行业指南：药品检测结果OOS调查指南

1、DoCGuidanceforIndustryInvestigatingOutofSpecification(OOS)TestResultsforPharmaceuticalProduction行业指南：药品检验结果ooS勺调查DRAFTGUIDANCE指南草案U.S.DepartmentofHealthandHumanServicesFoodandDrugAdministrationCenterforDrugEvaluationandResearch(CDER)September1998CP#D0CDOCTABLEOFCONTENTS录INTRODUCTION序1BACKGROUND.背景1I

2、DENTIFYINGANDASSESSINGOOSTESTRESULTSOOS检验结果的判断和评估.2ResponsibilityoftheAnalyst检验员的责TOC o 1-5 h z任2ResponsibilitiesoftheSupervisor主管的责任3INVESTIGATINGOOSTESTRESULTSOOS检验结果的调查一般调查原那实验室的调查阶调查结果的解GeneralInvestigationalPrinciples.么5LaboratoryPhaseofanInvestigation.段6CONCLUDINGTHEINVESTIGATION.调查结10A.Interp

3、retationofInvestigationResults.10B.Reporting.11ThisguidancehasbeenpreparedbytheOfficeofpliance/DivisionofManufacturingandProductQuality,CenterforDrugEvaluationandResearch(CDER)attheFoodandDrugAdministration.ThisguidancedocumentrepresentstheAgencyscurrentthinkingonevaluatingOOStestresults.Itdoesnotcr

4、eateorconferanyrightsfororonanypersonanddoesnotoperatetobindFDAorthepublic.Analternativeapproachmaybeusedifsuchapproachsatisfiestherequirementsoftheapplicablestatute,regulations,orboth.本指南由FDA勺CDE的达标办公室/制造、产品、质量分部起草，本指南说明了机构关于评估00检验结果的现行的想法。它不会创造或赠与任何人任何权力，也不会约束FDA或公众。如果其他可选择的相接近勺指南能满足适用勺法令和法规勺要求，也可

5、以使用。GUIDANCEF0RINDUST1RYInvestigating0utofSpecification(00S)TestResultsforPharmaceuticalProduction行业指南：药品检验结果00的调查INTR0DUCTI0N序言ThisguidanceforindustryprovidestheAgencyscurrentthinkingonhowtoevaluatesuspect,oroutofspecification(00S),testresults.Forpurposesofthisdocument,theterm00Sresultsincludesalls

6、uspectresultsthatfalloutsidethespecificationsoracceptancecriteriaestablishedinnewdrugapplications,officialpendia,orbythemanufacturer.本行业指南反响了FDA关于如何评估怀疑的或00检验结果的现行想法，本指南的目的，00S吉果包括超出了新药申请材料、法定标准、生产商建立的可承受标准或规格的所有可疑的结果。Thisguidanceappliestolaboratorytestingduringthemanufactureofactivepharmaceuticalin

7、gredients,excipients,andotherponentsandthetestingoffinishedproductstotheextentthatcurrentgoodmanufacturingpractices(CGMP)regulationsapply(21CFRparts210and211).Specifically,theguidancediscusseshowtoinvestigatesuspect,orOOStestresults,includingtheresponsibilitiesoflaboratorypersonnel,thelaboratoryphas

8、eoftheinvestigation,additionaltestingthatmaybenecessary,whentoexpandtheinvestigationoutsidethelaboratory,andthefinalevaluationofalltestresults.本指南适用于API、赋形剂和其它组分生产的实验室检验和CGM法规应用的成品检验。特别的，指南讨论了如何调查可疑的或00检验结果，包括了实验室人员的责任、实验室阶段调查、必须的额外试验，何时进展实验室X围外的调查和对所有检验结果的最终评估。BACKGR0UND背景FDAconsiderstheintegrityof

9、laboratorytestinganddocumentationrecordstobeoffundamentalimportanceduringdrugmanufacturing.Laboratorytesting,whichisrequiredbytheCGMPregulations(211.165),isnecessarytoconfirmthatponents,containersandclosures,inprocessmaterials,andfinishedproductsconformtospecifications,includingstability.TestingSpec

10、ificationsmustbescientificallysoundandappropriate(21CFR211.160(b),andtestproceduresmustbevalidatedastotheiraccuracy,reliability,andsuitabilityunderactualconditionsofuse(21CFR211.194(a)(2).ForproductsthatarethesubjectsofNDAs,ANDAs,orINDs,specificationsarecontainedintheapplication.Specificationsfornon

11、-applicationproductsmaybefoundinofficialpendia,orestablishedbythemanufacturer.FDA认为在药品生产中实验室检验和文件记录的完整性是根本重要的。CGM要求的实验室检验，必须确定成分、容器和封口材料、生产用辅料、成品符合标准要求，包括稳定性。检验标准必须是科学正确和适当的1CFR211.160(b)，检验方法必须验证现行使用条件下的正确性、线性和适应性(21CFR211.194(a)，如果是用于NDAs,ANDAs,或INDs申请的产品，申请材料中应包括标准。如果是非申请产品的质量标准，可以在法定标准中找到或由企业自已建

12、立。Althoughthesubjectofthisdocumentis00Sresults,muchoftheguidancemaybeusefulforexaminingresultsthatareoutoftrend.alsosupportsanalyticalandprocessvalidationefforts.GeneralCGMPregulationscoveringlaboratoryoperationscanbefoundinpart211,subpartsI(LaboratoryControls)andJ(RecordsandReports).Theseregulation

13、sprovidefortheestablishmentofscientificallysoundandappropriatespecifications,standards,andtestproceduresthataredesignedtoensurethatponentsandcontainersofdrugproductsconformtotheestablishedstandards.Section211.165(f)oftheCGMPregulationsspecifiesthatproductsthatfailtomeetestablishedstandardsandotherre

14、levantqualitycontrolcriteriawillberejected.尽管本指南的目的是00结果，其中许多局部对于调查超岀趋势以外的结果也是有用的，对分析和工艺验证也有用。有关实验室操作一般性CGM法规可以在211局部，分目I(实验室控制)和分目J(报告和记录)找到。这些法规规定了科学正确和适当的用于保证制剂的成分和容器符合建立标准的规格、标准和检验方法的建立。CGM的211.165(f)章节规定不符合既定标准和其它相关质量控制标准的产品不得放行。IDENTIFYINGANDASSESSING00STESTRESULTS00S检验结果的判断和评估FDAregulationsre

15、quirethataninvestigationbeconductedwheneveran00Stestresultisobtained.Thepurposeoftheinvestigationistodeterminethecauseofthe00S.Evenifabatchisrejectedbasedonan00Sresult,theinvestigationisnecessarytodetermineiftheresultisassociatedwithotherbatchesofthesamedrugproductorotherproducts.Batchrejectiondoesn

16、otnegatetheneedtoperformtheinvestigation.Theregulationsrequirethatawrittenrecordoftheinvestigationbemadeincludingtheconclusionsoftheinvestigationandfollow-up(211.192).FDA法规要求当00S佥验结果岀现时应该进展调查，调查的目的是确定引起00S勺原因。即使因00结果判断了不合格批，仍必须进展调查以确定该结果是否影响到同种产品其它批号或其它产品。法规要求对调查包括调查结论和随后采取的措施进展记录(211.192)。Tobemeani

17、ngful,theinvestigationshouldbethorough,timely,unbiased,well-documented,andscientificallydefensible.Thefirstphaseofsuchaninvestigationshouldincludeaninitialassessmentoftheaccuracyofthelaboratorysdata,beforetestsolutionsarediscarded,wheneverpossible.Thisway,hypothesesregardinglaboratoryerrororinstrume

18、ntmalfunctionsmaybetestedusingthesametestsolutions.Ifthisinitialassessmentindicatesthatnoerrorsweremadeintheanalyticalprocessusedtoarriveatthedata,apletefailureinvestigationshouldfollow.更有意义的，调查必须是完全的，与时的，不带有任何偏见的，记录是完整的和经得起科学推敲的。调查的最初阶段应该在试验溶液丢弃前，对实验室数据正确性进展最初评估，这样，假定认为是实验室错误或仪器故障，可以用原溶液测定。如果最初的评估评

19、估显示在得到该数据的分析过程中没有发生错误，必须立即开展一个完全的不合风格查。A.ResponsibilityoftheAnalyst检验员的责任Thefirstresponsibilityforachievingaccuratelaboratorytestingresultslieswiththeanalystwhoisperformingthetest.TheanalystshouldbeawareofpotentialproblemsthatcouldoccurduringthetestingprocessandshouldwatchforproblemsthatcouldcreateO

20、OSresults.从事测试的检验员的首要责任是取得正确实验室检验结果。检验员应该意识到在实验过程中可能发生的潜在的问题和应该注意可能产生00结果的问题。InaccordancewiththeCGMPregulations(211.160(b)(4),theanalystshouldensurethatonlythoseinstrumentsmeetingestablishedspecificationsareusedandthatallinstrumentsareproperlycalibrated.Certainanalyticalmethodshavesystemsuitabilityr

21、equirements,andsystemsnotmeetingsuchrequirementsshouldnotbeused.Forexample,inchromatographicsystems,referencestandardsolutionsmaybeinjectedatintervalsthroughoutchromatographicrunstomeasuredrift,noise,andrepeatability.Ifreferencestandardresponsesindicatethatthesystemisnotfunctioningproperly,allofthed

22、atacollectedduringthesuspecttimeperiodshouldbeproperlyidentifiedandshouldnotbeused.Thecauseofthemalfunctionshouldbeidentifiedandcorrectedbeforeadecisionismadewhethertouseanydatapriortothesuspectperiod.依据CGMP(211.160(b)(4)，检验员应该保证只有符合既定标准的仪器才能使用和所有的仪器都经过的校正。某些分析方法有系统适应性要求，不符合要求的系统不能使用。例如：在色谱系统中，在进展色谱

23、检测期间内间隔一段时间进样对照品溶液去测定漂移、噪声和重复性。如果对照品响应值显示该系统功能不正常，在可疑的时间内收集的所有数据应该被适当标识并不能使用。在决定是否使用可疑期间之前的数据前，应鉴别故障的原因并予以纠正。Beforediscardingtestpreparationsorstandardpreparations,analystsshouldcheckthedataforpliancewithspecifications.Whenunexpectedresultsareobtainedandnoobviousexplanationexists,testpreparationssho

24、uldberetainedandtheanalystshouldinformthesupervisor.Anassessmentoftheaccuracyoftheresultsshouldbestartedimmediately.在丢弃样品制备液和标准制备液之前，检验员应该核查数据对标准的符合性。当获得意想不到的结果且没有明显的理由时，应该保存样品制备液且检验员应该通知主管。应该立即开始评估检验结果的正确性。Iferrorsareobvious,suchasthespillingofasamplesolutionortheinpletetransferofasampleposite,thea

25、nalystshouldimmediatelydocumentwhathappened.Analystsshouldnotknowinglycontinueananalysistheyexpecttoinvalidateatalatertimeforanassignablecause(i.e.,analysesshouldnotbepletedforthesolepurposeofseeingwhatresultscanbeobtainedwhenobviouserrorsareknown).Thesesameresponsibilitiesextendtoanalystsatcontract

26、testinglaboratories.如果错误是明显的，如：样品溶液有洒出或样品成分的未完全转移，检验员应该立即记录所发生的情况。检验员不应该有意的继续这无效的分析也就是，当明显的错误发生了，不应该带着会得出什么结果的目的去完成分析。这些一样的责任适用于合同实验室的检验员。B.ResponsibilitiesoftheSupervisor主管的责任OnceanOOSresulthasbeenidentified,thesupervisorsassessmentshouldbeobjectiveandtimely.Thereshouldbenopreconceivedassumptionsas

27、tothecauseoftheOOSresult.Datashouldbeassessedpromptlytoascertainiftheresultsmaybeattributedtolaboratoryerror,orwhethertheresultscouldindicateproblemsinthemanufacturingprocess.Animmediateassessmentcouldincludere-examinationoftheactualsolutions,testunits,andglasswareusedintheoriginalmeasurementsandpre

28、parations,whichwouldallowmorecredibilitytobegiventolaboratoryerrortheories.一旦00结果被确定，主管应该客观的和与时的进展评估。不应该对00结果的原因进展预想的假定。应该迅速评估数据以确定结果是否属于实验室错误，或该结果是否显示是生产过程的问题。立即的评估应该包括重新检测原溶液、测试的单位、玻璃器具和溶液制备过程，这样可以进一步证实对00S结果属于实验室错误的假设。Thefollowingstepsshouldbetakenaspartofthesupervisorsassessment:下面是作为主管应评估的局部步骤：

29、Discussthetestmethodwiththeanalyst;confirmanalystknowledgeofandperformanceofthecorrectprocedure.与检验员讨论检测方法；确认检验员知道并执行了正确的程序。Examinetherawdataobtainedintheanalysis,includingchromatogramsandspectra,andidentifyanomalousorsuspectinformation.检查分析的原始数据，包括色谱和光谱，并识别出反常或可疑的信息。Confirmtheperformanceoftheinstru

30、ments.确认仪器性能Determinethatappropriatereferencestandards,solvents,reagents,andothersolutionswereusedandthattheymeetqualitycontrolspecifications.确定使用了适宜的参照标准品、溶媒、试剂和其它溶液，并且它们符合质量控制标准。Evaluatetheperformanceofthetestingmethodtoensurethatitisperformingaccordingtothestandardexpectedbasedonmethodvalidationd

31、ata.评估检验方法的执行情况，以保证是按照标准执行并有方法验证数据。Documentandpreserveevidenceofthisassessment.记录并保存评估的证据。TheassignmentofacauseforOOSresultswillbegreatlyfacilitatediftheretainedsamplepreparationsareexaminedpromptly.Hypothesesregardingwhatmighthavehappened(e.g.dilutionerror,instrumentmalfunction)canbetested.Examinat

32、ionoftheretainedsolutionscanbeperformedaspartofthelaboratoryinvestigation.如果与时检测了所保存的样品制备液，会极大的促进00S吉果原因确实定。可以对所发生的事情的假设如：稀释错误、仪器故障进展检测。对被留的溶液的检验可以作为实验室调查的一局部。Examples:例如!Solutionscanbere-injectedaspartofaninvestigationwhereatransientequipmentmalfunctionissuspected.Thiscouldoccurifbubbleswereintrodu

33、cedduringaninjectiononachromatographicsystem,whichothertestsindicatedwasperformingproperly.Suchtheoriesaredifficulttoprove.However,areinjectioncanprovidestrongevidencethattheproblemshouldbeattributedtotheinstrument,ratherthanthesampleoritspreparation.如果怀疑设备瞬间的故障，那么调查可以重新进样该溶液。如果在色谱系统一次进样期间引入了气泡，而其它的

34、试验进展正常，这是可能发生的。这样的推测很困难去证明。但是，重新进样能有力的证明问题与仪器有关，而不是与样品或样品的准备有关。!Forreleaseratetestingofcertainspecializeddosageforms,wherepossible,examinationofthedosageunittestedmightdeterminewhetheritwasdamagedinawaythataffecteditsperformance.SuchdamagewouldprovideevidencetoinvalidatetheOOStestresult,andaretestwo

35、uldbeindicated.对于某一特定制剂的放行检验，如果可能，检查一下被测的这份制剂可以确定是否在什么方面被损坏而影响了其性能。这样的损坏能提供证据使00检验结果无效，这样就可以进展重新检验。!Furtherextractionofadosageunitcanbeperformedtodeterminewhetheritwasfullyextractedduringtheoriginalanalysis.Inpleteextractioncouldinvalidatethetestresultsandshouldleadtoquestionsregardingvalidationofth

36、etestmethod.对一份制剂做进一步提取以确定在初次检验期间是否被充分的提取了。不完全的提取能使检验结果无效，将导致检验方法验证的问题。Itisimportantthateachstepintheinvestigationbefullydocumented.Thesupervisorshouldascertainnotonlythereliabilityoftheindividualvalueobtained,butalsothesignificancetheseOOSresultsrepresentintheoverallqualityassuranceprogram.在调查的每一步都

37、应做充分的记录，这是很重要的。主管不仅应该确定获得的个别值的可靠性，也应该确定在整个质量保证程序中说明的00结果的重要性。Supervisorsshouldbeespeciallyalerttodevelopingtrends.主管应该特别警惕开展趋势。Laboratoryerrorshouldberelativelyrare.Frequenterrorssuggestaproblemthatmightbeduetoinadequatetrainingofanalysts,poorlymaintainedorimproperlycalibratedequipment,orcarelesswor

38、k.Wheneverlaboratoryerrorisidentified,thefirmshoulddeterminethesourceofthaterrorandtakecorrectiveactiontoensurethatitdoesnotoccuragain.ToensurefullpliancewiththeCGMPregulations,themanufactureralsoshouldmaintainadequatedocumentationofthecorrectiveaction.Insummary,whenclearevidenceoflaboratoryerrorexi

39、sts,laboratorytestingresultsshouldbeinvalidated.Whenevidenceoflaboratoryerrorremainsunclear,afailureinvestigationshouldbeconductedtodeterminewhatcausedtheunexpectedresults.Itshouldnotbeassumedthatfailingtestresultsareattributabletoanalyticalerrorwithoutperforminganddocumentinganinvestigation.Boththe

40、initiallaboratoryassessmentandthefollowingfailureinvestigationshouldbedocumentedfully.实验室错误应该是极少发生的。频繁的错误暗示一个问题，那就是检验员培训不充分，设备维护保养不善或没有得到正确校正，或工作粗心。如果实验室错误被确定了，企业应该确定错误的来源并采取纠正措施去保证错误不再发生。为保证完全符合CGMP生产商应该保持充分的纠正措施的记录。总之，当实验室错误的证据已经存在，实验室检验结果应该作废。当实验室错误的证据仍然不清楚，应该进展不合格结果的调查以确定引起意外结果的原因。在没有进展调查与调查文件确定

41、之前不应该假定失败的检验结果属于检验错误。最初的实验室评估和下面的不合格结果的调查应该被充分的记录。INVESTIGATING00STESTRESULTS00S吉果调查Whentheinitialassessmentdoesnotdeterminethatlaboratoryerrorcausedthe00Sresultandtestingresultsappeartobeaccurate,afull-scalefailureinvestigationusingapredefinedprocedureshouldbeconducted.Theobjectiveofsuchaninvestiga

42、tionshouldbetoidentifythesourceofthe00Sresult.Varyingtestresultscouldindicateproblemsinthemanufacturingprocess,orresultfromsamplingproblems.Suchinvestigationspresentachallengebothtoemployeesandtomanagementandshouldbegiventhehighestpriority.如果最初的评估不能确定是实验室错误造成了00结果且实验结果是正确的，应按照预先确定的程序进展全方位的不合风格查。这样调查

43、的目的是确定00结果的来源，变化的试验结果可能显示是生产工艺的问题或取样问题导致的结果。这样的调查是对员工和管理者的挑战应给予极高度的重视。Theinvestigationshouldbeconductedbythequalitycontrolunitandshouldinvolveallotherdepartmentsthatcouldbeimplicated,includingmanufacturing,processdevelopment,maintenance,andengineering.0therpotentialproblemsshouldbeidentifiedandinves

44、tigated.调查应该由质量控制部门和所有其它相关的部门完成，包括生产部门，工艺研发部门，维护保养和工程部门。其它的潜在问题也应该被确定和调查。Therecordsanddocumentationofthemanufacturingprocessshouldbefullyinvestigatedtodeterminethepossiblecauseofthe00Sresults.应该充分调查生产过程的记录和文件以判断引起00结果的可能原因。A.GeneralInvestigationalPrinciples一般调查原那么Afailureinvestigationshouldconsistof

45、atimely,thorough,andwell-documentedreview.一个不合格结果的调查在于与时，彻底，和完善的记录审核。Thewrittenrecordshouldreflectthatthefollowinggeneralstepshavebeentaken.记录应该反映通常采取的以下步骤：Thereasonfortheinvestigationhasbeenclearlyidentified.调查的原因被清楚确实定。Themanufacturingprocesssequencesthatmayhavecausedtheproblemshouldbesummarized.对

46、可能引起问题的生产工艺流程进展了总结。Resultsofthedocumentationreviewshouldbeprovidedwiththeassignmentofactualorprobablecause.文件审核的结果提供了实际的或可能的原因。Areviewshouldbemadetodetermineiftheproblemhasoccurredpreviously.审核并判断是否以前发生过这类问题。Correctiveactionstakenshouldbedescribed.应该描述采取的纠正措施Thegeneralreviewshouldincludealistofother

47、batchesandproductspossiblyaffectedandanyrequiredcorrectiveactionstakenincludinganymentsandsignaturesofappropriateproductionandqualitycontrolpersonnelregardinganymaterialthatmayhavebeenreprocessedafteradditionaltesting.一般性审核应该包括可能受影响的其它批和产品的列表，任何必须的纠正行为包括对复检后进展返工的物料的适宜的生产和质量控制人员的评论和签名。B.LaboratoryPha

48、seofanInvestigation实验室阶段的调查Anumberofpracticesareusedduringthelaboratoryphaseofaninvestigation.Theseinclude:(1)retestingaportionoftheoriginalsample,(2)testingaspecimenfromthecollectionofanewsamplefromthebatch,(3)resamplingtestingdata,and(4)usingoutliertesting.在实验室阶段的一系列调查包括：1最初样品一局部的再检验，2从该批中重新取样样品的检

49、验，3重新取样的检验数据，4逸出值的检验1.Retesting重新检验Partoftheinvestigationmayinvolveretestingofaportionoftheoriginalsample.Thesampleusedfortheretestingshouldbetakenfromthesamehomogeneousmaterialthatwasoriginallycollectedfromthelot,tested,andyieldedtheOOSresults.Foraliquid,itmaybefromtheoriginalunitliquidproductorpos

50、iteoftheliquidproduct;forasoliditmaybeanadditionalweighingfromthesamesamplepositethathadbeenpreparedbytheanalyst.调查局部工作包括最初检验样品的再检验。用于再检验的样品应该是最初收集检验的、岀现00结果的样品均质物料的一局部。如果是液体，可以是液体成品的原始单位或液体成品的混合物。如果是固体，可以是检验员制备的一样混合物的额外的称量。Situationswhereretestingisindicatedincludeinvestigatingtestinginstrumentmalf

51、unctionsortoidentifyapossiblesamplehandlingintegrityproblem,forexample,asuspecteddilutionerror.Generally,retestingisneitherspecifiednorprohibitedbyapprovedapplicationsorbythependia.Decisionstoretestshouldbebasedontheobjectivesofthetestingandsoundscientificjudgment.Retestingshouldbeperformedbyananaly

52、stotherthantheonewhoperformedtheoriginaltest.再检验的情况标明包括检验仪器故障的调查或判断样品处理完整性。例如：一个可疑的溶解错误。一般地，在批准的申请中或药典中既没有限定也没有禁止，应该依据客观的检验和正确科学的判断决定再检验。再检验应该是另一检验员完成而不是最初检验人员。TheCGMPregulationsrequiretheestablishmentofspecifications,standards,samplingplans,testprocedures,andotherlaboratorycontrolmechanisms(211.160

53、).Theestablishmentofsuchcontrolmechanismsforexaminationofadditionalspecimens，formercialorregulatorypliancetestingmustbeinaccordancewithpredeterminedguidelinesorsamplingstrategies(USP23,GeneralNoticesandRequirements,p.9).CGM法规要求建立规格、标准、取样计划、检验程序和其它的实验室控制机制(211.160)。符加样品的检验、商业或调整的依从性试验的控制机制的建立必须依据预先确定

54、指南或取样策略。Somefirmshaveusedastrategyofrepeatedtestinguntilapassingresultisobtained(testingintopliance),thendisregardingtheOOSresultswithoutscientificjustification.TestingintoplianceisobjectionableundertheCGMPs.ThenumberofreteststobeperformedonasampleshouldbespecifiedinadvancebythefirmintheSOP.许多企业使用重复

55、的检验直到获得满意的结果的策略，然后没有科学的理由就无视了00结果。重复的检验直到获得满意的结果，CGM是不允许的。样品的再检验次数应该在企业的SO中被预先限定。Thenumbermayvarydependinguponthevariabilityoftheparticulartestmethodemployed,butshouldbebasedonscientificallysound,supportableprinciples.Thenumbershouldnotbeadjusteddependingontheresultsobtained.Thefirmspredeterminedtes

56、tingproceduresshouldcontainapointatwhichthetestingendsandtheproductisevaluated.If,atthispoint,theresultsareunsatisfactory,thebatchissuspectandmustberejectedorheldpendingfurtherinvestigation(211.165(f).检验次数可以根据使用的检验方法的可变性决定，但必须依据科学正确和可支持的原那么。检验次数不能依据结果进展判断。企业预先确定的检验程序应该包括何时检验完毕和进展产品评估。如果，就这一点，结果是不满意的

57、，这批产品就是可疑的，必须批准不合格或进一步调查后再作决定(211.165(f).。Inthecaseofaclearlyidentifiedlaboratoryerror,theretestresultswouldsubstitutefortheoriginaltestresults.Theoriginalresultsshouldberetained,however,andanexplanationrecorded.Thisrecordshouldbeinitialedanddatedbytheinvolvedpersonsandincludeadiscussionoftheerroran

58、dsupervisoryments.在明确确定了实验室错误的情况下，再检验结果将取代最初检验结果，应该保存最初结果和一个解释记录，这个记录应该有相关人员人员的签名、注上日期，并应包括对错误的讨论、主管的注释。Ifnolaboratoryorstatisticalerrorsareidentifiedinthefirsttest,thereisnoscientificbasisforinvalidatinginitial00Sresultsinfavorofpassingretestresults.Alltestresults,bothpassingandsuspect,shouldberepo

59、rtedandconsideredinbatchreleasedecisions.如果不能确定是首次的试验是实验室或统计错误，就没有为了让产品通过复检而废除最初00结果的科学根底。所有检验结果，不论是通过的和可疑的，都应该报告并供批放行结论中考虑。Resampling再取样DOCWhileretestingreferstoanalysisoftheoriginalsample,resamplinginvolvesanalyzingaspecimenfromthecollectionofanewsamplefromthebatch.Theestablishmentofcontrolmechani

60、smsforexaminationofadditionalspecimensformercialorregulatorypliancetestingshouldbeinaccordancewithpredeterminedproceduresandsamplingstrategies(211.165(c).涉与最初样品分析的再检验，再取样包括从该批产品抽取新的样品。应该按照预先制定的程序和取样策略来建立为符合商业或法规的试验的目的而对额外样品进展检验的控制机制。Insomecases,whenalldatahavebeenexamined,itmaybeconcludedthattheorig