GIS-MODELED INDICATORS OF MOBILE SOURCE EMISSIONS AND ADVERSE HEALTH EFFECTS AMONG CHILDREN IN EL PASO, TEXAS, USA

ISEE-135ANALYSIS OF REPEATED CROSS-SECTIONAL SURVEYS, OFPRE-SCHOOL RESPIRATORY SYMPTOMS AND TRAFFIC-RELATED AIR POLLUTION

Nevil Pierse,* Lesley Rushton,* Claudia E. Kuehni,† Mike Silverman,‡Catherine Mallon,§ Robert S. Harris,* Jonathan Grigg‡. *MRC Institutefor Environment and Health; †Dept. of Social and Preventive Medicine,University of Bern; ‡Dept. of Infection, Immunity and Inflammation,University of Leicester; §Leicester City Council, Leicester, UK

Introduction: Traffic-related air pollution is associated with respiratorymorbidity in school age children. However, there is very little informationon the effects of this pollution on younger children, the age at whichproblems start and whether problems in early childhood persist in laterlife.Aim: To determine if estimates of PM10 exposure at the home addressare associated with respiratory symptoms in young children and whetherthis association changes as the children get older.Methods: 4400 children were selected in a stratified random sample oflocal birth records and their parents were sent questionnaires in 1998(children aged one to five), 2001 (aged four to eight) and again in 2003(aged six to ten). The questionnaire asked about respiratory symptoms andrelevant confounders.

Logistic regression models were used to test the association between therespiratory symptoms and objective measures of exposure. The distancefrom home address to a major road (DHR) was used as a surrogatemeasure of PM10 exposure.Results:

Table 1: Association between distance to a major road (per 100m) andchronic respiratory symptoms during the past 12 months

Symptoms

Children Aged:

Odds ratio (95% CI)

1998 (N�3410)

1-5 years

2001 (N�2580)

4-8 years

2003 (N�2153)

6-10 years

Cough without cold 1.032 (1.013, 1.052) 1.029 (1.009, 1.049) 1.029 (1.007, 1.050)

Cough at Night* 1.018 (1.001, 1.035) 1.018 (1.000, 1.036) 1.017 (0.997, 1.037)

Nasal symptoms* 1.018 (1.002, 1.035) 1.027 (1.009, 1.044) 1.037 (1.017, 1.057)

Wheeze* 1.005 (0.988, 1.023) 1.023 (0.999, 1.048) 1.008 (0.983, 1.034)

*Questionnaire asked for symptoms in the last 12 months

From Table 1 we can see that in all three studies cough without a coldhas a significant association with DHR. Cough at night has a consistentassociation with DHR although this has declining significance due tofalling power. Nasal symptoms are associated with DHR in all surveys.Wheeze was not significantly associated with DHR in any year. Afteradjusting for relevant confounders the results were largely unchanged.Discussion: Both cough and nasal symptoms have a significantassociation with an objective exposure to traffic related air pollution inpre-school children. This association persists through early childhood.This association is less clear for wheeze.

ISEE-136HUMAN EXPOSURE TO BENZENE: THE ROLE OF URBAN AIRPOLLUTION AND TOBACCO SMOKE IN A COHORT OFTRAFFIC POLICEMEN

Luciano Maestri,* Deborah Traversi,† Sergio Ghittori,* Elisabetta Fea,‡Roberto Bono,† Tiziana Schiliro,† Giorgio Gilli†. *Fondazione SalvatoreMaugeri; †University of Torino; ‡University of Turin

Abstract: Benzene, a well known environmental and occupationalcarcinogen, is present primarily in air but, active smoke habit representsthe most important source of benzene exposure for humans. Fornonsmokers, however, benzene exposure ultimately derives from autoexhaust inhaled from the outdoor air. For these reasons, to evaluate thecontribution of traffic and tobacco smoke to benzene exposure in anurbanized population and to verify the level of risk for workers strictlybound up with urban traffic, an investigation on personal exposure tobenzene in 206 traffic policemen from the city of Turin city (SouthEurope) was carried out. Personal air samplers were used to measureexposures to benzene in the breathing zone during a work shift. Moreover,in a spot of urine collected at the end of the work shift, S-phenylmercapturic acid (S-PMA) and cotinine were measured respectivelyas biomarkers of benzene and as biomarker of tobacco smoke. S-PMAwas measured by HPLC, cotinine by gas chromatography. At the end ofthe work shift some information were also collected on: sex, age, minutesspent indoor, outdoor, directing traffic, cigarettes smoked per day.

The correlation between cigarettes/day and cotinine was positive andsignificant: p�0.703, r�0.000). The mean of benzene regarding the wholepopulation (n. 206) was 8.80 � 8.05 �g/m3 while the of S-PMA was1.48 � 1.14 �g/gr of Creatinine; the Pearson correlation between thesetwo parameter was significant (r�0.000) and positive (p�0.353),confirming the usefulness of this biomarker of benzene. The Pearsoncorrelation was significant (r�0.000) and positive (p�0.445), also for S-PMA versus cotinine but not for S-PMA versus the minutes spent bytraffic policemen in indoor environments, outdoor and directing traffic.

T-test carried out to compare means of S-PMA for the 52 smokers (2.340� 1.57) and the 154 non smokers (1.225 � 0.82) shown high significance(p�0.000), as well as the comparison of benzene means in smokers(12.677 � 8.600 �g/m3) and in non smokers (7.661 � 7.531 �g/m3):p�0.000. Finally, any relationship was recorded comparing means ofbenzene and S-PMA for subjects working only indoor versus only outdoorand subjects working only indoor versus traffic directors.

The present results highlight for traffic policemen the responsibility oftobacco smoke but not of the place of work. Furthermore, the benzeneintake for non smokers does not seem depend on outdoor air or trafficdirection, if compared to intake of a “normal” subject working in office.

This study was made possible by grants of from Regione Piemonte,ricerca finalizzata anno 2000.

ISEE-137EXPOSURE TO AIR POLLUTION FROM HEAVY TRAFFIC ISASSOCIATED WITH INCREASED EOSINOPHILIC ACTIVATIONIN DUTCH SCHOOLCHILDREN

Nicole A.H. Janssen, Gea de Meer, Bert Brunekreef. IRAS

Introduction: Epidemiological studies have shown associations betweentraffic-related air pollution and adverse effects on (respiratory) health.Some of these studies suggest that these effects are specifically related toexposure to truck (and not car) traffic, suggesting a role of diesel exhaust.Experimental studies have shown that exposure to diesel exhaust particlescan result in inflammatory reactions in the airways. Few epidemiologicalstudies, however, have included measurement of inflammatory markers toevaluate these effects in population studies. We therefore evaluated theassociation between eosinophilic activation and exposure to air pollutionfrom heavy traffic within our previously published study of childrenattending schools near highways.Methods: In 1997/1998 a study was conducted on the respiratory healthof children attending 24 schools within 400 m of a highway. Exposure totraffic-related air pollution was assessed using specific traffic-relatedcharacteristics (traffic counts for cars and trucks separately and distance ofthe homes and schools to the highway) as well as annual average

Epidemiology • Volume 15, Number 4, July 2004 Abstracts

© 2004 Lippincott Williams & Wilkins S65

concentrations of PM2.5, PM2.5-absorbance (“soot”) and NO2 outside theschools. Respiratory health status was assessed using the ISAAC IIprotocols, including blood withdrawal for IgE measurement. In addition,blood eosinophil counting and assessment of serum eosinophilic cationicprotein (ECP) was conducted. Eosinophil and ECP levels were log-transformed to normalize residuals. Associations were adjusted forpotential confounders, including age, sex, ETS, grass/pollen season, andparental history of asthma or rhinitis.Results: Blood eosinophil counting as well as assessment of serum ECPwas obtained for 521 children. No significant associations were foundbetween any of the exposure variables and the number of eosinophils.Levels of serum ECP, however, significantly (p�0.05) increased withincreasing soot concentration, with decreasing distance of the school tothe motorway and with increasing truck traffic density, but not with cartraffic density or concentrations of PM2.5 mass and/or NO2. Observedassociations correspond to a 50–60% increase in serum ECP whencomparing the highest with the lowest exposed children. Preliminaryfurther analysis suggests that these associations appear to be present bothin atopic as well as in non-atopic children.Discussion: To our knowledge this is the first study that evaluated theassociation between exposure to air pollution from heavy traffic andeosinophilic activation in an open population of school children. Increasedserum ECP was associated specifically with truck (and not car) traffic andsoot, suggesting that the effects are related to exposure to diesel exhaustparticles (DEP). The results are therefore in line with experimental studiesthat show that exposure to DEP can result in inflammatory responses.

ISEE-138GIS-MODELED INDICATORS OF MOBILE SOURCE EMISSIONSAND ADVERSE HEALTH EFFECTS AMONG CHILDREN IN ELPASO, TEXAS, USA.

Lucas Neas,* Melissa Gonzales,† Shabil Mukerjee,* Luther Smith,‡Erik Svendsen*. *U.S. Environmental Protection Agency; †University ofNew Mexico; ‡ManTech Environmental Technology, Inc.

Abstract: The El Paso Children’s Health Study examined ambientexposures to motor vehicle emissions and their effect on respiratory healthamong children. In the fall of 1999, a network of passive and activemonitors measured indicators of mobile source emissions, includingnitrogen dioxide and toluene, from which we developed a predictivegeographic information system (GIS) model. In the spring of 2001, theparents of all fourth and fifth grade children enrolled in the El PasoIndependent School District received questionnaires, and pulmonaryfunction testing was conducted on a subset of students. Completequestionnaires were obtained for 5763 children and acceptable pulmonaryfunction data were obtained for 1815 children. In preliminary analyses, wehave found different health associations among children attending schoolson the valley floor (n�2033) and among children attending schools athigher elevations (n�3730). Above 1170 meters, we found a significantassociation between GIS-modeled toluene exposure and decreased peakflow (15% across the range of modeled toluene (5 mg/m3), 95% CI26%, 2%) after controlling for sex, age, race, height, weight, ethnicity,an interaction between sex and height, interactions between ethnicity andheight and weight, parental education and smoking, length of El Pasoresidence, and the random effect of school. Above 1170 meters, we foundthat GIS-modeled nitrogen dioxide exposure (10 ppb) was associated witha physician’s diagnosis of bronchitis (odds ratio (OR) � 1.56, 95% 0.90to 2.69) and chronic morning cough (OR � 1.53, 95% CI 0.76 to 3.09)adjusted for questionnaire language, sex, age, race, ethnicity, length of ElPaso residence, presence of smoker, gas stove with pilot lights,respondent’s sex, single parent family, parental education and parentalallergies. Below 1170 meters, we found the suggestion of more adverse

health effects for these children as a whole, but no associations with GIS-modeled exposures. While not conclusive in themselves, these results forEl Paso, Texas are consistent with the findings of previous studies inEurope and southern California that have reported adverse health effectsin children associated with indicators of mobile source emissions. Thisabstract does not necessarily represent EPA policy.

ISEE-139TRAFFIC RELATED PARTICULATES AND CHILDREN’SRESPIRATORY HEALTH

Peter Rudnai, Eva Vaskovi, Miklos Naray, Janos Varro Mihaly,Zoltan Viragh, Maria Endrody, Timea Beregszaszi, Annamaria Macsik,Eszter Szabo. National Institute of Environmental Health, Fodor JozsefNational Center for Public Health, Budapest.

Introduction: Several studies provided evidence on the associationbetween traffic related air pollution and its impact on the children’srespiratory health. In the frame of a complex national programme weevaluated the health impact of airborne particles and their chemicalcomposition.Methods: An area along a busy road leading to a motorway and a quietsuburban region were selected. 24 hour PM10 samples were collected forone year. Besides gravimetric measurement, the absorption coefficient(black smoke) and the concentration of Pb, Be, Cd, Cr, Ni, Mn, V, Znand Tl was measured. We studied the health of 384 (225 in the busy areaand 159 in the control one) schoolchildren visiting 2nd and 3rd gradeclasses by physical examination and lung function tests. Furtherinformation was collected by a questionnaire completed by the parents.Within the one-year study term two three-week periods were chosen fortesting the children’s peak expiratory flow (mini-Wright peak flow meter)and evaluate the frequency of respiratory symptoms registered in a diary.The associations between the daily concentrations of the air pollutants andthe daily frequency of respiratory symptoms or the PEF values wereevaluated by multiple logistic or linear regression GEE model usingSTATA 7.0.Results: The level of pollution and most concentrations of the measuredcomponents were significantly higher in the busy area than in the controlone (e.g., Mn 6-fold, Pb and Cr 5-fold, Be and Ni 3-fold, V and Cd 2-fold increase). There was a significantly increased risk for the prevalenceof wheezing during the last 12 months in the busy area (adjustedOR�4.71, 95% C.I.: 1.05–21.19). During the 3 week testing periods theassociations between the various pollutants and the 1 lag frequency of therespiratory symptoms were more pronounced in the control area(especially in the case of PM10 and black smoke in the heating period).The PEF performance of the children was in some cases significantlyassociated with the black smoke concentration measured on the same day.Conclusion: The long term effects of the pollutants found in significantlyhigher concentrations in the busy area than in the control one may explainthe significant increase of some respiratory symptoms among schoolchildren. The short term changes in the children’s respiratory health couldbe explained by changes in the concentration of PM10 or black smoke butnot by their chemical composition.

The study was funded by the Hungarian National Programme forResearch and Development (NKFP 1/008/2001 and 3/005/2001,respectively).

ISEE-140EXPOSURE TO TRAFFIC AND THE RISK OF ATOPY INASTHMATIC CHILDREN

Kathleen Mortimer,* Elizabeth Matovinovic,† Fred Lurmann,‡Tami Funk,‡ John R. Balmes,§ Jennifer Mann,* Helene Margolis,�

Abstracts Epidemiology • Volume 15, Number 4, July 2004

© 2004 Lippincott Williams & WilkinsS66

Katherine Hammond,* Ira B. Tager*. *University of California, Berkeley;†University of British Columbia; ‡Sonoma Technology, Inc.; §Universityof California, San Francisco; �Air Resources Board

Introduction: Evidence regarding the association between pollution fromtraffic sources and the risk of atopy remains unclear. Most effects havebeen limited to people living near roadways with heavy-duty vehicletraffic, within 100 m of a freeway, which could be strongly confoundedby demographic factors. Few studies have adjusted for individual-levelcharacteristics that may confound the relationship. It is important toWeexamined the relationship within a population of asthmatics who may bemore at risk for atopyare at greater risk for being atopic.Methods: Among a population of 215 children with asthma enrolled inthe Fresno Asthmatic Children’s Environment Study (FACES), weevaluated various traffic exposure metrics and the prevalence of skin testreactivity to a series of 14 aero-allergens. In addition to self-reporteddistance from residence to freeway, major highway, major intersection, orstreet with heavy traffic, objective exposure metrics were obtained by geo-coding the current residences using the TeleAtlas service and assigningCALTRANS Annual Average Daily Traffic (AADT) volume data to theTeleAtlas roadway network. We considered distance from major andminor roadways, AADT, heavy-duty vehicle fraction (HDV), HDV trafficvolumes, and inverse-distance weighted traffic density. Atopy was definedas a positive skin test to at least one of 14 allergens. Individual-levelcharacteristics were obtained through in-person interviews. Settled dustsamples were collected from each home. Multivariate models weredeveloped to address confounding by demographic and home environmentcharacteristics.Results: Fewer than 1% of the children lived within 100 meters of amajor freeway (median�1283m), 61% of the children were atopic. Inunivariate models, only two metrics were associated with atopy: “heavy-duty vehicle fraction on closest major arterial (highest quartile)”(OR�2.2, 95% CI: 1.2, 4.3), and “heavy-duty vehicle traffic volume onthe closet major arterial (highest quartile)” (OR�1.7, 95% CI: 0.9, 3.2).Despite the fact that race was significantly related to atopy (African-American race, OR�6.5, 95% CI: 2.1, 20.2, Hispanic race, OR�1.8, 95%CI: 1.0, 3.3), traffic coefficients were only moderately affected byadjustment for these and other demographic factors (adjusted fractionOR�1.9, adjusted volume OR�2.2). Coefficients were similar whenchildren who had lived in the current residence less than one year wereexcluded. Future analyses will adjust for home allergen levels, refine thetraffic exposure metrics and include consideration of previous residences.Conclusion: In this asthmatic cohort, only heavy-duty vehicle fractionand volume on major arterials was associated with atopy. Race was a riskfactor for atopy but did not strongly confound these associations.

Funded by the California Air Resources Board.

ISEE-141EVALUATING TRAFFIC-EXHAUST EXPOSURE MODELS FORUSE IN EPIDEMIOLOGICAL STUDIES

Paul English,* Rusty Scalf,† Robert Gunier,* Steve Wall,*Svetlana Smorodinsky*. *California Department of Health Services;†Impact Assessment, Inc.

Introduction: Exposure to traffic exhaust is ubiquitous in the urbanenvironment and its components are associated with numerous humanhealth outcomes, including cancer, respiratory, and reproductive endpoints.Modeling of vehicle exhaust exposure has historically been dominated bytraffic planners, engineers, and government agencies charged with meetingair quality standards. Modeling has been underutilized for exposureassessment in epidemiological studies. Four air pollution models wereevaluated using NO2 field sampling data to assess model performance,predictability, and practical use for epidemiological studies.

Methods: The broad range of models chosen included: proximity(distance-weighted traffic counts), land-use regression (Briggs et al.),dispersion (BREEZE), and integrated meteorological-emission (ADMS-Urban) models. Ogawa samplers (dual NOx/NO2) and Palmes NO2diffusion tubes were used for field sampling. Palmes performance was firstvalidated against three real-time fixed chemiluminescence monitors. Forthe field monitoring, we selected previously used fixed sites with real-timemonitors and 36 public buildings, based on interquartile cutpoints oflevels of estimated census tract diesel PM data. At most sites, samplingwith two Palmes and one dual Ogawa sampler was conducted over a twoweek period, with sample analysis conducted by ion chromatography.Results: The first comparison of Palmes tubes with co-located real-timemonitors suggested a large residual difference between tubes andchemiluminescence monitors and was well correlated (r�0.9) withaverage wind speed. Wind shelters were designed to minimize the effectsof wind turbulence at the Palmes tube inlets. Wind tunnel testing indicatedthe minimum turbulence occurred with the diffusion tube openingrecessed 2 cm from the shelter edge. Repeat of the initial field samplingusing the new shelters yielded much better agreement (Spearmancorrelation � 0.77) between Palmes tubes and real-time monitors. Threeland-use regression models were initially developed based on 10 fixed sitemonitors that explained at least 80% of the variability in NO2 levels,although standard errors were large. We obtained data inputs for ADMS-Urban and generated surfaces for estimated NO2 levels.Discussion: Exhaust from traffic is a major contributor of overall airpollution burden in urban areas. Environmental epidemiologists need toemploy the simplest, most accurate, and least expensive measures oftraffic exhaust. Proximity and integrated meteorological-emission modelsrequire a range of data inputs and expertise. This study illustrates thepracticality of several methods of modeling traffic exposure which arecritical for air pollution studies.

ISEE-142HOME DAMPNESS AND MOULDS AND THE DEVELOPMENT OFASTHMA IN CHILDHOOD: A SIX –YEAR POPULATION-BASEDCOHORT STUDY

Bing-Fang Hwang, Niina Jaakkola. Department of Health Care Admin.,Diwan College of Management, Taiwan

Background: Previous studies of the effects of parental atopy andexposure to dampness and molds on the risk of asthma have mainly beencross-sectional or prevalent case-controls studies, where selection andinformation bias and temporality constitute problems. We assessedlongitudinally the independent and joint effects of parental atopy andexposure to molds in dwellings on the development of asthma inchildhood.Methods: We conducted a population based, 6-year prospective cohortstudy of 1984 children aged 1 to 7 years at the baseline in 1991 (follow-up rate 77%). The data collection included a baseline and follow-upsurvey. The outcome of interest was development of asthma during thestudy period. The studied determinants were parental allergic diseases andfour indicators of exposure at baseline including histories of waterdamage, presence of moisture and visible moulds, and perceived mouldodour in the home.Results: A total of 138 (7.2%) children developed asthma during thestudy period resulting in an incidence rate of 125 cases per 10,000 person-years (95% CI 104–146). In Poisson regression adjusting for confounding,parental atopy (adjusted incidence rate ratio �IRR� 1.52, 95% CI1.08–2.13) and the presence of mould odour in the home reported atbaseline (adjusted IRR 2.44, 95% CI 1.07–5.60) were strong independentdeterminants of asthma incidence, but no apparent interaction wasobserved.



Conclusions: The results of this cohort study with assessment ofexposure before the onset of asthma strengthen the evidence on theindependent effects of parental atopy and exposure to molds on thedevelopment of asthma.

ISEE-143SPATIAL DISTRIBUTION OF WHEEZING PREVALENCE AND AIRPOLLUTION ACROSS NORTH CAROLINA

Karin Yeatts, Marc Serre, Seun-Jae Lee. University of North Carolina atChapel Hill

Introduction: Spatial analyses can be a useful tool in investigatingrelationships between air pollutants common in urban environments andtheir possible health effects. Spatial maps can show the distribution acrossspace of (1) the prevalence of a health outcome and (2) the concentrationin the air of a potentially toxic environmental pollutant. We used theBayesian Maximum Entropy (BME), a modern spatiotemporalgeostatistics method, to analyze the spatial distribution of wheezingprevalence in middle school age children, and six criteria air pollutantsmonitored by the Environmental Protection Agency (EPA). The BMEapproach allows for versatile spatiotemporal analysis and mapping whichcan be used to integrate hard (accurate) and soft (imprecise) informationabout wheezing prevalence, as well as the composite space/time variabilityof criteria air pollutants (Serre et al., 2002). BME gives estimates ofspatial distribution that are more accurate than those obtained using theregression-based methods of classical geostatistics.Methods: Spatial maps of six EPA criteria pollutants concentrations andspatial maps of asthma and wheezing prevalence were generated using theBME framework. Wheezing prevalence data were collected with the NorthCarolina School Asthma Survey, a population-based survey of children(n�122,289) completed in the 1999–2000 school year. Exposures mappedinclude the annual median concentration of each of the six EPA criteriaair pollutants (ozone, NOx, SOx, PM10 and PM2.5, and lead).Results: The visual comparison of the color plots of ozone, PM10 andPM 2.5 with the color plot of the current wheezing data for year 1999suggest that the high prevalence of current wheezing seen in the easternpart of the western piedmont (around the cities of Charlotte and Winston-Salem) may correspond to the high concentration of ozone, PM10 and PM2.5 in those areas. High concentrations of SOx seem to correspond toareas with high prevalence of current wheezing responses in rural areassuch as the western most mountain area and the rural coastal area.Discussion: The high ozone, PM 2.5 and PM10 levels in the Charlottearea in 1999 may be due to urbanization and increased traffic. Theseassociations had not been previously detected. The pattern of high levelsof SOx in rural areas of NC also needs further investigation. Byrigorously processing soft information and space/time variability, the BMEmethod allowed for detection of spatial associations between wheezingprevalence and several EPA criteria air pollutants.

ISEE-144EXPOSURE ASSESSMENT OF MICROBIAL VOLATILE ORGANICCOMPOUNDS (MVOC) AND ORGANIC CONTAMINANTS INNORMAL BEDROOMS OF CHILDREN WITH AIRWAY DISEASE

Caroline Herr,* Simone Schenke,* Susanne Harpel,* Guido Fischer,*Tobias Rethage,* Fatma Ulu,* Alexander Bergmann,*Nikolaos Stilianakis,† Hermann Lindemann,‡ Thomas Eikmann*.*Institute for Hygiene and Environmental Medicine, Medical Center,University Giessen, Germany; †Joint Research Centre, EuropeanCommission, Ispra, Italy; ‡Pediatric Medical Center, Pulmology, MedicalCenter, University Giessen, Germany

Abstract: MVOC may be released indoors from microbial contaminationand other sources leading to the typical moldy odor. They might beassociated with health effects (e.g., mucous membrane irritation). Little isknown about background concentrations of MVOC in- and outdoors andof their distribution during different seasons. 18 children visiting thepulmonology department at the University of Giessen in 2002; 44% girls,39% older than 10 years, 66% with mild or moderate persistent asthma,83% with permanent inhalative medication participated in a year longstudy comprising home inspection, measurement of molds, MVOC(alcohols e.g., 3-octanol, ketones e.g., 3-octanone, sulfides e.g.,dimethlysulfide, and others e.g., geosmine), mite and cat allergens inchildren’s bedroom as well as health status (acc. to ISAAC-questionnaireand health related quality of life in KINDL, derived from SF-36). Thiswhole assessment was performed five times during 12 months in 2003.Active sampling of MVOC and GC/MS-determination and yielded a sumamounting to a minimum of 2.7, median 6.0 and maximum 18.6 �g/m3 atthe first measurement. Molds were differentiated after growth on DG18-agar. Ratios of viable molds (indoor/outdoor) during the year assessedwere: median 0.4, maximum: 56.4. Mite allergen was detected(DUSTSCREEN) more frequently in dust specimens of uncoveredmattresses. 28% of mattress dust samples contained cat allergen althoughno cats were held. The total concentrations for MVOC were higher thanreported previously. Hints for indoor mold growth according to theGerman Federal Environmental Agency were found in five cases inbedrooms during the year. In eight out of 19 bedrooms the concentrationsof Penicillium species were increased compared to background levels.Total MVOC concentrations did not correlate to the mold concentrations(r�0.01, p�0.99) or indoor temperature (r�0.24, p�0.33), as well asventilation habits, insulation, drying washing in homes, storage time ofbiowaste, and smoking. Indoor dampness showed the best correlation tothe sum of measured MVOC (r�0.36, p�0.14). Associations betweenhealth parameters (respiratory complaints) and indoor exposure (molds,MVOC, temperature, humidity) were analysed. No significant associationswere found. There was a tendency (r�0.33, p�0.22) for lower healthrelated life quality (parents reports in KINDL) with higher indoor airMVOC concentrations.

ISEE-145PARTICULATE AND GASEOUS POLLUTANT CONCENTRATIONSIN SOUTH BRONX, NY

Polina Maciejczyk,* Dritan Xhillari,* John H. Offenberg,†George D. Thurston,* Lung Chi Chen*. *New York University School ofMedicine, Nelson Institute of Environmental Medicine; †RutgersUniversity, Department of Environmental Sciences

Abstract: The objective of the South Bronx Environmental Study is tocharacterize the ambient air quality in this community having highconcentrations of diesel trucks and waste transfer facilities. Weemployed a mobile laboratory for continuous measurements of fineparticulate matter (PM2.5), black carbon (BC), oxides of nitrogen,sulfur dioxide, ozone and carbon monoxide at 6 locations for three tofour weeks each during the period of April 2001- July 2003. Integrated24-hr PM2.5 samples were also collected for elemental and PAHsanalyses. Bronx ambient PM2.5 and BC levels were compared tolevels at Bronx P.S. 154 (NYSDEC central monitoring site) and at aLower Manhattan site. Although the median daily PM2.5concentrations agreed within 20% among sites, the concurrent medianhourly BC concentrations were higher at all Bronx sites ranging 2.2 -3.8 ug/m3, compared to Manhattan ranging 1.0 - 2.6 ug/m3. Temporalvariability of both PM2.5 and BC concentrations was apparent at sitesvisited twice; however, the ratios of weekday to weekendconcentrations maintained similar. Median BC concentrations measured



along major highways during morning and evening rush hours over theperiod of 4 weeks at additional 27 sampling sites varied from 1.7 to 12ug/m3 on the weekdays, and were lower (0.5 to 2.9 ug/m3) on theweekends. The sum of 34 elements analyzed by XRF contributed 10.5to 17.3% of the PM2.5 mass. Elemental concentrations were higher atall Bronx sites than at Manhattan for all measured elements but Ni andV, and at the Hunts Point, an industrial location, were approximately2.5-fold higher for Cr, Mn, Fe, Cu, Co, Zn, Ba, and Pb. Analyses ofparticle composition at the Hunts Point by the aerosol time-of flightmass spectrometer yielded more than 370 individual particle classes,clustered in 10 main particle types. In contrast with the analyses ofparticles collected on filter media, nitrate-dominated particle classeswere identified. The mass spectra collected were classified according tothe time of sampling, particle dimension and wind direction. Theconcentrations of the sum of 35 gas phase and particle-bound PAHsranged from 106 to 374 ng/m3, with more than 97.5% of total PAHsfound in the gas phase. Particle-bound PAH composition wasdistinctive with 3,6-dimethylphenanthrene as highest contributor. Windspeed and direction were used to investigate the influence of localsources on gaseous pollutants levels. Correlations of gaseous pollutantconcentrations with BC and time-of-day were also found.

Supported by EPA (R827351, Agreement X-982152) and NIEHS(ES00260).

ISEE-146NEW OPPORTUNITIES FOR EPIDEMIOLOGIC STUDY OF AIRTOXICS IN THE UNITED STATES

Daniel Axelrad, Tracey Woodruff. U.S. Environmental Protection Agency

Abstract: The Clean Air Act designates 188 pollutants as Hazardous AirPollutants (HAPs, or “air toxics”). HAPs have potential to cause a widerange of adverse human health effects, such as cancer, birth defects,respiratory effects and neurodevelopmental effects. Findings regardinghealth hazards of HAPs are based largely on animal studies. There isrelatively little epidemiological evidence of the health effects, apart fromsome occupational studies for selected HAPs. This is largely due to lackof adequate data to characterize exposures. Measurements of ambientHAP concentrations are quite limited, in contrast to the extensivemonitoring data for the criteria air pollutants. Some epidemiologicalstudies have used proximity of residence to certain types of facilities or toroads as a surrogate for exposure to HAPs; however, such surrogates donot account for all sources of HAPs, and cannot distinguish among thevarious HAPs.

In recent years, the U.S. Environmental Protection Agency (EPA) hasdeveloped methods for estimating ambient concentrations of individual airtoxics at the census tract level, for every census tract in the contiguousUnited States. These methods first involve developing an inventory ofemissions from all identified sources of HAPs, which include largeindustrial facilities; smaller, widely dispersed facilities and activities; andboth onroad and nonroad vehicles. The emissions are then combined witha computerized atmospheric dispersion model. EPA has estimatednational, annual-average concentrations of air toxics for 1990 and 1996,with 1999 estimates forthcoming. Comparison of model estimates to thelimited available monitoring data indicate that the modeled estimates arereasonably accurate, with a general tendency to underestimate measuredconcentrations. The 1990 data set has been used in several publishedepidemiology studies.

This talk will discuss the development of these nationwide ambientconcentration data sets, present evaluations of the modeled estimates bycomparison with the available monitoring data, and review studiespublished to date with the 1990 data set. The emergence of these HAPconcentration estimates provide vast new opportunities for further study of

human health effects of exposure to hazardous air pollutants.

Disclaimer: The views presented are those of the authors, and do notnecessarily represent those of the U.S. Environmental Protection Agency.

ISEE-147CAPSAICIN EXPOSURE AS A RISK FACTOR FOR ASTHMA-LIKESYMPTOMS: GAZI� ANTEP RURAL AREA RESPIRATORYSYMPTOMS SURVEY

ZeynepDortbudak,* Ayten Filiz,† Oner Dikensoy,† Nazan Bayram,†Lucas M. Neas‡. *Koc Univ. School of Health Sciences; †GaziantepUniversity Medical School Department of Pulmonology; ‡U.S. EPAEpidemiology and Biomarkers Branch, Human Studies Div.

Abstract: Gaziantep Rural Area Respiratory Syptoms Survey wascompleted between November 2001–December 2002. The aims of thisstudy were to assess the prevalence and risk factors of asthma, chronicbronchitis, atopy and upper respiratory symptoms in a randomized sampleof rural villages with population of 200–400. The survey questionnairewas developed through an elaboration of existing standardized respiratorysymptoms questionnaires (2) and a pilot study conducted in the regionafter increased numbers of patients were observed to be referred to theGU Medical School’s Department of Pulmonology with asthma-likesymptoms following domestic processing of red pepper which are knownto contain Capsaicin, a bronchoconstrictor (1). The current report consistsof a questionnaire data and pulmonary function testing of a partialdatabase (672 subjects from 46 villages).

Logistic regression analysis of questionnaire data on symptoms related topepper processing and wheeze (WHEEZE) and asthma-like (AS-LIKE)symptoms consisting of at least three of wheeze without colds, night timeapnea as well as chest tightness symptoms has revealed a statisticallysignificant two-fold increase of these symptoms among study participantswith eye irritation (EI) and respiratory problems (RP) that reportedlylasted more than three days after pepper processing. GI (OR 2.06. CI1.45–2.91), RP (OR 2.00, CI 1.38–2.90). We attribute these findings tothe possible role of Capsaicin in the mechanism of allergic sensitisationleading to asthma-like symptoms and find them noteworthy. Twoadditional studies are planned to investigate this mechanism amongasthmatics that report to our clinic and among seasonal workers in thelocal pepper-processing factories in the region.

This study was supported by Gaziantep University Research Foundationand University of North Carolina, Center for Environmental Medicine andLung Biology. Disclaimer: This publication does not reflect the views orpolicy of the United States Environmental Protection Agency.

ISEE-148THE USE OF A CREATININE CORRECTION FOR REPORTINGCHILDREN’S URINARY PESTICIDE CONCENTRATIONS

Ruth Allen,* David Mage,† Gauthami Gondy,† Carol Christensen,*Dana Barr,‡ Larry Needham‡. *U.S. Environmental Protection Agency;†ISR Temple University; ‡CDC

Abstract: A ’spot sample’ urinary concentration of a pesticide ormetabolite in units of ug/liter has uncertain meaning because variable fluidintakes make large differences in the excreted concentrations. A correctionfor this variable hydration in adult subjects has been to report pollutantconcentration in urine as ug analyte / mg creatinine, which tends tostandardize the data because adults excrete creatinine at a relativelyconstant daily rate (mg/kg body mass) that is a function of age, genderand body surface area (BSA) given normal kidney function. Thevariability of this corrected value for differences in muscularity and diet(vegetarian vs. omnivorous) is small compared to the variability of



hydration, which is why this correction has been developed.

One objection to reporting data as ug analyte/mg creatinine ratio for allages is that children who are growing rapidly have developingmusculature and excrete creatinine at a much lower daily rate per kg bodymass than adults. This would make their standardized ratio much higherrelative to an adult for equal concentrations in units of ug analyte/liter.We propose that this objection can be removed if all such ug analyte/mgcreatinine ratios were multiplied by the subject’s individualized estimateof creatinine production (mg creatinine/kg body mass/day). The resultingquantity, ug analyte/kg body mass/day, is in units of U.S. EPA’s RfD fororal ingestions of pesticide compounds and can be compared to it whenthe mass of analyte, if a metabolite, is multiplied by the ratio of mass ofparent pesticide compound to the mass of the analyte.

We review the process of creatinine excretion and the various equationsfound in the peer-reviewed medical literature for infant – child –adolescent daily creatinine clearances that were developed for the purposeof computing the appropriate pediatric medicinal dose of pharmaceuticalsper kg body mass. These pediatric creatinine clearances are compared tothose we have developed in a similar manner for adults, and both theirasymptotes at age 18 are compared to demonstrate the consistency fortheir combined usage over all ages, if the age, gender and BSA of thesubject are known. We propose that such a consistent method of reportingall urinary concentration as ug pesticide/kg body mass/day allows all datato be compared on the same basis and that it has a direct relationship to ahealth effect for that pesticide.

ISEE-149TRACE ELEMENTS IN TOENAILS AS A PREDICTOR OFENVIRONMENTAL HEALTH RISKS IN AN ARAB AMERICANPOPULATION

Melissa Slotnick,* Jerome Nriagu,* Mary Johnson,* Aaron Linder,*Adnan Hammad,† Kathryn Savoie,† Hikmet Jamil†‡. *Department ofEnvironmental Health Sciences, School of Public Health, University ofMichigan; †Community Health and Research Center, ACCESS, Dearborn,MI; ‡Department of Family Medicine, Wayne State University, Detroit,MI

Introduction: Assessment of exposure to trace elements inepidemiological studies is complicated by multiple exposure pathways,and therefore is improved through the use of biological markers. Toenailsare often used as a biomarker of trace metal exposure, as many elementsbind to the fibrous proteins in nails. In contrast to use of serum or plasmametal levels, levels of trace elements in toenails reflect average exposureto metals over 6-12 months and may be a more suitable biomarker forexploring the relationship between trace element status and disease.Methods: Metropolitan Detroit, birthplace of the automotive industry, ishome to the largest Arab population in North America and continues toserve as a reception area for Middle Eastern immigrants. Exposure tocurrent and historic industrial sources, low socioeconomic status, andbarriers to health care place this population at high risk for environmentaldisease. Based on a previous study, 160 high risk households wereidentified in four Detroit Area communities. Respondents provided toenailclippings and completed questionnaires regarding health and economicstatus, demographics, and exposure to environmental risk factors. Toenailsamples were washed, hot-block digested, and analyzed for Al, V, Cr,Mn, Co, Ni, Cu, Zn, As, Se, Mo, Cd, Ba, Tl, and Pb using an InductivelyCoupled Plasma Mass Spectrometer (ICP-MS).Results: Logistic regression models, adjusted for age, gender andsmoking, were run for 18 doctor-diagnosed diseases and reveal severalsignificant relationships with toenail trace-element concentrations (Table1).

Table 1: Significant Associations between Trace Element Concentration inToenails and Disease Status

Disease Element OR (95% CI)

Allergies Cu 0.36 (0.14, 0.88)

Chronic fatigue Co 2.3 (1.2, 4.3)

Diabetes Co 2.1 (1.2, 3.6)

Diabetes Cu 2.8 (1.1, 7.2)

Diabetes V 6.7 (2.1, 21.7)

Diabetes Cr 5.5 (1.7, 17.2)

Diabetes Zn 5.3 (1.2, 24.37)

Diabetes Cd 2.9 (1.4, 6.2)

Emphysema/Chronic Bronchitis Mn 14.2 (1.1, 186)

High Blood Pressure Mn 2.2 (1.2, 4.0)

High Cholesterol V 2.6 (1.1, 5.8)

Other lung/breathing condition Co 3.2 (1.3, 7.7)

Reactive airway disease Se .002 (�.001, 0.323)

Significant Pearson correlation coefficients exist between metals releasedfrom local industry, including Cr and Cd, Mn and Pb, and Cr and Cu,suggesting a possible common source of exposure. Higher mean valuesfor Al, Ni, Co, Cr and Cd in this study compared with other USpopulations supports this conclusion. Discussion: Diet, metabolism,disease status, and exposure to environmental contaminants all contributeto trace element concentration in nails. The observed associations betweenconcentrations in toenails and disease status are likely reflective of acombination of these factors. Exposure to cobalt has previously beenassociated with lung diseases, and the association with selenium may lendinsight into the role of antioxidants in disease protection. The positiveassociation between diabetes and other trace metals is particularlynoteworthy, suggesting that the disease may be influencing the metaboliccycling of these elements. This study provides insight into the associationbetween trace element status and disease, establishing a basis for futureinvestigations in this and other communities. This project was partiallysupported by the NIEHS-Environmental Justice Program.

ISEE-150EFFECTS OF THE GSTM1 AND GSTP1 POLYMORPHISMS ONTHE LEVELS OF SERUM IGE IN RELATION TO AIR POLLUTION

Yun-Chul Hong,* Lee Kwan-Hee,* Heon Kim,† Eun-Hee Ha,‡Daehee Kang,§ Soo-Hun Cho§. *Inha University College of Medicine;†Chungbuk University College of Medicine; ‡Ewha Women’s UniversityCollege of Medicine; §Seoul National University College of Medicine

Introduction: The aims of this study were to evaluate the use of urinary2-naphthol levels as a means of measuring exposure to air pollution, andto investigate the genetic susceptibility of GSTM1 and GSTP1polymorphisms to allergic response and inflammatory reactions in relationto air pollution exposure.Methods: We measured urinary 2-naphthol levels for 289 healthyuniversity students. Serum immunoglobulin E, serum C-reactive protein,and white blood cell were measured to evaluate the allergic response andinflammatory reactions. We determined GSTM1 and GSTP1polymorphisms by PCR and RFLP.Results: Urinary 2-naphthol levels were significantly different dependingon smoking status (P�0.001). The urinary 2-naphthol levels for studentswho smoked 20 cigarettes or more were 61.86 mg/g creatinine, whereasthose for non-smoking students were 30.57 mg/g creatinine. Variables foroutdoor air pollution such as distance from main road, traffic volume, orpersonal NO2 exposure were not associated with urinary 2-naphthol levels



significantly. However, the outdoor air pollution factors contributed togive better explanation of the urinary 2-naphthol levels.

Individuals with GSTP1 variant genotypes showed significantly increasedserum IgE levels compared with students with GSTP1 wild genotype inthe group of higher urinary 2-naphthol levels (P�0.041). In the lowerurinary 2-naphthol levels, the difference of IgE levels between GSTP1wild and variant types did not reach statistical significance (P�0.05).GSTM1 null individuals had increased levels of serum IgE, however, theIgE levels did not differ significantly between GSTM1 wild and nullgenotypes (P�0.05). For serum C-reactive protein and white blood cellcounts, GSTM1 and GSTP1 polymorphisms did not show significantdifferences in any group of urinary 2-naphthol levels.Discussion: Our data indicate that urinary 2-naphthol is a biomarker ofexposure to various PAH sources, including smoking and outdoor airpollution. GSTP1 variant genotype with higher levels of urinary 2-naphthol was associated with increased serum IgE levels, suggesting theGSTP1 polymorphism identify a genetically susceptible people forenhanced allergic response due to air pollution.

ISEE-151SERUM LEVEL OF EPIDERMAL GROWTH FACTOR RECEPTORAS A BIOMARKER OF LIVER CANCER

How-Ran Guo,* Tzu-I Sung,* Ying-Jan Wang,* Tsui-Lien Hung,*Chi-Yi Chen†. *National Cheng Kung University; †Chia-Yi ChristianHospital

Introduction: Epidermal growth factor is associated with the occurrenceof various cancers, including those of the skin, lung, urinary bladder,prostate, breast, and stomach. Therefore, epidermal growth factor receptor(EGFR) may be used as a biomarker of malignancies. However, itsassociation with liver cancer has not been extensively studied. Weconducted a study in an endemic area of liver cancer to evaluate whetherserum EGFR may increase in patients of liver cancer.Methods: We recruited 100 patients of liver cancer and 100 age- andsex-matched controls from two hospitals in southern Taiwan from June2000 to April 2002. The incidence of liver cancer in this area is highbecause of environmental factors such as viral hepatitis. The levels ofEGFR in sera were determined by enzyme-linked immunosorbent assay(ELISA).Results: A significant higher serum EGFR level was observed amongpatients (668.1 vs. 243.1 fmol/mL, p�0.01). After adjusting for history ofhepatitis B and C through multiple regression, patients still had an averageserum EGFR level 406.1 fmol/mL higher than that of controls (p�0.01).Discussion: We concluded that EGFR is over-expressed in patients ofliver cancer, and therefore, serum EGFR level is a potential biomarker ofliver cancer. In particular, it is increased in patients of liver cancers notrelated to viral hepatitis and thus might be a biomarker of liver cancercaused by other environmental pollutants. Further studies should beconducted to define the role of EGFR in the development of liver cancerand identify environmental factors that may cause liver cancer throughpathways related to the activation of EGFR.

ISEE-152EPIDEMIOLOGICAL CHALLENGES OF USING PESTICIDEMETABOLITE MEASUREMENTS TO DEFINE HUMAN EXPOSURE

George E. Luber. Centers for Disease Control and Prevention, NationalCenter for Environmental Health

Introduction: The recent development of novel laboratory techniques tomeasure pesticide metabolite concentrations in humans has opened up newopportunities and challenges for the epidemiological study of pesticide

exposure. These new techniques come at an opportune moment as the useof pesticides to control vector-borne diseases, such as West Nile virus andEastern Equine Encephalitis, have become increasingly important in recentyears. Public concern about the risks of exposure to mosquito controlpesticides in residential areas has fueled interest in measuring thecontribution of broad-scale pesticide applications to individual pesticidemetabolite levels.Methods: This presentation will describe some recent investigations toquantify human exposure to mosquito control pesticides and discuss someof the epidemiological challenges that arose during these investigations.Results: One of the major challenges in quantifying exposure topesticides is the interpretation of non-specific pesticide metabolite data.While these laboratory techniques are extremely sensitive and accurate,they are sometimes unable to discern closely related pesticides as theyoften are metabolized into common breakdown products. Epidemiologicalsolutions to this problem are presented.

Previous research has indicated that household, dietary, and occupationalexposures to pesticides may be significant sources of the total bodyburden and need to be accounted for in all exposure assessments. Whileself-reported questionnaires have been used to characterize household andoccupational exposure, they are often of limited value due to recall bias.One solution lies in a prospective study design, with the collection of pre-exposure urine and environmental samples, which allows for thequantification of baseline metabolite levels prior to the exposure underinvestigation. Collecting data to verify the time and location of the sprayevent, as well as the location of the exposed population, is anotherimportant element to these studies. We discuss our use of GlobalPositioning Systems (GPS) and Geographical Information Systems (GIS)to model the spray swath and determine the time of exposure. Among thestatistical considerations we discuss are the analytic constraints incollecting multiple samples per location (i.e., household) while comparingpre- and post-exposure data.Discussion: While we present several epidemiological challenges andconsiderations for the collection and interpretation of these data, thesetypes of studies are important as they provide a unique opportunity toquantify a significant non-occupational exposure to a large number ofpeople.

ISEE-153THE EFFECT OF ARSENIC EXPOSURE ON A URINARYOXIDATIVE STRESS BIOMARKER IN WOMEN: STUDY DESIGN

Carrie Breton,* Mahnuder Rahman,† Golam Mahiuddin,†Andres Houseman,* Quazi Quamruzzaman,† David Christiani*. *HarvardSchool of Public Health; †Dhaka Community Hospital

Purpose: Arsenic ingestion has been linked to chronic cough andbronchitis, hypertension, ischemic heart disease, peripheral neuropathy,gangrene in the leg, and various cancers yet arsenic toxicity mechanismsin humans remain largely unknown. Arsenic may act by generatingreactive oxygen species and increasing oxidative stress, which can causeDNA damage and may contribute to carcinogenesis or cardiovasculardisease. This study aims to further the understanding of arsenic toxicity inhumans by evaluating the effect of arsenic exposure on whole bodyoxidative stress levels using the urinary biomarker 8-Hydroxy-2�-deoxyguanosine (8-OHdG). By measuring urinary 8-OHdG’s variabilityover three consecutive samples per person, its usefulness as a biomarkercan also be evaluated.Methods: A repeated measures cohort study design was employed. 113women with at least one urine sample were recruited from within anestablished cohort of 50 families from a district in a Southwestern Asiannation with known arsenic pollution in drilled well water. Three urinesamples per individual were taken on consecutive days to present a stable



estimate of oxidative stress. Total arsenic exposure was measured indrinking water samples from family tube wells and from toenail samplescollected at the time of first urine collection. Urinary total arsenic,speciated arsenic, and arsenic metabolites were measured from aliquots ofthe same samples used for 8-OHdG analysis.

A multilevel mixed linear regression model incorporating random effectswill be employed to allow for correlations due to clustering of familiesand repeated measures within individuals. The model will be run severaltimes for the different measures of arsenic exposure. The association oftotal arsenic consumed and 8-OHdG levels will also be evaluated using amixed model that dichotomizes arsenic exposure into high and lowcategories using the new WHO cut-off of 10 ug/L. Confounding bysmoking, age, BMI, sun exposure and education will be controlled.Expected Results: Individual arsenic levels measured in drinking water,toenails and urine are expected to be highly correlated. High arsenic levelsare expected to predict increased levels of urinary 8-OHdG. Urinary 8-OHdG may exhibit a fair degree of inter- and intra-individual variabilityover time.Conclusions: This study presents a novel approach to further understandarsenic toxicity in humans by evaluating whether high levels of arsenicaffect oxidative stress levels in the body. Arsenic may act in part througha reactive oxygen species pathway. In addition, the use of repeatedmeasures for biomarkers of exposure and outcome reduces within-personvariability, better enabling the measurement of true effects betweenindividuals.

Supported by NIH grants ES011622 and ES05947.

ISEE-154AN EPIDEMIOLOGICAL ASSESSMENT OF AN INDIVIDUALSBIOLOGICAL UPTAKE OF PARTICULATE AIR POLLUTION ANDRELATED HEALTH EFFECTS

Susan O’ Connell. University of Wales College of Medicine

Abstract: Epidemiological evidence shows that exposure to ambientenvironmental particulate air pollution has an adverse effect on the healthof the population, particularly respiratory and cardiovascular health, butmuch of this evidence is ecological. This study aims to show thatresidential exposure to particulate pollution as a result of vehicularemissions has an effect at an individual level.

An exposure comparison study has been designed to show whetherhealthy individuals residing in proximity to high volumes of traffic havebiological uptake of pollutants and suffer effects such as inflammatoryresponses, increased oxidative stress and changes in blood properties suchas coagulation and viscosity.

ISEE-155DIOXINS IN THE IRISH POPULATION, A STUDY OF HUMANBREAST MILK AND A CHALLENGE IN INFANT EXPOSUREASSESSMENT

Dominique Crowley,* Margaret Fitzgerald,† Mary O Mahony,‡Lousie Doherty, Elaine Scallan,§ Noelle Millar,‡ Nicki Clarke,�Michael Geary,¶ Iona Pratt,§ Wayne Anderson§. *Department of PublicHealth Medicine and Epidemiology, University College Dublin;†Department of Public Health, Eastern Regional Health Authority;‡Department of Public Health, Southern Health Board; §Food SafetyAuthority of Ireland; �National Maternity Hospital, Holles Street, Dublin;¶Rotunda Hospital

Abstract: Technological advances in the last century have resulted in anincreasing number of chemicals in the environment. Environmentalexposures of particular concern in Ireland are those substances knowncollectively as dioxins and the potential for infant exposure through breastmilk.

The aims of this study which was part of a larger WHO coordinatedproject were: (a) to produce reliable and comparable data on levels ofdioxins in human milk to further improve health risk assessment ininfants; (b) to provide an overview of exposure levels and to determinetrends in exposure levels in the participating countries and areas; and (c)to identify highly exposed local populations.

Four pooled samples were collected from first-time mothers in urban, sub-urban, rural and industrial areas. Each sample comprised 10 separatebreast milk samples from mothers living in the area for the past five yearswith a first-born infant aged between 2-6 weeks.

The median value for PCCDs/PCCDFs of 7.72 pg/g fat reflects arelatively low level compared to many other countries participating in thisstudy. Similarly, the level of dioxin-like PCBs, with a median value of4.57 pg/g fat, is also lower than most. The levels were also well belowthose found in most of the other European countries who participated inthis study.

Estimates of exposure to infants through breast-feeding, averaged over the6-month breast feeding period were was 23.65 pg WHO TEQ/kg b.w.,while the intake of dioxin-like PCBs was 13.24 pg WHO TEQ/kg. Theselevels are amongst the lowest in Europe. The benefits of breast-feeding faroutweigh any potential adverse effect and should continue to be promotedas the best option for both mother and infant. The absence of referencevalues for infant exposure makes these results difficult to interpret inrelation to lifetime exposure risk.

The results provide the first information relating to human exposure todioxins in Ireland. The body burden, as measured in breast-feedingmothers, was relatively low by European standards.

This study has been valuable in providing baseline measurements. Irelandshould continue to participate in future rounds of the international breastmilk studies.

Food is one of the main sources of exposure and it is important tocontinue to selectively monitor Irish foodstuffs, on an ongoing basis.Should municipal waste incineration be introduced, it would be equallyimportant to monitor emissions for the presence of these compounds andothers that could enter the food chain.

The authors would like to thank the support of the mothers in carrying outthis study. Also Dr. Declan Keane, National Maternity Hospital HollesStreet; Vincent Young, Dublin Public Analysts Laboratory; The ScientificCommittees of the Food Safety Authority of Ireland and Dr. RainerMalisch and Dr. Gerard Van Leeuwen co-ordinators of the internationalstudy, World Health Organisation.

ISEE-156MERKEL CELL CARCINOMA OF SKIN WITHOUT REGULAREXPOSURE TO SUNLIGHT AS A MARKER OF ARSENICEXPOSURE

Sheng-Yow Ho,* Helen H.W. Chen,† Yi-Chang Tsai,* How-Ran Guo†.*Sin-Lau Christian Hospital; †National Cheng Kung University

Introduction: Merkel cell carcinoma (MCC) is a rare cutaneousneoplasm with an aggressive biological behavior. MCC mostly developson the sun-exposed skin of elderly people, appearing on the trunk or atother sun-protected locations in less than 10% of cases. Little is knownabout specific factors in the pathogenesis of MCC. Arsenic exposure hasbeen implicated in the pathogenesis. We conduct a study in the balckfootdisease (BFD) endemic area, an endemic area of arsenic intoxication from



drinking well water in Taiwan, to evaluate the association between arsenicexposure and anatomical distribution of MCC.Method: We recruited cases from a teaching hospital near the BFDendemic area from 1988 to 2003 and also included cases ascertained froma teaching hospital in Taipei in a previous study. Taipei and its vicinityhad no remarkable arsenic in the drinking water. Demographic andclinical data were collected on each case. The diagnoses of all cases wereconfirmed by pathologists. We compare the distribution of lesions betweenthose who were from the BFD endemic area and those who from areaswithout arsenic exposures.Results: A total of 13 cases were included in the analyses. Among theeight cases from the BFD endemic area, five (62.5%) had lesions in areasnot regularly exposed to sunlight, including four on the chest wall and oneon the abdominal wall. In contrast, none of the cases from the areas had alesion in areas not regularly exposed to sunlight. The relative risk ofhaving a lesion in areas not regularly exposed to sunlight was about 17.3among those who came from the BFD endemic area.Discussion: The etiology of BFD was not clear for decades after thedocumentation of this disease. The observation of typical skin lesions ofarsenic intoxication in patients of BFD lead to speculation that BFD iscaused by arsenic intoxication, which was later supported byepidemiological studies. Therefore, these typical lesions, includinghyperpigmentation, hyperkeratosis, and some types of skin cancers such asBowen’s disease, basal cell carcinoma, and squamous cell carcinoma,were regarded as marker arsenic intoxication. Our study revealed thatMCC in areas not regularly exposed to sunlight is also associated witharsenic exposure. Therefore, the distribution of such skin lesions may beindicative of arsenic exposures.

ISEE-157CHROMOSOME ABERRATION ANALYSIS IN PERIPHERALLYMPHOCYTES OF GULF WAR AND BALKANS WAR VETERANS

Heike Schroeder,* Anna Heimers,* Wolfgang Hoffmann,†Rainer Frentzel-Beyme*. *University of Bremen, Germany; †University ofGreifswald, Germany

Abstract: British veterans of the Gulf War 1990/1991 or/and the BalkansWars 1995/1996 and 1999 who suffered from various medical complaintssince then volunteered to participate in the present pilot-study. Allreported situations during their active service associated with exposure todepleted uranium by inhalation of uranium oxide in its aerosol form.Exclusion criteria were previous radiation, medical use of cytotoxic drugs,heavy smoking, and previous work in badge-monitored occupations. Anappropriate control group with respect to all of the multiple warfareagents the investigated veterans had been exposed to (including multi-vaccinations) and ill with comparable complaints, but documented absenceof DU contamination, was not available, our own laboratory control waschosen to evaluate the findings of the study. Average ages were 40.1 (29– 57) and 35.1 (17 – 57), resp.

Average aberration yieldof mean dicentr. and

ring chromos.Conf. interval

(95%)

Veterans’ group 2.60/1000 (2.2–4.9/1000)

Laboratory control group 0.46/1000 (0.3–0.8/1000)

There was a 5.2-fold increase among the exposed compared to the control(p�0.001).

Among the volunteers no sample was without chromosomal aberrations.

Frequencies of dicentric and ring chromosomes in 8 from 16 cases differsignificantly in individual contrasts (Fisher’s exact test: p�.05).

Stratification according to deployment did not alter the results.

Major bias in the chromosomal analysis seems unlikely since allvolunteers and controls were analysed blindly by the same experiencedscorers.

Veterans were exposed besides DU-dust to different agents on thebattlefields. However, since dicentric chromosomes are reliable indicatorsof ionizing radiation, these findings contradict official releases fromIAEA, WHO, MOD and DOE, stating that the radiotoxicity of DU wouldbe negligible.

Computer simulations have also calculated only little radiological riskassociated with the use of DU weapons, hence, it can only be speculatedabout the mechanisms behind the observed cytogenetic effects,considering the relatively low specific radioactivity of DU. Modern airsampling techniques have shown hundreds of thousands of DU particles intwo selected samples from Kosovo in a few milligrams of contaminatedsoil, indicating that two years after the end of the war “spots” at differentsites hit by DU rounds remain and DU dust was widely dispersed into theenvironment, since it was discovered even in Hungary.

The environmental impact of DU dust has been pondered, our resultsindicating persistence of uranium in the organism, which was confirmedby animal experimental evidence.

The authors wish to thank the British National Gulf Veterans’ andFamilies’ Association for cooperation and for support by grants from theWorld Depleted Uranium Center, Berlin, Germany (WoDUC), as well asits director Prof. Albrecht Schott, for scientific advice.

ISEE-158MULTIPLE GENE POLYMORPHISMS AS HUMANSUSCEPTIBILITY FACTORS TO MULTIPLE OUTCOMES OFPERSISTENT ORGANIC POLLUTANTS (POPS)

Wenya Huang, Pei-Chien Tsai, Yueliang Leon Guo. College of Medicine,National Cheng Kung University

Purpose: Polychlorinated biphenyles (PCBs), dibenzofurans (PCDFs),and dibenzodioxins (PCDDs) are ubiquitous environment pollutants thatare persistent in the environment. In populations highly exposed to thesechemicals, several health outcomes, including skin, liver, thyroid, etc.have been reported. We examined whether genetic polymorphisms aresusceptibility factors for these outcomes in PCB/PCDF-exposedpopulation.Methods: We followed-up a population of 1837 people in centralTaiwan, who were exposed to polychlorinated biphenyls (PCBs) anddibenzofurans (PCDFs) due to accidental consumption of contaminatedrice oil in 1979. Chloracne, nail changes, abnormal liver function, andother symptoms were reported by this population. In 1993, a completereview of systems by phone interview was done. Recently blood wascollected and genetic materials obtained after informed consent. Geneticpolymorphisms on aryl hydrocarbon receptor (AhR) gene, cytochromeP450 (CYP) 1A1 gene (MspI and NcoI digestion sites), and twoglutathione S-transferase (GST) isoenzymes GSTM1/T1 genes were doneby polymerase chain reaction (PCR) and restriction fragment lengthpolymorphism(RFLP).Results: A total of 776 highly exposed subjects completed phone-interview for skin manifestations. Among them, 594 blood PCB levelswere previously measured in 1980s. We completed genotyping of 327such subjects. Skin allergy, chloracne, abnormal nails, abnormal liverfunction, and hyperkeratosis were reported in 27%, 22%, 16%, 8%, and5% respectively. Geometric mean of PCB levels in 1980s was 48.9 ppb.The prevalences of chloracne, abnormal nails, hyperkeratosis, skin allergy,and abnormal liver function were associated with blood levels of PCBs.The odds ratio for having chloracne was 2.13 (95% confidence interval



1.02-2.37) for CYP 1A1 MspI homozygous mutant type (m2/m2) only inhigher PCBs-exposed group, but not in lower exposed group. Liverdisease was associated with GST-T1 null type (odds ratio 4.8, 95%confidence interval 2.0-14.6) when chloracne was taken as an indicator forexposure. Other outcomes were not associated with genotypes tested inthis study.Conclusion: Genetic polymorphisms including CYP 1A1 and GST-T1were important predictors for health effects in PCBs/PCDFs exposedpopulation.

This study is supported by grants #NSC91-3112-B-006-006 from theNational Science Council, Taiwan.

ISEE-159EXPRESSION OF VON HIPPEL LINDAU GENE AS EXPOSUREAND OUTCOME MARKER FOR PCB/PCDF EXPOSEDPOPULATION

Pei-Chien Tsai,* Wenya Huang,* Yeu-Chin Lee,† Yueliang Leon Guo†.*National Cheng Kung University, Tainan, TAIWAN; †Environmental andOccupational Health, National Cheng Kung University

Purpose: Polychlorinated dibenzodioxins (PCDDs), polychlorinateddibenzofurans (PCDFs), and polychlorinated biphenyls (PCBs) areubiquitous environmental pollutants with multiple human and wildlifehealth effects. The effects of these chemicals on gene expression profilesin human peripheral lymphocytes have not been documented. We usedcDNA microarray approach to study a group of people highly exposed toPCBs/PCDFs to identify gene expression as biomarkers in the peripherallymphocytes.Methods: In 1979, a mass poisoning of more than 2000 people occurredin central Taiwan due to consumption of rice–bran cooking oilcontaminated with PCBs and PCDFs. The exposed people, later calledYucheng (“oil-disease” in Chinese) victims, suffered from fatigue,chloracne and other skin manifestations, peripheral neuropathies, liverfunction abnormalities, and other signs and symptoms. The exposedpeople had high serum levels of PCBs and PCDFs 20 years afterexposure. In 2002, peripheral blood of Yucheng men and unexposedcontrols were collected by venopuncture. The mRNAs extracted fromlymphocytes were reversely transcribed into cDNAs and labeled withbiotin, then hybridized to the cDNAs on the biochip. After thehybridization, the chips were applied to high resolution scanner for imageanalysis. We used the Millennium cDNA nylon membrance microarraychips, with 972 human cDNA, especially those involved in the pathwaysrelated to carcinogenesis and signal transduction, prepared by the Centerof Microarray of National Cheng Kung Medical College. Fourteenendogenous human housekeeping genes (such as G3PDH, beta-actin,HPR1, etc.) were used as comparison. The expression was furtherconfirmed by real time-polymerase chain reaction (PCR) by LightCycler(Roche Co.) in combination with the dsDNA binding dye SYBR Green I.Serially diluted PCR product of known concentration was used toconstruct a standard curve for quantification.Results: A total of 53 exposed and 35 control men participated in thisstudy, with age 42 years. Among the exposed men, 42 had ever hadserum levels of PCBs measured in 1980-1982, which averaged 81 ppb.The von Hippel Lindau (VHL) gene expressions were found to beincreased with older age, and suppressed in exposed population asscreened by microarray study, and confirmed by real time-PCR. Amongexposed subjects, VHL gene expression was negatively associated withhigher levels of PCBs measured in 1980-1982, and was reduced in thosewith chloracne, abnormal nails, or skin allergies, the most importantmanifestations associated with Yucheng.Conclusion: The VHL gene expression in peripheral lymphocytes maybe exposure and outcome biomarkers for PCBs/PCDFs exposure.

This study is supported by National Health Research Institute of Taiwan,Grant #NHRI-EX90-9025PL.

ISEE-160URINARY PAHS METABOLITES AND OXIDATIVE STRESSBIOMARKERS AS ENVIRONMENTAL AIR POLLUTION INCHINESE AND KOREAN

Kyoung Ho Lee,* Dong Ho Yoo,* Zhong Min Li,† Ho Jang Kwon,‡Yun Chul Hong,§ Soo Hun Cho,* Daehee Kang*. *Department ofPreventive Medicine College of Medicine Seoul Nat’l University, Institutefor Environmental Medicine SNUMRC, Seoul, Korea; †College of publicHealth Science, Jilin University, Changchun, China; ‡Department ofPreventive Medicine College of Medicine, Dankook University;§Department of Preventive Medicine, College of Medicine, InhaUniversity, Korea

Abstract: The aim of the study were to see the association betweenenvironmental particulate air pollution and urinary polycyclic aromatichydrocarbons (PAHs) metabolites and oxidative damage biomarker amongchildren and their mothers in South Korea and in China and to see if thereis any differences in urinary 1-hydroxypyrene glucuronide (1-OHPG), 2-naphthol and malondialdehyde (MDA) levels in children (6-15 years old)and their mothers living three cities in Korea (Seoul, Inchon and Pohang)and three in China (Changchun, Datong and Kunming), where the levelsof particulate air pollution varies.

The study subjects consisted of 346 Koreans (175 children and 171 theirmothers) and 308 Chinese (155 children and 153 their mothers). Urinary1-OHPG was measured by synchronous fluorescence spectroscopy afterimmunoaffinity purification using monoclonal antibody 8E11. Urinary 2-naphthol was measured by HPLC with fluorescence detector. And, MDAwas measured by HPLC with UV detector. Information on consumption ofdiet containing high PAHs, environmental tobacco smoke, type of cookingand heating fuels, and so on were collected by self-administeredquestionnaire.

Average ambient total suspended particulates (TSP) in Changchun inChina (0.316 mg/m3) were 3- to 6 fold higher than those in Seoul inKorea (0.080 mg/m3). Average urinary 1-OHPG levels in Chinese (1.61 �5.74 �mol/mol creatinine) were higher than those in Korean (0.26 � 0.45�mol/mol creatinine) (p�0.01). And urinary 2-naphthol levels (30.26 �148.67 �g/g creatinine) in Chinese were also 2- to 3-fold higher thanthose in Korean (10.28 � 40.81 �g/g creatinine) (p�0.01). Urinary 1-OHPG levels were significantly higher in winter (Korean: 0.44 � 0.62,Chinese: 4.18 � 7.96 �mol/mol creatinine) than in summer (Korean: 0.20� 0.26, Chinese: 0.88 � 1.54 �mol/mol creatinine) among Korean(p�0.01) and Chinese (p�0.01). Urinary 2-naphthol levels weresignificantly higher in winter (Korean: 13.53 � 46.71, Chinese: 38.53 �83.67 �g/g creatinine) than in summer (Korean: 7.13 � 31.23, Chinese:22.39 � 202.41 �g/g creatinine) among Chinese (p�0.02) and Korean(p�0.01). Average urinary MDA levels in Chinese (0.55 � 1.29 mg/gcreatinine) were higher than those in Korean (0.12 � 0.09 mg/gcreatinine) (p�0.01). There was a significant positive correlation inurinary 1-OHPG levels between mothers and children living at the samehousehold. Urinary 1-OHPG levels were significantly correlated withurinary 2-naphthol in Chinese (n�200, r�0.54, p�0.01) and Korean(n�192, r�0.45, p�0.01). Urinary 1-OHPG levels (Korean: n�118,r�0.49, p�0.01, Chinese: n�120, r�0.40, p�0.01) and urinary 2-naphthol (Korean: n�116, r�0.42, p�0.01, Chinese: n�112, r�0.30,p�0.01) were also significantly correlated with urinary MDA. Multiplelinear regression analysis indicated that use of well water (vs. tap water)were significant predictors for urinary 1-OHPG levels (overall modelr�0.41, n�134), 2-naphthol (r�0.23, n�93) and MDA (r�0.33, n�56)



in Chinese.

These results indicated that urinary 1-OHPG, 2-naphthol and urinaryMDA can be used as a biomarker of environmental particulate airpollution containing PAHs. And, urinary 1-OHPG levels, 2-naphthol andMDA were related with drinking water.

ISEE-161BIOMONITORING OF HUMAN EXPOSURE TO POLYCYCLICAROMATIC HYDROCARBONS AND DIESEL EXHAUST BYMEASUREMENT OF URINARY BIOMARKERS

Zheng (Jane) Li, Lovisa C.S. Romanoff, Kisha J. Young,Nelson C. Blakely III, Raymond Wei, Larry L. Needham,Donald G. Patterson, Jr., Courtney D. Sandau. Centers for DiseaseControl and Prevention

Abstract: Polycyclic aromatic hydrocarbons (PAHs) are a class ofpollutants that exist ubiquitously in the environment and are formedduring incomplete combustion processes of organic materials, such ascoal, gasoline, oil, wood, and cigarette. Among many sources, dieselexhaust contains not only PAHs, but also considerable amounts of nitro-PAHs that are potentially more carcinogenic and mutagenic. 1-Nitropyreneand 3-nitrobenzanthrone have been isolated from diesel fumes, and havebeen suggested as good markers for diesel exhaust exposure. Humans canbe exposed to PAHs and nitro-PAHs through inhalation, ingestion, ordermal contact from a wide spectrum of combustion products.Epidemiological data have shown that exposure to PAHs is potentiallycarcinogenic to humans, as exemplified by elevated cancer incidences incertain occupations such as coke oven workers.

Human exposure to environmental PAHs and nitro-PAHs can be assessedthrough biomonitoring levels of their urinary metabolites, i.e., hydroxy-PAHs and amino-PAHs. An automated urinary assay that simultaneouslymeasures 23 OH-PAHs, including metabolites of benzo�a�pyrene, has beendeveloped in our laboratory at the Centers for Disease Control andPrevention (CDC). This method is being expanded to measure dieselbiomarkers and other 11 amino-PAHs in human urine. The methodologyincludes enzymatic hydrolysis, automated solid phase extraction,derivatization, and final analysis by gas chromatography - high resolutionmass spectrometry (GC-HRMS). The method is robust, precise, and highlyreproducible. Limit of detection for most OH-PAHs are in low part-per-trillion level.

The measurement of urinary PAH metabolites allows the assessment of anindividual’s internal dose of PAHs, and provides epidemiologists andhealth scientists with unique information to evaluate health effectassociated with PAH exposure. This method will be applied for theanalysis of samples from the National Health and Nutrition ExaminationSurvey (NHANES), performed by CDC annually to assess exposure ofgeneral U.S. population to these environmental pollutants, as well as othercase-control studies.

ISEE-162URINARY PAH METABOLITES AND OXIDATIVE STRESSBIOMARKERS FOR THE ASIAN DUST EVENTS

Dong-Ho Yoo,* Kyoung-Ho Lee,* Zhong Min Li,† Soo-Hun Cho,*Ho-Jang Kwon,‡ Deahee Kang*. *Department of Preventive Medicine,Seoul National University College of Medicine, Institute forEnvironmental Medicine SNUMRC, Seoul, Korea; †College of PublicHealth Science, Jilin University Changchun, China; ‡Department ofPreventive Medicine, Dankook University College of Medicine, Chonan,Korea

Introduction: Asian Dust Events (ADEs) are the dust storm originatedfrom Mongolia and China, which can also get to the West Coast of USA.The ADEs were reported to be associated with increased daily mortalityin Seoul, Korea from previous epidemiological studies. This studyevaluated the potential usefulness of urinary 1-hydroxypyrene glucuronide(1-OHPG) and 2-naphthol as environmental aromatic hydrocarbons(PAHs), 8-hydroxy-deoxyguanosine (8-OHdG) and malondialdehyde(MDA) as oxidative stress markers related to the ADE.Materials and Methods: Urine samples were collected from 224subjects (112 children and 112 their mothers) from Seoul (n�60), Inchon(n�104) and Pohang (n�60) twice before and after the ADE (April 13,2003). Urinary 1-OHPG levels were measured by synchronousfluorescence spectroscopy after immuno-affinity purification usingmonoclonal antibody 8E11. Urinary 2-naphthol levels were measured byHPLC with fluorescence detector. Levels of urinary 8-OHdG weremeasured by using ELISA and levels of urinary MDA by HPLC with UVdetector.Results: Levels of urinary 1-OHPG and 2-naphthol after the ADE werehigher than those before the ADE, which were not significant. Levels ofurinary 8-OHdG after the ADE (7.76 � 2.57 ng/ml) were higher thanthose before the ADE (2.45 � 4.4) (n�30 paired, p�0.05). Levels ofurinary MDA after the ADE (0.44 � 1.82 umol/L) were higher than thosebefore the ADE (0.38 � 2.0) (n�60 paired, p�0.1). There were asignificant correlation between levels of urinary 1-OHPG and 8-OHdGboth after (n�78, r�0.44, p�0.001) and before the ADE (n�78, r�0.42,p�0.001). Urinary 1-OHPG and MDA levels were correlated with afterADE (n�178, r�0.22, p�0.01) and before ADE (n�178, r�0.17,p�0.05). Urinary MDA and 2-naphthol levels were also correlated withafter ADE (n�176, r�0.29, p�0.01) and before ADE (n�176, r�0.24,p�0.01).Conclusions: Although the ADEs were very mild in this year thanprevious years, our findings suggest that urinary 8-OHdG and MDA levelsincreased after the ADE and these biomarkers could be used as usefulbiomarkers for the ADE.

ISEE-163A CORRELATION STUDY OF ORGANOCHLORINE LEVELS INSERUM, BREAST ADIPOSE AND GLUTEAL ADIPOSE TISSUEAMONG BREAST CANCER CASES IN INDIA

Jennifer Rusiecki,* Aleyama Matthews,† Susan Sturgeon,‡Rashmi Sinha,§ Dalsu Baris*. *Occupational and EnvironmentalEpidemiology Branch, Division of Cancer Epidemiology and Genetics,National Cancer Institute, National Institutes of Health, Department ofHealth and Human Services; †Division of Epidemiology and ClinicalResearch, Regional Cancer Centre; ‡Department of Public Health,Biostatistics and Epidemiology Program; §Nutritional EpidemiologyBranch, Division of Cancer Epidemiology and Genetics, National CancerInstitute, National Institutes of Health, Department of Health and HumanServices; �Occupational and Environmental Epidemiology Branch,Division of Cancer Epidemiology and Genetics, National Cancer Institute,National Institutes of Health, Department of Health and Human Services

Background: In order to investigate the relationships betweenorganochlorine levels across different biological media, we utilizedmeasurements in serum, breast adipose and gluteal adipose tissue frombreast cancer patients enrolled in a pilot study carried out in Kerala, Indiain 1997. Our objective was to determine whether serum is a reliablebiomarker of lifetime body burden of organochlorines compounds. Thiswas a unique population to study, with respect to organochlorinecompounds, since some (DDT and �-BHC) were still commonly used inIndia until recently for malaria control, while others (PCBs) are not soabundant in Kerala.



Methods: Biological samples were collected from 37 untreated breastcancer patients, who were fasting at time of sample collection. Gas-liquidchromatography was used to determine serum, breast adipose and glutealadipose tissue levels of organochlorines. Summary statistics (mean,median, geometric mean), Spearman correlation coefficients (r) and ratiosof the summary statistics were calculated for both crude and lipid-corrected values of DDT, DDE, �-BHC, and one highly chlorinated PCBcongener, PCB-180.Results: Levels of DDT (medianserum �97.4 ng/g lipid) and itsmetabolite, DDE (medianserum �619.4 ng/g lipid), were high, comparedwith levels reported throughout the literature; levels of �-BHC(medianserum �2,818.2 ng/g lipid) were orders of magnitude higher;PCB-180 concentrations (medianserum �5.07 ng/g lipid) were the lowestreported. There were strong correlations between serum, breast adiposeand gluteal adipose tissue concentrations for all organochlorines measured,ranging from r�0.65 to 0.95. For each summary statistic, ratios wereclose to 1:1 for most of the comparisons between lipid-corrected serum,breast adipose, and gluteal adipose samples.Discussion: To our knowledge, this is the first study to investigate thecorrelations and ratios of organochlorines levels measured across threebiological media. The strong correlations between biological media didnot depend on the timing of exposure, since DDT and �-BHCexposures were likely recent, DDE reflects older exposures to DDT,and PCB exposure in this population was likely ambient, nor did theydepend on whether exposures were high or low. We concluded thatmeasurements in blood serum reflect lifetime body burden to the extentthat breast and gluteal adipose tissue do. This is important to confirm,since it is much less invasive to collect a blood sample than anadipose tissue sample.

ISEE-164

CURRENTLY THE DEPARTMENT OF DEFENSE (DOD) DOES NOTUSE EXPOSURE BIOMARKERS TO MEASURE SERVICEMEMBERS EXPOSURE TO ENVIRONMENTAL CHEMICALS

Jack Heller,* Coleen Weese,* Lisa May,† Arthur Lee*. *U.S. ArmyCenter for Health Promotion and Preventive Medicine; †UniformedServices University of the Health Sciences

Abstract: Currently the Department of Defense (DoD) does not useexposure biomarkers to measure Service Members exposure toenvironmental chemicals. Blood and urine exposure biomarkers forvolatile organic compounds (VOC), selected heavy metals, depleteduranium (DU), and chemical warfare agents are currently available buthave not been field tested or validated in military deployments as a tool todocument exposures by the DoD. The Military Deployment HumanExposure Assessment Study, a prospective cohort of 46 soldiers deployedto Bosnia, was designed to field test blood and urine exposure biomarkersas a mechanism to document exposures to these chemicals during militarydeployments. Blood and urine were collected pre-, during, and postdeployment. Standard questionnaire, environmental and occupationalmonitoring methods were conducted for comparison to the exposurebiomarker results. This paper compares and describes the pre-, during, andpost deployment blood VOC results, compares these to standard US bloodVOC levels, reports deployment environmental and occupationalmeasurements, and attempts to correlate environmental with blood VOCresults. The study outcomes indicate that exposure biomarkers may bevaluable tools to the DoD in exposures and risks from environmental andoccupational chemicals.

ISEE-165

ENVIRONMENTAL EXPOSURES AND OTHERCHARACTERISTICS ASSOCIATED WITH DETECTABLE PAH-DNA ADDUCTS AMONG A POPULATION-BASED SAMPLE OFHEALTHY WOMEN

Sumitra Shantakumar,* Marilie D Gammon,* Sybil M Eng,†Sharon K Sagiv,* Mia M Gaudet,* Susan L Teitelbaum,‡Julie A Britton,‡ Lian Wen Wang,§ Qiao Wang,§ Steve D Stellman,§Jan Beyea,� Maureen Hatch,¶ David Camann,** Alfred I Neugut,§Regina M Santella§. *Department of Epidemiology, School of PublicHealth, University of North Carolina at Chapel Hill; †Pfizer, Inc,Epidemiologic Resources, Safety Evaluation and Epidemiology;‡Department of Community and Preventive Medicine, Mt. Sinai School ofMedicine; §Mailman School of Public Health, Columbia University;�Consulting in the Public Interest, Inc.; ¶Department of Population andGenetics, National Cancer Institute; **Southwest Research Institute

Introduction: The presence of polycyclic aromatic hydrocarbon (PAH)-DNA adducts in human lymphocytes may be useful as a surrogateendpoint for individual cancer risk prediction. In this study, we examinethe relationship between environmental sources of PAH and otherparameters on the presence of PAH-DNA adducts in a large population-based sample of healthy women.Methods: Adult women over 20 years of age were randomly identifiedfrom the general population between August 1996 and July 1997. Amongthe 1,556 women who completed a structured questionnaire, 941 donatedsufficient blood (25� ml) to allow use of a competitive ELISA foranalysis of PAH-DNA adducts in peripheral blood mononuclear cells.Recent ambient PAH levels were estimated using geographic modeling(n�796). Environmental home samples of dust (n�220) and soil (n�197)were colleted on a subset of long-term residents (15� years). A predictivemultiple logistic regression was fit to estimate odds ratios for detectablelevels of PAH-DNA adducts (n�648) compared with non-detectablePAH-DNA adducts (n�293). Backward selection was used to identifyenvironmental measures or other characteristics that appreciablycontributed to the model. Three separate models were constructed basedon data from the: (A) questionnaire, including a dietary history; (B)environmental home samples; and (C) geographic modeling.Results: Season of blood donation and active cigarette smoking werestrongly associated with detectable PAH-DNA adducts. Compared withother seasons, women who donated blood in summer had over a two-foldincreased odds of detectable adducts (OR�2.65, 95% CI�1.69, 4.17).Current and past smokers had a 50% and 46% increased odds ofdetectable adducts, respectively. There were inconsistent associationsbetween detectable adducts and other known sources of PAH such asgrilled and smoked foods, or a summary measure of total dietary benzo-a-pyrene (BaP) intake during the year prior to the interview. Other factorsassociated with detectable adducts included increased age, increasedincome, age at menarche, and fewer months of breastfeeding. Detectableadducts were inversely associated with increased BaP levels in dust, butpositively associated with BaP levels in soil although confidence intervalswere wide. Ambient BaP estimates were not associated with dectectableadducts.Discussion: These data suggest that PAH-DNA adducts detected amonga population-based sample of adult women with ambient exposure levelsreflect some key PAH exposure sources assessed in this study, such ascigarette smoking, but not all.

For their valuable contributions to the Long Island Breast Cancer StudyProject the authors thank: members of the Long Island Breast CancerNetwork; the 31 participating institutions on Long Island and in NewYork City, NY; our National Institutes of Health collaborators, GwenColman, Ph.D., National Institutes of Environmental Health Sciences; G.



Iris Obrams, M.D., Ph.D. formerly of the National Cancer Institute;members of the External Advisory Committee to the population-basedcase-control study: Leslie Bernstein, Ph.D., (Committee chair); GeraldAkland, M.S.; Barbara Balaban, MSW; Blake Cady, M.D.; Dale Sandler,Ph.D.; Roy Shore, Ph.D.; and Gerald Wogan, Ph.D.; as well as othercollaborators who assisted with various aspects of our data collectionefforts including Gail Garbowski, MPH; H. Leon Bradlow, Ph.D.; MartinTrent, B.S.; Ruby Senie, Ph.D.; Carla Maffeo, Ph.D.; Pat Montalvan;Gertrud Berkowitz, Ph.D.; Margaret Kemeny, M.D.; Mark Citron, M.D.;Freya Schnabel, M.D.; Allen Schuss, M.D.; Steven Hajdu, M.D.; andVincent Vinceguerra, M.D. This work was supported in part by grantsfrom the National Cancer Institute and the National Institutes ofEnvironmental Health and Sciences (Grant nos. UO1CA/ES66572,P30ES09089, and P30ES10126), the Breast Cancer Research Foundation,and the Babylon Breast Cancer Coalition.

ISEE-166CAN DIOXIN BODY BURDEN DATA BE USED TO ESTIMATELONG TERM RESIDENTIAL EXPOSURE TO INDUSTRIALPOLLUTION?

Tanja Pless-Mulloli,* Richard Edwards,† Olaf PaepkeM,‡Thomas Herman‡. *University of Newcastle upon Tyne Medical School;†Evidence for Population Health Unit, School of Epidemiology andHealth Sciences; ‡ERGO Research Laboratory

Introduction: Body burden measures may act as markers for long-termexposure if biomarkers can be found with sufficiently long half lives inthe human body. We tested the hypothesis that women with long termresidence close to an industrial complex have a higher body burden and adistinct pattern of dioxins, furans (PCDD/F) and polychlorinated biphenyls(PCBs).Method Design: Observational study of 40 women living near (0.1-2.7km) and distant (5-40 km) from an industrial complex on Teesside, UK.Subjects and setting: Subjects were recruited from participants in acase-control study of residential exposure to air pollution and lung cancer.Residences were assigned to exposure zones: A�near, B�intermediate,C�distant from industry; informed by land-use surveys, dispersionmodelling and monitoring. We recruited 20 women from zone A (10 casesand 10 controls) and 20 from zone C (10 cases and 10 controls)sequentially between Jan 00 and Dec 03. Women with weight loss of�3.5 kg over last 3 months or who had lived for less than 20 yearsconsecutively in their current zone were excluded.Data collection: Information on: occupational exposure; residentialhistory; current and past consumption of local animal and vegetableproduce; breast feeding; current and typical diet; and self-reported heightand weight was collected at interview. Whole blood was analysed forPCDD/F and PCBs.Results: Mean age was 62.5 years for Zone A participants and 65.3years in zone C. Mean BMI, and lifelong dietary patterns were similarexcept Zone A participants had slightly higher consumption of fish andslightly lower consumption of meat.

Zone A(near

industry)

Zone C(away from

industry)

Differencein means(95% CI)

Number 20 20

Mean

WHO-TEQ (DF�PCB) 45.3 48.6 3.3 (13.2 to 6.7)

Zone A(near

industry)

Zone C(away from

industry)

Differencein means(95% CI)

WHO-TEQ (DF) 29.1 30.8 1.7 (8.0 to 4.6)

WHO-TEQ (PCB) 16.3 17.9 1.7 (5.9 to 2.6)

2378-TCDD 4.0 4.1 0.2 (1.2 to 0.8)

PCB 118 13993 18357 4364 (10878 to 1345

PCB 156 13180 12973 208 (2569 to 2984)

PCB 118:156 1.2 1.4 0.2 (0,7 to 0.3)

We found no significant difference in body burden for any of theindicators between zones A and C participants. All congener patterns wereconsistent with an urban background pattern. In multivariate linearregression no significant differences were found in any body burdenmeasures between participants form zones A and C, after adjusting forpotential confounders. The only consistently significant predictor ofPCDD/F and PCB body burden was age.Discussion: This was a population based study with much detail onconfounding information. The findings suggest that the intake of PCDD/Fand PCB through background food consumption were such a dominantsource of exposure that impacts from local point sources wereundetectable. Body burden of these biomarkers were not a useful indicatorfor long term residential exposure in this setting.

Funding was received from �Fight against cancer in Teesside’ and TeesHealth Authority. Table 1. Body burden of PCDD/F and selected PCBs inZones A and C in ng/kg lipid base.

ISEE-167URINARY ARSENIC METABOLITES IN RELATION TOEXPOSURE VIA FOOD AND WATER

Marie Vahter,* Tony Fletcher,† Peter Rudnai,‡ Walter Goessler,§Giovanni Leonardi,† Eugen Gurzau,� Kvetoslava Koppova,¶Anna-Lena Lindberg*. *Institute of Environmental Medicine, KarolinskaInstitute; †Public and Environmental Health Research Unit, LondonSchool of Hygiene and Tropical Medicine; ‡Division of EnvironmentalHealth Impact Assessment, National Institute of Environmental Health,�Jozsef Fodor’ National Center of Public Health; §Institut fur Chemie -Analytische Chemie, Karl-Franzens-Universitat; �Environmental HealthCentre, Cluj-Napoca; ¶State Health Institute, Banska Bystrica

Introduction: Against the background of the widespread exposure to lowlevel arsenic and the need for further scientific evidence on dose-responserelationships in the low dose range, we are presently conducting a threecountry case-control study to quantify the cancer risks (bladder, skin, andkidney) in relation to arsenic ingestion via drinking water. In addition toarsenic, we assess risk in relation to nutrition (including blood selenium)and inter-individual differences in arsenic metabolism and genotype (forselected cancer- and metabolism-related genes).Methods: For about 1000 cases and 600 controls, estimates ofcumulative dose via drinking water will be constructed. Evaluation ofcurrent exposure and metabolism pattern is based on measuredconcentrations of inorganic arsenic and its metabolites in urine, using ahighly sensitive arsenic speciation methodology (HPLC-HG-ICP-MS).Detection limits for the four arsenic compounds arsenite �As(III)�,dimethylarsinate (DMA), methylarsonate (MA) and arsenate �As(V)� were0.1, 0,1, 0.1 and 0.5 �g As/L, respectively. Speciation is being achievedacross a wide range of exposures, including a significant proportionexposed at close to zero water As concentrations.



Results: Data collection is underway, and will be complete by themeeting when fuller results will be presented, but preliminary results aregiven below (in �g As/L).Discussion: Preliminary results suggest that inorganic as well asmethylated arsenic species are present in urine of individuals exposed tovery low concentration of arsenic in water. Arsenic metabolite pattern willalso be evaluated in relation to type of cancer, geographical area,exposure level, genotype and selenium status.

ISEE-168

EYE DISCOMFORT, HEADACHE AND BACK PAIN AMONGWOMEN IN GUATEMALA USING OPEN FIRES FOR COOKINGAND HEATING

Tone Smith-Sivertsen,* Nigel Bruce,† Anaite Diaz,‡ Esperanza Diaz,*Morten Alexander Schei,* Daniel Pope,† Asheena Khalakdina,§John McCracken,� Byron Arana,‡ Kirk R. Smith,§ Robert Klein‡.*University of Bergen, Norway; †University of Liverpool, UK;‡Universidad del Valle, Guatemala; §University of California, Berkeley;�Harvard University, Boston

Introduction: Indoor air pollution from biomass combustion in openfires affects rural populations in developing countries all over the world.We will, in the first randomised control trial for air pollution ever donewith healthy populations, study the effect of installing improved chimneystoves (planchas) on women’s health. The plancha significantly reducesindoor air pollution levels. A positive side effect for the women may bethe change in working posture from use of an open fire on the floor to anupright position while cooking. Here we present preliminary results fromthe baseline assessment of self-reported eye discomfort, headaches, andback pain.Methods: This study is conducted in a Mayan Indian community inthe Guatemalan highlands and includes only households using an openwoodfire for cooking and spaceheating. It comprises 504 women �meanage 27.7 years (SD 7.5), 32.5% pregnant� visited in their homes bytrained native-speaking fieldworkers. As part of a more comprehensiveinterview, questions were asked about eye discomfort, headache andback pain. In addition, the women had their carbon monoxide (CO)levels in exhaled breath determined as a proxy for recent air pollutionexposure.Results: In total 63.3% of the women had experienced headacheduring the last month. Of these, 29.6% had symptoms every day, and32.3% reported their headache to be strong (Table 1). Sore or wateryeyes during the last month was reported by 41% of the women, and56% of these were bothered every day (Table 2). The risk of eyediscomfort increased with increasing levels of CO in exhaled breath,while no such relation was found for headache (Table 3). Back painduring the last month was reported by 48.6% of the women (Table 4),pregnant women having higher prevalence (OR 1.5, 95% CI 1.0-2.2).Of those with back pain, 30 (12.3%) could not perform daily dutiesdue to the pain. A number of factors were reported to make back painworse, most important being washing clothes (55.9%) and cooking(28.6%) (Table 4).Discussion: Headache and eye discomfort were common symptoms inthis female population exposed to high levels of indoor air pollution.Our results suggest that eye discomfort was linked to personalexposure levels. Back pain was common and reported to be associatedwith cooking by more than a quarter of the affected women. Animproved stove has the potential to prevent back pain in these women.

Table 1. Self-reported headache during the last month.1

N %

Headache, total 319 63.3

Among these2:

- Headache every day 94 29.6

- Headache most days of the week 42 13.2

- Headache a few days per week or less often 182 57.2

Severity of headache

- Strong 104 32.6

- Intermediate 185 58.0

- Mild 30 9.4

1Have you had headache during the last month?2Information on frequency missing from one woman.

Table 2. Self-reported prevalence of eye discomfort during the last month.1

N %

Eye discomfort, total 207 41.0

Among these:

- Eye discomfort every day 116 56.0

- Eye discomfort most days of the week 21 10.1

- Eye discomfort a few days per week or less often 70 33.7

1Have you had sore or watery eyes during the last month?

Table 3. Determinants of eye discomfort and headache.1

Eyediscomfort Headach

OR 95% CI OR 95% CI

Age (per year) 1.06 1.04–1.09 1.03 1.00–1.06

Pregnancy 1.78 1.20–2.66 1.43 0.95–2.14

CO in exhaled breath (per ppm increase) 1.06 1.01–1.11 1.03 0.98–1.08

1Logistic regression analysis. Odds ratios are mutually adjusted for eachother.

Table 4. Self-reported prevalence of back pain during the last month.1

N %

Back pain, total 245 48.6

Among these:2

- Backpain every day 94 38.7

- Symptoms most days of the week 35 14.4

- Symptoms a few days per week or less often 114 46.9

Aggravating factors

- Carrying wood 48 19.6

- Washing clothes 137 55.9

- Cooking 70 28.6

- Other3 73 29.8

Could not perform daily duties because of back pain 30 12.3

1Have you had back pain during the last month?2Information on frequency missing from two women.3For instance lifting/carrying child, carrying water, grinding corn.



GIS-MODELED INDICATORS OF MOBILE SOURCE EMISSIONS AND ADVERSE HEALTH EFFECTS AMONG CHILDREN IN EL PASO, TEXAS, USA

Documents