FDAQC實(shí)驗(yàn)室檢查指南

-.z.GUIDETOINSPECTIONSOFPHARMACEUTICALQUALITYCONTROLLABORATORIESNote:ThisdocumentisreferencematerialforinvestigatorsandotherFDApersonnel.ThedocumentdoesnotbindFDA,anddoesnoconferanyrights,privileges,benefits,orimmunitiesfororonanyperson(s).1.INTRODUCTIONThepharmaceuticalqualitycontrollaboratoryservesoneofthemostimportantfunctionsinpharmaceuticalproductionandcontrol.AsignificantportionoftheCGMPregulations(21CFR211)pertaintothequalitycontrollaboratoryandproducttesting.Similarconceptsapplytobulkdrugs.Thisinspectionguidesupplementsotherinspectionalinformationcontainedinotheragencyinspectionalguidancedocuments.Fore*ample,plianceProgram7346.832requiringpre-approvalNDA/ANDAinspectionscontainsgeneralinstructionstoconductproductspecificNDA/ANDAinspectionauditstomeasurepliancewiththeapplicationsandCGMPrequirements.Thisincludespharmaceuticallaboratoriesusedforin-processandfinishedproducttesting.2.OBJECTIVEThespecificobjectivewillbespelledoutpriortotheinspection.Thelaboratoryinspectionmaybelimitedtospecificissues,ortheinspectionmayenpassaprehensiveevaluationofthelaboratory'spliancewithCGMP's.Asaminimum,eachpharmaceuticalqualitycontrollaboratoryshouldreceiveaprehensiveGMPevaluationeachtwoyearsaspartofthestatutoryinspectionobligation.Ingeneraltheseinspectionsmayinclude--thespecificmethodologywhichwillbeusedtotestanewproduct--apleteassessmentoflaboratory'sconformancewithGMP's--aspecificaspectoflaboratoryoperations3.INSPECTIONPREPARATIONFDAInspectionGuidesarebasedontheteaminspectionapproachandourinspectionofalaboratoryisconsistentwiththisconcept.Aspartofourefforttoachieveuniformityandconsistencyinlaboratoryinspections,wee*pectthatple*,highlytechnicalandspecializedtestingequipment,proceduresanddatamanipulations,aswellasscientificlaboratoryoperationswillbeevaluatedbyane*periencedlaboratoryanalystwithspecializedknowledgeinsuchmatters.Districtmanagementmakesthefinaldecisionregardingtheassignmentofpersonneltoinspections.Nevertheless,wee*pectinvestigators,analystsandotherstoworkasteamsandtoadvisemanagementwhenadditionale*pertiseisrequiredtopleteameaningfulinspection.Teammembersparticipatinginapre-approvalinspectionmustreadandbefamiliarwithplianceProgram7346.832,Pre-ApprovalInspections/Investigations.RelevantsectionsoftheNDAorANDAshouldbereviewedpriortotheinspection;butiftheapplicationisnotavailablefromanyothersource,thisreviewwillhavetobeconductedusingthepany'scopyoftheapplication.Teammembersshouldmeet,ifpossible,priortotheinspectiontodiscusstheapproachtotheinspection,todefinetherolesoftheteammembers,andtoestablishgoalsforpletionoftheassignment.Responsibilitiesfordevelopmentofallreportsshouldalsobeestablishedpriortotheinspection.ThisincludesthepreparationoftheFDA483.TheCenterforDrugEvaluationandResearch(CDER)mayhaveissueddeficiencyletterslistingproblemsthatthesponsormustcorrectpriortotheapprovalofNDA/ANDA'sandsupplements.Theinspectionteamise*pectedtoreviewsuchlettersonfileatthedistrictoffice,andtheyaree*pectedtoasktheplantforaccesstosuchletters.Theteamshouldevaluatetherepliestotheseletterstoassurethatthedataareaccurateandauthentic.pletetheinspectioneventhoughtherehasbeennoresponsetotheselettersorwhentheresponseisjudgedinadequate.4.INSPECTIONAPPROACHA.GeneralInadditiontothegeneralapproachutilizedinadrugCGMPinspection,theinspectionofalaboratoryrequirestheuseofobservationsofthelaboratoryinoperationandoftherawlaboratorydatatoevaluatepliancewithCGMP'sandtospecificallycarryoutthemitmentsinanapplicationorDMF.Whenconductingaprehensiveinspectionofalaboratory,allaspectsofthelaboratoryoperationswillbeevaluated.Laboratoryrecordsandlogsrepresentavitalsourceofinformationthatallowsapleteoverviewofthetechnicalabilityofthestaffandofoverallqualitycontrolprocedures.SOPsshouldbepleteandadequateandtheoperationsofthelaboratoriesshouldconformtothewrittenprocedures.Specificationsandanalyticalproceduresshouldbesuitableand,asapplicable,inconformancewithapplicationmitmentsandpendialrequirements.Evaluaterawlaboratorydata,laboratoryproceduresandmethods,laboratoryequipment,includingmaintenanceandcalibration,andmethodsvalidationdatatodeterminetheoverallqualityofthelaboratoryoperationandtheabilitytoplywithCGMPregulations.E*aminechromatogramsandspectraforevidenceofimpurities,poortechnique,orlackofinstrumentcalibration.susesystemsthatprovidefortheinvestigationoflaboratorytestfailures.Thesearegenerallyrecordedinsometypeoflog.Asktoseeresultsofanalysesforlotsofproductthathavefailedtomeetspecificationsandreviewtheanalysisoflotsthathavebeenretested,rejected,orreworked.Evaluatethedecisiontoreleaselotsofproductwhenthelaboratoryresultsindicatethatthelotfailedtomeetspecificationsanddeterminewhoreleasedthem.B.Pre-ApprovalDocumentsrelatingtotheformulationoftheproduct,synthesisofthebulkdrugsubstance,productspecifications,analysisoftheproduct,andothersaree*aminedduringthereviewprocessinheadquarters.However,thesereviewsandevaluationsdependonaccurateandauthenticdatathattrulyrepresentstheproduct.Pre-approvalinspectionsaredesignedtodetermineifthedatasubmittedinanapplicationareauthenticandaccurateandiftheprocedureslistedintheapplicationwereactuallyusedtoproducethedatacontainedintheapplication.Additionally,theyaredesignedtoconfirmthatplants(includingthequalitycontrollaboratory)areinpliancewithCGMPregulations.Theanalyticalsectionsofdrugapplicationsusuallycontainonlytestresultsandthemethodsusedtoobtainthem.Sponsorsarenotrequiredtofileallthetestdatabecausesuchactionwouldrequirevoluminoussubmissionsandwouldoftenresultinfilingredundantinformation.Sponsorsmaydeliberatelyorunintentionallyselectandreportdatashowingthatadrugissafeandeffectiveanddeservestobeapproved.Theinspectionteammustdecideifthereisvalidandscientificjustificationforthefailuretoreportdatawhichdemonstratestheproductfailedtomeetitspredeterminedspecifications.Coordinationbetweenheadquartersandthefieldisessentialforapletereviewoftheapplicationandtheplant.E*periencedinvestigatorsandanalystsmaycontactthereviewchemist(withappropriatesupervisoryconcurrence)whenquestionsconcerningspecificationsandstandardsarise.Inspectionsshouldparetheresultsofanalysessubmittedwithresultsofanalysisofotherbatchesthatmayhavebeenproduced.Evaluatethemethodsandnoteanye*ceptionstotheproceduresorequipmentactuallyusedfromthoselistedintheapplicationandconfirmthatitisthesamemethodlistedintheapplication.Theanalystise*pectedtoevaluaterawlaboratorydatafortestsperformedonthetestbatches(biobatchesandclinicalbatches)andtoparethisrawdatatothedatafiledintheapplication.5.FAILURE(OUT-OF-SPECIFICATION)LABORATORYRESULTSEvaluatethepany'ssystemtoinvestigatelaboratorytestfailures.Theseinvestigationsrepresentakeyissueindecidingwhetheraproductmaybereleasedorrejectedandformthebasisforretesting,andresampling.Inarecentcourtdecisionthejudgeusedtheterm"out-of-specification"(OOS)laboratoryresultratherthantheterm"productfailure"whichismoremontoFDAinvestigatorsandanalysts.HeruledthatanOOSresultidentifiedasalaboratoryerrorbyafailureinvestigationoranoutliertest.Thecourtprovidede*plicitlimitationsontheuseofoutliertestsandthesearediscussedinalatersegmentofthisdocument.,oroverebyretesting.Thecourtruledontheuseofretestingwhichiscoveredinalatersegmentofthisdocument.isnotaproductfailure.OOSresultsfallintothreecategories:--laboratoryerror--non-processrelatedoroperatorerror--processrelatedormanufacturingprocesserrorA.LABORATORYERRORSLaboratoryerrorsoccurwhenanalystsmakemistakesinfollowingthemethodofanalysis,useincorrectstandards,and/orsimplymiscalculatethedata.LaboratoryerrorsmustbedeterminedthroughafailureinvestigationtoidentifythecauseoftheOOS.OncethenatureoftheOOSresulthasbeenidentifieditcanbeclassifiedintooneofthethreecategoriesabove.Theinquirymayvarywiththeobjectunderinvestigation.B.LABORATORYINVESTIGATIONSThee*actcauseofanalysterrorormistakecanbedifficulttodeterminespecificallyanditisunrealistictoe*pectthatanalysterrorwillalwaysbedeterminedanddocumented.Nevertheless,alaboratoryinvestigationconsistsofmorethanaretest.Theinabilitytoidentifyanerror'scausewithconfidenceaffectsretestingprocedures,nottheinvestigationinquiryrequiredfortheinitialOOSresult.Thefirm'sanalystshouldfollowawrittenprocedure,checkingoffeachstepasitispletedduringtheanalyticalprocedure.Wee*pectlaboratorytestdatatoberecordeddirectlyinnotebooks;useofscrappaperandloosepapermustbeavoided.Thesemonsensemeasuresenhancetheaccuracyandintegrityofdata.ReviewandevaluatethelaboratorySOPforproductfailureinvestigations.SpecificproceduresmustbefollowedwhensingleandmultipleOOSresultsareinvestigated.ForthesingleOOSresulttheinvestigationshouldincludethefollowingstepsandtheseinquiriesmustbeconductedbeforethereisaretestofthesample:otheanalystconductingthetestshouldreporttheOOSresulttothesupervisorotheanalystandthesupervisorshouldconductaninformallaboratoryinvestigationwhichaddressesthefollowingareas:1.discussthetestingprocedure2.discussthecalculation3.e*aminetheinstruments4.reviewthenotebookscontainingtheOOSresultAnalternativemeanstoinvalidateaninitialOOSresult,providedthefailureinvestigationprovesinconclusive,isthe"outlier"test.However,specificrestrictionsmustbeplacedontheuseofthistest.1.Firmscannotfrequentlyrejectresultsonthisbasis.2.TheUSPstandardsgovernitsuseinspecificcasesonly.3.Thetestcannotbeusedforchemicaltestingresults.AninitialcontentuniformitytestwasOOSfollowedbyapassingretest.TheinitialOOSresultwasclaimedtheresultofanalysterrorbasedonastatisticalevaluationofthedata.Thecourtruledthattheuseofanoutliertestisinappropriateinthiscase..4.Itisneverappropriatetoutilizeoutliertestsforastatisticallybasedtest,i.e.,contentuniformityanddissolution.DetermineifthefirmusesanoutliertestandevaluatetheSOP.DeterminethatafullscaleinquiryhasbeenmadeformultipleOOSresults.Thisinquiryinvolvesqualitycontrolandqualityassurancepersonnelinadditiontolaboratoryworkerstoidentifye*actprocessornonprocessrelatederrors.Whenthelaboratoryinvestigationisinconclusive(reasonfortheerrorisnotidentified)thefirm:1.Cannotconduct2retestsandbasereleaseonaverageofthreetests2.Cannotuseoutliertestinchemicaltests3.Cannotuseare-sampletoassumeasamplingorpreparationerror4.Canconductaretestofdifferenttabletsfromthesamesamplewhenaretestisconsideredappropriate(seecriteriaelsewhere)C.FORMALINVESTIGATIONSFormalinvestigationse*tendingbeyondthelaboratorymustfollowanoutlinewithparticularattentiontocorrectiveaction.Thepanymust:1.Statethereasonfortheinvestigation2.Providesummationoftheprocesssequencesthatmayhavecausedtheproblem3.Outlinecorrectiveactionsnecessarytosavethebatchandpreventsimilarrecurrence4.Listotherbatchesandproductspossiblyaffected,theresultsofinvestigationofthesebatchesandproducts,andanycorrectiveaction.Specifically:oe*amineotherbatchesofproductmadebytheerrantemployeeormachineoe*amineotherproductsproducedbytheerrantprocessoroperation5.PreservethementsandsignaturesofallproductionandqualitycontrolpersonnelwhoconductedtheinvestigationandapprovedanyreprocessedmaterialafteradditionaltestingD.INVESTIGATIONDOCUMENTATIONAnalyst'smistakes,suchasundetectedcalculationerrors,shouldbespecifiedwithparticularityandsupportedbyevidence.Investigationsalongwithconclusionsreachedmustbepreservedwithwrittendocumentationthatenumerateseachstepoftheinvestigation.Theevaluation,conclusionandcorrectiveaction,ifany,shouldbepreservedinaninvestigationorfailurereportandplacedintoacentralfile.E.INVESTIGATIONTIMEFRAMESAllfailureinvestigationsshouldbeperformedwithin20businessdaysoftheproblem'soccurrenceandrecordedandwrittenintoafailureorinvestigationreport.6.PRODUCTFAILURESAnOOSlaboratoryresultcanbeovere(invalidated)whenlaboratoryerrorhasbeendocumented.However,non-processandprocessrelatederrorsresultingfromoperatorsmakingmistakes,equipment(otherthanlaboratoryequipment)malfunctions,oramanufacturingprocessthatisfundamentallydeficient,suchasanimpropermi*ingtime,representproductfailures.E*aminetheresultsofinvestigationsusingtheguidanceinsection5aboveandevaluatethedecisiontorelease,retest,orreworkproducts.7.RETESTINGEvaluatethepany'sretestingSOPforpliancewithscientificallysoundandappropriateprocedures.Averyimportantrulinginonerecentcourtdecisionsetsforthaproceduretogoverntheretestingprogram.Thisdistrictcourtrulingprovidesane*cellentguidetouseinevaluatingsomeaspectsofapharmaceuticallaboratory,butshouldnotbeconsideredaslaw,regulationorbindinglegalprecedent.Thecourtruledthatafirmshouldhaveapredeterminedtestingprocedureanditshouldconsiderapointatwhichtestingendsandtheproductisevaluated.Ifresultsarenotsatisfactory,theproductisrejected.Additionally,thepanyshouldconsiderallretestresultsintheconte*toftheoverallrecordoftheproduct.Thisincludesthehistoryoftheproduct.Thecourtorderedarecallofonebatchofproductonthebasisofaninitialcontentuniformityfailureandnobasistoinvalidatethetestresultandonahistoryofcontentuniformityproblemswiththeproduct.,typeoftestperformed,andin-processtestresults.Failingassayresultscannotbedisregardedsimplyonthebasisofacceptablecontentuniformityresults.Thenumberofretestsperformedbeforeafirmconcludesthatanune*plainedOOSresultisinvalidorthataproductisunacceptableisamatterofscientificjudgment.ThegoalofretestingistoisolateOOSresultsbutretestingcannotcontinueadinfinitum.Inthecaseofnonprocessandprocess-relatederrors,retestingissuspect.Becausetheinitialtestsaregenuine,inthesecircumstances,additionaltestingalonecannotcontributetoproductquality.Thecourtacknowledgedthatsomeretestingmayprecedeafindingofnonprocessorprocess-basederrors.Oncethisdeterminationismade,however,additionalretestingforpurposesoftestingaproductintoplianceisnotacceptable.Fore*ample,inthecaseofcontentuniformitytestingdesignedtodetectvariabilityintheblendortablets,failingandnon-failingresultsarenotinherentlyinconsistentandpassingresultsonlimitedretestingdonotruleoutthepossibilitythatthebatchisnotuniform.Aspartoftheinvestigationfirmsshouldconsidertherecordofpreviousbatches,sincesimilarorrelatedfailuresondifferentbatcheswouldbeacauseofconcern.RetestingfollowinganOOSresultisruledappropriateonlyafterthefailureinvestigationisunderwayandthefailureinvestigationdeterminesinpartwhetherretestingisappropriate.Itisappropriatewhenanalysterrorisdocumentedorthereviewofanalyst'sworkis"inconclusive",butitisnotappropriateforknownandundisputednon-processorprocessrelatederrors.Thecourtruledthatretesting:omustbedoneonthesame,notadifferentsampleomaybedoneonasecondaliquotfromthesameportionofthesamplethatwasthesourceofthefirstaliquotomaybedoneonaportionofthesamelargersamplepreviouslycollectedforlaboratorypurposes8.RESAMPLINGFirmscannotrelyonresampling.ThecourtorderedtherecallofonebatchofproductafterhavingconcludedthatasuccessfulresampleresultalonecannotinvalidateaninitialOOSresult.toreleaseaproductthathasfailedtestingandretestingunlessthefailureinvestigationdisclosesevidencethattheoriginalsampleisnotrepresentativeorwasimproperlyprepared.Evaluateeachresamplingactivityforpliancewiththisguidance.9.AVERAGINGRESULTSOFANALYSISAveragingcanbearationalandvalidapproachwhentheobjectunderconsiderationistotalproductassay,butasageneralrulethispracticeshouldbeavoided.Thecourtruledthatthefirmmustrecallabatchthatwasreleasedforcontentuniformityonthebasisofaveragedtestresults.becauseaverageshidethevariabilityamongindividualtestresults.ThisphenomenonisparticularlytroublingiftestinggeneratesbothOOSandpassingindividualresultswhichwhenaveragedarewithinspecification.Here,relyingontheaveragefigurewithoute*aminingande*plainingtheindividualOOSresultsishighlymisleadingandunacceptable.Contentuniformityanddissolutionresultsnevershouldbeaveragedtoobtainapassingvalue.InthecaseofmicrobiologicalturbidimetricandplateassaysanaverageispreferredbytheUSP.Inthiscase,itisgoodpracticetoincludeOOSresultsintheaverageunlessanoutliertest(microbiologicalassays)suggeststheOOSisananomaly.10.BLENDSAMPLINGANDTESTINGThelaboratoryservesavitalfunctioninblendtestingwhichisnecessarytoincreasethelikelihoodofdetectinginferiorbatches.Blenduniformitytestingcannotbewaivedinfavoroftotalrelianceonfinishedproducttestingbecausefinishedproducttestingislimited.Onecourthasruledthatsamplesizeinfluencesultimateblendtestresultsandthatthesamplesizeshouldresemblethedosagesize.Anyotherpracticewouldblurdifferencesinportionsoftheblendanddefeattheobjectofthetest.Ifasamplelargerthantheunitmustbetakeninitially,aliquotswhichresemblethedosagesizeshouldbecarefullyremovedforthetest,retests,andreservesamples.Obviously,theinitiallargersampleshouldnotbesubjectedtoanyadditionalmi*ingormanipulationpriortoremovingtestaliquotsasthismayobscurenon-homogeneity.Multipleindividualblenduniformitysamplestakenfromdifferentareascannotbeposited.Howeverwhenvariationtestingisnottheobjectofassaytesting,positingispermitted.Iffirmssampleproductfromsitesotherthantheblender,theymustdemonstratethroughvalidationthattheirsamplingtechniqueisrepresentativeofallportionsandconcentrationsoftheblend.Thismeansthatthesamplesmustberepresentativeofthosesitesthatmightbeproblems;e.g.weakorhotspotsintheblend.11.MICROBIOLOGICALThereviewofmicrobiologicaldataonapplicabledosageformsisbestperformedbythemicrobiologist(analyst).Datathatshouldbereviewedincludepreservativeeffectivenesstesting,bioburdendata,andproductspecificmicrobiologicaltestingandmethods.Reviewbioburden(beforefiltrationand/orsterilization)frombothanendoto*inandsterilityperspective.Fordrugsubstancelabsevaluatemethodsvalidationandrawdataforsterility,endoto*intesting,environmentalmonitoring,andfilterandfiltrationvalidation.Also,evaluatethemethodsusedtotestandestablishbioburdens.RefertotheMicrobiologicalInspectionGuideforadditionalinformationconcerningtheinspectionofmicrobiologicallaboratories.12.SAMPLINGSampleswillbecollectedonpre-approvalinspections.FollowthesamplingguidelinesinCP7346.832,PartIII,pages5and6.13.LABORATORYRECORDSANDDOCUMENTATIONReviewpersonalanalyticalnotebookskeptbytheanalystsinthelaboratoryandparethemwiththeworksheetsandgenerallabnotebooksandrecords.Bepreparedtoe*amineallrecordsandworksheetsforaccuracyandauthenticityandtoverifythatrawdataareretainedtosupporttheconclusionsfoundinlaboratoryresults.Reviewlaboratorylogsforthesequenceofanalysisversusthesequenceofmanufacturingdates.Testdatesshouldcorrespondtothedateswhenthesampleshouldhavebeeninthelaboratory.Ifthereisaputerdatabase,determinetheprotocolsformakingchangestothedata.Thereshouldbeanaudittrailforchangestodata.Wee*pectrawlaboratorydatatobemaintainedinbound,(notlooseorscrapsheetsofpaper),booksoronanalyticalsheetsforwhichthereisaccountability,suchasprenumberedsheets.Formostofthosemanufacturerswhichhadduplicatesetsofrecordsor"rawdata",non-numberedloosesheetsofpaperwereemployed.Somepaniesusediscsortapesasrawdataandforthestorageofdata.Suchsystemshavealsobeenacceptedprovidedtheyhavebeendefined(withrawdataidentified)andvalidated.Carefullye*amineandevaluatelaboratorylogs,worksheetsandotherrecordscontainingtherawdatasuchasweighings,dilutions,theconditionofinstruments,andcalculations.Notewhetherrawdataaremissing,ifrecordshavebeenrewritten,orifcorrectionfluidhasbeenusedtoconcealerrors.Resultsshouldnotbechangedwithoute*planation.Crossreferencethedatathathasbeencorrectedtoauthenticateit.Productscannotbe"testedintopliance"byarbitrarilylabelingout-of-specificationlabresultsas"laboratoryerrors"withoutaninvestigationresultinginscientificallyvalidcriteria.Testresultsshouldnothavebeentranscribedwithoutretentionoftheoriginalrecords,norshouldtestresultsberecordedselectively.Fore*ample,investigationshaveuncoveredtheuseofloosesheetsofpaperwithsubsequentselectivetranscriptionsofgooddatatoanalystworksheetsand/orworkbooks.Absorbancevaluesandcalculationshaveevenbeenfoundondeskcalendars.Cutchartswithinjectionsmissing,deletionoffilesindirectdataentrysystems,indirectdataentrywithoutverification,andchangestoputerizedprogramstooverrideprogramfeaturesshouldbecarefullye*amined.Thesepracticesraisequestionsabouttheoverallqualityofdata.Thefirmshouldhaveawrittene*planationwheninjections,particularlyfromaseriesaremissingfromtheofficialwork-sheetsorfromfilesandareincludedamongtherawdata.Multipleinjectionsrecordedshouldbeinconsecutivefileswithconsecutiveinjectiontimesrecorded.E*pecttoseewrittenjustificationforthedeletionofallfiles.Determinetheadequacyofthefirm'sprocedurestoensurethatallvalidlaboratorydataareconsideredbythefirmintheirdeterminationofacceptabilityofponents,in-process,finishedproduct,andretainedstabilitysamples.Laboratorylogsanddocumentswhencrossreferencedmayshowthatdatahasbeendiscardedbypanyofficialswhodecidedtoreleasetheproductwithoutasatisfactorye*planationoftheresultsshowingtheproductfailstomeetthespecifications.Evaluatethejustificationfordisregardingtestresultsthatshowtheproductfailedtomeetspecifications.14.LABORATORYSTANDARDSOLUTIONSAscertainthatsuitablestandardsarebeingused(i.e.in-date,storedproperly).Checkforthereuseofstocksolutionswithoutassuringtheirstability.Stocksolutionsarefrequentlystoredinthelaboratoryrefrigerator.E*aminethelaboratoryrefrigeratorsforthesesolutionsandwhenfoundcheckforappropriateidentification.Reviewrecordsofstandardsolutionpreparationtoassurepleteandaccuratedocumentation.Itishighlyunlikelythatafirmcan"accuratelyandconsistentlyweigh"tothesamemicrogram.Thereforedatashowingthislevelofstandardizationorpatternissuspectandshouldbecarefullyinvestigated.15.METHODSVALIDATIONInformationregardingthevalidationofmethodsshouldbecarefullyevaluatedforpleteness,accuracyandreliability.Inparticular,ifapendialmethode*ists,butthefirmchoosestouseanalternatemethodinstead,theymustparethetwoanddemonstratethatthein-housemethodisequivalentorsuperiortotheofficialprocedure.Forpendialmethodsfirmsmustdemonstratethatthemethodworksundertheactualconditionsofuse.Methodscanbevalidatedinanumberofways.MethodsappearingintheUSPareconsideredvalidatedandtheyareconsideredvalidatedifpartofanapprovedANDA.Alsoapanycanconductavalidationstudyontheirmethod.Systemsuitabilitydataaloneisinsufficientforanddoesnotconstitutemethodvalidation.Inthereviewofmethodvalidationdata,itise*pectedthatdataforrepetitivetestingbeconsistentandthatthevaryingconcentrationsoftestsolutionsprovidelinearresults.ManyassayandimpuritytestsarenowHPLC,anditise*pectedthattheprecisionoftheseassaysbeequalorlessthantheRSD'sforsystemsuitabilitytesting.TheanalyticalperformanceparameterslistedintheUSP**II,<1225>,undertheheadingofValidationofpendialMethods,canbeusedasaguidefordeterminingtheanalyticalparameters(e.g.,accuracy,precis