FOOD AND DRUG ADMINISTRATION



















                         ADVISORY SUBCOMMITTEE
















                      Wednesday, September 8, 2004


                               8:00 a.m.



                              Holiday Inn

                          Versailles Ballrooms

                         8120 Wisconsin Avenue

                           Bethesda, Maryland





         Glenn Braunstein, M.D., Acting Chair

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

         Executive Secretary




         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




                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




                     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




                   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




                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




                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




                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


      Register, with an open comment period until late


      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




                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




                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




                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




                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




                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




                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




                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




                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




                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




                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.




                [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




                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




                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




                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




                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






                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




                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




                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




                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




                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




                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






                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




                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




                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




                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




                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




                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




                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






                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




                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




                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




                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




      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




      patients are covered on?  Is that why cash is so




                DR. GOVERNALE:  I think if I heard your


      question, it is why are most of these products not




                DR. WOOLF:  Yes, is the reason that cash


      accounts for three-quarters of the method of


      payment because they are being excluded from drug




                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




                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




      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




                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




                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




      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




      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




                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




                DR. GOVERNALE:  We did not look into the


      reasons for why some of these prescription drug


      products are being covered or not covered by




      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




                [Brief recess]




                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




      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




      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




                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




      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




      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




                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




      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




      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




      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




      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




      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




      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




      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




                Other barriers to obesity drugs are




      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.




                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




      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




      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




                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




                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




      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




                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




      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




                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




      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




      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




                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




      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.




                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,




      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




      can reproduce that kind of weight loss with one


      drug or combinations of drugs.  The surgery changes




                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




      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




      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




                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?




                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




                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




      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




      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




      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




      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




      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,