1
DEPARTMENT OF HEALTH AND HUMAN SERVICES
FOOD AND DRUG ADMINISTRATION
CENTER FOR DRUG EVALUATION AND RESEARCH
ADVISORY COMMITTEE FOR
REPRODUCTIVE HEALTH DRUGS
Thursday, December 2, 2004
8:00 a.m.
Hilton Hotel
The Ballrooms
Gaithersburg, Maryland
2
P A R T I C I P A N T S
AVISORY COMMITTEE REPRODUCTIVE HEALTH DRUGS
Linda C. Giudice, M.D. - CHAIR
Arthur L Burnett, M.D.
Scott S. Emerson, M.D.
W. David Hager, M.D.
Vivian Lewis, M.D.
Larry Lipshultz, M.D.
Charles L. Lockwood, M.D.
George A. Macones, M.D.
Valerie Montgomery-Rice, M.D.
Joseph B. Stanford, M.D.
Jonathan Tobert, M.B. [Industry Representative]
Lorraine J. Tulman, R.N.. D.N.Sc.
[Consumer Representative]
Teresa A. Watkins, R.Ph., Executive Secretary
CONSULTANTS AND GUESTS
SGE Consultants (voting)
Joanne F. Dorgan, Ph.D., M.P.H.
Diane F. Merritt, M.D.
Steven Nissen, M.D.
Donald L. Patrick, Ph.D.
Martha Solonche [Patient Representative]
SGE Consultants (non-voting)
Adrian S. Dobs, M.D.
Julia r. Heiman, Ph.D.
FDA PARTICIPANTS
Julie Beitz, M.D.
Donna Griebel, M.D.
Scott Monroe, M.D.
Daniel Davis, M.D.
Lisa Soule, M.D.
3
C O N T E N T S
PAGE
Call to Order and Introductions 5
Linda Giudice, M.D., Chair
Conflict of Interest Statement 8
LCDR Teresa Watkins, R.Ph.,
Executive Secretary
Welcome and Comments 10
Donna Griebel, M.D., Deputy Director,
Division of Reproductive and Urologic Drugs
Assessing Risks of Hormonal Interventions 19
Judith Hsia, M.D.,Professor of Medicine,
George Washington University
Sponsor Presentation
Introduction 36
Joan M. Meyer, Ph.D., Senior Director,
New Drug Development, Procter & Gamble
Pharmaceuticals
Phase III Clinical Efficacy Data 43
Johna D. Lucas, M.D., Medical Director,
Procter & Gamble Pharmaceuticals
Clinical Relevance of Treatment Effects 57
Leonard R. DeRogatis, Ph.D.,Director,
Johns Hopkins Center for Sexual Health
and Medicine
Phase III Clinical Safety Data 68
Johna D. Lucas, M.D.
Phase III Hormone Data 86
Glenn D. Braunstein, M.D., Professor
Geffen School of Medicine, UCLA;
Chairman, Department of Medicine,
Cedars-Sinai Medical Center
Hypoactive Sexual Desire Disorder 100
Jan L. Shifren, M.D., Faculty,
Harvard Medical School; Director,
Menopause Program, Vincent OB/GYN
Service, Massachusetts General Hospital
4
C O N T E N T S
PAGE
Sponsor Presentation - Continued
Phase IV Long-term Safety Plan 109
Michael Steinbuch, Ph.D., Director,
Pharmacovigilance and Epidemiology,
Procter & Gamble Pharmaceuticals
Closing Remarks 119
Joan M. Meyer, Ph.D.
FDA Invited Speaker
Safety of Exogenous Testosterone in Women 123
Adrian Dobs, M.D., Professor of Medicine
Johns Hopkins University
FDA Presentation
Efficacy Findings and Issues 147
Daniel Davis, M.D., Medical Officer,
Division of Reproductive and Urologic Drugs
Safety Findings and Issues 160
Lisa Soule, M.D., Medical Officer,
Division of Reproductive and Urologic Drugs
Open Public Hearing 182
Questions from the Committee to Sponsor and FDA 245
Committee Discussion 367
5
P R O C E E D I N G S
Call to Order and Introductions
DR. GUIDICE: Good morning. Would everyone
take their seats, please? Thank you.
I'm Linda Giudice, and this is the
reproductive Health Drugs Advisory Committee Meting
to the FDA. And today the committee will discuss
the new drug application, the testosterone
transdermal system, Procter and Gamble, indicated
for the treatment of hypoactive sexual desire
disorder in surgically menopausal women receiving
concomitant estrogen therapy.
I'd like to first remind the audience and
committee members to please put your cell phones on
silent or vibrate so that the proceedings are not
interrupted. And I'd like to begin this morning by
going around the table so that each of our
committee members may introduce himself or herself.
Please give your name and your affiliation, and
then we will move on with the program.
So I'd like to start on this side, please.
DR. BEITZ: I'm Julie Beitz, I'm the Deputy
6
in Office of Drug Evaluation III.
DR. GRIEBEL: I'm Donna Griebel. I'm the
Deputy in the Division of Reproductive and Urologic
Drug Products.
DR. SOULE: I'm Lisa Soule, Medical Officer
in the Division of Reproductive and Urologic Drug
Products.
DR. MONROE: I'm Scott Monroe, Clinical
Team Leader, Division of Reproductive and Urologic
Drug Products.
DR. DAVIS: Dan Davis, a medical reviewer
in the Reproductive Drug Products Division.
DR. MACONES: George Macones, Department of
OB-GYN, University of Pennsylvania.
DR. HAGER: David Hager, Obstetrics and
Gynecology, University of Kentucky and Central
Baptist Hospital, Lexington, Kentucky.
DR. TULMAN: Lorraine Tulman, School of
Nursing, University of Pennsylvania, Philadelphia.
DR. BURNETT: Bud Burnett, Urologist on
staff at Johns Hopkins in Baltimore.
DR. DICKEY: Nancy Dickey, Family and
7
Community Medicine, Texas A&M University System
Health Science Center.
DR. GIUDICE: I'm Linda Giudice,
reproductive endocrinologist at Stanford
University.
DR. WATKINS: I'm Teresa Watkins. I'm the
Executive Secretary with the Advisors and
Consultants staff.
DR. LOCKWOOD: Charles Lockwood, OB-GYN,
Yale University.
DR. LEWIS: Vivian Lewis, reproductive
endocrinology, University of Rochester.
DR. LIPSHULTZ: Baylor College of Medicine,
urology.
DR. SOLONCHE: Martha Solonche, patient
representative, New York City.
DR. PATRICK: Donald Patrick, Social and
Behavioral Sciences and Health Outcomes, University
of Washington.
DR. NISSEN: I'm Steve Nissen, and I'm a
cardiologist with the Cleveland Clinic.
DR. MONTGOMERY-RICE: Valerie
8
Montgomery-Rice, reproductive endocrinologist,
Meharry Medical College, Nashville, Tennessee.
DR. HEIMAN: Julia Heiman, Kinsey
Institute, Indiana University.
DR. TOBERT: Jonathan Tobert, Tobert
Medical Consulting, and the University of Oxford,
England.
DR. GIUDICE: Thank you.
Ms. Teresa Watkins will now read the
conflict of interest statement.
Conflict of Interest Statement
MS. WATKINS: Thank you.
The following announcement addresses the
issue of conflict of interest, and is made as part
of the record to preclude even the appearance of
such.
Based on the submitted agenda for the
meeting, and all financial interests reported by
the committee participants, it has been determined
that all interest in firms regulated by the Center
for Drug Evaluation and Research present no
potential for an appearance of a conflict of
9
interest at this meeting, with the following
exceptions.
Because Drs. Adrian Dobs and Julia Heiman
has past involvements with Proctor & Gamble related
to the product under discussion, the Agency has
decided to limit their participation.
Dr. Julia Heiman may participate in the
committee's deliberations. She is, however,
excluded from voting.
Dr. Adrian Dobs is permitted to give a
presentation to the committee and to answer
questions directly related to her presentation.
We would like to note that Dr. Jonathan
Tobert has been invited to participate as a
non-voting industry representative, acting on
behalf of regulated industry. Dr. Tobert's role on
this committee is to represent industry interests
in general, and not any one particular company.
Dr. Tobert owns Tobert Medical Consulting LLC.
In the event that the discussions involve
any other products or firms not already on the
10
agenda, for which an FDA participant has a
financial interest, the participants are aware of
the need to exclude themselves from such
involvement, and their exclusion will be noted for
the record.
With respect to all other participants, we
ask, in the interest of fairness, that they address
any current or previous financial involvement with
any firms they may wish to comment upon.
Thank you.
DR. GIUDICE: Thank you.
Donna Griebel will now give her welcoming
statement and comments.
Welcome and Comments
DR. GRIEBEL: [Off mike.] Good morning
everybody. Can you hear me?
Now can you hear me? I'll lean.
I'm Donna Griebel. I'm the Deputy
Director--as you've already heard--of the Division
of Reproductive and Urologic Drug Products. The
Division would like to welcome you all here
today--both committee members and guests--and we
11
would like to express our gratitude to you all for
traveling all the way to Washington at a very busy
time of the year to consider the application that
is before us.
As you have already heard, the NDA is for
testosterone transdermal system. The proposed
indication is for the treatment of a subtype of
female sexual dysfunction called hypoactive sexual
desire disorder. And it is specifically for a
sub-group of women with this disorder, which is
surgically menopausal women who are being treated
with estrogen therapy.
This application, if approved will be the
first product that will have been approved for
female sexual dysfunction.
I'm getting tired of leaning.
And the Division did designate this review
a priority review, which means it was given a
six-month review clock. We did so because there
are no products approved for female sexual
dysfunction, and a product that successfully treats
this disorder could have a major impact on a
12
woman's quality of life.
In your review of the FDA's background
document--and hen you hear the FDA's presentation
today you will conclude that the FDA concurs with
the applicant, Procter & Gamble, that they have
shown a statistically significant difference
associated with treatment with TTS in the primary
efficacy endpoint of interest, which is the number
of satisfying sexual events.
However, we do have questions about
whether the statistically significant change that
was produced by TTS, and the proportion of women
who experienced this improvement relative to
placebo is a clinically meaningful difference.
Analysis of efficacy is but one piece of a
risk-benefit analysis. You also have to look at
the safety data, and you have to evaluate the
safety data base for its adequacy.
The Division factored its experience with
the impact of the outcomes from the Women's Health
Initiative studies in their approach to evaluating
the safety data in this application. The Women's
13
Health Initiative studies were very large,
randomized controlled trials, prospective studies,
that were powered to show important differences in
safety outcomes. They led to changes in product
labeling for products that are intended for
treatment of menopausal symptoms in
post-menopausal, and they reinforced an initiative
to be sure that the lowest effective dose of these
products has been defined.
The results of the WHI studies became
available after launch of the studies that you will
be seeing presented today.
We've invited additional experts to sit at
the table today and participate in the committee's
deliberations. They include Dr. Donald Patrick,
who has expertise in evaluating health-related
quality of life instruments and endpoints; Dr.
Julia Heiman, who's director of the Institute; Dr.
Steven Nissen is a cardiologist; and Dr. Diane
Merritt was to join us today. She's an OB-GYN from
Washington University.
The Division has invited two guest
14
speakers to participate. They will be making
presentations on issues relevant to assessing risk
associated with hormonal products in
post-menopausal women.
Dr. Adrian Dobs is a professor at Johns
Hopkins University in endocrinology and metabolism.
She will join us later this morning to summarize
the existing clinical evidence in the literature of
potential safety issues associated with
testosterone use in women.
Our second guest speaker, Dr. Judith Hsia,
is a cardiologist at George Washington University.
She was a lead investigator in the WHI studies.
You will hear the FDA mention WHI over and over
again this morning and clearly we do believe that
the data from these studies are relevant to today's
discussion--for a number of reasons that include:
number one, if TTS is approved, it will be approved
for the population that was studied in this
application, and that is surgically post-menopausal
women who are taking estrogen. This implies that
women who take TTS will also be taking estrogen, as
15
well. And, presumably, they will be taking the
products chronically.
WHI demonstrates that millions of women
can be exposed to a drug for a well-accepted
off-label use, such as estrogen for prevention of
cardio-vascular disease, and only when studied in a
large, randomized controlled trial do we find that
the drug is not effective for the presumed benefit,
and is actually associated with substantial risk.
WHI also drives home some very important
practical issues. If you want to shoot for the
ideal study design to evaluate such risks--and that
is it takes enrollment of a lot of women to get
those answers. There are also hypotheses for the
safety outcomes from the WHI studies that may be
applicable to the data that you're reviewing today.
It's important for you to remember that
the guest speakers are only available to you for
dialogue and interaction between the committee
members and the guest speakers when they're at the
podium. So we would like to remind you to be sure
and ask your questions that you would like to ask
16
of them while they are at the podium.
Quickly, reviewing the agenda: Dr. Hsia
will actually be our first speaker, and we'll open
the meeting today. She will be followed by Procter
& Gamble's presentation of the data. There will be
a break, followed then by Dr. Adrian Dobs.
FDA will follow Dr. Dobs, and we'll close
the morning with the open public hearing speakers.
After lunch we will return and the
committee will have an opportunity to ask
clarifying questions of the applicant, Procter &
Gamble, and the FDA. And then you'll transition
into discussion of the FDA's questions, and we'll
vote on those questions.
When you're listening to the presentations
this morning, please keep the questions in mind.
And I'll briefly run through them.
[Slide.]
The first is an efficacy question: "Do the
efficacy data represent clinically meaningful
benefit above that of placebo for surgically
menopausal women with hypoactive sexual desire
17
disorder who are taking concomitant estrogen?"
Uh-oh. I am now missing two of my
questions.
[Comments off mike.]
I wasn't a good Boy Scout and didn't come
prepared to the podium.
The second question is with regard to the
safety database: whether there's been an adequate
number of women exposed, and whether the duration
of exposure is adequate; whether the duration of
exposure with the placebo control is adequate--with
the bottom-line question: "Is the exposure, the
total number of women treated, and duration of
treatment adequate to demonstrate long-term
safety?"
The third question is a three-part
question. The first is: "a) Are the safety
concerns or unanswered questions associated with
use of TTS in combination with estrogen that need
to be studied; for example, questions about
cardiovascular or breast cancer outcomes, or
questions about risks and benefits in populations
18
who are likely to use this product off-label?"
If you answer yes to that question, we
would like you to be sure to state what your
concerns are, what questions you have.
Part c) of the question is: "Should these
concerns or questions be studied prior to approval
of the product?" If you believe that these
questions need to be answered prior to approval, we
would like to know what studies you would
recommend, what the design would be, what are the
endpoints, who's the population.
If you think that the product can be
approved and that your questions could be answered
appropriately after approval, we would like to know
what the study designs would be and, specifically,
we would like you to comment on the applicant's
proposed claims-based cohort study.
And now the question that I do
have--question number four--it's a short one but an
important one: "Are the efficacy and safety data
adequate to support approval of TTS?"
Next, I would like to introduce Dr. Judith
19
Hsia from George Washington University, who will
speak today about the WHI, and assessing risks of
hormonal therapies in post-menopausal women.
Assessing Risks of Hormonal Interventions
DR. HSIA: Well, it's a pleasure to be
here, and I appreciate being able to speak first,
because the Women's Health Initiative steering
committee is meeting downtown today and, as you
know, if one is not present, there's a risk of
having undesirable tasks assigned.
[Laughter.]
So I want to try to avoid that.
So I'm here to talk about approaches to
assessing risks and benefits, and what we've
learned as a consequence of the WHI.
[Slide.]
I'm really going to sort of briefly cover
four areas: biomarkers; observational studies;
randomized trials; and looking at intermediate
outcomes.
[Slide.]
So, if you think about estrogen, there are
20
many biomarkers. One which was studied early and
often was lipids. So if you look at LDL--the bad
cholesterol--in the lower left here, you can see
that starting two years before menopause there's a
steady increase in LDL and currently, the HDL--the
good cholesterol--is falling. And it was known
that estrogen would improve this profile.
[Slide.]
Here's data from the Women's Health
Initiative estrogen-progestin trial. What is shown
is the difference between the active E+P and the
placebo groups with regard to change in the
biomarker from baseline to year one.
So if you look at the LDL cholesterol, you
can see that falls 12 percent--which is a good
thing. The HDL increase 7 percent, which also a
good thing. Triglycerides go up, which is not good
thing. Glucose and insulin both fall, which is a
good thing. And blood pressure goes up, which is
bad. So that you end up with a mixture of
desirable and undesirable characteristics, and it's
hard to know what the overall balance is going to
21
be with regard to risk and benefit.
So a great many observational studies have
been undertaken to try to assess the relative risks
and benefits with hormone therapy.
[Slide.]
This is a summary from an analysis that
was published relatively recently by Joy Manson and
co-authors in the New England Journal. It sort of
summarizes the observational literature so that for
breast cancer there was thought to be an increase
in risk with E+P, which was associated with
duration of therapy. There's a 25 percent
reduction in hip fracture. There was a 45 increase
in stroke; a doubling in the rate of pulmonary
embolism; and a 39 percent reduction in the risk of
MI and coronary death.
And it was on that basis of these type of
studies that women had been being prescribed
post-menopausal hormone therapy for coronary
prevention for many years.
[Slide.]
So at the time the WHI hormone program was
22
designed, that was the data was available, and it
was on that basis that the design was developed.
So it was anticipated that the benefits would
include coronary prevention. And although stroke
looked like it was increased, in those
observational studies, people didn't really quite
believe that. They thought if it prevented
coronary disease it would prevent stroke as well.
So this was sort of a question mark.
It was thought it would increase breast
cancer risk, and venous thromboembolic risk. It
would reduce hip fracture and possibly overall
mortality. So that was the plan.
[Slide.]
As the trials were conducted there were
two arms. It was stratified presence of a uterus
so that women who had no uterus were randomized
either to conjugated estrogens alone or placebo;
and those who had a uterus were randomized to CEE
with daily medroxy-progesterone acetate or placebo.
These were the sample sizes that were
built into the design. And let me just point out
23
that in order to achieve randomization of 27,000
women, that 373,000 women were screened--which is,
you know, a significant proportion of the
age-eligible population of the country.
[Slide.]
Now, if you just look at the E+P
trial--for which the data is fully complete and has
been published--in reality there was a 24 percent
increase in the risk of coronary heart disease.
The stroke data--there was a 31 percent increased
risk in stroke; venous thromboembolism was doubled.
There was an increase in breast cancer. There was
a somewhat unexpected reduction in colorectal
cancer, which was statistically significant. Hip
fracture was reduced.
And there was ancillary study called
"WHIMS"--the Women's Health Initiative Memory
Study--in which women who were 65 or older at study
entry were invited to have annual mini mental
status exams, and if those scores fell below a
certain level, then they went on and had subsequent
testing and clinical evaluation for cognitive
24
impairment. And almost everybody who was offered
participation in the ancillary study did
participate. And WHIMS demonstrated a doubling in
the rate of dementia if one was assigned to active
hormones.
[Slide.]
now, there's been a lot of discussion in
looking at the WHI data about the relationship
between relative risk and absolute risk. So these
are the absolute risks with E+P. It's per 10,000
woman-years. And I think of that as 2,000 women
treated for five years, since that was the duration
of follow-up at the time the trial was stopped for
safety reasons.
So if you treat 2,000 women for five
years, there will be 30 heart attacks or coronary
deaths in the placebo group, and 37 in the active
treatment group; so an excess of seven heart
attacks or coronary deaths associated with E+P
therapy.
Now, the trial, as a sort of study policy,
does not characterize any of these differences as
25
large of small. You know, it's in the eye of the
beholder, and we just put the data out there so
that health care providers and women can make an
informed decision--for the first time.
If you look at stroke, there were 21
strokes in the placebo group for 2,000 women for
five years, and 29 in the active treatment group.
And you can see similar numbers across the board.
For breast cancer, there were 30 breast
cancers in the placebo group, and 38 in the active
treatment group. And you see, actually, the
dementia numbers are in some ways are the most
dramatic.
[Slide.]
So, after the results became available,
the balance had shifted from what was anticipated,
so that on the benefit side, there was a reduction
in fracture--of both hip and total fracture--and a
reduction in colorectal cancer. And the risks
included increase in dementia, coronary heart
disease, stroke, venous thromboembolism and breast
cancer. So the trial was stopped 3.3 years early,
26
for these reasons.
[Slide.]
Now, if you go back to the Manson, et al.,
analysis in the New England Journal--this is just a
reproduction of those numbers. And if you put the
WHI E+P numbers up alongside, you can see that
actually the observational studies were quite
predictive, as far as breast cancer was concerned.
The hip fracture numbers are relatively similar.
The stroke numbers are relatively similar. The
venous thromboembolism numbers are really
identical. And the only thing that's off is
coronary heart disease.
So the question is: why is it? And one of
the things they tried to address in their analysis
was what some possible reasons might be.
[Slide.]
Some of it might be due to the "healthy
user" effect, where women who choose to take
estrogen in those observational studies--you know,
the way they're conducted is like the Nurses Health
Study, they fill out questionnaires, "Are you
27
taking estrogen?" And then they follow them along
and count up how many heart attacks people have.
And women who choose to take estrogen are known to
have other healthy practices, like they exercise
more, they're leaner, they ate more servings of
spinach and so on--and that these things, although
they're adjusted for in the models, you don't know
everything that you need to adjust for, and you
don't have the data on everything that might be
relevant. So that's one problem.
Another is compliance bias. The women who
are taking their hormones might also be adherent to
other healthful behaviors like taking their blood
pressure medicine, or whatever.
It was thought--this is a relatively new
idea--that there might be outcomes identification
bias, where their health care providers thought
that if they were taking estrogen that that symptom
they had couldn't possibly have been a heart
attack, and therefore they were under diagnosed for
the outcomes that they may have had.
And there may have been incomplete capture
28
of early events, where women who had adverse events
while taking estrogen stopped their hormones before
they joined the study.
[Slide.]
So, now what is the potential impact of
adding progestin to estrogen, which has relevance
to the topic today?
[Slide.]
If you look at the outcomes from the
Women's Health Initiative estrogen-alone trial,
compared to the E+P trial, you can see that they
are similar in some ways and different in others.
So that if you look at coronary heart disease,
rather than a 24 percent increase in risk, actually
it's neutral: estrogen-alone is neutral with regard
to heart disease risk.
Stroke, increase is relatively similar.
There is also an increase in venous
thromboembolism, although it's not as marked.
Breast cancer actually is neutral with estrogen
alone, whereas it's increased with E+P.
The colorectal cancer benefit appears to
29
disappear. Hip fracture is prevented by
estrogen-alone, and there is an increase in the
rate of dementia.
So, it does appear that there are some
impacts of adding progestin to the regimen.
[Slide.]
If you compare the absolute risks of
estrogen-alone in the upper panel, with E+P in the
lower panel you can see, for one thing, that the
characteristics of women who've had a hysterectomy
differ from those of women who still have a uterus,
so that the placebo event-rate is higher. These
scales are the same.
So the placebo event-rate is higher for
women with estrogen-alone. They have more
hypertension, diabetes, greater body mass index,
and so on, and this is an issue. However, we have
done analyses to try to assess the impact of this
on the difference between the trial outcomes, and
it does not appear that this accounts for them.
But you can see, for instance, the
absolute increase in the number of strokes here was
30
8, and the absolute number--an increase in
strokes--here is more like 11 or 12. So there are
some differences in the absolute event-rates.
[Slide.]
So if you look at the balance of risks and
benefits with estrogen-alone--which was a 6.8-year
follow-up at the time the trial was stopped early
because of the increased risk of stroke and
dementia--the benefit is fracture reduction; it was
neutral with regard to breast cancer and coronary
heart disease, and there was an increase in
dementia, stroke and venous thromboembolism.
[Slide.]
So the impact of adding androgen to an
estrogen regimen may be difficult to predict, and
may differ among the various types of progestins or
androgens that may be studied.
Now, the alternative to carrying out a big
randomized trial that may require screening 300,000
subjects to get your enrolled population, is to do
studies of intermediate outcomes.
[Slide.]
31
And there are number of different measures
that can potentially be used. This is coronary
calcium, which is assessed using a fast CT scan.
It's about 500 bucks. So it's a cheaper study to
do. It's non-invasive, and may potentially have
utility.
Carotid ultrasound is another approach
that one can take to evaluating drug effects or
intervention effects without waiting for hard
clinical outcomes.
And a third possibility is coronary
angiography, or--since Dr. Nissen is in the
audience--coronary ultrasound.
And there has been some investigation
using these intermediate outcomes, which is really
summarized here.
[Slide.]
There have been a limited number of
estrogen trials with intermediate outcomes. There
have been three angiographic trials, funded by the
NIH; two used conjugated estrogen, and one used
estradiol. And they all showed no benefit or harm
32
with post-menopausal hormone therapy. So, on the
one hand, they didn't show protection the way the
observational studies did, but on the other hand
they did not demonstrate the harm with the
combination E+P that was observed in the randomized
trial with clinical outcomes.
For carotid ultrasound there's been one
randomized trial using estradiol that did
demonstrate benefit, and that may be considered to
be consistent, in some sense, with the
estrogen-alone coronary outcomes, but it's not
consistent with the stroke outcomes, because there
was an increase in stroke in the estrogen-alone
trial which--you know, if you think carotid
ultrasound should be more predictive of
stroke--would not be consistent.
And there actually is no estrogen trial
data with coronary calcification.
The advantages of these trials are that,
you know, they're much cheaper to do. You can do
them with a sample size of 400, rather than 16,000.
And they can be done in a shorter period of
33
time--maybe three years rather than six. So the
advantages are considerable--if you think that the
data are going to be predictive.
[Slide.]
So, just to summarize: the approaches that
can be taken to evaluating risk are, I think,
basically four. There are biomarkers--which I
think have very limited utility, because they're
likely to give a mixed picture, and you're not sure
how to interpret those results. There are
observational studies which, I think, have been
demonstrated to be not reliable for assessment of
coronary risk, although possibly they may be
reliable for assessment of other types of risk.
And the other thing is that, of course, suitable
cohorts may not be always available.
And then there are randomized trials, with
intermediate outcomes, which may potentially be
useful, but their predictive value is still
somewhat suspect.
And there are randomized trials which, of
course, are the gold standard, but are expensive
34
and take a long time to undertake.
So, I'll close with that and--am I
supposed to take questions?
DR. GIUDICE: We have time for just a few
questions.
Yes, please?
DR. NISSEN: Judy, thank you for a very
lucid presentation. I had a couple of questions.
One is: what is known about the use of
estrogen, or estrogen-progestin, in women who have
existing coronary heart disease. That's question
one.
And question two is: using something like
Framingham Risk Score, is there an interaction
between the baseline risk and the risk of
adverse--or increases in event rates in women who
receive hormonal therapy? Can you predict any of
this, using something like Framingham?
DR. HSIA: Well, let me take your first
question first.
The HERS trial randomized women--which I
was principal investigator for--randomized women
35
with documented coronary disease to conjugated
estrogen with medroxy-progesterone daily, or
placebo, and demonstrated an increase in coronary
events in the first year--really, the first six
months--following randomization. And thereafter,
the curves came together and there was no risk or
benefit over longer-term therapy.
And the angiographic trials of course--the
three of them--all included exclusively women with
some coronary disease, and were neutral. But they
were smaller, and there may have been a power
issue.
With regard to whether risk
characteristics at baseline can predict safety, we
have done sub-group analyses looking not at the
Framingham score because, actually, although we
have blood on everybody, we don't have laboratory
measures on everybody for cost reasons. But if you
just count up risk factors, and look at women
with--you know, with or without more or less risk
factors at baseline, it does not help you predict.
There's no interaction between that and outcome.
36
And if you look at the women in the
randomized estrogen-alone and E+P trials who had
prevalent coronary disease at baseline--which was
only a few percent--they had the same relative risk
as women who did not.
DR. GIUDICE: Any other questions from the
committee?
[No response.]
DR. GIUDICE: Okay, thank you very much.
DR. HSIA: Thank you.
DR. GIUDICE: We'll now proceed with the
sponsor presentation. And the first speaker is Dr.
Joan Meyer, who is Senior Director of New Drug
Development at Procter & Gamble.
Sponsor Presentation
Introduction
DR. MEYER: Good morning. In addition to
being Senior Director of Drug Development at
Procter & Gamble, I'm also the Global Project
Leader for the Testosterone Transdermal System
Project.
Today we'll present data to you on the
37
testosterone transdermal system. We've been
granted the trade name Intrinsa, which I will use
from now on, because it's easier to say than
testosterone transdermal system.
So what is it?
[Slide.]
Intrinsa is a patch. It delivers 300 mcg
a day of naturally occurring testosterone, and the
patch is changed twice weekly.
Our proposed indication, as you've heard,
is for the treatment of hypoactive sexual desire
disorder in surgically menopausal women on
concomitant estrogen therapy.
So what is hypoactive sexual desire
disorder?
[Slide.]
Well, this is a recognized medical
condition that affects many women. It has both ICD
and DSM codes.
There are three key elements to HSDD that
are very important to keep in mind, especially when
considering the clinical relevance of this
38
condition. One is: the woman has a decrease in
sexual desire. This leads to a decrease in sexual
activity, and satisfying sexual activity. This, in
turn, causes the woman personal distress. So these
three things are very important: the desire, the
decrease in satisfying sexual activity, and the
increase in distress.
It affects all aspects of a woman's life,
as we can show you today; her health, well-being
and her relationship with her partner.
So what else will we be showing you today?
[Slide.]
After I give you a brief background of the
project, Dr. Johna Lucas, the Director of Clinical
Development at Procter & Gamble, will share with us
the clinical efficacy data, and she's show you the
highly significant improvements achieved by
Intrinsa in all the endpoints that were measured,
including desire, activity and distress.
Then, Dr. Leonard DeRogatis, the Director
of the Johns Hopkins Institute for Sexual Health
and Medicine, will show us the clinical relevance
39
in these data to the women who participated in the
trials.
Then Dr. Lucas will return to share with
us the safety data, and show us the favorable
safety profile we saw in the clinical studies with
Intrinsa.
Then we'll hear from Dr. Glenn Braunstein,
the Chairman of Internal Medicine at the Cedars
Sinai Hospital at UCLA Medical School. He will
share with us the data from our clinical trials on
the levels of testosterone, other hormones that we
measured, and discuss their implications for
safety.
Dr. Jan Shifren is an Assistant Professor
of Reproductive Endocrinology at Harvard Medical
School. She will put these data in context for us,
and she'll show us the impact HSDD has on women's
lives, and discuss the medical need for a drug like
Intrinsa.
Then Dr. Michael Steinbuch, who is the
Director of pharmacovigilance and Epidemiology and
Procter & Gamble, will go over with you our
40
comprehensive Phase IV program that we've designed
to follow the long-term safety of Intrinsa.
Finally, I'll return to wrap-up and share
concluding remarks with you.
But, also available for discussion today,
we have a variety of experts, both within and
outside Procter & Gamble, to discuss any questions
or data that you'd like additional information on.
[Slide.]
I just want to go over briefly the
clinical development program for Intrinsa. And
it's important to remember that this is, indeed,
the first drug for the treatment of hypoactive
sexual desire disorder. We worked closely with the
agency, through our development program. We
followed the FDA guidance on the female sexual
function drug development, and we also did a
variety of studies to design our Phase II and Phase
III programs.
[Slide.]
As you can see from this list, we did
dermal safety studies, PK studies. From these
41
studies we determined the optimal site for patch
application. We did dose-ranging studies. And we
also looked at the effect of route of estrogen
administration on the data.
Our Phase III program consisted of four
large, double-blind randomized placebo-controlled
studies: two of these in surgically menopausal
women. And these are the data we'll be sharing
with you today. Two of these are in naturally
menopausal women, one of which is just completing.
We will present safety data today from
these studies to round-out the safety picture of
what we know about Intrinsa.
But we also did some additional studies to
further understand the benefits of this treatment.
Again, because it's a new therapeutic area, we had
to do some non-traditional studies. And we had
significant input and interaction with the agency
to develop these.
[Slide.]
The currently available instruments that
are out there were not considered fully appropriate
42
really to measure the three key aspects of
HSDD--desire, activity and distress. So we
developed three now instruments to measure these
aspects, and we validated these in four separate
clinical studies.
We also included, in our Phase III
surgical menopause program, a couple of additional
studies. One is a blinded withdrawal, that showed
that the patch is indeed having the desired
pharmacological effect. This was covered in your
briefing book, and we won't be discussing it in the
formal presentations today, but we'll be happy to
answer any questions that you might have on that.
Also, as part of the Phase III surgical
menopause program, we conducted to establish what
level of change in these instruments was meaningful
to the women in the study. And Dr. DeRogatis will
discuss this with us this morning.
We've also done a lot of safety work.
We've evaluated the safety of Intrinsa in several
different ways. We've collected adrenergic adverse
events in a very systematic way during the studies,
43
and we've also collected spontaneous reports.
We've continued to extend the Phase III
studies, and we are now into year three of three of
these surgically menopausal studies.
We've held two scientific advisory groups,
one to discuss what do we currently know about the
safety of long-term use of testosterone, and the
other was on breast safety and testosterone. An
outcome of the latter study was another special
study that we did in conjunction with the
Karolinska Institute. And Dr. Braunstein will
discuss these results.
Also, as you'll hear from Dr. Steinbuch
today, we've proposed a comprehensive Phase IV
safety program.
But now I'll stop telling you what we're
going to tell you, and get started on the data.
I'd like to introduce Dr. Johna Lucas,
Medical Director at Procter & Gamble.
Phase III Clinical Efficacy Data
DR. LUCAS: Good morning. I will now
present the efficacy from our Phase III surgical
44
menopause program, using the 300 mcg per day
testosterone patch.
[Slide.]
I will show you how the drug increased not
just our primary, and important secondary, sexual
function endpoints, but every sexual function
endpoint that we measured--and most of them in a
highly significant manner.
[Slide.]
I want to begin by talking about the
instruments we developed, and explain how the
therapeutic goals that were important to patients
became our primary efficacy endpoints. I will
describe our Phase II and Phase III surgical
menopause program, and then I"ll present the Phase
III results.
[Slide.]
Because no tools existed to measure sexual
desire and distress associated with low sexual
desire in women with HSDD, we developed three
multi-national validated instruments for the
assessment of sexual function. We consulted with
45
more than 100 physicians and sex therapists to get
their perspective, in addition to our validation
program with patients.
Our instruments are: the Sexual Activity
Log. It is a weekly diary that quantitates the
numbers of sexual events; the numbers of orgasms,
and the numbers of satisfying sexual events, for
both intercourse and non-intercourse activity.
The PFSF--or Profile of Female Sexual
Functioning--is a 30-day recall inventory that
evaluates seven domains of sexual function.
The PDS--or Personal Distress Scale--is
also a 30-day recall that evaluates the distress
associated with low sexual desire.
The development of these three instruments
is the subject of two publications.
[Slide.]
First, to discuss the profile of female
sexual function and the PDS. We began by speaking
to more than 250 women who reported having normal
sexual functioning prior to the removal of their
ovaries, and had experienced a substantial of
46
sexual desire, which they found quite distressing,
after their surgery. We used their own specific
words to generate items to be used on our
measurement tools.
These interviews generated 450 items. We
categorized the items by content; removed the items
that were, for example, redundant, slang, had
compound concepts, or were not representative of
the general reading level. We kept 83 items that
retained both content and meaning in six languages,
with both forward and backward translation.
We then tested the ability of these items
to discriminate women with HSDD from their
age-matched control women who described themselves
as well-functioning sexually, with normal sexual
desire. We tested these in three trials.
[Slide.]
This took us down to 37 items for the
PFSF, and 7 for the PDS. These items were then
tested in a fourth validation trial, comparing,
again, surgically menopausal with HSDD to
age-matched control women. Both the PDS and all
47
domains of the PFSF discriminated between
surgically menopausal with HSDD< and their
age-matched controls.
[Slide.]
In parallel to the development of the PFSF
and the PDS, we developed the Sexual Activity
Log--or SAL--which is a weekly diary. We developed
the instrument with further interviews, revision
and retesting to confirm that it captured all
activity; that it avoided double-counting; and that
it could be easily and universally understood by
all.
The SAL also discriminated well between
surgically menopausal women with HSDD and their
age-matched controls
[Slide.]
Our surgical menopause program for
Intrinsa consisted of two Phase II trials to
determine dose, and two Phase III therapeutic
trials. All were of similar design, with an
eight-week baseline period, and a 24 week treatment
period to measure efficacy and safety.
48
In all trials, patients were on
concomitant estrogen, and required to meet similar
inclusion and exclusion criteria.
In the Phase II trials, 300 mcg per day
was found to be efficacious as well as safe, and
established to be the appropriate dose to take into
Phase III.
The Phase III trials included a safety
assessment of up to 12 months.
[Slide.]
Our Phase III program included two
concurrent multinational trials of more than 500
patients each, which we refer to as SM 1, and SM 2.
These two trials had similar numbers of sites,
geography of sites, inclusion and exclusion
criteria, visits and procedures.
[Slide.]
In choosing endpoints for the Phase III
trials, desire, distress and satisfying sexual
activity were all important endpoints to understand
the treatment effect of women with HSDD. Benefits
in each of these areas are essential to achieve a
49
relevant treatment response. In keeping with the
FDA's draft guidance for female sexual dysfunction,
we chose "satisfying sexual activity" as our
primary endpoint. We pre-specified "desire" and
"distress" as key secondary endpoints--not only
because patients reported this as their primary
therapeutic goal, but because they comprise the
definition of hypoactive sexual desire disorder.
After showing you these key three
endpoints, I will show you additional endpoints
from the PFSF that are also relevant for a
meaningful treatment effect, since they too reflect
the other important losses that patients with HSDD
identified.
[Slide.]
Inclusion criteria were designed to ensure
that patients had HSDD associated with their
surgery; that is, patients had considered
themselves to have normal sexual functioning prior
to their surgical menopause, and had experienced as
substantial decrease in sexual desire, and
50
accompanying distress, after their surgery. It was
also important that patients be on estrogen, and in
a stable monogamous relations with the partner
there most of the time.
[Slide.]
Exclusion criteria were designed to rule
out other causes of low sexual desire, either
physiologic or psychologic. Exclusion criteria
were also designed to exclude patients with
conditions or medications that could confound
efficacy, or for whom testosterone might be
inappropriate.
[Slide.]
This schematic shows the number of
patients that were screened--about 1,700; who were
randomized--about 1,100; and who completed--about
870. This represents about 80 percent of patients
completing the double-blind portion of the trial.
About 96 percent chose to enter the open-label
period of the trial, and of those, 76 percent
completed the 12 months.
[Slide.]
51
Data will be shown in this format, with
the first two columns representing SM1, the second
two columns SM2, placebo testosterone, placebo and
testosterone.
The completion rate of 80 percent was
similar in both arms, in both trials. Of the 20
percent who discontinued, reasons for dropping out
were very similar between studies and treatment
groups. Of importance, drops for adverse events
were similar at about 8 percent, regardless of the
study, for testosterone or placebo arm.
[Slide.]
Patient characteristics were also very
similar between studies and arms. Mean age was 49,
with about 80 percent of patients being between age
40 and 59. The average length of relationship was
about 20 years, and patients on average had been
surgically menopausal about half of that.
About three-quarters of patients were on
oral estrogen, and about a quarter of patients were
on transdermal estrogen.
[Slide.]
52
Patients were also balanced at baseline
with regard to disease characteristics. This
baseline sexual desire score of about 20
corresponds to patients' reporting "seldom"
interested in sex. The personal distress score of
about 60 translates to patients "often" being
distressed about their lack of interest in sex.
Baseline satisfying sexual activity was about 3
episodes per four weeks.
Now, to review the results of the Phase
III trial.
[Slide.]
This graph shows the man change in total
satisfying sexual activity experienced by patients
with treatment at 24 weeks. Yellow represents
testosterone treatment; blue, placebo treatment.
Improvement seen with testosterone compared to
placebo were highly statistically significant in
both studies.
At the bottom of the slide--to help you
understand how the patients experienced their
change in satisfying sexual activity, I have
53
provided you with a percent change from baseline in
the respective arms. On average, in the two
trials, we see that the change experienced with
Intrinsa treatment was about double that seen with
placebo.
[Slide.]
Here, you see the changes in sexual desire
at 24 weeks for both Phase III studies. Again,
there was a highly significant increase in sexual
desire with the use of the testosterone patch
compared to placebo. Averaging over the two
studies, the increases seen in placebo were about
24 percent. AS you know, this magnitude of placebo
effect is commonly seen in behavioral studies. On
the other hand, patients who were treated with
Intrinsa experienced, on average, a little more
than 50 percent increase in sexual desire--again,
roughly double that seen with placebo.
[Slide.]
This graph shows you the changes seen in
distress. Distress associated with low sexual
desire is the hallmark HSDD. Distress
54
significantly decreased with Intrinsa treatment
compared to placebo in both trials.
Patients receiving testosterone treatment
reported approximately a 65 percent decline in
distress compared to their baseline.
In addition to assessing sexual desire,
distress and satisfying sexual activity, we also
pre-specific all other domains of the PFSF as
endpoints. Recall that these domains reflected the
other important losses that women with HSDD
complained of and said they would like to have
corrected.
[Slide.]
Shown here is SM1. With Intrinsa therapy,
patients experienced an improvement in every domain
of sexual functioning over placebo; that is, they
had an improvement in arousal, orgasm, improvement
in sexual pleasure, reduction of sexual concerns,
improved sexual responsiveness, and improved sexual
self image.
[Slide.]
Here are the results of SM2. Improvements
55
were very consistent to that seen in SM1. Again,
all PFSF domains improved significantly with
testosterone therapy compared to placebo.
[Slide.]
This graph shows increases in the other
sexual activity endpoints measured on the Sexual
Activity Log. Not only were satisfying sexual
episodes increased within Intrinsa treatment--which
I showed you previously as our primary
endpoint--but patients treated with testosterone
also experienced significant increases in numbers
of sexual episodes and in numbers of orgasms.
Note here, with the Sexual Activity Log,
we are assessing numbers of orgasms, while with the
PFSF domain of orgasm, we were assessing the ease
and reliability of orgasm.
[Slide.]
In summary, after listening to what was
important to women with HSDD, we developed and
validated instruments to measure clinically
meaningful endpoints. Both of our pivotal trials
showed strong and consistent efficacy using these
56
rigorously developed instruments.
Surgically menopausal women with
hypoactive sexual desire disorder experienced a
significant increase over placebo in sexual desire
and satisfying sexual activity. They also
experienced a significant decrease in distress
associated with their low sexual desire. All three
of these endpoints are critical to assessing a
relevant treatment response in this clinical
condition.
Patients further experienced benefits in
sexual arousal, orgasm, pleasure, sexual
responsiveness, sexual self image, and a decline in
sexual concerns. Patients also had increased
number of sexual events, and numbers of orgasms.
The consistent significant improvement in
all areas of sexual functioning identified by
patients as important demonstrates the clinical
meaningfulness of Intrinsa therapy to these women
with HSDD.
Dr. DeRogatis will now provide additional
perspective on the clinical relevance of these
57
results.
Thank you.
Clinical Relevance of Treatment Effects
DR. DeROGATIS: Good morning.
As the Director of the Johns Hopkins
Center for Sexual Health and Medicine, I see and
treat a full spectrum of patients with sexual
disorders. I also have a strong professional
interest in the methods and techniques determining
clinical relevance of the results of clinical
treatment trials.
Dr. Lucas has shared with you the efficacy
results from the Phase III trials. Now I'd like to
speak with you about the clinical relevance of
those results.
[Slide.]
To begin with, I'd like to emphasize that
the results from the Phase III clinical trails
implicitly speak to clinical relevance in a number
of ways. First, recall, as Dr. Lucas pointed out,
all of the endpoints for both the Phase II studies
and the clinical relevance study are derived
58
directly from conversations, interviews, etcetera,
with patients who expressed their concerns,
experiences and what was important to them.
Instruments developed in this way, by the very
nature of the process, have clinical relevance for
the patients.
Second, the decrease in personal distress
indicates in a very direct manner the impact that
the testosterone patch treatment had on the
patients' feelings about heir disorder.
Third, the fact that all of the PFSF
domains--not just desire--show improvement with
testosterone patch treatment gives us confidence
that there is a meaningful treatment benefit
associated with the patch.
In addition to this implicit evidence from
the Phase III trials, the sponsor also did a formal
Clinical Relevance Study which employed the
anchoring technique, and I'd like to tell you more
about that now.
[Slide.]
The anchoring technique for examining
59
clinical relevance is very well established through
broad use in numerous disciplines of medicine.
Advantages of the method are that it's readily
understood. Perhaps the most important advantage
involves the fact that it's patient-based; it
utilizes direct questions of the patient as to the
patient's perception of clinical benefit. These
patient perceptions of benefit are then tied,
through statistical analyses, to the major study
endpoints. The statistical analysis in this case
is Receiver Operating Characteristic Analysis--or
ROC analysis.
The results then define meaningful change,
in terms of what are referred to as "minimum
clinically important differences," which helps to
establish clinical relevance.
[Slide.]
What you're looking at now is a model of
the anchoring technique that was used to relate
patient perceptions of meaningful benefit to
changes in major outcomes measures in the Phase III
trials. The Clinical Relevance Study essentially
60
had three stages. In the first stage patients were
queried directly as to their experience, or not, of
what they perceived to be a meaningful benefit.
Those that answered "yes" were indicated as
"responders," those that answered "no" were
indicated as "non-responders."
In step to, these data were then utilized
in receiver operating characteristic analysis, with
each of the three main study endpoints: the PFSF,
the SAL and the Distress Scale.
Third, each endpoint MCID score was then
applied back to the total Phase III population to
establish proportions of responders and
non-responders in the Phase III trials.
In addition to the anchoring technique, we
also asked patients if they wanted to continue on
the treatment. This is a very telling question in
regard to patient's perceived benefits. And as
we'll see a little later, the results were quite
interesting.
[Slide.]
Let me tell you a bit more about the
61
details of the Clinical Relevance Study. The
Clinical Relevance Study was performed during a
two-week period immediately subsequent to the
24-week double-blind randomized trial. 132 women
from the Phase III trials were the sample for the
Clinical Relevance Study. All of the interviews
were done by a single female interviewer to reduce
variability--potential variability--across
interviewers.
The interviews began with rather
open-ended questions about experiences in the
clinical study and prior to the clinical study, and
progressed to more specific questions about
perceived benefits, if any, on the part of the
patient.
The specific question for the anchoring
analysis was: "Overall, considering everything
we've talked about today, would you say that you
experienced a meaningful benefit from the study
patches?" Patients answered either "yes" or "no."
Subsequently, when the blind was broken,
we learned that 52 percent of the patients on
62
testosterone, and 31 percent of the patients on
placebo indicated they had experienced a meaningful
benefit or were responders. Now, these rates of
response are actually fairly typical of a number of
areas in which clinical trials are done in
medicine. They're comparable to the rates we see
in irritable bowel studies, in incontinence
studies, in studies of depression, and so were not
unusual and really not unexpected.
Following the relevance question
concerning benefit, these data were then utilized
in the ROC analysis to essentially determine
responders and non-responders.
[Slide.]
Focusing on this distinction for a moment,
this table contents the mean change from baseline
in the responder and non-responder group in major
study outcomes.
As you can see, they're very different.
If you take satisfying sexual activity for a
moment, we see a mean change of 4.4 satisfying
sexual activities per month on the part of the
63
responders, and a fractional 0.5 change for the
non-responders.
In terms of sexual desire, there's a 21
point increase in sexual desire in the responder
group, which translates into moving them from
"seldom" feeling sexual desire, to "sometimes"
feeling sexual desire--which, by the way, is the
modal frequency for women in this age group.
Distress also shows a dramatic shift in
the desired direction: almost 37 point reduction in
distress in the responder group. And, again, this
translates into feeling personal distress "often"
to feeling personal distress "seldom." So, some
fairly dramatic results in our responder group.
[Slide.]
Now let me share with you an ROC
analysis--okay? And let me talk about this just
briefly for a moment. What we see in our ROC
analysis is a vertical axis which has our true
positives on it. Our horizontal axis has false
positives. And the ROC curve is actually the
entire distribution of changes from baseline on
64
this outcomes measure--in this case "satisfying
sexual activity"--in the sample.
Now, where the white diagonal from
upper-left to lower-right intersects the ROC curve
is the point of optimum discrimination; that is,
the value that optimally correctly assigns
responders and non-responders, and minimizes false
positive and false negative errors. This value for
satisfying sexual activity, is 1.11 satisfying
sexual activities a month.
It's a little difficult to conceptualize
fractional sexual activities. And so the sponsor
decided to modify this value a little bit to
"greater than 1," thereby retaining the accuracy of
the result, but sparing us the difficult task of
trying to conceive of fractional sexual activities.
ROC analysis has a summary statistic
associated with it that tells us, essentially, how
good the discrimination was--in this case, between
responders and non-responders. This is referred to
as the "area under the curve"--AUC
coefficient--which on this particular analysis is
65
.77.
Now, this is very good, since experts in
the field believe that .80 is actually an excellent
discrimination. Analogous analyses were done for
sexual desire and for personal distress. The
coefficients in those instances were .77 and .78,
respectively.
So, we wind up with excellent
discrimination in our constellation or set of
outcomes measures.
[Slide.]
Now, lets look at our anchoring model with
all the data filled in.
Over on the left we see our information on
patient perception of meaningful benefit: 52
percent of those on testosterone, and 31 percent of
those on placebo have a meaningful benefit. These
data, when integrated with ROC analysis produce
these optimum cutting scores--or MCIDs--which, when
applied to the Phase III population generate these
proportions of responders--okay?--on each of our
study outcomes measures.
66
Now, the three main points to be made here
are the following: first, these percentages are
approximately equal across the three major study
endpoints; second, each shows a statistically
significant advantage for the active treatment;
third, the Phase III results or rates of responders
are very analogous to the rate of responders we saw
in the Clinical Relevance Study.
[Slide.]
Now I'd like to share one final set of
results from the Clinical Relevance Study involving
interest in continuing treatment. In this graph
what we see is that 80 percent of those who
indicated they had a meaningful benefit also
indicated that they probably or definitely would
continue treatment with the patch if it were
available. Conversely, an approximately equivalent
proportion of individuals who did not experience
meaningful benefit indicated they would not wish to
continue treatment with the patch.
Now, these results are an important
measure of clinical relevance because they help
67
confirm that the patients' assessment of meaningful
benefit is truly valid in terms of intended
behavior, not simply perception.
[Slide.]
So, in summary, what did we learn from the
Clinical Relevance Study?
First, we learned that significantly more
testosterone patients than placebo patients
experienced meaningful benefit.
Second, anchoring, using minimum clinical
important difference values, confirms similar
proportions of responders in the Phase III studies.
Third--and perhaps most importantly--these
results are consistent across all study endpoints:
sexual desire, satisfying sexual activity, and
personal distress.
[Slide.]
So, in conclusion, I think I have to
conclude that a consistent pattern of outcomes
shows strong evidence of a clinically meaningful
benefit, which translates into observable clinical
relevance. Both as a clinical scientist and a
68
practicing clinician, I would conclude that these
data strongly support the sponsor's contention that
the efficacy outcomes achieved by the testosterone
patch are, in fact, clinically relevant.
Thank you.
Phase III Clinical Safety Data
DR. LUCAS: Now to turn to safety.
[Slide.]
As we started to plan the safety
evaluations to be included in the surgical
menopause program for Intrinsa, we identified these
areas of potential concerns for testosterone in
women. We reviewed the literature for known
androgen effects in animals and humans, including
hyperandrogenism in women, and adverse effects
reported with androgen treatment in men.
We also consulted with experts
knowledgeable in two areas where there was less
literature: high dose androgen used in
female-to-male transsexuals, and androgen abuse
with athletes and body builders. We then sought to
assess each one of these areas in our safety
69
evaluations.
[Slide.]
Recall--as I told you earlier--the SM
trials--SM1 and 2--were essentially of identical
design: six-months double-blind, followed by
open-label. Baseline health risks were balanced in
the treatment arms and across trials.
The primary safety data I will present are
from the double-blind portion of the SM1 and SM2
trials--seen here in brackets--since, as you know,
a placebo-control offers the best opportunity to
determine a drug effect. I will also show you dat
on patients who dropped from the study during the
open-label period--which you see here in white.
Finally, I will present double-blind adverse event
data from the still ongoing natural menopause
program. These data are very similar to those seen
in the surgical menopause trials, and provide the
opportunity to see double-blind data with up to 12
months of testosterone exposure.
[Slide.]
1,800 patients in the Phase II and III
70
treatment trials have been exposed to 300 mcg per
day testosterone or greater. Over 1,300 patients
have received at least six months of therapy, and
over 600 have received over 12 months of therapy.
This represents an exposure of more than 14,000
months, or over 1,000 patient years to the 300 mcg
per day testosterone patch in post-menopausal
women.
[Slide.]
I will show you the safety data in this
order, starting with adverse event profile, weight
and vital signs, laboratory evaluations, and,
finally, breast cancer.
[Slide.]
The numbers of patients reporting AEs
overall--in both Phase III studies--were similar
between the active and the placebo groups, and
between the two trials. About 75 to 80 percent of
patients had at least one health occurrence during
the six-month period. The small number of serious
adverse events, and dropouts due to adverse events,
were consistent between active and placebo
71
treatment, and between the two studies, as well.
No deaths occurred in the SM1 trial.
There was one death in the placebo group of the SM2
trial.
Looking at the most common AEs about
one-third of patients reported an application-site
skin reaction, including mild redness or itching at
some time during the six-month trial. About 75
percent of these reports were considered "mild."
Application site reactions were not higher in
patients receiving patches containing testosterone.
About 3 percent of patients dropped out because of
these site reactions in all of the arms.
[Slide.]
Now, to look at these known androgenic
effects with ADs and clinical assessments.
In the testosterone-treated arms, 94
percent of the androgenic AEs were assessed as
"mild." 78 percent of patients reporting
androgenic AEs experienced only one of them during
testosterone therapy. Further, time to even was
not different between the active and
72
placebo-treated groups for any of the androgenic
AEs.
[Slide.]
Acne was evaluated in two ways: clinical
assessment and AEs. Clinicians were trained in the
use of the Palatsi scale to evaluate patients for
facial acne at scheduled visits. About 98 percent
of patients in both active and placebo groups had
no positive change in their acne score.
Looking at acne AEs--which were reported
by both physicians and patients--there were no
differences between active and placebo in SM1. In
SM2 there were fewer reports of acne in the placebo
group than in the testosterone treated group.
[Slide.]
Facial hair was also evaluated in two
ways: clinical assessments and AEs. We also
trained our investigators to use a modified
Ferriman-Gallwey scale to assess facial hair
growth. With these evaluations, we again saw no
difference between active and placebo in SM1. In
SM2 we did see a mild increase in active over
73
placebo, primarily in chin-hair growth.
Patient-reported AEs mirrored the
objective evaluations for facial hair growth, with
no differences seen in SM1, and higher reporting
with testosterone in SM2.
[Slide.]
Now, to look at the less common androgenic
AEs--investigators specifically asked patients
about hair loss and voice deepening at regular
visits. Answering "yes," as well as spontaneous
reports, were included as AEs. Again, we saw no
differences between groups in SM1, and a slightly
higher incidence in SM2 in the active group over
placebo.
The one reported case of clitoromegaly in
SM2 was reported by phone after 12 weeks of
therapy. This was not confirmed by physical exam.
She withdrew from the study and, when contacted
five weeks later, she reported the condition had
resolved.
[Slide.]
This slide shows the withdrawals due to
74
androgenic AEs. Please note that some patients
reported more than one androgenic AE, so this
represents an overall withdrawal rate due to
androgenic AEs in the two trials of 1.2 percent,
and 1.8 percent in the testosterone arms.
[Slide.]
We also looked at other adverse events
that could be important to this population. We
looked at breast tenderness and hot flushes to
confirm that we were not potentiating estrogenic
adverse events, and saw no increase with
testosterone therapy.
AEs of weight gain were examined and were
minimally higher in the testosterone-treated
groups.
[Slide.]
When we looked at weight gain objectively,
we saw a similar small difference. Mean weight
gains of about 1/4 of a kilogram--or - pound,
compared to similar mean weight losses of about -
pound in the placebo arm were seen. We also
examined vital signs and saw no change from
75
baseline in mean systolic or mean diastolic blood
pressure in either SM trial, in either arm.
[Slide.]
We also looked at other adverse events
that were reported in the male testosterone labeled
products. We saw no difference in anxiety, edema
or aggression.
[Slide.]
Because we saw liver function testing AEs
in the testosterone arms only, we evaluated each of
these patients individually. This represents six
patients, which I would like to tell you about.
Of the six, two of the patients had
isolated bilirubin AEs--increases. Both patients
wree just at the upper limit of normal, and
increased very minimally at 24 weeks.
Four patients had transaminase increases.
Three of the four had mild increases, with less
that two times upper limit of normal. Of these
three patients, two patients remained on
testosterone therapy, and their levels returned to
normal. One patient was still mildly
76
elevated--just above the upper limit of
normal--after completed 52 weeks of therapy, and
when she was examined after completing the trial,
and off testosterone, remained just about the upper
limit of normal where she had been before.
The final AE--the patient with a moderate
transaminase elevation--her ALT went from 9 to 91,
her AST went from 12 to 94--she also returned to
normal while remaining on testosterone therapy
after she discontinued chetachinasol, a drug known
to be associated with elevated liver function
testing.
[Slide.]
When we examined liver function testing
overall, we saw no difference in the mean changes
from baseline in any parameter of liver function
testing--as shown here--for either placebo or
testosterone. When we looked for outliers, no
difference was seen between treatment arms for any
parameter.
[Slide.]
Now, moving on to hematology, there was a
77
small--very small--clinically insignificant mean
increase in hemoglobin and hematocrit, seen in both
testosterone groups. This averaged about - percent
of hematocrit, and about 1/6 gram per dL of
hemoglobin.
[Slide.]
To confirm this was of no concern to
patients at highest risk for polycythemia, we
evaluated all patients on testosterone therapy with
this scattergram. Baseline hemoglobin levels are
on the x-axis; 24-week hemoglobins are on the
y-axis. Therefore, for a patient who did not
change, her dot would fall on the 45 degree line.
Generally, we saw no large increases.
Of particular concern, though, are
patients who started off with a high hemoglobin at
baseline. These patients, like patients generally,
show no evidence of concerning increases.
[Slide.]
Now to talk about assessing changes
associated with cardiovascular risk. We've already
talked about blood pressure and weight; now to talk
78
about carbohydrate and lipid metabolism.
To assess carbohydrate metabolism, we
measured fasting glucose, fasting insulin, and
HbA 1c. As you can see here, mean
increases were not
different in the testosterone and placebo-treated
groups in either trial. When we looked for
outliers in each of these parameters, there was no
difference seen between the treatment groups.
[Slide.]
Additionally, no drug effect was seen on
any parameter of the lipid profile that we
measured, either: total cholesterol, HDL, LDL, or
triglycerides. Again, outliers were not different
in the two treatment groups for any of these lipids
measured.
[Slide.]
We also evaluated changes in coagulation.
Here you see mean changes in laboratory evaluations
for clotting. Again, we saw no evidence of adverse
changes with testosterone therapy.
[Slide.]
We also looked at women who were at
79
potentially the highest risk for cardiovascular
risk, with four adverse laboratory changes.
When we looked at 150 women who were
positive for at least three of these five criteria,
that included obesity; adverse lipid profile or on
a lipid-lowering agent; hypertensive or on an
anti-hypertensive; or had elevated fasting
glucose--our best surrogate for metabolic syndrome
in the trials.
We found no evidence of more adverse
laboratory changes in this subpopulation.
[Slide.]
Here we see the AEs reported in the
open-label period of weeks 25 through 52, and the
extension, 53 through 78. Patients who were on
placebo for the first six months, and then on
open-label the second six months are noted
"P->TTS," and as they enter the second year,
"P->TTS->TTS."
Patients who were randomized initially to
testosterone, and then went into open-label
testosterone are noted "TTS->TTS," and then, again,
80
"TTS->TTS->TTS."
During the open-label period, patients
with up to 12 months of exposure in the TTS->TTS
group showed only minimal differences in overal
serious AEs or withdrawals due to AEs, compared to
patients who were having up to six months of
exposure in the P->TTS group.
Incidents of AEs were similar in the week
53 to 78 extension to those seen earlier with up to
18 months of exposure.
Average withdrawal rates due to androgenic
AEs were also very similar, regardless of exposure
up to 18 months.
[Slide.]
Looking at some of the same parameters in
the natural menopause population, we see a very
similar picture to what we just saw in the surgical
menopause. NM1 is the six-months double-blind
trial. NM2 is a one-year--52-week-double-blind
trial.
This trial is still ongoing, so that this
data is interim data. You will notice that the
81
numbers are not equal because this reflects a
one-to-two randomization.
Like surgical menopause, about 70 to 80
percent of patients in natural
menopause--regardless of arm--experienced an
adverse event during the study. Serious adverse
events and study withdrawals were very similar for
patients with placebo and active treatment. Two
deaths occurred from motor vehicle accidents in
NM1.
Breakthrough bleeding AEs were numerically
lower with testosterone treatment in both natural
menopause trials.
[Slide.]
The natural menopause trials also give us
an opportunity to see androgenic AEs with up to 12
months of exposure in a double-blind trial
situation. Minimal differences in androgenic AEs
were seen between the active and placebo trials in
NM1--with the most differences seen in acne and
hirsutism.
Additional blinded exposure of up to 12
82
months in NM2 did not appear to increase the
incidence of reported androgenic AEs compared to
placebo, or compared to six months exposure in NM1.
[Slide.]
Looking specifically at withdrawals due to
androgenic AEs, minimal differences were seen
between the active and placebo arms, with similar
rates to that that we saw in the SM trials.
Additional blinded exposure of up to 12 months in
NM2 did not appear to increase withdrawals due to
androgenic AEs compared to placebo, or to six
months exposure in NM1.
[Slide.]
The final parameter that I would like to
discuss is breast cancer. We performed mammograms
in all patients over 40 years of age at entrance
and at on year, or exit, if it was greater than six
months.
In our overall program of 2,200 patients,
four patients have been diagnosed with breast
cancer; one DCIS, and three invasive cancers. All
were in the surgical menopause program, and all
83
were initially randomized to placebo. Remember
that in our Phase III program, patients received
testosterone therapy in the open-label second six
months, regardless of what they were randomized to
initially.
[Slide.]
We evaluated these four cases. Case 1, we
saw in Phase II, and she did not receive any
testosterone. Case 2 was a 63-year-old who
presented with an axillary mass determined to be
metastatic adenocarcinoma, after she received five
weeks of open-label testosterone therapy.
[Slide.]
Case 3 was a 56-year-old who was diagnosed
with tubolobular carcinoma after 37 weeks of
testosterone treatment. Her diagnosis was made
after findings seen in her baseline mammogram
became more prominent.
Case 4 was a 50-year-old patient who was
diagnosed with ductile carcinoma in situ after 24
weeks of testosterone therapy.
This diagnosis was made based on a new
84
mammographic finding.
[Slide.]
The number of breast cancers observed in
our clinical program is not unexpected and, based
on calculations of expected rates, based on our
number of women and their risk profile. Further,
detecting breast cancer only in patients who
received the least amount of testosterone because
they were initially randomized to placebo, is not
consistent with an association with testosterone.
Also, in patients who have received
open-label testosterone for up to a second year of
exposure, no additional breast cancers have been
identified.
[Slide.]
In summary, after examining safety
evaluations in more than 1,300 women for six
months, and 600 women for a year, we found overal
AEs, serious AEs, and withdrawals due to AEs were
generally similar in the active and placebo-treated
groups.
Small increases in androgenic AEs and
85
assessments were seen primarily in one surgically
menopausal trial. Mean weight gains over placebo of
about a pound were seen at 24 weeks. And, except
for a very small increase in red-cell mass, no
changes were seen in any mean lab parameter, or
evidence of adverse changes in outliers of any lab
parameter.
So, in conclusion, in surgically
menopausal women with HSDD, the 300 mcg per day
testosterone patch was very well tolerate. No
serious safety concerns have been identified to
date. In both Phase III surgically menopausal
trials, we continue to monitor patients receiving
open-label therapy, most now in their third year of
therapy with no further safety signal being seen.
The safety changes which have been
identified are generally mild, were rarely
associated with withdrawal, and can generally be
easily diagnosed and easily monitored by the
patient herself.
Thank you. And now, Dr. Glenn Braunstein
will address hormone delivery.
86
Phase III Hormone Data
DR. BRAUNSTEIN: Good morning.
I'm an endocrinologist and clinical
investigator with over 30 years of experience in
treating women with hyperandrogenic disorders, as
well as androgen insufficiency.
Dr. Lucas has shown you today that the
testosterone transdermal patch is both safe and
efficacious. But one of the concerns that we
clinicians have is the potential impact of
long-term exposure to androgens administered
therapeutically.
So my role is to examine the hormone
levels achieved with Intrinsa, and to show the
relationship between the androgen levels and the
efficacy, and especially the safety parameters.
Hopefully, my presentation will provide additional
reassurance that the transdermal testosterone
system is safe and efficacious.
[Slide.]
Here is my agenda. To put the levels of
testosterone achieved with Intrinsa into
87
perspective, I will first review how testosterone
circulates in the blood, because it is important to
realize that not all the testosterone that is in
the blood is available to the tissues. It is also
important to understand that one of the main
determinants of how much testosterone is available
to the tissues is the level of sex-hormone-binding
globulin--or SHBG. Therefore, I will show you the
effect of Intrinsa on the levels of SHBG, as well
as testosterone.
Then I will show how the free testosterone
levels correlate with the efficacy parameters and
the safety parameters in patients receiving
Intrinsa.
And, finally, I'll also address some
specific safety issues concerning estrogen-related
target tissues: the breast and the uterus.
You will see that the median levels of
free and bioavailable testosterone achieved fall
within a physiological reference ranges for
pre-menopausal women; that there is no accumulation
of testosterone over 12 months of therapy; that the
88
testosterone levels correlate with the efficacy
parameters; and that there do not appear to be any
many hormone safety issues with Intrinsa.
[Slide.]
This is a diagram of how testosterone
circulates in the blood. Approximately 98 percent
of the total testosterone in the blood is bound to
serum proteins, and the major protein is
sex-hormone-binding globulin, or SHBG, to which
approximately two-thirds of the testosterone is
bound. This binding is very tight and, basically,
the testosterone on SHBG is unavailable to the
tissues.
Approximately a third of the testosterone
is bound to albumin. And this albumin-bound
testosterone is very loosely bound and may diffuse
off of the albumin and enter target tissues.
Only about 1 to 2 percent of the total
circulating testosterone exists in the free or
biologically active state. And we refer to the
combination of the free and the albumin-bound,
weakly bound, testosterone as being bioavailable.
89
[Slide.]
To put the testosterone levels in
perspective for you, we have established a
reference range, shown by the dotted lines on the
right side of the screen. Note that at the present
time there are no accepted reference ranges for
androgens in women. Testosterone levels in women
vary widely. There is an age-related change, with
the levels peaking in the 20s and then declining.
There is a diurnal variation, with levels being
higher in the morning than in the evening. And
there's also changes throughout the menstrual
cycle, with levels being highest at the mid-cycle
time.
We elected to use a range based upon 161
women aged 18 to 49 years, who had multiple blood
samples attained across the menstrual cycle, and
the hormones were measured and the levels averaged.
Since this is a range that women achieve naturally
during their reproductive life, we felt that the
androgen levels within or near this physiological
range would be safe.
90
On the next several slides, I will show
you the free, bioavailable and total testosterone
levels, before and after Intrinsa therapy. Because
the levels achieved in both the SM1 and SM2 trials
were virtually identical, we have combined the
results for easy presentation. However, the
individual results are shown in the briefing book,
and I'll be happy to discuss the results during the
discussion period.
[Slide.]
So we will begin by looking at the free
testosterone levels as shown here. And that's
because these are the most important levels, as far
as the cells in the body are concerned.
The circles represent the median hormone
levels, while the whiskers represent the 10
th and
90 th percentiles.
The baseline free testosterone levels
were, as expected, at or below the lower limit of
the reference range in both the patients who were
assigned to placebo, and those randomized to the
testosterone system. This was what was expected,
91
since these are all women who had undergone
surgical oophorectomy, were receiving estrogen, and
had hypoactive sexual desire disorder.
At the end of 24 weeks, the median level
of free testosterone in the patients on Intrinsa
had increased into the reference range, while there
was no change in the placebo group. And, very
importantly, the levels were very similar at 24 and
52 weeks in the Intrinsa group, indicating no
continued accumulation of testosterone over time.
[Slide.]
Similar results were find for bioavailable
testosterone, which again represents the
combination of free testosterone as well as the
loosely bound albumin testosterone. And, again,
there was no significant accumulation of
bioavailable testosterone between 24 and 52 weeks.
[Slide.]
This slide shows the total testosterone
data. And, although the median total testosterone
levels achieved with Intrinsa were above the
reference range--in part, due to the relatively
92
high levels of SHBG, which I'll show shortly--it is
important to emphasize that this was not seen as a
cause for concern because the free and bioavailable
testosterone levels--the ones that are most
important as far as the tissues are concerned--were
in the reference range for a majority of the
patients.
We also found that the total testosterone
levels did not change between 24 and 52 weeks.
[Slide.]
Median levels of serum SHBG were stable
over a one-year period of time, and they were all
at the upper limit of normal for the reference
range, reflecting the fact that three-quarters of
the patients in the study were receiving oral
estrogens, while one-quarter were receiving
transdermal estrogens.
[Slide.]
Now let's look at the correlations between
testosterone levels and the efficacy parameters.
This slide shows the consistency of
results across the Intrinsa clinical studies. We
93
find statistically significant correlations between
the changes in free testosterone and the changes in
the efficacy parameters, including total satisfying
sexual activity, desire, and a decrease in personal
distress.
Very similar correlations were also found
for bioavailable testosterone and total
testosterone.
This finding is also consistent across all
of the trials. We have pooled the Phase II trials,
as shown initially. We have pooled the Phase III
trials, and also in the NM1 trial--the natural
menopause trial--the results show a tremendous
amount of consistency.
Now let's look at the changes in androgen
levels and safety parameters.
[Slide.]
We will examine the effect on estrogens,
either through aromatization to
testosterone--aromatization of testosterone to
estradiol, or displacement of estradiol off of SHBG
94
by testosterone. We will also look at the effect
of Intrinsa on estrogen-responsive tissues,
including the breast and the endometrium.
We will show the relationship between the
reported androgenic adverse events and the observed
androgenic effects to she testosterone levels, and
we'll show you the results with the free
testosterone correlations and clinical laboratory
measurements.
Although the data linking estrogen use to
breast cancer is unclear, the relationship between
unopposed estrogen use and endometrial hyperplasia
and neoplasia has been established. Therefore, it
was important to examine the effect of Intrinsa on
the serum estrogen levels and on estrogen target
tissues--including the breast and the uterus.
[Slide.]
As shown here the serum total estradiol
levels were similar in patients who received
placebo and those who received Intrinsa, and there
were no changes through 52 weeks on Intrinsa. The
levels were identical between those on placebo, and
95
those receiving Intrinsa.
Similar results were found for free
estradiol in the serum, as well as serum estrone.
Thus, there does not appear to be any major
increase in aromatization--at least as assessed by
levels in the serum.
Currently available data would also
indicate that aromatization does not appear to be
an issue in the tissues either, as we'll summarize
on the next slide.
[Slide.]
A study was carried out at the Karolinska
Institute in Sweden, in which women who had not
undergone hysterectomy, and who were receiving
estrogen and norenthindrone were randomized to
receive either placebo or 300 mcg ga day of
transdermal testosterone through the patch system.
They underwent mammograms and fine-needle
aspirations of the breast, both at the baseline
state and after six months of therapy. The results
showed that there were no differences between the
two groups in respect to mammographic breast
96
density, or in breast epithelial proliferation.
There was, however, a significant decrease
in the stromal cell proliferation in patients
receiving transdermal testosterone. And this is
important because the stromal cells are major
source of aromatase enzyme activity in breast
tissue.
In addition, preliminary analysis carried
out on paired endometrial biopsy samples--that's
approximately 300 samples--obtained from women in
the Natural Menopause Study were examined. And two
cases of endometrial hyperplasia were found. Now,
that study is still blinded, and therefore we don't
know whether the samples were obtained from women
who were receiving placebo, or receiving
transdermal testosterone. But even if we assume
that both patients were receiving transdermal
testosterone, that would give an incidence of
approximately 1 percent of endometrial hyperplasia,
which falls within the FDA guidance for
estrogen-progestin combination products.
So these results are consistent with the
97
lack of any estrogen-related side effects relative
to placebo observed in the safety data presented by
Dr. Lucas.
Now we'll turn to the androgenic side
effects.
[Slide.]
Although the androgenic adverse events in
each of the Phase II and Phase III studies were not
different between women on active therapy versus
placebo, we pooled the results from the Phase II
and the Phase III trials on surgically menopausal
women in order to examine the issue with a larger
group of patients.
We used a trend test to look at the
incidence of androgenic adverse events against the
highest pre-testosterone level measured in each of
the women. The only statistically significant
finding was with facial hair. This was also borne
out with the combined observations of the
investigators at the different trial sites.
Importantly, there were no specific associations
between androgen levels and acne, alopecia, voice
98
deepening, or clitoromegaly.
And, in a similar manner, we examined the
relationship between changes in the laboratory
parameters and the maxim free testosterone obtained
in each of the women. And these were examined
across both the Phase II and Phase III studies,
which were combined.
These laboratory parameters included tests
of liver function, lipid analysis, carbohydrate
metabolism, hematology values and clotting factors.
These parameters were chosen because they had been
associated with abnormalities with the use of
pharmacologic doses of testosterone, based on
literature studies carried out in both men and
women. And, again, the changes for patients within
the highest decile of free testosterone, compared
to placebo, were small and clinically
insignificant.
[Slide.]
So, in summary, surgically menopausal
women with hypoactive sexual desire disorder had
low testosterone levels at baseline. 300 mcg a day
99
of transdermal testosterone increased the
concentrations of free, bioavailable and total
testosterone, with no evidence of continued
accumulation over 12 months of dosing.
Serum concentrations of total and free
estradiol, estrone, and SHBG were not affected bu
Intrinsa administration over the year of study.
Higher exposure of free and total
testosterone was not associated with clinically
significant laboratory changes.
[Slide.]
Intrinsa raised the median free and
bioavailable testosterone levels to within the
pre-menopausal reference range that we established,
and these correlated with an increase in the number
of satisfying sexual events, the increase in sexual
desire, and a decrease in personal distress.
Based upon the currently available hormone
data, the one-year safety profile of Intrinsa shows
no cause for concern. Higher free testosterone
levels are associated with small increases in
facial hair in the pooled trials, and the
100
appearance of such androgenic effects would allow
the patient to make a personal risk versus benefit
decision.
Thank you.
It's now my pleasure to introduce Dr.
Shifren.
Hypoactive Sexual Desire Disorder Medical Need
DR. SHIFREN: Good morning.
As a gynecologist and director of the
Vincent Menopause Program at the Mass General
Hospital, I see many women with sexual concerns
after menopause.
Hypoactive sexual desire disorder is the
most common sexual problem that I see in my
practice. It's a particularly poignant problem for
our younger patients who've had their ovaries
removed. After removal of the ovaries, a woman
loses almost all of her estrogen, and approximately
half of her testosterone. We have many options
available for estrogen replacement, but currently
there are no approved testosterone products to
treat our surgically menopausal women who present
101
with sexual dysfunction post-operatively.
Today, I'll tell you about the negative
impact that HSDD has on menopausal women, and talk
also about why I'm an advocate for a low-dose
testosterone patch.
Let me begin by describing some of the
research that helps us understand HSDD.
[Slide.]
The Women's International Study of Health
and Sexuality--or the WISHeS Study--was
specifically designed to better understand HSDD in
women, as available studies didn't completely
capture all important aspects of this disorder--in
particular, distress. This was a self-report
survey of more than 4,500 women in the United
States and Europe between the ages of 20 and 70.
I'll be discussing a U.S. sub-population of 520
surgically and naturally menopausal women with
partners.
This survey used three validated
instruments: the Short Form-36, a measure of
overall health status; and the Personal Distress
102
Scale; and Profile of Female Sexual Function.
So what did these studies teach us?
[Slide.]
Well, we learned that women with HSDD do
engage in sexual activity. As these data
illustrate, partner-initiated sexual activity is
constant for women, whether they have no or normal
desire. But, importantly, women with HSDD are
significantly less likely to initiate sexual
activity. This confirms that focusing only on the
frequency of sexual activity can be misleading when
trying to understand the sexual experience of women
with HSDD.
[Slide.]
Not surprisingly, we also learned that
sexual desire correlates with a woman's overall
level of satisfaction with her sexual life. This
graph shows that women with less desire were less
satisfied with their sex lives. Women with higher
desire scores were more satisfied overall.
[Slide.]
With respect to the effect of low desire
103
on couples, for women, sexuality typically exists
within the context of the important relationships
of their lives. As you can see in this graph,
sexual satisfaction and relationship satisfaction
are intricately related for women. Women who are
less satisfied with their sex lives are
significantly less satisfied with their personal
relationships or marriage.
[Slide.]
As we see in these data from WISHeS--and
as my patients tell me often--the distress women
feel with HSDD extends beyond their loss of desire.
Compared to women with normal desire, women with
HSDD report feelings of low self-esteem, shame and
failure. My patients often are very upset
regarding the impact that their low desire has on
their relationships. And, of note, nearly 90
percent with HSDD report feelings that they're
letting their partner down.
[Slide.]
It's important to realize that HSDD not
only affects a woman's sexual health, but is
104
associated with overall diminished health status.
The WISHeS investigators used the Short Form-36--a
well known and validated measure--to measure study
participants' general health status. Compared to
women with normal desire--in green--women with HSDD
were significantly more likely to report decreased
physical function, general health and vitality.
They also reported lower levels of social
functioning, emotional and mental health.
The SF-36 scores we saw in women with HSDD
are very similar to scores seen in people with
other medical conditions, such as arthritis or
diabetes.
Now we have an understanding of HSDD and
the impact this has on a woman's life. But why
does it happen?
[Slide.]
Female sexuality is a complicated
interplay of physiology, psychology, interpersonal
relationships, and socio-cultural influences. As a
reproductive endocrinologist, my focus is on the
physiologic factors that affect a woman's sexual
105
health. When a woman presents with HSDD, the first
thing I do is rule out non-physiologic causes, as
these couples often benefit from education,
counseling, sex-therapy and lifestyle changes.
Physiologic factors, including medical,
neurologic, gynecologic and urogenital problems all
may have a negative impact on healthy sexual
functioning. For example, I've seen many
menopausal women in my practice with vaginal
atrophy and dysprunia associated with estrogen
deficiency. For these women, sexual arousal and
response often improves greatly after starting
vaginal estrogen therapy.
[Slide.]
But the most common physiologic cause of
HSDD that I see in my practice is androgen
insufficiency--often associated with oophorectomy.
As you can see in this study, removing the ovaries
of both pre- and post-menopausal women results in
an approximate 50 percent decline in testosterone
levels. Many studies over the past 20 years have
shown significant improvements in sexual desire,
106
response and intercourse frequency in surgically
menopausal women treated with testosterone. But no
testosterone product currently is available for
treating our surgically menopausal patients with
HSDD.
So what are physicians and patients doing?
[Slide.]
Unfortunately, we're using products that
have been formulated for men, putting women at risk
for receiving high doses of testosterone.
Compounded products, with limited quality control
and dosing consistency, also are being used.
As you can see, in 2003, approximately 20
percent of total prescriptions for branded male
testosterone products were actually written for
women. And in that same time period there were
over 1 million prescriptions written for compounded
or generic testosterone products, for women.
These data show the need for a quality
testosterone product specifically formulated for
menopausal women, with proven efficacy for all
aspects of HSDD.
107
[Slide.]
Before concluding, I'd like to address a
concern that some of you may have regarding the
clinical meaningfulness of the changes in sexual
function that we saw with testosterone patch
treatment.
An increase of about two satisfying sexual
events per week may not seem like a lot, but for
women whose baseline activity level is only three
events in a four-week period, this increase
represents an important change.
In addition, desire scores increased on
average by approximately 11 points, that means that
women went from "seldom" having desire closer to
"sometimes" having desire. And, most importantly,
patients' distress scores decreased 23 points with
treatment, which means went from being "often"
distressed about their lack of desire in sex, to
"sometimes" being distressed. And this represents
a clinically significant change for women.
It's also important to remember that these
are mean changes with treatment. Some women had
108
little benefit, while others had much greater
improvements in sexual activity and desire. As you
may recall from Dr. DeRogatis' presentation,
responders had greater increases in all aspects of
sexual desire, activity and distress measured in
this study. In addition, women who derive little
benefit from treatment are unlikely to continue
therapy.
[Slide.]
So what I see as a clinician, and what
we've shown you today, is that HSDD is an important
medical condition that has an important impact on
women's lives. Low-dose transdermal testosterone
treatment is a meaningful treatment option for
patients with HSDD, as it improves all aspects of
this disorder.
Intrinsa also is a much-needed option for
patients and physicians who currently have no
approved therapies with demonstrated safety and
efficacy with which to treat our patients.
Thank you.
The next speaker will be Dr. Michael
109
Steinbuch.
Phase IV Long-term Safety Plan
DR. STEINBUCH: Good morning. My name is
Michael Steinbuch, and I'm responsible for
pharmacovigilance and epidemiology at Procter &
Gamble.
Let me begin by saying that P&G is
committed to monitor the long-term safety of
Intrinsa. We believe we can best accomplish this
by conducting a prospectively designed
observational safety that will be comprehensive in
its design, and have the necessary data to detect a
possible safety signal faster than any other
method.
[Slide.]
As you've heard from Dr. Lucas, despite
extensive and systematic patient monitoring during
the Phase III program, there were no serious safety
signals that merit specific follow-up.
Nevertheless, we plan to monitor women exposed to
Intrinsa for longer periods of time.
[Slide.]
110
In addition to routine post-marking
surveillance and Phase III extension studies, we
considered three options to address the long-term
safety of Intrinsa: observational studies;
randomized controlled trials; and patient
registries. Each of these options has advantages
and disadvantages.
We carefully evaluated all options and now
believe that the best way to generate timely
long-term safety data is through a rigorous,
prospectively-designed observational study.
[Slide.]
Observational studies offer several
advantages. They provide a robust method for rapid
signal detection in a real-world setting; an
opportunity to study large numbers of patients' the
ability to adjust for potential confounding
variables; and the opportunity to evaluate all
patients that fill scrips for Intrinsa.
[Slide.]
The FDA reviewed an earlier version of our
proposed Phase IV observational study. The agency
111
raised a couple of issues regarding the potential
limitations of such a study. FDA questioned
whether data in women over 65 would be captured
when post-menopausal women treated with Intrinsa
shift to medicare; and, whether the study would
provide adequate power to detect an excess risk of
safety events of interest with women in this age
group.
However, in our clinical trials, where we
actively recruited women ages 20 to 70, the women
over 65 represented only 2 to 3 percent of the
study subjects. As you just heard from Dr.
Shifren, women are currently using various forms of
prescription testosterone. Among the 30,000 female
testosterone users in our proposed study database,
only 3 percent are over 65. Also, the
disenrollment rates are comparable, whether they
are younger or older than 65.
Partly in response to FDA's response, we
updated our study design. I'll show later that the
updated study will be powered to detect potential
safety signals.
112
Now, let's take a look at what we're
proposing.
[Slide.]
We plan to contract with Ingenix, the
research affiliate of United Healthcare, to conduct
this research. Ingenix LabRx is a large,
comprehensive insurance claims database. It
represents approximately 5 percent of the U.S.
population. It covers a range of health care
services, and includes hospital, physician,
pharmacy and laboratory data.
United Healthcare has an open formulary.
In this multi-tier system, virtually all
prescription drugs--including drugs for sexual
dysfunction--have at least partial reimbursement.
As a result, drug usage will be captured in the
database.
Upon IRB approval, Ingenix will validate
endpoints by identifying and reviewing all relevant
records. About 85 percent of those are available
for abstraction, and 90 percent of all medical
claims are processed within four months.
113
United Healthcare has a stable enrolled
population, with 85 percent per year. In addition,
Ingenix has an experienced research staff and a
proven track record with studies of this type.
[Slide.]
For example, the FDA has accepted numerous
observational studies using the Ingenix database as
a method to assess product safety. This database
has demonstrated utility across a range of drugs
and study endpoints in post-marketing safety
studies. These include allergic reactions and GI
outcomes, among others.
[Slide.]
The objective of our study is to compare
event rates in Intrinsa users versus non-users,
using a prospective cohort design with three-to-one
matching. Matching variables will be carefully
selected, and may include a propensity score, which
is a proxy for overall health status. Appropriate
matching minimizes biases between groups.
We propose the study be conducted for a
period of five years. All patients exposed to
114
Intrinsa will be included in the analyses. Of
note: there are no exclusion criteria. Endpoints
of interest will include both cardiovascular and
cancer events.
Let's take a closer look at the Ingenix
database.
[Slide.]
There are 10 million patients in the
database; of those 600,000 are menopausal, and
135,000 are menopausal women taking estrogen.
Based on estimates of HSDD disease prevalence from
the literature, we expect approximately 19,000
potential users of Intrinsa. Assuming 30 percent
of these women fill a prescription for Intrinsa,
there would be about 5,500 Intrinsa-treated
patients in the first year following launch.
[Slide.]
To provide estimates of the size of the
patient population and power to detect safety
signals over the course of the five-year
observation period, we needed to make certain
assumptions. For example, we have assumed a fixed
115
rate of 5,500 new patients per year; a .15 percent
event rate per year for cardiovascular events--as
was observed for the 50 to 59-year-olds in the WHI
study; a 50 percent discontinuation rate per year;
15 percent disenrollment per year; an alpha of .05,
and a one-sided test, which has more power to
detect a safety signal.
[Slide.]
Given these assumptions, if a major safety
signal emerged, we would be able to have 82 percent
power to detect a relative risk of 1.9 as early as
two years post-launch. AS person-years of
observation accrue over time, our ability to detect
smaller differences will increase.
[Slide.]
We recognize that observational research
encompasses a broad range of techniques, and varies
greatly in terms of their robustness. We've
proposed an approach that will maximize its value
as a signal detection method. This comprehensive
design will have a number of components not
typically found in observational research. These
116
include a collaborative protocol development with
external experts and FDA involvement and approval;
a blinded medical expert panel to adjudicate events
ascertained from medical record abstraction. And,
importantly, an independent safety review board
will be established with no P&G participation or
representation on the board. The independent board
will be responsible for identifying possible
issues, as well as analyze, interpret and report
results to FDA and P&G.
Initial data will be available at 18
months post-launch. Results and analyses will be
available for review about two years post-launch.
[Slide.]
In addition to observational studies, we
considered randomized controlled trials and patient
registries.
[Slide.]
Randomized controlled trials are idea for
testing hypotheses and determining cause and
effect. They offer the advantage of random
allocation of study subjects, which minimizes
117
confounding. However, RCTs such as large simple
trials are typically conducted in restricted
patient populations, based on a set of inclusion
and exclusion criteria.
The reality is, they may not reflect
actual use of the product in the marketplace. We
believe the major issues with a randomized safety
trial for Intrinsa are recruitment, retention and
adherence to treatment. In the WHI, for example,
the participants were presented with a potential
for a cardiovascular benefit, and yet 80 percent
refused randomization.
As you heard this morning, RCTs are very
large, and would require screening extremely large
numbers of women. In our clinical trials, we
enrolled 1,200 women from 100 sites in six months.
Based on this experience, an RCT would take several
years for patient enrollment, and five more years
for follow-up.
We believe these factors would be major
barriers to executing an RCT and detecting a signal
in a timely fashion.
118
[Slide.]
While patient registries have the
advantage of allowing the study of large numbers of
exposed patients in a real-world setting, there is
no practical method to identify a relevant
comparison group.
[Slide.]
Procter & Gamble Pharmaceuticals will
monitor the long-term safety of Intrinsa and will
continue working with the FDA and Ingenix to refine
the study plan.
Ingenix's robust infrastructure will be
instrumental in executing a successful
observational study.
Our approach is novel, in that we plan to
implement the study at launch. We will have input
from external experts to help us design the study,
and we'll have this independent safety review board
to execute, analyze and report study results to the
FDA.
We believe this is the best study design
for detecting a possible safety signal quickly, and
119
we're committed to making this happen.
Thank you for your attention, and I'll now
turn the podium over to Dr. Meyer.
Closing Remarks
DR. MEYER: For my wrap-up, what I'd like
to do with you is go over the questions that the
FDA posed to you this morning. Due to time
considerations, I'll be happy to answer any
questions on our plan to maximize the safe use of
Intrinsa with physicians and patients during the
discussion period this afternoon.
The first question: The first is an
efficacy question: "Do the efficacy data represent
clinically meaningful benefit?"
Yes. We assessed three related but
independent endpoints, all critical in HSDD.
They're concordant but they're not redundant. They
measure different aspects of the disease. These
efficacy assessments were patient-centered. The
patients told us the results were relevant. And
independent observers confirmed these results were
relevant. And the statistics told us these results
120
were relevant.
The primary endpoint and all the secondary
endpoints were highly statistically significant.
The results were consistent across studies and
across endpoints.
In addition, the randomized withdrawal
trial that we ran reinforced that the
pharmacological effect of the drug was better than
that of placebo.
Question 2: "Is the patient exposure
adequate to demonstrate long-term safety?"
Yes. As Dr. Lucas showed you our exposure
table, we have over 14,000 total patient-months of
exposure--and these data are from June, when we
were preparing our 90-day safety update for the
agency. We have an additional 180 women in the
surgical menopause program with 12 months of
exposure; an additional 100 women to include in the
18 months of exposure. We currently have 80 people
already in year three of the surgical menopause
extension.
Because of the study that we're finishing
121
up enrolling, looking at the patch in women not on
concomitant estrogen, one year from now we will
have an additional 2,500 patient years of exposure
to 300 mcg of Intrinsa.
Question 3: "Are the safety concerns or
unanswered questions that need to be studied?"
Well, as you are all aware, it is not
uncommon to have unanswered safety questions at
approval. As you think about this relative to
Intrinsa, it's important to keep in mind thus far
we have seen no significant safety signals. And
there's substantial experience already, in the real
world, with concomitant androgen and estrogen use.
Testosterone is not a new drug.
Importantly, we have committed to a
strong, independent post-marketing safety study to
be put in place at launch. Also we welcome the
opportunity to hear more ideas from the committee
and the agency about how to strengthen our labeling
to address safety issues.
Number 4--this is an easy one for you:
"Are the efficacy and safety data adequate to
122
support approval of the transdermal testosterone
system?"
Yes. We feel very strongly that the
efficacy and the safety data are adequate to
support approval of Intrinsa to provide women with
HSDD and their physicians this important treatment
option.
Thank you for your time.
DR. GIUDICE: I'd like to thank the sponsor
for their presentation this morning.
We're scheduled to take a break, but
before we do, for those who are going to
participate in the open public hearing, please be
sure that you have registered outside, and please
do this before the end of the break, otherwise you
will not be able to participate in the open public
hearing.
So we will take a break, and let us
return, please, at 10:15. Thank you.
[Off the record.]
DR. GIUDICE: Back on the record.
Please take your seats and we can continue
123
with the morning session. Thank you.
We will now continue with the FDA invited
speaker, Dr. Adrian Dobs, who's professor of
medicine from Johns Hopkins University, who will be
talking on safety of exogenous testosterone in
women.
FDA Invited Speaker
Safety of Exogenous Testosterone in Women
DR. DOBS: Good morning. My name is Adrian
Dobs, and I'm an endocrinologist at Johns Hopkins.
And when discussing the safety issues,
it's really important to think that we're not
talking only about testosterone, but also its
metabolites.
[Slide.]
Testosterone is reduced through
5-alpha-reductase to dihydrotestosterone, and it's
likely the DHT that has the effect on facial and
body hair, scalp hair loss, acne and hirsutism.
It could also act directly on muscle,
bone, causing virilization with clitoromegaly,
brain and sexual function. And then it can be
124
aromatized to estrogen, and estrogen is likely
acting directly on the breast and the uterus, and
on the bone, brain and libido.
[Slide.]
So, what I'd like to do today is go
through with you, as an outline, of what are the
safety concerns with testosterone administration.
A lot of this has been derived from a review
article we just published in the Mayo Clinic
Proceedings, which is in your packet for the
committee members.
So what I'd like to discuss is the
androgenic effects, cardiovascular
effects--particularly lipids, vascular reactivity,
glucose tolerance and hematopoietic. Then I'd like
to discuss the endometrial and breast effects, and
finally try to come up with some kinds of
recommendations.
[Slide.]
When discussing the androgenic effects,
the main three androgenic effects is that of acne,
hirsutism, and virilization. The usual clinical
125
presentation of this is acne, increased hair
growth, clitoromegaly, temple baldness and lowering
of the voice. Testosterone is involved in
thickening of the vocal chords, that's why men have
lower voices than women.
In general, one would say that at higher
doses these are extremely common. At low dose of
androgen replacement, it's probably more rare or
mild. It has been shown as a mild effect--of acne
and hirsutism--in some of the studies with
methyltestosterone.
In general, I would make the statement
that it is dose and duration dependent, and most of
these effects are reversible. The "most" would
refer to temple balding. It's unclear if that will
be fully reversible, or the clitoromegaly may take
years to see any kind of resolution. But the acne
and the hirsutism in a period of months will
probably resolve.
[Slide.]
This is a study looking at hirsutism
scores with a methyltestosterone done by Elizabeth
126
Barrett-Connor. And she looked at varying
combinations of estrogens with testosterone.
Here's androgen, a low dose. Here's androgens at a
high dose. And in the light orange color is the
percentage of subjects who claimed that they got
worse by taking a high dose of androgens.
So you see there's a trend there that
actually was not statistically significant. So
that hirsutism and acne would not be surprising
observations in women taking testosterone.
[Slide.]
Now, a large concern is obviously
cardiovascular effects, and here there really is a
great deal of question.
First, to begin with, this is an example
of a study looking at the relationship of
endogenous hormones to cardiovascular risk. And
this is looking at the odds ratio of developing a
cardiovascular event, compared to the four
quartiles of hormones. So here is testosterone in
green, adjusted for risk and the free androgen
index. And what you could see here is that
127
individuals who are in the fourth quarter--the
highest quartile--of androgens have an increased
risk of developing some kind of cardiovascular
event. And this is data from the Women's Health
Study.
So there is this interesting relationship
between endogenous hormones.
So what is the specifics of this?
[Slide.]
Well, there have been a few studies
looking at testosterone in women. Generally, with
all androgens, there's a clear significant
reduction in HDL cholesterol. It's likely neutral
on LDL, and it does result in a lower of
triglycerides. The reason for this is probably the
first pass through the liver, and the changes in
hepatic lipase.
The data from women state that it's very
dependent on the route of administration--that is
oral testosterone will have a greater effect than
transdermal, and also the type of
testosterone--whether or not its methylated or an
128
anabolic steroid.
In general, aromatizable androgens--that
is, testosterone that's converted to estrogen--have
a neutral effect.
[Slide.]
This is some of our data looking at
changes in lipids with a methyltestosterone, and as
you can see, there's a pretty impressive decline in
the HDL cholesterol, as there is a decline in the
triglycerides.
[Slide.]
However looking at Shifren's data, there
was really very little change across all the lipid
parameters when used in a transdermal preparation.
[Slide.]
Now there's also questions about vascular
reactivity. This has to do with endothelial
dysfunction. It has been found to be an earlier
marker of cardiovascular disease; that is the
stiffness may predict the development of
cardiovascular disease. This has been studied in
small non-invasive publications, looking at
129
flow-mediated vasodilatation. And it turns out
that epidemiologically, there is a decline in this
flow-mediated dilatation in women as they go
through the menopause.
The best way to study this is by looking
at brachial artery reactivity. And this has been
done by looking at endothelial independent
dilatation, which is a glyceryl trinitrate induced
kind of a procedure.
[Slide.]
So this is just one example of a study,
looking at the effects of testosterone on vascular
reactivity in women. This is flow-mediated
dilatation before and six weeks after testosterone,
to show that there was increased flow. This is the
control group, and this is women that were given a
type of nitroglycerin that would vasodilate. And
there is a statistical increase in vasoreactivity.
There's some mixed data on this. I would
probably make the statement that testosterone has a
very minimal effect on vascular reactivity, and
perhaps it could be beneficial.
130
[Slide.]
Plasma viscosity is of great concern when
discussing cardiovascular effects. We know from
epidemiological studies that increased plasma
viscosity is a risk factor for cardiovascular
disease, and does predict coronary artery disease
development. The physiology here is not very
clear. It seems to be affected by fibrinogen and
triglycerides. And the only thing I could say is
there's just been a few small studies evaluating
this, and in one study they did show that there's
actually improvement in viscosity when women were
given testosterone.
[Slide.]
When it comes to hematopoietic factors,
we've known for many years that testosterone is
involved with erythropoiesis. It, in men, can be
associated with polycythemia, and polycythemia is a
risk factor for cardiovascular disease.
In looking at some large epidemiological
studies with increasing hematocrit, there is an
increase in cardiovascular risk.
131
The mechanism for this is that
testosterone is involved in stimulating production
of erythropoietin, and also in erythroid colony
units. In men, it's very clear that testosterone
can cause erythrocytosis. This is dose-related.
It's area-under-the-curve-related, so that men who
have been given injectable testosterone are much
more likely to be found to have increase
hematocrits. In men given transdermal types of
testosterone, there's essentially very little
problem.
In women, few studies have been done. In
a study of 22 young women using a testosterone
implant there was no effect on clotting factors.
And you heard some data earlier this morning about
the new transdermal compound.
Essentially, I could not find any reports
of true polycythemia in women.
[Slide.]
Now, glucose metabolism is clearly a risk
factor for cardiovascular disease, but for
hyperglycemia and hyperinsulinism. There's very
132
little data here. There seems to be no evidence of
changes in fasting glucose or in insulin
sensitivity, although I would say there's little
data in both men and women.
I think the big caveat to this is what
we're learning more and more about, which is that
of polycystic ovarian disease and metabolic
syndrome.
[Slide.]
This is just some data from Shifren
looking at glucose and insulin across groups. And
there was no difference. This is just looking at
fasting studies.
The big problem and the big question here
is that there seems to be a fairly consistent
relationship between endogenous testosterone and
cardiovascular risk. When we're talking about
polycystic ovarian syndromes or metabolic syndrome,
metabolic syndrome is being increasingly recognized
as a risk factor for the development of
cardiovascular disease. In this situation, it's
noted to have obesity, hyperinsulinism,
133
hyperandrogenism, and hyperlipidemia.
The mechanism for the hyperandrogenism in
this complex of metabolic syndrome and PCO is not
very clear. I've just written down one postulate,
and that is the hyperinsulinism stimulates
testosterone production from the ovary, and this
works to decrease SHBG, and that works to increase
free testosterone.
So I think of challenges that are ahead is
really to determine what is the relationship here
of testosterone, metabolic syndrom, PCO, and how to
put this into context. Is it testosterone per se
that's having a cardiovascular risk, or is it its
metabolism to estradiol?
[Slide.]
I'd like to next talk about the potential
side effect of endometrial or breast effects.
Again, from endogenous hormone levels in
epidemiological studies, there seems to be a
relationship between hormones and the development
of disease. So this is looking at the odds ratio of
endometrial cancer, by quartiles of steroid
134
hormones in post-menopausal women. And the green
is estradiol, and blue is testosterone. And you
see here that women with increasing doses of
endogenous hormones will have an increased risk of
developing endometrial cancer.
[Slide.]
There's been several reports of
hyperplasia and cancer with the use of high doses
of testosterone. Most of this data comes from
women who were given high doses of testosterone for
transsexualism. This may not be applicable in this
particular discussion. But clearly in that
population when the serum testosterone level gets
into the level of a male level, that testosterone
will be aromatized to estrogen and run the risk of
unopposed estrogen and the endometrium.
With low doses there is essentially no
cancers that have been reported. There has been,
in one study, endometrial hyperplasia that was seen
in one of 107 women given methyltestosterone. But
it was also seen in one of 111 women given
estrogen. So there clearly will be this issue of
135
unopposed estrogen in women who have a uterus and
might not be taking progesterone.
In one study there was a 6 percent
incidence of cystic endometrial hyperplasia. And
in another study, though, they did study vaginal
cytologies. And these were all stable throughout
the course of the study.
[Slide.]
Breast cancer risk is something that needs
to be discussed. Epidemiologically it's similar to
the endometrial data in that there seems to be a
relationship between high endogenous testosterone
and breast cancer.
With hyperandrogenism itself, it seems to
be related to the association with metastasis. The
physiology here, it's been postulated to be due to
the fact that there is androgen receptors that are
found in 50 to 90 percent of breast tumors. This
testosterone therefore may act directly to
stimulate breast epithelium, or it may be
aromatized and therefore it's the estrogen that's
acting on the breast tissue.
136
In women there have been no reports of
breast cancer that came from exogenous testosterone
treatment. We did hear this morning some cases of
active and placebo-treated women.
[Slide.]
now, there are some other possible
effects, but these are really very mild, but I'll
just mention them in passing.
In men given alkylated androgens there's
been associated hepatotoxicity and hepatic
adenomas. This has not been the case with
transdermal testosterones. In women, there's
essentially no evidence of abnormal liver
functions, either with pellets or transdermal.
There's a theoretical risk of some fluid
retention with testosterone, but really not in the
doses used even in men, and certainly no one would
expect it in the doses for women.
[Slide.]
There's a great deal of question about
anger and hostility. This is a very difficult
parameter to measure. Some studies have asked
137
questions like "Do you have an interest in smashing
things?"
[Laughter.]
So it's really a very tough parameter to
get one's hands around.
And the physiological explanation would be
that there are certainly androgens and estrogen
receptors in the brain. In men, there's some
questionable data--I mean "questionable" because of
the study design--to say that there may be a
relationship between testosterone and violent
behaviors. I really question this data. In the
clinical trials done in men, even given high doses
of testosterone, it's been very hard to elucidate
whether or not there's any relationship here.
And in women there was one study that
showed that there was some increase in hostility
scores in women given high doses. I think, in
general, this is not a clinical problem. We
generally live in a fairly controlled society.
[Laughter.]
So, I'd like to end up with talking about
138
some recommendations.
[Slide.]
I would say there are some absolute
contraindications for the use of testosterone, and
this would certainly include pregnancy and
lactation, endometrial cancer or any unexplained
vaginal bleeding, and breast cancer.
There might be some relative
contraindications, as well. And that would be
moderate to severe acne or hirsutism, androgenic
alopecia, sever insulin resistance, and anyone with
an anger management disorder.
[Laughter.]
[Slide.]
These are some recommendations.
Obviously, there will be lots of discussions about
this if and when things go further. But these are
some monitoring that I would recommend, in that
acne and hirsutism should be evaluated at each
visit. Virulization should be evaluated at each
visit. Anger and hostility can be asked at each
visit.
139
Breast exams and serial mammograms should
be done conscientiously--likely following strictly
the recommendations, and be done annually.
It's important to ensure that women
treated with testosterone should have regular
OB/GYN exams, and if there's vaginal bleeding this
should be discontinued. Obviously, this is the
assumption that if a women has an intact uterus and
is taking testosterone.
Hematocrits should be evaluated, I think,
after the first three months and then, likely,
annually.
Serum lipids--because transdermal has very
little effect on serum lipids--can really be done
as indicated.
And measuring of serum total testosterone
I think is extremely important to ensure that the
levels do not get very high since, as I've stated
before, the side effect profile for testosterone is
related to the type of testosterone, the route of
administration, the dose used, and the length of
time that the women is being treated.
140
I might recommend that women be evaluated
at six weeks, and then I'd put question marks about
whether or not this needs to be done every six
months thereafter.
[Slide.]
So I think there are several remaining
questions about the safety of testosterone therapy.
With more women being treated for longer periods of
time there will be questions about the long-term
effects on androgenic signs and symptoms.
Conceptually, there is still a lot of
question about the overlap of endogenous
testosterone versus exogenous treatment and
cardiovascular risk, and how this relates to such
things as abdominal obesity in the metabolic
syndrome.
The other remaining questions have to do
with the use of testosterone alone versus being
done in combination with estrogen and progestins.
And really my last point that I think is of
importance is better evaluation for the risk of
breast and uterine tissue.
141
Thank you very much.
DR. GIUDICE: Thank you, Dr. Dobs.
I'd like to point out to the committee
that this is now an opportunity to ask Dr. Dobs any
questions, because there will not be an opportunity
later.
Dr. Emerson?
DR. EMERSON: Have there been any studies
that looked at the relative balance of estrogen
versus testosterone and whether that's predictive?
Or has it always just looked at testosterone levels
versus estrogen levels?
DR. DOBS: They've basically been looked at
separately, although various combinations of free
androgen index, free testosterones. And the issue
of measuring free testosterone in women is a
difficult one. It's difficult enough in men, but
the assay and the level of detection for women can
be quite problematic.
So--no, I'm not aware of ratio differences
as much as the absolute numbers.
DR. GIUDICE: Yes.
142
DR. NISSEN: I wonder how much is known
about triple therapy--that is, estrogen,
progesterone and testosterone. You know, whenever
a drug gets out on the market there tends to be
off-label use, etcetera. So we need to have some
understanding about whether anything is known about
that?
DR. DOBS: Well, triple therapy is what had
been used in the Intrinsa data, in that most--in
one of your studies--didn't you have--right--where
they did use progesterone in women that had a
uterus on board--sorry. That's a terrible way of
phrasing that.
[Laughter.]
I meant estrogen on board--then being
given progesterone. And the data was about the
same there.
So there is a small study in which there's
triple therapy--which would have to be recommended
if a woman has her uterus and is being given
estrogen and being given testosterone. She would
have to take progestin.
143
DR. GIUDICE: Dr. Rice, and then Dr.
Lockwood.
DR. MONTGOMERY-RICE: I raise a question--a
concern--about the duration effect. I think what
they share with us is that we're looking at low
doses of testosterone. However, we know that when
we look at our low-dose trials of
estrogen---particularly, I'm thinking the Hope
trial, when we looked at .3 mg of CEE, and that
first year we saw minimum cases of endometrial
hyperplasia. And in the second year we saw more
cases of endometrial hyperplasia, even though we
didn't see an increase in estrogen levels.
And so should we be concerned here about a
duration effect, even in the presence of these low
doses of testosterone?
DR. DOBS: Well, there's really no good
data for this, because the only long-term
treatments of testosterone have been with higher
doses, and mainly in transsexuals. Goren, in
Amsterdam has very nice data on uterine hyperplasia
in that population.
144
So it's really hard to say what will be
the long-term effect. There doesn't appear to be a
cumulation of testosterone in the skin, as an
example. But I think the issue of dosing is
extremely important, and monitoring of doses.
I mean, I treat a lot of men with
hypogonadism and testosterone, and their levels can
be all over the place when given any kind of
transdermal testosterone--whether it be patches or
gels, there's a great amount of variability. And I
think that might be the case here, when
testosterone is going to be used in larger numbers
of women is: what is the likelihood that the women
will get to testosterone levels above the normal
range?
So that's why I feel strongly that does
need to be monitored carefully.
But if we're talking about greater then--I
think they have two-year data--there's really very
little that's out there to suggest there would be a
problem on safety.
DR. GIUDICE: Dr. Lockwood.
145
DR. LOCKWOOD: I have a comment--and I
apologize, Dr. Dobs, for having my back to you--
DR. DOBS: Yes, where are--oh, I see.
Okay.
DR. LOCKWOOD: I can't twist my head around
180 degrees.
My comment is that we've also looked at
the endometria of women that have been exposed to
high doses of testosterone in preparation for
transsexual surgery. And the marked effect that
we've observed is decidualization, which suggests
that the predominant effect is actually more
pre-gestational than estrogenic.
So, you know, I'm not sure I would be
convinced one way or the other about the risk of
endometrial cancer--particular at very high doses
of testosterone. But there may be individual
variations in that response, depending on the level
of aromatization and so forth.
My question to you is that there is a
fairly, now, long experience with the use of
danazol in women with mastalgia and fibrocystic
146
disease, etcetera, and the question--and I don't
know whether you have the answer to this or
whether, in fact, anyone does--but since that's a
natural group to look at in terms of the risk of
androgen-induced breast cancer, is there any
evidence that such therapy is associated with a
higher rate breast cancer?
DR. DOBS: That's an interesting question,
and I've never heard of a case of breast cancer
with the use of danazol. But I don't know about
breast biopsies or any intermediate changes that
might occur. I think that's an interesting
question.
DR. GIUDICE: We have time for one more
question.
Dr. Stanford?
DR. STANFORD: I was just wondering if you
are aware of any data on androgen levels or insulin
levels in women with PCOS who have undergone
oophorectomy?
DR. DOBS: Well, I could--no, I was going
to answer that--there are certainly a few taken
147
with PCOS, and they lose weight, or be given a
medication to affect this, the insulin levels, they
will drop their testosterone levels. So this could
be modulated.
The oophorectomies wouldn't be done.
Years ago, when it used to be called
Stein-Leventhal syndrome, and there was wedge
resections performed at that time, the testosterone
levels did drop. And that's probably why they were
able to get pregnant, and why it worked, is they
were taking out a mass of the ovary, and that
resulted in normal hormones and ovulation.
DR. GIUDICE: Thank you very much.
Going on now with the FDA presentation,
our first speaker is Dr. Daniel Davis who is a
medical officer in the Division of Reproductive and
Urologic Drugs, and he will be talking on efficacy
findings and issues.
FDA Presentation
Efficacy Findings and Issues
DR. DAVIS: I'm Dan Davis, and I'm one of
two primary reviewers for this application. I
148
reviewed the efficacy data, and Dr. Lisa Soule
reviewed the safety data.
My talk this morning will focus on the
efficacy data and the efficacy issues.
[Slide.]
There are three important points that are
listed here on this slide. First is to note that
this is the first application the FDA has ever
received for a female sexual dysfunction
indication. The division did have our draft
guidance for female sexual dysfunction, which was
written in May of 2000 for helping to evaluated
this application for hypoactive sexual desire
disorder.
The second point is the issue of the
relatively small treatment effect seen with
testosterone treatment. The division agrees with
the sponsor's analyses that the endpoint changes
associated with testosterone were statistically
significant compared to the placebo effect. But
the key issue for us is really the clinical
significant of the findings.
149
The primary endpoint, as already stated,
was a change in satisfactory sexual events--noted
in my presentation as "SSE." And the two secondary
endpoints were sexual desire and personal distress.
Small mean changes were noted in all three
endpoints. And relative to the testosterone
treatment--we'll note it as the "TTS" in this
presentation--there was a strong placebo effect
that persisted throughout the two six-month blinded
trials.
The third point regards the findings of
the applicant's study for determining the minimal
meaningful clinical change in the endpoints that
will be discussed briefly a little later in the
presentation.
[Slide.]
The FDA Draft Guidance for sexual
dysfunction, and the Division's advice were very
closely followed by Procter & Gamble. As noted
earlier, in the applicant's presentation, they
developed three instruments across different
cultures and languages to assess three different
150
efficacy endpoints, namely: satisfactory sexual
events, desire and distress that was associated
with HSDD. The two placebo-controlled Phase III
trials of six months' duration were completed, and
each with over 500 subjects. And a clinical study
to determine the magnitude of change in these three
endpoints that would be clinically meaningful to
the individual woman herself was performed.
[Slide.]
The key inclusion and exclusion criteria
are summarized here. All of the women were
surgically menopausal and on stable doses of
estrogen. As noted earlier, approximately 77
percent of the women were on oral estrogen, and 23
percent on transdermal.
The diagnosis of acquired HSDD was made
primarily by answering "yes" to five questions that
are listed on the next slide.
All of the women were in good general
health, and did not have major medical or
psychiatric illnesses. And there were no specific
serum testosterone criteria for inclusion or
151
exclusion.
[Slide.]
The five questions are listed here. I'm
not going to read them, but they were used
primarily for the diagnosis of acquired HSDD. The
women themselves--as opposed to a clinician or a
sex therapist--answered the questions, so that this
was purely determined by the individual women. In
essence, the subjects has a satisfying sex life
before surgery, followed by a decrease in desire
and activity after surgery, that cause them
personal distress and a wish to have an increase in
their sexual desire and activity.
Data from the Sexual Activity Log--which
is abbreviated as SAL in some of the applicant's
presentations--and the baseline scores on the two
instruments for measuring desire and distress were
not part of the entry criteria for the study.
[Slide.]
The primary endpoint was, as noted before,
the change from baseline in satisfactory sexual
events per four weeks. The individual's sexual
152
activity was recorded retrospectively each week on
the Sexual Activity Log, and was collected at the
clinic sites every month.
The secondary endpoints were the mean
change from baseline in the Personal Distress
score, and the mean change in the sexual desire
score. The two instruments that were used to
measure these endpoints had questions about how the
individual felt over the previous 30 days. Each
instrument was completed at baseline, and weeks 4,
8, 12 and 24 of the Phase III trials.
The answers then, from both of these
instruments, were normalized to a scale of 0 to 100
points. And I think it's extremely important to
remember that. So we're talking about a scale of 0
to 100 points.
A decrease in the distress score meant
less distress, and an increase in the sexual desire
score meant an increase in sexual desire.
[Slide.]
This slide summarizes the findings for the
primary endpoint of satisfactory sexual events. On
153
average, the baseline SSEs for all subjects in the
two trials was three satisfying events--which is
shown, really, in this column: an average of about
three events. The range is noted here.
A the end of the treatment period the TTS
group increased by approximately two SSEs--that
would be change, the change of 2 in basically the
1.6, compared to the placebo group, which had a
change of the 1 and .7 events per four weeks.
The difference was shown by the applicant
to be statistically significant. The clinical
significance, however, of this mean increase from
three to five events, and the difference of one
event, placebo, compared to testosterone treatment
is not clear to the Division.
[Slide.]
The secondary endpoints, the mean change
from baseline in personal distress is summarized on
this slide for the two trials. On a scale of 0 to
100, the mean baseline score was
65--approximately--here for all participants. And
this corresponds, on average, to an answer of
154
"often" to the questions related to distress.
A decrease of 16 to 18 points was seen
with placebo--that is here, 16 and 18 points--and a
decrease of about 24 points was seen with
testosterone treatment. The difference between the
testosterone response compared to the placebo
response in both trials, with 6 to 7 points on a
scale of 100. So here's the difference of
testosterone, compared to placebo.
We are not sure of the clinical
significance of this change from baseline, and the
relative difference between the placebo effect and
testosterone effect on personal distress.
[Slide.]
The same concern is seen in the next
slide, with the other secondary endpoint, for
sexual desire.
The overall mean score was 21--mean
baseline score, yes, was approximately 21 points on
this 100-point scale. The placebo group increased
an average of 6 to 7 points, while the testosterone
group increased 11 to 12 points.
155
The difference between testosterone
treatment and placebo is approximately 5 points in
the first study, and 5.2 in the second study. And
this is on a scale of 100.
Once again, the clinical significance of
this small numeric change is the issue.
[Slide.]
The next slide simply shows a summary of
the overall events. It just shows, for SSEs, just
your change, placebo from baseline. And I really
don't need to make any more comments, except that
it's showing that there is one more satisfactory
sexual even per four weeks; five more points on the
desire scale of 100 points, and six to seven point
greater decrease in the distress scale of 100
points--when we compare the placebo effect with the
testosterone effect.
[Slide.]
To put these findings into perspective, I
will now show data collected by the applicant for
age-matched normal women with no female sexual
dysfunction, and with a normal sexual desire.
156
This slide comes from basic data collected
by Procter & Gamble early in their development
program, at the time of the initial validation of
their endpoint instruments. Data was collected
from over 300 women with HSDD, and was compared to
over 250 "normal" age-matched controls.
[Slide.]
The data presented here shows the
"normals" for 146 U.S. women, and it's shown in the
blue bars. So we have a baseline of SSEs of 12
events per four weeks; for desire, the baseline of
"normal" women is 65 for desire; and for distress,
it's 5 points on a scale of 100.
The Phase III data of the combined trials
of all the women is also shown on this slide. So,
for SSEs our baseline was 3, and with testosterone
treatment increased to 5 for the SSEs. For desire,
our baseline was 21 points and increased to 33,
with a normal of 65 and a baseline here for
distress was 65, and then that decreased to 41.
Again, the normal is at 5.
Although there was a clear treatment
157
effect with both placebo and testosterone in the
two Phase III trials, we can easily see that the
testosterone effect over the six months of
treatment did not return to the values of the
normal age-matched women, as determined by the
applicant.
[Slide.]
This entire study was covered very nicely
by Dr. DeRogatis earlier. It summarizes the study
performed by the applicant to determine a
meaningful change in the endpoints as defined by
the women themselves.
All of the 132 interviews were done within
two weeks of stopping treatment. But the subject
and the interviewer were blinded to the actual
treatment received during the treatment, and the
key question--I won't read it again, because
DeRogatis showed it to you--but "--did you have a
meaningful benefit from the study patches?"
Of the women receiving TTS treatment, 52
percent felt they had a meaningful benefit, while
31 percent of the women on placebo felt they had a
158
meaningful benefit.
The Receiver Operating Characteristics
analysis was then used to determine those changes
that best separated the group of women that felt
they had a meaningful benefit from those who felt
there was no benefit.
And the final results on that analysis
showed that these were the minimal meaningful
treatment changes, namely: greater or equal to 1.1
sexually satisfying even per four weeks; a desire
score change of 8.9 or greater on the scale of 100;
and a distress scale decrease of 20 or more in the
distress score of 0 to 100.
The main importance, however, of this
study was so that then these values could be used
to perform a series of responder analyses, as shown
on the next slide.
[Slide.]
I'm really just going to focus on the
first set of bars here, because this was the
primary endpoint for the study; that is, the
satisfactory sexual events. And the responder
159
analysis--and this is for all of the subjects in
the Phase III studies combined--so this is over
1,000 women--showed that the mean responders for
placebo was 30 percent of the placebo women,
compared to 44 percent of the testosterone women.
If the parameters are changed to greater
than 2 SSEs or 3, we see a slight lowering of the
bars. But what is important is that the difference
between placebo and testosterone treatment range
from the 12 to 14 percent; and specifically, this
is a 14 percent difference here.
[Slide.]
In summary, the clinical efficacy findings
show a small but statistically significant
testosterone treatment effect seen in the three
efficacy endpoints. There was a mean increase in
the TTS users of one more satisfactory sexual event
per four weeks, compared to the placebo response
for four weeks.
For the secondary endpoints, the distress
score decreased 6 to 7 points more with
testosterone, compared to the placebo effect, on a
160
100-point scale. And the desire score showed a
difference of 5.1 units, comparing placebo to the
testosterone treatment.
For the responder analysis of the primary
endpoint "satisfactory events," there was a 14
percent difference in the number of events per four
weeks for testosterone treatment compared to
placebo.
These changes in the testosterone
treatment do not approach the normal values seen in
the age-matched women without hypoactive sexual
desire disorder as determined by Procter & Gamble
and shown in slide number 11.
We look forward to the Advisory
Committee's input concerning the clinical
significance of these efficacy findings.
This concludes my remarks. And next we'll
hear from Dr. Soule on the safety findings.
Safety Findings and Issues
DR. SOULE: Good morning. I'm Lisa Soule.
I'd like to highlight some points raised in our
safety review of the two Phase II and two Phase III
161
studies on transdermal testosterone in surgically
menopausal women.
[Slide.]
High level concerns about safety of this
product are two-fold. The adverse effects of
long-term or chronic use of this product cannot be
characterized from the current safety database.
Events of potential concern may have a long latency
from exposure to occurrence.
Also, the addition of testosterone to
estrogen may increase risks of estrogen-associated
adverse events such as breast cancer and
cardiovascular disease--as is seen when
progesterone is combined with estrogen. And it's
worth noting again that the target
population--surgically menopausal--will be
concurrently using estrogen, a product with known
risks, and may potentially use it on a long-term
basis.
Current guidelines on the use of estrogen
products are discussed on the next slide.
[Slide.]
162
Recommendations from the FDA and
professional societies concerning the use of
hormone therapy have changed significantly since
publication of the Women's Health Initiative
findings. And you can see the FDA boxed warning
recommends that estrogen and E+P products be
prescribed at the lowest effective doses, and for
the duration, consistent with treatment goals and
risks for the individual woman. And, similarly,
the American College of OB/GYNs says the lowest
effective estrogen dose should be used for the
shortest possible time to alleviate symptoms, and
further recommends that use should be reassessed
annually.
[Slide.]
Review of the literature on testosterone
use in women suggests a spectrum of risks. Some
are documented to occur with some certainty at
doses like the TTS patch, such as androgenic
adverse effects: acne, alopecia and hirsutism.
Others are not clearly or consistently found, but
are suggested from data on women with endogenous
163
hyperandrogenism, such as changes in lipids,
hypertension. And, finally, there are very limited
data addressing the outcomes of greatest
significance, such as increased cardiovascular
morbidity and breast cancer.
[Slide.]
In evaluating the safety database from
these studies, we must consider the extent to which
results obtained from the study population can be
generalized to the target population for this
product. For example, African-Americans make up 13
percent of the U.S. population, but only 6 percent
of the study population. We do know that
African-American women are more commonly surgically
menopausal than Caucasian women, but the prevalence
of HSDD in African-American women is not known.
In addition, there were small numbers of
older women in the trial. And, finally, women who
may be at the highest risk for cardiovascular
morbidity were either under represented--such as
African-Americans and older women--or were
completely excluded from these studies, as is the
164
case for women with existing diabetes or
cardiovascular disease.
[Slide.]
In reviewing the safety database for TTS,
the following safety endpoints were of particular
interest, because they bear on specific areas that
are potentially of concern in considering chronic
use of testosterone in women. These include
adverse events related to the use of TTS and to the
duration of exposure to TTS; laboratory data that
may reflect cardiac risk; changes in blood pressure
and weight; and occurrence of breast cancer.
Before discussing safety outcome data, I
want to show you some data on testosterone levels
that were obtained in the women in these studies.
Although, as you heard earlier, mean testosterone
levels remained within the range for
reproductive-age women, a significant proportion of
the treated subjects had level of free and
bioavailable testosterone above this range.
[Slide.]
This slide shows the proportion of
165
subjects with free testosterone values outside the
upper limit of normal range, which is 7.3 pcg per
ml for reproductive-aged women, aged 18 to 49. The
percent shown here are for placebo subjects--shown
in red here--and women on TTS for varying
durations.
The green bars show subjects who were on
TTS throughout the duration of the study, and had
blood sampling done at the noted weeks here.
The purple bars are the subjects who were
initially on placebo back here in the early part of
the study and then switched to TTS. So at these
points all of these women are on TTS.
It can be seen that virtually no placebo
subjects had levels outside the reproductive-age
range. At weeks 52 and 78, there was very little
difference between groups, according to their
duration of TTS. But you can see that almost a
quarter of the women on TTS for longer than a year
developed free T levels beyond the range for
reproductive-aged women.
[Slide.]
166
And here's the similar data for
bioavailable testosterone which, as you've heard,
comprises both free and albumin-bound testosterone.
And, again, virtually no placebo subjects are
beyond the reference range.
And within subjects on testosterone,
there's not a great difference according to the
duration of testosterone. But here, up to 43
percent of women who used TTS for up to a year
developed bioavailable T beyond the reference
range.
[Slide.]
As you see, the trial design is complex,
with different arms receiving different durations
of testosterone exposure. So before I present any
more data, I'd like to try to clarify how the
subsequent data will be presented.
This figure shows the movement of subjects
through the different phases of the trials, and the
TTS exposure that each group received at the
various phases. And what these "durations" here
are. The N's in each box, shown here, are the
167
number of women who entered that phase.
The data presented from here on will
primarily compare even frequency between women on
placebo and women who received between 0 to 6, 6 to
12, and 12 to 18 months of testosterone. These
exposures to TTS did not necessarily occur in the
same phase of the trial, as the exposure of women
who were initially randomized to placebo--this line
here--and then went on to TTS always lagged six
months behind those who were initially randomized
to TTS.
The events to be reported on each slide
from here on occurred in the study phase
corresponding to the exposure interval for each
group. So, for example, events reported for TTS
subjects with up to six-months exposure occurred
here in the double-blind phase, while events for
the placebo-to-TTS subjects with the same amount of
exposure occurred over here, in the open-label
phase.
[Slide.]
Some adverse events show an increased
168
frequency in women who received TTS compared to
subjects on placebo, including androgenic adverse
events overall, as well as the individual
components: acne, hirsutism, alopecia and voice
deepening. The percents shown here are the percent
of subjects at each of the three phases of the
study who experienced androgenic events. They are
grouped by the total duration of TTS exposure that
the subjects received.
Androgenic adverse events overall occur
more frequently in TTS-exposed subjects than
placebo subjects, but don't show a steady increase
with increasing duration of exposure. Some adverse
events did occur with increased frequency, though,
as duration increased.
[Slide.]
For example, alopecia. And here you can
see, starting with placebos and no exposure, that
the rate of these events increases as the women are
on longer and long durations. And, remember, these
events are not cumulative but, rather, show the
occurrence of new cases of alopecia in each of the
169
three phases. And each phase, remember, is about
six months long.
Since the associate of androgenic adverse
events with the use of TTS seems clear, the
relationship between free T levels and the
occurrence of these adverse events was explored.
[Slide.]
The incidence of androgenic adverse events
are examined here according to quartiles of free
testosterone obtained during the double-blind
phase. And the T values used here are the maximal
values obtained at either the week-12 or the
week-24 sampling.
There appears to be an association between
higher levels of free T and greater frequency of
acne and hirsutism. And, as you hear this morning,
there was only a significant trend test here for
hirsutism.
But if you look at acne, you can see that
the women in the upper two quartiles--that is,
women above the median--do appear to have higher
rates than placebo women or women down here in the
170
lower two quartiles of free T.
In hirsutism, you can see more of an
exposure response, which may plateau up about the
third quartile. And if you look at the ranges of
the third quartile, you can see that these are
still below the reference range for
reproductive-aged women.
[Slide.]
In contrast, alopecia and voice deepening
do occur more frequently among TTS than placebo
subjects, but don't seem to increase with higher
free T levels. And it may be that these events
occur where a threshold free T level is exceeded.
[Slide.]
Getting back to our risk spectrum, I want
to review some data bearing on possible impact of
TTS on cardiac risk factors. We're concerned about
the potential impact of TTS on a number of cardiac
risk factors and, as you've heard, several of these
are believed to be linked in a common
pathophysiologic process, and form a constellation
known as the "metabolic syndrome"--which is an
171
independent risk factor for cardiovascular disease.
There are several diagnostic schemes, but
generally the components include glucose
intolerance, dyslipidemia, hypertension, and
central obesity.
A community-based study of 16,000 mean and
women found that the free androgen index was
statistically significantly higher in women with
metabolic syndrome, and also found that the
prevalence of metabolic syndrome in
African-American women was twice that in Caucasian
women.
Let's look at the lipid data first.
[Slide.]
While, on average, lipid values showed
little mean or median change, over the course of
the studies some parameters were of concern in
terms of percent of subjects who developed values
outside the normal reference range. This slide
demonstrates the proportion of subjects with
abnormal values on LDLs and triglycerides. These
labs were measured in all three phases of the
172
trials. The placebo group is in red.
Abnormal values were defined at the levels
shown here. So, for LDL, above 160 mg/dL; for
triglycerides, about 250 mg/dL--but, additionally,
required that subjects have an increase from
baseline of greater than 30 percent. So women who
entered the trial with elevated lipids may not even
be represented here. And thinking back to the
metabolic syndrome criteria that you just saw, 3 to
4 percent of these subjects may meet the
triglyceride criteria for the dyslipidemia
component.
Virtually no subjects developed abnormal
HDL levels in these trials, and the mean and median
changes were neutral, or even showed a slight
increase from baseline for HDL.
[Slide.]
Although, as you heard, mean glucose
levels were similar between treatment arms during
the double-blind phase, the change from baseline in
glucose level appears to increase with duration of
TTS exposure. And here you can follow the trend in
173
the groups, characterized by their initial
randomization. So here are women on placebo at the
beginning, and you can see that they have a slight
drop in their glucose levels. But as they go on to
six months and then 12 months of TTS, you can see
that the glucose levels are rising.
Similarly, here are subjects receiving TTS
in the beginning of the trial and, again, as their
exposure increases, so too does their glucose
level.
[Slide.]
Insulin was measured only during the
double-blind and open-label phases, and not in the
extension phase. The mean increases in the
TTS-exposed subjects at both time periods--and
again, remember, both of these, even though we have
red coloring, are on TTS in the
open-label--exceeded those down here in the placebo
subjects.
Although the changes in these markers of
carbohydrate metabolism are small, these small
trends may be magnified when the full target
174
population, including who may already be glucose
intolerant or insulin resistant, is exposed to TTS.
[Slide.]
Fibrinogen is an independent risk factor
for coronary artery disease. The lab value was
also assessed only for 12 months in the
double-blind and open-label phases. The data on
mean change from baseline is suggestive of an
increase with TTS exposure for six to 12 months.
And the median change data are similar.
The effect of TTS on blood pressure was
also of interest as a risk factor for
cardiovascular disease, as a component of metabolic
syndrome, and in regard to the occurrence of
hypertension as an outcome in itself. In the
double-blind period, hypertension was recorded as
an adverse event for 1.3 percent of placebo
subjects, and 2 percent of TTS subjects.
A 2002 meta-analysis of 61 studies
conducted by the Prospective Studies Collaboration
in the U.K. found a two-fold increase in deaths
from vascular disease and ischemic heart disease in
175
40 to 69-year-olds for each increase of 20 mm in
systolic blood pressure, and 10 mm in diastolic
blood pressure.
[Slide.]
And here you can see: a rise of 10 to 19
mm occurs in 5 percent more subjects who received
TTS for 6 to 12 months, as compared to the placebo
subjects.
[Slide.]
Similarly, there are 4 percent more
subjects who had rises of diastolic blood pressure
from 10 to 19 mm in the group who received TTS for
up to six months, as compared to placebo subjects.
We don't have data that could speak to the
issue of central obesity, but we can look at the
changes in weight over the course of the studies.
[Slide.]
Although weight gain appeared to occur at
higher frequency with greater TTS exposure, so too
did the equivalent amount of weight loss.
[Slide.]
A short-term clinical trial database can
176
provide only limited information about risks with
longer latency, and a significant background
incidence; for example, no MIs occurred in these
trials.
In regard to breast cancer, epidemiologic
studies have suggested that androgen levels in
women may be linked to the risk for developing
breast cancer. And, similarly, there have been
studies suggesting that increased insulin levels
may increase breast cancer risk. Some authors have
suggested that androgen's potential role in breast
cancer risk may be through its impact on insulin
resistance.
[Slide.]
This slide describes the four cases of
breast cancer that occurred in the trails--as
you've previously heard. But given the dual
exposures to estrogen and to testosterone--all
aside from this one placebo subject--and the
relatively short duration of testosterone use--5 to
37 weeks--there are insufficient data to assess the
causal role of TTS. And this highlights the
177
limited ability of short-term clinical trials to
answer questions of causality, particularly in a
population with a moderately high background
incidence of the outcome of interest.
[Slide.]
The current safety database is unable to
answer many questions about the safety of TTS for
several reasons.
First of all, placebo-controlled data is
available only for six months, and even long-term
exposure is limited to 12 months--in under 500
women, and 18 months--in 127 women. And, as you've
heard, women with diabetes and cardiac disease were
not studied.
For naturally menopausal women, our
concerns would include safety of TTS in women who
retain their uterus, and the known risks of
estrogen and progestin, which might be used for
greater duration by women using the TTS than they
would otherwise be.
[Slide.]
Procter & Gamble has a number of studies
178
in progress that will provide some additional
useful data, and these include--as you've
heard--the surgically menopausal studies: currently
321 subjects have entered this extension phase,
which is in year two of what will be, ultimately, a
three-year extension. In addition, the two
placebo-controlled studies you've heard about in
naturally menopausal women, there's a six-month
study completed, and a 12-month study close to
completion, each enrolling about 400 women. And of
these naturally menopausal studies, 281 women have
enrolled in safety extension phases.
These studies will also ultimately provide
293 paired endometrial biopsy samples.
And, finally, there's in progress a study
of the TTS patch being used alone in women who are
not taking estrogen or system
estrogen-plus-progestin--although they are allowed
to use vaginal preparations. And this study has a
projected enrollment of 750 women in three arms,
which are placebo, a dose of 150 mcg a day--half of
what we're hearing about here, as well as the 300
179
mcg per day dose.
[Slide.]
As you've heard, Procter & Gamble has
proposed a post-marketing pharmacovigilance study.
And, briefly, to review, they propose to do a
prospective cohort study in a claims database, with
three-to-one matching of current and recent users
with control subjects, planned outcomes, including
cardiovascular disease and breast cancer; endpoints
to be adjudicated by a panel of medical experts
blinded to treatment exposure; and they propose
that the first analysis will be available at 24
months post-launch.
Further, they estimate that their power to
detect cardiac events occurring with a relative
risk of 1.5 will reach 85 percent by year five.
[Slide.]
We have a number of concerns about the
utility of this proposed plan, however.
First of all, to answer safety questions,
does a claims database and cohort study provide the
same level of evidence as a randomized
180
placebo-controlled trial?
We're concerned that the project sample
size is inadequate. A study powered to detect a
relative risk of 1.5 for cardiovascular disease may
miss important but lower risks. And, just to
remind you, the risks seen in the WHI
estrogen-plus-progestin study were on the order of
1.2 for total cardiovascular disease, 1.3 for
breast cancer, and 1.4 for stroke. And to detect
risks of this size, a sample size of almost 17,000
was needed.
In addition, we're not given any
information on the power to detect an increased
risk of breast cancer.
We're also concerned that events with long
latency may not be detected. And, again, in the
WHI E+P study, breast cancer rates did not diverge
until year four, suggesting that the effect of
hormone exposure may not manifest above the
background incidence until that time.
In addition, we're concerned about
recruitment goals, which have not been met
181
previously using this database. And we're also
concerned about turnover in plan coverage. And
although you've heard that this plan retains 85
percent per year, when you take that out to year
five you can see that you're retaining only 44
percent of the original population.
[Slide.]
The WHI had far-reaching effects on our
assessment of risks associated with long-term
hormonal treatment. We learned from this that the
data was discrepant from that we'd previously known
from observational studies, and that reinforced the
value of prospective, randomized controlled studies
of adequate duration to be able to define
attributable risk.
And, ultimately, WHI indicates the need to
give heightened scrutiny to hormonal therapy in
post-menopausal women.
[Slide.]
To summarize the issues we must consider:
the sample size and duration of treatment is
inadequate to exclude serious risks, including
182
cardiovascular disease and breast cancer, with this
treatment; and the population studied is inadequate
to identify important risks in naturally menopausal
women using estrogen and progestin, and in
sub-groups at higher risk for cardiovascular
morbidity.
We look forward to your discussion and
assistance in resolving these issues and the
efficacy issues raised earlier by Dr. Davis.
Thank you.
DR. GIUDICE: Thank you, Dr. Soule. And
the committee will have an opportunity to ask
questions of both of the FDA presenters after lunch
this afternoon.
Open Public Hearing
This is now time for the open public
hearing. And I have a statement, first, to read
regarding this.
Both the FDA and the public believe in a
transparent process for information gathering and
decision-making. To ensure such transparency at
the open public hearing session of the Advisory
183
Committee meeting, FDA believes that it is
important to understand the context of an
individual's presentation.
For this reason, FDA encourages you--the
open public hearing speaker--at the beginning of
your written or oral statement to advise the
committee of any financial relationship that you
may have with the sponsor, its product and, if
known, its direct competitors. For example, this
financial information may include the sponsor's
payment of your travel, lodging, or other expenses
in connection with your attendance at the meeting.
Likewise, FDA encourages you, at the
beginning of your statement, to advise the
committee if you do not have any financial
relationships.
If you choose not to address this issue of
financial relationships at the beginning of your
statement, it will not preclude you, however, from
speaking.
So I'd like to begin with the open public
hearing speakers. I would like to advise the
184
presenters to please state your name and your
organization. Presentations will be strictly
limited to three minutes--and we have a timer up
here. The light will be green for the first two
minutes and 30 seconds, then yellow for the
remaining 30 seconds--and this will be the warning
to conclude your talk. And the light will turn
red at the three minute mark, at which time the
microphone will cease to work.
[Laughter.]
So I'd like to call now--with that
proviso--Ms. Lisa Martinez, please.
MS. MARTINEZ: First, I'd like to state
that the Foundation just received a $4,000
charitable donation from Pfizer.
Now, on that point--good morning. I'm
Lisa Martinez, a nurse and an attorney, and the
executive director of the Women's Sexual Health
Foundation, an international non-profit
organization based in the U.S.
Our primary mission is to educate the
public and health care professionals in the area of
185
female sexual health, including FSD.
We have heard from many women and their
partners relating to female sexual health problems.
These stories are heart-wrenching, and have a
common theme: women are devastated, suffer in
silence, feel very much alone in their journey to
find the right answers, care and treatment; and
wish that their sexual health would be taken
seriously.
For women in relationships, this impacts
not only them but their partners, who often feel
equally helpless and devastated.
Sexual problems are not an easy subject to
discuss. Women may feel embarrassed, and yet they
don't give up. Some have gone for years looking
for help from various providers, sometimes with
success and sometimes not. It's not unusual for us
to hear that women have been told by their provider
that their problems are all in their head, or that
a hysterectomy or bilateral oophorectomy could
never be he physical cause of sexual health
difficulties, and that any such problem would be
186
purely psychological.
The Foundation believes that a
multi-disciplinary approach should be used to
address sexual health problems. This would include
both physical and emotional assessments.
As part of this complete approach to
women's sexual health complaints, a serious effort
must be made to determine if there are physical
causes, such as hormone insufficiencies. Health
care providers need to follow well-recognized
workups that will leave no stone unturned, so that
treatment plans are specifically targeted at the
underlying causes of sexual dysfunction.
Consideration should be given to pharmacologic and
counseling.
Currently, there are no FDA-approved
treatments for FSD, and providers are using
off-label medications that have not been studied in
women under FDA oversight.
There is a need for such treatment,
including testosterone. But, more importantly, FSD
is a serious health issue, and not just a lifestyle
187
issue.
Thank you.
DR. GIUDICE: That you for your comments.
The next presenter, Mark Klein.
DR. KLEIN: Yes, I am a Procter & Gamble
shareholder and a physician.
If we balance Intrinsa's unimpressive
modest results from very short-term studies,
against my estimate that 50 to 80 percent of all
the Intrinsa sold will be abused by non-menopausal
girls and women--including some pregnant and
nursing--it's a no-brainer: this is too dangerous
to license for any use.
Over a decades time we could be looking at
many tens of thousands of girls, women, fetuses and
newborns permanently injured by anabolic steroids.
There is no way to avoid such abuse. Once
approved, Intrinsa will be available off the
internet, off-label and on the black market.
In my 40 years of medicine I have never,
ever seen government action prevent abuse of a hot,
popular drug.
188
As the manager of a large asset family
investment office holding Procter & Gamble, I feel
major losses--like what happened to us with Wyeth's
phyn-phen, and Merck's Vioxx. They happened
because today's very weak FDA allowed these
companies to cut scientific and ethnical corners.
Almost as certainty, Intrinsa will result in mass
tort class actions that could drive Procter &
Gamble into bankruptcy.
I suspect there are many very savvy,
seasoned long-term investors like myself unwinding
big pharma holdings. I've reduced ours over the
past two to three years from 15 percent to 6
percent.
The core problem is the big pharma's
willingness to sacrifice scientific integrity to
make their earnings numbers. And this is sure the
case with Procter & Gamble.
As an investor and trustee for family
accounts, I will sell our Procter & Gamble should
Intrinsa be approved. The potential lawsuit risks
for the company are so great that, in my opinion,
189
as a fiduciary, holding Procter & Gamble violates
the prudent-investor rule.
In conclusion, I believe Intrinsa is the
most hazardous non-narcotic drug ever presented for
FDA approval. I urge it to be rejected for any
use. And, if approved overseas, banned from sale
in this country under threat of severe criminal and
civil sanctions.
I have one word of advice to Procter &
Gamble. I am personally absolutely shocked that
you have gone into this business. We are in the
business of selling soap, we are in the business of
selling implicit promises not overt promises. And
I hope you keep in mind, "Hell hath no fury."
DR. GIUDICE: Thank you for your comments.
The next presenter is Rosalyn Washington.
MS. WASHINGTON: Good morning. My name is
Rosalyn Washington, and I have no financial
affiliation with Procter & Gamble.
I am a wife, a mother, and a woman who
suffers from low libido. Almost 10 years ago I had
a hysterectomy with removal of both of my ovaries.
190
This surgery--which, in my opinion, is performed
far too frequently on thousands of women in the
U.S. on a yearly basis--robbed me of my sexual
desire.
Unless you have experienced the lack of
sexual desire you cannot completely understand the
feeling of frustration and sense of inadequacy I
have. When I learned of this research being
conducted to help women with low libido I jumped at
the chance of being a volunteer for the clinical
trial study.
Overall, my experience in the study was an
excellent one. I did not grow a mustache or a
beard or develop large muscles. My voice did not
deepen, and I did not grow hair on my chest.
However, there was a noticeable increase
in my libido. I had first-hand experience with
the positive effects of the Intrinsa testosterone
patch, as a participant in the clinical trial
studies, and I would like to experience those
feelings again.
It is a known fact that women are much
191
more complicated than men. In a lot of cases our
libido is directly linked to our emotions and
mental state of mind. But this is not always the
case. Physical factors, like a hysterectomy can
affect libido by removing the ovaries which produce
testosterone in a woman's body.
Intrinsa is a drug therapy that I believe
the studies have shown to be effective in raising
the levels of testosterone in a woman's body with
little or no side effects.
It is my hope that Intrinsa receives FDA
approval and be made available countless women like
myself, who are seeking a solution for our sexual
dysfunction.
A healthy and satisfying sex life is
important to a woman's physical and mental
well-being. I believe Intrinsa will help restore
the level of hormones necessary for me to once
again have a healthy sex life.
Thank you.
DR. GIUDICE: Thank you.
The next presenter is Kathleen Kelly.
192
MS. KELLY: In 1982 my mother died of
ovarian cancer, setting in motion ready-made
decisions for me and my sister with ovarian cancer
familial risks.
My name is Kathy Kelly. In the summer of
1998 I had a hysterectomy and a bilateral
oophorectomy. I searched on-line and found very
little that was helpful to me, so I gathered
materials, resources and added a discussion board,
and launched a website called "Hyster-Sisters."
Now, over six years later, Hyster-Sisters
is the largest on-line hysterectomy community, with
over 55,000 members.
The Hyster-Sisters site is neither
anti-hysterectomy, nor pro-hysterectomy; rather, it
is a on-line community of women who give and
receive support for hysterectomy decisions and
recovery.
The Hyster-Sisters have sent me today to
share their stories with you.
Many members recover from their surgery
and head back into their lives without much
193
fanfare. For others, they return to our site
months or even years later, in search of additional
support for new health issues. Predominantly, they
return in search of support for hormone therapy,
many complaining of a missing libido.
And while we have thousands of personal
posts from frustrated members regarding their loss
of desire, time constrains me only to share with
you a few.
A 47-year-old woman wrote: "Since my
hysterectomy, I've tried all forms of hormone
therapy and some herbal treatments. I'm beginning
to wonder if it's simply impossible to get back my
libido."
A woman in Arkansas wrote: "I have a
check-up appointment with my OB/GYN next month, and
my husband is going with me so we can talk about my
libido. What is wrong with me? I am 24 years old.
I love my husband. Has anyone else experienced a
change in libido? How do you go about correcting
it?"
A 43-year-old woman in Southern California
194
wrote: "Prior to my surgery my libido was
tremendous. My husband and I have had 25 years of
awesome sex. My husband has been very, very
patient with me. We have tried just about
everything, but to no avail."
The Hyster-Sisters have sent me, because
it is our hope that the medical community find
better treatment so that the hysterectomy is truly
a last resort. But for those hundreds of thousands
who have had an oophorectomy, we would like hormone
options to better restore what we have lost.
Our next generation of women is depending
on us.
The Hyster-Sisters have sent me to ask
that you approve this drug--the testosterone
patch--as one option for the libido needs of the
surgically menopausal woman.
Thank you.
DR. GIUDICE: Thank you.
The next presenter is Leonore Tiefer.
DR. TIEFER: First slide, please.
My name is Leonore Tiefer--no money, no
195
way.
AS a psychologist with over 30 years of
teaching, research, awards and publications in
sexuality, I see today as a perilous moment in the
history of women's sexuality.
Next slide.
[Slide.]
A few random credentials, which I offer
because this is the first women's sexuality drug
that the FDA has ever reviewed--and there is no
sexuality drug committee. Input may be useful from
someone who has spent decades immersed in issues of
sexual nomenclature, measurement, motivation,
behavior and biology.
Third slide, please.
[Slide.]
Here are my concerns--and I have handouts
on these points, since you can only say just so
much in 180 seconds.
The Intrinsa trials are grossly inadequate
to assess the risks of extended steroid hormone
treatment. And I hope we don't have to go through
196
another HRT scandal to learn this again. That's
point one.
Point two is that assessing sexual
experience is subtle and complex and arbitrary.
Experts in sexology agree that there are numerous
ways to define and measure desire and satisfaction.
Methods chosen in every study must be closely
examined for what they leave out, as well as what
they include.
Point three: these Intrinsa trials
excluded women with medical problems, relationship
problems and life stress. It's no wonder it took
52 trial sites to find a meager 1,095 subjects.
How representative are these carefully
selected subjects of the millions P&G is hoping to
interest in it's new medicine?
Which brings me to point four: Intrinsa is
not a glass of Chardonnay, and yet we have already
seen that it may well be promoted, with a giggle
and a wink, as the female Viagra.
Not so. This is a steroid hormone women
must continuously take for weeks before getting an
197
effect. Yet P&G's promotional materials encourage
the attitude that millions of women are walking
around under-androgenized, in danger of imminent
sexual withering away. It's a revival of menopause
as a deficiency-disease, only this time it's
testosterone riding to the rescue.
Fourth slide.
[Slide.]
So here are my recommendations.
First, postpone the application until
there are longer studies on more appropriate
populations.
Second: if women with low desire are
testosterone-deficient, we must have an affordable
assay to measure that deficiency, and there is none
now.
Third: good sex research should always
have a qualitative component.
And, finally, the FDA's DDMAC needs to
carefully monitor the P&G materials for bias and
boundary violations.
Last slide.
198
[Slide.]
I am representing a large group of experts
who couldn't be here today.
Thank you very much.
DR. GIUDICE: Thank you.
Our next presenter is Wayne Shields.
MR. SHIELDS: Hi. My name is Wayne
Shields. I'm president and CEO of the Association
of Reproductive Health Professionals.
Thank you for holding this hearing and
inviting us to be here today.
We are a 501(c)(3) organization, and we do
receive support from foundations and companies, and
we have in the past received unrestricted grants
from Procter & Gamble.
It's important to know, I think, that ARHP
has been around a while, and that ARHP is a medical
organization of over 12,000 health care providers,
multi-disciplinary, mostly OB/GYN and family
practice. We have researchers and on-the-ground
practicing clinicians.
On behalf of our members I'm glad to be
199
here today.
I also have to say I realize I'm a guy,
and this conversation is about sexuality. I do
represent mostly a female constituency. But I
think the topic of sexuality is key, and it's
under-addressed in America. So this is a great
forum, and thank you for providing that
opportunity.
ARHP is an organization accredited by the
ACCME, and we provide CME for health care
providers, and other credits. And we advocate for
evidence-based research. And we support the
availability of a wide range of safe and effective
and appropriately used treatment options on women.
I'm here today to support this
application, if it's appropriately used--and I know
you'll have that discussion. And I'm here today to
let you know that we believe this medication is
appropriate for enhancing sexual desire in a very
particular subset of surgically menopausal women.
We do support--and this is important for
us in our mission at ARHP--careful clinical
200
screening to ensure that this medication is given
only to appropriate candidates. I think that
that's key.
In the past, sexual health research has
focused mostly on men's sexual health, as you know.
We feel it's important that female sexual health be
represented more prominently. And while men have
benefitted from a number of products and the
ensuing attention on their sexual disorders, this
is an important conversation and an important
product, because it allows a forum for conversation
about female sexuality, and their very unique and
very different sexual disorders.
We support focusing on the sexual health
of women. We see the introduction of a safe and
effective medication for women as a great
opportunity to be able to discuss, in an
evidence-based and appropriate manner, female
sexuality to enhance health care provider
communication with patients.
We believe that hypoactive sexual desire
disorder is a real condition, and that surgically
201
menopausal women who suffer from it deserve a range
of treatment options, whether they be behavioral or
medical. Many women with HSDD can benefit from
counseling and lifestyle adaptation and other
non-medical treatments--very true. Let's talk
about that more. But there remain some women for
whom a safe and effective medical intervention such
as this one will be of benefit.
And we think this is especially relevant
for surgically menopausal women.
The data and research we have examine
indicates that while it's not appropriate--
[Time expired. Microphone turned off.]
[Laughter.]
DR. GIUDICE: Thank you.
The next speaker is Karen Hicks.
DR. HICKS: First slide, please.
[Slide.]
I'm Dr. Karen Hicks, a sexual health
educator and founder of the Dalkon Shield
Information Network--with no financial relationship
to the company.
202
I'm here to request that approval of the
Intrinsa patch be delayed until relevant safety
issues have been fully reviewed and documented.
Drug safety issues and scrutiny of the FDA have
dominated the business news lately, due to the
Vioxx scandal and risks of antidepressants for
children. Today, as you deliberate the dawn of a
whole new class of sexual medicines for women, it's
time to consider some new precedents for
considering the safety issues relevant to a drug
like Intrinsa.
Next slide.
[Slide.]
I raise four questions: One, what is the
safe dosage for individual variations among women
who may be very different from women in the
clinical trials, particulary with regard to ages,
differing weight profiles, general health status
and possible ethnic backgrounds?
Two: what is or isn't know about the
long-term use of this drug? In the clinical
trials, subjects used Intrinsa for time ranges
203
between 14 and 24 weeks, yet it's intended to be
used continuously over the long term, and possibly
for years.
Three: what potential adverse reactions
has ben anticipated, and what might likely
unanticipated outcomes be? Experience with
testosterone in pill or injectable form, and other
reproductive hormones prescribed to women, include
cancers of the breast and other tissues, liver
ailments, excessive facial hair growth and skin
problems--to name a few.
Four: how will problems in prescribing and
dispensing be prevented or minimized. Based on the
excitement being generated in the press for this
drug already, I predict that off-label use with
soon follow.
Next slide.
[Slide.]
This week's Journal of the American
Medical Association has two relevant editorials.
The first speaks to the weaknesses of the current
post-marketing surveillance process at the FDA.
204
The second explores the potential for conflict of
interest in the evaluation of suspected adverse
drug reactions.
The journal editors recommend that an
independent entity located outside the FDA be given
primary authority for this task. The Vioxx and
Dalkon Shield IUD scandals hinge on the long
suppression of information on dangers they posed to
their users.
Next slide.
[Slide.]
I off five recommendations for setting new
precedents.
One: admit full disclosure of the clinical
trials to the public.
Two: initiate a user registry under the
purview of the FDA to all users who volunteer to be
kept informed.
Three: upload all documentation on
efficacy and safety to the FDA website and announce
the URL and telephone numbers widely on pharmacy
patient package inserts and information sheets.
205
Four: if warranted, contact all users
early through pharmacy databases about discoveries
of dangers relative to this drug's use.
Five: include label warnings about
duration of use beyond the length of the clinical
trials.
Next slide.
[Slide.]
The public perception, reflected by
substantial press coverage, already suggests that
Intrinsa is "Viagra for women." I find this notion
distorted and disturbing. This treatment is not
equivalent in manner or duration of use. It acts
on different body systems and has different
effects.
Please consider these recommendations as
you deliberate today.
Thank you.
DR. GIUDICE: Thank you.
The next presenter is Jean Koehler.
DR. KOEHLER: Hi. I'm Dr. Jean Koehler.
First of all, I want to let you know that I was a
206
paid consultant in the persistence of benefit phase
of the Intrinsa trials, and am currently a paid
regional consultant with Procter & Gamble, and a
long-term stockholder.
Additionally, I'm a paid consultant for
Ortho-MacNeill Pharmaceuticals.
It is because of my clinical experience,
and experience interviewing women on Intrinsa, that
I felt the need to take te time out of my practice
to be here today, at my own expense, to support the
approval of this product. I am a licensed marriage
and family therapist, and certified sex therapist
in private practice. Additionally, I'm a faculty
member at the University of Louisville School of
Medicine, and have held both of these positions
since 1976.
I am also the immediate past president of
the American Association of Sex Educators,
Counselors and Therapists--or AASECT. With over
2,100 members, AASECT is the oldest, largest
certifying organization of professionals in the
sexuality fields. While AASECT does not endorse
207
any particular product, I have been authorized by
the organization to support the kind of research
that is being presented here today.
In my own professional opinion, as a sex
therapist and medical educator for over 28 years, I
also see a great need for this product. Female
hypoactive sexual desire disorder is not only one
of the most common presenting complaints in
psychotherapist and physicians offices, it also has
been the most difficult of all the sexual
dysfunctions to treat successfully.
While I have successfully treated many
psycho-social causes of this disorder, my multiple
years of experience tell me that without
concomitant testosterone therapy, psychotherapy and
relationship therapy has failed with women whose
testosterone levels remain low.
I'll give you just one of many examples of
positive use of testosterone in my practice.
The nine-year marriage of my client--whom
I'll call "Laura"--was about to break up because
she had totally lost her sexual desire and
208
willingness to have sex, after a good previous sex
life with her husband. Her free testosterone
levels were very low, and her husband as so
sexually frustrated that he was becoming
emotionally abusive.
After the combination of psychotherapy and
testosterone replacement therapy restored her
drive, Laura reports two years later, on follow-up,
that she is still using her prescription, and that
her marriage has more than stabilized, and she
reports no adverse events. She now enjoys sex
again, and two little children were spared the
trauma of impending divorce.
Similar reports came from the women I
interviewed in the persistence of benefits stage.
And not only did they notice important increases in
desire and function, but improved and increased
emotional closeness with their partners.
So the importance of this product is not
just about one more sexual experience per month.
It's also about a generally improved quality of
life for these women.
209
The patch will no doubt only work for a
carefully selected group of women, but as a woman
and as an advocate for patients like Laura, I
maintain they deserve--
[Time expired. Microphone turned off.]
DR. GIUDICE: Thank you.
The next presenter is Anne Kasper, and I'm
making these remarks, prepared by Breast Cancer
Action, on behalf of both Breast Cancer Action, and
Our Bodies Ourselves.
Both organizations work in the public
interest, and do not accept funding from the
pharmaceutical industry as a matter of
principle--and neither do I.
Breast Cancer Action opposes approval of
the proposed indication for NDA-21-769. While
reduced libido and vaginal dryness are serious
concerns for women with breast cancer who are put
into menopause by chemo treatments, the solution
does not lie in the approval of this therapy, which
has only been briefly evaluated, and not in
populations of women who may be at increased risk
210
from hormonal exposures due to their cancer
history.
We recognize that the proposed indication
of this NDA is for women whose menopause is
surgically induced, but we are deeply concerned
about the enormous potential for off-label use of
the therapy in inappropriate populations.
It is now widely accepted that breast
cancer is largely a hormonally driven disease.
Most of the known risk factors for breast cancer
have to do with lifetime exposure to endogenous
hormones--particularly estrogens. The skyrocketing
incidence of breast cancer--and the Women's Health
Initiative results on hormone replacement therapy
have raised concerns about the implications of
exogenous hormonal exposures as well. While much
remains unkonw about the etiology of breast cancer
and other hormonally driven women's cancers, there
is great concern that any treatment that interferes
with the eodocrine system will ultimately stimulate
some aspect of cancer development. Breast Cancer
Action therefore urges that all women at risk for
211
breast cancer--and the agencies charged with
protecting their health--proceed with extreme
caution before pursuing hormonal treatments of
other medical conditions.
For women who have already been diagnosed
with breast cancer, this caution cannot be stated
too strongly. A very recent study published in
September in the International Journal of Cancer
indicates that high serum testosterone levels
predict a greater likelihood of breast cance
incidence.
The drug application is based on a small,
six-month trial. Yet we know from both the Women's
Health Initiative and the experience with DES that
the long-term effects of hormonal therapies may not
be known for many years. Approval of the NDA will
lead to one more instance when women become guinea
pigs in an uncontrolled experiment that may have
serious implications for their long-term health.
Thank you.
DR. GIUDICE: Thank you.
The next presenter is Sidney Wolfe.
212
DR. WOLFE: I'm Sidney Wolfe of the Public
Citizen's Health Research Group.
Is an increase in approximtely one
sexually satisfying encounter a month--not from
zero to one, but approximately from four to
five--worth the possibility of an incrased risk in
breast cancer or a coronary heart disease?
Is the FDA actually considering the
approval of this product?
I was interested to hear Dr. Davis'
skepticism--I think is the proper way of raising
it--as to what clinical significante there is in
going from four to five sexual encounters a month,
or having an increase of just 5 to 6 points more
than a placebo on a scale of 100, in terms of
sexual desire.
What is known is that there's a fairly
good consensus among epidemiologist who work in the
area of endogenous--or body-produced--hormones and
breast cancer risk, that increasing levels of the
common estrogens and adrogens such as testosterone
are associated strongly with increasing levels of
213
breast cancer risk. This is best illustrated by a
recent pooled analysis of nine high-quality
prospective studies which were published in 2002 in
the Journal of the National Cancer Institute. In
the pooled analysis, the risk of breast cancer for
post-menopausal women increased by approximately
two-fold, with a quadrupling of blood testosterone
levels.
In the data on the transdermal
patch--which you saw some of this morning--the
average blood level of testosterone in these 18 to
49-year-old women rose from a pre-treatment level
of 176 nanograms per liter, to 797 nanograms by
week 52. This is a 4.5-fold incrase in blood
testosterone levels by 52 weeks, which is slightly
higher than the four-fold increase in blood
testosterone levels in the pooled study mentioned
above that was associated with a two-fold increased
risk of breast cancer.
Other concerns about the chronic use of
testosterone, as mentioned by the FDA, is the
increased risk of coronary artery disease. A study
214
entitled "The Relationship Between Serum Sex
Hormones and Coronary Artery Disease in
Post-Menopausal Women," found "--evidence of a
positive relationship between the serum free
testosterone level and the degree of coronary
artery disease in women." And, again, in this
study there was a four-fold increase in free
testosterone with the patch.
Decreased sexual desire is a very
complicated problem, as discussed by Dr. Tiefer
today. There is little question that a large
proportion of women with this complaint respond
very well to counseling that may reveal underlying
problems, such as a history of being sexually
abused, current unstable or unhealthy
relationships, depression, or other causes better
dealt with directy rather than being glossed over
with a testosterone patch.
The journalist H.L. Mencken has said that
for every complicated problem there is a simple
solution--which is usually wrong."
I urge you and the FDA to reject the
215
application for approval of the testosterone patch.
And, in closing, I have no financial
conflict of interest.
DR. GIUDICE: Thank you.
The next presenter is John Grossman.
DR. GROSSMAN: Good afternoon. I want to
thank the FDA and the panel for allowing me to
participate in this important process that will
serve the interests and advance the health of
women.
My name is Dr. John Grossman, and I am
professor of Obstetrics and Gynecology,
Microbiology and Tropical Medicine, Prevention and
Community Health, and Health Services Management
and Leadership at the Geroge Washington Univeristy.
I'm also the Executive Vice President of the
Society for Gynecologic Investigation.
For the record, my comments do not reflect
the positions of either of these organizations. I
am here, rather, today to present my own
perspective, based on being a practicing
gynecologist in Washington, D.C. since the mid
216
1970s.
I have no personal financial relationship
with the sponsor of this product, nor with their
competitors. And I have no financial interest in
this product or any comparable products.
My 29-year practice has become what my
father, who was a founding Fellow of the American
College of Obstetricians and Gynecologists, and a
medical practitioner for 46 years, would have
called "a mture practice." As years have gone by,
I have had the good fortune to have my patients
place increasing trust in our professional
relationship. And during this period, I have
learned that many women have problems with sexual
intimacy for a variety of reasons. Perhaps one of
the most difficult problems to quantify and to
treat is diminished sexual drive.
My interest and concern about this issue
prompted me to request a copy of the product
briefing document that has been submitted to this
panel. My review of the information currently
available leads me to believe that Intrinsa is safe
217
and effective for increasing satisfying sexual
activity in these women who have had
hysterectomies, and I believe it could potentially
benefit some other groups of women, as well.
I becan my medical education during the
sexual revolution of the 1960s Even then, I would
never have envisioned a time in America when
television commercials would address what we
currently call "erectile dysfunction." Relatively
recently, health professionals have expressed
concerns about gender bias, and the public has
strongly supported gender equity in all areas of
life.
I believe that Intrinsa safety and
effectively increases sexual desire and the
frequency of satisfying sexual activity, while
reducing sexually-related personal distress. And I
urge the panel to recommend approval of this
product, not only to provide gender equity in
issues of sexuality, but also to address the needs
of my patients, and those of many other women who
might not otherwise be heard.
218
Thank you.
DR. GIUDICE: Thank you. The next
presenter is Lenore Pomerance.
MS. POMERANCE: Good morning. I have no
financial relationship with Procter & Gamble.
My name is Lenore Pomerance. I'm a
psychotherapist in Washington, D.C., and work with
mid-life women and their partners on menopause,
sexual relationships, and healthy lifestyles.
The approval of Intrinsa is premature
until long-term data have proven its safety.
What short memories some of us have. 28
months ago the Women's Health Initiative halted its
unprecedented trial of estrogen and progestin
because risks outweighed the benefits for
post-menopausal women.
The study was undertaken to test whether
the estrogen preparations that millioins of women
had already been taking for over 30 years helped or
harmed them. If we knew then what we know now,
would we have let doctors and drug companies
convince us that menopause was a disease? Would we
219
have let ourselves be fooled into believe what Dr.
Robert Wilson--who claimed in his book Feminine
Forever--that "menopause is a hormone deficiency
disease, curable and totally preventable, and that
every woman, no matter what her age, can safety
live a fully-sexed life for her entire life?"
The cure for this disease was "hormone
replacement therapy," replacing what had been lost.
And those 30 years saw efforts to make HRT reverse
the ravages of old age, from wrinkled skin and weak
hearts, to addled brains.
The results of the WHI have taken the "R"
out of HRT. Researchers and practitioners don't
talk of "replacement" anymore. The new term is
"HT," and it is to be used only for menopausal
symptoms of hot flashes and vaginal dryness, for
the shortest time, and in the lowest dose possible.
The new guidelins are not based on proven safety,
but on women's wilingness to live with the risks.
Getting rid of the "R" is a backhanded way
of admitting that menopause is a natural condiiton;
a physiological process that every woman will
220
experience if she lives long enough. Many women
are never bothered by the physiological changes.
Some even welcome them.
In my practice, women for whom menopause
is a crisis are often experiencing other
life-changing events like divorce, widowhood, dying
elderly parents, children leaving home, retirement.
Does history have to repeat itself? Will
Intrinsa become the Premarin of the 21
st century? I
don't believe for a minute that Intrinsa
prescriptions will be confined to surgically
menopausal women with low libido.
The FDA has approved off-label
prescribing, but it must foresee that, as happened
with the Viagra boom, this new drug will be
requested by many people on whom it has not been
tested. Believing that it's safe, these women may
well become the guinea pigs of the 21
st century.
Let's not let that happen.
Thank you.
DR. GIUDICE: Thank you.
The next presenter is James Simon.
221
DR. SIMON: I'm Dr. James Simon. I'm a
reproductive endocrinologist from Washington, D.C..
I'm not speaking as past president of the North
American Menopause Society, I'm speaking as a
practitioner.
My conflicts or potential conflicts are
listed for you in my handout.
Slide--could I have the next slide?
[Slide.]
This is a concerned physician's view of
the subject, and I'm representing myself and only
myself.
Next.
[Slide.]
The issue before the panel really is: is
low sexual desire really a clinically important
problem? And what about the risks.
Next.
[Slide.]
in this study from Laumann, oftentimes
brought into question, they found in a large cohort
32 percent of women experience low sexual interest.
222
And I say--this is from 1999--they have been
criticized by this high number.
Next.
[Slide.]
A recent study, published only this year,
looked at 29 countries, with 27,000 men and women,
aged 40 to 80, and defined sexual dysfunction as
"frequent or persistent problems." Those problems
studied in women were--most importantly apropos of
this application--lac of sexual interest.
Next.
[Slide.]
In this study, 65 percent of the women
were having sex. The most common dysfunction was
lack of sexual interest, found in 21 percent of the
women, similar to their previous findings. They
concluded, overall, that overall 39 percent of all
women in their study--remember, they studied 27,500
men and women--had at least one sexual dysfunction.
Next.
[Slide.]
Another study, looking at safety--and,
223
importantly, most women make the decisions about
hormones and safety of hormones based on the issue
of breast cancer. And in this study, recently
published in Menopause by the group at NIH, they
looked to see if testosterone could protect--could
protect--against the added risk of standard hormone
therapy already mentioned by previous speakers.
This was a retrospective, observational study of a
small cohort of women, average age 56 years, and
they were studied for 5.8 years. Outcomes were the
incidence of breast cancer.
Next.
[Slide.]
They found, actually, that in women on
testosterone, the risk was 238 per 100,000
women-years, lower than--not higher than--lower
than women who were in the WHI, lower than women in
the Million Women Study. And, in fact, consistent
with women who had never used hormones.
Their conclusions were: the addition of
testosterone to conventional therapy for
post-menopausal women does not increase and may
224
indeed reduce the risk of breast cancer. These
same investigators found this in pre-clinical
scientific experiments.
[Time expired. Microphone turned off.]
DR. GIUDICE: Thank you.
The next speaker is William Petok.
DR. PETOK: Good morning. I have no
financial relationships with the producers of this
product.
Good morning, members of the Advisory
Committee. Thank you for allowing me time to share
some of my thoughts about the TTS.
I'm Bill Petok, a psychologist and a sex
therapist in private practice. I'm also the Chair
of the Mental Health Professional Group of the
American Society for Reproductive Medicine. In
addition to my clinical activities, I also teach
obstetrics and gynecology residents at Baltimore's
Sinai Hospital, specifically about human sexuality.
As you are aware, sexual dysfunction in
American women occurs at a significant rate. HSDD
is the most frequent problem I see in my clinical
225
practice, and the problem most frequently seen by
the physicians that I teach. It is also one of the
most difficult problems to treat because it can
have many determinants.
In addition to the problem that HSDD
presents for an individual, this disorder can also
have an impact on relationships in which sexual
interaction is a vital aspect. Low desire on the
part of one partner can lead to frustration and
dissatisfaction for both. Some of the women that I
treat report they do not understand why they have
lost interest in sexual interaction with their
partners, especially when other aspects of their
relationship are good. The partner can be at a
loss to understand the changes in relationship, as
well.
It is important to note that not all HSDD
is hormonally related. AS I said before, it can
have many determinants that include the quality of
the relationship and other psychological factors.
It is a complicated problem that requires a careful
analysis and intervention. When hormonal factors
226
are implicated as part of the etiology, as for
surgically menopausal women, testosterone
supplementation can be an effective therapeutic
addition.
Over the years that I have treated sexual
dysfunctions, I have had many women report their
physician has prescribed testosterone in one form
or another. Frequently, the prescription is
offered as a cream that is to be applied topically.
Often the directions given to the patient are less
than adequate, and she asks me why she isn't
getting a result that increases her desire level.
She may be unclear in her understanding of
whether she is to use the product just prior to
sexual relations, daily, or with some other
frquency. A delivery system that makes sense and
has little room for misinterpretation would be
useful to these and other patients. I believe the
advantage of a product like the TTS is that it is
easier to apply and therefore less likely to be
misused.
I have two reservations--not so much about
227
TTS as about how it is reported in the press or
promoted.
One, I'm concerned that this product will
be inappropriately described as a "female Viagra"
and be viewed as a cure-all for all female sexual
dysfunction of any kind. The researchers have been
careful to describe its success with a specific
group of women: surgically menopausal women who are
on a stable dose of estrogen, and who are in
long-term established relationships.
It would be inappropriate to generalize
these findings to a wider group of women.
Second, I am concerned that women not have
expectations that are out of line with reality.
This is not a cure--
[Time expired. Microphone turned off.]
DR. GIUDICE: Thank you.
The next presenter is Raymond Rosen.
DR. ROSEN: Can I have my slides, please?
My first slides, please?
[Slide.]
Okay. Thank you.
228
I'm Raymond Rose, professor of psychiatry
at Robert Wood Johnson Medical School. I've been a
psychologist and sex researcher for approximately
30 years.
I serve as a research consultant, and have
received research support from P&G, Pfizer and
Solve Pharmaceuticals.
I'm here representing myself, and my
travel expenses wer paid by my department.
Next slide, please?
[Slide.]
I want to use my brief time to just to
comment on two important documents. One of them
has been discussed quite a bit today, and that's
the FDA Draft Guidance document on FSD from 2000.
The other document hasn't been mentioned, but I
think is quite important, which is a consensus
document that a number of people here participated
in in 2001.
Next slide, please.
[Slide.]
Just to begin with a few comments about
229
the guidance document. This is available on the
FDA website.
Next slide, please.
[Slide.]
And the guidance document , which is very
important--for people who haven't read it--covers
four areas of sexual dysfunction in women: as has
ben mentioned, inhibited or hypoactive desire is
the most common area. The guidance document
indicates how these are defined. And I think the
sponsor's done an excellent job in following those
guidelines.
Next slide, please.
[Slide.]
in terms of clinical trial endpoints,
which has been a majoar area of discussion today,
it's important for us to understand that the
satisfactory sexual events aspect of it really
comes from that guidance document. And the sponsor
has really done everything they can to meet those
guidelines.
Many of us in the field believe that,
230
particularly in the area of sexual desire
disorders, satisfactory sexual events is not
necessarily the optimal primary endpoint. And I
want to urge all of us to move beyond that. I have
spoken to members of the FDA and, hopefully, we'll
be moving past that.
Nonetheless, I think the sponsor has shown
great consistency in the effects across differenet
endpoints.
Next slide, please.
[Slide.]
Just to comment quickly on the Princeton
Consensus Conference--about 16 international
experts, many of whom are here today, participated
in this meeting in Princeton in June 2001.
Next slide, please.
[Slide.]
Of importance, we defined female androgen
insufficiency. It's discussed and defined at great
length.
Next slide, please.
[Slide.]
231
And we identified four important
etiological sub-types, the most important of which,
of course is the ovarian sub-type, including
oophorectomy or effects of radiation.
I recommend this document because it
provides very clear guidelines for diagnosis and
classification of androgen insufficiency in women.
And I hope that this document and other similar
guidelines will be used in further developing this
product if it's approved.
Thank you.
DR. GIUDICE: Thank you.
The next presenter is Amy Allina.
MS. ALLINA: Thank you. I'm here from the
National Women's Health Network, which is a
nonprofit organization that works to improve the
health of all women by influencing health policy
and supporting consumer decision-making. We accept
no financial support from pharmaceutical companies
or medical device manufacturers.
I'm going to talk primarily about my
concerns about the limitations of the data. But
232
putting those limitations aside for a minute, I do
want to acknowledge that it appears Intrinsa offers
some benefit to the women in whom it's been
studied. And the need in that group is real, so
the chance to provide real help to women with the
problem of low sex desire is hard to pass up.
As women's health advocates, however, we
can't consider this product in a vacuum, and
neither can the FDA. The world changed when the
Women's Health Initiative revealed the long-term
negative health effects of hormone therapy. And a
six-month study of a testosterone patch that would
be the first drug of its kind may have seemed
adequate before, but it's not today.
Women who might stand to get a benefit
from the testosterone patch also need to know about
its long-term effects on their healht. In the wake
of the WHI, it's appropriate and necessary to
exercise special caution about long-term hormone
use without long-term safety data.
The patch hasn't been studied for an
adequate period of time to find out whether it
233
might increase risk of breast cancer. Some early
indication might have come from a mammographic
study but, as FDA noted in its medical review,
there are several limitations that make it
difficult to determine what effect the testosterone
patch might be having on breast tissue.
The short-term that P&G has collected so
far are also not able to provide reassurance about
the effect of their product and risk of heart
disease. The events occurring in the extension
phase, which FDA noted, could reflect
cardiovascular events are potentially important.
Although there is no placebo comparison for the
extension phase, the average age of women in the
combined trials was 49. So these problems can't
simply be dismissed as expected background.
The fact that lipid profiles were similar
in the testosterone and placebo groups isn't
adequate reassurance, since lipid levels failed to
predict the the cardiovascular problems that were
eventually found to be associated with hormone
therapy in the WHI.
234
In addition to these long-term safety
concerns, I want to urge FDA and all of you to
balance the benefit that the testosterone patch
might offer to a small group of women, with the
health risks it may pose to many more.
It would be naive and irresponsible, I
think, to pretend that this drug will only be
promoted anda prescribed to women who are exactly
like those in the trials.
A cursory scan of health websites and
books that deal with sexual health issues shows
that the recommendation of testosterone for
treatment of women's sex problems is not directed
solely to those women. It includes advice to
younger women, including women in their
reproductive years--which raises a whole new set of
questions, for example, about the effect of
testosterone on future fertility.
Even the company's proposed patient
information leaflet blurs the line somewhat,
defining Intrinsa
[Time expired. Microphone turned off.]
235
DR. GIUDICE: Thank you.
The next presenter is Neil Goodman.
DR. GOODMAN: My name is Dr. Neil Goodman.
I'm speaking today representing the American
Association of Clinical Endocrinologists, where I
serve as Chairman for Reproductive Endocrinology.
The American Association of Clinical
Endocrinologists is a professional medical
organization with over 5,000 members throughout the
United States and 70 foreign countries. ACE is
devoted to the enhancement of the practice of
clinical endocrinology, and the betterment of care
for patients with endocrine diseases.
ACE supports the approval of the new drug
application for the transdermal testosterone system
from Procter & Gamble for the treatment of
hypoactive sexual desire disorder in surgically
menopausal women receiving concomitant estrogen
therapy.
As you've heard, menopausal women
frequently experience low sexual desire, which can
cause substantial distress and negatively affect
236
quality of life. Representing about a third of
post-menopausal women, surgically menopausal women,
in particular, experience a greater degree of
sexual dysfunction than any other group of
menopausal women.
For more than 20 years scientific research
has supported the use of testosterone for the
treatment of female sexual dysfunction in
surgically menopausal women. However most studies
have been limited in the number of participants and
the duration of treatment. In addition, there has
not been a form of testosterone--which should be
emphasized here--that gives women a simple and
patient-friendly means of delivering physiologic
levels of testosterone. The testosterone
transdermal system is the first such testosterone
product that has proven to be both efficacious and
safe for the treatment of the hypoactive sexual
desire disorder through randomized controlled
trials. These are statistically significant, based
on the information provided to the FDA.
It is the opinion of ACE, in reviewing the
237
study submitted to FDA by Procter & Gamble, that
the transdermal testosterone system can achieve
imporved sexual functioning at physiologic levels
of testosterone, with a minimal incidence of
adverse effects. This opinion is based, in part,
on the extremely comprehensive and well-validated
instruments for the measurement of female sexual
dysfunction that's been developed by Procter &
Gamble. These instruments, which include the
Profile of Female Sexual Functioning, the Sexual
Activity Log, and the Personal Distress Scale, are
mandatory for proving efficacy of testosterone
therapy. And I believe that these instruments have
proven their effectiveness, and have been validated
in the studies you've seen today.
The statistical analysis is highly
significant, taking into account the number of
women studied and the duration of treatment. Based
on these studies, combined treatment with estrogen
and transdermal testosterone has proven to induce
increased motivational aspects of sexual behavior,
not just frequency but, in fact, the desire and the
238
impact of orgasm in sexual intercourse.
ACE believes that the scientific data
provided to the FDA is sufficient to prove efficacy
and safety of the transdermal system, and should be
approved without further studies.
Thank you.
DR. GIUDICE: Thank you.
The next presenter is Doug Ronsheim.
REV. RONSHEIM: Thank you very much for
giving me the opportunity to speak.
I am the executive director of the
American Association of Pastoral Counselors. It is
a national counseling organization of licensed and
credentialed professionals providing clinical
services to individuals, families and couples.
The membership also attends to spiritual
and faith issues which clients might wish to
address in the context of care.
Professionally, I'm a licensed marriage
and family therapist, a Fellow in the American
Association of Pastoral Counselors, a clinical
member and approved supervisor in the American
239
Association of Marriage and Family Therapy, also a
Presbyterian minister, and have had a previous
faculty appointment at the University of Pittsburgh
Medical School Department of Psychiatry.
I initially became aware and interested in
the topic of hypoactive sexual desire disorder as
it was presented in a variety of counseling
sessions with couples. Decreased sexual desire
usually was not the presenting problem, but often
emerged during the sexual history that sexual
incompatibility post-surgery was a stressor--and
for good, understandable reasons, and not due to
any long psychiatric history.
The stress was due and exacerbated, may
time, because the couples' previous sexual
relationship had been quite satisfactory. In
addition, awareness related to this has emerged in
this past year in conversations with Dr. Larry
Nelson, who does research at NIH at the other end
of the reproductive spectrum, with premature
ovarian failure, where women often, at an early
age, for inexplicable reasons, lose the functional
240
capacity of their ovaries.
HSDD has intra- as well as inter-personal
ramifications which can be significant. Dr.
Nelson's research has shown that the loss of one's
capacity to give birth to children, in addition to
decreased sexual desire, is a double loss for them.
A high preponderance of these patients exhibit
varying degrees of depression and anxiety.
What is common here is a theme of loss.
Interpersonally, decreased sexual desire is bound
to affect a marital relationship, even though both
spouses understand that this is caused by a medical
or physiological reason, it can be experienced
personally, and can be a factor for increased
marital stress.
Having a resource for the treatment of
HSDD can be a significant benefit and assist
specifically in the improvement of a woman's sexual
response and the general improvement of one's
marital relationship which could again embrace a
natural, more responsive sexual relationship.
The importance of addressing HSDD is not
241
only to curtail an already difficult and
frustrating malady, but preventing the potential of
the further spiraling down of intra- and
inter-personal functioning, and decreasing the
chances of the co-occurring development of the
symptoms of depression and anxiety.
Thank you.
DR. GIUDICE: Thank you.
And the final presenter is Kim Wallen.
DR. WALLEN: Hi. I'm Kim Wallen from Emory
University, and the Yerkes National Primate
Research Center. I have no financial interests in
any of the drug companies involved in these
studies.
I do have an interest in female sexual
desire, and my primary research interest is in how
hormones affect female sexual desire in non-human
primates. And there, I think, are some parallels
to humans.
As a researcher who's studied this for
over 25 years, i was not surprised by the minimal
effects that Intrinsa produced in women. These are
242
not striking effects, and part of the reason, I
think, is because this is an example of a drug that
is prematurely coming to closure on something where
we don't understand the basic science.
Having looked at the literature in humans
on androgen insufficiency, I think it is quite
unconvincing that we understand what is the
hormonal basis of female sexual desire.
My concern of approval of this
drug--besides the potential health risks of this
drug--is that it will prematurely close the
investigation of understanding how female sexual
desire is influenced by hormones. There certainly
are many factors that affect female sexual desire,
but hormones are one of them, and we need to
discover what the basic mechanism is. And I think
it is quite clear from the data that Procter &
Gamble submitted for this application that they
don't know what the mechanism is. And I think from
the other published literature, it's clear we don't
know what the mechanism is.
So I would urge the committee to not
243
approve this because I think it is a premature drug
that simply does not solve what is a very important
and critical problem that we need to understand.
And it is important to resolve this issue. But
this is not the answer.
DR. GIUDICE: Thank you.
I would like to thank all of the
presenters for their thoughtful comments.
And Ms. Watkins has a statement to
make--please, before we leave for lunch.
MS. WATKINS: I'd like to remind the
committee that, in the spirit of Federal Advisory
Committee Act, and the Sunshine Amendment, that
discussions about today's topic should take place
in the forum of this meeting only, and not occur
during lunch or in private discussions.
We ask that the press honor this
obligations of the committee members as well.
We'll break for lunch now and reconvene at
1:15.
Thank you.
[Off the record.]
244
DR. GIUDICE: Back on the record.
Would everyone please sit? And would all
of the members of the committee please take their
places around the tables?
[Pause.]
This afternoon we will first begin with
questions from the committee, initially to the
sponsor, and then we will have questions to the
presenters from the FDA. And then we will address
the questions directly that were given to us by the
FDA, and give them our recommendations.
DR. GIUDICE: So I'd like to open this up
now to members of the committee. And--yes, and for
those members who came in a little late, if you
wouldn't mind introducing yourselves. This is not
to point the finger to show that you were late--
[Laughter.]
--but just to familiarize everyone with
who you are.
Dr. Emerson.
DR. EMERSON: I'm Scott Emerson. I'm a
biostatistician from the University of Washington
245
in Seattle.
DR. GIUDICE: Dr. Stanford.
DR. STANFORD: Joe Stanford, family
physician from the University of Utah, Salt Lake
City.
DR. GIUDICE: Dr. Dorgan.
DR. DORGAN: I'm Joanne Dorgan. I'm an
epidemiologist at Fox Chase Cancer Center in
Philadelphia.
DR. GIUDICE: Thank you. And Dr. Merritt
did arrive, but she's not in the room right this
minute.
Questions from the Committee to Sponsor and FDA
DR. GIUDICE: So, I would now like to open
the committee questions, and if you would just
raise your hand and--Dr. Emerson? Please.
DR. EMERSON: I guess--I don't know how you
want to interleave these, or if you want subject
matter questions, or whatever--but one of the first
questions I have is for Dr. Braunstein, who
presented reference ranges for the testosterone
levels in normal women.
246
Could you tell me what those reference
ranges represent? It was like around slide 83 of
your presentation.
DR. BRAUNSTEIN: Sure. 161 women in the
reproductive age range, having normal menstrual
cycles, not receiving any hormones at all, gave
multiple samples across the menstrual cycle. So
they weren't weighted towards mid-cycle time, or
luteal phase, or follicular phase. So they were
across the menstrual cycle.
Each individual's--the results from each
of those samples for an individual was summed, and
then the data was averaged. And these lines
represent the lower 2-1/2 percentile and the upper
97-1/2 percentile.
DR. EMERSON: Okay. Then I just want to
bring the committee's attention--slide 84, for
example--while those reference ranges are for the
central 95 percent of the data, those whiskers are
the central 80 percent of the data. And we do
have, in our materials, tables of how high the
maxima actually go.
247
But the concept of what is being presented
here is not at all comparable to the reference
range, in terms of--that the central 95 percent
would be a much, much higher range. In fact some
of those measurements go up to 100 or so.
DR. GIUDICE: Dr. Nissen. And then Dr.
Montgomery-Rice.
DR. NISSEN: Although I'm a cardiologist
and here, I think, primarily to look at the
cardiovascular issues, I did want to ask a couple
of questions about efficacy.
The main one I want to understand is:
you've gone to a lot of trouble to validate tools
for assessing this. But what I didn't see anywhere
in here--and I'd be interested in whether you can
provide us with any information--on what the effect
is of non-hormonal interventions; that is, if you
take women with this disorder, and you give them
counseling, you give them other kinds of supportive
therapies that don't involve giving a systemic
hormone, how much improvement, on the same scale,
do you get?
248
You know, we saw about a 5 percent
increase, for example, on one of your scales.
Well, I'd like to know what that is for
non-hormonal interventions. And if anybody can
answer that I think it would be very helpful to me
to put this efficacy into the context of the safety
issues.
DR. MEYER: Although we did not ever test
our scales in any other type of interventional
therapy, I think Dr. Jan Shifren can give us some
additional perspective on the efficacy of other
types of interventions.
And one thing I did want to mention--as
Dr. Shifren will reiterate--is the Intrinsa patch
is not for everyone. As she has already told us,
there are other types of therapies that are
appropriate and have worked for women prior to
androgen therapy.
DR. SHIFREN: I am not a sex therapist.
I'm a reproductive endocrinologist. And certainly,
for women with non-physiologic causes of HSDD,
other interventions, such as counseling, sex
249
therapy, education, and lifestyle changes can be
very effective.
What's important to realize is that the
women in our trial--and, again, it was a select
group of women. These are the women we would like
to see the patch used in--all had healthy sexual
functioning before surgical menopause, and reported
significant decreases in desire and activity
associated with distress, following surgical
menopause.
To be entered in the study these women
couldn't have depression or relationship conflict,
or all of the other things that the other
non-pharmacologic interventions are so effective at
treating.
DR. NISSEN: I don't think you answered my
question, though.
My question is: what do we know about
whether those therapies are effective, and what is
the magnitude of their efficacy? You have a drug
therapy here, and there are other therapies out
there. And so I need more information to
250
understand how large is this efficacy in relation
to what can be offered women via conventional,
non-drug therapies.
DR. SHIFREN: I think it really depends on
the cause of the sexual dysfunction. There are
studies that make it clear that counseling and sex
therapy are very effective therapies, and they
should be used widely. But when you select a group
of women for whom the major issue is, let's say,
surgical menopause, those are less effective.
The other thing I did want to point out is
that our placebo-treated women--it's hard to say
that was a placebo. Wearing a
non-testosterone-containing patch is a very active
intervention. And in some ways, the response in
the placebo-treated women may almost address your
question. These women were receiving active
counseling by the physicians and the nurses that
they saw regularly at the study site. They were
wearing a patch, which was a clear reminder to both
the patient and her partner that she was concerned
about her sexual activity, that it distressed her,
251
and that she was committed to making it better.
So that just being in the non-testosterone
treatment arm was an active intervention.
DR. GIUDICE: Dr. Montgomery-Rice.
DR. MONTGOMERY-RICE: I want to go back to
slide 83-84, where Dr. Braunstein compared it to
some data that Dr. Soule presented for the FDA.
Because I'm confused--and maybe, Scott, you can
help me interpret this.
When I look at slide 84, and it talks
about his median free testosterone level, and we
are still talking about a level of 7 picagrams, it
doesn't appear that there are that many subjects.
But you're saying that that is 80 percent of the
subjects had a level that was greater than that.
So would that correspondent, then to what
Dr. Soule is saying, where she points to data in
slide 8 of her presentation, that more than 35
percent of the people were beyond this median value
of 8.6
DR. EMERSON: So, if I could clarify what I
was trying to get across--in the "normals" that
252
line that is drawn at 7, only 2-1/2 percent of
normals are above that line. In the larger
study--you know, we have 500 subjects in this--they
have 10 percent of the subjects above the top
whisker there.
And so we're seeing that there's
something--and what Dr. Soule presented would be
representative numbers, suggesting that between the
dashed line and the top whisker might be 10 to 15
percent of the patients. And then another 10
percent are above that top whisker.
But the point I was trying to make is that
this graph is actually quite misleading, in trying
to give the impressing that most of the
measurements--that the difference between the
normal range and what they observed isn't that
great, when they're using two different measures
there.
DR. BRAUNSTEIN: In actuality, at any time,
the patients on testosterone, about 15 to 20
percent or so were above the upper limit of the
free testosterone level here. Again, this is a
253
reference range: young, pre-menopause women--young
women. And we're using that just to show you that
the majority of women fall within that range. Some
will go above that range--but, certainly, not so
terribly above the range that one would get serious
concerns.
For instance, when Dr. Dobs was talking
about the therapeutic doses that are used in women
who are transgendered to males, they receive very
large doses, and the levels are quite a bit higher.
If you look at female athletes that are abusing
androgens, the levels are substantially higher.
DR. EMERSON: Well, but the scale--if you
went up to that 97-1/2 percentile on that--I can't
say exactly what it is, but your maximum, in the 24
week is closer to 100. Okay. So the scale of
that.
Now, again, maybe the majority of patients
aren't up there, but some patients have very, very
high levels of free--
DR. BRAUNSTEIN: Yes, there were a total of
11 patients that had a level of 21 picagrams per
254
ml. And of those 11 patients, there were, I
believe, 7 who had androgenic adverse events,
primarily hirsutism. And there were no other
significant adverse events--no liver function
abnormalities, kidney function abnormalities,
etcetera. Again, they were androgenic types of
adverse events--the types of things you would
expect with very high free androgen levels.
And if we look at the highest decile--the
highest 10 percent--really, there's no statistical
difference if you look at the women who were in the
highest decile versus those who were on the
placebo.
DR. EMERSON: Where you sample size is
already down to--
DR. BRAUNSTEIN: The sample size, the upper
decile is small.
DR. EMERSON: And the other thing that I
would also like to point out on this is that you
had significant dropout in your patient population.
20 percent of your patients dropped out during the
initial six months. And then at each stage, as you
255
went from the 6 months to 12 months, 12 months to
18 months, more patients dropped out. And so
actually the highest level that you observed in
free testosterone was in an intermediate
measurement. And we can imagine that perhaps that
patient dropped out because of that high
level--perhaps not. We don't have that
information. But that's something to keep in mind.
DR. BRAUNSTEIN: Well, in actuality, there
was no relationship between when the highest level
was found--because it could have been found at any
time the bloods were sampled during the
trial--versus when the androgenic adverse event,
for instance--if one appeared--did appear.
Could I have that last slide, please?
Just to show the point in the upper
decile--this is 6-99--you can see that in the
comparison to the placebo group, the patients in
the upper decile of free testosterone, 10.3 percent
had acne, versus 7 percent in the placebo group;
3.5 alopecia, versus 2.7; 5.2 facial hair versus 5;
no voice deepening versus 1.7.
256
So, again, the reference range--and your
point is very, very well taken--but the reference
range is just to provide you sort of a baseline
sense of security. This is not being--you know, we
aren't talking about this as a replacement therapy.
This is a drug therapy--but to show you the
relative levels of androgens achieved with Intrinsa
versus the normal physiologic range in reproductive
women was the purpose for establishing this
reference range.
DR. EMERSON: Do you know whether the
patient who had a measurement at 107.7 at 24 weeks
is represented in the 52-week population where the
maximum was 63?
DR. BRAUNSTEIN: I don't know that.
DR. EMERSON: Thank you.
DR. BRAUNSTEIN: But we can find out and
come back to you on that.
DR. GIUDICE: In order: Dr. Lipshultz, Dr.
Dr. Dickey, Dr. Judice, Dr. Stanford, Dr. Hager.
And we will go around the table.
[Laughter.]
257
So--Dr. Lipshultz.
DR. LIPSHULTZ: Yes, I have a question.
We see, repeatedly, a comparison of the
TTS-treated to the placebo-treated. But I'm
interested in Dr. Davis' slide, in which he showed
the frequency of satisfactory sexual encounters
over a four-week period, comparing the placebo to
the treated to the normal. And my question is: if
these people had a normal sexual relationship prior
to the oophorectomy, and following oophorectomy
they had androgen replacement therapy, why are they
not going back to their pre- to their normal--quote
"normal"--state. If the answer is androgen
replacement, then why aren't they back to normal,
rather than back to one increased encounter in a
month?
DR. MEYER: Well, if I could have slide
252, please--although I don't know what their
normal level may be.
But what this shows you is "total
satisfying activity," "desire" and "distress."
Here's the baseline level of the women in our
258
study--in the surgical menopause study. Here's the
normal level that we got from our instrument
validation studies in total satisfying activity in
a four-week period.
The self-identified responders from our
clinical relevance study got about 50 percent of
the way back to the normal women--these are
different women--but the normal women in the
validation.
DR. LIPSHULTZ: Exactly, and my question
is: what is your hypothesis, if this is due to
something related to the oophorectomy, and you're
replacing what you think is the target hormone, why
wouldn't they go back to the validation level?
DR. MEYER: Because these are different
women. Although we did not ask them what their
normal level was. They could be back at their
normal level.
The women who are self-identified
responders in this study had a mean satisfying--an
increase in satisfying sexual activity of 4.4
activities per month. So that's one per week.
259
DR. LIPSHULTZ: Right. But I think--
DR. MEYER: So that would get them back to
seven per month, which might be their normal,
although that is lower than the women in the
validation study.
DR. LIPSHULTZ: But isn't it important for
you to have that data? I mean, as to what you
consider normal sexual activity for that group of
women? Otherwise, as was pointed out by several
people from the audience, that, you know, the
androgen may not totally explain what's going on
here, or else you would expect to see a
normalization.
DR. MEYER: And Dr. Shifren can address
this. But I think another important thing to look
at is the distress. So, again, it's not just the
activity, but--
DR. LIPSHULTZ: I'm not disagreeing that
the patients are better. My question is why aren't
they back to normal if the answer is androgens.
They should say, "I feel exactly the way I did
before because you've replaced my androgens."
260
DR. SHIFREN: I may be able to help a
little bit here. I was not involved in the U.S.
validation study. But I can tell you, from my
typical healthy sexually functioning menopausal
women, 12 events in four weeks is actually quite a
lot. That's three times a week, and that's
actually more than what we consider the typical
average for even pre-menopausal Americans.
So I can't really speak to the validation
study and normalization. But I think it's very
important to realize: this isn't blood pressure.
And there actually is no "normal." And it's
actually normative that as couples age, with both
duration of the relationship and aging, that
frequency declines, but satisfaction can remain
very high.
So there's not really a good correlation,
for most couples, between frequency and
satisfaction, and distress. The very first thing I
say to almost all of my patients when they come in
with sexual complaints is: "Let me start by saying
there is no 'normal' frequency or set of behaviors.
261
And that the typical American experience has very
little to do with Sex in the City. And if you are
comfortable in your relationship, and it's working
for you and your partner, that's all that matters.
And if it means that you're doing non-intercourse
events because your husband has erectile
dysfunction, but it's a loving, close and intimate
relationship and you have no concerns, there's
nothing to treat."
So a lot of what I do is really validating
for women that there is no "normal" for sexual
function. The most important thing is that it's
working for the woman, and that there's no
associated distress.
DR. GIUDICE: Dr. Dickey.
DR. DICKEY: I'm going to change gears a
little bit. I have a question for Dr. Steinbuch.
I think that's how you pronounce it.
I'd like a little more information about
the proposed long-term safety plan. It would
appear that you're going to collect your
262
information from claims data. And I'm a little
concerned about privacy issues, and a little
concerned about whether claims data are actually
going to give you the kinds of information that
we've heard about through the morning, in terms of
risk factors.
DR. STEINBUCH: Yes, with regard to privacy
issues, the Ingenix database--as I described
earlier--is comprised of all of these claims of
patients throughout the United Healthcare System.
And there really is no privacy issue there in
regards to the fact that every report that we get
is de-identified. So there's no--we can't identify
any individual per se. At least our company.
Ingenix, when they go after the medical
records for abstraction, it will be their company
employees who would be actually going out and
getting the information. So there really shouldn't
be an issue there.
And, as I said, we're going well beyond
claims, because we're going to be contracting with
them to actually go after medical charts for review
263
of all relevant events of interest.
DR. DICKEY: And there's no privacy issue
there?
DR. GIUDICE: This is an issue, I think,
that needs to be further discussed--the whole issue
of HIPPA. And I think what Dr. Dickey is getting
at is can you explain to us exactly what the
process is, so that patient privacy is not
violated.
DR. STEINBUCH: I think it would be best if
Dr. Alec Walker, who's here in the audience, who
represents--he's a senior vice president of
Ingenix. Perhaps he could address this issue with
regard to how it works in the United Healthcare
System.
Dr. Walker?
DR. WALKER: We have a number of
FDA-mandated post-marketing safety
studies--Alexander Walker, Senior Vice President,
Ingenix.
We have, I believe, five of these studies
going on now. They are HIPPA compliant. They are
264
done under IRB approval.
The data are manipulated in a
de-identified fashion, to the extent that they can
be. Then, at the last minute, if you have a
potential event that you want to look at from the
claims data, then it's de-identified to the extent
necessary to go to the medical record, and it comes
back.
DR. DICKEY: So, patients will give
permission to participate in the study?
DR. WALKER: No. Under HIPPA, with an IRB
approval--or, indeed, for an FDA-mandated
study--you don't need individual informed consent.
You do need the IRB or privacy board to examine the
protocol and procedures to verify that there's
adequate confidentiality in force and maintained.
DR. DICKEY: Thank you.
DR. GIUDICE: I have two questions. One
has to do with the data and the analysis. And
perhaps Dr. Meyer can answer this.
Have any of the data been analyzed with
regard to BMI, because it appears that there's one
265
dose, and women--a large range of BMI. So can you
give us any insight into that?
And the second is that there have been
several comments about potential risks g
with regard to breast cancer and cardiovascular
disease. And in our briefing documents I did not
see any pre-clinical data. And I'm wondering if
you could share with us any insight into that?
DR. MEYER: Certainly. With respect to
BMI--and we did break out these data by sub-groups,
and I think they are in your briefing book--but in
women with a BMI less than 25, between 25 and 30,
and greater than 30, the data were essentially the
same--the response on sexual desire. And for
satisfying sexual activity, the women in the
highest BMI group--that greater than 30--the
response was slightly less--the median response was
slightly less than those in the lower BMI groups.
DR. GIUDICE: And what about free
testosterone in those groups?
DR. MEYER: Let's see, I think Dr.
Braunstein--I don't know that we broke--we can
266
check. We'll look and see if we broke out free T
by BMI.
DR. GIUDICE: Okay.
DR. MEYER: And with respect to
non-clinical data in breast cancer, Dr. Mike Winrow
is our non-clinical toxicology specialist, and has
done an extensive literature review of all that's
known on the non-clinical data of testosterone.
And although-- there are lots of in vivo and in
vitro studies that he can address that have looked
at that.
DR. WINROW: Thank you. If you look at the
labeling on current testosterone products, they're
all labeled as potential carcinogens. That's based
on data generated by the International Association
for Research on Cancer, based on rodent studies
with very high doses, many of them administered to
neonates, over a period of years.
There's no doubt that at very high doses,
in rodents, that can occur.
The more interesting data--and this is
fairly recent data, and it follows from the
267
discussion that came this morning on the work being
done at NIH--if you're looking at rodent data
you're dealing with intact animals, estrus cycle
not menstrual cycle, and all the other complicating
factors--and extremely high doses.
What's being done at NIH, and published by
Dr. Demitri Kakis, is shown on slide 644--if you
could show that, please. And in this study,
they've used probably the best model that could be
used, which is an ovary-ectomized Rhesus monkey.
And what they've done ovary-ectomized Rhesus
monkeys, and then supplemented those animals with
either estrogen--in the second line down--or with
estrogen-progesterone, or
estrogen-plus-testosterone.
I realize the numbers are small, but
recognize that using Rhesus monkeys is not a
trivial event these days. Their intention was to
replace these hormones at levels that would mimic
the pre-ovary-ectomy levels. So they are within a
reasonable range of normal for both non-human
primates and primates.
268
What they were looking at was
proliferative response in breast tissue, which they
were using as a marker for potential
carcinogenicity.
It's a short-term study. But, at the same
time, the numbers are quite interesting. They used
this KI-67 antibody, which picks up an antigen that
only is produced in proliferating cells. And with
hormones, continuous cell proliferation is a
requirement for the development of cancer. And so
you normally see hormonally-induced tumors in
tissues that respond in a proliferative way to the
particular hormone under consideration. So if you
get a lot of testosterone, you'll get prostate
cancer--as an example.
And, as you can see here in these female
Rhesus monkeys, the controls had KI--that's a
percentage of cells that are proliferating--of
about 8 percent. If you--following ovary-ectomy,
if you then administered estrogen, that number shot
up to 30.
So, administering estrogen to
269
ovary-ectomized monkeys produces a very significant
increase in cellular proliferation in breast
tissue. If you added progesterone to that--very
little difference; not surprising because the
progesterone's really there to address endometrial
issues. If you added testosterone, however, that
level dropped to 16.7, which--while the number is
small, statistically it's no different from
control. But, of course, as I say, the numbers are
small.
So, adding testosterone to ovary-ectomized
monkeys, in the presence of estrogen, significantly
reduced the level of proliferation produced by
estrogen alone, and moved them back towards normal.
The other things that this group have done
along the same lines, they ran another study where
they administered flutamide to intact Rhesus
monkeys for a period of six months. So what
they're doing there is block testosterone
chemically, but leaving all the other hormones
alone.
In that case you get a doubling of
270
proliferation in breast tissue. You've removed the
testosterone, which is no longer offsetting the
proliferative response of the estrogen, and so
proliferation rate goes up--in the absence of
testosterone.
Two other studies that they've run, and
looked at other biological endpoints, they've shown
that in these testosterone-treated animals, you can
get a reduction in the expression of the oncogene
MIK, and you can also see a change in the estrogen
receptor alpha-beta ratio away from the ratio
that's considered of concern for breast cancer.
So in these small studies, using
ovary-ectomized Rhesus monkeys, all the data point
towards testosterone not having a proliferative
effect, or not raising the level of concern for
breast cancer. In fact, it's the exact opposite if
you look at the numbers.
And it was on the basis of this work that
they did the clinical study in the Australian
population which was published fairly recently.
So, you've got those two different sets of
271
data.
DR. GIUDICE: Thank you.
Now, starting over here with Dr. Stanford.
DR. STANFORD: Two questions--the first
one, back to the reference range for free
testosterone from the 161 pre-menopausal women.
You mentioned hat hey were only women with
regular cycles. Was there any other effort to
exclude those that may have indicators of metabolic
syndrome? My concern is that if you just take 100
women off the street and look at testosterone
levels, some of them may have unhealthy levels,
because some estimates are that PCOS with
hyperandrogenism may be around 10 percent of the
population.
So we may not have a healthy range,
perhaps. I wondered if any other exclusion
criteria were applied to those 161 women. That's
my first question.
DR. MEYER: And I would like--Dr.
Braunstein can address the exclusion criteria, and
then Dr. Ricardo Azziz is with us, and he's an
272
expert on PCOS and normal androgen levels in women.
And I would like him to come up and address what is
normal and not a normal androgen level in women.
DR. BRAUNSTEIN: Actually, the inclusion
criteria was that they had normal menstrual
cycles--ovulatory menstrual cycles. So if a woman
with PCOS was in there, it would be one who was
having normal ovulatory cycles.
I do think, though, since this issue of
PCOS has been raised--and that condition is
different from the model of giving back
testosterone to oophorectomized women, I think it
would be important to have Dr. Azziz address that
issue.
DR. AZZIZ: Just as a disclosure, I own no
stock and I was not a participant in these clinical
trials. I'm simply a consultant as an
androgenexist I direct the Center for Androgen
Related Disorders at Cedars-Sinai Medical Center,
and I'm faculty at UCLA.
There's a number of concerns and questions
that the committee has brought up very well, and I
273
think these need to be addressed. One was the
issue of polycystic ovary syndrome as an example of
an androgenized patient.
I need to make sure that the committee
understands that polycystic ovary syndrome has an
underlying insulin resistant syndrome etiology.
Over 70 percent of the patients with polycystic
ovary syndrome are insulin resistant, and our
research, and that of others, indicates that it's a
cell-signaling defect that actually causes--a
cell-signaling defect in the insulin-signaling
pathway that causes PCOS.
So polycystic ovary syndrome is actually
not a very good example of a hyperandrogenic
effect. And one of the things that I think is
important to not bring up is not to mix in the
insulin-resistant syndrome or PCOS in this
discussion.
In fact, androgens have very little impact
on glucose metabolism in and of themselves, which
is why they're not the etiology for the
insulin-resistance in PCOS.
274
Secondly is the issue of androgen levels
in normal women. It is important for the committee
to understand that androgens, as opposed
to--say--thyroid are not regulated very closely in
a human body. You can quadruple or actually
increase the androgen levels 10-fold in humans--and
that's males and females--and LH levels, which are
the primary responder, do nothing. They change
absolutely nothing--which is why, in humans, the
normal range of androgens in males can go from 150
nannograms per dL to 1,000 nannograms per dL. And
these are all normal males here in the audience,
and the same thing for women.
So one of the reasons that the levels of
"normal" which were presented as a normal example
are so wide. And absolutely correct--Dr.
Emerson--that, in fact there are a number of
patients that are above this--quote "normal limit"
is that, in fact, physiologic effects of androgens
do not correlate directly to the levels; and, in
fact, there's a wide variability in androgen levels
in the normal population.
275
Evolutionarily, there has been no pressure
to select people with high or low androgens,
otherwise women, of course, would be extinct, and
so on and so forth.
[Laughter.]
And it is not the case.
So it's important for the committee to
understand that this is not regulated like thyroid.
Androgens go up, LH goes down. That is not the
case in humans.
The last one is the issue of a "normal"
population. A number of issues have been brought
up related to what is a "normal sexual function,"
and why--and Dr. Lipshultz brought up very
clearly--why isn't it that these people didn't
become "normal."
I should point out that part of what we
do, of course, is look at normal sexual function.
And part of our center looks at androgen
deficiency. 49-year-old women--couples--who are in
their 50s do not, on average, have 12 intercourse
encounters a month. The vast majority of surveys
276
of sexual function in the United States today put
the number of intercourse activities--or acts--in
normal couples in their 50s at between 1-1/2 and 2
encounters per week, which is somewhere between
seven and eight per month.
In the validation study you had to recruit
people who were willing to talk about sex. And
that may have biases the--quote--"normalicy." So
it's not actually proper to actually compare the
small population of people used in the validation
study in this study to the response. We need to
look at the response of our patients here to what
is normally assumed to be normal in the U.S.
population. And in 50-year-olds, that's somewhere
between seven and eight encounters a month.
I don't know if that answers some of the
questions.
DR. STANFORD: Thank you.
My second question was for Dr. Steinbuch
about--if I gathered correctly, you're estimating
that you'd capture about 5,500 women would be
prescribed Intrinsa in the Intrinsa in the
277
long-term safety monitoring plan.
What are you projecting that they would be
prescribed Intrinsa for? Would these all be
surgically menopausal, or would you be mixing
surgically menopausal and naturally menopausal
women? And how would you address that issue?
That's one of a number of issues, I think,
about this safety monitoring which are really key
to understand--whether it would be adequate.
DR. STEINBUCH: Yes, the 5,500 estimated
number of Intrinsa users per year is a combination
of both surgical and natural menopause women. And
that's actually one of the advantages of the
observational setting, in that all women who would
be prescribed this medication would be included in
the analysis, and there would be no group that
would be precluded from inclusion in the sort of
full statistical analysis at the end of each time
period.
DR. STANFORD: Okay. So maybe I could ask
just one other question.
Would chart reviews be done on all of
278
those women? What's the trigger for a chart
review? You mentioned having the company from
United Healthcare go back--Ingenix, I think, go
back and do the chart reviews.
What would be the trigger for that?
DR. STEINBUCH: As I indicated, there would
be a three-to-one match, and the chart review would
be triggered by any one of the events of interest
that will be ultimately determined by the panel
would be involved, and the independent safety
review board. And then once that final decision
has been made, any of those events would trigger a
medical chart review--throughout the system.
DR. GIUDICE: So just to follow up with
regard to the patient population, it will be
surgical menopause, natural menopause, with and
without a uterus? And therefore the treatment will
be Intrinsa alone? Or also estrogen and also
progestin? I mean, what's the plan?
DR. STEINBUCH: Yes--well, the plan is to,
as I said, bring all women who receive Intrinsa
scrips--whoever they may be. In terms of the
279
matching criteria, there would be some very careful
consideration to do appropriate matching and for
perhaps stratifying by if they're estrogen only, or
estrogen-plus-progestin, that might be a reasonable
thing to match on, for example.
DR. GIUDICE: Dr. Hager, did you have a
question? And then Dr. Macones, and then Dr. Rice,
and then Dr. Lewis.
DR. HAGER: I have three fairly brief
questions.
Regarding the anchoring technique--and I'm
not an epidemiologist, and don't claim to
understand that well--but as I understand it, you
evaluated data and you had a fairly significant
dropout that increased as you progressed through
the study. You used the anchoring technique, and
you used the last available interview among those
who dropped out to go back and recapture that
information. Is that correct?
And in so doing, do you have information
from those individuals before you recaptured that
information? Do you have data up to that drop-out
280
point? Because you had such large
drop-outs--without extending that to 24 weeks? If
a patient dropped out--a subject dropped out at 12
weeks, you captured that information and extended
it out to 24 weeks--is that correct?
DR. MEYER: That's correct. We used the
last observation carried forward.
DR. HAGER: Okay--so do you have the data
without that extension?
DR. MEYER: Yes, we do. And when we looked
at both ways, the very conservative LOCF method to
get the proportion of responders. And Dr.
DeRogatis showed you--what?--that was about 51, 31
percent.
When we take out the people who dropped
out and do just the protocol analysis, we get the
same--
DR. HAGER: Okay.
DR. MEYER: --effect.
DR. HAGER: Okay. And can you tell me why
the African-American and Hispanic population was
under represented in this study?
281
DR. MEYER: Yes--it's notoriously difficult
to try and recruit a sufficient number of
minorities into all clinical studies that are truly
representative of the patient population. And we
knew that going into this. And we were aware of
the data in African-American women on hysterectomy.
So we took extra care to try and recruit as many
minorities as possible. For example, we did a
patient recruitment in the media, and we would
target media that targeted various minority
populations. We talked to minority investigators.
We took as many steps as we could think of to
recruit minority populations. And 6 percent is not
representative of the U.S. population--although
better than a lot of clinical trials I've been
involved with.
But, again, we chose sites with access to
large numbers of African-Americans and Hispanic
people. We had all our instruments translated into
U.S. Spanish. We ran ads in the media--again--that
targeted minorities.
And so one of the things that we're doing
282
at P&G is we have an ongoing effort to continue to
try and increase our ability to recruit minorities
into our clinical studies.
DR. HAGER: So are you concerned about the
long-term follow-up in those sub-populations?
DR. MEYER: Well, again, with the large
observational study, if we move forward with that,
these minorities, if given a prescription, will be
able to be followed, and in larger numbers.
And the other thing that we have done is
also we have discussed our plans with a variety of,
for example, African-American clinical
researchers--OB-GYNs--to try and understand our
data in the context of these patient populations.
DR. GIUDICE: Dr. Macones?
DR. MACONES: A question for Dr. Steinbuch,
please.
I actually have two questions. First, to
follow up on Dr. Hager's point--the UHC data that
you're going to be using, can you tell us about the
ethnic background of patients that are included in
that data set? Is it a very generalizable group?
283
DR. STEINBUCH: The Ingenix database?
DR. MACONES: Yes, I'm sorry--from your
agency.
DR. STEINBUCH: Yes, it is, actually.
We've looked at some comparisons to the U.S.
Census. And demographically, they're reasonably
comparable.
DR. MACONES: My second question was about
your sample size estimate for your post-marketing
study, which you had on your slide number 120.
And my concern is that just at first
blush, I thought that a study of 5,500 patients was
going to be pretty small to look at some of the
events that you're going to be interested in.
And it made me wonder about some of your
assumptions. And the one that struck me the most
was your event-rate per year of .15 percent. And I
believe you said that that was based on WHI data
for cardiovascular events in women who were 50 to
59. Is that right?
DR. STEINBUCH: That's correct.
DR. MACONES: That's a curious choice,
284
because I think in your clinical trials about half
of the patients were less than 50.
So I wonder why you chose an estimate on
the older population, rather than being more
conservative, choosing perhaps a lower event rate,
having a bigger study, with more power and more
precision in the older patients.
I mean, I look at this now and I could
predict it's going to be a negative, under-powered
study. And that's a concern.
DR. STEINBUCH: Actually, the event rate
that we used was the lowest that was possible
within the WHI. As the FDA has indicated, they've
been using the WHI as sort of an anchor for this.
And we thought that was the best that we could do.
If you could please put up slide 691?
[Slide.]
This slide shows, broken out in the WHI,
broken out by decades here, and when we look
at--and this is for the estrogen-only arm--you
could see that this is about where the line is for
hazard ratio of 1. In the 50 to 59, most
285
of--either at or below 1 for this younger group.
Now, with regard to what the FDA has shown
and shared is I believe they've been using the E+P
for the full 50 to 79 year age group. And so,
since our mean age was about 50, this was the
closest that we could get using WHI, that would
actually be the closest estimate.
Does that make sense?
DR. MACONES: It does, but I disagree with
it. Again
DR. STEINBUCH: Okay.
DR. MACONES: I mean, the issue is that
you're going to have no power to look at events in
younger women, and that's likely to be half of the
population.
I mean, again, based on the enrollment in
your clinical trials, which the mean age was 50,
which roughly means that about half of the people
were less than that. And you're just not going to
have a lot of power to look at those patients. And
if you had a bigger study--again, power to look at
events in younger patients--you're going to have
286
lots of power to look at event rates in the older
population.
DR. LOCKWOOD: But to be fair--I mean, we
don't want to get into a debate about WHI--but if
you look at the data--and the argument's been made
that, in fact, it may be protective in that younger
age group. Look at the odds ratios--or the hazard
ratios that are presented there. They're not quite
significant, but there's certainly a strong, strong
tend toward a protective effect between 50 and 59.
So I think, to be honest, that the group
that you're thinking may have to be much larger to
detect a potential adverse event in fact might be
the opposite. It might be protective and it might
be additively protective.
DR. MACONES: I think the point is "might
be." And we don't know the "might be."
And I think for charged issues like this,
I think we're better off being conservative and
designing a bigger study rather than taking a
chance and designing a smaller study that might
miss an important effect.
287
But I appreciate the point that you make.
DR. GIUDICE: Dr. Dorgan.
DR. DORGAN: First of all, on this
question--could you tell us what the power to
detect increased breast cancer among these women
would be?
DR. STEINBUCH: Yes.
Could you please put up slide 507?
[Slide.]
And as was mentioned earlier with regard
to latency, I would direct your attention--the
event rate here was again using WHI, 50 to 59 year
age; .3 percent per year. All the other
assumptions were the same as before, with regard to
disenrollment, etcetera--discontinuation. And
getting down to the four to five year, which I
think is the most reasonable place to be looking
here, we have an 84 percent power at four years to
detect a relative risk of 1.4 with regard to breast
cancer.
DR. DORGAN: Okay. I have another question
that's based on the information in the briefing
288
book. In the briefing book it states that
"--epidemiologic studies examining the relationship
between endogenous testosterone and breast cancer
risk have yielded equivocal results." And the
reference for this is the Endogenous Hormones and
Breast Cancer Collaborative Group paper in 2002.
Now, as a co-author on that I was kind of
surprised by this interpretation of the findings.
I brought a copy of the manuscript with
me. Testosterone was very strongly, actually,
related to breast cancer risk in these women.
Women in the highest quintile of testosterone
levels--and this is within the normal range for
post-menopausal women. So it's much lower than
what we're talking about here.
Women in the highest quintile were at
two-fold--let's see exactly--2.2-fold increased
risk of developing breast cancer. This was highly
significant. And the trend was also highly
significant. With increasing levels of
testosterone, risk increased significantly, going
from the first quintile is the reference of 1, to
289
the second quintile of 1.3, to the third quintile
of 1.6--up to as high as 2.2 in the fifth quintile.
Again, I want to say that these levels are
endogenous levels in post-menopausal women. They
are much lower than the levels that we're looking
at here.
When we're talking about the concern for
the women whose levels are above the 90
th
percentile--based on that graph you keep looking
at, bioavailable testosterone. When we're focusing
only on the women whose levels are above the 90
th
percentile for pre-menopausal women, I think we're
missing the point.
The women who are even at the median, they
had levels, based on some of the data that you
provide us, that were three, four and sometimes
five-fold higher than your control placebo group.
This could translate into an increased risk of
breast cancer in this group as a whole, of maybe 70
percent going up to--it possibly could even go up
to a doubling of risk.
Could you comment on that?
290
DR. MEYER: Dr. Robert Reid, Chairman of
the OB/GYN department at Kingston will comment on
this.
DR. LOCKWOOD: On that, could I ask you a
question?
DR. DORGAN: Sure.
DR. LOCKWOOD: Did your study adjust for
potential confounding from BMI?
DR. DORGAN: We didn't adjust for BMI, but
we did actually adjust for estradiol, because of
concerns that the testosterone--this is endogenous
levels, we were concerned that there might be--the
testosterone levels might be just elevated
secondary to estradiol, and all the effect might
really be due to estradiol.
But when we did that--
[Pause.]
--sorry, I have to flip through
here--okay--and this is looking at a doubling, as
opposed to looking at quintiles.
So, if testosterone was looked at
unadjusted for estradiol, a doubling of
291
testosterone was related to a relative risk of 1.42
for breast cancer. So it's about a 40 percent
increased risk of developing breast cancer. It was
statistically significant.
Estradiol, on its own--the doubling of
estradiol--was related to--had a relative risk of
1.31--or a 31 percent increased risk.
When you include both testosterone and
estradiol together, the increased risk associated
with the doubling of testosterone was decreased a
little to 32 percent; whereas the increased risk
associated with a doubling of estradiol was
decreased to 18 percent. So it's not being
explained solely through estradiol.
DR. LOCKWOOD: I guess the big issue,
though, would be if these patients are obese
they're likely to have both elevated estradiol and
testosterone levels, and to be at higher risk,
independent of either, for breast cancer because of
the obesity. It increases aromatase activity,
potentially, in the breasts themselves.
So, you know, I hear you, but the
292
potential for confounding is so great that it does
call into question those--
DR. DORGAN: I disagree. I think that--I
don't know where you're coming--most of the
association of obesity, we have shown in subsequent
work, we can explain a lot of the association of
obesity with breast cancer risk by effects on serum
estradiol levels. And since we have adjusted for
estradiol in these analyses, I think that it's
showing that what appears to be a significant
independent effect of serum testosterone with
breast cancer in post-menopausal women.
Whether this effect is due to
aromatization of the testosterone in the breast, we
don't know. But I'm saying that these data surely
don't suggest that the results are
equivocally--epidemiologic studies that look at
associations show a very significant and strong
association of serum testosterone levels with
breast cancer risk in post-menopausal women.
We don't know if they're causal. You
would need a clinical trial to establish that.
293
DR. GIUDICE: Yes--please let's have a
response.
DR. REID: May I enter the foray?
I don't think we have a clear answer to
the question you've raised. I mean, the criticism
of studies that have looked at a single isolated
testosterone value in people who subsequently went
on to develop breast cancer have been challenged
based on problems with the assay sensitivity in the
range for women, because it's at the low end of the
assay. The fact that it's a single value, and
there may be variations due to a variety of
different life events and stresses.
The issue about aromatase activity in the
breast--and biopsy studies around breast cancer
have shown that there's often very high aromatase
activity in the quadrant where the cancer is
compared to other quadrants of the breast that
don't have a cancer. So local effects may be much
more profound than what you see in the circulation.
So it's a finding that merits
consideration and concern, but, you know, the other
294
types of data that we see, for example, in Rhesus
monkeys, showing an inhibition of proliferation, is
just the opposite. And it's very reassuring in
that context.
And I can't comment more than that on that
specific issue. I could make some comments about
relative risk and some of the other things that
affect breast cancer risk if it's of interest,
because I've heard a few things presented
about--the WHI is constantly being cited here, and
we heard a number of explanations for why the
observational data did not match the randomized
clinical trial data for cardiovascular disease.
But the one explanation we didn't hear put
forward--we heard about volunteer bias, and health
user bias and so on--there's been a lot of
discussion in the literature about the fact that
women who were involved in the WHI, many of them
were several years to many years post-menopausal.
And probably the biggest difference between
observational and randomized trial in that
circumstance that would explain, to a large degree,
295
the cardiovascular outcomes is, in fact, the
difference in age of the populations. There are
observational studies from Nurses Health Study was
in women who were 50 years of age or younger. And
WHI, as you know, two-thirds of them were over 60
in the combined arm. In the estrogen-only arm, 50
percent of them were 70 or older when the study was
closed.
So that's a very old population compared
to the younger women. I think it's a point that
maybe is lost in some of the discussion.
Would you tolerate a couple of slides? To
clarify it? Or not--it's up to you.
DR. GIUDICE: Io think we need--we have
many other questions, and we need to move on.
DR. REID: Okay.
DR. GIUDICE: So--in the queue--thank
you--is Dr. Rice, then Dr. Lewis, and then Dr.
Tulman--and others.
DR. MONTGOMERY-RICE: I have a couple of
questions.
One of the things I want to make sure I
296
understand, that in the questionnaire where you ask
about the satisfying sexual activity, that
"satisfying" was being used as an adjective,
meaning that were the women asked, well, did they
have a sexual activity? Did they have sexual
activity and maybe it wasn't satisfactory? Or did
they only have the option to check that they had a
satisfying sexual--
DR. MEYER: No, they had the option for
both.
DR. MONTGOMERY-RICE: So they did have the
option for both.
DR. MEYER: Activity and satisfying
activity--
DR. MONTGOMERY-RICE: Okay, so--
DR. MEYER: And we also, in the instrument
validation, took great pains to make sure that
these women understood what a "satisfying sexual
event" was. It's not handholding. And we
validated that in the--
DR. MONTGOMERY-RICE: Okay. So, but when
women said they had a satisfying sexual experience,
297
you took the mean--used means for your baseline and
for your increases. So this 1.9, or whatever we're
seeing, is a mean over all of--so how would an
individual have rated that if that one increase in
the number of sexual events was satisfying to them?
Was that enough for them to rate that they had
improvement?
And I ask that question because when I
look at Dr. De--ahh-
DR. MEYER: Dr. DeRogatis?
DR. MONTGOMERY-RICE: Right--in his slide
number 43, when you asked this question of interest
in continuing the treatment, if this was a
meaningful experience for them, and that a large
percentage of them got 1 point of that increase in
activity made a difference, how am I to interpret
when I see 70 percent "definitely not," or 60
percent "definitely." Tell me how I am to
interpret that, if these people really rated--they
had the option to rate whether they had sexual
activity, or whether it was satisfying? Why
weren't more people interested in continuing?
298
DR. MEYER: Oh, actually, what this shows
on this graph--in the white are the women who
reported no meaningful benefit. Now this--again,
the data were blinded to everyone when this
question was asked of them. So these are the women
who were the non-responders. So most of them--if
you look at the first "definitely not" or "probably
would not," 95 percent of these women would not use
this patch.
If you look at the women who reported
having a meaningful benefit, using the cut-off that
we used, a similar number--about 95 percent--90 to
95 percent of these women have said that they
probably or definitely would be interested in
continuing treatment.
DR. MONTGOMERY-RICE: And then the people
who are in the light blue, who would definitely
not, those were mixed--some of those on placebo,
and some of those were--or all of these are
treatment people?
DR. MEYER: All of these are women who said
they had a meaningful benefit. So some of them--
299
DR. MONTGOMERY-RICE: No, no.
DR. MEYER: --some are placebo and some
are treatment.
DR. MONTGOMERY-RICE: Okay. So what
percentage of the people who were "definitely not"
and "probably not" were the people who received the
patch?
DR. MEYER: The definitely and probably
would--this is about 51 percent of the women who
were on the patch. So it's about 49 percent are
going to be in the white bars that had no
meaningful benefit.
DR. MONTGOMERY-RICE: So 49 percent of the
people who were in this study, who were on the--49
percent of the people using the patch--
DR. MEYER: In the clinical relevance
study.
DR. MONTGOMERY-RICE: In the clinical
relevance--did not say they wanted to--they
probably would not continue treatment.
DR. MEYER: Yes. Well--no, this is placebo
also.
300
DR. MONTGOMERY-RICE: Okay. Well I'm going
to ask the question again. I want to make sure
we're clear about this.
What percentage of the people in the
"definitely not" and "probably not" were people who
used the patch?
DR. MEYER: What percent of these--
DR. MONTGOMERY-RICE: Yes.
DR. MEYER: It's about 49 percent.
[Comments off mike.]
DR. RODENBERG: 49 percent of the people on
active therapy--I have numbers to address your
question, maybe not completely, the way you're
asking it. But let me see if this addresses your
question.
49 percent of the women on active therapy
stated that they would probably not or definitely
not use the patch. 64 percent of the people on
placebo therapy stated that they would definitely
not, or probably not use the patch.
Does that answer your question?
DR. MONTGOMERY-RICE: Yes. Thank you.
301
DR. GIUDICE: Dr. Lewis?
DR. LEWIS: I have a couple of questions.
One has to do with the instrument that you
designed. And I think that certainly it helped to
make the study population very well-defined--which
I think it was well-defined in this case.
But I would like you to address a concern
that it will lead to less stringent selection of
candidates for this treatment when it reaches the
general population; that is, if it's already being
touted as a female Viagra in the general press, how
many physicians, and how will physicians be
educated to select a proper population with
hypoactive sexual desire disorder? That's one
question.
And the second question really is for our
committee consultant. Could you comment on a
correlation between testosterone levels and
hypoactive sexual--or sexual dysfunction in a
menopausal population? I'm really only aware of
the Australian study, which showed no correlation
between testosterone levels and sexual dysfunction.
302
And that was like a population-based study in
Australian women--maybe a couple thousand women,
something like that.
DR. MEYER: Okay, let me first address the
educational plans that we have to ensure that
patients and physicians understand how to maximize
the safe use of the patch.
If I could have slide 707 projected,
please?
[Slide.]
We have several plans. We have both a
package insert that we have tested with physicians
for clarity of understanding; that they understand
what this patch is indicated for, and how to use
it. And we have a patient information leaflet.
Again, it's tested which surgically and naturally
menopausal women to ensure that they understand how
to use the patch, and who it's intended for.
We're also developing tools to help both
clinicians and patients understand and recognize
HSDD, and for clinicians to diagnose HSDD and
identify appropriate patients. And these are
303
based, in part, on the five questions that we used
in our clinical study to enroll the
patients--again, to ensure they had desire, they
had the surgery, they lost desire, and they're
distressed about it.
So we have--we're targeting both the
patients and the physicians for this.
We also have a web-based educational
program in development for physicians on the
appropriate use of the product; the prevalence of
HSDD; and the clinical implications. And, for
example, one way that we are ensuring that
physicians get training in HSDD and the appropriate
use of the patch is if they would request a sample
to be sent to them they will have to fill out a
questionnaire on the disease--successfully fill out
a questionnaire, I should add--so that we can be
sure that they understand that these are for the
right people.
The other aspect of this education program
is to facilitate the dialogue between patients and
physicians, because it's not always occurring now.
304
Physicians don't have treatment options all the
time. We've done a lot of research with
physicians, and a lot of them don't discuss it with
their patients because they can't do anything
anyway. So what's the point? A lot of patients
don't really know whether or not what they're
feeling is normal, hence the Hyster-Sisters
website, which I think has helped an awful lot of
people.
So this is to facilitate dialogue between
patients and physicians so that everybody knows who
should have the patch, and who should not.
And the other thing that we're doing is a
CME program supported by unrestricted grants to
educate physicians on female sexual function.
DR. LEWIS: But your post-marketing
follow-up is with a mixed population. It's not
just with the surgical menopause patients.
DR. MEYER: Right. And we also will have
educational tools for naturally menopausal women,
contingent upon that.
DR. LEWIS: Okay. Thank you.
305
DR. GIUDICE: In the queue--
DR. MEYER: Oh, there was a second
question--
DR. GIUDICE: Oh, I'm sorry. That's
correct. Yes. Dr. Heiman.
DR. HEIMAN: Yes, I can make just a couple
of comments on the--you were talking about
endogenous levels in post-menopausal women, not
necessarily surgical menopausal?
So, these correlations typically--they're
often not significant. And there's actually going
to be a couple new studies coming out shortly that
I actually can't comment on at the moment.
But they are not significant. Typically,
DHEA, believe it or not, tends to be more
correlated with desire.
But there are subgroups of women for whom
it's--you know, that's really the problem with this
area, from our side of the fence. Number one, the
definition of sub-groups of low sexual desire, of
which there are sub-groups. There are, but they
haven't been clearly identified. And, number
306
two, the subgroups of women for whom low
testosterone actually is correlate: who are they?
How are they different?
So I'm not sure the question has been
finally answered, even with the new studies that
aren't quite out yet coming out.
DR. GIUDICE: Thank you.
DR. MEYER: Oh--Dr. Shifren would like to
comment on that.
DR. SHIFREN: One thing--if I could have
slide 500 projected, please. This will show you
some of what we're up against.
[Slide.]
Here are the levels of free testosterone
in women with low libido and women with normal
libido. And these are from our surgical menopause
population validation studies. And you can see the
extensive overlap in free testosterone between
these groups.
[Comments off mike]
The libido in oophorectomized women?
Well, as they're trying to pull that up I'll just
307
describe the study. But basically, as a
reproductive endocrinologist, I was very interested
in trying to find some data that showed us that
physiologically, testosterone is important for
libido. We clearly have treatment studies--many
before the Phase III studies you've seen today.
But what do we have as background data/
And, really, the best studies to look at are those
in which women have had their ovaries removed.
And, of course, you need to use hysterectomy,
because the majority f women who have oophorectomy
have concurrent hysterectomy.
So this is a very nice slide out of
Sweden, where they basically send questionnaires to
a group of women who had undergone hysterectomy at
one institution. They then asked the women, "Since
your hysterectomy, tell us if your libido is the
same or better or worse?"
And what you can see is that regardless of
whether women received estrogen therapy or not, if
you compare women who underwent bilateral
oophorectomy at the time of hysterectomy, to women
308
who did not undergo oophorectomy at the time of
hysterectomy, you can see that women who underwent
oophorectomy concurrently were significantly less
likely to say that libido was same or better, or
were significantly more likely to say that libido
was worse.
So I think this is one of the true more
natural experiments that does show that for
menopausal women--for surgically menopausal women,
the testosterone produced by their ovaries really
does affect libido.
DR. GIUDICE: Thank you. Dr. Tulman.
DR. TULMAN: Yes, thank you.
Do you have--and I don't know whether this
is for Dr. Shifren or for Dr. Braunstein--you've
shown us the testosterone levels for pre-menopausal
women. You've shown us testosterone levels through
the placebo group and the baseline treatment group
for women with HSDD who are, by definition,
post-menopausal.
Are there any norms you can show us for
women who are naturally menopausal, and their
309
testosterone levels, and how they might differ from
women with HSDD?
[Pause.]
And I have a part two of the question.
DR. BRAUNSTEIN: [Off mike.] I don't have a
slide for it, but--
DR. TULMAN: Well, can you tell me?
DR. BRAUNSTEIN: [Off mike.] Yes--
DR. TULMAN: I can't hear. I don't know if
anyone can hear you.
DR. BRAUNSTEIN: Yes. Levels were very
similar in the natural menopause versus surgical
menopause. They were both at sort of the lower
level of detection even with an assay that is
highly well validated by FDA standards against GC
tienna mass spec.
DR. TULMAN: And these are women both with
HSDD and women who report their sex lives as being
satisfactory?
DR. BRAUNSTEIN: No, specifically these are
the women with HSDD.
DR. TULMAN: How does that compare with
310
women whose--
DR. BRAUNSTEIN: There's a lot of overlap.
So if you look at the women in the validation study
with normal sexual function, they tended to have,
on the average, about 1 picagram per ml of free
testosterone levels higher than similar women who
had been oophorectomized.
DR. TULMAN: Okay. So that--and I guess my
part two of the question goes back to some basic
theory which one of the people asking the question
at the public forum part of the meting asked: how
does the theory--or where is the theory, or what is
the mechanism for testosterone to produce the
effect of improving a woman's sex life?
DR. BRAUNSTEIN: Well, there is certainly
pre-clinical animal data that shows that if you
remove ovaries and give testosterone there's
increased sexual activity. In regards to humans,
some of the best studies on the effect are te ones
that were carried out by Dr. Cherlyn Gelfand--and
Dr. Gelfand's here--looking at women before and
after oophorectomy, and women who either received
311
placebo, estrogen, testosterone alone or estrogen
plus testosterone, and showing that those on
placebo dropped down; those who were maintained on
testosterone stayed up as far as libido.
Now, as far as the theory of where it
works, there's a couple places that testosterone
works. One, I think--if I'm not mistaken, Dr.
Heiman's group has shown that there's increased
vaginal blood flow with testosterone
administration.
But probably most importantly, there are
receptors--there are androgen and estrogen
receptors in the brain. And testosterone probably
works primarily by increasing desire. It's not a
female Viagra. Viagra works mechanically on the
erectile function in the male. This is more of
something that works centrally on desire. And I
think the desire goes up, and then sexual activity
goes up.
DR. TULMAN: And what part of the brain is
that in?
[Laughter.]
312
DR. MEYER: The highest concentration of
steroid receptors in the brain is in the area known
as HTSM--hypothalamus thalamus septum and midbrain
area--and it's long been shown that, especially in
the hypothalamic area, this is where sexuality
lies. You can lesion that area and get
Kleuver-Bussey-like syndrome and things like that.
And that's where these receptors primarily are.
DR. GIUDICE: Dr. Patrick.
DR. PATRICK: Yes, I have some questions
for Dr. DeRogatis.
The first one has to do with--I'd like to
sort of compare your slide 42 on the clinical
relevance results--the summary of the MCID with
Dr. Davis' slide 14, which is the summary of the
results, and just make sure I understand, since
it's very hard to relate the responder analysis to
the change scores.
First, in the anchoring study--if I
understood this correctly--that a single question
was asked on whether they found the change to
be--they found a change, and they found the change
313
to be meaningful.
DR. DeROGATIS: A meaningful benefit.
DR. PATRICK: And cognitively, that's a
pretty difficult task. Did you do any debriefing?
DR. DeROGATIS: I'm sorry, I didn't
understand your question.
DR. PATRICK: Did you do any debriefing of
how women thought through that question? Did the
perceive a just-noticeable difference and then call
it "important," or did they--how did they know a
change was "meaningful?"
DR. DeROGATIS: Ahh--I think it was--the
question had to do with clinical benefit: "Did you
experience a meaningful benefit?"
DR. PATRICK: Okay.
DR. DeROGATIS: I don't know that they
were--and if I'm wrong there's someone who can tell
me this--I don't know that they were debriefed
about the details of that.
DR. PATRICK: Okay.
DR. DeROGATIS: However--
DR. PATRICK: So it all sort of hinges on
314
that one question.
DR. DeROGATIS: Well--yes, that's right.
But--and that--you know, there are many approaches
to clinical relevance, and the anchoring, and all
of them--there's not a definitive approach. All of
them have strengths and weaknesses.
DR. PATRICK: Right.
DR. DeROGATIS: And perhaps, if there is a
weakness in the anchoring approach, it's the
reliability of that single question, and that
playing an important part.
However, I think it's important to
recognize for clinical relevance, I think there's a
little confusion about it, clinical relevance is
predominantly established--it's an individual
patient characteristic, and it's predominantly
established by proportions of responders--as
opposed to comparisons of means, which is the basis
for statistical significance.
And so once the MCID is established in an
optimal fashion, then the magnitude of difference
can appear small to anyone, but it's actually the
315
proportion of responders. And the significant
difference between those proportions that is the
basis for clinical relevance statement.
DR. PATRICK: Right. But the clinical
trial results, with the statistical significance
are still changes in mean group scores. And we have
to be able to interpret those, as well as a
responder analysis. And if I understand the
results, we have a change of just about one event
in four weeks from a baseline of three events to
four events.
DR. DeROGATIS: No, it's actually two
events; that is, the treatment group changed two
events, making it really a 66 percent change--
DR. PATRICK: Yes, but I'm interested in--
DR. DeROGATIS: --and the placebo group--
DR. PATRICK: --taking away the placebo
effect here.
DR. DeROGATIS: Yes, well then it's one--
DR. PATRICK: That's right.
DR. DeROGATIS: --it's one event, but still
a 33 percent change.
316
DR. PATRICK: Okay.
DR. DeROGATIS: And I--no, go ahead.
DR. PATRICK: And then with the distress,
that's 6.7 units, and with desire, that's 5.1
units. And that's what we're left with: the
difference between treatment and placebo.
And I think that the instrumentation in
this study was really well known. The validity
studies are really well done.
And our issue is interpreting those mean
group scores. And although the responder analysis
gives us a little bit of a hint there, I wonder
what you would say to: if I took that difference in
your meaningful benefit versus no meaningful
benefit and applied it to the mean differences,
which is taking the difference between two groups
and applying it to a change score--and I wasn't
quite sure about the validity of doing that,
although that's what one of the slides from Dr.
Davis does.
If that's the case--I mean, how--is that
fair to apply the MCID?
317
DR. DeROGATIS: On hundred-point
scales--and now this is a little Kentucky windage,
because I don't have actual data, although there
are some very recent reports--Guy Att and his
group, and Sloan and his colleagues have both done
reports in which they have used as example
hundred-point scales. And this technique that
they're talking about is referred to as "effect
size empirical rule checking." And I won't bore
you all with the details of that because everyone
will fall asleep.
But, in fact, they wind up with magnitudes
of change that are only slightly greater--I think
they're in the range of seven or something like
that--as opposed to the five and six here.
So it's on the edge. But, again, I'm
flying by the seat of my pants.
DR. PATRICK: Well, the Dyatt approach is
actually using the anchoring approach, but it's
using it with a 15-point scale and not a "yes/no."
So it's actually from a minimally big change to a
great deal of change.
318
DR. DeROGATIS: But in his review he
contrasts--
DR. PATRICK: Right--back to the
distributional approach. But we're not talking
about distributional approach, because that was not
presented in our briefing package.
We don't know the SEM, we don't know the
center deviation difference. So I don't want to
get into that.
But my question is: is it fair--because
this is the important thing for me--to say that the
difference between the TTS and the placebo was one
event per four weeks; your minimally clinical
important difference was one--okay. So you win on
that.
DR. DeROGATIS: I'm with you so far.
DR. PATRICK: But for distress, it was
greater than 8, and we got a difference of 7--so
that we're just on the margin there.
Is that fair to do?
DR. DeROGATIS: No, I don't think it is.
DR. PATRICK: That what I wanted to know.
319
DR. DeROGATIS: And let me tell you why.
DR. PATRICK: Because that's what's going
on in--
DR. DeROGATIS: Because we're talking about
two very different distributions here. The
distribution--the MCID is an optimal discriminator
between distributions of change from baseline,
whereas the mean scores you're referring to are a
very different distribution. I don't know that
there's necessarily a relationship between those
distributions.
And I think the important thing is that
clinical relevance here is established
through--clinical relevance obvious as a
hypothetical construct is established through some
operational paradigms that we say we're going to do
this, we're going to do that. They're logical and
we agree consistently a science.
And so this particular operational
paradigm, using the anchoring technique which is
very traditional, well established, tied back to
patient perceptions, establishes that these are, in
320
fact, these are differential proportions of
responders which significantly favor active
treatment. Now, that's a clinical relevance
statement. It's an accepted one. It's by the
book, so to speak.
I can't tell you what the relationship is
between this traditional operational definition of
clinical relevance and Dr. Davis'.
DR. PATRICK: Well, what's a little bit
confusing is this "greater than or equal to 8.9,"
because that intimates that that's change. And
it's actually a different score.
DR. DeROGATIS: That's right. It's
change--
DR. PATRICK: [Simultaneous comment
inaudible.]
DR. DeROGATIS: That's right.
DR. PATRICK: And I'd be very interested in
the statistical comment on this.
I still think, in integrating, we are left
with 50 versus 34, 44 versus 30, 51 versus 39. And
then we are left with 1, 6 and 5. And somehow,
321
putting those two things together is a big
challenge of the committee.
DR. MEYER: I have just--I have a
slide--could I have slide 705, please?
[Slide.]
DR. GIUDICE: I don't want to curb the
discussion, but I just want the committee to be
aware that we have four other people who have
questions. We have questions also for the FDA.
And then we have the list of questions that we need
to get done, and we have three committee members,
apparently, who need to leave a little early.
So--could the sponsor please be very
succinct.
DR. MEYER: Very.
Let's look at the data in a different way.
Your "satisfying sexual activity," "desire" and
"distress"--the key endpoints of HSDD. Here's our
MCID of greater than 1--8.9 for desire, and less
than 20 for distress.
Now, what did we see in the TTS patients
from baseline? Because when you're out in the real
322
world there's no placebo group. And Dr. Shifren
told us the placebo is an intervention. Women who
have HSDD who are not getting treatment are not
spontaneously necessarily going to increase their
satisfying sexual activity by one event in four
weeks.
So we have a change of almost 2 in
satisfying sexual activity in the TTS-treated
patients; 10.8--almost 11--in desire; minus 23 in
distress.
Now let's look at the women who said they
were responders. These are the ones who said they
will use the patch: 4.4 per four weeks; 21 increase
in desire; and almost 37 unit decrease in distress.
And, just for perspective, the MCID on the Western
Ontario and McMaster-Womack score for measuring
pain in arthritis is 3 or 100 mm scale.
DR. PATRICK: Thank you. I simply don't
think you can discount placebo in this. But we'll
go on.
DR. GIUDICE: I think that's an important
point that the committee, I hope, has heard.
323
In the queue--and if you've already had
your question answered, then please let me know.
Don't feel obligated to ask it.
Going through the queue--Dr. Heiman did
you have something else?
DR. HEIMAN: Yes. I'll try to be brief.
It's in regards to efficacy. What I was
curious about is if there's information on how
quickly the treatment group got up to a clinically
significant level of change. For example, did it
happen at one month, or two months or three months?
So that's one question.
And whether there was any diminution--even
in six months. I mean, I wish we had more one-year
data. But since there's older clinical information
that, for example, estratase, there's an initial
effect--could be some placebo in there, of
course--and then it quickly diminishes. And so I'm
interested in that--that question.
And related to that: any thoughts about
why--do you still feel as strongly this needs to be
used in chronic dosing, as opposed to intermittent
324
dosing.
So those things are related a little bit.
DR. MEYER: Yes--and let me take your first
question first.
If I could have slide 167, please?
[Slide.]
And this sows the time course for
satisfying activity, sexual desire, and distress.
The blue represents placebo, the yellow represents
TTS.
And we saw a difference between placebo
and TTS on sexual desire and distress that was
statistically significantly different from placebo
as early as four weeks; also in satisfying
activity.
The maximum was reached at three
months--for all three endpoints. Placebo maximized
and it pretty much stayed the same by about four
weeks. TTS continues to go up. And on the
distress it continues to go down.
DR. HEIMAN: Do you see the change in
satisfying activity as being a significant decrease
325
from the four-week mark?
DR. MEYER: No. It's not.
DR. HEIMAN: Okay. Thanks.
And chronicity--any thoughts?
DR. MEYER: What we found in the
persistence of benefit study was that if you remove
the patch the effect goes away, basically.
DR. HEIMAN: Okay.
DR. GIUDICE: Dr. Emerson?
DR. EMERSON: I have two questions. The
first was just hit upon a couple minutes ago--this
idea of ignoring the placebo group, and assuming
that the activity in the placebo group is to be an
indication of placebo effect; that there's this
concept that, no, that was just what was going to
happen over time, and it had nothing to do with a
placebo effect--as was brought out.
I notice that not only in the slide 108,
where Dr. Shifren presented the effect, she gave
the whole effect from baseline. And I also noticed
that in the package insert the only information
that's given is baseline to follow-up in the
326
testosterone group.
I'd like some comment about why you would
disregard that placebo effect so strongly, when you
don't know whether this is just a general time
course. In fact, I would actually argue that that
last slide, where you're starting to see a
diminution of effect in both groups is actually
just possibly related to the fact that you have
incident cases of patients going to the doctor for
HSDD, and they are actively trying, and during that
period they're trying to increase the frequency of
intercourse, and then after a while they give up.
DR. MEYER: I'll have Dr. Shifren comment
on the clinical consequences of that. Course
comparison to placebo is always appropriate in a
randomized clinical trial. But then in the real
world--
DR. EMERSON: And in the real world, the
question is: would this have happened anyway?
DR. SHIFREN: I think none of us are saying
disregard the placebo. That would be absurd.
What we found exciting about exciting
327
about these results is that every single aspect of
HSDD that we measured improved in a statistically
significant way compared to placebo. That, I
think, is a very appropriate comparison to placebo.
But when you're actually looking at level
of change--and I think I'm going to use hot flashes
as an example--we expect a 30 percent decrease in
hot flashes if a woman takes a placebo tablet. But
let's say with estrogen, we expect 60 to 70
percent.
When a woman walks into my office and
says, "I'm feeling better on estrogen. My hot
flashes have decreased," she doesn't say, "Well, it
was 70 percent, but I'm going to take away the 30
percent that would have been placebo, so I've only
have a 40 percent reduction in hot flashes."
When you're actually looking at treatment
effect for the patient, it is the increase from
their baseline event rate.
So I think it's very important to use
placebo when we're looking at the statistical
significance--
328
DR. EMERSON: Do you not look at trade-offs
between the benefit, relative to potential harm,
for the toxicity, that the placebo--
DR. SHIFREN: Benefits and risks are a
crucial discussion that we have with every patient.
Actually, I did want to bring out--I mean,
I think actually the exciting thing about WHI is
that it has informed all physicians and patients
that we will never take hormone therapies lightly.
I think that will actually benefit this product
because I think the likelihood of off-label use has
been significantly decreased, given both
physicians' and patients' concerns about the
long-term risks of hormone therapy.
The comments from the audience during the
open session were excellent, and I really
appreciate them. But I think we were sometimes
undermining patients' abilities to weigh and
balance risks and benefits, and to physicians'
abilities to carefully read package inserts and
patient information and advise their patients.
I see women in my practice every day with
329
debilitating hot flashes who are making very
complex decisions.
DR. EMERSON: But, to that issue, don't you
think it would be useful in the package insert to
point out what the placebo effect was?
DR. SHIFREN: Certainly placebo effect
should be in package inserts.
DR. EMERSON: Okay. I do have--
DR. SHIFREN: I completely agree.
DR. EMERSON: Okay. And I'm sorry--I'm
going to be cut off in just a second, so--
Dr. DeRogatis, I have a question about
this anchoring technique--two things. One is I'm a
little bit bothered by this idea if we just go with
where the 45 degree line intersects that ROC curve,
without really thinking through--it's possible that
whatever you're looking at is not even predictive.
And then the other question is: the way
that this question was worded to the patients, it
was post-randomization, it was just an overall
question of "How was the treatment going," which
meant the patient could have been considering in
330
these risk-to-benefit ratios.
How did this analysis proceed, in terms of
the clinical benefit, if you looked at it
treatment-group versus placebo?
DR. DeROGATIS: I'm not sure I understand
the question, to be honest.
You're having problems with the notion of
the single anchor?
DR. EMERSON: That's correct..
DR. DeROGATIS: Okay. Well--
DR. EMERSON: And basing it on an ROC
curve, no matter what. Because an ROC curve could
be just no better than flipping a coin.
DR. DeROGATIS: Ahh--
DR. EMERSON: You had an area under the
curve of .77--which you said you felt was near
excellent.
DR. DeROGATIS: Well, yes, actually it's
much better. If we had put a diagonal line going
the other way, which the line of no information on
the ROC curve, the coefficient would have been .5.
That would have been no better than flipping a
331
coin.
So the fact that we had .77--
DR. EMERSON: But this idea of choosing the
anchor that would have intersected with that 45
degree line of no-benefit, as well. And so I'm
just questioning whether there is really any
scientific rationale there. And I'm also
questioning whether this analysis shouldn't have
been done separately for the placebo group and the
treatment group--
DR. DeROGATIS: Oh, I see--I'm sorry--
DR. EMERSON: --given the generality of
your questioning of patients.
DR. DeROGATIS: Yes, I see. Yes.
This is a very traditional way of doing
anchoring methodology, because what you're
attempting to do--and so placebo and active
treatment are kept together. They're not separated
out.
And the reason for this is what you're
trying to do is establish a meaningful change, as
opposed to looking at treatment mechanism or that
332
sort of thing. And so by keeping them together,
you're not artificially separating and pulling out
separate groups, but you're establishing in a
presumably representative sample that this is the
magnitude of change that represents the minimal
clinically important difference.
They are often--in many, if not most
anchoring techniques--kept together, placebo and
treatment.
DR. RODENBERG: Can I make a quick comment,
please?
In doing the analyses--I just think it
might be important to know that whether we're
talking about those that said they had a meaningful
benefit or didn't--placebo and the active therapy
groups responded very similarly, in that whether it
was due to being in active therapy or placebo, if
you had a large change, you considered it
meaningful, and if you did not have a large change,
you did not.
The difference--this is the proportion of
people on active therapy that had a meaningful
333
benefit compared to placebo. But once a person had
a meaningful benefit, they had very similar
changes, regardless of the mechanism behind it.
DR. EMERSON: Using your anchoring
technique, would you have come up with the exact
same threshold of 1 unit for both groups?
DR. RODENBERG: I haven't done that
analysis, but I can tell you that the means and the
medians; the distribution for the placebo and the
active group--for both the responders and
non-responders--are almost identical. They're not
statistically significantly different, but there is
a sample size issue--you could bring that up.
But looking at them, for active therapy
the mean was 4.4 on the responders, and it was 4.3
on the placebo group.
So I believe--yes, I didn't do the ROC
analysis because of the low sample size. But yes,
I do believe you'd actually get a very similar
cut-off if you did this for just the active therapy
group.
Also, the 45-degree line, it's actually
334
like the inverse 45-degree line. Where that
intersects the ROC curve, that was the point that
was used as the optimal cut-off, because it
balances different types of misclassification:
misclassification of true responders and
misclassification of true non-responders. And so
we were looking for something that basically
false-positives and false-negatives were treated
equally. And that's where the ROC curve intersects
that, that was the optimal cut-off kind of
balancing misclassification rates.
DR. EMERSON: But is there any rationale
that says that where that cut-off occurs is what's
truly clinically meaningful?
DR. RODENBERG: I think you can
always--right. We get the patients that say, you
know--when we look at the two different groups,
that seems to differentiate the two, in terms of
misclassification rates.
DR. DeROGATIS: Yes, you can apply
different utility functions to false-positive,
false-negative and true-positive, true negative.
335
But in a balanced equation, as Dr. Roderberg is
pointing out, that's the optimum correct
classification--
DR. EMERSON: That's one definition of
optimum. There are many definitions of optimum.
DR. DeROGATIS: Well, it's one definition,
but it minimizes misclassification of responders
and non-responders.
DR. GIUDICE: I think we could go on all
afternoon. But thank you.
Dr. Lockwood had a question.
DR. LOCKWOOD: A quick question, and it's
for Dr. Lucas, who's thus far escaped unscathed.
[Laughter.]
And I'm sure she was relaxing there.
And it's slide 67.
I think reasonable people will
debate--probably in perpetuity--whether or not
there are cardioprotective or cardiotoxic effects
in the younger age group of the WHI study. And
some people might even still debate--I have this
debate with my ex-chair in my institution all the
336
time--about potential for there being a causative
relationship between estrogen and progesterone and
breast cancer, or estrogen not playing a role.
But everyone would agree that hormone
replacement therapy increases the risk of
thromboembolism. And there's just no doubt and no
debate and no discussion about that.
So I think if there is one single element
of safety that deserves the most scrutiny it is the
potential role of this patch in promoting
thromboembolic disease. Now, since the prevalence
is so low you would really need a very, very large
study of a WHI-type to prove this.
But a reasonable surrogate is to look at
coagulation indexes. And I would posit that the
ones you've looked at aren't particularly useful.
They're not, in fact, often affected even by oral
contraceptives, if you look at the literature; and
that the most sensitive single indices is probably
protein-S--free protein-S and protein-S activity,
which is most affected by ovarian steroids and
potentially by androgens.
337
So my question to you is: in fact, do you
have that data someplace? Were protein-S activity
levels looked at? Or, in the absence of that, were
any real indices of thrombin activation looked
at--for example, prothrombin, fragment 1.2,
thrombin, anti-thrombin complexes--something to
give us a sense of whether or not TTS actually
increased the generation of thrombin.
DR. LUCAS: We did look at prothrombin
fragments 1 and 2 in the Phase II program. We also
looked at protein-C resistance, and we looked at
plasma viscosity, platelet aggregation, and we saw
nothing. And then the values that you see here are
what we then did in Phase III. And we've not seen
anything in any of the measures.
DR. LOCKWOOD: But no protein-S.
DR. LUCAS: We did not do protein-S.
DR. GIUDICE: And a follow-up to that--not
with regard to coagulation, but if you look at the
data on Danazol, with regard to its being an
immunosuppressant, have you looked at all at any
indices of suppression of the immune system in
338
women on the Intrinsa?
DR. LUCAS: In what measures would that be?
DR. GIUDICE: One would have to--
DR. LUCAS: We looked at white counts.
DR. GIUDICE: No, one would have to do in
vitro studies--mixed lymphocyte cultures.
VOICE: [Off mike.] Measuring TH 1 and TH 2
[inaudible.]
DR. LUCAS: No, we've not looked at that.
DR. GIUDICE: There are two other
questions, and then we need to move on.
One is from Dr. Merritt, and the other
from Dr. Burnett.
DR. MERRITT: n the Phase III study design
you allowed patients to continue after 52 weeks
into a persistence of benefit arm. Are those
randomized to placebo and others were randomized to
the TTS system? And you said there was loss of
benefit. Is that only in the placebo arm?
DR. MEYER: No, not entirely. For the
persistence of benefit study, after 52 weeks there
were about--slightly over 200 women who were asked
339
if they wanted to participate in a 13-week study,
knowing that they would get either a placebo or a
testosterone patch. Now, recall, they had all been
on testosterone prior to this.
Those who agreed to participate were
randomly assigned to either placebo or the 300 mcg
testosterone patch. And no one knew what they
were getting.
Following 13 weeks on therapy they were
interviewed by a trained interviewer, with an
extensive script--and it was a script which we also
used in the clinical relevance stud. It wasn't
just a single question, it was a lot of data that
we gathered.
And we asked, in these interviews, about
the same sorts of questions that were covered in
the instruments that they were filling out: "Did
you have a decrease in desire for sexual activity,"
etcetera, etcetera.
And, again, interestingly enough, those
randomized to placebo has a statistically
significant decrease in all their sexual activity
340
indices relative to TTS. The only area where the
two groups were the same was in a noticeable
decrease in willingness for partner-initiated
activity--again, consistent with what Dr. Shifren
showed us.
So, again, it was about 60 percent of the
placebo patents showed a decrease in these indices.
And it was about a 35 to 40 percent people in the
TTS group that shoed a decrease in these indices.
All the p values were statistically significant.
DR. MERRITT: Thank you. So then your
proposal would be that this system would need to be
used chronically and long term, and at the same
time the subjects would need to be on chronic and
long-term estrogen?
DR. MEYER: We have no data on women not on
background estrogen. So, yes, they would need to
be on concomitant estrogen therapy. The patch does
need to be worn continuously, and the duration
needs to be discussed between the woman and her
physician as to what's most appropriate for her
treatment.
341
[Pause.]
And Dr. Shifren will tell us--
DR. SHIFREN: I do just want to add that as
clinicians, since we really are using much less
systemic estrogen in our menopausal
patients--clearly, the major indication is
bothersome hot flashes--we as a group of clinicians
were very concerned about the potential for
off-label use in women not on systemic estrogens,
and wanted to know whether it was truly safe and
effective in that group.
And so the sponsor was very responsive to
our needs. Very quickly we jointly designed a
trial of transdermal testosterone in surgically and
naturally menopausal women on no systemic estrogen
therapy. And that trial is currently ongoing at
the Mass General and multiple other sites.
So, hopefully, we will have answers for
that. And I think we won't know until the study is
done.
DR. GIUDICE: Dr. Burnett, and then Ms.
Solonche.
342
DR. BURNETT: Thank you. This is a safety
question--safety issue question.
I guess with many things that are good you
always--at least in America--want to up-size. And
I just wonder what the potential for--or potential
consequences may be--for somebody who may take the
patch everyday, or use it perhaps in ways it's not
intended to be.
And I guess the background for that
concern relates to some of the data presented by
Dr. Soule. I think there's a trend towards some
effects here, particularly the metabolism syndrom
indices.
So I'm just curious about--do you have any
comments or any data with regard to that sort of
concern?
DR. MEYER: For metabolic syndrome Dr.
Braunstein will address that.
With respect to abuse, we are delivering
23 to over 300 times less testosterone than these
women would need if they wanted to have steroid
abuse.
343
If I could have slide 255 projected?
[Slide.]
This is what Dr. Lucas does in her spare
time.
[Laughter.]
This is how many patches a woman would
have to wear for about six months in order to get
any significant type of abuse potential out of
these patches. And there are way cheaper ways to
do that now, if you go to the drugstore and get
DHEA or something.
So it's not very practical or cost
effective.
DR. BRAUNSTEIN: Actually, let me review a
couple of safety issues, because I think there may
have been some misconceptions during the FDA
presentation.
If we can go to slide 337.
[Slide.]
I'll start off with some of the glucose
data, and then I'm going to show one-year data in
the natural menopause study, because the surgical
344
menopause was double-blind, placebo-controlled for
six months. The natural menopause was
double-blind, placebo-controlled for one year. And
I think that gives a much more meaningful type of
evaluation of both safety and efficacy over a
longer period of time.
If we look over here, this is the glucose
data. And one can see that basically, if you look
at the exposure to testosterone over months--here's
the double-blind period, and here's the open-label
extension--that there basically is no major change
in the glucose levels in these patients versus the
baseline. There is, obviously, some scatter, but
there's no major change.
If we go to 338--
[Slide.]
--we can look at the insulin levels. And,
again, very little change over time, either in the
double-blind placebo-controlled trial, or in the
open-label extension.
And then if we go to 334--
345
[Slide.]
--the glyco-hemoglobin data is shown here.
Again, no difference in glyco-hemoglobin levels in
these patients.
So there's no evidence of really any
metabolic deterioration. Because in the
presentation that you saw, the scale of glucose
changes really was quite expanded. But the changes
were very, very small, and really non-significant.
If we go to slide 228--
[Slide.]
--which is the natural menopause study,
there are two studies--one was a 24-week
double-blind, placebo-controlled study; another was
a 52-week placebo-controlled study. And if we just
look at the 52 weeks--because the data is really
very much the same--systolic blood pressure showed
no significant change; diastolic blood pressure
showed no significant change between placebo or the
testosterone group.
There were no significant
346
differences--both lost, although the placebo group
lost a little bit more weight than the testosterone
group.
If we go to slide 330-
[Slide.]
--we'll show the lipid changes in the
52-week double-blind, placebo-controlled trial in
naturally menopausal patients. And don't forget,
these patients are also on a progestin as well as
estrogen.
And so in the placebo group we can see
there's baseline cluster of 210, going up a total
of 2 at 52 weeks, versus testosterone group,
staring off at 208, going up 6.5--again, somewhat
of a regression to the mean.
HDL levels were very similar types of
changes; both going up, same direction.
LDL--again, the levels--the final levels
were very similar; some degree of regression of the
mean; this going up, because it starts off lower;
this going down.
And triglyceride levels, again, were very
347
similar.
If we go to 344, which shows carbohydrate
levels at one year--
[Slide.]
--we can see in the placebo group the
glucose levels went up 1.1 mg/dL in both the
placebo and the testosterone patch group.
Glyco-hemoglobin levels went down to a similar
degree, and insulin levels actually went up a
little bit higher in the placebo group than in the
testosterone group.
And here's the six-month double-blind
control group--and, again, showing very similar
types of results; trying to give some degree of
confidence in the longer-term safety issues.
And then if we--
DR. GIUDICE: I need to ask you how many
more slides, because we need to move on. We still
have the questions for the FDA.
DR. BRAUNSTEIN: I will stop. But I will
summarize by saying: similarly, there are no
changes in LDL, no changes in the coagulation
348
parameters in the natural menopause study.
And, again, protein-S was not measured.
DR. GIUDICE: Thank you.
Ms. Solonche, you had a question.
MS. SOLONCHE: Yes, first, are there any
differences in your trial results related to the
reason why the subject had an oophorectomy in the
first place?
Second, how do you differentiate between
women who are distressed about HSDD, and those who
might be depressed, and therefore out of the
running?
And, third, has any work been done on the
possible negative psychological effects of using a
medication to increase libido?
DR. MEYER: Okay. Let me start with your
first one.
We did not gather any data with respect to
differences in the women as to why they had the
oophorectomy. They all had been oophorectomized
about the same amount of time, but did not do any
statistical analyses on that.
349
With respect to depression versus
distress--everyone--an entrance criteria, because
we wanted to rule out depression, all women had to
take the Back Depression Inventory, and could not
be depressed according to that inventory, to rule
that at as a possible cause of HSDD. And then they
had to score positively--or negatively, as the case
may be--on the distress scale.
And then the negative psychological
aspects of using hormonal therapy, we did not ask
any questions about that. The fact that the women
who had a positive effect remained in the trial
and/or said they would continue the patch, would
suggest there is a positive benefit, at least for
some women. It was not listed as a withdrawal
criterion. No one gave that.
DR. GIUDICE: Dr. Dorgan, the last
question.
DR. DORGAN: Two very quick questions.
For the women following bilateral
oophorectomy and hysterectomy, could there not be
any psychological component to some of these
350
problems that are being attributed to decreased
testosterone levels?
Dr. Shifren, would you like to address
this?
And also recall that by the time the women
were in this trial they were, on average, eight or
nine years post-oophorectomy. I mean, they've lost
their ovaries and their uterus has been removed. I
would think that there could be a psychological
component, but I could be wrong.
DR. SHIFREN: If you think back to the
hysterectomy study--the observational study--in
general what we see is that the majority of women
who undergo hysterectomy actually have an improved
sex-life post-operatively. And that's been shown
in many large studies.
Of course, that makes a lot of sense.
Women only have a hysterectomy if they have
underlying pathology. So typically they have
bleeding, fibroids, endometriosis. And the removal
of that problem often leads to increased sexual
activity and function.
351
What was so interesting--I thought about
that Nathorst-Buhst study that I showed you
earlier, is that within that group of women, you
were still significantly less likely to get that
increase in libido if your ovaries were removed
concurrently, and significantly more likely to have
lowered libido post-operatively.
DR. DORGAN: A second question--just to
follow up--in your 2(b) studies, I see why you
chose the 300 mcg per day dose. But you looked at
a placebo, 150 mcg dose, a 300 mcg dose and a 450
mcg dose. And of all the parameters that we're
looking in terms of efficacy, the greatest effect
was with the 300 mcg dose.
If it's the testosterone per se--I'm not a
pharmacologist--but if it's the testosterone per se
that's responsible--if testosterone replacement per
se would improve libido in these women, why aren't
we seeing a linear effect with an increased--well,
sexual increased, number of satisfying episodes of
sexual intercourse, and improved personal distress
when we go up to the 500 mcg dose.
352
I was kind of concerned that we're not
seeing--we're not even maintaining the increase.
It's no longer significant when we go up to 450.
Could somebody comment for me?
DR. MEYER: If you could project slide 181,
please?
[Slide.]
Actually, what we found in the Phase II
study is placebo and 150 were essentially the same;
450 had an effect. It was just not statistically
significantly different from placebo; 300 was
better.
But as you can, in some domains of the
PFSF--placebo's in blue, 150 in yellow, 300 in the
darker yellow, and 450 in the orange--for example,
orgasm, sexual pleasure, sexual responsiveness--450
sometimes did better than 300.
But what we were interested in was the
lowest effective dose. So it was sometimes but not
consistently. And when we did our population PK
studies, we do get dose proportionality in the
doses. So you get higher blood levels of
353
testosterone with the 450--which, again is a bit of
a conundrum. It doesn't explain the whole story
unless we're at the top of the dose-response curve
with 300.
But we chose 300 as the lowest effective
dose.
DR. DORGAN: Okay. Thank you.
DR. GIUDICE: Thank you.
Now I'd like to ask the committee for
questions to the speakers from the FDA.
Dr. Nissen.
DR. NISSEN: I had one very brief question,
and that is: there were a couple of patients that
had bilirubin elevations, and I want to know if any
of the patients in the study--as far as the FDA can
determine--met High's rule of a concomitant
transaminase elevation and bilirubin elevation?
We know that testosterone has potentially
hepatotoxicity, and I"m just looking for any signal
there. Any of the hyperbilirubinemic patients also
have elevations in liver enzymes?
DR. SOULE:: I'm unfortunately unable to
354
answer that question. It's possible that Procter &
Gamble might be able to provide you with the
answer.
DR. LUCAS: Two patients who had markedly
abnormal bilirubin--could we see the box and
whisker plots for bilirubin?
[Slide.]
DR. NISSEN: But what I need to see is
their transaminases.
DR. LUCAS: Okay.
DR. NISSEN: You're aware of High's rule,
and why it's important, I assume.
DR. LUCAS: Yes.
Could I see the bilirubin? Oh, we can't
project.
We saw no difference in outliers with
testosterone compared to placebo for any of the
transaminases or bilirubin.
DR. NISSEN: Okay, the two patients that
had elevated bilirubins, did they also have
elevated transaminases?
DR. LUCAS: No, they did not. They
355
isolated bilirubin.
DR. NISSEN: Okay. Thank you.
DR. GIUDICE: Dr. Tobert.
DR. TOBERT: Yes, I have a question for Dr.
Davis, and it refers to the FDA's first question
about clinically meaningful differences.
And I thought it might illuminate the
question a bit to consider other drugs, or other
drug classes that the FDA considers meaningful that
act on the brain, as we've heard that this product
does.
For example, if you take an SSRI for
depression, what kind of differences would you get?
I mean, if I refer to your slides 11 and 12, you're
showing that the testosterone patch sort of gets
you some way back to normal, but only about a
quarter of the way. And you have like a 52 percent
responder rate versus 31 percent.
Now, I think that 52 versus 31 would be
acceptable for something like an antidepressant.
Can you--obviously you have access to a ton of data
on this--could you comment on that, please?
356
DR. DAVIS: In our division we have not
had prior drugs that really have been evaluated on
a quality of life or patient-reported outcome
analysis.
DR. TOBERT: But the FDA looks at this
question all the time.
DR. DAVIS: Yes, but in our division we
have not handled drugs of that sort, so perhaps--
DR. TOBERT: How about the PDE-5s, the
newer PDE-5s? Did you have to deal with that?
DR. DAVIS: Let's have Dr. Griebel or Dr.
Monroe answer that.
DR. GRIEBEL: I think each one of them is
dealt with individually. And this is our first
experience with female sexual dysfunction, these
endpoints. And we're asking--that's why we brought
it to committee. We're asking for your input on
this.
DR. TOBERT: Okay. All right, I would just
state that my impression is that for other drugs,
this would be considered pretty good: 52 versus 31
percent. But if I'm wrong, please correct me.
357
DR. MACONES: This is for Dr. Soule.
Again, to go back to the post-marketing
plan--which I'm having a tough time with--on your
slide number 29, your second bullet-point says that
recruitment goals were not previously met using
this database, which seems to be a really critical
point.
Could you give us a little more
information about that?
DR. SOULE:: I don't know how directly I
can discuss a plan that involves another product at
this committee--except to say that there has been
some experience with the database and goals have
not been met, in terms of recruitment and
timeliness.
DR. MACONES: Goals with this company?
DR. WALKER: Alexander Walker.
What we can offer and promise is a number
of people. So we've got covered lives. If a
product doesn't sell, if it's displaced by other
products, a projection that's based on marketing
will fall short.
358
DR. MONROE: Dr. Walker is correct. If the
exposure to the drug is below their expectations,
they won't be able to recruit at the rate that is
predicted, so that there's two options: they'll
either miss the recruitment target, or the study
will have to run for a longer period of time.
And as Dr. Soule said, we were somewhat
skeptical of the time-lines, based on an experience
that we've had in our division. And I think we're
both in agreement, Dr. Walker.
DR. GIUDICE: Dr. Patrick, and then Dr.
Hager, and then Dr. Lipshultz.
DR. PATRICK: The sponsor faithfully
followed the 2000 draft of the female sexual
dysfunction guidance. Given the difficulties in
interpreting the number of satisfactory sexual
events--particularly, as DeRogatis pointed out, a
fractional number is not easy to interpret--has
there been a thought of trying to look at these as
combined endpoints? Or what is the relationship?
And why is satisfactory sexual event thought to be
the primary endpoint in this area?
359
DR. MONROE: Well, as everyone has heard
from the discussion that has ensued earlier today,
this is a new area, I think, for everybody. It's
an area that doesn't have clean endpoints. There's
a lot of active research that's going on.
We recognize that many of the
investigators in this area do have questions about
this being the primary endpoint. At the time that
the draft guidance was done, it reflected the best
assimilation of the available data of the people
that drafted the guidance, based on their
interactions with various investigators in the
field.
We are considering looking at this entity
of hypoactive sexual desire disorder in somewhat
different ways. But at the time this study was
started, these were the rules, and the sponsor did
follow the advice that we gave to them. And I
think it's only fair, today, to evaluate their
application primarily on the way the rules were set
up at that time.
It doesn't mean that, as we go forward and
360
gain more experience in this area, we won't give
greater or lesser weight to several of the
secondary endpoints.
DR. GIUDICE: Dr. Hager.
DR. HAGER: Granted that this is a new area
for the agency, but we do have rather long
experience with a product of combined
dequon-estrogen and testosterone that has been
available.
Are there data regarding not reaching the
primary endpoints of efficacy, but side effects, as
far as breast cancer and cardiovascular risk?
DR. MONROE: Would you like to--someone is
with us from our office of Drug Safety. And I
think she's going to show you our experience using
the adverse event reporting system--the AERS
database from the agency, the spontaneous serious
adverse event reports that the agency has received,
perhaps over the last decade.
DR. GELPERIN: I'm Kate Gelperin. I'm a
medical officer in the Office of Drug Safety. And
in the interest of time, I'll just say briefly that
361
the type of reports that are in the AERS safety
database at the FDA are generally spontaneous
reports from consumers or health professionals, and
so they are not clinical trial results.
With Estratest, we did run a search of the
AERS safety database for serious adverse event
reports--"serious" is a regulatory definition that
includes death, life-threatening, requires or
prolongs hospitalization, congenital anomaly, and
then there's a category called "other" that's other
medically important events.
And when the search was run in that way
there were a total of 226 reports in the database.
Of the raw counts, the most frequently reported
include breast cancer, depression, headache and
acne.
A review was done of the breast cancer
reports with Estratest, including any report of
breast cancer in which Estratest was a suspect or
concomitant drug. This search showed that between
the years 1992 and 2004, four unduplicated cases of
breast cancer had been received, with Estratest
362
indicated as a suspect drug--which were not from
legal sources. Patient age ranged from 31 to 56
years.
We also found a total of 69 unduplicated
cases of breast cancer that were received via legal
sources. The first legal case was received on
October 17, 2003, and all others were received
thereafter.
In each of these cases, other suspect
drugs number between three and nine, and included
other HRT therapies. Patient age ranges from 45 to
70 years.
There were also eight unduplicated cases
of breast cancer in which Estratest was considered
a concomitant drug by the reporter.
A search was done for serious events with
Estratest considered suspect, which did not include
the outcome called "other," since that was a way of
zeroing in on perhaps some of the more serious
effects that might have been required of prolonged
hospitalization. And when that search was done,
the most frequently reported events included CVA,
363
coronary artery occlusion, dizziness, headache,
breast cancer, chest discomfort, depression,
glaucoma, hypoesthesia, pain, and ovarian cancer.
The following category of events--cases
were reviewed for cardiovascular events, including
MEDRA-preferred terms, which is a way of coding
spontaneous reports. We included the MEDRA Pts,
cardiovascular disorder, coronary artery occlusion,
coronary artery re-occlusion, myocardial
infarction, chest discomfort, and chest pain.
There were a total of six cases in the
data base that were serious: two myocardial
infarction, three chest pain, one coronary
occlusion.
In the two reports of myocardial
infarction, one case occurred in a 78-year-old
female who was participating in a clinical study.
The even was considered by the investigator to be
possibly related to Estratest.
The other report was in a 57-year-old
female who was taking multiple concomitant
medications, including opiates.
364
In the three cases of chest pain, one was
attributed to cholelithiasis, and one was
attributed to hypophosphatemia. And one case of
coronary occlusion in a 58-year-old female was
treated with a stent.
With regard to cerebrovascular effects,
these included the MEDRA PT cerebrovascular
disorder, cerebrovascular accident, and headache.
There were a total of six unduplicated serious
reports, with Estratest considered a suspect drug.
There were two reports of stroke, one in a
48-year-old female, and one in a 58-year-old
female. There were three reports of serious
headache, one in a 37-year-old female who was
admitted to the hospital with depression, and was
later diagnosed with multiple sclerosis.
The other two reports were for a
49-year-old female and a 45-year-old female. There
was one report of unspecified cerebrovascular
disease in a 64-year-old female.
Now, with regard to serious reports of
depression, there were three unduplicated reports,
365
with Estratest considered a suspect drug. All
three reports included other suspect drugs.
Just in summary I would say that the sense
of these reports is that although spontaneous
reports are important for hypothesis generation,
that in this case I don't think we could regard
these as in any sense confirming a hypothesis.
DR. GIUDICE: Thank you.
Dr. Lipshultz, you had a question.
DR. LIPSHULTZ: I just had a question for
the FDA speakers, and that was: it seems to me that
you were somewhat--at least in the briefing
document--insinuating that you were not 100 percent
happy with the endpoints here, in terms of change
over placebo. And you've said that this is a new
area, and we're looking at for the first time
quality of life drug.
But that's not true. I mean, you have
looked at the PD-5 inhibitors, with three
instruments that are almost the exact same as the
three instruments here. The names have been
changed and the questions are different, but it's
366
the same thinking.
And what I'm asking is: is reviewing the
data from the two newest PD-5 inhibitors, were
those changes more robust than what we're seeing
here, in terms of quality of life changes?
DR. SHAMUS: Ben Shamus. I'm the Director
of Reproductive and Urologic drugs.
You know, we can look at this many
different ways. One way to look at this is that it
requires us to treat 100 women--expose 100 women
for 15 of them to have a sort of borderline
clinically meaningful effect attributable to the
testosterone. I mean, there's lots of ways of
looking at this, but that's basically what it boils
down to.
So that is not to say there is not a
clinically meaningful effect in some women, and
that there is a mean statistical difference. The
thing, as you know, we have to grapple with is, in
a population setting, which is what we deal with
here--in a population--is that benefit to the
population worth the risk, whatever that may be, in
367
the situation.
Of course, the PD-5s or a whole different
thing, in a sense, that we have a lot more
experience, etcetera, in terms of the risk. And if
I recall the data, actually, many of the people
returned almost to normality in terms of erectile
dysfunction--I don't have the figures in front of
me--as opposed to here, where it's not at all the
case.
But we the benefit--it is what it is, and
that's what it is. And then the risk--we have to
weigh those two things and make some kind of
decision.
DR. GIUDICE: Thank you.
Committee Discussion
DR. GIUDICE: That's an excellent segue
into question number 1, which is:"Do the efficacy
data represent a clinically meaningful benefit
above that of placebo for surgically menopausal
women with hypoactive sexual desire disorder who
are taking concomitant estrogen?"
Our task is to give our recommendations to
368
the FDA, with a yes or no answer.
Does the committee feel that it is ready
to take a vote?
[Pause.]
Dr. Emerson?
DR. EMERSON: Just one point of
clarification: you do want this question answered,
basically, with that risk-benefit trade-off that
you just spoke to--the concept--
VOICE: [Off mike.] That's the last one.
DR. EMERSON: Well, but the concept was
saying that 15--that extra 15 responses relative to
the potential risks.
So--do we want to answer the question of
whether people should want to have one extra
episode per month, or do we want to answer the
question of is this the cost of getting this one
extra episode per month is too much.
DR. GRIEBEL: The risk-benefit question,
weighing in the safety, bottom-line, is at the end.
DR. EMERSON: Okay.
DR. GRIEBEL: Four.
369
DR. EMERSON: So just comment on efficacy.
Whether the one is worth it--I mean, whether the
one is something that you'd like.
DR. GRIEBEL: Mm-hmm.
DR. EMERSON: Okay.
DR. GIUDICE: We're voting on the 1, 6 and
7--essentially.
Dr. Rice?
DR. MONTGOMERY-RICE: I want to make sure I
understand something.
On this "clinically meaningful benefit"
was that a term that you all developed? Or is that
a scientific term that I've missed in statistics?
DR. GRIEBEL: Well, there's a whole
science of clinically meaningful benefit and
minimal important difference that Dr. Patrick might
want to comment on.
DR. PATRICK: Well, you would have learned
this as clinical significance. But the term
clinical is sort of odd here, because this is
defined by the women. And so it's really the
minimum important difference. Forget the clinical.
370
But you can think of it as parallel to clinical
significance if you had a clinical anchor here.
DR. MONTGOMERY-RICE: But that's not the
same as statistical--
DR. PATRICK: No.
DR. MONTGOMERY-RICE: --difference.
DR. PATRICK: So we're got statistical
differences, and what the committee's being
asked--and while I have the mike--I didn't
understand this incorporating risk. The question
asks about clinical benefit. It doesn't say risk
in the question at all. So I'd like that clarified
before we vote, because we have all these questions
about risk later. And so I thought we were
answering this one at a time, rather than
integrating this all.
DR. MONTGOMERY-RICE: And I just want to
ask one other question to the FDA.
If the event had been five more, would
that have changed the question, versus it being 1.4
or whatever it was more? Would that have changed
the question of being a clinically meaningful
371
benefit?
DR. GRIEBEL: Well, it could work both
ways. If the study that was designed to define the
clinically meaningful difference that you had
observed was five extra events, and then the
average was one for each women, then clearly it
wasn't met.
DR. MONTGOMERY-RICE: But you didn't define
that to begin with, did you?
DR. GRIEBEL: No.
DR. MONTGOMERY-RICE: Exactly. So you just
said that it needed to be statistically different
than you chose to use from baseline, and compared
it to groups. So you didn't define that it had to
be five more events over a four-week period of time
to be clinically meaningful. Or maybe you did.
DR. MONROE: We did not, but we told the
sponsor clearly that a statistical change in and of
itself would not be sufficient; that they would
also have to provide evidence--as they've attempted
to do with this study--that the change that would
be observed would be of clinical benefit to the
372
patients; the patients would conclude that they
derived a true benefit. How they're making that
decision is up to the individual that has the
disorder. And that's what was attempted to be done
in this case in that study.
Now, the question is--that was one of the
issues we raised to the panel is there are many
ways of trying to do these studies. Was that study
done in an acceptable manner so that the numbers
that they generated from that study, do they carry
credence with you as a committee member.
In other words, we've tried to present to
you the information that the sponsor has generated,
and we're asking you for your independent
assessment and interpretation of those data.
DR. GIUDICE: Yes, Dr. Heiman.
DR. HEIMAN: I just wanted to make a
comment on this, because I do think it's tricky.
If you come from a psychotherapy side of
doing interventions as opposed to a drug side, one
of the things "clinical significance" can mean is
you jump from the dysfunctional range into the
373
functional range--somewhere into the functional
range. That's a way of quantifying, but it's not
quite statistical--"clinical significance."
I think, in this area, that's very
difficult to do, because though we have some data,
as you've seen, some data on what is the normal
range, I don't think we should, at this stage of
the development of the field, rely 100 percent on
that idea.
So what we have in this case is what
they've tried to do is look for clinical relevance.
And when you go there, it looks like the figure
comes out somewhere around 50 percent when you try
to take that table 43 apart. So that's another
piece of evidence.
And, finally, if you look at one even over
four weeks, I would just be careful as you consider
that, to not treat that casually. That could be
quite important for the women in this trial. In
some ways--it just could be quite important. It
may seen insignificant, and that's also, by the
way--all those things actually are difficult to
374
judge. We don't have a lot of objective standards
from which to do that.
So that's just, by the way, on the event
issue. It's almost as if each of these--I'm not
asking that you consider them separately, but
they're very different measures, these three
measures. It's hard to put them all into one, I
think, and really consider what it means.
DR. GIUDICE: Thank you.
So are we ready to vote?
Okay, we're going to start on this side.
Dr. Macones.
DR. MACONES: Yes.
DR. DORGAN: No.
DR. EMERSON: I'll say yes.
DR. HAGER: In light of the significant
placebo effect, and in light of apparent
discontinuation of a large number of users, and the
maintenance of benefit over time not increasing but
staying stable--in spite of those things, I do
think that there is statistically significant
benefit, and so I would vote yes, with some
375
reservation.
DR. TULMAN: I would vote no.
DR. BURNETT: Yes.
DR. DICKEY: Yes.
DR. GIUDICE: Yes.
DR. LOCKWOOD: Yes.
DR. LEWIS: Yes.
DR. LIPSHULTZ: Yes
MS. SOLONCHE: No.
DR. PATRICK: Yes.
DR. NISSEN: Let me just qualify a little
bit here--
[Laughter.]
--it's been too easy for you.
I think that the agency set a bar here for
what had to be shown with regard to efficacy, and
that bar was met. And so you can't change the
rules--in my view.
And I think that they did--they set out to
do this, they did it very carefully. They showed
efficacy. Now the efficacy, I must tell you, is
fairly marginal. And what didn't really come
376
through there--and I decided not to prolong this by
questioning--but it looks to me like about 36
percent of the placebo patients would really like
to continue the therapy, and about 50 percent of
the treated patients would really like to continue
the therapy.
And I want the sponsor to consider
marketing the placebo--
[Laughter.]
--because that's a pretty good outcome.
So my answer here is yes, but it's not a
very big effect, in my view.
DR. GIUDICE: Thank you.
DR. MERRITT: Yes.
DR. MONTGOMERY-RICE: Yes.
DR. GIUDICE: Thank you.
We are now onto safety. Question No. 2
is:
in the safety database, 494 surgically menopausal
women were treated with TTS in combination with
estrogen for 12 months. Of these 127 were treated
for 18 months. There are no long-term, placebo
377
comparative safety data beyond six months.
The expected TTS use will be chronic in
the intend population.
Is this exposure--total number of women
treated and duration of treatment--adequate to
demonstrate long-term safety?
Does the committee feel that it is ready
to vote on this, or does it need some discussion?
Dr. Tobert.
DR. TOBERT: I had to prolong this, but I
think some few--we haven't really asked any safety
questions of FDA yet--or very few. And I do have
one or two.
DR. GIUDICE: Well, now is the time.
[Laughter.]
DR. TOBERT: Well, firstly, I don't totally
follow the logic here. The WHI studies were
disappointing, but they showed what you might have
expected, considering the history of oral
contraception; you know thromboembolic effects were
no surprise. And they were only disappointing
relative to the epidemiology.
378
But the point is: does testosterone have
any biological effects that are the same as
estrogen progestin? Other than the fact, of
course, that a small amount is aromatized to
estradiol.
Just because the WHI studies were
disappointing, I don't quite see why there's an
issue with testosterone. So maybe you could
explain that.
And my second question is: with regard to
the randomized controlled trial that you are
putting on the table, then you see I'm a big fan of
randomized controlled trials. I sit on the
steering committee of two of them--both of them
about 10,000 patients. But I would question
whether it's possible to do a study that big. And
I heard the number 17,000, but I think that was
not--doesn't include evaluating the risk of breast
cancer, and would be in much older women.
So could you clarify those two points?
What, actually, if you were to study 50-year-old
women, how many you would need in a randomized
379
trial to answer the questions you want answered.
And the other is about the biology of
testosterone versus estrogen and progestin.
Thank you.
DR. SOULE:: To take your second question
first, that's really one of the questions that we
have posed to you as our advisory committee. And
we would like your thoughts and suggestions on what
sort of trial would be optimal; what design, what
duration, what sample size would be optimal.
As far as potential for risk with
testosterone--as you mentioned, there is the
concern about aromatization to estrogen, with the
following of risks that we know to be attributable
to estrogen. But I think the biggest point is that
we simply don't have enough data on women taking
testosterone on a chronic basis to be able to look
into a crystal ball and see what we may see in a
population.
DR. TOBERT: But the labels for the male
products, where you're giving 20 times as much, are
pretty benign. They talk about prostate cancer as
380
potential risk, but they're not loaded up with
warnings and black boxes.
DR. SOULE:: But women are not
physiologically exposed to the sorts of levels.
DR. TOBERT: Well, but this patch is
providing levels that would be found in a young
woman.
DR. MONROE: Well, I think you took us back
to WHI in the sense that a lot of what was being
target there was to bring levels back to what was
in young women. And you created outcomes that were
a surprise to, certainly, many people. Some of
them were apparent, as we heard from the
epidemiologic studies. Some were not.
And I think we feel that there may be some
risk in making assumptions just because something
hasn't been shown in the past in limited numbers of
women treated for short periods of time relative to
the anticipated clinical use.
DR. TOBERT: But that really goes to my
point. Because testosterone is not thrombogenic
like estrogen and progestin are. The female sex
381
hormones tilt the hemostatic balance towards
coagulation. Testosterone does not. Men don't
have to, you know--when man was evolving, women
faced major hemostatic challenge every two years.
That's not the case with men, so testosterone isn't
procoagulant. And that, I think--you know, a lot
of the findings in the WHI probably are
attributable to that.
DR. GIUDICE: My interpretation of this
question is that it is whether the exposure has
been adequate for long-term safety evaluation,
including the fact that this is a patient
population that has been studied with estrogen.
So I don't think it's just the risk of
testosterone. I think it's also the fact that the
indication that the sponsor is going for is the use
of the TTS in the setting of the surgically
menopausal woman with estrogen replacement.
DR. TOBERT: Just to clarify--
DR. GIUDICE: Is the rest of the group--
DR. TOBERT: --we're not debating the
safety of estrogen here. It's only--
382
DR. GIUDICE: I'm not so sure if we are or
are not, because this is an indication for
long-term testosterone therapy, potentially, in the
setting of an obligate long-term estrogen therapy.
And so that is something I think that the
committee needs to discuss, because this, I think,
is part of the underlying issue of the long-term
safety.
DR. HAGER: I fully agree. I think we are
evaluating the product as presented by the sponsor;
and that is in combination, in surgically
menopausal females who are using estrogen.
DR. HEIMAN: Yes, and the way I also
understood this question is in terms of long-term
safety. And so that's very crucial, I think, in
this case. Is that correct?
DR. GIUDICE: Any further discussion on
this?
Yes, Dr. Lipshultz.
DR. LIPSHULTZ: As a
non-gynecologist--getting back to the estrogen use
over a long period of time, I mean, that's already
383
been discussed, decided, published and has
become--it's a done deal.
So, I mean, we're not rediscovering the
wheel here. I don't understand this. I mean, you
know, if they're mandating these patients have to
go on long-term estrogen because it's
incorporated--and then use the patch, then
basically aren't you back to where you were with
the discussion of the safety of estrogen?
DR. LOCKWOOD: I think what Linda's
saying--if I can put words in her mouth--is that we
know that estrogens are thrombogenic. This drug is
being used with estrogen. So the question is: does
it make the estrogen more or less thrombogenic.
And there's certainly some evidence, even from the
WHI, that progesterone may actually make it more
thrombogenic. But we don't know because, 1) there
are not enough numbers and, b) I'd like to have a
few more of these--
DR. LIPSHULTZ: But then you're putting a
product on top of a base that already has warnings.
DR. LOCKWOOD: Right, but it would be like
384
using nitroglycerine and a calcium channel blocker
for angina. And the question is, all right,
nitroglycerine works, why not give another
vasodilator? That should work even better.
But what happens when you do that is--I'm
making this up. The drugs are perfectly fine.
[Laughter.]
I'm on a roll, though. But when you use
them together there's some cardiotoxic effect that
was not predicted, or you stop perfusing the brain
and something bad happens.
And so, in fact, since you're labeling it
to be used together, you have to understand the
potential for synergistic or additive effects that
are toxic.
DR. LEWIS: And it's not just that, but
it's also that there all these warnings about
"long-term"--quote-unquote--usage, because that's
what's been linked to breast cancer. And we really
don't have any data on that from the sponsor.
And--you know, do we have concerns? Obviously, the
Women's Health Initiative, if there's one thing it
385
taught us it's that estrogen is different than
estrogen plus progestin. So what is estrogen plus
androgen? We really don't know.
DR. GIUDICE: Dr. Tobert and then Dr.
Dickey.
DR. TOBERT: Well, I think this discussion
is actually very important. Ad maybe we can just
clear it up.
My understanding--and maybe I'm completely
wrong about this--is that the sponsor is
recommending that if a woman is taking estrogen
anyway--and I think they mean oral or transdermal
estrogen. The label isn't totally clear on that.
If she is taking that anyway, she can take the
patch. If she stops taken systemic estrogen, she
should stop taking the patch.
And the intent, at least, is not to
encourage any more use of estrogen than would
otherwise occur. And to that extent, it's not--the
safety or otherwise of estrogen and progestin is
not relevant.
DR. LEWIS: I don't think we can conclude
386
that. I mean, all the data we have on efficacy are
from the combine usage. So, to me, looking at that
data, if I'm going to prescribe the product, I
would assume that there's something about having
adequate estrogen on board that helps make this
product work, and I would replace the patient with
estrogen first--or I would consider it in somebody
who's already placed on estrogen.
So--
DR. TOBERT: Would you treat the patient
more or longer with estrogen? Or would you put a
patient on estrogen just so you could use the
patch? Because if you would, then it's a different
question.
DR. LEWIS: If you're a literal,
evidence-based person, you would only use it
somebody who's already on estrogen. Obviously,
they're studying it in patients who are not on
estrogen, but we don't have those data to judge
yet.
DR. TOBERT: Because if it causes more use
of estrogen, that's a whole different and important
387
question. But it's a separate question.
DR. GIUDICE: Dr. Dickey has a comment.
DR. DICKEY: I think this alludes to the
same thing we were just discussing--but this new
drug application says it's only fro the surgically
menopausal women, and yet when we talked about the
long-term safety, it was clear that we're going to
be looking at both surgically menopausal women and
naturally occurring menopausal women--I presume
without redoing any of the baseline data? So we'll
begin to look then at some of these women may be
getting three drugs: progestin, estrogen and
androgen?
DR. MONROE: Well, the sponsor has parallel
studies going on in women that are naturally
menopausal who are presently on estrogen and
progestin, and are now taking testosterone on top
of that. And that's a group that I presume that
they may eventually want to expand the claim for
use in, depending on how the data is.
Similarly, they've indicated that they're
looking at a population that's just taking
388
testosterone alone, or the drug might work equally
well without the risk of estrogen.
So those are all things that we'll learn
perhaps in the future.
DR. GIUDICE: But our charge is to look at
the data that have been presented today, in the
context of the indication.
DR. NISSEN: Folks, let's not make this any
more complicated than it already is--you know? The
question, I think, is very clearly stated. I mean,
what we know is exactly what was done in these two
trials. This is the database that we have. And
we're being asked whether, in this population,
treated in this way, whether we have an adequate
safety database in order to make a decision.
I think that couldn't be more clear. And
I don't think we have to dance around it.
VOICE: [Off mike.] And we can answer yes
or no.
DR. GIUDICE: In fact, let's do that right
now.
[Laughter.]
389
Now we'll start on this side of the table.
Dr. Montgomery-Rice.
DR. MONTGOMERY-RICE: No.
DR. STANFORD: No.
DR. MERRITT: No.
DR. NISSEN: And I must qualify again--
[Laughter.]
DR. GIUDICE: A simple yes or no will do.
DR. NISSEN: No, I really am--you know,
first of all, I have to earn my $164 salary for
coming here.
[Laughter.]
And, you know, these hormones have
widespread biological effects, affecting virtually
every tissue. And the heart is obviously one of
the target organs, which is why you have a
cardiologist sitting on the panel.
And I just have to review for a
moment--just give me two minutes--to say that we
have at least four or five pieces of data to
suggest that there is a high probability of an
excess cardiovascular risk with this product. They
390
include the evidence that endogenous testosterone
levels are associated with coronary disease. We
have the evidence from Dr. Soule's evaluation that
shows that outliers are much more likely to have
elevated lipids, elevated blood sugars, worse
insulin resistance, increasing insulin levels;
blood pressure changes of several mm/Hg, and
sometimes in the range of 10 to 19 mm--which is
highly associated with cardiovascular risk;
elevated fibrinogen.
We also have the data on the known risks
of estrogen-progesterone in the WHI study.
And so given that, the safety data base
that we have of 500 patients, in my estimation, is
at least an order of magnitude. I'm talking about
10-fold too small for us to assess a therapy that's
likely to be used in millions of patients.
And so I think this answer is very, very
clear. This is a much too small of a safety
database--
VOICE: That's a no?
[Laughter.]
391
DR. NISSEN: --that's a no--for any
reasonable assessment of cardiovascular risk.
But I do think--I want to put this on the
record, because there's a very specific reason why
it's too small. If we didn't know anything about
these biological effects of hormones, maybe it
would be okay. But we know a lot, and what we know
doesn't suggest that it's a particular safe
approach.
DR. GIUDICE: And your vote is?
[Laughter.]
DR. NISSEN: No.
DR. GIUDICE: Thank you.
DR. PATRICK: Simply no.
MS. SOLONCHE: No.
DR. LIPSHULTZ: No.
DR. LEWIS: No.
DR. LOCKWOOD: No.
DR. GIUDICE: No.
DR. DICKEY: No.
DR. BURNETT: No.
DR. TULMAN: No.
392
DR. HEIMAN: No.
DR. EMERSON: No.
DR. DORGAN: No.
DR. MACONES: No.
DR. GIUDICE: For this vote it was
completely unanimous for no.
The previous question was three no's and
14 yeses.
Thank you.
The third question has three parts. The
first part is: Are the safety concerns or
unanswered questions associated with use of TTS in
combination with estrogen that need to be studied;
for example, questions about cardiovascular or
breast cancer outcomes, or questions about risks
and benefits in populations who are likely to use
this product off-label?"
So ask the committee--I think we've
actually discussed this at quite a bit of length.
But I think the agency is asking us to state what
these particular concerns are.
So would someone like to begin the
393
articulation of these?
Dr. Lockwood.
DR. LOCKWOOD: So I guess the answers would
be: yes, there are safety concerns; and yes, they
haven't been addressed.
And the three I'll focus on--and I'm sure
other people will add--are the risk of venous
thrombotic events; pregnancy exposure--obviously,
given my concerns; and breast cancer.
And I think that in terms of venous
thrombotic events, the follow-up that would be
required--minimum follow-up--would be I'd like to
protein-S activity values assessed. You probably
have the data---you probably have the blood samples
ready to be run. And I think it would be
reasonable to follow up in the context of the
actually post-approval study, the incidence of
venous thrombotic events, a) because I am not
overly concerned if protein-S activity levels are,
in fact, normal, that there will be a strikingly
elevated occurrence of venous thrombotic events,
over that already anticipated with estrogen. So I
394
would like to see protein-S activity measure. If
it were normal, then I think it would be reasonable
to not change placebo versus treatment to do a
post-approval follow-up in the context of the
larger study proposed.
Secondly, I'm very concerned about the
potential for use in pre-menopausal women, and
concerned about the potential for exposing fetuses.
Animal studies, particular in primates; extended
virilization--levels, if possible, by labeling the
testosterone; core blood; and then, obviously, some
registry to follow up the fetuses that are exposed.
Lastly, breast cancer. And I think,
again, the evidence is, in my mind, inconsistent
toward an association. My bias is actually that
androgens are protective, to be honest with you.
But, again, I think follow-up in a broad
post-approval study, looking specifically at the
incidence of breast cancer.
DR. GIUDICE: Before we go around with
additional recommendations, we need to go around
just to answer a yes or a no question. That's part
395
a). Part b) is the actual--and I didn't make that
clear--part b) is the actual delineation of what
the recommendations are.
So I would like to start on this side of
the table.
Dr. Macones?
DR. MACONES: Yes.
DR. DORGAN: Yes
DR. EMERSON: Yes.
DR. HAGER: Yes, there are safety concerns.
DR. TULMAN: Yes.
DR. BURNETT: Yes.
DR. DICKEY: Yes.
DR. GIUDICE: Yes.
DR. LOCKWOOD: Still yes.
DR. LEWIS: Yes.
DR. LIPSHULTZ: Yes
MS. SOLONCHE: Yes.
DR. PATRICK: Yes.
DR. NISSEN: Simple yes.
DR. MERRITT: Yes.
DR. MONTGOMERY-RICE: Yes.
396
DR. GIUDICE: Yes.
That was unanimous. Thank you.
Now we'll do 3-b). In addition to Dr.
Lockwood's comments, are there other suggestions?
Yes, Dr. Nissen.
DR. NISSEN: It seems to me that we'd need
to know about the post-menopausal patient--not the
surgically menopausal patient, but the naturally
menopausal patient before approval. Because the
likelihood that those patients would be exposed is
very, very high. And keep in mind that the
naturally menopausal patient has coronary disease
risks that are now approaching that of men. And so
now you're talking about an entirely different risk
category of patients.
And so until we've studied--until the
sponsor has studied that population--you know,
we're going to end up potentially exposing very
large numbers of post-menopausal women to hormonal
therapy for which we really don't have any evidence
of whether it does or not increase cardiovascular
risk.
397
DR. GIUDICE: Yes, Dr. Lewis.
DR. LEWIS: Yes, I'd like to see greater
effort be made to enroll a larger population of
African-American women. They're disproportionately
affected by problems that engender hysterectomies
and oophorectomies. They're at high risk for
cardiovascular disease. And so I think it's going
to be crucial to have a much better representation
in the studies.
DR. GIUDICE: Thank you.
Dr. Hager.
DR. HAGER: Yes, and I would add Hispanics
to that. Only 3 percent of the population were
Hispanic. So I think minority groups need to be
enrolled.
And I would just add to the comments that
was previously made--and I'm dropping down a little
bit, I realize--but I think that in light of the
potential for off-label use of this product, we
must have information from pre-menopausal women,
and menopausal women. We must have information
from women who are not only taking estrogen, but
398
are also taking with the progestin.
DR. GIUDICE: Thank you.
Yes, Dr. Dorgan.
DR. DORGAN: I'm concerned about the
potential for breast cancer risk. Regrettably, in
animal models--in rodents, particularly--the
hormonal relationships like DHEA and breast cancer
are not the same as you see in humans. And so
using animal models might not suffice.
Also, we don't know the mechanism--we see
observational data that women with elevated
testosterone are at increased risk of breast
cancer, but I agree: we don't know if it's causal,
we don't even know a potential mechanism.
And so we don't have any really good
intermediate markers that we could suggest to you.
I would hate to see FDA approving it and
then using post-marketing drug surveillance as our
only way of evaluating potential risk for breast
cancer, because breast cancer doesn't occur
just--you know, it's a long process. And once that
process is started, stopping the testosterone is
399
not going to make the risk disappear like it's
disappearing for acne and some of the other adverse
effects. It's going to take a while before the
woman's risk is lower.
I'd love to see randomized, controlled
clinical trial, if people are interested in
pursuing this further. I think that's the only way
to really answer the question.
We'd prefer not to have any surprises like
we did with the WHI.
DR. GIUDICE: Thank you.
Dr. Patrick.
DR. PATRICK: We didn't discuss this in the
committee, but one of the exclusion criteria was
that there could be no ongoing personal
disturbances in the relationship of these couples.
And I would like to see that a little bit loosened;
meaning, how would this go down in normal life,
where there are few disturbances in menopausal
relationships--living through that.
[Laughter.]
DR. GIUDICE: Yes, Dr. Montgomery-Rice.
400
DR. MONTGOMERY-RICE: I do agree that many
of these things need to be studied. But I also
think you've got to put some of this in
perspective. And putting it in perspective is that
when you think about the indication for the
product, and the indication of the product would be
a symptomatic patient who has tried other ways to
improve her libido, etcetera. And so she then does
a risk-benefit analysis. And women do this every
day when they come in they're having hot flushes.
And they know that, based on the WHI, etcetera,
that there's a potential of an increased risk of
breast cancer if they take estrogen. But their
symptoms are bad enough that they're willing to
take that risk.
So I don't--even though there may be some
increased risk--which I don't believe there
probably is--with testosterone and breast cancer, I
don't think that patients would do any different of
an analysis than if they were coming in for
estrogen for hot flushes, because this--if you were
prescribing this correctly, and if we could control
401
how this was going to really be prescribed, then
here you would have an indication of something that
may actually give patients some--what are we
calling it?--clinical meaningful benefit, in a
patient who is actually symptomatic.
And so patients will make that same
risk-benefit analysis.
However, I do agree: we don't have enough
information for long-term safety.
DR. GIUDICE: Thank you. And that, then,
brings us to part c) which is:"Should these
concerns or questions be studied prior to approval
of the product?"
So we'll go around the room again,
starting with Dr. Rice.
[Pause.]
Yes or no.
DR. MONTGOMERY-RICE: I have to--I mean,
everybody else has had a preference, so I have to
give my little say beforehand.
I do think this drug is going to require
402
more study. I think that it should. However, I
also say that in light of the fact that in an
everyday sense, when I see patients who are coming
in already using therapies that have not been
tested in any type of market, etcetera, I was
hopeful that we would have a product that we would
be able to give patients that we know the risk and
the benefits. And this product does allow that
potential.
But I do, at the minimum believe that we
need to look at it in the natural menopausal
patient, who is going to have estrogen and
progesterone on board already, to understand the
concomitant hormone risk.
DR. GIUDICE: So your answer is--
DR. MONTGOMERY-RICE: Yes.
DR. GIUDICE: Thank you.
DR. STANFORD: And I will say yes, too, but
I think it's an interesting issue for FDA policy,
in a sense, a moving target in terms of what level
of study is required for a drug.
I think the WHI has changed what probably
403
should be that level. So that's why I say yes.
DR. MERRITT: I would say yes.
DR. NISSEN: Yes.
DR. PATRICK: Yes
MS. SOLONCHE: Yes.
DR. LIPSHULTZ: I want to say yes, but I
just want to add something here.
I mean, I understand there must be
tremendous pressure about this drug. I mean, it's
been expedited. It's the first drug available for
the treatment of sexual dysfunction in women. And
what concerns me the most--aside from what's been
stated--is that there's going to be tremendous
off-label use. And I just don't see how this can
go to market without the data that the company is
already--it looks like they're almost finished
with, on natural menopausal women, because they're
the ones who are going to be taking this drug, as
well.
So my answer is yes, and I'd like to see
that data. And I think it's to everybody's good to
wait a little bit.
404
DR. LEWIS: I agree. I would say yes, and
that's the data I want to see: the data in women
who have a uterus.
Is it going to be affected by adding a
progestin? Do you need a progestin? Maybe you
don't. Maybe if you get an androgen effect on the
uterus, maybe estrogen and androgen alone are
enough. But we don't know. We just need more
data.
DR. LOCKWOOD: Yes, I agree with my two
colleagues, plus I'd like to see the protein-S data
and the animal studies in pregnancy.
DR. GIUDICE: I vote yes.
DR. DICKEY: Yes.
DR. BURNETT: Yes.
DR. TULMAN: I vote yes. And also I don't
think 20, 24, 26-week data is "long-term data.
DR. HEIMAN: The time to gain these data
are before approval. So I do vote yes. I do think
we need adequate long-term data that demonstrate
both efficacy and safety. I do think we need it in
women who are on estrogen-testosterone, as well as
405
estrogen-progestin-testosterone. And we need an
adequately powered study.
DR. EMERSON: Yes.
DR. DORGAN: Yes.
DR. MACONES: Yes.
DR. GIUDICE: Thank you.
The first part of--oh--what were the
results? Unanimous.
The first part of 3-c) is: "If yes,
what studies do you recommend? And please comment
on study populations, designs, endpoints, etcetera.
We have heard about increasing minority
populations. Perhaps one or two people from the
group could--I'm assuming you're asking for a study
design, and what kinds of studies to be done before
approval can be given. Is that correct?
DR. MONTGOMERY-RICE: Do we know what this
natural menopause study that they're doing--do we
know what the enrollment is on that? The one
that's currently going on?
DR. GIUDICE: Do we know more details about
it?
406
DR. MONTGOMERY-RICE: Do we know more
details of that study?
DR. MONROE: Why don't we let the sponsor
give you those details.
DR. LUCAS: We have two trials. The first
trial was the one that I presented--the Natural
Menopause I. We had about 550 patients that were
randomized one-to-one.
The second trial, which is still ongoing
but nearing completion, is a two-to-one
randomization, and has just a little over 600
patients. So 400 patients will be on testosterone,
and about 100 will be on placebo, and those
patients have endometrial biopsies.
We are to just about 300 matched pairs.
So that would be 200 in testosterone, and 100 in
placebo.
DR. MONTGOMERY-RICE: Is it a year-long
study?
DR. LUCAS: Yes, it is.
DR. MONTGOMERY-RICE: Okay.
DR. GIUDICE: And is that with or without--
407
DR. LUCAS: It is being extended, like our
surgical menopause, so the patients are rolling
over into an extension.
DR. GIUDICE: And is this with or without
estrogen?
DR. LUCAS: It is with and progestin.
DR. MERRITT: And what dose of estrogen,
please?
DR. LUCAS: Like our surgical menopause
patients can use, you know, any approved estrogen
does. But it is with continuous progestin.
DR. LEWIS: They both have continuous
progestin?
DR. LUCAS: Both the Natural Menopause I,
and the Natural Menopause II--both have continuous
estrogen-progestin.
DR. GIUDICE: Yes, Dr. Nissen.
DR. NISSEN: Again, this is setting the bar
very high, but I think it's important that the
committee understand, and the FDA understand this
issue.
Post-menopausal women, you know the
408
leading cause of death is cardiovascular disease.
And even a hazard ratio of 1.1 or 1.2, when a
million or more women are likely to be exposed,
represents a huge burden of morbidity. And
therefore I believe that you need a prospective,
adequate sized, long-term study. I would do it in
aspirin-eligible women; in other words, women who
have enough risk that they would require aspirin
for prophylaxis.
Now, we can discuss Framingham Risk Scores
and all that, but I think you have to have a high
enough risk population to simulate what could
happen in the general population if this agent were
used in a widespread way.
And I'd be happy to work with you, and
I'll give you some thoughts from cardiovascular
side about how you do that. But it's not going to
be 600 patients. It's going to be 5,000 patients,
or 10,000 patients.
And I recognize how high I'm setting the
bar, but I must say this very clearly: the risk
that was seen with Vioxx was very modest. But when
409
you translated that to 105 million prescriptions in
20 million Americans, it represented an enormous
burden.
The potential for this agent to increase
the risk of cardiovascular morbidity and mortality
is substantial, and can only be answered--not the
post-market surveillance. We know how badly that
works. It has got to be done prospectively.
And I'm not devaluing the importance of
this symptom and its treatment. But I also don't
want to expose several million American women to
the risk of heart attack and stroke, with their
devastating consequences, in order to have one more
sexual experience per month increase.
It is not an acceptable trade-off, and we
cannot allow this to move forward until we have
such data.
DR. GIUDICE: Dr. Tobert.
DR. TOBERT: Well, I hear what Dr. Nissen
is saying. I think the committee needs to consider
the practicability of doing a randomized controlled
trial. I do have quite a bit of experience with
410
long-term, randomized trials to look at
cardiovascular outcomes. And I can tell you to do a
study that would rule out a 10 percent increase in
the risk of major vascular events would probably
require 100,000 patients--especially since the
patients--I mean, really, to make it possible,
you'd have to use elderly women, but elderly women,
we've heard, do not use this product very much.
I think it's simply undoable. You cannot
do this study. It's not practical--unfortunately.
DR. GIUDICE: I have a question for the
sponsor, and that is: since the use of estrogen in
post-menopausal women--and even peri-menopausal
women--is decreasing, with the largest use in the
peri-menopause, have you considered a trial with
just testosterone alone, as opposed to adding the
estrogen? Because, in reality, again, one is then
committing long-term use, if one is going to be an
outcomes-based type of prescriber.
DR. LUCAS: Yes, we are. We're just
finishing recruiting a study that's enrolled both
surgical and naturally menopausal women not on
411
concomitant estrogen or progestin. Vaginal
estrogen is allowed.
Part of the reason is to rule out any
other cause of sexual dysfunction.
DR. GIUDICE: And how large is that?
DR. LUCAS: About 750 women will be
enrolled in this study, and we're looking at two
doses of testosterone--the 150 mcg patch, and the
300 mcg--and placebo. And it will be one year in
duration.
Double-blind.
DR. GIUDICE: Thank you.
DR. TOBERT: May I just add to my previous
remarks the fact that--I mean, there is an
implication that if a huge trial like this is
demanded for a testosterone patch--testosterone, of
course, is a natural hormone that we all have. And
the implication of that is that any product--any
new product--that binds to a hormone receptor--an
agonist, an antagonist, not necessarily in the
reproductive area--the same sort of thing could be
demanded.
412
I don't quite see what is so hazardous, a
priori, about testosterone that one should demand
such huge and--as I say--in any case, impossible
trials.
DR. GIUDICE: Dr. Nissen.
DR. NISSEN: Well, those of us that are
male, in addition to suffering from alopecia, which
I fortunately have enough testosterone to have, we
have a cardiovascular risk which is substantially
greater than that of women. Why? And a lot of us
think that testosterone is an atherogenic
substance. And there's lots of evidence to suggest
that. And until proven otherwise, it must be
assumed to be the case.
And so when you have a therapy here that
you're going to, again, expose a lot of people to,
Jonathan, you have to know this. Because on a
population basis, this can involve tens of
thousands of myocardial infarctions and potentially
deaths.
So I just don't think we have the evidence
of the safety of giving women testosterone that we
413
need in order to know that we're not going to make
them have cardiovascular risk rates that look like
those rates of men.
And until we know that, I think we've got
a real problem here.
DR. TOBERT: Can I respond to that/ I mean,
there is no evidence that men have higher rates of
coronary disease because of their circulating
androgens. All attempts to correlate circulating
androgens--whether it's free or total--to the risk
of coronary or cardiovascular disease have failed.
I mean, there's just no correlation. It may have
something to do with the testosterone surge that
occurs perinatally, but certainly it doesn't have
anything to do with the levels that we have as
adults.
In any case, as I say, Steve--how are you
going to do this trial? It's undoable.
DR. NISSEN: I don't think it's undoable,
because what you're doing whenever you face a
situation like this is you go to a well-defined,
high-risk population, and you find out--if you can
414
show that the agent--what the hazard ratio is in
the higher-risk individuals, then you can feel very
comfortable in giving it to lower risk individuals.
And so you define a group of people that are at
relatively high risk: women's with lots of risk
factors. And you can have an enriched population.
And that's what's done in lipid-lowering trials all
the time. And you know, you're involved the SEARCH
trial. Those are not done in normal, low-risk
people. They're done in high-risk people.
DR. TOBERT: All right. They all have an
MI. And you can't do--the people who have an MI
are not going to be taking this patch. There's an
incompatibility.
DR. GIUDICE: Dr. Rice.
DR. MONTGOMERY-RICE: I mean, I think we
need to be realistic here. People who are at high
risk for cardiovascular disease are not people who
are concerned about--necessarily--increasing their
level of libido to the point where they would take
that risk--
DR. NISSEN: That--
415
DR. MONTGOMERY-RICE: --let me finish,
please--where they would take that risk of being
enrolled in a trial. And I don't even know how
realistic this is to get through the IRB.
So, while I hear what you're saying, that
there is some risk--or potential risk--I don't want
us to offer or suggest to the FDA that we should do
some trial that's unrealistic to ever have
performed.
If you think about participation in
clinical trials, there are major barriers getting
women to participate in clinical trials. So you're
talking about not only finding women who meet this
criteria of this defined sexual dysfunction, but
then on top of that we want you to also be at risk
for having a heart attack so we can make sure that
we are not increasing your risk of having the heart
attack.
I want us to be realistic here. And I
don't think you're being realistic.
DR. NISSEN: Let me reassure you that
patients that I see that have coronary heart
416
disease, that have had an angioplasty, by-pass, or
myocardial infarction frequently--one of the first
questions they ask when they come back in to see me
is when can they resume sexual activity. So the
idea that sex stops with heart disease is simply
wrong.
And I think we can--I'm convinced that we
could define for you a population, not of whom
would be post-MI. Some of them would be multi-risk
factor patients, where the risk would be high
enough that you could get a signal in a reasonable
size trial.
Now, where you set the hazard ratio has to
be discussed. But it seems to me that that's the
prudent thing to do.
DR. GIUDICE: Does the FDA feel that it has
had sufficient suggestions--
[Laughter.]
--for approaches to clinical trials?
And we would assume that you would be
working with the sponsor in this, as well.
DR. GRIEBEL: Yes.
417
DR. LEWIS: could I just bring up one more
little point along those lines?
I tend to agree with Dr. Montgomery-Rice
about the--you know, it's a different population.
But maybe just one small thing to throw in there:
women who have had a hysterectomy do tend to be at
somewhat higher risk for cardiovascular disease
already. So if there's any way to work within that
population, perhaps you could prove the point a
little bit easier.
Another surrogate that you could look at
is C-reactive protein levels.
DR. GIUDICE: Yes, Dr. Heiman.
DR. HEIMAN: I would just like to speak to
women's brains--and that would be since there were
some significant effects with cognitive change in
the WHI study, that at least some study along in
here tries to track that. It hasn't been raised.
There may be no risk, there may even be a benefit.
But somebody should tract that because if anything
that women are worried about--other than their
overall life and life span--it has to do with their
418
cognitive functioning.
DR. LIPSHULTZ: Didn't you have some data
on that? Wasn't there some data on cognitive
function in one of your slides?
DR. HEIMAN: Whatever there was, there
isn't enough, in my opinion.
DR. LOCKWOOD: I think it was aggression,
and anxiety of something.
DR. HEIMAN: That's not what I mean. I
mean in the dementia direction, as opposed to
simply mood. Mood is something else which is
interesting, but I'm talking about actual cognitive
functioning in terms of processing information.
DR. GIUDICE: Thank you.
Yes?
DR. HAGER: Could we also suggest that DHEA
be used as a marker, in addition to testosterone,
total and free?
DR. GIUDICE: We can suggest that. Okay.
DR. LIPSHULTZ: [Off mike.] [Inaudible.]
DR. HAGER: Well, we have heard today that
testosterone is the indicator of altered sexual
419
function in women. And when they have low
testosterone, they have low sexual function. And
that really is not true.
And I think that there's some information
that DHEA may be more tightly attached to altered
sexual function. And I'm just saying that we could
use that as a marker.
DR. GIUDICE: Okay.
3(c) part (ii) is if we had an answer of
"no" to question 3(c)--and it was unanimously
"yes"--however, I think there is the issue of the
claims-based cohort study, which we have discussed.
And I want to bring to the attention of the
committee, and ask you if we need to have any
further discussion about this?
Yes, Dr. Stanford.
DR. STANFORD: Well, I'd just say that
ideally a study would be done along the lines of
what Dr. Nissen is suggesting. I am not totally
clear on how feasible it is. And I guess that
would have to be a judgment call. But if it is
feasible, and you did that, I don't think you'd
420
need the post-marking study.
DR. GIUDICE: The claims-based study that
was proposed.
DR. STANFORD: The claims-based study--yes.
DR. GIUDICE: Yes, Dr. Emerson.
DR. EMERSON: I just felt--and I think this
is pretty much a moot point now--but I felt that
the level of detail that was provided about how
such a study would be done made it just completely
impossible to judge whether that would have been
adequate or not, because there would be a whole
lot, in terms of how you would match patients, and
whether you're getting comparable patients who were
on Intrinsa versus not. And without further
information I just don't think anything could be
said.
DR. GIUDICE: Okay. Thank you.
So we now reach our fourth and final
question, and that is: "Are the efficacy and safety
data adequate to support approval of TTS?"
And I will begin on this side of the
table.
421
Dr. Macones.
DR. MACONES: No.
DR. DORGAN: No
DR. EMERSON: No.
DR. HAGER: No.
DR. TULMAN: No.
DR. BURNETT: No.
DR. DICKEY: No.
DR. GIUDICE: No.
DR. LOCKWOOD: No.
DR. LEWIS: No.
DR. LIPSHULTZ: No
MS. SOLONCHE: No.
DR. PATRICK: No.
DR. NISSEN: No.
DR. MERRITT: No.
DR. MONTGOMERY-RICE: No.
DR. GIUDICE: It's unanimous.
I want to thank the committee for their
hard work, and also our participants in the open
public forum.
And this now concludes our Advisory
422
Committee meeting. Thank you.
[Whereupon, at 4:22 p.m., the meeting was
adjourned.]
- - -