【文档说明】FDA药物流行病学_真实世界研究-课件2.ppt，共(100)页，2.046 MB，由小橙橙上传

转载请保留链接：https://www.ichengzhen.cn/view-233866.html

以下为本文档部分文字说明：

October2009Hangzhou1Real-lifeStudiesofPharmaceuticalAgents陳建煒,MD,ScD,FISPEHarvardSchoolofPublicHealthi3DrugSafetyOctober17-18,2009October200

9Hangzhou2October2009Hangzhou3Someterminologyaboutsafety⚫TherapeuticandNon-therapeuticuseofmedicalproducts

⚫Medicationerrors(adversedrugevents)andbiologiceffectsattherapeuticdose(adversedrugreactions)⚫SafetysignalOctober2009Hangzhou4Safe

tyofmedicalproducts⚫Patientsafety⚫Asaresultofhumanerror/systemfailure⚫Drug/Vaccine/Devicesafety⚫Toxicityoc

curredwhentheproductwasused‘appropriately’⚫Grayarea⚫Inappropriateuse(notonlyoff-labeluse)October2009Hangzhou5ByDr.RobertNelson,

inHartzema,Tilson,ChanOctober2009Hangzhou6MethodsandobjectivesInterventionalObservationalstudies(clinicalstudiestrials)Benefi

tEfficacy(Comparative)PhaseII/IIIEffectivenessRiskPragmatictrialReal-lifeuseofLarge&SimpletrialmedicalproductsOctober2009Hangzhou7Applicationsofph

armacoepidemiology⚫Safetyassessment⚫Relativeandabsoluteriskestimates⚫Riskfactorsforadverseoutcomes⚫Usagepatterns⚫Safetysignaldetection⚫Comp

arativeeffectiveness⚫Drugutilization⚫(Developmentofmedicalproducts)October2009Hangzhou8Threeexamplesfortodayandtomorrow⚫

Statinsandrhabdomyolysis⚫RiskfactorsforStevensJohnsonSyndrome⚫Aprotininincardiacsurgeryandincreasedri

skofend-organdamageOctober2009Hangzhou9⚫Why?⚫Pharmacoepidemiology⚫(Therapeutic)riskmanagementOctober2009Hangzhou10Journaloft

heNationalCancerInstitute2005;97:872-3⚫“Therealprobleminsafetyisoftennotinthedrugbuthowitisused,”--RaymondWo

osley,MD,PhD,CriticalPathInstitute⚫“Inmanyways,afteradrugisreleased,therealtrialbegins,”--PaulSeligman,MD,MPH,

OfficeofPharmacoepidemiologyandStatisticalScience,FDAOctober2009Hangzhou11ABriefreview:fromlaboratorytobedside

(T1andT2)⚫Fromlaboratorytoapproval(FDA)⚫Translationalresearch⚫http://www.fda.gov/oc/initiatives/criticalpath/⚫http://www.c-path.org/⚫Fromappr

ovaltobedside/community(AgencyforHealthcareResearchandQualityAHRQ)⚫Sometimescalleddiffusionoftechnology⚫TheroleofepidemiologyinthisprocessOctobe

r2009Hangzhou12R&Dintheindustry⚫Research⚫Basic/benchresearch⚫Animal,cellular,andmolecularmodels⚫Development⚫‘Bringamolecule

tothemarket’⚫EvaluationfromclinicalandbusinessperspectiveOctober2009Hangzhou13Clinicaldevelopment⚫Theroleofsurrogatemarkersandsurrogateen

dpoints⚫Solidepidemiologydataarerequiredtosubstantiatethecorrelationbetweensurrogateendpointsandimportantcl

inicaloutcomes⚫End-pointtrialsandOutcometrials⚫MayormaynotbegoodenoughforsafetyassessmentOctober2009Hang

zhou14Clinicaldevelopment⚫PhaseIIIbandIV⚫Newdosageandformulation(supplementaryNDA)⚫Newindications(supplementaryNDA)⚫Long

termoutcomes⚫Effectiveness(vs.efficacy)⚫Safety⚫Quality-of-life/Functionalstatus/Cost-effectivenessoutcomesOctober200

9Hangzhou15What’swrongwiththecurrentsystemfromasafetyperspective?⚫Limitedgeneralizabilityofanimal,cellular,andmolecularmodels⚫http://

content.nejm.org/cgi/content/abstract/333/17/1099⚫Limitedutilityofsurrogateendpoints⚫Foreffectiveness⚫Forsafety⚫Insufficientknowledgea

boutbackgroundratesandriskfactorsforadverseoutcomesOctober2009Hangzhou16What’swrongwiththecurrentsystemfromas

afetyperspective?⚫LimitationsofPhaseIIItrials⚫Comparisongroup⚫Duration⚫Patientpopulation⚫Pre-marketingvs.Post-marketing⚫Clinicaltrialsvs.Observ

ationalstudies⚫Thechallengeofcombiningstudies(meta-analysis)⚫MisuseofstatisticsOctober2009Hangzhou17TheScienceofSafe

ty“Weareseeingtheemergenceofascienceofsafety.Thissciencecombinesthegrowingunderstandingofdiseaseandits

originsatthemolecularlevel(includingunderstandingofadverseeventsresultingfromtreatment)withnewmethodsofsignaldetection,data

mining,andanalysis,enablingresearcherstogeneratehypothesesabout,andconfirmtheexistence,andcausalfactors,ofsafetyproblemsinthepopu

lationsusingtheproducts.”October2009Hangzhou18http://www.fda.gov/Safety/FDAsSentinelInitiative/ucm089474.htmOctober200

9Hangzhou19ScienceofEfficacyandScienceofSafety⚫Differentparadigm⚫Twosetsofrelatedmethodology,datasources,interpretationoffindings,andregulatoryi

mplications⚫Scienceofefficacy⚫Hypothesistesting⚫Controlledexperiments⚫Singleendpoint/outcome⚫Scienceofsafety⚫Riskidentification/Signald

etection(hypothesisgeneration)⚫Publichealthsurveillance⚫Riskassessment(hypothesistesting)⚫Experimental(clinicaltrial)andobservational(epidemiolog

y)methodsOctober2009Hangzhou20ScienceofSafety⚫Riskidentification/signaldetection⚫Hypothesisgeneration⚫Riskassessment/signalconfi

rmationorrefutation⚫Hypothesistesting⚫Riskmitigation/managementOctober2009Hangzhou21Datasourcesforsafetysignals⚫Ani

mal/cellularmodel⚫Clinicaltrials(pre-andpost-marketing)⚫Adverseevents⚫Laboratorydata(assurrogatemarkers)⚫Numericalimbalancevs.statis

ticalsignificance⚫Spontaneousreports(Pharmacovigilance)October2009Hangzhou22SpontaneousAdverseDrugReactionsReportingSystems⚫Voluntary⚫Fourelemen

ts⚫Anidentifiablepatient⚫Anadverseevent⚫Asuspecteddrug/vaccine/device⚫Anidentifiablereporter⚫SensitivityoverSp

ecificityOctober2009Hangzhou23SeriousAdverseEvents⚫http://www.fda.gov/medwatch/report/DESK/advevnt.htm⚫Death⚫Lifethreatening⚫Hospitalization

(initialorprolonged)⚫Disability⚫Congenitalanomaly⚫RequiresInterventiontoPreventPermanentImpairmentorDamageOctober2009Hangzhou24Str

engthsofthespontaneousreportssystem⚫Moderatecosts⚫Acaseorcaseseriescantellacompellingstory(temporalsequence,de-challenge&re-challenge

)⚫Minimaldelay⚫Workswellwithpublichealth-orientedcliniciansOctober2009Hangzhou25Limitations⚫Anecdotalinformation⚫“Thep

luralofanecdoteisnotdata”⚫(Usually)poordataquality⚫Cannotconclusivelyestablishcausality⚫Reportingratesaffectedbymanysecul

arfactors⚫Hypothesisgeneration,notconfirmatoryOctober2009Hangzhou26Hypotheticalcase#1Suddendeath⚫Year

1988,62year-oldmale,noCVriskfactor,survivedacuteMI⚫Frequentprematureventricularcomplexes(PVCs)afterMI⚫Patientwasgivenencainide,aclassIcan

ti-arrhythmicagent,tosuppressthePVCs⚫6monthslater,patientdiedsuddenly⚫Backgroundrate:one-yearmortality~10%October2009Hangzhou27Wouldanyonesuspec

tthatdeathwasprecipitatedbythedrug?⚫WhatiftherewasnoCASTtrial?⚫LearyWE.WarningIssuedon2HeartDrugsAfterDeathso

fPatientsinTest.NewYorkTimes,April26,1989,page1A.⚫CardiacArrhythmiaSuppressionTrial(CAST)Investigators.Preliminaryreport:…NEng

lJMed1989;321:406-12.(Aug10issue)October2009Hangzhou28Hypotheticalcase#2Liverfailure⚫Year1999,53year-oldmale,BMI31,socialdrinker,on

glyburide,troglitazone,atenolol,simvastatin,anddiclofenac⚫Patientdevelopedacuteliverfailureanddiedwhilewaitingforlivert

ransplantation⚫Liverfailuremaybecausedbyadrug⚫butwhichdrugorcombinationofdrugs?October2009Hangzhou29Whatisthedifferencebetweencase#1andcase#2?⚫Suspic

ionandreportingofAdverseDrugReactions⚫Severityofevent⚫Priorknowledge⚫Alternativeexplanations⚫Backgroundrateoftheadverseevent⚫Rare⚫NotsorareOctob

er2009Hangzhou30Rareandnot-so-rareadverseevents⚫Rare⚫Rhabdomyolysis,DILI,Stevens-JohnsonSyndrome,agranulo

cytosis,Churg-StraussSyndrome,…⚫Notsorare(especiallyamongcertainpopulations)⚫Myocardialinfarction⚫TuberculosisamongRApatients⚫Chr

onicactivehepatitisBexacerbation⚫Suicideamongadolescents⚫,…October2009Hangzhou31Hypotheticalcase#3Acutemyocardialinfarction⚫Ye

ar2000,49year-oldmale,smoker,BMI33,mildhypertensionandLDLelevation⚫Leftkneeosteoarthritis⚫Givenrofecoxib25mg/dayfo

rtwoweeks,thenthedoseincreasedto50mg/day⚫9monthslater,hadnon-fatalmyocardialinfarction⚫BackgroundrateofAMI:severalcasesper1,000person-yea

rsOctober2009Hangzhou32Aninterestingquestion⚫CouldtheexistingsystemidentifyCVsafetysignalsassociatedwithrofecoxibiftherewasnoVIGORtrial?

⚫PUBMEDsearchonOct24,2005,publishedcasereportsofrofecoxib:hepatitis,interstitialnephritis,colitis,angioedema,anaphylacticshock,gynecomastia,acut

erenalfailure,delirium,Stevens-Johnsonsyndrome,congestiveheartfailure,transientvisualimpairment,aquage

nicwrinklingofthepalms,…October2009Hangzhou33Epidemiology,ageneraloutline⚫MeasuresofDiseaseOccurrence⚫

IncidenceandPrevalence⚫Cumulativeincidenceandincidencerate⚫Conceptofdynamicpopulationandpersontime⚫MeasuresofAssociation⚫Relativ

erisk⚫Rateratio⚫OddsratioOctober2009Hangzhou34Epidemiology,ageneraloutline⚫DescriptiveEpidemiology⚫Caserepo

rts⚫Caseseries⚫Frequencyofevents⚫Studiesprovidingsomeevidence,butnotdefinitive⚫Cross-sectionalstudies⚫EcologicalstudiesOctober2

009Hangzhou35Epidemiology,ageneraloutline⚫Observationalstudies,selectedmethods⚫Cohortstudies⚫Case-contr

olstudies⚫Case-crossoverstudies⚫Interventionalstudies⚫Clinicaltrialsforindividualsubjects⚫ClusteredrandomizedtrialsOctob

er2009Hangzhou36Epidemiology,ageneraloutline⚫Combininginformationfrommultiplestudies⚫Meta-analysis⚫DatapoolingOctober2009Hangzhou37Systematica

ndrandomerrors⚫Validityconsiderationinepidemiology⚫Confounding⚫Bias⚫Informationbias⚫Selectionbias⚫Assessmentofrandomerrorinepidemiology:statistics

October2009Hangzhou38CommonlyusedtermsinepidemiologyConfounder(s)ExposureOutcomeOctober2009Hangzhou39Exposuresinpharmacoepidemiology⚫Dr

ugs⚫Vaccines⚫Medicaldevices⚫Procedures⚫Programs⚫e.g.screeningprograms⚫CodingsystemsforeachofthemOctober2009Ha

ngzhou40AscertainmentofOutcomes⚫Operationaldefinitions⚫Sensitivityandspecificity⚫Objectivecriteria⚫E.g.hemorrhagicst

roke⚫BlindedadjudicationifpossibleOctober2009Hangzhou41PerformanceofICD-9codesinthestudyofrhabdomyolysisOctober2009Hangzh

ou42Andradeetal.Healthplanadministrativedatabasescanefficientlyidentifyseriousmyopathyandrhabdomyolysis.JC

linEpi2005;58:171-4October2009Hangzhou43Cohortstudiesinpharmacoepidemiology⚫Sourcepopulation⚫Ascertainmentofexposure⚫Comparisongroups/

cohorts⚫Confounding?⚫Ascertainmentofoutcomes⚫InformationBias?⚫Controlforconfounding⚫Evaluationofeffectmo

dification⚫StrengthsandlimitationsOctober2009Hangzhou44ExposureandOutcome⚫Exposuresofinterest/Comparisongroups⚫Head-to-headcomparison⚫Drugvs.D

iseasecomparison⚫Multiplecomparisongroups⚫Outcomesofinterest⚫ClinicallyrelevantOctober2009Hangzhou45ArchInternMed2002;162:936-42.⚫Threestudygrou

ps⚫Alendronateusers(n=6,432)⚫Womenwithosteoporoticfracture(n=9,776)⚫Age-gendermatchedunexposed(n=33,176)⚫Compari

ngalendronateusersandunexposed⚫Adjustedrateratio=1.8,95%confidenceinterval0.8–3.9⚫Comparingalendronateusersandwomenwithfracture⚫Adjuste

drateratio=1.1,95%confidenceinterval,0.6-2.2October2009Hangzhou46Definingcohorts⚫Matching⚫Individually-match⚫Frequency-mat

ch⚫Matchingincohortstudyvs.matchingincase-controlstudy⚫Restriction/Selectioncriteria⚫New-userdesign⚫Incidentusersvs.PrevalentusersOctober2

009Hangzhou47Biasandconfounding⚫Confounding⚫Channeling,Confounding-by-Indication,andConfounding-by-Severity⚫

Howtomeasureindicationandseverity?⚫Otherknownriskfactorsfortheoutcomeofinterest(age,gender,smoking…)⚫Bias⚫Detectionbiasasaformofinformationbia

sOctober2009Hangzhou48Analyticapproaches⚫‘Atrisk’person-time⚫Immortalperson-time⚫Cumulativeincidence,incidence,andcorrelatedout

comes⚫Count,person-time,andtime-to-event⚫Controlforconfounding⚫Multipleregression,propensityscore,instrumentalvariable

s,andothers⚫Accountforunmeasuredconfounders⚫Nestedcase-controlstudyOctober2009Hangzhou49Stratifiedanalysis⚫Cont

rolforconfounding⚫Effectmodification⚫Specificsubgroup(s)⚫Drug-druginteractionOctober2009Hangzhou50Cohortstudiesinpharmacoepidemiol

ogy⚫Strengths⚫RelativeriskandAbsoluteriskestimates⚫Multipleoutcomes(safetysurveillanceformultipleoutcomesandrisk-benefitassessment)⚫Limitations⚫Mitig

atedbytheavailabilityofsecondaryand‘hybrid’datasourcesOctober2009Hangzhou51Clinicaltrials⚫‘ClinicalRegimen’trialsvs.‘Clin

icalpractice’trials⚫InclusionandExclusioncriteria⚫Titration,switching,andstopping⚫Follow-up/monitoring⚫NewDrugApplicationvs.Clinicalpractice⚫Placeb

o-controlledvs.head-to-headcomparison⚫Internalvalidityvs.Generalizability⚫Comorbidityandpotentialdrug-druginteractionOctober2009Hangzhou52Explanato

ryvs.PragmaticRolandandTorgerson.Understandingcontrolledtrials:Whatarepragmatictrials?BMJ1998;316:285“Trialsofheal

thcareinterventionsareoftendescribedaseitherexplanatoryorpragmatic.Explanatorytrialsgenerallymeasureefficacy—th

ebenefitatreatmentproducesunderidealconditions,oftenusingcarefullydefinedsubjectsinaresearchclinic.Pragmatic

trialsmeasureeffectiveness—thebenefitthetreatmentproducesinroutineclinicalpractice.”October2009Hangzhou53Somelarge

trials⚫ALLHATstudy⚫http://allhat.uth.tmc.edu/⚫PROVE-IT/TIMI-22study⚫http://content.nejm.org/cgi/content/abstr

act/350/15/1495⚫CATIE⚫http://www.nimh.nih.gov/press/catie_release.cfm⚫http://content.nejm.org/cgi/content/abstract/353/12/1209October

2009Hangzhou54Clinicaltrialsforsafety⚫LargeandSimpleTrial⚫LargeandDirtytrial(Meinert)⚫Ahybridapproach⚫Randomiza

tion⚫Individuallevel⚫Grouplevel⚫‘Real-life’managementandfollow-up⚫ImplicationsforcomparativeeffectivenessresearchOctober2009Hangz

hou55Exampleofalargeandsimpletrial⚫http://clinicaltrials.gov/ct2/show/NCT00418171⚫LargeSimpleTrial(LST)OfCardiovascularSafetyOfZi

prasidoneAndOlanzapine-(ZODIAC)⚫AnInternational,Multicenter,LargeSimpleTrial(LST)ToCompareTheCardiovascularSafetyOfZiprasidoneAndOlanzapine⚫T

heprimaryobjectiveofthestudyistoestimatetherelativeincidenceamongusersofziprasidoneandolanzapineofnon-suicidemortalitywithin12months.⚫

18,239enrolledpatientsOctober2009Hangzhou56EligibilitycriteriaforZODIAC⚫InclusionCriteria:⚫Patientsnewlytreatedf

orschizophreniaandthosereceivingcontinuingtreatmentwillbeeligibleifthetreatingpsychiatristisreadytoinitiate

anewantipsychoticmedication,andwouldconsiderusingeitherziprasidoneorolanzapineasanappropriatetherapy.⚫Age18andolder⚫Maleand

femalepatientswhomeetallcriterialistedbelowareeligibletobeenrolledinthisstudy:⚫Diagnosedwithschizophrenia⚫Willingtoprovideinformationo

natleastonealternatecontactpersonforstudystafftocontactregardingpatient'swhereabouts,shouldthepatientb

elost-to-follow-upoverthecourseofthestudy⚫ExclusionCriteria:⚫Progressivefataldiseaseofalifeexpectancywhichprohibitst

hemfromparticipatinginaoneyearresearchstudy⚫Previouslyrandomizedtostudymedicationandenrolledinthisstud

yOctober2009Hangzhou57Clinicaltrial,statisticalperspective⚫One(andonlyone)pre-specifiedprimaryendpoint⚫TypeIandTypeIIerrors,samplesizeest

imation⚫Erronwhichside?⚫Subgroupanalysis⚫Hypothesisgeneration,notconfirmation⚫Superioritytrialvs.Non-inferioritytrial⚫

Implicationsforthechoiceofcomparisongroup⚫Multiplecomparison(no-no!)October2009Hangzhou58Subgroupanalysis⚫HortonR.Fromstarsignstotrialguidelines.Lanc

et2000;355:1033-34⚫ISIS-2Collaborationgroup.Randomisedtrialofintravenousstreptokinase,oralaspirin,both,orneitheramong17,

187casesofsuspectedmyocardialinfarction.Lancet1988;ii:39-60⚫“Theanalysisshowedthataspirinwasbeneficialinallpatientsexceptthosewiththestarsign

sofLibraandGemini.”October2009Hangzhou59Differencebetweenpharmacoepidemiolgyandothertopicsinepidemiology⚫Focus⚫Exposurevs.Outcome⚫Confounder

s⚫Bias⚫Agreementwithclinicaltrials⚫RegulatoryimplicationsOctober2009Hangzhou60Whatiftheclinicaltrialsresultsdiffe

rfromtheobservationalfindings?⚫Women’sHealthInitiative,hormonereplacementtherapy,andcardiovascularoutcomes⚫http://jama.ama-assn.org/cgi/co

ntent/abstract/288/3/321⚫http://jama.ama-assn.org/cgi/content/abstract/297/13/1465⚫http://content.nejm.o

rg/cgi/content/abstract/356/25/2591October2009Hangzhou61Clinicaltrialsandobservationalstudies⚫“Rethinkingthehierarchyofevidence”–Vandenbroucke⚫O

bjectivesofstudy⚫Intendedeffectoftherapy(confirmation)⚫Unintendedeffectoftherapy(discovery)October2009Hangzhou6

2Aspecialmethodinpharmacoepidemiology:Registrieshttp://effectivehealthcare.ahrq.gov./reports/topic.cfm?topic=0&sid=2&rType=11&sType=3Aregist

ryisbasicallyaprospectivelyidentifiedcohort,definedby⚫Disease⚫Exposure⚫OutcomeOctober2009Hangzhou63Registriesinpharmacoepidemiol

ogy⚫Diseaseregistry⚫e.g.cancer,diabetes,rheumatoidarthritis,multiplesclerosis,myocardialinfarction(NationalRegistryofMyocardia

lInfarction)…⚫Exposure(medicalproductsandprocedures)⚫Coronaryrevascularization⚫Specialpopulation⚫Pregnancyregistry⚫Adv

erseoutcomes⚫AcuteliverfailureOctober2009Hangzhou64Lancet2002;359:1173–77October2009Hangzhou65Someregistryexamples

October2009Hangzhou66Antiepilepticdrugpregnancyregistry⚫http://www.aedpregnancyregistry.org/⚫WyszynksiD.F.,NambisanM.,SurveT.,AlsdorfR.

M.,SmithC.R.,HolmesL.B.Increasedrateofmajormalformationsinoffspringexposedtovalproateduringpregnancy.Neurology2005;64

:961-965.⚫HolmesL.B.,WyszynskiD.F.NorthAmericanAntiepilepticDrugPregnancyRegistry.Epilepsia2004;45(11):1465.⚫HolmesL.B.,Wy

szynskiD.F.,LiebermanE.S.,TheAED(AntiepilepticDrug)PregnancyRegistry:A6-YearExperience.ArchNeurol.2004;6

1:673-678.October2009Hangzhou67Registry,surveillance,andsafetyassessment⚫Registryofdisease⚫Internalcomparisongr

oups⚫Seculartrendintermsofclinicalepidemiologyandtreatment⚫Treatmentregistry⚫Limitedbyavailabilityofexternalcomparisongroups

⚫Adverseoutcomeregistry⚫Well-suitedforcasecontrolstudyOctober2009Hangzhou68Casecontrolstudiesinpharmacoepidemiology⚫Identifyingthecases

⚫Completecaseascertainment?⚫Selectionofcontrols⚫Nestedcase-controlstudies(fromadefinedstudypopulation)⚫Non-nestedcase-contro

lstudies⚫Hospitalbased⚫Communitybased⚫Ascertainmentofexposure⚫AscertainmentofpotentialconfoundersOctobe

r2009Hangzhou69Case-controlstudies⚫Confounding⚫Confoundingbyindication/severity⚫Bias⚫Selectionbias⚫Inform

ationbias(recallbias)October2009Hangzhou70Asthmadeaths⚫http://www.nytimes.com/1995/05/10/obituaries/teresa-teng-singer-40

-dies-famed-in-asia-for-love-songs.html⚫Pearce&Hensley.Epidemiologicstudiesofbetaagonistsandasthmadeaths.Epid

emiolRev1998;20:173-86October2009Hangzhou71Outcomeregistryandcase-controlstudyNEnglJMed1995;333:1600-7October2009Hangzhou72NEn

glJMed1996;335:609-16October2009Hangzhou73Case-controlstudiesofcongenitalanomaly⚫NEnglJMed2003;349:2556-59October2009Hangz

hou74NEnglJMed2000;343:1608-14.October2009Hangzhou75SloneEpidemiologyCenter“Since1976,theSloneEpidemiologyUnitBirthDefectsStudyhas

interviewedmothersoflive-borninfantswithmalformationsin80maternityortertiarycarehospitalsinthegreatermetropolitanareasofBoston,Ph

iladelphia,andToronto,and,between1983and1985,inpartofIowa.Since1988,womenwhohadstillborninfantsandtho

sewhosefetuseswereabortedbecauseofamalformationhavealsobeenincluded.Studysubjectsareidentifiedthroughthereviewof

admissionsanddischargesatmajorreferralhospitalsandclinicsandthroughregularcontactwithnew-bornnurseriesincommunityhospitals(toidentifyinfantswithm

alformationswhosemothersmaynothavebeenreferredtomajorcenters).”October2009Hangzhou76Case-crossoverdesign⚫Similartoacrossovertr

ial⚫Evaluationofshortterm/‘triggering’effect⚫Mittlemanetal.TriggeringofAcuteMyocardialInfarctionbyHeavyPhysicalExerti

on--ProtectionagainstTriggeringbyRegularExertion.NEnglJMed1993;329:1677-83⚫Hunteretal.Recentdiureticuseandtheriskofrecurrentgoutattacks:theonlinecase

-crossovergoutstudy.JRheumatology2006;33:1341-5October2009Hangzhou77October2009Hangzhou78JRheumatol2006;33:1341–5October2

009Hangzhou79Datapoolingandmeta-analysisJAMA2001;286:954-9October2009Hangzhou80Datapoolingandmeta-analysis⚫JAMA2005;294:258

1-6.October2009Hangzhou81PostingofclinicaltrialdatainthepublicdomainOctober2009Hangzhou82www.gsk-clinicalstudyregis

ter.com/files/pdf/3104.pdf⚫Rationale:AspartofGSK’songoingpharmacovigilanceprogram,anevaluationoftheassociation(ifany)betweenrosiglitazone(R

SG)andcardiovasculareventswasundertakenforeventsofcongestiveheartfailure(CHF)andforeventsrelatedtomyocardia

lischemia.Twodistinctapproacheshavebeenusedinthisevaluation.⚫OneapproachexaminedeventsofCHFaswellaseventsrelatedtomyocardialischemiautilizingaretros

pectivestatisticalanalysisofintegratedclinicaltrials.ThemajorityofpatientsinthisanalysiswererandomizedtoafixeddoseofRS

Gaddedtobackgroundtherapyvsmaintenanceofthebackgroundtherapyalone.⚫Thesecondapproachexaminedeventsrelatedtomyocardialischemia

(hospitalizationsformyocardialinfarctionand/orcoronaryrevascularization)utilizingalarge,well-balancedretrospective

observationalstudy.Comparisonsweremadebetweenmatchedcohortsofpatientsinitiatingantidiabetictherapyasoralmonotherapyororaldualtherapyo

rcombinationswithinsulininwhichpatientsweretreatedinaUSclinicalpracticeenvironment.October2009Hangzhou83Lancet200

9;373:2125-35.⚫321peopleintherosiglitazonegroupand323intheactivecontrolgroupexperiencedtheprimaryoutcomeduringamean5.5-yearfollow-up,…⚫HRwas0.84(0

.59-1.18)forcardiovasculardeath⚫1.14(0.80-1.63)formyocardialinfarction⚫0.72(0.49-1.06)forstrokeOctober2009Hangzhou84http://www.fda.gov/bbs/

topics/NEWS/2007/NEW01597.htmlThroughoutFebruaryandMarch2007,NovartisreportedtoFDAtheresultsofanewanalysisof29sh

ort-term(1-3months)randomized,controlledclinicaltrialsofZelnorm.FDAhasconcluded,basedonthesedatathatform

ostpatientsthebenefitsofthisdrugnolongeroutweightherisks.Theanalysisincludedmorethan11,600patientstreatedwithZelnormandover7000

patientstreatedwithplacebo.Thedatashowedthattheriskofseriouscardiovascularadverseevents(e.g.,angina,hear

tattacks,andstrokes)associatedwithuseofZelnormishigherthanwithplacebotreatment.ThirteenZelnorm-treatedpatients(or0.1%)hadconfirmedcardio

vascularischemicevents,andonly1placebo-treatedpatient(or0.01%)withanevent.October2009Hangzhou85Datapooling

conductedbydrugcompaniesOctober2009Hangzhou86Tegaserodwww.fda.gov/bbs/topics/NEWS/2007/NEW01597.html⚫"ThisdecisionreflectstheFDA'scommitmentt

ocontinuouslymonitorapproveddrugsthroughouttheirmarketinglife,andtakeactionwhenwebelievetherisksexceedthebenefits,"saidD

r.DouglasThrockmorton,DeputyDirectorfortheCenterforDrugEvaluationandResearch."Here,apotentialriskofveryseri

ousharmtopatientswhohavethisnon-life-threateningconditionwasrecentlyidentified,makingthisactionnecessary."October2009Hangzhou87Risk-benefitcalculu

s⚫Alternatives⚫Naturalhistoryofdisease⚫Characterizeandquantifythebenefit⚫EfficacyANDEffectiveness⚫Characterizeandquantifytherisk

s⚫Utility/Valueassessment⚫Populationlevelvs.individuallevelOctober2009Hangzhou88Anotherdatapoolingexercise:Tiotropiuman

dstrokeOctober2009Hangzhou89http://www.fda.gov/cder/drug/early_comm/tiotropium.htmBoehringerIngelheimreportedtotheFDAthatitha

sconductedananalysisofthesafetydatafrom29placebocontrolledclinicalstudies(“pooledanalysis”).In25ofthec

linicalstudies,patientsweretreatedwithSpirivaHandiHaler.Intheother4clinicalstudiespatientsweretreatedwithanotherformu

lationoftiotropiumapprovedinEurope,SpirivaRespimat.The29clinicalstudiesincludedapproximately13,500patientswithCOP

D.Basedondatafromthesestudies,thepreliminaryestimatesoftheriskofstrokeare8patientsper1000patientstreatedforoneyearwithSpiriva,and6patientsper

1000patientstreatedforoneyearwithplacebo.ThismeansthattheestimatedexcessriskofanytypeofstrokeduetoSpirivais2patientsforeach1000patientsusingSpir

ivaoveraoneyearperiod.October2009Hangzhou90http://www.fda.gov/cder/drug/early_comm/tiotropium.htmItisimportanttointerp

retthesepreliminaryresultswithcaution.FDAhasnotconfirmedtheseanalyses.Pooledanalysescanprovideearlyinformati

onaboutpotentialsafetyissues.However,theseanalyseshaveinherentlimitationsanduncertaintythatrequirefurtherinvestigationusingotherdat

asources.ThisearlycommunicationisinkeepingwithFDA’scommitmenttoinformthepublicaboutitsongoingsafetyrev

iewsofdrugs.October2009Hangzhou91EarlyCommunicationaboutanOngoingSafetyReviewofTiotropium(marketedasSpirivaHandiHaler)⚫Updat

e10/07/2008:FDA’sEarlyCommunicationAboutanOngoingSafetyReviewissuedonMarch18,2008statedthatBoehringerIngelheim,themak

erofSpirivaHandiHaler(tiotropiumbromide),hadconductedapooledanalysisof29trialsthatsuggestedasmallexcessri

skofstroke(2casesper1000)withtiotropiumbromideoverplacebo.FDAhasnowreceivedpreliminarydatafromUPLIFT(Und

erstandingthePotentialLong-TermImpactsonFunctionwithTiotropium),alarge,4-year,placebocontrolledclinicaltrialwithSpirivaHandiHalerinapproxim

ately6000patientswithchronicobstructivepulmonarydisease(COPD).ThepreliminaryresultsofUPLIFTreportedbyBoehringerIngelheimtotheFDAs

howedthattherewasnoincreasedriskofstrokewithtiotropiumbromide(SpirivaHandiHaler)comparedtoplacebo.October2009Hangz

hou92Potentialissueswithdatapoolingandmeta-analysis⚫Publicationbias⚫Whetherthestudywaspublished⚫Whetherthesafetyend

-pointwasreported⚫Studypopulationofthetrials⚫Exclusion/exclusioncriteria⚫Highriskpatients?⚫Real-lifepractice?⚫Durationoffo

llow-up?⚫Methodstocombinedata⚫Fixedeffectassumption⚫Randomeffectmodel⚫ZerocellsOctober2009Hangzhou93Regulatoryoptionswhentherei

sasafetyconcern⚫Delayedapprovalandwithdrawal⚫Tremendoussocietalcosts⚫Conditionalapprovaland(Therapeutic)RiskManagemen

t⚫Productlife-cycleassessment⚫Riskidentification(signaldetection)⚫Riskassessment(incidenceandriskfactors)⚫Riskcommunication⚫RiskManagementPl

ans(RMP)/RiskEvaluation&MitigationStrategy(REMS)October2009Hangzhou94InresponsetotherofecoxibeventandInstituteofMedicinerep

ortFDAinitiativesincludedevelopingnewscientificapproachestodetecting,understanding,predicting,andpreventingad

verseevents,developingandincorporatingnewquantitativetoolsintheassessmentofbenefitandrisk,andconductin

gapilotprogramtoreviewthesafetyprofilesofcertainnewlyapproveddrugsonaregularlyscheduledbasis.http://www.f

da.gov/bbs/topics/NEWS/2007/NEW01551.htmlOctober2009Hangzhou95Someexamples⚫Isotretinoinandpregnancy⚫Clozapineandagranulocytosis⚫Thalidomideform

ultiplemyeloma⚫Alosetronforirritablebowelsyndrome⚫…⚫http://www.fda.gov/cber/genadmin/rems.htm⚫Needobje

ctiveevidencetoevaluatetheeffectivenessofRMPsOctober2009Hangzhou96Restrictedaccess:treatmentINDOctober2009H

angzhou97Arethereeffectivetoolstoidentifyhighriskpatients?October2009Hangzhou98ATRMtool:blackboxwarning⚫BlackBoxWarningand“DearDoctorLett

er”⚫Howeffectivearethey?⚫Smalleyetal.ContraindicatedUseofCisapride:ImpactofFoodandDrugAdministrationRegulatoryActionJAMA2000;284

:3036-9.⚫Woosley.DrugLabelingRevisions—GuaranteedtoFail?JAMA2000;284:3047-9.October2009Hangzhou99October2009Hangzhou100Smalleyetal.JAMA2000;284:303

6-9.