ajh 101

1 public health problem.

2 Successful intervention depends on pharmaceutical

3 and psychotherapeutic treatment approaches', as well as a two

4 fold approach to education in professional and treatment

5 settings, as well as in the patient population and general

6 public.

7 Since primary care physicians and community mental

a health staffs are most likely to see people with PTSD first,

9 they must learn to ask about trauma exposure, recognize the

10 symptoms, and treat or refer patients appropriately.

11 Educating professionals first is xramount to

12 managing the influx of clients that will certainly follow

13 the public awareness programming that might come with this

14 indication for PTSD medicine.

15 Thank you.

16 DR. TAMMINGA: The committee appreciates your

G 17 &.&arks, Ms. Giller, and thank you for appearing before us.

ia The second public speaker we have is Ms. Bonnie

19 Green, who is representing the..'??!ternational Society for

20 Traumatic Stress.

21 Ms. Green.

22 DR. GREEN: Good morning. My name is Bonnie

23 Green. I am a Professor of Psychiatry at Georgetown

24 University Medical School.' I am here today as president-

25 elect of the International Society for Traumatic Stress

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 102
1 Studies, the ISTSS, which is an international organization

2 of approximately 2,500 mental health professionals who study

3 and treat survivors of traumatic events.

4 I am here to speak today on behalf of the ISTSS to

5 the importance to posttraumatic stress disorder, PTSD, as a

6 public health issue, and of the necessity of identifying

7 treatments for this potentially debilitating disorder.

8 PTSD is an anxiety disorder that is experienced

9 following a traumatic life event. An event that can

precipitate PTSD is usually a direct or indirect

confrontation with death, or with serious bzly injury,

which produces an overwhelming experience of fear,

helplessness, or horror.

Traumatic events, such as rape, assault, domestic

violence, accidents, and disasters are, unfortunately,

relatively common in the general population. Estimates are

t:.& one-half to three-quarters of Americans have

18 experienced a traumatic event in their lifetime.

.-.
Individuals with a PTS6 diagnosis following such

events reexperience th aumatic event in a number of ways

including intrusive recollections, having disturbing dreams

about the event, and becoming very upset when they are

23 reminded of the event.

24 Trauma survivors with PTSD also try to avoid

reminders of the event, theytfeel emotionally numb, and they

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 103
1 lave difficulty being close to others. Finally, PTSD

2 involves symptoms of physiological arousal, such as

difficulty with sleep and concentration, exaggerated startle

response, and hypervigilance. These symptoms can cause

substantial disability and disruption of interpersonal

relationships.
PTSD is a relatively frequent disorder in the

8 general population. There have now been several studies in

.
9 Jeneral community samples including a very large

10 epidemiologic study of over 8,000 people between the ages of

11 15 and 54 in the United States, the Nationaji_Comorbidity

12 Survey, which I am sure you have already heard about this

13 norning, that have assessed exposure to traumatic events and

14 10 PTSD.

15 In spite of very different methodologies, these

16 studies have produced remarkably similar estimates of the

c. -.7 +&valence of PTSD in the general population. Specifically,

18 this diagnosis occurs on a lifetime basis in about 10 to 12

19 percent of women and 5 to 6 percgnt of men.

\
20 Point preval estimates, estimates of who would

21 have PTSD at any given time are about 5 percent for women

22 and 2 to 3 percent for men in the United States.

23 Heidi Resnick and her colleagues, in their

24 national study of women, estimated that nearly 10 million

' 25 women would have PTSD at some point in their lives, and that

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 104
1 wer 4 million had PTSD at the time of the study.

2 If left untreated, PTSD can last for decades.

3 decent studies have found high prevalence of PTSD half a

4 :entury later in Holocaust survivors, World War II

5 combatants, and prisoners of war. The National Comorbidity

6 Survey found that among those people who developed PTSD

7 Eollowing a traumatic event, one-third of them continued to

8 aave the diagnosis 10 years later.

9 In addition to the mental anguish that PTSD

10 causes, it also contributes significantly to problems with

11 physical health, as Esther just mentioned. Studies have

12 oeen accumulating for the past decade that have documented

13 the relationship between exposure to traumatic events and

14 increased levels of physical health complaints, physical

15 illness conditions, physician diagnosis, visits to

16 physicians, and cost of health care.

*, '7 / Only in the past few years, however, have

18 researchers begun to investigate the mechanisms for these

,-
19 relationships. It turns out that there is convincing

20 support for PTSD as an ortant link between trauma and

21 poor physical health.

22 This means that among those traumatized in various

23 ways, it is the development of PTSD that predicts poor

24 health and higher utilization of care. PTSD also impact in

25 the economic realm, with findings from a recent study

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 105

1 ndicating that having PTSD is associated with high rates of

2 inemployment and with significantly lower wages. The

3 zombined impact of PTSD on emotional, physical, and economic

4 tell-being, therefore, makes it a significant public health

5 lroblem.

6 While PTSD has been associated historically with

7 :ombat trauma, studies in the past decade have clarified its

8 frequency in non-military populations, indeed, PTSD occurs

9 nost often outside of military settings.

10 In the National Comorbidity Survey and in other

11 studies, PTSD was most likely to develop inzth women and

12 nen following rape. Physical abuse was very likely to lead

13 :o PTSD in both genders, as well.

14 For women, being sexually molested and being

15 threatened with a weapon were also important predictors of

16 ?TSD. Since PTSD is more common in women than in men, it is

*. ._'7 ,Aar that PTSD is an important concern,'not only for

18 nilitary veterans, but for civilians in all walks of 1 ife.

19 3h other psychiatric
PTSD often coexists wl

20 disorders. The Nation omorbidity Survey found, for

21 example, that half of men and women with a lifetime history

22 of PTSD also had a lifetime history of major depression.

23 However, although anxiety and depression often coexist with

24 PTSD, in recent years, it has become increasingly cleat that

25 PTSD has a distinct neurobiology that can been

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 106
differentiated from depression.

Some of the more compelling evidence includes the

3 observation that levels of the stress hormone cortisol are

4 lower than normal in PTSD, whereas, they are consistently

5 higher than normal in major depression.

6 Moreover, in PTSD, the negative feedback

7 inhibition of cortisol, which regulates the sensitivity of

8 the stress response mechanism in humans and animals, is

9 altered in such a way as to produce an increase

10 responsiveness to stress.

11 This has been established with nuzrous studies

12 demonstrating an increased sensitivity of the glucocorticoid

13 receptor, evidenced by an exaggerated cortisol suppression

14 lfollowing dexamethasone administration, and an augmented

15 ACTH response to the cortisol inhibitor, metyrapone, in

16 PTSD.
** 17 ./ In contrast, depressed individuals typically show

18 a decreased sensitivity of the glucocorticoid receptor as

,<c-'$
19 evidenced by escape from dexametnasone suppression. This

20 evidence strongly sug that PTSD is a distinct

21 psychiatric disorder. .
22 Although PTSD first appeared in the Diagnostic and

23 Statistical Manual of Mental Disorders of the American

24 Psychiatric Association as'recently as 1980, there is

25 already enough preliminary information about potentially

MILLER REPORTING COMPANY, INC.

507 C Street,.N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 107
1 efficacious strategies to warrant publication of a Treatment

2 Guideline for PTSD, spearheaded by our organization, the

3 ISTSS.

4 This guideline reviews different treatment options

5 for PTSD including both psychotherapeutic and pharmacologic

approaches. As the guidelines indicate, there are indeed

efficacious treatments for PTSD.
With regard to medications, the SSRIs, in

9 particular, appear to be frequently used in clinical

10 practice and are well tolerated by patients. Medication

11 trials have indicated that the SSRIs are asziated with

12 reduction of symptoms in all of the PTSD symptom clusters,

13 reexperiencing symptoms, numbing symptoms, and physiological

14 arousal.

15 In closing, the ISTSS wishes to be present today

7
16 to speak to PTSD as a significant public health problem, and

i -7 t-,+nderscore the importance of developing effective

18 treatments for it.

19 ._ We believe an approve?medication for PTSD would

h
20 serve to encourage the lit to seek and receive treatment

21 for this disorder, and would add significantly to our

22 treatment options when we treat patients suffering from this

23 serious health condition.

24 Thank you.

25 DR. TAMMINGA: Thank you, Ms. Green, for speaking

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 108
1 :o the committee and sharing your concerns with us.

2 We have a third speaker here this morning, Ms.

3 Jerilyn Ross from the Anxiety Disorders Association of

4 qmerica.

Iis. Ross.
MS. ROSS: I am Jerilyn Ross. I am president of
zhe Anxiety Disorders Association of America, or ADAA. I am
director of the Ross Center for Anxiety and Related

9 lisorders here in Washington, and I am author of a book

10 called "Triumph Over Fear."

11 Thank you, Mr. Chairman, and membz of the

12 2dvisory Committee for the opportunity to speak to you here

13 this morning.

14 For those of you who don't know us, the ADAA is a

15 national nonprofit organization, and we are dedicated to the

16 aarly prevention, identification, and treatment of anxiety

-. '.7 l'rorders. We were established in 1980, and we are a

18 partnership of researchers, clinicians, patients with

19 anxiety disorders, and their f&?ly members and other

't&J-
20 interested individuals.-*

21 Together, we work towards the prevention and the

22 cure of anxiety disorders by supporting research and by

23 helping consumers gain early access to diagnosis and

24 treatment. We also seek to reduce stigma, we stimulate

25 ongoing research, and we educate health care professionals

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 109
1 and consumers about effective treatment.

2 Increasing access to safe and effective treatment

3 for people with posttraumatic stress disorder, as well as
4 with other anxiety disorders, is a major concern of our

5 organization, and therefore we hope for a positive outcome

6 to your deliberations today.

7 I am here on behalf of more than 19 million

8 Americans who suffer from an anxiety disorder, specifically,

9 today, the 8 million Americans who suffer from PTSD,

10 posttraumatic stress disorder, which is a severe and

11 potentially debilitating mental health problem.

12 People with PTSD come from every walk of life,

every social class, every educational level, and every

professional achievement. These are people who have been

15 exposed to an extreme trauma, maybe an accident, a natural

16 disaster, been raped, criminally assaulted, or exposed to

4. '.7 c &at or physical or sexual abuse.

18 And these are people who may at one time have been

19 hea-lthy, productive individual&?who,now, following exposure

20 to this trauma, are su ing real life-altering, but

21 treatable disorders.

22 People suffering from PTSD reexperience the

23 traumatic event in the form of flashbacks, nightmares,

24 intrusive, distressing recollections, and they develop

25 avoidance behavior, they develop increased arousal, and

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 110
1 iumbing, where they can't feel anything, positive or

2 iegative, emotionally.

3 They become vulnerable also to secondary problems,

4 panic attacks, depression, substance abuse and suicidal

5 thoughts and attempts, just to name a few, and many of them,

6 nost of them are unable to receive an accurate diagnosis for
7 their illness, and many of them end up being dismissed as

8 hypochondriacs or eccentrics or malingerers without getting

9 of the help that they need and so desperately deserve.

10 Each year at ADAA, we receive tens of thousands of

11 requests for information from people with an&ety . disorders.

12 As a matter of fact, we are,currently experiencing more than

1 3 43,000 people per month who spend a minimum of 10 minutes on

14 our web site seeking information, and we also get letters

15 and phone calls from people who describe their heart-

16 wrenching pain, their suffering, as well as their fear,

c :7 LLRir confusion, and their despair.

18 What we hear, what we find most frustrating from

19 these people is that they are r&F able to find health

20 professionals in their munities who are both

21 knowledgeable about anxiety disorders, and able to provide

22 effective treatment. <Sadly, at this time, particularly for

23 PTSD patients, there'are no approved medications and

24 millions of people with PTSD are suffering, Oftentimes

25 silently in the dark, with ignorance, frustration, and

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 111

shame.

According to a study that we published recently in

3 the Journal of Clinical Psychiatry, called "The Economic

4 Burden of Anxiety Disorders in the 199Os,ll PTSD was found to

5 be one of the two anxiety disorders with the highest rates

6 of risk factors for psychiatric service usage. PTSD was

7 also among the anxiety disorders associated with most

8 substantial impairment in workplace performance.

9 The good news is that thanks to new scientific

10 understandings of the biochemical component of PTSD, and

11 studies demonstrating the efficacy of speciz biological

12 and psychological treatment, things are beginning to change

13 for the better.

14 Our association has joined with other mental

15 health professional, as well as with other advocacy groups,

16 in hopes of spreading the word, getting the word out that

+, - 17 I&D is a bio-psycho-social disorder that is real, that is

18 serious, and that it is treatable.

e
19 ._ Improving physician e&cation about PTSD and

20 increasing the availad% y of safe and effective

21 medications, as well as of psychological treatments, are

22 vitally necessary, so that those suffering from PTSD are

23 better able to manage their illness and go on to lead full

24 and productive lives.

25 We have seen the difference that this has made as

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 112

1 effective treatments have become available for other anxiety

2 disorders , like panic disorder, obsessive compulsive

3 disorder, most recently for social anxiety disorder, and I

believe that your deliberations here today can contribute

greatly to achieving the obj,ective that all people with

anxiety disorders can get the diagnosis and the treatment

7 that they need and deserve.

8 I thank you very much for your consideration

9 today.

10 DR. TAMMINGA: Ms. Ross, thank you very much for

11 your remarks to the committee.

12 Advisory Committee Discussion and Deliberations

13 With this presentation, we conclude the open

14 public hearing portion of our meeting, and we begin the

15 Advisory Committee deliberations about sertraline for PTSD.

16 We have heard this morning from Pfizer, who

*. '.7. p+duced th e data about sertraline, you had an opportunity

18 to ask them some questions. We have heard from the FDA

. :: J&
19 about their analysis. _
20 We have a nu * of questions in front of us by

21 the FDA, and the questions that the committee has in front

22 of us today are questions that are not only about safety and

23 efficacy of the compound for the indication, but actually

24 questions about the indication itself, those questions that

25 Laughren posed to us earlier.

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 113

1 I would like to suggest that the committee begin

its deliberations by addressing some of those questions

shout the diagnosis that Dr. Laughren put to us, about PTSD

BS a new indication, how widely recognized and accepted is

:he entity, can PTSD be considered an independent diagnosis,

and then more practical questions about actually doing
studies in PTSD.

8 I wou.ld like to invite the committee to begin a

9 discussion on that.

10 DR. DOMINGUEZ: I will make a very general

11 statement to begin with. It is a disorder tit is hard to

12 ignore, although I think refinements will continue to take

13 place in the definition of the disorder. I think that there

14 are clusters of symptoms that are distinct enough that

15 indeed it is recognized within our field.

16‘ so, I would like to immediately begin by

c - -7 &&essing my opinion that yes, this is a distinct disorder

18 where we should be seeking specific forms of therapy, both

.L'T. ,s
19 psychosocial and pharmacotherapy. Ihave no problem with

20 that.

21 DR. TAMMINGA: Thanks. We have three PTSD experts

22 here, and perhaps the committee could hear from them.

23 Dr. Southwick.

24 DR. SOUTHWICK: I also feel this is a distinct

25 disorder that has had a very'long history and gone by many

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 114
1 different names throughout history, and many of the early
2 names were derived from combat experiences like shell-shock,
3 irritable heart of soldiers, combat fatigue, et cetera, and

4 as DSM was formed in 1980, PTSD became a formal diagnosis,

and I think what we have seen, although some of the symptoms

have changed since DSM-III, they are relatively stable

1between DSM-III-R and IV.

There has been really very little change with the

9 core symptoms, suggesting that with experience and research

10 and clinical input, that the disorder has been more

11 carefully and rigorously defined over the lz number of

12 years.

13 Also, as mentioned earlier, there are a number of

14 very distinct PTSD symptoms, I think there are eight, that

15 are specific to the trauma, which helps to differentiate

16 from other comorbid diagnoses.

DR. TAMMINGA: Thank you.

18 Dr. Brewerton.

DR. BREWERTON: yes. .'.':3 think among the questions

19 ..-
20 that are posed to us t this is probably the easiest

21 one. In my mind, there is no doubt that PTSD exists. It

22 certainly fits with all of my clinical experience, and I

23 think also the science is at a point now that does, in fact,

24 confirm its existence and distinction from depression.

I would add, amongjthe comments made already, that

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 115
1 :here are several psychiatric disorders that have

2 Jignificant overlaps with major depression. Certainly PTSD

3 -s just one of many.

4 We think about the anxiety disorders, eating

5 disorders, substance use disorders, somatoform disorders,

6 dissociative disorders, personality disorders, all of those

lave strong and important links to depression, but yet

remain as fairly distinct entities, and I think PTSD is just

{et another that fits that bill.

10 so, I would think that this is the easiest

11 question and I think the overwhelming evidexe is in favor

12 of its independent existence.

13 DR. TAMMINGA: Dr. Brewerton, would you comment a

14 little bit more on the nature of the evidence that it is an

15 independent disorder?

16 DR. BREWERTON: Well, I know at the Medical

4 17 Ldversity of South Carolina, in the National Crime Victims

18 Research and Treatment Center dataset, which is in reference

19 "'?oday.by Heidi Resnick,

to..one of the studies mentioned‘
.a
20 which is the National n's Study, which included over

21 4,000 women randomly selected across the United States,

22 there have been detailed cluster analyses of the symptoms

23 generated from this study, clearly again showing the links

24 between PTSD and depression, but that they do, in fact,

25 separate out in terms of factor analyses as clustering

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 116

1 Logether and separately.

2 I know that there are a number of other studies

3 like that, that show its independence.

4 DR. NORTH: There is considerable controversy

regarding the diagnosis of PTSD among clinicians, and I

:hink part of the force responsible for that is the

zomorbidities and confusion among diagnoses and the

preexisting disorders, but I myself come from epidemiology

9 If disasters, and I can say that what we see after disasters

10 often appears very different from much of what we see in

11 other populations, and that is because we ca&study PTSD in

12 2 more pure form after disasters because in other

13 populations, PTSD is confounded with vulnerability to a

14 traumatic event, whereas, disasters select populations

15 actually unselected for previous psychopathology.

16 In this setting, I can say after interviewing very

4.. _ ,7 n,,$iy disaster survivors, that I have seen many people with

18 PTSD without any previous or coexisting comorbidity, and I

"'3
19 am--definitely a believer in PT& from-my own research

20 experience, and I beli that PTSD looks different in

21 different populations, and that may be a source of the

22 disbelief among many clinicians, but in my experience as a

.2 3 researcher and a clinician, it is apparent to me from the

24 data and from clinical experience that this is an important

25 disorder. I

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
 ajh 117
1 DR. KATZ: I have a related question. Everything

2 )r most everything we have seen today, both in terms of the

3 .iterature, the previous literature that was discussed, and

4 :he data that we have had presented from this application

5 suggests that veterans, people whose particular traumatic

6 event was war or combat, don't seem to respond to treatments

7 :hat perhaps others do respond to, raising the question as

8 :o whether or not that is a fundamentally different thing,

9 lrhether that is a variant of PTSD, and it raises the sort of

10 generic question of does the event, does the specific

11 :raumatic event ‘have anything to do with whswe are calling

12 ?TSD. I just wonder what people think.

13 DR. HAMER: That is an interesting question, and I

14 :hink it relates directly to what happened in the clinical

15 :rials here. There a number of events or characteristics

16 :hat are clearly very confounded, at least in the databases

I... ^ -'7 ,/have - being a veteran, being male, the type of trauma,

18 zhe age of exposure to trauma, and the length of time since

..::.B
19 trauma and the duration of reported PTSD.

20 We are focus you are focusing at the moment on

21 :he veteran versus non-veteran issue. Pfizer and the

22 reviewers tended to focus on the gender issue, but to some

23 extent it could be any of them. I mean it could be that if

24 we had the data to find a cohort of males who had been

: : 25 sexually assaulted at roughly the same age as the women in

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 118
1 our cohorts have, and it has been that duration of time

2 since the assaults, we could find a similar pattern in terms

3 of response to medication or we might not, but given the

4 data that we have at hand, trying to separate out gender

5 effect, trauma effect, veteran, duration, age, and all that

sort of material seems to me to be fundamentally difficult.

DR. TAMMINGA: I am wondering if people who have

had experience with treating PTSD veterans versus non-

veterans, or combat trauma, could speak to that.

10 Dr. Southwick.

11 DR. SOUTHWICK: I think it is a czplicated issue.

12 One factor is in combat, one is typically exposed to

13 multiple repetitive traumas that may last, go on for years

14 or a year or whatever, so that one of the most important

15 questions I think is looking at the nature of the trauma,

16 how repetitive it is, that sort of thing.

4 :7 A It is also true that how you sample, which

18 patients you select, I think is very important because if

19 the patients are selected from.'cfie VA.now, 30 years later,

20 as opposed to a communi sample, advertising for veterans

21 who may have some of these symptoms, you may see a different

22 response because most of the veterans who are coming to the

23 hospital now have very severe PTSD and have been coming for

24 a long period of time, and have probably been in treatment,

25 and that sort of thing. I

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 119

1 so, I think sampling is very important issue. I

2 am not totally convinced that veterans do not respond to

3 medications. For example, 15 years ago, there were some

4 studies done on veterans at outreach centers, and so forth,

5 and they had not been in treatment for as long a period of

6 time, and I think some of those results were more promising.

7 DR. BREWERTON: I definitely agree it's a most

8 complicated issue, and another factor that I wonder about is

-
9 the issue of service-connected disability and what

10 percentage of the veterans had service-connected disability,

11 which becomes a disincentive to improvement.

12 DR. HAMER: That is another confound. I would

actually be curious to ask Pfizer, since I haven't seen the

14 protocol, what kind of exclusion criteria there were for

15 either involvement in some sort of a legal process, that is,

16 ,whether a lawsuit was ongoing, or whether the subject was

‘Ir '7
i li&eiving or about to receive some sort of disability

18 payment that would be an incentive to continue to report

<y. ,I'-
19 PTSD symptoms. .

20 DR. TAMMIN would like to broaden that

21 question just a little bit to include the question of

22 whether these veterans were, like Dr. Southwick implied,

23 recruited from a VA hospital or.were they veterans recruited

24 generally from the community.

25 DR. FARFEL: The e%clusion criteria for all four

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 120

1 If the trials, which include the veteran study, excluded

2 subjects who were currently in litigation, but did not

3 exclude veterans or anyone who was currently receiving

4 lisability benefits related to their PTSD, only if it was in

5 :erms of litigation.

6 DR. EJAMER: And, in fact, was there a higher rate

7 If people receiving disability payments for PTSD in the

8 reterans sample as opposed to the community sample?

.
9 DR. FARFEL: That would be somewhat of a logical

10 zonclusion, but we did not actually collect the data.

11 I am sorry, could you repeat youraestion?

12 DR. TAMMINGA: Where did you recruit from, did you

13 recruit from the hospital?

DR. FARFEL: Primarily, as I understand it, the VA

nedical centers recruited from their hospital patient base,

16 xt they were permitted to advertise and, in some cases,

t;&+z!y did. In addition, several of the Vi medical centers

18 Mere allowed to enroll subjects who were not veterans that

,...:yB
they found through their recruitment,-so there are

approximately, if I am rect, about 20 percent of subjects

who met that criteria.

DR. WINOKUR: While we are on this tack, were

23 there differences in this study with respect to prior

24 treatment attempts and also treatment failures as compared

25 to the other studies? I

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 121

1 DR. FARFEL: From what we could determine, there

2 was no difference in the prior treatments' success or

3 failure in Study 641 compared to the two positive general

4 population trials, however, as noted in your briefing

5 document, we did not in the most rigorous way collect the

6 prior treatment history data, so we backed into it looking

at using the data that we did collect regard the patient

self-report of psychiatric medication or psychotherapy

administered within the past five years, and we used the

10 indications of PTSD, depression, sleep, and I believe

11 anxiety to approximate those who might have2en treated for

12 symptoms related to this disorder, so it was not the most

rigorous collection of prior treatment history.

14 DR. BREWERTON: In response to your question, Dr.

15' Katz, regarding the type of trauma and what might account

16 for the differences in the males, in the veterans, there has

'-. L7 L&n a number of studies that have shown that life threat is

18 a powerful predictor of PTSD and the degree of life threat,

;,yJg
19 and I think, by definition, combat-related PTSD is probably

20 in general --certainly e are exceptions--but in general,

21 a more life-threatening situation and trauma than assault

22 even though they certainly can be life-threatening, but not

'23 necessarily so. That is one possible explanation for the

24 findings that we hear today.

25 DR. TAMMINGA: Do we know anything about whether

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 122
1 the dose of life threat is related to the severity of the

2 illness as related to the treatment .outcome?

3 DR. SOUTHWICK: There are many studies that

4 clearly show that level of traumatic exposure is related to

5 development of PTSD symptoms., so the more combat, the more

6 life-threatening experiences, the more likely one is to

7 develop PTSD, and not just combat, but other civilian

8 traumas, as well.

9 DR. TAMMINGA: Is that related to treatment

10 response?

11 DR. SOUTHWICK: I am not sure. ILssume that it

12 is related to treatment response. I can't think of specific

13 studies, but that is my impression.

14 DR. HAMER: I also have a question for Dr.

15 Southwick. One fundamental difference between combat and

16 civilian assault or rape is that in combat, you are part of

4,. Y.7 a-cohort which is being assaulted somewhat impersonally. By

18 and large there is not a specific individual out there

.‘Jg
19 trying specifically to hurt you; while an individual

20 assaults or rape, ther

21 Do you think that relates, do you have any data to

22 think that relates in any way to the potential difference in

23 efficacy that we have seen in these trials?

24 DR. SOUTHWICK: I don't know data specifically to

25 answer your question. I think that one of the variables

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 123
1 ;hat is felt to be very important with regard to stress

2 disorders is how uncontrollable the stress is, and combat is

3 lighly uncontrollable. If you are sitting in a foxhole,

4 rou cannot control whether the mortar around is going to hit

5 !ou or not.

6 so, there is a huge literature on the effects of

7 rncontrollable stress on later development of symptoms, and

8 [ think combat is the perfect example of stress that you

9 :annot in any way control or have very little control over

10 it, at least at times.

11 DR. COOK: I would just like to p&t out that

12 Erom the data that we are looking at in terms of efficacy

13 loday, most of it that is positive seems to not be the

14 combat related,' and a very large group seems to be post-

15 child and sexual abuse.

16 This may be something different, so I have no

4. A7 $&stion about the existence of PTSD, but having seen lots

18 of victims of child physical and sexual abuse at the time,

'-'g
19 it---is remarkable that there is“@Ate.a bit of disconnect
;/
20 between the literature;;

21 What I see are--again, not knowing which factor is

22 which and perhaps from a skewed perspective--it seems like

23 there may be a relationship between onset and later

24 treatment.

25 Now, this is worth'pursuing because many times

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 124
1 people would assume, well, if the onset is early, it may be

2 harder to treat. It is actually possible that the assault

during a different t ime, the nervous system may have a

different consequence that may have a relationship to this

treatment, and not to the other.

In terms of the specific question in terms of the

7 data, I am not sure that we have evidence that postcombat-

8 related PTSD responds to Zoloft. It may respond better to

9 something else. I don't know that we could say that yes or

10 no, but I raise the question.

11 DR. BREWERTON: Along those same *es, I thought

12 that the data were interesting that showed that the men who

were physically or sexually abused as children did respond

14 to sertraline versus the men who had non-childhood sexual or

15 physical abuse. So, I think it supports the notion that the

16 type of trauma is important in response and perhaps more

*. -7 i,,@ortant than gender.

18 DR. TAMMINGA: You are suggesting that the gender

19 effect may be an epiphenomenon‘a out.the type of trauma.

2 0 DR. BREWERTO That's right. You know, they are

21 embedded within each other. Clearly, the males have much

22 more combat related, and the females have much more civilian

23 related.

24 DR. TAMMINGA: Dr. Temple.

25 DR. TEMPLE: It ju$t seems worth mentioning, as

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 125
1 ?fizer pointed out, that that is a tiny number of patients,

2 I think maybe eight in the treated group.

3 DR. TAMMINGA: It is a good point remembering what

4 2r. Smith cautioned us about.

5 In addition to the more general discussion of PTSD

6 and its status as an independent entity, the FDA also would

7 like us to comment about the specific study of PTSD, the

8 kinds of protocols, the duration of studies, the need for

9 long-term studies, the appropriateness of the outcome

10 neasures.

11 I would invite some comment on thse practical

issues now. Dr. Southwick.

13 DR. SOUTHWICK: With regard to duration of

14 treatment, I think there is mounting evidence that the

15 trials need to be perhaps somewhat longer than in other

16 conditions or some other conditions anyway, and I would

-k. ,7 L&nk a minimum of eight weeks and more, as we saw in the

18 sertraline, 12 weeks, there was a difference in their other

19 pharmacologic studies that seem.'!0 have shown similar

20 results, that the effe may take a while to be seen.

21 DR. TAMMINGA: Dr. Brewerton.

22 DR. BREWERTON: I would very much like to second

23 that. I know from the Yale group, there was a study by

24 Goodman and Price, I believe, about OCD and fluoxetine, and

25 if you followed out the patients to 16 and maybe even 20

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 126
1 Ieeks, I think, you have gotten a significant amount of

2 yesponders out at that end, whereas, if you just cut it off

3 it 8 or 12 weeks, you don't get as much of a response, and

4 it may very well be true with this anxiety disorder, as

5 uell. When you have got patients being ill for 12, 18

Tears, it may be unreasonable to expect them to improve in

such a short time.

DR. TAMMINGA: Surely, if PTSD is a chronic

condition, one would ask the question whether the acute

10 symptom response to drug treatment predicts long-term

11 response. One would want to have some infozation about

12 Ihat. The treating physician would want-to have some

13 information about that.

14 DR. NORTH: Along those lines, it would be

15 important to have data on acute PTSD as opposed to chronic

16 PTSD as defined as DSM-IV.

A,-. &7 44 DR. TAMMINGA: You might suggest how one would get

18 those data. They may only come from the kind of PTSD
1$
19 populations that you run into.‘. Would-that be true?

20 DR. NORTH: n't have the exact statistics on

21 what percent of people showing up for treatment show up

22 shortly after a trauma, but the data seem to indicate that a

23 considerable majority of people have onset of symptoms

24 acutely after trauma, but that might be one way of obtaining

25' subjects short of going to a'disaster and doing a study

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 127
1 .here.

2 DR. WINOKUR: In this study, as I recall the data,

3 .t was, of course, a dose titration study, and the dose

4 gradually crept up, as would be expected, to about 150 mg in

5 ~11 four studies, and also it was at 12 weeks that some of

6 :he measures started to be clearly different, so it does

7 raise the possibility that perhaps a higher dose for a

8 .onger period of time may have brought out even more clearly

9 Some differences that were apparent at 12 weeks, but clearly

Right have been more evident with a more continued period of

:reatment.
DR. TEMPLE: Actually, I was curious about the

:itration design. Here is a condition that seems to

nctually respond very late and people are titrating every

couple of weeks in terms of response. It doesn't make a

16 whole lot of sense.

17 r I would be curious as to why that design was

18 chosen. If it was chosen to avoid adverse effect, that

19 would make some sense, but ordin%arily.I think you would

20 learn more from a randg ation to fixed doses even if you

21 inched your way up to those doses, and you didn't really get

22 any of that kind of information here.

23 Now, of course, you could analyze this to see if

24 there is a dose/response hidden in there, but I would be

25 curious about that. I

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 128
1 DR. HAMER: Actually, I think it is unfortunate

2 that there were no fixed dose studies done as part of the

set, because it makes it really utterly impossible to

discern in any decent way a dose/response effect.

Given a particular side effect profile, it is more

6 than possible that the people in whom there is a lack of

7 efficacy could be the ones who get inched up to the higher

8 dose, so you wind up showing an inverse dose/response effect

9 if you analyze these data naively.

10 so, I would have a hard time leaving dose/response

11 out of any set of purely flexible dose triakunfortunately.

12 DR. TEMPLE: You would say then that we should be

13 advising people to utilize fixed dose designs in this

14 situation as we do in most others, of course, frequently

15 ignored?

16 DR. HAMER: Yes, I was really surprised that there

43.. '. 7 ~-4 not one flexible dose and one fixed dose study in terms

18 of the set we were really asked to examine, because it is

19 true , in almost all of the other'things you do, you strongly

20 advise people to do bog ypes of studies, and there is a

21 good reason for that, so that you get a handle on dose and

22 dose/response, and this makes it more difficult.

23 DR. TAMMINGA: In the current clinical research

24 climate, one would have to recognize there is some skew

against doing dose/response studies and going to doses that

II
MILLER REPORTING COMPANY, INC.
507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 129
1 are most efficacious for most people to be compared with

2 placebo. I was just trying to state the other point of

3 view.

4 Dr. Dominguez.

5 DR. DOMINGUEZ: I would like to make two points.

6 Again, I was also surprised at the absence of fixed dose

7 studies. I think that the excuse that previous applications

8 for other indications have not found a relationship between

9 dose and response is a very weak excuse not to do it.

So, even though previous applications have not

shown that, that.does not justify not having that

information available.

One more comment regarding the duration of

treatment or the duration of the acute phase of a study. I

personally believe that 12 weeks may be the optimal, and I

16 disagree with you. Having considerable experience in the

t...aatment of OCD, the vast majority of patients, if you

18 treat them aggressively with pharmacotherapy, will respond

19 well within 12 weeks of treatme.;?. YQU only get the

20 outliers at week 8, we 0 or week 12.

21 You have to balance that against the human

22 subjects issue, the continued exposure of the individual to

23 placebo for an extended period of time. So, I personally

24 believe that a l%-week trial, and when I received the

information initially, was optimal in duration.

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh
DR. TAMMINGA: Dr. Temple.

DR. TEMPLE: Just an observation. I am sure any

3 sponsor that is interested in pursuing this sort of claim

4 will note that one of the two favorable studies would have

5 been much less persuasive if it had stopped prior to 12

6 weeks. That is an important lesson I think people will pick

7 up very quickly.

8 DR. TAMMINGA: Would any of the PTSD experts like

9 to comment on the dose/response question?

10 DR. BREWERTON: I would tend to agree with the

11 sentiment about 'having fixed dose studies. I think that are

12 some, even though not with soloft, there are precedents with

13 other SSRIs, notably OCD tending to respond at higher doses

14 in depression, and bulimia nervosa, as well, tending to

15 respond at higher doses than normal antidepressant doses.

16 DR. TAMMINGA: Dr. Hamer.

'_7 / DR. HAMER: I want to get back'slightly to the

18 issue earlier of gender difference, type of combat

.."#
19 difference, and so forth. .

20 We haven't sz them in our handouts, but you did

21 Phase I trials prior to this, and furthermore, did you

22 collect blood levels during the Phase III trials? Was there

23 any sort of a difference in pharmacology, pharmacokinetics,

24 pharmacodynamics between men and women, and were there

25 different dose blood level curves which might explain a

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 131
1 piece of whatever gender differences we see here?

2 DR. RYAN: We did not collect plasma samples

3 during our Phase III clinical program with sertraline,

4 however, if my memory serves me correct, for the panic

disorder program in a randomized fixed dose design, patients

randomized to 50, 100, and 200 mg at steady state, when

trough levels were taken, and when we evaluated the levels

8 in males versus females, there were no significant

9 differences in those concentrations.

10 DR. HAMER: What about the Phase I, even though we

11 are going back a while, pharmacokinetics ank

12 pharmacodynamics data, does anyone remember those?

13 DR. RYAN: Dr. Alderman, could you come forward

14 and speak to that, please.

15 DR. TAMMINGA: Could you identify yourself,

16 please, and your relationship to Zoloft.

A, 1 7 -8 DR. ALDERMAN: My name is Jeff Alderman. I am

18 with Clinical Pharmacology in Pfizer.

19 ..- We did have one Phask-?f study that looked at

20 differences in gender ~ age as it happened. If I could

21 have Slide No. 6, please.

22 [Slide.]

23 These are results from 11 subjects in each group,

24 young and elderly, as you See, the young being 18 to 45,

25 elderly 65 and over. In each case, male and female, equal

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 132
numbers were looked at. If you look across the

gharmacokinetic parameters, you can see some differences,

3 but if you look only for statistically significant ones, the

4 young males were somewhat less than any of the other groups.

5 There were no gender-specific differences.

6 DR. TAMMINGA: Could you tell us what young and

7 elderly are in terms of years?

8 DR. ALDERMAN: 18 to 45 for young, and 65 and

*
9 older for elderly. All of these, by the way, I don't think

10 I mentioned, this was the top dose of sertraline 200 mg per

_-
11 day for more than two weeks.

12 DR. TAMMINGA: And these Phase I data are

13 similarly manifest in your other studies with sertraline,

14 your depression studies or whatever?

15 DR. ALDERMAN: These levels are consistent, yes.

16 DR. HAMER: So, to interpret this correctly, you

4,
*7 h&e an area in the young, you have an area under the curve

18 that is 50 percent higher in the females, and you have a

.-
19 half-life that looks like it's's itout.50 percent longer.

20 DR. ALDERMAN n this particular group of 11

21 each, yes.

22 DR. HAMER: Which, with a little bit of

23 interpretation, would mean that there is sort of far more

24 sertraline hanging around in the blood of the females than

25 the males. I

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 133
1 DR. ALDERMAN: There is the difference that you
2 pointed out.

3 DR. TAMMINGA: Dr. Winokur.

4 DR. WINOKUR: I was interested in any information
5 from the depression clinical trials with Zoloft in terms of

6 even hints of gender differences in terms of either

7 magnitude of response or rate of response or different

8 doses, anything that we can kind of think about in

9 considering this issue here.

10 DR. RYAN: Yes. For the other three currently

11 approved indications for Zoloft, depression obsessive
I A.
12 compulsive disorder, and panic disorder, there was no hint

13 of this sort of gender by treatment interaction in any of

14 those pivotal studies which supported those indications.

15 DR. TAMMINGA: Has the company done any dose

16 analysis of the PTSD effect?
4 i.7 iv DR. GAFFNEY: Are you asking whether we attempted
18 :o do a dose/response within these four studies?

DR. TAMMINGA: yes. ..3 .

DR. GAFFNEY?' , we did not do that for the

reasons that were pointed out, that it is very difficult to

22 yet a dose/response effect when you are doing a titration

23 study such as this.

24 DR. TAMMINGA: Thank you.

25 Dr. Lacey. I

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
 ajh 134

DR. LACEY: Part of the inclusion criteria

squired young women to be on contraceptives or practicing

ome form of birth control. In raising the question about

4 he pharmacological kind of differences and seeing the

5 ender differences, I am curious as to whether or not--well,

6 hen I looked at the medication list, contraceptives were

7 ot listed either place as an included or excluded

8 edication, so I am just curious as to whether any look was

9 ade at those types of medications in terms of effect?

10 DR. FARFEL: Oral contraceptives were permitted,

11 nd, no, we have not looked at any analysis&f subjects who

12 tere or were not on oral contraceptives.

13 DR. WINOKUR: I wanted to ask Dr. Farfel, since,

14 us we talked about before, the doses did creep up in all

15 Iour of the studies pretty much to the same level, and I

16 zhink that is acceptable, do you have a sense or I am not

6-. ^ 17 ,,&e what kind of instructions the clinical investigators

18 lad in terms of was dose increase, especially later,

19 primarily driven by lack of or Inadequate response, or do

20 you have any other se bout why dose was continually.

21 Jpward titrated to the end of the study?

22 DR. FARFEL: No, I do not have a specific sense of

23 uhy the dose continued to be moved upward. They were only

24 instructed, the investigators, to titrate in terms of

: " 25 considering both efficacy and tolerability.

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 135

DR. TAMMINGA: One of the questions that the FDA

sked us to consider is the kind of trial designs that might

3 e optimal to use to suggest long-term efficacy and whether

4 r not those trial designs should be required before

5 pproval or whether we need those data eventually, but not

6 t the time of approval, and what kind of trial designs

7 ight be optimal.

8 Any comments on those kinds of questions from the

9 ommittee?

10 DR. DOMINGUEZ: Just briefly, I think that any

11 .ype of crossover trial carries with it so xch baggage,

12 .hat I was even surprised to see it as a question in this

13 .ight as a possibility for a chronic disorder.

14 I cannot think of any sort of crossover design

15 :hat would be convincing.

16 DR. SOUTHWICK: I think one of the other problems

., ,7 *I&h a crossover design in PTSD is that the symptoms do wax

18 ind wane, for example, people talk about anniversary

19 reactions where their symptoms%?e worse at a particular

20 zime of year, and it w be really impossible to factor

21 zhat out.

22 DR. TEMPLE: I guess that was a reference to the

23 initial study being crossover design, and certainly what

24 everybody said makes sense. A maintenance trial in which

25 there is a withdrawal is, technically speaking, a crossover

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 2.0002
(202) 546-6666
 ajh 136

1 esign although it is not a randomized order crossover, and

i
2 hose are commonly used now to show there are persistent

3 ffects in depression, and we eventually see those for most

4 .rugs.

5 Actually, Pfizer does have a trial of that design

i; kere. I guess the question is how do you feel about those,

7 nd we are still interested in whether you think that is so

8 .mportant it ought to be done prior to approval, which is

9 lot the normal standard in this country although it is in

10 Surope actually.

11 DR. HAMER: First of all, I want 2 say how much I

12 tppreciate seeing a physician argue eloquently against the

13 rse of crossovers, because they do carry with them so much

14 statistical and methodological baggage that it is really

15 difficult to figure out just what your generalizing to and

16 low you are generalizing.

. * _ '.7 .Y In terms of Dr. Temple's comments, the kinds of

18 sustained efficacy/relapse prevention trials that we get

19 with these re-randomizations, t$&y are not crossovers in the

.
20 same sense, because we‘% really restricting the

21 generalizations we make to the population in some sense that

22 we are using, and so there real1y.i.s much less difficulty in

2 3 those in making those generalizations.

24 DR. WINOKUR: I think longer term studies for the

2: treatment of PTSD will be important eventually certainly in

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 137
1 .ight of the chronicity of the illness, but I think in

2 Fairness, the same points and issues apply to so many of the

3 lisorders that we treat, and such longer term studies for US

4 ind our colleagues clinically have been very important in

5 establishing guidelines for Once treatment and remission of

6 ;ymptoms has been established, how to best manage patients

7 in the longer term, but I think it was pointed out very

8 nicely in the introduction, I think by Dr. Marmar, that

9 :here have been so few studies even looking at acute

treatment under controlled conditions, that for this

disorder, this seems like a very key point zestablish

oefore going on to longer and more complex designs.

DR. TAMMINGA: I think the committee may be ready

to move on to the specific questions of sertraline in PTSD.

We have had several quite specific parts of this

16 question addressed to us by Dr. Smith. I could just

c ;.amarize a couple of those, that we have two out of three

18 studies in the general population that show an effect, but

one. that doesn't show an effecff%and.we could have some

20 discussion of that, so dditional discussion, although we

21 have had a lot about the gender by treatment interaction,

22 and then just a carrying on of our more general discussion.

23 Dr. Cook.

24 DR. COOK: As far as the general question we

25 didn't address, there is one/that I thought was very

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 5 4 6 - 6 6 6 6
ajh 138

1 important, particularly based on the data today, is the

2 question of is PTSD found in pediatric populations and

3 should sponsors of drug products be encouraged to study this

4 disorder in pediatric patients.

5 I think we have already made those comments, but I

6 wanted to have that fully discussed, because in a sense,

7 more than 40 percent of the population is being treated as

8 adults, when they should have been treated as children or

9 were treated as children, but without this.

10 DR. TAMMINGA: So, the committee certainly

. 11 supports early and aggressive studies of PTztreatment in

12 children. Yes, Dr. Dominguez.

13 DR. DOMINGUEZ: I know that the Agency can do just

14 so much in the encouragement of development of a product or

15 an agent for a specific indication, but as I was reading the

16 materials that were provided prior to the meeting, I thought

*a '7 t_@myself wouldn't it be nice to have been able to dissect

18 the pharmacologic effect of the drug in the context of a

19 study which would include at so&? well-established

20 psychosocial intervent? to run concurrently with either

21 medication or placebo.

22 I think this is particular germane to a disorder

23 with so much comorbidity and where psychosocial

24

25
I interventions have been shown, in my opinion, to have a more

robust response than the pharmacologic response that I am

.
MILLER REPORTING COMPANY, INC.
507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 139

1 itnessing from this application.

2 DR. TAMMINGA: There may be some psychosocial

3 reatment buried in the design of the trial. In other

4 ords, if you have a newly diagnosed PTSD person that hasn't

5 one through,the list of symptom response, talked to

6 omebody extensively about their trauma, it seems that just

7 he conduct of the study itself will include some

8 sychosocial treatment.

9 Dr. Southwick.

10 DR. SOUTHWICK: I think this is a very important

31 .ssue. It has to deal with recruitment and&w subjects are

12 -ecruited, are they recruited from a clinic where a person

13 .s accustomed to the idea of PTSD, are they recruited by

14 advertising, someone who has never been in treatment, and as

15 rou said, part of the response --and we.saw some pretty big

16 llacebo responses --may be education.

4,.
-. 17 a@ The person becomes educated, perhaps they have

18 lever been in a relationship with a therapist who is really

-:;- 3
19 ittending to them, and in some ways you could see the

20 repetitive asking of ions about PTSD as a form of

21 exposure.

22 So, it seems to me it is important to really

23 understand how recruitment is done and exactly who the

24 patients are, and how closely the patients that are being

25 studied will match the patients that you are actually going

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 140
1 0 treat.

2 DR. DOMINGUEZ: Let me also add that that is my

3 eeling, that the sociodemographic profile of the population

4 hat was studied, it does not appear to be representative of

5 he individuals with PTSD. It is representative of those

6 ndividuals who will sign an informed consent for a double-

7 llind trial.

8 In general, outreach must take place in order to

9 nclude a more mixed racial population, a more mixed

10 minority population, and yet let me just personalize this

11 jar a second.

12 It is quite different to have gone through

13 Iurricane Andrew in Miami in 1992, and have your roof blown

14 )ff, knowing that you have insurance, knowing that they are

15 foing to put you up, knowing that you have a mother who has

16 1 home, that you can stay there for a while, versus various

'.7 ;-dkets of the population in South Florida which did not

18 lave the social support system, did not have these

19 recourses, and you may indeed gee a differential response to

20 ?harmacotherapy when actor in those social demographic

21 Jariables.

22 Again, it is an issue of average. It is an issue

23 of getting out there and expending more effort to try to

24 recruit those populations into studies that many of these

25 populations 'are very wary to$participate in for various

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 141

masons, but they are absolutely necessary and certain

ostmarketing, they are necessary.

DR. TAMMINGA: Dr. Lacey.

DR. LACEY: I would like to I guess follow up on

5 r. Dominguez's comment there about the recruitment of

6 opulations. Race, as we heard, is mandated as a

7 onsideration in these studies, but at the same time, as we

8 aw the analysis of the data, the number ended up being so

9 mall that we got no meaningful differences there,

10 meaningful findings there.

I.1 As we discussed posttraumatic strx disorder, we

12 .alk about sort of like the combative disorders in males, on

13 :he one hand, as has been pointed out here, and the sexual

14 Lssault, on the other, and we also talked about disaster

15 :hings, yet, we know within this society,. for all of this

16 :entury at least, there is a type of violence that is

'.7 :_&petrated and has been perpetrated and continues to be

18 :hat causes some of the same kinds of things, and they end

19 rp..showing up in people of vari6es racial makeup other than

-..
20 :he white majority. --

21 so, I am once again just saying that as we recruit

22 for persons in studies, I think those kind of considerations

23 need to go into the formula. Otherwise, we end up with a

24 definition that says we have something that works, but we

25 haven't studied it in various parts of our population, yet,

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 5 4 6 - 6 6 6 6
ajh 142
1 hen those persons with that definition come in, they may

2 ot respond at all to what is going on.

3 so, I would want to follow that as much as

4 ossible.

5 DR. TAMMINGA: Thanks, Dr. Lacey.

6 Dr. Hamer.

7 DR. HAMER: I also what to emphasize that for

a i.ological reasons, that is, we know that there are

9 ometimes vast ethnic and racial difference in

10 letabolization and processing by the cytochrome P45O

11 .soenzyme systems, and it is entirely possip that

12 different doses may be required in different subgroups, and

13 .t is important that we know that.

14 DR. TAMMINGA: Any ideas or opinions about Study

15 ;a2? That was the study in the general population which

16 ;howed no difference between placebb and drug. Any comments

-1. '.7 :hut it? Dr. Brewerton.

18 DR. BREWERTON: One of the things that I noticed

19 easthat it had a lower rate of'sssault in terms of the

20 percentages. The othe o were 62, 63 percent physical,

21 sexual assault, whereas, this one I think was 54 percent, so

22 I am not sure how significant that difference is, but that

23 is one thing that jumped out at me.

24 Again, it gets back to the issue of type of trauma

25 and the role that that playsi

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
 ajh 143

DR. TAMMINGA: Dr. Smith, we are not giving you

pery much help on this one.

3 Dr. Hamer.

4 DR. HAMER: Well, as a statistician, then, I will

5 rive Dr. Smith some help. You know, things happen.

[Laughter.]
.DR. TAMMINGA: We need more than that.

DR. HAMER: NO, but it is true, sometimes clinical

:rials fail. Sometimes placebo groups respond, sometimes

10 irug groups especially in psychiatry trials tend to fail to

11 respond. You know, a failed clinical trial2 not

12 larticularly unusual, and not particularly really

13 lisconcerting.

14 You know, if we saw a pattern of eight clinical

15 trials of which only two were successful and six failed,

16 zhat would be very different, but I don't have any--I know

I -7 :1&t that is not real help, but, you know, probability is

18 such that sometimes these things happen.

19 DR. TAMMINGA: Dr. &Gzb. . .

20 DR. KATZ: I t want point out, just sort of

21 naybe enlarge the context, we have asked the question about

22 how do we reconcile 682 with the other two positive studies,

23 but the reality is there are two negative studies out of

24 four, and we have sort of assumed--I am sure we will have

i. ,? 25 more discussion about this later--that that is because it is

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 144
1 . different population, it is not the general population,

2 jut, you know, that is an explanation after the fact even

3 :hough it seems sort of fairly obvious, but reality is just

4 .ooking at the results, two out of four are positive, two

5 jut of four are not, so it is not really two out of three.

6 DR. HAMER: I didn't say it was two out of three.

7 : just concentrated on the one in the general population if

a :or no other reason than in some psychological sense, we

9 lave sort of pushed the other one off the books.

10 DR. KATZ: Right. I just sort of. want to.put it

11 ,ack ofi the page.

12 DR. WINOKUR: But for a perspective with the two

13 2ut of four, what we have clearly heard is there appears to

14 oe a significant gender effect, or at least that is

'15 connected to something else that we need to try to

16 understand better, and there is something very strikingly

4. 17 Ldferent about the veteran population study.

ia So, at least on the face of it, the gender effect

cr.';@
19 is...as robust as the data we have, we.have reason to focus
L
20 primarily on the three'% dies that would have more of a

21 chance of being evaluable in terms of a. response.

22 DR. SMITH: If I might follow up to Dr. Hamer's

23 comment that things also don't happen, as well, so what our

24 concern is, is that could the trend be in the other

25 direction in which we have two unusual results in 640 and

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 145

1 71, as a regulatory agency, we want to protect the public

2 rom something that doesn't work.

3 DR. HAMER: Although, of course, since we attempt

4 o rig statistics so that we don't say something happened

5 nless there is a whole lot of evidence that it did, to some

6 xtent there is a difference in weight between two things

7 hat happened and two things that didn't happen or failed to

a rove that they happened.

9 DR. TAMMINGA: We have already said a lot about

10 he gender issues. Is there anything more that we have to

11 omment on about the gender issue? I am sure. Yes, Dr.

12 leller.
DR. GELLER This actually is a question for the

'DA. Are there any rules about the ratio of positive to

15 negative studies that are desirable at approval time?

16 DR. LAUGHREN: There are no strict rules about

4 '7 .IAt . One thing that we like to see for‘an indication that

ia .s more mature in some sense than this is, from a regulatory

19 standpoint, we like to see an a‘?'-2lve.control

* arm in a trial
-
20 :o help us in interpre it, so that if an active standard

21 irug, which is believed to work, also fails, we are more

22 inclined to discount that study.

23 That obviously is not a strategy that you can use

24 early on in the development of a new indication, but there

25 are no strict rules about what the ratio has to be.

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 146
1 DR. TAMMINGA: Dr. Geller.

2 DR. GELLER: What in the experience would be a

3 lesirable ratio?

4 DR. LAUGHREN: Are you asking what is the worst

5 :ase?

6 DR. GELLER: Or the best case.

DR. LAUGHREN: I really can't give a number. It

.s. always a judgment based on the entire set of evidence

jrovided. You can look at individual studies, even those in

10 rhich you don't have an active standard to rely on, as is

11 )eing done here with the veteran study, to xy and explain

12 thy that study might have failed. But there isn't any

13 lrecise number that one can rely on. It is always a

14 judgment based on the entire set of evidence.

15 DR. TAMMINGA: The phrase I recall is a

16 preponderance of evidence?
/ DR. LAUGHREN: Yes, it's an art more than a
I-
1E 3cience.
'. %',,S
1s -_ DR. TAMMINGA: Dr. Brewerton.
-.,
2c DR. BREWERTO Along those same lines, are there

21 any guidelines in terms of sample sizes or sheer numbers in

2; the studies, however many they are?

2: DR. KATZ: No, not in terms of determining

24 effectiveness in any event, that we often say that the

2E trials need to be as big as they need to be, and it is going

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 147

1 o depend on the variability, of course, it is going to

2 epend on the treatment effect, the population, the placebo,

3 ou know, presumed placebo response.

4 It is hard to say. Certainly, there are

5 onditions where we have considered studies positive or

6 pproved drugs on the basis of fairly small studies, but in

7 lhich the treatment has been shown to be statistically

8 significantly different from the control.

9 Of course, the smaller the studies, the more

10 .ikelihood that there is some bias creeping in or that there

11 .s some imbalance in important characteristz that you

12 ion/t really know how to test for, you don't even know what

13 :hey are necessarily.

14 So, we like to see larger,studies, but there is no

15 specific requirement for numbers. The standard in law for

letermining effectiveness is substantial evidence of

:Qectiveness, which is ordinarily considered to be at least

18 zwo trials.

19 . I am not even sure the@standard is preponderance.

20 It is just that the pr ption is if two adequate and well-

21 designed trials give you statistical significance, that is

22 pretty unlikely by chance that the drugs actually don't

23 work. So, how many studies out of how many? At least two

24 ordinarily.

25 DR. TAMMINGA: Sertraline is a little different

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 148

1 rom some of the other drugs that come before the FDA, that

2 ave broad safety database and other indications available,

3 hich we didn't see, but which we can understand is solid

4 nd reasonable.

5 Dr. Geller.

6 DR. GELLER: This bears on the question that you

7 ad raised, that actually I was indirectly addressing in

8 erms of gender. Are there precedents in a case like this

9 or recommending approval just for one gender or the other?

10 DR. KATZ: Apparently, there are cases in which a

11 pecific indication has.been approved for one sex, but we

12 .ave no personal experience in the Division as far as I know

13 rith that, and in those cases, I think it is usually because

14 nly one sex has been studied. This is a different

15 situation.

16 DR. LAUGHREN: If I can just add a comment on

‘r '7 :'-#At , it is a little bit problematic in my view from a

18 regulatory standpoint to entertain aj?proving a claim on what

.ea*,p
19 iould essentially be a subgroup-analysis if we were to focus
f ugh it is obvious after you see
20 )nly on the women even

21 :he overall effect, and you go back and subgroups, it

22 appears that the effect is coming .largely from women.

23 There are ways of handling that in labeling. This

24 is the situation we faced before. We have a drug Luvox,

25 which is now approved for use in children with OCD based on

MILLER REPQRTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 149

1 study which stratified on the basis of age, children and

2 Ldolescents. The study was positive overall, but again if

3 'ou go back and look at the subgroups, it appears that the

4 !ffect is coming virtually entirely from the children in

5 :hat sample even though it is not a power question. There

6 lere more adolescents in that trial.

7 We ended up approving that claim, but went on in

8 ;he clinical trial section to describe where the effects

9 appeared to be coming from, but more descriptively, but

10 tgain, the question is whether the Agency wants to approve a

11 rery specific indication that is based,.on azbgroup

12 analysis.

13 DR. TAMMINGA: But I would wonder whether the

14 committee would even want to recommend that. I was somewhat

15 impressed listening to the data presentation, the efficacy

16 data presentation this morning, that, in fact, there were

4. :7 ',.&e subgroups, there were subgroups within the male

18 population that responded significantly, the males with a

:,:..-p
19 history of drug abuse, although-they.are very small and it's

20 an early analysis and. oratory'and all that.

21 In my opinion, it would be a bit rash to say that

22 the drug is only active in women, even though the

23 preponderance of its effect, it seems to be most active

24 there.

25 DR. KATZ: We are actually very interested in the

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 150
1 zommittee's view on this question, because we are always

2 :oncerned about what is the label going to look like, what

3 .s the actual indication going to be. It is a critical

4 pestion for us.

5 Just to sort of close the loop, there is no

6 lrohibition against indicating a drug for one sex or

7 Inother, or one subgroup or another. It is just that

8 particular subgroup, it would be very unusual to do that.

9 DR. TAMMINGA: Dr. Hamer.

10 DR. HAMER: However, I think I would personally

11 nave some difficulty with concluding reallyzat this drug

12 #as only effective in women, simply because of the fact that

13 gender was so confounded with so many other.things that it

14 night be connected to, and it would be--you know, I would

15 strongly urge the sponsor to do some clinical trials to

16 attempt to address those issues.

e. :7 I would like to see clinical trials of males with

18 childhood sexual abuse. I would like to see clinical

.-
19 trials, powered appropriately, .'& females with and without

20 histories of drug abus nd so on, and so forth, to attempt

21 to get a handle on what may be driving this in addition to,

22 or instead of, sex itself.

23 DR. TAMMINGA: Dr. Katz.

24 DR. KATZ: Again, I would be very interested to

25 know what the committee thinks about whether or not there is

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 151
1 any hint of anything going on in men and what the basis for

2 granting a global claim, if the committee decides to

3 recommend that the drug ought to be approved, I would be

4 interested to know what the evidence would be that it should

5 be indicated for everybody, and beyond that, as we have

6 heard, there are other studies ongoing, and the question I

7 put to you is whether or not you think it would be necessary

8 to.have one or another of those studies in hand before you
-
9 recommend approval, so maybe they will shed light on the

10 male/female question.

11 DR. TAMMINGA: Dr. Geller.

12 DR. GELLER: This goes back to what Dr. Cook was

13 saying before about the importance of child studies.

14 Another reason is you can study males before the onset of

15 drug abuse, take out some of the confounds.

16 DR. LAUGHREN: Can I raise a question that has

+ ._'7 t&e up at several points during our discussion. My

18 impression is'that the sponsor has data from a relapse I

_.,-i 1
19 prevention trial, and we have noe seen those data yet, and

20 ordinarily, we wouldn'?? ve those discussed at this meeting

21 if we hadn't had a chance to candle them in some sense, but

22 I am wondering if it would be useful to take a peek at those

23 data from the standpoint bf this gender issue..

24 I mean, for example, if there were another source

25 of evidence, even if it was in a relapse prevention context

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 152
1 .hat addressed possibly efficacy in men versus women. Maybe

2 :hat would be helpful to the committee.

3 DR. TAMMINGA: The committee would be interested

4 .n seeing whatever we were allowed to see.

5 [Laughter.]

6 DR. LAUGHREN: Well, I think it is in sort of an

7 exploratory nature, you know, given that we haven't had a

8 :hance to candle it, but it may shed some light on this

9 lurning question about men versus women.

10 DR. TAMMINGA: Dr. Katz.

11 DR. KiTZ: I just want to make a xeat about

12 :hat, which is that, as Tom pointed out, we haven't looked

13 2t it, and the question that I am interested in is whether

14 or not, if we do see some preliminary discussion of it here,

15 Mhether or not it would be critical, ,whether or not the

16 committee thinks it would be critical for us to look at that

4” 17 L&sely and establish that there is or is not an effect on

18 males, let's say, from that study before we take an action.

19 '. 3 to.know whether or not you

In other words, we nee

20 think that data would ritical for an action or for a

21 specific indication.

22 DR. TAMMINGA: So, the company should understand

23 when it shows it‘to us that we could recommend that you be

24 given the data and take a careful look at it.

25 DR. KATZ: Right, and that anything you recommend

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 153

.l bout ultimate action on the application, whether it should

2 e approved at all or whether or not it should be approved

3 or a specific subgroup would depend upon what we think of

4 he data when we actually look at it closely. Is that

5 lear?

6 DR. TAMMINGA: Under those conditions, would the

7 ompany like to present any additional data?

8 DR. RYAN: Dr. Farfel will present a very brief

9 verview of these additional studies.

10 [Slide. 1

11 DR. FARFEL: Pfizer conducted twozng-term

12 ixtension trials. They were extension studies of two of the

13 iour double-blind, placebo-controlled 12-week trials. The

14 nitial studies that fed into the long-term studies were

15 studies 671 and 682, and I will remind you that Study 671

16 showed a treatment effect in favor of sertraline while Study

-4.; '.7 ;:ii! did not.

18 Subjects who completed Study 671 or 682,

.,;?-3
19 regardless of treatment group or response data', were
z
20 sntitled to enroll in 2 xtension study 672, which was a

21 six-month open label study with open label treatment of

22 sertraline in a flexible dose format.

23 At the end of the six months in Study 672,

24 subjects 'who met response criteria,which I will elaborate

25 in a minute, were allowed tojenroll in the relapse

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
 ajh 154
1 revention study 703, in which subjects were re-randomized

2 o either sertraline or placebo.

3 [Slide.]

4 It is important to note that in these two feeder

5 tudies, 671 and 682, there were 380 randomized subjects and

6 75 completed, so were eligible to enter open label

7 reatment. Of the 275 completed, 252 entered the open label

8 rial. 155 completed the six months of open label

9 reatment, and of those 155 completers, 139 met the

10 mesponder criteria, so were eligible to enroll in the re-.

11 *andomization trial.

12 Of the 139 who enrolled in the re-randomization

13 :rial, 96 actually chose to enroll in the re-randomization

"
14 :rial, and so 50 were randomized to placebo and 46 were

15 randomized to sertraline.

16 [Slide. 1

6. '7 B Just to restate, the eligibility criteria for

18 entering the six-month open label trial was simply

19 completion of one of the doubl&%lind.feeder studies. The

.,
20 eligibility criteria f ntering the re-randomization

21 study, the additional six months, was to meet responder

22 criteria for two consecutive visits, the subject's last two

23 consecutive visits, and this is where the responder

24 criteria, as we discussed earlier, were developed, a 30

25 percent decrease in the CAPSY~ total severity score from the

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 155
1 subject's initial baseline from the double-blind studies,

2 irom the first visit that they came to this investigator.

3 In addition, the subject had to have a CGI

4 improvement score'of 1 or 2 at both of the final visits.

5 [Slide.]

6 This slide shows the mean daily dose by selected

7 visit week, it was a long trial, in Study 672 for the safety

8 analyzable subjects. Again, we start with 252 subjects and

9 end with 158 subjects.

10 Because subjects coming into the open label trial

11 were either on sertraline or placebo, all szjects began

12 sgain at 25 mg per day at we,ek 1, and then were flexibly

13 zitrated between 50 and 200 mg, and the mean dose at week 14

14 #as 138 mg, which is consistent, which was also seen as the

15 nean dose in the 12-week studies.

16 [Slide.]

'-7 d This slide shows the mean change on the CAPS total

18 severity score during the six-month open label trial, and

_%.
19 this point here, about 74, repre%ents.the mean on the CAPS
.
20 for these same subject en they entered the initial

21 double-blind, 12-week study, Study 671 and 682.

22 SO, at the beginning of the la-week trial, they

23 had CAP scores of about 74. After 12 weeks of treatment,

24 and this is the placebo and sertraline groups I believe, Or

25 just sertraline--1 am sorry, 'we will clarify that in a

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
 ajh 156

1 inute --they had a mean CAPS of about 45, and then this was
h
2 heir improvement over the six months of open label

reatment. This is the observed cases, and then this is the

OCF at endpoint.
[Slide.]

The readings on the Davidson Scale and the CGI

7 mprovement Scale followed the same pattern.

8 In Study 703, which was the re-randomization

9 ;tudy, the double-blinded continuation trial, the primary

10 efficacy parameters were the time to relapse, so that the

11 Kaplan-Meier estimates of the time to rela and then the

12 )roportion of subjects who actually met relapse criteria.

13 Relapse criteria had to be met for the last two

14 :onsecutive visits in order to be called a relapse patient,

15 snd then the other primary endpoint was a combination of

16 subjects who met relapse criteria, as well as discontinuing

.* .- 17 1-d to insufficient clinical response, which is the ICR

18 abbreviation, because some subjects, when suspecting they

19 were on placebo and beginning ti'relapse, may have chosen to

20 exit the study after on-e eek of meeting relapse criteria or

21 one visit rather than two.

22 The secondary efficacy measures for these trials

23 were the mean changes from baseline to endpoint on the

24 efficacy rating scales.

25 For two consecutive visits, subjects had to have a

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 157

1 GI improvement rating of 3 or greater. We had considered a

2 esponder one who had a CGI improvement of 1 or 2, so all

3 ubjects in Study 703 began the trial with CGI ratings of 1

4 r 2. In addition, they had to have had their CAPS-2 score

5 ncreased by at least 30 percent, which was a minimum of 15

6 loints, from the baseline of the relapse trial, not the

7 baseline of the original feeder study, but they had been

8 considered responders when they entered the relapse trial if

9 .hey had their CAPS increase by 30 percent in addition to

10 .he CGI change, that was considered relapsing, and the

11 nvestigator had to concur with the rating rle.

12 [Slide. 1

13 This slide shows the doses across selected visit

14 reeks in the study for sertraline and placebo. In this

15 :ase, subjects began the trial on the same doses that they

16 had been on at the end of the open label trial, and the mean

a. _ 1.7 :';&e at endpoint is similar to what we have been seeing in

18 the'-other studies.

19 ._ You can also note [email protected] in N's from 46

20 to 28 in the sertralin eated group and from 50 to 20 in

21 the placebo-treated group.

22 [Slide. 1

23 This is a slide of the first primary efficacy

24 parameter, the Kaplan-Meier estimate of the probability of

25 not relapsing, and the red line indicates the sertraline-

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 158

1 reated subjects who had an extremely low probability or

2 elapsing during the course of this trial, and the subjects

3 n placebo had a significantly greater probability of

4 elapsing.

5 [Slide.]

6 The proportion of subjects who actually did

7 Lscontinue- -and discontinuation if you met relapse

8 riteria, you were required to be discontinued from the

-
9 itudy-- the proportion of subjects who discontinued due to

10 leeting relapse criteria, there were 2 of 38 in the

11 ;ertraline-treated group and 12 of 46 in thAplacebo-treated

12 rroup, and this difference was statistically significant.

13 [Slide. 1

14 This is the probability of two things, not

15 relapsing and not discontinuing due to insufficient clinical

16 response, and again, the sertraline-treated subjects had

6.. A7 .&nificantly lower probability of these events than the

18 llacebo-treated subjects.
:;;.g
19 [Slide.] .

20 This slide the proportion of subjects who

21 discontinued for either of these two reasons in both groups.

22 Six of 38 sertraline-treated subjects compared to 21 of 46

23 placebo-treated subjects, and again this difference is

24 statistically significant.'

25 [Slide.] I

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 159
1 These are the main changes in some of the efficacy

2 arameters, the CAPS, the DTS, and the IES. Sertraline-

3 reated subjects are in the red bars, placebo-treated

4 'ubjects are the blue bars, and some of you may be realizing
5 .hat the fact that the change is always positive indicates
6 .hat, in general, all of the subjects were having a

7 rorsening of their symptoms at endpoint--the mean, the mean

8 :hange was a worsening of symptoms at endpoint.

9 I would like to go to the next slide and put this

.n perspective for you.

[Slide; 1

At the beginning of Study 703, subjects had CAPS

scores of about 74. When they finished 12 weeks of double-

14 )lind treatment and entered the six-month open label trial,

15 :hey had CAPS scores of about 38.

16 For those who elected the double-blinded

.,dtinuation study, their CAPS scores at the beginning of

18 :hat--and all of them were de facto .defined as responders--

.:+
:heir CAPS scores here were be%: 20..

So, what you here is the fluctuation in the

ZAPS scores over the course of this additional six months of

zreatment including those who might have discontinued due to

23 relapsing on placebo, and then who were responding on

24 placebo, who continued in the trial.

25 so, although there'was an increase in symptom

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
 ajh 160

1 cores for the group as a mean in both cases, relative to

2 low they were when they came to the trial, is in not

3 remarkable.

4 [Slide.]

5 To conclude from these two trials, we feel that

6 sertraline was shown to be safe and effective in maintaining

7 % response in PTSD symptoms over the course of a year and

8 :hat it was more effective than placebo in relapse.

9 Dr. Katz.

10 DR. KATZ: I thought the reason we wanted to see

11 preliminary results of this related to the zder question.

12 DR. FARFEL: We felt it was hard to skip right to

13 zhe gender question without the trial that we have.

“?i
14 DR. KATZ: Well, again, as I say, my concern or

15 the question that we need to have answered from the

16 committee is whether or not additional data on the gender

4 '7 y-x&ion from this trial or'some other trial, is critical

18 for us to have in hand before we make a final decision on

-::i% -
19 the_ application.

20 That is real he question, not whether or not

21 the study is positive or negative. The committee has

22 already said that long-term data may not be necessary for an

23 approval, but it's the gender question we thought this was

24 trying to get at.

25 DR. FARFEL: We have that. Do you have the slide

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 161

1 f the-men, the change over the course of 672 and 703?

2 [Slide.]

3 This is for Study 672, the six-month open label

4 reatment study. Is this men? This is not men. Okay.

5 In Study 672, of the 244 subjects who entered, 67

6 ere male. Their baseline CAPS score was a 42, which was

7 imilar to the baseline of the females who chose to enter

8 he study, and their endpoint mean CAPS score was a 27

9 compared to a 28 in the female cohort.

10 so, the mean point change for the males compared

11 .o the females was comparable, and in this zmewhat enriched

12 copulation, the mean percentage change of the males who were

13 .n this six-month open label trial,'compared to the females

14 las also comparable, approximately a 36 percent change.

15 DR. KATZ: Maybe you will get to it, but this is

16 >pen label data.

‘* i I 7 / DR. FARFEL: The next slide.

18 DR. KATZ:I want to make it clear that this

.::p
19 doesn't address the question that we,are asking.

20 DR. FARFEL: you have the similar slide for

21 Study 703 by gender?

22 [Slide. 1

'23 In Study 703, the discontinuation study, there

24 were nine males in the sertraline group and 18 males in the

2s placebo group. The mean change from baseline to endpoint in

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 162

1 he sertraline-treated group for the males was a decrease in

2 core of 12.8 points, so that is improvement in symptoms to

3 he tune of 13 points, whereas, on the placebo group, the 18

4 ales had an increase in score of 17.5 points.

5 DR. KATZ: What about the primary outcome, which

6 as time to relapse or proportion of relapse?

7 DR. FARFEL: Could you bring up the Kaplan-Meier

8 or the males in 703 ?

9 DR. GELLER : Could go back to the slide you just

10 .ad?

11 DR. FARFEL: Bring the slide backLp.

12 DR. GELLER: On this slide, it may just be I am

13 .ooking at it quickly, but there is a negative change for

14 Iales and a positive change for females?

15 DR. FARFEL: Yes. That goes to the females, as a

16 group mean, were actually increasing slightly in symptoms.

I. _
-7
-I Q&in, the mean score on the CAPS for the study cohorts when

18 they entered at the beginning of the double-blind trials was

..I,.‘@ .
19 about a 75.
.
20 When they be after six months of open label

21 treatment and were called responders, they had a CAPS score

22 of about 18, so they are fluctuating now around what may be

23 floor effect.

24 DR. HAMER: While we have the slide up, those p-

25 values, exactly what are they testing?

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 54.6-6666
ajh 163
1 DR. FARFEL: This is quite a back-up slide. It's

2 , significant difference in terms of the sertraline group

3 compared to the placebo group. I am not sure which analysis

4 ras used. But the asterisks next to the placebo numbers are

5 !xtra.

6 DR. TAMMINGA: So, in both of these genders, male

7 ind female genders, placebo causes relapse significantly

8 iifferent from sertraline, which causes less relapse.

9 DR. FARFEL: Yes.

10 DR. TAMMINGA: Dr. Winokur.

11 DR. WI-NOKUR: Where were the maleiat the start of

12 :his, that they improved or they had a further decrease of

L2.8 points? This was at the point that they were

responders and then went into the--

15 DR. FARFEL: Could you back up to the slide

16 previous to this one? They were not in order. Because -_

*i-. '.7 ,:&se were males who elected to enter the double-blinded

18 continuation, their levels of symptoms on the CAPS were

I .-.
19 roughly the same, so they had m&n scores of about 20 when

20 they entered this doub linded continuation study, and

21 then they decreased further by 12 points compared to

22 increasing by 17 points on placebo.

23 DR. WINOKUR: SO, in effect, they were almost

24 super-responders, they improved to close to zero,

25 DR. FARFEL: Yes. /The numbers of subjects in the

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 164
1 iale sertraline group was 9.

2 DR. KATZ: I really hate to sound like a broken

3 yecord, but we are not going to be able to adequately

4 analyze these data here. That is why we ordinarily don't

5 lave a sponsor present data. This just is not the

6 tppropriate forum to do that.

7 I will ask it again. What we really need to know

8 is whether or not data of this sort are necessary in order

9 ior you to be able to recommend a particular action. If you

10 zhink it is necessary, we will have the sponsor submit it,

11 se will review it, and if we confirm what tzy say, then, we

12 Mill take an appropriate action.

'13 That is really the question that we need to have

14 addressed by the committee, not so much whether or not at

15 the moment we think this study is positive or negative. We

16 are not going to be able to do that.

_ '-7 $r DR. TAMMINGA: But thank Pfizer for your

18 presentation on the spot.

&==p
19 .._ So, the committee needs to.really continue talking
:.
20 about the gender quest We have seen what additional

21 kinds of data the company has and can submit.

22 The issue that I would like to see the committee

23 discuss is gender issue, to what extent gender is a factor

24 in drug action as we see it in the data presented here. We

25 have said a lot about it already.

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 165
1 Dr. Hamer.

2 DR. HAMER: Gee, I wasn't going to say very much

3 Loday. I have got two small pieces of slightly opposing

4 information, if you will. One is whatever difference we saw

5 in the pharmacology leading to different blood levels,

6 different areas under the curve, and different half-lives

7 oetween the two sexes, and not knowing how that might be

8 related to differential response rates, but it is something

9 zhat bears investigating, and then the other is the confound

10 oetween sex and all of these other things which may well be

11 related to differential response rates.

12 The only way to ge,t at those is to do some studies

13 with sufficient sample sizes in various subgroups, so you

14 can ask the questions. As I said earlier, I would have a

15 hard time concluding that this drug is effective in women,

16 and not effective in men in the absence of being able to

I* 17 h&ribute that difference to these other'confounders.

18 DR. TAMMINGA: Dr. Katz.

19 DR. KATZ: But in the'gbsence of that additional

20 information which we all love to have, what evidence

21 is there that it should be indicated in men?

22 DR. TAMMINGA: We have, in fact, seen data from

23 several subgroups of men that showed significant--in

24 exploratory analysis, showed significant responses. I

25 wouldn't guess that clinicians would like to have their male

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
 ajh 166

1 patients denied this. I don't know if any of our experts

2 want to speak to that.

3 DR. HAMER: But it's all a matter of labeling.

4 They could approve it, but put in the appropriate clinical

5 trial information, so that the physicians had information

6 that there was more evidence currently in females than

7 males, which would then perhaps provide some motivation to

8 the pharmaceutical company to provide other evidence.

9 DR. TAMMINGA: Dr. Winokur.

10 DR. WINOKUR: In his presentation and analysis,

11 Dr. Smith advised us or questioned us aboutsking

12 conclusions about subgroups for which studies were not

adequately powered, and I am struck that we are needing to

make calls, if we do get down to the gender difference, on

15 sample sizes that are strikingly low.

16 DR. KATZ: 'Of course. .I am we certainly are very

6
l. 7 w&y about doing that, as Tom said, and lots of others have

18 said. There is one slide that Dave showed, that I think

*. .3J
19 that is interesting in,this regard, whatever you make of it,

20 which was a slide I g that looked at the menI not in the

21 VA study, but maybe it was just the men in the two studies

22 that were positive.

'23 If you look at that slide, there was.--it could

24 still be, I suppose theoretically it could still be a power

, 25 question-- there are I think 50 men on drug and 50 men on

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 167
1 llacebo or something like that. The scores on sertraline

nd placebo were identical except for the global, which

3 lctually had the same response as the women did, but all the

lthers were right on top of each other.

It was suggesting that it's not really a power

Iuestion. They didn't have the same treatment size effect,

it just wasn't significant. Really, nothing was going on in

8 :hat analysis anyway. I just throw that out, and I would be

9 interested what people think about that.

10 DR. TAMMINGA: Dr. Smith.

11 DR. SMITH: If I might just clarif the slide

J2
12 :hat Dr. Katz is referring to is the slide titled women

13 ;Jersus men, PTSD-specific symptoms 640-671.

14 It is lots of columns of p-values. If I might

15 repeat what Dr. Katz said, it sounded like the column for

16 women did show an improvement, the column for men were

-. 17 c.&entially the same, so if you go across the rows for each

18 of the specific symptoms, women showed an improvement, men

_..
19 were essentially the same for s&traline versus placebo. I

20 think that was his po

21 DR. KATZ: Right. In fact, in some cases, it goes

22 in the wrong direction. That would just suggest then to me

23 that it is not really a power question, it's maybe something

24 else is going on.

DR. TAMMINGA: Dr./Cook.

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 168
1 DR. COOK: This relates to the way that I was

2 aught that things happen in statistics, meaning that you

3 'et different results when you pull different samples out of

4 he barrel, so I am still reluctant to call that a gender

5 :ffect knowing that genders don't have the same--there is a

6 rood chance that we have sampled different populations.

7 so, if we had, let's say, a pathologically defined
8 lisorder in the, sense of under the microscope, a very

9 specific condition, sampled from exactly the same clinics,

10 ior exactly the same problems, exactly the same histories,

I.1 tge of onset, then, I might say yes, these z the same.

12 I get a strong sense of apples and oranges by

13 Tender.

14 DR. TAMMINGA: Thank you.

15 I am going to ask one more question first. Is

16 :here anybody on the committee who would take the opposite

G '.7 ;-dition to the one just articulated by Dr. Cook?

18 DR. DOMINGUEZ: I am not going to take the

*-
19 opposite position, but I wonder-tf perhaps this is the time

20 to make a motion for t& ommittee to consider the question

21 whether indeed the gender issue has been inadequately

22 studied versus whether the study drug has failed to show an

23 effect in men, and I submit that motion to the chairperson.

24 DR. TAMMINGA: In my opinion, we are not quite

25 ready to address that motion/yet since I would like some

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 169
/I
1 I more discussion on both sides of the question first.

2 There are a number of people on the committee that

3 would suggest that with the data at hand, these is not

4 enough data to suggest that this drug would only be

5 effective in PTSD in women.

6 Is that a proposition that is shared by the

7 committee?

8 DR. HAMER: Just one other point about the

9 subgroup analyses which applies less of an extent to the

10 gender issue simply because that is really mandated

11 beforehand, so to some extent, it's a planned hypothesis,

12 but there is a multiplicity issue, and in addition to the

13 power. issue in subgroups, the more subgroups we have, the

14 more tests we do, and the more tests we do, the more likely

15 we are to pop one up by chance alone. So, we need to filter

16 that issue in, as well as the power issue.

0 '7 P DR. TAMMINGA: Dr. Smith.

18 DR. SMITH: I understand what Dr. Hamer said.

19 Because of the multiplicity.issug, the results that we are

-"$,
20 seeing would be dilute&&sentially, is that correct? Okay.

21 DR. TAMMINGA: Dr. Katz.

22 DR. KATZ: Certainly, multiplicity is a concern,

23 but this is a difference. It seems to be fairly robust in

24 II the sense that wherever you look for it almost, certainly at

25 least in the positive studies, it is clearly there.

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 170

1 I am wondering what are the differences, what

2 lakes us think that there is lots else going on besides

3 lender. is it the type of initiating event? If you look at

4 .he slide that we just talked about, that Dave put up, those

5 Iere men who, again small numbers, but those were men, who

6 lad not been in combat. Those are men who had the same

7 ;orts of initiating trauma largely as the women did.

8 So, duration of the disease in men, I don't think

9 -t is really much different, certainly the non-traumatic

10 nen, so I am just sort of wondering what the differences are

11 :hat would suggest to people that there is zally--we really

12 :an't say anything about the difference.

13 DR. TAMMINGA: Dr. Brewerton.

14 DR. BREWERTON: Is there, in fact, a data slide

15 zhat shows us what exactly the types of trauma are in the

16 oy-gender?

C. 17 /, DR. TAMMINGA: We haven't seeti it.

18 DR. BREWERTON: I don't think we have.

19 DR. TAMMINGA: I ha&% procedural question for

20 the committee. I am g to gauge whether we ought to

21 move forward or stop for a lunch break. If people could

22 make a slight nod of their heads one way or the other.

23 Move forward? We will move forward.

24 We do have one issue that we haven't approached

25 yet, and that is the independence of Zoloft's antidepressant

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 171

1 ffect, the independence of the sertraline effect on PTSD

2 'rom depression.

3 Nobody really responded to Dr. Katz's last

4 Iuestion, and I am not sure that we have anymore to say to

5 :hat question than we have already said. I would like to

6 lave us focus some on whether the committee saw this PTSD

7 :ffect as an independent effect on PTSD symptoms separate

8 ?rom its antidepressant effect.

9 Dr. Southwick.

10 DR. SOUTHWICK: Obviously, that is a very

11 :omplicated question as the data suggests. kthink that

1 2 )ecause of the overlap, for example, the HAM-D, I think is

13 obviously picking up a lot of PTSD sorts of symptoms.

14 I found one of the analyses to be illuminating,

15 and that was to look at PTSD with major depression compared

1E to PTSD without major depression, and which showed that the

_ r
% ^ -1 i&D without major depression responded as well.

18 I also think it is important that PTSD-specific

19 symptoms responded, but I think'?o pull them apart is

2c Dbviously very complic

21 DR. WINOKUR: One other point that at least would

2: make me cautious about some of the analysis that Dr. Smith

'22 provided, and I did feel that some of the presentation from

2L the Pfizer investigators in terms of separating out patients

2: diagnosed with primary or significant major depression and

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 172

Lso the improvement being across the spectrum of including

ore, quote, unquote, "PTSD-specific symptoms," some of the

3 nalysis that you presented took the item from the Hamilton

spression Scale, the depressed mood, and I think one needs

o be very cautious about using that item as clinical

epression.
There are so many circumstances in which people

ight acknowledge points on that particular item, where they

ould not be, by most clinician's or researcher's judgment,

10 ignificantly clinically depressed, and I think that it

11 ould also be expected that with overall imxovement in the

12 rimary disorder, that that item might well be expected to

13 hange.

14 so, I think that some caution in terms of

15 ,eneralizing from data with that item to depression per se

16 n proving is crucial to PTSD responding would really be in

^ '-7 Aer, and my overall weight of things was to feel that the

18 vidence more supported specific'effect.

19 We heard from severaf'$eople including Dr.

20 lrewerton on our commi that( anxiety disorders tend to be

21 .mportantly overlapping with.depression commonly, so I think

22 :his is a challenging issue that is difficult to sort out,

23 lot just with PTSD, but with virtually all of the anxiety

24 Disorders that have been ldoked at in terms of drug

25 efficacy. I

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 173

1 DR. SOUTHWICK: I just want to make it clear, if I

2 idn't, that is what I was actually trying to say, that I

eel that, reading and listening to all the data, to me it

ould appear that sertraline has an independent effect on

TSD symptoms.

6 DR. HAMER: I have two comments about that issue.

7 he first is that when we looked at the slide with the

8 orrelations between the Hamilton Depression Scale scores

9 nd the various PTSD indices, they were all correlations in

10 he 0.6 range, which means that in terms of shared variants

11 letween the two scales, the R-squares for tee are about

12 1.36.

13 So, to put it another way, even though there is

14 some overlap between at least the scales that measure the

yeverities of the two diagnoses, there is two-thirds non-

16 lverlap, so that is to me some evidence that, in fact, there

I, '7 .,&at least a difference between depression and PTSD. There

18 .s an awful lot of non-shared variation.

::/S
19 .- The other was that the analysis in which you
.
A&
20 .ooked at the differen&%etween the two treatments on the

21 ?TSD scales after covariating out the Hamilton, there still

22 remained a significant sertraline effect, and I find that to

23 3e reasonably convincing that there is a sertraline effect

24 on PTSD above and beyond that which may be due to

25 depression. 1

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
 ajh 174
1 Now, I may be wrong. One of the things about

2 tatisticians is that statisticians demand the right to be

3 rang 5 percent of the time, and so I could be wrong, but

4 rhat I see is that it certainly looks like it is not just

5 lepression.

6 DR. TAMMINGA: Any more discussion or comment on

7 :his question?

8 I think, Dr. Dominguez, we may be ready to

9 zonsider what your question was, if you could restate it for

.i
10 1 s .

11 DR. DOMINGUEZ: I would like to me the motion

12 for the chairperson to consider that based only on the acute

13 lata,,the 12-week data, whether the issue of the gender

14 differences were either inadequately studied or whether

15 indeed that was a failure to show a response based only on

16 zhe la-week acute studies, the positive studies.

4. '.7 8 DR. TAMMINGA: So, perhaps the committee can

18 discuss the proposition that we cannot fully answer the

19 gen.der question from the data prgsented. Is that it?

20 DR. DOMINGUE It's one opinion versus another,

21 whether it was again inadequately studied, in other words,

there have to be more studies to answer the question, or

23 whether indee,d the data that was presented failed to show an

24 effect in men

\ 25 DR. TAMMINGA: Gan'we have some

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 175
1 lominguez's proposition? I understand that these studies

2 lere not powered to separately demonstrate a PTSD effect in

3 Ien and women, so the idea that that be studied gave us

4 .ndependent information on sertraline effect in men and

5 lomen is true.

6 DR. DOMINGUEZ: Yes, I agree with you. On the

7 &her hand, as I understand it, it's a free country, and the

8 agency may,indeed wish to hear the opinion of the committee

9 rith regards to those two questions.

10 In fact, I think that the question was presented

11 1s such in Dr. Smith's presentation, correcz Okay. Wasn't

12 :hat a question in your pres,entation?

13 DR. KATZ: I am sorry. Could you repeat that? I

14 yould appreciate it.

15 DR. DOMINGUZZ: Versus whether it was inadequately

16 studied versus whether it failed to show an effect in men.

a, ,7 .+a DR. KATZ: To answer the question whether or not

18 :he committee thinks the drug can be approved, and if it can

.i
19 m-approved, what would the committee-propose as an
-.a+$:
20 indication, in other WC --and I think we have sort of been

21 talking about this a lot, whether or not it should be

22 approved, if you believe it should be approved as a general

23 treatment for PTSD, and then some statement later on about

24 where the data came from, or whether it should be restricted

25 to approval in women only, if you think it should be

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 176

1 approved, whether or not you think more data of any kind,

2 )ut specifically with regard to the gender question needs to

3 )e reviewed before you can even recommend an action.

4 DR. TAMMINGA: I think it is important for the

5 zommittee to keep in mind, I think that everybody here would

6 recommend more data in gender, more information since all of

7 .LS are of that bent in any case.'

8 The question that you want our opinion on is

9 whether or not they need more data before approval. Is that

10 correct?

11 I think it is time for me to addrxs the first

12 question, which actually requires a vote and take some

13 discussion on this:

14 Has the sponsor provided evidence from more than

15 one adequate and well-controlled clinical investigation that

16 supports the conclusion that Zoloft is effective for the

a. 17 L,&atment of posttraumatic stress disorder?

18 Do we have any additional discussion around this

‘. _

19 question?
9
20 [No response.

21 DR. TAMTCINGA: Then, I would like to call for a

22 vote on the question. There is three non-voting members of

'23 the committee - Dr. Brewerton, Dr. North, and Dr. Southwick,

24 so I think we should just go around the table and vote on

25 the question: Has the sponsor provided evidence that Zoloft

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 177
1 s effective for the treatment of posttraumatic stress

2 isorder?

3 Dr. Geller, might you start?

4 DR. GELLER: I think yes with some label

5 ronsideration to the need for further studies in certain

6 Lreas, types and gender.

7 DR. TAMMINGA: Dr. Cook.

8 DR. COOK: Yes with the caveat that it is only in

9 :he age groups studied.

10 DR. TAMMINGA: Dr. Lacey.

11 DR. LACEY: I think the sponsor ha&provided

12 evidence from more than one adequate study, but I don't feel

13 comfortable voting that there is sufficient data in the

14 studies presented, so I would be no.

15 DR. TAMMINGA: Dr. Winokur.

16 DR. WINOKUR: I would vote yes overall for

I, 27 LBonstration of efficacy, and I agree with the need to

18 discuss labeling guidelines in light of the limitations of

.>..p
19 :he information that we have. .

20 DR . TAMMIN r. Hamer.

21 DR. HAMER: I vote yes also with the trust that

22 the FDA will be judicial and careful in its labeling.

23 DR. TAMMINGA: Dr. Dominguez.

24 DR. DOMINGUEZ: My vote is yes, as well. My bias

25 is toward my belief that the'sponsor failed to show efficacy

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 178

n men.
DR. TAMMINGA: My vote is also yes, that the

ponsor has showed that sertraline is effective for the

reatment of PTSD with all the caveats that people have

lready submitted.
Dr. Katz.

DR. KATZ: I take to heart Dr. Hamer's hope and

mish that we will do the right thing, and we will try. But

gain, is there a general sense from the committee--I don't

10 .now that this needs a formal vote--that the specific

11 ndication should not be limited to women?

12 I just want to sort of make this explicit. There

re ways that labeling can be written. The indication

.tself could say approved for PTSD in women or it could just

15 lay approved for PTSD as a treatment for PTSD, and then in

16 nother place describe where the data come from.

IS
17 a@ I am just trying to get a sense from the committee

18 which of those two or perhaps some other option would be

._ 3
19 preferable. .

20 DR. 3hat I would say is that we just

TAMMINGAC-"

21 voted on the general efficacy of sertraline in posttraumatic

22 stress disorder and previously had a discussion about the

23 gender question, and that in the gender discussion, although

24 there was a universal call for more data, the majority of

25 the group was not willing to'exclude men from the efficacy

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 179
1 question. If I am misstating that, would somebody from the

2 :ommittee--

3 DR. BREWERTON: I don't think that we really took

4 L formal vote on that.

5 DR. TAMMINGA: We didn't take a vote on it. .I

6 said that was the gist of the discussion.

7 DR. BREWERTON: I would feel more comfortable

8 Limiting it to women personally. I see absolutely nothing

9 in the data that would supports its efficacy in men.

10 DR. TAMMINGA: More discussion on this issue?

11 DR. LAUGHREN: If I can just clariJy, the question

12 zhat I framed in my mind clearly was focused on the claim

13 generally, not limited to men or women or any other

14 subgroup.

15 Again, we have great flexibility in writing

16 labeling to describe the findings in the clinical trial

17 t.&tion, but the question that everyone voted on in my mind

18 was a question on a general claim for PTSD.

19 DR. TAMMINGA: Dr. Ka;;"z. . .

20 DR. KATZ: T is I think what we meant when we

21 wrote the question, but again, it is useful to us to know

22 explicitly how people feel about that, because I am not sure

23 that everybody, when they vote on it, voted on it the way it

24 was technically worded.

25 For example, Dr. Brewerton suggested he would like

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 180

1 t to be restricted, the indication itself to be restricted

2 ust to women.
DR. TAMMINGA: Would you like us to take a formal

ote on that/
DR. KATZ: Or at least poll the committee, I don't

are really about a vote, but if you could poll the

7 ndividual members of the committee, this way we would have

8 t on the record, we would know what people think.

9 DR. TJWMINGA: I would like to then repeat this in

10 mequence, not taking a vote, but having a statement from

11 lath member of the committee on their posit& on the

12 lender.

13 Dr. Geller, could you start?

14 DR. GELLER: I would like it not restricted by

15 render because my experience is the FDA will do its usual

16 outstanding job of including information on the trials in

-7 .L.C labeling.

18 DR. TAMMINGA: Dr. Cook.

19 ._ DR. COOK: I would agrge that it should be labeled

20 and obviously, data neg to come in. I really want to

21 emphasize that although we aren't presented with anything

22 that suggests stratification by previous history, age of

23 onset, that this should be looked for, maybe more important

24 in terms of getting the truth of what should be labeled.

25 So, for example, women with combat experience, we

MILLER REPORTING COMPANY, INC.

507 C Street, N.E:
Washington, D.C. 20002
(202) 546-6666
ajh 181

1 don't know that they respond any better than men with child

2 tbuse. So, it's just a little caveat. We can't say what we

3 ion/t know, and I guess we have to label for what is

4 )resented.

5 DR. TAMMINGA: Dr. Lacey, would you like to make a

6 statement?
7 DR. LACEY: I certainly strongly recommend that

8 :he differentiation be clearly specified about differences

9 in populations even though we might leave it open.

, 10 DR. TAMMINGA: Dr. Winokur.

11 DR. WINOKUR: Taking literally or&s expressed,

12 zhe question that we were asked to address, I, as others,

13 Eelt that we have been presented with convincing data from

14 zwo'good, well-controlled studies that support the general

15 efficacy question.

16 I feel personally, as a committee member, unable

A 17 %/this point, with the information available to address or

18 project whether in the long run, men will be shown to

.%':s
19 respond differently or other factors.will come out, as well,

20 but to me, this is an e that we commonly face in our

21 field where we are dealing with disorders that were

22 primarily describing syndromally or phenotypically, there is

23 heterogeneity in response, I think having treatments that

24 are well studied and shown to respond in at least a fair

25 percentage of individuals gives us a chance then to go

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
 ajh 182

1 forward and do more specific studies to factor that out, so

2 gender or age of exposure or type of exposure may well down

3 the road turn out to be very important dimensions for

4 further research, but I don't feel that we have the

5 information at this point to. really make appropriate

6 decisions about that.

7 DR. TAMMINGA: Dr. Hamer.

8 DR. HAMER: One of the first things we learn in

9 statistics is to not over-interpret a null hypothesis that

10 we failed to reject. We certainly failed to reject the null

11 hypothesis that there is no difference betwxn sertraline

12 and placebo in males here.

I will make the statement that in my opinion, the

sponsor has failed to show us that sertraline is effective

in males, but that is a vastly different thing from saying

16 that we were shown that it is not effective in males.

i _. '.7 4 So, however the labeling gets written, whether the

18 labeling is written in such a way that says the indication

19 :.'* Rer the labeling is written

is-only given in females or whet
3
20 with the indication sa the indication is for PTSD in

21 general and then appropriate labeling saying the sponsor

22 failed to demonstrate that it was effective in males, again,

23 I leave up to the FDA.

24 They do a wonderful job of this sort of thing, in

: : 25 negotiating with the pharmaceutical company, and then just

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 183

1 LS long as you are asking for opinions, if I were

2 constructing an ideal world, I would want to construct the

3 aorld 'in such a way that the sponsor was motivated to

4 investigate the issue of effectiveness in males.

5 DR. TAMMINGA: Dr. Dominguez.

.
6 DR. DOMINGUEZ: Deleting that very, very last

7 statement that you made, since English is not my primary

8 Language, I will ditto everything that Dr. Hamer said. I

9 zhink the sponsor has failed to show an effect in men.

10 DR. TAMMINGA: And I also, adding my opinion in as

_.
11 the chair of the committee, think that thisLndication ought

12 to be in PTSD without reference to gender, but that the

13 data, as it has been presented, go into the labeling.

14 The next question that we have to consider is the

15 safety question. The voting committee can assume that the

16 safety question of sertraline as a treatment in humans has

"I,. ,7 t&n already answered by the FDA, and we need to consider

18 the safety question of sertraline in PTSD.

;.-;3j
19 ..- Is there any discussion, any specific discussion
h&-.3
20 by anybody about safety- ncerns of sertraline in PTSD?

21 Comments by the voting and the non-voting members of the

22 committee.

23 Any concerns that any committee member has of

24 sertraline's safety in PTSD other than the concerns that we

might have about human use in general?

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 184
1 [No response.]

2 DR. TAMMINGA: I think I would like to go around

3 gain and then take the final vote on has the sponsor
4 rovided evidence that sertraline is safe when used in the

5 reatment of PTSD.

6 Dr. Geller.

7 DR. GELLER: Yes.

8 DR. TAMMINGA: Dr. Cook.

9 DR. COOK: Yes.

10 DR. TAMMINGA: Dr. Lacey.

11 DR. LACEY: Yes.

12 DR. TAMMINGA:' Dr. Winokur.

13 DR. WINOKUR: Yes.

14 DR. TAMMINGA: Dr. Hamer.

15 DR. HAMER: Yes.

16 DR. TAMMINGA: Dr. Dominguez.

A,
17 DR. DOMINGUEZ: Yes.

18 DR. TAMMINGA: Dr. Tamminga. Yes.

<,.--i%
19 .._ I cannot believe it, we had-no additional
.*..
20 1iscussion.
.
21 With answering these two questions for the FDA and

22 providing additional and extensive, if you will, or at least

23 some discussion of PTSD as an independent diagnosis worthy

24 of an indication, I would like to thank the committee, both

25 the voting and the, non-voting members, for their

MILLER REPORTING COMPANY, INC.

507 C Street, N.E.
Washington, D.C. 20002
(202) 546-6666
ajh 185

1 xticipation in this discussion, and wonder if the FDA has

2 ly final comments.

3 DR. KATZ: I also would like to thank everybody

4 3ry much. I think it was an interesting discussion, some

5 ough issues, and I appreciate it very much. We will take

6 our advice to heart.

7 DR. TAMMINGA: I would like to bring the meeting

8 o a close and thank everybody very much.

9 [Whereupon, at 1 :15 p.m., the meeting was

10 djourned.1

11

MILLER REPORTING COMPANY, INC.

507 c street, N.E.
Washington, D.C. 20002
(202) 546-6666
 186

CERTIFICATE

I, ALICE TOIGO, the Official Court Reporter for Miller Reporting Company,

Inc., hereby certify that I recorded the foregoing proceedings; that the

proceedings have, been reduced to typewriting by me, or under my direction and

that the foregoing transcript is a correct and accurate record of the proceedings

to the best of my knowledge, ability and belief.

ALICE TOIGO