Gastrointestinal Review Highlights of the CLASS Study

1
Gastrointestinal Review Highlights of the CLASS
Study

• Lawrence Goldkind M.D.

2
Outline of Gastrointestinal Review of CLASS study

• Review study design highlights
• Review study results
• 1. Primary analysis Complicated Ulcer
  (CSUGIE)
• a. ITT
• b. Subgroup analyses aspirin
  and non-aspirin
• 2. Composite endpoint- Symptomatic as
  well as
• Complicated ulcers
  (SX/Complicated)
• a. ITT
• b. Subgroup analyses of aspirin
  and non-aspirin

3
Outline of Gastrointestinal Review of CLASS
study(continued)

• 3. High risk populations
• Conclusions

4
Design Highlights

• The null hypothesis being tested is that there
  is no difference in the incidence of clinically
  significant UGI events between Celebrex and each
  of the NSAID groups (ibuprofen and diclofenac)
  protocol November 1998
• Complicated Ulcer (CSUGIEs)

5
Statistical Plan

• Two primary treatment comparisons will be
  performed celecoxib vs. ibuprofen and celecoxib
  vs. diclofenac.
• A stepwise procedure will be used to strongly
  control type-1 error. In this procedure, the
  first step is to test the overall hypothesis
  whether celecoxib and the pooled NSAIDs are
  different.

6
Statistical plan (continued)

• If the test is not significant, the null
  hypothesis is retained and the procedure stops.
• If the test is significant, the second step will
  be the pairwise tests between celecoxib and each
  of the two NSAIDs.
• (As long as the above conditions were met no
  alpha adjustment was considered necessary )

7
Endpoint DefinitionComplicated Ulcer

• Perforation
• Obstruction
• Upper Gastrointestinal Bleeding
• Traditionaldefinition of UGI bleed clear
  evidence of some blood loss with evidence of
  gastroduodenal injury
• Alternate evidence of imminently
  life-threatening bleed with evidence of
  gastroduodenal injury (transfusion, orthostasis,
  2g/dl drop in Hgb

8
Traditional Definition

• Documented Gastroduodenal ulcer or erosion in
  addition to one of the following
• 1. Hematemesis
• 2. Active bleeding at time of endoscopy or blood
  within the stomach at endoscopy
• 3. Stigmata of recent bleed (adherent clot or
  visible vessel)

9
Traditional definition of bleeding (continued)

• 4. Melena
• 5. Hemocult positive stool fall in Hct of gt5 or
  Hgb of gt 1 g/dl
• 6. Hemocult positive stool orthostasis
• 7. Hemocult positive stool need for transfusion
  on clinical grounds

10
Design highlights (continued)

• Dose selection 2X for RA 800mg
  
• Proof of COX-2 hypothesis Dose dependency of
  GI safety
• Dose creep potential phenomenon in painful
  chronic conditions (open-label studies in
  original NDA suggested majority of patients
  increase dose when allowed)

11
Design highlights (continued)

• Margin of overall safety when organ specific
  major safety claim being considered ( if
  overall safety not maintained with higher dose ?
  value of organ specific findings)
• 800 mg/day is 1X for chronic use in FAP
• Future indications and doses are unknown

12
Design Highlights(continued)

• Generalizability
• Population OA and RA
• 2 comparators
  
• Minimal exclusions
• a. renal or hepatic dysfunction or lab
  abnormality considered to be clinically
  significant by the investigator
• b. baseline occult GI bleed
• c. aspirin allowed ( patients with underlying
  significant CV disease included)

13
Design Highlights

• Study Duration
• The trial will continue until the anticipated
  number of clinically significant UGI events have
  been observed in both studies. Minimum
  participation for an individual patient is 26
  weeks and maximum study participation is 52 weeks

14
Design Highlights Conclusions

• Chronic exposure to assess chronic safety
• High dose to assess robustness of any safety
  claim
• Multiple comparators to address generalizability
• Rigorous and well-defined endpoints
• Large trial size allowed comparative data on
  overall safety including uncommon toxicities

15
Results
16
Primary endpoint Complicated Ulcer ITT
17
Time to Complicated Ulcer (CSUGIE )

• Traditional definition
• Intention to treat population

18
(No Transcript)
19
Subanalyses Complicated Ulcers Non-ASA and ASA
users
20
Non-ASA
21
Time to Complicated Ulcer (CSUGIE)

• Traditional definition
• Non-aspirin users

22
(No Transcript)
23
Complicated Ulcer ASA

24
Summary of Findings for Complicated Ulcers

• Primary analysis No differences between Celebrex
  and NSAIDs combined or individually
• Non-ASA Strong trend favoring Celebrex compared
  to ibuprofen. No difference between Celebrex and
  diclofenac
• ASA No differences between Celebrex and
  diclofenac. Paradoxical trend favoring ibuprofen
  compared to both Celebrex and diclofenac (smaller
  sample size, study not stratified based on ASA
  use)

25
Other Relevant AnalysesComposite
Endpoint(SX/Complicated)

• Original protocol
• Symptomatic UGI ulcers, documented by endoscopy
  or UGI barium x-ray with no evidence of
  perforation, bleeding or obstruction will be
  categorized and summarized separately.
• Composite endpoint of SX/complicated ulcers not
  pre-specified

26
Composite Endpoint(SX/Complicated)

• Clinically relevant endpoint
• Pre-specified ascertainment of data

27
Composite EndpointSX/Complicated UlcerITT

• Celebrex diclofenac
  ibuprofen
• of events 43/3987 26 /1996
  36/1985
• Cumulative 1.95 1.91 2.84
• rate
  (ns) (p0.017 uncorrected)
• No./100 pt yr. 1.85 2.41
  3.21
• log-rank test

28
Time to Composite SX/Complicated Ulcer

• Intention to treat population

29
(No Transcript)
30
CompositeSX/Complicated UlcerNon-ASA

• Celebrex diclofenac
  ibuprofen
• of events 21/3105 10/1551
  28/1573
• Cumulative 1.13 0.92
  3.00
• rate
  (p0.3) (p lt0.001 uncorrected)
• No./100 pt yr. 1.16 1.19
  3.20
• log-rank test

31
Time to Composite SX/Complicated Ulcer

• Non-ASA

32
(No Transcript)
33
Composite SX/Complicated UlcerASA

• Celebrex diclofenac
  ibuprofen
• of events 22/882
  16/445 8/412
• cumulative 4.94 5.31
  3.33
• rate
  (p0.15ns) (p0.46)
• 
  (uncorrected)
• log-rank test

34
Conclusions of Composite SX/Complicated Ulcer

• Pre-specified ascertainment of data but not a
  pre-specified endpoint
• Clinically relevant endpoint
• Strong trend in favor of Celebrex compared to
  ibuprofen in non-ASA users
• No difference between Celebrex and diclofenac in
  non-ASA users

35
Conclusions of Composite SX/Complicated Ulcer
(continued)

• In ASA users there was a paradoxical trend
  favoring ibuprofen compared to both Celebrex and
  diclofenac. Similar to pattern seen in primary
  endpoint complicated ulcer

36
Alternate Definition of Bleed/Major Bleed

• Hematemesis, melena or Hemocult positive stool in
  the face of a gastroduodenal ulcer or erosion
  plus
• 1. drop in Hgb of gt 2g/dl with adequate hydration
  or if urgent transfusion required, final
  hemoglobin after equilibration of lt pre-bleed
  level
• OR
• 2. Orthostatic hypotension or supine BP under
  90/60

37
Complicated Ulcer Using Pre-specified
Alternate more serious definitions of
Bleed/Major Bleed

• Celebrex diclofenac
  ibuprofen
• events 17/3987
  5/1996 9/1985
• Cumulative 0.68 0.35
  0.61
• rate
  (ns) (ns)

38
High risk populations Relative RiskComplicated
ulcer

• Age gt 75 Hx of UGI
  bleed ASA
• Celebrex 5 3.6
  4.0
• NSAID 5.8 7.1
  1.8
• (comparators)

39
High risk populations Relative Risk Composite
SX/Complicated ulcer

• Age gt 75 Hx of UGI bleed
  ASA use
• Celebrex 3.5 4.3
  3.7
• NSAID 3.7 3.4
  2.3
• (comparators)

40
High Risk Populations

• If age and history of ulcer complication are
  independent risk factors for ulcer
  disease...Findings of high risk in association
  with a therapy may represent intrinsic risk
  rather than drug effect (no causality) ?
• OR
• Interaction between underlying and drug related
  risk may produce an exaggerated/ higher risk
  attributable to therapy (causality) ?

41
Overall Conclusions

• No statistically significant differences were
  shown for the entire population for the primary
  endpoint of complicated ulcer between Celebrex
  and the NSAID comparators- combined or
  individually
• Relevant endpoint of the composite of
  SX/Complicated Ulcers suggested a difference
  between Celebrex and ibuprofen in favor of
  Celebrex. No difference was seen between Celebrex
  and diclofenac

42
Overall Conclusions(continued)Hypothesis
Generating Findings

• Co-administration of aspirin was associated with
  an increased and similar risk of complicated
  ulcers in both Celebrex and diclofenac groups
  ( _at_ 4-fold)
• The same trend was seen at the broader Composite
  endpoint of SX/Complicated Ulcer
• .

43
Overall Conclusions (continued)Hypothesis
Generating Findings

• The ibuprofen group requiring low dose aspirin
  experienced a lower rate of complicated ulcers
  than either of the other two groups. This trend
  was also seen in the composite endpoint of
  SX/complicated ulcer

44
Overall Conclusions(continued)

• It is unclear whether the paradoxical findings
  associated with the concomitant use of aspirin
  and ibuprofen represent random findings or
  whether they represent a true differential
  interaction between aspirin and NSAIDs in terms
  of UGI toxicity.

45
Overall Conclusions

• Further study is needed to clarify the safety of
  co-administration of aspirin and NSAIDs/COX-2
  selective agents
• No conclusions regarding safety of Celebrex
  compared to traditional less selective COX
  inhibitors as a group are possible

46
(No Transcript)