FDA指南 混合均匀性接受标准

G:\5831dft.doc10/27/03GuidanceforIndustryPowderBlendsandFinishedDosageUnits—StratifiedIn-ProcessDosageUnitSamplingandAssessmentDRAFTGUIDANCEThisguidancedocumentisbeingdistributedforcommentpurposesonly.Commentsandsuggestionsregardingthisdraftdocumentshouldbesubmittedwithin60daysofpublicationintheFederalRegisterofthenoticeannouncingtheavailabilityofthedraftguidance.SubmitcommentstoDocketsManagementBranch(HFA-305),FoodandDrugAdministration,5630FishersLane,rm.1061,Rockville,MD20852.AllcommentsshouldbeidentifiedwiththedocketnumberlistedinthenoticeofavailabilitythatpublishesintheFederalRegister.ForquestionsregardingthisdraftdocumentcontactJonE.Clark,301-594-5613orMikeGavini,301-827-9053.U.S.DepartmentofHealthandHumanServicesFoodandDrugAdministrationCenterforDrugEvaluationandResearch(CDER)October2003PharmaceuticalCGMPsG:\5831dft.doc10/27/03GuidanceforIndustryPowderBlendsandFinishedDosageUnits—StratifiedIn-ProcessDosageUnitSamplingandAssessmentAdditionalcopiesareavailablefrom:OfficeofTrainingandCommunicationDivisionofDrugInformation,HFD-240CenterforDrugEvaluationandResearchFoodandDrugAdministration5600FishersLaneRockville,MD20857(Tel)301-827-4573http://www.fda.gov/cder/guidance/index.htmU.S.DepartmentofHealthandHumanServicesFoodandDrugAdministrationCenterforDrugEvaluationandResearch(CDER)OfficeofPharmaceuticalScience(OPS)OfficeofCompliance(OC)October2003PharmaceuticalCGMPsContainsNonbindingRecommendationsDraft—NotforImplementationG:\5831dft.doc10/27/03TABLEOFCONTENTSI.INTRODUCTION.................................................................................................................1II.BACKGROUND...................................................................................................................1III.SCOPE...................................................................................................................................2IV.CORRELATIONOFIN-PROCESSSTRATIFIEDSAMPLINGWITHPOWDERMIXANDFINISHEDPRODUCT..............................................................................................4A.AssessmentofPowderMixUniformity........................................................................................4B.CorrelationofPowderMixUniformitywithStratifiedIn-ProcessDosageUnitData............5C.CorrelationofStratifiedIn-ProcessSampleswiththeFinishedProduct.................................6V.EXHIBIT/VALIDATATIONBATCHPOWDERMIXHOMOGENEITY...................6VI.VERIFICATIONOFMANUFACTURINGCRITERIA.................................................7A.In-ProcessDosageUnitSamplingandAnalysis..........................................................................7B.CriteriatoMeettheReadilyPassClassification.........................................................................8C.CriteriatoMeettheMarginallyPassClassification...................................................................8D.SampleLocationsforRoutineManufacturing............................................................................9VII.ROUTINEMANUFACTURINGBATCHTESTINGMETHODS............................9A.StandardCriteriaMethod(SCM)................................................................................................91.Stage1Test....................................................................................................................................102.Stage2Test....................................................................................................................................10B.MarginalCriteriaMethod(MCM)............................................................................................10C.SwitchingtoStandardTestMethodfromMarginalTestMethod..........................................11VIII.REPORTINGTHEUSEOFSTRATIFIEDSAMPLING..........................................11A.ApplicationsNotYetApproved..................................................................................................11B.PostapprovalChange...................................................................................................................12GLOSSARY.................................................................................................................................13ATTACHMENT1:VERIFICATIONOFMANUFACTURINGCRITERIA....................14ATTACHMENT2:ROUTINEMANUFACTURINGBATCHTESTING..........................15ContainsNonbindingRecommendationsDraft—NotforImplementationG:\5831dft.doc10/27/031GuidanceforIndustry112PowderBlendsandFinishedDosageUnits—StratifiedIn-Process3DosageUnitSamplingandAssessment456Thisdraftguidance,whenfinalized,willrepresenttheFoodandDrugAdministration's(FDA's)current7thinkingonthistopic.Itdoesnotcreateorconferanyrightsfororonanypersonanddoesnotoperateto8bindFDAorthepublic.Youcanuseanalternativeapproachiftheapproachsatisfiestherequirementsof9theapplicablestatutesandregulations.Ifyouwanttodiscussanalternativeapproach,contacttheFDA10staffresponsibleforimplementingthisguidance.IfyoucannotidentifytheappropriateFDAstaff,call11theappropriatenumberlistedonthetitlepageofthisguidance.12.131415I.INTRODUCTION1617Thisguidanceisintendedtoassistmanufacturersofhumandrugproductsinmeetingthe18requirementsof21CFR211.110fordemonstratingtheadequacyofmixingtoensureuniformity19ofin-processpowderblendsandfinisheddosageunits.Thisguidancedescribestheprocedures20forassessingpowdermixadequacy,correlatingin-processdosageunittestresultswithpowder21mixtestresults,andestablishingtheinitialcriteriaforcontrolproceduresusedinroutine22manufacturing.2324FDA'sguidancedocuments,includingthisguidance,donotestablishlegallyenforceable25responsibilities.Instead,guidancesdescribetheAgency'scurrentthinkingonatopicandshould26beviewedonlyasrecommendations,unlessspecificregulatoryorstatutoryrequirementsare27cited.TheuseofthewordshouldinAgencyguidancesmeansthatsomethingissuggestedor28recommended,butnotrequired.293031II.BACKGROUND3233ThisguidanceistheresultofanAgencyefforttoachieveascience-basedpolicyandregulatory34enforcement.Expertsfromindustry,academia,andtheFDAdevelopedtheprinciples35underlyingthisguidanceafterextensivepublicdiscussion.Abriefhistoryoftheevolutionof36thisguidanceisprovidedinthefollowingparagraphs.371ThisguidancehasbeenpreparedbytheOfficeofPharmaceuticalScienceandtheOfficeofComplianceintheCenterforDrugEvaluationandResearch(CDER)attheFoodandDrugAdministrationincooperationwiththeProductQualityResearchInstitute(PQRI)(seefootnote3).ThisguidancedocumentrepresentstheAgency'scurrentthinkingonassessmentoftheuniformityofpowderblendsandfinisheddosageunitsintheabsenceofnewtechnologydevelopmentorimplementation.ContainsNonbindingRecommendationsDraft—NotforImplementationG:\5831dft.doc10/27/03238Inresponsetoindustryconcernsregardingregulationsfordemonstratingtheadequacyofin-39processpowdermixing,theFDApublishedadraftguidanceforindustryonblenduniformity40analysisinAugust1999.2Commentssubmittedtothedocketresultedintheformationofthe41BlendUniformityWorkingGroup(BUWG)bytheProductQualityResearchInstitute(PQRI).342ThePQRIBUWGconductedapublicmeeting,PQRIWorkshoponBlendUniformity,on43September7and8,2000.4445Usingtheconsensusreachedbyparticipantsinthisworkshop,theBUWGdevelopedadraft46recommendation,TheUseofStratifiedSamplingofBlendandDosageUnitstoDemonstrate47AdequacyofMixforPowderBlends.Thedraftrecommendationreceivedexaminationandpeer48reviewinmultiplescientificandpublicvenues.Inaddition,theAdvisoryCommitteefor49PharmaceuticalScience(ACPS)reviewedthedraftrecommendationandreceivedpublic50commentduringscheduledmeetingsofthecommittee.4Thedraftrecommendationwasrevised51toincorporatetheresultsofpeerreviewandpubliccommentandwaspresentedtoCDER's52CenterDirectorinfinalformonDecember30,2002.Therecommendationwassubsequently53publishedinthePDAJournalofPharmaceuticalScienceandTechnology.5Thisdraftguidance54reflectsCDER'sefforttoincorporatethedraftrecommendationintoregulatorypolicy.555657III.SCOPE5859Stratifiedsamplingistheprocessofsamplingdosageunitsatpredefinedintervalsandcollecting60representativesamplesfromspecificallytargetedlocationsinthecompression/fillingoperation61thathavethegreatestpotentialtoyieldextremehighsandlowsintestresults.Thesetestresults62areusedtomonitorthemanufacturingprocessoutputthatismostresponsibleforcausing63finishedproductvariability.Thetestresultscanbeusedtodevelopasinglecontrolprocedureto64ensureadequatepowdermixanduniformcontentinfinishedproducts.6566Themethodsdescribedinthisguidancearenotintendedtobetheonlymethodsformeeting67Agencyrequirementstodemonstratetheadequacyofpowdermix.Traditionalpowderblend68samplingandtesting,inconjunctionwithtestingforuniformityofcontentinthefinished69product,canbeusedtocomplywithcurrentgoodmanufacturingpracticerequirements702TheFDAwithdrewtheguidanceforindustryANDAs:BlendUniformityAnalysisonMay17,2002.3PQRIisacollaborativebodyinvolvingFDA'sCenterforDrugEvaluationandResearch(CDER),industry,andacademia.SinceitsinceptioninJanuary1996,themissionofPQRIhasbeentogeneratescientificinformationinsupportofregulatorypoliciesthroughresearch.AdditionalinformationaboutPQRIisavailableatwww.pqri.org.4ThePQRIBUWGrecommendationappearedonthepublicACPSagendaonNovember28,2001(introduction),May8,2002(distributionandcomment),andOctober22,2002(finalcomment).5GBoehm,JClark,JDietrick,LFoust,TGarcia,MGavini,LGelber,JGeoffry,JHoblitzell,PJimenez,GMergen,FMuzzio,JPlanchard,JPrescott,JTimmermens,andNTakiar,"TheUseofStratefiedSamplingofBlendandDosageUnitstoDemonstrateAdequacyofMixforPowderBlends,PDAJ.Pharm.SciTechnol,.57:59-74,2003.ContainsNonbindingRecommendationsDraft—NotforImplementationG:\5831dft.doc10/27/033(CGMPs).Useofat-,in-,oron-linemeasurementsystemscanalsobeappropriateandare71describedinotherguidancedocuments.67273Thisguidanceprovidesrecommendationsonhowto:7475•Conductpowderblendsamplingandanalyses.76•Establishinitialcriteriaforstratifiedsamplingofin-processdosageunits7andevaluation77oftestresults.78•Analyzethestratifiedsamplesandevaluatedata.79•Correlatethestratifiedsampledatawiththepowderblenddata.80•Assesspowdermixuniformity.81•Correlatethestratifiedsampledatawiththefinisheddosageunitdataandassess82uniformityofcontent.83•Testexhibitandvalidationbatchesforadequacyofpowdermix.84•Testandevaluateroutinemanufacturingbatches.85•Reporttheuseofstratifiedsamplingintheapplication.8687Themethodsdescribedinthisguidancecanbeusedtomonitoractiveingredienthomogeneityof88powderblendsandtoensureuniformcontentofthefinishedproductforsolidoraldrugproducts.89ThesemethodsareonlyonewaytosatisfytheCGMPandapplicationreviewrequirementsfor90in-processtestingtodemonstrateadequacyofpowdermixanduniformcontentofthefinished91product.Themethodassumesappropriatemonitoringofallmanufacturingstepsasrequiredby92theregulationsorapplicationcommitments.Thisguidancedoesnotdiscusstheassessmentof93thepotencyandotherattributesthatcanaffectthefinisheddosageunits,orthehomogeneityof94inactiveingredients.Formulationswithextremelylowdoseand/orhighpotencymaycallfor95morerigoroussamplingthanthatdescribedinthisguidancetoassesstheuniformityofpowder96blendsortheuniformityofcontentofthefinisheddosageunits.9798Whenusingthemethodsdescribedinthisguidance,certaindataortrendsmaybeobserved.We99recommendthatmanufacturersscientificallyevaluatethesetypesofresearchdatatodetermineif100theyaffectthequalityofaproductand,ifso,how.TheFDAdoesnotintendtoinspectresearch101datacollectedonanexistingproductforthepurposeofevaluatingthesuitabilityofproposed102methods.AnyFDAdecisiontoinspectresearchdatawouldbebasedonexceptionalsituations1036InAugust2003,theAgencyissuedthedraftguidanceforindustryPAT–AFrameworkforInnovativePharmaceuticalManufacturingandQualityAssurance.Oncefinalized,itwillrepresenttheAgency'sperspectiveonthisissue.7Thein-processdosageunitisacapsuleortabletasitisformedinthemanufacturingprocessbeforeitiscoatedorpackaged.ContainsNonbindingRecommendationsDraft—NotforImplementationG:\5831dft.doc10/27/034similartothoseoutlinedinCompliancePolicyGuideSec.130.300.8Thosedatausedtosupport104validationorregulatorysubmissionswillbesubjecttoinspectionintheusualmanner.105106107IV.CORRELATIONOFIN-PROCESSSTRATIFIEDSAMPLINGWITHPOWDER108MIXANDFINISHEDPRODUCT109110Ifyouplantofollowtheproceduresdescribedinthisguidancedocument,werecommendthat111youfirstcompletetheprocessdevelopmentproceduresdescribedinthissectionbeforeusingthe112methodsdescribedinsectionsV,VI,VII.Thesubsectionsbelowdescribehowtoassessthe113adequacyofpowdermix,uniformityofcontentofthein-processandfinisheddosageunits114throughcorrelationandassessmentofdatafromdevelopment,validationandmanufacturing115batches.Theseprocedurescanrevealdeficienciesintheblendingoperationthatmaynothave116beenpreviouslydetected.Werecommendthatmanufacturerscorrectdeficienciesinthe117blendingoperationbeforeimplementingtheroutinemanufacturingcontrolmethodsdescribedin118thisguidance.119120A.AssessmentofPowderMixUniformity121122Werecommendtheassessmentofpowdermixuniformityusingthefollowingprocedures:123124•Conductblendanalysisonbatchesbyextensivelysamplingthemixintheblenderand/or125intermediatebulkcontainers(IBCs).126•Identifyappropriateblendingtimeandspeedranges,deadspotsinblenders,andlocations127ofsegregationinIBCs.Determinesamplingerrors.128•Definetheeffectsofsamplesize(e.g.,1-10Xdosageunitrange)whiledevelopinga129techniquecapableofmeasuringthetrueuniformityoftheblend.Samplequantitieslarger130than3Xcanbeusedwithadequatescientificjustification.Appropriateblendsampling131techniquesandproceduresshouldbedevelopedforeachproductwithconsiderationto132variousdesignsofblendpowdersamplingandthephysicalandchemicalpropertiesof133theblendcomponents.134•Designblend-samplingplansandevaluatethemusingappropriatestatisticalanalyses.135•Quantitativelymeasureanyvariabilitythatispresentamongthesamples.Attributethe136samplevariabilitytoeitherlackofuniformityoftheblendorsamplingerror.Significant137within-locationvarianceintheblenddatacanbeanindicationofonefactorora138combinationoffactorssuchasinadequacyofblendmix,samplingerror9or1398FDA/ORACompliancePolicyGuide,Sec.130.300,FDAAccesstoResultsofQualityAssuranceProgramAuditsandInspections(CPG7151.02)9Ifblendsamplingerrorisdetected,moresophisticated,statisticalanalysesshouldbeappliedtoassessthesituation,suchastheuseofmethodsdescribedinJBerman,DEElinski,CRGonzales,JDHofer,PJJimenez,JAPlanchard,RJTlachac,PFVogel,“BlendUniformityAnalysis:ValidationandIn-ProcessTesting.”TechnicalReportNo.25,PDAJPharm.Sci.Technol.51(Suppl3i-iii),S1-99,1997.ContainsNonbindingRecommendationsDraft—NotforImplementationG:\5831dft.doc10/27/035agglomeration.10,11Significantbetween-locationvarianceintheblenddatacanindicate140thattheblendingoperationisinadequate.141142B.CorrelationofPowderMixUniformitywithStratifiedIn-ProcessDosage143UnitData144145Werecommendthefollowingstepsforcorrelation:146147•Conductperiodicsamplingandtestingofthein-processdosageunitsbysamplingthemat148definedintervalsandlocationsthroughoutthecompressionorfillingprocess.Usea149minimumof20appropriatelyspacedin-processdosageunitsamplingpoints.There150shouldbeatleast7samplestakenfromeachoftheselocationsforatotalminimumofat151least140samples.152•Take7samplesfromeachadditionallocationtofurtherassesseachsignificantevent,12153suchasfillingoremptyingofhoppersandIBCs,startandendofthecompressionor154fillingprocessandequipmentshutdown.Thismaybeaccomplishedbyusingprocess155developmentbatches,validationbatches,orbyusingroutinemanufacturingbatchesfor156approvedproducts.157•Significanteventsmayalsoincludeobservationsorchangesfromonebatchtoanother158(e.g.,batchscale-upandobservationsofundesirabletrendsinpreviousbatchdata).159•Prepareasummaryofthedataandanalysisusedtocorrelatethestratifiedsampling160locationswithsignificanteventsintheblendingprocess.Werecommendyousubmitthis161summarywiththeapplicationasdescribedinsectionVIIIofthisguidance.162•Comparethepowdermixuniformitywiththein-processdosage-unitdatadescribed163above.164•Investigateanydiscrepanciesobservedbetweenpowdermixanddosage-unitdataand165establishrootcauses.Atleastonetrouble-shootingguideisavailablethatmaybehelpful166withthistask.13Possiblecorrectionsmayrangefromgoingbacktoformulation167developmenttoimprovepowdercharacteristicstoprocessoptimization.Sampling16810OSSudah,PEArratia,D.Coffin-Beach,FJMuzzio,"MixingofCohesivePharmaceuticalFormulationsinTote(Bin)-Blenders,”DrugDev.Ind.Pharm,28(8):905-918,2002.11VSwaminathan,DOKildsig,“Polydispersepowdermixtures:effectofparticlesizeandshapeonmixturestability,”DrugDev.Ind.Pharm.,28(1):41-48,2002.12Asignificanteventisanyoperationduringthesoliddosageproductionprocessthatcanaffecttheintegrityofthein-processmaterials–seesectionIXGlossary.13JKPrescott,TJGarcia,"ASolidDosageandBlendContentUniformityTroubleshootingDiagram,"Pharm.Technol.,25(3):68-88,2001.ContainsNonbindingRecommendationsDraft—NotforImplementationG:\5831dft.doc10/27/036problemsmayalsobenegatedbyuseofalternatestate-of-the-artmethodsofinsitureal-169timesamplingandanalysis.170171C.CorrelationofStratifiedIn-ProcessSampleswiththeFinishedProduct172173Werecommendthefollowingsteps:174175•Conducttestingforuniformcontentofthefinishedproductusinganappropriate176procedureorasspecifiedintheAbbreviatedNewDrugApplication(ANDA)ortheNew177DrugApplication(NDA)forapprovedproducts.178•Comparetheresultsofstratifiedin-processdosageunitanalysiswithuniformcontentof179thefinisheddosageunitsfromthepreviousstep.Thisanalysisshouldbedonewithout180weightcorrection.14181•Prepareasummaryofthedataandanalysisusedtoconcludethatthestratifiedin-process182samplingprovidesassuranceofuniformcontentofthefinishedproduct.Werecommend183yousubmitthissummarywiththeapplicationasdescribedinsectionVIIIofthis184guidance.185186187V.EXHIBIT/VALIDATATIONBATCHPOWDERMIXHOMOGENEITY188189Thissectiondescribessamplingandtestingthepowdermixofexhibitandprocessvalidation190batchesusedtosupportimplementingthestratifiedsamplingmethoddescribedinthisguidance.191192193Werecommendthatduringthemanufactureofexhibitandprocessvalidationbatches,youassess194theuniformityofthepowderblend,thein-processdosageunits,andthefinishedproduct195independently.Werecommendyouusethefollowingstepstoidentifysamplinglocationsand196acceptancecriteriapriortothemanufactureoftheexhibitand/orvalidationbatches.151971981.Carefullyidentifyatleast10samplinglocationsintheblendertorepresentpotentialareas199ofpoorblending.Forexample,intumblingblenders(suchasV-blenders,doublecones,200ordrummixers),samplesshouldbeselectedfromatleasttwodepthsalongtheaxisof201theblender.Forconvectiveblenders(suchasaribbonblender),aspecialeffortshould202bemadetoimplementuniformvolumetricsamplingtoincludethecornersanddischarge203area(atleast20locationsarerecommendedtoadequatelyvalidateconvectiveblenders).2042052.Collectatleast3replicatesamplesfromeachlocation.Samplesshouldmeetthe206followingcriteria:20714Weightcorrectionisamathematicalcorrectiontoeliminatetheeffectofpotentiallyvariabletabletweightonmeasurementofmixadequacy—seeGlossary,SectionIX.15ThisisdescribedinSectionIVofthisguidance.ContainsNonbindingRecommendationsDraft—NotforImplementationG:\5831dft.doc10/27/037208•Assayonesampleperlocation(numberofsamples(n)≥10)209(n=20forribbonblender).210211•RSD(relativestandarddeviation)ofallindividualresults≤5.0percent.212213•Allindividualresultsarewithin10.0percent(absolute)ofthemeanoftheresults.214215Ifsamplesdonotmeetthesecriteria,werecommendthatyouinvestigatethefailureaccordingto216theflowchartinAttachment1.Wealsorecommendthatyounotproceedanyfurtherwith217implementationofthemethodsdescribedinthisguidanceuntilthecriteriaaremet.218219Samplingerrorsmayoccurinsomepowderblends,samplingdevices,andtechniquesthatmake220itimpracticaltoevaluateadequacyofmixusingonlytheblenddata.Insuchcases,we221recommendthatyouusein-processdosageunitdatainconjunctionwithblendsampledatato222evaluateblenduniformity.223224Somepowderblendsmaypresentunacceptablesafetyriskwhendirectlysamp