FOOD AND DRUG ADMINISTRATION

































                         Friday, March 11, 2005


                               8:00 a.m.








                             Salons A and B

                Hilton Washington DC North/Gaithersberg

                           620 Perry Parkway

                         Gaithersburg, Maryland



                        P A R T I C I P A N T S


      Janet A. Englund, M.D., Chair


      Anuja M. Patel, M.P.H., Executive Secretary


      Committee Members:


      John A. Bartlett, M.D.

      Victor G. DeGruttola, Sc.D.

      Douglas G. Fish, M.D.

      John G. Gerber, M.D.

      Richard H. Haubrich, M.D.

      Victoria A. Johnson, M.D.

      Robert J. Munk, Ph.D. (Consumer Representative)

      Lynn A. Paxton, M.D., M.P.H.

      Kenneth E. Sherman, M.D., Ph.D.

      Eugene Sun, M.D. (Industry Representative)

      Maribel Rodriguez-Torres, M.D.

      Lauren V. Wood, M.D.

      Ronald G. Washburn, M.D.


      Special Government Employee Consultants (Voting):


      Samuel K. So, M.D., B.S.

      Kathleen Schwarz, M.D.


      Government Employee Consultants (Voting):


      Beth P. Bell, M.D., M.P.H.

      Ronald Herbert, D.V.M., Ph.D.

      Leonard B. Seeff, M.D.


      SGE Patient Representative (Voting)


      Brett Grodeck


      FDA Participants:


      Mark J. Goldberger, M.D., M.P.H., CDER

      Debra B. Birnkrant, M.D., CDER

      Linda L. Lewis, M.D., CDER

      James G. Farrelly, Ph.D., CDER



                            C O N T E N T S


      Call to Order and Opening Remarks,

         Janet Englund, M.D., Chair                              4


      Conflict of Interest Statement, Anuja Patel, M.P.H.

         Executive Secretary, FDA                                7


      Overview of Issues, Debra B. Birnkrant, M.D.,

          Director, DAVDP                                       10


      Sponsor Presentation:


      Introduction, Elliott Sigal, M.D., Ph.D.                  16


      Background, Richard Colonno, Ph.D.                        20


      Nonclinical Safety, Lois Lehman-McKeeman, Ph.D.           28


      Clinical Efficacy and Safety, Evren Atillasoy, M.D.       37


      Resistance, Richard Colonno, Ph.D.                        58


      Pharmacovigilance and Summary, Donna Morgan Murray,

      Ph.D.                                                     70


      Questions from the Committee                              77


      FDA Presentation:


      Carcinogenicity Issues, James G. Farrelly, Ph.D..        108


      Clinical Issues, Linda L. Lewis, M.D.                    119


      Discussion                                               151


      Advisory Committee Discussion of Questions

                Question 1:                                    185

                Question 2:                                    202

                Question 3:                                    204

                Question 4:                                    221

                Question 5:                                    235

                Question 6:                                    267




  1                      P R O C E E D I N G S


  2                Call to Order and Opening Remarks


  3             DR. ENGLUND:  Good morning.  Welcome,


  4   everyone.  My name is Janet Englund.  I am the


  5   acting chairperson today and I would like to


  6   welcome you to the Antiviral Drugs Advisory


  7   Committee.


  8             Today we are going to discuss the new drug


  9   application 21-797 and 21-798 for entecavir tablets


 10   and entecavir oral solution, respectively, by


 11   Bristol-Myers Squibb Company.  These drugs are


 12   proposed for the treatment of patients with chronic


 13   hepatitis B infection.


 14             With that, I would like to call the


 15   meeting to order and introduce the committee


 16   members.  In fact, I will have you introduce


 17   yourselves because that would be better.  I would


 18   like to just remind everyone on this committee that


 19   this is being transcribed and so, before you speak,


 20   you are going to need to identify yourself but, for


 21   now, if we could just start maybe with Dr. Sun and


 22   just introduce yourself and your affiliation.


 23             DR. SUN:  Eugene Sun, Abbott Laboratories.


 24             DR. GERBER:  John Gerber, University of


 25   Colorado Health Sciences Center.




  1             DR. WASHBURN:  Ron Washburn, Shreveport VA


  2   and LSU.


  3             DR. FISH:  Douglas Fish, Albany Medical


  4   College, Albany, New York.


  5             DR. HERBERT:  Ron Herbert, National


  6   Institutes of Environmental Health Sciences and the


  7   National Toxicology Program.


  8             DR. SHERMAN:  Ken Sherman, University of


  9   Cincinnati.


 10             DR. JOHNSON:  Victoria Johnson, University


 11   of Alabama at Birmingham.


 12             DR. PAXTON:  Lynn Paxton, Centers for


 13   Disease Control and Prevention.


 14             DR. WOOD:  Lauren Wood, National Cancer


 15   Institute.


 16             MR. GRODECK:  Brett Grodeck, patient


 17   representative.


 18             MS. PATEL:  Anuja Patel, Executive


 19   secretary for the Antiviral Drugs Advisory




  1   Committee, the Food and Drug Administration.


  2             DR. ENGLUND:  I am Janet Englund, from


  3   Children's Hospital and University of Washington,


  4   in Seattle.


  5             DR. DEGRUTTOLA:  Victor DeGruttola,


  6   Harvard School of Public Health.


  7             DR. BARTLETT:  I am John A. Bartlett, from


  8   Duke University.


  9             DR. HAUBRICH:  Richard Haubrich,


 10   University of California in San Diego.


 11             DR. MUNK:  Bob Munk, consumer


 12   representative.


 13             DR. SEEFF:  Leonard Seeff, Liver Disease


 14   Branch, NIDDK, National Institutes of Health.


 15             DR. BELL:  Beth Bell, Centers for Disease


 16   Control and Prevention.


 17             DR. SCHWARZ:  Kathy Schwarz, Johns Hopkins


 18   University.


 19             DR. FARRELLY:  Jim Farrelly, Division of


 20   Antiviral Drugs, FDA.


 21             DR. LEWIS:  Linda Lewis, Division of


 22   Antiviral Drugs, FDA.


 23             DR. BIRNKRANT:  Debbie Birnkrant, Division


 24   Director, Division of Antiviral Drugs, Food and


 25   Drug Administration.




  1             DR. ENGLUND:  And Dr. Mark Goldberger,


  2   from the FDA, will be joining us momentarily.  At


  3   this point I would like to have Anuja Patel read


  4   for us the conflict of interest statement.


  5                  Conflict of Interest Statement


  6             MS. PATEL:  Thank you.  The following


  7   announcement addresses the issue of conflict of


  8   interest and is made part of the record to preclude


  9   even the appearance of such at this meeting.  Based


 10   on the submitted agenda and all financial interests


 11   reported by the committee participants, it has been


 12   determined that all interests in firms regulated by


 13   the Center for Drug Evaluation and Research present


 14   no potential for an appearance of a conflict of


 15   interest, with the following exceptions:


 16             In accordance with 18 USC Section


 17   208(b)(3), full waivers have been granted to the


 18   following participants, Dr. Johnson for her


 19   employer's contract with a federal agency to




  1   provide virology laboratory support for the adult


  2   AIDS clinical trials group.  The contract is funded


  3   for greater than $300,000 per year.  Dr. Gerber for


  4   consulting on unrelated matters for the sponsor and


  5   a competitor.  He receives less than $10,001 per


  6   year per firm.  Dr. Bartlett for serving on


  7   speakers bureaus for two competitors.  He receives


  8   greater than $10,000 from one firm and between


  9   $5,001 to $10,000 per year from the other.  Dr.


 10   Sherman for serving on speakers bureaus for two


 11   competitors.  He receives from $5,001 to $10,000 a


 12   year from each firm.  Dr. Munk for consulting on


 13   unrelated matters for a competitor.  He receives


 14   less than $10,001 a year.


 15             Dr. Schwarz has been granted waivers under


 16   (b)(3) and 21 USC 355(n)(4) for her employer's


 17   grant to study competing products.  Each grant is


 18   funded for less than $100,000 per firm per year.


 19   Dr. Haubrich has been granted a (b)(3) waiver for


 20   consulting on unrelated matters for a competitor


 21   and the sponsor.  He receives less than $10,001 per


 22   year per firm.  Brett Grodeck has been granted a




  1   355(n)(4) waiver for owning stock in a competitor,


  2   valued at less than $5,001.  Because the stock in a


  3   competitor does not exceed $25,000, 5 CFR


  4   2640.202(a)(2) exception applies and a (b)(3)


  5   wavier is not required.  Dr. DeGruttola has been


  6   granted a (b)(3) waiver for consulting on unrelated


  7   matters for two competitors.  He receives less than


  8   $10,001 a year from each firm.


  9             A copy of the waiver statements may be


 10   obtained by submitting a written request to the


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


 12   of the Parklawn Building.


 13             In the event that the discussions involve


 14   any other products or firms not already on the


 15   agenda for which an FDA participant has a financial


 16   interest, the participants are aware of the need to


 17   exclude themselves from such involvement and their


 18   exclusion will be noted for the record.


 19             We would also like to note that Dr. Sun


 20   has been invited to participate as an industry


 21   representative, acting on behalf of the regulated


 22   industry.  Dr. Sun is employed by Abbott




  1   Laboratories.


  2             With respect to all other participants, we


  3   ask in the interest of fairness that they address


  4   any current or previous financial involvement with


  5   any farm whose products they may wish to comment


  6   upon.  Thank you.


  7             DR. ENGLUND:  Thank you, everyone.  With


  8   that done, I would like to introduce Dr. Debra


  9   Birnkrant who will now proceed to give us an


 10   overview of the issues and our plan for today.


 11                        Overview of Issues


 12             DR. BIRNKRANT:  Good morning.  I would


 13   also like to welcome our advisory committee members


 14   and consultants to this meeting.


 15             Today, as was mentioned, we will be


 16   discussing the new drug application for the tablet


 17   and solution formulations for entecavir for the


 18   treatment of chronic hepatitis B infection.


 19             The last time this committee met to


 20   discuss a similar topic was back in 2002 when we


 21   presented the new drug application for adefovir,


 22   and on the second day of that meeting we discussed




  1   general drug development for hepatitis B.  Today's


  2   meeting gives us another opportunity to discuss


  3   this serious problem.


  4             The next two slides were downloaded from


  5   cdc.gov.  This slide shows the geographic


  6   distribution of chronic hepatitis B infection.


  7   What you can see in red are high andemic areas in


  8   Africa and Asia with hepatitis B prevalence at a


  9   rate more than 8 percent, and this is considered


 10   high.  In gold we have medium prevalence areas, and


 11   in green we have low prevalence areas, such as the


 12   United States, excluding Alaska.  In the high


 13   prevalence areas the lifetime risk of acquiring


 14   hepatitis B infection approaches 60 percent and is


 15   acquired mainly during childhood, whereas in the


 16   low prevalence areas the lifetime risk is much


 17   lower and occurs in adolescents, adults and


 18   well-defined risk groups.


 19             This slide shows hepatitis B incidence by


 20   year through the years 1966 through 2000 in the


 21   United States.  What this is dramatic for is the


 22   decline in hepatitis B occurring soon after




  1   licensure of hepatitis B vaccine.  You can see that


  2   the incidence drops dramatically over the years in


  3   the late '80s and beyond after public health


  4   programs adopted hepatitis B vaccination.


  5             Although we see this dramatic decrease in


  6   the United States of acute hepatitis B it still


  7   remains a major problem.  It has been estimated


  8   that chronic hepatitis B infection affects 350-400


  9   million subjects worldwide and approximately 1.25


 10   million subjects in the United States.  It accounts


 11   for, it is estimated, approximately one million


 12   deaths per year due to complications of the


 13   disease, namely cirrhosis and hepatocellular


 14   carcinoma.  The treatment options are quite


 15   limited.  As you can see, there are only three at


 16   this point, interferon, lamivudine and adefovir


 17   dipivoxil.


 18             I will briefly touch on the pros and cons


 19   of these therapies.  Interferon is used in a


 20   limited patient population, however, it is used for


 21   a definite period of time and in the limited


 22   population the effect is durable.  However, the




  1   side effect profile is somewhat limiting.  With


  2   interferon we see flu-like syndrome, depression,


  3   alopecia and exacerbation of autoimmune disorders.


  4             Lamivudine, a nucleoside analog, is much


  5   better tolerated, however, subjects taking


  6   lamivudine develop resistance at a rate approaching


  7   20 percent per year.


  8             Adefovir dipivoxil, a prodrug of adefovir,


  9   a nucleotide analog, was approved in 2002.  It is


 10   active against lamivudine-resistant virus, and is


 11   tolerated well except for nephrotoxicity that


 12   appears in decompensated patients, more so, and


 13   other advanced patients such as those undergoing


 14   transplant.


 15             Let's turn now to today's subject, that


 16   is, entecavir.  Entecavir is also a nucleoside


 17   analog.  It has activity against HBV polymerase,


 18   and in vitro it inhibits lamivudine-resistant virus


 19   at concentrations 8-32-fold greater than that


 20   required for wild type virus.


 21             Its antiviral activity has been


 22   demonstrated in established animal models.  In




  1   woodchuck, hepatitis virus infected woodchucks with


  2   that disease, 67 percent treated with entecavir


  3   survived 3 years compared to a 4 percent survival


  4   rate in infected historic controls.  So, it appears


  5   quite active in this established animal model.


  6             Now I will describe pertinent nonclinical


  7   pharm/tox findings briefly.  There was an increased


  8   incidence of tumors in rodent carcinogenicity


  9   studies.  Lung tumors were observed at low


 10   multiples of entecavir exposure relative to humans


 11   and it is thought that these tumors may be species


 12   specific.  Other tumors occurred at much higher


 13   multiples of entecavir exposure relative to humans.


 14   This topic will be discussed extensively by


 15   Bristol-Myers Squibb and Dr. Farrelly of the Food


 16   and Drug Administration.  What we have to keep in


 17   mind here is that the animal data needs to be


 18   interpreted in the context of the clinical data,


 19   the severity of the disease and the available


 20   treatment options.       Turning to the clinical


 21   studies, I would like to commend Bristol-Myers


 22   Squibb for their drug development program for




  1   entecavir.  They studied a wide population in


  2   e-antigen positive, e-antigen negative and


  3   lamivudine-resistant subjects.  Their trials were


  4   multicenter and multinational, using an active


  5   control, lamivudine.  The endpoints used were


  6   similar to other approved therapies.


  7             At today's advisory committee meeting we


  8   will be asking you to discuss the clinical trial


  9   data in the context of these animal carcinogenicity


 10   findings and the implications for human use.  In


 11   addition, we will be asking you to discuss the


 12   adequacy of the proposed pharmacovigilance study.


 13   We will also pose a question related to pediatric


 14   usage.


 15             If in the afternoon session when questions


 16   are posed you vote that this drug should be


 17   approved, we will then proceed to discuss labeling


 18   implications and further post-marketing studies.


 19             With that, I would like to just briefly


 20   review the agenda.  Following my comments,


 21   Bristol-Myers Squibb will present.  This will be


 22   followed by a break.  Then FDA will present and the




  1   presentations will be discussed prior to lunch.  At


  2   one o'clock there is an open public hearing.


  3   Following that hearing, we will continue the


  4   discussion and then pose our questions to the


  5   advisory committee.  Thank you very much.


  6             DR. ENGLUND:  Thank you very much.  Now I


  7   think we would like to begin with the sponsor


  8   presentation by Bristol-Myers Squibb.


  9                       Sponsor Presentation


 10                           Introduction


 11             DR. SIGAL:  Thank you, Dr. Englund and


 12   members of the committee and FDA.  Good morning.  I


 13   am Elliott Sigal.  I am head of research and


 14   development and chief scientific officer for


 15   Bristol-Myers Squibb.  Today it is our pleasure to


 16   bring you data on entecavir for the treatment of


 17   patients with chronic hepatitis B infection.


 18             As you heard from Dr. Birnkrant, this


 19   disease affects well over actually a million people


 20   in the United States and accounts for approximately


 21   5,000 deaths here a year.  Outside the United


 22   States another 400 million people are chronically




  1   infected with hepatitis B so it represents a


  2   worldwide public health issue of great importance.


  3             We, at Bristol-Myers Squibb, have


  4   concluded, based on the data you will hear today,


  5   that entecavir represents an important therapeutic


  6   advance.  Our application is being considered first


  7   here, in the U.S., but we have filed in Europe and


  8   in China, and intend to file elsewhere around the


  9   world as part of a larger global commitment.


 10             All new therapies present a need to assess


 11   both benefits and risks.  Years ago, knowing this


 12   compound to be a nucleoside analog, we


 13   intentionally completed and analyzed rodent


 14   carcinogenicity studies before initiating a Phase


 15   III program.  Then we continued to explore the


 16   mechanisms of these rodent findings and we


 17   collaborated with health authorities around the


 18   world on how to characterize clinical benefit.  The


 19   goal has been to determine benefits seen in the


 20   clinic and weigh those against the potential for


 21   risk raised by nonclinical studies.


 22             Entecavir has clinical benefits based on




  1   its antiviral potency and these are superior


  2   suppression of viral replication; a favorable


  3   resistance profile; and improvement in both liver


  4   histology and in biochemical abnormalities.  To


  5   establish all of this we conducted an extensive


  6   Phase III program, the first in this field with an


  7   active comparator.  As the sponsor, we concluded


  8   that the benefits in the clinic, including the


  9   resistance profile, outweigh the potential seen of


 10   risk in nonclinical studies and entecavir, to us,


 11   represents an important therapeutic option for


 12   patients with chronic hepatitis B infection.


 13             However, as with any new medicine, an


 14   assessment of benefit-risk at the time of approval


 15   can only be an estimate.  Therefore, our company is


 16   committed to further defining therapeutic benefits


 17   and to understanding any potential human risk with


 18   entecavir.


 19             To accomplish this we have submitted to


 20   FDA draft pharmacovigilance plans, approaches and


 21   observational studies that we plan to conduct to


 22   allow for a continuous benefit-risk assessment once




  1   entecavir is available for patients.  For the


  2   medical community these studies will advance the


  3   overall scientific knowledge about this disease.


  4   Bristol-Myers Squibb has a history of antiviral


  5   clinical research in the treatment of patients with


  6   HIV infection.  Now with entecavir we are expanding


  7   that commitment to advance the medical science of


  8   chronic hepatitis B infection.


  9             Furthermore, let me say that our efforts


 10   in the marketplace will be directed to ensure the


 11   appropriate use of this new medicine.  We will


 12   create a U.S. field organization solely dedicated


 13   to entecavir.  It will combine medical


 14   professionals and representatives who will be


 15   specifically trained in chronic hepatitis B.  Their


 16   focus will be on a relatively small number of


 17   physicians, 3,500, that provide care for nearly all


 18   the U.S. patients treated for chronic hepatitis B.


 19   This focused approach will ensure high quality


 20   interaction with prescribing physicians and


 21   appropriate use of entecavir for patients.


 22             Dr. Rich Colonno will now begin the data




  1   presentation.  Dr. Englund, two of our speakers


  2   fell ill over the last 36 hours so you will see a


  3   few different names on the program.  One of our


  4   internal hepatologists, Dr. Atillasoy, will be the


  5   one presenting our clinical data.  Dr. Colonno?


  6                            Background


  7             DR. COLONNO:  Good morning.  Sorry for the


  8   confusion.  Entecavir is under review for the


  9   proposed indication shown here, the treatment of


 10   chronic hepatitis B disease in adults with evidence


 11   of liver inflammation.  The usual dose will be 0.5


 12   mg daily and a higher 1.0 mg dose is proposed for


 13   patients who are lamivudine-refractory.


 14             Our presentation will follow the outline


 15   shown on this slide, covering nonclinical safety,


 16   clinical efficacy, clinical safety, resistance and


 17   pharmacovigilance.  We have been assisted in


 18   evaluating our data by a number of experts who are


 19   listed on the next slide.  These consultants,


 20   covering hepatology, health policy, toxicology,


 21   pathology and biostatistics, are here and available


 22   to the committee.


 23             Dr. Birnkrant and Dr. Sigal outlined the


 24   disease burden and consequences of chronic HBV


 25   infection.  Only about 10-30 percent of people




  1   currently affected with HBV go on to develop a


  2   chronic infection.  But the millions who do, it is


  3   sometimes decade-long process that for a


  4   substantial number of patients ends with cirrhosis,


  5   liver failure, hepatocellular carcinoma, transplant


  6   or death.


  7             This is a viral disease and the clinical


  8   course of liver injury is driven by the continuous


  9   replication of the virus perpetuating a cycle of


 10   inflammation.  HBV is not inherently cytopathic but


 11   liver cells support a continuous cycle of viral


 12   replication that triggers the inflammatory response


 13   that over time leads to fibrosis, cirrhosis and


 14   liver cancer.  HBV has recently been designated a


 15   carcinogen, in recognition that HBV-induced


 16   hepatocellular carcinoma is the fifth most frequent


 17   single type of cancer.


 18             It has now been shown that the outcome of


 19   this long course of chronic infection with HBV is




  1   not just caused by the initial infection but is


  2   related to the degree of continued viral


  3   replication.  This was supported by a prospective


  4   Taiwan cohort study in which three key points


  5   emerged:  The incidence of hepatocellular carcinoma


  6   and liver cirrhosis correlated with baseline HBV


  7   DNA levels.  The higher the baseline, the higher


  8   the incidence.  Two, persisting elevation of the


  9   viral load over time has the greatest impact on


 10   hepatocellular carcinoma risk.  Viral load


 11   predicted risk of future hepatocellular carcinoma


 12   independent of e-antigen status and serum ALT


 13   levels.


 14             The concept that viral replication drives


 15   disease process is depicted in the schematic shown


 16   on this slide.  Viral replication, monitored by


 17   serum HBV DNA levels, drives the downstream


 18   inflammation, measured by ALT levels and by


 19   histology assessments.  These were our week 48


 20   endpoints, and we will be referring to this


 21   simplified schematic later in our presentation.


 22             Currently, three drugs are approved to




  1   treat chronic hepatitis B infection, interferon,


  2   lamivudine and adefovir.  Interferon is an


  3   immunomodulator while adefovir and lamivudine are


  4   antivirals whose demonstrated antiviral activity


  5   led to their approval.  In their clinical studies


  6   both lamivudine and adefovir were shown to be


  7   superior to placebo using the endpoints of liver


  8   histology, viral suppression and ALT normalization


  9   at week 48.  They decreased viral load, the first


 10   stage of the schema, and interrupted the process


 11   measured by ALT and histology, in the center


 12   section.  Beyond the week 48 data points,


 13   lamivudine has now shown superiority to placebo in


 14   affecting some of the long-term outcomes seen in


 15   the far right-hand slide of the schema,


 16   characterized as disease progression.


 17             In the recent landmark paper by Liaw et


 18   al., lamivudine treatment was prospectively


 19   compared with placebo in patients with compensated


 20   cirrhoses who are at greatest risk for disease


 21   progression, including HCC and worsening cirrhosis.


 22   With lamivudine treatment by 32 months the rate of




  1   disease progression was significantly reduced


  2   relative to placebo, 8 percent versus 18 percent.


  3   This study confirmed the hypothesis that effective


  4   antiviral therapy results in a better long-term


  5   clinical outcome than indicated by the week 48


  6   histology, virology and ALT endpoints.


  7             The study also pointed out that a


  8   development of resistance to a particular antiviral


  9   therapy limits its benefit.  By the end of the


 10   study roughly half of the lamivudine-treated


 11   patients who had developed lamivudine resistance,


 12   or YMDD virus, and these patients had twice the


 13   percentage of disease progression when compared to


 14   those where the virus remained fully susceptible,


 15   11 percent versus 5 percent respectively.


 16             So, while lamivudine is effective and


 17   lacks the tolerability concerns of interferon and,


 18   unlike adefovir, does not require careful


 19   monitoring of renal function, resistance impacts


 20   the ability of lamivudine to deliver long-term


 21   benefits.  While the study confirmed that antiviral


 22   treatment provides benefit, it also suggested that




  1   a more effective antiviral with both greater


  2   potency and less resistance will be more


  3   efficacious in preventing downstream clinical


  4   disease.


  5             This morning you will see that entecavir,


  6   by the accepted and proven histologic, virologic


  7   and biochemical endpoints of our studies, was


  8   superior to lamivudine.  We will demonstrate that


  9   entecavir is effective, safe and well tolerated;


 10   has excellent potency and very low rates of


 11   resistance; and maintains future options because it


 12   doesn't select for lamivudine or adefovir


 13   resistance and is, therefore, an important advance


 14   in therapy for chronic HBV disease.


 15             The activity of entecavir results from its


 16   being a cyclopentyl guanosine analog.  It is a


 17   selective and potent inhibitor of HBV replication.


 18   It has no significant activity against HIV.  The


 19   selectivity contributes to its safety since it is a


 20   poor substrate for sailor DNA polymerases and does


 21   not inhibit human mitochondrial or gamma


 22   polymerase.  Its potency reflects the fact that it




  1   inhibits all three functional activities of the HBV


  2   polymerase, priming, DNA-dependent synthesis and


  3   reverse transcription.  It is also a function of a


  4   highly efficient conversion of entecavir to its


  5   active form entecavir triphosphate, seen


  6   consistently in a wide variety of cell types.


  7             Entecavir undergoes rapid and efficient


  8   phosphorylation by sailor enzymes at low


  9   concentrations, and can be detected within one


 10   hour.  Once formed, the intracellular half-life of


 11   entecavir triphosphate is approximately 15 hours.


 12   With an EC                                                50 of 4 nM it

is the most potent inhibitor


 13   of hepatitis B virus.  Entecavir is greater than


 14   300 times more potent than either of the available


 15   agents, lamivudine or adefovir, or two newer agents


 16   under development dibividine[?] and tenofovir.


 17             Animal models of HBV have been developed


 18   using woodchucks and ducklings and entecavir


 19   demonstrated impressive potency in these systems as


 20   well.  The woodchuck model is of particular


 21   importance because it has been predictive of the


 22   efficacy and safety of drugs subsequently used in




  1   humans to treat hepatitis B virus.  The antiviral


  2   susceptibility of the woodchuck hepatitis B virus,


  3   or WHBV, is similar to the human virus.  In this


  4   model greater than 95 percent of chronically


  5   infected animals will development HCC and die, and


  6   less than 5 percent will survive to age 4.


  7             In our study, animals standard established


  8   chronic infection were dosed with entecavir at 0.5


  9   mg/kg, a dose that results in exposure levels of


 10   approximating the exposure in humans with the 1 mg


 11   dose.  The drug was initially administered daily


 12   for 2 months and then weekly for a total of 14-36


 13   months.  In both groups entecavir treatment


 14   resulted in viral DNA levels being reduced by as


 15   much as 8 logs to undetectable levels.  The


 16   reductions were sustained for up to 3 years, with


 17   no evidence of virologic rebound or resistance.


 18             The study compared the improvement in


 19   survival versus historical controls, shown in grey.


 20   The 11 woodchucks, represented by the yellow bars,


 21   started treatment at 8 months of age as soon as a


 22   chronic infection was verified.  They had 4-year




  1   HCC-free survival of 50 percent and 80 percent


  2   respectively for the 14- and 36-month treatment


  3   groups.  The non-concurrent historical control had


  4   a survival rate of 4 percent.  Although the numbers


  5   of animals were small, these results were of high


  6   statistical significance.  Surviving animals were


  7   also shown to have no histological evidence of HCC


  8   development upon subsequent examination.


  9             In summary, the nonclinical data and the


 10   expected benefit of antiviral treatment supported


 11   going forward with development of entecavir for


 12   treatment of chronic HBV infection.  As with any


 13   drug being developed for long-term chronic dosing


 14   in humans, the carcinogenicity potential of


 15   entecavir was evaluated in lifelong dosing studies


 16   in rats and mice.  Dr. Lois Lehman-McKeeman will


 17   now present this data.


 18                        Nonclinical Safety


 19             DR. LEHMAN-MCKEEMAN:  Today's discussion


 20   of the nonclinical safety of entecavir is focused


 21   on the rodent carcinogenicity studies.  Entecavir


 22   was identified as a carcinogenic hazard in rats and




  1   mice, and the benefit-risk evaluation for entecavir


  2   must consider this risk identified in animals


  3   relevant to the human clinical benefit.


  4             For background on the rodent data, I will


  5   briefly describe the design, conduct and


  6   interpretation of these studies.  Rodent


  7   carcinogenicity studies are lifetime studies,


  8   typically 2 years, and group sizes are large with


  9   50-60 animals per sex per group.  Dose selection is


 10   critical, and highest dosage is expected to


 11   represent a maximum tolerated dose, or MTD.  The


 12   simplest definition of an MTD is a dose that causes


 13   no more than a 10 percent decrease in body weight


 14   gain relative to controls.  The lower dosages


 15   studied, typically 2 additional levels, are


 16   selected to be some fraction of the MTD or some


 17   multiple of the relevant human clinical exposure.


 18             At the end of the study all tissues are


 19   evaluated microscopically for tumors.  Several


 20   tissues in rats and mice are prone to spontaneous


 21   tumor development.  For example, in mice there was


 22   a relatively high background rate of tumors in




  1   liver and lung, while in rats liver, pituitary and


  2   mammary gland tumors occurred at high spontaneous


  3   rates.  So, finding tumors in animals, including


  4   controls, is not surprising and we rely on


  5   statistical methods and an understanding of


  6   historical control tumor rates to identify those


  7   that are drug related.


  8             Statistical significance in rodent tumors


  9   is established by sequentially testing for a linear


 10   dose-dependent trend starting with all dose groups.


 11   Tumor incidence is adjusted for survival and the


 12   time and cause of death and the level of


 13   statistical significance varies with whether a


 14   tumor is common or rare.  The more common the


 15   tumor, the more rigorous the statistical analysis.


 16   When the results identify a positive trend, data


 17   are reanalyzed by dropping the highest dose and


 18   repeating the test.  This cycle is repeated until


 19   no significant trend is observed.


 20             With that as an overview on rodent


 21   carcinogenicity studies, let's review the results


 22   for entecavir.  These results have been reviewed




  1   with the FDA's Executive Carcinogenicity Assessment


  2   Committee, or CAC, and the full CAC and a number of


  3   tumor sites were concluded to be relevant to human


  4   safety.


  5             Entecavir-induced tumors followed two


  6   distinct patterns.  The first pattern was observed


  7   in tissues that showed preneoplastic changes, that


  8   is, sites were early changes, consistent with the


  9   increased likelihood of tumor development, were


 10   observed.  The only site that showed this pattern


 11   was the mouse lung.


 12             The second pattern of increased tumors was


 13   in tissues that showed no evidence of preneoplastic


 14   changes and occurred at high exposure multiples


 15   relative to anticipated human exposure.  These


 16   tumors included liver carcinomas in male mice;


 17   vascular tumors in female mice; gliomas in male


 18   rats; and gliomas, liver adenomas and skin fibromas


 19   in female rats.


 20             In addition to listing the tumor sites,


 21   let's look at the incidences observed in these


 22   studies.  Entecavir was dosed to mice across a dose




  1   range of 0.004 mg/kg to 4 mg/kg.  To orient you to


  2   this slide, the dosages are shown in the top line


  3   and the exposure multiples are noted below the


  4   dosages representing the comparison of the plasma


  5   area under the curve in mice relative to human


  6   exposure at the 0.5 mg or 1 mg dose.  The exposures


  7   are presented as those in the males, followed by


  8   the females.  4 mg/kg was an MTD and this dose


  9   represented at least a 40-fold multiple over the


 10   human exposure at 1 mg.


 11             The mouse lung is a major target organ for


 12   tumor development following entecavir treatment.


 13   Lung tumors are common in mice.  There was a 12


 14   percent incidence of tumors in the control males in


 15   this study.


 16             Entecavir increased the incidence of lung


 17   adenomas with a statistical increase in tumors,


 18   here noted in yellow, observed at the 0.4 mg/kg


 19   dose in males.  This dose is 3-5 times higher than


 20   human clinical exposure.  Lung adenomas were


 21   further increased at the 2 higher dosages and at 4


 22   mg/kg entecavir increased the incidence of lung




  1   carcinomas.


  2             In female mice lung tumors occur at a


  3   higher spontaneous rate than in males, with a


  4   background incidence of 20 percent in this study.


  5   Entecavir increased pulmonary tumors in female mice


  6   but the statistical significance was noted only at


  7   the highest dose.


  8             Other toxicology studies indicated that


  9   entecavir elicited unique changes in the mouse


 10   lung, and we conducted experiments to define these


 11   changes and to determine whether they were linked


 12   to the increased susceptibility to tumor


 13   development.  The results showed preneoplastic


 14   changes in the mouse lung that consisted of


 15   increased numbers of macrophages and Type II


 16   pneumocyte hyperplasia.  Cell proliferation is a


 17   recognized risk factor for tumor development and


 18   entecavir caused a sustained proliferation of Type


 19   II pneumocytes.  Most mouse lung tumors arise from


 20   Type II pneumocytes and these cells were identified


 21   as the progenitor cells for entecavir-induced lung


 22   tumors as well.  The increased numbers of




  1   macrophages was required to support the


  2   proliferation of the Type II pneumocytes and


  3   entecavir increased the number of alveolar


  4   macrophages in the lung because it was chemotactic


  5   for mouse monocytes.


  6             In contrast to the mouse, no similar


  7   changes were observed in the lungs of rats, dogs or


  8   monkeys treated with entecavir.  Finally, although


  9   entecavir was chemotactic for mouse monocytes, it


 10   was not chemotactic for human monocytes, suggesting


 11   that an accumulation of macrophages in the human


 12   lung would be unlikely to occur.  The results


 13   suggest that entecavir causes unique effects in the


 14   mouse lung and lung tumors observed in mice may be


 15   species specific.


 16             The second presentation of entecavir-


 17   induced tumors in mice was in organs that, unlike


 18   the lung, showed no evidence of preneoplastic


 19   change.  In males entecavir increased the incidence


 20   of liver carcinomas and in females entecavir


 21   increased the incidence of vascular tumors,


 22   specifically hemangiomas.  In both cases there was




  1   no dose response relationship noted, with tumors


  2   observed only at the highest dosage.


  3             We have not explored mechanisms underlying


  4   the high dose tumor findings on an organ by organ


  5   basis, but we have looked at whether a common mode


  6   of action may contribute to tumor development.


  7   Entecavir is phosphorylated to entecavir


  8   triphosphate, the active form that inhibits viral


  9   replication, and we determined that, likely by


 10   competing for phosphorylation as depicted here,


 11   entecavir disrupts deoxynucleotide triphosphate


 12   pools, dNTP pools, in male mouse liver.  Data in


 13   the scientific literature demonstrates that such


 14   perturbations disrupt the fidelity of DNA synthesis


 15   and repair.  We conclude that changes in the dNTP


 16   pools may explain tumor findings, particularly when


 17   there is a high dose response for tumor


 18   development.


 19             Moving on to rats, in Sprague-Dawley rats


 20   entecavir was dosed to males at dosages up to 1.4


 21   mg/kg or to females at dosages up to 2.6 mg/kg.


 22   The 4 dosage levels are noted here along with the




  1   exposure multiples as were presented on the mouse


  2   slides relative to the 0.5 mg or 1 mg clinical


  3   dose.  Maximum exposures were at least 35 times


  4   human exposure in male rats or 24 times human


  5   exposure in female rats.  In rats all tumors


  6   observed were consistent with the second pattern of


  7   tumor presentation, that is, no evidence of


  8   development of preneoplastic change.


  9             In males and females entecavir increased


 10   the incidence of gliomas with statistical


 11   significance only at the highest dosage.  In


 12   females entecavir increased the incidence of liver


 13   adenomas and skin fibromas.  As determined in mice,


 14   we have postulated that the dNTP pool perturbations


 15   resulting from high doses of entecavir that


 16   overwhelm the strict regulation of nucleotide


 17   metabolism may explain entecavir-induced tumors in


 18   rats.


 19             Carcinogenicity studies in rodents


 20   identify whether a compound is a carcinogenic


 21   hazard.  In the absence of data in humans it is


 22   assumed that carcinogenic effects in rodents




  1   suggest a possible carcinogenic risk in humans.


  2   However, to extrapolate these findings to humans


  3   other relevant data, such as genetic toxicity and


  4   species differences in biological response, along


  5   with dose-response relationships and exposure


  6   comparisons, are important considerations that may


  7   increase or decrease the likelihood of human cancer


  8   risk.  For entecavir there is evidence suggesting a


  9   unique biological response in the mouse lung and


 10   mouse lung tumors may be species specific.


 11             Extrapolation of the other tumor findings


 12   is more difficult, but the weight of evidence


 13   suggests that human risk is minimal because rodent


 14   tumors were observed at dosages that greatly exceed


 15   human clinical exposure.


 16             Dr. Evren Atillasoy will now review the


 17   benefit of entecavir as determined from the Phase


 18   III clinical trials.


 19                   Clinical Efficacy and Safety


 20             DR. ATILLASOY:  Thank you and good


 21   morning.  The entecavir clinical development


 22   program is comprehensive and assesses the efficacy




  1   and safety of entecavir for the treatment of


  2   chronic hepatitis B infection.  The experience was


  3   broad with major disease patterns well represented.


  4   Studies addressed hepatitis B e-antigen positive


  5   patients and e-negative disease, and assessed


  6   entecavir in lamivudine-refractory as well as


  7   nucleoside-naive patients.


  8             The global program recruited patients from


  9   5 continents in over 30 countries.  Separate


 10   programs are in progress in China and Japan.  The


 11   studies that contribute to the NDA review provide


 12   analyzed data on approximately 1,500


 13   entecavir-treated patients.  Entecavir is the first


 14   nucleoside program to be evaluated for HBV using an


 15   active comparator, lamivudine, which was the only


 16   approved HBV nucleoside at the time that the


 17   program was initiated.


 18             The map of the clinical program


 19   illustrates the sense of the size, breadth and


 20   complexity.  The core of the program is represented


 21   by the green box and includes the three Phase III


 22   studies you will be hearing about today.  Small




  1   studies in special populations include experiences


  2   in liver transplant patients, co-infected


  3   HIV-positive patients and decompensated patients,


  4   the trial which we are still actively enrolling.


  5             Two long-term rollover studies provide for


  6   prolonged observation and data collection.  Study


  7   901, at the bottom left, provides an ongoing


  8   treatment option for those patients in whom


  9   long-term treatment is appropriate.  Study 049 is a


 10   post-treatment observational study, designed to


 11   collect long-term safety and efficacy information.


 12   All Phase III patients have the opportunity to


 13   enroll in these trials.  These data in 049 have not


 14   yet been analyzed.


 15             Dose selection for entecavir anticipated


 16   that lamivudine-refractory patients would require a


 17   higher dose than naive patients because of the


 18   higher EC                                              50 of

lamivudine-resistant virus in vitro.


 19   An earlier proof of principle study testing doses


 20   over a range from 0.5 mg to 1 mg daily hinge on


 21   overlapping responses for the highest doses of 0.5


 22   mg and 1 mg daily.  Therefore, these doses were




  1   used as the highest ones tested in dose selection


  2   studies, 0.5 mg in naive patients, in yellow on the


  3   left graph, and 1 mg refractory patients, in orange


  4   on the right graph.  The lamivudine control is


  5   represented in blue in both graphs.


  6             A dose response was demonstrated in each


  7   population, with the greatest responses occurring


  8   at the two highest doses with diminishing


  9   incremental benefit at the last increase.


 10   Entecavir 0.5 mg daily and 1 mg daily were taken


 11   forward as the doses to be tested for Phase III for


 12   naive and refractory patients respectively.


 13             Clinical efficacy--Phase III included


 14   trials in three disease settings, nucleoside-naive


 15   e-antigen positive patients, nucleoside e-antigen


 16   negative patients and lamivudine refractory


 17   e-antigen positive patients.  The definition of


 18   lamivudine refractory was that patients must have


 19   clinical failure after at least 6 months of


 20   lamivudine, or earlier failure with the


 21   confirmation of lamivudine-resistant virus.


 22   Clinical failure was defined as detectable viremia




  1   using the bDNA assay.  Today's presentation of


  2   clinical results will be by treatment population


  3   rather than study number.


  4             Lets turn to study design across Phase


  5   III.  Patients were screened and randomized 1:1 to


  6   either entecavir or lamivudine in a double-blind


  7   fashion and were treated for a minimum of 52 weeks.


  8   Lamivudine-refractory patients who were required to


  9   have breakthrough viremia while on lamivudine were


 10   switched on treatment day 1 directly from


 11   lamivudine to blinded study drug without a period


 12   either of overlap or washout.  Liver biopsies were


 13   obtained at baseline and at week 48 for assessment


 14   of the primary efficacy endpoint, histologic


 15   improvement.  Patient management at week 52 was


 16   based on lab results using data from the week 48


 17   visit, with results of the 24 follow-up period


 18   presented in the briefing document that you have.


 19             Inclusion criteria, let's talk about these


 20   for the three studies.  Inclusion criteria required


 21   that patients needed to have compensated liver


 22   disease, together with an elevated ALT, or were




  1   required to have detectable viremia by bDNA.  The


  2   different virologic characteristics of the


  3   e-antigen positive and e-antigen negative disease


  4   patients resulted in different minimal requirements


  5   for enrollment by HBV DNA.


  6             The baseline demographics of each study


  7   population are consistent with the characteristics


  8   expected for the patient population.  In the


  9   presentations that follow results for the naive


 10   e-antigen positive patients will appear on the left


 11   of the slide.  In the middle you will see data for


 12   the naive e-antigen negative patients and on the


 13   furthest right you will see results for the


 14   lamivudine-refractory e-antigen positive


 15   population.  Within each study the


 16   entecavir/lamivudine study groups were well matched


 17   for demographic characteristics.


 18             Turning to baseline HBV characteristics,


 19   these are also expected to differ according to the


 20   pattern of disease studied.  Again, within each


 21   study the entecavir/lamivudine treatment groups


 22   were well matched for baseline HBV disease




  1   characteristics.  Looking across studies, HBV


  2   e-antigen positive patients, whether


  3   nucleoside-naive or lamivudine-refractory, had mean


  4   HBV DNA values that were approximately 2 logs


  5   higher than the mean value for the e-antigen


  6   negative population.


  7             Finally baseline histology across the


  8   studies showed a higher mean necroinflammatory


  9   score, using Knodell, than nucleoside-naive


 10   subjects.  Only a minority had biopsy evidence for


 11   cirrhosis as classified by Knodell fibrosis score


 12   of 4.  This is because participants were selected


 13   to have compensated liver disease.


 14             Patient disposition--patient disposition


 15   for the first 48 weeks across the three studies


 16   demonstrates high retention rates, with at least 94


 17   percent of entecavir-treated patients completing 48


 18   weeks of treatment in each of the three studies.


 19   Lamivudine retention rates ranged from 87-95


 20   percent, with the lowest rate in the


 21   lamivudine-refractory study.


 22             In all three studies, paired biopsies were




  1   scored using a single reader, who was Dr. Zachary


  2   Goodman.  Dr. Zachary Goodman was blinded to drug


  3   assignment as well as the temporal sequence of the


  4   paired biopsies.  Dr. Goodman also read the


  5   biopsies for lamivudine and adefovir registrational


  6   programs.


  7             Overall, paired baseline and week 48


  8   biopsies were available for efficacy assessment in


  9   88 percent of patients.  Histologic improvement at


 10   week 48 as compared to baseline is the primary


 11   efficacy endpoint in these trials.  Histologic


 12   improvement was defined as at least a 2-point


 13   reduction in the Knodell necroinflammatory score


 14   with no concurrent worsening in Knodell fibrosis.


 15             In order for a biopsy pair to be


 16   evaluable, the baseline sample must have had enough


 17   tissue pathologically and it also must have had a


 18   necroinflammatory score of at least 2, and 89


 19   percent of patients had a baseline biopsy that fit


 20   these criteria and constitute the evaluable


 21   baseline histology cohort.  Patients from the


 22   evaluable cohort who had missing or inadequate week




  1   48 specimens were considered to have no


  2   improvement.  Therefore, the primary analysis for


  3   histologic improvement is analogous to a


  4   non-completer or equal failure analysis but is


  5   applied to the evaluable cohort rather than the


  6   all-treated population.


  7             The nucleoside-naive studies were designed


  8   with two-stage testing.  The first test was for


  9   non-inferiority and, if that was met, then


 10   superiority was tested.  Non-inferiority is


 11   established if the lower confidence limit is above


 12   minus 10 percent.  Superiority is met if the lower


 13   confidence limit is above zero.  In comparing two


 14   active treatments it was expected that differences


 15   in histologic improvement, a downstream endpoint,


 16   might take longer than 48 weeks to emerge.


 17   Nevertheless, at week 48 entecavir 0.5 mg daily was


 18   superior to lamivudine 100 mg daily for histologic


 19   improvement in both nucleoside-naive populations.


 20   Entecavir achieved a 72 percent response rate in


 21   naive e-antigen positive patients and a 70 percent


 22   response rate in the naive e-negative population.


 23             Looking to the study in


 24   lamivudine-refractory patients, this was designed


 25   for superiority.  Two independent co-primary




  1   endpoints were evaluated because histologic


  2   response hadn't been characterized in this


  3   population previously.  The first co-primary


  4   endpoint is histologic improvement, as we have


  5   discussed.  The second is a composite reflecting


  6   both virologic response and hepatic inflammation as


  7   measured by serum ALT.  Entecavir 1 mg daily was


  8   superior to continued lamivudine 100 mg daily for


  9   both co-primary endpoints, and 55 percent achieved


 10   the endpoint of histologic improvement; likewise,


 11   55 percent achieved an HBV DNA below the detection


 12   of the bDNA assay, together with an ALT less than


 13   1.25 times the upper limit of normal.  Changes in


 14   fibrosis are expected to follow changes in


 15   necroinflammation.  While the primary endpoint,


 16   histologic improvement, assessed primarily


 17   necroinflammation, secondary histologic endpoints


 18   included an assessment of changes in fibrosis using


 19   the Ishak scoring system.


 20             The numbers in the circles along the zero


 21   line represent the proportions with no change,


 22   while the bars above and below the line represent


 23   the proportions with improvement and worsening


 24   respectively.  In the two naive studies entecavir


 25   and lamivudine are comparable.  This is not




  1   unexpected as week 48 is relatively an early time


  2   point for assessing this downstream endpoint,


  3   especially when comparing two active treatments.


  4   The effect of large differences, however, can be


  5   seen in lamivudine-refractory patients.  Here


  6   entecavir was superior to lamivudine for


  7   improvement in fibrosis.  The distribution of


  8   responses in entecavir-treated patients mirrors


  9   that in the naive studies and 34 percent had


 10   improvement while only 11 percent worsened while on


 11   entecavir.  This compares to only 16 percent


 12   improvement and 26 percent worsening for continued


 13   lamivudine.


 14             Non-histologic secondary endpoints were


 15   also assessed at week 48.  These included


 16   virologic, biochemical and serologic endpoints. 




  1   These assessments are all used routinely in the


  2   clinical management of patients with chronic HBV.


  3   Treatment comparisons were made using a


  4   non-completer or equal failure analysis, and all


  5   treated patients were counted in the denominator.


  6             Results for virologic endpoints


  7   demonstrate superiority for entecavir in all three


  8   populations studied.  The proportion of patients


  9   achieving an HBV DNA less than 400 copies/mL by PCR


 10   is presented here as a function of time on


 11   treatment, and 69 percent of naive e-antigen


 12   positive patients treated with entecavir achieved


 13   an HBV DNA of less than 400 copies/mL as compared


 14   to 38 percent for lamivudine, an absolute


 15   difference of 31 percentage points.


 16             The lower baseline viremia and e-antigen


 17   negative patients is associated with higher rates


 18   of viral suppression.  Here, 91 percent of


 19   entecavir-treated patients achieved an HBV DNA less


 20   than 400 copies as compared to 73 percent for


 21   lamivudine, an absolute difference of 18 percentage


 22   points.  In both populations there is an early




  1   separation response, with superiority for entecavir


  2   as early as week 24.  This was the first time point


  3   in which a PCR measurement was taken.


  4             In the lamivudine-refractory population


  5   entecavir was also superior to continued


  6   lamivudine, with early separation during the first


  7   24 weeks of treatment, and 21 percent of


  8   entecavir-treated patients achieved an HBV DNA less


  9   than 400 copies.


 10             An additional way of assessing virologic


 11   response is looking at the mean log reduction in


 12   HBV DNA from baseline.  For this analysis results


 13   depend upon the characteristics of the population


 14   studied and the HBV DNA used.  The maximum


 15   reduction possible for a particular population


 16   depends on the starting baseline values for those


 17   individuals.  In a responder the endpoint will


 18   reflect the lower limit of detection for an assay.


 19   Therefore, comparisons of this endpoint across


 20   different populations must account for differences


 21   in baseline characteristics and HBV DNA assay.


 22             Entecavir is superior to lamivudine across




  1   all three populations.  Naive e-antigen positive


  2   patients who started out with an HBV DNA of 9.7


  3   logs in wild type virus demonstrate--so that


  4   entecavir demonstrates its full potential with a


  5   mean decrease of nearly 7 logs at week 48,


  6   differing by 1.5 logs or 30-fold from lamivudine.


  7   In the e-negative population the 5-log decrease for


  8   entecavir approximates the maximal change possible


  9   given the lower starting HBV DNA and the PCR limit


 10   of quantitation at 2.5 logs, or 300 copies/mL.  In


 11   the lamivudine-refractory population entecavir


 12   achieves a substantial 5.1-log decrease in HBV DNA.


 13             Viral suppression also leads to reduced


 14   hepatic inflammation as judged by ALT.  Here,


 15   entecavir is superior to lamivudine for


 16   normalization of ALT in all three populations.  As


 17   expected, the largest treatment difference is seen


 18   in the refractory population.


 19             Reduced viral replication may also induce


 20   an immunologic response resulting in HBe antigen


 21   seroconversion.  The precise biology of this


 22   interaction is poorly understood.  In the naive




  1   e-antigen population entecavir and lamivudine are


  2   comparable for seroconversion with response rates


  3   of 21 and 18 percent respectively.


  4             In summary, across the three Phase III


  5   studies entecavir is consistently superior to


  6   lamivudine for histologic improvement, virologic


  7   response and ALT normalization.  For the four key


  8   endpoints across the three studies there were 11


  9   efficacy comparisons.  Entecavir demonstrates


 10   statistical superiority to lamivudine in 9 of these


 11   11, with confidence intervals for treatment


 12   differences lying to the right of zero.  The two


 13   seroconversion endpoints favor entecavir


 14   numerically and establish non-inferiority with


 15   confidence intervals lying above the minus 10


 16   boundary.  In addition, the mean log reduction is


 17   consistently superior for entecavir, ranging from


 18   5-7 logs across the three populations.


 19             Let's move to safety.  The clinical


 20   profile of entecavir has been extensively


 21   characterized.  The format for the safety


 22   presentation will differ slightly from that of the




  1   efficacy presentation.  These analyses use


  2   augmented patient cohorts and integrate data across


  3   studies in order to increase the sensitivity to


  4   possible safety signals.


  5             The nucleoside-naive lamivudine-refractory


  6   populations are considered separately, primarily


  7   because the exposure to entecavir differs with


  8   dose.  The safety cohort includes patients from 10


  9   analyzed Phase II and Phase III studies.  For the


 10   Phase III populations mean treatment duration was 5


 11   weeks longer for entecavir-treated naive patients


 12   and 17 weeks longer for entecavir-treated


 13   refractory patients.  The follow-up observations


 14   were consistently longer for entecavir than for


 15   lamivudine across all populations.


 16             Follow-up is defined as the period of


 17   post-treatment follow-up during which no


 18   alternative HBV therapy was given.  Its duration


 19   was shorter in refractory patients as compared to


 20   naive patients due to earlier initiation of


 21   alternative therapy or early enrollment into an


 22   entecavir rollover trial.  Observation periods for




  1   the safety cohort are expanded to include


  2   open-label treatment and post-treatment observation


  3   on alternate HBV therapy.


  4             The safety presentation is divided into


  5   three sections, general safety, hepatic safety and


  6   malignant neoplasms.  General safety analyses


  7   provide standard assessments for rates of clinical


  8   adverse events and laboratory abnormalities.  All


  9   analyses use data from all treated patients in the


 10   selected studies.  Analyses are cumulative from the


 11   first day of dosing through the last contact with


 12   each patient.  Therefore, year 2 data are included


 13   for some patients.


 14             Rates for three standard safety


 15   assessments--discontinuations due to an adverse


 16   event, serious adverse events and deaths, were low


 17   for both treatments across both populations.  The


 18   types of serious events reported for entecavir and


 19   lamivudine were comparable, and no individual


 20   serious adverse event occurred in more than one


 21   percent of patients.  None of the events leading to


 22   death was considered related to study drug.


 23             In terms of adverse events, on treatment


 24   adverse events were generally mild to moderate in


 25   severity and were common, reflecting the long




  1   duration of study observation.  The frequencies of


  2   individual events and the types and distribution of


  3   these events were comparable for both treatment


  4   groups across both populations.


  5             Hepatic safety--hepatic safety focuses on


  6   hepatic flares because these can represent an


  7   important clinical risk in the treatment of


  8   hepatitis B regardless of the specific therapy


  9   which is used.  ALT flares were defined as


 10   increases in ALT greater than 10 times the upper


 11   limit of normal and 2 times the patient's own


 12   reference value.  The reference value was the


 13   baseline value for on-treatment flares.  For


 14   off-treatment flares the reference was the lower of


 15   the baseline or the end of treatment value.


 16             Rates for on- and off-treatment flares are


 17   consistently less than 10 percent for entecavir.


 18   Of note, the median time from stopping therapy to


 19   an off-treatment flare is substantially longer for




  1   entecavir.  The delayed time course for


  2   off-treatment flares for entecavir may be related


  3   to the extent of virologic suppression achieved on


  4   treatment.


  5             ALT flares are frequently asymptomatic.  A


  6   deterioration in hepatic function can, however,


  7   occur without ALT changes that meet this flair


  8   definition.  Therefore, we performed analyses to


  9   identify individuals meeting flair criteria who had


 10   associated relevant laboratory abnormalities or


 11   relevant hepatic clinical events, or those who had


 12   a serious hepatic adverse event without meeting


 13   flair criteria.  These events were infrequent among


 14   both naive and refractory patients, with the number


 15   of individual cases summarized here.


 16             Safety surveillance of the entecavir


 17   development program involved the assessment of


 18   comparative incidences for new or recurrent


 19   malignancy diagnoses in entecavir- and


 20   lamivudine-treated subjects.  Use of the larger


 21   safety cohort database increases sensitivity in


 22   this analysis of events that are infrequent.  A new




  1   diagnosis or a new recurrence of malignancy was


  2   counted from the time of first study dose to the


  3   time of the last patient contact regardless of


  4   whether the event was diagnosed on or post


  5   treatment.  In the safety cohort the


  6   entecavir/lamivudine treatment groups differed in


  7   size and the duration of observation.


  8             Event rates are presented as incidences of


  9   patients diagnosed per 1,000 patient-years of


 10   observation.  Hepatocellular carcinoma is the


 11   single most frequent type of cancer identified, not


 12   unexpectedly, due to the underlying HBV disease.


 13   Incidences across the treatment groups are


 14   comparable whether assessed for any malignancy, any


 15   malignancy excluding non-melanoma skin tumors or


 16   the category of great interest, non-hepatocellular


 17   carcinoma, non-skin malignancies.


 18             Further analyses in the entecavir program


 19   demonstrate that the distribution of new or


 20   recurrent non-skin malignancy diagnoses over time


 21   is comparable for entecavir and lamivudine.  In


 22   both treatment groups the greatest number of new




  1   diagnoses occurred between weeks 24 and 48.  This


  2   temporal clustering may reflect tumors that were


  3   latent at the time of study enrollment.  There is


  4   an apparent leveling off for new diagnoses after


  5   week 48.


  6             In order to establish a comparative


  7   context for the observed tumor rates in the


  8   development program, Bristol-Myers Squibb provided


  9   grants to two independent research groups.  These


 10   groups identified cohorts of chronic HBS antigen


 11   positive patients within their established


 12   databases.  The results are provided in the two


 13   right-hand columns.  The Taiwan cohort had been


 14   prospectively identified as part of an established


 15   cancer incidence study which started in 1991 and is


 16   sponsored by the Taiwan Ministry of Health.  The


 17   rates of malignancy in the entecavir-lamivudine


 18   arms are comparable to the Taiwan and the Kaiser


 19   observational cohorts.


 20             In summary, the safety profile of


 21   entecavir is consistently comparable to that of


 22   lamivudine.  Also, the safety of entecavir is




  1   comparable across the nucleoside-naive and


  2   lamivudine-refractory populations, and across the


  3   two doses of 0.5 mg and 1 mg daily.  Importantly,


  4   the malignancy incidences among approximately 1,500


  5   entecavir-treated patients are comparable among


  6   those observed in the lamivudine-treated control


  7   group.  Dr. Richard Colonno will now present the


  8   resistance profile for entecavir.


  9                            Resistance


 10             DR. COLONNO:  Thank you.  For all


 11   antivirals there is a direct relationship between


 12   potent viral suppression and absence of viral


 13   resistance emergence because viruses require a


 14   minimal threshold level of replication to select


 15   for resistant variants.  Sustained suppression of


 16   viral DNA undetectable levels in the woodchuck


 17   model, described earlier, resulted in the absence


 18   of virologic rebound and no evidence of resistance


 19   over the 14- and 36-month treatment periods.


 20             To ascertain whether the potent and


 21   sustained suppression of viral replication achieved


 22   by entecavir in our clinical studies results in a




  1   favorable resistance profile, a comprehensive


  2   resistance evaluation was conducted that included


  3   both in vitro and in vivo studies, along with


  4   characterization of over 1,500 clinical samples


  5   from entecavir-treated patients.


  6             In vitro studies showed entecavir


  7   susceptibility was reduced when viruses contained


  8   the two primary lamivudine-resistant substitutions,


  9   a leucine thymodin[?] change at residue 180 and a


 10   methionine to valine or isoleucine change at


 11   residue 204.  Despite this reduction, entecavir


 12   remains greater than 50-fold more potent than


 13   adefovir against lamivudine-resistant viruses.


 14   There was no cross-resistance between entecavir and


 15   adefovir since adefovir-resistant viruses


 16   containing resistant substitutions at residues 181


 17   or 236 remain fully susceptible to entecavir.


 18             During Phase II studies two extensively


 19   pretreated patients, designated as patient A and


 20   patient B, exhibited virologic rebounds on


 21   entecavir therapy.  Following at least 76 weeks of


 22   entecavir, virologic rebounds noted in two patterns




  1   of genotypic resistance emergence were identified.


  2   Entecavir resistance emergence in patient A


  3   required two additional substitutions, an


  4   isoleucine change at residue 169 and a valine


  5   substitution at residue 250.  Patient B needed


  6   glycine and isoleucine substitutions at residues


  7   184 and 202 respectively, along with a subsequent


  8   change at residue 169.  In both cases these changes


  9   occurred in the background of preexisting


 10   lamivudine-resistant substitutions.  Both isolates


 11   were growth impaired and remained fully susceptible


 12   to adefovir.


 13             The impact of substitutions at each of


 14   these four residues of entecavir's susceptibility


 15   are shown on this slide.  Recombinant viruses


 16   containing the indicated substitutions at residues


 17   169, 184 and 202 alone had no significant impact on


 18   entecavir's susceptibility relative to wild type


 19   virus, while a change at residue 250 reduced


 20   entecavir's susceptibility levels by less than


 21   10-fold, about the same as when


 22   lamivudine-resistant substitutions alone are




  1   present.


  2             The 169 substitution appears to act as a


  3   secondary mutation and did not further reduce


  4   entecavir's susceptibility in the


  5   lamivudine-resistant viruses.  However, when


  6   lamivudine-resistant substitutions are combined


  7   with the entecavir-resistant substitutions at


  8   residues 184, 202 and 250 significantly higher


  9   levels of entecavir resistance are observed.


 10   Presence of multiple entecavir-resistant


 11   substitutions further decreased entecavir's


 12   susceptibility levels.


 13             An extensive resistance monitoring program


 14   was undertaken.  In the nucleoside-naive trials all


 15   available entecavir-treated e-antigen positive and


 16   two-thirds of randomly selected e-antigen negative


 17   patients were genotyped at study entry and at week


 18   48, a total of 550 pairs of patient samples.  For


 19   the lamivudine-refractory population all available


 20   patient samples were genotyped.  All emerging


 21   changes identified were tested for their potential


 22   impact on entecavir susceptibility.


 23             In addition, samples from all patients


 24   experiencing a virologic rebound, defined as any


 25   greater than or equal to 1 log increase from nadir




  1   identified by PCR, were genotyped and subjected to


  2   population phenotyping to determine if they


  3   harbored circulating viruses resistant to study


  4   drug.  In nucleoside-naive patients treated with


  5   entecavir there was no evidence of genotypic or


  6   phenotypic resistance by week 48.


  7             The figure plots the distribution of


  8   patients with the HBV DNA levels indicated at study


  9   entry and at week 48 for both entecavir and


 10   lamivudine.  The size of each circle corresponds to


 11   the percentage of patients and each column of


 12   circles adds up to 100 percent.  And, 81 percent of


 13   entecavir-treated patients achieved viral DNA


 14   levels of less than 300 copies/mL, represented by


 15   the bottom circle, compared to only 57 percent for


 16   lamivudine-treated patients.  Overall, 88 percent


 17   of patients, represented by the bottom two circles


 18   in each case, achieved viral DNA reductions below


 19   1,000 copies/mL on entecavir by week 48.


 20             Genotyping identified 76 emerging changes


 21   but no distinctive patterns were observed, and no


 22   change was present in more than three isolates,


 23   representing 0.6 percent of those treated.


 24   Phenotypic analysis of these emerging changes show


 25   that their presence did not result in a significant




  1   decrease in entecavir susceptibility.  There were


  2   11 virologic rebounds on the entecavir arms of


  3   these studies compared to 88 rebounds on lamivudine


  4   therapy.


  5             This slide shows the origin and frequency


  6   of rebounds by study.  When genotyped, nearly all


  7   of the observed virologic rebounds on lamivudine


  8   therapy coincided with the emergence of resistance


  9   substitutions at residues 180 and 204, yielding a


 10   confirmed resistance frequency of 8-18 percent by


 11   week 48.  In contrast, none of the entecavir


 12   virologic rebounds observed in nucleoside-naive


 13   patients could be attributed to emergence of


 14   resistance.


 15             A close examination of the individual


 16   patient profiles showed that all 11 patients




  1   exhibiting a rebound on entecavir had at least a


  2   3-log reduction in viral DNA levels and 7 of the 11


  3   had greater than a 5-log reduction.  Most


  4   importantly, all patients had viral populations


  5   that were full susceptible to entecavir at the time


  6   of rebound, and there was no evidence of emerging


  7   genotypic changes that reduced entecavir


  8   susceptibility.


  9             From this comprehensive analysis we


 10   conclude that there was no evidence of emerging


 11   genotypic or phenotypic resistance to entecavir in


 12   any of the nucleoside-naive patients by week 48, a


 13   result that is most likely due to the high degree


 14   of sustained viral suppression observed.  We


 15   continue to monitor these patients for resistance


 16   in subsequent treatment years.


 17             Let us now turn to the


 18   lamivudine-refractory patient population where


 19   previous studies indicated that entecavir


 20   resistance emergence can occur.  Similar to


 21   nucleoside-naive patients, entecavir was highly


 22   effective in lamivudine-refractory patients




  1   enrolled in study 026 and in the 1 mg arm of study


  2   014.


  3             The figure again plots the distribution of


  4   lamivudine-refractory patients having the HBV DNA


  5   levels indicated at study entry, week 24 and week


  6   48.  While reductions were somewhat less than those


  7   observed in nucleoside-naive patients, 22 percent


  8   of entecavir-treated patients achieved viral DNA


  9   reductions below 300 copies/mL by week 48.  There


 10   was a clear trend of sustained and increasing


 11   reductions from week 24 to week 48, and superiority


 12   to continued lamivudine therapy.


 13             As part of our comprehensive resistance


 14   evaluation, all patients, regardless of treatment


 15   arm, were genotyped at study entry and week 48.


 16   There were 5 virologic rebounds among the


 17   lamivudine-refractory patients treated with


 18   entecavir.


 19             The figure plots the HBV DNA levels for


 20   the first two patients, labeled 1 and 2.  Both


 21   exhibited only modest reductions in HBV DNA levels


 22   on entecavir therapy.  Evidence of entecavir




  1   resistance substitutions at residue 184 were noted


  2   in both patients and population phenotypes


  3   indicated a 15-19-fold decrease in entecavir


  4   susceptibility, consistent with resistance


  5   emergence.


  6             In contrast, the three other patients,


  7   labeled 3, 4 and 5, all experienced at least a


  8   4-log reduction in viral DNA levels and further


  9   reductions following rebound either on continued


 10   therapy or off treatment, with no evidence of


 11   genotypic or phenotypic changes beyond those


 12   expected for lamivudine-resistant viruses.


 13             Based on this evaluation, only two


 14   patients or one percent of lamivudine-refractory


 15   patients treated with entecavir experienced


 16   virologic rebound due to resistance by week 48.


 17   Entecavir-resistant substitutions were, however,


 18   noted in 12 entecavir-treated patients by week 48,


 19   all with a background of lamivudine-resistant


 20   substitutions.  These patients continue to be


 21   monitored for virologic rebounds in subsequent


 22   years.  Emerging substitutions at 14 other residues




  1   were also identified, but none were present in more


  2   than 3 patients or reduced entecavir susceptibility


  3   beyond those expected for lamivudine-resistant


  4   viruses.


  5             An unexpected finding was that lamivudine


  6   can preselect for entecavir-resistant


  7   substitutions.  This was further supported by the


  8   observation that lamivudine-treated patients showed


  9   evidence of emerging changes at residues 169 and


 10   184 in study 026.  Among the greater than 360


 11   lamivudine-refractory patients genotyped, at least


 12   22 had detectable changes at entecavir-resistant


 13   substitutions at study entry.  Nine were randomized


 14   to an entecavir treatment arm, where two progressed


 15   to have resistance-induced virologic rebounds


 16   described earlier.  Only 2/9 patients were able to


 17   reduce viral DNA levels below 300 copies/mL.  This


 18   observation, along with the other results described


 19   in this presentation, indicate that extended use of


 20   lamivudine will not only select for the primary


 21   lamivudine-resistant substitutions at 180 and 204,


 22   but can also select for a number of secondary




  1   substitutions that can significantly reduce


  2   entecavir susceptibility and clinical efficacy.


  3             This slide summarizes our current


  4   understanding of the entecavir resistance profile


  5   at week 48.  There was no evidence of genotypic or


  6   phenotypic resistance in any studied


  7   nucleoside-naive patients treated with entecavir.


  8   Entecavir did not select for lamivudine-resistant,


  9   or entecavir-resistant substitutions, or other


 10   novel substitutions that result in decreased


 11   entecavir susceptibility and there were no


 12   virologic rebounds due to resistance.


 13             Among the patients having primary


 14   lamivudine-resistant substitutions at residues 180


 15   and 204, 7 percent exhibited emerging


 16   entecavir-resistant substitutions while on


 17   entecavir therapy, and only 1 percent of


 18   lamivudine-refractory patients exhibited a


 19   virologic rebound due to resistance by week 48.


 20   The preexistence of entecavir-resistant


 21   substitutions appears to be a marker for decreased


 22   efficacy and potential virologic rebound.


 23             In summary, the potent and sustained


 24   suppression of viral replication by entecavir


 25   likely accounts for the absence of resistance




  1   emergence in nucleoside-naive patients.  An


  2   extensive analysis of nucleoside-naive patients


  3   showed no evidence or resistance.  Entecavir was


  4   also effective in lamivudine-refractory patients


  5   where only 1 percent of patients experienced a


  6   virologic rebound due to resistance by week 48.


  7   Substitutions correlated with entecavir resistance


  8   were identified at primary residues 184, 202 and


  9   250 and the secondary residue 169.


 10   Lamivudine-resistant substations are a prerequisite


 11   for achieving high level entecavir resistance and


 12   lamivudine treatment can preselect for some


 13   entecavir-resistant substitutions.


 14             We conclude that this virologic profile


 15   provides critical information to physicians


 16   regarding the placement of entecavir in the


 17   armamentarium of drugs available to treat chronic


 18   hepatitis B infection.  Dr. Donna Morgan Murray


 19   will now conclude our presentation with




  1   pharmacovigilance and final summary.


  2                  Pharmacovigilance and Summary


  3             DR. MORGAN MURRAY:  As you have heard this


  4   morning, the entecavir clinical development program


  5   was extensive.  It was the largest HBV program


  6   conducted to date and the only antiviral HBV


  7   program to use an active comparator in Phase III


  8   trials.  That comparator was lamivudine, the only


  9   agent available at the time of initiation of the


 10   trials and the most common HBV therapy used to


 11   date.


 12             Entecavir demonstrated substantial


 13   clinical benefit in Phase III and was superior to


 14   lamivudine in the prespecified primary endpoint of


 15   improved histology.  Entecavir was also superior to


 16   lamivudine in most of the secondary endpoints.


 17             Based on the rodent tumor findings,


 18   entecavir is a rodent carcinogen.  The lung tumors


 19   appear to be species specific, and the other tumors


 20   occur at high exposure multiples.  The


 21   investigative data submitted to the carcinogenicity


 22   assessment committee do not definitively eliminate




  1   a risk for humans.  With more than 2,300 patients


  2   treated with entecavir, there is no safety signal


  3   related to malignancy in the clinical development


  4   program.  While this is reassuring, we recognize


  5   that the observation period is short.


  6             As Dr. Sigal mentioned, we are committed


  7   to continuously assessing the benefit versus risk


  8   profile of entecavir, and have proposed a


  9   post-marketing pharmacovigilance plan with three


 10   main components.  In addition to routine


 11   post-marketing surveillance, the pharmacovigilance


 12   plan also includes real-time monitoring of special


 13   events, specifically malignancies and hepatic


 14   events.  We have designed special questionnaires to


 15   aid in collecting follow-up information for reports


 16   of both malignancies and hepatic events.  We will


 17   periodically review post-marketing and clinical


 18   trial adverse event data, using quarterly aggregate


 19   frequency reports, and we will review these events


 20   of special interest.


 21             There are three ongoing long-term safety


 22   studies and we have proposed an additional large,




  1   prospective, randomized safety study to be


  2   conducted post-marketing.  First let's review the


  3   ongoing studies.


  4             The clinical development program included


  5   one- to two-year treatment studies and long-term


  6   safety studies with careful observation for the


  7   development of malignancies.  Responders from the


  8   Phase II/III trials were encouraged to enroll in an


  9   observational study that was aimed to gather safety


 10   data off treatment.  Malignancy was the primary


 11   focus of this observational study.  Some patients


 12   from the Phase II treatment studies were eligible


 13   to enroll in open-label treatment studies, and


 14   these patients were also encouraged to enroll in


 15   the observational study.


 16             To date, more than 80 percent of patients


 17   from Phase III have enrolled in at least one of the


 18   long-term safety studies, and the observational


 19   study has more than 400 patients enrolled, with the


 20   expectation that we will enroll up to 1,500


 21   patients and all patients will be followed for 5


 22   years.  In addition to the ongoing studies, we




  1   propose initiating a large safety study post


  2   approval.


  3             Given the limitations of pre-approval


  4   clinical studies, we recognize that we cannot rule


  5   out a cancer risk in patients treated with


  6   entecavir.  Pre-approval studies do not provide


  7   sufficient numbers of patients to rule out such


  8   uncommon events.  We considered several options for


  9   further assessment and concluded that a randomized,


 10   prospective study would permit rigorous analysis of


 11   these events of special interest--mortality,


 12   neoplasms and progression of liver disease.


 13             The draft protocol for this study calls


 14   for patients to be randomized 1:1 to entecavir


 15   versus another standard of care nucleoside or


 16   nucleotide; to be stratified as naive or previously


 17   treated; and to be followed for at least 5 years.


 18   It is our intent to engage an external, independent


 19   data safety monitoring board to conduct periodic


 20   reviews of the data from this study.


 21             We propose to conduct the study globally


 22   and to recruit patients via their own physicians. 




  1   Patients who are starting a new HBV therapy or are


  2   changing their therapy will be eligible to enroll.


  3   We expect to enroll a total of 12,500 patients.  We


  4   will report annually on rates of all-cause


  5   mortality, malignancy and progression of liver


  6   disease.  While other common nucleosides also have


  7   rodent tumor findings, and the benefit-risk


  8   assessment was favorably concluded based on the


  9   serious nature of the disease, such as AZT for HIV,


 10   few have been the subject of the rigorous


 11   assessment that we propose here.


 12             However, the proposed study does have


 13   several challenges.  First, the planned primary


 14   analysis is intent-to-treat and, as patients will


 15   inevitably switch therapies over the course of the


 16   study, the primary analysis may be confounded.


 17   However, we will not limit our review of the data


 18   to this analysis and we will look at the data in


 19   several different ways.


 20             Second, there may be limited ability to


 21   detect treatment group differences for events of


 22   variable latency.  Since all patients will be




  1   studied for at least 5 years, and many may well be


  2   studied for up to 8 years, we should detect a


  3   signal if there is an increased risk.


  4             Third, the study is designed to detect


  5   differences in overall malignancy rates and in


  6   rates of HCC, but is not designed to detect


  7   treatment group differences for individual


  8   malignancy types.


  9             Finally, attrition will occur but this


 10   does not mean that patients will be lost to


 11   follow-up.  We will implement tactics to enhance


 12   follow-up, and we have developed strategies to


 13   address these challenges listed on this slide, and


 14   conclude that the proposed study will provide


 15   important data on both the benefits of entecavir


 16   and on further risk assessment.


 17             Adequate data exist to demonstrate the


 18   substantial benefit of entecavir over existing


 19   therapies.  Entecavir provides superior viral


 20   suppression in both nucleoside-naive and


 21   lamivudine-refractory patients.  Specifically,


 22   treatment with entecavir resulted in up to a 7-log




  1   decrease in HBV DNA.


  2             Entecavir results in superior


  3   normalization of ALT in both nucleoside-naive and


  4   lamivudine-refractory patients.  Up to 78 percent


  5   of patients achieve normal ALT.


  6             Entecavir also provides superior


  7   improvement in histology in both nucleoside-naive


  8   and lamivudine-refractory patients.  Treatment with


  9   entecavir resulted in up to 72 percent reduction in


 10   necroinflammation.


 11             Entecavir has a favorable resistance


 12   profile compared to lamivudine.  As you heard from


 13   Dr. Colonno, no resistance substitutions emerged in


 14   nucleoside-naive patients and resistance


 15   substitutions were uncommon in


 16   lamivudine-refractory patients.


 17             Given the demonstrated superiority of


 18   entecavir in viral suppression, ALT normalization


 19   and improved histology, and the favorable


 20   resistance profile both in nucleoside-naive and


 21   lamivudine-refractory patient populations,


 22   long-term benefits of entecavir might include a




  1   reduction in disease progression, such as lower


  2   rates of liver failure, liver cancer, liver


  3   transplant and liver-related deaths.


  4             We conclude that the demonstrated benefits


  5   of entecavir represent an important treatment


  6   advance for HBV infection.  The demonstrated


  7   benefits of entecavir against HBV, a known


  8   carcinogen, are indeed substantial and outweigh the


  9   theoretical risk posed by the rodent tumor data.


 10   Thank you for you attention this morning.


 11                   Questions from the Committee


 12             DR. ENGLUND:  Thank you very much, Dr.


 13   Murray.  I would like to thank the Bristol-Myers


 14   Squibb people for a very clear, concise and timely


 15   presentation.  It was very nice.  Thank you.


 16             This is the time that we are going to open


 17   up for questions to the panel, but I would like to


 18   caution people that the questions are supposed to


 19   be directly related to the information presented


 20   today.  We will have discussion time later on but


 21   if there are clarifications or questions about


 22   specific points related to the presentation we just




  1   heard, now is the time to begin so I will open it


  2   to the panel for questions.  Dr. DeGruttola?


  3             DR. DEGRUTTOLA:  Yes, I have two


  4   questions.  The presentations mentioned that the


  5   studies in dogs and rats did not find an increased


  6   risk of lung cancer associated with entecavir.  I


  7   was wondering how long those studies had gone on;


  8   were they powered to be able to detect such an


  9   effect?  Then, regarding the post-marketing study


 10   to try to determine an effect on cancer in humans,


 11   I was wondering what the power will be in that


 12   study; what magnitudes of effects is the study


 13   powered to detect?


 14             DR. MORGAN MURRAY:  First I will ask Dr.


 15   Lois Lehman-McKeeman to address your first question


 16   about the duration of studies in dogs and rats.


 17             DR. LEHMAN-MCKEEMAN:  I will speak to the


 18   rats first because they were, in fact, one of the


 19   species used in the lifetime carcinogenicity study.


 20   So, in two years, for the lifetime of the rat,


 21   there were no tumors in the lung that developed.


 22             The dog the studies were not conducted to




  1   be carcinogenicity studies; they were chronic


  2   toxicology studies and they were three months in


  3   duration.  However, what we understand about the


  4   lung lesion in the mouse is that it develops very


  5   quickly and the early preneoplastic change that I


  6   described occurs within the first two weeks of


  7   dosing.  In the course of a three-month study in


  8   dogs we saw no early preneoplastic change.


  9             DR. DEGRUTTOLA:  Thank you.


 10             DR. MORGAN MURRAY:  And for your second


 11   question about the power of our post-marketing


 12   study to detect differences, Dr. Phil Pierce will


 13   address that.


 14             DR. PIERCE:  The primary goal of the large


 15   safety trial is to investigate the potential


 16   treatment effect on the development of non-HCC


 17   malignancies.  First we had to establish what the


 18   background rate in this population is, and we


 19   utilized the data from the Taiwan cohort that was


 20   presented, as well as the background rates that we


 21   saw in the BMS studies.


 22             The background rate was approximately 4




  1   non-HCC cancers over 1,000 patient-years of


  2   follow-up.  We estimated from that that there would


  3   be 16 non-HCC malignant events per 1,000


  4   patient-years per arm over 5 years.  Also, the


  5   total accrual of time will be 65,000 patient-years.


  6   Our study was designed to show a 30 percent


  7   increased risk of malignancy.  That translates into


  8   5 additional cancers per 1,000 patient-years over


  9   the 16 that I mentioned earlier.  I believe BMS


 10   concludes this is a reasonable assessment of that


 11   risk.


 12             Slide 1-520, please.  I gave you a lot of


 13   numbers with that and I want to show the expected


 14   events in the untreated population over the 5


 15   years.  The rate that I mentioned for the non-skin,


 16   non-HCC cancers is 16 as the expected rate and we


 17   would have a power to detect, with this sized


 18   population, an increase of 5 over that 16.  The


 19   additional benefit of this study is that we will


 20   also be able to analyze the impact on the other


 21   events of interest which, obviously because of the


 22   large size of those, we are adequately powered to




  1   show whether we have an impact on the rates of HCC


  2   and on the progression to cirrhosis.


  3             DR. DEGRUTTOLA:  Thank you.


  4             DR. ENGLUND:  Thank you.  Dr. Washburn?


  5             DR. WASHBURN:  It is very interesting that


  6   the study drug is chemotactic for mouse monocytes


  7   but not human monocytes.  I wonder if there is any


  8   work that can be shared that would discuss some


  9   mechanism of that difference.  Does it relate to


 10   complement activation, or a macrophage chemotactic


 11   peptide, or other?  The question is of potential


 12   relevance in the carcinogenicity of disease.


 13             DR. MORGAN MURRAY:  Dr. Lehman-McKeeman


 14   will address that.


 15             DR. LEHMAN-MCKEEMAN:  At this point in


 16   time we don't know the molecular basis of that


 17   difference.  What we know is that based on the fact


 18   that macrophages were accumulating in the lung and


 19   were not proliferating to accumulate, we looked


 20   specifically for a chemotactic event and we tested


 21   that in some standard in vitro systems.  When we


 22   did that work, there is clear chemotactic activity




  1   to the mouse with no effect in the human at all.


  2             Now, to go further, we have looked, in


  3   doing some investigative work, at whether or not


  4   altering macrophage recruitment alters the


  5   progression of this lesion.  To do that, we have


  6   looked at a CCR2 knockout, so chemokine receptor to


  7   a knockout animal, and we found that that mouse


  8   does, indeed, have a very different response to the


  9   drug.  It is no unequivocal proof that this is


 10   mediated through CCR2, but it suggests that it


 11   plays a role.


 12             I want to add one other factor though, and


 13   that is that the lesion that we see involves


 14   accumulation of macrophages but, based on our


 15   assessment, those macrophages don't appear to be


 16   activated.  They are simply accumulating.


 17             DR. WASHBURN:  Thank you.


 18             DR. ENGLUND:  Dr. Fish?


 19             DR. FISH:  I didn't hear my name earlier


 20   in the disclosure statement and I just need to add


 21   that though I signed the disclosure waiver, I have


 22   been on the speakers bureau for the sponsor and two




  1   competitors.


  2             The question that I have is on the study


  3   were there pregnancies and, if so, the outcomes of


  4   those pregnancies in entecavir-treated patients?


  5             DR. MORGAN MURRAY:  I am going to try out


  6   Dr. Brett-Smith's voice here.  So, Helena?


  7             DR. BRETT-SMITH:  The studies were


  8   designed that if pregnancy was determined to occur


  9   during the course of the study the patient was to


 10   immediately stop study drug.  Indeed, pregnancies


 11   do occur.  The majority of these actually resulted


 12   in elective termination of pregnancies.


 13             If we could show slide 5-79, this includes


 14   the various treatment combinations that have been


 15   used across our entire program to date with


 16   entecavir alone, lamivudine alone, entecavir in


 17   combination with lamivudine, for the initial period


 18   of the 901 long-term rollover study and also in


 19   placebo.


 20             As you can see, the majority of


 21   pregnancies identified resulted in elective


 22   termination.  There was a small number of




  1   spontaneous abortions.  There have been 6 live


  2   births.  The 4 outcomes that are listed as


  3   "unknown" are progressions that are currently under


  4   way and for which we are actively pursuing


  5   follow-up on those deliveries.


  6             With respect to the live births, across


  7   those live births there were no reported defects in


  8   5 out of the 6 cases.  There was, indeed, 1 live


  9   birth where the mother had received entecavir 0.5


 10   mg for a total of 44 weeks but the diagnosis of the


 11   pregnancy was made at approximately week 7 of


 12   gestation.  That had a fairly complicated history.


 13   The child was born with what has been reported to


 14   us as a severe cerebral cortex defect.


 15   Unfortunately, despite repeated contact with the


 16   site, the family has not wished to provide us with


 17   further data.


 18             The details of the early pregnancy are a


 19   little complex so let me walk you through those.


 20   The patient had discontinued entecavir immediately


 21   at the time that pregnancy was diagnosed, as I


 22   said, about week 7.  The patient then experienced




  1   what was clinically diagnosed as a spontaneous


  2   abortion and was told by the gynecologist that no


  3   fetus had been present.  A subsequent ultrasound


  4   actually did reveal a live fetus, but in the


  5   interim entecavir had been briefly restarted by the


  6   clinician for 2 weeks and the moment the ultrasound


  7   became available it was discontinued.  So, that


  8   represents the sum of our experience to date in the


  9   program with pregnancy.


 10             DR. ENGLUND:  Dr. Haubrich?


 11             DR. HAUBRICH:  It is clear that emergence


 12   of viral resistance to therapy is dependent on the


 13   degree of viral suppression and, clearly, drugs


 14   that have greater suppression will have less


 15   emergence of resistance.  It is also clear from


 16   extensive experience in AZT that after 15-20 years


 17   of nucleoside therapy we are still identifying new


 18   mutations.  So, perhaps I didn't follow it well,


 19   but if you could clarify the emergence of mutations


 20   that may have occurred with entecavir.  Although


 21   they may not lead to phenotypic susceptibility


 22   since the number of mutations is few at this point,




  1   you know, they may in the future be defined when


  2   greater numbers of samples are available.


  3             So, just a comment that it is clear that


  4   the resistance profile is better with greater


  5   suppression, but it seems a little premature to be


  6   saying that there is no resistance that develops on


  7   therapy when the number of specimens is low and it


  8   may be a bit early.  So, if you could comment on


  9   that I would appreciate it.


 10             DR. MORGAN MURRAY:  I will ask Dr. Colonno


 11   to comment but first I would like to note that the


 12   original NDA and the safety update--at that time we


 13   only had 48-week data available and that is the


 14   only data that have been submitted for review.  But


 15   very recently we did complete the analysis on


 16   patients who have been treated for two years and


 17   Dr. Colonno can perhaps share those data as well.


 18             DR. COLONNO:  Let me just deal with the


 19   first part first in terms of the number of


 20   mutations, just to give you a sense of what


 21   mutations were found.


 22             Can I have slide 1-315, please?  This is a




  1   list of all the mutations that have been found or


  2   identified in all patients examined that have taken


  3   entecavir--as you can see, a very wide range.  The


  4   vast, vast majority of these, again, have occurred


  5   at polymorphic sites.  We call them new emerging


  6   substitutions because they have not been described


  7   previously at those particular sites.


  8             Again, I will point out that these


  9   mutations do not occur in any more than three


 10   patients.  Most of these occur in a single patient,


 11   again, representing less than one percent.  We have


 12   tested all of these different mutations and


 13   substitutions not only by themselves but also in


 14   the context of their preexisting clinical


 15   background and, as you can see by the EC                                 

                                                                50s that are


 16   present, they really do not alter the normal wild


 17   type susceptibility.


 18             Now if I can just move to your statement,


 19   which I think is a correct one and, again, as a


 20   virologist having worked in resistance for many,


 21   many years, there is no such thing as no


 22   resistance.  So, we have gone out to the second




  1   year, and this is real-time data and the data


  2   continues to come in, and I would like to just


  3   share with you some very encouraging data for the


  4   second year.


  5             This is the second year data as it


  6   currently stands.  On the left-hand side, again,


  7   are the bubble charts and the first thing I want to


  8   point out is this is study 022 where we have the


  9   most data.  You can see that the continued


 10   progression in decreasing DNA from week 48 to 96,


 11   where we have 65 undetectable now, we continue to


 12   drive viral load down with 81 percent of patients


 13   now with undetectable virus.


 14             That correlates with the table on the


 15   right where, again, despite the fact that we have


 16   treated now for 2 years, we have a very similar


 17   profile to what we saw in year 1.  In year 2 we


 18   have a total of 7 rebounds, virologic rebounds


 19   using the definition I described earlier but,


 20   again, looking at their genotypes and phenotypes we


 21   see no evidence of any genotypic or phenotypic


 22   resistance.  So, out to 2 years in the




  1   nucleoside-naive population with that type of viral


  2   suppression we have not observed any resistance to


  3   entecavir.


  4             DR. ENGLUND:  Dr. Johnson, do you have a


  5   specific question about that?


  6             DR. JOHNSON:  Victoria Johnson, University


  7   of Alabama at Birmingham.  As a virologist and


  8   viral resistance person, I share concerns that


  9   despite the elegant data presented, given this


 10   compound's potency, as you realize, two years may


 11   not be enough, and I want to just ask is this part


 12   of the pharmacovigilance monitoring plan?  That is


 13   one question.


 14             The second question is, if you can go to


 15   your second to last slide of your previous


 16   presentation--


 17             DR. MORGAN MURRAY:  Let me answer your


 18   first question first around the pharmacovigilance


 19   plan.  Several of our studies are ongoing, as I had


 20   mentioned, and in all of the ongoing clinical


 21   studies we do continue to monitor for resistance.


 22   Acknowledging that the pharmacovigilance plan is




  1   very large, we will have many centers and it will


  2   be usual practice, we feel it will be impossible


  3   for us to get resistance data on all of the 12,500


  4   patients.  But what we do propose is to have a


  5   sub-study, a subset of patients, a center in the


  6   U.S., a center, you know, here and there that we


  7   will get much more data including resistance data.


  8   I will let Dr. Colonno address your second point.


  9             DR. COLONNO:  We will continue to look for


 10   resistance until we find it.  Again, there is


 11   always going to be resistance at some point.  But


 12   the key point of this slide, which we don't have


 13   with HIV, unfortunately, even with combination


 14   therapy, is the ability to drive viral load down by


 15   6 or 7 logs, 8 logs in some cases and to maintain


 16   that for a very long period of time.  Those viruses


 17   require a minimal amount of replication to give


 18   rise to resistance.  So, we are encouraged.  Again,


 19   that is not to say there will never be resistance


 20   but we are highly encouraged with that kind of


 21   suppression and with the limited ability of the


 22   virus to actually replicate that a large amount of




  1   resistance will all of a sudden come up.  We will


  2   continue to monitor these patients for the


  3   foreseeable future.


  4             Another interesting point is that these


  5   particular patients do not give rise to any


  6   evidence of resistance substitutions being


  7   selected.  We know lamivudine resistance is a


  8   stepping stone to becoming clinically relevant


  9   resistance to entecavir.  But the fact that we,


 10   again in that population, see none of those changes


 11   really coming up again is encouraging but, again,


 12   it is only two-year data for a large number of


 13   patients, but not a tremendous amount, so we will


 14   continue to monitor in subsequent years.


 15             DR. JOHNSON:  My second question is on


 16   your second to last slide, just for clarification.


 17             DR. COLONNO:  My second to last slide?


 18             DR. JOHNSON:  Yes, from your earlier


 19   presentation.  It was called summary of viral


 20   resistance data at week 48.  So, just to clarify,


 21   and I think part of this got answered, the title is


 22   week 48 but the bottom data are presented on two




  1   patients who had greater than 76 weeks.


  2             DR. COLONNO:  Those two patients were from


  3   the Phase II study.  They are not included here;


  4   they were Phase II.


  5             DR. JOHNSON:  So, they are different than


  6   the two on this slide that are on the bottom?


  7             DR. COLONNO:  These two are from the Phase


  8   III evaluation.


  9             DR. JOHNSON:  At week 48?


 10             DR. COLONNO:  At week 48.


 11             DR. JOHNSON:  And that is different than


 12   the other two patients you described with virologic


 13   rebound resistance?


 14             DR. COLONNO:  That is correct.  One was in


 15   the 015 study which was a transplant study, and the


 16   other one was in 014.


 17             DR. JOHNSON:  But they appear to select


 18   the same signature mutations?


 19             DR. COLONNO:  They select the same


 20   signature mutations.  Those three mutations appear


 21   to be the key primary resistance markers for


 22   entecavir.


 23             DR. ENGLUND:  Dr. Sherman?


 24             DR. SHERMAN:  The presentation indicated


 25   that phosphorylation was required for this product.




  1   Could you comment on any data you have regarding


  2   interactions with anti-retrovirals that also


  3   require phosphorylation in vitro?  I know you have


  4   limited in vivo HIV-positive patients, but is there


  5   any pharmacokinetic analysis and any issues of


  6   changes in resistance to HIV or susceptibility


  7   because of the interaction?


  8             DR. MORGAN MURRAY:  I will let Dr. Colonno


  9   follow up on that.


 10             DR. COLONNO:  We have done an extensive


 11   analysis of the interactions because it is a


 12   nucleoside analog and there are many nucleoside


 13   analogs that are used in HIV, interactions based on


 14   the phosphorylation patterns of these various


 15   combinations.  What I can tell you is that because


 16   the concentration of entecavir is so low relative


 17   to other nucleoside analogs and the efficiency is


 18   so high, when one does in vitro cell culture


 19   combination studies to look for the effect of




  1   entecavir on the antiviral potency of the HIV


  2   nucleoside analogs, or in the opposite direction in


  3   the presence of the HIV and RTIs and does it have


  4   an impact on entecavir activity, we find, using


  5   concentrations of both sets of compounds up to five


  6   times their C                                                     max,

clinical Cmax, we see no


  7   interactions whatsoever; no antagonism; and no


  8   decrease in the activity.  Again, that is a big


  9   plus for entecavir because entecavir is very


 10   selective for hepatitis B and so it literally also


 11   can be used in a co-infected patient but not having


 12   to worry about any kind of selective pressure on


 13   HIV.


 14             DR. ENGLUND:  I am going by the order that


 15   I saw the hands come up, which may be wrong, and we


 16   are only going to have time for about four or five


 17   more questions.  But the first question was Mr.


 18   Grodeck's.


 19             MR. GRODECK:  In terms of marketing


 20   antivirals, one of the biggest games I have seen


 21   pharmaceutical companies play is the sequencing


 22   game--my drug should come before your drug.  In




  1   your description of the resistance profile of


  2   entecavir, it seems to me that you are setting up


  3   the drug to be positioned as a first-line


  4   treatment.  Is that your position?  How does it fit


  5   in terms of the range of other treatments available


  6   to chronic hepatitis B patients today?


  7             DR. MORGAN MURRAY:  I will ask Dr.


  8   Dienstag to comment on how entecavir might fit into


  9   current treatment guidelines and the physicians'


 10   armamentarium.  I will just remind you, from our


 11   data, that we have demonstrated that entecavir is


 12   superior to lamivudine.  We have substantial


 13   benefits in both nucleoside-naive and


 14   lamivudine-refractory patients.


 15             DR. DIENSTAG:  Jules Dienstag,


 16   Massachusetts General Hospital.  I think if we


 17   consider hepatitis B a viral disease, then the drug


 18   that suppresses HBV most profoundly is likely to


 19   have the most benefit.  That has been shown in this


 20   study for histology, biochemical markers and


 21   especially for the profundity of suppression of HBV


 22   DNA.  In almost 90 percent of patients you can




  1   achieve an undetectable level of HBV DNA, which no


  2   other antiviral comes close to at this point.


  3             So, it is not unreasonable to suggest that


  4   this would be a first-line therapy.  When you add


  5   the resistance profile and when you consider the


  6   potential that, for example, a drug like lamivudine


  7   sets you up for lamivudine resistance in the future


  8   and also sets you up for resistance to any other


  9   nucleoside, it makes sense to start with this drug.


 10   It is a very reasonable suggestion.


 11             DR. ENGLUND:  Dr. Paxton, did you get your


 12   question answered?


 13             DR. PAXTON:  Yes, it was.  Thank you.


 14             DR. ENGLUND:  Dr. Wood, or were you first,


 15   Dr. Seeff or Dr. Schwarz?


 16             DR. SCHWARZ:  I have two questions


 17   relative to future applications of entecavir.  You


 18   said that in the animal carcinogenicity models in


 19   the organ involved with the tumor there were


 20   ETV-induced dNTP pool perturbations.  In either the


 21   animal studies or in the human studies, was there


 22   evidence of peripheral blood lymphocytes--the same




  1   phenomenon occurring in peripheral blood


  2   lymphocytes that might be a useful non-invasive


  3   surrogate marker for malignant potential?


  4             Then the second question is I assume in


  5   these lifetime exposure studies that the drug was


  6   not started in the immediate newborn period.  So,


  7   at what age of the animal was it started, and can


  8   you make an educated guess about the human


  9   equivalent age?


 10             DR. MORGAN MURRAY:  I will ask Dr.


 11   Lehman-McKeeman to address the data that we have in


 12   animals around dNTP pool perturbations and also


 13   about the rodent studies.  I will just comment that


 14   we do not have any human data around dNTP pool


 15   perturbations.  As Dr. Lehman-McKeeman will


 16   describe, these perturbations in animals occur at


 17   much higher doses than we administer in humans.


 18             DR. LEHMAN-MCKEEMAN:  I will actually


 19   address the second question first for you.  The


 20   studies that are conducted in rodents basically


 21   start when they are approximately 5-6 weeks of age.


 22   For perspective, that is when a rodent reaches




  1   sexual maturity.  So, in a 2-year life span, if I


  2   had to extrapolate, I will just say at sexual


  3   maturity so it would be roughly teenage.


  4             To your first question about the dNTP


  5   pools, in the work that we did we specifically


  6   looked at target organ effects related to


  7   carcinogenicity.  So, we specifically looked at the


  8   liver and we don't have any data on another system.


  9   Those analyses are actually quite, I will say,


 10   difficult to do, as it were, simply because the


 11   pools themselves are really quite fleeting.  So, it


 12   really is prohibitive for us to collect more then


 13   one sample and we targeted the liver.  However,


 14   what we know, based on the work we have done, is


 15   that that is a high dose phenomenon.  So, at


 16   dosages where we saw carcinogenic activity we saw


 17   perturbations in pools, and at a dose below a


 18   carcinogenic effect we did not seriously disrupt


 19   pools.  So, I think it is a function, again, of the


 20   maximum tolerated dosage that we are administering


 21   in the carcinogenicity studies.


 22             DR. ENGLUND:  Dr. Wood?


 23             DR. WOOD:  My question has to do with


 24   analysis of rates of malignant neoplasms according


 25   to ethnicity.  This is related to the fact that I




  1   believe I read that Asians have a higher


  2   pharmacokinetic exposure to entecavir and I was


  3   wondering whether or not an analysis had been done


  4   on that basis.


  5             DR. MORGAN MURRAY:  I will ask Dr.


  6   Brett-Smith to come up again.  I will try and spare


  7   her voice a bit and comment that while we have not


  8   seen PK differences on the basis of race in


  9   particular, the differences that we do detect are


 10   related to weight more than to race.  Dr.


 11   Brett-Smith, on the malignancies?


 12             DR. BRETT-SMITH:  At this point we have


 13   chosen not to look at any subpopulations in terms


 14   of the overall rates in malignancies because the


 15   total numbers remain low, and we believe that the


 16   rates would be sort of unreliably variable.  That


 17   may become an option later in terms of the


 18   pharmacovigilance.


 19             DR. ENGLUND:  Dr. Seeff?


 20             DR. SEEFF:  I thought that the efficacy


 21   data that were presented were fairly impressive but


 22   there are a couple of questions that I just need


 23   some clarification on.  Perhaps you presented them


 24   and I missed them.


 25             The primary endpoint for your study was




  1   histologic using the Knodell score.  I gather that


  2   this is not unusual; this is fairly routine.  Is


  3   this the 18-point HAI score?


  4             DR. MORGAN MURRAY:  Yes.


  5             DR. SEEFF:  In other words, the drop for


  6   example from 10 points to 8 points would represent


  7   an endpoint having been achieved.


  8             DR. MORGAN MURRAY:  Correct.


  9             DR. SEEFF:  What was the average drop?  Do


 10   you know what the average decline in points was,


 11   just to get a sense of how much improvement there


 12   was in histology?  Do you have those data by any


 13   chance?  I mean, you have the percentage of people


 14   who achieved a 2-point reduction, but what I am


 15   interested in knowing is by how much of a


 16   reduction.


 17             DR. MORGAN MURRAY:  Dr. Brett-Smith?


 18             DR. BRETT-SMITH:  Yes, we do have data and


 19   I can present it for you if we can show slide 2-66.


 20   Overall, in the naive patients it was approximately


 21   a 4-point drop in the mean score, and in the


 22   refractory patients it was approximately a 3-point


 23   drop.


 24             DR. SEEF:  That is fine.  Thank you.  The


 25   second thing is your secondary endpoints,




  1   essentially a reduction in HBV viral load and


  2   normalization of ALT, do you have a composite score


  3   taking into account the virologic, histologic,


  4   biochemical reduction?  Do we have a score of using


  5   those three parameters?


  6             DR. MORGAN MURRAY:  Dr. Brett-Smith?


  7             DR. SEEFF:  And is it the same between


  8   entecavir and lamivudine for example?


  9             DR. BRETT-SMITH:  If I can just repeat the


 10   factors that you are interested in, you are


 11   interested in combining histology with virology--


 12             DR. SEEFF:  And with biochemical response.


 13             DR. BRETT-SMITH:  With ALT.


 14             DR. SEEF:  ALT.


 15             DR. BRETT-SMITH:  Those three.  We have


 16   looked at a number of ways of combining virology


 17   with ALT.  I will ask my colleagues to confirm


 18   whether we have yet completed the analysis


 19   combining with histology.  I do not have that at


 20   this time.


 21             DR. MORGAN MURRAY:  So, we will confer


 22   during the break and see if we can quickly pull


 23   something together to answer that.


 24             DR. ENGLUND:  Last question, Dr. So?


 25             DR. SO:  There is a common belief by many




  1   clinicians that, you know, if you have e-antigen


  2   seroconversion you pretty much, you know, have a


  3   good response and you might be cured.  So, I notice


  4   that earlier in this handout to us, the committee


  5   members, you did describe some follow-up on the


  6   patients in your study 22 where they have so-called


  7   complete response.  Some of the patients were taken


  8   off drugs.  Do you have two-year follow-up


  9   information regarding how many of those patients


 10   have so-called sustained response and what




  1   sustained response means?


  2             DR. MORGAN MURRAY:  I will ask Dr.


  3   Brett-Smith to comment, and I need to make the


  4   statement that in the NDA and the NDA update we


  5   only had the 48-week data so, again, these data


  6   have not been submitted for review.


  7             DR. BRETT-SMITH:  I heard a two-part


  8   question there.  Let me just clarify.  I heard


  9   first for patients who, at the end of year one,


 10   went off dosing--you were interested in the


 11   sustained response off treatment.


 12             DR. SO:  Right.


 13             DR. BRETT-SMITH:  Also, did I hear an


 14   interest in what happens to the portion of partial


 15   responders who have a virologic response--


 16             DR. SO:  No, I am just interested in your


 17   so-called complete responders.


 18             DR. BRETT-SMITH:  Let me first summarize


 19   for you the design of the studies at the week 52


 20   endpoint.  A clinical decision was made based on


 21   laboratory results from week 48 as to the


 22   management of the patient, which was simply a




  1   management algorithm that was modeled on guidance


  2   at the time and it differs for each population.  In


  3   the e-antigen positive population we required, in


  4   order to go off therapy, that patients have lost


  5   e-antigen and have an HBV DNA less than the bDNA


  6   assay level of detection, so less than 0.7.  In the


  7   e-negative population patients had to meet the


  8   virologic requirement of bDNA less than LOQ, and


  9   they had to have an ALT less than 1.25 times the


 10   upper limit of normal.  In the refractory antigen


 11   positive patients we again required that the


 12   patients achieve the virologic endpoint in


 13   association with e-loss.  In that last group there


 14   were very small numbers of patients going off


 15   treatment, therefore, we will not discuss that


 16   further; the numbers were substantially small.


 17             With respect to the two naive patient


 18   populations, if we could show slide 2-380, the


 19   studies were designed to follow people out to 24


 20   weeks of off-treatment follow-up.  If during that


 21   time patients went on alternative therapy or into


 22   the rollover study they were considered failures to




  1   maintain that endpoint.  These represent the


  2   respective percentages in the naive e-antigen


  3   positives on the left, 82 percent for entecavir and


  4   73 percent for lamivudine, who maintained their


  5   study-defined response rate at week 24 off


  6   treatment.  Likewise, in the naive e-antigen


  7   negative population we had 48 percent for entecavir


  8   and 35 percent for lamivudine.


  9             DR. SO:  But I don't think you answered my


 10   question.  How many of those patients who were off


 11   treatment actually were followed up, like actually


 12   48 weeks off treatment, are still off treatment?


 13   You know, it could be very misleading for a lot of


 14   clinicians when you say sustained response, not


 15   knowing, you know, for how many of those patients


 16   actually their histologic improvement was


 17   sustained?  Was the virologic improvement sustained


 18   at 48 weeks?  So, I feel that the 24-week off


 19   treatment, so-called sustained response, could be


 20   misleading.


 21             DR. BRETT-SMITH:  Point taken, 24 weeks is


 22   what had been agreed upon with regulatory




  1   authorities in the design of the original study.


  2   All patients are encouraged to enroll on completion


  3   of the original study in the 049 long-term rollover


  4   study which remains currently enrolling at this


  5   time and has not undergone its initial analysis.


  6             DR. SO:  Just one last question, how does


  7   your company plan to talk to those clinicians who


  8   say, you know, if my patient seroconverted--these


  9   are naive patients before treatment, if they


 10   seroconverted I am planning to stop the treatment?


 11   How do you plan to advise those clinicians?


 12             DR. MORGAN MURRAY:  Our current proposed


 13   labeling reflects how the studies were conducted,


 14   and in that regard, for those patients who were


 15   determined to be responders therapy was stopped at


 16   48 weeks and they were monitored.  Patients who


 17   were partial responders continued on therapy.  Our


 18   current trials cannot define the definitive


 19   duration of dosing for entecavir, which is in


 20   general in flux for HBV therapy.  Dr. Dienstag, do


 21   you have any further comments?


 22             DR. DIENSTAG:  Jules Dienstag, Mass.




  1   General Hospital.  No one really knows what the


  2   sustained responsiveness or the durability of an


  3   e-antigen response is, but in the experience we


  4   have for interferon, lamivudine and adefovir if a


  5   person maintains that serologic response for 6


  6   months after stopping therapy the durability is 80


  7   percent.  That is the experience in Asia and in the


  8   West.  I assume that that will be repeated in this


  9   experience but that remains to be seen.


 10             DR. ENGLUND:  Thank you, everyone, for


 11   asking questions, answering questions.  We will now


 12   take a 15-minute break.  We will be back at 10:25


 13   to resume the FDA portion of this morning's


 14   presentation.


 15             [Brief recess]


 16             DR. ENGLUND:  Thank you.  Welcome back


 17   from coffee.  We are now going to have an FDA


 18   presentation led by Dr. James Farrelly, the


 19   pharmacology team leader, and he will begin his


 20   presentation.


 21                         FDA Presentation


 22                      Carcinogenicity Issues


 23             DR. FARRELLY:  Good morning.  My name is


 24   Jim Farrelly.  I am the pharmacology team leader in


 25   the Division of Antiviral Drugs.




  1             Today our purpose is to present some of


  2   the data relating to the genetic toxicity and the


  3   animal carcinogenicity of entecavir.  Entecavir is


  4   a nucleoside analog and, as such, is a member of a


  5   class of molecules which are in general expected to


  6   be genetically toxic.  Its 5-prime hydroxyl can be


  7   phosphorylated to the nucleotide triphosphate and


  8   as a guanosine triphosphate analog can be


  9   incorporated into the growing DNA chain.  It has


 10   the three-prime hydroxyl group and is, therefore,


 11   not an obligate chain terminator as are many other


 12   nucleoside analogs.


 13             However, after incorporation of entecavir


 14   into the growing DNA chain, it halts DNA synthesis


 15   after the addition of a small number of subsequent


 16   bases.  Its mechanism of action is essentially as a


 17   chain terminator, which is consistent with its


 18   being a clastogenic compound or having the ability


 19   to break chromosomes.  Indeed, entecavir has been




  1   shown to be clastogenic in an in vitro assay in


  2   human lymphocytes.


  3             It is negative in a number of genetic


  4   toxicity tests both in vitro and in vivo.  These


  5   include an Ames test, an in vitro assay in Chinese


  6   hamster ovary cells, in the Syrian hamster embryo


  7   cell transformation assay, and in an in vivo rat


  8   micronucleus assay, and in an unscheduled DNA


  9   synthesis assay.  In general, most of the battery


 10   of genotoxicity tests can be used only for hazard


 11   identification.  They are not used for risk


 12   assessment but have indicated that entecavir can be


 13   a possible genetic toxicity hazard.


 14             In an effort to place the results of the


 15   genetic toxicity studies into perspective, one can


 16   compare the outcome of the studies used to evaluate


 17   entecavir with the outcome of the studies used to


 18   evaluate the genetic toxicity of the three entities


 19   approved for the treatment of hepatitis B.  The


 20   three are adefovir, lamivudine and interferon.


 21             Adefovir is a nucleotide analog rather


 22   than a nucleoside analog, and was found to be




  1   mutagenic and to induce chromosomal aberrations in


  2   two in vitro genetic toxicology studies.


  3   Lamivudine, or 3GC, is a nucleoside analog and was


  4   found to be mutagenic in two in vitro assays as


  5   well.  Interferon was not an active genetic toxin.


  6   Since it is a protein one would not expect


  7   interferon to be positive in the screening battery


  8   used to test for genetic toxicity.  However, most


  9   of the nucleoside analogs approved as antiviral


 10   antigens are positive in genetic toxicology


 11   batteries of tests.


 12             Now, as is usual for a drug that is going


 13   to be administered chronically to humans, entecavir


 14   was evaluated in two-year carcinogenicity studies


 15   in rats and mice.  The design and outcome of the


 16   study in rats can be seen in the next slide where


 17   the data for male rats are shown.


 18             Entecavir was administered by gavage to


 19   rats at four doses, 0.003, 0.02, 0.2 and 1.4


 20   mg/kg/day.  They were administered for 96 weeks.


 21   There were two identical vehicle controls in the


 22   study.  The doses in male rats represent the human




  1   equivalent exposure of much less than 1, 0.3, 5 and


  2   35 times the clinical dose at the 1 mg proposed


  3   clinical dose, which you see under MHD.


  4             In male rats at an exposure 35-fold that


  5   in the clinic entecavir caused the appearance of a


  6   low level but significant incidence of brain


  7   gliomas.  A no-level of tumors was seen, or very


  8   low level, at 5-fold the exposure, and below no


  9   significant number of tumors was seen in the study.


 10             The next slide shows the results in female


 11   rats.  As can be seen from this slide, entecavir


 12   was administered at doses of 0.01, 0.06, 0.4 or 2.6


 13   mg/kg/day for two years.  Dosing was again by


 14   gavage and drug groups as well as two identical


 15   vehicle control groups were treated for 104 weeks.


 16   As can be seen in the slide, entecavir again


 17   induced the appearance of brain gliomas at the high


 18   dose.  It also induced the appearance of skin


 19   fibromas at the high dose, and increased the


 20   incidence of liver tumors at the high dose from 1-8


 21   adenomas and from 0-3 carcinomas.  The exposure to


 22   entecavir at the high dose in which these tumors




  1   were seen was approximately 24-fold higher for


  2   females than that measured in the clinic at the 1


  3   mg dose.


  4             Mention should be made regarding the


  5   exposure multiples at which tumors were seen in the


  6   study.  Although a multiple of 24 in exposure is a


  7   high multiple of the human exposure, it should be


  8   remembered that there were no significant induction


  9   or increase in tumors at the 4-fold for females and


 10   5-fold level for males.  The real cutoff,


 11   therefore, is somewhere between the high dose and


 12   the next lower dose, and the no-observed effect for


 13   tumors was at the 4- and 5-fold human dose.


 14             The results of the mouse carcinogenicity


 15   study were more complicated.  In the next slide it


 16   is shown that male mice were treated in a similar


 17   manner as were the rats.  The doses of entecavir


 18   used actually in both the males and the females


 19   were the same on a milligram per kilogram per day


 20   basis. The doses were 0.004, 0.04, 0.4 and 4


 21   mg/kg/day.


 22             As seen here, entecavir caused a




  1   dose-related increase in common bronchoalveolar


  2   adenomas in the males, significant at the three


  3   higher doses.  The lowest of the three doses


  4   produced an exposure only 3-fold higher than the


  5   clinical exposure.  Also increased in the males was


  6   the incidence of hepatocellular carcinoma at the


  7   high dose, going from 1 in one of the controls to 8


  8   at the high dose.  The exposure in the latter case


  9   was 42-fold higher than the clinical exposure.  For


 10   the hepatocellular carcinomas no increase was seen


 11   at an exposure 40-fold the clinical dose, very low;


 12   not significant.


 13             The next slide shows the female mice in


 14   which entecavir induced a significant increase in


 15   the lung tumors only at the high dose, giving an


 16   exposure in the animal study 40-fold the exposure


 17   in the clinic.  There was no significant increase


 18   at 11-fold the exposure.  Also in female mice there


 19   was an increase in ovarian and uterine vascular


 20   tumors, again at the high dose.  If one combined


 21   all the vascular tumors, as is commonly done, there


 22   was a significant increase in combined hemangiomas




  1   and hemangiosarcomas at the high dose.


  2             We have heard the sponsor make a good case


  3   for the proposition that the pulmonary tumors seen


  4   in the mouse are mouse specific.  No cellular


  5   proliferation was seen in the lungs of rats and no


  6   lung tumors in rats, as well as no cellular


  7   proliferation in the dog and monkey studies.  If,


  8   indeed, the tumors were mouse specific the outcome


  9   would be that the only tumors seen in the two


 10   studies were at the high dose only.


 11             Again, putting the results of the


 12   carcinogenicity studies into perspective with the


 13   other approved regimens for hepatitis B, no


 14   carcinogenicity studies were carried out with


 15   interferon.  Studies were carried out with adefovir


 16   and lamivudine; they were not carcinogenic.


 17   However, because of kidney toxicity in the


 18   carcinogenicity studies, the exposures of the


 19   animals in the adefovir studies relative to the


 20   clinical exposures were 10-fold for mice and 4-fold


 21   for rats.  The maximum tolerated dose cannot go any


 22   higher than those.  So, if entecavir was examined




  1   only at those exposures, it would have been


  2   positive only for the lung tumors in mice and for


  3   no other tumor types.


  4             The exposures in lamivudine studies were


  5   high relative to the exposures in the clinic, up to


  6   34-fold in the mice and 200-fold in the rats.  At


  7   those exposures the entecavir results would have


  8   been at least identical to those which we have seen


  9   in these studies.  However, many nucleoside analogs


 10   approved as antivirals have been positive in


 11   carcinogenicity studies.


 12             The results of the two carcinogenicity


 13   studies were presented to the CDER Executive


 14   Carcinogenicity Assessment Committee, which we call


 15   the executive CAC, as well as to the full CAC for


 16   evaluation.  The CDER CAC committees were formed in


 17   the late 1980s to examine the protocols of


 18   carcinogenicity studies, as well as to examine the


 19   outcomes of the same studies.  The committees were


 20   founded so that the interpretation of the


 21   carcinogenicity data would not be inconsistent


 22   depending on which division reviewed them.  Two




  1   committees exist, the executive CAC, as I said, and


  2   the full CAC.


  3             The executive CAC consists of four


  4   members, the associate director for


  5   pharmacology/toxicology in the center; one


  6   permanent expert in the evaluation of


  7   carcinogenicity studies; the supervisor whose


  8   division is presenting the data; and another


  9   supervisor from another division chosen on a


 10   rotating roster.  The executive CAC meets every


 11   Tuesday and evaluates a great number of protocols


 12   and studies in a year, usually somewhere between


 13   150 and 200 either protocols or carcinogenicity


 14   studies in a year.


 15             The next slide shows the makeup of the


 16   full CAC which is empowered to review the studies


 17   when members of the executive CAC cannot


 18   unanimously agree on the interpretation of the


 19   data, or when requested by the sponsor of the drug.


 20   The full CAC consists of the associate director for


 21   the center; three associate directors for the


 22   offices; and each of the supervisors from the




  1   individual divisions in the center.  The full CAC


  2   is a fairly large committee and meets only rarely.


  3   In fact, the meeting for this drug was the first


  4   one in over a year for the full CAC.


  5             Both the executive CAC and the full CAC


  6   agreed that the tumors seen in the studies were


  7   probably relevant to a safety evaluation for


  8   humans.  The full CAC in general voted that the


  9   tumors seen in the carcinogenicity studies were


 10   relevant to human safety evaluation.


 11             The questions asked of the committee were


 12   does the CAC agree that the lung tumors in mice


 13   were relevant to human safety evaluation?  The


 14   committee voted yes, 16; no/probably not, 2; and 2


 15   answered they don't know.


 16             Does the CAC agree that, one, the liver


 17   tumors in male mice and, two, the vascular tumors


 18   in female mice are relevant to human safety


 19   evaluation?  The vote was 17 yes; 3 no.


 20             Does the CAC agree that, one,


 21   hepatocellular adenomas and carcinomas in female


 22   rats, two, the skin fibromas in female rats and,




  1   three, the brain gliomas in male and female rats


  2   are relevant to human safety evaluation?  The


  3   answer was yes, 17; 3 no.


  4             Now, in our division many carcinogenic


  5   nucleoside and nucleotide analogs have been


  6   approved for the treatment of viral diseases.


  7   Among these are ganciclovir which gives rodent


  8   tumors at very low doses relative to the human


  9   exposure; zidovudine; abacavir and cidofovir.


 10   Cidofovir causes palpable mammary adenocarcinomas


 11   in rats after as few as six weekly doses and is


 12   closely related in chemical structure to adefovir.


 13   Some of the reverse transcriptase inhibitors as


 14   well as the HIV protease inhibitors are positive


 15   for animal carcinogenicity.  Other drugs, such as


 16   8-methoxy psoralen, which has been approved for the


 17   treatment of psoriasis, are carcinogens.  In fact,


 18   this compound has been identified as a human


 19   carcinogen in epidemiology studies.  Dr. Linda


 20   Lewis will continue the division presentation.


 21   Thank you.


 22                         Clinical Issues


 23             DR. LEWIS:  Good morning.  My name is


 24   Linda Lewis, and I was the lead clinical reviewer


 25   for the entecavir review team.  I would like to




  1   give you the perspectives of the entire team on our


  2   review of entecavir for the treatment of chronic


  3   hepatitis B.


  4             My presentation is outlined in this slide.


  5   First I will go over a little bit of the


  6   development program for entecavir, which you have


  7   heard presented earlier by Bristol-Myers Squibb.


  8   Then I would like to go over the results of our


  9   reviews of the efficacy, safety and


 10   virology/resistance data that were contained in the


 11   NDA submission.  At that point I will turn my


 12   discussion to an assessment of the risk-benefit of


 13   entecavir and the applicant's proposed


 14   pharmacovigilance plan.  I will end the


 15   presentation with a preview of the questions that


 16   we would like the advisory committee to consider


 17   later this afternoon.


 18             As you heard this morning, the treatment


 19   options for chronic hepatitis B are somewhat




  1   limited.  Interferon was approved for treatment of


  2   hepatitis B in 1992.  Its requirement for


  3   parenteral administration and its significant side


  4   effect profile have somewhat limited its use.


  5   Lamivudine was the first effective oral therapy,


  6   and it was approved in 1998.  Its usefulness has


  7   been limited by the predictable emergence of


  8   resistance in relatively short periods of time.  A


  9   most recent addition, adefovir, was approved in


 10   2002.  It has known renal toxicity that may limit


 11   its use in some populations.


 12             The entecavir development program included


 13   a diverse patient population.  The clinical studies


 14   were drawn from multinational sites in North and


 15   South America, Europe and Asia.  Among these


 16   studies, patients from the United States made up


 17   about 10 percent of the pivotal trials.  The


 18   entecavir studies were made up of about 20 percent


 19   women.  There was a good mix of Asian and non-Asian


 20   patients in the populations.  However, Black or


 21   African American patients were under-represented in


 22   the clinical trials, making up only 2 percent of




  1   the pivotal studies.  The development program


  2   enrolled patients at different stages of disease


  3   and treatment.  Although there is a study in


  4   progress, the data were insufficient to review the


  5   use of entecavir in patients with decompensated


  6   liver disease during this review cycle.


  7             BMS submitted study reports and electronic


  8   data sets for the four key studies that they have


  9   mentioned in their presentation earlier.  To go


 10   over these again, study 022 was the Phase III study


 11   enrolling nucleoside-naive, e-antigen positive


 12   adults.  Study 027 enrolled nucleoside-naive


 13   e-antigen negative adults.  Both of these studies


 14   used a dose of 0.5 mg of entecavir given once


 15   daily.  Study 026 enrolled patients with persistent


 16   HBV viremia despite lamivudine treatment.  These


 17   are termed lamivudine-refractory subjects.


 18   Patients in this study were e-antigen positive and


 19   received a dose of 1 mg of entecavir given once


 20   daily.


 21             In order to expand the safety database for


 22   lamivudine-refractory patients we included in our




  1   review patients from study 014, the dose-finding


  2   study in that patient population, and used the


  3   cohorts that received either 1 mg of entecavir or


  4   the standard dose of lamivudine. As has been


  5   pointed out, all of the pivotal trials were


  6   compared to the standard dose of currently approved


  7   lamivudine.


  8             For all of the Phase III studies, studies


  9   022, 027 and 026, the primary endpoint was the


 10   overall histologic improvement in liver biopsy


 11   after 48 weeks of treatment.  This histologic


 12   improvement was defined as greater than or equal to


 13   a 2-point decrease in the Knodell necroinflammatory


 14   score, with no worsening in the Knodell fibrosis


 15   score compared to the baseline biopsy.  A series of


 16   secondary endpoints were also evaluated and


 17   included a number of virologic, serologic,


 18   biochemical and composite endpoints.


 19             The applicant also submitted data from


 20   several important studies in special populations.


 21   These included study 015.  This was a small pilot


 22   trial in post-liver transplant patients who had




  1   recurrent hepatitis B.  Study 038 enrolled a cohort


  2   of HIV/HBV co-infected patients.  Study 048


  3   compares the use of entecavir to adefovir in


  4   patients we decompensated liver disease.  This


  5   study is still enrolling and the data were not


  6   sufficient for us to conduct any meaningful interim


  7   analysis during this review cycle.  In these


  8   studies histologic endpoints were not used.  They


  9   relied on a series of virologic, serologic and


 10   biochemical endpoints.


 11             Now I would like to turn to the efficacy


 12   review of entecavir.  You will probably notice in


 13   these slides that many of our slides look very


 14   similar to those presented by the applicant earlier


 15   this morning.


 16             The FDA statistical review, conducted by


 17   Dr. Tom Hammerstron, confirmed the applicant's


 18   primary efficacy analysis.  A review of secondary


 19   efficacy analyses, using the virologic, serologic


 20   and biochemical endpoints, was also in agreement


 21   with BMS's conclusions.  Multiple sensitivity


 22   analyses and subgroup analyses were performed and




  1   all supported the primary analysis.


  2             This table displays the results of the


  3   primary efficacy analysis and some of the other


  4   histologic endpoints for each of the Phase III


  5   studies, study 022, 027 and 026.  The top line of


  6   the study shows the primary analysis, the overall


  7   histologic improvement after 48 weeks.  As you can


  8   see, in each of the three studies entecavir


  9   performed better than lamivudine in each study, as


 10   highlighted--these are supposed to be pink I don't


 11   know exactly how it projects.


 12             The next two lines display the two


 13   individual components that make up the overall


 14   histologic improvement score.  Again, you can see


 15   that patients receiving entecavir achieved that


 16   endpoint significantly more often than those who


 17   received lamivudine.  The last line of the study


 18   shows the secondary histologic endpoint of the


 19   Ishak fibrosis score.  This is another method of


 20   evaluating liver histology.  In this analysis


 21   entecavir was superior to lamivudine only in the


 22   lamivudine-refractory study, study 026.  In the




  1   treatment-naive studies the proportion of patients


  2   achieving an improvement in their Ishak fibrosis


  3   score was similar across the treatment arms.


  4             This table displays some of the


  5   sensitivity analyses that were done by our


  6   statistical reviewers.  The top line is a carryover


  7   from the previous slide and shows the primary


  8   analysis.  In the primary analysis the only


  9   subjects who had evaluable baseline biopsies were


 10   included in the analysis.  Subjects with missing or


 11   inadequate week 48 biopsies were counted as


 12   treatment failures.  The sensitivity analyses, done


 13   by Dr. Hammerstron, included a series of different


 14   methods to impute the missing data for each of the


 15   Phase III studies.  I am going to show you just two


 16   of the many analyses that he did.


 17             In FDA sensitivity analysis C, missing or


 18   inadequate baseline or week 48 biopsies were


 19   excluded from the analysis.  In this analysis, in


 20   study 022, the results were similar between


 21   entecavir and lamivudine and this is due primarily


 22   to the fact that more patients in the lamivudine




  1   arm were excluded because they did not have week 48


  2   biopsies.  In the other two studies, again,


  3   entecavir achieved the primary endpoint


  4   significantly more often than patients who received


  5   lamivudine.


  6             In sensitivity analysis D, this analysis


  7   includes all patients who were treated, not just


  8   those who had evaluable biopsies, but missing or


  9   inadequate week 48 biopsies were still counted as


 10   failures.  Although the numbers are lower for all


 11   of these analyses, the difference between entecavir


 12   and lamivudine remains evident in each of the three


 13   pivotal trials.


 14             This slide displays some of the analyses


 15   of secondary virologic, serologic and biochemical


 16   endpoints for the three pivotal trials.  Again, the


 17   significant values are highlighted in the pink


 18   cells.  In the Phase III studies a greater


 19   proportion of patients receiving entecavir than


 20   lamivudine achieved an HBV DNA PCR less than 400


 21   copies/mL.  Similarly, patients who received


 22   entecavir had a greater mean log decrease in PCR




  1   from baseline to week 48 than did patients who


  2   received lamivudine.  In the two studies that


  3   included e-antigen positive patients, studies 022


  4   and 026, the proportions of patients who had a


  5   seroconversion were roughly the same.  You will


  6   notice that in study 026 a relatively small number


  7   of patients actually met this criteria.  Finally,


  8   in terms of the proportion of patients who reached


  9   a normalized ALT, again, entecavir was shown to be


 10   superior to lamivudine in each of the three pivotal


 11   trials.


 12             We also conducted a number of subgroup


 13   analyses for baseline covariates of demographic or


 14   disease characteristics.  The treatment effect of


 15   the primary endpoint was comparable for the


 16   covariates gender, race, age, geographic region,


 17   HBV subtype, baseline ALT, baseline bDNA or PCR, or


 18   by prior treatment with lamivudine or interferon.


 19             Similarly, more limited subgroup analyses


 20   were performed to assess some of the key secondary


 21   endpoints.  The treatment effect measured by the


 22   proportion of patients of subjects who achieved a




  1   normalization of HBV DNA or those who achieved a


  2   viral load less than 400 copies/mL at weeks 24 and


  3   48 were similar according to gender, race and age.


  4             This slide displays a composite of the


  5   subgroup analysis for the Phase III studies.  I


  6   really show you this for pattern recognition more


  7   than to display any kind of specific results.  This


  8   slide plots the mean difference in treatment effect


  9   for the primary endpoint between entecavir and


 10   lamivudine, with 95 percent confidence intervals,


 11   for the three pivotal trials.  This cluster


 12   rep