1
DEPARTMENT OF HEALTH AND HUMAN
SERVICES
FOOD AND DRUG
ADMINISTRATION
CENTER FOR DRUG EVALUATION AND
RESEARCH
PSYCHOPHARMACOLOGIC DRUGS
ADVISORY COMMITTEE
WITH THE PEDIATRIC
SUBCOMMITTEE
OF THE ANTI-INFECTIVE DRUGS
ADVISORY COMMITTEE
Holiday Inn
2
PARTICIPANTS
Matthew Rudorfer, M.D., Chair
Anuja M. Patel, M.P.H., Executive
Secretary
PSYCHOPHARMACOLOGICAL DRUGS ADVISORY
COMMITTEE
MEMBERS
Tana Grady-Weliky, M.D.
Irene E. Ortiz, M.D.
Richard
P. Malone, M.D
Wayne K. Goodman, M.D.
James J. McGough, M.D.
Jean
E. Bronstein, R.N., M.S.
(Consumer
Rep)
Andrew C. Leon, Ph.D.
Philip S. Wang, M.D. M.P.H.,
Dr. P.H.
Dilip J. Mehta, M.D., Ph.D.,
(Industry Rep)
ANTI-INFECTIVE DRUGS ADVISORY COMMITTEE
MEMBERS
Steven C. Ebert, Pharm. D.
(Consumer Rep)
Mary P. Glode, M.D.
Samuel
D. Maldonado, M.D., M.P.H.
(Industry
Rep)
PEDIATRIC SUBCOMMITTEE OF THE
ANTI-INFECTIVE DRUGS
ADVISORY COMMITTEE MEMBERS
P. Joan Chesney, M.D.
Mary Glode, M.D.
Steven Ebert, Pharm. D.
(Consumer Rep)
Robert Nelson, M.D., Ph.D.
Richard Gorman, M.D., FAAP
Robert J. Fink, M.D.
Susan Fuchs, M.D.
David Danford, M.D.
Victor Santana, M.D.
Mark Hudak, M.D.
Judith R. O'Fallon, Ph.D.
SGE CONSULTANTS (VOTING)
Elizabeth B. Andrews, Ph.D.
Norman Fost, M.D., M.P.H.
Charles E. Irwin, Jr., M.D.
Lauren K. Leslie, M.D., FAAP
James
M. Perrin, M.D.
Cynthia
R. Pfeffer, M.D.
SGE PATIENT REPRESENTATIVE (VOTING)
Gail W. Griffith
3
PARTICIPANTS
(Continued)
GOVERNMENT EMPLOYEE (non-voting)
Daniel
S. Pine, M.D.
FDA
Robert Temple, M.D.
Russell G. Katz, M.D.
Thomas Laughren, M.D.
M. Dianne Murphy, M.D.
Susan Cummins, M.D., MPH
Anne Trontell, M.D., MPH
4
C O N T E N T S
PAGE
Call to Order and Opening Remarks:
Matthew Rudorfer, M.D. 6
Introductions 8
Conflict of Interest Statement:
Anuja M. Patel, MPH 15
Overview of Issues:
Russell Katz, M.D. 19
Pediatric Drug Development Program:
Dianne Murphy, M.D. 26
Pediatric Depression and Its Treatment:
Cynthia R. Pfeffer, M.D. 39
Suicide and Related Problems in
Adolescents:
David Shaffer, FRCP (Lond) FRC
Psych 60
Open Public Hearing
Irving Kirsch and David
Antonuccio 79
Lisa
Van Syckel
82
Ann Blake Tracy, Ph.D. 83
Tom Woodward 85
Mark Miller 87
Corey and Jay Baadsgaard 90
Joyce Storey 91
Jame Tierney 93
Donna and Mark Taylor 95
Shannon
Baker
97
Dawn Rider 98
Sara Bostock 100
Vera Hassner Sharav 103
Cynthia Brockman 104
Todd and Eileen Shivak 107
Andy Vickery 109
Rosie Carr Meysenburg 111
Rachel Adler 112
Pepper Draper 115
Donald Marks, M.D., Ph.D. 117
Leah Harris 119
Donald Farber 121
Matthew Piepenberg 125
Terri Williams 127
Glenn McIntosh 129
Delnora Duprey 132
Joe Pittman 133
Richard Mack 135
Noah Wright Smith 137
Marion
Goff
139
5
C O N T E N T S
(Continued)
PAGE
Open Public Hearing (Continued)
Gary Cheslek, M.D. 142
Sherri Walton 144
Peter
R. Breggin, M.D. 146
Robert Fritz 148
Suzanne Vogel-Scibilia, M.D. 152
Dennis
Winter 155
Steve Cole 157
Allan Routhier 158
Daniel J. Safer, M.D. 161
Julie Magno Zito, M.D. 163
Joseph Glenmullen, M.D. 164
Linda Cheslek 165
Jeff Avery 167
Harry Skigis 169
Pamela Wild 170
Karen Barth Menzies 172
Amy Coburn 174
Sharon McBride 175
Thomas Moore, M.D. 178
Pediatric and Adolescent Antidepressant
Drug Use
in the
Gianna C. Rigoni, Pharm.D., M.S. 181
One-Year Post-Exclusivity-Mandated
Adverse Event
Review for Paroxetine and Citalopram:
Solomon Iyasu, M.D., MPH 195
Office of Drug Safety Data Resources
for the Study of Suicidal Events:
Andrew D. Mosholder, M.D.,
MPH 215
Open Public Hearing
David Fassler, M.D. 225
Regulatory History on Antidepressants and
Suicidality and Update on Current Plans
for
Analysis of Pediatric Suicidality Data:
Thomas Laughren, M.D. 230
Suicidality Classification Project:
Kelly Posner, Ph.D. 265
Plans for Analysis of Patient Level Data
for Pediatric Studies:
Tarek Hammad, M.D., Ph.D.,
M.Sc., M.S. 273
Open Committee Discussion 291
6
1 Call to Order and Opening
Remarks
2 DR. RUDORFER: I am Dr. Matthew Rudorfer,
3 a research psychiatrist at the National
Institute
4 of Mental Health, today wearing my hat as
Chair of
5 the
Advisory Committee.
6
As you settle in, please take this
7
opportunity to put into silent mode your cell
8
phones and any other devices that ring, beep, or
9
play show tunes.
10
I have some official language to read.
11 All
committee members and consultants have been
12
provided with copies of background materials from
13 the
FDA and with copies of letters from the public
14
that were received by the January 26th deadline.
15 The
background materials have been posted on the
16 FDA
web site. Copies of all these materials
are
17
available for viewing at the FDA desk outside this
18
room.
19
We have a large table and a full house as
20 you
can see and a very important and exciting topic
21 to
discuss, so we would like to start with a few
22
rules of order. FDA relies on its
advisory
23
committees to provide the best possible scientific
24 advice
available to assist us in a discussion of
25
complex topics. We understand
that issues raised
7
1
during the meeting may well lead to conversations
2
over breaks or during lunch.
3
However, one of the benefits of an
4
advisory committee meeting is that discussions take
5
place in an open and public forum.
To that end, we
6
request that members of the committees not engage
7 in
off-record conversations on today's topic during
8 the
breaks and lunch.
9
Whenever there is an important topic to be
10
discussed, there are a variety of opinions. One of
11 our
goals today is for this meeting to be conducted
12 in
a fair and open way where every participant is
13
listened to carefully and treated with dignity,
14
courtesy, and respect. Anyone whose behavior is
15
disruptive to the meeting will be asked to leave.
16
We are confident that everyone here is
17
sensitive to these issues and can appreciate that
18
these comments are intended as a gentle reminder.
19 We
look forward to a productive and interesting
20
meeting.
21
Just to reiterate a couple of points.
22
This is an unusual meeting in that we have two
23
advisory committees represented here,
24
Psychopharmacologic Drugs and a subcommittee that
25 is
equivalent of a Pediatric Drugs Advisory
8
1
Committee chaired by Dr. Joan Chesney here to my
2
left.
3
Suppose we begin by going around the table
4 for
introductions. Can we start at that end,
5
please.
6 Introductions
7
DR. TEMPLE: I am Bob Temple. I am the
8
Office Director for Office of Drug Evaluation I.
9
DR. KATZ: Russ Katz, Division
Director of
10 the
Division of Neuropharmacological Drug Products,
11
FDA.
12 DR. LAUGHREN: Tom Laughren, Psychopharm
13
Team Leader in the Neuropharm Division.
14
DR. MURPHY: Dianne Murphy, Office
15
Director, Office of Counterterrorism and Pediatric
16
Drug Development.
17
DR. CUMMINS: Susan Cummins,
Medical Team
18
Leader with the Division of Pediatric Drug
19
Development.
20
DR. TRONTELL: Anne Trontell,
Deputy
21
Director, Office of Drug Safety.
22
DR. FUCHS: Susan Fuchs, member of
the
23
Pediatric Subcommittee of the Anti-Infective Drugs
24
Advisory Committee.
25
DR. FINK: Bob Fink, pediatric
9
1
pulmonologist, Dayton, Ohio.
2
DR. ORTIZ: Irene Ortiz, geriatric
3
psychiatrist, Albuquerque VA and the University of
4 New
Mexico.
5
DR. LESLIE: Lauren Leslie, behavioral
6 and
developmental pediatrician and health services
7
researcher in San Diego.
8
DR. LEON: Andrew Leon, Professor
of
9
Biostatistics and Psychiatry at Cornell Medical
10
College.
11
DR. GOODMAN: Wayne Goodman,
Professor and
12
Chairman, Department of Psychiatry at the
13
University of Florida.
14
DR. PFEFFER: Cynthia Pfeffer,
Adolescent
15
Psychiatrist and Professor of Psychiatry at Weill
16
Medical College of Cornell University.
17
DR. GORMAN: Rich Gorman,
pediatrician in
18
private practice in Ellicott City and member of the
19
Pediatric Advisory Subcommittee.
20
DR. GLODE: Mary Glode, Professor
of
21
Pediatrics, Pediatric Infectious Disease Specialist
22 at
Children's Hospital, University of Colorado at
23 Denver.
24
DR. HUDAK: Mark Hudak,
neonatologist and
25
Professor of Pediatrics, University of Florida at
10
1
Jacksonville, and member of the Pediatric
2
Subcommittee.
3
DR. MALONE: Richard Malone, child
4
psychiatrist, Drexel University, College of
5
Medicine, and I am a member of the Psychopharm
6
Advisory Committee.
7
DR. SANTANA: Victor Santana,
pediatric
8 hematologist/oncologist,
St. Jude's Children's
9
Research Hospital and University of Tennessee at
10
Memphis, Tennessee.
11
MS. PATEL: Anuja Patel, Executive
12
Secretary, Advisors and Consultants Staff.
13
DR. RUDORFER: Dr. Matthew Rudorfer,
14
Acting Chief, Adult Interventions Branch, National
15
Institute of Mental Health and Chair of the
16
Psychopharmacologic Drugs Advisory Committee.
17
DR. CHESNEY: Joan Chesney,
Professor of
18 Pediatrics at the University of Tennessee in
19
Memphis, and at St. Jude's Children Research
20
Hospital, and the Pediatric Subcommittee.
21
DR. McGOUGH: Jim McGough,
Associate
22
Professor in Child and Adolescent Psychiatry at
23
UCLA and member of the Psychopharm Drugs Advisory
24
Committee.
DR.
25
GRADY-WELIKY: Tana Grady-Weliky,
Associate
11
1
Professor of Psychiatry at the University of
2
Rochester, School of Medicine and Dentistry, and
3
member of the Psychopharm Advisory Committee.
4
DR. WANG: Philip Wang,
psychiatrist and
5
epidemiologist, Harvard Medical School.
6
DR. O'FALLON: Judith O'Fallon, recently
7
retired from the Cancer Center Statistics Unit of
8 the
Mayo Clinic. I am a member of the
Pediatric
9
Subcommittee.
10
DR. NELSON: Robert Nelson,
Pediatric
11
Critical Care Medicine at the Children's Hospital,
12
Philadelphia.
13
DR. ANDREWS: Elizabeth Andrews,
14
pharmaco-epidemiologist at Research Triangle
15
Institute and the University of North Carolina
16
Centers for Educational Research and Therapeutics,
17 and
I am a consultant.
18
MS. GRIFFITH: Gail Griffith. I am a
19
writer. I live in
Washington. I am the Patient
20
Representative, a parent of a child suffering from
21
MDD, and a patient who suffers from MDD.
22
DR. FOST: Norm Fost, Professor of
23
Pediatrics and Director of the Bioethics Program at
24 the
University of Wisconsin.
25
MS. BRONSTEIN: Jean Bronstein,
nurse with
12
1 a
background in psychiatry, retired, and I am the
2
Consumer Representative for Psychopharm.
3
DR. EBERT: Steve Ebert,
pharmacist and
4
infectious diseases, Professor of Pharmacy at the
5
University of Wisconsin/Madison, member of the
6
Pediatric Subcommittee.
7
DR. DANFORD: David Danford,
Professor of
8
Pediatrics and cardiologist in the Joint Section of
9
Pediatric Cardiology, University of Nebraska,
10
Creighton University, member of the Pediatric
11
Subcommittee.
12
DR. PINE: Daniel Pine, child
13
psychiatrist, National Institute of Mental Health,
14
Intramural Research Program.
15
DR. MALDONADO: Samuel Maldonado,
Chair of
16 the
Pediatric Working Group at PhRMA and
member of
17 the
Pediatric Subcommittee.
18
DR. MEHTA: Dilip Mehta from New
York. I
19 am
the Industry Representative on the
20
Psychopharmacologic Advisory Committee.
21 DR.
RUDORFER:
22
Thank you. Our session today is
actually the first
23 of
two planned advisory committee meetings convened
24 to
address recent concerns about reports of
25
suicidal ideas and behavior developing in some
13
1
children and adolescents during treatment of
2
depression with an SSRI or similar newer
3
antidepressants.
4
Our goal is to gather information from a
5
variety of sources and perspectives to help us
6
understand this complex situation and ultimately to
7
offer the best possible recommendations to the FDA.
8
I would like to thank the many groups,
9
individuals, and families that submitted written
10
statements in advance of this meeting, many of
11
which were quite informative as well as moving.
12
Much of today's meeting will be devoted to
13 a
two-part open public hearing during which dozens
14 of
people from around and even beyond the country
15
will have the opportunity to present their own
16
personal or professional experiences and ideas
17
about the relative risks and benefits of
18
antidepressant medications in children and
19
adolescents.
20
Although the necessary consideration of
21 the
clock will permit only a short time at the
22
microphone for each speaker, I can assure you that
23 the
committee welcomes and values input from all
24
viewpoints and feels it essential to our work that
25 all
voices be heard.
14
1
Major depression remains an
2
underdiagnosed, understudied, and undertreated
3
serious and even life-threatening mental disorder
4
among thousands of our nation's youth, leading to
5
considerable dysfunction, disability, and
6
heartbreak in many families.
7
I am hopeful that with a fair and
8
open-minded review of the evidence in hand and that
9
still emerging, this advisory committee can
10
constructively address the challenges we all share
11 to
assure that interventions for this deadly
12
disorder are available for those young people who
13
desperately need them and that those treatments
14
meet high standards for both effectiveness and
15
safety.
16
Now, I will ask Anuja Patel, of the FDA
17
Center for Drug Evaluation and Research, to review
18
some of the ground rules for the open public
19
hearing.
20
MS. PATEL: Good morning. As you know, we
21
have a very full open public hearing today and in
22 the
interest of both fairness and efficiency, we
23 are
running it by some strict rules.
24
Due to the vast majority of requests by
25
registered speakers to speak in the morning
15
1
session, we will lengthen the morning session of
2
open public hearing and shorten the afternoon
3
session accordingly.
4
To make the transitions between speakers
5
more efficient, all speakers will be using the
6
podium in front of the audience.
Each speaker has
7
been given their number and the order of
8
presentation, and when the person ahead of you is
9
speaking, we ask that you move to the nearby next
10
speaker chair.
11
Individual presenters and families have
12
been allotted two minutes for their presentations.
13 The
three combined groups' presentations have been
14
allotted three minutes. We will
be using a timer
15 and
speakers who run over their time limit will
16
find that the microphone is no longer working.
17
We apologize for the need for the strict
18
rules, but we wanted to give as many people as
19
possible an opportunity to participate.
Thank you
20 for
your cooperation.
21
I will now state the Conflict of Interest
22
Statement for the record.
23 Conflict of Interest
Statement
24
The following announcement addresses the
25
issue of conflict of interest with respect to this
16
1
meeting and is made a part of the record to
2
preclude even the appearance of such at this
3
meeting.
4
Based on the agenda, it has been
5
determined that the topics of today's meeting are
6
issues of broad applicability and there are no
7
products being approved at this meeting.
Unlike
8
issues before a committee in which a particular
9
product is discussed, issues of broader
10
applicability involve many industrial sponsors and
11
academic institutions.
12
All Special Government Employees have been
13
screened for their financial interests as they may
14
apply to the general topics at hand.
To determine
15 if
any conflict of interest existed, the Agency has
16
reviewed the agenda and all relevant financial
17
interests reported by the meeting participants.
18
The Food and Drug Administration has
19
granted general matter waivers to the Special
20
Government Employees participating in this meeting
21 who
require a waiver under Title 18, United States
22
Code, Section 208.
23
A copy of the waiver statements may be
24
obtained by submitting a written request to the
25
Agency's Freedom of Information Office, Room 12A-30
17
1 of
the Parklawn Building.
2
Because general topics impact so many
3
entities, it is not prudent to recite all potential
4
conflict of interests as they apply to each member
5 and
consultant and guest speaker.
6
FDA acknowledges that there may be
7
potential conflicts of interest, but because of the
8
general nature of the discussion before the
9
committee, these potential conflicts are mitigated.
10
With respect to FDA's invited industry
11
representatives, we would like to disclose that Dr.
12
Dilip Mehta and Dr. Samuel Maldonado are
13
participating in this meeting as industry
14
representatives acting on behalf of regulated
15
industry. Dr. Mehta is retired
from Pfizer and Dr.
16
Maldonado is employed by Johnson & Johnson.
17
In addition, FDA would also like to note
18
that one member of the Psychopharmacologic Drugs
19
Advisory Committee, Andrew Leon, and an FDA
20
speaker, David Shaffer, were members of the
21
American College of Neuropsychopharmacology ACMP
22
Task Force that has recently issued a preliminary
23
report on SSRIs and suicidal behavior in youth.
24
This task force reviewed published and
25
unpublished data from controlled trials in youth,
18
1
data from epidemiological studies, and data from
2
autopsy studies.
3
Based on their preliminary review, they
4
concluded that the available evidence does not
5
suggest that SSRIs increase the risk of suicidal
6 behavior
in youth and with depression, however,
7
they acknowledge that their conclusions are
8
preliminary and they recommend that the pertinent
9
data available to pharmaceutical companies and FDA
10 be
rapidly made available to ACMP and others, so
11
that they may be independently evaluated.
12
In the event that the discussions involve
13 any
other products or firms not already on the
14
agenda for which FDA participants have a financial
15
interest, the participants' involvement and their
16
exclusion will be noted for the record.
17
With respect to all other participants, we
18 ask
in the interest of fairness that they address
19 any
current or previous financial involvement with
20 any
firm whose product they may wish to comment
21
upon.
22
Thank you.
23
DR. RUDORFER: Thank you.
24
To put the meeting in context, I would now
25
like to turn to Dr. Russell Katz, Director of the
19
1 FDA
Division of Neuropharmacologic Drug Products,
2 who
will provide a brief overview of the background
3
leading to today's deliberations and the likely
4
next steps.
5 Overview of Issues
6
DR. KATZ: Thank you, Dr.
Rudorfer, and
7
good morning. I would like to
also add my welcome
8 to
all of you here for this joint meeting of the
9
Pediatric Subcommittee of the Anti-Infective Drugs
10
Advisory Committee and the Psychopharmacologic
11
Drugs Advisory Committee.
12
In particular, I would like to welcome our
13
invited guests who are not members of the
14
committee, but who have graciously agreed to help
15 us
grapple with the difficult problem that we bring
16 to
you today.
17
As you know, we are here to discuss with
18 you
an issue of enormous importance and interest,
19
namely, the relationship, if any, between treatment
20 of
pediatric patients with antidepressant drugs and
21
suicidal behavior.
22
This has been an issue of extreme
23
complexity and we are here both to inform you of
24 our
efforts to date to examine the question and our
25
plans for further examination of the data, as well
20
1 as
to ask for your comments and advice about these
2
plans.
3
We come to you at this time for several
4
reasons. Under current law, the Agency is required
5 to
present postmarketing adverse event data to the
6
Pediatric Subcommittee for the first year of
7
marketing for those drugs granted market
8
exclusivity under the pediatric exclusivity
9
provisions of the Act.
10
At this time, therefore, the Agency is
11
meeting its obligation under the law to present
12
this data for Paxil and Celexa.
More importantly,
13
however, given the intense interest in the Agency's
14
efforts to examine the question of antidepressant
15 use
in pediatric patients and suicidal behavior, we
16
concluded that it would be appropriate to inform
17 you
about these latter efforts at this time, as
18
well.
19
As you know, we most recently became aware
20 of
a potential signal of concern during the review
21 of
the controlled trial data for Paxil. In
the
22
course of that review, we became aware that the
23
sponsor had categorized some events that could have
24
represented suicidal behavior or suicidal thinking
25
using a description that seemed somewhat
21
1
inappropriate.
2
We asked them to clarify their
3
presentation of the data, and their response raised
4 a
concern that such a signal existed.
Based on
5
these concerns, the Agency issued a public
6
statement in June of last year recommending that
7
this drug not be used to treat pediatric patients
8
with depression, but based on the Paxil data and
9 the
problem of idiosyncratic characterization of
10
events of potential concern identified in that
11
application, we asked the sponsors of the other
12
antidepressant drugs to search their controlled
13
trial databases in a more formal way to identify
14
potential cases of suicidal behavior.
15
Our review of their responses resulted in
16 a
second Agency statement that alerted
17
practitioners to a similar potential signal for
18
other drugs in this class, and recommended that
19
these drugs be used with caution in these patients.
20
Our continued review of these data,
21
however, convinced us that the data submitted from
22 the
various companies involved may not have been
23
collected or reported to us in a form that would
24
permit us to adequately evaluate the potential
25
relationship between these drugs and suicidal
22
1
behavior.
2
Indeed, we became convinced that with the
3
data before us at that time, we could not
4
adequately answer the question of whether there was
5
such a relationship for any specific drug or
6
whether there were any differences between drugs.
7
You will hear in greater detail later the
8
deficiencies with these data as previously
9
submitted and why we have therefore continued to
10
work with the sponsors involved to submit to us
11
data in the form that will permit us to adequately
12 and
comprehensively address the critical question
13
before us.
14
It is because we are not yet able to do
15
this that we could not present definitive analyses
16 at
this time. It is absolutely critical, in
our
17
view, that we make every effort to provide the best
18
answer possible to this question. The wrong answer
19 in either
direction, prematurely arrived at, could
20
have profound negative consequences for the public
21
health.
22
However, we now believe that we have
23
obtained from the sponsors all of the relevant data
24
collected during the trials, presented in a
25
standardized manner that will permit us to perform
23
1
analyses that will give us the best possible chance
2 to
address this question.
3 Before we embark upon these analyses,
4
however, we are taking this opportunity to inform
5 you
and the public about the problems we have
6
encountered in trying to answer this question, how
7 we
have attempted to address those problems, and to
8
describe our plans for analyzing the data.
9
We are primarily interested in your views
10
about our proposed approaches to the data and are
11
eager to hear if you believe we should request
12
additional data from the sponsors and whether you
13
believe we should perform additional analyses
14
beyond those we will describe to you later today.
15
In our efforts to further evaluate the
16
data, we have enlisted the help of outside experts
17
with particular expertise in the issue of pediatric
18
depression and suicide, and in particular, we have
19
enlisted a group from Columbia University, who will
20
objectively reclassify potential cases of
21
suicidality from all the drug development programs,
22 so
that we may move forward with our more
23
definitive analyses. You will
hear about this from
24 Dr.
Kelly Posner in more detail later.
25
We will also present the postmarketing
24
1
adverse event data for the drugs in question, but
2 as
you will hear, and for the reasons you will
3
hear, we do not believe that this data can
4
reasonably inform our judgment about any
5
relationship between these drugs and suicidal
6
behavior.
7
It is the controlled trial data that we
8
believe is best able to help us provide an adequate
9
answer to this question, but as you have heard, and
10 you
will hear throughout today's presentations, we
11 do
not believe that this data until now has been
12
provided to us in a way that would permit us to
13
interpret it fully.
14
It should be noted that this view of the
15
data has not been a unanimous one among Agency
16
staff. Some within the Agency
have examined the
17
data and concluded that the data, as currently
18
submitted, do permit definitive analyses and that
19
these analyses support the conclusion that this
20
class of drugs is associated with a risk of
21
suicidal behavior in pediatric patients.
22
However, the staff of the
23
Neuropharmacological Drugs Division has examined
24 the
individual cases reported by the sponsors that
25
allegedly represent suicidal behavior, and we are
25
1
convinced that the categorization of these events,
2 as
performed idiosyncratically by the individual
3
sponsors, is not entirely reliable.
4
Examples of these categorizations will be
5
presented to you later today, and we are confident
6
that this conclusion will become clear to you.
7
Further, the pattern of these potential
8
signals is also difficult to understand, for
9
example, arising from one single study out of
10
several similarly size studies for a given drug.
11
This unusual pattern gives us further reason to
12
more closely examine the data.
13
We are, of course, aware that there is
14
great concern among the families of children and
15
adolescents with depression about whether or not
16
these drugs can be used safely.
For them, I am
17
sure answering this question has already taken too
18
long.
19
We, too, are frustrated with the time it
20 has
taken to come to a definitive answer to this
21
question. Indeed, we had
originally hoped to be
22
able to present to you today more definitive
23
analyses and conclusions, however, as I have
24
described, closer examination of the data at each
25
step of our analyses convinced us that it would be
26
1
premature to arrive at a conclusion without
2
additional work, the plans for which we will
3
present to you later today.
4
We are firmly convinced that we serve no
5
one's goals or needs by rushing to a judgment that
6 has
not considered all reasonable sides to the
7
question. We are committed to,
and fully expect
8 to,
come back to the committee in late summer with
9 the
results of the analyses we will discuss today.
10
At that time, we expect to be able to
11
present the best possible answer that the current
12
data can provide to the question of whether or not
13 any
of these drugs, all of these drugs, or none of
14
these drugs increase the risk of suicidality in
15
pediatric patients.
16
With that as an introduction, I will turn
17 it
back to Dr. Rudorfer.
18
DR. RUDORFER: Thank you, Dr.
Katz.
19
We will now hear from Dr. Dianne Murphy,
20
Director of FDA's Office of Counterterrorism and
21
Drug Development, who will speak about the
22
Pediatric Drug Development Program.
23 Pediatric Drug Development
Program
24
DR. MURPHY: Welcome. Thank you very much
25 for
taking time to make this endeavor an important
27
1
part of your scientific and academic life. We hold
2
your advice very important and look very much
3
forward to your discussion.
4 [Slide.]
5
I am going to ask you to step back for a
6
moment. My comments are not going to focus directly
7 on
the topic of depression or the therapies for
8
that. The goal of my presentation
is to provide
9 you
some background on pediatric drug development
10
because I think you will see that is the process
11
that has brought us some of this data and we need
12 to
make sure everybody understands how this
13
evolved.
14
It is also an example of watch out what
15 you
ask for because we now finally, in the last few
16
years, are beginning to get the kind of information
17
that we wanted for a long time to be able to
18
understand how we could better treat children with
19 the
therapies that we have.
20
Of course, we will be reviewing FDA's
21
specific responsibilities during these activities.
22
[Slide.]
23
Acronyms. Throughout the day, you
will be
24
hearing these potentially. You
have FDAMA. That
25 is
the Food and Drug Administration Modernization
28
1
Act. This is important because
this is the
2
legislative initiative that provided the Agency
3
with the ability to provide an incentive that has
4
been a tremendous -- I call it the engine that has
5
really been driving this process for being able to
6
develop information on how to use these products in
7 children.
8
Remember, before this, most children, if
9 it
was not a pediatric disease like otitis media,
10
these products were not being studied in children,
11 and
each child was an n of 1 in which we did not
12
learn anything, and that was not an approach we
13
thought useful. That's FDAMA.
14
Best Pharmaceuticals for Children, renewal
15 of
the legislation basically expanding not only the
16
legislative mandate to look at products that have
17
patents remaining where the incentive will work,
18 but
a process which mandates FDA and NIH to work
19
together to develop the same sort of data for
20
products that are older and would not benefit
21
because that was an area that was not being
22
developed.
23
The way that is done is important to
24
understand because it is done via what is called
25 the
written request in which FDA -- and this is
29
1
distinctive from most other drug development -- FDA
2
determines what the public health need is and
3
issues a written request defining the studies that
4
they think need to be done, so that we can better
5
understand how to dose children or if it works in
6
children, or what are the distinctive adverse
7
events that occur in children, because as we all
8
know, the variability between a preemie and a
9
fullback is tremendous, and we have that in
10 children,
and evolving developmental processes.
11
PREA was the recently legislation that in
12
essence said yes, FDA, you have the authority to
13
require that if a sponsor submits an application
14 for
a disease -- I am going to call it indication
15
throughout the rest of this -- for an indication
16
that exists in children for which this product will
17
likely be used, you are to study it in children
18
also. You are not just to market
it for adults.
19
This proposed pediatric
study is a process
20
that applies to the written request, which if
21
industry is interested in studying a product, they
22 can
submit it to FDA, and we can look at that.
23
That is important because what you need to
24
understand is that this whole exclusivity process
25 is
voluntary, so it is up to the sponsor whether
30
1
they want to participate or not.
This process is
2
not.
3
[Slide.]
4
The interesting thing about pediatric drug
5
development is that many of the legislation that
6 has
developed has developed because of misfortunes
7 and
severe tragedies that have happened in
8
children, and yet every time new legislation would
9 be
mandated, it would apply to adults, and not to
10
children.
11
Many of you have heard this talk, so I am
12
just quickly putting these up here to remind
13 everybody.
14
[Slide.]
15
We have for decades been trying to have
16
products that are being used in children studied,
17 and
this is just to give you really the benchmarks,
18
starting in the '70s, in which the Academy of
19
Pediatrics issued a statement saying we ought to be
20
studying these products we are using in children,
21 why
do we think that children are going to be less
22
variable than adults. All reason
and information
23
would say they are going to be more variable, and
24 we
need to.
25
The Agency actually issued a statement
31
1
saying we think children should be studied, and we
2
would like you to conduct two adequate trials also
3 for
children, to evaluate the safety and efficacy
4 in
children.
5
What happened was not much, and as
6
everybody has heard, the majority of products were
7 not
studied in children until really here.
8
In 1994, FDA published a regulation which
9
basically said we understand that there are times
10 in
which you can extrapolate efficacy only.
If the
11
disease is similar enough, the pathophysiology, and
12 the
expected response have been defined well
13
enough, that you might be able to extrapolate
14
efficacy, hoping to incentivize in a way the
15
interest in developing information and conducting
16
trials in children. Safety and
dose finding were
17
still trials that you would need to conduct in
18
children.
19
Again, minimal response. So,
bottom line,
20 the
first incentive program was the major push.
21 The
FDA published a regulation, which was then
22
enjoined by a court saying we didn't have the
23
authority to require it, so Congress came back in
24
2003 and said, yes, FDA, you do.
25
So, right now here are the two things that
32
1 are
driving pediatric drug development, so that we
2 can
better understand how to use these products in
3
children.
4
[Slide.]
5
It has been a tremendous response.
This
6 is
just simulated to exclusivity. We have
received
7
over 300 proposals. You could
have counted the
8
number of products developed on your fingers and
9
toes before this that weren't primarily pediatric
10
diseases.
11
We have issued over 283 written requests
12
where FDA has determined what needs to be developed
13 in
the way of studies, and has issued sponsors'
14
requests. This is updated from your handout, by the
15
way, these numbers are slightly different because
16 we
updated it for the slides.
17
The important thing about exclusivity
18
determinations, it means that over 100 products
19
have been brought in with the studies that have
20
been requested, and you are discussing some of
21
those today, with the type of information that
22
helps us better understand.
23
We have an entire one-hour talk on some of
24 the
very significant findings that have been
25
developed, that we have discovered in this process.
33
1
Today is another example of we are finding out what
2
more information we need if we are going to
3
properly use these products.
4
I only put these numbers up because once
5
exclusivity is granted, you can see some were
6
denied, even though it may have been denied, it
7
still could have been approved.
It just meant that
8
they didn't meet the terms completely that we asked
9
for.
10
There are now 63 new labels, so products
11
that are being used in children, there are now 63
12 of
them that have new labels, new important dosing
13 and
safety information in them including
14
information that says they don't work in kids with
15
these studies.
16
[Slide.]
17
These are the products that were mandated,
18 not
the individual products, but the process that
19 was
mandated by the Best Pharmaceuticals, the BPCA.
20 I
point this out because one of these, our set of
21
data you are going to hear today is the result of
22
BPCA saying FDA, one year after a product has been
23
granted exclusivity, you will follow all of the
24
adverse events that are reported for that product,
25 and
you will present it to the Pediatric Advisory
34
1
Subcommittee that will soon be a full committee,
2 and
that this is an area which BPCA wanted to make
3
sure that additional attention was paid to the
4
process of reviewing what happens.
5
The thing to understand about that is that
6 a
product could be approved way back 10 years ago,
7 and
it could then be studied later in its life for
8
pediatrics, so that the one-year post-safety
9
assessment is at varying stages of these different
10
products, they are not all the same, and the
11
Division has tried to standardize that for you
12 today in looking at the safety assessments at
more
13
standardized times because each product is coming
14 in
at a different time.
15
[Slide.]
16
The only other thing I really wanted to
17
point out to everybody, to bring us back to the
18
topic at hand today, is that this drug development
19
process that has begun to occur really since 1998,
20
five, six years, has brought forth not only new
21
information that challenges some of our
22
preconceived thoughts about safety and how children
23
respond, it has been a tremendous bounty of
24
information because children are finally getting
25
studied.
35
1
We are beginning to have to figure out how
2 do
you measure that endpoint in children.
That
3
type of science was not being developed.
We are
4
also dealing with the ethical issues that come up,
5
that are different for kids who cannot consent, so
6
this is a whole different process, and I just want
7 to
make sure that you all knew that we have brought
8
various ethical issues to the committees, and we
9
have a wonderful cadre of ethicists who are Special
10
Government Employees, who work with the Pediatric
11
Advisory Subcommittee, who attended these meetings
12 and
advised us on such topics as should children be
13
enrolled in trials in which they are not going to
14
receive direct benefit, should children be enrolled
15 in
placebo-controlled trials, should children who
16 are
especially vulnerable -- most people think of
17
children as a vulnerable population, but in truth,
18
there are subsets, subpopulations that are even
19 more
vulnerable, and this was a population of
20
children with CP, how do you develop a product in
21
that population. These are
difficult issues.
22
[Slide.]
23
This is, quickly, and I am not going to go
24
over every one of these, but to give you an idea of
25 the
broad array of products that are being
36
1
developed in children and the questions that have
2
come up.
3
Actually, Neuropharm, the Division of
4
Neuropharmacological Drug Products, has brought a
5
number of these issues to the committee, including
6 how
do we develop pediatric products -- NIMH also
7
participated in this meeting -- from such issues as
8 --
also, this was another Neuropharm Advisory
9
Committee meeting with the Pediatric Committee --
10
chronic hepatitis, reflux in infants, HIV drugs,
11 how
do you approach the whole field of developing a
12
product that may be put in almost every newborn who
13
develops hyperbilirubinemia, tremendous issues,
14
long term study issues.
15
Again, more, what do you do about some of
16
these products. Most of our
products' safety
17
databases are collected on weeks, usually, maybe
18
months, but certainly not years, what do you do
19
with products that we know can potentially suppress
20
your adrenal axis or products that we know can be
21
oncogenic, but have to be used.
22 [Slide.]
23
Some of the ongoing lessons that we have
24
learned during this process -- which we think is a
25
positive process, it is much better than ignorance
37
1 --
it is that children are even more variable than
2 we
really thought.
3
We are finding, for certain classes, you
4 may
have to have dosing based on clearance in three
5
different age groups that is very different, and it
6 is
not just the preemies, it is not just the
7
neonates. It is actually children
of all ages,
8
from adolescence, preschool, et cetera.
9
Adverse reactions that are
10
pediatric-specific are being defined.
Clearly,
11
growth is one everybody would expect would be
12
defined, that we are finding that products, and
13
Prozac was an example of that, are having an effect
14 on
growth. But there are many other
products that
15 we
are beginning to look now, and beginning to look
16 in
a more systematic way, that we are finding that
17
they do have an effect on growth.
18
But there are other issues - school
19
behavior problem, other products where aggression
20 and
behavioral changes have been seen. So,
this is
21 a
very important area that we are trying to look at
22 as
we develop these products.
23
Trial designs are being modified as we
24
learn, and I think that is probably why we are here
25
today. We are learning. We take the best
38
1
knowledge we have, we get the best experts, we
2
issue the type of study we think will be the best,
3 and
sometimes something happens in the meantime,
4
more data becomes available, we need to update
5
that, or what we thought we were going to be able
6 to
evaluate didn't turn out to be as valuable as
7
something else in the study.
8
We learn from these studies.
Remember,
9
there is a huge amount of science that has not been
10
developed, that is now being developed for
11
children, and, as I said, the ethical issues have
12 to
be reassessed from the pediatric perspective.
13
[Slide.]
14
I just got the signal that my time is up,
15 so
I will leave you with the general principles
16
that we have developed from the International
17
Conference on Harmonization on how one should
18 approach
the whole process involving children in
19
trials, and this is a group that involves European
20
nations, Japan and the United States, and I think
21
that it is a shared responsibility.
That is why we
22
thank you for being here today.
Thank you.
23
[Slide.]
24
This is where you can go onto the web.
25
There is a tremendous amount of information posted
39
1 on
pediatric numbers, stats, and studies.
2
Thank you.
3
DR. RUDORFER: Thank you, Dr.
Murphy.
4
As Dr. Katz pointed out, an important way
5 to
put issues of drug safety in context is to
6
understand more about the disorder being treated,
7 so
we are pleased to have a couple of experts in
8 the
area of depression in young people to address
9 us
on the latest understanding of this complicated
10
disorder.
11
First, from Weill Medical College of
12
Cornell University, we are pleased to have Dr.
13
Cynthia Pfeffer, who will address Pediatric
14
Depression and its Treatment.
15
Pediatric Depression and its Treatment
16
DR. PFEFFER: I want especially to
provide
17 an
overview of pediatric depression, which in fact
18 is
a major mental health problem in the United
19
States and probably worldwide.
20
[Slide.]
21
There is a tremendous need to develop
22
treatments for these problems and also prevention
23
efforts primarily because these disorders,
24
particularly major depressive disorder, dysthymic
25
disorder, and for that matter, other mood disorders
40
1 are
very prevalent and recurrent, they have high
2
rates of morbidity and comorbidity, they are often
3
accompanied by very poor psychosocial outcomes for
4
children and adolescents. They
are associated with
5 high
risk for suicide and also for substance abuse.
6
[Slide.]
7
There are a number of problems which I
8
will touch on in my talk in reducing major
9
depressive disorder in children and adolescents,
10 and
these include problems in actually diagnosing
11
children and adolescents. There
are developmental
12
variations that need to be considered.
13
There is a complexity of factors that are
14
associated with the clinical course of children who
15
have such mood disorders and a need for specificity
16 of
treatments.
17
[Slide.]
18
Epidemiologically, we know that the
19
prevalence of major depressive disorder in children
20 who
are prepubertal is approximately 2 percent, and
21 it
increases in adolescents to a rate of between 4
22 and
approximately 8 percent.
23
The male-to-female ratio for younger
24
people, prepubertal children, is about equal, but
25 in
adolescents, females outnumber males who have
41
1
major depression 2 to 1.
2
By the time a youngster reaches the age of
3 18,
there is approximately a 20 percent prevalence
4 rate of those who are depressed, who show
major
5
depression, and since prior to World War II, each
6
successive generation seems to have a higher risk
7 for
major depressive disorder.
8
If we look at dysthymia, the prevalence
9
rate is somewhat lower although something to be
10
concerned about, with the highest rate of
11
approximately 2 percent in children, and in
12
adolescents, ranging from almost 2 to 8 percent.
13
Dysthymia is a condition that is often
14
under-recognized.
15
[Slide.]
16
There are a number of complexities in
17
diagnosing major depression in children and
18
adolescents. These include an
overlap of a variety
19 of
the mood symptoms, and in addition, the symptoms
20
often overlap with comorbid disorders.
21
There are developmental variations in the
22
symptoms and how they are manifest.
There are
23
etiological variations of mood disorders that do
24
involve gene and environmental interactions, and
25
there is a question of whether some of these issues
42
1 are
actually spectrum related or categorical
2
disorders.
3
Finally, the effects of medical
conditions
4 on
the prevalence and incidence of major depression
5 and
other mood disorders needs to be considered.
6
[Slide.]
7
The DSM criteria for major depressive
8
disorder involves a pervasive change in mood, which
9 is
manifest for at least two weeks by either being
10
depressed or irritable or having a loss of interest
11 in
pleasure.
12
There are other symptoms that are
13
necessary in making the diagnosis, that include
14
changes in appetite, weight, sleep, activity
15
levels, concentration, and sometimes
16
indecisiveness, changes in energy level,
17
self-esteem, including worthlessness and excessive
18
guilt, changes in motivation, and recurrent
19
suicidal ideation and acts.
20
These symptoms should represent a change
21
from the child or adolescent's previous functioning
22 and
produce impairment. These symptoms are
not
23
attributable to substance abuse, medications, or
24
other psychiatric illness, bereavement, and medical
25
illness.
43
1
[Slide.]
2
There are developmental variations which
3
have been identified. For
example, in children,
4
they tend to have a greater number of symptoms of
5
anxiety, including phobias and separation anxiety,
6
more somatic complaints, and if they do occur,
7
auditory hallucinations.
8
They express irritability with temper
9
tantrums and behavioral problems, and the children
10
tend to have fewer delusions and fewer serious
11
suicide attempts, however, adolescents tend to show
12
more sleep and appetite disturbances, if they
13
occur, delusional thinking, greater degrees of
14
suicidal ideation and acts, and greater impairment
15 of
functioning.
16
Compared to adults, however, adolescents
17
have more behavioral problems and fewer
18
neurovegetative symptoms.
19
[Slide.]
20
The diagnostic criteria for dysthymia
21
involves a persistent long-term change in mood
22
which is less intense, but more chronic than major
23
depressive disorder. These
children in adolescence
24
have extensive psychosocial impairment.
25
The depressed mood or irritability occurs
44
1
most of the time during the day for at least one
2
year, and there are at least two other symptoms
3
that are associated in making the diagnosis. These
4
include again changes in appetite, sleep, lowered
5
self-esteem, problems with concentration, problems
6 with
decisionmaking, changes in energy level, and a
7
sense of hopelessness.
8
People who have no symptoms for more than
9 two
months at a time, and do not have a major
10
depressive disorder in the first year of
11
disturbance, may be considered to have dysthymic
12
disorder, and these are also youngsters who never
13 had
manic or hypomanic episodes.
14
[Slide.]
15
Other symptoms tend to go along with
16
dysthymic disorder. These include
feelings of
17
being unloved, angry outbursts, self-depreciation,
18
somatic complaints, anxiety, and often
19
disobedience.
20
[Slide.]
21
There are a variety of variations that the
22
symptoms of major depressive disorder involve. For
23
example, psychotic depression, bipolar depressive
24
states, atypical depression, seasonal affective
25
disorder, subclinical or subsyndromal depression,
45
1 and
treatment-resistant depression.
2
[Slide.]
3
I will touch on some of these variants now
4
more specifically. Psychotic
depression includes
5
major depressive disorder symptoms that are
6 associated with mood-congruent or incongruent
7
hallucinations and/or delusions, and unlike
8
adolescents, children tend to manifest more
9
hallucinations.
10
Psychotic depression occurs in up to about
11 30
percent of those youngsters with major
12
depressive disorder. It is associated with more
13
severe depression, greater long-term morbidity,
14
resistance to antidepressant monotherapy, a low
15
placebo response, increased risk for bipolar
16
disorder, and a family history of bipolar and
17
psychotic depression.
18
[Slide.]
19
Bipolar depression presents similarly to
20
unipolar depressive disorder. The
risks for
21
bipolar disorder is indicated by psychosis,
22
psychomotor retardation, psychopharmacologically
23
induced hypomania, and a family history of bipolar
24
disorder.
25
Adolescents are likely to have rapid
46
1
cycling or mixed episodes, and an increased suicide
2
risk and difficulty in treatment compliance. There
3 is
a need to rule out bipolar II disorder, which is
4
more prevalent in adolescents and often overlooked
5 and
misdiagnosed.
6
[Slide.]
7
Atypical depression has not yet been
8
studied in children and adolescents, and it usually
9 has
an onset in adolescence, and it is manifest by
10
increased lethargy, appetite and weight changes,
11 and
reactivity to rejection.
12
There is hypersomnia and often
13
carbohydrate craving. In adults,
it tends to be
14
genetically distinct from major depressive
15
disorder.
16
[Slide.]
17
Seasonal affective disorder usually has
18 its
onset in adolescence in those living in regions
19
with distinct seasons. The symptoms are similar to
20
those of atypical depression, but are more
21
episodic. They do not include
increase reactivity
22 to
rejection.
23
This disorder should be differentiated
24
from depression precipitated by school problems and
25
school stress since it usually overlaps with the
47
1
school calendar.
2
[Slide.]
3
Treatment-resistant depression is not
4
clearly defined for children and adolescents. It
5
occurs in approximately 6 to 10 percent of
6
depressed children and adolescents who suffer
7
chronic depression.
8
In adults, treatment resistance is defined
9 as
patients who have had at least two trials with
10 two
different classes of antidepressants which are
11
administered at approximately similar doses for at
12
least six weeks each.
13
[Slide.]
14
Another issue that needs to be thought
15
about in understanding the mood disorders and
16
especially major depression is that they may be
17
affected by the complexity of comorbid disorders
18
which may affect the recognition and diagnosis of
19
major depression, the types and efficacy of
20
treatments, and various psychosocial outcomes.
21
[Slide.]
22
Comorbidity tends to be present in 40 to
23 90
percent of youth with major depression.
Two or
24
more comorbid disorders tend to be present in
25
approximately 20 to 50 percent of youth with major
48
1
depression.
2
Comorbidity in youth with major depression
3
involves dysthymia or anxiety disorders with a rate
4 of
approximately 30 to 80 percent, disruptive
5
disorders with a rate of approximately 10 to 80
6
percent, and substance abuse disorders with a rate
7 of
approximately 20 to 30 percent.
8
Major depressive onset is usually after
9 the
comorbid disorders except for substance abuse
10 in
which major depression tends to antedate
11
substance abuse disorders. Conduct problems may be
12 a
complication of major depression and may persist
13
after the major depressive episode resolves.
14
Children may manifest separation anxiety
15
comorbid disorders, while adolescents may tend to
16
manifest social phobia, generalized anxiety
17
disorder, conduct disorder, and substance abuse.
18
[Slide.]
19
In terms of differential diagnosis of
20
major depressive disorder, the complexities tend to
21 be
with an overlap of symptoms with other
22
nonaffective disorders, such as anxiety states,
23
learning problems, disruptive disorders, and
24
personality disorders and eating disorders.
25
The overlapping symptoms may include poor
49
1
self-esteem, demoralization, poor concentration,
2
irritability, dysphoria, poor sleep, appetite
3
problems, suicidal thoughts, and being overwhelmed.
4
[Slide.]
5
One should consider in the differential
6
diagnosis the nonaffective psychiatric disorders,
7
which include anxiety disorders especially
8
separation anxiety, generalized anxiety, and other
9
anxiety states, disruptive and attention deficit
10
disorders, learning problems, substance abuse,
11
eating disorders especially anorexia nervosa,
12
personality disorders, and premenstrual dysphoric
13
disorder.
14
[Slide.]
15
Another disorder that needs to be
16
considered and understood is an adjustment disorder
17
with depressed mood. This includes a mood change
18 and
impairment of functioning within about three
19
months of a stressor, and this does not meet the
20
criteria for major depressive disorder.
21
Adjustment disorder with depressed mood
22
tends to be self-limited, there are less mood
23
disturbances associated with it, fewer symptoms,
24 and
no relapse, which is an important issue.
25
Consider other disorders if the symptoms
50
1
last more than six months or meet the criteria for
2
other disorders, for example, dysthymia.
3
[Slide.]
4
General medical conditions may be another
5
complexity in understanding and diagnosing major
6
depressive disorder. These
medical conditions may
7 be
accompanied by symptoms of depression.
They may
8
also impact the course of major depressive
9
disorder.
10
Major depression can be diagnosed if the
11
depressive symptoms preceded or are not solely due
12 to
the medical condition or to medications used to
13
treat the medical condition.
14
The incidence of major depression tends to
15 be
higher in certain medical illnesses.
Chronic
16
illness may affect sleep, appetite, and energy.
17
Guilt, worthlessness, hopelessness, and suicidal
18
ideation are usually not attributed to the medical
19
illness, but do suggest the symptoms of major
20
depressive disorder.
21
Medical conditions that are often
22
associated with major depressive disorder include
23
cancer, hypothyroidism, lupus erythematosus, AIDS,
24
anemia, diabetes, and epilepsy.
25
Chronic fatigue syndrome is another
51
1
disorder that needs to be considered, but its
2
symptoms are similar to major depression, but there
3
tends to be more somatic symptoms, less mood,
4
cognitive, and social symptoms.
5
Medication-induced symptoms involve those
6
induced by stimulants, neuroleptics, cortical
7
steroids, and contraceptives.
8
[Slide.]
9
Bereavement is another issue that needs to
10 be
considered because there are a similarity of
11
symptoms with major depressive disorder.
The
12
diagnosis of major depression can be made if the
13
bereaved child or adolescent has moderate or severe
14
functional impairment, psychosis, suicidal thoughts
15 or
acts, and a prolonged course.
16
Following bereavement, a predisposition to
17
major depression may be related to prior major
18
depression or a family history of
major depressive
19
disorder. In general,
uncomplicated bereavement
20
often remits in 6 to 12 months after a death.
21
[Slide.]
22
I would like to focus now on some issues
23 of
clinical course for major depressive disorder.
24 The
median duration for clinically referred
25
children and adolescents tends to be 7 to 9 months,
52
1 and
in community samples it has been reported to be
2
shorter, approximately 1 to 2 months.
3
Predictors of a longer course or duration
4
involve the severity of depression, the degree of
5
comorbidity, the presence of negative life events,
6
parental psychiatric disorders, and poor social
7
functioning.
8
Remission of major depression is defined
9 as
a period of 2 weeks to 2 months in which there
10 is
one clinically significant symptom only.
Ninety
11
percent of children and adolescents with major
12
depression remit in 1 to 2 years after the onset of
13 the
major depressive episode.
14
[Slide.]
15
Approximately 6 to 10 percent of those
16
with major depression have a protracted course. A
17
relapse is an episode of major depression during
18 the
period of remission, and predictors of relapse
19
include the natural course of major depression,
20
namely, the nature of the way it manifests, lack of
21
compliance with interventions, negative life
22
events, rapid decrease, or discontinuation of
23
therapy.
24
Forty to 60 percent of youth with major
25
depression tend to have a relapse after successful
53
1
acute therapy, it's a high rate.
This indicates
2 the
need for continuous treatment.
3
[Slide.]
4
Recurrences occur also, and this is an
5
emergence of major depressive symptoms during a
6
period of recovery, which is an asymptomatic period
7 of
more than two months. Clinical and non-clinical
8
samples have a probability of recurrence of
9
approximately 20 to 60 percent within one or two
10
years after recovery, and 70 percent after five
11
years of recovery. So, this is a
chronic disorder.
12
Predictors of recurrence include the
13 earlier age of onset of major depressive
symptoms,
14
increased number of prior episodes of major
15
depression, the severity of an initial episode, the
16
presence of psychosis, the degree of psychosocial
17
stressors, the presence of dysthymia and other
18
comorbidities, and the lack of compliance with
19
therapy.
20
[Slide.]
21
In terms of the clinical course, children
22
with major depression, 20 to 40 percent develop
23
bipolar disorder in 5 years after the onset of
24
major depressive disorder, and predictors for the
25
bipolar disorder onset would be early onset of
54
1
major depression, the presence of psychomotor
2
retardation, psychosis, a family history of
3
psychotic depression, a heavy family loading for
4
mood disorders, and psychopharmacologically-induced
5
hypomania.
6
[Slide.]
7
Other factors that affect the clinical
8
course of major depression is that the risk for
9
depression increases 2- to 4-fold after puberty, a
10
very important developmental issue, and that
11
various genetic, as well as environmental, factors
12
influence the pathogenesis of major depression.
13
For example, shared family environmental
14 or
not extra-environmental non-shared issues tend
15 to
be very important in affecting the course, as
16
well as those youngsters who have high genetic risk
17 are
more sensitive to various environmental
18
stressors.
19
Children with depressed parents are three
20
times more likely to have a lifetime episode of a
21
major depressive disorder.
22
[Slide.]
23
The prevalence of children's first-degree
24
relatives when children have major depression tends
25 to
be 30 to 50 percent. In addition,
parents also
55
1 may
have major depression and anxiety disorders,
2
substance abuse, as well as
personality disorders.
3
[Slide.]
4
The clinical course of children with major
5
depression is also associated with poor school
6
success, low parental satisfaction with the child,
7 a
very important parent-child problem, learning
8
problems, other psychiatric disorders that
9
interfere with the child's learning.
10
The course may also be affected by various
11
personality traits, such as the child being
12
judgmental, having angry outbursts frequently, poor
13
self-esteem, and dependency.
Cognitive styles and
14
temperament, such as negative attributional styles,
15 may
affect the course of major depressive disorder.
16
Early adverse experiences, such as
17
parental separation or death, may affect the
18
course. Recent adverse events may
affect the
19
course, family conflicts, neglect, and abuse,
20
biological factors, such as inability to regulate
21
emotions, and/or distress.
22
[Slide.]
23
The relation of dysthymia in major
24
depression is quite important because dysthymia is
25
associated with an increased risk for major
56
1
depressive disorder. Seventy
percent of youth with
2
dysthymia tend to have major depressive disorders.
3
Dysthymia has a mean episode of
4
approximately 3 to 4 years for
both clinical and
5
non-clinical in community samples.
A first major
6
depressive episode usually occurs 2 to 3 years
7
after the onset of dysthymia, which may be
8
considered a gateway to the developing recurrent
9
major depressive disorder.
10
The risk for dysthymia is associated with
11
chaotic families, high family loading for mood
12
disorders particularly dysthymia.
13
[Slide.]
14
Another important issue in terms of course
15 of
children with major depression is that they are
16 at
very high risk for suicidal tendencies.
There
17 are
a few studies, some of which I will highlight,
18 one
by Marika Kovacs, which is a 9-year follow-up
19 of
prepubertal children. She had various
groups
20
that she studied.
21
At the time of follow-up, children who had
22
major depression had a 74 percent rate of suicidal
23
thinking and a 28 percent rate of suicide attempts.
24
Those who initially had dysthymia, also had a 78
25
percent rate of suicidal thinking, and close to 20
57
1
percent rate of suicide attempts.
2
Compared to children with adjustment
3
disorder or other types of psychiatric disorders
4
that are not mood disorders, these rates for
5
children with mood disorders, namely, major
6
depression and dysthymia, are significantly greater
7 for
suicidal thinking and suicidal attempts.
8
Our own follow-up study of 6 to 8 years
9 for
prepubertal inpatients indicated that there is
10 a 5
times risk for suicide attempt when the
11
prepubertal children reach adolescence if they had
12 a
prepubertal mood disorder.
13
[Slide.]
14
A community sample study indicated that
15 the
1-year incidence of suicide attempts in
16
adolescence was associated with a 12 to 15 times
17
greater risk if the youngster had major depressive
18
disorder.
19
[Slide.]
20
There are various concerns about treating
21
major depressive disorder. The
treatment research,
22
first of all, is relatively sparse in children and
23 adolescents. There are varied opinions about
24
whether psychotherapy or pharmacotherapy, or a
25
combination should be the first-line treatment.
58
1
The initial acute treatment often depends
2 on
the severity of symptoms of major depression,
3 the
number of prior episodes, the chronicity, the
4
age, contextual issues in the family, school, and
5
other environmental features, the degree of
6 negative
life events, the compliance with
7
treatment, prior treatment responses, and the
8
motivation for treatment.
9
[Slide.]
10
Some general principles that clinicians
11
have thought about is that psychotherapy may be
12
considered for the more mild or moderate major
13
depressive symptoms. Empirical
effect of
14
psychotherapies that we now know of include
15
cognitive behavioral therapy and ITP, interpersonal
16
psychotherapy.
17
Antidepressants may be used
for youngsters
18 who
have symptoms of major depressive disorder,
19
nonrapid cycling by polar states, psychotic
20
depression, depression with severe symptoms that
21
prevent effective psychotherapy or that fail to
22
respond to psychotherapy.
23
Also, due to the psychosocial context,
24
frequently pharmacotherapy alone may not be
25
effective.
59
1
[Slide.]
2
The treatment of children with major
3
depression, there are very few studies of acute
4
treatment using medication. There
are few
5
pharmacokinetic or dose-range studies with children
6 and
adolescents.
7 The SSRIs are thought to perhaps
induce
8
mania, hypomania, behavioral activation, which
9
might include impulsive behavior, silly or agitated
10
daring, and there are no long-term studies for the
11
treatment of major depression.
12
I am going to actually conclude, and not
13 go
over some of these studies, which you will hear
14
about I am sure today, and to say again that major
15
depressive disorder in children and adolescents is
16
complex and heterogeneous regarding its clinical
17
course, comorbidities, predictors, of course, need
18 for
specificity of treatment, and the developmental
19
variations.
20
It is a chronic condition that recurs with
21
serious morbidity including suicidal tendencies.
22
There are few treatment studies, which limit our
23
knowledge of the methods to reduce these symptoms
24 and
the morbidities.
25
There is a need to clarify the indications
60
1 for
pharmacotherapy, as well as psychotherapy
2
whether alone or used in combination, as well as
3
that to maintain youngsters who have already
4
exhibited major depressive disorder.
5
Thank you.
6
DR. RUDORFER: Thank you, Dr.
Pfeffer.
7
We will now turn to Dr. David Shaffer of
8
Columbia University who will speak on the topic of
9
Suicide and Related Problems in Adolescents.
10 Suicide and Related Problems in
Adolescents
11
DR. SHAFFER: Good morning.
12
[Slide.]
13
I am going to review the epidemiology of
14
youth suicide and also some of its phenomenology as
15 it
may be relevant to the discussion that you are
16
going to be having for the rest of the day. It is
17 a
topic that I have been involved in for a number
18 of
years, and I hope that it is helpful.
19
[Slide.]
20
In the United States, in 2001, the last
21
year for which we have statistics of this kind,
22
about 1,600 15- to 19-year-olds committed suicide.
23 You
will see that that is the third leading cause
24 of
death in the United States, and in most
25 countries, it is the second leading cause of
death,
61
1 but
in the United States and a few other countries,
2
homicide comes between that.
3
You can also see that suicide accounted
4 for
more deaths, over twice as many deaths as from
5
cancer, in fact, more deaths than all of the other
6
major physical conditions combined.
7
[Slide.]
8
The methods by which children commit
9
suicide are, by and large, very similar to those --
10
with children, young people -- are very similar to
11
those which are used by adults.
The main
12
difference is that hanging is somewhat more common
13 in
young people, and the figures that I have got
14
here on the left are the 5- to 19-year-olds, on the
15
right, over the rest of the population.
16
You will see a few other things of
17
interest. Ingestion is primarily
a cause of death
18 in
females, firearms are more common in
males than
19 in
females, and carbon monoxide poisoning is one of
20 the
few conditions where there have been any
21
changes in causes of death, so that the proportion
22 of
suicides attributable to carbon monoxide
23
poisoning has declined since the introduction of
24
catalytic converters. The
proportion of suicides
25
attributable to firearms, even though there has
62
1 been
a general decline in access and use of
2
firearms, has not declined.
3
You can also see from this slide that
4
cutting, which there is often a lot of debate about
5
cutting, whether that is or is not a form of
6
suicide, in fact, accounts for a very negligible
7
number of deaths. I think most
people would view
8
cutting as not being part of the suicide syndrome.
9
[Slide.]
10
This is a chart which shows the
11
distribution of suicide by different genders and
12
ethnic groups across the life cycle, and the top
13
line represents white males. That
is followed by
14
African-American males, then white females and
15
black females. Where the vertical
arrow is, is the
16
rate for adolescents.
17
You can see several things from this
18
chart. First of all, I should say
that this chart
19 is
remarkably similar in one country to another, so
20
there is something about this pattern of mortality
21
which seems to be almost independent of cultural
22
influence.
23
You do get very big differences in parts
24 of
Asia, but apart from that, it is remarkably
25
similar. That is to say that
there are very, very
63
1 few
suicides that occur before puberty, that
2
adolescents occupies an intermediate position
3
between childhood and adulthood, and then one gets
4
this very striking increase in the rate in elderly
5
males and relatively little variation by age in
6
females.
7
[Slide.]
8
If we deconstruct this a little more and
9
thus look at adolescents, what you can see is that
10
here, most 10- to 15-year-old suicides actually are
11
occurring amongst 14- and 15-year-olds, and that
12
suicide before puberty is very, very rare.
13
Sometimes you will read about big
14
increases or big changes in the young child rate,
15 but
the rates are very low and very unstable as a
16
result of that, and I don't think that one can draw
17
very many conclusions about suicide before puberty.
18
That may also be relevant to the matters
19
that you are considering today, because both
20
suicide and depression are relatively uncommon,
21
very uncommon before puberty, and that may mean
22
that what we should be looking at is what are the
23
differences between adolescents and adults.
24 [Slide.]
25
The United States ranks around about in
64
1 the
bottom of the top tier of rates in the world.
2
Most countries with the highest rates of suicide
3 are
in Northern/Eastern Europe, but the United
4
States is 16th as far as males are concerned, and
5
ranks 22nd as far as females.
6
There are quite big differences in gender
7
mainly in China, where suicide is the 7th country
8 for
female deaths, but much lower for male deaths,
9
but, in general, the United States is not
10
distinguished by having a particularly high or a
11
particularly low rate.
12
[Slide.]
13
We know quite a lot about the frequency of
14
suicidal ideation and attempts from large community
15
studies, particularly the Youth Risk Behavior
16
Study, which is a study that is carried out by the
17
National Center for Health Statistics every two
18
years, for which different states volunteer, and a
19
broad population of between 15- and 20,000 high
20
school students are interviewed using self-report
21
measures every two years.
22
[Slide.]
23
What one has been able to see from that
24
really was a big eye-opener. That
is to say, that
25
suicidal ideation in high school students is
65
1
extraordinarily common. Almost 20
percent of
2
American high school students will think about
3
suicide during the past year.
4
Suicide attempts are also very common, so
5
that the overall rate is about 9 percent, and if
6 you
track these YRBS results, they don't show an
7
awful lot of variation from one year to another.
8
I have highlighted by color the difference
9
between the self-reported attempts and attempts
10
that received medical attention, because only about
11 a quarter
of attempts do receive medical attention
12 or
are brought to medical attention.
13
I think what is important about this is
14
that adolescents may not disclose even suicidal
15
attempt behavior, let alone suicide ideation, and
16
that is frequently not known to either their
17
parents or to others, and that also has to be a
18
consideration, I think, in what you are
19
considering.
20
Both ideation and attempts, and attempts
21
which receive medical attention, are far, far more
22
common than completed suicide, and if you were to
23
array these out by gender, we estimate that there
24 are
about 4,000 suicide attempts for every female
25
suicide death, but about 400 male attempts for
66
1
every male death, so that you do get these big
2
gender discrepancies with attempts being more
3
common in females and deaths being more common in
4
males, but you can see that the ratio of attempts
5 to
deaths is extreme particularly in females.
6
[Slide.]
7
Not only do many adolescents attempt and
8
think about suicide, but they do it quite often, so
9
that from the studies that we have, about half of
10
suicide attempters will make only
one attempt a
11
year, and nearly a half will make two or more, in
12
many instances, four or more deaths per year.
13
We get similar findings in clinical or
14
community studies, and we do know from follow-up
15
studies that having made one attempt will increase
16 the
probability of another 15-fold, so that can be
17
quite an important consideration if you are
18
planning a medication study or any other kind of
19
therapeutic study, because maybe what you need to
20
find out about is not so much the state of
21
suicidality at the time of inception into the
22
study, but the history of suicidality as well
23
because that could be an important factor in either
24
stratifying for suicide risk or for filtering it
25 out
or filtering it in.
67
1
The episodes of ideation, again, you can
2 see
that most youngsters who think about suicide do
3 so
more than once a year, and in many instances, it
4 is
several times a year.
5
[Slide.]
6
With respect to how suicidal adolescents
7 are
excluded from psychopharm studies, because in
8
general, the studies of depression have excluded
9
suicidal instances, there have been variations in
10 the
techniques that have been used, there has been
11 no
uniform approach, and that may be a
12
consideration that the committee would want to look
13 at
in weighing up different studies and trying to
14
compare them.
15
[Slide.]
16
Finally, with epidemiology, I just want to
17
show you how the suicide rate has changed over the
18
last century. This is the 20th century youth
19
suicide profile.
20
What you can see is that starting I guess
21 in
the late '50s, the top line are males and the
22
bottom are females, the male youth suicide rate
23
started to increase, and it increased and increased
24
3-fold, finally, reaching some sort of asymptote
25
around in the late '80s, peaked a little bit more
68
1
towards the end, and then started to decline.
2
So, starting in 1994, we have had an
3
extraordinary decline in the youth suicide rate,
4
which is very interesting. It has
been parallel
5
twice before, once coinciding with World War I and
6
once with World War II. We don't
know what this
7
could be due to, and that will be something that I
8 am
going to return to in a second or two.
9
[Slide.]
10
As far as the causes of suicide, far and
11
away the most common finding in psychological
12
autopsy studies, which interview friends and family
13
after a death has taken place, are the very high
14
rates of diagnosable psychiatric illness that are
15
present, and in studies done in a variety of
16
locations, 90 percent of completed suicides were
17
diagnosable with a DSM diagnosis prior to their
18
death, and the rates are
extraordinarily similar
19
from location to location.
20
[Slide.]
21
The most common diagnoses are depression,
22
antisocial behavior, substance abuse, and some form
23 of
anxiety, and most teen suicides occur in 16- to
24
19-year-olds, and in that group, in 16- to
25
19-year-old male suicides, it is important to know
69
1
that two-thirds meet the criteria for substance or
2
alcohol abuse.
3
So, the occurrence of completed suicide is
4
very closely linked to the occurrence of
5
particularly alcohol abuse.
6
[Slide.]
7
As Cynthia Pfeffer outlined, and I won't
8
repeat this, suicidality is extraordinarily common
9 in
depressed children and teens, both at the time
10 of
diagnosis -- and this is a meta-analysis from
11 six
studies -- ideation was present in about 60
12
percent, a previous attempt in 30 percent, and
13
during the follow-up period, attempts also occurred
14
frequently, so that when you find ideation and
15
attempts during the course of treatment of
16
depression, as I say, this is a well-reported
17
phenomenon.
18
[Slide.]
19
There are other factors that predispose to
20
suicide. Imitation is one that is
particularly
21
worrying because it means that public information
22
campaigns may have a double-edged sword, because we
23 do
know that you do get suicide epidemics in the
24
young.
25
There is a contagion factor, and the
70
1
Centers for Disease Control are very actively
2
engaged in trying to find ways of reducing this,
3 and
there are now a host of studies in adults, but
4 not
yet in children or adolescents, that show that
5
biological abnormalities may predispose to
6
impulsive responses to stress and a family history
7 of
suicide.
8
[Slide.]
9
We can devise a schema, which you have got
10 in
your handout, which can show the route from any
11 of
these disorders to suicide ideation and from
12
there to suicide, but I don't
think that there is
13
time to get into that model in this presentation.
14 [Slide.]
15
I just want to go back to changing rates,
16
because they may be very relevant to today's
17
discussion.
18
[Slide.]
19
As I showed you, there has been this very
20
striking and encouraging reduction in male suicide
21
males amongst young males 15 to 24.
It is even
22
more striking actually if you look at 15- to
23
19-year-olds.
24
What is important is that this has not
25
been a United States phenomenon only.
It has been
71
1
reported in a large number of other industrialized
2
nations.
3
In the list that I have given here, three
4
nations, Austria, Germany, and Switzerland, have
5
been experiencing a decline which well predated the
6
introduction of any of the newer groups of
7
antidepressants, but in all of the other countries,
8 the
decline started sometime after 1988.
9
There is only one country
which seems to
10
have a stable or rising rate, which is Scotland,
11 and
there are a number of possible reasons that
12
have been debated to explain these reductions.
13
One is that during the '90s, at least in
14 the
United States, there was economic prosperity, a
15
decline in unemployment, and other social indices
16
tended to improve, but rates also started to
17
decline in high youth unemployment countries in
18
Europe, and the relationship between SES and
19
suicide is not strong, and, in fact, it hasn't
20
really been established.
21
The first thought was if so many suicides
22 are
associated with drug and alcohol abuse, maybe
23
exposure to drugs and alcohol would have been
24
reduced during this time, and this is certainly my
25
first guess. However, use and
abuse rates have not
72
1
changed, if anything, they have continued to inch
2 up.
3
[Slide.]
4
Reduced firearm availability, the Brady
5 Act
was introduced in 1994, and there is evidence
6
from tracking studies that ownership and use of
7
firearms started to decline around about 1980, but
8 the
proportion of suicides by firearm has gone
9
unchanged, and although there have been very
10
striking declines in accidents attributable to
11
firearms, it is not clear that we can point to the
12
reduction in suicides as being caused by that.
13
Also, the declines have been noted in
14
countries in which there are almost no firearm
15
suicides, so this doesn't seem to be a very
16
plausible explanation.
17
[Slide.]
18
More psychotherapeutic treatment is a
19
possibility, but, in fact, the data seem to suggest
20
that visits for psychotherapy have declined
21
consistently over the past 10 to 12 years, more
22
psychopharmacologic treatment, and you will have
23
heard that there has been an enormous increase in
24
exposure to antidepressants during this period in
25
many countries, or it could be a nonspecific
73
1
finding, a better recognition of adolescent suicide
2
with some nonspecific interventions or some
3
combination of the above.
4
[Slide.]
5
A word or two about treatment.
There have
6
been some useful Cochrane analyses looking at
7
effective treatments for suicide attempts. These
8
have mainly been done in adults, and only two
9
treatments emerged as being successful.
10
One is dialectical behavior therapy, which
11 is
a very specific form of therapy which is
hard
12 to
come by because very few people are trained in
13 it,
and one study looking at flupenthixol, which is
14 an
antipsychotic or neuroleptic, in multiple
15
attempters.
16
There have also been studies showing
17
lithium or at least discontinuation of lithium
18
results in an increase in the suicidality, and
19
Clozaril seems to have a specific suicide sparing
20
effect in schizophrenia.
21
But apart from that, we don't have much to
22
guide us, and there is nothing out there which
23
tells the clinician what to do with this very
24
common problem.
25
[Slide.]
74
1
Maybe that is why, but, in general, teens
2 who
do commit suicide tend to be relatively
3
undertreated compared to adults, so that, for
4
example, the top three lines show that between 30
5 and
60 percent of adults who commit suicide will
6
have had mental health treatment, but in
7
adolescents, very few have had that, so it is
8
getting between 7 and 21 percent, they are an
9
undertreated group.
10
[Slide.]
11
Furthermore, one of the things that has
12
been interesting to epidemiologists over this
13
current debate is do you find antidepressants in
14
toxicologic studies of completed suicides, and Exen
15
[ph] in Sweden has done a study showing that the
16
findings in autopsy studies suggest that suicides
17 are
significantly undertreated with SSRIs compared
18 to
the rest of the population.
19
There has only been one study in youth,
20 and
that is from the Utah Youth Suicide Study by
21 Dr.
Gray, and he has looked at 50 psychological
22
autopsies, all of whom had careful toxicology
23
investigations.
24
A quarter of those had been prescribed
25
antidepressants, but in none of those cases were
75
1
antidepressants found at autopsy, so we know that
2
teenagers often don't take their medication, and
3
certainly they didn't seem to be taking it in this
4
case.
5 [Slide.]
6
So, I would just like to conclude with
7
some cautions and considerations.
Ideation and
8
attempts are very common in depressed teens, and
9
they recur frequently, so finding them in
10
youngsters being treated for depression is, of
11
course, not surprising. That
doesn't address any
12
treatment effect that might be found.
13
A methodological point. Teenagers
often
14
conceal ideation and attempts unless they are asked
15
about them directly. Self-report
facilitates
16
disclosure. It is my
understanding that we are
17
heavily dependent upon event reports in these data,
18 and
event reports may be influenced by the mode of
19
elicitation.
20
They are not used with a
glossary which
21
precisely defines how things should be classified,
22 so
misclassifications can occur.
23
Self-harm is a term that is used by some,
24 but
not others in the mental health profession.
It
25 is
a very heterogeneous descriptor and not all
76
1
types of self-harm are associated with suicidal
2
intent.
3
There have been no direct studies with
4
frequent and careful measurement examining whether
5
SSRIs increase, decrease, or have no effect on
6
suicidal ideation and behavior, so that we are
7
dependent very much on inference, but maybe that is
8
always the case.
9 I just would like to conclude with the
10
following. After increasing for 35 years, teen
11
suicide rates have been declining consistently in
12
many countries. During this
period, there has been
13 a
marked increase in exposure of teens to SSRI
14
antidepressants.
15
These trends could be related.
This is
16
ecologic, and we don't know whether they are
17
related, but at the moment we don't have a better
18
explanation for the turnabout of a condition that
19 led
to the death of tens of thousands of young
20
people.
21
I would like to stop at that point.
22
DR. RUDORFER: Thank you very
much.
23
At this time, just before our break, I
24
have one announcement to make.
Any open public
25
hearing speakers who have not yet signed in, please
77
1 do
so immediately. We will only be able to
call
2
upon speakers who have formally signed in, so we
3
wouldn't want you to miss your chance.
4
We have time for a 15-second break, but I
5 am
told that may not work, so why don't we take 5
6
minutes or as close to that as we can work, and we
7
will come back for our open public hearing.
8
Thanks.
9
[Break.]
10 Open Public Hearing
11
DR. RUDORFER: There is specific
guidance
12
from the FDA that I would like to read.
This
13
applies to all meetings or considered general
14
matters meetings, and as we heard earlier from
15
Anuja, since we are not focusing on one specific
16
product here, that encompasses this joint meeting.
17
Both the Food and Drug Administration, or
18
FDA, and the public believe in a transparent
19
process for information gathering and
20
decisionmaking. To ensure such
transparency at the
21
open public hearing sessions of the Advisory
22
Committee meeting, FDA believes that it is
23
important to understand the context of an
24
individual's presentation.
25
For this reason -- and I am addressing the
78
1
speakers this morning -- FDA encourages you, the
2
open public hearing speaker, at the beginning of
3
your oral statement to advise the committee of any
4
financial relationship you may have with any
5
company or any group that is likely to be impacted
6 by
the topic of this meeting. For example,
the
7
financial information may include a company's or a
8
group's payment of your travel, lodging, or other
9
expenses in connection with your attendance at the
10
meeting.
11
Likewise, FDA encourages you at the
12
beginning of your statement to advise the committee
13 if
you do not have any such financial
14
relationships. If you choose not
to address the
15
issue of financial relationships at the beginning
16 of
your statement, it will not preclude you from
17
speaking.
18
As I mentioned earlier, the clock dictates
19
only a limited amount of time for each speaker. I
20
would like to run all night, but I hear an ice
21
storm is coming, so in the interest of time, we
22
have a light warning system, and each speaker,
23
please be advised, when you see the yellow light,
24 you
have 30 seconds remaining, so please start to
25
wrap up.
79
1
The flashing red light means you are out
2 of
time and the microphone will go off. I
have
3
asked them to let you finish your sentence for
4
three or four words, but it is out of our hands.
5
We have two speaker-ready chairs, so I am
6
asked to remind you that when your two away from
7
your number, please be sure you are in one of
8
those.
9
Speakers are assigned by number and we
10
will begin with Number 1.
11 Irving Kirsch and David
Antonuccio
12
DR. KIRSCH: My name is Irving
Kirsch.
13
Baum, Hedlund has paid for my air tickets. I
14
decided to come before knowing that.
15
Dr. David Antonuccio, Amanda Drews, and I
16 are
reviewing the published literature evaluating
17 the
efficacy of antidepressants in depressed
18
children. A total of 12
randomized, controlled
19
clinical trials have been published.
20
Two-thirds of these trials failed to find
21 any
significant benefit of medication over inert
22
placebo. Only 4 trials reported
significant
23
differences, and these did so only on
24
clinician-rated measures, not on patient-rated
25
measures.
80
1
When the data from these trials are
2
combined, the placebo response is found to be 87
3
percent of the drug response.
This means that the
4
drug effect is only 13 percent of the drug
5
response. This is not a
clinically significant
6
effect.
7
Many children get better when given
8
antidepressants, but the data indicate that this is
9
largely a placebo effect. These conclusions
are
10
consistent with those found in 7 previous published
11
reviews.
12
To summarize, the published clinical trial
13
data show that the therapeutic benefits of
14
antidepressants for children is negligible at best.
15
David.
16
DR. ANTONUCCIO: These results
were drawn
17
from studies with design flaws that typically favor
18 the
study drug. For example, they frequently
19
exclude placebo responders before random
20 assignment,
rely on ratings by clinicians who have
21 a
vested interest in the outcome, and are likely to
22 be
unblinded by medication side effects.
23
Furthermore, these results are drawn from
24 the
published literature which is subject to
25
publication bias and file drawer problems meaning
81
1
that many studies with negative results do not get
2
published. Adding unpublished
studies, most of
3 which have negative results, will surely
shrink the
4
difference between antidepressants and placebo even
5
further.
6
In order to evaluate the cost
7
effectiveness of antidepressant use in children,
8 the
committee must consider the benefits, as well
9 as
the risks. Clinically meaningful
benefits have
10 not
been adequately demonstrated in depressed
11
children, therefore, no extra risk is warranted.
12
An increased risk of suicidal behavior is
13
certainly not justified by these minimal benefits.
14
Neither are the established increased risks of
15
other commonly reported side effects, which include
16
agitation, insomnia, and gastrointestinal problems.
17
The highest possible standard should be
18
applied to scientific data involving drug treatment
19 of
children, because children are essentially
20
involuntary patients. Those of
you on the
21
committee who are parents know this to be true
22 because
when your children have prescription
23
medication for something that ails them, you make
24
them take it as prescribed whether they want to or
25
not.
82
1
Children given antidepressant
medication
2
often do get better, but so do children given
3
placebo. Thus, the clinical data
suggest the
4
improvement is due primarily, if not entirely, with
5
placebo effect.
6
Please be careful to ensure that our
7
children are not exposed to risk without
8
commensurate benefit.
9
DR. RUDORFER: Thank you.
10
May we have the next speaker, Number 2.
11 Lisa Van Syckel
12
MS. SYCKEL: Good morning, ladies
and
13
gentlemen. My name is Lisa Van Syckel, and my
14
daughter, Michelle, at the age of 15, was placed on
15
Paxil. She was diagnosed with
depression and
16
anorexia nervosa. It turned out that
that
17
diagnosis was wrong, she actually had Lyme Disease.
18
My daughter self-mutilated, became
19
psychotic, became violent, attempted suicide twice.
20 My
daughter survived those two suicide attempts,
21 not
because of the drug, because of the police
22
officers who were summoned to my home.
23
Michelle has suffered severe withdrawal.
24 She
is constantly ill with flu-like symptoms.
She
25 has
had rectal bleeding, she has vomited blood.
83
1 She
has had her friends at school call her
2
"Psycho," all because she was misdiagnosed and all
3
because everyone has withheld from the public the
4
adverse effects of Paxil.
5
I am a parent. It is my right to
make an
6
informed decision on behalf of my daughter. You
7 did
not allow me to make that informed decision and
8 she
was harmed. We are blessed because
Michelle
9 did
not die, and Michelle is now attending
10
university and doing beautifully.
11
Please, have respect for our children,
12
make sure that you put proper warnings on these
13
medications. Our children's lives
are at stake
14
here, because not only does it cause suicide, it
15
also causes them to become violent, very, very
16
violent.
17
Thank you.
18
DR. RUDORFER: Thank you.
19
May we have the next speaker, Number 3.
20 Ann Blake Tracy, Ph.D.
21
DR. TRACY: I would like to say,
first of
22
all, that this is a meeting that should not be
23
taking place today. I testified
at an FDA hearing
24
similar to this in 1991, and these drugs should
25
have been banned at that time in my opinion.
84
1
I am Dr. Ann Blake Tracy, a Ph.D. in
2
health sciences with emphasis on psychology. I
3
have spent the last 14 years researching the SSRIs
4 and
working with patients who are having adverse
5
reactions to these medications. I
am also the
6
author of Prozac: Panacea or Pandora, Our Serotonin
7
Nightmare.
8
I have testified in criminal and civil
9
cases for 12 years concerning these medications,
10 and
I am greatly concerned about the use of these
11
drugs among children, with developing brains, who
12
have far more reactions than the general public
13
would, as I am the elderly who are having severe
14
adverse reactions.
15
What I presented to the FDA in 1991, I
16
would like to present again. Each
of you will get
17 a
copy of this. This is a 31-year-old
patient on
18
Prozac for six months, shows the patient, although
19
appearing alert and functioning, in a total
20
anesthetic sleep state while dreaming.
I believe
21
technically, you could call that a REM sleep
22
behavior disorder.
23
The research now shows, this many years
24
later, that 86 percent of the cases being diagnosed
25
with this REM sleep behavior disorder are patients
85
1 on
antidepressants, 80 percent of those on SSRI
2
antidepressants.
3
There are some very famous cases that I
4
believe manifest that very clearly, and in
5
representing those families today, I would give you
6
Andrea Yates, who drowned her five children while
7
taking Effexor and Remeron.
8
DR. RUDORFER: Thank you. I am afraid we
9 are
out of time now.
10
DR. RUDORFER: Thank you.
11
Number 4, please.
12 Tom Woodward
13
MR. WOODWARD: My name is Tom
Woodward.
14 My
wife Kathy and I have been married for 19 years
15 and
until 6 months ago had 4 children. Our
oldest
16
child, Julie, hung herself after 7 days on Zoloft,
17 and
she was only 17, was a cautious child, and had
18 no
history of self-harm or suicide, nor was there
19 any
history of depression or suicide in our family.
20
The doctors we spoke with stressed that
21
Zoloft was safe and had very few side effects. The
22
possibility of violence, self-harm, or suicidal
23
acts was never raised. The two
and a half pages we
24
received with the Zoloft never mentioned self-harm
25 or
suicide.
86
1
Julie began experiencing akathisia almost
2
immediately. We now know from a
blood test from
3 the
coroner's office that she was not metabolizing
4 the
drug.
5
We are 100 percent convinced that Zoloft
6
killed our daughter. We are here
because we
7
believe the system we have in place is flawed. It
8 is
clear that the FDA is a political entity and its
9
leadership has protected the economic interests of
10 the
drug industry. Under the Bush
administration,
11 the
FDA has placed the interests of the drug
12
industry over protecting the American public.
13
Dr. McClellan understands how important
14
political contributions are particularly since his
15 mother has headed up the Republican
fund-raising in
16
Texas. Eighty-six percent of the
$14 million in
17
political contributions given by drug companies has
18
gone to the Bush administration Republican
19
candidates - what did Pfizer, Eli Lilly, and
20
GlaxoSmithKline Beecham buy?
21
The FDA should be a jealous advocate in
22
protecting the American people.
Those in
23
leadership positions within the FDA must be beyond
24
reproach. FDA's chief counsel Daniel
Troy has
25
spent his career defending the drug industry.
87
1
Suppressing unfavorable data may be legal, but is
2 it
ethical?
3
If the trials don't favor a drug, the
4
public never hears of them. Legal
maneuverings
5
have thrown out the scientific method.
The drug
6
industry must be compelled to produce all of their
7
findings and studies. I also
believe public
8
funding of these trials is warranted.
9
Our daughter, Julie, had been excited
10
about college and scored 1,300 in her SATs several
11
weeks before her death. Instead of picking out
12
colleges with our daughter, my wife and I had to
13
pick out a cemetery plot for her.
14
Instead of looking forward to visiting
15
Julie at school, we now visit her grave.
The loss
16 we
have experienced is horrific. We don't
want
17
another innocent child or family to suffer this
18
tragedy.
19
DR. RUDORFER: Thank you, Mr.
Woodward.
20
May we have the next speaker, please.
21 Mark Miller
22
MR. MILLER: My wife Cheryl and I
23
desperately hope that our story, along with others
24
that you will hear today, and I so proud of the
25
teens and the young adults who you will hear from
88
1
today, that they have the courage to come forward
2 and
talk with you personally. I wish our son
3
could, he cannot.
4
There is a serious problem with the way
5
SSRI medications are being prescribed today and
6
how, in many cases, they can directly cause
7 violence
and suicidal behavior in those we love and
8
treasure the most, our children.
9
You see, we lost our 13-year-old son,
10
Matt, in the summer of 1997. He
died after a
11
psychiatrist we did not know gave him three sample
12
bottles of a pill we had never heard of, for a
13
perceived illness that his doctor could only guess
14 at.
15
We were advised with great authority that
16
Matt was suffering from a chemical imbalance that
17
could be helped by a new, wonderful medication
18
called Zoloft. It was safe,
effective, only two
19
minor side effects were cautioned with us -
20
insomnia, indigestion.
21
Now, I don't know if Matt had a chemical
22
imbalance. I do know this. We had moved into to a
23 new
neighborhood a year before, a new school
24
setting, he was uneasy. He didn't
have the friends
25 he
had grown up with in our old neighborhood.
Yes,
89
1 our
son was unhappy.
2
So, Matt's doctor, a man we know through
3
court testimony to have been a well-paid spokesman
4 for
Pfizer, gave us Zoloft. He said,
"Take these
5 for
a week, call me back when you know how Matt is
6
doing."
7
Matt didn't have a week. He
became
8
agitated on the pills. He did not
sleep. He did
9 not
eat. He could not sit still. That night, a
10
Sunday, before leaving on vacation, after taking
11 his
7th Zoloft tablet, he took his own life.
12
This is important for you to know.
Matt
13
hung himself from a bedroom closet hook, barely
14
higher than he was tall. To
commit this
15
unthinkable act, something he had never attempted
16
before, never threatened to any family member,
17
never talked about, he was actually able to pull
18 his
legs up off the floor and hold himself that way
19
until he lost consciousness and forced himself to
20
leave us.
21
Matt's autopsy showed the levels of
22
sertraline in his blood were three times the
23
therapeutic minimum levels.
24
You have an obligation today, this panel,
25 to
prevent this tragic story from being repeated
90
1
over and over and over again. I
hope you will do
2 the
right thing.
3
DR. RUDORFER: Thank you, Mr.
Miller.
4
If we could have the next speaker, please.
5 Corey and Jay Baadsgaard
6
MR. COREY BAADSGAARD: Good
morning. My
7
name is Corey Baadsgaard. Four
years ago I was
8
diagnosed with having social anxiety disorder, and
9 my
family practitioner doctor, he prescribed Paxil
10 20
milligrams.
11
After about 8 1/2 months, I started taking
12 40
milligrams of Paxil because it was not working
13 at
20 milligrams. A few months after that,
I went
14
back. The same problem, it wasn't
working, and he
15
suggested I start taking a new medication called
16
Effexor.
17
He abruptly discontinued the Paxil and put
18 me
immediately on Effexor at 75 milligrams, and I
19 was
supposed to work up to 300 milligrams over a
20
3-week period. The day that I
took the 300
21
milligrams, I didn't feel very well and I stayed
22
home from school.
23
I went back to sleep and that evening I
24
woke up in a juvenile detention center.
Unaware of
25
what I had actually done, I asked one of the
91
1
members of the juvenile detention center, and I
2
found out that I had taken my high-powered rifle
3 that
I use for hunting to my third period class,
4
took 23 of my classmates hostage and 1 teacher
5
hostage.
6
I spent 14 months in jail, not really
7
knowing why I had been there, not really
8
remembering anything that I had done.
9
This whole thing has changed my whole
10
family, it changed me, myself. We
were forced to
11
move. I cannot even go back to
the same town that
12 I
lived in, I have to stay at least 25 miles away
13
from city limits.
14
These drugs are ridiculous. They
should
15 not
be prescribed unless it's absolutely last
16
resort.
17
MR. JAY BAADSGAARD: These drugs
are hell.
18
Look at what they have done to my son.
19
DR. RUDORFER: Thank you.
20
May we have the next speaker, please.
21 Joyce Storey
22
MS. STOREY: My son, Brian Storey,
was 17
23
years old in 1997. Our family
doctor diagnosed him
24
with severe depression. He took blood, checked
25 for
drugs or any medical condition. He found
92
1
neither. He gave me 14 Zoloft
pills and said come
2
back in two weeks. He never told
me they had side
3
effects and he even said if a person is drinking or
4
doing drugs, that Zoloft works well with them.
5
Five days later, my son killed a woman.
6
When they arrested him, he was drug-tested. They
7
found no illegal drugs, he was only on Zoloft.
8
During his trial, the kids that testified with him
9 and
against him said he did no drugs or alcohol.
10
The psychiatrist that examined him was Dr.
11
James Merkangis from Connecticut.
He is also a
12
Doctor of Neurology and is on the faculty at Yale
13
University. He said Brian had a
manic reaction to
14
Zoloft. He testified Brian told
him it was like
15
being in a dream.
16
The news media called my son the
17
All-American boy, and he was. He
is now serving
18
life without parole. Six months
later, another boy
19 at
my son's high school, Jeff Franklin, 17 years
20
old, on Prozac, took an ax to both his parents and
21
three of his brothers and sisters.
Both of his
22
parents died. He is serving two
life sentences.
23
This is not a coincidence. There
is a
24
common denominator, teenager, severely depressed,
25 on
an SSRI antidepressant. What is scary is
that
93
1 you
are only hearing from a few of us that this has
2
happened to, and there are a lot more out there.
3
I am praying you will look at these drugs
4
very closely and, at the very least, take them out
5 of
the hands of pediatricians and GPs.
These
6
doctors are not psychiatrists, and they do not have
7 the
knowledge and experience in treating mentally
8 ill
children.
9 My son never had a chance. There are 13
10
million people on these drugs, 6 to 8 million are
11
children. The question is why are
we handing these
12
drugs out like candy, and the answer is $17 billion
13 a
year business. It is always about
money. Please
14
help before more families are destroyed.
15
Thank you.
16
DR. RUDORFER: Thank you.
17
Next speaker, please.
18 Jame Tierney
19
MS. JAME TIERNEY: Good
morning. My name
20 is
Jame Tierney. I was 14 years old when I
was
21
prescribed 75 milligrams of Effexor for migraine
22
headaches. I took this for about
a year. At the
23
time, the drug lost its effectiveness and my doctor
24
doubled the dose.
25
For the next 9 months, my life as I had
94
1
known it was gone. I thought
daily about suicide
2 and
hurting myself. I felt void of normal emotions.
3 I
was so belligerent, agitated, and filled with
4
hate - hate for my family, my friends, and most of
5 all
myself. Rage consumed me. I felt trapped.
6
I said and did things I had never done
7
before and never would do now. I
had little
8
control and little inhibition. It
was as if I was
9
watching a movie and some villain was destroying
10 all
the relationships around me. I spent my
time
11
alone and viciously fighting with my parents. They
12
would ask what was wrong and what had happened to
13
me. I could not answer them
because I did not know
14 or
understand myself. I was terrified.
15
I thank God my parents knew that wasn't
16
really me and continued to search for answers.
17
They found the answer to my uncharacteristic
18
behavior. It was the Effexor that
my neurologist
19 had
prescribed for my migraine headaches. I
was
20
not, repeat not, prescribed this drug for
21
depression. I have had no history
of depression
22
prior to or after I was off the Effexor.
For me,
23
this drug caused the very symptoms it's supposed to
24
alleviate.
25
Due to the severe withdrawal symptoms,
95
1
Prozac was used to get me off Effexor.
It worked,
2 but
the same personality and behavior problems
3
reemerged. Effexor and Prozac
affected me the same
4
way. I had never had these feelings
before I took
5
Effexor, I have never had these feelings since I
6
stopped taking the Effexor and Prozac.
7
Effexor took three years from me and I
8
will never get them back. The
horror of what these
9
drugs did to me is ineffable.
These drugs are
10
destroying lives everywhere.
11
I implore you to please protect the
12
children from these drugs.
13
DR. RUDORFER: Thank you very
much.
14
If we can have speaker Number 9, please.
15 Donna Taylor and Mark Taylor
16
MS. TAYLOR: Hi. My name is Donna Taylor.
17 My
son was shot at Columbine. He took 7 to
13
18
bullets though his chest and nearly died. I also
19
have other members of the family that have died
20
since then on these drugs, but we can't get into
21
that right now, and many, many people that we know,
22
that families have been divided and separated, and
23
there is just all kinds of divorces and all that
24
going on from these drugs.
25
I will let Mark speak.
96
1
MR. TAYLOR: First of all, I would
thank
2 you
for allowing me to come and speak on behalf of
3 the
thousands of innocent Americans that have died
4 as
a result of these drugs.
5
I would like to start with an opening,
6
very famous statement, and it says, "The measure of
7 a
man is not his strength or how much money he has,
8 or
how good he looks or how strong he is, or how
9
powerful he is. The measure of
the man is how
10
noble he is."
11
I want to ask you guys, are you really
12
being noble with your choices, or are you just
13
allowing the drug companies to squeeze by you just
14
because they have a big pocketbook.
This is
15
ridiculous.
16
Do you people have children, do you, do
17 any
of you? Have any of you had anyone that
has
18 died
on these drugs? If you have, I am amazed
that
19 you
guys are even standing here supporting these
20
drug companies.
21
I mean this has never happened in the
22
history of America. This is a
shame and it ought
23 to
be stopped today, not next week.
24
MS. TAYLOR: And God says the same
thing.
25
It's in the Bible, Revelations 18, 19 through 24
97
1
makes it clear, sorcery means anarchy in the last
2
days and blood will be running all over the
3
streets.
4
MR. TAYLOR: Say yes to America's
health
5 and
no to the drug companies.
6
DR. RUDORFER: Thank you both.
7
We are going to move on to speaker Number
8 11,
Shannon Baker.
9 Shannon Baker
10
MS. BAKER: My name is Shannon
Baker and I
11
have no financial ties to the pharmaceutical
12
industries, nor am I here to complain about my
13
daughter's side effects, adverse reactions, or
14
withdrawal symptoms. I am here
because she is no
15
longer alive.
16
I know you have all got pictures.
I am
17
here because today, I am representing the love that
18 my
daughter had for life and to be her voice and
19 the
voice of all the other children who their
20
voices have been silenced by these drugs.
21
Their deaths have been so senseless and
22
needless. I am here speaking in front
of you,
23
hoping that you will go the right direction and ban
24
these drugs for children. There
needs to be no
25
more senseless and needless deaths because of these
98
1
drugs.
2
Thank you.
3
DR. RUDORFER: Thank you.
4
Our next speaker, Number 12, please.
5 Dawn Rider
6
MS. RIDER: My name is Dawn Rider
and I am
7
here to tell you my story, and I represent, as
8
president of ASPIRE, more than 11,000 persons who
9 are
all named on the Eli Lilly and Prozac petition,
10
which a copy has been given to the panel.
11
We have been educated to believe that
12
mental, emotional, and behavioral disorders are
13
caused by chemical imbalances in the brain. The
14
fact is that this is only theory, and this theory
15 is
pushed on us as if it were the absolute truth.
16
The reality is that the best of scientists
17 do
not completely understand the complex inner
18
actions of the myriad chemicals in our brains.
19
Those of us who elect to believe this theory and
20
subject ourselves to treatment become guinea pigs
21 in
an ongoing experiment.
22
I know this from personal experience.
I
23
trusted our family doctor when he explained that
24
depression is caused by a chemical imbalance. We
25
trusted him when he determined that Paxil was right
99
1 for
my husband, and Prozac for my son.
2
We weren't educated enough at that time to
3 ask
him to provide us with the test results that
4
proved which chemicals were being balanced.
5
I am not going to go into details of what
6
happened to our family. I have
given you all
7
documentation, it's very painful.
Suffice it to
8 say
that my beautiful 14-year-old son is now dead,
9 and
when we discovered the problems with these
10
drugs, we decided it would be better for my husband
11 to
suffer through depression than end up dead like
12 our
son, and we found out that he could not get off
13 of
Paxil.
14 He went through over a year of hell
before
15 he
was able to finally withdraw from the drug, and
16 in
the process it destroyed our marriage of over 20
17
years.
18
I say with no apology whatsoever that
19
these SSRI drugs destroyed what was once a loving
20 and
vibrant family. Why do we believe that street
21
drugs like heroin and LSD can lead to outcomes such
22 as
this, yet, we won't accept that legally
23
prescribed drugs, working on the same
24 neurochemicals, can result in horrific crimes
25
against persons and property?
100
1
Why do we accept that a drug like
2
penicillin, beneficial as it is for some, can prove
3
fatal for others? We fail to
accept that these
4
drugs can have paradoxical effects.
These drugs
5 are
not safe for everyone.
6
They should be labeled with the strongest
7 of
precautions and dispensed only by trained
8
physicians who have time to adequately monitor the
9
patient. Most doctors do not have
time for this
10
level of care.
11
Also, patients should be required to sign
12
letters of informed consent.
Please carefully
13
consider the documentation that I have left with
14 you
and look at the faces of those that are here
15
today and the faces that out in the hall, those
16
children who cannot speak for themselves because
17
they are dead. They are not
merely anecdotal
18
evidences.
19
There is a preponderance of evidence that
20
will be presented before you today.
Please
21
consider it carefully and do the right thing.
22
Thank you.
23
DR. RUDORFER: Thank you.
24
We are up to Number 13.
25 Sara Bostock
101
1
MS. BOSTOCK: I have slides, so
please
2
look at the screen.
3
My daughter Cecily had only been taking
4
Paxil for two weeks before she died, during which
5
time her condition greatly worsened.
6
By the day of her death, was pale, unable
7 to
sleep, almost unable to converse, and in a
8
frightened, agitated state, jumping at the
9
slightest noise. That night she
got up and without
10
turning on any lights, went into our kitchen only
11 40
feet from where I was half asleep. She
stabbed
12
herself twice in the chest with a large chef's
13
knife. The only noise was a
slight yelp and a
14
thump when she fell on the floor.
15
This was a young woman who had everything
16 to
live for. She had just completed
applications
17 to
grad school and received a large pay increase
18 the
month before.
19
She had a boyfriend who loved her and
20
scores of wonderful friends. She
had never been
21
suicidal. To die in this violent,
unusual fashion
22
without making a sound after the
marked worsening
23 of
her condition led me to believe that Paxil must
24
have put her over the edge.
25
Her autopsy revealed she had a very high
102
1
blood level of Paxil, which reflects poor
2
metabolization and is a feature common to many of
3
these suicides. I believe this
induced an
4
intensely dissociative state, perhaps even
5
sleepwalking. SSRIs suppress
rapid eye movement
6 and
block the muscle paralysis which occurs in this
7
stage of sleep.
8
The whole regulation of waking, sleeping,
9
dreaming occurs in the brain stem where the
10
serotonin neurons are clustered and where SSRIs are
11
having their impact. Patients
taking SSRIs had
12
rapid eye movement during non-REM sleep and while
13
awake when they were not paralyzed.
This atypical
14 REM
is often associated with strange behaviors
15
including hallucinations.
16
The effects of SSRIs on sleeping, waking,
17
unconsciousness itself are ill understood. From
18
accounts of people under the influence of these
19
drugs, I believe SSRIs can alter consciousness in
20
some mysterious and frightening way that is not
21
normally seen even in mental illness.
I am certain
22
this is what happened to my daughter.
23
Untold thousands have died because of the
24
drug companies and the FDA's failure to heed the
25
evidence over the past 17 years.
103
1
DR. RUDORFER: Thank you.
2
Again, I apologize for the short time.
3
Number 14, please.
4
Vera Hassner Sharav
5
MS. SHARAV: I am Vera Sharav and
I am
6
president of the Alliance for Human Research
7
Protection.
8
The family testimonies that you are
9
hearing today are not anecdotes. They
are
10
corroborated by a Harvard review of children's
11
medical charts, which found that within three
12
months of treatment on an SSRI, 22 percent suffered
13
drug-induced adverse psychiatric effects, and
14
overall, 74 percent of children suffered adverse
15
events during the course of treatment.
16
The FDA has known for years, but failed to
17
reveal that antidepressants consistently fail to
18
demonstrate a benefit in children.
At least 12 of
19 15
trials failed. The FDA has known and
failed to
20
warn physicians and the public that SSRIs increase
21 the
risk of suicide and hostility in children.
22
FDA's 1996 Zoloft review found "7-fold
23
greater incidence of suicidality in children
24
treated with Zoloft than adults."
The British Drug
25
Regulatory Authority reviewed the evidence, which
104
1 is
not being shown in this meeting, and they
2
determined that the risks far outweigh any
3
benefits. They took action to
protect children.
4
When is the FDA going to take action?
5
The FDA is foot dragging, equivocating,
6 and
tinkering with definitions while children are
7
dying. The San Francisco
Chronicle reports that
8 the
FDA has barred its own medical reviewer who
9
reviewed more than 20 trials involving 4,000
10
children, and his findings confirmed the British
11
finding, which is that SSRIs increase the risk of
12
suicide.
13
DR. RUDORFER: Thank you.
14
If we could have speaker 16, please.
15 Cynthia Brockman
16
MS. BROCKMAN: Thank you for
allowing me
17 to
address you about the 1999 Zoloft-induced drug
18
reactions that my son Chris had at 16, resulting in
19 a
woman's death and a life sentence for him.
20
My son and I want to express sincere
21
sorrow for that death. Our
sympathies also extend
22 to
all victims of SSRI's deadly mind-altering
23
effects.
24
The medical community has tolerated mental
25
health care in which patients are worse off after
105
1
treatment than before with the worst cases ending
2 in
death.
3
I urge you to ban SSRI use in children,
4 and
not to let another life be destroyed by lack of
5
adequate SSRI regulation.
6
Chris took Zoloft or Adderall,
7
deteriorated from drug-induced akathisia, could not
8
bear adverse symptoms of inner turmoil, loss of
9
conscious behavior. He described
overpowering drug
10
effects, his uncontrollable fits of anger, pitches
11 and
voices setting him off, not wanting to be
12
touched, feeling horrible all over his body, not
13
being in reality.
14
After his offense, his drug reactions
15
stopped, went off all SSRIs for about a year, but
16 restarted
when depressed and put on Zoloft again.
17
Prison doctors ignored warnings, forced him to take
18
harmful drugs drugging him into hallucinating,
19
irrational, suicidal state.
20
May 2002, I met with the Texas House
21 Committee
on Corrections who ordered prison doctors
22 to
correct this health crisis caused by these
23
drugs. Various drugs had
triggered severe
24
suicidal, homicidal symptoms for about two years in
25 a
clinical setting of doctors starting and stopping
106
1 his
meds.
2
When doctors stopped all drugs, all
3
symptoms disappeared. Doctors
released Chris as
4
recovered from the prison psych hospital to a
5
regular unit May 2003. Chris has
not had any psych
6
drugs since.
7
These clinical events show dangerous
8
reactions caused by SSRI-induced psychosis through
9
challenge, de-challenge, re-challenge.
Medical
10
experts said Chris would not have been suicidal,
11
homicidal had he not been reacting to SSRI drugs.
12
Dr. O'Donnell concluded Chris' offense was
13
from combined toxic drug effects which altered
14
behavior, enhanced violent thoughts and actions,
15
impaired judgment, was unable to form intent.
16
Citizens Commission on Human Rights
17
confirmed SSRIs caused his symptoms.
Now Chris
18
take omega-3 fatty acids and fish oil to restore
19 his
mental health that was damaged from SSRIs.
He
20 is
doing well without medications, and I thank
21
Jesus Christ for that.
22
Please ban these drugs and their use in
23
children.
24
Thank you.
25
DR. RUDORFER: Thank you.
107
1
We will move on now to Number 18, please.
2 Todd Shivak
3
MR. SHIVAK: Good morning. We are Todd
4 and
Eileen Shivak. We do not have any
financial
5
relationship to anyone here.
6
Our story is much like the cases everyone
7
else here today is bringing forward to you.
8
Our son Michael was 11 when he was
9 prescribed
Paxil for what was diagnosed as
10
depression. The consequences of
this still live
11
with us today. Thank God he is
alive and with us
12
today, but Michael is afraid of his doctors, how
13 can
he trust what they will give him next.
14
He is afraid of the police. He
has been
15
wrestled down, handcuffed and taken to jail. The
16
police are supposed to protect us and look what
17
they have done to him.
18
It is difficult for him to trust his
19
teachers. They still look at Michael as a
20
troublemaker even though he currently is an A/B
21
student with much improving grades.
His peers
22
still think of him as a freak, the kid who tried to
23
slash his wrists while in class.
24
As parents, our most important job is to
25
protect our kids. We thought we
were doing the
108
1
right thing. The doctors
convinced us that taking
2
these drugs was the only thing that we could do for
3
Michael. Now, Michael wonders
whether we are going
4 to
have him arrested, sentenced, physically
5
restrained and punished again. If
he can't trust
6 his
parents, who can he trust?
7 Our daughter, Catherine, was 5
years old
8 at
the time. She witnessed firsthand some
of the
9
most terrifying sights that I have ever had to deal
10
with. Our family is finally
getting back to the
11
loving family we once were, but the fear of what
12
happened still haunts us.
13
Worse yet, how could all the doctors not
14
recognize what was happening?
Michael saw three
15
different social workers, two different
16
psychiatrists, and went through at least four
17
different emergency room psychological evaluations
18 in
two different hospitals.
19
We are here to plead that you do something
20 to
stop the prescriptions of these drugs, so that
21 no
one else has to go what we are all going
22
through. It is impossible to
describe the pain and
23
utter helplessness we all felt watching Michael
24
suffer, watch him cry, take up weapons against us,
25 and
beg us to let him die. How do you erase
the
109
1
picture of your child trying to run in front of a
2
moving car?
3
Please save our children from this drug.
4
DR. RUDORFER: Thank you.
5
If we can have speaker 19,
please.
6 Andy Vickery
7
MR. VICKERY: Good morning. My name is
8
Andy Vickery and I am a trial lawyer from Houston,
9
Texas. For the last eight years,
I have
10
represented parents who lost their children to
11
suicide induced by these drugs.
You have heard
12
from two of my clients this morning already and
13
will hear from another.
14
I only have two minutes and I can tell you
15 a
lot more than two minutes. The title of
the
16
paper that I filed with you is "Needle in the
17
Haystack." I applaud your
desire to look at the
18
randomized clinical trials comprehensively to see
19 if
they confirm the signal that Dr. Katz
20
acknowledged exists.
21
I applaud that, however, I am concerned as
22
Lilly was told in 1990 that you are looking for a
23
needle in a haystack, you are off on a wild goose
24
chase. These trials were not
designed to detect
25
suicidality, they did not use the Beck Suicide
110
1
Ideation Scale which would make the kind of refined
2
measurements that the epidemiologist gentleman who
3
spoke earlier said are needed.
They did not use
4 the
Barnes Scale, as Dr. Mann himself had
5
recommended in a '91 article to measure treatment
6
emergent akathisia.
7
They weren't designed to answer the
8
problem, and in 1990 or '91, when Lilly met -- and
9 you
have the handwritten notes of this in the
10
materials I gave you -- when they met with outside
11
consultants including Dr. Jerold Rosenbaum, he
12
said, "There is a data problem, you are looking for
13 a
needle in a haystack."
14
Find these vulnerable people and
15
rechallenge them. Please look at the way Lilly
16
sought to study this issue in 1990.
They followed
17 a
protocol by Charles Beasley that said don't use
18 RCTs,
don't use epi studies, find these people and
19
rechallenge them. That was done
by Anthony
20
Rothschild who said these patients need to be
21
reassured it's not them.
22
In the meantime, because the signal is
23
there, please issue warnings; while you look at the
24
data, issue warnings.
25
DR. RUDORFER: Thank you.
111
1
We are up to speaker 20.
2 Rosie Carr Meysenburg
3
MS. MEYSENBURG: My name is Rosie
Carr
4
Meysenburg. I am from Dallas, Texas.
I have no
5
financial ties with anybody but my husband of 40
6
years.
7
In my handout, I have highlighted what I
8 am
speaking about here.
9
The first paper is a personal letter from
10 Dr.
Peter S. Jensen. At that time, he was
the head
11 of
Child & Adolescent Disorders Research Branch of
12 the
NIMH, the National Institute of Mental Health.
13 He
said that research indicates that
14
antidepressants for depressed adolescents are not
15
very effective.
16
The second paper is a personal letter from
17 Dr.
Larry S. Goldman, Director of the AMA, the
18 American Medical Association. He writes physicians
19
have known for many years the dangers of giving any
20
antidepressant to patients with certain disorders.
21
There is a substantial risk of precipitating mania
22 or
psychosis.
23 The last item is a journal article
from
24 the
Journal of Clinical Psychiatry researched at
25
Yale University. It states that
11 percent of all
112
1
psychiatric hospital admissions were from
2
antidepressant-induced mania and psychosis.
3
It also states another area of research
4
that would be relevant to this issue is the work of
5
Winter and colleagues showing that Prozac and other
6 SSRIs can simulate the effects of LSD. In other
7
words, this is saying for some people, taking an
8
SSRI is the same as taking LSD.
9
About two million people enter a
10
psychiatric hospital every year, 11 percent then is
11
over 200,000 people a year who have an
12
antidepressant-induced psychosis and who are
13
hospitalized. Not all are
hospitalized. Some of
14
them have either committed suicide, a homicide, or
15 a
murder/suicide.
16
DR. RUDORFER: Thank you.
17
Number 21, please.
18 Rachel Adler
19
MS. ADLER: Mr. Chairman, I
respectfully
20
request that my entire remarks be entered in the
21
record. My name is Rachel Adler. I am on the
22
board of directors of the Child and Adolescent
23
Bipolar Foundation, CABF, a parent-led,
24
not-for-profit organization, that is the leading
25
source of public information for pediatric bipolar
113
1
disorder.
2
Board members Sheila McDonald and John
3
Adler are here with me, as well.
4
Bipolar disorder may emerge with an
5
episode of major depression, an illness which often
6
causes suicidality even in preschoolers.
Children
7
with depression are at a high risk to switch to
8
bipolar disorder.
9
We surveyed 17,000 members last month and
10
received a 15 percent response rate over a 5-day
11
period. Eighty-nine percent of
the respondents
12
report that their child had been treated with an
13
antidepressant.
14
We have received favorable comments, but
15
some responses indicate that in some subgroups of
16
children, suicidal ideation and behavior may emerge
17 for
the first time or worsen when a child is given
18 an
antidepressant. Some of these children
perhaps
19
have a vulnerability to a bipolar disorder.
20
For these reason, CABF urges the FDA to
21
require manufacturers to add a black box warning on
22 the
labeling for antidepressants to alert
23
clinicians and parents to the possibility that
24
antidepressants can trigger and worsen suicidality,
25 as
well as mania or rapid cycling bipolar disorder
114
1 in
some children.
2
CABF opposes any ban on the off-label use
3 of
these or other psychiatric medications in
4
children because many of our members report them to
5 be
necessary and even lifesaving for their children
6
with mood disorders especially when used in
7
combination with a mood stabilizer.
8
CABF also urges the pharmaceutical
9
industry and the Federal Government to fund
10
research to analyze what factors are shared by
11
those children who, according to parent reports,
12
became suicidal shortly after taking an
13
antidepressant.
14 Finally, CABF calls upon the
15
pharmaceutical industry and the National Institutes
16 of
Health to make public all safety and efficacy
17
data from unpublished studies in children.
18
I would also like to say that what I am
19
hearing is a lot of people blaming a medication for
20
what happened to their children, and have a direct
21
blame. What I would sort of like
to see is more
22
trained psychiatrists who actually know the side
23
effects as well themselves and who are talking to
24 the
parents, telling them about the possibility of
25
side effects, about that depression inherently, you
115
1
know, can result in suicidality and that the
2
medication might increase that.
3
But to blame the medication itself that
4 has
helped so many people and has also prevented so
5
many suicides, I don't think is the right way, but
6 we
do need to have much more clinicians guiding our
7
patients and parents, so that they know what kind
8 of
side effects are possible.
9
Thank you.
10
DR. RUDORFER: Thank you.
11
We are going to move on to Number 23.
12 Pepper Draper
13
MS. DRAPER: Good morning. My name is
14
Pepper Draper. I am a Director of
the
15
International Coalition for Drug Awareness. I have
16
absolutely no financial gain. I
do this completely
17 100
percent voluntary, and the reason for that is
18
because of my own son's problems.
19
My child was prescribed Ritalin, which
20
became very depressed, and we bought into the whole
21
serotonin theory, so we were naturally raising that
22
serotonin, which unfortunately started causing him
23 to
become severely depressed and suicidal.
24
Unfortunately or fortunately I should say
25 is
that we were able to finally understand the
116
1
truth about serotonin, that raising serotonin and
2
stopping the metabolism of it has caused suicide
3 and
aggression, and that is well documented.
4
Unfortunately, Dr. Tracy was not able to
5
talk about that, but what I want to share with you
6 is
that there is going to be others here from
7
Arizona who are going to share with you how
8
wonderful these drugs have been for the State of
9
Arizona, but I am here to tell you that I deal with
10
these people every day who are tired of their
11
mental health workers putting them on another
12
medication and another medication and another
13
medication, until these children are now being put
14 in
mental hospitals at an enormous rate.
15
They are being given electric shock
16
therapy and it is very tragic what I am seeing, and
17 I
just want to share with you that I know that if
18 we
will teach them the right ways to take care of
19
their bodies and cut out the things that are
20
addictive, like these medications are, that we can
21
help our youth learn to deal with what is going on
22 in
their lives, and I just want to share with you
23 one
last thing.
24
I am really saddened that the fact that
25
every single parent cannot share what has happened
117
1 to
their child because if they could, my mother
2
would be here, standing up here, sharing what has
3
happened to her adult son.
4
DR. RUDORFER: Thank you.
5
If we could have speaker 24, please, Dr.
6
Marks.
7 Donald Marks, M.D., Ph.D.
8
DR. MARKS: Good morning. My name is Dr.
9
Donald Marks and I address your subcommittee as a
10
prescribing physician, as a father, and as a former
11
associate director and director for clinical
12
research for two multinational pharmaceutical
13
companies. I am here at my own
expense because I
14
believe in the importance of these issues.
15
SSRI manufacturing and sales is serious
16
business with tens of millions of patients in the
17
U.S. and a market in the tens of billions of
18
dollars.
19
My experience working for pharmaceutical
20
companies is that any attempt to decrease sales by
21
increasing warnings will be met with severe
22
organized resistance. SSRI drugs
are mostly
23
prescribed by primary care physicians who have
24
limited time with patients, limited training in
25
childhood and adolescent neuropsychiatry and
118
1 neuropsychopharmacology,
and minimal time to
2
evaluate properly patient suitability and response
3 to
pharmacologic versus non-pharmacologic
4
interventions.
5
The seriousness and severe adverse event
6
effects of SSRI drugs make their use hardly
7
justified in the majority of cases because SSRIs
8 are
well known to have limited efficacy over
9
placebo and against non-pharmacologic treatments.
10
There are many studies in the peer
11
reviewed medical literature supporting the causal
12
role of serotonin in disinhibition and violence.
13 My
own prescribing experience with SSRI drugs and
14
evaluation of numerous cases referred to me has
15
revealed significant agitation and aggression,
16
akathisia, activation of mania and hypomania,
17
increased depression, serious dependency and
18
withdrawal difficulties, suicidal ideation, and
19
toxic interactions with other drugs.
20
It is important to be aware that these
21
symptoms of SSRI toxicity can be mistaken for the
22
progression of the underlying mental state being
23
treated, leading to use of more of the same and
24
other offending SSRI drugs rather than to
25
withdrawal of the causative SSRI agent.
119
1
This creates coding problems for
2
physicians, coding problems by clinical researchers
3 and
sponsoring companies reporting adverse events
4 in
SSRIs.
5
SSRI manufacturers, such as Glaxo and
6
Pfizer, have conducted clinical trials in depressed
7
children, many of which show no efficacy against
8
placebo, and this has led to an increased warning
9 in England
that Paxil should not be prescribed as
10 new
therapy for depressed children under the age of
11 18.
12
DR. RUDORFER: Thanks. I am sorry we are
13 out
of time, but thank you, Dr. Marks.
14
We are going to move on to speaker 25.
15 Leah Harris
16
MS. HARRIS: Good morning. My name is
17
Leah Harris and I am here at my own expense.
18
The two minutes I have to speak will not
19
permit me to go into the details of what I suffered
20
while taking Prozac, Paxil, and Zoloft from age 12
21 to
18. I provided additional information in
my
22
submitted written statement.
23
I went from being a shy and mildly
24
depressed, but never suicidal kid to being overcome
25
with thoughts of hurting and killing myself while
120
1 on
the SSRI drugs, thoughts which I acted on.
2
Since quitting SSRIs over a decade ago, I
3
have never again self-mutilated or had suicidal
4
thoughts. All other things being
equal, the
5
suicidality simply vanished. For me, this is clear
6
proof that the drugs must have played a role, and I
7 am
one of the lucky ones, I have survived to tell
8 the
tale.
9
I am not an anecdote and my story is not
10
anecdotal evidence. As a
tax-paying American
11
citizen who was hurt by these drugs throughout my
12
childhood, I demand that the FDA take seriously the
13
British decision of December 2003 banning all SSRIs
14
except Prozac for use in children.
15
Please consider all the evidence
16
especially that which the pharmaceutical industry
17
does not want you to see. The FDA must take action
18 now
regarding this grave issue of public health.
19
Yes, many people claim to be helped by
20
these drugs, and that is wonderful, but what about
21
those of us who are harmed?
Medical professionals
22 and
the public must be informed of the very serious
23
risks that are associated with SSRIs.
24
In light of these risks, at the very
25
least, isn't it time for the FDA to require that
121
1 the
drugs be labeled with clear warnings that might
2
save lives? Such warnings may
negatively affect
3
sales, as Dr. Marks referred, which may not please
4 the
pharmaceutical industry, but the FDA was
5
created as an independent regulatory agency to
6
serve the interests of the American public, not Big
7
Pharma.
8
American children are no less precious
9
than British children, and they are in need of our
10
protection, too.
11
Thank you.
12
DR. RUDORFER: Thank you.
13
We are up to speaker 26.
14 Donald Farber
15 MR. FARBER: I am Donald Farber of Marin
16
County, California. I am a
plaintiff's attorney.
17 I
have represented antidepressant victims for five
18
years.
19
As a lawyer, I look at the evidence, too.
20 I
hear the emotional stories, but I look at the
21
evidence.
22
On January 27th, six days
ago, I got a
23
writing that I have been waiting for the FDA for 15
24
years, from GlaxoSmithKline.
Attempted suicides on
25
Paxil during all premarketing testing were
122
1
frequent, placebo, it was actually rare, but due to
2 the
fact they manipulated the figures in the
3
re-analysis of the data, it was infrequent. So,
4
even by this standard, we should have had a warning
5 12
years ago. What do we have to do to get
a
6
warning?
7
Dr. Katz mentioned the re-analysis of the
8
data. I call it tinkering with
the data.
9
Here is what happened to Paxil to get it
10
approved. Dr. Laughren knows about these figures.
11
Here is what happened with the tinkering of the
12
data before and after.
13
Look at the difference. These are
not
14
lawyer figures, these are their figures.
They
15
manipulated the data, Paxil suicides went down,
16
placebo suicides, which is the key figure here for
17 you
mathematicians, went way up, so that the result
18 was
statistical insignificance by the time the PDAC
19 met
in October of '92.
20
Whether the drugs go on the market or not,
21
they have to be given a warning.
I am for full
22
disclosure. I am not for banning
these drugs, but
23 I
want full disclosure, and the FDA doesn't need a
24
citizens' petition to do their job.
25 Finally, I do object to this entire
123
1
meeting. I would venture that 95
percent of you
2 are
pro-industry and it is time for people like Joe
3
Glenmullen and Peter Breggin to sit on this
4
committee as well as you distinguished people.
5
Thank you.
6
DR. RUDORFER: Thank you, sir.
7
Could we have speaker 27, please.
8 Lorraine Slater
9
MS. SLATER: Informed parental
consent is
10
only possible as long as full disclosure is made by
11 the
pharmaceutical companies, the FDA, and the
12
medical community.
13
How can you imagine I feel as Dominique's
14
mother knowing now that I was slowly poisoning my
15
daughter every day as I was dispensing her
16
antidepressant medication including Celexa and
17
which she made her first suicide attempt after
18
being on it for almost one month, and effects of
19 the
last medication she was on when she did commit
20
suicide?
21
Yes, Dominique's mind and behavior were
22
slowly being altered to the point that she became
23
very agitated, irrational, ultimately suicidal,
24
because none of the so-called medical professionals
25
acknowledged the drug's role in her irrational and
124
1
suicidal behavior or properly withdrew her from
2
their suicidal effects.
3
Our lovely 14-year-old daughter is dead.
4
Dominique has been denied the unalienable right by
5 her
creator of the pursuit of life, liberty, and
6
happiness. She will no longer be
able to pursue
7 her
dreams of becoming either a computer software
8
engineer, computer graphics engineer, or marine
9
biologist, and someday an entrepreneur, she had
10
hoped.
11
Gone, too, is the ability to be able to
12
watch Dominique blossom into womanhood, as well as
13
motherhood, as she expressed the desire to someday
14
have five kids. Now, we will
never have the
15
opportunity to continue sharing our lives with
16
Dominique, whom we loved and cherished so much.
17 She was not only very intelligent,
18
humorous, delightful, insightful, and innovative,
19 she
was also very caring and thoughtful.
Dominique
20 had
a way of making others feel special and loved.
21 She
touched so many lives. For example,
Dominique
22
made 1,000 paper origami cranes and sent them to
23
Governor George Pataki of New York for the first
24
anniversary of 9/11.
25
It was because of Dominique's very loving
125
1 and
genuine nature that around 300 people showed up
2 to
her memorial service. They couldn't
believe
3
that for someone who was so loving and caring, she
4
would herself take her own life.
5
I submit to you today,
ladies and
6
gentlemen, that Dominique's life was taken from her
7 as
a result of drug-induced psychosis and suicidal
8
ideations, not to mention the probability of
9
experiencing akathisia, extreme agitation. As a
10
14-year-old adolescent, her brain was experiencing
11 the
second largest growth period, and her hormones
12
were unbalanced.
13
How can teenagers be allowed to be given
14
antidepressants that were never approved for
15
adolescent consumption, only for adults?
How come
16 the
medical profession doesn't fully disclose the
17
possible harmful and fatal effects of medication as
18
well as watch carefully for diverse effects on its
19
adolescent population?
20
DR. RUDORFER: I am sorry that we
are out
21 of
time, but thank you very much.
22
If we could have speaker 28, please.
23 Matthew Piepenburg
24
MR. PIEPENBURG: Well, there are
very
25
impressive credentials around this room and
126
1
certainly at this panel, and impressive schools and
2
qualifications and professorial positions at very
3 elite
institutions.
4
There are also a number of impressive
5
terms of art tossed around - morbidity,
6
idiosyncratic. I like Mr. Katz's
term controlled
7
data or controlled trial data.
8
What I would like to suggest is behind me
9 is
a number of things that do not show up in
10
controlled trial data that need to be heard, that
11 are
as important as what can be achieved
12
statistically.
13
I don't think for parents who spend a
14 great deal of time in cemeteries, controlled
trial
15
data is as pervasive or persuasive.
16
I do not suggest or believe that everyone
17
here has a negative or a grotesque motive or is all
18
greedy. I do think there are
legitimate motives
19
here, and I think these things do need to be
20
discussed without being incendiary.
21
Nevertheless, it is important to recognize
22 the
human dimension here. We had prepared a
23
two-page speech full of FDA talk papers, adverse
24
reporting events on Paxil in particular, my family
25
friend, Paul Domb, has suffered as a victim of
127
1
Paxil. It is just very hard to go
over that when
2 you
hear these stories.
3
Last night, we were at a restaurant.
We
4
gave the waiter our speech to print out for us off
5 of
a disk. He came back. He had suffered Paxil
6
side effects that led to suicidal thoughts, violent
7
thoughts after a 40-year marriage, and he saw our
8
speech and sat down for 20 minutes and basically
9
cried before us.
10
It is a pattern and epidemic that is
11
pervasive and has more importance to me than the
12
statistics we were going to read.
Let me just
13
suggest also that this individual had been to
14
Vietnam, lost most of his platoon and most of his
15
body in Vietnam, crawled for two and a half days
16
through the jungle to survive.
17
None of that caused him the depression or
18 the
desire to jump off a bridge like Paxil did.
If
19 he
could handle Vietnam with poise, how are 13- and
20
12-year-old kids supposed to handle Paxil?
21
Thank you very much.
22
DR. RUDORFER: Thank you.
23
Could we have speaker 29, please.
24 Terri Williams
25
MS. WILLIAMS: My son, Jacob
Williams, was
128
1
born on October the 15th, 1986.
Jacob was an
2
exceptional athlete who participated in football on
3
both the varsity and junior varsity football teams
4 in
his school.
5
In September of 2000, Jacob experienced a
6
loss of interest in his school activities. He
7
maintained his interest in football, however, there
8 was
a conflict with his grades and his attendance.
9
As a result of this issue, his father and
10 I attended a conference at his school on
October
11 the
11th, 2000 with various representatives from
12 the
school. The school administrator
suggested
13
that Jacob may be depressed and that we should seek
14
medical help.
15
I contacted Jacob's pediatrician
and made
16 an
appointment for 3:45 that afternoon. On
October
17 the
11th, 2000, his pediatrician prescribed 10
18
milligrams of Prozac, which was increased to 20
19
milligrams three weeks later.
20 Shortly after starting the initial
dose,
21
Jacob began to complain of having strange dreams,
22
which he had said were bad.
Shortly after the
23
dosage was increased, I began to
notice an
24
aggressive behavior, which had not been there
25
before. Jacob also became
destructive and
129
1
destroyed some of his favorite things.
2
His friends would later tell me they had
3
noticed the same behavioral change.
He also showed
4 a
verbal aggression and short temper, which had not
5
been present before.
6
When questioned about this behavior, he
7
stated I don't know what is making me do this. At
8
this time, I thought this could be a part of normal
9
adolescent behavior and did not pursue the matter
10 any
further.
11
On December the 5th, 2000, I discovered
12
Jacob's body hanging from the rafter in our attic.
13 He
had hung himself with his own belt. A
letter
14 was
placed on the ladder leading up to our attic
15
thanking us for giving him 14 years of a happy
16
life.
17
Something had to have gone wrong in the
18
thinking process to have brought this about. Had I
19
know that this was a potential side effect,
20
suicide, I would have never allowed my son to take
21 the
drug Prozac.
22
Thank you.
23
DR. RUDORFER: We are now going to
go to
24
speaker 32, please.
25 Glenn McIntosh
130
1
MR. McINTOSH: I would like to
introduce
2 you
to my daughter, Caitlin Elizabeth McIntosh.
3
Well, it is actually only a 2-dimensional image of
4
her, but it is all I have left.
She died of
5
suicide at age 12 years, 3 months, just 8 weeks
6
after being put on Paxil, and then Zoloft.
7
Caitlin was a straight "A" student in the
8
fifth grade, a talented musician, artist, and poet,
9 who
loved animals and wanted to be a veterinarian.
10 The
sixth grade began, and that, combined with the
11
onset of puberty, this bright, sensitive girl who
12 had
once loved going to school, started having some
13
trouble coping, as many kids do in the sixth grade,
14
it's a tough adjustment.
15
She was also having some problems sleeping
16 due
to a mild seizure disorder. We wanted to
help,
17 of
course, so we took her to our family physician,
18 who
prescribed her Paxil. He said it would
help
19
with her coping and her sleep.
20
She didn't do well on it at all, so he
21
took her off it cold turkey, which you are not
22 supposed
to do. When we saw a psychiatrist a week
23
later, he put her on Zoloft. She
then started
24
having strong suicidal ideations, along with severe
25
agitation known as akathisia and hallucinations,
131
1 and
she was put in the adolescent ward of a mental
2
hospital to "balance her meds."
3
Well, there, things only got worse, as she
4 was
put on other strong psychotropic drugs to treat
5 the
symptoms that we now know were actually caused
6 by
the SSRIs, and let me be very clear about
7
something. The dramatic and
severe symptoms that
8 led
to my daughter's suicide manifested only after
9 she
started taking antidepressant drugs.
10
The downward spiral continued until
11
January 5th, 2000, when she hung herself with her
12
shoelaces in the girl's bathroom in the middle
13
school she was attending.
14
We were told that antidepressants like
15
Paxil and Zoloft were wonder drugs, that they were
16
safe and effective for children.
We were lied to.
17 The
pharmaceutical companies have known for years
18
that these drugs could cause suicide in some
19
patients. Why didn't we?
20
I implore you, ban the use of
21
antidepressants here in the United States so that
22
other parents will not have to endure the pain I
23
felt and other children might be saved.
24
DR. RUDORFER: Thank you.
25
Speaker 33, please.
132
1 Delnora Duprey
2
MS. DUPREY: My name is Delnora
Duprey,
3 and
it has been well over two years since I have
4
seen my grandson play ball, ride a bike, talk on
5 the
phone, or run in to say, "Hey, grandma, what's
6 for
dinner?"
7
All the normal everyday things in his life
8 are
lost. He is not here to get his restricted
9
license in April, see his little sister start
10
school, to ride with his big sister when she
11
started driving, or just to go out and have pizza
12 and
see a movie.
13
A tall, thin boy, quiet and well liked and
14 respectful to everyone, a big heart and a
smile
15
that made you ask what are you up to, a boy who
16
loved his family dearly, had hopes and dreams for a
17
future. A future of uncertainty
now - he is locked
18
away in a detention center awaiting trial for the
19
murder of two people who he loved most in the
20
world.
21
A nightmare that started with a diagnosis
22 of
depression and placed on medication that was
23
never tested on children and never meant for their
24
use. He had no say in this. We, as adults, trust
25 our
doctors and the FDA to know what they are
133
1
doing. Even when we get
complaints, we say the
2
doctor said it will help you.
3
A sweet boy who never hurt himself or
4
anyone else went to live with his grandparents.
5 His
medication was changed from Paxil, which he had
6
been on a very short time, to Zoloft.
7
From a family physician, this medication
8 was
increased to 200 milligrams for an 80-pound
9
child. Within 48 hours, his
grandparents were
10
dead, and he is sitting, facing a life of
11
uncertainty, a life of maybe total incarceration
12 for
the rest of his life, a child that does not
13
even know what has happened to him.
14
I don't want to see any more families go
15
through this nightmare that we have all endured.
16 The
child's life changed forever. Next time
it
17
might be one of your own family.
We must stop
18
these drugs for children and strengthen our
19
restrictions on the doctors who prescribe them.
20
DR. RUDORFER: Thank you.
21
Number 34, please.
22 Joe Pittman
23
MR. PITTMAN: Hello. My name is Joe
24
Pittman.
25
My son, at the tender age of 12, killed my
134
1
parents. I am going to read you a
letter he wrote
2 to
me to you all.
3
"Dear FDA: My name is Chris
Pittman. I
4 am
now 14 years old. I would like to tell
you what
5
happened to me, what the medication did to me and
6 how
it made me feel.
7
"When I was taking Zoloft, I took the
8
lives of two people that I loved more than
9
anything, my grandparents. I went
to the doctor
10 and
he gave me a sample pack of Zoloft. He told me
11 to
take 50 milligrams once in the morning and
12
another 50 at night.
13
"I didn't notice a change in my behavior
14
until I was completely off the medication. It made
15 me
hate everyone. The smallest things made
me blow
16 up,
and I started getting into fights, which was
17 not
me. I would usually avoid fights. Before the
18
medication, I had only been in two fights my whole
19
life. I just hated the whole
world for no apparent
20
reason.
21
"A week after the doctor gave me the
22
sample packs, he increased my dosage to 200
23
milligrams a day. Everything just
kept getting
24
worse. Then, I snapped. I took everything out on
25 my
grandparents who I loved so very much.
135
1
"When I was lying in my bed that night, I
2
couldn't sleep because my voice in my head kept
3
echoing through my mind telling me to kill them
4
until I got up, got the gun, and I went upstairs
5 and
I pulled the trigger. Through the whole
thing
6 it
was like watching your favorite TV show.
You
7
know what is going to happen, but you can't do
8
anything to stop it. All you can
do is just watch
9 it
in fright.
10
"Because of my own personal experience on
11 the
medication, I would not want anyone to go
12
through what I have then and now, losing the lives
13 of
my loved ones for the effects of homicide or
14
suicide, or both, due to the medication.
15
"Thank you. Christopher
Pittman."
16
DR. RUDORFER: Thank you.
17
Number 35, please.
18 Richard Mack
19
MR. MACK: My name is Richard
Mack. I am
20 a
retired law enforcement officer and sheriff from
21
Arizona.
22
My expertise in that field was juvenile
23
delinquency, school violence, and narcotics
24
investigations.
25
My first experience with SSRIs was when I
136
1 was
a parent of a second grader, my wife and I were
2
called into the school, our son had a problem
3
staying in his chair. What was
the government
4
school's answer? Drug your son
into submission, so
5 he
will stay in his chair.
6
We refused and we thank God now that we
7
did. Our son turned out just
fine, played
8
basketball, baseball, and excelled at school and
9
sports.
10
I was a sheriff of a small community in
11
Arizona. We had an abnormal amount of high rate of
12
suicide and teen violence. I am
just an
13
investigator, I just present the facts.
One thing
14
that we could not ignore was the circumstantial
15
evidence that the common denominator in all of
16
these cases was the victims or perpetrators were on
17
SSRIs.
18
In investigating these events, it became
19
quite commonplace for all of us to ask the same
20
question as we got to the next event of horrified
21 and
traumatized people and families. You
have
22
heard from many of them today.
23
Some people don't have the adverse
24
reaction to these drugs, some do.
I learned the
25
same with LSD when I investigated that as an
137
1
undercover narcotics officer. I
can only say that
2 the
evidence is mounting over and over as did our
3
investigations.
4
We cannot, as law enforcement officials,
5
ignore such circumstantial evidence.
I doubt very
6
seriously if you could either. I
am an advocate
7 for
state's rights and I do believe that if the FDA
8
fails to take action, the state and local
9
authorities will have to.
10
Thank you.
11
DR. RUDORFER: Thank you.
12
Speaker 36.
13 Noah Wright Smith
14
MS. SMITH: My name is Noah Wright
Smith
15 and
I am a 15-year-old victim of legalized drug
16
abuse. My mother had me put on
Ritalin when I was
17
5. I felt sick all the time on
Ritalin and it was
18
just the beginning of bad things happening to me
19
because of drugs.
20
My grandparents won custody of me last
21
year. When they won, they got
upset because I was
22 in
bad shape and on a lot of drugs. They
picked me
23 up
at Broughton Mental Hospital in Morganton, North
24
Carolina, and learned I was on 1,000 milligrams of
25
drugs a day. In my lifetime, I
have been on 16
138
1
psychotropic drugs including Zoloft, Paxil, and
2
Effexor, and all of them made me feel sick and do
3
very bad things.
4
I wasn't a bad kid. I was a badly
abused
5
kid, abused by my mother and my stepfather. The
6
Department of Social Services knew I was being
7
abused, but they didn't do anything except put me
8 on
more drugs.
9
The drugs made me sick and do bad things
10
like trying to stab my teacher with scissors.
11
Sometimes it made me want to kill my parents, and I
12
told them that, and was put in a mental hospital.
13
Some drugs made me have bad nightmares, so
14 I
tried very hard not to sleep every night, so they
15
gave me drugs to make me sleep.
Some of the drugs
16
made me want to kill myself. I
couldn't stop
17
thinking about killing myself.
When I told the
18
doctors, they sent me to still another mental
19
hospital.
20
One day I tried to jump off a very high
21
railing to kill myself. I was put
in a mental
22
hospital again for doing that, but I really wanted
23 to
die. I really did want to, and I was so
scared
24 and
mad, too. In those mental hospitals,
they kept
25
giving me more drugs, and I got depressed. I got
139
1
diabetes and high blood pressure.
2
My grandparents won my custody and took me
3 to
a new psychiatrist. We have worked hard
4
together and he found I really don't need any
5
drugs. Last year he took me off
all of them, one
6 at
a time. No more nightmares or wanting to
hurt
7 or
kill other people, and I don't want to kill
8
myself anymore.
9
Drugs almost ruined my life and almost
10
killed me. What about the kids that have to take
11
these drugs? I don't want kids to
kill themselves.
12 Who
is taking care of them? Who really cares about
13 us
kids? I don't even know if you care, do
you?
14
Somebody had better listen to kids who say the
15
medicines make them want to kill themselves, and
16
make them sick, and do bad things, because they are
17
telling you the truth.
18
Thank you for listening to me.
Now,
19
please, help the other kids, so that they don't get
20
hurt by drugs, and so they don't kill themselves.
21 I
almost killed myself and I am glad I am alive.
22
DR. RUDORFER: Number 37, please.
23 Marion Goff
24
MS. GOFF: I do not have any
financial
25
ties. I am her with my daughter,
Alex. We are
140
1
here to tell you about her twin sister, Devon, when
2 she
was 9 years old. We are also joined by
Senator
3
Lincoln Chafee's wife Stephanie who is a friend of
4
ours.
5
In 2002, Devon developed an
6
obsessive-compulsive disorder very suddenly and
7
very severely. In a three-month
period, she lost
8 10
pounds. We consulted a specialist who
9
prescribed Zoloft on her second visit with him.
10
Soon thereafter, he increased the Zoloft to 50
11
milligrams or more, but it didn't help, so he
12
changed her prescription to Paxil.
13
She was hospitalized and Devon's medical
14
condition was compromised in that she had developed
15 a
cardiac arrhythmia and had to be placed on a
16
heart monitor. She was in the
hospital for one
17
month, and she was on the heart monitor and bed
18
rest for the entire time.
19
During this time, her Paxil was increased
20 to
20 milligrams. A few days later she was
started
21 on
Zyprexa also. Devon was not getting any
better,
22 in
fact, her behaviors grew worse. She
began
23
hitting her head against the metal hospital bed.
24 She
threatened to jump out of the window on two
25
occasions.
141
1
On two other occasions, we found a pair of
2
sharp scissors in her bed. Our
child was never
3
suicidal before these medications.
At one point,
4 my
9-year-old child, who weighed little more than
5 60
pounds, was on 30 milligrams of Paxil and 10
6
milligrams of Zyprexa.
7
Our gentle daughter would now fly into a
8
rage several times each day. It
became part of our
9
life to have my husband and myself restrain Devon
10 at
times for fear that she would truly hurt
11
herself.
12
During these times, she would try to
13
inflict injury upon herself by banging her head on
14
walls, beds, floors. She would
punch herself in
15 the
legs and arms. She grew extremely
violent
16
toward us. She would run to the
silverware drawer
17 and
get a knife and attempt to stab herself.
18
The worst moment happened when I looked in
19 on
her, in her room one night, to find her by her
20
open second floor bedroom window with one leg out
21 the
window in a position as if she appeared she
22
would jump.
23
Devon is presently being treated for Lyme
24
Disease. In summary, our
experience has been one
25 of
absolute terror to watch your 9-year-old
142
1
daughter suffer so much, so suddenly, and to be so
2
lost in helping her.
3
So often we would ask why this was
4
happening, and we were told to forget about the
5
etiology.
6
DR. RUDORFER: I am sorry, we are
out of
7
time. Thank you very much.
8
Number 38, please.
9 Gary Cheslek, M.D.
10
DR. CHESLEK: Actually, my wife is
11
speaking later.
12
My name is Gary Cheslek, and I am a
13
practicing dentist from Vicksburg, Mississippi, and
14 I
am speaking today, not just as a health care
15
professional, but also as a parent.
16
I am here today to tell you an anecdote.
17
Webster defines an anecdote as a short narrative of
18 an
interesting or amusing biographical event, an
19
anecdote or anecdotal. That is
the euphemism the
20
manufacturers of Prozac, Paxil, Effexor, and Zoloft
21 use
to describe the thousands of reported out of
22
character, violent, homicidal, suicidal events that
23
occur in a vulnerable subset of patients who ingest
24
their SSRI antidepressants. They
would have us
25
believe that these are mere coincidences and don't
143
1
prove anything.
2
My son, Justin, was a 20-year-sophomore at
3 the
University of Southern Mississippi when he went
4 to
the Student Health Clinic complaining of
5
insomnia. He was given a thorough
examination
6
including bloodwork. Significant
in the doctor's
7
note at that initial visit is the notation, "No
8
suicidal ideation."
9
Complaining that the sleep medication he
10 was
prescribed made him feel sedated and depressed,
11 he
was put on Paxil for two weeks. During
those
12 two
weeks, he repeatedly told his doctor he didn't
13
like the way the Paxil made him feel, so he was
14
switched to Effexor.
15
Within 24 hours of the switch to Effexor,
16 he
had a seizure. Five days later he hung
himself
17 in
his apartment. He didn't leave a
note. Beneath
18 him
was his laptop computer and a glass of Coke.
19 It
was as if some sudden impulse had made him do
20
this.
21
We grilled his girlfriend about his mood
22 and
behavior in the months prior to his death.
She
23
said his demeanor changed dramatically around her
24
birthday, February 22. Justin
started taking Paxil
25
February 21.
144
1
Last June, regulators in the UK and Canada
2
banned Effexor and Paxil for use in children and
3
adolescents, and recently expanded that ban to all
4
SSRIs except Prozac. Last August,
Wyeth issued a
5
Dear Doctor letter alerting the health care
6
professionals that the clinical trials had not
7
established the safety and effectiveness of Effexor
8 in
children, and revealed an increased risk of
9
suicidal ideation and self-harm.
10
The letter does not, however, indicate
11
that some of these trials were done seven years
12
ago.
13
DR. RUDORFER: Thank you very
much.
14
Sherri Walton
15
MS. WALTON: My name is Sherri
Walton and
16 I
am here as a volunteer advocate. This is
my
17
14-year-old daughter, Jordan. We
have traveled
18
here from Arizona at our own expense because we
19
know that public forums, such as this, usually only
20
hear from those who have had negative experiences.
21 We
felt it was important for us to share our story.
22
Jordan was diagnosed with Tourette's
23
syndrome when she was 7 years old.
As is typical
24 of
Tourette's syndrome, she also has OCD and ADHD.
25 She
was originally prescribed an SSRI medication to
145
1
relieve the anxiety that consumed her because she
2
could not control her thoughts or behaviors.
3
This medication allowed her to participate
4 in,
and understand, the cognitive behavior therapy
5
that gave her some semblance of normalcy. In
6
fourth grade, Jordan was still being hampered by
7 the
obsessive thoughts caused by her OCD. In
the
8
classroom, this was overwhelming and extremely
9
frightening for her.
10
Her medication was changed to a different
11
SSRI and within a few months, her obsessive
12
thoughts became less and less intense.
They were
13
still there, but now she was able to recognize what
14
they were and usually work through them.
15
Dance is Jordan's passion. It is
what she
16
wants to do with her life. In
November of 2002,
17 she
announced she wanted to quit dance. As
she
18
burst into tears, she said that she wanted to die,
19 she
wanted to kill herself.
20
She was diagnosed with clinical depression
21 and
her medication was changed from the SSRI she
22 had
taken for four years to a different SSRI to
23
treat both her OCD and depression.
24
As Jordan has struggled to find success in
25
school and in her relationships with peers, her
146
1
meds were sometimes the only thing she could count
2 on
to help her. The daughter I have here
now
3
standing next to me is a happy, healthy, successful
4
teenager. There is no doubt in my
mind that the
5
SSRI medication saved her life, and like the other
6
SSRI antidepressants she is taking gave her a
7
chance for a full and complete life.
8
With the greatest sympathy for any
9
families who have lost children to suicide, I ask
10
that you identify and fix any breakdown in the
11
system that could lead to such tragedy.
At the
12
same time, I ask that you appreciate and take into
13
account the enormous benefits that these
14
medications have had for children and their
15
families.
16
Please urge the FDA not to take away the
17
tools that have allowed my daughter and millions of
18
other sons and daughters out there to be successful
19 in
life, and, in fact, to have lives.
20
As a parent, I call on the FDA to take no
21
action that would harm my child.
22
DR. RUDORFER: Thank you.
23
We are up to speaker 40.
24 Peter R. Breggin, M.D.
25
DR. BREGGIN: Hello. I am Dr. Peter
147
1
Breggin. I am a psychiatrist and
one of the few
2
experts in the world on medications who isn't
3
involved in any way with the drug industry. I
4
think there are handful of us.
5
I have given you a peer-reviewed article
6
that came out just a few weeks ago that I wrote,
7
which is the most extensive review to date on
8
violence, suicide, and mania caused by the SSRIs,
9 and
it has just, I don't know, maybe hundreds of
10
citations.
11
Back in the 1980s when Prozac was being
12
approved, Richard Kapit, the chief medical officer
13 at
the FDA, identified a stimulant syndrome in
14
association with Prozac, and he repeatedly warned
15 in
in-house documents that this stimulant effect
16
would turn depression into agitated depression and
17
cause a deterioration in the individual.
18
Since then, we have been able to identify
19 a
continuum of stimulation that has at least four
20
syndromes involved, that I have now seen produce
21
violence and suicide in dozens of patients in my
22
clinical consultations and in my medical/legal
23
work.
24
The syndrome, first and foremost, includes
25
manic-like behavior. We know that
Luvox, for
148
1 example,
just in its label has a 4 percent rate of
2
mania. From Emslie's study, hidden in the fine
3
print, we know that Prozac, controlled clinical
4
trials, 6 percent rate of mania.
5
The second syndrome is the agitated
6
depression, it is hard to tell often clinically
7
from mania.
8
The third syndrome is this obsessive
9
suicidality and violence, and the fourth syndrome
10 is
akathisia, which we now know, and is even in the
11 old
DSM, can produce psychosis and agitation, and a
12
variety of other problems leading to suicide and to
13
violence.
14
The literature is extensive. You
have got
15 to
go beyond the needle in the haystack.
Please
16
look at my review.
17
DR. RUDORFER: Thank you, Dr.
Breggin.
18
Speaker 41.
19 Robert Fritz
20
MR. FRITZ: People have been
pleading with
21 the
FDA for 11-plus years to put warnings on
22 prescriptions
for antidepression medication to no
23
avail. The FDA has had people
present information
24
about suicidal tendency increase and numerous
25
completed suicides, and still no warnings of
149
1
increased risk of suicide were issued.
2
The people of the United States have a
3
right to know what risks are associated with taking
4
these drugs. I have a right to
know what risks are
5 associated
with taking these drugs, so I can make
6 an
informed decision as to whether or not I want my
7
children to take these drugs.
8
The need for a warning is compounded by
9 the
fact that doctors are prescribing these
10
medications off label. My
daughter, Stephanie Raye
11
Fritz was taking Zoloft. We
weren't told of any
12
risk of increased suicidal tendencies or increased
13
suicide attempts.
14
She hung herself on the evening of
15 November
11th in her bedroom after finishing her
16
homework. She showed no signs of
increased
17
depression or imminent suicidal thoughts, and, in
18
fact, was still recruiting people to see her sing
19 the
following month.
20
We had no warning of what Zoloft could do
21 to
our daughter, but you people, the FDA, certainly
22
did. On October 27th, two weeks
before she took
23 her
life, you put out a Public Health Advisory and
24
notified physicians about preliminary data from
25
studies suggesting an excess of reported suicidal
150
1
ideation and suicide attempts for pediatric
2
patients receiving certain of these antidepressant
3 drugs.
4
Why weren't we, the parents of the kids
5
taking Zoloft, notified with this advisory? It is
6 too
late for my daughter, but for the FDA to
7
continue to sit on this information and not let the
8
public know the risks associated with these drugs
9 is
a gross misuse of power.
10
I am not asking that these drugs be taken
11 off
the market. I don't know enough about
their
12
safety to recommend that. What I
am seeking is
13
that when the drugs are prescribed off label, or
14
when drugs are prescribed after an advisory is
15
issued suggesting new adverse side effects, that
16 the
FDA make it mandatory that the physicians
17
prescribing such drugs explain in plain English
18
what the risks are and that an informed written
19
consent be received from the parents or the
20
patient's guardian.
21
I hope that you will agree that all
22
Americans deserve to know what risks they are
23
assuming when they take medication.
I believe that
24
most Americans, including most elected officials,
25
agree with that.
151
1
How many more people have to die before a
2 warning gets issued?
3
DR. RUDORFER: Thank you.
4
We are going to move ahead to speaker 43.
5 Lawrence Greenhill, M.D.
6
DR. GREENHILL: My name is
Lawrence
7
Greenhill. I am a child
psychiatrist, Professor of
8
Child Psychiatry and Pharmacology at Columbia. I
9 am
speaking today on behalf of the American Academy
10 of
Child and Adolescent Psychiatry where I serve as
11
Chairman of the Program Committee and as Chair of
12 the
Pediatric Psychopharmacology Initiative
13
Committee.
14
First, I want to extend my sympathy to all
15 the
families who spoke so moving here today about
16
their losses. I think similarly, the membership,
17 who
are comprised of 7,000 child psychiatrists at
18 the
American Academy of Child and Adolescent
19
Psychiatry, are concerned about these families, and
20
they want to get the results of this review to help
21
their patients with safe and effective treatments.
22
In that regard, the American Academy of
23
Child and Adolescent Psychiatry supports the review
24
that is going on and it specifically supports the
25
reclassification of suicidal events using patient
152
1
charts, that is, patient level analysis, as the
2
category that turned up in Dr. Laughren's report of
3
possible suicide-related events was one most
4
subject to possible methodological bias that might
5 be
addressed by patient level analyses and
6
reclassification.
7
Furthermore, I support the mandatory
8
registration of all clinical trials as advocated in
9
JAMA by Dickerson and Rennie in July of 2003. That
10 is
because one of the greatest roadblocks to
11
understanding the safety and efficacy of trials is
12 the
lack of public access and its disclosure of
13
these data sets due to laws that treat some of the
14 data as proprietary trade secrets.
15
I join my colleagues at Columbia in
16
encouraging the field to carry out further
17
prospective placebo-controlled trials using methods
18
such as we have heard today, the randomized
19 withdrawal discontinuation or challenge,
20
de-challenge --.
21
DR. RUDORFER: Thank you, Dr.
Greenhill.
22
Number 46, please.
23 Suzanne Vogel-Scibilia, M.D.
24
DR. VOGEL-SCIBILIA: I would like
to have
25 my
remarks into the written record, and I want to
153
1 let
you know I am here at my own expense.
2
Good morning. My name is Dr.
Suzanne
3 Vogel-Scibilia. I a member of the NAMI board of
4
directors. As a person diagnosed
with bipolar
5
disorder, I am proud to serve on the NAMI Board and
6
proud that NAMI is the nation's voice on mental
7
illness representing both consumers and family
8
members. I am also proud to be
the mother of five
9
children, two who are diagnosed with mental
10
illnesses and one who is currently being treated
11
with an SSRI.
12
I am also a practicing clinical
13 psychiatrist
with no financial ties to the
14
pharmaceutical industry. I
represent thousands of
15
families across the country.
16
My son, Anthony, had a very severe mental
17
illness primarily depression and attention deficit
18
disorder as a manifestation of his bipolar
19
disorder, and another son has had treatment with
20
numerous antidepressant medications including
21
several SSRIs.
22
My children have had tremendous
23
improvement with their illnesses and lead very full
24 and
functional lives because of SSRI medication,
25
along with other psychotropic medications. I
154
1
shudder to think of their plight if these
2
medications were not available.
3
One of my sons has had suicide attempts
4 and
violent incidents with knives. He has
also run
5 out
of our house - in a fit of terror --in subzero
6
weather only to be found freezing and hypothermic
7 by
our local police department in the next town.
8
These incidents all occurred while his illness was
9 not
adequately treated with an antidepressant
10
medication.
11
My other son suffers from disabling
12
obsessive- compulsive disorder symptoms and
13
depression, and has had his life dramatically
14
improve from treatment with SSRIs.
15
I want to talk and speak about suicide and
16 the
consequences of untreated mental illnesses.
17
We are pleased that the FDA is looking
18
closely at the data related to SSRI use and
19
suicidality. NAMI is deeply
concerned with the
20
public health crisis and the number of youths who
21
commit suicide. The U.S. Surgeon
General reports
22
that up to 80 percent of our youth who need mental
23
health treatment receive none at all.
24
In summary, I would like to thank the
25
committee for allowing 200,000 members of NAMI to
155
1
share our views on this critically important issue.
2 I
hope and pray that this committee will render a
3
decision based, not on emotion-filled pleas of
4
individuals whose experience are not supported by
5
adequate research.
6
Thank you very much.
7
DR. RUDORFER: Thank you.
8
If we could have speaker 48, please.
9 Dennis Winter
10
MR. WINTER: I am Dennis
Winter. I am
11
here today with Karine Winter and Mary Lou Winter,
12
Beth's mom.
13
Four months ago or less than four months
14
ago, Beth, a 22-year-old recent graduate from the
15
University of Rhode Island, she graduated summa cum
16
laude, she was a child who was loving, from a very
17
tight, close family, never any instance of alcohol
18 or
drug abuse, never any problems, a wonderful
19
student, a wonderful girl, a loving sister to her
20 brothers and sisters, committed suicide after
being
21 on
Paxil for seven days.
22
Now, what I think is critical here is the
23
fact that she can go to her general practitioner on
24 the
first visit and be prescribed Paxil. I
think
25 it
is clear that you need to come out with warning
156
1
labels for practitioners and doctors, so the
2
lawyers in this room, when those labels are out
3
there, if the doctors continue to do it, will be
4
able to bring actions. If you
bring out the
5
warning labels, there is enough legal community in
6
this world that will police itself.
7
Let me go on. As we are sitting
here
8 today,
we heard a lot about idiosyncratic data, all
9
permitted data, requested data available, data we
10 are
permitted to evaluate fully, and it comes down
11 to
this data stream that we don't know that
12
happened 15, 20 years ago, the data stream you are
13
trying to analyze.
14
I don't know, like Mr. Farber said, if you
15 are
going to be analyze all that data and come out
16
with that data. You should put
out warning labels
17
because you are not going to get a clear answer.
18
I am running out of time, but Dr. Healy
19
provided testimony in federal court on May 22nd,
20
2001. Everybody needs to be read
that testimony.
21 He
gave it under oath, under threat of perjury, and
22 that
is very enlightening to anybody involved here,
23 and
you really need to read it.
24
Also, you need to look at confidentiality
25
agreements. A lot of families of
people who commit
157
1
suicide are embarrassed. When the
lawyers come,
2
they sign confidentiality agreements, and you don't
3
hear about what is really happening out there.
4
DR. RUDORFER: Thank you very
much.
5
We are going to move along
to speaker 51.
6 Steve Cole
7
MR. COLE: I am Steve Cole. I am here at
8 my
own expense.
9
My father committed suicide after 13 days
10 on
Prozac. He has absolutely no history of
mental
11
illness, in fact, quite the contrary.
He and my
12 mom
had just built a new house, a lot of the work
13 he
did himself. He and I and a friend built
a
14
cabin out of raw lumber.
15
These are not the type of things that you
16 do
if you are planning on dying. Let me
repeat
17
that. You do not do that.
18
He was looking forward to his new house.
19 He
was planning many activities. He was
upbeat, he
20
didn't drink or gamble, and he did not have any
21
recognized prerequisites for suicide unless you
22
want to consider all 70-year-old men suicidal, and
23 I
just don't buy that. Generally, he was
in very
24
good health.
25
Next slide.
158
1
He experienced some chest pains about a
2
month and a half after moving into the new house.
3 As
a precaution, he went to his
cardiologist. His
4 heart
tested perfectly well. He was upbeat and
had
5 a
new grandbaby on the way.
6
He was prescribed Prozac off label for the
7
chest pain. The doctor, who is an
outstanding,
8
wonderful man, stood behind us on this, and stated
9
that he has no doubt that it was Prozac induced.
10
Eleven days after he started, he demonstrated
11
symptoms of akathisia, he was jittery.
His fingers
12 and
his skin felt odd, he was easily agitated.
13
He told me, "I cannot stand the way this
14
drug makes me feel." Two
days later he committed
15
suicide.
16
Growing up, he watched a lot of westerns.
17 He
loved westerns, but he would turn the channel if
18 a
man was hung or lynched. This is the way
my
19
father died. He hung
himself. It was completely
20 out
of character. He died by means of his
own
21
nightmare.
22
Thank you very much.
23
DR. RUDORFER: Thank you.
24
Number 52, please.
25 Allan Routhier
159
1
MR. ROUTHIER: I am here to
request that
2
Wellbutrin be recognized as another dangerous drug.
3
Information was sent to this committee by some
4
researchers and myself as to the reasons for
5
inclusion. There are too many
cases of suicide and
6
deaths caused by this drug. It is
known to cause
7
akathisia, depression, psychosis, serotonin
8 syndrome, seizures, hallucinations, and many
other
9
serious adverse effects.
10
One suicide while on Wellbutrin for ADHD
11 was
9-year-old Carey Brooks, who had to kneel down
12 to
hang himself with his shoelace. There
are many
13
reasons these drugs are prescribed, and they can
14
cause suicide in non-depressed people.
15
Do not blame acts of drug-induced
16
psychosis on depression especially when this is
17
happening to people given these drugs for other
18
purposes. It is not only
SSRIs. SSRI is a
19
misnomer. None of them are
selective to serotonin.
20
When you affect one neurotransmitter, you affect
21
others.
22
Remeron, Serzone, Effexor are not SSRIs.
23
Effexor works on serotonin, norepinephrine, and
24
dopamine, as does Wellbutrin. FDA
Med Watch
25
reports hundreds of suicides on Wellbutrin.
160
1
Wellbutrin is structurally similar to amphetamine
2 and
overstimulates many people.
3
Six months ago my wife went to the doctor
4
sick and was sent home with Wellbutrin.
After six
5
days of serious adverse reactions and insomnia, she
6
shot herself. This was not her.
Forty years old,
7
beautiful, with two boys, she was a perfect wife
8 and
mother, married for 18 years, almost 25 years
9
working in the Welfare Office.
10
She was never depressed. She was
the most
11
loving, unselfish person anyone could know.
12
Immediately after starting Wellbutrin, she was not
13
herself. This was an act of
psychosis. This has
14
been happening for too long.
People are worth more
15
than profits.
16
How many more have to die
before something
17 is
done? Don't be fooled by manipulated
studies.
18
This was whitewashed in 1991, now they are trying
19 to
do it again. This happens to adults, as well as
20
children, prescribed for any reason, not just MDD.
21
My wife was murdered. The FDA is
supposed
22 to
protect us from these pill pushers.
23
Thank you.
24
DR. RUDORFER: Thank you.
25
Number 53, please.
161
1 Daniel J. Safer, M.D.
2
DR. SAFER: I am Daniel
Safer. I am a
3
child psychiatrist, and I have no conflict of
4
interest in coming here.
5
I think the major finding of the British
6
Committee on the Safety of Medicines was that most
7 of
the data that they got were unavailable to them
8
prior to the company coming in for an indication,
9 so
when they found the data, they were surprised to
10 see
that most of the studies were negative or
11
failed for the treatment of depression in children
12
using SSRIs. So, that was I think
the major
13
finding as far as I am concerned of the British
14
Committee.
15 The second finding indeed was that
most of
16 the
studies, the vast majority of the studies they
17
looked at were either failed or negative for the
18
treatment of depression in children.
19
The third finding had to do with the side
20
effects of particularly the suicidality issue,
21
which I consider a minor finding of the British
22
report. It was about 1 and a
quarter percent rate
23 for
placebo and about 3.5 percent for the active
24
medication.
25
I think that is fairly understandable
162
1
because the medication, the SSRIs are known, and
2
have been known, to increase the risk of agitation
3 and
activation and children. In fact, the rate is
4
about 15 to 20 percent when you look over about 40
5 or
50 studies on SSRIs.
6
It is a high rate, so you would expect
7
that children who were depressed might have an
8
increased rate of suicidality if they are agitated
9 or
anxious or activated under medication.
10
Now, there is a lot of concern about the
11
fact that a lot of these studies are not published,
12
they simply are put in a file drawer.
I think that
13 is
a big concern, it's a big concern for Eric Kahn
14
[ph] and Michael Thase and Norman Sussman, some of
15 the
major people in the field of psychiatry.
16
So, I think the focus of the meeting is
17
sort of unfortunate by focusing on suicidality
18
because I think the big issue here is that we don't
19
have access to the data that we need from the
20
controlled trials, that are simply put in a file
21
drawer by the companies.
22
So, I would like to close by quoting
23
Daniel Conner in the American Journal of American
24
Academy of Child and Adolescent Psychiatry this
25
month. Oh, I will leave the quote
out.
163
1
DR. RUDORFER: We will look it
up. Thank
2
you.
3
Speaker 54, please.
4 Julie Zito, Ph.D.
5
DR. ZITO: I am Julie Zito from
the
6
University of Maryland/Baltimore, and I bring to my
7
comments this morning 20 years' experience in
8
psychiatric pharmacoepidemiology.
9
I would like the committee to consider the
10
following drug safety issues in making their
11
recommendations.
12 First, symptoms like activation and
13
agitation are reported very inconsistently,
14
anywhere from no incidence in a clinical trial to
15 as
many as 55 percent of the children in an SSRI
16
trial. This information suggests
a lack of
17
standardization of measurements and methods with
18
which to assess these events.
19
Second, we need research on behavioral
20
toxicity in order to separate symptoms associated
21
with drug from those associated with the underlying
22
psychiatric disorder. I don't
think we can just
23
assume it.
24
Third, because suicide is a very rare
25
event, we need research that requires active
164
1
surveillance, not passive surveillance, active
2
surveillance in large, well-defined populations.
3 We
have the capacity to do that with research
4
methods in pharmacoepi, but as yet, there is no
5
federal mandate to go beyond Med Watch.
6
Thank you.
7
DR. RUDORFER: Thank you.
8
Speaker 55, please.
9 Joseph Glenmullen, M.D.
10
DR. GLENMULLEN: I am Joe
Glenmullen. I
11 am
a psychiatrist and clinical instructor in
12
Psychiatry at Harvard Medical School and the author
13 of
Prozac Backlash, which describes my experience
14
seeing patients become suicidal on SSRIs.
15
I am here at my own expense because there
16 is a specific side effect of SSRIs called
akathisia
17
that can make some patients so agitated that they
18
feel death would be a welcome relief.
19
This side effect is so well established
20
that it is clearly described with SSRIs in the
21
Diagnostic and Statistical Manual, the DSM, the
22
American Psychiatric Association's official
23
diagnostic manual.
24
If you look at the transcript of the FDA
25
hearing on this very side effect 10 years ago, you
165
1
will see the FDA saying repeatedly we don't know
2
what to do, we need more research.
It is a tragedy
3 to
be here 10 years later and hear the FDA saying
4 the
same thing.
5
The industry's response to this side
6
effect has been to blame the underlying psychiatric
7
conditions of patients, to dismiss legitimate
8
medical case reports as anecdotes, and to scare the
9
media away from the subject, claiming that it would
10
frighten patients away from treatment.
11
Indeed, there is a prevailing
12
authoritarian attitude don't warn patients, you
13
might scare them.
14
Well, I prescribe SSRIs and I warn
15
patients, and they are not frightened away from
16
treatment. Let's stop blaming
patient's underlying
17
psychiatric conditions. Let's
stop blaming the
18
victims and deal with this very real side effect.
19
Thank you.
20
DR. RUDORFER: Thank you.
21
Speaker 56, please.
22 Linda Cheslek
23
MS. CHESLEK: Hello. My name is Linda
24
Cheslek. I am a pediatric nurse
practitioner and I
25
have prescribed medications for pediatric patients
166
1 for
25 years.
2
In the past, I thought that when an FDA
3
drug was approved, that it had gone through a
4 rigorous
battery of independent tests and trials
5
under the auspices of the FDA, but I can longer
6
believe this.
7
Why? Well, this summer I received
this
8
letter from Wyeth. It is a Dear
Doctor letter. It
9
goes to all health care professionals, and it told
10 me
an update on Effexor, that the safety and
11
effectiveness in pediatric patients had not been
12
established, but there were reports of increased
13
hostility, suicide, adverse events, suicidal
14
ideation, and self-harm.
15
This letter that came to my home confirmed
16
what I already knew, that my son, who had a
17
three-week trial of Paxil and Effexor became very
18
much worse. He developed the
akathisia you have
19
been hearing about. He developed
serotonin
20
syndrome symptoms and a seizure.
21
Wyeth had this information for almost
22
seven years. Why did not the FDA require this trial
23
data to be submitted along with the other data?
24 The
FDA allows the drug sponsors to manipulate and
25
massage the data, to present it in a way that they
167
1
feel is promoting their drug, and not the truth.
2 I ask you to require them to submit
all
3 the
data and to give a warning about these
4
medications. When you go to bed
tonight, I hope
5 you
will see my face, the face of my son, and maybe
6 of
other faces of these people, and give a warning.
7
Thank you.
8
DR. RUDORFER: Thank you.
9
We are to speaker 57.
10 Jeff Avery
11
MR. AVERY: Hello. My name is Jeff Avery.
12
My 16-year-old stepson, Brandon Ferris,
13
committed suicide on July 22nd, 2001, about three
14
weeks after he began taking Zoloft.
Brandon was a
15
bright and socially outgoing teen who got along
16
well with others. He was a
black-belt instructor
17 in
Tai Kwon Do, active in the church's youth group,
18 and
held a part-time job.
19
His mother home-schooled Brandon and
20
worked at the Tai Kwon Do School, so she was very
21
active in Brandon's activities.
22
In June of 2001, Brandon
expressed that he
23 was
feeling down, and not his usual energetic self.
24 It
was decided that he should take some time off
25 and
see a counselor.
168
1
The counselor suggested that he see a
2
doctor. The doctor, who found no
physical
3
problems, prescribed Zoloft.
4
Sunday, July 22nd, Brandon and I went to
5
church. On the way home Brandon
volunteered to
6
make a cake for his mother's birthday.
He asked
7
permission to go on a boating trip.
He spent the
8
rest of the day with his friends and an older
9
brother Randy.
10
When he came home from his youth group
11 meeting
at 9:15, he seemed fine. At 9:45 he
asked
12 his
mother about the boating trip. At 10:30
he
13
went to check his e-mail, but his brother was using
14 the
computer. At 11 o'clock, he was found in
his
15
room hung by the neck from a belt in his closet.
16 We
called 911, we performed CPR to no avail.
He
17 was
pronounced dead at the hospital.
18
Reflecting on the day's events, I could
19 not
detect any indication of forethought to
20
suicide. However, later
conversations with others
21
close to Brandon inferred that he may have been
22
having problems with the medication.
23
The obvious question is what happened in
24
Brandon's mind between 10:30 and 10:45.
25
This was not the end of unspeakable
169
1
tragedy. Five months later,
Barbara, unable to
2
cope with the loss of her youngest son, took her
3
life.
4
Since then I have collaborated with
5
Brandon's biological father, Dan Ferris, to obtain
6
information that would point to the cause of
7
Brandon's death. We believe,
after having done
8
much research, that the drug Zoloft had a causal
9
effect in Brandon's final actions.
10
Thank you.
11
DR. RUDORFER: Speaker 58, please.
12 Harry Skigis
13
MR. SKIGIS: What can I say that
hasn't
14
really already been said, but I had a speech
15
prepared and decided to revamp it while sitting
16
here in the audience.
17
I tried to kill myself and luckily didn't
18
succeed. I am still on Paxil because I am hooked on
19 a
nonhabit-forming drug. I don't know if I
will
20
live long enough to see how this thing ends up, but
21 I
am going to try.
22
I have always believed that do unto others
23 as
you would have done to yourself. Would
you
24
people put your children on this drug?
Would you
25
take it yourselves? I doubt it.
170
1
Probably not all the statistics in the
2
world can't bring back the people that are dead
3
because of the irresponsibility of the FDA. How
4 can
I put in any faith in a government that still
5
somewhat denies that cigarettes are addictive?
6
I wonder if you people can sleep at night
7
while your decisions are killing innocent people
8
every day. I leave my life in
your hands and hope
9
that you will apologize to all the people here for
10
your decision and ignorance in this matter and how
11 it
has shattered so many people's lives.
12
I really hope you guys can do something
13
about this or at least tell us who will help us,
14
because a lot of people are dead here today, and
15
it's all in your hands. So good
luck.
16
DR. RUDORFER: Thank you.
17
Speaker 59, please.
18 Pamela Wild
19
MS. WILD: On September 9, 2001,
in a
20
state of confusion and hopelessness, I put a.38
21
Special, Smith & Wesson revolver under my chin and
22
pulled the trigger.
23
In going through withdrawal
from Paxil, I
24
lost all ability to cope and reason and without
25
realizing it, became suicidal. I
suffered from
171
1
sleeplessness, night sweats, light and sound
2
sensitivity, irritability, and dizziness.
3
I was in a constant state of terrible
4
anxiety and felt as though the only thing holding
5 me
together was my skin. I couldn't
understand why
6
others weren't seeing things my way, as though I
7 was
speaking in another language. I was told by my
8
therapist that I had drifted into a fantasyland.
9
She said it was though my system had been
10
poisoned somehow, I was told not to worry, the only
11 way
to die from this drug was to fill a tub with
12
Paxil and water and drown in it.
13
The side effects I experienced on Paxil,
14
even though I reported them to my doctor, were
15
dismissed because no one was warned that Paxil
16
could cause what I was experiencing.
17
If I, at 41 years old, could not
18
articulate what was happening, how do you expect a
19
child to?
20
There is no real medical explanation for
21 my survival.
The front of my face was blown away,
22
leaving a hole large enough to
encompass a man's
23
fist. The bullet miraculously
only took two-thirds
24 of
my tongue, most of my mandible and my cheek
25
bones. The maxilla was shattered.
172
1
The orbit of my left eye was broken and
2
forced the eyeball out onto what remained of my
3
left cheek. It completely
destroyed my hard and
4 soft
palate along with my nose and sinus cavity.
5
I was blessed, though. I may not
able to
6
taste or smell, but at least I lived.
I can see,
7
talk, and I can hear. But more
surprising than any
8 of
those, I have brain function. I truly
believe
9 my
life was spared for a reason. That reason is so
10 I
can prevent others from experiencing what I
11
experienced.
12
DR. RUDORFER: Thank you very
much.
13
We are up to speaker 60. Thank you.
14 Karen Barth Menzies
15
MS. MENZIES: Good morning. My name is
16
Karen Barth Menzies and I am an attorney for Baum,
17
Hedlund. We represent several
thousand SSRI
18
victims. We have been doing this
for 12 years.
19
The U.S. Code of Federal Regulations
20
201.57 mandates that you require the drug companies
21 to
warn when there is reasonable evidence, not
22
causation, reasonable evidence of an association of
23 a
serious risk.
24
The clinical researchers who did these
25
trials on kids and the drug companies themselves
173
1
confirmed that there are multiple events of
2
suicidality caused by the drug.
The methodology
3
that you are going to be using is designed to
4
explain away those events.
5
Even Dr. Laughren admits in the memo he
6
gave you for this hearing today that there is
7
evidence in these trials of an increased risk of
8
suicidality, reasonable evidence is there. If
9
there is reasonable evidence, you must make them
10
warn.
11
Serious risk, we certainly have that.
12
Akathisia, psychosis, mania. When
you are looking
13 at
this data, you are not just looking at the
14
suicide, also look for signs of akathisia and
15
psychosis and mania. These aren't
as easily
16
explained away by the drug companies, by blaming
17 the
disease, by blaming the victims.
18
When you take the potentially fatal risk
19 and
couple that with lack of efficacy of these
20
cases, why take that risk especially when it comes
21 to
our kids.
22
Paul Leber [ph] predicted this day when he
23
said that the FDA would come under attack because
24
they weren't as demanding as they ought to have
25
been when they were looking at the efficacy of the
174
1
antidepressant products.
2
Put me out of business for the right
3
reasons, warn about these drugs and disclose.
4
DR. RUDORFER: Thank you.
5
Speaker 61, please.
6 Amy Coburn
7
MS. COBURN: Hi. My name is Amy Coburn.
8 I
have flown here from Salt Lake City, Utah, at my
9 own
expense.
10
I am here on behalf of my father, myself,
11 and
my family. My father's name was Wayne
Coburn.
12
Most people remember him as a man full of life and
13
willing to help anyone in need.
14
I remember my dad as a man who loved his
15
family very much and was very loved in return, a
16 man
full of ideas and hope for the future, but like
17
many people, he found he got a little down in the
18
wintertime. He was diagnosed with
seasonal
19
depression without suicidal tendencies.
20
When I was 13 years old, he was put on
21
Paxil. Three weeks later he
pulled his car into an
22 old
factory garage, started his engine, and there
23
waited until he died of carbon monoxide poisoning.
24
This naturally shocked me and my family
25 and
we all had a hard time coping with his death.
175
1 I
started going to a counselor to work through my
2
grief, and I was put on Paxil, the same drug my
3
father was on.
4
I started acting differently, then very
5
soon after I started having suicidal thoughts, mood
6
swings, I was fighting with my friends, and the one
7
thing my mom noticed is that I wouldn't talk about
8 how
I was feeling. The only thing she could
get
9 out
of me was "I am fine, leave me alone."
10
Six weeks after I was put on the drug, I
11
stayed home from school, wrote my good-bye letters,
12 and
swallowed a cupful of poisonous bathroom
13
cleaner. I immediately got scared
and ran to my
14
neighbor's house. She called 911
and luckily I
15
survived and I am standing here today.
16
We soon found out that we weren't the only
17
ones who had problems with these drugs.
Hundreds
18 of
families have lost people they love because they
19 had
no idea of the effect they could have on a
20
person's mind. All me and my
family want are
21
warnings on these drugs.
22
DR. RUDORFER: Thank you. I am sorry, we
23 are
out of time. Thanks.
24
Speaker 62, please.
25 Sharon McBride
176
1
MS. McBRIDE: I am here as a
mother and I
2 am
here at my own expense.
3 When our daughter was 13 years old, she
4
came to me and said that something was wrong with
5
her. After discussion, I took her
to the emergency
6
room where she was diagnosed with depression.
7
After three years of intense psychotherapy
8 to
discover and help the cause, she experienced her
9
first manic episode. She was
hospitalized and
10
given lithium and a mild dose of antipsychotic
11
medication for a brief period of time.
12
The resulting acne and weight gain caused
13 her
further depression thereafter. Due to my
14
inability to accept the diagnosis, we took her to a
15
psychologist rather than a psychiatrist to get a
16
middle-of-the-road opinion.
17
Because she was so depressed, we did
18
eventually see a psychiatrist again, and she was
19
prescribed one of the SSRI medication, Zoloft.
20
Shortly after beginning this treatment, she had a
21
serious suicide attempt. The
doctor at the
22
hospital first thought that it was just another
23
attempt trying to get attention, but after he
24
interviewed her, his opinion changed.
25
While she had been depressed, she had
177
1
never attempted suicide before this time.
2
Eventually, she was prescribed three medications,
3 one
of which was Paxil. Three different
times in
4 her
life she abruptly stopped taking the
5
medications including Paxil, which resulted in
6
manic episodes.
7
Before her last episode, she had been
8
stable for five years. Then,
during a very
9
stressful time with her grandmother dying, she
10
abruptly stopped the Paxil and experienced her
11
worst manic episode with hallucinations and other
12
health problems.
13
She finally had to be court-ordered into
14 the
hospital and it devastated her life. She
lost
15 her
job as a security assistant at a hospital, and
16 her
roommates could no longer live with her because
17
this was not the person that they had known and
18
loved.
19
That was two years ago and she is just
20
beginning to put her life back together.
I would
21 encourage the committee to look very closely
at the
22
suicide attempt ratio for children and teenagers
23
taking these SSRI medications.
24
Thank you.
25
DR. RUDORFER: We are up to I
believe our
178
1
final speaker of the morning session, and that is
2 Dr.
Thomas Moore.
3 Thomas Moore, M.D.
4
DR. MOORE: Good afternoon. I represent
5
Drug Safety Research. I have
completed two studies
6
that raise additional questions about the safety of
7
antidepressant drugs, and both of those studies
8
should be in your binders.
9
The first of those concerns the medical
10 use
of these drugs, who are taking them, and the
11
headline finding is that in the four-period 1998 to
12
2001, use of antidepressant drugs in children
13
doubled.
14
The second finding is that less than 10
15
percent of these cases were these drugs being
16
prescribed for FDA-approved use, and the remaining
17 90
percent of the cases, they were for unapproved
18 use
or ones that raised safety concerns. Let
me
19
give you some examples of what I found.
20
Among boys 6 to 12 years old,
52 percent
21 of
the use was for treating attention deficit or
22
conduct disorders typically in combination with an
23
antipsychotic or a stimulant, such as Ritalin.
24
Now, I know of no scientific evidence that
25
says that combination therapy is effective in these
179
1
disorders, and I know of no evidence that it is
2
safe either.
3
As you go on, combination therapy was very
4
common in the real world.
Twenty-two percent were
5
taking two antidepressant drugs, 17 percent were
6
taking drugs that were ineffective in clinical
7
trials, 42 percent were taking two of more
8
antidepressant drugs.
9
So, what we are seeing is when drugs are
10
ineffective, rather than abandoning them or trying
11
alternatives, doctors increase the dose or combine
12 the
drugs in ways, the safety of which we are not
13 aware.
14
The second major study that I submitted to
15 you
today is of the adverse event experience,
16
largely the same data set, but different criteria
17
from what the FDA has conducted.
18
The two key findings there are, number
19
one, it appears based on the medical use of these
20
drugs that these drugs cause suicidal and related
21
behaviors at double the expected rate compared to
22
adults. So, they seem to be being
reported more
23
frequently in children.
24
The second finding is there appeared to be
25 no
difference in adverse event reports between the
180
1 two
drugs for which there were warnings, and those
2
four drugs for which we do not have warnings.
3
DR. RUDORFER: Thank you, Dr.
Moore.
4
We will now end our morning session.
I
5
want to thank all our open public hearing speakers
6 for
raising very important issues for the
7
committee. I believe we will have
two additional
8
public speakers during the afternoon session, but
9 we
are now going to take our lunch break.
10
We will reconvene at 1 o'clock.
11
[Whereupon, at 11:59 a.m., the proceedings
12
were recessed, to be resumed at 1:00 p.m.
181
1
A F T E R N O O N P R O C E E D I
N G S
2 [1:10 p.m.]
3
DR. RUDORFER: Good afternoon.
4
We are going to begin this afternoon's
5
session with several speakers from the FDA. What
6 we
are going to do is hear a total of six speakers
7 from
the FDA, as well as our two remaining public
8
speakers. There will be a break
along the way.
9
I am going to ask the committee to save
10
your questions until the end. We
will have a lot
11 of
time for discussion later this afternoon.
12
First, I would like to introduce Dr.
13
Gianna Rigoni from the Office of Drug Safety of the
14
FDA.
15
Pediatric and Adolescent Antidepressant
16 Drug Use in the U.S.
17
DR. RIGONI: Thank you, Dr. Rudorfer, and
18
good afternoon.
19
[Slide.]
20
Today, I would like to describe for you
21
antidepressant drug use trends in children and
22
adolescents in outpatient settings to provide a
23
context for further discussions this afternoon.
24
[Slide.]
25
First, I will describe the use of selected
182
1
antidepressant products by prescriptions dispense
2 in
the United States, followed by the proportion of
3
those prescriptions dispensed to 1- to
4
17-year-olds.
5
Next, I will examine the specialties of
6 the
physicians responsible for prescribing these
7
products to children and adolescents.
8
Finally, I will identify the primary
9
diagnoses for which these products are used in
10
these populations.
11
[Slide.]
12
The antidepressants examined in this
13
analysis include the selective serotonin reuptake
14
inhibitors, or SSRIs, as we refer to today, and the
15
atypical antidepressants seen on this list here.
16
Atypical include nefazodone, venlafaxine, and
17
mirtazapine.
18 These products will be presented at
the
19
molecule level, therefore, fluoxetine will refer to
20
Prozac, Prozac Weekly, Sarafem, and all generic
21
fluoxetine equivalents, and so on, for each
22
product.
23
All references to the term
24
"antidepressants" in this talk will refer only to
25
these 10 products. Tricyclic
antidepressants,
183
1
MAOIs, and other products used to treat depression
2
were not examined for this analysis.
3
[Slide.]
4
At this time, only three SSRI products
5
have FDA-approved labeling for use in pediatric
6
population. Fluoxetine is the
only product
7
approved for the treatment of pediatric major
8
depressive disorder at this time, while fluoxetine,
9
sertraline, and fluvoxamine are approved for the
10
treatment of obsessive-compulsive disorder in this
11
population.
12
Although only three products have
13
FDA-approved labeling for the treatment of MDD and
14
OCD, use of SSRIs and atypical antidepressants
15
outside of current FDA labeling in pediatrics is
16
endorsed by many in the medical community through
17
various clinical practice guidelines.
18
[Slide.]
19
I will now describe the methods that were
20
used in this analysis.
21
[Slide.]
22
Since data for 2003 was not complete in
23
time for this presentation, we will look at drug
24 use
trends from 1988, the year fluoxetine was
25
launched, through 2002.
184
1
When examining trends and prescriber
2
specialties and diagnoses related to prescribing
3
these products, trends over a five-year period of
4
time, from 1998 to 2002, were used.
Data on drug
5
utilization will be presented from sources FDA has
6
available under various contracts.
For this
7
analysis, outpatient data was obtained from two
8
IMS Health audits.
9
IMS is a source of marketing data commonly
10
used by the pharmaceutical industry and government
11
agencies, and is used to obtain numbers of
12
prescriptions dispensed, as well as diagnoses
13
related to the recommendation of pharmaceutical
14
products in physicians' offices in the U.S.
15
[Slide.]
16
The first IMS Health Audit examined the
17 National
Prescription Audit Plus, or NPA Plus, as I
18
will refer from now on, measures dispensed
19
prescriptions from the outpatient pharmacy settings
20
seen here. We have chain,
independent, mass
21
merchandisers, food stores with pharmacies, mail
22
order and long-term care pharmacies.
23
The number of estimated prescriptions
24
dispensed are obtained from a sample of
25
approximately 22,000 pharmacies in the U.S., and
185
1 are
projected nationally.
2
[Slide.]
3
Next, we examined data from the National
4
Disease and Therapeutic Index Audit, or NDTI, from
5 IMS
Health. NDTI collects data on drug
products
6 and
diagnoses mentioned during office-based
7
physician visits.
8
A mention is a physician's treatment
9
intention where they believe one of the selected
10
antidepressants is appropriate, and important to
11
remember is it could result in either a
12
prescription, a refill authorization, or samples
13
given to the patient.
14
Information on trends of diagnoses,
15
patients, and treatment patterns occurring during
16
these visits are linked to each drug.
NDTI data
17 are
obtained from a sample of 2,000 to 3,000
18
physicians representing approximately 100
19
specialties in the U.S., and are projected
20
nationally to reflect national prescribing
21
patterns.
22 The exact distribution of the
specialties
23
participating in the sample each year is
24
unavailable at this time, but is roughly
25
proportional to the distribution of office-based
186
1
practice specialties in the United States.
2
[Slide.]
3
We will now examine antidepressant
4
prescription trends, prescriber specialties, and
5
diagnoses from 1988 through 2002.
I will first
6
describe antidepressant use in the U.S. for all
7
ages and then zoom in more specifically on the
8
younger pediatric and adolescent age groups.
9
[Slide.]
10
It was estimated that over 157 million
11
prescriptions for SSRIs and atypical
12
antidepressants were dispensed in the United States
13 for
all ages in 2002. The market leaders
among
14
these 10 products were sertraline, accounting for
15
over 31 million prescriptions, followed closely by
16
paroxetine, with 30.5 million.
17
[Slide.]
18
I will now graphically show you the use
19
trends of these products since the launch of
20
fluoxetine. This graph has a lot
of information on
21 it,
but it displays the national estimates of
22
antidepressant use in the U.S. in millions of
23
prescriptions dispensed for all ages, so this y
24
axis here is in millions, and each product is
25
represented by a different color line.
187
1
Here, we see how the four products on the
2
previous slide make up the highest volumes
3
dispensed. Here, you see
paroxetine, sertraline,
4
fluoxetine, and citalopram. But
more importantly,
5 we
see that for the past 15 years, there is an
6
increasing and substantial number of prescriptions
7
dispensed in outpatient pharmacy settings for these
8
products.
9
We will now examine the estimated use of
10
these products in the younger pediatric and
11
adolescent populations.
12
[Slide.]
13
First, I must describe how we estimated
14
these numbers. Since NPA Plus
data does not
15
include the demographic information about the
16
patients receiving each prescription, we used NDTI
17 to
estimate the number of prescriptions dispensed
18 to
1- to 17-year-olds.
19
NPA Plus and NDTI were designed by IMS to
20 be
comparable in terms of volume of prescriptions
21
dispensed and the proportion of office visits
22
mentioning products dispensed in larger volumes.
23
So, to estimate the number of SSRI and
24
atypical antidepressant prescriptions dispensed to
25 1-
to 17-year-olds, the proportion of office visits
188
1 in
that population that involved the mention of one
2 of
these products were applied to the total number
3 of
prescriptions dispensed for that year.
4
[Slide.]
5
Applying the proportion of office visits
6 to
the national prescription estimates for 2002, I
7
present to you the top five selected
8
antidepressants in thousands of prescriptions
9
dispensed to 1- to 17-year-olds.
10
Approximately, 10.8 million total
11
prescriptions were dispensed for all SSRIs and
12
atypicals in this population, representing a
13
substantial 7 percent of the market in 2002.
14
Sertraline accounted for the highest
15
volume of prescriptions dispensed, at 2.9 million,
16 and
paroxetine followed closely with approximately
17 2.2
million, and this is for 2002.
18
Next, I will more closely examine these
19
patterns by breaking the 1- to 17-year age group
20
into the younger pediatric population, which will
21
represent 1- to 11-year-olds, and the adolescent
22
population, which will represent 12- to
23
17-year-olds.
24
[Slide.]
25
When we examined use in these
189
1
subpopulations, we can still see substantial use of
2
these products in both groups.
The younger
3
pediatric population accounted for approximately
4 2.7
million prescriptions dispensed in 2002.
5
Sertraline again was the most commonly prescribed
6
product, accounting for about 31 percent of
7
dispensed antidepressants, followed by paroxetine
8 and
then fluoxetine.
9
The adolescent population accounted for
10
approximately 8.1 million prescriptions dispensed
11 in
2002, and this is close to about 5 percent of
12 all
antidepressants dispensed in that year.
13
Again, sertraline was the most commonly
14
prescribed, accounting for 26 percent, but this
15
time followed closely by paroxetine, with 22
16
percent.
17
[Slide.]
18
Now that we better understand the trends
19 in
prescriptions dispensed for these products to
20
children and adolescents, we need to better
21
understand the specialties of the physicians most
22
often prescribing these products.
23
The top prescribers of SSRIs and atypical
24
antidepressants in 1998 were compared to those of
25
2002, and the top ranked specialties are listed
190
1
here by age group and by year.
2 Here, it makes sense to see
psychiatry as
3 the
top prescribing specialty over time since it is
4
hard to diagnose mental illness in younger
5
populations. There does appear to
be some shifting
6 in
prescribers over time, though, as the pediatric
7
specialty becomes responsible for a more
8
substantial proportion of mentions of these
9
products in 2002.
10
As you can see, the proportion of
11
pediatricians prescribing doubles over that
12 five-year period in both populations, or
nearly
13
doubles in adolescents.
14
[Slide.]
15
Now, we will examine the diagnoses most
16
commonly associated with these products in
17
office-based practices. All diagnoses naturally
18
fell into the following four categories:
19
Mood disorders, represented here by the
20
blue portion of the bar, include bipolar affective
21
disorders and all depressive disorders; anxiety
22
disorders are represented by the red portion of the
23
bar, and they include anxiety, obsessive-compulsive
24
disorder, and phobias.
25
Attention-deficit disorder is represented
191
1 by
the yellow portion of the bar, and Other
2
disorders are represented by the green portion.
3
Now, these Other disorders include other
4
diagnoses for psychiatric illnesses, such as
5
adjustment disorder, personality disorder, and
6
psychotic disorders, as well as including diagnoses
7 for
autism, migraine, convulsions, menstrual
8
symptoms, eating disorders, and drug and alcohol
9
dependency.
10
We see nearly 900,000 physician office
11 visits involved the mention of an
antidepressant in
12 the
younger pediatric population in 2002.
This
13
represents approximately 1.6 percent of all visits
14 in
the U.S. for these products across all ages.
15
We also see that anxiety and mood
16
disorders were the most common diagnoses in 2002,
17
accounting for 30 percent and 26 percent,
18
respectively, in this population.
19
Office visits involving the mention of one
20 of
these products in adolescents is much higher, at
21 2.6
million visits for 2002, and that represents
22
about 5 percent of the visits in the U.S. Mood
23
disorders were the most common diagnoses treated
24
with this product, accounting for nearly 60
25
percent.
192
1
Next, we will look at these bars more in
2
depth as we examine diagnoses trends for specific
3
drugs in younger pediatric and adolescent
4
populations.
5
[Slide.]
6
This slide contains a lot of information,
7 but
I believe it is important to show that not all
8 of
these products are used in the same way in the
9
younger pediatric population.
10
The following graph displays the
11
distribution of diagnoses for the top five
12
antidepressants mentioned in 2002 to this
13
population. Notice here the
percent scale on the y
14
axis. Each bar represents all
mentions for these
15 products
to this age group, and the percent is what
16
percent of the mentions for that drug were for each
17
disorder.
18
In the younger pediatric population, we
19 see
some variation in how these products are being
20
used, and from the previous slide, we saw that both
21
anxiety and depression or mood disorders were
22
primarily treated with these products. It is seen
23
right here in the graph.
24
When we look at the top five, we also see
25 that
bupropion has the distinctive use in treating
193
1
attention deficit disorders in this population, so
2
that middle bar signifies bupropion, and the yellow
3
portion is ADD.
4
[Slide.]
5
In the adolescent population, we see there
6 is
not much variation in prescribing of these
7
products. Mood disorders were the
primary
8
diagnosis being treated with all five products, but
9 we
do, however, once again see this distinctive use
10 of
bupropion for attention deficit disorders.
11
[Slide.]
12
Next, we wanted to determine if
13
prescribing trends for these products has changed
14
over the last five years. In the
younger pediatric
15
population, we saw a shift in prescribing from 1998
16 to
2002, from these antidepressants being used
17
primarily to treat mood disorders, which were
18
identified before as bipolar and other depressive
19
disorders, to being used more to treat anxiety
20
disorders, such as OCD and other anxiety or phobia
21
disorders.
22
We saw that in the adult population, there
23 was
no change in prescribing from 1998 to 2002, and
24
that continuously over this time period, these
25
products were used to treat mood disorders in this
194
1
population.
2
[Slide.]
3
Some limitations of our drug use data
4
analysis are, first, data on prescriptions
5
dispensed include prescriptions filled in
6
outpatient pharmacies only.
Inpatient and
7
institutional use of these products was not
8
included in this analysis.
9
Secondly, prescriptions dispensed to 1- to
10
17-year-olds were extrapolated from the proportion
11 of
these populations visiting a physician and
12
receiving a prescription sample or refill
13
authorization for one of these products, and this
14
methodology has not yet been fully validated.
15
Finally, data on diagnoses related to the
16 use
of these antidepressants reflects office-based
17
physicians prescribing based on a small sample of
18
physicians. The small sample size
may make these
19
numbers unstable and could underestimate the
20
prescribing patterns of certain subspecialists.
21
Also, since these patients are not
22
followed into the pharmacy after their appointment,
23 a
patient may not actually fill the antidepressant
24
prescription.
25
[Slide.]
195
1
In conclusion, use of SSRIs and atypical
2 antidepressants is substantial in children and
3
adolescents, and appears to be increasing rapidly
4
every year. Pediatric
specialists, pediatricians,
5 and
primary care providers continue to be the
6
leading prescribers of these products, and over the
7
past five years, the proportion of pediatricians
8
prescribing these products has nearly doubled.
9
Finally, diagnoses related to the use of
10
these antidepressants are slightly different among
11 the
younger pediatric population who are being
12
treated for mood and anxiety disorders, and the
13
adolescent population who are being treated mostly
14 for
mood disorders.
15
Thank you.
16
DR. RUDORFER: Thank you very
much.
17
This morning we heard from Dr. Murphy
18
about the mandated adverse event review associated
19
with one-year post-exclusivity for some
20
medications. Now, I am pleased to
welcome Dr.
21
Solomon Iyasu from the Division of Pediatric Drug
22
Development who will give us a review of that
23
information for paroxetine and citalopram.
24
One-Year Post-Exclusivity Mandated Adverse
25
Event Review for Paroxetine and Citalopram
196
1
DR. IYASU: Good afternoon.
2
Today, I am going to be presenting adverse
3
event reports that have been received by FDA and
4
reviewed as mandated by the Best Pharmaceuticals
5 for
Children Act.
6
[Slide.]
7
The Best Pharmaceuticals for Children Act
8 was
enacted January 4, 2003, and Section 17
9
mandates to FDA to review all adverse events for
10 one
year post-exclusivity determination, and then
11
report to the Pediatric Advisory Subcommittee for
12
their review.
13
[Slide.]
14
The data source for my presentation, as
15
well as Dr. Mosholder's presentation following
16
mine, is the FDA's Adverse Event Reporting System,
17
which is a spontaneous and voluntary reporting
18
system.
19
FDA maintains an electronic database of
20
postmarketing reports of adverse drug reactions,
21 and
reporters to this system include health care
22
providers, pharmacies, consumers, and
23
pharmaceutical manufacturers. A
large majority of
24
these reports come from manufacturers.
25
[Slide.]
197
1
To make today's presentation relevant to
2
today's topic, I will be focusing the later part of
3 my
presentation on the psychiatric adverse events
4
that have been reported during this one-year
5 post-exclusivity period.
6
[Slide.]
7
To give you some background about the drug
8
that I will be talking about today, paroxetine is
9 an
antidepressant that belongs to the class of
10
drugs which are called SSRIs, is marketed by
11
GlaxoSmithKline.
12
Adult indications that are approved by FDA
13
include major depressive disorder,
14
obsessive-compulsive disorder, panic disorder,
15
social anxiety disorder, generalized anxiety
16
disorder, and posttraumatic stress disorder.
17
The typical adult dose, which are
18
approved, are 20 to 60 milligrams per day. There
19 are
no approved pediatric indications, and the
20
exclusivity was granted January 27, 2002.
21
I have to point out here that exclusivity
22 to
a sponsor can be granted without getting an
23
approved indication as long as they do the study
24 set
that have been asked in the written request
25
that FDA issues, and that they have met the
198
1
criteria fairly as part of the written request.
2
[Slide.]
3
To give you some important information
4
that is on the label already, paroxetine is
5
Pregnancy Category C drug, which means that
6
paroxetine has not been studied in pregnancy and
7
therefore should be used only if potential benefit
8
justifies the risk to the fetus.
It also should be
9 used
with caution in nursing mothers.
10
There is also information on the
11
Precautions section of the label, suicide risk is
12
inherent in major depressive disorders especially
13
before remission occurs, therefore, high-risk
14
patients should be supervised very closely
15
especially during the initial phases of therapy.
16
There are also similar precautions about
17
mania and also about seizures, and recommendations
18 to
use this medication with caution in patients who
19
have a history of mania or seizures.
20
There is also, on the same section,
21
adverse events with abrupt discontinuation, which
22
includes symptoms like agitation, anxiety,
23
dizziness, sensory disturbance, that is related to
24
withdrawal, and therefore, the recommendation is to
25
taper it slowly.
199
1
[Slide.]
2
Now, I would just summarize the drug use
3
trends for paroxetine, extensively discussed by
4
Gianna before me, but paroxetine is the second most
5
commonly used SSRI in children.
Both pediatric and
6
adult prescriptions have steadily increased between
7
1999 and 2003.
8
The main diagnosis linked with its use
9
include depression, anxiety, and
10
obsessive-compulsive disorders in children.
11
Pediatric patients account for
12
approximately 3.5 percent of the total U.S.
13
prescriptions of Paxil between July 2002 and June
14
2003.
15
[Slide.]
16
To give you an overview of the adverse
17
event reports that have been received by FDA since
18 the
original marketing for this medication, there
19
were a total of 17,000 adult and pediatric reports
20
including domestic and foreign that were received
21 by
FDA. This included duplicates, as well,
and 68
22
percent of them were domestic.
Less than 5 percent
23 of
these reports were in pediatric patients
24
Looking at the top 20 pediatric adverse
25
events for this entire period,
the pediatric
200
1
adverse event in the top 20 was similar to those
2
reported in adults. The majority
were limited
3
events related to mostly the events that resulted
4
from maternal exposure, prenatal exposure.
5
[Slide.]
6
Looking at the annual reports of adverse
7
events for this drug since 1992, there was
8
distinctly an increase in 2002 compared to prior
9
years. These data, the bar graphs
represent raw
10
counts of adverse events that were received by FDA,
11 and
do not exclude the duplicate reports, and they
12 are
unadjusted for use.
13
You will notice that the last bar graph,
14
which is really representing the first half of the
15
year, the numbers were 87, which seems to suggest
16
that there is this continuing increase that was
17
observed in 2002.
18
[Slide.]
19
Just to provide some context, I want to
20
mention the timeline for some important events that
21 may
have some importance in this deliberation.
22
First, the yellow line as you see here is
23 the
period of that inclusive post-exclusivity
24
one-year period, and during that period, there was
25 a
BBC show, which is "The Secret of Seroxat," that
201
1 was
aired on October 2002 in the British TV, which
2
subsequently got very widespread media coverage
3
around the U.S. and other parts of the world.
4
In 2003, the British Government warned
5
against the use of Paxil, and FDA issued a talk
6
paper on Paxil for its treatment of depression in
7
June 2003. Following the
post-exclusivity period,
8 in
October 27, 2003, there was an FDA public
9
advisory for antidepressants and suicide.
10
The contents of this will be discussed
11
more fully, I think when Dr. Laughren presents his
12
talk.
13
[Slide.]
14
Now, focusing on the mandated period,
15
which is a one-year post-exclusivity
determination
16
period, after manual review of the reports, there
17
were a total of 127 unduplicated pediatric adverse
18
event reports. The gender
distribution was 61
19
females and 59 males.
20
The age distribution for these 127 reports
21
were zero to 2, about 32, which mostly represented
22
maternal exposures or prenatal exposures; 2 to 5,
23
about 6, the majority were actually in the older
24
kids.
25
The outcomes for the 127, 10 percent of
202
1 the
reports included outcomes of death, which were
2
13. Approximately, a third of
them also ended up
3 in
hospitals or at ER visits.
4
[Slide.]
5
The age distribution, to give you a flavor
6 by
type of exposure, is that in the
7
maternal/breastfeeding exposure, the majority were
8 in
males, and in the direct pediatric exposure, the
9
majority were females.
10
The age distribution, as expected, in the
11
maternal/breastfeeding group, 32 of them were less
12
than 2 years of age, which actually most of them
13
were in less than 1 month. In the
direct exposure,
14
most of the reports came from older kids, mostly 12
15 to
16, and 6 to 11.
16
[Slide.]
17
Looking at the pediatric exposures by
18
reasons for exposure to paroxetine, looking at 127,
19 33
of them were maternal exposure or breastfeeding
20
exposure, and the rest of them are described in
21
depression/dysthymia, 28; anxiety/panic or
22
posttraumatic syndrome disorder, about 15; ADHD, 2;
23
OCD, 1. There were about 18 of
them that had
24
multiple diagnosis of psychiatric conditions, and
25
then Others, which are a smattering of other
203
1
conditions which occurred in single digit. Unknown
2
were in 21, we did not have any information in the
3
reports about what the reason for exposure was.
4
[Slide.]
5
Looking again at the 127, concomitant
6
medications were described in 55 out of the 127
7
reports. Specifically, paroxetine
was mentioned as
8 the
only drug used in 5 cases. In most, it was
9
actually not described whether there was
10
concomitant medication or not.
11
Reporters for this 127, looking at the
12
type of reporter, one-third of the reports were
13
actually from health professionals, two-thirds of
14
them were from consumers, media, or litigation
15
sources, which is really atypical in the sense that
16
most of the reports that we get at FDA, two-thirds
17
often come from health professionals.
18
The dose range in the reports range from 5
19 to
60 mg/day. This excluded the
20
maternal/breastfeeding exposure. This was really
21
looking at the children that were exposed directly.
22
[Slide.]
23
The pediatric adverse events, looking at
24
them from predominant events, there were about 68
25
psychiatric adverse events, and discontinuation
204
1
syndrome or decreasing dose was observed in 7,
2
maternal exposure in 33 as previously described.
3
Today, I am going to be focusing more on
4 the
psychiatric adverse events, which are 68
5
reports that were received, and then the rest of
6 the
presentation in terms of describing the other
7
events will be in tomorrow's presentation which I
8
will be doing to the same committee.
9
[Slide.]
10
Looking at those 68 adverse events, and
11
looking at labeled and unlabeled events, there were
12
about 9 completed suicides reported, 17 suicide
13
attempts, and suicidal ideation in 11 patients, and
14
occurrence of other psychiatric symptoms that
15
included mania, impulsivity, disinhibition, or
16
obsessive behavior, and so forth.
17
Then, unlabeled events were self-injurious
18
behavior in about 10 patients, completed homicides
19 in
about 4, and then aggression, hostility,
20
homicidal ideation in about 8 patients.
21
[Slide.]
22
Looking more closely at the psychiatric
23
events, the gender distribution was 57 percent of
24
them were in females. The age
distribution, most
25 of
them were in the older children 12 to 16 years
205
1 of
age, 60 percent of them, and 35 percent in 6 to
2 11
years old.
3
Concomitant medications were described
4
only in 24 patients out of the 68, we did not have
5 any
information on the rest of them. In 20
of the
6 24
patients, there were other
psychotherapeutic
7
agents being used, as well.
8
Discontinuation or decrease in dose was
9
noted in about 11 of the 68 patients that were
10
reported.
11
[Slide.]
12
Going more in detail as to the
13
discontinuation or decrease in dose with respect to
14
psychiatric events, among the completed suicide, 1
15 out
of the 9, there was discontinuation or decrease
16 in
dose involved; suicidal attempts, 5 out of the
17 17,
and 2 out of the 4 for homicides, and then 3
18 out
of the 8 for the aggression/hostility/homicidal
19
ideation.
20
[Slide.]
21 Looking closely at the suicide
attempts,
22
which were about 17, the majority of them were
23
being treated for MDD or bipolar disorder.
24
Concomitant medications were mentioned in
25
approximately one-third of these patients, and
206
1
discontinuation or decrease in dose in
2
approximately one-fourth.
3
[Slide.]
4
Pediatric deaths, there were a total of 13
5 as
I mentioned before. Because of the topic
today,
6 I
will talk about the 9 completed suicides, and the
7
rest of the patients will be discussed in
8
tomorrow's presentation.
9
[Slide.]
10
Among the 9, the age distribution was 12
11 to
16 years, and then the gender distribution of 5
12
females and 4 males. Initial
diagnosis in these
13
patients, 5 of them was major depressive disorder,
14 1
explosive disorder, in 3 of them it was not
15
known.
16 Duration of treatment ranged from 14
days
17 to
1 year. Discontinuation was mentioned in
2
18
patients. Concomitant
medications, that included
19
also some psychotherapeutic agents, was mentioned
20 in
4 patients, and there was possible substance
21
abuse in 4 patients, and a history of prior
22
attempts in 3 of them.
23
[Slide.]
24
In summary, the causality assessment was
25
very difficult in many of the reviews that we have
207
1
done with these reports, and many of the
2
psychiatric events that were described in the
3
reports occurred in patients with underlying
4
psychiatric disorders, therefore, severity of
5
illness/underlying disease may play a role, and it
6 was
very difficult to disentangle its effect from
7
what might have been going on.
8
There is also a prior history of suicide
9
attempts in some of the patients, and in others,
10
there was no negative history of this.
The other
11
factors in terms of patient factors are concomitant
12
medications that were mentioned in several of these
13
patients, and also the lack in others.
So, there
14 is
the variability in terms of the type and the
15
quality of the reports that we got.
16
In terms of the reporting factors, there
17 was
inadequate detail in describing the event.
18
They also varied in terms of descriptions that were
19 in
the reports.
20
The timing of event in relationship to the
21
medication was not always clear in many of these
22
reports, and also ascertainment of reported events
23 by
medical professions was absent in many of these
24
reports. The lack of follow-up
information also
25
made it difficult to assess.
208
1
[Slide.]
2
I also want to mention the nature of the
3 data
system that we have, which is really a passive
4
spontaneous and voluntary system, and it suffers
5
from a number of limitations.
6
Often there is underreporting of important
7
events, and there may be also the reporting biases
8
that are influenced by either media publicity, and
9
also the well-known variability in terms of reports
10
that we get or the frequency of report related to
11 the
length of time that a drug has been in the
12
market. In the early period of
the marketing,
13
there are more reports than later.
14
The report quality, as I said, also may
15
vary, missing details, example, concomitant
16
medications is a common problem.
Also, because
17
this is really enumerated data, we could not really
18
estimate true incidence rate of events or exposure
19
risk for many of these medications that we have
20
reports for.
21
So, the AERS database has some serious
22
limitations in terms of interpreting the data that
23 we
have.
24
[Slide.]
25
In closing, the psychiatric events
209
1
described in the adverse event reports may actually
2
reflect to the underlying disease, because many of
3
these events are also unexpected in other natural
4
progression of the disease or part of the disease
5
picture.
6
It may also be a drug effect or other
7
concomitant medication, or it may actually be lack
8 of
effectiveness of the drug, and it is very
9
difficult from these reports to sort out what is
10
going on.
11
Therefore, evaluation of the controlled
12
trials is necessary to sort out causality in terms
13 of
the observed adverse events.
14
[Slide.]
15
I am going to continue with the next drug,
16
which is citalopram, but I would like to
17
acknowledge the following individuals for their
18
contribution for their review.
19
[Slide.]
20
Next, I will cover, as mandated by BPCA,
21
citalopram, and will be talking about the adverse
22
events in detail.
23
[Slide.]
24
To give you some background again about
25
citalopram, it's an antidepressant belonging to
210
1
SSRIs, and marketed by Forest Pharmaceuticals.
2
Its current approved adult indication is
3 for
major depressive disorder. The adult
dose
4
ranges from 20 to 40 mg/day.
There are no approved
5
pediatric indications.
6
The original market approval was July 17,
7
1998, and exclusivity was granted July 9, 2002.
8
[Slide.]
9
Again, to mention some of the relevant
10
safety labeling which already exists, Pregnancy
11
Category C, as I mentioned before, and also a
12
caution against the use in nursing mothers.
13
There is also a Precaution section that
14
mentions, similar to what is observed for Paxil,
15
suicide risk inherent in depression and also the
16
danger of activation of mania and hypomania.
17
Also, additional events mentioned in the
18
precautions, any psychoactive agent may impair
19
intellectual or psychomotor functions, and
20
therefore, care should be exercised in prescribing
21
these medications when individuals have to operate
22
machinery or other things that may require
23
intellectual and motor functions.
24
Seizures is another precaution that is
25
mentioned especially in those with history of
211
1
seizure.
2
[Slide.]
3
Additional safety information in the
4
Adverse Reaction section is about agitation with
5 the
use of citalopram, and also additional
6
premarketing reports which are frequent, impaired
7
concentration, depression, suicide attempt, and
8
confusion; and infrequently reported in premarket
9
reports are aggressive reaction, psychotic
10
reaction, delusion, paranoid reaction, emotional
11
lability, and panic reaction.
12
[Slide.]
13
To give you just a summary of the drug use
14
pattern, it is the fourth most commonly used SSRI
15 in
children. Again, use had been increasing
in
16
recent years. Pediatric patients
account for
17
approximately 3.3 percent of the total U.S.
18
prescriptions of Celexa.
19
Pediatric diagnoses most often linked with
20 its
use are depressive disorders,
21
obsessive-compulsive disorder, and attention
22
deficit order.
23
[Slide.]
24
Since marketing, there were over 6,000
25
reports which included also duplicates that were
212
1
reported to FDA, 79 percent of them were domestic.
2
Less than 5 percent of the reports were in
3
pediatric patients.
4
The top 20 pediatric adverse events were,
5
looking at that, all adverse events related to in
6
utero exposure were unlabeled, which actually
7
happened to be in the top 20 for pediatric adverse
8
events.
9
Adverse event reports for children
10
involving direct exposure were generally similar to
11
those reported for adults.
12
[Slide.]
13
After a manual review of the
one-year
14
post-exclusivity period, there were 42 unduplicated
15
reports that were pediatric.
Sixteen of them were
16 in
utero exposures, and resulted in unlabeled
17
events and one death.
18
There were 26 children involving
direct
19
exposure, 8 unlabeled events, and no deaths in this
20
group.
21
Looking at the outcomes, there were 16
22
serious outcomes, 10 hospitalizations, 4
23
life-threatening, and 2 was disability.
For the
24
direct exposure group, the dose range was typically
25 5
to 60 mg/day. The median dose was about
20
213
1
mg/day in these reports.
2
[Slide.]
3
Again looking at the age distribution, in
4 the
in utero exposure, most of them female, as well
5 as
in the direct exposure group, and age
6
distribution is 0 to 1 in 15 patients, and then
7
most of the direct exposure group, in older
8
children.
9
[Slide.]
10
Looking at the reasons for exposure to
11
citalopram, there were 26 direct pediatric
12
exposures and then 16 in utero exposures. I am
13
going to just focus on the adverse events
14
pertaining to psychiatric, but these are the
15
reasons for why they were exposed.
16
[Slide.]
17
There were only 5 psychiatric events, in 5
18
patients where there were psychiatric events, and
19
these are broken down by labeled and unlabeled
20
events.
21
In the labeled events are the cognitive
22
impairment, aggression, agitation, mania, and
23
delusions, suicidality, and psychotic reaction.
24
Unlabeled events are the violent/homicidal
25
behavior, which were observed in 2 of the patients.
214
1
[Slide.]
2
Looking at these 5 patients with
3
psychiatric events, there were 4 males and 1
4
female. The age distribution as 6
to 11 years with
5 2;
11 to 16, about 3 of them. Diagnosis in
4 of
6
them was MDD, and 1 case was oppositional defiant
7
disorder, ODD.
8
Concomitant medications were reported in 2
9
patients, Prozac in 1, and another, Keppra and
10
clonazepam. Symptom resolved once citalopram
11
discontinued in 4 according to the reports.
12
[Slide.]
13
In closing, I would like to say there were
14 few
psychiatric events that were reported during
15
this one-year post-exclusivity period, unable
16
really to determine causality due to limitations of
17 the
AERS database, therefore, we will continue to
18 monitor
these adverse events in children.
19
I would like to reiterate the same
20
limitations that I mentioned before with respect to
21
paroxetine when I talked about limitations of the
22
AERS database.
23
Thank you very much for your attention.
24
DR. RUDORFER: Thank you.
25
Dr. Andrew Mosholder will now speak on the
215
1
Office of Drug Safety Data Resources for the Study
2 of
Suicidal Events.
3
Andy.
4
Office of Drug Safety Data Resources for
5 the Study of Suicidal Events
6
DR. MOSHOLDER: Thank you very
much.
7
I am very pleased to be here this
8
afternoon. I am going to talk
about how we looked
9 at
some of our Office of Drug Safety data resources
10 to
see if they would be relevant to exploration of
11
this issue.
12
[Slide.]
13
It is very much a team effort and I want
14 to
start by acknowledging my colleagues who
15
assisted me.
16
[Slide.]
17
The objective of my brief presentation
18
will be to describe the data resources we have
19
available in the Office of Drug Safety at FDA that
20 are
relevant to this issue, and, in particular,
21
looked at two types of databases, the first being
22 the
postmarketing surveillance database that Dr.
23
Iyasu just described, and also some
24
population-based epidemiological databases.
25
Also, I will be describing the context of
216
1
spontaneous postmarketing reports of these
2
types of events with newer
antidepressants.
3
[Slide.]
4
Turning first to the postmarketing
5
surveillance data from the AERS system as you have
6
just heard about.
7
[Slide.]
8
We did a special search for these events,
9 and
I will describe the methods. The list of
drugs
10 is
shown here, and it is the same drugs we have
11
been discussing throughout the day.
We limited the
12 age
on the report to patients 17 years or
younger,
13 and
we looked at U.S. reports only.
14
[Slide.]
15
In the AERS database, the events are
16
classified under particular adverse event terms
17
according to the so-called MedDRA dictionary. We
18
chose a list of event terms that we thought would
19
capture suicidal behaviors and ideation.
I will
20 let
you read for yourselves the list, but that was
21 the
list of terms that we searched in the AERS data
22
base for those events.
23
[Slide.]
24
The results showed for all those drugs
25
over their full marketing history, there was a
217
1
total of 524 case reports, of which 110 were death
2 reports. I should add that these are raw counts,
3
which means there was no hands-on review for
4
duplicate reports.
5
Occasionally, the same case will be
6
reported by more than one health professional or
7 the
health professional and the consumer, and those
8 are
referred to as duplicate reports. So,
these
9 are
just the raw counts.
10
[Slide.]
11
Here they are broken down by drug, and you
12 see
they are ranked in order. You see
fluoxetine
13 has
the most, and, roughly speaking, the numbers of
14
reports parallels the prevalence of their use in
15 the
pediatric population, so that is not too
16
surprising.
17
[Slide.]
18
What this displays is the same totals
19
broken down by year of reports.
So, we see the
20
year the report was received down here on the x
21
axis, and the number of reports.
22
A couple of things to observe here.
First
23 of
all, for most of the drugs, we see that there is
24
between, say, zero and 10 reports annually, and
25
then, of course, there are these two sort of
218
1
exceptions to that. There is a peak
over here in
2 the
early '90s, and that is for fluoxetine, and
3
then in the last two to three years, there is
4
another peak, and that is for paroxetine.
5
The one thing to point out here relevant
6 to
the fluoxetine, as I am sure everyone is aware,
7
this peak coincides with the controversy in the
8
early '90s about whether fluoxetine can induce
9
suicidality. In fact, in 1991,
there was an
10
advisory committee about that topic.
11
To understand this increase with the
12
paroxetine reports, we looked at that in a little
13
more detail, as I will show you.
14
[Slide.]
15
This shows the proportion of reports
16
according to whether they were consumer or health
17
professional, and the interesting thing here is
18
that while the health professional reports have
19
remained fairly constant over the years, what we
20 see
in the last two to three years is an increase
21 in
the proportion of reports that are coming from
22
consumers, which, of course, doesn't mean that they
23 are
not legitimate reports, but it does illustrate
24
that there is some influence on the spontaneous
25
reporting that is encouraging consumers to report
219
1
more of the these events in the last few years, and
2
that seems to account for this increase.
3
[Slide.]
4
To go into things a little more in depth,
5 we
decided to look at reports from the first three
6
years of marketing and do an in-depth review.
7
We took reports for all 10 drugs from the
8
first three years that they were marketed in the
9
U.S. This is a standard way in
10
pharmacoepidemiology of comparing reports across
11
drugs to account for the so-called Weber effect
12
that applies during the first three years of a
13
drug's marketing history.
14
Even so, during this time period, there
15 was
limited pediatric use of these drugs, and
16
because of secular trends, changes in reporting
17
systems, and other variables, it is still very
18
difficult to make quantitative comparisons between
19
drugs.
20
[Slide.]
21
So, we looked at these reports, we
22
eliminated duplicate reports, and we chose four
23
suicide-related categories - suicidal ideation,
24
suicide attempt, completed suicide, and
25
self-mutilation, and classified the reports into
220
1 one
of those categories.
2
[Slide.]
3
This shows the results. There
were 94
4
reports retrieved from the AERS system.
After
5
review for duplicates, there were 78 unduplicated
6
reports, which gives you an idea of the proportion
7 of
duplication. It is something like 15
percent.
8
This was for 9 drugs, no cases for
9
nefazodone. Out of these 78
reports, most were
10
female, most were over 12 years of age, and that is
11
consistent with what we know about the epidemiology
12 of
suicidal behavior in adolescents. Most
of the
13
events were classified as suicide attempts.
14
There were 7 completed suicides, 6 with
15
fluoxetine, 1 paroxetine, 4 males, and 3 females.
16 We
found no reports of rechallenge with the same
17
drug, which is sometimes used as an indication of
18
evaluating the causality.
19
[Slide.]
20
This slide shows the numbers of reports by
21
category here and by drug. If you
look at the
22
total, you see that, as I already mentioned, 67 out
23 of
78 were in the suicide attempt category.
24
Again, these are ranked in terms of the
25
totals. You see that fluoxetine
again has the
221
1
most. Again, this sort of roughly parallels the
2 prevalence
of their pediatric use.
3
[Slide.]
4
So, interpreting these results, we would
5 say
that suicidality was reported with all drugs.
6 The
drugs with the largest numbers of reports
7
coincided, roughly speaking, with the greatest
8
amount of pediatric use.
9
The reporting is variable and appears to
10 be
influenced by various events and also because of
11 the
quality and variability and low pediatric use,
12 the
data really do not support quantitative
13
comparison between drugs.
14
[Slide.]
15
In general, AERS data are most useful for
16
distinctive or rare adverse drug reactions, such as
17
aplastic anemia. The problem
here, as Dr. Iyasu
18 has
already described, is that the outcome of
19
interest that we are tracking, which is
20
suicidality, is also an outcome of the indication
21 for
which the drug is prescribed, so that it is
22
very difficult to sort out whether the drug played
23 a
role or whether it was the underlying disorder
24
from evaluating data of this type.
25
[Slide.]
222
1
I want to move on to look at some other
2
data resources that we have in ODS and tell you
3
about that.
4
[Slide.]
5
We looked at four principal sources that
6
could be used, one, the Tennessee Medicaid. That
7 is
a health care claims database. We have
two
8
surveillance databases. I will
let you read the
9
descriptions, but they are maintained by CDC and
10 the
Consumer Products Safety Commission.
11
This one applies to hospital emergency
12
rooms, and this one applies to emergency rooms and
13
also ambulatory care.
14
Finally, there is the Oregon Adolescent
15
Suicide Attempt Data System. In
the State of
16
Oregon, adolescent suicides and suicide attempts
17 are
reportable conditions, so that the State Center
18 for
Health Statistics maintains a database on those
19
reports.
20
[Slide.]
21
To summarize briefly, there are
22
significant limitations in attempting to use these
23 data sources to evaluate this issue. One was
24
rarity of completed suicide, difficulty in
25
identifying individuals with outcome of completed
223
1
suicide. It may not generate a
health care claim,
2 for
example.
3
There is great difficulty in classifying
4
non-fatal suicidal behavior, as we have already
5
heard about, difficulty obtaining data on drug
6
exposure prior to the event, lack of suitable
7
control groups, confounding by indication, and
8
privacy restrictions.
9
[Slide.]
10
In conclusion, for the study of this issue
11 of
pediatric suicidal behavior associated with
12
antidepressant treatment, the available
13
pharmacoepidemiological data and postmarketing
14
surveillance data is of limited utility, and
15
randomized, controlled trial data should be
16
superior to these sources.
17
Thank you very much.
18
DR. RUDORFER: Thank you.
19 Open Public Hearing
20
DR. RUDORFER: We will now turn to
the
21
afternoon portion of our open public hearing.
22
I am mandated to read the ground rules for
23
meetings of general matters, so if you will bear
24
with me for a moment, I need to address our open
25
public hearing speakers.
224
1
Both the FDA and the public believe in a
2
transparent process for information gathering and
3
decisionmaking. To ensure such
transparency at
4
this open public hearing session of the Advisory
5
Committee meeting, FDA believes that it is
6
important to understand the context of an
7
individual's presentation.
8
For this reason FDA encourages you, the
9
open public hearing speaker, at the beginning of
10
your oral statement to advise the committee of any
11
financial relationship that you may have with any
12
company or any group that is likely to be impacted
13 by
the topic of this meeting. For example,
the
14
financial information may include a company's or a
15
group's payment of your travel, lodging, or other
16
expenses in connection with your attendance at the
17
meeting.
18
Likewise, FDA encourages you at the
19
beginning of your statement to advise the committee
20 if
you do not have any such financial
21
relationships. If you choose not
to address the
22
issue of financial relationships at the beginning
23 of
your statement, it will not preclude you from
24
speaking.
25
With that, we will turn to our first
225
1
afternoon speaker, David Fassler.
2 David Fassler, M.D.
3
DR. FASSLER: Thank you. My name is David
4
Fassler. I am a child and
adolescent psychiatrist
5
practicing in Burlington, Vermont.
I am speaking
6
today on behalf of the American Psychiatric
7
Association where I serve on the board of trustees.
8
The APA represents over 35,000 psychiatric
9
physicians across the country.
The APA receives
10
funding from a variety of sources including
11
pharmaceutical companies, but no pharmaceutical
12
funding was used in conjunction with my appearance
13
today or the preparation of my comments.
14
You have already heard lots of testimony
15
today, so let me try and briefly highlight and
16
underscore a few key issues.
17
First, childhood and adolescent depression
18 is
a very real illness which will affect between 3
19 and
5 percent of all young people. The good
news
20 is
that we can help most kids who suffer from this
21
disorder. Intervention is most
effective when it
22
begins early and when it involves a comprehensive
23
treatment plan individualized to the needs of the
24
child and family.
25
Because we care deeply about children, we
226
1
encourage parents to be advocates for their kids,
2 to
ask lots of questions about any proposed course
3 of
treatment. We also encourage the FDA to
develop
4
mechanisms to enhance access to data from clinical
5
trials including negative trials, as well as
6
unpublished research.
7
We believe that such access would
8
facilitate scientific discussion and dialogue and
9
help physicians and parents make fully informed
10
decisions about treatment options.
11
Second, with specific reference to
12
suicidal ideation, it is important to emphasize
13
that such thinking is always a very real concern,
14 and
as you have heard this morning, it is also not
15
uncommon.
16
From the Youth Risk Behavior Survey, we
17
know that 1 adolescent in 5 thinks about suicide
18
each year, and that by the end of high school, at
19
least 1 in 10 has made an actual suicide attempt.
20
Third, medications can be extremely
21
helpful and even lifesaving for some children, but
22
medication alone is rarely a sufficient treatment
23 for
complex child psychiatric disorders such as
24
depression.
25
Finally, we are concerned that the
227
1
publicity surrounding this issue may frighten some
2
parents and discourage them from seeking help for
3
their children. This would be a
real tragedy since
4 the
reality is that we really can help most of
5
these kids.
6
DR. RUDORFER: Thank you, Dr.
Fassler.
7
Our next speaker is Dr. Lawrence Diller.
8 Lawrence Diller, M.D.
9
DR. DILLER: Last but not
least. I am
10
behavioral developmental pediatrician who has
11
prescribed psychiatric drugs to children for 26
12
years. I have no financial
connections to the
13
industry.
14
I am the author of Running on Ritalin and
15
Should I Medicate My Child.
16
As a front-line practitioner, I have lost
17
faith in my research academic colleagues to provide
18 me
the data information, opinion, and conclusions
19 in
an objective and unbiased fashion. I
20
desperately need that information in order to
21
validate and augment the clinical decisions I must
22
make every day on who does and doesn't get
23
medication.
24
Unfortunately, in my quarter century of
25
practice, I have seen child psychiatry's biologic
228
1
revolution hijacked by a for-profit drug industry.
2
Drug companies so pervasively influence academic
3
research, professional education, now direct
4
consumer information, ultimately determining the
5
very way society views its own problems.
6
I see top research leaders in the field of
7
child psychiatry simultaneously publishing papers
8 in
scientific peer-reviewed journals while
9
appearing in press conferences for corporations
10
that have funded the research, which is then
11
reported in the Wall Street Journal.
12
We learn of nonpublication agreements of
13
negative finding studies and limited access to raw
14
data that potentially allows for completely
15
different interpretations or conclusions based upon
16 the
published information.
17
At this time, the conflict of interest
18
between my academic colleagues and the drug
19
industry rivals that of the stock analysts and the
20
brokerage firms. Doctors are at
risk of being
21
regulated by the government, but this is unlikely
22 to
happen soon since the public and the Congress
23
have been similarly influenced or bought by these
24
powerful corporations.
25
Unfortunately, it will take children dying
229
1
followed by trial lawyer class action suits to get
2
changes either in the practice or the regulation of
3 the
SSRIs. That is a heck of a costly way,
both
4 the
individual families and the public, for what
5
should be routine formal postmarketing drug
6
surveillance funded by neutral third parties.
7
Until then, I hope there is more
8
government-funded research, but as long as I only
9
have research funded or suppressed by drug
10
companies, I will remain quite cautious and
11
hypervigilant over what I prescribe the youth of
12
America.
13
Thank you.
14
DR. RUDORFER: Thank you, Dr.
Diller.
15
At this time we are going to take just a
16
very quick break and return for further speakers
17
from the FDA. Let's say five minutes if possible.
18
Thanks.
19 [Break.]
20
DR. RUDORFER: We have three
additional
21
speakers from the FDA who will address some of the
22
important data at hand and that is still emerging.
23
First, I am pleased to introduce Dr.
24 Thomas
Laughren, who is team leader of the Division
25 of
Neuropharmacologic Drug Products, who will
230
1
discuss with us the regulatory history on
2
antidepressants and suicidality, and give us an
3
update on current plans for the analysis of
4
pediatric suicidality data.
5
Regulatory History on Antidepressants and
6
Suicidality and Update on Current Plans
7
for Analysis of Pediatric Suicidality Data
8
DR. LAUGHREN: Thank you, Matt.
9
[Slide.]
10
I am going to talk very briefly about the
11
regulatory history of antidepressants and
12
suicidality, and then spend most of my time talking
13
about our current plans for looking at the
14
pediatric suicidality data coming out of the
15
controlled trials
16
But first I would like to thank the
17
families who came forward this morning to talk
18
about their very personal stories, both the
19
families that talked about tragic outcomes and
20
those who talked about children who appear to have
21
been helped by medications.
22
It is very hard to do that, and I think it
23 helps
us to put all of this discussion in context,
24 but
a very important point, and this has been made
25
several times, it is very difficult to assess
231
1
causality based on individual cases.
That is true
2
both of those cases where the outcome is tragic,
3 but
also true of the cases where the outcome is
4
good.
5
For either of those, we have to turn to
6
controlled trials, so my focus is going to be on
7 the
controlled trials.
8
[Slide.]
9
What I have given you in this slide is
10 the
standard language which is in all
11
antidepressant labeling, and has been in
12
antidepressant labeling for decades.
This is in
13 the
Precaution section. Essentially, it
warns
14
clinicians of the possibility of a suicide attempt
15 in
major depressive disorder, and advises
16
clinicians especially early in treatment to watch
17
patients very carefully.
18
Now, this statement does not explicitly
19
warn of the possible linkage between antidepressant
20 use
and the emergence of suicidality, but I think
21 it
allows for that interpretation and, in fact,
22
this idea that antidepressants may be associated
23
with the emergence of suicidality early in
24
treatment has been around for a very long time in
25
psychiatry.
232
1
[Slide.]
2
This is a statement from a textbook of
3
psychiatry published in 1960.
This was the time at
4
which the tricyclic antidepressants had just come
5 on
the scene. Let me read it.
6
It says, "With beginning convalescence,
7 the
risk of suicide once more becomes serious as
8
retardation fades."
9
[Slide.]
10
What this statement is referring to is
11
what is commonly known as the roll back phenomenon.
12
This is the observation again of emergent
13
suicidality early in treatment and the belief, the
14
belief that that is in some way linked to the use
15 of
the drug, and the view, the mechanism proposed
16 is
that antidepressants give patients increased
17
energy, particularly those with psychomotor
18
retardation, that allows them to act on their
19
suicidal ideas before the drug has had a chance to
20
affect mood.
21
So, this is one proposed mechanism for
22
this observation. In fact, it is
only one of
23
several proposed mechanisms. When
we met with the
24
advisory committee in 1991, to talk at that time
25
about Prozac and the possibility of suicidal
233
1
induction, Dr. Martin Teicher from Harvard
2
University reviewed a number of proposed mechanisms
3 to
explain this observation including the roll back
4
phenomenon.
5
But he also talked about the possibility
6 of
actually a paradoxical worsening of depression,
7 in
other words, the mood actually becoming worse
8
rather than better.
9
He talked about the possible role of
10
akathisia, which is associated with many of these
11
drugs, about the induction of anxiety and panic
12
attacks by some of these drugs, about the idea that
13
patients with bipolar depression may experience a
14
stage shift, in other words, moving from depression
15 to
a mixed state, and finally, even the induction
16 of
insomnia.
17
All of these ideas, the idea is that once
18
these behaviors are induced, there is then a link
19
from that behavior to suicidality, and all of these
20
proposed mechanisms have some plausibility, but it
21 is
quite a different matter between proposing a
22
mechanism and empirically establishing that there
23 is,
in fact, a link between the use of an
24
antidepressant and the emergence of suicidality.
25 [Slide.]
234
1
That is really the question that we are
2
dealing with here today and that is the question we
3
hope to be able to address with these clinical
4
trials data for these pediatric studies:
Is there
5 a
causal link between antidepressant drug use and
6
suicidality in pediatric patients with major
7
depressive disorder or with other psychiatric
8
disorders?
9
We agree that this is a
critically
10
important question to answer, but we also feel that
11 it
is important to answer it in a careful and
12
thoughtful manner because to err in either
13
direction has significant consequences.
14 Clearly, we do not want to miss a
signal
15 of
increased risk of suicidality, because that
16
would give us greater comfort in the use of these
17
drugs than would be warranted.
18
On the other hand, we don't want to reach
19 a
premature decision on the strength of the signal
20
because that could result either in the overly
21
conservative use of these medications or in their
22
lack of availability all together for treating
23
pediatric depression. So, it is
important to get
24 it
right.
25
[Slide.]
235
1
In this slide, what I have done is to list
2 the
9 drugs that are involved in our ongoing
3 review.
You have seen this list before today.
4
This involves a total of 25 studies in pediatrics,
5 16
of them in major depression, the others in
6
various other pediatric disorders, involving a
7
total of over 4,000 patients.
8 [Slide.]
9
Right now let me talk a little bit about
10 how
the signal came onto our radar screen.
We had
11
reviewed over the past three to four to five years
12
pediatric supplements for 8 drugs, and we looked at
13 the
safety and efficacy data for these drugs.
14
In the course of putting together a report
15 for
FDA, companies code their adverse event data,
16 and
they do this in their own ways. We don't
tell
17
them how to code the data, they choose their own
18
dictionaries and they set about coding the data
19
before they send it in.
20
This applied to any events suggestive of
21
suicidality, as well as any other adverse events.
22 We
reviewed those supplements over this period of
23
three to four years, and suicidality did not emerge
24 as
a matter of concern based on those reviews.
25
However, the Paxil review did raise a
236
1
question about data management in that events
2
suggestive of suicidality were coded under the
3
general preferred term "emotional lability."
4
This struck the reviewer as rather odd,
5 and
so in responding to GSK, we asked them to
6
separate out the verbatim terms suggestive of
7
suicidality under a term specific to suicidality.
8
[Slide.]
9
That request to GlaxoSmithKline resulted
10 in
additional work and ultimately resulted in a
11
report on paroxetine and pediatric suicidality.
12
That report went first to the MHRA -- that is FDA's
13
counterpart in the UK -- and shortly thereafter to
14 FDA
in May of last year.
15
That report indeed suggested an increased
16
risk of suicidality associated with paroxetine use
17 in
particular in one of the three studies done in
18
pediatric depression.
19
[Slide.]
20
What I am going to do in the next two
21 slides
is to quickly walk you through a timeline of
22 key
events that occurred over the past eight months
23 to
try and give you a sense of how we got from the
24
time of that initial report up to the present time.
25
So, that report was issued in May.
In
237
1
June, both FDA and MHRA issued regulatory
2
responses. As you heard earlier,
the MHRA
3
essentially contraindicated paroxetine in pediatric
4 depression.
FDA came out with fairly strong
5
language that recommended against its use in
6
pediatric depression, but stopped short of a
7
contraindication, and, in essence, we said that we
8
were continuing to look at the data.
9
In July, we issued a request to sponsors
10 of
the eight other antidepressant products asking
11
them to look at the suicidality data in their
12
databases using an approach similar to that, that
13 had
been used by GSK, and I will talk about that
14
approach a little bit later.
15
So, in essence, we wanted to look at
16
summary data from the other programs, similar to
17
what had been given to us for Paxil.
In August of
18
last year, we went back and relooked at the
19
suicidality data in the pediatric supplements.
20
In August, Wyeth, the manufacturer of
21
Effexor, having responded to our July request and
22
having looked at their data, decided that they did
23 have
a signal and they made a labeling change which
24
they are allowed to do under changes being effected
25
without our prior approval, so they changed their
238
1
labeling, adding information about that perceived
2
signal, and they also sent a Dear Doctor letter
3
which essentially recommended against the use of
4
Effexor in pediatrics.
5
Also, at that time, MHRA contraindicated
6
Effexor in pediatric depression.
7
In September of last year, we held an
8
internal regulatory briefing at FDA.
We hold these
9
briefings basically to update upper management on
10 key
issues that are before us, and this certainly
11 was
a key issue, and we have the briefing.
12
There were a number of recommendations
13
that came out of that briefing.
Two were of
14
critical importance to our ongoing review. One of
15
those was the suggestion that we think about
16
reclassifying the cases, because there was some
17
uncertainty about what this diverse array of events
18
coded under this broad term "possibly
19
suicide-related" actually meant.
So, there was a
20
suggestion that we do that.
21
There was also a suggestion that we think
22
about doing a more refined data analysis, allowing
23 the
use of adjustment for covariates.
24
[Slide.]
25
In September and October, we began to get
239
1
responses to our July requests for summary data for
2
other antidepressants, and it gave us some cause
3 for
concern, because we were seeing that sponsors
4 had
not used exactly the same approaches that we
5 had
suggested in our July request.
6
In October, we issued an updated Public
7
Health Advisory, at this time essentially
8
broadening the concern to all antidepressants. In
9
essence, we advised clinicians to use caution when
10
using these drugs in pediatric depression,
11
essentially, to pay attention to the language that
12 is
already in labeling.
13
In October, having thought more about a
14
patient level data analysis allowing us to look at
15
covariates, we issued a response to all
16
antidepressant manufacturers asking them to give us
17
patient level data sets to allow us to do this
18
analysis.
19
Also, in October, having thought more
20
about the reclassification effort, we decided,
21
instead of trying to do this inside FDA, we decided
22 to
go outside FDA and get an outside expert group
23 to
help us with this reclassification.
24
In November and December, having thought
25
more about this problem of case finding that I had
240
1
alluded to earlier in response to our July request,
2 we
issued a second and actually then a third
3
response to companies to give us cases to look at.
4
Finally, in December, as was pointed out
5
several times earlier today, MHRA, having completed
6 its
review of all the pediatric data, decided to go
7 ahead and contraindicate all the other new
8
generation antidepressants except for fluoxetine.
9
So, as I understand it, fluoxetine is the
10
only current generation antidepressant available
11 for
treating pediatric depression in the UK.
12
[Slide.]
13
I have used the terms summary data and
14
patient level data several times, and I want to
15
make sure that you understand what it is I am
16
talking about.
17
By "summary data," I am referring to data
18
tables that are provided to us by sponsors based on
19
their own analyses, that include only numbers of
20
patients with events as the numerators and either
21
total patients exposed or total accumulated
22 person-time as the denominators.
23
These are the data that we got from Glaxo
24
back in May and that we have since gotten from all
25 the
other sponsors. These are summary data.
241
1
"Patient level data" are data sets that
2 are
provided by sponsors in response to a detailed
3
request from FDA for electronic data sets that are
4
structured to include one row per patient
5
participating in each study, so that we have data
6 for
all patients participating in those trials, and
7 we
have multiple variable data for each patient.
8
These data sets allow us to do adjustments
9 for
covariates that may be important for any
10
particular event of interest, while summary data of
11
course do not.
12
In the next slide, I am going to summarize
13 for
you the suicidality risk data from the seven
14
programs for the antidepressants that were studied
15 in
pediatric depression. Before I do that, I want
16 to
clarify what the two event categories are that
17 we
are dealing with.
18
[Slide.]
19
The first event category is an umbrella
20
term, "possibly suicide related." This is the term
21
that Glaxo developed in looking at its own
22
database, and it is the term that we asked other
23
sponsors to look at in going through their data
24
sets.
25
Basically, it was intended to capture any
242
1
event in their databases that included any thoughts
2 or
behaviors that the sponsor considered to
3
represent possible suicidality, so it is a very
4 broad term.
5
The term "suicide attempt," as defined for
6
these analyses, was the subset of that umbrella
7
term, so a subset of these originally captured
8
events that met the conditions of having any
9
indication of self-harm. So, this
is how "suicide
10
attempt" was defined in this analysis.
11
So, the overall umbrella term "possibly
12
suicide related" and then the subset of those
13
events that had some indication of self-harm.
14
[Slide.]
15
This is, I am sorry, a very busy slide.
16
These are the risk data coming out of these seven
17
programs, and I am going to walk you through this.
18
Again, there were seven programs -
19
paroxetine, fluoxetine, sertraline, venlafaxine,
20
citalopram, nefazodone, and mirtazapine.
I have
21
divided these up into different colored rows so you
22 can
see the number of studies in each program, two
23 of
them involving three studies, the rest all
24
two-study programs.
25
This is risk data. So, this is
simply the
243
1
number of patients having one or more of these
2
events divided by the total number of patients
3
exposed. There is no adjustment
for time here.
4
This is crude risk. In
parentheses, I have got the
5
percent.
6
The way this is set up, first of all, the
7
overall umbrella category "possibly suicide
8
related," and then the subset of these events that
9 met
the criterion for "suicide attempt."
Again,
10
that criterion was any indication of self-harm.
11
Let's just walk through the individual
12
programs. Again, paroxetine had three trials. For
13 the
first trial, 329, you see a risk ratio
of
14
roughly 6, 6.5 percent for drug, 1.1 percent for
15
placebo, so definitely a signal of something.
16
However, if you look at the other two
17 studies
in this program, 377 and 701, these were
18
also fairly large studies, in fact, this one was
19
slightly larger, the risk ratio was around 1. So,
20 the
signal for paroxetine is essentially coming out
21 of
one study, a big signal, but the other studies
22
show essentially nothing.
23
If you look at fluoxetine, there really
24
isn't any signal coming out of the fluoxetine
25
program, the risk ratios are all in the vicinity of
244
1 1.
2
For sertraline, again, you have one study
3
which is suggestive of a signal, 4.1 percent versus
4
zero, drug versus placebo, but for the other study,
5
similarly sized, in fact, these were identically
6
designed studies, there is no signal. It's 2.2
7
percent for both.
8
If you look at venlafaxine, there appears
9 to
be a signal coming out of both studies in that
10
program. For citalopram, again,
you have two
11
studies, both large studies. One study, no signal,
12 in
fact, if anything, it is slightly in favor of
13
drug. The other study, a weak
signal, but many
14
more events, many more events in this study, and a
15
risk ratio of rough 1.6.
16
The number of events in the nefazodone and
17
mirtazapine programs is so small that it is hard to
18
know what to make of that.
19
There are two points that I want you to
20
take away from this slide. First
of all, I think
21 in
looking at these data, there is enough of a
22
suggestion of a signal of something that clearly it
23 is
worth pursuing this.
24
Everyone at FDA concluded that there is
25
obviously something going on here, we need to
245
1
pursue this, but one troubling thing about this set
2 of
data is the inconsistency in the signal across
3
studies within the programs.
4 In most of these programs where
there is a
5
signal except for venlafaxine, it appears to be
6
coming from one study. So, that
is something that
7 we
felt that we need to try and explore in some
8
way.
9
[Slide.]
10
In the remaining time what I am going to
11 do
is talk about the concerns we have had in
12
interpreting these suicidality data.
13
I should have mentioned at the outset, I
14 am
sorry, I made a number of changes in my slides
15
over the weekend, so I apologize.
I have had to
16
delete some of the material. I
didn't talk about
17
efficacy, and I am not planning on talking about
18
efficacy here, you know, in the discussion section
19 I
am happy to do that.
20
I thought it would be useful if I focused
21
instead on the clinical cases because one of the
22
concerns we have had is what these reported events
23
that are captured under this broad term "possibly
24 suicide
related" actually represent.
25
So, I put together a number of slides over
246
1 the
weekend to try and give you a better sense of
2
that, and that is why the slide package you have is
3
different than what I am presenting.
4
In any case, there are three concerns that
5 we
have looked at. One has to do with case
6
finding, and that is the first bullet, and I
7
alluded to that earlier. In
looking at the summary
8
data that sponsors gave us, it appeared that
9
somewhat different approaches were used to
10
capturing and presenting these cases to us. So, I
11
will talk about how we explore that.
12
Secondly, there is the issue I
talked
13
about of the question of how you classify these
14
cases into meaningful categories for the purposes
15 of
analysis and regulatory decisionmaking.
16
Finally, I have already alluded to the
17
issue of the inconsistency in the signal across
18
individual studies within the programs, and that
19 was
one of the findings that led us to want to do a
20
more refined analysis looking at covariates. It is
21 one
of several reasons, but that is one
22
justification for that analysis.
23
[Slide.]
24
Let me first focus on the issue of case
25
finding. This is a very busy slide, I apologize for
247
1
that, but I will walk you through it.
This is the
2
algorithm that was used initially by Glaxo and that
3 we
then asked the other companies to apply to their
4
databases in finding cases.
5
In essence, there were two components to
6
this. There was an electronic string search, which
7 I
will talk about, and what they were to do is to
8
apply this string search, and they were to blindly
9
look at the events that were turned up with that
10
search and decide whether or not those events were
11 of
interest from the standpoint of suicidality and
12
then give us those data.
13
So, the string search was one part of the
14
search. The other part was to do a blinded review
15 of
narratives for any deaths or other serious
16
adverse events in their databases.
Now, there were
17 no
deaths in any of these trials, so this part of
18 the
search focused on narratives for serious
19
adverse events.
20
So, let go back to the string search.
21
There were two components to the string search.
22
First of all, we asked companies to look at their
23
preferred terms. These are the
dictionary terms
24
that companies use in coding data.
We asked them
25 to
look at the text string "suic" and "overdos" to
248
1
pick up any instances of events that were coded
2
under either suicidality or overdose, or any
3
variation of that.
4
Now, the bullet underneath here suggests
5
that we ask for a separate listing for events coded
6 as
accidental overdose. Accidental overdose
is
7
usually, just to give you an example, where a
8
patient misses a dose on one day and then on the
9
next day thinks he should take two doses. So, that
10
would not be a suicide attempt, that is what is
11
usually considered an accidental overdose.
12
So, we didn't want those to be included
13
among the events, but we wanted to be able to see
14
them to see which ones were excluded.
15
The second part of this was to do a string
16
search for the actual verbatim investigator terms.
17
Here we used -- again, this is the
18
approach that was used by Glaxo, and we passed this
19 on
to the other sponsors -- a variety of terms
20
suggestive of either self-harm or of overdose or
21
suicidality.
22
So, this was to go through the
23
investigator terms and try and capture any events
24
that were suggestive either of suicidality overdose
25 or
some type of self-harm.
249
1
Again, we allowed exclusions from that
2
list for what I am calling false positives. A
3
false positive, for example, would be when the text
4
string inadvertently picks up a term that has
5
nothing at all to do with suicidality, so, for
6
example, the test string g-a-s, for gas, would pick
7 up
gastrointestinal, so we allowed companies to
8
exclude those events from their lists.
9
Once they came up with a list of events
10
that they considered representative of suicidality,
11 we
asked them to go through and blindly select out
12
from that overall group of possibly suicide-related
13
events, the events that were suggestive of suicide
14
attempt.
15
Again, the definition of that was any
16
indication of self-harm. So,
again, the overall
17
umbrella term and then the subset of suicide
18
attempts.
19
We asked them then to provide us a
20
narrative of all of those cases that had been
21
turned up. So, that was the
algorithm for finding
22
events.
23
[Slide.]
24
Now, we had hoped in doing that, that we
25
would get a fairly complete accounting of the
250
1
original list of events that had been turned up and
2 the
exclusions. Unfortunately, we weren't
explicit
3
about that, and it is not what we got.
4
Often, we got only the narratives for the
5
events that the companies had already decided
6
represented the suicidality set, and did not
7
include the exclusions. Often,
there was little
8
explanation for why certain events had been
9
excluded or what the criteria had been in excluding
10
events. So, that was one problem.
11
[Slide.]
12
Another problem was that we had failed to
13 ask
for narratives on accidental injuries. I
had
14
mentioned earlier that we had asked for a listing
15 of
accidental overdose, but not accidental
16
injuries. In talking to sponsors
about this, and
17
asking them to give us some of the accidental
18
injuries, we turned up a couple of events that
19
caused us some concern.
20
This was one particular example.
This was
21 a
child who had been excluded, this event had been
22
excluded from the list. It was a
patient who
23
stabbed himself in the neck with a pencil while
24
taking a test.
25
Now, this probably was an accident, but it
251
1
occurred to us that we wanted to see all of these.
2 We
wanted to see all of the events that had been
3
excluded as accidental injury, so that our experts
4 --
because at this point, we had already decided to
5 go
outside and have an outside group look at these
6
cases, we wanted to have a complete set of events
7 for
them to look at, so we asked for all the
8
accidental injuries.
9
[Slide.]
10
Another thing that we discovered when we
11
started talking to companies about the application
12 of
the search algorithm is that one company in
13
particular acknowledged that it had not done the
14 searching blindly of the narratives for
serious
15
adverse events, and this was a problem, because
16
again this had to be done blindly to be done
17
properly.
18
Another issue that turned up when we
19
started looking at these cases is that some
20
companies had excluded events that were not
21
"treatment emergent."
22
Now, when looking at adverse event data,
23 it
is entirely appropriate to be interested in
24
events that either occur for the first time on
25
assigned treatment, or if present at baseline, are
252
1
worse on treatment than at baseline.
That is what
2 we
mean by "treatment emergent."
3
So, it is not that it was improper to do
4
that. The problem was that we
wanted to see which
5
events were excluded for that reason, so that we
6
could assess ourselves whether or not it was an
7
appropriate exclusion. So, again,
in going back,
8 we
have now asked for all the events excluded as
9
treatment emergent.
10
Finally, in looking and comparing the
11
strength of the signal coming out of the pediatric
12
supplement re-review and the signal coming out of
13 the
summary data, in one particular case we noted a
14
fairly substantial discrepancy between the strength
15 of
the signal. That again raised a question
about
16
case finding.
17
[Slide.]
18
So, the bottom line is that having looked
19 at
these initial summary reports from companies, we
20 did
not have complete confidence in the case
21
finding, so we issued, as I mentioned, a second
22
request for clarification both of how the search
23 had
been done and then a complete accounting of how
24 the
companies winnowed down to the list of events
25
that they considered to represent the suicidality
253
1
set, so that we could see what events had been
2
excluded, for what reason, and so that we could be
3
confident that we had a complete set of data to
4
start with.
5
In addition, we asked for narratives for
6 all
serious adverse events rather than just the
7
ones that the companies decided represented
8
suicidality, so again our outside experts could go
9
through all of these data and independently and
10
blindly themselves decide which were representative
11 of
suicidality.
12
So, that is the case finding issue.
13
[Slide.]
14
Next, I want to talk about the issue of
15
reclassification. There were two
issues that again
16
caused us concern about the approach to classifying
17
these cases.
18
One was in looking at the events that got
19
captured, we noticed that there was an extremely
20
wide variability in the types of events that got
21
included under either the broad umbrella category
22 or
also under the narrower term "suicide attempt."
23
We also notice that companies appeared to
24
have used very different approaches to capturing
25 the
subset of events labeled "suicide attempt."
254
1
Some companies used a fairly conservative
2
approach, others essentially labeled all of the
3
events as suicide attempts even though there was
4
nothing in the case report to suggest self-harm.
5
[Slide.]
6
So, what I have done, and these are the
7
slides that I put together this weekend, I have
8
gone back to look at the 109 patients having one or
9
more possibly suicide-related events.
These were
10 the
patients who were included in the numerators
11 for
the table that I showed you earlier.
12
So, these are the cases, and the
13
collection of 109 patients goes across all studies,
14 not
just the depression studies.
15
A couple of points to make. First
of all,
16 the
point about there were no completed suicides
17
among these 109 cases. As I
mentioned, there was
18
very wide variability in the types of verbal
19
expressions and behaviors that were considered by
20
companies to be representative of suicidality.
21
Another problem with these cases is that
22 the
majority of them were not well described.
We
23 did
not have the level of detail in these cases
24
that one would have liked to do a rational
25
classification.
255
1
My goal in doing this is to provide you
2
with a sense of the range of events to consider.
3 You
know, this is not a formal classification.
4
Again, we have contracted with an outside group to
5 do
the classification, but I wanted you to have a
6
sense of the kind of variability in the case
7
material that we have, so you can appreciate why we
8
consider this a problem.
9
[Slide.]
10
There are two key questions.
First of
11
all, is it meaningful to subsume such diverse
12
events under this umbrella term "possibly suicide
13
related," and is it reasonable to define "suicide
14
attempt" as that subset of events that have any
15
degree of self-harm, is that a reasonable
16
definition of "suicide attempt."
17
I want to be very clear about this.
I am
18 not
attempting to trivialize in any way any of the
19
events that occurred. I mean
these are sick kids,
20 all
of these events have importance.
21
The question is what classification
22
approach is most useful and clinically meaningful
23 in
preparation for doing an analysis and in
24
preparation for taking regulatory action. That is
25
really my goal here.
256
1
[Slide.]
2 Let me describe how I approached
these
3
cases. For a small fraction of
them, patients had
4
more than one suicidality event, so for
5
consistency, I focused on the first one.
That only
6
applied to about 10 percent of these patients.
7
Then, I went ahead and I selected a subset
8 of
those events where there was any indication at
9 all
of self-harm. Again, this is to mimic
the
10
approach that the sponsors were supposed to use in
11
defining suicide attempt.
12
For those patients who had an indication
13 of
self-harm, I looked at whether or not they were
14
hospitalized for the event and whether or not there
15 was
any indication of suicide intent. By
that, I
16
mean either an active expression of intent in that
17
case narrative or I accepted any concurrent
18
indication of suicidal ideation.
19
For the remaining patients who had
20
suicidal ideation without self-harm, again, I
21
looked at whether or not they had been hospitalized
22 for
the event and whether or not there was a
23
suicidal plan, so there had to be an active
24
expression in the narrative of a suicidal plan in
25
association with that suicidal ideation.
257
1
[Slide.]
2
Overall, the hospitalization rate for
3
these 109 patients was 43 percent.
The subgroup
4
having suicidal ideation without any indication of
5
self-harm was 39 percent and the remainder -- these
6
were the patients who had some indication of
7
self-harm -- was 61 percent.
8
Again, there were no complete suicides,
9 all
patients were fully recovered from these
10
instances of self-harm. As sort
of an interesting
11
aside, in about 30 percent of these cases, the
12
self-harm event appeared to occur in the context of
13
some kind of interpersonal conflict.
14
A typical situation would be a child had
15 an
argument with a parent or a sibling or a peer or
16 a
girlfriend or boyfriend, impulsively engaged in
17
some kind of self-harm behavior, and the event was
18
over, and there was no indication of suicidal
19
ideation. That applied in about
30 percent of
20
these cases.
21
[Slide.]
22
In going through the self-harm case events
23 in
more detail, again, there were a total of 66 of
24
these. Nineteen of these involved
cutting
25
behavior. In almost all of these
cases of cutting,
258
1 it
appeared to be a superficial wound. There was
2 one
case where a young girl cut herself so deeply
3
that there was actually blood loss.
In another
4
case there was an indication that the patient
5
needed three stitches to suture the wound, but in
6 all
the other cases, they appeared to be
7
superficial.
8
There were 37 overdoses. Again,
there was
9 a
wide range of different types of behaviors that
10
were classified as overdose, ranging at the one
11
end, one patient was classified as an overdose for
12
taking 20 percent more medication than was
13
prescribed.
14
Ordinarily, this would not be considered a
15
suicide attempt, and there was no indication in
16
that case of suicidal ideation, but that was
17
classified as an overdose.
18
At the other extreme, there were patients
19 who
took fairly substantial quantities of either
20
study medication or usually
over-the-counter
21
medication, so a very wide range in terms of
22
amounts of drug that was taken.
23
There were two cases characterized as
24
hanging behavior. In both of
those cases, what
25
they really were, were interrupted attempts. These
259
1
were children who, in the presence of family or
2
parents, engaged in what was described as hanging
3
behavior, it was immediately interrupted, and so in
4
neither case was there any actual self-harm. So,
5
these were interrupted cases.
6
The case of burning was similar.
This
7
occurred in the context of family, and the child
8 was
immediately interrupted although in that case
9
there was some minor burns.
10
One case that was classified as a suicide
11
attempt was the case of a young girl who slapped
12
herself in the face, and that was it.
That was all
13
there was in that case, and there was no suicidal
14
ideation described in that case.
15
Then, there were six other cases where all
16
that the case indicated was minor self-mutilation.
17 It
was not specified what the self-harm behavior
18
was.
19
[Slide.]
20
Now, let me give you a breakdown of what I
21
found when I looked at, first of all, the cases of
22
cutting.
23
There were 19 of these. In most
of these
24
cases, in 16 out of the 19, there was no indication
25 of
either suicide intent or even any concurrent
260
1
suicidal ideation, and 4 of those 19 cases actually
2
ended up being hospitalized.
3
So, most of those cases did not involve
4
hospitalization and did not involve suicide intent
5 or
ideation.
6
[Slide.]
7
For the 37 cases of overdose, there were
8
more hospitalizations here, but again, if you
9
notice in this column, in almost every case there
10 was
no indication of suicide intent or suicidal
11
ideation.
12
A number of the
hospitalizations could be
13
characterized as an overnight hospitalization for
14
observation.
15
[Slide.]
16
Finally, for the remaining 43 patients who
17 had
suicidal ideation without self-harm, again, I
18
looked at whether or not there was a plan, an
19
expressed plan, and in most of these cases there
20 was
not a plan.
21
In the 7 where there was a plan, they were
22
hospitalized, but nevertheless, a majority of these
23
patients with suicidal ideation without self-harm
24
were hospitalized.
25
[Slide.]
261
1
So, I hope that gives you a little bit
2
better sense of the range of behaviors that we are
3
dealing with here and the difficulty we had in
4
including all of them under this one umbrella term
5 of
"possibly suicide related."
6
As I said, we have gone to an outside
7 group. What I want to do in this slide is talk a
8
little bit about the Columbia University
9
Suicidality Research Group and why we picked them.
10
I talked to a number of people about who
11
should help us with this, and most everyone I
12
talked to said that this group has the expertise to
13 do
this. They do have expertise, they have
been
14
doing this for almost 20 years.
15
In the last 5 years alone, they have more
16
than 40 funded grants to do this kind of research.
17
They are in the business of developing measures and
18
manuals and methodologies for evaluation of
19
suicidality.
20
They are a center for training on suicide
21
assessment, and research on both reliability and
22
validity. They are currently involved in the NIMH
23
study looking at adolescent suicide attempters.
24
This is the TASA study. They are
doing the suicide
25
assessment or the suicide classification for that
262
1
trial. As you can see, they have
a very large
2
number of publications over the 20 years they have
3
been doing this.
4
So, we think this is a good group to help
5 us
with this problem.
6
[Slide.]
7
I have two more slides left. What
I want
8 to
do in this slide is again remind you of what Dr.
9
Katz said earlier, is that we view this meeting
10
today as a preliminary meeting.
We are hoping to,
11 you
know, once we have had these cases
12
reclassified, and have done the analysis, to come
13
back to you with more definitive answers later in
14 the
summer.
15
You are going to hear next from Dr. Kelly
16
Posner from Columbia. She is
going to tell you
17
about the way they think about classifying suicidal
18
events and how they plan to approach these data.
19
Following that, you will be hearing from
20
Tarek Hammad from our Safety Group.
He is going to
21
tell you about our preliminary plans for an
22
appropriate patient level data analysis.
23
[Slide.]
24
Finally, these are the five topics for
25
which the Neuropharm Division would like to have
263
1
feedback from you. First of all,
three topics
2
pertinent to the analysis of suicidality data.
3
Again, I alluded to our concerns about the
4
approach to case finding and how we attempted to
5
resolve that, but we would be interested in knowing
6
what you think about that and whether you think
7
anything more needs to be done in terms of case
8
finding.
9
Secondly, you will be hearing
from Dr.
10
Posner about approaches to classifying these events
11
into appropriate categories before we do the
12
analysis. Since we are actively
engaged now in
13
discussing this with them, this would be a good
14
time to give us feedback on that.
15
Thirdly, if you have thoughts about our
16
plans for the patient level data analysis, we would
17 be
interested in hearing about that.
18
In terms of future concerns, again, one of
19 the
striking things about these cases is how poorly
20
they were described, and this may also indicate a
21
less than optimal approach to ascertainment in
22
these studies.
23
So, if you have thoughts, we are beginning
24 to
talk with Kelly Posner and others about
25
developing a guidance document for ascertaining
264
1
suicidality in future studies, if you have thoughts
2
about that, we would welcome them.
3
Finally, I didn't get a chance to talk
4
about efficacy, but obviously, the largely negative
5
results from the short-term trials in pediatrics is
6
clearly a concern.
7
We would be interested in knowing what
8
your thoughts are about that and whether or not you
9
think there are other possible designs that might
10
help us get at whether or not there are benefits
11
with these drugs.
12
One design that has been used in adult
13
studies is the randomized withdrawal design. This
14 is
a design where you take patients who have
15
responded to medication acutely, have been stable
16 for
some period of time, and are then randomized to
17 either
continue on drug or assignment to placebo,
18 and
you look at time to relapse as the event, as
19
another approach to trying to establish whether or
20 not
there are benefits.
21
I am going to stop there.
22
Thank you.
23
DR. RUDORFER: Thank you, Dr.
Laughren.
24
As Dr. Laughren said, we will now hear
25
from Dr. Kelly Posner of Columbia, who will
265
1
describe in more detail the suicidality
2
classification project.
3 Suicidality Classification
Project
4
DR. POSNER: Thank you.
5
[Slide.]
6
So, why is a methodologically sound,
7
research-supported classification warranted? Let's
8
back up a second and talk about the problem.
9
[Slide.]
10
The problem, as the cases that Dr.
11
Laughren discussed exemplified, there is a clear
12
lack of conceptual clarity about what suicidal
13
behavior means and a corresponding lack of
14
agreement on common terminology both in clinical
15
descriptions of suicidal acts, as well as research
16
descriptions of suicidal acts.
17
Given this lack of generally accepted
18
terms for referring to even the most basic suicidal
19
behaviors, the importance of using definitions that
20 are
both reliable, meaning we all define them and
21
assess them the same way, and valid, meaning there
22 is
some truth to them seems quite clear.
23
[Slide.]
24
So, what are these standardized
25
research-supported definitions? I
think it is
266
1 important to note that there really is
generally
2
agreement among suicide assessment experts on the
3
basics of these terms. So, we are
going to start
4
with the suicide intent.
5
A self-injurious act committed with at
6
least some intent to die. Intent
doesn't have to
7 be
100 percent. If there is any intent to
die, we
8
consider it an actual suicide attempt.
9
Intent does not have to be explicit and
10 can
be inferred. For example, if a patient
denies
11
intent to die, but thought that the behavior could
12 be
lethal, intent can be inferred.
13
A real case example includes a 12-year-old
14 who
is angry at her mother. She took 6 to 7
15
prescription pills, said she was aware that taking
16
that much medication might kill her, but she didn't
17
know if she intended to die by taking the pills.
18
That would clearly be categorized as a suicide
19
attempt.
20
Once again, it is important to note that
21
once there is any possibility of injury, the act is
22
defined as an attempt, meaning that if someone
23
pulled the trigger of a loaded gun, but
24
fortuitously missed, it is still a suicide attempt.
25
[Slide.]
267
1
Other classifications: suicidal behavior
2
without injury. Interrupted
attempts are defined
3 as
the individual is stopped by an outside
4
circumstance from starting the self-injurious act.
5
Examples of these: someone has
pills in their
6
hand, but they are stopped from ingesting. Once
7
even one pill is ingested, the event becomes an
8
actual attempt.
9
They have a gun pointed toward themselves,
10 the
gun is taken away by someone else or somehow
11
they are prevented from pulling the trigger. They
12 are
poised to jump, they are grabbed, taken down
13
from the ledge. All examples of
interrupted
14
attempts.
15
The next classification is what is called
16 an
aborted attempt in which an individual takes
17
steps toward making a suicide attempt, but stops
18
himself before engaging in any potentially
19
self-destructive behavior.
20
Remember, holding a loaded gun but not
21
pulling the trigger is a good example.
This could
22 not
possibly result in injury, therefore, it
23
constitutes an aborted attempt.
It is still
24
suicidal behavior, but it is not an actual attempt.
25
[Slide.]
268
1
I think it is worth focusing a moment on
2
suicidal intent, because again, intent here is the
3
determining factor when you are classifying
4
suicidality. It is the presence
of intent to die
5
that differentiates suicidal acts from self-injury.
6
One must determine whether the
7
self-injurious act was thought of as a means of
8
causing or facilitating death. Of course, we do
9
have research support for the validity of using
10
intent to define suicidality.
11
One example is that completed suicide is
12
predicted by previous intent measures, which was
13
demonstrated by Beck and his group in 1989.
14
[Slide.]
15
Some more case examples. These
are real
16
cases again. These are examples
of non-suicidal
17
self-injury.
18
A teenage girl reported her mother was
19
being cruel and neglectful and she wanted to escape
20
from her mother's home. She
states that she
21
researched lethal doses of ibuprofen to make
22
certain that she took an amount that would not be
23 life-threatening. She took 6, feeling sure it was
24 not
enough to kill her. She definitely did not
25
want to die, only to escape from her mother's
269
1
house. She was taken to the ER
and then admitted
2 to
a psychiatric hospital.
3
Another is the more common case of
4
self-mutilation where the person described 12
5
incidents of cutting himself, stated he did this
6
only "to relieve tension" and "to play with danger
7 to
see how far I would go" and no part of him
8
wanted to die. Thought about it
for hours before
9
acting on the urge and felt relieved of tension
10
afterwards, did not feel pain.
11
[Slide.]
12
So, what is our research support of these
13
classifications? We will start
with reliability.
14 We
have been able to demonstrate excellent
15
reliability utilizing these definitions and this
16
classification system in NIMH-funded treatment,
17
biological, and genetic trials across the life
18
span.
19
We have also been able to demonstrate
20
multi-site reliability with other expert centers in
21
family genetic studies and treatment trials, and
22
again particularly the treatment of adolescent
23
suicide attempters trials.
24
In short, across domains, across the life
25
span, and across institutions, we have been able to
270
1
demonstrate excellent reliability.
2
[Slide.]
3
Validity. How much truth is there
to
4
these definitions? Individuals
classified as
5
suicide attempters have as much as 2.5 times risk
6 of
future attempts or completions. So, we
know
7
this is a real category.
8
Similarly, interrupted attempters are
9
reported to be 3 times more likely to commit
10
suicide than uninterrupted attempters.
11
Finally aborted attempters are at risk for
12
eventual attempts and were more likely to have made
13 an
actual attempt in the past.
14
Again, all validating the classifications
15
that we are using.
16
[Slide.]
17 So, what is the classification
methodology
18
that we are proposing here?
19
To begin with, the data will be blinded by
20
experts not on the panel. It will
be blinded not
21
only to pharmaceutical information, but also to any
22
relevant clinical information that would bias an
23
event rating. For example, a
family history of
24
suicidality. An event
classification should stand
25 on
its own, and we want to make sure that it is
271
1
blinded in both domains.
2
Next, we have to determine the event
3
classifications based on these reliable and valid
4
constructs.
5
We are then going to do a training on the
6
classification system to establish reliability of
7
panel members who are all experts in the field.
8
Once the reliability study is done, the
9
expert panel will be divided into three subgroups,
10 and
the data will also be divided into three groups
11 in
order to do classifications.
12
There will be additional cases, and the
13
reason for the additional cases is to demonstrate
14
that the classifications are all being done in the
15
same way and to prevent what we call stratification
16
bias.
17
You want to exhibit a relationship between
18 the
groups and make sure it is not some other
19
factor that is causing a group to rate things in a
20
similar way, and then we will generate the
21
classified cases.
22
[Slide.]
23
What are the classifications that we are
24
proposing? Suicidal,
non-suicidal, and
25
indeterminate. Subclassifications
of suicidal
272
1
would include suicide attempt, suicidal behavior
2
without injury, which would include aborted and
3
interrupted attempts, suicidal ideation related
4
events.
5
Non-suicidal subclassifications
would
6
include self-injury or mutilation again with no
7
intent associated, and other categories, accidental
8
injuries or other psychiatric symptoms that we have
9
been hearing a lot about today, disinhibition,
10
akathisia, agitation.
11
Then, finally, the indeterminate category
12
either by non-consensus or inability to classify
13 due
to a paucity of data.
14
So, if, in fact, there is a signal, the
15
point is we just don't know yet what it is a signal
16 of,
and that is why a logical research-supported
17
approach is warranted. We want to
be able to look
18 at
the data consistently and logically across
19
trials in order to make some clinically meaningful
20
sense of it.
21
[Slide.]
22
I think it is also worth mentioning for a
23
moment future directions. We want
to develop
24
guidelines as to how to better capture data,
25
enabling appropriate classification and description
273
1 of
suicidality.
2
We will demonstrate, based on this
3
conceptual clarity, how to utilize research
4
assessment tools, what questions to ask, how to ask
5
them, and what measures aid in this, which will
6
then lead to consistency of terminology and
7
classification, as well as to improved, more valid
8
identification and documentation of suicidality.
9 In addition, as was mentioned earlier,
10
that will also enable more active appropriate
11
surveillance of suicidality, which is a great need
12
clearly.
13
Thank you.
14
DR. RUDORFER: Thank you, Dr.
Posner.
15
Our final formal speaker of the afternoon
16
will be Dr. Tarek Hammad from the Division of
17
Neuropharmacologic Drug Products, who will discuss
18
plans for analysis of patient level data for
19
pediatric studies.
20 Plans for Analysis of Patient
Level
21 Pediatric Studies
22
DR. HAMMAD: Good afternoon,
everyone.
23
I am here today to talk about our analysis
24
plan for the pediatric patients data.
25
[Slide.]
274
1
These are some of the elements that I will
2
cover in my talk. After a brief
description or a
3
statement of the objective of this work, I will
4
describe the data that we have and then I will go
5 on
to discussing the analysis plan.
6
[Slide.]
7
The objective of this work is to evaluate
8 the
risk of suicidality associated with the use of
9
antidepressants in pediatric patients using the
10
results of the blinded reclassification of cases.
11
I think you have heard enough about the
12
value of this reclassification.
13
In the process, we will address the
14
possible sources of imbalance in the data, for
15
example, trial design, duration of exposure, et
16
cetera, and also other potential confounders.
17
These efforts will help us understand the sources
18 of
inconsistency between trials or between drugs,
19 if
any.
20
[Slide.]
21
The source of all data is controlled
22
trials conducted in pediatric patients in nine drug
23
development programs. These are
the drugs that you
24
have seen before, that is the list of drugs and the
25
number of trials involving each drug.
275
1
For the analysis or at least for some
2
stages of the analysis, they will be grouped into
3 two
categories, an SSRI group and an Atypical
4
Antidepressant group.
5
[Slide.]
6
These trials were not done in one
7
indication, and for purpose of analysis again, they
8
will be categorized or divided into three different
9
subgroups - MDD, anxiety disorders, and attention
10
deficit hyperactivity disorder assuming, of course,
11 we
have enough cases within every category of
12
indication.
13
[Slide.]
14 As far as individual patients data that
we
15 are
requested, we developed a standard format to
16
guarantee the compatibility between data coming
17
from various sources. We actually
specified every
18
aspect of the desired database down to the variable
19
name and some description to clarify the contents,
20 and
some coding notes as appropriate.
21
In addition, we requested descriptive
22
information about every trial to evaluate the
23
similarity of these trials, which as you can
24
imagine is very important to determine if these
25
trials can be pooled or not to gain more power
276
1
while you are investigating this question.
2
[Slide.]
3
This is a list of the requested variables
4
that can be categorized in many subcategories -
5
demographics variables, disease-related variables,
6
drug-related variables.
7
[Slide.]
8
Outcome-related variables, psychiatric
9
history variables, and some treatment emergent
10
adverse events. As you can see,
this is not just
11
about having a second look at the data.
It is
12
about trying to understand and appreciate and
13
characterize the signal, if there is any.
14
[Slide.]
15
This is a list of some challenges we have
16
with the data, I wanted to mention here because of
17 the
important implications of these challenges on
18 the
proposed analysis and on the actual
19
interpretation.
20
They can be divided roughly into two
21
categories, a quality-related component and an
22
analysis-related component.
23
The first issue in the quality-related
24
component, which is pertinent to what Dr. Laughren
25 was
talking about, the case ascertainment, so I
277
1
will not belabor the issue more, but a similar
2
issue is pertinent to the other pieces of
3
information being collected, which is other
4
variables that we requested.
5
The mechanism of capturing these data
6
might be different from trial to trial or from
7 sponsor to sponsor, so we will investigate
this,
8 and
that is part of the challenge, trying to see if
9
these data can actually even be comparable or not.
10 But
for now, the rule that we will use is that we
11
will not use data with missing information more
12
than 10 percent. The second issue
is somewhat
13
detailed and I will address in the next few slides
14
The first point under the analysis-related
15
component is using the trial or the patient as the
16
unit of analysis. Pooling data
from different
17
trials, treating them as one large trial fails to
18
preserve the randomization effect and might
19
introduce bias and confounding.
20
That is because maintaining the
21 randomization
guards against the foreseen and
22
unforeseen imbalances between different treatment
23
groups in various trials.
24
The issue of trial similarity is not only
25
pertinent to having the same protocol, but it is
278
1
also pertinent to the implementation of those
2
protocols in reality. That is why
I believe the
3
trial-based approach is more appropriate.
4
However, we might be using some
5
information using the trial as the unit of the
6
analysis, because if we have zero events in one of
7 the
arms, for example, we have to impute some data,
8 but
if we have zero events in both arms, we will
9 not
be able to drive the information in this trial.
10
So, it depends on the eventual count of
11 the
actual cases that we would have. If we
are
12
losing too many trials, we might use the patient as
13 the
unit of the analysis, of course, after doing
14 the
appropriate adjustments.
15
[Slide.]
16
The second point is pertinent to the
17
limitations of pooling data in general whether we
18 use
the trial or the patient as the unit of the
19
analysis, because these trials have different
20
designs, patient populations, sometimes duration of
21
treatment, et cetera, and pooling them together
22
with the appropriate adjustment gives you an
23
average effect that is really dependent on the
24
proportions of different subpopulations in these
25
data.
279
1
This effect will be different subsequently
2 if
these proportions are different, so careful
3
evaluation of this has to be conducted and then
4
adjusting for it.
5
There is also the inherent class effect
6
assumption that is implied by pooling data across
7
drugs or within groups of indications even. Mind
8
you, we do this to try to gain more power, try to
9 see
some gathering of data instead of just looking
10 at
it trial by trial, but by doing this, if we pool
11
data from drugs within certain class assumption
12
here, the risk of suicidality is equal in all
13
drugs.
14
The problem comes in when we realize that
15 we
do have different size of data for different
16
drugs, and the smaller opportunity to observe an
17
event in one drug might lead to none being observed
18 or
very few.
19
The question becomes whether this is
20
because this drug is generally different from the
21
rest of the class or because we simply don't have
22
enough power. Unfortunately, this
will always be
23 an
open question, but I would report the results
24
both ways by individual drugs and by group data.
25
[Slide.]
280
1
The analysis plan would follow a standard
2
approach with initial exploratory phase, where we
3
will check for the compliance with our request and
4 the
completeness of data, check for coding errors,
5 and
the like, and then we will list all the risks
6 and
rates by drug, by indication, and by trial just
7 to
see what is going on in data, in all aspects of
8 the
subgroups before we pool anything, so we know
9
where the signal is coming from if there is any
10
afterward.
11
Then, we investigate the data separation,
12
which is an important component.
For example, if
13 all
cases were among men, for example, then, this
14
variable we will not be able to evaluate, and so
15
on. That is just part of the
process of
16
evaluation.
17
Then, we go to investigate interactions
18 and
potential confounders to try to understand what
19 is
going on and try to characterize the risk, as I
20
said before.
21
[Slide.]
22
This is just a sample of one of the tables
23
that will be produced, the rates and percentages
24 and
the risks of suicidality by drug and by
25
indication for every trial.
281
1
[Slide.]
2
To evaluate the estimate, to actually try
3 to
relate an overall effect, an estimate for an
4
overall effect, two approaches that I discussed are
5
options that we have. First,
which I believe is
6 the
more of an optimal approach, is using the trial
7 as
the unit of analysis.
8
In this analysis, I will adjust the
9
confounders on a trial level. We
are basically
10
looking for a randomization failure in if all of
11 these randomized trials, but in case there
might be
12
some failure in randomization, any small imbalances
13 can
actually be reflected on the apparent risk.
14
Then, I will have done everything by trial
15 and
by drug. In this particular analysis, I
will
16
pool trials for drug groups that I should do
17
initially within indication groups.
Trials will be
18
excluded if there are no cases reported in both
19
arms.
20
Now, depending on the heterogeneity of
21 the
trials' findings, the variability between
22
trials will be considered in a fixed effect order
23 in
random effects model.
24
The premise behind the fixed effects model
25 is
that the real effects we are trying to evaluate
282
1 is
fixed, and the observer variation between trials
2 is
just by chance. The premise behind the
random
3
effects model is that there is an average of these
4
effects that is the full distribution with a
5
variation affected by the observer trials.
6
Many times you will have both approaches
7
yielding the same results, but I am going to do it
8
both ways with some of the conditions for which
9
approach is more appropriate given the actual data
10 or
the heterogeneity of the data.
11
Now, if we opted to use the patient as the
12
unit of analysis in the situation I mentioned
13
before, which is a situation where we will not have
14
that many cases, and we would be losing trials
15
right and left, so we will try to pool and get some
16
slightly more power, pooling patients as the unit
17 of
the analysis.
18
We will use the Poisson regression to
19
model the rates of suicidality, adjusting for
20
potential confounders, and then again will pool
21
patient data for drug groups within indication
22
groups, and, of course, will adjust for trial in
23 the
model because these patients are coming from
24
different trials.
25
[Slide.]
283
1
As you know, these trials were not
2
designed to capture these particular events, so
3
there is some inherent uncertainty about the
4
finding. It depends on what kind
of feedback we
5 get
from our experts in the Columbia University.
6
Some sort of sensitivity analysis might be
7
warranted, stratifying by the amount of uncertainty
8 in
this particular finding.
9
[Slide.]
10
There are some limitations on the
11
interpretation of data that we should know upfront.
12
Just to put the limitations in context, I have here
13 the
first bullet to remind you about the goal of
14
this particular effort, which is to evaluate the
15
risk of suicidality associated with the use of
16
antidepressants in pediatric patients.
17
Now, after everything is said and done,
18 the
observed rates will not reflect the actual
19
patients in the general population.
Why? Because
20
there are some exclusions in some trials of
21
patients with some baseline suicidality, so the
22
observed rates will not reflect what is going on in
23
real life, and this might hamper our efforts in
24
trying to investigate the risk because it will lead
25 to
underestimation in all the arms, so we might not
284
1
have enough power to be able to detect the actual
2
thing.
3
Now that we only have short-term exposure
4
data, we will not be able to extrapolate this to
5
what happens after long-term exposure to these
6
drugs.
7
Now, we don't really have any information.
8 The
next bullet is that we don't have any
9
information on the patterns for discontinuation.
10
Considerably, there might be some informative
11
censoring going on with patients with suicidality
12
tendencies, might be likely to be discontinued. If
13
that happened more in the placebo group, then,
14
there might be some apparent underestimation of the
15
signal in the placebo, and this might lead to some
16
spurious finding, but we don't have information on
17
this which would be very hard to overcome.
18
My last point is that it remains to be
19
seen if we have enough statistical power to detect
20
differences in the risk of suicidality among
21
various drugs because of the issue that I alluded
22 to
before, which is there is no data for some of
23 the
drugs.
24
[Slide.]
25
In closing, there are our ideas and some
285
1 of
them were informed by our experience analyzing
2 the
data on the completed suicides in adults.
3
So, your feedback on our approach will be
4 greatly appreciated.
5
DR. RUDORFER: Thank you, Dr.
Hammad.
6
At this point, we are going to open up for
7
discussion by the committee. If
anyone has
8
questions for our speakers, now is the time to
9
raise them.
10
Dr. Laughren.
11
DR. LAUGHREN: Matt, I had in my
original
12
talk planned on giving a brief summary of the
13
efficacy data, and it sounds like a number of
14
people are disappointed that I didn't do that. I
15
have those data and I could, if you wanted me to
16
take five minutes and do that, I would be happy to
17 do
that.
18
DR. RUDORFER: Yes, please do so.
19
[Slide.
20
DR. LAUGHREN: What this slide does
is
21
summarize very briefly the outcome on the 15 trials
22
that we looked at for the 7 programs in pediatric
23
major depression.
24
Again, there are 3 studies in the
25
paroxetine program and 2 studies in each of the
286
1
other programs, and what this slide does is to
2
simply summarize in very crude form what the
3
outcome was on the primary endpoint.
The protocol
4
specified primary endpoint for those trials, and
5
this gives the age range in these studies here.
6
So, for example, for the paroxetine
7
program, there were 3 studies, all negative. For
8
sertraline, 1 trended in trend, for the purposes of
9
this slide, indicates a p value on that primary
10
endpoint of between 0.05 and 0.1.
A negative trial
11 is
indicated by a p value of greater than 0.01.
12
So, for paroxetine, all 3 studies were
13
negative, fluoxetine, both were positive and, as
14 you
know, this was the one program for which we
15
concluded that there was sufficient data to support
16 a
claim.
17
Our standard, and I believe the standard
18 of
most other regulatory agencies for pediatric
19
major depression, is 2 positive studies.
20
For the sertraline program, 1 trended and
21
then 1 negative. Venlafaxine,
both were negative.
22 For
citalopram, 1 positive and 1 negative.
23
Nefazodone, 1 trend, 1 negative, and both negative
24 for
mirtazapine.
25
Now, the one point I want to make in this
287
1
slide is that this was our fairly conservative view
2 of
these data. Others have looked at these
same
3
data and have reached different conclusions.
4
For example, for the paroxetine study 329,
5
this was the basis for a publication by Keller, et
6
al. They acknowledged that that
trial was negative
7 on
the primary endpoint, however, they pointed out
8
that it was positive on virtually all secondary
9
endpoints, and on that basis, they and many others
10
consider that to be a positive study.
11
Similarly, for the sertraline program,
12
although if you look at the individual trials,
13
neither one makes it. One of the
secondary
14
analyses in the plan for these identically designed
15
studies was to pool them, and when that is done,
16 the
pooled analysis is very positive, so some view
17
that -- and again this was the basis for a
18
publication by Wagner, et al. -- some view the
19
sertraline program as providing support for
20
efficacy in major depression.
21
Again, as I pointed out, the
citalopram
22
program had 1 of 2 studies that was clearly
23
positive.
24
[Slide.]
25
Now, I want to talk a little bit about
288
1
this largely negative outcome. If
you look at
2
adult major depression studies, and if you look at
3
drugs which we believe work and which have been
4
approved for depression in adults, about half the
5
time studies that on face look like they should
6
make it, fail.
7
These are studies that are done in what
8
appears to be the right population.
The sampling
9
size is appropriate, the doses appear to be
10
appropriate, assessments are appropriate, but for
11
whatever reason, about half the time, these studies
12
fail.
13
Now, if you assume that that failure rate
14 can
be applied to pediatric major depression
15
studies, and you look at the possible outcomes for
16 2
trials, for programs that involve 2 trials, you
17 can
very quickly reach the mathematical result that
18
only about 25 percent of the time would you expect
19 to
get 2 positive studies.
20
Most of the time you would expect either 1
21 or
both trials to fail if the failure rate were the
22
same as is true in adults. So, in
retrospect, it
23
perhaps was not as surprising as it turned out to
24 be
here that you get a lot of negative results.
25 On the other hand, the overall success
289
1
rate here of 3 out of 15 studies making it at 0.05
2 on
the primary endpoint is clearly, clearly a
3
concern.
4
[Slide.]
5
There are a couple of other things to keep
6 in
mind. If you look at the history of
short-term
7
trials with tricyclic antidepressants in pediatric
8
depression, it is uniformly negative, and there are
9 several
possible interpretations of that.
10
One is that the drugs don't have any
11
benefit. Another possibility is
that the extent of
12
heterogeneity in pediatric patients who are
13
captured under these major depressive disorder
14
criteria may capture patients who are even more
15
heterogeneous than we believed to be the case in
16
adults, and the greater the heterogeneity in that
17
sample, the more likely you would end up with
18
negative studies. So, that is one
possibility.
19
Another thing to keep in mind is that the
20
regulatory context for doing these studies was
21
somewhat unusual. In every other
case, when a
22
company does a study, the only gain they are going
23 to
get out of that study is if it turns out
24
positive.
25
In this case, these studies were done
290
1
primarily for pediatric exclusivity.
As was
2
pointed out earlier, there was no requirement that
3
they get positive studies to get exclusivity.
4
Either way, if they did the trial according to the
5
terms of the written requests, they would get
6
exclusivity.
7
I am not suggesting in any way that
8
companies set out to do inadequate studies, but
9
having that somewhat unusual mind-set could operate
10
against a trial in subtle ways, in terms of, for
11
example, recruitment of patients.
So, it is just
12 another
thing to keep in mind in terms of
13
interpreting these largely negative data.
14
Finally, at the time that the written
15
requests for these studies were issued, we were not
16
routinely asking for Phase II dose finding studies,
17 as
we are now in all of our written requests.
18
Again, to the extent that appropriate dose
19
finding was not done, that would work against
20
positive studies.
21
So, just in summary on the efficacy side,
22 I
think there are several plausible explanations
23 for
failure to find efficacy in these trials other
24
than the obvious possibility that maybe the drugs
25
have no benefits in pediatric major depression.
291
1
In any case, the failure to meet FDA's
2
fairly high standard of having 2 positive trials,
3 in
most of these programs, we do not consider proof
4 of
the lack of benefit. So, it is true they
didn't
5
meet the standard, but that is not quite the same
6
thing as saying that it has now been proven that
7 the
drugs have no benefit. That is a very
8
different conclusion.
9
On the other hand, the failure to show a
10
benefit in major depression in most of these
11
trials, obviously heightens the concern about any
12
adverse events, in particular, in this case, the
13
possibility of the induction of suicidality.
14
Clearly, the burden is on those who believe that
15
these drugs do have benefits to show it, to design
16 and
conduct studies that show this.
17
Again, one of the questions that I have
18 for
the committee is what your thoughts are about
19 how
to go about this, in particular, the
20
possibility of using a very different kind of study
21
design, for example, using the randomized
22
withdrawal design, which has been fairly successful
23 in
showing longer term benefits in adult studies.
24
I will stop there. Thank you.
25 Open Committee Discussion
292
1
DR. RUDORFER: I think we will now
open
2
this up for questions and discussion by the
3
committee.
4
DR. SANTANA: Can you clarify
something
5 for
me, so under the exclusivity rule, if the
6
results are positive, those studies can then be
7
used by the sponsors to make a supplemental claim,
8 and
that could then become part of a new indication
9 in
pediatrics, is that correct?
10
DR. LAUGHREN: I am sorry?
11
DR. MURPHY: Yes.
12
DR. LAUGHREN: The answer is yes.
13
DR. MURPHY: Yes.
14
DR. SANTANA: I was trying to
answer this
15
issue of whether there was some bias in these
16
studies because they were requested under the
17
exclusivity rule. I have never
interpreted it that
18
way.
19 DR. MURPHY: I think what has been a
20
concern from the very beginning with exclusivity,
21 we
think the intent of Congress was that they want
22
more information. If studies are
going to be
23
conducted, they want that information to be known,
24 and
therefore, they want to say to companies we
25
want you to go out and get this information. It
293
1
doesn't mean you have to reach the bar of having an
2
approval, because a negative study can be just as
3
important as a positive study.
4
But the other concern here is that you --
5 and
no one is saying this, so, please, I don't want
6
this quoted out of context -- but there is the
7
concern that it is easy to design a sloppy study,
8
fail, and still get your exclusivity. That is
9
always a concern, and it is our job to try to not
10
allow that to happen.
11
DR. RUDORFER: Dr. Temple.
12
DR. TEMPLE: I just was going to
emphasize
13 the
same thing. Tom isn't suggesting that
anybody
14 was
totally indifferent to the outcome, but the
15
sense of urgency that comes when you have launched
16 a
very expensive program to develop a drug, you
17
really must win or it's all toast, and that's not
18
true here. You can win anyway,
different
19
incentives.
20
DR. RUDORFER: Dr. Fink, you had a
21
question.
22
DR. FINK: This is sort of an
overriding
23
question, not to a specific speaker.
In looking at
24 the
questions that are being asked of the
25
committee, we have heard very little about the data
294
1 set
that is being used.
2
Are the inclusion and exclusion criteria
3 for
these various studies appropriate in terms of
4
drug history, history of substance abuse, family
5
history of psychiatric diagnoses?
Because these
6
were placebo-controlled trials, they probably
7
enrolled less severe disease as evidenced by the
8
lack of completed suicides, and finally, as has
9
been mentioned, there was no need of efficacy.
10
I am concerned that no amount of analyses
11 of
a possibly flawed or suboptimal data set will
12
answer the question. If there is
shown to be a
13
relationship to suicidality, we may take away drugs
14
that are useful in pediatric depression with
15 different
trial designs.
16
If the studies come out negative, we may
17 be
falsely reassured. So, I am not sure
that these
18
re-analyses are going to answer the question that
19 has
been brought forward to the committee by
20 particularly the audience and that maybe we
need to
21
start with designing what are the optimal pediatric
22
trials to answer this important issue.
23
DR. RUDORFER: Does someone from
the FDA
24
want to respond?
25
DR. TEMPLE: Well, Tom sort of opened that
295
1
question to a degree. One of the
things that no
2 one
will let you do probably is treat somebody very
3
severely ill in a placebo-controlled trial, they
4
would be uncomfortable, although since it is not
5
clear what works, maybe they shouldn't be that
6
uncomfortable.
7
Nonetheless, an alternative design which
8 in
pediatric studies has been proven very
9
attractive is to take people who appear in one way
10 or
another to be doing well on a particular
11
therapy, and in this case it really won't be as
12
critical how severe they were before, and do a
13
randomized withdrawal study in which people are
14
very, very closely observed for the first
15
recurrence of any symptom that is worrisome.
16
The Pediatric Committee has discussed this
17 at
considerable length, and there is more comfort
18 in
pediatric trials in using that design where you
19 do
need a placebo to interpret the trial.
So, that
20 is
one of the questions Tom raised, and I am sure
21 we
would be interested in some discussion on that.
22
DR. GOODMAN: I am also sharing
the
23
concern about the ability to get the answer to the
24
suicidal risk associated with these drugs based
25
upon the existing data set. I
think the signal is
296
1 not
going to be strong enough although we are
2
clearly most interested in suicide or suicide
3
attempts as the outcome.
4
I wondered if one could look at these data
5
sets for other possible evidence of behavioral
6
toxicity that might be antecedents of suicidality.
7 I
think there was some allusion to that earlier,
8 but
there wasn't much detail on it. I wonder
9
specifically if one could look at some of the
10
items, like of the HAM-D or the CDRS, looking for
11
agitation or irritability.
12
If those are being induced by the
13
medications particularly early in the treatment
14
trial, perhaps those are creating a behavioral
15
state that places that individual at risk for
16
suicidal behavior.
17
One could, of course, validate that by
18
first looking at those subjects in whom there was
19
evidence of suicidality to see if it was correlated
20 or
associated with other symptoms, but if it is,
21
then go on to look at those variables, which would
22
allow you to maybe get a more sensitive measure of
23 the
effect of the drugs.
24
DR. RUDORFER: Dr. Nelson and then
Dr.
25
Katz.
297
1
DR. NELSON: Two questions. The first is
2
about the data set. At the end of
the day, when
3 you
receive the data that you are asking for, will
4 you
be looking at the same data set that were
5
reviewed by the MHRA? I mean are
we going to be
6
drawing conclusions on similar data sets?
7
The second question goes to the issue of
8 the
interpretation of Appendix 1 and Appendix 2A.
9 I
am struck, if you remove fluoxetine, that you
10
have got 1 out of 13 trials for effectiveness
11
positive and 5 out of 13 for increased risk of
12
suicidality positive, and does assay sensitivity
13
apply to risks as well and why would we not
14
interpret that as a pretty strong signal if, in
15
fact, we accept that on the efficacy side?
16
DR. LAUGHREN: Regarding the
question
17
about the UK data, I can't be certain that they
18
have the same data, however, if we look at the
19
numbers that are presented on the UK web site, they
20 are
very familiar numbers. They appear to be
21
coming from the same summary data that we had
22
access to in looking at this data.
23
So, I am reasonably confident that we are
24
dealing with the identical data sets.
The only
25
difference is that we have gone beyond accepting
298
1 the
data at face value. It appears that the
UK
2 simply
accepted the summary data analyses done by
3 the
various companies, and on the basis of a
4
suggestion of a signal, and the admitted lack of
5
efficacy for most of these programs, have decided
6 to
contraindicate these drugs.
7
We have chosen on the safety side to look
8
more closely at what that signal is, and that is
9
really the question. The question
is -- and this
10
gets in reference to your second question about
11
Appendices 2 and 2A -- I agree with you that if you
12
look across these trials, even though the signal is
13 not
consistent from study to study within programs,
14 on
balance, it appears like there is an excess of
15
something for drug relative to placebo.
16 The question is what is that. You have
17
this very broad term, "possibly suicide related,"
18 but
when you dig deeper and look at what those
19
events are, they range all the way from something
20
that everyone would agree does not represent
21
anything close to a suicide attempt to very serious
22
suicide attempts.
23
So, that is why we think it is important
24 to
go back and reclassify those events, so we can
25
figure out, first of all, if there is a signal, and
299
1
secondly, a signal for what. But
I believe that
2 the
UK had the very same data that we have, and it
3
doesn't appear to me as if they did any analysis of
4
those data other than to just accept what the
5
companies have done already.
6
DR. RUDORFER: Dr. Katz, did you
have a
7
comment?
8
DR. KATZ: Just to say that the
suggestion
9
about looking at other behavioral symptoms that
10
might be premonitory to suicidal behavior, we are
11
very interested here whether or not there are
12
specific events we should be looking at that we
13
haven't looked at yet along the lines of how we
14
intend to look at the suicidal behavior data.
15
That might involve going back and asking
16
sponsors to resubmit data sets, but we are very
17
interested to hear that. Of
course, the question
18 of
the link between those symptoms and suicidal
19
behavior is also still an outstanding question,
20
it's not straightforward.
21
DR. RUDORFER: Dr. Chesney.
22
DR. CHESNEY: I also felt that
perhaps
23
just looking at suicide attempts, basically what
24 you
just said and what Dr. Goodman said, may not be
25 all
the answer. I am most impressed from
what we
300
1
heard in the public hearing this morning about the
2
stimulant syndrome and the number of individuals
3 who
had demonstrated psychoses, akathisia, mania,
4
agitation, and so on.
5
I was also impressed at one young lady who
6
said that she wouldn't disclose that she had had
7
suicidal ideation, and then particularly impressed
8
with the three people we heard from whose children
9 at
autopsy had very elevated levels of the drug,
10
which leads to my second question.
11
That is, what do we know about
12 pharmacokinetic data in children and in
individuals
13 who
develop this stimulant syndrome. I
suspect
14
someday that we will have pharmacogenomics to tell
15 us
maybe who to predict might have that, but do we
16
have any information about pharmacokinetics in
17
children, number one, and number two, in these
18
individuals who develop these stimulant syndromes,
19 is
there any relationship at all?
20
DR. KATZ: Well, in the written
requests,
21 as
a general matter, we ask sponsors to obtain
22
pharmacokinetic information in the relevant
23
pediatric population, so I believe we have probably
24
asked for that information.
25
I don't believe we know or have had
301
1
submitted to us in any event data linking plasma
2
levels in any individual patient and particular
3
adverse events. That might be
available somewhere,
4 but
I don't think we have it.
5 DR. LAUGHREN: Could I ask Daniel Pine to
6
comment on this construct stimulation syndrome and
7
whether or not that has been reasonably well
8
defined in some way that is agreed to by different
9
individuals?
10
DR. PINE: Sure, and then actually I have
11 a
couple other comments. I don't know if
you want
12 me
to wait until after this issue.
13
But I would say across a range of
14
pediatric mental syndromes, it has been fairly
15 frequently described that a strong minority of
16
children will get activated with SSRI medications,
17 and
not just children with major depression, and
18
that in most studies, if it is not statistically
19
greater than it is in placebo, that it is a fairly
20
consistent observation across most studies, that it
21 is
higher on SSRIs than it is on placebo.
22
DR. LAUGHREN: Are we talking
about
23
something other than the anxiety and agitation
24
which is well known as a drug-related risk with all
25 of
these drugs? That is something that we
see in
302
1
most of these trials, but it is not quite the same
2
thing as saying that someone has a stimulant
3
syndrome.
4
DR. PINE: I think if you look
across the
5
trials and you look at the range of terms that
6
people have used to describe this so-called
7
stimulant syndrome, you see that the problem that
8 you
were talking about with the relatively narrow
9 set
of behaviors, self-harm behaviors or suicidal
10
behaviors, becomes even worse because across
11
different trials or trials for the same medication
12
done by different individuals, really a broad range
13 of
behaviors have been kind of linked together.
14
It remains unclear the degree to which
15
different investigators are talking about the same
16
phenomenon or different medications are producing
17
similar phenomenon. The one thing
that is clear is
18
that there is an array of what some people have
19
called, and Dr. Goodman referred to, as behavioral
20
toxicities that are not that infrequently observed
21
with SSRIs, and it might extend beyond suicidal
22
ideation, and it also needs to be better
23
categorized.
24
I would also add that when one looks at
25
those events in most of the efficacy trials, they
303
1
tend to be uniformly mild. I am
most familiar with
2 the
anxiety trials where, while they are more
3
prevalent, they tend to not cause sometimes even
4
discontinuation of the medication.
5
DR. LAUGHREN: If I could just
follow up
6 on
this. If we were, Daniel, to look for
this, I
7
guess the question would be how would one define it
8 in
a way that we could hope to find examples of it?
9
DR. PINE: I think if you look at
most of
10 the
publications for most of the SSRI trials, you
11 can
see relatively broad categories that describe
12
something that people would call activation, so,
13 you
know, in the original sertraline trial, I think
14 it
was called hyperactivity. In the
fluvoxamine
15
trial, it was called activation.
16
In the recent sertraline trial, I think it
17 was
called impulsivity. So, there is a whole
range
18 of
terms that I think you would have to canvass the
19
field in terms of thinking about what are the most
20
appropriate terms to include, much the way that you
21
have done with suicidal ideation.
22
That is not to say that this is
23
necessarily related to suicidal ideation, though.
24
DR. RUDORFER: Dr. Goodman, did
you have a
25
follow-up comment?
304
1
DR. GOODMAN: Yes. In at least adults, I
2
think as clinicians as well as clinical
3
researchers, we have a good sense of kind of the
4
array you were talking about of activation-like
5
problems that can occur with the administration of
6
SSRIs.
7
Even in the labeling, we have seen that
8
there are warnings that you can induce bipolarity,
9
mania. In fact, it has often been
said that an
10
antidepressant is probably not effective unless it
11 can
induce bipolarity in some patients.
12
We also know about psychosis and anxiety
13
induction particularly in panic disorder patients.
14 So,
I think in adults, we have it a little better
15
characterized. What I am
concerned about with the
16
children is, one, their characterization probably
17 is
somewhat overlapping, and also because of what
18 is
special about children is maybe that they are
19
more likely to manifest these problems in a less
20
differentiated fashion, which includes suicidal
21
behavior.
22
DR. PERRIN: To follow up on two
of these
23
issues, one is we hear about serious adverse events
24
being beyond suicide, but potentially also murder
25 and
other such events.
305
1
I think maybe Dr. Pine is talking about
2
strategies that might help to elicit those sorts of
3
ideas in the data set, but it seems to me that is a
4
critical issue to go beyond suicidality as a
5
potential serious adverse event.
6
I want to get back to the pediatric rule
7
question for a moment, too, if I could, and just
8 ask
-- I show my naivete and ignorance here -- but
9
what purview, what surveillance does FDA have of
10
pharmaceutical companies carrying out their trials?
11
The little I know about them is that these
12 are
often multi-site trials, fairly complicated
13
data collection among a variety of providers.
14
Do you have any surveillance as to how
15
well this is carried out, or do you sort of rely on
16 the
pharmaceutical companies to say we did it
17
reasonably well, and might that be a source for
18
variation between pediatric rule trials compared to
19 the
sort of getting the drug on the market trials?
20
DR. TEMPLE: Others may want to
comment.
21
Usual rules apply. We can inspect
any of the
22
studies. As you can imagine,
inspecting a study
23
after the fact gives you only limited insight into
24 how
well those things went on.
25
The companies are expected to provide
306
1
oversight, the rules require that they do so, but
2 our
ability to know whether it is perfect or not is
3
difficult at best. What I can't
tell you is how
4
often we have inspected these sites.
We do
5
sometimes, I don't know if we have on these.
6
DR. MURPHY: I would like to say
that this
7 is
an issue at the level of the Office of
8
Commissioner, that there is an Office of Good
9
Clinical Practices, that they are addressing to
10
make sure that we do have adequate surveillance and
11
criteria. I don't think we can
provide you a lot
12 of
information right now, but it is an issue that
13
they are looking at.
14
DR. TEMPLE: It would be
relatively
15
unusual for us unless we had a concern about
16
whether they picked up adverse reactions, which we
17
might in this case, first, to inspect a study that
18 the
company agrees is negative. On the
whole, that
19 is
not where you go to look.
20
DR. RUDORFER: Dr. Gorman, you
have been
21
waiting.
22
DR. GORMAN: One of the themes
that struck
23 me
through the morning was the interruption or
24
potential interruption of information flow through
25 the
system.
307
1
I would like to continue on the thread of
2 the
study and then go to the information flow
3
question I have.
4
On the study, I think going back and
5
looking at that 109 out of 3,000 patients that were
6
being studied for efficacy and looking for them for
7
suicidal ideation or attempts will be trying to
8
make a silk purse out of a sow's ear.
I think that
9
would be an adventure in futility.
10
I don't think it's an unreasonable thing
11 to
do if it's the data that you have available, but
12 I
think the example used this morning was the
13
needle in the haystack.
14
I think we have stepped on the needle and
15 we
have either got to see if it's really there or
16 if
it is really not there, and design studies
17
prospectively either using randomized, controlled
18
clinical trial crossover designs or withdrawal of
19
effective therapy designs.
20
One of those three designs could be
21
designed looking specifically at the questionnaires
22
that our psychiatric and psychology colleagues tell
23 us
look for suicidal ideation.
24
To look for suicides as a rare event I
25
think is going to be again a futile search, but
308
1
looking for suicidal ideation induced by these
2
medications, I think should be relatively
3
straightforward, and I would not use the set of
4
mildly or majorly depressed people.
5
I would look at groups of individuals on
6
these medications who are not depressed, so that we
7
could separate that issue out, is it the disease or
8 is
it the medicine.
9
So, take the ODDs and take the
10
post-distress syndromes and study them for suicidal
11
ideation being initiated on these medications.
12
Now, back to the pediatric rule, which I
13
think I understand, and the Best Pharmaceuticals
14 for
Children Act, I thought there was a provision
15 in
there, I thought, that when we fixed it after
16
1997, that if you went for pediatric exclusivity,
17
when you finish the trial, whether the results were
18
positive or negative, you got exclusivity.
19
But I also think there was a requirement
20 for
the pharmaceutical companies to make those data
21
available in a public place. Is
my understanding
22
confused?
23
DR. MURPHY: No, that was a slide
this
24 morning.
The BPCA does say that within 180 days of
25 the
submission of the application, that the study
309
1
results, clinical trial and a summary of the
2
clinical aspects and the pharmacology report will
3 be
posted on the web by FDA.
4
DR. GORMAN: Is there a
dissemination
5
issue then that it comes as a surprise to me and
6
other people in this room that 12 out of 15 studies
7 --
and I am sorry if I got that number wrong --
8
were negative in their scope, or is that just
9
something that hasn't quite made its way to the web
10
site yet?
11
DR. MURPHY: No, there was an
issue in
12
that there was a window after BPCA was enacted in
13
which the sponsors had to be informed that they now
14 --
because they had been issued the written
15
requests earlier -- that they now were under the
16 new
legislation. They had been issued their
17
written requests under prior legislation.
18
In that window, a number of these studies
19
came in.
20
DR. RUDORFER: Dr. Leon.
21
DR. LEON: After hearing the
speakers
22
today, I think there is at least three avenues to
23
pursue simultaneously for being informed about this
24
topic, and all three will provide important
25
information about the public health risk and
310
1
benefits.
2
First, is looking at the existing clinical
3
trial data, or using the experts from Columbia
4
University, that is a good start.
I think what is
5
very important in looking at those data is we
6
haven't yet heard what percentage of people who
7
were screened to be in those clinical trials
8
actually were enrolled. Was it 5
percent, 10
9
percent, 80 percent? We have no
idea. That
10
certainly affects the generalizability of those
11
results.
12
The people we heard from this morning
13
might have been those who these data don't apply
14 to,
who would be excluded from trials, and we need
15 to
learn about those, and I will comment on that in
16
just a minute.
17 Also, in re-analyzing those data, I
would
18
really discourage the last speaker from dropping
19
data in which there were no suicide attempts. It
20
provides a false sense of risk actually.
It
21
inadequately characterizes exposure to the
22
medication.
23
The second avenue to pursue would be new
24
clinical trials, which were alluded to by a few
25
people today, and those should be designed with
311
1
very comprehensive assessments of suicidality,
2
agitation, hostility, akathisia, and assessments
3
that are sanctioned by the FDA, maybe with expert
4
advice again from Columbia and other universities,
5
other academic centers with expertise in assessing
6
those constructs.
7
They should carefully consider the
8
comparison group. That is
probably the hardest
9
part of designing those studies, whether it's a
10
withdrawal study, as Dr. Laughren alluded to, or
11
suggested, or maybe psychotherapy versus active med
12
versus combination, I am not quite sure, but that
13
certainly deserves discussion, and in those trials,
14
much broader inclusion criteria should be used than
15
have been used in the clinical trials to date.
16
The third avenue I would encourage is the
17 use
of existing observational data sets.
Now,
18
observational data sets, at the expense of internal
19
validity, at the expense of the association between
20
treatment and outcome provide wider
21
generalizability, and a much broader inclusion
22
criteria.
23
Dr. Pfeffer, her slides referred to at
24
least three different ongoing longitudinal
25
observational studies of children, depressed
312
1
children, and if those observational studies are
2
used, appropriate methods for adjustment and
3
stratification should be used.
4
They could consider some of the methods
5
used in the Division of Devices that are used to
6
adjust for observational differences.
7
I will stop there.
8
DR. RUDORFER: Dr. Katz and Dr.
Temple.
9
DR. KATZ: I just want to comment
on the
10
notion of how to better design trials in the future
11 to
look at this question. It is a very
important
12
question, it is one of the questions actually that
13 Tom
has drawn up, that we would like you to
14
discuss, and a number of people have already
15
mentioned it.
16
We think it's a good idea, too.
The
17
problem is I am not sure how to get those trials
18
done. As you have seen, pretty
much most of the
19
drugs in this class have been studied already,
20
their trials have been done under the pediatric
21
exclusivity provisions, and I am not sure we have
22 the
authority to require sponsors to go ahead and
23
redesign trials of the same treatments to have a
24
better look at trying to capture these events.
25
I would be very interested to know if
313
1
people have an idea about that, whether or not
2
there should be an NIMH-sponsored trial perhaps of
3
most of the drugs in this class, because I don't
4
think we can require the sponsors to do these
5
studies other than for the ones that might not yet
6
have studied them under the pediatric exclusivity
7
provisions, whichever ones those are, and there
8
aren't many.
9
DR. MURPHY: The one possibility
would be
10
that these products would be put on the list of
11
products that need to be studied.
We have an
12
off-patent list, but we can also reissue a written
13
request to a sponsor for a product which is on
14
patent, and if they refuse to do it, we could send
15
that request to the foundation at NIH.
16
Remember, I described earlier
this morning
17
there is a collaboration between NIH and FDA to
18
develop products for the off-patent including the
19
list of products that need to be studied. Some of
20
these products have come off patent, some will be,
21 and
even if they haven't, there is another
22
mechanism which FDA can issue a written request and
23
then if the sponsor doesn't want to do it, even if
24
it's still on patent, and it has a high enough
25
rating, it can be sent to NIH foundation.
314
1
I have to tell you, though, that the
2
problem is that funding for that foundation to do
3
studies is very small, so it would be getting in
4
line for a number of studies for which the funding
5 is
very limited at the moment, but those are the
6
possibilities I am aware of at this point.
7
DR. RUDORFER: Also, I wanted to
mention
8
that there is, in fact, an NIMH study that is
9
nearing completion, the treatment of adolescent
10
depression study, or TADS, that includes a
11
controlled trial of fluoxetine and placebo, as well
12 as
cognitive behavior therapy alone or with drug.
13
That is a 36-week acute trial followed by a 1-week
14
follow-up study in a total of 400 adolescents
15
coordinated at Duke.
16
I understand that the results should be
17
available by the beginning of June, so hopefully,
18 in
time to inform the FDA analysis.
19
Dr. Ebert has been waiting patiently.
20
DR. EBERT: It appears in some
ways that
21
many of these clinical trials may not reflect the
22
typical use for these agents. We
saw some data
23
that showed that many of these agents are used
24
other than for major depressive disorders, are
25
prescribed by physicians other than psychiatrists,
315
1 and
I am wondering if there is some way that we can
2
measure the adverse effects that we are seeing in
3 the
typical use.
4
We currently have the AERS system, which I
5
think admittedly is somewhat limited because of the
6
voluntary reporting that is necessary, but I am
7
wondering if the Agency could comment about some
8
other type of a postmarketing program that could be
9 set
up that might focus on this more rigorously.
10
DR. TEMPLE: The people from the
Office of
11
Drug Safety need to comment, too.
I just wanted to
12
make the observation that the most difficult
13
epidemiological situation you can identify probably
14 is
where the events you are looking for, both the
15
product of the disease and the potential product of
16 the
drug you are worried about, it is hard to think
17 of
anything more difficult, but some
18
epidemiologists ought to comment further on that.
19
I wanted to make one observation about
20
randomized withdrawal studies, which I like very
21
much. They are not a good way to
discover whether
22
these drugs cause suicidal thinking, because, by
23
definition, the people on those drugs are people
24 who
are doing well on them.
25
It is a possible way to show that the
316
1
drugs work in a situation that is somewhat
2
different from the high-intensity, high-support
3
setting of the acute trial, but I don't think that
4 is
going to get us the answer on suicidal thinking.
5
These are all bona-fide do-gooders or do-wellers if
6 you
like, so I don't think it is going to help on
7
that.
8
DR. RUDORFER: Dr. Pfeffer.
9
DR. PFEFFER: I wanted to comment
on
10
something that struck me, and that is the placebo
11
response rate seems to be relatively high in these
12
populations in these studies, and I wondered how
13 they
did compare to placebo rates in adults.
14
My sense is they are high and I would
15
assume maybe higher, and I wonder if that leads to
16 us
needing to think about other covariates, for
17
example, as will be done in the analyses, such as
18 the
environmental circumstances in which the
19
children are living, and to see what that feature
20 may
impact on not only the suicidal state, but the
21
potential for recovery.
22
I wonder certainly with the placebo rate
23
being a narrow range between the treated state, if
24 our
concerns about efficacy need to be rethought in
25
terms of developmental issue.
317
1
DR. RUDORFER: Does someone from
the FDA
2
want to comment on the placebo response rate in the
3
pediatric studies versus these same drugs in
4
adults?
5
DR. LAUGHREN: I don't have the
data in
6
front of me. My general sense is
yes, that the
7
placebo response rate in fact is an issue for both
8
adult and pediatric studies, perhaps even more of
9 an
issue in pediatric studies, and that may get at
10 the
issue I was raising earlier about heterogeneity
11
that you see when you try and capture a population
12
using the MDD criteria, but yes, it is definitely a
13
problem in both areas, but perhaps even more so in
14
pediatrics.
15
DR. RUDORFER: Dr. Laughren, is
there a
16 standard
way of assessing diagnosis in these MDD
17
trials? I mean as a matter of
just the sponsor
18
will say these subjects met DSM-IV criteria, do we
19
know if they used any kind of structured interview?
20
DR. LAUGHREN: They almost always
use
21
some kind of structured interview.
22
DR. RUDORFER: So, presumably, if
there
23
were comorbidities, those would be captured?
24
DR. LAUGHREN: Yes, and there
often is
25
comorbidity.
318
1
DR. PINE: Related to that
question, could
2 I
make one comment about it. When one
looks
3
particularly across the recent studies, while every
4
study will say that they used a standardized
5
assessment, there is really a quite marked
6
variability across studies in terms of the way in
7
which they documented the rigor of that approach.
8
So, if you read the recent letter in JAMA
9 from Wagner, that talks about the process of
10
establishing the diagnosis and the reliability
11
study, that reads very differently from some of the
12
other studies that maybe had a lower placebo
13
response rate or smaller samples.
14
So, I was wondering if it might be
15
possible to in some way evaluate or rate the rigor
16
with which both the diagnosis and the outcome
17
variables were assessed across the studies, paying
18
particular attention to issues of training and the
19
demonstration of reliability by those investigators
20
conducting the trial and using the instrument.
21
DR. LAUGHREN: It would be very
difficult
22 to
do that after the fact. If they claim to
have
23 done
it in a particular way, to document whether or
24 not
it had been done in that way, involve an
25
enormous amount of work, and given the time at
319
1
which these studies were done, you know, going back
2
four, five years, it would be hard to imagine how
3
that would be helpful.
4
DR. RUDORFER: Dr. Wang.
5
DR. WANG: I think in addition to
6
considering what the optimal design would be, if we
7 all
had our choosing, we should keep in mind that
8 at
the best, it will take a long time to do them
9
even sorting out all the other logistics, so I
10
think in the meantime, it is important to consider
11 how
to enhance the use of this existing data set,
12
which will be arriving soon enough, to study this
13
question.
14
One thing I am particularly concerned
15
about is you may lose an effect in the overall data
16 set
that you would otherwise be able to see in a
17
high-risk population.
18
I think in that list of covariates that
19 you
are asking the sponsors to all submit, to also
20 add
variables that will allow you to identify
21
high-risk populations, such as people -- some that
22
come to mind, kids that have insomnia at baseline
23 or
high anxieties, severity symptoms, or family
24
histories of bipolar illness, things that allow you
25 to
sort of concentrate on a group that is likely to
320
1
potentially show an effect.
2
DR. RUDORFER: Dr. Gorman.
3
DR. GORMAN: Again back to the
theme of
4
information flow, I was wondering if someone from
5 the
FDA could explain what bars we would have to
6
meet for changing the labeling on these substances
7
today. The labels get made when a
new product is
8
approved, but then are modified through many
9
mechanisms, I have no idea.
10
What do we need to do to put a precaution,
11
warning, or black box, or side effect or adverse
12
event that lists these as potential -- what bar do
13 we
have to meet to potentially include these in the
14
label?
15
DR. KATZ: As you heard from one
of the
16
speakers in the open session, it isn't required,
17 for
example, when we are contemplating putting
18
something in the Warning section that we have
19
absolute proof that the drug causes a particular
20
adverse event, but reasonable suspicion.
I forget
21
exactly what the words are.
22
On the other hand, of course, two points,
23
one, it is obviously a judgment as to whether or
24 not
there is reasonable evidence that a drug is
25
linked to a particular adverse event.
So, if you
321
1 ask
what the bar is, it is hard to say. It
is
2
highly case-dependent.
3
On the other hand, even though the law
4
permits us to include things in the Warning section
5
that we have not yet proven to be associated with
6 the
drug, there is always the risk of including
7
such events when we aren't really sure or almost
8
sure that the drug did it, because, number one, it
9 is
distracting, but beyond that, you might be
10
giving false information.
11
So, as a general matter, we tend to put
12
adverse events in the Warning section when we are
13
pretty sure, when we think we have pretty good
14
evidence that the drug actually does it as opposed
15 to
its just being associated with it.
16
A boxed warning again is a judgment, but I
17
would say, as a general matter, as well, we don't
18 put
a description of adverse events in a boxed
19
warning, which is sort of the most stringent
20
warning you can apply in a labeling unless we
21
really believe that the drug is causally related to
22 the
adverse event.
23
Then, of course, we don't put all causally
24
related adverse events in boxed warnings, only
25
those which we think are particularly serious, not
322
1 to
say that suicidal behavior would not be one of
2
those events, but really boxed warning and pretty
3
much warning, we like to have pretty good evidence
4
that the drug actually did it.
5
Of course, the type of evidence that is
6
brought to bear on the question of whether or not
7 the
drug is causally related varies. We like
to
8
have controlled data. It isn't
always controlled
9
data.
10
Sometimes for rare events, as you heard
11
earlier, events not associated with the condition
12
that you are looking at, postmarketing data,
13
comparing reporting rates to what we know about
14
background rates usually suffices.
I am not sure
15 we
can apply that sort of reasoning to this case.
16
DR. TEMPLE: We are particularly
17
interested in telling people of things they can do
18 to
avoid problems, if there is such a thing, that
19
seems reasonably likely to do it.
Current labeling
20
already does tell you that early after treatment
21
starts is the time to watch out.
22
It doesn't attribute that to the drug, but
23 it
doesn't seem out of the question that wording
24
like that could be enhanced and made clearer.
25
Everyone seems to agree that that is a dangerous
323
1
time whether you agree on why it is a dangerous
2
time or not.
3
So, there are things like that that can be
4
done. Another option that we have not used for most
5 of
these drugs is to provide information as best we
6 can
for patient or caregiver use. Those are
all
7
possibilities.
8
DR. MURPHY: Could I put a
pragmatic
9
response to that answer? That is,
that I think one
10 of
the other things that we need to consider --
11
it's in your questions -- is that to get a change
12 in
label, let's just assume for some reason that
13
people walked out of here today and wanted to
14
change the label.
15
It takes a while to get all that done and
16 by
the time you came back this summer, you might
17
want to change the label again.
18
So, I think what we are going to be asking
19 you
or in one of the set of questions is what are
20
your recommendations about what FDA may or could
21
possibly do in the interim, because I think that
22
everyone is very interested in what additional
23
information we can get, and we would like to make
24 it
the most efficient way of transmitting
25
information, which would be together instead of
324
1
trying to change things maybe two times.
2 So, I think the pragmatics of it are
what
3 can
we do to help better inform people before the
4
late summer meeting in which we hope to have a more
5
definitive response.
6
DR. RUDORFER: Dr. Nelson.
7
DR. NELSON: I think I will continue this
8
conversation about labeling rather than what I was
9
originally going to say.
10
I have two suggestions. What
struck me in
11
your remarks about the timing was the delay it
12 appeared
to take for you to actually get the data
13 you
were asking the sponsors to provide.
That
14
could be inadvertent or it could be, in fact,
15
duplicitous.
16
So, I would suggest that you tell them
17
that, in fact, if they don't provide the data you
18
want, that you will label it based on just the
19
British decision, with a warning, would be the
20
first suggestion.
21
But the second is, to answer Dianne's
22
question about a notice, I think you could honestly
23
take Appendix 2A and put that in the letter to both
24
health care professionals and to patients on the
25
medication saying the FDA is really worried about
325
1
this signal and we want to look at this data, and
2 if
you are worried, too, you ought to talk to your
3
clinician that prescribed it and discuss those
4
concerns and name the drugs.
5
It is unclear to me why someone couldn't
6
have the opportunity to see that signal and to make
7
their own evaluation as to whether or not they
8
would want to be slowly tapered and put on the one
9
drug that seems to be so far a winner in all of
10
this, which is fluoxetine.
11
DR. LAUGHREN: Let me just clarify
one
12
thing. We have the data from the
companies, the
13
ball is now in our court, so we are not waiting for
14
anything at this point from companies unless the
15 committee
feels that there is some deficiencies
16
here in terms of case finding, but we are satisfied
17
that we have what we need.
18
It is now a question of working on a
19
reclassification and designing an analysis. We
20 have
what we need.
21
Regarding the second issue of
22
disseminating Appendix 2 to prescribers, I am not
23
sure what purpose would be served in doing that. I
24
mean we have already issued a health advisory in
25
October saying that we are concerned, that we can't
326
1
rule out an increased risk of suicidality.
2
If we are not comfortable with what is in
3 the
numerators for these risks that are displayed
4 in
that table, I am not really sure what purpose is
5
served in disseminating that.
6
DR. NELSON: One brief response to
that
7 and
then I will be done. What bothered me in
8
listening to the testimony this morning is the
9
amount of off-label use, and the amount of times
10
that people mentioned that they were given samples.
11 I
would even go so far as to wonder if the handing
12 out
of samples is marketing outside of an
13 indication
where you could even come after a
14
company.
15
So, part of my desire to inform clinicians
16 is
to try to scare them away from off-label use
17
frankly. That bothers me, the
amount of off-label
18 use
that appears to be going on in this particular
19
market.
20
DR. RUDORFER: Dr. Griffith.
21
MS. GRIFFITH: I need to clarify,
I am not
22 a
doctor, I am a consumer, I am a parent, and as a
23 lay
person, the most troubling outcome I think of
24
this morning's and this afternoon's presentations
25 was
the urgency with which this needs to be
327
1
resolved.
2
After the presentation by Dr. Hammed, I
3 was
really struck by, in covering the analysis
4
plan, the last statement it remains to be seen that
5 if
we have enough statistical power, whether or not
6
there is enough statistical power.
7
My question is what happens then, if there
8 is
not enough evidence to make a conclusion, how
9
does the FDA inform the public, because as you say,
10 you
put out an advisory on October 27th, which I,
11 as
a parent and as a consumer, read, found it
12
terribly confusing.
13
It was reported on very contradictorily,
14 and
what I am suggesting is I think the FDA is
15
going to have a credibility problem if it does not
16 get
out ahead of this with some very public
17
statements about where it is going with these
18
studies and with the data.
19
DR. RUDORFER: Dr. Goodman, did
you want
20 to
respond?
21
DR. GOODMAN: I think it is going
to be
22
some time until at least I am comfortable that we
23
have enough data and analyze it properly to be sure
24 of
the connection with suicidality, however, I
25
think that myself -- and my guess is there are
328
1 other
people around the table -- are more
2
comfortable with the assumption or, to use these
3
other terms, have a reasonable suspicion that there
4 is
a subgroup of children who develop an
5
idiosyncratic reaction to SSRIs, that include
6
symptoms like insomnia, agitation, maybe
7
suspiciousness, hostility, and could possibly lead
8 to
violent behavior including self-harm.
9
I think a lot of clinicians are aware of
10
this already. I think that my
colleagues in child
11
psychiatry and pediatricians who are informed on
12
this issue are very attentive when they are
13
starting medication, if they are seeing any of
14
these signs, they adjust the dosage, they may stop
15 the
medication, they certainly don't increase the
16
dosage.
17
So, there are measures that can be taken
18 now
by clinicians as long as they are aware of it,
19 and
by parents who are made aware of it, to take
20
steps that may reduce the development of this
21
syndrome, whatever we want to call it, in a
22
susceptible group of kids that may or may not
23
increase risk for more serious adverse events that
24
include suicide.
25
MS. GRIFFITH: Just to follow up, I
don't
329
1
disagree and I feel that I have always been well
2
informed by clinicians, but I think that there is a
3
group of people who have not been able to either
4 look
at the data or not had access to good
5
therapeutic care, and I think that it is going to
6
become a public relations problem very quickly.
7
If the data comes back, if you are unable
8 to
use it when it comes back prior to this meeting
9 in
the summer, you are extending some sort of
10
reasonable period by which you can reasonably
11
inform the families, and it will snowball and get
12
completely out of control.
13
DR. RUDORFER: Dr. Katz.
14
DR. KATZ: One of the questions we
have of
15 the
committee is what, if anything, should we do in
16 the
interim while we are waiting to get the final
17
analyses. Of course, as a number
of people have
18
suggested, it is possible that come this summer
19
when we do the analyses based on these
20
resubmissions of the data, that we won't be able to
21 say
anything definitive.
22
What we really want to know from you
23
folks, first of all, in the interim, what, if
24
anything, we should say, and it sounds like at
25
least some people think we should do something
330
1
although I am not yet sure if and what other people
2
think should be done.
3
But it is possible that come this summer,
4 we
really won't be in a position to say anything
5
more definitive.
6
What we really want from you folks is, in
7
part, whether or not there is anything else you
8
think we can get from the data or whether or not
9
there are any other additional analyses that we
10
should do, so that we get as much as we possibly
11 can
out of the data, so that if we do come back in
12 the summer and say, look, we can't give you a
13
definitive answer, at least we can know that we
14
have done everything that we possibly could with
15 the
data that we have in front of us at the moment.
16
So, those are things we definitely want to
17
hear from you about.
18
DR. RUDORFER: Dr. Temple.
19
DR. TEMPLE: I just want to sort
of remind
20
everybody that what provoked the most recent
21
interest in this subject was those data, the 127
22
cases. If those prove to be
uninterpretable, we
23 are
back where we were.
24
What we then have is very impressive
25
individual reports of bad outcomes.
Those have
331
1
always been impressive when people have tried to
2
look at those in controlled trial environments and
3
things like that, and pooling our study data, they
4
haven't turned up at least so far.
5
There have been some criticisms of the way
6
that was done, but leaving that aside, they haven't
7
turned up. The difficult question
always is what
8 to
do with reports that have considerable cogency
9 to
them. I mean it sort of looks like something
10
happened when the person started the drug, it does,
11
that you can't really confirm in controlled trials,
12 and
that is always a problem with the postmarketing
13
data we get.
14
Sometimes the events aren't the very thing
15
that you are worried about happening in people with
16
that diagnosis. In this case, as
I said before, it
17 is
particularly difficult because people who are
18
depressed are the very people who have some of
19
those events.
20
Now, whether it looks like they were
21
accelerated or not are the kinds of things we have
22 to
think about, so as Russ said, we are very
23
interested in views as to what we can say that
24
would be useful now, apart from waiting for the
25
results of the trials, if there is such a thing.
332
1
DR. RUDORFER: We have several
speakers
2
lined up to continue the discussion on Question 1
3
regarding capturing all events of potential
4
interest, and I will ask everyone else to hold your
5
questions, and then we will move on to Question 2.
6
There is a lot of overlap, and I have been asked to
7 try
to keep these separate and distinct.
8
If we could turn to Dr. Maldonado,
9
followed by Dr. O'Fallon, please.
10
DR. MALDONADO: I am sorry to
bring you
11
back to the BPCA and rule. I want
to clarify the
12
point, the failed trials that the FDA is seeing
13
right now has been an issue of cost of doing
14
business for the pharmaceutical industry for
15
generations, is that when those so-called negative
16
trials happen, the pharmaceutical industry doesn't
17 even
bother to come into the FDA with those trials
18
because they know they are not going to get
19
anything out of that.
20
Now, you are seeing it in the context of
21 the
BPCA because it is necessary to disclose, and
22
because there is incentive to disclose it. So,
23
this is not a new phenomenon and I think that the
24
comment that the pharmaceutical industry is not
25
making the efforts that they should make is
333
1
unfounded.
2
A lot of these trials -- that is why they
3 are
called trials -- a lot of these drugs failed,
4
failed repeatedly, and those failures actually had
5 to
do more with the ignorance of the people
6
developing the compound than with the drug itself.
7
It is a process of learning until the
8
researchers fine-tune what they want to find. Not
9
only that, if there is a doubt that these studies
10 are
being done according to GCPs, the FDA has the
11
authority to have that oversight.
12
Not only that, the FDA has a very
13
historical authority now given by the government to
14
issue the written request. So,
those studies are
15 in
response to written requests issued by the FDA.
16
So, if those responses are not accurate
17 and
are not fulfilling the demands, then, there has
18 to
be a corrective that should happen there, just
19 for
clarification.
20
DR. RUDORFER: Dr. O'Fallon.
21
DR. O'FALLON: We are talking
about three
22
major topics here, and we keep flipping around
23
among them. One of them is the
potential that we
24 can
get out of this re-analysis. The second is
25
suggestions, advice as to what to do for future
334
1
studies. The third is the
labeling issues.
2
The questions that we are getting are
3
primarily focused on this re-analysis, at least the
4
ones that I saw. I want to say as
a statistician
5
that I don't have a whole lot of hope for your
6
being able to get good information out of the
7
planned re-analysis. I think it
should be done,
8 but
I don't think it is going to be because you are
9
going to get the information.
10
As a statistician again, I have learned a
11
long time ago that if you don't get your data right
12 the
first time, that it is very, very difficult to
13 go
back and get the information after the fact, and
14 I
am afraid that you are going to find that is a
15
problem.
16
If the data were not collected very well,
17 for
whatever reason, in those original studies, you
18 are
going to have a hard time finding it, and there
19 is
no such thing as being able to go back.
20
For example, if something is a genetic
21
defect, if there is really a genetic defect that is
22
underlying the ones that flip out, the kids that go
23
crazy, no one will ever know because we don't have
24 the
information, we didn't ask about it, and there
25 is
no way to go back and get it.
335
1
I am afraid that is what is going to
2
happen with this study.
Nonetheless, I think it is
3
worth going forward because I think you are going
4 to
learn a whole lot about methodologic issues when
5 you
struggle to analyze it, and I think that will
6 be
valuable information for writing future written
7
requests for evaluating future studies, so I think
8
there is a lot to be learned about it.
9
I don't even want to go on the labeling,
10 but
I have got a whole list of stuff there.
11
DR. RUDORFER: We will come back
to that.
12
Dr. Chesney.
13
DR. CHESNEY: Two issues with
respect to
14
Question 1. The first one, I
think we have already
15
gone over several times, but I would really
16
strongly encourage, if it is possible to go back
17 and
look at every patient, to look at this
18
stimulant syndrome issue, this mania, this
19
irritability, and so on, which I must say I was not
20
fully apprised of at all until we came today, and I
21 am
most impressed when I hear from, again in the
22
open session, about how some of these events
23
occurred very quickly.
24
I know the potential explanation of being
25
stimulated out of lethargy, but this sounds like
336
1
something different to me, which brings me to my
2
second question.
3
I wondered of any of the psychiatrists
4
could tell us if there is any association with drug
5
levels, because certainly in my field, which is
6
infectious diseases, drug levels are imperative or
7 you
wouldn't know what you were treating or how
8
well you were treating it or whatnot, but certainly
9 we
heard levels at autopsy referred to as being
10
three times I guess what was expected, and then Dr.
11
Goodman made the comment about adjusting dosage.
12
Do we have any idea of what the dosages
13
were in these studies and how they correlated with
14
body weight or levels? For those
of us not in the
15
field, I just don't know anything about the value
16 of
pharmacokinetic studies in these drugs.
17
DR. LAUGHREN: Just to comment on a couple
18 of
your questions. For the most part, blood
levels
19
were not obtained in these trials.
Any
20
pharmacokinetic data for these pediatric programs
21
were done in other smaller studies. For the most
22
part, I don't think we are going to have much luck
23 in
getting PK data here.
24
In terms of dosages, these were mostly,
25
virtually all flexible dose studies, so patients
337
1
were dosed within a range, usually the recommended
2
range for that drug. They were
not fixed dose
3
studies. We have dose
information, but without
4
something to link it to, it probably is not going
5 to be very productive.
6
But I wanted to come back to your first
7
point because now several people have raised this
8
question about some kind of a stimulation syndrome
9 and
linking that in some way with mania. If
we are
10 to
look for that, we have to know what it is that
11 we
are looking for.
12
I mean there has to be some kind of
13
definition. Are we talking about something that is
14
linked specifically to suicidal behavior or
15
something that occurs independent of suicidal
16
behavior. I am not sure if this
entity can be well
17
enough defined for us to search for it.
18
We have over 4,000 patients involved in
19
these trials. To head off looking
for a syndrome,
20 we
have to know what it is that we are looking for.
21
DR. CHESNEY: Can I just respond
to one
22
comment. I think on several
occasions we heard
23
that it was actually homicidal behavior that seemed
24 to
arise from mania, and if we just look at
25
suicide, maybe that is not all we want to know
338
1
about.
2
DR. RUDORFER: Tom, I wonder if I
could
3
interject a question for you. I
think the concept
4 of
akathisia, which again has come up repeatedly,
5
captures a lot of what various speakers are talking
6
about, and I wonder if the Agency's experience with
7
antipsychotic drugs would be helpful in that regard
8 in
terms of definition.
9
DR. LAUGHREN: We could certainly
search
10 for
akathisia. That term is reasonably well
11
understood I think clinically and would very likely
12
appear in the electronic database, or one I suppose
13
could come up with related terms that might get at
14
akathisia if it wasn't specifically named.
15
But again, my question is are we looking
16 for
that symptom by itself or are we looking for
17
that in association with some other behavior.
18
Again, there is a very widespread belief that
19
akathisia is linked to suicidal behavior, but I am
20 not
sure how strong the data are supporting that
21
belief, that is really the question.
22
But again, if we are going to
search this
23
database for something other than what it has
24
already been searched for, we have to have some
25
fairly specific guidance about how to do that.
339
1
MS. BRONSTEIN: My comments are
about
2
labeling and if you want me to wait, I will, or I
3
would like to get them off my chest now if I could.
4
If I heard nothing from this morning's
5 testimony,
I heard repeatedly that people feel the
6
need for patients and family to have more
7
information than they have currently.
8
I think that is really our responsibility
9 to
do something about it whether it is after this
10
meeting or after the summer meeting.
I think we
11
need to get something out there that describes
12
akathisia in a way that patients can embrace it and
13
understand it, and family members can watch for
14
this radical change in behavior.
15
I am seeing it as an apparent link to
16
either homicidal or suicidal behavior from the
17
testimony this morning and from what I have read,
18 as
well.
19
DR. RUDORFER: Dr. Ebert.
20
DR. EBERT: Most of my comments
also had
21 to
do with labelings. I just briefly wanted
to
22
react to what was stated earlier, though, again
23
about the issues of going beyond just the suicidal
24
behavior and whether it's akathisia or whether
25
there may be some other characteristics which
340
1
clearly indicate that -- and I am not in the area
2 of
psychiatry, so you will have to indulge me for a
3
second -- but just the whole issue of kind of a
4
concept of self versus others, whether it's through
5
homicide or it's hostile behavior or
6
aggressiveness.
7
To me, these things all seem to be a
8
constellation of the same types of syndrome that we
9
would be looking at.
10
DR. RUDORFER: Dr. Fink.
11
DR. FINK: Another sort of global
concern
12 --
and I think it may be particularly apropos to
13
this class of drugs -- is that when these clinical
14
trials are performed, they are usually performed by
15
experts in the field, yet much of the usage today,
16
particularly in the managed care environment, is
17
prescription of these drugs by non-mental health
18
trained professionals.
19
The results of a clinical trial performed
20 by
mental health professionals where you are
21
already using a highly select audience and highly
22
select practices may bear little relationship to
23
what you see with the drug in use in the real
24
world.
25
From a labeling standpoint, it would make
341
1
sense potentially to say that at least off-label
2 use
of these drugs really should be highly
3
restricted to mental health professionals or make
4
some kind of wording that would imply that, because
5 I
think that off-label use of these drugs by
6
non-mental health trained professionals seems to be
7
problematic, and it may well be that much of the
8
placebo effect that we are seeing in the clinical
9
trials is because they are receiving counseling
10
about mental health.
11
I am more familiar with asthma trials.
12 When we do asthma trials, we see a tremendous
13
placebo effect which is asthma education. My guess
14 is
in mental health trials, there is a tremendous
15
placebo effect because you are seeing a mental
16
health professional.
17 DR. RUDORFER: Dr. Leon.
18
DR. LEON: It would be interesting
to know
19
what items were captured in the severity ratings,
20
because if we knew the items that were there, then,
21 we
could see which ones correspond to the symptoms
22 we
heard of this morning, and look at treatment
23
emergent symptoms, symptoms that weren't there at
24
baseline, on the severity rating, that were
25
exacerbated during the course of this trial, so
342
1
looking at changed scores on a handful of a
2
priori-defined symptoms from the rating scales
3
would be very helpful.
4
DR. GOODMAN: Along those lines,
as I
5
mentioned earlier, the Hamilton has an item on
6
agitation, the CDRS has an item on irritability, so
7
that could be a first quick look, and you wouldn't
8
have to look at treatment emergent, you can look at
9
rating scale items.
10
I agree that one needs to give careful
11
thought into what symptoms or how we are describing
12
this constellation of symptoms, because it could be
13
very problematic.
14
For one reason, a number of symptoms you
15
would expect to get better with the SSRIs, and what
16 we
are really looking for is a minority of patients
17 in
whom you see a paradoxical increase in those
18
symptoms.
19
So, I think we need to take a very careful
20
approach to this analysis.
21
DR. RUDORFER: We have four more
questions
22 on
this topic.
23
I am sorry. Dr. Laughren.
24
DR. LAUGHREN: Just one follow up
on a
25
suggestion that has come up from several committee
343
1
members now about looking at items from the rating
2
scales. That was actually done
here, and it turned
3 out
not to be very helpful.
4
Now, this was a similar analysis that had
5
been done with the adult data years ago, for
6
example, looking at patients who move from looking
7 at
the suicide item on the HAM-D and looking at
8
patients who move from zero to 1 to a 3 or 4.
9
That did not detect a signal in these
10
trials, and part of the problem may have been that
11
these events often did not occur at a time when the
12
HAM-D would be done, because the HAM-D is done at
13
regular intervals.
14
If the event occurs between visits, which
15 it
almost always does, and then the patient is
16
discontinued at that point, you never get a HAM-D
17 or
whatever other instrument is being used.
18
So, companies did try that approach, and
19 it
was not particularly productive.
20
DR. RUDORFER: We are now going to
turn to
21
Drs. Malone, McGough, Pfeffer, and Ortiz, and then
22
move on to Question 2 more specifically.
23
DR. MALONE: I am sorry, I just
stepped
24
out, so I may have missed things that were just
25
discussed, but I was thinking that looking at
344
1
agitation would be an important thing if you think
2
about the way we use the recent meetings on
3
antipsychotics and agitation.
4
Agitation often leads to harming of self
5 or
others, and it might be a proxy for looking at
6
suicidal behavior. So, searching
the electronic
7 database
for agitation, violence, and trying to
8
construct an agitation -- I don't know what to call
9 it
-- but try to construct agitation and see if it
10
does differ in those who are having suicidal
11
ideation or having other such problems.
12
The other thing, I end up currently
13
treating children with autism, and I think this
14
whole activation syndrome is something that anyone
15 who
treats children with autism worries about if
16
they are going to consider giving an SSRI.
17
There is some sense in which I think you
18
could look at fairly quickly in a controlled trial
19
whether populations other than depressive
20
populations get agitation or get activated, and
21
then get some information whether these drugs in
22
children, in fact, cause this activation syndrome.
23
DR. RUDORFER: Dr. McGough.
24
DR. McGOUGH: This is really a
segue I
25
think to the labeling issue which keeps coming up
345
1
again and again. First, as far as off-label use
2
goes, child psychiatrists could not treat severely
3 ill
kids without off-label prescriptions, there is
4 no
doubt about that.
5
Secondly, even in the absence of
6
scientific clinical trial evidence, a physician
7
needs to be free in specific instances to choose to
8
take the risk of using a medicine even in the lack
9 of
a controlled study. Again, there is no
way to
10
meet the needs of these really severe kids without
11
this.
12
To your point, unfortunately, there aren't
13
enough child psychiatrists trained and available to
14 do
this, so it is left to other practitioners, and
15
what I was really struck with, hearing the stories
16
this morning, is many of the cases we heard were
17
kids just naively given adult titration regimens at
18
adult doses with no consideration to slow
19
metabolizing, in Caucasian kids particularly, with
20 no
concern about the need to monitor for akathisia
21 and
early onset activation, so I see we can't
22
restrict non-psychiatrist prescribing, we now have
23
pediatricians, family docs, nurses, psychologists,
24 all
of whom will be prescribing these medicines.
25
There has to be some way to really notify
346
1
people or put people on notice that at least in the
2
absence of efficacy data, you have to be very
3
concerned about safety, and if there is any
4
labeling tweaking to be done, that is what I would
5
want to see put in.
6
DR. RUDORFER: Dr. Pfeffer.
7 DR. PFEFFER: I have a number of questions
8
that have to do with the analysis issues and
9
perhaps my concern is having heard the families and
10 the
sense of their urgency, if while the Columbia
11
group is evaluating the suicidality question, if
12 one
might look at the data in a variety of other
13
ways that might inform us about, for example, who
14
improved and who didn't improve.
15
Who improved within the placebo group and
16 who
improved within the treated group, and what are
17 the
predictors of that or vice versa, what are the
18
predictors of a poor outcome, and we might find
19
that that might give us some very important clues
20 as
to the way that this population are responding
21 to
the drugs.
22
The question also that I have, and I
23
assume it must have been done, but I am not sure,
24 and
that is whether or not randomization really
25
worked, and especially did randomization work, for
347
1
example, in the suicidality issue.
2
I don't know if that has been looked at,
3 and
certainly once Columbia group looks at the
4
definition of suicidal behavior, it will be looked
5 at
again, but that would be an important question
6 to
also look at.
7
Then, if I might contribute some
8
information, for example, I know in the venlafaxine
9
studies, they were doing blood levels of
10
venlafaxine because they were looking at the
11
question of slow metabolizers or not, so I wonder
12 if
that data might be able to be looked at to, to
13
give us some clues about issues of metabolism.
14
DR. RUDORFER: Thank you.
15
Dr. Ortiz.
16
DR. ORTIZ: My comments, I think
are in
17
response to a couple of things that Dr. Chesney
18
brought up. As far as levels in
psychiatry, what
19 we
certainly know is that the Sinemet kinds of
20
medicines, which are dopaminergic, can cause
21
psychosis, and it is at different doses for
22
different individuals, the same thing with
23
amphetamines, they also can cause psychosis.
24
Again, it is not predictable in each individual.
25
I would also like to follow up on your
348
1
suggestion to specify the adverse effects and the
2
descriptions of them a little better.
3
As a psychiatrist, when I am watching
4
someone that I am concerned about, that may be
5
developing hypomania or mania, I am watching how
6
their speech patterns change, I am watching their
7
activity levels, I am monitoring their sleep, and I
8
think a little more precision in those kind of
9
descriptions might be helpful.
10
DR. RUDORFER: Dr. Andrews will
ask the
11
final question related to Question 1.
12
DR. ANDREWS: I have some concerns
about
13 the
exploration of this activation syndrome in the
14
context of the existing clinical trial data.
15
First of all, as has been said, we may not
16
know what the elements of that syndrome are, but in
17
addition to that, do we know whether the elements
18 of
that potential syndrome were collected
19
diligently, frequently, and similarly across all of
20 the
studies, and I think that needs to be addressed
21
before going into that expedition.
22
If not, I would encourage the
FDA and the
23
analysts to look at more objective endpoints, which
24 I
think are the ones that were established for
25
suicide events.
349
1
I have a bit of concern that the study may
2 not
answer all of the questions because of the
3
issue that was raised earlier regarding
4
generalizability. These patients
may not resemble
5 the
patients who are treated with these drugs.
6
They are probably treated in a different
7 way
in terms of dose titration in the context of a
8
clinical trial, and in the context of a clinical
9
trial, patients tend to be
monitored more
10
carefully, so that perhaps those at highest risk of
11
suicide or suicidal ideation might have been
12
identified earlier with other symptoms and
13
withdrawn from drug or had drug titrated down.
14
DR. RUDORFER: Thank you.
15
I think we will come back to some of these
16
issues. The sense I have from the committee is that
17
while people have reservations about the
18
limitations of the existing database, the sense
19
seems to be that we would endorse going ahead with
20 the
Columbia reclassification, but with some
21
additional measures.
22
Dr. Laughren had also specifically asked
23 us
about the appropriate categories in terms of the
24
definition of "possibly suicide related" and
25
"suicide attempt," and I wonder if anyone has any
350
1
feedback for the FDA on those questions.
2
Dr. McGough.
3
DR. McGOUGH: I was just speaking
from
4
experience and also the work Dr. Shaffer showed.
5 You
know, my view about cutting is that it is not a
6
suicidal behavior, and others might disagree, but
7
that would be my approach to that.
It would be not
8 to
classify cutting or superficial cutting
9
certainly as a suicidal behavior.
10
DR. RUDORFER: Dr. Chesney.
11
DR. CHESNEY: I was interested
again this
12
morning to hear in a number of instances that
13
people took a drug, took a dose and then found
14
themselves in jail and did not know what had
15
happened in the interim.
16
How is that described in psychiatric
17
terms, is that confusion of thought or absence of
18
presence, or is that something that you could pull
19
out? That seems a fairly profound
confusion to
20
just absent oneself from the situation and yet do
21
some fairly striking things.
22
DR. LAUGHREN: It is
phenomenologically an
23
amnestic syndrome of some sort. I
did not see that
24 in
these trials. At least it was not
described as
25
such.
351
1
DR. GOODMAN: Also,
phenomenologically, it
2
would be a dissociative or fugue state.
3
DR. CHESNEY: Was that asked for
in the
4
trials? Was that a question that
was on the --
5
DR. LAUGHREN: No, I am sure it
was not.
6
DR. LESLIE: I wanted to add two
comments.
7 One is on the Question No. 1, which is I
think part
8 of
this is a process question of where we go from
9
now. When I look at Dr. Hammad's
variables that he
10 has
listed, I think there are some that are missing
11 and
it would be good to redistribute that list to
12 the
committee for review.
13
For example, I only see after
14
discontinuation. I don't see on
an increase of
15
dose or decrease of dose. The
issue of family
16
history has come up.
17 I think all of us or there is a good
18
majority here that are concerned about aggressive
19
instances, and some of the family stories this
20
morning were not of kids who were feeling down.
21
They were of kids who acted suicidally because of
22
impulsivity, and not because of a suicidal
23
symptomatology that had been ongoing.
24
So, I think those things are important and
25 I
also worry about what is hidden in some of the
352
1
other neurological, et cetera, categories that are
2
listed.
3
So, again, I don't know the process here
4 and
how you all feel about doing this, but I think
5
redistributing this list for some suggestions of
6
some of the risk factors and things that might be
7
important to be looking at, as several of the
8
speakers have said, would be important.
9
I also wanted to say that I am impressed
10
that the American Academy of Child and Adolescent
11
Psychiatry has been here and other groups, but
12
there is no one here from the American Academy of
13
Pediatrics, representing the American Academy of
14
Pediatrics, although several of us are
15
pediatricians and on that committee, and there is
16 no
one from the National Association of Nurse
17
Practitioners, and there is no one from the
18
American Academy of Family Practice Doctors, and
19
reaching out to those organizations on an official
20
level, since so many of us are the ones that are
21
giving those medications, would be an important
22
step to be taking.
23
DR. LAUGHREN: In terms of the
lists of
24
variables, all committee members have that. It is
25
attached to a memo that I wrote.
We would be happy
353
1 to
accept suggestions at any point, it wouldn't
2
have to be at today's meeting, of additional
3
covariates that you think might be important to add
4 to
this database, so please free to do that.
5
DR. RUDORFER: Dr. Gorman.
6
DR. GORMAN: If all of these 15
studies
7
that we are going to re-review were not intent to
8
treats, analysis based on intent to treat, we would
9 not
be able to answer Dr. Chesney's questions.
10
DR. RUDORFER: Dr. O'Fallon.
11
DR. O'FALLON: One process
question. Do
12 you
have data for all the patients in all of those
13
studies? Do you have the detailed
data for all of
14 the
patients in all of the studies?
15
DR. LAUGHREN: What we have right
now are
16 in
terms of data sets. We have the data
sets for
17 the
variables that we specifically asked for.
18
Again, those are listed in an appendix to my
19
review. So, that is what we have
in terms of an
20
electronic data set for all patients, but it is
21
limited to those variables that we asked for.
22
DR. TEMPLE: Just with respect to
intent
23 to
treat, we expect to see all patients randomized
24 who
at least got some treatment. It is
typical in
25
symptomatic treatments not to include people who
354
1
don't get a treatment. You can
debate that, but it
2 is
usually not a big loss, but anybody who was
3
treated should be in those analyses.
4
DR. PERRIN: It does seem, having
read
5
that list of variables on the way down this
6
morning, that there are some important gaps. They
7 do
include again some of the factors Dr. Pfeffer
8
mentioned before which are really in the social
9
environmental phenomena that might influence rates
10 of
responsive treatment or might influence rates of
11
suicidal behaviors.
12
It does seem like you don't have a lot of
13
sort of data over time. It is
almost like an
14
adverse event reporting system, if I am reading the
15
data set right. In other words,
you don't have a
16 lot
of information on other response to treatment.
17
We have heard, for example, a lot of
18
discussion without a lot of evidence that the first
19
week or two or three of treatment is really
20
critical, so one would wonder a lot about what kind
21 of
things happened during that time that you do
22
have data on, and you talked about the notion that
23
maybe the next clinical trials might be a
24
withdrawal trial.
25
Again, there is a moderate amount of more
355
1
anecdotal than good evidence base that withdrawal
2 is
a very high-risk time, as well, for kids on
3
SSRIs, and again there, having some sense of what
4
happens relatively immediately in that two or
5
three-week time would be extremely helpful, but I
6
have a feeling you don't have those data for even
7 the
start-up time.
8
DR. LAUGHREN: Could you say a
little bit
9
more about how you would characterize that early
10
response? Are you talking about
looking at formal
11
assessments, HAM-D, and so forth?
I mean clearly,
12 we
have that. What we might not have is
more
13
anecdotal information about particular ways in
14
which a patient didn't do well.
15
DR. PERRIN: Well, then, maybe you
do have
16 it,
but on what periodicity do you have things like
17 the
HAM-D?
18
DR. LAUGHREN: Every week, you
know, early
19 on
certainly.
20
DR. PERRIN: Then, you may have
the
21
information, okay.
22
DR. RUDORFER: As I understand the
23
situation, Dr. Laughren's Question 3 on patient
24
level data analysis, I think we have been
25
discussing essentially on important covariates that
356
1
should be considered in the re-analysis.
2
Dr. Laughren, would you want us to address
3
anything else specifically on that before we turn
4 to
future directions?
5
DR. LAUGHREN: No, but again let
me just
6
reiterate if committee members, as you continue to
7
look at this list, if you have additional ideas,
8
please feel free even after this meeting to submit
9
them, because we want this to be as comprehensive
10 as
it can be. So, if there are important
11
covariates we have left out, let us know.
12
DR. RUDORFER: Dr. O'Fallon.
13
DR. O'FALLON: Looking at that
list again
14
with fresh eyes after this morning, you don't have
15 any
data that will help you to get at the
16
temporality of the various things.
17
For example, I look at that dose, and you
18 are
looking at the max of the mods, and things like
19
that, but you don't have -- you know, there is no
20 way
in your data set then to get at whether the
21 incidents
occurred when the dose was raised,
22
lowered, or discontinued.
23
So, one of the key questions is not going
24 to
be able to be assessed.
25
DR. LAUGHREN: That is something
that we
357
1
clearly have that information. We
don't have it
2
now, but we could get that information and add it
3 to
the model.
4
DR. RUDORFER: Dr. Katz.
5
DR. KATZ: I have a question of
6
clarification on Question 1, I guess it is, which a
7 lot
of people have been talking about, trying to
8
look at these other behavioral symptoms that are
9 not
explicitly suicide related, like the
10
stimulation syndrome, so called, or an activation
11
syndrome.
12
Again, you have seen what we have done to
13 try
and capture the explicitly suicide related
14
events. You know, we had these
text strings, I
15
think we had 15, we had to go back and forth with
16 the
sponsors and ask them to look at their verbatim
17
terms, you know, that took some time.
But we spent
18 a
lot of time trying to figure out exactly how to
19
ascertain those cases.
20
Is it the committee's desire for us to
21
attempt to recreate that process with regard to
22
this sort of stimulation syndrome, in other words,
23
look for multiple different sorts of terms that
24
might be subsumed reasonably under this syndrome,
25 in
other words, try to cast as broad a net as
358
1
possible?
2
DR. RUDORFER: Yes.
3
DR. KATZ: Is that a general sense
of the
4
committee?
5
DR. RUDORFER: Yes, the sense of the
6
committee is affirmative.
7
Dr. Laughren.
8
DR. LAUGHREN: Bearing in mind
that going
9
back to search the database involves a fair amount
10 of
additional time, now, we could proceed with our
11
analysis based on the data that we have now, and in
12
parallel, go back and ask for additional searches
13 for
other kinds of events like this activation
14
syndrome if it can be better defined.
That is
15 something we clearly could do.
16
I wouldn't want to hold up the suicidality
17
analysis waiting for that additional searching
18
because that does introduce a lot of additional
19
time to go back to companies and ask them to search
20
again.
21
DR. RUDORFER: Dr. Temple.
22
DR. TEMPLE: I just want to be
sure I
23
understand. I think everyone's expectation is that
24
there will be evidence of an activation syndrome or
25
hyperactivity or those things because the drugs are
359
1
labeled to do that.
2
What use would one make out of that if it
3
wasn't linked to some or one of the suicidal terms?
4 I
mean I guess it is more information and that is
5
never bad, but is it more than that, would it help
6 us
understand things?
7
DR. RUDORFER: I think if I may
speak for
8 the
committee, as I understand the discussion and
9 the
concerns, there are two issues.
10
One is that the activation or agitation or
11
akathisia may be what is actually more accessible
12
both to the patient and to the family and to the
13
clinician in terms of it seems, again going back to
14
some of the cases we heard this morning, it sounded
15 as
if we heard more instances of an individual
16
complaining of akathisia-like symptoms as opposed
17 to
volunteering suicidal ideation.
18
I think that there is concern that the
19
akathisia may be what is driving self-destructive
20
behavior at least in some cases, and that might
21
actually be more informative for the clinician to
22 be
watching for than actual more overt suicidality.
23
I also wonder if, in fact, don't we need
24
that information to see if in this database there
25 is
a link.
360
1
DR. TEMPLE: So, you think it would
be
2
useful. I mean obviously if it
sort of went along
3
with suicidal thinking and behavior, that would be
4
certainly of interest as a possible early signal of
5
that consequence.
6
Suppose there isn't any link to suicidal
7
thinking and all you found was a reasonable
8
estimate of the rate of that in a pediatric
9
population, do you think that would be useful all
10 by
itself? You could then say how likely it
is and
11 you
would know that.
12
DR. GOODMAN: I think the way I
would
13
approach it, as you described, as two parallel
14
processes where you continue the work, looking for
15 the
signal and suicidality. You then develop some
16
criteria that help describe this activation
17
syndrome which may occur in a subset of
18
individuals, and then you would test the validity
19 or
clinical meaningfulness of it by then plugging
20 it
back into seeing whether it is those individuals
21
that are more likely to go on to suicide as defined
22 by
the first part of your study.
23
So, I would agree -- one way of saying
24
that -- I agree that for the purposes of our
25
discussion, it would be more of an academic
361
1
exercise and not worthwhile unless we could then
2
find that that subgroup in which there is an
3
activation syndrome are also more likely to go on
4 to
be the ones that were identified as exhibiting
5
suicidal behavior.
6
DR. RUDORFER: Dr. Hudak and then
Dr.
7
Gorman.
8
DR. HUDAK: I have a question and
a few
9
comments.
10
The question involves the quality of the
11 data
that you currently have for analysis.
12
Basically, the 15 studies that are presented here
13
involved 7 drugs, and I am not knowledgeable about
14
these drugs and pharmacological companies, I
15
presume at least 7 drug companies are doing these
16
things. They are using different
protocols, they
17
have different outcome measures, and they have
18
different data acquisition tools, and all those
19
differences, and so forth.
20
The question I have specifically, the
21
information that was presented in Appendix 2,
22
looking at the difference between the "possibly
23
suicide related" versus the "suicide attempts," as
24 I
understand it, that in this population of kids
25 who
might be sick, you are going to have more
362
1
suicide-related type reporting, because that is
2
thoughts and behaviors in excess of suicide
3
attempts, which is just behavior, I mean the data
4
that was presented.
5
Looking at the information here, there are
6 a
number of these studies that basically, within
7
both the drug group and the placebo group, the
8
suicide related thought and behavior is exactly
9
equal to the suicide attempt, which I find
10
inconsistent.
11
Is this the final plumbing of the data, or
12 is
this before the word strings were done on the
13
other data?
14
DR. LAUGHREN: This is one of the
problems
15
that I was alluding to earlier.
When we sent out
16
this request in July of last year, we asked
17
companies to follow basically the same algorithm
18
that Glaxo had used in looking at the Paxil data,
19
which included, first of all, a general search for
20 any
term suggestive of possibly suicide related,
21 and
then an attempt to subgroup patients from that
22
larger set who had any indication of self-harm. I
23
mean that is how it was defined.
24
What we found is that
companies, in
25
carving out that subset of suicide attempt, in some
363
1
cases appeared to count every case as a suicide
2
attempt even though, if you looked at the
3
individual cases, there was not any clear
4
indication of self-harm, and that was one of the
5
reasons why we felt it was very important to have
6
these data completely reclassified by an outside
7
group.
8 Basically, our position is that
neither
9 one
of these categories, either possibly suicide
10
related or suicide attempt, as it has been carved
11 out
and defined by the companies, is particularly
12
meaningful, and that is specifically the reason why
13 we
want to have an outside group look at this broad
14
group of events that were captured as possibly
15
suicide related and help us figure out what kinds
16 of
bins to put those into.
17
As you saw from Dr. Posner's presentation,
18 we
will very likely end up with different
19
categories and different data than what we have
20
here. I mean this table is really
a very
21
preliminary table and we have very little
22
confidence in what these numbers mean because we
23 are
not confident in what the numerators are.
24
DR. HUDAK: I understand, but even
within
25 the
Paxil studies, there are three studies, and two
364
1 of
them show no difference, and one would think
2
that the studies were constructed in somewhat the
3
same way and the query was done in somewhat the
4
same way, and therefore at the end, even going back
5 and
having Columbia group look at this, you may
6
have very imperfect data to look at.
7
DR. LAUGHREN: That is undoubtedly
true,
8 and
that is a problem that we can't fix with these
9
studies. You know, if
ascertainment was poor,
10
there is no way to fix it at this point.
11
DR. HUDAK: I have two additional
12
comments. One is with respect to
this general
13
issue here. I think the big
picture that I take
14
away from this is the really unexplained doubling
15 or
tripling of suicide rates in particularly
16
vulnerable populations that occurred over the past
17 15,
20 years, which is really quite impressive.
18
So, whatever socioenvironmental type
19
etiology there is to this is a very significant
20
public health issue. To put this
sort of into
21
context, this is a doubling or tripling.
When we
22
have a one-point difference in infant mortality, we
23
have major committees sort of looking at why this
24
occurs.
25
Infant mortality over the past 20 years
365
1 has
gone down very substantially, but differences
2 in
infant mortality on the order of 1 in 1,000,
3
which is about 10 percent of the entire infant
4
mortality rate, are treated very significantly.
5 And
this is a huge problem, and I guess with one
6
teenager and one incipient teenager is something
7
that is dear to my concern.
8
The other comment I have is in relation to
9
looking at treatment and the amount of
10
prescriptions that are written, and so forth. I am
11
struck by the fact that we have so much drug
12
prescription done in a population that the efficacy
13 is
not established.
14
I fight that every day in the nursery, to
15
come around and see patients on 10 drugs, of which
16
maybe 2 have been shown to be effective and trying
17 to
withdraw therapy, but it must be -- I have no
18
problem with they are children who are clearly very
19 ill
and anything that can be done should be done,
20 and
I agree with that, but on the other hand, there
21
must be a large population of children -- a lot of
22 the
people who spoke this morning, the picture that
23 was
presented of their child or someone they knew
24 was
not someone who was very, very ill.
25
It was someone who had relatively minor
366
1
type findings, who were put on these drugs with
2
terrible consequences, and I agree with every
3
speaker who said that something needs to be done to
4
educate practitioners and the public that these
5
things may not at all be benign.
6
The fact that we don't find these things
7
that are reported among the audience and the
8
controlled trials is not surprising.
It may be a
9
very, low incidence phenomena that you are not
10
going to find unless you have got randomized
11
controlled trials, you know, 10,000 or more.
12
But each of these events, each of these
13
anecdotes, and I have heard enough of them to think
14
that, you know, you hear enough of these anecdotes,
15
there must be some truth in it. I
mean I am
16
willing to believe that there is an idiosyncratic
17
reaction that some patients have with these drugs,
18 and
I think that warning needs to go out in the
19 very
strongest terms from the Agency as soon as
20
possible.
21
DR. RUDORFER: If we can hear from
Dr.
22
Gorman and Dr. Chesney, please.
23
DR. GORMAN: I would like to pick
up on
24 the
thread of where we are data mining. One
of the
25
things that struck me in one of the slides that was
367
1 put
up was that in August, there was the request
2
from the pharmaceutical companies to relook at
3
their data and present it to the FDA.
4
Within a month, one of the pharmaceutical
5
companies, I am not sure they looked at their data,
6 but
they decided to change their labeling and
7
withdraw it from the market.
8
I would ask the FDA to investigate what
9
signal that pharmaceutical company found in their
10
data that made them want to change their label
11
without going through the FDA, and ask other
12
pharmaceutical companies to look in their data in
13 the
same way.
14
DR. RUDORFER: Dr. Laughren.
15
DR. LAUGHREN: Yes, can I just
respond to
16
that. That company was Wyeth and
the drug is
17
Effexor and Effexor XR. Having
gotten our request
18 in
July, they did go back and look for suicidality,
19 and
they also looked for hostility, and they found
20 a
signal, and on their own, as I explained, they
21 are
allowed to do that on their own if it
22
strengthens labeling under changes being effected.
23
What they did is to add mention of that
24
signal in the Pediatric Use section of their label.
25
They did not contraindicate the drug.
They did
368
1
send a letter out along with that label change
2
recommending that clinicians not use the drug in
3
pediatrics, but the labeling does not in any way
4
contraindicate it. It simply
mentions the signal,
5 and
it is the same signal that we have seen and are
6
currently evaluating.
7
You know, we have their analysis, I showed
8 it
to you, in fact. The question is if you
go back
9 and
do the kinds of work that we are now proposing
10 to
do in terms of looking at the actual events that
11 got
included under those broad categories, what
12
signal will you see.
13
That is really the question, and that is
14 why
we have not acted independently to approve that
15 label
change, but it is basically the same data.
I
16
mean there is nothing we haven't seen.
Again, it
17 is
not as if the drug has been pulled from the
18
market. They have simply added
mention of that
19
signal in one sentence in their label.
20
DR. RUDORFER: Dr. Chesney,
please.
21
DR. CHESNEY: This is in response
to the
22
question from the FDA about why look at activation
23
syndrome if it is not known whether it is directly
24
related to suicidality.
25
But what I heard this morning or the way I
369
1
interpreted what I heard this morning is that the
2
activation syndrome is associated or can be
3
associated with very violent and very hostile
4
behavior. Whether that results in
anybody's death
5 or
not, several of the families said that that
6
became an extremely difficult issue to live with.
7
Where we are dealing with a drug with no
8
apparent benefit, it seems to me that any risk
9
becomes incredibly important, so that is one
10
additional reason that I would say it is important
11 to
look at this activation syndrome that some of us
12
have just learned more about this morning.
13
DR. LAUGHREN: Can I just respond
to that?
14
Again, we are very happy to do that.
It would be
15
extremely helpful if the committee could come up
16
with a little bit more definition of what that is
17 to
help us in searching for it.
18
But independent of finding it in this
19
database, if there is a view that this syndrome is
20 so
well described and does exist, put together the
21
case. Send us literature,
whatever else, and it is
22
possible to make labeling changes about clear
23
events that are idiosyncratic in some way.
24
Again, the problem here has been that the
25
events we are looking at are part and parcel of the
370
1
disease. If there is an
activation syndrome that
2 is
unusual in its nature, and is not part of the
3
disease that is being treated, it could be
4
described in some way in labeling if there is
5
enough even non-controlled data to support the
6
existence of that syndrome, especially if it can be
7
linked to, as you suggest, hostility and violence
8 and
suicidality.
9
DR. RUDORFER: Dr. Trontell.
10
DR. TRONTELL: Thank you. I have a
11
question for Dr. Posner and perhaps other members
12 of
the committee because of looking at your
13
proposed reclassification of the cases.
14
I have a concern, as we have all been
15
discussing, that a very large number of cases may
16
well fall into the indeterminate category using the
17
very clear definitions you laid out for us.
18
Is there any mechanism you can suggest in
19
that category that there might be some
20
classification broadly, you know, low, medium, or
21
high, that might allow some sensitivity analysis?
22
I am a little concerned that data that
23
have been volunteered, you know, since this wasn't
24 a structured
inquiry into potential suicidal
25
behavior, might otherwise be lost.
371
1
DR. POSNER: I think it was
suggested
2
before that we do a level of certainty variability
3 and
analysis, and I think that that is a very good
4
point and something that we will take into account
5
when we are doing those classifications.
6
DR. RUDORFER: Dr. Maldonado is
next,
7
please.
8 DR. MALDONADO: This is a quick question.
9 I
am not trying to generate more work for the
10
people who are doing this work, but I also have the
11
concern that Dr. O'Fallon had, that these data may
12 not
yield what you are looking for.
13
Actually after hearing the comments in the
14
morning of some of the testimonies, it appears that
15
some of these reactions were very similar in adults
16
also, not only in children.
17
I understand that the signal is much less
18
evident and that is probably why adults have been
19
excluded, but since the database in adults, I
20
assume it is much larger and the disease appears to
21 be
less heterogeneous, I don't know if there will
22 be
a value in looking systematically into that data
23 to
see if there is a signal.
24
But again not knowing the data, it may not
25 be
warranted, but that is something that might
372
1
actually help to understand. I am
not talking
2
about only suicides and suicide attempts, I am
3
talking about all the other signals, the wide net
4
that has been proposed here that appears to happen
5
also in adults.
6
DR. RUDORFER: We are going to
hear from
7
Drs. Wang, Leon, and Fost, and then look towards
8 the
future.
9
DR. WANG: I just wanted to follow
up in
10
terms of the utility of studying this
11
akathisia-like symptom. I think
there is actually
12 a
lot of utility particularly if you focus on sort
13 of
the synchrony of change, not just whether there
14 is
a link, but also if there is, you know,
15
presumably this akathisia-like syndrome or
16
activation is just more frequent, so you should
17
have some power to study it, but see if there is a
18
time relationship, because there are so many
19
questions raised about, you know, these potentially
20
abrupt onsets of suicidality after developing some
21
kind of activation-like symptom.
22
Anyway, I would argue that there is some
23
utility in studying it.
24
DR. RUDORFER: Dr. Leon.
25
DR. LEON: A point of
clarification. Dr.
373
1
Laughren said the HAM-Ds or whatever severity
2
rating is available from the trials.
Are those
3
available for each week of the trial or just for
4
endpoint, and are those available at the item
5
level?
6
DR. LAUGHREN: They are available
by week,
7 and
they are available by item level. What I
was
8
pointing out earlier is that companies did try to
9 do
a similar analysis with the suicidality item
10
from the HAM-D, Item 3, similar to what has been
11
done with adults, and it did not generate a signal
12 in
general.
13
DR. LEON: But do they look at the
14
agitation item? I wouldn't expect
the suicide item
15 to
be very sensitive, and I expect it to be even
16
less sensitive in kids who are probably less
17
inclined to disclose their ideation.
18
DR. LAUGHREN: I think we probably
already
19
know that there is an excess of anxiety and
20
agitation both in adults and children with SSRIs.
21
The question is what is it linked to,
and
22
that is why we need help in trying to define the
23
syndrome that everyone is talking about and may
24
well be a real thing, but we already know about
25
agitation by itself.
374
1
DR. LESLIE: I think part of what
you may
2 be
raising, though, is using it as an independent
3
variable, and not as an outcome variable. I mean
4 one
thing would be is this is a sign of increased
5
aggression on the item, on the HAM-D or increased
6
irritability linked then later as an independent
7
variable or a predictor variable, so not as an
8
outcome variable, but as an independent variable.
9
DR. LAUGHREN: We already have
agitation
10 in
the model. That is one of the variables,
11
agitation on drug as opposed to a baseline
12
variable. We have already
included that in the
13
model. So, we should be able to
look at that.
14
The question is are there other things
15
like that, that might be combined in some way to
16
look at as some sort of a stimulation syndrome or
17
activation syndrome other than just agitation by
18
itself.
19
DR. MALONE: Do you have
hyperactivity in
20 the
model?
21
DR. LAUGHREN: I am not sure that
22
hyperactivity is a term that was even coded for. I
23
would have to go back and look at the dictionaries
24 and
see what preferred terms were used.
25
Are you thinking of hyperactivity as a
375
1
term for subsuming other investigator terms or as a
2
descriptive term in itself? I am
not sure what you
3
mean by "hyperactivity."
4
DR. MALONE: Increased motor
activity. In
5
addition to them just being described as agitated,
6 they may be described as having increased
motor
7
activity, sleeplessness, all as part of a syndrome.
8
DR. LAUGHREN: Or restlessness?
9
DR. MALONE: Restlessness, yes.
10
DR. LAUGHREN: Again, to the extent
that
11
committee members can put these thoughts together
12 and
help us identify something to look for, it
13
would be very helpful.
14
It doesn't have to be now. Again,
you can
15
think about this, and if you want to send us your
16
thoughts about this, we will be happy to entertain
17
them. This is the time to do it,
because now is
18 the
time, if we are going to ask for additional
19
variables, now is the time to do it.
20
Dr. Fost and then Dr. Pfeffer.
21
DR. FOST: Thank you. I have some
22
comments that have to do with Questions 5, 6, and
23 7,
and I think they cover all three issues.
24
There have been some comments both in the
25
public session and among the committee and the FDA
376
1
people that there are two problems here.
2
One is the possibility of causing harm to
3
children by prescribing these drugs that may induce
4
suicide, and the other problem is that we may be
5
scaring people away from prescribing them and there
6 may
be inadequate prescribing.
7
That is presented as if they are sort of
8
commensurate or symmetrical, but I think that is
9 not
quite right. There is a reason for the
first
10
principle of first do no harm. It
is almost the
11
whole raison d'etre of the FDA.
12
The reason for that is that it is widely
13
thought that it is more important not to harm
14
people than to fail to help people.
There is an
15
infinite number of people we maybe can help, and we
16
can't do all of it. It is unclear
whether we can
17 do
it, but we know we shouldn't harm people.
That
18 is
our first responsibility.
19
What is odd about this situation is that
20 we
may be doing both. That is, there is not
just
21
concern about causing harm to children, but there
22 is
tremendous ambiguity about whether anyone is
23
being helped.
24
So, as several people have said, if there
25 is
any risk of harm, even if it is a very small
377
1
risk, it is not worth it if there is nothing on the
2
benefit side of the scale.
3
So, it seems to me equally urgent to try
4 to
get some better information about the benefit
5
issue, as well as the harm issue.
6
Now, Bob Temple said that withdrawal
7
studies can't tell us anything about harm, which I
8
agree with, but they can tell us a lot about
9
benefit. In fact, they may be
more powerful than
10
prospective trials in showing benefit.
11
So, it seems to me encouraging, however
12 you
can get it done, getting some withdrawal trials
13 to
occur might take us a long way towards assessing
14 the
benefit issue. That can be done and it
is not
15 all
that expensive to do.
16
That seems to me equally urgent as
17
whatever can be done mining the database to find
18 out
about the harm. So, that is the first
point.
19 I
think both of those are important.
20
Second, in terms of what to do while we
21 are
waiting for these things to happen, while it is
22
correct that this long-standing section of the
23
label that says be especially careful when you
24
start people on treatment can be interpreted to
25
mean they might get worse.
378
1
I don't think an ordinary person, it is
2 all
counterintuitive, but I don't think it occurs
3 to
most parents and maybe not even to doctors who
4
aren't really highly informed about this, that that
5 may
happen, that an antidepressant can make you
6
more depressed or at least more suicidal.
7
I think that word needs to get out as soon
8 as
possible, first, that that is a real
9 possibility,
that the British FDA thinks it is a
10
very real possibility, that the FDA, the American
11 FDA
is very concerned about it, seriously concerned
12 Dr.
Laughren has said several times, that the level
13 of
concern that exists among everybody in this
14
room, public and committee members and FDA, is not
15
adequately out there.
16
For doctors, maybe psychiatrists, I can't
17
speak for them, but I doubt that pediatricians are
18
aware, or family practitioners, the level of
19
concern about this potential problem.
20
So, it seems to me while we are waiting,
21 it
would be very important to get that word out
22
through the AAP and the AAFP, through national
23
meetings, through pediatric news, through
24
newsletters, through panel discussions,
25
presentations at national meetings, and so on, and
379
1
second, to parents, so that when they make what are
2
ideally collaborative decisions with their doctors
3
about whether to put their children on these drugs,
4
they understand completely that there is at least
5
serious concern and that while it is not a settled
6
issue and FDA is looking into it, and you may
7
withdraw the serious concern by the summer, or you
8 may
enhance it, but I don't think that is so
9
terrible to say we are looking at it, it may take
10 us
another 6 or 12 months to figure it out, but
11
while we are waiting, you should be very alert to
12 the
risk of these drugs, you should be very alert
13 to
this activation syndrome in your children, here
14 are
some signs of it.
15
We don't know for sure whether it leads to
16
suicide or not, but there is a lot of smart people
17 who
think it may very well, so you need to be
18
hypervigilant about it.
19
Oh, and a last point. Just to
pick up on
20
something Skip Nelson said a couple of hours ago,
21
there is only one drug that has really been shown
22 to
be effective in children, and while you haven't
23
disproven efficacy, it hasn't been really well
24
established either for all the other drugs, so it
25
seems to me at least part of the education campaign
380
1 to
physicians is if they are going to prescribe
2
anything, why not prescribe the one that we know
3 the
most about and have the most confidence about.
4
That is not to say they may not also cause
5 the
suicidal problem, but at least we have efficacy
6
data for fluoxetine that is stronger than for the
7
other, so why mess around with these other drugs
8 for
which there is less encouraging data on the
9
efficacy side.
10
DR. RUDORFER: Drs. Nelson,
O'Fallon, and
11
Pine, please.
12
DR. NELSON: I want to just make
the
13
observation that that point about fluoxetine
14
complicates how you might then design a trial going
15
forward to look at the efficacy of the other drugs,
16
because you need to evaluate the alternatives that
17 the
child would not be on.
18
So, if you are proposing to start off with
19 an
open-label, non-randomized treatment of a drug
20
that has already been shown to not be effective in
21
your short-term trials, and not put that child on
22
fluoxetine, unless that child is a non-responder or
23 has
had an adverse effect to where you think the
24
profile of the drug you are going to put them on
25
would have some advantage, it is not clear to me
381
1
that that would be a trial that would get through
2
5052 on your IRB in evaluating whether it ought to
3 go
forward.
4
DR. O'FALLON: I recall that Dr.
Murphy
5
told us this morning that FDAMA was needed in order
6 to
basically motivate the drug industry to do the
7
studies of these in the children.
8
When I first went on the subcommittee, I
9 was
appalled to realize that a great many of the
10
doctors feel they pretty much have to prescribe off
11
label because there isn't anything on the label for
12 an
awful lot of different things.
13
So, I think that harm, being able to
14
identify harm in children may actually be more
15
important than being able to identify benefit,
16
simply because the physicians are often having to
17 --
are often having to work off, you know, just try
18 to
figure it out on the fly.
19
So, given that fact, one of the things
20
that really bothers me is the fact that the
21
exclusion criteria are trying to get rid of kids
22 who
are taking more than one drug for whatever
23
reason, but the kids out in the community who are
24
getting it are generally on more than one drug.
25
I think that your future studies have to
382
1
include children who are on other medications, as
2
well. They probably would have to
be stratified
3 and
treated carefully, but you should be getting
4 the
data on adverse events in those populations, as
5
well, because the physicians need to know what bad
6
things can happen.
7
I think placebos are needed because you
8
aren't going to be able to sort out the stuff that
9 is
coming off of the disease from the stuff that is
10
coming off of the treatment if you don't have a
11
placebo for at least some part of the time.
12
So, the forward studies, I mean there are
13 a
lot of things that you have got to do for future
14
studies, but it seems to me you must be looking at
15
these things in multi-polypharmacy, or whatever you
16
call that, group of patients, as well.
17
DR. RUDORFER: Dr. Pine.
18
DR. PINE: I have a couple of
comments in
19
light of a couple of things that have been said
20
over the last few minutes.
21
The first thing is in discussing the data
22 on
efficacy, I think it is important to point out
23 two
things, the first of which is that a number of
24
people have noted that the data are quite
25
discrepant for fluoxetine relative to the other
383
1
SSRIs in pediatric major depression.
2
Non-psychiatrists might not be aware that
3
that is highly unusual. The data
in adults, to the
4
extent that SSRIs have been compared, really do not
5
find that, and I think that one possibility is that
6
kids are very different, and fluoxetine works, and
7 the
other SSRIs don't.
8
Another possibility is that maybe there
9 are
systematic differences in terms of how the
10
studies were done, and I think it is important,
11
particularly from a labeling perspective, not to
12
jump too quickly to say, well, fluoxetine is okay
13 and
nothing else is, number one.
14
Number two, we spent a lot of time talking
15
about the efficacy data for major depression. As
16 was
said in a number of presentations throughout
17 the
morning, that particularly in young children,
18
major depression is not the leading condition for
19
which medications are prescribed, it's anxiety
20
disorders.
21
When one looks at the efficacy data for
22 the
anxiety disorders, for the SSRIs, one gets a
23
very different picture, at least to the extent that
24
those data have been made public and have been
25
published, that the efficacy data really looks much
384
1
stronger there.
2
So, I think again it is very important to
3 not
rush to judgment in terms of saying that SSRIs
4
have no benefits for children who present with
5
various types of psychiatric disorders, because the
6
fact of the matter is that a high proportion of
7
individuals who present with major depression will
8
also have anxiety, and I think it is very important
9 to look
at that issue.
10
Two other quick points. You know,
I think
11
that there are problems with the withdrawal design,
12 and
the FDA mentioned them. Probably the
biggest
13 one
is it doesn't do much for clinicians, for
14 patients,
or for parents to answer the specific
15
question if my child is depressed right now, and
16
they need treatment, is it better to give them an
17
SSRI or not. That is really the
question that we
18
need to answer.
19
The last brief comment, you
know, I know
20 you
guys are asking a lot about could we better
21
define what this activation syndrome is.
Something
22
that we need to consider very carefully is not only
23 is
it known at least among psychiatrists that this
24
syndrome occurs, but usually it is mild.
So,
25
usually, at least to the extent that it has been
385
1
studied in trials, the activation syndrome that
2
occurs is relatively mild.
3
So, to the extent that you are going to
4
look at it, it will be very important to not only
5
assess the type of behaviors that are manifest, but
6 to
all say, well, what is the difference between a
7
mild syndrome which might be relatively common and
8 a
severe syndrome which might be relatively rare.
9
DR. RUDORFER: Dr. Temple, would
you like
10 to
respond to that?
11
DR. TEMPLE: Partly respond to a number
of
12
things that have come up.
Actually, I wanted to
13 ask
Dr. Fost something first.
14
The proposed addition to labeling about
15 the
possibility of an immediate deterioration,
16
would that, in your view, be based on the results
17 of
the controlled trials that we have heard about,
18 or
on the observation from various personal
19
experiences that this seems to occur?
20
I ask that because, as you have heard, the
21
first of them were a little uncertain what it says,
22 and
the second is confounded by the difficulty that
23
some of the consequences that have been described
24 are
potential consequences of the underlying
25
disease, as well.
386
1
That doesn't mean we couldn't say watch
2 out
without necessarily acclaiming the state of the
3
evidence for it. As you pointed
out, we already do
4 say
this is a time to be careful when you start
5
therapy, but I am just interested in what you think
6 the
basis for expanding that would be.
7
DR. FOST: Yes, I think there are
multiple
8
reasons why the FDA called this difficult meeting
9
today, which is very challenging to put together
10 and
very stressful for a lot of people, but there
11 are
several streams of data that I am guessing
12
triggered it.
13
First, there are the data from the trials
14
themselves and the reexamination of it that is
15
going on, and the British conclusions from it, so,
16
first, it is that.
17
Second, it's, as Dr. Hudak pointed out,
18
this epidemic of suicide and what is causing it,
19 and
maybe -- it happens to be concurrent with the
20
rise of SSRIs -- maybe that has got something to do
21
with it.
22
DR. TEMPLE: Wait, you must have
seen
23
different data than what I saw.
What I saw was
24
that in recent years, approximately coinciding with
25 the
SSRIs, the rate of suicide is going down.
I am
387
1 not
saying that proves anything, but I don't see it
2 --
you didn't show it going up.
3
DR. FOST: So be it. The public concern,
4 I
mean the increasing number of anecdotes, I mean
5
obviously, you think that is important or you
6
wouldn't have spent so much time on it listening to
7 it
today.
8
I mean I think there are several things
9
that trigger it, but if nothing else, the data
10
alone, I mean the original trials themselves have
11
stimulated concern among scientific people.
12
DR. TEMPLE: As you heard, we have
13
considerable reservations about what the state of
14 the
trials themselves mean at the moment. I
am not
15
saying this is a bad idea, I am just trying to
16
figure out the basis of it, because if we propose
17
something, we will certainly be asked.
18
DR. FOST: I accept that you are
uncertain
19
about it and that is why you are going to a lot of
20
trouble to look at it much more carefully and in
21
much more detail, but while you are looking, I
22
think sharing this concern, given the seriousness
23 of
it if it turns out that way, is a relatively low
24
cost thing to do.
25
DR. TEMPLE: I just wanted to also
say
388
1
something about randomized withdrawal studies.
2
They are not the whole nine yards obviously.
3
I don't think most people would say that
4 it
is a good state to have only one possible drug.
5
Prozac is a fine drug and everything, but it stays
6
with you more or less permanently, when you stop
7 it,
it is very hard to get off, has a very long
8
half-life with active metabolites.
9
If there were other drugs that were
10
effective, it would be useful to know that. Now,
11 at
the moment, you can't say that there are any
12
other effective drugs.
13
The interest in a randomized withdrawal
14
study is that you take people who, in one way or
15
another, through off-label use, are on a drug
16
already, and you put people into a trial because
17
they seem to be doing well, not because they seem
18 to
be doing badly, and because the current standard
19 of
therapy isn't to keep kids on therapy forever,
20 at
some point you take them off and see how they
21 do.
22
Therefore, a randomized withdrawal study
23
approximates or may approximate clinical practice,
24 and
that would be the case for saying that it's an
25
ethically designed trial.
Obviously, people are
389
1
going to look closely at all this and see if they
2
agree with everything I said.
3
But it can tell you that a drug -- again,
4 you
taper the drug slowly, you don't do an abrupt
5
withdrawal or anything silly like that -- it can
6
tell you I think that the drug was having a
7
favorable effect. It confirms the
clinical
8
observation that led people to keep the patient on
9 the
drug in the first place. So, I wouldn't
rule
10 it
out.
11
DR. RUDORFER: I wonder if I could
12
interject a comment on the labeling.
We have,
13
under Question 5, a quotation from the usual
14
labeling about watching out for the risk of suicide
15
early in treatment.
16
I am thinking, in that small paragraph,
17 the
second sentence reads, "Prescriptions for Drug
18 X
should be written for the smallest quantity of
19
tablets consistent with good patient management, in
20
order to reduce the risk of overdose."
21
I am wondering if that space could be
22
better served. I think that is a
legacy from the
23
tricyclic era and I don't think clinicians today
24
really worry so much about their patients
25
committing suicide by antidepressant overdose.
390
1
I am wondering if instead we had a
2
statement that encompassed two thoughts, one, that
3
patients should be monitored frequently early in
4
treatment, and, two, that any change in behavior,
5
particularly early in treatment, should be reported
6 to
the clinician promptly, to avoid getting into
7
issues of causality, which we have not settled
8
since we don't have all the data yet, but I think
9 --
correct me if I am wrong, committees -- but I
10
think what we are saying is we want to put a speed
11
bump in the road, that, in fact, the sense of the
12
committee is that clinician should take these
13
medications more seriously, and not dispense them
14
overly liberally with inadequate monitoring.
15
I think our state of knowledge is such
16
that we don't have the data we want in terms of
17
showing efficacy and in terms of some of the
18
adverse effects, notably suicidality, obviously,
19
that the analysis is very much underway and we are
20
saying maybe there are other kinds of data to look
21 at,
but I think the concern that many of us felt
22
today was that the way SSRIs and other newer
23
antidepressants are being used now is such that the
24
warnings, as they exist in the current labeling,
25 are
not adequate and/or not being taken seriously.
391
1
My final thought is I wonder if it's time
2 to
reconsider the bolded warning about avoiding
3
combinations with MAO inhibitors, which again I
4
think that is a very important interaction to
5
avoid, but I am not sure how relevant that is to
6
practice today.
7
Dr. Fost.
8
DR. FOST: I just want to add I
think that
9
last sentence adds to the confusion about that
10
paragraph, because the way I read it, frankly, is
11
your patient is depressed, may be suicidal, you
12
have just started him or her on treatment, be
13
careful how many pills you give him because it may
14
take a while for the treatment to kick in and
15
during that time he may take too many of them.
16
It makes it look as if the message is
17
don't give your patient too many pills until he is
18
over the hump, he or she. So, I
agree completely
19
with your sentiment. I mean maybe
that is
20
important, too, but these are not major causes of
21
death, overdose of these pills we have heard.
22
So, it seems to me the more important
23
issue is watch for this other thing where the
24
patient may kill himself in some other way.
25
DR. NELSON: To continue on the
labeling,
392
1
looking through most of the labels, it says simply
2
that efficacy has not been established.
Even
3
though that is a true statement, I think most
4
general physicians and pediatricians have been
5
socialized into thinking that means that the
6
studies have not been done, where the reality here
7 is
they were done and did not show efficacy.
8
So, I would say you need to actually say
9
that, in fact, the studies were done and didn't
10
show efficacy, not that it has not been
11
established, because that is often read as the
12
studies weren't done.
13
DR. RUDORFER: We have time for
Dr.
14
Malone, Dr. Glode, and Dr. Irwin, and if we stay
15
longer than that, we will have to pass the hat for
16
rent, so we may have to wrap up.
17
DR. MALONE: I will just try to be
brief.
18 I
wanted to reiterate what Dr. Pine had said, that
19 a
lot of this discussion is about efficacy in
20
depression, but there is a lot of data about
21
efficacy in anxiety disorders. In
fact, three of
22 the
drugs are labeled I think for OCD, which is an
23
anxiety disorder in children.
24
The second thing is if you are doing a
25
discontinuation study, if the problem is that you
393
1
have such a high placebo response rate that it is
2
hard to separate drug from placebo, and you have a
3 lot
of placebo responders in your study group and
4
then you do the discontinuation, might it be
5
difficult to find an effect.
6
DR. TEMPLE: Can I comment on our
7
experience. That is not our
experience. As Tom
8
said, at least half of all conventional depression
9
trials in adults fail to distinguish drug from
10
placebo. This includes only drugs
we believe are
11
effective because they are successful in other
12
trials.
13
When you do the other, when you do a
14
randomized withdrawal trial, I am aware of only one
15
drug that has ever failed to be successful in that
16
setting. The reasons are fairly
obvious. One, you
17 are
only putting in people who do well. It
is an
18
enriched population for people who are likely to do
19 well.
It is almost -- you know, okay, that's one.
20
The second is that the support system that
21
probably helps the placebo response in the acute
22
episode isn't there here. These
are just people
23 out
in the community, they aren't seeing anybody or
24
chatting with anybody. I mean
they might be, but
25
they are generally not.
394
1
So, the history is that those trials are
2
much more successful, much more at showing
3
effectiveness. Tom can I am sure
elaborate, but I
4
think we have seen only one fail out of a lot.
5
DR. MALONE: I am not sure,
though, that
6 the
placebo response rates are the same in adults
7 as
they are in children. That would be my
only
8
concern.
9
DR. RUDORFER: Dr. Glode.
10
DR. GLODE: I just wanted to add
my
11
support to the recommendations, if I understood
12
them correctly, by Ms. Bronstein and Dr. Fost.
13
I am impressed, if again I have these
14
numbers right, that there were 8 million
15
prescriptions in adolescents for these drugs in
16
2002, so between now and June, let's say another 4
17 or
5 million prescriptions may be written, and
18
these may or may not be for children who were the
19
same as the 3- to 4,000 children with major
20
depression who were studied, again without knowing
21 the
exclusions for all of those studies, if
22
suicidal children were excluded.
23
Then, one comes to the risk of
24
overinforming people because I am going to support
25
additional information to be provided to parents,
395
1
patients, and providers, so that what is the risk
2 of
informing versus the benefit of informing.
3
So, the risk of informing, as mentioned,
4 is
that parents or patients could refuse to take
5 the
medicine that might possibly help them,
6
although again we have the limited efficacy data.
7
The benefit of informing them is that then
8 if
you gave them the right information, they would
9
re-present to their provider when they develop
10
these symptoms and be re-evaluated as opposed to
11
here is your two weeks of samples, you know, I hope
12 you
do well.
13
So, it seems to me that the benefits of
14
informing them probably outweighs the risks of
15
informing them, and my own advice to the FDA would
16 be
to immediately request that information be
17
provided to parents and patients at the time the
18
drug is prescribed. You know,
that just gives them
19
more information about this and ask them to
20
re-present --.
21
DR. RUDORFER: Dr. Irwin.
22
DR. IRWIN: I would argue that the
23
patients may be ahead of the curve than the
24
clinicians are, and I am a person who specializes
25 in
caring for adolescents, I run a large adolescent
396
1
medicine program at the University of
2
California/San Francisco.
3
I would argue that most of the
4 pediatricians
who prescribe these agents are not as
5
familiar as the psychiatrists are about the side
6
effects. I think in the way that
pediatricians --
7
when I was in training, you know, you treated
8
everybody that walked through the door who had a
9 red
ear -- now, we don't do that. We
basically do
10 a
lot of watchful waiting.
11
What I heard today from patients and
12
parents, as they stood up and talked about issues,
13
that many of them went to primary care physicians,
14 and
there was not any watchful waiting, in fact,
15
there was immediate response, and the immediate
16
response was based upon I think inadequate
17
information that is going to clinicians who are
18
acting in good faith and really committed to
19
improving the lives of young people, of which,
20
known in an adolescent medicine clinic, a primary
21
care clinic, about 1 in 5 kids that walk through
22 the
door have a behavioral disorder, so you are
23
really confronted with a big problem.
24
So, I think it is imperative I would say
25
that the FDA get something out to clinicians as
397
1
quickly as possible, and it can be done through a
2
variety of ways that have been mentioned here,
3
because I think those are the individuals that are
4
really acting in ways that we need to really try to
5
encourage them to be acting in a more responsible
6
manner when we are coming up with what really the
7
issues are.
8
Thanks.
9
DR. RUDORFER: Dr. Leslie, do you
have a
10
word, and the we will wrap up.
11
DR. LESLIE: I wanted to echo what
Dr.
12
Irwin was saying as a fellow pediatrician, and also
13
comment that one of the large pressures that many
14 of
us in primary care are under is that we cannot
15
access other types of mental health services.
16
There aren't mental health providers to see kids or
17
they are not able to get services through managed
18
care.
19
So, many primary care providers are trying
20 to
do what they can to help families and children
21 by
giving these medications. So, the other
thing
22 we
need to do -- and I am not sure what the role of
23 the
FDA in this is -- demand parity for mental
24
health services.
25
DR. RUDORFER: Thank you. I think we have
398
1
been identifying some very crucial issues. As Dr.
2
Laughren pointed out in his handout, the FDA does
3 not
control the practice of medicine, so that we
4
here have under the FDA's jurisdiction a limited
5 part of the overall scheme.
6
Nonetheless, I think the sense of the
7
committee is that the FDA has a very important role
8 to
play, and this challenge is an opportunity to
9
further protect the health of young people with
10
depression while the further studies we discussed
11
proceed.
12
If I can sum up the sense of the
13
committee, I think I have 18 seconds, I can distil
14
this to two major bullets.
15
First, we concur with the plan to have the
16
expert group at Columbia re-analyze the data from
17 the
efficacy trials that were presented and some
18
ideas were offered.
19
We could do this in a more formal way in
20
terms of other covariates, issues, such as family
21
history, the activation or overstimulation,
22
restlessness, akathisia spectrum, we discussed as
23
useful information to have.
24
It will be particularly helpful if it is
25
linked with the suicidality measures, but we think
399
1
nonetheless that is important to have established.
2
Correct me if I am wrong, committees, but
3 I
think our sense is that we would like in the
4
interim the FDA to go ahead and issue stronger
5
warning indications to clinicians regarding
6
possible risks of these medications, which we don't
7 see
as contraindicating their use, but we think
8
such warnings are required to elevate the level of
9
concern and attention that practitioners use in
10
prescribing them.
11
I think, as a group, we were recognizing
12 the
limitations of uncontrolled data. We
were all
13
concerned about the stories we heard of the actual
14 use
of these very powerful, potentially very
15
effective medications, but in many instances, being
16
used without adequate monitoring.
17
DR. TEMPLE: I would just add to
your
18
summary, information to physicians and to parents.
19
DR. RUDORFER: Thank you. I would now
20
like to turn the mike over to Dr. Chesney
21
representing the Pediatric Drug Subcommittee.
22
DR. CHESNEY: I just wanted to
thank the
23 FDA
for bringing this issue to all of us and for
24
being so open and listening and for asking us to
25
continue to provide them with additional
400
1
information.
2 I think it really brings home to all of
us
3 the
importance of looking at all drugs very
4
carefully in children. I also,
again on behalf of
5 the
Pediatric Committee want to thank all the
6
parents and children and individuals who came to
7
share their experiences with us today.
8
DR. RUDORFER: Dr. Katz.
9
DR. KATZ: I would like to thank
very much
10 the
committee. I think this is a very
complicated
11 and
important issue and through all of that, I
12
think ultimately, your recommendations have been
13
very clear, and I think we have a very good
14
understanding of what you think we should do and
15 how
we should proceed at this point.
16
I also would like to thank the families
17 for
coming forward and telling us your stories.
18
That was courageous and we know it was painful, but
19 I
believe we heard you, I believe the committee
20
heard you, and we appreciate it very, very much.
21 DR. RUDORFER: In closing, I would like to
22
thank the members of the two committees, I would
23
like to thank the FDA staff. It
is obvious what
24
time, effort, and hard work has gone into this
25
important issue, we appreciate that, and I want to
401
1
thank everyone in the audience who came,
2
particularly people who told us their painful
3
stories.
4
The FDA staff can attest to the fact I
5
kept arguing about the time limit.
I am sorry, but
6 we
would probably still be in the open public
7
hearing if we didn't have that red light.
8
Thanks all for coming and obviously, this
9
discussion is to be continued.
10
Get home safely.
11
[Whereupon, at 6:05 p.m., the meeting was
12
adjourned.]
13 - - -