FDA Experience with End of Phase IIa Meetings: An Attempt to Improve Drug Development Decisions - PowerPoint PPT Presentation

Loading...

PPT – FDA Experience with End of Phase IIa Meetings: An Attempt to Improve Drug Development Decisions PowerPoint presentation | free to download - id: a6fad-NTE3O



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

FDA Experience with End of Phase IIa Meetings: An Attempt to Improve Drug Development Decisions

Description:

FDA Experience with End of Phase IIa Meetings: An Attempt to Improve Drug Development Decisions ... Millenium. Eli Lilly. ABBOTT. PFIZER. FDA. CDISC ... – PowerPoint PPT presentation

Number of Views:112
Avg rating:3.0/5.0
Slides: 31
Provided by: powe96
Learn more at: http://www.fda.gov
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: FDA Experience with End of Phase IIa Meetings: An Attempt to Improve Drug Development Decisions


1
FDA Experience with End of Phase IIa Meetings An
Attempt to Improve Drug Development Decisions
Acknowledge Larry Lesko, Don Stanski, Joga
Gobburu, Peter Lee, Yaning Wang, Jenny Zheng and
many others
  • Bob Powell, Pharm.D.
  • Office of Clinical Pharmacology
  • FDA
  • powellr_at_cder.fda.gov
  • (301) 796-1589

2
When to Influence Clinical Drug Development ?
  • Flexibility
  • Learning
  • RD Expense
  • Revenue

NDA
Market
Phase 1
Discovery
Phase 2
Phase 3
Preclinical
3
50 Clinical Trial Failure RateIs it true? What
to do?
  • Root Cause
  • Ø Efficacy
  • ? Toxicity
  • Placebo
  • Baseline
  • Dropouts
  • Patient Selection

Account for known failure sources (prior
information) in clinical trial design
4
Reasons for Poor Decisions(Definition an
outcome which should/could have been anticipated)
  • Conspiracy of optimism
  • Framing the problem too narrowly to bring it
    inside my own comfort zone
  • Not involving the right people
  • Avoiding uncertainty
  • Ignoring information I do not understand
  • Being attached to sunk costs high spent
    development costs
  • Ignoring risks
  • Assuming no uncertainty in potential outcomes
  • Making decision alone

BIAS
Hammond, Keeney, Raiffa. Smart Choices Harvard
Business School Press, 1999
5
Model Based Drug DevelopmentWhat is it?
  • Objective improve decision quality by employing
    drug-disease models clinical trial simulation
  • Model mathematical explanation of relationships
    thought to explain outcome over time period of
    interest
  • Drug-Disease Model (empiric mechanistic)
  • Disease model relationship of patient (e.g.,
    gender, age, genotype), biomarker (e.g.,
    biochemical, imaging) relationship to disease
    morbidity and mortality
  • Drug-Disease model addition of drug (dose,
    concentration, combination, placebo) and patient
    (e.g., size, age, adherence, dropout) effects and
    adverse effects to the disease model
  • Simulation- Target
  • Clinical trial design- optimal
  • New designs-enrichment, randomized withdraw,
    adaptive
  • Dosage regimen(s) selection
  • Go/No go- Sponsor /or FDA
  • Labeling- Sponsor /or FDA

6
End of Phase 2a Meetings
  • Purpose ? Late phase clinical trial (2b, 3)
    unnecessary failure
  • Format non-binding scientific interchange.
    Marketing issues should be in the development
    plan, not at this meeting.
  • Deliverables
  • Perform modeling (relevant phase 1/2a data)
    simulation of next trial design employing
  • Mechanistic or empirical drug-disease model
  • Literature estimates for comparative drug effects
    if relevant
  • Placebo effect (magnitude time-course)
  • Rates for dropout and compliance. (prior FDA
    experience)
  • Recommendation on sponsors trial design
    alternative including patient selection, dosage
    regimen,
  • Code from FDA work, Sponsor can extend work
    (EOP2, NDA)
  • Answers to other questions from the clinical and
    clinical pharmacology development plan
  • Time-course 6 weeks
  • Key sponsor FDA participants physician,
    biostatistician, clinical pharmacology
    (pharmacometrics), project management

7
EOP2a Meeting Process
Meeting Request Letter Questions
FDA Receive Briefing Package, Data, Next Trial
Design
Sponsor Phone Meeting Explain Process Data
Needed
FDA Evaluation Approval
6 week start
  • Final Meeting
  • Focus on Drug-disease modeling Clinical trial
    simulations
  • 30-40 min Presentation
  • 1 hour dialogue focused on trial design, dosage
    regimens, patient selection
  • Simulation Strategy
  • Trial Design Alternatives
  • Dosage Regimens
  • Sample Size
  • Patient Selection
  • Sponsor or FDA?

FDA begin data analysis
Sponsor Questions
Answer other question in writing before meeting
8
Roles Responsibilities
  • Project manager
  • Sponsor communication
  • FDA meetings
  • Documentation
  • Physician
  • Primary endpoints
  • Disease information
  • Trial design
  • Draft guidance
  • Clinical Pharmacology/ Pharmacometrics
  • Drug-disease modeling
  • Dosage regimens
  • Drug interactions
  • Simulations
  • Statistician
  • Trial Design
  • Prior trial information
  • Placebo
  • Dropout rates
  • Simulation

9
Case Study HIV Phase 2a Meeting Key Questions
for Drug X
  • Is the target AUC of 950 ngh/mL (based on
    relationship to viral load suppression at day 11
    of Phase IIa trial) reasonable to select the best
    dose for Phase III?
  • Is testing BID and QD regimens appropriate?
  • Are 4 weeks adequate to select the dosage regimen
    for Phase III?

10
RNA Change from BL vs AUC(0-24)
An AUC(0-24) of 950 ng-h/mL is predicted to
achieve a 1.5 log10 decrease in HIV-1 RNA from BL
11
A Mechanistic PK/PD Model
  • Mechanistic viral dynamic model offers potential
    advantages over empirical models
  • Time course of virus load described
  • Schedules (bid vs. qd) can be differentiated
  • Drug-drug interactions
  • Different design scenarios can be evaluated
  • Adherence
  • Pharmacodynamic interactions
  • Drop out
  • Resistance

12
Proposed Phase IIb Design
  • Cohort 1 (N50)
  • Drug X 1 mg BID LPV/r 400/100mg BID
  • Cohort 2 (N50)
  • Drug X 2 mg BID LPV/r 400/100mg BID
  • Cohort 3 (N50)
  • Drug X 4 mg QD LPV/r 400/100mg BID
  • Standard of Care (SOC) (N25)
  • COMBIVIR (ZDV/3TC 300/150mg BID)
  • LPV/r
    400/100mg BID

13
Virus Dynamic Model (virus load vs time)
p
d2
PI
Active Infected
l production rate of target cell d1
dying rate of target cell c dying rate
of virus b infection rate
constant d2 dying rate of active
cells d3 dying rate of latent cells p
production rate of virus
l
fAbVT
CD4 Cells
(N)NRTI
Virus
a

fLbVT
Latent Infected
(N)NRTI
d1
c
d3
fA0.96 and fL0.03
J Acquir Immun Defic Syndr 26397, 2001
14
Drug X Observed and Model Predicted Mean Virus
Load vs Time Dose
2 mg QD 4 mg QD 2 mg BID 6 mg BID
15
Individual Patients Fit for Drug X
16
Drug-Disease Model ComponentsApplication
  • Adherence
  • ?Dose ? GI Adverse Effects ? ?Adherence
  • Dropout rate biphasic
  • Prior experience from other drugs
  • Drug-drug interaction
  • PK
  • PD
  • Verified model with data from prior drugsdid not
    share with sponsor

17
Viral Dynamics During 4 WeeksDrug X Kaletra
Kaletra BID
HIV RNA (Log10 copies/mL)
2Log Drop
Drug X 0.5, 1, 2 mg BID and 4QD
Log10(50)
Time (week)
18
Dropout Model Prior Study Submission
19
20 Simulated trials (2 log drop, 90 Adherence,
No Drop-out) suggest 2 BID is most likely the
winner
100
90
80
70
EQUAL
60
4 QD
Chances of Being the Winner
50
2 BID
40
1 BID
30
20
10
0
4 8 12 16 20 24 28 32
36 40 44 48
Week
20
HIV Phase 2a Meeting Key FDA Response to
Questions for Drug X
  • Is the target AUC of 950 ngh/mL reasonable?
  • Concentration targeted dose selection is more
    appropriate to compare schedule (BID vs QD).
  • Is testing BID and QD regimens appropriate?
  • BID regimen is preferable
  • 0.5 mg BID, instead of 4 mg QD, is worth
    considering
  • Are 4 weeks adequate to select the dose?
  • Low power to discern dose-viral load response
    thr 96wks. Selection of doses based on toxicity
    might be possible.
  • No, weeks 12-16 acceptable for preliminary
    assessment (pick dose for Phase III trial) and
    week 24 for confirmation based upon prior
    experience. Continue trial through week 48 for
    all doses.
  • In addition
  • Kaletra effect is so strong that it may be
    difficult to demonstrate Drug X dose-response
    in combination
  • Phase IIb trial was adequately designed to
    determine dose-response

21
EOP2a Meeting Metrics
  • Completed 5 over past year, 3 in progress
  • Therapeutic area problem
  • HIV new mechanism, dosing
  • Prostate Ca Formulation/dosing
  • Type 2 Diabetes Genotype, dosing
  • Anticonvulsant New mechanism
  • VMS (hot flashes) New mechanism, dosing
  • Pain receptor specificity/adrs, dosing
  • Weight-loss new mechanism, dosing
  • Workload 5-7 person-months/project
  • Sponsor evaluation (post-meeting) 4.1-4.3
    (1worthless, 5pivotal)

22
Sponsors Comments on the Experience (Abstracted
from their Senior RD Meeting slides)
  • FDAs Clinical Pharmacologists are very serious
    about leveraging Clin Pharm to
  • Aid selection of dosage regimens for Phases 2b
    3
  • Reduce attrition in Phase 3
  • Design better Phase 3 studies
  • FDA is inviting sponsors to participate for
    certain drugs
  • Drugs with reliable, quantitative measures of
    response and reasonable pk-pd models
  • Projects in early Phase 2
  • Particular interest in novel compounds
  • FDAs preparation was very extensive
  • Analyzed our exposure-response data
  • Applied pk-pd models (based on literature)
  • Had very detailed feedback
  • Our preparation must be extensive
  • Need high caliber Clin Pharm expertise
  • Be ready to submit datasets programs from all
    pk-pd studies - be ready for urgent queries
  • Work with our scientific team to prepare
    thoroughly for the meeting

23
IND/NDA Data Review AnalysisFDA Clinical
Pharmacology Work Plan
Data Visualization Data Set Creation
(I-Review)
Data Warehouse (PKS)
Clinical Trial Simulation (TS2)
End of Phase 2a Recommendation (2b/3 trial design)
24
GREEK
FRENCH
FARSI
Maori
DUTCH
Zulu
MANDARIN
VIETNAMESE
SWEDISH
DUTCH
Swahili
CANTONESE
  • UNITED NATIONS
  • WORLD BANK
  • INSEAD
  • NOVARTIS

MONGOLIAN
German
RUSSIAN
ARABIC
PORTUGESE
ENGLISH
ITALIAN
NORWEGIAN
TURKISH
Japanese
HINDI
ENGLISH
SPANISH
KOREAN
25
Ferring
GENENTECH
MERCK
GENZYME
AMGEN
GSK
ROCHE
Elan
NovoNordisk
JJ
FDA
Barr
Millenium
Daiichi
IDEC
Berlex
Eli Lilly
CDISC
AstraZeneca
Biogen
MYLAN
Clinical Data Interchange Standards Consortium
www.cdisc.org
MEDIMMUNE
ALLERGAN
PFIZER
ABBOTT
Sanofi
26
Drug-disease models at FDA
  • Primary sources literature, scientists, prior
    NDAs
  • Types
  • Mechanistic
  • Empirical
  • Diseases over past year
  • HIV
  • Diabetes Mellitus
  • Parkinsons Disease
  • Vasomotor Symptoms (Hot Flashes)
  • SLE-renal flare
  • Prostate Cancer- chemical castration
  • Kidney transplant rejection
  • In Development
  • Osteoporosis
  • Non-small cell lung cancer
  • Considering
  • How to share models some data on public
    website. Public dialogue on growing models

27
Diabetes
28
Modeling Results for FPG HBA1C Drug X in 1,000
patients
FPG
HbA1c
29
Save a Disease Prototype
  • Objective Quantitative library ? ? trial outcome
  • 3 levels clinical trial data, derived
    quantitative information, drug-disease models
  • Derived quantitative information (mean,
    variability, time-course) Create Standards
  • Disease, clinical trial, patient description
  • Biomarker/1 endpoint(s)
  • Placebo
  • Dropout rate
  • Drug (biomarker, 1 endpoint(s), adverse
    effects)
  • Covariates
  • 2 diseases chosen for prototype
  • Process Prototype ? small to larger groups
    (FDAgtSponsorsgtAcademics) Test value, set
    standards, fund

30
In Summary
  • End of Phase 2a Meetings pilot started
  • Program evaluation progression request within
    2 months
  • Draft guidance publication
  • Literature publication(s)
  • Model based drug development
  • Save a Disease prototype
  • Send an academic friend to FDA to create a
    disease model (funding available)
  • Software system for Clinical Pharmacology
  • CDISC standard needs to accelerate. This is a
    rate limiting step!
  • Your comments recommendations?
About PowerShow.com