1

DEPARTMENT OF HEALTH AND HUMAN SERVICES

FOOD AND DRUG ADMINISTRATION

CENTER FOR DRUG EVALUATION AND RESEARCH

ENDOCRINOLOGIC AND METABOLIC DRUGS

ADVISORY SUBCOMMITTEE

Wednesday, September 8, 2004

8:00 a.m.

Holiday Inn

Versailles Ballrooms

8120 Wisconsin Avenue

Bethesda, Maryland

PARTICIPANTS

Glenn Braunstein, M.D., Acting Chair

LCDR Dornette Spell-LeSane, M.H.A.,

Executive Secretary

MEMBERS:

Lynne L. Levitsky, M.D.

Louis J. Aronne, M.D.

Katherine Flegal, Ph.D., M.P.H.

Melanie G. Coffin (Patient Representative)

Frank L. Greenway, M.D.

Jules Hirsch, M.D.

Jack A. Yanovski, M.D., Ph.D.

Susan Z. Yanovski, M.D.

Paul D. Woolf, M.D.

Thomas O. Carpenter, M.D.

Dean A. Follmann, Ph.D.

Steven W. Ryder, M.D.

David S. Schade, M.D.

Morris Schambelan, M.D.

Nelson B. Watts, M.D.

Margaret E. Wierman, M.D.

C O N T E N T S

Call to Order and Introductions, Glenn Braunstein,

M.D., Cedars-Sinai Medical Center, UCLA School

of Medicine 4

Conflict of Interest Statement,

LCDR Dornette Spell-LeSane, Executive Secretary 6

Welcome and Introductory Comments, David Orloff,

M.D., Director, DMEDP 9

The Regulatory History of Weight-Loss Drugs,

Eric Colman, M.D., Medical Team Leader, DMEDP 14

The Epidemiology of Overweight and Obesity,

Katherine Flegal, Ph.D., Senior Research

Scientist, National Center for Health Statistics 38

Current Status of Weight-Loss Drugs, Frank

Greenway, M.D., Director, Pennington Biomedical

Research Center 71

Patterns of Weight-Loss Drug Use,

Laura A. Governale, Pharm.D., MBA, Drug

Utilization Specialist, Team Leader, Division of

Surveillance Research and Communications

Support, ODS 87

Role of Drugs in the Treatment of Obesity:

Current and Future, Richard L. Atkinson, M.D.,

Director, Obetech Obesity Research Center 106

Charge to the Committee, David Orloff, M.D.,

Director, DMEDP 150

Committee Discussion 186

P R O C E E D I N G S

Call to Order and Introductions

DR. BRAUNSTEIN: We will call the meeting

to order. This is the Food and Drug

Administration, Center for Drug Evaluation and

Research, Endocrinologic and Metabolic Drugs

Advisory Committee meeting, on September 8, 2004.

The agenda today is to discuss the FDA draft

guidance document entitled, "Guidance for the

Clinical Evaluation of Weight Control Drugs." The

original guidance was dated September 24, 1996.

I am Glenn Braunstein, Professor and

Chair, Department of Medicine, Cedars-Sinai Medical

Center. I am an endocrinologist. I would like to

go around the table and ask people to introduce

themselves and tell us where they are from. We

will start with Dr. Orloff.

DR. ORLOFF: I am David Orloff. I am

Director, Division of Metabolic and Endocrine Drugs

at FDA.

DR. COLMAN: I am Eric Colman, a medical

officer from Metabolic and Endocrine Drugs at FDA.

DR. HIRSCH: Jules Hirsch, Rockefeller

University, New York.

DR. SCHAMBELAN: Morris Schambelan, from

the University of California, San Francisco.

DR. FOLLMANN: Dean Follmann, from NIH.

DR. YANOVSKI: Jack Yanovski, from NIH.

DR. LEVITSKY: Lynne Levitsky, Pediatric

Endocrinology Unit, Massachusetts General.

MS. COFFIN: I am Melanie Coffin, patient

representative.

LCDR SPELL-LESANE: Dornette Spell-LeSane,

executive secretary for the committee.

DR. GREENWAY: I am Frank Greenway, from

the Pennington Center.

DR. FLEGAL: I am Katherine Flegal, from

CDC's National Center for Health Statistics.

DR. YANOVSKI: Susan Yanovski, NIH.

DR. CARPENTER: Tom Carpenter, pediatric

endocrinology at Yale University.

DR. WIERMAN: I am Maggie Wierman,

endocrinologist at the University of Colorado.

DR. WOOLF: Paul Woolf, endocrinologist,

Crozer Chester Medical Center.

DR. WATTS: Nelson Watts, endocrinology,

University of Cincinnati.

DR. SCHADE: Dave Schade, endocrinology,

University of New Mexico.

DR. ARONNE: Louis Aronne, New York City,

Weill Cornell Medical Center.

DR. RYDER: Steve Ryder, Pfizer Research

and Development. I am the industry representative.

DR. BRAUNSTEIN: Thank you. I will now

turn the meeting over to LCDR Dornette

Spell-LeSane.

Conflict of Interest Statement

LCDR SPELL-LESANE: Good morning. The

following announcement addresses the issue of

conflict of interest with respect to this meeting

and is made a part of the record to preclude even

the appearance of such at this meeting.

Based on the agenda, it has been

determined that the topic of today's meeting is an

issue of broad applicability, and there are no

products being approved at this meeting. Unlike

issues before a committee in which a particular

product is discussed, issues of broader

applicability involve many industrial sponsors and

academic institutions.

All special government employees have been

screened for their financial interests as they may

apply to the general topic at hand. To determine

if any conflict of interest existed, the agency has

reviewed the agenda and all relevant financial

interests reported by the meeting participants.

The Food and Drug Administration has granted

general matters waivers to the special government

employees participating in this meeting who require

a waiver under Title 18, United States Code,

Section 208.

A copy of the waiver statements may be

obtained by submitting a written request to the

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

of the Parklawn Building.

Because general topics impact so many

entities, it is not practical to recite all

potential conflicts of interest as they apply to

each member, consultant and guest speaker.

FDA acknowledges that there may be

potential conflicts of interest but, because of the

general nature of the discussion before the

committee, these potential conflicts are mitigated.

With respect to FDA's invited industry

representative, we would like to disclose that Dr.

Steven Ryder is participating in this meeting as a

non-voting industry representative, acting on

behalf of regulated industry. Dr. Ryder is

employed by Pfizer Global Research and Development

as senior vice president and global

cardiovascular/metabolism/GI/GU development head.

And, although Pfizer conducts research in

therapeutic areas possibly covered by today's

discussion, Dr. Ryder's role on this committee is

to represent industry interests in general and not

any one particular company.

In the event that the discussions involve

any other products or firms not already on the

agenda for which FDA participants have a financial

interest, the participant's involvement and their

exclusion will be noted for the record.

With respect to all other participants, we

ask in the interest of fairness, that they address

any current or previous financial involvement with

any firm whose product they may wish to comment

upon. Thank you.

DR. BRAUNSTEIN: Thank you. Dr. David

Orloff with give the welcome and introductory

comments.

Welcome and Introductory Comments

DR. ORLOFF: Good morning. The first

thing I want to say actually before I get to the

introductory comments is that I believe, Dornette,

if I am not mistaken, we do not have any speakers

for the open public hearing. Is that correct?

LCDR SPELL-LESANE: That is correct.

DR. ORLOFF: As a result of that, time

permitting, we may try to push Dr. Atkinson's talk

up to the morning before we break for lunch. I

leave that up to Dr. Braunstein and to the clock.

I want to wish everyone a good morning and

welcome our advisors, our guest consultants, FDA

staff and interested public.

The purpose of today's meeting of the

Metabolic and Endocrine Advisory Committee is to

revisit the division's 1996 draft guidance on the

development of drugs for the treatment of obesity.

As everyone present knows, the FDA's public health

mission includes the charge to assure that safe and

effective drugs are efficiently, but without

cutting crucial corners, brought forward through

development to the marketplace in order to

diagnose, cure, treat, prevent or mitigate disease.

That, broadly speaking, explains our purpose here

today.

More specifically, in August of 2003 the

then Commissioner McClellan established the FDA

obesity working group and asked that group to

develop a plan of action to address critical

aspects of the burgeoning obesity problem in the

U.S. within the authorities of the Food and Drug

Administration.

Germane to our work here today, he charged

the so-called therapeutic subgroup, which was led

by our division, to assess real or perceived

barriers to obesity drug development and to make

recommendations on ways to encourage the

development of new or enhanced therapeutics for

obesity.

In the face of this growing public health

problem, advances in the understanding of the

physiology and pathophysiology of obesity and the

activity within the pharmaceutical industry in this

therapeutic area, we took the opportunity to plan a

discussion of the current guidance and its

potential modification. Today's meeting is further

timely in light of the recent release by NIH,

announced on August 24th, of the final version of

its own strategic plan for obesity research which

includes intensification of efforts on

pharmacologic approaches to the prevention and

treatment of obesity in both children and adults.

In early 2004 we published a formal call

for comments on the guidance in the Federal

April. Today, with the help of our advisors and

consultants, we will review what we consider to be

the salient issues raised in the comments we

received, as well as others that we believe are

critical to ensuring that FDA's evidentiary

standards for safety and efficacy of obesity drugs

are, to the extent possible, in line with the

science of the day.

We look forward to the formal

presentations and what we trust will be a fruitful

discussion to follow. I should make clear, as I

believe is apparent from the agenda, that this

meeting is not intended to discuss any specific

drug products, approved or in development.

Furthermore, in a manner distinct from a meeting of

that type, we will not ask the committee and guests

to vote per se on the questions we will pose.

These are intended to raise the issues that we wish

to hear discussed. We, the FDA staff, will listen

and contribute as we see fit and, of course,

respond to questions directed at us as we are able.

We intend to take the information we have gleaned

today back for consideration in drafting possible

revisions to our guidance for industry.

It is important for those participating

and listening today to understand that by this

meeting we make no formal commitment to changes in

the guidance. We view this as an information

gathering step in a process that may lead to

changes. I know you all have the agenda and I will

not review it. I have already announced the

potential changes.

Finally, before we start, I would like to

thank Dr. Lynne Levitsky, valued advisor

particularly on pediatric matters whose term has

recently expired, for service to us over the last

term. She is here today formally speaking as a

consultant. By her agreement to stay on in that

capacity, we hope to continue to engage her in the

future and look forward to her additional input

into the work of the division and the agency.

Finally, special recognition goes to Dr.

Glenn Braunstein who has kindly agreed to serve as

the chair of today's meeting. Dr. Braunstein's

second term as a member expired in June, and having

him here today is particularly fitting given that

he chaired the 1995 meeting that led to the

drafting of the '96 guidance under discussion.

Glenn's association with the committee and the

division dates to late 1991, including two stints

as an extremely effective chair. We thank him once

again for his invaluable service. Indeed, we are

not releasing him. He too has agreed to remain a

consultant. Glenn, thank you for your generosity

with your precious time and for your contributions

over many years to this committee, to the division

and to the work of the agency. With that, I will

turn it over to you.

DR. BRAUNSTEIN: Thank you, David. I

appreciate that. The first speaker this morning as

far as presentations are concerned is Dr. Eric

Colman, who is medical team leader, and he is going

to speak about the regulatory history of weight

loss drugs.

The Regulatory History of Weight-Loss Drugs

DR. COLMAN: What I plan to do for the

next 20 minutes or so is to give you an overview of

the FDA regulation of obesity drugs from roughly

the years 1938 through 1999. Before I get to that,

I did want to mention two milestones in the history

of drug regulatory.

These were, first, the signature in 1906

of the original Food and Drugs Act and that was

signed by President Teddy Roosevelt. Roughly

30-plus years later another Roosevelt, Franklin,

signed the Food, Drug and Cosmetic Act of 1938.

This Act has quite an influence on drug regulation.

It affected the labeling provisions, the

advertising provisions for drugs, and it was also

the first time that drug companies had to submit

evidence of a drug's safety before it was allowed

to go onto the market. It also marked the

beginning of the new drug application, or NDA,

process that we have all come to appreciate over

the years.

Getting started with the obesity drugs, in

1938 Myerson and colleagues reported the paper in

The New England Journal of Medicine, "Benzedrine

sulfate as an aid in the treatment of obesity."

These two colleagues treated roughly 17 obese

patients with 30 mg of amphetamine sulfate, which

is what Benzedrine is. They reported that the

patients lost anywhere from 9-54 lbs. This was

just one of a number of studies that started to

appear in the medical literature that suggested

that amphetamines may be an effective way to treat

obesity.

The following year, in 1939, the agency

approved Benzedrine for a host of different

indications, however, obesity was not one of them.

Several years later the agency approved another

amphetamine. This one was desoxyephedrine. Again,

there was a list of indications--narcolepsy, mild

depression, alcoholism, even hay fever at one

point, but again obesity was not in the list.

Now, it took four more years before the

agency finally felt comfortable and granted an

obesity indication for desoxyephedrine. To the

best of my knowledge, this was the first drug that

the agency approved for the treatment of obesity.

I have shown you here some of the language

from the labeling at that time to give you a flavor

of what people were thinking. For this drug the

labeling stated that the sympathomimetic amines

have been found of value, when administered under

the supervision of a physician, as an adjunct to

the dietary management of obesity. That was the

indication section.

The labeling also warned, however, against

its use in persons with cardiovascular disease,

hypertension or insomnia, and in those who were

neurotic or hyperexcitable. So, clearly, there was

an awareness that these drugs were stimulatory to

the central nervous system, to the cardiovascular

system.

On this last point regarding the

amphetamines I want to just highlight--this is just

to remind you that I will be talking a lot about

amphetamines and I will be talking about

amphetamine-like drugs in a moment. But when I

refer to the amphetamines I am including a large

number of compounds which include amphetamine

sulfate, desoxyephedrine, also referred to as

methamphetamine, dextroamphetamine and a number of

amphetamine-barbiturate combinations.

Soon after the amphetamines were approved

in the '40s and early '50s companies began to work

to try to develop compounds that, on the one hand,

maintained the anorectic properties of the

amphetamines but had less of the stimulatory

properties. They were successful to varying

degrees.

By 1960 the agency had approved five new

what I refer to as amphetamine-like drugs. These

are also referred to as the amphetamine cogeners.

These drugs were phenmetrazine, phendimetrazine,

phentermine, benzphetamine and diethylpropion.

Again to give you a sense of what people

were thinking during this time, I have shown you

some of the language from the labeling for

diethylpropion. This drug was indicated for any

obese patient, including the adolescent, the

geriatric and the gravid, as well as the special

high-risk situations of the cardiac, hypertensive

and diabetic patient. That is probably an

indication section that most drug companies would

die for at this point.

The labeling also stated that because

tolerance, habituation or addiction did not

develop, this drug was ideal for long-term use.

Again, it is interesting to look at the labeling

language from back in the late '50s.

Against the backdrop of the approval of

the amphetamines and amphetamine-like drugs, there

was a problem growing in this country and some

people were referring to it as an epidemic. That

was an epidemic of the abuse of amphetamines.

I have shown three figures here to give

you a sense of the amount of use of these

compounds. In 1958 there were approximately 3.5

billion tablets of amphetamines manufactured

legally in this country. Approximately a decade

later that had more than doubled to 8 billion

tablets. Expressed another way, in 1967 there were

approximately 23 million prescriptions for

amphetamines, 80 percent were for women and of all

the indications, these drugs were most commonly

prescribed for obesity.

The government tried to intervene to slow

or stop the spread of this abuse by passing two

laws, one was the Drug Abuse Control Amendments, in

1965. The second was the Controlled Substances Act

of 1970. This is when the scheduling of drugs was

introduced.

Moving from 1970 back to the early '60s,

in 1962 there was a very important addition made to

the '38 Food, Drug and Cosmetic Act. These were

the Kefauver-Harris Amendments, also known as the

Drug Efficacy Amendments. This legislation for the

first time mandated that new drug applications

contain substantial evidence of a dug's

effectiveness.

You recall, I mentioned that in '38 the

law said you had to have evidence of safety. This

law now said you had to have evidence of efficacy.

So the loop had now been closed. And, this

effectiveness was to come from adequate and

well-controlled investigations.

This raised a problem however. This

legislation took care of drugs approved in '62

forward but there were literally thousands of drugs

that were approved between '38 and '62. The

question came up what do we do about the efficacy

assessment of these drugs approved before 1962?

The answer came when the Commissioner called upon

the National Research Council of the National

Academy of Sciences to take this task on. This

endeavor became known as the Drug Efficacy Study

Implementation, or the DESI review process.

The formal portion of the Drug Efficacy

Study was conducted between 1966 and 1969. There

were a host of different drug panels, depending on

expertise, and it was the psychiatric drug panel

that was charged with reviewing the available data

on the efficacy of the amphetamines and

amphetamine-like drugs. They were told after they

completed their analyses of the available data that

they should classify the efficacy using one of

these five descriptors, starting at the top with

effective; effective but; probably effective;

possibly effective; or ineffective.

They completed their analyses in 1969 and

sent the results outcome the FDA Commissioner, and

this is what they concluded. They felt the

efficacy data supported a statement saying that the

amphetamines were possibly effective for the

treatment of obesity.

Regarding the amphetamine-like drugs, a

little bit better--they thought that this was

effective but... so, again, one step below

effective. The reasons they cited for not

classifying these compounds as effective were the

following: Many of the studies that they looked at

were of short duration. There was no evidence

available that the drugs altered the natural

history of obesity. There was some evidence that

the anorectic effects may have been strongly

influenced by the suggestibility of the patient.

And, there were concerns about the adequacy of the

controls in some of the clinical studies.

What were the regulatory consequences of

DESI review of the obesity drugs? In 1970 the FDA

concluded that the amphetamines were, indeed,

possibly effective in the treatment of obesity, and

basically mimicked what the DESI review panel

recommended. However, because this was short of

the category of effective, the FDA directed

industry to submit evidence of weight-loss efficacy

from adequate and well-controlled trials and,

ideally, of more than a few weeks duration. I

would point out here that at this time the FDA made

no comment about the efficacy of the

amphetamine-like drugs. That wouldn't come for a

few more years.

After the DESI review process was finished

in '69, in the early 1970s the Division of

Neuropharmacology Drug Products at the agency--that

was a division that had the regulatory purview of

these agents--clearly felt the need to develop a

policy whereby they could develop and regulate

obesity drugs. So, flowing from the DESI review

process, three important actions occurred in the

early '70s. These were the Prout Consultant Group,

the Neuropharmacology Drugs Advisory Committee, and

the conduct of the Amphetamine-Anorectic Drug

Project. Let me go through each one of those

briefly.

The Prout Consultant Group was put

together by folks in the Neuropharmacology Drug

Division at the FDA. It consisted of eight

external consultants and was headed by a physician

named Thaddeus Prout who was an endocrinologist

from Johns Hopkins. This group of eight

individuals met in April of 1071 to discuss

specific issues related to obesity drugs and

regulation and development of these compounds.

They issued these four statements back to

the Neuropharmacology Division: They felt that, in

fact, weight-loss drugs did have some potential

value. They felt that the efficacy trials for

these drugs should be at least 12 weeks in

duration; that the long-term follow-up of patients

was not the responsibility of drug companies; and

that the efficacy of the weight-loss drugs should

be defined as statistical superiority of drug to

placebo. This is an interesting point. This group

was specifically asked to define clinically

significant weight loss. Either they could not do

it or they did not want to do it but, in any event,

they said you should consider this as efficacy. In

other words, if the weight loss on a drug is more

than the weight on the placebo and the difference

is statistically significant, then you have a drug

that works .

About five months after the Prout Group

met and made their recommendations, the

Neuropharmacology Drug Division convened its own

advisory committee, in September of '71, and again

they wanted to get input about how to develop and

regulate the obesity drugs. They were also asked

to provide a definition of clinically significant

weight loss. They did not venture an answer.

Instead, they referred back to Prout's

recommendation that efficacy be defined as

statistical superiority of drug to placebo. This

was another group that could not define clinically

significant weight loss.

So, after two groups deliberated on this

the agency still had no working definition of

clinically significant weight loss. The

Amphetamine-Anorectic Drug Project somewhat

approached this problem in a backward direction.

This was a meta-analysis conducted by members of

the Neuropharmacology Drug Division, along with

agency statisticians. The overall goal was to try

to, once and for all, quantitate the efficacy of

the amphetamine and the amphetamine-like drugs. At

this point there were data available for

fenfluramine and sanorex.

This meta-analysis was quite large. It

included 200 clinical studies. These studies

ranged in duration from one month to six months. I

would say that the average study was six to 8 weeks

in duration. There were about 10,000 patients

involved in the whole analysis. At the end of the

day, when they got done analyzing these data, they

issued two conclusions. The first one doesn't

sound very impressive but this is what they said:

Patients treated with active medication did, in

fact, lose some fraction of a pound a week more

than those on placebo. The second conclusion was

that the data did not suggest that one drug was

superior to another, nor that the amphetamines as a

class were more effective than the amphetamine-like

drugs. This would have major implications, as we

will see in a few minutes.

What were the consequences of this

meta-analysis? In 1973 the agency officially

declared that the amphetamines and the

amphetamine-like drugs were effective for the

treatment of obesity. You will recall that in 1970

they said amphetamines were possibly effective and

they didn't say anything about the amphetamine-like

drugs. So, from doing this meta-analysis, they

felt comfortable in declaring that these two sets

of compounds were both effective for the treatment

of obesity.

The second thing that came out of this

project was class labeling. I mentioned the abuse

problem, the speed epidemic that had continued

through the '60s and into the '70s. So, the abuse

of the amphetamines was still very much on the

minds of the senior leadership at the FDA. So,

people started to reason, well, if you limit the

use of these drugs just for a few weeks you can't

get abuse. So, if we limit their use to only a few

weeks we take care of the abuse problem and that

way we tidy up the risk/benefit profile for these

drugs. So, they made a blanket case and not only

were the amphetamines indicated for short term and

a few weeks actually shows up in the label. People

have often referred to this as a few months but the

label actually says a few weeks.

Instead of just limiting it to the

amphetamines, they threw it over to the

amphetamine-like-like drugs as well so at this

point all these drugs became indicated only for

short-term use, a few weeks use, and I would submit

that was largely driven by concerns about abuse,

street abuse.

The next notable event in this history

came in 1979 when the agency announced its plans to

remove the obesity indication from the

amphetamines. They still hadn't had enough; they

wanted more. They felt that they had good reason

to propose this removal. One of the things that

backed them up, they believed, is that there was

continued evidence of abuse of amphetamines. They

knew it was largely coming from this database

referred to as DAWN, which stands for the Drug

Abuse Warning Network.

The other point has to do with, as I just

mentioned, the risk/benefit profile of the

amphetamines relative to the amphetamine-like

drugs. The FDA had clearly said that they don't

think the efficacy is any different for the

amphetamines than the amphetamine-like drugs but we

do believe that the abuse potential was more of a

problem for the amphetamines than the

amphetamine-like drugs. Therefore, amphetamines

have a less favorable risk/benefit profile versus

the amphetamine-like drugs. If you took the

obesity indication away from the amphetamines

people in this country would not suffer at all;

they had have the amphetamine-like drugs that

worked just as well.

The industry had a chance to respond to

this proposal and they did so. I have listed four

of their rebuttals here. For one thing, industry

felt that the FDA analyses of the DAWN data were

incorrect. They just didn't believe that there was

evidence of continued abuse.

Secondly, they argued that if illicit

production and use of the amphetamines was a real

problem, that was the purview of the state medical

boards and the Department of Justice; it wasn't

something the FDA should get involved in.

Thirdly, they said, wait a minute, abuse

requires use beyond a few weeks and our drugs are

only approved for a few weeks. So, if this is a

problem we are talking about off-label drug use

and, once again, that is not something the FDA gets

involved in.

Finally, the risk/benefit issue--they felt

that the risk/benefit equation should be made on

its own merits, in other words, relative to

placebo. In this case, the agency was saying that

the risk/benefit profile of the amphetamines was

less favorable than the risk/benefit profile for

the amphetamine-like drugs. According to industry,

they didn't think that was a legitimate or legal or

regulatorily tenable reason to take away the

indication.

I have to say that after all this

bickering the industry won out because this planned

action never took place. The agency never removed

the obesity indication from the amphetamines and,

to this day, I know of one amphetamine that still

has in its label its use for short-term treatment

of obesity.

We now enter the 1980s, and I think the

1980s in terms of obesity drug development really

should focus on one particular happening, and that

was the start of the phen-fen studies. In the

early 1980s, a clinical pharmacologist from the

University of Rochester reasoned that the stimulant

effects of phentermine would counter the sedative

effects of fenfluramine such that the two together

would provide a very tolerable combination that

could be used over long-term use. So, he and his

colleagues started these studies in the '80s.

In 1992 they published a number of papers

citing the main results of these trials. They

concluded that yes, indeed, the combination was

tolerable and that people could take these drugs

over the course of years, and that it was safe and

effective.

Again, these were published in 1992. They

had a major impact on subsequent use of these

drugs, as I have shown here, in this table. These

are the estimated total number of prescriptions for

phentermine in 1992, 2 million. For fenfluarmine

there were about 70,000 prescriptions in

1992--again, the year the papers were published.

Four years later these numbers had gone from 2

million to 11 million and from 69,000 to 7 million.

I am not saying all of this was due to these papers

but a large part of it was.

There was another event that happened

around 1992, and that was the transfer of the

regulatory responsibility of the obesity drugs from

neuropharmacology to the Division of Metabolic and

Endocrine Drugs, where they are now. When the new

drugs arrived in the new division there was fairly

strong feeling that effective drug treatment

required long-term or indefinite treatment.

Therefore, why don't we have long-term pre-approval

trials? There were other thoughts within the

division. There was a strong sense that we need to

get this formulated into a guidance policy. They

convened their advisory committee in a two-day

meeting in 1995 to discuss how to develop and

regulate obesity drugs, with an eye to issuing an

obesity guidance document.

They had a successful meeting. The

obesity draft guidance was issue in 1996. I just

show you two of the more important components of

that guidance document, and these will be issues

that we will be discussing later today.

In terms of efficacy, a 5 percent

benchmark was chosen. At that time, people could

point to the fact that if people lose as little as

5 percent of weight they could get improvements in

lipids, blood pressure and cholesterol and,

therefore, this was a clinically significant weight

loss. So, now we finally have a definition for

clinically significant weight loss.

On the other side, in terms of the size

and duration of phase 3 trials, I think most people

felt comfortable that we had agreed that one year

of a placebo-controlled trial would be an adequate

exposure to assess efficacy and some degree of

safety. A lot of people felt though that of these

1500 patients who made it out to a year, 200-500

should be rolled over into an open-label exposure

for a following year, again, to get another sense

of safety. We will be talking about these issues

as well later today.

Just briefly, long-term treatment of

obesity, from FDA's perspective, came about when

dexfenfluramine was approved in 1996. We all know

it was removed from the market the following year

because of valvulopathy. A couple of months after

the removal, sibutramine, or Meridia, was approved.

I have shown you here the actual labeling for the

indication. Meridia is indicated for weight loss

and weight maintenance. Xenical, the most recently

approved drug, in 1999, has the same indications,

weight loss and weight maintenance, but it also has

an additional indication and that is to reduce the

risk for weight regain after prior weight loss.

These are issues that we hope committee members

will engage in a dialogue later this afternoon in

terms of what these terms mean; how they should be

defined, etc.

So, if I could provide you with a global

summary, I think it is safe to say that defining or

quantitating the efficacy of weight-loss drugs has

been problematic. It certainly has been a

challenge from a regulatory perspective. It wasn't

until the mid-1990s that we had a workable

definition of clinically significant weight loss,

and that is the 5 percent benchmark. We still

don't have a definition of clinically significant

drug-induced weight loss--that is a different

issue.

On the other side of the coin, I also

think it is safe to say that the regulatory history

of the obesity drugs has seen its share of highly

publicized safety problems. Beginning with the

abuse of the amphetamines in the '40s, '50s, '60s

and beyond, primary hypertension became an issue

with a drug called aminorex that was used in Europe

in the '60s. It was never in this country.

But this condition was subsequently linked

to fenfluramine and it was a major issue at the

time that dexfenfluramine was approved. It was

well-known that this drug increased the risk of BPH

in people who took dexfenfluramine. That was

before dexfenfluramine was approved. These

concerns were only later overshadowed by the

cardiac valvulopathy that showed up a year after

their approval. These were all very, very highly

publicized events, basically so many in the

population were exposed to these drugs.

Finally, the approval of Meridia or

sibutramine, back in '97, was accompanied by very

strong warnings, precautions and concerns regarding

the effect of that drug on blood pressure and

pulse.

Let me close. Since the topic of today's

discussion is the obesity guidance document, I

thought I would just provide a visual reminder of

the goals of not only this guidance document but I

think of all guidance documents, and that is

obviously, on the one hand, to facilitate

industry's development of safe and effective drugs

but, just as importantly, to provide regulators

with the best available evidence upon which to

judge a new drug's risk/benefit profile before the

drug is approved. Obviously, those two things

require a certain amount of compromise and juggling

but I will leave you today with that thought. Keep

that in the back of your mind as we deliberate the

various proposals to change the guidance document.

Thank you.

DR. BRAUNSTEIN: Thank you, Dr. Colman.

Are there any questions from the panel for Dr.

Colman?

[No response]

Thank you. We will move on then to Dr.

Katherine Flegal's discussion of the epidemiology

of overweight and obesity.

The Epidemiology of Overweight and Obesity

DR. FLEGAL: This is the outline. I am

going to give a very brief overview of trends in

obesity and overweight in the United State,; a

history of regulation of weight-loss drugs, a brief

history of definitions of overweight; some

population estimates; prevalence of overweight

categories and comorbidities; and kind of a brief

discussion of some of the aspects of possible

benefits and risks of weight change in mildly

overweight people with comorbid conditions.

Most of the data I am going to present

today come from the series of National Health and

Nutrition examination surveys in the U.S., NHANES.

Many of you are familiar with this but I know some

of you aren't. These are a series of

cross-sectional national representative surveys,

conducted by CDC's National Center for Health

Statistics, in which weight and height are measured

and many other actual measurements are taken. We

have a series of these dating back to the 1960s up

until today. So, we have a little over 40 years of

data on the U.S. population from these surveys.

The most recent one began in 1999 and is

continuous, representing some data from 1999 up to

2002 in that survey.

This slide shows the age-adjusted trends

in obesity, defined as a body mass index of 30 or

above in the United States. Starting back in 1960,

the prevalence was only about 10 percent for men

and today it has gone up to almost 30 percent. As

you see, the prevalence was really fairly constant

from 1960. In '71 to '74 and '76 to '80 there were

not large changes for either men or women. In the

'89 to '94 survey the prevalence went up sharply

and somewhat unexpectedly, and in the most recent

survey it has gone up again so we see this

continuing trend.

This is the same setup. This is for

overweight defined as a body mass index of 25 or

above so it includes the obesity data I just showed

you. Again, the prevalence was relatively stable

over the first three surveys and then increased.

One thing to note is that the prevalence of

overweight with these definitions has been pretty

high since 1960. Almost 50 percent of men and 40

percent of women were overweight in 1960 according

to this definition.

As you have just seen, the definitions of

overweight and obesity that I am using are based on

body mass index which is calculated as weight in

kilograms divided by height in meters squared.

There are two definitions of overweight in this

system. One is a body mass index of 25 up to 29.9

or a body mass index of 25 or above. Obesity is

then defined as a body mass index of 30 or above

and a healthy weight as a BMI of 18.5 but less than

25.

These definitions have been a long time

getting systematized and standardized. This is a

very brief overview, but basically definitions of

overweight up to the early '80s really were not

systematized and there were very wide international

variations. In the United States there was a lot

of use of weight-height tables like the insurance

company tables that you have probably seen. There

is a whole set of issues of skinfolds measurements,

different kinds of prediction equations; a lot of

different kinds of weight-height indices. There is

the Broca index, ponderal index. You can see these

used in different literature and they are used in

different metric systems as well so you never knew

whether it would be kilograms and meters or

centimeters or pounds and inches. So, if you look

at the literature back in the '70s, say, and before

it is very difficult to make any comparisons.

There are a lot of different definitions that were

being used and there were a lot of differences

between countries as well.

I think during the 1980s epidemiologic

consensus began to form around body mass index,

which is also called Quetelet index after the great

Belgian statistician in the 19th century. So, you

can see that this index has been around for a long

time and has been used somewhat, but it began to be

really more the index of choice. An NIH consensus

conference in 1985 recommended the use of body mass

index. But at that point the cut-off values still

were somewhat varied.

The 1959 Metropolitan Life tables in the

U.S. had a range of desirable weights for a given

height. There was a practice that had grown up of

taking the midpoint of that range as kind of the

ideal weight and then saying if you are at or about

120 percent of that midpoint, then that was the

beginning of the definition of overweight of the

median frame weight range.

At the NIH consensus conference, in '85,

there was data presented from NHANES, as I have

already shown you, about the 85th percentile values

for men and women age 20 to 29. Those were a value

of 27.8 for men, 27.3 for women. The consensus

conference decided to adopt those as some kind of

definition of overweight because they actually

correspond pretty closely to the Met Life, to the

120 percent definition based on Met Life. We, in

fact, used these values as recently as 10 years

ago. We would have been publishing data using

those particular cut-off points.

Meanwhile, BMI cut-points of 25 and 30

began to be recommended by expert committees.

These were not suggested originally by these expert

committees. I think the earliest suggestions I am

aware of were by George Bray and George Garrow,

back around 1980 probably or perhaps before. But

these were thought to be more systematized. There

was a 1995 report from an expert committee of the

World Health Organization that suggested these

cut-off points. That was followed in 1998 by the

Clinical Guidelines on the Identification,

Evaluation and Treatment of Overweight and Obesity

in Adults that NHLBI put out, which is really more

or less the basis for our current use of these

values.

Why these values? Here is what it says in

the 1995 expert committee report that they proposed

a classification with cut-off points of 25, 30 and

40. This is based principally on the association

between BMI and mortality.

They go on to say the method used to

establish these kind of points has been largely

arbitrary. In essence, it has been based on visual

inspection of the relationship between BMI and

mortality: the cut-off of 30 is based on the point

of flexion of the curve.

So, in this report and in others there is

not a careful study of the criteria for using

exactly 25 or 30 as opposed to, say, using 30.5 or

27.8. These are kind of general and, as I say,

largely arbitrary. Here is kind of a typical

relation between mortality and BMI curve that would

have been available to that committee. This is

from the American Cancer Society studies.

You see a couple of things here. First of

all, the point of lowest mortality tends to hover

around a BMI of 25. You see this curvolinear,

somewhat U-shaped relationship with much higher

risk out here. Also, body mass index is not a

physiologic measure; it is just an index and you

can kind of intuit that the choice of cut points of

20, 25, 30, 35 and 40 are because these are round

numbers and they vary by 5. These are not really

physiologically based cut points. So, these are

approximations. They are very useful

approximations, by the way. We are very glad to

have internationally standardized definitions that

we can all use. Now you can compare one person's

data with another person's data so these are quite

valuable to have.

The 1998 NHLBI clinical guidelines also

offer the same definitions. Here overweight is 25

to 29.9 and they say the rationale was based on

epidemiological data that show increases in

mortality with BMIs above 25. This increase tends

to be modest until a BMI of 30 is reached. So, you

see that this language also is somewhat imprecise.

I think this is on the following page.

They describe quite a few studies. Very often the

point of minimum mortality is around a BMI of 25.

This is a study of NHANES I where they show the

lowest mortality in the range of 25-30, and they

found, by race and sex, the lowest mortality at

24.5 for white men, 26.5 for white women, and even

higher values for black men and women. There is

other information presented in the same NHLBI

report which also has somewhat similar analyses.

So, definitions of overweight have changed

quite a bit over time. We have pretty much settled

down now to using these standard definitions but

that is a little bit of the history.

Getting back to definitions, overweight is

a BMI of 25-29.9. These are slides like the ones I

showed you but now these are really just that

range. There are two things you can see from this.

One is that the prevalence of overweight by these

definitions has really changed very little over

time. It is almost constant. Another thing you

can see is that the prevalence of overweight by

these definitions is quite a bit higher in men than

it is in women, which is less true of the

prevalence of obesity. It is about 38-40 percent

for men and about 25 percent for women.

Just looking at the numbers of people, and

I am going to try to divide this by separate

categories. One is BMI to under 27 and 27 up to 30

because that is one of the cut points used in the

current guidance document. This is just to show

you the number of people in the U.S. population who

fall into these various categories. I also

included the next lowest category as kind of a

comparison point. In this lowest category of 23 to

less than 25 the numbers of men and women are

approximately equal, about 14 million in each. In

the range of 25 to under 27 there are a few more

men than women. There are about 16 million men and

12 million women. As you go up to the range of 27

to 30 there are more people in this category, which

is actually a broader category, of course, also.

There are 21 million men and about 17 million

women.

Looking at that by age as well, I have

divided this into 4 age groups, 29-29, 40-59, 60-79

and 80 and above. You can see that for a BMI of 25

to 27, men and women both in that BMI range are in

the age groups 20 up to 59. When you get to the

60-79 year-old age range there are fewer people but

you see that in the younger ranges there are more

men than women in these categories. When you get

up to this age range there are actually almost

equal numbers of men and women in the older ages.

The same is true for the next category of a BMI of

27-30. So, there is a definite age pattern with

these numbers.

Now I am going to talk about

comorbidities. There are five listed plus "other"

in the guidance document: hypertension,

hyperlipidemia, glucose intolerance, cardiovascular

disease, sleep apnea and other obesity-related

conditions. I don't really have good data to show

you on cardiovascular disease or on sleep apnea or

the other conditions so I am just going to talk

about these three from what we have, hypertension,

high cholesterol and glucose intolerance.

One thing I was asked to do is to consider

the question of the point of inflection of the

curve of the relationship of these comorbidities to

BMI. So, I have presented the data this way and I

have a whole series of slides, all laid out the

same way.

The yellow line is men--this is for men,

20-39; the green line, 40-59; the pink line, 60-79;

and then 80 and above. This shows the body mass

index categories along this axis and the

prevalence. There are a couple of things you can

see from this. First of all, you don't see a very

clear point of inflection, for example, between

23-25, 25-27. Here you see a little increase but

in this case it was a decrease here and an increase

there so these bounce around somewhat. So, you see

a gradual increase in the prevalence of these

conditions with the BMI level in all age groups.

You don't visually see an obvious point of

inflection.

The other thing to notice is that although

we talk about these as obesity-related

comorbidities, this shows you pretty clearly that

they are also age-related comorbidities and, in

fact, the prevalence of any of these conditions in

people with a BMI of 30 who are young is far, far

lower than the prevalence even in people at the

lowest BMI level who are older. So, you need to

keep that in mind. Again, there are other risk

factors for these conditions and age, in

particular, is a very strong risk factor.

This is the same picture now for women.

Again, you see at the old age range a very high

prevalence of hypertension at all BMI levels. You

see in most age groups a slight increase in the

prevalence by BMI level and you don't see a strong

inflection point. By the way, I defined

hypertension as a measure of blood pressure

systolic over 140 or diastolic over 90, or using

medications for hypertension.

This is for high cholesterol, which I

defined for this purpose as total cholesterol of

above 240 mg/dl or using medication. Here you see

a somewhat similar picture. The prevalence is not

as high even in the oldest age group and our data

are somewhat sparse in the older age group. It may

be one of the reasons this curve is not estimated

that well. Again, you see some tendency for

increase in cholesterol with BMI, also a tendency

to increase with age--not a terribly clear

inflection point.

Here is the same information for women.

Again, sort of the same comments would apply.

Finally diabetes--this is just based on

diagnosed diabetes and this is self-report of

diagnosed diabetes so this is not based on

measurements of glucose tolerance or looking at

undiagnosed diabetes. This is people who say that

they have been told that they have diabetes. I

also excluded people who had age at onset below 30

and have used insulin since diagnosis,

approximately since diagnosis, to try to limit this

proximally to type 2 diabetes. Again you see the

increase with BMI. You see the age differential

and you, again, don't really see a strong

inflection point. The same thing for women.

This is just the prevalence of any

comorbidity. I should say any selected comorbidity

because I am only looking at three. This is by age

and body mass index group for men. This has

somewhat smoothed out the lines because there are

more comorbidities involved. Again, there is this

big age differential--you know, fairly smooth

curves; they go up and down some but there is no

obvious inflection point between 25-27 and 27-30 or

23-25.

This is the same diagram for women.

Again, big age differences; increasing prevalence

of comorbidity with BMI group; fairly smooth lines.

So, how many millions of people are we

talking about? I will show you some other data but

this is when people have 2 or more comorbidities by

BMI categories. For comparison purposes, I put a

lot of BMI levels in here. In this range, which is

the range of interest for this purpose I think,

there are roughly speaking about 4 million people

in the U.S. who have a BMI at that level and have 2

or more comorbidities. That is in contrast to

about 6 million in the 27-30 range who have 2 or

more comorbidities. So, there is a ratio here.

This is about two-thirds of that. I have left out

some comorbidities so presumably these numbers

could be higher so this is just selected

comorbidities.

For comparison, even at the next lower

level there are almost 3 million people who would

fall into that category, even the lowest BMI

category. So, these comorbidities, again, are not

limited only to people in these overweight and

obese ranges. At the BMI level of 30-35 the

numbers are really much higher. Also, the numbers

of men and women are pretty equal in these

categories of interest in the overweight range.

There is a difference by age again. This

shows the same slide but now it is just limited to

people in the age range of 20-59. Here there is

about one and a half million people who fall into

this category, which is BMI 25-27 and one or more

comorbidities, and now the numbers of men and women

are no longer equal. There are about twice as many

men as women in this younger category.

This is for ages 60 and above. Remember,

the total here is a little under 4 million so

almost 2.5 million of those people are in the age

range of 60-70 and now we see that there are, not

unexpectedly in this case, more women than men in

this age range in this BMI category with

comorbidities. That is true along the whole

spectrum of BMI levels.

This is sort of changing the design here

but this shows you the number of million of people

with 1 or more, 2 or more or 3 comorbidities. This

is for BMI 25-27 so this is 1 or more, 2 or more or

3, and this shows the total with the different

components of the bar showing the age ranges. So,

what you can see is that, for example, is people

with one or more comorbidity about equal numbers of

people in the 40-59 and 60-70 age ranges and those

make up the majority of people with a smaller

contribution from people 20-39, even though many

people in the population in this 20-39 age range

don't fall into the comorbidity range.

So, we see about 12 million total with one

or more comorbidities as compared to 18 million in

the higher BMI range. When we get down to 2 or

more comorbidities, which is the number I just

showed you, this is approximately 4 million. The

largest group is going to be people in the 60-79

age range and people above 60 make up the majority

of this group, although not everybody in this

group. That is true also for BMI 27-29. So, the

age structure of these age groups is not the same

as the age structure of the population.

What about weight loss for people with BMI

25-27? I have tried to review the literature. I

have probably not reviewed everything, by a long

shot. As far as I can discern, there is not very

much information about the benefits of weight loss

in this particular BMI range. Most studies of

weight loss don't include that many people in this

level. Again, up to 10 years ago we would not have

considered people in this range to be overweight so

that might be one of the reasons why they were not

really going to be included. Some of them may

actually explicitly exclude people when they study

a BMI of 27 or a BMI or 28.

That is also true of studies of the

benefits of weight loss in the control of

conditions such as hypertension or hyperlipidemia.

They may explicitly exclude people who have BMIs as

low, so to speak, as 25-27 or may include few, if

any, participants.

In kind of a mirror image, I also read an

article which was complaining that drug trials for

hypertension are conducted in people who are

overweight but not obese so we know very little

about it. So, basically, trials of weight loss and

hypertension are conducted in obese people and in

trials of drug use and hypertension are studied in

overweight people but not the converse. So, we may

have missing information on both sides of that.

The NHLBI clinical guidelines

recommendations for a BMI of 25-29 overweight

recommend treatment only when patients have 2 or

more risk factors or a high waist circumference.

Other than that, weight maintenance is actually

recommended. So, the guidelines here for

overweight treatment do not recommend treatment for

everybody but just for people with other risk

factors. They also mention--I didn't put this on

the slide--that treatment of the other risk factors

is also just as important and should also be

considered.

You will see this statement on another

slide, but there are a lot of studies that show

that short-term weight loss has beneficial effects

on risk factors such as high blood pressure and

cholesterol. That is really very well established.

Most studies suggest that these are monotonic

relations but there is no obvious threshold. So,

you would infer from this that weight loss is very

likely to improve blood pressure and other risk

factors, certainly in the range of BMI of 25-27 as

well and perhaps at any weight level. We don't

really know but there is not that much evidence on

the specific BMI range. This is a fairly

reasonable inference.

How much benefit would that have? What

would be the net result? That is very hard to

judge in the literature. This is one very

approximate way of looking at it. You have already

seen these data but in a different format. What is

the prevalence of having 2 or more comorbidities by

age group for BMI 23-25 versus 25-27? If you think

that weight loss in the BMI group 25-27 puts you

into this next lower group, which is a very

plausible assumption, roughly speaking what would

the expected prevalence be?

You can see that effect--this is an

approximation again--by just comparing these 2 bars

which show the prevalence in the 23-25 BMI group

versus the prevalence in the 25-27 for different

age groups. In the youngest age group in which BMI

is probably a stronger risk factor, relative risk

for hypertension associated with BMI stronger are

stronger in the youngest group, you see a pretty

big potential difference of about half the number

of people in this lower BMI group. The number with

2 or more comorbidities is about half. So, that

would suggest that you get a fairly noticeable

prevalence effect by this kind of change in weight.

At the older age ranges the prevalence is high.

So, just looking at these data you would

suspect that if you had people with a BMI of 25-27

and they reduced their weight to 23-25 it is not

likely that they are going to end up down here

where the 20-39 year-olds are. They are more

likely to be approximately where people in their

same age group are. So, the prevalence of having 2

or more comorbidities is likely to be high even

after weight reduction. So, while there is likely

to be a beneficial effect, the net effect on

prevalence may not be that great.

Weight loss is just kind of part of the

therapeutic armamentarium for treatment of various

conditions. There is a whole non-pharmacologic

treatment or therapeutic lifestyle changes which

include weight loss, physical activity and

healthful eating habits, which may mean more fruits

and vegetables, less sodium, less saturated fat, a

whole different range of possible changes. These

are an important part of the treatment of diabetes

and cardiovascular risk factors obviously. Drug

treatment is also often used in managing these

conditions.

So, you might ask what is the relative

contribution of weight loss in this panoply of

treatments. As far as I can find out, that is not

well established. For example, what would be the

probability that non-pharmacologic treatment alone

versus drug treatment would have on management of

hypertension? There are review articles and

summary data on this but they tend to start at a

higher BMI level, at BMI of 27 or above or 28 or

above. So, it is somewhat difficult to assess.

Also, for example, there is one paper by

Ed Gregg using the national health interview survey

data that suggests that the intention to lose

weight is associated with improved mortality

regardless of actual weight loss, and the intention

to lose weight may be accompanied by some of these

other changes, such as increased physical activity

and changes in eating habits. So, it is hard to

judge and usually weight loss by itself is not the

only part of it. Therapeutic lifestyle changes

include more, and clinical trials will also look at

lifestyle changes. So, they include more than

weight change and try to assess where weight change

itself falls in the pictures. I couldn't find any

data that really spoke very clearly to this issue.

There are a couple of concerns. This is

from the Look Ahead Action for Health and Diabetes

study, I guess. This is from their website. This

is the sentence I already had on the other slide.

Although we know that weight loss improves risk

factors and clearly improves blood pressure and

glucose tolerance, there are these observational

studies that suggest some association of weight

loss with increased rather than decreased

mortality.

These studies do not differentiate

intentional from unintentional weight loss so they

definitely have limitations but they can't be

completely ignored either. Because of this, there

is actually a randomized clinical trial of

intentional weight loss going on. There are some

questions we don't really have the answers to about

this possibility of increased mortality with weight

loss so that is one concern, looking at weight loss

in this BMI range.

Another possible concern is, again, that a

lot of the people who are in this BMI range who

have comorbid conditions are elderly and more of

the elderly, not surprisingly, are women rather

than men and there are, you know, some possible

adverse effects of weight loss in this age range in

the elderly and particularly perhaps for women.

One of these is the possibility that weight loss as

adverse effects on bone health and can result in

lower bone density or greater risk of hip fracture.

This is a report from the study of

osteoporotic fractures where these women were close

to this range and the median and they had an

increased risk of hip fracture with weight loss.

In fact, this study found also an increase in thin

women as well, although the increase was not as

great. They did look at intentionality versus

unintentionality or lack of intention to lose

weight. In this study, and this is not the only

study on this topic but just something to kind of

keep in mind as a possible issue, regardless of

current weight or intention to lose weight there

was an association of weight loss with hip bone

loss and risk of hip fracture. So, they concluded

that even voluntary weight loss in overweight women

increases hip fracture risk.

Just to summarize, definitions of

overweight have varied a lot over time, and

epidemiologically useful consensus definitions do

not necessarily represent physiological

differences.

The prevalence of selected comorbidities

rises with BMI and doesn't have, at least in my

analysis, clear inflection points. There are about

12 million adults with a BMI 25-27 with at least

one selected comorbidity and about 4 million have

at least 2 selected comorbidities. So, it is a

large group of the population.

Half or more of the adults with BMI 25-27

and selected comorbidities are age 60 and above.

Weight loss, lifestyle changes and drugs are all

used to manage these and other comorbidities. So,

weight loss is part of a whole package of possible

treatment modalities.

Weight loss is associated with some

possible adverse consequences in observational

studies. So, I would conclude that the benefits

and risks of weight loss for people with BMI 25 to

under 27 have not been clearly established. Thank

you.

DR. BRAUNSTEIN: Thank you. Are there any

questions from the panel members? Dr. Follmann?

DR. FOLLMANN: I just had a comment. I

hadn't seen the relationship between BMI and

overall mortality before and I was really struck by

the nadir at 25. Many of these documents we have

been reading before this meeting were talking about

a cut point of 25-30 for definition of overweight,

and it just strikes me as maybe curious as to why

you would recommend or why people would consider

having someone who has to be above 25.1, which is

close to optimal, lose weight. So, I was wondering

if you could comment on that.

DR. FLEGAL: Well, I guess I think of this

from an epidemiological perspective. We have

prevalence estimates that use 25 and, you know,

different studies show the nadir at different

points so I don't think you can say that it is

exactly at 25. But the recommendations of NHLBI

are really not to lose weight at a BMI of 25.1

unless you have comorbid conditions. So, avoidance

of weight gain is probably more important in that

range.

DR. BRAUNSTEIN: Yes?

DR. RYDER: Yes, I just have one quick

question. On the two graphs that I you showed

earlier on the age-adjusted trends in obesity, you

used two categorical definitions, one of 25 and one

of 30 with somewhat different patterns.

I have a two-part question. One is if you

use 27 instead of 25 or 30, because I have seen

that put forward, would the display be more like 30

or more like 25?

DR. FLEGAL: I think it would be more like

30 but I haven't actually looked at data.

DR. RYDER: And the second part is the

average weight in the United States over this time

period I believe has been going off in somewhat of

a linear way, or maybe even more than a linear way.

Has the distribution pattern, Poissant

distribution, been maintained or is it just one arm

skewing out?

DR. FLEGAL: That I can't answer.

Basically, the whole distribution of body mass

index is shifting to the right a little bit. But

the distribution is becoming much more skewed so

there are much larger changes at the higher tail of

the distribution. The median has shifted somewhat

but the 90th percentile has shifted a lot. So, the

distribution is both shifting to the right and

becoming much more skewed.

DR. RYDER: Thank you.

DR. CARPENTER: I was struck by the large

impact of age on the comorbidities and, at the same

time, struck by the fact that in your later slides

you demonstrate that the effect on comorbidities

with weight loss is much greater at the young ages.

I wonder if anybody has looked at the duration of

carrying a certain BMI as being more important than

the current BMI as a risk factor for comorbidities.

DR. FLEGAL: There are studies like that.

I don't think they would explain those age

differences. I think basically a lot of people,

even at the lowest BMI in the age range of

60-79--you know, a lot of people have hypertension

regardless of BMI. So, any duration or changes

can't affect that. You know, at every BMI level

you have like 70-80 percent of people with

hypertension so, although duration may very well

have an impact, I don't think that can be the

explanation for those prevalence figures.

DR. BRAUNSTEIN: Dr. Follmann?

DR. FOLLMANN: Have there been studies

done that look at the pattern of weight gain over,

say, a 10-year period and how that might affect

mortality? I am thinking of someone who, say,

weighs 200 lbs at 40 and goes up to 250 lbs in a

steady linear fashion as one kind of trajectory,

and the other where they repeatedly diet and their

weight fluctuates a lot over that 10-year period

but they end up at the same weight. So, steady

versus erratic weight velocities--have there been

studies looking at the risk associated with those

two possible trajectories?

DR. FLEGAL: Well, there have been studies

of weight cycling. Sue Yanovski probably knows

more about that than I do. But I believe that a

kind of consensus is that weight cycling probably

doesn't have a large impact on mortality. Is that

right, Sue?

DR. S. YANOVSKI: Yes, the difficulty with

these kinds of studies is that they are all

observational studies, and weight cycling in itself

is associated with a lot of psychiatric morbidity,

a lot of other comorbidities and it is really

difficult to tease out cause and effect in those

kinds of studies.

DR. FLEGAL: Again, there are

observational studies that suggest that weight loss

is associated with increased mortality. There are

a lot of questions about intentionality; why do

people change their weight. As Sue was saying,

there are other issues. So, this whole area is a

very tangled and confused area to really sort out.

DR. BRAUNSTEIN: Dr. Hirsch?

DR. HIRSCH: I think you have just about

answered what I was going to ask. The 1997

recommendation concerning the issue that a

randomized clinical trial of intentional weight

loss is the only way we could prove whether there

are dangers inherent in weight loss--no such trial

that fits any of those issues has been carried out.

Is that true?

DR. FLEGAL: Of intentional weight loss--

DR. HIRSCH: Yes, randomized, prospective

trial. You are saying that is the only way you

could find out what the inherent harms of weight

loss might be.

DR. FLEGAL: Look Ahead is the only one I

am aware of. Is that right, Sue?

DR. S. YANOVSKI: Yes. NIDDK is

sponsoring the Look Ahead clinical trial, which is

5000 individuals with diabetes who are randomized

to intentional weight loss or a controlled

condition.

DR. HIRSCH: But no data are available?

DR. S. YANOVSKI: Not yet.

DR. BRAUNSTEIN: Why do you think the

mortality curve is J-shaped? That is, that the

mortality goes up as you start getting to a lower

BMI at a time when all the comorbid risk factors

seem to be lowest?

DR. FLEGAL: Well, again, there are a lot

of issues that are really unresolved. It could be

that at older ages there is some association--at

all ages there is an association of low BMI with

mortality as well as with high BMI. It may have to

do with issues like not having adequate nutritional

reserves; people going in for surgery at age 65 and

you lose weight in the course of being in a

hospital and deplete your nutritional reserves.

You may be at a higher risk of hip fracture. The

pattern of the causes of mortality may be different

at different BMI levels at different ages. There

are also issues of smoking. Most of these studies

adjust in some way for smoking but smokers tend to

have lower body mass index and be at higher risk.

So, there are a lot of different issues. I don't

think it has really been sorted out very clearly in

the literature.

DR. BRAUNSTEIN: Thank you. We will move

on to Dr. Frank Greenway's discussion of the

current status of weight-loss drugs.

Current Status of Weight-Loss Drugs

DR. GREENWAY: I was asked to speak on the

safety and efficacy of the drugs that we have for

weight loss at the present time. Obesity, before

the 1985 consensus conference, was felt to be bad

habits rather than a chronic disease, which is the

way we now understand it. At least, it was my

understanding that eating habits can be retrained

over a period of a few weeks and that this at least

was another reason why the older recommendation for

obesity drugs was over a shorter period of time.

The drugs approved before 1985 were,

therefore, approved for periods up to a few weeks,

and tested over that period of time. Mazindol and

fenfluramine are no longer available; phentermine

and diethylpropion are. Dr. Colman already

reviewed the analysis of the FDA information on new

drug applications that were reviewed in the 1970s

that showed that these drugs approximately doubled

the weight loss seen with the placebo groups.

Just a few overview comments about

treating obesity as a chronic disease with

medications, first of all, the drugs work only when

they are taken and I will show you a slide to

demonstrate that. The average weight of the

participants in those studies is 100 kg. So, one

can look at these weight loss graphs as percent

weight losses or kilograms of weight loss since it

is 100 kg.

The placebo group in these trials has

always required some type of treatment because IRBs

feel that placebo groups need to get some form of

treatment as well. So, people in these trials are

really getting two different treatments. Weight

loss in these trials usually plateaus at about 6

months. The primary criteria for approving drugs

in Europe is a 10 percent weight loss that is

greater than placebo. A primary criterion in the

United States is a weight loss that is 5 percent

greater than placebo and is statistically

significant.

This is a slide of a study done in a

practice situation where patients were given

fenfluramine, a drug no longer approved. They were

seen monthly for a year. As you can see, the

weight loss plateaus at about 6 months. The

one-year people in this trial had their drug

discontinued but they would continue to be followed

for the following year. When they were followed

off the treatment the weight loss just about went

away by the time they got to the second year.

Another point I wanted to make was about

the ancillary treatment that goes on in these

clinical trials. Back in the early '70s behavior

modification was a new treatment. There was a

trial that was done to approve mazindol and two of

the sites did it in the standard way, which is

demonstrated on this slide. Everybody got a

tear-off diet sheet and the placebo and drug groups

were given pills each week and were weighed each

week. As one can see, the placebo group really

lost no weight over 6 weeks and the mazindol group

lost 6.5 lbs over that period of time.

In another site in that trial behavior

modification was superimposed upon all groups. The

mazindol group in that site lost 8.5 lbs rather

than 6.5 lbs but the difference between drug and

placebo was reduced considerably, to the point

where, at least in this particular site, the

difference was no longer significant.

To sort of carry that forward, because

this is sort of the difference between the European

and U.S. kinds of criteria, here you can see that

an orlistat trial in Europe had 11 percent weight

loss but only a 2 percent difference from placebo.

This is because there was presumably a larger

ancillary program that was superimposed upon this

weight loss program.

This is a sibutramine trial that was done

in the United States where the difference was to

get a spread between the two groups. You have a 7

percent weight loss with sibutramine and a 2

percent loss with placebo, and there was,

therefore, a 5 percent difference.

In talking about the safety and efficacy

of the drugs that are presently available, the Rand

Corporation was commissioned to prepare an evidence

report on the pharmacologic treatment of obesity by

the agency for Health Care Policy and Research of

our federal government. Some new meta-analyses

were done during that process using the studies

that were at least 6 months in duration. Although

this hasn't yet been published, they have given me

permission to present some of that data.

There are several categories that one can

put the drugs available into. One would be

phentermine and diethylpropion which are approved

for obesity but for short-term use. The second

would be orlistat and sibutramine which are

approved for obesity for long-term use. Then there

are drugs that are approved for other indications,

not for obesity, things that are approved for

depression, like fluoxetine and bupropion; things

that are approved for epilepsy such as topiramate

and zonisamide which also give weight loss. Then,

there are 2 drugs that are in phase 3 clinical

trials, Axokine and rimonabant which have some

public information available on them.

The data presented here on efficacy

presents the data in the way the FDA evaluates

drugs, that is, the difference between the placebo

group and the drug group. In trials of phentermine

up to 6 months in duration, using 30 mg/day, the

difference between drug and placebo was about 3.5

kg. With diethylpropion, in studies that went up

to about a year, the difference was 3 kg.

One might ask how can one, in this day and

age when we understand obesity to be a chronic

disease, find a use for these medications that are

only approved over a period of a few weeks. This

is a study that was done comparing the green line,

which shows continuous use of phentermine, against

the yellow line, which showed 1 month on 1 month

off; 1 month on, 1 month off.

As you can see, the line is more jagged

but they end up at approximately the same place at

9 months compared to the red line, which is

placebo. So, there are still ways that these drugs

can be useful.

Orlistat, at 120 mg 3 times a day, gave a

2.5 kg difference compared to placebo at 6 months

in the 11 studies in this meta-analysis, and about

2.75 kg at 1 year in 21 studies.

Orlistat is an inhibitor of pancreatic

lipase. It causes a third of dietary fat to be

lost in the stool. The relative risks for diarrhea

were 3.4, for flatulence 3.1, and dyspepsia 1.5.

So, one can see that these side effects result from

the mechanism of action.

These trials showed a reduction in total

LDL cholesterol and in blood pressure. There was a

slight reduction in glucose and glycohemoglobin in

diabetics, and it was shown that one could prevent

diabetes in those with impaired glucose tolerance.

Sibutramine, in doses of 10-20 mg/day,

showed a 3.5 kg difference from placebo at 6 months

in 12 trials, and about a 4.5 kg difference at 1

year in 5 trials. Sibutramine is a norepinephrine

and serotonin reuptake inhibitor. It had

dose-related dry mouth, insomnia and nausea

associated with it. The heart rate went up 4

beats/minute in these trials, and there was no

consistent effect on blood pressure or lipids.

There was a slight improvement in glucose and

glycohemoglobin in diabetics.

One could logically ask, since we have

these two drugs that are approved for long-term use

in obesity and they work by different mechanisms,

could one combine them and get better weight loss.

This is one trial that tried to address that issue.

The yellow line shows sibutramine treatment for a

year. You can see that the weight loss plateau'd

at 6 months and remained stable for the next 6

months. When orlistat was added to sibutramine

there was no further weight loss.

Fluoxetine is a medication that was

approved for depression, not for obesity. It was

studied for obesity, however, and at 60 mg/day, a

higher dose than is typically used for depression,

it caused about a 4.5 kg difference from placebo at

6 months. But, as you probably will notice as

something different compared to the other slides,

there is less difference at 1 year than there was

at 6 months, in this case 3 kg.

Fluoxetine is a reuptake inhibitor of

serotonin. The relative risks of nervousness,

sweating and tremors was 6.6; of nausea and

vomiting 2.7; fatigue and somnolence 2.4; insomnia

2.0; and diarrhea 1.7. There was regain of weight

between 6 months and a year. That is presumably

the reason that it was not approved.

This is a slide to graphically demonstrate

that fact. You can see that the weight loss came

down and plateau'd at around 6 months, but in the

last 6 months of that year there was obvious weight

gain in the fluoxetine group and not in the placebo

group.

Bupropion is a drug that is approved for

depression and smoking cessation. At 200 mg twice

a day in 2 6-month trials there was about a 2 kg

difference from placebo. In one trial at 1 year

there was about a 5 kg difference.

Bupropion is a reuptake inhibitor of

dopamine and norepinephrine. The 6-month studies

were both in depressed patients. The 12-month

study was in obese patients that were not

depressed. So, these may represent 2 different

groups in terms of response. The relative risk for

dry mouth was 3. There was also an increased

incidence in insomnia, and there were no increases

in pulse or blood pressure in those studies.

Topiramate is a drug approved for

epilepsy, not for obesity. At 192 mg/day there was

a trial that showed a 6.5 kg difference between

that drug and placebo. The mechanism or weight

loss with this drug is not clear. The relative

risk of paresthesia was 4.9. Taste perversions was

9.2. There were other central nervous system and

gastrointestinal side effects with this medication.

Zonisamide is another anti-epileptic drug,

not approved for use in obesity. A 16-week trial

showed a 5 kg difference between that drug and

placebo.

Axokine is a large protein that is

injected subcutaneously and is in development in

phase 3 for the treatment of obesity. There is one

study that is in the public domain that shows a 3.5

kg difference from placebo at 1 year. Axokine

appears to activate the leptin pathway distal to

the place where leptin acts since it acts in

animals that don't have leptin. It has injection

site reactions, nausea and a dry cough associated

with its use. Over 30 percent of the people in the

trial that I mentioned developed antibodies to

Axokine. Those patients who developed these

antibodies lost less than 1 percent of their body

weight compared to placebo at a year.

Rimonabant is the other medication on

which there is public information in the phase 3

trials for the treatment of obesity. The one trial

that was reported talked about uncomplicated

obesity. It was a 16-week trial and I took the

liberty of projecting the weight loss consistent

with other weight loss curves of these types of

drugs. If one projects that out to 6 months, one

gets just slightly less than a 5 kg difference,

assuming no weight loss in the placebo group which

was not reported on that website.

There is a second trial that used

rimonabant in dyslipidemic patients. The

difference from placebo was 5 kg at 6 months and

6.5 kg at a year.

Rimonabant is an antagonist of

cannabinoid-1 receptor. In other words, it blocks

the receptor that is thought to be effective in

causing the munchies when people smoke marijuana.

Nausea and diarrhea were greater than 5 percent

above placebo. There was a 10 percent increase in

HDL, a 15 percent reduction in triglycerides and a

reduction in the 2-hour post glucose load insulin,

and no significant effects on pulse or blood

pressure in these dyslipidemic patients.

I put in this slide to put into context

the blue line, which is a typical drug where there

is weight loss of 10 percent, compared with the

gastric bypass which has weight loss of 30 percent

which is durable over 14 years.

In summary, there are short-term weight

loss medications that are approved for treatment of

obesity, such as phentermine and diethylpropion.

There are drugs that are approved for the long-term

use in the treatment of obesity, that is, orlistat

and sibutramine. There are other medications

approved for epilepsy or depression, i.e.,

bupropion, fluoxetine, topiramate and zonisamide,

which are not approved for use in treating obesity

but which seem to give weight loss. And, there are

two drugs, Axokine and rimonabant, about which

there is public information that are presently in

phase 3 trials for the treatment of obesity.

In conclusion, all these drugs give

between a 2 and 6.5 kg greater weight loss than

placebo in trials that last up to a year, and the

amount of weight loss appears to be medically

significant. The weight loss between these

different drugs is not different statistically and

the choice, therefore, revolves around side

effects. The weight loss and the difference from

placebo are two different things, which I hope I

demonstrated, and data beyond 2 years essentially

does not exist, with a couple of exceptions. Thank

you.

DR. BRAUNSTEIN: Thank you. Questions

from the panel? Yes, Dr. Woolf?

DR. WOOLF: There was a report in "New

York Times" on Monday, I think it was, of results

of a one-year or a two-year trial in Europe with a

drug that they didn't specify, other than saying it

was a receptor blocker of some sort that had, I

think, 19 lbs weight loss and 3.5 inch reduction in

waist and a 24 percent increase in HDL. Do you

know anything about that?

DR. GREENWAY: That was rimonabant. I saw

that article and that was about rimonabant.

DR. WOOLF: Sorry?

DR. GREENWAY: I read the article and it

was reporting on rimonabant, a new study of

rimonabant, not the one that was reported by Frank.

DR. WOOLF: Thank you.

DR. GREENWAY: Actually, those results

are on the website. I checked it yesterday, 1

year, 52 weeks, 5 and 20 mg.

DR. BRAUNSTEIN: Yes?

DR. ARONNE: Frank, can you talk a little

bit about the problem with dropouts in weight-loss

drug studies, and some of the pros and cons of the

type of analyses used, last observation carried

forward versus completers?

DR. GREENWAY: Well, it seems as though

people in weight-loss studies appear to have a

feeling of being stigmatized when they drop out of

studies because they don't want to come back. It

is very difficult to get final data on people who

drop from weight-loss studies. Weight-loss studies

that go out to a year usually have something like a

30 percent dropout rate.

The traditional way of analyzing these

studies, as Susan suggested, has been the last

observation carried forward, and what that does is

it dilutes the effect of the drug because it

assumes that the reason the people dropped out is

because they didn't lose weight. Actually, what

the physician treating a patient is interested in

is more what happens to the patient that I treat

who stays in treatment, rather than the more public

health perspective of this last observation carried

forward which looks at the entire group. If you

treat everybody, what does the total group gain

from this experience? So, from the way in which

these medications are used, it is much more

informative to me, as a clinician, to have the

analysis of completers rather than the last

observation carried forward.

DR. BRAUNSTEIN: Dr. Schambelan?

DR. SCHAMBELAN: Just a quick question

about Axokine. You said it worked distal to

leptin. Do you know if it works distal to the

leptin receptor or just distal to leptin? Is its

actual site of action known?

DR. GREENWAY: The site of action of

Axokine is in the leptin pathway. It is probably

in that signaling pathway but it is distal to the

site where leptin acts.

DR. BRAUNSTEIN: Frank, can you describe,

in the studies that were carried out for one year

with these drugs, what the effect was on the

comorbid states and whether there were any

differences among the drugs? For instance, did

some lead to lowering of blood pressure and others

didn't? Did some lead to lowering of cholesterol

while others didn't? Or were they all fairly

consistent?

DR. GREENWAY: You are asking me what was

the effect on comorbidities in these studies?

DR. BRAUNSTEIN: Yes.

DR. GREENWAY: Of the two drugs that are

approved for treatment of obesity in the United

States, orlistat seems to have a disproportionate

beneficial effect on lipids, probably because it

enforces a low fat diet. Sibutramine doesn't have

the expected beneficial effect on blood pressure

that one might expect, probably because of its

norepinephrine reuptake mechanism of action.

Otherwise, one gets the expected benefits that one

would expect with weight loss with these drugs.

DR. BRAUNSTEIN: Other questions?

[No response]

Thank you. Our next speaker will be Dr.

Laura Governale, who is going to speak about

patterns of weight-loss drug use.

Patterns of Weight-Loss Drug Use

DR. GOVERNALE: Good morning. To begin, I

would like to briefly state that the Division of

Surveillance Research and Communications Support in

the Office of Drug Safety is responsible for the

procurement, management and analysis of drug

utilization databases for the FDA's use. The

information contained in these slides has been

approved for this meeting.

The topics I will be discussing today are

the patterns of prescription weight-loss drug use

and the patient demographics associated with

weight-loss drug use. For this presentation

weight-loss drugs are defined as dexfenfluramine,

sibutramine and orlistat and amphetamine congeners

such as phentermine and dimetrazine diethylpropion,

phendimetrazine, diethylpropion, benzphetamine,

mazindol and fenfluramine. We did not include

amphetamines in this analysis. Also not covered in

this analysis are over-the-counter drugs and

nutritional supplements. The analysis is conducted

using proprietary databases at the agency's

disposal.

Two databases were used in this analysis

from IMS Health. IMS health is a pharmaceutical

marketing usage company that collects prescription

drug use information worldwide. The agency uses

these databases as well in order to obtain drug use

information and trends in the U.S. The two

databases from IMS Health were the National

Prescription Audit Plus and the National Disease

and Therapeutic Index.

NPA, or the National Prescription Audit

Plus, measures the retail outflow of prescriptions

from pharmacies into the hands of consumers by

formal prescriptions. The number of dispensed

prescriptions is obtained from a sample of

approximately 22,000 randomly selected pharmacies

around the country and projected nationally. The

pharmacies in the database account for

approximately 40 percent of all pharmacy stores and

represent approximately 45 percent of prescription

coverage in the U.S. The pharmacies include the

following retail channels such as chain,

independent, mass merchandisers and food stores

with pharmacies, and also include mail-order and

long-term care pharmacies.

The National Disease and Therapeutic Index

is a survey of roughly 3000 office-based physicians

around the country. The data gathered in NDTI are

designed to provide descriptive information on the

patterns and treatment of disease encountered in

this setting. The data are collected and projected

to provide a national estimate of use. However, in

certain instances the small sample size tend to

make these data unstable and sometimes these

results should be interpreted with caution.

Patterns for prescription weight-loss

drugs dispensed were obtained from NPA Plus. Here

I will present the trends in prescription

weight-loss drug use dispensed from 1966 to 2003

and also the method of payment for these

prescription weight-loss drugs from 1999 to 2003.

This slide, which is based on NPA data,

represents the total number of prescriptions

dispensed for prescription weight-loss products

from 1966 to 2003. The total number of

prescriptions represents new prescriptions as well

as refill prescriptions.

The yellow-shaded area here represents the

total added prescription weight-loss products of

all the individual weight-loss products as shown

here. The individual lines represent individual

active ingredients in some of these weight-loss

drug products. Again, this slide does not include

any amphetamine products.

As you can see, over the last 38 years

there have been fluctuations in prescription

weight-loss products. As you can see, there are

two major spikes in prescription drug use. These

fluctuations in use have been largely due to two or

three prescription drugs at any given time.

The first spike, which occurred during the

early 1970s, around the decade of the '70s, was

most likely due to the enactment of the Controlled

Substances Act in 1970. This was also presented by

Dr. Colman in a previous presentation. This

legislation in essence restricted the production

and distribution of amphetamines which, throughout

the 1960s, were commonly prescribed for weight

loss. When these restrictions were placed on

amphetamines the amphetamine congeners were used

more frequently.

Also in 1973, the agency declared that

amphetamine and amphetamine-like compounds were

effective for the treatment of obesity. This led

to a large spike in use for diethylpropion and

phentermine products. The number of prescriptions

here peaked at 12.5 million in 1976.

However, we see a decline in use around

1979. In 1979 there was a Federal Register notice

calling for the removal of the obesity indication

in amphetamines. This led to a sharp decline in

use in weight-loss drugs, namely, for phentermine

and diethylpropion. However, the proposal to

remove the obesity indication from the amphetamines

never materialized. Since then, the use of

weight-loss products had steadily declined until

the mid-1990s.

I will focus now on the last 13 years for

prescription drug trends. Looking at the last 13

years, the number of total prescriptions dispensed

for weight-loss drugs reached its lowest point

around the 1990s, early 1990s, with approximately

3.3 million prescriptions dispensed.

Then, in early 1995, 1996, we began to

notice an increase in usage. This was most likely

due to the result of a publication, in 1992, of a

series of papers that concluded that the

combination of phentermine and fenfluramine, or

phen-fen, was safe and effective for long-term

weight loss. In 1996 the FDA approved

dexfenfluramine for the treatment of obesity. The

number of anti-obesity prescription drugs dispensed

reached its peak in 1996 with 21 million

prescriptions. The compounds responsible for this

increase include phentermine, fenfluramine and

dexfenfluramine.

Again, dexfenfluramine was marketed under

the name of Vidoxx and fenfluramine was marketed

under the name of Pondimin. During its peak use in

1996 fenfluramine held 33 percent of the market

share with 7 million prescriptions dispensed,

whereas dexfenfluramine held 11 percent of the

market share with 2.3 million prescriptions

dispensed. Phentermine held 52 percent of the

market share with approximately 11 million

prescriptions dispensed.

This large spike in use was followed by a

market decline over the next two years when, in

1997, the FDA announced a voluntary withdrawal of

fenfluramine and dexfenfluramine following

increased reports of cardiac valvulopathy in

patients treated for obesity. The total number of

prescriptions dispensed went from a peak of 21

million prescriptions down to approximately 7

million prescriptions ion 1998, which represents

approximately a 67 percent decline. Since then the

number of prescriptions dispensed for weight-loss

drugs has declined to approximately 5.8 million

prescriptions in the year 2003.

Orlistat was released into the market

around 1997, and sibutramine in 1999. Currently,

or in year 2003, they hold second and third place

in the market with 1.3 million prescriptions

dispensed for orlistat or 22 percent of the market

share, and 760,000 prescriptions dispensed for

sibutramine, which represents 13 percent of the

market share.

Phentermine continues to predominate the

market with approximately 3 million prescriptions

dispensed, which represents over 50 percent of the

market share. Other products, such as the

amphetamine congeners, have steadily declined in

use since the mid-1990s and collectively account

for less than a million prescriptions per year.

This slide, in contrast to the previous

slides, represents only new prescriptions

dispensed. Furthermore, this analysis excludes the

mail-order and long-term care channels. Therefore,

the numbers of prescriptions reported in this

analysis are smaller than in the previous slides.

This graph is an analysis of method of

payment for prescription weight-loss drugs. As you

can see, the number of new prescriptions paid by

cash has declined steadily over the past 5 years,

from approximately 5 million in year 1999 to 2.6

million in year 2003. However, the number of

third-party payment for new prescriptions has

remained steady at approximately 1.1 to 1.6 million

prescriptions over the 5-year period surveyed. The

drop in cash payment, in effect, has increased the

proportion of third-party payment for these drugs

from 20 percent in year 1999 to approximately 30

percent in year 2003. So, the main message from

this slide is that cash payment remains an

important mechanism for the payment of these

weight-loss prescription drugs.

Next I will discuss the patient

demographics associated with prescription

weight-loss drugs. The data are based on IMS

Health, National Diseases and Therapeutic Index.

Again, the data are projected nationally. However,

it does not represent disease burden, nor is it

representative of all disease states in the nation.

Rather, the data reflect a population of ambulatory

patients, visiting physicians and office-based

practice settings during which a weight-loss drug

is mentioned during the visit. Again, due to the

limitations in data sampling in this database, any

perceived trends must be interpreted with caution.

The topics I will be discussing for

patient demographics include the principal

diagnoses associated with prescription weight-loss

drugs, the gender distribution, age distribution

and race distribution.

This table represents the principal

diagnoses associated with prescription weight-loss

products for ambulatory patients. Not

surprisingly, obesity is the diagnosis most often

mentioned with weight-loss drug products, with

approximately 89 percent or 1.8 million projected

diagnosis visits.

This slide represents the number of

mentions associated with the use of weight-loss

drugs as reported by office-based physician

practice settings. This is a measure of drug

mentions again and is not reflective of disease

burden in the nation.

As you can see, females account for a

clear majority in use for prescriptions of

weight-loss products, with an average of 2.3

million drug appearances or 85 percent over the

time period surveyed.

Taking a closer look at the most recent

calendar year, we see that the adult age group,

18-44, accounts for the largest majority of drug

use for prescription weight-loss products, with

approximately 1.2 million or 62 percent of total

drug appearances. This is followed next by age

45-64 category, with 624,000 mentions or 32.6

percent of total drug appearances in the year 2003.

In conclusion, the majority of weight-loss drug

products is in the young, female, adult and middle

age adults.

This graph represents the race

distribution of patients associated with the use of

prescription weight-loss drugs as reported by

office-based physician practice settings. Again,

the reporting in this database, NDTI, is reported

by the physician and not is not self-reported by

the patient. The key take-away from this graph is

that a proportion represented by each race group

has remained constant over the time period

surveyed. Approximately three-quarters of use is

from Caucasian patients.

Now that I have represented the data, I

will present the limitations on each of these

databases. The NDA Plus data provide only limited

demographic information on prescription use.

Therefore, we did not use this database for this

analysis. Instead, we used NDTI to obtain

demographic information which has these

limitations: As you can see, the small sample size

makes some projections unstable. Again, the data

are not generalizable to all of these patients.

And, due to the limitations, any perceived trends

must be interpreted with caution.

In conclusion, over the last 38 years

there has been a fluctuation in the total number of

prescriptions dispensed for prescription

weight-loss products. These fluctuations have been

largely due to two or three drugs at any given

time.

The second point is that cash payment

remains an important mechanism for payment for

these products. Also the primary users of these

products are Caucasian women between the ages of 18

100

and 44. That is the end of the presentation.

DR. BRAUNSTEIN: Thank you. Any questions

from members of the panel? Yes?

MS. COFFIN: On your slide that talks

about the race distribution of the patients you can

see huge differences and, of course, the Caucasian

patients are shown as the largest amount. How does

that normalize to the population as a whole? Is

the population as a whole from '98 to 2003 more

greatly Caucasian than it is Asian American or

African American?

DR. GOVERNALE: Again, this database is

not supposed to represent any epidemiology of

obesity. It represents patients visiting

office-based physicians and it could reflect just

that there are more Caucasian patients visiting

these physicians.

DR. BRAUNSTEIN: Dr. Woolf?

DR. WOOLF: Do you know if the heavy use

of cash for these drugs is because they are being

excluded by drug plans that the patients have? Are

they being excluded from the formularies that the

101

patients are covered on? Is that why cash is so

prevalent?

DR. GOVERNALE: I think if I heard your

question, it is why are most of these products not

covered?

DR. WOOLF: Yes, is the reason that cash

accounts for three-quarters of the method of

payment because they are being excluded from drug

plans?

DR. GOVERNALE: Yes, that is the

limitation with these products. Most of these

products are not covered by third-party payers and,

therefore, that is why they are being paid for by

cash.

DR. BRAUNSTEIN: Dr. Watts?

DR. WATTS: I was interested in your

demographics. I may be making wrong inferences but

it seems to me if the largest use of these drugs in

the real world is by younger white women, that may

be more for cosmetic benefits of weight loss. This

is a question then for Dr. Greenway. I didn't get

from your presentation the demographics of the

102

subjects that are studied in the typical weight

loss trial. Are they different from the people who

are using these drugs as we heard from this

presentation?

DR. GREENWAY: The patients in the regular

weight-loss trials primarily have a BMI between 30

and 40. So, they aren't in the trials because they

have just cosmetic concerns, but I think that what

you have observed is correct, that obesity is

stigmatized in our society, particularly

stigmatized in regards to women, and that is

probably the reason that we have 80 percent of

these obesity trials that are composed of women.

Clearly, 80 percent of the population isn't women.

DR. WATTS: To extend that though, is

there a particular age of the subjects in the

studies that you showed? Were they different,

older, from the use of these drugs in the real

world?

DR. GREENWAY: The average age of the

people in the trials is usually around 40. So,

they may be slightly older than this group but I

103

think they are probably fairly representative.

DR. BRAUNSTEIN: Dr. Yanovski?

DR. S. YANOVSKI: These data did not

report out BMI. They might have the physician

diagnosis for obesity correct but you couldn't tell

how many of the patients in these studies had BMIs

above a certain range. Is that correct?

DR. GOVERNALE: Correct. There is no

linkage of BMIs to the diagnosis of obesity.

DR. S. YANOVSKI: So, in preparation for

this I pulled an article by Laura Kettle Conning

and colleagues at CDC that looked at use of

prescription weight-loss pills in U.S. adults from

1996-98 that I think addresses your question. They

used the behavioral risk factors surveillance

survey and they looked at all patients who reported

use of prescription weight-loss drugs. They then

looked at the proportion of patients who reported

using prescription weight-loss drugs who had a BMI

of less than 27, which was the lower limit for

indication with comorbidities. What they found was

that 5 million U.S. adults had used prescription

104

weight-loss drugs in that 2-year period. Of that

group, 25 percent reported that they had a BMI of

less than 25. So, it looks like there is

substantial use of these medications for cosmetic

purposes.

DR. BRAUNSTEIN: Dr. Schambelan?

DR. SCHAMBELAN: I just want to pursue Dr.

Woolf's question about the reason that the cash

payment is decreased. Do we know that there has

been a systematic change in policy of third-party

payers as to what they will approve for weight-loss

drugs? Perhaps you don't know but other panelists

may know.

DR. BRAUNSTEIN: What will the effect of

the recent change in coverage of the drugs by

Medicare have on all this? I guess that is part of

the question.

DR. SCHAMBELAN: Well, Medicare or other

payers.

DR. GOVERNALE: We did not look into the

reasons for why some of these prescription drug

products are being covered or not covered by

105

third-parties, but that could be a very interesting

question to look into for future analyses.

DR. BRAUNSTEIN: Dr. Aronne?

DR. ARONNE: While there hasn't been a

systematic change in the number of plans which are

covering drugs, what we have seen is that

practitioners of obesity medicine where we focus on

obesity to treat someone's diabetes, sleep apnea or

other complications, is a steady increase in the

willingness of insurance companies to pay for drug

therapy in an appropriate setting. So, with prior

approval, if the patient is in a medically

supervised program, the insurance companies will

pay for the drugs. Right now in the New York area

it is more than 40 percent. The last number I

heard was that 44 percent of patients who have

insurance get coverage for these types of drugs.

DR. BRAUNSTEIN: Any further questions?

[No response]

We will take a 15-minute break. Thank

you.

[Brief recess]

106

DR. BRAUNSTEIN: We are changing the order

a bit. We are going to ask Dr. Richard Atkinson,

who is director of Obetech Obesity Research Center,

to speak on the role of drugs in the treatment of

obesity: current and future. Following that, we

will then move to the open public hearing, followed

by Dr. Orloff's talk.

Role of Drugs in the Treatment of Obesity:

Current and Future

DR. ATKINSON: Thank you, Dr. Braunstein.

Thank you, Dr. Orloff and Dr. Colman for inviting

me to speak. I am coming today wearing two hats.

One is the president of the American Obesity

Association and the second is a physician/clinician

who has literally treated thousands of obese people

over the years.

From that perspective, I have looked into

the eyes of these people and seen the pain and

heard their pain as they talk, and I have failed

them and I think we have all failed them. The

physicians and scientists have failed them. The

drug companies have failed them and the government

107

has failed them. That sounds like a negative

message and I am going to spend a little time

talking about why I think we have all failed. But

I think the promise of the future is really very

bright and I will try to end up on that.

I always like to start off with something

that is not unique to me; I probably stole it from

someone, but obesity is a chronic disease of

multiple etiologies characterized by the presence

of excess adipose tissue. Everybody has excess

adipose tissue but the critical word I think here

is "disease." I think we have heard in this

discussion this morning even some questioning of

the idea of obesity as a disease. But I believe

obesity is a chronic disease and if you think of

other chronic diseases, try to think of one that is

not treated with drugs.

If obesity is a chronic disease and most

other chronic diseases are treated with drugs, why

not obesity? We know that the biochemistry of

obese individuals is different from that of lean

people. That is very well known. Bob Eckle and

108

others have some data that when obese people lose

weight their biochemistry does not become the same

as lean people's. For example, lipoprotein lipase,

the major clearer of triglycerides out of the

bloodstream, in adipose tissue lipoprotein lipase

goes up; in muscle it goes down. So, people who

were formerly obese are poised to regain their fat.

What do we do with drugs? We change the

biochemistry. So, the rationale for using drugs is

to change the biochemistry of the bodies of obese

people.

There have been, as you have heard, a

number of barriers to the use of drugs. The first

one I am going to put up here is discrimination

against obesity. I am going to spend several

slides talking about this.

When Dr. Orloff asked me to talk, we

talked about the fact that we were going to have a

very nice bunch of scientific presentations and I

am going to come with a more emotional part with

this presentation. But as president of an

organization that is advocating for these people, I

109

want to point out some of the discrimination

against obesity. The fact of physician and

clinician ignorance of obesity, an particularly of

obesity drugs; economic factors; policy and

political barriers, and I have come much more to

appreciate that. We have had several meetings with

people from the FDA, NIH and others and I have come

to appreciate more some of the barriers. There is

a lack of advocacy about obese people and, finally,

currently there is a modest effectiveness of

obesity drugs, as we have heard.

I am going to talk a little bit about

discrimination. Obesity is the last bastion of

socially acceptable bigotry. If you are a radio

announcer or a TV announcer and you tell a joke, a

race joke or an ethnic joke, or a joke directed

against homosexuals, you will get fired. Fat jokes

are told all the time. Look in your comic pages

and virtually every day there is some slam against

fat people and nothing is done.

This discrimination against obesity is in

the people who are in our field. Stan Heshka and

110

David Allison are two very good friends, two very

bright people, good scientists, but "labeling

obesity a disease may be expedient but it is not a

necessary step in a campaign to combat obesity and

it may be interpreted as a self-serving advocacy

without a sound scientific basis." Well, those are

pretty strong words for somebody who is in the

field.

There is a lack of medicalization of

obesity. Think about obesity compared with some

other chronic diseases. For example, newly

diagnosed type 2 diabetes, newly diagnosed

hypertension--a very high percentage of those

patients will respond very well to diet and

exercise. it goes away. I did a study about 20

years ago and it goes away in 80 percent of the

people. But the first words our of the mouth of a

primary care physician are not "I'm going to put

you on a diet and exercise program;" it is "I'm

going to put you on drugs."

The primary treatment for obesity is diet

and exercise and drugs are an adjunct. As we have

111

heard from Colman's talk, that has been true for

many, many years. Many patients must demonstrate

that they have failed diet and exercise before they

can get either drugs or surgery. There is no other

disease where that happens.

Physician and clinician

ignorance--obesity, obviously, is not thought to be

a real disease. As many of you know, we have been

doing some work on viruses that cause obesity and I

have gotten up and had people shake their fist at

me and say, "you're trying to give these fat people

an excuse." Wow! Physicians are uncomfortable

about counseling overweight or obese patients. I

have a talk on discrimination against obesity to

document many papers in the literature where this

has been shown.

Physicians and clinicians are not

knowledgeable about nutrition, physical activity,

and particularly about obesity drugs. This is a

disease that is killing 400,000 people per year

according to the CDC. At the University of

Wisconsin I was able to get a clinical nutrition

112

course on the curriculum and we had exactly three

lectures on obesity. Since I left, that has now

been cut to one. That is pretty much all they get

about obesity in the whole curriculum.

Physicians are unaware of referral

information. If you have a fat person, what do you

do? They feel helpless and there is a feeling that

if you refer a patient to an obesity physician you

are sort of sending them to a charlatan. There is

a bias. We have heard a little bit about that

today, that drug treatments are dangerous,

ineffective and somehow not worthy.

There are economic factors that are

barriers to obesity drugs. I was flabbergasted to

hear Lou Aronne's comment that in New York 40

percent of third-party payers are starting to pay

for drugs. That is super. We looked in Wisconsin

and in our population it was between 10-15 percent.

They had a very high percentage of HMOs and these

HMOs simply didn't cover obesity or obesity drugs.

Some of the reasons for that are that the

treatment is fairly expensive. This, after all, is

113

a chronic disease. The insurance companies and

employers are worried about breaking the bank.

There are a large number of overweight and obese

people. We heard Katherine Flegal's talk. Over 30

percent of the entire adult population is obese,

has a BMI of 30 or above. So, one, there is a

large population that might want to use those drugs

or use that treatment and, secondly, given a

choice, they will.

We heard from Susan Yanovski about maybe

as many as 25 percent of people that are using

these drugs are using them for cosmetic purposes.

That is a little bit of a discrimination in itself.

There is a whole industry that makes drugs for

cosmetic purposes, like skin rashes and so forth

and so on. So, what is so bad about somebody

wanting to lose some weight when, if you are

overweight, you have a harder time getting a job,

getting promoted in a job, finding a spouse. If

you are a small kid other kids don't want to play

with you. It is not just a cosmetic problem; this

is a socioeconomic huge discrimination problem

114

against obese people.

For many of the insurance companies and

insurance plans and HMOs savings in the future

costs are too remote compared to current expenses.

Their bottom line is a year or less. If it doesn't

pay for itself in a year, let's don't pay for it.

Turning to the government, honest to God,

at a Harvard CME course put on by George Blackburn,

I am happy to say a former member of the FDA, he

made the statement, "gaining weight doesn't hurt

you and losing weight doesn't help you." I am

embarrassed to say I got into a shouting match with

him in front of 400 people.

Obesity drugs I think have been held to a

different standard in the past than drugs for other

diseases. I will just bring up the phen-fen

debacle versus troglitazone. Within two months of

the first unconfirmed, uncontrolled case series

that was prematurely reported by The New England

Journal--it wasn't even published yet but what was

released as a press release--within two months of

that fenfluramine and dexfenfluramine were taken

115

off the market. It was not really sure that anyone

had died from fenfluramine or dexfenfluramine.

Sixty people had died from troglitazone and it

stayed on the market two more years.

Now, the cynic in me says that is because

diabetes is a real disease and obesity is not.

That may not be fair and there are other factors,

but from my side, coming from an advocacy

organization, it looks like that is discrimination

against obesity. I am not saying that fenfluramine

and dexfenfluramine should not have been taken off

the market, but the timing was interesting.

The recent experience of obesity

drugs--dexfenfluramine had quite a hard time

getting approved. Sibutramine was initially turned

down and only upon appeal was approved. Orlistat

had what apparently was a spurious association with

cancer so they had to go back and do a great many

more trials to look at the patients to show that

there was not a correlation with cancer.

As many of you know, and as many of you

here have participated in, the American Obesity

116

Association has sponsored a series of meetings with

people from the FDA, the NIH, other government

agencies, scientists and many representatives of

the pharmaceutical industry who have obesity drugs

or are interested in obesity drugs. And, one of

the things I have been impressed with is that the

people at FDA have a huge load on their shoulders

because if anything goes wrong, it is their

problem. We have, you know, 100 million people who

might be wanting to take these drugs and if even a

few of them start to have problems, it is the FDA's

fault for having not been more careful.

Fenfluramine had been on the market since

1973. It was not until 1997 that it was found to

have cardiac valve problems. The problem with

pulmonary hypertension, as Eric noted, was there

but it was really pretty rare. So, I have a much

better understanding of the pressures, both

political and from media and from scientists, on

the FDA and why they simply have to be cautious.

From the Medicare/Medicaid perspective, we

have already heard today that until just a month or

117

so ago the language in the Medicare and Medicaid

regulations was "obesity is not a disease" despite

the fact that it was called a disease in 1985 by an

NIH consensus development conference. Apparently,

efficacy standards, in contrast to drugs and

treatments for most other chronic diseases, will

have to have some sort of proving that this

treatment works. Now, as I said earlier, we

have failed these people but obese people fail

themselves. The expectations and the behavior of

obese people contribute to the problem because many

do not believe they are worthy or respect. Obese

people discriminate against obese people actually

more than thin people discriminate against obese

people. They do not bind together for action.

Trying to get people to join this advocacy group

has been absolutely amazing. I thought everybody

in the world who was obese would sign up. They

don't. They are ashamed to be associated with the

world of obesity. They simply do not act as

advocates.

Other barriers to obesity drugs are

118

limited choices and poor efficacy. There are only

two drugs still on patent. Why haven't the drug

companies done more in the past? They are

certainly doing it now. There are really only

three categories of drugs, as you have heard, the

adrenergic agents, sibutramine which is in a

category by itself and orlistat which is in a

category by itself. We still have an infantile

understanding of the etiology of obesity and

mechanisms of action of drugs. We heard about

topiramate. We don't have a clue as to how it

works. It causes weight loss but we don't know how

it works.

As we have heard, typical weight-loss

agents, single agents at least, cause only about a

10 percent loss from initial body weight, and there

has been very limited use of combinations of drugs.

I will come back to that.

This is the data on dexfenfluramine, the

index study from Europe, the best study that

dexfenfluramine had, and there was about a 10

percent weight loss at a year.

119

Sibutramine, about an 8 percent or 9

percent weight loss at a year with 15 mg.

Sibutramine out over 2 years, again over in

Europe--again, this is about a 13 percent weight

loss. This is the Storm trial.

With orlistat, about a 10 percent weight

loss at one year. This is a 2-year trial. The

2-year data was about 8 percent.

If you are a 220 lb woman and you lose 22

lbs, your physician can tell you all he or she

wishes, "oh, you're healthy; your blood pressure's

better; your blood sugar's better, your lipids are

better." That woman or that man who is obese is

still suffering the slings and arrows of

discrimination by society. As a matter of fact,

when we showed the data from 2000, John Monroe's

group in BMJ and in Practitioner back a long time

ago, these were 36-week trials and the percent

weight loss was about 13 percent in each. That is

pretty much all there is with phentermine which is

the most commonly used drug.

That is sort of, if not the bad news, at

120

least the mediocre news. Let's look at what is

going on for the future. I apologize, I am sure

some of the companies out there have some areas

that I have left out here. But we know gut

peptides are a very fierce focus of action with CCK

analogues and enterostatin and so on; opioid

antagonists, the ones in phase 3 trials; various

neurotransmitter agonists and antagonists;

thermogenic agents, the Holy Grail--increase your

metabolic rate, keep eating and increase your

muscle mass and reduce your fat mass. Growth

hormone and growth factors have been disappointing

to date but maybe there is something there. Things

that enhance lipid oxidation I think will be of

particular interest; and nutrient partitioning

agents will be very interesting agents for the

future. I am sure this isn't all. There are many

more areas in which we may be able to affect food

intake, body weight or body composition.

These are just some of the potential

agents. Again, several people and particularly

Frank have talked about bupropion, topiramate and

121

zonisamide which are already out there. There are

several in clinical trials, and then there are a

number of others here. From my understanding,

there are about 350 different drugs in the

pipeline.

This is the data on bupropion. Frank

already showed this, about a 10 percent weight loss

at a year.

Topiramate--I put this slide up because it

shows 5-year data in epilepsy patients. As Frank

showed you, it has a pretty reasonable comparison

against placebo at 6 months.

This is the data from Gadde on zonisamide,

again showing a 32-week weight loss of about 9

percent.

However, single drugs are not likely to be

very effective or much better than about 10 percent

or 15 percent because there are so many redundant

systems regulating food intake and body weight,

something so critical to life as that. So, I think

I am not sure we are even born yet with our use of

drug combinations. I talked about the infancy of

122

drubs. Obesity is a chronic disease. Most chronic

diseases are treated with drugs. Most chronic

diseases require more than one drug. How many

chronic diseases can you think of that are treated

with just one drug?

You heard about phentermine and

fenfluramine. Combinations of drugs may have

additive or synergistic effects. As Weintraub

showed with phen-fen, some of the side effects may

even be offset.

Here is the original Weintraub data

showing about a 15 percent weight loss at a year.

As you know, he took those data out to 4 years. He

had a pretty good dropout rate but still had

efficacy.

Here is another combination of ephedrine

and caffeine. Either one alone is not terribly

effective but the combination causes about a 16

percent weight loss that persisted out to a year.

That is, of course, not on the market anymore.

Here is some data that we did at the

University of Wisconsin comparing phen-fen to

123

phentermine and fluoxetine and the slope is about

the same. These are 6-month data. Again, we had a

pretty good dropout. We did not have a control

group.

This meeting is all about the guidances

and what is going to happen to the guidances. So,

let me put on my helmet, get my lance out and tilt

at this windmill for a while to talk about some of

the things that I think would be very useful to

have from an obesity advocacy point of view.

Obesity is a major public health problem.

We have an epidemic here. There have been 10 times

more people dying of obesity-related causes than

are dying of AIDS in this country alone. Why

shouldn't obesity drugs be fast track as they are

for many other drugs? As Dr. Greenway pointed out,

in virtually all the drugs that we see the weight

loss plateaus certainly by 6 months.

Why should we need to show efficacy? Why

should the trials go out past that? Why not have

safety? You know, safety is what is really

important. If you show efficacy, and almost all

124

the drugs show a 5 percent weight loss in 6 months

or better, virtually all the safety issues have

been seen by then, and when we have a drug on the

market--fenfluramine--for 20-some years and we

don't pick up that it has a problem, it is just a

crap shoot. Why not go ahead and allow the drug

companies to cut those massive costs of research to

get the drugs on the market earlier, and then have

a much more rigorous long-term safety evaluation as

the drugs are on the market and they can begin to

recover some of their costs?

This extended run-in period--I see very

little usefulness for the run-in period. I am

raising some of the questions that were brought up

at the discussions that we have had, four

discussions so far. One of the things that most

people feel is really pretty useless is a run-in

period. In one of the original guidances people

were supposed to show weight loss and only those

people who showed weight loss would then be allowed

to go on to the clinical trial. That makes no

sense at all. We know that long-term diet and

125

exercise don't work. The question is do drugs work

long term? Trying to get people to change their

behavior is very, very difficult. If you can

change their biochemistry maybe we can get

somewhere.

Frank Greenway showed a trial on mazindol

with and without behavior modification, and when

you throw in behavior modification you reduce the

apparent efficacy of the drug but you can only lose

weight so fast. If you starve yourself you can

only lose weight so fast. So, as you have a better

diet and exercise program you wash out the effect

of the drug. Why have a run-in period at all?

Another thing that I think would be

useful--I was quite interested in Eric's comments

about what used to be the acceptable standard, any

statistically significant difference from placebo.

Drugs almost certainly will have to be used in

combination. Unfortunately, sibutramine and

orlistat don't work in combination but phentermine

and fenfluramine did. I, and I know others in this

room, have used phentermine and topiramate together

126

and appear to get a little better weight loss than

with either one alone. That has not been studied

in any organized fashion. So, if safe, these drugs

not only cause modest weight loss, many hold

promise that they could be used in combination with

other drugs and I would love to see that in the

guidances somehow. Varied indications for use are

justified. Others have shown that rapid weight

loss initially is associated with better response

of blood pressure, blood sugar. Jim Anderson has

done two meta-analyses showing that rapid weight

loss early, no matter what the time period, no

matter what the outcome measure--the people who

have lost a lot initially have at least as good or

better outcome variables. So, maybe drugs that

only cause a short-term weight loss might be useful

and then you switch to something else.

So, I think there are many varied

indications for use. Some for short term; some for

long term; some for use after a very low calory

diet, and perhaps the committee could consider some

of those indications.

127

One of the things that David Orloff and I

spent some time talking about on the phone is how

desperate the American public for drugs to treat

obesity. So, I think rational expectations for the

media and for patients--current drugs are only

modestly effective. Drugs in the pipeline appear

to be similar in terms of efficacy. The maximum

weight loss that I have heard is about 17 percent.

When companies over-hype the weight loss

or try to convince people that a 5 percent or 10

percent weight loss is wonderful, it is not. So, I

think over-hyping is bad on the part of the drug

companies. On the other hand, over-caution by the

FDA and scientists is detrimental. Hundreds of

thousands of people are dying of obesity-related

causes. Some drugs cause some problems. We have

to be safe but there is that tradeoff and Eric's

balance at the end I thought was particularly

appropriate. The media has not always been

responsible. In fact, I would say the media has

been predominantly irresponsible. I still remember

the Redux revolution--the cover on Time magazine,

128

this is going to solve all your problems, America.

The general public is desperate. They need

perspective and understanding of obesity as a

disease. Physicians have not really given them

that perspective. Unfortunately, I think long-term

lifestyle changes are needed. Trying to change

behavior is very difficult but that is what we are

stuck we right now.

For the future, I believe drugs of the

future of obesity treatment has the offer of

virtually every other chronic disease. Obesity is

due to biochemical differences. Drugs change

biochemistry. And, why am I so optimistic? Frank

Greenway showed you the data on obesity surgery is

somewhere between 25-40 percent of anethole body

weight. Surgery doesn't work because it makes a

little gastric pouch. It works because it changes

the biochemistry of the body. There are starting

to be lots of papers on changes in metabolic rate

and multiple different hormones. And, if surgery

can do it I have no doubt that the smart people at

the drug companies are going to figure out how they

129

can reproduce that kind of weight loss with one

drug or combinations of drugs. The surgery changes

biochemistry.

At least 350 drugs are in the pipeline, I

understand, and I think that bodes very well for

the future. Combination treatment I think is going

to be necessary, and I think the future is

extremely bright.

So, I will just end up by showing the

slide for the American Obesity Association. It is

a lay advocacy group. Its mission is to improve

the quality of life of obese people. I guess you

got copies of these slides but this is my contact

information, here. Thank you very much.

DR. BRAUNSTEIN: Thank you, Richard.

Questions from the panel? Let me start off with

one. You mentioned that fenfluramine had been on

the market for some time before the valvulopathy

was uncovered. If we look at the previous

speaker's slides on the use of fenfluramine, it

really didn't pick up greatly until the Weintraub

papers had come out. So, I wonder if what we are

130

talking about in terms of safety is a numbers game;

if you really do need a large number of patients to

pick up some of these potentially disastrous

complications. I would like your thoughts on that.

DR. ATKINSON: Yes, I notice there was

something like 70,000 prescriptions of fenfluramine

per year over a long period of time. That went up

to several million later. But Weintraub's paper

came out in 1992. By 1993 those numbers were going

up dramatically and it still took until 1997 before

it was identified, and there were millions of

people taking it obviously, and fenfluramine and

dexfenfluramine had been used in Europe.

Obviously, dexfenfluramine had been approved 10

years earlier. So, it is not just here. It was

all over the world that it was being used and it

wasn't picked up.

So, you know, I think drugs are going to

have consequences and obviously we need to look

very carefully at the drugs and study them, but I

would argue for shorter initial trials and more

intensive longer-term trials. I noticed in one

131

slide that the FDA was not charged with showing the

safety of drugs after they have been approved.

That probably ought to change.

DR. S. YANOVSKI: I would just like to

comment on your excellent question about

dexfenfluramine and why it hadn't been picked up

earlier with the fenfluramine. Before the

Weintraub papers came out these drugs were used

only exclusively short term, often 30 days or less

and it was never more than 90 days. It was only

after the Weintraub paper came out that they

started getting used for months and months and, in

some cases, even years. Since there was a length

of treatment response relationship with the

valvulopathy, that is likely why it wasn't seen

earlier.

DR. ATKINSON: Yes, that is an interesting

point, however, there were a number of people that

were reported that had valvulopathy who used it for

a relatively short period of time. It is probably

an idiosyncratic reaction.

DR. BRAUNSTEIN: Dr. Woolf?

132

DR. WOOLF: I am unclear. Are you

proposing that drugs for the treatment of obesity

be held to a different standard in terms of

evaluation of efficacy and safety than drugs in

general? At least the drugs that we discussed in

this committee before have had trials longer than 6

months, and certainly the clinical trials that I

participated in have been longer than 6 months.

So, are you proposing a different standard for

obesity drugs, or that the FDA change its modus

operandi?

DR. ATKINSON: I couldn't hear that very

well, but what I heard is am I proposing different

standards for obesity drugs? I think there are

different standards for different drugs. For

example, drugs for Alzheimer's disease, drugs for

AIDS, after relatively limited safety and efficacy

evaluations, are allowed to go on the market. The

point I am making is we have an epidemic of obesity

and a third of the population is affected. I think

it is not unreasonable to say how can we improve

the delivery of drugs, new and better drugs and

133

more drugs so we can try some of those

combinations? No, I don't want to have different

standards but I think there are different standards

for drugs and I would like to put obesity with sort

of the ones that get handled expediently.

DR. BRAUNSTEIN: Dr. Schade?

DR. SCHADE: I have a question about

weight loss. If one assumes that the drugs overall

result in, let's say--I am going to be

optimistic--10 percent weight loss, if you look at

the mortality or morbidity curve, if somebody has a

BMI of 35 and they lose 10 percent of the weight so

they drop to a BMI of 32, is their mortality then

exactly the same as a group that doesn't lose

weight but has a BMI of 32?

DR. ATKINSON: Yes, that is a very good

question. I don't know the answer to that.

Katherine Flegal's data were mainly focused on the

lower BMI groups. As you start up, when you start

getting to 30 and above, those curves start going

up fairly dramatically I think, if that is right,

Katherine. But I can't tell you that if you have

134

lost 2, 3 or 5 BMI units, do you then assume the

mortality and the morbidity of people who have

never been above that. I just don't know that.

DR. SCHADE: Well, the reason I ask that

is when we treat diabetes we treat hemoglobin A1C

and we assume, through our treatment, that we then

reduce the hemoglobin A1C and we can plot on the

curve the benefit. I just wondered whether the

curve for obesity is similar.

DR. ATKINSON: Yes, and I think that is

good. As you heard, there are some trials that are

ongoing to try to look at these sorts of things.

Again, this is a disease that

affects--what?--100-some million people in the U.S.

and we know almost nothing about it.

DR. BRAUNSTEIN: Dr. Aronne?

DR. ARONNE: Can I comment on the last

question? I think that the benefit from small

amounts of weight loss is disproportionate to the

amount of weight lost because of the initial loss

of visceral fat. When you look at the composition

of weight that is initially lost, it is the

135

riskiest fat that is lost first so small amounts of

weight loss appear to have disproportionate

benefit, out of proportion of what you would expect

from that small amount. What some people have

suggested is following something like the

C-reactive protein and that in the future that

could turn out to be our version of the hemoglobin

A1C because it is a measure of the inflammatory

burden of fat, and a lot of people believe that

visceral fat is where a lot of the C-reactive

protein is coming from.

DR. BRAUNSTEIN: Dr. Orloff?

DR. ORLOFF: Can you reiterate your

position on the run-in aspect of trial designs, and

specifically address whether you are proposing that

all run-ins of any duration, of any type, be

dispensed with?

DR. ATKINSON: No, certainly not. I think

a run-in period in the trials that I have designed

and gone out and done, investigator initiated type

clinical trials, we have put in a 2-week run-in

period that was not a treatment period but we

136

stretched out the initial evaluation. What that

does is get out the people who are not serious, who

don't want to come or it is too difficult to come,

or whatever. But in terms of requiring a weight

loss or requiring people to show that they can

adhere to a diet before they are allowed to go on

drugs, that is not true for other kinds of

diseases, for example, diabetes and hypertension,

and there may be companies that would want to do

that and would want to have a run-in, or that would

be what they think their drug is going to be useful

for--in other words, get the weight off and then

this is going to be their weight maintenance drug.

Fine, they can have a run-in. But I think the

mandate that all companies have to have an extended

run-in I don't agree with.

DR. ORLOFF: Again, a bit more

clarification. Do you still advocate diet and

exercise and continued reinforcement of those

lifestyle aspects for treatment of obesity in the

context of the trial?

DR. ATKINSON: Yes, I guess it was fairly

137

dramatically shown here. Dr. Greenway showed the

difference in weight losses that are achieved in

the United States versus over in Europe. The

companies design the trial to get well over

whatever the standards are. So, I think that can

be manipulated.

I make the statement often that everybody,

whether they are skinny or fat, needs to have a

good diet and lifestyle. I think as people come in

they need to be informed of what is a healthy

lifestyle and the exercise and the vegetables, and

all those things. Again, I am speaking for myself

not for anybody else, but I think the idea of

allowing the drug companies individually to figure

out where in the spectrum they want to put that is

not unreasonable, but at least some lip service

ought to be given, if for no other reason, because

people will do things very differently. I mean,