TROGLITAZONE: PRESENTATION TO ADVISORY COMMITTEE

1
TROGLITAZONEPRESENTATION TO ADVISORY COMMITTEE

• Murray M. Lumpkin, M.D.
• Deputy Center Director
• Center for Drug Evaluation Research
• US Food and Drug Administration
• 19 May 2000

2
Background

• Troglitazone - (Rezulin)
• Approved January 29, 1997
• Marketing in US began March 97
• Adv Cmt (12/11/96) 8-0 vote for AP
• Priority review (first cycle) AP
• Total time to approval around 6 mo.
• 7 mo. later first reports of LVR death
• Several Dear Dr letters relabeled
• Adv Cmt (3/26/99) 11-1 BR for concomitant use
  with insulin 12-0 BR for concomitant use with
  sulfonylureas 3-9 BR for monotherapy addition
  ideas for labeling, study, etc

3
Background

• Rosiglitazone (Avandia)
• AP in May 25, 1999
• 6 mo P review
• Pioglitazone (Actos)
• AP in July 15, 1999
• 6 mo P review

4
CDER Definition of Liver Failure

• Hepatic insufficiency that results in
• Death
• Transplant
• Recovery after encephalopathy or on transplant
  list
• Coagulopathy not sufficient alone
• None of the 3 products had such pts in their NDA
  databases

5
Troglitazone Analysis of AERS Reports March
2000

• Liver failures (n90)
• 36 months of marketing
• 60 deaths without transplant
• 3 deaths with transplant
• 7 alive with transplant
• 10 recovered without transplant
• 10 with unknown outcomes
• probably deaths (but cant confirm)
• Thought at least possibly or probably related to
  troglitazone at time of market WD

6
Rosiglitazone Analysis of AERS ReportsMarch
2000

• Liver failure (n2)
• 10 months of marketing
• 1 death without transplant
• 1 recovered
• Thought at least possibly or probably related to
  rosiglitazone at time of troglitazone WD

7
PioglitazoneAnalysis of AERS ReportsMarch 2000

• Liver failure (n0)
• 8 months of marketing
• At time of troglitazone WD

8
Questions

• Do the benefits of troglitazone no longer
  outweigh its risks given the availability of
  rosiglitazone and pioglitazone?
• Is troglitazone less safe than rosiglitazone or
  pioglitazone?
• Does troglitazone have any unique efficacy or
  niche population(s)?

9
Possible Comparisons

• No head-to-head data available
• Launch vs launch
• date of disease onset
• all start therapy within 9 mo. of launch
• Contemporaneous
• date of disease onset
• all start therapy after June 99 launch of
  rosiglitazone)
• Entire exposure vs entire exposure
• date of disease onset in last 9 months
• start therapy anytime after launch

10
Contemporaneous of cases based on date of
onset of sADR onset of rx from June, 1999
Entire Exposure of cases based on date of
onset of sADR onset of rx from March, 1997
months
0
9
Rezulin (6/99 thru 2/00)
of cases based on date of onset of sADR onset
of rx from launch date
of cases based on date of onset of sADR onset
of rx from launch date
months
0
9
0
months
Initial launch
9
Initial launch
Avandia
Rezulin
(6/99 thru 2/00)
(3/97 thru 11/97)
11
Data Limitations

• Spontaneous data
• Not clinical trials data
• Voluntary reporting
• Reporter motivation
• New drug biases
• No real denominator/numerator
• Under-reporting/magnitude?
• No control groups
• Confounding factors
• Veracity of causal attribution?

12
Spontaneous Reports

• Great signal generator/detector
• Rare, serious, unexpected sADRs
• Poor at determining the depth/strength of the
  signal
• Situational - especially poor with high
  background rate of occurrence
• Can NOT do incidence calculations
• Can only do reporting rates
• Soft data at best, but all we often really have

13
COMPARISON 1LAUNCH vs LAUNCH
14
Contemporaneous of cases based on date of
onset of sADR onset of rx from June, 1999
Entire Exposure of cases based on date of
onset of sADR onset of rx from March, 1997
months
0
9
Rezulin (6/99 thru 2/00)
of cases based on date of onset of sADR onset
of rx from launch date
of cases based on date of onset of sADR onset
of rx from launch date
months
0
9
0
months
Initial launch
9
Initial launch
Avandia
Rezulin
(6/99 thru 2/00)
(3/97 thru 11/97)
15
Launch vs LaunchTotal US Prescriptions

• Total US prescriptions for Rezulin Avandia
  (NPA)/ PER MO.
• Rxs in thousands
• First 3 quarters of marketing

prescriptions
Rezulin Avandia
months
16
TroglitazoneFirst 5 Qtrs experience LVR
failure onset (cumulative)

• 1st Qtr 0 cases 421,000 Rxs
• 2nd Qtr 2 cases 1,228,000 Rxs
• 3rd Qtr 16 cases 2,402,000 Rxs
• 4th Qtr 24 cases 3,586,000 Rxs
• 5th Qtr 40 cases 5,000,000 Rxs

(7 mo 5)
prob represents 800K pts
12 (onset reporting thru 3rd qtr)
17
RosiglitazoneFirst 3 Qtrs experience LVR
failure onset (cumulative)

• 1st Qtr 1 case 369,000 Rxs
• 2nd Qtr 1 case 769,000 Rxs
• 3rd Qtr 2 cases 1,563,000 Rxs

7 mo 2
(prob represents 500,000 pts
Questionsall reported for 3rd qtr? Is level of
underreporting similar?
18
Troglitazone vs RosiglitazoneLiver Failure
Experience from Launch(Case onset by Quarter)
(cumulative number of cases reported)
19
End of 3rd QTR comparison

• TROGLITAZONE
• 16/800,000 pts
• 1/50,000
• 12/800,000 pts
• 1/67,000

• ROSIGLITAZONE
• 2/500,000 pts
• 1/250,000

20
Arguments AGAINSTLaunch/Launch Comparison

• Dynamic in the community possibly quite different
  in 1997 and 1999
• People possibly now more educated, more aware of
  possibility of hepatotoxicity, and possibly
  acting accordingly
• ? Monitoring helping
• ? More appropriate patient selection

21
Comparison 2Contemporaneous
22
Contemporaneous of cases based on date of
onset of sADR onset of rx from June, 1999
Entire Exposure of cases based on date of
onset of sADR onset of rx from March, 1997
months
0
9
Rezulin (6/99 thru 2/00)
of cases based on date of onset of sADR onset
of rx from launch date
of cases based on date of onset of sADR onset
of rx from launch date
months
0
9
0
months
Initial launch
9
Initial launch
Avandia
Rezulin
(6/99 thru 2/00)
(3/97 thru 11/97)
23
TroglitazoneCase onset in those starting therapy
since 6/99

• N0
• Around 150,000 new starts (patients)
• Aware of one that comes as close as possible
  without actually making the definition. Also
  working on one that might ultimately make
  definition.

24
RosiglitazoneCase onset in those starting
therapy since 6/99

• N2
• 500,000 new starts (patients)

25
Arguments FOR Contemporaneous comparison

• Contend little underreporting of troglitazone
  because of increased publicity
• Physicians now monitoring appropriately
• Physicians better selecting patients for
  glitazone therapy
• People now stop therapy earlier when s/sx of
  liver tox begin

26
Arguments AGAINSTContemporaneous Comparison

• New drug reporting effect
• Zenith - rosiglitazone Nadir - troglitazone
• Liability concerns with reporting
• People more likely to report rosiglitazone
• Not convinced monitoring helps or even being done
  by physicians
• Not convinced one can identify or select
  appropriate patients

27
Arguments AGAINSTContemporaneous Comparison
(contd)

• Often patients quite ill within 30 days from
  having normal liver enzymes drawn
• 9/12 showed rapid rise within 30 days
• Jaundice is often the first evidence of liver
  involvement in most of the cases

28
Comparison 3Total Experience vsTotal
Experience
29
Contemporaneous of cases based on date of
onset of sADR onset of rx from June, 1999
Entire Exposure of cases based on date of
onset of sADR onset of rx from March, 1997
months
0
9
Rezulin (6/99 thru 2/00)
of cases based on date of onset of sADR onset
of rx from launch date
of cases based on date of onset of sADR onset
of rx from launch date
months
0
9
0
months
Initial launch
9
Initial launch
Avandia
Rezulin
(6/99 thru 2/00)
(3/97 thru 11/97)
30
Case Onset since 6/99 (troglitazone could have
been started anytime)

• Troglitazone n11
• Rosiglitazone n2
• no deaths since June 99

31

• OTHER DATA/PRIORS
• Clinical Trials Experience
• AERS Hepatitis/Jaundice
• UHC Epidemiology Study
• All other Population- based Studies
• DPP Study at NIH

32
Other Data/PriorsClinical Trials Experience

• Clinical trials database (pre-approval) no
  liver-related deaths or transplants in any
  application
• Overall hepatotoxicity picture better with
  rosiglitazone and pioglitazone
• without cases of jaundice and less liver enzyme
  elevation

33
Clinical Trials Database Troglitazone

• 2 patients treated with troglitazone developed
  reversible jaundice
• 1 had liver Bx consistent with idiosyncratic drug
  reaction
• 1 other patient treated with troglitazone also
  had a liver bx consistent with idiosyncratic drug
  reaction
• Info in original approved label

34
ELEVATED ALT NDAs
35
Other Data/PriorsHepatitis/jaundice AERS
reports

• Post-marketing Cases of hepatitis or jaundice
  that didnt meet previous definition of LVR
  failure, but did result in hospitalization or
  were considered medically significant or
  life-threatening

36
Other Data/PriorsHepatitis/jaundice AERS
reports

• Troglitazone
• n150 (75/1 million Rxs)
• Rosiglitazone
• n25 (16/1 million Rxs)
• Entire database from launch of product through
  March 2000

37
Other Data/PriorsUHC Epidemiology Study

• Cohort study in UHC database
• N 9369 patients
• Avg use around 6 months
• 4873 person-years of exposure
• 1 case of fatal hepatotoxicity
• Rate 205/ 1 million patient- years

38
Other Data/PriorsOther population-based
studies

• All PD studies (March Adv Cmt)
• N15,006 patients
• Most treated for 1-3 months
• 6227 person-years of exposure
• 1 case of fatal hepatotoxicity
• Rate 161/1 million person years

39
Other Data/PriorsDPP Study at NIH

• Prevention study in patients thought to be at
  higher risk to develop Type 2 diabetes not Type
  2 diabetics yet
• N 585 patients
• Most treated for 12 months
• 1 case of fatal hepatotoxicity
• Rate 1724 / 1 million person years

40
CONCLUSIONS March 2000

• Incidence of troglitazone-induced liver failure
  possibly around
• 18,000 - 20,000 (acute)
• 12000 (chronic) modeling
• Clearly a bracket with reasonable arguments that
  it could be larger and smaller

41
CONCLUSIONS March 2000

• Efficacy -- no consistent evidence that has
  demonstrated that the glitazones are not
  equivalent in efficacy (no special niches)

42
CONCLUSIONS March 2000

• Other safety concerns with other products fluid
  retention and CHF with rosiglitazone -
• Does not appear worse than with troglitazone
  (post-marketing reporting)

43
CONCLUSIONS March 2000

• Rosiglitazone and pioglitazone appear to have a
  better safety profile than troglitazone and offer
  patients same efficacy benefits
• Troglitazone should come off US market (outmoded
  drug)

44
OUTMODED DRUGS

• Benefits outweighed risks at time of initial
  approval but with newer drugs with better safety
  profiles, that perspective changes
• will be seeing more and more
• another example is terfenadine
• not a failure of the system
• would be a failure not to take off once convinced
  that a safer alternative available