FDA关于PAI检查指南

1、13preapprovalinspections: the critical compliancepathto successmartind. hynesiiieli lilly and company,indianapolis,indiana,u.s.a.carmenmedinaprecisionconsultants,inc., coronado,california,u.s.a.1preapprovalinspections1.1introductiona preapprovalinspection(pai) is a visit by one or more food and drug

2、 inves-tigators to review the adequacy and accuracy of the informationprovidedina regulatorysubmissionthe fda compliance program guidance manual onpre-approvalinspections/investigations(program7346.832). the programwas a direct result of the generic drug scandal of the late 1980s. priorto thistime,

3、the food and drug administration(fda)relied on rmsto provideaccuratedata in supportof their submissions.priorto the implementationof the pai program,companiesessentially operated on an honor system withthe fda.thishonor system was e?ectively terminatedwith the generic drugscandal. in fact, in 1990 d

4、r. david kessler, then commissionerof the fda,was quoted assaying, whati learned most from the generic drug scandal is463copyright ? 2004 marcel dekker, inc.that in the end, the data this agency acts on has to be audited.thehonor sys-tem is out the window 1.as a result, the objectives of the pai wer

5、e to:ensure that facilitieslisted in the new drug applications(nda)havethe capabilitiesto ful ll the commitmentsto manufacture,process,control,package, and label a drug productfollowinggood manufac-turing practices (cgmps).ensureadequacyandaccuracyofanalyticalmethods(validatedmethods).ensure that th

6、e manufacturingprocess for clinicaltrial material,bio-availabilitystudy material,and stabilitystudies correlateswith theledprocess.ensure that scienticevidence supportsfull-scaleproductionproce-dures and controls.ensure that rmshave submittedfactual data.ensure protocolsare in place to validate the

7、manufacturingprocess.ensure equipmentis adequate and suitable for use (equipmentquali -cation).the paiprogramwas rstimplementedin the fda smid-atlanticregion. this region has the highest density of pharmaceuticalcompaniesinthe unitedstates. a 1990 publicationauthored by henry avalloneentitled mid-at

8、lanticregion pharmaceuticalinspectionprogram was the rstformal notice of the programprovidedto the pharmaceuticalindustry. notlong after the publicationof the avallonepaper, the fdaissued the rstcompliancemanual entitledthe fdacomplianceprogram guidance manualon pre-approval inspections/investigatio

9、ns(program7346.832). the manualoutlineda new review step that was being added to the new drug approvalprocess, in whichboththe centerfordrugevaluationand research(cder)and the districto?ceswould play pivotalroles. for the rsttime,the fda districto?ce was to become involved in the ndaapproval process

10、.the role of the districto?ce was to ensure compliancewith cgmps,aswellas audit the data submittedto the fdato ensure that it was truthful,ade-quate, and accurate.1.1.1historicaloverviewthe historicalroutes of fda pais can be traced back to the preamble of thecgmp published in the september 1978 fed

11、eral register and the draft guide-line on the preparationof investigationalnew drug (ind)products,whichwas issued in february1988.thepreamble to the cgmp issued in september1978 containedtwo statements that are highly relevant to the pai program.464hynes and medinacopyright ? 2004 marcel dekker, inc

12、.the relevant section reads as follows:thecommissionernds thatas statedin211.1 thesecgmpregulationsapplytothepreparationofany drugproductforadministrationto humans or animals, includingthose still in inves-tigationalstages. it is appropriatethat the process by which a drugproductis manufacturedin th

13、e developmentphase be well docu-mented and controlledin order to assure the reproducibilityof theproductfor furthertesting and subsequent commercialproduction.the commissioneris consideringproposingadditionalcgmpreg-ulations speci cally designed to cover drugs in research stages 2.the rstconcept mad

14、e explicitin this documentis that the fdahasjurisdictionover materialsthat are used in clinicaltrialsprior to approvaland market launch; thus some 12 years prior to the start of pais by the fdait had establishedan industrystandardthat complianceto cgmps,wasrequiredduring the manufactureof clinicaltr

15、ial materials.the preambleto the 1978 gmpalso containsa key second concept;thatis, that the expectationsrelatingto the manufactureof clinicaltrialmaterialsare di?erentfrom those of commercialmaterials.support for thisassertion comes from the statement that the fdawasconsideringproposingadditionalcgm

16、ps to cover drugs in the research stage. despite this sugges-tion, the fda has yet to issue cgmps speci callyfor clinicaltrial materials,rather;it has elected to set standards througha variety of di?erentmechan-isms, such as complianceprograms,inspectionguides, and podiumpolicypresentations.thereare

17、 also ichguidelinesthataddress myriadcom-pliance issues related to clinicaltrial activities.speci cally, the ich guide-lines provide directionon the followingtopics:for the investigatorinvestigatorsquali cations and agreementscommunicationwith irb/ieccompliancewith protocolrandomizationprocedures an

18、d unblindinginformedconsent of trial sutbjectssafety reportingprematureterminationor suspension of trialfor the sponsor (or companyresponsible for conductingthe trial)qualityassurance (qa) and qualitycontrol aspects of trialcontractresearch organization(cro) responsibilitiestrial designinvestigators

19、electionpreapprovalinspections465copyright ? 2004 marcel dekker, inc.allocationof duties and functionsnoti cationor submission to regulatoryauthority(or authorities)additionally,the roots of the pai programcan be traced to the 1988draft guideline on the preparationof indproducts.theseguidelines clea

20、rlystated that cgmps applied to drugs that were being made for clinicaltrialsinvolvinghuman subjects.the need for proper documentationduring the drug developmentpro-cess was stronglyemphasizedin these draftindguidelines.in addition,controlof components,productioncontrols,process controls,equipmentid

21、entication, packaging,and labeling were speci cally addressed.theideathat controls should increase throughoutthe course of the drug developmentprocess as additionalexperience was outlinedin these guidelines.the fdahad thus rmlyestablished its jurisdictionover clinicaltrialmaterials throughthe cgmps

22、preamble and the indguidelines prior to the1989 generic drug scandal.the rstdocumentissued by the fda outliningthe pai program wasthe mid-atlanticregion pharmaceuticalinspectionprogram. its goal wasto clarifyfda sexpectationsfor the compliancebranch of the fda as wellas the pharmaceuticalindustry.th

23、e programwas then formalizedinto twocompliancedocuments:fda compliance program guidance manual on pre-approval inspections/investigations(program7346.832), issued in octoberof 1990 and fda complianceprogramguidance manualon pre-approvalinspection of new animaldrug applications(nada)(program7368.001)

24、,issued in february of 19shiftin inspectiontrendsover timethe pai program has evolved over the past10 years.thisis evidenced by therevision of the fda scompliancemanuals as well asby inspectiontrends.since the rsteditionof the various compliancemanuals referencedabove, 7346.832 was revised o

25、nce, in august 1994. this revisionprovidedimprovedguidanceforall phases of the inspection,samplecollection,laboratoryevaluation,and assessment of ndings.additionally,this revisionoutlinedroles for cderand districto?cesin the inspectionprocess. thefda also issued the guide to the inspection of dosage

26、 form drug manufac-turers cgmp in october 1993 and the guide to inspection of pharmaceuticalquality control laboratories in july 1993. both of these guides in uencedtheevolution of the pai program into what companies experiencetoday.the food and drug administrationmodernizationact (fdama)of1997,whic

27、h was implementedin february of 1998, also altered the pai pro-cess.the act signicantly altered the balance of authoritywithinthe fda.at the time the inspectionprogram was implemented,fda eldo?ces were466hynes and medinacopyright ? 2004 marcel dekker, inc.given the primaryrole in cgmpand data integr

28、ityissues. the pre-eminentrole was shifted to the reviewing divisionby the1997 modernizationact. oneof the provisionsof the act indicatedthat no action by the ndareview divi-sion maybe delayed because of the unavailabilityof informationfrom orbecause of action by eldpersonnel.this indicatesthat the

29、ndareviewdivisionmay override the concerns of the districto?ce. additionally,it sug-gests that the inabilityof the districto?ce to performascheduled inspectionmay be an inadequatebasis for delaying approval.it should be noted that atthis writing,the pai programhas not been speci cally altered to re

30、ect thechanges includedin the modernizationact.additionalevidence that supportsthe change in the inspectionpro-gram over time comes from lookingat the number of inspectionsthat havebeen conducted in the years since the start of the program aswell asthe num-ber of approvals that have been withheldas

31、aresult of the actual inspection.the number of pais conductedin the years since the program s incep-tionis shown in fig. 1. as can be seen from fig. 1, the fdais conductingslightlyfewer inspections over time.additionally,the numberof inspectionsthat resultedin a delay inapprovalhas decreased over ti

32、me, as also can be seen in fig. 1.this change in the numberof approvalswithheldis in large measuredue to the fact that fda-regulatedindustrieshave invested a great deal oftime and energy in preparingfor these inspections.this level of improvedperformanceis evidenced by the fact that the fda srecomme

33、ndationstofigure1preapprovalinspections7-year summary.preapprovalinspections467copyright ? 2004 marcel dekker, inc.withholdapprovals have declined signicantly, from a high of 60% in1990 tolessthan 30% in 1996.in the late 1990s, the majorreasons that companieswere still failingfda pais were cgmpdevia

34、tions,disparitiesfrom commitmentsmade inthe application,discrepanciesin records, failure to report adverse ndings,and suspicion of fraud.thefollowingis a listof the majorcgmpdeciencies thathaveresulted in failed pais 3.standardoperationgprocedures(sops) incomplete,not current,ornot available to the

35、operators in the productionareabatch productionrecords incomplete,not recordedat time of opera-tion, or not speci cenough to documentsignicant process stepscleaningproceduresnotvalidatedor not includingall processingequipmentand transfer implements(scoops, etc.)failureto establishyields or acceptabl

36、elevels of rejects for both in-process and nishedproductfailureto conduct stabilitystudiesmanufacturingequipmentnot identiedand/orquali edinadequatetrainingof employees workingin aseptic operationsinadequateprocess change proceduresvalidationprotocolsthat lack acceptance criteriaincompleteinvestigat

37、ionsof laboratoryfailuresfailureto followunitedstates pharmacopeia(usp) proceduresforthe bacterialendotoxintest1.1.3differencebetweenpai and other typesof inspectionsthe fda conductsmany di?erenttypes of inspections;some are very tai-loredto thereasonfortheinspection.forexample,a for-causeor investi

38、gator-directedinspectionoccurswhen the agency has receivedspeci ccompliantsfrom the publicor the trade about a rmsproductorpractices. this type of inspectioncould be triggeredby a consumercom-plaintrelatedto an adverse drug experienceor misleadinginformationon the product spackaging. what makes this

39、 type of inspectionalong withmost others di?erentfroma paiis the fda slack of noti cationto therm. theinvestigatorarrivesunannouncedpreparedtoconductanassessment of the rmand its product(s)and practices. whereas the paiis a scheduledinspectionarrangedby an fdapaimanager and a com-pany representative

40、.468hynes and medinacopyright ? 2004 marcel dekker, inc.the pai will focus on the rmscommercializatione?orts related to theproductfor whichapprovalis being sought. other types of inspectionsaremore unpredictablein nature and tend to be more general (i.e., the focus maynot necessarily be on one produ

41、ct).1.2trendsovertime:the fda modernizationact, teambiologics,newmedicaldeviceregulations,and qualitysystemsinspection(qsi)withthe passingof the fdama,the agency s inspectionlatitudewasincreasedand its inspectionalresponsibilityover activepharmaceuticalingredient(api)manufacturerswas reiterated.the

42、agency did increase itsinspectionactivityover rmsthat manufactureapis and excipients.addi-tionally,fda sreview of foreignapi manufacturersincreased slightly. theonus, however, remains with the sponsor company to ensure that the manu-facturerof their api is cgmp compliantand the materialsuitable for

43、use.1.2.1teambiologicssimilarto the fdama,team biologicswas born of vice presidentgoresreorganizationof government(rego)initiativein the mid-1990s. teambiologicswas a rationalresponse by the fda so?ceof regulatorya?airs(ora)to controlvariousbiologicproducts(vaccines,bloodproducts,invitrodiagnostics)

44、more e?ectively.whatresultedfromredirectingthemanner in which biologicsinspectionsoccurredas well as changing whichfdao?cewas responsiblefor eldinspectionswas an inspectionprocessthat more closely mimickedthe eldinspectionsfor cder-regulateddrugs.this means a far greater assessment of cgmp complianc

45、eto biologicsman-ufacturingand testing, as well as an increase in administrative,regulatoryand judicialactions against them. the biologicsindustryhas never been thesame since team biologics has imposed its inspectionalstrategies and techni-ques on an industrythat was unaccustomedto comprehensivecomp

46、liance-focused inspectionsby the fda.thenumber of warning letters and consentdecrees levied against a large numberof biologicproducts manufacturerssince the inceptionof team biologicsapproach is unprecedented.1.2.2newmedicaldeviceregulationsthe new medical device regulationswere codi ed in1996,signi

47、cantly alter-ing the way in whichmedical device manufacturerswere inspectedby thefda.the qualitysystems inspectiontechnique(qsit)was launchedaspart of the new strategy for inspectingthe device industry. this inspectiontechniqueallowed fda to move closer to global harmonizationguidelinespreapprovalin

48、spections469copyright ? 2004 marcel dekker, inc.for regulatoryauditing.this new inspectionapproachmandated that inves-tigators assessseven subsystems withina device rm.the seven substancesare asfollows:1.executivemanagement2.design controls3.materialcontrols4.records, documents,and change controls5.

49、equipmentand facilitycontrols6.productionand process controls7.correctiveand preventiveactionsthis focus helped eldinvestigatorsconductmore e?ective, e?cient,and comprehensiveinspectionsof medical device manufacturersby evaluat-ing key elements of a rmsoverallqualitysystem and compliancestatus.this

50、new and improvedinspectionapproachalso led to an unprecedentednumber of administrative,regulator, and judicialactions against the medicaldevice industry.1.2.3qualitysystemsinspectioninterestingly,along with the changing trends in biologicsand medical deviceinspections,the approachemployedby cderduri

51、nga pharmaceuticalinspectionhasalso changed.thereis atrend toward assessingoverall qualitysystems and major compliancecategories withina pharmaceuticalmanufac-turing and quality controllaboratoryfrom the top down rather than from thebottom up. more emphasis is being placed on management responsibili

52、tyaswell as identifyingweaknesses withina particularsystem, along with thegenesis of the weaknesses,versus identifyingareas of nonconformance.2preparingfor a preapprovalinspectionfrom acorporateperspectivethe preparationsfor pais must be incorporatedinto the overall process ofdrug development.prepar

53、atione?orts must start with the initialformationof the drug developmentteam prior to the start of phase i clinicaltrials. oneof the rsttasks the team needs to completeis to author the drug develop-ment plan, which integratesthe work that is to be performedby the variousfunctions,such as medical,toxi

54、cology,metabolism,and process and pro-duct development.the workto be conductedby the process and productdevelopmentgroups must followgood developmentpracticesthat willen-sure a successful pais. thus, the preparationwork for a pai needs to begin470hynes and medinacopyright ? 2004 marcel dekker, inc.w

55、hen the developmentteam rstconvenes to begin planningthe relevantdevelopmentactivities.prior to asking developmentscientists to work on teams and to preparefor a pai, scientistsshould be knowledgeableabout the fda pai strategy,how to prepare for it, and the consequences of afailed inspection.thiscan

56、 beaccomplishedby sending them to trainingsessions or relevant industry meet-ings. additionally,it is always helpful for scientists who are new to the indus-try to talk with coworkers who have survived previous fda inspections.the chances of passing a paiare greatlyenhancedif the relevantdevelopment

57、is carried out in conformancewith well-dened quality princi-ples and some applicable cgmps.thesequality principlesshould cover suchtopic as batch disposition,stability,process validation,training,deviations,managementnoti cation,documentationchange, and historyof develop-ment. these principleshave b

58、een describedin detailin chapter2 and byhynes 3.if the developmentwork is planned with the goal of passing the pai andthe work complies with cgmpquality principles,there should be only mini-mal preparationsneeded for the fda in the weeks and months just prior tothe inspectionitself. a number of di?e

59、rentmethodologieshave been devel-oped to help with the short-termpreparatione?orts. justice and co-workersat eli lillyand companyhave describeda 10-stepprocess to help guidethese preparatione?orts 4.thesesteps can be found in table 1.2.1the use of internalauditsmany companieshave utilizedmock fdains

60、pectionsto help with short-term preparatione?orts3.these mock inspectionshelp not only identifyissues priorto the inspectionbut also providehands-onfdainspectionexperienceto personnel.2.2froman fda investigatorsviewpointbeforean fdainvestigatorcan make any kindof recommendationforapproval,he or shem