1 PHARM Placebo in Hypertension Adverse Reaction Meta-analysis Principals at FDA Albert DeFelice Jim...

Post on 21-Jan-2016

212 views 0 download

Tags:

Transcript of 1 PHARM Placebo in Hypertension Adverse Reaction Meta-analysis Principals at FDA Albert DeFelice Jim...

1

PHARM

Placebo in Hypertension Adverse Reaction Meta-analysis

Principals at FDAAlbert DeFeliceJim WillardJames HungJohn Lawrence

Principals at IREFDennis Mangano

A Collaborative Effort Effort Between:The Division of Cardio-Renal Drug Products

IREF (Ischemia Research and Education Foundation) And

Stephen Glasser

All Companies were formally contactedAnd

Formally gave permission to use the dataAnd

Agreed to retrieve & supply missing CRFs

2

28 Years of Placebo-ControlledAntihypertensive Development Trials

540 Individual Protocols86,137 Randomized Patients (21,699 P, 64,438 D)

12,657 Patient Years (3,221.5 P, 9,436 D)42 Chemical Entities, 6 Drug Classes

A Blinded (from original CRFs)Meta-Analysis of (Deaths & Dropouts), by study :

9,636 Dropouts (3,056 P, 6,580 D)Mean age = 54.1 yearsMean Sitting Blood Pressure - 157.6/102.4 mm HG

Singular Primary End Point, RR (P/D) of Dropping Out

RR = 1.33 (1.28, 1.39; p<10-15)RR (Mortality, Stroke, MI) = 1.03 (0.71, 1.47; p = 0.86)

3

31.9%

30.4%

33.3%

4.4%Administrative

Lack of Blood Pressure Control

Adverse Effects

Angina, MI, CHF,Stroke, TIA, Death

4

Administrative OT = OTher than medical Moved, Surgery, non-compliance, etc.

Lack of Blood Pressure Control TF = Therapeutic Failure Investigator or Patient Judgment

HE = Hypertensive Emergency Diastolic BP > 110 mm Hg, increase by >10 mm Hg and/or evidence of new end organ involvement

Adverse Effects OC = Other Cardiac adverse events Angioedema, edema, low blood pressure, non-specific EKG changes

OAE = Non-cardiac, Other Adverse events Laboratory abnormalities, headache, nausea/vomiting

5

All Cause Mortality Death

Non-Fatal Myocardial Infarction MI

Non-Fatal Stroke CVA

Transient Ischemic Attack TIA

Congestive Heart Failure CHF

Angina Pectoris AP

Non-Fatal Arrhythmia AR

Unscheduled Visits ER Hospitalization

6

Administrative DropOuts (OT)3,081 Events (840 P, 2,241 D)*RR (placebo/Drug) = 1.09 (1.01, 1.18, p =0.031)Not a primary, 1 of many comparisons

*= Maximum Likelihood Estimation

Therapeutic Failure (TF)2,650 Events (1,266 P, 1,384 D)*RR = 2.53 (2.35, 2.73) p<10-15)Not a primary, 1 of many comparisons

Together These Two CategoriesAccount for 5,731 EventsOr 62.3% of All Drop Outs

7

TF Placebo TF Drug OT Placebo OT Drug0.00%

0.50%

1.00%

1.50%

2.00%

2.50%

3.00%

3.50%

4.00%

4.50%

5.00%

Percent of DropOut Group with ER/Hosp

ER

Hosp

22 Events Total

116 Events Total

Unscheduled Visits

8

Other Adverse Events (OAE)2,734 Events (653 P, 2,081 D)*RR (placebo/Drug) = 0.87 (0.79, 0.95, p =0.0017)Not a primary, 1 of many comparisons

*= Maximum Likelihood Estimation

Other Cardiac Adverse Events (OC)469 Events (52 P, 417 D)*RR = 0.33 (0.24, 0.44) P=<10-15)Not a primary, 1 of many comparisons

Together These Two CategoriesAccount for 3,203 EventsOr 33.3% of All Drop Outs

9

Other Adverse Events (OAE) Usually a combination of 2, 3 or more various kinds of signs, symptoms or laboratory abnormalities (that included skin rash, nausea and vomiting, headache, fatigue/drowsiness, etc. Rarely a single descriptor.

Other Cardiac Adverse Events (OC) Hypotension was most prominent (140 D, 8 P) with 56% being postural hypotensive phenomena, non-specific EKG changes were next most frequent, then in decreasing frequency edema of extremities, non-descript chest pain.

10

TF Placebo

TF Drug

OT Placebo

OT Drug

OAE Placebo

OAE Drug

OC Placebo

OC Drug

0.00%

2.00%

4.00%

6.00%

8.00%

10.00%

12.00%

14.00%Percent of DropOut Group with ER/Hosp

ER

Hosp

Unscheduled Visits

116 Events Total

222 Events Total

54 Events Total

22 Events Total

11

Hypertensive Emergency (HE)279 Events (134 P, 145 D)*RR = 2.75 (2.19, 3.57) p<10-15)Not a primary, 1 of many comparisons

Meant to require new end organ damage with high blood pressureBUT

Defined as New End Organ Damage OR

Diastolic > 110 mm HgIncrease of Diastolic >10 mm Hg

Less than 25% of the 279 patients in this categoryhad clear “in words” new end organ organ involvement

12

QuickTime™ and aNone decompressor

are needed to see this picture.

Baseline Cumulative Distribution Frequency

Sitting Diastolic Blood Pressure

All 9,636 drop outs (placebo & drug groups)

13

QuickTime™ and aNone decompressor

are needed to see this picture.

BaselineCumulative Distribution Frequency

Supine Diastolic Blood Pressure

All 9,636 drop outs (placebo & drug groups)

14

Change From BaselineCumulative Distribution Frequency

Supine Diastolic Blood Pressure

QuickTime™ and aNone decompressor

are needed to see this picture.

All 9,636 drop outs (placebo & drug groups)

15

Change From BaselineCumulative Distribution Frequency

Sitting Diastolic Blood Pressure

QuickTime™ and aNone decompressor

are needed to see this picture.

All 9,636 drop outs (placebo & drug groups)

16

TF Placebo

TF Drug

OT Placebo

OT Drug

OAE Placebo

OAE Drug

OC Placebo

OC Drug

HE Placebo

HE Drug

0.00%

5.00%

10.00%

15.00%

20.00%

25.00%

Percent of DropOut Group with ER/Hosp

ER

Hosp

Unscheduled Visits

22 Events Total

116 Events Total

222 Events Total

54 Events Total

52 Events Total

17

TF Placebo

TF Drug

OT Placebo

OT Drug

OAE Placebo

OAE Drug

OC Placebo

OC Drug

HE Placebo

HE Drug

TIA Placebo

TIA Drug

AR Placebo

AR Drug

CHF Placebo

CHF Drug

AP Placebo

AP Drug

Death Placebo

Death Drug

MI Placebo

MI Drug

CVA Placebo

CVA Drug

0.00%

20.00%

40.00%

60.00%

80.00%

100.00%

120.00%

Proportion of Drop Out Category With ER/HOSP

ER

Hosp

7.47% of the drop out population had an unscheduled visit.

***

*= Statistically Significant RR

18

Higher Risk on Treatment Higher Risk on Placebo

Relative Risk (Placebo/Drug)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

OTOAETFOCHEAPARMI

CHFDeathCVATIAVT

All dropout

ERHOSP

ER or HOSP

CVA, MI, DeathHE, AP, MI, CHF, CVA

OC, AR, Death, VT

p<10-15

p<10-16

p<10-15p<10-15

p=1.7X10-3

There were 19 different analyses performed, Inferential p of 0.001 requires p<10-6

19

Favor PlaceboFavor Drug SUM-50

-40

-30

-20

-10

0

10

20

30

40

Difference (Placebo - Drug) per 1,000 Patient Years

OAE

OC

AR

Death

HE

CHF

CVA

MI

TIA

AP

RR=2.75

p=<10-15

RR=0.87p=0.0017

RR=0.33

p<10-15

RR=0.89p=0.001

All Drop Outs Without OT and TFSUM

20

Favor PlaceboFavor Drug SUM-50

-40

-30

-20

-10

0

10

Difference (Placebo - Drug) per 1,000 Patient Years

OAE

OC

AR

Death

CHF

CVA

MI

TIA

AP

RR=0.80

p<10-8

RR=0.87p=0.017

RR=0.33

p<10-15

All Drop Outs Without OT, TF and HE

SUM

21

Favor PlaceboFavor Drug SUM-0.5

0

0.5

1

1.5

2

Difference per 1,000 patient years

Death

MI

CVA

Irreversible Harm

RR = 1.03p = 0.86 Sum

22

All Cause Mortality11 trials + HVET Pilot + PHARM179,766 patient years of follow-upVAH-NIH had no systolic pressurereported

Solid Colors = PlaceboStippled = Drug Treated

Points to PHARM

23

Non-Fatal Stroke

Solid Colors = Placebo

Stippled = Drug Treated

11 trials + PHARM178,334 patient years of follow-up

Points to PHARM

24

Non-Fatal MISolid Colors = PlaceboStippled = Drug Treated

11 trials + PHARM178,334 patient years of follow-up

Points to PHARM

25

All Cause Mortality11 trials + PHARM + HVET-pilot Solid Colors = Placebo

Stippled = Drug Treated

Points to PHARM

26

Non-Fatal Stroke11 trials + PHARM (HVET-pilot had onlyMortality)

Solid Colors = PlaceboStippled = Drug Treated

Points to PHARM

27

11 trials + PHARM (HVET-pilot hadonly Mortality)

Non-Fatal MISolid Colors = PlaceboStippled = Drug Treated

Points to PHARM

28Lewington et al Lancet 360:1903-1913, 20002

29Collins & MacMahon Brit Med Bul 50:272-298, 1994

30

So, Where Does That Leave Us?In My Opinion:

30 years and 590 trials (12,657 patient years)have produced no evidence that there is an increase of irreversible harm produced by the utilization of placebo in short-termtrials

The population represented by PHARM appears to be like that of all publishedstudies, but of relatively low absolute risk(low systolic blood pressure and low chronological age)

31

Equipoise can be maintained (mainly because of the low absolute risk), suchtrials can continue; after another 30 to 50 years, one should check on the wisdom of continued equipoise

The only meaningful discussion shouldrelate to could these trials be made “safer” by limiting chronological age toSome arbitrary limit (say 50), and systolicBlood pressure to some arbitrary limit(say 145)

32

Back-up Stuff, unorganized

33

AP AR CVA CHF Death HE MI OT OAE OC TF TIA VT-15

-10

-5

0

5

10

15

-15

-10

-5

0

5

10

15

Change From Baseline (mm Hg) for Event Group

Difference (Drug-Placeebo) for the group (mm Hg)

Drop Out Category

Change (Base-Last) for Group

Diff (Drug - Placebo)

Sitting Diastolic Blood Pressure

* *

*

*

*

*

*

*

*

*

*

34

Baseline >= -7 Days <= -1 Treatment >= 12 Days <= 63

By Patient Summary Entire SessionsAll Measurements in a Session, 24 or more Hours

Error Bars are 1 SEM, n varies from 364 patients Placebo to 1050 patients Drug

Placebo Drug

-12

-10

-8

-6

-4

-2

0

2

Treatment Effect (Treatment - Baseline) mm Hg

Sys

Dia

Sys Dia

p < 0. 000000000

p < 0. 000000000

p = 0. 6

1st Analysis, not to be trusted

p = 0. 5

p values are that Difference

is = 0

May 1999From 12/98

35

Baseline >= -7 Days <= -1 Treatment >= 12 Days <= 63

Error Bars are 1 SEM, n varies from 364 patients Placebo to 1050 patients Drug

Morning By Patient Summary6:00 A.M. to 9:59 A.M.

Placebo Drug

-12

-10

-8

-6

-4

-2

0

2

Drug Effect (Treatment - Baseline) mm Hg

Sys

Dia

SysDia

p < 0. 000000000

p < 0. 000000000

p = 0. 9

1st Analysis, not to be trusted

p = 0. 6

p values are that Difference

is = 0

Value = -0.05

May 1999From 12/98

36

Placebo Drug

-14

-12

-10

-8

-6

-4

-2

0

2

Treatment Effect (Treatment - Baseline) mm Hg

Baseline >= -7 Days <= -1 Treatment >= 12 Days <= 63

Error Bars are 1 SEM, n varies from 364 patients Placebo to 1050 patients Drug

DayTime By Patient Treatment Effect10:00 A.M. to 5:59 P.M.

Sys

Dia

Sys Dia

p < 0. 000000000

p < 0. 000000000

1st Analysis, not to be trusted

p = 0. 6

p values are that Difference

is = 0

p = 0. 3

May 1999From 12/98

37

Baseline >= -7 Days <= -1 Treatment >= 12 Days <= 63

By Patient Summary Entire SessionsAll Measurements in a Session, 24 or more Hours

Error Bars are 1 SEM, n varies from 364 patients Placebo to 1050 patients Drug

Placebo Drug

-12

-10

-8

-6

-4

-2

0

2

Treatment Effect (Treatment - Baseline) mm Hg

Sys

Dia

Sys Dia

p < 0. 000000000

p < 0. 000000000

p = 0. 6

1st Analysis, not to be trusted

p = 0. 5

p values are that Difference

is = 0

May 1999From 12/98

38

diastolic systolic

-14

-12

-10

-8

-6

-4

-2

0

2Mean Change From Baseline (mm Hg)

placebo abp

Rx abp

placebo cuff

RX Cuff

Preliminary Results from 9 Trialsn = about700 for cuff, 250 for abpm

SE

May 1999From 12/98

39

closest 5 min 10 min 30 min 60 min 120 min60

70

80

90

100

110

120

130

140

150

160

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Standing Systolic

abp Systolic

Standing Diastolic

abp Diastolic

Correlation Systolic

Correlation Diastolic

Blo

od P

ress

ure

(m

m H

g)C

orrelation C

oefficient

Simultanious (?) ABPM and Cuff Blood Pressures

May 1999From 12/98