TheRelationshipbetweenEnvironmentalMonitoringandBiologicalMarkersinExposureAssessmentS.M.Rappaport,'E.Symanski,1J.W.Yager,23andL.L.Kupper''SchoolofPublicHealth,UniversityofNorthCarolina,ChapelHill,NorthCarolina;2SchoolofPublicHealth,UniversityofCalifornia,Berkeley,CaliforniaThepoorqualityoftraditionalassessmentsofexposurehasencouragedepidemiologiststoexplorebiologicalmonitoringinstudiesofchronicdis-eases.Yet,despitetheoreticaladvantages,biomarkershavenotbeenwidelyusedinsuchapplications.Thisarticlecomparesthegeneralutilityofabiomarkerwiththatofthemeasurementofexposureperse.Pointsareillustratedwithalongitudinalstudyofboatworkersinwhichlevelsofstyreneinthebreathingzoneandinexhaledairwerecomparedtosisterchromatidexchanges(SCEs)inperipherallymphocytes.First,thelinearrelationshipisexploredbetweenpersonalexposureandthelevelsofabiomarkerinthecohort.Agoodfittothestraight-linerelationshipreflectedbyacorrelationcoefficientwhichiscloseto1,suchasobservedwithstyreneinexhaledair(r2=0.83),suggestslinearkinetics,thattheappropriaterouteofexposurewasmeasuredbypersonalmonitoring,smallinterindividualdifferences,adequatesamplesizes,andaspecificbiomarker.2However,asmallcorrelationcoefficient,asobservedbetweenSCEsandstyreneexposure(r=0.1),indicatesthateitherkineticswerenonlinearorthatmorecomplexissueswereinvolvedwithoneormoreofthesefactors.Second,environmentalandbiologicmeasurementsarecomparedforuseasindependentvariablesinestablishingastraight-linerelationshipbetweenexposureandthehealtheffect.Iftheratioofthewithin-persontothebetween-personcomponentsofvarianceoftheindependentvariableislarge,thensignificantattenuationresultswhenestimatingtheslopeoftheline.Sincesuchattenuationcanbereducedbymakingrepeatedmeasurementsoneachpersoninthecohort,thesamplesizesrequiredtoreducethebiasto afixedlevelcanbeusedtocomparethevariousmeasuresofexposure.Usingdatafromthestyrene-exposedworkers,itisshownthattheslopebeingestimatedwouldbewithin10%ofthecorrectslopeparameterwith3personalmeasurementsofexposurecomparedto4samplesofexhaledair(12measurements)and20assaysofSCEs.Thus,inthiscase,themeasurementofairborneexposurewouldbemoreefficientthanthatofeitherexhaledairorSCEsforepidemiologicpurposes.-EnvironHealthPerspect103(Suppl3):49-54(1995)Keywords:biomarkers,exposureassessment,styrene,exhaledair,SCEs,measurementerror,dose-responserelationship,epidemiology,variancecomponentsIntroductionAmajorgoalofoccupationalandenviron-mentalepidemiologyhasbeentoevaluatetherelationshipsbetweenexposurestohaz-ardoussubstancesandtherisksofdisease.Unfortunately,fewexposure-responserela-tionshipshavebeenelucidatedbecauseoftheappallinglackofhistoricexposuredata.Thispaucityofinformationaboutexpo-surehasfosteredapplicationsofindirectmethodstodefinepastexposures(1)andThispaperwaspresentedattheConferenceonHumanTissueMonitoringandSpecimenBanking:OpportunitiesforExposureAssessment,RiskAssessment,andEpidemiologicResearchheld30March-1April1993inResearchTrianglePark,NorthCarolina.ThisworkwassupportedbygrantR010H02221fromtheNationalInstituteforOccupationalSafetyandHealthoftheCentersforDiseaseControlandPreventionandbygrantP42ES04705oftheNationalInstituteofEnvironmentalHealthSciences.TheauthorsappreciatetheassistanceofL.DionnefororganizingthefieldsurveysandofR.C.Yu,whomodeledtheexposure-doserelationships.AddresscorrespondencetoProfessorS.M.Rappaport,CB#7400,UniversityofNorthCarolina,ChapelHill,NorthCarolina27599-7400.Telephone(919)966-5017.Fax(919)966-4711.hasencouragedthenotionthatbiomarkersshouldbeusedprospectivelytodefineexposuresratherthanlevelsofcontami-nantsinair,water,orfood(2).Inthiscon-text,theterm"biomarker"referstoameasureofexposureintheformof"...anexogenoussubstanceoritsmetaboliteortheproductofaninteractionbetweenthexenobioticagentandsometargetmoleculeorcellthatismeasuredinacompartmentwithinanorganism"(3),andnottomea-suresofeffectorofsusceptibility,whichhavedistinctdefinitions(3).Inweighingtheadvantagesanddisad-vantagesofbiologicandenvironmentalmonitoringtodefineexposure,itisclearthatbiomarkersenjoyatleastthreetheo-reticaladvantagesandonemajordisadvan-tageovermedia-specificmeasurements.Ontheplusside,somebiomarkerscansub-stantiallysmooththeextremevariabilityinexposuretypicalofenvironmentaltoxicantsandtherebyreducethemonitoringeffort(4-6).ThisisillustratedinFigure1,whichshowstheexposuresofsixworkerstoinorganiclead(FigureIA)andtheasso-ciatedlevelsofbloodleadamongthesameindividuals(Figure1B)[datatakenfromCopeetal.(7)].Itisalsotruethatbiologi-calmonitoringaccountsforallroutes,i.e.,inhalation,ingestion,anddermalabsorp-tion,andtherebyprovidesameasureofthetotalexposurereceivedbytheindividual(3,8,9).Finally,variousbiomarkersaccountfordifferencesintheuptake,elim-ination,metabolism,andrepairoftoxicsubstancesamongexposedpersons(3,8,9).However,ontheminusside,thesamplingandanalyticdemandsofbiomarkersaregenerallygreaterthanthoseassociatedwithenvironmentalmeasurementsandcanleadtoreductionsinsamplesizesatagivencost.Thus,itremainstobeseenwhetherbiolog-icalmonitoringwillsupplantenvironmen-talmonitoringinfutureinvestigations.Sincefewstudieshavebeenpublishedwhichcomparelevelsofbiomarkerswithexposuresinthesamepopulation,it