Get documents about "In other words
Document Sample
1
U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES
+ + + + +
PUBLIC HEALTH SERVICE
+ + + + +
FOOD AND DRUG ADMINISTRATION
+ + + + +
CENTER FOR DRUG EVALUATION AND RESEARCH
+ + + + +
CARDIOVASCULAR AND RENAL DRUGS
ADVISORY COMMITTEE
+ + + + +
88TH MEETING
+ + + + +
THURSDAY,
APRIL 29, 1999
+ + + + +
The meeting took place in the Jack Masur
Auditorium, Clinical Center, Building 10, National
Institutes of Health, 9000 Rockville Pike, Bethesda,
Maryland at 9:00 a.m., Milton Packer, M.D., Chairperson,
presiding.
PRESENT:
MILTON PACKER, M.D., Chairperson
JOAN C. STANDAERT, Executive Secretary
ROBERT CALIFF, M.D., Member
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
2
PRESENT: (CONT'D.)
THOMAS GRABOYS, M.D., Consumer Representative
CINDY GRINES, M.D., Member
MARVIN KONSTAM, M.D., Member
JoANN LINDENFELD, M.D., Member
LeMUEL MOYÉ, M.D., Ph.D., Member
ILEANA PIÑA, M.D., Member
UDHO THADANI, M.D., FRCP, Member
J. THOMAS BIGGER, M.D., Guest Expert
MICHAEL CAIN, M.D., Guest Expert
ROBERT FENICHEL, M.D., FDA Representative
PRAN MARROTT, M.D., Sponsor Representative
PETER KOWEY, M.D., Sponsor Representative
ALSO PRESENT:
Lloyd Fisher, Ph.D.
Ed Pritchett, M.D.,
John Williams, M.D.
Daniel MacNeil, M.D.
Alexandra Kapatou, Ph.D.
Judy Jin, Ph.D.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
3
I-N-D-E-X
Call To Order and Opening Comments,
by Chairman Packer 4
Introductions, 4
Administrative Matters, 5
Open Public Hearing,
NDA 19-865, Betapace (sotalol HCL), Berlex Laboratories
Sponsor's Presentation,
Introduction, by Dr. Marrott 9
Clinical Pharmacology, by Dr. Kowey 13
Efficacy, by Dr. Kowey 16
Safety, by Dr. Kowey 231
Dosing Recommendations, by Dr. Kowey 248
Concluding Remarks, by Dr. Marrott 297
Committee Discussion and Review, 299
Committe Recommendation, 381
Adjourn, 412
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
4
1 P-R-O-C-E-E-D-I-N-G-S
2 (9:03 a.m.)
3 CHAIRMAN PACKER: This is the 88th meeting
4 of the Cardiovascular and Renal Drugs Advisory
5 Committee. At today's meeting, we have the usual
6 members of the committee. We have, also, two experts
7 who have been invited specifically to join us for today's
8 deliberations. And just so that we can do this in the
9 appropriate fashion, I'll ask the -- those who are seated
10 at the -- on the podium today to simply go down and
11 introduce themselves.
12 Lem, why you start. And just name and
13 affiliation.
14 DR. MOYÉ: Sure. Lem Moyé, University of
15 Texas, School of Public Health.
16 DR. BIGGER: Tom Bigger, Columbia
17 University.
18 DR. GRABOYS: Tom Graboys, Brigham and
19 Women's Hospital, Harvard.
20 DR. KONSTAM: Marv Konstam, Tufts
21 University, New England Medical Center.
22 DR. CALIFF: Rob Califf from Duke
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
5
1 University.
2 MS. STANDAERT: Joan Standaert, Executive
3 Secretary.
4 CHAIRMAN PACKER: Milton Packer, Columbia
5 University.
6 DR. LINDENFELD: JoAnn Lindenfeld,
7 University of Colorado.
8 DR. CAIN: Michael Cain, Washington
9 University in St. Louis.
10 DR. PIÑA: Ileana Piña, Temple University,
11 Philadelphia.
12 DR. THADANI: Udho Thadani, Oklahoma
13 University Health Sciences Center.
14 DR. FENICHEL: Bob Fenichel, Division of
15 Cardiorenal Drug Products, FDA.
16 CHAIRMAN PACKER: We'll ask Joan Standaert
17 to read the administrative matters for today.
18 Joan.
19 MS. STANDAERT: Yes, the following
20 announcement addresses the issue of conflict of interest
21 with regard to this meeting and is made a part of the
22 record to preclude even the appearance of conflict at
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
6
1 this meeting.
2 Based on the submitted agenda for the
3 meeting and all financial interests reported by the
4 participants, it has been determined that all interest
5 in firms regulated by the Center for Drug Evaluation
6 and Research, which has been reported by the
7 participants, sees that no potential for a conflict of
8 interest at this meeting with the following exceptions.
9
10 In accordance with 18 USC Section
11 208(b)(3), waivers have been granted to Dr. Milton
12 Packer, Dr. Cindy Grines, and Dr. Marvin Konstam. A
13 copy of these waiver statements may be obtained by
14 submitting a written request to the Agency's Freedom
15 of Information Office, Room 12A30 of the Parklawn
16 Building.
17 In addition, we would like to disclose for
18 the record that Dr. Robert Califf and Dr. Lemuel Moyé's
19 employers have interests which do not constitute a
20 financial interest in the particular matter within the
21 meeting at 18 USC 208, but which would create the
22 appearance of a conflict. The Agency has determined
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
7
1 notwithstanding these interests, that the interest in
2 the government in Dr. Califf's and Dr. Moyé's
3 participation outweighs the concern that the integrity
4 of the Agency's program and operations may be
5 questioned. Therefore, Doctors Califf and Moyé may
6 participate fully in the committee's discussions and
7 vote concerning Betapace.
8 With respect to FDA's invited guests, there
9 are reported interests that we believe should be made
10 public to allow the participants to object and
11 reevaluate their comments. Dr. Michael Cain would like
12 to disclose that he has been invited to attend an
13 arrhythmia board meeting sponsored by Proctor & Gamble.
14 In the event that the discussions involve any other
15 products or firms not already on the agenda for which
16 an FDA participant has a financial interest, the
17 participants are aware of the need to exclude themselves
18 from such involvement and their exclusion will be noted
19 for the record.
20 With respect to all other participants, we
21 ask in the interest of fairness that they address any
22 current or previous involvements with any firms or
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
8
1 products they may wish to comment upon.
2 And that concludes the statement for April
3 29th, 1999.
4 CHAIRMAN PACKER: Thank you, Joan.
5 We'll call for any public comment.
6 There being none, we'll move on to
7 evaluation of today's NDA. It's NDA 19-865, sotalol
8 or Betapace. The sponsor is Berlex Laboratories.
9 Proposed indication for the treatment of, or prevention
10 of, recurrence of atrial fibrillation/atrial flutter.
11 And I think that Dr. Marrott that will being the
12 presentation, please.
13 DR. MOYÉ: I'm just asking what the
14 preference is for asking questions today?
15 CHAIRMAN PACKER: Well, I think the sponsor
16 would always like to have the questions held or
17 segregated in distinct groups and I think that in general
18 we have followed that policy. If there are certain
19 issues of immediacy in clarification that you feel
20 shouldn't or cannot be held to a specific break in the
21 presentation, simply ask for a clarification.
22 DR. MOYÉ: But questions should occur at
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
9
1 the conclusion of each presenter's session?
2 CHAIRMAN PACKER: Well, we're going to
3 probably divide the presentation this morning into the
4 distinct categories which are listed on the agenda and
5 we'll take questions after each of them.
6 DR. MOYÉ: Thank you.
7 DR. MARROTT: Mr. Chairman, members of the
8 advisory committee, and Dr. Fenichel, good morning.
9 I would like to thank you, first of all, on behalf of
10 Berlex Laboratories, the sponsor, for inviting the
11 sponsor to make a presentation. Details of our
12 presentation can be seen on the slides.
13 After a brief introduction, Dr. Peter
14 Kowey, Professor of Medicine at Jefferson Medical
15 College, will provide an overview covering clinical
16 pharmacology, efficacy, safety, and dosing
17 recommendations.
18 The conclusion will be presented by myself.
19 Betapace or sotalol, or d,l-sotalol as our
20 products will be referred today, has been approved in
21 57 countries worldwide and is being used in both the
22 beta blockers as well as the arrhythmia indications.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
10
1 And NDA's files by Bristol-Myers Squibb, the previous
2 owner, was approved by the FDA in October '92 for the
3 indication life threatening ventricular arrhythmia.
4 Soon thereafter, the product was licensed in the U.S.
5 only to Berlex and Berlex launched Betapace in January
6 1993.
7 You will see from this slide that between
8 1993 and 1998, a considerable proportion of total
9 prescriptions, 60 to 77 percent, have been written for
10 patients suffering from supra ventricular arrhythmia,
11 chiefly atrial fibrillation and flutters. Thus, of the
12 total 3.6 million prescriptions, or thereabout, 2.5
13 million have been written for this disease.
14 This degree of use in atrial flutter and
15 fibrillation does not come as a total surprise to the
16 sponsor. Published articles in peer review journals
17 provide evidence of efficacy, safety, and benefit risk
18 to the physician of d,l-sotalol in atrial fibrillation.
19 Leading physicians have participated in investigation
20 trials undertaken by Bristol-Myers Squibb in this
21 population.
22 And last, treatment algorithms for atrial
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
11
1 flutter and fibrillation presented and discussed by
2 academic cardiac electro-physiologists at heart
3 meetings emphasize the use of sotalol in patients with
4 and without structural heart disease but in the absence
5 of heart failure.
6 Ever since we heard of this use in atrial
7 fibrillation, we have begun to consider what steps the
8 company should take because we would have liked to be
9 in a position to provide detailed information regarding
10 the safety of our product to the physicians in this
11 disease population. The next logical step for us,
12 therefore, was to complete the clinical program of
13 studies initiated by Bristol-Myers Squibb and which,
14 by the way,w as well underway. This, we did, and we
15 filed a supplemental NDA in June of 1998 for the atrial
16 fibrillation flutter indication.
17 Our proposed indication reads as follows.
18 d,l-sotalol is indicated for extending the time to
19 symptomatic recurrence of chronic or paroxysmal atrial
20 fibrillation or flutter in patients without or with
21 structural heart disease in the absence of heart
22 failure. We have present here today our consultants
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
12
1 who will participate in today's discussion. I have
2 already mentioned that Dr. Kowey will present the
3 overview on our behalf. In addition, participating in
4 the discussions are Doctors Pritchett, Fisher, and
5 Barbey. The titles and the affiliations of these
6 experts is mentioned on the slide.
7 We also have here today Dr. Dan MacNeil,
8 Executive Director of Clinical Research at
9 Bristol-Myers Squibb. Dr. MacNeil was responsible for
10 some of the clinical trials undertaken by Bristol-Myers
11 Squibb for d-sotalol; d,l-sotalol.
12 That concludes the introduction, Mr.
13 Chairman. I thank you for your attention. And with
14 your permission, I would like to ask Dr. Kowey to come
15 forward to present his overview.
16 Thank you.
17 CHAIRMAN PACKER: As Dr. Kowey is coming
18 forward, let me just, to facilitate communication, I
19 think it would be entirely appropriate for the committee
20 to refer to this drug as sotalol as opposed to continuing
21 to say d,l-sotalol unless someone wants to. And that
22 when specific reference is made to d-sotalol, that a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
13
1 clear distinction be made. But I think it would be
2 perfectly okay just to refer to sotalol all through
3 today's presentation except when the distinction is
4 important.
5 DR. MARROTT: Thank you very much.
6 DR. KOWEY: Mr. Chairman, Dr. Fenichel, Dr.
7 Lipicky, welcome back, members of the advisory committee
8 and ladies and gentlemen. It is with a good deal of
9 pleasure that I represent the sponsor this morning to
10 present information regarding the use of sotalol in
11 patients with atrial fibrillation and atrial flutter.
12 I will present this in four distinct sections and as
13 Dr. Packer already said, we will pause between sections
14 in order to take questions. But if you have any points
15 of clarification when the slides are up, please feel
16 free to let me know.
17 We're going to talk about clinical
18 pharmacology first, followed by efficacy, safety, and
19 dosing recommendations. We'll start with clinical
20 pharmacology.
21 A good deal of this information that I'm
22 going to show you this morning is already contained in
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
14
1 the package insert for sotalol since the
2 pharmaco-kinetics are the same for the compound that's
3 currently being used for patients with ventricular
4 arrhythmias.
5 This is a drug which has linear dose
6 proportional and predictable pharmacokinetics. It is
7 nearly 100 percent bioavailable. It's t-max is 2.5 four
8 hours. In cases of normal renal function, the half life
9 of the drug is 12 hours. In case with abnormal renal
10 function, the half life is prolonged.
11 Notably, the drug is not metabolized by any
12 enzyme system in the liver. Most importantly, not by
13 the P-450 enzyme system. It is excreted -- More than
14 75 percent of the drug is excreted in urine. It's renal
15 elimination is mainly by glomerular filtration and
16 protein binding is negligible.
17 We would like to make a few comments about
18 special populations because this is important in dosing
19 the drug. Most importantly are patients who have renal
20 dysfunction. Remember, the plasma clearance is reduced
21 and the half life is prolonged in patients who have renal
22 dysfunction described by creatinine clearance.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
15
1 Therefore, in all of the clinical trials, dose
2 adjustment was needed and was carried out in patients
3 who had reduced creatinine clearance, or patients were
4 excluded from the clinical protocol on that basis.
5 The observed effects in patients who are
6 old, and males versus females, are almost entirely
7 accounted by differences in renal function. Hepatic
8 dysfunction has no effect on the kinetics of the drug.
9 Finally, a statement regarding the
10 pharmacokinetic drug interactions: There is a 20
11 percent reduction in area under the curve in patients
12 who have been fed. There is a specific drug interaction
13 with Maalox and not to our knowledge with other antacids
14 which causes about a 20 to 25 percent reduction in C
15 max in area under the curve.
16 There are no demonstrable interactions
17 between hydrochlorothiazide, warfarin, or digoxin. I
18 would point out that for hydrochlorothiazide and
19 warfarin, there is no effect either on sotalol or
20 warfarin or hydrochlorothiazide blood concentrations.
21 Digoxin levels are not increased in patients who
22 receive d,l-sotalol but there have not been sufficient
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
16
1 studies to document what happens to d,l-sotalol in the
2 presence of digoxin.
3 Dr. Packer, that concludes my section on
4 clinical pharmacology.
5 CHAIRMAN PACKER: Okay. I don't see any
6 questions. Why don't you proceed.
7 DR. KOWEY: Thank you.
8 I'll now cover efficacy. This is a
9 somewhat longer part of the presentation. We're going
10 to be presenting information regarding a number of the
11 clinical trials in the d,l-sotalol efficacy database.
12 I want to point out that we will be discussing the eight
13 control trials in the database and in addition, we will
14 be presenting a bit of information regarding the use
15 of sotalol as it occurred in dofetilide, database Study
16 345, which you're familiar with and was presented at
17 the last advisory committee meeting in January.
18 On the top, I have listed the categories,
19 the broad categories, of atrial fibrillation type,
20 prevention, for chronic atrial fibrillation prevention,
21 which really, according to the indication that we've
22 listed, doesn't really mean prevention but extension
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
17
1 of time in recurrence. For paroxysmal atrial
2 fibrillation, one study that considered not only
3 prevention of chronic atrial fibrillation but also
4 conversion of the arrhythmia. And then finally two
5 studies which examine the interaction between the drug
6 and digoxin.
7 The studies which are in pink are those
8 studies for which I will provide fairly detailed
9 information. We do have information regarding Study
10 G which is a subpopulation study in AF and for the two
11 digoxin studies, and we have that available if you have
12 questions about those trials. They will be included
13 in the safety database, but for efficacy I won't be
14 covering them this morning.
15 Let me start with Study 004 which was a study
16 in patients with chronic atrial fibrillation and atrial
17 flutter. And "chronic" in this study, and in most of
18 the clinical trials I'll describe to you this morning,
19 is defined as greater than two weeks in duration and
20 less than one year. These patients were cardioverted
21 and were in normal sinus rhythm at the time that they
22 were randomized. And they needed to be in normal sinus
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
18
1 rhythm for greater than two hours before they were
2 randomized.
3 This was a study that was done in out-
4 patients and patients were randomized to placebo
5 d,l-sotalol. The d,l-sotalol dose was 80 to 160
6 milligrams twice per day. And this drug was given in
7 a blinded titration fashion. Or d-sotalol in doses
8 between 100 and 200 milligrams twice per day. Again,
9 this dose was blindly titrated.
10 I would point out in this study, patients
11 who had a creatinine clearance of less than 50 ccs per
12 minute were excluded from the study. Patients not
13 tolerating d,l-sotalol at a BID regimen received the
14 drug 80 milligrams once per day.
15 This dose titration process went on for two
16 weeks and was followed by 22 weeks of maintenance at
17 the fixed titrated dose. There was an opportunity to,
18 again, titrate to tolerance. I want to point out that
19 discontinued patients were followed in this study for
20 the full six months of the trial.
21 This study had three distinct primary
22 endpoints: time to recurrence of symptomatic ECG
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
19
1 documented atrial fibrillation, the time to recurrence
2 of ECG documented atrial fibrillation including
3 patients who did and did not have symptoms (so
4 asymptomatic patients were found on routine telemetry
5 monitoring), and the number of patients remaining in
6 sinus rhythm after six months of therapy as
7 proportioned. There was a secondary endpoint, change
8 in defibrilar rate in patients prior to therapy and on
9 therapy, which I won't discuss in detail but we can show
10 you, if you'd like.
11 Let me discuss each of these primary
12 endpoints when we get to the efficacy evaluation.
13 First of all, I want to point out, in the
14 statistical analysis of efficacy, for this and for most
15 of the subsequent studies that I'm going to show you,
16 that the pre-specified analysis was done by log rank
17 with Kaplan Meier survival. Included in your briefing
18 document and in the analysis is a second statistical
19 test, a generalized Wilcoxon test called the Gehan
20 statistic. The Gehan statistic is useful for
21 demonstrating efficacy in the early portion of the
22 Kaplan Meier. Whereas, the log rank is more valuable
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
20
1 in the latter portions of the Kaplan Meier.
2 This is a chi squared for the number of the
3 patients remaining in normal sinus rhythm, which was
4 one of the endpoints of the study. I want to point out
5 in this and several of the subsequent studies that we
6 carried out a Cox proportional hazards model to describe
7 the relative risk of sotalol use compared to placebo.
8 And we also used this analysis to determine the effect
9 of prognostic risk factors, which I'll show you. And
10 then finally, for quantitative data, we used an analysis
11 of variance, an ANOVA which was a one-way analysis of
12 variance.
13 These are the demographics for Study 004:
14 age, gender, race, and creatinine clearance, pointing
15 out that there were patients in the trial with creatinine
16 clearance of less than 60 ccs per minute who were not
17 excluded from the study because its cut off, as you'll
18 recall, was 50 ccs per minute. So these patients were
19 in sort of the borderline range.
20 The groups were well matched according to
21 the clinical characteristics. Similarly, they were
22 well matched with regard to the cardiac history.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
21
1 Majority of the patients in this study were a New York
2 Heart Association Class I and II. About half the
3 patients had structural heart disease. You see the
4 percentage here: patients who had coronary artery
5 disease; and a smaller subset of those patients who had
6 a previous myocardial infarction; and a 20, 30 percent,
7 40 percent incidence of having had hypertension.
8 Remember, this was a study in which the
9 endpoint of the study was symptomatic recurrence of
10 atrial fibrillation or atrial flutter. This is a slide
11 showing you what the symptoms were in these patients,
12 and what their arrhythmia history had been. This, on
13 the top line, is the number of months since the first
14 episode of atrial fibrillation that the patient
15 reported. This is the duration of the atrial
16 fibrillation episode that got the patient into the
17 study. And as you can see, it was about four months.
18 These are sort of the typical symptoms that
19 you would expect in patients who have atrial
20 fibrillation: weakness, palpitation, shortness of
21 breath, and dizziness, chest pain being the most common.
22 Now, remember, in the sotalol arm of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
22
1 study, patients were titrated between 80 and 160
2 milligrams twice per day. And so it's important for
3 you to know that the majority of patients, two-thirds
4 of the patients, in the maintenance phase of the study,
5 during that 22 week period, were actually on 160
6 milligrams twice per day. Smaller percentage on the
7 lower doses.
8 This is the Kaplan Meier curve for the first
9 primary pre-specified endpoint in the clinical trial
10 which was time to first ECG-documented recurrence of
11 symptomatic atrial arrhythmia since randomization.
12 And you can see how the groups are colored here: sotalol
13 in blue, d-sotalol in yellow, placebo in red. And these
14 are the statistics for the analysis. This is the log
15 rank statistic, and this is the Gehan statistic. And
16 in all the Kaplan Meier curves that I'll be showing you,
17 you'll be seeing this kind of a lay out for the
18 statistical analysis.
19 Following several of the Kaplan Meier
20 curves I'm going to show you, I'll also show you tabular
21 data which comes from the same data set. This is medium
22 time to recurrence in days with placebo group, for the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
23
1 d,l-sotalol group, and for the -- I'm sorry, Milton.
2 I'll try not to do that too many times. For the sotalol
3 group and for the d-sotalol group.
4 The reason why this is greater than 180 days
5 is because fewer than 50 percent of the patients had
6 a recurrence of arrhythmia in those groups at the
7 endpoint of the study.
8 Percentage of relapse-free patients. This
9 is the p value you've already seen. And this is the
10 relative risk by the Cox method that I describe in the
11 statistical slide. And these are the confidence
12 intervals for those observations. Point 56 for sotalol
13 at these confidence intervals.
14 Let me just back up to that. Can I back
15 up to that slide? I'm sorry.
16 I just want to point out that two deaths
17 did occur in this study. Neither one was on sotalol.
18 One was on d-sotalol, and one was in placebo. And it's
19 important for you to know that they were censored in
20 the analysis at the time of the death for the Kaplan
21 Meier curve that I showed you.
22 This is effect of prognostic factors on the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
24
1 hazard risk of sotalol versus placebo at six months after
2 randomization. What this slides lends is the fact that
3 the covariates did not provide an alternative
4 explanation of the clinical benefit. This is the
5 unadjusted clinical benefit. This is the clinical
6 benefit adjusted for the baseline factors. And you can
7 see that they line up, indicating that there was balance
8 in the randomization.
9 I also want to show you a subgroup analysis
10 of these data using what we consider to be important
11 clinical variables; and that is age, gender, structural
12 heart disease, New York Heart Association class, years
13 since the development of the arrhythmia. And you can
14 see that there is good consistency of the data with the
15 point estimates lining up on the side favoring sotalol.
16 I do want to point out that this consistency
17 held for patients older and younger than 65, for men
18 as well as women, and for patients who did and did not
19 have structural heart disease in this clinical trial.
20 I also want to point out that this is the
21 remainder of that same subgroup analysis. This is part
22 two. I want to point out that it also held up for
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
25
1 patients who had a creatinine clearance less than 60
2 ccs per minute and greater than 60 ccs per minute.
3 This is the Kaplan Meier curve of similar
4 data from Study 004. This is time to first ECG-
5 documented recurrence of symptomatic atrial
6 fibrillation or atrial flutter. We now have added in
7 death or discontinuation since randomization. Since
8 there were very few deaths in the study, and since there
9 were actually very few discontinuations in the study,
10 the log rank p value looks very similar to what you had
11 already seen and so does the statistical Gehan analysis.
12 You remember that the second primary
13 endpoint in this clinical trial was time to ECG-
14 documented recurrence of any atrial fibrillation or
15 atrial flutter since randomization. This is the Kaplan
16 Meier analysis for that data set, again showing
17 separation between sotalol, d-sotalol, and placebo; and
18 these are the p values for that observation.
19 Finally, the third primary endpoint in
20 Study 004 was the percentage of patients in normal sinus
21 rhythm at six months as a proportion. There were 32
22 percent of the placebo patients in normal sinus rhythm
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
26
1 at six months compared to 50 percent of patients in the
2 sotalol group with this p value.
3 You have received the communication from
4 the Food and Drug Administration and the staff regarding
5 a possible concern about Study 29. Study 29 was a
6 center in Stockholm which enrolled patients in the
7 latter phases of the trial. And as you can see from
8 these numbers, for d,l-sotalol and for d-sotalol, that
9 there was a robust treatment effect for Study 29 or for
10 Center 29.
11 We have a difficult time understanding why
12 data are being extracted for a single center. And this
13 is more or less to play chance on a clinical trial.
14 We want to point out that there was a center in this
15 study, Center 24, that had a particularly bad effect.
16 And in fact, if the data for Center 24 and Center 29
17 are both taken away from the analysis, the best and the
18 worst, the p values remain statistically significant.
19 We've prepared more of a discussion
20 regarding this issue which we'd be very happy to have
21 with you. Dr. Lloyd Fisher, who is here with us today,
22 has looked at these data very carefully and is prepared
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
27
1 to offer some of his interpretation of the data as well.
2 The second study in the efficacy database
3 which I'd like to address briefly is Study 345 which,
4 again, is a study that you've seen in January, which
5 was the dofetilide Study 345. In Study 345, which
6 consisted of 671 patients, 137 patients received
7 d,l-sotalol, and the same number of patients received
8 placebo. As you'll recall, these are patients who had
9 chronic atrial fibrillation or atrial flutter at entry.
10 The duration was one week to two years which looks
11 familiar to the enrollment criteria for our trials.
12 And these were all patients who had been successfully
13 converted to normal sinus rhythm either
14 pharmacologically or electrically.
15 This study was a 12-month randomized
16 parallel group, double-blind, placebo and active
17 control study. And again, the active comparator in the
18 study was racemic sotalol. The primary endpoint of the
19 study was time from conversion to normal sinus rhythm.
20 So once the patients were in normal sinus rhythm, it's
21 the time it took for them to recur with atrial
22 fibrillation, atrial flutter. The statistical
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
28
1 analysis for the Study 345 is the same as the statistical
2 analysis that we had used for our data.
3 I apologize. This slide is not colored in
4 the same manner as the slides that we've used this
5 morning, but that's because we obtained this information
6 from the Freedom of Information and we weren't able to
7 really do much with it. It was scanned. But I just
8 want to point out that we have put a red arrow on this
9 for you so you can see the 80 milligrams twice per day
10 of sotalol dose arm, and this is the placebo arm.
11 Remember that this is the lower end of our dose range
12 for our clinical trials but it was the dose that was
13 included in the dofetilide experience. And this is the
14 p value for the observation of the difference between
15 sotalol and placebo. This is the percentage of patients
16 in normal sinus rhythm at 12 months.
17 I should point out that we will return to
18 Study 345 in the safety analysis because we do have some
19 safety information to show you also from that trial.
20 I want to now move from the chronic atrial
21 fibrillation cohort to move into the patients in the
22 paroxysmal atrial fibrillation cohort.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
29
1 CHAIRMAN PACKER: Peter, if I could just
2 have you pause. If there is anyone from the sponsor's
3 point of view for dofetilide 345, I think it would be
4 appropriate for us to hear their comments later on.
5 I just want to give everyone a heads up on that.
6 Second is just a clarification. Freedom
7 of Information normally applies to access of information
8 for drugs that have been approved. I don't know of any
9 specific action on the approval of dofetilide. How does
10 Freedom of Information apply here?
11 DR. KOWEY: I don't know. Milton, it was
12 presented at a public hearing in January. So I would
13 have assumed that that means that it is in the public
14 domain, but I --
15 CHAIRMAN PACKER: No, I --
16 DR. KOWEY: I'm not an attorney, so I can't
17 really tell you.
18 CHAIRMAN PACKER: I think the reason I'm
19 bringing it up is that I think it is in the public domain.
20 But I don't think it could possibly have been obtained
21 by Freedom of Information.
22 DR. KOWEY: Okay. I stand corrected. But
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
30
1 I don't think there's anything wrong with having shown
2 the information on the other hand. Do you agree?
3 CHAIRMAN PACKER: I'm sorry?
4 DR. KOWEY: There's nothing wrong with
5 having shown the information?
6 CHAIRMAN PACKER: No, no. There's nothing
7 wrong. I just want to clarify.
8 DR. KOWEY: And there's nothing wrong with
9 the FDA taking this into account in the approval process
10 route for racemic sotalol.
11 So I apologize if I misspoke.
12 Actually, all I was trying to do was tell you why
13 it was such a crappy slide. I probably should have just
14 kept my mouth shut.
15 Let me move on to Study 05 which is the
16 paroxysmal atrial fibrillation cohort. Again, this
17 says prevention. I want to make sure that everybody's
18 very clear. We read the indication. Milton read the
19 indication. Pran read the indication. It's
20 prolongation to time to recurrence. Not overall
21 prevention of the arrhythmia.
22 Study 05 was a study that included patients
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
31
1 who had atrial fibrillation within the last three
2 months. But at the time that they were actually
3 enrolled in the clinical trial, they were in normal sinus
4 rhythm. The majority of these patients, the vast
5 majority of these patients, had spontaneous reversion
6 to normal sinus rhythm. It did not require
7 cardioversion in order to have them in sinus rhythm at
8 the time of randomization.
9 I'd like to point out that this is a unique
10 study in the database because it is the only study in
11 which inpatient dosing was mandatory, was mandated.
12 And it was mandated for patients who had structural heart
13 disease. Investigators had the option of using the drug
14 outpatient for patients who did not have structural
15 heart disease but they didn't have to use it outpatient.
16 So inpatient was mandatory; outpatient wasn't.
17 Patients were randomized to placebo and to one of three
18 doses of sotalol, 80 milligrams twice per day, 120
19 milligrams twice per day, and 160 milligrams twice per
20 day.
21 Now in this study, in contrast to 04,
22 patients who had creatinine clearances of 40 to 60 ccs
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
32
1 per minute received the drug once a day rather than being
2 excluded from the protocol. If they were under 40 ccs
3 per minute, they were out. Open label treatment, as
4 I would point out here, was optional for the remainder
5 of the 12 months if the patients had a recurrence. So
6 the patients could have treatment for 12 months open
7 label after recurrence; and the duration of the study,
8 as you can see here, was 12 months.
9 I want to point out and it's important to
10 realize and remember that these patients were titrated
11 to their dose and that was the dose that they had to
12 receive. If they couldn't tolerate the dose, they were
13 dropped from the study.
14 The primary pre-specified endpoint in the
15 analysis was the time of the first recurring symptomatic
16 episode of atrial fibrillation or atrial flutter during
17 the efficacy evaluation period. What does that mean?
18 That means that after the patients had been dosed for
19 three days if they were receiving the drug twice a day,
20 or six days if they were receiving the dose once a day,
21 to get to a presumed steady state plasma concentration.
22
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
33
1 There were a number of secondary endpoints
2 in this trial. We will present you some of this
3 information. For example, time to the first recurring
4 symptomatic episode of arrhythmia after the first dose
5 of study medication, which has been referred to by some
6 people as the intention-to-treat analysis. Also, the
7 proportion of patients free of recurring symptomatic
8 atrial fibrillation and flutter at six and 12 months,
9 another secondary endpoint. And time to occurrence in
10 patients who were receiving the drug twice a day or once
11 a day.
12 Again, I won't go through this. It's
13 exactly the same statistical methods that were used in
14 004. These are the pre-specified analyses. The Gehan
15 was not pre-specified. It was used post hoc in order
16 to examine the data because of the high incidence of
17 early recurrence. And this is the Cox proportional
18 hazards for relative risk, prognostic risk factors, and
19 dose response relationship.
20 These are the demographics of the study.
21 I want to point out that about a quarter to a third
22 of the patients had creatinine clearances of less than
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
34
1 50 ccs per minute and may have therefore received a once-
2 a-day dose of the medication. These are the patients'
3 race, female, gender, et cetera, which are well matched.
4 This is structural heart disease by dose
5 groups. Pointing out patients, again, a relatively
6 similar percentage of patients with coronary artery
7 disease. This is the subgroup with myocardial
8 infarction. And as I pointed out already, the majority
9 of patients in this clinical trial had been designated
10 by the investigator to have had defined paroxysmal
11 atrial fibrillation. The remainder had what the
12 investigator called chronic atrial fibrillation.
13 This is the time for the first ECG-
14 documented recurrence of symptomatic arrhythmia from
15 presumed steady state. So this is the primary
16 pre-specified analysis. Again, looking at log rank and
17 Gehan statistic, the Gehan showing a more robust p value
18 than the log rank, 120 milligrams used in this analysis,
19 showing a more robust p value than the 160 milligram
20 group.
21 These are the tabular data. We begin by
22 the number of patients in the trial who discontinued
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
35
1 because of adverse events. Again, I would point out
2 that because the patients were placed in a dose arm and
3 could not leave that dose arm or be titrated, there was
4 a higher, and an expectedly higher, discontinuation rate
5 in patients who received 160 milligrams, either twice
6 a day if they had normal creatinine clearances or once
7 a day. Median time to recurrence: Again, you can
8 confer from the Kaplan Meier that it would have been
9 longer for the d,l-sotalol group. This is the
10 percentage of patients at the end of the 12 month period
11 who are relapse free, one of the secondary endpoints.
12 Because the 160/120 milligram lines crossed towards
13 the end of the study, it turned out that more patients
14 for 160 milligrams group by that analysis were in sinus
15 rhythm.
16 These are the p values by log rank and Gehan.
17 And these are the same point estimates for relative
18 risk of the confidence intervals for each of the dose
19 groups.
20 This is time to first ECG-documented
21 recurrence of symptomatic atrial fibrillation or atrial
22 flutter since the patient had been randomized. So this
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
36
1 starts from the first time the patient took a dose which
2 is obviously earlier than the time of presumed steady
3 state.
4 There were actually more patients in this
5 analysis because the number of patients were dropped
6 during that initial phase of the study. And because
7 of that, the p values are perhaps a bit more robust for
8 both the Gehan and the log rank. And you can see here
9 that there is a step up to 160 milligrams with a -- there
10 is a significant p value attached to the 160 milligram
11 dose for the log rank and also for the Gehan.
12 This is an analysis which is a very
13 draconian look at the data. It involves taking patients
14 who not only had recurrences in the study, but were also
15 were discontinued since randomization. It is an
16 analysis that I showed you for 004 which happened to
17 show a better outcome because there were few dropouts.
18 In this analysis, it cancels out the statistical
19 benefit by log rank although not by Gehan. Since,
20 again, in the higher dose groups, as expected, there
21 were a larger number of dropouts. This is a question
22 that's been addressed to the committee, and we think
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
37
1 it's a very important statistical question. And
2 actually, Dr. Fisher also has some comments perhaps we
3 could have later during the discussion regarding this
4 question of handling of discontinuations in patients
5 in clinical trials of this nature.
6 This is the same evaluation to determine
7 the covariant. Covariants do not provide an
8 alternative explanation for the clinical benefit.
9 We're looking at several of the clinical characteristics
10 that I showed you from the last study: age, gender,
11 structural heart disease, coronary artery disease, et
12 cetera, in- versus outpatient initiation. And you can
13 see that because of balance there really isn't much of
14 a difference from the unadjusted point estimate.
15 This is the subgroup analysis as I showed
16 you in the last study. Looking at important clinical
17 variables -- age, gender, structural heart disease,
18 paroxysmal versus chronic atrial fibrillation -- most
19 of the patients here in this analysis obviously were
20 paroxysmal. I'd point out that this particular
21 analysis is done in patients who had received 120
22 milligrams of sotalol versus placebo, and this is at
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
38
1 12 month since randomization.
2 This continues the subject analysis I did
3 in the last study showing that whether or not the
4 creatinine clearance was high or greater than 60 or less
5 than 60, the benefit treatment effect was consistent.
6 I want to move on to Study 9A which was also
7 a study in patients with paroxysmal atrial fibrillation
8 as a subpopulation of a larger clinical trial.
9 CHAIRMAN PACKER: Peter, hold on one
10 second, please.
11 DR. KOWEY: Yes.
12 DR. CALIFF: Peter, just a point of
13 clarification. On the odds ratios that you're showing,
14 those are the non-intention to treat odds ratios?
15 DR. KOWEY: That is the non-intention to
16 treat. That was -- let me go back. Can I go back a
17 slide.
18 DR. CALIFF: That's been true for all the
19 odds ratios you've shown?
20 DR. CALIFF: Yes. Well, this one is -- you
21 can see here. If you call intention to treat from
22 randomization -- is that what you mean, Robert?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
39
1 DR. CALIFF: It's usually what we call
2 intention to treat.
3 DR. KOWEY: Yes. No, this is not for that
4 analysis. This is for the time-to-presume steady state
5 which was the primary analysis in the trial.
6 The primary analysis in the trial -- Oh,
7 this is from randomization. I'm sorry, Robert. This
8 is from randomization. So this is the intention to
9 treat.
10 DR. CALIFF: But it includes patients who
11 came off the drug, even if they weren't --
12 DR. KOWEY: This does not have the
13 discontinuation. Correct. Yes. These were where the
14 discontinuations were censored.
15 DR. CALIFF: Right. And we'll come back
16 to it later, but I wouldn't call that intention to treat.
17 I just want to clarify which analyses were being shown
18 as odds ratios.
19 DR. KOWEY: Yes, this analysis is -- let
20 me just clarify so everybody understands. It is from
21 the time of randomization, and it does not include
22 patients who were discontinued for adverse effects.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
40
1 CHAIRMAN PACKER: Peter, just a
2 clarification. You referred to the p value for the log
3 rank as 160 milligrams, .029 as being statistically
4 significant. The alpha assigned to that is .025.
5 DR. KOWEY: That's correct. You're right.
6 You're right.
7 CHAIRMAN PACKER: It is statistically
8 significant.
9 DR. KOWEY: That's correct. You are
10 correct.
11 The next study in the paroxysmal atrial
12 fibrillation strata is patients with -- in Study 9A which
13 was a sub-study of a larger study of patients with
14 paroxysmal supra ventricular tachycardia. This had a
15 relatively complicated baseline period. Let me just
16 explain to you how this was done.
17 Patients were observed for the first week.
18 If they had one episode of arrhythmia within the first
19 week, those patients then went through two more one-
20 week observation periods. These were patients that
21 obviously had fairly frequent arrhythmia, and therefore
22 the period of observation was shorter.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
41
1 On the other hand, at the other end of the
2 extreme, patients who had an episode every four weeks,
3 or once a month, in the first four-week period, then
4 went on to have two more four-week periods of observation
5 for a total of 12 weeks.
6 Once they had this baseline quantification
7 of arrhythmia frequency, they were randomized and
8 stratified by their baseline observation period to
9 d,l-sotalol, regular sotalol; or d-sotalol; or placebo.
10 Just to show you that about 60 percent of the patients
11 were in the yellow group, about 25 percent of the
12 patients were in the green group in terms of the
13 frequency of arrhythmia, and about 15 percent were in
14 the red group, randomized to these drugs during a dose
15 escalation phase, and then for the last two periods of
16 the study they were observed.
17 The endpoints of the study were time to
18 first recurrence of supra ventricular arrhythmia and
19 the percentage of patients without recurrence. The
20 statistical analysis for this was a Kaplan Meier
21 survival curve and a Cox proportional hazards model,
22 as you've seen before. This is the Kaplan Meier curve
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
42
1 for the intention-to-treat analysis for the patients
2 with supra ventricular arrhythmia, including patients
3 with PSVT as well as atrial fibrillation, showing you
4 each of the doses of sotalol and d-sotalol lining up
5 compared to placebo. These are the p values for those
6 overall observations which was the primary
7 pre-specified analysis.
8 On this slide, we've shown the
9 subpopulation of patients with paroxysmal atrial
10 fibrillation by history looking at sotalol, d-sotalol,
11 and placebo. What is striking about the results is the
12 relatively short time to relapse in patients in the
13 placebo group and the difference between that
14 observation and the time to relapse in patients on
15 sotalol. Yielding a p value which was highly
16 statistically significant with these confident
17 intervals for the relative risk observations, which are
18 here, the point estimates.
19 There was a single study in patients with
20 chronic atrial fibrillation that examined two issues.
21 One is conversion of atrial fibrillation or flutter
22 in normal sinus rhythm. And the second was exploration
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
43
1 of a higher dose of the drug. That was Study 014 of
2 161 patients.
3 This is a study, as I said, in patients with
4 chronic atrial fibrillation, again, defined the same
5 way that the other studies were defined. And these
6 patients were randomized between sotalol and placebo.
7 There was a dose titration phase in the first part of
8 the study. Patients were started on 160 milligrams
9 twice a day and then titrated with 320 milligrams, twice
10 per day at three-day intervals. If intolerance
11 occurred at the 160 milligram twice per day dose, they
12 could be titrated downward to 80 milligrams twice per
13 day. Patients with creatinine clearances in this study
14 of less than 50 ccs per minute were excluded from the
15 protocol. Patients following this period of titration,
16 if they had not converted to sinus rhythm on the drug,
17 underwent direct cardioversion.
18 After completion of the double-blind
19 treatment phase, an open label treatment for one year
20 was an option for the patients.
21 There were three endpoints that were
22 described in the protocol for this particular study.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
44
1 A portion of patients achieving sinus rhythm with
2 double-blind treatment by Fisher exact tests, the time
3 between restoration in sinus rhythm and relapse into
4 atrial fibrillation and atrial flutter analyzed by log
5 rank, and the proportion of patients remaining in sinus
6 rhythm at the end of six months of double-blind
7 treatment.
8 These are the demographics: New York Heart
9 Association class, percentage of patients with
10 structural heart disease, percentage of patients with
11 coronary disease or previous myocardial infarction.
12 This should look very familiar. This is for the placebo
13 group and the d,l-sotalol group.
14 Now remember that the first endpoint of the
15 trial and the unique endpoint of the trial was conversion
16 of atrial fibrillation or flutter to normal sinus rhythm
17 with drugs. So this is the pharmacologic conversion
18 rate during the dose titration phase of the study showing
19 a 30 percent conversion rate for d,l-sotalol compared
20 to one percent of patients on placebo. The 30 percent
21 value is very much in the range of what we've seen with
22 oral Class III for conversion of atrial arrhythmia to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
45
1 sinus. This is the p value like Fisher's exact.
2 This is the Kaplan Meier analysis, time to
3 relapse of atrial fibrillation or flutter since
4 restoration of normal sinus rhythm. And these are the
5 p values which are attached by log rank and by Gehan.
6 Deaths were censored in this particular Kaplan Meier
7 analysis.
8 These are the tabular data from those
9 observations. We're looking at number of patients who
10 were discontinued due to adverse events. There was a
11 fairly large number of patients who were discontinued
12 in this trial, remembering that we were using doses in
13 this trial which are higher than the doses which we are
14 recommending today for treatment of patients with this
15 arrhythmia. This is median time to recurrence in days,
16 percentage of patients relapse free, and the statistical
17 tests, log rank and Gehan, and the point estimate with
18 confidence intervals for the relative risk.
19 This is adding deaths or discontinuations
20 to relapse of atrial fibrillation or flutter in patients
21 in Study 014. These were the statistical results by
22 log rank and Gehan, remembering, again, that there was
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
46
1 a high discontinuation rate in patients who received
2 the drug at these doses.
3 Finally, the last study that I'd like to
4 outline for you is Study H which was in patients with
5 chronic atrial fibrillation and which sotalol was
6 compared to quinidine.
7 Again, the same group of patients with
8 atrial fibrillation and atrial flutter now more than
9 two months and less than one year. These patients were
10 cardioverted and needed to remain in normal sinus rhythm
11 following cardioversion for more than two hours before
12 randomization. This was an open label study. And
13 treatment was randomized between sotalol at a dose of
14 80 to 160 milligrams twice per day including sulphate,
15 400 to 600 milligrams twice per day. These are the
16 number of patients that were treated with each of these
17 regiments.
18 This is a cardiovascular history. It's
19 worth pointing out that in this trial, as not in the
20 other trials that I've shown you, that there were
21 patients who had congestive heart failure by history.
22 There were more patients who had cardiomegaly. And
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
47
1 in fact, there were about 16 to 18 percent of patients
2 who had Class III New York Heart Association class.
3 The distribution of patients with coronary disease and
4 previous myocardial infarction should look familiar.
5 This is the Kaplan Meier analysis in which
6 we have analyzed the time to recurrence of atrial
7 fibrillation or atrial flutter, or discontinuation for
8 an adverse effect. And although this was not powered
9 to be an equivalent study, it's clear that these lines
10 are very near each other with this p value.
11 This is number of patients who are relapse
12 free. We begin with the number of patients who are in
13 sinus rhythm at six months on study drugs and the p value.
14 These are the number of patients who relapsed. These
15 are the number of patients who were discontinued for
16 adverse events. Seventeen percent in the quinidine
17 arm, 10 percent in the sotalol arm. There was one death
18 on quinidine due to a stroke and there was one death
19 on sotalol due to myocardial infarction.
20 You would expect that a drug that had beta
21 blocker effect as part of its electrophysiologic profile
22 to slow heart rate at the time of a rhythm relapse.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
48
1 And so another endpoint in this trial which was examined
2 was the mean resting ventricular rate when patients
3 relapsed back into atrial fibrillation or atrial
4 flutter. These are the data for d,l-sotalol.
5 Borderline statistical difference between the value for
6 relapse and baseline. These are the data for quinidine
7 baseline relapse. There was a highly statistically
8 significant difference between relapse heart rate on
9 sotalol versus quinidine with this p value by two sample
10 t-test.
11 Another unique part about this protocol is
12 that patients were interrogated for symptoms at
13 baseline, and they were then re-interrogated for
14 symptoms at one month after treatment. I would point
15 out that the Ns for these observations are lower than
16 the Ns for the patients that were actually randomized
17 into these arms because a patient dropped out of the
18 protocol either because of adverse effects or inefficacy
19 within this one-month time period. Twelve for sotalol,
20 24 for quinidine.
21 I would point out that symptoms of
22 palpitation and weakness decreased in both sides of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
49
1 study, both for quinidine as well as for sotalol. It
2 would be expected in patients who were achieving some
3 kind of a therapeutic effect with these antiarrhythmic
4 drugs.
5 DR. CALIFF: Peter, just to make sure I
6 understand. What you're saying is that, for people who
7 didn't have side effects, they had -- they looked better?
8 DR. KOWEY: They felt better if they were
9 in sinus rhythm.
10 DR. CALIFF: If they didn't drop out
11 because of side effects?
12 DR. KOWEY: Correct. That's correct.
13 CHAIRMAN PACKER: Let me see if I
14 understand. They were in sinus rhythm at the start at
15 the trial?
16 DR. KOWEY: Yes.
17 CHAIRMAN PACKER: They were at sinus rhythm
18 at one month and they felt better?
19 DR. KOWEY: No, no. The symptoms were --
20 the way they were interrogated at baseline was, What
21 were your symptoms when you were in atrial fibrillation,
22 not, What were your symptoms when you entered the study.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
50
1 CHAIRMAN PACKER: So it's not history.
2 DR. KOWEY: It was not concurrently with
3 randomization.
4 CHAIRMAN PACKER: It's not the baseline.
5 DR. KOWEY: It is not the baseline
6 symptoms. It's the symptoms the patient had before they
7 were treated and when they were in atrial fibrillation.
8 DR. KOWEY: Again, a further comparison.
9 These data you've already seen on a preceding slide.
10 I just want to point out that this is the proportion
11 of patients in the clinical trial who were reported to
12 have had adverse events. These are the
13 discontinuations. These are the values for patients
14 who actually had adverse events. Fifty percent in the
15 quinidine arm, 28 percent in the sotalol arm.
16 I want to just summarize, Milton, if I may,
17 with just a couple of slides and then we can answer
18 questions about efficacy.
19 What's done in this slide, and the two
20 succeeding slides, is look at the clinical trials that
21 I presented to you in each category and describe on the
22 slide the percentage of patients relapse free, the p
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
51
1 values for the log rank and the relative risk versus
2 control. This is for the chronic atrial fibrillation
3 and atrial flutter strata. And we're looking at
4 treatment versus control. I'd point out for Study H,
5 the control was not placebo; the control was quinidine.
6 This is for Study 004. These were also the
7 primary analysis, .56 with this p value. We do not have
8 a point estimate for the dofetilide experience. All
9 we have is the p value for the log rank which you can
10 see here. This is 014, which was the high dose study
11 showing you the relative risk versus control of the point
12 estimate. And this is the quinidine/sotalol comparator
13 study in which there was really not much to choose
14 between the two therapies.
15 This is an analysis for the paroxysmal
16 atrial fibrillation, atrial flutter cohort. I've
17 actually -- the reason why there's two slides for PAF
18 is because in the first slide, I'm showing you the data.
19 Rob, this is from randomization. So this is the from
20 randomization analysis, the relative risk versus
21 placebo point estimates and log rank for sotalol at 80;
22 d,l-sotalol at 120, and d,l-sotalol at 160. Remember,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
52
1 this could be once a day or twice a day in this study.
2 And these are the log rank p values.
3 This is the analysis for 9A at the low dose,
4 80 milligrams twice per day; and at the higher dose,
5 160 milligrams twice per day. Again, these are the
6 relative risk point values and confidence intervals.
7 Let me show you --
8 DR. CALIFF: These are again -- this is
9 censoring patients when they stop taking the drug?
10 DR. KOWEY: Yes, that's correct. That's
11 correct.
12 And let me just show you, Rob, the other
13 analysis which was from presumed steady state plasma
14 concentration for Study 5. So this was what was the
15 pre-specified analysis in the protocol. The relative
16 risk estimates are a bit different here. The p values
17 are a bit less small. And the reason, again, is because
18 there are patients who were lost during the early
19 phases of the trial. These data here are exactly
20 the same as what I've just shown you.
21 So I would just like to conclude the
22 efficacy portion of this presentation by pointing out
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
53
1 that d,l-sotalol extends the time to symptomatic
2 recurrence of these arrhythmias in patients with both
3 chronic and paroxysmal atrial fibrillation and atrial
4 flutter. It would appear that patients with and without
5 structural heart disease obtained a similar benefit.
6 The study had doses ranging between 80 and
7 160 milligrams twice per day in some of the trials, with
8 once a day dosing in patients with ultra creatinine
9 clearance. And these appear to be effective.
10 Conversion rates to sinus rhythm is 30
11 percent in Study 014. However, doses above what we are
12 recommending for clinical use were required in order
13 to achieve that clinical benefit. Dose dependent
14 increase in recurrence-free rate was seen in Study 05
15 which was the randomized comparison of dose.
16 That concludes my efficacy presentation,
17 and I'd be happy to take questions.
18 CHAIRMAN PACKER: What I'd like to do is
19 to pause here for questions from the committee. In all
20 cases, we're going to begin our questions with JoAnn
21 Lindenfeld who is the primary reviewer for this NDA.
22 I also think it would be very useful to take
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
54
1 the questions in a systematic fashion per study. And
2 so what I would ask the committee not to do is jump around
3 from study to study. We're going to go through all the
4 studies individually. Some of the studies have common
5 issues. Some of the studies have distinct issues. And
6 let me begin, as the briefing document does, with Study
7 05. So we're going to start the questions with Study
8 05.
9 JoAnn.
10 DR. LINDENFELD: I just want to start--
11 this will involve the entire discussion. Could you give
12 us a rough idea of the average age and the average percent
13 of women with atrial fibrillation in the United States?
14 DR. KOWEY: In the United States?
15 DR. LINDENFELD: Just what's the average
16 age of these patients and what percentage are women?
17 DR. KOWEY: I'm going to take a wild stab
18 at this, JoAnn. I don't know. I don't have precise
19 data, but I do know that it's an elderly population.
20 So this is a group of patients that should be greater
21 than 65 for the most part. And although men may have
22 more disease when they're younger because of their
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
55
1 coronary disease, women certainly catch up with them
2 and they have a higher incidence in the elderly.
3 DR. LINDENFELD: Yes. I think one thing
4 that will go through all of these studies is that this
5 is a relatively young population for this disease and
6 a relatively high percentage of men, I think, for atrial
7 fibrillation. And this becomes important because
8 creatinine clearance becomes so important with age, I
9 think.
10 DR. KOWEY: I agree.
11 DR. LINDENFELD: But that's just to start
12 off.
13 Now in terms of excluded drugs in Study 05,
14 I want to just address this issue of calcium blockers.
15 Am I correct in saying that dotiazam and verapamil were
16 excluded drugs?
17 DR. KOWEY: That is correct.
18 DR. LINDENFELD: And I think that is in all
19 of these studies; is that correct? That will become
20 an important point later on as we talk about adverse
21 effects and bradycardia. At least in 00 -- we'll stick
22 to 05, but I believe that's true in 004 as well.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
56
1 Let's come back to that because I think it's
2 important dotiazam would be a commonly used drug in this
3 population of patients.
4 DR. KOWEY: Can we have back up, please,
5 slide 190? This is, JoAnn, specifically Study 05,
6 concomitant therapy.
7 DR. LINDENFELD: These calcium channel
8 blockers, that excluded dotiazam and verapamil; is that
9 correct?
10 DR. KOWEY: That was in the protocol, yes.
11 DR. LINDENFELD: Because I think the point
12 would be that those would be relatively common drugs
13 that these patients might be taking. So just an
14 important point for the future.
15 And is it also true that the therapy was
16 blinded but the dose was not? In other words, the
17 physicians and patients didn't know which therapy, but
18 the potential dose of therapy was known?
19 DR. MARROTT: In 35, that is correct.
20 DR. LINDENFELD: So just as a point of --
21 people might know that the dose was higher but that would
22 apply to both, of course.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
57
1 Now I want to get to this issue that I think
2 everybody wants to get to about the dropouts and we'll
3 come back to the dropouts. In including the dropouts,
4 if a worse case scenario is included, that is all the
5 dropouts are considered failures, then the study is
6 non-significant, at least, I think, according to the
7 FDA analysis. And the analysis that we read suggested
8 that perhaps the truth lies somewhere in between.
9 But I'm a little bit concerned, and I want
10 to get some opinions from everyone because I'm a little
11 bit more concerned than what I saw in the FDA document
12 that the people who drop out may actually be the people
13 -- people who drop out on sotalol for adverse events
14 may actually be the people at highest risk of recurrence
15 for atrial fibrillation. Particularly, I know, that
16 at least in a few of the studies when it was documented,
17 those were clearly more often elderly people. So maybe
18 we could have some comments on that and maybe from the
19 committee, too. I'm concerned that actually the worse
20 case scenario may apply here.
21 CHAIRMAN PACKER: Let me just outline what
22 the issue is so that it is clear to everyone what we're
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
58
1 talking about because, as JoAnn says, this is an issue
2 which is on the minds of I think every member of the
3 committee, as well as noted as an important issue in
4 the FDA review.
5 And that is, in Study 05 as well as in nearly
6 all the trials presented on behalf of sotalol, patients
7 who discontinue the drug were not observed to the end
8 of the planned therapy for the occurrence or recurrence
9 of atrial arrhythmias. Consequently, we do not know
10 whether patients assigned to a specific dose of sotalol
11 or placebo had a recurrence of atrial fibrillation.
12 In other words, the data were censored at the time of
13 discontinuation, and we are all concerned that that
14 censoring is informative. That is, it's not random,
15 that censoring was not random.
16 The FDA reviewer had asked the sponsor to
17 try to gauge the degree of difficulty created by this
18 by including the time of discontinuation in the
19 analysis. And this was done for both treatment arms.
20 This is referred to as the so-called "worst case"
21 analysis.
22 Let me just make a comment here. This is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
59
1 not a worst case analysis. A worst case analysis would
2 be to censor all of the placebo and to assign events
3 at the time of discontinuation to all the patients
4 receiving active therapy. That would be worst case.
5 So an analysis in which all patients are
6 considered to have an event at the time of
7 discontinuation is not a worst case analysis.
8 DR. KOWEY: Right. Because it will pick
9 up the even worse.
10 CHAIRMAN PACKER: It could even be worse
11 than that.
12 DR. KOWEY: Right.
13 CHAIRMAN PACKER: And therefore, the
14 analysis presented here is not the most conservative
15 analysis. And one could be more conservative than the
16 analysis being presented. But I think it's important
17 to talk about this because it has implications not only
18 for sotalol but it also has implications for almost every
19 long-term trial this committee sees, for any drug, for
20 any indication.
21 And it is also an issue that is brought up
22 in the committee questions. And so I would like to take
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
60
1 JoAnn's lead here and have the committee spend a little
2 bit of time on this because it is so important. And
3 let me do so by --
4 Rob, do you want to comment on this?
5 DR. CALIFF: I can just make a few comments
6 because in general, I think, that whenever we
7 discontinue follow up in patients who have been
8 randomized, then we have a violation of the intention-
9 to-treat principle. And we're left with some
10 uncertainty about the implications that has for the
11 analysis.
12 I certainly agree -- Lloyd, just sit down
13 for a minute here.
14 Someone mentioned that Dr. Fisher was going
15 to be on the edge of his seat within milliseconds and
16 indeed he is.
17 I certainly agree that when a few patients
18 get lost, then that would be a reason to censor. But
19 I think the problem that I'm seeing is that trials are
20 being designed where, by design, patients are no longer
21 followed when they stop taking the drug, which I think
22 is a very dangerous approach in doing clinical trials
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
61
1 but it seems to be the norm rather than the exception.
2 Now one could also argue, and I think it
3 has been reasonably argued, that in practice you try
4 a drug and if the patient has a side effect, stop the
5 drug and try something else. So it's likely that the
6 right answer is somewhere in between but it's certainly
7 not -- the right answer is certainly not at the point
8 of censoring because I think as you and JoAnn pointed
9 out, the patients most likely to drop out -- and this
10 is another thing that worries me -- are not only the
11 ones most likely to fail therapy but in the case of the
12 drug that may cause toxicity -- most particularly
13 related to things like renal function, drug
14 accumulation, electrophysiologic property -- that
15 toxicity is likely to be very much concentrated in a
16 very small group of patients who are at high risk.
17 And so it leaves you uncertain about judging
18 both the efficacy and safety, I think, of what will
19 happen when this thing is unleashed on the public.
20 CHAIRMAN PACKER: Now, Lem, do you want to
21 comment?
22 DR. MOYÉ: Yes. I think that sometimes
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
62
1 investigators don't really know what they're getting
2 into when they start a clinical trial. Much of the
3 emphasis is placed on randomization. There's a
4 tremendous full-court press to randomize patients. And
5 sometimes what gets lost is that when a patient is
6 randomized, that investigator essentially buys that
7 patient for the duration. Essentially, the study pays
8 a price for having that patient enter into the study
9 because the study analysis assumes that patient is going
10 to be followed until the very end of the experiment.
11 Now sometimes people are fooled by the fact
12 that we randomize so many patients in this study.
13 Sometimes we randomize thousands or tens of thousands
14 of patients. And there tends to be a sense that there's
15 some play in these numbers, that because you randomize
16 so many patients, you can afford to lose a few and still
17 not wind up vitiating the findings.
18 This is a trap. This is a great trap.
19 Because the findings in the end come down to the delta,
20 the difference in the number of patients who have the
21 endpoint in the placebo group versus the number of
22 patients who have the endpoint in the active group.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
63
1 And even though you may randomize hundreds of thousands
2 of patients, the delta winds up being five or ten, or
3 15, or 20 patients. So the entire efficacy of the
4 study pivots on what happens to those ten or 15 people.
5 If the investigators, of course, knew who
6 these people were in the beginning, they would give them
7 tremendous care. But the investigators don't. So the
8 best that they can do is treat each patient like that
9 patient is the patient that's going to make the
10 difference. That translates to following everybody for
11 as long as you can, or certainly for the duration of
12 the experiment, perhaps longer if possible.
13 If that does not happen, you have what, to
14 me, is a discordancy. That is to say, that the protocol
15 essentially specified there would be one mode of
16 execution and in fact the actual execution was
17 different. Now, in some sense, the investigators have
18 let us down because we haven't been able to -- we cannot
19 look at the data as we would have expected to see it
20 from the protocol.
21 And so the question then becomes, Can you
22 make some kind of adjustment? Well, here you really
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
64
1 can't make a persuasive adjustment. I think that
2 probably Lloyd and I could spend all day throwing
3 scenarios back and forth at one another about what would
4 be reasonable and what is not. Essentially, the
5 computation for the effect side and the p value is beyond
6 adjustment. In my word, it's corrupted. There's no
7 way you can compute the p value which actually assesses
8 what is the truth in this experiment in that what --
9 by that I mean, what it actually mirrors what it tells
10 us about the population.
11 The best we could hope for, of course, which
12 we don't have here, is the absolute worst case analysis,
13 and I second what Milton's comments were here, and that
14 is we assume the active patients who are lost were the
15 ones who had the bad clinical outcomes and the placebo
16 patients do not.
17 If you don't have this extreme worst case
18 analysis which lines up with the initial analysis, I
19 think that we must go by the most conservative analysis.
20 And the most conservative analysis here is that the
21 p values in fact are not significant. This is a
22 conclusion that I am reluctant to reach. However, since
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
65
1 the investigators are not able to follow patients as
2 they had initially planned, then I think in terms of
3 -- since the implications of what we decide here are
4 not just for this trial, but the implications are for
5 what the side effects are going to be in the community,
6 the most conservative approach here I believe is the
7 best one.
8 CHAIRMAN PACKER: Yes. Maybe, let me just
9 see if I can get a clarification.
10 Lem, you're suggesting that the
11 investigators sort of violated a commitment to the
12 trial. But if I understand correctly, the trial
13 protocol actually said they wouldn't be followed. So
14 it wasn't the investigators violating their commitment.
15 It was the design of the study that encouraged the lack
16 of follow up in the patients who dropped out because
17 of an adverse event.
18 Is that correct?
19 DR. KOWEY: That's correct.
20 CHAIRMAN PACKER: But a different
21 philosophy was followed for Study 04. Why were they
22 different?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
66
1 DR. KOWEY: 04 was a study that predated
2 05 and was not done by the sponsor. And I don't have
3 an explanation for why there was a change in the
4 philosophy of follow up. I wasn't privy to that.
5 DR. KONSTAM: Could we get -- I'd just like
6 this clarified. The analyses that we saw for 004 was
7 a true intent-to-treat analysis without censoring of
8 dropouts?
9 DR. FISHER: Could I make one comment about
10 terminology? I won't go into my other comments. But
11 I think it would be useful, the term "intent to treat"
12 means everybody's included in the group to which are
13 randomized. And I would maintain that -- for example,
14 you will later see an ICD trial. I don't actually mind
15 an ICD trial which considered a person to have an
16 endpoint at the time there's a discharge for VF to
17 consider that the equivalent of an endpoint had they
18 not had a defibrillator. I would call that "intent to
19 treat," although you can follow them further for
20 subsequent discharges.
21 So I would like to distinguish between
22 "intent to treat" where everybody is included in the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
67
1 groups to which they're randomized and maybe we can call
2 it "complete follow up", being even if they discontinue,
3 they go to the end of the study period, just for logical
4 consistency because I can think of situations where I'm
5 relatively happy with discontinuations and others where
6 I am not.
7 DR. KONSTAM: So in 004, all patients,
8 whether they were discontinued or not, are included in
9 the efficacy analysis that we saw; is that correct?
10 And in 05, the primary efficacy analysis
11 excluded patients who were excluded because of adverse
12 events, discontinued drug because of adverse events.
13 DR. MARROTT: In Study 004 -- can you hear
14 me now? In Study 004, all patients were followed until
15 the end of the trial. In Study 05, patients were not
16 followed if they discontinued due to side effects or
17 if they had a relapse. The sponsor's point of view was
18 that being a fixed-dose trial, we expected a larger
19 number of side effects; and it was difficult to ask the
20 investigator that the patient be followed because it
21 is not easy to follow patients who have discontinued
22 in the atrial fibrillation population because they would
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
68
1 have gone on to other treatments and other types of
2 management.
3 CHAIRMAN PACKER: It is actually easy to
4 follow them. It is very easy to follow them for the
5 planned duration of therapy. It is not hard to do the
6 right thing. What is hard is to accept the consequences
7 of what happens after the discontinuation of treatment.
8 Because what we're really talking about
9 here is not an intention-to-treat issue. It's an issue
10 of informative censoring, whether the censoring here
11 is informative or non-informative. And I guess that
12 would be the correct terminology. What we're
13 concerned about here is not only is the censoring
14 informative, but it is frequent and it was planned.
15 Udho.
16 DR. THADANI: There are several issues
17 which come to mind here. Investigators in some of the
18 trials have medical knowledge. For example, Karl
19 Rillow's study got around this issue by dropping
20 patients who had adverse effects during open label
21 phase. And those patients were dropped out. So that's
22 one way to do the trial. Unfortunately, this trial was
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
69
1 not designed by this, so I think you're stuck with it.
2 You randomize a patient. You will count them all.
3 Had you taken the, say, one month period and done a study
4 in those patients with side effects were the reason they
5 were not randomized, you would not be arguing with this.
6 Also, there are problems there, too.
7 CHAIRMAN PACKER: Udho, I think --
8 DR. THADANI: I'm just mentioning --
9 CHAIRMAN PACKER: You might be fixing a
10 problem with another problem.
11 DR. THADANI: I'm not fixing. I'm raising
12 some of the issues. So I think what I'm trying to say
13 is that once you randomize, intent to treat analysis
14 should include all the patients. And you might have
15 been actually benefited had these patients who had side
16 effects gone on to other therapy, which is helpful in
17 a fib, might have had less arrhythmia.
18 You could not have stopped the provisions,
19 and so you said you that you did not follow because it's
20 hard to follow patients. I don't think I buy that.
21 I think probably you would have come out beneficial had
22 they gone on to ALD drugs which also prevent a fib.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
70
1 But that is still intent to treat and that's a real
2 medicine.
3 So I don't think the FDA reviewing all the
4 files and what is presented, it's not a worse case
5 scenario hasn't identified it out. I think that's
6 intent to treat. So you're going to have to live with
7 it. One is a bit shaky and I sympathize with your
8 patients. The trial was written that way and the
9 investigator didn't follow that.
10 But, I'm concerned that intent to treat did
11 not show a difference and if you drop the patient --
12 and it seems like the higher doses, the higher drop out
13 rate. And since the 80 milligram did not work, you are
14 recommending 120 or a higher dose with, unfortunately,
15 a higher drop out rate. And it's possible that had these
16 patients been seen at months 3 to 6, they would have
17 had a higher recurrence and they would have neutralized
18 the effects.
19 So I think there are some concerns and all
20 of us have concerns when we looked at the FDA documents
21 as well as the database.
22 CHAIRMAN PACKER: Lloyd.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
71
1 DR. FISHER: I wonder if I could make a few
2 comments. First, it's wonderful to see Dr. Lipicky back
3 and he has gone home to watch our festivities on the
4 internet. And hopefully he's there and hears this.
5 I was actually glad to see this point
6 brought up and I think it deserves very careful
7 consideration. And I would take a very parochial view.
8 I view the cardiorenal committee as the best division
9 and committee within the Bureau of Drugs. And I've now
10 had enough experience with other committees that that
11 may be a true statement. I certainly haven't seen every
12 committee.
13 First, I'd like to note that the question
14 the FDA stated is not, strictly speaking, correct. The
15 question talks about the assumptions being violated
16 because there's different drop out rates at different
17 doses. That in itself is not enough to invalidate the
18 censoring.
19 What has to happen is you have a
20 differential treatment effect, as Milton implied,
21 associated with this censoring. So the real issue is
22 how differential that treatment effect would have been.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
72
1 How robust are the findings. And what Dr. Moyé was
2 talking about is absolutely correct. You're left with
3 mathematically what's called an unidentifiable problem.
4 You can hypothesize different sorts of scenarios, none
5 of which you can tell from the data once you have the
6 censoring. It's sort of like a Zen poem, sound of one
7 hand clapping. What would have happened if these people
8 could have tolerated the drug and taken it.
9 And I'd like to note also that many areas
10 where we usually don't think about this problem it truly
11 exists. And that is where we measure things
12 continuously and have our last observation carried
13 forward analyses. When the people go out, those
14 observations definitely could have changed also. So
15 this is a problem that cuts across almost every drug
16 development area.
17 I think when take Dr. Moyé's hard line, as
18 he tends to take on each and every issue, we will kill
19 drug development in all kinds of areas. It just will
20 not be conceivable because of the tolerability of drugs
21 to get a positive study.
22 So I think it would be very poor policy to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
73
1 make discontinuations, or even worse, the worse case,
2 a primary analysis that had to be satisfied necessarily
3 for approval, despite all those caveats I've just
4 spread. Normally what we do in observing the amount
5 of discontinuation is, if it's quite low, we tend to
6 ignore it. And in the past, even when it's somewhat
7 moderate like this, we tended to ignore it perhaps
8 inappropriately.
9 Parenthetically, I push the sponsor, and
10 Dr. Pritchett who is here said I shouldn't introduce
11 his name in my comments, why weren't these people
12 followed up. And he assured -- I don't want to start
13 a big fight among the Duke medical faculty, but he
14 assured me that this was not possible and I'm mentioning
15 him, hopefully, hoping that he will give this comment
16 first-hand because I think it's very relevant in this
17 particular discussion.
18 Anyway, what you have to do, then, is use
19 your judgment of the biological plausibility of what's
20 gone on. I would first note, if we could get slide 195.
21 For this particular study, the discussion is going as
22 if this analysis destroyed everything. But in fact,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
74
1 you can see the 120 p value of .034, the sponsor said
2 they would look at the two higher doses and use a
3 Bonferroni correction, which is a little too
4 conservative, of .025. So this is not significant but
5 it's not as if you've destroyed the whole study. And
6 certainly, you would have to agree there's a very strong
7 trend, if not significant. And I think that's important
8 when you integrate all this data in your mind, that you
9 not think of this as a study where when you considered
10 the discontinuation failures, everything fell apart.
11 That's not true. What did happen is the statistical
12 significance dropped.
13 So what you have to do is use your biological
14 judgment of plausibility and you people are the medical
15 and biological people and I'm the statistician. But
16 I would suggest that there's a lot of comfort in 04 which,
17 because they had upward titration or because the
18 Europeans are more stoical, I mean, there are a variety
19 of reasons. But they had very few dropouts. And they
20 have the same type of pattern.
21 So there was a case where this problem
22 didn't enter in and the data are somewhat consistent.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
75
1 And there's a number of other ethics to these studies
2 that you heard about from Peter. So what you would have
3 to do in your own mind is make a judgment about just
4 how robust the findings are, how much the
5 discontinuation might have effected things, and it's
6 hard for a statistician. If we take a very hard line,
7 we would take the absolute worse case. I think in
8 practice, actually I consider this actually a fairly
9 draconian correction and I would agree with, I think,
10 Dr. Califf's comment, the truth is somewhere in between
11 this and the center's situation. And precisely where
12 that truth lies is a very difficult matter of judgment.
13 DR. THADANI: Rob, before -- sorry, I
14 believe it's 05, this is 05, right? Right. If I'm
15 looking at the FDA document, I believe on page 10,
16 there's a small booklet. The p value there is point--
17 is different than what you're showing here.
18 DR. FISHER: I don't have the document.
19 DR. THADANI: Point 042. Whether the
20 statistician reviewed the work as which is intent to
21 treat, he found no significance at all. And p values
22 there are -- given are much different. Log rank is .62.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
76
1 DR. FISHER: I think the difference is the
2 --
3 DR. THADANI: Why there are differences in
4 that?
5 DR. FISHER: Okay. I think the difference
6 is the following. The investigators preferred
7 analysis, which is not my preferred analysis nor I think
8 the agency's, was from steady state. Because the theory
9 being until you reach steady state, what could happen.
10 If you -- So they probably have this value.
11 If you do it from steady state, because some
12 people drop out before steady state, you lose the
13 protection of the randomized process, just as we do for
14 the discontinuation.
15 DR. THADANI: No, this isn't
16 randomization. All patients were included and the p
17 values for log rank is point -- this is on page 10 of
18 the document, p value is .62 for 80, .098 for 120.
19 DR. FISHER: Yes, that's --
20 DR. THADANI: And .912 for this. And if
21 you look at it against it's .042. It's not what you're
22 showing here with significant. But there is no
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
77
1 significant --
2 CHAIRMAN PACKER: Udho, those numbers are
3 the same numbers that are on the slide. Can you just
4 check?
5 DR. FISHER: Oh, no, I had a prior -- I
6 apologize. I can explain the difference.
7 The difference was in the slide that I
8 hastily picked up, did not refer to symptomatic. But
9 it referred to any ECG documented. And .098, which I
10 still say is the symptomatic return on a randomization.
11 I apologize. I didn't do that knowingly
12 with knowledge.
13 CHAIRMAN PACKER: I think we have to --
14 there's two separate issues here. The issue number one
15 is what does the committee think the right kind of design
16 and analysis should be as a general issue. The second
17 is, how does that feeling influence the interpretation
18 of the data on sotalol. And I think, Roy, I think you've
19 actually said that.
20 DR. FISHER: I would just suggest, before
21 you begin your general assessment, we hear from Dr.
22 Pritchett about the possible practicality. I mean,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
78
1 maybe his views are incorrect but --
2 DR. CALIFF: But, Lloyd, I would at least
3 like you to make a statement as to what you think the
4 proper study design is, if at all possible.
5 DR. FISHER: If at all possible, the proper
6 design, the advice I give to people, is you follow them
7 until they -- everybody, even if they discontinue their
8 study drug, they're in the study. You follow them to
9 the end of the study and you try to select all the interim
10 observations. If it's a survival trial, that's usually
11 very possible. It may be a little difficult depending
12 upon the reason that people discontinue things, and so
13 on, and --
14 DR. CALIFF: Not to sound like a lawyer,
15 but it can be difficult but it should be the goal to
16 follow every patient to the end of the --
17 DR. FISHER: Well, I don't know. I mean,
18 I'd like to hear from Dr. Pritchett who has given me
19 --
20 DR. CALIFF: He can say all the reasons why
21 it's hard to do it, but as a principle, it should be
22 done.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
79
1 DR. FISHER: If you can -- to the extent
2 you can do it, you should do it. I agree with that.
3 CHAIRMAN PACKER: I don't know if we want
4 to get into extreme detail as to why it's hard and I'm
5 sure, Ed, you would tell us why it's hard. But, I'm
6 not certain it would change the underlying principles
7 that are important here which apparently, from what I
8 can tell, there's unanimity of opinion. And that is
9 that there should always be a concerted, systematic
10 effort to follow all patients until the planned end of
11 the duration of therapy for all factors related to the
12 primary or secondary analyses.
13 And although it could be difficult, and in
14 fact doing so might end up diluting the treatment effect,
15 that is the most interpretable way of looking at the
16 data from any trial. It's a principle which has been
17 exceedingly well established for mortality analyses.
18 And all I think we're saying, and Lloyd, you're
19 agreeing, and I don't hear anyone disagreeing, is that
20 if it's good enough for a fatal outcome, it's good enough
21 for non-fatal outcome.
22 DR. FISHER: Well, I have no problems with
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
80
1 a perfect world. All I'm saying is, I think you have
2 to consider it by particularly the reality of what you're
3 doing. I'm not -- You can judge better than I can.
4 CHAIRMAN PACKER: Lloyd, we're not looking
5 for perfection. What we're looking for --
6 DR. FISHER: I understand perfectly. As
7 the guideline.
8 CHAIRMAN PACKER: What we're looking for
9 is to send out a -- the strongest possible message that
10 to design trials so that data are systematically omitted
11 is wrong.
12 DR. FISHER: No, I -- In fact, I have said
13 the same thing, not to the sponsors, not to the sponsor.
14 I was not around when these studies were planned and
15 so on, so I don't -- And one of the replies I get, well,
16 why do that? They're off the drug. They're
17 discontinued. We're going to be censored, then, and
18 people accept that. And I think it is a good thing to
19 change that perception and in the future do things.
20 And it would have been nice if this trial had been done
21 the same way.
22 DR. THADANI: Lloyd, also, I think that if
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
81
1 you allow censoring and excluding patients, we are
2 saying why bother looking at intent to treat and might
3 as well throw it down the drain. Because what you're
4 encouraging then the patient drops out, don't have to
5 follow them. So, I think we can't set the principle
6 that patients who have side effects withdraw them from
7 the study are not counted, even when you're shown they
8 are not statistically significant at the end it's
9 included. So, I think we must insist that once a patient
10 in a trial, whether it's a mortality trial or not a
11 mortality trial, they should be followed. And what
12 happens, other treatment does it, I think that's
13 different. But otherwise, we would be violating all
14 the rules and at the end as a clinician, I would not
15 be sure. I know, my practice would not change if I have
16 the trial results or no trial results. I'd be doing
17 the same because I would not be any better off looking
18 at this result or any other result if you're censoring
19 patients.
20 DR. FISHER: Well, just to complicate the
21 issue, it is not always fair what is the best analysis.
22 Let's say you're doing an equivalence trial and
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
82
1 everybody who discontinues, then gets put on the drug
2 you're trying to show equivalence to. And then you do
3 the intent to treat analysis and take the entire time
4 period where an awful lot of the people are on the drug
5 to which you're trying to show equivalence. One would
6 argue, then, that probably the best analysis is to censor
7 at the time they went on the actual drug to show an
8 equivalence.
9 So, this is a complex issue where it's hard
10 to make across the board general statements. But as
11 a general principle, I think it is valuable, I agree,
12 to follow up for endpoints as best you can after people
13 discontinue study medication.
14 DR. CALIFF: Yes, I guess my point is we're
15 cheated of the opportunity to deal with the uncertainty
16 in the most rational way if we don't have all the data.
17 I would recognize that the answer is somewhere in
18 between.
19 DR. FISHER: And I certainly agree with
20 that statement.
21 CHAIRMAN PACKER: We rarely achieve
22 unanimity of opinion amongst the members of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
83
1 committee or with the committee and sponsor. This is
2 a special moment.
3 DR. FISHER: We haven't heard from Dr.
4 Pritchett yet.
5 CHAIRMAN PACKER: Ed, is it okay if we just
6 say it's hard or do you want to tell us how hard?
7 DR. PRITCHETT: Rob, I think that's the
8 point, that it's hard. If you're following a mortality
9 endpoint, you can get that always, even from the national
10 death registry if you need it. If you're following some
11 non-fatal endpoint, particularly one like documentation
12 of a symptomatic arrhythmia that required a certain
13 amount of cooperation from the patient, it is much more
14 difficult. And I think that that is one of the
15 principles that has guided the design of these trials
16 as we've worked on them over the past several years,
17 along with the attempt that the outcome of the trial,
18 the conclusions drawn from the trial, could mimic the
19 way clinicians think. And I don't think that clinicians
20 equate a recurrence of an arrhythmia with “I
21 discontinued the drug because the patient didn't like
22 it.” Just like I don't think physicians equate a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
84
1 recurrence of an arrhythmia on the second day of therapy
2 when the patient has only had a couple of doses of
3 medication with one that occurs two or three weeks later
4 when they're at steady state.
5 So, these trials were designed to mimics
6 the clinical practice as well as to deal with the
7 practical difficulty trying to document an outcome in
8 patients who won't cooperate with you any more.
9 DR. CALIFF: But there are clinicians
10 don't, when a patient fails a drug, don't say I'm going
11 to forget about you for the rest of your life. They
12 actually follow the patients after that, maybe on
13 another therapy. So, I think your analogy is actually
14 flawed in that respect.
15 DR. PRITCHETT: I think that sometimes the
16 physicians following patients in clinical trials are
17 the physicians who follow the patients forever and other
18 times they're not. So, I think we do see many patients
19 in clinical trials that, once they're finished with the
20 clinical trial, are lost to the investigator after they
21 lose their ability to document the outcome.
22 CHAIRMAN PACKER: I think we're actually
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
85
1 all saying the same thing. I think that we're saying
2 that to systematically design the trial so that the
3 information is not collected is not a good idea. That
4 every effort must be made to obtain the information and
5 we understand it would be hard to do so under some
6 circumstances. And some trials may lend itself to more
7 completeness, to greater completeness, of data than
8 others. And certain indications might lend itself to
9 that.
10 But, I think we're -- it's really important
11 to emphasize the principle. Having emphasized the
12 principle, maybe we should go back to sotalol.
13 DR. KOWEY: Just one point about what could
14 have been done in this study was with something that
15 was like what we have done in 004, which was to allow
16 a titration down on the dose for the patients who have
17 been randomized for maximum dose. That's a much more
18 conventional way of handling it. Then you do your first
19 period analysis. You have your data. And the patient
20 may be able to stay on this drug and you can follow
21 forever its effects. And that probably should have been
22 the design and unfortunately we all agree that it wasn't.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
86
1 But it was -- Just to make sure that
2 everybody understands that the intention of the study
3 was a dose ranging study and patients report on a dose,
4 and they had no option once they were at that dose.
5 CHAIRMAN PACKER: But, Peter, even in
6 normal, usual, conventional dose ranging studies, the
7 dose to which a patient is randomized is not considered
8 to be an all or nothing phenomenon. It's usually
9 considered to be a target dose. It's the intention to
10 achieve that dose, not the perfection of achieving that
11 dose.
12 DR. KOWEY: I agree.
13 CHAIRMAN PACKER: So, this trial is doubly
14 confounded in the sense that there was no follow up after
15 discontinuation and that in fact clinical practice was
16 subverted here. There would normally be down
17 titration. One normally designs dose response trials
18 with an intended target as opposed to, “gee, if you don't
19 take this dose, you're out.” So, there are lots of
20 things that could have been done better with this trial.
21 DR. KOWEY: Which was really what was done
22 in 9A, which allowed the people to have one or two doses
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
87
1 in a titration mode trial.
2 CHAIRMAN PACKER: JoAnn, while we continue
3 with comments on 05.
4 DR. LINDENFELD: This is just a point of
5 clarification. I think when drugs have beta blocker
6 effects and we measure time to symptomatic recurrence,
7 one has to wonder if that's because the -- when they
8 recur, the rate is slowed and they have less symptoms,
9 or in fact the duration is prolonged.
10 So, can you, just to clarify for me, show
11 me the difference in this study in terms of time to
12 symptomatic recurrence and time to EKG documented
13 recurrence? Were those substantially different?
14 In other words, I believe when I look at
15 this, that if one just takes the EKGs that were measured,
16 I think by telephone line every two weeks, that the study
17 was less significant if one measured the time to EKG
18 documented recurrence irrespective of symptoms. But,
19 I just want to be sure that's a correct statement.
20 DR. KOWEY: This is for the symptomatic.
21 Can we have all ECG -- all A fib flutter recurrences
22 for 05. I don't believe that we have that analysis in
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
88
1 our back ups. We'll look for it, JoAnn.
2 DR. LINDENFELD: I just want to know if some
3 of this beta blocking as opposed to the other.
4 DR. KOWEY: We do have that analysis for,
5 you saw, for 04. And there was really no difference
6 between the two analyses. That is, no real big
7 differences.
8 DR. LINDENFELD: I thought 04 was less
9 significant for the EKG as opposed to --
10 DR. KOWEY: We show you the two for 04 if
11 you'd like.
12 DR. LINDENFELD: We can probably come back
13 to that.
14 DR. KOWEY: Would you like to see those?
15 Yes?
16 DR. THADANI: JoAnn, while on this point,
17 this slide really shows the patient had a symptomatic
18 and then were documented to be an A fib on ECG. I didn't
19 see any data when I was reading that you might have that
20 is data showing that repeat ECGs were done was recurrence
21 rate. I didn't see that either. These are patients
22 who complain of symptoms and they were collaborated to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
89
1 be an A fib and the ECG showed that.
2 Is that correct, Peter, or they were
3 symptomatic and then you happen to do the ECG. And I've
4 not seen any in my reading. You might have seen it
5 because you are the primary reviewer, but I didn't.
6 DR. LINDENFELD: No, I think that's really
7 what I'm asking because I think that patients are less
8 likely to be symptomatic in atrial fib if they're on
9 sotalol than placebo. So, less likely to be picked up.
10 DR. KONSTAM: Could we get this clarified
11 some more. So, if the patient was symptomatic, the
12 first endpoint, symptomatic A fib, I assume, and maybe
13 you can correct me if I'm wrong, that, then, was
14 confirmed by ECG?
15 DR. KOWEY: Correct.
16 DR. KONSTAM: So, the symptoms, what we're
17 calling symptomatic A fib was symptomatic, then
18 confirmed by electrocardiogram?
19 DR. KOWEY: Correct.
20 DR. KONSTAM: Now, this is -- now, this one
21 is different. How -- were you doing screening
22 electrocardiograms? That's what this is.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
90
1 DR. KOWEY: Right. That's in the 04 study
2 which we already showed.
3 You have the slide number? Number 24.
4 DR. LINDENFELD: People need to use the
5 microphone here so that we can get this.
6 DR. KOWEY: And go forward one.
7 DR. KONSTAM: Well, I guess I'm asking for
8 clarification about what these endpoints mean.
9 DR. KOWEY: One more. And then one more
10 after that. This is any atrial fibrillation, atrial
11 flutter, even without symptoms. The patients were
12 being periodically monitored in the absence of symptoms.
13 DR. KONSTAM: Can --
14 DR. KOWEY: We don't have the data for 05.
15 DR. KONSTAM: Well, wait a minute. Can you
16 tell us more about that? In other words, tell us about
17 these screening EKGs. Tell us whether there were
18 additional EKGs done if the patients were symptomatic.
19 I think I'm concerned about the same --
20 DR. KOWEY: Yes. We'll go through that
21 methodology for you.
22 Ed.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
91
1 DR. PRITCHETT: This study, this is the one
2 that was done in Scandinavia, Europe. It did not use
3 trans-telephonic monitoring. It used -- patients were
4 asked to come back periodically. If they had symptoms
5 of atrial fibrillation, they could come back to the site
6 to have an ECG recorded. They were also asked to come
7 back at specific months of follow up. And if they were
8 symptomatic and had an atrial fib, they went into the
9 symptomatic. If they were asymptomatic or symptomatic,
10 they wound up in this analysis.
11 DR. KONSTAM: So, this endpoint includes
12 the time when patients showed up?
13 DR. KOWEY: Yes. This is any atrial
14 fibrillation.
15 DR. KONSTAM: So, it's still, I think-- I
16 think the concern that JoAnn raised still stays. I
17 think that this is influenced by the fact that patients
18 are experiencing symptoms. Then we have to get into
19 interpretation about the implication with regard to
20 heart rate.
21 But, it doesn't dispel the problem, I don't
22 think, based on what I'm hearing about how this endpoint
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
92
1 was derived.
2 DR. GRABOYS: Can I --
3 CHAIRMAN PACKER: Yes, Tom.
4 DR. GRABOYS: Let me just comment on that.
5 Those of us taking care of large volumes
6 of patients with AF realize that a large fraction of
7 patient with this rhythm present with stroke. I think
8 it's extremely soft data in terms of what the actual
9 time from the recurrence of the arrhythmia is based on
10 clinical grounds in which frequently we see patients,
11 whether or not their rate is slowed or not by concomitant
12 therapy or not, who present in atrial fibrillation
13 having never had any kind of symptoms.
14 So, it's, to me, disquieting and very
15 difficult that we're basing judgments upon data
16 potentially coming from very soft sources.
17 CHAIRMAN PACKER: This -- let me see if I
18 understand it. The concern, Marv and Tom, just so we
19 can clarify this specifically, is indeed a sampling
20 approach in the trials. Are you convinced that either
21 the 05, and I hate to move into 04 because we're not
22 quite there yet, but it's a related question, that either
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
93
1 05 or 04 did it the way you think it should be done?
2 In other words, in order to avoid the issue of bias?
3 DR. KONSTAM: I'm not sure what you're
4 asking.
5 JoAnn raised the point that, the question
6 is, are we actually -- if I may paraphrase, are we
7 actually looking for -- looking at recurrences A fib
8 or are we looking at something else, perhaps, and
9 recurrence of a fib driven by the fact that patients
10 had to be symptomatic to be identified. This drug has
11 a beta blocker effect. It slows, we believe, it slows
12 heart rate at time of recurrence. Therefore, a
13 recurrence of atrial fibrillation may tend to be less
14 symptomatic than if the drug were not on board.
15 DR. GRABOYS: And therefore, what may
16 happen is if you've got a patient presenting who is now
17 symptomatic but in fact had the onset of atrial
18 fibrillation, a good bit of time prior to when they were
19 symptomatic.
20 CHAIRMAN PACKER: But I don't understand
21 how -- I understand the issue that you're bringing up
22 and I don't disagree with the issues. I'm just trying
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
94
1 to figure out how would the sponsor have fixed this
2 problem?
3 DR. GRABOYS: Well, short of continuous
4 monitoring, there's no way he can -- there's no way that
5 we can be assured that this data is in fact what it is.
6 CHAIRMAN PACKER: But specificity --
7 DR. GRABOYS: It is only dependent upon
8 symptoms --
9 CHAIRMAN PACKER: No, I understand. But
10 you can say this is a problem here with the approach,
11 the sampling approach. But I'm just trying to figure
12 out, is there an approach they should have used.
13 DR. KONSTAM: Let's, before answering that
14 question, I mean, let me just point out. This isn't
15 generic to all antiarrhythmics, necessarily, because
16 this agent has a beta blocker effect. So, I think it's
17 more important in this drug than perhaps in other drugs,
18 first of all. And, secondly, it is a very important
19 problem for the reasons that Tom raised.
20 Now, in terms of how it might have been dealt
21 with, I don't, maybe JoAnn had an answer to that.
22 DR. LINDENFELD: Well, didn't 05 look at
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
95
1 monitoring every two weeks? Is that correct?
2 DR. MARROTT: Yes.
3 DR. LINDENFELD: And 004 was just
4 symptomatic recurrence plus a much longer duration of
5 routine follow up. So, if we could have the same data
6 for 05, I think we all probably would give a couple of
7 weeks as a reasonable time.
8 DR. KONSTAM: Well, I'm not sure about it
9 but that's why I was asking about the nature of this
10 ECG derived endpoint in 05.
11 DR. MARROTT: Mr. Chairman, if you don't
12 mind, I would just like to clarify. In both studies,
13 004 and 05, there's a TTM when the patient was
14 symptomatic. But there were routine TTMs performed
15 every two weeks. In addition to that, of course, when
16 the patient visited the outpatient clinic, there was
17 an opportunity to see what the rhythm was.
18 DR. LINDENFELD: So, in 004, I thought you
19 said earlier that this wasn't done every two weeks.
20 It was done every two weeks?
21 DR. MARROTT: Yes, it was.
22 DR. THADANI: One of the issues, I think,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
96
1 comes up even when you're just doing monitoring, the
2 05 trial is a paroxysmal method trial. And you know,
3 to be confident, I realize that symptomatic isn't shown,
4 would have liked to see at least 48 hours Holter data
5 because these patients were in sinus rhythm from zero
6 to three months before they were entered in the trial.
7 And it's possible, even if you're just doing
8 a routine ECG, they could be in sinus rhythm and once
9 -- even I realize Holter is not adequate because even
10 48 hour Holter will miss it unless you've got the now
11 cardio-beeper device or something else.
12 DR. KOWEY: But with the information that
13 you now know that the patients were monitored --
14 DR. THADANI: But we haven't seen any data
15 on that.
16 DR. KOWEY: Well, no, we have it right here.
17 We have Kaplan Meier for --
18 DR. THADANI: No, no. 05. This is 004.
19 DR. KOWEY: Well, we don't have the data,
20 unfortunately. We can -- I'm sure we can get the data.
21 We don't have it. But for 004, the Kaplan Meier values
22 are exactly the same for any AF recurrence --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
97
1 DR. KONSTAM: Peter, how often were ECGs
2 done in 004 routinely?
3 DR. KOWEY: There was trans-telephonic
4 monitoring, I'm told, every two weeks in 004. This is
5 the European trial.
6 Is that true -- Every two weeks.
7 DR. THADANI: 004 is the chronic effort
8 trial, right?
9 DR. KOWEY: That's correct.
10 DR. THADANI: And this is the paroxysmal
11 --
12 DR. KOWEY: No, this is --
13 DR. THADANI: 05 is -- 05 is the paroxysmal.
14 The data we're discussing at the moment is
15 05.
16 DR. KOWEY: The data you're looking at here
17 --
18 DR. THADANI: No, no. On the board, yes.
19 But we were discussing 05.
20 DR. KOWEY: I understand what you're
21 saying. Yes, that's correct.
22 DR. THADANI: And I'm still a bit leery
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
98
1 because symptomatic with a heart rate flowing, you're
2 driving the episodes lower. If the patient doesn't have
3 palpitation, he doesn't complain.
4 DR. KOWEY: Well, let me just put up --
5 DR. THADANI: But, if you have the data,
6 we'd really like to see the data. That's what I'm
7 saying.
8 DR. KOWEY: This is not a -- I would take
9 issue with the fact that this is a problem only with
10 the drug that has beta blocker problems. This is a
11 problem in all clinical trials of all antiarrhythmic
12 drugs and that is that patients, when they have
13 recurrences, many times don't have symptoms. And you
14 can look at Holter monitors and people on drugs for
15 atrial fibrillation and be astounded.
16 DR. KONSTAM: Yes, that's true, Peter.
17 But here we have a mechanism that specifically will
18 reduce heart rate.
19 DR. KOWEY: Sure enough. But the
20 protection for the patient, as Tom's concern is a stroke.
21 I'm going to point out for example, on AFRM, the NIH
22 trial, that nobody's telling anybody to take anybody
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
99
1 off an anticoagulant drug, even when they think that
2 their rhythm controlled with an antiarrhythmic drug
3 because of this fear of an asymptomatic recurrence no
4 matter what drug you use.
5 DR. LINDENFELD: Well, the concern isn't
6 just stroke. The concern is what we're showing to be
7 statistically significant here is the time to onset of
8 atrial fibrillation. And so, if this impacts the time
9 to onset.
10 DR. KOWEY: You sort of can't have your cake
11 and eat it, too, because in one aspect of this, is that
12 the agency is very concerned about showing that there's
13 some clinical benefit to reducing atrial fibrillation
14 episodes. Well, the benefit that's most demonstrable
15 in these studies is the reduction of symptoms. So,
16 there has to be -- there has to be an analysis of symptom
17 in order to derive its clinical benefit.
18 So, you have --
19 DR. LINDENFELD: I don't think any --
20 DR. KOWEY: We agree that not -- having
21 atrial fibrillation that's subclinical is not a good
22 thing. But on the other hand, the goal of the trial
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
100
1 is to prove that there was some clinical benefit to the
2 patient which is reducing symptoms.
3 DR. LINDENFELD: I think the goal was to
4 show a difference in atrial fibrillation.
5 DR. KOWEY: No --
6 DR. LINDENFELD: That's the primary
7 endpoint.
8 DR. MARROTT: Mr. Chairman, I do have the
9 answer to the question.
10 CHAIRMAN PACKER: That would be very
11 helpful.
12 DR. MARROTT: We do have a Kaplan Meier
13 curve addressing that issue which I -- we don't have
14 a slide.
15 CHAIRMAN PACKER: We'll have our primary
16 reviewer look at this.
17 What we are looking at is time to first
18 symptomatic or asymptomatic. This is any recurrence
19 of atrial fibrillation and flutter. This is in Study
20 05. This is, I think, the issue which is at hand which
21 is what asymptomatic recurrences look like. We'll just
22 remind everyone that presumably the asymptomatic
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
101
1 recurrences were picked up on the trans-telephonic
2 monitoring done every two weeks.
3 Is this correct?
4 DR. MARROTT: Yes.
5 CHAIRMAN PACKER: And the -- this is 05.
6 05. 04, you already have seen.
7 DR. BIGGER: The only problem is that a
8 symptomatic recurrence does not exclude the fact that
9 the patient may have had an asymptomatic.
10 CHAIRMAN PACKER: This is time to either.
11 This is time to either.
12 DR. BIGGER: But short of continuous
13 monitoring, you're not going to know that.
14 CHAIRMAN PACKER: Well, but, it's every two
15 weeks. I mean, you could make it every one week. You
16 could make it every four days. You can make it every
17 -- You can put a Holter on for the rest of the life time
18 of the patient --
19 DR. BIGGER: I know. We're getting back
20 in terms of what are the clinical indications going to
21 be potentially using a drug that has potential toxicity.
22 And what is the -- what's going to be the clinical
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
102
1 indications for use of the drug.
2 CHAIRMAN PACKER: It's a totally separate
3 issue. Let's focus on the issue.
4 The issue is, the question that was raised
5 was, if you included asymptomatic arrhythmias, what
6 would the data look like.
7 DR. FENICHEL: That's always going to be
8 biased, Milton, by even in the case where they are
9 monitoring every two weeks if there is, in addition,
10 sort of supplemental monitoring at the times that
11 symptoms are perceived. So, it is, of course, going
12 to be biased in the direction of a drug which is either
13 bradycardic, or amnestic, or analgesic, or has some
14 other censorium confounding properties that keeps
15 people from bringing these symptomatic events forward.
16 This is the problem with any symptomatic
17 claim that on the one hand when we approve antianginal,
18 we want to say, okay, people are having fewer symptoms.
19 Usually under a fixed stress, a treadmill or something
20 like that. And then the reason that we don't approve
21 ketamine and morphine, and lots of other --
22 benzodiazepines, lots of drugs that might confuse people
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
103
1 and leads them to either override or ignore, or fail
2 to perceive their anginal symptoms is say, no, no, no,
3 it has to also be ischemic.
4 Now, here it seems to me a pure symptomatic
5 claim is, and you'll be asked the question as to this,
6 but it is certainly a possible claim saying this makes
7 people feel better. We say, well, we could give people
8 morphine. We could get people high on benzodiazepines
9 all the time, they would probably would have fewer
10 symptoms.
11 All right. You're got to then show this
12 is an electrically active drug. That it indeed in some
13 way can be shown to cause a reduction in electrical
14 problems. But you don't have to show that in every
15 patient. This is a symptomatic claim.
16 Now, what I guess Tom has pointed out is
17 there are other possible claims. And if the change in
18 the frequency of -- if the change in the frequency of
19 atrial fib really didn't -- if there really were no
20 change in atrial fib, that the only thing that happened
21 were symptomatic changes, maybe the rate's a little bit
22 lower or some other reason people don't perceive it,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
104
1 the presumably the risk of stroke has not really changed
2 and you haven't effected that. Well, that's right.
3 You haven't effected that. This is very much claim
4 dependent.
5 CHAIRMAN PACKER: Yes, Bob, I agree with
6 you. I think that the only reason, and I -- and trying
7 to read the committee's intentions here, that we want
8 to see the asymptomatic arrhythmias to understand
9 whether the reduction in symptoms is related to a
10 suppression of arrhythmias or maybe some other property
11 of the drug. It's more of a internal set of mechanisms.
12 DR. KONSTAM: No, I would say it stronger
13 than that, Milton. Because I think it will, to me, cut
14 in the end directly to the issue of approvability. I
15 think for a couple of reasons. Mostly because when we
16 get into the safety profile of this agent, there are
17 going to be major issues raised. And I think we're going
18 to want to know that its safety profile is acceptable
19 given the specific mechanism that it achieves. If it's
20 achieving its reduction in the recurrence of symptomatic
21 atrial fibrillation because it's a beta blocker, then
22 that will be important.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
105
1 And on the other side of the coin is the
2 issue that Tom raised. I think one could argue, well,
3 all that matters is the symptomatic recurrence. But
4 if that is interpreted by the clinician as meaning that
5 the patient is not in atrial fibrillation, then that
6 may influence the issue of anticoagulation.
7 So, I think this goes beyond an
8 understanding. I think it will directly influence the
9 approvability.
10 CHAIRMAN PACKER: Just so that everyone
11 knows what we're talking about here because this slide
12 isn't available and we are referring to it so it's
13 important to know what we are referring to, the sponsor
14 has presented to the committee, and we'll circulate this
15 up and down so everyone can see it, a slide that includes
16 asymptomatic as well as symptomatic episodes. Time to
17 first event in Study 05.
18 And I think that JoAnn can differ from the
19 interpretation but I think what we're looking at is an
20 overall splay of the curve, first time to event curve,
21 which is fundamentally pretty similar to what we've seen
22 for symptomatic events, with p values that, if anything,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
106
1 are probably a little bit smaller than the p values for
2 the symptomatic events.
3 Would you agree with that? Okay.
4 And we will copy this and send it up and
5 down, or we can just pass this up and down. Why don't
6 we just pass it up and down.
7 DR. KONSTAM: So, I think, then, though it
8 would be worthwhile spending a couple of minutes
9 dissecting this out in terms of methodology. And I
10 guess, and I think Bob spoke to this, is that it's even
11 in the presence of the monitoring, it's an endpoint
12 that's influenced by whether or not the patient has
13 symptoms. And is there a way of sorting that out, and
14 maybe there is and maybe there isn't.
15 DR. PIÑA: Could I ask a question.
16 CHAIRMAN PACKER: Yes.
17 DR. PIÑA: Peter, you said in 05 that the
18 patients who were -- who had impaired creatinine
19 clearance for the most part weren't excluded. And
20 patients with structural heart disease needed to be
21 hospitalized to get into the study. Now, the patients
22 that were hospitalized to get into the study, were they
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
107
1 not on continuous monitoring? Where they not on
2 telemetry?
3 DR. KOWEY: They were.
4 DR. PIÑA: So, is that data included here?
5 I haven't seen these graphs yet, but you would have
6 obvious --
7 DR. KOWEY: Yes. They would have been
8 captured in the Kaplan Meier.
9 DR. PIÑA: So, this would include this
10 monitoring here. It would have been before steady state
11 but --
12 DR. KOWEY: But it would have been only in
13 the early portion of the curve. But we're showing --
14 Milton, is that slide since randomization or, since I
15 didn't see it, or presumed steady state? Do you recall?
16 CHAIRMAN PACKER: I don't have it in front
17 of me.
18 DR. KOWEY: If it was since randomization,
19 they would have been on a monitor. If it was presumed
20 steady state, they may not have been on a monitor.
21 CHAIRMAN PACKER: Before we leave this, let
22 me just emphasize that, and this is, I think, a correct
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
108
1 observation. If you compare the slide which is behind
2 us, which is the -- well, this is 04. If you look at
3 the data in 05, on only symptomatic recurrences, and
4 compare it to the graph with symptomatic or asymptomatic
5 recurrences, and you just look at the exempt rate over
6 time, it would appear to me that more than 90 percent
7 of the events were symptomatic. Is that correct?
8 So, the number of asymptomatic events here
9 is exceedingly small. Is that what one would expect
10 or is that surprising? Tom.
11 If you look at the data and compare it to
12 page 26 of the briefing document, and you look at the
13 percentage, the actual events, the shape of the curve,
14 and how far -- how many patients are free of an event
15 at any given point in time, it would appear as if the
16 vast majority, more than 90 percent of the events in
17 the curve which includes symptomatic and asymptomatic
18 actually are symptomatic because they're already
19 included in the graph on page 26. In other words, the
20 curves don't come down. There isn't a greater failure
21 rate because there's a lot of asymptomatic episodes
22 being included. Is that something you would expect or
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
109
1 not expect? In other words, the vast majority of
2 recurrences here are symptomatic?
3 DR. THADANI: I think because it's a
4 paroxysmal A fib, you're not surprised really. If you
5 look at -- one way to look at it, look at the slide,
6 see how many of these patients paroxysmal if they were
7 symptomatic. Then you can find the incidence of some
8 kind of symptoms were only 40 percent, 60 percent are
9 going to be asymptomatic and not -- maybe (a), they are
10 not in a fib when you're screening them every two weeks.
11 You may not pick it up.
12 DR. KOWEY: Udho, the reason why it may not
13 be just novel to the paroxysmal is the same data are
14 demonstrable for the chronic. If you look at -- this
15 is slide 30. Can I have slide 24, please.
16 This is the symptomatic since randomization
17 in 04. And this is the same point Milton was just making
18 which is that, if you look at the numbers here, patients
19 in each of the curves. And then, if I could have slide
20 30. The numbers are nearly identical. So, this is for
21 a chronic AFD. These are patients who had had a longer
22 duration of AF. So, I don't think it's just because
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
110
1 it was a sort snapshot of their arrhythmia that you
2 weren't capturing.
3 I think patients who have chronic AF and
4 have recurrences usually don't go in and out of AF.
5 They go in AF. They stay in Af.
6 DR. KONSTAM: Well, let me ask whether that
7 -- I'm not sure whether that reassures me or worries
8 me. I mean, I guess, does somebody have some comment
9 about -- I think Tom said something. The frequency with
10 which recurrence of a fib is in fact symptomatic. I
11 think you in fact spoke a moment ago to the fact that
12 most, or that a big proportion of patients with
13 recurrence with a fib in fact is not symptomatic.
14 DR. KOWEY: But I'm encouraged --
15 DR. KONSTAM: And if that's the say -- Well,
16 I'm saying if that's the case, I guess it raises concerns
17 about the effectiveness of the monitoring process here.
18 DR. CALIFF: Ed has written a bunch of
19 papers about this. It would seem like it would be worth
20 hearing from somebody who has actually studied it rather
21 than having opinion.
22 DR. PRITCHETT: The study that Rob's
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
111
1 referring to is the study that Rick Page did in my
2 laboratory in which we took patients who had symptomatic
3 atrial fibrillation and were trained to use a
4 trans-telephonic monitor. And discontinued all the
5 antiarrhythmic therapy and put them on a Holter monitor
6 once a week for five weeks. So that we had an estimate
7 of the rate at which they had asymptomatic atrial
8 fibrillation as well as an estimate of the rate at which
9 they had symptomatic atrial fibrillation documented by
10 the trans-telephonic monitor.
11 And we defined asymptomatic event as any
12 episode of atrial fibrillation lasting 30 seconds or
13 more on the Holter monitor and the symptomatic event
14 which one documented by trans-telephonic monitoring.
15 In the population of patients that were
16 studied there for over that month period, we estimated
17 that for every symptomatic episode that was documented,
18 there were 12 asymptomatic episodes documented. So,
19 there's a lot that. I think if you look at data coming
20 out of pacemakers, which are collecting information on
21 the occurrence of atrial fibrillation in patients, there
22 appears to be a lot of asymptomatic atrial fibrillation.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
112
1 But as Peter has pointed out, the claim here
2 is for symptomatic atrial fibrillation. It's not for
3 stroke and it's not for asymptomatic atrial
4 fibrillation. It's for symptomatic atrial
5 fibrillation.
6 DR. KONSTAM: Well, I guess in an attempt
7 to reassure me, you've worried me more. What you're
8 saying, Ed, is that most -- Well, everybody heard what
9 you said. That's not what we see in the data here as
10 represented. What's represented in the data here, as
11 Milton and others have pointed out, is that there are
12 -- were not very many asymptomatic episodes picked up.
13 DR. PRITCHETT: Well, of course, that
14 relates to the technique you use as a surveillance
15 methodology.
16 DR. KONSTAM: So, that's right. And then,
17 the next -- the point that follows is, maybe there was
18 a problem with the surveillance technique here. And
19 the reason I think that that is important, I mean, I
20 guess this will come for discussion later on, but I,
21 for one, have problems with saying that the indication
22 is going to be for symptoms, symptomatic a fib, because
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
113
1 I'm going to be stuck if I really believe that the big
2 part of that results from rate control. I'm going to
3 have a big problem with that.
4 So, I'm not really reassured by this.
5 DR. KOWEY: I'm having a difficult time --
6 I mean, I managed a very, very large number of people
7 with AF. And if I can get a patient who has very frequent
8 symptoms on a drug that reduces their symptoms in a
9 significant way, I'm not necessarily looking that gift
10 horse in the mouth.
11 Now, with a proviso, Marv, that I do protect
12 those patients against stroke risk if I have any inkling
13 that they are having asymptomatic recurrences. So,
14 clinical practice is that you try to make people feel
15 better.
16 DR. KONSTAM: Right. I agree with that,
17 Peter. The problem for me is going to become, is this
18 better than a beta blocker? That's the problem that
19 I'm going to be -- Milton is shaking his head but this
20 is obviously not a problem for him.
21 But for me, it's going to be a problem
22 because of the side effects of that profile. So, I'm
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
114
1 going to have problem in the end of the day approving
2 the drug for prevention of symptomatic recurrent a fib
3 if I wind up thinking that the vast majority of that
4 effect is a beta blocker effect.
5 DR. PRITCHETT: Milt, this is all germane
6 to the further discussions. I think you have to bear
7 with us. But, Peter, I think would agree that
8 essentially the indications, the three issues with
9 atrial fibrillation are one, stroke, and two is
10 ventricular response. And three is symptoms. And if
11 we can ameliorate symptoms with drugs that are less
12 potentially toxic, and we're looking at risk benefits
13 of these drugs, then obviously common sense would
14 dictate we go with the least noxious drug that can effect
15 a reduction in symptoms, if you're covering the patient
16 for rate control and covering the patient for stroke.
17 CHAIRMAN PACKER: Let me see if I got this.
18 The job of the advisory committee is to evaluate data
19 and to determine if there's evidence that establishes
20 risk to benefit is in the patient’s favor. We generally
21 do not, although we are specifically invited to today
22 by the question, to perform hypothetically or
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
115
1 practically a comparison of the choice under discussion
2 today to the other choices that might be available.
3 There are exceptions to that rule.
4 DR. KONSTAM: No, I agree completely. The
5 goal is going to be assess risk to benefit ratio and
6 if -- so, are we going to wind up having to do a thought
7 experiment at the end as to whether it's better than
8 available therapy that is less toxic. I mean, that's
9 going to come into play.
10 CHAIRMAN PACKER: Let me do this. There
11 are undoubtedly -- I think we have discussed this issue
12 thoroughly. And we will undoubtedly discuss this issue
13 more later on. Let me see if there are other issues
14 related to Study 05. We're still on 05.
15 DR. THADANI: Yes. There are a couple more
16 issues.
17 CHAIRMAN PACKER: JoAnn is first.
18 DR. LINDENFELD: Just one other issue. I
19 think that the time to symptomatic atrial fibrillation,
20 we all want our patients to have less symptoms. But
21 the fact that someone notices a few palpitations, that
22 would, I think, have precipitated a call here because
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
116
1 it doesn't necessarily mean they were bothered by those
2 symptoms or bothered significantly by those symptoms.
3 In other words, that's a subtle difference but -- a
4 few palpitations.
5 So, if I can go on to another, just as a
6 corollary to this. I wonder what the compliance rate
7 with the trans-telephonic monitoring was and what the
8 quality of those were? One, what was the compliance
9 rate? How many were actually successfully done? What
10 was the quality? Where they interpretable? And then,
11 the third part of that is, I'd like to know if the QT
12 interval was evaluated on the trans-telephonic
13 monitoring and if changes were made based on that?
14 Changes in dosing. Or drop outs were effected.
15 DR. MARROTT: I think the answer to you
16 first question is no, we cannot give you that
17 information.
18 DR. LINDENFELD: But then, the results of
19 the study are -- we don't know how many of those
20 trans-telephonic monitors were actually done
21 successfully?
22 DR. MARROTT: No, I just realized that was
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
117
1 your --
2 DR. LINDENFELD: So then, we can't really
3 interpret --
4 DR. MARROTT: -- question. Could you
5 please repeat your first question?
6 DR. LINDENFELD: Yes, I'm sorry. How many
7 of the trans-telephonic monitorings that were to be done
8 every two weeks were successfully completed? What
9 percentage?
10 DR. MARROTT: Well, we can give you that
11 information but we cannot give you that information just
12 now. We'll have to go --
13 DR. LINDENFELD: I think we sort of have
14 to have it.
15 DR. MARROTT: And to provide you that
16 information. But we do have the results of the advanced
17 telephonic monitoring recording as to what were the
18 symptoms and then what was the resulting ECG on that
19 occasion. So, we do have that information.
20 DR. LINDENFELD: I think we'll need to know
21 to that.
22 DR. MARROTT: It will be contained in the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
118
1 report of the study. But we do not have that detail
2 just now.
3 DR. LINDENFELD: At least I would need to
4 know that to be able to assess this time to ECG recurrence
5 is to know how many were successfully done. There would
6 be a big -- if it were 95, that would great. But if
7 it were 60 percent --
8 DR. KOWEY: It is not a low percentage,
9 JoAnn. We can get it for you but it was the over 90
10 percent of them were successfully transmitted.
11 DR. LINDENFELD: Were successfully
12 transmitted with an interpretable result? More than
13 90 percent?
14 DR. KOWEY: Yes. It was virtually in the
15 90s but I don't have the numbers. We can get that.
16 DR. LINDENFELD: And then the other part
17 of that question is, did you evaluate -- I couldn't tell
18 in the protocol, was QT interval evaluated on these
19 monitors and were changes made, drop outs or changes
20 in dosage made, on that? The reason I ask this is because
21 the patients were monitored every two weeks. It's
22 important to know if a significant number of changes
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
119
1 were made based on two week monitoring.
2 DR. MARROTT: The QT was certainly
3 monitored in patients in whom the PDM was recorded on
4 an outpatient basis during the earlier part of the
5 initiation of treatment.
6 DR. KOWEY: The answer is yes. Yes, JoAnn,
7 the QT was monitored by trans-telephonic and if patients
8 made the cut off for the QT in the study, they were
9 dropped.
10 DR. LINDENFELD: Can you tell us what-- The
11 reason this concerns me is because these patients were
12 monitored every two weeks and obviously safety is an
13 important issue. So, I think I need to know to sort
14 of think about how often patients should have
15 electrocardiographic check. I need to know how many
16 patient's doses were changed or were dropped out just
17 solely because of trans-telephonic monitoring.
18 DR. KOWEY: Can I have back up slide 329,
19 please.
20 This is the number of patients with QT
21 intervals greater than 520 milliseconds in all of the
22 studies, including 05, who are then discontinued.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
120
1 DR. LINDENFELD: So, then, less than one
2 percent of patients were withdrawn based on
3 trans-telephonic monitoring, is that --
4 DR. KOWEY: That's placebo.
5 DR. LINDENFELD: And how many of those were
6 after the first, the initial, monitoring period? What
7 I'm trying to just get at is how often do these patients
8 need to be monitored and how many are withdrawn?
9 DR. KOWEY: I don't have -- we don't have
10 data as to when they were withdrawn from the trial.
11 By trans-telephonic, we don't have those data.
12 CHAIRMAN PACKER: Hold on one second.
13 JoAnn, do you have any more questions about
14 05?
15 DR. LINDENFELD: No, I have -- let me just
16 ask one other question.
17 DR. KOWEY: JoAnn, I don't have it
18 specifically for 05. Do you want to see it for the entire
19 data set?
20 DR. LINDENFELD: That would be great.
21 DR. KOWEY: Can I have back up slide,
22 please, 289. 289.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
121
1 This is -- if you look at QT greater than
2 520 milliseconds, JoAnn, this is for the controlled
3 studies 05, 004, 014, and 9A. This is when the patients
4 were dropped.
5 DR. LINDENFELD: Good.
6 Now, this is just -- this I think includes
7 05. But it's just sort of a general question. Assuming
8 that the recommendations will be for treatment in
9 patients with creatinine clearances greater than 40,
10 that was I think what the final -- is that correct?
11 These studies didn't include patients with creatinine
12 clearance less than 40?
13 DR. KOWEY: No, there were no patients in
14 the studies less than 40.
15 DR. LINDENFELD: So, I just would sort of
16 -- I've had one comment. That would mean that your
17 average 75 year old ladies who weighed 70 kilograms with
18 a creatinine of 1.4 would be excluded?
19 DR. KOWEY: Damn right. I wouldn't put
20 that patient on sotalol now or if the drug was approved
21 for the indication. I think that's a high risk patient.
22 DR. LINDENFELD: And I can't give you an
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
122
1 exact number but as we discuss safety, my guess would
2 be that that probably is about 30 percent of the patients
3 with atrial fibrillation in the United States.
4 Somewhere near there.
5 DR. KOWEY: Not in my practice. That may
6 be true in some parts of the country but I can tell you
7 that that's not --
8 DR. KONSTAM: But you don't see patients
9 in nursing homes.
10 DR. KOWEY: No, I don't do -- well, no,
11 actually, there are actually two nursing homes that are
12 attached to our hospital. So, I do see patients in
13 nursing homes.
14 DR. LINDENFELD: I think we know that 70
15 percent of patients are over the age of 65 with atrial
16 fib in the United States.
17 DR. KOWEY: I don't disagree with you at
18 all, JoAnn. There is clearly a subset of patients who
19 are not candidates for this drug and never will be.
20 And it turns out that a little old lady with a creatinine
21 of 1.4 is probably one of them. But what that percentage
22 is of somebody's practice really depends on where you're
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
123
1 practicing and who you're seeing.
2 DR. LINDENFELD: And I don't --
3 DR. KOWEY: It is a significant percentage
4 of patients. I wouldn't argue about that.
5 DR. LINDENFELD: I think there will be --
6 those concerns will be written in. But I do think that
7 many people would not consider that without looking at
8 a carefully, necessarily, a contraindication to these
9 drugs. I think a lot of people would look at it the
10 75 year old woman with a creatinine of 1.4 is a reasonably
11 health person.
12 DR. KOWEY: I've had people who have put
13 patients on sotalol who are anephric. That doesn't mean
14 that that's right. That's a clinical mistake. I think
15 we're not up here arguing about what good clinical
16 practice is. The question is, is there a definable
17 population of patients who can receive the drug? And
18 as we'll see when we get to the safety discussion, there
19 is.
20 CHAIRMAN PACKER: Let me suggest that this
21 is really a safety issue and we'll come back to it.
22 And maybe we'll have an opportunity to see more safety
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
124
1 data. So, let's hold -- I just want to hear efficacy
2 issues related to 05.
3 And, Cindy, you had one?
4 DR. GRINES: Actually, I just wanted to
5 maybe ask any of the panel who want to comment on this,
6 but I was kind of struck by the high rate of asymptomatic
7 atrial fibrillation but the definition in the study
8 quoted earlier by the Duke University. And I guess the
9 question is, if the definition is only 30 seconds of
10 atrial fibrillation and the patient is asymptomatic,
11 is that of any clinical relevance?
12 At our very last meeting, we talked about,
13 it was one of our panel questions. And we questioned
14 the panel whether keeping patients out of atrial fib
15 was clinically self-evident. And it was my
16 recollection that virtually everybody on this panel
17 answered yes, it was clinically self-evident. We did
18 not have to demonstrate a reduction in symptoms.
19 So, that being taken into consideration,
20 I guess I wonder what has changed at this particular
21 meeting and why the opinions are so different?
22 CHAIRMAN PACKER: That's a good question.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
125
1 I'm not certain how to get everyone's view on this.
2 I think maybe if with your permission, Cindy, what I'd
3 like to do is bring this up specifically when we talk
4 about the first question in the panel. Because I think
5 you bring up a very important issue.
6 And the only reason I'd like to postpone
7 it is that we're still on the first efficacy study.
8 DR. GRINES: Well, I guess the second
9 question I have relates to these two studies. And
10 there's been a lot of discussion about this, whether
11 they're symptomatic or asymptomatic and whether every
12 two weeks of monitoring was monitoring frequently
13 enough. And if one looks at -- I guess it was slide
14 25 and slide 44, at least in our paper copies, it details
15 the median time to recurrence and the percentage of
16 relapse-free patients. And they're pretty striking
17 differences between placebo and the proposed dosing
18 group. And I guess question is it pertinent to discuss
19 and spend so much time discussing whether we're missing
20 any relapses when the time to recurrence is so different?
21 And could it be missed if we're monitoring every two
22 weeks?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
126
1 I mean, if we look at Study 004, the median
2 time to recurrence of placebo was 84 days versus greater
3 than 150 days for sotalol. And for Study 05, the median
4 time to recurrence was 25 days in placebo and 226 days
5 on the 120 milligrams of sotalol.
6 DR. LINDENFELD: But part of the problem,
7 I think, with that is that, on 05 at least, at the end
8 of two months, only 40 percent of patients were left
9 in any of the groups.
10 In other words, if you look at .05 and you
11 take the drop outs and the recurrences of atrial fib
12 in each group, it's almost all 40 percent of the patients
13 who are left at two months.
14 DR. GRINES: But that shows the majority
15 of them dropped out because of recurrences, though,
16 isn't that correct? I mean, the SAD is --
17 DR. LINDENFELD: No, that's true in the
18 placebo group but it's not true -- most of them -- many
19 of them were drop outs in the other groups. So, part
20 of the whole question here becomes, I think we all would
21 like to see a drug that prolonged time to atrial
22 fibrillation by six months or eight months. If the time
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
127
1 in the average patient is a couple of weeks, maybe it's
2 not -- But that data is based on only 40 percent of the
3 patients remaining by two months.
4 DR. KOWEY: But, JoAnn, I'd point out that
5 in 9A, drop outs were accounted. And although it was
6 a small study, I'd point out that there are a lot of
7 drugs that have been approved by this advisory committee
8 on less than 100 patients in a data set. Like flecainide
9 for example. But in 9A, which is a relatively small
10 number of patients, it was a very robust p value, the
11 difference between placebo, 80, and 160 milligrams.
12 DR. CALIFF: Don't blame us for flecainide.
13 DR. KOWEY: And drop outs were counted as
14 treatment failures.
15 DR. LINDENFELD: I wasn't born for
16 flecainide.
17 DR. GRINES: I guess I'm still confused,
18 then, as to why these patients did drop out. Because,
19 we saw slides of talking about side effects, and how
20 many dropped out due to side effects. And I guess I
21 was assuming that if they didn't drop out due to side
22 effects, they dropped out due to the fact that they had
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
128
1 a clinical event that was counted.
2 Now, is there a third category as to why
3 the drop out rate was so high?
4 CHAIRMAN PACKER: Did people drop out for
5 reasons other than lack of efficacy or adverse events?
6 DR. KOWEY: I'm sorry, what was the
7 question again?
8 CHAIRMAN PACKER: Did people drop out for
9 reasons other than recurrence or adverse events?
10 DR. KOWEY: Back up slide 205, please.
11 This is the number of discontinuations by treatment
12 group. This is in the outpatients and I can show it
13 to you also for the inpatients.
14 The next slide, 206.
15 CHAIRMAN PACKER: So, I guess it is
16 somewhere around, in the inpatients, about 5 to 10
17 percent. Is that about right?
18 DR. KOWEY: Yes.
19 CHAIRMAN PACKER: And this was
20 administrative issues?
21 DR. KOWEY: I don't have those details.
22 John, do you know? Drop outs for other?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
129
1 DR. WILLIAMS: The other category was a
2 miscellaneous group. Either patients moved away from
3 the center or they were protocol violators. So, that
4 was the usual group of non-compliant study patients.
5 For the AE drop outs, most of them were for
6 beta blocking side effects, bradycardia, weakness,
7 dizziness, and so forth.
8 CHAIRMAN PACKER: We'll get into side
9 effects later. We're just talking about how side
10 effects influence the effect in an efficacy.
11 DR. THADANI: Is that the correct slide?
12 I'm having a hard time following it now. You're saying
13 there were 50 patients on placebo, 40 dropped out?
14 CHAIRMAN PACKER: Yes, don't forget.
15 Discontinuation includes a recurrence here.
16 DR. THADANI: That's a --
17 DR. KOWEY: This lack of efficacy was 60,
18 35 of those 40 were lack of efficacy.
19 DR. THADANI: And the
20 discontinuation--you're including everything?
21 DR. KOWEY: Yes.
22 CHAIRMAN PACKER: Anyone have any other
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
130
1 points about 05?
2 Mark.
3 DR. KONSTAM: Yes. You know, the issue
4 about whether patients were started in hospital or out
5 of the hospital, I guess if I understood in 05, patients
6 with structural heart disease were mandated to be in
7 the hospital. Those without structural heart disease
8 were not mandated to be in the hospital. Do I got that
9 right?
10 DR. KOWEY: That's right.
11 DR. KONSTAM: But could be in the hospital?
12 DR. KOWEY: Yes. A very small percentage
13 of those patients were in the hospital.
14 DR. KONSTAM: What percent?
15 DR. KOWEY: It was less than 10 percent.
16 DR. KONSTAM: So, overall, over the entire
17 population, what percentage was in hospital and what
18 percentage was out of hospital?
19 DR. KOWEY: Well, can I have those two
20 slides I just had. You can count the numbers.
21 DR. LINDENFELD: Twenty-three percent.
22 DR. KOWEY: I'm sorry, 27 percent.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
131
1 DR. KONSTAM: Twenty --
2 DR. KOWEY: Twenty-seven percent in the
3 hospital.
4 DR. KONSTAM: Twenty-seven percent in the
5 hospital. Thanks.
6 CHAIRMAN PACKER: On that same --
7 DR. THADANI: I've got a question.
8 CHAIRMAN PACKER: Yes, Udho, hold on.
9 On the same issue, the outpatients were
10 generally -- the indications without structural heart
11 disease were generally viewed as outpatients. But when
12 they were outpatients, they still underwent in all cases
13 trans-telephonic monitoring. For how long? Can you
14 clarify?
15 DR. KOWEY: When they were out of the
16 hospital?
17 CHAIRMAN PACKER: Initiated, when they
18 initiated on therapy as an outpatient, they underwent
19 TTM for a certain period of time continuously during
20 initiation of therapy.
21 John.
22 DR. WILLIAMS: TTM monitoring was done
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
132
1 continually until they either had a relapse or they
2 finished six months of treatment.
3 CHAIRMAN PACKER: No, no, no. It was done
4 intermittently.
5 DR. WILLIAMS: During the outpatient
6 initiation, they had -- they sent in telephonic
7 monitoring, I think maybe about three days at the time
8 of reaching steady state. We weren't getting daily --
9 a daily TTM. We have practice TTMs to teach them how
10 to use the device and then they were -- during the
11 initiation, we had more frequent TTMs.
12 CHAIRMAN PACKER: Again, I'm sorry, I just
13 want to clarify the point. The TTM is recorded for a
14 relatively brief period of time. People hook
15 themselves up and send it in over telephone lines for
16 a brief period of time.
17 DR. WILLIAMS: The TTM is recorded for just
18 a few seconds. And then during follow up, it was every
19 two weeks. And that's why when you have such a short
20 period of ECG documentation, you don't pick up a lot
21 of asymptomatic occurrences.
22 CHAIRMAN PACKER: Just so I understand,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
133
1 that in the proposed labeling, when the concept of where
2 therapy should be initiated is discussed and it says
3 patients without structural heart disease can be
4 initiated outpatient, outpatient with TTM or outpatient
5 with daily ECGs, or outpatient without either?
6 DR. KOWEY: The way it was done in the trial
7 was outpatient with both TTM and periodic
8 electrocardiograms. The way I think it should be done
9 in practice is with TTM. I'd probably, to be honest
10 with you, Milton, I probably would do it more frequently
11 in the initial phases than every three days. I commonly
12 get one every day for the first seven to ten days the
13 patient is being titrated at that dose.
14 CHAIRMAN PACKER: I'm sorry, Mike. Yes.
15 DR. CAIN: Peter, two questions which would
16 also include all the studies. Atrial fib and atrial
17 flutter are grouped together.
18 DR. KOWEY: Yes.
19 DR. CAIN: My bias would be that most of
20 that was atrial fibrillation.
21 DR. KOWEY: Yes.
22 DR. CAIN: But just for the record, was that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
134
1 the case or were there differences among the trials where
2 some trial picked up more of pure flutter versus atrial
3 fib?
4 DR. KOWEY: No, in fact, Mike, we went back
5 and looked at that very carefully because we were
6 concerned about the same issue. The vast majority of
7 patients in these trials had atrial fibrillation. And
8 somewhere between 10 to 20 percent, and this was really
9 consistent across the trials, also had atrial flutter.
10 There were, for example, these were the patients that
11 at least had some period of atrial fibrillation. So,
12 the numbers, where it's less than 100, means that there
13 were 10 percent for example in 9A that had only flutter.
14 So, when you see 100 percent, that means that they had
15 atrial fibrillation and about 10 to 20 percent of those
16 patients across the trials also had some period of
17 flutter. So, it was a typical AF population in all
18 respects.
19 DR. CAIN: And the second is just a point
20 of clarification about the duration of the
21 trans-telephonic monitoring in NO5, what the definition
22 of a recurrence was. So, specifically, if someone had
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
135
1 symptoms and had 25 seconds of what looked like atrial
2 fibrillation, was that counted as a recurrence or did
3 you use the 30 second definition, or did it vary?
4 DR. WILLIAMS: The recurrence was
5 diagnosed with ECG documentation of a fib or flutter,
6 plus they had to have symptoms of a fib or flutter.
7 DR. CAIN: And the duration of the ECG strip
8 that showed the fib and flutter was two minutes? Thirty
9 seconds?
10 DR. WILLIAMS: No, there was no definition
11 of duration but they had to have symptoms with it.
12 DR. CAIN: So, a failure could be someone
13 who had marked symptoms of an irregular palpitation
14 feeling in their chest. And yet the ECG strip could
15 have shown 10 seconds of atrial fibrillation?
16 DR. WILLIAMS: Theoretically possible.
17 DR. PRITCHETT: Remember, Mike, that the
18 -- it takes a little bit of time to get the device on.
19 It could be anywhere from a minute to several minutes
20 to get the thing on. It isn't -- if you captured it,
21 it's more -- and it lasted six seconds on the recording,
22 the thing lasted more than six seconds.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
136
1 CHAIRMAN PACKER: But on the other hand,
2 I guess it's possible that they would have had a burst
3 of palpitations and by the time they got the device on,
4 nothing was recorded. And that wouldn't count at all.
5 DR. KONSTAM: If somebody had -- so this
6 is in terms of following to the endpoints. If somebody
7 had a trans-telephonic monitor routinely done, not
8 because of any reported symptoms, and the patient was
9 in a fib on this monitor, and still recorded no symptoms,
10 how was that patient handled in terms of an endpoint?
11 DR. WILLIAMS: Without symptoms, they
12 would be continued in the trial.
13 DR. KONSTAM: And they would not have been
14 considered an endpoint?
15 DR. WILLIAMS: No, asymptomatic a fib was
16 not an endpoint for the study.
17 DR. KONSTAM: But you tracked it. You do
18 have asymptomatic. You do have ECG documented a fib.
19 DR. WILLIAMS: The number of asymptomatic
20 documentation that we got from routine monitoring was
21 a very small number. As you will see from the Kaplan
22 Meier curves, the difference in the ends for
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
137
1 asymptomatic plus symptomatic was very -- almost the
2 same as for symptomatic.
3 DR. KONSTAM: But we just heard that for
4 every asymptomatic not what we would have expected for
5 every symptomatic recurrence of a fib, we would have,
6 what did you say, 14 --
7 I'm struck by the fact --
8 DR. PRITCHETT: Captured by Holter --
9 captured by continuous monitoring, now. This is not
10 trans-telephonic. This is continuous monitoring that
11 you do for five days over the course of a month. You
12 capture a lot more asymptomatic stuff than you do by
13 sampling for 30 seconds every couple of weeks.
14 The every two week sampling is not a good
15 way to measure the rate at which asymptomatic atrial
16 fib occurs. It is a way to estimate the relative rate
17 that occurs in different groups.
18 DR. KONSTAM: Well, the two curves are
19 almost identical.
20 DR. PRITCHETT: Yes.
21 DR. KONSTAM: So, this is what you're
22 saying. Is that you picked up very, very few -- through
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
138
1 trans-telephonic monitoring, you picked up very few
2 episodes of a fib that was asymptomatic.
3 DR. KOWEY: Ed, is this something that we
4 saw in the flecainide experience?
5 DR. PRITCHETT: Those data were conducted
6 in the infancy of this technique and I think we did not
7 look nearly as closely at that time at those data. So,
8 I don't think we know what went on in the flecainide
9 group.
10 DR. KONSTAM: But you still must have
11 captured, even though those few, as endpoints because
12 you kept track of them. And they were -- they do appear
13 in those curves that we have that are called symptomatic
14 or ECG. Is that right, that those patients were then
15 just followed. Could they then, could those few
16 patients have subsequently developed an episode of
17 symptomatic a fib?
18 DR. PRITCHETT: Of course. I mean, what
19 happened, the trans-telephonic electrocardiograms are
20 provided to the investigator so that the investigator
21 knows that that patient had atrial fibrillation and can
22 do something about it if he thinks it's important. In
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
139
1 point of fact, most times that asymptomatic episode of
2 atrial fibrillation resolved spontaneously and the
3 patient goes on and has a symptomatic episode at some
4 point later.
5 DR. KONSTAM: Let me just follow up on that,
6 then.
7 If I'm an investigator and I've got a
8 trans-telephonic monitor, and it shows that the
9 patient's in a fib. Now, does that not bias me in terms
10 of interpretation of the subjective endpoint of
11 symptomatic a fib? In other words, when I speak to the
12 patient next, maybe the next day or maybe my nurse calls
13 him on the phone, isn't it more likely that I'm going
14 to solicit symptoms of a fib because I know that that
15 patient is in a fib?
16 DR. PRITCHETT: Are solicited in using the
17 trans-telephonic technique are recorded by the
18 technicians who handle the calls at the time the patients
19 make them. So, the patient calls in, transmits the
20 recording, the technician says, what symptoms have you
21 had, have you had any of these.
22 DR. KONSTAM: I understand. But then you
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
140
1 just said that the investigator then has that
2 information. And what I'm saying is couldn't it bias
3 the likelihood that the next day in the physician's
4 conversations with the patient or somebody else's, now
5 we know that that patient is in a fib. It would seem
6 to me they would be more likely to document a symptomatic
7 endpoint in that circumstance.
8 What I'm sort of getting the feeling for
9 is that the distinction between symptomatic and
10 asymptomatic here is very murky. I mean, I think you
11 were doing the trans-telephonic monitoring. The
12 numbers, in fact, are almost identical. You say you
13 picked up very few additional. But I'm wondering
14 really, and we're dealing with a subjective assessment
15 of a patient if they have symptoms of a fib. I call
16 the patient up and then they say, yes, I -- see, I have
17 been having some palpitation. I don't know.
18 DR. KOWEY: Marv, does this help you at all?
19 We just thought maybe if you look at the patients who
20 were or were not taking beta blockers in addition to
21 the study drug, and I believe that there area bout 30
22 percent.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
141
1 DR. THADANI: I was going to ask you that
2 question. Why was it two -- sotalol is a beta blocker.
3 DR. KOWEY: Well, it was a blinded study.
4 DR. THADANI: Yes, but 30 percent. I'm
5 surprised at the start of the study beta blockers were
6 not withdrawn. How often in practice --
7 DR. KOWEY: They are.
8 DR. THADANI: -- two beta blockers.
9 DR. KOWEY: If you know you're giving
10 somebody a beta blocker, you don't use another beta
11 blocker.
12 DR. THADANI: Yes, I'm surprised.
13 DR. KOWEY: But they may have been getting
14 a low dose of the study drug or they may have been getting
15 placebo.
16 DR. THADANI: No, but this was a dual
17 response study, 05. Thirty percent of the patients.
18 I know JoAnn asked the question on calcium channel
19 blockers. But 30 percent on beta blockers.
20 DR. KOWEY: Right.
21 DR. THADANI: You would have thought they
22 would have been withdrawn before they entered the study.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
142
1 DR. KOWEY: Well, if you look at this, Udho,
2 most -- There's a larger percentage of patients getting
3 beta blockers in the placebo and in the low dose, as
4 you would have expected, if that's the reason.
5 DR. THADANI: But you've still got --
6 DR. KOWEY: I'm not arguing --
7 DR. THADANI: -- 21 percent. So, you've
8 still got 21 percent even in the highest dose. I think
9 my feeling is beta blockers would never stop because
10 the protocol was not design to withdraw the beta
11 blockers. There happened to be -- there might have been
12 post MI patients. And although sotalol has been used
13 for them in European trials, they were never withdrawn.
14 DR. KOWEY: Maybe this has boomeranged a
15 little bit, but basically the reason I put this on--
16 if I could have the other slide back, the hazard ratio.
17 I was going to try to help Marvin. Now it
18 looks like I've hurt you. But, if you look at the use
19 of beta blocker, Marv, you see that if it were a beta
20 blocker effect, then maybe people who were getting more
21 beta blocker would have had a better outcome. I don't
22 know. It's a way of looking at it.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
143
1 DR. KONSTAM: Yes, it could be. But then
2 again, the concomitant beta blocker use may have
3 influenced the ability to exceed the dose.
4 DR. KOWEY: It's not a perfect answer. I'm
5 just trying to help you to be comfortable.
6 But now I've made Dr. Thadani extremely
7 uncomfortable and I don't know how to deal with that.
8 DR. THADANI: No, but I think in practice
9 I don't use two beta blockers. If I see a patient who
10 is in a fib and he happens to be in the beta blocker,
11 if I switch him to sotalol, although not approved, but
12 I never have. It's very rare. I think only last week
13 I saw a patient that was on both drugs.
14 So, the only thing could be that you could
15 argue the other way around, the patients on placebo and
16 beta blocker might have a less asymptomatic -- less
17 symptomatic a fib because their heart rate could have
18 been slower.
19 So, I was surprised that the study design
20 was -- went through all the time and this drug was not
21 withdrawn.
22 The other question I had that might be
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
144
1 relevant, is if you look at the decay curve of recurrency
2 survival, pick up, you know, 05 study, it seems like
3 placebo decays very quickly and then flattens. Had you
4 followed these patients for a longer time, all probably
5 would have recurred.
6 DR. KOWEY: I think that's a fair
7 statement.
8 DR. THADANI: So, if that is true, so really
9 all we're talking about, symptomatic recurrence for a
10 period of six months or eight months. And you follow
11 them for -- maybe it's relevant because both patients
12 are symptomatic.
13 DR. KOWEY: Clinically, again, I think the
14 reason why the claim was worded the way it was, is because
15 we never expected antiarrhythmic drugs to be harmful.
16 We expected it would be better. And so, we expect
17 people on active therapy or on placebo to ultimately
18 have a recurrence. It's the commonality. So, that's
19 why the wording was the wording.
20 DR. THADANI: The other question is now a
21 lot of patients in this study with structural heart
22 disease. How many patients really had, say, a Class
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
145
1 II, III, or IV heart failures? Very few if I remember
2 correctly. Six, seven.
3 DR. KOWEY: There were a number of Class
4 IIs. There were very few Class IIIs. I can go back
5 to Study 05. If I could have the core slide.
6 CHAIRMAN PACKER: I just want to note that
7 the sponsor is specifically requesting an exclusion on
8 patients with overt heart barrier from any indications.
9 DR. THADANI: Yes, I think it was. That's
10 why I asked this question, if I remember correctly.
11 CHAIRMAN PACKER: For study -- Do we have
12 the slide? I don't know if we have it on the core.
13 It was a small percentage. There was a very small
14 percentage of Class IIIs.
15 DR. THADANI: So, they were excluded, if
16 they had a heart failure.
17 DR. KOWEY: Yes, we don't have the
18 breakdown on these slides but it was a very small number
19 for Class IIIs. Most of them were Class Is and IIs.
20 DR. THADANI: I think that becomes
21 relevant, too, now with the changing therapy for heart
22 failure. These are going to take a role, and if they
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
146
1 are in heart failure, they will not necessarily -- we
2 don't have any data --
3 DR. KOWEY: No, no.
4 DR. THADANI: -- from this study and heart
5 failure?
6 DR. KOWEY: No, we definitely do not. We
7 don't have it anywhere. The only place where we have
8 any heart failure patients in Class III was in the
9 quinidine comparative study. This is the data from 004
10 that looks at the distribution. It's the same for 05.
11 CHAIRMAN PACKER: Tom.
12 DR. BIGGER: I just had a point of
13 clarification about what was called structural heart
14 disease in these studies. And was it the same for each
15 of the studies?
16 DR. KOWEY: Yes, we have -- we can show you
17 that. That's on a back up. We'll try to get it out
18 for you.
19 Do we have the definition on a back up?
20 What is it?
21 Here you go. This is the definition and
22 it was the same across all the studies. If you had any
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
147
1 one of those, you were classified as having structural
2 heart disease.
3 CHAIRMAN PACKER: All right. Does anyone
4 else have any comments on 05?
5 DR. GRINES: Did left atrial enlargement
6 qualify as structural heart disease? No?
7 DR. KOWEY: No.
8 CHAIRMAN PACKER: Any other comments on 05?
9 I want to, before we break for lunch, try
10 to get through the questions or issues on the other
11 studies so that we can begin safety after the break.
12 let me, in doing so, simply make note of the fact that
13 a lot of the issues that we have brought up on 05 apply
14 to the other trials. And therefore, we need not
15 reiterate all of the, or revisit, all of these issues,
16 the issues of informative censoring, the issues of
17 trans-telephonic monitoring symptoms. We have covered
18 these, I think, fairly thoroughly in the last long period
19 of time.
20 And consequently, I think that we can assume
21 that whatever concerns applied to 05, whether or not
22 they've been resolved, will in fact apply, resolved or
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
148
1 not, to the other trials. So, let me just ask everyone
2 to, when they review the other trials, try to bring up
3 issues unique to those studies and not reiterate the
4 same issues.
5 I'm going to ask JoAnn to initiate the
6 discussion of 04 next. But before, perhaps, doing that,
7 it would be, I think, important to mention the first
8 issue that I think everyone has already identified from
9 the FDA review which was the possibility of a treatment
10 by center interaction for Center 29.
11 JoAnn, do you want to ask a specific
12 question about that? And I only want to begin that way
13 because it was highlighted in the FDA review.
14 DR. LINDENFELD: Yes, I think maybe the way
15 to begin first is this issue that's highlighted in the
16 briefing booklet about the actual intent for numbers
17 of patients. We're told that the study was originally
18 designed for 200 patients and somewhere increased to
19 349. The reason I want to -- and then address the site
20 specific issue. Because I'm concerned not just that
21 there was so much different in one site, but that the
22 difference in those first 200 patients and the last 150,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
149
1 that specific site, I think entered a very large
2 disproportionate number of patients in the last 149
3 patients of that.
4 So, maybe you could address all of that
5 together.
6 DR. KOWEY: Dr. Marrott will address that
7 question.
8 DR. MARROTT: Mr. Chairman, with regards
9 to the sample size issue on Study 004, somewhere in
10 January of 1992, at a time point when about 110 patients
11 or thereabouts had been entered into the drug, the
12 sponsor, that is Bristol-Myers Squibb, the personnel
13 that were responsible for undertaking the trial, that
14 is the physician, the biostatistican, and the other
15 support team, came to the conclusion from looking at
16 certain other trials, for example, the control relapse
17 information in the Coplan analysis, if you remember,
18 there were six studies in the Coplan analysis. And they
19 looked at the response of the recurrence rate in the
20 control arm of the quinidine control evaluation
21 meta-analysis as you may remember.
22 They also looked at the sotalol and the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
150
1 quinidine study, that is, Study H, and they looked at
2 the results of Study 014 where the placebo response was
3 33 percent.
4 I think what has happened there, Mr.
5 Chairman, is that there's been only a small fine tuning
6 of the sample size for a different assumption that was
7 30 percent initially at the start of the trail to 20
8 percent. So that the assumption at the start of the
9 trial, if I remember it right, was 25 percent for placebo
10 and 55 percent for active group, both d-sotalol and
11 d,l-sotalol. Whereas, it was the other group 30 percent
12 for placebo and 50 percent for sotalol based on the 014
13 for the placebo and sotalol. Based on the Coplan
14 analysis for the standard -- sorry, the control group.
15 And based on the age study, again, for sotalol.
16 Now, there was a perception with the
17 division that something was not quite clear about this
18 and then an analysis was done with the first 200
19 patients. But we did point out with the division at
20 the time of the amendment of the protocol, we only had
21 110 patients recruited. And as you know from clinical
22 trials that are done by pharmaceutical companies, if
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
151
1 110 patients are enrolled, possibly those 110 patients,
2 they are not available in house.
3 So, I think that there was a fair assessment
4 of what was going to be a better assumption with regards
5 to placebo and with regards to the active groups. So
6 that is our response to the issue of sample size and
7 the issue of the first 200 patients analysis.
8 I think there was a third one that Center
9 29 where you rightly point out that Center 29 we think
10 by a play of chance, is performing very well for the
11 active and very badly for the placebo.
12 And I would, if you don't mind, Mr.
13 Chairman, request Dr. Fisher to please put forward his
14 point of view.
15 CHAIRMAN PACKER: Before we do that, see
16 if there's any additional clarification which is needed
17 on the expansion of sample size. They are two separate.
18 They're a little bit related but I think JoAnn's
19 question was specifically on the expansion of sample
20 size.
21 DR. MacNEIL: I think from the company
22 point of view --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
152
1 CHAIRMAN PACKER: Can you identify
2 yourself, please, for the record.
3 DR. MacNEIL: Sorry, Dr. MacNeil from
4 Bristol-Myers Squibb.
5 At the time we were conducting the study,
6 people become aware of the fact that the original sample
7 size was based on a treatment effect of 30 percent
8 difference assuming that there would be a 25 percent
9 placebo freedom of recurrence and a 55 percent freedom
10 of recurrence for the active drug.
11 It was recognized from the meta-analyses
12 that the estimate of people free from recurrence should
13 have been higher based on what we knew from placebo.
14 So, the recommendation was that we considered that the
15 placebo effect would be 30 percent free from a recurrence
16 and the active drug be 50 percent. And that's what led
17 to the increase in the sample size. It was just felt
18 that the study was under powered to show a difference.
19 And we were blinded. We didn't have, you know,
20 unblinded information upon which to make that judgment.
21 CHAIRMAN PACKER: Is the concern of the
22 division the lack of documentation of this? I just want
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
153
1 to see if I understand what the issues are because the
2 explanation that you have provided is slightly different
3 than the concern that has been raised. And I just want
4 to make sure we reconcile these so that we can move
5 forward.
6 Abe, Dr. Karkowsky.
7 DR. KARKOWSKY: Abraham Karkowsky,
8 Cardiorenal.
9 We read the protocols as they come in. This
10 was an non-IND protocol. We saw no protocols. We have
11 no record as to when things were done and when things
12 were changed.
13 It's hard to retrospectively say what would
14 have happened if this study would have found in fact
15 200 patients. Would they have continued to enroll
16 patients?
17 CHAIRMAN PACKER: But they say they didn't
18 unblind. We have to take their --
19 DR. CALIFF: Can I ask a question about
20 that? And I've frequently wondered this. In these
21 kind of small studies there's no DSMC. You're telling
22 me there's no one who has access to the code or is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
154
1 monitoring the study?
2 DR. MacNEIL: The company does have access
3 to the code but it's really -- it's restricted. It's
4 not available to anyone. So, it's basically locked up.
5 DR. CALIFF: Restricted to whom I guess is
6 what I'm asking.
7 DR. MacNEIL: Well, it's the person who
8 generated the randomization code. It's basically in
9 the statistical department.
10 DR. CALIFF: So there's a statistician
11 who's monitoring the trial?
12 DR. MacNEIL: Not in an unblinded fashion.
13 In other words, the randomization code is generated
14 and then in essence it's not available to anybody to
15 review.
16 DR. CALIFF: And adverse events are not
17 monitored by anyone?
18 DR. MacNEIL: Adverse events are monitored
19 but they're monitored in the blinded way unless there's
20 a specific reason to unblind. And then there has to be
21 -- there's a formal mechanism by which there's a request
22 made to the statistician for unblinding of a specific
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
155
1 patient. But the statisticians themselves don't follow
2 unblinded events.
3 DR. CALIFF: Okay, so I'm not saying
4 there's a problem here. I'm just trying to understand
5 how it worked.
6 So, there were just like a safety committee,
7 there was someone who was looking at the data as it came
8 in. It was a statistician. There were no clear rules
9 for when the statistician might say there were too many
10 adverse events. And that was restricted purely to that
11 statistician.
12 DR. MacNEIL: Well, let me clarify. The
13 statistician generates the randomization code and then
14 it is kept in a secure place. And the statistician
15 doesn't otherwise look at any of the trial data. The
16 clinical persons responsible for the trial review all
17 of the adverse events as they are received on an ongoing
18 basis. In general, these studies remain blinded
19 despite the fact that there's a serious adverse event
20 that might occur.
21 But in the unique instance, the
22 investigator may request in order to manage the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
156
1 patients, to know what the specific drug was that the
2 patient was receiving. When that happens, then within
3 the company there's a procedure by which the signatures
4 of several individuals are involved in order to get the
5 statistician to go back into the randomization code and
6 tell specifically what that person was on. In an
7 emergency basis, all of our drugs are labeled such that
8 the investigator could unblind on site but we -- that
9 would invalidate the patient and that we don't encourage
10 investigators to do.
11 DR. MOYÉ: I think I'd just like to just
12 ask you to elaborate like that in responding to my
13 question. I understand that you have serious adverse
14 event monitors to look at the individual case before
15 they come in. And they have reporting responsibilities
16 based on severity of event.
17 But let me ask you specifically, was there
18 anybody in the company who was monitoring the trial on
19 a per group basis? Anybody who's looking at placebo
20 event rates, active group event rates, or efficacy for
21 the first 110 patients? Was there anybody anywhere
22 doing that?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
157
1 DR. MacNEIL: No.
2 DR. MOYÉ: Thank you.
3 CHAIRMAN PACKER: Let me see if I can just
4 clarify. I think you just said that the amendment that
5 expanded the trial was made in January of 1992?
6 DR. MacNEIL: Yes.
7 CHAIRMAN PACKER: I guess the reason for
8 the physician's concern is in the annual report dated
9 July 7th, 1992, six months after the amendment, this
10 study is still referred to as a 200 patient trial. In
11 other words, if the decision had been made in January
12 to expand the study, the question that the division has
13 is why six months later the annual report does not refer
14 to the expanded patient population?
15 DR. MacNEIL: I would have to say that was
16 an error because the amendment does exist increasing
17 the sample size.
18 DR. KARKOWSKY: We have the information.
19 The sponsor sent us -- Berlex sends us through Bristol
20 some information from Bristol-Myers Squibb as to the
21 rationale for modifying their sample size. I can give
22 that to you to look at it.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
158
1 There was -- It did not seem like it was
2 definitive that they would go to do, but it was clearly
3 a discussion amongst people. We have no additional
4 information. If it's important, DSI could go out and
5 look at that information and convince everybody as to
6 the timing of amendments, whether they received by all
7 centers. And from the vantage point of the division,
8 we can't do any more. And we will treat the information
9 as if it is not any way unreasonable.
10 CHAIRMAN PACKER: And I just want one more
11 clarification because it's also raised by the division
12 when the discrepancy or an explanation was first sought
13 during a meeting with the division. Apparently the
14 division was told that the sample size was increased
15 not because the expected event rates were adjusted, but
16 because the initial intent was to do a comparison of
17 d-sotalol versus placebo as opposed to a comparison of
18 either treatment. The explanation you have just
19 provided is different than the explanation in the
20 divisional record. We just want to be able to make sure
21 what the story is.
22 Did I say that correctly?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
159
1 DR. MacNEIL: I don't know. The
2 information we have, the discussion amongst the
3 statisticians at Bristol Myers was mostly with respect
4 to d-sotalol. Okay, as far as I remember, looking at
5 the primary endpoints in the study, it was a comparison
6 of d-sotalol with placebo and d,l-sotalol with placebo.
7 So, I can't address your issue specifically but I would
8 have presumed that it's d and d,l versus placebo.
9 CHAIRMAN PACKER: Maybe I can rephrase the
10 question.
11 I don't think that anyone in the division,
12 and I don't think anyone in the committee is saying that
13 there is a problem. I think what we just want to do
14 is clarify the actual sequence of events. And maybe
15 I should ask.
16 Lem, if there were a problem, what would
17 you do with the p values in order to make an appropriate
18 adjustment? Because, it could be that the p values for
19 this trial are sufficiently small that it just doesn't
20 make any difference. I just want to be able to clarify.
21 Let me just clarify the intent of my
22 question is not to assume there's a problem. This is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
160
1 more of a hypothetical question as opposed to a question
2 specifically. Lem?
3 DR. MOYÉ: Right. So, we're talking about
4 a hypothetical circumstance here it sounds like, where--
5 CHAIRMAN PACKER: I just want to -- I want
6 to emphasize that because there is -- the division is
7 not saying that there is a problem. All, and you know
8 we all want to be very, very careful here. My question
9 is simply if there were in fact an expansion in the trial
10 based on an interim analysis, what would one then do?
11 DR. MOYÉ: I think that if this interim
12 analysis was not prospectively specified, and it was
13 not in the protocol at the inception of the trial, that
14 there would be a potential expansion of the sample size.
15 Then we're looking at essentially letting the data from
16 the trial determine the analysis plan. And I think that
17 that has severe implications. I think the best thing
18 to do in that circumstance is to analyze the data as
19 the investigators planned to collect it.
20 Then it's as simple as planning what you
21 mean to do and then doing what you plan. And if the
22 initial plan was to evaluate 200 patients, then the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
161
1 efficacy analysis for 004 would be solely on the 200
2 patients. To -- It is very -- I will say it is likely
3 that you can get a subsequently randomized sub-cohort,
4 which would have a different effect than that seen by
5 the first 200. One reason would be your sampling
6 variability, and I've seen that happen before. Another
7 reason is that the patients who are randomized later.
8 Perhaps they come from different centers. Perhaps
9 they don't meet the exact same exclusion criteria. And
10 so, that would be another reason why the effect might
11 be different in one later randomized cohort than
12 another.
13 CHAIRMAN PACKER: Tom and then Mark.
14 DR. BIGGER: Yes, if I understood it, that
15 sounds right hypothetically. But that wasn't what was
16 done here at all. The data from the trial wasn't
17 examined to increase the sample size. As I understood
18 Mr. MacNeil. But, information coming from outside the
19 trial, that the event rates they used to estimate the
20 sample size were not the best estimates at the time they
21 were reviewing the sample size situation. And the basis
22 for decision to increase the sample size didn't come
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
162
1 from the data inside the trial but from information
2 coming from outside being reviewed into the cumulative.
3 That's what he said anyway.
4 DR. MOYÉ: I don't dispute that. It was
5 a hypothetical.
6 CHAIRMAN PACKER: It is a hypothetical
7 question. Marv and then Rob.
8 DR. KONSTAM: Yes, let me just -- this is
9 interesting because Rob and I were just on a panel
10 yesterday where this very issue came up. And it came
11 up from the perspective of an interim analysis was done.
12 There was a concern that there was not sufficient power
13 based on that interim analysis. There was no
14 pre-specified plan in the protocol to expand the sample
15 size. Nevertheless, that was the decision that was
16 taken at the time of the interim analysis to expand the
17 sample size. And so, that was very clear that that had
18 occurred. And the discussion took place in the room
19 with several statisticians present, what was the level
20 of penalty that should be imposed. And the consensus
21 of opinion was a very small penalty. That it would not
22 substantially influence the interpretation of the p
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
163
1 value.
2 So, we can go back over the minutes of that
3 meeting. But just in the interest of bringing
4 consistency to this discussion, this is the way it took
5 place yesterday.
6 CHAIRMAN PACKER: Why don't we move forward
7 since we are not saying that this is an issue. Let me
8 reiterate, we are not saying that this is an issue.
9 DR. CALIFF: Milton.
10 CHAIRMAN PACKER: Oh, Rob. I'm sorry.
11 DR. CALIFF: I think it would, from my
12 perspective having been on the panel yesterday and
13 having been outvoted eight to one on this issue, it would
14 be useful, I think, to at least hear the cardiorenal
15 perspective on this issue of engineering clinical
16 trials; that is, doing an interim analysis, looking at
17 the unblinded event rates in the two groups, observing
18 the observed difference, and then recalculating the
19 sample size based on the observation. The two other
20 Divisions yesterday said that it mattered a little bit
21 but not much.
22 CHAIRMAN PACKER: Can I make a suggestion?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
164
1 This sounds like a great idea for a symposium. I think
2 that we are ill-equipped to deal with this in any kind
3 of definitive fashion today.
4 DR. FISHER: Just a quick aside, Rob, I have
5 published a sequential method but one of the
6 consequences is you can look at the unblinded data
7 continuously, make decisions on how far you are going
8 in such a way to preserve the type 1 error. I mean,
9 there are ways of dealing with it now.
10 CHAIRMAN PACKER: Okay. JoAnn, we're
11 going to go back to you on any other issues. Do you
12 want to deal with the issue of Center 29?
13 DR. LINDENFELD: I know we're going to hear
14 some discussion about Center 29 but the other concern
15 I have about Center 29 is that they entered a substantia
16 by greater percentage of the patients in the study in
17 the second half than in the first half. Not only were
18 their results different, or more impressive at least,
19 but they entered 20 percent of the patients in the 149
20 set and only five percent of the patients in the 200
21 set. Maybe we could hear -- ease our minds about both
22 of those problems.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
165
1 DR. FISHER: I haven't thought about the
2 timing of enrollment but maybe we could get slide 178
3 out. As a general principle in analyzing data, we would
4 like to include all of the data. We have to be very
5 careful about excluding things.
6 Having said that, it's perfectly fair game
7 to look for recruitment interaction. We all know
8 certain situations where data are appropriately
9 excluded. For example, investigator fraud, incredibly
10 poor quality data of such that it is just virtually
11 unbelievable, and so on.
12 In this particular case, the first point
13 I would like to make is even if there is a certified
14 treatment interaction, which I think is plausible
15 looking at these data, it appears to be quantitative.
16 In other words, if you remove Center 29,
17 in the remainder there is also an estimated favorable
18 effect. Of course, you are losing a lot of power. If
19 one center differs, there is an issue of what is a quote
20 and what is the truth. I don't think it is necessarily
21 clear they are a good or bad center and necessarily
22 somehow is so unrepresentative we shouldn't consider
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
166
1 it.
2 It's clear statistically that if the data
3 are homogeneous, the best center is going to look a lot
4 better than the average treatment effect. I mean,
5 that's obvious. It's also totally clear statistically
6 if things are homogenous and you remove the best center,
7 you are going to underestimate treatment effect.
8 What we have here are three possible ways
9 of dealing with things. The top is excluding the best
10 site, 29. As you can see, when you do this, if you look
11 at symptomatic AFAL statistical significance is lost,
12 if you look at any occurrence, symptomatic or
13 asymptomatic, if the log rank is just borderline and
14 the median test statistic is better, I ask them to throw
15 out the worst site just to illustrate the point.
16 I mean, once you start playing this game,
17 one great way to improve things is to say, "Well, gee,
18 the worst site looks so bad we want to toss this out."
19 If you do that -- I'm not advocating this but if you
20 do this, you can see, of course, you have greatly
21 improved things. No big shocker. Just like it's no
22 shocker when you throw out the best point and things
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
167
1 get worse.
2 If you treat things sort of symmetrically
3 down below and move in the best and the worst, things
4 are still statistically significant. This is despite
5 the fact that, of course, if you're going to do this,
6 you tend to lose statistical power because you are
7 decreasing the sample size as you do things. I actually
8 find this one a relatively easy part of the discussion.
9 We ought to include all the data.
10 It's possible there is an interaction but
11 it doesn't look to be qualitative enough. In other
12 words, there's no good indication I can see that in some
13 centers the drug works and in other centers it actually
14 has an adverse effect. I think everybody in medicine
15 assumes that when we approve a compound it works better
16 in some patients than others.
17 In other words, if this overall summary
18 relative risk is not a true effect in every single
19 patient but the continuity of human biology is such that
20 we are willing to say, well it works overall. There
21 may be differences and they are probably not that great.
22
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
168
1 DR. CALIFF: Lloyd, did you calculate the
2 probability that a result as seen in 29 could have been
3 -- what was the probability that could have occurred
4 taking into account the overall treatment effect? In
5 other words --
6 DR. FISHER: You mean if you only looked
7 at that one clinic?
8 DR. CALIFF: No. If you take the overall
9 effect and the total sample size, you ought to be able
10 to calculate the likelihood that you would get one center
11 that far in the extreme. Is it like 1 in 1,000 chance?
12 DR. FISHER: I hesitate to do that because
13 it's a totally data driven thing where you look at it
14 and say, "Oh, my goodness. In this study it looks like
15 we may have an interaction." I have made the statement
16 and I intend to write a paper in the next few years,
17 I've never seen a totally consistent clinical
18 development program. Never. There are so many things
19 going on. There is something weird. Well, how come
20 you don't get the same dose response here as there,
21 etcetera, etcetera.
22 I did do that using as a major, the spread
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
169
1 of the largest clinic which will almost always occur
2 in one of the few clinics. It's almost like a P value
3 on that one site. That's at about a one in 100 level.
4 That's why I say it's certainly conceivable to me that
5 there is a quantitative interaction.
6 DR. CALIFF: I would have thought the main
7 purpose of bringing this up would be to go look at the
8 site and just make sure there is nothing funny going
9 on there.
10 DR. FISHER: Yes.
11 DR. CALIFF: We've got some documentation
12 that that was done.
13 DR. FISHER: I agree that is always very
14 prudent.
15 DR. THADANI: Before you leave that point,
16 obviously the doctor in Center 29 has better healing
17 power than other physicians obviously. We go for
18 observations onward, but I was looking at the table
19 provided from the center and --
20 DR. FISHER: I'm worried about the turkey.
21
22 DR. THADANI: I realize that.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
170
1 DR. FISHER: The turkey is in 24 or it
2 doesn't work. How do we keep the drug out of evidence?
3 DR. THADANI: If you look at the large
4 centers -- I'm looking at Table 4 on the document I was
5 sent by the Center, faxed to me on 4/23. I don't know
6 if you have it or not. They look at large centers with
7 sample size of 40, 42, 32, 30, 25, and 24 subjects, fairly
8 large sample size. In that group the P value for Center
9 29 is 0002 and 00003, but none of the other centers
10 approach really anything like it.
11 So if you exclude that center the
12 significance literally disappears. I think we could
13 argue the center is very good. I don't know how you
14 are interpreting it, the symptoms at that center may
15 be very peculiar unless you audit the center.
16 DR. FISHER: As we can see from the log
17 rank, I mean, the significance does disappear.
18 DR. THADANI: The question always comes up
19 why is that center so peculiar which is driving the whole
20 database. Plus we have heard, for the first 200
21 patients there is no significance. Then you have the
22 next database which is highly significant but none of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
171
1 it is being driven by just one center which is obviously
2 recurrence rated variable.
3 CHAIRMAN PACKER: Let me just make sure
4 that we are not driving ourselves off a cliff here.
5 The only reason that I know of that the Center 29 issue
6 came up is not because -- tell me if I'm right--Center
7 29 is so materially different than the other centers
8 because we see that in clinical trials.
9 That is, some centers do have a greater
10 estimated treatment effect than others. I think the
11 only reason this came up -- correct me if I'm wrong --
12 is that Center 29 out recruited the other centers by
13 far after the 200 patient extension. If the 200 patient
14 extension didn't exist, would Center 29 be an issue?
15 DR. FISHER: Milt, my impression was it
16 wasn't that but it was the fact that a sizable segment
17 of the improvement in the time to recurrence was at that
18 center.
19 DR. THADANI: Milton, actually the
20 question is raised because I'm reading what the FDA sent
21 us, "Excluding the single study site, which is Center
22 29, the P value is no longer significant."
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
172
1 CHAIRMAN PACKER: No, but, Udho, it's not
2 fair. You can take a whole bunch of databases and throw
3 out the best site and the treatment effect disappears
4 because you're not only throwing out the vector but
5 you're throwing out -- you are reducing sample size.
6 I don't know how many databases would stand that kind
7 of assault.
8 DR. THADANI: I'm not saying to throw the
9 samples out. I'm just saying it raises some issues that
10 you've got excluding these 40 patients in the 360
11 patients. There's no benefit.
12 DR. FISHER: But it's not just reducing the
13 sample size. You are biasing in the estimate against
14 your study drug by taking out the best.
15 DR. KARKOWSKY: The initial protocol said
16 they would look at investigator cross-site
17 interactions. We had not seen that analysis. It is
18 fair game to do an investigative cross-site interaction.
19 The reason it was done is because we did not understand
20 why the study size was increased and we saw a disparity
21 in the first half and the last half of the study.
22 That was not data dredging. That was based
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
173
1 on the only analysis we did. We don't have the
2 facilities to data dredge. We have one statistician
3 and she's dredging the stuff that is supposed to be
4 dredging. Nevertheless, it did come up and then the
5 question was why were those last 150 patients different?
6 Now the protocol stipulated that they would do
7 investigative cross-site interaction so we felt very
8 comfortable looking at that.
9 Having found it we said let's take a
10 conservative analysis which would be take out that site
11 that looks the most deviant. We thought that was a
12 reasonable analysis. The things that might be
13 determined are, (1) the study doesn't find anything;
14 (2) the study may have found something but it's certainly
15 less robust than we initially thought it was and that's
16 why you guys are up there.
17 DR. KOWEY: Can I just make a suggestion
18 that Rob already made which is I'm sure the sponsor would
19 be very happy to have the site audited in detail to make
20 sure there were no irregularities at the site. If there
21 were no irregularities at the site, I think this
22 discussion probably is moot.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
174
1 CHAIRMAN PACKER: I assume the site has not
2 been audited.
3 DR. KOWEY: No. It was not a sponsor study
4 and it has not been audited. It's in Stockholm if
5 anybody wants to take a ride.
6 CHAIRMAN PACKER: Okay. JoAnn, other
7 issues on 04?
8 DR. LINDENFELD: This is just a general
9 issue. Is this on? Yeah. Maybe you can just give me
10 some insight if these are not beta-blocker effects that
11 are prolonging the time to atrial fibrillation, why is
12 it that d-sotalol is not effective? As we talked
13 earlier about some of the things we see beta-blocker
14 effects. Are all of them beta-blocker effects? We
15 don't know. I was just wondering if you could give me
16 some insight into that.
17 DR. KOWEY: Actually, my feeling about this
18 particular kind of drug is that having a beta-blocker
19 incorporated into the molecule is an important aspect
20 of the electrophysiological effect of the drug. It's
21 a combination of effects which I think are important.
22 I don't know why d-sotalol wasn't effective
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
175
1 in the study. I would have predicted that it should
2 have had some efficacy. In fact, it did. It wasn't
3 ineffective. It was just not as effective as the
4 racemic one. I think that the drug works best when you
5 do have some beta-blocking effect in addition to the
6 Class III effect. I think it's a composite effect.
7 DR. KONSTAM: Can I follow on that? What
8 percentage of the effect do you think -- I mean, you
9 made that statement so I guess it must be based on
10 something. What percentage of the effect that we see
11 here do you think is on the basis of beta-blockade?
12 DR. KOWEY: Yes. If I can have the first
13 Kaplan Meier curve from study 004 that we showed which
14 was slide 24. I don't want to be flip but I think it's
15 probably about half way between here and here. This
16 is what d-sotalol did as a pure Class III, this is what
17 the combination drug did, and this is placebo. So it's
18 probably an equal contribution of both parts.
19 I don't really know. I do know that when
20 you use the drug at 80 milligrams twice a day it's a
21 beta-blocker. I know when you use it at 160 milligrams
22 twice per day it's more than a beta-blocker.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
176
1 You are beginning to see Class III effect.
2 We'll show you that later. Clearly since there is a
3 difference between 160 and 80 in 9A, for example, you
4 need to have both effects. Since there's a difference
5 between d-sotalol and racemic sotalol, I think you need
6 to have both.
7 DR. KONSTAM: I'll just point out what's
8 missing from this slide is a pure beta-blocker.
9 DR. KOWEY: Yes. That's right. There is
10 no where in this database unfortunately anything that
11 I can show you that is a pure beta-blocker study.
12 DR. KONSTAM: Can you just clarify for us
13 non-electrophysiologists, you said that d-sotalol does
14 not have beta-blocker effect. It's completely devoid
15 of beta-blocker effect?
16 DR. KOWEY: Yes. A pure IKL blocker.
17 DR. KONSTAM: And with regard to the
18 complement of electrophysiologic effects is it
19 identical to d,l-sotalol or are there other differences?
20 DR. KOWEY: No. The d,l isomer's Class III
21 properties are exactly the same as the d isomer's Class
22 III properties.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
177
1 DR. THADANI: Before you leave that point
2 now on the -- sorry. The question is on d-sotalol.
3 Looking at table 10 provided by the center, d-sotalol
4 was not different than placebo in all 349 patients.
5 The P value was .206 and d,l-sotalol is 0003.
6 DR. KOWEY: No. I'm not saying it's
7 sophistically significant. I'm saying that it's not
8 the same as placebo.
9 DR. THADANI: The reason again I'm bringing
10 the center into action, it seems again that Center 29
11 had a P value of 0003 and 0004 even with d-sotalol.
12 I'm just wondering if the center is very peculiar,
13 although Milton's point is well taken that one center
14 can influence. It seems like a very peculiar center
15 showing a marked efficacy with the two active drugs.
16 DR. KOWEY: Can I have the core slide,
17 please, No. 54?
18 DR. FENICHEL: Actually, before --
19 DR. KOWEY: I'm sorry. Hold that a moment.
20 DR. FENICHEL: Hang on just a second,
21 Peter.
22 DR. KOWEY: I'm sorry.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
178
1 DR. FENICHEL: What we do know about that
2 center is that the performance of d,l-sotalol was not
3 all that different from that seen in other centers.
4 What was remarkable in that center was the placebo was
5 different. The placebo performed especially badly.
6 It wasn't that the active drugs performed especially
7 well. So you would expect to see a marked increase in
8 the apparent efficacy of both active agents.
9 DR. THADANI: So the question is could an
10 investigator somehow have known what his placebo because
11 it has symptomatic recurrence and he would say, "Okay,
12 if you had symptoms, he could ignore it.", you know,
13 I'm just not -- the investigator just makes me a bit
14 uncomfortable with two active drugs by showing a very
15 similar thing and placebo incidents were very low.
16 DR. FISHER: I just want to comment as a
17 statistician with a lot of multiple comparisons, when
18 you get your biggest effect, if they were all equal
19 sample size, you expect it at a clinic where the placebo
20 effect by chance is greater than expected and the active
21 therapy is better than expected. That's all a valid
22 part of the statistics taken into account in the total
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
179
1 analysis.
2 DR. MARROTT: Mr. Chairman, I would like
3 to make a couple of points. First is that in study 004
4 Center 29 when you look at the relapses for d-sotalol,
5 the comment was made that study 29 seems to favor both
6 d,l-sotalol and d-sotalol. That is not true because
7 if you look at the 14 patients that were recruited by
8 the center in the d-sotalol group, there were four
9 relapses in the symptomatic category.
10 Then you look at any category and it was
11 eight relapses in the any category. In fact, in the
12 active group, albeit sotalol, there were eight relapses
13 out of 14. I would consider that it wasn't like the
14 investigator knew that there were some symptoms so it
15 was an active group and the results came out because
16 of the bias towards the active group. So, I don't agree
17 with that comment.
18 The other point I would like to make is the
19 reason we are making all this discussion is because the
20 two groups, the d,l-sotalol group and the placebo group
21 and fortunately for the sponsor are heading in the
22 opposite direction. That is, whereas the active group
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
180
1 has benefitted with less relapses, the placebo group
2 has suffered with more relapses.
3 You know, if that would not have happened,
4 we would not have had this discussion because then we
5 would have been dealing with only one part of the
6 equation which is deviated from normal. Here what has
7 happened is that by chance I suspect that the two sides
8 have deviated on the opposite side.
9 I would also like to point out that as every
10 reputable company does, Bristol-Myers Squibb and
11 ourselves are not an exception, we go through the
12 monitoring of sites very diligently and very seriously.
13 We do this more so since this was one of the key trials
14 of the company.
15 When we looked at the listings and we looked
16 at some of the case report information, we could not
17 detect that anything deviant had happened at the center.
18 I can tell you that with the greatest degree of
19 diligence that we have tried to be very objective.
20 I will add one more comment, that the number
21 of patients who had structural heart disease was more
22 in the placebo group but I can't elaborate further on
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
181
1 that issue. I don't want to make any claims about that.
2 One more point, Mr. Chairman, very quickly.
3 A point was made that the d-sotalol did not benefit
4 but it has benefitted in study 9A so the data in study
5 9A does show that d-sotalol 200 milligrams BID does as
6 well as the d,l-sotalol. Thank you.
7 DR. THADANI: Pran, the question I was
8 saying to study this study. I did not say the other
9 study. I was only referring to this database.
10 CHAIRMAN PACKER: Let me just see if I
11 understand. Abe, just clarify. There was or was not
12 a statistically significant treatment by center
13 interaction in 04?
14 DR. KARKOWSKY: We did not see an analysis
15 of treatment by center interaction.
16 CHAIRMAN PACKER: Okay. Has the sponsor
17 performed such an analysis?
18 DR. KOWEY: No. It's not been done yet.
19 CHAIRMAN PACKER: Okay. And the division
20 has not performed an analysis?
21 DR. KARKOWSKY: I certainly don't know how
22 to do it. I've spoken to the statisticians and it may
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
182
1 not be that easy to do.
2 DR. FISHER: The reason it wasn't done was
3 there was so many sites with such small numbers of
4 patients. At least the asymptotic statistics would be
5 greatly endowed. There might be somebody somewhere who
6 could have a program to do an exact analysis.
7 CHAIRMAN PACKER: Isn't the way that one
8 deals with a small number of centers, a small number
9 of patients, is to create pseudo centers where the
10 centers are pulled?
11 DR. FISHER: Well, it was done by
12 geographic region. There was no interaction. I said
13 that's not what the agency is talking about because then
14 you are to some extent washing out the 29 effect. The
15 problem is doing this all post-hoc.
16 If you know you have a problem at a big
17 center, then you'll be lumping almost everything else,
18 or a huge number of everything else, into this one
19 mega-center which is almost like saying the mean, which
20 might have -- if somebody had written up in the protocol
21 perspective, that might have some merits for looking
22 at big centers. After you look at the data and see it's
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
183
1 triggered by that, it's very hard to know what would
2 be the right thing to do.
3 CHAIRMAN PACKER: Alexandra?
4 DR. KAPATOU: Yes. Another point I wanted
5 to make in this was that since we actually did a
6 nonparametric analysis. The log rank test is a
7 nonparametric analysis. There is no direct way of
8 studying the treatment by center and direction except
9 making tables like the ones I presented. If we had
10 parametric model, then we could have put an additional
11 term and that would have taken care of it.
12 DR. FISHER: Well, if we had bigger numbers
13 there is a way to address it. The Cox model is
14 semiparametric. It's nonparametric with time to event
15 but parametric with respect to covariates and you could
16 put in indicator variables for clinics and look at the
17 sum degrees of freedom and interaction.
18 The problem here, as I mentioned, is the
19 small numbers in many of the sites. If there were five
20 centers all of which had 20 or more people enrolled,
21 then I think it would be fairly clear how to attack it.
22 CHAIRMAN PACKER: Maybe I can ask Bob
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
184
1 Fenichel. Bob, we don't see treatment by center
2 interactions on primary endpoints all that common but
3 we have seen a few examples over the past five to 10
4 years of this observation. How has the division
5 approached this in the past when a treatment by center
6 interaction occurred?
7 Obviously if a treatment by center
8 interaction occurred in a non-major trial or on the
9 secondary endpoint, people probably wouldn't spend a
10 whole lot of time talking about it. Just suppose
11 something like that was seen in a major trial on its
12 primary endpoint. What approach has division taken or
13 has the policy been not clearly defined even in the past?
14 DR. FENICHEL: I think it's very hard to
15 establish a perspective policy on this. I think some
16 of what Lloyd has just said is pertinent that it's quite
17 difficult when the outstanding center is indeed also
18 the biggest center in which one also has the sort of
19 systemic feel that, well, there's a time interaction
20 in that the late arriving patients in this trial look
21 different from the early ones. Well, that's because
22 -- that is because, in some sense of because.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
185
1 This center became more active as the trial
2 went on. This center was the biggest center and
3 increasingly so as time went on. There are all these
4 things confounded and the usual strategies are sort of
5 worst-case analysis, if you like, where you say just
6 throw it out. There's a small center that is the outlier
7 and maybe it's because there has to be somebody who is
8 the outlier.
9 Well, that's fine. We'll just throw it out
10 and you'll see the results are kind of the same. You've
11 lost a little power but it's all heading in the right
12 direction. Here, you know, that doesn't quite apply.
13
14 Certainly the only recent experience with
15 a dramatic effect that was somewhat similar to this was
16 in one of the epolifibitide studies where there were
17 dramatic region by treatment interactions -- region by
18 treatment by gender interactions, three ways, that the
19 stuff seemed to work a little bit better in men than
20 women all over the place.
21 But then in Latin America it was actually
22 worse than placebo in women. Well, what are you going
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
186
1 to make of that? We don't understand it. We don't
2 regulate drugs in Latin America. We regulate them here
3 and so we just put it into labels and say, "Look at this.
4 What do you think of that?" That's where it stands.
5 I don't know that some succinctly
6 describable policy can be distilled from what we've
7 done. Certainly it has never been set forth
8 perspectively as a guide to our behavior.
9 CHAIRMAN PACKER: Mike. Mike has been
10 waiting for a long time.
11 DR. CAIN: Can we have slide 24 put back
12 on the screen, please? I just wanted to get
13 clarification both from the sponsor and from the
14 division about a point that was made earlier, and that
15 was study 004 seems to be the only one in which we do
16 have follow-up of individuals who dropped out of the
17 study because of adverse effects. We do know the
18 natural history then of what happened to those people.
19 The sponsor made a comment, if I remember
20 it, earlier that the data that are presented for study
21 004 takes into account that additional data. This has
22 been a very key slide. If this is the perfect world
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
187
1 of where we've taken into account follow-up of those
2 people who were discontinued, then some of the issues
3 we're talking about this morning are resolved. If it's
4 not, then I would like to know that.
5 DR. KOWEY: Mike, this is slide 24 and let
6 me show you slide 29 which is adding in all the patients
7 who were discontinued or died during the course of the
8 study. Again, this is the analysis that we have been
9 talking about, the 05, all morning. Looking at the same
10 kind of analysis, I won't say it's the worst case. It's
11 the semi-worst case analysis, showing that these people
12 died.
13 CHAIRMAN PACKER: But, Peter, just to make
14 the point, although Michael's point is important because
15 04 did follow people all the way through, the percentage
16 of people who are dropping out here is much smaller than
17 all the others because it's six percent versus 29
18 percent.
19 DR. KOWEY: Yes.
20 CHAIRMAN PACKER: Consequently, even if
21 one didn't get complete follow-up, and they did, but
22 even if one didn't, the impact of a worst-case analysis
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
188
1 here, even my proposed worst-case analysis, would be
2 very small because the number of patients for whom data
3 would be missing is very, very small.
4 DR. FISHER: The sponsor does have what was
5 being asked for and it's on the slide including all of
6 the follow-up data even after people discontinued and
7 it looks the same.
8 CHAIRMAN PACKER: Any other issues on 04?
9 I have one other question on 04. You showed, Peter,
10 a subgroup analysis of patients above and below a
11 creatinine clearance of 60.
12 DR. KOWEY: Yes. That's slide 28.
13 CHAIRMAN PACKER: Was there a interaction,
14 a P value for the difference of the estimated treatment
15 effect in the people above and below 60? Is there a
16 P value associated with this?
17 DR. KOWEY: You have that as a backup. We
18 have efficacy by creatinine clearance.
19 CHAIRMAN PACKER: Because the sense is, and
20 one usually sort of gets a poor man's estimate of an
21 interaction by seeing to what degree the competence
22 interval is overlapped. The competence intervals here
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
189
1 don't overlap very much. The implication, if that were
2 to be the case, was that this drug primarily works in
3 a population which would be the smallest subgroup of
4 this trial. That would be a very strange kind of
5 conclusion.
6 DR. FENICHEL: Milton, do you think this
7 might be just a concentration proxy because it does seem
8 to work better in women? Not in every slide that Peter
9 showed but in most where it was separated out by gender
10 there was sort of a trend working better in women which
11 may just be size.
12 Here it might be that the stratagem they
13 used to correct creatinine clearance and reduce the dose
14 did not completely correct for creatinine clearance and,
15 therefore, the people with lower creatinine clearance
16 or poor renal function were, in fact, getting more drug
17 or a higher AUC anyway because certainly there is a
18 strong tendency toward a dose response with the drug.
19 I think this is all just a proxy for that.
20 CHAIRMAN PACKER: I hadn't actually
21 considered that because I guess I assumed that the
22 algorithm they used corrected adequately. But if, in
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
190
1 fact, it is a dose dependent phenomenon, it would suggest
2 the possibility that the target dose in this call which
3 was 160 BID, which is the highest recommended dose in
4 the proposed labeling, is an inadequate dose. I don't
5 want to go there.
6 DR. FENICHEL: Yes. I know. You're
7 getting more adverse effects as you go up, too. It's
8 a tradeoff.
9 DR. KOWEY: Let me just say, Milton, I think
10 Bob's explanation is accurate because the exclusion was
11 50 CC per minute. You know as well as I do that there
12 is a lot of error within that measurement. We're
13 dealing with a very tightly defined patient population
14 that probably works in people that did get more of the
15 drug.
16 DR. FISHER: I just ask what happened in
17 005. If we could see slide 49. This suggests to me
18 the other slide is a proxy for the multiple comparison
19 problem. I don't see that shown but we don't see the
20 same thing here.
21 DR. FENICHEL: Well, it could be a
22 different range of observed creatinine clearances so
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
191
1 an effect would be amplified in one population and not
2 so much in the other. I don't know. I don't think there
3 is any explanation.
4 DR. FISHER: But then I would suggest that
5 the agency, of course, can have them explore this further
6 by looking at estimated creatinine clearance and going
7 into it and body weight.
8 DR. THADANI: But wasn't the trial also
9 different in the first study? He said the criteria was
10 60 ml, although some patients just happen to fall in
11 because people were not very careful and didn't know
12 how to calculate creatinine and they were 50s. This
13 one brought it out in 40. That might be the difference.
14 DR. KOWEY: That's right. In that range
15 of 40 to 60 they got it once a day.
16 DR. THADANI: So that could have the
17 difference.
18 DR. KOWEY: Or they should have gotten it
19 once a day.
20 DR. THADANI: Another thing I think looking
21 at these trials, the side effect profile was much lower
22 in all four while it was much higher because of the dose
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
192
1 response design study in the 05. That might have
2 practical implications because here you've titrated
3 them.
4 DR. KOWEY: You'll see that.
5 DR. THADANI: That might be very relevant
6 because if you have intent to treat here, it looks very
7 highly significant. Yet, in the other one if you have
8 intent to treat it falls apart because of the large
9 dropout rate because of side effects.
10 DR. KOWEY: That's a possible explanation.
11 DR. THADANI: I think it's worth keeping
12 it in mind.
13 CHAIRMAN PACKER: Ileana.
14 DR. PIÑA: I want to come back to a point
15 that Marvin was marking before and that, Peter, you
16 alluded to, the beta-blocking effects of the drug may
17 be more significant at a lower dose and as you go up
18 on the dose you start to get the QT prolongation.
19 Actually, they should have data on this because
20 there were comparative studies with atenolol and timolol
21 when the drug was being studied for ventriculary
22 arrhythmia so there should be some comparative data on
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
193
1 beta potency or beta-blocking effects to answer the
2 question as to where would a beta -- obviously it would
3 be theoretical as to where a beta-blocker would fit in
4 there.
5 I have not reviewed the studies in great
6 detail but I know that they exist and this was very early.
7 Some of the doses are higher but some of the doses are
8 at 160 and 180. DR. KOWEY: At the end of the
9 presentation we'll talk about dose recommendations.
10 We're going to show you some of the data relating to
11 that question.
12 DR. THADANI: Isn't part of the data is by
13 80 milligram dose was not more effective?
14 DR. KOWEY: It depends on what study you're
15 talking about. I know in A it was.
16 DR. THADANI: No. In the first study you
17 show --
18 DR. KOWEY: In 05 it was.
19 DR. THADANI: It was marginal. It was not
20 effective.
21 DR. KOWEY: Right. Borderline effective.
22 CHAIRMAN PACKER: I think the study that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
194
1 primarily supports the efficacy of 80 BID is dofetilide
2 345.
3 DR. KOWEY: Do you like that one?
4 CHAIRMAN PACKER: Rob?
5 DR. CALIFF: This maybe is headed to where
6 I'm going eventually anyway.
7 DR. KOWEY: But, Milton, the reason I said
8 9A and you said 345 is we were talking about paroxysmal
9 population in 05. You're right in terms of the
10 robustness of the feedback. I'm sorry, Rob.
11 DR. CALIFF: We're picking apart each
12 individual study. Are you going to show us any
13 composites of the entire database? For example, this
14 question about creatinine clearance. It seems silly
15 to me to look at each individual study when you've got
16 a provided database with a much larger sample.
17 CHAIRMAN PACKER: Are you going to be going
18 over safety as it relates to creatinine clearance?
19 DR. KOWEY: Yes.
20 CHAIRMAN PACKER: We'll get some of it
21 then. But you want to know efficacy. Right?
22 DR. KOWEY: Yes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
195
1 CHAIRMAN PACKER: Do you have efficacy as
2 it relates to creatinine clearance combined across the
3 trunks?
4 DR. KOWEY: We have it for 04 and 05.
5 CHAIRMAN PACKER: Individual.
6 DR. KOWEY: Unfortunately, a lot of these
7 other studies, Rob, were not in an electronic database
8 so it's really hard to pull that kind of detail.
9 DR. KONSTAM: The problem with doing that
10 is, as Bob is suggesting, which I concur, is if it's
11 likely to be related to the levels and effect on levels,
12 then it's an interaction between the renal function and
13 the dosing regimen. Since the dosing regimes are
14 different, you would have to really think about that
15 a little bit.
16 I just want to say that I am concerned about
17 this creatinine clearance break in 004. Let me say it
18 the way I see it. It looks like the treatment effect,
19 which I think to me is more robust in 04 than it is in
20 05, is driven principally by the group with the low
21 creatinine clearance, if you want to say it that way.
22
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
196
1 The only rationale that comes to mind to
2 explain that is Bob's rationale that that's a group that
3 the correction per dose didn't work perfectly so you
4 wind up saying that it's conceivable then that it works
5 only with where you have higher concentrations. We do
6 believe that the adverse effect profile is going to be
7 influenced by the concentrations.
8 DR. KOWEY: These are the data from 004
9 looking at the covariate adjustment by Cox or creatinine
10 less than 60. Does that help you?
11 DR. FENICHEL: This is not the pertinent
12 analysis.
13 DR. KOWEY: This is not the subgroup
14 analysis.
15 DR. FENICHEL: No. This is a justified --
16 DR. KOWEY: This is a justified baseline.
17 DR. FENICHEL: That's right. If you had
18 something that was driven by some little subgroup, then
19 you might come up with a different result here and
20 something might stand out. What you want is to take
21 any of these various categories such as sex or age or
22 creatinine clearance and show as you have in some --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
197
1 DR. KOWEY: The next slide.
2 DR. FENICHEL: -- how it looks with people
3 with low creatinine clearance and how it looks to people
4 with higher creatinine clearance. You never have
5 weight. Weight would be good. How it looks in big
6 people, how it looks in small people. Of course, gender
7 is kind of a proxy to that but not perfect.
8 DR. KOWEY: I agree. Both of these, I
9 think, help a little bit. This one helps a little bit
10 and the other one helps a lot. The subject analysis
11 we showed you already.
12 CHAIRMAN PACKER: Any other questions on
13 04? If not, I think we need to re-energize. We're going
14 to break for lunch. The study will talk about and take
15 questions on after lunch until 3:45. We'll reconvene
16 at, let's say, 1:30.
17 (Whereupon, the meeting was recessed at
18 12:44 p.m. to reconvene at 1:30 this same day.)
19
20
21
22
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
198
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
199
1 A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N
2 (1:33 p.m.)
3 CHAIRMAN PACKER: We will be starting in
4 the next minute. Is the sponsor ready? We have an
5 addition to the administrative issues for this morning
6 so Joan will complete that at the present time.
7 MS. STANDAERT: Yes. For the record, in
8 a memorandum I read earlier the exclusions were omitted.
9 Two committee members were excluded from the
10 discussions. That would be Dr. Roden and Dr. DiMarco.
11 Thank you.
12 CHAIRMAN PACKER: Okay. Thank you.
13 Before we start a discussion on dofetilide 345 I think,
14 JoAnn, you had one other question you wanted to address.
15 DR. LINDENFELD: I do. Just a general
16 question about both studies. As we talk about
17 symptomatic recurrence, I want to clarify this point
18 because I think there's a difference in -- I think I
19 said this earlier but I didn't say it strongly enough
20 -- there's a difference in a patient who has symptoms
21 that are bothersome to a patient and the patient who
22 that just notices they have a rhythm change.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
200
1 I wonder if there was any collection of
2 symptoms that we might consider important more than
3 occasional palpitations such as shortness of breath,
4 fatigue, dyspnea at the time of atrial fibrillation is
5 noticed? Or when we say symptomatic here, is all we
6 mean is that the best we know is that the patient noticed
7 a change in their heart rhythm?
8 DR. KOWEY: We actually have that on a back
9 up -- what's the number?
10 DR. LINDENFELD: I know we have baseline
11 symptoms but I haven't really seen --
12 DR. KOWEY: No. We have symptoms. 173?
13 This is from 004. This is change from baseline to
14 endpoint for 004.
15 DR. LINDENFELD: So there's no difference
16 between placebo and sotalol? The symptoms were not
17 different?
18 DR. KOWEY: That was an endpoint. That was
19 looking at it from change from baseline to endpoint.
20 This is the one. This is probably a better one. This
21 is from 05.
22 DR. LINDENFELD: I sort of like the other
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
201
1 one. Can we look at that again?
2 DR. KOWEY: Sure.
3 DR. LINDENFELD: The first one. If we take
4 all the patients, there was no difference in these
5 important symptoms from baseline to endpoint.
6 DR. KOWEY: No. There was a 13 percent
7 reduction in the d,l-sotalol for any symptoms compared
8 to nine percent placebo. There was a 14 percent
9 reduction.
10 DR. LINDENFELD: But that's not
11 significant. Is it?
12 DR. KOWEY: I don't think that there were
13 P values calculated for these observations.
14 DR. LINDENFELD: You don't think it would
15 be significant if it were probably.
16 DR. KOWEY: You are welcome to look at this
17 if you want to look at the symptoms which I think is
18 what you want is the other slide.
19 DR. LINDENFELD: The reason why I'm making
20 this point is because as we talk about approving this
21 drug for symptomatic atrial fibrillation, this doesn't
22 really evaluate symptoms as we usually think about them
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
202
1 and study them; that is, shortness of breath, fatigue.
2
3 This evaluated whether or not the patient
4 noted primarily that they had a difference. I guess
5 we could look at that and you, if I'm wrong about that,
6 can show me. So from what symptoms we have, at least
7 we knew before there was no difference in the two groups,
8 no major difference.
9 DR. KOWEY: Wait, wait, wait. The
10 endpoint of the trial was the time to symptomatic
11 occurrence of AF.
12 DR. LINDENFELD: Right. But I think
13 there's been some confusion in here in the fact that
14 we are ameliorating symptoms. In other words, there's
15 a difference between symptomatic recurrence of atrial
16 fibrillation which is not serious. It may have been
17 in some patients. They may have had more shortness of
18 breath and more fatigue.
19 DR. KOWEY: Here is the percentage of
20 patients who had specific symptoms during their return
21 to symptomatic atrial fibrillation flutter by dose in
22 05. This is the question you asked which is what was
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
203
1 the symptom that they had at the time that they had their
2 recurrence. This is it.
3 DR. LINDENFELD: And it doesn't look like
4 to me there are any major differences.
5 DR. KOWEY: No, no, no, no.
6 DR. LINDENFELD: Okay. Make sure I
7 understand.
8 DR. KOWEY: This is not an endpoint. This
9 is just telling you what the symptom was when the patient
10 had their recurrence.
11 DR. FENICHEL: JoAnn, suppose this were a
12 mortality trial then you find that at endpoint everybody
13 is dead.
14 DR. FISHER: If you look at the ends at the
15 top there are different numbers experiencing the
16 recurrence. See this at recurrence.
17 DR. KOWEY: You had to have a recurrence
18 to get on this slide.
19 DR. LINDENFELD: At recurrence. Okay.
20 CHAIRMAN PACKER: Okay. I just want to
21 keep moving and move on to dofetilide 345. Just to
22 clarify the record, Dr. Kowey indicated during the break
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
204
1 that the information on dofetilide study 345 was
2 actually obtained from the Internet. I don't know how
3 many of you know that everything you see today can be
4 accessed through the Internet. I guess that shouldn't
5 be too surprising. You can access anything in the world
6 through the Internet these days. I just wanted to --
7 DR. KOWEY: Including what happened at
8 Center 29.
9 CHAIRMAN PACKER: Peter, it's not common
10 for a sponsor to utilize another sponsor study to support
11 approval. There are a lot of reasons for that. One
12 is that most commonly sponsors don't do comparisons
13 against drugs not approved for the indication that is
14 being pursued.
15 DR. KOWEY: Correct.
16 CHAIRMAN PACKER: And also I think
17 frequently a lot of times the studies that are carried
18 out by a sponsor tend not to demonstrate that the
19 competing drug works. Consequently, there is little
20 enthusiasm to use it. I think the concern that I have,
21 and maybe the other members of the committee share it,
22 is that when a trial is reviewed by the FDA, that has
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
205
1 an active component that is being reviewed in the NDA
2 and has the comparator.
3 There is a lot of attention given to the
4 quality of a comparison for the active drug being
5 considered and not necessarily a lot of attention being
6 given to the comparator. In other words, checks of
7 integrity, completeness, all the things that the FDA
8 does are frequently not done, for example, in dofetilide
9 345 but sotalol arm because the sotalol arm isn't the
10 arm on which a claim is being sought.
11 I guess one question that I have is to what
12 degree can we utilize the sponsor's presentation in what
13 may literally be for the purposes of today's discussion
14 a study in which the integrity of the sotalol database
15 has not been as carefully evaluated as the integrity
16 of the dofetilide database in study 345.
17 DR. KOWEY: Can we ask Bob what his opinion
18 is about that?
19 DR. FENICHEL: I was hoping you wouldn't
20 ask that. I think the answer is that this committee
21 can use anything it wants. I have heard members of this
22 committee refer to their vast clinical experience.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
206
1 Indeed, we recruit members of this committee in a
2 slightly different sense of that phrase because of their
3 vast clinical experience. So even if that experience
4 is not explicitly referred to in your every remark, it
5 is taken to carry the weight to a certain extent.
6 Now, so may you use 345 in supporting
7 sotalol? Yes, you may. Is the FDA able to use 345 in
8 supporting sotalol? I'm not really sure. I think the
9 answer is probably not. What I'm drawing on is we
10 certainly have experience with, say, competing
11 sustained release products for common chemicals like
12 verapamil, propriadin.
13 The question is, well, can the new sponsor
14 make use of the animal toxicology data for the existing
15 product? The answer is no, not without a right,
16 presumably a purchased right, to refer to that data which
17 is owned by the original sponsor or sponsors.
18 I imagine something like that applies here
19 and that if there were some intended claim supported
20 only by data from 345, I think that would be very
21 problematic because that's the strict analogy to the
22 animal toxicology case where there isn't anything that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
207
1 people know about verapamil in rats except from those
2 studies that were done on rats.
3 Someone who comes along with a new sustained
4 release form really doesn't have any intention of doing
5 experiments in rats if he or she doesn't want to and
6 doesn't have to. What in some sense we have done is
7 they have to unless they can get right to reference the
8 previous work.
9 Here it might be true that this is part of
10 a big picture that there isn't some unique fact that
11 can be found only in 345. If there is a unique fact
12 only in 345, my guess is that this sponsor can't use
13 it. I'm not sure that is correct.
14 CHAIRMAN PACKER: The reason for bringing
15 this up is, (1) it's unusual and, therefore, we need
16 to talk about it. It may or may not be relevant and
17 the committee will probably make clear in its subsequent
18 comments how important 345 might be in their
19 deliberation.
20 One concern I have is that it's difficult
21 for us to ask you questions about 345 that you can answer.
22 In fact, it's almost impossible for us to ask you the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
208
1 same kind of questions about 345 that we're asking about
2 04 or 05 or the other studies. I mean, if one were to
3 ask about dropouts, completeness of follow-up, issues
4 related to symptomatic or nonsymptomatic, it's hard to
5 get those answers because my presumption is that the
6 details of that, which are so important to our
7 deliberation of 04 and 05 are known primarily to the
8 sponsor of 345 and may not be known to either you or
9 the committee.
10 DR. KOWEY: Let me just say a couple of
11 things about that. First of all, I don't disagree with
12 you at all. I also agree with Bob that if this were
13 coming out of the blue as a totally novel concept, that
14 we would be concerned and would not have presented it.
15 Two issues; (1) It is being used in this
16 context to provide reassurance that a dose of 80
17 milligrams twice per day is effective. In fact, I could
18 turn this argument around a bit that you are using about
19 credibility of data and say that as a positive comparator
20 it was the last thought on Pfizer's part that they wanted
21 to show up the 80 milligrams twice a day. In fact, when
22 they presented it to the advisory committee in January,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
209
1 I remember looking at it and thinking, "Okay." I passed
2 it over. I didn't even think about it.
3 It was only after it was pointed out to me
4 later that, gee, 80 milligrams twice per day did really
5 well in that study. I think in reverse fashion,
6 although I agree there are problems with validation,
7 it actually provides me some reassurance that the
8 observations that we've made in studies where we have
9 an interest or this sponsor has an interest were made
10 by somebody who really didn't have an interest. It is
11 a novel concept and I agree with you, Milton. I don't
12 remember ever having seen this before, this drug
13 dofetilide is not yet approved.
14 CHAIRMAN PACKER: The concern I have is not
15 its novelty. The concern I have is it is our ability
16 to interrogate the data and the issues with the same
17 degree. I mean, our reflex would be to say, gee, that
18 P value looks pretty small.
19 DR. KOWEY: Is there some way for the agency
20 if they thought they needed to do that to interrogate
21 the data on a more rigorous basis?
22 CHAIRMAN PACKER: The problem is --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
210
1 DR. FENICHEL: If you buy the rights to the
2 data from Pfizer, I'm sure he'll do it.
3 DR. KOWEY: I don't have a check with me
4 but I'm sure --
5 DR. FENICHEL: Look, I think that is a real
6 problem but what I would recommend to the committee is
7 I don't think this is different from the problem that
8 arose with Center 29. I don't think it's different from
9 the problem that arises implicitly all the time in
10 looking at files, which is to say, well, maybe these
11 results will not survive audit. Maybe the agency was
12 convinced after looking at the data. But then when we
13 send the DSI person around to the site, he finds that
14 the patients were made up or whatever.
15 The committee should proceed with the data
16 on its face. There may be questions that cannot be
17 answered by the sponsor because of peculiar
18 circumstances here. There will be other questions that
19 are similar to the questions that always arise in terms
20 of audit not having been done. Of course, that's true.
21
22 I think as regards the rights to this data,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
211
1 if that ever becomes pivotal in trying to establish some
2 otherwise unsupported fact that the sponsor wants to
3 assert about sotalol, I think that's going to go
4 ultimately to -- it's not going to be my opinion. It's
5 going to be something that the FDA general counsel deals
6 with. The way to deal with lawyers is not to ask them
7 what to do. It's tell them what you want to do and then
8 see how they will allow you to do it. That's the same
9 thing here.
10 DR. FISHER: Could I ask Bob a
11 hypothetical? What would happen if sotalol in 345
12 actually looked harmful or it had a lot of adverse event
13 data. You could not explicitly consider that? I mean,
14 that would be a horrible thing to happen.
15 DR. FENICHEL: Well, you're right. I
16 don't know what the answer is. I think the answer is
17 that in that setting this would be like the sponsor is
18 obligated under law to inform FDA of whatever it knows
19 about the drug. Very often people tell us, "Well, here
20 is some literature of some guy's study reported
21 somewhere. We don't have the data and we don't know
22 very much about it but we found it so we're passing it
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
212
1 on."
2 Most of the time that stuff doesn't amount
3 to much. Sometimes it is. Well, sometimes it's stuff
4 like adverse reaction. Here is a report from some
5 minor journal somewhere of an adverse reaction to our
6 drug. We don't even know if it's true. We don't know
7 very much about it.
8 Sometimes on the strength of that we tell
9 the firm, "Look. This sounds suspicious. You ought
10 to study that." The firm goes back and studies it.
11 We do collect adverse things and we are able to consider
12 them in that light.
13 You know, I suppose to carry your
14 hypothetical to the extreme, suppose that 345 were very
15 large and what it had demonstrated had been not that
16 sotalol 80 milligrams was pretty good, but rather that
17 80 milligrams of sotalol killed everyone who took it.
18 Well, what would we do with that? That would be very
19 tough. We would have to deal with that. That's not
20 at all like the current situation.
21 CHAIRMAN PACKER: Okay. I think we have
22 at least noted the issues. I think we have addressed
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
213
1 the issue as much as we can. I would ask the committee
2 that to the extent that 345 influences any of your votes
3 or deliberations, you should make it clear so that it
4 is clear to the division. If it's irrelevant, that's
5 fine. If it's relevant, please make clear that it is
6 relevant. It's just that we can't really do a whole
7 lot about asking the sponsor questions about a study
8 they didn't do.
9 JoAnn, I think you had one more question
10 before we move on to the next thing. You're fine? Okay.
11 We'll move on to study 014. JoAnn.
12 DR. LINDENFELD: We're told that three
13 subjects were entered twice into 161. Is that correct?
14 I was wondering if you could tell me what the study
15 looks like without those patients who were considered
16 twice, or did it make any difference?
17 DR. KOWEY: The company will address that.
18 DR. LINDENFELD: I don't know if it helps.
19 Page 69 of our briefing document says that three of
20 the subjects were withdrawn and then re-enrolled. It
21 seems like an unusual --
22 DR. MARROTT: We will check by the board
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
214
1 and come back to you on this question.
2 DR. LINDENFELD: And then just clarify for
3 me subject dropouts were followed, not followed?
4 DR. MARROTT: They were not.
5 DR. LINDENFELD: They were not found just
6 as in 05.
7 DR. KOWEY: Is that correct?
8 DR. MARROTT: That is correct.
9 CHAIRMAN PACKER: Just to clarify, about
10 20 percent of the patients had AEE's and didn't have
11 follow-up. Is that about right?
12 DR. KOWEY: That's correct.
13 CHAIRMAN PACKER: Any other questions --
14 DR. KOWEY: It is actually 15 percent. Do
15 you want to see it?
16 CHAIRMAN PACKER: No. Any other questions
17 from any other member of the committee on study 014?
18 Okay. Let's move onto study 9A.
19 DR. LINDENFELD: One question on 9A. Let
20 me see if I got this right. The time to recurrence of
21 arrhythmia meantime was six days with placebo and 13
22 and 18 days in the two sotalol groups. Is that correct?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
215
1 DR. KOWEY: Give me one second and we'll
2 put the slide up.
3 DR. LINDENFELD: Just a rough idea if that
4 is correct.
5 DR. KOWEY: It's slide 54, please.
6 DR. LINDENFELD: I guess maybe other people
7 want to comment on this. It brings up what is
8 statistically significant and what's clinically
9 significant in terms of -- I know these are recurrent
10 arrhythmia so maybe this makes more of a difference.
11 DR. KOWEY: Yes. Remember this had a
12 fairly arcane running so that we were collecting data
13 by frequency of occurrence. It was stratified by the
14 amount of time the patients were watched. In addition,
15 the analysis was also adjusted for that period of
16 observation. This is clearly a group of patients who
17 have very frequent arrhythmias judging from the placebo
18 time to relapse.
19 Because there was such an enriched patient
20 population, P values for the differences between the
21 groups are significant. Not only for d,l-sotalol but
22 also for d-sotalol. Clearly it was a different patient
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
216
1 population.
2 DR. FENICHEL: Peter, let me take off on
3 something in JoAnn's question and that is, of course,
4 each of these trials was analyzed as a survival trial.
5 If this were literally survival and the endpoint was
6 death, one might say they all died within a couple of
7 weeks. Does it really make that much difference?
8 Then, of course, it is recurrent arrythmia
9 and what we see in each of the trials going back to,
10 I think, the flecainide trials is that the analysis has
11 been a survival analysis with the assumption that the
12 Poisson parameter, if you like, that the time to
13 recurrence to the first recurrence is somehow typical
14 and representative of the subsequent time between
15 recurrences which is a plausible thing to assert.
16 But I don't know, and maybe Ed Pritchett
17 wants to speak about this, because I think this is a
18 conceptual issue in this area. Is that well validated
19 that this kind of analysis is, in fact, predictive of
20 what would happen over the course of months or years
21 of continuing therapy.
22 DR. PRITCHETT: There are several lines of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
217
1 data that support that. First, we know that in
2 individual patients that the individual occurrences,
3 serial occurrences of a symptomatic supra ventricular
4 arrythmia constitute a Poisson process as you said.
5 Secondly, we know that in a group of
6 patients -- by the way, these two observations were
7 published in the same paper in Circulation in 1981 --
8 in a group of patients if you measure the time to first
9 event and then the time between the first and the second,
10 the distributions are identical. They sit right on top
11 of each other.
12 Finally, you may recall from the flecainide
13 program where in that program there was an attempt to
14 measure to record four events during the follow-up
15 period and the primary analysis was timed to first event.
16 If you look at the average time between events for four
17 events and the ratio between the flecainide and the
18 treatment group, it came out to be the same as for the
19 time to first event. The time to first event is a good
20 estimate apparently of what this does to the rate of
21 recurrences over time.
22 DR. THADANI: On this study on the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
218
1 recurrences, by how frequent ECG is monitored? Is it
2 Holter or again just the transmission of 25 seconds?
3 Because if it is not Holter, as you said, the Holter
4 shows 10 to 12 times more than this one. How much
5 reliance can one place if it's not the Holter?
6 DR. MARROTT: It's not Holter.
7 DR. THADANI: So it's just transtelephonic
8 monitoring. As we have heard, the reliance that could
9 vary from incidents 12 times less than the Holter. How
10 much reliance with a sample size so small one can't
11 compare anything on it? Is it possible that if done
12 more frequently you'll have more episodes than other
13 groups?
14 DR. KOWEY: It's certainly possible. By
15 the way, the endpoint here, the ECG documented
16 recurrence of atrial fibrillation. It doesn't
17 necessarily mean that it has correlated its symptoms.
18 DR. THADANI: Sure. It could be just on
19 that 20 second recording which is negative and the event
20 could have been positive.
21 DR. KOWEY: That's right.
22 DR. PRITCHETT: I might just comment with
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
219
1 respect to, as you just said, the important thing is
2 that this is a symptomatic outcome. Also with respect
3 to the treatment effect here, I think Peter pointed out
4 this is a very, very active group of patients with a
5 median recurrence time of six days.
6 What you see with the 80 milligram dose is
7 a doubling of the median time and then the 160 a tripling
8 by the standards that the committee has used in the past
9 which said that the minimum useful effect would be a
10 doubling. This certainly meets that.
11 CHAIRMAN PACKER: Abe?
12 DR. THADANI: It's only a nine week study
13 and the frequency of monitoring could have been two or
14 three times.
15 DR. PRITCHETT: No, no. The patients were
16 called in when they had a symptomatic event.
17 DR. THADANI: It was only a nine week
18 double-blind study.
19 DR. PRITCHETT: At the median time of event
20 the placebo group was six days, it could have been a
21 lot shorter than nine weeks and all the useful
22 information would have been captured. This is a very,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
220
1 very active group of patients.
2 DR. KARKOWSKY: What I would do is like to
3 make a couple of things clear. The first thing is that
4 the analysis for the overall study if one includes
5 discontinuation as having bad outcomes I don't think
6 makes statistical significance. That's point number
7 one.
8 Number two, the a fib flutter subgroup was
9 never p specified as a subgroup in this study. No. 3,
10 that in this group if you look through the study there
11 were people who had arrhythmias classified as either
12 a fib and a flutter, a fib and atrial tachycardia.
13 Number three, one could have dissected the
14 group to decide that the a fib group included those
15 people, didn't include those people, or include some
16 of those people. If you look at the numbers, and if
17 you have one placebo patient or two placebo patients
18 who didn't have recurrences, the P value would have
19 probably gone away.
20 CHAIRMAN PACKER: Maybe we should ask a
21 question. In general we think that looking at subgroups
22 of studies are interpretable if the overall study was
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
221
1 positive. Was the overall study -- we don't know the
2 results of the overall study.
3 DR. KOWEY: We showed you the intention of
4 treatment on slide 53 for all patients. This is for
5 all patients.
6 DR. CAIN: Ed, can I ask one follow-up to
7 your question -- to your comment and that was if there
8 is a doubling and tripling in this population group,
9 the confidence that that doubling and tripling would
10 be applicable to other patient groups, how far can one
11 stretch that as opposed to the data simply being specific
12 for this particular population.
13 DR. PRITCHETT: I think that how
14 generalizable the data are depends on how generalizable
15 the entry criteria are. One of the things that made
16 this group so active was the screening that was done
17 when they set up that screening period to say put
18 patients in the one-week bin or the two-week bin which
19 developed a very enriched population of patients.
20 I believe that those numbers would likely
21 hold up if the entry criteria were related sort of to
22 the type of patient, not just -- in other words, I don't
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
222
1 think you have to screen your patients and see whether
2 they have one episode per week or three weeks in a row
3 to obtain this benefit.
4 DR. KOWEY: Michael, do you want to see the
5 inclusion/exclusion criteria?
6 DR. CAIN: It's okay.
7 DR. THADANI: On that issue of paroxysmal
8 a fib which is happening so often, I personally would
9 have liked to have seen the frequency of episode rather
10 than just -- I realize the study was designed for the
11 first episode, but if a patient is getting 10 episodes
12 on a 24-hour Holter or 20 episodes and then that
13 information is available, that would be very useful to
14 realize that not only you're reducing the onset of the
15 first episode but are you reducing the number of
16 episodes. These patients are obviously bothered with
17 recurring symptoms. Any data they have?
18 DR. PRITCHETT: Well, No. That is very
19 closely related to the question that Michael Cain
20 answered. What we know is that --
21 DR. THADANI: In this population was there
22 any data?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
223
1 DR. PRITCHETT: Well, it wasn't done in
2 this study. What we know in general about patients who
3 have recurring arrhythmias are that when you put them
4 on observation and measure time to first and then measure
5 time between the first and the second that that number
6 is the same. And from the previous clinical trials
7 program done with Flecainide presented to this committee
8 in October of 1989 and published in Circulation by Geoff
9 Anderson in 1991, we know if you follow people to the
10 fourth event and then look at the average time interval
11 between events compared with the median time interval,
12 that those numbers are nearly the same.
13 There is a substantial empiric body of data that
14 tells you that measuring treatment effects by looking
15 at time to first event is a good way to estimate long-term
16 effects.
17 The best study done to try and follow
18 patients for multiple events was the bidisomide study
19 conducted by Ciro which was published in Circulation
20 in 1995 which recruited 1,200 patients with atrial
21 fibrillation and 200 with PSVT, and we tried to follow
22 patients for a full year no matter how many events we
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
224
1 had.
2 What we found is you could keep them in the
3 trial for a couple of events before they demanded to
4 be taken out if the drug wasn't working. While it's
5 a nice idea to say let's follow patients until they have
6 eight events or let's capture every event over the course
7 of a year, in practical terms that's very, very difficult
8 to do. No one has been able to do it successfully.
9 CHAIRMAN PACKER: Okay. If there are no
10 other questions on 9A, let's move on. I guess the only
11 study remaining is study H. This is the open label
12 comparison of quinidine and sotalol. Any questions?
13 Okay. Are there any questions at all on any other
14 issues related to efficacy? Are we in a post prandial
15 lull? Let's proceed to safety, please.
16 DR. GRABOYS: There's three groups of
17 patients. This may be extraditable to I think all of
18 the studies but there are three groups of patients here
19 that are missing in terms of being able to make a decision
20 about risk benefit.
21 Two have already been alluded to. One is
22 the octogenarian population. We are increasingly
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
225
1 seeing people in their 80's who are presenting with
2 atrial fibrillation. We question then in that
3 population what kind of data do you have to support
4 safety in that population.
5 The second is women, particularly in view
6 of the fact that QT prolongation seemed to occur more
7 commonly in women with sotalol, and the fact that has
8 already been alluded to, women seem to be
9 under-represented in the data.
10 The third is the African-American and Black
11 population in which you see that 99 to 100 percent or
12 even 90 percent in some of the studies are all white.
13 How do we then interpret that in terms of, again, safety
14 utilization in the Black population.
15 DR. KOWEY: Tom, I'm going to address, in
16 the safety presentation, I'll talk about age and gender.
17 I do have some other backup slides that I'll show you
18 on age and gender if you want to see them.
19 Unfortunately, many of these studies were done in
20 Scandinavia.
21 A couple of these studies, as you've heard
22 already, were done in countries where there are no
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
226
1 African-Americans, and so the database, as you point
2 out correctly, does not contain them. It is something
3 I'm sure that the sponsor would consider doing in a
4 post-marketing effort.
5 CHAIRMAN PACKER: Okay. There being no
6 further questions about efficacy, let's move on to the
7 presentation. Rob, yes?
8 DR. CALIFF: Just one question. This is
9 a mixture of studies not all of which were done
10 specifically for marketing for this indication. Out
11 of the universal study looking at sotalol for atrial
12 dysrhythmias, is this 100 percent of those studies?
13 DR. KOWEY: The only studies that we have
14 in the database that we did present to you for efficacy
15 where G, which was a subpopulation study of MSPT cohort
16 and then the two Stige studies which were really not
17 done specifically to look at these studies. Stige II
18 sort of was but it was stopped very early and there was
19 no useful data.
20 DR. THADANI: There are no other studies
21 which have been done, have negative results not
22 published or not presented or not shown here?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
227
1 DR. KOWEY: I'm sure there are studies that
2 have been kind of not published but not by the sponsors.
3 DR. THADANI: There is no end date on those
4 studies?
5 DR. KOWEY: No.
6 DR. THADANI: All the end dates have been--
7 DR. KOWEY: There are compassionate use
8 studies but they are not --
9 CHAIRMAN PACKER: Any ongoing trials at the
10 present time?
11 DR. KOWEY: No.
12 CHAIRMAN PACKER: Let's move to safety.
13 DR. KOWEY: I asked Milton and he agreed
14 so I'm going to hold him to it. I only have few slides
15 on dosing recommendations, so I'm going to cover these
16 last two topics. Neither of these areas are nearly as
17 long as the efficacy discussion so we should be able
18 to get through it fairly quickly.
19 This is the composite of the clinical
20 information we are going to use for the safety
21 presentation, 2,184 patients. I'll be pointing out
22 that there are four studies that are in an electronically
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
228
1 pulled database that you've already heard about this
2 morning, 05, 004, 014, and 9A.
3 We have safety information for the most
4 serious adverse events in the unpulled and I'm not going
5 to be discussing specifically issues in the
6 compassionate use studies, although I will present you
7 a bit of information about the most serious adverse
8 events in the total database.
9 I do want to point out, Milton, that we are
10 going to show you data for deaths and torsade only from
11 the 345 study of dofetilide. This will be grouped
12 basically into three parts of this presentation. The
13 first part is to look at the most common adverse events.
14 This will be, as I said earlier, from the pulled
15 database.
16 I have a discussion of clinically important
17 adverse events from a larger database. Then I'm going
18 to present supportive studies showing no access for
19 structural heart disease. One is the post-MI Julian
20 study which the agency felt very strongly that we should
21 show you today, an old sotalol study. Then the ICD study
22 that was recently completed and the results are
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
229
1 currently in review.
2 Again, I want to make sure that everyone
3 understands that we will be using different denominators
4 for the safety discussion. For the most common AE's
5 we'll be using a pulled placebo controlled database.
6 We will be showing you data from the double-blind phase
7 of these four pulled trials. So we will have 415
8 patients in the sotalol arm and 282 patients in the
9 placebo arm for the most common AE's.
10 For heart failure, stroke, and myocardial
11 infarction, we have data from the controlled phase of
12 eight controlled studies. That's the four pulled and
13 the four older studies for an end of 656 and a placebo
14 group of 358. We've added in for death and torsade data
15 from the dofetilide of 137 patients in placebo and in
16 the sotalol arm. That's where these numbers come from.
17 This is the most common adverse events in
18 the clinical trials. This is looking at the pulled
19 placebo controlled trials. And I think the numbers
20 speak for themselves. You probably would expect for
21 a beta-blocker to see fatigue and dizziness,
22 bradycardia, dyspnea, and palpitations as adverse
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
230
1 events in the d,l-sotalol arm at a higher frequency than
2 in the placebo arm. These are the discontinuation
3 rates; in the d,l-sotalol arm 17 percent compared to
4 five percent in the placebo arm.
5 I'm now moving into more serious adverse
6 events. You are going to see this kind of format on
7 several of these slides that will be coming up. We have
8 put the studies on the left-hand side that generate the
9 information. The d,l-sotalol placebo, a comparator if
10 one is in the study or one is included in these particular
11 studies. That would be d-sotalol or quinidine.
12 And this is for deaths in the controlled
13 phase of the eight controlled studies; the dofetilide
14 trial, 245. Again, we are adding these numbers. This
15 is the percentage of patients who died in the program,
16 0.5 percent compared to 0.4 percent in the placebo arm.
17 These are the numbers for d-sotalol and quinidine.
18 These are the patients who died in these
19 trials. In the d,l-sotalol arm three of the deaths were
20 in study 014. These three patients received doses of
21 the drug which were in excess of the dose that we are
22 recommending for this particular patient. You notice
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
231
1 that there are no deaths in 05 or 04.
2 There was one death, as I already mentioned
3 to you, in H which was a myocardial infarction in a
4 patient who got 160 milligrams per day. The cause of
5 death, like I said, was myocardial infarction. In the
6 placebo arms, in 004 there were two deaths, as I have
7 already mentioned. Both patients have structural heart
8 disease.
9 This is torsade in the controlled phase of
10 the eight controlled trials and dofetilide trial 245.
11 The total is four for 0.5 percent, placebo arm 0.2
12 percent, none in the d-sotalol, and one in the quinidine
13 arm. I want to emphasize that in this entire data set
14 all these patients who had suffered torsade, there were
15 no deaths.
16 These are the torsade cases themselves.
17 This is, again, the controlled phase of the controlled
18 trials. Study 014 was where three of the sotalol
19 related torsade events occurred. These are the number
20 of days that the patient had been on the dose that led
21 to the torsade.
22 I want to point out this little cross here.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
232
1 That patient required a cardiovert. These three
2 patients had self-terminating torsade.
3 In this group of controls, again, I
4 mentioned earlier that there was one patient we studied
5 each who developed torsade and that person required a
6 cardioversion. The patient in 004 had received a
7 placebo and developed torsade, actually had torsade,
8 or taken the marketed anti-arrhythmic drug after
9 stopping placebo. So this is a torsade experience in
10 controlled clinical trials.
11 I want to show you an analysis of the back
12 end torsade in the controlled phase of the eight
13 controlled trials and a dofetilide study segregated on
14 the basis of dose in the study. This is within the
15 recommended dose for this indication of 320 milligrams
16 per day. This one, this is in excess of 320 milligram
17 dose. There were 62 patients in the controlled phases
18 of controlled trials who received a dose greater than
19 320 milligrams.
20 There were 734 patients who received the
21 dose that we are recommending. This is death. There
22 was 4.8 percent in this group, 0.1 percent in the less
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
233
1 than 320 milligram group. In the placebo the death rate
2 was 0.4 percent. This P value is Fisher's Exact test
3 comparing these two columns. Not comparing placebo,
4 comparing this column with this column.
5 This is the torsade rate, 3.2 percent for
6 patients who received greater than 320 milligrams.
7 There was 0.3 percent in patients receiving less than
8 320 milligrams. This is the placebo rate which was a
9 little tiny bit less. This is the P value with the
10 difference between that group and that group. This is
11 important because of recommendation of dose for this
12 indication.
13 This is heart failure in the controlled
14 phase of the eight controlled trials. Here we do not
15 have a dofetilide information so at the end the smaller
16 656 patients, 1.5 percent in the d,l-sotalol group.
17 You can see it's 0.8 percent in the placebo group, 0.6
18 percent for d-sotalol, 1.3 percent for quinidine.
19 This is stroke in the controlled phase of
20 eight controlled studies. d,l-Sotalol, 0.9 percent;
21 placebo, 0.6 percent; d-sotalol, 0.5 percent;
22 quinidine, 2.3 percent.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
234
1 This is myocardial infarctions in the
2 controlled phase of the eight controlled studies; 0.5
3 percent for d,l-sotalol, 0.6 percent for placebo, 0.6
4 percent for d-sotalol, and there were no cases in the
5 quinidine group.
6 This is a slide now from the entire safety
7 database, not just from the eight controlled studies,
8 showing the overall incidents of death, torsade, heart
9 failure, stroke, and myocardial infarction in the entire
10 database including the compassionate use studies in the
11 open label experiments following the controlled phase
12 and these are the percentages. Remember
13 that for death and torsade, the data includes patients
14 in the dofetilide trial.
15 I would like to point out that there were
16 28 cases of torsade in the entire database. Of those
17 28 patients there were two deaths and torsade.
18 I very briefly want to run through the two
19 supportive studies for no excess mortality and
20 structural heart disease. Again, this was an agency
21 request and this one is a study which has just recently
22 been completed and I'll do this briefly.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
235
1 The Julian study was a post-myocardial
2 infarction study very much in the genre in the early
3 1980's of beta-blocker trials after myocardial
4 infarction. The goal of this study was to evaluate the
5 efficacy of sotalol in reducing all-cause mortality
6 reinfarction following acute myocardial infarction.
7 There were 456 patients, the usual age group enrolled
8 five to 14 days after MI. The primary endpoint was
9 all-cause mortality and reinfarction.
10 We are only going to show you data for
11 all-cause mortality there's been some question about
12 the way that the reinfarctions were quantitated. We
13 have those data if you want to see it. A statistical
14 test and it was a one-year study which was published
15 in Lancer.
16 Persons, as I said, who had a recent
17 myocardial infarction were randomized to a very, very
18 novel dose of sotalol, 320 milligrams delivered as a
19 single dose in the morning compared to placebo,
20 double-blind treatment for 12 months. For those
21 patients who could not tolerate 320 milligrams a day,
22 they could have their dose reduced to 160 milligrams
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
236
1 per day but in the protocol specifically indicated that
2 that was for patients with bradycardia less than 50 beats
3 per minute.
4 This is patient characteristics. There
5 were mostly males in the study. This is the history
6 of coronary arteries. There is hypertension prior to
7 the acute event. I want to point out that during the
8 infarction a substantial number of patients had had
9 heart failure, an increased cardiothoracic ratio,
10 relative hypotension. There's a relatively even split
11 at both ends of the study as to anterior, interior
12 infarct.
13 This is the Kaplan Meier. Well, this is
14 the cumulative mortality total for the study for
15 d,l-sotalol and for placebo. There has been some
16 discussion about the early mortality that was seen in
17 the d,l-sotalol group. I want to point out that at no
18 time during the first 10 to 30 days of this study was
19 there a statistically significant difference between
20 the two groups in terms of mortality and there are many
21 explanations that you can discuss as to why that may
22 have happened.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
237
1 But in any case, the study showed that there
2 was not only no excess mortality with d,l-sotalol for
3 a one-year time frame, but in fact there was a reduction
4 in mortality although it by no means met any kind of
5 statistical significance.
6 The second study, the so-called ICD study,
7 was a test of the hypothesis that d,l-sotalol would be
8 effective in place of the placebo preventing all-cause
9 ICD shocks. As many of you know, patients who have
10 defibrillators are prone to frequent device disrupt
11 causing a significant amount of morbidity and, in some
12 cases, mortality.
13 Population of patients with 202 patients
14 with ICD's who were implanted for the indication of
15 life-threatening ventricular arrhythmias. I'll show
16 you a breakdown of that in a moment. The primary
17 prespecified endpoint was time to first all-cause ICD
18 shock with that after randomization. And it was a
19 Kaplan Meier survival curve with a log rank test.
20 As I said, these are patients with
21 life-threatening ventricular arrhythmias. The
22 randomization was stratified for ejection fraction.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
238
1 Patients were distributed by ejection fraction greater
2 than and less than 30 percent. In the presence of renal
3 insufficiency in this study, the find is a creatinine
4 clearance between 30 and 60 cc's per minute. Patients
5 received the once daily dose similar to what was done
6 in 05. Patients less than 30 cc's per minute were
7 excluded. The therapy was continued double-blind for
8 12 months.
9
10 And in the sotalol arm of the study -- well, in both
11 arms of the study there was the opportunity for changing
12 dose which was done blindly.
13 This was the inclusion criteria. These
14 were patients who were undergoing first implantation
15 or placement of an ICD within three months of enrollment.
16 For those patients who had had replacement, it was
17 necessary for them to have had at least one shock during
18 the preceding six month period in order to be certain
19 that these were not simply quiescent patients.
20 Tiered therapy ICDs were used in all cases
21 and all devices had electrogram storage and logging of
22 shock and other types of cardio pacing episodes for us
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
239
1 to be able to retrieve the information and make a
2 judgment as to whether or not the shocks were
3 appropriate.
4 I would also point out that predistress
5 testing was carried out in this cohort to guarantee that
6 shock energies were at least 10 joules below the maximum
7 ICD ouput. This is fairly standard clinical practice
8 for ICD input.
9 These are the baseline characteristics of
10 the patients in the placebo arm and in the sotalol arm,
11 even number of patients. This is male being two
12 percent, age matched. I would point out that a number
13 of these patients had undergone coronary interventions.
14 A substantial number of these patients had previous
15 myocardial infarctions. There was a small percentage
16 of patients who had Class II New York Heart Association.
17 I would also point out that as a typical
18 ICD patient population, a third of the patients that
19 had aborted sudden death, two-thirds of the patients
20 had ventricular tachycardia either symptomatic and/or
21 inducible in the physiology laboratory.
22 This is the primary three-step spot
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
240
1 analysis for the ICD study upon the first all-cause shock
2 or death internalization with the intention to treat
3 analysis. This is the true value for the observation
4 and difference between sotalol and placebo.
5 If one relied on the investigator's
6 interpretation of the electrograms to determine whether
7 or not the ICD shock was appropriate for VT/VF -- that's
8 the first shock was appropriate for VT/VF. These are
9 the data. The P value is 0.007.
10 Finally, this is all-cause mortality and
11 intention to treat analysis or placebo, 4.6 percent,
12 and for d,l-sotalol, 2.6 percent. There were no sudden
13 deaths from this clinical trial.
14 This is just to make you aware of the fact
15 that when we look at the study based on the
16 stratification of less than and greater than 30 percent
17 ejection fraction, there was consistency of the results
18 across those two strata.
19 I will conclude from this entire safety
20 discussion that doses between 80 and 160 milligrams
21 twice per day are safe. In fact, in study 05 and study
22 04, where they were the doses used, there were no deaths
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
241
1 and there were no extensive torsade. Discontinuation
2 due to adverse events, as we've seen before lunch, is
3 dose related, but when titration is permitted, it
4 maximizes the benefit to this ratio.
5 The incidents of death, torsade, and other
6 serious AEs in the entire database as well. It appears,
7 therefore, justified that in patients who have
8 structural heart disease, outpatient therapy may be
9 safely undertaken.
10 I just want to make, Milton, if I may, just
11 the three or four slides on the characteristics of dosing
12 recommendations because this does have to do a good deal
13 with safety. This gets to a question that we discussed
14 earlier, and that is the electrophysiologic and
15 pharmacodynamic effect of this drug, vis-a-vis its
16 affect on electrophysiologic parameters and then on
17 efficacy. These are data from the
18 randomized dose ranging study 05 looking at heart rate
19 QT and QTc. Heart rate is illustrated on the slide in
20 pink and you can see that with increasing dose of the
21 drug, there is a progressive fallen heart rate. But
22 there is also a more profound increase in the continued
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
242
1 interval if one gets the 160 milligram dose, which,
2 again, in these patients for the most part was taken
3 twice per day and with a renal impairment was taken once
4 a day. If you examine the information regulating QTc
5 to the dose response in study 05, the pink bars are
6 change. This is delta now PTC in
7 milliseconds. Our presumed study states from baseline
8 and QTc. And as you would expect, as you increase the
9 dose you increase the effect on QTc. The green is a
10 Kaplan Meier estimate of relapse-free intervals, a
11 relapse-free rate at 12 months showing a progressive
12 effect by dose.
13 Based on all this information, we would make
14 the following recommendations about dose. First, as
15 in the clinical trials, it is extremely important that
16 careful attention be paid to identify and correct the
17 risk factors for coarrythmics effects of sotalol which
18 include hypokalemia, tachycardia, and QT prolongation,
19 either a congenial or acquired on the basis of use, for
20 example, of other drugs which come on in the interval
21 which are well described.
22 Sotalol may be initiated on an outpatient
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
243
1 basis, as I said, about structural heart disease, but
2 doses greater than 160 milligrams twice a day, or once
3 a day in patients with renal dysfunction, are not
4 recommended. The titration is an extremely important
5 part of using sotalol. It's the way we use it in clinical
6 practice. It's the way it was done in several of the
7 clinical trials.
8 And we think that the data adequately
9 supports these recommendations. Treatment should be
10 initiated with 80 milligrams twice per day. We have
11 data from study 345 and study 05 which provides, we
12 think, comprehensive evidence that the drug works at
13 that dose and has a good safety profile. Remember, in
14 study 345 there were no deaths and no torsade, and
15 neither were there in study 05.
16 Study 05 provides evidence of efficacy and
17 safety for 120 milligrams twice per day, which should
18 be the second step in the titration process. Many
19 physicians routinely go to 120 milligrams twice per day
20 even if the patient has had no recurrences with 80
21 milligrams twice per day on the principal that this may
22 be the most effective dose.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
244
1 And then, finally, for patients who do not
2 respond to 120 milligrams dose, study 004 suggest that
3 d,l-sotalol at 160 milligrams twice per day is effective
4 and safe.
5 That concludes my comments.
6 CHAIRMAN PACKER: Peter, could you just go
7 back one moment.
8 DR. KOWEY: Can we go back?
9 CHAIRMAN PACKER: Yes. Statement No. 1,
10 you say study 05 provides evidence for efficacy and
11 safety at this dose?
12 DR. KOWEY: I would say that there is a
13 better effect but not statistically significantly
14 better effect for 80 milligrams in 05.
15 CHAIRMAN PACKER: First of all, I don't
16 think there was anything that one could talk about at
17 80 milligrams versus placebo in study 05.
18 DR. KOWEY: Can we go back at least two
19 slides? Okay, I agree. The reason I said it is because
20 there are patients who will respond to an 80 milligrams
21 twice per day dose. We don't know how to preselect those
22 patients necessarily, but I think, as I said in clinical
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
245
1 practice, we usually start at 80 milligrams twice per
2 day.
3 CHAIRMAN PACKER: I'm sure 80 milligrams
4 twice per day is effective with people of creatinines
5 of 4. I'm joking.
6 DR. KOWEY: Or for people with small body
7 size, as Bob was talking about earlier.
8 CHAIRMAN PACKER: The issue here is not to
9 question the recommendation of where to start but to
10 question your conclusion that that starting dose is
11 effective or has been shown to be effective instead of
12 the 05.
13 DR. KOWEY: Okay. I will concede, Milton,
14 that most of the efficacy data for 80 milligrams has
15 to come from 345. But we do have safety data from 05
16 and that was a compound sentence that said safety and
17 efficacy. So, maybe I can hide behind that.
18 DR. THADANI: Also, I think you can't say
19 some of the patients respond; so did the placebo
20 patients, 28 percent, so that's a nonstatement.
21 DR. KOWEY: Well, again, it depends on how
22 much confidence you place in study 345.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
246
1 DR. THADANI: I realize that but there's
2 only 2 percent difference. I think Milton's point is
3 very valid.
4 DR. KOWEY: I have to concede that was
5 overstated.
6 CHAIRMAN PACKER: Why don't we begin with
7 JoAnn in the conventional way.
8 JoAnn, questions about safety?
9 DR. LINDENFELD: Just to start off, could
10 you show us a list of the drugs that were excluded in
11 these studies? I guess the question is were they the
12 same. I know anti-arrythmics were excluded. And that
13 ties in with verapamil. It just said in the protocol
14 there was a list of excluded drugs. Just for the
15 purposes of how we use these drugs, do you know what
16 that included? Erythromycin, bactrim?
17 DR. KOWEY: Yes. The investigators were
18 instructed in the protocol to exclude the use of any
19 drug that prolonged the QT.
20 DR. LINDENFELD: Could you just show us a
21 list of that so we --
22 DR. KOWEY: We don't have a list of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
247
1 actual slides.
2 DR. LINDENFELD: So any drug that might
3 prolong QT but there was the specific --
4 DR. KOWEY: That was specifically an
5 exclusion criteria.
6 DR. LINDENFELD: And approximately how
7 many were there on there, 30, 20? Was there a large
8 number?
9 DR. KOWEY: I'm sorry?
10 DR. LINDENFELD: There were a fairly large
11 number on that list?
12 DR. KOWEY: Of drugs? Oh, yes.
13 DR. LINDENFELD: That the investigators
14 looked at.
15 DR. KOWEY: It's a big list.
16 DR. LINDENFELD: Again, that doesn't take
17 away from the efficacy but there's a large group of drugs
18 that these older patients might be taking that interfere
19 here.
20 DR. KOWEY: Absolutely true.
21 DR. LINDENFELD: Just to be sure, I
22 understand that even in the 004 study that dotiazam and
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
248
1 verapamil were required to be withdrawn prior to the
2 use of sotalol so that those drugs were specifically
3 excluded because that gets back to the risk of
4 bradycardia and there is a fairly substantial risk of
5 bradycardia. What I wanted to ask was, do we know was
6 that primarily following conversion to sinus rhythm,
7 the bradycardia?
8 DR. KOWEY: Yes.
9 DR. LINDENFELD: Okay. Many of those were
10 classified as serious adverse effects. Is that
11 correct?
12 DR. KOWEY: Which? I'm sorry, JoAnn.
13 DR. LINDENFELD: The bradycardia.
14 DR. KOWEY: Yes.
15 DR. LINDENFELD: A number was quite
16 serious.
17 DR. KOWEY: A number of them were
18 classified as serious. Yes.
19 DR. LINDENFELD: And in terms of adverse
20 effects --
21 DR. KOWEY: Do you want to see the
22 percentages?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
249
1 DR. LINDENFELD: That would be great.
2 DR. KOWEY: Can we have the common adverse
3 events greater than 6 percent slide in the core?
4 DR. LINDENFELD: I just want to emphasize
5 here that the safety things will include not only the
6 QT prolonging drugs but drugs that may cause
7 bradycardia.
8 DR. KOWEY: Here it is. Actually, you
9 know, I take it back a bit. This is greater than 6
10 percent incidence bradycardia. I take back what I just
11 said. It's actually a small percentage that were
12 considered severe.
13 DR. LINDENFELD: And then can you give us
14 some information about specific subgroups, the ones we
15 know are high risk for torsade?
16 DR. KOWEY: Yes.
17 DR. LINDENFELD: Heart failure, female
18 gender, age greater than 70. Specifically, I wondered
19 LVH wasn't specifically included but let's get back to
20 the group Tom talked about, the African-Americans with
21 LVH who might be considered high risk. Do you have some
22 specifics for those groups?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
250
1 DR. KOWEY: I don't have LVH but I can show
2 you a lot of the other stuff.
3 DR. LINDENFELD: Was LVH excluded? I
4 didn't see that as one of the --
5 DR. KOWEY: No. It was not excluded.
6 There were a smattering of about 20 or 30 percent of
7 the patients in the trials that had hypertension so LVH
8 was not an exclusion.
9 Can I have backup slide, please, No. 361.
10 This is population less than and greater than 60 in
11 the controlled phase of the controlled trials. These
12 are the deaths in the greater than 65 and less than 65
13 group. Heart failure, stroke, torsade. There was more
14 dizziness in the older patients and there was more
15 bradycardia in the older patients.
16 DR. GRABOYS: Peter, those over age 65,
17 what was the average age? We need to know about the
18 older population.
19 DR. KOWEY: I'll tell you, Tom, I don't have
20 the data but I can tell you that there was not a large
21 number of very, very old in the study. The
22 octogenarians that you were talking about, I'm afraid
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
251
1 there really weren't very many in the trial. I don't
2 know exactly what the number was.
3 DR. LINDENFELD: Do you have torsade by
4 gender of creatinine clearance?
5 DR. KOWEY: Yes. Which way do you want it?
6 You want gender first or you want creatinine clearance
7 first?
8 DR. LINDENFELD: Either one first.
9 DR. KOWEY: Okay. How about if we do
10 gender. And do you want torsade? This is torsade by
11 gender and treatment in a controlled phase of the
12 controlled class. This is female, male. This is
13 comparative data with quinidine and there's some
14 d-sotalol, a very small number of patients. What was
15 the other one, JoAnn, heart failure?
16 DR. LINDENFELD: Creatinine clearance.
17 DR. KOWEY: Creatinine clearance. We can
18 do that. Can I have 367, please. This is creatinine
19 clearances greater than 60, less than 60. This is
20 torsade, deaths, heart failure, stroke. There was more
21 bradycardia in the patients who had well preserved
22 creatinine clearances and there was more dizziness in
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
252
1 patients with low creatinine clearances.
2 DR. FENICHEL: You know, I don't see how
3 this is interpretable really because suppose that the
4 people in the trials generally all had creatinine
5 clearances that were in the range of, say, 55 to 65.
6 Well, then what you'd find is the people -- and the
7 threshold, I think, was to cut the dose at 60.
8 Well, then all of a sudden the people who
9 all essentially had the same creatinine clearance, the
10 ones with the slightly lower creatinine clearance were
11 actually getting a much lower dose so it looked like
12 it was much safer in that group. I think this is
13 hopelessly confounded.
14 CHAIRMAN PACKER: I think it's also
15 hopelessly confounded by the fact that none of the ADCR
16 people corrected. Consequently, elderly people could
17 get more side effects on placebo than younger people.
18 That wouldn't be too surprising. So that I think that
19 in order to really interpret this one has to adjust for
20 the corresponding incidence in the specific subgroups
21 in the placebo group.
22 DR. KOWEY: Can I have slide 296. This is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
253
1 the placebo group, Milton, and this is the d-sotalol
2 group. This is the breakdown for creatinine clearance.
3 I can do this basically for whatever parameter you would
4 like.
5 CHAIRMAN PACKER: I think this actually,
6 sort of, makes a point. Look at dizziness. If you
7 compared only the sotalol group, you might expect that
8 there was not a lot of difference in dizziness above
9 and below 60. If you compare it to the corresponding
10 placebo group, which is less than 60 on sotalol, less
11 than 60 on placebo, there is a substantial difference
12 in risk of dizziness which is not present if you do a
13 placebo correction on the group with more normal renal
14 function, something which one would never have picked
15 up if one only did a comparison above and below 60 in
16 the sotalol group.
17 So if you--One, I think one should always
18 do a placebo correction, and second is that this would
19 indicate that dizziness is an issue in the group. Much
20 more of an issue than not in a group with more borderline,
21 you know, functioning. The group with normal renal
22 function, literally there is no increase in dizziness.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
254
1 DR. KOWEY: Everything you said I
2 completely agree with. That's why we analyze the data
3 for common adverse events with placebo in the controlled
4 phase of the controlled trials because you had placebo
5 information.
6 CHAIRMAN PACKER: Bradycardia is not
7 differentially distributed.
8 DR. KOWEY: That's what I was just going
9 to say. So you can that with the placebo correction,
10 it's really not an issue. I can do that if you would
11 like. I have data for other subgoups, but I think what
12 you'll see is that it comes out as a wash in many of
13 these studies.
14 CHAIRMAN PACKER: Does anyone want a
15 specific subgroup not--that they haven't seen for this
16 kind of placebo corrected data. I think that we would
17 like to spare all of use having to see every single
18 permutation and combination of these.
19 DR. KOWEY: Did you want to see gender?
20 DR. THADANI: You might show gender in QTc.
21 I thought my reading, if I remember correctly, QTc was
22 more prolonged in men than women. Am I correct?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
255
1 DR. KOWEY: I can show you data on that on
2 slide --
3 DR. THADANI: Am I reading it wrong? I
4 thought--
5 DR. KOWEY: I'll show you slide 320.
6 DR. THADANI: Torsade is the opposite
7 around. No, I realize this doesn't go together.
8 DR. KOWEY: These are data male versus
9 female, male in yellow and female in orange. This is
10 change in heart rate. You can conclude anything. I
11 mean, you can look at these and decide what you think.
12 This is QTc data and this is QT uncorrected for 80,
13 120, 160 milligram dose groups in study 05.
14 DR. THADANI: There's less prolongation in
15 woman of QT and QTc, which--You know, normally we talk
16 about incident is torsade greater in woman. My
17 indication would have been that QTc is more prolonged
18 in those groups.
19 DR. KOWEY: For reasons that are not clear
20 to me -- you are right, by the way, that torsade is more
21 frequent in women. You are also right that you tend
22 to get more QT prolonging effect in women. We didn't
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
256
1 see either of those in these data. I suspect part of
2 the reason is because we were at the low end of the dose.
3 If we had gone higher on the dose, we may have seen
4 that effect.
5 DR. CALIFF: And, I might add, at the low
6 end of the number of patients.
7 DR. KOWEY: Yes. Also true.
8 DR. THADANI: Also probably the
9 beta-blocker has different effect. Doesn't it?
10 DR. KOWEY: Yes.
11 DR. THADANI: Where you find the QT because
12 other drugs don't have beta-blockade.
13 DR. KOWEY: But these are the bases that
14 we want to use, and that's the data on the QT.
15 CHAIRMAN PACKER: Michael and Ileana
16 after.
17 DR. CAIN: Just one methodologic question.
18 The QT measurements that are reflected in these data
19 were made off the 12 week ECG?
20 DR. KOWEY: Yes.
21 DR. CAIN: Because someone earlier had said
22 you were also measuring them off of the transtelephonic.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
257
1 DR. KOWEY: The transtelephonic was used
2 as an indicator that the patient had to come in for a
3 12 week so that these data we're seeing here are from
4 12 weeks.
5 DR. PIÑA: I'm having a bit of a time
6 figuring out why the ICD trial was shown under the safety
7 considerations. Since it was and there were patients
8 entered who had ejection fractions of below 30 percent,
9 were they dosed differently? Were they dosed
10 in-hospital versus outpatient?
11 DR. KOWEY: The majority of the patients
12 in this study were in the hospital. The majority, I
13 don't know what the percentage was. All of them--
14 that's a good majority--were all in the hospital for
15 initiation of the drug, but there was no dose adjustment
16 by ejection fraction. There was a dose adjustment by
17 creatinine clearance but not by ejection fraction.
18 DR. THADANI: But that's a VT population.
19 Right? ICD.
20 DR. KOWEY: You know, to explain to Ileana,
21 her first question was, "Why did you see that? Because
22 you really like it?" I think, first of all, it's the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
258
1 first study that has ever been shown. I mean, it's a
2 little bit extraneous but it's the first study that shows
3 that there is a benefit to using the drug as adjuvant
4 therapy in ICD patients.
5 But, in addition, the question that a lot
6 of you have been asking is if you give this drug to
7 patients who don't have good ventricular function, does
8 it have an adverse effect? Well, the FDA wanted us to
9 show you the Julian data, which we did, and this is
10 another group of patients who have bad ventricles.
11 That's the reason why the data was shown. That's the
12 only reason why the data was shown.
13 DR. THADANI: What were the incidents of
14 torsade in that ICD group? Because here you showed that
15 sotalol far less often so there's no incident of torsade.
16 DR. KOWEY: If you can tell me an ICD
17 patient from an electrogram whether something is torsade
18 or polymorphic--
19 DR. THADANI: We can't tell.
20 DR. KOWEY: --then you have to really tell
21 me how to do that.
22 DR. THADANI: How many were polymorphic
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
259
1 versus monomorphic VT?
2 DR. KOWEY: I don't have a breakdown. I
3 do not have a breakdown. Do we have a breakdown of poly
4 versus mono VT?
5 DR. WILLIAMS: There was one patient that
6 had torsade diagnosed from a Holter. We had in the
7 protocol a requirement for a Holter at one month.
8 Protocol required a Holter recording at one month, and
9 one patient had a polymorphic VT documented on that
10 Holter QT prolongation which was called torsade. The
11 patient was taken out of the study and was on placebo.
12 We had another patient who from the electrogram --
13 DR. KOWEY: While you are talking, John,
14 can I have slide 357, please?
15 DR. WILLIAMS: The electrogram suggested
16 torsade but I'm not sure if you can diagnose torsade
17 from an electrogram of an ICD.
18 DR. KOWEY: One of those patients was a
19 placebo patient and one of these patients was a
20 d,l-sotalol patient, but to tell you the truth, I mean,
21 it's possible that some of these discharges could have
22 been for torsade.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
260
1 DR. THADANI: But you were able to
2 interview. That's the good thing about the ICD. You
3 can go back and integrate the --
4 DR. KOWEY: Yes. It was done.
5 DR. THADANI: That means the incidence of
6 torsade mostly is asymptomatic in patients, unless they
7 die, that might have important implications. Now you
8 are showing me that QTc in women doesn't change very
9 much. Why you won't admit these patients in the data
10 of cost containment for hospitalization realizing that
11 adverse effect could have happened up to 10 days of
12 therapy, not necessarily 24 or 48 hours. If you're
13 going to see it, even in structural heart disease if
14 you feel comfortable that's there's no torsade in that,
15 why you won't admit the patient at all.
16 DR. KOWEY: I hope I didn't misspeak.
17 DR. THADANI: One of your slides said
18 nonstructural heart disease patients can be started
19 outpatient.
20 DR. KOWEY: Nonstructural.
21 DR. THADANI: Why nonstructural heart
22 disease? With the incidents so low why are you
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
261
1 recommending that? Why can't you make a bold stab that
2 you're not going to hospitalize anybody and give a drug
3 and monitor them just on ICD.
4 DR. KOWEY: There was torsade and there
5 were deaths in patients who had structural heart disease
6 in this database. So we're not horribly comfortable
7 with saying that you can start the drug out of hospital.
8 In 05 and 04 there were no events in patients that
9 received the low end of the dose. That's the basis for
10 that recommendation, but I think that's an arguable
11 point. We can argue that as a clinician but I'm not
12 sure that the data would support either way.
13 DR. THADANI: You're suggesting I start a
14 patient on 80, send them home and bring them, and back
15 before I do 160 or 120 rehospitalize them?
16 DR. KOWEY: In my practice patients that
17 have structural heart disease, they are rehospitalized
18 for every step in the titration. I know that sounds
19 difficult and it is always difficult for the patient
20 but it's the way I practice. Yes, the answer is yes.
21 CHAIRMAN PACKER: Dr. Karkowsky.
22 DR. KARKOWSKY: A quick point. I expected
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
262
1 Bob to offer the usual agency disclaimer. We haven't
2 reviewed the ICD study neither for efficacy or safety.
3 So to the extent that's pivotal and requires an FDA
4 review, that needs to be deferred.
5 CHAIRMAN PACKER: Thank you. Marv.
6 DR. KONSTAM: Yes. I'd like to comment a
7 little bit about the issue about the Julian study and
8 the ICD study. I want to make the point, and I feel
9 fairly strongly about this, that actually I'm not helped
10 in the least by these studies. I made the same point
11 with regard to the dofetilide data set in Diamond
12 studies.
13 The issue is you have a population that is
14 targeted for the approval, for the indication, and a
15 very different population for which you are making some
16 comment vis-a-vis survival. And you are doing it in
17 the context of the drug that is fairly complicated that
18 has antiarrhythmic effects, proarrhythmic effects, and
19 beta-blocker effects. And when you look at what's--
20
21 I think the ICD study is particularly
22 problematic. I won't even go there. I think with
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
263
1 regard to the Julian study, a post-MI population where
2 it looked the incidence of mortality was overall was
3 about 8 percent, far more than we expect in the
4 population that is going to be targeted by this
5 indication. And a population in whom we know very well
6 in hindsight that they are going to be benefitted by
7 the beta-blocker effect and a fair likelihood that they
8 are going to be benefitted by the antiarrhythmic effect
9 of the drug as well.
10 And in contrast, the population that is
11 being targeted by the indication being asked for here
12 where in the absence of, well, I think much less
13 likelihood, significantly less likelihood of
14 benefitting from the beta-blocker effect per se in terms
15 of survival, and certainly no rationale for likelihood
16 that they are going to benefit from the antiarrhythmic
17 effect of the drug again with regard to survival.
18 But, you know, a concern, and that's what
19 the concern is that I don't think is allayed by those
20 other trials, that there will be some small but
21 significant excess mortality perhaps in the 1 to 2
22 percent range from the proarrhythmic effect of this
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
264
1 drug. I think that the solace that has been taken, and
2 I think this was particularly relevant in the dofetilide
3 data set and equally relevant here with the Julian trial.
4 I actually take absolutely no solace from the findings
5 of those two studies that you showed at the end.
6 DR. KOWEY: Let me just make two comments.
7 No. 1, it was the agency that really wanted us to show
8 the Julian study but we didn't mind doing that because,
9 to be perfectly honest with you, Marv, as a clinician
10 if I'm going to use a drug on somebody with a beat-up
11 ventricle for atrial fibrillation and I know there's
12 a study out there that randomized 1,400 patients with
13 beat-up ventricles, and if anything the drug showed a
14 positive effect, not a negative effect, and this is true
15 in spades for amiodarone, in spades, I'm much more
16 likely to use that drug than another drug.
17 You can argue you shouldn't use any drug.
18 You can argue that maybe you need some other experience,
19 but the fact is you've got to use a drug in many patients
20 and would you rather use a drug for which there is no
21 mortality data or would you rather use one where there
22 is some data that shows that's it's at least neutral?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
265
1 DR. KONSTAM: Well, I have to say we could
2 get into this in all sorts of directions but I don't
3 look at it that way. What I'm looking at is I just really
4 want to ask a question and let's just focus on the
5 question. Is there a potential for excess mortality
6 in the target population here under consideration and
7 what is that level of excess mortality? That's really
8 the question that I need to figure out. I'm going to
9 say to you I am not helped in the least about that
10 question from that study.
11 DR. KOWEY: How about the data that we
12 showed you in the less than 320 milligram group where
13 there was less mortality in those patients than in the
14 placebo ones in the trial? Did that compel you?
15 DR. KONSTAM: In which population?
16 DR. KOWEY: Can we have the core slide?
17 DR. KONSTAM: Okay. In which population?
18 DR. KOWEY: In the population we're looking
19 at here. Can I have the core slide?
20 DR. KONSTAM: That gets into -- that's the
21 next set of data. I'm just focusing on the Julian.
22 DR. FENICHEL: May I say why the agency
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
266
1 thought the Julian study was pertinent? I don't mean
2 dispositive but certainly pertinent. That is suppose
3 the Julian study had come out the other way? Suppose
4 in this relatively fragile population compared to your
5 usual run of atrial fib patients, d,l-sotalol had been
6 extraordinarily toxic and had resulted in increased
7 mortality across the board? That would certainly make
8 one very nervous.
9 DR. KONSTAM: Bob, I don't disagree with
10 that.
11 DR. FENICHEL: Well, okay. But if you say
12 that here was a randomized controlled trial, one of whose
13 possible outcomes was very bad, then that result is
14 informative. Now, the fact that it was not very bad,
15 is it fabulously reassuring? Do we have any reason
16 whatsoever to believe that the life-giving effect that
17 one may read into this, unexamined by FDA studies in
18 that population, should be expected to occur also in
19 the AF population? Of course not. In that respect it
20 doesn't help.
21 But as a means of looking for proarrhythmic
22 effects in a population which has demonstrated in the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
267
1 past its capacity as mind canaries, if you like, who
2 detect those effects, I think that is perfectly find
3 if, once again, it survives FDA review which it has not
4 seen yet.
5 DR. FENICHEL: But it's also a population
6 in which I would construe has more of a potential for
7 benefitting in terms of survival from the antiarrhythmic
8 effect. I guess just the bottom line about my feeling
9 is I don't object to looking at the data for the reasons
10 that you indicated. If there were something worrisome,
11 we've got to look at it to get worried, but all I'm saying
12 is that the absence of seeing something worrisome does
13 not reassure me.
14 CHAIRMAN PACKER: Michael.
15 DR. CAIN: I think the additive part is that
16 it is neutral at best. It's an old study, but I think
17 one has also been faced with the clinical scenario that
18 if you have someone who is recovering from an infract,
19 the beta-blocking effects of sotalol did not achieve
20 the statistical significance in improving mortality.
21
22 If you now had a post-MI patient who had
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
268
1 atrial fibrillation in 1999, could you be doing he or
2 she a disservice by putting them on sotalol and not
3 putting them on a primary beta-blocker that has been
4 shown to have a favorable effect?
5 DR. FENICHEL: Oh, I wouldn't for a minute
6 use those results as the basis for a post-MI claim, which
7 is not being requested.
8 CHAIRMAN PACKER: Let me see. I don't
9 think anyone here is saying anything that is different
10 than anyone else. I think that the definitive database,
11 if you want to be reassured about outcomes, would be
12 an outcome study in patients who are specifically
13 targeted for treatment. Given the event rate in such
14 individuals, which I think is probably, especially if
15 you include those without structural heart disease, is
16 an event rate which is much lower than the post-MI or
17 ICD trials.
18 We're talking about trials of substantial
19 numbers of patients. I'm not saying that should or
20 shouldn't be done. Clearly summation of the mortality
21 data from the existing trials is difficult to interpret
22 because the number of events is so small the confidence
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
269
1 interval is stretched to eternity.
2 Consequently in the absence of -- in an
3 effort to provide some data, just some, they said, "Well,
4 you know, we did these trials for another purpose."
5 They didn't do it for atrial fib. It's clear they didn't
6 do it for atrial fib. They are putting this forward
7 and I don't think they are putting this forward to say
8 that this should be persuasive that there is no excess
9 mortality in atrial fibrillation.
10 I think they are putting it forward to say
11 that, "We did these trials and we want to tell you about
12 them. They are the only long-term outcome trials we
13 have." Maybe they are hard to interpret and maybe they
14 are reassuring but I don't think they are uninformative.
15 They've got to be informative. They've got to do
16 something.
17 DR. CALIFF: I want to speak out in great
18 opposition to Marv here and his feelings about this
19 meaning nothing. To me this is much more meaningful
20 than a 1,000 patients put in atrial fib trials with 10
21 pages of exclusion criteria to take out most patients
22 who are actually going to get the drug in practice.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
270
1 To me this is much closer to the segment of patients
2 where most of the action in terms of cause of death is
3 going to be. This is very important data to me and I
4 would hate to not see it shown.
5 I do agree it is not definitive. The best
6 thing would be 3,000 or 4,000- patients with atrial fib
7 that represented the true population including
8 80-year-olds that are likely to be treated with the drug
9 when it gets in practice but we never get to see that
10 in these meetings.
11 CHAIRMAN PACKER: Abe.
12 DR. KARKOWSKY: Let me bring up one more
13 study which may or may not be relevant and that is the
14 Sword study. Now the Sword study people got d-sotalol
15 and not d,l-sotalol and people here are getting
16 d-sotalol, too, but just having an l-sotalol to
17 counteract it. To the extent that one has comfort, one
18 can diminish that comfort by looking at the Sword study
19 if one believes that is relevant.
20 CHAIRMAN PACKER: Can we address the issue
21 of Sword and the issue of the Julian trial head on in
22 the following way? I mean, you mentioned, Peter, that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
271
1 there is an early apparent increase in mortality in the
2 Julian trial. Admittedly the dose is 320 milligrams
3 once a day and it's an atypical dose.
4 Maybe others can comment on this but there
5 is a dose dependent prolongation at QTc interval up to
6 21 milliseconds with this drug at a dose of 160
7 milligrams BID which is within the recommended dosing
8 range. In the past when we've seen databases of drugs
9 that increase QTc by 21 milliseconds, there's a fairly
10 good torsade signal in those databases with 21
11 millisecond increase.
12 DR. THADANI: Is it really, Milton?
13 CHAIRMAN PACKER: Yes.
14 DR. THADANI: I thought you had to be, you
15 know, chained from baseline by x percent above 520 and
16 20 milliseconds you start with 420 and only go to 444.
17 CHAIRMAN PACKER: I was looking at the
18 dofetilide database and there was a 20 millisecond
19 increase with dofetilide, I think, at their highest
20 dose. They had a torsade signal at 500 milligrams.
21 DR. FENICHEL: Milton, this is a little bit
22 of a digression but what Udho has raised is something
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
272
1 that people don't understand about this. It may be
2 worthwhile. What we have seen with bad actives has been
3 on average QT prolongation, exactly as Milton has said,
4 of on the order of 20 milliseconds.
5 If you look at cisapride it's 21
6 milliseconds or 18 milliseconds. If you look at
7 terfenadine at doses of 200 milligrams, which was higher
8 than were recommended for that drug. It was 23
9 milliseconds or something like that. And across the
10 board there really aren't drugs in common use that raise
11 the average QT an awful lot.
12 And the reason is that, of course, there
13 is highly varying susceptibility to QT prolongation and
14 the average of 20 milliseconds reflects a few outliers
15 who are people who are hypokalemic and people who are
16 women. People who are hypokalemic who start out with
17 long baselines and who, therefore, somehow, unjustly
18 perhaps, seem susceptible to further prolongation, and
19 so forth and so on. But 20 milliseconds of average
20 prolongation is plenty.
21 Now, who are the people who get into
22 trouble? They are not, by in large, the people whose
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
273
1 prolongation is 20 milliseconds, so that was the issue
2 that I think Udho spoke of.
3 CHAIRMAN PACKER: Yes, but I think that's
4 the point. The 21 millisecond average increase here
5 represents a fair number of people who have more than
6 30 millisecond increases who, at least based on the
7 experience with other drugs that have average increases
8 of 20 to 22 milliseconds, usually produces a barely
9 recognizable signal of torsade.
10 I guess what I'm asking is: one, what was
11 the average increase in QTc with d-sotalol at the dose
12 that increased mortality in Sword, and if it was 20
13 milliseconds or 22 milliseconds, is the reason that
14 we're not seeing that signal here because the
15 beta-blocking properties of l-sotalol?
16 DR. FENICHEL: Well, you know, Craig Platt
17 is here and I'm diffident about speaking about Sword
18 but my recollection is that there were five cases of
19 torsade in all of Sword. I mean, there was a hugely
20 increased death rate in the patients who were treated
21 with d-sotalol in that trial but there were only five
22 identifiable cases of torsade, some of which were in
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
274
1 the placebo group.
2 CHAIRMAN PACKER: I'm just curious. So
3 the drug didn't kill people because of --
4 DR. KOWEY: Can I have slide 338, please?
5 This is an analysis that was done of the early deaths
6 in the Julian study. I don't know whether you read about
7 all this but there was a tremendous amount of interest.
8 In fact, Ronnie Campbell, the late Ronnie Campbell,
9 chaired at least two meetings in which there was a very,
10 very intense examination of the death that occurred
11 early in the Julian study. This is the total deaths
12 that occurred in the sotalol and placebo groups.
13 Inflectors, by the way, were people who had
14 a abrupt change in their course on sotalol and,
15 therefore, were considered to be people that would
16 probably have something bad happen from the drug.
17 That's the best I can explain that. It's a very
18 complicated definition.
19 But if you look at what they thought the
20 mode of death was, it was very interesting. The
21 electrical deaths that you would have thought would have
22 been likely because of such a large dose of sotalol,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
275
1 320 milligrams is a single dose, which does have a very
2 high C max and should produce a good deal QT prolongation
3 didn't occur except in the placebo group. There were
4 more what they thought were mechanical deaths in the
5 patients who were receiving sotalol.
6 CHAIRMAN PACKER: The only problem is that
7 I can't believe -- you can't tell how people die. I
8 mean --
9 DR. KOWEY: Well, you know, we're looking
10 at a study that is 16, 18 years old and --
11 CHAIRMAN PACKER: No, no, no. Oh, please.
12 I'm not asking you to do better than this. What I'm
13 asking is, and I guess this was raised by Bob Fenichel's
14 question, there's a database with d-sotalol raising
15 concerns. d-Sotalol increased QTc. Let assume for a
16 moment an increased QTc of about 20 to 25 milliseconds.
17 DR. KOWEY: Okay.
18 CHAIRMAN PACKER: This drug at 160 BID,
19 which is in the recommended dosing range, increases QTc
20 20 to 25 milliseconds, doesn't appear to produce the
21 same torsade signal, and in a patient population of the
22 Julian study was not associated with the same increase
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
276
1 in mortality.
2 That indicates to me support for Marv's
3 hypothesis that there may be two countervailing
4 influences here. One a beneficial one which is limiting
5 the clinical consequences of a prolonged QTc interval
6 and maybe reducing mortality hiding an adverse signal.
7 I hope I expressed that accurately.
8 DR. KONSTAM: Yes. But, you know, I want
9 to ask another question taking that around the primary
10 safety gate is set and, you know, we're saying there's
11 a low torsade signal and a low deathrate. Tell us about
12 the time frame associated with this experience. In
13 other words, we can't really tell from this in terms
14 of the ends. What is the median time of exposure, what
15 is the total number of patient years or what have you?
16 What is the median exposure time that we're looking
17 at in terms of the denominator for the effect?
18 DR. FISHER: Maybe while you are getting
19 it, I can insert one thing. Answering JoAnn's question
20 on 014 about the three people who were entered twice.
21 Because they started from or they had to get into normal
22 sinus rhythm, actually those three people were only
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
277
1 counted once in the original analysis. But since then,
2 the sponsor has re-run it without those three people
3 at all, and the log rank P value goes from .017 to .030.
4 And it was not significant as you consider
5 discontinuations, and that is still true. It goes from
6 .275 to .334. So basically things are the same.
7 DR. GRINES: I have a question about the
8 differences between d-sotalol.
9 DR. KONSTAM: Wait a minute, can we get the
10 median time?
11 DR. KOWEY: Did you want that, Marv, in
12 Julian, or did you want that --
13 DR. KONSTAM: No, no. I am not interested
14 in Julian. In the primary data set.
15 DR. KOWEY: Can I have slide 277, please?
16 Thank you. That was fast. This is number of patients
17 by duration of exposure in weeks.
18 DR. KONSTAM: Can you explain that more?
19 DR. KOWEY: I am sorry. The top is double
20 blind and the bottom is double blind and open label
21 combined.
22 DR. KONSTAM: Right.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
278
1 DR. KOWEY: This is less than 320 and
2 greater than 320. And this is any Sotalol and this is
3 placebo. So this is one week of exposure, 4, 12, 24,
4 and 56. It is the percentage of patients -- the
5 percentage expressed as the number of patients in the
6 group.
7 DR. KONSTAM: Okay. So most of the
8 patients -- I mean I don't know how to -- I mean what
9 this study does. I mean, I guess the easiest way to
10 express this, I think, would be to look at the median
11 time of exposure. So that when we have a denominator
12 in there of the number of patients, how many patient
13 months are we actually talking about here. I get the
14 feeling it is very short median exposure. You know?
15 It seems like most of patients are taken care of with
16 the 4 week group.
17 DR. THADANI: Between one month and four
18 months.
19 DR. KONSTAM: The four-month group. So we
20 are talking about -- you can't lose -- the four-week
21 group. It is weeks, right? The four-week group. The
22 four-week. All right. So we are talking -- do you know
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
279
1 the median time?
2 DR. KOWEY: I don't have it in front of me.
3 But we can see that. We have got it somewhere in here.
4 DR. KONSTAM: I think it is very important
5 when we are talking about one percent death rates and
6 not all one percent death rates are the same. Certainly
7 a one percent death rate over a four-week exposure is
8 not the same as the one percent death rate in a typical
9 survival study. So just to make that point.
10 DR. THADANI: In that context, could he
11 show the death rates to the time too or what? I know
12 there were only two deaths. Did they occur early or
13 late?
14 CHAIRMAN PACKER: But the numbers are so
15 small, what are you going to do with them? I mean, how
16 many ways can you cut four deaths?
17 DR. KONSTAM: I think that is the point.
18 I think the point is the numbers are really small and
19 not in the least reassuring, therefore.
20 CHAIRMAN PACKER: Tom?
21 DR. KOWEY: To put this in some
22 perspective, Marv. You weren't on the committee when
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
280
1 other antiarrhythmic drugs were approved for this
2 indication. But the flecainide data base was
3 substantially smaller than this data base. And for
4 flecainide, there was a mortality MI study that actually
5 went in the wrong direction. And flecainide was
6 approved for a defined patient group at a defined dose.
7 So it is a little bit inconsistent to be saying that
8 this is a tiny number of patients. It is not a tiny
9 number.
10 DR. KONSTAM: No, I know.
11 DR. KOWEY: It is actually a substantial
12 number and there is a mortality study -- two of them
13 in fact -- that go in the direction of benefit.
14 DR. KONSTAM: Peter, that is -- I
15 understand. You are getting into questions of
16 interpretation. I just want to say I don't -- I mean,
17 I am not reassured by any of that. I mean, I guess I
18 am just trying to say if you want to make a case that
19 sotalol is associated with no excess mortality of
20 importance, I am only making the point that I can't
21 conclude that at all. That is all.
22 DR. KOWEY: But, Marv, if you can tell me
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
281
1 what antiarrhythmic drug you can conclude that for --
2 DR. KONSTAM: I am not trying to argue that
3 point.
4 DR. KOWEY: No. But it is a problem with
5 every single antiarrhythmic drug we have for AF because,
6 as Rob said earlier, we don't have 3,000 patient studies
7 in the appropriate patient population. What we have
8 is a defined data set, and then you have some other
9 studies tacked on. I agree it is not perfect, but it
10 really isn't all that bad compared to what we have seen
11 in the past.
12 CHAIRMAN PACKER: Cindy?
13 DR. GRINES: I guess I wanted to just point
14 out -- I know you don't want to hear about the Julian
15 study, but they did treat them for 12 months and actually
16 the withdrawal rate, according to the article, is around
17 25 percent of the patients receiving sotalol and 21
18 percent of the placebo. So that withdrawal rate isn't
19 quite as high as the ones in the a fib trial. And I
20 just wondered, if we are going to talk about d,sotalol
21 and the Sword trial, I guess maybe I need some
22 clarification about how it differs from the d,l variety.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
282
1 Because it seems in the atrial fibrillation trials that
2 d,sotalol was not effective at reducing the incidence
3 of a fib. And I don't know enough about the two
4 preparations.
5 DR. KOWEY: Well, one is -- one contains
6 a beta blocker and is an IKL blocker. That is sotalol.
7 And one is an IKL blocker, which is a common garden
8 variety antiarrhythmic drug.
9 DR. GRINES: So the d,sotalol is just an
10 antiarrhythmic drug?
11 DR. KOWEY: It is just an anti -- it has
12 no beta blocker.
13 DR. GRINES: And the d,l is the beta
14 blocker?
15 DR. KOWEY: Yes.
16 DR. GRINES: Okay .
17 CHAIRMAN PACKER: Or the d,l is both.
18 DR. KOWEY: Both.
19 DR. GRABOYS: Peter, you know I think we
20 are going over this in such picayune detail because it
21 is not simply using an antiarrhythmic drug in terms of
22 using an antiarrhythmic drug and acknowledging the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
283
1 toxicity of the drug and excess mortality. We are
2 talking about using an antiarrhythmic drug that is
3 potentially toxic for a population that is soft in its
4 indication for the use of the drug. For me to take or
5 to use this drug in a patient, in a 75-year-old patient,
6 because I am hoping that maybe they will have less
7 palpitations at 6 months than with placebo they had less
8 palpitations at 3 months, it makes me very concerned
9 and we get back to "first do no harm." And that is really
10 why there continues to be a lot of this discussion.
11 DR. KOWEY: I would say, Tom, that in
12 clinical practice if someone merely has a few
13 palpitations now and again, I think that it is probably
14 wrong to use a drug that has a powerful effect on
15 repolarization to treat them. Especially if it is in
16 a group of patients for whom there appears to be some
17 excess chance of harm. I don't disagree with you. But
18 there is a universe of patients who have atrial
19 fibrillation that is very symptomatic and they want to
20 have it treated. They want to have those symptoms
21 reduced. And in order to do that, we have to use an
22 antiarrhythmic drug. The question is in the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
284
1 antiarrhythmic drugs we have available, where does this
2 fit in? There is a definable patient population for
3 whom this drug may be useful. It is just that it is
4 not -- it is not the universe of AF. It may not even
5 be the majority of AF. But it is a definable patient
6 population. It has been defined in the clinical trials.
7 And I think that to say that this drug can't be used
8 for anybody with AF because we are concerned about a
9 group of patients at the end of the spectrum of risk,
10 I don't think that is right. That is where the rub comes
11 in.
12 CHAIRMAN PACKER: Marv?
13 DR. THADANI: Can I ask a question of Bob
14 Fenichel here? You might remember when we discussed
15 the aspirin issue -- going off the track here. Aspirin
16 was recommended for approval on the basis of separate
17 trials, which is a Scandinavian -- Sotalol plus aspirin
18 versus Sotalol plus placebo. Do you recall? I can't
19 remember now what the incidence of torsade was. Because
20 they had a 3,000 patient population. It was a very neat
21 study. They were not on anything else.
22 DR. FENICHEL: No one who doesn't use
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
285
1 Holters has any idea of the incidence of torsade. You
2 can't find torsade in a spot check population. Torsade
3 comes and goes. So I don't remember the study, but
4 unless they did that, they don't know how much torsade
5 there was either.
6 DR. THADANI: But could you also say that
7 torsade based on the -- I realize the IC data, people
8 don't like it. But we may not necessarily care, like
9 you said. The reason the incidence that Milton was
10 pointing out in some studies is higher is because they
11 did the Holters. And in this study, they never did the
12 Holters, so you never know the true incidence of torsade.
13 It may be much higher than the reported incidence
14 sometimes that you read in the literature as 4 to 6
15 percent.
16 CHAIRMAN PACKER: Udho, I don't think you
17 are quoting me correctly. The torsade signal comes from
18 clinical events not from Holters. So that the --
19 DR. THADANI: Not --
20 CHAIRMAN PACKER: No, not from Holters.
21 Not from Holters. It is from clinical events. There
22 is a discernible signal. I think in many antiarrhythmic
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
286
1 drugs -- I think Bob outlined some of the examples.
2 At 20 millisecond increases, you are getting a signal
3 for lots and lots of drugs.
4 DR. THADANI: I don't -- Milton, I think
5 I have seen Holters and patients don't complain of
6 anything and they have 40 seconds. So it looks like
7 polymorphic as it moves around.
8 CHAIRMAN PACKER: It may be. But that is
9 -- but --
10 DR. THADANI: I think you pick up more on
11 a Holter than on a random transtelephonic monitoring.
12 So I think the incidence probably is underestimated.
13 CHAIRMAN PACKER: No, no. But the data
14 bases that Bob Fenichel cited were clinical event data
15 bases. Unequivocally. They were not -- the risk
16 identified with those drugs was based on clinical
17 events, not on Holters. Not Holter-detected
18 asymptomatic torsade.
19 DR. THADANI: On the Vaprodil data base,
20 I remember when we were doing those studies, we had
21 Holters. And on aspirin buffered Elvaclo, we had PVCs.
22 We put Holters on and we stopped the study because there
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
287
1 were a couple of -- very few deaths, but I think there
2 were some incidences of torsade which was higher. So
3 there is some data there.
4 DR. FENICHEL: Even depodril. Deprodil is
5 probably the worst proarrhythmic drug around. It is
6 approved as second line therapy only. And I think the
7 number of identified arrhythmias in the preclinical --
8 in the preapproval data base was -- you could count it
9 on your hand. It was very, very few. You know, it is
10 very hard to find these signals. We care about these
11 signals, but that doesn't mean there are so many of them
12 that they will show up in samples of this size.
13 CHAIRMAN PACKER: Marv?
14 DR. FISHER: Marv, can I make one quick
15 comment just for your information? From the slide up
16 there, you can get an underestimate of the average
17 exposure, which is 12.6 weeks. So it probably is around
18 16 to 18, I would guess, from --
19 DR. KONSTAM: What's -- I am sorry, I don't
20 understand.
21 DR. FISHER: This is the average exposure
22 of the people in the sotalol group. You are talking
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
288
1 about numerators without denominators. You know, what
2 is the death rate for --
3 DR. KONSTAM: Right, right, right.
4 DR. FISHER: I am just giving you a rough
5 idea of the --
6 DR. KONSTAM: 16 to 18 weeks?
7 DR. FISHER: That is a guess. A
8 mathematical under bound. But it is certainly correct
9 to 12.6.
10 DR. KONSTAM: A question about the dosing,
11 Peter. What is the evidence that 160 mg bid is better
12 than 120 mg bid?
13 DR. KOWEY: The 9A study had a better
14 outcome in patients that were at 160. And in addition,
15 the 004 study had two-thirds of the patients on 160 mg.
16 DR. KONSTAM: Right. But it didn't have
17 a 120 mg dose.
18 DR. KOWEY: And if you are just talking
19 about efficacy, and we don't want to get back to where
20 we were this morning, 160 beat 120 -- I am sorry, 160
21 had a higher number of patients event-free at 12 months
22 in 05. So it was in one parameter better than 120.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
289
1 In the other parameter, it was less good than 120. But
2 120 was an effective dose in 05, and was significantly
3 better than 80.
4 CHAIRMAN PACKER: Does anyone have any
5 other questions about safety or dose response? Can we
6 move on to a conclusion of the sponsor's presentation?
7 DR. MARROTT: Mr. Chairman, members of the
8 Advisory Committee, and Dr. Fenichel, Dr. Kowey has
9 presented with clarity a balanced overview of the
10 clinical data and the potential of sotalol for treating
11 patients with atrial fibrillation.
12 I guess no clinical data base is squeaky
13 clean, and the sotalol data base is certainly not an
14 exception. However, it appears that the efficacy data
15 from studies presented point in a similar direction.
16 All are positive and significant to varying degrees.
17 This is true whether patients have chronic or
18 paroxysmal atrial fibrillation or structural or no
19 structural heart disease. In the latter group of
20 patients, the data suggests that outpatient
21 administration can be safely undertaken.
22 The risk from serious adverse events is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
290
1 present as with any other antiarrhythmic drug. But it
2 is extremely low. We recognize that safety is of
3 paramount importance when treating atrial fibrillation
4 or flutters, arrhythmias which in the majority of
5 patients are not life-threatening. The sponsor bears
6 the responsibility for providing conditions that will
7 minimize the safety risks, and we would like to discharge
8 this responsibility diligently.
9 We have discussed the possibility of a risk
10 management program with the FDA. The three key
11 objectives of this program shown on the slide are to
12 differentiate our product when used in atrial
13 fibrillation or flutter and when used in ventricular
14 tachycardia or fibrillation.
15 Second, to provide patients with atrial
16 fibrillation valid information and support. This is
17 in keeping with the current notion that patients should
18 understand their treatment and be allowed to have a
19 greater say in their health and well-being.
20 And third, to ensure that the healthcare
21 professional is better informed and better educated.
22 We will continue to develop this theme and discuss our
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
291
1 ideas with the FDA. This concludes the sponsor's
2 presentation. I would like to thank you, Mr. Chairman
3 and members of the Advisory Committee, for your
4 attention and for the courtesy you have given to us.
5 Thank you.
6 CHAIRMAN PACKER: Any other comments or
7 questions from any member of the committee?
8 DR. CALIFF: There was some discussion
9 about post-marketing surveillance data related to
10 sotalol. Is that going to be discussed?
11 CHAIRMAN PACKER: There were
12 post-marketing surveillance data that were summarized
13 in the briefing document briefly.
14 DR. CALIFF: Yes.
15 CHAIRMAN PACKER: It hasn't been formally
16 presented. Did you have any questions about it, Rob?
17 DR. CALIFF: Well, it seemed confusing.
18 I didn't know what to make of it. But there seemed to
19 be a lot written about it by the FDA.
20 CHAIRMAN PACKER: Most of the comments are
21 in the FDA review -- some of the supplemental reviews
22 the committee received within the last 10 days.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
292
1 DR. CALIFF: And to me it is very important
2 because we don't know a lot about post-marketing
3 surveillance, but the public believes that there is some
4 way that we can tell if drugs are safe or not once they
5 get on the market through this methodology. Here we
6 have a drug that has been used for many years. It has
7 been used a lot for the indication being sought here
8 and we have post-marketing surveillance data. And the
9 question is it of any value?
10 DR. KOWEY: We can present a few slides if
11 you want to see them. Dr. Jin, do you want to hop in
12 and show some information?
13 DR. JIN: Okay. Slide 381, please? Okay.
14 Post-marketing adverse events recorded did not show
15 anything really surprising than what you have seen from
16 clinical trials. The most common adverse events were
17 fatigue, weakness, wheezing, shortness of breath and
18 bradycardia. But the most significant ones probably
19 or the ones of concern were torsade, VT, VF, cardiac
20 arrest and syncope. They are presented on this slide.
21 In this five years -- six, I am sorry -- six-year period
22 from 1993 to 1998, FDA has received a total of 46 case
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
293
1 reports of torsade, VT/VF, and cardiac arrest. And
2 Berlex laboratory had received 28 case reports because
3 some of the cases we reported to the FDA directly. And
4 10 of these 46 cases died.
5 Based on our marketing research data from
6 IMS and from other sources, we estimated that roughly
7 about 2,000 person year exposures of Betapace in this
8 six-year period. And this is presented in a rate of
9 per 100 person years. So that means the reporting rate
10 of domestic torsade, VT/VF and cardiac arrest is about
11 2 per 10,000 person years of exposure. So it is very
12 rare. But obviously we don't know the under-reporting
13 rate and the spontaneous report is subject to the
14 under-reporting. Not every case is reported to us or
15 to the agency.
16 For foreign cases, there were 27 cases in
17 FDA it has said. We know of 72 through Bristol-Myers,
18 because it is marketed by Bristol-Myers, and we don't
19 have denominator information to provide a reporting rate
20 of adverse events.
21 Can I skip the next one and go to the third
22 slide, 383? The other significant adverse event from
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
294
1 post-marketing reports is bradycardia. Again, FDA has
2 said there were 43 case reports of bradycardia and 8
3 of these cases died. And actually 5 of these 8 cases
4 are included in the torsade slide I just presented.
5 Again, Berlex received 32 case reports of bradycardia.
6 And the adverse event recording rate of bradycardia
7 is 2 per 10,000 person year exposures. So both show
8 that the reporting rate is extremely rare. And I guess
9 nobody can really tell you exactly the magnitude of
10 under-reporting to give you an incidence rate of torsade
11 or bradycardia from the commercial use of the product.
12
13 DR. THADANI: Do you know what the death
14 rate will be in a similar patient population who is not
15 on the drug? Because the event rate looked very low.
16 So is there some data available? I am sure there are
17 some population-based studies available to show that.
18 DR. JIN: I do see that for torsade I saw
19 about one-third would die not on this product.
20 DR. THADANI: But without this product, if
21 you took the same patients, what would one see?
22 DR. CALIFF: So, Udho, what you are asking
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
295
1 is what seems like just a -- I mean, it makes you wonder
2 why we go to all this trouble. What you end up with
3 is a numerator with no denominator. And even the
4 numerator doesn't have a denominator for the likelihood
5 that the numerator would ever be reported. So what --
6 Bob, maybe you can tell us what value -- I mean, we are
7 evaluating clinical trials done in highly selected
8 populations, not representative of patients who will
9 actually be treated. Then we put the drugs out there.
10 Some information comes in and we can count up the things
11 that come in. But we have no earthly idea what the
12 denominator is or what the control population would have
13 been. Is this any better than just how the doctor feels
14 on that day about the drug? Or what is the value of
15 all this?
16 DR. THADANI: I don't think it is valuable.
17 DR. FENICHEL: Well, I think data like this
18 are extremely hard to interpret. I think we collect
19 these looking for unusual events -- events whose rate
20 compared to the background rate can be defined so that
21 it is much more interpretable, although still difficult
22 to see TTP or agranulocytosis or fulminant hepatic
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
296
1 failure or something else whose background rate is on
2 the order of 1 in 100,000 or 1 in a million. When we
3 look at death, which in an ordinary middle-aged
4 population occurs at the rate of 1 percent per year,
5 and in an ordinary elderly population occurs at higher
6 rates than that. And in a population with substantial
7 organic heart disease, it is even higher, perhaps into
8 the 6 to 8 percent rate. I don't know what to make of
9 this.
10 CHAIRMAN PACKER: Okay. All right.
11 Thank you very much. Any final questions? Okay, thank
12 you. We will proceed to the questions. Question 1,
13 atrial fibrillation/flutter may be associated with
14 disabling symptoms or with no symptoms at all. Whether
15 or not it is accompanied by symptoms, atrial
16 fibrillation is associated with an increase in the risk
17 of stroke. Without regard to the data about sotalol,
18 what sort of data should be required with respect to
19 any drug for atrial fibrillation? Is deferral of
20 relapse into atrial a fib sufficient, or must some more
21 immediate patient benefit, for example reduced symptoms
22 or reduced incidence of stroke, be part of any approvable
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
297
1 claim?
2 The question here focuses on what
3 constitutes an approvable package for a claim for atrial
4 fibrillation. JoAnn?
5 DR. LINDENFELD: I think that at least
6 recently we have considered deferral of relapse as a
7 reasonable basis for a claim. But I do think this whole
8 study brings up that ideally what it would be nice to
9 see is if that change in the incidence of atrial
10 fibrillation results in some measurable symptomatic
11 outcome -- exercise capacity, symptomatic benefit,
12 fatigue. I think that would be ideal in this study.
13 And we don't really know that from this study. That
14 was the point I was making earlier about actual symptoms.
15 We don't really know that at the end of the day the
16 symptoms were different.
17 CHAIRMAN PACKER: Yes. I am just trying
18 to -- can you hold that thought for a moment? I am trying
19 to figure out how we get from A to B. Because if you
20 have a patient who has a paroxysmal atrial fibrillation
21 and so they are enrolled in trial when they are in normal
22 sinus rhythm. And they are arbitrarily reevaluated at
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
298
1 a fixed point in time after or during the course of double
2 blind therapy. Also in normal sinus rhythm, it would
3 be hard to know how one evaluates exercise tolerance
4 or symptoms or fatigue or anything because they are in
5 the same rhythm. I guess ideally one would -- well,
6 I don't know ideally what one would do. How do you get
7 from A to B? How do you actually evaluate something
8 that is a transient recurrent intermittent event that
9 you are trying to put a symptomatic measure on, aside
10 from the symptoms that the patient would report while
11 they have the event.
12 DR. LINDENFELD: Yes, I don't know how to
13 do that either in the paroxysmal arrhythmias. But I
14 think in the patients who relapse into chronic atrial
15 fibrillation, it would be nice to know if the percentage
16 is higher of those who remain in sinus rhythm. Do they
17 actually feel better than the ones that have reverted
18 to atrial fib. And that is a measurable outcome. So
19 in that fairly large group of patients at least that
20 would be measurable. I don't know that I know how to
21 do it in the paroxysmal. I doubt quality of life would
22 really capture that.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
299
1 CHAIRMAN PACKER: All right, Udho?
2 DR. THADANI: Then you could do that by
3 giving a diary. That would be one way. If you really
4 believe in exercise tolerance, you can put them on a
5 pedal speedometer and see how the patient walks every
6 day.
7 CHAIRMAN PACKER: When would you measure
8 exercise? When they are in sinus rhythm?
9 DR. THADANI: Well, in a day how much
10 distance they walked. Because a lot of paroxysmal a
11 fib, at least some of the patients I see in coronary
12 disease are induced by exercise too. So if you could
13 have a daily record. I mean, I could give you an idea,
14 but it is an impossible task. And I think if you really
15 believe in the -- since the question is combined with
16 strokes and symptoms, one way would be to have put one
17 patient on say Coumadin, and then another patient group
18 with a paroxysmal a fib not on Coumadin and see the stroke
19 difference rate, which will be a tough issue.
20 CHAIRMAN PACKER: You can't do that. Your
21 patients are on anticoagulation.
22 DR. THADANI: But most of the time now what
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
300
1 we are doing is we put patient on anticoagulants and
2 after six months if you see them in sinus rhythm, they
3 are taken off. And I think there are ways you could
4 do the trial. The question is open. So I am just giving
5 you some of the issues one could address. But it is
6 tough to document it in a trial.
7 DR. GRABOYS: I don't think the Coumadin
8 issue is germane at all. The standard of care now with
9 an increasingly elderly population is that regard --
10 once they have declared themselves in having AF, the
11 standard of care mandates that they be on anticoagulants
12 period.
13 CHAIRMAN PACKER: Marv?
14 DR. KONSTAM: No, but there is another side
15 to this coin, which is that if you believe, as I happen
16 to believe, that there is something good about being
17 in sinus rhythm. So take that as a potential useful
18 endpoint. But then ask the question, however, is there
19 something adverse going on simultaneously. So I think
20 one of the issues really here is are there -- is there
21 something about this drug that is driving adverse
22 effects on symptoms, which would make it -- if you were
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
301
1 concerned about that and I think we ought to be concerned
2 about that -- then you would want to measure symptoms
3 across the population to at least reassure yourself that
4 that is not going on. So I understand that doesn't
5 directly address the question of symptomatic a fib, but
6 it is another point.
7 CHAIRMAN PACKER: Okay. But let's directly
8 address the question. The question says what
9 constitutes an approvable claim. What data base
10 constitutes an approvable claim? Who on the committee
11 would suggest that an approvable claim require evidence
12 for reduction in the risk of stroke? Anyone? Who would
13 require an improvement in exercise tolerance? Who
14 would require -- and stop me when I get to something
15 you like.
16 DR. THADANI: You are talking about
17 paroxysmal a fib now or chronic a fib?
18 CHAIRMAN PACKER: Even in the chronic a fib
19 studies, they are cardioverted.
20 DR. THADANI: But say if you don't
21 cardiovert and the patient is in a fib?
22 CHAIRMAN PACKER: That is a different
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
302
1 claim.
2 DR. THADANI: No, no. I realize that.
3 You are talking about if he stays out of the a fib, he
4 benefits. But if he is in chronic a fib, the exercise
5 tolerance might go down and it could even slow the heart
6 rate and you could improve the exercise tolerance. And
7 I think you probably have to dissociate between chronic
8 a fib and paroxysmal a fib on that issue.
9 CHAIRMAN PACKER: Michael?
10 DR. CAIN: I think for the way that they
11 have defined paroxysmal and chronic for this study and
12 for the question that you are asking, it doesn't matter.
13 You are talking about someone who has been in atrial
14 fibrillation for some period of time and is now through
15 whatever mechanism in sinus rhythm. And now what you
16 are trying to judge is do you feel better or worse in
17 sinus rhythm than the way you feel should you happen
18 to go back into atrial fibrillation. And so I think
19 for the way that it is being defined here, chronic and
20 paroxysmal groups are essentially the same thing.
21 DR. KONSTAM: That is right.
22 CHAIRMAN PACKER: Okay. Let me see if I
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
303
1 got it. If a sponsor comes to this committee with a
2 data base that shows a reduction in time to first
3 symptomatic atrial fibrillation, is that okay? Anyone
4 thinks -- does anyone think it is not okay?
5 DR. PIÑA: Did you just put the word
6 symptomatic in there now?
7 CHAIRMAN PACKER: No. The episode is
8 symptomatic. The episode is symptomatic. The concept
9 that the differential here is that the episode is
10 symptomatic, but under usual clinical trial
11 methodology, one assesses symptoms at a fixed point in
12 time. They may or may not be in atrial fib. It is hard
13 to know how to assess that if they are in normal sinus
14 rhythm. You are not actually addressing the question.
15 You are actually addressing a safety issue rather than
16 an efficacy issue under those circumstances.
17 DR. FENICHEL: Milton, let me clarify the
18 question and also respond to something that Marvin said
19 much earlier in the day, raising a whole new version
20 of the question. The idea when the question was written
21 was could one submit a claim -- would it be an approvable
22 claim to come in with data consisting entirely of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
304
1 electrocardiographic measurements? In other words, to
2 show that the actively treated population had better
3 looking electrocardiograms than the group treated with
4 placebo and independent of any demonstrated effect upon
5 symptoms or upon the risk of stroke. So is this a --
6 is that laboratory finding, if you like, sufficient,
7 or does it need to be accompanied by some clinical
8 benefits such as the patient says he feels better.
9 Despite all the other miscellaneous unrelated effects
10 of the drug, it is still so much nicer to be in sinus
11 that he is willing to put up with the diarrhea and
12 vomiting or whatever it is.
13 Now the other side of the question which
14 Marvin raises, which I had never considered, is not so
15 much is the electrocardiographic victory sufficient,
16 but Marvin raises the question, as I understand his point
17 earlier today, of is the electrocardiographic benefit
18 necessary? Suppose that through rate control or some
19 other means, the patient was made to feel so much better
20 despite ongoing or perhaps even increased fraction of
21 time in atrial fibrillation because he is flipping in
22 and out but he does it at such a low rate or in such
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
305
1 a numb state that he does not object to it. Is that
2 okay? Because he does feel better, even though
3 electrically he is worse. Now as I understand what
4 Marvin brought up earlier is he said, no, no, that would
5 not be acceptable. The symptoms are indeed a surrogate
6 for the real benefit, which is an electrical benefit.
7 Well, that is a respectable point of view too. So I
8 guess the two questions are or would be is the
9 electrocardiographic benefit sufficient? That was the
10 original question. And the new question prompted by
11 Marvin is the electrical benefit necessary?
12 CHAIRMAN PACKER: Rob?
13 DR. CALIFF: Yes. As far as the
14 electrocardiographic benefit, I think that is a nice
15 scientific benefit, but not one that is germane to the
16 public health. So while I would like to see that as
17 part of any effort, and certainly since we are attempting
18 to base our eventual clinical benefit on science, I think
19 it should be a part of the protocol. But essentially
20 I don't think people are taking drugs because of their
21 electrocardiograms. They want to have their symptoms
22 alleviated. So for me, the critical things are that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
306
1 there be some demonstration of improvement in symptoms
2 and that there be enough patients representing the kinds
3 of patients in whom the drug will be given to rule out
4 any plausible unacceptable increase in risk of bad
5 things happening. And I think in order to do that, what
6 I would really like to see being done would be to push
7 the research community to do more inclusive studies,
8 particularly in atrial fibrillation. I really agree
9 with Tom. We need to be including people over age 80.
10 We have this term therapeutic orphans now for children
11 because randomized trials have not been done in children
12 and therefore we don't know how these drugs work. But
13 I think the elderly now are in exactly the same
14 situation. And the trials need to be larger. When you
15 are looking at trials of 100 patients in each type of
16 atrial fibrillation, it is going to give us some nice
17 information. But that should be a subset of a larger
18 study that can really allow us to estimate even by
19 Marvin's strictly sticking to the atrial fibrillation
20 group of what the real risk is.
21 CHAIRMAN PACKER: Ileana?
22 DR. PIÑA: Rob, are you saying that if the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
307
1 drug doesn't convert a fib into sinus but the patient
2 doesn't have symptoms anymore because they are feeling
3 better and maybe their rate when they are in a fib is
4 slower, which is something that Bob was saying, it is
5 okay and they don't have to demonstrate efficacy by
6 keeping the patient in sinus? In other words, is being
7 in sinus rhythm better than being in a fib, even if you
8 don't have symptoms?
9 DR. CALIFF: Well, I think in this case we
10 are fortunate that we have a methodology that will tell
11 us about the ECG and can be done in a smaller sample
12 size than what it takes to have an adequate safety data
13 base. So I think we can do both here. But Marvin was
14 actually raising the converse of that, which is that
15 they may be worse. They could still be in atrial fib
16 and not know it because you may be tempted not to take
17 your anticoagulant drug. But here again that is where
18 an adequate safety data base in a population that was
19 really at risk of having strokes, for example, would
20 be very helpful in knowing whether it was better or
21 worse. But being symptomatically better, I think, with
22 a slower rate, to me that would be a nice thing to have
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
308
1 as long as it didn't increase your risk that something
2 else bad happened.
3 DR. THADANI: On paroxysmal, there is a
4 trial ongoing. NIH is doing a firm trial just to address
5 that issue. The patients in one is very controlled.
6 The other one is keeping them in sinus rhythm post
7 cardioversion. So I think those issue are very relevant
8 and they are looking at outcome in a very large sample
9 size. So the symptoms are important because we get
10 patients reporting with rigor and they fail, who are
11 ending up having oblation of their AV node just to get
12 rid of the symptoms. So I think it depends on how
13 symptomatic a patient is. I think that is very
14 critical. If the symptoms are mild or the symptoms are
15 bad enough that he would need hospitalization. So I think
16 the question is being addressed in a very large NIH
17 trial.
18 CHAIRMAN PACKER: Yes. It is not the
19 question --
20 DR. THADANI: It is not the relevant issue
21 here.
22 CHAIRMAN PACKER: It is not the question
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
309
1 we are talking about. The analogy here, although Bob
2 Fenichel hasn't made it, is probably the analogy of how
3 this committee and the agency evaluates antianginal
4 drugs. The data base for antianginal drugs -- and I
5 understand we haven't seen one for a long time -- is
6 that there are two kinds of data bases. One is a
7 symptomatic data base and the other is a physiologic
8 data base -- a prolongation of exercise tolerance or
9 a prolongation of time to a specific ST segment
10 depression. The Agency -- Bob Fenichel, please correct
11 me if I am wrong -- has taken the point of view that
12 reduction in symptoms per se is insufficient because
13 one might achieve that with morphine. That one requires
14 both a -- well, I should say one requires a reduction
15 in the -- it requires the physiologic response and it
16 would be nice to have the symptomatic response, would
17 that be correct?
18 DR. FENICHEL: No. No, that is not right.
19 The basic claim is the symptomatic claim. You need
20 to have a lot of evidence of the symptomatic claim and
21 then you need to have what I think would fairly be
22 described as merely supporting evidence of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
310
1 anti-ischemic claim. So it is not strictly analogous.
2 Well, what you said is right, but it is strictly
3 analogous. In other words, what you would -- as I
4 described it, yes, it is a two component claim and it
5 is a symptomatic claim. It is not an outcome claim,
6 for example. But there is this physiologic or
7 electrophysiologic component which is necessary. It
8 is a non-ischemic component. The situation here, and
9 indeed that is the analogy which I think the original
10 question was meant to draw out -- is this something where
11 you could simply get a drug approved for ischemia in
12 the case of antianginal drugs? And of course that is
13 an acceptable indication in some parts of the world.
14 In Europe, the drugs are approved as anti-ischemic.
15 And a demonstration of symptom relief or exercise
16 improvement or something like that is not required.
17 Well, we don't do that. The question here is this like
18 angina, where you need symptoms and the cardiogram
19 should go in the right direction? Or is the model,
20 Marvin's model, where the electrocardiogram is really
21 the disease and the symptoms are a surrogate for true
22 patient factor, that is a fairly radical idea. But it
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
311
1 is not --
2 DR. KONSTAM: Since this is my model, can
3 I kind of refine it a little bit?
4 DR. FENICHEL: Yes.
5 DR. KONSTAM: I think you set up the
6 discussion just I think in the right way, Bob. But let's
7 maybe focus on it from the perspective of an efficacy
8 claim. And I guess take them in two steps. So, first,
9 just the ECG, which is the way the question is posed.
10 And let me give you my answer to that. It is I would
11 be accepting of that on the grounds that -- and we
12 discussed this at length around elfedalide, and there
13 were some very persuasive arguments made and I
14 personally accepted them that if you knew that a drug
15 kept somebody out of atrial fibrillation. And I
16 personally would accept that from a perspective of
17 efficacy. And I can justify it on the grounds of the
18 stroke story. You know, not everybody can take
19 anticoagulation. There are a variety of reasons. But
20 I think that to me is sufficient to say that that could
21 be a claim for efficacy. So that is my opinion about
22 that.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
312
1 With regard to the other side of it as you
2 have raised it, I would refine really what I really want
3 to say here. I actually would be accepting of an efficacy
4 claim for, as the sponsor has set it up, prevention of
5 symptomatic atrial fibrillation, accepting the fact
6 that some of that is going to be contributed to by rate
7 control and some of it may be contributed to by
8 preventing the a fib. On the grounds that that is
9 preventing symptoms. So I can accept that. I guess
10 where I am going to get into issues around it is when
11 we get -- when we move from the question of efficacy
12 to the question of approvability. Because to me it
13 makes a big difference if we wind up concluding that
14 the way this drug works is by reducing heart rate when
15 the patient is in a fib. Then that makes me look at
16 the data set very differently and it makes me say, well,
17 we have other drugs that do that. And so when we get
18 to the benefit/risk ratio, to me it shifts it. And that
19 is really -- it wasn't so much to say that I couldn't
20 accept that as an efficacy endpoint. It is just that
21 I think that the mechanism is important in looking at
22 the totality of the question.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
313
1 CHAIRMAN PACKER: Okay. Let's -- this is
2 not a specific issue or question about sotalol. So why
3 don't we just go down the committee and phrase the
4 question as follows. Should the primary basis for the
5 approvability of a drug for atrial fibrillation be the
6 ability to suppress symptoms? Should the primary basis
7 be the ability to suppress ECG recurrence of the
8 arrhythmia? Or are both -- or should both be required?
9 In other words, the first possibility is that symptoms
10 are very important and the ECG evidence is supportive.
11 The second is ECG is important. Suppression is
12 important. And the symptomatic relief is supportive.
13 Or do you feel that both are required? Bob, would that
14 answer the question?
15 DR. FENICHEL: Yes, that is it.
16 CHAIRMAN PACKER: Good. Lem, why don't
17 you start -- we will start -- oh, I am sorry, JoAnn,
18 our primary reviewer, tell us what you think.
19 DR. LINDENFELD: I think EKG evidence is
20 sufficient without symptoms. I think symptoms would
21 be supportive.
22 CHAIRMAN PACKER: Okay. Lem, while we --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
314
1 DR. MOYÉ: I would say both.
2 DR. BIGGER: I have a different answer.
3 I would say either. Either by itself without the other.
4 CHAIRMAN PACKER: Okay. We didn't include
5 that, but one could. Would anyone change their answer
6 based on that possibility? No. Okay.
7 DR. GRINES: Actually, I might change my
8 choice if you had either.
9 CHAIRMAN PACKER: Okay. Tom?
10 DR. GRABOYS: I would like to see both.
11 CHAIRMAN PACKER: Marv?
12 DR. KONSTAM: I'd say either.
13 CHAIRMAN PACKER: Okay. Michael?
14 DR. CAIN: Either. And the scenario would
15 be that you could have somebody who is dizzy all the
16 time but the drug puts them in sinus rhythm. Although
17 they are still dizzy, they are in sinus rhythm. And
18 that would be either.
19 CHAIRMAN PACKER: Ileana?
20 DR. PIÑA: I would go with either.
21 CHAIRMAN PACKER: Udho?
22 DR. THADANI: I would go both, especially
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
315
1 based on the past experience. You could suppress PVCs
2 and patients could die. So if the patient is both
3 symptomatic and the ECG is better, are you going to say
4 that drug is effective? So I will go for both.
5 CHAIRMAN PACKER: Okay. I would vote for
6 both as well. Let me just say that, Joan, from what
7 we have outlined, there is a slight preference for either
8 for the committee. Let me just say that I think it is
9 important to, although both Tom Bigger and Michael Cain
10 don't officially vote, this is more of a general
11 question. And, Bob, you are getting a little bit of
12 a mixed message here, but I think you are getting a sense
13 of where the committee stands.
14 DR. FENICHEL: Yes. That's fine.
15 CHAIRMAN PACKER: That means the committee
16 feels the way it does today and would be anxious to look
17 at other data bases tomorrow. Okay, question number
18 2. It is also a generic question. I am not going to
19 read the question, but we all know the issue of drop-out
20 rates and the issue that we talked about of informative
21 censoring. Let me --
22 DR. FENICHEL: Milton, I think that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
316
1 question was really adequately discussed.
2 CHAIRMAN PACKER: Fine.
3 DR. FENICHEL: And since this doesn't bear
4 upon the specific drug, I think we can skip it now.
5 CHAIRMAN PACKER: Fine. I think it would
6 be fair to say, Bob, that the committee was unanimous
7 in indicating what they thought was the right path to
8 follow. Number three, most of the patients in --
9 question 3 deals with paroxysmal atrial
10 fibrillation/flutter. There are two studies. There is
11 study 05 and there is study 9A. And considering both
12 study 05 and study 9A, did these trials have specific
13 features that might render them more or less than
14 normally persuasive. This really actually, JoAnn,
15 gives you an opportunity to highlight the issues that
16 are of concern or basically highlight the major
17 strengths or weaknesses of the trial that would lead
18 you to think that one should put more or less weight
19 on them.
20 DR. LINDENFELD: I think study 05 was a
21 reasonably good study with the one problem that we
22 discussed that question 2 would have addressed, this
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
317
1 drop-out rate. And I think there is some concern that
2 the drop-out rate here was among people who would have
3 been most likely to revert to atrial fibrillation. I
4 think the second study, 9A, isn't very persuasive, just
5 very small numbers. So I wouldn't put much weight on
6 that. I think that overall these studies are --
7 particularly 05 is a reasonably persuasive study. Once
8 again, I think the specific features of concern are the
9 very short follow-up, the drop-out rate, which we
10 discussed, and then I think also this just is not really
11 terribly representative of the population of patients
12 we would treat.
13 CHAIRMAN PACKER: Okay. Any -- what we
14 should do is find out how the committee feels. Bob,
15 let me just ask -- a distinction is made here because
16 the sponsor made the distinction between paroxysmal
17 atrial fib/flutter and a chronic atrial fib that has
18 been converted to normal sinus rhythm. I think I heard
19 Michael Cain earlier suggest that that distinction is
20 somewhat artificial and impractical. Can we -- if that
21 is the case, then the distinction between 3 and 4 is
22 totally artificial. Does the Agency believe that these
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
318
1 are distinct indications? Because some of us might feel
2 that they are not distinct.
3 DR. FENICHEL: I think the best solution
4 for those of you who do believe they are not distinct
5 is to cast the same vote on each question. I think we
6 do have products approved for one condition and not the
7 other. So at least at one time or another we and you
8 must have been convinced that they are distinct.
9 CHAIRMAN PACKER: Okay. Why don't we
10 rephrase -- I think we have all discussed the issues
11 related to 05 and 9A. I think that one could ask the
12 question 3(A) in the following manner. And that is, let
13 us ask whether one finds from 05 and 9A, can one conclude
14 from looking at these, given their strengths and
15 weaknesses, that there is an effect on sotalol on
16 recurrent paroxysmal atrial fib, and whether you would
17 consider that data to be persuasive for that indication.
18 Because that is really what we are asking here.
19 DR. THADANI: Are you going to take a vote
20 whether people on the board here believe paroxysmal is
21 different than chronic?
22 CHAIRMAN PACKER: No.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
319
1 DR. THADANI: People are obviously -- you
2 said they might be the same. But --
3 CHAIRMAN PACKER: All right, Bob?
4 DR. THADANI: Do you want us to vote on that
5 yes or no, just to know how many people on the panel
6 differ them? Yes? No?
7 CHAIRMAN PACKER: No.
8 DR. FENICHEL: First of all, I think it is
9 a religious discussion and it is not going to get
10 anywhere. Secondly, I think that to the extent it could
11 be made a rational discussion, I don't think it can be
12 made rational separate from the drug at hand. So one
13 might -- the committee members might vote that, oh yes,
14 they are the same as regards sotalol. But that might
15 give the wrong impression to those hearing the
16 discussion because the same committee might, with equal
17 rationale, believe that they were different when
18 considering the next drug that comes along for the same
19 general area of indication or indications. So I just
20 don't think it is a useful discussion.
21 DR. THADANI: But, Bob, they really are
22 different because the patient --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
320
1 DR. FENICHEL: Look, I don't think it is
2 a useful discussion.
3 DR. THADANI: But they are different.
4 Chronic a fib don't go into sinus rhythm by definition.
5 CHAIRMAN PACKER: Yes, they do because --
6 DR. THADANI: After cardioversion or
7 something.
8 CHAIRMAN PACKER: They are converted.
9 Anyway, I think the question that is before the committee
10 with 3A evolves into do you consider the data with
11 sotalol in paroxysmal atrial fibrillation to be
12 persuasive. You can -- that is really the question.
13 I don't think it is the identification of the issues
14 because we spent a lot of time identifying the issues.
15 I think the question is do you consider it persuasive.
16 You can use any criteria you want to answer that
17 question. And I guess if you wanted to, you could say
18 a little bit or a lot. I don't want to restrain the
19 range of responses. So, JoAnn, do you consider the data
20 supporting an effect of sotalol in patients with a
21 history of paroxysmal atrial fibrillation and flutter
22 to be persuasive based on studies 05 and 9A?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
321
1 DR. LINDENFELD: I think we often talk
2 about this as do we consider this similar to a single
3 reasonable trial, and I would say I consider this as
4 these two together. I don't really consider 9A. So I
5 think I would be -- I think this is as persuasive as
6 a single significant clinical trial.
7 CHAIRMAN PACKER: Can I suggest the
8 following? Because I think that the previous format
9 that we have used, which is equivalent to two trials,
10 between one and two trials, one trial, et cetera, there
11 are certain situations where that becomes particularly
12 useful. This is sort of different than that because
13 we sometimes use that scale when we have one very, very
14 big study and we are trying to gauge the level of
15 persuasiveness. Here we have got 8 studies of varying
16 degrees of issues. So I think one sort of has to look
17 at the totality of the data base. And so I think that
18 is why the usual scale is avoided here. Is that true,
19 Bob? And consequently I think -- you can still hedge
20 your bets. But rather than say one or two, do you
21 consider it to be persuasive to support an effect of
22 the drug?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
322
1 DR. LINDENFELD: I would consider it
2 normally persuasive.
3 CHAIRMAN PACKER: Normally -- I like that.
4 Normally persuasive. Lem?
5 DR. MOYÉ: I would consider it
6 unpersuasive. I think that study 05 is fatally flawed
7 because of the problem of drop-outs. And 9A is
8 essentially a subgroup analysis. So I would say no.
9 CHAIRMAN PACKER: Okay. Tom, we are going
10 to ask you to vote. Joan will not officially record
11 the vote. But I think it is important for everyone to
12 hear how you feel.
13 DR. BIGGER: Yes. I think they are
14 reasonably persuasive.
15 CHAIRMAN PACKER: Okay. Cindy?
16 DR. GRINES: I agree.
17 CHAIRMAN PACKER: Tom?
18 DR. GRABOYS: No, I don't think they are
19 persuasive.
20 DR. KONSTAM: I don't know how to quantify
21 it. I think they are a little bit less persuasive than
22 I would like them to be. I don't know how to quantify
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
323
1 that any more than that. And I think that the concern
2 that is driving Lem really all the way to say it is
3 useless doesn't drive me anywhere near that far, but
4 it does raise concerns in my mind.
5 CHAIRMAN PACKER: I guess here is the way
6 to judge the question. Assuming that the committee were
7 to vote that the drug were approvable, and that is a
8 much more globally comprehensive question, would you
9 include in the concept of what the drug was approvable
10 for those with a history of paroxysmal atrial
11 fibrillation? I mean, that is really the question.
12 DR. KONSTAM: Well, that is more
13 complicated. I mean, you know if -- I am not sure how
14 to answer that. I mean, if we really were looking at
15 these two things as different entities, then I would
16 have a lot of problems. Because I would be left with
17 this data set as the only evidence and I would not feel
18 that that in and of itself would be sufficient just
19 taking 05 as the single trial. So I guess I am going
20 to be stuck unless we say that these are sort of one
21 condition that take different forms.
22 CHAIRMAN PACKER: Okay, that is fair.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
324
1 DR. KONSTAM: And I am happy to do that.
2 So in that light, I think the answer would be I think
3 that the studies, particularly 05, is useful and would
4 move me -- would under those circumstances move me toward
5 an approvability.
6 CHAIRMAN PACKER: I understand. I
7 understand and the intent of the question is not to
8 create a black and white situation.
9 DR. KONSTAM: Right.
10 CHAIRMAN PACKER: The intent, in fact, is
11 to gauge it. And I think you have accurately portrayed
12 your feelings about it. Michael?
13 DR. CAIN: Again, I would lump them
14 together for the purpose that both groups of patients,
15 paroxysmal and chronic were in atrial fibrillation and
16 are now in normal sinus rhythm. And it separates that
17 from the chronic persistent, which means no matter what
18 you do, you cannot restore sinus rhythm. If you lump
19 them together, then I think 05 is normally persuasive
20 in the group of patients that have been studied and
21 discussed, with the footnote that there are still
22 several groups of patients that have not been included.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
325
1 CHAIRMAN PACKER: That actually is a
2 specific question later on. So we will highlight it at
3 that point in time. Ileana?
4 DR. PIÑA: I share Marv's concerns about
5 05. And with that caveat, I would say I am somewhat
6 persuaded.
7 CHAIRMAN PACKER: Udho?
8 DR. THADANI: I think 05 we discussed the
9 problems because of intent to treat versus drop-out
10 issues. And I think that is a concern, although the
11 data is going in the right way on the whole. And again,
12 the problem with the subgrouping in the other study.
13 So I am marginally persuaded, but the evidence is not
14 as strong as I would like to see. But there is a
15 suggestion it is going in the right track. So
16 marginally persuaded.
17 CHAIRMAN PACKER: Okay. I guess I agree
18 with what the -- the way the committee is actually
19 emerging is they feel there is evidence for an effect
20 on paroxysmal atrial fibrillation, which is less than
21 the kind they would like to see. But they believe that
22 in effect does exist. They would not like to conclude
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
326
1 an effect doesn't exist. And I think that Michael Cain
2 said it probably best by saying that the decision might
3 be easier if one were to consider paroxysmal and chronic
4 together as a combined data base, but that is for --
5 I think we will discuss that in just a few more minutes.
6 Bob, I think this will become more clear in just a few
7 minutes. JoAnn, did they identify a dosing regimen that
8 convincingly defers relapse into atrial fibrillation?
9 DR. LINDENFELD: I think 120 mg bid was the
10 minimum dose that convincingly does that.
11 CHAIRMAN PACKER: Okay. Does anyone --
12 would anyone suggest a different answer? Would anyone
13 suggest 80 mg bid? Would anyone suggest 160 mg bid?
14 Would anyone agree with 120 mg bid?
15 DR. KONSTAM: Yes.
16 CHAIRMAN PACKER: Okay. I just want to
17 make sure you are awake. Okay. JoAnn, did they
18 identify a regimen that convincingly alleviates
19 symptoms or reduces the incidence of stroke?
20 DR. LINDENFELD: Well, the answer to stroke
21 is no. I think symptoms -- the symptomatic recurrence
22 as an isolated symptom, yes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
327
1 CHAIRMAN PACKER: Okay. Would anyone
2 disagree?
3 DR. KONSTAM: Yes, I would.
4 CHAIRMAN PACKER: Marv?
5 DR. KONSTAM: You know, I mean in the sense
6 that I think -- I guess calling this symptomatic a fib
7 is a little bit different than saying -- I don't know
8 how to -- this isn't going to sound right. Saying that
9 it alleviated symptoms of a fib in the sense that I think
10 what the investigators were really detecting was
11 palpable a fib. In other words, patients who knew they
12 were in a fib. And I think that that is somewhat
13 different from saying that they were experiencing a
14 limiting symptom from the a fib. So I would -- I guess
15 I would question that this is clearly an effect on
16 symptoms per se.
17 DR. LINDENFELD: I totally agree with that.
18 This is just symptomatic atrial fibrillation but not
19 other symptoms.
20 CHAIRMAN PACKER: Okay. So the proposal
21 that has been put forward is that symptoms here -- that
22 there should be no claim for symptomatic relief. The
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
328
1 symptoms here are a marker for recurrence and
2 consequently evidence for a drug effect. Is that
3 correct? I just want to make sure what you are saying.
4 Okay.
5 DR. KONSTAM: I can accept that. I mean,
6 I don't really have a problem with the term symptomatic
7 a fib. I just think it has a little different meaning
8 than what we are usually looking for when we talk about
9 symptoms.
10 CHAIRMAN PACKER: Yes. I mean,
11 symptomatic a fib here is the name of a disease.
12 DR. KONSTAM: Right.
13 CHAIRMAN PACKER: As opposed to for the
14 reduction of symptoms in patients with atrial
15 fibrillation.
16 DR. KONSTAM: Right. A subtle point.
17 CHAIRMAN PACKER: We are saying -- and let
18 me make sure that everyone is in accordance -- that the
19 disease being treated here is symptomatic atrial fib.
20 That was the entry criteria. And that the reduction
21 of symptoms is evidence for a drug effect but not
22 evidence for a claim for reduction in symptoms. Is that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
329
1 fair?
2 DR. THADANI: Yes. I think the table we
3 saw, the overall symptom rate was not much different
4 -- the totality of the symptoms. So I think if we are
5 talking about symptomatic a fib, we should separate it
6 from the totality of the symptoms. Because I didn't
7 see any significant P values.
8 CHAIRMAN PACKER: I think you are agreeing,
9 Udho, right? You are agreeing. Okay. Let us move on
10 and address exactly the same questions, and then I am
11 going to take the liberty of trying to get a consensus
12 of 3 and 4 together. The studies under consideration
13 are those that randomize patients after being converted
14 from chronic AF. The studies are 004 and 014. And,
15 JoAnn, do you consider, considering all the weaknesses
16 and strengths of these studies, that the evidence this
17 drug has an effect in preventing or reducing the risk
18 or extending the time to recurrence in patients with
19 a history of chronic AF that have been cardioverted,
20 do you think that you consider the evidence that sotalol
21 has such an effect to be persuasive -- normally
22 persuasive?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
330
1 DR. LINDENFELD: Normally persuasive, yes.
2 CHAIRMAN PACKER: Normally persuasive.
3 And why don't we start at the other side for this
4 question. We went this way. Udho?
5 DR. THADANI: Yes, I think the 004 study
6 is pretty persuasive. I will go along with the vote.
7 CHAIRMAN PACKER: Ileana?
8 DR. PIÑA: I agree.
9 CHAIRMAN PACKER: Michael?
10 DR. CAIN: I agree.
11 CHAIRMAN PACKER: Marv?
12 DR. KONSTAM: Well, I guess the only thing
13 -- 004 is the study that was extended, right?
14 CHAIRMAN PACKER: Yes.
15 DR. KONSTAM: So I guess just in terms of
16 the answer to 4A -- I mean, there is a feature that raises
17 questions and I am not sure how much that should affect
18 things.
19 CHAIRMAN PACKER: That is why you get paid
20 big bucks.
21 DR. KONSTAM: Yes, right, big bucks. So
22 I guess I am in statistical limbo about this. Because
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
331
1 I have heard very different advice from different
2 statisticians about this point. And digesting all
3 that, I am going to still say that I am persuaded by
4 the study.
5 CHAIRMAN PACKER: Tom?
6 DR. GRABOYS: I agree.
7 CHAIRMAN PACKER: Okay. Cindy?
8 DR. GRINES: Agree.
9 CHAIRMAN PACKER: Tom?
10 DR. BIGGER: I agree.
11 CHAIRMAN PACKER: Lem?
12 DR. MOYÉ: I don't agree. I think 014 has
13 the same lethal flaw that 05 has. And I think 004 is
14 much too small and doesn't have the -- it is not as all
15 inclusive of important demographic subgroups as it
16 should be to be persuasive. So my answer is no.
17 CHAIRMAN PACKER: The lethal flaw you
18 identify in 004 is not the informative censoring, since
19 that wasn't an issue. I just want to make sure that
20 we know what lethal flaw in 004 you are referring to.
21
22 DR. MOYÉ: No, 014 had the lethal flaw.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
332
1 CHAIRMAN PACKER: I am sorry.
2 DR. MOYÉ: 014.
3 CHAIRMAN PACKER: Does 004 have a lethal
4 flaw?
5 DR. MOYÉ: No. 004 -- the problem I have
6 with 004 is that it is small and that it does not include
7 -- it is not all-inclusive of the demographic subgroups
8 I would like to see in order to base a decision. That
9 is going to affect many communities.
10 CHAIRMAN PACKER: Okay. My vote is that
11 it is persuasive. JoAnn, do you want to identify what
12 dosing regimen you believe has been shown to have an
13 effect which is being discussed in question 4?
14 DR. LINDENFELD: Well, I think this was a
15 dosing regimen, I believe, of 80 bid up-titrated to 160
16 bid. And so I think that would have to be the
17 recommendation. It is hard to identify a single dose
18 out of that.
19 CHAIRMAN PACKER: I am sorry?
20 DR. LINDENFELD: This was a regimen that
21 started off at 80 bid going up to -- wasn't that correct,
22 160 bid?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
333
1 CHAIRMAN PACKER: Yes.
2 DR. LINDENFELD: So there was a single
3 regimen that was tested rather than a specific dose.
4 CHAIRMAN PACKER: All right. A dosing
5 strategy that was tested. So I guess we would need to
6 be empiric here and say it was the dosing strategy that
7 was evaluated in the trial.
8 DR. LINDENFELD: Right.
9 CHAIRMAN PACKER: Okay. And I assume that
10 the answer that everyone has to 4C is identical to the
11 answer they provided for 3C. Does anyone disagree?
12 Okay. Now let me --
13 DR. THADANI: You know on the stroke issue
14 -- because one other concern one has is I don't think
15 everybody was on anticoagulants. When I was reading
16 it, some 40 percent or 50 percent of the patients were.
17 So I don't think you can address the issue at all not
18 knowing the details of how many -- you know the sample
19 size is small.
20 CHAIRMAN PACKER: No. JoAnn said to
21 question 3C that there were no data whatsoever on stroke.
22 DR. THADANI: Oh, okay. So the same
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
334
1 applies.
2 CHAIRMAN PACKER: And in fact the response
3 of the committee to 3C was there was no data that showed
4 it alleviated symptoms. What it did was prevented a
5 disease called symptomatic atrial fibrillation.
6 DR. THADANI: Okay.
7 CHAIRMAN PACKER: Okay. I think we
8 clearly enunciated that principle in 3C, and I think
9 we are reiterating that principle in 4C. Okay. Let
10 me just -- I want to get two more questions here which
11 are to be inserted between 4 and 5. To what degree,
12 if any, are your opinions on 3 and 4 influenced by the
13 results of dofetilide 345? It is not asked. But I think
14 the Agency probably would like to know because there
15 is an operational issue which is important here. The
16 question 3 and question 4 did not ask you to consider
17 dofetilide 345. So, okay, you answered 3 and you
18 answered 4. Now the question is do you need dofetilide
19 345 to get to where you want to go based on your answers
20 to 3 and 4, and the answer could be it doesn't matter
21 or it helped a little or it helped a lot. JoAnn?
22 DR. LINDENFELD: It helped a little.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
335
1 CHAIRMAN PACKER: Okay. Udho?
2 DR. THADANI: I think it helped a little
3 because I am still not convinced that we know the true
4 incidence of torsade in the absence of Holter
5 monitoring. I think we got a very in-random sequencing
6 of the data base. I am not sure I can be very -- you
7 know, the incidence of torsade might have been
8 underestimated.
9 CHAIRMAN PACKER: Are you sure you are
10 answering the question?
11 DR. THADANI: Well, the question is are you
12 sure with the --
13 CHAIRMAN PACKER: No, no. This is an
14 efficacy issue on atrial fibrillation and the question
15 is did 345 influence you in a material manner. And the
16 possibilities are no, a little, or a lot. The efficacy.
17 There is no torsade issue here.
18 DR. THADANI: I think a little.
19 CHAIRMAN PACKER: Ileana?
20 DR. PIÑA: It helped me very little simply
21 because it was going in the right direction.
22 CHAIRMAN PACKER: Michael?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
336
1 DR. CAIN: No real help.
2 CHAIRMAN PACKER: Marv?
3 DR. KONSTAM: Yes, I am -- I would have
4 answered these questions identically had I not seen 345.
5 So in that light, I will say not at all.
6 CHAIRMAN PACKER: Tom?
7 DR. GRABOYS: Not at all.
8 CHAIRMAN PACKER: Cindy?
9 DR. GRINES: It helped a little.
10 CHAIRMAN PACKER: Tom?
11 DR. BIGGER: Yes, it helped a little
12 because a small dose had a definite efficacy signal.
13 CHAIRMAN PACKER: Now, Lem, you can
14 actually use this as an opportunity to change your vote
15 because the real question being asked is to what degree
16 are you looking at 345 in a meaningful fashion. I think
17 you have already said for questions 3 and 4 that you
18 are not persuaded. So you could say that if you included
19 345, you would be persuaded.
20 DR. MOYÉ: I am appreciative that at this
21 last meeting I can attend that the chairman gives me
22 an opportunity to change my mind. I think I will
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
337
1 decline.
2 CHAIRMAN PACKER: Okay. All right, you
3 can't say I didn't make the offer. And personally I
4 think that Marv has said it the way I would feel about
5 it, which is that I would have voted the same way without
6 345. But in all honesty, it probably helped a little.
7 Okay. Let me see if I can clarify one other issue before
8 going on to 5. One concern that I have
9 -- this is to Bob Fenichel -- is that physicians might
10 get the impression that sotalol is a treatment for
11 chronic atrial fibrillation in patients who remain in
12 chronic atrial fibrillation. I have -- I must say I
13 have a major concern about that. And in fact this is
14 a treatment for patients in normal sinus rhythm. And
15 the proposed wording that the sponsor has in my view
16 does not make that clear, which is why I tend to favor
17 Michael Cain's suggestion that the distinction between
18 paroxysmal and chronic atrial fibrillation here is more
19 of the history of the patient as opposed to the state
20 that the patient is in at the initiation of therapy or
21 the intent of therapy. The state that the patient is
22 in is normal sinus rhythm at the initiation of therapy
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
338
1 and the intent of therapy is to prevent recurrence.
2 And the only difference between 3 and 4 is whether their
3 previous history of atrial fibrillation was
4 intermittent or continuous. And because of that, my
5 sense is that what we are really talking about is that
6 this is a treatment for patients in normal sinus rhythm
7 with a -- and I hate to say this -- either a history
8 or recent history of atrial fibrillation or flutter as
9 opposed to paroxysmal or chronic. And I just wanted
10 to make sure that I got a sense of the committee that
11 Michael Cain's view on this, which is that the
12 distinction here is somewhat artificial, and therefore
13 the two data bases can be viewed as being mutually
14 supportive, is the consensus view here. Because I
15 understand, Bob, there is a history of making these
16 distinctions. But the intent in fact is to treat
17 patients who are in normal sinus rhythm. The risk I
18 see here is that if the wording of the indication
19 includes for the treatment of chronic atrial
20 fibrillation, that patients who are in atrial
21 fibrillation will get this drug to either convert them
22 or for some unbelievably undefined goal, which I think
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
339
1 is a significant risk.
2 DR. FENICHEL: Well, we know certainly that
3 an awful lot of the quinidine that is used is used in
4 patients who have been in atrial fibrillation for
5 decades and physicians say when asked, oh yes, well
6 quinidine is used for atrial fibrillation. So, of
7 course, the risk you are alluding to is a serious one.
8 And in my fantasy, we get the labeling wording right
9 and that solves problems. But that is a different
10 world. So I think we are going to work on getting the
11 labeling right to send the message that indeed this is
12 for people who are now in sinus rhythm but who have
13 histories of one thing or another. I think there is
14 -- well, the reason that this distinction has been made
15 in the past is not that it is to be given either -- that
16 other drugs are to be given either to people in chronic
17 fibrillation or paroxysmal fibrillation, but rather
18 that drugs have tried to demonstrate efficacy in
19 patients of both kinds and have succeeded in only one.
20 So what are you going to do? Well, you've plainly got
21 to label it for the one.
22 CHAIRMAN PACKER: I think the key
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
340
1 operational distinction here, and I will just try to
2 reach a contrast with dofetilide, is that if I remember
3 correctly, the sponsor of dofetilide was actually
4 seeking a claim and provided data that in the doses that
5 they were recommending that there was -- that they would
6 use the drug for conversion.
7 DR. FENICHEL: That is correct. We have
8 pharmacological conversion labeling for -- well, for
9 ibutolide and for quinidine certainly and submitted for
10 dofetilide.
11 CHAIRMAN PACKER: No such claim is being
12 requested here. And in the only study to evaluate it,
13 the doses used are outside the recommended range. So
14 all of this, I think, underscores the concept that this
15 is a drug to be given in patients in normal sinus rhythm
16 for the prevention of recurrence. This is not a
17 treatment for atrial fibrillation. This is in somewhat
18 contrast to dofetilide that actually had presented the
19 committee a conversion data base with the intent that
20 the drug could be used for a conversion at the same doses
21 that it was used for the prevention of recurrence. I
22 think the distinction here is important. Tom?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
341
1 DR. GRABOYS: Yes. See what you have done
2 though is you have raised the issue, the fundamental
3 issue, which is the translation from what goes on in
4 here or what goes on in the office when Dr. Kowey is
5 managing a patient in the umpteenth degree perfectly
6 is different than translating it into the community.
7 And the concern that we have is being this is not going
8 to be used for an indication as a life-saving event where
9 you are willing to accept some risk. We are back to
10 the same fundamental issue of how the physician in the
11 community is going to be dealing with this. And the
12 fact that I frankly don't trust the physician in the
13 community in terms of managing these patients,
14 regardless of how much so-called educational material
15 they are going to try to give that physician.
16 DR. THADANI: Milton, also I think perhaps
17 one of the issues could be the wording could be
18 completely changed. Because the trials we have seen
19 from my point of view would be to delay rather than use
20 the word prevent. I think prevent is the wrong term
21 here. We should use the word delaying the reversion
22 to atrial fibrillation after a patient in a fib has been
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
342
1 converted to sinus rhythm.
2 CHAIRMAN PACKER: The sponsor has already
3 incorporated that concept in their proposed labeling.
4 DR. THADANI: So that might be easier.
5 That means you have to convert the patient into sinus
6 first before you start the drug. And then all you are
7 doing is claiming a delay of reversion into a fib.
8 CHAIRMAN PACKER: I think that the sponsor
9 has already incorporated that. I think the committee
10 is in favor of the emphasis on the use of this drug in
11 normal sinus rhythm. And, Tom, I wish we could address
12 your concern in a meaningful fashion. But we probably
13 -- unless you can come up with a specific suggestion,
14 that dissociation is commonplace, not only -- and does
15 not only apply to this drug. And I don't know how to
16 fix the problem.
17 DR. PIÑA: I have a question.
18 DR. GRABOYS: We have got the only other
19 scenario --
20 DR. PIÑA: I am sorry, I have a question
21 of the sponsor. You have a lot of numbers of patients
22 already being treated for atrial fibrillation with this
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
343
1 drug. Do you have any in-house data as to how many of
2 those patients are being given the drug for conversion
3 versus keeping them in sinus rhythm? I mean, you showed
4 a large -- a very large use of this drug in the community.
5
6 DR. KOWEY: As a point of fact, Ileana, they
7 do not have any data. All I can tell you is that in
8 clinical practice, this drug is rarely used for
9 conversion to sinus rhythm. And the reason is because
10 it rarely works. At the doses that we are recommending,
11 it just simply doesn't have enough of an effect. Where
12 it is used, however, which is a little bit different
13 than what you are talking about and which has not been
14 studied in all honesty, is it is used in a fashion where
15 the patient is loaded with the medication in the hospital
16 in atrial fibrillation and then are cardioverted on the
17 drug to sinus rhythm. That has not been studied in the
18 clinical trials we have presented today. But that
19 actually is probably
20 -- if you are talking about the way doctors use the drug
21 which is not what you are talking about, that is the
22 way doctors use the drug, the way you are not talking
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
344
1 about.
2 CHAIRMAN PACKER: Tom, I am wondering if
3 we can address your concerns in part by suggesting that
4 whatever approved labeling -- and I don't want to
5 wordsmith this too much, but I think that some guidance
6 here is appropriate. Maybe the indication section
7 should say this drug is not a treatment for atrial
8 fibrillation and should not be given to patients with
9 atrial fibrillation. That is probably too strong. I
10 am trying to figure out how to phrase it.
11 DR. THADANI: You could say unless they
12 have been converted to normal sinus rhythm. Because you
13 are not going to use it for a fib. And also I think
14 you could go further to allay Tom's fear. If the patient
15 has a first recurrence, the drug should be stopped.
16 Because that is how the studies were done. If the
17 patient had a recurrence, the patient was withdrawn.
18 And that is not an intent to treat. So you could say,
19 okay, you start a drug if the patient is in sinus rhythm
20 and paroxysmal and continue it. When the first
21 recurrence occurs, the patient is out of the drug or
22 if the patient is in chronic a fib, you cardiovert him
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
345
1 and put him on the drug and if he has a relapse, the
2 drug should be stopped. Because if there is a relapse,
3 why do you want to continue it? And I think that would
4 be an important thing. I know we have not done it, but
5 that would be one way to avoid indiscriminate use of
6 a drug which may be questionable to use once the patient
7 is in a fib other than controlling the rate.
8 CHAIRMAN PACKER: Tom, you raised the
9 concern and I think we all share your concern. Does
10 the proposal that I just made go somewhat to addressing
11 your concern or would you modify it in any way?
12 DR. GRABOYS: No. I am going to have
13 difficulty approving this drug based on the points that
14 I have already raised.
15 CHAIRMAN PACKER: Are you trying to say --
16 DR. GRABOYS: The fact is that this
17 population will all end up in atrial fibrillation. All
18 you are doing is delaying it, as has already been
19 mentioned, or deferring it for a few months. And you
20 are deferring it under the concept that somehow I am
21 going to feel better for those few months. But
22 ultimately they are going to end up in AF and AF will
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
346
1 be the rhythm of choice and you are going to control
2 their rate and they are going to feel better.
3 CHAIRMAN PACKER: Okay. Why don't we do
4 this. The appropriate time to discuss this at length
5 would be after question 8. So let's go on to question
6 5. JoAnn, what do you think is the likely incidence
7 of QT prolongation and torsade in various populations
8 if the patients are treated with sotalol using the dosing
9 regimens suggested by the clinical trials?
10 DR. LINDENFELD: This is a broad question
11 in various populations, but I think that if the drug
12 is used as specified in these trials and these doses
13 in these patients, the incidence will be low, probably
14 under 1 percent. I just think several people have made
15 the point that this is not necessarily the population
16 or the doses in which it will be used or the doses for
17 creatinine clearance.
18 CHAIRMAN PACKER: Okay. A question for
19 Bob Fenichel, is that a good enough approximation?
20 DR. FENICHEL: Yes.
21 CHAIRMAN PACKER: Okay.
22 DR. KONSTAM: Can I chime in, Milt?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
347
1 CHAIRMAN PACKER: All right.
2 DR. KONSTAM: I mean, again, I just don't
3 know the answer to this question. And the thing is that
4 we can't answer this question without reference to the
5 time frame of our observation in this population. So
6 I am not sure what the median time of exposure that we
7 have in terms of our data set is, but it is relatively
8 short. It is measured in weeks. And so I think if you
9 were going to ask the question of what is the incidence
10 of torsade -- all the caveats that you mentioned, JoAnn,
11 I fully agree with. But I would just add to that the
12 time frame in that we don't know what it would be in
13 a year or two years.
14 DR. THADANI: But surely the incidence of
15 QT is not small. QT interval prolongation is dose
16 response related.
17 DR. KONSTAM: Oh yes, QT prolongation is
18 a lot more than 1 percent.
19 DR. THADANI: So I think there are two
20 separate questions here. The incidence of torsade is
21 small, but the incidence of QT is proportional to the
22 dose. So if you look at the 05 study or even 04 when
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
348
1 you dose titrated, you are saying average QT
2 prolongation is 21 or 22 seconds. That is the mean
3 value, so there are a lot of patients falling outside.
4 So I think the QT prolongation is uniform. That is
5 what the drug does. While the incidence of torsade is
6 small. So I think my reading is there is a dissociation
7 between the two phenomenon. And I am also more puzzled
8 now the more I hear about QT. Women show less
9 prolongation and have a higher incidence of QT in
10 general. So I am more confused than ever what the real
11 relationship is.
12 CHAIRMAN PACKER: Ileana?
13 DR. PIÑA: Yes. I think that more than
14 dose, it is probably serum concentration which varies
15 according to renal clearance. So that the higher --
16 I mean the lower the renal clearance, the longer the
17 QT. But that is part of the drug effect as well. So
18 you are going to live with it if you approve the drug
19 this way. And I think we just have to be cognizant of
20 the fact. But I still think that the rate of torsade,
21 at least if dosed appropriately as recommended,
22 continues to be small.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
349
1 DR. THADANI: Also, I think since the
2 question also addresses population, there is some
3 discomfort. We don't have a large sample size in patients
4 with a diminished LV function. In the first study, the
5 disclaimer was 60 -- a creatinine clearance of 60. In
6 the second study, it was 40 to 60. But the sample size
7 was so small in all that we are using or the ones that
8 we are dosing. I think we would like to really see a
9 bit larger sample size to be sure than in these patients,
10 even in the once daily dosing. Say if you used 160,
11 you might end up having more prolongation of the QT than
12 one could be reassured from this smaller data base.
13 CHAIRMAN PACKER: Okay. I think what we
14 have right now is -- just reading everyone's comments,
15 the following conclusion. That the incidence of QT
16 prolongation is dose dependant and has, in fact, been
17 described and quantified in the existing trials. And
18 I think what has been added in general is that of course
19 the incidence of torsade has been defined only in the
20 patients who were treated and has not been defined in
21 the patients who were not treated. So the incidence,
22 I guess, JoAnn, of less than 1 percent applies to the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
350
1 patient populations that were adequately represented
2 in the clinical trials. Does anyone disagree with that
3 conclusion? Okay. Does sotalol cause significant
4 side effects other than QT prolongation and torsade?
5 DR. LINDENFELD: Yes I think it causes
6 side effects we would expect from a beta blocker --
7 bradycardia, fatigue, exacerbation of heart failure.
8 CHAIRMAN PACKER: I don't think anyone
9 disagrees. Any other side effects anyone believes
10 should proceed?
11 DR. THADANI: What about the age group?
12 You know, there was some concern about dizziness in
13 people who were above 65. I think we should mention
14 that because again the sample size might be small and
15 Tom brought up patients who are 70 and 75. And you don't
16 want them to have syncopal attacks or something or
17 whatever. So I think probably we need more information
18 on that.
19 CHAIRMAN PACKER: Okay. Let me just -- for
20 the record, Dr. Califf had to catch a plane, but he voted
21 yes on questions 3 and 4 for persuasiveness and also
22 agreed with the committee on the dosing issues and on
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
351
1 the symptom and stroke issues. Any other points on 6?
2
3 DR. BIGGER: Well, just the bradycardia can
4 be very severe and cause hypotension and even death.
5 So it shouldn't be -- it should get a little bit of
6 a highlight.
7 DR. THADANI: Milton, on that question I
8 think also probably raise the issue. Because I was
9 surprised in these trials that Digoxin was allowed and
10 most other drugs as background therapy. We know that
11 especially in patients with a fib, if they are dig-plus
12 another beta blockers, sometimes -- especially when they
13 are in a fib, the rate really goes slow. And I have seen
14 pauses of three to four seconds, especially with the
15 two combinations. From the data base, since the trial
16 was done on background Digoxin on most of the patients,
17 are we going to recommend that Dig must be used or what?
18 Because I know it won't come up in the questions until
19 the bradycardia is used. Because none of the trials
20 -- I think they wanted the Dig background. And the
21 problem is they rated this background so that the
22 patients don't have too many symptoms. I don't know
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
352
1 why. Because I was a bit puzzled. I know Dig controls
2 the rate in some patients, but not in all. So would
3 you have to say this drug should be only used in patients
4 who despite Dig remain symptomatic?
5 CHAIRMAN PACKER: That is question 9, Udho.
6 DR. THADANI: But in the bradycardia issue,
7 can you dissociate the two?
8 CHAIRMAN PACKER: We will bring it up in
9 9.
10 DR. THADANI: Okay.
11 CHAIRMAN PACKER: Okay. Number 7, do
12 there appear to be differences in safety and efficacy
13 between d,l-sotalol and available therapies. I think
14 the Agency -- the division emphasizes that it may be
15 hard to make this assessment because there are no direct
16 comparative studies or there are very few comparative
17 studies. It is still relevant to ask the question
18 whether the risk/benefit relationship for this drug
19 differs materially from what one might think or one might
20 deduce would be the risk/benefit relationship for other
21 drugs. I assume that that comparison, Bob, is to be
22 made for drugs that are approved for the indication.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
353
1 DR. FENICHEL: Well, I think it would be
2 problematic if there were a drug which -- well, I don't
3 think it should be limited to stuff approved for the
4 indication.
5 CHAIRMAN PACKER: Okay. That is fine.
6 DR. FENICHEL: Well, I think people might
7 think, well, there is another drug even in the pipeline
8 that the rest of us haven't heard about. And I can
9 imagine that people might think, oh well, there is this
10 secret drug which is so much better and it would be a
11 shame to have this one out. Well, that is a respectable
12 point of view. Or people might think they know
13 something about dofetilide, even though of course it
14 is not approved but it was discussed at this meeting
15 just a couple of months ago. And so I don't think people
16 should limit themselves to approved therapy.
17 CHAIRMAN PACKER: Okay. JoAnn, 7A? From
18 what you can deduce, do you think that sotalol is
19 markedly more or less effective than other treatments?
20 The word available here is to be converted to the word
21 other.
22 DR. LINDENFELD: Other. Given
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
354
1 everything, I think that sotalol is equivalent to other
2 available therapy.
3 CHAIRMAN PACKER: Does anyone disagree?
4 Okay. 7B. Does sotalol appear to be more or less
5 proarrhythmic than other therapy?
6 DR. LINDENFELD: Compared to other drugs
7 that cause torsade, I think sotalol appears --
8 CHAIRMAN PACKER: Other drugs for the
9 treatment -- for the prevention of recurrence of atrial
10 fibrillation. This is not a treatment for atrial
11 fibrillation.
12 DR. LINDENFELD: I think the rate of
13 torsade is equivalent or the same.
14 CHAIRMAN PACKER: Okay.
15 DR. GRABOYS: Is that what you are focusing
16 -- is that the definition you are looking at as
17 proarrhythmic or torsade rather?
18 DR. LINDENFELD: Yes, I think torsade is
19 the same if you want to include bradycardias as
20 proarrhythmias.
21 DR. GRABOYS: You are saying with regard
22 to available therapy.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
355
1 DR. LINDENFELD: I guess I was counting --
2 DR. GRABOYS: You are not saying available
3 therapy of beta blockers and calcium channel drugs.
4 DR. THADANI: Quinidine and --
5 CHAIRMAN PACKER: Yes, I think --
6 DR. GRABOYS: That needs to be clarified.
7 DR. LINDENFELD: We should clarify that,
8 yes.
9 DR. KONSTAM: I think that is the critical
10 thing. And this is where the mechanism comes in.
11 Because what is this drug and how is it working and what
12 is it doing? And we believe that a significant portion
13 of the drug's effect is beta blockade. We are not
14 exactly sure how much. But I am not sure we know for
15 sure that it is anything other than a beta blocker.
16 And so obviously I think if we're comparing it to another
17 beta blocker, it is far more proarrhythmic. So that
18 is where the quandary comes in.
19 DR. THADANI: But surely here we are not
20 asking about indications to control the rate, right?
21 A beta blocker --
22 DR. KONSTAM: Yes, but Udho -- I guess we
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
356
1 don't know really how much of the effect in terms of
2 preventing recurring symptomatic a fib is an influence
3 on heart rate when the patient goes into a fib.
4 DR. THADANI: I think perhaps would it be
5 reasonable to insert something by that that the rate
6 has to be controlled and other drugs should be used?
7 Because beta blocker has been approved to control the
8 rate in a fib, right?
9 DR. LINDENFELD: I think if we compare it
10 to drugs that we would use to prevent recurrent atrial
11 fibrillation, and given the drugs we would use to prevent
12 it, I would consider the risk of torsade to be
13 equivalent.
14 CHAIRMAN PACKER: Yes. I don't think we
15 have to make this too complicated. I think that
16 compared to drugs that block the AV node, this would
17 be more proarrhythmic when compared to the drugs that
18 we use to prevent recurrence of atrial fibrillation,
19 which I think is what is being asked here.
20 DR. KONSTAM: But beta blockers, I am sure,
21 prevent recurrence of symptomatic atrial fibrillation.
22 I don't know if it has ever been studied quite that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
357
1 way.
2 DR. THADANI: But they are not approved,
3 though, for that.
4 DR. KONSTAM: I understand.
5 CHAIRMAN PACKER: I don't think we can get
6 from where we are to -- you understand the problem.
7 We don't have any data base.
8 DR. KONSTAM: Right. Right. I also want
9 to just -- can I just add that even with reference to
10 this issue in terms of other drugs that prevent atrial
11 fibrillation recurrence per se, I concur with your
12 thought although I just don't feel that we have the data.
13 DR. LINDENFELD: No.
14 DR. KONSTAM: Even to say that.
15 DR. CAIN: And I think the only other
16 caveat is that if you take propathenone and flecainide,
17 they are indicated and used, I think, because of CAST,
18 hopefully exclusively in people without structural
19 heart disease, where as at least sotalol here there is
20 some people who have had structural heart disease who
21 have received the drug. So I think that comparison
22 between propathenone and flecainide in people without
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
358
1 structural heart disease and sotalol without structural
2 heart disease versus with structural heart disease, we
3 don't have the full story on that.
4 DR. THADANI: Milton, available means
5 approved?
6 CHAIRMAN PACKER: No. We went through
7 this already.
8 DR. THADANI: We said approved, right?
9 CHAIRMAN PACKER: No.
10 DR. THADANI: Because then I think I want
11 to make a little --
12 CHAIRMAN PACKER: The operative word here
13 is other.
14 DR. THADANI: Yes, I think I want to make
15 one comment. The incidence probably is higher than what
16 is reported with amiodarone. Because available therapy
17 -- they are using a lot of amio for prevent of recurrence
18 of a fib. And if I look at the literature data base,
19 the relapse rate of maintaining sinus is 60 or 70
20 percent. Again, not an approved indication. But amio
21 is the only drug which doesn't cause torsade. Because
22 we have used amio in patients who have had torsade on
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
359
1 type 1A and other drugs. So if you are putting a --
2 when JoAnn said it is the same as others, the only
3 exception I will put probably is amiodarone. I don't
4 know if the committee members would agree or not. But
5 that is at least my experience from the literature data.
6 CHAIRMAN PACKER: Tom and Marv?
7 DR. KONSTAM: No, I just think Udho made
8 a good point.
9 CHAIRMAN PACKER: Okay. Let me just make
10 sure that I understand. The consensus here is that
11 maybe with the possible exception of amiodarone and with
12 the possible exception of beta blockers if they were
13 to work for this condition, that d,l-sotalol is not any
14 better or worse compared to other drugs for prevention
15 of recurrence in terms of its proarrhythmic effects.
16 Is that what the committee feels? I just want to ask
17 one question. Just so that I understand. The dose that
18 appeared to be a reasonably effective dose here was 120.
19 If you -- I didn't see a lot of proarrhythmias at 120.
20 Am I missing something? I mean, I would almost be
21 tempted to think about the possibility that it looks
22 the same as others if you get up to 160 bid or higher.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
360
1 But is there a dose that is effective here that is less
2 proarrhythmic, or is that a conclusion which absolutely
3 just can't be reached from the available data? Cindy?
4 DR. GRINES: I agree that it looks like the
5 torsade is dose-related in that I was struck also by
6 the low incidence of torsade in the proposed dosing.
7 But I think there are just so few patients that we have
8 to conclude that it is probably equivalent.
9 DR. CAIN: And I think, Milt, the other
10 thing is the type of patient. We are excluding people
11 with large infarcts and congestive heart failure, which
12 are the ones that --
13 CHAIRMAN PACKER: Right. And those were
14 not excluded from the dofetilide data base. Okay. So
15 is everyone comfortable with comparable with plus or
16 minus amiodarone and plus or minus beta blockade? Okay?
17 You've got it. All right.
18 DR. THADANI: Plus or minus heart rating
19 lowering calcium blockers. Like we still use --
20 CHAIRMAN PACKER: That is a different
21 indication.
22 DR. THADANI: No, no. The beta blocker is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
361
1 the same to control the rate. A different indication,
2 yes.
3 CHAIRMAN PACKER: A different indication.
4 Okay. Is -- are there non-proarrhythmic side effects
5 that are more or less prominent with this drug than with
6 other drugs that would be used to treat the same
7 condition? JoAnn?
8 DR. LINDENFELD: I think there are some
9 non-proarrhythmic side effects that are more common.
10 But I think if you take all non-proarrhythmic side
11 effects that this drug has a reasonable side effect
12 profile.
13 CHAIRMAN PACKER: Comparable?
14 DR. LINDENFELD: Comparable, yes.
15 CHAIRMAN PACKER: Comparable?
16 DR. KONSTAM: I mean the bradycardia. Tom
17 Bigger made the point. I mean the bradycardia is more
18 than certain other drugs that are available.
19 DR. LINDENFELD: But other drugs have, for
20 instance, more diarrhea and more --
21 DR. THADANI: Less GI side effects than
22 quinidine, I guess.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
362
1 CHAIRMAN PACKER: I guess the question is
2 is the non-proarrhythmic side effect of this drug so
3 clearly distinguishable from others that you would use
4 it as a factor to sway your opinion as to whether this
5 drug should be made available? I think the answer or
6 the sense that I got from your response, JoAnn, is no.
7 What you gain with one, you lose with another?
8 DR. LINDENFELD: Correct.
9 DR. THADANI: If the patient was
10 bradycardic, they were excluded. So if the heart rate
11 is already 50, you are not going to put those patients
12 on the drug. You know, the drugs which don't lower the
13 heart rate, they could be put on it. So I think we will
14 have to absolutely make sure that if you've already got
15 a bradycardia that you probably -- that would be an
16 exclusion. So that would be a different issue to be
17 considered on starting the therapy to start with.
18 CHAIRMAN PACKER: Okay. Tom?
19 DR. BIGGER: It has got less organ toxicity
20 than some of the drugs that are used for conversion and
21 delay of recurrence. For example, you don't see
22 thrombocytopenia and you don't see glucocyte reaction
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
363
1 and things of that sort. That is in its favor.
2 DR. LINDENFELD: Another advantage is no
3 Dig --
4 CHAIRMAN PACKER: No Dig interaction.
5 DR. LINDENFELD: No cytochrome problems.
6 You just have to really watch renal function here.
7 CHAIRMAN PACKER: Less drug interactions,
8 more bradycardia. I think the Division gets the idea.
9 Okay. Number 8, should sotalol be approved to reduce
10 the frequency of relapse of atrial fibrillation in
11 patients in normal sinus rhythm with a history of atrial
12 fibrillation? I think that that is sort of the concept
13 that we were discussing before. And, JoAnn, why don't
14 we get your answer.
15 DR. LINDENFELD: Yes, I think given all of
16 the data that we have seen, I feel reasonably comfortable
17 lumping these two groups of patients together. But I
18 would say that it should be approved to delay the onset
19 of atrial fibrillation.
20 CHAIRMAN PACKER: Discussion?
21 DR. FENICHEL: Wait, you don't have to
22 anguish about the distinction between chronic and
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
364
1 paroxysmal. I just want to reassure members of the
2 committee that if you think it should be approved for
3 anything, say yes. Obviously if you say it should be
4 approved for neither indication, assuming you think that
5 they are two independent indications, then that takes
6 care of it. But as long as you say yes for any, there
7 is a question down the way, 9B or something, that says
8 do you want to make that distinction. So don't make
9 it now. Carry on.
10 DR. LINDENFELD: Yes.
11 CHAIRMAN PACKER: Okay. Discussion? We
12 always have a discussion for a critical issue of approval
13 or non-approval. No one -- does anyone want to discuss?
14 No one wants to --
15 DR. KONSTAM: Well, I will just say that
16 here is -- I think we wind up being influenced by
17 cost/benefit ratio as opposed to just the pure question
18 of efficacy and the pure question of safety. Here is
19 where we have got to put it all together. And I think
20 to me -- I think that is why I just say that I am going
21 to wind up being strongly influenced by my response to
22 7B and where does it really fall and what am I really
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
365
1 comparing it to. So that is my discussion.
2 CHAIRMAN PACKER: Okay. Let me just
3 emphasize the point we always emphasize at this type
4 of question, which is you need not modify your answer
5 or a restriction to a specific indication or a
6 subpopulation or a requirement for post-marketing or
7 anything. All of that is in question 9. So the answer
8 to 8 should be if you can think of any restricted or
9 unrestricted, modified or unmodified reason for
10 approval, the answer to 8 should be yes. And then you
11 should clarify what your concerns and limitations should
12 be in question 9. Okay? So you might think that sotalol
13 could be approved for one person and 8 would be yes and
14 9 would be for one person. Okay? Let's start with
15 JoAnn. I am sorry, JoAnn, you did say yes.
16 DR. LINDENFELD: Yes.
17 CHAIRMAN PACKER: Okay. Lem, we will
18 start with -- we will go down with you.
19 DR. MOYÉ: I would say no. And just very
20 briefly, I think we have to proceed very gingerly and
21 certainly in the case of an antiarrhythmic therapy.
22 These drugs have been shown to have such dangerous
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
366
1 stingers in their tails. We need to really have a very
2 solid data base from which to draw conclusions. They
3 are often used in patients for whom they were not
4 initially studied. They are used in very fragile
5 communities. We need to have some assurances that in
6 fact this drug is going to be safe and effective in those
7 communities. And with all the discussions today, I don't
8 think we have that information. So my answer is no.
9 CHAIRMAN PACKER: Tom, we will ask you to
10 vote, but your vote will not count here.
11 DR. BIGGER: Well, I am not voting. I am
12 just making a comment. I think it should be approved.
13 It is a little like getting married after your children
14 are in college. It has been used for many years for
15 this indication and much more broadly than the
16 indication the sponsor is asking for. Considering what
17 else is available and becoming available and how we are
18 suggesting it should be used, I think it would be
19 appropriate to approve it.
20 CHAIRMAN PACKER: Getting married after
21 your children are in college, huh? Tom, I really have
22 to think about that. Cindy?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
367
1 DR. GRINES: I agree with Tom.
2 CHAIRMAN PACKER: About children in
3 college?
4 DR. GRINES: No. I see it very commonly
5 used for this indication and I believe it should be
6 approved.
7 CHAIRMAN PACKER: Tom?
8 DR. GRABOYS: For all the reasons I have
9 mentioned already, I don't think it should be approved.
10 CHAIRMAN PACKER: Okay. Could you just
11 clarify those reasons again? Because this is the
12 appropriate time to do it.
13 DR. GRABOYS: The concern is you are using
14 a drug that is potentially proarrhythmic that is
15 non-proarrhythmic for an indication that is not to
16 prolong life or prevent sudden death. It is an
17 indication for "quality of life" for a rhythm problem
18 that inevitably is going to end up in atrial fibrillation
19 anyway. So why risk one of our patients' potential
20 lives for that soft an indication.
21 CHAIRMAN PACKER: Do you think -- Tom, I
22 just want to clarify. Do you think that the drug --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
368
1 that no drug should be approved for that indication or
2 that the drug would have to be safer than this one?
3 DR. GRABOYS: Well, safety is the prime --
4 I think is the prime concern. But if I were going to
5 review all of the membrane active antiarrhythmic drugs,
6 I would like to hold them to the same criteria.
7 CHAIRMAN PACKER: Marv?
8 DR. KONSTAM: I am going to vote no and I
9 just want to make a few points. One is I just -- Peter
10 made the comment earlier about previous drugs approved
11 and whether we would be holding this to a higher
12 standard. And just in general terms, it is always sort
13 of an agonizing problem. But in the end, I think you
14 wind up having to say, okay, what about the drug before
15 us today. So I think that is the take-home that I wind
16 up making. And beyond that, I think it differs again
17 with regard to the mechanistic questions that I will
18 mention in a moment.
19 The other point I want to make is I am not
20 sure how I really should be influenced by the fact that
21 this is already an available drug with off-label use.
22 The sponsor feels that approval is needed for the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
369
1 purpose of doing education. I understand that point,
2 but I am also persuaded by the opposite that taking the
3 drug now and the FDA giving it the label to say yes but
4 it is specifically safe and effective in a fib I think
5 is going beyond what I would like to do, and I think
6 the bottom line is people will be able to use this drug
7 off-label if they feel they want to do it.
8 I think the thing that I wind up coming home
9 on is the problem I am facing by the fact that we don't
10 know what exactly this drug is doing. And it is on both
11 the mechanistic level as well as on a clinical level.
12 The drug is a beta blocker, and we think that on a
13 mechanistic level the beta blockade has some significant
14 contribution to its effect and it may be all of its
15 effect. And I think likewise the corollary of that is
16 on the clinical level, it may in fact be working
17 predominantly by lowering heart rate in patients who
18 go into atrial fibrillation. And I think from a pure
19 efficacy perspective, that probably doesn't matter.
20 But I think it does matter relative to the risk. I don't
21 think the risk would be out of the acceptable range if
22 I really knew that I was providing the medical community
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
370
1 with a new mechanistic agent. But I guess I am not
2 convinced of that from the data set, and that is really
3 why I wind up coming down negative on the cost/benefit
4 ratio. Because I do think there is a risk and I don't
5 think we know what it is and I don't think we know what
6 it is particularly from studies like the Julian and other
7 studies.
8 The only other point I wanted to add that
9 was made earlier but we haven't focused in on is the
10 absence of experience with this agent in the presence
11 of dilthiazim or verapamil. I think there is going to
12 be widespread use with these two drugs, and we know
13 absolutely nothing about the safety and efficacy of the
14 agent with those two agents. So I think that is another
15 negative.
16 CHAIRMAN PACKER: Michael, your vote won't
17 count, but we would like to here what you think.
18 DR. CAIN: In both drugs, I would approve
19 it for the indication used, although when we get to
20 number 9, it may be one patient.
21 CHAIRMAN PACKER: I understand. Ileana?
22 DR. PIÑA: I would vote to approve.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
371
1 CHAIRMAN PACKER: Udho?
2 DR. THADANI: I would put the word
3 symptomatic, because this one doesn't say it.
4 CHAIRMAN PACKER: You can -- we will talk
5 about modification -- and I think everyone on the
6 committee has very specific recommendations for
7 limitation, restriction and modification. So let us
8 postpone that until question 9. If you think it should
9 be approved for anyone, the answer should be yes.
10 DR. THADANI: Yes, I think for one of the
11 nine, I would vote yes. Because there are certain
12 reservations I would like to make.
13 CHAIRMAN PACKER: Okay. My vote is also
14 yes. Califf is yes. It is 6 to 3. Okay. Now, JoAnn,
15 can you outline for us the specific restrictions that
16 should apply? 9A is the approval should be limited to
17 specific individuals. Who should it be limited to, if
18 at all?
19 DR. LINDENFELD: Well, we have some
20 specifics just by the exclusion criteria. It should not
21 probably be given in overt heart failure. And we know
22 about patients with bradycardia or any
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
372
1 contraindications to beta blockers. And probably not
2 at least preliminarily in patients on rate-lowering
3 calcium channel blockers.
4 CHAIRMAN PACKER: Anything else?
5 DR. LINDENFELD: Those are the main ones.
6 CHAIRMAN PACKER: Okay. The question also
7 contains should it be restricted to those who have severe
8 or disabling symptoms as part of their symptomatic
9 atrial fibrillation?
10 DR. LINDENFELD: Well, I would like to see
11 its use restricted to patients who have significant
12 symptoms, but I don't know that I can recommend that
13 on the basis of this data.
14 CHAIRMAN PACKER: You could recommend that
15 based on an assessment of risk to benefit.
16 DR. LINDENFELD: Then I would probably
17 recommend that at least patients with significant
18 symptoms, yes.
19 CHAIRMAN PACKER: Okay. I think that
20 everyone on the committee would agree that there should
21 be specific mention of rate-lowering calcium channel
22 blockers. That there be mention of no use in overt heart
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
373
1 failure. The sponsor has already proposed that. Other
2 beta blockers -- let me also suggest that the patients
3 who should not receive the drug includes the one that
4 JoAnn mentioned earlier, which is elderly women because
5 almost all of them have creatinine clearances less than
6 the cut-off. I mean, when they have a certain
7 creatinine. I don't know how you phrase that. My sense
8 is that --
9 DR. FENICHEL: Well, Milton, do you think
10 it is essential to phrase that as in addition to the
11 restriction in terms of creatinine?
12 CHAIRMAN PACKER: I don't think that
13 physicians translate a creatinine of 1.4 into a
14 creatinine clearance of less than 50. I think that is
15 what JoAnn's point was. But let me -- JoAnn, what do
16 you think?
17 DR. LINDENFELD: No, I think so. I mean,
18 this is not in keeping with current labeling practices,
19 but it might even be reasonable to say that a 70 kg,
20 75-year-old woman with a creatinine of 1.4 or higher
21 is not eligible for this drug by creatinine clearance
22 criteria. I just think that brings home to the doctor,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
374
1 if they read that box, that I think that is a patient
2 that most people wouldn't be terribly worried about.
3 But when you do the calculations, that creatinine
4 clearance is below 40. And that is a lot of people with
5 atrial fib.
6 CHAIRMAN PACKER: Bob, let me just get a
7 sense of the committee and then find out how many in
8 the committee would disagree with a restriction based
9 on something more directed than -- I guess we are --
10 how many share the concern that it should be something
11 more specific than a calculated creatinine clearance?
12 Because the way the Division would normally do this
13 would be creatinine clearance and JoAnn says, well gee,
14 that is true but the creatinine clearance cut-off here
15 isn't 20 or 30. The creatinine clearance cut-off is
16 40 and 50. And 40 and 50 cuts of a lot of people. Does
17 the committee -- how does the committee feel about this?
18 Cindy?
19 DR. GRINES: Well, I guess I agree that we
20 don't really know how to calculate creatinine clearance.
21 And I think if you are talking about putting in
22 restrictions, you probably should put a chart based on
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
375
1 gender, age, and body size, and serum creatinine.
2 Something that is more than just a specific patient.
3 But I also wonder whether we have enough data to even
4 make that suggestion since there is a lower dose
5 available. And if anything, if one looks at the data
6 on creatinine clearance of less than 60, those handful
7 of patients actually had higher efficacy. So I am not
8 as concerned and perhaps it should just be cautioned
9 that a lower dose be given in patients with low
10 creatinine clearance.
11 CHAIRMAN PACKER: Well, the sponsor is
12 actually suggesting that such patients not receive the
13 drug.
14 DR. LINDENFELD: And I think also there is
15 some question about what the half-life is when you get
16 the creatinine clearance down there. There was one
17 suggestion that under 40 that the time interval of drug
18 dosing might be 36 hours. So I think we don't have any
19 way of telling what to do there when we get that low.
20 CHAIRMAN PACKER: Okay. Let me -- again,
21 how many of you would restrict the drug to patients whose
22 symptoms were severe or disabling?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
376
1 DR. THADANI: I think I would like to
2 because if a patient is not symptomatic -- because the
3 whole data base I have seen in symptomatic patients.
4 So I think I would like to restrict it given the
5 potential side effects to restrict it to that. So if
6 you are going to use it asymptomatic or mildly
7 symptomatic patients, I have not seen any overall
8 benefit. And with the noise of some worry, I would
9 probably restrict it to the patient who still remains
10 symptomatic despite, you know, whatever. So I think I
11 would go for the labeling that since, you know, severely
12 or disabling fibs.
13 CHAIRMAN PACKER: All right, Tom? This is
14 Tom Graboys. Tom, the assessment of risk to benefit
15 here I think was very typical to your thinking process.
16 How would you feel -- and this is to try to understand
17 what you were saying earlier -- how would you feel if
18 the labeling specifically said to patients with a
19 history of severe and disabling symptoms when they were
20 in atrial fib?
21 DR. GRABOYS: You mean as the prime
22 indication?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
377
1 CHAIRMAN PACKER: As the prime indication.
2 DR. GRABOYS: Well, I think that should be
3 -- by definition, yes, I think that should be the sole
4 prime indication. That still doesn't change my vote.
5 CHAIRMAN PACKER: No, I understand that.
6 That is okay. Okay, how many would disagree with that?
7 DR. BIGGER: I think that is too
8 restrictive. I think that language is too restrictive.
9 I think someone with significant aggravating symptoms,
10 not necessarily disabling or life-threatening. The
11 wording sounds overly restrictive to me.
12 CHAIRMAN PACKER: Okay. What I would like
13 to do is take two votes, because this is really
14 important. I think everyone agrees about overt heart
15 failure, rate lowering calcium channel blockers,
16 concomitant beta blockers. There is agreement on the
17 creatinine clearance or a renal function exclusion.
18 I want two votes. Vote number one is on severe and
19 disabling symptoms. Vote number 2 is with or without
20 structural heart disease. Those are specific issues
21 asked for by the Division. So we will take severe and
22 disabling symptoms first.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
378
1 DR. FENICHEL: Milton, those were just
2 examples.
3 CHAIRMAN PACKER: No, I know.
4 DR. FENICHEL: We just --
5 CHAIRMAN PACKER: But they are good
6 examples.
7 DR. FENICHEL: Okay.
8 CHAIRMAN PACKER: So the first question is
9 do you believe the drug -- the approval should be
10 restricted to patients with severe or disabling symptoms
11 at the time of their atrial fibrillation? The answer
12 would be yes or no. JoAnn?
13 DR. LINDENFELD: Yes.
14 CHAIRMAN PACKER: Okay. Udho?
15 DR. THADANI: Yes.
16 CHAIRMAN PACKER: Ileana?
17 DR. PIÑA: Can you repeat that question
18 again?
19 CHAIRMAN PACKER: Yes. Should the
20 approval -- should the indication for the drug include
21 a restriction or use only in patients with severe or
22 disabling symptoms at the time of their atrial
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
379
1 fibrillation. The wordsmithing will be worked out by
2 the Division.
3 DR. PIÑA: Yes. I think that is too
4 restrictive. These patients that -
5 CHAIRMAN PACKER: That is what we are
6 asking.
7 DR. PIÑA: By the studies that we have used
8 today to say, yes, the drug should be approved included
9 patients with symptoms. It didn't say disabling and
10 severe. So I think that that is too restrictive.
11 DR. LINDENFELD: I think we are partially
12 basing that on the fact that we were concerned about
13 the overall risk and that the drug -- that people feel
14 that we wouldn't like to necessarily recommend this drug
15 just for everyone to prevent atrial fibrillation, but
16 rather those that have substantial symptoms with their
17 atrial fibrillation. I think that is the --
18 DR. PIÑA: I agree. I mean, we have been
19 going back and forth with this all day that the patients
20 that dropped out were probably the patients who perhaps
21 needed the drug more or the population we may see more
22 often. I just don't think that we have any data to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
380
1 specifically say only severe or disabling.
2 CHAIRMAN PACKER: No, no, no. The
3 severe/disabling can be imposed as a way of assessing
4 the concept of risk to benefit. This is Tom's point.
5 DR. PIÑA: I would say -- and I certainly
6 understand Tom's point and I agree with him that drugs
7 are used not as they should. But I would say symptomatic
8 a fib.
9 CHAIRMAN PACKER: Okay. So just to make
10 sure I've got it correct, JoAnn, I think you voted yes
11 for severe/disabling. Udho, you voted yes for severe
12 disabling. Ileana, you are voting no for
13 severe/disabling. I just want to keep it clean.
14 Michael?
15 DR. CAIN: I would vote for severe and
16 disabling.
17 CHAIRMAN PACKER: Okay. Marv Konstam
18 voted for severe and disabling. Tom?
19 DR. GRABOYS: Yes.
20 CHAIRMAN PACKER: Yes. Cindy?
21 DR. GRINES: I would vote yes if we could
22 relabel all the existing drugs for the exact same
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
381
1 indication. Because I don't think it is fair. This
2 drug has no worse of a safety profile than anything else
3 I have seen. And to -- I think it is unfair to label
4 this one for severe and disabling and have a wide open
5 indication for other drugs.
6 DR. FENICHEL: Well, let me remind you of
7 what the labeling for quinidine says. The labeling for
8 quinidine describes the meta-analysis showing that
9 quinidine triples the mortality in those who receive
10 it. And then it says this drug is for people whose
11 symptoms are so frequent and severe that they in
12 discussion with their physicians are willing to accept
13 that increase in mortality in exchange for the
14 symptomatic benefit which is presumed to come from the
15 use of quinidine. So it is not an altogether
16 unprecedented thing to describe the requirement in terms
17 of severe and disabling symptoms. On the other hand,
18 I would hasten to point out, and I am sure the sponsor
19 will point it out if I do not, that there is no allegation
20 here that mortality is tripled by the d,l-sotalol.
21 DR. GRINES: Well, I guess the other drugs
22 that are approved -- flecainide and other -- exactly.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
382
1 That if we are going to say severe and disabling symptoms
2 for d,l-sotalol, I think that we have to be consistent
3 with all drugs that maintain sinus rhythm.
4 CHAIRMAN PACKER: It is really two separate
5 questions, and the question is what do you think should
6 be done with d,l-sotalol. And you could say to the
7 Division that they should seek a similar --
8 DR. GRINES: Right. Well, a phrase like
9 that, I do believe that all antiarrhythmic drugs for
10 atrial fibrillation should be used only for severe.
11 CHAIRMAN PACKER: Okay. I think that is
12 fine. Tom? I think you have said it is too severe.
13 DR. BIGGER: Yes.
14 CHAIRMAN PACKER: I would agree with severe
15 and disabling. So the vote on that, for people who
16 count, is 6 to 1, Joan, for severe and disabling. The
17 next point is structural -- Moye didn't vote. I do have
18 Marvin's vote and I don't have any comment from Rob
19 Califf on this. So we can only use the votes we have.
20 Structural heart disease? Who would favor restricting
21 this drug to patients without structural heart disease?
22 JoAnn, would you favor restricting the drug to patients
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
383
1 without -- this would be a flecainide type labeling?
2 DR. LINDENFELD: No, I don't think I would
3 restrict it.
4 CHAIRMAN PACKER: Would anyone restrict it
5 to patients without structural heart disease?
6 DR. GRABOYS: Yes, I would. I just think
7 again the data on proarrhythmia continues to be so
8 impressive in terms of the dichotomy of proarrhythmia
9 dependent upon the presence or absence of structural
10 heart disease. Again, I am concerned with the whole
11 concept of the trickle down. We are trying to come up
12 with some indication for it that is going to incorporate
13 physicians' practice. And if we open it up for across
14 the board, you are going to have patients with ischemic
15 disease, recent infarct. I mean, there is going to be
16 a whole panoply of problems.
17 CHAIRMAN PACKER: Okay. Does anyone want
18 to vote along with Tom for a restriction to no structural
19 heart disease? If not, then the vote is 6 to 1 in favor
20 of phraseology with and without structural heart
21 disease. The next consideration is, let's see, should
22 the approval distinguish between chronic and paroxysmal
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
384
1 fibrillation? We have discussed this already. JoAnn,
2 what do you think?
3 DR. LINDENFELD: No, I don't think so.
4 CHAIRMAN PACKER: Does anyone think there
5 should be a distinction? Okay. Bob?
6 DR. THADANI: Before you go further, I
7 think one of the issues -- the strongest evidence was
8 in patients with a chronic who were converted and then
9 relapse rate was delayed. I have some concern with the
10 paroxysmal because of the -- as we discussed in the study
11 because intent to treat did not show a difference. I
12 don't know. I have some of my reservations in that
13 situation because unless you are doing repeated Holter
14 monitoring. Plus, the patients who were dropped from
15 there. It is only on one study. So I feel more
16 comfortable with patients who are in chronic a fib are
17 converted and on this drug until the first relapse rather
18 than paroxysmal. So I will have some concern there.
19 CHAIRMAN PACKER: But it sounds like
20 everyone else -- Michael?
21 DR. CAIN: I just think it is important that
22 if you use the word paroxysmal and chronic, I think you
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
385
1 will increase the risk that people will misuse the drug,
2 and I would recommend that you not get into that trap.
3 And what you are really talking about is the treatment
4 for people who had a recent history of atrial
5 fibrillation who are now on a sinus rhythm. And leave
6 paroxysmal and chronic out of it.
7 CHAIRMAN PACKER: Okay. If I get a sense
8 -- I just want to make sure that what we are talking
9 about resembles the following, which would be something
10 like the reduction or a delay in the onset of or a
11 reduction in the risk of recurrence of atrial
12 fibrillation or atrial flutter in patients in normal
13 sinus rhythm with a recent history of atrial fib or
14 flutter that produced severe or disabling symptoms.
15 DR. THADANI: And have been converted into
16 sinus rhythm.
17 CHAIRMAN PACKER: No, no. We already said
18 that. In sinus rhythm.
19 DR. THADANI: No, but recent -- in sinus
20 rhythm at the time of start.
21 CHAIRMAN PACKER: No, it says in sinus
22 rhythm. What I just said was in normal sinus rhythm
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
386
1 with a recent history of atrial fib/atrial flutter that
2 produced -- was associated with or produced severe
3 disabling symptoms. Okay, Peter?
4 DR. KOWEY: Just a brief comment, Milton,
5 as a point of order. When the sponsor came for the
6 pre-meeting meeting with Ray, who is unfortunately not
7 here today, the sponsor really didn't differentiate
8 these two arrhythmias. It really was Ray who asked us
9 to present the data to the specific subsets. And the
10 reason was because of the recent dofetilide experience.
11 I personally agree with what you are saying. The only
12 thing I would ask, and I am sure this will happen with
13 Bob, is that the words be crafted carefully so that it
14 is clear what the data in the data set showed, and I
15 think that is what Michael was saying, rather than saying
16 you can use it in an arbitrarily defined subgroup that
17 we really have a hard time defining anyway. So I think
18 we are all in agreement with that.
19 CHAIRMAN PACKER: Okay. I think we
20 actually have consensus on this. And I think Michael's
21 point that if you include paroxysmal -- the words
22 paroxysmal or chronic -- you are going to increase the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
387
1 likelihood that the application of the drug would be
2 misunderstood.
3 DR. CAIN: Electrophysiologists can't
4 agree on how to pronounce the arrhythmia, let alone
5 define it.
6 CHAIRMAN PACKER: Fibrillation, right?
7 Nevermind. All right. Okay, a lot of the other issues
8 are straightforward. But let me -- there is one -- there
9 are two other very important issues here that need to
10 be addressed. Should the data -- should the drug be
11 started -- who should be hospitalized for initiation
12 of the drug? This is a very important consideration.
13 The sponsor is suggesting outpatient use for patients
14 without structural heart disease and inpatient
15 initiation in patients with structural heart disease.
16 JoAnn, what do you think?
17 DR. LINDENFELD: I am not comfortable yet
18 starting this drug as an outpatient for several reasons.
19 One, I think there is still a reasonably small number
20 of patients, 349 in 04 and 25 percent of 05, which is
21 a small number. But also we have discussed over and over
22 again the population that will actually be treated, and
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
388
1 those are older people. And I think expecting -- seeing
2 the risk of taking an extra drug, I just am not yet
3 comfortable with the safety of this drug as an
4 outpatient. So I would say no to that.
5 CHAIRMAN PACKER: Okay. Anyone would
6 disagree with JoAnn? Okay. Then it is the consensus
7 of the committee as well as Marv Konstam, who also said
8 that all patients should be hospitalized for treatment.
9 So it would be in-hospital initiation. Any other
10 comments? Okay.
11 DR. THADANI: Also, I think it would be nice
12 to collect the data base on patients to give more comfort
13 in Tom's question of patients who are elderly and
14 patients with relatively poor LV function. So at least
15 we will have a bit more objective data collection after
16 the approval process.
17 CHAIRMAN PACKER: Okay. Let's go through
18 the other issues very rapidly because most of them are
19 fairly straightforward. I assume that everyone would
20 agree that there should be adjustment based on renal
21 function, and I think, JoAnn, you specifically indicated
22 that specific clinical examples of what constitutes a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
389
1 patient who is not a candidate based on renal function,
2 elderly women for example, be specifically mentioned.
3 DR. LINDENFELD: Right.
4 CHAIRMAN PACKER: Anyone disagree? Okay.
5 Should any recommendation be made about
6 anticoagulation with respect to the use of this drug?
7 DR. THADANI: Yes, I think -- oh, sorry.
8 CHAIRMAN PACKER: JoAnn?
9 DR. LINDENFELD: I don't think so. I think
10 that just as when we discussed dofetilide, those are
11 clear indications now for anticoagulation.
12 CHAIRMAN PACKER: Okay. Therefore no
13 mention? Anyone disagree?
14 DR. THADANI: I think all patients should
15 be anticoagulated who are--
16 CHAIRMAN PACKER: That is not the question.
17 The question is should the labeling say so? We said
18 not to say it in dofetilide. So the same here as for
19 dofetilide? Tom? That is fine. Okay. Last two
20 questions, should a -- what program should be instituted
21 to determine what fraction of patients are receiving
22 sotalol in accordance with the dosing regimen that would
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
390
1 be recommended? This is a requirement for a formal
2 comprehensive post-marketing surveillance program.
3 Something like that was discussed with dofetilide. It
4 wasn't entirely clarified. Do you think that kind of
5 surveillance is important for this drug? JoAnn?
6 DR. LINDENFELD: This is a hard one. I
7 think that the things that every physician should know
8 are obviously the calculated creatinine clearance and
9 the QT interval. And that is what everybody should know
10 with dofetilide too, in addition to some other things.
11 So I think probably they should be required.
12 DR. FENICHEL: Well, that is not the
13 question unfortunately, because that is easy. I mean,
14 the question that you are answering is --
15 CHAIRMAN PACKER: The question is what is
16 it.
17 DR. FENICHEL: Well, the question that
18 JoAnn was answering was how should the drug be given.
19 Well, of course, that is what you have been answering
20 for much of the day. The question that we are asking
21 here, and it is the same question that came up with
22 dofetilide, is there something that should be part of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
391
1 the approval package that guarantees -- or guarantee
2 may be too strong -- that makes it more likely that this
3 advice that will be in the labeling is in fact being
4 heeded. And the extreme example that is given in the
5 question is that used with Clozaril, where the problem
6 in that case is a matter of getting repeated CBC's to
7 look for neutropenia, which is caused by that drug in
8 something like 1 percent of the patients. And what is
9 done there is that patients may not obtain the drug
10 without demonstrating that they have shown up and got
11 their blood drawn. Now that is about the most radical
12 case that I know about in terms of making sure that the
13 drug is being given right. Actually, there is another
14 one that is in progress for the reappearance for some
15 age-related and other indications for thalidomide,
16 where you really want to make sure that people are
17 getting the drug right. So that was the question that
18 was raised with dofetilide. Not how do you give it,
19 but rather what should be done to make sure that people
20 are doing that or following those instructions that you
21 just made up so nicely. It is a toughie. You said it
22 was a toughie when you thought it was the easier one.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
392
1 DR. LINDENFELD: I know. That is why I
2 answered the easier one. Boy, I don't know. I think
3 whatever we do would have to be done with dofetilide
4 as well, the same issues. I think that what could be
5 done, of course, is to require at least for an initial
6 prescription a QT interval and a calculated creatinine
7 clearance. I think people would be upset by that, but
8 probably -- physicians -- but probably they shouldn't
9 be because that is what you should have to prescribe
10 it. And if you haven't -- as I said about with the
11 example of the 70-year-old lady, if you haven't sat down
12 and calculated the creatinine clearance, you are going
13 to get a surprise in a lot of these people. So if we
14 are going to do a program, at least -- now it is going
15 to be a problem, because I am not sure that has to be
16 done for every single recurrent prescription. I don't
17 think it does. But at least for an initial
18 prescription, a calculated creatinine clearance and a
19 QT interval for the prescription.
20 CHAIRMAN PACKER: I am curious, I
21 understand something like this was discussed for
22 dofetilide as well. There is a difference here, and
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
393
1 the difference is that this is a drug which is already
2 on the market. And the requirements that you are
3 talking about are not imposed for the present use of
4 the drug although the present use of the drug includes
5 the possibility of doses higher than the ones being
6 recommended for atrial fibrillation, and no such
7 surveillance is mandated at the present time. So we
8 would be in a sort of interesting situation of requiring
9 greater surveillance for a lower dose in a lower risk
10 patient population, but no surveillance for a higher
11 dose in a higher risk patient population. Now we have
12 done crazy things like that before. I just want to know
13 whether you think we should do a crazy thing like that
14 again.
15 DR. LINDENFELD: Well, I think maybe we
16 should. I think I understand the point you are making.
17 But also I think as Tom has said several times, this
18 is a population of people -- a population of people with
19 life-threatening ventricular arrhythmias. Indeed, the
20 risk may be higher, but they have a substantial benefit.
21 And here we may have a bigger risk/benefit ratio than
22 in the other population. I think that is possible.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
394
1 So although it is a conundrum, if we were approving this
2 for the first time, I think I would say -- if it were
3 not on the market, I think I would say yes. I mean,
4 we want to do everything we can do. We have said we
5 are not sure this drug has an overall benefit and it
6 does have a risk.
7 DR. KOWEY: Milton, if you do what you said
8 in the preceding part, which was to mandate an
9 in-hospital start for all patients, then doing what
10 JoAnn is suggesting isn't really such a big deal. I mean,
11 they are in the hospital and they get a creatinine and
12 they have an EKG. So it is -- this really is a sting,
13 especially for the initial dosing, is if you ever let
14 somebody do it out of hospital. But let me just tell
15 you that having said that you don't want it started out
16 of hospital, unfortunately there will be that that will
17 happen. And I guess the question I have is admitting
18 that, do you want to talk about out-of-hospital starts
19 even though you are not telling people to do that?
20 Because it is going to happen. You see, JoAnn's
21 question is -- JoAnn's answer is easy if you are doing
22 it in the hospital. But what happens if you are starting
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
395
1 it out of the hospital, as you are not supposed to be
2 doing. It is sort of an interesting kind of conundrum.
3 CHAIRMAN PACKER: Right. In other words,
4 the patients who require the greatest -- who would be
5 the source of the greatest concern would be the ones
6 in which the physician isn't doing the right thing in
7 the first place and therefore is more likely to not --
8 to continue to do the wrong thing?
9 DR. KOWEY: Correct. That is correct.
10 DR. PIÑA: I have a question for Bob.
11 DR. KOWEY: That is what you have set up
12 basically.
13 DR. PIÑA: I have a question for Bob. How
14 is the Clozaril program being handled? Is it the
15 pharmacist who has to dispense the drug but can't
16 dispense it unless he or she sees the white count, and
17 in this case it would be the pharmacist who wouldn't
18 dispense the drug until they see the EKG and the
19 creatinine clearance and know the QT and know the
20 creatinine clearance?
21 DR. FENICHEL: Yes. It is a good question
22 and the answer is I don't know it. I don't know the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
396
1 answer. Clozapine is part of -- I mean, there is
2 something of the sort I described that is in use right
3 now and it is part of Agency folklore, but I personally
4 don't know what the details are.
5 DR. THADANI: You know, Milton, there are
6 almost 3 million prescriptions written already on the
7 drug, which is greater than the indicating use of VT.
8 You know, it is mind boggling the numbers. Obviously
9 if you are saying that this drug must be used as an
10 inpatient, then I think we also should say that the
11 patient must have creatinine clearance measured,
12 formula given provided by the company on a little caliper
13 or whatever, and also make sure that the ECG is done
14 before any dose escalation to safeguard the patient,
15 which should not be difficult. Now, if the patient --
16 if people are going to -- how are you going to collect
17 data on people who are going to use it outside unless
18 the Agency or the company is going to track all the
19 prescriptions outside. It would be impossible. So I
20 think at least inpatient you could try it.
21 DR. PIÑA: You know, what I do think the
22 company would need to do to my satisfaction -- and even
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
397
1 though I do a lot of teaching and sometimes I don't think
2 that physicians always listen to everything that we have
3 to say -- that the company does have to embark on a very
4 strong educational program to teach physicians who are
5 likely to use the drug how to use it and how to use it
6 appropriately. And I don't care if you have to hand
7 them something to show them how to calculate a creatinine
8 clearance. Even though it is so simple, most people
9 don't know how to calculate a creatinine clearance
10 unless they actually order the 24-hour urine clearance.
11 So other than that, I don't see how you would enforce
12 this.
13 DR. THADANI: It might work negative
14 against the company. They are already using our
15 prescription and now you are going to decline them.
16 CHAIRMAN PACKER: Yes. I don't actually
17 think we have to go further with this. I think, Bob,
18 you have a sense of the kinds of discussions that one
19 could have, and I think we have reached the limit as
20 to how precisely we can define it. The last question
21 to the committee, and that is what post-marketing
22 commitments should be made? This would include any of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
398
1 the above-listed in the question or action studies and
2 studies in patient population, head-to-head
3 comparisons. JoAnn, what do you think?
4 DR. LINDENFELD: Well, I think it would be
5 wonderful to have a study in the actual population of
6 patients that will be treated average age 75, half or
7 close to half women. That was the one I would like to
8 see most with enough patients at least followed up for
9 a minimum of six months and probably a year.
10 CHAIRMAN PACKER: We are talking about
11 things that would be required. Would you require that
12 study?
13 DR. LINDENFELD: I don't think I would
14 require it, no.
15 CHAIRMAN PACKER: Okay. Who would require
16 -- there are all sorts of studies that one could imagine
17 here and maybe the best thing --
18 DR. FENICHEL: Well, Milton, we didn't use
19 the word require.
20 CHAIRMAN PACKER: Soft?
21 DR. FENICHEL: In part because we have no
22 legal authority to use the word require.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
399
1 CHAIRMAN PACKER: I understand.
2 DR. FENICHEL: If it is approved -- and this
3 is important and perhaps I should have mentioned this
4 before you voted about approval, but I don't know that
5 it would have changed the decision or should have. But
6 the fact is that if you think, well it should be approved
7 but only because we know they are going to do such and
8 such study, well then you shouldn't vote to approve it.
9 We don't have the facility to make conditional
10 approvals. And so all we can do is seek such studies.
11
12 CHAIRMAN PACKER: Okay. I think -- let me
13 -- I think the best way, therefore, to do it is to answer
14 the question the way it is framed, which is should
15 certain studies be sought. And let me just propose the
16 following, only because they came up during the course
17 of the meeting. Elderly patients -- should such a study
18 be sought? Anyone disagree? Okay. Interaction
19 studies with calcium channel blockers and/or beta
20 blockers, anyone disagree?
21 DR. THADANI: I don't know whether we would
22 need with all the calcium channel blockers. I would
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
400
1 go for the calcium channel blockers which reduce or
2 affect the AV node. I am not sure the four hydroproteins
3 would make a difference, unless you believe they are
4 both cardio-depressants.
5 CHAIRMAN PACKER: Okay. Any other
6 populations that -- yes, please, Michael?
7 DR. CAIN: I think the other one is the NIH
8 is putting out another application for scores in sudden
9 death in African Americans because of the high incidence
10 or higher incidence of sudden death in blacks, and one
11 of the presumed mechanisms of that is left ventricular
12 hypertrophy, which fits into hypertension, left
13 ventricular hypertrophy, atrial fibrillation, and there
14 really are no or very few data on non-whites. And so
15 I think that becomes critical.
16 DR. THADANI: One question didn't come up.
17 At least sometimes we use beta blockers and amiodarone
18 in certain populations. There is no data on it. So
19 are we going to say this should not be used concomitantly
20 with amiodarone? Because both could be used for the
21 same indication. And I could see a patient with
22 coronary artery disease goes into a fib and gets put
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
401
1 on this drug for whatever reason and later on he might
2 have VT and gets put on amiodarone. Do we need more data?
3 How comfortable do you feel? Or we should make a
4 recommendation that there is no data and should not be
5 used?
6 CHAIRMAN PACKER: That is really actually
7 more of an addition to the list of calcium channel
8 blockers that there is also no concomitant data on.
9 Any other suggestions or modifications or any other
10 comments? Bob, have we addressed the questions from
11 the Division?
12 DR. FENICHEL: Yes.
13 CHAIRMAN PACKER: We are adjourned.
14 (Whereupon, at 5:28 p.m., the meeting was
15 adjourned.)
16
17
18
19
20
21
22
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
402
1
2
3
4
5
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
