1

DEPARTMENT OF HEALTH AND HUMAN SERVICES

FOOD AND DRUG ADMINISTRATION

CENTER FOR DRUG EVALUATION AND RESEARCH

GASTROINTESTINAL DRUGS ADVISORY SUBCOMMITTEE

Wednesday, July 14, 2004

8:30 a.m.

ACS Conference Room

Room 1066

5630 Fishers Lane

Rockville, Maryland

PARTICIPANTS

Ronald P. Fogel, M.D., Acting Chair

Thomas H. Perez, M.P.H., R.Ph.,

Executive Secretary

MEMBERS:

Alan Lewis Buchman, M.D.

Byron Cryer, M.D.

Alexander H. Krist, M.D.

John T. LaMont, M.D.

Robert A. Levine, M.D.

David C. Metz, M.D.

Weichung Joe Shih, Ph.D.

David B. Sachar, M.D.

Jose M. Vega, M.D., Industry Representative

DRUG SAFETY AND RISK MANAGEMENT

ADVISORY COMMITTEE (Voting):

Brian L. Strom, M.D., M.P.H.

Curt D. Furberg, M.D., Ph.D.

Arthur A. Levin, M.P.H., Consumer Representative

CONSULTANTS (Voting):

Ralph D'Agostino, Ph.D.

Allen Mangel, M.D., Ph.D.

FEDERAL EMPLOYEE (Voting):

Maria H. Sjogren, M.D.

FDA STAFF:

Robert Justice, M.D., Director, Division

of Gastrointestinal and Coagulation

Drug Products

Robert Prizont, M.D., Medical Officer

Garry Della'Zanna, D.O., M.Sc., Medical Officer

Julie Beitz, M.D., Deputy Director, ODE III

C O N T E N T S

Call to Order, Introductions, Ronald Fogel, M.D., 5

Meeting Statement, Thomas H. Perez, M.P.H. 8

Opening Comments, Robert Justice, M.D., Director,

Division Gastrointestinal and Coagulation Drug

Products 11

Novartis Presentation, Zelnorm, NDA 21-200:

Introduction, John R. Cutt, Ph.D., Novartis

Executive Director Global Head GI, Drug

Regulatory Affairs 14

Chronic Constipation: An Unresolved Problem for

Many Patients, Charlene M. Prather, M.D., St.

Louis University School of Medicine 20

Zelnorm: Efficacy and Safety in Chronic

Constipation,

Eslie Dennis, M.D., Novartis Senior Medical

Director, GI Clinical Develop and Medical

Affairs 40

Zelnorm: Safety Overview, Bo Joelsson, M.D.,

Novartis Senior Medical Director, Clinical R&D 69

Fatality Cases, Michael Shetzline, M.D., Ph.D.,

Novartis Senior Medical Director, U.S. Clinical

Development and Medical Affairs 82

Zelnorm: Safety Overview (continued),

Bo Joelsson, M.D. 92

Benefit/Risk Assessment, Philip Schoenfeld, M.D.,

University of Michigan School of Medicine 94

Questions on Presentation 116

FDA Efficacy Presentation, Robert Prizont, M.D.,

Medical Officer, Division of Gastrointestinal

and Coagulation Drug Products 158

C O N T E N T S (Continued)

Questions on Presentation 173

FDA Safety Presentation, Gary Della'Zanna, M.Sc.,

Medical Officer, Division of Gastrointestinal

and Coagulation Drug Products 199

Questions on Presentation 220

Open Public Hearing:

Jeffrey D. Roberts, B.Sc., IBS Self Help Group 229

Constance Hill 235

Linda Roepke 241

Clarification of Issues 249

Discussion of Questions 295

Adjournement 364

P R O C E E D I N G S

Call to Order, Introductions

DR. FOGEL: Good morning. My name is Ron

Fogel. I am acting chair for today's meeting of

the Gastrointestinal Drugs Advisory Committee.

Today's meeting deals with the new drug application

of Zelnorm for the proposed indication of the

treatment of patients with chronic constipation and

relief of associated symptoms of straining hard or

lumpy stools and infrequent defecation.

There has been one change to today's

agenda. The agenda has been pushed back half an

hour so the tentative time of adjournment is five

o'clock. Why don't we start by going around the

table and introducing ourselves? If we could start

on my far left?

DR. VEGA: Jose Vega, from Amgen in

California.

DR. LEVIN: Arthur Levin. I am a member

of the Drug Safety and Risk Management Advisory

Committee. I am a consumer representative and I am

a guest as a consumer representative here today.

DR. STROM: Brian Strom, University of

Pennsylvania. I am a recent graduate of the Drug

Safety and Risk Management Advisory Committee--I

have already forgotten the name of the committee!

I am here as a special government employee, though

that is not what is says there.

DR. FURBERG: I am Curt Furberg, from Wake

Forrest University. I am an active member of the

Drug Safety and Risk Management Advisory Committee.

DR. D'AGOSTINO: Ralph D'Agostino, from

Boston University, statistician, consultant to the

FDA.

DR. LAMONT: I am Tom LaMont. I am a

member of the GIDAC. I work at Beth Israel

Hospital in Boston and Harvard Medical School.

DR. LEVINE: I am Bob Levine, State

Medical University, Syracuse, New York, and I am a

member of the GI advisory committee.

DR. METZ: David Metz, University of

Pennsylvania. I am on the GI drug advisory

committee.

DR. PEREZ: Tom Perez, Executive Secretary

to this meeting.

DR. FOGEL: Ron Fogel, Henry Ford Health

System, in Detroit.

DR. SACHAR: I am David Sachar, from Mount

Sinai School of Medicine, in New York--my maiden

voyage on the GI drug advisory committee.

[Laughter]

DR. BUCHMAN: Alan Buchman, from

Northwestern University, in Chicago, and this is

also my first cruise with today as well.

DR. MANGEL: Allen Mangel, Research

Triangle Institute. I am a special government

employee.

DR. CRYER: I am Byron Cryer, from the

Dallas VA Medical Center and UT Southwestern

Medical School. I am a member of the GI advisory

committee.

DR. DELLA'ZANNA: Garry Della'Zanna,

medical officer in the GI and Coagulation Drug

Product Division.

DR. JUSTICE: Robert Justice, Director,

Division of Gastrointestinal and Coagulation Drug

Products.

DR. BEITZ: Julie Beitz, Deputy Director

in the Office of Drug Evaluation III.

DR. FOGEL: Thank you, all. At this point

Tom Perez will read the meeting statement.

Meeting Statement

DR. PEREZ: Thank you and good morning.

The following announcement addresses the issue of

conflict of interest with regard to this meeting,

and is made part of the record to preclude even the

appearance of such at this meeting.

Based on the submitted agenda for the

meeting and all financial interests reported by the

committee participants, it has been determined that

all interests in firms regulated by the Center for

Drug Evaluation and Research present no potential

for an appearance of a conflict of interest at this

meeting, with the following exceptions:

In accordance with 18 USC Section

208(b)(3), full waivers have been granted to the

following participants, Dr. Ronald Fogel has been

granted a waiver for serving as a member of the

sponsor's speakers bureau. His lectures are

unrelated to the matter at issue and he receives

less than $10,001 per year.

Dr. Ralph D'Agostino has been granted a

waiver for serving on a competitor's advisory board

on unrelated matters. He receives less than

$10,001 per year.

Dr. Byron Cryer has been granted a waiver

under 21 USC 355(n)(4), amendment of Section 505 of

the Food and Drug Administration Modernization Act,

for ownership of stock in a competitor. The stock

is worth less than $5,001. Because this interest

falls below the de minimis exemption allowed under

5 CFR 2640.202(a)(2) a waiver underlying 18 USC

208(b)(3) is not required.

Dr. David Metz has been granted waivers

under 18 USC Section 208(b)(3) and 21 USC 355(n)(4)

for his spouse's ownership of stock in a competitor

valued from $50,001 to $100,000.

Lastly, Dr. Allen Buchman has been granted

waivers under 18 USC Section 208(b)(s) and 21 USC

355(n)(4) for owning stock in a competitor valued

from $25,001 to $50,000.

A copy of the waiver statements may be

obtained by submitting a written request to the

agency's Freedom of Information Office, Room 12A-30

or the Parklawn Building.

In the event that the discussions involve

any other products or firms not already on the

agenda for which an FDA participant has a financial

interest, the participants are aware of the need to

exclude themselves from such involvement and their

exclusion will be noted for the record.

We would also like to note that Dr. Jose

Vega has been invited to participate as an industry

representative, acting on behalf of regulated

industry. Dr. Vega is employed by Amgen, Inc.

With respect to all other participants, we

ask in the interest of fairness that they address

any current or previous financial involvement with

any firm whose product they may wish to comment

upon. Thank you.

DR. FOGEL: Thank you. At this time I

will turn the meeting over to Dr. Justice, of the

FDA, for opening comments.

Opening Comments

DR. JUSTICE: Good morning. I would like

to welcome everyone to today's meeting of the

Gastrointestinal Drugs Advisory Committee, and I

would especially like to welcome members of the

committee and special government employees for

taking the time to provide us with your advice.

As you have heard, today we will be

discussing the application for Zelnorm tablets for

the proposed indication of treatment of chronic

constipation. Before going on to the

presentations, I would just like to briefly go

through the questions so you will have them in mind

as you listen to the discussions.

The first item is that we would like you

to discuss the appropriateness of a primary

efficacy endpoint of an increase of equal to or

greater than 1 complete spontaneous bowel movement

per week versus a total of greater than 3 complete

spontaneous bowel movements per week.

The second question is, is the population

studied representative of patients with chronic

constipation? If not, how do the populations

differ?

The third question is only 9 to 16 percent

of subjects were greater than or equal to 65 years

of age and the treatment effect was significantly

smaller in older patients. Are these data adequate

for an indication that is common in the elderly?

The fourth efficacy question is that only

9 to 14 percent of the subjects were male and the

treatment effect was smaller in males than females.

Are these data adequate to support approval of

Zelnorm for use in the treatment of chronic

constipation in males?

The next question is are the clinical

trial data adequate with respect to the population

of non-irritable bowel syndrome patients with

chronic constipation that is likely to be treated

with Zelnorm?

The next efficacy question is, is Zelnorm

effective for the treatment of chronic constipation

and associated symptoms?

As far as the safety questions,

postmarketing access of ischemic colitis and

serious complications of diarrhea were not limited

to patients with irritable bowel syndrome. What

are the implications of these adverse events from

patients with chronic constipation?

The next safety question is that the

incidence of diarrhea and discontinuation due to

diarrhea was higher in patients 65 years of age or

older. Is there sufficient information that

Zelnorm is safe for use in this age group?

The next safety question is do the adverse

event data from the clinical trials and

postmarketing surveillance provide adequate

evidence of safety of Zelnorm for the treatment of

chronic constipation?

The next safety question is should the

information on the postmarketing cases of ischemic

colitis and intestinal ischemia be moved from the

"precautions" section to the "warning" section of

the package insert?

Then, the final question will be the

overall question of should Zelnorm be approved for

the proposed indication of the treatment of

patients with chronic constipation and relief of

associated symptoms of straining, hard or lumpy

stools, and infrequent defecation?

With that, I will turn it back over to Dr.

Fogel for Novartis' presentation.

DR. FOGEL: Thank you very much. At this

juncture I will turn the meeting over to Dr. John

Cutt, Global Head of GI Drug Regulatory Affairs for

Novartis, who will introduce the speakers and the

presentations. Thank you.

Novartis Presentation

Introduction

DR. CUTT: Thank you. First I would like

to thank Dr. Beitz, Dr. Justice--Dr. Prizont is not

here yet--and Dr. Della'Zanna and the rest of the

FDA reviewers, and Dr. Fogel and the rest of the

advisory committee, and say good morning to you.

My name is John Cutt. As Dr. Fogel said,

I am the executive director and the global head for

the gastrointestinal regulatory group at Novartis,

and it is my pleasure to share with you today the

clinical data on chronic constipation that we have

generated.

Let me start out with the objectives, as

we see them today for the meeting. We would like

to share the Zelnorm Phase 3 clinical data in

support of a new indication. Dr. Fogel read that

before, I will read it again. Zelnorm is indicated

for the treatment of patients with chronic

constipation and the relief of associated symptoms

of straining, hard or lumpy stools, and infrequent

defecation.

The second topic that we are going to

review today is the postmarketing safety data that

we have generated since the approval of the drug in

the United States in July of 2002. This approval

was for patients with irritable bowel syndrome with

constipation.

A brief introduction of the compound,

Zelnorm is tegaserod maleate. It is 5-HT

4 receptor

partial agonist with affinity for the 5-HT

receptor in the GI tract. For its pharmacologic

activity in the GI tract, Zelnorm enhances

intestinal motility; increases intestinal

secretion; and inhibits visceral sensitivity. We

have also demonstrated in clinical trials that

Zelnorm can improve the constipation symptoms in

patients with irritable bowel syndrome with

constipation.

So, these data together are the basis for

the hypothesis that Zelnorm could be effective to

treat patients suffering from chronic constipation.

Novartis-designed clinical development

program for chronic constipation included two

randomized, placebo-controlled pivotal studies.

Both the studies were 12 weeks in duration to

assess the efficacy, safety and tolerability of the

drug. We studies both the 2 mg dose and the 6 mg

dose BID versus placebo. In total, there were

2,612 patients that were studied. The program also

included one extension phase study which was added

on to one of the pivotal studies. This was a

13-month extension for assessing the long-term

safety of the compound. The other pivotal studied

included a 4-week withdrawal period. Today what we

are going to do is show you the results of these

pivotal studies.

We will also share with you the

postmarketing clinical data that we have collected

since the approval of the drug in the United States

in July of 2002. That approval was for the

short-term treatment of women with irritable bowel

syndrome whose primary bowel symptom is

constipation. The recommended dose is 6 mg BID for

a period of up to 12 weeks. At the time of the

approval we demonstrated the efficacy, safety and

tolerability in 5,319 patients in the clinical

trial program. At this point in time, now, we have

generated data on 11,600 patients in clinical

trials. These patients were all treated with

Zelnorm. What this means is that it translates to

approximately 3,456 patient-year exposure to the

drug in the clinical trials.

In terms of the worldwide clinical

experience for the drug, Zelnorm is now approved in

56 countries for the indication of IBS with

constipation. We have also received approval for

the drug in 10 countries for the indication of

chronic constipation that we are seeking from the

advisory committee and the FDA.

We first made the drug available to

patients suffering from IBS-C in January of 2001 in

the rest of the world. So, at this point we have

over 3 years of postmarketing experience with the

drug in patients. What this means is that we have

treated approximately 3 million patients globally

with the drug and about 2 million of those patients

have been treated in the United States. This now

translates to about 362,000 patient-years of

experience to Zelnorm.

The safety data from the clinical trial

setting and the postmarketing environment we

believe supports a favorable safety profile for

Zelnorm. So, our conclusion from the data that you

will see today during the presentations is that

Zelnorm, at the recommended dose of 6 mg BID, is

efficacious and safe for the treatment of patients

with multiple symptoms of chronic constipation.

What I want to do is review the agenda

very briefly, the people that will be presenting

for us. First we have Dr. Charlene Prather. She

is from the St. Louis University and will speak

about chronic constipation. Her presentation is

title an unresolved problem for many patients in

clinical practice.

She will be followed by Dr. Eslie Dennis.

Eslie is from the Novartis clinical development and

medical affairs department. Dr. Dennis will

present the efficacy and safety data from the

pivotal studies.

That will be followed by Dr. Bo Joelsson.

He will present our overall clinical safety data

and review some of the adverse events of special

interest that we have agreed to talk about with the

FDA.

Finally, Dr. Philip Schoenfeld, who is the

chief of the gastrointestinal group at the Veterans

Hospital in Ann Arbor, at the University of

Michigan, will conclude historical presentation

with a benefit/risk assessment for the drug.

We also have four consultants that have

joined us today to answer questions that you may

have. The first one is Dr. Felix Arellano. He is

from the Risk Management Resources group. His

expertise is pharmacovigilance, epidemiology and

risk management. Then we have Dr. Gary Koch. Dr.

Gary Koch is from North Carolina, Chapel Hill. He

is an expert in biostatistics. We have Dr. David

Lieberman. He is from the Oregon Health and

Science University. He will be here to answer any

questions you have on the core database which is

part of the presentation.

Then we have Dr. Walter Peterson, a

gastroenterologist from the University of Texas

Southwestern.

We have also a number of scientists and

clinicians from Novartis who can answer any of the

specific questions that you have on Zelnorm.

Now I would like to invite Dr. Prather up

to the podium.

Chronic Constipation:

An Unresolved Problem for Many Patients

DR. PRATHER: Thank you, Dr. Cutt. Dr.

Fogel, committee members, ladies and gentlemen, my

name is Charlene Prather. As you heard, I am from

St. Louis University. I am a gastroenterologist.

I have been in practice for over ten years. My

career has been dedicated to the clinical

investigation and, importantly, the clinical

treatment of patients with functional bowel

disorders and gastrointestinal motility disorders.

Chronic constipation is one of the very common

problems that I see in my clinical practice and it

is, indeed, an unresolved problem for many of the

patients that come to see me.

First I would like to review my

presentation objectives. I will begin with a

definition of chronic constipation. I will discuss

epidemiology and resource utilization that is

associated with chronic constipation. I will

review available therapies and the limitations that

some of these therapies may have. I will also

summarize for you my feelings regarding the unmet

medical need associated with chronic constipation.

First, beginning with the definition of

chronic constipation, there are a variety of ways

to define constipation. I have decided to define

constipation into either primary causes of

constipation or secondary causes of constipation.

First let's discuss the secondary causes

of chronic constipation. Secondary causes include

things such as drug-induced constipation. We are

certainly familiar with this with the narcotics we

may give our chronic pain patients. Metabolic

factors--hypothyroidism, hypocalcemia may be

associated with chronic constipation. Importantly,

co-morbid medical conditions. We are certainly

familiar with a variety of neurological disorders,

such as Parkinson's disease, multiple sclerosis, in

which constipation is an important complaint that

many of these patients may bring to us. However,

this is not what I am here to discuss today.

Today I would like to review primary

constipation. Again, with primary constipation we

have learned much in the past several years

regarding what causes primary constipation. There

may be impaired colonic transit or motor function,

certainly an area that I am very interested in. We

often call this slow transit constipation. This

may result from a failure of the neurenteric

function of the digestive system or from the

gastrointestinal reflexes that are involved. It

may also result because there is a problem with the

muscle, a failure of the muscular apparatus.

We also can look at chronic constipation

as a subgroup having ineffective defecation. We

also may call this functional outlet obstruction.

This is really where there is a poor coordination

in the muscular apparatus that is involved in the

defecation process. There are some other

terminologies that may be used as well, such as

pelvic dyssynergia or anismus may be a term that

you have also heard. Most cases of primary chronic

constipation fall into neither of these categories.

They are actually normal transit constipation.

Constipation really isn't defined by

physiologic testing; it is defined on the basis of

symptoms. In my practice the most common reported

symptoms that I see coming from my patients are

complaints of hard or lumpy stools; increased

straining. They may complain of infrequent bowel

movements, but often the sensation of incomplete

evacuation, really outcome having a satisfactory

bowel movement and, not uncommonly, the complaint

of gloating or fullness. Typically, the longer

they have gone since they had a bowel movement, you

know, they are feeling more full and they may

describe that as a bloated sort of sensation.

When I think about chronic constipation,

this is more persistent than intermittent or

episodic constipation. We are familiar with

transient constipation that may occur as a result

of a dietary change. We may also see it in

relation to travel. When I think about what is the

definition I will use for chronicity, it needs to

have been present for several months duration and

quite commonly, in my practice, these patients have

had their constipation for years, frequently dating

back to early adolescence or sometimes even

childhood.

Well, how valid are my ideas about what my

patients bring me with those symptoms? There

actually have been some studies that have taken a

look at this. One of the first studies, performed

by Dr. Robert Sandler, in North Carolina, took a

look at a group of young adults, those around the

university community. These were individuals who

had constipation and the symptoms they reported

most often were, indeed, straining 52 percent of

the time; hard stools, 44 percent of the time;

wanted to have a bowel movement but were unable to,

34 percent of the time; with infrequent stools

being reported just 32 percent of the time.

Now let's think about this. Physicians

were often called upon to think very quantitatively

so we often think about the frequency as being the

most important symptom in constipation. But,

clearly, our patients seem to be telling us

something a bit different. Now, this was not a

population-based study so what actually happens in

the population when we discuss symptoms and

constipation?

I have two studies to review with you.

First is a study on the left, a large

population-based, epidemiologic study by Stewart.

He took a look at the symptoms most commonly

reported in constipation. Again, at the top we see

the complaint--an incomplete bowel movement 83

percent of the time. Unsuccessful bowel movement,

being called a stool but being unable to 65 percent

of the time. We see complaints of abdominal

discomfort, needing to press on the abdomen in

order to have a bowel movement; some abdominal

bloating in a group of patients; but, again, down

at the bottom of this list is frequency, with less

than 3 bowel movements per week being reported by

only 13 percent of this cohort.

On the right hand of the slide is another

population-based study by Pare. Again we see

similar findings, with straining right at the top.

Again, near the bottom less than 3 bowel movements

per week being less frequent in this case, in this

population, 36 percent of the time.

Now, I previously mentioned the subtypes

of primary constipation that I considered. Might

it be slow transit constipation; might it be an

outlet problem; or is it normal transit

constipation? Well, unfortunately, the symptoms

don't help me differentiate between those different

physiologic groups. Fortunately, that is not

necessary because in practice we don't use

physiologic testing, nor do we use the patient

symptoms to define which subgroup they belong in.

This has been information we have really found out

over the past several years. We would like to

think their symptoms will tell us exactly which

subgroup they belong into but that just hasn't been

the case. In clinical practice and in clinical

trials we really don't try to define the subtype of

constipation based on either their symptoms or on

physiologic testing.

However, it is important that we have

criteria for the use of clinical testing and having

a relatively homogenous group of patients that we

can take a look at. An important stab at this has

been the Rome criteria. I would like to review

with you the Rome II criteria that are used in the

diagnosis of functional constipation.

The Rome II diagnostic criteria include at

least 12 weeks, which not be consecutive, in the

past 12 months of 2 or more of the following

symptoms: These symptoms include straining, lumpy

or hard stools, a sensation of incomplete

evacuation, a sensation of anorectal obstruction or

blockage, or having to use manual maneuvers, such

as digitation, to facilitate defecation. You see

an asterisk by these because these need to have

been present on at least a quarter of defecations.

The other criterion is less than 3 defecations per

week.

Looking back at the top line, we see that

it is 2 or more of the following symptoms. So, a

patient may have straining, lumpy or hard stools 25

percent of the time and this would be consistent

with the Rome II criteria for chronic constipation.

Or, it could be that they do have less than 3

defecations per week and a sensation of incomplete

evacuation. For this criterion loose stools must

not be present and there should be insufficient

criteria for irritable bowel syndrome.

A caveat with the Rome criteria is that

one of the criteria that they say is that I cannot

use these criteria if my patients are already on

laxatives. So, these are criteria that are really

appropriate for individuals who are not currently

using laxatives.

Using these definitions and, again, the

Rome I definition was for the criteria from before,

how common is chronic constipation in the general

population? These are some population-based

studies and, depending on the criteria that have

been used, we see a prevalence in the range of 4

percent up to 16 percent. This is a lot of

patients that are complaining of constipation.

However, not all of these patients are actually

coming to see us. A few of these studies actually

looked at how many of these patients or individuals

had actually sought physician care for the

evaluation and treatment of constipation.

What we see is that it is only about 25

percent that actually come in to the physician's

office in order to seek some sort of treatment.

A little bit more about the prevalence,

when we think about constipation we need to know

what age groups might be affected. In the Pare

study he was able to divide this out. In the green

bars we see the Rome I criteria and in the magenta

bars the Rome II criteria. There are some slight

differences, again, depending on the definition

that has been used. This is true of all of the

epidemiologic studies, that we really need to

understand the criteria.

Well, what we see is that actually

constipation is a bit more common in the younger

age group, the 18-34 year-old age group. This is

consistent with my clinical practice. We see that

the prevalence is relatively flat when we take a

look at the 35-49 year-old age group; the 50-64

year-old age group; and even the over 64 year-ole

age group. So, constipation really affects all age

groups.

To summarize the epidemiology related with

constipation quite briefly, chronic constipation

is, indeed, common in the general population.

Again, not all of these individuals come to see us.

Approximately 25 percent actually seek physician

care. In data I have not presented, it is slightly

more common in women than it is in men. The female

to male ratio ranges from 1.3 to 2.5. Importantly,

constipation affects all age groups.

Now, how does this really reflect with

what I see in my clinical practice? In my clinical

practice generally I see females. Now, this may be

because I am a female gastroenterologist, that is

the obvious. However, when I discuss this with my

male colleagues they tell me that they too are

seeing predominantly women that come to see them

for chronic constipation. Most of my patients have

been symptomatic for many years, typically over 10

years. The majority have tried life style changes.

They have tried fiber. They have used

over-the-counter laxatives prior to even seeking

initial care from their primary care physicians.

Most of them manage their constipation with

combinations of these, a combination of fiber and

laxatives. The patients that come to me in my

practice are predominantly referred to me by

primary care physicians and I am also going to see

patients that come from other gastroenterologists.

Again, I really like my patient population

and I like taking care of patients with functional

bowel disorders. These are not all crazy patients

as I know some gastroenterologists think. They

actually cope reasonably well with their condition,

however, they are not completely satisfied and they

are looking for something a little bit better.

So, what is the impact of this condition?

Is it just kind of a minor annoyance to my

patients? I would like to present some data

related to the impact this condition has in

patients. First I would like to take a look at

quality of life. There haven't been that many

studies. I will present three of the four studies

I am aware of.

In Olmstead County, Minnesota, individuals

with chronic constipation reported a significant

impact in quality of life with reduced SF-36

scores. Similarly, in Canada, people who have

either self-reported or Rome II constipation also

had worse SF-36 scores compared to the normal

population. In Australia, people with constipation

had significantly worse SF-12 scores on both the

mental and the physical components. So, there is

certainly an impact on these individuals' quality

of life.

Not only does chronic constipation impact

quality of life, but it is also associated with

increased healthcare utilization. In this next

slide we see that 5.7 million constipation-related

outpatient visits do occur annually; 4.1 million of

these are physician office-based visits. However,

there are 991,000 emergency room visits and 587,000

hospital outpatient visits each year.

The cost is a little bit difficult to get

at as it relates to how expensive is this

condition. The one study that I found was from

1997, a study by Rantis and colleagues, who found

in patients who had been referred for tertiary care

evaluation had costs for additional testing on the

average of $2,752. Again, in 2004 dollars I am

sure that may be a bit more. But the point is

really that this disorder does have an impact both

from a quality of life and from a healthcare

utilization perspective.

So, if this is affecting my patients'

lives I would like to be able to treat them

appropriately, and my goal of therapy is that I

would like to be able to improve GI function in

order to obtain relief of the key symptoms that my

patients are bringing to me, and we have reviewed

what these symptoms are.

Well, what do we have available in order

to do this? Certainly we have fiber; laxatives, be

they osmotic or stimulant laxatives. We can use

enemas or suppositories and we do have some

miscellaneous agents that we use. We can use a

cholinergic agonist, such as bethanchol. I don't

think too many of the gastroenterologists have used

that actually for treatment of constipation but it

is available for us and we have used it in the

past. We may use a prostaglandin analog called

misoprostil and we have also used

colchicine--again, certainly not ideal agents but

they are things that we do have available for us.

Well, there are some challenges with these

agents. My patients tell me that they really

aren't consistently getting relief. There is a

variable treatment response. Importantly, for the

constellation of constipation symptoms that we see

the efficacy really has not been evaluated or

demonstrated for most of these agents.

Importantly, chronic constipation is just that, it

is a chronic problem and most agents are indicated

for less than or equal to 2 weeks of therapy. I

would certainly like to be able to offer my

patients something on an ongoing basis.

There are other limitations associated

with these agents. First is the worsening of some

of the constipation symptoms that I am actually

trying to treat. I mean, who among us has not

given fiber to patients only to have them come back

a week or two later complaining of increased

bloating and gas? Likewise, these agents can also

cause cramping, abdominal pain or colicky stools.

Fortunately, complications are not common with the

treatments that I use for treating constipation. I

do worry in some patients should they develop

severe diarrhea which can result in hypovolemia or

electrolyte disturbance. Metabolic disturbances

can occur, such as hypokalemia or hypomagnesemia

depending on the agent I may have used.

There are also other adverse effects

which, fortunately, also are not too common. We

can see interference with concomitant drug

absorption. For instance, some laxatives when

given with cipro may result in poor absorption of

that medication or with theophylline.

I am not too concerned about the

structural changes that may occur in the gut

mucosa, things such as melanosis coli or the abuse

potential or dependency, although I can tell you my

patients and many physicians do consider these to

be obstacles to use of many of the agents that are

currently available. My patients certainly tell me

that there is diminished therapeutic effect that

they see that occurs over time when using these

agents, causing them to have to escalate the drug

usage and often with additional side effects

associated with this.

I am talking to my patients not being

satisfied, and can I get at is there truly a

quantitative effect that tells me how satisfied are

patients or physicians with these therapies? Well,

there really isn't much out there. Fortunately, a

colleague of mine, Dr. Larry Schiller, has shared

with me an abstract that he has submitted to The

American College of Gastroenterology. This is an

Internet-based study that was done, and a group of

physicians were asked are your patients completely

satisfied with treatments for constipation? The

physicians overwhelmingly, 82 percent, said no, my

patients are not completely satisfied.

If you take a look at the box on the

right, the reasons for dissatisfaction included

lack of efficacy, 93 percent; safety or side

effects, 57 percent; or other reasons such as taste

or compliance in 27 percent. In this group of

physicians, 60 percent of physicians agreed that

they do not have adequate products for treating

their patients with chronic constipation, and 90

percent of these physicians wanted better treatment

options. Physicians cited frustration with the

current treatments as one of the top 3 reasons

patients state for seeking care for their

condition.

Another study, a population-based study,

also Internet based, took a look at patients who

had seen a physician for constipation within the

past year. In this group of 557 patients, they

were asked are you completely satisfied with your

treatment for constipation? Nearly half said no,

they were not completely satisfied. So, again,

these are patients that have seen a physician

within the past year that were obtained through a

national database survey. The reasons for

dissatisfaction included efficacy, similar to the

physicians, in 82 percent. Patients weren't quite

as concerned as physicians were about safety or

side effects but still an important concern of

theirs in 16 percent. Other reasons, such as taste

or not wanting to take the agents regularly, in 17

percent.

At the bottom of the slide we see two

other references, one from Irvine and one from

Ferrzzi. These data support the findings that they

found in their studies related to patient concerns

regarding the currently available treatments for

constipation.

In conclusion, chronic constipation, in my

opinion, is a condition that is truly in need of a

better approach. Constipation is characterized by

a constellation of symptoms and we need to

recognize what the symptoms are that our patients

bring to us as being most important, including the

complaints of straining and incomplete evacuation.

Certainly, we want to remember frequency but this

is not our patients' primary concern. Chronic

constipation is associated with high resource

utilization and does have a significant negative

impact on our patients' quality of life. The

current pharmacologic agents have some limitations

and many patients and their physicians are not

completely satisfied with the available therapies.

I truly believe that better treatment options are

needed for this condition.

Thank you for allowing me to share with

you today my thoughts about chronic constipation.

This is obviously a topic of great importance to me

and to my patients. I am looking forward to

hearing more from the other speakers today what I

am sure will be a very lively and interesting

discussion.

Zelnorm: Efficacy and Safety in

Chronic Constipation

DR. DENNIS: Thank you, Dr. Prather. Dr.

Fogel, members of the advisory committee, FDA

representatives, ladies and gentlemen, good

morning. My name is Eslie Dennis and I am one of

the senior medical directors for gastroenterology

at Novartis Pharmaceuticals. I am delighted to be

here today to be able to share with you our chronic

constipation program, and I thank you for the

opportunity to do so.

Over the next 30 minutes I will provide

you with our rationale for studying patients with

chronic constipation. I will then highlight the

study objectives of our Phase 3 program, walk you

through the study design, and provide more

specifics around the patient population that was

studied. Then I will present the efficacy data

from our primary and secondary endpoints and,

finally, the safety data for the 12-week

double-blind, placebo-controlled studies and the

safety data from the 13-month blinded extension

study.

Zelnorm is a 5-HT

4 receptor partial

agonist. It is representative of a new class, the

aminoguanidine indole, and it was designed

specifically to act at 5-HT

4 receptors in the GI

tract. The molecular structure of Zelnorm is based

on serotonin which we know plays a crucial role in

the normal functioning of the GI tract. We also

know that the action of serotonin at 5-HT

receptors is prokinetic.

Our mechanism of action and preclinical

data have demonstrated that tegaserod is, indeed, a

promotility agent. Tegaserod has been shown to

augment peristalsis, thereby enhancing gut motility

and decreasing transit time. Furthermore, animal

studies have shown that tegaserod increases

chloride and water secretion which would improve

stool consistency independent of the promotile

effect of the drug. In addition, we have the data

from our IBS with constipation studies that confirm

the significant improvement with Zelnorm compared

to placebo on stool frequency, stool consistency

and straining--all important benefits when treating

chronic constipation.

On the basis of our IBS with constipation

studies, we felt that we could proceed directly to

Phase 2 trials in chronic constipation without a

formal Phase 2 program, and that we would use the

same doses that were tested in our IBS-C Phase 3

trials.

Let me now walk you through the Phase 3

chronic constipation program. The study objectives

were to evaluate the efficacy, tolerability and

safety of 2 doses of Zelnorm, 2 mg BID and 6 mg

BID, compared to placebo over a 12-week treatment

period in patients with chronic constipation.

We had 2 large randomized, double-blind,

placebo-controlled clinical trials in our program.

Study 2301 was conducted in 128 centers in 16

countries in Europe and in Australia and South

Africa. The design consisted of a 2-week baseline

period, followed by a 12-week treatment period with

either Zelnorm 2 mg BID, 6 mg BID or placebo.

One thousand, two hundred and sixty-four

patients were randomized. We chose the time line

of 12 weeks of treatment for the core studies in

keeping with the Rome committee guidelines

regarding duration of clinical studies in

functional bowel diseases for chronic therapies.

The 2 doses of Zelnorm and the BID regimen were

based on our experience with the previous

dose-ranging and Phase 3 studies that were

conducted in IBS-C patients.

The initial 12-week treatment period was

then followed by an optional 13-month extension

period. This extension period was double-blinded

but there was no placebo arm. So, patients who had

received Zelnorm 2 mg BID or 6 mg BID remained on

these doses and patients who had received placebo

then received Zelnorm 6 mg BID in the extension,

and 842 patients entered the extension study.

The primary aim of the extension study was

to provide long-term safety data for the 2 doses of

Zelnorm. Study 2302 was conducted in 105 centers

in 7 countries in North and South America. The

study design was very similar, with a 2-week

baseline period and a 12-week treatment period.

However, in this study the 12-week treatment period

was followed by a 4-week drug-free withdrawal

phase. A similar number of patients were

randomized, 1,348.

Patient inclusion and several of the

endpoints, including the primary endpoint, were

based on the number of complete, spontaneous bowel

movements of CSBMs. Let me clarify this

terminology that we have used. BMs refer to all

bowel movements. SBMs refer to spontaneous bowel

movements. Spontaneous means a non-laxative

induced stool, in other words, no laxative or enema

in the preceding 24 hours. These can be stools

with either complete evacuation or incomplete

evacuation. CSBMs refer to complete spontaneous

bowel movements. Complete is a subjective

definition of a bowel movement that results in a

sensation of complete evacuation. We know that

there are constipated patients out there who have

more than 3 bowel movements a week but these are

often small amounts of hard and lumpy stools with

straining and incomplete evacuation, and these are

patients that are not satisfied with the quality of

their bowel movements. So, complete spontaneous

bowel movement captures the quality of a bowel

movement that is not laxative induced and is a

measure of both the quality and frequency. We felt

that this endpoint best captured what patients

complain of, based on expert opinion and the

published literature.

A recent state-of-the-art review on

chronic constipation in The New England Journal of

Medicine referred to the large study that Dr.

Prather has shown you, stating that constipation

had been identified in this study as an inability

to evacuate stool completely and spontaneously 3 or

more times a week.

Given that there is no recognized gold

standard for endpoints in chronic constipation, it

seemed very reasonable to use the SCBM endpoint as

our primary endpoint. In fact, this is a more

stringent endpoint than using just bowel movements

alone. However, we recognize that different

experts might request different analyses and

different endpoints and so we defined a priori a

number of secondary endpoints representing the

multiple symptoms of chronic constipation that I

will also be presenting to you today.

We included males and females over the age

of 18 years with chronic constipation. Chronic was

defined as at least 6 months of consistent

symptoms. Constipation was defined as less than 3

complete, spontaneous bowel movements per week and

one or more of the following 25 percent of the

time, very hard or hard stools, sensation of

incomplete evacuation, or straining at defecation.

These criteria were based on the well-established

Rome criteria.

Patients were also required to have had a

normal endoscopic of radiological evaluation of the

bowel within the past 5 years and after the onset

of symptoms. In addition, there had to be no

history or evidence of alarm features such as

weight loss, rectal bleeding or anemia since the

evaluation was performed.

Patients were excluded if they had

constipation for which the cause was known, in

other words, secondary constipation as you can see

listed on the slide. So, we studied patients with

chronic constipation of unknown cause. In

addition, patients on concomitant medications that

could affect GI transit were excluded, as well as

patients with fecal impaction requiring surgical or

manual intervention. These criteria were excluded

based on a comprehensive history, thorough physical

examination, as well as baseline ECG and laboratory

assessments.

At the end of a 2-week baseline period and

just prior to randomization additional exclusion

criteria were applied. Patients were excluded if

constipation could not be confirmed by the number

of CSBMs, straining and/or very hard or hard stools

recorded in daily diaries. They were also excluded

if they had loose or watery stools for more than 3

of the 14 days and if they used laxatives outside

of the guidelines for more than 2 of the 14 days.

Patients were deemed to be non-compliant with diary

completion if they entered less than 11 days in the

daily diary and were subsequently also excluded.

On a daily basis we assessed a number of

parameters related to bowel habits--straining,

stool frequency, stool form that we measured using

the Bristol Stool Scale, and whether evacuation was

complete or incomplete. Patients were required to

collect this data for each individual bowel

movement, and we determined which bowel movement

was spontaneous based on the daily diary data

reflecting the time of administration of any rescue

laxatives.

On a weekly basis we asked about

satisfaction with bowel habits, as well as

bothersomeness of constipation, bothersomeness of a

bowel movement distention or bloating and

bothersomeness of abdominal discomfort or pain.

Let's look at the patient disposition.

Over 80 percent of randomized patients completed

the study, with fewer than 20 percent

discontinuations. The most common reason for

discontinuation was for unsatisfactory therapeutic

effect, with the largest percentage being in the

placebo group, as we might expect. Adverse events

accounted for similar numbers of discontinuations

in the placebo and Zelnorm 2 mg BID groups, with a

slightly higher percentage in the 6 mg BID group

which was not statistically significant. The

pattern of disposition was similar in the 2 trials.

Let's look at the results, starting with

the demographic data. These were very similar

between the two studies. The vast majority of

patients, 86 and 90 percent, were female. The mean

age was 46 and 47 years, with a similar age range,

from 18-88 years. Fourteen percent and 12 percent

were 65 years or older, and just under half the

female population was postmenopausal. The vast

majority were Caucasian, more so in the European

study.

Patients were required to have had at

least a 6-month history of constipation symptoms.

As you can see, the mean duration of symptoms was

considerably longer, 14.7 and 19.5 years.

The means of the characteristics of bowel

habit by history and during the 14-day baseline

period are shown on the slide here. However, as

these baseline parameters would not be normally

distributed in a constipated population it may be

more relevant to look at median data at baseline.

When we do so, we see from the history the duration

of symptoms was 10 and 15 years, with hard or very

hard stools 90 percent of the time and a median

number of one SBM per week. Now, in clinical

practice the Rome criteria are applied to the

history to make a diagnosis of chronic

constipation.

From the baseline diary data we see that

the median number of CSBMs was zero and SBMs 2.5

and 2.9. So, these patients fulfilled the

inclusion criteria of having less than 3 CSBMs per

week. In fact, they had less than 3 SBMs per week

by history and by baseline median data. In

addition, the number of SBMs with straining per

week was 2.0 and 2.5 so the majority of spontaneous

bowel movements were associated with straining.

This confirms that the patient population was,

indeed, constipated and, indeed, had chronic

constipation.

I have outlined the study design, the

patient population, demographics and the baseline

characteristics. Now let's look at the primary

efficacy variable. For this endpoint we defined a

responder as having an increase of at least one

complete spontaneous bowel movement per week on

average during the first 4 weeks of the treatment

compared to the 2 weeks at baseline. They had to

have had at least 7 days of treatment.

The results were positive. In study 2301

the responder rates for Zelnorm were 35.6 percent

for 2 mg BID, 40.2 percent for 6 mg BID compared to

26.7 percent for placebo.

In study 2302 the responder rates were

41.4 percent with 2 mg, 43.2 percent with 6 mg BID

compared to 25.1 percent on placebo. The p values

were significant. The 6 mg BID dose was

consistently more efficacious, with deltas of 13.5

percent and 18.1 percent for the 2 studies.

Now, in order to confirm sustained

efficacy of Zelnorm we analyzed those patients with

an increase of at least one complete spontaneous

bowel movement per week on average over the entire

12-week trial duration, compared to the baseline

period of 2 weeks.

Again the results were positive and

consistent with the results for the primary

efficacy variable. A treatment effect for the 6 mg

BID dose over placebo of 13 and 18 percentage

points for the 2 trials respectively was

demonstrated.

When we look at weekly responder rates

using this responder definition, we can see that

the effect of Zelnorm is seen early, within the

first week, and is sustained throughout the entire

treatment period. In study 2302 we can see that

the treatment effect is lost once the drug is

withdrawn. The percentage of responders in both

Zelnorm groups reached the level of placebo within

2 weeks after termination of treatment. The

results with the 6 mg BID dose are consistently

superior to placebo, and here I am showing you the

data for this dose alone so that you can more

clearly see the benefit.

When we look at the number of CSBMs, there

is a marked increase within the first week of

treatment with a significant improvement over

placebo. The number of CSBMs decreased on

withdrawal of the drug, approaching but not

reaching the level observed during the baseline

period. There was no rebound effect demonstrated.

The effect was again more consistent with the 6 mg

BID dose, which you can see more clearly on this

slide.

In order to further assess the benefit of

Zelnorm, we conducted analyses on other

constipation assessments which we defined a priori.

Let me share these results with you. Let's start

with satisfaction with bowel habits. Now, this is

an important endpoint and an important measure of

clinically relevant benefit. The Rome committee

has advocated the use of global endpoints and

satisfaction really is a composite subjective

assessment by the patient. We asked the question

how satisfied were you with your bowel habits over

the past week? We used a 5-point ordinal scale,

with zero being a very great deal satisfied and 4

being not at all satisfied. So, improvement was

represented by a decrease in the satisfaction

score.

Here we defined a responder as having a

mean decrease of 1 or more on the 5-point scale

over 12 weeks compared to baseline. We

subsequently have validated this data relating

satisfaction scores to a relative shift in

distribution, and we have looked at baseline

standard deviations and week 12 standard deviations

and a 1-point change on this score is associated

with significant effect sizes. We saw significant

superiority of both doses of Zelnorm compared to

placebo in this satisfaction endpoint.

Stool form is another important marker of

constipation, and also showed significant

improvement

on Zelnorm. Stool form was 2.5 and 2.8 at baseline

in the 2 studies respectively and on treatment with

Zelnorm. On treatment with Zelnorm this was

maintained at around a score of 3.5 on the Bristol

Stool Scale.

On this slide you can see the change from

baseline in stool form, which showed significant

benefit over placebo for nearly all weeks. Again,

we can see the loss of benefit during the

withdrawal period.

What about straining, yet another

important symptom of constipation? For each bowel

movement we asked the question did you have any

straining? This was a 3-point scale and the

possible responses were zero, no straining; 1,

acceptable straining; and 2, too much straining.

We did not capture straining in the absence of a

bowel movement. We subsequently analyzed straining

scores for spontaneous bowel movements and saw

significant improvement on Zelnorm compared to

placebo which was consistent over time, as we saw

with the other variables.

Now, what about the bothersomeness

questions? On a weekly basis patients were asked

to evaluate the bothersomeness of constipation.

Now, this is a global assessment in keeping with

the global satisfaction assessment.

In addition, we looked at the

bothersomeness of a bowel movement bloating and

distention and bothersomeness of abdominal

discomfort. As you heard from Dr. Prather earlier,

patients with chronic constipation can present with

bloating and abdominal discomfort, and we can see

significant improvement in the bothersomeness of

constipation on Zelnorm for both doses in both

studies. For abdominal bloating and distention and

abdominal discomfort and pain we saw improvement in

these symptoms in both studies, reaching

statistical significance for the 2 doses in study

2302.

As you also heard from Dr. Prather, many

of the currently available therapies for

constipation in fact aggravate the symptoms of

abdominal bloating and abdominal discomfort so this

is another important benefit of Zelnorm.

So, I have presented several secondary

endpoints. Now the question we asked ourselves was

is there an association between responders for the

primary endpoint and responders for the secondary

efficacy variables. Well, as you can see on this

slide, there is a strong positive association

between responders for the primary endpoint and

response to secondary variables. Remember, the

primary responder definition was an increase of at

least one CSBM per week compared to baseline over

the first 4 weeks of the treatment.

Improvement on stool form is represented

by a positive increase, while improvement on the

other variables is represented by a decrease in

scores. You can see the clear-cut difference

between responders and non-responders, which is

significant for each endpoint, which supports the

CSBM primary endpoint.

Now I will walk you through some of the

additional analyses that were done. In discussions

with the FDA early last year, some other responder

analyses were requested prior to database lock.

One of these define a responder was having at least

3 CSBMs per week for the first 4 weeks of the

study.

Now, this was a fixed definition with no

comparison to baseline, and this was a high hurdle

to achieve considering that the patient population

had a median number of CSBMs of zero and a mean

number of CSBMs of 0.5 at baseline. So, reaching a

level of greater than or equal to 3 CSBMs per week

represents on average a 6-fold increase required to

meet this responder definition. As you can see

though, Zelnorm was significantly better than

placebo. Both doses in both studies were

significant, with deltas of 9 percent for the 6 mg

BID dose in the 2 studies. We saw similar results

using this responder definition over the 12-week

treatment period, with deltas of 9 and 11 percent

for the 6 mg BID dose.

So, we have demonstrated significant

benefit of Zelnorm compared to placebo for our

primary endpoint, and we have demonstrated

significant benefit of Zelnorm compared to placebo

in these analyses that were requested by the FDA.

Let us now look at the effect of the

number of bowel movements at baseline on response.

Now, bearing in mind that we used the concept of

complete spontaneous bowel movements, which is a

relatively new concept, we wanted to see if

baseline number of bowel movements, and that is

all-comers, would affect our primary efficacy

variable.

So, we looked at patients who had less

than 3 bowel movements per week at baseline. What

you can see is that Zelnorm is equally effective in

the group that has less than 3 bowel movements per

week as it is in the group that has more than 3

bowel movements per week at baseline.

We also did various subgroup analyses.

These were planned prospectively but we did not

attempt to meet a minimum number of patients in any

subgroup. Subsequently, some of these groups had

very few subjects and this is reflected in the wide

confidence intervals. It is important to remember

that the purpose of subgroup analyses is not to

demonstrate efficacy as these analyses are not

powered to detect statistical significance. The

purpose of subgroup analyses is to ensure that the

effect in any subgroup is consistent with the

overall effect and that we are not seeing any

negative trends.

Here I am showing you the data for the 6

mg BID dose, and we can see the positive odds

ratios for almost all the subjects that we

analyzed. In the group 65 years and older there

was a total of 88 patients in the 6 mg BID group

and 117 patients in the placebo group, with an odds

ratio of 1 for the overall population.

For the male patients, there were 106 on 6

mg BID and 93 on placebo. The odds ratio was

positive at 1.36, and improvement on Zelnorm was

seen for most variables in men.

One of the issues I want to address now is

the question how many of these patients in this

chronic constipation program were, in fact, IBS

patients, and did this have an effect on our

results. We did not actively exclude IBS patients

from the study but when we went back to the patient

history only 4 percent of patients had a diagnosis

of IBS in their history.

As we did not administer the Rome

questionnaire for IBS, we decided to take a

conservative approach to try and identify patients

that we thought may be potentially IBS-like. So,

we identified all patients in whom abdominal pain

was the main complaint at baseline, and this was

about 12 percent of our patients. In addition, we

included patients who had abdominal pain that may

not necessarily have been their predominant symptom

but they also had diarrhea together with this

abdominal pain. So, criteria (a) and (b) on this

slide come from the history and criteria (c) comes

from the baseline diary data.

We came up with a total of 22 percent of

our patients as possibly having IBS. These were

equally represented in the 3 treatment arms. So,

we felt confident that almost 80 percent of our

patients were, indeed, chronic constipation

patients and did not have IBS. However, we were

interested to see what the efficacy results would

look like if we excluded those patients who were

IBS-like.

Here is the pooled data. In this group,

without IS-like features, in other words, the pure

chronic constipation population, you can clearly

see the benefit of Zelnorm with improvement over

placebo of 40 percent in the 2 mg BID group and 18

percent in the 6 mg BID group. We can compare this

to the deltas in the pooled ITT primary efficacy

analysis in which there was a 13 percent delta in

the 2 mg BID group and 16 percent delta in the 6 mg

BID group. So, in this pooled subgroup analysis

the results in the pure chronic constipation group

are even more robust than in the ITT analysis.

So, I have presented data here that

demonstrate the efficacy of Zelnorm in patients

with chronic constipation. The onset of action is

early. The effect is sustained, and there is no

rebound phenomenon.

We have measured the efficacy of Zelnorm

using a number of parameters and Zelnorm is

efficacious for multiple symptoms of chronic

constipation which include straining, hard stools

and infrequent stools, with overall improved

satisfaction. The 6 mg BIT dose has emerged as

consistently more efficacious than the 2 mg BID

dose.

Now let's look at the safety data. I am

going to go through the 12-week data looking at

exposure, adverse event profile, serious adverse

events, and laboratory evaluations, and then the

long-term safety profile from the extension study.

This was a 13-month extension study, providing a

total of 16 months of data for the groups that

received Zelnorm in the core study.

Let's start with overall exposure. The

intended study duration was 84 days. Exposure was

comparable across all treatment groups. The mean

duration of treatment was 80 days, and 84 percent

of patients completed at least 11 weeks of

treatment and 69 percent had more than 85 days of

exposure in this 12-week period. The total number

of patients who experienced any adverse event was

60 percent in the placebo group, 57 percent in the

6 mg BID group and 56 percent in the 2 mg BID

group. The most frequent adverse events were

headache, nasopharyngitis , diarrhea, abdominal

pain and nausea. The only notable adverse event

seen more frequently with Zelnorm was diarrhea, as

you would expect given the pharmacodynamic action

of this drug.

When we look at the most frequent adverse

events leading to discontinuation, we see abdominal

pain, diarrhea, abdominal distension, nausea and

headache. On this table we have included all

discontinuations in which there were at least 5

patients on any dose of Zelnorm. Overall, the only

one where discontinuations appeared to be

dose-dependent was diarrhea, with a discontinuation

rate of less than 1.0 percent on the 6 mg BID dose

and 0.3 percent for the 2 mg BID dose.

Let's look at the diarrhea in more

detail--4.2 percent with the 2 mg BID dose, 6.6

percent with the 6 mg BID dose versus 3 percent

with placebo. Over 80 percent of patients who

experienced diarrhea had only a single episode, and

the median duration of the first episode was about

2 days. Most diarrhea occurred on the first day of

treatment.

When we look at the characteristics of the

stool on the first day of diarrhea, the median

number of bowel movements was 3 in the placebo

group, 2 on Zelnorm 2 mg BID and 3 in the mg BID

group. The median stool form was essentially

similar across all treatment groups at 5.7 for

placebo and 6 and 6.3 for the 2 Zelnorm doses

respectively.

Most patients who had diarrhea continued

on their medication and took no action. There were

more patients on 6 mg BID who adjusted their dose

or temporarily interrupted therapy but there were

very few patients who discontinued permanently

because of diarrhea. None of these cases met the

definition of a serious adverse event or the

definition of clinically significant consequences

of diarrhea such as hypovolemia, hypokalemia or the

need for IV fluids or electrolyte replacement.

Let's look at serious adverse events.

Incidence rates were comparable across all

treatment groups. There were very few

discontinuations due to these serious adverse

events. There were no deaths during the course of

the study but there was one death 67 days after the

last dose of study medication in study 2302. This

was an 85 year-old man who had been on Zelnorm 2 mg

BID. He died from respiratory failure and

mesothelioma secondary to preexisting asbestosis.

We also evaluated a number of laboratory

parameters. There was a low frequency of notable

abnormalities which were essentially similar across

all treatment groups.

The incidence of any abdominal or pelvic

surgeries in the Zelnorm-treated group was lower

than in the placebo group. One patient in study

2302, on Zelnorm 6 mg BID, had a cholecystectomy.

The investigator assessed the event as not related

to study medication. The incidence of other

surgeries was well balanced between Zelnorm and

placebo.

Now let's look at the 13-month extension

study, and 842 patients entered the extension

phase. And, 61.7 percent were exposed to at least

12 months of Zelnorm; 46 percent of patients

discontinued over the 13 months. Most

discontinuations were for unsatisfactory

therapeutic responses, with very few

discontinuations for adverse events. The same

adverse events predominated as during the core

period. Frequencies followed the same pattern as

seen in the core studies, although the incidence

rates were generally higher due to the longer

duration of exposure. No relevant differences were

seen in the rates between the 2 doses of Zelnorm.

There were no deaths in the 13-month extension.

So, to summarize our safety conclusions,

the incidence of adverse events on Zelnorm in

chronic constipation is similar to placebo, except

for diarrhea, which is what we would expect from

the pharmacodynamic profile. There were low

discontinuation rates due to adverse events. The

long-term safety profile was similar to the profile

in the core 12-week studies. Zelnorm, therefore,

as been demonstrated to be safe and well tolerated

in patients with chronic constipation.

What are our final overall conclusions?

Zelnorm is effective in the treatment of multiple

symptoms of chronic constipation, with the 6 mg BID

dose consistently more efficacious than the 2 mg

BID dose. Zelnorm improves satisfaction with bowel

habits; straining; hard and lumpy stools; and

infrequent bowel movements. In addition, Zelnorm

has a favorable safety profile.

Therefore, we are asking for an approval

for Zelnorm for the treatment of patients with

chronic constipation and relief of associated

symptoms of straining, hard or lumpy stools, and

infrequent defecation.

That concludes my presentation. Thank you

very much. I would now like to introduce Dr. Bo

Joelsson, vice president and head of clinical

research and development for gastroenterology, who

will present the general safety overview. Dr.

Joelsson?

Zelnorm Safety Overview

DR. JOELSSON: Good morning, Dr. Fogel,

advisory committee, representatives from the FDA,

ladies and gentlemen. My name is Bo Joelsson and I

am the head of GI research and development at

Novartis.

Today I will review with you the overall

safety experience with Zelnorm, and demonstrate to

you that Zelnorm is a safe and well-tolerated drug.

This is what I am going to review with you today.

First I will show that the positive safety profile

of Zelnorm that was established at the time of

approval in July, 2002 is confirmed in our chronic

constipation clinical program. Secondly, I will

briefly present a few safety topics that we have

agreed with the FDA to discuss at this meeting:

serious consequences of diarrhea; rectal bleeding;

ischemic colitis and other forms of intestinal

ischemia; biliary tract disorders and ovarian

cysts. Finally, I will summarize our experience

demonstrating that Zelnorm is a safe and well

tolerated drug.

The three most common adverse events that

were reported at approval in July, 2002 were

headache, abdominal pain and diarrhea, and the

incidence of diarrhea was higher on Zelnorm.

This slide shows that the adverse event

data from the chronic constipation clinical trials

compared favorably to the IBS constipation data.

Headache incidence is similar between the treatment

arms. Abdominal pain was less common in the

chronic constipation studies than in the IBS

trials, demonstrating that the chronic constipation

population is different from the IBS population

which is characterized by abdominal pain. The

incidence of abdominal pain in Zelnorm and placebo

treated patients indicates that abdominal pain as

an adverse event is not related to Zelnorm

treatment. As in IBS, the reported incidence of

diarrhea is higher on Zelnorm. Adverse events

leading to discontinuation are also low in the

chronic constipation clinical trials.

At approval in July, 2002 the incidence of

serious adverse events was low. This was 1.6

percent on Zelnorm as compared to 1.1 percent on

placebo. Serious adverse events leading to

discontinuation of study drug were 0.7 percent on

Zelnorm and 0.6 on placebo.

The incidence of serious adverse events in

the chronic constipation clinical trials was

similar to that in the IBS clinical program. The

incidence of serious adverse events leading to

discontinuation was lower and identical in Zelnorm

and placebo treated patients.

The clinical trial adverse event data

collected in chronic constipation clinical program

supports and strengthens the positive safety

profile established at the time of approval. With

the exception of diarrhea, the Zelnorm safety

profile is similar to that of placebo.

We have at this time experience with use

of Zelnorm both from clinical trials in patients

with IBS, chronic constipation and upper GI

indications, as well as postmarketing clinical use.

In clinical trials more than 15,000 patients have

been included and more than 11,000 subjects have

taken Zelnorm, which corresponds to 3,456

patient-years of Zelnorm experience. More than

10,000 patients have been involved in controlled

clinical trials and close to 7,000 of them have

been on Zelnorm.

Zelnorm is currently registered in 56

countries worldwide. It has been available since

January, 2001 and here, in the United States, since

July, 2002. Approximately 3 million patients have

been treated, 2 million in the United States. That

corresponds to more than 350,000 patient-years of

treatment and more than 230,000 patient-years in

the United States.

We have agreed with the FDA to discuss

several specific safety topics at this advisory

committee meeting. The first of these is serious

consequences of diarrhea. Diarrhea is an expected

effect of Zelnorm in some patients due to the

mechanism of action. The diarrhea is generally

mild, is generally transient and self-limiting, and

rarely leads to serious consequences.

A patient is defined as having a serious

consequence of diarrhea if one or more of the

following took place: A serious adverse event was

reported as defined by regulatory requirements; if

hypokalemia occurred; if hypovolemia was diagnosed;

if IV fluids were administered; or any medically

significant events related to diarrhea occurred,

such as hypotension, syncope or cardiac effects.

We carefully reviewed our clinical trial

experience of more than 11,000 patients using this

definition in order to identify cases of serious

consequences of diarrhea, and this is our clinical

trial experience. Six cases of serious

consequences of diarrhea were found in the clinical

studies on Zelnorm with more than 11,000 patients.

Four of these patients required hospitalization.

Two received IV fluids. Two had actually possible

other causes. One reported gastroenteritis and one

reported antibiotic-induced diarrhea. All patients

recovered without complications and four of them

were actually able to continue on study medication

after these episodes.

From the postmarketing experience in

approximately 3 million patients, 30 cases have

been reported; 16 were hospitalized; 11 received IV

fluids; 8 exhibited hypotension; 4, syncope; and 4

were considered life-threatening by the reporting

physician; 1 had hypokalemia. One fatality from

aspiration pneumonia was reported in a patient with

acute pancreatitis and chronic liver cirrhosis.

This demographic information on the 30

patients with serious consequences of diarrhea in

the postmarketing experience. There was a wide age

spectrum, 18 to 82 years. The median age was 49

years and only 9 patients were older than 65,

indicating that this is not an elderly specific

issue. As is expected, most were women, reflecting

the label in most countries. Serious consequences

of diarrhea mostly occurred in patients on 12 mg of

Zelnorm per day but were also reported in some

patients on a lower dose.

In clinical trials diarrhea usually

occurred during the first days of treatment, as we

heard before. This was also true for serious

consequences of diarrhea in the postmarketing

experience. It occurred within 5 days in all cases

and the median time was 1 day.

In conclusion, serious consequences of

diarrhea are rare in clinical trials and very

rarely reported in the postmarketing experience.

All cases resolved without sequelae.

Rectal bleeding is of special interest

because of its possible relationship with serious

colon conditions. We have carefully reviewed our

clinical trial data for terms that indicate the

presence of rectal bleeding, such as rectal

hemorrhage, melena, hematochezia, etc., etc. We

found that the presence of rectal bleeding was very

similar in patients on Zelnorm and placebo in our

controlled clinical trials.

From the postmarketing experience of

approximately 3 million patients, 82 cases of

rectal bleeding were reported. Twenty-one were

reported in conjunction with suspected ischemic

colitis; 1 from another form of intestinal

ischemia; 3 from other forms of colitis. In 23

cases hemorrhoids were a possible source of

bleeding. The rest of the cases listed on this

slide show varying etiologies. Fifteen of the

patients were not investigated.

Our clinical trial data indicated a

similar reporting rate in Zelnorm- and

placebo-treated patients. There are rare reports

of rectal bleeding from postmarketing experience,

which indicates that Zelnorm therapy is not

causally related to the rectal bleeding.

Now, the occurrence of ischemic colitis

and other forms of intestinal ischemia is a concern

with drugs used to treat IBS with diarrhea. These

drugs block the 5-HT

3 receptor and, thus, have a

very different mechanism of action than Zelnorm,

which is a 5-HT 4

receptor agonist used to treat IBS

with constipation. Nonetheless, since these are

potentially serious conditions and Zelnorm affects

the serotonin receptor, it is important to

carefully assess whether Zelnorm therapy could

increase the risk of intestinal ischemia.

Ischemic colitis is a rare condition in

the general population. When it occurs, it is

potentially serious but is generally mild and

transient. It is characterized by mucosal erosions

seen at colonoscopy, with rectal bleeding and

abdominal pain being the most common clinical

presentations. Usually no specific treatment is

needed and surgical intervention is rarely

indicated.

While ischemic colitis is very rare in the

general population, it is more commonly reported in

IBS patients. In a study from the Medi-Cal claims

database, 179 cases per 100,000 patient-years were

found in IBS patients versus 47 cases per 100,000

patient-years in non-IBS patients. In a study from

the United Health Care claims database, with a

younger patient population, the corresponding

numbers were 43 in IBS patients and 7 in non-IBS

patients. In the CORI database which collects data

from endoscopy units all over the United States,

ischemic colitis was found in 93 per 100,000

colonoscopies in patients with IBS-like symptoms

while there were 21 cases per 100,000 screening

colonoscopies in asymptomatic individuals. All of

these cases were endoscopically verified and in

most cases supported by histology.

It has been suggested by Dr. Brinker et

al., from the FDA, that the increased incidence of

ischemic colitis in IBS is the result of

misdiagnosis. Dr. Brinker published this analysis

from the patients from the United Health Care

study. He found evidence for misdiagnosis during

the first 3 weeks after IBS diagnosis. However,

when patients with IBS were followed for more than

a year, he still found an increased rate of

ischemic colitis, 53 per 100,000 patient-years.

Thus, ischemic colitis can be misdiagnosed

as IBS during the first weeks of treatment, but

patients with a stable IBS diagnosis for more than

1 year still have an increased risk of ischemic

colitis diagnosis.

Now, there are two, maybe more, possible

hypotheses why the rate of ischemic colitis in IBS

is increased. Ascertainment bias because of the

documented fact that IBS patients are investigated

two to three times more than the general

population, and/or because there are currently

unknown common pathophysiological mechanisms that

we don't know about today.

We carefully reviewed all cases of rectal

bleeding, colonoscopy and reports of colitis in our

clinical trials to identify possible cases of

ischemic colitis and there were no cases of

ischemic colitis identified in any of our clinical

trials involving more than 11,000 patients on

Zelnorm. However, one placebo case with ischemic

colitis was identified in our clinical trials.

From postmarketing experience as of June

1, 2004, 26 cases of suspected ischemic colitis

have been reported. This corresponds to a

reporting rate of 7 cases per 100,000 patient-years

worldwide and 12 per 100,000 patient-years in the

United States. This rate is consistent with the

background rate incidence in IBS patients.

The reported cases of ischemic colitis do

not exhibit any distinct pattern with regard to

duration of treatment, dose of drug, age of

patients, co-morbid conditions, or other

demographic subgroups. The absence of ischemic

colitis cases in clinical studies and the low

reporting rate in postmarketing experience suggest

that Zelnorm treatment does not increase the risk

of ischemic colitis.

Now, these findings are not surprising

since Zelnorm is not expected to cause

vasoconstriction as there are no 5-HT

4 receptors in

the human vascular system. This is further

supported by preclinical studies. In vivo animal

studies have demonstrated no effect on colonic

vascular conductance which is a measure of vasal

activity. In addition, although tegaserod has

negligible affinity for the 5-HT

3 receptor,

tegaserod has affinity for the 5-HT1B receptor but

it does not cause vasoconstriction, as illustrated

in this graph.

This graph depicts the results of adding

sumatriptan and ergotamine, which are two known

5-HT1B agonists, and tegaserod to a preparation of

isolated coronary arteries from non-human primates.

As expected, ergotamine and sumatriptan cause

marked contraction while tegaserod has no effect.

In conclusion, there is no evidence for a

causal relationship between Zelnorm and ischemic

colitis. Preclinical studies have clearly

demonstrated that tegaserod has no vasoconstrictive

potential. There have been no cases of ischemic

colitis in clinical trials on Zelnorm, and the

reporting rate in the postmarketing experience is

consistent with the background rate of IC in the

IBS population even taking under-reporting into

account. These data indicate that Zelnorm does not

increase the risk of ischemic colitis.

Now, there have been four spontaneous

reports of fatalities in patients with intestinal

ischemia from postmarketing reviews. We take these

reports very seriously and have investigated them

thoroughly. Based on our individual and careful

review of each case, we are confident that Zelnorm

did not cause these fatalities.

The four fatalities are as follows: One

case with untreated central line sepsis and

ischemic colitis; one case of untreated chronic

abdominal angina; one case with untreated

hypothyroidism leading to severe fecal impaction;

and one case of multi-organ failure from unknown

cause.

Dr. Shetzline has carefully investigated

these cases and he will now discuss them in some

detail with you. Please, Dr. Shetzline?

Fatality Cases

DR. SHETZLINE: Thanks, Dr. Joelsson. I

am Michael Shetzline, an gastroenterologist and a

senior medical director at Novartis, responsible

for Zelnorm in the United States. Myself and Dr.

Christian Avery are clinical safety experts

responsible for evaluation of these cases.

I would like to review these cases in some

detail. Given the complicated nature of the cases,

it is important for us to go into some detail in

order to separate out and look at the medical

issues and make them clear. The first case is a 76

year-old woman. Her past medical history is

significant for 16 years of constipation. She had

IBS with constipation diagnosed in the year 2000

and started on Zelnorm in November of 2002. She

also had dementia of the Alzheimer's type.

In late August of 2003, after 282 days of

Zelnorm use the patient was found "down" at home.

She presented at the emergency department and was

admitted with abdominal pain, vomiting,

hypotension, hypothermia and altered mental status.

Her urine eventually grew E. coli and an abdominal

CT noted dilated loops of small bowel, consistent

with partial small bowel obstruction,

diverticulosis and focal ischemic changes of the

left colon. She was treated with antibiotics and

hydration.

During this admission she had a

colonoscopy for an incidental episode of guaiac

positive stool, and this revealed sigmoid and

splenic flexure ulcers with areas of regeneration

and healing, consistent with ischemic colitis.

Zelnorm was discontinued at this time. Biopsies

were consistent with ischemic colitis and she was

placed on bowel rest and provided total parenteral

nutrition.

She was eventually transferred to an

extended care facility and had two colonoscopies on

September 17th and 19th. Both noted improved

colonic mucosa, resolving ischemic colitis. This

is the usual expected course of ischemic colitis.

However, she remained on TPN, total parenteral

nutrition. She became hypotensive with E. coli UTI

and was readmitted on September 26th for failure to

thrive, febrile and more acutely ill.

After discussion with the family and

patient, no heroic surgical interventions and/or

CPR were to be performed; only supportive care.

She was diagnosed with central line sepsis and,

given her medical co-morbidities and discussion

with the patient and family, a "do not resuscitate"

order was initiated. Her antibiotics were

discontinue on October 1st and she expired. In

summary, this event of ischemic colitis resulted

from hypotension and possibly urosepsis.

The second case is a 66 year-old female

who had a past medical history of hypertension,

chronic obstructive pulmonary disease and tobacco

abuse. She had a prior stroke in 1997 due to small

vessel disease, and carried a prior diagnosis of

non-specific chronic colitis. Significantly, she

had symptoms of abdominal angina for 2-3 years

characterized by chronic abdominal pain with food

intake, and this resulted in 36 lbs of weight loss

during this interval. Her IBS was diagnosed in

January of 2000.

In October of 2003 she had continued and

more severe post-prandial abdominal pain and

constipation. On October 10th she was given

samples of Zelnorm, 6 mg BID, by her primary care

physician. She was not given a prescription and

her caregiver, who was responsible for

administering all her medications, the medications

taken by the patient, does not recall the patient

taking Zelnorm. He does specifically recall her

increasing use of Vicodin due to this more severe

abdominal pain.

On October 15th she was admitted to the

hospital with severe abdominal pain and bloody

diarrhea. Zelnorm was not listed as an active

medication in any of her admission documents.

On the 19th she developed acute abdomen

and had an exploratory laparotomy for, quote,

probable chronic intestinal ischemia, acutely

worse, end quote. The laparotomy revealed

infarcted bowel from the ligament of Treitz to the

terminal ileum, cecum, and proximal ascending

colon, consistent with occlusion of the superior

mesenteric artery. Given the extensive bowel

necrosis, comfort measures were provided and she

expired. The cause of death was listed as bowel

infarction due to peripheral vascular disease. In

summary, this is the natural history of end-stage

chronic abdominal or mesenteric angina and it is

likely Zelnorm was not taken by this patient.

The third case is a 41 year-old woman who

had a very significant past medical history of

chronic obstructive pulmonary disease. She had a

very extensive history of tobacco abuse, with 60-90

pack-years of tobacco use for a 41 year-old woman.

She also had asthma, prior alcohol and illicit drug

use, as well as obsessive-compulsive disorder. She

had peripheral vascular disease with claudication,

constipation, recurrent urinary tract infections

and hypothyroidism. She also had a significant

abdominal event due to appendectomy with a rupture

which resulted in abscess formation and a partial

colectomy. She had medical and medication

non-compliance, as noted in a primary care visit

from November of 2003.

This individual was presumably prescribed

Zelnorm in March of 2003, however, these documents

were not available for review. We have no

follow-up from the March presumed prescription and

the November primary care records which document

her non-compliance. She developed severe abdominal

pain and she collapsed with a cardiorespiratory

arrest. After an emergency medical service call

she was resuscitated and admitted.

It is important to note that admission

documents list only her Lithobid and Seroquel as

active medications. They do not list Zelnorm or

her thyroid supplement. These documents include

emergency medical service notes at home, admission

notes and multiple physician evaluations.

On admission, her abdominal x-ray revealed

free air in the abdomen and an exploratory

laparotomy demonstrated a rectal sigmoid densely

packed with rock-hard stool. She had ischemic

colitis and enteritis involving the colon and

terminal ileum and early gangrene of the distal

bowel. She had marked dilatation, a picture

consistent with toxic megacolon.

She had a sub-total colectomy with

ileostomy and was treated with ventilatory support,

broad-spectrum antibiotics and vasopressors. A

subsequent neurology evaluation revealed anoxic

brain injury with diffuse edema and a suspicion of

herniation. She developed multi-organ failure and

expired three days after admission. In summary,

this patient had a bowel obstruction from likely

untreated hypothyroidism due to her medication

non-compliance and a secondary perforation. Given

her medical and medication non-compliance, it is

likely she never took Zelnorm.

The last case is a 67 year-old woman who

had a very significant history of heart disease,

with known coronary disease, a prior coronary

bypass graft procedure, angioplasty with stent

placements, known occluded grafts, congestive heart

failure, hypotension, atrial fibrillation and

diabetes. She had chronic and acute renal failure.

She was on Zelnorm 6 mg BID for an unknown

indication from June 16th to August 7th, the date

of this event.

She as admitted at this time with

progressive shoulder and chest pain, as well as

shortness of breath, and was hospitalized, on

telemetry for a rule-out myocardial infarction, on

August 7th. On admission, her abdomen was soft,

non-tender. Her lungs had few bibasal rales and

her extremities showed trace pedal edema.

It is important to note that at this time

she had no diarrhea, no melena and no bright red

per rectum. On hospital day 3 she complained of

abdominal pain and nausea, and surgical consult

indicated a soft abdomen which was not distended.

She did, however, have left lower quadrant

tenderness and a questionable diverticulitis. An

abdominal x-ray at this time showed a large amount

of fecal material in the colon There was no

gaseous distention or free air.

On the same day laboratory results

indicated an amylase of over 7,000 and a lipase of

over 400. A pulmonary consult for dyspnea

indicated respiratory failure and she required

mechanical ventilation. At this time she was

evaluated for bronchitis, pneumonia, rule-out

abdominal sepsis, rule-out ischemic colitis,

coronary disease and hypotension.

A clinical evaluation noted, quote, in

view of her acute deterioration and chronic medical

problems, her prognosis is extremely poor.

Consequently, continuation of heroic interventions

may be inappropriate, end of quote. A cardiologist

summary indicated hypotension and it was felt that

the patient had a catastrophic abdominal event.

This may have included ischemic bowel, possible

perforation, pancreatitis, acute renal failure, all

in addition to her known co-morbidities of ischemic

cardiomyopathy, congestive heart failure, renal

failure and diabetes. The patient was made a "no

code" and died on hospital day 4.

The death certificate listed

cardiorespiratory failure as the primary immediate

cause of death. Other factors included shock,

pancreatitis and inflammatory bowel disease. In

summary, this patient experienced cardiovascular

collapse with a history of coronary disease and

congestive heart failure, as well as other medical

co-morbidities. This was likely unrelated to

Zelnorm.

Now I would like to return the safety

update to Dr. Joelsson.

Zelnorm: Safety Overview (continued)

DR. JOELSSON: Thank you, Dr. Shetzline.

In summary, these four cases, as you may

understand, are very complicated. In two cases it

is actually unclear if the patients actually took

Zelnorm in the first place. In our opinion and

those of external experts that have reviewed these

cases, the evidence does not support that the death

of these patients was caused by or contributed to

by Zelnorm.

At the time of approval in July, 2002,

biliary tract disorders were discussed because

there was an imbalance of cholecystectomies

introduction he clinical trials. When pooling the

clinical trial data, there is still an imbalance in

favor of placebo, although smaller than was seen in

the approval trials.

An adjudication was performed with outside

experts, resulting in a rate of 0.06 percent in

Zelnorm-treated patients versus 0.03 percent on

placebo.

In the postmarketing experience there were

30 reports of biliary tract events in approximately

3 million patients, and 18 were cholecystectomies;

2 were cholelithiasis; and 10 were other events.

There were no serious sequelae reported from these

patients.

In order to further elucidate the possible

effects of Zelnorm on gallbladder function a very

thorough study was performed using dynamic

ultrasound measurements. No effect on gallbladder

function was detected. There was no impact on

ejection fraction, ejection rate and period, or

maximal emptying. There was no impact on fasting

and residual volume, and there were no stimulus

effects on gallbladder contraction during fasting.

Based on this data, it is unlikely that Zelnorm

affects gallbladder function.

At approval there was also discussion

about ovarian cysts. However, ovarian cysts are

very well balanced in the clinical trials and there

are very rare reports from the postmarketing

experience. Our conclusion from these data is that

Zelnorm treatment does not increase the risk of

ovarian cysts.

Zelnorm has been extensively studied in

clinical trials and postmarketing experience, and

the safety profile of Zelnorm is very favorable.

In fact, Zelnorm has the overall safety profile of

placebo, with the only exception being diarrhea.

However, serious consequences of diarrhea are very

rare and do not result in significant clinical

sequelae. Evidence from either clinical trials or

postmarketing experience does not suggest that

Zelnorm increases the risk of rectal bleeding,

ischemic colitis, other forms of intestinal

ischemia, cholecystectomies or ovarian cysts.

Zelnorm is a safe and well-tolerated drug

that has a safety profile that supports its use in

chronic constipation patients. Thank you. I would

now like to introduce Dr. Schoenfeld who will

discuss with you his benefit/risk assessment.

Benefit/Risk Assessment

DR. SCHOENFELD: Well, good morning, Dr.

Fogel, members of the advisory committee, FDA

officers, audience members. I am Philip

Schoenfeld, a gastroenterologist at the University

of Michigan School of Medicine. It is my pleasure

to present a risk/benefit analysis of the use of

tegaserod and traditional therapies for the

management of constipation.

Now, I sympathize with the members of the

advisory committee. You have been sitting here now

for over an hour and a half. I imagine that I

should keep this presentation brief but also as

stimulating as possible to maintain your attention.

I am going to present an evidence-based medicine

analysis, revising the randomized, controlled trial

data about the efficacy for tegaserod and

traditional therapies in the management of

constipation, and review the best available

clinical trial data about the safety of tegaserod

and traditional therapies in the management of

constipation.

I think it is particularly important to

consider the risk/benefit analysis not only for

tegaserod but also for alternative therapies for

constipation because, as a practicing

gastroenterologist, when I am treating a patient

with constipation I have to consider the

risk/benefit analysis for all of these possible

treatments when I select the best possible

treatment for my patient, and I certainly think

this is an important topic. As Dr. Prather pointed

out during her presentation, constipation is

common. It negatively impacts the quality of life

for patients who actively seek medical care, and

many constipated patients are dissatisfied with

available treatments.

Let's stop for a moment and think about

that last statement, and look at the randomized,

controlled trial data about traditional therapies

for constipation to try to determine why

constipated patients might not be satisfied with

available therapies.

This is a partial list of the commonly

used treatments for constipation. They include

surface-acting agents like dioctyl sodium

sulfosuccinate--I had to practice saying that and

hereafter I will refer to it as Colace; bulking

agents like psyllium; stimulant laxatives and

osmotic laxatives like PEG-3350. These are all

FDA-approved treatments for constipation.

Dr. Prizont, in his efficacy section in

the FDA briefing document, provided a brief but

very balanced review about traditional therapies

for constipation. He specifically noted that there

are some randomized, controlled trials looking at

the benefits of traditional therapies for

constipation but many of these were conducted under

deficient designs. In other words, many of these

studies did not meet the Rome committee criteria

for appropriate design of a functional GI disorder

trial. They had inappropriately small sample

sizes. They had inadequate blinding. They had

very vague or imprecise criteria to identify

patients with constipation.

Now, in the briefing document Dr. Prizont

concluded that these trials revealed little

differences between laxatives and modest

improvement over placebo. He actually referenced

the most recent and most comprehensive

meta-analysis about traditional therapies for

laxatives, conducted by Jones and Nick Talley and

colleagues. In fact, the actual title of that

meta-analysis is "The Lack of Objective Evidence of

Efficacy of Laxatives in Chronic Constipation."

That is quite a provocative title.

Let's delve into that study a little bit

further to see how they came up with that. In

brief trials of 4 weeks or less in duration, they

found that laxatives increased stool frequency by

about 2 stools per week compared to baseline.

Placebo increased stool frequency by about 1 stool

per week compared to baseline. But as you note,

the 95 percent confidence intervals here for

placebo and laxatives are superimposed, not clearly

demonstrating a difference in efficacy.

For trials of 5-12 weeks in duration the

results are less impressive. Laxatives increased

stool frequency by only 1 bowel movement per week

versus placebo-treated patients who had an increase

in stool frequency of 1.5 bowel movements per week.

Now, I think we should be cautious about

interpreting these results. This is a

meta-analysis that provides a single summary

statistic and combines the results from bulking

agents, stimulant laxatives and osmotic laxatives.

So, it might be more beneficial to look at a

systematic review that at least separated out

bulking agents from other types of laxatives.

That is actually available. The other

well-designed, systematic review about traditional

therapies for constipation comes from Tramonte and

Cindy Mulrow and colleagues, at the Cochrane Center

in San Antonio, Texas. They separated out bulking

agents versus laxatives and found that bulking

agents increase stool frequency by about 1.4 stools

per week compared to baseline and laxatives

increase stool frequency by about 1.5 stools per

week compared to baseline. So, their study

conclusions were that fiber and laxatives do appear

to modestly increase stool frequency over placebo.

100

They also concluded that it was unknown if these

agents would improve general well being or global

satisfaction because this endpoint wasn't examined

in many of these trials.

Now, it is beyond the scope of my

presentation to individually review each of the

randomized, controlled trials looking at

traditional therapies and I am sure you wouldn't

want to sit through all of that. But I will

conclude by noting that the randomized, controlled

trial evidence for psyllium, PEG-3355 and lactulose

consistently demonstrates significant increases in

stool frequency versus placebo. On the other hand,

other commonly used and FDA-approved treatments for

constipation, such as bisacodyl, Surfak, Colace,

consistently do not demonstrate a significant

increase in stool frequency versus placebo. It

doesn't necessarily mean that these drugs are

ineffective. As Dr. Prizont noted, most of these

RCTs were carried out under deficient designs and

if appropriately designed studies that met the Rome

committee criteria were conducted, we might be able

101

to demonstrate efficacy. Nevertheless, when I am

selecting a treatment for constipation I have to

consider not just my clinical experience but also

the randomized, controlled trial data of efficacy

as well as the clinical trial data of safety.

There are several other issues for

discussion today. First, whether or not the

clinical trial data are adequate with respect to

the chronic constipation population that is likely

to be treated with tegaserod. I would just

reemphasize part of Dr. Dennis' presentation, the

two Novartis randomized, controlled trials

contained inclusion criteria that are very similar

to the Rome II committee criteria for functional

constipation. In fact, in some ways they are more

stringent.

Patients had to have greater than 6 months

of symptoms by the Novartis criteria, whereas the

Rome criteria require only 12 weeks, which need not

be consecutive, of symptoms in the previous year.

The Novartis criteria required that patients have

fewer than 3 spontaneous bowel movements per week.

102

That is not an actually requirement to meet the

Rome committee criteria for functional

constipation. A patient, for example, who just had

straining and lumpy, hard stools for 12

non-consecutive weeks would meet the Rome committee

criteria for functional constipation.

Certainly, I think it is very true that 78

percent of the patients in these RCTs appear to

have chronic constipation while as many as 22

percent had some symptoms of abdominal discomfort

that might have led them to be classified as IBS

with constipation. Nevertheless, Miss Mealey, the

FDA statistical reviewer, in her very thorough and

comprehensive statistical review noted that the

responder rates for the constipated patients in

these trials who didn't have IBS-like symptoms were

similar to the overall responder rates. In fact,

they tended to do better than the overall response

rates that were recorded.

Certainly, the issue has been raised about

whether or not we may have subtypes of patients

with slow transit constipation included in this

103

trial. As a clinician, my main point about that

would be that the AGA's medical position statement

about that provides essentially identical treatment

algorithms whether somebody has normal transit

constipation or slow transit constipation. So, my

choice of therapy wouldn't necessarily differ based

on whether or not there might have been some

patients with slow transit constipation included in

these trials.

Another issue for discussion is the

appropriateness of a primary endpoint of an

increase of 1 or more complete spontaneous bowel

movements per week compared to baseline versus the

percentage of patients who attained 3 or more

complete spontaneous bowel movements per week.

This is a difficult issue. The Rome II committee

actually couldn't come to a consensus about what

was the most appropriate endpoint for trials of IBS

and functional constipation. They recognized that

there are multiple symptoms present in patients

with these lower GI functional disorders. They

actually stated that in addition to whatever

104

primary endpoint is chosen, there should be a

select number of a priori defined secondary

endpoints that reflect the multiple symptoms that

are present in patients with these functional GI

disorders.

In fact, Sander Van Zanten and his

colleagues, who were on the subcommittee of the

Rome committee who laid out the appropriate design

of a trial of a functional GI disorder, actually

stated that global improvement in satisfaction may

be the most appropriate endpoint. That is an a

prior defined secondary endpoint in the 2 Novartis

randomized, controlled trials, that patients on

tegaserod 6 mg BID were significantly more likely

to be responders for global satisfaction than

patients that were on placebo. This is not only a

significant difference but, in my opinion, a

clinically important difference where the magnitude

of benefit is 9-12 percent more for patients on

tegaserod 6 mg BID who were responders for global

satisfaction compared to patients on placebo.

Again, there were a select number of a

105

priori defined secondary endpoints included, to try

to contrast that with traditional therapies that

patients on tegaserod 6 mg BID had 1.9 to 2 more

spontaneous bowel movements per week compared to

patients on placebo who had about 0.9 to 1 more

spontaneous bowel movements per week. These are

statistically significant differences.

If we do decide to apply the FDA's

criteria that patients have to have 3 or more

spontaneous bowel movements per week, and we look

at the proportion of patients who attained what is

a pretty high therapeutic goal, we still see that

almost twice as many patients on tegaserod

experienced 3 or more complete spontaneous bowel

movements per week compared to patients on placebo

and the magnitude of this difference, whether we

look at 4 weeks or the entire 12-week trials, is

approximately 10 percent. Again, in my opinion,

that is a clinically important difference.

So, in conclusion for efficacy, I would

state that the randomized, controlled trial data

about the efficacy of tegaserod is very robust and

106

precise. These are the best designed, most

comprehensive trials about treatments for

constipation that are available among all the

treatments that we have available for constipation.

The study population does reflect patients with

chronic constipation and the a priori defined

primary and secondary endpoints do reflect the

multiple symptoms that patients with constipation

have, and these RCT data consistently demonstrate

that tegaserod produces significant and clinically

important improvement in the multiple symptoms of

constipation.

Let's move on to safety. Unfortunately,

there is very little data about the safety of

traditional therapies for constipation. The most

recent and comprehensive meta-analysis by Jones,

Nick Talley and colleagues actually didn't even

address the issue because the data were so scant.

If we do go back to the systematic review performed

by Tramonte and Cindy Mulrow and colleagues from

the Cochrane Center in San Antonio, Texas, they

specifically noted that few studies used

107

standardized techniques to assess adverse events.

They also did note that they did not identify any

significant differences in adverse events between

laxatives and placebo. So, they concluded that

although there is no evidence that laxatives are

unduly harmful, the data available are very limited

and short-term.

Thus, we are really left with looking at

the prescribing information and case report data to

try to get an idea about what adverse events are

associated with commonly used laxatives. We see

for bulking agents that fecal impaction and large

bowel obstruction have been reported, and even

acute esophageal obstruction when bulking agents

aren't taken with adequate amounts of water.

Anaphylaxis has been reported with psyllium. Among

osmotic agents all different types of electrolyte

abnormalities have been reported, specifically with

magnesium-based agents that are used on a regular

basis. Stimulant laxatives have been associated

with both electrolyte imbalances as well as

abdominal cramps. All of these agents have been

108

reported to have been associated with diarrhea.

So, another issue for discussion is

whether or not the clinical trial data and

postmarketing surveillance data provide adequate

evidence of safety. I pause here for a moment,

looking at the title of this slide, to just note

that the clinical trial safety data where patients

are followed per protocol probably provides at

least the most precise safety data that we have

available to us. When we look at the clinical

trial safety data available for tegaserod we see

that in the Novartis 2 randomized, controlled

trials over 2,600 patients with constipation were

enrolled. Over 1,700 received tegaserod. In the

whole clinical trial safety database you have over

11,000 patients treated with tegaserod and over

3,400 patient-years of tegaserod use followed

within the context of clinical trials. I would

suggest that infers that there is very robust and

precise clinical trial safety data for tegaserod,

certainly more robust and precise clinical trial

safety data than what we have available for any

109

other treatment of constipation.

That very precise data allows us to

estimate what is the likelihood of serious adverse

events for constipated patients using tegaserod or

placebo. We see essentially similar rates. And,

that very robust and precise safety data let's us

quantify the likelihood of diarrhea as an adverse

event. Among constipated patients we see that it

is reported as an adverse event in 5 percent of

patients in clinical trials versus 3 percent in

patients on placebo. We see that 0.6 percent of

patients treated with tegaserod actually

discontinued the medication because of the severity

of their diarrhea. When we look at the entire

clinical trial database we can estimate that the

likelihood of clinically serious consequences of

diarrhea--going to the emergency department because

of dehydration and getting IV fluids, virtually

being hospitalized because of syncopal

episode--occurs in 0.04 percent or 1 in 2,500

patients treated with tegaserod.

To conclude, let's turn to the safety

110

issue about ischemic colitis. Obviously as

gastroenterologists, as primary care providers for

patients, we are concerned about the issue of

ischemic colitis because it has been brought to our

attention by our clinical experience with

alosetron. Alosetron, again, is an antagonist of

the 5-HT 3 serotonin receptor

as opposed to

tegaserod that is an agonist of the 5-HT

4 serotonin

receptor. We know from the clinical trial data

that there were 17 cases of ischemic colitis among

the 10,805 alosetron-treated patients in those

clinical trials. That calculates out to a rate of

5.9 cases of ischemic colitis per 1,000

patient-years based on the clinical trial data.

I would also like to point out the fact

that among placebo-treated patients with IBS the

rate of ischemic colitis was 1.1 cases per 1,000

patient-years. Even in the context of this

clinical trial, there was a background rate of

ischemic colitis among patients treated with

placebo.

So, what can we do about comparing the

111

issue of ischemic colitis with tegaserod patients

treated for constipation versus other patients

treated for constipation? The only other treatment

for constipation that has any breadth of clinical

trial safety data is PEG-3350. In their new drug

application to the FDA they reported a rate of

ischemic colitis in their clinical trial safety

data of 3 cases per 1,000 patient-years. I want to

emphasize that that is an extrapolation. The exact

number is that there was 1 case of ischemic colitis

among 300 patient-years of clinical trial safety

data when they submitted their new drug

application. That is the only other treatment for

constipation where we have any breadth of clinical

trial safety data to estimate the likelihood of

patients experiencing ischemic colitis.

What is the data for tegaserod? Zero

cases among 11,640 tegaserod-treated patients

studied over 3,400 patient-years of exposure

versus, among all the clinical trial database for

placebo-treated patients, 1 probable case of

ischemic colitis among 4,267 placebo-treated

112

patients followed for 780 patient-years of

exposure.

Now, the FDA officials wanted to identify,

based on that clinical trial data--zero cases among

all the patients followed in the clinical trial

database--what would be the maximal rate of

ischemic colitis that still could be occurring

within 95 percent confidence intervals. So, they

did their statistical analysis based initially on

the 7,000 tegaserod-treated patients in clinical

trials that they had reviewed and they came up with

a maximal ischemic colitis rate, within the

confines of 95 percent confidence intervals, of 1

case in approximately 2,000 patients, based on the

fact that there were zero reported cases among over

7,000 patients.

If we give the up to date analysis based

on all 11,640 tegaserod-treated patients, then a

similar statistical analysis would shoe that the

maximal rate within 95 percent confidence

intervals, considering there are zero cases among

11,640 patients, would be 1 case in 3,883 patients.

113

In order to be balanced, I think we should

consider the placebo patients too. The same

statistical analysis shows that their maximal rate

would be 1 case in 867 placebo-treated patients.

This analysis is still based on very few

cases of ischemic colitis. So, I certainly

understand the need to look at postmarketing

surveillance data. In the U.S. over 2 million

prescriptions, accounting for over 233,000

patient-years of use; 26 reported cases of possible

ischemic colitis, equating to a rate of

approximately 12 cases per 100,0000 patient- years.

Again, as pointed out during Dr.

Joelsson's presentation, patients with irritable

bowel syndrome seem to be diagnosed with ischemic

colitis more often than the general population. It

may be an ascertainment bias because these patients

tend to be scoped more frequently. It may be due

to an unknown pathophysiologic factor. Obviously,

there is recent research to indicate there are true

pathophysiologic differences among IBS patients.

But regardless of which epidemiologic study we look

114

at, all the available epidemiologic data indicates

that patients with IBS are 3-4 times more likely to

be diagnosed with ischemic colitis than is the

general population, and the rates vary depending on

the age of the population that is examined.

Obviously, the Medi-Cal population tended to be

older than the patients studied in the United

Health Care study and, thus, we are seeing a higher

rate of ischemic colitis both in the general

population and in the IBS population.

I certainly comment Dr. Brinker. He did a

very interesting analysis of the United Health Care

base and identified that, clearly, when a patient

is diagnosed with IBS their rate of getting

subsequently diagnosed with ischemic colitis within

the next 3 weeks is very high. Those patients

almost certainly are patients that are misdiagnosed

with IBS when they really have ischemic colitis.

Nevertheless, the same analysis found that patients

who had a stable IBS diagnosis for over 1 year

still had a rate of 53 cases per 100,000

patient-years compared to the general population

115

where it was 7 cases per 100,000 patient-years.

I do want to make particular note here.

When I talked about the clinical trial data my

denominator was 1,000 patient-years. We have now

shifted. All the postmarketing surveillance data

is based on a denominator of 100,000 patient-years

of use.

Postmarketing surveillance data for IBS

patients treated with tegaserod is 12 cases per

100,000 patient-years, which is 4- to 15-fold lower

than the expected rate, although I certainly

understand there may be some under-reporting, and

it very difficult to get an estimate for how often

that occurs.

So, in conclusion, I certainly think that

the clinical trial safety data for tegaserod is

more robust and more precise than it is for any

other treatment that we have available for

constipation. This safety data allows us to have a

very precise estimate of the likelihood of

clinically serious consequences of diarrhea, but

the evidence doesn't support an association between

116

tegaserod and ischemic colitis.

When I do a risk/benefit analysis I see

the benefits being this robust clinical trial data

that demonstrates that tegaserod is efficacious for

the treatment of constipation, especially the

multiple symptoms of constipation, and that the

safety data is more robust than it is for any other

treatment I might choose and that safety data from

clinical trials demonstrates to me that there is a

very low but finite risk of clinically serious

consequences of diarrhea.

So, that analysis demonstrates to me that

tegaserod has a very favorable risk/benefit profile

in the management of chronic constipation, and that

it compares very favorably with the risk/benefit

analysis for any other therapy that I might choose

to use to treat patients with constipation. Thanks

very, very much for your attention and I will turn

the program back over to Dr. Fogel.

Questions on Presentations

DR. FOGEL: I would like to thank the

presenters for their informative presentations. At

117

this juncture we turn the meeting over to the

committee for questions to the presenters. Dr.

D'Agostino I think had his hand up first, and then

Dr. Sachar.

DR. D'AGOSTINO: When I saw there was a

two-hour presentation I said, my God, they will

never take that long but it actually was a great

presentation. Thank you very much.

I have a comment about the subset

analysis, which we will have to address later. I

understand that you look at subsets for consistency

and not necessarily expecting to see significant

results, but shouldn't we be concerned that we are

not seeing the effect lying on the right side with

the elderly greater than or equal to 65 and the

males, and the Blacks? Can you give us some words

on how we can feel comfort that you aren't seeing

the effect in greater than or equal to 65 year-old

individuals and also males, and I would like

something on the Blacks also.

DR. DENNIS: Thank you for that question.

Yes, absolutely. Can I have slide AQ-16, please?

118

We will start off with the elderly population since

that was the question that you asked initially. As

you know, we only randomized 13 percent of our

patients that were 65 years or older, and this is

the responders by age group looking at our primary

efficacy analysis.

What we can see in the group that is 65

years and older is that we are seeing a treatment

effect in the patients that are on Zelnorm. We are

also seeing an effect in patients on placebo as

well. So, the interesting thing though is that we

can break this down further by looking at the older

population by age and by gender.

If I could have the next slide, please,

which is AQ-17, let me show you what happens when

we break it down into further subsets. On the top

row we are seeing female patients and on the bottom

row we are seeing male patients. The patients that

are less than 65 years old--if we start with that

column on the left-hand side, we can see that the

effect in the younger female population is similar

to the effect in the younger male population.

119

Remember that we have much fewer numbers in terms

of the male group so we don't reach statistical

significance because, as I said before, these

subgroup analyses are not powered to detect

statistical significance. So, I think we are

seeing a consistent effect in the men that are 65

years and younger that we are seeing in the female

population.

If we look at the slide on the right-hand

side, and let's turn to the elderly population, we

do see an effect in female population. Of course,

the effect size is slightly smaller--again, small

numbers of patients, and when we will look at the

male patient population that are 65 years and older

we are seeing that there really is no effect

looking at it on this particular analysis.

But I am going to take it one step further

and take out those patients that we felt were

probably IBS-like because, if you remember, in our

overall efficacy analysis when we took that group

out the efficacy was slightly more robust.

So, if I can have the next slide, which is

120

AQ-18, this shows you what happens when we take out

those patients that are IBS-like. I am going to

focus your attention on that male population that

is 65 years and older. You can see that in the

previous analysis--here we have really small

numbers of patients. We are dealing with 20

patients in that group that are on 6 mg BID. So,

the responders that we saw in the previous analysis

were all chronic constipation patients and when we

take out the other patients that have IBS obviously

our denominator changes and, you know, we see a

much more different effect looking at these

numbers. But I do want to caution that these are

very small numbers when we are looking at these

subgroup analyses.

But if we look at the four different

quadrants I think we can see the effect in the

patients less than 65 is similar in men as it is in

women. I think we are seeing an effect in elderly

females, and I think we are seeing an effect in

elderly males when you take out the IBS-like

subset.

121

DR. FOGEL: Dr. Sachar?

DR. SACHAR: With the permission of the

chair, if I could address some questions to each of

the four major presenters, Dr. Schoenfeld, when you

presented your efficacy data in slides 13 and 14

you appeared to have limited your analysis only to

the highest dose tegaserod of 6 mg BID. Yet, when

you discussed the adverse effects you combined the

2 mg and the 6 mg doses. It would seem to me if

you really want to look at a benefit/risk ratio we

really ought to look at a comparison for the same

doses. If we were to do that, we would find in

your slide 21 that it really isn't 5.4 percent of

patients but is actually 6.6 percent of patients

who had some adverse effect with diarrhea at the

equivalent dose, at the 6 mg dose.

I am not a professional statistician but

if we go a little further we might say that since

the number needed to treat--to get some benefit,

some demonstrated benefit from this drug is

approximately 10. It ranged between 9-11 in all

the analyses. It is approximately 10. The number

122

needed to treat to see some adverse effect from

diarrhea is actually about 2.8 at the equivalent

dose. So, would it be fair to say that for every 3

patients who get some benefit from this drug 1 will

experience some diarrhea?

DR. SCHOENFELD: No, I would not go along

with that and I think there are multiple points

there to address. The first one is that all of us

have conducted clinical trials and, as we

recognize, reporting an adverse event in the

context of a clinical trial is not the same as

suffering a clinically important adverse event.

When these patients are followed in the context of

clinical trials, to paraphrase, your study nurse

may say, "anything unusual happen in the past

week?" And, if the patient says, "I had a little

bit of diarrhea last Thursday," that becomes an

adverse event. What is probably a much more

appropriate adverse event category to assess,

clinically important adverse events, is how often

patients stop their medication due to diarrhea.

Obviously, here it is 0.6 percent.

123

Having said that, I certainly take your

comment appropriately, that if you look at the 6 mg

BID dose the rate at which diarrhea was reported as

an adverse event is about 6.6 or 6.7. For the

broader issue though of safety, my own

experience--and there are other experts here that

have far more experience in safety issues--is that

when we look at efficacy we want to look at what is

going to be most likely the dose that is utilized.

But in safety we tend to look at multiple different

dose ranges to find out what the benefit is.

Having said that, I think a subgroup

analysis about what the rate would be for 6 mg BID

versus 2 mg BID would be helpful, although I will

mention for the most serious adverse events that we

are concerned about here, which in my mind are

really ischemic colitis, when you look at 6 mg BID

or 2 mg BID it is still going to be zero events.

You are just going to change your denominator a

bit. And, the majority of patients in these trials

were treated with 6 mg BID.

DR. SACHAR: Agreed. When you talk about

124

the diarrhea issue that brings me to the one

question for Dr. Joelsson, and that is simply that

you did discuss the physiologically serious

consequences and those were obviously very, very

low. Did anybody record whether any patient with

diarrhea had any episode of incontinence?

DR. JOELSSON: We have not that recorded.

I cannot answer that.

DR. SACHAR: Because that is sort of an

impact thing.

DR. JOELSSON: Yes.

DR. SACHAR: And for Dr. Dennis, in slide

46--

DR. DENNIS: I will flash it up on the

screen.

DR. SACHAR: Yes, it is very important, as

everybody has indicated, that when you excluded IBS

from the analysis you still showed efficacy. I

think that is a very important point. But you

showed us the data for doing that only at week 1-4.

DR. DENNIS: Yes.

DR. SACHAR: Do you have any data on that

125

for the 12-week point?

DR. DENNIS: It looked similar. I don't

have the data on a slide but it does look similar

over the 12-week treatment period.

DR. SACHAR: It is the same approximately?

DR. DENNIS: Yes.

DR. SACHAR: Great, fine. In slide 13 you

showed us that, in terms of the inclusion criteria,

they had to have a bowel evaluation within the past

5 years. Do I take that to mean that if some

patients had early constipation symptoms 3 years

ago or 4 years ago or 5 years ago and they had a

barium enema, and then more recently the symptoms

persisted or worsened and they represent that they

would be eligible to go in this study without any

new reexamination?

DR. DENNIS: If they had had a bowel

evaluation that was after the onset of symptoms and

the symptoms remained the same within the past 5

years there was no need for them to have a

reevaluation. However, if there was any change in

the symptoms or if there were any alarm features,

126

as I said, anything that suggested rectal bleeding,

hemorrhage, anemia or any change in the pattern,

those patients would have had to have a new

evaluation. But it was stable patients who had

been having symptoms that had remained the same

within the time they had the evaluation.

DR. SACHAR: Great! My last question is

for Dr. Prather. I am not actually familiar with

the Canadian study of Pare et al., but you

indicated it was a population study. Does the

population in that study represent the group

receiving and taking medications for chronic

constipation? Is it a clinic-based or a true

population-based study? Because if it is truly

population based it doesn't reflect people who are

taking medications for their constipation.

DR. PRATHER: It was, indeed, a

population-based study but that would include

all-comers with constipation that were actually in

the population. So, it didn't discriminate against

individuals who may or may not have seen a

physician for their constipation.

127

DR. SACHAR: Right, so that means that in

Larry Schiller's study that you showed in slides 19

and 20 with all the dissatisfaction, that included

patients who had taken over-the-counter

preparations or had seen a GP, or something, and

had been perfectly satisfied? Or, was it only the

dissatisfied patients who sought out GI specialists

who were in that study?

DR. PRATHER: This included

individuals--it was a study that was done through

the Internet that was representative of the U.S.

population, but with the initial questions,

actually to get into the study they had to have

seen a physician within the past 12 months for

constipation.

DR. SACHAR: A physician or a

gastroenterologist?

DR. PRATHER: Actually, these were primary

care physicians predominantly, yes.

DR. FOGEL: To increase the number of

questions that we can ask during our time frame, I

would like the members of the committee to keep

128

their comments brief. I am going to take the

prerogative of the chair and ask my questions of

Dr. Dennis.

Can you provide us additional details

regarding the question that you asked for

subjective global assessment, and can you tell us

how the data was analyzed and whether responders

had a persistent response over the 12 weeks of the

study?

DR. DENNIS: We asked the question how

satisfied were you with your bowel habits over the

past week? And, the responses were a very great

deal satisfied; a good deal satisfied; moderately

satisfied; hardly satisfied; and not at all

satisfied. We defined a responder as having a

decrease of 1 on the satisfaction score.

I am going to first show you some data on

the persistence of satisfaction response and then I

will call one of the statisticians to actually come

up and explain to you the statistical analysis that

was done.

If I could have slide AQ-58, please? This

129

is an analysis that we did where we said to those

patients that met the score of a very great deal

satisfied and a good deal satisfied, so zero and 1,

for at least 50 percent of the weeks of the whole

trial, which is 12 weeks. What we can see on the

study is definitely a significant benefit of

Zelnorm versus placebo when we look at patients

that had satisfaction over 6 weeks of the 12-week

treatment period. So, I think we are seeing

persistence of the satisfaction result.

I am going to call upon Dr. Jeen Liu, who

is our statistician, to come and respond to your

question about how these were actually calculated.

DR. LIU: My name is Jeen Liu. I am the

statistician from Novartis responsible for this

project. The slide that Dr. Dennis just showed was

a response rate that we defined--actually, she

showed two slides for two time intervals, weeks 1-4

weeks and 1-12. What we did was we took the

patient score at each week, averaged them over the

respective time intervals, either 4 weeks or 12

weeks, and compared that with the baseline score

130

that each patient had during the 2 weeks prior to

treatment, and got the difference and compared with

it was a decrease of 1 or more. If it is 1 or

more, it is a responder; otherwise the patient was

a non-responder. Thank you.

DR. FOGEL: Thank you. Dr. Metz?

DR. METZ: Great, thanks. Just in the

interest of time, I am going to float a few

questions to you. I want to thank you for a nice,

comprehensive presentation. Three areas to

address, first of all, the problem with the

subgroup analysis, as has been alluded to. Can you

perhaps tell me why you chose 65 years? I would be

more interested in actually seeing a median age

above and below, perhaps divided into quartiles

above that and see where you actually see your

cut-off. I am not sure why 65 is necessarily

relevant.

The second question will be a little bit

about the loss of efficacy in one of your two

pivotal trials in the 2 mg group. It appears to me

more because of the placebo effect increasing up to

131

reach the 2 mg, but it makes one wonder a bit about

a tolerance response, and that brings me to why you

really chose the first 4 weeks. This is something

people are going to be using way beyond 12 weeks.

So, why the first 4 weeks; why not 12 weeks and

beyond as your primary outcome measurement?

The third question is use of surrogate

measurements, which you actually have in your

binder but didn't talk about at all today. That is

the use of rescue medication and seeing any

difference there as a sort of idea of, you know,

you are seeing an effect because of using less

rescue? Can you address those three points,

please?

DR. JOELSSON: While Dr. Dennis is

thinking about the second question I can take the

first question. The analysis of patients above 65

years and below 65 years is a very traditional

analysis that we do, which is based on what the FDA

wants us to do. This is kind of the cookbook thing

you do. So, it is not that it was anything that we

came up with; this is the traditional way of doing

132

it, and we don't have the data the way that you

describe. I am sorry about that.

DR. METZ: Do you think that would be a

useful examination to go through?

DR. JOELSSON: Yes, I agree.

DR. DENNIS: let me address the other

questions. I am first going to tackle the question

that you asked about loss of efficacy. I think

what we see in these clinical studies is that the

treatment effect of Zelnorm was sustained

throughout the entire treatment period. We did not

see a decrease in the number of responder rates.

There is certainly nothing to suggest that we saw a

loss of efficacy in terms of the drug response

itself. Placebo responses, as we know, are not

uncommon in clinical trials and we see them in all

clinical trials. You know, the issue is in some

clinical trials placebo responses continue to rise.

If I could go back to my core slide CE-28,

this shows you the weekly responder definition over

the 12-week treatment period, and I just want to

really point out again that we are seeing that the

133

efficacy is sustained throughout the entire

treatment period.

Maybe I can get a clarification, Dr. Metz.

Were you referring specifically to the 2 mg dose in

the 2301 study?

DR. METZ: That is correct. Clearly, you

don't see that in the 2302 but you do see it in the

2301.

DR. DENNIS: Absolutely. You know, I

think the 6 mg BID dose has emerged consistently as

being more efficacious and that is why we are going

for that dose as an indication. We have actually

looked at what are the reasons that could have, you

know, determined why we are seeing this in 2301 and

not 2302 because these two studies were essentially

identical in the core period. The only differences

that we can find are geographical. 2301 was done

mainly in Europe and 2302 was done in North and

South America.

To address the question of why we chose a

4-week duration period, I think that was because

when physicians prescribe the drug they want to

134

look at the effect size within 4 weeks. They want

to know is this drug going to work within 4 weeks

or not. So, we looked at 4 weeks as our primary

endpoint but we also looked at it over 12 weeks to

make sure that we would see sustained efficacy.

So, we have the data for both of those two

endpoints.

DR. METZ: Right, but the point I am

making is that this is a chronic problem. You are

defining chronic constipation as something that has

been around for more than 6 months--

DR. DENNIS: Sure.

DR. METZ: --and you are not going to

treat for 4 weeks and then stop.

DR. DENNIS: And that is why we have a

12-week treatment duration.

The last question that you had was

laxative use. There were very strict guidelines

for laxative use in these studies. Patients were

only allowed to take laxatives if they had not had

a bowel action for 96 hours. So, they had to wait

96 hours from the time of their last bowel action

135

before they could have a laxative. When we looked

at laxative intake in these particular studies, we

measured how many patients took at least one dose

of a laxative throughout the entire 12-week

treatment period, and we found that about 50

percent of patients in the study took a laxative at

some point during the study. But when we really

break this down and we say how frequently were

laxatives actually being taken, we find that

laxative intake, in fact, was quite infrequent.

This slide I am going to show you is going

to show you laxative use by mean number of days.

If you look at the baseline period, we see that

laxatives were used about once every 11-12 days.

In the double-blind, placebo group we see that

laxatives were used about once every 14 days and

about once every 18 days on Zelnorm.

If I could have the next slide, which is

AQ-62, this is going to show you the median days

data. Here we are seeing by median data of use

that the baseline laxative use was about 14 days a

week. The median use of laxatives in the placebo

136

group goes down to 0.11, which translates to 1

every 64 days. When you look at the

Zelnorm-treated group we are seeing that that goes

down to 0.08, which translates to once every 88

days. So, when you really look at it, laxative

intake is really very infrequent.

However, to speak to your point, we are

seeing that there is more laxative use in the group

on placebo than there is on Zelnorm, and if we are

expecting to see a confounder because of that, we

would expect to see it more in the placebo than we

would in the Zelnorm.

DR. FOGEL: Dr. Cryer?

DR. CRYER: Dr. Dennis, I would just like

to follow-up on this theme of the subgroup analysis

in those who were greater than 65 and those who

were men. You very strongly make the point that

Zelnorm, as you just showed us, has maintained

efficacy over the 12-week period. However, all of

your slides that you showed us to support its

observations in the subgroups of those who were

greater than 65 or those who were men were the

137

4-week data points. So, I am wondering whether you

can show us that, in fact, the sustained 12-week

data in that subgroup of men and those who were

greater than 65.

DR. DENNIS: Right. The reason why we did

the subgroup analysis on the 4-week data initially

was because that was our primary endpoint and so

that is why we determined to do that.

I don't actually have the slides with me

right now to show you the actual week 12 but I will

just confer with my colleagues and make sure we

have those before the end of the presentation.

DR. CRYER: I think this is a very

important point because when you consider the

potential target group for therapy, many of them,

as we have learned from Dr. Prather, are going to

be greater than 65 year-old age population. So, I

think in the assessment that we are making today it

would be very, very helpful for us to specifically

look at the effects in a target population.

DR. DENNIS: Right, and I will make sure

we have those slides and we will come back to that.

138

DR. FOGEL: You can present those slides

later.

DR. DENNIS: Right, thank you.

DR. FOGEL: The next question is by Dr.

Buchman.