Monday, March 23, 2006

8:00 a.m.

Hilton Hotel
The Ballrooms
Gaithersburg, MD


Wayne K. Goodman, M.D., Chair
Cicely Reese, Pharm.D., Executive Secretary


Jorge Armenteros, M.D.
Jean Bronstein, R.N., M.S.
Andrew C. Leon, M.D.
Daniel S. Pine, M.D.
Delbert Robinson, M.D.
Philip Wang, M.D., Dr.P.H.
Barbara Wells, Pharm.D.


Michael Bigby, M.D.
Deborah Dokken, MPA
Richard Malone, M.D.
Cynthia Pfeffer, M.D.
Marsha Rappley, M.D.


Dilip Mehta, M.D., Ph.D.


Robert Temple, M.D.
Thomas P. Laughren, M.D.
Paul J. Andreason, M.D.
Glenn Mannheim, M.D.

Call to Order and Opening Remarks,
Wayne K. Goodman, M.D., Chair 5
Conflict of Interest Statement,
Cicely Reese, Pharm.D.,
Executive Secretary 8
FDA Introductory Remarks,

Thomas P. Laughren, M.D. 11
FDA Presentation:
FDA Clinical Review,

Glenn Mannheim, M.D. 16
Modafinil for Treatment of ADHD,

Paul J. Andreason, M.D. 37
Serious Adverse Cutaneous Reactions to Drugs,

Michael E. Bigby, M.D. 64
Sponsor Presentation:

Victor Raczkowski, M.D. 82
Overview of ADHD,
Joseph Biederman, M.D. 87
Clinical Pharmacology and Efficacy,
Lesley Russell, M.R.C.P. 103
General Safety,
Srdjan Stankovic, M.D. 121
Benefit-Risk Conclusions,
Lesley Russell, M.R.C.P. 163
Questions from the Committee to the
FDA and Sponsor 168

C O N T E N T S (Continued)
Open Public Hearing 210
Committee Discussion 229
Questions to the Committee 257


Call to Order and Opening Remarks

DR. GOODMAN: We expect a few more people
to join us around the table but I want to make sure
that we start on time. Welcome, everyone, to the
Psychopharmacologic Drugs Advisory Committee, or
the PDAC. We have been asked today by the FDA to
advise them on a new drug application for modafinil
in the treatment of attention deficit hyperactivity
disorder, ADHD. Most of the questions, as will be
articulated by the FDA, concern safety issues.

Yesterday there was a meeting of the
Pediatric Advisory Committee which discussed a
range of safety issues concerning medications used
in the treatment of ADHD, the stimulants as well as
Strattera, and actually some data emerged on
modafinil as well during those discussions. I was
present as an observer during those meetings. I am
glad I was there. Some of the members of the
committee that are here today were also present
yesterday so I think a lot of heavy lifting was
done yesterday on some of these important side

effect issues that will help inform us in our
deliberations today.

My remarks are going to be unusually
brief, in part because my voice is strained. My
voice has not been cooperating for the last few
days. In fact, sometimes I am not sure it is my
voice--I don't know what kind of symptom that would
mean. But we have a backup plan. Danny Pine, when
he comes here, in case my voice fails, he will
become my voice.

I also want to put you on notice that
Cicely Reese may deliver at any moment! I am not
kidding! So, we have plans for her transportation
and replacement should that occur. Please bear
with us under these circumstances.

Now I would just like to go around the
table and ask everyone to introduce themselves.
Let's start from the FDA end.

DR. LAUGHREN: Tom Laughren, from the
Division of Psychiatry Products.

DR. ANDREASON: Paul Andreason, Division
of Psychiatry Products.

DR. MANNHEIM: Glenn Mannheim, Division of
Psychiatry Products.

DR. BIGBY: I am Michael Bigby,
dermatologist from Boston.

DR. RAPPLEY: Marsha Rappley,
Developmental Behavior Pediatrics, Michigan State

DR. WANG: Phil Wang, psychiatrist and
epidemiologist from Harvard Medical School.

DR. REESE: Cicely Reese, executive

DR. GOODMAN: Wayne Goodman, chair of this
committee as well as chair of the Department of
Psychiatry, University of Florida.

DR. LEON: I am Andrew Leon, professor of
biostatistics at Cornell Medical School.

DR. ROBINSON: I am Delbert Robinson. I
am a psychiatrist at the Zucker Hillside Hospital
and the Albert Einstein College of Medicine.

DR. PFEFFER: I am Cynthia Pfeffer, child
and adolescent psychiatrist at Weill Medical
College of Cornell University.

DR. ARMENTEROS: Jorge Armenteros, child
and adolescent psychiatrist in Miami, Florida.

DR. WELLS: Barbara Wells, I am dean of
the School of Pharmacy at the University of

MS. DOKKEN: I am Deborah Dokken. I am
the patient family rep. on the Pediatric Advisory

DR. MALONE: I am Richard Malone, a child
psychiatrist from Drexel University College of

DR. MEHTA: Dilip Mehta, retired physician
from the drug industry. I am the industry
representative on the committee.

DR. GOODMAN: Thank you all very much. I
think Daniel Pine will be joining us shortly. I
would now like to turn the microphone over to
Cicely Reese to go over some housekeeping,
particularly the conflict of interest statements.

Conflict of Interest Statement
DR. REESE: The following announcement
addresses the issue of conflict of interest and is

made part of the record to preclude even the
appearance of such at this meeting. Based on the
submitted agenda and all financial interests
reported by the committee's participants, it has
been determined that all interests in firms
regulated by the Center for Drug Evaluation and
Research present no potential for an appearance of
a conflict of interest at this meeting with the
following exceptions:

In accordance with 18 USC, Section
208(b)(30, Dr. Wayne Goodman has been granted a
full waiver for his employer's related contract
with a competitor, funded between $100,001 and
$300,000 per year. His employer also has related
contracts with another competitor, funded for less
than $100,001 per year.

Dr. Andrew Leon has been granted a waiver
under 21 USC, 355(n)(4) for his ownership of stock
in a competitor. This stock is valued from $5,001
to $25,000.

A copy of the waiver statements may be
obtained by submitting a written request to the

agency's Freedom of Information Office, Room 12A-30
of the Parklawn Building.

We would also like to note that Dr. Dilip
Mehta has been invited to participate as an
industry representative, acting on behalf of
regulated industry. Dr. Mehta's role on this
committee is to represent industry interests in
general and not any one particular company. Dr.
Mehta is retired from Pfizer.

In the event that the discussions involve
any other products or firms not already on the
agenda for which the 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 whose products they wish to comment upon.
Thank you.

DR. GOODMAN: Dr. Daniel Pine just joined

us so I wonder if you could introduce yourself.

DR. PINE: Danny Pine, Chief of
Developmental Studies, Mood and Anxiety Disorders
Program, National Institute of Mental Health
Intramural Research Program and I am a child and
adolescent psychiatrist.

DR. GOODMAN: In a moment I will turn over
the floor to Dr. Laughren who will give us the
introductory remarks. I think for all of us who
have read through these briefing materials one of
the issues that emerges, that didn't surface during
yesterday's discussions, are questions about
dermatological reactions. I see that we will also
have the benefit of an intensive review of those
issues as well to help us in our decision-making
today. So, Tom, would you please come forward?
Thank you.

FDA Introductory Remarks

DR. LAUGHREN: I would like to welcome
everyone to today's meeting. Before I introduce
the topics for today's meeting I would like to
acknowledge the service of one of your colleagues

on this committee whose term is ending in June, and
that colleague is Wayne Goodman. This has been a
particularly busy time for the committee, as you
know, and Wayne has, of course, been the chair of
the committee for much of this time. Serving on
this committee, again as all of you know, is a very
demanding and sometimes stressful task and I hope
that you all understand how much we appreciate the
help that you give us.

Now, Wayne told me after the September,
2004 meeting on antidepressants and suicidality in
pediatric patients that he didn't have any friends
anymore in the academic and clinical community. I
just want to assure him that he always has friends
here, at FDA.


So, thank you, Wayne. This is a small
token of our appreciation.

DR. GOODMAN: Thank you very much. I used
to have a voice before I started this!


DR. LAUGHREN: Now, on to the topic for

today's meeting, we are going to focus on NDA
20-717, supplement 19. This is for modafinil in
the treatment of attention deficit hyperactivity
disorder. As you know, modafinil is marketed as
Provigil to improve wakefulness in adults with
excessive sleepiness associated with narcolepsy,
obstructive sleep apnea syndrome and shift work
sleep disorder. It is a Schedule IV drug and the
recommended dose in these disorders in adults is
200 mg.

Now, Cephalon has provided us data in
support of a claim for the safety and the
short-term effectiveness of modafinil in the
treatment of ADHD at a slightly higher dose, at a
dose of 340 mg per day in children less than 30 kg
and 425 mg per day in children greater than 30 kg.
This supplement was submitted in December of 2004
and, as you know, we issued an approvable letter in
October of last year.

Though we did issue an approvable letter,
the letter addressed three concerns that we wanted
to have further addressed. One of those was

serious skin rashes; a second was psychiatric
adverse events; and, finally, there were three
patients with transaminase elevations for which we
wanted additional data.

The sponsor responded to our approvable
letter, in November of last year, and today you
will hear from several FDA staff. You will hear
from the primary reviewer, Glenn Mannheim who, as
you know having seen his review, has recommended
against approving this drug based on his concerns
about rash and several other adverse events.

You will also hear from Dr. Paul Andreason
who will provide some additional comments on
safety. Our presentations are going to focus
entirely on the safety issues because we agree with
the company on efficacy. But you will hear from
the company on efficacy and, as well, you also have
our reviews.

In addition, we have obtained advice on
the dermatologic problems from our own internal
consultants from Dermatology. You have their
reviews, and Dr. Markham Luke, from the Dermatology

Division, is here to address any questions you
might have. In addition, we have Dr. Michael
Bigby, who is the chair of the Dermatology Advisory
Committee, who will be making a presentation on
serious drug-related rashes and he will be
participating in the discussion as well.

Now, I want to be clear that the Division
of Psychiatry Products has not reached a conclusion
yet about this application. We have these concerns
and that is precisely why we are coming to you to
ask for your advice. After you have heard the
findings and the arguments we are going to ask you
to vote on two questions. The first question is
focusing on efficacy questions, whether or not you
believe that the company has demonstrated that this
product is effective in the treatment of ADHD.
Secondly, we will ask you to vote on the question
of whether or not it has been shown to be
acceptably safe in the treatment of this disorder.

In addition, we are going to be asking for
your comments on several other issues related
mostly to rash. First of all, if the drug were to

be approved for this indication we would like your
advice on a risk management plan. We would like
your advice on labeling, particularly for rash.
Finally, we would like your advice on any
postmarketing studies that you think might be
useful to further clarify this problem. I think I
will stop there and Dr. Mannheim will be presenting
his findings. Thank you.

FDA Presentation

FDA Clinical Review


DR. MANNHEIM: As Dr. Laughren explained,
I reviewed the initial submission for modafinil for
this indication. I will review with you today the
information specific to safety.


Here is an outline of what I will be
covering. I will be reviewing a little bit of the
background and history of modafinil; an overview of
the safety database; common adverse events in
Cephalon's clinical trial; other adverse events of
significance; psychiatric adverse events; and, most

importantly, rashes and what I think the potential
public health impact may be. Then I will give you
some closing comments, and I will be followed by
Dr. Andreason.


In 1998 modafinil was approved as a
wakefulness-promoting agent in adults with
excessive daytime sleepiness associated with
narcolepsy. Additional indications were granted by
FDA in 2003 for excessive daytime sleepiness
associated with obstructive sleep apnea/hypopnea
syndrome and shift work sleep disorder.

The important thing that I would like you
to notice from this slide is the dose. The
recommended dosing was 200 mg once a day which,
based on a 65 kg adult, comes to about 2.67 or 2.7
mg/kg. I want you to remember those numbers since
we will come to it in other slides.
Recommendations were to give modafinil, Provigil,
as a single morning dose for narcolepsy or
obstructive sleep apnea or for shift work sleep
disorder one hour prior to the start of the work

shift. No additional benefit was shown for doses

more than 200 mg.


The current application is for use of
modafinil in children and adolescents with ADHD.
Two doses have been proposed by Cephalon. For
children less than 65 lbs or 30 kg, they would be
getting a single daily dose of 340 mg. For
children or adolescents more than 65 lbs or 30 kg,
they would be getting a dose of 425 mg a day.

Now, remember the number 2.6. For the
highest dose in children, on a milligram/kilogram
basis, the children would be getting 21 mg/kg or
about 8 times higher than the adult dose. For
those over 65 lbs or 30 kg the highest dose would
be 5.3 times higher than the adult dose. Cephalon
is recommending that children or adolescents start
the drug at initial doses of 85 mg and slowly
titrate up, based on tolerability, by incremental
steps of 85 mg to the targeted dose of 340 mg or
425 mg a day.


This shows the population which was
studied in the submission. It was children and
adolescents 6-17 years of age with DSM-IV ADHD who
attended a full-time school. These were moderately
to severely ill children. They had minimal
learning difficulties. As it relates to adverse
events, children with psychiatric comorbidities
were excluded. Stimulant non-responders were not
allowed in the trial. Those with abnormal
laboratory or medical conditions one month prior to
the start of the study were also excluded.


There are three studies which are called
the pivotal studies for this. Study 309 and 311
were 2 9-week, double-blind, flexible dose, weekly
titration studies. Study 310 was a 7-week,
double-blind, fixed dose study, followed by a
2-week randomized withdrawal to modafinil or
placebo. Children less than 65 lbs went on 340 mg
a day and those over 65 lbs went on 425 mg a day.


This slide shows the total number of

subjects and doses used in the Phase 3
double-blind, placebo-controlled trial. Of note,
420 subjects were treated with modafinil and 213
subjects were treated with placebo. The important
thing to note here is the numbers 102, 256 or 358.
Children and adolescents only received the proposed
labeled efficacious doses.


This slide comes from Cephalon's briefing
document which was submitted for your consideration
by Cephalon. It summarizes the pediatric trials.
The 420 comes from the Phase 3 double-blind
exposure. The number I want to show here is the
number 933 because this constitutes the core safety
database of this supplemental NDA. This slide
indicates that additional 303 children were exposed
to modafinil in an open-label ongoing Phase 3
trial. About 400 other children for obstructive
sleep apnea and narcolepsy are the legacy studies.

As far as the purposes of this submission,
we are only considering the number 933 since we
don't have an integrated safety database for the

other 689 children and they were not part of this
submission. It certainly would be reassuring if
there were no adverse events in these subjects but
we really don't know at this point.


This slide is a little busy and I
apologize for that. It shows exposure to modafinil
and modafinil metabolites and compares it with what
one sees in clinically used doses in adults with
those proposed for children. What I want to bring
your attention to is the exposure of the modafinil
sulfone as measured by the total exposure area
under the curve. In adults, with an initial dose
of 200 mg, the average area under the curve is 38
or close to 40. Going to the highest child,
receiving 425 mg, the area under the curve of the
sulfone is about 250. This is 6.5 times higher
than the exposure seen in adults. Going to the
lowest dose of children receiving 340 mg, the
average area under the curve is around 630. This
is 16 times higher than that seen in adults with
clinical dosing. This cannot be explained by

differences in dosing on a milligram/kilogram
basis. These are clinically used doses and with
them one sees that the sulfone metabolite is much
higher compared to adults.


Now we are going to look at the adverse
event data.


The incidence of common treatment-emergent
adverse events in the Phase 3 double-blind,
placebo-controlled trial is listed. Of note,
insomnia occurred in 27 percent of subjects on
modafinil and 4 percent of subjects on placebo.
Anorexia occurred in 16 percent of subjects on
modafinil and 3 percent of subjects on placebo.
Perhaps associated with that, there was weight loss
in 4 percent of subjects on modafinil and 1 percent
of subjects on placebo.


Notable psychiatric adverse events include
psychosis in 0.5 percent, as listed here, and
suicidal events in 6 subjects, 0.6 percent. The

suicidal events included 5 ideations; 1 attempt.
None were completed. Yesterday Dr. Mosholder
reported on a pooled analysis of the ADHD trials
and that suicidal behavior was infrequent among the
non-medicated ADHD placebo subjects.


Other clinically significant adverse
events which were noted in this trial consisted of
gastric or duodenal ulcers in 2 subjects. One case
of note was a child who was admitted to the
hospital with a moderate metabolic acidosis who had
an H. pylori infection.

There were 9 cases of syncope in the total
exposure. Of note is a child who, according to the
vignette information, had a 40-minute bradycardia,
hypotensive syncopal episode and one week later an
EKG was performed which showed AV dissociation with
adjunctional rhythm. There were 24 children who
were quoted as having asthma.

Of note is a subject in one of the pivotal
trials, 310, who, 8 days after being started on
modafinil at a dose of 340 mg collapsed at school

during gym, stopped breathing momentarily and was
given an inhaler and began breathing normally.
This was diagnosed as an acute asthma attack and
the child was discontinued from the study on day 9.

There were 3 subjects who had dehydration,
and of note is a subject in the open-label
continuation trial who, on day 147 of treatment,
was admitted to hospital with severe dehydration,
moderate ketoacidosis and hypoglycemia which was
found secondary to a strep. throat.

Sixteen subjects had laboratory evidence
of hepatocellular injury based on transaminases
being greater than 3 times the upper limit of
normal. Of note, there were no cases of jaundice
or liver failure, or no significant bilirubin


Now I am going to talk to you about the
rashes but I am not a dermatologist and I am
relying on FDA's dermatologist, Dr. Porres who did
a consult, and someone from FDA from Dermatology is
here to answer some questions.

When you look at all the subjects who were
exposed, rashes were present in 5 percent of all
subjects compared to 4 percent that you saw in the
Phase 3 double-blind, placebo-controlled trial
versus 2 percent in placebo. Only 1 subject
dropped out in the double-blind, placebo-controlled
Phase 3 trial, which was an 8-week, study, because
of a rash.

When you look at all the studies,
including the open-label safety study, 101 subjects
dropped out because of an adverse event, of which
26 percent were noted to have a rash in their
vignettes although it may not have been coded as a
reason for discontinuation. In one-half of these
subjects, or 13 subjects, the rash was coded as a
primary reason for discontinuation. The rashes
varied in spectrum of severity. Eight with rash
also had fever; 2 with rash had elevated liver
function tests, one with a transaminase of 17 times
the upper limit of normal.


I am now going to discuss some of the

serious skin rashes, primarily the erythema
multiforme, Stevens-Johnson which, from my standing
as a pediatric neurologist are usually
hypersensitivity reactions to drugs. There were 2
rashes which were thought to be erythema
multiforme, Stevens-Johnson.

One subject had peeling and blistering
over the entire body, with lips and urinary tract
involvement, in study 311. The drug was stopped
but the rash progressed to involve peeling,
blistering, mucosal involvement over days. In
another subject in study 207 the drug was stopped
but the rash progressed. The child was

Other rashes of note included a child in
study 207 with vesiculobullous cheeks with severe
lip blisters. In study 312 another subject had a
rash where there is no clear description but the
rash was obviously severe enough that he was
treated with systemic steroids, prednisone and
given Benadryl. The rash recurred when restarted
at 85 mg on day 34. There are two cases of

positive, you know, rechallenge.


Other skin reactions of note--there were
possible allergic events in about 22 of the
subjects out of the total exposure of 933 subjects,
at 2.4 percent. They included hives, urticaria;
facial edema; pruritus; allergic reactions; red
lips; eczema with increased LFTs. There were some
non-allergic events of alopecia, tongue blotches,
Herpes zoster, plantar warts and ringworm.


Now I would like to give some more details
about the index cases here. Case number one was a
young girl with an unremarkable medical history who
had attention deficit disorder. She was started
and then titrated over 2 weeks to a target dose of
either 340 mg or 425 mg a day, but it differs in 2
different vignettes. Two days later, on day 16,
the child developed a fever of almost 102, sore
throat, mild rash which was described as red bumps.
The next day the child was seen in the emergency
room. My understanding is they thought the child

had strep. throat and they gave one dose of
amoxicillin which was subsequently stopped. The
next day, day 18, the modafinil was stopped. Over
the next 4 days the rash worsened and progressed.
There were multiple pruritic areas over the arms
and stomach. On day 22 the rash progressed to
involve the face. On day 23 mucosal involvement
was said to be present in 2 areas. It burned when
the child urinated so there was involvement of the
urethra. The child had swollen and crusty lips.
At some time later--the exact course is uncertain
from the vignettes--there was extensive skin
peeling involving the palms and soles. No new
lesions were said to be present by day 30 and the
event was said to be resolved. By day 31 or day
26--it differs in 2 vignettes--the child was given
1 more dose of modafinil by the mother for unclear
reasons and the itching returned. On day 44 the
child was withdrawn from the study and the vignette
indicates the Stevens-Johnson syndrome resolved but
the erythema multiforme continued.


This case involves a young child with
inattentive deficit disorder who also had Turner's
syndrome and bed-wetting, who was on somatotropin
for the Turner syndrome for 7.5 years prior and
desmopressin for the bed-wetting for 4 months
prior. She was started, titrated on modafinil 200
mg a day for week 1 of the study and then 100 mg a
day for week 2 of the study. By day 4 she
developed fever, abdominal pain and diarrhea. This
lasted for 9 days. By day 14 the child was seen in
the emergency room for pruritic urticaria involving
the face and chest. The drug was stopped. The
child was treated with diphenhydramine. The rash
worsened by day 15. The child was then
hospitalized with a provisional diagnosis of
Stevens-Johnson. The child was seen by a
dermatologist who found no evidence of mucosal
involvement but was diagnosed as a moderate
morbilliform rash. The child was treated with
hydroxyzine. This rash resolved in 1 week. This
case was accepted by Cephalon as being compatible
with Stevens-Johnson syndrome.


Another subject of note is a young boy who
was started on modafinil at 400 mg a day for 2.5
weeks, and on day 14 developed fever and a moderate
rash on the cheeks. The rash progressed. By day
17 there was severe blistering on the lips. The
rash was described as vesiculobullous. On day 19
the modafinil was stopped. The time course of
everything else was not specified in the vignette
and no more information is available. The child
was treated with cephalexin for the rash and
Tylenol with codeine for fever and pain.


Dr. Porres, of the Division of Dermatology
at FDA, reviewed the 21 cases identified in my
initial review and the entire safety database of
this submission. He divided the cases into three
categories, definite cases representing erythema
multiforme, Stevens-Johnson. There are 2 subjects
there or 0.2 percent; subjects who had a history
consistent with early prodromal erythema multiforme
and Stevens-Johnson, there were 3 subjects, 0.32

percent; and then there were 7 additional subjects
who had a history suggestive of prodromal erythema
multiforme, Stevens-Johnson. So, 10 more subjects
plus the 2 subjects, or 12 subjects, so this is a
total of 1.25 percent of subjects with definite and
potential erythema multiforme, Stevens-Johnson.


When one looks at the postmarketing
experience with modafinil, there were 6 reports of
serious skin reactions. All occurred in adults 18
and over. There were 5 biopsy confirmed cases of
erythema multiforme, Stevens-Johnson. Four were
hospitalized and 1 died, but this case was really
confounded by other medications and medical
conditions. There was 1 dermatitis bullous.
Because of the under-reporting, the true number of
cases is probably likely to be greater. But the
take-home message that I would like to say is that
this slide shows that biopsy confirmed
Stevens-Johnson syndrome occurred in adults at
lower exposures than those received by children.


Erythema multiforme or Stevens-Johnson
syndrome is generally thought to be
hypersensitivity reactions to drugs.


One of the cases which was really
interesting involved a child who developed
urticaria, facial edema, fever and a 17-fold
elevation in transaminase between 10-14 days after
starting the drug. The child had a history of
allergy to sulfamethoxazole trimethoprim.
Sulfamethoxazole is a sulfonamide and is one of the
drugs known to cause Stevens-Johnson. It is
structurally similar to modafinil sulfone, which
raises the question of a possible cross-sensitivity
to the sulfone metabolite.


What is the potential public health impact
of these findings? Two recent estimates of the
background rate for erythema multiforme,
Stevens-Johnson was 1-2/million/ year. In this
submission there were 2 subjects with erythema
multiforme and 10 other possible cases of a

significant rash. The total range of risk is

anywhere between 0.2 percent to 1.3 percent.


A recent CDC study estimated that 2.5
million children, ages 4-17, were on ADHD
medication. Now, if we assume that only 10 percent
of these children will try modafinil at some point,
then we ask the next question, how many cases would


We estimated that there would be a range
between 500 and 3,000 cases which will occur based
on the 0.2 percent to the 1.3 percent incidence
among the 10 percent who are switched to modafinil.
Based on the known mortality associated with
erythema multiforme, Stevens-Johnson, we would
expect from 15 to over 400 deaths to occur. We
conclude that even though a crude estimate can only
be made at this time, a potential exists for a
significant number of cases to occur post-approval
since ADHD is so prevalent.


The question is can one label for this?
Can we prevent this? Dr. Le Grenade and her
co-authors at FDA recently published a paper on
Stevens-Johnson syndrome and toxic epidermal
necrolysis in association with selective COX-2
inhibitors. I quote from her and I italicized
certain areas: There is no satisfactory method for
determining who is at greatest risk for developing
drug-associated Stevens-Johnson syndrome and toxic
epidermal necrolysis and hence of preventing it,
short of avoiding drugs altogether. There has been
a single study suggesting that early withdrawal of
the agent at the first sign of illness may improve
the outcome. Although this intuitively makes
sense, this study needs to be replicated. Even if
it is proven correct, its practical application
will be limited because it is very difficult to
identify the very earliest lesion in a timely
manner because of the rapidly progressive nature of
this illness and the non-specific features of its

In the cases observed with modafinil in

this submission in children no deaths occurred.
The rash progressed after the drug was stopped and
the children recovered. It may not be so next


ADHD is a serious condition--I will give
you closing comments--it is a serious condition
which is usually not considered to be associated
with a fatal outcome. Exposure to a sulfone
metabolite is significantly greater, up to 16 times
more in children than in adults. This raises
questions about the relevancy of the adult safety
experiences to pediatric use.


The relationship of this metabolite to
rash is purely speculative but it has structural
similarities to drugs known to cause erythema
multiforme and Stevens-Johnson syndrome which can
be fatal.

The incidence of erythema multiforme,
Stevens-Johnson syndrome observed in these studies
is, at a minimum, hundreds of times the background.

The age ranges of the rashes appear skewed towards
subjects less than 12 years, those having a higher
sulfone exposure. Doses lower than 340 mg have
been shown to limit efficacy, hence, dose reduction
is not a reliable option.


Although some cases with rash got better,
there were 2 positive rechallenges and one case
progressed after discontinuing the drug. One
subject with rash was hospitalized but there was
disagreement about the diagnosis. One child with a
history of reactions to sulfa drugs developed a
hypersensitivity reaction with transaminase
elevation 17 times the upper limit of normal, with
urticaria, fever and facial edema 10 days after
starting modafinil, which raises the hypothesis of
cross-sensitivity with sulfa drugs.


Psychosis and suicidality, although not
standardly significant, were more frequent in
subjects on modafinil than with placebo. Insomnia
was present in 27 percent of subjects on modafinil

versus 4 percent in placebo, and anorexia occurred
in 16 percent of subjects on modafinil versus 3
percent on placebo in the double-blind, Phase 3


This review was very much a team effort of
my many colleagues at FDA, some of whom I am
blessed to call my friends. Thank you.

DR. GOODMAN: Before you step down, could
you review any cardiovascular effects, effects on
heart rate and blood pressure?

DR. MANNHEIM: Dr. Andreason is going to
do that.

Modafinil for the Treatment of ADHD

DR. ANDREASON: Good morning.


My name is Paul Andreason and I am the
Acting Deputy Director of the Division of
Psychiatry Products. I would like to talk to you
this morning about modafinil in the treatment of


Dr. Mannheim has outlined the concerns
that he has about modafinil in the treatment of
ADHD, and I think what we are faced with as we look
especially at the skin rashes is what I like to
call incongruity of data. I will get into that in
a little bit. I would also like to acknowledge the
Neurology Products Division where the drug
resides--it is kind of its home since it was
approved there first--and the safety team for
helping us out with the background rates for
Stevens-Johnson and looking at the adverse event
reports through the Adverse Event Reporting System
and their epidemiologic expertise.

Glenn did the primary review on the first
submission. June Cai helped out with the review of
the response to the approval letter. In the
Division of Dermatology, I would like to thank Joe
Porres and Markham Luke, who is here with us
today--Markham is here with us today. Joe took
another job and he is not with the Division of
Dermatology right now. Then, in the Division of
Drug Risk Evaluation, I would like to thank Andy

Mosholder and Kate Gelperin who, as part of their
presentation yesterday, did an analysis of the
psychiatric adverse events that are associated with
modafinil use.


Just as a quick review of how we workup
safety problems with drugs or safety profiles of
drugs, I should say, when we look at a drug we look
at deaths, serious adverse events, adverse
dropouts, potentially clinically significant labs,
ECG and vital signs and then we develop information
on comparative common and drug-related adverse
events, all these things from controlled trials.
We also do special searches, especially in this
case with modafinil and many psychiatric
drugs--well, all psychiatric drugs, for psychiatric
adverse events; in this case, for Stevens-Johnson
syndrome and neutropenia because these were things
that kind of popped up; and then the recent
interest in blood pressure, pulse and
cardiovascular adverse events. Then with the
response to the approvable letter we get a safety

update. In that safety update we focus on serious
adverse events and deaths, if they occur. We
develop our profile of the common and drug-related
adverse events from the controlled trial data, as
well as the comparative information on labs, ECGs
and vital signs.


Modafinil is a marketed product and we got
some information from Verispan about the exposure
to modafinil at this point. These are numbers not
in patient-years but in unique patients. At all
ages at this point between the years 2002 to 2005
there were 1,087,000, roughly 1,088,000 exposures
in all ages, and for children ages 0-17 there were
roughly 36,000 exposures.

I kind of want you to keep that in mind
because this is the first piece of what I would
call inconsistent data about rash. It is almost
unheard of to see cases of Stevens-Johnson syndrome
in controlled trial data and here we have at least
nominally 2 cases that have been identified as
such. At that kind of a rate you would expect to

see something in the adverse event reporting data.
Dr. Mannheim said, well, based on these numbers,
these would be the projected number of cases that
we would see after marketing. The piece of
incongruity here is that the drug is already
marketed. We have 36,000 exposures in the age
ranges that were studied, and in the 0-12 group
right around 11,000 and we have no AERS-reported

Now, one of the cases from the controlled
trial data actually is a duplicate case. It got
reported in AERS but there are no spontaneous
reported cases. So, given that kind of projection,
I would expect to see some cases reported in AERS
and we haven't seen that yet.


This is, again, a review of patient
exposure in the controlled trial database. In the
safety update we did get some information on
serious adverse events and dropouts, as well as
deaths, and Stevens-Johnson syndrome would be
considered in that group. So, as more and more

information comes in, you know, that denominator of
cases reported per amount of exposure changes,
however, even with 2 cases in 1,600 that would
still be a large number.


I think the problem that comes about when
we look at Stevens-Johnson--and we will hear more
from Dr. Bigby in a moment about how that
ascertainment is made--is that in these two cases
one was hospitalized and one was not. Neither of
them were in a burn unit or the ICU and we don't
have biopsy information on those kids.


These are tables that you have already
seen. It reviews the numbers of patients exposed
in the three pivotal trials.


Now, this is a table that shows you the
common adverse event profile. Our usual definition
of common and drug-related is adverse events that
occur at least 5 percent of the time and occur at a
rate that is twice placebo. In italics you will

see that anorexia and insomnia meet that criteria.
There are a couple of other adverse events that are
close but don't quite make that cutoff. This is
the table actually that is proposed in labeling and
is the usual kind of table that we have in


Just as a quick overview of the safety
results from the controlled trials, there were no
deaths and, of the adverse events of note, there
were these 2 cases that were identified as either
Stevens-Johnson or erythema multiforme. There were
no new cases of leukopenia in the AERS system
update, and we could see no real signal for
leukopenia in the controlled trial data.

As far as psychiatric events, there were 4
suicide-related adverse events, no completed
suicides. I will talk more about those in a
moment. There were none in the placebo group.

As far as mean blood pressure changes,
modafinil actually showed a slight decrease
compared to placebo in mean blood pressure.

However, the numbers of patients that met the
outlier criteria of systolic blood pressure of
greater than 130 and an increase in greater than 20
mmHg were 9/420 for modafinil and 1/213. With
pulse there was no difference in the mean value
either, and the numbers for outliers are 6/420
versus 2/213. The 6 versus 2 in those 2 groups is,
in my opinion, not terribly different. There was
some weight loss, 0.7 kg weight loss with modafinil
versus 1.0 kg mean weight gain in the placebo

So, did that answer your question about
blood pressure?




As far as psychiatric adverse events, this
is drawn from one of Andy Mosholder's slides from
yesterday. These were the comparative numbers with
patient-year exposure, and these are real years.
They are not multiplied. So, with 33 patient-years
exposure in placebo you have no cases of mania or

psychosis or suicide-related adverse events. But
there were 5 cases of aggression, spontaneously
reported aggression. Zero cells are kind of tough
to deal with when you are doing statistical
analysis, but oftentimes you can use a Fisher's
exact test to get at least some idea of whether or
not something is statistically significant. I will
show you that for the suicide-related adverse
events on the next slide.

You will notice that numerically the cases
of aggression are slightly less with modafinil than
they are with placebo. As Dr. Mosholder stated
yesterday, that was not a significant difference
but it is not, by the same token, greater. In the
open-label data it shows that the rates are lower.
That doesn't necessarily mean--well, let me put it
this way, these are patients, once they reach
open-label, who have tolerated the drug and I think
that probably the best comparison for this is the
controlled trial data.


These are the results of the Fisher's

exact test for suicide-related adverse events. You
see here that you have the 4 cases in the modafinil
and that is compared to no suicide-related adverse
events in the 660 in placebo. So, that ends up
with a 2-tail value of p of 0.31 and 1-tailed p of

Just to give you kind of a comparison

with, say, Strattera that has received labeling for
suicide-related adverse events, with Strattera
there were 6/1357 versus 0/851. Because of the
increased sample size, those numbers ended up being
statistically significant. There were 5 cases of
ideation and 1 attempt in the FDA defined cases. I
would like to note that Eli Lilly has slightly
different numbers because they had a slightly
different definition of the suicide-related adverse
events. They had 7 versus 1 out of 1357 and 851
respectively. That p value ended up being 0.07.
Traditionally, for safety-related topics we don't
necessarily use a p of 0.05 like we do for
efficacy. We use a p of 0.1. So, using a cutoff

of a p of 0.1, the 0.7 would still be statistically
significant. And, Strattera has a boxed warning.

Now, with the modafinil there are 4/664
versus 0/308. This is not statistically
significant by Fisher's exact and all were cases of
ideation and 3 of the cases actually resolved
without discontinuation of the drug. The sponsor
proposes warning language in labeling as opposed to
boxed language.


As far as the cases of severe rash that
are identified as Stevens-Johnson syndrome, we will
hear more about that, again, from Dr. Bigby, and
Dr. Markham Luke is here today to talk about the
cases individually if people have questions on

The problem that we have with almost any
adverse event report ascertainment, there was no
histopathology with either of these cases. With
Stevens-Johnson syndrome admission to burn units
and ICUs is common. One of the kids was
hospitalized but not in an ICU or burn unit. The

other child was treated as an outpatient. You have
heard about the cases. I can go back to those if
anyone has any questions. Again, there were no
children in the postmarketing Adverse Event
Reporting System, other than the one case that is
the duplicate from the controlled trial.

There were 4 adults in the AERS
postmarketing database, and it turned out that 3
had confirmatory histopathology and the other one
was erythema multiforme without histopathology.
There were no adults with Stevens-Johnson
identified in the adult controlled trial database.


So, what we are left with from this
controlled trial database, along with the
open-label material that goes with it, is 2 serious
cases, one admitted to the hospital, neither to the
ICU or burn unit; none in the placebo group; 10
dropouts due to rash versus no dropouts due to rash
in the placebo group. Spontaneous adverse events
in the controlled trial, about 4 percent for
modafinil versus 1 percent for rash, for all kinds

of rash.

But then, the incongruity here is that
there are no other children with either
Stevens-Johnson syndrome in the postmarketing
adverse event database with about 36,000 kids
exposed. Again, with that kind of exposure and the
projected numbers of cases of SJS, based on 2/933
or even 2/1,600, one would expect to see more in
the Adverse Event Reporting System.


So, just to compare and contrast labeling
with Lamictal that carries a boxed warning for
Stevens-Johnson syndrome, in that boxed warning
there are some fairly hard numbers. For example,
they did a prospective registry study and there was
one death due to Stevens-Johnson syndrome with
Lamictal out of 1983 patients. There was also
information on rates in adults with different kinds
of diagnoses, for example, 8/1000 in children with
Lennox Gasteau and 3/1000 adults. Then in the
bipolar population it was 1.3/1000 adults on
adjunctive therapy for bipolar disorder. So, those

are some fairly hard numbers.

On the modafinil side, the sponsor
proposes warning language and I have listed under
here the points to compare and contrast with
Lamictal. There are no deaths reported. Actually,
on one of the slides that Dr. Mannheim presented,
he said there was one death. That case was a
fellow who came in to the hospital and had a
subarachnoid hemorrhage and was treated with
several drugs, one of which was phenytoin, known to
be associated with Stevens-Johnson syndrome. He
developed Stevens-Johnson syndrome as part of the
course of his hospitalization and apparently was
treated with modafinil prior to that
hospitalization. So, I think that case is terribly
confounded and I wouldn't count that as drug
related, or I don't think we could count that one
as a good drug-related case. So, based on my
exclusion of that case, there would be no deaths so
far due to Stevens-Johnson.

The child cases were not severe enough to
require burn unit or ICU. Now, again back to this,

there are at least 2 nominally identified cases out
of the 933 in the submission that were identified
but there is no biopsy confirmation. Back to the
other part of the incongruity, there are 36,000
exposures already with no cases. Then back to the
other side, you have 3 confirmed cases of adult SJS
in the postmarketing but that is with 1.5 million.
So, that is getting close to the background.
Depending on how you factor in under-reporting, you
know, there could be association and increased risk
for Stevens-Johnson.


So, I guess in the end the question that
you need to think about is if there is an increased
risk of Stevens-Johnson syndrome associated with
modafinil, what would be an acceptable risk. And,
if modafinil were considered for approval, what
kind of risk management program would you want to
implement, and how should the concern about serious
skin rashes be addressed in product labeling.
Again, you have the examples of labeling and I can
go over those a little more if you would like.

And, should there be a requirement for
postmarketing studies, if approved, to better
understand the skin rashes?

There was one slide, as a bit of a
digression that Dr. Mannheim showed about liver
enzymes, he included GGT in a slide, along with ALT
and AST as a percentage of increased liver enzymes
under the heading of hepatocellular injury.
Usually we look at ALT, AST and bilirubin as signs
of hepatocellular injury and don't necessarily
include GGT. Excluding GGT, there were 3/420 cases
of elevation of ALT and AST of greater than 3 times
the upper limit of normal, for a percentage of 0.7.
In placebo there was 1/213, for a percentage of
0.5, and I don't see that as meaningfully

So, that concludes my remarks about
modafinil and I would be happy to entertain any

DR. GOODMAN: Thank you. I understand
there is going to be a change of technology before
the next presentation. Is that correct? Yes? So,

why don't we start doing that but I wonder if you
can stay for questions that this committee may
have. Let me start that off.

Of the 35,000 or so children who have been
exposed to modafinil postmarketing, how many of
them were on the doses that are proposed to be used
in ADHD? I would suppose further that that would
be mostly for Mexico where it is already approved.

DR. ANDREASON: I am not sure. By the
way, those numbers are for the United States only.

DR. GOODMAN: So, if they are for the
United States only let me go back and rephrase it.
How many of them would be in the dose range that is
proposed for ADHD?

DR. ANDREASON: Yes, we tried to track
that down and found that it was impossible to get
that kind of information. I think the only thing
you could do would be to assume that they had
received the maximum recommended dose, which was
only 200.

DR. GOODMAN: We have one of our committee
members that I would like to see introduce herself,

Jean Bronstein.

MS. BRONSTEIN: Good morning. I am Jean
Bronstein. I am a nurse and consumer
representative for the committee, and I apologize
for being a little late. But I do have a question.
I am trying to understand some 300 patients, I
believe, that have dropped out of the study, and I
am wondering if the rash numbers are also
representative of all patients having dropped out
of the study at some point. Is that clear?

DR. ANDREASON: The patients who dropped
out due to rash are included in those numbers, yes.

MS. BRONSTEIN: They are? Thank you.

DR. GOODMAN: Any other questions around
the table or are we ready to proceed with the

DR. WANG: It is actually a follow-up to
Wayne's question. In the negotiations for these
pivotal trials how were the doses chosen? Why was,
you know, 300 mg, 400 mg chosen? For the
wakefulness indication it is 200 with no additional

DR. LAUGHREN: Paul, you may be able to
say more about this, but my understanding is that
there was a Phase 2 study, a fixed dose study that
looked at different doses, I think running from
100-400, that basically showed effects only at the
higher doses and that was the basis for focusing on
the higher doses in the pivotal trials.

DR. ANDREASON: I am sorry, I thought you
were looking at Cephalon when you asked that
question. But, yes, that is correct. Two of the
studies were flexible dose studies in the pivotal
trials and one of them was a fixed dose study.

DR. WANG: There is no data in here to
suggest something that you have raised, Dr.
Mannheim, that there is no potential benefit to
trying lower doses. They don't look to be, you
know, sort of clinically efficacious, which these
data don't suggest.

DR. MANNHEIM: My understanding from the
earlier Phase 2 trials was that lower doses didn't
work and they had to get to these doses in order to
show efficacy in ADHD. Cephalon can respond to


DR. GOODMAN: Yes, maybe we can hear more
about that from the sponsor during their
presentation. We have Dr. Leon and then Dr. Mehta.

DR. LEON: I would like to follow-up on
the number of exposed. Is that based strictly on
the 7- or 9-week clinical trials out of which maybe
about 40 percent dropped out? Or, does that
include the follow-up as well? Is there slide on
person-week exposure?

DR. ANDREASON: Oh, as far as the number
of patients that are exposed for an adequate amount
of time to count, I think what we came up with was
as an estimate that was around 600. Again, if you
say 2 cases out of 600, that makes the rate of it
seem even higher. Then, it even makes it seem more
implausible that we don't see anything in
postmarketing. So, I think you are right, if
Stevens-Johnson is something that shows up in the
first 2-8 weeks of treatment the numbers in the
controlled trials would be right around 600. The
900 includes patients that dropped out. It

includes patients that were in Phase 2
studies--excuse me, I take that back. The 933
includes patients that were in the open-label
trials, so patients that were exposed for much

When we looked at postmarketing data and
estimates of exposure we didn't look at
patient-year exposure because we wanted to focus on
the fact that Stevens-Johnson we probably likely to
show up in the first 2-8 weeks and if we looked at
patient-year exposure the rate of background would
start to drop with extended exposure if you looked
at patient years instead of unique patients. That
is why we chose to look at it that way.

DR. GOODMAN: Dr. Mehta?

DR. MEHTA: Is the dose of the drug
relevant to the occurrence of Stevens-Johnson
syndrome? I would have thought that this is a
sensitivity reaction so it doesn't matter what came
out of the drug that is used, that is, the dosage
is irrelevant for the occurrence of SJS.

DR. ANDREASON: You will see from the

presentation coming up from the company that it
didn't appear that the sulfone metabolite was
increased any higher, than in other patients that
didn't have rash, in the 2 patients that were
identified as having rash. The expert is here to
talk about that.

DR. BIGBY: Can I just make a quick
comment about that?

DR. GOODMAN: Go ahead.

DR. BIGBY: It shouldn't make any
difference in terms of the incidence of the rash.
I think you are correct in that regard. The only
way that I think it could affect the disorder is in
outcome in terms of at a higher dose it will take a
little bit longer to be cleared from the body so
that the prognosis might be worse if you start with
a higher dose because it may take longer to be

DR. GOODMAN: Dr. Malone?

DR. MALONE: I just had a question about
dosing. The stimulants are also used for daytime
sleepiness disorders. Is the dose used for ADHD a

lot different than the dose used for daytime

DR. ANDREASON: Yes, it is higher.
Daytime sleepiness and obstructive sleep apnea
doses are 200. Right, Glenn? Then, for ADHD it is
300 and 425.

DR. RAPPLEY: No, I think he is asking
about comparing the other stimulants? So, are the
doses for Adderall or Ritalin higher for ADHD than
they are--

DR. ANDREASON: I am sorry, I don't have
that right at the top of my head.

DR. GOODMAN: Dr. Pfeffer?

DR. PFEFFER: I have a question about the
pharmacokinetics of this drug, with the dose being
so much higher in children, especially younger
children, how is the drug metabolized? Also, is
there a way of understanding if there are children
who are slower metabolizers of this drug and,
therefore, this is higher? Can we understand that?
Is there a way of understanding also if there are
certain children that might be assessed in terms of

the metabolism and understand that relative risk?

DR. ANDREASON: I can't really answer that
question. I think we have somebody who can answer
that question. Let me preface it by this though, I
don't necessarily believe that ruling out the
sulfone would be necessarily a guarantee of safety.
In my opinion, I think the sulfone may be a bit of
a red herring. I think with Lamictal we have no
real idea why SJS occurs and it is much more common
in kids than it is in adults. If they could come
up with some kind of marker that would show what
the risk was, that would be wonderful. I don't
think in this case the sulfone reaches that kind of
state. First of all, we have to identify whether
or not there is a signal. Then, if one came to the
conclusion that there was a signal, I still don't
think that the sulfone would give us any assurance
of safety regardless at this point.

DR. GOODMAN: Dr. Temple and then I want
to go to Dr. Bigby's presentation.

DR. TEMPLE: I just wanted to observe that
on the dose relatedness matter it is sort of

unusual to have almost an order of magnitude
difference between what one group gets and what the
other group gets. So, a lot of our experience with
drugs will be looking at, you know, one- or
two-fold differences and things like that. I am
not sure one really could say that a marked
difference in blood levels or exposure might not be
related to rate. It could.

DR. ANDREASON: Ron Cavanaugh was the
human biopharmacologist on this.

DR. GOODMAN: But then definitely Dr.

DR. CAVANAUGH: Thank you. I agree with
Dr. Andreason that at this point for the sulfone we
really don't know any relationship for certain. It
is purely at this point a plausible hypothesis. It
is very structurally similar to the one drug which
causes the highest incidence of Stevens-Johnson,
blethamide, which is slightly different than many
of the other sulfonamides in that it has a third
oxygen in addition to the nitrogen and the sulfone.
Modafinil also has that third oxygen in the same

position. We do not know. Basically, the only
reason we looked at the sulfone was because of the
dramatic higher amounts, as well as when Glenn
asked me about the sulfone I immediately thought
Stevens-Johnson and immediately thought
sulfonamide. So, at this point we don't know.

In terms of the kinetics, from what I have
seen, the metabolism does not seem to be
particularly well defined. So, I really do not
know at this point, you know, anything in terms of
could specific metabolic pathways result in higher
sulfone concentrations in some kids versus others.
The concentrations that I see, and it is very poor,
do not lead me to believe that the sulfone
concentrations in these particular children--and
there was only one child who had any measurement of
the sulfone who had any of the Stevens-Johnson or
other severe dermatologic reactions, and that
sample was taken several days after the drug was
discontinued. If I back calculate, it basically
seems that for that one child the concentration was
in the approximate range.

In terms of dose response, which was
raised, we really have too few numbers here. You
are also dealing with, you know, population numbers
and you are dealing with 0.1 percent difference,

0.2 percent differences. There is no way you can
be certain.

In terms of a dose related to hapten and
degree of what is the likelihood of Stevens-Johnson
or a hypersensitivity of any sort, I am not an
immunologist; I do not know. One of the reasons we
focused on AUC is because that gives you an idea of
total exposure. When someone has already developed
hypersensitivity and they have already had a
history with a sulfonamide, they can get
Stevens-Johnson with the very first dose of a drop,
and there have been deaths in cases like that.

Really we are more talking about the
development of hypersensitivity, and what you are
developing as it being a hapten, my understanding
is that it is the combination of, you know,
developing hypersensitivity to the combination of
the drug bound to certain proteins, or other things

in the body. So, with the higher exposure you are
going to get more of this binding and, therefore,
more antigenic exposure. The numbers are so small,
we don't know. Also, with longer duration you
would expect more stimulation.

I would really refer you to an
immunologist. I really don't know, but the whole
issue of dose response and everything else in terms
of developing hypersensitivity, to me, is not

DR. GOODMAN: Thank you. Thank you for
being patient, Dr. Bigby. We are running behind
schedule but the way I hope we may be able to make
up some time is that I am going to cut lunch and I
don't think we have that many public speakers. But
I am determined to end at least at our scheduled
time. During the sponsor presentations I would ask
the committee members to restrict their questions
to ones of clarification.

Serious Adverse Cutaneous Reactions to Drugs

DR. BIGBY: Good morning.


I am always impressed when I come down to
sort of work in FDA committee meetings about the
seriousness of what goes on here, and also the
dedication that people have to trying to make
rational decisions, and I hope my comments are
helpful in your deliberations.


What I was going to talk about is serious
adverse cutaneous reactions to drugs, and in order
to do so I am going to cover three things. One is
how to identify a drug eruption as a drug eruption
and pin it down to a specific drug. We will look
specifically at common eruptions, the serious
eruptions, and I will end by showing you some
things that are commonly mistaken for drug

This is sort of a gold standard for
determining that a rash is due to a drug. First
you have to be sure that the rash you are looking
at is a morphology that can be caused by drugs.
You have to exclude alternative causes. You have
to examine the relationship between the exposure to

the development of a rash in terms of the time
interval; note the response to drug withdrawal,
i.e., the rash will go away.

For many drugs there is information known
about their proclivity to produce rashes, so what
the frequency of rash is for a particular drugs.
Then, in those rare cases where you actually do
have a re-exposure, to determine what happens on
re-exposure so you can be positive that it is a
drug rash and looks like an eruption that is a
classic eruption for drugs. You have excluded
alternative causes; the interval from exposure to
the development of a rash is correct in terms of
what is known about that drug and that eruption.
It goes away. Often I think the mistake people
make about the response to withdrawal is that you
expect it to go away very quickly when you withdraw
the drug. For most eruptions that is not the case
and the rash will actually take much, much longer
than most people think to go away after you
withdraw the drug. Then, oddly enough, re-exposure
doesn't always produce the rash but when it does,

then you can be absolutely sure that you are
looking at a definite cause. Where some of these
things are missing, are judged to be probable or
unlikely or not due to the drug at all.


Very quickly, we are going to look at
these three common drug eruptions: exanthem which
is the most common, urticaria and fixed-drug
eruption. This is a patient with a widespread
exanthematous drug eruption. It usually starts
within the first 3 days after exposure to the drug.
For some drugs like antibiotics and allopurinol
that exposure window can be up to about 2 weeks.
The rash is best described as small, erythematous
papules that may coalesce. These patients have
pruritus but they are not generally ill. Mucous
membrane involvement is rare. It is a benign
condition in that as the drug is withdrawn the
patient gets better. They can often later in the
course of the disease desquamate but they don't
develop blisters and they don't have epidermal


There is very good data about this type of
reaction. It has been studied in several
prospectively collected data sets. One was the
Boston Collaborative Drug Program. There is data
on I think something like 35,000 exposures over
about, you know, a 10-year period, collected in
many hospitals.


You can say with fair certainty that there
are certain drugs that have higher rates than
others, and in this list the highest tends to be
antibiotics. The highest rates are for amoxicillin
and co-trimoxazole.


It is also helpful to know that there is a
large list of drugs that are almost never
associated with reactions. So, if a patient is on
multiple drugs, which they often are, it is useful
to refer to this type of list to exclude the ones
that are least likely to be the culprit.


Urticaria you all know how to recognize.
It is areas of swelling. There are usually plaque
type lesions, and the key about urticaria is any
individual lesion generally will last for less than
24 hours. Here is one of the perfect examples. If
you identify the cause and you withdraw it patients
will often have urticaria after such an exposure
for weeks and even months even though you have
identified the correct drug and withdrawn it.


The list of drugs that cause urticaria is
very similar to the ones that produce exanthem.


Lastly, this is a fixed-drug eruption. A
fixed-drug eruption is a really peculiar thing in
that it tends to occur only on certain areas and to
recur in those areas on re-exposure. It is the one
instance where people will often be re-exposed
because it is not so clear to the providers that
this was, in fact, a drug eruption. The other
reason that this is quite relevant is that the
histopathology of a fixed-drug eruption is very

similar to what you see in erythema multiforme and,
to a lesser degree, in Stevens-Johnson syndrome and
toxic epidermal necrolysis.


Again, if you look at the drugs that cause
fixed-drug eruptions, there is a lot of overlap
between the drugs that most commonly cause all of
these types of eruptions.


The three serious drug reactions that I
want to talk about are the ones that I think are
the most relevant to this question that you are
asking today, and that is toxic epidermal
necrolysis, Stevens-Johnson syndrome and the drug
hypersensitivity syndrome.


Of all of the things which I have to say
today, this is the slide that I want you to
remember the most. These are two patients that I
saw personally. These are people with toxic
epidermal necrolysis. The most obvious and
important thing about these patients is, number

one, that they are sick. They often have multiple
mucous membranes involved. In severe cases not
only are the sort of distal mucous membranes
involved, but it can affect the trachea and even
the bronchi.

The second most important thing is that
they have widespread areas of cutaneous involvement
and, in the case on the right, they often shed
full-thickness necrotic skin over very large areas,
and they have basically the equivalent of a
widespread burn. The summary of toxic epidermal
necrolysis in terms of its clinical features is
also a prodrome of fever and malaise. This usually
lasts one to two days. The eruption is
predominantly on the face and torso. The lesions
are best described as pruritic plaques. They can
have bullae. Multiple mucosa are commonly
involved. Patients with toxic epidermal
necrolysis, however, do not have true target
lesions. Probably by the best definition of toxic
epidermal necrolysis, it has to involve at least 30
percent of the body surface area, and the mortality

for such described toxic epidermal necrolysis is
quite high, around 30 percent. The majority of
deaths are due to either infection or problems with
respiratory distress that are either due to
pneumonia or to the fact that the airway linings
are involved.


As has already been mentioned, it is a
relatively rare phenomenon so that in most
population studies the incidence is about one case
per million patient years; 95 percent of the cases
clearly have a drug etiology, and there are certain
drugs for which the incidence is much higher.


Based on a case-controlled study that was
published in the New England Journal ten years ago,
this was a study that carefully ascertained cases
in France, Germany and Italy and to drug exposure
histories from patients in three age and gender
matched controls, and came up with an estimate of
the number of cases per million exposures that one
would see per week. It was highest for

sulfonamides. If you do the arithmetic, this turns
out to be something in the order of 1 case in

200.00 or 250,000 for some of these drugs.
The drugs commonly associated with TEN are

listed here. Again, these lists are very similar
to the ones that cause benign eruptions and the
same sort of drugs keep showing up: sulfonamides,
hydantoins, some but not all of the nonsteroidals
and allopurinol.


This is a patient with Stevens-Johnson
syndrome, Stevens-Johnson syndrome and toxic
epidermal necrolysis are dissimilar disorders in a
continuum. The difference between Stevens-Johnson
and TEN is one of degree of epidermal detachment.
The symptoms are very similar. There is prodrome
often of fever and malaise. The lesions are very
similar. In Stevens-Johnson syndrome the area of
involvement is usually defined as being less than
10 percent. It has a much lower mortality.

The other interesting thing is that if you

look at the etiology of Stevens-Johnson syndrome,
it can be attributed to drugs in only about 50
percent of cases. Now, that seems to be in
congress with TEN and SJS being part of a spectrum.
I think the problem is that there is a lot of
confusion about mixing up cases of erythema
multiforme, which I think is a quite separable
disease, with Stevens-Johnson syndrome. I think
that explains why drug etiology is less commonly
identified. I will have more to say about erythema
multiforme at the very end.


Again, the incidence is about one per
million per year, drug induced in about 50 percent.
There is a higher incidence with some drugs and it
is that same list of drugs, you should note.


Now, what I was saying about the
relationship between SJS and TEN, TEN is defined as
those cases where the area of involvement is more
than 30 percent. SJS is less than 10 percent.
Then, there are people who are kind of in the

middle, between 10-30 percent, that are called
SJS/TEN overlap. The other thing to note is that
erythema multiforme is not mentioned anywhere on
this slide or in my previous comments because, as I
said, I think it is a distinct disorder.


The last serious reaction that I wanted to
talk to you about is the hypersensitivity syndrome.
That is what this slide is an example of. It looks
very similar to exanthem except for two things.
When you have seen a few of these patients it
always strikes you that the color in the
hypersensitivity syndrome is a much brighter and
darker red and the amount of confluence of the rash
is much higher.


Symptomatically, these people have
exanthem. They have fever, lymphadenopathy, often
have hepatitis, some of them have arthritis. This
is a disorder that has a significant mortality. It
is not clear how patients should be treated and,
again, the list of drugs that cause this that are

already known and sort of identified as such is
very similar to the list of drugs that cause drug
rashes in general.


This is a slide from a paper that was done
by Roujeau and Stern, in the New England Journal,
and it is a very busy slide. The only thing I want
you to note is that the fatality rate for the
hypersensitivity syndrome is about 10 percent. For
TEN it is about 30 percent. For Stevens-Johnson
syndrome, because of the area of involvement and
severity it is much less; it is lower.


Finally, exposure to rash for TEN and SJS
is in the order of weeks, so 1-3 weeks is noted in
the third column in this slide. Skin biopsies are
very helpful because in TEN and SJS they tend to
show full-thickness epidermal necrosis, detachment
of the skin at the dermal/epidermal junction, and
often there is very, very little inflammation in
the dermis associated with the rash.


Finally, I would like to conclude by
showing you examples of things that are often
called drug eruptions but are not, primarily
erythema multiforme. Now, erythema multiforme--you
can't make that diagnosis unless patients have
typical target lesions. Typical target lesions
have three rings, either a dusky or bullous center,
an area of erythema around that and then a
surrounding area of edemous skin. You can often
actually have rings beyond that but if you have the
three rings it is I think easily identifiable as
erythema multiforme.

In terms of the distribution, another
thing that is helpful is that erythema multiforme
predominantly affects the face and the extremities.
The torso is much less commonly and much less
extensively involved. The majority of cases of
erythema multiforme are actually associated with
infection, herpes simplex being the most common one
and, although it can be caused by drugs, drugs are
a much, much less common etiologic factor for
erythema multiform.

The other thing is that erythema
multiforme, by and large, is a benign disorder.
Patients recover and deaths are extremely rare for
erythema multiforme. I think people shouldn't
combine erythema multiforme, even erythema
multiforme major with mucosal involvement with
Stevens-Johnson because I think they are distinct

Grover's disease is another one that looks
to the world like a drug eruption. It tends to
occur commonly in hospitalized males on their back
and, you know, a lot of the times we get called for
drug eruptions and it turns out to be just this
scenario, elderly men with Grover's disease
predominantly on their back.

Lastly, extensive cases of pityriasis
rosea can be mistaken for drug eruptions. The key
there is that, you know, the history is usually
pretty classic. The distribution is classic as
well and if you have the herald patches, as noted
in the right-hand slide, there is not a lot of

The last two things on this list, the
viral exanthem and graft versus host disease, no
one can really distinguish those from drug
exanthems or several other drug eruptions and it is
a matter of great difficulty.


That is where I will stop. Thank you.

DR. GOODMAN: Thank you. Questions from
the committee? Dr. Pine?

DR. PINE: I guess I am struggling a
little bit, kind of like Dr. Andreason was. On the
one hand, you know, the rashes were very concerning
that were described. On the other hand, there are
no cases in the spontaneous reporting. I was just
wondering, given your background as somebody who
sees this kind of thing every day presumably, or
frequently, what was your impression when you
reviewed the cases in terms of how convincing they
were, number one and, number two, when you combine
that with what you would expect to see how
concerned, as a dermatologist who spends a lot of
time thinking about this, were you about the data

that you saw and the cases that you saw?

DR. BIGBY: I think that the 7 year-old
child that was described, to me, was a probable
case of SJS that was drug related. After looking
at the material, I think that the drug is going to
be, and probably already is, associated with sort
of an excess of cases of SJS/TEN.

DR. PINE: Thanks.

DR. GOODMAN: Dr. Rappley?

DR. RAPPLEY: I looked over 26 cases, I
think it was, that had rash and I noted that many
of those cases presented on a continuum that
included fever, pharyngitis, rash, and it went from
very mild to very severe. That is something in
pediatrics that we see as a reaction with
immunosuppression or reaction that, you know,
reminds of Kawasaki's--not exactly but it makes me
think of that. It reminds me of drug reactions.
It reminds me of neutropenia. So, my question is
do you see that as a continuum, those symptoms as

DR. BIGBY: I am actually not sure I

understand your question. I mean, of the material
that I saw, I think that there was one case that
probably had SJS. I think that the other sort of

rashes described-
DR. RAPPLEY: So, you don't see that as a
DR. No-  
DR. RAPPLEY: You see Stevens-Johnson as a
very discrete-  

DR. BIGBY: Yes, right. You know, I think
that eruptions are sort of specific things to
dermatologists and these things don't sort of fit
together as a gestalt for a kind of reaction to
that drug, no.


DR. GOODMAN: What I would like to do is
take an unscheduled quick break, seven minutes.
Before we do that, just an admonishment, I would
like to remind the committee that, in the spirit of
the Federal Advisory Committee Act and Sunshine
Amendment, discussions about today's topic should
take place in the forum of this meeting only and

not during lunch breaks or in private sessions. We
ask that the press honor the obligations of the
committee as well. If you will allow the committee
members to exit the room first to take their break,
we will reconvene in seven minutes. Thank you.

[Brief recess]

DR. GOODMAN: We are resuming now with a
series of presentations from the sponsors. Please,
committee members, restrict any burning questions
to those of clarification. I think that we will go
to lunch at 12:30 instead of 12:00, which means
that we save time for more detailed questions of
the sponsor after the public presentations. Please
go ahead.

Sponsor Presentation


DR. RACZKOWSKI: Good morning.


Dr. Goodman, members of the advisory
committee, Dr. Laughren, FDA representatives,
ladies and gentlemen, today we will be discussing
the application for Sparlon tablets for approval

for treatment in pediatric patients with ADHD.


My name is Victor Raczkowski and I am the
vice president for worldwide regulatory affairs at


Our proposed indication for Sparlon is for
the treatment of ADHD in children as well as
adolescent patients. We filed our application in
December of 2004 and we received an approvable
letter about ten months later. We submitted a
complete response then to the agency in November of
last year.


Sparlon contains the active ingredient
modafinil which is also contained in Provigil
tablets. So, modafinil is not a new chemical
entity. Sparlon tablets have been formulated to
facilitate administration to pediatric patients.
That is, on a milligram/kilogram basis of modafinil
they are smaller than Provigil tablets. They come
in dosage that ranges in strength from 85-425 mg

and are intended for once daily administration in

the morning.


Provigil has been marketed in the United
States since 1999 and is currently marketed in 28
countries worldwide. Provigil is a wakefulness
promoting product and it is approved in the United
States in adults with excessive sleepiness
associated with narcolepsy, obstructive sleep
apnea/hypopnea syndrome or shift work sleep
disorder. We have estimated exposure globally
through the end of February 2006 as being
approximately 780,000 patient-treatment years, of
which 30,000 patient-treatment years are in
pediatric patients. Modafinil is also listed in
Schedule IV of the Controlled Substances Act.


You have been asked to address a number of
questions today and the two voting questions are on
the efficacy and safety of Sparlon. We hope to be
able to show you with our data today that not only
is Sparlon effective for the treatment of ADHD, but

it has also been shown to be acceptably safe for

the treatment of ADHD in pediatric patients.


We have also been asked to address the
dermatological safety of Sparlon tablets. Just by
way of orientation, this slide represents cases of
SJS in pediatric patients in clinical trials as
well as in our postmarketing experience. Other
speakers are going to go into this in much more
detail. This is just to orient you that an earlier
review by our experts indicted that there was one
case of probable SJS in our clinical trial program
out of 1622 patients exposed. However, that case
is of uncertain etiology. In addition, if the
committee has questions or interest in the clinical
course of that patient, we do have the investigator
here at the meeting today who can describe the
clinical course of that patient.

In our pediatric postmarketing experience
we have seen no pediatric cases of SJS in over
30,000 pediatric patient-treatment years.


We will have a number of presentations
today, beginning with an overview of attention
deficit hyperactivity disorder by Dr. Joseph
Biederman. That will be followed by a review of
both the clinical pharmacology and efficacy by Dr.
Lesley Russell. Dr. Srdjan Stankovic will then
provide an overview of the safety and then Dr.
Russell will conclude with an overall benefit-risk


We have a number of consultant experts in
the field with us today representing various
disciplines including psychiatry/ADHD, dermatology,
addiction medicine, cardiology, child development
as well as epidemiology. I would just like to
highlight one name since dermatology is a major
issue in today's presentation, we do have an
individual, Dr. Neil Shear, with us today who has
published extensively in peer reviewed journals on
SJS as well as other dermatological disorders. I
would also like to note, as you can see on this
slide, that we have a number of investigators with

us today.


With that, I would now like to introduce
Dr. Joseph Biederman, who is professor of
psychiatry at Harvard Medical School. Dr.

Overview of ADHD

DR. BIEDERMAN: It is a pleasure to be
here. I would like to offer you a very brief
overview of ADHD as a very serious illness of
genetic etiology affecting the brain that has a bad
prognosis. I strongly believe that without
understanding the assessment of benefit-risk
alternative treatment is impossible.


First of all, I think that it is important
to note that ADHD is a highly heterogeneous illness
like all psychiatric illnesses. We know that genes
are important, as I am going to show you in
moment--perhaps the most important risk factor. We
know quite a bit about heterogeneous neuroanatomy
and neurochemistry. We know that CNS insults, if

affecting key regions of the brain like the
prefrontal cortex, can produce very similar
problems as those produced by genes. Even
environmental factors can be important in ADHD. We
need to consider that environmental factors are not
bad mothers or bad teachers, like frequently
thought, but include things like poverty, exposure
to parental psychopathology, etc., etc., things
that in themselves can be driven by genes. So,
heterogeneous illness requires different treatment.
Different patients require different alternative
therapeutic options.


Another thing that has been highlighted
today but I would like to stress again is that ADHD
is a worldwide condition, not only an American
invention. It affects children in the 5-10 percent
range worldwide. Data are coming from Asia now,
from China and Japan and data from South American,
Western and Eastern Europe and, of course, North
and South American point to the fact that no matter
what criteria you want to use, it is an

extraordinarily common disease. You have to
remember that at least 50 percent--at least 50
percent of the children of today are going to be
adults tomorrow and we now know that ADHD affects
at least 4 percent or 5 percent of adults in this
country, not only that it affects them but it is
very morbid and dysfunctional.

I would like to stress that the yardstick
of considering the severity of an illness just by
mortality may not be an adequate yardstick. Many
conditions are devastating to our patients, even
though they are not necessarily lethal in the
traditional sense, like malignancies. This is a
condition that profoundly affects the lives of
those affected and everybody around them.


Many of the MRI studies that have been
conducted have been small and many of the children
participating in them have been medicated, creating
the suspicion that perhaps what we see in MRI
studies may reflect the toxic effects of
medications. Therefore, this study that I briefly

want to review for you is extraordinarily
important. This study is large. It was published
in a very prestigious medical journal, JAMA. It
was done by colleagues at the National Institute of
Mental Health. The lead author is Dr. Castellanos.
And, 152 children and adolescents of both genders
and a similar number of controls of both genders
were assessed longitudinally. The specific
objective of the study was to assess the issue of
medication status, whether medication is important
in brain abnormalities.


What this study found is that the brain of
children and adolescents with ADHD was
significantly smaller, in the 3 percent range,
independently of medication status. These
volumetric abnormalities were persistent over time
so this is not a neurodegenerative disease. It is
early disease that persists into adult life. There
were no gender differences and there was some
evidence of an association between severity of ADHD
and brain findings.


The visual of this study shows the
brains--these are males up to the age of 20;
females out to age 15. You can see that both
genders have significantly smaller brains and the
lines are flat over time.


The conclusion of this study is that
either genetic or early environmental influences on
brain development are operant in ADHD. These are
fixed, nonprogressive and unrelated to stimulant


If you look at key regions of the brain
that are involved in attention and executive
function, anyone of us in this room irrespective of
having or not having ADHD, we have this area of the
brain--this is the cingulate gyrus; this is the
dorsal anterior cingulate associated with executive
control; dorsolateral prefrontal cortex associated
with selective attention; and the right frontal
lobe associated with alerting. These are

interconnected areas, key regions for cognition and
attention. Their disruption will cause symptoms of


This is a recent study that we just
submitted for publication from our program at the
Mass. General. This is a three-dimensional
reconstruction of the anterior cingulate. What you
see here is a study of adults with ADHD. The
anterior cingulate area is 13 percent smaller in
individuals with ADHD compared with controls.


With imaging studies you can do not only
volume, as I just showed you with the
three-dimensional reconstruction so you can measure
volume of this region--this is the cingulate gyrus
again--but you can also measure the thickness of
the cortex, how thick or thin is the cortex in
critical brain regions.


So, this is another study that we have
done in our program of cortical thickness. This

has not yet been published but I promise you it
will be published. What you see, first of all, in
red here is a statistical comparison between the
brains of ADHD individuals compared with controls.
What this depicts, in orange and yellow, is
significant differences in cortical thinness.
These areas are selectively thinner in these
regions. So, you don't see thinness across the
entire brain; you see thinness in critical cortical
regions involved with executive control and
attention. This is the dorsal anterior cingulate.
This region hovers between the cognitive and
emotional division of the anterior cingulate--very
important issues for clinical understanding of the
symptoms of this condition. This is, of course,
the dorsolateral prefrontal cortex that is clearly
involved in cognition and in the symptoms that
subserve this illness.


We have also done this analysis. It is
very exciting. This is diffusion tensor imaging
that measures white tracts. What you see here is

that we are documenting disruptions in the
perigenual area of the anterior cingulate and
dorsal anterior cingulate selectively. So, this
area of the brain that is involved in cognition,
executive functions and regulatory controls is
smaller in volume, is thinner in cortical thickness
and has other abnormalities as well. I am not
aware of many other psychiatric illnesses can claim
such conversion of information, focusing on the
same brain regions that could certainly account for
the clinical picture.


If you look at functional MRI in the same
region, if you look at the coronal view of the
brain, if you put people without ADHD on the
scanner you can very nicely activate anterior
cingulate doing a very mild cognitive task. If you
put adults with ADHD, they fail to activate the
same region and, instead, they activate insular so
these adults can do the task but they are not using
the part of the brain that is specifically wired to
do the task at hand. We have very exciting new

data that you can correct this malfunction with



These findings on neuroimaging are
extraordinarily congruent with conceptualization
from neuropsychology. As you know, ADHD is
considered a neuropsychological disease. What is
called directed attention, the circuit that allows
people to pay attention to things that they are not
interested in is disrupted. Inhibitory deficits,
the person fails to inhibit when destructions
occur; and executive dysfunction issues of planning
and organization, working memory, etc., etc. are
disturbed. These are the regions that are in this
part of the brain where we are documenting

Another circuit that is involved in ADHD
is called fascination reward circuit. People with
ADHD have difficulties with delayed gratification;
difficulties with regulating mood. This kind of
hot temper that characterizes people with ADHD and
some of the road rages that lead to accidents, and

so on and so forth, may be accounted for by these

neuropsychological deficits.


Another key aspect to document that ADHD
is a neurobiological disorder is genetic research.
ADHD clearly runs in families. There is a 5-7-fold
increased risk of ADHD in first-degree relatives of
children with ADHD. Of course, that is not
evidence for genetics so we need to have additional
information to make a genetic hypothesis or genetic
claim. Twin studies are very important because
twins come from two varieties, monozygotic and
dizygotic twins. For genetic illnesses, you expect
that monozygotic twins will have a higher level of
concordance than dizygotic or fraternal twins.

Twin studies are also very important
because they can allow us to compute coefficients
of heritability that I will tell you about in a
moment. Adoption studies are important because
with genetic illness you expect to have a higher
rate of the disease in biological rather than
adopting relatives. Finally, molecular genetic

studies will look at specific genes that may be

associated with this condition.


Coefficients of heritability are based on
twin studies. I would like to point out to you
that there are a lot of twin studies in ADHD.
Coefficients of heritability range from zero/zero
percent of the variance accounted for by genes to
100/100 percent of the variance accounted for by
genes. The congruence of the genetic studies for
the coefficient of heritability in ADHD is
remarkable. Even though the studies use different
methods, parent support, teacher support,
structured interviews questionnaires, look how
consistent this is. On average, coefficients of
heritability are close to 80 percent, in other
words, 80 percent of the variance of ADHD can be
accounted for by genes. For example, height, a
very genetic trait, is about 90 percent genetic;
schizophrenia and bipolar illness, very genetic.
Recognized genetic illnesses are equally genetic as
ADHD, 80 percent. Panic disorder and major

depression are genetic as well, about 50 percent,

0.5 for coefficient of variability; other medical
conditions like asthma or breast cancer at 0.3, 0.4
coefficients of heritability. So, we are dealing
with a very genetic brain disorder here.


Specific genes have been associated with
ADHD. The first genes that were looked at in ADHD
are genes that are associated with the dopamine
system, candidate genes because the drugs that we
usually use to treat this condition are
dopaminergic drugs. Mutation in a dopamine
transporter gene, what is called DAT1 or DAT10;
mutation in the dopamine receptor of D4 and D5
genes--these are cortically distributed receptors.
There is also an association between a very rare
thyroid disease on chromosome 3. People that have
this mutation also have ADHD, but this probably
accounts for very little of ADHD out there.

So, the genes that have bee associated
with ADHD in a consistent fashion are the dopamine

transporter genes that bring back dopamine to the
presynaptic neuron. This gene over-expresses the
dopamine transporter, make more it active or more
transported, too much dopamine to the presynaptic
neuron. Cortically distributed genes are the D4
and D5. There is also a gene, SNAP25, that is
involved in the presynaptic encapsulation of


Well, I see this slide is totally
degenerated. I am very sorry. But what I would
like to say is that Dr. Faraone published in
Biological Psychiatry a review of a meta-analysis.
There are certain genes, about seven genes that
consistently have been associated with ADHD,
several genes in the serotonergic system, DBH,
dopamine transporter gene and also dopamine
receptor DR D4 and DR D5. These genes have been
found in multiple studies in meta-analyses of these
specific genes to be associated with ADHD. You
probably have these details in the handout, as well
as the reference for the paper that support these



The treatments that we have available are
clearly effective. This study was not done by
industry. It was funded by the NIMH on the
multi-modal treatment of ADHD. This study studied
close to 600 children 7-9 years of age in this
country, in 5 sites. This was a study in which
children were randomized to very aggressive
medication management, very aggressive,
comprehensive behavioral treatment, a behavioral
treatment that was so comprehensive that you could
not possibly improve on it, and it is so expensive
that it is not doable. But that was the purpose of
the study. Children received both medication and
behavioral management and community-based


What this study found, and I would like to
point this out to you, is that these two arms, the
very aggressive medication arm and combination
treatment were superior to just behavioral

treatment and community based treatment that
consisted largely of communication too. This
group, here, is a group that received the best
treatment that we can offer, very aggressive
pharmacotherapy and very aggressive behavioral
treatment. And, even using the best we can, the
response is 60 percent, leaving 40 percent of our
patients inadequately treated with intolerable side
effects or with difficulty tolerating this


Stimulants are Schedule II drugs. About
40 percent or perhaps more do not tolerate or to
not respond to these treatments. The side effect
profile of sleep, appetite, difficulties with mood
and anxiety can seriously hinder our ability to
treat all the patients that otherwise could
benefit, and concerns about growth suppression and
tic development continue to plague stimulant
treatment in ADHD, Although the data is generally
reassuring, people continue to be concerned about
these issues.


ADHD not only affects school and school
performance, it is clearly a life disease. For
example, ADHD has been shown to be associated with
a significantly increased risk for automobile
accidents, and automobile accidents, as you know,
are the leading cause of death in our young. So,
patients not only have poor grades in school or may
not reach college but may not reach adult life
altogether or may kill somebody in the process of
driving and not paying attention to the road.


So, if you look at the morbidity of ADHD
as a serious--perhaps not a lethal illness but a
very devastating illness to the individual, the
family and society secondary to under-achievement,
under-employment, marital difficulties, drugs and
substance abuse, legal difficulties, or morbidity,
we are dealing with an extraordinarily morbid
disease that can profoundly affect those afflicted
with this condition.


So, in summary, ADHD is a neurobehavioral
disorder with complex etiology. It is a disorder
that affects the brain; has a strong genetic
component, as I alluded to; affects millions of
people worldwide, both sexes. It is highly
persistent in the majority of those affected. It
has a negative impact on the life of the individual
affected and everybody around. Although the
stimulants are clearly an effective treatment for
ADHD, a sizeable number, in the order of magnitude
of 40 percent, are non-responsive or not tolerating
this treatment, calling for alternative treatment
for this condition. Thank you very much for your


The next presentation will be by Dr.
Lesley Russell who will be talking about clinical
pharmacology and efficacy.

Clinical Pharmacology

DR. RUSSELL: Thank you, Dr. Biederman.

Today I would like to briefly overview the
clinical pharmacokinetics of modafinil in children,

and then summarize for you the efficacy findings of

our program in ADHD.


Just to begin with, here is an outline of
the development program that Cephalon undertook in
children and adolescents with ADHD. The three
Phase 3 studies, studies 309, 310 and 311, formed
the basis of the efficacy and safety that will be
discussed today. In addition, we had two
pharmacokinetic studies, studies 113 and study 206,
that outline the pharmacokinetic parameters of
modafinil in children. In addition, we conducted
two studies, 207 and 213, to help us define the
dose required for the Phase 3 studies. All
patients in the Phase 3 studies and some from the
Phase 2 studies were allowed to enroll into study
312 which is an ongoing open-label extension
program. Following submission of the sNDA, we
initiated a further study, study 3044, in which 303
patients were enrolled. This study is still


I would now like to briefly summarize the
pharmacokinetics of modafinil in children.


As shown on this slide, the
pharmacokinetics and exposure are dose-proportional
over the dose range studied. The absorption is
rapid, with a maximum concentration observed 2-3
hours after administration. When administered with
food, there is an approximate 1 hour delay in the
time to Cmax although the overall absorption is not
affected. The volume of distribution increases in
children linearly with their weight. The
metabolism of modafinil is primarily hepatic, with
less than 10 percent excreted unchanged in the
urine. There are 2 primary metabolites, modafinil
acid and modafinil sulfone. As you heard earlier,
we did observe higher levels of modafinil sulfone
in the younger children. The elimination of
modafinil is time- and age-dependent. We observe a
decrease in clearance over time with steady state
being reached by about week 6 of treatment. There
is a gradual decrease in clearance with age, with a

pronounced shift between 9-11 years of age. So, we
see that there is a half-life in the younger
children of approximately 7 hours which is compared
to a half-life of 15 hours in the adults.


I would now like to outline for you the
basis of the dose selection that we used in the
Phase 3 studies. Study 207 was a relatively small,
double-blind, randomized, 4-period crossover study,
and this was the first program undertaken to assess
the efficacy of modafinil in the treatment of ADHD.

The results shown for you are the total
scores on the ADHD rating scale as assessed by the
parent. With the caveats of this being a small
study, you can see that those patients who received
100 mg barely discriminated from placebo. A
slightly larger treatment effect was seen with the
200 mg dose group and a larger treatment effect was
seen with the 300 mg or 400 mg dose group. And, I
should reiterate here that the 300/400 mg doses
were administered based on weight, with the 300
given to children less than 30 kg and the 400 to

those children weighing at least 30 kg or more.


The next study we undertook was study 213,
and this was designed to see the best way of
administering a single dose of 300 mg either as a
single dose administered first thing in the
morning, which is depicted in blue, or as a split
dose of 200 mg in the morning and 100 mg at lunch,
depicted in orange, 100 mg in the morning and 200
mg at lunch time, depicted in pink and compared to

As you can see from the slide, there
appeared to be little benefit to splitting the dose
and the largest response we saw was in the 300 mg
dose group administered as a single dose. As you
can see on the right-hand side of the slide, this
is for all patients, but when we looked at it
stratified by weight you can see that it is clearly
the younger and lighter children that had the
larger response. So, from this study we concluded
that older and heavier children may require a
higher dose.


This slide identifies the systemic
exposure that we saw following these dosing
regimens. In the middle, here, are those children
weighing less than 30 kg who received 300 mg. Here
are those children weighing more than 30 kg who
received. You can see that the systemic exposure
associated with the lighter children is clearly
higher than the systemic exposure with the heavier


Using these data and the clinical efficacy
results from the Phase 2 studies, we developed a
pharmacokinetic/ pharmacodynamic model and
estimated that the systemic exposure which would be
associated with a consistent pharmacodynamic effect
would be in the order of 150 mcg/hour/ml, and that
the doses that would be required to achieve this
exposure at steady state would be 340 mg for those
children weighing less than 30 kg and 425 mg for
those children weighing 30 kg or more.


Following the Phase 3 program which
included sampling from population pharmacokinetics,
we went back to test this hypothesis. As you can
see from this slide, we pretty much got it right in
that here are the children weighing less than 30 kg
who received 340 mg and here are those heavier
children who received 425 mg, and the systemic
exposure in those groups is pretty similar, around
150 mcg/hour/ml. As you will see from the next
slides, these doses were associated with
substantial efficacy.


Here are the designs of the 3 pivotal
studies that were undertaken. All studies were
double-blind, randomized, placebo-controlled and
had a 2:1 randomization. Study 309 and 311 were
identical in design. Both were 9 weeks in duration
and employed a flexible dose titration regimen
whereby children could be titrated from a minimum
dose of 170 mg to a maximum dose of 425 mg based on
perceived efficacy and their tolerability to
treatment. The dosing increments occurred on a

weekly basis at 85 mg.

Study 310 was slightly different. This
study had a 7-week efficacy period and a 2-week
period that assessed abrupt discontinuation of the
drug, the results of which I will not show you
today. This study was also a fixed dose study and
patients were titrated at 85 mg increments every 2
days to their target dose based on weight, so 340
mg for the children weighing less than 30 kg and
425 mg for those children weighing 30 kg or more.


The patients enrolled in the study were
very similar. All patients were 6-17 years of age
with a diagnosis or ADHD according to the
Diagnostic and Statistical Manual of Mental
Disorders. The children were required to be at
least moderately ill on the Clinical Global
Impression of severity and have an ADHD rating
scale which was at least 1.5 standard deviations
above the norm for age and gender. Patients were
required to be of normal intelligence with no
learning disability and attend school full time.

Patients were excluded from study if they
failed to respond to 2 or more adequate courses of
stimulant therapy, although it should be noted that
patients who had failed one stimulant therapy were
allowed to be enrolled. Patients were also
excluded if they had psychiatric comorbidities
requiring current pharmacotherapy and were well
controlled with their current ADHD therapy and had
no good reason to change treatments.


The efficacy assessments were identical
for each of the 3 studies. The primary outcome
measure was the change from baseline in the Total
Score ADHD Rating Scale as assessed by the teacher.

Secondary outcome measures included a
change from baseline in the Home ADHD Rating Scale
as assessed by the parents in the evening between

6:00 and 8:00 at night; the Clinical Global
Impression of change as assessed by the treating
physician; the Conners' Parent Rating Scale as
assessed by the parent; the test of Variables of
Attention, which is a continuous performance test;

the Social Skill Rating Scale and the Child Health



As you can see, the average age of
patients entering into the program was around 10,
with the majority of patients being less than 12
years of age. The majority were boys and white,
and about two-thirds of the patients actually
weighed 30 kg or more.


As per inclusion criteria, patients were
required to be at least moderately ill on the
Clinical Global Impression of severity and, as you
can see from the slide, about 50 percent of the
patients were considered to be moderately ill and
the other 50 percent were considered to be markedly
or severely ill. Around two-thirds of the patients
had the combined inattentive and hyperactive
subtype of ADHD. About a third were predominantly
inattentive and very few were purely hyperactive.
The baseline ADHD rating scale at entry was on
average 37, which is well above the norm for a

10-year boy which is roughly 18.


A total of 638 patients were randomized
into the study and 630 received treatment, 420 in
the modafinil treatment group and 213 in the
placebo treatment group. Around two-thirds of the
patients completed the double-blind treatment
period, with the reasons for discontinuation
outlined here. As you can see, some of the main
reasons for discontinuation were lack of efficacy
with a much higher proportion in the placebo
treatment group, and adverse events with the higher
proportion in the modafinil treatment group. The
other reasons are listed for you here.


The following three slides will show the
outcomes of the primary efficacy variables for each
individual study. Here are the results for study
309, the first of the flexible dose studies. Just
to orient you, on the Y axis is the Total ADHD
Rating Scale with the lowest score showing benefit,
and across the X axis is the duration of the

treatment period in weeks.

Just as a reminder, the primary outcome
variable was the change from baseline to endpoint
using the last observation carried forward
analysis. As you can see from this slide, there
was a statistically significant difference in favor
of those patients being treated with modafinil.
More specifically, the treatment effect on the
modafinil treatment group was 17.5 points with the
treatment effect on the placebo group of 9.7 points
for an effect estimate, which is the difference
between the 2 treatment groups using the Lee
squared means of 7.4. In addition, statistically
significant results were seen using the observed
cases analysis.


A similar result was seen in study 311
which is the second flexible dose study. At
endpoint the treatment effect on the modafinil
treatment group was 15 points and a treatment
effect on the placebo group was 7.3 points for an
effect estimate of 8. Again, this is statistically

significant both at endpoint and using the observed

case analysis.


The last study is study 310 and, again, a
very consistent treatment effect was seen in this
study, with a treatment effect of minus 17.2 points
on the modafinil group versus 8.2 points on the
placebo treatment group for an effect estimate of

Outlined for you on this slide is just

another way of viewing the data. On the left-hand
side of the screen is the modafinil treatment group
at baseline and at endpoint. On the right-hand
screen is the placebo group at baseline and at
endpoint. This slide illustrates the remarkably
consistent effect seen not only in the modafinil
treatment group but in the placebo treatment group.
The red line depicts what would be considered to be
a normative value on the ADHD rating scale for a 10
year-old boy. As you can see, those patients
treated with modafinil are beginning to approach

this normative value.


We also assessed the effect of treatment
using a responder analysis on the School ADHD
Rating Scale with those patients who had at least a
30 percent reduction in their scores from baseline
to endpoint or a 50 percent reduction from baseline
to endpoint. As you can see, in all 3 studies a
significantly higher proportion of patients treated
with modafinil had either a 30 percent or a 50
percent reduction in their ADHD symptoms.


This slide shows for you in all 3 studies
the home version of the ADHD Rating Scale. As a
reminder, this was assessed by the parents in the
early evening. The results seen here very much
mirror the results we saw using the school version
of the ADHD Rating Scale, with significant
differences seen both at endpoint and in the
observed case analyses in all 3 studies in favor of
the modafinil treatment group.


Depicted here is the responder analysis on
the Clinical Global Impression of improvement.
Outlined for you are those patients who were either
considered to be much or very much improved by the
treating physician. Again, in all 3 studies we see
a very consistent treatment effect, with a
significantly higher proportion of patients
considered to be much or very much improved on this
scale by the treating physician.


Another commonly used scale for assessing
ADHD and their response to medication is the
Conners' Parent Rating Scale. Again, you can see
in each of the 3 studies, using this scale, a very
similar effect to the observation seen using the
ADHD rating scale, with improvements on the
modafinil treatment group in cognitive problems and
in attention, hyperactivity and their total ADHD

This scale also allows the assessment of
treatment on the oppositional behavior. As you can
see, in all 3 studies there appears to be a

treatment effect in favor of modafinil, although
this is only statistically significant in study

The one inconsistent effect that we saw

was using the Test of Variable Attention. Outlined
for you in this study is the pooled analysis using
data from all 3 studies. Although you can see that
those patients treated with modafinil tend to do
better than those patients treated with placebo, it
should be noted that this is actually a decline in
performance rather than an improvement in
performance over time.


Children with ADHD often have poor peer to
peer relationships and difficulties with
socialization. We wanted to assess the effects of
treatment on these parameters and we used the
Social Skills Rating Scale. Again, this is the
data from all 3 studies pooled. The individual
studies did show a consistent treatment effect. As
you can see, there appears to be an improvement in

many of these parameters when treated with

modafinil, including the Social Skills Total Scale.


A similar improvement was seen in other
problem behaviors as measured by this scale. It
should be noted that these results were only seen
in the children in grades kindergarten to 6th grade
and we observed no major differences between
treatment groups in the older age groups.


Lastly, here are the results of the Child
Health Questionnaire, a global sort of quality of
life instrument that assesses many behaviors that
can be impaired with ADHD. Again, this is the
pooled analysis of all 3 studies. As you can see
from this slide, there appears to be an improvement
in many of the behavioral aspects seen for those
patients treated with modafinil--


--including an improvement in the total
psychosocial summary. We did not see significant
improvements in the physical functioning domain,

although it should be noted that these values were

normal at baseline.


We have undertaken many subgroup analyses,
many of which are outlined in your briefing
document. Here I just want to show for you the
subgroup analysis for those patients who were
either stimulant naive at study entry, and that was
for about 50 percent of the patients, and those
patients who had received a prior stimulant before
enrolling into the study, which was again about 50
percent of the patients.

Here you can see that treatment with
modafinil was effective even in those patients who
had had prior stimulant therapy, although it should
be noted that the treatment effect appears to be
larger in those who were stimulant naive.


In conclusion, we saw consistent efficacy
results across 3 pivotal studies. The improvement
in ADHD symptoms was seen by the teachers, the
parents and the treating physicians. Improvements

were seen at school, at home and across the day.
As well as improvement in the core ADHD symptoms,
we did observe improvement in other psychosocial
domains. Finally, we saw efficacy in stimulant
naive patients and in patients who had had prior
stimulant experience.

I would now like to hand over to Dr.
Srdjan Stankovic who will outline safety for you.
General Safety

DR. STANKOVIC: Thank you. My name is
Serge Stankovic and I am with the Cephalon clinical
research group.


My presentation this morning on modafinil
safety is organized as follows: I will review
overall modafinil exposure in clinical trials.
Following that, I will review the safety data for
the modafinil ADHD program in children and
adolescents, and this will include review of
general safety and events of special interest such
as skin reactions and psychiatric events. In the

balance of my presentation I will briefly summarize
high level safety information from our development
program in excessive sleepiness in pediatric
patients. Finally, I will review modafinil
information coming from our postmarketing safety


Overall, safety data for 933 patients with
ADHD were included in the supplemental NDA,
submitted in December of 2004. Of these, in the 3
Phase 3 placebo-controlled trials, 420 patients
were treated with modafinil and 213 patients were
treated with placebo. Please note these numbers as
I will often refer back to them when I am
presenting data from our controlled trials.

Following the sNDA submission, one
additional open-label study in children with ADHD
was initiated. With that, as of February, 2006, a
total number of pediatric ADHD patients exposed to
modafinil was 1236. Additional pediatric exposure
comes from our development program in excessive
sleepiness, 270 pediatric patients, and from

pediatric patients exposed to a variety of foreign
studies for various indications, 116 patients.

Finally, just a reminder that 4000 adult
patients were exposed to modafinil in the
development program for excessive sleepiness and in
other clinical trials.


Looking at patient exposure in the
pediatric ADHD program, this slide presents
exposure by modal dose and duration for 933
patients as of February 1, 2006. A total of 246
patients were treated with modafinil for a minimum
of 12 months, and as many as 164 were on drug for
18 months or longer. About half of the patients
received modafinil at the modal dose of 425 mg a
day, while about one-third at the modal dose of 340
mg a day. The total exposure to modafinil in the
pediatric ADHD program is 575 patient-years.


Next I will discuss adverse events
observed in ADHD studies of children and


A general overview of adverse events
reported in 3 Phase 3 placebo-controlled studies is
presented on this table. While the majority of
patients in both groups experienced at least one
adverse event, a higher incidence was observed in
the modafinil treatment group. Relatively few of
these events were reported to be severe, were
reported to be a reason for study discontinuation
or were reported to be a serious regulatory
definition of that word.


The most commonly observed adverse events
in the Phase 3 placebo-controlled studies were
insomnia, headache and anorexia. The COSTART term
of anorexia used here includes both loss of
appetite and decreased appetite. In fact, about 70
percent of patients reporting anorexia experienced
decreased appetite. Insomnia and anorexia were
reported at a substantively higher rate in the
modafinil group compared to placebo. Review of
these two events indicated that very few were

reported as severe, specifically 9 out of 115
events for insomnia and 1 event for anorexia.
Likewise, only 5 events in insomnia led to
discontinuation, while 2 patients reporting
anorexia discontinued study due to that adverse
event. In most instances, these 2 events first
occurred in the initial 2 weeks of treatment and
the median duration reported was about 2 weeks.


Out of 933 patients included in the sNDA,
18 patients experienced at least one serious
adverse event by the time of the 10-month safety
update submitted in November of 2005. Four of
these patients were enrolled in the 3
placebo-controlled Phase 3 studies and all of them
were in the modafinil treatment group. From the 2
ongoing pediatric studies in ADHD, 3 patients
experienced a serious adverse event during the
period up to February, 2006. Discussion of serious
skin adverse events as well as psychiatric events
in more detail with be part of the discussion of
special safety.


In the next four slides I will review
relevant information related to laboratory
evaluations from the pediatric ADHD studies. Data
for selective hematology and blood chemistry
parameters will be reviewed in more detail.
Although included in your background package, data
for other laboratory parameters did not raise
questions or concerns and, therefore, will not be
presented here today.


Based on some early observations from the
Phase 2 studies, concern was raised regarding
modafinil treatment effects on absolute neutrophil
count and white blood cell count in children. Our
Phase 3 controlled data did not show a meaningful
difference in mean change from baseline or
incidence in clinically significant values between
modafinil and placebo. Furthermore, as presented
on this slide, when the lowest on treatment values
are grouped by range there was no meaningful
difference between modafinil and placebo treatment



With respect to serum chemistry, as in
adults, we did observe a difference in mean change
from baseline between modafinil and placebo for
alkaline phosphatase and GGT. In the Phase 3
placebo-controlled studies there were few patients
experiencing a clinically significant change on any
of the parameters, with no apparent imbalance
between treatment groups. On the next slide we
will discuss LFT elevations highlighted in the
background document as cases of possible concern.
These cases are included in this table in the
column for all modafinil studies.


In the FDA approvable letter it was stated
that although controlled trials data did not reveal
a signal for drug-related mean increase in
transaminase values or in drug-related outliers,
there were 3 modafinil-treated patients who had
transaminase increases of concern, but insufficient
other information to further assess the

significance of these changes. Details related to
these 3 patients are presented on the slide.

In all 3 cases, total bilirubin values
both at the time of observation of abnormal LFT
values and throughout the study were normal. In
one case laboratory abnormalities returned to
normal while patients continued treatment with
modafinil. In the second case treatment was
continued for an additional 6 months prior to study
discontinuation. At that time, all abnormal LFT
values returned to normal except for a mild
elevation in ALT. In the third patient abnormal
values returned to normal after withdrawal of
modafinil. This case will be discussed later in
relation to possible hypersensitivity reactions.


The next segment of the safety
presentation is focused on cardiovascular safety.
I will review blood pressure and pulse data, ECG
information including QTc interval and
cardiovascular adverse events from the Phase 3
placebo-controlled trials. It should be noted that

the vital signs measurements in ECGs were recorded
in these studies at variable time points during the
day and in relation to the intake of study


With respect to blood pressure, no notable
effects in sitting blood pressure were observed in
the Phase 3 controlled studies. Presented on this
slide are box plots for systolic blood pressure on
the left side of the screen and diastolic blood
pressure on the right side of the screen in
modafinil and placebo treatment arms respectively.

Changes from baseline for both systolic
and diastolic blood pressure were similar in the 2
treatment groups with respect to both mean values,
overall distribution and extreme outliers.


This graph presents the distribution of
observed change from baseline in sitting pulse for
the 2 treatment groups. As presented, we observed
similar distribution between the 2 treatment arms
and the occurrence of outliers.


Review of the ECG tracings from the ADHD
pediatric studies did not reveal specific concerns
both with respect to morphology or interval
measures. This slide presents an overview of QTc
interval data from the 3 placebo-controlled trials
expressed as maximum change from baseline or as
maximum duration observed. The slide presents data
for QTc using the Fridericia correction, but the
findings are similar when other corrections are
used. Either way, there is no apparent effect on
QTc interval or imbalance between treatment arms.


Finally, when reported adverse events are
reviewed, we observe relatively few cardiovascular
events. Only a small fraction of these, 2 patients
on modafinil and 1 on placebo, reported events
leading to treatment discontinuation. In all 3
cases the stated reason for discontinuation was
tachycardia. None of the reported cardiovascular
events were reported to be serious.


Important consideration in the safety
evaluation of any ADHD compound is assessment of
its effects on growth. [Slide]

In the placebo-controlled Phase 3 studies
modafinil treatment of up to 9 weeks duration led
to relative weight loss compared to weight gain
observed in the placebo group. Similarly, a
significantly higher proportion of
modafinil-treated patients experienced clinically
significant weight loss, defined as at least 7
percent in weight reduction. To be precise, 9
percent of modafinil-treated patients versus 1
percent of placebo-treated patients experienced
significant weight loss during the study.


Naturally, we did look at the longer term
treatment data related to weight and growth in
general. As you know, for accurate evaluation of
growth effect in children, we need to evaluate them
relative to norms. To achieve this, we expressed
changes in weight and height using Z-scores. Just
a quick reminder, Z-score is a statistical measure

that quantifies the distance measured in standard
deviations of a patient data point, in this case
individual weight or height, from the population
mean, in this case CDC growth norm for
corresponding age and gender.

This graph presents mean weight and height
Z-scores over 12 months of treatment with
modafinil. A decline in Z-score is observed
initially in the first 3 months of treatment
consistent with the reported weight loss in our
short-term trials. In the following months the
line remains horizontal, meaning that the normative
pattern of growth is regained. Using the same
presentation, it is apparent that there was no
indication of adverse effects on height over the 12
months of treatment with modafinil.


In the course of the modafinil ADHD
pediatric development cases of serious skin
reactions were reported. Some of these were
indicative of a possible Stevens-Johnson syndrome
or hypersensitivity reaction, generally a rare but

very serious complication of treatment. Cephalon
shares the important concerns raised by our
colleagues at FDA in regard to these events.
Therefore, I will review skin events in greater


To bring everybody on the same page with
respect to cases of interest, I will start with the
list of events, included in the FDA briefing
document, in the second dermatology consult report
dated February 27, 2006. In this report the events
were grouped in 3 categories based on the level of
diagnostic confidence. The 3 groups are events
representing EM, SJS or TEN; events somewhat
suggestive but lacking confirmation; and events
resembling prodromal presentation but without
sufficient information for diagnosis. Cephalon has
performed a similar review and in the next two
slides I will review cases from the first two
groups. With respect to the third group, our
review did not support the conclusion that any of
these cases should be classified as SJS or

prodrome. We based this on the low specificity and
low predictive value of reported symptoms.
Additionally, many of the symptoms are quite common
and many were not reported concomitantly or


First, we will review the clinical trial
cases. Patient number 1 is a 7 year-old boy who,
on day 16 of treatment with a 340 mg dose,
presented with symptoms described by the
investigator as erythema multiforme,
Stevens-Johnson syndrome and both FDA and Cephalon
reviewers agreed that the diagnosis of
Stevens-Johnson syndrome is likely accurate, with
less consensus on the possible etiology. I am sure
that this case will be discussed in more details
later and, as Dr. Raczkowski said, we have the
investigator here who was treating the patient, as
well as members of our panel of dermatologists who
can talk more about the case.

Patient number 2 is an 11 year-old girl
reported with morbilliform rash on day 15 with

treatment of a 200 mg dose of modafinil. This
patient was hospitalized and the SJS diagnosis was
excluded. FDA review indicated that this was a
case representative of EM/SJS. Cephalon's panel of
independent reviewers, on the other hand, was
unanimous that the reported diagnosis of
morbilliform rash is probably correct and the event
did not represent Stevens-Johnson syndrome.

Patient number 3 is a 6 year-old boy who
reported rash, fever and vomiting 2 weeks after
initiation of treatment. Review of the source
documentation received from the investigator
indicated that this event was diagnosed as fifth

Patient number 4 is an event in an 8
year-old boy described as rash on the cheeks and
blisters on the lips, and was reported as erythema
multiforme. The event occurred on day 23 of
treatment with a 300 mg dose of modafinil. This
case is considered by the FDA reviewer as somewhat
suggestive but not representative of EM/SJS or TEN.
Cephalon's reviewers, on the other hand, agreed

that this is unlikely erythema multiforme, but did
not agree on the alternative diagnosis. One
considers this event to be possible SJS. A second
reviewer considered it to be probable herpetic
gingivostomatis and a third independent reviewer
attributed to the event as either viral etiology or

Patient number 5 is a 9 year-old boy with
reported symptoms of urticaria, fever and facial
edema. This patient also had elevated
transaminases. Cephalon's review indicates that
this is a possible case of hypersensitivity
reaction and it is not consistent with SJS.


In the review of postmarketing reports
both FDA and Cephalon concluded that there were 4
reports of serious skin reactions, 1 SJS/EM and 3
SJS reports. Of the 12 suggestive but not
confirmed cases on the FDA list, Cephalon has
identified 8 reports considered suggestive of
possible hypersensitivity but not indicative of EM,
SJS or TEN spectrum. The other 4 cases were also

not considered suggestive of SJS.


In the Phase 3 placebo-controlled trials
the incidence of rashes coded by the COSTART coding
system was 4 percent in the modafinil treatment
group and 2 percent in placebo. As we all know,
the preferred term "rash" in the COSTART coding
system does not include many terms that could be
considered non-urticarial rash. Therefore,
Cephalon undertook an additional analysis to
ascertain the incidence of non-urticarial rash. In
collaboration with 2 external dermatology experts,
we defined a category of non-urticarial rash which
included all adverse events indicative of rash,
excluding urticaria and related reactions.

Using this definition, cases of
non-urticarial rash in the pediatric ADHD studies,
as well as in the pediatric studies for excessive
sleepiness and in all adult studies with modafinil
were identified and frequency tables were
constructed. Additionally, all reported adverse
events of urticaria, hypersensitivity reactions and

all allergic reasons in the pediatric ADHD studies
were reviewed for possible underlying causality and
prior medical history.


Based on the described methodology, we
calculated the incidence of non-urticarial rash
across treatment groups in controlled pediatric
ADHD trials and in all pediatric patients. This
table presents the incidence in the
placebo-controlled trials. We also present the
incidence of those described as severe and those
leading to treatment withdrawal. The overall
incidence of rash was higher in the modafinil
treatment groups, with few being described as
severe or leading to treatment discontinuation.


In the ongoing open-label study in ADHD
initiated after the supplemental NDA submission a
total of 303 additional newly exposed patients
entered the study, with 188 receiving modafinil for
at least 4 weeks. Presented on this slide is the
observed incidence of non-urticarial rash in that

study. As in the previous slide, we also present
the incidence of events described as severe or
those leading to discontinuation.

As seen on this slide, the reported
incidence is somewhat lower compared to the
modafinil group in the controlled studies. One
patient reported a severe rash on day 10 and
discontinued the study on day 13 due to this rash
which was described by the investigator as rash.


The overall incidence of non-urticarial
rash reported in the controlled pediatric studies
for excessive sleepiness was similar between
modafinil treatment groups and placebo. These are
much smaller studies Additionally, the observed
incidence was lower compared to ADHD pediatric
studies. Only one event was reported as severe for
events described by the investigator as fifth
disease. No events led to discontinuation or were
serious by regulatory definition.


The observed imbalance in incidence of

non-urticarial rash in the controlled ADHD
pediatric studies prompted further evaluation for
possible association with treatment. We approached
this in 3 ways. We evaluated the relationship
between rash and dose; relationship between rash
and modafinil plasma exposure; and, finally, we
evaluated the relationship between rash and
modafinil sulfone exposure, one of the metabolites
known to be present in higher concentrations in
children. With respect to relationship of
non-urticarial rash and dose, we conducted a
case-control analysis where patients with rash were
matched with controls based on the study protocol,
time in the study to event and weight. Based on
this analysis, we did not find statistical evidence
for association between rash and modafinil dose.


A second analysis looked at the modafinil
plasma exposure by comparing area under the curve
between patients reporting non-urticarial rash, in
the far left box on the slide, controls, in the
middle box, and overall patient population in Phase

3 studies, in the far right box. Areas under the
curve were calculated based on sparse sampling data
from the Phase 3 trials and PK modeling. As
presented on this slide, no difference was apparent
between the 3 groups.


An assessment of the relationship between
non-urticarial rash and exposure to modafinil
sulfone was also conducted. Here we graphically
depict the distribution of modafinil sulfone
concentrations in patients developing rash--small
red boxes at the bottom, and in patients not
developing rash--blue boxes. On the Y axis the
number of patients is depicted and different
modafinil sulfone concentration ranges are depicted
on the X axis.

We observed that the distribution of
sulfone concentrations in patients with rash
appears to closely mimic the distribution of
sulfone concentrations in the full population of
treated patients in placebo-controlled studies. We
conclude, therefore, that there appears to be no

correlation between non-urticarial rashes and
systemic exposure to modafinil sulfone. One
additional piece of information is that 2 cases in
question had modafinil sulfone concentration of
less than 6 mcg/ml.


We have also examined the adverse events
database from the controlled ADHD pediatric studies
for COSTART preferred terms indicative of
urticaria, hypersensitivity reactions or allergies.
This slide presents a tabular summary of the
reviewed preferred terms and associated medical
history reported prior to treatment initiation.
One can easily see from the table that the vast
majority of these events was reported in patients
with prior history of seasonal allergies or asthma.


Psychiatric adverse events related to ADHD
treatment have enjoyed special interest in the
recent months, culminating in some important
discussions as recently as yesterday. In response
to the request from the Division issued to all ADHD

drug manufacturers, Cephalon has performed a full
analysis of psychiatric events from all pediatric
studies and from our pharmacovigilance database as
per prespecified methodology.

In addition, we have reviewed serious
adverse events occurring after the last safety
update cutoff in October, 2005, covering the period
through February 1, 2006. The results will be
presented in the next several slides.

DR. GOODMAN: Excuse me just a moment, I
want to ask a question of clarification on the
previous metabolite levels that you showed. What
was the relationship between the timing of
obtaining the sulfone metabolite level and the
dosing? Obviously, there can be a lot of noise
contributed by relationship between time of assay
and dose.

DR. STANKOVIC: We obtained the values for
concentrations of modafinil sulfone closest to the
event for those patients that reported a rash.

DR. GOODMAN: But it might not have been
the same relationship to the time the dose was

actually taken. Right?

DR. STANKOVIC: That is right, yes. That
is correct.


A brief introduction on methodology of the
psychiatric evaluation, all adverse events reported
in the ADHD and excessive sleepiness pediatric
programs were subject to a review by a string
search for COSTART preferred terms of investigator
verbatim terms indicative of psychiatric events.
Once identified, all events are classified in the
following groups, psychotic events including mania,
suicidal ideation and behavior, aggressive and
violent behavior and miscellaneous psychiatric
events that were serious by regulatory definition.
A similar string search approach was employed in
the review of our psychovigilance reports. Event
terms and narratives from the ongoing pediatric
studies for serious adverse events were reviewed in
order to identify psychiatric events as well.


We present here psychiatric adverse events

from the ADHD pediatric program. Just a quick
note, this table includes both events that occurred
during treatment as well as those that occurred 48
hours following last dose of modafinil. As I will
be discussing these cases, we put them together.
This is somewhat different than the methodology
applied in the tabulations presented yesterday.

In the controlled studies all psychotic
events, as well as all events of suicidal ideation
or behavior were reported in modafinil treatment
groups. Reports of aggression or violent behavior
were relatively balanced between treatment groups,
with a slight higher proportion of these events
occurring in placebo. Additionally, no serious
miscellaneous events were reported in either group.
When the smaller pediatric program in excessive
sleepiness was examined, no psychotic or suicidal
events were found. Obviously, even few events or a
psychotic or suicidal nature are a great concern so
we will review them in more detail.


A total of 5 patients reported psychotic

symptoms while on modafinil treatment, all within
48 hours post last dose. Three of these events
were relatively short in duration and why patients
continued modafinil in one case or following
withdrawal of the drug in two cases. One
additional case, described as psychotic disorder
aggravated, was also relatively short in duration
but did require hospitalization and led to
withdrawal from study. This case, also in the
narrative, we learned reported as suicidal
verbalization but it is included in this table in
the psychotic disorders. The fifth case was an
interesting case of reported ideas of reference
that apparently did not require any specific
treatment--yes, sir?

DR. GOODMAN: We have a question.

DR. PINE: I want to understand both of
these cases because the last two cases don't really
make sense to me and I am wondering if you could go
into them in a little detail, really the last case
more than the second to the last one. When it says
psychotic disorder aggravated, that implies to me

that there was either a preexisting psychotic
disorder or some other factor that was contributing
and it sounds concerning that the child was
hospitalized. So, that is one question.

The second question is that this is a case
of ideas of referential control which, again,
sounds somewhat concerning and the event lasted ten
months, which is also somewhat concerning if those
are really ideas of referential control, but the
action taken was to continue with modafinil. So,
that doesn't make any sense to me. I wondered if
you could explain those situations.

DR. STANKOVIC: Yes, I can talk a bit
about those cases additionally. The psychotic
disorder aggravated is an 8 year-old boy with ADHD.
He presented with severe psychosis beginning on day
19 of the open-label study. He was hospitalized
and at the time of hospitalization we learned that
there was a prior history of a psychotic disorder
that was not reported at the time of the entry to
the study.

The second case is a very interesting case

to us as well. Unfortunately, we do not have quite
a clarification of continuing modafinil treatment
in ten months of continued ideas of reference. We
don't have any additional details. It is
interesting and somewhat confusing but that is what
happened. The investigator continued treatment for
an additional ten months.

DR. PINE: Just to make a comment about
that, I mean, not only does that raise questions
about this case but it raises questions about the
nature of the data in general because it just
wouldn't make sense that somebody would see
something like this, and idea of reference, that
would be ongoing for ten months but not feel the
need to take any treatment. Anyway, I guess it
speaks for itself.


DR. STANKOVIC: We have here a similar
presentation for the 5 patients reporting adverse
events classified as suicidal ideation or behavior.
The first 3 patients experienced brief episodes of
suicidal ideation, described as suicidal statement.

In 1 patient this happened on 2 occasions. None of
these events required either treatment for the
event or study drug discontinuation. One patient
verbalized a suicidal threat which was resolved
after study drug was discontinued.

One case, however, is a case of aggressive
behavior reported initially as normal behavior.
The case narrative described suicidal behavior in a
6 year-old girl with a psychiatric history and
possible family history. The event occurred 2 days
following the last dose of study medication and
required hospitalization and prolonged treatment.


Between the last safety update in
November, 2005 and February, 2006 4 serious adverse
events indicative of suicidality of psychotic
symptoms were reported in the ongoing pediatric
studies. These include both ADHD studies and
ongoing pediatric studies in excessive sleepiness.
Three patients reported events that were classified
as suicidal events, ideation or gesture. In 2 of
these cases no treatment intervention was required

and the patients continued in the study. Treatment
was withdrawn for 1 patient. One additional
patient reported paranoid reaction following 16
days of treatment. The event lasted 5 days and the
study drug was withdrawn.


In the request from the Division for
analysis of psychiatric adverse events, we have
been asked to review postmarketing reports received
during the period January, 2000 to June, 2005. We
estimate that for this particular period the total
pediatric exposure approximates 24,700
patient-treatment years. A total of 7 psychiatric
reports were received during this period.


These are the events reported. The events
were reported in a wide ranges of ages, as you can
see, from 6 to 17 years, and across both genders.
Four events involved psychotic symptoms. One event
was reported as a suicide attempt. However, in
this case modafinil was not taken prior to the
event but was only taken as a part of the cocktail

of drugs used in the multi-drug overdose. The 2
remaining cases are events of aggressive symptoms
and violent behavior.


You may wonder at this time how does the
safety profile of modafinil observed in pediatric
ADHD studies compare to other programs in children.
We have one additional program, smaller, completed
as a part of the pediatric retail request in
narcolepsy and obstructive sleep apnea for
excessive sleepiness. I will review here the
general safety profile observed during the
pediatric development program in this indication.


Overall, a similar safety profile was
observed in the small patient population; a similar
AE profile, effects in vital signs or laboratory
parameters were observed. Notably, no adverse
effects on weight were observed during the
short-term trials in this patient population.
Lower incidence of non-urticarial rash was observed
compared to ADHD studies, and no events led to

discontinuation or were serious in nature.


One serious adverse event from the
pediatric studies in excessive sleepiness requires
discussion as it was mentioned as a point of
concern in the FDA clinical review as a possible
case of Reye's syndrome.

The clinical picture in a 6 year-old boy
was that of a non-specific viral syndrome--nausea,
vomiting, pharyngitis, followed 3-4 days later by a
change in mental status characterized by
somnolence, delirium, hallucinations and seizures.
The patient had elevated serum ammonia but not

The case was reviewed at Cephalon's
request by two external consultants, one pediatric
neurologist and one pediatrician. The consensus
opinion was that the most likely diagnosis was
viral encephalitis or inborn error of metabolism.
Urea cycle disorder was mentioned. Reye's syndrome
was considered unlikely because of normal LFTs.
According to the FDA briefing package, the FDA

consultant also concluded that this case is not

drug related.


Some of the postmarketing information has
been reviewed earlier as part of the discussion on
skin and psychiatric reactions. Here we will
review the profile of the reported events through
our pharmacovigilance system from the perspective
of different system organ classes.


First, review of estimated postmarketing
exposure, we estimate that as of February, 2006
total postmarketing exposure to modafinil was
780,000 patient-treatment years. This includes
worldwide exposure for the period since drug
approval in the first country in 1999. As it
appears, based on the prescription data market
research that we have, 4 percent of these exposures
included individuals less than 18 years of age so
we estimate that the overall pediatric exposure is
about 30,000 patient-years. Based on some
information that we have available, the estimated

median duration of treatment with Provigil in the
market is approximately 3 months. So, using those
numbers, one can estimate exposure to modafinil to
be higher than a million, up to 3 million adults
and in excess of 100,000 children.


Presented on this graph are comparative
profiles of postmarketing adverse drug reactions
reported for adult patients, in blue rectangles,
and pediatric patients, in orange. The total
number of reported adverse drug reactions in a
particular system organ class is presented on the Y
axis while different system organ classes are
presented on the X axis. We had a total of 105
adverse drug reaction reports for all pediatric

As you can see, although it is a little
hard on this slide, the two profiles appear largely
similar across different organ systems. It should
be noted, however, that the we do not have reliable
information on how the two populations relate with
respect to underlying indications for which the

drug is prescribed or doses used.


Based on the postmarketing reports, the
Provigil label is continuously reviewed and updated
as deemed necessary. This slide is a reminder of 3
label changes initiated by Cephalon within the past
3 years. As you can see, some of the safety events
observed in the pediatric ADHD program are fairly
consistent with the postmarketing experience that
resulted in label changes.


I have reviewed a considerable amount of
safety information and will try in the next two
slides to briefly summarize the main points. We
believe that it is fair to say that modafinil is
generally well tolerated at doses studied. Not
unusual for ADHD medication, the most frequently
reported adverse events were insomnia, headache and
anorexia. These events were seldom severe and few
led to treatment discontinuation. Likewise, few
significant laboratory abnormalities were observed.
No effects on mean systolic blood pressure,

diastolic blood pressure, pulse or QTc interval

were observed in the controlled trials.


Beyond initial weight loss, there were no
consistent adverse effects on growth observed over
12-month treatment with modafinil. We did observe
events of suicidal ideation and psychotic events in
the ADHD pediatric patients treated with modafinil.
These events were short in duration in general and
did not require additional treatment in many cases.
We believe that there is one case of probable
Stevens-Johnson syndrome reported in the pediatric
clinical program so far at this point, at an
exposure of 1622 patients. This case resolved
without any adverse sequelae. As I mentioned, I am
sure that there will be more discussion of this
case and we will hear from the investigator and
consultants on this.


In the next presentation Dr. Lesley
Russell will review--

DR. GOODMAN: Before you go to that

presentation, Dr. Rappley?

DR. RAPPLEY: I have two questions; the
first, in the 30,000 children that you expect were
exposed in the postmarketing period, do you expect
that most of them received the 200 mg dose?

DR. STANKOVIC: As I said, it is very
difficult to know exactly what dose was prescribed
and for what indication it was prescribed so I
cannot comment on that. I don't really know. As
Dr. Temple mentioned earlier, one can assume a
variety of things. Whether it was 200 mg or
higher, we don't know.

DR. RAPPLEY: And, in your study 310 it
was cited for not obtaining hematologic values, and
one of the sites was with 21 patients. In looking
at your table on slide 76 which describes
neutrophilia, to what extent did those missed
samples affect your data? How many samples were

DR. STANKOVIC: I can't give you the exact
number; I don't know it off the top of my head, but
I think that the number of analytes may be 390 or

maybe 20 or 30 patients that don't have all of the
analytes, but I am not positive about that. I can
find you that number.

DR. RAPPLEY: Thank you.

DR. GOODMAN: Dr. Leon?

DR. LEON: I would like clarification on
the case control analysis you did. In the sponsor
book it is on page 64-65. You very briefly made
reference to your analyses--


DR. LEON: --in your slides, that you
found no risk of a variety of dosing factors for
the rash. It was a dependent variable. It was a
case control where you had 39 cases and 3 times
that number, 117, controls apparently matched on 3
variables. I have some questions.

First of all, it looks like you entered
about 14 variables that were very highly correlated
simultaneously. I mean, that is what it says here.
The effects were measures of dose entered at one

DR. RAPPLEY: Will you show us the

document you are referring to? Is it this one?

DR. LEON: Yes. Sponsor's book page
64-65. I mean, this is being used as evidence of
no association when the analyses were not conducted
in the way that I believe an association would be
detected. So, my first question has to do with
entering all those variables simultaneously, very
highly correlated measures of dosing.

My second question has to do with what is
the statistical power you would have with this
sample size? You would have statistical power to
detect what effect? The sample size is only 39
versus 117. Would that be an odds ratio of maybe 2
or 2.5? You could miss some pretty substantial

Third, did the analyses account for the
clustering of these sets of 4 who were matched? In
what way did it account for it?

DR. KINGSBURY: Let me address these one
at a time. First of all, let me inform you that we
did not use all 14 variables at a time. This was
just different approaches to explore those and they

were done one at a time. Okay?

First of all, let me describe the matched
control analysis that we did. There were 39 cases.
We found 3 matched controls, as described in the
briefing document. As already indicated they were
matched by the study they came from; by the weight
stratum they were in; and also by having been in
the study at least as long as the time taken for
the event to take place. So, in that set, using
each of those 14 variables one at a time, we looked
at the distribution of whatever the dose was in
quartiles and tried to ascertain whether there was
a relationship, but understanding the limited
power. This is more of an exploratory analysis--

DR. LEON: What was the way that you
accounted for the clustering of these quartets of
case controls there? What was the analysis?

DR. KINGSBURY: I am sorry?

DR. LEON: Well, you have groups of
people, as you would in a paired T-test if you had
diads, and you have sets of 4 people who are
matched on these criteria that you just described,

and I want to know what is the statistical analysis
that was used to account for this clustering, the
correlation among these sets, these quartets.

DR. KINGSBURY: We did essentially a
conditional logistic regression in which we defined
this stratum as the case. We identified each case
and the corresponding matched controls. Then we
looked at the odds ratios of each of the various
increasing quartiles relative to the fist quartile
just to get a sense--I mean, this was very much a
descriptive statistical approach to see if there
was any evidence of a consistent dose response. We
did not find that.

DR. LEON: So, you acknowledge limited
power. You have power here with 150 subjects total
to detect what size odds ratio? Just so you can
let us know the magnitude that might have been
missed there.

DR. KINGSBURY: Because it was not an a
priori designed analysis, we did not focus on that
issue. We actually did not test anything; we were
just obtaining confidence intervals because that is

all we felt would be appropriate. As I mentioned
before, although the confidence intervals
overlapped 1, the odds ratios extended from 0.09 to
a little over 2.

DR. LEON: But when you are looking to see
if confidence intervals are overlapping 1, then you
are doing tests, exactly the same as looking at p
values. You are getting more information as well
about the magnitude of the change and about the
variability of that change, that association.

DR. KINGSBURY: We don't claim to have
shown no association. All the conclusion we are
making is that--by the way, consistent with the
limitation in the numbers that we have no
compelling evidence of an association, we did an
additional analysis based on the randomized
clinical trial data, and from that analysis we
found an odds ratio of 1.4 with a confidence
interval extending from 0.678 to 3.094. Going back
to the case control analysis--

DR. GOODMAN: Thank you very much. Let's
go on to our next speaker. Thank you.

DR. STANKOVIC: The next speaker is Dr.
Lesley Russell.
Benefit-Risk Conclusions

DR. RUSSELL: Thank you, Dr. Stankovic.
We have presented a lot of information this morning
regarding the efficacy and safety profile of
modafinil in the treatment of ADHD.


Following your deliberations, you will be
asked to answer two questions, the first being has
modafinil been shown to be effective for the
treatment of ADHD in children and adolescents?

We believe that the answer to this first
question is yes. In the 3 pivotal studies
consistent benefit of treatment with modafinil was
seen in all 3 studies, with these effects observed
by the teacher, the parent and the treating
physician across different rating scales and
instruments, and with effects being observed both
at home and at school.


You are also going to be asked today
whether modafinil has been shown to be acceptably
safe for the treatment of ADHD in children and

In the Phase 3 clinical program modafinil
was generally well tolerated. The most common
adverse events reported, insomnia and anorexia,
were generally mild to moderate in severity and
rarely a cause for treatment discontinuation. No
adverse signals were observed in the Phase 3
program with respect to pulse, blood pressure or

We were asked in the approvable letter to
provide more information on 3 cases of liver
transaminase elevations. As outlined in our
response to the approvable letter and presented
here today, in 2 of these cases the transaminase
elevations were resolving on continued treatment
with modafinil with, in 1 case, ALT values
returning to normal whilst continuing treatment.
In the third case the transaminase levels were
returning to normal on discontinuation of

treatment. We do not believe that an adverse
signal with respect to liver function has been

Concerns have been raised over the
reporting of psychiatric adverse events. As you
are aware, these events were fully discussed
yesterday at the Pediatric Advisory Committee for
all ADHD products. Although no consensus was
reached on how to label aggression, psychosis,
mania and suicidality, Cephalon has proposed
language in the warning section of the label which
we believe provides appropriate information
regarding these events seen in our clinical


Concerns have also been raised over the
reporting of serious skin reactions, and in the
approvable letter we were asked to provide you with
more information on 3 cases of interest seen in the
clinical trials and 4 cases reported in adults in
the postmarketing setting.

As suggested by FDA, these cases were

reviewed by experts in the field and there appeared
to be general concurrence reached by these reviews
and Dr. Porres, from the FDA, with respect to the
first case, the 7 year-old boy with possible SJS.
But there does appear to be some diversity of
opinion regarding the other 2 clinical trial cases.
This seems to be in keeping with the diagnostic and
etiologic uncertainty surrounding the diagnosis of
these types of skin reactions. However, we
acknowledge that an association with modafinil
cannot fully be excluded. In all 3 of these cases,
however, the events did abate following
discontinuation of drug and no adverse sequelae

In assessing the risk for SJS and
reviewing the totality of the data in the clinical
trials and postmarketing database for both adults
and children as reviewed, we believe that the risk
for SJS is low. However, we have proposed language
to be included in the warning section of the label.
Based on your deliberations today, we will be happy
to modify this as appropriate in order to provide

patients and healthcare providers with adequate
information concerning these events.

Lastly, modafinil is not a new chemical
entity and to date there have been 780,000
patient-years of exposure which, when looking at
actual patients exposed, may equate to
approximately 3 million exposures since
introduction of the drug in France, in 1994.

Pharmacovigilance is undertaken to assess
risks associated with modafinil usage and, as you
have heard today, this has led to 3 labeling
changes, one regarding the incidence of severe skin
reactions. Cephalon is committed to improving
these risk assessments further by undertaking a
more structured case ascertainment with respect to
skin adverse events.


So, in conclusion, we believe we have
shown you today that modafinil is an effective
treatment for ADHD with an acceptable safety
profile, with the benefits of treatment outweighing
its risks. Thank you for your attention.

Questions from the Committee to FDA and Sponsor

DR. GOODMAN: Thank you. May I suggest
that your team stay at the podium to address some
questions? I am going to assume that most of the
committee members are going to have questions for
you. If we start to run out of time, we are going
to have more opportunity to ask those questions
later this afternoon.

Let me start off with what may be the
easier of the two questions we are asked to vote on
today, the one regarding efficacy. From the FDA
standpoint and what I read, they were satisfied
with the efficacy data. I certainly feel satisfied
from what I have seen. Yet, before we move on to
the harder question of evaluating issues of safety,
it is very important to have the context in mind of
the benefit.

So, I want to give you an opportunity to
answer, from your perspective, where you see this
medication fitting in; where is it going to add
value or options in the marketplace? Is it going
to be advantages in the area of efficacy,

tolerability? I wonder if you could just expand on
those issues to give a little bit of a framework to
think about the benefits of this medication.

DR. RUSSELL: Well, as you heard from Dr.
Biederman with the MTS study, despite treatment
with drugs that are considered to be very
effective--and we certainly don't doubt that--there
does remain a group of patients that still either
cannot tolerate drugs or don't respond to them. We
saw in our program that, although maybe not
considered refractory, patients who had failed on a
prior stimulant therapy did appear to benefit from
the drug. We also saw that if you are stimulant
naive you respond slightly better to the drug.

So, we see this as a viable treatment
alternative to other drugs that are obviously
commonly used and considered to be effective
agents. However, I would like to have a treating
physician in the field come up and maybe give you
that from his perspective. So, if I could ask Dr.

DR. GOODMAN: Sure, go ahead.

DR. BIEDERMAN: I think that in clinical
practice we need alternative treatments to treat
our patients. The idea that the most efficacious
treatment treats all our patients is not true to
life. So, clinicians in practice need to have
options to allow us to better serve the people that
consult with us.

The issue of adverse effects is a
statistical issue. That means that even if side
effects are similar within a class of drugs, some
patients clearly tolerate one versus another even
if on average they have a similar spectrum of
adverse effects. So, patients that have poor
tolerability may benefit from a drug that may have
on average similar issues but may be better
tolerated for them.

Finally, the issue of scheduling--I think
that even though many of the new generation
stimulants that are available today are clearly
less of an issue for diversion and abuse, many
clinicians and many families do not want their
children to be on a Schedule II drug. So, I think

that this gives an option for clinicians to use a
lesser scheduled drug in cases where they choose
not to use a scheduled compound.

DR. GOODMAN: Joe, before you step down,
has it been your impression so far that there is
less abuse potential, less potential for diversion
as, say, compared to stimulants?

DR. BIEDERMAN: Yes. I am not an expert
on abuse and we have here a colleague that
specializes in that. The abuse and
diversion--first of all, let me comment on abuse
and diversion. There are different publics that
use these drugs recreationally and therapeutically.
Our battles in clinical practice are to encourage
our patients to remain in treatment. There is a
very severe problem of non-adherence to these
treatments. So, it is not something that our
patients look forward to taking.

The attraction of the stimulants is when
the tablet can be crushed and snorted for an
IV-like experience. It is the parenteral intake
that produces the euphoria, not the oral intake.

So, this drug is not snortable, injectable, and so
on and so forth, so it is not a drug that the
addict community on the street would pay a high
price for to get it. But maybe we can get some of
our colleagues that are here with better expertise
than mine on diversion and abuse to give a

DR. RUSSELL: Does that answer your

DR. GOODMAN: I would like to hear a
little more on that issue.

DR. RUSSELL: Dr. Dackis?

DR. DACKIS: With regard to the abuse
potential of modafinil, I think it is important to
note that it is chemically unrelated to central
stimulants and has a very weak effect on the
dopamine transporter so that it is extremely
unlikely to increase dopamine levels, except in
very high dosages.

There have also been a number of studies
in humans to assess what the subjective effects of
this agent are and these studies, which have been

conducted by Jasinski demonstrate that in males
there is no effect of modafinil. There was a
smaller study in females that did show some
stimulant effects using these various rating
scales. Two other studies, again, showed that
there was not a significant high; that the subjects
were not willing to pay money for modafinil, etc.

In addition, animal studies, looking at
things like self-administration and condition-place
preference showed very weak stimulant-like effect
of this agent. So, there is some reinforcing
quality but it is very, very weak. DR. GOODMAN: I
thought monkey studies showed preference.

DR. DACKIS: Yes, that is correct. Gold
and Balster's study did show that monkeys, trained
to self-administer cocaine, if given modafinil
would continue to self-administer large doses of
this agent, as they would with other compounds like
ephedrine. So, large doses are required to
continue to self-administer.

DR. GOODMAN: Thank you. Dr. Temple?

DR. TEMPLE: I am sympathetic to the idea

that drugs with different pharmacology may have
different usefulness, but I want to address the
question of whether they have documented the
ability of this drug to work in people who are
resistant to stimulant drugs, and the answer is
that they have not.

There is a perfectly simple, never done
kind of study design to do that. You take people
who fail on whatever it is you want to test and
then you randomize back to that drug and to the new
drug. It is a perfectly simple study. That is how
clozapine came to the market because we wouldn't
have approved clozapine unless it worked in
failures because of the 1.5 percent
agranulocytosis. That study could be done. You
might even think about whether it is something that
ought to be done, but it has not been done. The
mere fact that people given a second drug after
failing the first respond to it tells you nothing
at all. We have many examples where drugs don't
particularly work in non-responders to other
therapy but the second time around the people do

better. So, I just want to make it clear they have
not shown that. It might be true. It is plausible
even but it hasn't been shown.

DR. PINE: Can I ask a question about
that? Of course, there have been other medications
that have been discussed over the last couple of
years for new indications for ADHD and I am sure
that that issue came up. I think that those
studies have not been done and what was the
thinking and discussion around that?

DR. TEMPLE: Well, they are almost never
done. We don't usually have a reason to say, for
example, only use this drug in people who have
failed on other therapy, if one thought that was an
appropriate thing because I am not saying you
should or not--you are going to get to that. I am
just making the point that they have not documented
in a rigorous way that the drug would actually work
in those people. You might think that there is a
little evidence that it does, and you might think
the pharmacology difference suggests that it might,
all of which I agree with but that hasn't been

studied and it can be studied, and it never is

DR. PINE: For what it is worth, my take
on it would be that that would only be one of the
potential uses of the medicine clinically, and it
seems like some of the other issues are, you know,
kind of bigger in terms of thinking about the
medicine as opposed to, you know, is it primarily
for people who don't respond to stimulants.

DR. REESE: We are going to get to
everyone's questions. First we are going to have
Dr. Bronstein and then we will have Dr. Wang.
Thank you.

MS. BRONSTEIN: My question is a fairly
straightforward, easy one. On slide 93, in the
Phase 3 study you have one person who had a severe
event and withdrew from the study. What kind of
rash was this?

DR. RUSSELL: Unfortunately, the only
description on the case report form, which reflects
the source documents, is just a verbatim of rash so
I am unable to describe it further for you.

MS. BRONSTEIN: We can assume though that
it was severe.

DR. RUSSELL: It certainly led to
discontinuation of the drug. That is all the
information I can give you.

MS. BRONSTEIN: Thank you.

DR. REESE: Ms. Dokken?

MS. DOKKEN: Yes, I apologize, I thought
we were supposed to hold our questions until the
end so my question really goes back to slides 30
and 31 and this issue of the 40 percent who are
non-responders or had intolerable side effects. I
am wondering whether anyone can sort of unpack, you
know, how many people are in which category because
it seems to me that what we have been hearing is
that one of the marketing messages for modafinil
will be that it is an alternative. If it is an
alternative and we are talking about whatever
percentage of that 40 percent are ones who suffered
"intolerable" side effects, certainly this
particular drug--and those of us who were fortunate
or unfortunate enough to be present yesterday, you

know, the side effects are present in almost all.
Then that leads me to the worry about the next step
which is, you know, if it were approved how is it
marketed and what are the messages because probably
it was the Pediatric Advisory Committee that has
seen, you know, other situations where something is
marketed as being free of something else,
suggesting that there are no risks and to say that
because it is a non-stimulant it has no risk would
be a concern for me.

DR. TEMPLE: Drug advertising reports to
me so I have to worry about this. We are fairly
careful about making claims when you don't have a
direct comparison and there aren't any direct
comparisons. However, if one is scheduled at a
different place, or something like that, that is
true and they would be allowed to claim that.

There are some cases in which the
difference in certain side effects is so
obvious--like it never happens with this and it
happens all the time--where we might allow
something like that. But we are very careful about

comparisons in the absence of actual comparative
data across study comparisons and treat it with

DR. REESE: Dr. Pfeffer?

DR. PFEFFER: Thank you. I am not
questioning the efficacy but I have some questions
on slides 52, 53 and 54, please. Maybe you can
help us understand the longitudinal process of the
three studies. For example, it looks as if in
slide 52 I guess efficacy was being demonstrated by
week 5. Then in slide 53 and 54 it seems that it
was earlier, although on slide 53 at week 5 there
was perhaps less of that. I don't know if that is
due to dropouts and then resumption.

So, my question is on the early phase of
these, week 3 and even week 2 on slide 54, what
were the general doses that the children were on at
that point in time? Then, if you can tell us what
happened in week 5, on slide 53? Finally, if you
could tell us a little bit about when were blood
tests taken in the process of the study and when
did the side effects emerge, especially skin

reactions, etc.? I am trying to link the time
course with the doses and the longitudinal course.

DR. RUSSELL: In study 10, which is the
slide up here, this is the fixed dose study so that
by the second week patients would have been
titrated to that target dose. That would have
occurred by day 7 for those randomized to 340 and
day 9 respectively.

DR. PFEFFER: I thought I understood that
but my concern is if, in slide 53 and 54, you see
earlier efficacy is that at the target dose or less
than the target dose?

DR. RUSSELL: In this study, which is the
fixed dose study, they would have been at target

Could you go back to the previous slide
for 311, please? This is one of the flexible dose
titration studies. So, in the earlier weeks they
would have still been titrating up.

DR. PFEFFER: Do you know approximately
the average doses at the early phase?

DR. RUSSELL: Probably around 255 mg by

the second week and up to the 340 mg by the fourth

DR. PFEFFER: And on slide 54 it is
similar. Is that right?
  DR. RUSSELL: Slide 54, which I think is

study 310, is where they titrated up more quickly
so they would have been at target dose by day 7 and
9 respectively.

DR. REESE: Dr. Armenteros and then Dr.

DR. ARMENTEROS: Just to follow-up a
little bit on the dosing question, I understand the
model that you used to dose the two groups of
children, you know, below 30 kg and above. Now,
most of the children that got into the study were
above 30 kg, like 68 percent that you mention here.
Now, when you presented data on efficacy there
wasn't a differential response between these two
groups by weight.

The reason that I ask that question is
that we already know from your previous studies
that at lower doses you do get response for daytime

sleepiness, and so forth. So, I don't know if we
may be missing perhaps different points in dosing
at which these kids may respond. Because at the
end of the trial I come out with a very fuzzy
impression of what the actual dosing should be and
I hope I can get a better understanding.

DR. RUSSELL: First let me answer the
excessive sleepiness programs first because what we
do find is a very different pharmacodynamic
response when we are treating excessive sleepiness
than when we are treating ADHD. So, in the
excessive sleepiness programs and the pediatric
narcolepsy, although we looked at doses of 100 mg
through 400 mg, doses of 400 mg were clearly
efficacious in that model. Then we did some PK/PD
work and the target exposure needed for an effect
in narcolepsy is substantially lower than the
target plasma exposure associated with effect in
ADHD--so very different pharmacodynamic response
which I don't think I can explain, but it is very

In terms of looking at the doses and how

did they respond to efficacy, what we did was to
look at the different quartiles of dosing and in
the third and fourth dosing quartiles, which are
the higher dose groups, you see numerically a
slightly higher response but it is only a point or
two. So, I would say that the dose response, with
all the caveats because we were titrating to a
target dose, is flat in the doses that we looked at

DR. REESE: Dr. Malone?

DR. MALONE: I have two questions. One is
on efficacy. The stimulants wear off every day by
the end of the day. Is that true for this drug? I
am just wondering if it is like the stimulants,
that you have to dose it every day; you dose it in
the morning and then it wears off by the evening.

DR. RUSSELL: The only data we have with
respect to that is actually in the 2-week
withdrawal period where the patients who had
received modafinil during the double-blind
treatment period were randomized to either stay on
modafinil or were randomized to receive placebo.

What we see is not an immediate return to baseline
in symptoms but a more gradual return towards
baseline and their symptoms. So, based on the
limitations of that data which I acknowledge here,
there doesn't appear to be a sort of complete
rebound effect.

DR. REESE: Dr. Bigby?

DR. BIGBY: I have a question about the
ADHD rating scale. If you gave this test to a
group of normal kids who don't have ADHD, what
would their score be?

DR. RUSSELL: The average for a 10
year-old boy I think is 18.8, and the children
going into our study had an average of around 37.
So, they were clearly much higher than what would
be considered to be normative for a 10 year-old
boy, which was the average population in our study.
It does differ a little bit based on whether you
are a boy or a girl or your age, but that appears
to be the average for a 10 year-old boy.

DR. GOODMAN: Dr. Temple?

DR. TEMPLE: In one of the studies you

actually did a withdrawal phase but I believe the

data weren't shown.

DR. RUSSELL: That is right.

DR. TEMPLE: You must have a slide of it.
That would answer the question of how soon it wears

  DR. RUSSELL: If I could have the slide,

This is over the 2-week withdrawal period.
You can see on the right-hand side that the placebo
at the end of the 7-week period and the end of the
9-week period obviously stays the same. In the
modafinil group there is a point difference, but
for those who were on modafinil and then got
changed to placebo you can see that there is a
beginning of deterioration of their symptoms over
that 2-week period. It is not huge but there is a
deterioration and it looks like they are returning
towards baseline. But there doesn't appear to be a
sort of instantaneous effect.

DR. TEMPLE: And you don't have it day by

day or anything like that?

DR. RUSSELL: Unfortunately, we don't.

DR. REESE: Dr. Malone, your second
question and then Dr. Rappley.

DR. MALONE: It was really I guess partly
answered. It had to do with the abuse potential
for modafinil. I think, from the reading, it did
say that it can cause euphoria and that animals
would work for this drug. If that is true, I just
have a question why would a stimulant be a Class II
and this a Class IV? How do they decide that?

DR. GOODMAN: Dr. Temple or Dr. Laughren?
I have a very similar question about the
classification. Currently this drug is classified
Schedule IV compared to the stimulants which are
Schedule II. Could you just explain that
distinction? It would be in the context of a quick
follow-up I was going to do and ask sponsor how
they would best characterize or classify their

DR. LAUGHREN: Actually, FDA doesn't
decide that classification. The decision is made

by the Drug Enforcement Administration. They do an
8-factor analysis. I haven't looked at that.
Maybe the company could respond to, you know, how
it is that the DEA arrived at a Class IV rather
than a Class II.

DR. TEMPLE: There is a very sharp
distinction between the level of control. I think
we are about to hear about that. II is, you know,
locked cabinets and all the rest; IV is much less.

DR. GOODMAN: Yes, please, could we hear
about that?  
DR. RUSSELL: The difference between a

Schedule II and a Schedule IV, is that what you are

DR. LAUGHREN: How it got a Schedule IV
rather than a Schedule II.

DR. RUSSELL: I wasn't with the
organization at the time of the original
scheduling. Perhaps I could ask Dr. Vaught, who
was here, to explain how that happened.

DR. VAUGHT: Good morning. My name is Dr.
Jeff Vaught, executive vice president for research

and development for Cephalon.


I would like to very briefly just go over
the aspects of scheduling which, certainly the
agency knows as well as I do, has to do not only
with the physical chemical characteristics of the
compound but also testing that is done in human
beings to suggest that there is a reinforcing
property. So, if we look at the overall physical
chemical activity of modafinil, it has very, very
low water solubility which is incompatible with
intravenous injection. It is very unstable at high
temperature, therefore, it is incompatible with
smoking. Importantly, it is structurally unrelated
to other agents that are known to be abused. While
it does have a very, very weak--and it is really
the only neurochemical effect that we have been
able to demonstrate in blood receptor binding
assays, etc.--with dopamine. It doesn't appear to
cause elevations of dopamine of nucleus accumbens,
which is markedly related to drugs of abuse, as
well, it has not releasing properties as do other

Schedule II stimulants. There is also lack of
activation, as I mentioned, of reward centers, and
really the results, as Dr. Dackis described to you,
from preclinical studies suggest that if there is a
signal it is very, very weak.

Now, all this is theoretical because that
is all nonclinical data. Perhaps more importantly
and something that we undertook at Cephalon
spontaneously, is a postmarketing surveillance,
starting in 1999 with the Haight Ashbury group.
The Haight Ashbury group monitors a variety of
areas worldwide where drugs may be diverted to,
including rave scenes, medical professionals, etc.
Now that we have had six years we still have
reporting on this. There have been limited to no
reports of euphoric effects. There are no reports
of reinforcing effects. There has been a very
large increase since the drug has been approved for
wakefulness for mainstream publicity regarding the
use of modafinil, including in The New Yorker
magazine, college newspapers, etc., and across the
Internet every now and then we will see postings of

potential use but nothing that is consistent. In
fact, the Haight Ashbury concludes after evaluating
this for the last six or seven years that if there
is abuse potential for modafinil at all, it is
very, very low.

So, all of this is consistent with what is
seen as an agent with low abuse potential. We now
have considerably more experience with the
substance than we did five or six years ago when we
were getting approval and we thought that was
consistent with the regulatory standards for
Schedule IV.

DR. GOODMAN: Thank you. Apart from how
DEA will classify your drug, how would you
internally classify it? Would you say it is a
stimulant or is it distinct based upon its
mechanism of action, which I understand is unknown.
Although I know at one time it was thought to be
mediated through orexin receptors, I guess that is
not as firmly established at this point.

The reason I ask is not just a semantic
question but whether it gets counted or considered

a stimulant may have labeling implications. As
revealed by discussions yesterday, for example,
Strattera, should that be considered a stimulant
and, if so, should it have certain warnings
attached to it that go with the rest of the class
of stimulants? So, I would just like the sponsor's
perspective on whether you would classify this
medication as a stimulant or not.

DR. VAUGHT: We approach this from a
couple of levels. One is the preclinical data that
we have, as well as the clinical information. In
direct answer to your question, I would not
classify it as a traditional sympathomimetic
stimulant. It is a CNS activating agent and we
have all been taught, prior to the introduction of
modafinil, that, in fact, most of our CNS
activators are psychostimulants. Nonclinically,
modafinil has a profile of wake-promoting activity
that, unlike the classical stimulants--its
wake-promoting activities are not blocked by
haloperidol which has been characteristic of wake,
if you will.

As far as the orexin component that is
involved, we have been able to demonstrate it has
no interaction with the orexin system because in
knock-out animals, as well as human beings and dogs
it is highly effective. When we move to human
beings, we similarly don't see the typical types of
profile that one sees with the stimulant
population. If we include this with
methylphenidate and amphetamines this includes
sympathomimetic-like effects as well as generalized
excitation reinforcing properties, euphoric
effects, etc. So, overall the pharmacology would
suggest that if we want to classify it as CNS
activating agent it is certainly a non-traditional

DR. GOODMAN: Would you say that it has
less peripheral--if you look at the relationship
between CNS, there is relatively more CNS to
peripheral activation?


DR. GOODMAN: That was my last question.

DR. REESE: Dr. Rappley?

DR. RAPPLEY: My question goes back to the

safety area and pharmacodynamics. Dr. Mannheim
noted that we don't have information about steady
state for the sulfone metabolite. We know it
accumulates to a much greater extent in children
but we don't know quite when that steady state is
achieved and I wonder if you have more information
about that.

DR. RUSSELL: The sulfone metabolite
appears to reach steady state at about 2 weeks and
then it actually plateaus thereafter.

DR. REESE: Dr. Wang?

DR. WANG: I have one more housekeeping
question about efficacy. Are these effect sizes
and response rates--I guess this is either for the
sponsor or maybe our pediatric colleagues--are
these response rates comparable to what is seen
with other treatments for ADHD or is there some
differential response here?

DR. RUSSELL: Dr. Biederman?

DR. BIEDERMAN: I believe that the
computed effect size is about 0.7, very similar to

the effect size of Strattera; lower than the effect
sizes of the stimulants that are about 0.9. So, it
is lower than the stimulants but potent enough to
treat ADHD.

DR. WANG: Then this is actually a
question for the FDA. The sponsor is already
proposing warning language and I am curious what
are the potential actions you can take. I mean,
bolded warning; black box warning? Are those the
same thing? Are there other intermediate warning
language actions you can take? Because the sponsor
is already proposing potential language.

DR. ANDREASON: I am sorry, I missed the
first part of your question.

DR. WANG: Firstly, I should know this but
is there a difference between bolded warning
language and a black box warning? And, are there
intermediates between them and what other options
are there?

DR. ANDREASON: Yes, they are different.
Which adverse event are we talking about here?

DR. WANG: They are already proposing

language for, it sounds like, psychiatric adverse
events and also for skin rash.

DR. LAUGHREN: Well, there is a difference
between bolded language and unbolded language in
warnings. I mean, sometimes if we want to give
particular emphasis to something we will bold it.
That is different than a box. A box goes as the
first thing in labeling and it is surrounded by a
box. So, that is very different than just bolding
language in warnings. So, there is a continuum.

DR. TEMPLE: In the context of the CNS
warnings, you need to think about it in the setting
of the consideration of all of the drugs that went
on yesterday, and so on. The skin is their own
baby. So, if we were very worried about it we
could put it in a box. Usually you put things in a
box when you want to be very sure that the doctor
absolutely, positively considers this before
prescribing it.

There are other things you can do.
Ziprasidone, because of the QT prolongation, says
you really think should think about using other

drugs before you do this. You can go further, you
can say this is absolutely only for people who fail
other therapy. Sometimes we do that even if we
don't know for sure, as I said earlier, that it
absolutely works in people who failed other
therapy. You know, because of its different
properties, you assume there might be a population
that responds that way. There are a variety of
things you can do to try to direct therapy. We
like to say we don't practice medicine but we do
sometimes try to influence the way a drug is used
if we are worried about its safety. The black box
is the loudest statement. There is at least a
perception that it affects use because it scares
people. That is why some people like it and some
people don't like it. Bolding is more prominent
than non-bolding, and so on.

DR. REESE: Dr. Leon?

DR. LEON: Dr. Biederman, I would like to
clarify what you said about the Strattera effect
relative to what we saw in this trial. It is my
understanding that the Strattera effect size was

about 0.80 and in these data it was 0.56. So, this
is quite a bit smaller. It is still a minor effect
size but it is not as large as was seen in the
Strattera trial.

DR. BIEDERMAN: To my knowledge, and I
don't remember those numbers by memory, but I think
between 0.6 to 0.7 is the effect size of modafinil.
The company may have that information better than
me. I understand as well that the effect size of
Strattera on average is very much similar at about

0.7. In the meta-analysis of non-stimulants that
Dr. Faraone did a relatively short time ago, that
is shared by other non-stimulants as well, like
tricyclics and things of that type, on the order of
magnitude of 0.7, a low effect size of stimulants
at about 0.9.

DR. REESE: Dr. Pine?

DR. PINE: I would like to go to slide
number 89. I guess the thing that I am struggling
with most, and I think a lot of people might be, is
the dermatologic issue. On the one hand, I don't
want to start a fight but, on the other hand, I

guess I am struggling a little bit with some of the
inconsistencies in terms of the way three of the
cases on slide 89 are being discussed. So, I guess
what I want to do is point out what I see as the
inconsistencies and then maybe hear from Dr. Bigby
about do I have it right; do I not have it right;
and then maybe also try to clarify some of those

So, the way that I heard it is that case
number 1 or patient number 1 everybody agrees had
Stevens-Johnson but there is disagreement about the
etiology, I heard, which confuses me a little bit
because I don't understand what the etiology
possibly could have been except for the medicine or
except for the modafinil. So, I would like to hear
discussion about that.

For patient number 4, at least what I
heard was that Dr. Bigby did think it was
Stevens-Johnson and I heard that two out of the
three experts at Cephalon thought it was at least
possible Stevens-Johnson. So, at least in the way
I am thinking about it, I would think of those as

two at least likely cases.

Then, for patient number 5 I am a little
blurry in terms of the magnitude of concern as a
non-dermatological clinician. If I see a possibly
suggestive hypersensitivity reaction or whatever
Dr. Bigby classified it as, is that equally
concerning, or slightly less concerning, or how
much less concerning than Stevens-Johnson?

So, do we have three cases where everybody
would agree that these are concerning dermatologic
issues? Do we have one case? Do we have two that
are somewhat concerning and a third that is
suggestive? You know, can we get some agreement on

DR. REESE: Dr. Bigby?

DR. BIGBY: What I would say is that case
number 62338 is a case of Stevens-Johnson syndrome
and, based on the information that is provided, I
would say it is drug related.

DR. PINE: What about the other two? For
case 18004 would you also say that? And, what is
the disagreement?

DR. BIGBY: I would say that that case is
more likely to be due to something else other than
a drug. So, I don't actually count that as a
drug-related case. What was the third one?

DR. PINE: The third one was case 056003.
You said fever, urticaria, swollen eyes, vomiting,
increased ALT/AST and the Cephalon review said
possibly suggestive of a hypersensitivity
reaction--I guess level of clinical concern in
terms of a serious adverse effect related to the

DR. BIGBY: You are going to have to give
me a little time for that one.

DR. PINE: Okay. Dr. Goodman is
whispering in my ear that he wants to know what
made you conclude on case number 18004 that it was
not medication related.

DR. BIGBY: For that case it is just not
so clear to me what the diagnosis is. I mean, it
is hard in sort of spottedly reported case reports
to figure things out and I just am not convinced
that that is a drug rash at all.

DR. PINE: Then I guess the last question,
when I asked you before about your level of concern
you seemed fairly clear that there is, quote, a
signal here in terms of dermatologic risk. Based
on what you just said, my conclusion would be that
you are basing it on this one confirmed case out of


DR. BIGBY: Plus, there is a signal for
exanthems. Those aren't serious reactions but
there is also a signal of exanthems occurring with
the drug.

DR. PINE: But I also understood you to
say that there is not necessarily a relationship
between exanthems and incidence of Stevens-Johnson.

DR. BIGBY: This is correct.

DR. PINE: So, again, I guess what I am
hearing is that it is really the one case out of
the 923.

DR. BIGBY: I think that that is a good
summary of how I feel about it.
DR. RAPPLEY: But there are also the four
cases in adults. Is that correct?

DR. PINE: I think those were in adults.

DR. RAPPLEY: That is right, in adults.

DR. PINE: And it was consistent with the
base rate. When we looked at the patient-years
exposure it was consistent with the base rate of
Stevens-Johnson syndrome, the four adults.

DR. REESE: Dr. Robinson?

DR. ROBINSON: Could we go to slide 112?
I just want to clarify a few things because in Dr.
Andreason's presentation it said that we were
finding some dermatologic signal within the
clinical trials but not in the postmarketing, and I
just want to clarify a few things on this slide.

In the pediatric subgroup you didn't find
a signal for rash in the postmarketing. Is that

DR. RUSSELL: Certainly, in the
postmarketing setting in children we have had no
reports of any serious skin reactions. That would
be correct.

DR. ROBINSON: Okay. Then, one of the
questions about that is, is that because there is

none or is it that you are unable to detect that?
So, that is why I would like to ask a question
about the psychiatric signal that you do have in
the pediatric subgroup because in the clinical
trials it seemed that there is some signal about
suicide and psychosis, and in the postmarketing
data for pediatrics were you picking up that

DR. RUSSELL: In the postmarketing data we
saw seven cases that Dr. Stankovic highlighted for

DR. ROBINSON: That was in pediatrics?

DR. RUSSELL: That was in pediatrics, yes.

DR. ROBINSON: And it was which ones?
Psychosis or suicide?

DR. RUSSELL: If I remember right, there
were three psychosis, one suicidal ideation.
Perhaps you can clarify?

DR. STANKOVIC: There were four cases of
psychotic symptoms. There were two cases of
aggression and violent behavior and there was one
case of a suicide attempt. That was the patient

that overdosed and used modafinil as one of the
cocktail drugs but it was not modafinil prior to
the event.

DR. RUSSELL: Thank you for clarifying.

DR. ROBINSON: Thank you.

DR. REESE: Dr. Temple?

DR. TEMPLE: Back to derm., I think it
would be helpful to be clear on what the
appropriate denominator is because there seems to
be one case everybody agrees on. This 933 number
that has been used includes some very short
exposures. Dr. Bigby can tell us what kind of
exposure is enough, but let's say we wanted to say
how many of those 933 or some of the people from
the other studies had, say, at least two weeks or
whatever the right amount is. That would help.
Maybe it doesn't matter whether it is one out of
900 or one out of 600 but it would be good to have
a number. So, how many people who were on it long
enough to have had a nasty skin reaction actually
were there for that one to be the numerator for?

DR. BIGBY: That is a very good question.

You know, I think that the best data available
about the window of exposure where TEN/SJS is going
to occur comes from that study that I cited. It
was sort of a consensus panel in three countries,
and the majority of cases occur within the first
one to four weeks. It is probably one to three
weeks. And, if you sort of include in your
denominator patients that have been on it steadily
for months and months and months you actually
probably come up with a lower rate than the actual
because the time that you are going to get it in is
in that first month.

DR. REESE: We can have the response and
then Dr. Armenteros.

DR. RUSSELL: I can get Dr. Shear to come
up and comment on these cases with respect to
etiology and all the other aspects we have been

DR. SHEAR: Thank you very much. From a
dermatologic point of view from somebody who has
been doing this for 20 years, first of all, I would
like to thank Dr. Bigby for his excellent

presentation because I agree with what he said and
this is an area that has been really messy over the
years and you can see the confusion that led us

So, I would really focus on that one case
of Stevens-Johnson syndrome. Going through that
case extensively--the panel went through it but I
also went through it with the panel again, with Amy
Paller who was the leader of the panel--to try and
figure out exactly what was going on with that case
and how we could best characterize it.

I think we see enough to call it either
Stevens-Johnson syndrome or maybe erythema
multiforme major. You could then argue about which
it is, and does it really matter since both of
those can be viral induced? Speaking with the
investigator and looking through the case records,
there were clear viral-looking lesions that
suggested Coxsackie very highly in the pharynx
prior to the patient getting this. The clinical
course was very compatible with a viral-induced
either erythema multiforme major or Stevens-Johnson

syndrome because actually the patient was not that
sick and was able to continue going to school and
continue with other activities. Part of the
problem was getting the full history, and much of
it was retrospective and there was a language
barrier, but the patient wasn't sick enough to be
admitted to hospital or really to be seen very
carefully during the actual event. But still,
piecing it together, I would certainly put viral
etiology well within the mix. I don't know what
percent I would give it but, you know, drug is in
there and virus is in there so it is not a
completely clear case of either Stevens-Johnson
syndrome nor is it a completely clear case that it
was drug induced.

DR. GOODMAN: Dr. Bigby, would you concur?

DR. BIGBY: I think the patient had SJS.

DR. PINE: But the suggestion is that it
could potentially have been Coxsackie virus induced
SJS, which would be a very different thing. Again,
I mean I get the impression that you do not think
that that is likely.

DR. BIGBY: You know, it is really

impossible, never having seen a patient, to do
this. I don't think you should call things EM if
the patient doesn't have typical targets. There is
no description--the data is inadequate to be very
dogmatic or firm about this. I mean, I would say
that none of the dermatologists involved here would
go out and have a big fight about what this case is
because the description is just not good enough.

DR. SHEAR: Yes, I should mention that in
one of the papers it did describe target lesions.
So, that was helpful but, again, there are all
these bits and pieces in trying to look at the
source documents. From the source all the way to
the narrative, you get different bits and pieces.
Some are quite extensive. The MedWatch report has
different data, but piecing it all together, there
is uncertainty but it is in that EM major and
Stevens-Johnson spectrum that overlap, if you will.

DR. GOODMAN: Hold your questions. We are
going to break for lunch and come back at one
o'clock. We will have the public hearing component

at that time.

[Whereupon, at 12:30 p.m., the proceedings
were recessed for lunch, to reconvene at 1:00 p.m.]


Open Public Hearing

DR. GOODMAN: We are going to begin the
afternoon proceedings. I am going to ask Dr. Pine
to read the description of the process for the
benefit of the individuals who are presenting at
the public hearing segment of today's proceedings.

DR. PINE: Both the Food and Drug
Administration 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
committee meeting, the FDA believes that it is
important to understand the context of an
individual's presentation.

For this reason, the 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 any company or any group that is
likely to be impacted by the topic of this meeting.
For example, this financial information may include

a company's or a group'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 such
financial relationships. If you choose not to
address this issue of financial relationships at
the beginning of your statement, it will not
preclude you from speaking.

DR. REESE: We will have the first
speaker, who will have five minutes and when there
is one minute remaining we will let you know your

DR. RAVENEL: Dr. Ravenel. As a
pediatrician with 36 years experience--by the way,
I have no financial disclosure; no connections; no
funding. I am here independently as a private
practitioner. With 36 years combined experience in
academic and private practice with a heavy emphasis
on behavioral pediatrics, I want to share some
concerns with the committee concerning the pending
new indication for Cephalon's modafinil, to be

marketed as Sparlon.

My concerns include the potential for
abuse and diversion, as well as data questioning
its effectiveness for ADHD, along with
counterbalancing risk of adverse effects. Although
being promoted as a drug with low potential for
abuse, a substantial risk is actually suggested by
the following: One, the FDA posted a warning
letter on January 14, 2002 which compared the abuse
potential of modafinil with that of methylphenidate
in an inpatient study of individuals experienced
with drugs of abuse. Quote: Results from this
clinical trial demonstrated that modafinil produced
psychoactive and euphoric effects and feelings
consistent with other scheduled CNS stimulants

Number two, an Internet drug information
database source states, quote: Modafinil may be
habit forming. You should discuss the abuse and
dependence potential of modafinil with your doctor.

Number three, as reported in "The New York
Times," the United States Olympic Committee

includes modafinil in a list of banned stimulants
and raises the spectra of widespread diversion and
even more problem with, quote, lifestyle and
cognitive enhancement and recreational use as has
already been seen with traditional stimulants.
Several psychiatrists and other professionals with
experience with substance abuse by teenagers and
young adults have warned that off-label use of this
drug is, quote, staggering already, and warned that
modafinil is very likely to become the next popular
drug for its perceived cognitive enhancement or
other perceived benefits enabling users to remain
awake and alert for prolonged periods.

Marketing claims for the drug's
effectiveness for ADHD appear to be exaggerated. A
recent study in the Journal of Pediatrics of the
American Academy of Pediatrics proclaims that at
the final visit 48 percent of the modafinil-treated
subjects were rated as much or very much improved
compared to 17 percent of placebo subjects. One
can see that 52 percent of subjects were not
improved significantly. This compares to 75-85

percent comparable improvement on traditional
stimulants. Insomnia was reported in 29 percent of
the treated subjects, and it is noteworthy that
drug tolerability was evaluated only by
spontaneously reported adverse events. This can be
expected to minimize adverse events significantly
below their actual occurrence.

"The New York Times" article quotes
experts as being concerned that manipulating
natural sleep by reducing it may have serious
consequences such as chronic sleep depravation
damages health, immune system and is associated
with life span. All of these references are
provided in my speech.

The aforementioned FDA warning letter to
Cephalon pointed out that the putative mechanism of
action being claimed by the company was misleading,
noting that the PI states that, quote: The precise
mechanisms of action through which modafinil
promotes wakefulness is unknown, period.

In summary, claims for potential
effectiveness are exaggerated and the risks of

adverse events are minimized for a drug which has
been shown to have a potential for abuse and for
recreational use that far exceeds even that for
traditional stimulants. Approval at this time for
ADHD is premature considering the emerging
controversy and public awareness of issues of
adverse events, diversion and abuse related to

It is ironic that this very phenomenon is
being used by those promoting modafinil for ADHD.
That is--

DR. REESE: One minute.

DR. RAVENEL: --looking at the problems
with the stimulants. The FDA would be better
served by exercising caution and by opening the
door to even more of the same criticisms that have
emerged recently about stimulant drugs. Thank you
for your consideration.

DR. GOODMAN: Thank you.

DR. REESE: We will have speaker number

DR. JACKSON: If you could hold on putting

up the first slide, I might do those towards the
end. Thank you.

My name is Dr. Grace Jackson. I am here
independently as a private practice psychiatrist
from eastern North Carolina. I am here today to
actually begin by correcting some of the
misinformation which has been disseminated to
committee members over the past 48 hours.

The first point I would like to make is
about some of the concerns I have as a doctor and
who has actually worked in the prison system and as
a former Naval physician. One of the first things
I would like to point out is that I think that the
precautions which should be described are basically
the elephants that nobody seems to be looking at in
the room. I would like to talk about some of those

The first elephant has to do with the fact
that stimulants rewire the brain. This is what
Harvard University and McLean Hospital clinicians
have referred to as neuronal imprinting.
Basically, what this means is that we shouldn't be

focusing just upon the potential for current
diversion or current recreational abuse, but we
should be looking at the fact that these drugs are
altering the plasticity of the brain in children
and adolescents in a way which increases the
likelihood of future chemical dependencies,
particularly to nicotine and to cocaine.

I would like to direct your attention to
the papers by Nadine Lambert at the University of
California Berkeley, papers published by Russell
Berkeley in which statistical manipulations have
been used to try to conceal this correlation, and
also a recent publication from the University of
Michigan which has demonstrated the same kinds of
findings, that people who are arriving on college
campuses who have received stimulants in middle
school, high school or college have a 3-7 times
higher likelihood of taking prescription stimulants
illicitly, and a higher rate of actually using
cocaine in the past year.

The second elephant that I would like to
talk about which we haven't really been hearing

enough about, I don't believe, is the effects of
stimulants on growth suppression. While it used to
be the case that doctors took seriously the growth
curve, it seems that this is now something which is
casually dismissed. I believe it is time for the
FDA and physicians to begin seriously considering
the suppression of growth not only on the long
bones of the legs and the arms, but also potential
impact on the skull which continues development
through adolescence and particularly the growth
effects upon the brain, a point to which I will
return in a few moments.

The third elephant I would like to talk
about is the fact that no one yet here, at the FDA
or at these hearings in the past two days, has
discussed the effects of stimulants on cortical
blood flow, specifically frontal cortex, parenteral
cortex and temporal cortex. I believe that if you
will actually pay attention to the medical
literature there is a real vascular effect which
actually deserves a black box warning, at the
least, so that physicians and family members are

aware of the fact that these stimulants have the
potential to shrink the cortex, especially the
frontal cortex, and if they are not doing that, at
the very least, they are not benefiting children
who, some practitioners believe, begin life with
smaller brain volumes initially.

The next point I would like to make is
that we hear so much about the FDA needing to
balance the risks and the benefits of drugs. Well,
I would like to just point out the fact that I have
heard numerous references to the MTA study in the
past two days. Fourteen-month outcomes have been
emphasized repeatedly. I would like to just say
something that was misrepresented yesterday and
again today. If you will actually go into a paper
which was in the Archives of General Psychiatry in
1999, called "Mediators and Moderators of the
Outcomes of the MTA Study," you will find about two
sentences in that whole article where they actually
have done a subgroup analysis of the children who
began that study in an unmedicated condition and
who remained in an unmedicated condition. Those

children actually had superior numerical
improvements compared to the children who began
unmedicated and were placed on stimulants. While
that finding was not standardly significant, that
may have been an effect of the study being

Even more important though is the fact
that in the Pediatrics journal, in the year 2004,
24-month outcomes were published for the MTA study.
The findings at that point demonstrated that the
effects of medication deteriorated; that the
trajectories for symptomatic improvement reversed;
and, in fact, the benefits of behavioral therapy--a
modality that consisted mostly of one 8-week summer
camp--actually had enduring effects.

So, I would like to suggest that this
implies that a lot of the treatments that we are
hearing so much about as being so necessary are, in
fact, futile when one carries out the studies to a
long enough duration of time.

Finally, I would like to return to the
misinformation which continues to surround the

classification of medications that we keep hearing
as they are not really stimulants; they are just
central nervous system activators--

DR. REESE: One minute.

DR. JACKSON: I would like to suggest to
the committee that they need to talk to the World
Health Organization. Stimulants are classified not
on the basis of potential addictiveness; they are
classified as stimulants on their potential to be
CNS activators. Actually, the World Health
Organization classifies drugs on the basis of three
properties: One, chemical structure. You ignore
the fact that atomoxatine is a chemical derivative
of phenylpropanolamine, a chemical structure which
was removed from the market by the FDA in 2000
because it caused hemorrhagic stroke.

I would like to point out the fact that
the World Health Organization also classifies
stimulants on the basis of pharmacological
properties, none of which require dopamine
re-uptake inhibition to meet the criteria of being
a stimulant, and I would like to see that everybody

drinking coffee here to recognize the fact that
caffeine, which is an adenosine-2 antagonist, is
not something that you would classify by Dr.
Andreason's standards as a stimulant. Yet, I think
Starbucks would say something else.

Finally, I would like to say--

DR. REESE: Time.

DR. JACKSON: Thanks.

DR. GOODMAN: Thank you very much.

DR. BAUGHMAN: I am Fred Baughman, a
neurologist. I have discovered and described a
handful of real neurological and genetic diseases.
I am speaking on the chemical imbalance lie as it
applies to modafinil and other ADHD drugs.

If one goes to a physician or takes a
child or parent to a physician, if there is a gross
microscopic or chemical abnormality a disease is
present. If there are no abnormalities no
physician should say that there is a disease. In
psychiatry there are no physical abnormalities,
which means there are no actual diseases and here
we speak of the risk side of the risk-benefit

equation. Psychiatric drugs appeared in the '50s.
Psychiatry and big PhRMA married and gave birth to
the chemical imbalance lie. At a 1970
congressional hearing the chemical imbalance
strategy was already in place. Lippman, of the
FDA, argued hyperkinesis is a medical syndrome. In
1994 Leber, of the FDA, in a letter to me confessed
no pathophysiology has been delineated. At the
1998 consensus conference William Kerry concluded
ADHD appears to be a set of normal behaviors. At
the consensus conference James Swanson reviewed
anatomic MRI research, concluding ADHD subjects
have on average 10 percent atrophy.

I challenged Dr. Swanson, saying why
didn't you mention that virtually all of the ADHD
subjects were on stimulant therapy? The research
had proven 14 times over that the drugs, not the
fictitious disease ADHD, had caused the brain
atrophy. Caught in this lie, the consensus
conference panel confessed, quote: We do not have
an independent valid test for ADHD. There are no
data to indicate that ADHD is a brain malfunction.

Unlike real epidemics, once psychiatric diseases
are found not to exist, the epidemics flourish

In 2002 Castellanos published the one and
only MRI study of an ADHD untreated group but,
inexplicably, they failed to use matched controls,
voiding the study. So, I would hope that this
study is not referenced as the proof that ADHD is
an actual disease.

While the FDA's Goodman acknowledged that
claims that SSRIs correct serotonin and imbalance
go too far, he had the gall to suggest, quote, this
is reasonable shorthand for expressing that this is
a chemical or brain-based problem. Saying any
psychiatric diagnosis is a brain-based problem and
that medications are normalizing the function is an
absolute lie.

There is nothing more despicable than a
physician or physicians who knowingly tell normal
patients that they are diseased for profit--

DR. REESE: One minute.

DR. BAUGHMAN: --yet, this has become

standard practice throughout medicine and FDA, APA,
AMA, AACP, AAP, AANCNS, AAFP. The right to
informed consent universally abrogated must be
restored. You are mandating the medical treatment
of ADHD. Where is the proof that ADHD is a
disease? Give us that reference, that citation
right now, please. Give us the reference citation
to the test that demonstrates an objective
abnormality child by child, please.

DR. REESE: Calling for speaker number

MR. HANSON: Good afternoon. My name is
Ben Hanson. I am from Traverse City, Michigan. In
the interest of full disclosure, I suppose I should
mention a few things. In March, 2000 I was
appointed to the Michigan Department of Community
Health Recipient Rights Advisory Committee, a state
watchdog panel that meets in Lansing. I received
no compensation for serving on this committee,
other than mileage for travel expenses. Also, I am
the Michigan contact person for Mind Freedom
International, a non-profit organization which

advocates for the rights of individuals stigmatized
by psychiatric labels. This is a volunteer
position for which I receive no compensation. I am
also a proud member of the International Center for
the Study of Psychiatry and Psychology,,
and I was one of the principal organizers of
yesterday's ICSPP press conference here, in the
Hilton. It is possible I may be reimbursed for
some of my travel expenses by ICSPP but to date I
have not received one dime of compensation from
ICSPP, which is fine. I am happy to do this work
for free. Finally, I have been contracted on a
part-time basis by another non-profit organization,
the Law Project for Psychiatric Rights,, founded by Alaska's attorney Jim
Godstein. To date, I have received a total of not
over $1000 for various services like updating the
mailing list, working on the web page, etc.

I want to make it clear that I am speaking
on my own behalf today. I am not speaking for
anyone else, including these organizations I just
mentioned. I am here before you as a private

citizen, a taxpayer of the U.S.

I want to say a few words about the drug
Sparlon, also known as Provigil, also known as
modafinil. My interest in this drug began a few
years ago when I learned that modafinil had been
approved for treatment of a new disease called
shift work sleep disorder. This interested me
personally because for nine years, from 1995
through 2003, I worked for the Michigan Department
of Natural Resources as a ranger in a state park
located in northern Michigan. I worked the night
shift, from 7:00 p.m. until 4:00 a.m. five nights a
week. Basically, my job was to walk around in the
woods after dark, which I loved because I love the
outdoors. It was a dream job except for those
hours and I never got used to it. I can testify to
the fatigue, to the irritability, to the general
clumsiness and inattentiveness which is caused by
working those late night hours, especially that
last hour from 3:00 a.m. to 4:00 a.m. The rangers
call it the "dead hour," the dead of night when the
whole world except you seems to be asleep, nothing

stirred, not even the crickets, not even the

I can testify to this mental dullness. In
fact, I would say if you work those hours and you
don't grow a little groggy and a little clumsy, the
only reason I can think is that you are probably on
some kind of drug. I believe consenting adults
should have the right to take any drug they wish
but I condemn the FDA for endorsing a fictitious
disease, created most likely by some pharmaceutical
marketing department as a way to sell more drugs.
What is next, FDA? Are you going to approve jet
lag as an official disease? Perhaps it is only a
coincidence but I understand the formulary patent
on modafinil expires this month, March, 2006--a
minor inconvenience to Cephalon and its
stockholders. But one way around that problem
would be to change the name of the drug, call it
Sparlon and approve it for the treatment of ADHD,
which is a larger market than shift work sleep
disorder anyway. Isn't that what Eli Lilly did
when Prozac's patent was about to expire? They

changed the color of the pill from green to pink.
They changed the name to Sarafem and they marketed
it for another invented disease, PMDD, and the FDA
approved that. No problem.

I flew down here from Michigan--

DR. REESE: One minute.

MR. HANSON: --because I couldn't
believe--I can't believe that you people are really
going to approve this pep pill, which reportedly
allows people to get by on two hours of sleep a
night, for children diagnosed with ADHD. If you do
this I want to be here to see you do it with my own
eyes. Thank you for this opportunity to express my

Committee Discussion

DR. GOODMAN: Thank you. At this point I
would like to invite our committee members to ask
questions of both the FDA and the sponsor. At a
time when I think it is probably the appropriate
moment we will put the questions up on the screen,
but before we do that let's have more free-ranging
questions, including some that may have been

carried over during lunchtime. I remember that a
few people didn't get an opportunity to ask their

DR. REESE: Dr. Armenteros?

DR. ARMENTEROS: Yes, a question to the
sponsor pertaining to the most common side effect
in the list, which is insomnia, could you tell us a
little bit more? For example, does this start
happening in the very beginning? Do kids get used
to it? Does it change during the time of treatment
under observation? And, does it have anything to
do with dosing?

DR. RUSSELL: In the main, the insomnia
appears to start with treatment initiation, and the
highest incidence of first reports of insomnia
occurred during the first two weeks of treatment
and then it does appear to taper off. As Dr.
Stankovic mentioned, I think we had seven
withdrawals from the drug because of insomnia. So,
there appear to be people who either learn to get
used to the insomnia or habituate to it, as with
many of the other drugs that I think have this as a

side effect. I am sorry, I know there was a third
part to the question but I have forgotten it.

DR. ARMENTEROS: Yes, what is the
relationship to the dose?

DR. RUSSELL: We did look at the doses and
there doesn't appear to be a major difference with
the doses of 340 mg or 425 mg.

DR. REESE: Dr. Mehta?

DR. MEHTA: Actually, it is just a comment
on Dr. Temple's earlier remark. All the studies
are two-week or longer. There is only one study in
24 subjects which is a single dose. So, the
denominator should be somewhere around 920 or
something like that.

DR. GOODMAN: I have a question for the
sponsor about the pharmacokinetics. If you took
two children, same age, and one was being
administered 400 mg modafinil, the other 200 mg
modafinil obviously the plasma levels would be
higher in the one that is receiving the 400 mg, but
would the levels of the metabolite be proportionate
or disproportionate to those levels as well? You

may have covered that and I may have missed it.

DR. RUSSELL: Yes is the answer.
DR. GOODMAN: They would be proportional?
DR. GOODMAN: In a linear fashion?
DR. GOODMAN: Thank you.

DR. REESE: Dr. Leon?

DR. LEON: Could the sponsor please show
us a slide of the weekly retention rates? I notice
there is a big difference between the LOCF results
and the endpoint and I am curious to see how those
retention rates look, and if there are differences.
I know maybe about 50 percent more people dropped
out of placebo than active medication.

DR. RUSSELL: The biggest time of dropout
was between weeks three and five. This largely may
have had something to do with the design of the
protocol that did allow patients who were going to
come off for an adverse event to roll over into the
open-label program at that time. The reason for
that allowance was based on a lot of input from

investigators who found that it would be difficult
to keep children on a placebo for that length of
time. So, there was a dropout between week three
and week five.

DR. LEON: Do you have a slide that shows
the weekly retention rates? Could we see that,

DR. RUSSELL: I am looking at my
colleagues and they don't seem to have it.

DR. LEON: I didn't see it in the
materials. Is it in the book maybe? It is pretty
important when we are trying to draw inferences
about efficacy.

DR. RUSSELL: I think in your briefing
document there are by week numbers.

DR. LEON: I didn't see it. Maybe you
could tell us what pages to look on.

DR. RUSSELL: Let me try and find the
page. If you look at figure 4 on page 31 of the
briefing document, there are the numbers for the
CGI that are actually the numbers--sorry, they are
not; I am misleading you. I am afraid we don't

have it.

DR. REESE: Dr. Wells?

DR. WELLS: I have a question about the
source of the postmarketing adverse events that
were reported, a question to the sponsor. These
postmarketing events, do these come from all of the
postmarketing studies, all events from all studies?
Also, do they include voluntary reports from
practitioners in a more naturalistic setting?

DR. RUSSELL: I am sorry, could you repeat
the question?

DR. WELLS: The question is about the
source of the postmarketing reports of adverse
events. Where do these comes from? Presumably
postmarketing studies are included of the drug used
in other indications. Would it also include
voluntary reports of practitioners--

DR. RUSSELL: Yes, it would. The
spontaneous reports would be reports from
healthcare providers, consumers. Any study that we
undertake we include in our clinical trials

DR. WELLS: So, it is data from studies as
well as voluntary reports from practitioners?

DR. RUSSELL: In the postmarketing it
doesn't include the studies; it includes the
voluntary reports.

DR. WELLS: Just the voluntary reports?

DR. RUSSELL: If I could clarify the
previous question, on figure 3 on page 29 of the
briefing document there are numbers at the bottom
of each of the graphs.

DR. REESE: Dr. Pine?

DR. PINE: I guess two issues, and one of
them I think maybe we will just come back to. That
is the issue that Dr. Leon just raised about the
sample sizes for each week on page 25 efficacy
data. I realize you don't have it now but, you
know, I think a few of us are a little concerned
about differential attrition in terms of the
efficacy data and it would be very helpful to see
those data but, again, I know that you don't have
that right now but maybe if you could get them and
give them to us sometime in the next half hour or


The second issue is a question on
psychosis. I guess there are really two things
that--yes, that slide, right there. If you could
just give us the Ns in each group at each data


DR. RUSSELL: This is the ADHD rating
scale. Actually, the numbers are here the teachers
versions so you can see that there are dropouts as
the weeks go by.

DR. PINE: Yes, they are very small.

DR. RUSSELL: They are pretty small. I am
afraid I can't see those from here.

DR. PINE: You can barely make them out in
the document but you can see them; they are there.

But the issue of psychosis, there was an
extensive discussion about this yesterday, for
people who weren't here and I don't know that we
need to repeat the whole discussion. I guess I
would just like to raise two issues. One is that I
seem to recall on one of the slides from yesterday

that there was a hint--and I can't remember which
event it was, that one of the adverse psychiatric
events looked to be more prevalent in modafinil or
Sparlon relative to the other agents. If somebody
from the FDA could either point that out or bring
it up, that would be helpful. Then I have one
other point about that.

DR. GOODMAN: Which event was it? Dr.
Mosholder would know. Do you remember?

DR. MOSHOLDER: Andy Mosholder, Office of
Drug Safety at FDA. For the suicidal event
category there were four events in my analysis of
modafinil, zero on placebo. I wonder if that is

the one. For psychosis there were two and none on
placebo. Those are just the double-blind.
  DR. PINE: And I guess my take on it is

that I don't feel any differently looking at the
data here for this compound than I felt about the
broader discussion yesterday, on the one hand. On
the other hand, I think it is important,
particularly for people who weren't here yesterday,
to know that similar concerns that were raised in

general for other compounds should also be
acknowledged or discussed here.

I guess the last thing to say is that the
quality of the adverse event reports in general
always concerns me. But I guess on slide 102, case
number 312592271 with 10 months of ideas of
reference concerns me. I realize it is one case
and I don't think we should make too much out of
it, on the one hand. On the other hand, in terms
of discussing the medication I do think that we
need to at least bring up the point again that
there needs to be some acknowledgement that these
are potentially concerning adverse events.

DR. GOODMAN: As long as you are on that
subject, Dr. Pine, it reminds me that in the review
of the correspondence between the FDA and the
sponsor there was a description of one case that
seemed to be misclassified or mis-coded. It was an
individual who was said to have had a personality
change or personality disorder and, in fact, they
had a noose around their neck. Could somebody from
either the sponsor or the FDA side clarify?

Obviously, if you read a case like that it harkens
back to early concerns we had about previous
problems in appropriate reporting of those kinds of

DR. RUSSELL: Yes, I can comment on this
case. This was a six year-old girl who after two
days of discontinuing the drug--she stopped the
drug on day 91 and then on day 93 engaged in what
her mother calls bizarre behavior but there was
some suicide intent by putting a rope around her
neck. The patient was hospitalized. The inpatient
assessment says that the patient had major
separation anxiety and admitted to trying to hurt
herself with grave references to suicide. This was
a girl who had a history of mood swings and a
family history including maternal depression and a
suicide attempt. So, that is that case which was
originally thought by the investigator to be
abnormal behavior but, as you saw today, we
included it on the slide with the suicidal cases.

DR. REESE: Dr. Bigby?

DR. BIGBY: I have a question about

response of the placebo group. You have a figure
that was in the CD that you sent and it was a
summary slide for the three studies looking at the
ADHD rating scale school version for all three
studies. I think it went out to eight weeks, and
it is really striking how much the placebo curve
drops down. Also, if you can find and put that
slide up--I don't know if you have that slide, it
has the numbers of people still in the study at the
various time points.

So, I have two questions. The first one
is for anybody who knows about ADHD trials. Is
this kind of effect in the placebo group sort of
universally seen in ADHD trials?

DR. RUSSELL: Dr. Biederman?

DR. BIEDERMAN: I am not sure whether I
know what you are asking, but placebo response in
ADHD is on the order of magnitude of 30 percent on
average in the literature. You are asking if this
placebo effect is typical of other studies of ADHD.
It is pretty much within that range.

DR. BIGBY: Then, the second question I

think sort of goes back to the question about
numbers of dropouts. Now, at each of these time
points you have listed the number of patients in
the treated and the control group. For the control
group you start out with 210 and by the time you
get to the seventh week there are 71. So,
basically two-thirds have dropped out. Is the
score with the bracket at each one of those time
points just the people still in the trial, in which
case the same would be true for the treatment group
and it is sort of a per-protocol trial and not an
intent-to-treat trial.

DR. RUSSELL: Depicted on these figures is
the by week analysis so those are the patients that
are actually in the study at that particular time.
The endpoint is the last observation carried
forward analysis so all values are included in the
endpoint analysis.

DR. REESE: Dr. Wang?

DR. WANG: Yes, I want to explore some
more this differential in effect size between
modafinil and other ADHD treatments. Particularly,

you mentioned Strattera. What I am interested in
is the clinical significance of that decrement. Is
this decrement of clinical significance? Is it of
a size where it would warrant making modafinil a
second-line treatment? I think it will have some
bearing on how desirable we think it is to warn
about the safety issues.

DR. RUSSELL: The overall effect size
across all three studies was 0.69.

DR. WANG: I am talking about the
differential between the effect size in your
pivotal trials and what is known about the effect
sizes for other ADHD treatments and what is the
clinical significance of the difference.

DR. PINE: Can I ask a question about that
0.69? That is a Cohen's D for the difference in

active versus placebo? Is that what that
difference is?  
DR. RUSSELL: Dr. Kingsbury, can you

DR. PINE: You know, typically most people
go by Cohen's D criteria so stimulants have Cohen's

D effect of somewhere in the low 1s, 1 to 1.2.
Strattera I think is frequently quoted as 0.7 to

0.8. So, if it is 0.7, if that is a Cohen's D for
the difference in the change of active that
accounts for placebo and that would be a reasonable
effect size. But I would like to hear if that
really is the effect size they are quoting.

DR. KINGSBURY: Specifically calculated as
the difference in treatment effect divided by the
pooled estimate of the standard error.

DR. PINE: The pool of the placebo?

DR. KINGSBURY: The placebo.

DR. PINE: So, that is a reasonable
effect. It is not huge but it is within the realm
of an effective agent.

DR. WANG: Would you say it is getting to
the point where this would be a second-line

DR. PINE: I would not say that.
Clinically, based on an effect size of 0.69, I
would not say that that would make it a second-line
agent necessarily. It is clearly less than what

you would expect in stimulants but I would think
about it similarly as I would about atamoxatine.

DR. LAUGHREN: I think if you are going to
be comparing effect sizes for different drugs you
ought to be looking at it in the same trial because
it varies a lot from trial to trial. It is going
to depend on the sample size and on the placebo
response. So, I think it is really hazardous to
compare effect sizes, whatever measure you are
using, whether it is Cohen's D or anything else, to
cross-study comparisons.

DR. PINE: On the other hand, since we all
know of multi-drug trials where people obviously
are grappling a little bit with the efficacy data I
think we have to say something about, you know, in
the universe of studies of ADHD, is this in the
realm of a reasonable treatment or not. Again, I
would agree; I wouldn't quibble with what you said.

DR. WANG: And I am not arguing. I am
just trying to kind of qualitatively understand
whether if there is a warning, whatever shape or
form it takes, and it drives down use or

effectively turns modafinil into a second-line
treatment, is that a terrible thing? Is it a good
thing? Is it neutral? That is what I am trying to
kind of understand.

DR. REESE: Dr. Rappley?

DR. RAPPLEY: I want to go back to the
skin issue so if anybody else wants to talk about
the effect--keep going? Okay. I would like to ask
Dr. Bigby if he might have some insight about how
we might think about the spontaneous reports of
Stevens-Johnson syndrome and how that compares to
the actual incidence.

DR. BIGBY: Actually, that is a very good
question. Neil actually did a study in Canada
where he ascertained cases of TEN, and he can give
you the details of the study, and compared it to
the spontaneous reporting system they have in
Canada. I think that it is vastly under-reported.
I looked at that paper last week and I think it was
10 percent or less than the cases that he found
that had actually been reported. I think the same
is true for other researchers that have looked at

reporting of TEN vis-a-vis drug usage in an attempt
to try to determine rates of reactions.

DR. SHEAR: That is correct. We tried to
look at patients with TEN and contacted burn units
across the country to see patients who were coming
in versus what was actually reported to Health
Canada through the spontaneous reporting system,
and we came up with a number of around 10 percent
that were actually reported. So, we realized it
wasn't necessarily the burn doctors but probably
hospital pharmacists who were reporting it or other
healthcare professionals but still it was about a
10 percent reporting rate.

DR. REESE: Dr. Pfeffer?

DR. PFEFFER: Again a clarification, Dr.
Bigby. Maybe you can help us. You said that the
Stevens-Johnson was dose related.

DR. BIGBY: No, no.

DR. PFEFFER: Then if you could clarify a
little bit more about the onset of this type of
skin problem.

DR. BIGBY: I mean, I think SJS and TEN

are idiosyncratic hypersensitivity reactions. I
tend to shy away from talking about mechanism
because I don't think anybody really knows what the
mechanism is. Developing the disorder I think is
not dose related but the point I was trying to make
about the dose perhaps being a factor is that I am
aware of at least two studies that have shown that
the patient prognosis is better if the drug is
identified and stopped, and it has mostly to do
with the half-life of the drug and the body's
ability to clear the drug. I think that if you
start with a greater concentration it will take you
longer to have undetectable levels and it might
affect prognosis. I don't think it has anything to
do with the incidence.

DR. PFEFFER: One other question about the
syndrome, if a child develops this on a medication
such as modafinil, would that child be at increased
risk in the future for the syndrome? In other
words, would the exposure to this particular drug
increase the risk or would that not be an issue?

DR. BIGBY: Increase the risk if they were

exposed to any drug?

DR. PFEFFER: Either any drug or whatever
causes the syndrome, yes. Does it lead to

DR. BIGBY: Well, the only definite thing
that i can tell you is that if they got the same
drug again it might be that they would have the
same reaction. Whether it identifies them as
someone who is more likely than the general
population to do develop TEN to other drugs,
particularly drugs that are known to be associated
with TEN, I can't answer that question although
there is some suggestion, not entirely convincing,
that that might be the case. But it is not clear
to me that the exposure to the drug and the fact
that they developed the TEN as the cause of that
identifies them as someone who has that potential.
So, I don't know if I am answering your question.
I think that a patient that develops TEN to a
drug--there is some evidence that they are more
likely to develop that type of reaction to drugs.
But I think probably they were that way before the


Actually, again, this is a subject that
Neil has done more work on than anyone I know. I
mean, I think it would be useful for you to hear
his comment on it as well.

DR. SHEAR: Thanks, Michael. It is a
difficult question because you are not going to get
enough data ever to really do that, especially if a
child has had TEN. For every drug they get in the
future the parents ask can this drug cause TEN, it
is no longer a hypothetical possibility and if the
answer is yes, but don't worry that couldn't
possibly happen in a billion years, you know they
are not going to get the drug. So, you are not
going to collect that data.

What we did show was that among the
aromatic anticonvulsants there was a risk of
cross-reactivity and that is even hard to explain
structurally. We don't know why that is but we
showed in vitro and in vivo that it does seem to
exist. But otherwise, usually people who have had
Stevens-Johnson don't get it again.

DR. REESE: Dr. Malone?

DR. MALONE: I know you said that dose
probably was not related to Stevens-Johnson but I
guess it is a similar question, is there any
mechanism that will explain why a group of children
getting 340 mg of the drug might have a higher rate
of Stevens-Johnson than those being treated for
daytime sleepiness getting 200 mg or less?

DR. BIGBY: The only thing I can do is
repeat I don't think the development of TEN, as far
as anyone knows, is a dose-dependent phenomenon.

DR. GOODMAN: I think you are being
appropriately cautious but there is the other
factor we discussed of the sulfone metabolite.
Although there is no proven relationship, there is
certainly a suggestion based upon other compounds
that have been associated with Stevens-Johnson that
have that sulfone group. So, is it at least
conceivable or plausible that the higher levels of
that metabolite could pose a greater risk for
development of Stevens-Johnson syndrome?

DR. BIGBY: Is it conceivable? Yes. But,

I mean, I think the threshold for whatever it is
that is the mechanism for developing TEN is
exceeded by all of the doses that you are talking
about here. I mean, yes, what you said is
hypothetically true. The problem is I don't have
any evidence to say that it is or isn't.

DR. VAUGHT: Mr. Chairman, if I could
address that for you, please?

DR. REESE: You may.

DR. VAUGHT: Thank you. What I would like
to do is just perhaps orient the panel a bit.


Because of the inference of the sulfone to
agents that have been directly associated with the
occurrence of SJS--I am not going to do a chemistry
lecture today but on the right-hand side of the
slide is the modafinil sulfone. With the
structural characteristics there is, in fact, a
similarity across two agents that have been
directly related to SJS. Obviously with the sulfa
drugs and the aryl-sulfonamide valdecoxib, the only
similarity is the sulfone group. I think what you

will notice is--and I agree with Dr. Bigby that
while the mechanism is not well-known, in general
with the sulfa drugs the amine group becomes
activated and it is believed to be one part of the
overall syndrome that is created, as well as the
fact of the close association of the sulfone group
to the aromatic ring. Valdecoxib is similar to
this in that it has a sulfonamide group again
associated with the phenyl group. While this is
not a conclusive relationship, there seems to be a
very broad preponderance of this type of structural
feature being associated and directly related to
SJS. We can see that with the modafinil sulfone
moiety this is structurally simply not similar to
these agents.

DR. MANNHEIM: We have a similar slide I
would like to show.


DR. CAVANAUGH: We also looked and I have
to say I am very impressed with the level of
discussion today from everybody. When Dr. Mannheim
asked me about Stevens-Johnson and I heard that

there was a sulfone, I said, well, you know,
sulfonamide, as you know, is classically thought of
and it is in the labeling. Sulfonamides in general
are labeled as 0.1 to less than 1 percent. You can
again see the sulfone here with the amine, and that
is the sulfonamide; this is sulfanilamide and you
can see it here and, again, sulfamethoxazole.

As you pointed out, there is some
similarity with the sulfone but the amine is
separated by two carbons and there is also a ketone
here. If you look at sulfacetamide, and here you
see a 3D structure rendering so you can see it a
little clearer with the two oxygens, two carbons
separated, a third oxygen and then the nitrogen
which is going to be withdrawing electrons. If you
look at sulfacetamide, the difference is that
instead of the nitrogen being on this side, it is
basically substituting for this carbon. So, this
is a sulfonamide but it has a third oxygen, it is a
third atom away. The interesting thing about
sulfacetamide is that it has been reported to cause
Stevens-Johnson, at least in the labeling, at 3

percent--I believe it is in the labeling; it might
be in other places--and that is an eye drop. You
know, people have died with even the first dose of
eye drop where they have a history of sensitivities
to sulfonamide.

So, you know, this whole issue of is it
the sulfone, isn't it, I think what you have been
hearing is that it is very, very muddy. We don't
know. You have heard factual information and we
don't know. The same with the dose. It is too
small numbers. We don't know. It is plausible.
Is there cross-reactivity? Maybe yes, maybe no; we
don't know. So, these are some of the issues that
we have been struggling with and I am glad the
committee is dealing with them.


Let's see, was there any other point I
wanted to make? The only thing that I wanted to
point out is, you know, we have been talking about
Stevens-Johnson and there has been talk about other
hypersensitivity reactions, and I went through the
various cases of rash and you heard earlier about

the PK and the exposures not being any different,
and that is about what you would expect. But a lot
of those rashes also were just general rashes.
When you look at cases that could be possible
hypersensitivity, you have several cases of
allergic reactions. You have the vesiculobullous
possible SJS. You have increased LFTs. One of
them was a hypersensitivity. You don't have
anything up here in the teens but it also could be
due to the numbers. But in general the percentage
for possible hypersensitivity is kind of


These are the individual cases and you can
see the combination of symptoms. Here is an
allergic reaction. Here is an allergic reaction
with nothing else. Here is an allergic reaction
with a rash. Here is the increased LFTs with
eczema and that is the individual--oh, I am sorry,
here is the increased LFTs with edema and urticaria
and that is the individual who was 17-fold higher.
You also have hives, fever, whatever. So, there is

some evidence of additional hypersensitivity. We
are just arguing about numbers at this point in

DR. GOODMAN: That is helpful. Thank you
very much. I have a more global question for
anybody on the FDA side about the safety data. Is
the FDA in general satisfied that there is
sufficient long-term safety data at the doses being
used for the pediatric population? We have focused
a lot on the acute trials, individual areas of
concern, but just in terms of a kind of more
panoramic view do we have sufficient long-term
safety data at this point for this dose in this

DR. LAUGHREN: I think I remember from the
earlier slide that we now have about 240 patients
greater than six months. Is that right? That is
about as much as we usually have. Again, this is
not a new compound. If there were some adverse
event that we thought was related that had a long
latency we might be more worried about it. I think
the event that we are most concerned about here is

one that probably has a short latency.

Questions to the Committee

DR. GOODMAN: That satisfies me. I would
like to turn to the questions, if we could have
those projected. There are two questions for which
we must take a vote. The first question is has
modafinil been shown to be effective for the
treatment of ADHD in children and adolescents?

Number two, has modafinil been shown to be
acceptably safe in the treatment of attention
deficit hyperactivity disorder in children and

When we get to number two I would like to
break that down in the following way, starting
first with dermatological issues because those have
been the most salient features; then with cardiac,
growth and psychiatric. Let's begin with the first
question pertaining to efficacy.

I think I already shared my view earlier
and that has not changed, that I am satisfied that
there is sufficient efficacy data as we have heard.
You know, we don't have a direct head-to-head

comparison with an active comparator. That is
unfortunate but it is not an atypical situation.
When we have looked at the effect sizes, I think
most of the experts in the room said it is probably
not quite at the level of the stimulants. It is
probably closer to the range of Strattera, yet it
is still quite effective and has certain features
that I think would make it a valuable addition to
the armamentarium. I just shared my opinion but I
want us to have a discussion and hear from around
the room, from all of you, regarding the efficacy
question. Anybody can volunteer.

DR. REESE: Dr. Wang?

DR. WANG: I would just second that one
caveat about the comparative efficacy. I think it
would be useful to have additional data just to
understand where in the armamentarium this would
fall. That is number one.

Number two, another big area of a question
mark is the dose. I think it is unfortunate and it
would be helpful if there were more data to suggest
whether you have fathomed the lower bound of the

dosing range because, as Dr. Bigby said, maybe the
development of these skin rashes isn't necessarily
dose dependent but the prognosis may be dose
related. So, getting kids on the smallest dose
possible would be optimal. I don't know if this is
additional trials but some way to understand if a
lower dose might be useful.

DR. REESE: Dr. Pfeffer?

DR. PFEFFER: I share that idea and
concern and I asked the question previously. It
seems that trial 309 was a fixed dose and is the
dose that was proposed, which was the higher doses.
Trial 311 was a flexible dose and it looked to me,
in slide 53, that there was demonstrated efficacy
early on and I am assuming it is at a lower dose.
Then, on slide 54, while it was a flexible dose we
heard that it was a very rapid increase of dose
early on. So, there wasn't enough sense in that
trial if a lower dose might also have been

So, while I think there is definitely
demonstration of efficacy, the efficacy is

demonstrated on the high dose and the question
about would a lower dose serve the purpose is not
answered clearly.

DR. REESE: Dr. Temple?

DR. TEMPLE: Well, I think it would be
helpful if, maybe with another look, you took a
look at the Phase 2 study that led to the
conclusion that you need the high dose, number one.
I mean, we press people for dose-response data all
the time but apparently we were satisfied that that
had ruled out usefulness of lower doses.

The other thing to do is look at the
average dose or maybe even dose groups in the
titration studies to see whether, while the dose is
still quite low, there is some separation. I mean,
that wasn't the planned analysis but the company
may have that. Early on there is not much
separation so at least for the earliest part of the
titration you really don't see much. Then at
either three weeks, four weeks, five weeks you do,
but we don't know the doses or the average doses or
the subsets of dosing by that time. So, perhaps

one could look at that and see if we have an answer


DR. REESE: Dr. Armenteros?

DR. ARMENTEROS: That is fine, but it
could also be an artifact of the time lag between
administration of the drug and response.

DR. TEMPLE: I totally agree. If you
didn't see something you wouldn't really know
whether a lower dose might--but it was the Phase 2
study that I think is what convinced the Division
that the dose-finding was sufficient. So, I think
if you don't think that is adequate we need to know

DR. REESE: Dr. Pine?

DR. PINE: I guess just briefly to second
some of the statements, it looks at least to me
fairly clear that there are not a lot of questions
about efficacy. It sounds like the data have been
reviewed a few times. Just in looking at the three
studies, on the face of it there can be a
reasonably strong case made for efficacy here, and
I don't know that I have a big need to discuss it

much further although I would be happy to hear

other people's thoughts.

DR. REESE: Dr. Rappley?

DR. RAPPLEY: Speaking from the point of
view of a clinician, I would say that this is a
medication that looks to be somewhat less effective
than the other options available to me to treat
attention deficit hyperactivity disorder, and it
has the common side effects, common and mild side
effects that are very similar to the other agents.
So, it would probably be perhaps a fourth or even a
fifth line of medication that I might turn to in
order to treat a child who was not responding to
the other medications.

DR. GOODMAN: I wondered if you would
revise that positioning of the medication if you
had a sense of abuse potential, diversion
potential. We can talk about that a little bit
more today, but I have heard a variety of different
views on this. It would appear that the abuse
potential is less than with some of the stimulants
but it is certainly true we often don't find out

about abuse potential until a medication becomes
widely available in a particular population. So, I
was just wondering if you might revise that if you
felt that the abuse or diversion potential was

DR. RAPPLEY: Well, I would like to answer
that in two ways. One is that we have heard that
one reason a physician might want to use this is so
that the physician would not have to deal with
controlled substances. I don't like that argument.
That is not about what is best for my patient in
terms of their condition and their treatment. That
is about a system that makes it difficult for me to
deliver care effectively. So, I would rather
educate my families that this is not a narcotic and
it is controlled for some legitimate reasons and it
is the best set of medications I can use and,
therefore, I will work with that. So, I don't see
that as persuasive.

The other suggestion that it would be less
likely to be abused as an agent itself, I think
that might be attractive to me if, in fact, I was

looking at a family where I thought abuse by other
family members or my patient was possible, which is
not an unusual case for my practice. But I have
other agents in the classification of stimulants
that I could turn to for that purpose.

DR. GOODMAN: Other comments on the issue
of efficacy before we take a formal vote? Dr.

DR. TEMPLE: I am actually embarrassed to
have to ask this, but outside of maybe psychiatry
this Cohen's D is not widely used. Could somebody
dilate on that a little bit? To divide effect size
by some kind of measure of variance seems to give
you something that doesn't have tangibility.

DR. PINE: It is not dividing the effect
size. Maybe Andy can talk more about this. It is
dividing the mean. So, it is a difference in means
divided by pooled standard deviation. It kind of
goes back to the in the social sciences, in a
widely cited book in the mid '80s, about
statistical power for that particular metric, which
was the difference in means divided by the pooled

standard deviation in the two groups. Standards
were kind of put forth that were somewhat arbitrary
at the time, in the mid '80s, for a small, medium
and large effect. And, there are standard
deviation units so up to 0.3 was a small
difference; from 0.3 to 0.8 was medium; and above

0.8 was large.

Then, what has happened over the last 15
years, particularly among pediatric
psychopharmacologists but also adult
psychopharmacologists, is that those standards have
been applied and they tend to fit in terms of how
people think about medications clinically.
Typically, medications that physicians tend to
think about as powerful tend to have large
standardized differences or a difference in
standard deviation of approximately one unit
between an active treatment and an inactive
treatment. Similarly, medium treatments tend to
follow in the 0.5 to the 0.8 range.

DR. TEMPLE: It sounds, for example, like
making your study larger makes your effect size

look bigger.

DR. PINE: No, it will not.

DR. TEMPLE: Won't decrease the standard

DR. PINE: No, it will not do that. In
fact, one of the nice things about the Cohen's D is
that it is independent of sample size.

DR. TEMPLE: We will talk off-line.

DR. GOODMAN: Other comments before we
call the vote?

[No response]

In that case, we are voting on the first
question on efficacy. We have three options, yes,
no or abstain. Let's start with Dr. Mehta.
Although officially his vote doesn't count, in my
mind his non-vote is extremely persuasive.

DR. MEHTA: On this drug it is not an
issue but I think it is a pleasure not to be able
to vote on most of the drugs! I think there is
clear and persistent evidence of efficacy so
efficacy-wise I don't think I have an issue.

DR. MALONE: I don't have any issue with

efficacy either. I think that all the studies were

positive and overall it looks effective.
DR. REESE: Ms. Dokken?
MS. DOKKEN: Yes on efficacy.
DR. REESE: Could you say your name before

you give your vote? Thanks.
DR. WELLS: Barbara Wells, yes.
DR. ARMENTEROS: Dr. Armenteros, yes.
DR. PFEFFER: Dr. Cynthia Pfeffer, yes.
DR. ROBINSON: Delbert Robinson, yes.
DR. LEON: Andrew Leon, yes.
DR. PINE: Danny Pine, yes.
DR. GOODMAN: Wayne Goodman, yes.
MS. BRONSTEIN: Jean Bronstein, yes.
DR. WANG: Phil Wang, yes with those two

caveats earlier.
DR. RAPPLEY: Marsha Rappley, yes.
DR. BIGBY: Michael Bigby, and if you

really want my opinion about the efficacy of a

psychiatric drug, yes.
DR. GOODMAN: Do you want to recap for us


DR. REESE: Well, "the yes" have it. It
is unanimous.

DR. GOODMAN: Let's turn to question
number two, which is a bit more thorny--

DR. BIGBY: Could I ask a question?

DR. GOODMAN: Who has a question?

DR. BIGBY: Me. Is there a definition for
acceptably safe?

DR. GOODMAN: It is the same that you
would use in dermatology!


DR. REESE: Dr. Temple?

DR. TEMPLE: Well, this goes back to the
law and various elaborations of it. What the law
asks is that safety be assessed by all tests
reasonably applicable--a very broad standard that
you could drive any sort of truck through; and that
it show the drug to be safe for its effective use,
which has generally been interpreted to mean that
the benefits appear to outweigh the risks. But it
goes on to make it clear that something can be

unacceptable either because it shows something bad
or because you haven't done enough. We have
elaborated on that in various risk management
things but it is always the same--have you done
what you need to do or enough of what you need to
do? A judgment call obviously. And, can you
conclude that in light of what it does that is good
for you, you have acceptable risk? That is what it
always means.

DR. GOODMAN: As we return to this
question, I would like to break it down to
different categories. Let's start with the
dermatological issues first. I wonder if I could
turn to you, Dr. Bigby, to offer your opinion on
whether you think this drug is reasonably safe in
this population, given what we have heard today
about possible dermatological complications?

DR. BIGBY: I think that the drug should
be put in the context of other currently available,
marketed and highly used drugs where over time it
has become clear that they are associated with the
development of severe adverse skin reactions, such

as TEN and SJS, and I think that this drug will
find itself among that group.

DR. PINE: Can I ask you a question about
that? I actually found, and I don't know what the
number of the slide is from Dr. Andreason--I found
the slide that gives the labeling of Lamictal
interesting and relevant and I wondered if you
might comment on that. For Lamictal it says
Stevens-Johnson syndrome--it gives 8/1000, that is
what it gives in children. It is page 8, on the
bottom. Because I do think your comment about
placing it in the context of other medications is
very helpful and lamotrigine is a medication that
there is some familiarity with and I wondered if
you might comment on the comparison. It has a
black box below age 15, lamotrigine. Is that
right? I think that is right. That is my
recollection, anyway.

DR. BIGBY: You know, the problem that I
think you are going to have is that you are going
to have a difficult time coming up with and
agreeing on a number, but I do think that the drug

should be labeled as one where people should be
aware that it could be associated with SJS/TEN.
Actually, I am quite surprised at the 8/1000 number
because you are pretty close to one percent. You
know, that is a pretty high rate for TEN. So, I
don't know, I mean, I have a hard time believing
that the number is really that high. Is it really
that high?

DR. PINE: I don't know. I am just
looking at what Dr. Andreason--

DR. GOODMAN: The numbers have come down
over time. Is that correct?

DR. ANDREASON: Well, it is hard to say.
If you look at the prospective registry study that
was done, there was one death of Stevens-Johnson
syndrome in that prospective registry with 1983.
So, the idea that it is more common in children
than in adults is fairly well accepted. I think
that the numbers are reasonable from what we know.
I think they are reasonable estimates.

DR. GOODMAN: It was placed in a black box
and even though over time it would appear that the

incidence is lower for lamotrigine it has
maintained its black box position. Is that

DR. ANDREASON: I am not aware that the
numbers have actually been documented to come down.

DR. TEMPLE: But if you had one death in a
thousand people you wouldn't remove the black box.
That is an impressive number for most drugs.

DR. GOODMAN: Dr. Bigby, I have a
follow-up question too. I agree with your position
but I just want to clarify the basis of it. It
seems like it is largely on one case, one case in
which you have fair degree of suspicion or
confidence that there is a bona fide case of
Stevens-Johnson syndrome and, given the
denominator, that was enough for you to be
concerned. Is that fair?


DR. GOODMAN: And there was some exanthem
as well.

DR. BIGBY: That is fair. I mean, that is
a fair statement of my position.

DR. GOODMAN: Let me just follow up then.
There is this disconnect that we have all talked
about--the real concern here is the extrapolation
to large numbers and there is the disconnect with
the postmarketing surveillance. But it would seem
to me that that could in part be explained by
dosing. I think I understand that dosing may not
determine the incidence but it may have played a
role in the persistence of the problem. So, we
don't know whether the sulfone metabolite is
relevant or not, nevertheless, we don't have a lot
of postmarketing data in that age range at that
dose and that could, indeed, explain the lower than
expected rate in that population, in my mind at
least. I just want to see others' reflection on
that position.

DR. BIGBY: Can I just make a comment?
You know, I have been involved in quite a few of
these discussions about incidence of side effects
postmarketing, and one of the things that is really
striking about postmarketing studies is that unless
they are very rigorous they don't detect much. So,

when you are relying on spontaneous reports I think
that you are going to miss a lot of the cases that
were, in fact, cases and it is striking how poor,
in terms of pickup of adverse reactions,
postmarketing studies are unless they are really
done with some sort of design in mind.

DR. TEMPLE: Obviously a problem is that
nobody can answer the question of what the degree
of under-reporting is and it is estimated widely.
However, there is a lot of reason to think it is
less bad when events occur that are likely to be
drug related. So, for example, we have been pretty
good at picking up acute hepatic necrosis in cases
like that because when that happens the drug is
highly suspect. When we approved a drug that was a
major 3A4 inhibitor we got cases of rhabdommyolysis
because it inhibited the metabolism of a couple of
statins. We get cases very rapidly. Now, I don't
know whether we got them all but these kinds of
things you probably do better than things that
happen regularly in the background--seizures,
things like that--where why would a person decide

that the drug did it. So, it is not that
discouraging for things that are obvious and that
deepens the mystery to me because, you know, the
fact that the dose is about half what you would
recommend now, that doesn't make it seem like there
shouldn't be any cases. I mean, that is why it is
here, because we find it a little surprising that
there are no cases and yet there was one.

DR. CAVANNAUGH: In terms of your
question, Dr. Andreason showed a slide where he
estimated about 11,000 children 6-12 years old who
were probably getting modafinil from the
postmarketing experience. If you take that 1/900,
that is just about 0.1 percent. So, if you take

0.1 percent of 11,000, that would be about 10

Now, it is commonly quoted that reporting
rates are about 10 percent. That is based upon
drugs where they may have been on the market a
while but, all of a sudden, somebody publishes an
article with a case-control series and then
everybody else starts reporting it. In that case,

after people are, you know, kind of all reporting
cases, that is where you get the 10 percent. You
know, if you even took 10 percent of 11 cases or 10
cases you might expect one case to be reported.

Now, yesterday you heard that psychosis
and aggression was about one percent consistently
with the various drugs used for ADHD. Back in
June, we also discussed this with Concerta
specifically and you have about 1.25 million kids
on Concerta and we know now that it is about one
percent in terms of psychosis. Yet, you were
dealing six months ago--say, one percent out of

1.25 million is 1250 and yet you were only dealing
with--I can't remember the number but maybe 30
cases. So, it was less than one-half of one
percent that was the reporting rate.

MS. BRONSTEIN: My question is to the FDA.
What kind of requirement does manufacture of
Lamictal have for postmarketing studies and
reporting of incidence?

DR. ANDREASON: I am not sure what the
requirements are. Right now, they have already

completed the registry study. That is in labeling.
There is already a black box. I think that the
risk has been capped. I am not sure exactly what
more one would want. It is also noted in the black
box that it is only approved in children for Lennox
Gasteau even though it is approved for other things
in adults. I think that is about all we could

MS. BRONSTEIN: Thank you.

DR. REESE: Dr. Laughren?

DR. LAUGHREN: I want to come back to the
point that Dr. Cavanaugh was making. I think there
is a real problem in knowing what the extent of
under-reporting is and it probably varies so much
depending on what the event is. With something
like psychosis, especially depending on how you
define it and if you are defining it just as
hallucinations, a lot of those probably aren't
going to get reported because it is a fairly common
event in the background. Something like
Stevens-Johnson, which is an extremely unusual
event, a very alarming event, is probably much more

likely to get reported. But the truth is we don't
know what the extent of under-reporting is so you
have to factor that into this. It is hard to know
what it means that you don't have any reports among
roughly 35,000 kids who have been exposed to it
postmarketing but it is a disconnect and you just
have to figure that in, in your overall
deliberations on this matter.

DR. REESE: Dr. Mehta?

DR. MEHTA: I think it is just a comment
to Dr. Cavanaugh too, essentially reiterating what
Dr. Laughren said. I can't believe that 90 percent
of the Stevens-Johnson syndrome which occurs in
patients, either in Europe or in this country, is
not reported.

DR. ANDREASON: Also, those numbers on
exposure are unique patients between the years 2002
and 2005 only in the United States.

DR. REESE: Dr. Temple?

DR. TEMPLE: This is right at the heart of
all this. If you really believe the one case is
likely to be drug related you are talking about a

rate with a point estimate of something like 1/1000
and a lower bound that is a lot worse than that.
That is one problem.

One question is how reassured to be by the
fact that the pediatric use in the outpatient
setting hasn't produced any, and I guess if you
follow what Dr. Cavanaugh said you shouldn't take
any reassurance from that at all because people
report so poorly. My own view is that I take a
little bit of reassurance but it is very hard to
know. But that is what is at the nub of this.
Just to make it obvious in case it isn't from the
questions, the things you can do is try to manage
that risk, taking some estimate of it, or ask for
more data. That is the question. That is what
question two is about.

DR. PINE: I guess thinking out loud a
little bit, and in many ways my comments are
similar to what Dr. Temple just said, I think if
you listen to anybody who knows about dermatologic
issues and who has talked about it today there is
clearly a concern among everybody I think--you

know, the sponsor's dermatologist was concerned;
Dr. Bigby is clearly concerned. I guess the thing
I am struggling with is, you know, what is the
level of concern. I think the other thing we would
say, and I think everybody would agree with this
and Dr. Bigby himself said this, that we really do
not have enough data clearly to specify what the
level of concern would be because there is this one
case out of 923 but, when pushed, I totally agree
with what you said, that you haven't examined the

So, I guess what it brings things down to
and it makes me somewhat uncomfortable is that
there is a lot of judgment call going on here for a
potentially incredibly important decision. I just
feel somewhat uneasy with that because, you know,
you miss it either way and you could screw up big
time. I don't know if that says we need to get
more data. I don't know what that says but it just
seems to me that we are stuck in a way.

DR. GOODMAN: Let me take it from there.
So, I think the real question I would like to

discuss now about this issue among the committee is
whether it warrants a black box for that concern
about Stevens-Johnson syndrome. I think that is
really what you are alluding to there, Danny. We
have efficacy. We have agreed upon that. We have
already voted it. We have concerns about
Stevens-Johnson but we have only one case that we
can really hang our hat on. We don't have the
postmarketing yet. It might be appropriate use for
a black box given that it is something that will
alert the prescriber and the patient to recognize
it early. I think it is that early recognition
that could make a difference in terms of outcome.

I am not emphatic about it. I think that
there might be other ways of addressing the warning
without it being put in a black box because we have
so little data at this point. Perhaps the
highlighting would be a step below that. There is
no question I think at this point that it should be
included among the warnings. So, it is really a
matter of does it wind up in a box or is it
highlighted. Those are probably the two choices in

my mind. Dr. Temple, help us.

DR. TEMPLE: Well, I would say, not to try
to preempt the discussion, it at least gets a black

DR. PINE: Why do you say that?

DR. TEMPLE: That is the least because the
only data we have says the rate is something like
1/1000. It is life-threatening. Everybody has to
know about this and we don't know the rate. It
could be 1/300; it could be worse. I have
discussed this with Tom and I am virtually certain
that would be what we would do.

There are two other things to do though
that you need to think about and address for us.
One is whether it should be in some form or another
recommended as not first-line therapy or think
about other things first. There are various levels
of subtlety in how to do it. We also are going to
ask you whether we should ask for more data before
we say yes. But maybe you think we are just wacky
about the black box. That is all right, feel free
to tell us.

DR. GOODMAN: Dr. Wang?

DR. WANG: I think there are several lines
of argument that all point towards at least a black
box. I mean, Lamictal sets the standard. If you
are willing to put a black box for one
Stevens-Johnson death out of 2000, here our best
estimate is about 1/1000. The fact that, you know,
with Lamictal the case was a fatal one doesn't
really hold much weight. I mean, there are black
box warnings for suicidality even though none of
the cases were fatal. So, the fact that this one
case didn't die is just fortunate I think.

DR. GOODMAN: I am sorry to interrupt, but
the big difference there is we also didn't have
efficacy, or at least very much efficacy.

DR. WANG: Granted. I think this whole
issue of should this, on efficacy grounds, be a
second-line treatment again pushes you. There
appears to be less of a downside in putting
something like a black box because if that de facto
has the effect of causing it to be used second
after failing a stimulant, then maybe that is, on

efficacy grounds, also justified.

DR. TEMPLE: I have to say we would be
very uncomfortable without a direct comparison
asserting--I mean, even though everybody loves this
measurement and everything, we would be very
uncomfortable asserting that it is second line
because it is not as good, without direct
comparisons. You can come back and say why don't
you ask for direct comparisons all the time. That
is another story.

DR. REESE: Dr. Rappley?

DR. RAPPLEY: Have we rendered an opinion
about whether or not this is acceptably safe? I
think in some ways we have taken a jump here and
talked about what kind of labeling.

Also, something you said made me think the
black box label, or whatever warning is on the
label is not related to efficacy and that is not a
risk-benefit judgment. That is just a statement of
risk. Am I correct?

DR. PINE: The way it was discussed
yesterday, and it would be nice to hear that again,

in what makes a black box it was a risk-benefit
consideration and efficacy does go into the
consideration, at least the way it was discussed

DR. RAPPLEY: So, approval for use in
children I see as weighed risk and benefit.

DR. TEMPLE: The decision to include a box
has something to do with what the drug is for. If
there were only one treatment for this and it was
considered urgent to treat it, I don't know whether
you would put a black box in. We don't box most
anti-cancer drugs, but they are all lethal in one
degree or another, because that is an expected part
of the deal. So, what it is for and what it does
has at least something to do with it. So, there
are several other classes of drugs that work; you
have choices; and here is one particular liability.
But feel free to tell me I am all wet.

DR. GOODMAN: Let me stay with Dr.
Rappley's comment. I didn't mean to short-circuit
the discussion. I was offering my opinion but you
are welcome to express the opinion if you feel,

based upon the dermatological issues, it is not
reasonably safe.

DR. RAPPLEY: The way I am thinking about
this is I understand that there is a particular
metabolism of this medication in children and we
have one case of Stevens-Johnson, perhaps 1/1000.
We have plausibility that this medication can be
linked to this serious condition. My understanding
of under-reporting is that it is significantly
under-reported and it is more likely to be common
and I am reflecting comments from Dr. Bigby that we
will find it associated in the future, and my faith
in postmarketing studies is somewhat small.

So, given those things, I think that
children are at risk for serious side effects with
this medication and, if you ask me to do the cost
benefit analysis, I think it is not adequately
balanced by what we have to offer in bringing this
to treatment of children for ADHD.

DR. GOODMAN: I would like to hear if
others would share that point of view.

DR. REESE: Dr. Malone?

DR. MALONE: I think I partly share that
point of view. I don't think it is safe enough to
recommend it as a first-line treatment, especially
when we have a number of effective, well-known
first-line treatments--with the data that we have
right now. It may turn out that this isn't going
to be an issue but I think with the data that we
have now it is hard to recommend as a first-line
treatment something that could have such a
dangerous side effect.

DR. REESE: Dr. Laughren?

DR. LAUGHREN: I just want to come back to
a point that Dr. Temple was making earlier about
what acceptably safe means. Part of what is
inherent in that concept is having enough
information to make a judgment about safety so I
really want to make sure that you consider the full
range of options. You might, looking at what you
have, decide that you don't have enough information
to make a judgment about safety but if you are
going down that path, then tell us what more
information you would like to see. But I just want

to make sure you consider the full range of options
other than, you know, black boxes and whatever.

DR. REESE: Dr. Bigby and then Dr.
DR. BIGBY: I actually enjoyed the comment

at the end of the table here because I don't know
if you figured this out but I like to try to make
things simple. You know, thinking about it that
way does actually make it simpler. The statement
about we don't have enough information to say that
it is safe, I would actually say it the other way
and that is that we have reason to worry but we
don't actually have enough data to say it is not

DR. REESE: Dr. Leon?

DR. LEON: Well, I am concerned about a
couple of comments I have heard in the last 10 or
15 minutes. Dr. Pine said right now it is just a
judgment. Without putting words in his mouth, I
think we are basing this without enough data. Dr.
Bigby is predicting that once this is used widely
we will see more Stevens-Johnson; Dr. Temple is

saying we need more data and suggesting we should
look at more data. I don't feel comfortable saying
it is safe until we have more data. There is at
least one ongoing study. When are those results
going to be in? There are 303 children, if I am
correct, being followed right now. It is certainly
worth waiting for them, and that is still a very
small number.

DR. BIGBY: But those children aren't
going to help you with the issue that you have.

DR. LEON: That is a good point, yes. But
in my opinion we just have inadequate data. In the
first 1000 there was a case. Is the next 1000
going to have 20 cases or zero cases? I don't
think we can guess yet.

DR. REESE: Dr. Robinson?

DR. ROBINSON: Actually, it is
interesting, what Andy is saying. I guess my
question is, okay, we have 1/1000, how many more
kids do we have to do to where we really say the
estimate really changed dramatically, either going
down or going up, that would be clinically

meaningful either down or up? Are we talking about
having another 1000 kids? Another 10,000 kids?
Because we are dealing with what seems to be a rare
event with all drugs. So, that is the question.
It is always good to say we would like more data
but is that in the actual realm of doability?

DR. LAUGHREN: Actually, you can figure
out how many patients you need to follow to cap the
risk at whatever level you want to be comfortable
with. We have this rule of 3 which, you know,
estimates the upper bound of the confidence
interval for the finding of no cases. For example,
if you wanted to be comfortable with a level of
1/1000 you would have to follow 3000 for whatever
period of time was of interest. If you found no
cases, that would cap the risk at 1/1000. So, you
can use that method to calculate how many patients
you would have to look at, at the doses of interest
and for the time period of interest, with the
finding of no events that would cap the risk. Now,
if you wanted to cap the risk at somewhere near the
background rate, that is not a doable experiment

but you could at least figure out, say, with 3000
that the risk is no greater than 1/1000 if you
found no cases.

DR. REESE: Dr. Pine?

DR. PINE: I guess two things. I want to
bring up one point that we haven't spent much time
talking about, and that is kind of the need for
more treatments in ADHD. You know, the important
thing to remember is, yes, clearly stimulants are
effective. No question, and they are good
treatments and there are other treatments around.
Again, no question. But even when medications are
effective the amount of improvement that you get
even when treatments work well is often not
necessarily what you want, and there are not nearly
enough treatments available for kids with ADHD.
You know, I think it is hard to say where this is
going to fit in and I would totally second what Dr.
Temple said, you know, to base a decision on
limited use on efficacy would not be a good thing
to do because it is a yes/no question. The
medication clearly works. And, I am uneasy about

withholding treatment that could be potentially
efficacious given the availability of treatments,
such as they are, for ADHD. So, that is the first
thing. I don't think we have spent enough time
recognizing that fact, that there are clearly needs
for other treatments. Number one.

Number two, thinking about that on the one
hand, with capping the risk on the other hand, just
personally, off the top of my head, I would be much
more comfortable if we could cap the risk at
1/1000. I would feel much more comfortable about
making a statement or decision or conclusion about
whatever the word--what is the word, relatively
safe?--acceptably safe. If I knew that a good
estimate of the risk was 1/1000 I would feel a lot
better. If you are saying that 3000 cases treated
for two weeks openly and we see no cases would
answer that question, I would feel a heck of a lot

DR. TEMPLE: That is our rule of 3, and I
am sure Dr. Leon can explain why it is not quite
right but it has been considered close enough.

Just another way to look at this, suppose you
thought that the risk could be as great as 1/500--I
mean, the data we have now has a confidence
interval and it probably goes down to 1/300 or
something like that, where would you be
comfortable? You just said 1/1000 properly labeled
and everybody knowing it would probably be okay.
But I think it is important to discuss that.

DR. PINE: One in 950.

DR. REESE: Dr. Pfeffer?

DR. PFEFFER: I think there are several
other considerations. I certainly agree that if we
can enlarge our treatment spectrum for this
disorder it would be wonderful. But I also think
that we in a way have concern about the potential
risks in this case without sufficient data, and I
am thinking also about what happens in the real
world once a drug is approved. Many of the
children with this disorder are treated with
multiple medications, unfortunately, and I would
wonder about what drugs might have potential for
cross-reactivity that might increase the risk for

these children. And, I tend to think that we have
a disorder that is severe, there is no doubt. We
have carefully tried to develop approaches to treat
these children and perhaps a careful approach is to
ask for more data and to sort of place that in
abeyance for the time being until we can answer
this question with a little bit more assuredness.
It just raises a new issue because we did talk
about some medications, one of which I think is
commonly used, which could have cross-reactivity.

DR. GOODMAN: Let me clarify. We have
1/1000 and there was an estimate of--what was
it?--5 percent of cases of Stevens-Johnson lead to


DR. GOODMAN: Let's start with that just
as a figure. How many open cases would you need to
treat, for what period of time, in order to gather
those data with some degree of confidence?

DR. TEMPLE: I don't think you could
imagine getting good mortality data--

DR. GOODMAN: No, I am not talking about

mortality data.

DR. TEMPLE: Well, to take a simple task,
as Tom said, if you wanted reasonable reassurance
that it was not greater than 1/1000, if you had
data on 3000 people and no cases that would provide
that. I mean, the tension we have had is here is
this one case in 1000. Here are 30,000 people
treated, no cases. Is this just some wild, weird
fluke or is that close to the true rate? You might
even decide--I mean, you did say even if that is
the true rate, that might be okay. Maybe you would
make it second line or do something else. That
might be okay. But at the moment, one of the
reasons this was brought to you is we don't know
what the rate is. We don't have enough data to
know what the rate is and it could be rather high
or maybe it is really low and this is just a fluke
and that is our uncertainty.

DR. PINE: Speaking only for myself, that
is what I would want to know and I would be
comfortable with that, but until I know that it is
going to be hard for me to make a decision.

DR. REESE: Dr. Rappley?

DR. RAPPLEY: I could ask it another way.
Is anybody comfortable with the amount of data that
we currently have? Then we could move to
discussing what additional data we need if there is
further discussion on that. I don't mean to push.

DR. GOODMAN: No, that is good.

DR. TEMPLE: Just one thing, as you
discuss that I think it is important to think of
enough data for everybody, enough data for a fairly
scary statement that says this is only for people
who haven't responded well to other things, not
that we have data on that but, I mean, there are a
number of things to think about as you discuss

DR. PINE: Again, related to the
discussion we have had I don't think the questions
are really about efficacy or what the niche is
going to be so, personally, I would care less about
who receives the medication in terms of what narrow
type of condition they have, and I would be more
concerned with capping the risk estimate. Based on

what Dr. Bigby said, it seems to me 3000 patients
treated for a month openly would be what you would
want to do.

DR. GOODMAN: We are not finished with
this obviously. I would like to move on to some of
the other concerns we have and see if we can go
through a list and perhaps even identify where we
think that this medication might have some
advantage, some possible niche.

In terms of cardiac issues, those were
discussed at some length yesterday. I think in the
context of stimulants it was decided that an
individual who had known structural cardiac
abnormalities should not be prescribed a stimulant.
Would we be having similar concerns about this
agent? In the data that I have seen there wasn't
very much evidence for increases in cardiac
parameters such as heart rate or blood pressure
and, therefore, would it be in that context perhaps
a safer alternative?

DR. REESE: Dr. Andreason?

DR. ANDREASON: I just wanted to add that

in the Provigil labeling already it warns against
using modafinil in patients with hypertrophy and
bicuspid aortic valve.

DR. GOODMAN: So, you would already put it
in the same category with the stimulants?

DR. ANDREASON: Well, it kind of already
is. It is already in labeling. Unless you felt
that the data that was presented should remove

  DR. GOODMAN: I don't see anyone saying
  DR. REESE: Could you come to the

microphone and state your name? Thank you.

DR. HERSKOWITZ: Norman Herskowitz,
medical officer in DMP. In the labeling, as I
recall, it really discusses the limitation--I think
this is the initial studies--to issues of mitral
valve stenosis and regurge type of syndromes, but
not to any other sorts of cardiac history. So,
that is just for information sake.

DR. ANDREASON: I am pulling up that
labeling for you; I am not as fast as I thought I

would be.

DR. HERSKOWITZ: It mentions some very
subtle changes in blood pressure, but extremely
subtle. In the adult studies there seemed to be a
pattern of increase in anti-hypertensive use
although no changes in mean blood pressure.

DR. ANDREASON: I have it. This is under
cardiovascular system in the Provigil labeling. It
says in clinical studies of Provigil signs and
symptoms, including chest pain, palpitations,
dyspnea and transient ischemic T-wave changes on
ECG were observed in three subjects in association
with mitral valve prolapse or left ventricular
hypertrophy. It is recommended that Provigil
tablets not be used in patients with a history of
left ventricular hypertrophy or in patients with
mitral valve prolapse who have experienced the
mitral valve prolapse syndrome in previously
receiving CNS stimulants. Such signs may include
but are not limited to ischemic ECG changes, chest
pain or arrhythmia.

DR. GOODMAN: And that is at a lower dose

than is being proposed.


DR. REESE: Dr. Rappley?

DR. RAPPLEY: The discussion yesterday
from Dr. John Moore who is a pediatric cardiologist
on the Pediatric Advisory Committee and, Deborah,
add to this if you can, we talked about how the
increases in blood pressure and pulse were perhaps
not clinically significant for children but
statistically significant and, yet, the concern
persists because of the idiopathic hypertrophic
subaortic stenosis being a condition that really
cannot be detected in the population until the
serious adverse event occurs, and that it is
plausible that increasing sympathetic tone could
contribute to that in the same way that running
track or becoming dehydrated does.

DR. GOODMAN: After Dr. Pine makes a
comment I would like to take a ten-minute break
before we come back for further discussion and vote
on the second question. I need a few minutes to
deliberate. Dr. Pine?

DR. PINE: I guess with a lot of these
secondary adverse effects--the cardiac effects, the
psychiatric sequelae, the growth effects--for some
of the same reasons that people were uncomfortable
making statements about comparative efficacy, I
would be uncomfortable making statements about
comparative adversity unless there have been
head-to-head trials, which there haven't been. You
know, my take from looking at all the other data,
besides the dermatologic data, I am slightly
concerned with the psychiatric adverse effects, no
more concerned here than the discussions yesterday,
and I just think it is probably not fair, given the
data, to make statements that this is better or not
better than any other agent unless they have been
compared head-to-head. I think it is, you know, is
it safe enough or not for all of these secondary
issues and, again, in my mind it seems safe enough,
whatever that means.

DR. GOODMAN: I would agree. Let's take a
ten-minute break.

[Brief recess]

DR. GOODMAN: It seems to me that a lot is

hinging on one case and I still haven't decided
which way I want to go based upon that pivot point.
So, let me just go back to that case for a moment.
First I would like to hear from Dr. Bigby. I think
I have already heard, but I need him to repeat,
that there was definitely a case of SJS but I would
like to hear again his opinion on the association
between the drug and that case of Stevens-Johnson

DR. BIGBY: My opinion about that reported
case is that it is probably a case of
Stevens-Johnson syndrome related to the drug. Now,
that doesn't mean that it is definitely related to
the drug. And, I think that the difficulty would
be for anybody to say with any certainty that it is
not drug related. But, you know, am I absolutely
certain that it was due to the drug? No. But I
wouldn't want to be put in the position to argue
that it is not drug related. That is the problem
we have.

DR. GOODMAN: I understand that the

investigator who treated that subject is here.
Could that person identify himself or herself?
Would you mind coming forward and just describing
your impressions of the case?

DR. REESE: Could you please be sure to
identify yourself, sir?

DR. BELNOR: I am Samuel Belnor, a
pediatric neurologist and I was the principal
investigator on this case that was a 7 year-old
Asian boy who is perhaps the most compelling case
for Stevens-Johnson. My impression on this
patient--and then I will go into detail, but my
impression was that the most likely diagnosis was
erythema multiforme, possibly Stevens-Johnson. The
most likely etiology was a viral infection,
possibly drug related.

The patient had shown improvement in the
clinical symptoms of ADHD after one week on drug
and was seen on the 14th day. On the 14th day the
patient presented with fever of 101.9, a sore
throat and feeling bad. I was out of town but the
sub-investigator, a pediatrician, saw the patient.

The mother complained of two lesions on the leg
which she thought were possibly a brown spider bite
but there was no rash. The next day the
sub-investigator put the patient on amoxicillin and
did a rapid screen for strep. which was negative.
The throat, he felt, looked like a viral throat
infection. There was no exudative pharyngitis but
papules in the throat. The patient was seen the
next day by a pediatric group locally. The
pediatric group saw typical lesions of Coxsackie B
virus in the posterior pharynx and diagnosed this
patient as having a Coxsackie B virus infection.
The rash was over most of the areas of the body but
it was more marked on the face and extremities.
Also, they felt that the two lesions on the legs
were the target lesions of erythema multiforme.

The patient did not develop any
apparent--there were no lesions in the mouth at
that time and no mucous membrane involvement.
About six days later--I apologize, we have a real
lack of data because the mother did not bring the
patient back to us until four weeks after the rash

developed, in spite of being called on numerous
occasions. She had a single family business and
was the only employee and would not bring him back.
He went to school many of these days. We told the
family the day of the rash to stop the drug. The
teacher recommended, on day 23, that she felt that
he should go back on the drug because his behavior
was much worse and the mother gave him one dose of
the drug and nothing really changed much except
that she felt that he was maybe pealing more and
did not give any more.

No one saw the lesions in the mouth, other
than the mother, and she thought that there were
lesions in the mouth because he would not eat well.
No physician see mucosal involvement. He did
complain of burning when he urinated, which is a

The patient really felt quite good during
this four-week period from the onset of the rash
until we saw him next. He went to school about
half the time. The mother was really unconcerned.
When I saw him four weeks after the rash onset I

saw no lesions in the mouth. There was no evidence
of any previous lesions in the mouth. He was
happy; no stress. And, his skin was pealing.
There was no evidence of any dermal involvement
other than just some pealing of the skin, mainly on
the extremities. There were no lesions in the
posterior pharynx of the Coxsackie B virus.

If we had seen the patient earlier we
obviously would have done a skin biopsy. We did a
RAS test later to modafinil and to amoxicillin--of
course, it is of limited value, but it did not show
any positive reaction.

DR. GOODMAN: I would like comments on
what you have just heard from either Dr. Rappley or
Dr. Bigby. Does that help one way or the other in
the diagnosis?

DR. BIGBY: Given the description, I don't
think that anybody can say that that was not a case
of SJS. You know, it would be nice to know if the
patient had typical targets or not but I don't
think you are going to get that described in this

DR. RAPPLEY: I guess for me it is the
degree of uncertainty that we have at every point;
it is sort of the added uncertainty that makes me
uneasy; that makes me unwilling to say that it is
just fine, let's go forward and treat everybody.

DR. GOODMAN: Can you repeat that?

DR. RAPPLEY: It is the degree of
uncertainty that we have that makes it difficult
for me to say that it is fine or perfectly
acceptable to proceed with just having people make
sure they report rashes.

DR. GOODMAN: Dr. Pine?

DR. PINE: I want to go back to the
statement Dr. Temple made. You obviously seemed
very taken with this when you said it is at least
going to get a black box and we moved away from you
fairly quickly. Could you just spell out your
thinking, what made you react that way? I mean, I
think it is more than just this one case or maybe
it is just this one case but I would like to hear

DR. TEMPLE: Well, it goes without saying

that I have no credible, sensible view about
whether this is a bona fide case or not. I am
listening to people who do though. So, what we are
seeing is that in something like 1000 people, but
perhaps when you look at exposure it may be 700 or
800, you have one case that is at least
statistically compatible with rates that are high
enough to be worrisome, you know, down to one in a
few hundred and up to whatever, and a condition
that is very scary and is life-threatening. So, it
all turns on believing the case. I mean, if this
was dismissable I wouldn't have said that but
everything I have heard up to now, both internally
and even from the company, says that this is a
plausible case.

So, when your best estimates of something
very worrisome are in the neighborhood of 1/1000,
you know, of it was agranulocytosis or something we
are accustomed to taking full note of those. That
is really all I meant.

DR. PINE: Yes, that is helpful. Thanks.

DR. GOODMAN: Dr. Mehta?

DR. MEHTA: Actually, I have two questions
of the investigator. One is was the patient
hospitalized? And, what surface area of the body
was affected by Stevens-Johnson syndrome?

DR. BELNOR: I am sorry, will you repeat
the first question?

DR. MEHTA: Was the patient hospitalized?

DR. BELNOR: No. Although most of the
areas of the body were involved, the total surface
area of the body involved, according to the
mother's history and the pediatrician that saw him
and our examination when he came back, was less
than 10 percent.

DR. MEHTA: Can I ask Dr. Bigby a

DR. GOODMAN: Sure, go ahead, Dr. Mehta.

DR. MEHTA: What percentage of
Stevens-Johnson syndrome patients would be

DR. BIGBY: Excellent question to which I
do not know the answer.

DR. GOODMAN: Ms. Bronstein?

MS. BRONSTEIN: I think there is another
signal we can't forget and that is the adult
population on the low dose having three cases in a
little over a million, which is two cases more in a
million than the general population on the same
drug. So, you know, even if there is some question
there is also some other linking stuff, at least in
my mind.

DR. GOODMAN: Dr. Laughren?

DR. LAUGHREN: Could we get some clarity,
maybe from Dr. Bigby, on what the background rate
of Stevens-Johnson is? Then, what the reporting
rate is in this experience with this drug in

DR. BIGBY: If you look at sort of
population-based studies the estimate is one case
in a million or 500,000. If you look at the
case-control study that was done in Germany, Italy
and France where they sort of specifically tried to
identify all of the cases over a period of time and
they took detailed drug histories from the patients
and they limited the definition to SJS and TEN the

way I defined it in my talk, it was in the order of
1/100,000 to 1/400,000.

DR. LAUGHREN: So, it sounds like it
varies anywhere from 1/100,000 to 1/million. Do we
have clarity on what the reporting rate is for
those three or four cases in adults? Maybe the
company would know that.

DR. CIVIL: Yes, for the person taking the
transcript, my name is Rich Civil, C-i-v-i-l. Our
reporting rate for events coded as SJS and TEN, the
number of cases we have has been discussed. There
are five. Each of them can be looked at
individually and, indeed, the discussions up to
this point have already excluded largely from
consideration one of the cases, that being the
patient with subarachnoid hemorrhage who developed
the cutaneous skin reaction in association with the
apparent initiation of treatment we phenytoin and
phenobarbital. Subtracting that case out, we have
four cases in approximately 750,000 adult
patient-treatment years of exposure.

Given the described hazard profile which

suggests a greater risk in the first four weeks
perhaps, we recognize that a better denominator for
that exposure would be patients rather than
patient-treatment years. Based on what we have
estimated as an average treatment duration of
approximately 2.5 months on average in the
postmarketing environment, we would then calculate
that the 750,000 patient-treatment years translate
to the rough equivalent, based on IMS estimates and
survey data, of approximately three million
patients treated.

DR. LAUGHREN: So, the reporting rate with
that denominator is roughly one per million. So,
there you have it. I mean, you have a reporting
rate of one in a million; background rate somewhat
less than that.

DR. TEMPLE: Well, that doesn't take in
the under-reporting.

DR. LAUGHREN: Right, but we usually
compare reporting rates to background rate,
understanding that there is under-reporting. We
generally take some comfort if the reporting rate

is well below the estimated background rate.

DR. PINE: I guess my question is what is
the downside of capping it at 1/1000 by studying
3000 more patients? As far as I can hear, the only
downside is that we are going to delay putting the
treatment on the market for six months or a year,
which seems like a risk worth taking if we really
want to be sure that, you know, 1/1000 is really
the risk of Stevens-Johnson. I mean, that seems
like a fairly fair trade, you know, to be sure that
the rate is really no higher than 1/1000 and we
delay approving a treatment for however long that
takes, six months or a year. DR. LAUGHREN: That
is precisely what we are asking the committee.

DR. TEMPLE: Right, and people have to
weigh the cost of the delay and the consequences of
not doing that.

DR .MALONE: Most of the talk has focused
on the one Stevens-Johnson. What about the other
case? I didn't quite understand the case where
there was urticaria. Was that thought to be drug
related or a signal of anything else related to

serious skin reactions?

DR. BIGBY: I think that that case is less
than 50 percent likely to be drug related.

DR. GOODMAN: If this were a new molecular
entity with no prior marketing experience and I was
presented with these data, with the degree of
uncertainty that we are all facing, I would say we
needed additional data, for sure. And, I think one
of the reasons that I have been on the fence in the
last hour or so is because it is an agent that has
been out there for a long period of time. But,
given the fact that it is at a higher dose and it
is going to be given to a population that
metabolizes it differently, perhaps I should be
taking it more as if it were new rather than a
different indication for the same compound in the
same population.

So, I would have to say I am leaning at
this point to recommend additional safety testing.
I don't feel, as I re-read this question, that
modafinil has been shown to be acceptably safe
given the doubts that we have in our minds. If it

turns out to be 2/1000 I think we would all regret
the decision to go forward. I don't want to do
that experiment in the postmarketing arena.

I think, that said, if we are to recommend
the studies--hopefully, the FDA would be the ones
to really design this--that we don't set the bar
too high. I don't want to be disingenuous. I
think that this is a drug that we all agree is
efficacious. There may be certain advantages over
existing compounds. Some of those are yet to be
proven. I would like to see an opportunity for the
company to come back with those additional data
that would give us an extra degree of assurance
that this case was a fluke, and that could exactly
be what it was.

DR. LAUGHREN: I think it is important to
be clear about what level of comfort we could
gather from the study that I proposed earlier. The
most you would be able to do is to cap the risk at
1/1000. So, even if you did that and you were
comfortable with that as a cap, I think the drug
would still have fairly strong labeling. I just

want to be clear about that. It is not going to
make the problem go away.

DR. PINE: Related to that, there probably
would be some discussion about, you know, let's say
you could do a larger study and cap the risk even
lower. You know, maybe people would want to do
that. There could be some discussion about that.
I think that is probably going a little far based
on the data we have right now. I think the
question is, is it safe enough or not and that is
kind of what we are debating.

DR. GOODMAN: Dr. Wang?

DR. WANG: I am a little bit less
sanguine. You feel that with a study of another
1000 patients maybe you will cap it at 1/2000. It
quantitatively gives you reassurance; it won't
qualitatively necessarily give you maybe the
reassurance we are looking for. One thing in favor
of additional studies is an active comparator, a
study that actually could maybe sort out sort of
where in the armamentarium this might fit in.

DR. GOODMAN: Dr. Bigby?

DR. BIGBY: Somebody has to clarify to me
then what is the black box labeling for Lamictal.
If they have a rate of 8/1000 in children what does
the label say?

DR. ANDREASON: Let me put it up.

DR. PINE: The other thing to remember
about the labeling for Lamictal is that it is for
Lennox Gasteau syndrome so it is a different
disorder fundamentally.

DR. TEMPLE: And for which I believe there
is no other treatment.

DR. PINE: There is no other treatment,
that is right.

DR. TEMPLE: It makes a difference. I am
confident if there were no other treatment here our
discussion would be different.


DR. ANDREASON: Here is the lamotrigine
black box, or at least the part that has the data
and the warning up front. This is for Lennox
Gasteau in kids and then adjunctive therapy for
epilepsy in adults and bipolar in adults.

DR. GOODMAN: Dr. Temple?

DR. TEMPLE: Well, I am sure you are going
to want people to discuss what you just said but I
just want to throw one other thing into the mix,
and that is, suppose the company did a study
showing definitively the way I want it shown that
it really did work in people who failed on other
therapy--a properly designed study, not that hard
to do if it really does work in that setting, would
that make any difference in all this?

DR. GOODMAN: It would make a difference.
It would definitely wind up with a black box.

DR. PINE: I don't think it is an
either/or though. I would want to cap the risk.

DR. TEMPLE: Well, that is what I am
asking. What you have just been discussing is
capping the risk before it gains approval for this
use. What I am asking is if, before doing that,
they knew that it unequivocally works in people who
failed on other therapy would that make you want to
make it available even before you capped the risk,
with an appropriate box, or not?

DR. PINE: You know, that is a theoretical
debate. I would have to see how well does it work;
does it really beat a stimulant head-to-head; who
are these people--

DR. TEMPLE: That is the test. It would
have to beat the drug they supposedly failed on,
presumably a stimulant, in a randomized trial and
it would have to beat it.

DR. PINE: That would be great. I mean, I
can't tell you that I would definitely say forget
about the risk if you show me that, but it
definitely changes the discussion we are having
right now quite appreciably.

DR. GOODMAN: I would echo Dr. Wang's
point earlier that not having the comparative data
I think is a weakness of this application.

DR. PINE: But, personally, I don't think
you have to have that. I mean, I think if it
worked just the way it does work and you knew that
the risk was 1/1000, again just speaking for
myself, I would be comfortable with that.

DR. TEMPLE: No, I just meant whether you

could truncate further characterization of the risk
if you knew that thing about it. Maybe that is
such a hard study nobody is even interested but you
don't know until you ask.

DR. GOODMAN: Dr. Bigby, has your question
been answered?


DR. REESE: Dr. Malone?

DR. MALONE: If you did such a study, then
would the label reflect that it was approved for
the treatment of patients who failed other
treatments or would it not include that in the

DR. TEMPLE: Excellent question. With
clozapine where we had those data the labeling said
you should have failed on other therapy because the

1.5 percent rate of agranulocytosis was considered
unacceptable in a first-line population. So, if
there were no further characterization of risk you
might very well say that it is for people who
failed other therapy, and maybe you wouldn't have
to wait for the further characterization of risk.

If the risk were then further characterized and
everybody was comfortable, maybe then we would feel
it could just be thrown into the mix and they would
have this particular piece of information.

DR. MALONE: How well do post-approval
registries help resolve a question like this
because it won't be that easily resolved? And, how
would you do that?

DR. TEMPLE: Yes, that is a hard question.
There are probably people better able to answer.
The most successful registries are ones where you
are controlling distribution so, for example, the
clozapine registry, in my view, is a huge success
because you can't get the drug without going to the
right pharmacies and your name goes in it, and one
of the purposes of it is to keep people who have
already gotten agranulocytosis from ever getting
the drug again and, as near as we can tell, it has
been very, very successful and there have been
analyses, but that is because you have to sign up
to get the drug.

Registries ordinarily in many other cases

are voluntary and whether people stick to them or
not is uncertain, and they have varying degrees of
success. We would have to get some people who know
more about it than I do to answer that though.

DR. MALONE: Currently, with stimulants
you almost have to see a patient fairly regularly
because you have to keep writing the prescription.
Could a registry be developed by requiring a script
from a doctor? At least they could ask if they had
a rash.

DR. TEMPLE: Well, any system that limits
distribution--first of all, it is very difficult if
the drug is already available in another form.
Second, they are a lot of trouble. I mean, we do
those things for drugs we are really worried about.
There is one being set up for Accutane that is more
rigorous than before; thalidomide--I mean, those
are the things we are talking about. You don't do
them lightly because there is actually some
evidence that they interfere with use. We have
distribution system for a drug called dofetilide
that is used to maintain normal sinus rhythm and a

study, I guess out of Duke, showed that people are
using solatol or quinidine instead. Well, that was
not what we had in mind. So, you have to fit it
into the system and it has to work out.

DR. GOODMAN: Dr. Malone, by this line of
questioning, are you suggesting that our target
would be to vote in favor of the compound but to
put in place a rigorous registry program to monitor
for rashes, particularly Stevens-Johnson?

DR. MALONE: No, I wasn't trying to
suggest that. I don't think it is going to be easy
to answer how often a rare event occurs if you do
more patients. So, I think in the end you are
going to have to have a longer way of answering
that question. I wasn't trying to suggest that you
would approve it and then handle it that way.

DR. GOODMAN: Any members of the committee
that would like to argue in favor of this being
shown acceptably safe, and we are focusing on the
dermatological complications?

[No response]

I would like to give a representative of

the company a chance to argue that point before we
take a vote.

DR. RUSSELL: I would just like to ask
Neil Shear to give his opinion on the risk of

DR. SHEAR: Well, I guess I can perhaps
add strength to your difficulty. The question of
this single case is exactly the way I would have
explained it, that there was a single case that was
sort of convincing. It didn't meet a definition of
Stevens-Johnson because the body surface area of
epidermal detachment was not high enough. It would
probably meet a definition of erythema multiforme
major, and it probably is post viral.

The other issue you can look at is it is
not 1/1000 because it was 10/10,000--it was one and
that one could easily be zero and that one could be
two. So, in terms of it perhaps being a fluke, I
think there is some strength to that argument.
Then trying to do the balance that you are talking
about, I think, you know, you have raised various
possibilities. I don't feel it is up to me to tell

you what to do on that. But keeping track of
reactions has been done before for other drugs.

I would also say that because of its
already accumulated experience, the pediatric dose
notwithstanding, this is not Lamictal. This is not
a drug that had started right from the
beginning--Lamictal, when it was started in England
out of Burroughs Wellcome, was causing problems
immediately and continued to cause issues. Now,
some of those are probably over-ascertainment
because people were jumping on the bandwagon in
terms of diagnosis but, still, it is a drug that
has a very different risk and I think that has been
managed over the years, actually many years now.
Here is a drug that was on the market. It is not a
new chemical entity but is being used in a broader
population in children so you have the balance

What I have seen so far has not convinced
me. I think where I would differ from Dr. Bigby is
that I don't feel that I can absolutely--and I
don't think he said absolutely, but I don't know if

I can really confidently say that there are going
to be cases of Stevens-Johnson/TEN with this. I
just don't see that based on the exposures we have
but that is, again, just personal after looking at
many of these drugs for many years. If you look at
dilantin, if you look at sulfonamide, they were
recognized in the '30s. When they first came on
the market it was clear that these drugs were
causing these kinds of problems right away.

I do want to make one more comment since I
have the microphone for a second, the sulfonamide
allergy story--for the severe reactions to
sulfonamide it is the aromatic amine at the end of
the molecule and not the sulfonamide moiety that is
considered to be responsible. There is certainly
no evidence to the contrary and the only evidence
that exists on a metabolic basis is that the
aromatic amine is hydrolyzed to a hydroxylamine
which goes on to become a nitroso, which is a P450
pathway through 2C19, and that is what appears to
lead in vivo and in vivo to toxicity.

DR. PINE: I would like to ask you a

question about your statement about where you would
disagree with Dr. Bigby. How confidently would you
assert that you doubt that we would see additional
cases? I understand that you said that there is
not a lot of evidence to support that.

DR. SHEAR: Well, I guess what I would do
is look at the cases that exist. What is real?
What am I comfortable with? We do see that there
are some cases in the adult literature. It is hard
to tell but, you know, we do have some numbers that
are low, like background, and they are in the
1/100,000 to 1/million type of range. Though
adults don't usually get Stevens-Johnson syndrome
that often, we do see it. We do see people come
in; they have no drug and they get a real
Stevens-Johnson syndrome. So, that is probably
out there. The pediatric exposures of at least
30,000 children--Dr. Andreason showed the numbers
for people who were getting the drug through
various programs, and they had none in 30,000
exposures. Again, if this case was rock-solid
Stevens-Johnson, which it doesn't appear to really

be by the usual case definitions that we use
nowadays, but if it is erythema multiforme major,
which is something that kids do get and something
that suggests viral from what we heard about the
case, the more you dig into this the more I am
getting more comfortable that it isn't. And, until
I had a chance to actually talk to the investigator
I don't think I would have been saying this, but
looking at it in its totality and trying to balance
it against the other known hard-core data, that is
what makes me more comfortable and I think I have
had that information maybe hours longer than Dr.
Bigby, but not much more, and I think you do get
more comfortable, and we sat down as a group of
experts to talk about it and we did become more and
more comfortable where that probably fit.

DR. REESE: Dr. Bigby?

DR. BIGBY: Do you have a response to the
question that was asked about what percentage of
patients with Stevens-Johnson syndrome get admitted
to the hospital?

DR. SHEAR: Yes, your answer was a good

answer; it is a good question. I mean, we do
sometimes see people who come in who we think have
Stevens-Johnson admitted to the emergency
departments, but I would say that if they actually
had some real epidermal detachment they would be
admitted not only to hospital but probably to a
burn unit. I mean, we are talking about some
pretty sick people and if you see a kid with truly
Stevens-Johnson syndrome, well, you are not going
to send them home. Unless you want to, you know,
not only potentially kill the child but end your
career, you are not going to do that. This is a
serious event and it is easy to recognized. This
is not a subtle diagnosis really. I mean, these
people have mucosal blistering that is not only
horribly painful but is hemorrhagic, and that is
not what we saw in this case and, again, we have
not seen any reports in the larger pediatric
population or the postmarketing surveillance.

DR. GOODMAN: Further discussion? Dr.

DR. MEHTA: I have worked in the drug

industry for about 40 years and I must say that I
have worked with a lot of different drugs and I
have seen during clinical studies about 20 patients
with toxic epidermal necrolysis or Stevens-Johnson.
I don't recall a single patient not being
hospitalized. Every single patient is
hospitalized. It is such a serious disease because
mortality now is about 5-15 percent. Ten or 20
years ago it used to be 50 percent. So, every
patient was hospitalized.

  DR. GOODMAN: Dr. Rappley, do you have a
  DR. RAPPLEY: All day we have dealt with

the uncertainty before us but now we hear a lot of
confidence that it is not Stevens-Johnson. I,
myself, am not changing my view on this.

DR. GOODMAN: Dr. Temple?

DR. TEMPLE: Well, let me offer a
suggestion or a question. My assumption is that to
the extent confidence that this case really
represented Stevens-Johnson, you would be more
comfortable with going directly to approval perhaps

with language in the labeling, and I don't think we
are going to be able to fully do that here. So,
let us tell you--you know, we might telephone you
or something, but we will look more at this. We
have experts around even though none of us
personally knows about it. If the case starts to
look very weak, that is going to change things and
i think we understand what you think about that.
But if the case stays reasonable strong, not 100
percent but reasonably strong then I think we have
heard your advice.

DR. GOODMAN: I find that acceptable. I
would like to call the vote on the question based

upon what we know now.
DR. RAPPLEY: Will you clarify what it is
that we are voting on?  
DR. GOODMAN: Has modafinil been shown to

be acceptably safe in the treatment of attention
deficit hyperactivity disorder in children and
adolescents? You have a comment, Dr. Andreason?

DR. ANDREASON: Yes, Dr. Luke had a
question about the case report versus the report

given by the investigator.

DR. LUKE: Yes, in the original written
case report it stated that it covered the entire
body. It was described very differently from how
the investigator describes it today. I think that
contributed somewhat to the relative uncertainty
that we are now hearing within the last half hour
or so. So, the question is what is the real story,
was it the written report provided by the sponsor
or is it the investigator's testimony given now at
today's meeting?

DR. BELNOR: I don't think we have changed
the story. The implication was that it was on all
areas of the body but it didn't cover every area of
the body completely. It was less than 10 percent
of the total body surface area. It was on the
trunk, the face, the extremities and the back.

DR. LUKE: So, you are saying the pealing
is less than 10 percent but perhaps the rash

DR. BELNOR: No, the rash. The rash was
around 10 percent by the history that we obtained

from the pediatrician.

DR. LUKE: Oh, so it is by history. You
did not yourself observe this?

DR. BELNOR: No, the peeling looked like
it was obviously less than 10 percent when we saw
the patient.

DR. LUKE: So, then there is still some
doubt. It is really hard for a dermatologist, and
I know other dermatologists in the room can attest
to it, to make an assessment from hearing a story,
especially if it is not carefully written up.
Photographs are often helpful and biopsies are
helpful but, again, it is lack of information that
adds to uncertainty.

DR. GOODMAN: Let's go ahead with the
vote. I am going to start with Dr. Bigby.

DR. BIGBY: So, is this a yes or no

DR. GOODMAN: Or abstain.

DR. BIGBY: I would say yes, it is
acceptably safe.

DR. GOODMAN: And explain your reason.

DR. BIGBY: You know, I think that this is

an instance where we are being asked to make a
decision on the basis of a single case that is
probable but not definite. I mean, I have concern
that when the drug is more widely used over a
longer period of time you are going to see cases of
SJS but you see that with lots of other drugs that
are already marketed.

DR. GOODMAN: Before I go on with the
vote, I actually expected a different response and
I am assuming others did too. So, maybe there is
room for further discussion, given the opinion you
just rendered, before we go on with the vote. Does
that change anybody's mind around the table?

MS. BRONSTEIN: I have one question of the
investigator. It was my impression that you did
not see this patient yourself until four weeks
after the very final time the mom brought the child
in. Is that correct?

DR. BELNOR: The mother refused to bring
the child back in from the second visit until the
last visit.

MS. BRONSTEIN: My question is when did
you lay eyes on the patient.

DR. BELNOR: I saw the patient at the
first visit.

MS. BRONSTEIN: Before the rash?

DR. BELNOR: Before the rash.

MS. BRONSTEIN: And did you see the
patient on the last visit?


MS. BRONSTEIN: But not when the rash was
in its fullest--

DR. BELNOR: No, none of the investigators
saw the patient when the rash was present. We told
the referring doctor to stop the medicine and send
the patient to us for a biopsy.

MS. BRONSTEIN: And who did the write-up
of the patient that was received?

DR. BELNOR: I did.

MS. BRONSTEIN: The first write-up that
was received to the company?

DR. BELNOR: I did most of the write-ups.
I don't know. There are a lot of errors in the


MS. BRONSTEIN: Thank you.

DR. GOODMAN: Do you have any further

MS. BRONSTEIN: I am left with a lot of
questions and a lot of lack of confidence, and I
feel like erring on the side of conservatism,
either longer testing or saying no but as the
consumer representative I feel like the public
needs to be protected and we have a lot of
questions here. As a working mom, I really can
relate to this mom not bringing the kid in; I did
it myself. And, I don't know that you are going to
get good anecdotal reporting. I also don't have a
lot of confidence in non-dermatologists reading
rashes. So, that is where I am with all this.

DR. GOODMAN: Dr. Wells?

DR. WELLS: While it may be true that the
case for lack of safety has not been made, it is
also true that the case for safety has not been
sufficiently made, and I think that is what we have
to have in order to make a statement that it is

adequately safe and I am not there.

DR. GOODMAN: Dr. Pfeffer?

DR. PFEFFER: Am I voting or making a

DR. GOODMAN: You are just making a

DR. PFEFFER: I wanted to ask, I just
began to remember, isn't it true, Dr. Bigby, that
this problem, Stevens-Johnson for example, doesn't
always appear on the entire body simultaneously?
Isn't there a course that goes from head to foot?
So, I was wondering about this case. How
frequently did the pediatrician see the child once
the rash occurred?

DR. BIGBY: So, Stevens-Johnson does
normally evolve over a period of several days, and
it is true that it is not full-blown at its onset.
People can continue to get lesions over several
days. I would say in the majority of cases you
start getting new areas of involvement after about
a week or so.

DR. GOODMAN: Dr. Robinson?

DR. ROBINSON: Well, I think it is very
striking that the one case that we are debating
about came in the context of somebody who was in a
controlled trial. Even in a controlled trial,
which wasn't designed to look at this issue, we are
in the situation where experts can debate back and
forth, and I think that says that if we approve the
drug and say there is going to be postmarketing we
would not get data that was really usable because
even in a controlled trial we are debating. I
think that argues for us getting a study design to
look at this specifically so if somebody has a
suspected case of it the proper information is
obtained, like photographs and expert dermatologic
consultation so that we can actually say what is an
estimate. It is just striking that even in this
sort of controlled trial we are not getting the
information we need and I think that argues for a
specific study.

DR. GOODMAN: Dr. Armenteros?

DR. ARMENTEROS: I also have a concern
that I am not so sure that even a controlled trial

moving forward from this point would still resolve

our doubts. I am concerned about that.

DR. GOODMAN: Dr. Andreason?

DR. ANDREASON: I suppose for something
that is as rare as Stevens-Johnson or, say,
something like acute liver failure you don't even
need a controlled because the historical control is
so rare that if you pick up a case in an open-label
trial of, say, 3000 patients that is significant.
So, that would be an acceptable design to look at
something like this.

DR. GOODMAN: Let's start with the vote
again, and this time I will begin with myself. I
am going to vote no. I have been persuaded by my
colleagues around the table and my comfort level is
not sufficient that this has been shown to be
acceptably safe. I don't know what to make exactly
of that one case and, frankly, I don't think we are
ever going to be sure. It certainly raises a
sufficient number of doubts about a serious adverse
event that should not have occurred even at the
rate of 1/1000 or less that we saw in this trial.

Perhaps the other factor that has led to
my decision is the absence of other strong,
convincing reasons to consider this drug having
advantages in other areas of safety or tolerability
or efficacy so I am not willing to find the risk
acceptable of going forward without additional data
that would rest some of my concerns about the
dermatological reactions. Now we can go back to
Dr. Bigby.

DR. BIGBY: I voted.

DR. RAPPLEY: I do not think it is
acceptably safe and I think you all have
articulated my feelings.

DR. WANG: I think it is just unknown.
Can I abstain until we have more information? I
mean, it could be everything from this things
shouldn't be approvable if this is a real signal to
there is no warning needed at all if this isn't the
case. We don't even know what to make of this
case. There is no temporal or inter-rater
reliability even within this meeting.

DR. GOODMAN: Dr. Laughren?

DR. LAUGHREN: Actually, I don't

understand an abstention in this situation. I
mean, we are asking you if you feel there is not
enough information to make a judgment, then I think
the answer would be no.

DR. WANG: No, okay. It is no, we don't
have enough information.

MS. BRONSTEIN: My vote is no unless more
information is obtained.

DR. PINE: I guess I will make two
statements. I found Dr. Temple's statement about
you will look into it and, the more doubtful this
diagnosis becomes, everything changes, and I would
agree with that. You know, just sitting here today
it has to be obvious to anybody that knows nothing
about Stevens-Johnson syndrome that there is a
reasonable suspicion. I think everybody would
agree with that, that there is a reasonable
suspicion that this was a case of Stevens-Johnson
syndrome related to the medication exposure. So,
that is the first thing.

The second thing is that I really don't

think it is that big a deal to cap the risk at
1/1000. So, I am going to vote no and what I would
recommend is a study of 3000 patients that is not
an efficacy study, that is simply designed to make
sure that there is not a single case of
Stevens-Johnson syndrome, you know, treated for a

DR. LEON: I will vote no, based on the
data we have seen that modafinil has not been shown
to be acceptably safe for children and adolescents
with ADHD.

DR. ROBINSON: I am voting no because I
think that we do need a study specifically designed
to at least get a good estimate of what the rate
is, and especially in a therapeutic area where
modafinil hasn't shown a specific efficacy that is
greater than with the already available agents.

DR. PFEFFER: I am voting no also. I
think that we need more information which I think
will be extremely helpful in guiding not only the
clinician but enhancing perhaps compliance of
patients. My feeling is that if this were approved

now, regardless of how the clinician might feel or
try to explain it, I think the compliance of
parents for the children would not be as good
perhaps than if there were a clearer view of the
risks where they could make a more informed
decision. I think we need more data and I think
that it is worth that wait.

DR. ARMENTEROS: Well, based on the
confusion that I have been exposed to through the
whole day, I am going to vote no and I am hoping
that given this ADHD diagnosis we can identify
readily and do studies to bring on the data. We
are not talking about a condition that is rare so
we should be able to move ahead at a later stage
with much more clear information that in everyone's
mind will be better at that stage.

DR. WELLS: Barbara Wells, and I will vote
no. I don't believe the case for safety has been
adequately made and, in addition, I don't believe
we were convinced that it is more effective than
available treatments and perhaps not as effective
as available treatments. We also have reason to at

least suspect that the incidence of even the common
side effects is higher with this drug than with
available treatments.

MS. DOKKEN: Deborah Dokken, I also vote
no on the question of safety. I mean, the
uncertainty about all of this today has been almost
painful and on those grounds I think we do need
more information before we can put it out for the

DR. MALONE: I vote no also. I think that
the potential population who would get the drug is
fairly big, especially considering the safety risks
that we have been talking about today and the
apparent lack of any safety advantages for this

DR. MEHTA: I know I cannot vote but if I
were to vote I feel like the California voter in
the presidential elections where my vote doesn't
count because it has already been decided. Anyway,
let me make a couple of comments. One is that I am
not convinced that this is a patient with
Stevens-Johnson syndrome. I have heard enough

discussions and I have been faced with and reviewed
patients like that. I am not a dermatologist but
still I do not believe that this is a case very

Secondly, the case for the dose
relationship, that a higher dose will lead to a
higher incidence of Stevens-Johnson syndrome
certainly has not been made. If that is the case,
then one should use as a denominator something like
3 million patients. So, we have about four or five
patients with Stevens-Johnson syndrome in an
exposure of 3 million people, which is no different
than anything else. So, from my point of view, if
I had to vote I would have voted differently with a
lot of strictures about how to get more data to
make sure that the real incidence is not more than
what we already see.

DR. GOODMAN: Could you give us the tally?

DR. REESE: Yes. There is one yes and 12
no. Going back to question one, it was 12 yes.
Dr. Andreason?

DR. ANDREASON: Just for note-taking, I

heard some discussion that a risk cap at 1/1000
would be something that you would like to know
about. Did you want to talk about that or make a
vote on that?

DR. PINE: Again, I would emphasize that
really the question of efficacy is not on the table
here, that we have been convinced of efficacy and I
can imagine that a fair amount of time and energy
and patients experiences have already been invested
in doing that. I would not want to reinvent the
wheel there. I think the main thing would be to
know definitively what the risk is from a ballpark
sampling. Again, just speaking for myself, if we
were to see open-label treatment, treated by
pediatricians who are seeing patients regularly,
that there was not a single case that would raise
any dermatological concerns about Stevens-Johnson
syndrome in 1000 cases, then I would vote yes.

DR. REESE: Dr. Rappley?

DR. RAPPLEY: I would support that, and I
think it is clear that the medication is
efficacious and the comparison studies can be done

postmarketing and I would be happy with that.

DR. GOODMAN: What I am about to say is
not necessarily a criticism of this particular
sponsor but I think there is a lesson in here about
the need for better assessment of these
dermatological adverse experiences, and I think a
lot of these issues would have been settled and
perhaps even the outcome would have been different
if we had better documentation that would have
allowed our dermatological colleagues to make a
more definitive conclusion. So, I think we are
dealing with some fuzzy information but, given that
this wasn't a compelling enough story here, both on
the efficacy side and on the safety side, to reach
a comfort level by which this committee could
endorse this compound moving forward to market.
So, I think we did err on the side of consumer
protection and I would hope sincerely that the
company would find the means by which it could
gather the additional data to collect the necessary
safety data and the outcome could be different
under those circumstances.

I want to thank everybody for attending
for the long day, and most of all for putting up
with my scratchy voice. Thank you.

DR. LAUGHREN: And I want to thank the
committee again for a heroic effort in helping us
with our job. Thank you.

[Whereupon, at 4:25 p.m., the proceedings
were adjourned.]
- - -