Fundamentals of Trial DesignBruce E. Sands, MD, MSDr. Burrill B. Crohn Professor of MedicineChief of the Dr. Henry D. Janowitz Division of GastroenterologyIcahn School of Medicine at Mount SinaiMount Sinai Medical CenterNew York, NY1

DisclosuresResearch grantsAbbott Immunology, Amgen, GlaxoSmithKline, Janssen Biotech, Pfizer, Prometheus Laboratories, TakedaConsultant and received honorariaAGA Institute, Atlantic Healthcare, Avaxia Biologics, Baxter, Bristol-Myers Squibb, Curatio CME, Dainippon Sumitomo Pharma, Elan, IMEDEX, Janssen, Millennium, New York School of Medicine, Pfizer, Puretech, Resolvyx Pharmaceuticals, Sigmoid Biotechnologies, Takeda, Teva, University of Western Ontario, Washington UniversityStock in a non-publicly traded companyAvaxia Biologics2

OverviewGoals of trial designSpecial challenges in designing studies in Crohn’s diseaseHow can trial design support clinical practice?Trial design elementsUse of placeboOther design considerations3

The Goals of Trial DesignTo evaluate efficiently —and if appropriate, bring to approval—new therapies that provide to patients meaningful benefit with acceptable riskEssentially, to support approval throughSelection of right dose(s) and schedulesAppropriate assessment of efficacyAppropriate assessment of safetyAt the right time/period of observation4

Special challenges in designing and performing studies in Crohn’s diseaseNot a common disease—limited numbers of patients to enroll in studiesIncreasingly effective therapeutic choices availableHistorically high placebo response rateDepends upon outcome measures, disease activity at baseline, duration of disease, parenteral vs. oral, number of study visits, efficacy definitions, timing of endpoint, concomitant/prior treatment, patient populationHeterogeneity of disease expressionDefining what is meaningful to patientsClinically relevant outcome, effect sizeDefining what is acceptably safe for patientsCarryover effect in crossover designDepending upon PK, PD, and patient population5

Symptoms reach treatment thresholdStop or continue base therapy?New treatment begunAppropriate treatment durationYesNoSTOPCONTINUEt0t1Observe on treatmentWhat happens today in clinical practice?Patient better?No limiting side effect?6

Trial Design Elements to Consider7

Trial Design Elements to Consider8

Trial Design Elements to Consider9

Use of placeboBy far the most common designAs clinically meaningful effect sizes can be relatively modest, active comparator would require large sample sizeEthically accepted, as usually layered over established therapy (albeit ineffective, by definition)Important in safety analysis, though exposure to drug or placebo may change in seamless studies, or induction-maintenance combination studiesEMEA sentiment is for active comparison10

What about placebo in other conditions?In RA, it is felt that keeping patients on placebo with uncontrolled disease for long periods raises ethical concernsExposure to placebo and ineffective therapy should be kept to a minimum (~12 weeks)Short-term outcome assessment is feasible, but this limits ability to assess long-term outcomesFDA: Draft Guidance for Industry, Rheumatoid Arthritis: Developing Drug Products for Tratement, May 2013, Revision 1.11

What trends will affect clinical practice?Increasing willingness to integrate objective evidence of responseBut little consensus about which modalities to incorporate, thresholds for therapeutic decisions, or timing of response measuresLittle understanding of the cost-benefit propositionIncreasing incorporation of pharmacokinetic measures (drug levels, antibodies to drug) in therapeutic decisionsLittle prospective data on timing, frequency and definitions of “therapeutic window”12

What designs have we yet to see in registrational studies in Crohn’s disease?Dosing adapted to target drug level (“tailored dosing regimen”)Selection of population by evidence of a specific disease mechanism (“personalized medicine”)Studies of drug algorithmCluster randomization studies13

Considerations in early phase trial designPhase II can be relatively large if traditional disease activity definitions of response/remission are usedNovel trial designs used in other disciplines (cancer) Phase II designs that allow early stopping for lack of efficacyBayesian designs—but dependence on knowing “prior distribution” to interpret “posterior distribution” can be problematicLikelihood approaches (“probability of misleading evidence”) to make inferencesMultiple outcome studies that incorporate safety and efficacy—but safety is rarely limiting with our agentsActive comparator, or “reference arm”Biomarker as primary outcome of interestAdaptive randomization design (“play the winner”)14

Challenge QuestionsWhat is the proper use of placebo in clinical trials in Crohn’s disease?What features of Crohn’s disease trial design best translate to clinical practice?15