What Is Benchmark Dose (BMD) and How to Calculate BMDL

Usually we use benchmark dose (lower confidence limit) (BMDL) to calculate human health guidance value since it is more conservative.  > Topics > CRA > ToxicologyandHealthRiskAssessment CRA WhatIsBenchmarkDose(BMD)andHowtoCalculateBMDL LittlePro on 2018-08-01 Inchemicalriskassessment,bothno-observed-adverse-effectlevel(NOAEL)andstatisticalbenchmarkdose(BMD)canbeusedaspointofdeparture(POD)toderivehumanhealthguidancevaluesuchasreferencedose(RfD)orderivedno-effectlevel(DNEL) oracceptabledailyintake(ADI).NowadaystheBMDapproachhasbecometheUSEPA’spreferreddose–responseassessmentmethod.OtherauthoritiessuchasEuropeanFoodSafetyAuthority(EFSA)alsousetheBMDmethodforfoodsafetyriskassessment.Inthisarticle,wewillgiveyouanintroductiontothebenchmarkdose(BMD)approachandshowtohowtocalculatebenchmarkdose(lowerconfidencelimit),alsoknownasBMDL. WhatIsBenchmarkDose(BMD)andBenchmarkResponse(BMR) Abenchmarkdose(BMD)isadoseorconcentrationthatproducesapredeterminedchangeintheresponserateofanadverseeffect.Thispredeterminedchangeinresponseiscalledthebenchmarkresponse(BMR). Normally,thedefaultBMRis5%or10%changeintheresponserateofanadverseeffectrelativetotheresponseofcontrolgroupdependingonwhetherresponsedataiscontinouousorquantal(dichotomous). ResponseData Examples DefaultBMR Continuousdata Thevaluesofabiologicalparameterinindividualanimals Bodyweight Cellproliferation Countofredbloodcells 5%(EFSA) 10%(EPA) Quantaldata Thefractionofanimalswithaspecificeffectinapopulation Tumourincidencerate Percentageofanimalsdevelopinganadverseeffect 10% AfteracriticaleffectandtheBMRhavebeendetermined,differentmathematicalmodelsareappliedtofitthedose-reponsedataandestimatethebenchmarkdose(BMD). Itshouldbenotedthateachfitted(andaccepted)modelresultsinaconfidenceintervalfortheestimatedBMD.TheBMDwegetfromstatisticalmodelisarange,ratherthanafixednumber(pleaseseepicturebelow).Usuallyweuse benchmarkdose(lowerconfidencelimit)(BMDL) tocalculatehumanhealthguidancevaluesinceitismoreconservative.TheBMDLcanberegardedasadoseatwhichtheeffectissmaller thantheBMR(withdefinedconfidence). HowtoUseBMDLinRiskAssessment ThemaingoalofcalculatingBMDListouseittoestimateadailyoralordermalexposureleveltothehumanpopulation(includingsensitivesubgroups)thatislikelytobewithoutanappreciableriskofdeleteriouseffectsduringalifetime.  RfDorDNELorADI=POD(NOAELorBMDL)/UF(Uncertaintyfactors) Note:UFrangesfrom100to10,00.Readmore.. AfterRfDorDNELhasbeenderived,itwillbecomparedwithacutualorestimatedhumanexposureleveltodetermineifexposureriskisacceptableornot.  ComparisonofBMDMethodandNOAELApproach BMDLismorereliablethanNOAELsinceitislessdependentondoseselectionandsamplesize.J.AllenDavis,JeffreyS.GiftandQ.JayZhaofromtheUSEPAhavemadeaverygoodcomparisonofBMDmethodandtraditionalNOAELapproachintheirresearchpaper[Read..] BMDadvantages NOAELlimitations •Notlimitedtoexperimentaldoses •Highlydependentondoseselection •Lessdependentondosespacing •Highlydependentonsamplesize •Appropriatelyaccountsforvariabilityanduncertaintyresultingfromstudyquality •Doesnotaccountforvariabilityanduncertaintyintheexperimentalresults(e.g.,doesnotaccountforstudyqualityappropriately) •Takesintoaccounttheshapeofthedose–responsecurveandotherrelatedinformation •Dose–responseinformation(e.g.,shapeofdose–responsecurve)nottakenintoaccount •Correspondstoconsistentresponselevelandcanbeusedtocompareresultsacrosschemicalsandstudies •Doesnotcorrespondtoconsistentresponselevelsforcomparisonsacrossstudies •Flexibilityindeterminingbiologicallysignificantrates •ALOAELcannotbeusedtoderiveaNOAEL BMDlimitations NOAELadvantages •AbilitytoestimateBMDmaybelimitedbytheformatofdatapresented •CanbeusedwhendataisnotamenableforBMDmodeling •Timeconsuming •Easytoderive •Morecomplicateddecisionmakingprocess •HasbeenthestandardmethodforderivingaPODfordecades(e.g.,isfamiliartomostriskassessors) HowtoCalculateBMDL WecancalculateBMDorBMDLusingUSEPA'sbenchmarkdosesoftware(BMDS)orRIVM’sPROASTpackage.Bothsoftwarepackagesareinternationallyrecognized.PersonallyIpreferEPA'ssoftwaresinceitiseasiertouse. BeforewetaketheBMDapproach,wemustevaluatewhetheravailabletoxicologydataissuitableforBMDmodelling.AvailabledatamustmeetthefollowingcriteriaforBMDmodelling. Reportedresponsedatamustbeeitherquantaldataorcontinuousdata.  Thereisacleardose-responsetrend. Therearesufficientdosegroups(minimum3dosinggroups+1controlgroup).DatasetsinwhichresponseisonlyobservedatthehighdoseareusuallynotsuitableforBMDmodeling Thedoseresponse-modelshouldfitthedataadequatelybysomepredefinedcriterion(e.g.P-value>0.1). Oncewehavedeterminedwhetherreportedresponsedataisquantalorcontinuous,wecanchoosedichotomousorcontinuousmodelstofitthedoseresponsecurveinBMDS.TheBMDSsoftwarewillthenperformthefittingandcalculation(seepicturebelow). Often,morethanonemodelormodelingoptionswillresultinanacceptablefittothedata,currentEPAguidance(2012TG)providesguidanceonhowtopickasingle“best”model: Whatis“sufficientlyclose”canvarybasedontheneedsoftheassessment,butgenerallyshouldnotbemorethan3-fold. IfBMDLsarenotsufficientlyclose,EPArecommendspickingthemodelwiththelowestBMDL IfBMDLsaresufficientlyclose,EPArecommendsselectingthemodelwiththelowestAIC IfmultiplemodelshavethesameAIC,EPArecommendscombiningBMDLs BMDLvsNOAEL ThevalueofcalculatedBMDLcanbebiggerorsmallerthanNOAEL.Whentestsample(i.e,no.oftestedanimals)isverybig,BMDLis biggerthanNOAEL.Whenthesizeofsampleissmall,BMDLmaybesmallerthanNOAEL. References IntroductiontobenchmarkdosemethodsandU.S.EPA'sbenchmarkdosesoftware(BMDS) WorkshopconfirmsBMDapproachasthebestmethodfordose-responsemodellinginriskassessment-EFSA HavingQuestions? Wedonotprovideconsultancyservices.Ifyouhavequestionsorneedanyhelp,pleasecontactoursponsor.YoumayalsofindanexpertinCSPbusinessdirectorybelow.Ifyouareaconsultant,youmaygetyourselflistedinCSPbusinessdirectory(free)or sponsorthispagetoleaveyourcontactinfoonthispage..  Tags:Topics-CRA, ToxicologyandHealthRiskAssessment WhatIsNew?   SubscribetoFreeNewsletter  ViewPastNewsletter RecommendedforYou TopNewsoftheMonth TSCA Brexit EUChemicalStrategyforSustainability ChemicalsinElectronics PFASRegulation BiocidesandBiocidalProductRegulations MoreInfo ReportsandGuides EventsandTraining Webinars Podcasts HotArticles