Potential Conflicts of Interest

1
Potential Conflicts of Interest

• IREFNonprofit (NIH FDA prior Industry)
  Personal Pharmaceutical ConsultingNone
  Personal Pharmaceutical HonorariaNone Personal
  Pharmaceutical Speakers PanelNone Personal
  IREF Pharmaceutical StockNone IP /
  LicenseEPI I, II, Adenosine Regulating Agents
  (acadesine GP531)

2
PHARMPlacebo in Hypertension Adverse Reaction
Meta-analysis
3
PHARMPlacebo in Hypertension Adverse Reaction
Meta-analysis
A Collaborative Investigation of the Safety of
Placebo-Controlled Trials in Hypertension
FDA Division of Cardio-Renal Drug Products
IREF The Ischemia Research and Education
Foundation
4
Principal Investigators Raymond Lipicky, M.D.
(FDA) Dennis T. Mangano, Ph.D., M.D.
(IREF)Co-Investigators A. DeFelice / R.
Fenichel / J. Girton / S. Glasser / M. Gordan /
J. Hung / A. Karkowsky / J.
Lawrence / T. Sherpa / B. Stertz / S. Targum /
D. Throckmorton / J. Willard
Industry Contribution 93 NDAs (SNDAs) MOU
Collaborative Relationship / Data Access /
Publication / Data Sharing / IREF
875,000 Grant to PHARM (2 fulltime employees x
8.5 years) / 102,000 Internal
IREF Costs

COLLABORATIVE STRUCTURE
5

MINORITY PRESENTATION
Emphasis Literal Per-Protocol-Specified
Interpretation
Evidence Weight
? Primary Analysis (nearly 100 weight)--Basis
for conclusion ?
Secondary Analyses (minimal weight, but may
provide insight) Philosophy ? Neither
Placebo-control Orthodoxy, nor Active-control
Orthodoxy (Both
views discount the ethical and methodologic
complexities of clinical research.)
? Risk-averse When effective
therapies exist, there must be compelling
methodological reasons to conduct a
PCT. ? When an
effective therapy exists, The placebo-control
trial may be considered if and only if
placebo-treated patients are not be more
likely to (1) die, (2) suffer irreversible
morbidity, (3) suffer reversible but
serious harm, or (4) experience severe discomfort.
When effective therapy exists Placebo-control
Trials are unsafe until proven otherwise.
6

PROTOCOL
Specific Aim ? To determine the
relative risk of adverse clinical events among
patients receiving placebo versus
those receiving anti-hypertensive therapy.
? The relative risk will be determined for
three adverse event spheres
1. Overall Morbidity
2. Cardiovascular Morbidity
3. Neurologic Morbidity
? The adverse event spheres are prospectively
defined as Overall
Morbidity Cardiovascular Morbidity or
Neurologic Morbidity
Cardiovascular
Morbidity Angina AP or Arrhythmia AR or MI
or CHF Neurologic
Morbidity Stroke CVA or TIA or Hypertensive
Emergency HE
If the Relative Risk P/D is significantly
1.0, in any of the three spheres,
thenReassessment of the of placebo-controlled
trials of anti-hypertensive drugs is indicated.
Death also will be assessed independently
however, the power to discern difference (placebo
versus drug) is 7

Methods
As described by Dr. Lipicky.

8

Methods
As described by Dr. Lipicky.

Here, I will Focus on the Primary Outcomes
Overall Morbidity Cardiovascular Morbidity or
Neurologic Morbidity Cardiovascular Morbidity
Angina or Arrhythmia or MI or CHF Neurologic
Morbidity Stroke or TIA or Hypertensive
Emergency
9

Methods
As described by Dr. Lipicky.

Here, I will Focus on the Primary Outcomes
Overall Morbidity Cardiovascular Morbidity or
Neurologic Morbidity Cardiovascular Morbidity
Angina or Arrhythmia or MI or CHF Neurologic
Morbidity Stroke or TIA or Hypertensive
Emergency
And, on the Primary Aim / Hypothesis /
Study-Inference
If the Relative Risk P/D is significantly
1.0, in any of the three spheres,
then Reassessment of placebo-controlled trials
of anti-hypertensive drugs is indicated.
10

Results

OVERALL FINDINGS86,137 Patients
11
Consort Diagram of Study Patients
93 NDAs / SNDAs
12
Consort Diagram of Study Patients
93 NDAs / SNDAs
1973-2001
13
Consort Diagram of Study Patients
93 NDAs / SNDAs
1973-2001
540 RCTs
14
Consort Diagram of Study Patients
93 NDAs / SNDAs
1973-2001
540 RCTs
20 Companies
15
Consort Diagram of Study Patients
93 NDAs / SNDAs
1973-2001
540 RCTs
20 Companies
86,137 Patients enrolled
16
Consort Diagram of Study Patients
93 NDAs / SNDAs
1973-2001
540 RCTs
20 Companies
86,137 Patients enrolled
9,636 Patient Dropouts (11.1)
17
Consort Diagram of Study Patients
93 NDAs / SNDAs
1973-2001
540 RCTs
20 Companies
86,137 Patients enrolled
9,636 Patient Dropouts (11.1)
54 Years of Age 40 Women 30
Minority 159/102 BP dropout
18
Drop-Out Characteristics (9,363 Patients)
Sitting BP
Sitting
Supine BP
Women.............40 --------------------------
----------------- Caucasian.........70 AA
.22 Hispanic.....4 Other..4
19
Consort Diagram of Study Patients
93 NDAs / SNDAs
1973-2001
540 RCTs
20 Companies
86,137 Patients enrolled
54 Years of Age 40 Women 30 Minority 159/102 BP
dropout
9,636 Patient Dropouts (11.1)
PRIMARY AES Cardiovascular
Neurologic Angina (AP)99 0.11
TIA30 0.03 Arrhythmia (AR)82 0.10
Stroke (CVA)40 0.05 MI77 0.09
Hypertensive Emergency (HE)279
0.32 CHF44 0.05
Death43 0.01
20
Primary Outcomes Incidence(All Patients 86,137)
21
Consort Diagram of Study Patients
93 NDAs / SNDAs
1973-2001
540 RCTs
20 Companies
86,137 Patients enrolled
54 Years of Age 40 Women 30 Minority 159/102 BP
dropout
9,636 Patient Dropouts (11.1)
PRIMARY AES Cardiovascular
Neurologic Angina (AP)99 0.11
TIA30 0.03 Arrhythmia (AR)82 0.10
Stroke (CVA)40 0.05 MI77 0.09
Hypertensive Emergency (HE)279
0.32 CHF44 0.05
Death43 0.01
OTHER AES Other Cv (OC)489 5.4 VT8
0.00 Tx Failure (TF)2650 3.08 Other AE
(OAE)2734 3.17 Admin (OT)3081 3.58
22
Other Outcomes Incidence(All Patients 86,137)
23
Consort Diagram of Study Patients
93 NDAs / SNDAs
540 RCTs
64,438 Drug Patients 74.8
21,699 Placebo Patients 25.2
86,137 Patients enrolled
24
Consort Diagram of Study Patients
93 NDAs / SNDAs
540 RCTs
21,699 Placebo Patients (25.2)
64,438 Drug Patients (74.8)
86,137 Patients enrolled
3,056 Placebo Dropouts (14.1)
6,580 Drug Dropouts (10.2)
25
Primary Outcomes Incidence
Note Three Primary Outcomes (1) Overall
Morbidity Cardiovascular Morbidity or
Neurologic Morbidity (2) Cardiovascular
Morbidity Angina or Arrhythmia or MI or
CHF (3) Neurologic Morbidity Stroke or
TIA or Hypertensive Emergency
26
Primary Outcomes Incidence
27
Primary Outcomes Incidence
28
Primary Outcomes Relative Risk
29
Primary Outcomes Relative Risk
30

Conclusions

31
Primary Conclusion
Overall Morbidity Cardiovascular Morbidity or
Neurologic Morbidity Cardiovascular Morbidity
Angina or Arrhythmia or MI or CHF Neurologic
Morbidity Stroke or TIA or Hypertensive
Emergency
If the Relative Risk P/D is significantly
1.0, in any of the three spheres,
then Reassessment of placebo-controlled trials
of anti-hypertensive drugs is indicated.
32
Primary Conclusion
The Relative Risk P/D is significantly 1.0,
in two of the three spheres, therefore Reasses
sment of placebo-controlled trials of
anti-hypertensive drugs is indicated.
33

A Secondary Consideration

Counter-argument
34
(No Transcript)
35
EQUIPOISE
36
EQUIPOISE
37
EQUIPOISE
38
EQUIPOISE
39
EQUIPOISE
40
EQUIPOISE
41
EQUIPOISE ?????
?
42


HE OC BalanceDrug HE-OC Events 145417
526 / 64,438 0.82 Placebo HE-OC Events
13452 186 / 21,699 0.86
43
EQUIPOISE
44


Conclusion Equipoise
45


Conclusion EquipoiseBut 1..
46
PRIMARY ENDPOINT COMPONENT
47
Protocol-Excluded EP Components
Primary Endpoint Components
48


Conclusion EquipoiseBut 2..
49
ARITHMETIC EQUIPOISE
HE-OC Events Drug 0.82 (526 / 64,438)
Placebo 0.86 (186 / 21.6990
50


But
51
But, are they Comparable???
52


ButAre they comparable?
53


ButAre they comparable?Certainly not
prospectively.
54


ButAre they comparable?Certainly not
prospectively. Even putting that aside,
their severities may be different.
55
HE and OC
56
HE and OC
57
HE and OC
58
BACKGROUND

• Clearly, not all clinical events carry like
  severity, even when classified as adverse events,
  or serious adverse events (with WHO definitions).
  
• PHARMA analysis suggested that two
  discontinuation-related events distinguished the
  study groups (1) hypertensive emergencies and
  (2) other cardiovascular events.
• Regarding the formerHypertensive Emergencythe
  gravity of the events appears obvious however,
  subjectivity in the interpretation of a
  "hypertensive emergency" may exist and
  therefore, even these events should be subjected
  to a severity grading.
• Regarding the latterOther Cardiovascular
  Eventsit is clear that a broad spectrum of
  conditions may exist for this category, and that
  in certain circumstances, Other Cardiovascular
  Events may be viewed as a "default category."
• Therefore, a grading system was constructed and
  applied blindly.

59
Hypertensive Emergencies HE279 patients
discontinued for HE. Patient identifiers study
group were redacted, then each HE event was
graded from 0 to 10 0No emergency and No
hypertension 1-4No hypertension
5Hypertension (200/100) No s/s
brain/eye/heart/kidney 6Hypertension
(200/100) 1 Organ s/s brain/eye/heart/kidney
7Hypertension (200/100) 2 Organs s/s
brain/eye/heart/kidney 8Hypertension
(200/100) 3 Organ s/s brain/eye/heart/kidney
9Hypertension (200/100) 4 Organ s/s
brain/eye/heart/kidney 10Hypertension
(200/100) Death
Hypertensive Emergency Severity Assessment
60
Other Cardiovascular Events OCE 469
patients discontinued for OCE. Patient
identifiers study group were redacted, then
each OCE event was graded from 0 to 10
0No event 1No ischemia of any
organ 2-5Minor symptoms
6Myocardial Ischemia, Stupor, Loss of
Consciousness 7Documented two-organ
ischemia 8Documented MI or Stroke
9Documented MI or Stroke and second
organ ischemia 10Death.
Other Cv Events Severity Assessment
61
Hypertensive Emergency Events
Drug
Placebo
62
Hypertensive Emergency Events
Drug
Placebo
63
Other Cardiovascular Events
Drug
Placebo
64
Other Cardiovascular Events
Drug
Placebo
65
Conclusions HE versus OC
66
Conclusions HE versus OC

• The arithmetic balance between OC and HE
  events argues for equipoise.

67
Conclusions HE versus OC

• The arithmetic balance between OC and HE
  events argues for equipoise.
• However, two arguments exist for discounting
  that inference.
• 1. The arithmetic balance analysis is
  a secondary construct.

68
Conclusions HE versus OC

• The arithmetic balance between OC and HE
  events argues for equipoise.
• However, two arguments exist for discounting
  that inference.
• 1. The arithmetic balance analysis is
  a secondary construct.
• 2. Given that severe events likely
  have greater short and long-term
• physical consequence, the findings for both
  severe HE and severe OC
• support the hypothesis that harm
  is associated with PCT for HTN.

69
Conclusions HE versus OC

• The arithmetic balance between OC and HE
  events argues for equipoise.
• However, two arguments exist for discounting
  that inference.
• 1. The arithmetic balance analysis is
  a secondary construct.
• 2. Given that severe events likely
  have greater short and long-term
• physical consequence, the findings for both
  severe HE and severe OC
• support the hypothesis that harm
  is associated with PCT for HTN.
• Given that the primary hypothesis was
  satisfied, namely that RR1 for
• 2/3 spheres (neurologic morbidity and all
  morbidity), then I still conclude
• 
  that

70
Conclusions HE versus OC

• The arithmetic balance between OC and HE
  events argues for equipoise.
• However, two arguments exist for discounting
  that inference.
• 1. The arithmetic balance analysis is
  a secondary construct.
• 2. Given that severe events likely
  have greater short and long-term
• physical consequence, the findings for both
  severe HE and severe OC
• support the hypothesis that harm
  is associated with PCT for HTN.
• Given that the primary hypothesis was
  satisfied, namely that RR1 for
• 2/3 spheres (neurologic morbidity and all
  morbidity), then I still conclude
• 
  that

Reassessment of placebo-controlled trials of
anti-hypertensive drugs is indicated.
71

• Thank You

72

My Inferences forClinical Trial Design

73

CONTROVERSY
Considerations 1 ? Declaration
of Helsinki The benefits, risks burdens, and
effectiveness of a new method should be used
against those of the best current prophylactic,
diagnostic, and therapeutic methods. This does
not exclude the use of placebo, or no treatment,
in studies where no proven prophylactic,
diagnostic, and therapeutic method exists.
? International Conference for
Harmonization Considers use of PCT ethical,
even if effective treatment is available for the
condition under study, if withholding the
treatment leads to no serious harm and if
patients are fully informed about available
therapies and consequences of no or delayed
treatment.
74

CONTROVERSY
Considerations 2 Placebo OrthodoxyPCT
are appropriate, even when an effective therapy
exists. No therapy should be
approved unless it is clearly superior to
placebo / no therapy.
Strengths a. Methodologically rigorous.
b. New therapy standard
therapy may still be clinically valuable.
Limitations a. Criteria for ethical
use of PCT are not always clearly stated. b.
Focus on irreversible physical harm.
(But, temporary,
reversible conditions ? permanent harm) c.
Permits intolerable suffering.
75

CONTROVERSY
Considerations 2
Active-Control OrthodoxyPCT are not
appropriate, when an effective therapy
exists. PCTs are inappropriate ? the
clinically relevant question is not whether
a new therapy is better than nothing, but
whether it is better than standard treatment.
Strengths a.
Methodologically rigorous. b.
New therapy standard therapy may still be
clinically valuable.
Limitations a. Harm/discomfort associated with
placebo-assignment may be non- existent/
trivial. b. Power of the placebo response is
underestimated (?attention). c.
Active-control harm may placebo-control harm
(?saes equivalence ?.
76

CONTROVERSY
Considerations 3 ? PCTs clearly
unethical if withholding therapy?death,
life-threat, serious harm. Ex. PCT for TPA
was not ethical (SK was effective). ? PCTs
clearly ethical if withholding therapy?minimal
chance of harm/suffering.
Ex. Rogain / allergic rhinitis. ? PCT
Orthodoxy Active-control Orthodoxy
Both views discount the ethical and methodologic
complexities of clinical research. ? PCTs
clearly ethical if withholding therapy?minimal
chance of harm/suffering.
Ex. Rogain / allergic
rhinitis.
77

CONTROVERSY
Considerations 4 Principle When
effective therapies exist, there must be
compelling methodological reasons to
conduct a PCT. Methodological Criteria
? High placebo response
rate. ? Conditions wax
wane. ? Existing Txs are
only partially effective, or have serious
side-effects. ? Low
frequency of condition (Equivalence)Ethical
Criteria ? Placebo
patients should not be more likely to
die
suffer irreversible
morbidity
suffer reversible but serious harm
experience severe
discomfort