Post-Marketing Studies from the Industry Perspective

1
Post-Marketing Studies from the Industry
Perspective

• Gretchen S. Dieck, Ph.D.
• Vice President, Risk Management Strategy

2
Key Risk Management Assumptions

• Each new drug is unique
• No drug is risk free
• Safety-related decisions must be evidence-based
• Industry is responsible for bringing
  evidence-based information forward
• No individual information source should be viewed
  in isolation
• Good communications with regulators and medical
  community are essential
• Ensure patient safety
• No surprises

3
Value of Risk Management Across the Lifecycle
Exposure (Potential Denominator)
Ph IV
Ph III
Ph II
Ph I
FIM
Product Life Cycle
Approval
Post Marketing Pharmacovigilance
Drug Discovery/Preclinical
Clinical Development

• Estimate potentialmarkets
• Predict potentialcandidates

• Initiate studies to understand treatment
  population, risks, concomitant diseases, etc.
• Promote faster approval
• Convert non-approvable to approvable
• Achieve appropriate label

• Understand absolute and relative risks
• Maintain patient safety

4
Tools for Evaluating Drug Safety Profile

• Clinical data
• Epidemiology
• Spontaneous reports

DecreasingScientific Rigor
Data from all sources are necessary to evaluate
a drugs safety profile
5
Risk Identification Pre- Post- Approval
Public expectation of risk knowledge at approval
is greater than reality
Risk Identification
1/100,000
Risk Identification
1/10,000
1/1000
Post Approval1 Million Patients
1/1000
1/500
1/100
Pre-Approval10,000 Patients
Reality
Public Expectation
How to Get There from Here
6
Building Risk Knowledge Base
Post-approval experience allows for
identification of smaller risks
SpontaneousReports
ObservationalStudies
Background Epidemiology
Other Activities
POTENTIAL SIGNALS OF RARE EVENTS
HYPOTHESIS TESTING
INCIDENCE OF COMMONLY OCCURRING EVENT FROM
CLINICALTRIALS
INCIDENCE OF EVENT IN GENERAL POPULATION IDENTIF
ICATION OF RISK FACTORS
1/1,000,000
1/100,000
1/500,000
1/10,000
1/1,000
1/50,000
1/100,000,
1/500
1/10,000
1/5,000
1/100
1/1,000
Risk Management/Assessment Strategy
Building knowledge through experience
7
Peri-Approval Risk Assessment
Risk Assessment Important part of Risk
Management

• Risk Management should start early in drug
  development
• Covers the entire drug life cycle
• Issues may arise during review period
• Assessment of risk
• Understanding the population being treated
• Obtaining information to put AEs into context

Important to identify subgroups at risk
8
Post-Approval Studies

• Classical epidemiological studies
• Cohort
• Case-control
• Case-crossover
• Large Simple Trial
• Registries
• Replicate findings using different study designs
  and populations

9
Post-Approval StudiesRisk Assessment
Example Geodon Large Simple Trial

• First launched September 2000 in Sweden and
  launched March 2001 in the United States
• Modestly prolongs the QTc interval - unknown
  whether modest QT prolongation results in an
  increased risk of serious cardiac events

10
Post Approval Safety Study Large Simple Trial

• Designed to study cardiovascular outcomes (sudden
  cardiac death and hospitalization due to
  cardiovascular events) in a real life setting
• A large, naturalistic, prospective study with
  random assignment of patients to antipsychotic
  treatment, to control for channeling bias
• 18,000 patients randomized to ziprasidone or
  olanzapine
• No additional study-required monitoring or tests
  after randomization
• Follow up during usual care

11
Why a Large Simple Trial?

• Strongest of observational study designs
• Random allocation eliminates possibility that
  patients are high risk of cardiovascular disease
  are preferentially prescribed ziprasidone, i.e.
  channeling bias
• Eliminates others forms of selection
  bias/confounding by indication
• Use of External Scientific Committees for highest
  scientific standards
• Scientific Steering Committee
• Safeguard interests of participating patients and
  act on recommendations of DSMB
• Scientific oversight of study conduct
• Data Safety Monitoring Board (DSMB)
• Biannual review of study endpoints and
  recruitment, discontinuation and retention rates
• Endpoint Committee
• Medical records reviewed by blinded reviewers
  using standard criteria - Reviewer consensus in
  cases of disagreement

12
Study Status

• Patient enrollment on projected schedule
• First patient enrolled February 2002
• gt than 6000 patients enrolled in US, Brazil and
  Sweden
• 360 sites enrolling patients in US
• More than 80 of sites are private practice
• Other sites are primarily mental health centers
  and Veterans Administration/psychiatric
  hospitals
• Further recruitment ongoing

13
Post-Approval StudiesUnderstanding about the
underlying population
Example Relpax

• Triptan for use in treating migraines
• Due to a vasoconstriction effect, there has been
  concern whether triptans can cause serious
  cardiovascular and/or cerebrovascular disease
• Lack of data on incidence of cardiovascular and
  cerebrovascular diseases, or mortality among
  migraine triptan users and non-triptan users
• Epidemiologic Studies
• Incidence of serious cardiovascular and
  cerebrovascular disease and mortality among
  migraine patients (triptan users and non-users)
• GPRD study (UK)
• UnitedHealthcare Research Database study (US)

14
Relpax - United HealthCare Study Study Design

• Retrospective cohort study 1995-99
• All patients with a diagnosis of migraine or a
  dispensing for a triptan
• Age-, sex-, and health plan-matched controls were
  randomly selected from individuals w/o migraine
  diagnosis and dispensing of triptans or ergot
  alkaloids

Non Migraineurs n130,411
Migraineurs n130,411
Non-triptan users n80,028
Triptan users n50,383
15
Relpax - United HealthCare Study Results
Rate per 1000 person-years. Rate ratio
comparing migraineurs to nonmigraineurs, adjusted
for age, gender, year of cohort entry,
comorbidities in year prior to study entry, oral
contraceptive use, and estrogen replacement
therapy use.
16
Relpax - United HealthCare Study Results
Rate per 1000 person-years. Rate ratio compared
to periods of non-use among migraineurs, adjusted
for age, gender, year of cohort entry, current
and recent use of ergot alkaloids, comorbidities
in year prior to study entry, oral contraceptive
use and estrogen replacement therapy use.
17
Post-Approval StudiesUnderstanding about the
underlying population
Conclusions

• Increased risk of IHD/unstable angina and
  stroke/TIA is observed in migraine patients.
  However, the increased risk is does not appear to
  be associated with the use of triptans
• The use of triptans is not associated with
  increased risk of acute MI, IHD/unstable angina,
  ventricular arrhythmias, stroke/TIA, all-cause
  mortality or cardiovascular mortality

18
Post-Approval StudiesPutting adverse events in
perspective - Geodon

• Epidemiology of schizophrenia
• Understanding underlying disease or condition
• Providing background rates for outcomes of
  interest
• Cardiovascular morbidity and mortality in
  schizophrenics
• Epidemiologic studies
• Saskachewan Health Databases
• United HealthCare Research Database
• Swedish National Board of Health and Human Welfare

19
Study Design
Geodon Saskatchewan Health Study

• Retrospective cohort using longitudinally
  collected data
• Individuals diagnosed with schizophrenia between
  1994 and 1995 (n3022)
• General population matched by age and sex (14)
  for comparison (n12,088)
• Prevalence study period (baseline risk factors)
  1994-1995
• Incidence follow-up period 1996-March 1999

Curkendall SM, Mo J, Stang MR, Jones JK, Glasser
DB. Cardiovascular disease in patients with
schizophrenia, Saskatchewan, Canada. Submitted to
The Journal of Clinical Psychiatry

20

Relative Risk of CV Disease and Diabetes
Schizophrenia Patients vs. General Population
1996-March 1999, Saskatchewan, Canada
RR 95 CI P-value Acute MI
0.86 0.61-1.22 NS Arrhythmias 1.16 0.96-1.40 NS
Vent. Arrhythmias 2.03 1.10-3.77 lt0.05 Syncope/
collapse 1.51 0.84-2.72 NS Stroke 1.34 1.01-1.7
7 0.05 TIA 0.86 0.58-1.26 NS Diabetes 1.62 1.1
9-2.20 lt0.01

21


Relative Risk of Mortality Schizophrenia
Patients vs. General Population 1996-March 1999,
Saskatchewan, Canada

RR 95 CI P-value All-Cause
2.69 2.31-3.13 lt0.0001 Suicide 7.00 2.55-19
.27 lt0.0005 Non-suicide 2.62 2.24-3.07 lt0.0001
Sudden Death 3.44 1.43-8.30 lt0.05 CV
Death 2.04 1.60-2.60 lt0.0001 Non-CV
Death 3.08 2.48-3.81 lt0.0001

22
Post-Approval StudiesPutting adverse events in
perspective - Geodon
Conclusions

• Schizophrenics have a significantly increased
  risk of cardiovascular morbidity and mortality
  compared to general population
• These findings were generally consistent across
  three geographically different study populations
• These factors need to be considered when
  evaluating serious cardiovascular AEs for Geodon

23
Registries Function of a Pregnancy Exposure
Registry

• Provide an estimate of an increased risk over
  background or to provide margins of reassurance
  regarding lack of risk
• Monitor ongoing postmarketing situation for
  suspected and emerging risks
• Identify factors that affect known risks
• Serve as hypothesis-generating tools
• Avert consequences of poor information
• Patients denied treatment
• Wanted pregnancies terminated

Overall goal to provide clinically meaningful
data to help prescribers and patients make
decisions about drug use during pregnancy
24
When is a Product a Good Candidate for a
Pregnancy Exposure Registry?

• Likely to be used during pregnancy as therapy for
  a new or chronic condition
• High likelihood of use by women of childbearing
  age
• Potential to cause harm during pregnancy, as
  identified from toxicology studies,
  structure-activity relationships, class effects,
  human case reports
• New product for a disease for which therapy has
  historically been teratogenic

25
Ideal Registry Design

• Actively collects information
• Study population exposed and unexposed groups
  with the same baseline risk, identified
  prospectively
• Exposure measurement dose, duration, and timing
  of all relevant medications
• Outcome measurement complete ascertainment of
  outcome in both groups
• Covariates complete information on all potential
  confounders and effect modifiers
• Follow-up complete follow-up for time period
  sufficient to observe outcomes
• Power ability to detect or rule out an increased
  risk of X over the exposed or observed in the
  unexposed group

26
Post-Approval StudiesChallenges

• Low recruitment of physicians and patients into
  study
• Carry out survey to identify potential barriers
• Viagra Changed questionnaire, went into
  additional countries
• Geodon LST Went into additional countries,
  added additional investigator sites
• Cant answer risk question
• Relpax What is risk of ischemic events in
  migraine population sample size needed made
  study not feasible. Proposed focused monitoring
  of ischemic AEs instead
• Viagra What is unique risk of CV events due to
  Viagra alone cant separate out effect of
  Viagra from risk with sexual activity. Upfront
  agreement with regulators over interpretation of
  study results

27
Post-Approval StudiesChallenges

• Risk questions cant be answered using
  observational methods
• Geodon what is risk of Qt prolongation.
  Suggested harder endpoints such as CV death and
  hospitalization
• Investigational drug what is risk of decreased
  pulmonary function cant measure lung function
  in a purely observational setting. Need formal
  clinical trial
• Other challenges
• Lack of information on specific data needed
  past medical history, OTC medication use, etc.
  May need to carry out several studies to get
  around limitation of one study design

28
Post-Approval StudiesConclusion

• Post-Approval studies are an important source of
  information to round out safety profile of a drug
• Manufacturers need to work closely with
  regulators on design and interpretation of key
  studies
• Risk assessment can be used to identify subgroups
  of patients at risk gives a more favorable
  benefit-risk balance
• Industry is interested in working to improve the
  tools available for post-marketing studies