1
DEPARTMENT OF HEALTH AND HUMAN
SERVICES
FOOD AND DRUG
ADMINISTRATION
CENTER FOR BIOLOGICS EVALUATION
AND RESEARCH
DRUG SAFETY AND RISK MANAGEMENT
ADVISORY COMMITTEE
IN JOINT SESSION WITH
THE
DERMATOLOGIC AND OPHTHALMIC
DRUGS
ADVISORY COMMITTEE
Hilton
2
PARTICIPANTS
DRUG SAFETY AND RISK MANAGEMENT ADVISORY
COMMITTEE:
Peter A. Gross M.D., Chairman
Michael R. Cohen, R.Ph., M.S., D.Sc.
Stephanie Y. Crawford, Ph.D., M.P.H.
Ruth S. Day, Ph.D.
Jacqueline S. Gardner,
Ph.D., M.P.H.
Arthur A. Levin, M.P.H.
Robyn S. Shapiro, J.D.
Brian L. Strom, M.D.,
M.P.H.
DERMATOLOGIC AND OPHTHALMIC
DRUGS ADVISORY
COMMITTEE:
Roselyn E. Epps, M.D.
Robert Katz, M.D.
Paula Knudson, Consumer Representative
Sharon S. Raimer, M.D.
Eileen W. Ringel, M.D.
Kathleen Y. Sawada, M.D.
Jimmy D. Schmidt, M.D.
Elizabeth S. Whitmore, M.D.
Michael G. Wilkerson, M.D.
CONSULTANTS (Voting):
Wilma F. Bergfeld, M.D.
Michael E. Bigby, M.D.
Margaret Honein, Ph.D.,
M.P.H.
Arthur H. Kibbe, Ph.D.
Sarah Sellers, Pharm.D.
Amarilys Vega, M.D., Ph.D.
Jurgen Venitz, M.D.,
Ph.D.
GUEST SPEAKER (Non-Voting):
Richard K. Miller, Ph.D.
3
PARTICIPANTS
(Continued)
FDA STAFF:
Jonca Bull, M.D.
Steven Galson, M.D., M.P.H.
John Jenkins, M.D.
Sandra Kweder, M.D.
Paul Seligman, M.D., M.P.H.
Anne Trontell, M.D., M.P.H.
Jonathan Wilkin, M.D.
4
C O N T E N T S
PAGE
Call to Order and Introductions, Peter
Gross, M.D.5
Conflict of Interest Statement,
Shalini Jain, PA-C, M.B.A., Executive
Secretary 7
Effectiveness of the Isotretinoin Risk
Management
Program for the Prevention of Fetal
Exposure to
Accutane and its Generic Equivalents and
Consideration of whether Changes
to this Isotretinoin Risk Management
Program would
be Appropriate:
Charge to the Committees, Steven Galson,
M.D.,
M.P.H., Acting Director, CDER 12
Background and Regulatory History,
Jill Lindstrom, M.D., Division of
Dermatologic
and Dental Drug Products, FDA 15
Questions to the Speaker from
Committee 49
Open Public Hearing:
Robert A. Silverman, M.D. 68
Sidney Wolfe, M.D.,
Public Citizen Research Group 74
Curt D. Furberg, M.D., Ph.D. (Letter
Read by
Dr. Sherri Shubin, M.D., MPH 83
Hoffmann-La Roche, Inc. Presentations:
Introduction, Joanna Waugh, Group
Director,
Regulatory Affairs 90
Benefit/Risk, Martin H. Huber, Vice
President,
Global Head Drug Safety Risk
Management 94
Regulatory Overview, Joanna Waugh 96
5
C O N T E N T S
(continued)
PAGE
Overview of the S.M.A.R.T. Program,
Susan Ackermann Shiff, Ph.D., Global
Head Risk
Management, Drug Safety Risk
Management 101
Evaluation of S.M.A.R.T. Program,
Martin H. Huber, M.D., Vice President,
Global Head Drug Safety Risk
Management 116
Generic Firms' Presentations:
Isotretinoin Risk Management Program,
Background
Information, Frank R. Sisto, Vice
President,
Corporate Regulatory Affairs,
Mylan Laboratory, Inc. 140
Isotretinoin Survey, Allen A.
Mitchell, M.D.
Slone Epidemiology Center, Boston
University 152
Isotretinoin Enhanced Risk Management
Program,
Program Elements for which Advisory
Committee
Input is Requested, Robert W. Pollock,
Vice
President, Lachman Consultant
Services, Inc. 169
Questions to Roche and Generic Firms from
Committee
174
Isotretinoin Pregnancy Exposure:
Spontaneous
Reports 1 Year Pre- and 1 Year Post-Risk
Management Program,
Marilyn Pitts, Pharm.D.,
Office of Drug Safety, FDA 218
Isotretinoin Pregnancy Prevention Program
Evaluation,
Allen Brinker, M.D., M.S.,
Office of Drug Safety, FDA 237
Kaiser Presentation, Richard A. Wagner,
Pharm.D.,
Kaiser Permanente Drug Use
Management 265
Questions to Kaiser from the
Committee 289
6
C O N T E N T S
(Continued)
PAGE
Organization of Teratology Information
Services,
Interim Report, North American
Isotretinoin
Information and Survey Line, Richard Miller,
Ph.D., University of Rochester 296
Questions to OTIS from the Committee 313
Risk Management Options for Pregnancy
Prevention,
Kathleen Uhl, M.D., Pregnancy Labeling
Team, FDA 321
Selecting Risk Management Tools:
Considerations and
Experience, Anne Trontell, M.D.,
M.P.H. Deputy
Director, Office of Drug Safety,
FDA 338
Questions to Speakers from the
Committee 366
7
1 P R O C E E D I N G S
2 Call to Order and
Introductions
3
DR. GROSS: Good morning. I am Dr. Peter
4
Gross. I am Chair of the Drug
Safety and Risk
5
Management Advisory Committee. I
would like to
6
thank you all for coming this morning, and the
7
first order of business is for us to go around the
8
room and introduce everybody at the table. So, I
9 am
Dr. Peter Gross. I am Chair of the
Department
10 of
Internal Medicine at Hackensack University
11
Medical Center and New Jersey Medical School.
12
MS. JAIN: Shalini Jain, Executive
13
Secretary, FDA, Center for Drug Evaluation and
14
Research.
15
DR. WILKERSON: Michael Wilkerson,
MD.,
16
private practice, Tulsa, Oklahoma.
17
DR. RINGEL: Eileen Ringel, I am
in
18
private practice in Waterville, Maine.
19
DR. DAY: Ruth Day, I direct the
Medical
20
Cognition Laboratory at Duke University and I am on
21 the
Drug Safety and Risk Management Committee.
22
DR. KIBBE: Art Kibbe, Chairman of
the
8
1
Pharmaceutical Sciences Department, Wilkes
2
University School of Pharmacy and Chairman of the
3
Pharmaceutical Sciences Advisory Committee to the
4
FDA.
5
DR. GARDNER: Jackie Gardner,
Professor of
6
Pharmacy, University of Washington, and Drug Safety
7 and
Risk Management Advisory Committee.
8
DR. KATZ: Robert Katz, I am in
private
9
practice in Rockville, Maryland, and Clinical
10 Assistant Professor of Dermatology at
Georgetown
11
University.
12
DR. SELLERS: Sarah Sellers,
Pharm.D. I am
13 a
Masters in Public Health Candidate at Bloomberg
14
School of Public Health.
15
DR. TRONTELL: Anne Trontell,
Deputy
16
Director of the Office of Drug Safety in the FDA
17
Center for Drugs.
18
DR. SELIGMAN: Paul Seligman,
Director of
19 the
Office of Pharmacoepidemiology and Statistical
20
Science, also in the Center for Drugs at the FDA.
21
DR. WILKIN: Jonathan Wilkin,
Director of
22 the
Division of Dermatologic and Dental Drug
9
1
Products in CDER, FDA.
2
DR. BULL: Good morning. Jonca Bull,
3
Director, Office of Drug Evaluation V in the Office
4 of
New Drugs, Center for Drug Evaluation and
5
Research.
6
DR. KWEDER: Sandra Kweder, Deputy
7
Director of Office of New Drugs in CDER.
8
DR. GALSON: Steve Galson, I am
the Acting
9
Director of the Center for Drug Evaluation and
10
Research.
11
MR. LEVIN: Art Levin, I am the
consumer
12
representative on the Drug Safety Committee.
13
DR. SAWADA: Kathleen Sawada,
14
dermatologist, private practice in Lakewood,
15
Colorado.
16
DR. VENITZ: Jurgen Venitz,
Associate
17
Professor, Virginia Commonwealth University and
18
Chair of the Clinical Pharmacology Subcommittee.
19
DR. STROM: Brian Strom, I am
Chair of the
20
Department of Biostatistics and Epidemiology at the
21
University of Pennsylvania School of Medicine, and
22 I
am a member of the Drug Safety and Risk
10
1
Management Committee.
2
DR. BERGFELD: I am Wilma
Bergfeld,
3
dermatologist and dermatopathologist, head of
4
Clinical Research Department of Dermatology at the
5
Cleveland Clinic.
6
DR. RAIMER: Sharon Raimer,
Chairman of
7
Dermatology at the University of Texas in
8
Galveston.
9
MS. KNUDSON: Paula Knudson, I am
the IRB
10
administrator for the University of Texas at
11
Houston, and I am with the Dermatology Advisory
12
Committee.
13
DR. BIGBY: I am Michael
Bigby. I am a
14
dermatologist at Beth Israel Deaconess Medical
15
Center and Harvard Medical School.
16
DR. HONEIN: I am Peggy Honein. I am an
17
epidemiologist with the Birth Defects Center at the
18
Centers for Disease Control and Prevention.
19
DR. COHEN: Mike Cohen, I am a
pharmacist
20
with the Institute for Safe Medication Practices,
21 and
I am with the Drug Safety and Risk Management
22
Advisory Committee.
11
1
DR. WHITMORE: Beth Whitmore, I am
in
2
private practice in Wheaton, Illinois.
3
DR. SHAPIRO: Robyn Shapiro, I am
4
Professor and Director of the Center for the Study
5 of
Bioethics at the Medical College of Wisconsin,
6 and
I am on the Drug Safety and Risk Management
7
Advisory Committee.
8
DR. EPPS: Roselyn Epps, Chief of
the
9
Division of Dermatology in Children's National
10
Medical Center, and also a member of the
11
Dermatologic and Ophthalmic Drugs Advisory
12
Committee.
13
DR. SCHMIDT: I am Jimmy Schmidt,
in
14
clinical practice from Houston, Texas and I am on
15 the
clinical faculty of University of Texas and
16
Baylor Medical School.
17
DR. CRAWFORD: Good morning. Stephanie
18
Crawford, Associate Professor, University of
19
Illinois at Chicago College of Pharmacy, and I am a
20
member of the Drug Safety and Risk Management
21
Advisory Committee.
22
DR. GROSS: Thank you all, and now
I would
12
1
like to ask Shalini Jain to read the conflict of
2
interest statement.
3 Conflict of Interest
Statement
4
MS. JAIN: The following statement
5
addresses the issue of conflict of interest with
6
respect to this meeting, and is made a part of the
7
record to preclude even the appearance of such at
8
this meeting.
9
The topics to be discussed at today's
10
meeting are matters of broad applicability. Unlike
11 issues
before a committee in which a particular
12
sponsor's product is discussed, issues of broad
13
applicability involve many sponsors and their
14
products. All FDA participants
have been screened
15 for
their financial interests as they may apply to
16 the
products and companies that could be affected
17 by
the committee's discussions.
18
Based on this review, it has been
19
determined that there is no potential for an actual
20 or
apparent conflict of interest at this meeting,
21
with the following exception: In
accordance with
22 18
U.S.C. 208(b)(3), Dr. Ruth Day has been granted
13
1 a
waiver that permits her to participate fully.
2
A copy of the waiver statement maybe
3
obtained by submitting a request to the Food and
4
Drug Administration's Office of Management
5
Programs, Division of Freedom of Information,
6
HF1-35 5600 Fishers Lane, Rockville, Maryland
7
20857.
8
Because issues of broad applicability
9
involve many sponsors and their products, it is not
10
prudent to recite all potential conflicts of
11
interest as they may apply to each member,
12
consultant and guest speaker. In
addition, there
13
will be no industry representatives at today's
14
meeting. As you may be aware, the
Food and Drug
15
Administration has appointed industry
16
representatives that currently serve on each of
17
these committees but Annette Stemhagen, Dr.PH., the
18
industry representative to the Drug Safety and Risk
19
Management Committee, and Peter Kresel, M.B.A., the
20
industry representative to the Dermatologic and
21
Ophthalmic Drugs Advisory Committee, work with
22
sponsors that are directly impacted by the matters
14
1
before the committee. FDA has
contacted three
2
industry representatives from other Center for Drug
3
Evaluation and Research committees that have
4
experience with risk management issues and with FDA
5
advisory committee processes.
However, none were
6
available to participate in this meeting. Dr.
7
Stemhagen and Mr. Kresel are present in the
8
audience and attending as interested observers.
9
Further, we would like to note that Dr.
10
Louis Morris, a member of the Drug Safety and Risk
11
Management Committee, has been recused from
12
participating in today's meeting.
Dr. Morris is
13
also present in the audience and attending as an
14
interested observer.
15
We would like to remind the FDA
16
participants not to discuss the issues at hand
17
outside the advisory committee meeting.
In the
18
event that the discussions involve any other
19
products or firms not already on the agenda for
20
which FDA participants have a financial interest,
21 the
participant's involvement and exclusion will be
22
noted for the record. With
respect to all other
15
1
meeting participants, we ask in the interest of
2
fairness that they address any current or previous
3
financial involvement with any firm whose product
4
they wish to comment upon. Thank
you.
5
DR. GROSS: Thank you. The topic for
6
discussion for the next two days is the
7
effectiveness of the isotretinoin risk management
8 program for the prevention of fetal exposure
to
9
Accutane and its generic equivalents, and to
10
consider whether changes to this risk management
11
program would be appropriate. Dr.
Steven Galson
12
will give our committees the charge.
He is Acting
13
Director of the Center for Drug Evaluation and
14
Research.
15 Charge to the Committees
16
DR. GALSON: Thank you very much,
Dr.
17
Gross. I want to thank all of the
committee
18
members for being here. Your
commitment to public
19
service, indicated by the time commitment that you
20
have agreed to make to this subject, is extremely
21
important for the Food and Drug Administration and,
22
indeed, very important for all the patients taking
16
1
this drug and our decision-making process.
2
Today and tomorrow you are going to hear
3
details about the regulatory history of
4 isotretinoin. You are going to review data that
5 has
been collected over the last few years about
6 the
Pregnancy Prevention Program, and you are going
7 to
help us by giving us advice about where this
8
program should go in the future.
These
9
perspectives are extremely important to us. We can
10
spend a lot of time talking to each other and
11
tossing ideas around about what is the best course
12 of
action but when we have outside observers who
13
have taken a fresh look at these programs it is
14
enormously helpful to us as we move down the path
15 to
make decisions.
16
Isotretinoin has been on the market for
17
about 22 years and it may take the record for the
18
single drug with the most advisory committee
19
meetings. I don't know if that is
true but it is
20
certainly very close. When Roche
established the
21
current S.M.A.R.T. program in consultation with the
22 FDA
in 2001, the agency established several goals
17
1 for
the program. They were that no person
should
2
begin isotretinoin therapy if pregnant and that no
3
pregnancy should occur while a woman is taking
4
isotretinoin.
5
I want to just note that although those
6
were the goals, the agency is very cognizant of the
7
fact that setting a zero goal as a metric for
8
something that really depends on human behavior for
9
success and is probably not possible to attain. It
10 is
good to set that goal but when these issues are
11
totally out of the control of manufacturers,
12
physicians or the agency it is really impossible to
13
actually meet that, and we have been criticized for
14
saying our goal is zero. I want
to make it clear
15
that we recognize that it is probably not
16
attainable but we still think it is important to
17 set
these important goals because it helps us set
18 the
stage for figuring out what steps we want to
19
take and we think that is very important.
20
Setting these goals and establishing
21
metrics to get there is very consistent with one of
22 the
evolving foundations of CDER's risk management
18
1
program which is that risk management programs must
2 be
periodically evaluated for effectiveness.
3
Efficiency in risk management is very important
4
and, without measuring the effectiveness of the
5
program and knowing whether we are getting adequate
6
preventive power for the resources devoted we
7
really don't know where to go in the future with
8
this program, and it doesn't help us in terms of
9 establishing
and setting up new programs for
10
additional drugs.
11
Manufacturers of isotretinoin have been
12
challenged by the agency to work together to
13
minimize adverse events related to this drug, and
14 we
are really extremely heartened by the degree of
15
collaboration that has taken place to date and by
16 the
way the manufacturers are working together to
17
look towards the future. We
really expect this
18
collaboration to continue and we think that the
19
goal of minimized the teratogenic risk of this drug
20 is
something that we all share with all the
21
manufacturers and we, again, want to congratulate
22 and
are very heartened by the degree to which these
19
1
groups have been working together.
We look forward
2 to
hearing about how the S.M.A.R.T. program has
3
worked and how the companies have been working in
4
detail together.
5
I want to just talk about the
committee
6
now. We ask you to really remain
focused on the
7
purpose of this meeting, the risk management
8
program for the prevention of fetal exposure. We
9 are
aware that there are other important safety
10
issues related to this drug but we really are going
11 to
focus on prevention of fetal exposure in this
12
meeting. We would like you to
consider the data
13
presented. We want you to
consider the past risk
14
management programs and their achievements, and we
15 are
really looking forward to your recommendations
16 as
to whether the program, as it now exists, should
17
continue; whether it is as effective as it could
18 be;
and how we should enhance it or establish new
19 or
different tools. So, with that I will
close and
20
pass it back to the Chair. Thank
you very much.
21 We
are looking forward to a great meeting.
22
DR. GROSS: Thank you, Dr.
Galson. You
20
1 are
keeping us on time, setting a high target.
The
2
next speaker is Jill Lindstrom, a medical officer
3 for
the Division of Dermatologic and Dental Drug
4
Products at the FDA, who will talk about the
5
background and regulatory history of this
6
medication.
7 Background and Regulatory
History
8
DR. LINDSTROM: Good morning.
9
[Slide]
10
My objectives this morning are to set for
11 you
a clinical context for the use of isotretinoin;
12 to
outline the history of risk management efforts
13 for
this drug; to describe the current risk
14
management plan in some detail; and to provide the
15
committee with some rough guidelines for their
16
assessment of the data that will be presented.
17
[Slide]
18
Isotretinoin is an oral retinoid that is
19
indicated for the treatment of severe recalcitrant
20
nodulocystic acne. It is the only
drug moiety
21
approved for this indication, although there are
22
other oral related products in development. The
21
1
innovator was approved in 1982 and three generic
2
products have recently entered the market.
3
[Slide]
4
This patient has nodular acne, a
5
devastating disease that can result in significant
6
scarring and permanent disfigurement.
You can see
7 that he has many lesions, to include large
8
fluctuant nodules on his forehead, his cheeks, his
9
chin and his nose.
10
[Slide]
11
This patient also has nodular acne and,
12
again, you can see the many lesions on his face,
13 the
large fluctuant nodules extending down onto his
14
trunk.
15
[Slide]
16
This is the same patient, a view of his
17
back.
18
[Slide]
19
Again, a view of that patient's face prior
20 to
isotretinoin therapy--
21
[Slide]
22
--and following conclusion of a course of
22
1
isotretinoin therapy--he is dramatically improved.
2
[Slide]
3
And a third clinical example of a patient
4
with severe nodular acne. Again,
you can see the
5
nodules, sinus track formation and scarring. This
6 is
the patient prior to a course of isotretinoin
7 therapy--
8
[Slide]
9
--and at completion of his course of
10
therapy.
11
[Slide]
12
Because of its unique effectiveness,
13
current practice standards have expanded the use of
14
isotretinoin to the setting of non-nodular but
15
still scarring acne.
16
[Slide]
17
This patient does not have nodules, does
18 not
have classic nodular acne. She has
severe
19
papulopustular acne and her disease is scarring.
20 You
can also imagine that, in addition to the
21
cutaneous morbidity, she has significant
22
psychosocial morbidity from her disease.
This is
23
1 her
presentation prior to treatment with
2
isotretinoin--
3
[Slide]
4
--and her result at conclusion of therapy.
5
[Slide]
6
And a second patient, again without
7
nodular acne but with severe scarring papular acne.
8
This is a front view--
9
[Slide]
10
--and a side view prior to treatment with
11
isotretinoin--
12
[Slide]
13
--and the patient's result at conclusion
14 of
therapy, again dramatically improved.
15
[Slide]
16
Now, isotretinoin is unique among the
17
therapies in the acne armamentarium in that it
18
addresses all four of the known pathogenetic
19
mechanisms of acne. It decreases
sebum production
20 and
shrinks the size of the sebaceous glands.
It
21
normalizes follicular hyperkeratinization and
22
reduces follicular plugging. It
decreases P. acnes
24
1 colonization, although not through a direct
2
antibacterial mechanism but probably through making
3 the
micro climate of the follicle inhospitable to
4 the
organism. Finally, it is mildly
5
anti-inflammatory.
6
[Slide]
7
These events can be seen in this
8
histological specimen, this biopsy of a comedo
9
prior to isotretinoin therapy.
You can see the
10
dilated follicle filled with keratinous debris, the
11
large sebaceous glands. Not well
appreciated in
12 the
black and white photograph is the
13
perifollicular inflammation and the numerous
14
bacteria in the follicle.
15
[Slide]
16
In a biopsy of a follicle following
17
isotretinoin therapy the sebaceous glands--again, I
18
regret that I don't have a pointer but the
19
sebaceous glands are much smaller in size; the
20
follicular lumen is narrow. There
is no follicular
21
plugging and there is an absence of perifollicular
22
inflammation.
25
1
[Slide]
2
Isotretinoin is also unique in that a
3
course of therapy is temporally circumscribed.
4
Other anti-acne agents have no long-term impact and
5 are
effective only while they are being used.
A
6
course of isotretinoin, however, can result in
7
complete and prolonged disease remission. Thus,
8
patients with severe scarring acne like the
9
clinical examples that I just showed you prior to
10 the
approval of isotretinoin would have faced
11
years, perhaps even decades, of therapy with oral
12
antibiotics in combination with topical agents.
13 Now
such patients, after a course of isotretinoin
14
therapy, will see their disease become quiescent
15 and
the progression of their disfigurement halted,
16 and
they are spared the risk, the expense and the
17
inconvenience of years of oral and topical
18
therapies.
19 [Slide]
20
However, isotretinoin does present its own
21
risks. It is a known human
teratogen. In utero
22
exposure to isotretinoin can result in an increased
26
1
risk of spontaneous abortion and premature births,
2 as
well as structural abnormalities.
Approximately
3 28
percent of exposed fetuses will have sufficient
4
stigmata at the time of birth to be diagnosed with
5
retinoid embryopathy.
Additionally, many babies
6 who
are exposed to isotretinoin in utero will
7
appear normal at birth and will go on later in life
8 to
manifest neurodevelopmental deficits.
9
[Slide]
10
What has been done to manage this risk?
11 At
the time of approval in 1982 it was understood
12
from animal data that isotretinoin was likely a
13
teratogen, and in labeling the drug was classified
14
pregnancy category X. Prescribers
and patients
15 were advised in the contraindications,
warnings and
16
precautions sections of labeling not to become
17
pregnant while using the drug.
18
[Slide]
19
The first report of a human malformation
20
following in utero exposure to isotretinoin was
21
published in 1983. In response,
red warning
22
stickers were distributed to pharmacies to be
27
1
affixed to each isotretinoin prescription that was
2
dispensed. Additional reports of
exposed
3
pregnancies were received raising the concern both
4 in
the agency and the manufacturer.
Multiple "dear
5
doctor" letters were issued to inform the medical
6
community of this risk and the label was revised as
7
information became available.
8
[Slide]
9
In 1988 the sponsor proposed a
10
multi-tiered program to augment the risk management
11
plan which they entitled the Pregnancy Prevention
12
Program. An advisory committee
was convened to
13
review this proposal. There were,
as I said,
14
multiple components. First, the
label was altered
15 to
include warnings printed directly on the
16
package, and the "avoid pregnancy" icon was
17
introduced, the familiar red circle with the slash
18 and
the pregnant figure. And, the packaging
was
19
changed to blister packaging.
20
[Slide]
21
The package insert was updated to include
22 a
boxed warning informing physicians and patients
28
1 of
a need for a negative pregnancy test seven days
2
before treatment initiation; the importance of
3
using two reliable forms of contraception; waiting
4 to
begin therapy until the second or third day of
5 the
next menses; and limiting the supply dispensed
6 to
30 days; and the importance of repeating
7
pregnancy testing and contraceptive counseling on a
8
monthly basis.
9
[Slide]
10
An informed consent form for females was
11
introduced in that program. A kit
for prescribers
12 was
provided to explain the details of the program,
13 and
the first iteration of the voluntary patient
14
survey was introduced at that time.
Additionally,
15
there was a tracking survey to assess prescriber
16 use
of the program. That advisory committee
17
recommended approval of the Pregnancy Prevention
18 Program
and the program was implemented in 1989.
19
[Slide]
20
What was the impact of the program?
It is
21
somewhat difficult to say. From
the time of
22
approval of isotretinoin in 1982 pregnancies have
29
1
been reported to the agency. At
the time of the
2
introduction of the Pregnancy Prevention Program we
3
gained a new tool to gather information about
4
pregnancy reports, the patient survey.
Those
5
pregnancy reports are represented by the light blue
6
bars from 1989 on.
7
Both of these reporting mechanisms,
8
spontaneous reports as well as reports through the
9
survey, are voluntary reporting mechanisms and so
10 it
is difficult to ascertain an accurate pregnancy
11
rate. I want to remind you that
this is a
12
historical view prior to the implementation of the
13
current risk management program, but what we can
14 say
is that the public health burden from exposed
15
pregnancies continued to be large.
16
[Slide]
17
Additionally, during this time or during
18 the
'90s Accutane use was increasing significantly.
19
Because of these reasons, the large public health
20
burden from exposed pregnancies as well as the
21
increasing use, an advisory committee was convened
22
again to consider augmentation of the risk
30
1
management plan.
2
[Slide]
3
This advisory committee was convened in
4
September of 2000 and they determined that there
5
was, indeed, a compelling need for augmentation of
6 the
risk management plan. The agency agreed
and
7
this was communicated to the sponsor in a letter
8
dated October 6, 2000. This
letter has been
9
included in the briefing package for the committee.
10
[Slide]
11
In this letter risk management is
12
addressed from two perspectives, both pregnancy
13
prevention and potential neuropsychiatric adverse
14
events. Pregnancy prevention is
the focus of this
15
advisory committee. However,
since the letter was
16
included in your packet and does address
17
neuropsychiatric risk management I want to briefly
18
update the committee on the status of risk
19
management efforts with regards to potential
20
neuropsychiatric risk.
21
[Slide]
22
Three points of action were
recommended by
31
1 the
committee and communicated in that letter.
2
First, that the informed consent be amended to
3
inform patients of the potential for
4
neuropsychiatric adverse events, and this has been
5
done. Second, it was advised that
an educational
6
program for prescribers be implemented, and this
7 has
also been done. Third, it was
recommended that
8 a
comprehensive research program be undertaken to
9
include clinical trials.
10
The sponsor submitted clinical protocols
11 to
investigate neuropsychiatric risk to the agency.
12
When the agency reviewed them and gave the area
13 some
additional considered thought it was
14
recognized that more basic science groundwork
15
needed to be done before moving on to clinical
16
trials, and this basic science groundwork is now
17
being undertaken in collaboration with the National
18
Institute for Mental Health. As
that data is
19
accrued we will move on at the appropriate time to
20
clinical trials.
21
That is all I am going to say today about
22
risk management of neuropsychiatric risk. I want
32
1 to
remind both the committee and the public that it
2 is
not the subject of this advisory committee.
3
[Slide]
4
Moving on to pregnancy prevention, also
5
addressed in that letter, two goals, as Dr. Galson
6
already mentioned, were articulated.
The first,
7
that no one should begin isotretinoin therapy if
8
they are pregnant and the second, that effective
9
pregnancy prevention would occur throughout the
10
course of isotretinoin therapy.
Implied in these
11 two
goals is that we would have the ability to
12
assess whether or not they have been achieved.
13
[Slide]
14
To achieve these two goals, five points of
15
action were advised: augmentation of patient
16
education; registration of all patients;
17
registration of prescribers; implementation of a
18
pregnancy registry; and linkage of prescription
19
dispensing to adequate pregnancy testing.
20
[Slide]
21
The agency and the sponsor, having heard
22 the
committee's recommendations, entered into
33
1
extensive discussions and negotiations in an
2
attempt to design a plan that would incorporate the
3
five points of action to achieve the two goals that
4 had
been articulated.
5
However, obstacles were encountered,
6
particularly regarding patient privacy issues and
7
compliance with the newly passed Health Insurance
8
Portability and Accountability Act.
Eventually,
9
however, a plan was crafted and was approved in
10
October, 2001. The innovator was
the only product
11 on
the market at that time and they named their
12
risk management plan S.M.A.R.T., a System to Manage
13
Accutane-Related Teratogenicity.
I will refer to
14
their plan and the subsequent generic risk
15
management plans as the current risk management
16
plan so when I use the term the current risk
17
management plan, you can think of that as
18
interchangeable with S.M.A.R.T., S.P.I.R.I.T,
19
I.M.P.A.R.T., etc.
20
I want to now move and describe how the
21
plan that was crafted sought to incorporate those
22
five points of action and then I will describe for
34
1 you
the mechanics of the plan in some detail.
2 [Slide]
3
The first point of action articulated by
4 the
committee was a heightened educational program
5 for
each patient that included verifiable
6
documented written informed consent.
This is
7
fairly straightforward and is a component of the
8
current risk management plan.
9
[Slide]
10
The second point was complete registration
11 of
all patients, both male and female. This
was
12
intended to provide the denominator for
13
ascertainment of the pregnancy rate.
However,
14
registries raise issues regarding patient privacy.
15 The
sponsor proposed an alternative proposal to
16
estimate the denominator using pharmacy databases
17 and
survey data. This, of course, would
avoid
18
those patient privacy issues but the accuracy of
19 the
alternative proposal was dependent on
20
increasing the survey response rate.
The sponsor
21
felt that this would be achievable.
22
[Slide]
35
1
The third point of action was complete
2
registration and certification of all prescribers.
3 The
sponsor objected that they did not have the
4
authority to certify prescribers and so a plan of
5
voluntary registration was devised in which
6
prescribers self-attest that they possess the
7
relevant competencies needed to safely prescribe
8
isotretinoin. Additionally,
prescribers singed a
9
commitment to use the current risk management plan.
10 The
sponsor does provide prescribers with
11
information about the plan, but the responsibility
12 for
obtaining the necessary education to achieve
13 the
relevant competencies rests with the
14
prescriber. I will detail these
competencies in a
15 few
moments.
16
[Slide]
17
The fourth point of action was a
18
comprehensive plan to track fetal exposures to
19
isotretinoin to include a formal pregnancy
20
registry. This was intended to
provide the
21
numerator for ascertainment of the pregnancy rate.
22
Again, because it involved a registry, it raised
36
1 concerns
regarding patient privacy and issues
2
regarding compliance with the newly passed HIPPA.
3
Again, to avoid these obstacles and to
4
speed the implementation of augmented risk
5
management measures, the sponsor proposed
6
extrapolation of the numerator from survey response
7
data. Accurate extrapolation from
survey response
8
data would require an increased survey response,
9
which the sponsor identified as an increased
10
response rate of greater than 60 percent. Now,
11
they did feel that this would be achievable and, in
12
order to achieve the increased rate, they planned
13
targeted education of prescribers to increase
14
awareness of the survey and they increased
15
reimbursement for patient participation by 300
16
percent.
17
[Slide]
18
The final point of action advised by the
19
committee was the linking of dispensing of
20
isotretinoin to verification of adequate pregnancy
21 testing.
This is accomplished in the current risk
22
management plan through the use of yellow
37
1
qualification stickers. The
physician verifies the
2
negative pregnancy test and fills out the
3
qualification sticker. The
patient takes the
4
prescription with the qualification sticker to the
5
pharmacist who then verifies that the patient has,
6
indeed, been qualified. However,
in the current
7 plan
the pharmacist does not independently review
8 the
negative pregnancy test lab report.
Pharmacist
9
participation in the current plan is voluntary but
10
encouraged through the way that the plan is
11
designed.
12
[Slide]
13
I want to take a moment now and describe
14 in
some detail the mechanics of how the current
15
risk management plan works. It
can be a bit
16
complex if you haven't used it yourself in a
17
clinical setting. The program
begins with a
18
physician who decides that they would like to
19
prescribe isotretinoin and that they possess the
20
relevant competencies necessary to do so.
21
The physician will sign a one-time letter
22 of
understanding with the manufacturer, attesting
38
1
that they do possess the necessary knowledge and
2
experience in order to safely prescribe the drug,
3
specifically that they are knowledgeable about the
4
different forms of acne and its treatment; that
5
they are knowledgeable about isotretinoin and its
6
risks for teratogenicity; that they are
7
knowledgeable about the risks for and the
8
prevention of unplanned pregnancy; and finally,
9
that they are knowledgeable about the current risk
10
management plan and that they agree to use its
11
mechanisms.
12
When the manufacturer receives this signed
13
letter of understanding, they then forward to the
14
prescriber the qualification stickers and separate
15
educational materials for both the prescriber as
16
well as for patients. Prescriber
educational
17
materials consist of things like best practices
18
guides that inform the prescriber how to use the
19
components of the current risk management plan.
20
Educational materials for patients include things
21
like brochures and videos.
22
The physician then encounters a patient
39
1 for
whom they believe treatment with isotretinoin
2 is
indicated. From this point forward, as I
am
3
describing the mechanics when I refer to a patient
4 I
am speaking specifically of a female patient.
5 So,
when the prescriber encounters a patient for
6
whom isotretinoin is indicated the first thing that
7
they will do, having made the preliminary decision
8 to
prescribe the drug, is obtain a screening
9 pregnancy test. They would also provide
10
educational materials to the patient and the
11
informed consent forms, which I will talk about in
12 a
minute.
13
Also at this time, contraception
14
counseling and contraception would be provided.
15
This can be accomplished in one of two ways, the
16
prescriber him or herself, if they possess the
17
necessary expertise, can provide the counseling
18
themselves or they can refer to a reproductive
19
health specialist such as a gynecologist for
20
provision of the contraception counseling and the
21
contraception. The female
patient, unless they
22
select complete abstinence, must be on two forms of
40
1
contraception, at least one of which must be a
2
primary form, for 30 days prior to the initiation
3 of
isotretinoin therapy.
4
The patient reads the educational
5
material, obtains the contraception counseling and
6 the
contraception and reads through the informed
7
consent documents, signs those and returns them to
8 the
physician. There are actually two
informed
9
consent documents. The first is
an informed
10
consent/patient agreement which is given to both
11
male and female patients. This
outlines the risks
12 for
teratogenicity, as well as the potential risk
13 for
psychiatric adverse events, and also elicits
14
agreement from the patient that they will abide by
15 the
risk management principles of the current risk
16
management program, such as that they will not
17
share their isotretinoin with other people; they
18
will not give blood until at least 30 days after
19 the
conclusion of their therapy; that they will
20
return to their physician on at least a monthly
21
basis. The second informed
consent document is
22
specific for female patients and goes into much
41
1
greater detail about the risks of unplanned
2
pregnancy and the risk of teratogenicity with
3
isotretinoin therapy.
4
Both of those informed consent forms and
5 the
informed consent/patient agreement need to be
6
signed and returned to the physician.
7
Additionally, before prescribing isotretinoin the
8
physician must obtain a second pregnancy test, this
9
time timed to the woman's cycle within the first
10
five days of the menses or, if the patient is
11
amenorrheic, at least 11 days after the last
12
episode of unprotected intercourse.
After these
13
steps have been accomplished the physician then
14
fills out the prescription form, affixes the
15
qualification sticker and fills that out with the
16
date of qualification signifying that two negative
17
pregnancy tests have been obtained; that the
18
patient understands the risk management program;
19
that adequate contraception, either two forms or
20
absolute abstinence, have been initiated.
21
The patient then takes the prescription
22
with the qualifying sticker affixed and filled out
42
1 to
the pharmacist. The pharmacist verifies
that
2 the
sticker has been affixed, has been properly
3
completed, and also that the receipt of this
4
sticker and the dispensing of the isotretinoin
5
occur within seven days of the date of the
6
physician's qualification of the patient. If all
7 of
those criteria are met the pharmacist dispenses
8 the
isotretinoin along with a medication guide
9
which is an information brochure for patients
10
which, by law, must be dispensed each time
11
isotretinoin is dispensed that describes in
12
layman's language the risks of the drug and the
13
steps that need to be taken to minimize those
14
risks.
15
The patient then initiates their course of
16
isotretinoin therapy and on a monthly basis will
17
return to the prescriber to be requalified.
18
Requalification consists of repeating the pregnancy
19
test and verifying that the test is negative;
20
re-counseling the patient regarding contraception;
21 and
ensuring that the risk management program is
22
being abided by.
43
1
We receive data about the program from
2
several sources, first, spontaneous adverse events
3
reports come to the agency from physicians, the
4
manufacturer, from patients as well as from
5
pharmacists. Additionally, the
patient is
6
encouraged to participate in the voluntary patient
7
survey and data is gathered through that mechanism.
8
Finally, pharmacies are surveyed and the
9
prescriptions are audited to check for compliance
10
with the sticker program.
11
[Slide]
12
The risk management plan, as I have
13
described, was approved for the innovator in
14
October of 2001. Since that time
three generic
15
products have been approved and have entered the
16
market. Their risk management
plans are identical
17 in
the essential elements that I have just
18
described to the innovator plan.
So, again, when I
19
speak of the current risk management plan, that
20
would be interchangeable for either the innovator
21
plan or the plan of the three generic products.
22
[Slide]
44
1
However, while the four risk management
2
plans are identical in their essential elements and
3 can
be considered interchangeable, there are some
4
differences that have caused marketplace confusion.
5
Besides having different trade names for the four
6
drugs, each manufacturer has elected to name their
7
risk management program by a different name so for
8
Accutane with have S.M.A.R.T., the System to Manage
9
Accutane-Related Teratogenicity.
For Amnesteem we
10
have S.P.I.R.I.T, the System to Prevent
11
Isotretinoin-Related Issues of Teratogenicity. For
12
Sotret it is I.M.P.A.R.T., Isotretinoin Medication
13
Program Alerting you to the Risks of
14
Teratogenicity. For Claravis it
is A.L.E.R.T, the
15
Adverse Event Learning and Education Program
16
Regarding Teratogenicity.
Additionally, different
17
survey contractors have been employed by the
18 innovator
who uses Degge/SI and the generic firms
19 who
all use the Slone Epidemiology Unit.
Finally,
20
mid-course changes by the patient's pharmacy
21
provider in brand of isotretinoin dispensed can
22
result in patient confusion and perhaps multiple
45
1
enrollment in the voluntary survey.
2
[Slide]
3
When this current risk management plan was
4
approved the sponsor was instructed to submit a
5
comprehensive report on the metrics of the program
6
after one year of implementation.
This advisory
7
committee has been convened to comment on those
8
data. The advisory committee in
2000 did not
9
address benchmarks nor define success.
Indeed, to
10 do
so is challenging. But at this time I
want to
11
provide you with some rough guidelines that you can
12 use
as you are thinking about three parameters in
13
particular, the survey response rate, the sticker
14 use
and the number of fetal exposures.
15
[Slide]
16
The survey response rate, by the sponsor's
17 own
assertion, would need to be greater than 60
18
percent. The success of the
current risk
19
management program in terms of accurate estimation
20 of
that numerator for the pregnancy rate is
21
dependent on this higher survey response rate. The
22
agency's approval of the current risk management
46
1
plan was based on the sponsor's assertion that they
2
would be able to achieve this threshold.
3
[Slide]
4
The qualification stickers serve as a
5
surrogate endpoint for the use of the current risk
6
management plan. When the agency
approved the plan
7 it
was understood that the stickers were an
8
imperfect surrogate and, in fact, as the data has
9
come in they may be more imperfect than we had
10
realized, and other speakers will describe to you
11 the
linkage between the stickers and various
12
components of the program such as pregnancy
13
testing. However, at the time of
approval the
14
sponsor was informed that because the sticker
15 served
as a surrogate, and an imperfect surrogate
16 at
that, the threshold for success would be very,
17
very high and, in fact, would approach 100 percent
18 in
terms of sticker use.
19
[Slide]
20
Finally, and perhaps most importantly,
21
fetal exposures--it would be difficult to identify
22 an
acceptable number for fetal exposures.
In
47
1
considering what success would look like in terms
2 of
fetal exposures the committee may want to think
3 of
this in parallel with the two goals that were
4
articulated by the 2000 advisory committee, the
5
first goal being that no one initiate isotretinoin
6
therapy if pregnant. This goal,
the responsibility
7 for
which rests largely on the shoulders of
8
prescribers, may best be achievable.
9
The second goal, that no one become
10
pregnant while on isotretinoin therapy, is more
11
complex because it depends on patient behavior.
12
Again, in considering the threshold of success in
13
terms of fetal exposure you may want to think of
14
these two populations independently, and also in
15
considering what risk management tools would impact
16
these populations you may want to consider them
17
separately as different tools may be appropriate.
18
[Slide]
19
In summary, isotretinoin is a uniquely
20
effective drug for the treatment of severe,
21 scarring acne, a truly devastating
disease. There
22 has
been a long history of risk management efforts
48
1 to
prevent fetal exposures to this drug which were
2
built sequentially. The current
risk management
3
program has introduced some new tools and the
4
advisory committee is being asked to comment on the
5
effectiveness of these new tools and the current
6
program.
7
I and my colleagues look forward to
8
hearing your considered input on the data and how
9 we
can optimize the public health by ensuring that
10
isotretinoin is available to the patients who
11
needed it in a context that minimizes and best
12
manages the risks. So, I thank
you for your
13
attention this morning and I would be happy to take
14
your questions.
15
DR. GROSS: Thank you very much,
Dr.
16
Lindstrom. Before the questions,
I would like to
17
introduce an additional consultant who will be
18
participating in our joint advisory committee
19
session, Dr. Vega. Dr. Vega,
would you please
20
introduce yourself?
21
DR. VEGA: Yes, good morning. I am a
22
Board-certified pediatrician with a Masters in
49
1
Public Health and a Fellowship in
2
Pharmacoepidemiology from the Food and Drug
3
Administration. I am also a
former medical
4
epidemiologist from the Office of Drug Safety, with
5
extensive experience with the isotretinoin
6
pregnancy prevention issue. I
presented at the
7
last advisory committee the data on the different
8
options to modify the Pregnancy Prevention Program.
9 I
currently work for PSI International in their
10
adverse event reporting project.
11 Questions from the Committee
12
DR. GROSS: Thank you. Now Dr. Lindstrom
13
will entertain questions from the committees. Yes?
14
DR. CRAWFORD: Dr. Lindstrom,
thank you
15 for
the overview. In terms of considering
possible
16
risk management tools to enhance pregnancy
17
prevention, one thing I am not sure of after
18
reading all the materials we were provided is
19
whether the reasons for failure have been
20
identified. So, has there ever
been any thought
21
given to some type of failure mode analysis
22
determining for those patients who do become
50
1
pregnant, exactly what went wrong so efforts could
2 be
targeted on preventing those failures in the
3
future?
4
DR. LINDSTROM: That is an
excellent
5
question. The speakers that
follow will be
6
addressing the data and I believe also, as much as
7 we
know, the reasons for failures. So, if
you
8
don't mind, I think I will defer the answer to that
9
question to the presentations that will follow
10
mine.
11
DR. GROSS: Dr. Gardner?
12
DR. GARDNER: Dr. Lindstrom, could
you
13
give us some idea of the epidemiology of the severe
14
acne for which these drugs are both specifically
15
indicated and also for which they are being used?
16 For
example, can you tell us the incidence or even
17 the
prevalence of the condition in the population
18 and
the distribution by gender and by age, if you
19
know?
20
DR. LINDSTROM: I will do my best
to
21
answer that question. Acne is
extremely common,
22
particularly in the adolescent age range. The
51
1
incidence has been reported to be 80 percent in the
2
12-20 year-old group and falling to about 3 percent
3 in
the over 45 year-old age group. You can
sort of
4
extrapolate the decrease during that time.
5
DR. GARDNER: Is that severe acne?
6
DR. LINDSTROM: No, that is all
acne.
7
There is not an ICD-9 code for severe acne so it is
8
difficult--I don't actually know and I couldn't
9
find, in preparing for this committee meeting, an
10
incidence or a prevalence for severe acne. I can
11
tell you that recalcitrant nodular acne is not the
12
majority of acne. Severe scarring
acne is a larger
13
proportion of acne patients. As a
practicing
14
dermatologist, it was not uncommon.
I saw scarring
15
acne on essentially a daily basis but I don't have
16
incidence or prevalence figures for you, other than
17 the
prevalence of acne in the population at large.
18
DR. GROSS: Sarah Sellers?
19
DR. SELLERS: A quick question on
the
20
qualification in the current program, the
21
qualification sticker that goes to the pharmacy has
22 a
qualification date on it?
52
1
DR. LINDSTROM: Yes.
2
DR. SELLERS: And, is that date
the date
3 of
the confirmed negative test?
4
DR. LINDSTROM: Yes, it is. For
5
initiation of therapy it would be the date of the
6
second confirmed negative pregnancy test and for
7
ongoing therapy it would be the date of the
8
repeated negative pregnancy test.
9
DR. SELLERS: It is not the date
that a
10
sample was taken for a pregnancy test?
11
DR. LINDSTROM: No, I believe it
is the
12
date--I am sorry, I didn't follow actually your
13
question.
14
DR. SELLERS: The qualification
date is
15
actually when the negative result is received--
16
DR. LINDSTROM: That is my
understanding.
17
DR. SELLERS: --not the date a
sample is
18
drawn for analysis to go to the lab?
19
DR. LINDSTROM: Correct.
20
DR. SELLERS: Thank you.
21
DR. GROSS: Yes, Robyn??
22
DR. SHAPIRO: I guess I am curious
about
53
1 the
HIPPA problem that you have found with some of
2 the
registry ideas. Why couldn't the
patients
3
simply authorize release of particular information
4 in
order for them to get the drug and, therefore,
5
make that information available?
6
DR. LINDSTROM: At the time of the
prior
7
advisory committee and at the time that the agency
8 and
the sponsor were working to craft the plan,
9
HIPPA had just been approved and towards the end of
10
that time period was being implemented.
In working
11
with consul from the company as well as consul
12
within the agency, working out the details of HIPPA
13
compliance proved difficult and while it probably
14
would have been achievable, it was taking a lot of
15
time. So, the sponsor proposed
and the agency
16
approved these alternative methods in order to have
17 a
plan in a more timely fashion that could be
18
implemented that could augment the risk management
19
program. As understanding of
compliance of HIPPA
20 has
matured, I think it would be much easier to
21
navigate those waters at this time but at that time
22 the
Act had just been passed and was in the process
54
1 of
being implemented and understanding was not yet
2
mature.
3
DR. GROSS: Dr. Bigby?
4
DR. BIGBY: I have two
questions. The
5
first one is that you stated that some patients who
6
take Accutane never have acne again.
Are you or
7
someone else going to actually tell the committee
8
what the actual numbers are in terms of the
9
long-term efficacy of Accutane?
10 DR. LINDSTROM: What I had hoped to state
11 was
that patients may achieve complete and
12
long-term remission. I have read different figures.
13
Approximately 10-20 percent of patients who are
14
treated with Accutane never require treatment with
15
Accutane again. Another way to
state that would be
16
that 10-20 percent of patients who undergo a course
17 of
isotretinoin therapy do require a second course
18 of
isotretinoin therapy. Of the 80-90
percent that
19
only require one course of isotretinoin therapy, a
20
portion of those are then able to be maintained
21
with no treatment at all. A
portion would require
22
only topical therapy and some may require oral
55
1
antibiotic therapy.
2
DR. BIGBY: I just think that it
is
3
important for the committee to know actually what
4
those proportions are and I just hope somebody
5
brings that data to the table.
6
DR. LINDSTROM: I don't have those
7
numbers. All I can tell you is
that between 10-20
8
percent of isotretinoin patients do undergo a
9
second course of therapy.
10
DR. BIGBY: Well, those numbers do
exist
11 and
I just hope it is sort of made known to the
12
committee what those numbers are.
13
The other question I had was of the
14
pregnancies that occurred prior to S.M.A.R.T. and
15
during S.M.A.R.T., is there any data about who the
16
prescribers were?
17
DR. LINDSTROM: I am sorry, can
you repeat
18
your question?
19
DR. BIGBY: You presented
information
20
about pregnancies that occurred for the year prior
21 to
S.M.A.R.T. and during a year of S.M.A.R.T.
What
22 I
would like to know is who the prescribers of
56
1
Accutane were for those women who got pregnant.
2
DR. LINDSTROM: Yes, actually I
did not
3
present any data about pregnancies during
4
S.M.A.R.T. My objectives at this
point of the day
5
were to set the historical context so the slide
6
that I showed was that reported pregnancies to the
7
agency were from 1982 through 1999.
Speakers later
8
today will update you with the current pregnancy
9
data, the more recent data during the
10
implementation of the current risk management
11
program.
12
Now, there were two parts to your question
13 and
I only answered half. Can you tell me
again
14 the
second part of that question?
15
DR. BIGBY: No, you answered it.
16
DR. LINDSTROM: Okay.
17
DR. GROSS: Dr. Michael Cohen?
18 DR. COHEN: Earlier you mentioned that
19
there may occasionally be some confusion between
20 the
various risk management programs for
21
isotretinoin that exist and perhaps also the brand
22
names. Are you saying that that
occasionally
57
1
contributes to some of the problem that we are
2
seeing with isotretinoin and the way that it is
3
handled? Also, who actually does
the selection?
4 Is
it the prescriber or the pharmacist? Is
it a
5
substitution that is made? I
didn't understand
6
that.
7
DR. LINDSTROM: In stating the
various
8
names and alluding to confusion, my point is just
9 to
give the perspective of patients and
10
prescribers. It is a somewhat
complex plan and
11
there are various names out there, and to just make
12 the
committee aware that that is a potential source
13 of
confusion, the multiple names for the risk
14
management plans. I did not mean
to imply that
15
there should not be different trade names for the
16
products of the various manufacturers but, rather,
17
that the risk management plan having multiple names
18
does present some confusion for patients. The
19
second part of your question?
20
DR. COHEN: Well, I guess I am a
little
21 bit
confused about who actually selects the brand
22
that will be used. You mentioned
that occasionally
58
1 a
patient can go from one brand to another--
2
DR. LINDSTROM: Right.
3
DR. COHEN: --does that contribute
to any
4
confusion that we should be concerned about? I
5
understand the plans are pretty much the same.
6
DR. LINDSTROM: Right.
7
DR. COHEN: They have the same
baseline
8
requirements but are there any errors that this
9
contributes to that, you know, might have an
10
adverse outcome that we should know about?
11
DR. LINDSTROM: Sure.
12
DR. COHEN: In other words, should
there
13 be
one plan?
14
DR. LINDSTROM: I think that is an
15
excellent question and one that the committee will
16
need to be considering as the day goes forward.
17
Other speakers will present to you the details of
18 the
data that has been obtained from the current
19
risk management plan and will be in a better
20
position to address confusion from the agency's
21
perspective in terms of data collection from
22
multiple plans.
59
1
As far as whether a patient receives one
2
particular manufacturer's isotretinoin or another,
3 a
physician can specify that as they write the
4
prescription but I think in many instances it is
5 the
pharmacy provider that makes that determination
6 of
which patient receives which brand. So,
it is a
7
little bit outside of the prescriber-patient
8
relationship.
9
DR. GROSS: Dr. Kweder?
10
DR. KWEDER: Yes, I think I can
clarify a
11
little bit. We do not have
specific data on the
12
frequency of switching between brands.
We have
13
heard for patients and providers that this is a
14
potential source of difficulty but we do not have
15
data saying how common it is for patients to be
16
required to switch mid-course.
Just like any
17
medication, the source of imposing a change could
18 be
anything from the patient wanting a cheaper
19
brand to the pharmacist pressing for that, or the
20
physician or even the health insurance plan that
21
will only pay a certain amount.
22
DR. GROSS: Dr. Trontell?
60
1
DR. TRONTELL: I was going to just
2
elaborate on Dr. Kweder's remarks.
We don't yet
3
have any data to document that confusion has
4
occurred between these programs.
5
DR. GROSS: Thank you. Dr. Whitmore, did
6 you
have a question?
7
DR. WHITMORE: The answer came up
already,
8
thank you.
9
DR. GROSS: Dr. Day?
10
DR. DAY: Was any provision made
for
11
providing the risk management plan for mail order
12
prescriptions? I assume that
originally Accutane
13 was
available through mail order.
14
DR. LINDSTROM: The prior risk
management
15
plan did allow for mail order prescriptions. For
16 the
current risk management plan, as I understand
17 it,
a mail order prescription might be challenging
18 in
that the drug needs to be dispensed within a
19 seven-day
window of qualification. Not only that,
20 but
there are other features of the plan that might
21 not
happen. So, it is not allowed.
22
DR. GROSS: Dr. Honein?
61
1
DR. HONEIN: I just want to
follow-up with
2
some questions on the multiple risk management
3
programs. I wondered if there was
any data on how
4
often women get one set of information from a
5
prescriber and a different set of information from
6 the
pharmacist at the time it is dispensed, and if
7
there are any reports of that contributing to
8
confusion.
9
DR. LINDSTROM: The information
that the
10
patient receives from the pharmacist would be the
11
medication guide which would be the same for all of
12 the
manufacturers' products, the innovator as well
13 as
the generic. The pharmacy has the option
of
14
providing additional patient education information
15 that is not part of the current risk
management
16
plan that would be in addition to that.
17
DR. HONEIN: Don't they get
enrollment
18
forms both from the prescriber and the pharmacy,
19 and
wouldn't those be different if they got
20
different sets of material?
21
DR. LINDSTROM: Thank you. That is a good
22
point. The enrollment forms are
included with each
62
1
prescription that is dispensed and the enrollment
2
form for the innovator uses one contractor and the
3
enrollment forms for the generics utilize a
4
different contractor so you are correct that that
5
would be another potential source of confusion for
6 a
patient.
7
DR. GROSS: Dr. Knudson?
8
MS. KNUDSON: I am curious about
the age
9
distribution of the women taking the drug. I would
10
like to know does the enrollment form or the survey
11
form or the qualifying sticker carry the age?
12
DR. LINDSTROM: The qualifying
sticker
13
does not. Age may be obtained by
the pharmacy as
14
part of an independent pharmacy data collection
15
with age, date of birth and so forth to ensure that
16 the
correct prescription is dispensed to the
17
correct patient. Age is a
component of the
18
voluntary patient survey.
19
DR. GROSS: Dr. Ringel?
20
DR. RINGEL: This is a quibbling
point
21
from the "nothing in life is perfect" department.
22 You
mentioned that it should be possible to prevent
63
1
initiation of isotretinoin therapy before a
2
pregnancy, and there are ways you can actually
3
manage it if you consider that there is a certain
4
number of false-negative pregnancy tests,
5
particularly early in pregnancy, and also there can
6 be
confusion with bleeding at implantation and
7
bleeding for other reasons with menses.
If you put
8
those together, in fact, it would be possible to be
9
pregnant, despite all of our efforts, before
10
initiating Accutane.
11
DR. GROSS: Dr. Strom?
12
DR. STROM: In the era of
increasing
13
computerized data entry, how would this risk
14
management plan work?
15
DR. LINDSTROM: I am sorry, can
you
16
elaborate on your question?
17
DR. STROM: Sure. The current risk
18
management plan, as I understand it, relies on a
19
sticker program.
20
DR. LINDSTROM: Yes.
21
DR. STROM: There is increasing
use of
22
computerized prescribing and a big push nationwide
64
1 to
increase that.
2
DR. LINDSTROM: Yes.
3
DR. STROM: How could this be
4
operationalized? How could this
plan possibly work
5 in
that context?
6
DR. LINDSTROM: The current risk
7
management plan does not allow for computerized
8
prescriptions.
9
DR. STROM: Just to clarify, given
the
10
current environment in pharmacy, neither mail order
11 nor
computerized prescriptions are compatible with
12 the
current plan.
13
DR. LINDSTROM: Computerized
prescriptions
14 are
not compatible with the current plan and I
15
think mail order would be difficult with the
16
current plan. Again, I have set
the historical
17 context and described the current plan.
18
DR. GROSS: Dr. Kibbe?
19
DR. KIBBE: I have just a question
about
20 the
two figures that you gave us and the data that
21 is
contained therein. Have you taken the
number of
22
reports of pregnancies for the years from '91 to
65
1 '99
and divided them by the number that you show
2 for
the number of female patients during those same
3
years and gotten, even though it is an inaccurate
4
number, at least an estimate of number of
5
pregnancies per 1,000 patients over that time
6
frame?
7
DR. LINDSTROM: I believe that you
are
8
bringing up the issue of pregnancy rate.
While the
9
absolute number of pregnancies reported to the
10
agency was relatively constant, the number of women
11
receiving isotretinoin prescriptions was rising. I
12
don't want to belabor this point but there are two
13 issues
related to deriving a rate from the data
14
that I showed. First, pregnancy
reporting is
15
voluntary, both the spontaneous reports and those
16
received through the survey. They
are voluntary.
17
Both are voluntary mechanisms. We
know that
18
adverse event reporting declines over time and we
19
know that it does not capture all events so it is
20 an
imprecise number.
21
Second, even if that numerator in terms of
22 the
number of pregnancies reported was reflective
66
1 of
the total number of exposed pregnancies that had
2
occurred, even if that number, indeed, did stay
3
flat the public health burden of those exposed
4 pregnancies, of those affected babies, was
not
5
declining. Those two slides were
actually
6
presented to the advisory committee in 2000, and
7 for
those reasons it was determined to be important
8 to
increase the risk management for this drug
9
because the public health impact has remained
10
significant.
11
DR. KIBBE: So, your answer is no?
12
DR. LINDSTROM: Yes.
13
DR. GROSS: Dr. Bull?
14
DR. BULL: I just wanted to remind
you,
15
going back to the issue of computerized
16
prescriptions, that this whole risk management plan
17 is
predicated on a high level of interaction
18
between the patient and the healthcare provider.
19
These are non-refillable prescriptions.
The
20
patient has to return to the healthcare provider
21 for
an interaction, hopefully a face-to-face
22
evaluation of how the acne treatment is
67
1 progressing,
such that because of the fact that
2
these are not prescriptions that are automatically
3
refilled it is not a course of therapy where you
4 are
given a prescription that you renew for five
5
months. It is one where every month
during that
6
course of time there is a need to return to the
7
healthcare provider of record.
8
DR. GROSS: I am going to take the
9
prerogative of the chair and declare a break at
10
this particular time. We have no
breaks scheduled
11 for
the morning and I think we will hold questions
12
until a little bit later. Thank
you. We will
13
reconvene at 9:30.
14
[Brief recess]
15 Open Public Hearing
16
DR. GROSS: Both the Food and Drug
17
Administration and the public believe in a
18
transparent process for information gathering and
19
decision-making. To ensure such
transparency at
20 the
open public hearing session of the advisory
21 committee meeting, which we are about to
start, the
22 FDA
believes that it is important to understand the
68
1
context of an individual's presentation.
For this
2
reason, the FDA encourages you, the open public
3
hearing speaker, at the beginning of your written
4 or
oral statement to advise the committee of any
5
financial relationship that you may have with the
6
sponsors of any products in the pharmaceutical
7
category under discussion at today's meeting. For
8
example, the financial information may include the
9
sponsor's payment of your travel, lodging or other
10
expenses in connection with your attendance at the
11
meeting. 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 this
15
issue of financial relationships at the beginning
16 of
your statement it will not preclude you from
17
speaking.
18
We have two registered speakers for the
19
morning, Dr. Robert A. Silverman is first. Dr.
20
Silverman?
21
DR. SILVERMAN: Dr. Gross, members
of the
22
advisory committee, thank you for giving me the
69
1
opportunity to speak about the continued
2
availability of isotretinoin. My
statement will
3
focus on the benefits of this drug and the impact
4 of
pregnancy prevention risk management efforts on
5 its
availability to patients.
6
I have been practicing pediatric
7
dermatology for nearly two decades.
At first I was
8 in
Cleveland at Rainbow Babies and Children's
9
Hospital. Since 1989 I have
maintained a private
10
practice in Northern Virginia and a dermatology
11
clinic in the Department of Pediatrics at
12
Georgetown University. For the
record, I have not
13
participated in any pharmaceutical company
14
sponsored acne drug studies, nor am I taking any
15
reimbursement from the AADA, and the only thing I
16
have taken today is one bottle of water.
17
[Laughter]
18
I am a physician who treats patients, not
19 a
healthcare provider who sees clients. I
make the
20
distinction to emphasize the trust and close
21
relationship between a physician and patient that
22 is
necessary for obtaining the best results when
70
1
treating acne while minimizing side effects of any
2 of
the medications that we use. As a
pediatrician,
3 I
recognize the social and psychological impact
4 that an acne-scarred body image has on
teenagers.
5 I
know of no drug that has changed the lives of my
6
patients with acne more than isotretinoin. It has
7
been a Godsend to adolescents and to young adults
8
with recalcitrant, nodular, nodulocystic and
9
scarring disease.
10
Unlike dermatologists entering the medical
11
work force today, I remember how we used to treat
12
severe nodulocystic acne. One of
the most painful,
13
gruesome procedures that I learned in my training
14 at
the Children's Hospital in Boston was the
15
incision and drainage of multiple purulent
16
abscesses, like you saw earlier, on the faces of
17
young men and women afflicted with recalcitrant
18
nodulocystic acne. The procedure
is nearly a
19
historical footnote since we have the availability
20 of
isotretinoin.
21
There is not a week that goes by in my
22
practice that a concerned parent, with facial scars
71
1
themselves, brings in a preadolescent with minimal
2 or
no acne for anticipatory guidance in hopes of
3
their child avoiding the same fate that they had
4
when they were growing up. Of
course, the vast
5
majority of these children never-ever reach the
6
point of needing isotretinoin.
But for the few who
7
progress and require it, I am thankful that I have
8 the
option to use this medication. The
reason I am
9
here today is to keep this drug available to all
10
people who need it.
11
Let me share a story that perfectly
12
illustrates the wonders that can be worked by this
13
drug. In 1982, when I was in
Boston, the year that
14
isotretinoin first became available in the United
15
States, I met a beautiful young lady who had a
16
beautiful complexion. During that
year she
17
developed inflammatory acne that then rapidly
18
progressed to severe painful, nodulocystic disease.
19 She
was being cared for by an excellent
20
dermatologist at one of the nation's first and
21
premier HMOs. Minocycline,
benzoyl peroxides,
22
Retin-A and oral contraceptives made no difference
72
1 in
her appearance.
2
Isotretinoin was not widely prescribed and
3 it
was not until 1986 when she saw her fourth
4
dermatologist, after moving to Washington, D.C.,
5
that Accutane was offered to her.
The years
6
between 1982 and 1986 were for her filled with
7
anxiety and self-consciousness. I
know this
8
because this woman is now my wife.
She took
9
isotretinoin safely. She was
aware of the
10
teratogenic risks and used two forms of birth
11
control. We now have two healthy
boys who were
12
conceived well after my wife-to-be's finishing the
13
drug. This story is obviously
close to my heart
14 but
it also illustrates the fact that female
15 patients
of childbearing potential can and do use
16
isotretinoin safely.
17
I have treated many teenaged girls and
18
young women with isotretinoin. I
have personally
19
prescribed isotretinoin since 1986 and have used it
20 according
to the risk management guidelines with
21
utmost caution, and since the S.M.A.R.T. program
22 has
been in effect I have complied with it to the
73
1
best of my ability.
2
As a clinician in the trenches, I am
3
familiar with the difficulties and weaknesses that
4
were outlined that may impede optimal participation
5 in
the S.M.A.R.T. program. Complicating and
6
restricting access will only drive needy patients
7 to
obtain isotretinoin through illicit channels or
8
those that circumvent well-established
9
doctor-patient relationships.
This would be a
10
travesty of monumental proportions.
In grade
11
school I learned the acronym KIS--keep it simple.
12 The
more complicated you make the process of
13
obtaining this medication the more mistakes are
14
going to be made.
15
I would be happy to help in any way that I
16 can
to keep this medication available to all who
17
need it and to address the small, but unfortunate,
18
number of pregnancies that have occurred while on
19
this drug. Thank you for your
time and
20
consideration and I would be happy to entertain any
21 questions if we have a few seconds. Thank you.
22
DR. GROSS: Thank you very much,
Dr.
74
1
Silverman. The next speaker is
Dr. Sidney Wolfe of
2 the
Public Citizen's Health Research Group.
3
DR. WOLFE: Helping out in this
4
presentation is Dr. Sherri Shubin who is a
5
pediatrician and currently doing a preventive
6
medicine residency at Johns Hopkins.
She is
7
spending part of her residency with us.
8
[Slide]
9
I will take a minute or so to go over the
10
first couple of slides. Our
involvement really
11
started shortly after the drug came on the market
12 in
September of '83. We submitted a
petition
13
urging patient package inserts and black box
14
warnings about birth defects and life-threatening
15
adverse events. Prior to
approval, as many of you
16
know, there was a pretty comprehensive program to
17
make sure that no one who got the drug got
18
pregnant, and a number of those strictures were
19
dropped at the time of initial marketing and slowly
20
some of them were reintroduced.
21
On April 26, ADA testified before this
22
committee describing Accutane as an imminent public
75
1
health hazard and saying that unless certain
2
restrictions were imposed it should really come off
3 the
market. The restrictions are listed
there, one
4 of
the most important of which is, of course,
5
limiting prescribing to dermatologists who file
6
sworn affidavits stating they will adhere to the
7
stated indications for the drug.
That is supposed
8 to
be happening, not the sworn affidavit part but,
9
obviously the amount of prescriptions belies the
10
fact that that is what it is limited to.
We then
11
filed a petition to the FDA in May of 1988 with
12
recommendations, saying it should come off the
13
market and only be allowed back on with these
14
restrictions.
15
[Slide]
16
Just finishing up a little bit on that, we
17
continued urging removal from the market unless
18
restrictions were put in and, thus far, these
19
restrictions just have not been put in.
Most
20
recently, in September, 2000, we testified that at
21
that time the issue of depression and suicide had
22
arisen and again we proposed restrictions.
76
1
[Slide]
2
This is testimony before this committee by
3 Dr.
David Erickson, who was then Chief at the
4
Centers for Disease Control and Prevention of the
5
Genetics and Birth Control Branch.
His statements
6 are
very poignant because 15 years later the same
7
issue is there: "The birth of babies with defects
8
caused by fetal exposure to Accutane is
9
unnecessary. FDA decision to
allow the marketing
10 of
Accutane is a failed regulatory experiment. A
11
decision to depend on better contraception alone,
12
without active intervention to reduce the number of
13
users, is a decision to leave the number of
14
affected babies at an unacceptably high level."
15
Finally, one of his suggestions was, "perhaps a
16
formal IND," investigational new drug, "would be a
17
suitable mechanism to reduce the frequency of
18
Accutane embryopathy."
19
[Slide]
20
Now, these are data from the package that
21 was
provided to you a couple of weeks ago--it
22
should have been included. This
is an FDA
77
1
presentation before this committee back in
2
September of 2000. What they said
was that as of
3
that time these are just the reported cases and, as
4
several people said this morning and it understates
5 the
actual magnitude of the problem with 1,995
6
exposed pregnancies; 1,214 elective abortions; 383
7
live births; and 162 infants with birth defects.
8
[Slide]
9
These now are the more recent data from
10 the
first year of the S.M.A.R.T. program.
Again,
11 the
first two points are taken directly from the
12
package that was handed out and 156,800
13
"unique"--the phrase used in there--women were
14
given the drug. Secondly, the
estimated pregnancy
15
rate, and I am sure this is on the low side but
16
that is what was in this information set is 0.35
17
percent. If you take that rate
and apply it to the
18
number of "unique" women given the drug in that
19
first year it means that there have been 548
20
pregnancies and this is 4.6 times higher than the
21
voluntarily spontaneously reported pregnancies that
22 are
also listed in the package, which is a measure
78
1 of
the under-reporting.
2
[Slide]
3
Of the 61 pregnancies with known
4
outcomes--remember, about half of them had outcomes
5
unknown--48 of 61 or 78.7 percent resulted in
6
elective abortions. Again, if you
apply this to
7 the
more likely estimate of the actual number of
8
pregnancies, 548, this means that there would have
9
been 431 elective abortions in that one year ending
10 in
March of 2003.
11
[Slide]
12
Again estimating the number of deliveries,
13 of
61 pregnancies with known outcomes, 7 of 61 or
14
11.5 percent resulted in deliveries.
There were
15
some spontaneous abortions, and so forth that make
16 up
some of the other ones aside from the elective
17
abortions. Again, applying this
to the 548
18
estimated pregnancies, there would have been 63
19
deliveries. Using FDA's and the
CDC's figure,
20
which is probably on the low side, 25 percent birth
21
defects and the 50 percent mental retardation is as
22
close--there hasn't been any really careful study
79
1 on
it but applying those figures to this estimate,
2 we
are talking about 16 infants with birth defects
3 and 31 with mental retardation just in that
one
4
year.
5
[Slide]
6
The reason the S.M.A.R.T. program and even
7 the
new Roche proposal do not seriously address the
8 two
major issues here are as follows: In
1989 CDC
9
estimated that there were no more than 4,000 women
10 of
childbearing age with severe cystic acne.
They
11 did
not even get into the recalcitrant or other
12
therapy. Adjusted for population
growth because
13
these were 1987 data, the number may now be 6,000.
14
Given that there were 156,800 "unique" women of
15
childbearing age who got the drug in that first
16
year of S.M.A.R.T., this represents a 26-fold
17
excess in prescribing over the number of on-label
18
prescriptions.
19
The second point is that unless there is
20
something more than a sticker and an assurance but
21
there is actually the provision of a lab test
22
showing that the woman, in fact, was not pregnant
80
1 and
at least a description of the contraceptive
2
methods--unless that happens, then people are going
3 to
have stickers that are misrepresenting what has
4
actually happened.
5
[Slide]
6
The reason why we are about to file a
7
petition in the next week or two asking for this
8
drug to be taken off the market and made available
9
through an IND is that there have been 20 years of
10
failed voluntary and even more recently some
11
mandatory restrictions, and they have led to
12
actually a total of more pregnancy exposures
13
because the actual amount of prescriptions has gone
14
up. I think it was estimated in
'88 or '89 that
15
there may be 70,000 "unique" women of childbearing
16 age
getting the drug and it is now some 150,000.
17
As we recommended in '88 and the CDC
18
itself suggested as an option the next year, as I
19
showed you in Dr. Erickson's presentation, we now
20
propose a ban on marketing with subsequent
21
availability only under a tightly controlled IND as
22 the
only feasible way to significantly reduce
81
1
prescriptions and pregnancy exposures.
2
[Slide]
3
These would be the main elements of the
4
restrictions in an IND:
Photographic proof of
5
severe cystic acne confirmed by an independent
6
group of dermatologists. Digital
cameras make this
7
kind of process relatively easy to set up.
8
Secondly, a written record for each
9
patient that there, in fact, is adequate previous
10
treatment of the disease with antibiotics and other
11
treatments and that there is recalcitrance to it.
12
Third, a written statement of
13
contraceptive practices and provision of a copy of
14
this and a negative pregnancy test in order for the
15
drug to be dispensed each time.
16
[Slide]
17
In summary, the S.M.A.R.T. program is
18
clearly a failure. Without these
proposed IND
19
restrictions, this administration and this advisory
20
committee will continue to put its imprimatur on
21 the
reckless use of a drug that each year causes
22 the
need for hundreds of abortions and many
82
1
seriously deformed infants with birth defects
2 and/or mental retardation. This is one of the two
3
worst epidemics of preventable serious birth
4
defects ever seen in the U.S. I
would just point
5 out
the other one is two defects where there is
6
deficiency of folic acid, as you know.
The odds of
7
neural tube defects are a couple of orders of
8
magnitude lower than the odds of a birth defect
9
with a live birth. Of course, it
is different not
10 to
have enough folic acid as opposed to be
11
administering one of the more potent teratogens we
12
have ever seen. It is time to end
the more than 20
13
years of voluntary restrictions and some mandatory
14
ones that have failed to reduce its prescribing for
15
more than 20 times as many women as would be using
16 the
drug if it were limited to the approved
17
indications. Thank you. I would be glad to try
18 and
answer any questions.
19
DR. GROSS: Thank you very much,
Dr.
20
Wolfe. The final speaker in the
open public
21
hearing will be Dr. Sherri Shubin, who will read a
22
letter from Dr. Furberg, which is in your packet.
83
1
DR. SHUBIN: Thank you. I have no
2
financial conflicts of interest.
As a member of
3 the
public, I would like to read the statement that
4 was
written by Dr. Curt Furberg, a member of this
5
committee who could not be here today:
6
Due to an unexpected family health
7
problem, I will not be able to attend the upcoming
8
advisory committee meeting on Accutane risk
9
management. Based on long
observation and careful
10
study, I feel very strongly about this issue and
11
regret that I will not be there to express my views
12 and
participate in the committee's discussion and
13
deliberation. As a member of the
committee, I
14
would ask that the following be read aloud at the
15
meeting after all testimony has been presented but
16
before the committee begins its consideration and
17
discussion of the issue and the questions presented
18 it
by the agency.
19
To be candid, the history of Accutane is
20 an
example of inadequate and ineffective risk
21 management
by the FDA and the manufacturer of
22
Accutane to the detriment of thousands of women.
84
1
Examples are numerous. Although
Accutane was a
2
known animal teratogen and a suspected human
3
teratogen at the time of its approval, the company
4 did
not recommend and the FDA did not insist upon
5
labeling that emphasized the importance of
6
contraception or abstinence while under treatment
7 with
the drug. The consequences of this
omission
8
become more apparent when one understands that five
9
women became pregnant while taking Accutane during
10
pre-approval clinical trials despite following the
11
contraception requirements of the study.
12
In 1988, a highly publicized FDA advisory
13
committee meeting was held to discuss the high
14
level of pregnancy exposure to Accutane and the
15
overuse of the product and its contribution to the
16
pregnancy exposure problem. The
Pregnancy
17
Prevention Program, PPP, emerged following this
18
meeting as the primary means of managing Accutane's
19
teratogenic risks. Several
advisory committee
20
meetings were held to monitor the progress of the
21 PPP
between 1989 and 1991. It was clear from
these
22
meetings that the majority of women taking Accutane
85
1
were not volunteering to participate in the PPP,
2 that
even in the group that did volunteer pregnancy
3
testing was infrequently performed and that
4
pregnancy exposure to Accutane was still occurring
5 at
a high level.
6
Remarkably, no advisory committee meeting
7 on
the Accutane pregnancy exposure and the
8
performance of the PPP was convened until
9
September, 2000. At this meeting,
it was shown
10
that enrollment in the PPP was low and falling,
11
that pregnancy testing was still often not being
12 performed and that recommendations about
13
contraception or abstinence were often not adhered
14
to. Even more alarming, the use
of Accutane in
15
women had increased three-fold during the preceding
16
ten-year period when one would have expected it to
17
decline substantially because of successful
18
treatment of prevalent cases of severe nodular
19
acne. The committee's response to
this evidence
20 was
to declare the PPP a failure and to recommend
21
that a comprehensive risk management program that
22
included patient and physician registration, as
86
1
well as mandatory pregnancy testing, be
2
established. None of these has
been implemented.
3
Instead, the S.M.A.R.T. program was
4
introduced. It is an effort that
added yellow
5
stickers to the existing PPP, but had no means of
6
determining if pregnancy testing was actually
7
performed or of how many pregnancy exposures
8
actually occurred. Unfortunately,
S.M.A.R.T. had
9 the
same basic design limitations as the PPP and
10
this should have been recognized.
Now, after
11
almost four years and thousands more of unnecessary
12 pregnancy
exposures to Accutane, this committee is
13
once again asked to advise the FDA.
14
Simply put, I believe that the system is
15 not
safe and cannot be used in a safe manner.
To
16
minimize the number of pregnancy exposures to
17
isotretinoin an IND-like process could be
18
implemented that ensures universal pregnancy
19
testing, registration of all pregnancy test results
20 and
incorporates a mechanism whereby the drug
21
cannot be dispensed without a negative pregnancy
22
test. This coupling of a negative
pregnancy test
87
1
with dispensing of the drug would be analogous to
2 the
policy that has been successfully employed with
3 the
antipsychotic clozapine and has been summarized
4 as
"no blood, no drug." An added
benefit of such
5 an
approach would be that we would have more
6
accurate information regarding the actual number of
7
pregnancy exposures to the drug.
The numbers we
8
have now, coming from a relatively small and
9
self-selected group of volunteers, is undoubtedly a
10
gross underestimate of reality.
11
Other features of this IND-like approach
12
could include limiting the number of
13
isotretinoin-dispensing centers, mandatory
14
pregnancy avoidance counseling at each visit and
15 the
proviso that dispensing centers would be
16
audited periodically. An
important objective of
17 our
risk management should be to reduce the overuse
18 of
isotretinoin. Therefore, I would
recommend that
19 the
IND-like process I have briefly described
20
include some means of documenting the presence of
21
severe nodular acne in patients being considered
22 for
isotretinoin treatment. In clinical
trials for
88
1 the
approval of Accutane only patients with severe
2
cystic acne were enrolled, and photographs of at
3
least some of these patients were taken and used in
4
advertisements and at professional meetings.
5
Perhaps a photograph, documenting the patient's
6
severe cystic acne, could be required prior to
7
approval for treatment.
8 The S.T.E.P.S. program for
thalidomide has
9
been talked about as a possible model for
10
isotretinoin risk management, but it would not be
11
adequate. If my understanding is
correct, there is
12
actually no current coupling of a negative
13
pregnancy test with dispensing of the drug and
14
there is no central registry of the pregnancy test
15
results. Under this system,
thalidomide
16
prescribers answer several questions over the
17
telephone in response to automated prompts in order
18 to
receive a number authorizing use of the drug for
19 the
next month. This system is very similar
to the
20
yellow sticker system under S.M.A.R.T. in that it
21
relies on prescriber self-attestation.
There is no
22
validation that what the prescriber has answered is
89
1
true and there is no comprehensive or reliable
2
means of knowing how many pregnancy exposures have
3 occurred.
4
The occurrence of pregnancy exposures with
5 the
original Accutane pre-approval clinical trials
6
and, more recently, within a clinical trial for
7
another formulation of isotretinoin raises an
8
uncomfortable question, should this drug have ever
9
been released on the open market?
I think it was
10 and
is unethical to allow isotretinoin to be
11
available for use outside of the protections that
12
would be afforded by a controlled and documentable
13
process of distribution.
14
I do have copies of this statement for
15
anyone who would like one.
16
DR. GROSS: Yes, there are copies
of the
17
statement in the committee's folders.
Thank you
18
very much, Dr. Shubin. Shalini
Jain has a comment
19 she
would like to make now.
20
MS. JAIN: I just want to make a
comment
21 for
a point of clarification with regards to Dr.
22
Furberg's letter. Dr. Shubin was
reading the
90
1
letter on behalf of Dr. Furberg.
In functioning as
2 an
FDA committee member representative, he has not
3
been cleared for this meeting for purposes of
4
conflict of interest but solely as a representative
5 of
the public today. Thank you.
6
DR. GROSS: Thank you. We will now move
7 on
to the next set of presentations. From
8
Hoffmann-La Roche, Joanna Waugh, Group Director for
9
Regulatory Affairs, is first; Dr. Martin Huber,
10
Vice President, Global Head, Drug Safety Risk
11
Management; and Dr. Susan Ackermann Shiff, Global
12
Head, Risk Management, Drug Safety Risk Management.
13
Hoffmann-La Roche, Inc. Presentations
14 Introduction
15
MS. WAUGH: Good morning.
16
[Slide]
17
I am Joanna Waugh, from the Regulatory
18
Affairs Department at Hoffmann-La Roche, Nutley,
19 New
Jersey. Thank you to the FDA and the
committee
20 for
giving us the opportunity to present today.
21
[Slide]
22
What I would like to do first is to just
91
1
give you an overview of the framework of our
2
presentation today. Having heard
the FDA
3
presentation, there is some overlap with our
4
presentation so we will go fairly rapidly through
5
some of the areas where there is duplication.
6 Following myself, Dr. Martin Huber
will
7
provide a brief overview of the risk/benefit
8
profile for isotretinoin. I will
then briefly
9
summarize a regulatory overview, focusing on risk
10
management milestones for Accutane since its launch
11 in
the U.S. in 1982. Dr. Susan Ackermann
Shiff
12
will then provide an overview of the S.M.A.R.T.
13
program, comprising a description of what that
14
program entails, as well as an assessment of some
15 of
the data with particular reference to metrics
16
which were predetermined in agreement with FDA.
17 Dr.
Martin Huber will then review the pregnancy
18
data from the S.M.A.R.T. program and move on to
19
discuss our recommendations for program
20 modification.
21
[Slide]
22
In addition to the team of presenters
92
1
which are listed on the left-hand side of this
2
slide, Roche does also have available, for
3
responding to questions in the question and answer
4
session, some additional colleagues, Miss Kay Bess
5
from our Drug Safety Risk Management Department,
6 Dr.
Karen Blesch, from the same department, Miss
7
Tammy Reilly, Vice President of Dermatology and
8
Oncology, and Dr. Susan Sacks, from the Drug Safety
9
Risk Management Department.
10
[Slide]
11
Additionally, we have available the
12
following outside experts for responding to
13
questions, Dr. Diane Berson, from Cornell
14
University; Dr. Judith Jones, from the Degge Group;
15 and
Dr. Victor Strecher, from the University of
16
Michigan School of Public Health.
17
[Slide]
18
As this slide shows and was referred to by
19 the
FDA in their presentation, the S.M.A.R.T. risk
20
management program was approved in 2001 and it was
21
subsequently implemented early in 2002.
Since 2002
22
generic isotretinoin has also been available on the
93
1
U.S. marketplace and this slide shows the
2
respective manufacturers' products and risk
3
management programs, which are all equivalent to
4 the
Accutane S.M.A.R.T. risk management program.
5
[Slide]
6
The conditions for the approval of the
7
S.M.A.R.T. risk management program included the
8
requirement to develop a backup program for a
9
mandatory registry, as well as the understanding
10
that a follow-up advisory committee would be
11
convened when more data was available to discuss
12 the
effectiveness of the program, which is why we
13 are
here today. When more data was
available,
14
Roche evaluated that data and developed a specific
15
proposal for program enhancement based on the data
16 we
saw emerging.
17
[Slide]
18
In December of 2003, the FDA, Roche and
19 the
generic companies reviewed the data across our
20
respective risk management programs.
Roche and the
21
generic companies subsequently worked together on
22
recommendations for program modification. All
94
1
companies agree on the need for one single program
2 and
the recommendation that you will hear put
3
forward today is generally agreed to by all the
4
companies. The details of the
implementation
5
require some further refinement and discussion and
6 we
look forward to the discussion from the advisory
7
committee today on the proposal that we will put
8
forward to you.
9
I will now hand over to Dr. Martin Huber.
10 Benefit/Risk
11
DR. HUBER: Good morning.
12
[Slide]
13
What I would like to briefly review for
14 you
is the benefit/risk. As a first step in
any
15
risk management approach there needs to be an
16
assessment of both the benefit and the risk that we
17 are
addressing.
18
[Slide]
19
Just to remind you, isotretinoin is
20
indicated for severe recalcitrant nodular acne. It
21 is
indicated only for patients who are unresponsive
22 to
conventional therapy, including systemic
95
1
antibiotics. Finally, it is
indicated only for
2
those females who are not pregnant and agree to not
3
become pregnant.
4
[Slide]
5
The medical need for isotretinoin is
6
because it is a serious disease with profound
7
consequences. Inadequately
treated severe
8
recalcitrant nodular acne can lead to disfiguring
9
scarring. Fortunately, it is a
uniquely
10
efficacious therapy for this condition and there
11 are
currently no alternative therapies for these
12
patients.
13
[Slide]
14
To briefly remind you, this is the
15 concern.
Inadequately treated SRNA can lead to
16
disfiguring scarring which is life-long.
17
[Slide]
18
However, there is a specific challenge for
19
isotretinoin, as has been indicated by the previous
20
speakers. Isotretinoin is known
to be a human
21
teratogen. The majority of the
female patients who
22 use
this drug are of childbearing potential.
96
1
Therefore, pregnancy prevention measures, including
2
proactive risk management, are essential. But the
3
specific challenge of this program is that we must
4
change the behavior of patients in order to have
5
them comply better with these risk management
6 programs.
7
[Slide]
8
The public health goals, as previously
9
stated, remain the same. Our
vision is that no
10
woman who is pregnant should receive isotretinoin
11
therapy; no woman should become pregnant during or
12 for
one month after receiving isotretinoin therapy.
13
[Slide]
14
I will now turn it over to Miss Waugh who
15
will review the regulatory history and the risk
16
management program to date.
17
Regulatory Overview
18
[Slide]
19
MS. WAUGH: The teratogenic risk
of
20
Accutane has been known since the approval of the
21
drug in 1982 in the U.S. Because
of this known
22
risk, we have taken a variety of risk management
97
1
steps throughout the product life cycle with the
2 aim
of reducing pregnancies as far as possible.
3 The
proposed program that you will hear today
4
includes risk management enhancements in response
5 to
data that we have seen in the S.M.A.R.T. risk
6
management program.
7
[Slide]
8
This slide provides an overview of some
9
examples of steps Roche has taken throughout the
10
product life cycle to minimize pregnancies. Since
11
product launch in 1982, the product had a pregnancy
12
category X, i.e., it was contraindicated in
13
pregnant women. In 1984 a black
box warning was
14
introduced to increase the prominence of warnings
15
surrounding pregnancy.
16
In 1988 the Pregnancy Prevention Program
17 was
introduced which FDA alluded to in the earlier
18
presentation. This was the first
risk management
19 program
of its kind which used mechanisms over and
20
above labeling as tools for risk management. Some
21
components of the Pregnancy Prevention Program are
22
listed on this slide and I will just go through
98
1
them briefly, the requirement for two forms of
2
contraception to be used simultaneously for one
3
month before, during and after Accutane treatment.
4
Additionally, the requirement for negative monthly
5
pregnancy testing; the addition of an "avoid
6
pregnancy" symbol in the packaging.
Educational
7
materials were introduced regarding contraceptives
8 and
pregnancy avoidance, and a female informed
9
consent form was introduced.
10
Further evaluation tools to assess the
11
effectiveness of this program were introduced which
12
included the Accutane survey.
This survey was
13
developed by the Slone Epidemiology Center at
14
Boston University and provided information about
15
women's understanding of the risk issues related to
16
teratogenicity, as well as some information about
17 the
pregnancy rate based on the number of women
18
enrolled in the survey.
19
[Slide]
20
in 1990 we added information to the U.S.
21
product information concerning a description of
22
birth defects that could occur, as well as a
99
1
recommendation that prescribing should be limited
2 to
a one-month supply.
3
In 1994 the patient informed consent form
4 was
updated to include additional requirements.
In
5 May
of 2000, amongst additional requirements, one
6 of
the requirements was to have two negative
7
pregnancy tests prior to the initial prescription.
8
In September of 2000, as has been
9
mentioned earlier, an advisory committee was
10
convened which discussed pregnancy prevention. I
11 will come back to that in a little bit more
detail
12
later. Subsequent to that
advisory committee,
13
Roche worked in collaboration with the FDA to
14
determine how best to implement their
15
recommendations. The result of
these discussions
16 was
the ultimate approval for the S.M.A.R.T. risk
17
management program in October, 2001.
18
[Slide]
19
As I mentioned, the 2000 advisory
20
committee discussed pregnancy prevention. The
21
recommendations from that advisory committee, as
22
mentioned by the FDA, included the recommendation
100
1 for
the introduction of patient and prescriber
2
registry. In subsequent discussions
with the
3
agency, Roche put forward various proposals which
4
included mandatory patient and prescriber
5
registration. In discussions with
the agency about
6 the
best way to implement these recommendations and
7 in
view of some of the issues that FDA alluded to
8
earlier, Roche and FDA agreed that the critical
9
issue was to link a negative pregnancy test with
10
each prescription and dispensing of Accutane.
11
[Slide]
12
The S.M.A.R.T. program introduced a link
13
between dispensing and negative pregnancy testing
14 via
the Accutane qualification sticker.
15
Additionally, the S.M.A.R.T. program included
16
enhanced education, enhanced informed consent, and
17 the
requirement for prescribers who wish to
18
prescribe Accutane to be registered into a
19
database.
20
[Slide]
21
Dr. Susan Ackermann Shiff will now provide
22
more details about the S.M.A.R.T. program.
101
1 Overview of the S.M.A.R.T.
Program
2
DR. ACKERMANN SHIFF: Thank you.
3
[Slide]
4
What I would now like to briefly do is
5
overview the S.M.A.R.T. program or the System to
6
Manage Accutane-Related Teratogenicity.
The
7
program was developed with and approved by the FDA,
8 and
went into effect on April 10 of 2002.
9
[Slide]
10
This high level overview slide provides
11 two
important features, first that the registered,
12
qualified physician, the qualified patient and the
13
pharmacist work together in the dispensing of the
14
product. Second, the
qualification sticker is the
15 one
area where the negative pregnancy test and the
16
dispensing of the product is linked.
17
[Slide]
18
Different from the Pregnancy Prevention
19
Program, all prescribers must be enrolled in the
20
program in order to prescribe the product. A
21
prescriber will read the guide to best practices,
22
sign a letter of understanding and receive the
102
1
qualification stickers.
2
[Slide]
3
When the prescriber signs the letter of
4
understanding they attest to the fact that they
5
know the risk and severity of fetal injury and
6
birth defects; that they know how to diagnose and
7
treat various forms of acne; that they know the
8
risk factors of unplanned pregnancy and they will
9
properly follow the S.M.A.R.T. procedures. In this
10
case, it includes education, pregnancy testing,
11
contraception, informed consent and offering of the
12
Accutane survey.
13
[Slide]
14
Once the prescriber has been registered
15
within the system, they can educate the patient on
16 the
appropriate use of the product. The
"Be Smart,
17 Be
Safe, Be Sure" educational brochure that is
18
shown on this slide contains elements of education
19
about the product; contraceptive information; the
20 two
informed consents, the all-patient informed
21
consent and the female patient informed consent; an
22
enrollment card for the Accutane survey; and
103
1
educational reinforcement. The
purpose of this
2
brochure is that it be used at the initial office
3 visit
and all subsequent office visits.
4
[Slide]
5
As Roche understands that patients learn
6 in
different ways, we have also provided a variety
7 of
other educational materials. There are
story
8
boards in both English and Spanish and two
9
educational videos, one about contraception and one
10
about the risks of unplanned pregnancy.
There are
11 two
1-800 lines, one for Accutane information and
12 one
for contraception. In addition, there is
an
13
"avoid" blister pack pregnancy symbol and a
14
medication guide that is now packaged in the
15
blister pack that has information about the product
16 and
is patient friendly.
17
[Slide]
18
There is also a patient education brochure
19 for
men that contains product information, informed
20
consent and educational reinforcement.
21
[Slide]
22
The qualification sticker signifies that
104
1
there is a qualification date that, in this case,
2 is
the date of the last negative pregnancy test,
3 not
the date that the pregnancy test was received.
4 The
pharmacist must dispense within seven days of
5 the
qualification date and can't dispense more than
6 a
30-day supply. No refills are
allowed. Both
7
males and females have a qualification sticker
8
attached to their prescription.
9
[Slide]
10
The qualification criteria or what the
11
sticker represents on actual presentation is that
12 the
female patient has had the negative pregnancy
13
testing, two at the start of therapy and one every
14
month during therapy. In
addition, she has
15
selected and committed to use two safe and
16
effective forms of contraception.
She has signed
17
all-patient informed consent and the female
18
informed consent, and has been offered the
19
opportunity to participate in the Accutane survey
20 and
knows of its importance.
21
[Slide]
22
Again, the qualification sticker is the
105
1
actual sticker that links the dispensing of the
2 product with the negative pregnancy test. The
3
pharmacist will allow no more than a 30-day supply;
4
will dispense within seven days of the
5
qualification date or the date of the last negative
6
pregnancy test; and no refills are allowed. In
7
addition, no telephone, computerized or mail order
8
prescriptions are allowed. The
pharmacist also has
9 the
opportunity to verify that the physician has
10
been entered into the system by calling a 1-800
11
number.
12
[Slide]
13
What I would like to do now is to review
14 the
data from S.M.A.R.T. year one, or April 1 of
15
2002 through March 31, 2003. In
some cases I will
16 be
comparing these data to the year previous to
17 S.M.A.R.T. or the last year of the Pregnancy
18
Prevention Program which is April 1, 2001 through
19
March 31, 2002.
20
[Slide]
21
We used three specific data sources to
22
evaluate the S.M.A.R.T. program in year one. The
106
1
first is the prescription compliance survey; the
2
second, the Accutane survey; and, three, pregnancy
3
reports. I will be reviewing the
first two data
4 sources
and Dr. Huber will be reviewing the
5
pregnancy reports in addition to the corresponding
6
failure analyses.
7
[Slide]
8
The prescription compliance survey is a
9
quarterly survey of a random sample of pharmacies
10
pertaining to the use and completion of the
11
qualification stickers. In
addition, Roche
12
conducted a quarterly audit of anonymous Accutane
13
prescriptions from a random sample of these
14
participating pharmacies. While I
will not be
15
discussing the quarterly audit, what I can say is
16
that the results are consistent with the quarterly
17
sample of the random sample of pharmacies.
18
[Slide]
19
There is one major objective of the
20
prescription compliance survey, that is, to assess
21
prescribers' and dispensing pharmacists' compliance
22
with the appropriate use of the qualification
107
1
sticker.
2
[Slide]
3
During our discussions with the FDA, we
4 had
decided on two specific sets of metrics with
5
regard to the prescription compliance survey. The
6
first is that by the end of S.M.A.R.T. year one 90
7
percent of all physicians would use the
8
qualification stickers. The
secondary metrics
9
included that 90 percent of all physicians would
10
completely and correctly fill out the stickers, and
11
that 90 percent of all prescriptions would be
12
dispensed with a medication guide.
In October of
13
2002 Roche started packaging the medication guides
14
within the blister packs so the secondary metric is
15 no
longer applicable.
16
[Slide]
17
The results of the
prescription compliance
18
survey are as follows: Over the
six waves of the
19
survey an average of 97 percent of all
20
prescriptions had a qualification sticker affixed.
21 Of
those, 96 percent were correctly or completely
22
completed. There were no
differences between the
108
1
survey waves and there were no differences between
2 the
age of patient, the gender of patient, the
3 location
of the dispensing of the prescription or
4 the
payer type. In conclusion, we have met
and
5
exceeded our metrics for stickers and the mechanics
6 of
the stickers are working well. [Slide]
7
Now I would like to review some of the
8
high-level results from the Accutane survey.
9
[Slide]
10
As Ms. Waugh noted previously, the
11
Accutane survey was developed by Slone Epidemiology
12
Center of the Boston University School of Public
13
Health. It was initially
implemented with the
14
Pregnancy Prevention Program in 1989 and, to date,
15
Roche has had two vendors for the survey. From
16
1989 to the presentation of these data, Slone
17
Epidemiology Center was our primary research
18
organization. In October, 2002 we
switched
19
research organizations to SI International and the
20
Degge Group.
21
While I won't go into detail about the
22
methodology of the survey, and I know that Dr.
109
1
Mitchell is presenting later, what I do want to
2
note is that there are two specific arms within the
3
Accutane survey, the Accutane after treatment arm
4 and
the during and after treatment arm. The
5
presentation of these data deal only with the
6
during and after treatment arm.
In addition, I
7
would also like to note that the research
8
organization SI/Degge did implement the
9
questionnaire that was modified to include
10
components of S.M.A.R.T.
11
[Slide]
12
There are four specific objectives of the
13
Accutane survey. It was a
voluntary survey to
14
determine female patient awareness of the
15
teratogenic risks of Accutane. In
addition, it is
16
used to measure compliance with key components of
17
S.M.A.R.T., in this case informed consent, the
18
medication guide, pregnancy testing, contraceptive
19 use
and the qualification sticker.
Historically,
20 we
have used data from the Slone Epidemiology
21
Center to calculate a rate of pregnancy among
22
female Accutane users and to identify risk factors
110
1
that occur with pregnancy.
2
[Slide]
3
Again, during our discussions with the FDA
4 we
had agreed upon a variety of primary and
5
secondary metrics, the primary metric being that 60
6
percent of all women would enroll in the Accutane
7
survey by the end of S.M.A.R.T. year one. We have
8
several data specific secondary metrics including
9
female patient representativeness; recall of
10
qualification sticker; recall of pregnancy test;
11
medication guide; the use of two forms of safe and
12
effective forms of contraception; and enrollment in
13 the
Accutane survey via the prescriber's office,
14
from the blister pack or by calling a toll-free
15
number.
16
[Slide]
17
Before I go on to specific review of the
18
data, I would like to give you a high-level
19
overview of our findings. We were
successful in
20
increasing enrollment in the Accutane survey by
21
approximately 10 percentage points but missed the
22 60
percent metric.
111
1
We found that females recalled the use of
2 the
qualification sticker and that percentage was
3 almost
100 percent. In addition, almost 100
4
percent of all women knew the risks of taking
5
Accutane while pregnant and were told to avoid
6
pregnancy during Accutane.
However, they did not
7
receive the pregnancy testing or were not using
8
contraception according to the package insert.
9
With regard to the enrollment rate, we
10
calculated an enrollment rate by dividing the
11
number of enrollees by the number of new patient
12
female starts. The result for the
first year of
13
S.M.A.R.T. is 28.2 percent of enrollment of all
14
female Accutane users, which was up from 17 percent
15 in
pre-S.M.A.R.T. year one. While, again,
we
16
increased the enrollment rate, we did not succeed
17 in
meeting the 60 percent metric.
18
[Slide]
19
However, when you look at the method of
20
enrollment, we were very successful in shifting the
21
method of enrollment from the blister pack to the
22
physician's office. Again, if you
remember, the
112
1
enrollment card is within the "Be Smart, Be Safe,
2 Be
sure" educational brochure and we see this as a
3
marker that education was occurring within the
4
physician's office.
5
[Slide]
6
When we asked females at the start of
7
therapy what might Accutane do if it is taken
8
during pregnancy, and did your doctor tell you the
9
importance of avoiding pregnancy while on Accutane,
10
almost 100 percent of all women indicated that it
11
causes birth defects and that their physician told
12
them the importance of avoiding pregnancy while on
13
Accutane.
14
[Slide]
15 However, when we look at two
important
16
components of the S.M.A.R.T. program, pregnancy
17
testing and contraceptive compliance, we found that
18
only 64 percent of all women at the start of
19
treatment indicated that they had received two
20
pregnancy tests. We were
successful in reducing
21 the
women that reported no pregnancy tests from 18
22
percent to approximately 9 percent, but a large
113
1
proportion of women did not receive the pregnancy
2
testing according to the package insert.
3
[Slide]
4
When we looked at the risk category at the
5
start of treatment, we found that 50 percent of all
6 women
were not sexually active but 44 percent of
7 all
women reported some sort of sexual activity.
8
[Slide]
9
When we looked at sexual activity by use
10 of
two forms of contraception, we found that 41
11
percent of these women indicated that they were not
12
using two safe and effective forms of contraception
13 at
the start of their treatment.
14
[Slide]
15
When we looked at non-compliance with
16
contraception by age, we noticed that females 12-19
17
reported the highest percent of not sexual
18
activity. However, women 20-29,
30-39 and 40-44
19 who
were sexually active reported high levels of
20 not
using two forms of contraception, 20 percent,
21 35
percent and 40 percent respectively.
22
[Slide]
114
1
We noted previously from the prescription
2
compliance survey that a large percentage of
3
prescriptions had the sticker affixed.
In this
4
case, the percentage is similar, 97 percent of all
5
women indicated that a qualification sticker was
6
affixed to their prescription.
However, when we
7
asked them about baseline pregnancy testing,
8
receipt of two or more pregnancy tests and sexual
9
activity by using two safe and effective forms of
10
contraception, the percentages were no different
11
between those women who reported a qualification
12
sticker affixed to their prescription and those
13
women who did not. In fact, when
we look at the 22
14
cases of pregnancy, 20 of those cases of pregnancy
15
occurred in women who claimed to have a
16
qualification sticker attached to their
17
prescription.
18
[Slide]
19
Further, when we looked at these same data
20
during treatment, 21 percent of all women during
21
treatment indicated that they had not received a
22
pregnancy test. Only 63 percent
of these women
115
1
indicated that they had received two or more
2
pregnancy tests. Again, during
this time in the
3
course of their treatment they should have received
4 at
least three pregnancy tests.
5
[Slide]
6
Forty percent of all women indicated they
7
were sexually active during treatment.
However, 52
8
percent of these women indicated that during
9
treatment they were not using two safe and
10
effective forms of contraception as outlined in the
11
S.M.A.R.T. materials.
12
[Slide]
13
However again, we found that almost 100
14
percent of all women said that they had seen a
15
qualification sticker on their prescription during
16 the
course of their treatment.
17
[Slide]
18
In summary, we believe we were successful
19 in
increasing the enrollment of the Accutane survey
20 by
10 percentage points, however, we did not meet
21 the
60 percent metric set out. We increased
the
22
proportion of patients enrolling vis-a-vis the
116
1
prescriber's office. For us, that
was an
2 indication
that education is occurring within the
3
prescriber's office. And, we
believe that the
4
mechanics of the sticker are working well.
5
[Slide]
6
In addition, from the percentages in the
7
Accutane survey, women do understand the need to
8
avoid pregnancy and the consequences of becoming
9
pregnant while on Accutane.
However, there was
10
incomplete compliance with both pregnancy testing
11 and
with contraception. In fact, we found little
12
relationship between the qualification sticker,
13
pregnancy testing and contraception.
14
[Slide]
15
Dr. Huber?
16 Evaluation of S.M.A.R.T.
Program
17
[Slide]
18
DR. HUBER: I would now like to
briefly
19
review the pregnancy case reports that we have
20
received at Roche and put them in perspective--as
21 Dr.
Crawford was asking earlier, the case value
22
analysis basically.
117
1
[Slide]
2
First, I would like to go briefly through
3 the
methodology. These reports come from
multiple
4
sources. These are exposed
pregnancies that are
5
reported via either of the vendors for the Accutane
6
survey or via spontaneous reports from healthcare
7
professionals or consumers.
8
[Slide]
9
In order to compare the pregnancy numbers
10
from S.M.A.R.T. and the pre-S.M.A.R.T. year we set
11 up
the following metrics so that the numbers would
12 be
somewhat comparable. What we refer to
here as
13
pre-S.M.A.R.T. is that treatment was started so
14
isotretinoin or Accutane was started between April
15 1,
2001 to March 31, 2002. But, because
there are
16
delays in receiving some of these reports, we
17
allowed that the report was received by August 15,
18
2002. The S.M.A.R.T. data is
essentially these
19
same definitions but one year later.
20
The other issue we have to deal with in
21 the
analysis of these data is that there are
22
numerous reports that come in, in which there is no
118
1 therapy
date stated on the report. We don't know
2
when the therapy started. In
fact, when you review
3
these, in some of these it is fairly explicit that
4 the
therapy was years ago. So, we include
this
5
category of therapy start dates unknown and,
6
because we don't have a therapy start date, we
7
assign them to the period in which the report was
8
received.
9
[Slide]
10
To give some context to these
11
reports--there have been numerous questions about
12
rates, etc.--what I would like to do is remind you
13 of
the overall use of the product. First,
the
14
majority of these reports are from spontaneous
15
reporting sources, not from the survey.
Also,
16
overall Accutane use has been declining since 2000.
17 I
would like to focus on the estimated number of
18
females treated. This is female
patients in total,
19 not
just childbearing, and this is Accutane.
So,
20 you
see 278,000, 253,000, 218,000. I would
like to
21
note that generics were introduced in 2002. So,
22
when you see 2003 here, the 128,000 reflects purely
119
1 the
Accutane, not isotretinoin, data.
2
[Slide]
3
These are the pregnancy case reports we
4
have received according to the cut-offs I defined
5
earlier. For pre-S.M.A.R.T. there
was a total
6
number of 150 pregnancies; for S.M.A.R.T., 183. If
7 we
focus on those in which there is a treatment
8
initiation date known to occur in the period, it is
9
essentially 94 and 94. Where the
biggest increase
10 has
been is in this group of patients, these 89
11
with treatment initiation date unknown.
I will go
12
into a little more explanation of why we think this
13
occurred in the next few slides.
14
[Slide]
15
We think it is unlikely that the true
16
number of pregnancy case reports is a true
17
increase. In other words, we
don't believe it is
18
possible that an increased educational program,
19
with increased monitoring and with the
20
qualification sticker actually led to more exposed
21
pregnancies. Rather, as has been
noted by several
22 of
the previous speakers, in a spontaneous
120
1
environment there is a percentage of reports that
2 you
receive and a percentage you don't know about.
3 We
believe that this is most likely what is
4
occurring here and that with the first year of
5
S.M.A.R.T. we have actually seen an increased
6
proportion of reporting.
7
Why did that occur? Of note,
there was
8
increased awareness among physicians with
9
S.M.A.R.T. There was also, as Dr.
Ackermann noted,
10
increased participation in the survey.
Finally,
11
there is increased education and awareness among
12
patients.
13
[Slide]
14
To go through the details of these cases
15
now, I will start with what is the source of these
16
reports. We follow the convention
of this in
17
pre-S.M.A.R.T. with a known therapy start date;
18
S.M.A.R.T. with a known therapy start date; this is
19
pre-S.M.A.R.T. and S.M.A.R.T. with an unknown
20
therapy start date cases. This
bottom color here
21 is
those cases that came in via the Accutane survey
22
from either vendor. The green is
direct to Roche
121
1
from a healthcare professional.
This orange is
2
direct to Roche from consumers or others.
3
What you see here is that the most
4
substantial increase in number of pregnancy reports
5 is
this 10 to 33 in association with the Accutane
6
survey. Also consistent with
increased awareness
7
from consumers, while we didn't see an increase
8
here, what we did see was a substantial increase
9
from 19 to 30 of these cases coming to Roche from
10
consumers that had this unknown therapy start date.
11
[Slide]
12
When we start looking at this, as has been
13
noted, there are really two issues here.
There are
14
those patients who are pregnant prior to starting
15
Accutane therapy and then those patients who become
16
pregnant on Accutane therapy.
These data try to
17
break this down. We looked
specifically at the
18
patients who were pregnant prior to starting
19
Accutane therapy. For the
pre-S.M.A.R.T., of the
20 150
pregnancies, 28 or 19 percent occurred in the
21
pre-S.M.A.R.T. year; S.M.A.R.T. year one, 24 of 183
22 or
13 percent occurred prior starting Accutane
122
1
therapy. If we look at the number
that became
2
pregnant while on Accutane therapy, 51 percent, 41
3
percent with approximately the same numbers.
4
Patients becoming pregnant within 30 days after
5
stopping, 44 or 29 percent, 58 and 31 percent. The
6
biggest increase is in this unknown category but,
7 as
I stated earlier, this does include a large
8
number of cases that had unknown treatment
9
initiation and they also had unknown pregnancy
10
date.
11
[Slide]
12
Looking at the demographics of these
13
patients, these are the same patients, 153
14
S.M.A.R.T., 183 S.M.A.R.T., broken down by age. Of
15
note, when you look at the 16-19 group or from
16
19-29, 12-15 percent. What is
interesting is the
17 age
group 20-29 declined from 41 percent to 24
18
percent but please note that 20-29 remains the
19
largest category of patients, and 30-39 is
20
essentially similar, 16 and almost 15 percent and
21
once again a large number of unknown in S.M.A.R.T.
22
year one. The mean or median did
not shift
123
1
significantly.
2
[Slide]
3
Now coming to why are these people getting
4
pregnant, we looked for evidence of educational and
5
compliance understanding of patients.
These are
6 not
a linkage of survey data to these case reports.
7
Rather, this information is gathered as part of our
8
follow-up procedure for the pregnancy case reports.
9
The green is yes, the orange is no and the
10
light, pale color here is unknown.
What I would
11
like to do is focus on the signed female informed
12
consent, received a spiral notebook and enrolled in
13 the
Accutane survey. The axis here is the
number
14 of
pregnancy case reports that qualified for each
15
category. These are now the
S.M.A.R.T. year one
16 cases only.
17
What we see here is that only three
18
patients stated no to recall of a signed female
19
informed consent. Only five
patients stated no to
20
receiving a spiral notebook and this is in the
21
group that is our worst outcome group in that they
22 got
exposed pregnancy, and seven said no to
124
1
enrolling in the survey. So, of
those that
2
answered, the interpretation of the data is they
3 are
getting the educational materials. This
is
4
also consistent with what Dr. Ackermann talked
5
about in the survey where 99 percent of the
6
patients know they are not supposed to get
7
pregnant. They do receive the educational
8
materials.
9
[Slide]
10
The problem, as we see it, is linking it
11 to
compliance with the behaviors in the program.
12
Using the same format, this is once again
13
S.M.A.R.T. year one, the number of pregnancy case
14
reports are on this axis, two baseline pregnancy
15
tests, monthly follow-up pregnancy tests, used two
16
forms of contraception, and was the qualification
17
sticker attached.
18
I will start on the right first and 58
19
versus zero recalled the qualification sticker and
20
this is among the patients who became pregnant.
21
What is most disturbing is that 16 said no to the
22
question of baseline pregnancy tests; 8 said no to
125
1
monthly follow-up pregnancy tests; and 6 said no to
2
using two 2 forms of contraception.
So, what we
3
detect in this data is a pattern of failure to
4
comply with the educational materials that they
5
received.
6
[Slide]
7
I would like to review briefly the methods
8 of
contraception in these cases. Now we are
9
pre-S.M.A.R.T. and S.M.A.R.T. and these were
10
focusing on the 94 with the known start date in
11
both groups. Of note, 10 were
pre-S.M.A.R.T.; 11
12 of
S.M.A.R.T. were using abstinence as a primary
13
method of contraception. Of note,
of these 11
14
cases that reported abstinence, 4 did report
15 additionally
using condoms.
16
For the two forms of contraception, 17
17
pre-S.M.A.R.T., which increased dramatically to 30
18 in
the S.M.A.R.T. reports, reported using two
19
forms, one primary and one secondary.
No one
20 reported
in either year using two forms of
21
secondary contraception. With
regards to one form
22 of
primary, 18 and 18; one form secondary, 14 and
126
1
11. Unknown declined from 27 to
19.
2
[Slide]
3
To put these numbers in perspective I am
4
going to use some data from the Accutane survey.
5 Dr.
Mitchell will talk in more detail about these
6
later. But this is the one set of
data we have in
7
which we have a numerator--the number of
8
pregnancies via the survey, and a denominator--the
9
number of patients who enrolled in the survey.
10
However, this applies only to the Slone Accutane
11
survey participants. We have not
calculated this
12
rate for year one of S.M.A.R.T. in the SI because
13
there is an issue with the follow-up necessary to
14 get
the patients in and sufficient follow-up is not
15
there yet.
16
The other thing is that pregnancy rates
17 get
reported in multiple ways. So, you are
going
18 to
see some numbers potentially through the course
19 of
this day kind of flying around. I would
like to
20
show you two ways to try and help you understand.
21 One
approach has used Accutane exposed pregnancies
22 per
1,000 of the 140-day Accutane treatment
127
1
courses. Given that the normal
treatment course is
2 140
days, one approach has been to analyze the
3
exposed pregnancies by treatment courses. So, when
4 you
see the 140-day treatment course, this is what
5 we
are referring to. The other way which
you will
6 see
used is the number of Accutane exposed
7
pregnancies per 1,000 patients per year.
8
[Slide]
9
On this slide we are looking at these data
10 by
both methods. On the left vertical axis
here,
11
this is the number per treatment courses. This is
12 when we refer to the 140; zero on the bottom,
4 on
13 the
top. This is the blue line, over time
within
14 the
survey and enrollment date year 1989 to 2002
15 and
we decline from 4 down to a rate of about 2.9.
16
If you take these same data and do it by
17
number of pregnancies per 1,000 patients per year
18 you
go from this axis, over here, around 10 to
19
around--sorry, that is 2.9; the other one was 1.2.
20 I
apologize.
21
The basic message is that we have seen a
22
decline in the rate within the survey as we have
128
1
gone through a series of risk management steps.
2
That is important to remember when we discuss next
3
steps. There has been some
progress and I want to
4
make sure we don't lose that in our next
5
activities.
6
[Slide]
7
So, what do we conclude on the basis of
8
these pregnancy data? We believe
that there were
9
moderate decreases in the number of women who
10
initiated Accutane therapy while pregnant. Those
11 are
those 19 percent declining to 13 percent,
12
basically those women who were pregnant before
13
S.M.A.R.T. We think this reflects
a slight
14
improved intervention with S.M.A.R.T.
15
However, there was a relative increase in
16 the
number of pregnancies reported. We
believe
17
this is likely due to increased awareness. This,
18 in
fact, probably reflects an improved assessment.
19 You
are getting more data. But the fundamental
20
problem is that pregnancy is associated with
21
incomplete compliance with risk management
22
parameters.
129
1
[Slide]
2
The goal of the enhanced program remains
3 the
same as we started with initially. No
woman
4 who
is pregnant should receive isotretinoin
5
therapy. Our specific proposal, which
I will
6
outline for you, is that we need to further enhance
7 the
link of a negative pregnancy test to dispensing
8 of
the product.
9
With regards to the second public health
10
goal, no woman should become pregnant during
11
isotretinoin therapy, we believe this is best
12
accomplished by attempts to enhance patient
13
compliance with a behavior component, more
14
specifically, increased use of contraceptives, and
15 we
will also cover that in our new proposal.
16
[Slide]
17
Now for our recommendations, our proposal
18 is
for a single information system that provides a
19
verifiable link between a registered physician with
20 the
results of laboratory conducted pregnancy test,
21 a
registered patient including patient interaction
22
with the educational and risk management evaluation
130
1
component of the system, and a registered pharmacy
2
with a link to a product dispensed.
3
[Slide]
4
I will now try to walk you through the
5
path for females of childbearing potential. After
6 I
have completed this I will come back and walk
7 through
what we propose for non-childbearing
8
potential and males.
9
[Slide]
10
First, a potential candidate for Accutane
11 is
identified. They see a registered
physician.
12 The
physician determines if a patient is an
13
appropriate patient. They
determine if they are of
14
childbearing potential. They
perform a screening
15
pregnancy test. This can be done
in the office.
16
They educate the patient, provide them materials
17 and
they provide informed consent. In the
current
18
proposal the majority of these materials are
19
generally consistent with what is currently
20
available in the S.M.A.R.T. program.
The
21
difference is that the physician then enters this
22
information and confirmation into a central
131
1
registry. The registry, in
return, will give the
2
physician a unique patient identifier number for
3
that patient.
4
[Slide]
5
The patient then will leave the office
6
and, using their unique patient identification
7
number, interact with the system.
What they will
8 be
doing is interacting with educational and risk
9
management components.
Potentially this is
10
interactive voice recognition.
There are
11
alternative approaches we can do, but the focus is
12 on
reinforcement of compliance with the two forms
13 of
contraception that they have originally been
14 educated on by the provider.
15
[Slide]
16
There are two visits involved for a woman
17 of
childbearing potential. She comes back
to the
18
dermatologist's office, consistent with our current
19
approach, and then will have a laboratory pregnancy
20
test obtained. What we were
recommending here,
21
because of the concerns about the pregnancy testing
22
deficiencies in the current approach, is that all
132
1
women have a laboratory pregnancy test and that the
2
results of that laboratory test be entered into the
3
system.
4
At this visit there is further education
5 of
the patient with a focus on the reinforcement of
6
compliance with two forms of contraception. At
7
this time the patient receives the prescription
8
with the qualification sticker with patient ID.
9
[Slide]
10
The patient then goes to a registered
11
pharmacy that verifies the qualification sticker
12 and
verifies the treatment is authorized.
They do
13
this by calling into the registry and basically
14
receiving a yes or no, the patient is qualified or
15
not. If the patient is not
qualified, which means
16
that they have either not bee appropriately
17
registered, something has happened on their
18
interaction with the educational, or there is not a
19
negative pregnancy test in the system that falls
20 within
the prescribed dates, that patient is told
21 no
and is asked to contact the physician.
If it is
22
authorized and they provide product information,
133
1
obtain a confirmation number, the medication guide
2 is
dispensed, they dispense the medication.
3
[Slide]
4
The patient will continue to receive only
5 a
30-day supply, as noted here. Thirty
days later
6
they will loop back into this system in which they
7
will do further interaction with the educational
8
risk management component, have a laboratory
9
confirmed test, further education and receive the
10
prescription for the next 30 days.
11
[Slide]
12
For males and females of non-childbearing
13
potential it is essentially the same but the
14
requirement for the pregnancy test is out of the
15
system. It makes it a little
simpler.
16
You have the same determinant
17
qualification. You educate
patients generally on
18 the
product; the informed consent. There is
a male
19
informed consent as well. Part of
the education
20
here is focused on not sharing of the pills with a
21
female partner. The patient
receives the
22
prescription with qualification sticker.
The
134
1
patient is registered and receives an ID. The
2
reason for putting them into the system is that in
3
order to control the dispensing it is important
4
that every patient go through the same process and
5
there not be two parallel dispensing routes for
6
males and females. The registered
pharmacist does
7 the
same process. Once again, it is a 30-day
8
renewal.
9
[Slide]
10
In addition to this, there will be a
11
centralized pregnancy registry which will provide a
12
system for reporting, confirming and follow-up of
13 all
pregnancies in a uniform fashion. This
should
14
enhance our failure analysis efforts and it will
15
facilitate calculation of a risk-exposed pregnancy
16
rate.
17
[Slide]
18
How has this evolved from our current
19
S.M.A.R.T. program? This is PPP,
S.M.A.R.T., and
20
this is the proposed new program.
There is now
21
registration of patients, registration of
22
pharmacies. More important than
the actual
135
1
registration act itself is what this allows us to
2
do. What it allows is a stronger
check on the
3
prescriber because there is now this registration
4
into the system. It allows a hard
link of the
5 patient's
interaction with the system on education
6 and
risk management to dispensing. If they
don't
7
interact with the system appropriately they cannot
8 get
the product. There is also a hard link
to a
9
laboratory pregnancy test which we see as an
10
enhancement. This increases your
ability to audit;
11
potentially gives you more opportunities for
12
interaction with regards to contraceptives and,
13
finally, we will have a centralized pregnancy
14
reporting process.
15
[Slide]
16
When we consider this we see benefits and
17
challenges. From a benefit point
of view, we see
18 two
important improvements. It improves the
link
19
between dispensing and compliance with both
20
pregnancy testing and pregnancy prevention
21
activities. The interaction with
the system on the
22
educational and risk management components of this
136
1 is
important to try to address the compliance with
2 the
necessary behavior regarding pregnancy
3
prevention.
4
This should also improve data quality
5
reporting. We will now have 100
percent of the
6
patients in the system and gathering data on these
7
patients.
8
It also provides opportunities for
9
real-time patient qualification assessment linked
10 to
dispensing. The questions are asked as
part of
11 the
loop on compliance with various behaviors prior
12 to
dispensing.
13
Because this data will now be gathered on
14 all
patients on an ongoing basis, it should provide
15 us
opportunities to enhance risk management
16
evaluation activities. Various
questions exist on
17
what are the risk factors for failure.
This should
18
allow us to better address these on an ongoing
19
basis.
20
What are the challenges? One
concern is
21
that we are now interfering with the primary
22
relationship between the prescriber and the
137
1
patient. This is still the
primary basis for
2
prescribing of medications. What
we see this
3
system as is an enhancement of that relationship.
4 We
will certainly, however, have to remain cautious
5
with regards to privacy issues in this setting.
6
The other potential concern is the size of
7
this program. There are other
risk management
8
programs for pregnancy currently out there but the
9
logistics, the size and scope of this is
10
substantially larger than the other programs.
11
That, in and of itself, is not a major issue but if
12 any
of these other issues becomes a barrier to the
13
patient's access to the product, if we make it so
14
burdensome that the patients choose not to go with
15
this mechanism, the concern is that patients will
16
pursue alternative sourcing. If
patients move
17
dramatically to alternative sourcing we will
18
undermine the overall public health goal.
19
[Slide]
20
In conclusion, we propose the program
21
enhancements that establish a verifiable link
22
between a prescriber, a patient, a pharmacy, the
138
1
negative laboratory-conducted test, the patient
2
interaction with the educational and risk
3
management system and the product dispensed.
4
[Slide]
5 This should reduce the number of
women who
6 are
pregnant when they receive their initial
7
prescription. Also, the
educational component
8
would reduce the number of women who become
9
pregnant during isotretinoin therapy.
It will also
10
enhance pregnancy detection and follow-up.
11
There is one word of caution here.
If you
12
enhance your assessment, increase the proportion
13
that you find because 100 percent of the patients
14 are
in the system, the number of pregnancies
15
reported--not that occur but that are reported--may
16
initially increase.
17
[Slide]
18
I would like to put this now into
19
perspective of the FDA's risk management model to
20
kind of get some guidance on where we are in this
21
process. We have identified the
issues. We see it
22 as
two major issues for this committee today, those
139
1
patients who are pregnant before receiving an
2
isotretinoin prescription and those patients who
3
become pregnant during they.
4
We assessed the risks and benefits.
We
5
have gone through the benefit of this product. It
6 is
an essential product for which there is not an
7
alternative therapy. However,
there is a risk
8
which is specific to a subpopulation of the
9
patients.
10
We have identified and analyzed options.
11 We
have proposed to you today our recommended
12
strategy. Assuming that is
implemented, then we
13
will need to evaluate the results.
14
Today we sit here, in this middle circle,
15
engaging you and your advice on how we can do this
16 in
a better fashion. I would like to thank
you for
17
your time and attention. If I
understand it
18
correctly, we are going to do questions after the
19
other manufacturers. Thank you.
20
DR. GROSS: Thank you very
much. I am
21 going to pass the chair now to Dr. Stephanie
22
Crawford. I have to step out for
about an hour.
140
1 She
will introduce the generic firms'
2
presentations. Thank you.
3
DR. CRAWFORD: Thank you. Dr. Gross seems
4 to
think I am going to relinquish the chair back to
5 him
when he returns.
6
[Laughter]
7
At this point in the program, we welcome
8 the
opportunity to hear from the generic firms'
9
presentations and the three speakers will be Dr.
10
Frank Sisto, Dr. Allen Mitchell and Mr. Robert
11
Pollock.
12 Generic Firms Presentation
13
Isotretinoin Risk Management Program, Background
14 Information
15
MR. SISTO: Good morning.
16
[Slide]
17
My name is Frank Sisto and I am Vice
18
President of Regulatory Affairs for Mylan
19
Laboratories, which is one of the companies that
20 are
currently involved in the marketing of generic
21
isotretinoin in capsules.
22
The presentation which we have put
141
1
together for you today from the generic companies
2
involves three parts. The first
part, which I will
3
give, provides some background information
4
regarding the first year of marketing for the
5
generic isotretinoin products, including
6 identification
of the pregnancy cases that have
7
been reported to the various generic companies
8
which we have then subsequently reported to FDA.
9
The second presentation will be given by
10 Dr.
Allen Mitchell, from the Slone Epidemiology
11
Center at Boston University. Dr.
Mitchell will
12
provide information with regards to the voluntary
13
isotretinoin survey that has been conducted for the
14
various generic companies from December, 2002 when
15 the
first generic product came on the market,
16
through December, 2003.
17
The third part of our presentation will be
18
given on behalf of all the generic companies by Mr.
19
Robert Pollock. Mr. Pollock will
present three of
20 the
elements of a proposed enhanced risk management
21
program for which we are interested in getting some
22
additional input from the committee members.
142
1
[Slide]
2
With regard to my presentation, I would
3
like to first reiterate the fact that there are
4
currently three generic isotretinoin products
5
approved and marketed which are therapeutic
6
equivalents to Accutane. Amnesteem,
which was
7
approved under an ANDA from Genpharm and is
8
marketed by Mylan and Bertek Pharmaceuticals, was
9
approved in November of 2002 and was first marketed
10 in
December of 2002, a little over a year ago.
11
Sotret, which is the generic isotretinoin capsule
12
product from Ranbaxy Pharmaceuticals, was first
13
marketed in March of 2003, and Claravis, which is
14 the
generic isotretinoin capsule product from Barr
15
Laboratories, was first marketed in May of last
16
year.
17
It is also important to note--and you have
18
heard this a couple of times today in
19
presentations--that although the risk management
20
programs for each of these products has a different
21
name, and that was an issue which had to do with
22
trade names and copyrights which is the reason for
143
1
that, all of these risk management programs are
2
equivalent and substitutable or interchangeable and
3
they are equivalent to the S.M.A.R.T. risk
4
management program that was approved by FDA and
5
implemented by Hoffmann-La Roche in early 2002.
6
As the risk management program is
7
considered by FDA to be part of labeling, it is
8
required that these programs be the same as for the
9
Accutane or innovator product, and that FDA find
10
these components acceptable as a condition of
11
approval for the generic products.
12 [Slide]
13
Since I am going to be talking about the
14
number of pregnancies that have been reported to
15 the
various generic companies and subsequently
16
reported to FDA, I wanted to put that in
17
perspective by providing a little information with
18
regards to the prescriptions dispensed by each of
19 the
generic companies since the marketing has
20
begun.
21
Going back to December of 2002, which was
22 the
first month when the first generic product,
144
1
which is Amnesteem, came on the market, there are
2
approximately 104,000 prescriptions dispensed per
3
month. This went down to about
91,000 in the time
4
frame around August of '03 and in December of '03
5
when all three generic products were on the market
6 the
total number of prescriptions was around
7
117,000. Of those prescriptions
in December of
8
2003 for Accutane from Hoffmann-La Roche and
9
Amnesteem from Bertek Pharmaceuticals, there were
10
about 43,000 or 45,000 prescriptions a month. For
11 the
Claravis product from Barr, there were about
12
17,000 prescriptions per month and for the Sotret
13 product
from Ranbaxy there were about 8,500
14
prescriptions per month.
15
[Slide]
16
In terms of market share, it is important
17 to
see that within the first 4 months of marketing
18 of
the Amnesteem product, the first generic product
19
that was approved in December of '02, Amnesteem had
20
acquired about 42 percent of the prescriptions in
21
April of '03 and at the end of December of '03 this
22
number had increased--well, 60 percent of the total
145
1
prescriptions in December of '03 were dispensed for
2 the
generic isotretinoin products. Of that,
a
3
little under 40 percent was for Amnesteem, about 14
4
percent was the Claravis product and about 8
5
percent was the Sotret product from Ranbaxy.
6
[Slide]
7
Just going back, about two and a half
8
years prior to any generic product being on the
9
market, I put this slide in here just to show that
10
there has been somewhat of a decrease in the
11
overall number of prescriptions over time. If you
12 go
back to April of '02, there were approximately
13
170,000 prescriptions per month at that time and
14 Accutane
was the product that was on the market.
15
There were some enhancements to the program during
16 the
months that followed and that may have, in
17
part, caused some of the difference in prescribing
18
habits and some of the decrease that was seen in
19 the
number of prescriptions dispensed on a monthly
20
basis.
21
In January of '02 the S.M.A.R.T. program
22 was
implemented and, in April of '02, that program
146
1
became mandatory. Again, these
enhancements may
2
have also affected in some way the prescribing
3
habits for isotretinoin and caused somewhat of a
4
decrease in the overall prescriptions per month.
5 As
previously reported, when all the generics were
6 on
the market at the end of December of '03 the
7
prescriptions were approximately 113,000 per month.
8
[Slide]
9
Now, with regards to the number of
10
pregnancies that have been reported to the generic
11
companies, which we have subsequently reported to
12 the
FDA, there has been a total of 19 pregnancies
13
reported since the first generic came on the market
14 in
December of '02 through February 5, 2004.
Of
15
these, 18 have been reported through Bertek
16
Pharmaceuticals, the Genpharm Amnesteem product.
17 One
has been reported by Barr Laboratories for the
18
Claravis product and zero have been reported for
19
Ranbaxy. Of the 18 that were
reported by Genpharm,
20 9
of those were captured in the Accutane survey
21
which Slone is conducting and 9 were reported
22
directly to the company.
147
1
[Slide]
2
With regards to the sorts of reports, as
3
indicated, nine of those were reported through the
4
Slone survey. Nine had a known
therapy start date.
5
Three were reported directly to the company by
6
healthcare professionals. They
also had known
7
therapy start dates. The six that
were reported
8
directly to the company either by consumers or
9
others had a known therapy start date, and there
10 was
one reported that did not have a known therapy
11
start date.
12
[Slide]
13
With regards to the timing of exposure to
14
isotretinoin therapy relative to pregnancy, it was
15
found that three patients were pregnant when
16
isotretinoin was started, when they started
17
isotretinoin therapy. Six
pregnancies occurred
18
after the start of isotretinoin treatment. Five
19
occurred within 30 days of the completion or
20
stopping of isotretinoin treatment.
Three occurred
21
later than 30 days after completion or stopping of
22
isotretinoin treatment. Although
this is outside
148
1 of
labeling and does not need to be reported, they
2
were reported to us and we are just including them
3
here for completeness. Also, two
of the timing of
4
exposures relative to the pregnancy were unknown.
5
[Slide]
6
With regards to the timing of exposure
7
relative to pregnancy outcome, of the three
8
patients that were pregnant when they started the
9
isotretinoin treatment, one of those was lost to
10
follow-up and two were terminated in therapeutic
11
abortions. Of the six where
pregnancy occurred
12
after the start of treatment, one of those
13
pregnancies is still ongoing; four ended in
14
therapeutic abortion and one was unknown. Of the
15
five pregnancies that occurred within 30 days of
16
completion or stopping of isotretinoin treatment,
17
three were lost to follow-up, one is ongoing and
18 one
ended in therapeutic abortion. Of the
three
19
that occurred greater than 30 days after completion
20 or
stopping of treatment, two are ongoing and one
21
ended in therapeutic abortion and two are still
22
unknown.
149
1
[Slide]
2
With regards to the pregnancy outcome
3
versus offspring status, the only thing I really
4 wanted
to mention here is that there have been no
5
deliveries to date so there is really nothing to
6
report with regards to the offspring status.
7
[Slide]
8
With regards to the age group of the
9
female patients for which pregnancies were
10
reported, 2 were in the age range of 16-19; the
11
majority, 11, were in the age range of 20-29; 3
12
were in the age range of 30-39 and 3 had unknown
13
ages. Both the mean and median,
which corroborates
14
very well with the information provided by
15
Hoffmann-La Roche, was around 25 years old and the
16
range was 17-39 years.
17
[Slide]
18
The last thing that I wanted to mention
19 has
to do with the prescription compliance survey
20 or
the prescription audit that was conducted.
21
Mylan/Bertek has conducted this audit, the only
22
company to date since we have been on the market
150
1 the
longest, and we conducted this audit in March
2 of
'03 and assessed isotretinoin prescriptions from
3
April, 02 through December 31, '02.
4
What is important to note here is that in
5
reality the only product that was on the market at
6
that time was Accutane, except for the last two
7
weeks of December. So, this
really provides
8
information with the risk management program and it
9
provides a good baseline for the generics also.
10 The
objective of that program was to collect,
11
analyze and validate the data pertaining to the
12
dispensing of isotretinoin prescriptions under the
13
current risk management program.
14
[Slide]
15
The primary endpoint for the survey or the
16
audit was to determine the total number of
17
stickered prescriptions, those prescriptions having
18 a
yellow qualification sticker, in the total pool
19 of
evaluable isotretinoin prescriptions.
This
20
survey was done for us by Express Script, Inc., or
21
ESI.
22
The secondary endpoint was to determine
151
1 the
total number of correctly completed
2
isotretinoin stickered prescriptions in the total
3
pool of evaluable prescriptions containing yellow
4
stickers. Those were stickers
where the
5
male/female box was appropriately checked and which
6 had
the appropriate or specific date on them.
7
[Slide]
8
With regards to the results from the
9
survey, 13,510 prescription-specific surveys,
10
meaning information that was based on a review of
11
actual prescriptions and not from memory
12
representing 2,939 pharmacies, were returned. Of
13
these, 96 percent reported that a yellow
14
qualification sticker was present on the
15
prescription, which was the primary objective of
16 the
survey. The audit and validation process
with
17
regards to this part of the survey revealed that 96
18
percent of the responses were correctly answered.
19
For the secondary objective, the survey
20
revealed that a little over 97 percent of all
21
prescription-specific surveys with a qualification
22
date were filled out properly.
The audit and
152
1
validation process for this part of the survey
2
revealed that 96 percent of the responses were
3
correctly answered. This very
much corroborates or
4
compares with the data which has been provided for
5 the
survey which was conducted for Hoffmann-La
6
Roche.
7
With that, I would like to turn it over to
8 Dr.
Mitchell who will present information on the
9
isotretinoin survey that he is conducting.
10 Isotretinoin Survey
11
DR. MITCHELL: Thank you.
12
[Slide]
13
It is a pleasure to be here this morning,
14 as
it has been over the past 14 years before FDA
15
advisory committees. I do feel
the necessity to
16
make a disclosure, which is that I am a special
17
government employee, which in English means that I
18 am
a consultant to the FDA, but that I am not
19
engaged in any way in consultation with the FDA on
20
these matters.
21
We will be presenting the results of the
22
isotretinoin survey, by which I mean the survey
153
1
conducted for the generic manufacturers.
2
[Slide]
3
The survey extends the design, as noted
4
before, that the Slone Epidemiology Center
5
developed in 1989 for the Accutane survey under the
6
sponsorship of Hoffmann-La Roche.
We conducted the
7
Accutane survey until July of 2003.
Through that
8
point we had enrolled approximately 592,000 women
9 in
that survey.
10
I should point out that the FDA, in the
11
briefing materials, has reviewed some of the
12
earlier survey findings, some of which we would
13
concur with and others we would not.
The schedule
14 for
this morning does not provide us an opportunity
15 to
respond to those critiques and we would welcome
16 the
opportunity to do so sometime during the day.
17 If
the committee would wish, we would do that
18
briefly.
19
We have conducted the isotretinoin survey
20 for
the three generic sponsors since December of
21
2002 and we have enrolled through that period, to
22
December 2003, 8,625 women. The objectives of the
154
1
isotretinoin survey, as was the case for the
2
Accutane survey, are to assess compliance with
3
pregnancy prevention efforts and specifically, as
4 you
have heard, the awareness of the teratogenic
5
risk; patient and physician behaviors; pregnancy
6
rates; pregnancy outcomes; and risk factors for
7
pregnancy. I might add, and
reinforce the point
8
that has been made earlier, that when we say
9
pregnancy rate we mean a meaningful pregnancy rate
10
based on an identifiable and reliable denominator.
11
[Slide]
12
By way of background, I think it is useful
13 to
keep in mind that in the early phases when there
14 was
a single product we observed in our survey
15
pregnancy rates in 1989 of approximately 4/1,000
16
courses of therapy and that rate, as has been
17
noted, steadily declined, such that in the year in
18
2002 it was just a little bit over 1/1,000 courses.
19
I should also point out that we are in the
20
midst now of conducting a risk factor analysis
21
based on this very large sample of pre-S.M.A.R.T.
22
data to see if we can identify any factors that
155
1
might predict women at risk of pregnancy that might
2 not have been identified to date.
3
[Slide]
4
In 2001, in anticipation of S.M.A.R.T.,
5 the
Slone Epidemiology Center modified the survey
6
design and questionnaires with input both from
7
Roche and the Food and Drug Administration. I
8
should point out that a number of the questions
9
that were included in the questionnaire were
10
included at the request of FDA.
Also, we provided
11 the
content of the revised questionnaire to Roche
12
who, in turn, provided it to SI/Degge.
The data
13
that followed were collected for the generic
14
sponsors, post-S.M.A.R.T., using the modified
15
design and questionnaires.
16
[Slide]
17
Needless to say, there is a ramp-up period
18
that is inherent in the marketing of different
19
products. The generics were first
introduced in
20
December of '02 and we are covering essentially the
21
one-year period until December 31 of '03. The
22
number of enrollments received by quarters are
156
1
presented here. Obviously, it is
increasing each
2
quarter.
3
[Slide]
4
The method of enrollment by quarter is
5
presented in this slide. As you
can see, initially
6
virtually all enrollments came from the enrollment
7
form included in the medication package.
That was
8 a
part of the design we created back in 1989.
Only
9 a
small fraction at the outset, close to zero, came
10
through doctor-generated enrollment forms. Those
11
patterns began to shift over time and have
12
continued to shift. This is
actually what had been
13
expected in contrast to the innovator's product
14
where the physicians had in their hands and in
15
their offices the Roche enrollment forms, the
16
doctor enrollment forms. The
primary opportunity
17 for
enrollment in the generic survey really comes
18
when the patient fills the prescription and in that
19
prescription, being a generic product, finds the
20
generic enrollment form. So, as
the generic
21
enrollment forms come into greater use in the
22
doctor setting, we see an increase in
157
1
doctor-generated enrollments.
2
[Slide]
3
The DAT1 questionnaire, as has been
4
pointed out, reflects responses to the
5
questionnaire at the onset of therapy.
We will be
6
looking at the DAT1, which is at the onset of
7
therapy and DAT2, which is in the midst of therapy.
8 The
pregnancy risk categories which we established
9 at
the outset are reflected here: 5 percent of
10
women had hysterectomy or were postmenopausal; 31
11
percent were not sexually active but using birth
12
control; 23 percent were not sexually active and
13 not
using birth control; 39 percent of the sample
14
were sexually active and using birth control; and
15
only 1 percent was sexually active and not using
16
birth control. If you look at the
denominator
17
restricted to sexually active women, the 1 percent
18
becomes approximately 2 percent or 3 percent.
19
[Slide]
20
We created this photograph card in this
21
multi-product environment to help respondents
22
identify both the product that they were taking and
158
1 the
educational materials that were presented to
2
them. These become relevant in
the subsequent
3
slides.
4
[Slide]
5
The reported source of informational
6
materials received at the outset of therapy really
7 reflects
the marketplace in the first year of the
8
availability of generics, with 58 percent of the
9
women reporting that the informational materials to
10
which they were exposed were the Roche product; 18
11
percent Amnesteem; 4 percent Sotret; and 1 percent
12
Claravis; in terms of none, 9 percent of women.
13
[Slide]
14
The medication guide was reported to have
15
been received by 92 percent of women.
16
[Slide]
17
The information received in terms of
18
pregnancy prevention, as has been indicated in the
19
previous talks--99 percent reported that they were
20
told to avoid pregnancy; 80 percent read the guide
21 to
contraception; 75 percent read the contraception
22
knowledge self-assessment; 68 percent read the
159
1
emergency contraception information; only 7 percent
2
reported watching the video about pregnancy
3
prevention.
4
[Slide]
5
Forty-four percent knew about the
6
isotretinoin information telephone line; 38 percent
7
about the contraception counseling telephone line.
8
Eighty-two percent reported that the doctor
9 discussed
contraception with them, and 19 percent
10
reported that the doctor discussed emergency
11
contraception.
12
[Slide]
13
Seventy-nine percent reported signing two
14
consent forms. Recall that there
are two consent
15
forms for women, one a general consent and one
16
specific to women. Six percent
reported signing
17 one
consent form. Seven percent said they
didn't
18
sign any consent form and eight percent simply
19
weren't sure.
20 [Slide]
21
We anticipated that many women would
22
switch from Accutane to generic once the generic
160
1
drug became available. We also
expected that many
2 of
these women would enroll in the isotretinoin
3
survey during treatment prompted by the enrollment
4
form in the generic package.
5
The DAT1 questionnaire was designed to
6
assess compliance at the onset of treatment, not
7
during treatment. For that
reason, some questions
8 in
the questionnaire may be confusing to those
9
enrolling during treatment and some of their
10
responses may be inaccurate.
11
[Slide]
12
For that reason, we stratified the women
13
according to their responses, at least for the next
14 two
slides, to this question: when in the past 12
15
months did isotretinoin treatment begin?
Women who
16
said my treatment began with my most recent
17
prescription were called new users, and that is
18
roughly half. The women who said
that my treatment
19
began prior to my most recent prescription were
20
called prior users.
21
[Slide]
22
Because timing of pregnancy testing is
161
1
such a critical variable, we restricted it to the
2
category of new users. Among
those, 28 percent
3
reported 1 test; 41 percent 2 tests.
A significant
4
proportion reported 3 or more and 9 percent of
5
women reported none.
6
[Slide]
7
In terms of the timing, was the
8
pre-treatment pregnancy test properly timed
9
relative to prescription receipt, 82 percent would
10 be
considered appropriately timed; 10 percent had
11 the
test after the prescription was received; and 9
12
percent, as we said, had no test.
13
[Slide]
14
Primary contraceptive methods among the
15
nonsurgical contraceptives is presented in this
16
slide. Clearly, the overwhelming
choice of
17
contraception in all age categories is the oral
18
contraceptive, representing roughly 70 percent in
19 the
15-24 year-olds and 25-34 year-olds and still
20
close to 60 percent among the 35-44 year-old women.
21
[Slide]
22
The number of contraceptive methods among
162
1
women who were sexually active since starting
2
isotretinoin is presented here.
The yellow line
3
represents one method and there is a slight decline
4 in
the three most recent quarters, and that decline
5 is
actually reflected in an increase in the number
6 of
women reporting two methods.
7
[Slide]
8
Now I am going to talk about the
9
qualification sticker. We are
violating all rules
10 of
slide preparation quite deliberately because
11
this is a verbatim text. It
happens to be from the
12
S.M.A.R.T. program but, as has been indicated, it
13 is
standard for all programs.
14
This is what the physician is indicating
15
when he or she signs the qualification sticker. We
16 have
highlighted in yellow those portions that we
17
think are the most relevant to where we are going,
18
that a woman must have had two negative urine or
19
serum pregnancy tests with a specified sensitivity
20
before receiving the initial Accutane prescription,
21 and
that the second pregnancy test, a confirmation
22
test, should be done during the first five days of
163
1 the
menstrual period immediately preceding the
2
beginning of Accutane therapy.
3
Where the first bullet reflects pregnancy
4
testing, the second bullet reflects contraception
5 in
that the woman must have selected and have
6
committed to use of two forms of effective
7
contraception simultaneously, at least one of which
8
must be a primary form unless absolute abstinence
9 is
the chosen method or the patient has undergone a
10
hysterectomy.
11
[Slide]
12
Well, when we look at whether the sticker
13 was
present on the prescription, we see response
14
rates that are comparable to what has been seen in
15
other reports including the pharmacy audit, and 94
16
percent of women reported seeing a sticker on their
17
prescription. Only 2 percent were
clear that there
18 was
no sticker. Then, there was about 4
percent
19 who
couldn't be sure.
20
[Slide]
21
If we look at the presence of the sticker
22 as
a reflection of the patient-reported behaviors,
164
1 and
I do stress this is patient-reported behaviors
2 not
necessarily the truth; there may be some
3
misclassification, but among the women who reported
4 a
qualification sticker on their prescription, 68
5
percent reported that they were using two or more
6
forms of contraception. Among the
women without a
7
sticker, 49 percent; 51 percent of women without
8 stickers
reported that they were using less than
9
compliant forms of contraception; and 32 percent of
10 the
women with a sticker also reflected that they
11
were non-compliant.
12
[Slide]
13
When we look at the pregnancy testing, and
14
particularly the pregnancy testing and timing
15
requirements rolled into one, which is what the
16
qualification sticker...
17
[Pause for technical difficulties]
18
DR. CRAWFORD: While we are
waiting for
19 the
audiovisual equipment, I would just like to
20
state that you had mentioned a desire to provide a
21
critique. If you can write up a
critique and make
22
sufficient copies for the joint committee and the
165
1 FDA
staff representatives and give them to the FDA
2
staff by early tomorrow morning it will be
3
distributed to the committee for consideration.
4
DR. MITCHELL: If the committee
were
5
interested, I would welcome the opportunity to take
6 ten
minutes at some point to present it as well as
7 to
provide you copies.
8
DR. CRAWFORD: The chair tomorrow
will
9
decide on that. Thank you.
10
DR. MITCHELL: Okay, we would be
happy to
11 do
that. Thank you. Someone made a comment and
12
"DNK" is "do not know." I am sorry for not
13
clarifying that.
14
I had just put this slide on when the
15
goblins turned the computer off.
Remember, we were
16
talking about how the qualification sticker
17
reflects patient-reported behaviors.
We looked at
18
contraception practices in the previous slide.
19
In this slide we are looking at compliance
20
with pregnancy testing, and particularly with the
21
timing requirements which are obviously not simple.
22
This slide would suggest that only 26 percent of
166
1 the
women reported compliance with the timing of
2
pregnancy testing, no different from the women who
3
didn't have a sticker. These data
would suggest
4
that the sticker itself is not a very accurate
5
predictor of compliance.
6
[Slide]
7
Now focusing on the DAT2 or the
8
interaction with patients in the midst of therapy,
9 96
percent, a high rate, continue to report the
10
presence of a qualification sticker on their last
11
prescription and 86 percent reported receiving a
12
medication guide with their last prescription.
13
[Slide]
14
Interestingly, the behaviors prompted by
15
information in the medication guide as reported by
16 the
women include changing their contraception
17
method in 1/8; deciding not to have sexual
18
intercourse with a male partner while taking
19
isotretinoin, 16 percent; decided to have sexual
20
intercourse with a male partner less frequently
21 while taking isotretinoin, 6 percent; having a
22
pregnancy test while taking the drug, 22 percent;
167
1
requesting more information on the drug, 5 percent;
2 and
requesting more information on contraception, 2
3
percent.
4
[Slide]
5
The pregnancy risk category comparison
6
between the beginning of therapy and the midst of
7
therapy is done for a couple of reasons, not the
8 least
of which is to try to obtain some reassurance
9
that things aren't deteriorating once the initial
10
education has been completed. Not
surprisingly,
11 the
proportion of women who had a hysterectomy or
12
were postmenopausal had not changed much, but there
13 was
some decline in the women who were not sexually
14
active but using birth control. I
would point out
15
that that decline was, to some extent, made up by
16 an
increase in the proportion of women who were
17
using birth control who had become sexually active.
18
Again, the concern is that a woman who
19
declares that she is not sexually active at the
20
outset of therapy may become sexually active and it
21
would appear that that does happen, but those women
22 had
chosen to use birth control from the beginning
168
1 of
therapy and, indeed, if you look at the small
2
proportion we had shown before of sexually active
3
women who were not using birth control, that has
4 not
changed from the beginning to the midst of
5
therapy.
6
[Slide]
7
How many pregnancy tests in the past two
8
months were reported by women still taking the drug
9 at
DAT2? One test by 12 percent; 50 percent
10
reported 2 tests; and 13 percent reported no test.
11
[Slide]
12
Dermatologists accounted for 93 percent of
13 the
prescriptions; 2 percent by general
14
practitioners; 3 percent by family practitioners;
15 and
less than 1 percent by other specialists.
16
[Slide]
17
In this slide we are presenting only the
18 six
total pregnancies that we identified through
19
December 31. Recall that
additional three
20
pregnancies have been identified and forwarded to
21 the
manufacturers, who presented them today.
Of
22 the
six, two women reported being pregnant at the
169
1
start of treatment; four women reported becoming
2
pregnant during treatment, for a total of six.
3
We present the number of women completing
4 the
DAT1 and the number of women completing the
5
DAT2 not to suggest that it would be a simple
6
matter to calculate the pregnancy rate, and we
7
would be happy to talk about the difficulties in
8
deriving a meaningful pregnancy rate, but we
9
present them here just to give you a sense of the
10
numbers that we are dealing with.
Clearly, more
11
time needs to go by, we estimate 18 months from
12
enrollment of the women in treatment until you will
13
really know the pregnancy rate.
This has been
14 reflected
by the FDA as well. If you estimate a
15
5-month course of pregnancy, a 6-month period in
16
which to identify the pregnancies--11 months--for
17 the
woman to identify the pregnancy, and another 6
18
months to gather that information and query the
19
women in detail about some of the factors related.
20
With that, I will stop and turn the podium
21
over to Bob Pollock.
22
Isotretinoin Enhanced Risk Management Program
170
1
Program Elements for which Advisory Committee Input
2 is Requested
3
MR. POLLOCK: Thanks, Allen. I appreciate
4 the
opportunity to address the committee on behalf
5 of
the generic manufacturers.
6
[Slide]
7
First I would like to thank the generic
8
manufacturers and Roche for the method in which, in
9
such a short period of time, they worked together
10 to
bring a proposal to you, and I want to stress
11
that it is a proposal. I would
like to further
12
state that the generic companies fully support the
13 16
points of a consensus agreement that is
14
described in the briefing document, and are
15
committed to support whatever position is adopted
16 by
the joint advisory committee and ultimately by
17 the
FDA.
18
I would like to stress that this is a very
19
unusual, and perhaps precedent setting, situation
20
where it is being proposed that a widely
21
distributed drug product is being brought under a
22
mandatory registry system and represents the
171
1
introduction of a more complex program on top of a
2
fairly recent change in the voluntary isotretinoin
3
risk management program.
4
The agency's directive in our initial
5
December, 2003 meeting was to evaluate the program
6 to
see how best to reduce the risk of fetal
7
exposure. In that regard, the
proposed risk
8
management program proposed requires a mandatory
9 100
percent registry of physicians and all female
10
patients, and provides a hard link to dispensing of
11
isotretinoin at the pharmacy level to females of
12
childbearing potential only when there is
13
documented evidence of a current negative
14
laboratory-based pregnancy test.
15
We request that the advisory committee
16
provide advice on certain issues in an effort to
17
help us strike a balance that will assure that the
18
elements of the proposed program, when implemented,
19
will not be too complex and burdensome as to cause
20
practitioners to abandon the therapy that for
21
certain patients is extremely valuable, and will
22
also not create a potential access problem for
172
1
patients or make it impossible for them to navigate
2
through, yet is stringent enough to significantly
3
reduce the risk of fetal exposure.
4
[Slide]
5
In that regard, we would like the advisory
6
committee to provide and discuss certain issues
7
that we think are extremely important.
One has to
8 do
with the registration of male patients in
9
addition to female patients.
Secondly, the
10
component of the patient interaction with the
11
educational and risk management evaluation
12
component and, third, to discuss for us a firmer
13
link between the registry and the pharmacist,
14
perhaps through a pharmacy registration program.
15
[Slide]
16
In relationship to the registration of
17
male patients, we would like the committee to take
18
note of the fact that our focus was to reduce the
19
risk of fetal exposure.
Approximately 50 percent
20 of
the patients, as you have seen in data presented
21
earlier, are male. We would also
like to note that
22 the
sticker program, which would continue,
173
1
differentiates between male and female patients. I
2
would also like to emphasize the fact that the
3
educational components associated with male
4
patients would continue as they are today.
5
[Slide]
6
Secondly, we would like your thoughts on
7 the
patient interaction with the educational and
8
risk management evaluation component of the
9
program. We would like you
specifically to address
10
what should be the purpose of this interaction.
11
Should it be to reinforce education, or should it
12 be
to define and enforce compliance? By
that, I
13
mean if a female patient were asked a question, how
14
many forms of birth control do you use and she
15
answered one or none, should there be a "no drug"
16
provision for an inappropriate response to an
17
interaction? And, if there should
be, how would we
18
deal with the impact on the potential interruption
19 of
treatment program with the "no drug" provision
20
confounded by a negative pregnancy test finding?
21
[Slide]
22
Lastly, in relationship to an issue
174
1
brought up by the agency in regard to a firmer link
2
between the registry and the pharmacist, we would
3
like you to address whether this is sufficiently
4 addressed
by the requirement of a hard link to a
5
laboratory-based negative pregnancy test.
6
We would like you to also consider whether
7 it
would be appropriate to revise the isotretinoin
8
label to state that it would permit dispensing only
9 if
the prescription is authorized by the registry
10 for
female patients and that the pharmacist must
11
verify patient eligibility and obtain an
12
authorization number for each prescription for a
13
female patient through an interaction with the
14
registry.
15
We would also like you to address whether
16 or
not the control factor in this regard should be
17 the
practice pharmacy or should there be a more
18
restrictive requirement that may result in a
19
restricted distribution system.
20
We appreciate very much your views on this
21 and
look forward to hearing your input.
Thank you.
22
Questions to Roche and Generic Firms from Committee
175
1
DR. CRAWFORD: Thanks to each of
the
2
speakers. At this time we would
like to open up
3 the
forum for questions from any members of the
4
committee to any of the speakers who represent the
5
sponsors. Dr. Strom?
6
DR. STROM: I have two
questions. One is
7
that we heard that in the S.M.A.R.T. system or its
8
generic equivalents mail order wasn't allowed. Are
9
there data about how much is being dispensed by
10
mail order now despite the fact that it is not
11
allowed?
12
My second question is for Dr. Mitchell but
13 I
don't know if you want me to proceed or wait for
14 the
answer.
15
DR. HUBER: I am Mary Huber from
Roche.
16 To
answer your first question, mail order is not
17
allowed.
18
DR. STROM: I understand it is not
19
allowed. My question is how much
is happening.
20
DR. HUBER: We are not aware of
any
21
dispensing via mail order.
22
DR. STROM: I mean, are you not
aware or
176
1 do
you have data? Given the increasing
proportion
2 of
dispensing nationally that is happening by mail
3
order, I am trying to get a sense of whether your
4
system is being bypassed that way.
5
MS. REILLY: Tammy Reilly. The data that
6 we
get through sourcing of prescription information
7 demonstrates
to us that there is not mail order
8
prescription coming through for the Accutane
9
product.
10
DR. STROM: Okay. The second question is
11 for
Dr. Mitchell. Obviously, it is early in
your
12 new
survey but if I did my seat-of-the-pants
13
calculations correctly, it looks like there is on
14 the
order of 60,000 prescriptions per month
15
generically being dispensed. If
you assume the
16
normal course is four months, that is about 15,000
17 new
people a month getting it. Across 12
months
18
that is about 180,000 patients who began a course
19 of
Accutane prescriptions, generic Accutane
20
prescriptions, which would seem consistent with the
21
data that FDA was presenting us in terms of total
22
numbers in the market share data we were obtaining.
177
1
What we saw in the survey was 8,600 subjects. With
2 the
survey that is looking at 8,600 out of 180,00,
3 I
wonder if you might want to speculate, if you
4
can, about how you think the people you are getting
5
data from may be different in their compliance, and
6 so
on, from the much larger number of people you
7
have been unable to get.
8
DR. MITCHELL: Clearly, that is a
critical
9
question and it is something that I would like to
10
speak to more, if I am allowed to do that, but the
11
brief answer is that the fraction of the target
12
population that is enrolled, while informative, as
13
anyone who understands epidemiology understands
14
obviously, is not necessarily itself a reflection
15 of
whether it is representative.
16
The question of representativeness is a
17 tough one to identify without having really
firm
18
information available and, as I said, I hope
19
tomorrow to be able to touch on that.
I would
20
argue that in our older data where we had large
21
numbers and an opportunity to do some comparisons,
22 we
actually found the data to be not
178
1
unrepresentative in the sense that women at higher
2
risk were not preferentially being excluded from
3 the
survey. In fact, to give the punch line,
we
4
would argue that the survey might preferentially
5
include women at higher risk, for reasons that I
6
could touch on.
7
The issue of whether we can do much at
8
this point with 8,600 women out of this universe is
9 one
that I simply can't answer because we are
10
ramping up in our enrollments, as you have seen.
11 The
first enrollment came in essentially a year ago
12 and
we have been increasing enrollment
13 substantially each quarter. I think what is
14
necessary is to do the same kind of comparison on
15 the
data coming in to the generic survey that we
16 had
the opportunity to do for the Accutane survey
17
with the previous 14 years.
18 Can it be representative? Yes, I would
19
argue it could be. Is it? I think it is early to
20
know and I actually shy away from rate estimation
21
based on such a small numerator and denominator for
22 all
sorts of reasons.
179
1
DR. CRAWFORD: Dr. Cohen?
2
DR. COHEN: I would like to get
back to
3
something that you actually brought up right at the
4
beginning of the meeting today, and that is what
5
really is behind these failures?
I am not hearing
6
that yet. I think it is great to
look at all the
7
aggregate data from surveys, etc.
That is
8
important. But I know I have
learned, and others
9 have learned over the years, that if you can
drill
10
down and really learn through root cause analysis,
11
asking at several levels why did this happen, you
12
learn an awful lot.
13
I think I am hearing some of the
14
recommendations or allusions to recommendations
15
about, for example, an IND program under which this
16
drug might be made available. I
think, no matter
17
what you do, if you don't have those reasons for
18
what is going wrong you are going to have the same
19
risk at least to some extent of pregnancy, etc.
20 So,
I really think it is about, you know, learning
21
more about what is actually going wrong that causes
22
these failures and not having two forms of birth
180
1
control, not having pregnancy tests done, etc.,
2
etc. It is not just about
stickers. It is what is
3
going on, and I think if we are going to make
4
decisions like that, at least at some point before
5 FDA
reacts to this, this morning, we need to have
6
more information.
7
DR. CRAWFORD: Thank you. We have five
8
speakers in the queue right now.
The next is Dr.
9
Kibbe.
10
DR. KIBBE: I have some questions
for
11
Roche, if somebody feels like taking them. Some of
12
them are pretty easy and some may be a little more
13
complicated.
14
Currently do you have any estimate of how
15
many countries worldwide allow the marketing of
16
your product?
17
DR. HUBER: Most countries
worldwide.
18
DR. KIBBE: Is there any country
that has
19
asked you to withdraw the product from the market?
20 One
of the suggestions here that we heard from one
21 of
our speakers was to take it off the market.
22
DR. HUBER: I don't specifically
recall
181
1 any
withdrawal in my memory.
2
DR. KIBBE: How many countries worldwide
3
have a risk management system to help control this
4
particular risk of teratological effects?
5
DR. HUBER: Most countries have
some form
6 of
risk management. It ranges from labeling
to
7
other approaches. If I could
actually have the
8
slides on, please?
9
[Slide]
10
One of the issues is we see this wide
11
variation, and what we do have in common and we try
12 to
put in place in most places is a patient
13
education component, recommendations for pregnancy
14
testing, and recommendations for contraceptive use.
15 But
how this then gets implemented into a program
16 and
the mechanisms of the distribution of the
17
product, is widely divergent. For
example,
18
something like contraception, in some countries
19
there is very well done contraceptive counseling by
20
OB/GYNs for almost every patient who gets
21
contraceptions; in other countries that doesn't
22
exist. In some countries there is
much better
182
1
certification of specializations than in others,
2
plus, we are also dealing with a different health
3 authority
in each country.
4
So, what we have taken the approach of is
5 we
have identified the key themes here, the
6
education, pregnancy testing, contraceptive usage,
7 and
then it gets adapted working with the local
8 health
authority.
9
DR. KIBBE: Do you have at your
disposal a
10
kind of a measure of which countries seem to be
11
being effective relative to the problems that we
12 are
facing? I think I would be most
interested in
13 knowing
what you see as the differences between how
14
that country handles it and the way we handle it.
15
Because, if we are going to make changes, it would
16 be
nice to see a system that works better and learn
17
from it.
18
DR. HUBER: We have the numbers
for
19
pregnancies from the various countries.
We will
20 get
those for you in a moment. But the thing
I
21
would like to do is urge caution in interpretation
22 of
these data.
183
1
[Slide]
2
First of all, I don't have the sales for
3
each of these countries broken down so this is just
4 raw
numbers of pregnancies. What you see is
that
5
there are pregnancy case reports, exposed
6
pregnancies, occurring in each of these.
One point
7 I
would like to make is that on an individual
8
country basis the actual use of the product is much
9
lower.
10
The other thing is that the tendency for
11
reporting of pregnancies in many of these countries
12 is
substantially different than in the United
13
States. Then, the thing that
makes it even more
14
complicated is that the background pregnancy rate,
15 for
example in western Europe, is lower than the
16
pregnancy rate of women in the United States. So,
17 at
the end of the day, while there are some numbers
18 and
we can look at them and these rates look
19
better, what is interesting is that many of these
20
programs are actually less restrictive than the
21
current U.S. proposal but the number of reports is
22
lower.
184
1
DR. KIBBE: Does that--
2 DR. CRAWFORD: Thank you.
Dr. Kibbe, we
3
need to move on.
4
DR. KIBBE: One more?
5
DR. CRAWFORD: Yes, the last one.
6
DR. KIBBE: Does that mean that if
we
7
wanted to have a more positive impact it might not
8 be
by directly impacting this drug but by impacting
9
some other sets of behaviors?
10
DR. HUBER: We believe the
fundamental
11
thing we need to try to impact on now is the basic
12
behavior around a patient becoming pregnant.
13
DR. CRAWFORD: Dr. Wilkerson?
14
DR. WILKERSON: A couple of
questions,
15
first of all to Dr. Huber also, what was your
16
worldwide volume of sales prior to introduction of
17 the
generic substitutions? What was the
dollar
18
amount per year for this product?
19
DR. HUBER: I don't know
off-hand. The
20
majority was the U.S. from a dollar point of view.
21
DR. WILKERSON: And within a
hundred
22
million dollars how much was that?
185
1
DR. HUBER: I don't know.
2
DR. WILKERSON: Can you get that
data?
3
DR. HUBER: I will see if I can
get that
4
number for you.
5
DR. WILKERSON: Okay. Now, the second
6
part of my question is since the introduction of
7 the
product what studies have you done
8
post-introduction to try to address this problem?
9 What kind of clinical studies have been done
to see
10
what kind of best clinical practices are available
11 to
reduce this risk?
12
DR. HUBER: I am sorry, I don't
follow
13
your question. Clinical studies?
14
DR. WILKERSON: Since the
introduction of
15 the
drug in 1982, post-introduction, what clinical
16
studies have been done to determine best practices
17 for
reducing the incidence of pregnancy?
18
DR. HUBER: Clinical studies are
not the
19 way
to address the behavioral issue. By
20
definition, clinical trials are conducted in a very
21
controlled environment. We know
that in a clinical
22
trial we can influence pregnancy avoidance,
186
1
contraception, etc. The problem
is that when you
2 go
into the marketplace there are much less
3
restriction; there is much less control.
So, what
4 we
believe is the more appropriate approach is
5
through things like the Accutane survey and now
6
through our new proposal to collect data from the
7
post-market environment and use that as a basis for
8
changes to the program.
9
DR. WILKERSON: But we have tons
of data.
10 You
still have to have best practices for community
11
application of these programs.
So, you don't have
12 any
studies in other words?
13
DR. HUBER: We do not--
14
DR. WILKERSON: In a community
setting,
15
what is the best practices--you don't have that?
16
DR. HUBER: Yes, sir.
17
DR. CRAWFORD: Dr. Whitmore?
18
DR. WHITMORE: Dr. Huber, I have a
19
question for you. With regard to
the
20
implementation of the new plan, that sounds
21
terrific. I think one thing that
needs to be
22
stated once again is that only 13 percent of people
187
1
were pregnant when they began Accutane so, in
2
effect, the new program will only affect that 13
3
percent theoretically, unless you can come up with
4
some data that your educational program will affect
5
pregnancy rates during therapy.
6
DR. HUBER: We agree that
pregnancy at the
7
start is the smallest component.
The 13 percent is
8
probably a slight underestimate because one of the
9
things we do see is pregnancies occurring in the
10
first cycle. So, you will hear
some people saying
11 we
think maybe a quarter of them are actually
12
pregnant at the beginning because some of those
13
that are occurring at the first cycle are actually
14
within five, ten days. Those are
patients who
15
probably were pregnant at the time of initiation.
16 We
can debate that. But you are right, that
is the
17
smallest subset.
18
What do we think we are going to do for
19 the
latter part? The behavior part is
definitely
20 the
hardest component. There are some
precedents.
21 I
mean, for the S.T.E.P.S. program it is a
22
different population but they do have this
188
1
interaction with the patients on education. What
2 we see this as is that the patients have one
3
educational opportunity now in the physician's
4
office. We are seeing this as an
opportunity to
5
actually supplement that. Can I
give you data that
6
says this will improve contraceptive compliance?
7 No,
I cannot give you that data today. But
we
8
think that there is broader data in other settings
9
where, if you increase the number of interactions
10 and
these other types of approaches, you can modify
11
some behaviors.
12
DR. WHITMORE: May I ask one more
question
13
about that? With regard to
behavior, we keep
14
stating that the reason for pregnancy is behavior
15 of
the women who are on this drug. Are we
certain
16 of
that? Is there evidence to support that
such
17
that those who are on Depo Provera do not get
18
pregnant?
19
DR. HUBER: We have limited data
on Depo
20
Provera. The problem with Depo
Provera is that it
21
actually aggravates the acne so it is not widely
22
used in this population. What we
see as evidence
189
1 in
the behavior, and I am answering your question
2
indirectly and I apologize, is if you look at
3
compliance with forms of contraception the current
4
data shows that the women are not complying with
5 the
two forms of contraception. We clearly
see
6
that, and we have multiple sources.
We see that in
7 the
Accutane survey data. We see that also
with
8 the
pregnancy case failures. So, we think
there is
9
room for improvement in compliance in the need for
10 two
forms of effective contraception.
11
DR. CRAWFORD: Dr. Bergfeld,
followed by
12 Dr.
Bigby.
13
DR. BERGFELD: Thank you. I have a
14
question regarding the generic presentation. It
15
implied that the inclusion of the survey in the
16
package dispensed drug was better than the
17
physician survey. Is that a
correct statement for
18 the
population that was sampled? I mean, the
19
population was relatively small.
Is that a correct
20
assumption?
21
DR. MITCHELL: I am sorry, I
apologize if
22 I
gave that impression. Could you restate
what
190
1
your concern was?
2
DR. BERGFELD: In one of your
graphs you
3
demonstrated that including the survey in the drug
4
package dispensed by the pharmacist the compliance
5 of
filling out the survey was improved with that
6
method over the physician.
7
DR. MITCHELL: No, no, that was
not the
8
intent at all. I was simply
trying to reflect the
9
source of the survey enrollment.
10
DR. BERGFELD: For that limited
11
population, that limited sample--
12
DR. MITCHELL: Well, women in the
generic
13
survey, women who receive generic drug, whether
14
they are prescribed brand Accutane or another
15
brand, when they receive from the pharmacy generic
16
drug, that is when they are most likely in the
17
early months or implementation--that is changing
18
over time but in the early months that is really
19 their
only source of an enrollment form for the
20
generic survey.
21
DR. BERGFELD: At that time the
drug
22
Accutane could be substituted with a generic and
191
1 the
information to the patient went through the
2
pharmacist and the dispensing of the survey with
3 the
drug. Is that correct?
4
DR. MITCHELL: Well, it is my
5
understanding that at that time and to this day a
6
substitution is permitted, and often encouraged by
7 the
payers but the pharmacist merely dispenses the
8
medication package for the appropriate brand. In
9
that package is an enrollment form.
In the Roche
10
brand there is an enrollment form that goes to the
11
SI/Degge folks. In the generic
brands there is an
12
enrollment form that goes to our survey.
The same
13 is
true in the materials provided to physicians.
14
Physicians have materials provided to them by Roche
15 which include enrollment in the Roche Accutane
16
survey. If they use the generic
educational
17
materials there are enrollment forms in those
18
materials that will go to our survey.
If it is
19
confusing, I apologize but it is confusing.
20
DR. BERGFELD: Thank you.
21
DR. CRAWFORD: Dr. Bigby? I would like to
22
remind the speakers to please turn off your
192
1
microphone after you have spoken.
2
DR. BIGBY: I actually have three
3
questions. But I have a question
for the chair for
4 the
moment. Is this the only opportunity
that we
5 are
going to have to ask questions?
6
DR. CRAWFORD: Oh, no. We have a full day
7
with all the representatives from the sponsors here
8
tomorrow as well.
9
DR. BIGBY: Okay. So, for my first and
10
most important question I would like a response
11
from Dr. Huber and also from Mr. Sisto.
That is,
12
what is your goal of the new proposed program? Put
13
another way, if we convened a year from now or a
14
year after the program is implemented, at what
15
number or rate of pregnancies would you consider
16 the
program unacceptable and a failure?
17
DR. HUBER: The overall goal
remains a
18
decrease in exposed pregnancies.
With regards to a
19
specific number, we are not prepared to do that
20
today. That is something that is
going to need to
21 be
discussed with the FDA and with other bodies.
22
Your input would be helpful on that.
What we are
193
1
concerned about is we have all recognized that
2 there
is an under-reporting element here. If
we
3 put
in a 100 percent assessment program the actual
4
numbers of pregnancies that are reported--not the
5
number that are necessarily current but the numbers
6
reported is likely to increase in the first year or
7 so
of the program. What we wouldn't want to
do is
8
design an effective program that we basically make
9 a
decision on, doing something negative about, on
10 the
basis of actually a much improved assessment
11
activity. So, we realize that is
the long-term
12
goal but to provide a specific number that we would
13
target for a year from now, we are not comfortable
14
with that at this point.
15
MR. SISTO: And I would agree with
what
16 Dr.
Huber said. We were given the charge of
trying
17 to
come up with something that perhaps could
18
enhance the existing program to try to reduce the
19
number of pregnancies but, again, with having 100
20
percent control over the number of patients it is
21
important to just understand that the reporting of
22
those pregnancies may go up. We
don't know how
194
1
long that may occur. Therefore,
to be able to
2
determine what an acceptable rate may be at this
3
time, I just don't think is possible.
4
DR. CRAWFORD: Before we continue,
I need
5 to
make an announcement, please. The
Federal
6
Register notice for this meeting announced that the
7
open public hearing session would take place today
8
between 11:00 a.m. and 12:00 noon.
However, three
9
registered open public hearing speakers gave their
10
presentations at an earlier time this morning. The
11
Executive Secretary has informed me that no
12
additional speakers have signed up to speak during
13
today's open public hearing session.
In the
14
interest of fairness, I would like to provide the
15
opportunity at this time for any other member of
16 the
public to speak. Seeing none, we will
continue
17
with the questions from the committee.
I do want
18 to
say that it appears that almost all the other
19
members of the committee are in the queue and we
20 will
take as many questions and comments as we can
21
before our lunch break.
22
DR. BIGBY: I wasn't done. I had three
195
1
questions.
2
DR. CRAWFORD: I am sorry, Dr.
Bigby will
3
continue, followed by Dr. Gardner.
4
DR. BIGBY: To the same two
responders,
5
what component of your new program will actually
6
result in a decrease in the number and rate of
7
pregnancies?
8
DR. HUBER: There are two things
that are
9
aimed specifically at reducing the pregnancy rate.
10 The
first is the smaller component, granted, the 13
11
percent of women who are pregnant when they receive
12
that prescription. We feel very
confident that
13
requirement of a laboratory-certified test,
14
pregnancy test, within a specified window will
15
close that for almost all patients.
So, we feel
16
very strongly that we can have an impact on the
17
majority of those.
18
[Slide]
19
What we need to look at though is that
20
other component. What we are
building on--if you
21
notice here, that was that first patient visit.
22
This is analogous to the current S.M.A.R.T. The
196
1
issue with the current S.M.A.R.T. program is what
2 it
does is it gives a one-time opportunity to the
3
physician to educate the patient but there is no
4
testing or follow-up for whether the patient
5
retained that information; have they understood
6
that information; are they complying with that
7
information. We see this
interaction with the
8
registry here as an opportunity to actually assess
9
that knowledge.
10
We think this will have an impact in two
11
ways. One, as we have heard
several times, the
12
need for more data. What is the
root cause, is I
13
believe how you stated it. What
we are looking for
14 is
are there certain specific behavioral
15
components? Are there answers to
questions? Are
16
there specific things that we can identify when we
17
gather this data on all patients on an ongoing
18
basis that tells us who are the patients at risk so
19 we
can do a specific intervention?
20
But the broader benefit is that this
21
serves as a reminder to the patient.
You have now
22
added an additional repetition of the message to
197
1 the
patient every cycle through the treatment.
2
That is what we see as the primary impact of the
3
program.
4
DR. CRAWFORD: We need to move
on. Right
5 now
I am going to ask that the committee members
6
please try to ask just one question, perhaps with a
7
quick follow through. I believe
one of the
8
sponsors wishes to reply. Dr.
Gardner, please get
9
ready.
10
MR. SISTO: The generic companies
agree
11
that the two biggest things would be the constant
12
interaction for the reaffirmation of the
13
educational component of the program, and also the
14
hard link to the pregnancy test.
Those two items
15
would be very critical to that enhanced issue.
16
DR. GARDNER: Dr. Huber showed us
a slide
17
that had two points about isotretinoin indicated
18 for
severe recalcitrant nodular acne for people who
19 are
unresponsive to conventional therapy. I
am
20
still trying to understand who is at risk here.
21 So,
my question is do you have an estimate of how
22
many people with that characteristic there are and
198
1 how
many of those are women of childbearing age.
2
Secondly, we have heard a lot about
3
behavior and I would like to ask about corporate
4
behavior. Does your company, or
any of them,
5
currently have a direct to consumer advertising
6
program for these products?
Thirdly, I think you
7
were going to respond to Dr. Wilkerson's third
8
question.
9
DR. HUBER: I may have Ms. Reilly
respond
10 to
your second question about direct to consumer
11
advertising. If I remember your
first question, it
12 was
regarding the incidence or prevalence of severe
13
recalcitrant nodular acne. As was
pointed out by
14 the
previous speaker, unfortunately, that data is
15 not
readily available for incidence or prevalence.
16 The
issue, as she pointed out, is that the ICD-9
17
code does not distinguish between moderate and
18
severe recalcitrant nodular acne.
So, when you try
19 to
go back in the databases, most of them are based
20 on
this coding so what you get is a mixture of both
21 of
those diagnoses.
22
With regard to your second question
199
1
regarding direct to consumer advertising, Ms.
2
Reilly?
3
MS. REILLY: The answer is simply
no, we
4 do
not have direct to consumer advertising for this
5
product.
6
DR. CRAWFORD: Dr. Katz?
7
DR. KATZ: I have a question to Dr.
Huber
8 and
a comment on Dr. Bigby's comment. Could
you
9
just clarify what would be involved with the
10
patient ID proposal and the interaction with the
11
registry that you are proposing?
12
DR. HUBER: The patient number
would be a
13
unique number generated by the system.
We do not
14
want to use things such as social security numbers,
15
etc. because that gets us into other issues. So,
16 our
intent is that the physician would interact
17
with the system.
18
DR. KATZ: How would that be? Telephone
19
call?
20
DR. HUBER: Well, the details of
that are
21
something we need to work on. It
would probably be
22 an
interactive voice system but there are also
200
1
web-based technologies that are doing this. It may
2 be
a mixture of the two that allows whichever would
3 be
the more convenient--
4
DR. KATZ: In other words, it
would be
5
picking up the telephone and saying this patient is
6
qualified, has had a pregnancy test--
7
DR. HUBER: Right, and there would
8
probably be a few questions they would answer and,
9
because they are a registered physician, they would
10
enter their physician code so that the system knows
11
that it is a registered physician and in response
12
back they would get a number for that patient.
13
DR. KATZ: And what about the interaction
14
with the registry? You said that
the
15
physician--you used the term health provider, but
16 the
doctor initially has the interaction with the
17
patient and that is the last time.
I heard implied
18
that that is the last time that there is real
19
interaction and then the patient would interact
20
with the registry. What would
that involve,
21
patient interaction with the registry?
22
DR. HUBER: Could I have the slide
on,
201
1
please?
2
[Slide]
3
I know this is tough to follow from a
4
distance. This is patient visit
one. This is the
5
initial visit. This is what I am
referring to
6 where the physician enters his information
and gets
7 the
patient ID number back.
8
This second interaction is outside of the
9
physician's office, or it can be outside of the
10
physician's office. This is a
patient interaction
11 via
web or telephone with the registry. The
12
registry is this amorphous box across the top here.
13
DR. KATZ: Who would initiate
that? The
14
patient?
15
DR. HUBER: Our intent at this
point in
16 time is that the patient would take their
17
identification number, call in, identify themselves
18
using their number and initiate it.
There are
19
actually technologies we are investigating which
20
would allow the system, on a periodic basis, to
21
trigger the call back. That would
be something we
22
would work on.
202
1
DR. KATZ: Why would the patient
prefer to
2
call the registry rather than call the doctor with
3 a
question? I mean, ordinarily as a
practicing
4
physician, it is not too outlandish when we finish
5 a
visit to say, by the way, I will see you in four
6
weeks. If you have any questions
before then, give
7 me
a call, and generally they do. Why is
there
8
this interaction with the registry?
9
DR. HUBER: The intent of this is
not to
10
replace the patient asking questions of the
11
physician. What the purpose of
this middle visit
12
here, and there is another physician visit here,
13 is,
shall we say for lack of a better word, almost
14
like a testing. It is ensuring
through a series of
15
interactions that the patient has understood what
16
they have been instructed on.
Clearly, we still
17
think the primary relationship is between the
18
physician and the patient and if the patient has a
19
question they should always go back to their
20
physician. The interaction with
the system is
21
merely to indicate before dispensing that the
22
patient has some understanding of what they have
203
1
been told.
2
DR. KATZ: As a physician of 34
years,
3 that
is very foreign to me.
4
DR. CRAWFORD: Dr. Katz, can you
hold it
5
until tomorrow because we have a lot of time
6
tomorrow? Thank you. Dr. Sellers?
7
DR. SELLERS: We have acknowledged
that
8 the
data from the survey is very limited, and in
9 the
materials we received prior to the meeting the
10 FDA
illustrates why this data is not generalizable.
11 So,
at best, the information that we have seen is
12
merely hypothesis generating.
That concerns me
13
because the proposed risk management programs that
14 we
are hearing about today rely heavily on the
15
qualification stickers. During
the presentations
16 by
Roche and also by the generic manufacturers we
17
heard reports of 96 percent and 97 percent where
18 the
qualification stickers were filled out properly
19 or
correctly completed. But exactly what
that
20
means was not defined. In fact,
Dr. Ackermann
21
Schiff referred to the last negative pregnancy
22
test, not whether it was a baseline or follow-up
204
1
test.
2
So, my concern is, number one, how this
3
assessment of the efficacy of the qualification
4
stickers was made and the fact that we are using
5
that as a basis for the new programs may be
6
problematic. In fact, in the
S.M.A.R.T. package,
7 and
as it related to the question I asked earlier,
8
under the qualification date--this is in the
9
S.M.A.R.T. briefing package, Table 1, on page 62,
10 the
use of pregnancy tests and Accutane
11
qualification stickers, under the qualification
12
date it is listed as that date that a sample was
13
taken for the confirmatory negative pregnancy test,
14 not
the date when a negative pregnancy test was
15
actually obtained. So, I do have
concerns over the
16
qualification process and I hope that we address
17
this further. Thank you.
18
DR. CRAWFORD: Did you want a
response
19
right now? No? Dr. Honein?
20
DR. HONEIN: Thank you. I had a question
21 for
Roche about the transition that went on in the
22
fall of 2002. If I understood
correctly, around
205
1 six
months into the S.M.A.R.T. program you
2
transitioned from Slone to the Degge Group. I was
3
wondering if you could comment on how that may have
4
impacted evaluation of the first-year S.M.A.R.T.
5
and, in particular, how long you think or you know
6
dual materials might have been on the marketplace,
7
such as in prescribers' offices or in the
8
medication packets, for those two Accutane surveys?
9 DR. HUBER: There was a transition during
10 the
period described. With regards to the
impact
11 on
pregnancies, we don't have any data that says
12
there was an impact.
13
DR. HONEIN: Sorry, I wasn't
asking about
14 an
impact on the pregnancies but an impact on
15
evaluating the first year of the S.M.A.R.T. program
16 and
how the various metrics were working.
17
DR. HUBER: With specific regards
to the
18
metrics? The primary metric failed. That was the
19 one
that 60 percent of the patients would enroll in
20 the
survey. We did not achieve 60
percent. There
21 was
an increase in survey participation with the
22
transition of the program but we still did not
206
1
achieve the 60 percent. The
metrics of the
2
Accutane survey I believe is what your question
3
was, correct?
4
DR. HONEIN: Well, I want to know
in
5
general how you think it affected your evaluation
6 of
the first year of S.M.A.R.T. to have this
7
transition about six months into that year. One
8
example that I was wondering about is for how long
9 in
some prescribers' offices did they still have
10 the
older Accutane survey S.M.A.R.T. materials and
11 not
the new ones, and how all of that was handled.
12
DR. ACKERMANN SHIFF: When we
switched we
13
were in a single source environment so the switch
14
between the packaging of the Accutane survey
15
through SEC to SI/Degge occurred in a single source
16
environment. In addition, the
switch happened at
17 the
pharmacy level where the pharmacist provided
18 new
enrollment cards with the correct enrollment
19
information.
20
DR. HONEIN: And the prescriber
enrollment
21
forms, was there any issue there?
22
MS. REILLY: When we switched to
the new
207
1
program all of the enrollment materials were
2
exchanged at the physician office by the sales
3
representatives that were working for Roche at the
4
time. So, they would go in,
basically offer them
5 new
materials and retrieve the old materials.
That
6 was
the process.
7
DR. CRAWFORD: Thank you. We are going to
8
move on now. The Executive
Secretary has told me
9
that we can go until 12:10 and we have about five
10
more speakers who have requested to speak. The
11
next ones will be Dr. Epps and then Dr. Schmidt.
12
DR. EPPS: Thank you. I will try to be
13
brief. As a pediatrician and
dermatologist, I will
14
tell you that adolescents are a unique population
15 and
surveys can be helpful but I think the survey
16
that was enclosed here--if you all haven't read it
17
very carefully--is extremely personal and explicit:
18
have you interacted with someone in the last three
19
months? Was the male fertile? Has he had a
20
vasectomy? Those are questions
that not a lot of
21
young people are going to want to answer and not a
22 lot
of adults are going to want to answer either.
208
1 So,
they sometimes may tell you what they want you
2 to
hear.
3
Also, a frame of reference would be the
4
concept of time. Sexually active
may be ever; it
5 may
be within the last two weeks. That is
also
6
true of contraceptive use. I used
two the last
7
time; I may not use any the next time.
So, at the
8
time of the survey you may be getting the best
9
information that you can get and that will
10
certainly affect the results of your survey not
11
only for the concept of time but also for honesty
12
because sometimes they will put down what they
13
think you want to hear and they will put down what
14
they think their parents want them to put down.
15
I am also very concerned about
16
confidentiality with the registry identification
17
numbers. Most people will tell
you, certainly in
18 the
D.C. area, that they don't think that
19
confidential means confidential.
20
DR. SCHMIDT: There is something
about
21
drinking the water of the Potomac that makes me
22
want to ask just definitions of simple words, and
209
1 the
one word is pregnancy. In one of our
handouts
2 a
person was considered pregnant by a urine
3
pregnancy test at home. It has
been my feeling
4
that urine pregnancy tests can have 10 percent
5
false positives if there is protein or blood in the
6
urine. I would like to ask in
these defined
7
pregnancies how are they defined as pregnancies.
8
The second thing is that I would like to
9
back up how unusual teenagers can be.
I know; I
10
have had some. I gave my son
Accutane and I almost
11 had
a heart attack when I went to the drugstore to
12 pay
for it. One of the challenges of all
this is
13 how
are we going to pay for it. I just throw
this
14 out
to the group.
15
Then, my third thing is the glossary on
16 Pub
Med that I reviewed before I came here.
It has
17
almost 140 papers on this and it sounded like we
18 are
trying to go into a registry like thalidomide.
19
And, there was a reprint that I couldn't find
20
because we don't have it in the Texas Medical
21
Center Library that says "thalidomide and
22
isotretinoin--why treat them differently?" and it
210
1 is
in Reproductive Toxicology. And, I would
like
2 to
ask if Hoffmann-La Roche or somebody could get a
3
copy of that for us to review while we are here.
4
DR. HUBER: We will certainly try.
5
DR. SCHMIDT: Thank you.
6
DR. CRAWFORD: Dr. Shapiro?
7
DR. SHAPIRO: I am not sure who I
am
8
directing these questions to, but I will be brief.
9
From an ethical and a legal perspective, clearly
10 the
most difficult aspect of everything we have
11
talked about is requiring and/or tracking
12
contraceptive behavior. That
leads me to emphasis
13 on
the informed consent process, which is more than
14 a
form as everybody knows. I am wondering
whether
15 any
analysis about the video and the other
16
available media for enhancing that process has been
17
done in terms of impact on contraceptive behavior
18 in
the group we are studying, and whether further
19
thought has been given to checking understanding
20
prior to receipt of the prescription but after the
21
informed consent process to see where we can make
22
that better and more effective.
That is one group
211
1 of
questions.
2
The second has to do with the teeth in all
3 of
this, so currently and also with respect to the
4
proposed plan what happens theoretically and what
5
happens really and what needs to happen when a
6
prescription is filled without a sticker, or a
7 physician
qualifies a patient when that patient
8
should not be qualified, and so forth.
9
DR. HUBER: With regards to your
second
10
question first, with the qualification sticker
11
there is no punitive means available if a
12
pharmacist fails to report. We
are limited
13
basically to practice of pharmacy at state level.
14 So,
if pharmacists are not complying with the
15
labeling we can simply report them.
16
One of the advantages of our proposal that
17 we
put forward today is that that would then get
18
linked back very specifically prior to dispensing
19 to
the registered pharmacist. In the
current
20
approach the pharmacist does not do a letter of
21
understanding. In the new
proposal they would have
22 to
do a letter of understanding of what their role
212
1 was
in order to get a registration number.
2
DR. SHAPIRO: Can I just ask a further
3
question? With respect to both
the doctor and the
4
pharmacy, there are examining boards in every
5
state, as you said. So, do you
report or would you
6
report or do you think you should report?
7
DR. HUBER: I am not aware of us
8
specifically reporting anybody, but I am also not
9
aware--we have seen isolated examples of errors on
10
stickers. I am not aware of any
systematic failure
11
that we have seen where we have clearly become
12 aware
of somebody intentionally, willfully,
13
continually going around the program.
So, at this
14
point in time, we think the mistakes are happening
15 in
the current system. That is why we
propose to
16
change the overall system. If we
thought it was
17
limited to one or two shall we say bad apples, I
18
think we would probably propose a different
19
approach than we did today.
20
DR. CRAWFORD: Thank you. Mr. Levin and
21 our
final question before lunch may be Dr.
22
Ringel's.
213
1
MR. LEVIN: I will be brief. I guess I am
2
confused on the logistics of the Roche proposal. I
3
thought at first you were suggesting that the
4
registry would actually receive some sort of hard
5
copy, so to speak, of the pregnancy test result.
6
Then you talk about calling in and the
7
communication with the registry is by phone and the
8 two
don't match up.
9
DR. HUBER: The primary
interaction for
10
most of the interactions is, indeed, phone or
11
web-based. The laboratory test
result, because of
12
time compatibilities--we discussed things like
13
faxing the lab slip but you start creating some
14
logistical issues. So, this is a
detail we would
15 be
happy to hear your advice on, but our thoughts
16 are
at this time that basically it would be the
17
physician entering the lab test but then there may
18 be
some follow-up. In other words, you
would have
19 to
send the form in to reinforce that it was there.
20
That is one of the details we will be discussing in
21 the
design of the system.
22
MR. LEVIN: I think it is more
than a
214
1
detail because you are hanging an awful lot of this
2
program on that issue--
3
DR. HUBER: Yes.
4
MR. LEVIN: --and it is much more
than a
5
detail; it is central to that part which you are
6
suggesting is the big improvement of the program.
7
DR. HUBER: In the registry we
want
8
evidence that a laboratory pregnancy test was
9
obtained. How that gets in there,
via a fax or via
10
some other mechanism, is something that we need--we
11 are
not sure what the best mechanism is to do that
12
today.
13
DR. KWEDER: I can follow-up on
that.
14
There will be some discussion later today and
15
tomorrow, I believe, of some of the tools that can
16 be
used. There are other programs that
employ
17
them.
18
DR. CRAWFORD: Dr. Ringel?
19
DR. RINGEL: I suppose this is a
related
20
issue, but one of the problems with the program now
21 is
patient non-compliance and one is physician
22
non-compliance. Clearly,
physician non-compliance
215
1 is
at least in part due to physician frustration
2
with the difficulty of using the system.
It is
3
important, whatever system we put in place, to do a
4
walk through and try to imagine what is really
5
going to happen. I did that a
little in my head
6
with the system that is here.
7
I will go through this briefly.
In the
8
initial visit the physician is required to register
9 the
patient during that first office visit, which
10
means that he or she has to take time off from what
11
they are doing to get on the phone to do this
12
registration. I don't know how
long it is going to
13
take. It may take a while to get
through. That is
14
going to be difficult.
15
For the patient to have the second visit,
16
that can't be scheduled because it has to be done
17
during a patient's menstrual period so that is
18
going to be an urgent call to the physician who is
19
then going to have to fit the patient in.
20
The laboratory pregnancy test won't be
21
back that same day so the physician can't possibly
22
write the prescription on the same day that they do
216
1 the
test. So, that is yet another trip or phone
2
call, or whatever.
3
When you give a patient an Accutane
4
prescription for 30 days you can't see them back in
5 30
days because they have to get it filled, plus
6 the
fact that maybe 30 days is a Sunday. So,
to be
7 on
the safe side you have to make the appointment
8 for
26, 27 days and after a while you get into what
9 I
think of as appointment drift. You keep
on
10
making it earlier and earlier and patients keep on
11
collecting more and more pills until at the end
12
they are left with extra pills that they haven't
13
taken, which is something we really don't want.
14
That is not to even mention all the other
15
real-world problems such as it snowed on my
16 appointment day, I can't get there. The physician
17 is
on vacation that week. What do we do
then?
18
College students I have found to be an enormous
19
problem; rural settings where people travel over an
20
hour to get to their physicians.
Whatever system
21 we
have, we need to walk through it and make sure
22
that this is really something that is practical,
217
1
otherwise physicians are not going to cooperate
2
and, like most biologic systems, as much as the
3
test may be very good, the biologic systems find a
4 way
to wriggle around it.
5
DR. CRAWFORD: Thank you. You may notice
6
that Dr. Gross has returned and, as I said, I
7
wasn't giving it back easily. I
want to make an
8
apology to five of our committee members who had
9
questions or follow-ups and I am going to ask that
10 Dr.
Gross place you on the list for tomorrow.
11
At this time, if FDA has no additional
12
input, we will break for lunch and I have been told
13
that the meeting will reconvene promptly again at
14
1:00 p.m. Thank you.
15
[Whereupon, at 12:15 p.m., the proceedings
16
were recessed for lunch, to reconvene at 1:00 p.m.]
218
1
A F T E R O O N P R O C E E D I N
G S
2
DR. GROSS: The first speaker is
Dr.
3
Marilyn Pitts, who is a safety evaluator for the
4
FDA. She will talk about
isotretinoin pregnancy
5
exposure: spontaneous reports one year pre- and one
6
year post-the risk management program.
7
Isotretinoin Pregnancy Exposure: Spontaneous
8
Reports One Year Pre- and One Year Post Risk
9 Management Program
10
DR. PITTS: Thank you. Good afternoon.
11
[Slide]
12
Over the next hour the Office of Drug
13
Safety will provide two presentations.
I will lead
14
with a presentation entitled isotretinoin pregnancy
15
exposures: spontaneous reports one year prior to
16 and
one year after implementation of the current
17
risk management program. I will be followed by Dr.
18
Allen Brinker who will present the isotretinoin
19
pregnancy prevention program evaluation, to include
20 an
analysis of the prescription compliance survey,
21 as
well as an analysis of the isotretinoin surveys.
22
[Slide]
219
1
My presentation is from collaborative
2
reviews by myself, Dr. Claudia Karwaski and Dr.
3
Aaron Mendelsohn of the Office of Drug Safety.
4
[Slide]
5
I will provide the objectives of the
6
presentation, the methods that we used to analyze
7 the
data, a description of the limitations to the
8
data, as well as the results of the spontaneous
9
adverse event reports of the women who were
10
pregnant while using isotretinoin.
I will focus on
11
pregnancy testing, contraceptive use and pregnancy
12 and
fetal outcomes. I will also provide drug
use
13
data and offer conclusions.
14
[Slide]
15
My objectives are to compare the
16
spontaneous adverse event reports of women who were
17
pregnant while using isotretinoin.
I will compare
18
reports received one year before the implementation
19 of
the current risk management program to one year
20
after implementation.
Additionally, I will provide
21
information concerning isotretinoin drug use during
22 the
same time periods.
220
1 [Slide]
2
We identify cases for analysis by searing
3 the
AERS database, as well as requesting the
4
manufacturers of isotretinoin to submit pregnancy
5
exposure reports. We identified
all reported cases
6 of
maternal exposure where exposure occurred during
7
isotretinoin treatment or within 30 days of
8
discontinuation of isotretinoin.
All identified
9
cases were reported by August 15, 2003.
10
We categorized cases by conception date
11
into three groups. Differences in
categorization
12
resulted in the manufacturer having different cases
13 in
their case series. The prior risk
management
14
program cases included reports of conception dates
15
from April 1, 2001 to March 31, 2002 and the
16
current risk management program included cases with
17
conception dates from April 1, 2002 to March 31,
18
2003. We placed in the unknown
category those
19
cases where the conception date was unknown or
20 could not be determined.
21
[Slide]
22
We included in our analysis 325
221
1
self-reported cases of women who were pregnant
2
while using isotretinoin. We
believe these 325
3
cases represent a fraction of what occurs and the
4
true number of isotretinoin exposed pregnancies is
5
unknown. Of the cases that
provided sufficient
6
conception date confirmation or the conception date
7
could be estimated, we analyzed 127 cases during
8 the
prior risk management program and 120 cases
9
with the current program.
Seventy-eight cases
10
lacked sufficient conception date information to
11
categorize and were, therefore, placed in the
12
unknown category.
13
[Slide]
14
This slide is a little busy. The
15
important point on this slide is the "total" column
16
where we see that the majority of pregnancy
17
exposures were reported directly to the
18
manufacturers, followed by a smaller number
19
reported to the isotretinoin surveys.
We note an
20
increase in reporting to the isotretinoin surveys
21
when the current program is compared to the prior
22
program.
222
1
[Slide]
2
Again, to conduct our analysis we reviewed
3
spontaneous adverse event reports.
Before
4
proceeding to discuss our findings, I would like to
5
describe some of the limitations of using case
6
reports. Case reports are subject
to variable
7
reporting. Reporting can be
influenced by a number
8 of
factors, including but not limited to publicity
9
surrounding the drug product, as well as the length
10 of
time a drug product has been on the market.
11
Case reports are also subject to variable quality
12 and
information as well as variable completeness of
13 the
information provided. Additionally, for this
14
review in particular the case reports lacked risk
15
management program specific information to guide
16 the
reporter in providing data. As such, the
17
experience of the women who were pregnant in our
18
case series may not represent the general
19
isotretinoin user.
20
[Slide]
21
This slide provides information concerning
22 the
age of the women who were pregnant while using
223
1
isotretinoin. There is very
little difference in
2 the
ages of the women who experienced isotretinoin
3
exposure under the current risk management program
4
when you compare to the prior program.
5
[Slide]
6 In the women who were pregnant while
using
7
isotretinoin we analyzed the timing of conception
8 in
relationship to the time of isotretinoin
9
treatment. In other words, when
did the women
10
become pregnant while using isotretinoin? In the
11
cases that provided sufficient information, we
12
found that women became pregnant throughout
13
isotretinoin treatment. A small
number of women
14
were already pregnant when they started
15
isotretinoin. For the 20 women
who were already
16
pregnant when they started, we found fewer cases in
17 the
current program when you compare to the prior
18
program. For the women who became
pregnant during
19
treatment we found no appreciable difference
20 between
the programs. Moreover, when we looked
at
21 the
individual months when women became pregnant,
22 the
largest number of women became pregnant in the
224
1
first month of treatment under both programs. We
2 did
not find any appreciable difference in the
3
number of women who became pregnant within 30 days
4 of
discontinuation.
5
[Slide]
6
In the women who were pregnant we analyzed
7 the
duration of exposure of the pregnancy to
8
isotretinoin. Of note, less than
50 percent of all
9
reports provided sufficient information to make
10
this determination. Of those
reports that provided
11
sufficient information, we see a wide range of the
12
duration of exposure, from up to three months in
13 the
prior program to up to two months in the
14
current program. Although the
range of exposure is
15
wide, the median time of exposure was the same for
16
both programs.
17
[Slide]
18
I would like to shift our focus now to
19
pregnancy testing in the women who were pregnant
20
while using isotretinoin.
Currently, the
21
isotretinoin label requires two baseline pregnancy
22
tests prior to initiation of therapy.
The first
225
1
baseline pregnancy test is a screening test that
2
should occur at the time the decision is made to
3 pursue
treatment. The second baseline test is a
4
confirmatory test that should be obtained during
5 the
first five days of the menses immediately
6
preceding initiation of treatment.
7
[Slide]
8
We reviewed the cases with pregnancy test
9
information. We found that almost
50 percent of
10
reports in both series did not provide sufficient
11
information. Of the reports that
did provide
12
baseline pregnancy test information, we found that
13 50
percent of women in both series reported having
14 any
baseline pregnancy tests. However, when
we
15
looked at adherence to the label recommendations of
16 two
baseline pregnancy tests, we found that the
17
majority of baseline pregnancy testing did not
18
adhere to the label. The majority
of women who had
19
baseline pregnancy testing only had one test in
20
both the current program and the prior program.
21
Only a small number of women had two baseline
22
pregnancy tests as recommended by the label, and
226
1
more women in the current program had two tests
2
compared to the prior program. We
didn't find a
3
significant difference between the two programs in
4 the
number of women who reported not having a
5
baseline pregnancy test.
6
[Slide]
7
Baseline pregnancy tests should prevent
8
women from receiving isotretinoin if they are
9
already pregnant at the start of treatment. In an
10
earlier slide we saw that there were 20 women who
11
were already pregnant prior to starting therapy.
12
Twelve cases were in the prior program and 7 cases
13
were in the current program. We
looked closely at
14
those cases and found that 16/20 women reported
15
having at least one baseline pregnancy test and 9
16 of
those women reported having two tests.
Of 11
17
women who reported the results of the baseline
18
pregnancy testing, we found that 8 had negative
19
test results and 3 had positive test results. We
20
also found that 14/16 women reported not having a
21
baseline pregnancy test during the first 5 days of
22
menses before starting isotretinoin as recommended
227
1 by
the label.
2
[Slide]
3
In addition to baseline pregnancy testing
4
before starting treatment, once the patient has
5
started treatment the current label requires
6
monthly negative testing for the female patients to
7
continue receiving isotretinoin prescriptions.
8
Pregnancy testing while taking isotretinoin does
9 not
prevent pregnancies but serves to limit
10
exposure of the pregnancy by facilitating detection
11 and
limiting supply of the teratogen.
12
[Slide]
13
In the cases of the women who were
14
pregnant while using isotretinoin, we reviewed the
15
case reports for any pregnancy testing during
16
treatment. Pregnancy testing
during treatment is a
17
strategy for early detection and does not
18
necessarily prevent pregnancy.
Please note that 60
19
percent of the cases in both series did not provide
20
sufficient information. Of the
cases that did
21
provide pregnancy testing information during use,
22 we
found that slightly more women in the current
228
1 risk
management program reported any pregnancy
2
testing during treatment when compared to the prior
3
program. We also found that there
was no
4
difference in either program in the number of women
5 who
reported no pregnancy testing during treatment.
6
[Slide]
7
Now we turn our attention to contraceptive
8 use
in the women who were pregnant while using
9
isotretinoin. The current
isotretinoin label
10
requires at least two forms of safe and effective
11
contraception, one of which should be a primary
12
method. Primary methods include
hormonal
13
contraceptives including tablets, injections,
14
implants, patches and vaginal rings, as well as the
15 IUD
and male and female surgical sterilization.
16
Additionally, contraception should start one month
17
prior to treatment, continue throughout treatment
18 and
continue for one month after discontinuation of
19
treatment. There are two
exceptions to the
20 contraception
requirement, if a woman is practicing
21
absolute abstinence and if a woman has had a
22
hysterectomy.
229
1
[Slide]
2
In the women who were pregnant while using
3
isotretinoin, we found that one-third of all
4
reports did not provide information concerning
5
contraceptive methods. Of the
women who reported
6
using contraception, we found that slightly more
7
women in the current program reported using any
8
method of birth control compared to women in the
9
prior program. Additionally,
there were fewer
10
women in the current program who reported
11
abstinence or not using any birth control method.
12
[Slide]
13
However, when we look at those women who
14
were using any type of contraceptive method, we
15
found that the majority of women in both programs,
16
contrary to label recommendations, used only one
17
method of contraception, with that method usually
18
being the primary form. We also
found that a small
19
number of women in both programs, as recommended by
20 the
label, used two methods of contraception.
21
[Slide]
22 A small subset of women provided
230
1
additional information concerning their use of
2
contraception. We found that
slightly more women
3 in
the current risk management program reported
4
non-adherence to contraceptive directions when
5
compared to women in the prior program.
We also
6
found that a small number of women in both programs
7
equally reported contraceptive failure.
We define
8
contraceptive failure as a pregnancy occurring even
9
when the woman adhered to directions for use of the
10
contraception.
11
[Slide]
12
We now turn our attention to pregnancy and
13
fetal outcomes.
14
[Slide]
15
Although we present this information in a
16
comparative tabular format, once pregnancy exposure
17 has
occurred we recognize that no aspect of the
18
prior or current risk management program
19
necessarily had a direct impact on these outcomes.
20
Please note that in 50 percent of all women who
21
were pregnant while using isotretinoin the outcome
22 of
the pregnancy is unknown. We include in
this
231
1
unknown category cases that are lost to follow-up,
2
cases where the pregnancy is ongoing at the time of
3 the
report, and cases where the outcome was not
4
reported. In the cases where we
do know the
5
outcome, we found that the majority of pregnancies
6
ended in termination, either elective or
7
spontaneous. We find that 29
pregnancies, across
8
this row, ended in live births.
9
[Slide]
10
Again, aspects of the risk management
11
program are not expected to impact fetal outcome
12
once exposure to isotretinoin has occurred. We
13
found that at the time of reports the majority of
14
babies born were reported as normal.
It is unknown
15 if
any of the normal babies later exhibited
16
developmental delays. We also
found that 7 babies
17
were reported abnormal, with 4/7 having
18
abnormalities in organs affected by retinoids. The
19
remaining 3/7 abnormalities reported were not
20
consistent with retinoid effects.
Again, the fetal
21
outcomes were at the time of the report and may not
22
reflect any additional developmental delays that
232
1
have been reported with retinoid exposure.
2
[Slide]
3
We also looked at isotretinoin drug use
4 for
a one-year time period before implementation of
5 the
current risk management program and a one-year
6
period after implementation. We
obtained
7
prescription utilization data from IMS Health, Inc.
8 and
AdvancePCS.
9
We used IMS National Prescription Audit
10
Plus. National Prescription Audit
Plus data
11
measures the outflow of dispensing prescriptions
12
from pharmacies to consumers in retail stores, mail
13
order stores and long-term care facilities. Just
14 to
make a note, this is the source of the data.
It
15
doesn't necessarily mean that isotretinoin
16
prescriptions came from any of those particular
17
areas but this is where the data comes from.
18
Data are obtained from a sample of
19
approximately 22,000 pharmacies in the U.S., which
20
represents approximately 45 percent of U.S.
21
prescriptions. Data from NPA Plus
are projected
22
nationally.
233
1
[Slide]
2
We also obtained data from AdvancePCS.
3
Advance PCS is a large U.S. pharmacy benefits
4
manager that covers more than 50 million patient
5
lives and over 300 million prescriptions annually.
6
Data are obtained from paid prescription claims for
7
those patients with prescription drug benefits
8 administered
by AdvancePCS.
9
[Slide]
10
There are limitations to using this data.
11 The
data do not permit a more detailed analysis of
12 the
observed trends. Additionally, national
13
estimates from IMS Health may be variable due to
14
small numbers in certain subgroups, and AdvancePCS
15
data may not be nationally representative.
16
[Slide]
17
We found that for the year following
18
implementation of the current risk management
19
program the number of prescriptions decreased by 23
20
percent compared to the previous year.
We also
21
found that the number of refills decreased from 16
22
percent to 2 percent. This is
important because
234
1 the
current risk management program does not allow
2 for
automatic refills. Additionally, we see
the
3
arrival of generics to the marketplace during the
4
current program. However, the generic
product
5
labeling is consistent with the innovator's label
6
with respect to elements of risk management. The
7
level of prescribing by dermatologists remained
8
unchanged and the percentage of women who received
9
prescriptions remained unchanged.
10
[Slide]
11
In conclusion, we found that the number of
12
prescriptions dispensed for isotretinoin decreased
13 by
23 percent following implementation of the
14
current risk management program.
We also found
15
that the number of refills decreased from 16
16
percent to 2 percent, demonstrating increasing
17
adherence with the label recommendations of no
18
automatic refills. We found that
other utilization
19
variables such as gender or prescriber did not
20
appear to be influenced by the implementation of
21 the
program.
22
[Slide]
235
1
For pregnancy exposures we found that the
2
actual number of women who were pregnant while
3
using isotretinoin did not decrease appreciably,
4
especially since there has been a modest decline in
5 the
number of isotretinoin prescriptions dispensed
6 for
the same time period. We found in the
current
7
risk management program a slight decrease in the
8
number of women who reported being pregnant prior
9 to
starting treatment. We also found that
10
pregnancies continued to occur throughout
11
isotretinoin therapy for both risk management
12
programs.
13
[Slide]
14
In the women who were pregnant while using
15
isotretinoin we found no reported difference in the
16
duration of exposure of the pregnancy to
17
isotretinoin when you compare programs.
18
Additionally, we found no improvement in baseline
19
pregnancy testing, however, we did see a slight
20
improvement reported in pregnancy testing during
21
isotretinoin treatment as well as a slight
22
improvement in the reported use of one method of
236
1
contraception.
2
[Slide]
3
However, we found that only 15 percent of
4 the
325 women adhered to the label recommendations
5 of
using two safe and effective methods of
6
contraception. Finally, we found
that of the women
7 who
reported contraception information, 38 percent
8
reported non-adherence to the healthcare provider's
9
instructions for use.
10
Again, we used spontaneous adverse event
11
reports to conduct our analysis.
Although
12
valuable, there are limitations to the use of case
13
reports as previously discussed.
The experiences
14 of
the women in this case series who were pregnant
15
while using isotretinoin may not be representative
16 of
the general isotretinoin user.
17
DR. GROSS: Thank you very
much. The next
18
speaker is Dr. Allen Brinker, lead medical officer
19 for
epidemiology in the FDA. He will talk
about
20
isotretinoin Pregnancy Prevention Program
21
evaluation.
22 Isotretinoin Pregnancy
Prevention
237
1 Program Evaluation
2
DR. BRINKER: Good afternoon. Like Dr.
3
Pitts, I am with the Office of Drug Safety.
4
[Slide]
5
I will be discussing this afternoon the
6
isotretinoin Pregnancy Prevention Program
7
evaluation, both the prescription compliance survey
8 and
the patient survey.
9
[Slide]
10
First I would like to recognize our
11
collaborators, Cynthia Kornegay and Parivash
12
Nourjah. I would also like to
recognize the
13
contributions of Dr. Karen Lecter and Dr. Mark
14
Avignon.
15
[Slide]
16
I will begin by describing the
17
prescription compliance survey or PCS, and then
18
move to the patient survey. In
keeping with
19
previous speakers, I will utilize the expression
20
current RMP, current risk management plan, to refer
21 to
the risk management plan for isotretinoin as
22
implemented on April 1, 2002.
238
1
[Slide]
2
Beginning now with the PCS, first I will
3
present a brief summary of the survey design and
4
major findings. I will then
discuss the major
5
methodological issues of the survey and also talk
6
about issues with the audit portion of the PCS.
7
Finally, I will discuss the major conclusions of
8 the
analysis.
9
[Slide]
10
The PCS was conducted by the Accutane
11
sponsor for the Accutane brand isotretinoin
12
prescriptions. The primary
outcome of interest is
13
compliance with sticker use, which is defined as
14 the
presence of an Accutane qualification sticker
15 on
the prescription. Secondary outcomes are
the
16
completeness of the sticker and whether or not the
17
information on the sticker is correct.
18
It should be noted that the PCS was only
19
intended to measure compliance with qualification
20
stickers, which are just one component of the
21
current RMP. Compliance with
qualification sticker
22 use
was never intended to be used as a complete
239
1
surrogate for compliance with the totality of
2
changes implemented with the current RMP.
3
[Slide]
4
Design of the PCS--the PCS is a
5
retrospective, repeated-measure study which will
6
recruit in total some 6,000 randomly selected U.S.
7
pharmacies, stratified on selected criteria so as
8 to
be representative of all U.S. pharmacies.
Data
9
collection takes place four times a year for a
10
period of two years and 750 stores are selected for
11
each data collection period or wave.
A store can
12
only be selected once and if it refuses to
13
participate it is not back into the pool of
14
available stores.
15
[Slide]
16
Results--the results of the survey
17
demonstrate a very high rate of compliance, that
18 is,
the appearance of a qualification sticker on an
19
Accutane prescription throughout the survey period.
20 The
secondary objectives of correctness and
21
completeness were also consistently above 90
22
percent. This was true for the
audit as well.
240
1
Results were consistent across age, gender and
2
payer type. Although there were
some differences
3
between rural and urban stores and pharmacies with
4
high versus low prescription volume, both high
5
volume and rural pharmacies were more likely to
6
receive prescriptions with incomplete qualification
7
stickers.
8
[Slide]
9
I will now discuss selected methodological
10
issues with the PCS first based on results of the
11
pilot study at least 450 stores or 60 percent of
12 the
sample needed to respond during each data
13 collection period to ensure an adequate
sample for
14
study power. The observed
response rates for the
15
first five survey waves ranged from 25 percent to
16 59
percent.
17
Second, during the third survey wave five
18
major retail pharmacy chains asked and were removed
19
from the survey pool. These
chains represent
20
approximately 33 percent of all retail pharmacies
21 in
the U.S. However, they may process more
than 33
22
percent of all Accutane prescriptions.
241
1
[Slide]
2
Finally, based on the pilot study and
3
available data the sponsor estimated pharmacies
4
would have an average of 2.55 Accutane
5
prescriptions for each survey wave.
If 60 percent
6 of
the pharmacies responded, this would yield 1,150
7
prescriptions available for analysis.
However,
8
overall the responding pharmacies averaged less
9
than one prescription per pharmacy, with an average
10 of
268 prescriptions selected for analysis for each
11
survey wave. Thus, the actual
number of
12
prescriptions captured was much lower than
13
expected.
14
[Slide]
15
In addition to the main survey, the PCS
16
includes an audit component in which copies of
17
Accutane brand isotretinoin prescriptions are
18
obtained from 15 percent of PCS participants for
19
purposes of comparison. While the
actual
20
recruiting process is not described for us, the
21
audit does not appear to be a random sample.
22
Without more detail on the audit the utility and/or
242
1
applicability of the audit results is questionable.
2
[Slide]
3
In conclusion, the PCS reported a high
4
rate of compliance with qualification stickers.
5
This was seen across all survey waves for both the
6
survey and the audit. Again, it
should be noted
7
that qualification stickers are just one component
8 of
the current RMP, in this case the S.M.A.R.T.
9
program which incorporated many changes.
10
[Slide]
11
The PCS did not realize sufficient sample
12
size or power to definitively address
13
generalizability of the results.
Several factors
14
contributed to this, including a lower than
15
expected response from pharmacies; low number of
16
prescriptions captured; and the loss of the five
17
largest pharmacy chains in the U.S.
These problems
18
suggest alternate study designs should be
19
considered for future studies with similar goals.
20
[Slide]
21
I will now turn to the subject of
22
isotretinoin patient surveys.
This section, as
243
1
outlined on this slide, will include review of the
2
purpose of the patient survey; its methods and
3
limitations; the survey population and its
4
generalizability to the population of female
5
isotretinoin users at large; the results of FDA
6
analyses; and summary conclusions.
To repeat, I
7
will utilize the expression current RMP to refer to
8 the
risk management program for isotretinoin
9
implemented as of April 1, 2002.
10
[Slide]
11
Purpose--patient surveys were implemented
12 in
1989 in order to assess the compliance of
13
physicians and patients with the Accutane Pregnancy
14
Prevention Program and to identify the rate of
15
pregnancy during treatment with isotretinoin. The
16
sole isotretinoin patient survey up until the fall
17 of
2002 was administered by the Slone Epidemiology
18
Group from Boston University, and is often referred
19 to
as the Slone survey. In the fall of 2002
the
20
sponsor of Accutane brand isotretinoin shifted
21
conduct of patient surveys for Accutane from Slone
22 to
another provider, Degge/SI.
244
1
[Slide]
2
Data available for this FDA review
3
included review of the Slone Epidemiology Group
4
quarterly reports for the year prior to
5
implementation of the current RMP and continuing
6
into the year following initiation of the current
7
RMP, and primary independent analysis of an
8
Accutane brand patient survey data set, conducted
9 by
Degge/SI, for the Accutane sponsor started in
10 the
third quarter following implementation of the
11
current RMP.
12
[Slide]
13
In order to address revisions included in
14 the
current RMP, patients enrolled in the Degge/SI
15 survey
received a new survey instrument. To
16
review, the old survey instrument did not ask about
17 the
presence of a qualification sticker; did not
18 ask
about the presence of a MedGuide; and asked
19
only if any pregnancy test had been performed.
20
[Slide]
21
In contrast, the new survey instrument
22
included questions about both the qualification
245
1
sticker and the date and number of pregnancy tests
2
performed. The new survey was
introduced first to
3
Accutane recipients participating in the Degge/SI
4
survey but not until the third quarter of the first
5
year of the current RMP. Degge/SI
would enroll
6 some
6,000 participants through the end of that
7
first year.
8
[Slide]
9
Now I would like to discuss some
10
limitations of the survey. The
survey is a
11
self-administered mailed survey which would be the
12 preferred
method to gather information on sensitive
13
questions. However, this
advantage is compromised
14 by
the fact that this is not an anonymous survey
15
because it is a follow-up survey.
Historically,
16 the
isotretinoin patient survey has suffered from
17 low
enrollment. Furthermore, low enrollment,
in
18
combination with the voluntary nature of the
19
survey, increases the likelihood that patients
20
participating in the surveys may be different in
21
important compliance behaviors than those who do
22 not
participate in the surveys.
246
1
[Slide]
2
A further area of concern is measurement
3
errors. These would include
recall bias since we
4 are
relying on patient memory to recall the exact
5
dates of such events, menses, receipt of a
6
prescription and start of therapy.
As highlighted
7 on
the previous slide, the lack of anonymity in
8 this survey increases the possibility of
social
9
desirability bias to sensitive questions, such as
10
those regarding sexual behavior, birth control use
11 and
accidental pregnancy. FDA review also
12
concluded the survey instrument included both
13
complex questions and complex question skip
14
patterns that might be confusing to some
15
participants. In sum, these
biases reduce the
16
generalizability or inference of the results.
17
[Slide]
18 With these concerns noted, I will now
move
19 on
to selected FDA results. First, based on
FDA
20
analysis, absolute participation in isotretinoin
21
patient surveys, including both the Degge/SI and
22
Slone epi. group surveys, increased from a range of
247
1 16
percent to 19 percent in the year prior to
2
implementation of the current RMP to a range of 22
3
percent to 26 percent in the first year following
4
implementation of the current RMP.
5
[Slide]
6
These data are shown graphically on this
7
slide. Note that the X axis is
labeled as quarters
8 one
through eight, representing the four quarters
9
before implementation of the revised RMP, which was
10 in
April of 2002, shown here by this arrow, and the
11
following four quarters. As
shown, enrollment
12
appears to have started to increase before
13
implementation of the S.M.A.R.T. program and has
14
either peaked or is increasing very slowly.
15
[Slide]
16
FDA was also able to compare two
17
demographic characteristics of patients enrolling
18 in
the Degge/SI cohort to isotretinoin recipients
19 managed by AdvancePCS and appearing within the
IMS
20
National Disease and Therapeutic Index, on NDTI.
21 As
shown in this slide, which is from the ODS
22
review, in comparison to females within both
248
1
AdvancePCS and NDTI, it appears that the youngest
2
recipients of isotretinoin, shown by this row right
3
here, are under-represented in the Degge/SI cohort.
4 So,
we are comparing 35 percent in the Degge/SI
5
cohort to 43 percent and 45 percent, which you
6
would expect--with the reciprocal increase or
7
over-representation of participants aged 20-29,
8
which is this row right here, so 30 to 28 to 38.
9
[Slide]
10
In addition, FDA analyses of
the Degge/SI
11
cohort 94 percent of participants indicated that
12 the
prescriber was a dermatologist. In
comparison,
13
approximately 80 percent of recent isotretinoin
14
prescriptions were associated with a dermatologist.
15
Thus, it appears that survey participants receiving
16
care from non-dermatologists are slightly
17
under-represented within the population
18
participating in the Degge/SI survey.
19
I will now turn from issues of
20
generalizability and representativeness to apparent
21
adherence with current labeling, first around the
22
initiation of isotretinoin therapy and then, more
249
1
briefly, during isotretinoin.
2
[Slide]
3
The current isotretinoin RMP requires
4
women to sign two consent forms, one required of
5 all
patients and the other required only of
6
females. In FDA analysis of the
Degge/SI cohort 76
7
percent of participants reported signing two
8
consent forms. Four percent
signed only one form;
9 9
percent reported signing no consent forms; and 11
10
percent were uncertain or did not answer the
11 question.
12
[Slide]
13
As noted before, the current isotretinoin
14 RMP
requires all isotretinoin prescriptions to
15
carry a qualification sticker. In
FDA analyses of
16 the
Degge/SI cohort 92 percent of participants
17
reported their prescription to carry a
18
qualification sticker. This is
consistent with the
19
findings in the PCS.
Additionally, 2.5 percent of
20
participants reported no sticker and 5.5 percent
21 did
not know or did not answer the question.
22
[Slide]
250
1
The current RMP also requires women to
2
have two pregnancy tests before initiation of
3
therapy. In FDA analyses of
apparently fertile,
4
15-45 year-old participants in the Degge/SI cohort
5 91
percent reported at least one pregnancy test; 66
6
percent reported two pregnancy tests.
Performance
7
increased only slightly to 92 percent and 68
8 percent
with restriction to sexually active
9
participants. Participation
decreases slightly, to
10 89
percent and 63 percent, with restriction to
11
participants who reported no current sexual
12
activity.
13
[Slide]
14 Review data reported by the Slone
15
Epidemiology Group suggests that the rate of any
16
pregnancy testing prior to initiation of therapy
17
with isotretinoin increased from a range of 77
18
percent to 85 percent for the year prior to the
19
implementation of the current RMP to 91 percent to
20 92
percent in the first year of the current RMP.
21
[Slide]
22
These data are shown graphically on this
251
1
slide which highlights that the improvement began
2 to
take place shortly before implementation of the
3
current RMP, shown here again by the arrow, April
4 of
2002, and appears to have plateau'd.
5
[Slide]
6
According to the revised labeling,
7
sexually active women receiving isotretinoin should
8 use
two forms of birth control, consisting of one
9
primary and one secondary method.
In FDA analyses
10 of
the currently fertile and sexually active women
11
within the Degge/SI cohort, 95 percent reported use
12 of
some form of birth control. Almost 50
percent
13
reported use of appropriate birth control,
14
consisting of one primary and one secondary method.
15
[Slide]
16
I would like to highlight that the women
17
included in the previous slide represent only a
18
minority of patients within the patient survey as
19
this analysis was restricted to sexually active
20
women. As outlined in the review,
there were in
21
total some 5,300 women who started treatment with
22
Accutane in the Degge/SI cohort.
About 600 women,
252
1 or
11 percent, reported reproductive state or
2
postmenopausal status post hysterectomy or were of
3 age
less than 15 years of age or greater than 45
4
years. These women were excluded
from most
5
analyses for pregnancy specific behaviors. Of the
6
remaining 4,596 apparently fertile women, only
7
1,806, or 39 percent, reported current sexual
8
activity. The remainder, or 61
percent, denied
9
current sexual activity. If a
woman does not
10
consider herself sexually active and that
11
represents the majority of women within this
12
survey, these women may not be prepared for
13
contraception should the need arise.
14
[Slide]
15
I will now outline selected FDA bivariate
16
analyses conducted to address the relationship
17
between the presence or absence of a qualification
18
sticker and any pregnancy testing and any birth
19
control.
20
[Slide]
21
This table outlines the relationship
22 between
qualification sticker and pregnancy testing
253
1
based on FDA analyses of the Degge/SI cohort. As
2 can
be seen in this table, the overall effect of
3 the
qualification sticker did not appear to relate
4 to
performance of a pregnancy test as testing was
5
high both in the presence and absence of a
6
qualification sticker. Of
particular interest, 9
7
percent of issued qualification stickers were, by
8
patient reports, not linked to a pregnancy test.
9
[Slide]
10
Per the current RMP, the qualification
11
sticker is intended to document that the patient
12
received education and counseling on pregnancy
13 prevention.
This table outlines the relationship
14
between qualification sticker and any birth control
15
based on FDA analyses of apparently fertile,
16
sexually active, 15-45 year-old Degge/SI
17
participants. You will note that
the N falls--it
18 was
about 4,000 in the last table--to 1,788 in this
19
analysis as it is restricted to sexually active
20
women. As was appreciated in the
previous slide,
21
there does not appear to be a strong relationship
22
between the presence of a qualification sticker and
254
1
compliance with birth control.
Birth control use
2 was
high among the strata reporting current sexual
3
activity regardless of quality of life sticker.
4
This table also highlights that around 3 percent of
5
apparently fertile and sexually active participants
6
deny the use of any form of birth control.
7
[Slide]
8
FDA also has data on participants in the
9
Degge/SI cohort during therapy with isotretinoin.
10 In
contrast to the previous slides which centered
11 on
reports from around the initiation of therapy,
12
these data are generally consistent with results
13 seen
early in therapy with isotretinoin but also
14 may
show some signs of complacency. For
example,
15
report of a qualification sticker falls from 97
16
percent around initiation of therapy to 95 percent
17
during therapy. Monthly pregnancy
testing also
18
falls from around 92 percent around initiation of
19
therapy to 81 percent during therapy.
20
[Slide]
21
Review of data reported by the Slone
22
Epidemiology Group suggests that the rate of any
255
1
pregnancy testing during therapy with isotretinoin
2
increased from about 70 percent in the year before
3
implementation of the current RMP to around 85
4
percent in the first year of the current RMP.
5
[Slide]
6
These data are shown graphically in this
7
slide which suggests that the improvement appears
8 to
have taken place shortly before implementation
9 of
the current RMP, again as noted by the arrow,
10 and
thereafter appears to have plateau'd.
11
[Slide]
12
The Degge/SI cohort of Accutane users
13
includes 15 reports of pregnancy among 4,277
14
first-time isotretinoin users.
This translates to
15 a
rate of 3.5/1,000, which is very similar to the
16
historic rate reported for participants in the
17
Slone survey of Accutane users.
It should be noted
18
that this rate is censored and so could be an
19
underestimate of the true rate for this cohort,
20
calculated when all these women finished their
21
course of isotretinoin therapy.
22
[Slide]
256
1
In addition to data by supplied by the
2
Accutane sponsor, FDA has received individual
3
quarterly reports from sponsors of generic
4
isotretinoin and quarterly reports from the Slone
5
epi. group on patients enrolling from one of the
6
three generic brands of isotretinoin.
These data,
7
including quarterly reports for the second quarter
8 of
2003 and the third quarter of 2003, are
9
generally supportive of results for the interval of
10
April 1, 2002 through March 31, 2003.
11
Specifically, reported any pregnancy testing prior
12 to
initiation of therapy continues at about 90
13
percent. Report of two or more
pregnancy tests
14
prior to initiation of therapy continues at about
15 65
percent. Report of no birth control
among
16
apparently fertile, sexually active respondents
17
continues at about 3 percent; and reports of any
18
pregnancy testing during therapy continues at about
19 82
percent.
20
[Slide]
21
To summarize the findings from both my
22
presentation and the presentation by Dr. Pitts,
257
1
isotretinoin-exposed pregnancies continued to occur
2
after implementation of the current RMP.
3
Enrollment in isotretinoin patient surveys
4
increased only modestly after implementation of the
5
current RMP.
6
[Slide]
7
Despite their wide utilization,
8
qualification stickers have been issued to patients
9 who
have not undergone pregnancy testing.
10
[Slide]
11
Lastly, the observed pregnancy rate for
12 the
Degge/SI cohort recruited following
13
implementation of the current RMP appears similar
14 to
that reported for cohorts recruited before
15
implementation of the current RMP.
16
That concludes my presentation.
17
DR. GROSS: Thank you very
much. We have
18 a
few minutes for questions. I am going to
ask one
19
myself. The irrationality of
human behavior always
20
intrigues me. Is there any
information as to why
21
women didn't get regular pregnancy tests or why
22
they didn't use two contraception measures when so
258
1
advised? Do any of the FDA or
Slone surveys
2
approach those questions?
3
DR. BRINKER: I will personally
defer that
4
question to another member of the FDA staff if they
5 want
to comment on it. That wasn't
necessarily
6
included in our review.
7
DR. TRONTELL: FDA uses the
voluntary
8
information that has been supplied--I see Dr. Pitts
9 at
the microphone; she can say if any of that was
10 mentioned
in the reports that came through the
11
spontaneous reporting system.
12
DR. PITTS: Actually, I was going
to say
13
that, no, the spontaneous reports really don't
14
guide the person in providing the type of
15
information that we wanted for this particular
16
analysis so I don't have any further information.
17 It
is somewhat incomplete in that respect.
18
DR. GROSS: Yes, the only reason I
asked
19 is
if we are going to be designing new programs or
20
making recommendations it would help to have that
21
kind of information. We have some
questions from
22
earlier. Dr. Katz?
259
1
DR. KATZ: The data that we are
hearing
2 now
of people having two pregnancy tests or one
3
pregnancy test or pregnancy tests through therapy,
4
that data is derived from patient recall on the
5
survey? Is that correct?
6
[Dr. Brinker nods]
7
Well, that has to be considered seriously
8
because when they are asked that on this very
9
complicated survey that they have gotten from the
10
enrollment, many patients--a lot is going on in
11
their life, in real life, and when you ask did you
12
have a pregnancy test many patients would say, no,
13 I
don't remember a pregnancy test, be it a urine
14
test or a pelvic examination.
They forget that the
15
blood test includes a CBC, hepatic profile, lipids
16
and, by the way, a pregnancy test.
Now, they have
17
been told that at the beginning but all they may
18
know, I would imagine, is they got a couple of
19
blood tests--yes, the doctor gets a blood test
20
every month but they didn't get any pregnancy test.
21 So,
that has to be considered in that response,
22
that human response.
260
1
DR. BRINKER: I would admit that
this is a
2
very blunt tool and the data are what the data are.
3
DR. KATZ: As a follow-up, was any
attempt
4
made by the company or FDA for the patients who
5
said, no, they didn't get any pregnancy test to get
6
follow-up from the doctor's office?
That would be
7
relatively simple in a very small pilot manner, and
8 you
might find that 98 percent or 100 percent of
9
those people not getting a pregnancy test did get
10 two
pregnancy tests and got pregnancy tests every
11 month.
12
DR. PITTS: In the spontaneous
reports
13
there was a significant number of reports that
14
didn't mention the data at all.
Of the ones that
15
specifically said they did not get pregnancy tests,
16 we
took that as affirmative, that they truly did
17 not
get it. Of the ones that mentioned that
they
18
received some pregnancy testing throughout, we took
19
that and that is the data that we had to work with
20 in
the spontaneous reports.
21
DR. GROSS: Thank you.
Dr. Gardner?
22
DR. GARDNER: I have a question
for each
261
1 of
you. Dr. Brinker, could you clarify for
me, you
2 had
said in your PCS results slide, the
3
prescription compliance survey, that both high
4
volume and rural pharmacies were more likely to
5
receive prescriptions with incomplete stickers.
6
Doesn't your methodology look at prescriptions
7
after dispensing has happened?
8
DR. BRINKER: I am going to defer
that
9
question to the primary ODS FDA reviewer of the
10
PCS, Dr. Cynthia Kornegay.
11
DR. GARDNER: My question then to
Dr.
12
Kornegay is does this imply that high volume rural
13
pharmacies went ahead and dispensed in the face of
14
prescriptions with incomplete stickers?
15
DR. KORNEGAY: No, it does
not. This is
16
merely what the pharmacy received.
There were
17
other aspects of the pharmacy compliance survey
18
that did point out how many prescriptions were
19
received but not dispensed and that was
20
consistently fairly low.
21
DR. GARDNER: Thanks. My other question
22 for
Dr. Pitts is in using the AdvancePCS data, back
262
1 to
the issue of labeling compliance that we have
2
talked about several times today, using that
3
database resource, has there been any effort to
4
determine what proportion of people getting
5
Accutane had actually had, and presumably failed,
6
other dermatology therapies before?
7
DR. PITTS: I am going to defer
that to
8 Dr.
Mendelsohn.
9
DR. MENDELSOHN: That was actually
one
10
thing that we did not work at, but it would be
11
possible actually to examine that with the
12
AdvancePCS data but we didn't consider that yet.
13
DR. KWEDER: I believe the Slone
unit did
14
address that in one of Dr. Mitchell's backup
15
slides.
16
DR. MITCHELL: I don't know who
said that
17 but
thank you. If we could go to my
presentation,
18 if
that is possible--is it still booted up there?
19
[Slide]
20
The survey asked from the outset of the
21
onset of therapy about past treatments.
The
22
implication, of course is that these are failures.
263
1 One could argue that I suppose. I can't read it
2
from here but I think it is 93 percent that
3
reported that they had previously been on an
4
antibiotic 51 percent, on Ortho-Tricylen or vitamin
5 A
10 percent, Retin-A 68 percent and so forth,
6
benzoyl peroxide and so forth. So
a substantial
7
proportion of women have tried one or more
8
therapies prior to Accutane.
9
DR. GARDNER: For the survey? Thanks.
10
DR. GROSS: Dr. Honein?
11 DR. HONEIN: Yes, my question is for Dr.
12
Brinker. When you were
calculating the survey
13
participation rate for the first year of the
14
current risk management program were you able to
15
eliminate duplicate enrollments, meaning women that
16 may
have enrolled in both the Slone Accutane survey
17 and
the Degge Accutane survey?
18
DR. BRINKER: Well, remember, that
only
19
kicked in during the last two quarters so that is
20 the
only place that it is impacted, and we took our
21
numerator data from the sponsor.
So, if the
22
sponsor would like to elaborate, we used the same
264
1
numerator that they used in their.
So, to the
2
extent that they were able to do it, we did it.
3
DR. GROSS: Dr. Epps will ask the
last
4
question.
5
DR. EPPS: My question is for Dr.
Pitts
6
regarding the utilization. The
percentage of
7
refills with the current RMP fell to 2.4 percent.
8
That seems really low. I was
wondering whether a
9
change in the dose was considered a new
10
prescription or was it considered a refill because
11 the
patient was continuing on therapy. Sometimes
12
when prescribing Accutane you may start at one dose
13 and
modify your dose depending upon blood tests,
14
participant interactions or something of that sort.
15
DR. PITTS: Actually, as of the
16
implementation of the program all isotretinoin
17
prescriptions are considered new prescriptions.
18
Previously you could write a prescription that was
19
one, plus three refills, and every one is a new
20
prescription. Therefore, the
numbers should
21 decrease and approach zero at some point, or
at
22
least decrease. When you look at
that particular
265
1
aspect you should not have refills.
2
DR. GROSS: Thank you. The next speaker
3 is
Dr. Richard Wagner, Kaiser Permanente Drug Use
4
Management.
5 Kaiser Presentation
6
DR. WAGNER: Good afternoon.
7
[Slide]
8
I am from California. I am here
with a
9
colleague of mine, Craig Cheetham, who is in the
10
back, who has helped me put this presentation
11
together. What we would like to
do is actually
12
show our results for the past two years.
We have
13
actually managed Accutane this year's pregnancies
14
differently than the S.M.A.R.T. program.
So, we
15 are
going to, for the first time actually publicly,
16 lay
this information out for folks to see.
We
17
would be interested in your feedback on what things
18 we
could do together with the agency or others to
19
work to actually maybe even further validate the
20
results that we are going to present today.
21
[Slide]
22
For us it really boils down to the
266
1
previous advisory panel from about year 2000 saying
2
that no one should begin isotretinoin therapy if
3
pregnant and that no pregnancy should occur while
4 on
that therapy. To me that is very simple,
that
5
goal was zero.
6
Our charge in our organization was could
7 we
actually prove that we could get that to zero
8 and
prove that it is zero based on a program that
9 we
conceived to be different than the S.M.A.R.T.
10
program. Some of you have seen
the KP Med-SMART
11
logo. We also got a little bit
clever and
12
Med-SMART for us really talks about systematic
13
monitoring and assessment for risk of toxicity.
14 This
program could work not only for Accutane; it
15
could actually work for other drugs and we are in
16 the
process of figuring out how we are going to add
17
methotrexate into an equivalent program because, if
18
women on isotretinoin deserve to be protected,
19
women who are on methotrexate certainly deserve to
20 be
protected also. So, we would like to
come back
21 at
a future time and let you know how we are doing
22
with that one.
267
1
The "how"--the linkage is dispensing of
2
isotretinoin for female patients to verification of
3 a
negative pregnancy test. To us it was
simple.
4 The
pharmacist had to see, either electronically or
5 via
paper or some other documented communication,
6
that for that prescription for that woman that day
7
there was a negative pregnancy test before we
8
dispensed that prescription, and we needed to do
9
that every time for every woman in every situation
10
that we manage.
11
We developed a centralized female patient
12
registry with prescription level detail and
13
associated prescriber and pharmacy performance
14
reporting. You can't put together
a program and
15 not
report on people's performance. If you
don't
16
report, they won't improve or they won't improve
17
very much. Even the best people
who will try need
18
reporting and feedback in order to improve. I am
19 going
to show you how we do that.
20
We developed operational guidelines.
We
21
needed to have a consistent approach over 250
22
Kaiser outpatient pharmacies so we wanted to have a
268
1
consistent approach and have written operational
2
guidelines for how we do this in each of our
3
pharmacies. We do allow for some
reinvention, and
4 I
will talk about that later on, because there is
5
always some minor variation that is not going to
6
quite work if you tell everybody they have to do it
7 the
exact same way everywhere. But the goal
always
8 has
to be that nobody gets a prescription for
9
Accutane that is pregnant, or we are not going to
10
dispense an Accutane prescription if we don't have
11
that negative pregnancy test. The
patient will
12
have to go get a negative pregnancy test before we
13
dispense that prescription.
14
We have over 100 dermatologists that we
15
work with in southern California and northern
16
California, and there are guidelines for the
17
appropriate use of Accutane, isotretinoin. Those
18
guidelines have actually been in place for years
19 and
over time we have adjusted those guidelines to
20
reflect current reality. So, we
actually have a
21
living document that says this is how we think
22
Accutane or isotretinoin should be used in our
269
1
program for our patients.
2
[Slide]
3
I will only give you one Kaiser slide.
4
Typically the Kaiser folks show up with ten slides
5
because we have a very complicated organization.
6 But
in California we take care of about 6.1 million
7
Californians. We are an
integrated healthcare
8
delivery system. I am not up here
talking about us
9
being a PBM or an insurance company or any of that
10
other type of stuff. We really take
care of
11
patients and we do that in a variety of ways. We
12 are
linked financially; we are linked
13
operationally; we are linked technologically; and
14
maybe most importantly, we are linked culturally.
15
These efforts within Kaiser Permanente are common
16
efforts for us so I am talking about one example.
17 If
we were actually talking about many topics we
18
could come up with many examples of how this
19
approach actually works to provide, I think, very
20
good patient care.
21
As you know, Kaiser Foundation Hospitals
22 and
Kaiser Foundation Health Plan is a non-profit
270
1
community benefit organization.
In California we
2
work with two very large medical groups, one in
3
northern California and one in southern California.
4
They are an independent group of physicians that
5
contract exclusively to provide care to Kaiser
6
members. In contrast maybe to
folks who live in
7 the
Washington, D.C./Baltimore area, California is
8
really a hospital-based integrated healthcare
9
delivery system. We have about 29
hospitals and
10
medical centers associated with those hospitals,
11
over 200 medical officers and over 250 outpatient
12
pharmacies.
13
So, the picture I am trying to paint for
14 you
is that you really have to see that
15
laboratories, pharmacies, physicians--everybody is
16 in
the same building. I mean people work
together.
17
They are busy doing their own things, certainly,
18 but
it is a big advantage to actually know people
19 in
your building that are taking care of patients
20 and
have those folks interact with you to make sure
21
that we can actually take care of patients that
22
require this little bit of extra effort, the best
271
1
that we can.
2
In terms of prescription volume, we have
3
about 42 million outpatient prescriptions. Last
4
year about 24,000 of them were isotretinoin
5
prescriptions. That is pretty
consistent, about
6
2,000 prescriptions per month.
7
What we don't have in California but other
8
Kaiser regions do have--and it is going to come--we
9
don't have an automated medical record.
That is an
10
important thing to keep in mind.
When the
11
automated medical record comes-- little pieces of
12 paper
with drug names, putting little stickers on
13
them, it is just not going to make it.
There has
14 to
be a different way involving technology that is
15
going to support the equivalent, and that is
16
ensuring that patients are not pregnant when they
17 get
these medications. We don't have that
here.
18 One
of the advantages in us trying to perfect this
19
process outside of the yellow sticker is that we
20
wanted to learn how to do this so that when the
21
electronic health record comes, the automated
22
health record, we are going to incorporate that
272
1
learning into that tool.
2
The last thing, let me make a
3
connection--and I didn't quite figure this out
4
initially either--we have extensive experience in
5
care management or disease management
6
registries--diabetes, asthma, whatever you want to
7
call it. These are patients at
risk also and some
8 of
these registries that we manage have over
9
300,000, 400,000, 500,000 patients in the
10
registries. If you think of the
female patient as
11 an
at-risk population taking an at-risk drug, many
12 of
the learnings from that group are translated
13
into some of the efforts that we are presenting
14
today. So, what we are doing for
the isotretinoin
15
patients really is reflective of what we are doing
16 for
other at-risk populations.
17
[Slide]
18
Some descriptive statistics--we do have
19 IRB
approval from southern California to share some
20 of
the descriptive statistics. I will tell
you
21
that the north and the south are very similar. If
22 we
were to do further work with the agency or
273
1
others, we could go back and get IRB approval for
2
northern California data also.
But it is actually
3
interesting I think the results are somewhat
4
similar to what we have seen today.
5
This is January, 2002 through November,
6
2003 data reflected up here. The
average age for
7 the
female patients is about 25 years. It is
about
8 49
percent females. Unique patients in the
south,
9
2,376; unique patients in north California, 2,253.
10 You
can see the prescription counts for those
11
patients. In about 85 percent to
90 percent of the
12
time we can find an acne-related diagnosis in our
13 machine. The other 10 percent just might not be
14
findable or someone didn't code it right. It is
15
interesting to note that cancer diagnosis comes up
16
about half percent and we do have a small number of
17
patients with various cancers that we do treat with
18
isotretinoin.
19
Another interesting observation is about
20
prior therapy within the previous 6 months, and 95
21
percent of the patients had not been on
22
isotretinoin before; only 5 percent had had a
274
1
previous episode of therapy in the previous 6
2
months in our program. The range
of prescriptions
3 was
1-16, with a mean of 4; 4-5 prescriptions,
4
30-day supply. It all adds up to
about 124 days of
5
therapy per patient. The type of
therapies that
6 the
patients had before they started isotretinoin
7 are
listed there also. About 64 percent of
the
8
patients had one of those therapies listed down
9
below. Most patients are not
started on
10
isotretinoin as a first-line therapy.
11
[Slide]
12
Here is where the good stuff is.
It may
13 be
a little bit hard to see. January, 2002
through
14 December,
2003, what was going on here? This is
15
before the formal program started, the first four
16
months of 2002. We went back and
just measured
17
what was going on in our system during that time
18
frame. I did not do any chart
reviews or anything
19
like that, just data system stuff.
You can see
20
that about 50 percent or 60 percent of the time we
21
were at baseline compliant with a patient picking
22 up
their isotretinoin prescription within seven
275
1
days.
2
Part of that is misleading though because
3
what was happening is that we didn't have all the
4
pregnancy tests in the machine.
The dermatologist
5
practice was to do about half of those pregnancy
6
tests in their office, and the big change we had
7
with the dermatologists was saying we have to get
8 you
to do these tests so the electronic system can
9
pick them up, otherwise you can't get credit for
10
doing this. If one medical center
does it one way
11 and
another medical center does it another way we
12
really don't know what is going on.
So, the big
13
change we asked of the dermatologists was that at
14
least one test for every dispensed prescription
15
needs to come through the laboratory system because
16 we
and the pharmacy can then, from a technology
17
standpoint, see the results of that test. If you
18
test them in the office, that gets put in the chart
19 and
we may or may not see it and it won't get
20
reported as being compliant. One
of the things we
21
wanted to be able to demonstrate is, if we are
22
going to do something different than the S.M.A.R.T.
276
1
program, we want it to be at least as good or
2
better than the S.M.A.R.T. program, and that meant
3
that we had to be able to collect and show data
4
like we are doing today.
5 So, I don't want to leave anybody
with the
6
misunderstanding that somehow the dermatologists
7
weren't doing the right things. I
just did not go
8
back to their charts and pull out of all of those
9
urine or lab tests that were in the charts for
10
years and years because that is how they practiced.
11
But let's look after April. After
April,
12 you
can see that things got better. We went
from
13 50
percent or 60 percent to about 75 percent or 80
14
percent. I have to tell you, we
went back again
15 and
said 75 percent or 80 percent, guys, is just
16 not
going to make it either. We have to do
better
17 on
that one also. So, what did we do? I am going
18 to
show you some pictures of this in a minute.
We
19
actually developed a web-based tool that was the
20
centralized patient registry.
That tool collects
21
every prescription for Accutane for female patients
22 and
all of the associated lab values and makes it
277
1
available within near real-time back to the medical
2
center so that they can actually act on any
3
discrepancies or make any corrections, if necessary
4 or,
if there is a quality problem, actually
5
intervene and try and fix it or get it into the
6
quality system.
7
So, this web-based tool, and I will show
8 you
some screen shots, actually turned out to be
9 the
next strategy that we added in here. We
were
10
doing this early stuff by reporting, sending pieces
11 of
paper out to people and saying please look at
12
this report and please do better.
But they did go
13 out
two or three months late and, anyone who is
14
involved in taking care of patients and getting a
15
report two or three months late--it means nothing.
16 You
have to get a report about performance in terms
17 of
patient care within days or maybe a week,
18
otherwise you just don't remember and you are off
19 to
the next thing.
20
So, what happened, lo and behold, we had
21 to
go through a little bit of training. We
22
actually did this in the 29 medical centers across
278
1 the
state and we rolled the web-based tool out
2
about here. When you start
providing people with
3 the
necessary information and the reporting and the
4
technology to support good clinical practice, and
5
they can see how they practice over time and they
6 can
see how their peers and other medical centers
7
practice, the A students go right to the top. That
8 has
been sustained now for over a year.
9
I am going to show you what that tool
10
looks like but the thing that really drove us from
11
mid-level performance to I think pretty high
12
performance was the implementation of the tool in
13
addition to all the other things that we did. The
14
tool by itself wouldn't do it.
Without guidelines
15 by
the dermatologists and their interest in
16
following up, I don't think we would get there;
17
without the pharmacists paying active attention and
18
just being very rigorous--if you don't have a
19
negative pregnancy test, you have to have it before
20 we
fill out that prescription. You start to
put
21
those elements together, get people the reporting
22 and
feedback they need and they do perform.
279
1
I am going to show you that it also worked
2
this way in northern California.
This is a
3
different group of physicians. I
mean, the
4
physicians in the south and north are really two
5
separate medical groups but the impact was the
6
same.
7
[Slide]
8
I should have made it green or blue or
9
something different but, again, you can see the
10
kind of baseline performance before.
This is
11
really written policies, procedures and standards
12 on
how we wanted people to behave. In order
to
13
really take it from this 80 percent thing, we had
14 to
put the same electronic web-based tool in place
15
and, again, provide people real-time information,
16 and
we said if you are lagging on performance
17
everybody can see it; you can see it.
How are we
18
going to fix it? And, people tend
to fix those
19
things as quickly as they can, to tell you the
20
truth.
21
[Slide]
22
A couple of things about pharmacy
280
1
utilization, we didn't see a drop in pharmacy
2
utilization within our program.
It was very
3
fascinating to me that outside of Kaiser
4
utilization dropped 23 percent or so.
We didn't
5 see
a change in utilization.
6
There are a couple of interesting things
7
here. This is southern California
data, 2002 to
8
2003. So, the time frames are not
quite matched
9
but, in essence, we are filling the same number of
10
prescriptions this year and last year as we filled
11 in
2002. I really go back to my comments
about the
12
dermatologist practice and evidence-based medicine
13 and
having guidelines that they endorse.
That
14
didn't change with all the things we did when we
15
added our new technology and our new ability to
16
collect and analyze information.
I think the
17
practice stayed the same. The one
practice element
18
that changed really was to stop doing lab tests in
19 the
medical office because we can't track those
20
things. It was just one of those
things that had
21 to
change. But in terms of I think
prescribing to
22 the
appropriate people, we did not see a change
281
1
really in prescribing for the appropriate patients
2 for
Accutane, if you really believe in
3
evidence-based medicine and using the guidelines to
4
drive clinical performance.
5
It is interesting that there does seem to
6 be
some seasonality. I saw some discussion
in some
7 of
the documents about is there seasonality or is
8
there not. I would say there is
seasonality. It
9
looks like this is when kids are in school. It
10
looks like this is when they are at the beach. I
11
don't know if it has to do with the school or the
12 beach or fun, or maybe dermatologists take
vacation
13
during the summer, but something is going on there.
14 It
is actually very consistent with other pharmacy
15
utilization pictures. Things tend
to peak in the
16
wintertime and things tend to drop in the
17
summertime.
18
[Slide]
19
Northern California had a similar pattern.
20 You
can see that in 2003 we were actually filling
21
more prescriptions for isotretinoin than in the
22
previous year, but the patterns are very much the
282
1
same. So, we didn't see the
drop-off in terms of
2
utilization. In fact, if
anything, we are filling
3 a
few more prescriptions here.
4
[Slide]
5
Let me talk about the web-based
6
application just because I know there is interest
7 in
maybe a centralized repository or patient
8
registry for patients. I will say
again that
9
basically I really built this in concert with folks
10 who
have been managing patients with diabetes,
11
asthma, heart failure, etc. We
have extensive
12
experience in terms of managing large patient
13
databases and then using the databases to improve
14
performance. So, we are really
standing on the
15
shoulders of others that have come before us.
16
This is an at-risk population.
They may
17 be
only at risk for four or five months but the
18
consequences are severe. If it is
really the goal
19 to
have nobody become pregnant while on this
20
medication, then you need to have tools similar to
21
this I think to be successful.
22
It acts as a centralized patient registry.
283
1 It
is indexed on an isotretinoin prescription.
So,
2 we
only know about you if you got a prescription
3
filled in our Kaiser pharmacy.
There are some
4
problems a little bit with that too, but 95 percent
5 of
our patients have Kaiser drug benefits so we are
6
probably picking up at least 95 percent of those
7
prescriptions. Prescriptions can
go out of the
8
system and we would lose them in that case.
9
Prescriptions can come into the system and we would
10 add
them in also.
11
The system automatically links pharmacy,
12
laboratory and patient demographic data.
We
13
refresh it weekly. We could
refresh it daily but
14 we
expect people to take a look at the tool weekly
15
because that is usually soon enough for us. The
16
other aspect of the tool is that 93 percent of the
17
time we pull the data in for the people.
There are
18
folks in the medical centers, pharmacists and
19
dermatologists and nurses that are taking care of
20
patients and you don't want them having to type
21
data in the machines. Seven
percent of the time
22
there is a transactional input because we still
284
1
honor the yellow sticker. So, to
the degree we
2
have a manual prescription without the laboratory
3
coming through the Kaiser system, we will honor
4
that prescription whether it is from a KP
5
dermatologist or someone outside.
So, about seven
6
percent of the time we have to type that thing in.
7 You
know, that is a bit of a pain but that is how
8 you
get credit. If you don't type it in I
will see
9 a
gap in your performance and they will get a phone
10
call.
11
What is also important here--we do this
12
with every other tool that we have, we compare
13
medical center performance to medical center
14
performance during the same time frame, and we also
15
tell our medical centers to compare their
16
performance over time. It is up
to each medical
17
center to actually get into the nitty-gritty of
18
this stuff. That is where patient
care is
19
provided. They have the tools in
front of them to
20 see
their own performance. They know
everybody
21
else sees their performance; they see everybody
22
else's performance and it actually does make people
285
1
perform because, first, they always want to do the
2
right thing for the patient but then, if you put up
3
something like this, they absolutely want to do
4
well in terms of the performance that is out there.
5 Data is current. That is another key
6
thing for us. If data is not
current people can't
7 be
expected to act or react or to improve.
Getting
8
physicians' or pharmacists' or nurses' data six
9
months old means nothing. We download
the data
10
into Excel format for custom reporting so if you
11
want to do some custom reports at your medical
12
center, you can actually download that thing and
13
actually create all the funny little reports that
14 you
want, and you can actually drill down into the
15
prescription level data, the physician level data,
16 the
pharmacy level data, and patient level data and
17 we
know everything about everybody and we also can
18
roll it out so that people can see it at a high
19
level also.
20
Most importantly, security and limited
21
access to protect the confidential nature of this
22
data--it is heavily protected. It
is password
286
1
protected. People can't get in
there. People who
2 do
get in there we know about and if anybody, God
3
forbid, does something wrong, you know, they are
4
going to get fired during these days of
5
confidentiality.
6
[Slide]
7
I am going to give you a couple of screen
8
shots and then try to move pretty quickly. Here is
9 a
screen shot that I took. It was a screen
shot
10
from data refreshed through 2/13/2004.
This
11
actually shows you January of 2002 to December,
12
2003. The essential thing is that
we have this
13
web-based system. The data is
displayed. People
14 can
roll it out and with the tool itself you can
15
actually monitor managed performance.
16
[Slide]
17
I want to finish up with a couple of
18
things. For those of you that
read Paul Plsek,
19
these are complex human behavior things.
They are
20 not
complicated issues. One thing I think
the risk
21
management people and maybe the FDA could learn is
22
that people who do population management actually
287
1
manage patient care. The
difference is between
2
managing a complicated problem and managing a
3
complex problem. I am not going
to go through that
4
today but there are people out there who are
5
experts in this area, obviously.
6
[Slide]
7
There are some issues here that kind of
8
differentiate between a complicated problem and a
9
complex problem. I will leave you
that to read.
10
[Slide]
11
Another guru here is Don Berwick who has
12
laid out much of what it takes actually to change
13 the
healthcare system. If you are not
looking at
14
what he has to say in this area, you may be missing
15 an
opportunity to truly improve the healthcare
16
system in this country.
17
[Slide]
18
I will finish up with why this
is working
19 at
Kaiser Permanente. I probably have
already
20
convinced you of this, I hope.
Physician and
21
pharmacist buy-in. We work as a
team. It works
22
because people there are committed.
There is a
288
1
huge cultural alignment to make this work.
2
We also allow for reinvention locally.
If
3 a
lab locally and a pharmacist want to get together
4 a certain
way that is slightly different but the
5
pharmacist has that negative pregnancy result
6
before they dispense a prescription, it is allowed.
7 We
just want that documentation. You have
to have
8
some flexibility to help people work together
9
because things are just not set up exactly the same
10
every place.
11
I have also talked about technology and
12
data. It is readily
available. It is in real
13
time. That is obviously a big
advantage.
14
Performance reporting and monitoring is critical to
15
success. People do not get better
unless they have
16
real-time performance reporting and monitoring.
17
And, quality issues are reviewed through the normal
18
peer review quality channel at each of our medical
19
centers. In the event that there
is a quality
20
issue, that does get channeled normally through for
21
peer review and then the quality process takes
22
hold.
289
1
I am going to stop there and take
2
questions, and would be interested if anybody had
3 any
observations.
4
Questions to Kaiser from the Committee
5
DR. GROSS: Dr. Trontell?
6
DR. TRONTELL: Yes, with your
impressive
7
improvement in pregnancy test performance, I wonder
8 if
you can give us any information on differences
9 in
pregnancy outcomes, and whether the Kaiser
10
system captures information on contraception use as
11
well as various forms in which pregnancy might be
12
terminated.
13
DR. WAGNER: Yes, we would have to
do some
14
additional work because I pulled most of this data
15 off
our systems. But, in theory, we would be
able
16 to
go back and link this up. In essence,
you have
17 a
northern and southern California cohort of 5,000
18
patients during this time frame.
For example, what
19 we
haven't done is take a look 60 days out after
20 the
last fill so that a patient has a prescription,
21
three days worth of medicine and then add 30 more
22
days in and what was their pregnancy status 60 days
290
1 after
that dispense or 30 days after the last fill?
2
Those are things that we could actually work on
3
because the data is available. We
just haven't
4
done that level of work. In
theory, we could data
5
mine and find all sorts of things.
If we spend
6
some time and effort, I think we can actually tease
7 out
much of this.
8
DR. GROSS: Dr. Kibbe?
9
DR. KIBBE: The crucial question
was
10
asked. Without an endpoint of
knowing how many
11 pregnancies occurred in your system, I don't
know
12
whether it is any better than anything else.
13
DR. WAGNER: The one thing we can
14
ascertain looking back at 2003 data, and I have
15
done this myself in addition to the data people, is
16
that we had no patient who was pregnant and
17
received an Accutane prescription in 2003 in
18
northern California and southern California, and we
19 had
no patient who was on Accutane and became
20
pregnant during that time frame.
Now, we should do
21 a
little bit more work because people could become
22
pregnant, like I say, 30 days after and we haven't
291
1
picked up that data. We would be
willing to look
2 at
that also just in the event there was a
3
pregnancy after the last Accutane fill.
4
DR. GROSS: Dr. Day?
5
DR. DAY: We all have the same
question.
6
Could you just repeat that last part?
Are you
7
saying there were no fetal exposures in the initial
8
period and during the period of taking the drug,
9 and
you just haven't looked at the 30 days post?
10
DR. WAGNER: For 2003, north and
south, we
11 can
find no fetal exposures.
12
DR. GROSS: Dr. Vega?
13
DR. VEGA: Yes, I have the same
question
14 Dr.
Trontell had about the pregnancy occurrence
15
among your group. Have you done
studies to link
16
pregnancy status with drug use or pharmacy
17
prescription before?
18
DR. WAGNER: General studies?
19
DR. VEGA: Yes.
20
DR. WAGNER: No. Actually, this was
21
probably more work we have done related to
22
pregnancy status or potential pregnancy status than
292
1 any
other work we have done.
2
DR. GROSS: Dr. Wilkerson?
3
DR. WILKERSON: First of all, my
4
compliments. This is truly what I
was talking
5
about before as best of practices and it doesn't
6 get
any better than a controlled setting like this.
7
Having been a managed care director myself, I can
8
tell you that trying to get physicians to go in one
9
direction is like herding cats and you certainly
10
have mastered the art, probably from a monetary
11
reward, but it certainly helps change behavior.
12
This is the sort of stuff that we need as
13 a
committee before we go inventing the national
14
wheel to know in best practices does a particular
15
strategy work. If this data holds
up--I mean this
16 is
an ideal situation. Physicians are
controlled.
17 The
distribution is controlled. You have a
panel
18 of
patients that are interested in their health.
19
This is the sort of stuff that we need to know as a
20
committee if we are going to recommend something
21
that is not a band-aid, that in best practices it
22
does work. So, my compliments to
you.
293
1
DR. GROSS: I would echo that
question.
2
Your adaptive complex system works because you have
3 a
homogeneous culture. In the rest of the
world
4
where medicine is somewhat of a free-for-all, do
5 you
have any lessons for those systems, any
6
suggestions? You said you
actually publish the
7
results and show them to the doctors as a group so
8
they see everybody else's results.
9
DR. WAGNER: Yes.
10
DR. GROSS: What kind of feedback
do you
11 get
on that? Do they accept that?
12
DR. WAGNER: Absolutely. It happens all
13 the
time, not just with this report but whether you
14 are
managing an asthmatic patient or any other
15
groups of patients. Using
unblinded peer
16
comparison reports--they are protected from a
17
confidentiality standpoint; I mean, it is in a
18
departmental meeting or that type of thing--is a
19
very powerful tool to get people either to change
20 or,
you know, there are times when there may be
21
justifications for a practice but it is a very
22
effective technique that has been used over and
294
1
over.
2
One thing I would add, and I actually
3
just discovered this maybe in the last month, there
4 is
a clozapine registry that I think has many of
5 the
same elements. The Ivax company in
Florida
6
that manufactures clozapine requires white cell
7
counts before you can dispense it.
This clozapine
8
registry is an absolutely fascinating tool. It is
9
very similar to our Med-SMART tool in terms of the
10
functionality and capability. So,
if anybody is
11
interested in looking at that, and the agency
12
certainly has some review of clozapine--if they
13
could demonstrate similar levels of compliance,
14
because that would be outside of Kaiser Permanente,
15
then that might be a very powerful learning also.
16
DR. GROSS: Dr. Bigby?
17
DR. BIGBY: I would like to know
what was
18
your experience in terms of numbers and rates of
19 pregnancies
on Accutane prior to April of 2002.
20
DR. WAGNER: I don't have those
actual
21
numbers. Those numbers would be
rolled up at a
22
medical center level. They would
have been dealt
295
1
with in the confidential peer review quality
2
assurance process. So, I don't
actually have those
3
data. They weren't available
before we actually
4
started this program in a centralized fashion.
5 DR. BIGBY: So, it is possible you didn't
6
have any before.
7
DR. WAGNER: It is possible we
didn't have
8 any
before. Again, if you talk to the
9
dermatologists I work with, they will tell you that
10
this whole thing that they are going through they
11
don't think changed quality much.
But it actually
12 I
think demonstrates quality in a way that if you
13 ask
me that question today I actually have a lot
14
greater degree of confidence that, yes, the quality
15 is
here. But the dermatologists who do this
16
quality review process would tell you I think they
17
don't think they had much of a problem before.
18
DR. GROSS: Dr. Whitmore?
19
DR. WHITMORE: I think your
program
20
probably does prevent initiation of Accutane when
21
pregnancy is present. As far as
preventing
22
pregnancy during therapy, I would question how you
296
1
gather data about pregnancy. I
would also say that
2
some women may take abortifactants and never report
3
that to their doctors, and also go outside of your
4
system for an abortion.
5
DR. WAGNER: Exactly. The
6
program--because it indexes on Accutane
7
prescriptions, at the end it is open.
But I think
8
there are enough data elements within our system
9
that we could actually figure out how big a
10
problem, if any, does exist.
11
DR. GROSS: Thank you very much. The next
12
speaker is Dr. Richard Miller, Professor and
13
Associate Chair of Obstetrics and Gynecology. He
14
will talk about the Organization of Teratology
15
Information Services, and this is an interim
16
report.
17
Organization of Teratology Information Services,
18
Interim Report, North American Isotretinoin
19 Information and Survey Line
20
DR. MILLER: I appreciate the
opportunity
21 of
joining you today and reflect back on my first
22
visit with you all, back in the early '80s and
297
1
beyond.
2
As indicated, I am on the faculty of the
3 University of Rochester but I also am
Director of
4
PEDEX, NIH a New York teratogen information
5
services as well. So, I tend to
see the patients
6 who
do become pregnant and are, in fact, addressing
7
those sorts of issues today.
8
[Slide]
9
What I will be talking about is what OTIS
10
does. How we, in fact, have
gotten involved in the
11
Accutane issue, and I will share a couple of cases
12
with you and a study that we have been doing,
13
called the OTIS North American Isotretinoin
14
Information and Survey Line. This
is an interim
15
report. We were asked by the FDA
to come forward
16
with our data as we are in the middle of this
17
particular study.
18
[Slide]
19
The study is funded by the CDC in
20
partnership with the Association of American
21
Medical Colleges, and you can see the other
22
individuals that are involved with this program.
298
1
[Slide]
2
OTIS is a non-profit North American
3
network of 19 state or regional teratology
4
information services. These TIS
provide updates on
5
information regarding the effects of drugs and
6
chemicals on the human embryo and fetus via a free
7 of
charge telephone consultation. There are
more
8
than 100,000 calls per year that we receive. Half
9 of
them are from the general public and the other
10 half
are from healthcare professionals.
11
[Slide]
12
OTIS is organized to stimulate and
13
encourage research, education and the dissemination
14 of
knowledge in the field of teratology and,
15
hopefully, to improve the abilities or teratogen
16
information services to provide accurate and timely
17
information about prenatal exposure with the
18
overall objective or preventing birth defects and
19
improving the public health.
20
[Slide]
21
Now, in 2002 we provided testimony to the
22
Subcommittee on Oversight and Investigations of the
299
1
Committee of Energy and Commerce of the U.S. House
2 of
Representatives. I share here some of
the
3
recommendations that we have offered at that time.
4 In
brief, OTIS recommended safeguards that are
5
modeled on the thalidomide steps program, which
6
include several mandatory elements.
7
[Slide]
8
OTIS also recommended strict adherence to
9 the
approval indications for the use of
10
isotretinoin, limiting prescribing to
11
dermatologists and the S.M.A.R.T. program, improved
12
contraceptive counseling, and direct patient access
13 to
risk assessment and counseling and, very
14
importantly, the continued evaluation of the
15
effectiveness of these programs.
Because I have
16
limited time, I have highlighted only a few of
17 these
recommendations, and down here you can find
18 the
entire report, here at the web site.
19
[Slide]
20
We have been active and in 2000 the
21
California Teratogen Information Service
22
contributed to a study of 14 women whose
300
1
pregnancies were inadvertently exposed to
2
isotretinoin, reflecting the failure of the
3
Pregnancy Prevention Program.
4
[Slide]
5 At that time, we began cataloguing the
6
number of calls that we get at OTIS across the
7
United States and Canada, and here you can see the
8
numbers that we have gotten over the past three
9
years. This was at the same time
that we are now
10
beginning to look at the transition to new systems,
11 the
S.M.A.R.T. versus the old pregnancy prevention
12
system. In these particular
instances though we
13
have not found, at least in our calls of very
14
specific patients who call in, any decreases in the
15
number of calls that we have been receiving during
16
those times.
17
In collecting this information, we also
18
realized that we needed to gather much more
19
detailed information from these patients to better
20
understand why they were not successful in
21
preventing pregnancies.
22
[Slide]
301
1
In so doing, we initiated the isotretinoin
2
survey and, remember, these are women who
3
voluntarily called our service seeking help. So,
4 we
spend a lot of time with these patients,
5
counseling them up front and then, at the end, ask
6
them if they are willing to participate in the
7
survey. So, we have already
gained some confidence
8
with these women and begin to know them, at least
9
from that first meeting.
10
There is going to continue to be
11
enrollment through September, 2004, and we use a
12
very detailed and structured interview by a
13
research specialist and the participant is followed
14
until the known outcome of the pregnancy.
15
Our objectives for this study have been to
16
identify barriers to the successful implementation
17 of
the components of the pregnancy risk management
18
programs. Remember, we are
talking both about
19
Canada and the U.S.
20
[Slide]
21
As we have seen multiple times, there are
22 two
major goals to the S.M.A.R.T. program, one to
302
1
prevent pregnancy in women who are already taking
2
isotretinoin and to prevent embryonic exposure to
3 isotretinoin in women who are already
pregnant. I
4
would like to share two case examples with you. We
5
have been talking about general specifics, but
6
let's talk about two ladies who called our service.
7
[Slide]
8 I will call our first lady Ms.
A. She is
9 a
pregnant woman in her 30s and she has reported to
10 us
that she had exposure between the third and
11
seventh week of gestation. She
reported
12
discontinuing her birth control method one month
13
before starting isotretinoin and misinterpreted her
14
doctor's statement that it might be more difficult
15 to
get pregnant for a time afterwards to mean she
16
couldn't get pregnant.
17
She also reported to her dermatologist
18
that she couldn't get pregnant and, therefore, he
19
took that as a reason not to counsel the patient
20
about contraception or do pregnancy tests. She
21
also was reported to be taking samples of
22
isotretinoin to treat a chronic skin condition that
303
1 was
not acne. So, she was receiving these
from her
2
dermatologist; she didn't have to go to the
3
pharmacy.
4
So, here we are relying on a variety of
5
information from the patient that led to,
6
obviously, a very serious problem, the fact that
7 she
is pregnant and we do not know the fetal status
8 at
this point today.
9
I add as an additional comment that we
10
reviewed the patient's exposure history here
11
because we were not aware that physician samples
12
were available for isotretinoin.
All of her
13
responses though were consistent with that.
14
[Slide]
15
Let us turn to Ms. Z. Ms. Z. is a
16
pregnant teenager. She called one
of our Teratogen
17
Information Services to learn about potential
18
problems that her baby might have because of
19
isotretinoin exposure that occurred between five
20 and
six weeks for a total of about six days.
She
21 had
never taken isotretinoin before, however, she
22
wanted to clear up her acne. She
reported that she
304
1
learned she was pregnant following a phone call
2
from her dermatologist.
3
Because a family member was present when
4 the
phone call was received, she reported to her
5
dermatologist that she was not sexually active and
6
that the test must be incorrect.
Her dermatologist
7
reportedly accepted her statement and thought the
8
pregnancy test must be incorrect and gave her a
9
prescription for isotretinoin which she started the
10
very next day.
11
After taking isotretinoin for those six
12
days she decided to have another pregnancy test and
13
when, in fact, she discovered it was positive she
14
stopped her isotretinoin. Her
pregnancy is
15
continuing and the fetal status is also unknown.
16
[Slide]
17
So, from these case histories we can see
18
really an illustration of several missed
19
opportunities for prevention of exposure to
20
isotretinoin during pregnancy. The
errors arise
21
from multiple sources--miscommunication between the
22
healthcare provider and the patient;
305
1
misinterpretation of information by the healthcare
2 provider;
and denial of risk by the patient. Lack
3 of
adherence to the required components of the risk
4
management program remove safeguards that may have
5
prevented these exposures.
6
[Slide]
7
Let's turn to our study. We have
a
8
limited number of cases, a total of 23, half from
9 the
United States and half from Canada.
There are
10 key
differences here because in Canada we continue
11 to
have the Pregnancy Prevention Program whereas in
12 the
United States we have the S.M.A.R.T. program so
13 we
have a built-in comparison, you might say.
In
14
Canada 100 percent of the patients were taking
15
Accutane. In the United States 55
percent of the
16
patients were and one patient was actually given
17
both Accutane and a generic.
18
If we look at some of the data, especially
19
what may be considered why they were taking it, in
20
response to several questionnaires about the use of
21
isotretinoin we asked our patients about whether
22
they were being treated for severe, recalcitrant
306
1
nodular acne. What you can see
here is that 36
2
percent of the patients described their skin
3
condition as cystic or nodular in the United States
4 and
a similar number in the Pregnancy Prevention
5
Program. Even fewer recalled that
their doctor had
6
diagnosed this condition.
7
Somewhat more encouraging, which in fact
8 was
discussed a little earlier, is that 82 percent
9 of
the patients were, in fact, previously treated
10
with oral antibiotics; 57 percent in the Pregnancy
11
Prevention Program in Canada.
12
[Slide]
13
So, this helps to establish to some degree
14
what, in fact, was the usage but we were more
15
concerned with the elements of the S.M.A.R.T.
16
program and the survey included questions that
17
explore almost all aspects of women's pregnancy
18
exposure but today I am going to limit it to, and
19
highlight, these four elements.
These elements
20
pertain to monitoring for the pregnancy program and
21
they can be objectively evaluated, and are the
22 keystone
of an effective prevention program.
307
1
If we turn to "women must have two
2
negative pregnancy tests" and, obviously, you are
3 all
aware of that what we asked our patients about
4 was
whether, in fact, they had a second pregnancy
5
test during their menstrual period before beginning
6
isotretinoin. Interestingly, 27
percent did and 3
7
percent did. Thus, for the
S.M.A.R.T. program
8 alone it appears that 73 percent of the women
9
surveyed were not screened using two pregnancy
10
tests as required by the program.
In addition, 22
11
percent of these women reported that they had one
12
pregnancy test that indicated they were not
13
pregnant when, in fact, they were.
A second test
14
might have successfully prevented the exposure.
15
[Slide]
16
Our second element--according to the
17
S.M.A.R.T. program women must use two forms of
18
birth control simultaneously starting one month
19
before receiving the prescription.
Overall, 70
20
percent of the women surveyed said that they were
21
using at least one birth control method and this
22 was
similar in the two groups. However, only
36
308
1
percent of the women in the U.S. and 8 percent of
2 the
women in Canada reported they were using the
3 two
forms. Thus, 64 percent of women surveyed
4
indicated they were not following the S.M.A.R.T.
5
requirements to use two forms.
6
I must add though, in addition, when women
7 who
reported they were using contraception one of
8 the
responses was referring to unreliable methods
9
such as the rhythm method. This,
again, confirms
10 the
need for effective contraception counseling.
11
[Slide]
12
Our third element--according to the
13
S.M.A.R.T. program women must receive a pregnancy
14
test each month before refilling their
15
prescription. Only 36 percent of
our women in the
16
U.S. reported that they had monthly pregnancy
17
testing during the course of therapy.
Actually,
18 the
percentage in Canada was much higher.
Our
19
conclusion is that it appears that 64 percent of
20 the
women in our group were not screened for
21
pregnancy monthly as required.
22
[Slide]
309
1
Our fourth element--pharmacists must only
2
fill prescriptions that bear a yellow qualification
3
sticker. The responses from our
group, and maybe
4 it
is an unusual group, are that women who recalled
5
seeing a sticker on their prescription that they
6
took was about 30 percent. For
more than
7
two-thirds of the women surveyed there is doubt
8
about compliance with the use of the S.M.A.R.T.
9
yellow prescription program. So,
our data set of
10
women, which may be unusual, indicates that these
11 are
some of the issues that they believe they
12
noted.
13
[Slide]
14
In response to participation in any of the
15
manufacturer surveys, this is where we were
16
disappointed that about 13 percent of them reported
17
contacting the survey; 65 percent reported they did
18 not
participate in any of them and actually 18
19
percent couldn't be sure, but it is different than
20 the
reports that were seen so this may be a
21
subpopulation. This may be due to
the new
22
management for the change of the Accutane survey;
310
1
dual surveys for brands and generic preparations;
2 and
different names and appearances of the risk
3
program as it moved through the process.
But we
4 are
talking here 13 percent.
5
[Slide]
6
Strengths of our study--these interim
7
results presented here should be considered in
8
light of strengths and limitations.
All the staff
9 has
extensive experience in communicating with
10
women about their reproductive concerns.
This
11
experience was used in designing a survey that
12 elicited information while respecting the
difficult
13
issues that these women were facing.
The survey
14
used a detailed, structured interview instrument.
15
Most interviews were completed within three months
16 of
exposure and before the status of the fetus was
17
known.
18
Some of the limitations here are obviously
19 the
small numbers because it is an interim study.
20 All
of our estimates were based on women's recall
21 of
events which may have been influenced by a
22
number of factors. Finally, women
who call a
311
1
Teratogen Information Service and agree to
2
participate in our survey are likely not
3
representative of all women who take isotretinoin,
4 and
may not be representative of all women with an
5
isotretinoin exposed pregnancy but obviously we
6
have demonstrated a very unusual population.
7
[Slide]
8
So, our preliminary conclusions indicate
9
that the data collected in our studies of
10
isotretinoin exposure in the U.S. prior to
11
institution of the S.M.A.R.T. system found similar
12
rates of non-compliance with U.S. and Canadian
13
programs. These data illustrate
that preventable
14
exposures continue to occur due to non-compliance
15
with current requirements, and that perhaps this
16
survey information will be helpful to you and
17
others as one begins to use this qualitative data
18 to
identify risk factors for exposure.
19
One of the things that really came out to
20 us
in this survey is how important the counseling
21 to
this group of patients related to contraception
22 and
to pregnancy test is so critical. The
optional
312
1
need or opportunity to refer to a woman's
2
healthcare provider may be one that we need to look
3 at
even more closely.
4
[Slide]
5
Before taking questions, I would like to
6
close by saying that the desire for a healthy child
7 is
nearly a universal one. OTIS is
dedicated to
8
helping women and their doctors in this endeavor
9 and
we hope that this OTIS survey will be useful in
10
identifying missed opportunities for preventing
11
exposure to this powerful teratogen during
12
pregnancy.
13
Finally, as a personal note, one question
14 I
would like to raise with everyone around the
15
table here is when we are talking about getting a
16
dermatologist certified and we all are constantly
17
having to go through a recertification processes,
18 but
it hasn't been mentioned at all at the table
19
here whether in fact having updates in requiring
20
recertification, reinforced with those that are
21
prescribing, is of importance to what we are doing
22
here today. Thank you, and I am
open to questions.
313
1
DR. GROSS: Thank you, Dr.
Miller. Our
2
first question is from Dr. Epps.
3 Questions to OTIS from the
Committee
4
DR. EPPS: Yes, we recertify in
5
dermatology, and I can assure you a lot of the
6
things you have addressed as far as continuing
7
education goes on--the American Academy of
8
Dermatology at all of our meetings at the local
9
level, the national level and our publications we
10 talk
about this drug; we talk about Accutane all
11 the
time.
12
I would like to ask you, however, there
13 are
some other drugs and substances which are
14
associated with embryopathy, everything from TORCH
15
infections, radiation exposure, but specifically
16
medications which include anticonvulsants which are
17
commonly used, also aminopterin and other
18
medications reported. Someone
mentioned
19
methotrexate which is definitely a category X
20
medication. Do you all also
monitor pregnancy
21
rates with other category X programs?
We know
22
about thalidomide; we have talked about that. But
314
1 I
would counter that, you know, certainly
2
methotrexate is used for juvenile rheumatoid
3
arthritis, as well as some psoriasis and certainly
4
pregnancy potential is there for those patients.
5
Those happen in young people all the time, those
6
conditions, and are we treating this drug
7
equivalently?
8
DR. MILLER: I think you know the
answer
9 to
that one, that we are not treating them
10
equivalently. We are treating
this drug much more
11
like a thalidomide than we are treating them like a
12
valproic acid or diphenylhydantoin.
Perhaps some
13 of
the reasons for that can come from the fact that
14
many of these patients are on them chronically and,
15
therefore, someone who is taking an anticonvulsant
16 if,
in fact, she is ever going to become
17
pregnant--and for phenytoin we are down around the
18 10
percent level of the phenytoin syndrome, she may
19
never be able to get off that medication.
20
Obviously, we want to get folks off the valproic
21
acid because of the neural tube defects.
This
22
needs to be thought of up front with the healthcare
315
1
provider who is prescribing, the neurologist or
2 whoever,
and also the obstetrician/gynecologist in
3 her
planning for a pregnancy. In those
particular
4
instances that is certainly a bit different than
5
what we are talking about here about a rather
6
acute, even though only a five-month, exposure.
7
DR. EPPS: Also, I don't know if
we have
8
mentioned alcohol--
9
DR. MILLER: Yes, Sidney Wolfe
didn't do
10
that this morning. He said the
two most critical
11
ones and he left out alcohol. I noticed
that too.
12
DR. EPPS: Fetal alcohol syndrome
is
13
certainly a lot more common.
14
DR. MILLER: Yes.
15
DR. GROSS: The next question is
from Dr.
16
Gardner.
17
DR. GARDNER: Can we ask the
manufacturers
18 to
address the question of sampling?
19
MS. REILLY: Tammy Reilly. We absolutely
20 do
not sample isotretinoin. What we do is
we have
21 a
medical needs program that provides drug for
22
indigent patients. We do that for
all of our drugs
316
1 at
Roche Pharmaceuticals because we feel that any
2
patient who is under-insured or uninsured should
3
have the opportunity to get our drugs.
So, I can
4
only surmise that perhaps the sample that this
5
person is referring to might have been through that
6
program. What is required in that
program
7
specifically for Accutane, which is different from
8 any
other drug that we have and provide at Roche
9
through this program, is that in addition to the
10
form that the physician must fill out to qualify
11 the
patient from a financial perspective, we also
12
require that they actually include the sticker on
13 the
form that states that they have qualified the
14
patient according to the contraindications and the
15
warnings of the package insert when they send in
16
that form. So, we would not
expect that they have
17 not
gone through the rigorous process that they
18
would be expected to do through a normal
19
prescription process.
20
DR. GROSS: Dr. Whitmore?
21
DR. WHITMORE: Dr. Miller, when
you were
22
asking about recertification, did you mean with the
317
1
S.M.A.R.T. program?
2
DR. MILLER: Yes. If you have to do it
3
once, why not do it again?
4
DR. WHITMORE: You are right, we
only do
5 it
one time and I agree with you.
6
DR. GROSS: Dr. Bergfeld?
7
DR. BERGFELD: Thank you. I just wanted
8 to
have addressed somewhat of an inconsistency.
If
9
those two cases that you used as illustrations were
10
typical of those who called in and actually
11
reported to you so you could interview them, it is
12
hard for me to imagine that if they were able to do
13
that that they would not have read the information
14
that the physician gave them, signed it and sent it
15
because to make this call is quite a step. So, I
16 see
there is inconsistency here in the type of
17
person that we are talking about.
18
DR. MILLER: Well, I agree
wholeheartedly
19
with you that you have a subset of the general
20
population and you are using numbers of 96 percent
21
successful with writing the qualification. I
22
imagine the subset of women we are talking about
318
1
here is probably well below one percent of the
2
users of this drug. But these
are, at least in
3
this case, 24 women who have, unfortunately, not
4
been successful with the Pregnancy Prevention
5 Program and, for a variety of reasons along
these
6
lines, many of them shared with the healthcare
7
professionals in terms of misunderstandings but
8
also problems with not using the right form of
9
contraception. But these are the
ones that are
10
falling through the cracks.
Obviously, most women
11 are
doing well. We are trying to identify
what is
12
that population that isn't doing well.
This is
13
where we come up with all of these inconsistencies
14 because
they are a subset of the total group that
15 is
certainly different. I don't know if I
answered
16
your question.
17
DR. BERGFELD: I guess the bottom
line is
18
they didn't read the information given to them and
19
shared with them by their physician and then they
20 are
smart enough to call your group. There
is an
21
inconsistency in that they can understand the
22
directions. Compliance is another
thing but
319
1
consistency of understanding, and you indicated
2
that they didn't understand.
3
DR. MILLER: Well, the part that
they
4
didn't understand, that was when their OB did a
5
procedure that then said, gee, you probably will
6
have difficulty getting pregnant for the next year.
7 She
interpreted that as, well, I can't get
8
pregnant. She passed that on to
another healthcare
9
provider who said, well, if you can't get pregnant,
10 then you fall into that group and, therefore,
we
11 can
give you the drug. So, there was a
variety of
12
levels of misinterpretation for that patient. But
13 I
agree in general. If they have signed
all of
14
these consent forms and have managed to get this
15 far
along, they should be able to do better, but
16
this is the nature of that subgroup that is getting
17
pregnant and that is the group we are trying to
18
help.
19
DR. GROSS: Robyn Shapiro?
20 DR. SHAPIRO: Along those lines, and I am
21
assuming the answers here, but did you go back and
22
check with the providers about the accuracy of the
320
1
report of the patients about what exactly happened?
2
DR. MILLER: At this point in time
we are
3
providing you interim results. We
have not
4
contacted any providers at this point and this
5
survey is based upon what women are, in fact,
6
reporting to us. In this
particular situation one
7 may
defensively point fingers in different
8
directions too.
9
DR. SHAPIRO: So, do you have
plans to do
10
that?
11
DR. MILLER: As part of this
survey we do
12 not
have permission at this point to contact the
13
providers and we have contact with the patient.
14 So,
this may be a possibility in the future but our
15
IRBs have not approved us to do that, and certainly
16
Paula Knudson would not want us to do that without
17
approval.
18
DR. GROSS: Thank you all very
much. It
19 is
time for a break. We will reconvene at
3:15.
20
[Brief recess]
21
DR. GROSS: We have two more
presentations
22
before we close. You have all had
a chance to call
321
1
your offices and know what misery greets you on
2
Monday.
3
[Laughter]
4
I see a lot of nodding heads.
Okay, Dr.
5
Kathleen Uhl, working at the FDA on the pregnancy
6 and
labeling team, will talk about risk management
7
options for pregnancy prevention.
8
Risk Management Options for Pregnancy Prevention
9
DR. UHL: Thank you and good
afternoon.
10
[Slide]
11
The goals of this talk are, first, to
12
describe the general principles of what constitutes
13 a
teratogen, and then to describe the elements that
14 go
into the decision-making process regarding risk
15
management strategies to prevent fetal exposure to
16 a
particular drug and, lastly, to describe existing
17
strategies that are used to prevent pregnancy and,
18
therefore, prevent fetal exposure.
19
[Slide]
20
For most people, and in very simple terms,
21 a
teratogen is an agent or a factor that causes
22
birth defects or congenital malformation. There
322
1 are
some other definitions that are more
2
scientifically driven. One
example is here on the
3
slide, that a teratogen is an agent or factor that
4
causes the production or physical defects or
5
abnormal development in an exposed embryo or fetus.
6
One of the common misconceptions though about the
7
word teratogen is that people think that exposure
8 to
the agent or drug will always result in abnormal
9
fetal development.
10
[Slide]
11
It is probably more accurate to use the
12
terminology teratogenic exposure as opposed to
13
teratogen. Teratogenic exposure
implies that a
14
drug or agent has teratogenic potential at
15
clinically relevant doses that are used in humans.
16
Despite exposure to a particular drug or agent, at
17 the
right dose and at the right time in pregnancy,
18
this does not necessarily result in a birth defect
19 so
the teratogenic effect is not 100 percent.
20
[Audio technical difficulty with slide 5
21 and
6]
22
... a teratogen is with animal data.
323
1
Animal studies are typically performed in the
2 pre-marketing phase of drug development to
assess
3
reproductive risk. The animal
data are
4
particularly useful for generating signals about
5
whether a drug may be a human teratogen.
The
6
assessment is based on the totality of evidence
7
from animal data as well as what is known about
8
drugs with similar pharmacologic activity. If the
9
data are concerning one could conclude that a drug
10 is
highly suspected to be a human teratogen.
In
11
this case the drug is not yet proven to be a human
12
teratogen. Based on animal data
alone we will
13
never be able to conclude that a drug is a human
14
teratogen.
15
[Slide]
16
The way that we do that is with human
17
data. Typically, it takes years
even decades to
18
generate sufficient human pregnancy exposure data
19 to
conclude that a drug is a human teratogen.
20
There are some sources that are very useful in
21
assessing whether or not a particular drug is
22
associated with teratogenicity.
These include such
324
1
things as adverse event reports, those that are
2
reported to regulatory authorities and the example
3 here
would be the MedWatch forms. Case reports,
4
case series, case control studies that are found in
5 the
medical literature are very useful. In
the
6
case of isotretinoin there were case reports of
7
teratogenicity within the first year of marketing
8 in
the U.S.
9
Another source of information are the data
10
that come from pregnancy exposure registries which
11 are
prospective epidemiologic studies or data that
12 can
come from other post-marketing studies.
The
13
peer reviewed assessments that are done by the
14
Organization of Teratogen Information Services and
15
also by TERIS are typically not independent data
16
sources and they really serve as a resource that
17
helps pull together all the various data sources
18
into one location.
19
[Slide]
20
Now what I want to do is move on to
21
discuss how we go about a decision-making process
22
regarding risk management to prevent fetal
325
1
exposure.
2
[Slide]
3
When developing and implementing pregnancy
4
prevention strategies it is important to match the
5
strategy to the level of concern that you have for
6
that drug. Here we have listed
three levels of
7
concern just for simplification purposes. When the
8
data do not indicate fetal risk the level of
9
concern would obviously be quite low.
The next
10
level of concern would be when we have animal data
11 or
other sources that are very concerning and, in
12
that case, we would consider the drug to be highly
13
suspect to be a teratogen. The
highest level of
14
concern is for those products that are known human
15
teratogens.
16
It is important to keep in mind that not
17 all
teratogens are equal and you need to consider
18 the
frequency, the severity of adverse fetal
19
outcome, the reversibility and the timing of
20
exposure in the selection of pregnancy prevention
21
strategies.
22
[Slide]
326
1
Another aspect of the decision-making is
2 to
understand what is the risk. In order to
do
3
that you need to address some specific aspects for
4
fetal risk. For example, the
frequency of the
5
event, is the event of high frequency or low as it
6
compares to the background risk for congenital
7
anomalies?
8
What is the severity of outcome?
By
9
regulation, all birth defects are considered
10
serious. However, not all birth
defects are equal.
11 For
example, there are some birth defects that are
12
incompatible with life. There are
other birth
13
defects that are cosmetic. The
example of that
14
would be the tooth staining that can occur
15
following exposure to tetracycline.
There are
16
other birth defects that are reversible.
There are
17
some congenital heart malformations that can be
18
surgically corrected.
19
The type of abnormality is important to
20
consider, and those were reviewed on a previous
21
slide. The timing of exposure is
a critical item
22 to determine in assessing risk. When during
327
1
pregnancy is exposure associated with the highest
2
risks to the developing fetus?
For example, if a
3
birth defect is associated with first trimester
4
exposure, that actually rather broad range, could
5 the
time of exposure be narrowed to a very defined
6
time, for example the ninth to the eleventh week of
7
gestation? Or, as in the case
with ACE inhibitors,
8 is
the birth defect associated only with second and
9
third trimester exposure? With
the ACE inhibitors
10
therapy can be easily discontinued or changed prior
11 to
the critical time of exposure for adverse fetal
12
outcomes. The earlier in
pregnancy the adverse
13
fetal effects occur, the more likely it is that
14
pregnancy prevention strategies are needed. The
15
severity and the type of adverse fetal outcomes
16
affect our perception of "badness" and help to
17
drive the decision to implement pregnancy
18
prevention strategies.
19
[Slide]
20
Another element that goes into the
21
decision-making process is the consideration of
22
maternal disease. Maternal
disease itself may
328
1
carry increased risk for birth defects.
The
2
example that has already been given is that of
3
diabetes. Untreated maternal
disease may have
4
serious untoward consequences on the health of the
5
mother or the fetus. An example
would be untreated
6
seizure disorders. The benefits
of treatment are
7
important. The risk/benefit to
the mother and the
8
fetus colors our willingness to accept birth
9
defects or adverse fetal outcomes.
10
[Slide]
11
Despite the many drugs that are on the
12
market, there are only a few and some sources say
13
that there approximately 19 drugs or groups of
14 drugs
that are believed to be teratogenic in
15
humans. Several of the known
human teratogens do
16 not
have specific pregnancy prevention strategies.
17
What I would like to do here is contrast
18 two
well-known human teratogens, warfarin and
19
isotretinoin. The toxicity for
warfarin is
20
well-known and the warfarin embryopathy has been
21
well characterized. The
teratogenic risk is
22
relatively small and occurs at relatively low
329
1
rates. The window of exposure for
teratogenic risk
2 is
highest in the second half of the first
3
trimester and the window is even narrower, six to
4
nine weeks, of fetal life.
5
Females of childbearing potential
6
represent a small percentage of patients who are
7
taking the drug, and the patient-provider
8
relationship is typically more comprehensive owing
9 to
the long-term treatment of chronic medical
10 conditions such as continued anti-coagulation
for
11
artificial heart valves or for thromboembolic
12
disorders.
13
With isotretinoin some of the toxicity is
14
well-known, in particular the structural
15
malformations are well-known.
There are other
16
toxicities that are not as well recognized. The
17
risk is larger and occurs at higher rates. The
18
window of exposure is highest in very early
19
pregnancy and some have estimated that the highest
20 risk
period is from three to five weeks. The
21
overall use of isotretinoin is high in females of
22
childbearing potential and the patient-provider
330
1
relationship is of relatively short duration and
2
targeted to a single medical problem, that being
3
acne.
4
[Slide]
5
There is no question and we are not
6
debating whether isotretinoin is a human teratogen.
7
There are multiple developmental abnormalities that
8 are
associated with isotretinoin exposure to the
9
developing fetus. As illustrated
here, there are
10
multiple structural malformations including
11
craniofacial and ear findings. It
is estimated
12
that the full spectrum of retinoid embryopathy
13
occurs in 20-30 percent of exposed fetuses, as was
14
mentioned earlier this morning by Dr. Lindstrom.
15
However, it is believed that even higher
16
numbers of exposed fetuses have single structural
17
malformations. The work done by
Adams and Lammer
18 in
the early '90s demonstrated a high incidence of
19
intellectual deficits in children who were exposed
20
early in the first trimester. The
intellectual
21
deficits were found in children both with and
22
without major structural malformations.
331
1
Another developmental abnormality that is
2
associated with teratogenicity in general is fetal
3 and
infant mortality, and isotretinoin is
4
associated with increased spontaneous abortion and
5
premature birth.
6
The critical period of exposure is very
7
early in gestation and is reported to be from the
8
28th to the 70th day of fetal development, with the
9
most critical time being the third to fifth week.
10 It
is important to note though that no apparent
11
relationship has been found between the duration of
12 exposure
and resulting malformations. Teratogenic
13
outcomes consistent with retinoids have occurred
14
following only one dose of isotretinoin during
15
pregnancy.
16
Isotretinoin has unique pharmacokinetics.
17 The
half-life of isotretinoin and its major
18
metabolite are approximately 24 hours.
So, even
19
with immediate cessation of drug it will take
20
approximately two weeks for 99 percent of the drug
21 and
metabolite to be cleared from the body.
It is
22
important to remember that stopping the drug does
332
1 not
result in immediate cessation of exposure to
2 the
fetus.
3
[Slide]
4
To reiterate--as if we haven't seen this
5
slide enough times--the goals of pregnancy
6
prevention are that pregnant women do not receive
7 the
drug and that females of childbearing potential
8 do
not get pregnant while taking the drug.
In
9 addition,
this second goal would also encompass the
10
30-day period after drug has been stopped.
11
[Slide]
12
For products for which there is a concern
13 to
the developing fetus the agency has historically
14
labeled drugs in two different ways, taking into
15
consideration maternal disease, the population of
16
intended use and the frequency and severity of
17
adverse fetal outcome.
18
If it is felt that the benefits of
19
maternal drug use outweigh the drug's potential
20
risks, then the drug is labeled as category D,
21
pregnancy category D and, by regulation, there must
22 be
specific wording that is included in the warning
333
1
section of the label.
2
If it is felt that the benefits of
3
maternal drug use do not outweigh the drug's
4
potential risks, then that drug should not be used
5 in
pregnancy and it is contraindicated in
6
pregnancy. The product is labeled
as a pregnancy
7
category X and, by regulation, there is some
8
specific wording that must be included in the
9
contraindications section of labeling.
10
[Slide]
11
On this slide are just some
examples of
12
known human teratogens and highly suspect human
13
teratogens that are contraindicated for use in
14
pregnancy. This is not intended
to be an inclusive
15
list. Simply contraindicating the
drug alone does
16 not
equate with true pregnancy prevention
17
strategies.
18
[Slide]
19
Beyond contraindicating the drug in
20
labeling, there are additional pregnancy prevention
21
strategies that can be utilized. This slide has
22
several examples that are informational.
They
334
1
include the black box warning.
Any information in
2 the
black box warning must also go into product
3
advertising. There could be
wording in the warning
4
sections or in other sections of the labeling. The
5
informed consent documents can be included in
6
labeling and are either advised to be used or
7
included when the physician is writing for this
8
drug, and a medication guide which is required by
9 law
to be issued when the drug is dispensed.
10
[Slide]
11
There are other pregnancy prevention
12
strategies that are active interventions. These
13
include such things as pregnancy testing and
14
contraceptive use. These
typically require the
15
healthcare provider and the patient to actually do
16
something.
17
[Slide]
18
The strategy of pregnancy testing
19
addresses that first goal of pregnancy prevention,
20
that no pregnant woman will get the drug. It is
21
useful as a risk management strategy in preventing
22
that only with the first pregnancy test that is
335
1
done. Pregnancy testing does not
address the
2
second goal, that women do not get pregnant while
3
taking the drug. It allows for
early detection of
4
pregnancy and prevention of further exposure to the
5
developing fetus.
6
But simply stating that a pregnancy test
7
should be done is likely to be insufficient as a
8
pregnancy prevention strategy.
Other aspects of
9
pregnancy testing need to be addressed, such as
10
when to start pregnancy testing in relation to
11
beginning the drug and the number of pregnancy
12
tests that should be performed.
Also, how
13
pregnancy testing should be continued throughout
14
therapy is important. For
example, should the test
15 be
done monthly or periodically without any
16
specificity as to what that periodicity would be?
17
How should pregnancy testing be continued
18
after stopping the drug? It is
hoped that this
19
would be driven by the drug's pharmacokinetic
20
properties. Also, the test
specifics such as test
21
sensitivity, the types of tests or the setting for
22
testing.
336
1
[Slide]
2
This next strategy of contraception
3
addresses the second goal of pregnancy prevention,
4
that women do not get pregnant while taking drug.
5
Contraception does not address the first goal of no
6 pregnant woman getting the drug. Again, simply
7
stating that contraception should be used is likely
8 to
be insufficient as a pregnancy prevention
9
strategy and other aspects of contraception use
10
need to be addressed, such as when to start
11
contraception in relation to beginning therapy;
12
that contraception be used consistently throughout
13
drug therapy; how long to continue contraception
14
after stopping the drug, again a recommendation
15
that would be based on the drug's pharmacokinetic
16
properties; and the specifics of contraception such
17 as
the types and the numbers of acceptable methods.
18
[Slide]
19
There are numerous other pregnancy
20
prevention strategies that could be used. The
21
strategies that are listed here are not specific to
22
pregnancy prevention and have been used more
337
1
generally in other risk management programs. These
2
strategies will actually be discussed in more
3
detail in the next presentation by Dr. Trontell.
4
[Slide]
5
It is unlikely that the two goals of
6
pregnancy prevention will be met if the patient and
7 the
provider do not understand the risk to the
8
fetus and actively work to mitigate it.
Strategies
9
should include that females of childbearing
10
potential are definitely informed of risk to the
11
developing fetus but, in addition to being
12
informed, the woman must understand the risk and
13 she
must demonstrate behavior that is commensurate
14
with the understanding of that risk.
15
[Slide]
16
Strategies to prevent pregnancy are very
17
complex. The ultimate goal of
pregnancy prevention
18 is
to prevent fetal exposure to the drug both at
19
drug initiation and with continued use of the drug.
20 It
is important to remember that not all teratogens
21 are
equal, therefore, pregnancy prevention
22
strategies must be tailored to the specific drug.
338
1
This is not a "one size fits all" process.
2
DR. GROSS: Thank you, Dr. Uhl. The next
3
speaker is Dr. Anne Trontell, who is Deputy
4
Director, Office of Drug Safety at the FDA. She
5
will talk about selecting risk management tools:
6
considerations and experience.
After her
7
presentation we will take questions.
Thank you.
8 Selecting Risk Management
Tools:
9 Considerations and Experience
10
DR. TRONTELL: Good afternoon.
11
[Slide]
12
As Dr. Gross told you, I will be
13 describing some considerations and experience
that
14 FDA
has with selecting risk management tools.
15
[Slide]
16
I will start with two definitions of risk
17
management program goals and of risk management
18
program tools. Then I will state
some general
19
considerations in selecting tools for risk
20
management. I will then
recapitulate some of the
21
concerns that have been expressed today with the
22
current isotretinoin risk management program, and
339
1
then suggest some candidate tools that might
2
address those concerns. I will
describe two
3
related risk management programs and compare them
4 to
the isotretinoin risk management program, and
5
then give you some impressions about the relative
6
advantages and disadvantages of some of the tool
7
options available to us.
8
[Slide]
9
In the context of risk management
10
programs, goals are described as the ideal product
11 use
scenario or vision statement. We have
heard
12
about them many times today.
These are tailored to
13 the
product-specific risk concerns and, as goals,
14
stated in absolute terms may not be fully
15
achievable. An example of a goal
is that no fetal
16
exposure shall occur and the two goals of the
17
isotretinoin program have been articulated several
18
times today.
19
[Slide]
20
Again, in the context of risk
management
21
programs, tools are defined as those processes or
22
systems that are intended to enhance safe product
340
1 use
by reducing risk. These tools are chosen
2
considering the severity, reversibility and the
3
frequency of the risk that is attempted to be
4
minimized.
5
[Slide]
6
Some general considerations are available
7 in
selecting risk management tools.
Ideally, each
8
tool should add value in attaining the risk
9
management program goal. In
choosing tools, one
10
should seek those tools that have proven
11
effectiveness in reducing risk.
One should also
12 seek
acceptability by the healthcare system and by
13
patients as well as practitioners, and low burden
14 on
the healthcare system. In selecting
tools, one
15
should also seek to avoid those that place
16
unnecessary limitations on the beneficial uses of
17 the
product. In recognition of the potential
18
confusion and burden, one should also avoid the
19
creation of multiple customized tools that might
20 add
to confusion. Insofar as possible, one
should
21
seek, in designing tools for risk management, to
22
avoid unintended consequences or paradoxical
341
1
worsening of risk in attempts to reduce it.
2
[Slide]
3
For purposes of this presentation and
4
discussion, I am going to talk about tools in four
5
broad categories and, quite frankly, spend most of
6 my
time talking about those that go outside of the
7
product labeling that Dr. Uhl described so well for
8 you
just a few minutes ago. Just to be
clear,
9
product labeling, sometimes referred to as the
10
package insert or PI, is largely targeted to
11
healthcare practitioners--physicians and
12
pharmacists mostly.
13
So, the first category that
I would like
14 to
describe I will term education and outreach.
15
This involves any of a variety of educational
16
materials that might be given to patients or to
17
practitioners. This might take
the form of
18
brochures or videos or patient information such as
19
patient package inserts of medication guides.
20
The second broad category is one that is
21
difficult to capture in a single term.
For want of
22 a
better one, I will refer to them as reminder or
342
1
prompting systems. These are
voluntary systems
2
that make it easier for individuals, be it
3
physicians, pharmacists or patients, to do the
4
right thing and follow the necessary risk reduction
5
efforts. These may take the form
of stickers,
6
informed consent or limitations on product supply
7 or
unusual product packaging.
8
The last category that I will describe is
9
limited distribution systems.
These restrict the
10
prescribing, dispensing and use of a product to
11
selected groups of physicians, pharmacists or
12
patients, and they are often linked to mandatory
13
compliance with some form of risk management.
14
[Slide]
15
In terms of the experience that we have in
16 the
agency with tools being applied in risk
17
management, when it comes to product labeling and
18
education outreach, in fact, there is quite a large
19
number of programs. Obviously,
all products have
20
package inserts that have been approved by FDA.
21 So,
there is extensive use of educational material,
22
however, evaluation of these materials for their
343
1
effectiveness has been done in very few instances
2 and
some of those evaluations, looking at changes
3 in
labeling or at "dear healthcare practitioner"
4
letters have shown somewhat disappointing results
5 in
terms of their ability to affect behavior.
6
With respect to the reminder or prompting
7
systems, these have been used relatively
8
infrequently so we have relatively few models upon
9
which to base our experience.
Effectiveness of
10
these has largely been untested and, in fact, the
11
evaluation today of the isotretinoin risk
12
management program represents an important step
13
forward in evaluating such systems.
14
The last category of tools, those
15
involving limitations and distribution, are used
16
most uncommonly, downright rarely.
These typically
17 are
used for small patient populations that have
18
limited therapeutic options. In
these programs the
19
logistics of setting up the limited distribution
20
system involve some form of registration of the
21
participants and that registration, in fact,
22
enables daily collection that has allowed us to
344
1
evaluate their effectiveness and those have been
2
demonstrated effective at this point.
I will
3
elaborate further shortly.
4
[Slide]
5
Again, to cement these categories for the
6
discussion to come, some examples of drug products
7 in
the reminder or prompting system would include
8
isotretinoin, the closely related program that is
9 in
place for alosetron and yet another one to give
10 you
as an example is the drug product lindane,
11
where in the past year restrictions have been made
12 on
the amount of that product that can be dispensed
13 to
patients to minimize the likelihood of over-use
14
that has led to toxicity.
15
With limited distribution we have
16
approximately six programs listed here, bosentan,
17
clozapine, dofetilide, mifepristone, thalidomide
18 and
xyrem. Those that have the asterisk by
them
19 are
ones where laboratory testing is required as
20
part of these programs.
21
[Slide]
22
Let me now turn to some of the concerns
345
1
that have been expressed today with the performance
2 of
the current isotretinoin risk management
3
program. As Dr. Pitts indicated,
refills are not
4
supposed to occur. They have been
reduced in the
5
current program but still occur at a rate of
6
approximately 2.5 percent of all the prescriptions.
7 We
have evidence from the pharmacy compliance
8
survey and from patient reports that some
9
prescriptions, a relatively small amount, are being
10
filled without the stickers being present. Perhaps
11
more concerning is the self-reported data from
12
patients suggesting that stickers may be being used
13
without concordant pregnancy testing, estimated at
14 9
percent from our analysis of the Degge survey.
15
[Slide]
16
The other important issue however that we
17
really want to address is those issues of pregnancy
18
exposures that have occurred. We
have in our
19
analysis at FDA estimated that the number of
20
patients initiating therapy while pregnant is
21
approximately 6 percent. We have
heard other
22
estimates from the sponsors today that it may be 13
346
1
percent or perhaps as much as 25 percent. In some
2
instances this represents inadequate pregnancy
3
testing, perhaps two tests not being done, and
4
correct timing and other forms of errors.
5
There are other pregnancy exposures that
6
have been reported to us voluntarily, and the
7
majority of those are those that have occurred
8
during isotretinoin therapy or in the month
9
following. The reports that come
to us have
10
suggested these largely reflect problems either in
11
poor or absent contraception by patients. In some
12
instances the case reports have described
13
individuals who plan to be abstinent but had
14
unanticipated sexual activity and were unprepared
15
with contraception.
16
In addition to these concerns, one that
17 has
not been emphasized to any great extent today
18 but
which, through case reports, have been made
19
known to us is that this product is used in some
20
unknown percentage by individuals without the
21
benefit of medical supervision.
There are
22
individuals, some described in the MMWR article
347
1
described previously, who have obtained the product
2
illicitly through the Internet.
They may have
3
borrowed it from a friend or, in some instances,
4 may
have made use of leftover pills from a prior
5
course of therapy.
6
[Slide]
7
An additional concern, not about
8
performance of the system, gets at the issue of
9
evaluating the performance of the program. We are
10
limited in our ability to estimate the extent of
11
pregnancy exposures. We are
relying in our
12
presentations today on voluntary reports and on
13 voluntary patient surveys. As has been suggested,
14 the
existence of multiple surveys actually permits
15 the
possibility that a patient's information may be
16
counted more than once. Also, in
estimating the
17
extent of isotretinoin exposure it is important to
18
know that these are estimates and based upon
19
pharmacy data, and don't let us know the true
20
extent or duration of isotretinoin exposure among
21
females of childbearing potential.
22
[Slide]
348
1
Let me now turn to some of the tool
2
options that we might consider to address each of
3 the
concerns that I have just described.
Looking
4 at
sticker use and appropriate prescribing or
5
dispensing of the product with stickers, one might
6
look to the broad category of education and
7
outreach for opportunities for improvement. There
8
might be better education of pharmacists and
9
physicians to improve what we believe are good
10
faith efforts to prescribe and dispense this
11
product appropriately.
12
If we are to think of the next category of
13
potential tools, one might think of ways that we
14 could
increase the number or types of reminder
15
systems to make it more difficult for individuals
16 to
forget to do what is appropriate. In
this arena
17 we
have limited models to draw on for drug exposure
18
and, as our colleague from Kaiser described, it may
19
be that some of the disease
management models that
20
exist for chronic conditions may be helpful to us.
21
In the last category, where limitations
22
might be imposed upon prescribing or dispensing by
349
1
healthcare practitioners, one might imagine broadly
2
various scenarios where training, certification or
3
registration might be required of practitioners and
4 that
that might also make some requirements for
5
systems to be in place that obligate compliance
6
with key program elements and would actually allow
7 us
better monitoring of program performance.
8
[Slide]
9
Turning to concerns about the extent and
10
completeness of pregnancy testing, two options in
11
this area are very similar to what I have just
12
described. We might try more
effective education
13 for
healthcare practitioners. Similarly, we
might
14
have other mechanisms to try and do a job of
15
reminding them in a better fashion.
Similarly,
16
limitations may be imposed so that only those
17
individuals, through some process, who have been
18
documented to do a good job of pregnancy testing
19
would be allowed to prescribe this product. Again,
20
looking to the experience in the Kaiser program,
21
there may be some opportunity to consider whether
22
documentation of pregnancy test results may, in
350
1
fact, be one mechanism to improve compliance with
2
this form of pregnancy prevention.
3
[Slide]
4
Turning to contraception, I think it is
5 important
that we acknowledge the great
6
difficulties in intervening in what is a complex
7 and
private human behavior. It is clearly
very
8
sensitive for both patients and physicians to
9
discuss this. It is certainly a
great challenge in
10 the
instance of adolescents who may be receiving
11
drug therapy with their parents present, but I
12
would submit for patients of all ages the
13
assumptions and misinformation that might occur
14
around the sensitivities and awkwardness lead to
15
some errors in factual information.
16
The other important challenging factor
17
around intervention and contraception is that
18
behaviors of individuals are, not just
19
contraceptive behaviors, are influenced by
20
knowledge but they are not controlled by knowledge
21 in
that they are complex attitudinal and behavioral
22
components that should be addressed.
351
1
[Slide]
2
Around contraception, however, let's go
3
back to sort of our three-stage model of
4
intervention of tools. If we look
to improving
5
education and outreach to patients we might
6
increase their knowledge about the need for two
7
simultaneous effective methods of contraception and
8
perhaps we may need to address issues of what are
9
ineffective methods of contraception, particularly
10 in
light of the comments from our colleague from
11 OTIS.
Also, I might add to this slide that we may
12
also need to reinforce the importance of continued
13
contraception after terminating therapy.
14
[Slide]
15
If we were to look to the reminder or
16
prompts category of tools, we might look to some
17
form of counseling though that may take any of a
18
variety of forms. We might hope
that counseling
19
would allow some reinforcement of knowledge of
20
appropriate contraceptive behaviors and could
21
address attitudes that might influence appropriate
22
contraceptive use, issues about planned or
352
1
unplanned sexual activity and other sensitive
2
issues that may deal with partner compliance,
3
cooperation or resistance to the use of
4
contraception.
5
These reminder systems could be put in
6
place on a one-time basis. We
might suggest
7
periodic reinforcement would be a preferred
8
strategy but we invite your commentary.
As you may
9 be
aware, there are a number of methods that can be
10 put
in place involving counselors or some
11
technologies that are now available, and were
12
described earlier, such as interactive voice
13
recognition software, moderated chat rooms and so
14
forth.
15
[Slide]
16
Turning to the third category of tools
17
that we might use to address contraception, we are
18
again getting to a difficult and sensitive area but
19 one
might imagine ways that we might try to limit
20
patient access to drug to those who have
21
demonstrated the appropriate knowledge and skills
22 and
behaviors around the use of contraception.
I
353
1
will suggest this could happen perhaps through some
2
form of counselor certification that would attest
3 to
the patient's level of commitment and
4
demonstrated skills in their ability to use the
5
chosen contraception. Again, some
periodic contact
6 via
counselors or some IVR technology might allow
7
screening for high risk behaviors and the
8
opportunity to intervene to direct those people to
9 lower
risk strategies or to temporarily suspend
10
their exposure to the drug. I
think it is highly
11
extreme, but for purposes of discussion, there are
12
models in the case of tuberculosis testing with
13
directly observed therapy. Obviously,
in the area
14 of
contraception this is challenging. You
might
15
look to use of oral contraceptives or contraceptive
16
patches, pill counts or other methods that you
17
might track adherence with contraception.
18
[Slide]
19
In talking about tools, however, we need
20 to
address the issue that contraceptive failures
21
occur, and I use this term broadly to include
22
failures of the method as well as failures in
354
1
practice. One option to be
considered is if some
2
level of contraceptive failure is acknowledged to
3
occur is that we may want to explicitly limit
4
exposure of females of childbearing potential to
5
this product, and to do so perhaps based on the
6
severity of the acne that those individuals
7
experience. Again, the possible
mechanisms that
8
this could be done are through some form of
9
required documentation of acne severity, a prior
10
authorization mechanism, second opinion or some
11
other check mechanism, again, to assure that
12
females of childbearing potential who are exposed
13 to
isotretinoin are only those who have the most
14
severe forms of acne that we saw this morning.
15
[Slide]
16
Talking about medically unsupervised use
17
really taxes our imagination because we are talking
18
about people who, by definition, are going outside
19 the
system where we have the most influence.
But,
20
again, we might hope, back to our education model,
21
that we could educate individuals better about the
22
risks of using these products without medical
355
1
supervision, and perhaps there may be some
2
innovative mechanisms around product packaging
3
where we could make note of the risks of
4
unsupervised use, obtaining it through the Internet
5 or
sharing it with other individuals. One
other
6
possibility might be to, in fact, limit the amount
7
that is supplied to the patient to some amount less
8
than the typical 30-day supply that is currently
9
dispensed to decrease the opportunity for people
10
hoarding or sharing with other individuals.
11
Obviously, the manufacturers of isotretinoin share
12
with FDA a great desire to constrain illicit
13
Internet sales.
14
[Slide]
15
Let me now describe two risk management
16
programs briefly that have been alluded today that
17 may
be relevant for comparison to the isotretinoin
18
program. The first one is
clozapine, an appealing
19
comparison because it has significant safety risks
20 and it has a risk management program that is
21
administered by multiple manufacturers.
There are
22
inter-related data systems that are in place for
356
1
this product and the systems have allowed some
2
evaluation of the program's effectiveness. In
3
fact, that information has been used to relax
4
program requirements over the lifetime of this risk
5
management program.
6
Thalidomide is the other product which has
7 an
extensive and effective risk management program
8
also for teratogenicity. It is
important to note
9
that this program, although very appealing in terms
10 of
its experience, has had very limited use among
11
females of childbearing potential.
Only about 5
12
percent of the total population exposed to this
13
product have been women of childbearing potential,
14
only about 4,000 individuals.
15
[Slide]
16
Clozapine, briefly for those who may not
17 be
familiar, is an antipsychotic that carries the
18
risk of agranulocytosis, and this risk is managed
19 by
a program of weekly to biweekly blood testing to
20
assure that the white count is adequate.
The
21
pharmacist is required to view the white count
22
documentation. The white count
must be presented
357
1 in
order to dispense the product, and only
2
registered pharmacists, patients and physicians are
3
able to access this drug product; not every
4
pharmacy is able to fill these prescriptions.
5
[Slide]
6
For clozapine there is a centralized
7
registry of patients. This is
reserved for those
8
patients who are not to be rechallenged with this
9
drug based upon their prior experience of having
10 had
a lowered white count while taking it.
In
11
addition to the centralized "do not rechallenge"
12
registry, there are independent sponsor programs
13
that permit the weekly and biweekly testing to be
14
done. This program does not have
any patient
15
survey or education, considering that the
16
population receiving it is largely individuals with
17
refractory schizophrenia, but there is extensive
18
education for the providers, both the physicians
19 and
pharmacists, about what they are to do in
20
administering the program.
21
[Slide]
22
For thalidomide the goal is that no fetal
358
1
exposure should occur because of its teratogenicity
2
and, like clozapine, only registered patients,
3
pharmacists and physicians are able to access this
4
drug. Pregnancy testing is done
on female patients
5
according to their pregnancy risk category and that
6 is
based upon their age and fertility status.
7
Physicians report to a centralized database the
8
negative pregnancy status of their patients in
9
order to authorize that prescription being
10
released.
11
[Slide]
12
Patients must also report via IVR on their
13
risk factors for pregnancy exposure.
Those
14
individuals who give responses suggestive of high
15
risk behaviors are routed directly to a live
16
operator of action and potential intervention.
17
Pharmacists, at the time of dispensing
18
this product, check the central database and look
19 to
see if the information from the physician and
20 the
patient are both appropriate and permissive for
21
them to dispense the product. The
system allows
22
through its central database the ability to track
359
1
pregnancy exposures, at least those that are not
2
lost to follow-up. Much like
isotretinoin, there
3 is
an extensive educational program associated with
4
this product. There is a medication
guide,
5
informed consent, a video and many other materials.
6
[Slide]
7
Let me now make some comparison of the
8
three programs, isotretinoin, thalidomide and
9
clozapine. I have abbreviate each
here by their
10
first letter.
11
Warnings exist in physician labeling or
12
package inserts for all of these products. Patient
13
education materials, medication guides and patient
14
informed consent are extensive both for
15
isotretinoin and for thalidomide.
16
[Slide]
17
All programs do require laboratory testing
18 but
there are some subtle differences that I will
19
describe. In the case of
clozapine, actual
20
documentation of test results is required in order
21 to
receive the product. For thalidomide the
form
22 of
laboratory testing can come via the physician
360
1
report of the test being positive or negative. In
2 the
case of isotretinoin, a physician uses a
3
sticker to attest that pregnancy testing has been
4
done and that those test results are negative.
5
[Slide]
6
With patient registration, this is done
7 for
all patients, both male and female for
8
thalidomide. In the case of
clozapine, a central
9
registry is maintained only for those patients who
10 are
not to be rechallenged with the drug product.
11
There is no registry for patients taking
12
isotretinoin.
13
Physician registration is required for
14
prescribing for both thalidomide and for clozapine.
15 For
the case of isotretinoin, physicians enroll in
16 the
program. There is mandatory enrollment
to get
17
stickers, but physicians are technically allowed to
18
prescribe this product without stickers.
19
[Slide]
20
Pharmacy registration is required to
21
dispense both thalidomide and clozapine but not for
22
isotretinoin.
361
1
[Slide]
2
When it comes to tracking program
3
performance, patient behaviors are captured through
4 the
IVR component of the thalidomide system, and
5 for
the isotretinoin program it is captured for the
6
proportion of patients that respond to the
7
voluntary patient survey. Patient
exposures are
8
captured for all patients taking thalidomide and
9
clozapine is captured uniquely only for those
10
patients who are not to be rechallenged with the
11
drug. There is direct tracking of
outcomes by the
12
thalidomide and clozapine program, thalidomide of
13
pregnancy test results and clozapine of white
14
counts. For isotretinoin there is
voluntary
15
reporting, either directly to the agency or through
16 the
survey, of patient behaviors.
17
[Slide]
18
Let me now discuss some of the advantages
19 and
disadvantages we might think of broadly when we
20 are
thinking of the various categories of tools
21
that I have described. If we were
to consider
22
increasing education or outreach for isotretinoin,
362
1 we
could see as advantages that this is an option
2
that is likely to be acceptable to most of the
3
participants in this therapeutic relationship. It
4 is
feasible. It would involve no change in
product
5
access and might allow time for the current program
6 to
see if experience changes or improves.
7
Disadvantages in the area of education and
8
outreach are that education and outreach alone have
9 not
yet shown a good track record of effectiveness,
10 and
it may be particularly challenging to think of
11
changing contraceptive behaviors simply by adding
12
education and outreach.
13
[Slide]
14
If we were to turn to reminder and
15
prompting systems as a way to try and fortify the
16
response in the isotretinoin risk management
17
program, advantages of such an approach might
18
include the continued autonomy of physicians,
19
pharmacists and patients. There
is no mandatory
20
component. By definition, these
are guiding little
21
nudges to make sure you are supposed to do what is
22
right. There might be an
opportunity with these
363
1
reminder systems to allow ongoing education and
2
reminders about the risks of the product and how to
3 use
it most safely. In comparison to the
next
4
category of tools I will describe, limited
5
distribution is certainly much less intrusive.
6
Disadvantages in this category, as I have
7
stated previously, are that for drug therapies we
8
have limited experience and limited models to
9
borrow from. The effectiveness of
these are
10 unknown,
and you have seen the experience with the
11
current program outlined for you today.
As has
12
been mentioned by some individuals already, we need
13 to
consider the time and financial burdens that
14
might be entailed with counseling or some form of
15
disease management around this issue.
16
[Slide]
17
Turning now to the last category of
18
limitations imposed on the prescribing, dispensing
19 or
use of this product, potential advantages in
20
designing a program around such a system would be
21
that we could theoretically limit access of the
22
product to those individuals who would adhere to
364
1 the
critical risk minimization tools--counseling,
2
pregnancy testing, and so forth.
3
Again, the nature of limiting distribution
4
requires that you have to have some list of who can
5 and
can't give the product and that listing, in
6 fact, gives you the ability to register
individuals
7 and
have better data to evaluate the program in
8
terms of its baseline performance or changes over
9
time.
10
An additional advantage, whether it be
11
done explicitly or implicitly, is that the burdens
12 and
logistics of such programs are likely to limit
13 the
exposure of females of childbearing potential,
14 if
only for the perceived tassel of complying with
15 the
limited distribution program.
16
In terms of disadvantages for isotretinoin
17 in
comparison to at least the limited distribution
18
program that we have as a model for thalidomide,
19 the
effectiveness of the S.T.E.P.S. program in
20
young, fertile women is not yet documented. Again,
21 we
would consider that restricted or limited
22
distribution programs would impose time and
365
1
financial burdens upon the healthcare system and
2
would probably introduce a real possibility of
3
limiting access to the drug to those people who
4
might legitimately benefit from its use.
It is
5
also important to recognize that the creation of
6
substantial barriers around product access may
7
increase the incentive for individuals to go
8
outside of the system, obtain the product illicitly
9
and, in that way, bypass any access to some of the
10
safety measures that we are trying to promote.
11
[Slide]
12
Let me conclude by reminding you of the
13
general considerations I used at the beginning of
14 my
talk. If we are to consider modifying or
15
selecting new risk management program tools for
16
isotretinoin we want to keep these considerations
17 in
mind. First, that we seek evidence for
18
effectiveness of any tools and the high likelihood
19
that tools added would add value to the attainment
20 of
the risk management program goals; that drawing
21
upon effectiveness, familiarity and acceptability,
22 we
would like to try and stay as close as possible
366
1 to
familiar tools that are already being applied in
2
practice.
3
Where possible, we should avoid
4
unnecessary limitations on product use, and we
5
should anticipate that any of our interventions are
6
likely to impose time, cost and access burdens by
7
constraining access to the product, and we should
8 be
wary of the possibility of unintended
9
consequences of our very good intentions. Thank
10
you.
11
Question to the Speakers from Committee
12
DR. GROSS: Thank you very much,
Anne.
13
That was extremely helpful. The
floor is now open
14 for
questions from the committee members for these
15 two
speakers. Dr. Gardner?
16
DR. GARDNER: Dr. Trontell, I
don't think
17 it
really came home to me until I heard your words
18
literally saying that the registration of
19
physicians for isotretinoin is voluntary and that
20
physicians do it to get yellow stickers but
21
technically they can prescribe without yellow
22
stickers. So, that triggers two
thoughts in my
367
1
mind. One is that, therefore, the
burden of
2
overseeing this effective stopping is really at the
3
pharmacy when prescriptions come through without
4
stickers, in the first instance.
5
Then, second, it calls into question for
6 me
all of the information we have had today about
7 how
mail order prescribing doesn't happen, or is
8 not
allowed, or is unauthorized. Suddenly I
am
9
very confused about how much might be going on.
10
DR. TRONTELL: The use of the term
11
voluntary I think is in the strictest sense. I
12
think we have seen from the data presented today
13
that use of the stickers is high, and I think the
14
experience would suggest that individuals are
15
largely trying to do the appropriate thing.
16
Technically, pharmacy law administered at the state
17
level allows a pharmacist to honor any
18
prescription, and if a physician wanted to make a
19
case that they could prescribe this product without
20 a
sticker it might be that a pharmacist would
21
permit it to be so.
22
I think there is the implication in the
368
1
program, because of the risk management program
2
that is spelled out so explicitly in labeling, that
3
individuals would comply; that individuals who
4
didn't comply might, in fact, face certain risk, if
5
only from a liability standpoint.
So, it is
6
largely through influence that this program has its
7
extensive use, at least as we have seen it
8
described today. I am not sure if
that answers
9
your question sufficiently.
10
DR. GARDNER: Yes, I think
so. The grey
11
area for me still falls into the mail order
12
department.
13
DR. TRONTELL: In the area of mail
order,
14 we
were told that there is no known exposure. In
15 the
data that were presented today, some of the
16
data streams include prescriptions that might come
17
from mail order but we did not do an explicit
18
breakout of that. That is
certainly something that
19
could be done though but probably we would be
20
unlikely to give you an answer before the end of
21
this meeting tomorrow.
22
DR. GROSS: Dr. Schmidt?
369
1
DR. SCHMIDT: Helen Keller was
interviewed
2 by
a newspaper reporter one time and was asked what
3
could be worse than losing your sight, and she said
4
losing your vision. Certainly, I
think your talk
5
brings us to a vision. One of my
visions, and I
6
just reflect on this, is that one of the things we
7
could also do is find better retinoids and
8
retinoids that don't have teratogenicity. In
9
Maybock's book on dermatotoxicology there is a
10
chapter where he talks about the class of retinoids
11
called n-phenyl retinamide that appear to bear no
12
teratogenic effect risk although they have a good
13
effectiveness. So, hopefully, in
the future we
14
will be able to also solve this problem by finding
15 retinoids
that do not have this kind of risk.
16
DR. GROSS: Dr. Bergfeld?
17
DR. BERGFELD: I would like to
revisit the
18
pharmacy. We have heard a lot of
things evolve
19
around the pharmacist and in our information sent
20 to
us we also have been made aware that many major
21
pharmacies have dropped out of this program. There
22 was
no explanation but I would assume it was
370
1
because it is so time intensive.
If the pharmacist
2
will still continue, no matter who does it, to be
3 the
police in any program that is evolved, I think
4
that we should hear from the pharmacists and the
5
reality if they are taking on that responsibility.
6
DR. TRONTELL: If I may answer
that with a
7
clarification, the dropout problem that was
8
described in the review from the Office of Drug
9
Safety was dropout from pharmacies participating in
10 the
compliance survey. That wasn't meant to
imply
11
that those pharmacies weren't, in fact, complying
12
with the sticker program. They
declined to do a
13
separate survey of their prescriptions to see, in
14
fact, at what level of compliance they were
15
performing. So, there is a
difference between
16
consenting to be measured and consenting to
17
participate in the program. We
have no evidence to
18
suggest that any particular pharmacy chain has
19
elected not to use the sticker program.
20
DR. GROSS: Anne, I would like to
ask you
21 why
you would have confidence in more education.
22 It
hasn't seemed to work so far and many of the
371
1
quality improvement studies done on education show
2
that it is of limited benefit.
3
DR. TRONTELL: I think,
personally,
4
education is necessary but not sufficient to change
5
behaviors. I suspect in the arena
of contraceptive
6
behavior--but this is largely based on my clinical
7
experience as a pediatrician and dealing with the
8
vast challenge of teenage sexuality--that there is
9 a
lot of misinformation and, in fact, it may be
10
perhaps even more challenging for young adults in
11 an
office visit for a physician to presume that a
12
patient is less than well informed about what are
13
effective methods of contraception.
14
But I do agree. We have seen a
quite
15
thorough educational program put in place with the
16
current isotretinoin risk management program and I
17 am
not entirely clear what would be the best
18
mechanisms to improve that. We
might ask some
19
individuals who are particularly well informed
20
about education and information of individuals.
21
DR. GROSS: Dr. Katz?
22
DR. KATZ: Dr. Trontell, you in
part
372
1
clarified Dr. Gardner's question, but I want to
2
emphasize that the term voluntary would signify to
3
non-dermatologists that, "oh, that is just up to
4 the
doctor to do it." In real life you
don't get a
5
prescription filled without those yellow stickers.
6
Whether the patient, on the questionnaire,
7
remembers it or not, basically, at least around
8
here, if it goes to the pharmacist without a
9
sticker it doesn't get filled.
10
DR. TRONTELL: It is important to
note
11
that to my knowledge of the pharmacy compliance
12
surveys that have been done, those are for the
13
prescriptions that have been filled so there, in
14
fact, is an imperfect way of capturing the
15
prescriptions that have been refused by the
16
pharmacy.
17
DR. BIGBY: This is a simple
question for
18
anybody who can answer it. Why
did CVS and the
19
large chains drop out from the pharmacy compliance
20
survey?
21
DR. KIBBE: Money. In most cases large
22
chains are very tight with how they expend
373
1
resources and, unless they are compensated for
2
doing extra paperwork, they don't want to, and if
3 they are not mandated to do it or not
compensated
4 for
it, they would opt out of it. There was
no
5
contingency that meant that they would then lose
6 the
business of filling prescriptions because they
7
didn't do the survey and I don't think there was a
8
benefit to them.
9
DR. BIGBY: So, they were not paid
to do
10 the
survey?
11
DR. ACKERMANN SCHIFF: Susan
Ackermann
12
Schiff, Hoffmann-La Roche. Yes,
they were
13
compensated per prescription they audited to
14
participate in the survey, although the
15
compensation was minimal.
16
DR. GROSS: Dr. Ringel?
17
DR. RINGEL: I have two comments,
one is
18
normal and one is bizarre. The
normal one is that
19 I
do think education is helpful. I think
it is
20
also helpful for the physician and for the patient.
21 We
are given a list of regulations. That
doesn't
22
mean that we understand why those regulations are
374
1 in
place. For example, it used to be that
2
isotretinoin was taken on the second or third day
3 of
the menstrual period and then all of a sudden it
4
changed so that now the pregnancy test is done
5
during the menstrual period and the drug is started
6
within seven days. Now, why was
that change made?
7 No
one ever explained it to me and I remember
8
sitting down with this and saying I don't get it.
9 I
kind of thought it through and now it makes sense
10 to
me but it wasn't intuitively obvious. I
get the
11
impression that that is something that folks are
12
messing up a lot, the seven-day period and, you
13
know, why should we do the pregnancy test then.
14
Another thing is why two pregnancy tests?
15 I
mean, if the first one is positive the second one
16 is
going to be positive so why bother to do the
17
first one? I mean, somebody could
argue that. I
18
think people should have like a chart so this is
19 the
regulation and this is why. I think it
would
20 be
helpful for both physicians and for patients.
21
The bizarre one is this, I have an idea
22 for
a reminder or a prompting system. If
everyone
375
1
were on a primary form of contraception and did it
2
right and complied all the time we would have a
3
very low pregnancy rate. You can
document that
4
people have had a vasectomy or tubal ligation or
5 are
getting monthly Depo Provera shots or have an
6
IUD, but you can't monitor people's use of oral
7
contraceptions. I think somebody
might need to
8
tell me if the technology for this is there; I
9
think it is. This would have to
be an electronic
10
solution. Let's have an
electronic pill box and it
11 has
28 days, and you put one pill in each day.
12
There is a little microcomputer thing and each time
13 you
open up the little cell and take one the date
14 is
marked on this little computer card, kind of
15
like the thing you have in your digital camera. At
16 the
end of the month you take that to the
17
pharmacist and he puts it in a card reader and if
18 you
have taken your pill every day he will give you
19
your next Accutane. Wait, I am
not done.
20
[Laughter]
21
Now, if you don't take your pill though
22
there is this little buzzer that goes off and this
376
1
little annoying ding, you know, the ding-ding-ding
2 to
drive you crazy and remind you to take it.
3
Then--then, one more thing, if you don't do it two
4 days
in a row there is a red light that goes on and
5
underneath the red light it says "stop
6
isotretinoin; use alternative birth control; call
7
physician immediately." If
you have missed two
8
days in a row and the red light goes on and that is
9 it,
and then you would know if people are taking
10 it.
11
DR. GROSS: And then what do you
do when
12
they put the pill in the wastebasket?
13
DR. RINGEL: Well, you have to
assume that
14
nobody is trying to get pregnant on Accutane;
15
nobody says, "oh gosh, I really want to have a baby
16
with a birth defect." I
don't think people are
17
going to do that, I really don't.
18
DR. GROSS: Sarah Sellers?
19
DR. SELLERS: Dr. Trontell
addressed the
20
comparison of risk management programs including
21
clozapine. I think it is
important to point out
22
that under this program they require registration
377
1 of
pharmacists and not pharmacies. In the
2
presentations I have seen concerning the
3
effectiveness of other risk management programs,
4
specifically thalidomide, they have noted that the
5
system does break down at the level of the pharmacy
6
because if a pharmacy is registered and, for
7
instance, that pharmacy is open 24 hours a day
8
there may be pharmacists during the day that are
9
very well trained in the program but then a shift
10
pharmacist comes in for the midnight shift who is
11 not
well trained. So, this is where they
were
12
seeing breakdown of the program.
Registering
13
pharmacists, on the other hand, would eliminate
14
that risk.
15
DR. GROSS: Dr. Whitmore?
16
DR. WHITMORE: Dr. Gross, you had
brought
17 up
a couple of different times education and also
18 why
people wouldn't use two forms of contraception.
19 Mr.
Sisto from one of the generic companies had
20
shown us data on pregnancy rates and, just to be
21
complete in his reporting of the pregnancy rates,
22 he
actually gave us rates after persons had
378
1
completed Accutane. So, even in
women who were 30
2
days from Accutane when no effect of the drug
3
should be seen at that time what he showed in that
4
data was that one in three of those women who
5
became pregnant 30 days after the drug had
6 abortions.
7
So, that just brings up the obvious, that
8
women have abortions outside Accutane, and I don't
9
know what the rate of abortion is with pregnancy in
10
this country but I wonder if it is any different
11
with Accutane versus with women who get pregnant
12 who
are not on Accutane. It just brings up
the
13
obvious, that for some women abortion is okay and
14 for
others it is not. For those for whom it
is
15
okay, I think that they take the idea of two forms
16 of
contraception more seriously than others.
There
17
were women who did get pregnant when they were on
18
Accutane who did not have abortions so, obviously,
19
there are women who will not have abortions even
20 for
medical reasons.
21
In summary, if we want to control this
22
outside of what women want, the only way for us to
379
1 do
that is to insist that they have some form of
2 adequate contraception such as Depo
Provera. That
3 is
the only way we are going to get 100
4
percent--well, I don't know if it is 100 percent
5 but
that is the only way we are ever going to get
6
close to 100 percent prevention of pregnancy when
7
somebody is on this drug. You
know, this is
8
independent of whether abortion is okay or not
9
okay. If we really want people
not to get pregnant
10
while on this drug, we are going to have to insist
11
that we enforce that in some way.
12
DR. GROSS: Yes, I think we may
have an
13
answer.
14
DR. MITCHELL: Well, it is a
partial
15
answer. If I have the privilege
tomorrow I could
16
actually present the data themselves.
But we track
17 the
rates of pregnancy and the rates of elective
18
abortion not only during Accutane therapy and the
19
month following but in the five months following
20
that. I would be happy to present
it tomorrow. I
21
think it is sort of an internal data set that bears
22 on
the question you asked.
380
1
DR. GROSS: Yes, and my comments
on
2
education were two-fold. One is
that education
3
already takes place in this program and seems to
4
fail. The types of education that
traditionally
5
fail are passive education. Those
that are
6
interactive where the person who is being educated
7
somehow participates in the session, those studies
8
indicate that that type of interactive education is
9
much more successful.
10
MR. HUBER: I am sorry, I think we
may
11
have the answer to the previous question if I
12
understood the question correctly.
13
DR. GROSS: Go ahead.
14
MR. HUBER: I think we have the
data.
15
Could you repeat your question, please?
16
DR. WHITMORE: Well, my point was
that
17
women in this country have abortions and there are
18
probably women who are on Accutane who think I can
19
have an abortion if I get pregnant and a woman,
20
independent of being on Accutane or not, had she
21 got
pregnant would have had an abortion.
And, the
22
only way we are going to control this and say 100
381
1
percent we do not people getting pregnant while on
2
this drug is to insist that there is a mechanism by
3
which they can't get pregnant, and that would be
4
something like Depo Provera but, again, that is not
5 100
percent but at least we are eliminating the
6
behavioral failures of birth control pills and
7
things like that.
8
MR. HUBER: If your question was
the rate
9 of
abortion in the population in general, we do
10
have that data if you would like to see that.
11
DR. WHITMORE: Surely.
12
MR. HUBER: Could I have the slide
on,
13
please?
14
[Slide]
15
The data source is Henshaw Family Planning
16
Perspective, 1998. If you look at
this, there is
17 the
age, age at outcome and what these are is
18
estimated rates of unintended pregnancies,
19
unintended births and abortions per 1,000 women,
20 age
and marital status and percentage of unintended
21
pregnancies ended by abortion. If
you look, we
22
have unintended pregnancy, unintended birth,
382
1
abortion and percent ended by abortion.
The
2
numerator is the women who are in the group. The
3
denominator is the age range. As
you can see from
4
these data, abortion and percentage of unintended
5
pregnancies ended by abortion is quite high.
6
DR. GROSS: Dr. Honein?
7
DR. HONEIN: I had a question
about how
8 the
registration works with the thalidomide program
9 in
respect to whether that is similar or not to
10
what is being proposed for isotretinoin.
Is there
11 a
unique patient ID number? If so, does
the
12
registry also capture the patient's name that is
13
linked to that ID number? Or, how
do you eliminate
14
duplicates if you are using a unique ID number if
15
there isn't a name linked with that somehow?
16
DR. TRONTELL: Some of the details
we may
17
not, in fact, have for you. Bear
in mind, however,
18
that thalidomide is in a single source environment
19 so
that potential duplication is not an issue for
20
that product. The issue of how
the patient is
21
explicitly registered in the system, whether that
22
involves an identifier--
383
1
DR. HONEIN: Even in a single source
2
environment you could have duplication over time if
3
someone had multiple prescriptions but not
4
continuous, or no longer had their patient ID
5
number.
6
DR. UHL: We can get that answer
for you.
7
DR. TRONTELL: Right. Bear in mind too
8
that thalidomide's use is largely for the use of
9
oncology indications. These are
people who are
10
very ill, who typically are not on the product for
11 very
long and are receiving it in a way that--you
12
know, some of that in and out therapy that we know
13
happens with isotretinoin isn't perhaps likely to
14
occur. We will seek that answer
for you.
15
DR. GROSS: Dr. Raimer?
16
DR. RAIMER: Dr. Trontell, I was
just
17
wondering, and you may have partially answered the
18
question, with the S.T.E.P.S. thalidomide program I
19
know the numbers of women on that drug are very low
20 but
what is the incidence of pregnancy with the
21
S.T.E.P.S. program and thalidomide?
22
DR. TRONTELL: Dr. Uhl will
answer.
384
1
DR. UHL: We actually have some
backup
2 slides on this as well.
3
[Slide]
4
This is some information that we have on
5 the
S.T.E.P.S. program for prescribing of
6
thalidomide. Thalidomide was
approved before
7
September of 1998 but it was ready for marketing
8 and
implemented in September of 1998. So,
from
9
September of 1998 until April of 2003 there were
10
approximately 400,000 prescriptions in totality
11
written for thalidomide. There
are approximately
12
80,000 patients that were exposed to thalidomide
13 and
the majority of the use was for patients with
14
oncologic indications.
15
[Slide]
16
Here are the demographics of patients that
17
have taken thalidomide. As I
said, there were
18
80,000 patients. More than half
of those patients
19
were male patients and less than half were females.
20
Less than 5 percent, in the ballpark of about 4,000
21
patients were females of childbearing potential.
22 The
demographics of users of thalidomide are
385
1
considerably different than the demographics of
2
patients that use isotretinoin.
The mean age was
3 66
years and the range was anywhere from less than
4 one
year to over 100 years. For females of
5
childbearing potential the mean age was 42, with a
6
range of 13-59. The mean length
of therapy was
7
four months for thalidomide.
8
[Slide]
9
In that time period there was
one
10
pregnancy that has occurred in the S.T.E.P.S.
11
program and here is some specific information about
12
that case. This woman was a 44
year-old gravida 5,
13
para 3 with history of 2 previous spontaneous
14
abortions or miscarriages that were unrelated to
15
thalidomide therapy. This patient
had high risk
16
malignant melanoma.
17
What is pertinent about this patient's
18
case is that she was intolerant of oral
19 contraception
therapy and used two methods of
20
barrier contraception, condom and spermicidal foam.
21 In
the S.T.E.P.S. program the patient has to have
22 two
negative pregnancy tests before they are given
386
1
thalidomide and the second negative pregnancy test
2
needs to be done within 24 hours of getting that
3
drug. So, this patient had a
negative pregnancy
4
test on the day prior to starting thalidomide. The
5
S.T.E.P.S. program also has weekly pregnancy tests
6 for
the first month and subsequent weekly pregnancy
7
tests for this patient were negative on three
8
consecutive weeks. On week four
she had a positive
9
pregnancy test and then had a positive quantitative
10
test, and then she went on to have a spontaneous
11
miscarriage on day 63 of her menstrual cycle.
12
DR. GROSS: Robyn Shapiro?
13
DR. SHAPIRO: That actually was
half of my
14 question. Risk management is, of course, weighing
15 and
balancing benefits and burdens of what you may
16
decide to do so I guess, Dr. Trontell, I am
17
wondering if you could speak, while it is much less
18
analogous, to clozapine and the etiology of that
19
more restrictive risk management program? What
20
were the numbers of bad things happening there that
21
convinced you and the agency--you and the sponsor,
22
whoever it was--to go to the next level in terms of
387
1
restrictiveness of a risk management program?
2
DR. TRONTELL: The clozapine
program was
3 put
in place at the first marketing of this
4
product. I don't have first-hand
knowledge so I am
5
going to be reporting information that is
6
second-hand. I will invite any
who may wish to
7
correct me, but in the clinical trial experience of
8
this drug there was a significant and concerning
9 rate
of agranulocytosis. I recall it as being
on
10 the
order of three percent of so. Please
don't
11
quote me on that.
12
So, the concern with this product, which
13 did
show benefits for a particularly difficult
14
population to treat, was how that might be
15
prevented. So, from the start, at
the time of
16
approval, the system was put into place.
I am told
17
that advocates for individuals with this illness
18
actually advocated for the collection and
19
registration of patients. My
understanding is that
20 the
experience of agranulocytosis in the program in
21
practice actually is less than the percentage that
22 was
observed in clinical trials. The white
cell
388
1
testing doesn't prevent the occurrence of
2
agranulocytosis. Presumably you
see a reduction in
3 the
white cell count before it becomes critically
4 low
and it may be, in fact, that individuals are
5
operating cautiously when they see a downward trend
6 and
they operate proactively to remove the patient
7
from the drug product. So, that
was instituted
8
from the beginning.
9
Similarly, it is important to note that
10 the
thalidomide program was also instituted at the
11
time of approval. Once a drug
product is on the
12
market it is a little more challenging to redesign
13 the
methods in which individuals have become
14
accustomed to using it.
15
I will make one additional comment.
There
16 is
also probably a difference between going to a
17
pharmacist and showing them a white cell count,
18
which is really nothing that you individually
19
influence, and going to a pharmacist and handing
20
them a pregnancy test result.
21
DR. GROSS: Dr. Crawford?
22
DR. CRAWFORD: This question is
for either
389
1 Dr.
Uhl or Dr. Trontell. In looking at the
goal of
2 no
fetal exposures to isotretinoin for the at-risk
3
population as females of childbearing potential, in
4
some of the presentations this morning there were
5
suggestions that risk management programs also
6
include men. This would be a new
category of risk,
7 not
the at-risk population, knowing that people may
8
share drugs. Though, given what
was said about
9
looking at undue burden, I would like to ask the
10
representatives of the FDA to please comment on
11
that.
12
DR. TRONTELL: With thalidomide
there is
13
actually some potential concern for risks relating
14 to
paternal use of the drug and potential levels of
15 the
drug product in seminal fluid that may or may
16 not
present actually a risk to the female partner
17 of
those individuals. So, in fact,
pregnancy
18
prevention is directed to those individuals. They
19 are
advised to use barrier methods of
20 contraception.
21
With isotretinoin some of the concerns
22
that have extended the program to males gets at the
390
1
issue of having a parallel system for males that is
2
different than for females. In
fact, a pharmacist
3 who
is in a busy practice may have greater
4
opportunity to make an inadvertent error on a
5
female's prescription if, you know, they don't
6
attend to the name or the check box on the sticker,
7 and
so forth.
8
In terms of the issue of sharing drugs, I
9
don't know any evidence to suggest that individuals
10 in
intimate relationships are more likely to share
11
products than others but it clearly is something
12
that happens among female patients who are trying
13 to
improve their skin sometimes for a specific
14
event. I have heard of swap meets
occurring. I
15
think that was described in the MMWR article.
16
DR. GROSS: Dr. Whitmore?
17
DR. WHITMORE: In the Accutane
Roche
18
briefing package they have information on
19
pregnancies. Among 183
pregnancies, 32 percent
20
occurred in the 30 days after the Accutane dosing.
21 Is
there any proposal to address that because that
22 is
when they are lost to the sticker system and
391
1
other things like that? After the
completion of
2
Accutane there is no longer a requirement to
3
qualify in any way to get more drug and that 30-day
4
time period is a period in which we don't want them
5
becoming pregnant.
6
DR. GROSS: Would anyone from
Roche like
7 to
answer?
8
MR. HUBER: There is no specific
element
9
aimed at that post population. We
have been
10
struggling with how we would catch them in a
11
follow-up cycle. The problem is
we can put in the
12
educational components but because they don't have
13 to
do anything because they are not requesting more
14
drug, there is not the same incentive as you can
15
have in the regular follow-up.
So, our current
16
approach would be to probably put some type of
17
follow-up in the intervention but this is something
18
that if the committee has some guidance on, we
19
would be very appreciative to hear your thoughts on
20
that.
21
DR. GROSS: I would like to close
for the
22
day. I would like to thank the speakers
and thank
392
1 the
committee members for their attentiveness and
2
perceptive questions.
3
There are a couple of items of business.
4 The
committee members can leave all their paperwork
5
here. It will be safe
overnight. The committee
6
members do have some homework.
Attached to the
7
agenda, behind the agenda sheet for tomorrow, are
8 the
questions that we will be considering tomorrow.
9
They will be presented in detail by Dr. Seligman
10 and
then we will discuss them and give our opinions
11 and
recommendations.
12
For the committee members, you are all
13
invited to dinner. A bus will be
out front at 6:00
14
p.m. to take us back and forth to the restaurant.
15 So,
thank you all. I look forward to seeing
you
16
tomorrow.
17
[Whereupon, at 4:45 p.m., the proceedings
18
were recessed, to resume at 8:00 a.m., Friday,
19 February 27, 2004.]
20 - - -