搜索
Guidelinesonmethodvalidationtobeperformedinsupportofanalyticalmethodsforagrochemicalformulationsth21.SCOPETheseGuidelinesreferprimarilytoanalyticalmethodsforformulations,butmayalsobeappliedtoactiveingredientcontentoftechnicalmaterialandarepartiallyapplicabletophysicalmethods,(see3.2.2.).Theydonotapplytoresidueortraceanalysismethods.TherecommendationsintheseGuidelines,whilstintendedspecificallytoapplytostudiessubmittedtoPSDshouldsatisfythebroadrequirementsofmethodvalidationforformulationsaslaiddownintheDraftECUniformPrinciples.2.DEFINITIONS2.1ErrorsThetypeoferrorscoveredbyvalidationintheseGuidelinesare:2.1.1RandomerrorsTheseareerrors,usuallysmall,whichgiverisetoaspreadofresultsaroundtheaverageresults.Inotherwords,theydefinetherepeatabilityorreproducibilityoftheprocedures.2.1.2SystematicerrorsTheseareerrorswhichcauseabiasintheresultsobtained,suchthattheaveragevalueobservedisaboveorbelowthetruevalue(i.e.somefeatureofthemethodwhichleadstoinaccuracyintheresults).2.2PrecisionPrecisionisameasureofrandomerrors,andmaybeexpressedasrepeatabilityandreproducibility.ThesetermsaredefinedinISO5725-1986E:32.2.1Repeatabilityistheclosenessofagreementbetweenmutallyindependenttestresultsobtainedwiththesamemethodonidenticaltestmaterialinthesamelaboratorybythesameoperatorusingthesameequipmentwithinshortintervalsoftime.2.2.2Reproducibilityistheclosenessofagreementbetweentestresultsobtainedwiththesamemethodonidenticaltestmaterialindifferentlaboratorieswithdifferentoperatorsusingdifferentequipment.2.3AccuracyTheaccuracyofamethodisthedegreetowhichtheobservedresultscorrespondtothetruevalueoftheanalyteinthesample.2.4LinearityThelinearityofatestprocedureisitsability(withinagivenrange)toobtaintestresultsproportionaltotheconcentration(amount)ofanalyteinthesample.2.5SpecificityThespecificityofthemethodisadefinitionofthespeciesgivingrisetothesignalusedforquantitation.3.METHODVALIDATIONItisrecognisedthatmostAgrochemicalmanufacturerswillhaveinternalproceduresforanalyticalmethodvalidation.Itwillbethemanufacturer'sresponsibilitytoensurethattheirprocedurescomplywiththerequirementsoftheseGuidelines.Methodvalidationdatasubmittedshouldaddressthefollowingissues:-Linearityofresponsefortheanalyte(andinternalstandard,ifappropriate),inthemethod.-Anestimateoftheprecisionoftheprocedure.-Ademonstrationoftheaccuracyoftheprocedure.-Ademonstrationofnointerferencefromexcipients.-Adefinitionofthespeciesbeingdetermined.Althoughreproducibilityisimportantforamethodproposedforwidegeneraluse,itsassessmentisnotconsiderednecessaryaspartoftheseGuidelines.Itis4bestestimated(ifrequired)throughafullcollaborativestudybyCIPACorAOAC.3.1LinearityThelinearityofresponsetotheanalyteshouldbedemonstratedatleastovertherange(nominalanalyteconcentration±20%).Atleast3concentrationsshouldbemeasuredwithduplicatemeasurementsforeach.Thelinegeneratedshouldbesubmitted,togetherwithslope,interceptandcorrelationco-efficientdata..Themeasuredslopeshoulddemonstrateaclearcorrelationbetweenresponseandanalyteconcentration.Theresultsshouldnotshowasignificantdeviationfromlinearitywhichistakentomeanthatthecorrelationcoefficient(r)is0.99,overtherange(nominal±20%).Ifthisisnotthecase,thesubmittermustprovideanexplanationofhowmethodvalidityistobemaintained.Incaseswhereanon-linearresponseisdeliberatelyused,anexplanationmustalsobeprovided.3.2Precision3.2.1.ChemicalAnalysisFortheseGuidelines,asimpleassessmentofrepeatabilitywillbeacceptable.Aminimumoffivereplicatesampledeterminationsshouldbemadetogetherwithasimplestatisticalassessmentoftheresultsincludingthe%RSD.Ifconsideredappropriate,asuitabletestforoutliers(e.g.DixonsorGrubbsTest)maybeappliedtotheresults.However,itshouldbeclearlyindicatedifresultshavebeendiscardedandsomeattemptmadetoexplainwhytheoutliermayhaveoccurred.TheacceptabilityoftheresultsshouldbebasedonthemodifiedHorwitz1equation;RSDr2(1-0.5logC)x0.67whereC=concentrationoftheanalyteinthesampleasadecimalfraction.ThederivationandworkedexamplesoftheHorwitzequationaregiveninAppendixI.3.2.2.Physico-chemicalmeasurementsFormeasurementsofphysicalorphysico-chemicalproperties,nofurthervalidationisneededwhereofficialmethods(suchasCIPACorOECD)areused.Thisalsoappliestocloseadaptionsofsuchmethods.Whreanothermethodisused,itsrepeatability5mustbedetermined,butneednotcomplywiththeHorwitzEquation.3.3AccuracyTheaccuracyoftheprocedureshouldbeassessedbythepreparationandanalysisofatleastfoursamplesoflaboratory-prepared'synthetic'formulationcontainingknownweightsofanalyte.TheresultsmaybeassessedusingtheStudentst-statisticorotheracceptableapproaches,giveninAppendixII.3.4Non-analyteinterferenceThisiscoveredtosomeextentbytheassessmentofaccuracy,sinceanyinterferencefromexcipientswillconferasystematicerroronthemethod.However,ananalysisshouldbecarriedoutusinganexcipientblank,eithertodemonstratelackofinterferenceorquantifyanywhichisoccurring.Samplechromatogramsorotherresultsshouldbesubmitted.Wherespecificimpuritiesareknowntooccurinthetechnicalactiveingredient,itmustbedemonstratedthatthesedonotcontributemorethan3%tothetotalpeakar