FOOD AND DRUG ADMINISTRATION













                               Volume II

















                         Wednesday, May 4, 2005


                               8:30 a.m.









                CDER Advisory Committee Conference Room

                           5630 Fishers Lane

                          Rockville, Maryland





      Charles Cooney, Ph.D., Chair

      Hilda F. Scharen, M.S., Executive Secretary




      Patrick P. DeLuca, Ph.D.

      Paul H. Fackler, Ph.D., Industry Representative

      Michael S. Korczynski, Ph.D.

      Gerald P. Migliaccio (Industry Representative)

      Kenneth R. Morris, Ph.D.

      Marc Swadener, Ed.D. (Consumer Representative)

      Cynthia R.D. Selassie, Ph.D.

      Nozer Singpurwalla, Ph.D.




      Carol Gloff, Ph.D.

      Arthur Kibbe, Ph.D.

      Thomas P. Layloff, Jr., Ph.D.

      Marvin C. Meyer, Ph.D.




      Gary Buehler, R.Ph.

      Kathleen A. Clouse, Ph.D.

      Jerry Collins, Ph.D.

      Ajaz Hussain, Ph.D.

      Robert Lionberger, Ph.D.

      Robert O'Neill, Ph.D.

      Keith O. Webber, Ph.D.

      Helen Winkle

      Lawrence Yu, Ph.D.



                            C O N T E N T S



      Call to Order

                Charles Cooney, Ph.D.                            5


      Conflict of Interest Statement

                Hilda Scharen, M.S.                              5


      Parametric Tolerance Interval Test for Dose

        Content Uniformity


         Current Update on the Working Group

                Robert O'Neill, Ph.D.                            8


      Quality-by-Design and Pharmaceutical Equivalence


         Topic Introduction

                Ajaz Hussain, Ph.D.                             10


         Using Product Development Information to

           Extend Biopharmaceutics Classification

           System-based Biowaivers

                Ajaz Hussain, Ph.D.                             30


         Using Product Development Information to Address

           the Challenge of Highly Variable Drugs

                Lawrence Yu, Ph.D.                              79


         Using Product Development Information to Support

           Establishing Therapeutic Equivalence of

           Topical Products

                Robert Lionberger, Ph.D.                       140


         Summary of Plan

                Ajaz Hussain, Ph.D.                            169


      Committee Discussion and Recommendations                 192


      Criteria for Establishing a Working Group for

        Review and Assessment of OPS Research Programs


         CBER Peer Review Process for


                Kathleen A. Clouse, Ph.D.                      221


         CDER Peer Review Research

                Jerry Collins, Ph.D.                           243



                      C O N T E N T S (Continued)



      Committee Discussion and Recommendations                 258


      Conclusion and Summary Remarks

                Ajaz Hussain, Ph.D.                            273

                Helen Winkle                                   280




                         P R O C E E D I N G S


                             Call to Order


                DR. COONEY:  I would like to call the


      meeting to order this morning.


                     Conflict of Interest Statement


                MS. SCHAREN:  Good morning.  I am going to


      be going through the Conflict of Interest




                The Food and Drug Administration has


      prepared general matters waivers for the following


      Special Government Employees:  Drs. Charles Cooney,


      Patrick DeLuca, Carol Gloff, Arthur Kibbe, Michael


      Korczynski, Thomas Layloff, Marvin Meyer, Kenneth


      Morris, Nozer Singpurwalla, who are attending


      today's meeting of the Pharmaceutical Science


      Advisory Committee, to:


                1.  Receive an update on current


      activities of the Parametric Tolerance Interval


      Level PTIT Work Group;


                2.  Discuss and provide comments on the


      general topic of considerations for assessment of


      pharmaceutical equivalence and product design;


                3.  Discuss criteria for establishing a


      working group for review and assessment of Office


      of Pharmaceutical Science Research Programs.




                The meeting is being held by the Center


      for Drug Evaluation and Research.  Unlike issues


      before a committee in which a particular product is


      discussed, issues of broader applicability, such as


      the topic of today's meeting, involve many


      industrial sponsors and academic institutions.


                The committee members have been screened


      for their financial interests as they may apply to


      the general topic at hand.  Because general topics


      impact so many institutions, it is not practical to


      recite all potential conflicts of interest as they


      apply to each member.


                FDA acknowledges that there may be


      potential conflicts of interest, but because of the


      general nature of the discussions before the


      committee, these potential conflicts are mitigated.


                With respect to FDA's invited industry


      representative, we would like to disclose that Dr.


      Paul Fackler and Dr. Gerald Migliaccio are




      participating in this meeting as non-voting


      industry representatives acting on behalf of


      regulated industry.  Dr. Fackler's and Dr.


      Migliaccio's role on this committee is to represent


      industry interests in general, and not any other


      one particular company.  Dr. Fackler is employed by


      Teva Pharmaceuticals, Dr. Migliaccio is employed by




                In the event that the discussions involve


      any other products or firms not already on the


      agenda for which FDA participants have a financial


      interest, the participants' involvement and their


      exclusion will be noted for the record.


                With respect to all other participants, we


      ask in the interest of fairness that they address


      any current or previous financial involvement with


      any firm whose product they may wish to comment




                Thank you.


                DR. COONEY:  Thank you, Hilda.


                I am Charles Cooney, the chairman of the


      committee, and will preside over the schedule




      today.  We have several topics on the schedule.


      The first of these is a current update of the


      working group Parametric Tolerance Interval Test


      for Dose Content Uniformity by Robert O'Neill, who


      I believe has just come in.


                 Parametric Tolerance Interval Test for


                        Dose Content Uniformity


                  Current Update on the Working Group


                DR. O'NEILL:  I am here to just update you


      on what we promised you from the last meeting on


      October 19th.  As you know, there is a technical


      working group that has been put together with folks


      from FDA and folks from the IPAC group.


                We have been working diligently since


      then.  We thought we would have some


      recommendations for you today.  We do not, but we


      have had approximately eight get-togethers, five


      telecons and three face-to-face meetings, the last


      of which was about a week, week and a half ago.


                What those discussions have been about is


      the agreement of the statistical formulations of


      the problem. There were a number of discussions




      regarding the coverage probability and symmetry of


      coverage, off-target operating characteristic curve


      agreements, and things of that nature.


                There has been some computer programs that


      have been shared back and forth, validation of each


      other's methods, and I believe there is now


      agreement on that aspect of it and that the working


      group is turning towards the agreement on where the


      operating characteristic curve ought to be placed


      and how it might handle certain particular


      situations, particularly off-target means.


                That is essentially where we are.  I think


      everyone is optimistic that probably the next time


      we report to you, that there will be actual


      recommendations for you to respond to, but that is


      essentially where we are right now.


                I would be willing to take any questions


      if you have any.


                DR. COONEY:  Any comments or questions


      from the Committee?  If not, thank you very much


      and we look forward to the next step.


                DR. O'NEILL:  Thank you.


                DR. COONEY:  We will immediately begin


      with the second topic this morning, which is


      Quality-by-Design and Pharmaceutical Equivalence,




      to be introduced by Ajaz Hussain.


            Quality-by-Design and Pharmaceutical Equivalence


                           Topic Introduction


                DR. HUSSAIN:  Good morning.  I would like


      to introduce to you the Topic No. 2,


      Quality-by-Design and Pharmaceutical Equivalence.


      We believe we have an opportunity here to explore


      and what we plan to do with you is to share some of


      our initial thoughts and hopefully, engage the


      Advisory Committee in discussion to help us make


      sure we are on the right track.


                In many ways, the discussions continue


      from yesterday where, in essence, we are looking at


      opportunities that have been created and


      re-examining some of our current policies and to


      see how we can realize opportunities to move


      towards a desired state.


                I would like to put this in the context of


      moving from a reactive to a proactive decision




      system for pharmaceutical quality, and recognizing


      that this is only a journey, not a destination.  I


      think in the world of continuous learning and


      continuous improvement, really, continuous learning


      is your destination in some ways.


                I think over the last four years, we have


      focused on discussing some of the opportunities in


      general, but if I look at some of the reactive


      examples on this chart, yesterday, in some way, we


      talked about testing to document quality, repeating


      deviations in our specification investigations, and


      in some ways we will start focusing on the other


      aspects of prior approval supplement for process


      optimization and continuous improvement, multiple


      CMC review cycles, but more importantly, I think,


      how can we leverage the opportunity of quality by


      design for demonstrating therapeutic equivalence of


      generic products.


                I think we struggle often in this arena,


      and we have struggled and unable to sort of find


      ways of approving generic products and the degree


      of complexity has increased such as topical




      products and inhalation products, and so forth, so


      how can we leverage the pharmaceutical development


      information to seek out ways to find mechanisms for


      approving complex generic products, I think is a




                I think the proactive examples I think is


      quality by design and real time release, right


      first time, process optimization and continuous


      improvement within the facilities quality system.


                But I think in some ways, today, we will


      focus on single CMC review cycle, less so, but I


      think quality-by- design approach for demonstrating


      therapeutic equivalence of generic products would


      be a focus.


                Therapeutic equivalence.  The definition


      from the Orange Book is as follows.  Drug products


      are considered to be therapeutic equivalents only


      if they are pharmaceutical equivalents and if they


      can be expected to have the same clinical effect


      and safety profile when administered to patients


      under conditions specified in the labeling.


                Pharmaceutical equivalent products are




      drug products are considered to be pharmaceutical


      equivalents if they contain the same active


      ingredients, are of the same dosage form, route of


      administration, and are identical in strength or




                Pharmaceutically equivalent drug products


      are formulated to contain the same amount of active


      ingredient in the same dosage form  and to meet the


      same or compendial or other applicable standards of


      strength, quality, purity, and identity, but they


      may differ in characteristics such as shape,


      scoring configuration, release mechanisms,


      packaging, excipients, expiration time, and within


      certain limits, labeling.


                So, I think we have certain flexibility


      built into the issue of pharmaceutical equivalence,


      and one of the desired states is to leverage that


      and to say that we want to set specification based


      on mechanistic understanding, so if you have a


      different mechanism, that is perfectly fine, but


      then you set your specification on that.


                The discussion yesterday was not to force




      the generic to a particular specification, but


      recognize that as part of that.


                Today, I think I would like to put this in


      the context of risk, uncertainty, and variability,


      and I think that framework will help us think more


      clearly about the issues.


                FDA classifies as therapeutically


      equivalent those products:  That are approved safe


      and effective, so you have to have a reference


      product which is safe and effective, and has an


      approved pharmaceutical equivalents against the


      repetition of that, and are bioequivalent, that


      they do not present a known or potential


      bioequivalence problem, and they meet an acceptable


      in vitro standard, or if they do present such a


      known or potential problem, they are shown to meet


      an appropriate bioequivalence standard.


                In absence of pharmaceutical development


      information and quality-by-design aspect, we have


      to assume that they present a bioproblem, so we


      often go to the second bullet in most cases; are


      adequately labeled, and are manufactured in




      compliance to Current Good Manufacturing Practice




                In a sense, the last four years, our


      initiative goals have been to focus on


      science-based, risk-based approaches, and there are


      certain challenges I think in our articulation of


      the problem today.


                I want to sort of share with you what the


      challenges might be.


                Risk-based scientific decisions on


      pharmaceutical quality is clearly our goal.  Risk


      is a combination of the probability of occurrence


      of harm and the severity of that harm.  The reason


      for focusing on that is an aspect that I feel that


      we often get entangled in and unable to really


      articulate the problem carefully.


                Uncertainty with respect to severity of


      harm and/or probability of its occurrence and their


      modulating factors is that challenge that we face,


      what are the critical quality attributes, and what


      is an acceptable variability.


                I have argued, and I think I will ask the




      Committee to think about this, is my argument sort


      of correct, in some ways, in the current decision


      system which tends to be reactive, one contributing


      factor for the reactive decision system is we


      confound uncertainty, variability, and risk.


                This is, by nature, how we develop our


      products to a large degree.  The current paradigm


      for product and process development tends to do


      this, because our entire effort in new drug


      development, for example, is focus on safety and


      efficacy of a molecule, but we use a product to


      achieve that.


                The connection between product quality and


      the clinical is generally focused on the molecules


      rather than the product.  So, that is a part of


      this discussion.


                Often intrinsic safety and efficacy of a


      new molecular entity is confounded with its product


      and manufacturing process just by the nature of our


      product development and process itself.


                We have multifactorial aspects of


      pharmaceutical products and manufacturing




      processes, and there is increasing complexity, and


      if we can find a way to articulate our problem more


      carefully, this may help moving forward more




                Establishing constraints based on prior


      knowledge and limited development experiments that


      are done in the development cycle.


                What I have argued is there is a need to


      entangle or, as I call it, de-convolute


      uncertainty, variability, and risk, and then to


      achieve truly a scientific integration of these for


      quality decisions, how we set specification.


                Yesterday, for example, what we proposed


      was an assessment of variability, example, begins R


      and R.  That is an attempt to start characterizing


      the variability and attempt to start teasing out


      what comes from what.


                This may appear to be paradoxical, and it


      probably is without the concept of quality by


      design, and that sort of links to Dr. Woodcock's


      paper that we talked about.


                Let me illustrate that.  When we approve a




      new drug application, we bring assessment of all


      the disciplines to bear on a decision which says


      the risk to benefit ratio of this proposed drug


      product is acceptable when used in accordance to


      the label.  That is the pivotal decision criteria.


                Now, that is based on the pivotal clinical


      trials along with the toxicology, and along with


      all the studies that go along with that, but the


      pivotal clinical trials play the major role there.


                What do we use for the pivotal clinical


      trials?  We use a product, and we often do not have


      the opportunity or do not have the intent to gauge


      whether the product is modulating the safety and


      efficacy of the molecule beyond that of exposure


      less the bioavailability.


                So, from that aspect, the quality of the


      product has to be built in before you get into the


      clinical trials, otherwise you confound the


      clinical trials with quality problems, and


      actually, I will illustrate one example in my


      second presentation of that.


                As that happens, then, the product is




      approved and then manufacturing is transferred to


      production and you have seven production lots, but


      that goes to the patient population.


                So, you see the disconnect between


      pharmaceutical quality or product quality.  I am


      not talking about the safety and efficacy of a


      molecule, I am talking about how a product


      modulates that.


                So, in many ways, if I look at drug


      development program, and here is an illustration of


      that, this is actually a real case study.  The only


      difference I have here is I have a line that made


      it more linear.  That is the only change I did, but


      here is a development program.


                The initial formulation was a capsule.


      Then, they went to a table, wet granulation tablet,


      and each star is a bioequivalent study.  So, they


      qualified that change from capsule to tablet using


      a bioequivalent study.  Then, they add a film coat,


      then, change the site of manufacture of the drug


      substance, and so forth.


                You see the changes that were occurring in




      this drug development program, some before clinical


      trials were initiated, some during the clinical


      trials, site changes, and so forth.


                But this company actually qualified, what


      they call bridging studies using a bioequivalence


      trial.  Towards the end of the review time, to


      qualify and bridge the clinical trial, pivotal


      clinical trial material with the to-be-marketed


      product, they opted to use multiple-dose studies of


      the traditional thing, and that failed at the last


      minute.  Actually, it didn't fail, it just failed


      to establish bioequivalence, so they went back and


      actually repeated the study with a larger number of


      subjects, but it delayed them.


                The point here is all these changes are


      being qualified on the basis of their traditional


      bioequivalence trial.  In new NDAs, we generally


      see from 3 to 6 clinical bioequivalence studies,


      and that was our number.


                Here is a number that was shared with us


      by Jack Cook sometime, this is the year 2000, about


      7 per approved compound.  In a sense, if I look




      at--this is Gary Buehler's slide--I think in many


      ways, the difference between the generic drug


      approval and the new drug approval is minimal when


      it comes to bioequivalence trial.


                In many ways, we use bioequivalence to


      gauge information which we don't have in the NDA,


      but in reality, in some cases, it is simply the two


      are comparable.  So, from that aspect, I think when


      I look at oral products, immediate release


      products, and here is my demarcation in an attempt


      to sort of categorize what is uncertainty, what is


      variability, and what is risk.


                The goal of our generic drug approval


      process is to approve a generic product.  An


      approved product is expected to have the same


      clinical effect and safety profile when


      administered to patients under conditions specified


      in the labeling.


                Based on the previous slide, I have shown


      the characteristics of this decision system is,


      one, the first one, the product must be


      pharmaceutically equivalent, and here the questions




      are has the applicant demonstrated that it's the


      same active, identical amount, same dosage form,


      route of administration, and so forth, identity,


      strength, quality, purity.


                So, that is an assessment process of how


      good the identical methods are and how good they


      have qualified, so it's a knowledge-based decision,


      there is an uncertainty aspect associated with




                Then, we have to define acceptable


      variability for that product, and we rely on the


      compendial or other standards to do that.


                The risk in this case, I am talking about


      risk from a clinical perspective, is the prior


      knowledge that come from NDA.


                Need for bioequivalence assessment for


      oral products is that next question, and again the


      same words from the Orange Book is if you do not


      present a known or potential bioequivalence


      problem, acceptable in vitro standard is fine.  But


      you saw the debate with dissolution, we often don't


      go there, we often go to a bioequivalence study.


                If you go to in vitro standard compendial


      dissolution test method, if you go to a


      bioequivalence standard, then, the acceptable




      variability is our bioequivalence standard, 90


      percent confidence interval for test or reference


      ratio for rate and extent of absorption is within


      80 to 125 percent.


                It has to be adequately labeled and


      manufactured in conformance to cGMPs, and in that


      case, acceptable variability is what we tolerate in


      terms of deviation or specifications, and so forth.


                So, that is one way of looking at trying


      to partition uncertainty, variability, and risk, so


      that we can formulate the right questions.


                Clearly, I think the quality-by-design


      thinking is intended to focus on the intended use,


      and design is about doing things consciously.  I


      showed this slide to you before.


                I think what we would like to consider is


      in the context of pre- and post-approval changes,


      generic drugs, and even extending that to the


      concept of follow-on protein pharmaceuticals, the




      primary goal of a scientific decision framework--I


      am not talking with a regulatory process--the


      decision criteria that we bring, need to understand


      the complexity and uncertainty, but the decision


      process should be consistent.


                I think that is the fundamental basis that


      we work under.  Furthermore, I think our goal is


      also to identify and eliminate unnecessary human


      and animal testing for this decision framework,


      keeping in mind that most bioequivalent studies are


      done in normal healthy subject volunteers, new


      drugs and so forth.


                If we can avoid exposing normal healthy


      subject human volunteers, it is desirable, and that


      is part of the regulation.  I will share that


      regulation with you.


                So, Topic 2 today, the premise that we had


      in mind was that a quality-by-design approach via


      pharmaceutical development information can


      potentially provide an excellent means to address a


      number of challenges previously discussed at ACPS


      meetings without complete or satisfactory




      resolution, for example, bioequivalence of highly


      variable drugs, bio-in-equivalence criteria,


      pharmaceutical and therapeutic equipment of locally


      acting drug products, such as topical products.


                Today, Lawrence will bring his thoughts to


      you, and these are our preliminary thoughts, and I


      think we just wanted to put our preliminary


      thoughts on the table to engage you and engage the


      community to help us think about this, so that as


      we spend our time thinking about this, we already


      have some feedback and we are also on the right




                Rob Lionberger will come back.  He


      presented a decision tree to you before, but he


      will recast that decision tree for topical products


      in the context of quality-by-design.


                Yesterday, we focused on dissolution


      testing, and as the past chair of the BCS Working


      Group, I took it upon myself to sort of go back and


      re-examine what were my personal thoughts and what


      held me back to make the recommendations that you


      see in the guidance and see how we can recast that




      discussion into the quality-by-design thinking, so


      that I can take this discussion and give that as a


      recommendation to the current Technical Committee


      on Biopharmaceutical Classification System.


                Again, as I said, these are initial


      thoughts, and our goal is to engage you in a debate


      and discussion to hopefully give us some


      perspective are we on the right track.  The three


      topics today are:


                How can pharmaceutical development


      information help to extend the applications of


      BCS-based waivers of in-vivo studies for immediate


      release products?


                How can pharmaceutical development


      information be utilized to address the challenges


      of highly variable drugs?


                How do we use this to establish


      therapeutic equivalence of topical products?


                Those are the three topics that we would


      like to present, and the general question is are we


      on the right track, but then more detailed


      recommendations on how we should proceed with these




      three topics or other topics that we should have


      considered instead of this.


                So, that is the Topic 2.


                DR. COONEY:  Thank you, Ajaz.


                We might pause for a moment for any


      questions particularly around clarification of


      these opening comments from the committee.  Art?


                DR. KIBBE:  The question I always struggle


      with is how do we define highly variable drugs.


      Are we defining them in clinical outcomes, because


      then the dosage form might not be involved in it at




                DR. HUSSAIN:  The definition hinges on the


      bioequivalence drug, the variability that we


      estimate from the bioequivalence drug.


                DR. KIBBE:  But that could be a function


      of intersubject variability, subject population


      selection, and have nothing at all to do or


      minimally to do with the actual product that you


      are making.


                DR. HUSSAIN:  That is the discussion


      Lawrence will bring to you, so if you hold that




      question for Lawrence.


                DR. KIBBE:  Thank you.


                DR. COONEY:  Ken?


                DR. MORRIS:  The two comments is that the


      Topic 2 premise, the other part of that premise is


      that the proper development information is being


      generated at the companies which, of course, you


      have limited control over, and that is being


      shared, just as a caveat.


                DR. COONEY:  Nozer?


                DR. SINGPURWALLA:  Two comments.  Slide


      No. 7.


                DR. HUSSAIN:  I am sorry, I don't have




                DR. SINGPURWALLA:  I know.  Components of


      the Challenge.  The second bullet.  That second


      bullet is wrongly worded, it has to be changed, and


      I will tell you why.


                The more important reason is you, on your


      eight bullet, are talking about confounding of


      uncertainty, variability, and risk.  They should


      not be confounded.  Who is confounding them and






                DR. HUSSAIN:  The current system has a


      tendency to do that.  That is what I mean.


                DR. SINGPURWALLA:  But that simply means


      that the system is not educated enough, because


      variability is the cause of uncertainty, is one of


      the causes of uncertainty, and risk is a measure,


      is a way to measure uncertainty and its




                So, why is there so much confusion about


      these very elementary ideas in the industry and


      perhaps in the pharmaceutical community?


                DR. HUSSAIN:  I don't know how to answer




                DR. SINGPURWALLA:  Well, they need to be


      trained, educated.


                DR. HUSSAIN:  But let me propose this in


      the sense, uncertainty is not risk.


                DR. SINGPURWALLA:  I agree with you.  I


      agree with what you are saying completely.  What I


      am asking is, what is the cause of this confusion,


      and it is so easy to remove this confusion?


                DR. HUSSAIN:  I understand the concern you


      are expressing, and my premise is for years we have


      not utilized the pharmaceutical development, and we




      have treated that as an art rather than a science,


      and that is the way to get away from that


      confusion.  So, that is the premise.


                DR. COONEY:  Any other comments at this


      point?  We will have ample opportunity for further




                Using Product Development Information to


                Extend Biopharmaceutics Classification


                        System-Based Biowaivers


                DR. HUSSAIN:  Let me go on to the


      Biopharmaceutics Classification System.  In


      preparing for this presentation, I actually went


      back and reviewed all the publications that have


      occurred in this arena in this discipline, in this


      topic area for the last five years, and there has


      been a tremendous number of progress in this area.


                For example, more recently, Professor Les


      Benet's article was published on how you can


      actually start predicting metabolism, and so forth,




      and how you can sort of add that additional




                There has been a paper published in Pharm


      Research on quantitative instead of, you know,


      rigid boundaries, and so forth.


                But instead of sort of trying to summarize


      the progression signs, what I wanted to do was to


      go back and re-examine my own thoughts that were


      expressed to the Advisory Committee in the year


      2000, so I am actually going to repeat an old


      presentation, but in light of what we have learned


      in the last four years.


                My goal here is to share with you some of


      the concerns we had when we proposed the BCS-based


      waiver guidance in the year 2000, and to what


      extent those concerns remain, and to what extent


      quality-by-design may be able to alleviate some of


      this concern, and the discussion with you, I intend


      to use that as recommendation to the current


      Technical Committee on BCS.  So, that is the game




                This is an old presentation with some




      minor modification.  When I had made this


      presentation, I was completely focused on risk, and


      the title was "Biopharmaceutics Classification


      System: A Risk Management Tool."


                In light of the learning that I at least


      personally had, I want to sort of bring in the


      uncertainty and variability components to this.


                Since this is a presentation, probably my


      last presentation on the BCS topic before I handed


      over the reins of responsibility to Lawrence and


      Mehul, the new BCS Technical Committee was formed,


      when Helen asked me to move to OPS and the PAT


      process got started, so my focus went to PAT for a


      reason which connects back to this one.


                So, this BCS Technical Committee has been


      in place under the leadership of Lawrence and


      Mehul, and they have been diligently addressing a


      number of implementation issues trying to


      coordinate all the submissions, and so forth, and


      there has been significant activity on this


      guidance on the new drug side, very little, if any,


      on the generic side.


                You also heard from Mehul the database is


      now almost ready, is being audited, database and


      prospective research for extensions, links to PQRI




      and FIP, but the PQRI program really didn't take


      off, and our thoughts were we wanted to explore


      extension of BCS-based biowaivers to Class III and


      Class II drugs.


                Further research at the FDA, which we


      completed, and we did extend the BCS-based


      biowaivers to "fed" bioequivalence studies, and


      that was part of the thing, and that work was done


      with collaboration with Tennessee.


                Continuing education initiatives and


      practitioners and public, and the group has been


      busy.  International harmonization was an aspect,


      but to the extent the definition of high solubility


      and rapid dissolution, we got into ICH Q6A, the


      European Guideline also adopted much of the BCS


      recommendation to some extent.  There are certain


      differences, though, it is not fully harmonized.


                With that as a background, let me trace


      for you the evolution of the recommendations in the




      BCS guidance that we released in the year 2000.


                Regulatory Bioequivalence: An Overview,


      from my perspective, this is a graphical


      representation of our regulation.


                If you look at the dosage form that we


      deal with, solutions, suspensions, chewable


      tablets, conventional tablets, and modified release


      products, for solutions, we consider bioequivalence


      essentially is self-evident, bioavailability is


      self-evident, and biowaivers are granted, condition


      being excipients do not alter absorption, and that


      is an assessment based on historical data.


                For any product that contains drug in a


      solid form, we have a concern, and for pre-1962


      drugs, we call DESI drugs, in vivo evaluation for


      bio-problem, that was the original biopharm


      classification system, if you really look at it,


      that had many of the elements of therapeutic index,


      PK, the solubility, and so forth.


                For post-1962 drugs, generally, in vivo,


      some exceptions with IVIVC.  Then, we introduced a


      SUPAC-IR guideline in '95, and we introduced the




      elements of BCS guidance, BCS system in that to


      give a waiver for minor changes and moderate


      changes that Mehul talked to you about yesterday.


                For modified release, we don't have a


      classification system, bioequivalence has to be


      demonstrated in vivo with the exception of SUPAC


      modified release for within a product.  If you have


      in vitro bioequivalence, you can make changes.


      Again, Mehul summarized that.


                I want to trace back the discussion to a


      bioequivalence hearing, which I did not attend.  I


      was just graduating in '86, but I was connected to


      this because I made the slides of a number of


      people who presented here, so I knew what was




                This was a pivotal discussion and I think


      set the stage for what evolved as bioequivalence


      standard.  There were two comments that I want to


      share with you.


                One was Dr. Bob Temple.  He said, after


      the end of this discussion, it seems sensible to


      think that swallowing something that turns into a




      solution rapidly would be difficult to lead to


      differences from one product to another.


                So, the clinicians were arguing you don't


      need biostudies for everything.  Arnold Beckett had


      made that argument years at that time, so he said


      you shouldn't go with in vivo for everything, but


      we did.


                Milo Gibaldi, an eminent pharmacokinetic


      professor, "I have learned that there is no support


      here for attempting to provide such assurance


      solely with in vitro data."  So, that was a pivotal


      aspect, I think, and I went back and sort of tried


      to examine the thoughts and the concerns that were


      expressed at that session.


                The other aspect that I do want to put on


      the table is need to reduce our reliance on in vivo


      bioequivalence studies.  Why?  Ethical reasons.  21


      CFR 320.25 says, "No unnecessary human research


      should be done." Science continues to provide new


      methods to identify and eliminate unnecessary in


      vivo bioequivalence studies.


                Focus on prevention, "building quality




      into products - right first time."  So you see the


      PAT initiative and how this will connect to that


      was in the thought process and why we aggressively


      moved in that direction, "right first time," I used


      before Pfizer.


                Time and cost of drug development and


      review is a key, because if you see that we have


      three to six bioequivalent studies in our NDAs.  We


      actually at some point said we don't even review


      some of those because they are redundant, so why


      expose normal healthy subjects to a new drug which


      is under development with all the risks associated


      with that.


                So, prior to SUPAC-IR/BCS, in vivo


      bioequivalence assessments to justify a majority of


      manufacturing changes.  So, this was a significant


      hurdle, and that changed.  In the SUPAC-IR guidance


      in '95, we brought in the classification system and


      provided a tiered approach for changes based on in




                For example, highly soluble, highly


      permeable drug, the critical processes for gastric




      emptying, dissolution is not likely, and


      dissolution is not likely to be rate limiting, but


      we said 0.1 single point, 85 percent, and so forth.


                So, you can see that for each class, we


      recommended a tiered approach for waiver of


      biostudies for minor changes, and so forth.  We


      excluded the Narrow Therapeutic Index drugs from


      waiver consideration, but we never defined what


      narrow therapeutic index was, and we still haven't.


                The guidance in 2000 really extended that


      and put that as a waiver for first time approval,


      and also provided the methods to classify your


      drug, and so forth.  The pivotal recommendation in


      that was waiver for in vivo bioequivalence studies.


                I do want to stop here and say the title


      of this guidance was debated to the nth degree


      before we agreed on this internally.  The word


      "waiver" was to signify that we want an in vivo


      study for everything that is in solid, so the title


      was very carefully chosen to reflect that.


                Anyway, it's waiver for in vivo


      bioequivalence studies is recommended for a solid




      oral test product that exhibit rapid and similar in


      vitro dissolution under specified conditions to an


      approved reference product when the following


      conditions are satisfied:  products are


      pharmaceutically equivalent, drug substance is


      highly soluble and highly permeable and is not


      considered to have a narrow therapeutic range and


      excipients used are not likely to affect drug




                The class membership, the boundaries that


      you see, which are rigid, high solubility, the


      highest dose strength is soluble in less than or


      equal to 250 ml of aqueous buffers over the pH


      range that we had 1 to 6.5 or so, whatever that


      thing is I forgot.


                The reason for 250 ml is the glass of


      water that we take when we do a bioequivalence


      study.  High permeability, the extent of absorption


      in humans is determined to be greater than 90




                Rapid dissolution is 85 percent dissolves


      within 30 minutes in three different conditions,




      HCL, pH 4.5 and 6.8 buffers using traditional


      settings of dissolution apparatus.


                Now, clearly, I had approached this as a


      risk to bio-in-equivalence because since we started


      with the premise that you needed bioequivalence


      trials for approval of changes, and so forth, so


      the risk factor for me was the proposal the


      recommendations should not result in




                The risk factors that we had in mind were


      clearly manufacturing changes, poor process


      capability, high between and within batch


      variability, but the thing we focused on, reliance


      on in vitro dissolution tests and how reliable that


      is, single point specification, sampling,


      predictability were the issues.


                Clearly, the other factors were there,


      deficiencies in BE study design, Type II errors,


      and so forth.


                Now, assessment of risk, what is the risk


      of bio-in-equivalence between two pharmaceutically


      equivalent products when in vitro dissolution test




      comparisons are used for regulatory decisions?


      That was the question we asked.


                What is the likelihood of occurrence and


      the severity of the consequences?


                Severity was not meeting the


      bioequivalence criteria was unacceptable, but what


      was the likelihood, so we needed some information


      on that.


                Regulatory decision, whether or not the


      risks are such that the project can be pursued with


      or without additional arrangements to mitigate the


      risk, and clearly, acceptability of the decision,


      is the decision acceptable to society?  This took


      four years.


                We started working on this in '96, and if


      you think I was busy with the PAT presentations


      around the globe, that is exactly had to do the


      same thing for this one, too, because the mind-set,


      the paradigm was so entrenched in the old system.


                Professor Gordon Amidon spent some time


      with us, he and I.  I had the luxury of having the


      biopharm document room right outside my office in




      the Parklawn Building, so we went through a number


      of applications, about 160 applications at that


      point, to get a sense of what is happening.


                Roughly, what we found was on the new drug


      side only, because we have failed studies or we


      have all the studies submitted on the generic side,


      we couldn't use that database because you just have


      the passing studies in there.


                So, we looked at the new drug side and


      said when does dissolution signal


      bio-in-equivalence or does not signal


      bio-in-equivalence.  What we found there was


      generally, you see big differences in dissolution,


      no difference in blood levels.


                But, on the other hand, there were signals


      that dissolution fails to signal bio-in-equivalence


      about 30 percent of the time, and we wanted to ask




                So, minimizing risk of bio-in-equivalence,


      does in vitro dissolution process emulate in vivo


      dissolution process in vitro and in vivo?  Dosage


      form disintegration, dissolution, and stability




      were the concern.


                The gastrointestinal fluid volume,


      composition, and hydrodynamic conditions were the


      concern, and clearly, I think one thing which was


      pivotal for the oral discriminating part was the


      surface tension, and that could have been picked




                Residence time in the stomach and small


      intestine were an issue, so we did a lot of


      analysis actually of gastric emptying and what


      factors affect gastric emptying, and so forth.


                Impact of excipient differences on GI


      physiology and drug bioavailability were the




                The key question was how well this


      emulates in vivo, because this is our standard


      dissolution test.


                This was a cartoon that I prepared and to


      take a look at typical physiologic parameters in a


      single dose fasting BE study.  We had fairly good


      estimates of the gastric fluid plus the 8 ounces of


      water.  We knew what the gastric pH range is




      generally in the normal subject.


                We had the information on the gastric


      emptying time, which is highly variable, but


      approximately T50 is 15 percent.  The permeability


      is low, and that was an advantage in the stomach


      compared to small intestine, the surface tension is


      lower, and clearly, volumes in the small intestine


      were uncertain, and pH, and so forth, and the


      permeability was high.  Hydrodynamics was a big


      question in our minds.


                Lawrence summarized to you the debates


      that we have had for dissolution for the last 30


      years, and that dissolution tests are over


      discriminating, on one hand, and in the USP, the


      statement that products that dissolve about 70


      percent in 45 minutes have no medically relevant


      bioequivalence problems, what was the basis of




                Dissolution tests are not sufficient to


      assure bioequivalence, and demonstration of IVIVC


      is necessary, but when you do that, product


      specific, so those are two sides of the debate.


                I showed you this slide of the problems


      with the dissolution tests of false positives and


      false negatives, but then we also looked at things




      that we made decisions on.


                Here is a product.  The generic product


      was Product B.  We actually withdrew this product


      from the market after approval.  This is a pre-62


      drugs, it was approved on the basis of dissolution,


      meeting the USP specification criteria.


                This was a pre-62 drug, one of the older


      drugs, and we had a challenge from the innovator,


      which is Product A, that the generic is not


      bioequivalent, and that was the basis at which we


      had withdrawn this product, Product B, from the




                The Cmax, you can see is clearly high, but


      in many ways, Product B was more reliable, less


      variable, and it was more modern technology, but


      the constraints on us is you have to be equal, if


      not better.


                Here are examples of where there were real


      dissolution impacts on in vivo absorption.  Here is




      a weak acid where the initial formulation for


      clinical was capsule.  They went to


      wet-granulation, and the to-be-marketed was a


      direction compression with dicalcium phosphate, and


      the dissolution in this case was Q17 30 minutes in


      simulated intestinal fluid.


                That's the criss-cross you see if you do


      acid in alkaline conditions, you don't distinguish


      that, and this had to be reformulated.  But this


      was I think in my mind a signal that we probably


      are designing our products for dissolution rather


      than the intended use.


                I wanted to walk through this with you,


      and that was one of the reasons for the


      quality-by-design thinking.


                Here is a Drug X.  This is actually a


      clinical study, 100 mg dose, so it's a highly


      soluble drug by all definition.  It's a weak base


      exhibits a sharp decline in solubility with


      increasing pH above 3.


                Now, the clinical trial material in this


      case was wet granulation, drug particle size of D50




      of 80 microns, and this was a concern, because our


      particle size specification was very crude.  I mean


      what does D50 percent give me.  We had lactose,


      microcrystalline cellulose, and so forth.  You see


      that formulation there.


                But the point to focus there is the


      dissolution 0.1 normal HCL was 65 percent in 15


      minutes and 100 percent in 20 minutes.


      Disintegration time was 10 minutes.


                The way I had presented this, the


      to-be-marketed was the formulation of direct


      compression, but actually the wet granulation in


      this case was a U.S. formulation, a formulation


      using U.S. clinical trials, and the European


      clinical trials were done with the to-be-marketed,


      and we had to bridge those together.


                The to-be-marketed formulation, you can


      see what happened here.  Direction compression,


      drug particle size of 300 microns, dicalcium


      phosphate, and so forth, and the dissolution is


      more rapid now, 0.1 normal HCL - 85 percent in 15


      minutes, about 95 percent in 20 minutes, so the




      initial dissolution burst is very rapid,


      disintegration time is 1 minute.


                Clinical product exhibits poor dissolution


      at 7.4. Can you imagine, I mean this is a BCS Class


      I drug.  The Cmax or the rate or the exposure of


      this in terms of p concentration for the


      to-be-marketed formulation in this case was half,


      so you see half the blood levels in terms of


      height.  So, it simply was signal that if you don't


      get the physicochemical properties of drug


      understood, you will have these problems.


                So, in vitro and in vivo dissolution,


      dissolution methods evolved over the last 30 years.


      The year 2000, I said there were reproducible test


      methods for lot-to-lot quality assurance, so you


      can imagine my surprise of the calibration, which I


      was not aware of at that time.


                The dissolution media volume and


      composition selected to maintain "sink" conditions.


      In vivo dissolution is a complex process, and you


      have to consider pH profile, bile concentration,


      motility patterns, and so forth.


                In vivo, the "sink" condition is created


      due to intestinal permeability, and this was a


      contention, which Lawrence and others, we recently




      published to show how permeability actually impacts


      in vivo dissolution, so we have published on this


      now already.


                I will talk to you about in vitro-in vivo


      correlation.  The formulation specificity of IVIVC


      has been known since 1997.  This is drug


      spironolactone from a publication in J. Pharm


      Science in '97.


                So, you can see a change in particle size.


      You may have a correlation, but a change in


      particle size could be outside that.  So, a


      correlation itself why this formulation specific


      has to be really brought into context.


                So, reliance on current dissolution


      practice can pose an unacceptable level of risk


      from bio-in-equivalence perspective.  Compared to


      high solubility drugs, risk is higher for low


      solubility drugs.


                Products with slow or extended dissolution




      profiles pose a higher risk.  The dissolution is


      rate limiting.


                So, in a sense, we wanted to use


      dissolution test only to rule out that dissolution


      is not rate limiting.  So, that was the basis for


      our thought process.  So, we constructed a rapid


      dissolution criteria for that purpose.  We did not


      want to use dissolution tests for bioavailable


      decisions if there was a hint that dissolution is


      rate limiting.


                Potential for significant differences


      between in vivo and in vitro "sink" conditions


      higher for low permeability drugs, which we had to


      prove later on with a simulation study that


      Lawrence did.


                Now, to establish a boundary for rapid


      dissolution, we simply postulated that since


      gastric mucosa does not have high permeability to


      drugs, you have a 15-minute time, so you can take


      advantage of that.


                So, if the dissolution is rapid, then,


      much of it is complete before it empties into the




      small intestine where you have high permeability.


      So, in a sense, a very rapidly dissolving drug will


      behave essentially like a solution, and it does.


                So, here is a snapshot of dissolution


      versus AUC and Cmax ratios and the bioequivalence


      goalpost of a drug metoprolol.  The reason I did


      not take the name off here, because this is already


      an ACPS presentation, it is already on the website.


                So, you can see a solution versus all the


      other formulations that we have approved, generic


      and innovator, plus there are research


      formulations, which we deliberately made to be very


      slowly dissolving.


                You see that essentially, for the most


      part, the slope is zero.  We then did extensive


      simulation work to establish that if in vivo


      dissolution is rapid, as a function of different


      gastric emptying as a function of mean intestinal


      transit time, you are not likely to see a


      difference between solution and a tablet, and that


      was the basis for our 85 percent and 30 minute in


      vitro criteria for rapid dissolution, but we did




      not apply that to low permeability drugs because of


      the risk factors that we felt were coming from




                The question we asked was is the current


      approach for evaluating excipients for decisions


      related to biowaiver for oral solutions sufficient.


      That is the database I have.


                There were hints that the excipient


      effects were not fully appreciated.  For example,


      there was a study by Ian Wyling's group where you


      could show mannitol, 2.3 grams of mannitol clearly


      had a big impact on bioavailability of cimetidine,


      a low permeability drug, and on the other hand,


      Fassihi had shown that a high permeability drug had


      minimal impact of sorbitol, even 10 grams of




                So, that was a hint, and we actually


      conducted a study at the University of Tennessee,


      and we have now completed that study, even getting


      to a mechanistic basis for generalizing the


      permeability effect to other excipients, is to test


      this out.


                Metoprolol was our boundary drug for high


      permeability, so we did a study with ranitidine and


      metoprolol, and we did a crossover study in




      replicate design to get an estimate of subject, the


      formulation interaction also.


                So, it is a very simple formulation.  You


      have your drug.  You have sucrose or sorbitol, and


      you have 15 ml of water.  That is the simplest




                What we found--I will just show you a


      picture--just a confirmation for a low permeability


      drug like ranitidine, a dramatic effect.


                Now we have completed the study of the


      dose response, the amount of sorbitol that triggers


      this is about 1.2 grams.  For metoprolol, the Cmax


      was affected, but not to the same extent.  AUC was




                In addition, there were a tremendous


      amount of information coming out of other excipient


      effects on transporters, and so forth.  So, this


      was an evolving issue at that time, and it


      continues to be evolving issue, and methodologies




      were being proposed of in vitro evaluation of this.


      I won't get into that.


                But we also did an extensive evaluation of


      excipients and what we found was I think in most


      cases, excipients that are used in solid dosage


      form really do not have a significant impact, but


      the way we had to evaluate that was comparing the


      differences in formulation that we have approved,


      like, for example, verapamil, and so forth.


                But the risk factor was excipients.  Is


      the current approach for evaluating excipients for


      decisions related to biowaiver of oral solutions


      sufficient?  Well, I think we left the question




                For BCS-based biowaivers, a higher


      standard was adopted by limiting biowaivers to


      highly permeable drugs. Excipients used in solid


      oral products are less likely to impact drug


      absorption compared to liquid oral products,


      because it was simply the volume and amount in




                For example, we had products on the market




      that contains 23 grams of sorbitol in one dose, so


      you can see cross interaction possibly.


                High permeability attribute reduces the


      risk of bio-in-equivalence, decreased small


      intestinal residence time by osmotic pressure,


      because low permeability generally have a tendency


      to be absorbed in small segments of the intestine.


                Clearly, on the other hand, the boundary


      that we chose for high permeability to be 90,


      because there are other surfactants, and so forth,


      that could increase permeability, so if you set


      your boundary at 90, there is no risk of failing.


                There were other examples.  There were so


      many examples that we had not really paid attention


      to.  Here is a submission--not a submission--this


      is a graph that a student of mine sent me from a


      company, and they have seen such effect.


                Here is a drug, a tablet and a solution.


      The solution has almost half the bioavailability of


      a tablet, so you can see that, with sorbitol or


      mannitol for a low permeability drug, it can have


      lower bioavailability.


                So, that was the basis that we came up


      with the recommendations and the boundaries for the


      BCS classification system that gave the basis for




      waiver for new drug applications.


                So, Class I drug, you have jejunal


      permeability. This was our research that we


      classified the number of drugs, and the volume of


      water required to dissolve the dose on the x axis.


      Class I dissolution in vivo is not likely to be


      rate limiting, and well characterized excipients.


      So, dissolution itself is likely to be rapid


      inherently, and then we can rely on in vitro


      dissolution for that purpose.


                Class II dissolution is likely to be rate


      determining and complex in vivo dissolution, and


      solubilization process, so no, not going there.


                Class III was where the debate was.  Some


      hesitation with the use of current dissolution


      test, because the site specific absorption was a


      concern, and excipients.


                Class IV was generally problem drugs with


      in vitro dissolution may not be reliable.


                So, that was the basis for our




                So, wrapping up, in terms of quality


      design thinking, what can we do now?  If I


      summarize my concerns, one major concern was if we


      went towards a dissolution-based method, people




      will design products for dissolution rather than


      the intended use was a concern, and that example


      sort of illustrates that, the one to-be-marketed


      difference, that was a concern.


                So, I wanted to feel comfortable having


      some formulation assessment as part of this


      extension.  Clearly, I think excipients and the


      transporters, all were evolving issues at that


      time.  There is a lot more information available


      now than we had.


                So, in quality-by-design thinking, you


      really have focus on what are the critical


      variables that affect dissolution, and these are


      easily identifiable especially for immediate


      release dosage form.


                You easily can start thinking about




      excipients and what their impact might be on


      solubilization, and so forth, and Lawrence himself


      has done some work in this area, and so forth.


                So, with that in mind, what my thoughts


      are, in addition to thinking about evaluating the


      boundaries themselves, I would like to recommend to


      the group that, first of all, BCS should be, and


      probably will be, a key, too, in quality-by-design


      decision trees that we talked about.


                I mean solubility, permeability


      characterization has to be a starting point for the


      formulation, so clearly, I think we need to build


      these concepts in the decision trees we talked to


      you about yesterday, but also quality-by-design and


      design space with respect to pre- and post-approval


      by bridging studies.  Waivers for in vivo studies


      based on design space concept, sort of is that


      connection to extension concept.


                The challenge I think is from a generic


      industry perspective, there is a lot of hesitation


      to seek for approval, the first time approval, a


      biowaiver based on that, concerned with




      permeability assessment, and so forth.


      That concern probably will remain.


                It is not a scientific concern, it is a


      perceptional concern, it's a regulatory concern,


      and so forth, but that doesn't say that you cannot


      classify a drug for a generic product after


      approval.  The rest of the post-approval changes,


      they could be based on that, and that could be


      quality by design.


                I think those are the key connections that


      if the Technical Committee sort of starts building


      in, their efforts really get connected to the PAT


      and the quality-by-design thinking, so you see the


      connections sort of evolved.


                Clearly, they have already started


      developing in FDA's knowledge base, a knowledge


      base.  Drug-excipient interaction, I think is an


      increasing issue from chemistry and clinical


      pharmacology aspect, and I think you need to start


      connecting those dots.


                At the same time, drug substance and


      formulation variables and clinical performance that




      Mehul alluded to the variability, but with what is


      happening on the clinical side now, with focus on


      biomarkers, focus on surrogate markers, and so


      forth, I think we need to proactively keep an eye


      on that.


                I mean this is an evolving issue, but seek


      out some connectivity between quality and clinical,


      and be available to what is happening at least, and


      that is the point I made also to Jurgen yesterday


      when he gave his report.


                So, with that, let me stop and open it for




                DR. COONEY:  Thank you, Ajaz.


                Comments and questions from the Committee?


                DR. MEYER:  While I am formulating my


      question, I will ask another one, so I will give


      you some time to think.


                How confident are you that our knowledge


      base on excipients allows a reviewer to sit back


      and say, well, it has X, Y, and Z, and therefore,


      there is no problem?


                DR. HUSSAIN:  Well, I think the




      traditional excipients for immediate release dosage


      forms that we use as formulation aids for process


      ability, and so forth, I am fairly confident that


      there are very little concern there.


                There are other excipients that are


      necessary to aid in the dissolution process, such


      as surfactants and other aspect.  I think a close


      grouping of those would be necessary.  I think if


      we collect this information, it will start making


      the case, but if you have properties, such as high


      solubility, and so forth, you probably would not


      need those anyway.


                So, I think you can carve out excipients


      that we know are not an issue.


                DR. COONEY:  If I can just pick up on


      Marvin's point for a moment, your initial question


      was about knowledge base of the excipients per se,


      but it is really the relationship of the excipients


      to the drug substance, to the API, that seems to be


      the area of uncertainty, and that knowledge, it


      seems to me, is much less clear.


                DR. HUSSAIN:  If I amy sort of put that in




      the context, traditional screening experiments that


      are done for drug excipient compatibility, may


      provide not only information that will be useful


      for stability, failure modes of the product, but


      also hints about the interactions among excipients


      and drugs, how they might have a bearing on




                DR. COONEY:  Ken.


                DR. MORRIS:  The point you just raised is


      the source of my question, as well, since


      drug-excipient interactions are typically for


      chemical stability.


                Is it like particularly for something like


      BCS Class III, is it more of the concern that the


      interaction with the drug is changing absorption,


      or that the interaction of the excipient with the


      mucosa or the sites of absorption?


                DR. HUSSAIN:  I think they are both


      concerns in the sense, but the concerns we had when


      we were working on this were more on the impact on


      the GI membrane, transporters, and so forth, the


      concerns with excipient-drug interactions that




      might be physicochemical were less of a concern,


      because we did not really focus on that aspect.


                DR. MORRIS:  That is sort of where I would


      assume it, but I just don't know enough biopharm.


                DR. HUSSAIN:  The aspect I think is this.


      I think the draft guidance that we have issued on


      polymorphism, for example, I think is a


      concentration there, in the sense what we have said


      is you could have a different polymorph, but if you


      can design your product well and if it meets the


      criteria, it's fine.


                So, I think that is the flexibility that


      already sort of comes through that, is that ability


      to demonstrate that, you can be different, but yet


      meet the intended use.


                DR. COONEY:  Art, then Marvin.


                DR. KIBBE:  This is more complex than we


      can handle in one or two days.  The number of


      excipients you use exponentially increases the


      possibility that one excipient is reacting with


      another excipient, that is reacting with the API.


                On top of that, the processing of the




      product changes things.  I mean we know all sorts


      of problems with mag. stearate and overblending,


      but we know certain issues. The question that


      always sits in the back of my mind is we have been


      discovering these issues on a regular basis over


      the last 20 years, have we discovered them all, and


      how can I be naive enough to think I have


      discovered them all, and I don't think I have.


                So, that gives me a basic uncomfort level


      with just waiving stuff when I really want a part


      that works in my patients.  So, I am uncomfortable


      with that.


                I have a question, a substantive question.


      If I make a soft gelatin capsule which contains a


      solution, is it a solution or is it a capsule?


                DR. HUSSAIN:  Capsule.  That's the way it




                DR. KIBBE:  On your basic I guess third


      slide, I could argue it's solution.


                DR. HUSSAIN:  Yes.  That's the aspect, I


      think now we can start thinking about those aspects


      if you have not done so in the past.


                But let me go back to the concern you


      raised in the sense, impact of magnesium stearate,


      and on dissolution, it is clearly documented.




      Impact of magnesium stearate on in vivo absorption


      has not been done yet.


                All the studies we have done, we had no


      impact of magnesium stearate on immediate release,


      and so forth, on in vivo absorption.  I could not


      find a single paper that conclusively tells me that


      what we see in vitro, the big difference is


      translating in vivo differences.


                There are two reasons for that.  One, is


      that old study that was published in 1967, J. Pharm


      Science, by Professor Newton from the University of


      London, where he demonstrated for lithium carbonate


      that if you include a very, very small amount, 0.01


      percent or 0.001 percent of sodium sulfate in your


      formulation, you negate the effect of magnesium


      stearate that you see in the solution.


                So, that was a hint to me that suggested


      that the surface tension differences that we see


      between in vivo fluids and in vitro fluids probably




      are the reason, because all the studies we did at


      the University of Maryland, we actually probed this


      for a low solubility drug--I am forgetting the name


      of the drug--we didn't see in vivo relevance of




                So, now that is the reason we have to


      start thinking about is risk-based decision to


      really understand the behavior of things in vitro,


      because one of the concerns that we had earlier was


      you see big differences in vitro, how do you know


      this will or will not translate.


                If quality-by-design, we are thinking, why


      is this assessment, then, that provides a basis to


      think about it.


                DR. KIBBE:  But that argument, I think


      would logically lead us to the conclusion that we


      have to go to a bioavailability study, we have to


      go to a clinical trial.  We can't rely on any of


      the standard tests that we do that are surrogates,


      because they don't work out, because they either


      show a problem that isn't real, or they ignore a


      problem that is real.


                DR. HUSSAIN:  Right.  That second bullet,


      that is what I am really thinking about.  If you


      have to qualify your design space, your




      bioavailability studies, if you are a new drug


      applicant, that becomes your test of hypothesis is


      to say that we have looked at these are the big


      differences we see that has an impact.


                So, one category of BCS-based biowaivers


      would be SUPAC related where you have demonstrated


      this in vivo, and that becomes the basis for that,


      and not just rely on in vitro testing and lack of


      that information.


                So, waiver is an extension of SUPAC in


      terms of design space is a bigger opportunity


      probably in the quality-by-design thinking.


                DR. COONEY:  Marvin.


                DR. MEYER:  Ajaz, I think I asked this


      question yesterday, and I think you said you were


      going to address it and maybe you did and I missed


      it, the rigid fixing of the--


                DR. HUSSAIN:  Boundaries.


                DR. MEYER:  --boundaries.  I really don't




      have any problem with the rigid definition of high


      solubility, high permeability, I mean we have


      pretty well nailed that down, but then you have, of


      course, if it's soluble in greater or less than or


      equal to 250, well, if it's only soluble in 300, is


      that really poor solubility, and if it's only 89


      percent absorbed, is that really low permeability,


      and, if so, does it fall in that 30 percent


      probability of failing a product?


                How do you deal with--you have to draw the


      line, but on the other hand, you draw the line, it


      becomes somewhat arbitrary and capricious.


                DR. HUSSAIN:  Very good point.  This is I


      think an important point because the objective of


      this guidance was to make the decision.  This is


      the decision.  You meet this, there is no issue.


      If you don't meet, you always have an option to


      explain, but nobody uses that option.


                So, this, in my opinion, is an approach


      that we had before.  In the new paradigm, the


      decision trees that we developed opens the door in


      a sense.  Here, the decision is pre-made, but




      instead of premaking that decision, how can you


      allow your science to drive a decision process that


      can justify the recommendations that come, but that


      then becomes specific to a company.  It is not a


      general guidance.  It is a decision tree to arrive


      at a proper decision.


                So, that would be an extension concept for


      BCS, not a general decision recommendation, which


      is what we have been trying to achieve.  It has


      changed the boundaries, and so forth.  But to


      incorporate that as a decision process, it becomes


      demonstrate this, and the decision can be yours


      sort of a thing.


                DR. MEYER:  So, it's sort of a work in


      progress, so to speak.


                DR. HUSSAIN:  It's not.  I mean I think


      the group has been busy with a number of things,


      but this isn't a thing that they could start


      thinking about, we have not done so.


                DR. COONEY:  Ajaz, if I can get clarity on


      that point.  I think the point you just made is


      that the decisions on class membership will be




      integrated into the thinking about the decision




                DR. HUSSAIN:  Yes.


                DR. COONEY:  Is that correct?


                DR. HUSSAIN:  Definitely.  If you are a


      Class I drug and you exhibit the rapid dissolution


      with the conditions we have outlined, the decision


      is okay.  Anything else, the guidance does not


      recommend a waiver.  That's how it is.


                Based on what Marvin just suggested, and


      what I am formulating that as, this is a decision


      for every sponsor right now.  Their design and


      process understanding would vary from one end to


      the other end, but one way of extending this


      concept is not a general decision that this is


      where you get the waiver, but to define a decision


      tree and how you demonstrate to the degree of


      confidence that we need, that waiver would be you


      have demonstrated an understanding that the waiver


      will be granted.


                That will be Class I, that will be Class


      II, that will be Class III, so you have different




      levels of complexity in those, but the signs and


      the level of understanding could drive you to that,


      but that probably will become a post-approval as


      part of the design space.


                DR. COONEY:  Tom.


                DR. LAYLOFF:  I think I am not confused,


      but I don't understand some things like when a drug


      goes into the intestinal space, it is bound, not by


      water, but probably proteins and various other


      things that are present in the medium, and then it


      is absorbed through different sites depending on


      how it is wrapped in with the rest of the medium,


      and that is a drug-specific property, which then


      can be affected by an excipient, which might change


      the transport site, it may change the structure of


      the solution characteristics, is that correct?


                DR. HUSSAIN:  No, I think the basic


      premise is this, yes, you can have binding, you can


      have a number of other complexation reactions, and


      so forth, that occur.  Many of those are ionic and


      loose, so you establish equilibrium.


                For some drugs, you have complexation that




      really are almost very tight like with calcium and


      tetracycline, and so forth, there are a few such


      examples, but in this scenario, what we are talking


      about are equivalent behavior of the same drug


      molecule in two different formulations.


                So, if there are intrinsic properties of


      the drug molecule itself that will contribute, but


      that molecule is the same, that we are dealing with


      two formulations.  Now, how do formulations act


      with that behavior is a concern.


                I will sort of extend that concern to a


      paper that we had, a poster that we had, is that of


      precipitation.  A weak basis will dissolve very


      rapidly in acid conditions, but when they get


      emptied, there is a potential for precipitation,


      and so forth, and that could be a very complex


      process, and the size of the particles, not


      precipitation, crystallization may differ based on


      the excipients you have and the conditions you




                There is a potential that excipients could


      impact that.  So, that is generally Class II drugs,




      that's the boundary for Class I for high solubility


      was intended to prevent some of those things from


      occurring, too.


                DR. LAYLOFF:  Do you think the


      complexation and coordination around the drug


      substance would actually affect the transport site,


      change the site of transport, would change the


      properties of the system?


                DR. HUSSAIN:  Yes, it is clearly possible,


      but unlikely for an immediate release dosage form


      with the type of excipients we use.


                DR. MORRIS:  You know when I think about,


      it sort of makes my head hurt, but when I think




                DR. HUSSAIN:  It's complex.


                DR. MORRIS:  --the amount of time we spent


      working on developing design spaces for the


      processing end of things, which as Jerry says, may


      be a way off, but still in comparison relatively


      simple to the larger problem, is there an


      opportunity in the context of using development


      data to somehow leverage tox studies to be able to




      get early reads on, not the tox itself, but in


      terms of some of the dynamics that are going on


      with the dosage forms?


                DR. HUSSAIN:  I have a decision tree,


      which I did not present yesterday, but it was part


      of the handout.  That decision tree came out of our


      AAPS workshop on how to leverage that.  The paper


      is published, Diane Burgess [ph], Eric Duffy from


      FDA is on it, so it is there in your handout.  I


      don't have it in this one.  Take a look at it.


                That leverage is every bit of information


      coming from Pharm Tox, and so forth, to start


      building that case for that.


                DR. MORRIS:  For the design of the dosage




                DR. HUSSAIN:  Yes, for particle size


      dissolution and bioavailability concentration.


                DR. COONEY:  Marvin.


                DR. MEYER:  Ajaz, when a generic company


      sends in their ANDA, it was my understanding that


      the generic group does not go back to the NDA to


      review the contents of the NDA, so they don't look




      at the excipients and see which excipients are now


      different in the ANDA than were in the NDA product?


                DR. HUSSAIN:  I will let Lawrence answer


      that, but we do have a process of inactive


      ingredient guide that we consult, and so forth.  I


      put him on the spot here.


                DR. YU:  I guess this morning we talk


      about excipients, which emphasize how complex the


      process is.  Yes, with advances in molecular


      biology, we discovered I even don't know how many


      transporters going on.  As far as the PGP, at least


      32 and 64 is transporters, however, I want to


      emphasize that how those transporters impact


      absorption we rarely see in clinical settings.


                In other words, we very see excipients


      impact on absorption of Drug A, B, C, D, but in


      vivo setting, we have very, very few, two or three


      publications out there compared to tens of


      thousands of publications to show that excipients


      impact in vitro.


                So, I have to say that we still want to


      see more evidence to show the impact of excipients




      on absorption of drugs in general.


                Secondly, while we see the impact exceeds


      absorption, we very open to see the unique of some


      of the products out there.  The reality for, say,


      70 or 60 percent of immediate release products,


      people espouse the intensity, use very limited


      number of excipients, I would say 10, within 10.


      For example, Avacel almost uses the majority of


      products.  All those excipients impact, and have


      not seen in vitro, as well as in vivo.


                Certain, because of those common used


      excipients, since we have a sufficient knowledge to


      judge whether they are going to impact absorption


      or not, will generate and not see the formulations,


      however, in very few cases, some cases, we suspect


      potential impact of excipients absorption, we will


      look into it further before we make any scientific


      decisions about approvability of any NDAs.


                I hope that answers his question.


                DR. HUSSAIN:  The other aspect, just to


      build on what Lawrence said, traditionally, the


      composition, especially for immediate release, it




      is hardly any different than the quality.


                DR. MEYER:  My real question was do you go


      back and look at the NDA to see if Pfizer used


      Avacel, and Teva used who knows what, do you make


      that comparison, say, well, wait a minute, they are


      putting in two things instead of Pfizer's one




                Can that potentially make a difference?


      Do you review the NDA product composition?


                DR. YU:  Well, certainly, we will review


      any scientific literature out there and information


      available to us to make the best decisions.


                DR. MEYER:  But do you review the NDA?


                DR. HUSSAIN:  Marvin, often we don't have


      to, because it's in the label.


                DR. MEYER:  Well, that's true.  It didn't


      used to be.


                DR. HUSSAIN:  But definitely, the criteria


      there is to look at what has been approved and what


      has been used in dosage forms and inactive


      ingredient guide, and so forth.


                DR. YU:  I guess the answer is as long as




      are trying to build in the science base or any ANDA




      we will use any information which is available,


      whether scientific literature or not, to us.


                DR. COONEY:  Are there any other questions


      at the moment?  Thank you, Ajaz.


                We are running a bit ahead of schedule.  I


      think this would an appropriate time to take a


      break for 15 minutes.  We will reconvene at 8 past


      10:00 and then begin immediately with Lawrence Yu's






                DR. COONEY:  I appreciate everyone's


      diligence to staying on time.  It has worked very




                Lawrence Yu will proceed with a


      presentation on Using Product Development


      Information to Address the Challenge of Highly


      Variable Drugs.




            Using Product Development Information to Address


                 the Challenge of Highly Variable Drugs


                DR. YU:  The assignment to me today, this


      morning, is for me to discuss how to use


      pharmaceutical development information to address




      or potentially address the bioequivalence issues of


      highly variable drugs.


                Before I go on and talk about how to use


      or potentially use the pharmaceutical development


      information to highly variable drugs by equivalency


      issues, I want to give you a very brief overview or


      update what has been happening before.


                For highly variable drugs, this is not the


      first time, it's the second time we present it to


      you.  In the first presentation on April 14th of


      2004, we discussed the challenges and the


      opportunities for bioequivalence of highly variable




                At this meeting, the objective was to


      explore and define bioequivalence issues of highly


      variable drugs, for example, what is called highly


      variable drugs and discuss potential solutions to


      deal with the bioequivalence of highly variable




                We invited a number of speakers from


      industry, academia to address issues related to


      bioequivalence including why the bioequivalence of


      highly variable drugs is an issue, highly variable


      drugs a source of variability by Gordon Amidon from


      the University of Michigan, and the clinical




      implications of highly variable drugs by Leslie


      Benet, and from bioequivalence method include the


      skin method by Laszlo, as well as bioequivalence of


      highly variable drugs, we had a good discussion at


      this meeting.


                I want to highlight some of the things


      which have been discussed at this meeting,


      particularly the slides by Professor Leslie Benet


      from the USCSF.  His talk with implications of


      highly variable drugs, the argument was why highly


      variable drugs are safer.


                Specifically, he said for wide therapeutic


      index highly variable drugs, we should not have to


      study the excessive number of patients to confirm,


      to demonstrate that two equivalent products meet


      the preset statistical criteria or by equivalent






                This is because, by definition, highly


      variable approved drugs must have a wide


      therapeutic index, otherwise, there have been


      significant safety issues and lack of efficacy


      during Phase III.


                Highly variable narrow therapeutic index


      drugs are dropped in Phase II since it is


      impossible to prove either efficacy or safety.


                Now, for the benefit of some new members


      for this committee, I have two slides to briefly


      review why this issue, why the one-size-fits-all,


      what we are using today.


                In order to determine bioequivalence, we


      normally define as a rate of bioavailability,


      defined as a rate and extent of drug absorption.


      Bioequivalence is defined as absence of significant


      difference in the rate and extent absorption.


                In practice, when we give the drugs


      orally, for example, to a healthy volunteer, we


      will draw the blood.  We got the plasma


      concentration profile.  We are certainly not able




      to get exactly how much and how fast drug gets


      absorbed, therefore, in practice, we use AUC, area


      under the curve, to represent extent of absorption.


                We use Cmax as a surrogate for the rate of


      absorption, certainly in some cases we also look at


      Tmax, because indeed, if you look at Tmax and Cmax


      here, it represents the rate of absorption.


                So, from that, we will define what the


      bioequivalence study is passing or not passing.


      Basically, the bioequivalence criteria, either


      statistical criteria here is 80 to 125 percent.


                At this date, that is the


      one-size-fits-all regardless drug, drug product,


      regardless of therapeutic class, regardless for


      anything, that bioequivalence study, you have to


      use preset, so-called bioequivalence criteria,


      which is 80 to 125 percent.


                Now, let's look and explain why the highly


      variable drug is an issue.  Let's look at the red


      one.  If you use a highly variable or intersubject


      variability is high.  Statistical confidence


      interval, if you use the same number of subjects,




      when variability goes higher, the confidence is


      going wider.  When confidence gets wider, it


      becomes more and more difficult pass the confidence


      interval or bioequivalence interval if 80 to 125




                So, that explains when the variability


      goes higher, it gets more and more difficult to


      pass a study.


                On the other side, in order to narrow the


      confidence interval, for example, here it is fair


      to demonstrate bioequivalence for super red one


      here in order to make confidence interval narrower,


      you have to use a large number of subjects, because


      the higher the variability, the higher the


      confidence interval, the higher the number of


      subjects in general, the narrower the confidence




                Therefore, for highly variable drugs, you


      will have to use higher number of subjects.  Just


      to give you example here, for example, normally, we


      have a 20 percent or 30 percent intersubject


      variability.  You maybe use 18, 24, even sometimes




      for good product or good drugs, you only need to


      use 12, actually, they can pass the bioequivalence


      confidence interval.


                But this is not always true, because when


      the variability goes higher, now, this variability


      could be because of a drug, or it could be because


      of a product, so think about if variability goes


      100 percent--some of you think 100 percent, that is


      unrealistic, but we do have a drug, we do have


      examples--think about with 100 percent variability,


      assume test and the reference, there is 5 percent


      difference, you end up it could be 500 or more


      subjects, or 300 subjects, so this is certainly a


      large number of subjects in order to pass the


      bioequivalence study.


                So, Leslie argued at the previous meeting,


      from the clinical perspective, this is not


      necessary.  To give you a real example, these are


      slides from Leslie Benet.  Now, you would argue,


      you may ask how do we get intersubject variability.


                Certainly, you could get this number from


      literature or sometimes company conduct a pilot




      study, get some kind of estimate how many subjects


      need to be used to pass the bioequivalence study.


                Of course, in this case, based on


      intersubject variability, you need to use 300, now


      this is the drug.


                So, at the previous meeting, when we


      present the issue to you, and also we present some


      of the possible potential solutions including


      widening the confidence interval.  Now, that is


      very straightforward.  You said by confidence 80 to


      125 is too narrow for highly variable drugs, and


      your intuitive thinking is just widening the


      confidence interval, that is one of the potential




                Another solution is a scaling approach, in


      other words, based on the variability of reference


      list product, reference list drugs, and calculate,


      use statistical approaches to calculate the


      confidence interval, then, to determine whether the


      study is passing or not.


                Obviously, I have to say this.  The active


      approach because the confidence interval too wide




      in order for the study to pass, so let's widening


      the standards.


                You came in and asked to do that, that


      certainly the scaling approach, we ought to


      carefully look into, the Committee suggested--a


      quote here--"the need to understanding where the


      variability originated.  The members added that


      prior knowledge of all biostudies may help set more


      appropriate specifications or criteria to make




                So, you suggest that we have to understand


      the origin of the variability.  Now, to look at the


      mechanistic understanding of variability for drug


      substance obviously is the same, reference list


      product and the generic product, or any other


      product, but the potential difference could be


      formulation.  Certainly, the generic products could


      be narrower or could be wider, the variability.


                We believe at this point, in order to


      understand the origin of the variability, that


      pharmaceutical development report, or


      pharmaceutical development information can help us




      understand the source of variability, can help us


      make rigid scientific evaluation.


                Now, in order to see the utility of


      pharmaceutical development report to evaluation or


      reviews of NDAs, let me go back and review some of


      the basic fundamentals or the premises for ANDA




                Ajaz has mentioned about therapeutic


      equivalence. Basically, the products are considered


      to be therapeutic equivalents only if they are


      approved as safe and effective, they are


      pharmaceutically equivalent, they are


      bioequivalent, adequately labeled, and manufactured


      according to cGMP.


                Now, here, I want to emphasize the


      pharmaceutical equivalence.  When we define


      pharmaceutical equivalence, we basically have the


      same active ingredients, obvious.  I know we are


      managing about pharmaceutical solid polymorphism,


      which was presented to you two years ago, has a


      drugs guidance out there and published by FDA in


      December of 2004.


                You have to be same dosage form, same


      route of administration, and identical in strength


      and concentration, and may differ the other




      characteristics, such as shape, excipients,


      packaging, and so on, and so forth.


                Now, under the same dosage form, I think


      what the complexity of dosage form is particularly.


      Yes, I would say several decades ago, that dosage


      form is reasonably simple and in most cases I would


      say the immediate release product or solutions.


                Certainly, with advances of pharmaceutic


      industry and the pharmaceutic technologies,


      so-called dosage form gets more and more complex,


      and we now have the soft gel capsules, we have


      ricin [?] product, we have inhersion [?] product,


      presents additional challenge to us in terms to


      make scientific decisions, in terms of make


      scientific evaluations.


                We therefore would believe a


      pharmaceutical development report,


      quality-by-design, designed to be equivalent,


      become more and more significant in this regard.


                Why does pharmaceutical equivalence


      matter?  Because of user experience and


      expectations.  Then, bioequivalence test is


      normally conducted in healthy subjects.  Certainly,


      we have assumption that equivalence in healthy


      subjects equals equivalence in patients.  Now, we




      have many, many generic products out there which


      are safe, which is a high quality, which are


      effective, which is the equivalent to the reference


      list product.


                So, certainly, we have tremendous


      experience with that, and certainly the


      pharmaceutical equivalence presents more and more,


      become because you want to make sure the data from


      the healthy volunteer does the equivalent in


      patient, and against novel drug delivery systems


      presents a challenge.


                That is why we want to use more and more


      pharmaceutical development approach to make a


      judgment, pharmaceutical development information to


      make a scientific judgment.


                Highly variable drugs very often have, as




      I mentioned, a wide therapeutic index, and the


      clinical trials of reference list product have


      established the acceptable level of variability,


      because I said otherwise, these highly variable


      drugs, a big window index, they will be dropped in


      Phase II.


                So, under an ideal situation, you will


      think about variabilities very high, so you will


      think it should be easier to pass, easier to design


      equivalent product simply because they are so wide,


      the target is wide, so it is easier for you to




                Obviously, as I said before, if you use


      the preset 80 to 125 bioequivalence confidence


      interval, it is not the case.  While we explore our


      tentative approaches to deal with the


      bioequivalency issues, certainly, the design issues


      come out.


                So, now, how do we deal with


      pharmaceutical development for highly variable


      drugs?  Obviously, sponsor need to understand what


      are reference products supposed to do with origin




      of variability, and the purpose of design can be


      equivalent, and to evaluate and to verify the


      design and hopefully, in the future, use the


      bioequivalent study design for highly variable




                So, we put more emphasis on design in this


      regard to establish pharmaceutical equivalence, in


      order to establish therapeutic equivalence, which


      will be more appropriate.


                While we are looking for shared


      information for generics with us, there is a reason


      for doing that, not only for evaluation for highly


      variable drugs, certainly for pharmaceutical


      development is required.  It's one of the CTD


      format.  It is also outlined in ICH Q8, although


      they do not apply to us, but I think some


      principles should apply to generic industry also.


                Also, more significantly, OGD question


      based on review.  Now, this is still a work in


      progress, but I want to share some questions, I


      think it is important to ask to share.


                What is the formulation intended to do? 




      What mechanism does it use to accomplish this?


      Were any other formulation alternatives


      investigated and how did they perform?  Is the


      formulation design consistent with the dosage form


      classification in the label?


                So, those questions will help us get


      information about a pharmaceutical product, the


      report will help us, pharmaceutical product design


      and development, make more sound and appropriate


      scientific determination or evaluation.


                The question often comes up, why do we


      need to provide those things to the FDA?  That


      again is a quality-by-design paradigm, and


      pharmaceutical development report is where you


      demonstrate the drug is highly variable.


                Now, this demonstrate not necessarily to


      study, but you certainly use any source available


      to show why this drug or drug product is highly


      variable, and may use a different criteria other


      than 80 to 125 percent confidence interval.


                Also, the pharmaceutical development is


      where you justify equivalence of design, why do you




      think the product which you designed is equivalent


      to the reference list product.


                During the discussion, the members ask


      whether it is drug or drug product.  Now, for


      example, Product A, the variability is the active


      ingredient into exceptions, so formulation design


      could be rapid release, so demonstrated by


      dissolution comparison under physiologically


      relevant conditions, if this is BCS Class I drug,


      which is highly soluble, highly permeable, even


      though they are highly variable, you may still


      require biowaiver, otherwise, you will have to


      conduct some bioequivalence studies to demonstrate


      that they are bioequivalent.


                Certainly, the approaches to deal with


      highly variable drugs, to deal with the


      bioequivalence of highly variable drugs are still


      in discussion and still in investigation.  I am


      hoping in the near future we share with you some


      proposal or recommendation we have with respect to


      bioequivalence of highly variable drugs.


                Another drug could be drug product, the




      drug product could be highly variable, even drug


      substance is I would say low variability, and


      certainly design for equivalence begins with the


      characterization of the reference list product, and


      generic product should target the mean, and the


      current system again would have no reward for


      narrow or less variability of generic products.


      That is why we need to explore the alternative


      approaches or more appropriate approaches to deal


      with highly variable drug products.


                I just want to give you some examples of


      what we talk about here.  This is real data.  This


      is single subject replicate design, in other words,


      you give the same product to the same patient


      twice.  Here is the plasma level, obviously, I am


      sure that out of 80 to 125 percent confidence


      interval, by any standards, it probably does not


      need a statistician to figure this out.


                You can see here, this is the first period


      or this is the second period.  It is not in your


      handout or printout because this is in color.  If


      you look carefully, these two curves are




      significantly different, probably different by I


      even don't know how many folders.


                This is a single-dose study twice,


      replicate study design.  Sorry, I should do a


      better job next time, use red, so you can see it.


                DR. KIBBE:  I am just looking at the curve


      and wondering why we got the hump at the back end


      and whether the product is intended to have a


      second release.


                DR. YU:  No, it's simply by design, for


      whatever reason this peak has come out.  Obviously,


      a second dose, this peak is no longer there.  So,


      it's not purpose designed, it's simply because of


      physiology involved.


                This is happening because enteric coated,


      this is coated to release at the target pH, so when


      the physiological pH in the gastrointestinal tract


      may fluctuate, and those curves will change.


                Think about, for example, if we have a


      product designed will release a pH 7, so then in


      the terminal ileum, at one point is pH 6.8, you


      will not see the release.  But a second day,




      because of food or because of other reasons,


      terminal ileum pH becomes 7 or 7.2, you do not see


      the release.  Otherwise, you will not see it.


                So, simply pH effect or significant impact


      the absorption.


                DR. KIBBE:  The product had gotten on the


      market because it worked clinically?


                DR. YU:  Yes.  Even though we see the


      significant variation in pharmacokinetics, but we


      have no reason to believe this variation will


      impact safety and efficacy.


                So, in order to do more appropriate


      pharmacokinetic studies, we also look into what


      additional information, for example, develop


      information will help support those cases or


      bioequivalence cases, because you can see the


      bioequivalence obviously is very difficult to


      conduct variability probably up to 2 or 100


      percent, and the number of subjects very high.


                So, we want to see can we use any


      additional pharmaceutical development information


      to help us to make more proper scientific






                Again, for example, when we are looking,


      in many cases, we do get very consistent, the in


      vitro dissolution actually out to say the majority


      of cases, those help us out to make a more proper


      scientific decision or rational scientific decision


      when we recommend any method to demonstrate


      bioequivalence, but occasionally, we do get very


      strange results, and actually, the variability is


      extremely high and does not help you.


                I just want to show you another case here


      when we conduct the dissolution under physiological


      relevant pH condition, and you get dissolution all


      over the place.


                Now, this is a 6 tablet, same lots, same


      bottle, put in 6 vessels, you get a distortion




                So, the next question we ask, this is a


      large variability because of the operator or


      because of other reasons.  I think the answer is we


      are almost certain those difference is because of


      product, not because of other factors.


                So, what I can present to you today is we


      have challenges to deal with bioequivalence of


      highly variable drugs.  We use the clinical




      evaluation and sometimes we are also facing


      challenges when we are trying to use in vitro


      information to help us make decisions.


                DR. SINGPURWALLA:  Lawrence, what is the


      difference between each curve, different vessels?


                DR. YU:  Yes, six vessels.


                DR. SINGPURWALLA:  Six vessels, so it


      could be that the vessels are different.


                DR. YU:  I think I stated that the


      variability because other reasons, for example,


      vessel difference, media difference, degassing


      difference, operator difference, assay difference.


      We do not believe all these reasons can explain.


                DR. HUSSAIN:  Yesterday, this same figure,


      Cindy actually showed you the reason for this


      difference was the coating thickness, and so forth,


      so this is the same slide.


                DR. MORRIS:  You wouldn't get 2 1/2 hours


      difference in dissolution time from different




      vessels.  That is not the magnitude you would




                DR. SINGPURWALLA:  How am I supposed to


      believe that?


                DR. YU:  You have to believe in me.  You


      don't have any other options.




                DR. SINGPURWALLA:  I don't believe in




                DR. KIBBE:  This is a constant pH


      throughout, right, we haven't shifted pH during the


      process or anything, right?


                DR. YU:  Correct.


                DR. HUSSAIN:  The reason to believe that


      is I think it was done by Cindy, and with our


      stringent mechanical calibration.


                DR. LAYLOFF:  He demonstrated it with


      variable coating.  It's variable coating on enteric


      coating material, so if there is a crack in the


      coating, it disintegrates much more rapidly.


                DR. SINGPURWALLA:  I think I believe Tom.


                DR. YU:  Thank you very much.


                So, the objective, the case we presented


      to you is certainly difficult, I just want to say,


      variability issue, whether from clinical evaluation




      or sometimes for in vitro conditions, in vitro


      testing, target main performance question.  I am


      sure you will ask where is the main performance.


                I just want to show you that these are the


      challenges which we are facing, and certainly we


      are open to any suggestions or input from you.


                So, in summary, we believe pharmaceutical


      development information will help.  I quoted here,


      that's the conclusion made by you April 14th of


      2004.  Understanding what the problem is, as well


      as the real fundamentals, for example, physical and


      chemical parameters, and make coherent and


      scientific science-based decision based on


      pharmaceutical development information, I think I


      present to you the cases to see hopefully how we


      use pharmaceutical development information to help


      us in most cases, but in some cases, we still have


      challenges and we have opportunities for us to move




                Thank you and any comments are welcome.


      Thank you very much.


                DR. COONEY:  Thank you, Lawrence.


                We now have time for questions from the




                DR. SINGPURWALLA:  Lawrence, I have two




      kinds of questions.  Question No. 1.  Is it the


      purpose of this presentation of yours to ask the


      manufacturers, namely, the industry, to provide


      more information to you because there is so much


      variability and you are trying to get to the source


      of the variability, is that the objective?


                DR. YU:  Yes, very precise, certainly much


      better than I said.


                DR. SINGPURWALLA:  That is the political


      question. The scientific question, and I have heard


      this before, what does T/R percent mean in your


      Slide No. 8?


                DR. MEYER:  Test over reference.


                DR. SINGPURWALLA:  Test over reference.


                DR. YU:  Yes.


                DR. SINGPURWALLA:  How was this 80 percent




      and 125 percent figure arrived at?


                DR. YU:  Slide 9.  I am trying to get


      Slide 9.


                DR. SINGPURWALLA:  That's it, the picture.


      So, 80 percent and 125.


                DR. YU:  T is the test.


                DR. SINGPURWALLA:  No, forget that.  How


      did you get 80 and 125?


                DR. YU:  That's an excellent question, and


      we have been asked many times.


                DR. SINGPURWALLA:  It can't be excellent.


                DR. YU:  It's back to it was published


      when I was in high school, I would say, 20 years


      ago, or even more than 20 years ago, when the


      pharmacokinetics, the discipline was developed, and


      FDA developed the criteria.  Actually, this


      evolving process and trying to develop what kind of


      standards or criteria can we use to judge a


      bioequivalence study is okay or is not okay.


                I think at that time, the physicians get


      together, as we do today, and the physicians


      together made the determination that the 20




      percent, the difference between product would not


      be considered clinically significant, because the


      20 percent will not be considered significant


      difference, therefore, when translated into in vivo


      setting, you have 80 percent.


                So, you would think from 80 to 120 instead


      of 25. Now, in the normal processing of


      pharmacokinetic data, they used log normal to be


      much better to describe the distribution.  So, when


      you use log normal, 80 is still 80. When you have


      the 1 over 80 or 1 divided by 0.80, equals 1.25.


      That is why you see 80 to 125.


                Now, at the beginning, I would think 20


      percent instead of 19 percent or 21 percent, which


      is 20 percent, it was decided.  Then, the question


      come back to us now can we change 20 percent to 25


      to 15, 10, 5 percent, and I guess we have to use,


      say, over the 20 or 25 years, we approved product,


      they are all safe, they are all equivalent, they


      are all high quality, because of those experience


      or prior knowledge, determining 80 to 125 percent


      works fine.


                Now, this does not necessarily mean we


      cannot change it, but the criteria we have is very


      stringent criteria, we feel confident with that.




                Now, with a statistical interplay--


                DR. HUSSAIN:  Lawrence, if I may.


                DR. YU:  Yes, please.


                DR. HUSSAIN:  It's a "feel good" criteria,


      we felt good about it.


                DR. SINGPURWALLA:  I got the answer.  I


      think I got the answer.




                DR. SINGPURWALLA:  The answer is




                DR. HUSSAIN:  No, it's rational science.


                DR. YU:  It is rational science.  I think


      I proved it.


                DR. SINGPURWALLA:  Let me make a




                DR. YU:  Yes, please.


                DR. SINGPURWALLA:  That tradition with


      some dose of rationality was good 20 years ago when


      you were in high school.  Times have changed. 




      These kind of decisions to either prove equivalence


      or prove in-equivalence should be based on risk


      considerations and should be based on appropriate




                So, I think it is time to change, and I


      think I said that April 14th, 2004.  Has there been


      any progress made towards changing?


                DR. YU:  The answer is yes.


                DR. SINGPURWALLA:  Oh, good.  What?


                DR. YU:  Certainly, you said you want


      suggestion of change, and I think under the


      leadership of Gary Buehler, that we are exploring


      the confidence interval, for example, the window


      index drugs, and also we are exploring confidence


      interval for highly variable drugs.  In other


      words, in the future, I am hoping someday, with


      your support and agreement, we will have different


      criteria other than 80 to 125 to different class of


      drug in consideration of the risk interplay.


                Obviously, to make any changes, six months


      or one year is not enough.


                DR. COONEY:  Marvin, then Ken, then Paul.


                DR. MEYER:  Your talk I believe tried to


      marry the quality-by-design to the highly variable


      drug and show that you could, in part, solve the




      problem by quality-by-design, that's the objective.


                DR. YU:  Yes.


                DR. MEYER:  Personally, I think if you


      have a competent company, then, your highly


      variable drug is biological problem which the


      company can't solve.  You have to speak directly to


      a higher power to get rid of that variability.


                So, I think, yes, there is cases where,


      for example, I could cite failure by design if you


      want to put an enteric coating on something,


      because that is, in my view, not a good dosage form


      because it is so dependent on gastric pH and


      emptying, and all of that, so you are setting


      yourself up for failure.


                Now, you can say, well, I dealt with


      quality by design by not using enteric coated, I


      kind of took the reverse of that.  A competent


      company looking at Slide 24, the 6-vessel graph,


      would never go to a biostudy with a product that




      showed dissolution characteristics like that.


                So, indeed, some quality built in that


      says whoa, let's not spend $100,000 on a biostudy


      when our drug is all over the map in dissolution.


      So, I think you can deal with some variability, but


      that is fairly straightforward I think for a




                So, the issue that really faces us is the


      physiological variability and do we extend the


      confidence limits, do we have point estimate


      restrictions or just do we do 600-subject studies.


                DR. HUSSAIN:  Marv, may I just sort of put


      that in context a bit?  In some ways, what we are


      seeing here is this.  Since we are comparing two


      formulations of the same drug, the drug is the


      same, the variability, the physiologic, the


      variability that is coming is the same for the drug




                If we can compare formulations and say


      that all the conditions that are critical to


      exposure are well controlled, and so forth, and get


      confidence, what will give us the confidence to say




      that the inherent variability is the physiologic


      variability, not the quality variability, then, we


      can move forward.  I think that is the hope that we




                DR. MEYER:  Do you think practically, you


      can look at the restrictions and the SOPs--


                DR. HUSSAIN:  No.


                DR. MEYER:  --and just see how a company


      is formulating and designing and developing a




                DR. HUSSAIN:  Not with the traditional


      work we do about formulations, putting things


      together, and so forth, no, it has to be a


      structured design approach that goes through


      identifying the sources of variability in your


      materials, and so forth, and putting a convincing


      case together to say based on the assessment, in


      this case it's a generic product, and based on


      characterization of reference material and your


      test product, you can make the case that the


      variability that you are seeing in your product is


      no more different than of the best argument.


                That gives you a leverage to now make a


      rational decision with respect to what sort of a


      biostudy criteria would be necessary.




                You can build flexibility, and not go


      rigid with, say, the Japanese approach, which was


      in your background packet, was to say do we really


      need confidence interval criteria here.  We just


      want to confirm the mean values.  It's a


      confirmation rather than a complete full-fledged


      study.  One option could be that.


                DR. YU:  I think the message we are trying


      to convey is when we explore alternative approach,


      which could be a wide confidence interval, or your


      scaling approach to show or to demonstrate the


      bioequivalence is demonstrated with the additional


      information, which is pharmaceutical development


      information, will help us to make scientific


      coherent decisions.


                Right now we don't, we don't have those


      informations.  In other words, we are not able to


      see how dissolution variability here may change it,


      for example, in this case, if we change the pH,




      dissolution is very beautiful, so that is the data


      we got.  We have now seen this data I showed you on


      the screen.  Thank you.


                DR. COONEY:  Ken.


                DR. MORRIS:  A couple of points.  One is I


      agree with Marvin in the sense that you wouldn't


      expect a company to release dissolution, I mean


      going to a biostate with dissolution like that, but


      I think those studies were done under different


      conditions.  These were done here, so they wouldn't


      have seen that under normal dissolution conditions.


                My more general question is--


                DR. YU:  You are correct, yes, in normal


      conditions, especially, for example, USP


      dissolution, maybe you are not able to see.


      Actually, dissolutions are beautiful.


                DR. MORRIS:  Right, so that comes back to


      sort of our discussions yesterday in a sense.  The


      question I have is to what level or to what extent


      do the ICH initiatives, I mean including the CTD


      and Q8, impact on the ANDA, I mean is there an


      intent that they follow suit with NDAs?


                DR. YU:  Obviously, the basic principles


      from ICH and CTD, the CTD cure document for drug


      substance and drug products, it is not just for NDA




      only, for both NDAs and ANDAs.  ICH Q8 is, in


      principle, a part into NDAs, certainly the basic


      principle also apply to ANDAs.


                The way I actively look into this to


      document and to see what information will help us


      to make scientific decisions.  Certainly, as I said


      before, we are not looking for information which is


      nice to know, we are looking for information which


      is essential to know.


                DR. MORRIS:  I guess to that end, because


      this is something, of course, we have been


      discussing for several years, but the idea that


      rather than having checklists of what the companies


      have to do, if they can make scientific decisions


      based on the intended dosage forms and the


      properties of the API, which should be a lower


      hurdle, I mean that should be known more by the


      time you get to the generic.


                DR. YU:  Yes.


                DR. MORRIS:  Instead of having to do a lot


      of the other testing that might normally be done,


      if they can focus on the identification of the


      critical to quality attributes of the product and


      capture that in a development report, it seems like


      that is a reasonable way forward.




                DR. YU:  That is correct.  In fact,


      industry is coming forward and they share some of


      the pharmaceutical development report with us, we


      are actively looking into this to develop some kind


      of review templates which will incorporate


      pharmaceutical development information into our


      review process.


                Again, I said we are looking for


      information which is essential to know, not nice to




                DR. MORRIS:  Maybe this is for Paul, is


      that a reasonable stance as far as how you look at


      generic development?


                DR. FACKLER:  I am not sure exactly what


      you are asking.


                DR. MORRIS:  I can clarify if you want,




      but basically, if you could, instead of having to


      do sort of checkbox testing, if you could do


      testing that was largely prescribed by your need to


      establish certain scientific issues, rather than


      having to do as many, let's say, sort of--what is


      the word--statutory testing, if you will, is that a


      reasonable stance for you guys?


                DR. FACKLER:  I don't see a problem with


      that.  What I didn't hear here was that there are


      any different statutory requirements for highly


      variable drugs.


                If a generic company still needs to pass a


      bioequivalence study, and we are going to assume


      that the pharmaceutical equivalence is simple, I


      don't understand what the generic company


      understanding the origin of the innovator's


      variability has to do with the approvability of a


      lot of material that is shown to be


      pharmaceutically equivalent and therapeutically


      equivalent through a bioequivalence study.


                I guess that is the piece I am missing.


      Why is the burden on the generic company to




      understand the variability of the reference listed


      drug, and what value does that have if really all


      the generic company needs to do still is


      demonstrate a bioequivalent product?


                DR. MORRIS:  You are talking about BE


      variability now, not pharmaceutical?


                DR. FACKLER:  Yes.


                DR. YU:  Paul, if we use


      one-size-fits-all, which is 80 to 125 percent to


      some of drugs, you may have difficulty to pass the


      confidence interval.  So, when we are exploring the


      alternative approach including the scaling


      approach, you will have to demonstrate this product


      is highly variable or not highly variable.


                You have to know that because otherwise,


      suppose someday in the future, if the scientific is


      mature enough, we have a scaling approach, for


      example, for highly variable drugs, your submitted


      application did not show these are highly variable


      drugs, how would we know these are highly variable




                So, you have to show, in your development




      report, that is a highly variable drug before we


      move forward.


                DR. FACKLER:  Agreed, but wouldn't a


      replicate design bioequivalence study inherently


      capture the variability of the reference listed




                DR. YU:  Yes, if you choose to do so, use


      replicate design, certainly, you are able to


      demonstrate that reference list product is highly


      variable or not.


                DR. FACKLER:  But that is already part of


      an ANDA application is my point.


                DR. YU:  I guess, Paul, we have not


      reached a consensus or we have not made a


      determination you have to use replicate design.


                DR. MORRIS:  Is part of that the fact that


      you are still struggling with the concepts that are


      entailed in that dissolution plot where you can't


      factor into the pharmaceutical variability, factor


      the pharmaceutical variability from the clinical?


                DR. YU:  I guess the struggle we have here


      is, look, Lawrence, in order for you to get this




      direct for pass, whether you use scaling approach


      or you use widen the confidence interval, you


      simply widen the confidence interval, let them to


      pass.  You need to explain why.  You need to


      explain why you think that is a feasible approach,


      you think that is scientifically sound.


                So, when you say explain why the


      pharmaceutical development report can help us


      provide additional information to explain why.


                DR. FACKLER:  I agree that certainly you


      need to understand the variability of the reference


      listed drug especially if a generic applicant is


      claiming that the variability is an issue for this


      particular product.


                DR. YU:  Correct.


                DR. FACKLER:  I am not sure what value the


      steps that were taken has to that determination of


      variability.  Variability sometimes is listed in


      the label for a reference listed drug; other times


      applicants do replicate design studies or run


      reference versus reference to measure that inherent


      variability, but that would all be part of an




      application already, as I understand it.


                DR. YU:  Yes, in many cases actually


      lately for some of complex dosage forms.  Dosage


      form, we very often sent many, many deficiency


      letters.  Actually, company provide information


      during the cycles, and as I said, at the GPA Chair


      meeting, we have four or five or six cycles,


      provide additional information to us, and


      eventually, the product get approval.  I am not


      saying you not provide that information.


                What I am trying to say is with the arena


      of pharmaceutical development report in the ICHQ


      paradigm, can you provide that information in the


      application instead of for us to send many


      deficiency letters.


                When we see the OGD list receive 25


      percent or more of the applications every year,


      where do you want to put resource into those


      reviews.  Suppose you provide those additional


      information, which I believe will help us in our


      reviews, and they reduce the cycles, I see it's a


      win/win situation for you and for us.


                DR. COONEY:  Ajaz.


                DR. HUSSAIN:  I think look at it from this


      perspective in the sense the whole aspect is you




      are trying to make a decision and you are trying to


      choose the right measurement system here.


                Now, the Code of Federal Regulations


      essentially has a hierarchy of methods that you


      choose for bioequivalence.  Our current criteria is


      a PK crossover, PK-based, pharmacokinetic-based


      study is the most discriminating one.


                So, you are looking at, you are trying to


      now judge approvability of a generic drug, and for


      that you need to establish its pharmaceutical


      equivalence and its bioequivalence.  The


      bioequivalence measurement system that we have has


      inherent variability, and much of that variability


      is coming from the measurement system, and may not


      be coming from the test samples that you are doing.


                So, is this measurement system the ideal


      measurement system right now or not?  That is


      really the question.


                The dilemma that we have is the in vitro




      characterization and in vitro testing with


      dissolution often is not reliable enough by itself


      to make that call.  If it was, you would not be in


      this dilemma.


                So, if you really then look at it, what


      are we saying, is we have information generally


      that even if I give this drug intravenously or as a


      solution, and so forth, the variability is coming


      from the subjects, it is coming from physiology,


      which is inherently variable.  If I sleep on my


      lefthand side or righthand side, it will make a


      difference, I mean it literally happens.


                So, that is the measurement system, but


      then you are putting your product into that system


      and trying to see is there a difference of 20


      percent or not, and to meet that confidence


      interval criteria, you need 600 subjects or 300


      subjects, and so forth.


                Can we utilize the signs of design to say,


      to confirm, not necessarily to have a confidence


      interval, a confirmation that the new formulation


      actually is not contributing to that variability,




      is there sufficient science to do that or not.


                If it is, then it opens the door for


      saying that the bioequivalence assessment then


      could be tailored based on that understanding.


                DR. COONEY:  Before we go on to some other


      questions, I would like to see if your question,


      Paul, has been addressed.


                I think the question was--well, first,


      Lawrence is proposing that there be a


      pharmaceutical development report added to the


      information that is part of the application, and


      you are asking what will be the implications of


      providing that additional information and


      facilitating the next step, which is approval of a




                DR. FACKLER:  That is part of the


      question.  The other part was what would be in a


      pharmaceutical development report that isn't


      already part of an ANDA.  That is really what I am


      trying to understand, and, of course, then, what


      value would that provide.


                DR. COONEY:  Is there clarity to that




      question?  So, that is back to Lawrence, to Paul's


      question.  What would be in that pharmaceutical


      development report that is not already part of the




                DR. YU:  I thought that was a topic of our


      next advisory committee meeting.


                DR. HUSSAIN:  Let me put it this way.


      There is nothing there right now.  There is nothing


      there to even gauge the aspects of.  So, what we do


      is our decisions are made based on one batch test


      results and the biostudy.  That is what it is.


                DR. MORRIS:  Can I just weigh in?  I think


      part of this is that a lot of what would go in the


      development report is stuff that people are already


      doing, but doesn't just get included in a summary


      fashion, much like we have discussed earlier, that


      there is development studies you do, but you don't


      put together.


                That is what we were talking about


      yesterday, is that, as a reviewer, if you have to


      try to piece together a development rationale from


      data here and there, you end up with sort of a




      development rationale that the person filing really


      wouldn't want to be there displayed to the world,


      you know, sort of a Frankenstein development




                So, if the company does it, then, they can


      see the logic that you use.  Whether they agree or


      not is a different question.  So, in my sort of


      concept of this, which may be flawed, of course, if


      the company, let's say, had used Cynthia's


      dissolution method, because they said this is what


      has really mattered, and they got those curves to


      overlay, then, that is a big step forward to say


      that the variability that may come out of the BE


      studies are not due to our change.


                So, if you see the variability of the BE


      studies and you have demonstrated that it is not


      due to the lack of adherence to a design space, for


      lack of a better word, then, that has got to be as


      good as the innovator.  That is my concept.  This


      may be down the road, as Jerry said.


                DR. YU:  I want to make comments that when


      we say the pharmaceutical development information,




      I think I emphasize those information that is


      essential to know, not just for nice to know.


                We are looking into this, what additional


      information will help us in making decisions, and I


      think we are happy to share with you in the future,


      but at this point, we cannot say that for every


      single ANDA or for every single product, you need


      the pharmaceutical developed, because you have a


      prior knowledge, some of the information already


      there, so this need clarification when you are


      understanding what additional information is


      provided.  I think we need to discuss and work it




                DR. FACKLER:  I understand.  To Ken's


      point, you start over here and the bioequivalent


      product is over here, and sometimes you take a


      direct approach to it and sometimes you don't.  You


      are right, oftentimes it is over here and then you


      realize you need to be over here, and then finally,


      you get where you need to be.


                But I am not sure I understand the value


      or what it matters what path you took as long as




      you end up in the right place.  All this


      information does exist, of course, and the field


      inspectors have access to it, and we are just


      reluctant to expand the content of an ANDA in the


      fear that it will slow down an already overburdened


      review process.


                So, where the information is critical to


      understanding whether a product is pharmaceutically


      or therapeutically equivalent, of course, it ought


      to be submitted, but where it is not essential for


      that evaluation, I just question whether or not it


      ought to be added to the burden of the reviewers.


                DR. COONEY:  Art, then Marvin.


                DR. KIBBE:  Let's get back to what we are


      trying to determine, and that is whether or not a


      clinician who prescribes this medication for its


      effect has got a reasonable expectation of a


      therapeutically similar outcome when he uses the


      innovator or when he uses the generic.  That is


      where we are.


                If a product is inherently variable, as


      manufactured by the innovator, then, we ought to




      know that early on, and as Les correctly points


      out, if that was the case during development and


      prior to approval, it wouldn't make it on the


      market if it wasn't that that breath of variability


      was allowable for clinical outcome, because if the


      clinical outcomes wouldn't--there were times when


      there were failures and times when they were


      toxicities apart, never gets on the market. which


      means that we have already historically established


      large variability is okay, because we have that


      product on the market.


                Now, if I am a generic company, all I want


      to do is say that I am going to be no more than, or


      perhaps less variable, and I am going to get to the


      same therapeutic outcome.


                If I can test a replicate design that


      shows that my level of variability is lower than or


      equal to the variability of the innovator, and my


      means are on target, then, I can with reasonable


      assurance argue that my product used in the


      marketplace on patients is going to have the same


      efficacy and failure rate as the innovator.


                The second thing is we already have agreed


      that dissolution is a hammer when we need a


      surgical scalpel to figure out what is going on,




      and if you make a shift in a dissolution criteria


      and all of a sudden you can differentiate tablets


      from the same batch, but that batch used in people


      isn't differentiatable, then, you are making a


      differentiation which is of no value to anyone


      except if you want to go back and process improve.


                In fact, that is what it should be used


      for.  The companies ought to be investing time and


      energy in process improvement by looking for better


      differentiators for their own internal consistency,


      and perhaps they could narrow the variability if


      they found them.


                I think the justification for going to the


      study that you said that if they used the USP


      numbers, they would all pass, and going to your


      numbers, we have this high variability, but that


      high variability doesn't relate to clinical




                Now, I am coming on the market as a




      generic.  If I can establish that I am not more


      variable than they are, and my means are the same


      as they are in a biostudy, how much more


      information does the agency need?  I don't think it


      needs much more.


                DR. HUSSAIN:  Art, you are missing the


      point in the sense to demonstrate that your


      variability is acceptable, you actually have to do


      more now through a bioequivalence or replicate


      design, and so forth.


                What we are saying is in the sense, there


      are ways or there should be ways to sort of the


      justification that goes into a formulation that you


      move forward, could then become a basis to say you


      don't have to go through extraordinary means to say


      the variability is unacceptable.


                So, if we know a drug substance is highly


      variable, you mostly have that information that


      says you sort of at least definitely will when you


      approve the product, then, the signs of formulation


      design could provide you a basis for saying there


      is no reason your particle sizes, which are




      critical for your dissolution, your coating


      thickness, which are your release mechanism, are


      essentially being controlled, and so forth.


                So, why should a generic form then have to


      do a large study with replicate or with whatever?


      Isn't there an option available for something--


                DR. KIBBE:  So, what you are really


      talking about is a waiver of what we would say


      would be a standard replicate design to get around




                DR. HUSSAIN:  Exactly, so that is what we


      are suggesting.


                DR. KIBBE:  So, the company then would


      come with its own development data and show that a


      broad range of dissolution numbers are not highly


      variable or something.


                DR. HUSSAIN:  Yes, the way I would think


      about that is in a sense if it's a tablet, I will


      go to the basic mechanisms of what the dissolution


      will be affected, and here is my assessment of my


      particle size, here is my control strategy, here is


      the prior knowledge of similar dosage forms.  There




      is no apparent reason for this to be variable from


      that perspective.


                So, that becomes a basis for a decision


      criteria saying that why would we expose normal


      healthy subject volunteers, a large number of them,


      to simply get our numbers within the confidence


      interval criteria, which is somewhat arbitrary.


      That is the crux of this.


                DR. YU:  I don't know if I can clarify,


      the point we are trying to make is that if you can


      conduct bioequivalence study now to best pass the


      confidence interval, this is good enough.  I am not


      saying this is not good enough.  We are not asking


      additional information.


                The problem which we are facing is you


      will not have difficulty, it is not impossible if


      you have recruited 1,000 or 2,000 subjects, it is


      almost impossible to do by a current study, and


      this is scenario that pharmaceutical variability


      information may come into play and to help us out.


      That is what we are trying to convey.  Thank you.


                DR. MEYER:  I think part of my problem is




      that I believe what you are putting forth is a


      concept without any data, which obviously you can't


      have yet, because the concept hasn't even been


      implemented, it is just a concept.


                I think certainly from my perspective, if


      you have some ideas that might streamline the whole


      system, I would say go for it and then let's see


      the meat once the skeleton is exhumed, so to speak.


                That is the bottom line, but I think there


      are some other ideas in there that are perhaps


      easier for me to understand, characterize the


      reference listed drug and then presumably, if you


      have done that, FDA will take that into


      consideration to explain why you have confidence


      limits that aren't up to par perhaps.


                For example, a simple example, the RLD has


      an overage in it of 10 percent.  They claim that


      isn't released ever, so they just have it in there


      because their release mechanism doesn't allow for


      except 100 percent.


                You have some evidence that says well, in


      fact, it is released 110 percent sometimes, so the




      poor generic company is already 10 percent in the


      hole when it comes to AUC.  If that can be


      demonstrated in some reasonable scientific fashion,


      that ought to maybe taken into account.


                A better example maybe is with the osmotic


      pump.  We have done studies where you can harvest


      the ghost out of the feces, and sometimes it has 50


      percent drug in it, sometimes it has 10 percent,


      sometimes it has no drug in it. It seems to be a


      direct function of intestinal transit time.


                Well, if you are a generic trying to match


      without using an osmotic pump, you don't have a


      snowball's chance in hell of coming across and


      matching a product that sometimes is 50, sometimes


      is 100, sometimes is 10 percent.


                So, I think as long as you hit the means,


      and you bring that kind of data to FDA, they ought


      to have the latitude of saying yeah, we know that's


      a problem with the RLD, and we can therefore adjust


      our thinking when it comes to the generic.


                Obviously, that is going to take a fair


      amount of work, but I think that these things need




      to be thought of, as well as more statistically


      based ways of dealing with high variability.  That


      is kind of a short-term fix which ultimately once


      the statisticians get done fighting, then, the rest


      of us can agree, but the other is certainly a


      concept worth pursuing, I think.


                DR. SINGPURWALLA:  I would like to respond


      to that.


                DR. HUSSAIN:  If I may, there is an aspect


      what Marv just said in the sense a practice that


      all of us know exists is when you have variability,


      then, you pick and choose what your comparator is.


      I mean it bothers me in a sense to say that, you


      know, you can pick and choose what lot you will


      compare to, and so forth.


                Why do we have to sort of have those type


      of decisions where, you know, I think we can be


      better than that, so I think just to build on what


      Marv says, to say that I think we can really be


      confident in what we are doing, and not to feel a


      bit guilty that we are picking and choosing what we


      test, and so forth.


                DR. MEYER:  As you well know I am sure,


      there are a number of countries.  You do your


      dissolution on three lots of the RLD and then you




      pick the one in the middle, not the one that is


      closest to what your product happens to be.


                DR. YU:  I want to make comments about


      Marvin's comments.  Yes, in the case here, what you


      present, actually, those information is not in the


      original ANDA submission, but those information


      eventually is shared with us.


                So, go through many cycles, many, many


      months, or even several years to get us that


      information.  What we are seeing is that we think


      if those information, which you eventually shared,


      only a couple that go through the five or six or


      seven cycles, shared in first place will help us to


      make decisions, will help us to reduce cycles, will


      help us actually use the resource wisely.  That is


      what we are trying to say.  Thank you.


                DR. COONEY:  Nozer.


                DR. SINGPURWALLA:  General comments.


      First thing, Ajaz, don't use the word confidence




      limits for those two boundaries.  Call them control


      limits.  Confidence limits are completely




                The second thing is you are fighting, at


      least there is a lot of discussion because there is


      a lot of variability.  What you seem to have done


      is taken reactive approach, have said variability


      is there, what shall we do about it.


                Well, yesterday, you talked about 6 sigma


      in one of your slides.  Well, I think wherever you


      have these high variability issues, whether they be


      in industry or whether they be within your own


      system, I would encourage you to put into practice


      what you were preaching yesterday about 6 sigma.


                I would say, you know, has anybody thought


      about that, because 6 sigma came about in industry


      because there was a lot of variability, and they


      said how do we control it.  Well, you just don't


      control it by doing statistical methods.  You


      control it by proper management and proper


      procedures, and I would say that you should try to


      bring that into the arena.


                DR. COONEY:  Paul.


                DR. FACKLER:  The generic industry is just


      as interested in minimizing the number of 6 and 7




      cycle reviews on products.  Clearly, we have the


      same goal in mind.


                I guess what I would suggest is that for


      highly variable drugs, for instance, it would be


      useful for the agency to tell industry the kind of


      information that is generally lacking, but with 500


      applications a year, or 800, whatever the numbers


      might be these days, coming into the agency, I


      don't think it is wise to require this information


      on all of the applications.


                I would suggest maybe we clarify the


      additional information that is often being left out


      of submissions for highly variable products, and


      presumably, generic companies in the interest of


      having a minimum number of review cycles will


      submit it the first time rather than an iterative


      process to give you all the information that you


      need to make a fair decision.


                DR. COONEY:  Gary.


                DR. BUEHLER:  For the development reports


      in general, I thank you for not wanting to


      overburden us with additional information.  We do


      have a lot to look at.  If we do get additional


      information, we will look at it for sure.


                I know that we get some amount of this




      information sprinkled through the ANDAs and I would


      think Ajaz was a bit draconian when he said all we


      get is the batch record and whatever.  I mean there


      are explanations.  We do demand explanations when


      there aren't any deviations from what we normally


      see, that is in ANDAs and we do look at that.


                Lawrence and a group is working on a


      question-based review for the Office of Generic


      Drugs.  It is a very detailed project.  He is


      working with experienced reviewers in our office,


      and he is developing this in a very stepwise


      manner, both first by involving both the


      supervisors and reviewers in our own office, and


      then at a certain point we want to sort of unveil


      it to industry.


                We want to make sure that when we do bring




      this new review method and these new requirements


      or whatever you want to call them with respect to


      pharmaceutical development reports, the industry is


      very aware of what we want and why we want it, so


      that they will feel good about giving us this


      information, and like Lawrence said, it will


      hopefully reduce the number of cycles we have, it


      will not overburden the reviewers, but, in fact,


      reduce the burden on the reviewers, because they


      won't have to see the same applications four, five,


      or six times, and they will understand why we need


      this information.


                It is also a risk-based system, so that


      there are some applications that you won't have to


      provide this type of information, because there are


      some applications obviously that are easier than


      other applications, and the applications for


      complex dosage forms and unique dosage forms


      obviously, we are going to ask for more information


      than for the vary standard solid orals that are


      fairly easy to manufacture.


                But we are doing this over a two-year




      period and hopefully, sometime toward the end of


      this year, we will be able to begin to tell


      industry what we hope to expect in the future


      applications and industry will be comfortable with




                DR. MORRIS:  I just have a quick question


      for Gary.  I am assuming that development reports,


      as you say, depending upon the complexity of the


      dosage form, I mean they can be relatively brief if


      it's a very simplistic or simple dosage form, so I


      am not so sure that it's the burden if the payback


      is fewer review cycles or less clinical studies.


      Clinical studies are a lot more expensive than


      writing a development report and doing a few more


      development studies.


                Is that more or less the case, Gary?


                DR. BUEHLER:  I am not sure it is going to


      be able to be submitted in lieu of a study or




                DR. MORRIS:  No, I meant the extensiveness


      of a development report.


                DR. BUEHLER:  Some development reports




      will say we wanted to develop a bioequivalent


      formulation, and, you know, here it is, and it


      could be a page or two.  I mean clearly, it won't


      be very long for a generic, because the goal of a


      generic is pretty evident, but other development


      reports will be more extensive, so yeah, you know.


                DR. COONEY:  It sounds like there is a


      need for clarity on what will be requested and


      expected, and also for clarity on what the


      implications of that will be.  It sounds like that


      will be forthcoming.


                Ajax, what I would like to do is move on


      to the next presentation.


                DR. HUSSAIN:  Just go back to the original


      intent.  Our initial thoughts that we wanted to get


      the discussion started, so we never intended this


      to make a proposal, so these are initial thoughts


      and we are moving forward with this.


                If industry wants to be proactive, they


      had better start thinking about it and how they can


      use this opportunity instead of asking us what do


      we want.  I think it is equally burdensome on




      industry to think about how to develop the products


      for the intended use, and make the case, and grab


      that opportunity.


                If not, the system as it stays, we are


      perfectly happy with it.


                DR. COONEY:  So, there is an opportunity


      here for dialogue and there is no doubt from the


      last 45 minutes that there will be dialogue.


                I would like to ask Robert Lionberger to


      proceed with the next presentation.


            Using Product Development Information to Support


                Establishing Therapeutic Equivalence of


                            Topical Products


                DR. LIONBERGER:  Today, I am going to be


      discussing how to apply the concepts of


      pharmaceutical equivalence to topical dosage forms


      and look at how this is related to quality by




                Here, I am going to focus on topical


      dosage forms that are in the local delivery, so not


      products such as transdermal products that are


      trying to deliver drugs systemically.


                In the Office of Generic Drugs, as you


      have heard several times before this morning, our


      mission is to provide therapeutically equivalent




      products to the public.  When someone uses a


      generic drug, they should expect the same clinical


      effect and safety profile as the branded reference




                Just to summarize some things that Ajaz


      talked about in his introductions, the preface to


      the Orange Book explains how we do that.  Products


      must be pharmaceutically equivalent and


      bioequivalent.  But I want to dig a little bit


      deeper into this and ask why do we actually require


      both, why isn't bioequivalence by itself enough to


      determine that the products are the same.


                One aspect of that is that consumers have


      some expectation about product behavior.  If the


      reference product is a capsule, you don't want to


      replace that with a solution.  So, there is some


      user experience and expectation.


                So, pharmaceutical equivalence


      encapsulates concepts related to like the user




      interface of the product, but then there is another


      aspect to it, and I have tried to express it here,


      is that pharmaceutical equivalence supports the


      determination of therapeutic equivalence based on


      bioequivalence study.


                We don't say that just because two


      products pass a bioequivalence study, they are


      therapeutically equivalent products.  An example


      might be an oral solution and a tablet.  There can


      be many products for which those two dosage forms


      would be bioequivalent, but we wouldn't say that


      they are therapeutically equivalent products.


                One aspect of that is that our current


      determination of bioequivalence is really very


      strongly based on in vivo testing.  So, again,


      there are limitations to testing.  We test these


      products in a small population  and then we


      extrapolate that conclusion to all people who are


      going to use the products from all batches in the




                So, to sort of back up that extrapolation,


      there is some other information.  Right now that's




      the pharmaceutical equivalence between the products


      that supports that.


                In the occasions when we do equivalence


      studies in patients, there are other differences.


      Sometimes the clinical endpoints aren't very


      sensitive to small differences, bringing in


      examples from topical products, you can imagine


      there are cases where, say, a cream and an ointment


      formulation might have the same therapeutic effect,


      but they wouldn't be considered pharmaceutically


      equivalent products or therapeutically equivalent


      products even though the clinical endpoint study


      might show equivalent efficacy.


                Again, from the sort of pharmacokinetic


      studies for one of the challenges that is often


      made to some of our bioequivalent studies for


      topical products is since the skin is a barrier,


      you say, well, healthy subjects have healthy skin


      barriers.  There is a question.  Sometimes people


      will claim in patients, the skin might be diseased


      or damaged, so that is a common concern.  There is


      a common challenge to some of our bioequivalence




      determinations here.


                So, inside of the pharmaceutical


      equivalence concept, there is some idea of other


      things we want to know about the products to sort


      of generalize this idea of equivalence.


                If you think about this and want to sort


      of tie this to quality by design, one way that


      might be useful for you to think about this is that


      our current definition of pharmaceutical


      equivalence might be considered a first step toward


      a quality by design.


                If you were going to design equivalent


      products, the first things you would start with


      were some of the concepts that are in our current


      definition of pharmaceutical equivalence.  You


      would want to have the same active ingredient.  You


      would want to have the same strength, the same


      dosage form.


                So, if we look at sort of a different way


      of looking at our paradigm, maybe instead of a


      regulatory framework, a more scientific framework,


      what we are doing when we review a generic product,




      is we are looking to see is the product designed to


      be equivalent, and then does it demonstrate


      bioequivalence in an in vivo study.


                So, you can see sort of this combination


      sort of parallels our current sort of regulatory


      framework of pharmaceutical equivalence and


      bioequivalence leading to a determination of


      therapeutic equivalence, where we might say that on


      sort of a scientific level, what we might want to


      be doing in the future might be to say look at the


      quality by design, look at the generic product that


      is designed to be equivalent to a reference


      product, and then based on this design, choose the


      appropriate either in vitro or in vivo


      bioequivalence testing for this product to complete


      the determination of therapeutic equivalence.


                So, I want to bring this sort of


      conceptual framework and bring it into this sort of


      particular example for topical products.  Sort of


      to motivate that, I just want to outline some of


      the complex issues that we deal with that are


      related to pharmaceutical equivalence for topical






                Again, we have a lot more experience with


      immediate release, oral dosage forms in effective


      excipients, what excipients you can change, what


      excipients you can't change.  For topical products,


      a lot of times the excipients may or may not affect


      the barrier properties of the skin and drug




                We don't have as much experience about


      that, so a lot of times we are worried about what


      differences in formulation are appropriate for


      comparing a test in a reference product - is a


      change in solvent appropriate, what if the base of


      the formulation in ointment or cream has changed


      from being hydrophilic to lipophilic, how much


      water content should there be in the product.  You


      might affect evaporation, the feel of the product.


                A lot of these sort of differences in


      formulation get wrapped up into the question of are


      two products the same dosage form.  I will talk a


      little bit more in detail about that in the rest of


      the products.


                We also have questions, when we don't have


      good bioequivalence methods for use for topical


      products, what indications should be used for the




      clinical equivalence studies.  Perhaps the product


      has multiple indications, which one is the most


      appropriate one to use.


                These are the kinds of issues that we deal


      with in generic topical products.  Some of the


      implications of these for the ANDA sponsors are


      that the approval times for these products can be


      longer.  If there are these issues that we don't


      have a good understanding internally, we have to


      schedule meetings with the appropriate people, have


      to have internal discussions.


                When the sort of standards aren't clear,


      this is an opportunity for the reference listed


      drug sponsors to challenge correspondence to OGD or


      through the citizen petition process that we have


      to address the scientific issues there that aren't


      sort of clearly defined.


                A lot of times, at the end of these


      discussions, we will end up going back to the




      sponsors and asking them for more information to


      help us resolve these issues, and then usually that


      is done through deficiency letters to them, and it


      ends up with sort of multiple review cycles.


                So, as we heard in Lawrence's talk, there


      is the question of more product development


      information in the ANDA itself may help OGD deal


      with these issues more efficiently.


                This is sort of very similar to some of


      the things that Lawrence talked about, that there


      are harmonization efforts underway that describe a


      product development report, but I think it is clear


      that these are mainly aimed at new drug


      applications, so it is not sort of obvious or clear


      how these should apply to ANDA sponsors.


                I think the theme of this talk to see this


      as an opportunity, these development reports, as an


      opportunity to provide information that will help


      the agency set rational specifications for products


      that are complex, for immediate release oral dosage


      forms we have various standard systems set in


      place, but for topical products, where we have less




      experience, the more information that is provided


      about, say, why was an excipient changed, and why


      do you know that it is not going to have any effect


      could be very helpful to us in making efficient




                Again, the product development reports are


      the place in the application to emphasize the


      quality by design, that the product is designed to


      be equivalent.  That will help us set the right


      requirements for the bioequivalence testing for


      particular products.


                This is just a few examples of what some


      of these harmonization documents say about a


      pharmaceutical development report.


                In this case, again, the key part here


      might be to establish that the dosage form and the


      formulation are appropriate for the purpose


      specified in the application, or in the Q8


      document, it talks about an opportunity to present


      the knowledge gained through the application of


      scientific approaches.


                Here, it is talking specifically about




      sort of formulation and development for the topical


      products, that there is information that the


      company that is developing the generic product


      knows about why they made certain choices in the


      formulation.  It would be very helpful to us in


      deciding that that is acceptable, where the agency


      itself has less experience with particular dosage




                I have emphasized this concept of quality


      by design or, in the case of the generic products,


      quality by design means you are designing the


      product to be equivalent to the reference product.


                So, I want to try to be a little bit more


      specific about what that means.  There are two


      cases.  One, the mechanism of release.  Clearly,


      the mechanism of release between a generic product


      and the reference product can be different, but the


      intent of those different mechanisms ought to be to


      produce the same rate and extent of absorption.


      This is the bioequivalence criteria.


                Again, we also recognize that depending on


      the particular product, that the release rate from




      the product might not be the rate controlling step


      at absorption.  So, the determination of how close


      release rates might have to match would depend on


      the absorption process involved and what is the


      rate-limiting step in the absorption process.


                Again, between generic products and


      reference products, the excipients can be


      different.  Again, it is a good thing to understand


      the differences between the excipients.


                The IIG limits are a starting point.  They


      tell you that this excipient has been used in this


      dosage form up to a certain amount, and that really


      addresses, specifically in the case of topical


      products, safety-related exposures, so you know


      that level of exposure.


                The thing that complicates the topical


      products is when you change the excipients, the


      real question that we often deal with is do the


      changes in the excipients to the products affect


      the permeation of the drug through the skin. I


      think that is the sort of challenging question


      there for the topical products that we occasionally




      have to deal with.


                Again, as I said, the purpose of quality


      by design is to design the equivalent product.  I


      want to just give sort of three sort of examples of


      this process here.


                The first is talking about Q1 and Q2


      equivalent products for topical products, and then


      look at what happens when you make changes to the


      formulation, they become Q1 and Q2 different, and


      then this leads into the discussion of issues


      related to the dosage form classification and how


      product development information might help us make


      a better decision or more scientific based


      decisions on dosage form classifications.


                First, I want to start off with the


      definition of Q1, Q2, Q3.  So, products that are Q1


      have the same components, so both the generic and


      the reference product would have the same




                If products are Q2, they would have the


      same components, but they would also have the same


      amount of each ingredient.


                The Q3 concept is same components, same


      concentration, but here I am saying same


      arrangement of material or microstructure, and this




      is particularly important for topical products that


      are semisolid dosage forms, so non-equilibrium


      dosage forms, where you might have, say, an


      emulsion with exactly the same components, exactly


      the same concentrations, but say, for example, the


      droplet size might depend on how you have


      manufactured that product.


                So, there is potential differences for


      semisolid dosage forms depending on how they are


      produced even if overall the composition is exactly


      the same.


                A contrary example would be a solution.


      If a solution is Q1 and Q2, because the solution is


      at thermodynamic equilibrium, you would be able to


      say we know that this product has exactly the same


      arrangement of material in the product.


                The importance of the Q3 concept is when


      you know that the products have the same


      arrangement of material, you know that they are




      going to be bioequivalent, there is no question


      about that.  An example of that is again a solution


      where you know that the products are in


      thermodynamic equilibrium if the compositions are


      the same, the structure and arrangement of the


      material is the same.


                Unfortunately, for most topical semisolid


      dosage forms, they are not necessarily equilibrium


      arrangements of material, and so a direct


      measurement of Q3 level equivalence is challenging.


                So, if we have the topical products where


      Q1 and Q2 are identical, again, the only potential


      differences are differences in this Q3 parameter,


      which can come from differences in manufacturing


      processes, because they are not going to be


      manufactured by exactly the same process.


                We know for particular semisolid dosage


      forms, such as emulsions, that rheology and in


      vitro release rates can be very sensitive


      measurements of microstructure and are related to


      product performances.


                So, the sort of idea that sort of




      advancing here from a scientific point of view is


      when the products are Q1 and Q2, that in vitro


      tests should be equivalent to ensure bioequivalence


      of the two products, because again here, the issues


      are detecting differences due to differences in


      manufacturing processes, and the argument would be


      that in vitro tests are the best evaluation method


      to detect whether any differences in manufacturing


      process have significant differences in the product


      formulation or performance.


                Now, things get more complex when a


      generic product and a reference product have


      different compositions, and this connects with the


      dosage form classification, and these differences


      occasionally could be barriers to generic




                A generic company might want to formulate


      products that are Q1/Q2 different because the


      innovator has formulation patterns, so there might


      be either legal reasons or perhaps manufacturing


      process reasons why you might want to formulate a


      product that is not exactly identical in




      composition to the reference product.


                One of the products, again, one of the


      members of the Committee mentioned the sort of


      economic effect of uncertainty on product


      development if you don't know what dosage form the


      product is going to be classified as.  That adds


      cost to the development process because of


      uncertainty of what is going to happen to the




                In particular, if we think about methods


      by which we would classify the dosage form of


      topical products, here, I have generated a list of


      four possible ways that you could approach this.


                One is we would just use whatever the


      sponsor says their product is as long as it is


      consistent with some of the traditional definitions


      that are available in various sources, and we will


      look in sort of detail at some of those traditional




                You might say, well, the generic product


      is the same dosage form if it feels the same to me,


      so I will just try it out and see if it is the




      same, if it passes.  You know, if the look and feel


      of the products are the same. That is getting to


      the aspect of pharmaceutical equivalence, a sort of


      patient experience rather than sort of scientific


      issues related to product performance.


                Then, I am going to sort of discuss recent


      work that the FDA group led by Cindy has done on


      looking at a whole bunch of products and coming up


      with a quantitative decision tree to classify


      topical products.


                Then, sort of the fourth aspect of that is


      looking at whether or not issues about dosage form


      classification for complex issues would be


      something that you would want to include in a


      product development report, so justifying the


      formulation development as being the same as the


      reference product.  That sort of might be a more


      scientific way to look at these issues.


                First, if we look a some of the


      traditional definitions.  Here, I will just focus


      on the difference between a cream and an ointment.


      One source is the CDER's data standards




      definitions.  These are sort of similar to USP


      definitions of these products.


                The cream is a semisolid dosage form


      containing one or more drug substances dissolved in


      a suitable base, and then it says more recently the


      term has been restricted to products consisting of


      oil-in-water emulsions.  That is obvious what a


      company should do - does a cream have to be an


      oil-in-water emulsion or not.


                Then, it talks about products that are


      cosmetically and aesthetically acceptable, is part


      of the definition of the cream, so that is not very


      quantitative. It is hard to say is this product


      aesthetically acceptable. That is really opinion




                An ointment is a semisolid preparation


      intended for external application.  It seems to me


      that a cream could be a semisolid preparation and


      fit under the ointment definition.  So, it doesn't


      seem that those two definitions are really




                In another FDA guidance, this is the SUPAC




      semisolid guidance.  It has a glossary with


      definitions of dosage forms, but these aren't the


      same as the previous ones.  A cream is a semisolid


      emulsion, and an ointment is an unctuous semisolid


      and typically based.  So, typically based is not


      sort of a definition, it doesn't have to be based


      on petrolatum.


                This definition talks about an ointment


      being one phase, and not having sufficient water.


      Again, the USP definition is sort of similar to the


      one in the CDER data standards, but it is not


      word-for-word identical, and talks about four


      different classes of ointments.


                So, again, the problem with the


      traditional classifications is they are not really


      consistent, and not very quantitative.  So, a lot


      of the sort of decision process would depend on


      what your opinion was of a particular product, and


      they might be overlapping, like you might be able


      to call--under a particular definition of a


      particular product, you might be able to call it a


      cream or an ointment.


                So, the result of this Topical Working


      Group led by Cindy has been presented to previous


      advisory committee meetings, and they recently




      published a paper outlining this classification




                What they did was they surveyed existing


      products and devised a classification scheme, and I


      just included the classification scheme here just


      for reference in the presentation.


                These are just some slides from their


      previous presentations to give you the general idea


      of what they did. They measured particular aspects


      of products, say, creams and ointments, they


      measured viscosity.


                They looked at the loss on drying, and


      then based on these products that were either on


      the market or manufactured for them, they came up


      with a classification scheme that sort of put the


      products in the right category based on existing




                The real advantage of this is it is


      quantitative. If you take a product and you go




      through and you measure the things outlined in


      their decision tree, you will always end up in the


      same place, and it will always be consistent.


                In addition to that, this is a very data


      driven approach.  They looked at the products and


      then drew the lines.  It wasn't said here is sort


      of a mechanistic definition of what a cream or an


      ointment should do.


                So, the question is, could this be overly


      restrictive.  If you follow this classification


      scheme, you would be restricting products to


      essentially what has been done before, and then


      there is a question.


                They didn't survey every product that is


      on the market now, so there is a question, if a


      reference listed drug falls into a different part


      of this classification scheme, then, it's labeling.


      So, it might be labeled as a cream, but by the


      definition, it would be classified as an ointment.


      What should a sponsor do in that case?


                So, the final sort of approach to dosage


      form classification might be to look at a more




      scientific view of the formulation design.  I just


      want to point out that sort of the legal aspects,


      referring to topical use, sort of point toward this


      here from the CFR.


                It talks about inactive ingredient changes


      for topical products.  It says again that


      abbreviated applications can use different


      ingredients if they identify and characterize the


      difference and provide information demonstrating


      that the differences do not affect the safety or


      efficacy of the proposed drug products.


                So, a current way of looking at is a


      change in formulation acceptable, you should check


      the new excipients against the IIG.  As I said


      before, this looks at the safety of the individual




                We also really consider that passing


      bioequivalence tests are evidence that the


      formulation change is acceptable.  That is one


      strong piece of evidence against that.  But again


      the product development report is an opportunity


      for sponsors to characterize the differences.


                Again, this could be important, you know,


      if you are formulating a product and you are on the


      boundaries of these, we have this empirical




      decision tree, what happens if you are on the


      boundaries, how do you explain that this product


      should be considered the same as the reference


      product, you know, from a scientific point of view


      rather than empirical classification scheme, of if


      someone says, well, no, your product is not really


      an ointment because it doesn't meet a particular


      published definition.


                So, again, the product development report


      is the opportunity for a sponsor to characterize


      the difference that is sort of requested in the




                Again, also, in the statutes, they list


      reasons to reject ANDAs, and they talk about drug


      products for topical administration where there is


      a change in lipophilic properties of the vehicle.


                Again, in this case, a product development


      report is an opportunity for sponsors to explain


      why the changes, Q1 and Q2 differences are




      appropriate for this particular product.


                If this issue comes up, a lot of times we


      will actually have to go from the review process,


      you will have to go back and ask sponsors for more


      information about these particular issues.  So, the


      development report is sort of an up-front way to


      explain the reasons for doing that.


                Just to sort of conclude the discussion


      here, the first concept is the importance of Q1,


      Q2, Q3 classification to identify appropriate


      bioequivalent studies for the level of difference


      in the product design.  If a few products have


      exactly the same active and inactive ingredients,


      you might want to request different in vivo


      bioequivalence studies than for a product where


      there has been a change in inactive ingredient that


      may affect the absorption of the drug product.


                So, here again, we are looking at the


      second concept, we are looking at the evolution of


      the concept of pharmaceutical equivalence where we


      have these traditional dosage form definitions,


      maybe now backed up by empirical decision trees,




      but in the future, looking for a quality-by-design


      aspect where the determination of whether a product


      should be considered equivalent would depend on the


      mechanistic understanding and the formulation


      design rather than some traditional definitions,


      and that the ideal state would be that this


      understanding would reduce the need or allow us to


      set the appropriate in vitro testing for a


      particular product, and also to expand sort of the


      formulation design space beyond past experience.


                If you want to formulate a product that


      goes beyond, say, an empirical dosage form


      classification, this is the sort of way that you


      would approach it, by providing the scientific


      information to show that the formulation you have


      chosen gives equivalent performance in the key


      attributes as the reference product.


                With that, I will conclude my




                DR. COONEY:  Thank you.  I believe the


      purpose of your presentation today is to bring us


      up to date on the current thinking where you are




      and where you are going as opposed to requesting a


      specific action on our part, is that correct?


                DR. LIONBERGER:  Yes, that's right.


                DR. COONEY:  I would like to invite


      questions and comments from the Committee.  Yes,




                DR. SELASSIE:  This is a very general


      question.  With all these product development


      reports that you get, obviously, there is going to


      be a lot of information that is extraneous and


      won't be useful for that particular application,


      but will you all retain this information like in a


      database, so that it could have use down the line?


                DR. LIONBERGER:  I don't know if we would


      retain it in a database, but I would say that like


      as Lawrence said, we are looking at our review


      process, and in that, had the opportunity to read


      several product development reports.


                I find that they are a very useful way to


      get an overview of what is going on with a


      particular product.  You know, an hour of reading


      the development report, it seems like a very good




      way to start the review of the application in more


      detail, so I think it can be very valuable.


                We don't have much experience with using


      them yet, so in that sense, it could be valuable,


      but we don't know how we would use that information


      or store that information in the future.


                DR. HUSSAIN:  If you are suggesting that


      there is a need to capture and create databases, I


      think we do want to move in that direction, and we


      tried to do that.  Currently, our systems does


      capture some of the key aspects. The inactive


      ingredient guide is a process that we capture every


      inactive ingredient that comes, but developing a


      formal knowledge base would really be helpful, and


      I think we have been thinking about it.


                I tried to do that with immediate release


      dosage forms and actually did some modeling with


      that data that we have, and so forth, so we will


      look into that.


                DR. COONEY:  Are there any other questions


      from the Committee?  Marv.


                DR. MEYER:  Just a quick comment.  You




      have my sympathies.  I thought it was difficult to


      determine how to do the BE studies on topicals, and


      now I have been reinforced, you don't even know


      what slot to put them in the Orange Book if they


      are bioequivalent, so you have a big job ahead of




                DR. COONEY:  If there are no further


      comments, Robert, thank you very much.


                There have been no requests for


      participation in the open public hearing at 1


      o'clock, so we will proceed with the continuation


      of the discussion on quality by design precisely at


      1 o'clock when we come back from lunch.


                We will begin that by a presentation of a


      summary of the plan by Ajaz, and then we will


      continue the discussion that we began earlier this




                So, enjoy lunch and we will see you back


      at 1 o'clock.


                [Whereupon, at 11:59 a.m., the proceedings


      were recessed, to be resumed at 1:00 p.m.]




                A F T E R N O O N  P R O C E E D I N G S


                                                       [1:00 p.m.]


                DR. COONEY:  I would like to welcome


      everyone back from lunch.


                We will proceed with the afternoon


      schedule.  The first topic this afternoon will be


      Ajaz Hussain, who will provide a summary


      description of the plan to go forward.


                  Quality-by-Design and Pharmaceutical


                        Equivalence (Continued)


                            Summary of Plan


                DR. HUSSAIN:  I am going to go back to the


      slides I used in the morning instead of the ones I


      had for this session.  That was based on the


      discussion that occurred.


                Just on reflection, I just want to make a


      couple of points.  Yesterday, in a sense, as part


      of the tactical plan to start our journey towards


      the desired state, in a sense what we have done at


      this meeting is to take a look back last 10 years


      or so to see how our policies have evolved and how


      they could evolve with two tools that we have




      introduced, the PAT guidance and ICH Q8.


                Tom Layloff reminded me that in many ways,


      some of the topics we have discussed, we have been


      discussing for 30 years, and we keep discussing


      those topics again and again, and the difference


      that we have tried, at the training session that


      some of you attended, I am clearly cognizant of the


      fact that we are discussing topics that we have


      been discussing for 30 years, and the quote I had


      was the thing that if you tried to approach the


      problem with the same tools and the same approach


      again, we are bound to find the same solutions, so


      we need something new.


                What is new at the issue of this problem


      is the science of formulation design, of science of


      product design. The key aspect, much of that has


      always been considered as an art, and as the


      complexity of products is increasing, that art will


      not be sufficient to really achieve the performance


      we are trying to achieve.


                So, it is a reflection back of saying all


      right, 30 years of pharmaceutical sciences in




      particular pharmaceutics, industry, pharmacy, and


      so forth, what have we learned and what we need to


      learn more to do things differently.


                In some sense, that is the heart of the


      debate.  I also sort of mentioned to you, and this


      is my original starting point in the thought


      process was that you really need at FDA more people


      with that background to really make that happen.  I


      changed my thoughts over the last several years.


                What we have at FDA is scientists from


      many, many different disciplines who sort of work


      together.  The reason I changed my mind was I think


      looking at some of the practices and formulation