt (UH), Hasselt. (3) University of Louvain (UCL), Brussels.
Background: Cadmium accumulates in the human body, so that the 24 h urinary excretion is a biomarker of life-time exposure. Until now, no population-based study provided evidence in
support of the classification of cadmium as a human carcinogen.
Methods: In 1985–1989, we recruited a random population sample (n=994) from an area close to three zinc smelters and a reference area with lower exposure. At baseline (1985-1989), we
measured cadmium in 24-h urine samples and in the soil of the participants’ gardens. We followed the incidence of cancer until 30 June 2004. We used Cox regression to calculate hazard
ratios for cancer in relation to the internal and external exposure to cadmium, while adjusting for covariables.
Findings: The cadmium concentration in the soil ranged from 0.8 to 17.0 mg/kg. At baseline, the geometric mean urinary cadmium excretion was 59.7% higher in the polluted than in the
reference area (12.3 vs 7.7 nmol/day, p<0.0001). During follow-up (median 17.2 years). 50 fatal and 20 non-fatal cancers occurred, of which 18 and 1 were lung cancers. Total cancer risk was
significantly associated with cadmium excretion (p=0•026), because of the positive association between lung cancer and exposure to cadmium. The population-attributable risks for lung
cancer were 79% for residing in the polluted area and 82% for smoking. The hazard ratios for lung cancer adjusted for sex, age and smoking, were 1.70 (1.13–2.57, p=0.011) for a doubling
of the urinary cadmium excretion, 4.17 (1.21–14.4, p=0.024) for residence in the high vs low exposure area, and 1.57 (1.11–2.24, p=0.012) for a doubling of the cadmium concentration in
the soil. These findings were consistent after exclusion of 42 workers with a history of occupational exposure to cadmium and after additional adjustment for the 24 h urinary arsenic excretion,
which was measured in 734 participants.
Interpretation: The observation that environmental exposure to cadmium was associated with an increased incidence risk of lung cancer must have implications for environmental policies
worldwide.




SAA3-PD-10                  LOW LEVEL EXPOSURE TO LEAD AND COGNITION AMONG OLDER WOMEN
J WEUVE1, S KORRICK2, J SCHWARTZ1, H NIE2, F GRODSTEIN2, H HU1
(1) Harvard School of Public Health, Department of Environmental Health, Boston. (2) Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital,
Harvard Medical School, Boston.
Introduction:
Lead exposure is associated with cognitive impairments in children and occupationally exposed adults. Recent data indicate that chronic low-level exposure to lead is associated with
accelerated declines in cognition among older men, but the association has not been examined among non-occupationally exposed women.
Methods:
The authors examined the relation of biomarkers of lead exposure to performance on a battery of cognitive function tests among a cohort of 578 older women. When the women were aged
47-74 years, we measured their bone lead using K-shell x-ray fluorescence and their blood lead using graphite furnace atomic absorption. Bone lead levels indicate exposures over years and
decades, while blood lead levels indicate exposures over days and months. Five years after the lead assessments, on average, we evaluated women’s cognitive function by validated telephone
interviews.
Results:
Mean (standard deviation) lead levels in tibia, patella and blood were 11.0 ug/g bone (9.9), 13.3 ug/g bone (12.1), and 3.0 ug/dL (1.9), respectively, consistent with low-level exposures. In
multivariable-adjusted analyses, bone lead levels were generally associated with worse performance on the tests of cognitive function, although none of the associations was statistically
significant.
Tibia lead level was associated with worse performance on 6 of the 8 tests. For example, a 12-ug/g increment in tibia lead (interquartile range) was associated with a 0.20-unit lower score
on a test of working memory and attention (95% CI: -0.45 to 0.05). Similarly, patella lead level was associated with worse performance on 5 of the 8 tests, most strongly on a test of delayed
verbal recall, where a 13-ug/g increment in patella lead (interquartile range) was associated with a 0.16-unit lower score (95% CI: -0.37 to 0.05). In our data, this corresponded in the difference
in delayed verbal recall that we observed between women who were 4 years apart in age. We found similar results when we examined averaged z scores of the two bone lead measures.
Blood lead level was not consistently associated with cognitive test scores.
Conclusion:
In this cohort of older women with community-level exposure to lead, bone lead, an index of cumulative exposure, was adversely but not significantly associated with performance on several
tests of cognition. These preliminary findings provide impetus to conduct further research into indices of women’s cognitive susceptibility to lead.




                                                                                                71
        Symposia Session:

SAA3-PD-11                  REPRODUCTIVE AND DEVELOPMENTAL EFFECTS ASSOCIATED WITH
                            ENVIRONMENTAL ARSENIC EXPOSURE
C HOPENHAYN1, S GOEL1
(1) Department of Epidemiology, University of KY, Lexington.
Inorganic arsenic exposure can occur in occupational settings, yet environmental exposure from naturally contaminated drinking water constitutes the major source. Millions of people
worldwide are exposed, but the highest exposures known to date have been concentrated in areas of India, Bangladesh, Chile, China and some other countries. A number of health effects
have been documented to be associated with inorganic arsenic, particularly in the last couple of decades, in which many studies have been conducted and published. The effects range widely
from skin abnormalities such as keratosis, to skin cancer, peripheral vascular effects, cardio- and cerebro-vascular diseases, bladder and lung cancer, diabetes and neuropathies. The exposure
levels and duration also vary across populations. Surprisingly, the potential relationship between arsenic ingestion and reproductive and developmental risks has only recently begun to receive
the deserved attention. In the last several years, investigations were conducted in different countries, addressing such outcomes as infant mortality, stillbirths, reduced birthweight, prematurity
and spontaneous abortions. Nevertheless, the results are somewhat inconsistent, and the studies vary widely in study design, data sources and analysis.
This paper will present a comparative analysis of all published reports to date, as well as some ongoing analyses, on the reproductive and developmental effects attributed to arsenic exposure,
focusing on environmental, drinking water studies. It will also address related biological mechanisms such as variations in arsenic metabolism during pregnancy, tissue distribution of arsenic
and cofactors that may explain some of the observed variability and across studies.




SAA4-PD                     HEALTH AND METEOROLOGY
SAA4-PD-01                  SOCIAL AND ENVIRONMENTAL DETERMINANTS OF HEAT RELATED
                            MORTALITY: A SYSTEMATIC REVIEW
S KOVATS1
(1) London School of Hygiene and Tropical Medicine, London.
Background: Although much research has focused on quantifying the health impacts of the 2003 heat wave in Western Europe, there is relatively little epidemiology on individual or group
level risk factors for heatwave mortality. I reviewed the epidemological literature on social, environmental and medical determinants of heat-related mortality, to help identify high risk
individuals who can be targeted by public health interventions.
Methods: Systematic literature review of population-based epidemiological studies of determinants of heat- and heatwave-related mortality. A search strategy was developed to identify all
relevant published studies from databases (PubMed, EMBASE, AgeInfo, Medline), as well as direct contacts with researchers. Exclusion criteria included: studies in occupational, sports or school
settings, physiological studies and case reports. Studies were limited to high income countries in temperate latitudes, and those published since 1960.
Results: Approximately 50 papers were found that fulfilled the criteria. The studies were sub- divided by exposure type: the impact of individual heat wave events, heat-related mortality, and
certified heat deaths. Results were summarised for key vulnerability factors:
• Age: Greatest effects of heat and heat waves were in the elderly but effects also apparent for adults, particularly for external causes.
• Gender: The effect of heat and heat waves is greater in women than in men, in most age groups and populations, even after adjustment for age and other factors.
• Pre-existing disease: Consistent evidence that certain pre-existing medical conditions are risk factors for mortality in heat waves, particularly diabetes, fluid/electrolyte disorders, and some
  neurological and mental disorders.
• Institutions: Strong evidence that hospital inpatients and nursing home residents are at much higher risk of heat-related mortality.
• Socio-economic status: There is limited evidence for Europe that socio-economic factors are an important determinants of risk. The evidence for the US is different, where studies consistently
  show an effect of income at individual, small area and city-level.
• Housing factors: Very limited evidence indicates that residents of top floor flats at higher risk.
• Urban populations are more at risk than rural or suburban populations.
Conclusion: Heat waves and hot weather affect a range of different high risk groups and a range of public health interventions are required. There are important differences in vulnerability
between countries, regions, and age groups.



SAA4-PD-02                  MODELING THE TIME COURSE OF TEMPERATURE EFFECTS ON DAILY
                            MORTALITY
H MOSHAMMER1, HP HUTTER1, P WALLNER1
(1) Medical University, Vienna.
Introduction: Polynomial distributed lag models have been successfully applied in time series studies on air pollution and daily deaths. In that case one would expect an early increase in the
death rate following a high pollution episode. After this maximum a minimum might be expected as a result of “harvesting”. Such an expected time course could well be modeled by a
3rd degree polynom. The study of the influences of daily temperature on mortality is more complicated as temperature and mortality do not follow an approximately linear association.
Methods: It is well known that mortality increases both at or after hot and cold days. The range of the “optimal” temperature (with the lowest mortality rate) depends on the local average
temperature i.e. on the latitude of the city under study. Cold days have a longer lasting effect while the acute (same and next day) effects of hot days seem to be more pronounced.
To capture the non-linear association between temperature and mortality a 2nd degree polynom was used i.e. both temperature (daily mean) and temperature squared were entered into the
model. The distributed lags of both temp and temp squared were modeled in a 3rd degree model.
The calculations were run on the mortality and temperature data of the city of Vienna for the years 2000-2004. A generalized additive model (GAM) exact procedure (S-Plus, Vers. 6.2) was
applied. Seasonal effects were modeled by a natural spline (degrees of freedom were chosen to minimize residual autoregression) and day of week was modeled with dummies.
Results: Models with up to 60 lag days were run and compared with each other by ANOVA. Adding more than 14 lags did no more yield any better results. Thus a 14 day lag model is finally
used to visualize the time dependent influence of temperature on daily deaths.




                                                                                                72
                                                                                                                                                                                                    SUNDAY SEPTEMBER 3
        Symposia Session:

SAA4-PD-03                 SUMMER TEMPERATURES, AIR POLLUTION AND MORTALITY IN MOSCOW
B REVICH1, D SHAPOSHNIKOV2, A GOLUB3
(1) Center of Demography and Human Ecology, Moscow. (2) Center of Demography and Human Ecology, Moscow. (3) Environmental Defense, Washington DC.
Study objective: To test the following hypothesis: daily mortality rates increase with concentrations of pollutants and with air temperature during summer heat episodes.
Design: Daily data on all cause mortality and cause-specific (cardio-vascular and respiratory) mortality were obtained for the summers of 2003 and 2004. At the first stage, using t-test for
equality of means, we measured differences between daily mortality rates during hot days and average summer daily mortality rates. To quantify the relationship between temperature and
mortality, we compared mortality rates during the hot days with temperatures above 97th, 95th and 90th centile values of daily minimum temperatures. On the second stage, we investigated the
relationships between air temperature, NO2, PM10 and O3 pollution and daily mortality rates, using time-series Poisson regression technique with distributed lag of concentrations.
Main results: The rise in heat-related deaths in Moscow begins at about 21?C. For every one degree increase of daily minimum temperature above this value, total mortality increases by
1.9% (95% CI 1.2% to 2.6%), while cardio-vascular mortality increases by 4.1% (0.4% to 7.7%). Statistically significant increase of mortality from ischaemia and respiratory diseases with
temperature was not observed. Multiple correlation analysis showed that daily average concentrations of NO2, PM10 and O3 increased with temperature, and the power of association was very
strong: respective correlation coefficients R were 0.47, 0.57 and 0.49. Percentage change in all-cause mortality was 2.02% (1.12% to 2.92%) per 10 ppb increase in daily average measure of
ozone, and 1.23% (0.42% to 2.04%) per 10 ppb increase in daily average PM10 concentrations, using a lag of zero days. The corresponding values for mortality from ischaemia were respectively
3.0% (1.4% to 4.6%) per 10 ppb of ozone and 1.7% (0.2% to 3.2%) per 10 ppb of PM10, also with a lag of zero days. At the same time, daily mortality from cerebro-vascular and respiratory
diseases significantly increased only with ozone concentrations with a lag of 1 day. Corresponding relative risks were respectively 3.0% (1.3% to 4.7%) and 8.5% (0.7% to 16.3%) per 10 ppb
of increase of O3. Thus, the effect of pollution on respiratory mortality was the strongest.
Conclusions: In the conditions of autocorrelation among exposures to air pollutants and temperature, one may speak only about combined effects of hot temperature, ozone and PM10 on
mortality.




SAA4-PD-04                 IMPACT OF EXTREME WEATHER EVENTS AND CLIMATE VARIABILITY ON THE
                           HEALTH OF CANADIANS: A MULTICENTRE STUDY
W THOMPSON1, Y MAO1, S CHENG2, Y JIANG1
(1) Public Health Agency of Canada, Ottawa. (2) Environment Canada, Dufferin.
Relatively small changes in the average climate conditions could produce large changes in the frequency and magnitude of weather factors and events. In order to asses the effects of climate
change and climate variation on the health of a population, it is necessary to understand the relationship between health and climate under current and past conditions. It is important to
identify and quantify the health effects associate with climate and how these effects may vary by region or by population. The purpose of this project is to assess the prevalence of illness,
injury and death as a result of extreme heat and cold events through the collection and evaluation of administrative health data in the form of mortality, hospital separations, and emergency
department records from selected urban cities across Canada (Vancouver, Winnipeg, Edmonton, Ottawa, Quebec City, and Halifax).
The synoptic classification system by Environment Canada, assigns each day in terms of the various weather variables into a particular synoptic category based primarily on air mass
differentiation. Among the selected cities, Ottawa possessed the highest percentage of hot weather group (8%), followed by Quebec City (5.8%) and Winnipeg (3.6%), whereas Vancouver
had the least number of days in the hot weather group (0.6%). The cold weather group occurs most frequently in Winnipeg (16.9%) and least frequently in Vancouver (6.9%). The percentage
of air pollution days ranged from 46.6% (Quebec City) to 53% (Halifax). Elevated mortality and hospital separations are present when daily mortality exceeds the base-line and are associated
with extreme temperatures and acute exposures to air pollution. This pattern is consistent for all cities investigated. In addition to annual total counts, daily mean elevated mortality and
hospital separations breakdown by weather groups were also examine for the six cities. Daily elevated mortality was much higher for the extreme temperatures and air pollution-related groups
that it was for “other” (comfortable) weather groups.
Linking the regional/area health data to synoptic weather classifications of extreme heat and cols events over an approximate 10 year period (1990-2002) will provide new knowledge
regarding the vulnerability of certain populations and/or regions and establish the need for a surveillance system to monitor associated health impacts of climate variability. These results can
provide more accurate assessments of the health effects of climate change in Canada, a base measure for health service utilization during these extreme weather events and a scientific basis
for preventive and adaptation measures needed to policy and decision-makers. This is a joint project between Health Canada, Environment Canada, participating hospitals and universities
across Canada, partially funded by the Climate Change Adaptation Initiative, Natural Resources Canada.




SAA4-PD-05                 AN EVALUATION OF THRESHOLDS OF TEMPERATURE ON DAILY MORTALITY –
                           BASED ON SIX MAJOR CITIES IN SOUTH KOREA BETWEEN 1994 AND 2003
H KIM1, J HA1, J PARK2
(1) Seoul National University, Seoul. (2) Korea Environment Institute, Seoul.
Introduction
Dose response curve between temperature and morality is known to have “U” or “V” shape. And it has been well reported that the association between mortality and high temperature has
several forms based on the geographic area. This study was aimed to confirm the hypothesis that mild temperature has no effect on mortality but high temperature is a risk factor of mortality.
And the study was to compare thresholds of temperature on daily mortality for six major cities in Korea.
Methods
We collected mortality and meteorological data between 1994 and 2003 from the governmental agencies. After controlling for several confounders including the time-trends, relative humidity,
and other meteorological variables we evaluated the association between the total mortality excluding deaths due to accidents and the daily mean temperature. We fitted nonparametric
smoothing regression models to check the linearity of the association first and then fitted threshold models (combination of two linear models) to estimate the thresholds and the effects.
Results
Daily mortality was associated with the mean temperature with lag of one day. The threshold values are estimated as 29.7 C°, 28.9 C°, 27.9 C°, and 27.0 C° for four major cities. No threshold
effects were found in other two cities (in other words, linear relationships were confirmed in these two cities). Estimated percent increases of daily mortality (and their 95% CI) due to 1C°
increase in temperature with the range larger than the threshold are 16.25(14.17,18.39), 9.10(5.12,13.22), 7.01(4.42,9.66), and 6.73(2.47,11.17) for the four cities. Threshold values for those
who are elder than 65 were estimated 0.4~2.6 C° lower than those from the whole population. Results using heat-index (which uses humidity and temperature) are very consistent with our
results.
Discussion and Conclusions
Our research indicates the existence of temperature thresholds on mortality for several cities in Korea. It also indicates that threshold varies according to the geographical areas and
demographic characteristics of the population such as age. It is also noticed that in two cities (out of six) we didn’t observe the threshold effects of temperature on the total mortality. We
also estimated that up to 7.92 percent of the total deaths are related to the high temperature in 1994 in Seoul, Korea.




                                                                                              73
        Symposia Session:

SAA4-PD-06                 APPARENT TEMPERATURE AND INFLAMMATORY MARKERS IN A EUROPEAN
                           PANEL STUDY
A SCHEIDER1, D PANAGIOTAKOS2, S PICCIOTTO3, K KATSOUYANNI2, H LOEWEL1, B JACQUEMIN4, T LANKI5, M STAFOGGIA3, T BELLANDER6, W KOENIG7,
A PETERS1
(1) GSF-NATIONAL RESEARCH CENTER FOR ENVIRONMENT AND HEALTH, INSTITUTE OF EPIDEMIOLOGY, NEUHERBERG. (2) DEPARTMENT OF HYGIENE AND
EPIDEMIOLOGY, UNIVERSITY OF ATHENS, ATHENS. (3) LOCAL HEALTH AUTHORITY ROME, ROME. (4) IMIM-MUNICIPAL INSTITUTE OF MEDICAL INVESTIGATION,
BARCELONA. (5) ENVIRONMENTAL EPIDEMIOLOGY UNIT, NATIONAL PUBLIC HEALTH INSTITUTE (KTL), KUOPIO. (6) DEPARTMENT OF OCCUPATIONAL AND
ENVIRONMENTAL HEALTH, STOCKHOLM COUNTY COUNCIL & INSTITUTE OF ENVIRONMENTAL MEDICINE, KAROLINSKA INSTITUTE, STOCKHOLM. (7) DEPARTMENT
OF CARDIOLOGY, UNIVERSITY OF ULM, ULM.
Introduction: Inflammatory marker levels have been associated with cardiovascular risk, but the effect of climatological variables in this relationship is not well understood and appreciated.
Studies have shown an increase of fibrinogen in winter whereas C-reactive protein (CRP) has not been shown to be seasonally dependent so far. The objective of this study was to analyze the
association between apparent temperature, an indicator of perceived temperature based on ambient temperature and dew point temperature, and the inflammatory markers CRP, interleukin-
6 (IL-6) and fibrinogen in post-myocardial infarction (MI) patients.
Methods: In a multicenter panel study of MI survivors, the three inflammatory blood markers were measured repeatedly (up to 8 times) within a period of 13 months. In total 5835 blood
samples in 1006 subjects were collected in six European cities. Data on patient characteristics and disease history were gathered at the baseline visit. Meteorological data were obtained from
the country-specific network stations. The association was analyzed using a generalized additive model with random patient effects. The model variables were selected in each city separately
based on Akaike’s Information Criterion and consisted of time-invariant baseline variables and long-term time trend. Different lag-structures of the apparent temperature effect were tested.
For the analysis CRP and IL-6 had to be log-transformed. The effect estimates are presented as percent changes of geometric mean based on a 5°C decrease of apparent temperature.
Results: Patient mean levels of CRP ranged between 0.2 and 37.4 mg/l (mean ± standard deviation: 2.6±3.3), IL-6 had a range of 0.5 and 61.4 pg/ml (3.0±3.1) and levels of fibrinogen were
between 1.9 and 6.9 g/l (3.6±0.7). In the preliminary analysis the mean of the current and the previous day’s apparent temperature was associated with CRP. The city-specific effect estimates
were all positive except for one city. Although they ranged between -1.61% and 8.22%, no evidence for heterogeneity was observed. The pooled effect estimate was 1.52% change in CRP
(95%-confidence interval: [-0.03%; 3.10%]) in association with a 5°C decrease in apparent temperature. The effects on IL-6 and fibrinogen were heterogeneous between the cities and
therefore not pooled.
Discussion and conclusion: The results suggest that a decrease in apparent temperature leads to an increase in CRP-level. In susceptible patients this might lead to an additional risk for
cardiovascular events and provides a hint for the observed seasonal variation in death from ischemic heart disease and stroke in the elderly.



SAA4-PD-07                 EMERGENCY HOSPITAL ADMISSIONS OF CARDIOVASCULAR DISEASE AND
                           HEART FAILURE IN THREE AUSTRALIAN CITIES: ARE THEY ASSOCIATED
                           WITH CLIMATIC FACTORS?
R D’SOUZA1, H BAMBRICK1, R WOODRUFF1, A MCMICHAEL1
(1) Austalian National University, Canberra.
Introduction: Cardiovascular disease (CVD) is a major health and economic burden worldwide. It is a leading cause of death in Australia and individuals aged > 65 years account for 60% of
CVD hospitalisations. Congestive heart failure (CHF) caused 1.8% of deaths in Australia in 2003 and accounted for 9.2% of hospitalisations due to CVD. It is a major burden on the community
due to the costs of care and premature death. This study describes the seasonal pattern of emergency hospital admissions and mortality from CVD and CHF in three Australian cities, and the
association with weekly temperature.
Methods: We fitted log-linear models of daily hospitalisations and mortality for CVD and CHF (1994 to 2003) for people aged >65 years in three Australian cities (Brisbane, Canberra and
Melbourne) with the previous week’s average, minimum and maximum temperatures. A negative binomial chance model was used to accommodate over-dispersion and to control for day of
the week, season and long-term trends
Results: Seasonal variation in CVD hospitalisations differed less than mortality, which showed a marked increase during winter in all cities. Hospitalisations and CHF mortality also increased
during winter in Brisbane and Melbourne, while peaking earlier for hospitalisations in Canberra.
A decrement of one degree in the previous week’s minimum temperature was associated with a 1% increase in daily CVD admissions and a 2% increase in mortality in all cities. For CHF,
minimum temperature of previous week was associated with increased daily admissions in Brisbane (3%), Melbourne (2%). Minimum temperature of previous week was associated with
increased mortality in Canberra (6%), Melbourne (2%).
Average and maximum temperatures of the previous week were also significantly inversely associated with hospitalisations for CVD and CHF, and with CVD mortality. Average and minimum
temperatures had similar effect sizes, while maximum temperature had a smaller effect size.
Conclusions: Low temperatures were associated with higher CVD admissions and mortality. Cold temperatures may exacerbate CHF, induce myocardial ischaemia and precipitate arrhythmias.
Ischaemia and arrhythmias could further increase risk of heart failure decompensation. The elderly are less able to regulate body temperature and more at risk from cooling. Detecting and
correcting decompensation during cold periods may reduce the risk of infections that exacerbate CHF. The one week lag potentiates forecasting increases in risk and hospital use, and alerting
carers for the elderly to be watchful and to correct decompensation early.




SAA4-PD-08                 HEAT AND HOSPITAL ADMISSIONS AMONG U.S. ELDERLY
M O’NEILL1, A DIEZ ROUX1, A ZANOBETTI2, D BROWN1, H LEVY1, D SCHRAG2, R WILLIAMS3, J SCHWARTZ2
(1) University of Michigan, Ann Arbor. (2) Harvard University, Boston. (3) Cities for Climate Protection, Oakland.
This abstract was invited for a symposium organized by Sari Kovats and Bettina Menne. Dr. Kovats was made aware that results were not yet available but still considered the study design
and conceptual framework to be of interest. Approval for use of confidential human data has just been received and preliminary results could be available by September. We present a
conceptual approach and study design on multiple levels of vulnerability to heat-related morbidity among elderly people.
Introduction: Biomedical, socio-economic and community determinants of vulnerability to the impacts of climate change on health are all relevant for developing preventive programs. Yet
few studies have simulaneously evaluated these multiple levels, in multiple cities, with a consistent statistical approach. A new study of heat and hospital admissions among the elderly in U.S.
cities will do this in cooperation with a non-profit group representing 157 cities.
Methods: Poisson regression will be used to determine whether excess hospital admissions occur among elderly people during hot weather in 34 U.S. cities, from cardiovascular, respiratory,
diabetes and heat-related causes. Next, effect modification will be evaluated by stratifying the data by :
- co-morbid conditions (diabetes; cardiovascular, respiratory, renal disease; overall frailty)
- individual characteristics (race, gender, age, use of public vs. private hospital). Finally, a second stage analysis will examine whether the associations differ according to several city-wide
characteristics (percent poverty, percent with college education, percent of non-white population, air conditioning (AC) prevalence, percent green space, housing characteristics, air pollution
concentrations, weather variability, city preventive programs).
The economic impact of these admissions will be quantified under different climate change scenarios and adaptive/mitigative strategies, addressing equity concerns. Results will be
disseminated to city officials in order to foster and inform preventive actions and policies.
Results: Data on Medicare hospital admissions in 34 U.S. cities from 1985-2003 will be combined with environmental data and city-wide characteristics (Census socioeconomic data, American
Housing Survey information on housing type and AC prevalence, percent green space from satellite photos, government preventive programs, weather variability and air pollution levels).
The costs of admissions will be estimated from billing records, and the future effect of several potential climate change, and societal activities to reduce heat exposure, on these costs will be
estimated. Results will be incorporated into a software planning module for use by local governments and a workshop with local officials held to discuss results and preventive strategies.
Discussion and conclusions: Local public health officials have requested more quantitative information on how climate change may affect human health, and this study will provide data
to support risk assessment and efforts in prevention and outreach. This interdisciplinary initiative is a unique partnership between researchers and city officials, intended to enhance
understanding on how heat and morbidity associations vary by community, medical and socio-economic characteristics, and foster preventive action.




                                                                                               74
                                                                                                                                                                                                    SUNDAY SEPTEMBER 3
        Symposia Session:

SAA4-PD-09                 THE RELATIONSHIP BETWEEN SUMMER DAILY TEMPERATURE AND
                           MORTALITY IN FRANCE: 1975 TO 2003
A FOUILLET1, G REY1, E JOUGLA2, D HEMON1
(1) INSERM-U754, VILLEJUIF. (2) INSERM-CepiDc, LE VESINET.
Introduction: Several studies have modelled the relationship between climatic factors and mortality using a time-series approach or by analyzing specific heat waves. The results depend on
the site of those studies and its climatic characteristics.
Our objective was to define a combination of climatic factors enabling optimum prediction of the daily fluctuations in mortality in summer.
Material: The study addressed French mortality and temperature data from June to September for the years 1975-2003. Mortality was quantified in terms of the daily mortality rates for
subjects aged over 55 years.
The national daily minimum and maximum temperatures consisted in the population-weighted average values recorded by 97 weather stations representative of the French départements.
For each day, a cumulative variable for minimum/maximum temperature was defined as the sum of the numbers of degrees above a cut-off point over the preceding 10 days, providing that
the temperature remained higher than the cut-off.
The daily mortality rate was modeled using a Poisson regression and controlling for a long-term mortality trend. An autoregressive structure for the residuals and an over-dispersion term were
added. Lagged effects of temperature were accounted for by including the 5 preceding days in the model. A “backward” method was used to select the most significant climatic variables.
The goodness-of-fit was measured on the basis of the over-dispersion and the first-order autocorrelation of the residuals. The estimated and observed mortality rates were also compared using
a validation group of years.
Results: The climatic factors had both auto- and cross-correlative structures. The over-dispersion of the model (M0) without any climatic factors was close to 5.6 and of the order of 2.5 in the
final model. Residuals were shown to have a weak first-order autocorrelation.
The main part of the daily fluctuations in mortality in summer was explained by temperatures observed on the current and preceding days. Prediction of mortality during extreme events was
highly improved using the cumulative variables for maximum temperatures.
The correlation between the observed and estimated mortality rates increased from 0.65 with M0 to 0.92 with the final model.
Discussion and conclusions: The first-order correlation of the residuals suggests that other factors may have a significant impact on daily mortality in summer, such as humidity or air
pollution. A spatial-time model is also required in order to take geographic heterogeneity into account.
The final model may contribute to enhancing the alert system for extreme climatic events such as heat waves.




SAA4-PD-10                 MODIFIERS OF THE EFFECT OF EXTREME TEMPERATURES ON MORTALITY: A
                           MULTI-CITY CASE ONLY ANALYSIS
M MEDINA-RAMON1, A ZANOBETTI1, DP CAVANAGH2, J SCHWARTZ1
(1) Harvard School of Public Health, Boston, MA. (2) Massachusetts Department of Public Health, Boston, MA.
INTRODUCTION: An increase in average temperature and changes in the frequency of temperature extremes are predicted to occur over the next century. Extremes of temperature are
associated with short-term increases in daily mortality. We aimed to identify subpopulations with increased susceptibility to temperature extremes and explore differences in susceptibility
according to the primary cause of death. We also aimed to identify specific mortality causes that experience greater relative increases on extreme temperature days.
METHODS: We conducted a case-only analysis using daily mortality and hourly weather data from 50 US cities for the period 1989-2000, covering a total of 7,531,120 deaths. Distributions
of daily minimum and maximum temperature in each city were used to define extremely hot days (≥ 99th percentile) and extremely cold days (≤1st percentile), respectively. We examined
whether the effect of extreme temperatures on mortality was modified by socio-demographic characteristics, medical conditions, placement of death and specific mortality causes. For each
(hypothesized) effect modifier, a city-specific logistic regression model was fitted and an overall estimate was calculated in a subsequent meta-analysis.
RESULTS: In general, effect modification by personal characteristics was modest, with the strongest associations observed for extreme heat. Particularly, older subjects (odds ratio (OR) 1.018
(95% confidence interval: 1.004, 1.033), diabetics (OR 1.036 (1.009, 1.063)), blacks (OR 1.043 (1.019, 1.068)) and those dying outside a hospital (OR 1.068 (1.037, 1.100)) were more
susceptible to extreme heat. Differences in susceptibility were observed according to the primary cause of death and this was especially noticeable for black subjects, whose susceptibility to
extreme heat was particularly pronounced when dying from a cardiovascular disease (OR 1.119 (1.075, 1.164)), and not so much when dying from other causes (OR 1.015 (0.992, 1.039)).
Cardiovascular deaths (OR 1.052 (1.036, 1.069)), and especially cardiac arrest deaths (OR 1.134 (1.052, 1.222)), showed a greater relative increase on extremely cold days; while the increase
in heat-related mortality was higher for those with coexisting atrial fibrillation (OR 1.063 (1.002, 1.128)).
DISCUSSION AND CONCLUSIONS: we identified several subpopulations and mortality causes with increased susceptibility to extreme temperatures using a relatively new approach and in
a larger sample of cities than has been previously examined. Our results may contribute to establish health programs that would better protect the vulnerable, especially in the light of a
differential susceptibility according to the cause of death.




SAB3-PD                    DIETARY EXPOSURES
SAB3-PD-01                 ASSOCIATIONS BETWEEN GASTRIC CANCER AND DIETARY EXPOSURES
L DREGVAL1, J ZICKUTE1, L STRUMYLAITE1, J DUDZEVICIUS1
(1) Institute for Biomedical Research of Kaunas University of Medicine, Kaunas.
Introduction. Gastric cancer is the second cause of all cancers death in Lithuania. Dietary factors are estimated to account for up to 30% of cancers in developed countries making diet second
only to tobacco as a preventable cause of cancer. Diet is especially important factor in gastric cancer etiology.
Aim. To assess the risk of gastric cancer in relation to dietary exposures.
Methods. Hospital based case-control study was carried out in Lithuania. The study included 379 cases aged 30-75 with gastric cancer and 1137 hospital controls (1:3) that were cancer and
stomach diseases free. Cases and controls were matched by gender and age. A questionnaire was used to collect information and included questions on dietary habits. Diet estimated by
consumption frequency of different food items. Data were analyzed using the logistic regression model.
Results. 71.3% of cases and 43% of controls reported they liked salty food. Adding salt to prepared meal indicated 73.1% of cases and 46.5% of controls. According to univariate regression
the risk of gastric cancer among persons who liked salty food was 3 times higher than among those who did not: for salty food OR=3.3; 95% CI=2.7-4.3 and for adding salt to meal OR=3.1;
95% CI=2.4-4.1. After adjustments for alcohol use, smoking, family history on cancer, BMI at 20 yr of age, educational level, residence, physical activity and other dietary habits the risk
remained about 3 times higher for those who liked salty food (OR=2.8; 95% CI=2.0-3.8, p for trend<0.001) and added salt (OR=2.7; 95% CI=2.0-3.7, p for trend<0.001). The consumption
of salted meat, smoked meat and fish ?1-2 times/week after adjustments (mentioned above) was related to higher risk of gastric cancer: OR=2.2 (95% CI=1.4-3.4), OR=1.8 (95% CI=1.2-2.6)
and OR=1.7 (95% CI=1.1-2.5), respectively. A high consumption of raw vegetables in summer (?7 times/week) and winter (?3 times/week) was associated with a significantly decreased risk
of gastric cancer (OR=0.7; 95% CI=0.5-0.9, p for trend=0.036 in summer and OR=0.5; 95% CI=0.4-0.7, p for trend<0.001 in winter). The intake of raw vegetables such as broccoli (?1-3
times/month vs. almost never), cabbages (?1-3 times/month vs. almost never), carrots (?1-2 times/week vs. ?1-3 times/month), tomatoes (?1 time/day vs. ?1-3 times/month), garlic (?1-4
times/week vs. ?1-3 times/month) and beans (?1-4 times/week vs. ?1-3 times/month) after adjustments significantly decreased risk of gastric cancer from 2 (broccoli) to 4 (cabbages) times.
Conclusions. The associations between gastric cancer and dietary exposures were observed. Salt and salt processed food increased risk of disease. The association between gastric cancer
and salty food support previous findings in different countries. According to our data frequent consumption of raw vegetables was related to a lower risk of disease though some studies did
not find a significantly reduced risk of gastric cancer in relation to high intake of raw vegetables.
PLEASE DO NOT MODIFY THE TEMPLATE FRAME
otherwise the abstract will be rejected by the Programme Committee



                                                                                              75
        Symposia Session:

SAB3-PD-02                 IDENTIFICATION AND QUANTIFICATION OF URINARY METABOLITES FROM
                           SHORT TERM EXPOSURE TO ACRYLAMIDE
T BJELLAAS1, K JANÁK1, E LUNDANES2, L KRONBERG3, G BECHER1
(1) Norwegian Institute of Public Health, Division of Environmental Medicine, Oslo. (2) University of Oslo, Department of Chemistry, Oslo. (3) Åbo Akademi University,
Department of Organic Chemistry, Turku/Åbo.
Introduction
It is known that heat treated carbohydrate rich foods may contain high levels of acrylamide (AA), and concentrations up to 4000 µg/kg in potato crisps and 2000 µg/kg in french fries have
been reported. In order to get more information on human exposure to AA and its metabolism, a method for determination of known urinary metabolites from dietary exposure to AA has
been developed, utilizing solid phase extraction and liquid chromatography with positive electrospray MS/MS detection.
Methods
The urinary metabolites were synthesized and their structures were determined by NMR and MS. To test the method, a pilot study was conducted, in which all urine from 6 volunteers (1
smoker) was collected during 48 hrs starting with 24 hrs fasting. The chromatography was performed on a graphitized (Hypercarb) column with dimensions 2.1 x 50 mm and 5µm particles
size. For detection a triple quadrupole mass spectrometer operated in multiple reaction monitoring mode was used.
Results
The assay was validated in the range from 8.6 µg/L to 342.9 µg/L.The two urinary metabolites, N-acetyl-S-(3-amino-2-hydroxy-3-oxopropyl)cysteine (MA-GA3) and N-acetyl-S-(3-amino-3-
oxopropyl)cysteine (MA-AA), were found above the detection limit. Fasting during one day caused about a 50% decrease in the total level of the metabolites. However, after one day of a
normal diet, the urinary metabolite levels had returned to pre-fasting levels. The total amount of acrylamide excreted over the period acrylamide in the form of urinary metabolites was
estimated to be about 40 µg AA per day for the average non-smoker.
Discussion and conclusions
The urinary metabolites were baseline separated on HPLC, and the use of tandem mass spectrometry resulted in high selectivity and sensitivity. The significance of fasting on the reduction of
the urinary metabolites makes it plausible that the major part of the AA, excreted as urinary metabolites, originates from the diet. It has been demonstrated that the method has potential for
biomonitoring the short term exposure to acrylamide in humans.




SAB3-PD-03                 DIETARY ARSENIC EXPOSURE FOR FEMALE HEADS OF HOUSEHOLDS IN
                           BANGLADESH
M.L KILE1, E.A HOUSEMAN1, T.J SMITH1, J.J HARRINGTON1, Q. QUAMRUZZAMAN2, M. RAHMAN2, G. MAHIUDDIN2, DC CHRISTIANI1
(1) Harvard School of Public Health, Boston. (2) Dhaka Community Hospital, Dhaka.
Introduction: Arsenic-contaminated groundwater is widespread throughout Bangladesh and provides the majority of the nation’s drinking and irrigation water. However, little is known
regarding diet as a route of exposure or how it contributes to internal dose.
Methods: In 2004, a 6-day duplicate diet study was conducted in Pabna, Bangladesh in a cohort of female heads of households (N=47). Total arsenic was measured in 24-hour composite
food samples using an inductively coupled plasma mass spectrometer (ICP-MS) equipped with a dynamic reaction cell. Drinking water arsenic concentrations were measured using 12 samples
per household using ICP-MS. Intake rates, defined as the volume of water and weight of food consumed within a 24-hour period, were multiplied by their respective drinking water and 24-
hour food composite arsenic concentration to calculate dietary intake. Separate log-log multivariate regression models examined the relationship between toenail arsenic, a biomarker of
internal dose, drinking water, total dietary intake, and total dietary dose where dietary variables were adjusted using the residual method to provide estimates that were independent of water
arsenic concentrations.
Results: Median arsenic concentrations in food and water were 0.03 •g As/g wet weight (90th percentile: 0.09) and 1.5 •g As/L (90th percentile: 143.3). The median daily intake of food and
water was 1620 g wet weight and 2530 mls, resulting in median daily arsenic intakes of 48.0 •g/day (90th percentile: 133.6) from food and 4.2 •g/day (90th percentile: 356.1) from water. The
median total daily arsenic dose was 1.9 •g/kg body weight-day (90th percentile: 4.1). The regression model that used total daily arsenic intake from food (•=0.46; 95%CI: 0.18-0.73) and
drinking water (95%CI: 0.26-0.38) explained the most variability in toenail arsenic concentrations (R2a =0.71).
Conclusions: Drinking water exposure was highly skewed and was the dominant exposure route for the upper 25th percentile of the distribution. However, dietary exposure was high and
provided the primary source of exposure for three quarters of the participants. Overall, 38% (95% C.I:25-54%) of the participants exceeded the World Health Organization’s provisional
tolerable daily intake of 2 •g As/kg- day, of which 12% (95%C.I.:3 - 27%) used drinking water sources that were below the Bangladesh standard of 50 •g As/L. Furthermore, total daily arsenic
intake from food and drinking water had similar effect estimates suggesting that both sources have a similar contribution to internal dose.
Funding Source: Supported by ES 04957 and ES 00002.




SAB3-PD-04                 PREDICTION AND EVALUATION OF DIETARY ON ARSENIC EXPOSURES USING
                           THE SHEDS MODEL
J XUE1, V ZARTARIAN1, H ÖZKAYNAK1
(1) US EPA, Boston.
Arsenic (As) is one of the important environmental contaminants with many documented health hazards. Dietary intake is one of the most important exposure pathways for arsenic.
Approximately 40,000 person-days of data from the USDA’s 1994-1996 Continuing Survey of Food Intake by Individuals (CSFII) and arsenic residues from the FDA’s Total Dietary Survey (TDS)
were combined in EPA’s Stochastic Human Exposure and Dose Simulation (SHEDS) model to estimate daily average dietary As exposure. Results ranged from approximately 0.2 to 0.5 ug/kg/day
for different age groups, with higher As exposures predicted for the younger age groups (ages 1-5 years), mainly due to their lower body weight. Because of their special dietary habits, Asian
Americans were predicted to have the highest dietary As exposure among the other groups studied (1 ug/kg/day, more than three times that of Caucasians). The dietary As exposure for people
living in urban areas was estimated to be about 1.6 greater than those living in rural areas on average. The dietary As exposure for populations from the Northeastern U.S. was found to be
higher than those from other regions. The model was used to identify the 50 most important contributing food items, which contribute averagely more than 50% As dietary exposure. Seafood
and rice were found to be the two most important food categories contributing to dietary As exposure. Evaluation of the SHEDS dietary model with duplicate food samples for 102 subjects
from the National Human Exposure Assessment Survey (NHEXAS) showed that model results compare well against observed results.
Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy.




                                                                                             76
                                                                                                                                                                                                     SUNDAY SEPTEMBER 3
        Symposia Session:

SAB3-PD-05                 LOCAL DIETARY CONTRIBUTIONS TO SERUM DIOXINS, FURANS, AND
                           POLYCHLORINATED BIPHENYLS IN PERI PUBERTAL RUSSIAN BOYS
J BURNS1, P WILLIAMS1, O SERGEYEV2, E GITIN1, T DENISOVA3, S KORRICK4, M LEE5, B REVICH6, L ALTSHUL1, D PATTERSON7, W TURNER7, A
RONNENBERG8, R HAUSER1
(1) HARVARD SCHOOL OF PUBLIC HEALTH, BOSTON. (2) SAMARA STATE MEDICAL UNIVERSITY, SAMARA. (3) CHAPAEVSK MEDICAL ASSOCIATION, CHAPAEVSK. (4)
HARVARD MEDICAL SCHOOL, BOSTON. (5) UNIVERSITY OF MASSACHUSETTS MEDICAL SCHOOL, WORCESTER. (6) RUSSIAN ACADEMY OF SCIENCES, MOSCOW. (7)
CENTER FOR DISEASE CONTROL AND PREVENTION, ATLANTA. (8) UNIVERSITY OF MASSACHUSETTS, AMHERST.
Introduction: In Chapaevsk, Russia, where there is high environmental exposure to polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and polychlorinated
biphenyls (PCBs), we assessed the dietary contribution of locally produced foods to serum levels of these compounds in boys.
Methods: Our study sample included 123 boys aged 8-9 years enrolled in 2003. The boys’ initial study visit included physical exams, blood sampling, and completion of semiquantitative food
frequency questionnaires by parents or guardians. The blood samples were analyzed for dioxin-like compounds (PCDDs/PCDFs/PCBs) by the Centers for Disease Control and Prevention (CDC).
In preliminary analyses, we assessed the associations between categories of local food consumption (fruits and vegetables, meats, poultry, fish, dairy, and eggs) and the log10-transformed sum
of PCDDs/PCDFs/coplanar-PCBs using general linear regression models, adjusted for parental education and duration of breast feeding. We then translated estimated changes back to the
original scale to calculate percent change in the sum of serum PCDDs/PCDFs/coplanar-PCBs. For food categories with more than one food item, each food item was scored 0 (no local
consumption) or 1 (any local consumption) without consideration of quantity consumed. We then computed a percentage of consumed local food (%LF) in each food category for each boy.
For example, in the fruit/vegetables category (18 items) if a boy had local consumption of 7 fruit/vegetables items his %LF would be 7/18 (39%).
Results: The median (25th, 75th percentile) concentration for total PCDDs/PCDFs/coplanar-PCBs was 502.8 (430.7, 655.1) pg/g lipids. Fruits and vegetables had the highest %LF (median 39%).
Meats had the lowest %LF (median 0%), with only 7 boys eating any local meats. The consumption of local eggs was associated with a 26.1% (95% CI: 6.6%, 49.3%) higher serum total
PCDDs/PCDFs/coplanar-PCBs. A 10 unit increase in %LF of fruit/vegetables was associated with a 3.8% (95% CI: 0.9%, 6.8%) increase in serum total PCDDs/PCDFs/coplanar-PCBs. %LF of
meat, poultry, fish, or dairy food consumption was not associated with higher serum total PCDDs/PCDFs/coplanar-PCBs.
Discussion and Conclusions: In this cohort of Russian boys, consumption of local eggs and produce were predictors of serum PCDDs/PCDFs/coplanar-PCBs levels. The latter finding was
unexpected and may be due to confounding by correlates of produce intake and exposure misclassification since local food consumption was not quantified. Future multivariate analyses are
planned to better elucidate the potential independent role of diet as a source of exposure to PCDDs/PCDFs/PCBs, expressed as both serum concentrations and total TEQs.
Funded by EPA Grant R82943701 & NIEHS GRANT 5T32-ES007069-25.




SAB3-PD-06                 IMPACT OF SELF PRODUCED FOODSTUFFS ON THE MANAGEMENT OPTIONS TO
                           REDUCE FOOD EXPOSURE TO CHLORDECONE OF THE MARTINICAN
                           POPULATION
C DUBUISSON1, F HÉRAUD1, C FLAMAND2, JC LEBLANC1, S GALLOTTI1, A BLATEAU2, P QUENEL2, JL VOLATIER1
(1) Agence Française de Sécurité Sanitaire des Aliments, Maisons-Alfort. (2) CIRE Antilles - Guyane, Fort-De-France.
Introduction: Chlordecone is an organochlorine insecticide formerly used in the banana plantations of the French Antilles. Even though the use of Chlordecone was banned in 1993, recent
surveys revealed its wide presence in soils, rivers and domestic food products, raising questions on the risk posed to the Martinican population. The aim of this study is to assess this
population’s food exposure and then to analyse the impact of different management options. As an appreciable part of the food consumed in Martinique corresponds to subsistence production,
this study proposes to modulate the exposure models according to the food supply habits.
Methods: Consumption data derive from the ESCAL Survey carried out in 2004 on 1814 subjects aged over 3 years. This survey also provides information on the households’ food supply
habits. Residues data come from the National Monitoring Programmes 2002-2004 (n=822). Assumptions on food contamination levels are made depending on whether the subjects live on
a soil-contaminated place or not and on their supply habits (subsistence production, local market and groceries versus bigger stores). Food exposure is assessed though deterministic models
and then compared to a Toxicological Reference Dose previously established at 0.5 g / kg bw / day. The impact of various Maximum Limits introduced in the contamination distribution of
the main food contributors is then analysed on the mean and P95 exposures, firstly wherever the foodstuffs have been supplied, secondly only on commercialised products.
Results: Exposure is greater in children than in adults with respective mean exposures of 0.095 and 0.076 g / kg bw / day. The probability of exceeding the TRD is respectively 20.9% [6.2 ;
34.4] and 15.6% [9.6 ; 20.8] for children and adults living in a soil-contaminated area and null for the remaining population. Roots and tubers are the major foodstuffs contributing to the
total intake. MLs below 300 ppb would reduce significantly the exposure. When such MLs are applied to both self-produced and commercialised foodstuffs, the probability of exceeding the
TRD is non statistically different from zero. It is no more the case if they are implemented only on commercialised products.
Discussion and conclusions: For the first time, this study shows the supply habits, particularly subsistence production, may have significant impacts on food exposure to contaminants like
Chlordecone. It reveals that setting MLs, which can only be controlled on commercialised products, are not sufficiently protective and other management options like consumption
recommendations are needed to ensure the safety of local consumers population. Nevertheless this finding must be interpreted cautiously due to the assumptions made to link the product
origin with its contamination level. More investigations are needed to characterize it better.




SAB3-PD-07                 NUTRIENT DEFICIENCY CLASSIFICATION BIAS RELATING TO THE 24 HOUR
                           DIETARY RECALL: A CASE STUDY WITH DIETARY AND RED BLOOD CELL
                           FOLATE
L BARRAJ1, N TRAN1
(1) Exponent, Washington.
The 24-hour recall data from food consumption surveys, such as the US Continuing Survey of Food Intake by Individuals (CSFII) or the NHANES, are frequently used to estimate dietary exposure,
which is often compared with recommended daily intake (RDI) for micronutrient to classify “nutrient deficiency status” for various population sub-groups. However, the 24-hour recall method
is not ideal for estimating long term average daily food intakes since there are considerable variations in day-to-day intake of food within individuals. Single-day intake data usually result in
a distribution that is flatter and wider than the true distribution of usual intakes of individuals in the population. Thus, the prevalence of high or low intakes (under or over nutrition) is
overestimated with single 24-hour recall. The UK National Diet and Nutrition Surveys collect multiple days of intake data and could provide a mean to quantify the effects of intra-individual
variation on estimates of usual food intakes. Individual’s dietary folate intakes for day one through day 7 are estimated and folate status is established based on different combination of
intake days and number of days prior to blood sampling. Using the red blood cell (RBC) folate data collected in the UK survey, individual’s folate deficiency status is also defined. Red blood
cell (RBC) folate concentration is believed to be a better index of tissue folate storage than serum folate, i.e. longer-term measure of folate status. Classification of folate deficiency status
based on the dietary intake estimates is compared with classification based on RBC folates. Bias in the classification of folate status using the single 24-hour recall is discussed.




                                                                                               77
         Symposia Session:

SAB3-PD-08                  UNDERESTIMATION OF ADVERSE EFFECTS OF MERCURY EXPOSURE WITHOUT
                            ADJUSTMENT FOR BENEFICIAL EFFECTS OF SEAFOOD CONSUMPTION
E BUDTZ-JØRGENSEN1, P WEIHE2, P GRANDJEAN3
(1) Department of Biostatistics, Institute of Public Health, University of Copenhagen, Copenhagen. (2) Department of Occupational and Environmental Health, The
Faroese Hospital System, Torshavn. (3) Department of Environmental Medicine, University of Southern Denmark, Odense.
Introduction: Methylmercury is a common contaminant of seafood and occurs in varying concentrations in different types of seafood, which also contain essential nutrients in less variable
amounts. The effects of methylmercury exposure from the diet may be ameliorated to some degree by beneficial effects of nutrients. We have examined this issue in a mercury-exposed birth
cohort from the Faroe Islands.
Methods: A cohort of 1022 children was recruited at birth in 1987 and 1988 in the Faroe Islands, where the diet includes frequent seafood dinners and occasional consumption of whale
meat. The children’s prenatal mercury exposure was determined from mercury concentrations in maternal hair and cord blood. At 7 years and again at 14 years, the children underwent a
detailed neurobehavioral examination.
In the statistical analysis, we used structural equation modeling to include allowance for the beneficial effect of seafood intake, while at the same time adjusting for imprecise exposure
assessments. We thereby extended previous analyses of the data from the 7-year examinations and the recently collected data from age 14 years. The adjustment for exposure imprecision
was achieved by including three independent exposure variables (mercury in cord blood, mercury in maternal hair, and maternal whale meat consumption). Frequency of maternal fish dinners
was added as a separate covariate reflecting beneficial nutrient intakes.
Results: At 7 years, outcome variables were grouped into motor functions and verbally-mediated functions that were each joined into a latent variable in the structural equation. Both groups
of tests showed significant mercury-associated deficits. Tests conducted at 14 years were also split into groups. After adjustment for fish intake, at both occasions, the mercury-associated
effects increased, especially in regard to motor skills. The results also suggested a positive effect of fish intake. However, fish intake was based only on maternal recall of the frequency of fish
dinners and was likely to be an imprecise indicator of the true intake of beneficial nutrients. Sensitivity analyses showed that non-directional imprecision of this confounder added to the
underestimation of the mercury effect.
Discussion and Conclusions: The findings suggest that, without mutual adjustment for beneficial and toxic effects, and without allowance for imprecision, the opposite effects of mercury
and nutrient intakes from seafood are both underestimated. Assessment of adverse effects caused by toxicants therefore needs to adjust for both confounding and imprecision of the exposure
variables.




SAB3-PD-09                  PERSISTENT ORGANIC POLLUTANTS IN FATTY FISH – AN UNHEALTHY
                            CONCERN
T SANDANGER1, M BRUSTAD2, C HOVANDER2, E LUND2
(1) Norwegian Institute for Air Research, Tromsø. (2) Institute of community Medicine, University of Tromsø, Tromsø.
Introduction: Recent research has again raised the issue regarding the potential health effects of long term exposure to persistent organic pollutants (POPs) in the context of a risk-benefit
analysis compared to the health benefits of consuming fatty fish. There is clear evidence of the beneficial effects of fish consumption on public health and in particular fatty fish with its
essential fatty acids. In addition to providing fatty acids, fish fat constitutes an essential source to vitamin D and A. Norwegian Food Control Authorities have recently recommended that
children and women of childbearing age not to consume fish liver due to the risk associated with high intake of POPs.
Methods: The results presented here are from 4 cross sectional studies in selected rural coastal communities where consumption of fish, fatty fish and fish liver is considerable. The number
of participants was, 35, 33, 50 and 60 in the respective studies. The participants were of both genders and between the age of 29 and 70 years. The level of PCBs and pesticides were determined
plasma from all participants. All participants further answered a detailed dietary questionnaire in order to link dietary habits to the levels of POPs.
Results: These levels of PCBs and pesticides were not significantly affected by the reported intake of fish liver, or fatty fish when age and gender were considered. Age was a significant
predictor for most compounds and gender was significant for several of the compounds. When merging the datasets with an urban population with considerable lower intake of fish liver, the
intake of cod liver still did not significantly affect the levels of PCBs and p,p’-DDE when adjusted for age, gender and place.
Conclusions: The levels of POPs in human blood were not predicted by reported frequency of intake of fish liver or fatty fish but rather by age and gender in any of the study samples. In risk
assessments, that lead to advises to the population about intake of traditional marine food items, a approach that embraces both positive and negative health aspects regarding this diet is
recommended.




SAB3-PD-10                  A RISK ANALYSIS OF THE LEVEL OF RISK TO CRYPTOSPORIDIUM THROUGH
                            CONSUMING OR HAVING CONTACT WITH CONTAMINATED FISH AMONG
                            BALTIMORE URBAN ANGLERS IN MARYLAND
J ROBERTS1, E SILBERGELD2, T GRACZYK2
(1) CHEMRISK, INC, SAN FRANCISCO. (2) JOHNS HOPKINS BLOOMBERG SCHOOL OF PUBLIC HEALTH, BALTIMORE.
Introduction: Urban fishing, the practice of fishing in urban waterways for recreational or subsistence purposes, is an under-recognized environmental health issue. In urban areas, the
environmental factors that are especially troublesome are due to exposure from microbiological contaminants, such as Cryptosporidium. This study examines the risk of exposure to
Cryptosporidium based on fish and hand wash samples taken from Baltimore urban anglers during the summers of 2002 and 2003.
Methods: Using a questionnaire, information regarding fishing frequency and consumption were collected through on-site interviews at various urban waterbodies in the Baltimore area.
Anglers were identified by locating sites that were commonly visited for fishing. Hand wash samples of the anglers’ hands were collected using a solution of PBS eluting fluid with 0.1% Tween
80, 0.1% SDS, and 0.001% of antifoam agent. Fish samples were also collected from the anglers and caught by the researchers. The probability of infection of Cryptosporidium was determined
using the U.S. Environmental Protection Agency’s (EPA) dose-response model for Cryptosporidium and a Monte Carlo analysis.
Results: There were a total of 48 fish and hand wash samples collected and 56 anglers interviewed. Based on the hand wash samples and the information provided from the urban angler
questionnaire, it was determined that the probability of infection could be as high as one indicating that infection would occur, if proper hand washing did not occur. Even with the most
conservative estimates the probability of infection based on the fish and hand samples was 0.22 and 0.41, respectively.
Conclusions: Many of the anglers interviewed during this research were unaware of this risk because currently most states, including Maryland, provide no information in regard to exposure
to Cryptosporidium or any microbiological contaminant within their fish consumption advisories. This research provides enormous implications for the necessary changes and improvements
that need to be made in order to improve and adequately inform the anglers of these potential health risks. Furthermore, this research embarks on a new frontier of research that not only
Maryland can benefit from, but other states who share similar problems with microbiological contamination of their urban watersheds.




                                                                                                78
                                                                                                                                                                                                       SUNDAY SEPTEMBER 3
        Symposia Session:

SAB4-PD                     AIR POLLUTION÷ MECHANISMS AND HEALTH
SAB4-PD-01                  HEALTH IMPACT OF ROAD TRAFFIC EMISSIONS, FORECASTED DECREASE
                            ASSOCIATED TO IMPLEMENTATION OF EUROPEAN STANDARDS IN FRENCH
                            VEHICLE FLEET
V NEDELLEC1, L MOSQUERON1, H DESQUEYROUX2, N JEANNÉE3, B GUILLAUME4, C LIOUSSE4, R LAGACHE5
(1) Vincent Nedellec Consultants, Paris. (2) ADEME, Agence de l’Environnement et de la Maitrise de l’Energie, Paris. (3) Géovariance, Avon. (4) Laboratoire d’Aérologie,
UMR 5560, UPS/CNRS, Toulouse. (5) LRPC-Lille CETE Nord-Picardie, Lille.
Introduction: Health Impact Assessment (HIA) of road transports emissions, based on exposure response functions (ERF) from epidemiological studies, have influenced new European
standards for terrestrial vehicles emissions. This study aims at quantifying the health impact decrease attributable to European Standards Implementation in France between 2000 and 2010.
Methods: Urban exposure to road traffic emissions (RTE) was evaluated with data from French air quality-monitoring network during the year 2000. The concentrations in NO2 and PM10
were estimated by a geo statistic approach in grid cells of 4 km?. Modelling the emissions of stationary and mobile sources refined this approach. The second approach allows estimating the
contribution ratio of RTE in the total atmospheric concentrations. The population data in each cell and stratified by age, were extracted from the 1999 French census. HIA method followed the
WHO and InVS methodological recommendations. Health impacts (HI) were estimated with European epidemiological ERF derived for elderly, adults and children. For children, new health
outcomes such as pre-term birth or childhood cancers were included. The morbidity and mortality rates for general population came from French sanitary statistics. Comparison of the results
between 2000 and 2010 made it possible to quantify the reduction associated with the introduction of new standards of emission.
Results: In urban area, mean ambient concentration (all sources) should have decreased by 27% for NO2 and 28% for PM10 during the 2000-2010 period. However, RTE contribution rates
will follow a lower decreasing trend for both pollutants (respectively 18%, 20%). In 2000, main HI due to RTE are long-term mortality (7 700 cases per year), cardiovascular hospitalization
(10 600) and short-term mortality in adults (3 600). In children majors HI are bronchitis (258 000), asthma attacks (47 200) and prematurely birth (11 765). Contribution of RTE to HI results
will decrease in 2010 by 17% to 49% according to health outcomes.
Conclusions: Major uncertainties, except those in the exposure assessment, come under ERF used for children. Evidences of causality between certain health effects and the air pollution have
not yet been strongly shown. ERF for childhood cancer and pre-term birth are based on single epidemiological study. Moreover we were unable to take demographic and population health
evolutions into account. HIA results plead for accompanying actions on traffic flux that will increase emissions standard benefits for population health.




SAB4-PD-02                  THE ASSOCIATION BETWEEN PARTICULATE AIR POLLUTION AND
                            HOMOCYSTEINE: THE VA NORMATIVE AGING STUDY
SK PARK1, M O’NEILL2, P VOKONAS3, D SPARROW3, A SPIRO III3, K TUCKER4, H SUH1, H HU1, J SCHWARTZ1
(1) Department of Environmental Health, Harvard School of Public Health, Boston. (2) Department of Epidemiology, University of Michigan School of Public Health, Ann
Arbor. (3) VA Normative Aging Study, Veterans Affairs Boston Healthcare System, Boston. (4) Jean Mayer USDA Human Nutrition Research Center on Aging, Tufts
University, Boston.
Introduction: In the past decade, a number of epidemiologic studies have shown that homocysteine, a sulfur-containing amino acid formed during the metabolism of methionine, is a risk
factor of atherosclerosis, myocardial infarction, stroke, and thrombosis. Elevated blood levels of homocysteine produce endothelial cell injury, primarily through oxidation. Particulate air
pollution has been related to cardiovascular death and hospital admission, but the underlying mechanisms are not elucidated.
Methods: We examined the associations between ambient particulate air pollution and plasma levels of homocysteine in 639 older men. Total homocysteine in plasma was measured using
high-performance liquid chromatography with fluorescence detection. Concentrations of ambient particulate air pollutants (particulate matter less than 2.5 mm in aerodynamic diameter
(PM2.5), black carbon (BC), and sulfate) were measured at a stationary ambient monitoring site. Because of skewness, the log-transformed homocysteine was used as an outcome variable. To
account for lagged effects of air pollution, moving averages (MA) of air pollutant of up to 5-days were used. Linear regression analyses were conducted to assess the relation of homocysteine
with air pollution, adjusting for age, BMI, total cholesterol, cigarette smoking, alcohol consumption, season, and outdoor apparent temperature.
Results: The median concentration of total plasma homocysteine was 10.2 µmol/L. Interquartile range increases in 2-day MA PM2.5 (6.04 µg/m3) and 5-day MA BC (0.34 µg/m3) were
significantly associated with a 2.9% (95% confidence interval (CI), 0.6% to 5.2%) and a 4.8% (95% CI, 1.4% to 8.4%) increase in plasma homocysteine, respectively. Positive associations
with one-day average sulfate concentrations measured a day before plasma sample collection were also marginally significant (1.6%; 95% CI, -0.2% to 3.5% for a 2.43 µmg/m3 increase in
sulfate). These associations were stronger in persons with diabetes and ischemic heart disease (IHD), showing a 4-fold higher increases in homocysteine in association with PM2.5, as compared
with persons without diabetes and IHD. The associations with BC were also stronger in persons with diabetes and IHD, with almost 2-fold higher percent changes. Effect modification by folate
and other nutrients was not found.
Discussion and Conclusions: Exposures to ambient particles are associated with elevated plasma homocysteine levels. Persons with diabetes and IHD are more susceptible to particle
exposures in relation to homocysteine. This study suggests that homocysteine may be a critical component or biological marker of the oxidation pathways underlying the effect of ambient
particles on the cardiovascular system.




SAB4-PD-03                  SOURCE APPORTIONMENT OF FINE PARTICULATE MATTER IN THE U.S. AND
                            ASSOCIATIONS WITH LUNG INFLAMMATORY MARKERS IL-8, COX-2, AND HO-1
R DUVALL1, G NORRIS1, J BURKE1, J MCGEE2, MI GILMOUR2, R DEVLIN2
(1) US Environmental Protection Agency, National Exposure Research Laboratory, Research Triangle Park. (2) US Environmental Protection Agency, National Health and
Environmental Effects Laboratory, Research Triangle Park.
Introduction
Associations are well established between particulate matter (PM) mass concentrations and increased human mortality and morbidity. The toxicity of PM may depend on the PM source and
composition of PM which will vary depending on the sampling site. The association between fine PM sources and three lung inflammatory markers (interleukin-8, cyclooxygenase-2, and heme
oxygenase-1) was evaluated in multiple cities with contrasting PM sources as part of the Multiple Pollutant Study (MAPS).
Methods
Fine particles were collected in six sites in the U.S. during 2003-2004. Sampling sites included Washington (Seattle), Utah (Salt Lake City), Arizona (Phoenix), and New York (Hunter College,
Sterling Forest, and South Bronx). One-week samples (24-hrs a day) were collected at each site over a 4-week period using a High Volume Cascade Impactor. Trace metals were measured by
Inductively Coupled Plasma–Optical Emission Spectrometry (ICP-OES) and ions were measured by Ion Chromatography (IC). A chemical mass balance (CMB) model was used to quantify the
sources impacting each sampling site. Particles were applied to cultured human primary airway epithelial cells and the secretion of interleukin-8 (IL-8) was measured 24-hrs later by enzyme-
linked immunosorbent assay (ELISA). Heme oxygenase-1 (HO-1) and cyclooxygenase-2 (COX-2) levels were quantified by real time reverse transcriptase polymerase chain reaction (RT-PCR).
Results
The sources of fine PM and their relative contributions varied across the sampling sites: mobile sources (4-43%), residual oil combustion (1-4%), coal combustion (19-53%), lead smelting
(<1%), oil refinery (<1%), wood combustion (8-54%), soil dust (3-4%), and secondary nitrate (8-36%). Strong linear associations were observed between IL-8 and residual oil combustion
(r2=0.84) in addition to coal combustion (r2=0.70). COX-2 exhibited weak linear correlations between residual oil combustion and coal combustion (r2=0.38 for each source). A strong linear
association was observed between HO-1 and oil refineries (r2=0.99). It should be noted that residual oil combustion was only present in the Eastern U.S. sites (n=3) and oil refineries were
only present in the Western U.S sites (n=3).
Discussion and Conclusions
The inflammatory markers analyzed yield varying health responses to sources. IL-8 and COX-2 are correlated with the same sources, coal combustion and residual oil combustion, although at
differing levels. In contrast, HO-1 had no significant correlation with these sources, but rather was strongly correlated with oil refineries. Combustion and industrial sources appear to elicit an
inflammatory response which is consistent with previous health studies.



                                                                                                79
        Symposia Session:

SAB4-PD-04                 PARTICULATE EXPOSURES AND HEART RATE VARIABILITY DURING SLEEP IN
                           BOILERMAKER CONSTRUCTION WORKERS
J CAVALLARI1, E EISEN1, R HAUSER1, R HERRICK1, J SCHWARTZ1, D CHRISTIANI1
(1) Dept. of Env. Health, Harvard School of Public Health, Boston.
Introduction: Studies have shown associations between PM2.5 exposure and decreased heart rate variability (HRV), a marker of cardiac autonomic control. Limited data exists on the cardiac
effects of PM2.5 exposure during sleep, a period free of cardiac influences from extraneous activities. We investigated the association between HRV during sleep and daytime PM2.5 exposures
in a group of boilermaker construction workers exposed to welding fumes.
Methods: 30 male workers participated in a short-term prospective study with a repeated measures design. Participants were continuously monitored over two 24-hour periods, including a
non -workday (baseline) and a workday, for personal PM2.5 exposure and by ambulatory ECG. Health status and physical activities were assessed by questionnaire. Three time domain measures
of HRV were computed over an 8-hour sleep period: root-mean-square of differences between adjacent RR intervals (rMSSD), standard deviation of normal RR intervals (SDNN) and the mean
of the standard deviation of normal 5-min RR intervals (SDNNi). Personal PM2.5 exposures were averaged over the workday and evaluated in relation to HRV parameters during sleep in linear
regression models.
Results: The mean PM2.5 concentration at work was 0.88 ± 0.15 mg/m3 and ranged from 0.13 mg/m3 to 3.73 mg/m3. With increasing PM2.5 exposure, there was a decline in all three HRV
parameters during sleep. Among smokers (n=14), for each 1 mg/m3 increase in 8-hour daytime average PM2.5, there was a change of -9.7 ms (95% CI: -15.5 – -3.8) for rMSSD, -11.1 ms (95%
CI: -25.0 – 2.9) for SDNN, and -9.2 ms (95% CI: -14.8 – -3.5) for SDNNi after adjustment for baseline sleep HRV. Among non-smokers (n = 16), for each 1 mg/m3 increase in 8-hour daytime
average PM2.5, there was a change of -4.1 ms (95% CI: -12.9 – 4.7) for rMSSD, -28.1 ms (95% CI: -58.2 – 1.9) for SDNN, and -5.4 ms (95% CI: -16.4 – 5.6) for SDNNi after adjustment for
baseline sleep HRV.
Conclusions: Preliminary results suggest that high levels of daytime PM2.5 exposures may cause changes in HRV during the subsequent night’s sleep in healthy, working men. Findings further
suggest that smokers may be more susceptible to the cardiac alterations.
Supported by NIEHS ES09860 and ES00002.




SAB4-PD-05                 HAEMATOLOGICAL AND INFLAMMATORY EFFECTS OF SHORT TERM EXPOSURE
                           TO URBAN PARTICULATE MATTER – THE PM-CARE STUDY
L RUGGERI1, P CARRER1, B BISCARO2, E CORSINI3, G DE VITO4, S FOSSATI1, S FUSTINONI5, I INFUSINO6, A LATTUADA2, P LURASCHI6, M MARINOVICH3,
F METRUCCIO1, M TIBILETTI4, P URSO1, GC CESANA4, M PANTEGHINI7, M MARONI1
(1) Department of Occupational & Environmental Health, Hospital L. Sacco Unit, University of Milan, Milan. (2) Laboratory of Immunohematology and Blood Transfusion,
Hospital L. Sacco, Milan. (3) Laboratory of Toxicology, Department of Pharmacological Sciences, University of Milan, Milan. (4) Department of CLinical Medicine and
Prevention, Milano-Bicocca University, Monza. (5) Department of Occupational Health, Fondazione Policlinico Ospedale Maggiore, Mangiagalli e Regina Elena, Milan.
(6) Laboratory of Clinical Chemistry, Hospital L. Sacco, Milan. (7) Department of Clinical Sciences, University of Milan, Milan.
Background: PM-CARE is an acronym for Particulate Matter Cardio Respiratory Effects; the aim of the study is to investigate the mechanisms through which urban air particulate matter
causes harmful effects on cardiovascular and respiratory systems.
Aim: To assess haematological and inflammatory parameters in healthy subjects and in subjects with cardiac or chronic lung disease in Summer and Winter surveys of the PM-CARE Study
Methods: Three panels of non-smoking adult subjects 35 with cardiovascular disease (heart group), 21 with chronic lung disease (COPD or asthma) (lung group) and 27 healthy subjects
(healthy group) were recruited from the patients community of two Italian hospitals. Each subject was investigated for environmental and clinical parameters for one day during and after
his/her habitual activities in Summer 2005 and in Winter 2005/2006. Environmental parameters included ultrafine, fine and coarse fractions of particulate matter, CO, O3, NO2 (for environmental
data see abstract “Assessment of Personal PM Exposure in the PM-CARE Project”) Clinical investigation consisted in daily symptom/activity recording, measurement of heart and lung
functions, analysis of biomarkers of cardio-respiratory injury in blood and sputum samples. Biological samples were collected at the end of 24-h investigation: differential blood cells count,
fibrinogen, coagulation parameters (aPTT, INR, PF100, F1+2, vWF, t-PAI, D-dimer), inflammatory parameters (TNF-alfa, sR-I and II of TNF-alfa, IL-8, IL-10 measured both in plasma and in vitro,
following stimulation with phytohemagglutinin or lipopolysaccharide), hs-CRP. Coagulation and inflammatory parameters will be measured at the end of the survey.
Results: Preliminary environmental data showed higher personal exposure to particulate matter in Winter than in Summer survey (PM10 daily average mass concentration: Winter 71 mcg/m3,
Summer 44 mcg/m3). Preliminary Summer and Winter clinical data showed higher Winter levels of platelets (p=0.040), leukocytes (p=0.047), monocytes count (p=0.001). Analysing each group,
we found higher Winter levels of platelets (p=0.03) and monocytes count (p=0.046) in healthy group and higher Winter levels of monocytes count (p=0.022) in heart group.
Discussion and conclusions: Data from Winter survey suggested a systemic inflammatory response involving a variety cells type. More evaluations will be possible in the next months, when
the collection and analysis of results will be completed. Haematological and inflammatory results will be complemented with data on lung and cardiovascular functions and the relation
between clinical parameters and individual particulate matter monitoring results will be analysed.




SAB4-PD-06                 SHORT TERM ASSOCIATION BETWEEN AIR POLLUTION AND CARDIOVASCULAR
                           DISEASES IN SPAIN
C IÑIGUEZ1, MP FERNÁNDEZ1, F BALLESTER1, A DAPONTE2, A LERTXUNDI3, P ARIAS4, F ARRIBAS5, J BELLIDO6, A CAÑADA7, A FIGUEIRAS8, JJ GUILLÉN9,
E LÓPEZ10, C MORENO-IRIBAS10, T RUEDA10, C SAURINA10, S TORO10
(1) Unit of Epidemiology and Statistics, Valencian School of Health Studies (EVES), Valencia. (2) Escuela Andaluza de Salud Pública, Granada. (3) Grup de Recerca en
Estadística, Economía Aplicada y Salut (GRECS), Universidad de Girona, Girona. (4) DG Salud Pública. Comunidad de Madrid, Madrid. (5) Departamento de Sanidad
Aragón, Zaragoza. (6) Centro de Salud Pública, Castellón. (7) Dirección General de Salud Pública, Asturias, Oviedo. (8) Universidad de Santiago. Facultad de Medicina.,
Vigo. (9) Centro de Área, Cartagena., Cartagena. (10) DG Salud Pública Canarias, Las Palmas de Gran Canaria.
Introducction: The EMECAS project (Spanish Multicentric Study on the relationship between Air Pollution and Health) assesses the short-term effects of several pollutants on hospital
admissions and mortality in Spain. Results for cardiovascular mortality are presented and compared with those for cardiovascular admissions.
Methods: Fourteen cities, accounting for 9 million of inhabitants, participated. The study period ranged from 1993 to 1999. The daily number of deaths and emergency admissions for all
cardiovascular diseases (ICD-9: 390-459) were obtained from the mortality register and hospital records, respectively. Black smoke (BS), total suspended particles (TSP), PM10, SO2, NO2, CO and
O3 data were provided by the regional air pollution networks. Using these data, daily indicators for each city were calculated: 24 hour average of BS, TSP, PM10, SO2 and NO2, 8 hours maximum
moving average of CO and ozone, and 1 hour maximum of SO2, NO2, CO and O3. The magnitude of association was estimated using generalized additive models (GAM) of Poisson controlling
for confusion and overdispersion. Analyses were done using the S-Plus GAM function with stringent convergence criteria. Lagged effects up to three days were examined. For each cause and
lag, combined estimates were obtained under a “random” or “fixed” effects model depending on if heterogeneity was present or not. Analyses for ozone were restricted to the warm period
(May to October)
Results: Local estimates were, mostly, positive. Lags 2 and 3 were the most consistent ones for CVS mortality, while for admissions lags 0 and 1 were the most consistent. Combined analyses
showed an association with CVS mortality for 24-hour levels of BS, TSP, SO2 and NO2. An increase of 10 µg/m3 in BS levels was associated with a 1.13% (95CI: 0.58-1.67%) increase of the
number of CVS deaths, the same increase in TSP was associated with a 0.98% (95 CI: 0.01-1.97) increase, NO2 with 0.44% (0.09-0.80%) and SO2 with 1.4% (0.25-2.5%). Weaker non
significant associations were found for 1-hour maximum indicators and neither Ozone nor CO showed effect. For CVS admissions, estimates were in general of similar magnitude for 24-hour
averages. Also 1-hour maximum indicators showed the same impact, and both ozone and CO showed effect.
Discussion and conclusions: An association between air pollution and cardiovascular health has been described in Spanish cities. In mortality, compared with hospital admissions, the effect
was more delayed and daily peaks of pollutants showed no association.
Study funded by the Spanish Ministry of Heatlh (FIS-FEDER 00/0010).


                                                                                              80
                                                                                                                                                                                                         SUNDAY SEPTEMBER 3
         Symposia Session:

SAB4-PD-07                  TRENDS IN RELATIVE RISKS FOR BLACK SMOKE AND DAILY MORTALITY
                            OVER A 30 YEARS PERIOD IN THE NETHERLANDS
P FISCHER1, C AMELING1, M MARRA1
(1) National Institute of Public Health and the Environment, Bilthoven.
There is overwhelming evidence that daily variation in levels of air pollution is associated with increased daily mortality. However, uncertainty remains about the causal agents that are
responsible for the health effects. In a previous analyses of the relationship between daily SO2 levels and daily mortality in the Netherlands, it was shown that despite of the strong decrease
in the SO2 levels over time, the relative risks for mortality increased over time (Buringh E., Fischer P., Hoek G. Is SO2 a causative factor for the pm-associated mortality risks in the Netherlands.
Inhalation Toxicology 2000; 12(Suppl. 1): 55-60). The interpretation of this result was that the statistical significant association between daily variation in SO2 and daily mortality was did not
reflect a causal relation but it was just the result of the correlation between the daily variation in levels of SO2 with an unknown causal agent in the air pollution mixture.
In this paper we present the results of an analysis of daily levels of Black Smoke (BS) and daily mortality, spread over a 30 years period (1972 – 2002). During this period the BS levels decreased
in The Netherlands from 27 ug/m3 to 8 ug/m3. If, like SO2, BS was correlated with some causal agent in the air pollution mixture and not itself causally associated with daily mortality, we
expected that relative risks of BS would increase over time.
We applied penalized splines GAM analyses with adjustments for daily temperature, influenza counts, trend, day of the week, relative humidity and atmospheric pressure. We analysed daily
total mortality and daily cause-specific mortality (cardiovascular, pneumonia, and respiratory). Our results revealed no systematic change in the relative risks for BS over the 30 years period.
We also compared our results with results based on lowess GAM smoothers showing a tendency for lower effect estimates based on p-splines analyses.
We conclude that, in contrary to what previously was concluded for SO2, BS may not be ruled out as one of the components of air pollution which are causally related with premature mortality
in the Dutch population.




SAB4-PD-08                  THE RELATIONSHIP BETWEEN PARTICULATE MATTER (PM) COMPONENTS AND
                            SURVIVAL AMONG COHORT MEMBERS
Rr WYZGA1, F LIPFERT2, J BATY3, JP MILLER3
(1) EPRI (Electric Power Research Institute), Palo Alto, CA. (2) consultant, Northport, NY. (3) Washington University, St., Louis, MO.
Introduction: Several studies have reported statistically significant associations between air pollution, characterized by particulate matter (PM) and/or gaseous pollutant levels, and the
survivorship among cohort members. Studies to date have not systematically investigated the associations between specific components of PM and cohort survival. This study considers the
relationship between PM, components (metallic ions, organic and elemental carbon, sulfate and nitrate ions), gaseous pollutants, and other indices of environmental nuisance (e.g., traffic)
and cohort survivorship.
Methods: This study makes use of a cohort of 70,000 men who have been followed since 1976. Proportional hazard regression techniques were used to estimate the influence of several
external variables, including pollution values, on the survival times of the cohort after adjusting for demographic (age, race), behavioral (smoking history), and prior health status (body-mass
index, blood pressure) variables among cohort members. Several model specifications were considered, including models with multiple pollution variables. Model results were compared using
measures of statistical significance, effect size, and overall model fit as measured by the Akaike Information Criterion.
Results: Single pollutant models indicated statistically significant associations between cohort survivorship and elemental carbon (EC), nitrate ion, vanadium, and nickel. Peak ozone and iron
were positively associated with cohort survival, but the association was not statistically significant. Two pollutants models showed EC and the nitrate ion retained statistically significant
associations with survivorship when they were considered jointly with nickel ion concentrations or with peak ozone. Peak ozone, iron, vanadium, and nickel were statistically significantly
associated with survivorship in at least one model with multiple pollutants. When a traffic density variable was introduced into the multiple pollutants models, it appeared to dominate the
pollution variables in terms of predicting cohort survivorship and remained statistically significant in many models.
Discussion and Conclusions: The results indicate that some components of pollution and PM are much more highly associated with cohort survivorship than others. An indicator of traffic
density trumped these variables, however, when they were considered jointly with pollution measures. It is unclear whether this result implies that 1.) there is a non-measured component of
pollution associated with traffic that may explain the observed associations or 2.) some non-pollution aspect of traffic is a better predictor of cohort survivorship. This issue is complicated by
considerations of differential measurement error among the variables studied.




SAB4-PD-09                  TIME VARYING COEFFICIENT MODELS FOR THE ANALYSIS OF
                            TRAFFIC RELATED AIR POLLUTION AND MORTALITY
S BREITNER1, J CYRYS2, H KÜCHENHOFF3, M STÖLZEL1, J HEINRICH1, H-E WICHMANN1, A PETERS1
(1) GSF - National Research Center for Environment and Health, Institute of Epidemiology, Neuherberg. (2) WZU - Environmental Science Center of the University
Augsburg, Augsburg. (3) University of Munich, Department of Statistics, Munich.
Introduction: Numerous time series studies have reported positive associations between short-term variations in ambient levels of traffic-related air pollution and daily mortality counts. The
models used in these studies have typically assumed that the association between pollutants and daily mortality is constant over time.
The objective of this analysis was to test whether the relative risks for daily mortality in association with traffic-related air pollution concentration remained unchanged over a ten-year period.
Methods: Daily mortality counts were obtained between 1991 and 2002 in a central European city with 200,000 inhabitants. Gaseous air pollutants and meteorological data were collected
from official networks. Daily means of the particle size distribution including number concentrations of ultrafine particles (UFP) were obtained from our measurement site.
We investigated potential changes in the effect estimates over time using Bayesian hierarchical Poisson models. We applied a smooth time-varying effect consisting of an interaction between
a penalized B-spline function of time and the respective pollutant. One advantage of this approach is that it produces a smooth estimate of the changing air pollution effect over time that
does not follow any a-priori fixed parametric form specified by the investigator. We further included smooth functions of long-term time trend, season, influenza epidemics and meteorology
to control for confounding effects of these variables.
Results: The total study period was divided into three sub-periods: period 1 from October 1991 to August 1995, period 2 from September 1995 to February 1998, and period 3 from March
1998 to March 2002. Mean CO, NO2 and UFP concentrations decreased between periods 1 and 3 by a factor of 2.70, 1.46 and 1.23, respectively. All decreases were statistically significant.
We observed an association of CO at lag 4 days and daily mortality. We found that the magnitude of this association varied smoothly over time. Relative risks per interquartile range (IQR)
ranged between 0.98 (95% CI: 0.89, 1.07; in 2001) and 1.06 (95% CI: 1.01, 1.10; in 1997). NO2 exhibited less strong patterns. For UFP number concentrations, we found a time-varying
association with daily mortality whose magnitude shows a seasonal pattern with a relative risk per IQR ranging between 0.89 (95% CI: 0.68, 1.17) in summer and 1.09 (95% CI: 1.01, 1.18)
in spring.
Conclusion: Results from this analysis indicate an elevated mortality risk from short-term exposure to traffic-related air pollution whose magnitude may change over time along with changing
air pollution patterns.




                                                                                                 81
        Symposia Session:

SAB4-PD-10                 PUSH AND PULL LAGS
R KLEMM1, C GUST1, R WYZGA2
(1) Klemm Analysis Group, Inc., Washington, DC. (2) Electric Power Research Institute, Palo Alto, CA.
Introduction
Often researchers investigating the effects of air quality and weather measures on human mortality postulate models that assume that the air quality and weather measures are either
coincident or prior to the date of death (“push lags”). To better understand and interprete the meaning of these findings, it can be useful to estimate the effects of air quality and weather
measures of comparable time periods after the date of death (“pull lags”).
Methods
The team of researchers used daily mortality data for Atlanta, GA during 1998 - 2000 and Phoenix, AZ for 1995 - 1997. Counts of daily mortality were investigated for all deaths, all deaths
by whether under or over 65 years of age at time of death, and categories of causes of death (such as circulatory, respiratory or cancer). Particulate matter data consisted of approximately
50 individual measures from the ARIES database of Atlanta, GA discussed by Klemm, et al. (Inhalation Toxicology, 2004) during August 1998 - July 2000 and that discussed by Thurston, et al.
(Environmental Health Perspectives, 2005) for Phoenix, AZ during March 1995 - June 1998. Associations between daily human mortality were estimated using Poisson generalized linear models
(GLM) in SPLUS 2000.
Results
Systematic changes in the number, length and sign of lags demonstrate some potentially surprising results. Situations where “negative” lags are both positive in sign and either significant or
close to significant at the p=.05 level are compared to their “positive” lag counterpart models. In addition, adjustments from the standard model of coincident weather measures are presented
to better understand the otherwise un-anticipated “negative” lag results. Variances by the causes of death are also illustrated graphically and compared between the two locations, Atlanta
and Phoenix.
Discussion and Conclusions
Investigating “non-sensible” time-sequenced models can be useful in understanding models that are sensible in a time-sequenced manner. They can serve as “benchmarks” against which the
sensible models can be compared. Simple ways to present the “sensible” results that incorporate the “non-sensible” results are presented that strengthen the sensible findings.




SAB4-PD-11                 TEMPERATURE MODIFIES THE SHORT TERM EFFECTS OF PARTICULATE
                           MATTER ON CARDIOVASCULAR DISEASES IN BRISBANE, AUSTRALIA
C REN1, S TONG1
(1) School of Public Health, Queensland University of Technology, Brisbane.
Introduction: A large amount of literature shows that ambient particulate matter is associated with health outcomes. A few epidemiological studies have examined whether there is an
interactive effect between temperature and ambient particulate matter on cardiorespiratory diseases, but the results were inconsistent. This study was to examine whether maximum
temperature modified the association of ambient particulate matter less than 10 µm in diameter (PM10) with total daily respiratory hospital admissions (RHA), cardiovascular hospital
admissions (CHA), respiratory emergency visits (REV), cardiovascular emergency visits (CEV), all non-external cause mortality (NECM) and cardiovascular mortality (CM) in Brisbane between
1996 and 2001.
Methods: Three Poisson generalized additive regression models were used to separately examine whether there are interactions between temperature and PM10 on individual health
outcomes at each of 0-2 lags after adjustment for covariates, such as seasonality, days of the week, humidity, rainfall, influenza epidemic, nitrogen dioxide. A bivariate response surface model
was used to perspectively examine joint effects of temperature and PM10 on cardiovascular diseases. A non-stratification model was applied to quantitatively examine the interaction effect,
and a stratification model was performed to examine the heterogeneity of PM10 effects across different levels of temperature.
Results: The results show that there existed a consistent and statistically significant interaction between PM10 and maximum temperature on RHA, REV, CEV, NECM and CM at various lags,
but not obviously on CHA in Brisbane during the study period. The choice of the degree of freedom adjusting for confounders and the selection of arbitrary cut-offs for temperature slightly
affected the magnitudes of estimates, but did not change the overall conclusion.
Conclusion: There is consistent evidence that maximum temperature modified the short-term effects of PM10 on RHA, REV, CEV, NECM and CM in Brisbane between 1996 and 2001. PM10
exhibited the greater adverse health effects on warmer days. The results imply that it is important to control and reduce the emission of particles in Brisbane, particularly when temperature
increases.




SAB4-PD-12                 THE ROLE OF POLLUTION IN HEAT WAVE MORTALITY: SYSTEMATIC REVIEW
K KATSOUYANNI1, A ANALITIS1
(1) Univ of Athens Medical School, Athens.
Short-term heat wave effects on mortality as well as the short-term effects of specific air pollutants’ concentrations are well known. In situations of extreme heat the concentration levels of
pollutants and the pollutant mix are affected through various mechanisms. It is therefore necessary to control for the confounding effects of air pollution when investigating the heat-wave
effects and it is of significant interest to study the combined, possibly synergistic, effects of heat and other meteorological variables on the one hand and air pollutants’ concentrations on the
other.
A review of the above topics will be presented as part of a dedicated session to the social and environmental determinants of heat wave mortality. The pollutants of interest are those affected
by photochemical procedures, mainly ozone and ambient particulate matter. The heat wave effect is often assessed using composite indices, which combine a measure of heat and humidity
or more meteorological variables into synoptic weather categories. Comparability and interpretation of the results is sometimes made more difficult by the use of indices.
Papers addressing the issue of confounding are few, but represent a diversity of locations, for example Canada, Mexico, central and southern Europe. The confounding effects of pollution in
the heat-wave or heat related mortality appears to be small, but goes in either direction.
Only a few studies have addressed the possible interactive effects of temperature or heat waves and specific air pollutants on health outcomes. The outcome most readily studied is mortality.
Although the independent effects of heat waves and of pollution are established and generally accepted, the evidence for synergistic effects is sparse. From published studies, it appears that
there is interaction only for very high temperatures or in heat wave situations.
The study of interactive effects presents methodological challenges which will be discussed.




                                                                                               82
                                                                                                                                                                                                     SUNDAY SEPTEMBER 3
        Symposia Session:

SAB4-PD-13                 SHORT TERM ASSOCIATIONS BETWEEN AMBIENT AIR POLLUTION AND
                           ARRHYTHMIC EVENTS IN PATIENTS WITH IMPLANTED CARDIOVERTER
                           DEFIBRILLATORS, 1993 2002
M KLEIN1, P TOLBERT1, K METZGER2, S SARNAT1, J PEEL3, D FLANDERS1, J MULHOLLAND4, L DARROW1, J LANGBERG5
(1) (1) Emory University, Atlanta. (2) (2) NYC Department of Health and Mental Hygiene, New York. (3) (3) Colorado State University, Fort Collins. (4) (4) Georgia Institute
of Technology, Atlanta. (5) (5) School of Medicine, Atlanta.
Introduction: To investigate the effect of air pollution on cardiac arrhythmia, we have compiled one of the largest databases of patients with implanted cardioverter defibrillators (ICDs) from
three electrophysiology clinics for the time period 1993-2002. ICDs continuously monitor the heart rate for tachyarrhythmias, emit electrical therapy or shocks to regulate heart rhythm, and
record data on each event, including date, time and type of therapy. Our database includes 884 patients, of whom 518 had 13,108 recorded ventricular fibrillation and ventricular tachycardia
events. We previously examined associations between these events and daily air pollution concentrations; results were generally consistent with the null. In the current analysis, we used hourly
pollution data to refine our assessment of short-term associations.
Methods: We performed case-crossover analyses using conditional logistic regression to examine associations between the hour of tachyarrhythmic events and hourly ambient particle and
gas concentrations. This analysis design implicitly controls for time-trend, seasonality, and diurnal patterns. “Event-hours” were defined by truncating subjects’ event-times to the nearest hour
and excluding events occurring within 24-hours of a previous event. “Control-hours” were matched to events by subject, year, calendar month, day-of-week, and hour of day. We excluded
control-hours if a subject experienced an event within the previous 24-hours. After exclusions, we had information on 506 subjects contributing 5,193 event-hours and 15,250 control-hours.
We used several pollutant metrics in our models with event data, including hourly lags of 0 and 1, as well as 3-, 6-, 12-, and 24-hour moving averages prior to the event.
Results: Associations were generally consistent across the different time metrics for both particle and gaseous pollutants. Results were largely consistent with the null, although there was
some evidence of associations with nitrate, total carbon, nitric oxide and total reactive odd nitrogen (NOy). When we restricted our analysis to events that resulted in electrical therapy, these
associations persisted although there was a loss of power. For the most severe outcome group, events that resulted in high-energy shock, several estimated associations were negative.
Discussion: In this analysis, use of hourly pollution data allowed for a more refined assessment short-term associations. Our current results are consistent with our previous analyses, although
we observed some suggestive associations that need further exploration.




                                                                                               83

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:21
posted:10/16/2011
language:English
pages:82